Enzyme Catalysis in Organic Synthesis, Third Edition (2012) PDF

Edited by
Karlheinz Drauz, Harald Gröger,

and Oliver May
Enzyme Catalysis in Organic

Synthesis
Volume 1
Related Titles
Whittall, J., Sutton, P.

Hou, C. T., Shaw, J.-F.
Practical Methods for Biocatalysis
and Biotransformations Biocatalysis and Bioenergy
Hardcover
Hardcover
ISBN: 978-0-470-13404-7
ISBN: 978-0-470-51927-1
Gotor, V., Alfonso, I., García-Urdiales, E. (eds.)

Tao, J., Lin, G.-Q., Liese, A.
Asymmetric Organic Synthesis with
Biocatalysis for the Pharmaceutical
Enzymes
Industry
2008
Discovery, Development, and
Hardcover
Manufacturing ISBN: 978-3-527-31825-4
Hardcover
ISBN: 978-0-470-82314-9 Aehle, W. (ed.)
Crabtree, R. H. (ed.) Enzymes in Industry

Production and Applications
Handbook of Green Chemistry -
2007
Green Catalysis Hardcover
2009 ISBN: 978-3-527-31689-2
Hardcover
ISBN: 978-3-527-31577-2 Roberts, S. M.
Fessner, W.-D., Anthonsen, T. (eds.)

Catalysts for Fine Chemical Synthesis
Volumes 1-5. Set
Modern Biocatalysis 2007
Stereoselective and Environmentally Hardcover
Friendly Reactions ISBN: 978-0-470-51605-8
2009
Hardcover Enders, D., Jaeger, K.-E. (eds.)
SBN: 978-3-527-32071-4
Asymmetric Synthesis with Chemical
and Biological Methods
Garcia-Junceda, E. (ed.)
2007
Multi-Step Enzyme Catalysis Hardcover
ISBN: 978-3-527-31473-7
Biotransformations and
Chemoenzymatic Synthesis
2008
Hardcover
ISBN: 978-3-527-31921-3
Edited by Karlheinz Drauz,
Harald Gröger, and Oliver May
Enzyme Catalysis in Organic Synthesis
Volume 1
With a Foreword by Herbert Waldmann
Third, Completely Revised and Enlarged Edition

The Editors All books published by Wiley-VCH are carefully
produced. Nevertheless, authors, editors, and pub-
Prof. Dr. Karlheinz Drauz lisher do not warrant the information contained in
Evonik Industries AG these books, including this book, to be free of errors.
Innovation Management Readers are advised to keep in mind that statements,
Rodenbacher Chaussee 4 data, illustrations, procedural details or other items
63457 Hanau may inadvertently be inaccurate.
Germany
Library of Congress Card No.: applied for
Prof. Dr. Harald Gröger British Library Cataloguing-in-Publication Data

Bielefeld University A catalogue record for this book is available from the
Organic Chemistry I British Library.
P.O. box 100131
33501 Bielefeld Bibliographic information published by
Germany the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publica-
Dr. Oliver May tion in the Deutsche Nationalbibliografie; detailed
DSM Innovative Synthesis B.V. bibliographic data are available on the Internet at
Mijnweg 2 http://dnb.d-nb.de.
6167 AC Geleen
The Netherlands # 2012 Wiley-VCH Verlag & Co. KGaA,
Boschstr. 12, 69469 Weinheim, Germany
All rights reserved (including those of translation

into other languages). No part of this book may be
reproduced in any form – by photoprinting, micro-
film, or any other means – nor transmitted or trans-
lated into a machine language without written
permission from the publishers. Registered names,
trademarks, etc. used in this book, even when not
specifically marked as such, are not to be considered
unprotected by law.
Cover Design Formgeber, Eppelheim

Typesetting Thomson Digital, Noida, India
Printing and Binding betz-druck GmbH, Darmstadt
Printed in the Federal Republic of Germany

Printed on acid-free paper
Print ISBN: 978-3-527-32547-4

oBook ISBN: 978-3-527-63986-1
VII
Contents
Foreword V
About the Editors XLI
Preface XLIII
List of Contributors XLV
Contents to Volume 1
Part I Principles of Enzyme Catalysis 1
1 Introduction – Principles and Historical Landmarks of Enzyme

Catalysis in Organic Synthesis 3
Harald Gröger and Yasuhisa Asano
1.1 General Remarks 3
1.2 Potential of Enzymes as Catalysts in Organic Synthesis: Enzyme
Reactions Overview 4
1.2.1 Enzyme Catalysts: Three-Dimensional Structure and General
Properties 4
1.2.2 Overview of Enzyme Classes (EC Numbers) and Related Reactions 5
1.2.3 Overview of Coenzymes and Cofactors and Applications in
Organic Synthesis 11
1.2.4 Factors Affecting Enzymatic Reactions 13
1.2.5 Why Use Enzymes in Organic Synthesis? Factors Affecting
Enzymatic Reactions, Advantages and Drawbacks 14
1.3 The Early Steps: From Fermentation to Biotransformations
Using Wild-Type Whole Cells 17
1.3.1 Historical Development of Fermentation and First Microbial
Transformations 17
1.3.2 Development of Practical Synthesis of Chemicals via
Transformations Using Wild-Type Whole Cells in Non-Immobilized
Form 19
1.3.3 Development of Practical Synthesis of Chemicals via Transformations
Using Wild-Type Whole Cells in Immobilized Form 20
VIII Contents
1.4 Chemical Processes with Isolated Enzymes: The Impact

of Process Engineering 22
1.4.1 Historical Development of Transformations with
Isolated Enzymes 22
Transformations Using Isolated Enzymes in Immobilized
(Solid-Supported) Form 23
Transformations Using Isolated Enzymes in ‘‘Free’’ Form 24
1.5 Towards Tailor-Made Enzymes: Principles in Enzyme Screening
and Protein Engineering Methodologies 26
1.5.1 Tools for Enzyme Discovery 26
1.5.2 Protein Engineering Methodologies 28
1.6 ‘‘Hybridization’’ of Enzyme Catalysis with Organic Syntheses:
New Opportunities for Industrial Production of Chemicals
and Drugs 31
1.6.1 Applications of Tailor-Made Recombinant Whole-Cell
Catalysts in Organic Synthesis 32
1.6.2 Novel Retrosynthetic Approaches in Drug Synthesis: From Enzyme
Catalysis in Chemoenzymatic Multistep Processes towards New Drug
Production Pathways in Industry 34
1.6.3 Recent Aspects of Applications of Enzymes in Organic Synthesis 38
1.7 Summary and Outlook 39
References 39
2 Concepts in Biocatalysis 43
Eduardo García-Urdiales, Iván Lavandera, and Vicente Gotor
2.1 Introduction 43
2.2 Types of Biocatalytic Processes 45
2.2.1 Dealing with Racemates: Kinetic Resolutions (KRs) 46
2.2.2 Overcoming the ee Limitation of KRs: Parallel Kinetic
Resolutions (PKRs) 48
2.2.3 Overcoming the Yield Limitation of KRs 50
2.2.3.1 Dealing with Prochiral or Meso Compounds:
Desymmetrizations 50
2.2.3.2 (Cyclic) Deracemizations (CycDs) 52
2.2.3.3 Enantioconvergent Processes (ECPs) 54
2.2.3.4 Dynamic Kinetic Resolutions (DKRs) 54
2.2.3.5 Dynamic Kinetic Asymmetric Transformations (DYKATs):
Types I and II 56
2.2.4 Dealing with Diastereomers: DYKATs Types III and IV 58
2.2.5 Making it at Once: Cascade or Domino Processes 60
2.2.6 Novel Concepts 61
References 63
Contents IX
3 Discovery of Enzymes 67
Wolfgang Aehle and Juergen Eck
3.1 Introduction 67
3.1.1 Historical Overview 67
3.1.2 The ‘‘Ideal Enzyme’’ Concept 70
3.2 Exploiting Functional Sequence Space: Resources and Screening
Strategies 72
3.2.1 Resources for Enzyme Discovery 72
3.2.2 Screening Strategies 73
3.3 Enzyme Discovery Techniques 74
3.3.1 Gene Mining 74
3.3.2 Sequence Homology-Based Screening 75
3.3.3 Expression of Active Enzymes for Activity-Based Screening 76
3.3.4 Activity-Based Screening 78
3.4 Challenges in Enzyme Screening 81
3.5 Concluding Remarks 82
References 83
4 Rational Design of Enzymes 89

Jürgen Pleiss
4.1 Enzyme Design: Learn from Nature 89
4.2 Today: Find and Improve Enzymes 90
4.2.1 Data Mining: Find Appropriate Biocatalysts in Databases 90
4.2.2 Rational Evolution: Improve Efficiency of Directed Evolution 93
4.2.3 Molecular Modeling and Protein Design of Stability, Specificity,
and Selectivity 94
4.2.3.1 Prediction of Enzyme Structure 94
4.2.3.2 Prediction of Protein Stability and Solubility 95
4.2.3.3 Docking 96
4.2.3.4 Molecular Dynamics Simulations 97
4.2.3.5 Quantum Chemical Methods 99
4.2.4 Role of Solvent 100
4.2.4.1 Hydration of Enzymes 100
4.2.4.2 Enzymes in Organic Solvents 100
4.2.4.3 Solvent-Induced Conformational Changes 101
4.3 De Novo Design of Stable and Functional Proteins 102
4.4 Challenges and Outlook 104
4.4.1 Force Field, System Size, and Simulation Time 104
4.4.2 Enzymes are Nanomachines 105
4.4.3 Outlook 106
References 107
5 Directed Evolution of Enzymes 119

Manfred T. Reetz
5.1 Purpose of Directed Evolution 119
X Contents
5.2 Short History of Directed Evolution 119

5.3 Basic Principles and Challenges 121
5.4 Gene Mutagenesis Methods 122
5.4.1 Whole Gene Methods 122
5.4.2 Saturation Mutagenesis 128
5.4.3 Recombinant Methods 135
5.4.4 Other Methods 139
5.5 Strategies for Applying Gene Mutagenesis Methods 140
5.5.1 General Guidelines 140
5.5.2 Rare but Helpful Comparative Studies 143
5.5.3 Computational Guides 149
5.6 Screening Versus Selection 152
5.7 Engineering Enzyme Stability 156
5.8 Engineering Enzyme Stereoselectivity 160
5.8.1 General Remarks 160
5.8.2 Hydrolases 161
5.8.2.1 Nitrilase from an Environmental Sample 161
5.8.2.2 Epoxide Hydrolase from Aspergillus niger 161
5.8.2.3 Esterase from Bacillus subtilis 166
5.8.3 Oxidases 167
5.8.3.1 Monoamine Oxidase from Aspergillus niger 167
5.8.3.2 Baeyer–Villiger Monooxygenases 168
5.8.3.3 Cytochrome P450 Monooxygenases 170
5.8.4 Reductases 170
5.8.4.1 b-Keto Ester Reductase from Penicillium citrinum 170
5.8.4.2 Ketoreductase from an Environmental Sample 171
5.8.4.3 Enoate-Reductase YqjM 171
5.8.5 C–C Bond-Forming Enzymes 172
5.8.5.1 Aldolases 172
5.8.5.2 Benzoylformate Decarboxylase from Pseudomonas putida 173
References 175
6 Production and Isolation of Enzymes 191

Yoshihiko Hirose
6.1 Introduction 191
6.2 Enzyme Suppliers for Biotransformation 194
6.3 Origins of Enzymes 194
6.3.1 Microbial Enzymes 194
6.3.2 Plant Enzymes 195
6.3.3 Animal Enzymes 196
6.4 Fermentation of Enzymes 196
6.4.1 Liquid Fermentation 196
6.4.2 Solid Fermentation 196
6.4.3 Extraction of Enzymes 197
Contents XI
6.5 Extraction of Enzymes 197

6.5.1 Microbial Enzymes 197
6.5.2 Plant Enzymes 198
6.5.3 Animal Enzymes 198
6.6 Concentration 198
6.7 Purification of Enzymes 199
6.7.1 Chromatography 199
6.7.1.1 Ion-Exchange Chromatography (IEX) 199
6.7.1.2 Hydrophobic Interaction Chromatography (HIC) 203
6.7.1.3 Gel Filtration (GF) 206
6.7.1.4 Reversed-Phase Chromatography 207
6.7.1.5 Hydrogen Bond Chromatography 208
6.7.1.6 Affinity Chromatography 209
6.7.1.7 Salting-Out Chromatography 211
6.7.2 Precipitation 211
6.7.2.1 Precipitation by Salting Out 211
6.7.2.2 Precipitation by Organic Solvents 212
6.7.2.3 Precipitation by Changing pH 212
6.7.2.4 Precipitation by Water-Soluble Polymer 212
6.7.3 Crystallization 212
6.7.4 Stabilization During Purification 213
6.7.5 Storage of Enzymes 213
6.7.5.1 Storage in Liquids 213
6.7.5.2 Storage in Solids 213
6.8 Commercial Biocatalysts 214
References 214
7 Reaction and Process Engineering 217

John M. Woodley
7.1.1 Scope and Background 217
7.1.2 Role of Reaction Engineering 218
7.1.3 Applications 219
7.2 Reactor Options and Characteristics 219
7.2.1 Introduction 219
7.2.2 Ideal Reactor Types 220
7.2.3 Use of Multiple Reactors 223
7.2.4 Addition of Reagents 223
7.2.5 Alternative Reactors for Insoluble Enzymes 225
7.2.6 Alternative Reactors for Soluble Enzymes 225
7.2.7 Reactors for use with Multiphasic Systems 227
7.2.8 Reactor Scale-Up 228
7.3 Downstream Processing and Product Recovery 229
7.3.1 Downstream Schemes 229
7.3.2 Biocatalyst Recovery 230
XII Contents
7.4 Process Operation 231

7.4.1 Control of Operating Parameters 231
7.4.2 Reaction Control 231
7.5 Process Intensification 232
7.5.1 Process Metrics Required for an Effective
Process 232
7.5.2 Intensification Methods 234
7.5.2.1 Enzyme Immobilization 234
7.5.2.2 Use of Organic Solvents 235
7.5.2.3 Use of Resins 235
7.5.2.4 In Situ Product Removal 235
7.6 Process Intensification 236
7.6.2 Process Simulation 236
7.6.3 Environmental Assessment Tools 237
7.6.4 Operating Windows 238
7.6.5 Sensitivity and Uncertainty Analysis 238
7.6.6 Parameter Estimation 239
References 242
Part II Hydrolysis and Formation of C–O Bonds 249
8 Hydrolysis and Formation of Carboxylic Acid Esters 251

Monica Paravidino, Philipp Böhm, Harald Gröger, and Ulf Hanefeld
8.1.1 How Do Esterases (Lipases) Work? 251
8.1.2 Ester Synthesis versus Ester Hydrolysis 255
8.1.2.1 Ester Synthesis – Reactions in Organic Solvents 256
8.1.3 Stereochemistry 258
8.1.4 Reaction Concepts 262
8.2 Enantioselective Hydrolysis of Racemic Acyclic Carboxylates
(Resolutions) 265
8.2.1 Overview 265
8.2.2 Carboxylates with a Chiral Acid Moiety 265
8.2.2.1 Resolution of Carboxylates with a Non-functionalized Stereogenic
Center at the a-Position 266
8.2.2.2 Resolution of Carboxylates with an Amino-Functionalized
Stereogenic Center at the a-Position 270
8.2.2.3 Resolution of Carboxylates with a Hydroxy-(or oxo-)Functionalized
Stereogenic Center at the a-Position 273
8.2.2.4 Resolution of Carboxylates with Two Heteroatom-Substituted
Stereogenic Centers at the a,b-Positions 275
8.2.2.5 Resolution of Carboxylates with an Amino-Functionalized Stereogenic
Center at the b-Position 276
Contents XIII
8.2.2.6 Resolution of Carboxylates with a Hydroxy-(or oxo-)

Functionalized Stereogenic Center at the b-Position 280
8.2.2.7 Resolution of Carboxylates with a Stereogenic Heteroatom Center
at the b-Position 282
8.2.2.8 Resolution of Carboxylates with a Remote Stereogenic Center 282
8.2.2.9 Resolution of Carboxylates with Axial and Planar Chirality 283
8.2.3 Carboxylates with a Chiral Alcohol Moiety 284
8.2.3.1 Resolution of Esters with a Chiral Alcohol Moiety
(Non-heteroatom Functionalized) 285
8.2.3.2 Resolution of Esters with a Heteroatom Functionalized
Chiral Alcohol Moiety 286
8.2.3.3 Resolution of Esters with a Remote Stereogenic Center at
the Alcohol Moiety 289
8.2.3.4 Resolution of Esters with Axial Chirality at the Alcohol Moiety 291
8.3 Enantioselective Hydrolysis of Prochiral and meso-Carboxylates
(Desymmetrization) 292
8.3.1 Overview 292
8.3.2 Hydrolysis of Prochiral Carboxylates 293
8.3.3 Hydrolysis of meso-Carboxylates 296
8.4 Other Stereoselective and Non-stereoselective Hydrolysis of
Acyclic Carboxylates 298
8.5 Enantioselective Hydrolysis of Cyclic Esters (Lactones) and
Derivatives Thereof 299
8.5.1 Resolution of Lactones 300
8.5.2 Resolution of Azlactones 301
8.5.3 Resolution of Thiazolin-5-ones 302
8.6 Enantioselective Formation of Carboxylates via Esterification 302
8.6.1 Overview 302
8.6.2 Resolution of rac-Alcohols 302
8.6.2.1 Enzymatic Resolution of Primary Alcohols 304
8.6.2.2 Enzymatic Resolution of Secondary Alcohols 314
8.6.2.3 Enzymatic Resolution of Tertiary Alcohols 327
8.6.2.4 Enzymatic Resolution of rac-Diols 327
8.6.2.5 Enzymatic Resolution of rac-Acids and rac-Esters 334
8.6.2.6 Enzymatic Resolution of rac-Acids and rac-Esters with a Stereocenter
at the a Position 334
at the b-Position 337
8.6.3 Desymmetrization of Prochiral and meso-Carboxylates via
Transesterification 337
8.7 Enantioselective Formation of Carboxylates from Prochiral
and meso-Diols (Desymmetrization via Acylation) 339
8.7.1 Overview 339
8.7.2 Desymmetrization of Prochiral Diols 339
8.7.2.1 Desymmetrization of 2-Substituted 1,3-Diols 340
XIV Contents
8.7.2.2 Desymmetrization of 2,2-Disubstituted 1,3-Diols 342

8.7.2.3 Desymmetrization of 1,3,5-Triol Derivatives 344
8.7.3 Desymmetrization of meso-Diols 345
8.7.3.1 Desymmetrization of Primary Cyclic meso-Diols 345
8.7.4 Desymmetrization of Secondary Cyclic meso-Diols 348
8.8 Non-stereoselective Formation of (Fatty Acid-Based) Esters 350
References 351
9 Hydrolysis and Formation of Epoxides 363

Jeffrey H. Lutje Spelberg and Erik J. de Vries
9.1.1 Biocatalytic Strategies Towards Optically Pure Epoxides and
Derivatives 365
9.1.1.1 Epoxide Conjugation 365
9.1.1.2 Oxidation of Alkenes 366
9.1.1.3 Alcohol Dehydrogenases 366
9.1.1.4 Hydrolases and Other Enzymes Acting on an Ancillary
Functional Group 367
9.1.2 Scope and Outline of this Chapter 368
9.2 Conversion and Formation of Epoxides Catalyzed by Halohydrin
Dehalogenases 369
9.2.1 Classification, Structure, and Mechanism of Halohydrin
Dehalogenases 369
9.2.2 Discovery of Halohydrin Dehalogenases 371
9.2.3 Ring-Closure Reactions 373
9.2.3.1 Production of Chiral C3 Building Blocks Through Ring Closure 373
9.2.3.2 Production of Aromatic Building Blocks Through Ring Closure 375
9.2.4 Ring-Opening Reactions 376
9.2.5 Improving Halohydrin Dehalogenases by Mutagenesis and
Evolution 381
9.2.6 Towards 100% Yield 383
9.2.7 Cascade Reactions Using Multiple Enzymes 384
9.2.7.1 Haloperoxidase and Halohydrin Dehalogenase 384
9.2.7.2 Halohydrin Dehalogenase and Epoxide Hydrolase 385
9.2.7.3 Alcohol Dehydrogenase and Halohydrin Dehalogenase 386
9.2.8 Outlook on Halohydrin Dehalogenases 386
9.3 Hydrolysis of Epoxides Catalyzed by Epoxide Hydrolases 388
9.3.1 Epoxide Hydrolases in Nature 388
9.3.2 Discovery of Novel Microbial Epoxide Hydrolase Activity 389
9.3.3 Structure and Mechanism of Microbial Epoxide Hydrolases 390
9.3.4 Practical Application of Epoxide Hydrolases to the Synthesis of
Chiral Epoxides and Diols 391
9.3.5 Reaction Engineering 398
9.3.6 Improving Epoxide Hydrolases by Mutagenesis and Evolution 400
9.3.6.1 Epoxide Hydrolase Assays 401
Contents XV
9.3.7 Towards 100% Yield 403

9.3.7.1 Enantioconvergent Reactions Catalyzed by a Single Enzyme 403
9.3.7.2 Enantioconvergent Reactions by Employing Two Enzymes 404
9.3.7.3 Enantioconvergent Chemoenzymatic Reactions 405
9.3.7.4 Conversion of Meso-Epoxides 406
9.3.8 Outlook on Epoxide Hydrolases 406
References 408
10 Hydrolysis and Formation of Glycosidic Bonds 417

Daniela Monti and Sergio Riva
10.2 Glycosidases 419
10.2.1 Catalytic Mechanism 420
10.2.1.1 Inverting Glycosidases 420
10.2.1.2 Retaining Glycosidases 421
10.2.2 Glycosidases Inhibitors 421
10.2.3 Synthetic Applications of Glycosidases 421
10.2.3.1 Glycosidase-Catalyzed Hydrolysis of Glycosidic Bonds 422
10.2.3.2 Glycosidases-Catalyzed Formation of Glycosidic Bonds 423
10.2.4 Glycosynthases 427
10.3 Glycosyltransferases 428
10.3.1 Glycosyltransferases of the Leloir Pathway 429
10.3.2 Synthesis of Sugar Nucleoside Phosphates 432
10.3.3 Substrate Specificity and Synthetic Applications 438
10.3.4 New Glycosyltransferases from Microbial Sources 448
10.3.5 Non-Leloir Glycosyltransferases 453
References 454
11 Addition of Water to C¼C Bonds and its Elimination 467

Jianfeng Jin, Isabel W.C.E. Arends, and Ulf Hanefeld
11.2 Addition of Water to Isolated Double Bonds 469
11.2.1 Oleate Hydratase 469
11.2.2 Carotenoid Hydratases 469
11.2.3 Kievitone Hydratase 471
11.2.4 Acetylene Hydratase 471
11.2.5 Diol Dehydratase/Glycerol Dehydratase 472
11.3 Addition of Water to Conjugated Double Bonds 473
11.3.1 The Activating Group is an Acid Group 473
11.3.1.1 Fumarase 473
11.3.1.2 Malease and Citraconase 475
11.3.1.3 Aconitase 476
11.3.1.4 Urocanase 476
11.3.1.5 Dihydroxy Acid Dehydratase 477
11.3.1.6 Sugar Dehydratases 478
XVI Contents
11.3.1.7 2-Hydroxy-4-Dienoate Hydratases 478

11.3.1.8 Serine and Threonine Dehydratases 482
11.3.1.9 Hydratase-Tautomerase Bifunctionality 483
11.3.2 The Activating Group is a Ketone 484
11.3.2.1 Dehydroquinase 484
11.3.2.2 Scytalone Dehydratase 484
11.3.2.3 1,5-Anhydro-D-Fructose Dehydratase and Aldos-2-Ulose
Dehydratase 485
11.3.3 The Activating Group is a Thioester 486
11.3.3.1 Fatty Acid Biosynthesis 486
11.3.3.2 Fatty Acid Degradation, b-Oxidation 489
11.3.3.3 Hydroxycinnamoyl-CoA Hydratase Lyase (HCHL) 491
11.4 Outlook 492
References 492
12 Industrial Application and Processes Forming C–O Bonds 503

Lutz Hilterhaus and Andreas Liese
12.1 Processes Using Lipases 503
12.1.1 Processes Using Lipases in Hydrolytic Reactions 503
12.1.2 Processes Using Lipases in Esterifications 508
12.2 Processes Using Glycosyltransferases, Glycosidases, and
Carbon–Oxygen Lyases 513
12.2.1 Processes Applying Glycosyltransferases 514
12.2.2 Syntheses Using Carbon–Oxygen Lyases 517
12.2.3 Outlook 523
References 525
Part III Hydrolysis and Formation of C–N Bonds 531
13 Hydrolysis of Nitriles to Amides 533

Alexander Yanenko and Steffen Osswald
13.1 Nitrile Hydratases 533
13.1.1 Occurrence and Classification of Nitrile Hydratases 533
13.1.2 Protein Structure, Metal Cofactors, and Posttranslational
Modifications 534
13.1.3 Reaction Mechanism 535
13.1.4 Substrate Specificity 536
13.1.5 Enantioselectivity 536
13.2 Biocatalysts Containing Nitrile Hydratase 537
13.2.1 Whole-Cell Biocatalysts – Native Strains 537
13.2.2 Whole-Cell Biocatalysts – Recombinant Strains 541
References 542
Contents XVII
14 Hydrolysis of Nitriles to Carboxylic Acids 545

Steffen Oßwald and Alexander Yanenko
14.2 Nitrilases 545
14.2.1 Occurrence and Classification of Nitrilases 546
14.2.2 Protein Structure and Oligomerization 546
14.2.3 Reaction Mechanism 548
14.2.4 Side Activities 548
14.2.5 Substrate Specificity 550
14.2.6 Regioselectivity/Monohydrolysis of Dinitriles 550
14.2.7 (E)-/(Z)-Selectivity 551
14.2.8 Enantioselectivity 552
14.3 Nitrilase-Containing Biocatalysts for Hydrolysis of Nitriles to Acids 554
14.3.1 Whole Cell Biocatalysts 554
14.3.2 Enzyme Preparations 555
References 557
15 Hydrolysis of Amides 561

Theo Sonke and Bernard Kaptein
15.2 Enantioselective Hydrolysis of Carboxylic Acid Amides 561
15.3 Enantioselective Hydrolysis of Cyclic Amides 570
15.4 Enantioselective Hydrolysis of Amino Acid Amides 574
15.4.1 Synthesis of Enantiopure a-H-a-Amino Acids 575
15.4.1.1 L-Selective a-H-a-Amino Acid Amide Hydrolase 579
15.4.1.2 Leucine Aminopeptidases of the M17 Family 588
15.4.1.3 D-Selective a-H-a-Amino Acid Amide Hydrolase 594
15.4.2 Synthesis of Enantiopure a,a-Disubstituted Amino Acids 607
15.4.3 Synthesis of Enantiopure b-Amino Acids by b-Aminopeptidases 613
15.5 Enantioselective Hydrolysis of Hydroxy Acid Amides 618
15.6 Enantioselective Hydrolysis of Azido Acid Amides 620
15.7 Selective Cleavage of a C-Terminal Amide Bond 622
15.7.1 Peptide Amidase from the Flavedo of Oranges 622
15.7.2 Peptide Amidase from Microbial Sources 626
References 629
16 Hydrolysis and Formation of Hydantoins 651

Jun Ogawa, Nobuyuki Horinouchi, and Sakayu Shimizu
16.1 Overview of Microbial Hydantoin Metabolism and its
Application to Biotechnology 651
16.2 D-Hydantoinase 656
16.3 L-Hydantoinase 657
16.4 D-N-Carbamoylase 658
XVIII Contents
16.5 L-N-Carbamoylase 659

16.6 Hydantoin Rasemase 660
16.7 Biotechnology of Hydantoin-Transforming Enzymes 660
16.7.1 D-Amino Acid Production 660
16.7.2 L-Amino Acid Production 661
16.7.3 Recent Application of Hydantoin Racemase 662
16.7.4 Recent Applications of Hydantoinase 662
16.7.5 Recent Applications of N-Carbamoylase 663
16.7.6 Recent Application for b-Amino Acid Production 663
16.8 Structural Analysis and Protein Engineering of
Hydantoin-Transforming Enzymes 663
16.9 Diversity and Versatility of Cyclic Amide Transforming Enzymes
and its Application 665
16.10 Conclusion 669
References 669
17 Hydrolysis and Synthesis of Peptides 675

Timo Nuijens, Peter J.L.M. Quaedflieg, and Hans-Dieter Jakubke
17.2 Hydrolysis of Peptides 676
17.2.1 Peptide-Cleaving Enzymes 676
17.2.1.1 Introduction and Terminology 676
17.2.1.2 Catalytic Mechanism 680
17.2.1.3 EC Classification 684
17.2.1.4 Peptidase Families and Clans 684
17.2.2 Importance of Proteolysis 688
17.3 Synthesis of Peptides 692
17.3.1 Tools for Peptide Synthesis 692
17.3.2 Identification of the Ideal Enzyme 697
17.3.3 Principles of Enzymatic Peptide Synthesis 698
17.3.3.1 General Manipulations in Favoring Synthesis 699
17.3.3.2 Equilibrium-Controlled Synthesis 700
17.3.3.3 Kinetically Controlled Synthesis 701
17.3.3.4 Prediction of Synthesis by S0 Subsite Mapping 702
17.3.3.5 What Approach should be Preferred? 705
17.3.4 Manipulations to Suppress Competitive Reactions 706
17.3.4.1 Medium Engineering with Organic Solvents 706
17.3.4.2 Medium Engineering by Reducing Water Content 710
17.3.4.3 Substrate Engineering 714
17.3.5 Approaches to Irreversible Formation of Peptide Bond 715
17.3.5.1 Use of Nonpeptidases 715
17.3.5.2 Use of Proteolytically Inactive Zymogens 715
17.3.6 Irreversible C–N Ligations by Mimicking Enzyme
Specificity 717
17.3.6.1 Mechanism of Substrate Mimetic Hydrolysis 717
Contents XIX
17.3.6.2 Cationic Substrate Mimetics 720

17.3.6.3 Anionic Substrate Mimetics 721
17.3.6.4 Hydrophobic Substrate Mimetics 723
17.3.6.5 Chemoenzymatic Substrate Mimetic Approach 725
17.3.6.6 Highly Activated Acyl Donors 726
17.3.7 Planning and Process Development of Enzymatic
Peptide Synthesis 729
17.3.7.1 Stepwise Chain Elongation 729
17.3.7.2 Fragment Condensation 732
17.3.8 Enzymatic Modification of Peptides 737
17.4 Conclusion and Outlook 738
References 740
18 C–N Lyases Catalyzing Addition of Ammonia, Amines,

and Amides to C¼C and C¼O Bonds 749
Bian Wu, Wiktor Szymanski, Ciprian G. Crismaru, Ben L. Feringa,
and Dick B. Janssen
18.2 Addition of Ammonia and Amines to Fumaric Acid: L-Aspartase-
Fumarase Superfamily 750
18.2.1 General Properties 750
18.2.2 Structure and Catalytic Mechanism 751
18.2.3 Diversity 752
18.2.4 Biocatalytic Scope and Applications 752
18.2.5 Enzyme Engineering 753
18.3 Other Aspartase/Fumarase Family Members: Adenylosuccinate
Lyase, Argininosuccinate Lyase, and EDDS Lyase 754
18.3.1 Adenylosuccinate Lyase 754
18.3.2 Argininosuccinate Lyase 755
18.3.3 EDDS Lyase 755
18.4 Addition of Ammonia to Mesaconic Acid: L-Methylaspartase 756
18.4.2 Structure and Mechanism 757
18.4.3 Substrate Scope and Biocatalytic Application 758
18.5 Aromatic Amino Acid Ammonia Lyases 758
18.5.2 Structure and Mechanism 759
18.5.3 Distribution and Diversity 760
18.5.4 Biocatalytic Relevance and Applications 761
18.5.5 Engineering Studies 763
18.5.6 b-Alanyl CoA Ammonia Lyase 764
18.5.7 Serine Dehydratase, Threonine Dehydratase, and Other
Class IIPLP-Dependent Enzymes 765
18.5.8 L-Serine Dehydratase/Deaminase 766
18.5.9 D-Serine Dehydratase/Deaminase 767
XX Contents
18.5.10 L-Threonine Dehydratase/Deaminase 767

18.5.11 Threo-3-Hydroxy-L-Aspartate Ammonia-Lyase 768
18.5.12 Diaminopropionate Ammonia-Lyase 768
18.5.13 D-Glucosaminate Dehydratase 768
18.5.14 Fe-S-Dependent Serine Hydratases 769
18.5.15 Miscellaneous Lyases Adding Amines to C¼C Bonds 769
18.6 Conclusions and Outlook 771
References 772
19 Application of Transaminases 779

Matthias Höhne and Uwe T. Bornscheuer
19.2 Occurrence and Properties of Transaminases 781
19.2.1 Classification Based on Substrate Specificity 782
19.2.2 Classification Based on Sequence Similarities and
Three-Dimensional Structures 783
19.2.3 Mechanism 784
19.2.4 Methods to Assay Transaminase Activity and
Enantioselectivity 784
19.3 Strategies for Using Transaminases in Biocatalysis 788
19.3.1 Kinetic Resolution with Amine-TA 790
19.3.2 Asymmetric Synthesis with a-TA 792
19.3.2.1 Product Precipitation 793
19.3.2.2 Decomposition of the Keto Acid By-Product 793
19.3.2.3 Recycling of the Amino Donor via Reductive Amination 794
19.3.2.4 Coupling with v-Amino Acid TA 795
19.3.2.5 Synthesis of D-Amino Acids 795
19.3.2.6 Equilibrium Shift in Action 796
19.3.3 Asymmetric Synthesis with Amine-TA 798
19.3.3.1 Shifting the Equilibrium by Cyclization of the Amine Product 798
19.3.3.2 Shifting the Equilibrium by Removal of Coproduct 799
19.3.4 Amine-TA in Action: Optimization of Reactions for
Industrial Scale 802
19.3.4.1 In Situ Product Removal 802
19.3.4.2 Protein Engineering for Increasing Activity and
Thermostability 803
19.3.4.3 Protein Engineering for Decreasing Substrate and Product
Inhibition 804
19.3.5 Scope and Limitations of Amine-TA 805
19.3.5.1 Enantioselectivity 811
19.3.5.2 Substrate Scope 811
19.3.5.3 Enzyme Availability 813
19.4 Conclusions 813
References 814
Contents XXI
20 Industrial Applications and Processes Using Enzymes

Acting on C–N Bonds 821
Ruslan Yuryev, Lutz Hilterhaus, and Andreas Liese
20.2 Hydration of Nitriles to Amides 822
20.3 Hydrolysis of Nitriles to Acids 824
20.4 Hydrolysis and Formation of Amides 826
20.5 Processes Using Hydantoinases 839
20.6 Hydrolysis and Formation of Peptides 841
20.7 Processes Using C–N Lyases 845
20.8 Processes Using Transaminases 848
References 851
Part IV Formation and Cleavage of C–C Bonds 855
21 Aldol Reactions 857

Wolf-Dieter Fessner
21.1 Aldol Reactions 857
21.1.1 Classes of Aldolases 858
21.1.2 2-Deoxyribose 5-Phosphate Aldolase (EC 4.1.2.4) 861
21.1.3 Pyruvate/Phosphoenolpyruvate-Utilizing Aldolases 864
21.1.3.1 N-Acetylneuraminate (NeuNAc) Aldolase (EC 4.1.3.3) and
NeuNAc Synthetase (EC 4.1.3.19) 864
21.1.3.2 3-Deoxy-D-manno-2-octulosonate (Kdo) Aldolase
(EC 4.1.2.23) 872
21.1.3.3 2-Keto-3-deoxy-6-phosphogluconate (KDPG) Aldolase
(EC 4.1.2.14) and 2-Keto-3-deoxy-6-phosphogalactonate Aldolase
(EC 4.1.2.21) 873
21.1.3.4 SanM and 4-Hydroxy-3-methyl-2-keto-pentanoate Aldolase
(EC 4.1.3.39) 877
21.1.4 DHA/DHAP-Utilizing Aldolases 877
21.1.4.1 Fructose 1,6-Bisphosphate Aldolase (EC 4.1.2.13) 879
21.1.4.2 Fuculose 1-Phosphate Aldolase (EC 4.1.2.17), Rhamnulose
1-Phosphate Aldolase (EC 4.1.2.19) and Tagatose 1,6-Bisphosphate
Aldolase (EC 4.1.2.40) 880
21.1.4.3 Synthetic Strategies, Stereoselectivity, and Product Diversity Using
DHAP-Dependent Aldolases 882
21.1.4.4 Synthesis of Dihydroxyacetone Phosphate (DHAP) 895
21.1.4.5 Transaldolase (EC 2.2.1.2) and Fructose 6-Phosphate Aldolase
(EC 4.1.2.n) 898
21.1.5 Glycine-Utilizing Aldolases 901
21.1.6 Development of Novel Catalysts 908
References 909
XXII Contents
22 Acyloin and Benzoin Condensations 919

Martina Pohl, Carola Dresen, Maryam Beigi, and Michael Müller
22.1 Umpolung Reactions in Chemistry and Biology 919
22.2 Acyloin Condensations 920
22.2.1 Chemoselectivity of Enzymatic Acyloin Condensations 922
22.2.2 Stereoselectivity of Enzymatic Acyloin Condensations 923
22.2.3 Aliphatic–Aromatic Acyloins 924
22.2.3.1 Acyloin Condensations with Aliphatic Donor Aldehydes and
Aromatic Acceptors 924
22.2.4 Carboligation of Aromatic Donors and Aliphatic Acceptors 927
22.2.5 Araliphatic–Aliphatic Acyloins 928
22.2.6 Aliphatic Acyloins 929
22.2.7 Olefinic Aliphatic and Araliphatic Acyloins 929
22.2.8 2-Acyl-2-Hydroxy Acids 930
22.2.9 Sugar Derivatives 930
22.3 Benzoin Condensations 931
22.3.1 Benzoin Condensations 931
22.3.2 Cross Benzoin Condensations 932
22.4 Miscellaneous Acyloin Condensations 933
22.4.1 Stetter-Type Reactions 933
22.4.2 Acyloin Condensations with Ketones and Imines 935
22.4.3 Acyloin Condensations with Formaldehyde and
Formaldehyde Synthons 936
22.4.4 Racemic Resolution via Lyase/Ligase Reactions 938
References 940
23 Cleavage and Formation of Cyanohydrins 947

Mandana Gruber-Khadjawi, Martin H. Fechter, and Herfried Griengl
23.2 Hydroxynitrile Lyases Commonly Used for Preparative Application 948
23.2.1 (R)-Selective HNLs 948
23.2.2 (S)-Selective HNLs 951
23.3 Hydroxynitrile Lyase Catalyzed Addition of HCN to Aldehydes 953
23.3.1 (R)-Selective HNLs 953
23.3.2 (S)-Selective HNLs 955
23.4 HNL-Catalyzed Addition of Hydrogen Cyanide to Ketones 955
23.5 Transhydrocyanation 964
23.6 Mechanistic Aspects and Enzymatic Promiscuity 967
23.6.1 (R)-PaHNL (EC 4.1.2.10) 967
23.6.2 (R)-LuHNL (EC 4.1.2.46) 968
23.6.3 (S)-HbHNL (EC 4.1.2.47) 968
23.6.4 (S)-MeHNL (EC 4.1.2.47) 969
23.6.5 (S)-SbHNL (EC 4.1.2.11) 969
23.7 Improvement of HNLs by Enzyme Engineering,
Enzyme Stabilization 970
Contents XXIII
23.8 Resolution of Racemates 973

23.8.1 Hydroxynitrile Lyase as Catalyst 973
23.8.2 Esterase or Lipase as Catalyst 973
23.9 Follow-Up Chemistry of Enantiomerically Pure Cyanohydrins 975
23.10 Experimental Techniques for HNL-Catalyzed Biotransformations
and Safe Handling of Cyanides 977
23.10.1 HNL Catalysis in Aqueous Medium 978
23.10.2 HNL Catalysis in Organic Medium 978
23.10.3 HNL Catalysis in Biphasic Medium 979
23.10.4 Transhydrocyanation for HCN Generation 980
23.10.5 Technical Applications 981
References 981
24 Industrial Application and Processes Using Carbon–Carbon Lyases 991

24.1 Processes Using Carbon–Carbon Lyases 991
24.2 Syntheses Using Carboxy-Lyases 991
24.3 Syntheses Using Aldehyde Lyases 993
24.4 Syntheses Using Oxo-Acid Lyases 995
24.4.1 Synthesis of L-DOPA Catalyzed by Tyrosine Phenol Lyase
from Erwinia herbicola 997
24.5 Outlook 998
References 998
Part V Hydrolysis and Formation of P–O Bonds 1001
25 Hydrolysis and formation of P–O Bonds 1003

Ron Wever and Teunie van Herk
25.2 Biological Phosphorylating Agents, Phosphate Esters, and
Thermodynamic Considerations 1004
25.3 Enzymatic Phosphoryl Transfer Reactions and
Phosphorylated Intermediates 1007
25.3.1 Phosphorylation by Kinases 1007
25.3.2 Enzymes Used in the Regeneration of ATP 1007
25.4 Phosphate Hydrolyzing Enzymes: The Phosphatases 1009
25.4.1 Structural and Mechanistic Description of Alkaline Phosphatase 1010
25.4.1.1 Application of Alkaline Phosphatases in Dephosphorylation 1012
25.4.1.2 Transphosphorylation by Alkaline Phosphatases 1012
25.4.2 Structural and Mechanistic Description of Acid Phosphatases 1013
25.4.2.1 Dephosphorylation by Acid Phosphatases and 50 Ribonucleotide
Phosphohydrolases 1016
25.4.2.2 Transphosphorylation by Acid Phosphatases 1017
25.4.2.3 Formation of DHAP 1020
XXIV Contents
25.5 Phosphorylases 1021

25.6 Enzyme-Cascade Reactions in One Pot Using Phosphorylated
Intermediates 1022
25.7 Outlook 1027
References 1028
Part VI Reductions 1035
26 Reduction of Ketones and Aldehydes to Alcohols 1037

Harald Gröger, Werner Hummel, Sonja Borchert, and Marina Kraußer
26.2 Alcohol Dehydrogenases as Biocatalysts 1038
26.2.1 Overview of the Types of Alcohol Dehydrogenases 1040
26.2.1.1 Aldo-keto Reductases (AKRs) 1040
26.2.1.2 Medium-Chain Dehydrogenases/Reductases (MDR) 1040
26.2.1.3 Short-Chain Dehydrogenases/Reductases (SDR) 1041
26.2.2 Sources of Alcohol Dehydrogenases Useful for Biocatalysis 1042
26.2.2.1 (S)-Specific NADH-Dependent ADH from Horse Liver 1042
26.2.2.2 (S)-Specific NADPH-Dependent ADH from
Thermoanaerobacter sp 1042
26.2.2.3 (R)-Specific NADPH-Dependent ADH from Lactobacillus
kefir and L. brevis 1043
26.2.2.4 (S)-Specific NADH-Dependent ADH from Rhodococcus
erythropolis 1044
26.2.2.5 (S)-Specific NADH-Dependent ADH from Rhodococcus ruber 1045
26.2.2.6 (S)-Specific NADPH-Dependent ADH Gre2p from Saccharomyces
cerevisiae 1046
26.2.2.7 (R)-Specific NADH-Dependent ADH from Nocardia globerula 1046
26.2.2.8 (R)-Specific NADPH-Dependent ADH from Candida
magnolia 1047
26.2.2.9 (S)-Specific NADH-Dependent ADH from Sporobolomyces
salmonicolor 1047
26.2.2.10 NADPH-Dependent Glycerol Dehydrogenase (Gox1615) from
Gluconobacter oxydans 1048
26.2.3 Screening Methods to Obtain Novel ADHs 1049
26.3 Concepts of Biocatalytic Ketone and Aldehyde Reduction 1049
26.3.1 Overview of Process Concepts 1049
26.3.2 Ketone Reduction Based on Substrate-Coupled
Cofactor-Regeneration with Isopropanol 1054
26.3.2.1 Use of Isolated Enzymes 1054
26.3.2.2 Use of Whole Cells 1058
26.3.3 Enzyme-Coupled Cofactor-Regeneration Using a Formate
Dehydrogenase 1063
Contents XXV
26.3.4 Enzyme-Coupled Cofactor-Regeneration Using a Glucose

Dehydrogenase 1068
26.3.5 Enzyme-Coupled Cofactor-Regeneration Using a Glucose-6-
Phosphate Dehydrogenase 1074
26.3.6 Enzyme-Coupled Cofactor-Regeneration Using a Phosphite
Dehydrogenase 1075
26.3.7 Ketone Reduction Based on Wild-Type Microorganism and
Glucose in a Fermentation-Like Processes 1076
26.3.8 Cofactor Regeneration Using Chemocatalytic and Electrochemical
Methods 1079
26.4 Specific Synthetic Applications of Enzymatic Reductions 1081
26.4.1 Introduction and General Remarks 1081
26.4.2 Reduction of Ketones with Two Small Substituents 1081
26.4.3 Reduction of Multisubstituted and Hydroxy-Substituted
Acetophenone Derivatives 1083
26.4.4 Reduction of Bulky Ketones with Two Large Substituents 1085
26.4.5 Reduction of More Complex Cyclic Ketones 1090
26.4.6 Reduction of Steroid Ketones 1092
26.4.7 Reduction of Keto Esters 1095
26.4.8 Reduction of Aldehydes 1098
References 1101
27 Reduction of C¼C Double Bonds 1111

Despina J. Bougioukou and Jon D. Stewart
27.2 Alkene Reduction by Whole Microbial Cells 1111
27.2.1 Bakers Yeast 1112
27.2.2 Other Microbial Species 1114
27.3 Alkene Reductions by Isolated Enzymes 1116
27.3.1 Saccharomyces pastorianus Old Yellow Enzyme 1116
27.3.2 Fungal Old Yellow Enzyme Superfamily Members 1121
27.3.3 Bacterial Old Yellow Enzyme Superfamily Members 1124
27.3.4 Plant Old Yellow Enzyme Superfamily Members 1129
27.3.5 Enoate Reductases 1135
27.3.6 Medium-Chain Dehydrogenases 1138
27.3.7 Short-Chain Dehydrogenases 1143
27.4 Applications of Alkene Reductases 1143
27.4.1 a,b-Unsaturated Aldehydes and Ketones 1143
27.4.2 Acrylates and Acrylate Esters 1149
27.4.3 Nitroalkenes 1149
XXVI Contents
27.5 Accessing Both Product Enantiomers 1150

27.5.1 Using Wild-Type Enzymes 1150
27.5.2 Using Mutant Enzymes 1152
References 1153
28 Reductive Amination of Keto Acids 1165

Werner Hummel and Harald Gröger
28.2 Biochemical Properties of Enzymes Catalyzing Reductive
Amination Reactions 1170
28.2.1 L-Amino Acid Dehydrogenases 1170
28.2.1.1 Leucine Dehydrogenase 1171
28.2.1.2 Phenylalanine Dehydrogenase 1172
28.2.1.3 Glutamate Dehydrogenase 1177
28.2.1.4 Further Amino Acid Dehydrogenases 1177
28.2.2 D-Amino Acid Dehydrogenases 1180
28.2.3 N-Methyl-L-amino Acid Dehydrogenase 1180
28.2.4 Opine Dehydrogenases 1181
28.3 Synthetic Applications of Enzymes Catalyzing Reductive Amination 1181
28.3.1 Introduction and General Remarks 1181
28.3.2 Leucine Dehydrogenase Catalyzed Reductive Amination 1183
28.3.2.1 L-tert-Leucine 1183
28.3.2.2 L-Neopentylglycine (and Further Aliphatic Amino Acids) 1186
28.3.2.3 L-b-Hydroxyvaline 1188
28.3.2.4 Isotopically Labeled L-Amino Acids 1189
28.3.2.5 L-Amino Acids with two Stereogenic Centers 1190
28.3.3 Phenylalanine Dehydrogenase Catalyzed Reductive Amination 1191
28.3.3.1 Synthesis of (S)-2-Amino-4-Phenylbutanoic Acid
(L-Homophenylalanine) 1191
28.3.3.2 Synthesis of Allysine Ethylene Acetal ((S)-2-Amino-5-1,3-Dioxolan-2-
ylpentanoic Acid) 1192
28.3.3.3 Synthesis of the N-Terminal Amino Acid Portion of
Nikkomycins 1192
28.3.3.4 Synthesis of (S)-3-Hydroxyadamantylglycine 1193
28.3.4 Glutamate Dehydrogenase Catalyzed Reductive Amination 1194
28.3.4.1 Synthesis of L-6-Hydroxynorleucine with Glutamate
Dehydrogenase 1194
28.3.4.2 In Situ Synthesis of L-Glutamate as a Cosubstrate for
Transamination Processes 1195
28.3.5 D-Amino Acid Dehydrogenase-Catalyzed Reductive Amination 1195
28.3.6 N-Methyl-amino Acid Dehydrogenase 1196
28.3.7 Opine Dehydrogenase 1199
28.4 Summary 1199
References 1200
Contents XXVII
29 Industrial Application of Oxidoreductase catalyzed Reduction

of Ketones and Aldehydes 1205
Katharina Götz, Lutz Hilterhaus, and Andreas Liese
29.2 Reduction Processes Using Whole Cells 1205
29.3 Reduction Processes Using Isolated Enzymes 1211
29.3.1 Approaches for In Situ Cofactor Regeneration 1211
29.3.1.1 Substrate-Coupled Cofactor Regeneration 1212
29.3.1.2 Enzyme-Coupled Cofactor Regeneration 1214
29.4 Reductive Amination in Industry 1218
29.5 Summary 1220
References 1221
Part VII Oxidations 1225
30 Oxyfunctionalization of C–H Bonds 1227

Vlada B. Urlacher and Marco Girhard
30.2 Activation of Molecular Dioxygen 1228
30.3 Heme Metallo Monooxygenases 1229
30.3.1 Cytochrome P450 Monooxygenases 1229
30.3.2 Heme Peroxidases 1233
30.4 Non-heme Metallo Monooxygenases 1235
30.4.1 Non-heme Diiron Monooxygenases 1235
30.4.2 Tetrahydropterin-dependent Monooxygenases 1236
30.4.3 Other Metallo Monooxygenases 1236
30.5 Dioxygenases 1240
30.5.1 Rieske cis-diol Dioxygenases 1241
30.5.2 Iron(II)/a-Keto Acid-dependent Dioxygenases 1241
30.5.3 Lipoxygenases 1244
30.6 Oxyfunctionalization of C–H Bonds for Production of
Fine Chemicals 1245
30.6.1 General Aspects 1245
30.6.2 Oxidation of Fatty Acids 1246
30.6.3 Oxidation of Alkanes 1248
30.6.4 Oxidation of Terpenes and Terpenoids 1250
30.6.4.1 Monocyclic Monoterpenes: Limonene 1251
30.6.4.2 Dicyclic Monoterpenes: Pinene 1252
30.6.4.3 Sesquiterpenoides: Valencene 1253
30.6.4.4 Sesquiterpenoid Analogs: Ionone 1254
30.6.5 Oxidation of Steroids 1255
References 1258
XXVIII Contents
31 Oxyfunctionalization of C–C Multiple Bonds 1269

Bruno Bühler, Katja Bühler, and Frank Hollmann
31.2 Enzymes Capable of C–C Multiple Bond Oxyfunctionalization 1269
31.2.1 Binuclear Non-heme Iron Oxygenases 1270
31.2.2 Mononuclear Non-heme Iron Oxygenases 1270
31.2.3 Heme-Containing Monooxygenases 1274
31.2.4 Flavin-Dependent Oxygenases 1276
31.2.5 Peroxidases 1277
31.3 Epoxidation of C¼C Double Bonds 1278
31.3.1 Aliphatic Olefins 1278
31.3.2 Vinylaromatic Substrates 1290
31.3.3 Terpenes 1295
31.4 Dihydroxylation of C¼C Double Bonds 1302
31.4.1 Aliphatic Olefins and Conjugated Alkenes 1303
31.4.2 Terpenes 1305
31.5 Oxidative Cleavage of Double Bonds 1308
31.6 Triple Bond Oxyfunctionalization 1313
References 1316
32 Oxidation of Alcohols, Aldehydes, and Acids 1325

Frank Hollmann, Katja Bühler, and Bruno Bühler
32.2 Oxidation of Alcohols 1325
32.2.1 Alcohol Dehydrogenases (ADH) as Catalyst for the Oxidation
of Alcohols 1326
32.2.1.1 Commonly Used ADHs 1328
32.2.1.2 Horse Liver Alcohol Dehydrogenase (HLADH) 1328
32.2.1.3 Yeast Alcohol Dehydrogenase (YADH) 1330
32.2.1.4 ADHs from Thermophilic Organisms 1330
32.2.1.5 ADH from Rhodococcus ruber (ADH-A) 1331
32.2.1.6 Glycerol Dehydrogenases (GDHs) 1331
32.2.1.7 Other ADHs 1334
32.2.1.8 NAD(P)þ Regeneration Systems 1335
32.2.1.9 Miscellaneous (Non-enzymatic Approaches) 1338
32.2.2 NAD(P)-Independent Dehydrogenases 1341
32.2.2.1 Regeneration of NAD(P)-Independent Dehydrogenases 1344
32.2.3 Alcohol Oxidases 1345
32.2.3.1 Methods to Diminish/Avoid Hydrogen Peroxide 1345
32.2.3.2 Common Oxidases 1347
32.2.4 Peroxidases 1354
32.2.5 Laccases 1358
32.2.6 Aldehydes/Acids from Primary Alcohols 1363
32.2.6.1 Stopping the Oxidation at the Aldehyde Stage 1363
Contents XXIX
32.2.6.2 ‘‘Through Oxidations’’ 1369

32.2.7 Regioselective Oxidation in Polyols 1373
32.2.8 Kinetic Resolutions/Desymmetrizations 1379
32.2.9 Racemizations 1379
32.2.10 Deracemizations 1385
32.2.11 Stereoinversions 1391
32.3 Oxidation of Aldehydes 1392
32.3.1 Overview and Most Important Enzyme Classes/Applications 1392
32.3.2 Alcohol Dehydrogenases 1407
32.3.3 Aldehyde Dehydrogenases 1408
32.3.4 Monooxygenases 1410
32.3.5 Oxidases 1414
32.3.6 Aldehyde Oxidations with Intact Microbial Cells 1414
32.4 Oxidation of Carboxylic Acids 1418
32.4.2 Pyruvate Oxidase (EC 1.2.3.3) 1418
32.4.3 Formate Dehydrogenase (EC 1.2.1.2) 1420
32.4.4 Oxidations with Intact Microbial Cells 1421
32.4.4.1 Production of Benzaldehyde from Benzoyl Formate or
Mandelic Acid 1421
32.4.4.2 Microbial Production of cis,cis-Muconic Acid from Benzoic Acid 1422
32.4.4.3 Biotransformation of Substituted Benzoates into the Corresponding
cis-Diols 1422
References 1423
33 Baeyer–Villiger Oxidations 1439

Marko D. Mihovilovic
33.2 Mechanism and Enzyme Structure 1440
33.3 Cofactor Recycling and Preparative Operations 1443
33.4 Synthetic Applications 1448
33.4.1 Enzyme Platform 1448
33.4.2 Chemoselectivity 1452
33.4.3 Desymmetrizations 1453
33.4.4 Kinetic Resolutions 1456
33.4.5 Regioselectivity 1462
33.4.6 Application in Bioactive Compound and Natural
Product Synthesis 1469
33.5 Enzyme Engineering 1474
References 1478
34 Aromatic Oxidations 1487

David J. Leak, Ying Yin, Jun-Jie Zhang, and Ning-Yi Zhou
34.1 Enzymology of Aromatic Hydrocarbon Oxidation 1487
XXX Contents
34.1.1 Metabolism of Aromatic Compounds 1487

34.1.2 Dioxygenases 1490
34.1.3 Monooxygenases (Di-iron) 1496
34.1.4 Monooxygenases (Flavoprotein) 1498
34.1.5 Ring Cleavage Dioxygenases 1500
34.1.5.1 Intradiol Dioxygenase 1500
34.1.5.2 Extradiol Dioxygenase 1503
34.2 Biotransformations of Aromatic Compounds 1506
34.2.1 Whole Cell versus Cell-Free Reactions and Strategic
Approaches 1506
34.2.2 Dihydroxylations 1508
34.2.2.1 Substrate Specificity 1508
34.2.2.2 Reaction Selectivity 1509
34.2.2.3 Regioselectivity 1509
34.2.2.4 Stereoselectivity 1510
34.2.2.5 Effect of Ring Heteroatoms 1511
34.2.2.6 Using cis-Dihydrodiols in Synthesis 1511
34.2.2.7 Catechols 1515
34.2.3 Monohydroxylations 1515
34.2.4 Side Chain Oxidation 1517
34.2.5 Products from Ring-Cleavage Reactions 1517
34.2.6 Future Challenges 1519
References 1520
35 Oxidation of C–N Bonds 1535

Nicholas J. Turner
35.1 Introduction and Overview of Enzymes That Catalyze the
Oxidation of C–N Bonds 1535
35.2 L-Amino Acid Oxidase 1536
35.3 D-Amino Acid Oxidase (EC 1.4.3.3) 1537
35.3.1 Deracemization of Racemic Amino Acids Using Amino
Acid Oxidases 1539
35.4 Amine Oxidases 1542
35.4.1 Monoamine Oxidase MAO-N (EC 1.4.3.4) 1542
35.5 Amino Acid Dehydrogenases 1545
35.6 Flavin-Dependent Monooxygenase and P450 Monooxygenase 1547
35.7 Peroxidase, Laccase, and Tyrosinase 1548
35.8 Conclusions and Future Perspectives 1550
References 1551
36 Oxidation at Sulfur and Oxidation of Amino Groups 1553

Anke Matura and Karl-Heinz van Pée
36.1 Enzymes Oxidizing at Sulfur 1553
36.2 Oxidation of Sulfides 1554
Contents XXXI
36.2.1 Oxidation of Sulfides by Monooxygenases and by Whole

Organisms 1554
36.2.2 Oxidation of Sulfides by Peroxidases and Haloperoxidases 1557
36.3 Oxidation of Amino Groups 1559
36.3.1 Oxidation of Amino Groups by an Fe-Dependent Enzyme 1560
36.3.2 Oxidation of Amino Groups by a Mn-Containing, Radical-Mediated,
Hydrogen Peroxide-Dependent Enzyme 1561
36.3.3 Substrate Specificity of Amino Group Oxidizing Enzymes 1562
References 1563
37 Halogenation 1569
Karl-Heinz van Pée
37.1 Classification of Halogenating Enzymes and Their Reaction
Mechanisms 1569
37.1.1 Hydrogen Peroxide-Dependent Halogenases 1569
37.1.2 FADH2-Dependent Halogenases 1570
37.1.3 Non-heme Iron, a-Ketoglutarate, O2-Dependent Halogenases 1572
37.1.4 S-Adenosylmethionine-Dependent Halogenases 1573
37.2 Substrates for Halogenating Enzymes and Substrate
Specificity 1576
37.2.1 Haloperoxidases and Perhydrolases 1576
37.2.2 FADH2-Dependent Halogenases 1576
37.2.3 Non-heme Iron, a-Ketoglutarate, O2-Dependent Halogenases 1578
37.2.4 S-Adenosylmethionine-Dependent Fluorinase and Chlorinase 1578
37.3 Regioselectivity and Stereospecificity of Enzymatic Halogenation
Reactions 1579
37.4 Comparison of Chemical with Enzymatic Halogenation 1580
References 1581
38 Industrial Application and Processes Using Biocatalysts for

Oxidation Reactions 1585
38.1 Oxidation Processes Using Biocatalysts 1585
38.2 Oxidation by Oxidases 1586
38.2.1 Oxidative Deamination Catalyzed by Immobilized D-Amino
Acid Oxidase 1586
38.2.2 Kinetic Resolution by Whole Cells from Rhodococcus erythropolis 1587
38.2.3 Epoxidation by Oxidase 1589
38.2.4 Hydroxylation Catalyzed by Whole Cells 1590
38.2.5 Hydroxylation of Nicotinic Acid (Niacin) 1590
38.2.6 Reduction of Hydrogen Peroxide Concentration by Catalase 1591
38.3 Oxidation by Dehydrogenases 1591
38.4 Oxidation by Monooxygenases 1594
38.5 Oxidation by Dioxygenases 1598
References 1603
XXXII Contents
Part VIII Isomerizations 1607
39 Isomerizations 1609
Yasuhisa Asano and Kathrin Hölsch
39.2 Racemizations and Epimerizations 1610
39.2.1 Alanine Racemase (EC 5.1.1.1) 1610
39.2.2 Serine Racemase (EC 5.1.1.18) 1611
39.2.3 Phenylalanine Racemase (Gramicidin S Synthetase 1,
EC 5.1.1.11) 1612
39.2.4 L-to-D-Peptide Isomerase 1612
39.2.5 Dynamic Kinetic Resolution 1613
39.2.5.1 Amino Acid Racemase with Low Substrate Specificity
(EC 5.1.1.10) 1614
39.2.5.2 a-Amino-e-Caprolactam Racemase 1614
39.2.6 Synthesis by Enantiomerization (Deracemization) 1617
39.2.7 Cofactor-Independent Amino Acid Racemases and Epimerases 1618
39.2.7.1 Reaction Mechanism 1618
39.2.7.2 Glutamate Racemase (EC 5.1.1.3) 1619
39.2.7.3 Aspartate Racemase (EC 5.1.1.13) 1624
39.2.7.4 Diaminopimelate Epimerase (EC 5.1.1.7) 1625
39.2.7.5 Proline Racemase 1626
39.2.8 Other Racemases and Epimerases Acting on Amino Acid
Derivatives 1628
39.2.8.1 2-Amino-D2-thiazoline-4-carboxylate Racemase 1628
39.2.8.2 Hydantoin Racemase (EC 5.1.99.5) 1628
39.2.8.3 N-Acylamino Acid Racemase 1630
39.2.8.4 Isopenicillin N Epimerase (EC 5.1.1.17) 1631
39.2.9 Racemization and Epimerization at Hydroxy-Substituted
Carbons 1632
39.2.9.1 Mandelate Racemase (EC 5.1.2.2) 1633
39.2.9.2 Biocatalytic Racemization of Hydroxy Compounds Using
Microbial Cells 1637
39.2.10 Epimerases Acting on Carbohydrates and Derivatives 1637
39.2.10.1 N-Acylglucosamine 2-Epimerase (EC 5.1.3.8) 1637
39.2.10.2 Carbohydrate Epimerases Involved in Sugar Nucleotide Synthesis 1641
39.2.10.3 Ketohexose 3-Epimerases 1641
39.3 Cis–Trans Isomerases (EC 5.2) 1643
39.3.1 Maleate Cis–Trans Isomerase (EC 5.2.1.1) 1643
39.3.2 Linoleate Cis–Trans Isomerase (EC 5.2.1.5) 1644
39.3.3 Monounsaturated Fatty Acid Cis–Trans Isomerases 1646
39.4 Intramolecular Oxidoreductases (EC 5.3) 1646
39.4.1 Triosephosphate Isomerase (EC 5.3.1.1) 1647
39.4.2 D-Arabinose Isomerase (EC 5.3.1.3) 1648
39.4.3 L-Arabinose Isomerase (EC 5.3.1.4) 1651
Contents XXXIII
39.4.4 D-Xylose (Glucose) Isomerase (EC 5.3.1.5) 1652

39.4.4.1 Biochemical Properties and Reaction Mechanism 1653
39.4.4.2 Production of High-Fructose Corn Syrup 1655
39.4.4.3 Production of Bioethanol 1656
39.4.4.4 Production of Non-natural Sugar Derivatives 1657
39.4.5 L-Rhamnose Isomerase (EC 5.3.1.14) 1657
39.4.6 Isopentenyl-Diphosphate D-Isomerase (EC 5.3.3.2) 1660
39.5 Mutases (EC 5.4) 1661
39.5.1 Chorismate Mutase (EC 5.4.99.5) 1661
39.5.2 Aminomutases 1663
39.5.2.1 Lysine 2,3-Aminomutase (EC 5.4.3.2) 1663
39.5.2.2 Ornithine 4,5-Aminomutase (EC 5.4.3.5) 1666
39.5.2.3 b-Lysine 5,6-Aminomutase (EC 5.4.3.3) and D-Lysine 5,6-Aminomutase
(EC 5.4.3.4) 1666
39.5.2.4 Glutamate Mutase (EC 5.4.99.1) 1666
39.5.2.5 Tyrosine 2,3-Aminomutase (EC 5.4.3.6) 1667
39.5.3 Isomaltulose Synthase (EC 5.4.99.11) 1671
References 1672
40 Industrial Application and Processes Using Isomerases 1685

40.1 Isomerase Processes 1685
40.2 Syntheses Using Racemases and Epimerases 1685
40.3 Syntheses Using Intramolecular Oxidoreductases 1687
40.4 Syntheses Using Mutases 1688
40.5 Outlook 1689
References 1689
Part IX Extended Applications of Enzyme Catalysis 1693
41 Enzymatic Catalytic Promiscuity and the Design of New Enzyme

Catalyzed Reactions 1695
Uwe T. Bornscheuer and Romas J. Kazlauskas
41.2 Enzymatic Catalytic Promiscuity 1696
41.2.1 Hydrolysis and Other Substitutions at Carboxylic Acids
and Derivatives 1696
41.2.1.1 Serine Hydrolases 1696
41.2.1.2 Metallohydrolases 1702
41.2.2 Carbon–Carbon Bond Formation 1703
41.2.2.1 Aldol Additions: Enolate Formation 1703
41.2.2.2 Michael Additions and Related Additions to a,b-Unsaturated
Carbonyl Compounds 1707
41.2.2.3 Acyloin Condensation 1709
41.2.2.4 Cationic Polyene Cyclizations by Terpene Cyclases 1710
XXXIV Contents
41.2.3 Oxidation–Reduction 1712

41.2.3.1 Non-heme Iron(II) and 2-Oxoglutarate-Dependent Enzymes 1712
41.2.3.2 P450 Enzymes 1713
41.2.4 Sugar Coupling: Glycosynthases and Related Reactions 1713
41.2.5 Other Catalytically Promiscuous Reactions 1715
41.2.5.1 Racemases 1715
41.2.5.2 Apparent Catalytic Promiscuity 1716
41.2.5.3 Catalysis by Non-catalytic Proteins 1717
41.3 Design of New Enzyme Catalyzed Reactions 1717
41.3.1 Protein Engineering to Add New Catalytic Steps 1717
41.3.2 Amino Acid Changes Based on a Related Enzyme
(Mimic Divergent Evolution) 1718
41.3.2.1 Esterase to Epoxide Hydrolase 1719
41.3.2.2 Esterase to Oxynitrilase 1719
41.3.2.3 Adding Cofactors or Substituting Metal Ions 1720
41.3.3 Directed Evolution and Computational Design of New
Catalytic Activities 1722
References 1723
42 Catalytic Antibodies 1735

Ivan V. Smirnov, Alexey A. Belogurov Jr., Arina V. Kozyr,
and Alexander Gabibov
42.1 Obtaining of Catalytic Antibodies: Principles and Techniques 1735
42.2 Aldol and Retro-aldol Reactions 1738
42.3 Cyclization Reactions 1739
42.4 Hydrolysis of the Ester Bond 1741
42.4.1 Esterase Antibodies 1741
42.4.2 Phosphate Ester Hydrolysis 1747
42.4.3 Naturally Occurring Antibodies with Nuclease Activity:
Possible Role in Apoptosis 1747
42.5 Glycosidase Antibodies 1754
42.6 Formation and Opening of the Oxirane Ring 1755
42.7 1,2-Elimination Reactions (b-Elimination) 1756
42.8 Diels–Alder and Other Cycloaddition Antibodies 1756
42.9 Isomerization Reactions 1757
42.10 Hydrolysis and Formation of the Amide Bond 1758
42.10.1 Amidase and Ligase Antibodies 1758
42.10.2 Naturally Occurring Antibodies with Amidase Activity:
Beneficial and Pathogenic Impact 1759
42.10.2.1 Catalytic Antibodies and Hormonal Dysfunctions 1760
42.10.2.2 Blood Factors as Targets for Induced Catalytic Antibody
Response 1761
42.10.2.3 Infectious Diseases 1761
42.10.2.4 Catalytic Antibodies and Neurodegeneration 1762
Contents XXXV
42.10.3 Induction of Artificial Proteolytic Abzymes: Antigen-Directed

Abzyme Prodrug Therapy 1763
42.11 Oxido-reductase Antibodies: Implication in Innate Immunity 1765
42.12 Miscellaneous Reactions 1766
42.13 Applications of Antibody Catalysis – ‘‘Catalytic Vaccines’’
Based on Abzymes, Abzymes as Potential Scavengers
of Organophosphorous Poisons 1767
42.14 Abzymes in Non-aqueous Solutions: Application of
Nanocompartments 1770
References 1770
43 Chemoenzymatic Dynamic Kinetic Resolution and Related

Dynamic Asymmetric Transformations 1777
Ibrar Hussain and Jan-E. Bäckvall
43.2 DKR of Secondary Alcohols 1779
43.3 DKR of Secondary Alcohols with Two Large a-Groups 1784
43.4 DKR of Heteroaromatic Alcohols 1785
43.5 DKR of Cyanohydrins 1786
43.6 DKR of b-Halohydrins 1787
43.7 Dynamic Kinetic Asymmetric Transformation (DYKAT) of Diols 1789
43.8 DKR of Allylic Alcohols 1791
43.9 DKR of Primary Alcohols 1793
43.10 DKR of a-Amino Acid Esters 1795
43.11 DKR of Amines 1796
43.12 DKR of Axially Chiral Allenes 1801
43.13 Hydrolysis Reactions 1802
43.14 Concluding Remarks 1803
References 1804
44 Biocatalysis in Material Science 1807

Georg M. Guebitz
44.2 Synthesis and Functionalization of Synthetic Polymers 1808
44.2.1 Enzymatic Synthesis of Polyesters 1808
44.2.2 Limited Enzymatic Surface Hydrolysis 1811
44.2.2.1 Poly(alkylene terephthalates) 1811
44.2.2.2 Enzymatic Hydrolysis of Polyamides 1814
44.2.2.3 Enzymatic Hydrolysis of Polyacrylonitriles 1816
44.2.3 Surface Modification with Oxidoreductases 1817
44.2.4 Enzymatic Grafting 1818
44.3 Surface Functionalization of Biopolymers 1819
44.3.1 Enzymatic Modification of Lignocellulose Based Materials 1819
44.3.1.1 Transesterification and Transglycosylation Reactions 1819
44.3.1.2 Coupling Reactions by Using Oxidoreductases 1822
XXXVI Contents
44.3.2 Enzymatic Modification of Protein-Based Materials 1824

44.3.2.1 Hydrolytic Surface Modification 1824
44.3.2.2 Crosslinking and Grafting 1824
44.4 Conclusion 1825
References 1826
45 Industrial Applications of Enzymes in Emerging Areas 1837

Anne van den Wittenboer, Lutz Hilterhaus, and Andreas Liese
45.1 Industrial Processes Using Catalytically Promiscuous
Enzyme Activities 1837
45.2 Chemoenzymatic Industrial Processes 1838
45.2.1 Chemoenzymatic Dynamic Kinetic Resolution of Secondary
Alcohols 1838
45.2.2 Chemoenzymatic Deracemization of Amines and Amino Acids 1839
45.2.3 Chemoenzymatic Synthesis of Xolvone 1840
45.3 Industrial Application of Enzymes in Material Science 1841
45.3.1 Enzymatic Large-Scale Production of Poly(hexane-1,6-diyl
adipate) 1842
45.3.2 Enzymatic Synthesis of Aqueous Polyamide Dispersions 1842
45.3.3 Enzymes Applied in the Textile Industry 1843
References 1845
Part X Tabular Survey of Available Enzymes 1847
46 Tabular Survey of Available Enzymes 1849

David Rozzell
46.2 Almac Sciences 1850
46.2.1 Contact Information 1850
46.2.2 Carbonyl Reductase: CREDs 1850
46.2.3 Nitrile Hydratases and Nitrilases: NESK-1400 1851
46.2.4 Hydrolases: HESK-4600 1852
46.3 Amano Enzyme Company 1854
46.3.2 Lipases 1856
46.3.3 Proteases 1857
46.3.4 Acylases 1857
46.4 ASA Spezialenzyme GmbH 1857
46.4.3 Oxidoreductases 1859
46.4.4 Oxynitrilases 1860
46.5 Asahi Kasei Pharma Corporation 1860
Contents XXXVII
46.5.3 Oxidases 1862

46.5.4 Dehydrogenases 1863
46.5.5 Lyases, Decarboxylases, Ligases, and Miscellaneous Enzymes 1865
46.6 BBI Enzymes 1865
46.6.2 Miscellaneous Hydrolytic and Redox Enzymes 1867
46.7 Bio-Research Products 1868
46.7.2 Miscellaneous Enzymes 1868
46.8 Biocat Collection 1869
46.8.2 Enzyme Availability 1870
46.9 Biocatalysts Ltd 1870
46.9.2 Carbohydrate Hydrolases 1871
46.9.3 Lipases and Esterases 1872
46.9.4 Proteases and Peptidases 1872
46.10 ChiralVision BV 1873
46.10.2 Generic CaLB 1874
46.10.3 Genencor Proteases 1874
46.10.4 Extremely Thermostable Proteases 1875
46.10.5 Immobilization Support Material 1876
46.10.5.1 ImmobeadTM 1876
46.11 c-LEcta GmbH 1876
46.11.2 Lipases 1877
46.12 Codexis, Inc 1877
46.12.2 Catalog Enzyme Products: Screening Kits 1878
46.12.3 Cofactor Recycling Enzymes 1880
46.12.4 Other Enzymes 1881
46.13 Daicel 1881
46.13.2 Enzyme Screening Sets 1883
46.14 EnzBank (Korea) 1883
46.15 EnzySource 1885
46.16 Eucodis Bioscience GmbH 1885
46.16.2 Lipases 1886
46.16.3 Phospholipases 1887
46.16.4 b-Lactamases 1888
XXXVIII Contents
46.17 Evocatal GmbH 1888

46.17.2 Ketoreductases 1889
46.17.2.1 Ketoreductase Screening Sets and Kits 1889
46.17.3 NAD(P)H Regeneration Enzymes 1889
46.17.4 Other Alcohol Dehydrogenases 1890
46.17.5 Amino Acid Dehydrogenases 1890
46.17.6 Lyases 1891
46.18 Godo Shushei 1891
46.19 IMEnz Engineering 1892
46.19.2 Thermostable Proteases 1892
46.20 Libradyn 1893
46.20.2 Enzyme Products 1893
46.21 LibraGen 1893
46.21.2 Enzyme Screening Sets 1894
46.22 Meito Sangyo Co., Ltd 1895
46.23 Nagase ChemteX Corporation 1897
46.23.3 Other Enzymes 1899
46.24 Nzomics Biocatalysis 1899
46.24.2 Nitrile Hydratases 1900
46.24.3 Nitrilases 1901
46.24.4 Carbohydrate Hydrolyzing and Modifying Enzymes 1902
46.25 Osaka Saikin Kenkyusho 1904
46.26 Scientific Protein Laboratories 1904
46.27 Syncore Laboratories 1906
46.27.2 Ene Reductases 1906
46.27.4 Nitrile Hydratases 1908
46.27.5 Nitrilases 1908
Contents XXXIX
46.27.6 Hydroxynitrile Lyases (Oxynitrilases) 1909

46.27.7 Transaminases (also Known as v-Transaminases) (Amine
Forming) 1910
46.27.8 Nitro Reductases 1911
46.27.9 Amidases 1911
46.27.10 Glucose Dehydrogenase 1912
46.27.11 Formate Dehydrogenase 1912
46.27.12 Glycosidases 1913
46.27.14 Immobilized Lipases, Particle Size ¼ 150–300 mm 1915
46.27.15 Immobilized Proteases, Particle size ¼ 150–300 mm 1916
46.27.16 Other Immobilized Enzymes 1916
46.28 Toyobo Enzymes 1917
46.28.1.1 Distributors 1917
46.28.3 Oxidases 1921
46.28.4 Hydroxylases 1924
46.28.5 Dehydrogenases, Nicotinamide-Requiring 1924
46.29 Unitaka, Ltd. 1927
46.29.3 Dehydrogenases, Nicotinamide-Requiring 1928
46.30 Valley Enzymes 1930
46.31 X-Zyme GmbH 1932
46.31.3.1 Ketoreductase Screening Sets and Kits 1934
46.31.4 NAD(P)H Regeneration Enzymes 1935
46.31.5 Other Alcohol Dehydrogenases 1935
46.31.6 Amino Acid Dehydrogenases 1936
46.31.7 Decarboxylases 1937
46.31.8 Oxidases 1937
46.31.9 Lyases 1938
46.31.10 Glycosyl Transferases 1938
Index 1939
jV
Foreword
When the first edition of Enzyme Catalysis in Organic Synthesis: A Comprehensive

Handbook was planned and then published biocatalysts had made their first major
impact on research in organic synthesis both in academia and in industry. Significant
advances in the expression and isolation of enzymes as well as techniques for their
immobilization and stabilization had prompted a flurry of research activities and
publications from the organic chemistry community and the industrial sector. This
was adequately reflected in the handbook with its dominant central part summa-
rizing applications of enzymes in academic and industrial synthesis, arranged
according to reaction type.
In the following years, the number of applications – less so the kind of novel
transformations – steadily grew, but gradually the main emphasis in the community
shifted from the application to the biocatalysts themselves. Directed evolution,
molecular biology techniques, and biocatalyst design began to have a major impact
and to define new challenges and the frontier of research in the field.
Since the publication of the second edition the use of enzymes in organic
synthesis has again undergone a major shift in emphasis. Green chemistry, the
focus of the chemical industry towards increasingly environmentally friendlier
processes, is now a major driver in the establishment of novel techniques and
transformations, and clearly the advantages of biocatalytic transformations shape the
field to a major extent. Along with the scientific and economic pull arising from this
industrial demand, the scientific push provided by the rise of synthetic biology
offers completely novel opportunities for research and application.
This shift in emphasis since the publication of the first edition is mirrored by an
appropriate change in the editors (and of course the authors) of the handbook.
Karlheinz Drauz has a knowledge of the whole field and experience in research in
biocatalysis in more fundamental research as well as in industrial application that is
second to none. He guarantees the continuity and the reality check that the
practicing reader of a handbook expects. Harald Gr€ oger and Oliver May together
guarantee that the face-changing recent developments in academia and industry
are very well reflected in the scientific and technological expertise of the editorial
team.
VI j Foreword
The third edition of Enzyme Catalysis in Organic Synthesis accordingly has kept the
originally chosen classification of biocatalytic transformations by reaction type but it
also covers new developments that have changed and will change the face of the field.
It is not only an invaluable source of knowledge and references, it also embodies a
treasure that consists of a multitude of findings not yet exploited by the scientific
community. I have no doubt that – like the first and the second editions – it will fuel
numerous research projects and industrial applications.
Dortmund, November 2011

Herbert Waldmann
jXLI
About the Editors
Karlheinz Drauz Karlheinz Drauz was Vice Presi-

dent International Scientific Relations of Evonik
Degussa. He evaluated interesting topics in research
and development and networks globally with aca-
demic and industrial institutions. He did his Ph.D. at
the Technical University Stuttgart, Germany, and
started his career at Degussa 1980 and run through
various positions in research and development and
innovation management. Amongst his research
interests are amino acids, peptides, biological active
compounds, asymmetric synthesis, metal and bio-
catalysis as well as material science and process
chemistry. He holds more 160 patents and 100 scientific publications. Since 1992
he is honorary professor for organic chemistry at the University W€urzburg, Germany.
Since 2010 he functions as Senior Advisor.
Harald Gr€oger Harald Gr€ oger studied Chemistry at

the Universities of Erlangen-N€ urnberg and Olden-
burg and received his diploma degree in Chemistry
from the University of Oldenburg in 1994. His
doctoral thesis he completed at the University of
Oldenburg in 1997 under the supervision of Prof. Dr.
Martens. After staying as a postdoctoral fellow at the
University of Tokyo in the group of Prof. Dr. Shiba-
saki, he joined the research department Chemische
Forschung of SKW Trostberg AG in 1998. After the
merger with Degussa-H€ uls AG to Degussa AG in
2001, he became Project Manager in the Project
House Biotechnology of Degussa AG. From 2004 to 2006 he worked as a Senior
Project Manager at the research unit Service Center Biocatalysis of Degussa AG.
From 2006 to 2011 he was W2-Professor (Associate Professor) for Organic Chemistry
XLII j About the Editors
at the University of Erlangen-N€
urnberg, and since April 2011 he is W3-Professor (Full
Professor) for Organic Chemistry at Bielefeld University. Harald Gr€ oger has
authored more than 90 scientific publications and more than 30 patent applications.
He and his teams were awarded the Degussa Innovation Award 2003 (category: new
products) and the Degussa Innovation Award 2005 (category: new or improved
processes). In addition, he was awarded the Carl-Duisberg-Memorial-Prize 2008 of
the German Chemical Society (GDCh). His main current research areas center on
the use of biocatalysts in organic synthesis.
Oliver May Oliver May is R&D Director of DSM Bio-

based Products & Services and was responsible for
the Biocatalysis competence field within DSM as
Corporate Scientist. Before joining DSM in April
2006, he was with Degussa (now Evonik) in various
functions; latest as General Manager of Degussas
Service Center Biocatalysis. He was educated in
Germany, receiving a PhD degree in Technical Biol-
ogy from the University of Stuttgart where he worked
at the Institute of Biochemical Engineering for Prof.
C. Syldatk and at the German Center for Biotechnol-
ogy in Braunschweig with Prof. D. Schomburg. He
joined Caltech as a postdoc in 1998 where he worked
until 2000 in the group of Prof. F.H. Arnold on directed evolution of enzymes. Oliver
May has authored more than 40 scientific contributions in Journals and Books and
more than 15 patent applications. He and his team was awarded several research &
innovation awards on hydantoinase technology, whole-cell processes, recombinant
pig liver esterase in its application for production of a pharma intermediate and the
latest on advanced yeasts for production of bioethanol from cellulosic feedstocks.
XLIII
Preface
While biocatalysis experts refer to the previous two editions of this handbook as ‘‘The
Drauz–Waldmann’’ handbook its official title is of course Enzyme Catalysis in
Organic Synthesis and it is recognized as a reference work in the field of biocatalysis.
We hope this third edition will provide the same value as the previous two editions
and so become known as ‘‘The Drauz–Gröger–May’’ handbook. The fact that you are
holding this book in your hands shows your trust in our selection of the world
renowned experts who have authored this book. All the authors have put a lot of
effort into their individual chapters to secure high level contributions and to create a
reference work on biocatalysis.
We felt that a third edition is necessary as ten years have elapsed since the last
edition, which is a long time in such a dynamic field. Progress is reflected by the fact
that many of the chapters had to be completely rewritten and new chapters have been
added. To show the relevance of biocatalysis one fact was very important for us:
highlighting proven industrial applications by adding new coherently structured
application sections to the various chapters dealing with the different chemical
reaction types. We hope this will convince non-professionals in biocatalysis that
this technology is an established tool that should not be omitted from the repertoire
of any chemist working on the development of highly efficient syntheses in acade-
mia as well as industry.
There are also elements that we did not want to change. Again we have chosen to
keep the overall arrangement of three different volumes, of which the first provides a
comprehensive introduction to the field and important enabling tools. The other two
volumes focus on specific reaction types and emerging fields in biocatalysis. Many
evolutions, or even revolutions, have taken place over the last decade, especially in
the field of enabling tools. While directed evolution was just emerging when we
issued the second edition, in the first volume of this new edition genome sequen-
cing, gene synthesis, metagenomics, and bioinformatics are now much more
prominently featured as standard tools that have an enormous positive impact on
development speed and diversity of enzymes that can thus be created. The reader
will also observe many new developments in specific reaction types, for example, the
conversion of ketones into amines and alcohols triggered by the improved accessi-
bility of transaminases and dehydrogenases.
XLIV Preface
We hope this book will motivate a generation of open-minded chemists to capture

the full potential of biocatalysis and to collaborate closely with biologists to enable
urgently needed innovations for today in an alliance that enables the most successful
mastery of chemistry in the future.
Hanau, Bielefeld, and Aachen Karlheinz Drauz, Harald Gröger,

November 2011 and Oliver May
XLV
List of Contributors
Wolfgang Aehle Maryam Beigi

BRAIN – Biotechnology Research and Albert-Ludwigs University Freiburg
Information Network AG Institute of Pharmaceutical Sciences
Darmstädter Straße 34–36 Albertstr. 25
64673 Zwingenberg 79104 Freiburg
Germany Germany
Isabel W.C.E. Arends Alexey A. Belogurov Jr.

Delft University of Technology Russian Academy of Sciences
Department of Biotechnology Shemyakin and Ovchinnikov Institute
Biocatalysis and Organic Chemistry of Bioorganic Chemistry
Julianalaan 136 Ul. Miklukho-Maklaya 16/10
2628 BL Delft 117871 Moscow
The Netherlands Russia
Yasuhisa Asano Philipp Böhm

Toyama Prefectural University Bielefeld University
Biotechnology Research Center and Faculty of Chemistry
Department of Biotechnology Universitätsstr. 25
5180 Kurokawa, Imizu 33615 Bielefeld
Toyama 939-0398 Germany
Japan
Sonja Borchert
Jan-E. Bäckvall University of Erlangen-Nürnberg
Stockholm University Department of Chemistry and
Arrhenius Laboratory Pharmacy
Department of Organic Chemistry Henkestrasse 42
10691 Stockholm 91054 Erlangen
Sweden Germany
XLVI List of Contributors
Uwe T. Bornscheuer Erik J. de Vries

Greifswald University Codexis, Inc.
Institute of Biochemistry Corporate R&D
Department of Biotechnology & 200 Penobscot Drive
Enzyme Catalysis Redwood City, CA 94063
Felix-Hausdorff-Str. 4 USA
17487 Greifswald
Germany Carola Dresen
Albert-Ludwigs University Freiburg
Despina J. Bougioukou Institute of Pharmaceutical Sciences
University of Florida Albertstr. 25
Department of Chemistry 79104 Freiburg
102 Leigh Hall Germany
Gainsville, FL 32611
USA Juergen Eck
BRAIN – Biotechnology Research and
Katja Bühler Information Network AG
TU Dortmund University Darmstädter Straße 34–36
Department of Biochemical and 64673 Zwingenberg
Chemical Engineering Germany
Laboratory of Chemical Biotechnology
Emil-Figge-Str. 66 Martin H. Fechter
44227 Dortmund Austrian Foundry Research Institute
Germany Chemical Laboratory
Parkstraße 21
Bruno Bühler 8700 Leoben
TU Dortmund University Austria
Department of Biochemical and
Chemical Engineering Ben L. Feringa
Laboratory of Chemical Biotechnology University of Groningen
Emil-Figge-Str. 66 Stratingh Institute for Chemistry
44227 Dortmund Niejenborgh 4
Germany 9747 AG Groningen
The Netherlands
Ciprian G. Crismaru
University of Groningen Wolf-Dieter Fessner
Groningen Biomolecular Sciences and Technische Universität Darmstadt
Biotechnology Institute Department of Organic Chemistry and
Biochemical Laboratory Biochemistry
Niejenborgh 4 Petersenstrasse 22
9747 AG Groningen 64287 Darmstadt
The Netherlands Germany
List of Contributors XLVII
Alexander Gabibov Herfried Griengl

Russian Academy of Sciences Graz University of Technology
Shemyakin and Ovchinnikov Institute Institute of Organic Chemistry
of Bioorganic Chemistry Stremayrgasse 9
Ul. Miklukho-Maklaya 16/10 8010 Graz
117871 Moscow Austria
Russia
Harald Gröger
Eduardo García-Urdiales Bielefeld University
University of Oviedo Faculty of Chemistry
Faculty of Chemistry Universitätsstrasse 25
Department of Organic and Inorganic 33615 Bielefeld
Chemistry Germany
Avenida Julián de Clavería 8
33006 Oviedo (Asturias) Mandana Gruber-Khadjawi
Spain ACIB GmbH
c/o Graz University of Technology
Marco Girhard Institute of Organic Chemistry
Heinrich-Heine University Düsseldorf Stremarygasse 9
Institute of Biochemistry 8010 Graz
Universitätsstraße 1 Austria
40225 Düsseldorf
Germany Georg M. Guebitz
Graz University of Technology
Vicente Gotor Department of Environmental
University of Oviedo Biotechnology
Faculty of Chemistry Petersgasse 12
Department of Organic and Inorganic 8010 Graz
Chemistry Austria
Avenida Julián de Clavería 8
33006 Oviedo (Asturias) Ulf Hanefeld
Spain Delft University of Technology
Department of Biotechnology
Katharina Götz Biocatalysis and Organic Chemistry
Hamburg University of Technology Julianalaan 136
Institute of Technical Biocatalysis 2628 BL Delft
Denickestr. 15 The Netherlands
21073 Hamburg
Germany
XLVIII List of Contributors
Lutz Hilterhaus Nobuyuki Horinouchi

Technische Universität Hamburg- Kyoto University
Harburg Graduate School of Agriculture
Institut für Technische Biokatalyse Division of Applied Life Sciences
Denickestraße 15 Sakyo-ku
21073 Hamburg Kyoto 606-8502
Germany Japan
Yoshihiko Hirose Werner Hummel

Amano Enzyme Inc. Heinrich-Heine University Düsseldorf
Gifu R&D Center Institute of Molecular Enzyme
1-6, Technoplaza, Kakamigahara Technology
Gifu 509-0109 Research Centre Jülich
Japan Stetternicher Forst
52426 Jülich
Matthias Höhne Germany
Greifswald University
Institute of Biochemistry Ibrar Hussain
Department of Biotechnology & Stockholm University
Enzyme Catalysis Arrhenius Laboratory
Felix-Hausdorff-Str. 4 Department of Organic Chemistry
17487 Greifswald 10691 Stockholm
Germany Sweden
Frank Hollmann and

Delft University of Technology
Department of Biotechnology Scion
Biocatalysis and Organic Chemistry Te Papa Tipu Innovation Park
Julianalaan 136 49 Sala Street
2628 BL Delft Rotorua 3046
The Netherlands New Zealand
Kathrin Hölsch Hans-Dieter Jakubke

Toyama Prefectural University Leipzig University
Biotechnology Research Center and Department of Biochemistry
Department of Biotechnology Brüderstr. 34
5180 Kurokawa, Imizu 04103 Leipzig
Toyama 939-0398 Germany
Japan
and
Höntzschstr. 1a
01465 Langebrück
Germany
List of Contributors XLIX
Dick B. Janssen Iván Lavandera

University of Groningen University of Oviedo
Groningen Biomolecular Sciences and Faculty of Chemistry
Biotechnology Institute Department of Organic and Inorganic
Biochemical Laboratory Chemistry
Niejenborgh 4 Avenida Julián de Clavería 8
9747 AG Groningen 33006 Oviedo (Asturias)
The Netherlands Spain
Jianfeng Jin David J. Leak

Delft University of Technology Imperial College London
Department of Biotechnology Faculty of Natural Sciences
Biocatalysis and Organic Chemistry Division of Biology
Julianalaan 136 South Kensington Campus
2628 BL Delft London SW7 2AZ
The Netherlands UK
Bernard Kaptein Andreas Liese

DSM Innovative Synthesis B.V. Technische Universität Hamburg-
6160 MD Geleen Harburg
The Netherlands Institut für Technische Biokatalyse
Denickestraße 15
Romas J. Kazlauskas 21073 Hamburg
University of Minnesota Germany
Department of Biochemistry, Molecular
Biology & Biophysics Jeffrey H. Lutje Spelberg
1479 Gortner Avenue Tijmgaard 11
Saint Paul, MN 55108 6417HE Heerlen
USA The Netherlands
Arina V. Kozyr and

Russian Academy of Sciences
Shemyakin and Ovchinnikov Institute Julich Chiral Solutions GmbH
of Bioorganic Chemistry A Codexis Company
Ul. Miklukho-Maklaya 16/10 Prof.-Rehm Strasse 1
117871 Moscow 52428 Jülich
Russia Germany
Marina Kraußer Anke Matura

University of Erlangen-Nürnberg Technische Universität Dresden
Department of Chemistry and Allgemeine Biochemie
Pharmacy Bergstrasse 66
Henkestrasse 42 TU Dresden
91054 Erlangen D-01062 Dresden
Germany Germany
L List of Contributors
Marko D. Mihovilovic Monica Paravidino

Vienna University of Technology Delft University of Technology
Institute for Applied Synthetic Department of Biotechnology
Chemistry Biocatalysis and Organic Chemistry
Getreidemarkt 9 Julianalaan 136
1060 Vienna 2628 BL Delft
Austria The Netherlands
Daniela Monti Jürgen Pleiss

Istituto di Chimica del Riconoscimento University of Stuttgart
Molecolare, CNR Institute of Technical Biochemistry
Via Mario Bianco 9 Allmandring 31
20131 Milano 70569 Stuttgart
Italy Germany
Michael Müller Martina Pohl

Albert-Ludwigs University Freiburg Forschungszentrum Jülich GmbH
Institute of Pharmaceutical Sciences Institute of Bio- and Geosciences IBG-1:
Albertstr. 25 Biotechnology
79104 Freiburg Wilhelm-Johnen-Strasse
Germany 52425 Jülich
Germany
Timo Nuijens
DSM Innovative Synthesis B.V. Peter J.L.M. Quaedflieg
Urmonderbaan 22 DSM Innovative Synthesis B.V.
6160 MD Geleen Urmonderbaan 22
The Netherlands 6160 MD Geleen
The Netherlands
Jun Ogawa
Kyoto University Manfred T. Reetz
Graduate School of Agriculture Department of Synthetic Organic
Division of Applied Life Sciences Chemistry
Sakyo-ku Max-Planck-Institut für
Kyoto 606-8502 Kohlenforschung
Japan Kaiser-Wilhelm-Platz 1
45470 Mülheim an der Ruhr
Steffen Oßwald Germany
Evonik Degussa GmbH
Rodenbacher Chausee 4 Sergio Riva
63457 Hanau Istituto di Chimica del Riconoscimento
Germany Molecolare, CNR
Via Mario Bianco 9
20131 Milano
Italy
List of Contributors LI
David Rozzell and

Sustainable Chemistry Solutions
437 South Sparks Street University of Groningen
Burbank, CA 91506 Stratingh Institute for Chemistry
USA Niejenborgh 4
9747 AG Groningen
Sakayu Shimizu The Netherlands
Kyoto University
Graduate School of Agriculture Nicholas J. Turner
Division of Applied Life Sciences University of Manchester
Sakyo-ku School of Chemistry
Kyoto 606-8502 Manchester Interdisciplinary Biocentre
Japan 131 Princess Street
Manchester M1 7DN
Ivan V. Smirnov UK
Russian Academy of Sciences
Vlada B. Urlacher
Shemyakin and Ovchinnikov Institute
Heinrich-Heine University Düsseldorf
of Bioorganic Chemistry
Institute of Biochemistry
Ul. Miklukho-Maklaya 16/10
Universitätsstraße 1
117871 Moscow
40225 Düsseldorf
Russia
Germany
Theo Sonke Anne van den Wittenboer
DSM Innovative Synthesis B.V. Technische Universität Hamburg-
6160 MD Geleen Harburg
The Netherlands Institut für Technische Biokatalyse
Denickestraße 15
Jon D. Stewart 21073 Hamburg
University of Florida Germany
Department of Chemistry
102 Leigh Hall Teunie van Herk
Gainsville, FL 32611 Universiteit van Amsterdam
USA Van ’t Hoff Institute for Molecular
Sciences, Biocatalysis
Wiktor Szymanski Science Park 904
University of Groningen 1098 XH Amsterdam
Groningen Biomolecular Sciences and The Netherlands
Biotechnology Institute
Biochemical Laboratory Karl-Heinz van Pée
Niejenborgh 4 Technische Universität Dresden
9747 AG Groningen Allgemeine Biochemie
The Netherlands Bergstrasse 66
TU Dresden
D-01062 Dresden
Germany
LII List of Contributors
Ron Wever Ying Yin

Universiteit van Amsterdam Chinese Academy of Sciences
Van ’t Hoff Institute for Molecular Wuhan Institute of Virology
Sciences Xiao Hong Shan No. 44
Science Park 904 Wuhan 430071
1098 XH Amsterdam China
The Netherlands
Ruslan Yuryev
John M. Woodley Technische Universität Hamburg-
Technical University of Denmark Harburg
Center for Process Engineering and Institut für Technische Biokatalyse
Technology Denickestraße 15
Department of Chemical and 21073 Hamburg
Biochemical Engineering Germany
2800 Lyngby
Denmark Jun-Jie Zhang
Chinese Academy of Sciences
Bian Wu Wuhan Institute of Virology
University of Groningen Xiao Hong Shan No. 44
Groningen Biomolecular Sciences and Wuhan 430071
Biotechnology Institute China
Biochemical Laboratory
Niejenborgh 4 Ning-Yi Zhou
9747 AG Groningen Chinese Academy of Sciences
The Netherlands Wuhan Institute of Virology
Xiao Hong Shan No. 44
Alexander Yanenko Wuhan 430071
Institute of Genetics and Selection of China
Industrial Microorganisms
1st Dorozhny pr. 1
113545 Moscow
Russia
j1
Part I
Principles of Enzyme Catalysis
Enzyme Catalysis in Organic Synthesis, Third Edition. Edited by Karlheinz Drauz, Harald Gr€oger,
and Oliver May.
Ó 2012 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2012 by Wiley-VCH Verlag GmbH & Co. KGaA.
j3
1
Introduction – Principles and Historical Landmarks of Enzyme
Catalysis in Organic Synthesis
Harald Gr€oger and Yasuhisa Asano
1.1
General Remarks
Enzyme catalysis in organic synthesis – behind this term stands a technology that today
is widely recognized as a first choice opportunity in the preparation of a wide range
of chemical compounds. Notably, this is true not only for academic syntheses but
also for industrial-scale applications [1]. For numerous molecules the synthetic
routes based on enzyme catalysis have turned out to be competitive (and often
superior!) compared with classic chemical as well as chemocatalytic synthetic
approaches. Thus, enzymatic catalysis is increasingly recognized by organic
chemists in both academia and industry as an attractive synthetic tool besides the
traditional organic disciplines such as classic synthesis, metal catalysis, and
organocatalysis [2].
By means of enzymes a broad range of transformations relevant in organic
chemistry can be catalyzed, including, for example, redox reactions, carbon–carbon
bond forming reactions, and hydrolytic reactions. Nonetheless, for a long time
enzyme catalysis was not realized as a first choice option in organic synthesis.
Organic chemists did not use enzymes as catalysts for their envisioned syntheses
because of observed (or assumed) disadvantages such as narrow substrate range,
limited stability of enzymes under organic reaction conditions, low efficiency when
using wild-type strains, and diluted substrate and product solutions, thus leading to
non-satisfactory volumetric productivities. However, due to tremendous progress in
enzyme discovery, enzyme engineering, and process development, in recent years
numerous examples of organic syntheses with (tailor-made) enzymes have been
developed that avoid these disadvantages.
The achievements in microbiology and molecular biology have already led to a
broad range of widely applicable enzymes showing an excellent performance. Today
such enzymes are typically prepared in a highly attractive economic fashion by high-
cell density fermentation, and can be used in the form of tailor-made recombinant
whole-cell catalysts. This economically attractive access to highly efficient (bio-)
catalysts enables an excellent opportunity to realize the development of attractive
organic synthetic processes with enzymes as catalysts.
and Oliver May.
j 1 Introduction – Principles and Historical Landmarks of Enzyme Catalysis in Organic Synthesis
4
Benefiting from these achievements in microbiology and molecular biology,

organic chemists have applied these tailor-made biocatalysts (as isolated enzymes
or recombinant microorganisms, so-called designer cells) very successfully in a
broad range of organic syntheses. Many of those synthetic examples have been found
to be suitable even for industrial-scale productions and turned out to be superior to
competitive classic chemical or chemocatalytic approaches. In particular this is true
for the production of chiral compounds used as drug intermediates.
A further aspect that makes the research area enzyme catalysis in organic synthesis
both highly interesting and challenging from a scientific perspective is the high
interdisciplinarity of this field, which requires competencies from a broad variety of
disciplines, comprising, for example, microbiology, genetics, molecular biology,
organic synthesis, and reaction engineering. The hybridization of such compe-
tencies is certainly a key factor in the successful development of efficient biocatalytic
processes.
The following gives an overview of enzymes typically applied in organic synthesis
as well as some selected landmarks of the impressive development of enzyme catalysis
in organic synthesis towards a highly recognized synthetic technology in academia and
industry [3].
1.2
Potential of Enzymes as Catalysts in Organic Synthesis: Enzyme Reactions Overview
1.2.1
Enzyme Catalysts: Three-Dimensional Structure and General Properties
The unique functions of enzymes as catalytically active proteins are a result of their
complex three-dimensional structures and the active site integrated therein [4]. This
enables a highly specific recognition of specific substrates, leading to excellent
selectivities. Besides chemoselectivity, the stereoselectivity of enzymes is also in
general high to excellent and, furthermore, this is typically true for regio-, diastereo-,
as well as enantioselectivity. Figure 1.1 shows, as a representative example of the
impressive (and beautiful) three-dimensional structures of enzymes, the well known
and widely used lipase from Candida antarctica [5].
The unique properties of enzymes to stereoselectively recognize a substrate was
found by Fischer already at the end of the nineteenth century [6, 7]. Based on these
findings he postulated the lock-and-key theory, according to which the substrate
has to fit into the active site of the enzyme like a key into the lock. A further
theoretical milestone was the kinetic analysis of enzyme reactions conducted by
Michaelis and Menten a few years later [8]. Their theory is based on the formation of
an enzyme–substrate complex, and subsequent product formation and release of
the enzyme for the next catalytic cycle after the reaction has been conducted. In later
years this kinetic model has been further refined and today kinetic analysis [9] of
enzymatic reactions and characterization of the enzyme with such methods is a key
feature in biocatalytic research projects.
1.2 Potential of Enzymes as Catalysts in Organic Synthesis: Enzyme Reactions Overview j5
Figure 1.1 Three-dimensional structure of a lipase from Candida antarctica; image from the RCSB
PDB (www.pdb.org) of PDB ID 1LBS (Ref. [5])
In addition to the substrate the reaction conditions also play a very important role
in enzyme catalysis. It is difficult, though, to define properties under which in general
enzymes are able to operate as a catalyst. At the same time, however, it is evident that
enzyme catalysis requires specific suitable reaction conditions such as pH, temper-
ature, and solvent, which have to be considered in (bio-)process development. In the
following, some selected reaction parameters of specific importance for enzymatic
reactions are briefly discussed.
For enzymes pH and temperature are certainly highly important reaction para-
meters in terms of both activity and stability. Typically, enzymes operate in a more or
less neutral or weakly basic/acidic pH range, usually between pH 5 and 10, although
exceptions are known. The natural reaction environment for enzymes is water.
Interestingly, however, water as a reaction medium is not necessarily required and
many enzymatic transformations (including industrial processes) are run in organic
reaction medium. The factors affecting enzymatic reactions are described in more
detail below.
There are several ways to describe enzymatic activity; popular and widely used
criteria are the maximum reaction rate (vmax, measured, for example, in U mg1) and
the Km value.
1.2.2
Overview of Enzyme Classes (EC Numbers) and Related Reactions
Enzymes are typically classified according to the types of reactions they catalyze. In the
Enzyme Nomenclature classification [10] they are subdivided and categorized into six
main enzyme classes corresponding to the type of reactions such enzymes catalyze.
Table 1.1 gives an overview of this categorization, in particular the main enzyme
classes.
6
Table 1.1 Categorization of enzymes according to the general type of reactions they catalyze.
Enzyme class EC Selected reactions

number
Oxidoreductases 1 Reduction of C ¼ O and C ¼ C; reductive amination of C ¼ O;

oxidation of C-H, C ¼ C, C-N, and C-O; cofactor reduction/
oxidation
Transferases 2 Transfer of functional groups such as amino, acyl, phosphoryl,
methyl, glycosyl, nitro, and sulfur-containing groups
Hydrolases 3 Hydrolysis of esters, amides, lactones, lactams, epoxides,
nitriles, and so on, as well as the reverse reactions to form such
functionalities
Lyases (synthases) 4 Addition of small molecules to double bonds such as C ¼ C,
C ¼ N, and C ¼ O
Isomerases 5 Transformation of isomers (isomerizations) such as racemi-
zations, epimerizations, and rearrangement reactions
Ligases (synthetases) 6 Formation of complex compounds (in analogy to lyases) but
enzymatically active only when combined with ATP cleavage
With respect to applications of enzymes in organic synthesis, enzymes from nearly

all enzyme classes play an important synthetic role in organic chemistry. As an
exception, at least in part, one might regard enzymes from enzyme class 6 (ligases).
Since in situ regeneration of the cofactor ATP is still a challenge, ligases have found
limited use as catalysts for in vitro applications in organic syntheses. In contrast,
enzymes from enzyme classes EC 1–5 turned out to be highly efficient catalysts for a
broad range of organic synthetic transformations that, in part, are also suitable for
technical-scale applications.
With oxidoreductases (EC 1) many successful reduction and oxidation processes
have been realized.1) Scheme 1.1 summarizes selected oxidoreductase-catalyzed
reaction types that have gained broad interest in organic chemistry. With respect to
(asymmetric) reductions as a synthetically important reaction in organic chemistry,
the reduction of a carbonyl moiety to an alcohol (when using, for example, alcohol
dehydrogenases or a-hydroxy acid dehydrogenases as catalysts) or amino function-
ality (when using a-amino acid dehydrogenases in reductive aminations) has already
found a wide application range in organic chemistry as well as industrial applications.
A more recent trend is the increasing tendency to apply enzymes also for C¼C double
bond reductions. Although pioneering work in this area with so-called old yellow
enzymes was carried out many decades ago, expansion of the synthetic range as well
as the pool of available (robust and stable) enzymes has been a main focus of recent
research [11a]. Notably, commercialization of this technology also has been reported
recently [11b]. A further recent and current hot topic is the field of oxidation
reactions using suitable oxidoreductases. Key advantages of using enzymes as
1) Enzymatic organic syntheses with oxidoreductases, both academic and industrial contributions, are
covered in detail in, for example, Chapters 26–38.
O OH
O 1 2 1 NH2
R R R R2
R CO2H R CO2H
R2 R2
EWG oxidoreductases EWG
R1 (E.C. 1) R1
(EWG: electron
withdrawing group) OH
R1 R2 R1 R2
O
Ph
Ph
Scheme 1.1 Overview of selected reactions catalyzed by enzymes from EC 1 (oxidoreductases).
catalysts in redox processes are (i) the excellent selectivity even when using non-
functionalized compounds, as, for example, demonstrated for alkanes and cycloalk-
anes as substrates and (ii) the use of molecular oxygen as a cheap and sustainable
oxidizing agent. Oxidative selective functionalization of alkane moieties is still a
challenge for chemocatalysts and a limited number of efficient catalysts exist for such
transformations. In contrast there is a high demand for this reaction type, also from
an industrial perspective. Given the excellent stereoselectivity of enzymes, it is no
surprise that today hydroxylation of steroids is industrially carried out by means of
biocatalytic hydroxylation instead of chemical methods [12]. A remaining challenge
for enzymatic oxidations, however, can be seen in the limited activity of some
enzymes, which is often below 1 U mg1, such as, for example, in case of P450-
monooxygenases. As well as hydroxylation other oxidative processes with enzymes
are also of interest in organic syntheses, such as, for example, reactions with
Baeyer–Villiger monooxygenases (for Baeyer–Villiger oxidations leading to lactones
from ketones) and styrene monooxygenases (for epoxidation of styrenes). In sum-
mary, oxidoreductases are the second most used enzyme types in organic synthesis;
only the representatives of enzyme class EC 3 show more synthetic applications.
Representatives of enzyme class EC 2, so-called transferases, are further versatile
catalysts for organic synthetic transformations.2) In particular, transaminases have
attracted widespread attention with interesting applications for the synthesis of
amino acids and amines. Industrial applications have been reported as well. As a
starting material the corresponding carbonyl compounds are required. Scheme 1.2
gives an overview of reactions with transaminases and other transferases.
Without doubt, the most popular and most frequently applied enzymes in organic
chemistry are hydrolases (EC 3).3) In particular this is due to (i) the fact that many
hydrolases are commercially available, often in an attractive price range (e.g., in the
2) Enzymatic organic syntheses with transferases, both academic and industrial contributions, are
covered in detail in, for example, Chapters 19 and 20.
3) Enzymatic organic syntheses with hydrolases, both academic and industrial contributions, are
covered in detail in, for example, Chapters 8–10, 12, 14–17, 20, and 25.
8
O NH2
O NH2
R1 R2 R1 R2
R CO2H R CO2H
O OH
transferases R2
1
R1 (E.C. 2) R
O
2-
PO3
OH O
R1 CH3 R1
Ar OH Ar O
Scheme 1.2 Overview of selected reactions catalyzed by enzymes from EC 2 (transferases).
case of applications in the food and laundry detergent industries) [13], (ii) their
direct and often simple use without the need for, additional cofactor and cofactor
regeneration methods, (iii) numerous synthetic applications through modifications
of the carboxyl moiety, for example, in resolution and desymmetrization processes,
and (iv) their suitability (at least in part, particularly in the case of lipases) to use
hydrolases in an organic solvent as a reaction medium, which is often favored by
organic chemists [14]. Representative examples for hydrolases frequently used in
organic synthesis are proteases, lipases, and esterases. Typical transformations
include the hydrolysis of esters and amides, and their reverse reactions, namely
esterification and amidation. Scheme 1.3 gives an overview of selected examples of
hydrolase-catalyzed (stereoselective) processes. Often, hydrolases are used in res-
olution processes since many acids and functionalized derivatives thereof are easily
accessible in racemic form by means of standard chemical methods. A further
X R3 XH
rac
R2
rac R1 R2 R1 R2 R2
XR3 (X=O,NH)
R1 OH
R1
(X=O,NH) O
O
R rac
O
NH2
NH hydrolases
HN R CO2H
(E.C. 3)
O
CO2R3 R2
CO2H
R2
R1 O R1
CO2R3 XH CO2R3
rac
X R3
R1 R2
(X=O,NH) R1 R2
Scheme 1.3 Overview of selected reactions catalyzed by enzymes from EC 3 (hydrolases).

stereoselective process type of interest are desymmetrization reactions, for example,
of prochiral or meso-type diesters. Furthermore, non-stereoselective applications
have been reported as well, for example, cleavage of unwanted acid side-chains
under smooth hydrolytic conditions, thus avoiding the harsh chemical reaction
conditions of alternative chemical hydrolytic processes. This option has been
particularly used for derivatization of easily accessible natural products (e.g., in
the synthesis of antibiotics such as 6-aminopenicillinic acid). To date, hydrolases
have also attracted a lot of industrial interest and numerous examples of technical
applications of hydrolases have been demonstrated. Notably, representatives of this
enzyme class, in particular lipases, are also suitable for reactions in pure organic
media and this technology is used, e.g., for the large-scale production of fatty acid
esters starting from the fatty acid and an alcohol moiety [15–17].
Lyases, which are summarized in enzyme class EC 4, are characterized by the
formation of a new C(Nu)-XH bond (X¼C,N,O; Nu-: nucleophile) by an addition
reaction of a small molecule Nu-H to a C¼X double bond.4) Such a small molecule
can be, for example, water, ammonia, as well as a carbon nucleophile. Enzymes for
addition of water to activated C¼C double bonds are called hydratases, and ammonia
lyases are enzymes capable of adding ammonia to an enoate (typically in a highly
enantioselective fashion, albeit the substrate spectrum often is narrow). In addition,
the use of carbon nucleophiles enables the (stereoselective) formation of new CC
single bonds. Common carbon nucleophiles such as, for example, cyanide in hydro-
cyanation of aldehydes as well as aldehydes in umpolung reactions such as the benzoin
condensation have been widely used. Scheme 1.4 gives an overview of typical (selected)
reactions based on catalysis with lyases. Although the overall number of reactions with
lyases used in organic synthesis up to now is still not very broad (when compared with
the huge number of known classic organic CC bond formation reactions), those
O OH
O O + HCN OH
R R CN
+ R2
R1 R2 R1
O
CO2H
lyases R
CO2H
R (E.C. 4) XH
(X=O,NH)
O
O OH O
+ R2
R1 O OH R1 R2
OH
+ CO2H CO2H OH
R R
NH2 NH2
Scheme 1.4 Overview of selected reactions catalyzed by enzymes from EC 4 (lyases).
4) Enzymatic organic syntheses with lyases, both academic and industrial contributions, are covered in
detail in, for example, Chapters 11–13, 18, 21, and 24.
10
O O
R2 NH R2 NH
OH OH
R1 R1
NH2 O O NH2
NH2 NH2
R R
O O
R O R O
isomerases
HN NH (E.C. 5) HN NH
O O
OH O
H
H H
R 2 NH2 R
R
O R1 OH
R2
NH2 R1
Scheme 1.5 Overview of selected reactions catalyzed by enzymes from EC 5 (isomerases).
reactions known can be carried out highly efficiently. This is underlined by, for
example, technical applications of addition reactions of water and ammonia to enoates
as well as hydrocyanation and umpolung reactions.
Enzyme class EC 5 consists of those enzymes capable of catalyzing isomerization
reactions.5) The types of isomerizations are diverse, consisting of, for example,
racemizations, 1,2-migrations of functional groups (e.g., of amino functionalities)
and cis–trans isomerizations. Scheme 1.5 gives an overview of selected isomerization
catalyzed processes employed in organic chemistry. Interestingly, the largest bio-
catalytic application today is based on the use of an isomerase, namely, the production
of high fructose corn syrup via enzymatic transformation of glucose into fructose [18],
which is carried out on a >1 million tons scale. In organic chemistry, the use of
racemases has attracted most interest within the enzymes of EC 5, since the
combination of a racemase with a further biocatalyst for a resolution step enables
the development of dynamic kinetic resolution processes. Typically, such resolution
processes to be combined with racemases are reactions catalyzed by hydrolases, and
such resolutions run either in the hydrolytic or acylation direction.
Whereas enzymes from enzyme classes EC 1 to EC 5 are already widely used as
catalysts in organic synthesis and have enabled a broad range of highly efficient
synthetic processes (running, in part, already even on an industrial scale), the
application range of enzymes from EC 6 (ligases) is sill narrow. At first glance this
might sound surprising due to the numerous interesting reaction types these
enzymes can catalyze. However, these reactions require ATP as a cofactor, which
is efficiently regenerated in living cell processes, but its cofactor regeneration in situ
under in vitro reaction conditions remains a challenge. Although some methods
have been developed, applicability in organic syntheses (in particular with respect to
5) Enzymatic organic syntheses with isomerases, both academic and industrial contributions, are
covered in detail in, for example, Chapters 39 and 40.
large-scale processes) is still limited. Certainly, development of efficient and
economically attractive methods for the in situ regeneration of ATP is an attractive
goal in future research activities.
1.2.3
Overview of Coenzymes and Cofactors and Applications in Organic Synthesis
Cofactors are non-proteinogenic compounds that are required for the catalytic activity
of enzymes and which can bind to the enzyme either in a covalent or non-covalent
mode [4, 9]. A broad variety of cofactors are known, consisting of organic molecules and
inorganic ions. In the covalent mode, when the cofactor is permanently bound to the
enzyme, the cofactor is called a prosthetic group. In case of a non-covalent binding of
the cofactor to the enzyme it is called a coenzyme. Since the coenzyme is modified
during the catalytic process (by transferring electrons or chemical groups to the
substrate), its regeneration in a subsequent reaction is a key issue in order to use
the cofactor in catalytic amounts. Thus, the cosubstrate required for the cofactors
regeneration is needed in stoichiometric amounts. Figure 1.2 shows selected cofactors
that are often applied in organic synthetic processes with enzymes.
With exception of hydrolases (EC 3), members of all other enzyme classes (or at
least a part thereof) show a cofactor dependency, although in some cases (e.g., in case
of lyases) cofactors are not necessarily involved in the catalytic process. For most
enzymes belonging to enzyme classes EC 1–5, however, cofactors are involved in the
H H NH2 NH2
CONH2
N N
N N
N N N N N
O O O O O
OH OH
O P O P O HO P O P O P O
OH OH O HO OH OH O
O
OH OH OH OH
NADH ATP
O
H3 C N
NH
NH2 CH3
H3 C N N O
N N
OH
H3C N S O OH
OH P O OH
HO O P OH
ThDP O
O
P OH
O OH
FMN
Figure 1.2 Cofactors often applied in enzymatic organic synthesis.

12
catalytic process. Furthermore, the availability of an opportunity to regenerate such

cofactors efficiently under the chosen organic reaction conditions often decides
whether this method can be developed towards an attractive synthetic process.
In the following this shall be exemplified for the regeneration of a cofactor in its
reduced and oxidized form, namely, NAD(P)H and NAD(P) þ , which are used in
enzymatic redox processes (Scheme 1.6) [19]. Since the reducing agent for the
oxidoreductase in a reduction process is NAD(P)H, in total a stoichiometric amount
of such a reducing agent is need (as in classic organic chemistry a stoichiometric
amount of molecular hydrogen or borane or sodium borohydride is used). Taking
into account both the very high molecular weight and price of cofactors such as NAD
(P)H, the use of a stoichiometric amount of such molecules would not enable any
synthetically useful process. Thus, in situ cofactor regeneration of such cofactors,
enabling their use in catalytic amounts, is a prerequisite to conduct such biocatalytic
redox processes in a synthetically useful and attractive fashion. Such in situ cofactor
regeneration can be achieved through combination with a second enzymatic trans-
(a) "Reductive" cofactor recycling mode based on the use of a formate dehydrogenase
OH OH
formate NAD(P)+ or
R1 R2 R1 R2
formate alcohol
dehydrogenase dehydrogenase
O
CO2
NAD(P)H
R1 R2
(b) "Oxidative" cofactor recycling mode based on the use of an NAD(P)H-oxidase

O
O2 NAD(P)H R 1
R2
NAD(P)H- alcohol
oxidase dehydrogenase
OH OH
H2O or + 1 2 or 1
H2O2 NAD(P) R R R R2
Scheme 1.6 Selected regeneration processes of cofactors in their (a) reduced and (b) oxidized
form, exemplified for NAD(P)H and NAD(P)þ .
formation, which regenerates the cofactor. To make the cofactor regeneration
economically attractive it is important that the substrate consumed in this second
enzymatic process is cheap and readily available, since this substrate (the so-called
cosubstrate) is required in stoichiometric amount. From the perspective of the
reaction formula, this cosubstrate represents the reducing or oxidizing agent
required in stoichiometric amount. For example, in the selected cofactor regener-
ation methods shown in Scheme 1.6, the stoichiometric reducing agent is formate
(which is oxidized to carbon dioxide; process 1) and the stoichiometric oxidizing
agent is molecular oxygen (which is reduced to water; process 2).
To date, a broad set of cofactor regeneration methods have been successfully
developed. Notably, besides enzymatic cofactor regenerations, electrochemical and
chemocatalytic cofactor regenerations have also been reported. Processes with in situ
cofactor regenerations can be conducted using isolated enzymes as well as permea-
bilized whole-cell catalysts. In addition, in fermentation-like processes with intact
microorganisms cofactor regeneration is carried out within the metabolism in the
cell. In organic chemistry, all of these options for in situ regeneration of cofactors have
been realized in enzymatic syntheses with cofactors.
1.2.4
Factors Affecting Enzymatic Reactions
As with chemocatalysts, enzymes also have a typical application range with respect to
reaction parameters, which have to be considered in those transformations. These
typical reaction parameters are in general related to physiological conditions under
which the corresponding enzymes work. In particular, the pH and temperature
profile of enzymes should be determined prior to use in organic synthesis. For most
enzymes a pH in the range 6–10 and temperatures of 20–50 C are preferred although
many exceptions are known. Notably, enzymes suitable for catalyzing the reactions at
very low or high pH and also at elevated temperature exceeding 80 C have been
found. A typical natural source of such types of enzymes is the group of so-called
extremophilic microorganisms. For example, thermophilic microbial strains from
hot springs are an interesting source of enzymes that can be impressively active at
high temperatures.
A further important criteria when setting up an organic synthesis with (bio-)
catalysts is the choice of reaction medium. The preferred reaction media for
enzymes – when taking into account their natural function – are aqueous (buffered)
solutions. However, notably, many enzymes are highly tolerant towards the pres-
ence of organic solvents [14]. This has been demonstrated in particular for lipases as
catalysts. The reaction medium of choice for most of enzymes is nevertheless water
(or related buffer solutions). Since, however, organic substrates are often hydro-
phobic, water-miscible and water-immiscible organic solvents have been added in
biotransformations to ensure sufficient solubility of the substrate. Notably, many
enzymes turned out to be stable under such conditions, thus allowing the devel-
opment of efficient processes in aqueous-organic one-phase or two-phase solvent
systems.
14
1.2.5
Why Use Enzymes in Organic Synthesis? Factors Affecting Enzymatic Reactions,
Advantages and Drawbacks
Before discussing the advantages and drawbacks when using enzymes as catalysts in
organic synthesis, a brief overview is given of selected criteria for choosing a specific
synthetic route (based on, for example, chemo- or biocatalysts or classic resolutions to
form diastereomeric salt pairs). Scheme 1.7 summarizes selected criteria that are
relevant also for biotransformations. In general high conversion and enantioselec-
tivity (and/or regioselectivity/diastereoselectivity) are desirable. A high, ideally
quantitative (product-related) conversion has not only the advantage of consuming
the maximal amount of substrate (thus contributing to a decrease in substrate costs)
but also simplifies downstream-processing. This is particularly true for reactions in
which substrate and product show similar properties, for example, similar boiling
points, which make separation tedious. With respect to enantioselectivity, typically a
high enantiomeric excess of >99% e.e. (as required from the FDA for chiral drugs)
for the resulting product is desirable. Besides conversion and selectivity issues,
substrate and product concentrations as well as volumetric productivities are further
important criteria for (bio-)transformations in organic chemistry. As a rule of
thumb, a substrate input of >100 gl1 (in total, optional added in portions) is
desirable to reach economically attractive volumetric productivities in technical
production processes. Among further important criteria for realizing an efficient
synthetic process are an attractive access to the (bio-)catalyst component and the
technical feasibility of the process.
Scheme 1.7 Criteria for efficient (bio-)transformations in organic synthesis.
A major advantage of enzymes as catalysts in organic synthesis, which is often

regarded as the major advantage of biocatalysts, are the excellent selectivities
enzymatic reactions typically show. For numerous asymmetric reactions starting
from prochiral compounds the desired products are formed in excellent enantio-
meric excess of >99% e.e. High enantioselectivities are also typically observed in
enzymatic resolution processes, with enantioselectivities often exceeding E-values of
100. High to excellent stereoselectivities have also been observed in regio- and
diastereoselective enzymatic reactions, respectively. Even when unsatisfactory
stereoselectivities are observed for wild-type enzymes, several protein engineering
methodologies are available that have already turned out to be suitable for optimi-
zation of enzymatic performance in many examples.
Whereas chirality and a defined absolute configuration is also an important
criterion in nature, high volumetric productivities and a substrate input of, for
example, >100 gl 1 is a desirable feature for organic syntheses, but not for processes
in living organisms. Thus, unsurprisingly, for many enzymes the development of
synthetic processes running at a high substrate input and leading to a high
space–time yield turned out to be a challenging task (but at the same time it should
be added that many enzymes turned out to be able to do so efficiently!).
In addition, optimization of the specific activity of enzymes by means of protein
engineering is a further challenge in order to make the biocatalyst attractive for
synthetic purposes. To be suitable for organic syntheses, in general (as a rule of
thumb) specific activities exceeding 1 U mg1 are desirable. However, it also should
be mentioned that numerous enzymes show as wild-type enzymes excellent activity
data exceeding 100 U mg1 for specific substrates. Needless to say, such enzymes are
highly interesting catalysts, fulfilling a key prerequisite towards realizing technically
feasible and economically attractive biocatalytic processes.
Besides the specific activity, a further key feature for attractive biocatalysts is an
economic production method for their preparation. Certainly, this criterion has long
been a limiting factor and drawback in enzymatic chemistry when regarding the
decade long biotransformations with wild-type organisms. The use of wild-type
organisms has major drawbacks, for example, because the expression of the desired
protein is very low, thus requiring a large amount of biomass for the biotransformation.
Often substrate loading is below 1 g l1, accompanied by a high biomass loading of
>25 g l1 in such biotransformations (as can be seen, for example, in many reactions
with bakers yeast as a biocatalyst). A further consequence of low protein expression of
the desired enzyme in wild-type microorganisms is the significant impact of side-
reactions. Owing to impressive advances in molecular biology related to protein
engineering [20] today enzymes can be (mostly) made available in recombinant form.
Overexpression in host organisms such as, for example, Escherichia coli often exceeds
20%, which not only contributes to a high biocatalyst amount in the recombinant
whole-cell but also to suppressing side reactions due to the favored ratio of desired
enzyme over other enzymes catalyzing for competing side-reactions.
Furthermore, tremendous advances in bioprocess engineering have been made
that allows the production of such recombinant whole-cells with impressive biomass
concentrations. Thus, by means of high-cell density techniques biomass concentra-
tions of >200 g per litre of fermentation broth can now be reached for the required
recombinant cells bearing the desired enzyme in overexpressed form. Accordingly,
overexpression jointly with high biomass concentration in the fermentation process
16j 1 Introduction – Principles and Historical Landmarks of Enzyme Catalysis in Organic Synthesis
represents a valuable access to the (bio-)catalyst. When these criteria are fulfilled such
an economically highly attractive access to the biocatalyst is certainly a major
advantage.
Dependent on the process, such recombinant whole-cells can be used directly in
the biotransformation or, alternatively, after cell disruption and optional purification
the enzymes are used as biocatalyst component in free or immobilized form.
A further major advantage of biocatalysis over chemocatalysis is the possibility of
using substrates often without the need for protecting groups, which is due to the
high selectivity of enzymes for specific functional groups. This is underlined by two
examples in asymmetric synthesis, which are visualized in Schemes 1.8 and 1.9:
NH2
HCO2- NAD+ R CO2H
L-amino acid
formate
dehydrogenase,
dehydrogenase
ammonia
no preformation of
stable imine required
O no protection of acid
moiety required
CO2 NADH R CO2H
Scheme 1.8 Protecting group-free biocatalytic strategy for reductive amination.
(a) "Classic" multi-step approach using a chemocatalyst

:
O
R
1. esterification + H cleavage of both
O 2. imine O OH protecting OH
R groups R
formation CO2R CO2H
OH OR
O
NH2 N Ph chiral N Ph NH2
synthetic -
Ph Ph
Ph catalyst Ph mixture of
- ROH L-syn- / L-anti-
amino acid
(b) One-step biocatalytic approach:
O
R
+ H
O OH
R
CO2H
OH
NH2 L-enantioselective NH2
threonine aldolase,
PLP-dependent mixture of
L-syn- / L-anti-
amino acid
Scheme 1.9 (a) Multistep chemical and (b) protecting group-free biocatalytic strategies in an aldol
reaction [21].
1.3 The Early Steps: From Fermentation to Biotransformations Using Wild-Type Whole Cells j17
The first example is the biocatalytic reductive amination of a-keto acids (Scheme 1.8;
see also Chapter 28). Notably, this substrate can be used directly in the enzymatic
reductive amination process without the need to either protect the carboxylate moiety
as an ester or to prepare a stable imine prior to the reduction step. Such additional
steps might have to be considered when using a potential classic chemical process as
an alternative (based on, for example, the formation of an imine with a chiral auxiliary
and a diastereoselective reduction, followed by further steps for protecting group
cleavage).
The second comparison reflects the situation for an aldol reaction for the
asymmetric preparation of b-hydroxy a-amino acids (Scheme 1.9; [21]; also Chapter
21). Whereas in the biocatalytic step glycine can be used directly as a donor, in a
chemocatalytic reaction totally protected glycine is required. This requires two steps
prior to the asymmetric chemocatalytic key step, conducted, for example, by means of
a phase-transfer catalyst or a metal catalyst, as well as two cleavage steps after the
reactions. Thus, biocatalysis offers a straightforward access, requiring only one
synthetic step compared to five steps in the chemical approach [21]. In this case,
however, a challenge for biocatalysis is still the limited diasteresoelectivity of the
process, whereas enantioselectivity is excellent.
These examples demonstrate that biocatalysis offers many unique advantages over
chemical alternatives, thus representing an exciting complementary alternative to the
pool of classic chemical and chemocatalytic synthetic methods available to the
organic chemist today.
1.3
The Early Steps: From Fermentation to Biotransformations Using Wild-Type
Whole Cells
1.3.1
Historical Development of Fermentation and First Microbial Transformations
The finding of the production of ethanol from glucose is regarded as one of the first
discoveries of human beings in the field of biotransformations. In such processes
whole cells of microorganisms such as the yeasts Saccharomyces cerevisiae and
Schizosaccharomyces pombe and the bacterial producer Zymomonas mobilis were used.
In addition, the use of a-amylase for saccharification of starch began at least 5000
years ago in Mesopotamia or Egypt, applying this process for the production of beer.
Another earlier enzyme-catalyzed process is the production of cheese from milk
when kept in the stomach of sheep. Although the presence and function of this
enzyme chymosin was not understood at that time, people could use it for this
application. Another early product produced by biotechnology is vinegar (acetic acid),
which is produced by oxidation of ethanol by Acetobacter aceti or Gluconobacter
suboxydans [22]. Acetic acid has been produced at the surface of static cultures of
Acetobacter aceti or Gluconobacter suboxydans. Acetobacter has been also used in a
large-scale production of acetic acid in an aerobic reaction tower filled with a number
18
Table 1.2 Milestones in the history of applied microbiology.
Age (year) Development in microbiology Typical products
Prehistoric Alcohol
Middle Ages Vinegar, pickles, miso, soy sauce
First half of the Surface culture Organic acids
twentieth century
Static culture Acetone, n-butanol
Second half of the Aerobic culture Penicillin
twentieth century
1955 Screening Antibiotics
Biochemical auxotrophs Amino acids, nucleic acids
1965 Analog resistant mutants
1975 Screening for enzymes Steroids
Microbial transformation Natural and unnatural amino acids
Screening for substrates
Petroleum microorganisms Long-chain dicarboxylic acids
1985 C1 microorganisms Microbial protein
Recombinant DNA Interferon, human growth hormone,
and so on
Cell fusion Interleukin
Bioreactor Acrylamide
1995 Genome science Various
of different packing materials such as ceramics, hollow fibers, charcoal pellets and so
on shavings, onto which ethanol is sprayed. Table 1.2 gives an overview of milestones
in the history of applied microbiology.
The historical development of enzymatic synthesis is greatly related to the progress
of microbiology, because microorganisms have been the main sources of the
enzymes. One of the first questions to be asked was what does the mechanism of
alcoholic fermentation looks like. It is no exaggeration to say that biochemistry was
born to answer such questions. Notably, there was a big paradigm shift from anaerobic
culture to aerobic culture around middle of the twentieth century, caused by an
engineering development to cultivate the microorganisms aerobically by shaking
cultures and aerobic bioreactors, resulting in the discovery of varieties of new abilities
of wild-type microorganisms, enabling large-scale production of the products, as
compared with the static cultures that had only produced beer, sake, vinegar, yogurt,
miso, pickles, and so on. Another big lesson in the history of applied microbiology is
the notion of screening or microbial diversity, as evidenced by the fact that wide
varieties of new antibiotics were isolated by changing the microbial producers, and
also in the way some biochemists always start their research by finding the best
producers of the enzyme, to make the purification and characterization of the enzyme
much easier. Over the years, there have been many successful examples of microbial
biotransformations (Table 1.3). Some selected processes thereof are described in more
detail in the following to underline the potential and (early) industrial achievements of
microbial biotransformations with wild-type strains.
Table 1.3 Overview of selected milestones in industrial microbial biotransformations with wild-type
strains.
Product Biocatalyst Operating Company

since
Acetic acid Bacteria 1823 Various

L-2-Methylamino-1- Yeast 1930 Knoll, Germany
phenylpropan-1-ol
(ephedrine)
L-Sorbose Acetobacter suboxydans 1934 Various
Prednisolone Arthrobacter simplex 1955 Schering, Germany
L-Aspartic acid Escherichia coli 1958 Tanabe, Japan
7-ADCA Bacillus megaterium 1970 Asahi Chemical, Japan
L-Malic acid Brevibacterium ammoniagenes 1974 Tanabe, Japan
D-p-Hydroxy- Pseudomonas striata 1983 Kaneka, Japan
phenylglycine
Acrylamide Rhodococcus sp. 1985 Nitto (Mitsubishi
Rayon), Japan
D-Aspartic acid and Pseudomonas chlororaphis and 1988 Tanabe, Japan
L-alanine Pseudomonas dacunhae
L-Carnitine Agrobacterium sp. 1993 Lonza, Switzerland
2-Keto-L-gluconic acid Acetobacter sp. 1999 BASF, Germany
1.3.2
Development of Practical Synthesis of Chemicals via Transformations Using
Wild-Type Whole Cells in Non-Immobilized Form
An early example of a successful application of microbial-based biotransformations is

the synthesis of an intermediate for L-ephedrine ((1S,2S)-2-methylamino-1-phenyl-
propan-1-ol), established in the 1930s industrially by Knoll AG (Scheme 1.10) [23, 24].
The compound L-ephedrine and its diastereomer, pseudoephedrine, are pharma-
ceuticals used as decongestants and anti-asthmatics. The biocatalytically synthesized
intermediates by means of microorganism are L-phenylacetylcarbinol and its D-
enantiomer. Their synthesis is based on a condensation of an active acetaldehyde
derived from pyruvic acid and externally added benzaldehyde, and as microorgan-
isms yeasts such as Saccharomyces cerevisiae and Candida utilis have been used. This
type production technology, is which still applied today, is shown schematically in
Scheme 1.10.
O O pyruvate decarboxylase OH 1. +H2, Pt OH

CH3 from bakerýeast CH3 2. +CH3NH2 CH3
H + HO
O - CO2 O NHCH3
(in situ-formed
enzymatically) L-ephedrine
Scheme 1.10 Microbial biotransformation as a key step in the production of L-ephedrine.

20
1.3.3
Development of Practical Synthesis of Chemicals via Transformations Using
Wild-Type Whole Cells in Immobilized Form
Development of practical synthesis of chemicals via transformations using wild-type

whole cells has often been achieved based on the concept of treating microbial cells as
just a bag of enzymes for the use in organic synthesis. This research field led to
tremendous synthetic applications, often applied on an industrial scale [1], in
particular when immobilized microorganisms were used as biocatalyst. This tech-
nology of using (immobilized) whole microbial cells was made possible by combin-
ing the notions of screening unknown microorganisms, finding new enzymes from
nature, induction phenomenon of the enzymes, the uses of precursors or analogues
as substrates, the development of immobilization of whole microbial cells, and
biochemical engineering technologies.
One of the most successful achievements in biocatalytic organic synthesis is
the transformation of acrylonitrile into acrylamide catalyzed by a nitrile hydra-
tase [25, 26]. This biocatalytic transformation, running in the presence of an
immobilized microbial catalyst, led to an industrial production process that is
now estimated to produce more than 400 000 tons of acrylamide annually
worldwide. Notably, when the enzymes were fully induced in the microbial
cultures, as in the case of nitrile hydratase, they consisted of up to 50% of the
protein of the whole cells of Rhodococcus rhodochrous J-1. Scheme 1.11 shows
this process concept.
whole-cell biocatalyst
containing a
nitrile hydratase
CN + H2O CONH2
acrylonitrile aqueous reaction acrylamide
medium
Scheme 1.11 Nitrile hydratase-catalyzed transformation of acrylonitrile into acrylamide.
The historical development of this microbial process is also of interest since it

underlines that careful identification of metabolic steps and identification of the
individual enzymes involved has been required to develop an efficient, selective
biotransformation for acrylonitrile hydration: Based on earlier work, studies on
microbial degradation of polyacrylonitrile was started and further extended to
screening for microorganisms that degrade various low molecular weight nitrile
compounds. The microbial hydrolysis of nitriles was found to be catalyzed by two
enzymes, nitrile hydratase and amidase, or by a single enzyme nitrilase. Further-
more, the ability of microbial enzyme to synthesize acrylamide in a very high
concentration was discovered for the first time. By an enrichment culture tech-
nique, Rhodococcus rhodochrous (formerly Arthrobacer sp.) J-1 and Pseudomonas
chlororaphis B23, which have both become industrial strains, were isolated as
acetonitrile and isobutyronitrile utilizers, respectively. P. chlororaphis B23 was
found to catalyze the synthesis of acrylamide, forming up to 400 gl1 of acrylamide
from acrylonitrile [27]. From R. rhodochrous J-1, nitrile hydratase was discovered,
purified, and characterized.
Since only proteinogenic amino acids are produced by fermentation, attempts
were made at an early stage to produce non-proteinogenic amino acids also by means
of biotransformations with microorganisms starting from chemically synthesized
substrates. An elegant example of using (immobilized) wild-type microorganisms for
the production of non-proteinogenic amino acids is the biocatalytic synthesis of the
semisynthetic b-lactam side-chain D-p-hydroxyphenylglycine [28, 29]. This com-
pound is required for the preparation of amoxicillin, which is one of the commercially
most successful semisynthetic antibiotics and has a structure that consists of an
amide compound based on D-p-hydroxyphenylglycine and 6-APA (6-aminopenicil-
lanic acid). When the action of enzyme with the whole cells of microorganisms were
screened with substituted hydantoin compounds as substrates, D-stereoselective
hydrolysis of the hydantoins was, surprisingly, observed together with a simulta-
neous racemization of the substrate in slightly alkaline pH, thus enabling dynamic
kinetic resolution of the synthetic substituted hydantoins to form, for example,
N-carbamoyl D-p-hydroxyphenylglycine (Scheme 1.12). The enzyme responsible for
the D-stereoselective hydrolysis of the hydantoin was characterized as a hydantoinase
(dihydropyrimidinase). The product of the reaction, namely, N-carbamoyl D-p-hydro-
xyphenylglycine, can then be decarbamylated chemically (or as an alternative
enzymatically) to form D-p-hydroxyphenylglycine. The production process for this
D-amino acid was established at Kanegafuchi Chemical (now: Kaneka Corporation)
and runs highly enantioselectively in the presence of immobilized whole-cells of a
Bacillus brevis strain (Scheme 1.12). Notably, quantitative conversion is achieved, and
the production volume has been in the range 300–700 tons annually. The conden-
sation reaction of D-p-hydroxyphenylglycine, obtained after further cleavage of the N-
carbamoyl moiety by chemical treatment with sodium nitrite, and 6-APA can be
carried out by enzymatic process using a penicillin acylase from Klebsiella citrophila.
HO wild-type
HO
whole-cell catalyst,
containing classic HO
rac
O D-hydantoinase chemical
CO2H deprotection
HN NH HN CO2H
+ H2O NH2 H2N
O O
D-p-hydroxy
phenylglycine
Scheme 1.12 Microbial production of N-carbamoyl D-p-hydroxyphenylglycine from the

corresponding racemic hydantoin and subsequent classic chemical deprotection.
In summary, by means of (non-genetically engineered) wild-type strains a range of

efficient microbial transformations have been realized that, in part, turned out to be
suitable for industrial application.
22
1.4
Chemical Processes with Isolated Enzymes: The Impact of Process Engineering
1.4.1
Historical Development of Transformations with Isolated Enzymes
The historical development of biotransformations in organic synthesis has been

mainly driven from two perspectives. The first is based on earlier available fermen-
tation methods applied in food industry and this experience was then used for the
production of chemicals. Such microbial processes and their historical development
towards biotransformations for synthetic purpose are described and summarized in
the previous section. Notably, based on these earlier biotechnological production
processes, which have been established, an awareness of the occurrence and function
of enzymes arose around the nineteenth century. In 1811, Kirchhoff found that
maltose is formed from starch by the action of barley extract [30], and in 1833 Payen
and Persoz named the ethanol precipitate active toward starch as diastase, which
means separation in Greek [31]. In1836, Schwann named a substance in the
human stomach as pepsin, which hydrolyzes meat [32]. 1860, Berthelot found an
invertase activity in the cell-free extract of yeasts. The term enzyme was coined in
1876, by K€ uhne, meaning in (en) the yeast (zyme) [33].
Thus, besides the microbial-based biotransformation based on an increasing
understanding of the occurrence and function of enzymes as molecular catalysts,
a second trend began with the development of biotransformations using isolated
enzymes. This trend, however, is also a result of the perspective of the chemical
industry and in general organic chemists in developing and designing catalytic
synthetic processes. A key issue in the development of chemocatalytic processes is
the use of a single molecule or salt pair as a catalytic component. Such catalytically
active molecules or salt pairs are typically used in isolated and purified form.
Accordingly, the development of enzymatic processes based on such a philosophy
is based on the use and understanding of single isolated enzymes as molecular
catalysts rather than microbial systems (being aware that, of course, also in
microbial systems the same enzymes are responsible for the catalytic reaction).
However, compared to microbial systems, access to isolated enzymes is connected
with additional purification efforts that represent a costs factor. Whereas fermen-
tation of biomass as a catalyst is a cheap process, purification of enzymes (dependent
on the purification degree) makes the enzyme component increasingly expensive.
Owing to such higher costs for manufacturing isolated (purified) proteins over
microbial whole-cells obtained as biomass directly from the fermentation, during
process development there has been great interest in immobilization and recycling of
isolated (purified) enzymes and reaction engineering issues to attain an economically
attractive biocatalyst and economically favorable data for the biotransformation (in
particular) on a large scale [34].
The following describes selected concepts for recycling of isolated enzymes, which
enabled the realization of industrial manufacturing processes known today as
landmark processes in industrial biocatalysis. More recently, due to the tremen-
dous progress in molecular biology and the availability of efficient recombinant
1.4 Chemical Processes with Isolated Enzymes: The Impact of Process Engineering j23
whole-cells as production strains as well as high-cell density fermentation methods,
the production of isolated enzymes has also become economically very attractive, at
least when using them as a crude extract or in only partially purified form (which
often is suitable for biotransformations in contrast to diagnostic purposes). Thus,
dependent on the type of recombinant enzyme and degree of purification, even
biotransformations without recycling might represent an attractive option for todays
large-scale biotransformations. Selected examples of such types of biotransforma-
tions with free, non-immobilized enzymes are described below.
1.4.2
Development of Practical Synthesis of Chemicals via Transformations Using Isolated
Enzymes in Immobilized (Solid-Supported) Form
The option of recycling a catalyst has, in general, often been realized by means of
immobilization of the catalyst on a solid support, which enables simple separation of
the heterogeneous catalyst from the reaction mixture and its synthetic re-use [35, 36].
However, it also should be mentioned that many other techniques for catalyst
immobilization have been developed.
As for enzymes, one example of a very successful application of this concept of
heterogeneous enzyme catalysis is the established biocatalytic synthesis of 6-amino
penicillanic acid (6-APA) [37–39], which is applied with an annual production volume
exceeding 10 000 tons per year. The catalytic concept is shown in Scheme 1.13.
Compared with the alternative chemical route, the use of an immobilized Pen G
acylase enables cleavage of the unwanted side-chain without the need for significant
amounts of a range of hazardous chemicals. As a solid support, Eupergit beads
turned out to be highly efficient for the Pen G acylase catalyst. Notably, the
immobilized enzyme catalyst can be re-used more than 850 times, thus delivering
a highly efficient production process and very low overall enzyme loading of the
heterogenized enzyme catalyst per kg of 6-aminopenicillanic acid [40].
selective side-chain
cleavage immobilized
pencillin acylase
H (via covalent binding
Ph N on solid support) H2N
S OH S
CH3 Ph CH3
O +
N CH3 H2O O N CH3
O O
CO2H CO2H
6-aminopenicillanic acid
(6-APA)
Scheme 1.13 Immobilized penicillin G acylase (Pen G acylase) in the production of 6-APA.
Furthermore, heterogeneous enzyme catalysis has also been carried out very
successfully in organic reaction media. For example, when using immobilized lipase,
direct ester formation starting from an acid and an alcohol enables efficient
formation of the ester in a solvent-free medium. Such a process technology has
been industrially established for fatty acid ester manufacture, for example, at
Unichema Chemie and Degussa AG (now: Evonik Degussa GmbH) [15–17]. In the
24j 1 Introduction – Principles and Historical Landmarks of Enzyme Catalysis in Organic Synthesis
field of racemic resolution, the BASF process for the production of chiral amines is
based on an enantioselective acylation when starting from a racemic amine in the
presence of an immobilized lipase [41–43]. The production volume of this process
technology is in the >1000 tons per year range.
1.4.3
Development of Practical Synthesis of Chemicals via Transformations Using Isolated
Enzymes in Free Form
The use of solid-supported catalysts, however, also means that there is a switch from a
catalytic reaction in a homogeneous reaction medium (as in case of free enzymes)
towards a heterogeneous reaction medium. Thus, it has been a challenge to develop
immobilization-like systems that nonetheless enable both (i) the simple separation
and re-use of the enzyme component and at the same time (ii) the enzymatic reaction
to be carried out under homogeneous reaction conditions. This has been achieved,
for example, by means of a so-called enzyme-membrane reactor (Figure 1.3) [44–47].
In an EMR, the enzyme reacts as a free enzyme but is prevented from leaving the
reactor by a membrane. This membrane has a specific molecular-weight cut-off, that
L-aminoacylase
O O
H3C NH NH2 H3C NH

rac
water
H3C H3C + H3C
S CO2H S CO2H S CO2H
+ H2O
N-acetyl rac-methionine - CH3CO2H L-methionine N-acetyl D-methionine
Process concept of the enzyme membrane reactor for L-amino acid synthesis
L-product
D-substrate
sterile filter
enzyme
polarimeter
hollow
heat exchanger fiber
module
PC
sterile filter
DL-substrate
electronic feedback
Figure 1.3 Production of L-methionine in an enzyme-membrane reaction (EMR) [47].

1.4 Chemical Processes with Isolated Enzymes: The Impact of Process Engineering j25
is, only molecules that have a molecular weight below the cut-off can cross the
membrane. These membrane-permeable molecules are typically substrate and
product. The structure and concept of such an enzyme-membrane reactor is shown
in Figure 1.3 [47], exemplified for the synthesis of L-methionine by means of an
aminoacylase. Notably, the EMR runs also very successfully on an industrial scale,
producing L-methionine on a multi-hundred tons scale annually [45–47]. This
industrial process has been developed and established jointly by the Wandrey group
and researchers at Degussa AG (now: Evonik Degussa GmbH).
A further elegant example of the combination of organic chemistry with a
biocatalyst used in the form of an isolated enzyme is the synthesis of the artificial
sweetener aspartame in homogeneous reaction medium (Scheme 1.14) [48, 49]. This
process also represents an efficient and well-established industrial process running
at about the 2500 tons scale annually. The condensation reaction of the protected
aspartate and phenylalanine methyl ester is catalyzed by thermolysin from a Bacillus
strain in aqueous homogenous medium. In the downstream processing a filtration
step is included that ensures enzyme separation and recovery for re-use. The
reaction, which proceeds with excellent selectivity (>99.9%), originally had much
advantage as an efficient resolution of racemic phenylalanine, until the supply of
L-phenylalanine was started by the fermentative method.
NHZ NHZ
H
NH2 HO CO2H NH2 N CO2H
rac thermolysin
+ O + O
O O O O O O
(Z: benzloxycarbonyl
CH3 CH3 CH3
protecting group)
unwanted D-enantiomer Z-protected aspartame
(recycling after separation)
Scheme 1.14 Enzymatic synthesis of L-aspartame with thermolysin.
A further process option, which is of particular of interest when producing

hydrophobic molecules, is to run the reaction in a two-phase reaction medium
consisting of an aqueous phase and a water-immiscible organic solvent. This concept
is, of course, not only suitable for enzymes but also for whole-cell catalysts. Whereas
the enzymatic reaction proceeds in the aqueous phase, product accumulation takes
place in the organic phase. After completion of the reaction, simple phase separation
enables recovery of the enzyme (dissolved in the aqueous phase and ready for use in
the next reaction run) and isolation of the product (dissolved in the organic phase). A
prerequisite for such a process with enzyme recycling is a high stability of the enzyme
towards the reaction medium and substrates/products, leading to sufficient remain-
ing enzyme activity after the biotransformation and phase-separation steps. Today
many enzymes fulfill this prerequisite. Notably, among them are not only hydrolases
but also enzymes from other enzyme classes.
26
O (S)-oxynitrilase OH
O O
H + HCN CN
aqueous buffer /
methyl tert-butyl ether
(S)-m-phenoxybenzaldehyde
cyanohydrin
Scheme 1.15 Hydrocyanation process with free, non-immobilized enzymes.
A selected example of such a process concept with non-immobilized enzymes is

the hydrocyanation of m-phenoxybenzaldehyde in the presence of a recombinant
oxynitrilase from Hevea brasiliensis (Scheme 1.15) [50–52]. This biotransformation,
leading to an asymmetric CC bond formation, which has been developed by the
Griengl group, has found an industrial application at DSM, running at a hundred
tons scale per year. Notably, an impressive space–time yield of 1 kg l1 day1 has been
achieved. The enantioselectivity is also excellent, leading to the desired (S)-cyano-
hydrin with 98.5% e.e.
Thus, today, various options are readily available for the development of efficient
biotransformations with isolated enzymes. These methodologies consist of the use of
enzymes in immobilized, solid-supported form or as free enzymes. In the latter
case, enzyme separation from the reaction mixture and re-use for the next synthetic
cycle can be realized by means of, for example, enzyme-membrane reactor technol-
ogy or the use of two-phase systems and a phase-separation after the reaction is
completed.
1.5
Towards Tailor-Made Enzymes: Principles in Enzyme Screening and Protein
Engineering Methodologies
In the following, recent remarkable examples in the screening for new enzymes and
enzyme evolution are introduced and their merits discussed. In particular, the history
of enzyme discovery will be covered, consisting of screening strategies for enzymes,
for example, via enrichment culture technique, stock cultures, in silico screening, and
modeling of proteins. In Figure 1.4 a flowchart of the use of tools for the discovery and
development of industrial enzymes is shown.
1.5.1
Tools for Enzyme Discovery
The success in microbial transformation has been based on the screening for
microbial enzymes catalyzing new reactions or by screening known enzymes for
an unknown activity with synthetic substrates. Taking into account the advantages of
using enzymes in organic synthesis (which have been in part been described in the
sections above and which are illustrated in detail in the individual chapters of this
1.5 Towards Tailor-Made Enzymes: Principles in Enzyme Screening and Protein j27
Figure 1.4 Flowchart of the use of tools for discovery and the development of industrial enzymes.
book), more and more attention has been paid to the systematic exploitation of new
enzyme reactions and how to obtain enzymes with desired activities from various
enzyme sources and databases. In screening enzymes, it is extremely important to
make clear what the purpose of the experiment is. One would screen microorganisms
or databases of enzymes for an enzyme when (i) nothing is known about the specific
reaction but a homologous enzymatic reaction of the same category is known
(substrate specificity), (ii) novelty is necessary and the same enzyme is known only
in other biological sources, (iii) improved function, such as a better productivity, heat-
, solvent-, pH stability, and so on for practical use is required, and even when (iv)
almost no information is available for the desirable reaction.
Different strategies for enzyme screening are conceivable. One can simply buy an
enzyme from the suppliers or clone the known gene by polymerase chain reaction
(PCR) according to the information given in the genome database on the internet,
and express it in a versatile heterologous host, such as E. coli. On the other hand, if one
would like to establish an entirely new enzymatic industrial process, there should not
be a similar case in the literature and therefore the enzyme reaction should be
unique. This strategy is extremely interesting because such basic studies often
accompany the discovery of unforeseen biological phenomena and new materials
hidden in nature. Thus, a successful screening should focus on what is new in the
screening: substrate, product, gene, protein, property, screening source, method, and
so on [53].
Enzymes need to be made more robust under harsh conditions in order to further
expand their use in the varieties of practical chemical reactions. Because several
substrates need to be checked for the transformation, the enzymology established
with a certain enzyme for a physiological substrate sometimes does not supply
enough information. The enantioselectivity shown by the enzyme is one of the most
28
important properties, although such information has been sometimes limited

because of, e.g., the lack of the supply of synthetic substrates and analyses of the
products by fine analytical methods.
Enrichment culture is a technique to isolate microorganisms having special
growth characteristics. It is estimated that there are 108 living microorganisms in a
1-g soil sample. Many industrial strains or new reactions have been discovered with
this technique [54]. Some microorganisms grow faster than others in media with
limited nutrients, high temperature, extreme pH values, and so on. Microorgan-
isms that grow faster than other species become dominant after several transfers of
the culture. An adaptation to a synthetic medium containing a target compound
often results in the isolation of microorganisms having a new enzyme to degrade
the sources for their growth. For example, industrial acrylamide producers Pseu-
domonas chlororaphis B23 [55] and Rhodococcus rhodochrous (formerly Arthrobacer
sp.) J-1 [56] were isolated by using nitrile compounds as carbon and nitrogen
sources, or just as a nitrogen source. Nitrile hydratase was discovered from
Rhodococcus rhodochrous J-1 [26].
1.5.2
Protein Engineering Methodologies
Protein engineering is the term given to the alteration of the primary structure of
proteins or enzymes by substituting amino acid residues by changing the codon of its
corresponding DNA. It has become possible as recombinant DNA techniques and
structures obtained by X-ray crystallography of the enzymes have become widely
available. The amount of information on the DNA and primary structures of the
proteins and the X-ray structures is so huge that it is now available through databases
on the internet. The key residues causing hydrogen bonding, hydrophobic, and
electrostatic effects around the active sites are often subjected to alteration to different
amino acids. Furthermore, by saturation mutagenesis, the single residues can be
changed to one of 19 other residues. These point mutations are very effective in the
analysis of specific residues in proteins by changing their properties, although the
effect of the specific changes to the structure and the properties of the enzymes often
lack additivity and is still complex and sometimes unpredictable. Therefore, satis-
factory results for applications have often been obtained by a combination of
mutagenesis and screening of the mutants, similar to the screening that had been
made for the wild-type microorganisms or enzymes from the nature. The screening
has been extended to high-throughput screening by the development of robotics such
as colony pickers, various kinds of distributors, mini-scale cultivation, followed by
monitoring by micro plate readers, along with miniaturization of the biochemical
reactions and the use of appropriate indicators [57].
Directed evolution has become employed for the mutagenesis of enzymes since
the commercial availability of the thermal cycler and DNA polymerase from ther-
mophilic microorganisms for PCR in the late 1980s to early 1990s [58, 59]. The gene
for the enzyme is randomly mutated by error-prone PCR, in which the DNA
polymerase reaction is staggered by the inadequate reaction conditions such as very
1.5 Towards Tailor-Made Enzymes: Principles in Enzyme Screening and Protein j29
low concentration of some of the substrate dNTP, or using Mn2 þ instead of Mg2 þ in
the reaction mixture. The mutated genes are expressed in appropriate hosts such as E.
coli, or in vitro by DNA translation system from various biological sources. A number
of the E. coli transformants are picked up manually or by a machine called a colony
picker and assayed by various high-throughput robotic systems, with microtiter
plates of 96-wells or more. By choosing the best candidate as a parent, the next cycle of
mutagenesis is begun.
In DNA shuffling, which causes much mutation in PCR, the DNA-catalyzed
reaction is not started with primers but only with fragmented DNA molecules partly
digested with DNase I. Annealing reactions proceed with partially overlapped DNA
fragments and, finally, DNA of the same original size is constructed, creating
mixtures of DNA with different mutation sites. Stemmer et al. described an example
of generating TEM-1 b-lactamase that showed 16 000 times resistance (MIC (min-
imum inhibitory concentration) 320 mg ml1) against the wild-type enzyme (MIC
0.02 mg ml1) [60]. DNA shuffling of a family of genes from diverse species accel-
erates directed evolution. Family shuffling is a more powerful method for making
libraries of chimeric genes by random fragmentation of a pool of related genes, by
combining useful mutations from individual genes. Moxalactamase activity from
four cephalosporinase genes evolved separately from a mixed pool of the four
different genes. A single cycle of shuffling yielded 270- to 540-fold improvement
from the four genes shuffled together. The best clone contained eight segments from
three of the four genes as well as 33 amino-acid point mutations [61].
In the following, as a case study, the successful combination of a classic screening
with a subsequent directed evolution as a highly efficient tool for enzyme optimi-
zation is shown, and exemplified for a new industrial enzymatic method of selective
phosphorylation of nucleosides for the production of, for example, inosine-50 -
monophosphate (50 -IMP). This is one of the first examples of the application of
directed evolution for an industrial reaction combined with conventional screening
for a very selective phosphorylation reaction, catalyzed by an acid phosphatase.
Scheme 1.16 illustrates the corresponding target reaction.
O
Base Base
Base
Base PPi Pi HO P O CH2
HOCH2
O O
OH
acid phosphatase OH OH
OH OH
5´-IMP
0
Scheme 1.16 Acid phosphatase-catalyzed synthesis of 5 -IMP.
Inosine-50 -monophosphate (50 -IMP) and guanosine-50 -monophosphate (50 -GMP)

are important nucleotides because they give foods the umami taste. There is no
taste in other isomers such as 20 - and 30 -inosinic acids. Two phosphorylation methods
have been reported. One is a chemical phosphorylation process that uses phosphoryl
chloride (POCl3) and the other is an enzymatic process that uses inosine kinase.
To attain a greener and newer enzymatic method to produce 50 -IMP, microorganisms
30
120
5'-Inosinic acid (g/l)

100
80
60
40
20
0
0 4 8 12 16 20
Reaction time (h)
Figure 1.5 Time course of the synthesis of 50 -inosinic acid monophosphate with ðoÞ wild-type
enzyme and (.) mutant.
(3000 strains) that phosphorylate nucleosides using pyrophosphate (PPi) as the

phosphate donor regioselectively at the 50 -position were screened (Figure 1.5) [62].
Although many of the microorganisms screened were able to phosphorylate inosine,
phosphotransferase activity specific to the 50 -position was found to be distributed
among bacteria belonging to the family Enterobacteriaceae. Morganella morganii
NCIMB 10 466 was selected as a 50 -IMP producer.
A selective nucleoside phosphorylating enzyme was purified to homogeneity from
M. morganii NCIMB 10 466 crude extract. The enzyme appeared to consist of six
subunits identical in molecular weight (Mr, 25 000). It phosphorylated various
nucleosides at the 5-position to produce nucleoside-50 -monophosphates using PPi
as the phosphate source. Energy-rich compounds, such as carbamylphosphate and
acetylphosphate, were also very effective phosphate donors. The enzyme also
exhibited phosphatase activity, and dephosphorylated various phosphate esters, but
had a weak effect on nucleoside-30 -monophosphates. The M. morganii gene encoding
a nucleoside phosphorylating enzyme was isolated by a shotgun-cloning strategy. It
was identical to the M. morganii PhoC acid phosphatase gene. Using the purified
enzyme, 32.6 mM 50 -IMP was synthesized from inosine with a 41% molar yield, but
the synthesized 50 -IMP was hydrolyzed back to inosine due to its phosphatase activity
as the reaction time was extended.
To suppress the dephosphorylation reaction and increase the efficiency of the
transphosphorylation reaction, a random mutagenesis approach was used. By error-
prone PCR, one mutated acid phosphatase that increased the phosphotransferase
reaction yield was obtained. With the E. coli overproducing the mutated acid phospha-
tase, 101 g l1 (191 mM) of 50 -IMPwas synthesizedfrominosine in85% molar yield [63].
This improvement was achieved with two mutations, Gly to Asp at position 92 and Ile to
Thr at position 171. A decreased Km for inosine was responsible for the increased
productivity. Ile151Thr caused the improvement of the affinity toward inosine, and the
mutation at Gly72Asp caused not only increasing affinity toward inosine together with
Ile151Thr, but also lowered the dephosphorylating activity. An X-ray analysis of the
crystal of the enzyme revealed two mutation sites located near the active site, and the
Thr151 seems to form hydrogen bonds with inosine and enhance the affinity towards
1.6 Hybridization of Enzyme Catalysis with Organic Syntheses j31
inosine [64]. Now, 50 -IMP and 50 -GMP have been produced in multi-thousand tons per
year since 2003 by Ajinomoto Co. Inc., Japan.
A further efficient tool for protein design and engineering is based on compu-
tational methods. Here a range of successful modeling work has been done, leading
to impressive enzyme properties. Details about rational protein engineering are
given in Chapters 3 and 4. In the following the power of rational protein design and
engineering is exemplified by the exciting development of a new, de novo designed
enzyme suitable for the Kemp elimination, a reaction not found in biological systems.
This recently reported computationally designed enzyme uses two different catalytic
motifs to catalyze the Kemp elimination (Scheme 1.17) [65].
B:
δ+
B
H H
O2 N O2 N O2N CN
N N
O O δ- OH
HX
Scheme 1.17 Kemp elimination reaction.
These results demonstrate the power of combining computational protein design

with directed evolution for creating new enzymes, and we anticipate the creation of a
wide range of useful new catalysts in the future. The first step for designing new
enzymes is to assume a catalytic mechanism and then to use quantum mechanical
transition state calculations to create an idealized active site, with protein functional
groups positioned so as to maximize transition state stabilization. The next step is the
use of software to search for protein backbone positions capable of supporting these
idealized active sites from already reported high-resolution natural crystal structures.
Subsequently, residues surrounding the transition state are redesigned to maximize
the stability of the active site conformation. It was found that TIM (triose phosphate
isomerase) barrel scaffolds were enriched in the computer search. Eight of the
designs showed measurable activity in the Kemp elimination. Scaffolds of indole-3-
glycerolphosphate synthase from Sulfolobus solfataricus and deoxyribose-phosphate
aldolase from Escherichia coli were virtually chosen. Furthermore, enzymes with a
better activity were generated by an in vitro evolution procedure. It represented a rate
enhancement of 1 106 over the spontaneous reaction in a solution. The structures
of the enzymes were solved.
1.6
Hybridization of Enzyme Catalysis with Organic Syntheses: New Opportunities
for Industrial Production of Chemicals and Drugs
This section presents a selection of more recent trends in enzyme catalysis that
already have had or which (presumably) will have an impact on organic synthesis and
32
technical applications, while being aware that there will be numerous other inter-
esting application areas of enzymes as catalysts in organic synthesis as well (which
are also covered in different chapters of this book).
1.6.1
Applications of Tailor-Made Recombinant Whole-Cell Catalysts in Organic Synthesis
The tremendous progress made in molecular biology (as discussed in the previous
section) also had a significant impact on process development with biocatalysts.
Some trends resulting from these impressive achievements are given here. The
opportunity to optimize enzyme properties by protein engineering combined with
overexpression in recombinant host organisms and high-cell density fermentation
for their production enables access to both economically and synthetically highly
attractive catalysts. Thus, there has been an increasing tendency in organic
chemistry to directly use such types of tailor-made recombinant whole-cell
catalysts, so-called designer cells or designer bugs in organic (asymmetric)
synthesis. Such recombinant whole-cell catalysts have gained a high level of
popularity in particular in the field of those processes where more than one
reaction is carried out, since two or more enzymes are overexpressed in the
recombinant host strain. Thus, only one fermentation process is required to
produce the required enzymes. Furthermore, the biomass represents the cheapest
form of an enzyme, and costly enzyme purification steps can be avoided.
Figure 1.6 give a schematic comparison of process unit operations of a whole-
cell biocatalyzed (redox) process and an analogous process in the presence of
isolated enzymes [66], exemplified for the biocatalytic redox process in which two
enzymes as well as a cofactor are required.
Biocatalytic production processes that require more than one enzyme are, for
example, the multistep transformation of a hydantoin into an L- or D-a-amino acid and
Process option Isolated Whole Cell

Enzymes Catalysts
Unit operations
fermentation
cell disruption
purification
concentration
bioconversion
NAD(P)+ less/no NAD(P)+
increase increase
of of
enzyme biocatalyst
purity costs
Figure 1.6 Overview of process unit operations when using recombinant whole-cells and when
using isolated enzymes in a redox process [66].
the reduction of ketones towards chiral secondary alcohols. Both processes have
already been conducted in a very efficient way by means of recombinant whole-cell
biocatalysts and in the following these two types of processes are described in more
detail.
Starting with the biocatalytic dynamic kinetic resolution of hydantoins with a
recombinant whole-cell biocatalyst, this process is based on three enzymatic steps,
requiring a racemase, hydantoinase, and carbamoylase [67, 68]. Overexpression of all
three enzymes in an E. coli host organism delivered a highly efficient catalyst for this
process, enabling excellent productivity data. Such a process technology, enabling
access to L- as well as D-a-amino acids, has been established on an industrial scale at
Degussa AG (now: Evonik Degussa GmbH). Scheme 1.18 shows such a process
schematically for the synthesis of L-a-amino acids. Since – in contrast to wild-type
microorganisms – different strains can represent the original sources of the three
enzymes, the most suitable enzyme from all strains for each step can be utilized as
preferred enzyme component in overexpressed form in the recombinant host
organism.
O + H2O
R R - CO2
+ H2O CO2H - NH3 R
HN NH HN CO2H
NH2 L-carbamoylase H2N
L-hydantoinase
O O L-amino acid
L-hydantoin
racemase
recombinant
whole-cell
R O catalyst
HN NH
O
D-hydantoin
Scheme 1.18 Dynamic kinetic resolution of hydantoins using a recombinant whole-cell catalyst.
The use of recombinant whole-cell catalysts has also attracted high interest for the
asymmetric reduction of ketones, leading to secondary alcohols in enantiomerically
pure form (>99% e.e.) [69]. The desired reduction of the ketone is catalyzed by an
alcohol dehydrogenase, requiring NAD(P)H as a cofactor and reducing agent. The
oxidized cofactor NAD(P) þ is subsequently reduced and recycled to NAD(P)H,
which then can be used for the next catalytic cycle. One option for the in situ cofactor
recycling is based on the use of a glucose dehydrogenase, which oxidizes D-glucose to
D-gluconolactone, thus recycling the required cofactor. It turned out that a recom-
binant E. coli strain, containing an alcohol dehydrogenase and a glucose dehydro-
genase in overexpressed form, represents a highly efficient, tailor-made biocatalyst.
Such whole-cell catalyzed reductions of ketones have been established at, for
34
example, Kaneka and Degussa AG (now: Evonik Degussa GmbH), and different
process options can be used for this type of process (e.g., in the presence or absence of
an organic solvent) [69–72]. Notably, such processes run at a high substrate input,
exceeding 100 g l1, and lead to excellent productivity data. The process concept is
shown in Scheme 1.19 (and for process unit operations of the overall concept,
including access to the biocatalyst, see Figure 1.6 and the discussion above).
OH
NAD(P)+
D-glucono-
R1 R2
lactone
glucose recombinant alcohol

dehydrogenase whole-cell dehydrogenase
(GDH) catalyst (ADH)
D-glucose NAD(P)H R1 R2
Scheme 1.19 Asymmetric ketone reduction using a recombinant whole-cell catalyst.
A range of other types of efficient whole-cell catalyzed processes have been

reported as well, for example, the reductive amination of a-keto acids [66, 73, 74]
and reduction of C¼C double bonds [75, 76] as further technologies applied on a
technical scale.
Thus, in such modern biotransformations with recombinant whole-cells one can
see a renaissance of whole-cell catalysis with high industrial impact, since major
limitations when using wild-type strains such as low substrate input, formation of
side-products, and limited volumetric productivity can be overcome by means of
tailor-made recombinant whole-cell catalysts.
1.6.2
Novel Retrosynthetic Approaches in Drug Synthesis: From Enzyme Catalysis in
Chemoenzymatic Multistep Processes towards New Drug Production Pathways
in Industry
A major application area of enzyme catalysis is the synthesis of pharmaceuticals and

intermediates thereof, since such molecules often are chiral and, at the same time,
enzymes often show excellent stereoselectivities [77, 78]. Despite this ideal fit of
synthetic requirement and enzyme stereoselectivity properties, the number of
enzyme-catalyzed processes in organic syntheses has been rather limited for a long
time compared with classic chemical or chemocatalytic syntheses. One reason that
might represent in part an explanation for this is, on first glance, the surprising
observation that retrosynthetic strategies in natural product or drug synthesis have
often been designed mainly based on classic organic synthetic routes without
consideration of biocatalytic key steps. This might be because biocatalysis still needs
to be more implemented as a standard synthetic tool in teaching courses on organic
chemistry. Another reason for the limited use of enzymatic steps in natural product
and drugs synthesis is that for a long time there was a lack of process efficiency of
enzymes despite the excellent stereoselectivities (such as, for example, low volu-
metric productivities). Although highly efficient processes have long been reported
mainly when using hydrolases as catalysts (and in part oxidoreductases), other
enzyme classes have been used to a lesser extent with respect to the development
of highly efficient production processes.
However, in the last two decades tremendous progress has been made to overcome
these two above-mentioned limitations. In particular for pharmaceuticals exciting
retrosynthetic approaches have been designed that are now based on enzymatic key
steps (a selected example is given below; Scheme 1.20). In addition, process
efficiencies have also been increased in a broad range of biocatalytic processes with
enzymes other than hydrolases. Today, besides hydrolases-catalyzed syntheses a
range of other industrially applied reactions based on the use of oxidoreductases,
transferases, lyases, and isomerases are known. A very recent selected impressive
example (shown in Scheme 1.21) is also discussed below.
(a) First generation synthetic route to pregabalin based on "classic" resolution:

non-CH-acidic substrate
for resolution
(S)-mandelic
CN CN Ni, H2
rac KOH rac rac NH2 acid NH2
EtO2C CO2Et CO2K CO2H CO2H

pregabalin
(b) Second generation synthetic route to pregabalin based on enzymatic resolution:

CH-acidic substrate
for resolution
CN CN
rac nitrilase NH2
CN CO2H CO2H
+ pregabalin
CN
basic racemization
CN
Scheme 1.20 (a) First- and (b) second-generation multistep route to pregabalin.
Starting with the implementation of biocatalytic reactions as key steps in retro-

synthetic approaches for drugs synthesis, an impressive contribution has been, as an
example, by Tao and coworkers in developing alternative pathways towards prega-
balin as the active pharmaceutical ingredient (API) of Pfizers neuropathic pain drug
36
Route A): Route B):

Asymmetric metal-catalyzed F Asymmetric enzyme-catalyzed
hydrogenation as key step F transamination as key step
O O
N N
N
N F
NH4OAc
F3C transaminase, PLP
F + i-PrNH2
- acetone
F
O NH2
N N F
N F
N F
O NH2
F 3C N
1. Rh/ligand/H2 (250 psi) N
2. carbon treatment to remove Rh N
N F
F 99.95% ee
F3C
F
O NH2
N N
N
N F
97% ee
F 3C
H2PO4- F
F
1. heptane/i-PrOH O NH3+
H3PO4
2. H3PO4 N N
N
N F
F 3C
sitagliptin phosphate
Scheme 1.21 Comparison of original metal-catalyzed and new biocatalytic route to sitagliptin.
LyricaÒ (Scheme 1.20) [79, 80]. The original synthetic route, shown in Scheme 1.20,
starts from a b-cyano-malonate, which is then transformed by classic chemical
synthetic steps into the desired racemic c-amino acid. The final step is a classic
resolution based on diastereoselective salt formation with a chiral acid. By means of
this method pregabalin was obtained in an overall yield of 20%. Its main disadvantage
is the lack of a suitable racemization for the unwanted enantiomer, since the b-amino
acid is difficult to racemize due to the lack of a C-H-acidic functionality at the
b-position. This has been addressed in a second-generation route as an improved
alternative, which is based on a regio- and enantioselective hydrolysis of isobutyl-
succinonitrile as the starting material in the presence of a nitrilase as a biocatalyst,
leading to the required intermediate in 45% yield and with >97% e.e. This strategy
turned out to be superior to the first route since the resolution process now leads to an
unwanted enantiomer that can be easily racemized under basic conditions due to the
presence of an a-C-H-acidic functionality at the stereogenic center. Thus, this
chemoenzymatic multistep synthesis of pregabalin is a very elegant approach,
combining an efficient enzymatic resolution with the idea of implementing the
resolution at an early stage, thus enabling racemization of the unwanted enantiomer.
Notably, in addition, this example underlines the huge diversity of enzyme catalysis
with respect to resolution processes. Whereas enzymatic resolution of nitriles is
catalyzed enantio- and regioselectively by nitrilases, chemical resolution methods for
nitriles are less explored (compared with amines and acids).
The next example addresses the tremendous progress that has been made to
make enzymes suitable for the highly competitive industrial synthesis of complex,
pharmaceutically relevant molecules. Without doubt, one of the highlights in
recent years has been the development of a chemoenzymatic production process
for the drug sitagliptin phosphate by Merck and Codexis researchers
(Scheme 1.21) [81, 82]. Notably, this enzymatic process, based on a transaminase
as a biocatalyst, has turned out to be advantageous over the previously developed
and also industrially established chemocatalytic alternative. In the chemical
synthesis, using an asymmetric metal-catalyzed hydrogenation as a key step, first
the ketone used as a starting material is transformed into an enamine, which is
subsequently hydrogenated enantioselectively. The final step consists of salt
formation of the drug sitagliptin phosphate. Despite an excellent hydrogenation
process, the whole synthetic route possesses two drawbacks: First, direct trans-
formation of the ketone into the amine instead of a two-step process with an
enamine intermediate formation would be more desirable. Even more important,
however, is having a heavy-metal catalyzed process nearly at the end of the
multistep synthesis; the need to remove metal traces from a drug intermediate
at a late stage is tedious and disadvantageous. These drawbacks have been solved
by applying a direct enzymatic transformation with a transaminase, allowing
direct conversion of the ketone substrate into the desired amine. Furthermore,
heavy metals (required as catalyst component in the chemocatalytic asymmetric
hydrogenation step) are no longer involved. A major challenge, however, had to be
solved to realize this process, namely, the development of a transaminase showing
sufficient activity for the sterically demanding ketone substrate. The enzyme
optimization carried out also underlined todays tremendous opportunities in
protein engineering. Starting from a wild-type enzyme showing negligible activity
for the ketone substrate, eleven mutation rounds led to a highly efficient mutant.
In the presence of this optimized enzyme, a process has been realized that runs at
impressive substrate input and leads to the desired product with the excellent
enantioselectivity of 99.95% e.e. A detailed comparison of the chemical and
biocatalytic routes, describing the significant advantages biocatalysis offers, has
been reported recently [82].
These two recent examples in the field of chemoenzymatic multistep drugs
synthesis underline the tremendous potential of enzyme-catalyzed processes for
the multistep synthesis of complex molecules such as drugs and natural products.
Thus, in future an increasing tendency to integrate biocatalytic key steps into
multistep routes to such molecules can be expected, thus contributing to the
development of both economically attractive and sustainable production
processes.
38
1.6.3
Recent Aspects of Applications of Enzymes in Organic Synthesis
Furthermore, there is a range of other emerging fields within enzyme catalysis, for
example, running biocatalytic processes in non-conventional media, the develop-
ment of one-pot multistep processes by combining enzymatic reactions or enzymatic
and chemocatalytic reactions, and enzymatic promiscuity in organic synthesis. The
latter research area is briefly described here as a representative example of the many
exciting emerging research areas in biocatalysis by two selected processes.
The term enzyme promiscuity is used to describe unusual reactions enzymes are
able to catalyze besides their catalytic properties known from natural processes. One
type of enzyme promiscuity refers to new reaction types, in particular non-natural
reaction types that an enzyme turned out to be able to catalyze. It has been exciting to
see that by means of enzymes organic reactions such as, for example, the nitroaldol
reaction or hydroformylation can be conducted. Different strategies to obtain access
to such enzymes have been studied. Often, the starting point was an enzyme that
catalyzes a reaction with a similar reaction mechanism in nature. The usefulness of
this strategy has been demonstrated successfully, for example, by the Griengl group
for the development of the first enzymatic nitroaldol reaction, also known as the
Henry reaction (Scheme 1.22) [83–85]. As enzymes, oxynitrilases were studied that
catalyze the addition of cyanide as a nucleophile to aldehydes. It, however, also turned
out that an oxynitrilase from Hevea brasiliensis can accept nitromethane and
nitroethane as a nucleophile. The mechanism of both reactions is comparable, since
both additions are based on activation of the donor by proton abstraction of the a-C-
H-acidic donor. Although catalytic activity is lower when using nitromethane
compared with hydrogen cyanide, this enzymatic nitroaldol reaction is an elegant
example of a successful expansion of an enzymés reaction range based on the rational
concept of applying knowledge of the mechanisms of organic reactions.
O (S)-oxynitrilase OH
NO2
H + Me NO2
aqueous buffer/
methyl tert-butyl ether
(S)-nitroaldol
adduct
Scheme 1.22 Enzyme promiscuity in an asymmetric nitroaldol reaction.
A different approach of enzymes is based on rational protein engineering (see also

section above). The starting point for such approaches can be, for example, a model
based on a reaction mechanism (followed by a subsequent design of the enzyme by
molecular modeling) or a three-dimensional structure of the enzyme and subsequent
optimization by means of a rational and experimental approach. The latter strategy
has recently led to an enzyme capable of catalyzing hydroformylation as an example
for an industrially important non-natural reaction, when starting from a carbonic
References j39
anhydrase as a starting point in this rational protein engineering study and replacing
the zinc center by rhodium as a metal component in the active site [86].
1.7
Summary and Outlook
In conclusion, due to impressive interactions between biology, chemistry, and

engineering in recent decades enzyme catalysis has become an attractive synthetic
tool in organic chemistry, thus complementing existing classic chemical and
chemocatalytic approaches. Today a broad range of organic reactions such as, for
example, redox reactions, hydrolytic reactions, and CC bond formations can be
carried out very efficiently by means of enzymes. Furthermore, enzyme catalysis has
developed towards a broadly applied production technology in the chemical industry,
in particular in the fields of fine chemicals and pharmaceuticals.
In future, it can be expected that we will see many more biocatalytic reaction types
running in a highly efficient manner, thus being suitable for industrial-scale applica-
tions, too. It also can be expected that besides optimization of known biocatalytic
reactions expansion towards new type of reactions types will be possible by means of
protein engineering techniques. Another challenge in the future will be the further
implementation of biocatalytic reactions into multistep synthesis of (chiral) building
blocks such as pharmaceuticals. This field consists of the development of alternative
retrosynthetic approaches to drugs based on biocatalytic key steps as well as the
development of multistep one-pot syntheses with biocatalytic reactions. Further
improvements in molecular modeling of enzyme-catalyzed syntheses can be expected
as well, thus also enabling an increased number of biocatalytic processes based on
enzymes optimized or designed by rational protein engineering.
References
1 Liese, A., Seelbach, K., and Wandrey, C. Verlag GmbH, Weinheim,

(eds) (2006) Industrial Biotransformations, pp. 1–36.
2nd edn, Wiley-VCH Verlag GmbH, 4 Voet, D. and Voet, J. (2004) Biochemistry,
Weinheim. John Wiley & Sons, Inc., Hoboken.
2 For a brief recent overview about enzyme- 5 Uppenberg, J., Oehrner, N., Norin, M.,
catalyzed organic syntheses, see: Gr€oger, Hult, K., Kleywegt, G.J., Patkar, S.,
H. (2010) in Catalytic Asymmetric Synthesis, Waagen, V., Anthonsen, T., and Jones, T.A.
3rd edn (ed. I. Ojima), John Wiley & Sons, (1995) Biochemistry, 34, 16838–16851.
Inc., Hoboken, ch. 6, pp. 269–341. 6 Fischer, E. (1894) Ber. Dtsch. Chem. Ges.,
3 For a comprehensive, excellent 27, 2985–2993.
overview of the history of industrial 7 Fischer, E. (1894) Ber. Dtsch. Chem. Ges.,
biotransformations and enzyme 27, 3189–3232.
catalysis, see: Vasic-Racki, D. 8 Michaelis, L. and Menten, M.L. (1913)
(2006) in Industrial Biotransformations, Biochem. Z., 49, 333–369.
2nd edn (eds A. Liese, K. Seelbach, 9 Segel, I.H. (1995) Enzyme Kinetics, John
and C. Wandrey), Wiley-VCH Wiley & Sons, Inc., New York.
40
10 International Union of Biochemistry and 25 Yamada, H. and Kobayashi, M. (1996)

Molecular Biology (1992) Enzyme Biosci. Biotechnol. Biochem., 60,
Nomenclature, Academic Press, San 1391–1400.
Diego. 26 Asano, Y. and Kaul, P. (2011) Molecular
11 (a) Review: Stuermer, R., Hauer, B., Hall, Chirality, Elsevier, in press.
M., and Faber, K. (2007) Curr. Opin. Chem. 27 Asano, Y., Yasuda, T., Tani, Y., and
Biol., 11, 203–213; (b) BASF SE (2008) Yamada, H. (1982) Agric. Biol. Chem., 46,
news release, 2008, P 402/08 e. 1183–1189.
12 Review: Kardinahl, S., Rabelt, D., and 28 Ikemi, M. (1994) Bioproc. Technol., 19,
Reschke, M. (2006) Chem. Ing. Tech., 78, 797–813.
209–217. 29 Liese, A., Seelbach, K., and Wandrey, C.
13 Aehle, W. (2007) Enzymes in Industry, 3rd (eds) (2006) Industrial Biotransformations,
edn, Wiley-VCH Verlag GmbH, 2nd edn, Wiley-VCH Verlag GmbH,
Weinheim. Weinheim, pp. 411–416.
14 Review: Koskinen, A.M.P. and 30 Kirchoff, G.S.C. (1811) Acad. Imp. Sci., St.
Klibanov, A.M. (eds) (1996) Enzymatic Petersburg, Memoires, 4, 27.
Reactions in Organic Media, Blackie 31 Payen, A. and Persoz, J.-F. (1833) Ann.
Academic & Professional, Glasgow. Chim. Phys., 53, 73–92.
15 Liese, A., Seelbach, K., and Wandrey, C. 32 Schwann, T. (1836) Arch. Anatom. Physiol.
(2006) Industrial Biotransformations, 2nd Wiss. Med., 90–138.
edn, Wiley-VCH Verlag GmbH, 33 K€uhne, W. (1876) Verhandl.
Weinheim, pp. 313–315. Heidelb. Naturhistor. Vereins, Carl Winter
16 Hills, G.A., MsCrae, A.R., and Verlag, Universit€atsbuchhandlung
Poulina, R.R. (1990) (Unichema Chemie Heidelberg, Heidelberg.
BV), Eur. Pat. EP0383405. 34 For a review on reaction engineering
17 Hilterhaus, L., Thum, O., and Liese, A. issues in biocatalysis, see: Rao, N.N., L€
utz,
(2008) Org. Process Res. Dev., 12, 618–625. S., Seelbach, K., and Liese, A. (2006) in
18 Marshall, R.O., Kooi, E.R., and Moffett, Industrial Biotransformations, 2nd edn (eds
G.M. (1957) Science, 125, 648–649. A. Liese, K. Seelbach, and C. Wandrey),
19 For a review of in situ cofactor Wiley-VCH Verlag GmbH, Weinheim, pp.
regeneration when using redox enzymes, 115–145.
see: Weckbecker, A., Gr€oger, H., and 35 Kragl, U. (1996) in Immobilized Enzymes
Hummel, W. (2010) Adv. Biochem. Eng./ and Membrane Reactors, Industrial
Biotechnol., 120, 195–242. Enzymology, Macmillan Press, London, pp.
20 For a comprehensive review, see: Behrens, 275–283.
G.A., Hummel, A., Padhi, S.K., Sch€atzle, 36 Tanaka, A., Tosa, T., and Kobayashi, T.
S., and Bornscheuer, U.T. (2011) Adv. (eds) (1993) Industrial Application of
Synth. Catal., 353, 2191–2215. Immobilized Biocatalysts, Marcel Dekker,
21 This comparison of chemo- and New York.
biocatalysis for the synthesis of 37 Matsumoto, K. (1993) in Industrial
b-hydroxy a-amino acids is also discussed Application of Immobilized Biocatalysts
in: Baer, K., D€uckers, N., Rosenbaum, C., (eds A. Tanaka, T. Tosa, and
Leggewie, C., Simon, S., Kraußer, M., T. Kobayashi), Marcel Dekker, New York,
Oßwald, S., Hummel, W., and Gr€oger, H. pp. 67–88.
(2011) Tetrahedron: Asymmetry, 22, 38 Tramper, J. (1996) Biotechnol. Bioeng., 52,
925–928. 290–295.
22 Review: Raspor, P. and Goranovic, D. 39 (a) Reviews: Boller, T., Meier, C., and
(2008) Critical Rev. Biotechnol., 28, 101–124. Menzler, S. (2002) Org. Process Res. Dev., 6,
23 Neuberg, C. and Hirsch, J. (1921) Biochem. 509–519; (b) Katchalski-Katzir, E. and
Z., 115, 282–310. Kraemer, D.M. (2000) J. Mol. Cat. B:
24 Sch€afer, B. (2007) Naturstoffe der Enzym., 10, 157–176.
Chemischen Industrie, Elsevier, 40 Buchholz, K., Kasche, V., and
Munich, 389. Bornscheuer, U.T. (2005) Biocatalysts and
References j41
Enzyme Technology, Wiley-VCH Verlag 55 Asano, Y., Yasuda, T., Tani, Y., and
GmbH, Weinheim. Yamada, H. (1982) Agric. Biol. Chem., 46,
41 Balkenhohl, F., Ditrich, K., Hauer, B., and 1183–1189.
Ladner, W. (1997) J. Prakt. Chem., 339, 56 Asano, Y., Tani, Y., and Yamada, H. (1980)
381–384. Agric. Biol. Chem., 44, 2251–2252.
42 Ditrich, K. (2008) Synthesis, 2283–2287. 57 B€ottcher, D. and Bornscheuer, U.T. (2006)
43 Review: Breuer, M., Ditrich, K., Nat. Protocols, 1, 2340–2343.
Habicher, T., Hauer, B., Keßeler, M., 58 Cadwell, R.C. and Joyce, G.F. (1992) PCR
St€
urmer, R., and Zelinski, T. (2004) Angew. Methods Appl., 2, 28–33.
Chem., 116, 806–843;(2004) Angew. Chem. 59 Arnold, F.H. (1998) Acc. Chem. Res., 31,
Int. Ed., 43, 788–824. 125–131.
44 Wandrey, C., Wichmann, R., 60 Stemmer, W.P.C. (1994) Proc. Natl. Acad.
B€uckmann, A.F., and Kula, M.-R. (1980) in Sci. USA, 91, 10747–10751.
Enzyme Engineering 5 (eds H.H. Weetall 61 Crameri, A., Raillard, S.-A., Bermudez, E.,
and G.P. Royer), Plenum Press, New York, and Stemmer, W.P.C. (1998) Nature, 391,
pp. 453–456. 288–291.
45 Wandrey, C. and Flaschel, E. (1979) Adv. 62 Asano, Y., Mihara, Y., and Yamada, H.
Biochem. Eng., 12, 147–218. (1999) J. Mol. Catal. B: Enzym., 6, 271–277.
46 Bommarius, A.S., Schwarm, M., and 63 Mihara, Y., Utagawa, T., Yamada, H., and
Drauz, K. (1996) Chim. Oggi, 14, 61–64. Asano, Y. (2000) Appl. Environ. Microbiol.,
47 Gr€oger, H. and Drauz, K. (2004) in Large- 66, 2811–2816.
Scale Asymmetric Catalysis (eds E. Schmidt 64 Suzuki, E., Ishikawa, K., Mihara, Y.,
and H.U. Blaser), Wiley-VCH Verlag Shimba, N., and Asano, Y. (2007) Bull.
GmbH, Weinheim, ch. 1.8, pp. 131–147. Chem. Soc. Jpn., 80, 276–286.
48 Oyama, K. (1992) in Chirality in Industry 65 R€othlisberger, D., Khersonsky, O.,
(eds A.N. Collins, G.N. Sheldrake, and J. Wollacott, A.M., Jiang, L., DeChancie, J.,
Crosby), John Wiley & Sons, Inc., New Betker, J., Gallaher, J.L., Althoff, E.A.,
York, pp. 237–247. Zanghellini, A., Dym, O., Albeck, S.,
49 Liese, A., Seelbach, K., and Wandrey, C. Houk, K.N., Tawfik, D.S., and Baker, D.
(eds) (2006) Industrial Biotransformations, (2008) Nature, 453, 190–195.
2nd edn, Wiley-VCH Verlag GmbH, 66 Gr€oger, H., May, O., Werner, H.,
Weinheim, pp. 373–376. Menzel, A., and Altenbuchner, J. (2006)
50 Griengl, H., Hickel, A., Johnson, D.V., Org. Process Res. Dev., 10, 666–669.
Kratky, C., Schmidt, M., and Schwab, H. 67 May, O., Verseck, S., Bommarius, A., and
(1997) Chem. Commun., 1933–1940. Drauz, K. (2002) Org. Process Res. Dev., 6,
51 Liese, A., Seelbach, K., and Wandrey, C. 452–457.
(eds) (2006) Industrial Biotransformations, 68 May, O., Nguyen, P.T., and Arnold, F.H.
Wiley-VCH Verlag GmbH, Weinheim, pp. (2000) Nat. Biotechnol., 18, 317–320.
455–456. 69 Review: Gr€oger, H., Borchert, S.,
52 (a) Review: Purkarthofer, T., Skranc, W., Kraußer, M., and Hummel, W. (2010) in
Schuster, C., and Griengl, H. (2007) Appl. Encyclopedia of Industrial Biotechnology.
Microbiol. Biotechnol., 76, 309–320; Bioprocess, Bioseparation, and Cell
(b) Poechlauer, P., Scranc, W., and Technology(ed. M. Flickinger),
Wubbolts, M. (2004) in Large-Scale John Wiley & Sons, Inc., Hoboken,
Asymmetric Catalysis (eds E. Schmidt and pp. 2094–2110.
H.U. Blaser), Wiley-VCH Verlag GmbH, 70 Kataoka, M., Kita, K., Wada, M.,
Weinheim, ch. 2.1, pp. 151–164. Yasohara, Y., Hasegawa, J., and Shimizu,
53 Asano, Y. (2010) in Manual of Industrial S. (2003) Appl. Microbiol. Biotechnol., 62,
Microbiology and Biotechnology, 3rd edn 437–445.
(eds R.H. Baltz, J.E. Davies, and A. 71 Kizaki, N., Yasohara, Y., Hasegawa, J.,
Demain), American Society for Wada, M., Kataoka, M., and Shimizu, S.
Microbiology, pp. 441–452. (2001) Appl. Microbiol. Biotechnol., 55,
54 Asano, Y. (2002) J. Biotechnol., 94, 65–72. 590–595.
42
72 Gr€
oger, H., Chamouleau, F., Orologas, N., 79 Ran, N., Zhao, L., Chen, Z., and Tao, J.
Rollmann, C., Drauz, K., Hummel, W., (2008) Green Chem., 10, 361–372.
Weckbecker, A., and May, O. (2006) Angew. 80 Tao, J., Zhao, L., and Ran, N. (2007) Org.
Chem., 118, 5806–5809; (2006) Angew. Process Res. Dev., 11, 259–267.
Chem. Int. Ed., 45, 5677–5681. 81 Savile, C., Janey, J.M., Mundorff, E.C.,
73 Galkin, A., Kulakova, L., Yohimura, T., Moore, J.C., Tam, S., Jarvis, W.R., Colbeck,
Soda, K., and Esaki, N. (1997) Appl. J.C., Krebber, A., Fleitz, F.J., Brands, J.,
Environ. Microbiol., 63, 4651–4656. Devine, P.N., Huisman, G.W., and
74 Menzel, A., Werner, H., Altenbuchner, J., Hughes, G.J. (2010) Science, 329, 305–309.
and Gr€oger, H. (2004) Eng. Life Sci., 4, 82 Desai, A.A. (2011) Angew. Chem., 123,
573–576. 2018–2020;(2011) Angew. Chem. Int. Ed.,
75 Kataoka, M., Kotaka, A., Hasegawa, A., 50, 1974–1976.
Wada, M., Yoshizumi, A., Nakamori, S., 83 Purkarthofer, T., Gruber, K.,
and Shimizu, S. (2002) Biosci. Biotechnol. Gruber-Khadjawi, M., Waich, K.,
Biochem., 66, 2651–2657. Skranc, W., Mink, D., and Griengl, H.
76 Kataoka, M., Kotaka, A., Thiwthong, R., (2006) Angew. Chem., 118, 3532–3535;
Wada, M., Nakamori, S., and (2006) Angew. Chem. Int. Ed., 45,
Shimizu, S. (2004) J. Biotechnol., 114, 1–9. 3454–3456.
77 Patel, R.N. (ed.) (2006) Biocatalysis in the 84 Gruber-Khadjawi, M., Purkarthofer, T.,
Pharmaceutical and Biotechnology Skranc, W., and Griengl, H. (2007) Adv.
Industries, CRC Press, New York. Synth. Cat., 349, 1445–1450.
78 Dunn, P.J., Wells, A.S., and Williams, M.T. 85 Fuhshuku, K. and Asano, Y. (2011) J.
(eds) (2010) Green Chemistry in the Biotechnol., 153, 153–159.
Pharmaceutical Industry, Wiley-VCH 86 Jing, Q. and Kazlauskas, R.J. (2010)
Verlag GmbH, Weinheim. ChemCatChem, 2, 953–957
j43
2
Concepts in Biocatalysis
Eduardo García-Urdiales, Ivan Lavandera, and Vicente Gotor
2.1
Introduction
Enzymes catalyze a wide range of synthetic transformations (Scheme 2.1) with high
activity [1]. In their natural form many enzymes are sensitive catalysts that exert their
activity mainly in aqueous solution, and their handling requires some knowledge of
their biochemistry. These aspects have traditionally fostered a reluctance on the part of
organic chemists towards the employment of biocatalysts in their syntheses. However,
nowadays many enzymes can be acquired and used as readily as any other chemical
reagent. The following facts have contributed considerably to this situation: (i) some
enzymes can operate in non-aqueous media (Chapter 7), accepting a broad range of
substrates; (ii) immobilization techniques (Chapter 6) increase their stability and
simplify their handling; (iii) cofactor recycling techniques (Chapter 26, Chapter 27,
and Chapter 28) allow their usage in catalytic amounts, thus considerably lowering the
cost of these processes; and (iv) molecular biology, alone or in combination with
computational biochemistry, allows the modification of proteins at will and the
preparation of mutant enzymes in both rational (Chapter 4) and directed-evolution
(Chapter 5) fashions, with native or unnatural activities (enzymatic promiscuity,
Chapter 41). Moreover, the knowledge accumulated so far concerning the principles
of enzymatic catalysis has started to go beyond the biocatalysis field, which is reflected
in the numerous small-molecule catalysts inspired by biocatalytic mechanisms [2].
Selectivity is probably the most appealing property of enzymes applied in organic
synthesis. When enzymes catalyze the transformation of a given substrate they can
distinguish different functional groups (chemoselectivity), locations of identical
functional groups (regioselectivity), and stereoisomers or orientations of pro-
chiral/meso compounds (stereoselectivity) [1d]. This selectivity is a result of the
different energy of the transition states formed by the enzyme with each of the
different substrates or the orientations of a given substrate during those steps leading
to and including the rate-limiting step of the mechanism. The larger this energy
difference, the better the selectivity of the process. If we assume that a given
enzymatic reaction can be effectively described by Michaelis–Menten kinetics [3],
and Oliver May.
j 2 Concepts in Biocatalysis
44
Scheme 2.1 Enzymatic transformations most commonly used in organic synthesis.
during a selective biotransformation, a substrate (A) non-covalently binds to the active

site of the enzyme to afford the different possible Michaelis complexes (ES1, ES2. . .
ESn), which subsequently undergo reaction to irreversibly yield the different products
(P1, P2. . . Pn) (Scheme 2.2). In general, the enzymatic selectivity towards the major
product can be expressed as a function of the pseudo-second-order kinetic constants of
the whole transformation (also termed specificity constants) according to Eq. (2.1):
(kc at) 1
ES1 P1
(K S )1
(K m )1
(K S)2 (kc at) 2
A + E (K m) 2
ES2 P2
...
...
(K S)n
(K m)n
(kc at) n
ESn Pn
Scheme 2.2 General kinetic scheme for a substrate (A) selectively transformed into different
products (P1n) following Michaelis–Menten kinetics. P1 is the major product of the reaction.
2.2 Types of Biocatalytic Processes j45

kcat
KM 1
EðP1 Þ ¼ ð2:1Þ
kcat
KM 2 þ ... þ kcat
KM n
This selectivity ratio (E) is an inherent property of the enzyme that, unlike the
enantiomeric excesses of product(s) (eep) and substrate(s) (ees), is independent of the
degree of conversion of the reaction (c) [4]. Therefore, it is usually the parameter of
choice with which to quantify and compare selective transformations.
Among all types of enzymatic selectivity, stereoselectivity is, by far, the property
that has received most attention. IUPAC defines stereoselectivity as
the preferential formation in a chemical reaction of one stereoisomer

over another. When the stereoisomers are enantiomers, the phenomenon
is called enantioselectivity; when they are diastereoisomers, it is called
diastereoselectivity [5].
According to this definition, stereoselectivity and enantioselectivity (or diastereos-

electivity) are equivalent because only the nature of the products formed is taken into
account. However, there is some controversy in the scientific community, and some
authors prefer to use a substrate-based criterion to define stereo- and enantioselec-
tivity (or diastereoselectivity). Thus, if a substrate is either prochiral or meso and one
enantiomer is preferentially formed over the other the process is stereoselective but
not enantioselective since chirality is not present in the reagent. On the other hand,
when a mixture of enantiomers is used as substrate and one is preferentially
transformed, the process can be both enantioselective and stereoselective. According
to this substrate criterion, stereoselective biotransformations can be grouped into two
main different classes: asymmetric synthesis and kinetic resolution (KR) of racemic
mixtures. Conceptually, they differ in the fact that while asymmetric synthesis
implies the formation of one or more chirality elements in the substrate, a KR is
based on the preferential transformation of one of the two enantiomers of the
racemic mixture, which, subsequently, simplifies their separation. This conceptual
difference confers two practical disadvantages:
1) As long as the KR proceeds, the relative concentration of the slow-reacting
enantiomer in the remaining substrate increases (Figure 2.1). If the enantios-
electivity of the KR is not large enough, this seriously affects the performance of
the KR by lowering the optical purity of the product (eep).
2) In a KR only half of the starting material is transformed. When only one
enantiomer of a substrate is required this is a clear disadvantage.
2.2
Types of Biocatalytic Processes
This chapter describes the different stereoselective biotransformations known so far.

Thus, it starts with classical KRs of racemates. Next, parallel kinetic resolutions
(PKRs), aimed at overcoming the eep limitation of KRs, are described. However, the
46
Figure 2.1 Graphical representation of the Chen equations for two KRs with E values of 1057
(black lines) and 48 (grey lines). The eep (start above 90%) remains almost constant while ees
(start at 0) increases with the degree of conversion (c).
goal of most enzymatic syntheses is the quantitative production of a single isomer

of the target molecule. This fact is reflected in the numerous processes designed
to circumvent the yield limitation of KRs. Thus, the special case of the stereo-
selective enzymatic desymmetrization (SED) is described of meso and prochiral
compounds followed by the different approaches intended for racemic substrates.
With the exception of the dynamic kinetic resolution (DKR) of labile products, they
involve more than one catalyst and they either perform the divergent transformation
of the enantiomers of a racemate or their interconversion into a single enantiomeri-
cally pure compound. Finally, the case of dynamic kinetic asymmetric transformations
(DYKAT) of diastereomeric mixtures and some examples concerning biocatalyzed
cascade or domino sequences and concurrent processes are also shown.
2.2.1
Dealing with Racemates: Kinetic Resolutions (KRs)
Kinetic resolutions (KRs) are based on the different rate of the transformation of both
enantiomers of a racemic substrate into the corresponding enantiomeric products
(Scheme 2.3). The E of KRs is commonly termed the enantiomeric ratio [6]. Although
it is a ratio of kinetic constants, it can be reformulated to a function of values easy to
Scheme 2.3 General scheme of a kinetic resolution (KR). A and B are the enantiomers of a racemic
substrate; P and Q are the enantiomers of the corresponding product.
determine such as time (t), degree of conversion (c) of the reaction, and the
enantiomeric excesses of the substrate (ees) and the product (eep) [7]. In principle,
any combination of two out of the four aforementioned variables can be used to
calculate E values. In practice, the fact that methods that make use of t have not been
experimentally verified makes the employment of c, ees, and eep advisable. The Chen
equations (Eqs. (2.2)–(2.5)), which are limited to irreversible Michaelis–Menten
kinetics, relate E to these variables [6]. However, c values can be misleading if
unnoticed competing reactions occur. Although in Eq. (2.5) c values are not present
upon introducing Eq. (2.4) into Eqs. (2.2) or (2.3), according to Rakels et al. this is not
formally correct. Consequently, Rakels et al. derived an alternative formulation of E
directly derived from ees and eep (Eq. (2.6)) [8]. By using the Rakels method, E values
are obtained by means of a nonlinear regression of a data set of ees and eep values
measured at different reaction times:
ln½ð1cÞð1ees Þ
E¼ ð2:2Þ
ln½ð1c Þð1 þ ees Þ

ln 1c 1 þ eep
E¼ ð2:3Þ
ln 1c 1eep
ees
c¼ ð2:4Þ
ees þ eep

1ees
ln eep
ees þ eep
E¼ ð2:5Þ
1 þ ees
ln eep
ees þ eep
Rakels equation relating eep as a function of ees and E:

ees
eep ¼ " #E1
1 ð2:6Þ
ð1 þ ees ÞE
1
1ees
There is some debate about the threshold of E that leads to synthetically useful KRs,
which usually varies depending on the author and the success of the KR published.
In any case, if we take into account the Chen equations (Eqs. (2.2)–(2.4)), to obtain
both substrate and product with 99% ee an E value slightly higher than 1000 is
48
required (c ¼ 50%; Figure 2.1). But if this is not the case, high optical purities can still
be attained by tuning the extent of conversion of the KR. This is especially true for the
case of the substrate of the reaction. Thus, an E value of 48 is enough to obtain the
slow-reacting enantiomer of a racemate with 99% ee by stopping the reaction at a c
value of 55%. However, for the product of the reaction, E values slightly higher than
200 are required to obtain the fast-reacting enantiomer with 99% ee at a degree of
conversion higher than 10%.
Hydrolase-catalyzed KRs are, by far, the most frequent examples of reported KRs
(Chapter 8) [9]. In this case, both racemic nucleophiles and electrophiles (mainly
carboxylic acids and their derivatives) can be resolved. In particular, the KR of
secondary alcohols catalyzed by lipases/esterases is well studied – an empirical rule
has been formulated by Kazlauskas and coworkers on the basis of the size of the
substituents at the stereocentre [10]. According to this rule, the (R)-enantiomer
reacts faster and the bigger the difference in size between substituents the higher
the enantiomeric ratios.1) Structural studies have revealed that this behavior is due to
the presence of two pockets of different size in the nucleophilic binding site of these
enzymes [11]. Moreover, proteases usually show mirror image binding sites as
compared to lipases, and thus provide access to the opposite enantiomeric series [12].
As this pocket is highly conserved among the members of these families [13] the rule
can be extrapolated to most of these enzymes and to nucleophiles such as disubstituted
primary alcohols and isosteric primary amines [14]. Unfortunately, the structural
diversity of the binding sites of hydrolases that host the electrophiles [15] is responsible
for the fact that rules explaining the chiral preference of hydrolases towards this kind of
substrate have only been formulated for a few cases and cannot be generalized [16].
2.2.2
Overcoming the ee Limitation of KRs: Parallel Kinetic Resolutions (PKRs)
If the E of a KR is not high enough, the enantiopurity of the product dramatically

decreases, especially at conversion values close to 50%. To avoid this limitation, the
slow-reacting enantiomer can be removed by a parallel reaction, ideally at an identical
rate, thus maintaining the 1 : 1 ratio of the substrate enantiomers and yielding two
different products with substantially improved ee up to the 50% theoretical yield
(Scheme 2.4). This approach is called parallel kinetic resolution (PKR), for which it is
necessary that both reactions (i) occur without mutual interference, (ii) have similar
rates, (iii) have complementary enantiocontrol, and (iv) afford different and easily
separable products [17]. According to this last criterion, the processes can be divided
into chemodivergent, regiodivergent, and stereodivergent categories:
. Chemodivergent PKRs: include those reactions that yield two non-isomeric
compounds. In some cases, they are completely different, and one of them can
even be non-chiral (Scheme 2.5a) [18]. However, in most examples described to
1) The absolute configuration is assigned on the basis of the following Cahn–Ingold–Prelog priority
rules: the heteroatom is the substituent with the highest priority, followed by the large-sized
substituent, the medium-sized one, and the hydrogen atom.
Scheme 2.4 General scheme of a parallel kinetic resolution (PKR). A and B are the enantiomers of a
racemic substrate; P and Q are different products derived from A and B, respectively. The
obtainment of two chiral products is not mandatory.
date, the products of the chemodivergent PKR are pseudoenantiomers: two

products possessing all the stereocenters with opposite configuration and differ-
ing at a position far from them.
. Regiodivergent PKRs: include either those in which the substrate has the same
reacting functional group at different positions on the molecule or those in which a
single functional group leads to two regioisomeric compounds (Scheme 2.5b) [19].
. Stereodivergent PKRs: include those in which a new chiral center (of one fixed
configuration) is formed in both enantiomers of the molecule, thus generating
two different diastereomers (Scheme 2.5c) [20].
The quantification of enzymatic PKRs obviously differs from that of classical KRs,
and also differs among PKRs depending on the type of transformation involved. In
fact, E values are not the only parameters required to quantify them. The relative rates
of each resolution also play a key role. Nevertheless, if the E values of the transforma-
tions of a PKR are very high, the products are going to be obtained in high optical
purities. But when the E values are moderate, the eeps are more difficult to predict.
Several equations that relate the eeps to their concentrations have been developed [21].
(a)
O O O O O O
Baker's yeast
+
H3 CO H 3CO
O O O OH O
O
(b) O
Microorganism O
O + O
(c) O OH OH
PaHNL
H CN + CN
O O O
Scheme 2.5 Examples of (a) chemodivergent, (b) regiodivergent, and (c) stereodivergent
enzymatic PKRs.
50
However, these equations are specific for the transformation of a racemic substrate
containing an asymmetric center and a prochiral one.
Most examples of biocatalyzed PKRs published so far make use of oxidoreductases
as catalysts and ketones as substrates (either reduction – Chapter 26 – or Baeyer–
Villiger oxidation – Chapter 33 – to afford alcohols and esters, respectively). However,
examples of other enzymatic processes like the hydrolase-catalyzed nucleophilic
opening of epoxides (Chapter 9) and lyase-catalyzed addition of hydrogen cyanide to
racemic aldehydes (Chapter 23) have also been described.
2.2.3
Overcoming the Yield Limitation of KRs
So far, the processes shown start from a racemic mixture and, either only 50% of the
starting material is transformed into another compound (KRs) or 50% is transformed
in one product and the other 50% in another one (PKRs). Obviously, there is a
limitation concerning the atom efficiency of these transformations since half of the
material is discarded or has to be recycled (which comes with extra waste and
cost) [22]. One obvious way to avoid this problem is to use prochiral or meso
compounds, the symmetry of which allows a quantitative yield of a single enantiomer
of the product. However, racemic starting materials are more abundant in nature and
cannot always be avoided. This is the reason why, in recent years, organic chemists
have been trying to overcome the yield limitation of KRs, focusing rather on
answering the question of how to improve the economic balance of a KR by breaking
the 50% yield threshold of a single enantiomer.
2.2.3.1 Dealing with Prochiral or Meso Compounds: Desymmetrizations

The desymmetrization of symmetric compounds consists of a modification that
eliminates one or more elements of symmetry of the substrate. If the symmetry
elements that preclude chirality are eliminated, optically active products can be
obtained. For stereoselective enzymatic desymmetrizations (SEDs) a maximum yield
of 100% can be attained [23]. For this reason, they constitute a very interesting
alternative to KRs for the preparation of optically active compounds.
Meso and prochiral compounds have in common the presence of either two
enantiotopic groups or a planar trigonal group with two enantiotopic faces (Fig-
ure 2.2). An enantiotopic group is described as pro-R if the configuration of the
generated chiral center is assigned the stereodescriptor (R) as determined by
Cahn–Ingold–Prelog (CIP) priority [24]. Alternatively, the other group is described
as pro-S. Although this nomenclature is also applicable to the enantiotopic faces of a
trigonal system the Re and Si terminology is more often used. Thus, the stereo-
heterotropic face of a trigonal atom is designated Re if the ligands of the trigonal atom
appear in a clockwise sense in order of CIP priority when viewed from that side of the
face. The opposite arrangement is termed Si.
Despite the apparent differences, KRs and SEDs are quite similar from a
mechanistic point of view. Thus, although the substrate has no enantiomers, its
enantiotopic groups or faces also lead to diastereomeric transition states conceptually
Figure 2.2 Nomenclature of enantiotopic groups or faces of prochiral and meso compounds; X is a
C atom or a heteroatom and A, B, C, and D are substituents with decreasing CIP priority.
identical to those coming from the two enantiomers of a racemate. However,

experimental quantification of the enantioselectivity is different since the substrate
has no ee. For instance, Eq. (2.1) can, therefore, not be applied since it is not possible
to measure two KM values for the substrate. This fact is reflected in the different
names that have been used for the selectivity of a SED, like selectivity of the reaction
or prochiral selectivity. However, there is no contradiction in using the term
enantiomeric ratio, with the same sense as it is used in the enzymatic KR of racemic
mixtures. Unless the reaction is reversible or the product of the reaction is not stable
under the reaction conditions and it is further transformed, the enantiomeric ratio
can be determined as the ratio of the concentration of the major enantiomeric
product to that of the minor product because, in contrast to KRs, this ratio is constant
throughout the reaction [4]. However, an alternative formulation of E in terms of eep is
usually preferred (Eq. (2.7)). Indeed, eep itself can be simply employed as parameter to
compare selectivity among SEDs:
1 þ eep
E¼ ð2:7Þ
1eep
The reduction of prochiral ketones to afford alcohols (Chapter 26) is, by far, the
most explored and studied SED. Additionally, the hydrolase-catalyzed desymmetri-
zation of meso dicarboxylic acid derivatives with nucleophiles (such as water, alcohols,
ammonia, amines, hydrazine, peroxides, and thiols) (Chapter 8) has also been very
popular. Nevertheless, many other examples exist such as the Baeyer–Villiger
oxidation of prochiral ketones (Chapter 33) and many C–C bond formation reactions
(Chapter 10, Chapter 21, Chapter 22, and Chapter 23). However, in many cases, they
are not referred to as desymmetrizations [23].
52
2.2.3.2 (Cyclic) Deracemizations (CycDs)

Several strategies have been developed to allow the complete selective transformation
of one of the enantiomers of a racemic mixture into the other one, thus affording a
single stereoisomeric product in 100% theoretical yield. In general, it is acceptable to
refer to them as deracemization processes [25]. As shown in Scheme 2.6, dera-
cemization of enantiomer B affords untouched substrate enantiomer A as the sole
product. This process has also been denoted enantiomerization [26]. However, this
term has been more frequently used in the context of the interconversion of chirally
labile compounds, resembling racemization rather than deracemization.
Scheme 2.6 Deracemization processes via (a) direct stereoinversion and (b) cyclic
deracemization. A and B are enantiomers; I is an intermediate.
Of the various ways to perform a deracemization, a one-pot single-step protocol

via direct stereoinversion (Scheme 2.6a), is certainly the most challenging. Obviously,
the simplest and cleanest way would be the direct stereoinversion of one enantiomer
into the other in an irreversible manner (pathway A). However, in most cases, a one-
pot stereoinversion is achieved through a two-step sequence involving an interme-
diate and two (quasi)irreversible processes (pathway B).
A typical example of direct stereoinversion is the deracemization of secondary
alcohols. Although these compounds can be synthesized by reduction of the
corresponding prochiral ketones using chemical [27] or biocatalytic methods [28],
sec-alcohols are often more accessible than the corresponding ketones. Although the
number of examples for the single-step deracemization of these substrates employ-
ing whole microbial or plant cells has been steadily increasing over the past years,
there have been only speculations about the actual driving force of the stereoinver-
sion as well as its mechanism and the enzyme(s) involved.
For this type of processes, the most likely explanation is that the microbial
deracemization of secondary alcohols in a one-pot single-step system occurs via
an oxidation–reduction sequence through the corresponding prochiral ketone, which
is often detected in small amounts. Thus, with these facts in mind, it has been
proposed that in the cell machinery two (or even more) nicotinamide-dependent
alcohol dehydrogenases (ADHs) possessing opposite stereopreferences are
involved [29]. If one ADH shows (S)-preference, the other one has to exhibit (R)-
stereopreference [30]. Therefore, one enantiomer from the racemate is selectively
oxidized by a dehydrogenase (or oxidase), and the resulting ketone is reduced back
again by a different ADH displaying opposite stereopreference. In any event, based
on the data available to date, it appears more likely that the deracemization of sec-
alcohols follows different mechanisms in different microorganisms – the presence of
a single unique mechanism for all deracemization reactions seems rather unlikely.
Not all deracemizations of secondary alcohols, however, are carried out using
whole cells. Recently an impressive example of deracemization of sec-alcohols using
two purified enzymes with opposite stereopreference was reported that will be
discussed in more detail in Section 2.2.6 [31].
Many more examples have been described that employ cyclic deracemizations
(CycDs, Scheme 2.6b). In this case, enantiomer B is selectively transformed into
intermediate I, which is then non-selectively transformed into both enantiomers A
and B, resulting in a novel mixture with an increased concentration of enantiomer A.
Starting from a racemic mixture, one enantiomer can be obtained in high excess
(>95% ee) after only 4–5 cycles [32].
For instance, sec-alcohols can be deracemized by sequential highly selective
oxidation of one enantiomer followed by partial selective reduction of the ketone
intermediate with only one enzyme [33]. Although in theory, due to the reversibility of
the enzymatic reactions, the oxidation and reduction should show the same stereo-
preference, in these cases the amount of one enantiomer increases due to, for
example, a lower selectivity of the oxidation process followed by a perfect selective
reduction reaction (Scheme 2.7a). In this case, none of the reactions needs to be
irreversible since the net redox-balance of this process is zero and (in the ideal case)
no external cofactor recycling is necessary since NAD(P)H is internally recycled
between both steps. Other examples to deracemize secondary alcohols make use of
two enzymes, one reversible (e.g., ADH) and an irreversible one (e.g., alcohol
oxidase) [34].
Scheme 2.7 Examples of cyclic deracemizations to obtain (a) sec-alcohols using a single enzyme
and (b) chiral amines using an amine oxidase plus a chemical reducing agent.
The most common way to achieve cyclic deracemizations makes use of a

selective enzyme plus a non-selective chemical reagent. For instance, compounds
bearing a chiral amino group can be deracemized in a one-pot process by means
of an enantioselective oxidation of the rac-amine catalyzed by an amine oxidase,
yielding the prochiral imine, which in a subsequent step is (chemically) reduced
in a non-selective fashion to afford the enantioenriched amine. The cyclic
combination of both steps leads to a highly versatile deracemization technique
(Scheme 2.7b) [35].
Similar approaches have been used to deracemize secondary alcohols by coupling
alcohol dehydrogenases or oxidases with sodium borohydride [36], or a-amino acids
using amino acid oxidases plus chemical reducing agents [37].
54
2.2.3.3 Enantioconvergent Processes (ECPs)

Enantioconvergent processes (ECPs) furnish a single stereoisomeric product from a
racemate by a process in which one of the enantiomers of the substrate is trans-
formed through inversion of configuration. The second enantiomer reacts through
retention of configuration, thus proceeding through a stereochemically matching
pathway (Scheme 2.8). This means that the catalyst(s) must be enantioselective
(preferring one enantiomer over the other one), but also selective regarding retention
or inversion of the substrate configuration, what makes this type of transformations
not very common [38].
In contrast to deracemization processes both enantiomers of the substrate react to
produce a single final product (Scheme 2.8a), or one enantiomer is transformed into a
product with inversion of configuration while the other remains unmodified, thus
affording a homochiral final mixture (Scheme 2.8b). An example of the first case is
the two-enzyme catalyzed hydrolysis of styrene-type epoxides by using two different
epoxide hydrolases (EHs) employed as whole cell biocatalysts (Scheme 2.8c). In this
case Aspergillus niger transformed the (R)-oxirane with retention of configuration
whereas Beauveria bassiana hydrolyzed the antipode in an stereocomplementary
fashion yielding a single (R)-configured diol as the sole product [39]. More recent are
the corresponding examples of the second type of ECPs, where enantiopure sec-
alcohols can be obtained from the corresponding sulfate esters using sulfatases
(Scheme 2.8d) [40]. By the action of an enantioselective sulfatase, a mixture of the sec-
alcohol and the non-reacting sec-sulfate ester with the same configuration was
obtained [41]. Subsequently, chemical cleavage of the sulfate ester proceeding
through retention of configuration gave a single enantiomeric sec-alcohol as the
final product.
Scheme 2.8 Enantioconvergent processes homochiral mixture; (c) ECP to obtain chiral
(ECPs). (a) Both enantiomers afford the same diols using epoxide hydrolases; and (d) ECP to
product; (b) a single enantiomer reacts with obtain chiral sec-alcohols using sulfatases. A and
inversion of configuration, affording a B are enantiomers; P is the final product.
2.2.3.4 Dynamic Kinetic Resolutions (DKRs)

If, under the reaction conditions, both enantiomers of a racemic substrate can be
interconverted while the enantiomers of the product are stable, an enzymatic dynamic
fast
A P
slow
B Q
Scheme 2.9 General scheme of a dynamic kinetic resolution (DKR). A and B are the enantiomers of
a racemic substrate; P and Q are the enantiomers of the corresponding product.
kinetic resolution (DKR) can be carried out (Scheme 2.9) and a single enantiomer at
quantitative conversions can be attained [42]. Similarly to SEDs, under ideal condi-
tions, enzymatic DKRs afford the major product with a constant eep during the whole
process. Ideal conditions actually mean that the two enantiomers of the substrate
undergo equilibration much faster than the irreversible transformation of the fast-
reacting enantiomer of the substrate into the product. Therefore, the enzyme always
faces a racemic substrate regardless the degree of conversion of the reaction, and the
maximum eep of the product in a DKR is the initial eep of the corresponding KR.
Therefore, the Evalue of a KR limits the maximum eep of its dynamic version. However,
the rate of racemization is not always faster than that obtained for the enzymatic
transformation and, therefore, the eep usually varies with the degree of conversion of
the DKR. But, in practice, as long as the E value is high enough, the DKR has synthetic
utility. If this were not the case, the relative rates of the enzymatic transformation and
the racemization can be tuned by controlling the amount of enzyme and racemization
catalyst, and by making a sensible choice of the reaction conditions. Although a
mathematical treatment similar to that of classical KRs could be carried out, in general
the quality of DKRs is usually tested by comparing their eep and c (or yield) values.
The applicability of enzymatic DKRs is highly dependent on the ability of a given
substrate to undergo in situ racemization under conditions compatible with both the
stability and activity of the enzyme. Such racemization can be defined as the
equilibrium established by one enantiomer of a substrate with either the other
enantiomer or prochiral or meso species. This aspect constitutes the key difference
between racemization and deracemization concepts: the former proceeds towards
equal concentrations of enantiomers, while the latter does the opposite. Zwanenburg
and coworkers have classified the existing racemization processes according to their
mechanism into the following main categories [43]:
. Base-catalyzed: this can be in principle used with any compound bearing an acidic
hydrogen atom at the chiral center (i.e., a-substituted carbonyl groups).
. Schiff base-mediated: this is restricted to compounds bearing free primary amino
groups at the chiral center (i.e., a-amino acids).
. Thermal: This is in principle applicable to any compound that racemizes by
rotation or deformation of bonds (i.e., biaryls), by pyramidal inversion, or by
rearrangement of bonds. Its scope is, however, limited by the stability of the
substrate and, especially, the enzyme at the temperature of racemization.
56
. Enzymatic: The use of racemases to perform racemizations is limited by their

substrate specificity and their low stability in organic solvents. This approach has
been traditionally used with amino and hydroxy acids and their derivatives.
. Acid-catalyzed: this requires either (i) the protonation of a double bonded
heteroatom in an a-position to a chiral center followed by the abstraction of the
proton of the chiral center (i.e., keto-enol tautomerism) or (ii) the presence of a
group at the chiral center that, upon protonation, becomes a good leaving group
and affords a carbocation.
. Via redox and/or radical reactions: these are applicable to any compound that
undergoes a redox equilibrium via a prochiral or meso intermediate (i.e., alcohols
via carbonylic compounds).
. Nucleophilic substitution: this can be applied to chiral substrates that can undergo
SN or nucleophilic addition–elimination sequences. This method requires that
the substituent at the chiral center is a good leaving group as well as a good
nucleophile. An advantage of this method is that it can be used for compounds
without a hydrogen atom at the stereocenter.
Among all these methods, the base-catalyzed racemization is, probably, the most
widely applied in enzymatic DKRs. It has been used successfully with carboxylic
acid derivatives with a a-chiral center (a-substituted-b-keto esters and nitriles,
various 5-oxazolones, etc.) [42c]. However, since the discovery that transition metal
catalysts can be compatible with enzymatic catalysis, the scope of enzymatic DKRs
has dramatically changed, and the number of examples that include metal catalysts is
growing rapidly [42a,b,e]. In this case, the mechanism of racemization proceeds
mainly via either a redox or an addition–elimination process. Thus, alcohols can
efficiently be resolved via DKR by the combination of a lipase with oxidative Al, Ru,
and Rh species [42b]. Moreover, depending on the structure of the alcohol, the
racemization possibilities can differ. Thus, allylic alcohols can be racemized effi-
ciently via p-allyl-palladium complexes or vanadate intermediates. On the other hand,
amines can be resolved by combination of a hydrolase-catalyzed transformation with
a metal-catalyzed racemization either with Pd(0) [44] or Ru species [42b]. In this case,
the racemization proceeds via an imine intermediate.
2.2.3.5 Dynamic Kinetic Asymmetric Transformations (DYKATs): Types I and II

Recently, DYKAT (dynamic kinetic asymmetric transformation) processes have been
defined as the desymmetrization of racemic or diastereomeric mixtures involving
interconverting diastereomeric intermediates – implying different equilibration
rates of the stereoisomers – [45].
In this definition there are some remarkable points that should be addressed. First,
the DYKAT methodology cannot only be applied to mixtures of diastereoisomers but
also to mixtures of enantiomers if the intermediate(s) is(are) diastereoisomeric.
Consequently, although several processes have been frequently referred to as a DKR,
they should be more properly denoted as DYKAT, since they involve the (metal-
catalyzed) equilibration of diastereomeric (rather than enantiomeric) intermedi-
ates [46]. Second, these intermediates must have the ability to interconvert to obtain
an appropriate substrate equilibration. One could think of a diastereoisomeric
intermediate when a chiral acid interacts with a lipase to afford a diastereomeric
mixture of acyl-enzyme intermediates, but in this case both diastereoisomers cannot
interconvert, resulting therefore in a typical KR process (50% maximum product
yield) and not a DYKATone (100% maximum product yield). Finally, we point out that
during this and the following section we use the term deracemization not with the
meaning explained in Section 2.2.3.2, but rather with reference to the transformation
of a racemic mixture into a single enantiomer of a product. If the starting material is a
mixture of diastereomers, the term de-epimerization will be employed instead.
Although the different meanings of deracemization can cause some confusion
their usage is so widespread that, in practice, it is impossible to avoid.
Four types of DYKAT have been previously defined, two of them applied to
deracemization of enantiomers (types I and II, Scheme 2.10), and two applied to
the de-epimerization of diastereoisomers (types III and IV, Section 2.2.4). Trost et al.
introduced the first two types of DYKAT applied to the resolution of racemates
through diastereomeric intermediates, classifying them as type I (Scheme 2.10a) [47]
and II (Scheme 2.10b) [48].
In theory, the eep remains constant in ideal DKRs and is equal to the initial value of
the corresponding KR. This is because ees always remains zero, due to fast equil-
ibration of substrate enantiomers. In contrast, in a typical DYKAT system, substrates
are interconverted through diastereomeric complexes (e.g., in DYKAT type I through
Scheme 2.10 Dynamic kinetic asymmetric transformations (DYKATs): (a) type I and (b) type II.
A and B are enantiomers; P and Q are product enantiomers; ACat and BCat are diastereoisomeric
substrate–catalyst complexes; ICat is the chiral intermediate–catalyst complex.
ACat and BCat) and therefore the substrate enantiomers are not present in equal
amounts, due to the preferential formation of one of the substrate diastereoisomers.
From this, two different situations can be distinguished: if the equilibration step (e.g.,
kACat) leading to the major product of a DYKAT system (e.g., kP) is faster than that
forming the minor product (e.g., kBCat) the overall selectivity of the process will be
higher than the one observed for the KR, because the preferred transformation to
obtain the final product (P) matches to the enhanced formation of the correct
diastereomeric complex (ACat). In contrast, in a mismatch situation, where the
equilibration process does not favor the formation of complex ACat with regards to
BCat, the overall selectivity decreases (Figure 2.3).
For DYKAT type II, the intermediate ICat comes from both enantiomers; one chiral
center is then lost and thus stereoselectivity depends on the subsequent step that is
58
100
eep match DYKAT
eep DKR
eep mismatch DYKAT
ee [%]
E = 8.0
0
c [%]
Figure 2.3 Plot of ee versus c, showing the variation of eep for different types of DYKAT (dynamic
kinetic asymmetric transformation) and for DKR (dynamic kinetic resolution).
guided by the remaining chiral centers of the substrate and the enantiopure ligand. In
this case the first equilibration step shows stereoselectivity, but this selectivity is not
transferred into the final product. Although to date no biocatalytic example of this
type of system has been published, new approaches could fit in here in the near
future.
2.2.4
Dealing with Diastereomers: DYKATs Types III and IV
When chemists face a transformation on a mixture of diastereomers, a simple

selective reaction is not so desirable since the maximum theoretical yield of the final
product declines dramatically. Thus, a stereoselective transformation into a single
stereoisomer (denoted as de-epimerization) to enhance the conversion up to 100%
is highly valuable [25]. There is a fundamental difference between deracemizations
and de-epimerizations: while the former are isoenthalpic, for the latter a difference in
reaction enthalpy (DH) exists. This is because enantiomers have the same enthalpy
(DH ¼ 0) whereas diastereomers are different compounds with different physical
properties. Consequently, thermodynamically, de-epimerizations can be simpler
than deracemizations since the driving force can be provided by the difference in
DH. In contrast, kinetically, it is more difficult to achieve a good de-epimerization
than a proper deracemization since more stereoisomers are involved in the former
and, hence, the (bio)catalyst needs to be more selective. Historically, the first concept
of de-epimerization was developed for the selective crystallization of diastereomers
from in situ equilibrating mixtures of epimers [49].
Since the classification of DYKAT types I and II (Scheme 2.10), several novel
DYKAT protocols have been established and applied to the de-epimerization of
diastereomeric mixtures, which have been classified as type III (Scheme 2.11a) and
IV (Scheme 2.11b) [50]. The main difference between the four types is the resolution
of enantiomers (types I and II) or diastereomers (types III and IV).
Type III describes the de-epimerization of a diastereomeric mixture of enantio-
meric pairs. As all possible diastereoisomers are abundant, this system is much more
difficult to analyze than type I and II systems from a kinetic point of view. In this type
of transformations all diastereomeric substrate–catalyst complexes are in equilibri-
um and a second (quasi)irreversible reaction is responsible for the overall selectivity
(Scheme 2.11a). A typical example of biocatalyzed-mediated DYKAT type III is the
resolution of chiral 1,3- or 1,4-diols (Scheme 2.12a) using a ruthenium complex to
(a)
PRS PRR
k RS fast k RR slow
k RSCat k RS/RR k RRCat
ARS ARSCat ARRCat ARR
k RS/SS k RR/SR
ASS ASSCat ASRCat ASR

k SSCat k SS/SR k SRCat
k SS slow k SR slow
PSS PSR
(b)
PRS PRR
k RS fast k RR slow
k RSCat k RRCat
ARS ARSCat ARRCat ARR
k RS ' k RR '
C+D
achiral
k SS ' k SR '
ASS ASSCat ASRCat ASR

k SSCat k SRCat
k SS slow k SR slow
PSS PSR
Scheme 2.11 Dynamic kinetic asymmetric and ASSCat are diastereomeric

transformations (DYKATs): (a) type III and substrate–catalyst complexes; C and D are
(b) type IV. ARS, ARR, ASR, and ASS are achiral intermediates. For simplicity, it is
diastereomers; PRS, PRR, PSR, and PSS are assumed that substrate A has only two chiral
product diastereomers; ARSCat, ARRCat, ASRCat, centers.
60 j 2 Concepts in Biocatalysis
OH OH
(b) CO2H L-TyrDC
(a) NH2 NH2

OH OH OH
L-TA
Ph
HO O
H + H2N CO2H
HO OH
L-TA
OH OH
CO2H
NH2 L-TyrDC NH2
Scheme 2.12 (a) Examples of 1,3- and 1,4-diols resolved through DYKAT type III methodology; (b)
bi-enzymatic DYKAT type IV based on a reversible aldol reaction with L-threonine aldolase followed
by irreversible decarboxylation with L-tyrosine decarboxylase.
afford the interconversion of all diastereomers through an oxidation/reduction

process plus a stereoselective lipase-catalyzed acylation of the diol [51].
In DYKAT type IV, the de-epimerization of all possible diastereomers proceeds
through (reversible) destruction of both chiral centers, yielding two non-chiral inter-
mediates (Scheme 2.11b). In comparison to DYKAT type III, the biocatalytic type IV
systems described to date are simpler because the epimerization was already highly
stereoselective; consequently, only two out of four diastereoisomers were present in
measurable amounts. Thus, the only biocatalytic system of this type is the enzymatic
synthesis of amino alcohols starting from glycine and benzaldehyde using a L-threonine
aldolase (L-TA) coupled with a L-tyrosine decarboxylase (L-TyrDC) (Scheme 2.12b) [52].
2.2.5
Making it at Once: Cascade or Domino Processes
Clearly, there is growing industrial and academic interest in developing novel

strategies that respect the concepts of green chemistry and sustainability by
maximizing the efficiency of the new processes developed. An important contribu-
tion is the implementation of catalytic methodologies applied to fine chemicals
manufacture in a multistep synthetic fashion. The ultimate goal is to combine two or
more (bio)catalytic steps into a one-pot, multistep catalytic cascade (or domino)
process without isolation of the corresponding intermediates (Scheme 2.13) [53].
This elegant type of transformation presents several advantages such as the
optimized use of reagents and solvents, higher volumetric and space–time yields,
Scheme 2.13 One-pot cascade or domino processes.

Scheme 2.14 Recent examples of dehydrogenase (ADH) and a halohydrin

biocatalyzed-cascade processes: (a) synthesis dehalogenase (Hhe); (b) synthesis of chiral
of enantiopure b-azidoalcohols and a-hydroxyamides coupling a hydroxynitrile lyase
b-hydroxynitriles using an alcohol (HnL) with a nitrile hydratase (NHase).
and less waste. Thus, it is very appropriate from both environmental and eco-
nomical points of view, although it also has some drawbacks due to the difficulty in
finding suitable reaction conditions for all (bio)catalysts and reagents present in
the reaction medium. One remark must be made concerning the definition of
cascade (or domino) processes: all (bio)catalysts/reagents should be present at the
beginning of the overall transformation, because if other species must be added at
a later stage to the reaction vessel, it is more correct to mention this as a one-pot
process. Owing to this, many possibilities arise regarding the development of
cascade processes where at least one of the reactions is biocatalyzed. Thus,
Scheme 2.14 shows two very recent examples of biocatalyzed domino transforma-
tions. In the first one [54], several b-azidoalcohols and b-hydroxynitriles were
synthesized by Kroutil and coworkers starting from the corresponding a-chloro
ketones. In a first step these substrates were stereoselectively reduced by an ADH
to provide the enantiopure chlorohydrins that subsequently reacted with a halo-
hydrin dehalogenase (Hhe) to afford first the epoxide, which in the presence of a
nucleophile such as azide or cyanide reacted to yield the desired products due to
the action of the Hhe (Scheme 2.14a). In the second example [55], Sheldon and
coworkers reported the chiral synthesis of a-hydroxyamides or a-hydroxycar-
boxylic acids starting from several aldehydes, making them react in the presence
of cyanide with a selective hydroxynitrile lyase (HnL) to form the chiral cyanohy-
drins, followed by unselective nitrile hydratase (HNase)- or nitrilase (NLase)-
catalyzed hydrolysis to afford the final derivatives (Scheme 2.14b).
2.2.6
Novel Concepts
In recent years, and motivated by the same concern that has led to the development of
cascade or domino processes, the simultaneous use of biocatalysts to obtain enan-
tiopure compounds has become more relevant. This has resulted in routes that
somewhat resemble the cell metabolic pathways, and due to the ever-increasing
62
knowledge of biochemical routes and chemical biology tools more specific and
selective transformations are being developed. Some of these bioprocesses have been
defined as concurrent catalytic reactions and all the steps must carefully be balanced to
ensure comparable rates and that the different catalytic reactions do not interfere with
one another. Thus, these transformations circumvent the often time-intensive and
yield-reducing isolation and purification of intermediates in multistep syntheses.
However, it has to be kept in mind that cascade reactions are always linear reaction
sequences in which the substrate, intermediate(s), and final product are different
compounds, while concurrent processes can either yield the substrate of the reaction
as product or can involve parallel reactions with some kind of interconnection among
biocatalysts. For instance, some elegant examples of concurrent deracemizations
have been described recently. In a one-pot single-step process several a-aryl- and
a-aryloxy substituted propionic acids were deracemized by biocatalytic stereoinver-
sion via a three-step sequence: first the formation of an Acyl-CoA-derivative of the
acid, followed by epimerization of the latter to yield the opposite isomer, and finally
the hydrolysis back to the acid. The whole process involves an Acyl-CoA synthetase,
an epimerase, and a hydrolase, all enzymes that are present in the fatty acid
biosynthesis/degradation pathways [56]. Recently, Kroutil and coworkers published
a deracemization process of sec-alcohols by using either two enantiocomplementary
enzymes [31] or a microorganism plus an alcohol dehydrogenase [57]. Thus, in the
first case, by taking advantage of the different cofactor preferences of the enzymes
several secondary alcohols could be deracemized (even enantiopure substrates could
be stereoinverted) by using two ADHs (Scheme 2.15).
OH O OH
R * R' R R' R * R'

(R) (R)-selective ADH (S)-selective ADH (S)
+ + +
OH OH OH
R * R' NADP+ NADPH R * R' NADH NAD+ R * R'

(S) (S) (S)
NADP-selective NAD-selective
recycling system recycling system
Scheme 2.15 Deracemization of sec-alcohols by concurrent tandem oxidation–reduction cycles.
Finally, it should be mentioned that not all examples of concurrent catalysis imply
more than one catalyst. Thus, one single enzyme has been used in a one-pot tandem
biohydrogen transfer process to simultaneously obtain two enantiopure sec-alcohols
by using the racemic mixture of one of them and a prochiral ketone as starting
materials. The enzyme selectively both oxidizes one of the enantiomers of the alcohol
and reduces the ketone. The interconnection between the two processes is the
cofactor, which is reduced and oxidized in a cyclic way as long as the two reactions
References j63
proceed in parallel. Therefore, only catalytic amounts of the cofactor were required,
which minimizes dramatically the quantity of reagents employed for its recycling,
and several interesting building blocks could be easily obtained in an enantiocom-
plementary fashion [58].
2.3
Summary and Outlook
The first enzymatic resolution was described in 1903 by Dakin [59], but was not until
the 1980s that biocatalyzed kinetic resolutions were systematically studied to obtain
compounds in enantiopure form. This is now a well-established methodology since a
great number of biocatalysts are commercially available. Furthermore, molecular
biology tools are nowadays well implemented, providing a continuously growing
number of enzymes for organic syntheses. However, there is still a need to design
processes by which optically pure products can be obtained with 100% yield. In this
sense, chemists have in the last few years developed novel techniques that can
overcome the yield limitation of KRs. Initially, special attention was paid to prochiral
or meso compounds since they can be desymmetrized by a single catalyst to afford
enantiopure derivatives in quantitative yields. Unfortunately, when a chemist is
confronted with the synthesis of a target molecule, symmetric starting materials are
not always available. In most cases, racemic mixtures cannot be avoided and,
therefore, novel concepts had to be designed (and implemented) to fulfill the
requirements of industry. CycDs, ECPs, DKRs, and DYKATs are examples of such
processes, which are gaining in relevance due to recent work focusing on the
optimization of (bio)catalysts.
On the other hand, the need for more sustainable processes based on highly
efficient transformations is turning the attention of the scientific community to the
combination of several (bio)catalytic steps to perform one-pot, multistep catalytic
cascade processes without isolation of the corresponding intermediates. Thus,
inspired by biochemical routes, several biocatalysts can be combined and work in
a multistep concurrent and orchestrated fashion to afford the desired products.
Despite the inherent difficult setup of such processes, we are convinced that the
advantages they offer are going to make multistep biocatalysis of utmost relevance in
the near future.
References
1 (a) Bommarius, A.S. and Riebel, B.R. (eds) Amsterdam; (d) Gotor, V., Alfonso, I., and
(2004) Biocatalysis, Wiley-VCH Verlag Garcıa-Urdiales, E. (eds) (2008) Asymmetric
GmbH, Weinheim; (b) Patel, R.N. (ed.) Organic Synthesis with Enzymes, Wiley-VCH
(2007) Biocatalysis in the Pharmaceutical and Verlag GmbH, Weinheim; (e) Fessner,
Biotechnology Industry, CRC Press, Boca W.-D. and Anthonsen, T. (eds) (2009)
Raton, FL;(c) Matsuda, T. (ed.) (2007) Future Modern Biocatalysis, Wiley-VCH Verlag
Directions in Biocatalysis, Elsevier, GmbH, Weinheim.
64
2 (a) Jeannette, B., Rousseau, C., Microb. Technol., 19, 328–331; (b) Zuegg, J.,
Marinescu, L., and Bols, M. (2008) Appl. H€onig, H., Schrag, J.D., and Cygler, M.
Microbiol. Biotechnol., 81, 1–11; (b) Breslow, (1997) J. Mol. Catal. B: Enzym., 3, 83–98;
R. (2009) J. Biol. Chem., 284, 1337–1342. (c) Tuomi, W.V. and Kazlauskas, R.J. (1999)
3 Fehrst, A.R. (ed.) (1999) Structure and J. Org. Chem., 64, 2638–2647.
Mechanism in Protein Science, W. H. 15 Pleiss, J., Fischer, M., and Schmid, R.D.
Freeman and Company, New York. (1998) Chem. Phys. Lipids, 93, 67–80.
4 Jacobsen, E.E. and Anthonsen, T. (2009) 16 (a) Ahmed, S.N., Kazlauskas, R.J.,
Factors affecting enantioselectivity: Morinville, A.H., Grochulski, P.,
allosteric effects in Modern Biocatalysis Schrag, J.D., and Cygler, M. (1994)
(eds W.-D. Fessner and T. Anthonsen), Biocatalysis, 9, 209–225; (b) Arroyo, M. and
Wiley-VCH Verlag GmbH, Weinheim, Sinisterra, J.V. (1994) J. Org. Chem., 59,
pp. 96–107. 4410–4417; (c) Berglund, P., Homquist,
5 International Union of Pure and Applied M., H€ogberg, H.E., and Hult, K. (1995)
Chemistry (1996) Pure & Appl. Chem., Biotechnol. Lett., 17, 55–60; (d) Holmquist,
68 (12), 2219. Available at http://www M., Hæffner, F., Norin, T., and Hult, K.
.iupac.org/publications/pac/1996/pdf/ (1996) Prot. Sci., 5, 83–88; (e) Berglund, P.,
6812x2193.pdf. Holmquist, M., and Hult, K. (1998) J. Mol.
6 Chen, C-.S., Fujimoto, Y., Girdaukas, G., Catal. B: Enzym., 5, 283–287.
and Sih, C.J. (1982) J. Am. Chem. Soc., 17 Dehli, J.R. and Gotor, V. (2002) Chem. Soc.
104, 7294–7299. Rev., 31, 365–370.
7 Straathof, A.J.J. and Jongejan, J.A. (1997) 18 Brooks, D.W., Wilson, M., and Webb, M.
Enzyme Microb. Technol., 21, 559–571. (1987) J. Org. Chem., 52, 2244–2248.
8 Rakels, J.L.L., Straathof, A.J.J., and 19 Alphand, V. and Furstoss, R. (1992) J. Org.
Hiejnen, J.J. (1993) Enzyme Microb. Chem., 57, 1306–1309.
Technol., 15, 1051–1056. 20 Bianchi, P., Roda, G., Riva, S., Danieli, B.,
9 Bornscheuer, U.T. and Kazlauskas, R.J. Zabelinskaja-Mackova, A., and Griengl, H.
(eds) (2005) Hydrolases in Organic (2001) Tetrahedron, 57, 2213–2220.
Synthesis: Regio- and Stereoselective 21 Dehli, J.R. and Gotor, V. (2002) Arkivoc, V,
Biotransformations, 2nd edn, Wiley-VCH 196–202.
Verlag GmbH, Weinheim. 22 (a) Trost, B.M. (1991) Science, 254,
10 Kazlauskas R.J., Weissfloch, A.N.E., 1471–1477; (b) Sheldon, R.A. (2000) Pure
Rappaport, A.T., and Cuccia, L.A. (1991) Appl. Chem., 72, 1233–1246.
J. Org. Chem., 56, 2656–2665. 23 (a) Garcıa-Urdiales, E., Alfonso, I., and
11 (a) Hæffner, F., Norin, T., and Hult, K. Gotor, V. (2005) Chem. Rev., 105, 313–354;
(1998) J. Biophys., 74, 1251–1262; (b) Garcıa-Urdiales, E., Alfonso, I., and
(b) Orrenius, C., Hæffner, F., Gotor, V. (2011) Chem. Rev., 111,
€
Rotticci, D., Ohrner, N., Norin, T., and PR110–PR180.
Hult, K. (1998) Biocatal. Biotransform., 24 (a) Cahn, R.S., Ingold, C.K., and Prelog, V.
16, 1–15. (1966) Angew. Chem., Int. Ed. Engl., 5, 385;
12 For a recent review on mirror-image (b) Prelog, V. and Helmchen, G. (1982)
enzyme active sites see: Mugford, P.F., Angew. Chem., Int. Ed. Engl., 21, 567.
Wagner, U.G., Jiang, Y., Faber, K., and 25 (a) Faber, K. (2001) Chem. Eur. J., 7,
Kazlauskas, R.J. (2008) Angew. Chem. Int. 5004–5010; (b) Gruber, C.C., Lavandera,
Ed., 47, 8782–8793. I., Faber, K., and Kroutil, W. (2006) Adv.
13 Ollis, D.L., Cheah, E., Cygler, M., Synth. Catal., 348, 1789–1805.
Dijkstra, B., Frolow, F., Franken, S.M., 26 Soda, K., Oikawa, T., and Yokoigawa, K.
Harel, M., Remington, S.J., Silman, I., (2001) J. Mol. Catal. B: Enzym., 11,
Schrag, J., Sussman, J.L., Verschueren, 149–153.
K.H.G., and Goldman, A. (1992) Protein 27 (a) Noyori, R. and Ohkuma, T. (2001)
Eng., 5, 197–211. Angew. Chem. Int. Ed., 40, 40–73; (b) Ma,
14 €
(a) Ohrner, N., Orrenius, C., Mattson, A., Y., Liu, H., Chen, L., Cui, X., Zhu, J., and
Norin, T., and Hult, K. (1996) Enzyme Deng, J. (2003) Org. Lett., 5, 2103–2106.
References j65
28 (a) Patel, R.N. (2001) Curr. Opin. 35 (a) Turner, N.J. (2003) Curr. Opin.
Biotechnol., 12, 587–604; (b) Kroutil, W., Biotechnol., 14, 401–406; (b) Turner, N.J.
Mang, H., Edegger, K., and Faber, K. (2004) Curr. Opin. Chem. Biol., 8, 114–119;
(2004) Curr. Opin. Chem. Biol., 8, 120–126; (c) Fotheringham, I., Archer, I., Carr, R.,
(c) Buchholz, S. and Gr€oger, H. (2007) in Speight, R., and Turner, N.J. (2006)
Biocatalysis in the Pharmaceutical and Biochem. Soc. Trans., 34, 287–290.
Biotechnology Industry (ed. R. N. Patel), 36 Oikawa, T., Mukoyama, S., and Soda, K.
CRC Press, Boca Raton, FL, pp. 757–790; (2001) Biotechnol. Bioeng., 73, 80–82.
(d) de Wildeman, S.M.A., Sonke, T., 37 (a) Huh, J.W., Yokoigawa, K., Esaki, N.,
Schoemaker, H.E., and May, O. (2007) Acc. and Soda, K. (1992) J. Ferment. Bioeng., 74,
Chem. Res., 40, 1260–1266; (e) Goldberg, 189–190; (b) Alexandre, F.-R., Pantaleone,
K., Schroer, K., L€utz, S., and Liese, A. D.P., Taylor, P.P., Fotheringham, I.G.,
(2007) Appl. Microbiol. Biotechnol., 76, Ager, D.J., and Turner, N.J. (2002)
237–248; (f) Huisman, G.W., Liang, J., Tetrahedron Lett., 43, 707–710;
and Krebber, A. (2010) Curr. Opin. Chem. (c) Beard, T.M. and Turner, N.J. (2002)
Biol., 14, 122–129. Chem. Commun., 246–247.
29 (a) Nakamura, K., Inoue, Y., Matsuda, T., 38 Wallner, S.R., Pogorevc, M.,
and Ohno, A. (1995) Tetrahedron Lett., 36, Trauthwein, H., and Faber, K. (2004) Eng.
6263–6266; (b) Carnell, A.J. (1999) Adv. Life Sci., 4, 512–516.
Biochem. Eng. Biotechnol., 63, 57–72; 39 Pedragosa-Moreau, S., Archelas, A., and
(c) Chadha, A. and Baskar, B. (2002) Furstoss, R. (1993) J. Org. Chem., 58,
Enantioselective biocatalytic reduction of 5533–5536.
ketones for the synthesis of optically active 40 Gadler, P., Glueck, S.M., Kroutil, W.,
alcohols. Tetrahedron: Asymmetry, 13, Nestl, B.M., Larissegger-Schnell, B.,
1461–1464. Ueberbacher, B.T., Wallner, S.R., and
30 (a) Takemoto, M. and Achiwa, K. (1998) Faber, K. (2006) Biochem. Soc. Trans., 34,
Chem. Pharm. Bull., 46, 577–580; 296–300.
(b) Azerad, R. and Buisson, D. (2000) Curr. 41 Gadler, P. and Faber, K. (2007) Trends
Opin. Biotechnol., 11, 565–571. Biotechnol., 25, 83–88.
31 Voss, C.V., Gruber, C.C., Faber, K., 42 (a) Kim, M.-J., Ahn, Y., and Park, J. (2002)
Knaus, T., Macheroux, P., and Kroutil, W. Curr. Opin. Biotechnol., 13, 578–587;
(2008) J. Am. Chem. Soc., 130, (b) Pamies, O. and B€ackvall, J.-E. (2003)
13969–13972. Chem. Rev., 103, 3247–3261; (c) Pamies, O.
32 Kroutil, W. and Faber, K. (1998) and B€ackvall, J.-E. (2004) Trends
Tetrahedron: Asymmetry, 9, 2901–2913. Biotechnol., 22, 130–135; (d) Turner, N.J.
33 Cardus, G.J., Carnell, A.J., Trauthwein, H., (2004) Curr. Opin. Chem. Biol., 8, 114–119;
and Riermeir, T. (2004) Tetrahedron: (e) Martın-Matute, B. and B€ackvall, J.-E.
Asymmetry, 15, 239–243. (2007) Curr. Opin. Chem. Biol., 11,
34 (a) Azerad, R. and Buisson, D. (1992) 226–232; (f) Pellissier, H. (2011)
Stereocontrolled reduction of b-ketoesters Tetrahedron, 67, 3769–3802; (g) Kim, Y.,
with geotrichum candidum in Microbial Park, J., and Kim, M.-J. (2011)
Reagents in Organic Synthesis, vol. 381, ChemCatChem, 3, 271–277.
NATO ASI Series C (ed. S. Servi), Kluwer 43 Ebbers, E.J., Ariaans, J.A.,
Academic Publishers, Dordrecht, pp. Houbiers, J.P.M., Bruggink, A., and
421–440; (b) Ogawa, A.J., Zwanenburg, B. (1997) Tetrahedron, 53,
Xie, S.-X., Shimizu, S. (1999) Biotechnol. 9417–9476.
Lett, 21, 331–335; (c) Chadha, A. and 44 Reetz, M.T. and Schimossek, K. (1996)
Baskar, B. (2002) Tetrahedron: Asymmetry, Chimia, 50, 668–669.
13, 1461–1464; (d) Nie, Y., 45 Trost, B.M., Bunt, R.C., Lemoine, R.C.,
Xu, Y., and Mu, X.Q. (2004) Org. Process and Calkins, T.L. (2000) J. Am. Chem. Soc.,
Res. Dev., 8, 246–251; (e) Baskar, B., 122, 5968–5976.
Pandian, N.G., Priya, K., and Chadha, A. 46 (a) See, for instance: Luessem, B.J. and
(2005) Tetrahedron, 61, 12296–12306. Gais, H.-J. (2004) J. Org. Chem., 69,
66
4041–4052; (b) Mohr, J.T., Behenna, D.C., Reisinger, C., Fesko, K., Mink, D.,
Harned, A.M., and Stoltz, B.M. (2005) and Griengl, H. (2007) Angew.
Angew. Chem. Int. Ed., 44, 6924–6927. Chem. Int. Ed., 46, 1624–1626;
47 (a) Trost, B.M. and Toste, F.D. (1999) (b) Steinreiber, J., Sch€urmann, M., van
J. Am. Chem. Soc., 121, 3543–3544; Assema, F., Wolberg, M., Fesko, K.,
(b) Braun, M. and Kotter, W. (2004) Angew. Reisinger, C., Mink, D., and Griengl, H.
Chem. Int. Ed., 43, 514–517; (c) Hughes, (2007) Adv. Synth. Catal., 349,
D.L., Palucki, M., Yasuda, N., Reamer, 1379–1386.
R.A., and Reider, P.J. (2002) J. Org. Chem., 53 Garcia-Junceda, E. (ed.) (2008) Multi-Step
67, 2762–2768. Enzyme Catalysis, Wiley-VCH Verlag
48 (a) Trost, B.M., Patterson, D.E., and GmbH, Weinheim.
Hembre, E.J. (1999) J. Am. Chem. Soc., 54 Schrittwieser J.H., Lavandera, I., Seisser,
121, 10834–10835; (b) Trost, B.M. (2002) B., Mautner, B., and Kroutil, W. (2009) Eur.
Chem. Pharm. Bull., 50, 1–14. J. Org. Chem., 2293–2298.
49 Anderson, N.G. (2005) Org. Process Res. 55 van Pelt, S., van Rantwijk, F., and Sheldon,
Dev., 9, 800–813. R.A. (2009) Adv. Synth. Catal., 351,
50 Steinreiber, J., Faber, K., and 397–404.
Griengl, H. (2008) Chem. Eur. J., 14, 56 (a) Kato, D., Mitsuda, S., and Ohta, H.
8060–8072. (2002) Org. Lett., 4, 371–373;
51 (a) Edin, M., Steinreiber, J., and (b) Mitsukura, K., Yoshida, T., and
B€ackvall, J.-E. (2004) Proc. Natl. Acad. Sci. Nagasawa, T. (2002) Biotechnol. Lett., 24,
U.S.A., 101, 5761–5766; 1615–1621.
(b) Martin-Matute, B. and B€ackvall, J.-E. 57 Voss, C.V., Gruber, C.C., and Kroutil, W.
(2004) J. Org. Chem., 69, 9191–9195; (2008) Angew. Chem. Int. Ed., 47, 741–745.
(c) Edin, M., Martin-Matute, B., and 58 Bisogno, F.R., Lavandera, I., Kroutil, W.,
B€ackvall, J.-E. (2006) Tetrahedron: and Gotor, V. (2009) J. Org. Chem., 74,
Asymmetry, 17, 708–715. 1730–1732.
52 (a) Steinreiber, J., Sch€urmann, M., 59 Dakin, H.D. (1903) J. Physiol., 30,
Wolberg, M., van Assema, F., 253–263.
j67
3
Discovery of Enzymes
Wolfgang Aehle and Juergen Eck
3.1
Introduction
3.1.1
Historical Overview
The principle of enzyme catalysis was used by men before the underlying mechan-
isms were understood. The Greek author Homer describes milk clotting induced by
fig juice, which contains, as we know now, the protease ficin (EC 3.4.22.3) that
induces the clotting. In fact many traditional preparations of food depend on the use
of enzymes, including diverse processes like milk processing into cheese, the
maturing of meat, or the production of pickled herring.
It took until 1833 before two French scientists, Anselme Payen and Jean François
Persoz, discovered a starch liquefaction principle (diastase from barley), tiny
amounts of which could liquefy large amounts of starch, a feature that is one of
the characteristics of a catalyst. After evidence had built up that there were catalytic
materials that could be isolated from living matter, the German physiologist Wilhelm
Friedrich K€ uhne coined in 1878 the artificial word enzyme for these materials, a
word derived from the Middle Greek word enzymous, which means leaven or from
yeast.
Subsequently, enzymes were purified and crystallized, which made clear that
enzymes are catalytic proteins. A milestone in understanding the function and nature
of enzymes was the determination of the 3D structure of lysozyme (EC 3.2.1.17) by
David C. Phillips et al. in 1965 [1]. Thenceforward it was possible to decipher the
mechanism of enzyme catalysis on an atomic level not only through structure
determination with X-ray techniques but later also by using nuclear magnetic
resonance (NMR).
The assignment of enzyme function is one of the most cumbersome tasks in
biochemistry. Traditionally, the functionality was either clear from the function that
led to its discovery, such as the screening for amylolytic or proteolytic function, or
from the physiological context that suggested the catalytic role of the enzyme. In the
and Oliver May.
j 3 Discovery of Enzymes
68
1950s it became evident that the number of newly identified enzymes was increasing
rapidly and that a classification was needed to maintain an overview over the ever
increasing functionalities. Consequently, the International Union of Biochemistry
(IUB) in consultation with the International Union of Pure and Applied Chemistry
(IUPAC) decided to establish the International Commission of Enzymes, which
began operation in 1956. The task of the commission was laid down as follows:
To consider the classification and nomenclature of enzymes and coenzymes,

their units of activity and standard methods of assay, together with the
symbols used in the description of enzyme kinetics. [2]
As a result of these activities, today most enzymes can be grouped by their function
into one of six main groups and are identified by the enzyme classification numbers
(EC). The six groups of the enzyme classification system are oxidoreductases (EC 1),
transferases (EC 2), hydrolases (EC 3), lyases (EC 4), isomerases (EC 5), and ligases
(EC 6). The EC numbering system is a hierarchical system, which can be exemplified
by the protease tripeptide aminopeptidase with an EC number 3.4.11.4. It is a
hydrolase (EC 3) that acts on peptide bonds (EC 3.4) and cleaves off the amino acids
from the amino end (EC 3.4.11), while acting on tripeptides as substrates (EC
3.4.11.4).
Classification of enzymes according to the EC classification is the most exact
definition that can be used. A correct assignment of a function in line with the EC
classification requires a careful study of the enzyme, preferably the purified enzyme,
to avoid any misinterpretation or wrong assignment of a function. This classical
annotation of a protein sequence with a function is still the most reliable and correct
annotation method, but it is limited to enzymes that can actually be produced in
amounts that allow for their characterization. The Braunschweig Enzyme Database
(BRENDA) (http://www.brenda-enzymes.info) [3] is a comprehensive database of
enzyme properties; BRENDA is based on the more than 4900 different enzyme types
that are classified in the EC system. It contains functional data for all these enzymes
based on extensive literature evaluation.
The traditional annotation of enzyme function has though a capacity limit and it
can never account for the intrinsic promiscuity of enzymes with respect to substrate
specificity. The substrate promiscuity of enzymes describes the observation that a
particular enzyme can use its fundamental catalytic mechanism for the conversion of
more than one substrate into a different range of products. As a consequence the
activity-based annotation of enzyme function cannot be used for the tremendous
amount of protein sequences uncovered by the various sequencing activities in what
amounts to an exponential growth of data. It has therefore become necessary to
develop techniques to assign a function to a protein sequence based on the sequence
information alone. This functional annotation can use sequence similarity or 3D-
structural similarity as the starting point. The assignment process begins with the
identification of a sequence family or structural family that can accommodate the new
sequence. Sequence identity or structural similarities are complementary to each
other. The iterative usage of sequence-based and structural alignment enables the
3.1 Introduction j69
user to align sequences of an identity as low as 20%, which allows the grouping into
one of the enzyme families. The assignment of an enzyme to a given family – be it
structure-based or based on sequence identity – gives a hint as to the function of this
enzyme as long as the (catalytic) function of at least one member of that particular
family has been clearly defined.
The most comprehensive database for protein sequences is the Universal
Protein Resource Knowledgebase (UniProt) (http://www.uniprot.org) [4, 5] that
contains two sets of protein sequence data. The TrEMBL-dataset contains protein
sequences with computationally annotated functions, while the Swiss-Prot dataset
consists of protein sequences that have been manually annotated by evaluating
and reviewing the available literature data. The two databases in UniProt are not
redundant in the sense that every sequence that enters SwissProt is removed from
the TrEMBL set. A BLAST-search with a newly discovered sequence against the
UniProt protein sequence database can suggest a function for the new protein,
which is not necessarily its actual catalytic function in the proteome of the source
organism.
The annotation of protein function in UniProt is supported by tools such as Prosite
and Pfam. Prosite (http://www.expasy.org/prosite) [6] is a database of manually
defined sequence patterns or rules that enable the automated detection of these
patterns in newly discovered protein sequences. The presence of one or more
sequence patterns in a protein sequence allows us to classify the new sequence
into one of the Prosite families, which gives an indication about its possible function.
Pfam (http://pfam.sanger.ac.uk) [7] is a large collection of multiple sequence
alignments and hidden Markov models covering many common protein domains. It
helps in identifying enzyme function through the identification of domains in a
protein. Together Prosite and Pfam give useful hints for the annotation of a new
sequence.
The Structural Classification of Proteins (SCOP) (http://scop.mrc-lmb.cam
.ac.uk/scop/) [8] uses the evolutionary and structural similarity of proteins to
classify them. SCOP is mainly built on the visual inspection and comparison of
protein 3D structures. SCOP uses a hierarchical system. On the lowest levels of this
hierarchy are families of related domains. The related families are in turn grouped
into superfamilies. These first two levels cover homologous proteins sharing an
evolutionary origin. The next higher level of the SCOP hierarchy consists of groups
of superfamilies that share a given fold, but may include analogous sequences. The
highest level constitutes the secondary structure classes, embracing similar fold
types.
Other databases focus on a specialized area of enzyme function. MEROPS (http://
merops.sanger.ac.uk) [9], for instance, classifies proteolytic enzymes and proteina-
ceous protease inhibitors according to sequence identity. Each protease is assigned to
a protein family based on sequence identity. The family name contains a letter, which
represents the principal catalytic residue, and a number to separate families with
identical catalytic residues. Families are combined in clans based on homology. It is
assumed that all families of a MEROPS clan have evolved from the same origin. The
approximately 176 peptidase families belong to 42 protease clans.
70
The Carbohydrate-Active EnZymes (CAZY) database (http://www.cazy.org/) [10]

focuses on enzymes that act on carbohydrates. CAZY is broader in scope than
MEROPS, because it includes all enzymes that modify bonds in carbohydrates. Like
the other databases CAZY uses sequence similarity, domain composition, and
function to classify enzymes, which offers helpful information for the functional
annotation of new enzymes.
Even though the mentioned databases offer a very detailed and careful analysis and
classification of the known sequences, their reliability as predictive tools has its
limitations. The unambiguous identification of a sequence as belonging to a given
CAZY family does not necessarily allow us to predict its exact catalytic function. For
example, the most well-known members of the glycoside hydrolase family 13 (GH13)
are the a-amylases (EC 3.2.1.1), which – with only very few exceptions – all belong to
this family. However, not all members of the GH13 family are a-amylases. Besides
various glycoside hydrolases that gave the family its name, the GH13 family contains
glycosyl transferases and glycosyl synthases. These are catalytic functions that belong
to a different EC class than the name-giving hydrolases.
The vast amount of new sequences, whose number is still growing exponentially,
has resulted in the inability to classify approximately 15% of new protein sequences.
These 15% of all new protein sequences have been dubbed the dark matter of the
protein universe. It might be possible to reduce this level of dark protein matter by
increasing efforts to determine 3D structures by X-ray or NMR techniques or by using
ab initio modeling of the 3D structures of these new proteins [11].
Any annotation of protein function through the use of even the most sophisticated
computational tools available today, however, finds its limitation as soon as more
information about the substrate range, regiospecificity, or enantioselectivity of a new
enzyme is needed. There are still no tools available that reveal information about
enzyme parameters such as pH dependency or the influence of temperature or
solvents on the enzymes performance. To date, such parameters can only be reliably
determined by traditional wet-lab experiments.
Nevertheless the approach to the development of new enzymes has changed
dramatically. In the early days of enzyme discovery research had to depend on a
limited number of enzyme activities that was accessible through an activity-based
screen. It was the main task of enzyme development to find a compromise between
the requirements of the application and the limitations of the available enzymes.
Today it is possible to define an enzyme application with all its physical limitations
and select the right enzyme from a plethora of available functionalities.
3.1.2
The Ideal Enzyme Concept
In the chemical industries, researchers and process engineers were forced for a long
time to use enzymes derived from a relatively small range of cultivable organisms
(<1% of bacterial cells present in a given habitat; [12]). This often required the
adaptation of an existing process to the requirements of the enzyme such as its
stability or pH spectrum. As a result the conditions for the enzyme-catalyzed
Perfomance profile of two industrial enzymes
Parameter importance for process
6
0
temperature stability
pH stability
solvent stability
ingredient/by-product stability
substrate range
substrate specificity (KM, kcat/KM)
substrate selectivity (regio-, enantio-)
substrate conversion (%), yield
product inhibition
pH profile
temperature profile
specific activity
turnover frequency (kcat)
producibility/expression yield
by-product ingredient inhibition
space–time yield
Figure 3.1 Performance profiles of two synthesis purposes). The x-axis shows enzyme
industrial enzymes for different technical parameters. The y-axis displays the importance
applications (solid line: enzyme for detergent of a parameter for the process in arbitrary units
applications, dotted line: enzyme as catalyst for (1 ¼ low, 6 ¼ high).
process were often a compromise between process needs and enzyme capabilities.
This could lead to the unpleasant situation in which the enzymatic process was
economically not competitive despite potential benefits concerning the environ-
ment and sustainability.
With an increasing number of available enzymatic activities, however, it has
become increasingly feasible to determine optimal conditions for the process and
to deduce the specifications for the most suitable – the ideal – enzyme (a paradigm
shift [13]). As an example, Figure 3.1 shows the specifications for two typical technical
enzymes. While the detergent industry requires enzymes with broad substrate
specificity and no enantioselectivity at all, the requirement list for an esterase for
synthesis purposes is dominated by the demand for regio- and enantioselectivity and
of course high turnover to allow for a high space–time yield in a synthetic process.
A good definition of the required properties of the desired enzyme enables a
screening strategy that is directed to an enzyme that exactly fulfils the needs of an
ideal process. Such an enzyme will in general perform much better in that process
than an enzyme that is either already known as an industrial enzyme or was selected
originally for another application and only shows the required reaction type. An ideal
72
enzyme is an enzyme that is only limited by the diffusion of its substrate and/or
product to and from the enzyme. Such an enzyme, which has attained catalytic
perfection, has a kcat/KM between 108 and 109 M1 s1. A single enzyme will not be
catalytically perfect (i.e., ideal) for more than one process, unless the reaction
conditions are very similar.
3.2
Exploiting Functional Sequence Space: Resources and Screening Strategies
3.2.1
Resources for Enzyme Discovery
Traditionally, the resource for enzyme discovery has consisted of all proteins that are
accessible from living, cultivable organisms. Traditional enzyme sources include
mammals, such as bovine calves (chymosin), or plants (bromelain). Nowadays there
is a high prevalence of enzymes of microbial origin in technical enzyme applications.
For enzyme discovery, (micro)-organisms are cultivated and tested for the production
of the catalytic function of interest. This method has been very successful in the past.
Not only have most of todays industrial enzymes been discovered via this traditional
method but also a high number of secondary metabolites, which are used for instance
as antibiotic compounds in pharmaceutical applications.
The natural diversity of cultivable microorganisms has proven to be a good
resource for new proteins. The accessibility of this resource, however, depends on
the ability to cultivate organisms under laboratory conditions. It has been estimated
that less than 1% of all microbial species can be cultivated by man [14, 15].
Consequently, the vast majority of the existing microbial biodiversity cannot
be exploited with traditional approaches. Modern molecular biology has therefore
developed methods to assess the vast amount of non-cultivable biodiversity. These
technologies allow us to screen or sequence the so-called metagenome (a term coined
by Handelsman et al. [16]) of various habitats. A metagenome is the collective
genomic information of all microorganisms (bacteria, fungi, algae, protists, and
archaea) indigenous in a given habitat at a given (sampling) time point. There are
habitats with a wide genomic diversity like uncultivated forest soils or pasture land,
which contain the genomic equivalent of 6000 to 8000 genomes of Escherichia coli per
cm3 of soil. In contrast, the diversity in ecological niches with a high selective
pressure is generally much lower. For example, in a salt-crystallizing pond only seven
E. coli genomic equivalents were found per cm3 of soil [14]. The genetic information
of a given metagenome can be recovered by directly isolating metagenomic DNA
from environmental samples without the need of cultivation. The DNA can then be
deposited in gene libraries that are subsequently screened for desired enzymatic
activities. Lorenz and Eck have compiled enzymatic activities recovered by this
approach [17].
Acquiring genomic information from metagenome samples requires several
sophisticated technologies. Metagenomic sequences tags (MSTs) can not only be
3.2 Exploiting Functional Sequence Space: Resources and Screening Strategies j73
obtained from cloned DNA fragments but also directly from purified metagenome
DNA. Since, in this case, the remaining parts of the recovered genes are not physically
linked to specific clones and, consequently, cannot be obtained by sequencing, other
methods have to be employed to yield full-length genes. Various gene library-
independent methods for identification of up- and downstream regions in single
genomes or DNA from low-diversity habitats have been described previously,
including thermal asymmetric interlaced PCR [18], adapter-mediated PCR [19],
panhandle PCR [20], universal fast walking [21], and SON-PCR [22]. Recently, the
drawbacks of these methods concerning rare sequences and high background in
complex metagenomes have been addressed using magnetic bead capture of the
initial partial gene fragments in combination with inverse PCR or subtractive
hybridization strategies [23, 24].
The massive sequencing of nucleic acids as a way to approaching a global survey of
metagenomic DNA from environmental sources is technically feasible [25]. Elaborate
algorithms subsequently serve to identify open reading frames in silico and detect
related sequence entries in databases [26, 27].
Random mutagenesis on a whole gene or in defined segments of it is the
fundamental method for the creation of artificial enzyme diversity. It is possible to
introduce mutations into genes either during their replication in a PCR reaction, so-
called error-prone PCR [28], or by traditional methods like UV mutagenesis.
The introduction of new mutations in a natural gene can be performed either over
the whole length of that gene (in smaller, defined segments) [29] or at multiple or
single sites of a target gene [30]. It is also possible to recombine the genetic
information of existing genes into diverse new variants by shuffling the genes; this
gene shuffling process is performed either in vitro [31, 32] or in vivo [33].
Through the generation of artificial sequence space, variations of well-known
enzymes become available. These variations have the advantage that the basic
function of the original enzyme is maintained while various properties can be
changed and screened for as defined in the specifications of the enzyme. The
assessment of the artificial diversity is certainly complementary to the identification
of completely new enzymes and can be used to fine tune the properties of a new
discovered enzyme for the envisaged application. The methods for modification of
existing enzymes are combined under protein engineering approaches and are
discussed in Chapters 4 and 5.
3.2.2
Screening Strategies
The discovery of enzymes starts in almost all cases with the setup of a screening
strategy for the target enzyme. It depends very much on the functionality and the
desired properties as to how the screening process has to be organized. Two main
approaches can be distinguished: sequence homology-based screening and activity-
based screening. Sequence homology-based screening relies on existing knowledge
about sequence patterns of enzymes with a given functionality. A sequence
homology-based screening allows identification of variants of enzymes with known
74
functionality, but with different properties than the already known enzymes. It can
be performed as a gene mining experiment, which is basically an in silico screening
of sequence databases, followed by gene synthesis, cloning, enzyme expression, and
production for wet-lab testing. It can also be performed as a PCR-based screening of
single or multiple genomes or metagenomes for enzyme variants, which are
identified based on common sequence features of known enzymes that allow
the unambiguous detection of new, similar genes in a sequence databases. Again,
the newly discovered enzymes need to be expressed and produced for wet-lab
testing and final determination of their properties.
An activity-based screening system uses the identification of enzymatic activity
under defined conditions for the discovery of the new enzyme. Careful design of the
screening system allows for the identification of enzymes that have a high chance of
exhibiting excellent performance under the desired conditions. An activity-based
screening has advantages if the screening is directed to a functionality that has not yet
been identified on an enzymatic level or if the process conditions are very different
from those of already known enzymes having the target function.
Notably, the two fundamental strategies are not competitive or mutually exclusive,
but are complementary. The full power of the two strategies can only be fully exploited
when they are applied in a smart combination.
3.3
Enzyme Discovery Techniques
3.3.1
Gene Mining
Gene mining relies on bioinformatic approaches that allow us to identify genes of

interest via systematic inspection of nucleotide and protein databases. In a very
short period, a gene mining experiment may yield a huge amount of sequences,
which serve as a pool for further selection of sequences. Since the large number of
sequences often does not allow for cloning, expression, and testing of all candidates
it is necessary to prioritize them for wet-lab experiments. Analysis of the sequences
on a DNA and protein sequence level will give the necessary information. For
instance, it usually allows us to evaluate whether a protein can be secreted or
whether a particular host is suitable for expression of the sequence in question. In
an ideal situation, the researcher would also be able to predict whether the enzyme
fits into the physical conditions of the envisaged technical application. To check
parameters such as stability, solvent tolerance, pH profile, or substrate specificity it
would be necessary to profile the newly discovered set of sequences against
sequences of known and functionally characterized molecules. Unfortunately, the
required information is often lacking or not accessible via a comprehensive
database search. Protein sequences can also be used to generate 3D models by
automated homology modeling [34]. These models can serve as a basis for
structure-based evaluation of sequences. Such structure-based evaluations may
3.3 Enzyme Discovery Techniques j75
reveal information about substrate specificity or even regio- or enantioselectivity of
the putative enzymes [35, 36].
In addition to the wet-lab approaches it is possible to assess biodiversity in silico
by inspecting the sequence information that is deposited in protein sequence
databases such as the UniProt database or DNA databases such as GenBank (http://
www.ncbi.nlm.nih.gov/Genbank) [37] of the American National Center for Bio-
technology Information (NCBI). The sequences that are entered into these data-
bases originate from all sequencing efforts that are ongoing. This includes the
sequences of earlier identified proteins, whole genome sequences, and the
sequences from metagenome (environmental) sequencing projects. The pool of
available sequences in databases is huge. Owing to the exponential growth of the
sequences in databases, which serve as the bases for gene-mining, the chances of
identifying a variant of an existing gene are very good. Todays sequence databases
already contain sequences of the more than 1000 complete genomes. A detailed
summary of published genomes and active genome sequencing projects is available
via the Genomes Online databases GOLD (http://genomesonline.org) [38].
In addition to the genomic information from complete genomes the metagenomes
of several habitats that are difficult to assess by traditional screening strategies such
the metagenome sequencing experiment of the Sargasso Sea [25], termite gut,
whale carcasses, and so on are partly available in the databases. The CAMERA
database gives a complete list of these explored special habitats (http://camera
.calit2.net) [39].
The amount of publicly available sequence information is huge. Release 175 of
GenBank (15 December 2009) contains more than 110 million entries, each of them
coding for one protein. Since 1982 the number of entries per release has doubled
every 18 months and the growth of the known sequence space is still in an exponential
phase. Even with respect to one enzyme functionality the number of known
sequences is often very high. For example, the ninth release of the peptidase database
MEROPS contains already more than 135 000 classified peptidase sequences (release
9, 15 December 2009).
As well as the speed in discovering suitable enzymes for a given problem, gene
mining has the advantage of being able to deliver sequences of enzymes without the
need for laborious wet-lab experiments and additional sequencing efforts. Since the
genes can be easily obtained by gene synthesis and be sequence-optimized for any
given expression host, gene mining offers a relatively fast route to deliver prototypes
to test the feasibility of the chosen concept in wet-lab experiments.
3.3.2
Sequence Homology-Based Screening
Sequence homology-based screening techniques exploit the knowledge derived from

(translated) DNA and protein sequences of known enzymes. It can be performed as
an in silico experiment, by screening existing databases. Common sequence motifs
can be used to screen the databases for enzymes of the same type (presumably having
a similar functionality) – a process called gene mining (see above).
76
Alternatively, sequence homology-based screening can be executed as a wet lab

experiment by using the available genetic information to search for new variants of
existing genes. Degenerate primers or hydridization probes, which can be derived on
the basis of the identified sequence motif(s), can serve for experimental searches for
similar genes from a genome or metagenome.
Sequence homology-based screening approaches depend on sequence informa-
tion about conserved amino acid motifs that are characteristic for the targeted protein
family. In general, these primary sequence motifs are used to design degenerate PCR
primers or oligonucleotides for direct hybridization with isolated or pooled clones of
metagenome libraries. Using these oligonucleotides, complete gene libraries can be
screened by PCR or Southern blot for clones carrying a target gene and, depending on
the grade of primer/probe degeneration, to identify a panoply of clones encoding
different related gene fragments. After the identification of hit clones, full-length
gene sequences are usually determined by sequencing. Depending on the quality of
the metagenome libraries as well as on the design of oligonucleotides used for
screening, a surprisingly broad diversity of not yet described enzymes is often
accessed, such as demonstrated for xylanases [40], polyketide synthases [41], and
several other biotechnologically relevant enzyme classes [42]. In contrast to activity-
based screening methods, the whole targeted sequence space is theoretically
accessed, whereas only parts of it may be detected in a functional screening because
of sub-threshold expression levels in the chosen heterologous expression host.
In some cases, this can lead to significant discrepancies in sequence diversity
recovered by sequence- and activity-based screening approaches, as illustrated for
the example of galactosidases. Even retrieved partial gene sequences (metagenome
sequence tags (MSTs)) may provide valuable input for the creation of hybrid genes [43]
and gene shuffling experiments [44, 45].
3.3.3
Expression of Active Enzymes for Activity-Based Screening
A prerequisite for any activity-based screen is the production of active enzymes to

enable their testing in activity assays. Therefore, the respective genes have to be
transcribed and the genetic information must be translated into a polypeptide, which
subsequently needs to fold into the native conformation and possibly has to be
modified posttranslationally. All these processes rely to different extents on cellular
factors of the surrogate host, which might not be suited for the expression of the
heterologous genes. The design of more general expression systems is thus the
challenge of future metagenomic activity-based screening concepts. In other cases,
the cellular factors might not be ideal but allow a basal expression of the gene of interest
(GOI). Depending on the sensitivity of the used assay system, strong overexpression of
the GOI may not be necessary to identify the clone carrying the respective gene and
suboptimal conditions for the expression of the GOI can be accepted at this stage of the
enzyme development process. Given the enormous diversity provided by genomic or
metagenomic libraries, even the limited number of genes amenable to expression
represents a large enough resource to identify many interesting enzymes.
Concerning transcription of the GOI, one approach is to rely on the acceptance of
the native regulatory elements of the heterologous gene by the RNA-polymerases of
the surrogate host. However, practical experience has shown that only a limited
number of the desired genes can be identified this way. A more sophisticated
approach is to introduce defined promoter elements in the cloning vector used to
set up the metagenome library. The transcription then depends on both the
orientation and the distance of the GOI with respect to this promoter. The recently
introduced concept of vectors equipped with two oppositely arranged promoters
flanking the metagenomic insert DNA could indeed increase the hit rate, ideally by a
factor of two [46]. A further step taken to ensure the transcription of the GOI is the
transposon-mediated random insertion of strong uni- or bidirectional promoters
(K. Liebeton et al., BRAIN, unpublished results, and Reference [47]). This approach
seems to be useful especially for libraries with large insert sizes since here the
distance between GOI and vector-based promoter might become too large to ensure a
read-through of the GOI. This matter of distance is usually addressed by the careful
selection of the ratio of the size of the GOI and the insert size of the metagenomic
DNA. The probability that transcription is terminated before a complete read-
through of the GOI is obtained is proportional to the distance between promoter
and GOI.
While there are means to ensure the transcription of GOIs in metagenomic libraries,
the process of translation of the corresponding mRNA is far less susceptible to
manipulation. The rate of translation initiation depends on several parameters like,
for example, the nature of the start codon, the sequence of the Shine–Dalgarno
sequence (SD), the distance between SD and start codon, or the sequence of the
downstream box, that is, the first 4–5 codons following the start codon, all of which are
determined by the sequence of the metagenomic DNA. Since this sequence informa-
tion is usually not at hand, the opportunities to improve translation are rather limited.
The probability that the GOI is by chance inserted at an adequate distance to a
vector-located strong RBS or fused in the correct reading frame to a strongly
expressed, vector-based open reading-frame is too small to rely on. Since the
definition of efficient translational regulatory elements differs clearly among bac-
terial species it seems that variation of the expression host is the most promising
answer to the issue of translation.
Escherichia coli was used in many activity-based metagenomic screening projects as
the expression host, especially because of the high transformation rates necessary to
set up libraries as large as possible in order to approximate even faintly the complete
coverage of the microbial diversity. However, the use of E. coli as the expression host is
also reasonable with respect to the issue of translation. The canonical model of
translation initiation, with the efficiency of translation initiation depending directly
on the complementarity of the 30 terminus of the 16S-RNA and the SD, does not apply
to E. coli [48]. Therefore, E. coli can translate mRNA having very diverse translational
signals, something that is not valid for other commonly used expression hosts such as
Bacillus subtilis [49]. A large discrepancy in hit rates was observed in these host
systems, for instance, when prospecting for galactosidases in an activity-based
screening campaign (F. Niehaus et al., BRAIN, unpublished data). Here, several
78
enriched metagenome libraries derived from cold-adapted habitats with a total insert
size of nearly 0.6 GB (corresponding to >170 complete bacterial genomes) was
screened for lactolytic activity using X-Gal containing agar media. A high copy shuttle
vector was used to propagate identical libraries in E. coli and B. subtilis host strains.
A total number of about 270 000 clones were screened in each host, yielding 52 hit
clones with activity against X-Gal in E. coli, while only eight galactosidases were found
in B. subtilis, indicating a lower expression capacity of the latter host. Data mining for
galactosidase genes in the two above-mentioned DNA data pools revealed the
presence of more than 250 galactosidase sequences in the equivalent of 500 complete
bacterial genomes. Experimental hit rates in E. coli correlated relatively well with this
finding, while hit rates in B. subtilis were about one magnitude lower than the
calculated values. These results illustrate that, depending on the selected expression
host, a more or less restricted sub-set of genes can be harvested from the
metagenome.
The use of different non-E. coli expression hosts was established early in screening
approaches for the identification of new small molecule drugs produced by large
synthesis clusters [50, 51]. Here, the E. coli system seems to suffer from several
limitations such as, for example, the supply of precursors, although genes encoding
enzymes for the production of bioactive compounds were also identified using E. coli
as expression host [50]. The metagenomic libraries can be set up directly in a non-
E. coli surrogate host, such as for example Streptomyces lividans [51]. Alternatively,
Martinez et al. [52] constructed an environmental cosmid-library first in E. coli and
then shuttled the library by conjugation to S. lividans and Pseudomonas putida. The
establishment of vector systems for the integration of host specific attachment sites
and gene regulatory elements into the vector part of metagenomic libraries assem-
bled in E. coli by recombination allows one to transfer the same library to even a
broader range of expression hosts, that is, Pseudomonas, Bacillus, Streptomyces, and
Mycobacterium (G. Meurer et al., BRAIN, unpublished data).
3.3.4
Activity-Based Screening
In contrast to the sequence-based screening of gene libraries, activity-based

screening does not require any structural and sequence knowledge, but relies
solely on the activity of the enzymes sought after. However, in contrast to the
sequence homology-based approach, which enables the identification of novel
variants of already familiar protein families or known functional classes of
proteins [53], this approach affords the identification of completely new functional
entities for which the enzyme activity under investigation was not described before.
This is especially interesting with respect to enzyme promiscuity, that is, the fact
that enzymes might not be as specific with respect to substrate and even reaction
catalyzed as taught in introductory courses in biochemistry for many years. Since
enzyme catalytic promiscuity, where enzymes catalyze accidentally or induce new
reactions [54], has been recognized as being prevalent in several enzyme classes,
the activity-based screening approach might allow us to also identify biocatalysts
currently not expected to catalyze the reaction under investigation. Beloqui and
coworkers [55] identified by an activity-based screening approach of a bovine rumen
metagenomic library a new family of multicopper oxidases that is structurally
different from known multicopper oxidases. Since these enzymes were previously
annotated as conserved hypothetical proteins, a sequence homology-based screen-
ing approach would not have led to their identification.
When looking for an enzyme as a biocatalyst in an industrial application, it is
important to decide as soon as possible on the suitability of the enzyme pertaining to
process-relevant characteristics, such as thermal, pH and detergent stability, specific
activity, substrate specificity, or enantioselectivity, before starting into a lengthy and,
therefore, costly enzyme development process [56].
Activity-based screening of pure cultures of microorganisms represents the
traditional screening approach that has been used for enzyme discovery for more
than a century. A conventional enzyme discovery process relies on the growth of
microorganisms as the fundamental condition for enzyme discovery. The main
limitation of this dependency is the fact that most organisms do not grow under
conditions that have been established approximately a century ago in the laboratory
and have not been changed remarkably since then. The traditional activity-based
screening is limited to microorganisms as an enzyme source, since these are the only
organisms that can be cultivated in a way that allows for unambiguous detection of a
specific enzyme activity in an activity-based screening system. There have been,
however, successful attempts to increase the number of cultivable organisms by a
high-throughput variation of the cultivation conditions [57]. Nevertheless, most
living matter is not accessible as a resource for new enzymes when using the
traditional strategies. A powerful extension of the conventional activity-based screen-
ing of isolated microbial species is the expression of genes in a surrogate host to
generate material for an activity-based screening. For this purpose gene expression
banks have to be generated, which use either the genome of a single organism or a
metagenome as resource. This overcomes the limitations of the traditional
approaches by using a standardized well-controlled protein production system while
allowing for a very specific search for new enzyme functionality.
The parameters that define the properties of the new enzyme are very diverse and
can in many cases not be tested in a one-step screening. It is therefore often necessary
to organize the screening in multiple subsequent campaigns. Multilayered screening
campaigns will be set up like a funnel with multiple subsequent screens. Starting
with the screen with the highest throughput, but the least discriminating power, the
campaign will end with a low-throughput screen that allows a very good character-
ization of the enzymes, which often includes very sophisticated analytical methods.
The initial screening round should allow the highest possible throughput. This
primary screening aims to identify the desired catalytic functionality, ideally includ-
ing already one basic process parameter such as pH or temperature. The initial
screening generally reveals a candidate set that needs further characterization, most
often with a much lower throughput. In this stage much more sophisticated assays
need to be performed to collect data about, for example, stereo- or regioselectivity of
the enzyme towards the target substrate. Eventually, the most promising enzyme
80
candidates will be expressed in a production host and characterized in detail to fine

tune the selection of the ideal enzyme for that particular process, including the
definition of the most economic process parameters.
The success of an activity-based screening campaign depends largely on the
elegant combination of screening principles and methods for the detection of hit
candidates.
An activity-bearing clone can be detected in various ways. The easiest detection
uses the selection for growth of the surrogate hosts. The most widespread detection
of a new activity depends on the reaction of artificial substrates either on agar plates
or in a liquid assay system in microtiter plates or other liquid systems in different
volumes down to even the pico-liter scale. The reaction can be monitored by
absorption or fluorescent spectroscopy or in some cases simply by visual
inspection.
Screening systems that allow for the highest throughput rely on growth selection of
potential hit clones. In these systems, the growth of the screening host depends on
the presence and active expression of the targeted genes. The function of the target
enzyme complements the proteome of the screening host in such a way that the
organism only grows in the presence of that enzyme. Since the only detection is
growth of the host, the throughput of a selective screen is virtually unlimited. It finds
its main application in screenings for rare enzymes. A selective screening is limited
by the basic growth requirement of the host strain. It is, for instance, mostly
impossible to screen at extreme pH or temperatures. If these conditions, however,
constitute a process requirement, a secondary screening using an in vitro assay
system is required. However, the discovery of a new function even exhibiting the
wrong physicochemical properties may be very useful. The newly discovered hits of
the primary screening can be tested in an in vitro secondary screening under the right
physical conditions for the process. In any case, the sequence information of that new
enzyme can serve as a useful input for a set up of a subsequent sequence homology-
based screening.
Another high-throughput screening exploits the speed and selectivity of a cell
sorter. The so-called fluorescence accelerated cell sorting (FACS) screenings have a
high throughput, but require a fluorescent reporter substrate and the ability to
separate single cells and the substrate by advanced emulsification technology in
separate compartments to allow their sorting. A FACS machine does not just collect
cells with the desired activity but rather enriches them [58, 59]. As a consequence
several sorting cycles need to be performed to have a highly enriched population of
cells with the activity of interest. This population can be subjected to a traditional
activity-based assay system to isolate single clones, which contain the targeted
activity.
Most expression screenings use an artificial substrate for the identification of
clones that produce the desired enzyme. Activity on artificial substrates is a good
choice for a fast primary screen, but in most cases the activity on an artificial substrate
is not predictive for the performance of the enzyme in the real application. The hits
from the primary screening must therefore usually be subject of a more stringent and
laborious secondary screening.
3.4 Challenges in Enzyme Screening j81
3.4
Challenges in Enzyme Screening
The discovery of enzymes for multistep bioconversions such as the production of

1,3-propanediol from glucose [60] or methionine [61] is one of the major challenges in
enzyme screening. Multi-step bioconversions can be performed as biosynthetic
reaction processes in one or more subsequent reactors or as a fermentation process,
where a designer bug carries out all necessary reactions of a multistep synthesis
without the need for isolation of intermediate products. Fermentative multistep
bioconversions require many reaction steps with often instable intermediates [62]
that cannot be handled as isolated substrates in an activity-based screening assay.
Synthetic biology tools do not in all cases allow for an exact design of the synthesis
route, because there remains uncertainty about the exact substrate specificity of some
key enzymes in the designed pathway [63]. Another aspect that has to be accounted
for is the cofactor requirement of a fermentative bioconversion. It might be necessary
to incorporate a cofactor regeneration system or in rare cases to add a new synthesis
route for a less common cofactor system. A screening for a multistep biosynthesis
route has to able to tolerate the possible toxicity of substrates, intermediates, and
product towards the screening host [64]. The material flux through the synthesis
route needs to be optimized to avoid the accumulation of potential toxic compounds
during fermentation [65].
A good screening system for a multistep synthetic process copes with all the
challenges outlined above. The enzyme discovery process can be completely sepa-
rated into single-step screenings, if substrates and intermediates are available and
suitable for an assay. In a final step the complete enzyme set for the synthesis cascade
has to be assembled in the fermentation host. It will be difficult in a stepwise screen to
optimize the material flux already in the enzyme screening phase. It does thus carry
the risk that the enzymes are not compatible in their final assembly. The risk can be
diminished by using the future host as the screening host, for example, in a
complementary screening with growth selection [66].
The key criterion for the suitability of an enzyme for technical applications is the
ability to produce it in large enough quantities at a price that makes the usage of a
biocatalytic process competitive with an existing process or that makes the process
economically feasible at all.
It is tempting to perform the screening in the future production host. One of the
selection criteria for a new enzyme can be defined as being its producibility in
the production host. A production organism has a long history of being an effective
protein producer in specialized production equipment and fermentation conditions.
These requirements for a good production strain, however, might make it unsuitable
as a screening host. A screening host, for instance, needs to have high transformation
efficiency and needs to grow under conditions, often in microtiter plates, that are far
from ideal for the highly specialized and demanding production organisms. Fur-
thermore, it cannot be expected that the production of enzymes in a batch system – as
usually used for screening purposes – is predictive for the production of the same
enzyme under fermentative fed-batch conditions that are usually applied in a
82
commercial production plant. Consequently, the most important criterion for the
selection of an enzyme, namely, the economic criterion, remains uncertain until the
enzyme is finally produced in the production host to gain insight into its overall
suitability for the target process.
As for any project in enzyme discovery or optimization, the speed and predictability
of a project is a crucial factor for its success. In general, all steps until the enzyme
characterization and expression studies for production are fast and can be performed as
a streamlined, easy to plan process. Basics of the enzyme discovery process are the
establishment of a screening assay system on the one hand and the development of a
suitable screening host on the other hand. While these initial steps are executed it is
possible to select resources for the future screening. These are either the sequences
from databases or suitable metagenomic DNA-sources. As soon as the basics have been
established the hot phase of the screening can begin after three to four months.
Depending on the complexity and throughput of the assay system, between a hundred-
thousand and a few million clones are screened in a three to six months period in the
primary screen and initially characterized in a secondary screening system. The
resulting hit clones enter the final phase of the screening. It can take between six
and twelve additional months until the final candidate has been fully characterized and
is produced in quantities that allow for the commercial use of the process.
The above short outline makes it clear that the limiting factor of the enzyme
discovery process lies in the phase where enzyme characterization and expression
studies take place. To shorten the enzyme discovery process even further, tools are
needed that allow the prediction of enzyme properties as soon as the sequences of hit
candidates (or the results a database search) become available. These tools have not
been developed yet, but the first attempts show that some properties of enzymes can
already be estimated by analyzing protein sequences [67].
Another challenge of enzyme discovery is the economic production of the newly
discovered enzyme. At present, there are hardly any knowledge-based criteria for the
selection of a suitable production host. It is a well-educated guess of the scientists that
decides which of the available hosts are most suited to produce the ideal hit candidate
of a screening campaign. As a consequence many companies have a limited set of
very diverse production hosts and molecular biology constructs available as a plug-in
system that allows rapid identification of the right production system for the enzyme
of interest. The availability of one universal production strain for all enzymes would
of course be the ideal solution to this problem, but remains a dream until the
extremely complex expression and secretion machinery of microorganisms is fully
understood and becomes a tool with predictable behavior in the hands of the
scientists.
3.5
Concluding Remarks
The discovery of enzymes with new properties or new enzyme functionalities is

developing into a knowledge-based science. While the discovery of enzymes was
References j83
driven in the past by the pure chance of finding an enzyme based on its activity on a
given substrate, enzyme discovery today is using the knowledge and tool box modern
biotechnology has to offer. While classical screening processes required the culti-
vation of organisms followed by testing of cell material or fermentation supernatant,
it is nowadays possible to perform the initial steps of a screening in a computer even
before laboratory work starts. Even the necessity to cultivate the enzyme-carrying
source organism no longer exists, because researchers can access virtually any DNA-
sequence through the use of metagenomes from various sources. This means that
the accessible diversity of enzyme sequences is virtually unlimited, suggesting that
almost tailor-made enzymes may be discovered for any defined process.
The art of discovering new enzymes has become the ability to manage the high
number of available enzymes to make the discovery process as efficient as possible.
Consequently, an enzyme discovery process has become an activity that can be well
planned and its result is predictable in terms of the ability to find the required activity
or to define a decision point in order to stop a screening process in the unlikely case
that it has become evident that the desired activity cannot be found in nature.
Enzyme discovery has at least two challenges left. First, the screening process
needs to be streamlined more and more to shorten it and make enzymes available
even more quickly for new processes. Second, bioinformatics tools need to
be developed that allow us to predict whether a new protein performs and can be
produced within the economic boundaries of a desired application.
References
1 Blake, C.C., Koenig, D.F., Mair, G.A., and implementation of the UniProt
North, A.C., Phillips, D.C., and website. BMC Bioinformatics, 10, 136.
Sarma, V.R. (1965) Structure of hen egg- 6 Hulo, N., Bairoch, A., Bulliard, V.,
white lysozyme. A three-dimensional Cerutti, L., Cuche, B., De Castro, E.,
Fourier synthesis at 2 Angstrom Lachaize, C., Langendijk-Genevaux, P.S.,
resolution. Nature, 206 (986), 757–761. and Sigrist, C.J.A. (2007) The 20 years of
2 International Union of Biochemistry and PROSITE. Nucleic Acids Res., 36,
Molecular Biology (1992) Enzyme D245–D249.
Nomenclature, Academic Press, 7 Finn, R.D., Tate, J., Mistry, J.,
San Diego. Coggill, P.C., Sammut, J.S., Hotz, H.R.,
3 Chang, A., Scheer, M., Grote, A., Ceric, G., Forslund, K., Eddy, S.R.,
Schomburg, I., and Schomburg, D. (2009) Sonnhammer, E.L., and Bateman, A.
BRENDA, AMENDA and FRENDA the (2008) The Pfam protein families
enzyme information system: new content database. Nucleic Acids Res., 36,
and tools in 2009. Nucleic Acids Res., 37, D281–D288.
D588–D592. 8 Murzin, A.G., Brenner, S.E., Hubbard, T.,
4 The UniProt Consortium (2009) The and Chothia, C. (1995) SCOP: a
universal protein resource (UniProt). structural classification of proteins
Nucleic Acids Res., 37, D169–D174. database for the investigation of
5 Jain, E., Bairoch, A., Duvaud, S., Phan, I., sequences and structures. J. Mol. Biol.,
Redaschi, N., Suzek, B.E., Martin, M.J., 247, 536–540.
McGarvey, P., and Gasteiger, E. (2009) 9 Rawlings, N.D., Morton, F.R., Kok, C.Y.,
Infrastructure for the life sciences: design Kong, J., and Barrett, A.J. (2008)
84
MEROPS: the peptidase database. Nucleic cDNA: a rapid method for isolation of MLL
Acids Res., 36, D320–D325. fusion transcripts involving unknown
10 Cantarel, B.L., Coutinho, P.M., partner genes. Proc. Natl. Acad. Sci.
Rancurel, C., Bernard, T., Lombard, V., U.S.A., 97, 9597–9602.
and Henrissat, B. (2009) The 21 Myrick, K.V. and Gelbart, W.M. (2002)
carbohydrate-active EnZymes database Universal Fast Walking for direct and
(CAZy): an expert resource for versatile determination of flanking
glycogenomics. Nucleic Acids Res., 37, sequence. Gene, 284, 125–131.
D233–D238. 22 Antal, Z., Rascle, C., Fevre, M., and
11 Levitt, M. (2009) Nature of the protein Bruel, C. (2004) Single oligonucleotide
universe. Proc. Natl. Acad. Sci. U.S.A., nested PCR: a rapid method for the
106 (27), 11079–11084. isolation of genes and their flanking
12 Amann, R.I., Ludwig, W., and regions from expressed sequence tags.
Schleifer, K.H. (1995) Phylogenetic Curr. Genet., 46, 240–246.
identification and in situ detection of 23 Meyer, Q.C., Burton, S.G., and
individual microbial cells without Cowan, D.A. (2007) Subtractive
cultivation. Microbiol. Rev., 59, 143–169. hybridization magnetic bead capture: a
13 Burton, S.G., Cowan, D.A., and new technique for the recovery of full-
Woodley, J.M. (2002) The search for the length ORFs from the metagenome.
ideal biocatalyst. Nat. Biotechnol., 20, J. Biotechnol., 2, 36–40.
37–45. 24 Uchiyama, T. and Watanabe, K. (2006)
14 Torsvik, V., Ovreas, L., and Thingstad, T.F. Improved inverse PCR scheme for
(2002) Prokaryotic diversity–magnitude, metagenome walking. BioTechniques, 41,
dynamics, and controlling factors. Science, 183–188.
296, 1064–1066. 25 Venter, J.C., Remington, K.,
15 Torsvik, V. and Ovreas, L. (2002) Microbial Heidelberg, J.F., Halpern, A.L., Rusch, D.,
diversity and function in soil: from genes Eisen, J.A., Wu, D., Paulsen, I.,
to ecosystems. Curr. Opin. Microbiol., 5, Nelson, K.E., Nelson, W., Fouts, D.E.,
240–245. Levy, S., Knap, A.H., Lomas, M.W.,
16 Handelsman, J., Rondon, M.R., Nealson, K., White, O., Peterson, J.,
Brady, S.F., Clardy, J., and Goodman, R.M. Hoffman, J., Parsons, R.,
(1998) Molecular biological access to the Baden-Tillson, H., Pfannkoch, C.,
chemistry of unknown soil microbes: a Rogers, Y.H., and Smith, H.O. (2004)
new frontier for natural products. Chem. Environmental genome shotgun
Biol., 5, R245–R249. sequencing of the Sargasso Sea. Science,
17 Lorenz, P. and Eck, J. (2005) 304, 66–74.
Metagenomics and industrial 26 Krause, L., Diaz, N.N., Bartels, D.,
applications. Nat. Rev., 3, 510–516. Edwards, R.A., Puhler, A., Rohwer, F.,
18 Liu, Y.G. and Whittier, R.F. (1995) Meyer, F., and Stoye, J. (2006) Finding
Thermal asymmetric interlaced PCR: novel genes in bacterial communities
automatable amplification and isolated from the environment.
sequencing of insert end fragments from Bioinformatics, 22, e281–e289.
P1 and YAC clones for chromosome 27 Noguchi, H., Park, J., and Takagi, T. (2006)
walking. Genomics, 25, 674–681. MetaGene: prokaryotic gene finding from
19 Ochman, H., Ayala, F.J., and Hartl, D.L. environmental genome shotgun
(1993) Use of polymerase chain reaction to sequences. Nucleic Acids Res., 34,
amplify segments outside boundaries of 5623–5630.
known sequences. Methods Enzymol., 218, 28 Labrou, N.E. (2010) Random mutagenesis
309–321. methods for in vitro directed enzyme
20 Megonigal, M.D., Rappaport, E.F., evolution. Curr. Protein. Pept. Sci., 11,
Wilson, R.B., Jones, D.H., Whitlock, J.A., 91–100.
Ortega, J.A., Slater, D.J., Nowell, P.C., and 29 Reetz, M.T., Wang, L.W., and Bocola, M.
Felix, C.A. (2000) Panhandle PCR for (2006) Directed evolution of
References j85
enantioselective enzymes: iterative cycles associated metadata. Nucleic Acids Res., 36,
of CASTing for probing protein-sequence D475–D479.
space. Angew. Chem. Int. Ed., 45, 39 Seshadri, R., Kravitz, S.A., Smarr, L.,
1236–1241. Gilna, P., and Frazier, M. (2007)
30 Wells, J.A., Vasser, M., and Powers, D.B. CAMERA: A community resource for
(1985) Cassette mutagenesis: an efficient metagenomics. PLoS Biol., 5 (3), e75. doi:
method for generation of multiple 10.1371/journal.pbio.0050075
mutations at defined sites. Gene, 34, 40 Radomski, C., Seow, K., Warren, R., and
315–323. Yap, W. (1988) Method for isolating
31 Stemmer, W.P. (1994) Rapid evolution of a xylanase gene sequences from soil DNA,
protein in vitro by DNA shuffling. Nature, compositions useful in such method and
370, 389–391. compositions obtained thereby, US patent
32 Crameri, A. and Stemmer, W.P. (1995) number 5,849,491.
Combinatorial multiple cassette 41 Seow, K.T., Meurer, G., Gerlitz, M.,
mutagenesis creates all the permutations Wendt-Pienkowski, E., Hutchinson, C.R.,
of mutant and wild-type sequences. and Davies, J. (1997) A study of iterative
Biotechniques, 18, 194–196. type II polyketide synthases, using
33 Gomez, A., Galic, T., Mariet, J.F., Matic, I., bacterial genes cloned from soil DNA: a
Radman, M., and Petit, M.A. (2005) means to access and use genes from
Creating new genes by plasmid uncultured microorganisms. J. Bacteriol.,
recombination in Escherichia coli and 179, 7360–7368.
Bacillus subtilis. Appl. Environ. Microbiol., 42 Lorenz, P., Liebeton, K., Niehaus, F., and
71, 7607–7609. Eck, J. (2002) Screening for novel enzymes
34 Pieper, U., Eswar, N., Webb, B.M., for biocatalytic processes: accessing the
Eramian, D., Kelly, L., Barkan, D.T., Carter, metagenome as a resource of novel
H., Mankoo, P., Karchin, R., Marti- functional sequence space. Curr. Opin.
Renom, M.A., Davis, F.P., and Sali, A. Biotechnol., 13, 572–577.
(2009) MODBASE, a database of 43 Okuta, A., Ohnishi, K., and Harayama, S.
annotated comparative protein structure (1998) PCR isolation of catechol 2,3-
models and associated resources. Nucleic dioxygenase gene fragments from
Acids Res., 37, D347–D354. environmental samples and their
35 Ortiz, A.R., Gomez-Puertas, P., assembly into functional genes. Gene, 212,
Leo-Macias, A., Lopez-Romero, P., 221–228.
Lopez-Vi~ nas, E., Morreale, A., Murcia, M., 44 Crameri, A., Raillard, S.A., Bermudez, E.,
and Wang, K. (2006) Computational and Stemmer, W.P. (1998) DNA shuffling
approaches to model ligand selectivity in of a family of genes from diverse species
drug design. Curr. Top. Med. Chem., 6, accelerates directed evolution. Nature,
41–55. 391, 288–291.
36 Juhl, P.B., Trodler, P., Tyagi, S., and 45 Gibbs, M.D., Nevalainen, K.M., and
Pleiss, J. (2009) Modelling substrate Bergquist, P.L. (2001) Degenerate
specificity and enantioselectivity for oligonucleotide gene shuffling (DOGS): a
lipases and esterases by substrate- method for enhancing the frequency of
imprinted docking. BMC Struct. Biol., recombination with family shuffling.
9, 39. Gene, 271, 13–20.
37 Benson, D.A., Karsch-Mizrachi, I., 46 L€ammle, K., Zipper, H., Breuer, M.,
Lipman, D.J., Ostell, J., and Wheeler, D.L. Hauer, B., Buta, C., Brunner, H., and
(2008) GenBank. Nucleic Acids Res., 36, Rupp, S. (2007) Identification of novel
D25–D30. enzymes with different hydrolytic
38 Liolios, K., Mavrommatis, K., activities by metagenome expression
Tavernarakis, N., and Kyrpides, N.C. cloning. J. Biotechnol., 127, 575–592.
(2008) The Genomes On Line Database 47 Leggewie, C., Henning, H.,
(GOLD) in 2007: status of genomic and Schmeisser, C., Streit, W.R., and
metagenomic projects and their Jaeger, K.E. (2006) A novel transposon for
86
functional expression of DNA libraries. Golyshin, P.N. (2006) Novel polyphenol

J. Biotechnol., 123, 281–287. oxidase mined from a metagenome
48 Boni, I.V. (2006) Diverse molecular expression library of bovine rumen:
mechanisms for translation initiation in biochemical properties, structural
prokaryotes. Mol. Biol. (Mosk), 40, analysis, and phylogenetic relationships.
658–668. J. Biol. Chem., 281, 22933–22942.
49 Vellanoweth, R.L. and Rabinowitz, J.C. 56 Gabor, E., Liebeton, K., Niehaus, F., Eck, J.,
(1992) The influence of ribosome- and Lorenz, P. (2007) Updating the
binding-site elements on translational metagenomics toolbox. J. Biotechnol., 2,
efficiency in Bacillus subtilis and 201–206.
Escherichia coli in vivo. Mol. Microbiol., 6, 57 Zengler, K., Toledo, G., Rappe, M.,
1105–1114. Elkins, J., Mathur, E.J., Short, J.M., and
50 Courtois, S., Cappellano, C.M., Ball, M., Keller, M. (2002) Nonlinear partial
Francou, F.X., Normand, P., Helynck, G., differential equations and applications:
Martinez, A., Kolvek, S.J., Hopke, J., cultivating the uncultured. Proc. Natl.
Osburne, M.S., August, P.R., Nalin, R., Acad. Sci. U.S.A., 99, 15681–15686.
Guerineau, M., Jeannin, P., Simonet, P., 58 Yang, G. and Withers, S.G. (2009)
and Pernodet, J.L. (2003) Recombinant Ultrahigh-throughput FACS-based
environmental libraries provide access to screening for directed enzyme evolution.
microbial diversity for drug discovery ChemBioChem., 10, 2704–2715.
from natural products. Appl. Environ. 59 Bershtein, S. and Tawfik, D.S. (2008)
Microbiol., 69, 49–55. Advances in laboratory evolution of
51 Wang, G.Y., Graziani, E., Waters, B., enzymes. Curr. Opin. Chem. Biol., 12,
Pan, W., Li, X., McDermott, J., Meurer, G., 151–158.
Saxena, G., Andersen, R.J., and Davies, J. 60 Nakamura, C.E. and Whited, G.M. (2003)
(2000) Novel natural products from soil Metabolic engineering for the microbial
DNA libraries in a streptomycete host. production of 1,3-propanediol. Curr. Opin.
Org. Lett., 2, 2401–2404. Biotechnol., 14 (5), 454–459.
52 Martinez, A., Kolvek, S.J., Yip, C.L., 61 Kr€omer, J.O., Wittmann, C., Schr€oder, H.,
Hopke, J., Brown, K.A., MacNeil, I.A., and and Heinzle, E. (2006) Metabolic pathway
Osburne, M.S. (2004) Genetically analysis for rational design of L-
modified bacterial strains and novel methionine production by Escherichia coli
bacterial artificial chromosome shuttle and Corynebacterium glutamicum. Metab.
vectors for constructing environmental Eng., 8, 353–369.
libraries and detecting heterologous 62 Kirby, J. and Keasling, J.D. (2009)
natural products in multiple expression Biosynthesis of plant isoprenoids:
hosts. Appl. Environ. Microbiol., 70, perspectives for microbial engineering.
2452–2463. Annu. Rev. Plant. Biol., 60, 335–355.
53 Lorenz, P. and Schleper, C. (2002) 63 Yus, E., Maier, T., Michalodimitrakis, K.,
Metagenome – a challenging source of van Noort, V., Yamada, T., Chen, W.H.,
enzyme discovery. J. Mol. Catal., B Wodke, J.A., G€ uell, M., Martınez, S.,
Enzym., 19–20, 13–19. Bourgeois, R., K€ uhner, S., Raineri, E.,
54 Hult, K. and Berglund, P. (2007) Enzyme Letunic, I., Kalinina, O.V., Rode, M.,
promiscuity: mechanism and Herrmann, R., Gutierrez-Gallego, R.,
applications. Trends Biotechnol., 25, Russell, R.B., Gavin, A.C., Bork, P., and
231–238. Serrano, L. (2009) Impact of genome
55 Beloqui, A., Pita, M., Polaina, J., reduction on bacterial metabolism and its
Martinez-Arias, A., Golyshina, O.V., regulation. Science, 326 1263–1268.
Zumarraga, M., Yakimov, M.M., 64 Zhu, M.M., Lawman, P.D., and
Garcia-Arellano, H., Alcalde, M., Cameron, D.C. (2002) Improving 1,3-
Fernandez, V.M., Elborough, K., propanediol production from glycerol in a
Andreu, J.M., Ballesteros, A., Plou, F.J., metabolically engineered Escherichia
Timmis, K.N., Ferrer, M., and coli by reducing accumulation of
References j87
sn-glycerol-3-phosphate. Biotechnol. Prog., for chemical synthesis. Proc. Natl. Acad.
18, 694–699. Sci. U.S.A., 103, 1693–1698.
65 Dueber, J.E., Wu, G.C., 67 Kuipers, R.K., Joosten, H.J.,
Malmirchegini, G.R., Moon, T.S., Verwiel, E., Paans, S., Akerboom, J.,
Petzold, C.J., Ullal, A.V., Prather, K.L., and van der Oost, J., Leferink, N.G.,
Keasling, J.D. (2009) Synthetic protein van Berkel, W.J., Vriend, G., and
scaffolds provide modular control over Schaap, P.J. (2009) Correlated
metabolic flux. Nat. Biotechnol., 8, mutation analyses on super-family
753–759. alignments reveal functionally
66 van Sint Fiet, S., van Beilen, J.B., and important residues. Proteins, 76,
Witholt, B. (2006) Selection of biocatalysts 608–616.
j89
4
Rational Design of Enzymes
J€
urgen Pleiss
4.1
Enzyme Design: Learn from Nature
Enzymes are amazing devices: they have a complex structure, multi-timescale

dynamics, and exquisite biochemical properties, even though they are coded simply
by a linear chain of only a few hundred characters. They catalyze a broad range of
chemical reactions under mild conditions in aqueous solution and at room temper-
ature, but nature also provides us with enzymes that function in many non-aqueous
solvents [1] or at optimal temperatures exceeding 100 C [2]. They can accelerate
chemical reactions up to 1017-fold [3] and often show high regio-, stereo-, or
chemoselectivity. Without understanding the reason for their amazing properties,
enzymes have been used since the early days of mankind to prepare food, feed, and
useful materials, and a considerable toolbox of enzymes has become available.
However, the enzymes that have been characterized to date rarely have the combined
properties necessary for industrial chemical production such as high activity, high
selectivity, broad substrate specificity towards non-natural substrates, no inhibition
by substrate or product, and a high stability in organic solvents and at high substrate
or product concentrations [4]. The strategy of searching for appropriate enzymes in
nature is still promising, because the number of enzymes still to be discovered
outnumbers the known enzymes by far. Although the techniques to screen for new
enzymes and to characterize their properties have improved considerably since the
Bronze Age, the principle is still the same: go out and find the desired tool in nature.
Versatile methods of screening for new enzymes are presented in Chapter 3.
In a complementary approach, the biochemical properties of enzymes have been
intensively studied since the first report on an enzyme and its function was published
in 1833 [5]. Models to describe their function were developed to learn from nature and
ultimately to become able to design biocatalysts, reactions, and processes. Basic
models to explain recognition between enzymes and substrates or other ligands and
thus to explain specificity and selectivity were the lock and key model [6] and
the induced fit model [7]. To explain their catalytic proficiency, enzyme kinetics
were interpreted in the framework of transition state theory [8] assuming that the
and Oliver May.
j 4 Rational Design of Enzymes
90
rate-limiting step of a chemical reaction can be assigned to the passage through a high-
energy transition state. According to this model, the catalytic activity of enzymes is a
result of their highly specific binding of the substrate in its transition state as compared
to its ground state, thus lowering the energy difference between transition state and
ground state [9]. The most proficient enzyme, orotidine 50 -phosphate decarboxylase,
shows a rate enhancement of 1017. This rate enhancement is reflected by the amazing
specificity of the enzyme: the transition state is bound with an estimated dissociation
constant of 5 1024 M, while the substrate in its ground state is bound with a
dissociation constant in the order of 6 107 M as indicated by its Km [10].
Since the first high-resolution structure of proteins became available 50 years
ago [11], the structural basis of enzyme function has been thoroughly investigated
and molecular models at atomic resolution have been proposed to explain the
function of enzymes. By combining X-ray analysis, molecular dynamics simulations,
and quantum chemical calculations, previous ideas on the role of enzymes in binding
the transition state were substantiated and the stabilization of the transition state
could be quantitatively evaluated from first principles [12], thus providing a model of
enzyme function on the atomic and even electronic level. As an alternative approach,
data mining makes use of the large amount of data on sequence, structure, and
function of enzymes, and seeks to establish quantitative relationships between the
properties of a series of substrates or a library of enzymes and the activity, specificity,
or selectivity of the catalyzed reaction [13, 14].
In both approaches computational methods of modeling enzymes and analyzing
data are crucial; and both approaches are complementary, because each of them
allows us to establish hypotheses on the function of enzymes that can then be used by
the other method. In this chapter we discuss what we can learn from data mining and
molecular modeling of enzymes, how these methods can help us in enzyme design,
what their limitations are, and finally visions for the coming years.
4.2
Today: Find and Improve Enzymes
4.2.1
Data Mining: Find Appropriate Biocatalysts in Databases
Since the mid-1990s, the number of published gene sequences has been rapidly
increasing due to genome and metagenome sequencing projects. Currently, nearly
800 bacterial genomes are already sequenced, and 2400 bacterial genome projects are
ongoing [15]. Next-generation DNA sequencing techniques are expected to further
boost sequence throughput [16]. Especially, enzymes from extremophilic organisms
are interesting for industrial bioconversions [17]. Metagenomics is especially prom-
ising approach to the discovery of new biocatalysts [18]; however, it is challenging to
handle the complexity of metagenomics data [19]. While the sheer production of
sequence data is no longer a bottleneck for biocatalyst discovery, it is more and more
of a challenge to transform this data into knowledge.
4.2 Today: Find and Improve Enzymes j91
Identification of a member of a desired enzyme family in a DNA sequence is
usually a straightforward task, and preferably the genome sequence of a thermophilic
organism or of an organism which is already known to convert a given substrate is
selected. Sequence similarity searches can be performed by pairwise sequence
alignment, the basic local alignment search tool (BLAST) being most frequently
used for searches in large sequence databases [20]. For more distant similarities,
sequence profile methods are used [21, 22]. Both approaches, however, assume that
sequence similarity implies functional similarity. While this assumption is true in
many cases, matching sequences do not always infer similar functions and some-
times proteins with dissimilar sequences have similar functions. In addition,
individual biochemical properties such as stability, activity, specificity, or selectivity
can still not be deduced from the gene sequence alone. However, in many cases a
systematic comparison of sequences and known biochemical functions of the
members of a protein family allows us to derive rules and create fruitful hypotheses,
to focus the search, and to select the most promising candidates with a much higher
chance of success than obtained by random picking [23].
Simple rules might be patterns, multisequence alignments, or profiles that code
for a biochemical function. A collection of patterns and signatures are provided by
databases such as PROSITE [24] or PRINTS [25], of multisequence alignments by
BLOCKS [26] or ProDom [27], and of profiles by Pfam [28]. New consensus patterns
have be derived by analyzing sequence alignments of small protein families that code
for interesting biochemical properties. Using such property-specific sequence pat-
terns, lipases that have activity toward esters of tertiary alcohols could be distin-
guished from lipases that do not have this specificity [29], or laccases can be
distinguished from other multicopper oxidases [30].
While sequence information can provide information on catalytic function and, in
rare cases, on substrate specificity, the recognition of a substrate by an enzyme is
determined by the chemical structure of the substrate itself and by the shape of the
substrate binding site. Comprehensive analysis of the substrate structure is the basis
of QSAR methods that derive three-dimensional quantitative structure–activity
relationships [31]. These methods are widely applied in computer-aided drug design
to predict inhibitors [32]. QSAR approaches have also been extensively applied to
predict substrates of cytochrome P450 monooxygenases [33]. Various pharmaco-
phore models have been constructed by superimposing structures of substrates and
non-substrates to extract functional groups, so-called descriptors, which correlate
with specificity or selectivity of cytochrome P450 monooxygenases, or identify the
most efficient biocatalyst [31]. A more detailed view of the interaction between
enzyme and substrate has been obtained by combining QSAR analyses with
information on the structure of the binding site. Thus, substrate specificity as well
as regio- and stereoselectivity of the catalyzed reaction have been successfully
predicted [34].
To predict the binding affinity of lead structures to target proteins of known
structure, molecular docking is a widely used method in medicinal chemistry,
because a geometrical complementarity between protein and ligand correlates well
with experimentally observed binding affinity. However, it is not straightforward
92
to extend this concept to predict the catalytic activity of an enzyme toward a

prospective substrate, because the enzyme should specifically bind the transition
state of the substrate, but not the ground states of substrate or product. Therefore, to
obtain relevant results the prospective substrate has to be docked in its transition
state [35, 36]. Because the rate-limiting chemical step requires an optimal geometry
and small differences in structure might lead to huge effects in catalytic activity,
modeling of the flexibility of the enzyme is a major issue [37]. Recently, molecular
docking has been systematically applied to identify potential substrates of an enzyme
by performing a virtual screening of metabolite libraries [38] and to generate
hypotheses about sequence–structure–function relationships of large enzyme fam-
ilies [39]. Thus, proteins with previously unknown function were assigned to the
family of uronate isomerases or mandelate racemases. A subsequent biochemical
characterization confirmed this prediction. By analyzing the structure of the sub-
strate binding site, amino acids that mediate substrate specificity or enantioselectivity
were predicted.
For a more systematic understanding of enzyme-catalyzed reactions, a database
of enzyme reaction mechanisms (MACiE) was developed by the group of
Janet Thornton [40] and systematically analyzed. It turned out that most enzyme
reactions rely upon nucleophilic and general acid–base chemistry, while electrophilic
reactions are very rare. Most catalytic amino acid residues are histidine, cysteine, and
aspartic acid, and most amino acid residues perform stabilization roles or proton
shuttling roles. The most common metal ions are magnesium and zinc, which are
used in enzymes to stabilize negative charges and to activate substrates, and iron,
manganese, cobalt, molybdenum, copper, and nickel to act as Lewis acids or as redox
centers. In addition, redox-active metal ions are often associated with organic
cofactors [41]. Thus, the known enzymes show a surprisingly small spectrum of
active components.
To learn more about the evolution of enzyme families and to systematically study
the relationship between sequence, structure, specificity, and selectivity, specialized
databases have been developed for several enzyme families such as peptidases [42],
a/b hydrolases [43, 44], cytochrome P450 monooxygenases [45–48], and vitamin B6-
dependent enzymes [49]. Data warehouse systems have been developed and applied
to construct and analyze new enzyme families: the DWARF system developed in our
group [50] and the 3DM system by Henk-Jan Joosten [51].
Enzymes are thought to have optimized their astonishing catalytic power and
specificity by evolving their protein surface to complement substrate transition states.
However, the assumption that the shape of substrate binding pockets is related to the
shape of the ligand or substrate molecules is only partially true. A systematic
comparison of binding pockets of proteins that bind the same ligand demonstrated
that the binding pockets are more variable in their shapes than can be accounted for
by the conformational variability of the ligand. This observation indicates that shape
complementarity might not be the only driving force of molecular recognition [52].
On average, binding pockets of proteins are three times larger than their bound
ligand; the space between ligand and ligand-interacting protein atoms is occupied by
water molecules that contribute to affinity and specificity of binding.
4.2.2
Rational Evolution: Improve Efficiency of Directed Evolution
Directed evolution has proven to be an effective method to improve the properties of

enzymes (see Chapter 5). Assuming that the effects of single mutations are additive,
consecutive rounds of creation and screening of large random mutant libraries are
performed. However, this assumption is not always correct and the quality of the
library as well as the need for an appropriate high-throughput screening assay can
limit the applicability and efficiency of many directed evolution strategies. By
screening for activity, directed evolution experiments frequently lead to mutants
with improved expression level or solubility, because these properties are mediated by
a potentially large number of single, additive mutations.
More recently, two approaches have been suggested to improve the efficiency of the
directed evolution by enriching the library and reducing the library size substantially,
taking into account further information. The first approach seeks to improve the
efficiency of searching for mutants with improved specificity or selectivity. In this
approach structure information is applied to identify residues that are involved in
substrate binding. By focusing the library to a few residues in the substrate binding
site, lipase mutants were generated with substantially increased enantioselectiv-
ity [53], and the substrate specificity of a bacterial P450 monooxygenase was shifted
toward short-chain fatty acids in only a few iterative rounds [54]. The major advantage
of highly enriched and focused libraries is the simultaneous mutation of a few
hotspot residues. The more insight we gain into the molecular basis of enzyme
function, the smaller the number of hotspot residues will become and the more
efficiently the resulting library can be analyzed. A systematic analysis of P450
monooxygenases identified only two residues that were predicted to contribute to
selectivity in all enzymes of this family [55]. On this basis, a minimal, highly enriched
library consisting of only 25 mutants was created, which provided mutants with
improved selectivity toward a broad range of substrates [56]. Thus, the strategy of
creating focused, highly enriched libraries based on molecular modeling of sub-
strate–enzyme interactions seems to be most efficient for engineering of enzymatic
properties such as substrate specificity and selectivity, which are mediated by the
concerted, non-additive action of a small number of spatially close amino acids.
The second approach takes advantage of improved sequencing capabilities. In
classical directed evolution experiments only the best mutants in each round are
selected and sequenced; therefore sequence information on unfavorable mutants is
missing. By establishing a quantitative sequence–function relationship based on a
comprehensive statistical analysis of a large number of favorable and less favorable
mutants, hotspot regions that are prone to be beneficial were identified, and focused
libraries could be designed [57–60]. This strategy has been successfully applied to
improve specific activity, stability, expression level, or solubility, which usually are the
result of many (on the order of 20–40) additive single mutations.
Both approaches tightly integrate computational modeling and experiment. While
model-derived hypotheses aid in designing more focused libraries, the information
from library screening on successful and less successful variants teaches us about the
94
contribution of individual residues to the function of the biocatalyst. Thus, the two
methods are complementary and contribute to a deeper understanding of the
sequence–structure–function relationships of the enzyme under investigation.
4.2.3
Molecular Modeling and Protein Design of Stability, Specificity, and Selectivity
When the term protein engineering was coined [61], three methods were consid-
ered to be crucial: (chemical) synthesis of DNA, experimental determination of
protein structures, and computer modeling of structure, folding, and function. The
goal of protein engineering was declared as to control and improve in a predictable
fashion the biochemical properties of enzymes such as the kinetic properties of
enzymes, thermostability, temperature optimum, stability in organic solvents, and
substrate specificity. It was expected that combining structure determination and
modeling would directly result in the prediction of mutants with improved prop-
erties. The predictions would then be validated by recombinant expression of the
variant and by its biochemical and structural characterization. Subsequently, the
engineered gene product would be subjected to the next round of modeling, mutant
design, and characterization. This iterative approach via multiple protein engineer-
ing cycles should lead to a stepwise improvement of the design by using knowledge-
based procedures that exploit facts, rules, and observations about proteins of known
three-dimensional structure [62].
The principle of protein engineering is still valid, though many new experimental
techniques have been developed, new enzyme families and reactions have been
identified, and much insight into the molecular basis of enzyme function has been
gained since 1983. Protein engineering is still considered as the art of optimizing the
properties of a sub-optimal enzyme by the iterative application of mutations, aided by
computational methods such as sequence alignment, structure analysis, docking, or
molecular dynamics simulations. However, establishing a quantitative model of how
enzymes function is more than a mere enabling technique that allows us to re-
engineer proteins more efficiently. A model that allows us to derive measurable
quantities from first principles would be the crucial organizing principle of an
overwhelming amount of experimental data on enzymes such as sequences, struc-
tures, reactions, reaction mechanisms, specificities, selectivities, and the effect of
environment (solvent, temperature, pH).
4.2.3.1 Prediction of Enzyme Structure

Because DNA sequencing has become fast, cheap, and reliable, sequence informa-
tion on the enzyme of interest is readily available. In contrast, obtaining information
on posttranslational modifications or the three dimensional structure of an enzyme is
more challenging. Therefore, for most enzymes or enzyme variants an experimen-
tally determined structure is not yet available. Instead, the structure of the enzyme
has to be modeled based on its sequence. As a benchmark of structure prediction the
biennial competition Critical Assessment of Structure Prediction (CASP) has been
established. In an unbiased, objective procedure the progress of structure prediction
methods is measured and criteria on how to measure quality of predictions are
discussed [63]. Since the start of CASP1 in 1994 until the recently finished CASP8
competition significant progress has been seen in the two major prediction catego-
ries: template-based modeling [64] and template-free modeling [65], although it is still
disputed how the quality of a structure prediction should be measured.
Template-based modeling [66] relies on the observation that structure is generally
more conserved than sequence; thus even at a low sequence similarity the structure
can be reliably modeled (http://predictioncenter.org/casp8/). However, the general
assumption of template-based modeling that proteins with similar sequence have
similar structure has recently been challenged [67], as there is growing evidence that
there are homologous proteins with a significant degree of sequence similarity but
different fold [68, 69]. These structural drifts [70] can explain the change of protein
structures during evolution, and have important consequences for protein design.
Recently the group of Phil Bryan could demonstrated experimentally that two
designed proteins with 88% sequence identity have different structure and func-
tion [71]. While these examples demonstrate that there seems to be a small number of
critical amino acid positions that can cause a significant change in structure, in most
positions an exchange of amino acids is tolerated and random mutations lead to a
failure in folding or a loss of function in only 34% of the mutants [72].
While template-based modeling is based on the experimentally determined
structure of a homologous protein, template-free modeling predicts the structure
of proteins solely from their amino acid sequence. Thus, all protein structure
prediction targets that lacked substantial similarity to a protein in the PDB at
the time of assessment were considered to be template-free modeling targets [73].
The success of template-free methods is still limited, as less than 50% of the targets
were modeled satisfactory [65]. In CASP8, only single domains were considered for
template-free predictions, and only two can be classified as actually new folds [65].
The vast majority of 164 assessments were classified as template-based modeling,
only 13 targets as template-free modeling [74]. This ratio will further increase in the
future. As a result of structural genomics and structural proteomics initiatives the
protein structure space will be more densely filled with experimental structures, and
so the need for template-free modeling is expected to decrease.
4.2.3.2 Prediction of Protein Stability and Solubility

Several simple rules to predict the stability of a protein from its sequence or structure
have been formulated over the years [75]. Most sequence-based methods to predict
stability analyze a family of homologous proteins to construct a consensus sequence
or an ancestor sequence. It has been shown that back-to-the-consensus mutations
have a high probability to increase stability [76–78] or improve expression [79].
Similarly, ancestral mutations have been successfully used to improve protein
stability [80, 81]. Recently, ancestral mutations have been integrated with directed
evolution [82] to generate a stabilized starting point of highly diverse and evolvable
gene libraries [83]. Alternatively, multi-sequence alignments were analyzed to identify
correlated mutations. In connection with structure models correlation analyses have
been used to identify structurally or functionally relevant residues [84, 85] and to
96
predict mutations to improve substrate specificity, catalytic activity, or protein

stability [86].
Apart from stability and function, aggregation of proteins is a major topic not only
in understanding of folding-related diseases but also in the recombinant expression
of proteins. It has long been that the net electric charge of a protein correlates with its
solubility, because proteins are least soluble near their isoelectric point [87]. In
contrast, increasing the net charge of a protein decreases its tendency to aggregate.
Highly charged variants of green fluorescent protein remained soluble when exposed
to conditions that normally cause proteins to aggregate [88]. Apart from charge,
hydrophobicity and secondary structure propensity are major properties that deter-
mine aggregation [89]. Sequence-based methods to predict aggregation-prone
regions have been developed [90] and applied to design mutants with decreased
aggregation rate [91].
Prediction and engineering of stability and solubility can be greatly enhanced by
taking into account structure information, because it accounts for spatial interactions
that are hidden from sequence analysis. It is appreciated that there are three major
properties that contribute to stability: electrostatic interactions, packing, and
quaternary structure [92]. Many strategies for rational protein stabilization have
been proposed, among them are the optimization of the distribution of surface
charge–charge interactions [93, 94], improvement of core packing [95] and of the
protein surface [96], and rigidification by introduction of prolines, exchange of
glycines, or the introduction of disulfide bridges [97]. However, it is still challenging
to reliably predict mutations that stabilize the enzyme without affecting its activity
or selectivity.
4.2.3.3 Docking
Specificity and selectivity of an enzyme is a direct consequence of the molecular
recognition of the substrate by the enzyme. Therefore, modeling of the enzyme–
substrate complex by molecular docking methods is used to study the molecular basis
of specificity and selectivity, and to predict mutations in the enzyme or modifications
of the substrate structure that mediate specificity or selectivity [98, 99]. It is
recognized that shape and physicochemical properties of the active site and the
substrate binding site are the major driving forces to provide the specific interactions
between enzyme and the transition state of the substrate that lead to catalysis.
Moreover, there is increasing evidence that flexibility of the enzyme–substrate
complex is crucial to recognition, because minor structural adjustments can have
a big impact on the docking score [37]. Therefore it is crucial to the success of docking
to start from a reliable structure model that has been determined under the relevant
conditions by X-ray or NMR analysis, or derived by template-based or template-free
modeling. Because of the sensitivity of the docking score to structure, induced fit
effects upon binding of the substrate have to be considered carefully [36, 37].
In most docking approaches, it is assumed that the binding site of the enzyme is in
a well-defined conformation prior to substrate binding, and an X-ray structure or a
structural model of the enzyme can be safely taken as a starting point for docking.
However, there is evidence for a pre-existing population of conformations of the
binding site in the absence of substrate. A small fraction of them is similar to the
conformation of the binding site after binding of the substrate while others
differ from the conformation of the substrate complex [100]. In addition, the
population might be shifted by mutations that are located far from the binding
site [101]. Apart from the shape of the substrate binding site and the specific
interactions between enzyme and substrate, displacement of water molecules from
the active site by the substrate has been suggested as a principal source of binding free
energy [102].
Docking has been extensively used to predict substrate specificity and to identify
positions that mediate substrate binding. Amino acids that clash with the desired
substrate upon docking were exchanged, leading to an increase of catalytic activity of
the enzyme variant towards this substrate [103–105]. The catalytic activity of wild type
and two variants of human anhydrase II toward substrates was in agreement with
docking results upon binding of a transition state – analogous inhibitor [106]. Based
on a docking study, a double mutant of a hydantoinase was designed with 200-fold
increased activity towards a desired substrate [107].
4.2.3.4 Molecular Dynamics Simulations

Geometry optimization by molecular mechanics and molecular dynamics simula-
tions are versatile methods to explore the conformational space of complex molecular
systems and to derive thermodynamic properties such as density, enthalpy, or
entropy. At the same time the dynamics of the system is evaluated by time correlation
functions, diffusion coefficients, or the kinetics of conformational transitions [108–
110]. There is no principal limitation of size or complexity of the system, but due to
the limitations of available computer resources the size of enzyme–solvent systems
generally does not exceed a few million particles and a simulation time of a few
ms [111–113]. Small systems have already been simulated on a timescale of ms [114],
which corresponds to the timescale of substrate binding and turnover. Programs that
have been widely applied in the field of biomolecular simulations are Amber [115],
BOSS/MCPRO [116], Desmond [117], Folding@home [118], GROMOS [119],
GROMACS [120], and NAMD [121]. Traditional force fields assume pairwise
interaction potentials between the atoms, including bonded and non-bonded con-
tributions, from which the resulting force acting on each atom is derived. The
molecular dynamics simulation itself consists of an iterative solution of Newtons
second law of motion with typical time steps of 1–5 fs (1 fs ¼ 1015 s). Prior to the
simulation, interaction parameters such as force constants of bonded interactions or
partial atomic charges are assigned. It is assumed that these parameters do not
change during the simulation, and no bonds are broken or formed. However, the
assumption of fixed partial atomic charges does not reflect the dependence of
the electronic structure of a molecule on its environment. Therefore, to model the
influence of water or organic solvents on the polarization of the protein, more recent
developments allow for a polarizability of bonds [122]. The first stable 2 ns simulation
of a polarizable model of bovine pancreatic trypsin inhibitor (BPTI) in water was
performed [123]. While the overall structure of the protein was similar to a
nonpolarizable simulation, the hydrogen bonding pattern and the structural and
98
dynamic properties of the solvent surrounding the protein differed. However, further
studies will be necessary to analyze the accuracy of polarizable protein models. As an
alternative to parameterization of mechanical models, force fields are being devel-
oped that are described explicitly by a quantum chemical wave function [124]. In such
a force field, polarization and charge transfer are implicitly included, and the method
could be used to model chemical reactions. A 50 ps molecular dynamics simulation of
BPTI in water demonstrated that water has a significant polarization effect and that a
charge transfer occurs between amino acids. Thus, residues of the same type may
have average charges that differ by up to 0.1 atomic units depending on protein
sequence, and the instantaneous excess charges vary even more [124]. However,
further work is still needed to evaluate the relevance of charge transfer to biochemical
or biophysical properties of proteins.
Molecular dynamics simulations have been applied to investigate the effect of
mutations or solvent to the biochemical properties of enzymes such as stability,
specificity, or selectivity. Mutations in tightly packed regions of the proteins are
expected to change stability, because they lead to local changes of the non-bonded
interaction energies. Upon simulation of mutations in representative proteins of five
different fold families, not only local rearrangements of the protein structure near the
mutated site were observed, but also long-range cooperative changes [125]. Protein
stability is also achieved by salt bridge networks on the protein surface, especially at
higher temperatures. In simulations of proteins with salt bridge networks, the
increase of configurational entropy at higher temperature did not lead to a corre-
sponding increase of the root-mean-square fluctuations, but the disorder caused by
thermal motion was accommodated to produce thermostability and to prevent
unfolding [126].
The delicate balance between rigidity and flexibility near the active site has also
been associated with the different temperature activity profiles of psychrophilic and
mesophilic homologues. While in a mesophilic a-amylase the substrate-binding
loops are longer and the fluctuations occur mainly near the tip of the loops, far away
from the active site, the substrate-binding loops are shorter in the psychrophilic
homolog and showed a higher flexibility in the immediate neighborhood of the active
site [127]. Differences in flexibility were also observed for mutations in a psychro-
philic lipase that shifted its temperature optimum to higher temperatures [128].
Molecular dynamics simulations have been extensively used to study specificity
and selectivity of enzymes by modeling enzyme–substrate complexes assuming
complete flexibility. For various enzyme–substrate complexes, critical parameters
such as the distance between the catalytic site and the bound substrate were shown to
be predictive for activity, specificity, and selectivity of enzymes such as lipases [129–
133] and cytochrome P450 monooxygenases [55, 134, 135]. Even indirect long-range
effects of mutations in metallo-b-lactamases could be modeled successfully [136]. For
an esterase, catalytic activity toward novel esters was predicted by combining docking
and molecular dynamics simulations [137]. Especially for highly flexible binding sites
such as in cytochrome P450 monooxygenases, molecular dynamics simulations have
proven to be superior to molecular docking in reproducing experimentally deter-
mined selectivity [138] and binding affinity [139].
A topic of increasing relevance is the role of fluctuations, protein dynamics, and
coupled motions to binding of ligands and to the activity of enzymes. NMR studies in
the 1990s demonstrated for the first time the surprising role of conformational
entropy in protein stability and ligand binding [140]. While intuitively we would
expect that the flexibility of a protein decreases upon binding of a ligand (which is true
in most cases), the opposite effect was observed upon binding of a hydrophobic
ligand into the hydrophobic binding pocket of mouse major urinary protein, a small
lipocalin protein. Upon binding, the NMR-derived order parameters for backbone
NH groups of the protein decreased significantly, which corresponds to an increase
in backbone motion [141]. This observation is not unique, and an increase of protein
flexibility upon binding of ions, small molecules, peptides, or nucleic acids was
observed for other proteins, too [140].
These observations point to the crucial role of entropic contributions to binding of
ligands and substrate. It is supported by experimental evidence from careful measure-
ments of the enthalpic and entropic contributions to stereoselectivity of lipases and led
to surprising insights: while in most cases stereoselectivity was driven by enthalpy and
counterbalanced by entropy [142], in some cases it is driven by both enthalpy and
entropy [143]. Thus, rational design of enantioselective enzymes requires considera-
tions of entropy [144]. As a consequence, flexibility is a crucial, inherent property of the
substrate binding site of enzymes. This is consistent with the observation that binding
sites are located in regions most able to affect the cooperative network of interacting
amino acids, and it has been speculated that enzymes use conformational fluctuations
in carrying out their functions [145]. Comprehensive analyses of short- and long-range
effects of mutations in dihydrofolate reductase on protein motion and activity
supported this notion [146], and molecular dynamics simulations contributed con-
siderably to learning about coupled motions of residues. Although simulations are
powerful tools to interpret experimental results in retrospect, we are only at the
beginning of applying these methods for a predictive, rational design of enzymes.
4.2.3.5 Quantum Chemical Methods

The contribution of an enzyme in accelerating the chemical step of substrate
conversion, bond breaking and forming, is studied by methods based on quantum
chemistry. It is assumed that the effect of enzymes in accelerating the rate of a
chemical reaction can be derived from temporary, local interactions between the
substrate and a few catalytically active amino acids, metals, or cofactors [147], or by
electrostatic interactions between the enzyme and the substrate in its ground or
transition state [148]. By reducing the system size to a minimum, the best and most
precise methods can be applied to identify the reaction path and the most relevant
transition states, and to calculate energy barriers at a precision within 1 kcal
mol1 [149]. To explicitly take into account the influence of the enzyme environment,
a hybrid method was introduced more than 30 years ago [150]. QM/MM combines a
quantum chemical analysis of the active site and a molecular mechanics treatment of
the enzyme environment, and was successfully applied to create hypotheses about
the reaction mechanism, to calculate geometries and energy barriers, and to design
mutants with a changed chemistry [12, 147, 151, 152].
100
4.2.4
Role of Solvent
4.2.4.1 Hydration of Enzymes

A systematic analysis of high-resolution crystal structures demonstrated that in all
proteins a small number of water molecules are tightly bound to water binding sites
on the protein surface and inside a protein [153–155]. In addition to tightly bound
water molecules, a network of immobilized water molecules is formed at the surface
of all water-soluble proteins [156]. It has been suggested that protein-bound water has
multiple roles in enzymes: it contributes to stabilization of the protein structure,
mediates protein flexibility, and binds loosely to the substrate binding site. Because
fully dry enzymes in organic solvents are inactive, water has been assumed to play the
role of a lubricant unlocking the structure upon addition to a dry protein [157]. At
higher amounts of water, it has been shown that lysozyme is slowly denaturat-
ing [158]. In general, the catalytic activity increases with water activity (aw) for
different enzymes such as pig liver esterase [159], lipases [160], chloroperoxi-
dase [161], or bilirubin oxidase [162]. For immobilized lipases it has been shown
that the catalytic activity depends on aw and less on the amount of immobilized
enzyme or the properties of the support [160], while activation of alcohol dehydro-
genase and a-chymotrypsin by the presence of additives was observed at low water
activity [163]. In contrast to catalytic activity that increased with aw, the enantios-
electivity of Candida antarctica lipase B increased for low water activities until
aw ¼ 0.2, but decreased for higher water activities [164]. To investigate directly the
role of water in catalytic activity and to avoid interaction of the enzyme with non-
reacting organic solvent molecules, enzyme-catalyzed reactions have been studied in
a continuous solid/gas reactor [165]. The dependence of catalytic activity on aw was
similar to results obtained in organic solvents [166], but the entropic and enthalpic
contributions differed in gas phase and in organic solvent [167]. Recently, binding
of water to dry C. antarctica lipase B in the gas phase was quantitatively modeled
by molecular dynamics simulations, and the dependence of the amount of bound
water on water activity agreed with experimental results obtained by sorption
isotherms [168].
4.2.4.2 Enzymes in Organic Solvents

Changing the solvent of an enzyme from water to organic solvent has several
consequences. Apart from the effect on the thermodynamic equilibrium of a reaction
by changing the effective concentration of substrates and products, organic solvents
exert multiple effects to the properties of an enzyme–substrate complex: they mediate
water binding to the enzyme, decrease protein flexibility, and might change the
enzyme structure.
It was experimentally observed that nonpolar solvents with a high octanol–water
partition coefficient (log P) led to a high catalytic activity, while catalytic activity
decreases in more polar organic solvents. This observation was explained by the
partitioning of water between the protein surface and the solvent; thus more polar
solvents act by stripping off the essential water molecules from the enzyme and thus
lead to deactivation [157]. In addition, solvent molecules might also directly bind to
the substrate binding site and interfere with substrate binding [169].
The major effect, however, is linked to a reduction of protein flexibility in organic
solvents. Molecular dynamics simulation of proteins in organic solvents indicated
the molecular mechanism: at low water content in solvents of increasing log P, the
exchange rate of protein-bound water molecules decreased, and spanning water
networks were gradually formed, leading to a reduction of protein flexibility [170].
Increasing the amount of water increased the fluctuations of a cutinase simulated in
hexane–water mixtures [171]. Up to a water content of 10 wt.%, the protein structure
was native-like, but started to unfold for higher water contents.
Careful experimental studies in the group of Karl Hult pointed to the relevance of
protein flexibility and water binding by measuring the contribution of enthalpy and
entropy to substrate specificity and enantioselectivity of C. antarctica lipase B. For
most substrates, enantioselectivity was driven by enthalpy and counterbalanced by
entropy [143]. Interestingly, for some substrates both enthalpy and entropy drove
enantioselectivity. In a systematic study of the effect of solvent on enantioselectivity,
the same group observed a correlation between the log P value of the solvent and
enantioselectivity. Increasing the log P of the solvent increased both the enthalpic and
the entropic contribution, and led to a better discrimination between the two
enantiomers of a chiral substrate [172], which might reflect the decreased flexibility
of the binding site in nonpolar solvents.
To understand the molecular basis of the effects of organic solvent on proteins,
molecular dynamics simulations are valuable tools, because they allow the simul-
taneous exploration of the energy landscape and the dynamics of the protein–solvent
system. In simulations of lipases in organic solvents with several hundred water
molecules bound to the protein, their structure did not change, but their flexibility
was reduced in solvents with high log P [170, 173]. Two molecular mechanisms were
identified that led to an increased rigidity in organic solvents: formation of salt
bridges on the protein surface and formation of a spanning water network of water
molecules on the protein surface that were tightly bound to the protein and slowly
exchanging with the solvent, similar to simulations in the gas phase [168]. Previously,
simulations were performed to study the protective role of the co-solvent trimethy-
lamine N-oxide and the mechanism of denaturation by urea [174], and to identify the
molecular mechanism of protein stabilization or unfolding. In both cases, water plays
a central role. While urea disrupts the hydrogen bonding network of water which
leads to a hydration of the hydrophobic core of the protein, trimethylamine N-oxide
stabilizes water–water interactions, thus stabilizing the protein as a result of the
increased penalty for the hydration of hydrophobic residues.
4.2.4.3 Solvent-Induced Conformational Changes

Most lipases have a mobile lid that covers the active site and blocks it from substrate
access. In an environment with high log P or at a hydrophobic substrate interface, it
has been proposed that the lid opens. X-Ray structures of closed and open forms have
been determined for many lipases. To study the molecular basis of this solvent-
induced conformational change, molecular dynamics simulations of lipase in the
102
presence of organic solvent were performed. Short simulations of lipase with a

closed lid revealed an increased flexibility of the lid region in organic solvent [173,
175] or in the presence of its triglyceride substrate [176]. Only recently, full lid
opening was observed in 50 ns molecular dynamics simulations of lipases in
toluene [177], octane [178], or in the presence of triglyceride [179]. Upon binding
to a triglyceride aggregate, the lid served as an anchor of the enzyme to the substrate
interface.
4.3
De Novo Design of Stable and Functional Proteins
Since the early 1970s when the experimental structures of an increasing number of
proteins became available, it became apparent that homologous proteins have a
similar structure despite their sometimes considerable difference in sequence [66].
Based on this observation, in the mid-1980s the now widely used method of
homology modeling or comparative modeling was introduced [66, 180], which
assumed that if the sequences of a target and a template protein are similar the
structure of the target protein can be modeled on the basis of the template protein.
The high similarity of the structures of homologous proteins even at low sequence
similarity is a consequence of their high degree of plasticity, which allows them to
adjust for amino acid exchanges while maintaining their overall structure. This is not
only true for amino acids on the protein surface but also in the protein core [181]. For
orotate phosphoribosyltransferase, it was even possible to achieve a stable and
functional enzyme variant where 88% of all residues are from a reduced alphabet
of only nine amino acids (A, D, G, L, P, R, T, V, Y), while seven amino acids (C, H, I, M,
N, Q, W) are completely missing [182]. Thus, stable protein cores are sufficiently
forgiving to accommodate exchanges of side chains, and there might be many
different sequences that fold into one structure. This has led to the development of
algorithms to solve the reverse folding problem (Is a sequence compatible with a
particular structure?) by empirical potentials [183, 184]. While threading approaches
identify regions where sequence and structure are incompatible, they could not be
applied successfully to solve the folding problem or to design new proteins. For small,
fast-folding protein domains, direct molecular dynamics simulations were able to
model the folding pathway and the native structure [185–187]. Although these
methods allow in principle the prediction of structure and of the folding pathway,
they have not yet been used for re-design or de novo design of proteins.
A breakthrough in de novo protein design came in the mid-1990s when the group of
Stephen Mayo introduced the ORBITprotein design software. For a particular protein
structure, the software searched for the globally optimal sequence and the optimal
side chain conformations using transferable, atom-based potentials. Thus, thermo-
stability of a homodimeric coiled coil was increased by re-design of the buried
hydrophobic surface [188], and a hyperstable variant of the streptococcal protein Gb1
domain was designed [189]. At the same time the group of David Baker presented
their de novo design tool ROSETTA, which successfully predicted the structure of
4.3 De Novo Design of Stable and Functional Proteins j103
three small proteins in the range of 67 to 99 residues [190]. Using ROSETTA, the first
de novo design of a protein with a new fold that was not yet observed in nature was
presented four years later [191]. The design was highly accurate with a overall

deviation between modeled and experimental structure of only 1.2 A, and the protein
was exceptionally stable with a melting temperature far above 100 C.
Beyond the design of stable proteins, the ultimate challenge of enzyme design is
the design of catalytic function, either by transferring activity to a catalytically inactive
protein or by designing enzymes with new catalytic functions or selectivity. The
design of catalytic function was greatly aided by two observations: first, homologous
enzymes might have different functions, which have developed during natural
evolution. Thus, analyzing sequence–structure–function relationships in enzyme
families was applied successfully to (re)design enzymes [192]. Second, and more
surprisingly, even single enzymes might have multiple functions. While it has long
been recognized that most enzymes accept alternative substrates [193], an increasing
number of enzymes have been found to catalyze multiple chemical reactions, a
phenomenon that has been referred to as catalytic promiscuity [194]. Indeed, recent
experimental evidence suggests that catalytic promiscuity is not as rare as was
previously thought [195], but the evolutionary and mechanistic aspects of catalytic
promiscuity are barely understood [196]. However, understanding the molecular
basis of promiscuity would enable us to transfer catalytic activity to inactive protein or
to change the catalytic activity of an enzyme by re-design. Many successful examples
prove the power of this concept: engineering of a peptidyl-prolyl cis-trans isomerase
into a endopeptidase [197], redesigning a lipase into an aldolase [198] or into enzymes
that catalyze Baeyer–Villiger oxidation with hydrogen peroxide [199], epoxidation of
a,b-unsaturated aldehydes with hydrogen peroxide [200], Michael additions [201],
or hydrolysis of epoxides [202]. Therefore, using promiscuous enzymes and improv-
ing them by protein engineering is regarded as a promising strategy in biocataly-
sis [195, 203, 204].
The first successful steps have been made to design enzymes that catalyze a desired
reaction and have a catalytic function that has not been observed in nature, yet. In the
group of David Baker a retro-aldol enzyme was designed that showed a rate
acceleration of the catalyzed versus the uncatalyzed reaction of 104 [205], and an
enzyme that catalyzes a Kemp elimination reaction with a rate acceleration of
105 [206]. The latter enzyme could be further improved 200-fold by directed evolution.
Apart from the practical aspect of creating biocatalysts with new catalytic functions
that could be applied in chemical synthesis, the ability of designing a biocatalyst with
a desired catalytic function is a major challenge to our full understanding of
enzymatic function and the ultimate goal of protein engineering.
A recent systematic investigation of enzymatic mechanisms and active sites of
enzymes revealed that most enzyme reactions rely upon nucleophilic and general
acid–base chemistry, while radical reactions are rare and electrophilic reactions in
enzymes are very rare [207]. Most catalytic amino acid residues stabilize the substrate
or serve as proton shuttle. Two amino acids, histidine and cysteine, have an
extraordinarily high propensity for being a catalytic residue, and cysteine is the
most catalytically versatile residue, being involved in 70% of all reaction types. There
104
are only a handful of predominant catalytic mechanisms in the enzymes observed in

nature, and the catalytically active residues have a rather limited repertoire of
function. Thus, there seems to be room for more chemistry.
4.4
Challenges and Outlook
4.4.1
Force Field, System Size, and Simulation Time
Although we have seen impressive advances in our understanding of enzymes and in

our ability to design the structure and biochemical properties of enzymes, we are still
far from a situation that allows us to quantitatively predict the behavior of enzyme–
solvent systems. There are three technical limitations that restrict our ability to model
realistic systems. First, classical, non-polarizable force fields are a result of a
compromise between transferability of atom types and a parameterization that
optimally describes structure, thermodynamics, and transport properties. Therefore,
the force fields are continuously improved by comparing them to each other [208],
checking with experimental data [209], and especially by applying long timescale
simulations [210].
The second limiting factor is the time scale that is accessible to molecular
dynamics simulations and that limits the conformational space that can be sampled.
Currently, simulation times between 100 ns to 1 ms can be reached for medium-sized
protein–solvent systems (30 000–60 000 atoms) with available hard- and software,
and systems with production rates of more than 10 ms per day are under construc-
tion [211]. For small proteins, molecular dynamics simulations in the ms timescale
have been performed on the distributed computing system Folding@home
[185], which is within the biochemically relevant timescale of folding or catalytic
turnover.
The third factor is the complexity of the system to be investigated. In the
first molecular dynamics simulations of proteins, a single molecule was simulated
in vacuo, neglecting the dielectric and hydrodynamic effects of the solvent [212]. Since
then it has become clear that binding to a solid surface [213] and also the solvent, pH,
and temperature [214] have an influence on protein structure and dynamics. Up to
now mostly single protein molecules have been studied in their solvent environment.
However, at high protein concentration intermolecular interactions become relevant,
which can only be treated by including more than one protein into the simulation.
Therefore, a realistic description of biocatalysts demands the simulation of protein
systems far beyond the current system size of 60 000 atoms and simulation time of
107 time steps. However, these are merely technical limitations that will be overcome
as massively parallel computing architectures become available. In parallel, the force
fields will be gradually refined, and new force fields will be introduced, which have
the potential to reproduce structure, thermodynamics, and transport properties
simultaneously.
4.4 Challenges and Outlook j105
4.4.2
Enzymes are Nanomachines
More challenging than the technical limitations, however, are the limitations of our
current scientific understanding of how enzymes work. The current, widely accepted
concept of an enzymatic mechanism is based on the assumption of a single rate-
limiting transition state: for most enzymes it is assumed that there is a rate-limiting
step along the reaction path from substrate to product. According to modern
transition state theory, the acceleration of the reaction rate is achieved by lowering
the activation free energy, whereas the effect of the transmission coefficient can be
neglected [215]. Thus, binding of the transition state is assumed to determine
enzymatic activity, substrate specificity, and regio-, chemo-, and stereoselectivity.
However, it has become evident that enzymes are more complex. Single-molecule
kinetics of hydrolysis catalyzed by Candida antarctica lipase A revealed that the
enzyme exists in a broad spectrum of conformations, of which only certain con-
formations are catalytically active [216], and transitions between active and inactive
conformations play a crucial role in determining enzymatic activity. Thus, enzymes
behave like complex nanomachines [217].
Usually the term nanomachine is associated with proteins that function as
molecular motors and produce rotary [218] or linear [219] motion. Similarly, enzymes
function by a sequence of steps, each associated with a conformational change of the
enzyme, the substrate(s), or a cofactor: binding of the substrate, guiding it to the
active site, performing a series of local conformational changes and chemical
reactions that lead to the product(s), and finally releasing the product(s). Each of
these steps might be rate-limiting and relevant to specificity or selectivity. The
combination of high-resolution structure determination by X-ray crystallography or
NMR techniques, careful kinetic characterization, especially by single-molecule
experimentation, extensive modeling of the relevant steps in the catalytic cycle, and
validation by designing mutants has provided new insights into the complexity of
these nanomachines.
Further evidence that supports this view of enzymes as nanomachines is based on
the observation that specificity and selectivity is not exclusively determined by
binding of the transition state, but can be mediated by regions in the enzyme far
from the active site. For a haloalkane dehalogenase, activity- and specificity-deter-
mining residues have been identified that are located at the entrance of a tunnel
leading from the bulk solvent to the active site [220]. Similarly, mutations in the
substrate access channel of Burkholderia cepacia lipase [221] and Pseudomonas
fluorescens esterase [222] considerably increased stereoselectivity. In addition, in the
binding site of cytochrome P450 monooxygenases there are sites far from the active
site that are involved in substrate transfer and influence selectivity and
specificity [223].
In this view, the catalytic reaction cycle consists of multiple relevant steps [224].
Careful investigation of dihydrofolate reductase pointed out the relevance of confor-
mational changes and coupled motions on the ms to ms time scale. It has been
suggested that multiple sequential intermediates occur as catalysis proceeds, and
106
there might be multiple reaction paths [225]. The rate-limiting step changes with
reaction conditions: in dihydrofolate reductase, product release is rate limiting at low
pH but above pH 8.4 hydride transfer is rate limiting [226]. According to these
observations, the remarkable efficiency of enzymes is achieved by a fast passage
through a multidimensional free energy landscape with multiple minima and
transition states. Protein dynamics, coupling between the motions of residues, and
fluctuations play a crucial role. Enzymes are complex molecular machines with an
essentially stochastic behavior, and thus are different from man-made machines [227].
The need for a more dynamic view of enzyme catalysis is further underlined by our
current limitations in de novo design. Stable proteins that bind the transition state
along the supposed pathway of a chemical reaction with high affinity can be designed
successfully. These de novo designed enzymes led to a substantial enhancement of the
reaction rate of the catalyzed reaction by a factor of 104–106 relative to the uncatalyzed
reaction. This rate enhancement is of the same order of magnitude as catalytic
antibodies, which follow the same principle of catalysis. Thus, by stabilizing the
transition state, a rate enhancement of 106 can be reached, which is still far from the
rate enhancement of a real enzyme (of the order of 1010–1019 [228]).
4.4.3
Outlook
In the further development of biocatalysis, integration will play a dominant role:

integration of experiment with modeling, of basic research with application, and of
the cultures of biologists, chemists, physicists, and engineers. Five fields seem to be
crucial in our ambitious venture to design new biocatalysts with a catalytic efficiency
that is useful for synthetic applications, beyond the catalytic power of catalytic
antibodies:
1) Large-scale sequencing and comprehensive sequence analysis to explore the
sequence space of enzymes found in nature.
2) Structural characterization of the reaction path in enzyme-catalyzed reactions by
trapping of intermediates [229].
3) Molecular enzymology to provide reproducible kinetic data under defined
reaction conditions. Here new techniques such as microfluidics and single
molecule experimentation promise high-throughput, low volume, and highly
reproducible experiments.
4) Extensive modeling of sufficiently large and complex systems using long
timescale simulations and sampling of ensembles of trajectories instead of
single simulations to include contributions from entropy and the effect of
solvent, temperature, and mutations.
5) Synthetic capabilities of modern molecular biology such as synthesis of single
genes or gene libraries, reliable recombinant expression, and fast purification of
newly designed enzymes.
In our view, biocatalyst design will be able to predict quantitatively the properties of
realistic enzyme–solvent systems and to construct new biocatalysts. While most of
References j107
our current protein design efforts use amino acids as basic parts, de novo design is
not limited to proteinogenic amino acids. By using expression platforms with
expanded genetic code [230], single unnatural amino acids can be incorporated
by in vivo [231] or in vitro protein biosynthesis [232]. Combining amino acids with
other components from organic and inorganic chemistry is expected to enhance
considerably the synthetic potential [233–236]. Thus, enzyme engineering will
become truly synthetic, beyond imitating or improving naturally observed enzymes.
It is the vision of synthetic biotechnology to encourage human engineers to design
new biocatalysts in the same way we design cars, computers, and production
processes, by a complete and realistic simulation of the structure and the properties
of the desired biocatalyst.
In his Nobel award lecture in 1902, Emil Fischer presented his vision on the future
of biochemistry:
The organism is capable of performing highly specific chemical transforma-

tions which can never be accomplished with the customary agents. To equal
Nature here, the same means have to be applied, and I therefore foresee the
day when physiological chemistry will not only make extensive use of the
natural enzymes as agents, but when it will also prepare synthetic ferments for
its purposes.
(http://nobelprize.org/nobel_prizes/chemistry/laureates/1902/fischer-lecture.html)
After more than 100 years of successful biochemical research, we are beginning to
approach his vision.
References
1 Serdakowski, A.L. and Dordick, J.S. 5 Payen, A. and Persoz, J.-F. (1833)
(2008) Enzyme activation for organic Memoire sur la diastase, les principaux
solvents made easy. Trends Biotechnol., produits de ses reactions et leurs
26 (1), 48–54. applications aux arts industriels. Ann.
2 Unsworth, L.D., van der Oost, J., and Chim. Phys., 53 (2), 73–92.
Koutsopoulos, S. (2007) 6 Fischer, E. (1894) Einfluss der
Hyperthermophilic enzymes – stability, configuration auf die wirkung der
activity and implementation strategies enzyme. Ber. Dtsch. Chem. Ges., 27 (3),
for high temperature applications. FEBS 2985–2993.
J., 274 (16), 4044–4056. 7 Koshland, D.E. (1958) Application of a
3 Wood, B.M. et al. (2009) Mechanism theory of enzyme specificity to protein
of the orotidine 50 -monophosphate synthesis. Proc. Natl. Acad. Sci. U.S.A.,
decarboxylase-catalyzed reaction: 44 (2), 98–104.
effect of solvent viscosity on kinetic 8 Eyring, H. and Stearn, A.E. (1939) The
constants. Biochemistry, 48 (24), application of the theory of absolute
5510–5517. reaction rates to proteins. Chem. Rev.,
4 Luetz, S., Giver, L., and Lalonde, J. (2008) 24 (2), 253–270.
Engineered enzymes for chemical 9 Pauling, L. (1946) Molecular architecture
production. Biotechnol. Bioeng., 101 (4), and biological reactions. Chem. Eng.
647–653. News, 24, 1375–1377.
108
10 Radzicka, A. and Wolfenden, R. (1995) A 24 Bairoch, A. (1991) PROSITE: a dictionary

proficient enzyme. Science, 267 (5194), of sites and patterns in proteins. Nucleic
90–93. Acids Res., 19 (Suppl), 2241–2245.
11 Kendrew, J.C. et al. (1958) A three- 25 Attwood, T.K. and Beck, M.E. (1994)
dimensional model of the myoglobin PRINTS – a protein motif fingerprint
molecule obtained by x-ray analysis. database. Protein Eng., 7 (7), 841–848.
Nature, 181 (4610), 662–666. 26 Pietrokovski, S., Henikoff, J.G., and
12 Senn, H.M. and Thiel, W. (2009) QM/ Henikoff, S. (1996) The Blocks database –
MM methods for biomolecular a system for protein classification. Nucleic
systems. Angew. Chem. Int. Ed., 48 (7), Acids Res., 24 (1), 197–200.
1198–1229. 27 Corpet, F., Gouzy, J., and Kahn, D. (1998)
13 Urban, P., Truan, G., and Pompon, D. The ProDom database of protein domain
(2008) High-throughput enzymology families. Nucleic Acids Res., 26 (1),
and combinatorial mutagenesis for 323–326.
mining cytochrome P450 functions. 28 Sonnhammer, E.L. et al. (1998) Pfam:
Expert Opin. Drug Metab. Toxicol., 4 (6), multiple sequence alignments and
733–747. HMM-profiles of protein domains.
14 Wackett, L.P. (2004) Novel biocatalysis by Nucleic Acids Res., 26 (1), 320–322.
database mining. Curr. Opin. Biotechnol., 29 Henke, E., Pleiss, J., and Bornscheuer,
15 (4), 280–284. U.T. (2002) Activity of lipases and
15 Tettelin, H. and Feldblyum, T. (2009) esterases towards tertiary alcohols:
Bacterial genome sequencing. Methods insights into structure-function
Mol. Biol., 551, 231–247. relationships. Angew. Chem. Int. Ed.,
16 Ansorge, W.J. (2009) Next-generation 41 (17), 3211–3213.
DNA sequencing techniques. Nat. 30 Hoegger, P.J. et al. (2006) Phylogenetic
Biotechnol., 25 (4), 195–203. comparison and classification of laccase
17 Ferrer, M. et al. (2007) Mining enzymes and related multicopper oxidase protein
from extreme environments. Curr. Opin. sequences. FEBS J., 273 (10), 2308–2326.
Microbiol., 10 (3), 207–214. 31 de Groot, M.J. and Ekins, S. (2002)
18 Gabor, E. et al. (2007) Updating the Pharmacophore modeling of
metagenomics toolbox. Biotechnol. J., cytochromes P450. Adv. Drug Deliv. Rev.,
2 (2), 201–206. 54 (3), 367–383.
19 Raes, J., Foerstner, K.U., and Bork, P. 32 Gedeck, P. and Lewis, R.A. (2008)
(2007) Get the most out of your Exploiting QSAR models in lead
metagenome: computational analysis of optimization. Curr. Opin. Drug Discov.
environmental sequence data. Curr. Devel., 11 (4), 569–575.
Opin. Microbiol., 10 (5), 490–498. 33 Terfloth, L., Bienfait, B., and Gasteiger, J.
20 Altschul, S.F. et al. (1990) Basic local (2007) Ligand-based models for the
alignment search tool. J. Mol. Biol., isoform specificity of cytochrome P450
215 (3), 403–410. 3A4, 2D6, and 2C9 substrates. J. Chem.
21 Eddy, S.R. (1998) Profile hidden Inf. Model, 47 (4), 1688–1701.
Markov models. Bioinformatics, 14 (9), 34 Braiuca, P. et al. (2006) Computational
755–763. methods to rationalize experimental
22 Ohlson, T., Wallner, B., and Elofsson, A. strategies in biocatalysis. Trends
(2004) Profile-profile methods provide Biotechnol., 24 (9), 419–425.
improved fold-recognition: a study of 35 Hermann, J.C. et al. (2006) Predicting
different profile-profile alignment substrates by docking high-energy
methods. Proteins, 57 (1), 188–197. intermediates to enzyme structures.
23 Atkinson, H.J. et al. (2009) Using J. Am. Chem. Soc., 128 (49), 15882–15891.
sequence similarity networks for 36 Tyagi, S. and Pleiss, J. (2006) Biochemical
visualization of relationships across profiling in silico – predicting substrate
diverse protein superfamilies. PLoS One, specificities of large enzyme families.
4 (2), e4345. J. Biotechnol., 124 (1), 108–116.
References j109
37 Juhl, P.B. et al. (2009) Modelling classification of vitamin B6-dependent
substrate specificity and enzymatic activities and of the
enantioselectivity for lipases and corresponding protein families.
esterases by substrate-imprinted BMC Bioinformatics, 10, 273.
docking. BMC Struct. Biol., 9, 39. 50 Fischer, M. et al. (2006) DWARF – a
38 Kalyanaraman, C. et al. (2008) Discovery data warehouse system for analyzing
of a dipeptide epimerase enzymatic protein families. BMC Bioinformatics,
function guided by homology modeling 7, 495.
and virtual screening. Structure, 16 (11), 51 Joosten, H.J. (2008) Method of generating
1668–1677. a protein database. WO 2008/035970.
39 Pieper, U. et al. (2009) Target selection 52 Kahraman, A. et al. (2007) Shape
and annotation for the structural variation in protein binding pockets
genomics of the amidohydrolase and and their ligands. J. Mol. Biol., 368 (1),
enolase superfamilies. J. Struct. Funct. 283–301.
Genomics, 10 (2), 107–125. 53 Reetz, M.T. et al. (2001) Directed
40 Holliday, G.L. et al. (2005) MACiE: a evolution of an enantioselective enzyme
database of enzyme reaction through combinatorial multiple-cassette
mechanisms. Bioinformatics, 21 (23), mutagenesis. Angew. Chem. Int. Ed.,
4315–4316. 40 (19), 3589–3591.
41 Andreini, C. et al. (2008) Metal ions in 54 Li, Q.S. et al. (2001) Rational evolution of
biological catalysis: from enzyme a medium chain-specific cytochrome P-
databases to general principles. 450 BM-3 variant. Biochim. Biophys. Acta,
J. Biol. Inorg. Chem., 13 (8), 1545 (1–2), 114–121.
1205–1218. 55 Seifert, A. and Pleiss, J. (2009)
42 Rawlings, N.D. et al. (2008) MEROPS: the Identification of selectivity-determining
peptidase database. Nucleic Acids Res., residues in cytochrome P450
36 (Database issue), D320–D325. monooxygenases: a systematic analysis of
43 Fischer, M. and Pleiss, J. (2003) The the substrate recognition site 5. Proteins,
Lipase Engineering Database: a 74 (4), 1028–1035.
navigation and analysis tool for protein 56 Seifert, A. et al. (2009) Rational design of a
families. Nucleic Acids Res., 31 (1), minimal and highly enriched CYP102A1
319–321. mutant library with improved regio-,
44 Hotelier, T. et al. (2004) ESTHER, the stereo- and chemoselectivity.
database of the alpha/beta-hydrolase ChemBioChem, 10 (5), 853–861.
fold superfamily of proteins. Nucleic 57 Aehle, W. and Estell, D.A. (2008)
Acids Res., 32 (Database issue), Systematic evaluation of sequence and
D145–D147. activity relationships using site
45 Fabian, P. and Degtyarenko, K.N. (1997) evaluation libraries for engineering
The directory of P450-containing systems multiple properties. WO 2008/002472.
in 1996. Nucleic Acids Res., 25 (1), 58 Barak, Y. et al. (2008) Enzyme
274–277. improvement in the absence of structural
46 Fischer, M. et al. (2007) The Cytochrome knowledge: a novel statistical approach.
P450 Engineering Database: a navigation ISME J., 2 (2), 171–179.
and prediction tool for the cytochrome 59 Chaparro-Riggers, J.F., Polizzi, K.M., and
P450 protein family. Bioinformatics, Bommarius, A.S. (2007) Better library
23 (15), 2015–2017. design: data-driven protein engineering.
47 Lisitsa, A.V. et al. (2001) Cytochrome Biotechnol. J., 2 (2), 180–191.
P450 database. SAR QSAR Environ. Res., 60 Fox, R.J. et al. (2007) Improving catalytic
12 (4), 359–366. function by ProSAR-driven enzyme
48 Park, J. et al. (2008) Fungal cytochrome evolution. Nat. Biotechnol., 25 (3),
P450 database. BMC Genomics, 9, 402. 338–344.
49 Percudani, R. and Peracchi, A. (2009) The 61 Ulmer, K.M. (1983) Protein engineering.
B6 database: a tool for the description and Science, 219 (4585), 666–671.
110
62 Blundell, T.L. et al. (1989) Protein 75 Damborsky, J. and Brezovsky, J. (2009)

engineering and design. Philos. Trans. R. Computational tools for designing and
Soc. London, Ser. B., 324 (1224), 447–460. engineering biocatalysts. Curr. Opin.
63 Moult, J. et al. (2009) Critical assessment Chem. Biol., 13 (1), 26–34.
of methods of protein structure 76 Lehmann, M. et al. (2000) The consensus
prediction - Round VIII. Proteins, 77 concept for thermostability engineering
(Suppl 9), 1–4. of proteins. Biochim. Biophys. Acta,
64 Keedy, D.A. et al. (2009) The other 90% 1543 (2), 408–415.
of the protein: assessment beyond the 77 Steipe, B. et al. (1994) Sequence statistics
Calphas for CASP8 template-based and reliably predict stabilizing mutations in a
high-accuracy models. Proteins, 77 (Suppl protein domain. J. Mol. Biol., 240 (3),
9), 29–49. 188–192.
65 Ben-David, M. et al. (2009) Assessment of 78 Vazquez-Figueroa, E. et al. (2008)
CASP8 structure predictions for template Thermostable variants constructed via
free targets. Proteins, 77 (Suppl 9), 50–65. the structure-guided consensus method
66 Sutcliffe, M.J. et al. (1987) Knowledge also show increased stability in salts
based modelling of homologous solutions and homogeneous aqueous-
proteins, Part I: Three-dimensional organic media. Protein Eng. Des. Sel.,
frameworks derived from the 21 (11), 673–680.
simultaneous superposition of multiple 79 Dai, M. et al. (2007) The creation of a
structures. Protein Eng., 1 (5), 377–384. novel fluorescent protein by guided
67 Petrey, D., Fischer, M., and Honig, B. consensus engineering. Protein Eng. Des.
(2009) Structural relationships among Select., 20 (2), 69–79.
proteins with different global topologies 80 Miyazaki, J. et al. (2001) Ancestral
and their implications for function residues stabilizing 3-isopropylmalate
annotation strategies. Proc. Natl. Acad. dehydrogenase of an extreme
Sci. U.S.A., 106 (41), 17377–17382. thermophile: experimental evidence
68 Grishin, N.V. (2001) Fold change in supporting the thermophilic common
evolution of protein structures. ancestor hypothesis. J. Biochem., 129 (5),
J. Struct. Biol., 134 (2–3), 167–185. 777–782.
69 Taylor, W.R. (2007) Evolutionary 81 Watanabe, K. et al. (2006) Designing
transitions in protein fold space. thermostable proteins: ancestral mutants
Curr. Opin. Struct. Biol., 17 (3), 354–361. of 3-isopropylmalate dehydrogenase
70 Krishna, S.S. and Grishin, N.V. (2005) designed by using a phylogenetic tree.
Structural drift: a possible path to protein J. Mol. Biol., 355 (4), 664–674.
fold change. Bioinformatics, 21 (8), 82 Khersonsky, O. et al. (2009) Directed
1308–1310. evolution of serum paraoxonase PON3 by
71 Alexander, P.A. et al. (2007) The design family shuffling and ancestor/consensus
and characterization of two proteins with mutagenesis, and its biochemical
88% sequence identity but different characterization. Biochemistry, 48 (28),
structure and function. Proc. Natl. Acad. 6644–6654.
Sci. U.S.A., 104 (29), 11963–11968. 83 Bershtein, S., Goldin, K., and Tawfik, D.S.
72 Guo, H.H., Choe, J., and Loeb, L.A. (2008) Intense neutral drifts yield robust
(2004) Protein tolerance to random and evolvable consensus proteins.
amino acid change. Proc. Natl. Acad. Sci. J. Mol. Biol., 379 (5), 1029–1044.
U.S.A., 101 (25), 9205–9210. 84 Fodor, A.A. and Aldrich, R.W. (2004)
73 Jauch, R. et al. (2007) Assessment of Influence of conservation on calculations
CASP7 structure predictions for template of amino acid covariance in multiple
free targets. Proteins, 69 (Suppl 8), 57–67. sequence alignments. Proteins, 56 (2),
74 Tress, M.L., Ezkurdia, I., and Richardson, 211–221.
J.S. (2009) Target domain definition and 85 Halperin, I., Wolfson, H., and Nussinov,
classification in CASP8. Proteins, 77 R. (2006) Correlated mutations: advances
(Suppl 9), 10–17. and limitations. A study on fusion
References j111
proteins and on the cohesin-dockerin 97 Eijsink, V.G. et al. (2004)
families. Proteins, 63 (4), 832–845. Rational engineering of enzyme
86 Kuipers, R.K. et al. (2009) Correlated stability. J. Biotechnol., 113 (1–3),
mutation analyses on super-family 105–120.
alignments reveal functionally important 98 Marechal, J.D. et al. (2008) Insights into
residues. Proteins, 76 (3), 608–616. drug metabolism by cytochromes P450
87 Loeb, J. (1921) Chemical and physical from modelling studies of CYP2D6-drug
behavior of casein solutions. interactions. Br. J. Pharmacol., 153
J. Gen. Physiol., 3 (4), 547–555. (Suppl 1), S82–S89.
88 Lawrence, M.S., Phillips, K.J., and 99 Sousa, S.F., Fernandes, P.A., and
Liu, D.R. (2007) Supercharging proteins Ramos, M.J. (2006) Protein-ligand
can impart unusual resilience. docking: current status and future
J. Am. Chem. Soc., 129 (33), 10110–10112. challenges. Proteins, 65 (1), 15–26.
89 Chiti, F. et al. (2003) Rationalization of 100 Xu, Y. et al. (2008) Induced-fit or
the effects of mutations on peptide and preexisting equilibrium dynamics?
protein aggregation rates. Nature, Lessons from protein crystallography
424 (6950), 805–808. and MD simulations on
90 Tartaglia, G.G. et al. (2008) Prediction acetylcholinesterase and implications for
of aggregation-prone regions in structure-based drug design. Protein Sci.,
structured proteins. J. Mol. Biol., 380 (2), 17 (4), 601–605.
425–436. 101 Ma, B. et al. (2002) Multiple diverse
91 Buell, A.K. et al. (2009) Position- ligands binding at a single protein site:
dependent electrostatic protection a matter of pre-existing populations.
against protein aggregation. Protein Sci., 11 (2), 184–197.
ChemBioChem, 10 (8), 1309–1312. 102 Abel, R. et al. (2008) Role of the active-site
92 Robinson-Rechavi, M., Alibes, A., and solvent in the thermodynamics of factor
Godzik, A. (2006) Contribution of Xa ligand binding. J. Am. Chem. Soc.,
electrostatic interactions, compactness 130 (9), 2817–2831.
and quaternary structure to protein 103 Kapoli, P. et al. (2008) Engineering
thermostability: lessons from structural sensitive glutathione transferase for the
genomics of Thermotoga maritima. detection of xenobiotics. Biosens.
J. Mol. Biol., 356 (2), 547–557. Bioelectron., 24 (3), 498–503.
93 Perl, D. and Schmid, F.X. (2001) 104 Miura, S. et al. (2008) Development of
Electrostatic stabilization of a fructosyl amine oxidase specific to
thermophilic cold shock protein. fructosyl valine by site-directed
J. Mol. Biol., 313 (2), 343–357. mutagenesis. Protein Eng. Des. Sel., 21 (4),
94 Schweiker, K.L. et al. (2007) 233–239.
Computational design of the Fyn SH3 105 Zhu, D. et al. (2008) Inverting the
domain with increased stability through enantioselectivity of a carbonyl reductase
optimization of surface charge charge via substrate-enzyme docking-guided
interactions. Protein Sci., 16 (12), point mutation. Org. Lett., 10 (4),
2694–2702. 525–528.
95 Goldstein, R.A. (2007) Amino-acid 106 Host, G.E. and Jonsson, B.H. (2008)
interactions in psychrophiles, Converting human carbonic anhydrase II
mesophiles, thermophiles, and into a benzoate ester hydrolase through
hyperthermophiles: insights from the rational redesign. Biochim. Biophys. Acta,
quasi-chemical approximation. Protein 1784 (5), 811–815.
Sci., 16 (9), 1887–1895. 107 Lee, S.C. et al. (2009) Designing the
96 Braiuca, P. et al. (2007) Volsurf substrate specificity of D-hydantoinase
computational method applied to using a rational approach. Enzyme
the prediction of stability of Microb. Technol., 44 (3), 170–175.
thermostable enzymes. Biotechnol. J., 108 Adcock, S.A. and McCammon, J.A. (2006)
2 (2), 214–220. Molecular dynamics: survey of methods
112
for simulating the activity of proteins. 119 Christen, M. et al. (2005) The GROMOS
Chem. Rev., 106 (5), 1589–1615. software for biomolecular simulation:
109 Christen, M. and van Gunsteren, W.F. GROMOS05. J. Comput. Chem., 26 (16),
(2008) On searching in, sampling of, and 1719–1751.
dynamically moving through 120 Van Der Spoel, D. et al. (2005)
conformational space of biomolecular GROMACS: fast, flexible, and free.
systems: a review. J. Comput. Chem., J. Comput. Chem., 26 (16),
29 (2), 157–166. 1701–1718.
110 Karplus, M. and Kuriyan, J. (2005) 121 Phillips, J.C. et al. (2005) Scalable
Molecular dynamics and protein molecular dynamics with NAMD.
function. Proc. Natl. Acad. Sci. U.S.A., J. Comput. Chem., 26 (16), 1781–1802.
102 (19), 6679–6685. 122 Cieplak, P., Caldwell, J., and Kollman, P.
111 Arkin, I.T. et al. (2007) Mechanism of (2001) Molecular mechanical models for
Na þ /H þ antiporting. Science, organic and biological systems going
317 (5839), 799–803. beyond the atom centered two body
112 Voelz, V.A. et al. (2009) Probing the additive approximation: aqueous solution
nanosecond dynamics of a designed free energies of methanol and N-methyl
three-stranded beta-sheet with a acetamide, nucleic acid base, and amide
massively parallel molecular dynamics hydrogen bonding and chloroform/water
simulation. Int. J. Mol. Sci., 10 (3), partition coefficients of the nucleic acid
1013–1030. bases. J. Comput. Chem., 22 (10),
113 Beck, D.A., White, G.W., and Daggett, V. 1048–1057.
(2007) Exploring the energy landscape 123 Harder, E. et al. (2005) Efficient
of protein folding using replica-exchange simulation method for polarizable
and conventional molecular dynamics protein force fields: application to the
simulations. J. Struct. Biol., 157 (3), simulation of BPTI in liquid. J. Chem.
514–523. Theory Comp., 1 (1), 169–180.
114 Maragakis, P. et al. (2008) Microsecond 124 Xie, W. et al. (2009) X-Pol potential: an
molecular dynamics simulation shows electronic structure-based force field for
effect of slow loop dynamics on backbone molecular dynamics simulation of a
amide order parameters of proteins. solvated protein in water. J. Chem. Theory
J. Phys. Chem. B, 112 (19), 6155–6158. Comp., 5 (3), 459–467.
115 Case, D.A. et al. (2005) The Amber 125 Morra, G. and Colombo, G. (2008)
biomolecular simulation programs. Relationship between energy distribution
J. Comput. Chem., 26 (16), and fold stability: insights from molecular
1668–1688. dynamics simulations of native and
116 Jorgensen, W.L. and Tirado-Rives, J. mutant proteins. Proteins, 72 (2),
(2005) Molecular modeling of organic 660–672.
and biomolecular systems using BOSS 126 Missimer, J.H. et al. (2007)
and MCPRO. J. Comput. Chem., 26 (16), Configurational entropy elucidates the
1689–1700. role of salt-bridge networks in protein
117 Kevin, J.B. et al. (2006) Scalable thermostability. Protein Sci., 16 (7),
algorithms for molecular dynamics 1349–1359.
simulations on commodity clusters, SC 127 Pasi, M. et al. (2009) Dynamic properties
06Proceedings of the 2006 ACM/IEEE of a psychrophilic alpha-amylase in
Conference on Supercomputing, CD- comparison with a mesophilic
ROM, Article 84, Association for homologue. J. Phys. Chem. B, 113 (41),
Computing Machinery, New York. 13585–13595.
118 Zagrovic, B. et al. (2002) Simulation of 128 Gatti-Lafranconi, P. et al. (2008)
folding of a small alpha-helical protein in Unscrambling thermal stability and
atomistic detail using worldwide- temperature adaptation in evolved
distributed computing. J. Mol. Biol., variants of a cold-active lipase. FEBS Lett.,
323 (5), 927–937. 582 (15), 2313–2318.
References j113
129 Bocola, M. et al. (2004) Learning from 1A1: enzyme-substrate interactions and
directed evolution: theoretical substrate binding affinities. J. Biomol.
investigations into cooperative mutations Struct. Dyn., 20 (2), 155–162.
in lipase enantioselectivity. 140 Stone, M.J. (2001) NMR relaxation
ChemBioChem, 5 (2), 214–223. studies of the role of conformational
130 Henke, E. et al. (2003) A molecular entropy in protein stability and ligand
mechanism of enantiorecognition of binding. Acc. Chem. Res., 34 (5), 379–388.
tertiary alcohols by carboxylesterases. 141 Zidek, L., Novotny, M.V., and Stone, M.J.
ChemBioChem, 4 (6), 485–493. (1999) Increased protein backbone
131 Holzwarth, H.C., Pleiss, J., and Schmid, conformational entropy upon
R.D. (1997) Computer-aided modelling of hydrophobic ligand binding. Nat. Struct.
stereoselective triglyceride hydrolysis Biol., 6 (12), 1118–1121.
catalyzed by Rhizopus oryzae lipase. 142 Toth, K. et al. (2009) An examination of
J. Mol. Catal. B: Enzym., 3 (1–4), 73–82. the relationship between active site loop
132 Scheib, H. et al. (1998) Rational design of size and thermodynamic activation
Rhizopus oryzae lipase with modified parameters for orotidine 50 -
stereoselectivity toward triradylglycerols. monophosphate decarboxylase from
Protein Eng., 11 (8), 675–682. mesophilic and thermophilic organisms.
133 Schulz, T., Pleiss, J., and Schmid, R.D. Biochemistry, 48 (33), 8006–8013.
(2000) Stereoselectivity of Pseudomonas 143 Ottosson, J., Fransson, L., and Hult, K.
cepacia lipase toward secondary alcohols: (2002) Substrate entropy in enzyme
a quantitative model. Protein Sci., 9 (6), enantioselectivity: an experimental and
1053–1062. molecular modeling study of a lipase.
134 Branco, R.J. et al. (2008) Anchoring Protein Sci., 11 (6), 1462–1471.
effects in a wide binding pocket: the 144 Ottosson, J. et al. (2001) Rational design of
molecular basis of regioselectivity in enantioselective enzymes requires
engineered cytochrome P450 considerations of entropy. Protein Sci.,
monooxygenase from B. megaterium. 10 (9), 1769–1774.
Proteins, 73 (3), 597–607. 145 Liu, T., Whitten, S.T., and Hilser, V.J.
135 Stjernschantz, E. et al. (2008) Structural (2007) Functional residues serve a
rationalization of novel drug dominant role in mediating the
metabolizing mutants of cytochrome cooperativity of the protein ensemble.
P450 BM3. Proteins, 71 (1), 336–352. Proc. Natl. Acad. Sci. U.S.A., 104 (11),
136 Oelschlaeger, P., Schmid, R.D., and 4347–4352.
Pleiss, J. (2003) Modeling domino effects 146 Hammes-Schiffer, S. and Benkovic, S.J.
in enzymes: molecular basis of the (2006) Relating protein motion to
substrate specificity of the bacterial catalysis. Annu. Rev. Biochem., 75,
metallo-beta-lactamases IMP-1 and IMP- 519–541.
6. Biochemistry, 42 (30), 8945–8956. 147 Siegbahn, P.E. and Borowski, T. (2006)
137 Vistoli, G., Pedretti, A., Mazzolari, A., and Modeling enzymatic reactions involving
Testa, B. (2010) In silico prediction of transition metals. Acc. Chem. Res., 39 (10),
human carboxylesterase-1 (hCES1) 729–738.
metabolism combining docking analyses 148 Warshel, A. et al. (2006) Electrostatic basis
and MD simulations. Bioorg. Med. Chem., for enzyme catalysis. Chem. Rev., 106 (8),
18, 320–329. 3210–3235.
138 Bikadi, Z. and Hazai, E. (2008) In silico 149 Mata, R.A. et al. (2008) Toward accurate
description of differential barriers for enzymatic reactions: QM/
enantioselectivity in methoxychlor MM case study on p-hydroxybenzoate
O-demethylation by CYP2C enzymes. hydroxylase. J. Chem. Phys., 128 (2),
Biochim. Biophys. Acta, 1780 (9), 025104.
1070–1079. 150 Warshel, A. and Levitt, M. (1976)
139 Szklarz, G.D. and Paulsen, M.D. (2002) Theoretical studies of enzymic reactions:
Molecular modeling of cytochrome P450 dielectric, electrostatic and steric
114
stabilization of the carbonium ion in the hydrophobic organic media. Biotechnol.

reaction of lysozyme. J. Mol. Biol., 103 (2), Bioeng., 72 (5), 523–529.
227–249. 162 Alston, M.J. and Freedman, R.B. (2002)
151 Cohen, S., Kumar, D., and Shaik, S. The water-dependence of the catalytic
(2006) In silico design of a mutant of activity of bilirubin oxidase suspensions
cytochrome P450 containing in low-water systems. Biotechnol. Bioeng.,
selenocysteine. J. Am. Chem. Soc., 128 (8), 77 (6), 651–657.
2649–2653. 163 Adlercreutz, P. (1993) Activation of
152 van der Kamp, M.W. and Mulholland, A.J. enzymes in organic media at low water
(2008) Computational enzymology: activity by polyols and saccharides.
insight into biological catalysts from Biochim. Biophys. Acta, 1163 (2),
modelling. Nat. Prod. Rep., 25 (6), 144–148.
1001–1014. 164 Leonard, V. et al. (2007) A water molecule
153 Bos, F. and Pleiss, J. (2008) Conserved in the stereospecificity pocket of Candida
water molecules stabilize the Omega-loop antarctica lipase B enhances
in class A beta-lactamases. Antimicrob. enantioselectivity towards pentan-2-ol.
Agents Chemother., 52 (3), 1072–1079. ChemBioChem, 8 (6), 662–667.
154 Bottoms, C.A., White, T.A., and Tanner, 165 Barzana, E., Klibanov, A.M., and Karel, M.
J.J. (2006) Exploring structurally (1987) Enzyme-catalyzed, gas-phase
conserved solvent sites in protein reactions. Appl. Biochem. Biotechnol.,
families. Proteins, 64 (2), 404–421. 15 (1), 25–34.
155 Nakasako, M. (2004) Water-protein 166 Yang, F. and Russell, A.J. (1996) The role
interactions from high-resolution protein of hydration in enzyme activity and
crystallography. Philos. Trans. R. Soc. stability: 2. Alcohol dehydrogenase
London, Ser. B, 359 (1448), 1191–1204; activity and stability in a continuous gas
discussion 1204–1206. phase reactor. Biotechnol. Bioeng., 49 (6),
156 Smolin, N. et al. (2005) Properties of 709–716.
spanning water networks at protein 167 Graber, M. et al. (2003) Water plays a
surfaces. J. Phys. Chem. B, 109 (21), different role on activation
10995–11005. thermodynamic parameters of
157 Zaks, A. and Klibanov, A.M. (1988) The alcoholysis reaction catalyzed by lipase in
effect of water on enzyme action in gaseous and organic media. Biochim.
organic media. J. Biol. Chem., 263 (17), Biophys. Acta, 1645 (1), 56–62.
8017–8021. 168 Branco, R.J. et al. (2009) Molecular
158 Griebenow, K. and Klibanov, A.M. (1996) mechanism of the hydration of Candida
On protein denaturation in aqueous- antarctica lipase B in the gas phase: water
organic mixtures but not in pure organic adsorption isotherms and molecular
solvents. J. Am. Chem. Soc., 118 (47), dynamics simulations. ChemBioChem,
11695–11700. 10 (18), 2913–2919.
159 Degn, H., Bohatka, S., and Lloyd, D. 169 Graber, M. et al. (2007) Solvent as a
(1992) Enzyme activity in organic solvent competitive inhibitor for Candida
as a function of water activity determined antarctica lipase B. Biochim. Biophys. Acta,
by membrane inlet mass spectrometry. 1774 (8), 1052–1057.
Biotechnol. Tech., 6 (2), 161–164. 170 Trodler, P. and Pleiss, J. (2008) Modeling
160 Valivety, R.H. et al. (1994) Relationship structure and flexibility of Candida
between water activity and catalytic antarctica lipase B in organic solvents.
activity of lipases in organic media – BMC Struct. Biol., 8, 9.
effects of supports. Loading and enzyme 171 Soares, C.M., Teixeira, V.H., and Baptista,
preparation. Eur. J. Biochem., 222 (2), A.M. (2003) Protein structure and
461–466. dynamics in nonaqueous solvents:
161 van de Velde, F. et al. (2001) insights from molecular dynamics
Chloroperoxidase-catalyzed simulation studies. Biophys. J., 84 (3),
enantioselective oxidations in 1628–1641.
References j115
172 Ottosson, J. et al. (2002) Size as a 183 Jones, D.T., Miller, R.T., and
parameter for solvent effects on Candida Thornton, J.M. (1995) Successful protein
antarctica lipase B enantioselectivity. fold recognition by optimal sequence
Biochim. Biophys. Acta, 1594 (2), 325–334. threading validated by rigorous blind
173 Tejo, B.A., Salleh, A.B., and Pleiss, J. testing. Proteins, 23 (3), 387–397.
(2004) Structure and dynamics of 184 Sippl, M.J. (1995) Knowledge-based
Candida rugosa lipase: the role of potentials for proteins. Curr. Opin. Struct.
organic solvent. J. Mol. Model., 10 (5–6), Biol., 5 (2), 229–235.
358–366. 185 Ensign, D.L. and Pande, V.S. (2009) The
174 Beck, D.A. et al. (2007) Simulations of Fip35 WW domain folds with structural
macromolecules in protective and and mechanistic heterogeneity in
denaturing osmolytes: properties of molecular dynamics simulations.
mixed solvent systems and their effects Biophys. J., 96 (8), L53–L55.
on water and protein structure and 186 Lei, H. et al. (2007) Folding free-energy
dynamics. Methods Enzymol., 428, landscape of villin headpiece subdomain
373–396. from molecular dynamics simulations.
175 Peters, G.H. et al. (1996) Dynamics of Proc. Natl. Acad. Sci. U.S.A., 104 (12),
proteins in different solvent systems: 4925–4930.
analysis of essential motion in lipases. 187 Snow, C.D., Zagrovic, B., and Pande, V.S.
Biophys. J., 71 (5), 2245–2255. (2002) The Trp cage: folding kinetics and
176 Peters, G.H. and Bywater, R.P. (1999) unfolded state topology via molecular
Computational analysis of chain dynamics simulations. J. Am. Chem. Soc.,
flexibility and fluctuations in Rhizomucor 124 (49), 14548–14549.
miehei lipase. Protein Eng., 12 (9), 188 Dahiyat, B.I. and Mayo, S.L. (1996)
747–754. Protein design automation. Protein Sci.,
177 Trodler, P., Schmid, R.D., and Pleiss, J. 5 (5), 895–903.
(2009) Modeling of solvent-dependent 189 Malakauskas, S.M. and Mayo, S.L. (1998)
conformational transitions in Design, structure and stability of a
Burkholderia cepacia lipase. BMC Struct. hyperthermophilic protein variant. Nat.
Biol., 9, 38. Struct. Biol., 5 (6), 470–475.
178 Barbe, S. et al. (2009) Insights into lid 190 Simons, K.T. et al. (1999) Ab initio protein
movements of Burkholderia cepacia structure prediction of CASP III targets
lipase inferred from molecular dynamics using ROSETTA. Proteins (Suppl 3), 171–
simulations. Proteins, 77 (3), 509–523. 176.
179 Santini, S. et al. (2009) Study of 191 Kuhlman, B. et al. (2003) Design of a
Thermomyces lanuginosa lipase in the novel globular protein fold with atomic-
presence of tributyrylglycerol and water. level accuracy. Science, 302 (5649), 1364–
Biophys. J., 96 (12), 4814–4825. 1368.
180 Sutcliffe, M.J., Hayes, F.R., and 192 Gerlt, J.A. and Babbitt, P.C. (2009)
Blundell, T.L. (1987) Knowledge based Enzyme (re)design: lessons from natural
modelling of homologous proteins, Part evolution and computation. Curr. Opin.
II: rules for the conformations of Chem. Biol., 13 (1), 10–18.
substituted sidechains. Protein Eng., 1 (5), 193 Jensen, R.A. (1976) Enzyme recruitment
385–392. in evolution of new function. Annu. Rev.
181 Lim, W.A. and Sauer, R.T. (1989) Microbiol., 30, 409–425.
Alternative packing arrangements in the 194 OBrien, P.J. and Herschlag, D. (1999)
hydrophobic core of lambda repressor. Catalytic promiscuity and the evolution of
Nature, 339 (6219), 31–36. new enzymatic activities. Chem. Biol.,
182 Akanuma, S., Kigawa, T., and Yokoyama, 6 (4), R91–R95.
S. (2002) Combinatorial mutagenesis to 195 Nobeli, I., Favia, A.D., and Thornton, J.M.
restrict amino acid usage in an enzyme to (2009) Protein promiscuity and its
a reduced set. Proc. Natl. Acad. Sci. implications for biotechnology. Nat.
U.S.A., 99 (21), 13549–13553. Biotechnol., 27 (2), 157–167.
116
196 Khersonsky, O., Roodveldt, C., and 210 Freddolino, P.L. et al. (2009) Force field
Tawfik, D.S. (2006) Enzyme promiscuity: bias in protein folding simulations.
evolutionary and mechanistic aspects. Biophys. J., 96 (9), 3772–3780.
Curr. Opin. Chem. Biol., 10 (5), 498–508. 211 Klepeis, J.L. et al. (2009) Long-timescale
197 Quemeneur, E. et al. (1998) Engineering molecular dynamics simulations of
cyclophilin into a proline-specific protein structure and function.
endopeptidase. Nature, 391 (6664), Curr. Opin. Struct. Biol., 19 (2), 120–127.
301–304. 212 McCammon, J.A., Gelin, B.R., and
198 Branneby, C. et al. (2003) Carbon-carbon Karplus, M. (1977) Dynamics of
bonds by hydrolytic enzymes. J. Am. folded proteins. Nature, 267 (5612),
Chem. Soc., 125 (4), 874–875. 585–590.
199 Carlqvist, P. et al. (2003) Rational 213 Kubiak, K. and Mulheran, P.A. (2009)
design of a lipase to accommodate Molecular dynamics simulations of hen
catalysis of Baeyer-Villiger oxidation with egg white lysozyme adsorption at a
hydrogen peroxide. J. Mol. Model., 9 (3), charged solid surface. J. Phys. Chem. B,
164–171. 113 (36), 12189–12200.
200 Svedendahl, M., Hult, K., and Berglund, 214 Kony, D.B., Hunenberger, P.H., and
P. (2005) Fast carbon-carbon bond van Gunsteren, W.F. (2007) Molecular
formation by a promiscuous lipase. J. Am. dynamics simulations of the native and
Chem. Soc., 127 (51), 17988–17989. partially folded states of ubiquitin:
201 Svedendahl, M. et al. (2008) Direct influence of methanol cosolvent, pH, and
epoxidation in Candida antarctica lipase B temperature on the protein structure
studied by experiment and theory. and dynamics. Protein Sci., 16 (6),
ChemBioChem, 9 (15), 2443–2451. 1101–1118.
202 Jochens, H. et al. (2009) Converting an 215 Garcia-Viloca, M. et al. (2004) How
esterase into an epoxide hydrolase. enzymes work: analysis by modern rate
Angew. Chem. Int. Ed., 48 (19), 3532–3535. theory and computer simulations.
203 Bornscheuer, U.T. and Kazlauskas, R.J. Science, 303 (5655), 186–195.
(2004) Catalytic promiscuity in 216 Velonia, K. et al. (2005) Single-enzyme
biocatalysis: using old enzymes to kinetics of CALB-catalyzed hydrolysis.
form new bonds and follow new Angew. Chem. Int. Ed., 44 (4), 560–564.
pathways. Angew. Chem. Int. Ed., 43 (45), 217 Mantle, T.J. (2001) Enzymes: natures
6032–6040. nanomachines. Royal Irish Academy
204 Kazlauskas, R.J. (2005) Enhancing Medal Lecture. Biochem. Soc. Trans., 29 (Pt
catalytic promiscuity for biocatalysis. 2), 331–336.
Curr. Opin. Chem. Biol., 9 (2), 195–201. 218 Soong, R.K. et al. (2000) Powering an
205 Jiang, L. et al. (2008) De novo inorganic nanodevice with a
computational design of retro-aldol biomolecular motor. Science, 290 (5496),
enzymes. Science, 319 (5868), 1387–1391. 1555–1558.
206 Rothlisberger, D. et al. (2008) Kemp 219 Dubey, A. et al. (2004) Computational
elimination catalysts by computational studies of viral protein nano-actuators.
enzyme design. Nature, 453 (7192), J. Comput. Theor. Nanosci., 1 (1), 18–28.
190–195. 220 Chaloupkova, R. et al. (2003) Modification
207 Holliday, G.L. et al. (2007) The chemistry of activity and specificity of haloalkane
of protein catalysis. J. Mol. Biol., 372 (5), dehalogenase from Sphingomonas
1261–1277. paucimobilis UT26 by engineering of its
208 Rueda, M. et al. (2007) A consensus view entrance tunnel. J. Biol. Chem., 278 (52),
of protein dynamics. Proc. Natl. Acad. Sci. 52622–52628.
U.S.A., 104 (3), 796–801. 221 Guieysse, D. et al. (2008) A structure-
209 Hu, Z. and Jiang, J. (2009) Assessment of controlled investigation of lipase
biomolecular force fields for molecular enantioselectivity by a path-planning
dynamics simulations in a protein crystal. approach. ChemBioChem, 9 (8),
J. Comput. Chem. 31 (2), 371–380. 1308–1317.
References j117
222 Schliessmann, A. et al. (2009) Increased atomic resolution. Science, 287 (5458),
enantioselectivity by engineering 1615–1622.
bottleneck mutants in an esterase from 230 Wang, L. et al. (2001) Expanding the
Pseudomonas fluorescens. genetic code of Escherichia coli. Science,
ChemBioChem. 10 (18), 2809–2951. 292 (5516), 498–500.
223 Li, H.M. et al. (2008) Cytochrome P450 231 Stromgaard, A., Jensen, A.A., and
BM-3 evolved by random and saturation Stromgaard, K. (2004) Site-specific
mutagenesis as an effective indole- incorporation of unnatural amino acids
hydroxylating catalyst. Appl. Biochem. into proteins. ChemBioChem, 5 (7),
Biotechnol., 144 (1), 27–36. 909–916.
224 Smith, A.J. et al. (2008) Structural 232 Gilmore, M.A., Steward, L.E., and
reorganization and preorganization in Chamberlin, A.R. (1999) Incorporation of
enzyme active sites: comparisons of noncoded amino acids by in vitro protein
experimental and theoretically ideal biosynthesis, in Implementation and
active site geometries in the multistep Redesign of Catalytic Function in
serine esterase reaction cycle. J. Am. Biopolymers, Topics in Current Chemistry,
Chem. Soc., 130 (46), 15361–15373. vol. 202, Springer, pp. 77–99.
225 Benkovic, S.J., Hammes, G.G., and 233 Creus, M. and Ward, T.R. (2007)
Hammes-Schiffer, S. (2008) Free-energy Designed evolution of artificial
landscape of enzyme catalysis. metalloenzymes: protein catalysts made
Biochemistry, 47 (11), 3317–3321. to order. Org. Biomol. Chem., 5 (12),
226 Fierke, C.A., Johnson, K.A., and 1835–1844.
Benkovic, S.J. (1987) Construction and 234 Jing, Q., Okrasa, K., and Kazlauskas, R.J.
evaluation of the kinetic scheme (2009) Stereoselective hydrogenation of
associated with dihydrofolate reductase olefins using rhodium-substituted
from Escherichia coli. Biochemistry, 26 (13), carbonic anhydrase-a new reductase.
4085–4092. Chem.-Eur. J., 15 (6), 1370–1376.
227 Yanagida, T. (2008) Fluctuation as a tool of 235 Reetz, M.T. et al. (2008) A robust protein
biological molecular machines. host for anchoring chelating ligands and
Biosystems, 93 (1–2), 3–7. organocatalysts. ChemBioChem, 9 (4),
228 Wolfenden, R. and Snider, M.J. (2001) 552–564.
The depth of chemical time and the power 236 Ward, T.R. (2008) Artificial enzymes
of enzymes as catalysts. Acc. Chem. Res., made to order: combination of
34 (12), 938–945. computational design and directed
229 Schlichting, I. et al. (2000) The catalytic evolution. Angew. Chem. Int. Ed., 47 (41),
pathway of cytochrome p450cam at 7802–7803.
j119
5
Directed Evolution of Enzymes
Manfred T. Reetz
5.1
Purpose of Directed Evolution
The use of enzymes as catalysts in synthetic organic chemistry and white biotech-
nology has experienced rapid growth during the last three decades [1], yet these
biocatalysts have suffered traditionally from several limitations. These include in
many cases limited substrate scope (rate), poor stereoselectivity, insufficient stability,
and sometimes product inhibition. During the last 15 years, the genetic technique of
directed evolution, which simulates natural evolution in the laboratory (evolution in
the test tube), has been developed to such an extent that all of these problems can be
addressed and solved [2].
5.2
Short History of Directed Evolution
The term directed evolution refers to laboratory evolution, meaning experimental

platforms that mimic or simulate the process of evolution that occurs in nature. The
roots of directed evolution go back to the work of Spiegelmann, who in 1965–1967
performed a Darwinian experiment with a self-duplicating nucleic acid molecule
(RNA) outside a living cell [3]. Although later research showed that Spiegelmanns
RNA molecules are not really self-duplicating in the way that it is defined today, his
contributions mark the beginning of a fruitful area of research, fueled over the last
few decades by Szostak, Joyce, and others [4]. In contrast to Spiegelmann, they had
(have) access to modern molecular biological methods such as the polymerase chain
reaction (PCR), which allow experiments to be carried out that previously were not
possible. Areas of application include selection of RNA aptamers, selection of
catalytic RNA molecules, and evolution of an RNA polymerase ribozyme and of
ribozymes by continuous serial transfer [4].
The directed evolution of proteins, specifically of enzymes, is a different area of
research with totally different applications. It is the subject of this chapter, the focus
and Oliver May.
j 5 Directed Evolution of Enzymes
120
being on the evolution of enzymes as catalysts in organic chemistry and white

biotechnology. The term directed evolution in this field was used as early as 1972 by
Hansche, who evaluated naturally occurring mutants of a phosphatase in a popu-
lation of 109 cells over 1000 generations in a continuous manner [5]. This in vivo
mutagenesis is based on selection pressure used in the experimental setup, which led
to an increase in enzyme activity. Further contributions appeared sporadically, again
most of them being in vivo processes, as in the experiments of Hall [6]. In 1984 Eigen
predicted that it should be possible to construct an evolutionary machine for
producing mutant proteins, a laboratory process consisting of repeated cycles of gene
mutagenesis, amplification, and selection [7]. Around that time researchers were
beginning to apply mutagenesis methods to obtain libraries of mutant enzymes that
were selected or screened for some property, usually thermostability. For example,
using chemically induced mutagenesis (hydroxylamine), the thermal stability of
phage T4 lysozyme was increased, specifically by screening about 30 000 phage
plaques for their ability to withstand incubation at a temperature that inactivates the
wild-type (WT) enzyme [8]. In another case the thermostability of the protease
subtilisin BPN0 from Bacillus amyloliquefaciens was improved by chemical mutagen-
esis using bisulfite, a chemical mutagen that deaminates cytosine to uracil in single-
strand DNA [9]. In 1989 error-prone polymerase chain reaction (epPCR) was reported
[10a], which was improved in 1994 [10b], but applications came later. In one such
study, epPCR was used to increase the thermostability of aspartase, a catalyst in the
industrially important addition of ammonia to fumarate with formation of L-aspartic
acid [11]. However, in this report and in related papers, only initial libraries were
generated. Real directed evolution involves at least two cycles, that is, the gene of a hit
identified in the first round is used as a template for mutagenesis/screening (or
selection) in a subsequent step, and so on, until the desired degree of catalyst
improvement has been achieved. Going through several cycles creates evolutionary
pressure, which stands at the heart of this approach to protein engineering. An early
example of two cycles of mutagenesis/selection concerns the increase in thermo-
stability of kanamycin nucleotidyl transferase by an in vivo process using a mutator
strain [12]. Another step forward in in vitro directed evolution was set in 1993, when
Arnold described the successful attempt to increase the stability of the protease
subtilisin E toward hostile organic solvents by going through several cycles of epPCR
at low error rate [13].
In 1994 Stemmer reported the development of a fundamentally different molec-
ular biological method useful in directed evolution, namely, DNA shuffling, which
simulates sexual evolution in nature (Section 5.4) [14]. The activity of a b-lactamase
was increased by going through several rounds of shuffling. This seminal study set
the stage for the development and application of many other recombinant methods
(Section 5.4.3). Yet another important date is 1997 when the Reetz group published
the first case of directed evolution as a means to control the stereoselectivity of
enzymes, specifically by going through four cycles of epPCR [15] (Section 5.5.2). The
1993–1997 phase was followed by a burst of activities aimed at generalizing directed
evolution in the quest to increase thermostability [16], manipulate substrate scope,
and enhance stereoselectivity [17]. As hundreds of successful studies utilizing
5.3 Basic Principles and Challenges j121
various different enzyme types appeared, using many different mutagenesis meth-
ods, directed evolution emerged as a powerful way to engineer essentially any
property of enzymes [2]. However, even around the year 2000 efficiency was not a
major concern, and many studies had the character of proof-of-principle. As time
went on, this changed because researchers in academia and industry became
interested in developing fast laboratory evolution characterized by high efficacy
(quality, not quantity) [17, 18]. Comparative studies regarding the use of different
strategies for probing protein sequence space began to appear. Indeed, during recent
years methodology development in the quest to make directed evolution more
efficient than in the past has become a pressing issue. This is the main focus of
the present chapter.
5.3
Basic Principles and Challenges
Directed evolution studies begin by applying a gene mutagenesis method, followed

by insertion of the library of mutated genes into an organismic host (e.g., Escherichia
coli) that is then plated out on agar plates. Following a growth period, single colonies
appear, each originating from a single cell, which prevents the formation of
undesired mixtures of mutants. These are individually harvested with a colony
picker (or with toothpicks!), cultured, and subjected to an appropriate screening (or
selection) system [19, 20]. Assays for screening activity or enantioselectivity have been
automated, generally based on 96- or 384-well microtiter plates that are scanned
using some analytical tool such as UV–vis spectroscopy, fluorescence, mass spec-
trometry (MS), or IR spectroscopy (Section 5.5) [20]. Following the identification of an
improved mutant (hit), the corresponding mutated gene is collected and used as a
template for the next cycle (Figure 5.1). In some cases it is possible to utilize selection
in place of screening by constructing an experimental platform that ensures a growth
advantage of the bacterial host due to the presence of an improved enzyme. The
difference between screening [20] and selection [19] is delineated in Section 5.6.
Several reviews covering various aspects of directed evolution have appeared
recently [2].
The major challenge in developing efficient laboratory evolution touches on the so-
called numbers problem, which is directly related to the screening effort. To illustrate
expression
..
insertion . . . .. .. . colony screening
.
. . .
mutagenesis
into host . . . . .. .... . picking
. .. .
library of bacterial colonies bacteria producing visualization of

repeat mutant genes on agar plate mutant enzymes positive mutants
in a test tube grown in nutrient broth
Figure 5.1 Individual steps in directed evolution utilizing screening.

122
this, the following calculation is often cited [2]. The number of enzyme variants N at
the theoretically maximum degree of diversity is traditionally described by the
algorithm given in Eq. (5.1):
N ¼ 19M X !=ðX MÞ!M! ð5:1Þ
where M denotes the total number of amino acid substitutions per enzyme
molecule and X is the total number of residues. When considering an enzyme
composed of 300 residues, for example, 5700 different mutants are possible if
one amino acid is substituted randomly, 16 million are possible in the case of two
simultaneous substitutions, and about 30 billion if three amino acids are exchanged
simultaneously. Thus, even with only a few simultaneous amino acid exchange
events, the corresponding protein sequence space becomes unrealistically vast.
Typically, the size of libraries of enzyme mutants ranges between 103 and 106, but
even such small numbers as 103 transformants may cause screening problems. Two
solutions appear logical: (i) the development of better and faster screening (or
selection) protocols; (ii) the development of more efficient methods for probing
sequence, meaning the production of smaller but higher quality libraries.
5.4
Gene Mutagenesis Methods
Traditionally, gene mutagenesis for various purposes was performed using light,
chemicals, or mutator strains [21], methods that are occasionally still employed
today [2]. With the spectacular advance of molecular biology during the last few
decades, many different gene mutagenesis methods have been published, including
variations or important improvements of previous protocols. It is not always easy for
the experimenter to make the best choice. A method claimed to be superior, as in one
that causes little or no amino acid bias, may actually require considerably greater
experimental effort relative to some other option, which means that the respective
benefits need to be realistically assessed. Yet another aspect to be considered is the
question of how to apply a given mutagenesis method, that is, how to develop a
strategy for probing protein sequence space efficiently (Section 5.5). Currently, it is
impossible to assess precisely the known methods and strategies, because the
number of comparative studies is limited (Section 5.5.2). Nevertheless, trends
regarding practical protocols and applications have emerged.
5.4.1
Whole Gene Methods
To this day the most popular mutagenesis method is epPCR, which, in particular, can
be used when no structural information regarding the enzyme under study is
available [10]. The PCR process is carried out in such a way that mistakes in the
base pairings are made which encode point mutations on the protein level. The error-
rate can be controlled by varying the experimental conditions empirically, for
5.4 Gene Mutagenesis Methods j123
example, by changing the MgCl2 (or MnCl2) concentration, using unbalanced
amounts of nucleotides and/or employing increased concentrations of Taq DNA
polymerase so that on average one, two, or more amino acid exchange events occur.
Accordingly, 2–7 nucleotide substitutions per gene correlate with 1–4 amino acid
exchanges [22]. The mutation rate can also be influenced by the incorporation of
synthetic mutagenic dNTPs such as 8-oxo-dGTP, which are eliminated later in a
subsequent PCR reaction employing natural dNTPs [23]. Alcohol-mediated epPCR
has also been reported, and is claimed to have some practical advantages [24]. Further
mutagenic effects can be achieved by manipulating the fidelity of Taq polymerase (or
other polymerases) by protein engineering, or by using other polymerases displaying
high error-rates such as the Mutazyme polymerase as part of the GeneMorph
Random Mutagenesis Kit of Stratagene [25]. The size of epPCR libraries varies
considerably, depending upon the amount of laboratory work/screening that the
experimenter wishes to invest, with 103–106 clones (transformants) being typical.
Irrespective of the particular variation, epPCR is a shot-gun method that
addresses more or less the whole gene/protein. However, only single bases are
replaced within the triplet codon, which restricts diversity. Indeed, for several reasons
application of Eq. (5.1) (Section 5.3) to assess the diversity of an epPCR library of
mutants is not permissible [26]. First of all, the event of two or even three base-pair
exchanges per codon is highly unlikely on statistical grounds. At best, one nucleotide
of a given codon will be exchanged, thereby leading to just nine (instead of 64
possible) different codons encoding four to seven (instead of 20) different amino
acids. In reality, the number of amino acid exchange events aimed for depends on the
type of the original codon. Silent mutations are more likely for some types of codons
(e.g., CGA coding for arginine) than for other types (e.g., AAC coding for aspar-
agines). In a model calculation, the real number of enzyme variants obtained by
epPCR with one mutation introduced per gene was approximated by analyzing every
single codon of an enzyme (lipase from Bacillus subtilis) composed of 181 amino acids
[26a]. Accordingly, only about one-third of the theoretical number of variants is in fact
accessible. Comparable results were obtained upon analysis of several other
enzymes. Amino acid bias is always a problem and can have a variety of different
causes. For example, in most published studies that use Taq polymerase in MnCl2-
containing buffer, the respective protein preferentially introduced A ! T, T ! A
transversions and A ! G, T ! C transitions, while A ! C and T ! G transversions as
well as G ! A and C ! T transitions occurred at lower frequencies. The frequencies
of transversions G ! T and C ! A proved to be very low, and G ! C and C ! G
transversions hardly ever happened. If this mutational bias is also taken into account,
the calculated library diversity represents only about 20% of the theoretical size
(Eq. (5.1)), with the actual sizes also depending upon the GC content of the target
gene [26a]. Bias can also arise from the exponential nature of the amplification
process [27]. Several statistical analyses of epPCR and of other mutagenesis methods
illuminate further aspects of amino acid bias [26, 28]. Irrespective of the bias issue,
the question regarding the optimal mutation rate also needs to be addressed.
As an alternative to epPCR, error-prone rolling circle amplification (epRCA) has
been described as the simplest random mutagenesis method available [29]. Based
124
on earlier work regarding rolling circle replication of circular DNA molecules in

nature, rolling circle amplification (RCA) was developed in several laboratories, as a
method that delivers linear DNA consisting of tandem repeats of the circular DNA
sequence. epRCA is based on this technique but utilizes varying concentrations of
MnCl2 to introduce errors. Only one RCA step is needed followed by direct
transformation of the host strain, yielding mutants with typically 3–4 mutations
per kilobase. The advantage has been related to the fact that no restriction enzymes,
ligases, specific primers, or special equipment such as a thermal-cycler are needed.
The usual problem of bias remains, however. Use of Ø29 DNA polymerase or mutant
thereof favors C to T and G to A mutations (66%), which is quite different from Taq
polymerase used most often in epPCR [29].
Several approaches to alleviating the bias problem in whole gene mutagenesis have
been reported, an early one being the modification of dNTP substrate ratio in epPCR
[10b]. However, a notable degree of bias persists, and a decrease in yield as a
consequence of specific mutations cannot be prevented [30]. Alcohol-mediated
epPCR influences bias in that preferential replacements of Gs and Cs occur, opposite
to standard epPCR, for example, by using the VentrÒ (exo-)DNA polymerase in the
presence of propanol. Careful adjustment of reaction conditions are necessary,
allowing DNA molecules up to 2.8 kb in length to be amplified, which is longer
than what is achieved with normal epPCR using Taq-polymerase [24]. It was
suggested that the combination of VentrÒ (exo-) and Taq-polymerase should consti-
tute an epPCR variation with decreased bias. Indeed, a similar strategy using a
different combination was reported to be effective, namely, the combination of
libraries generated by Taq-based PCR and those obtained by application of the
Stratagene GeneMorph Kit, which is characterized by a different bias [18, 27].
A similar technique using other polymerases was also reported [31].
Another approach to whole gene randomization that also reduces amino acid bias
is sequence saturation mutagenesis (SeSaM), a method that requires four steps
[26b,32]: (i) generation of a pool of DNA fragments with random length, (ii) addition
of universal/degenerate base(s) at the 30 -termini of the DNA fragment pool,
(iii) elongation of the DNA fragment pool to full-length genes by PCR, and (iv)
replacement of universal/degenerate bases by standard nucleotides (Figure 5.2). It is
independent of the mutational bias of DNA polymerases and has the advantage that
the fragment distribution of a DNA library can be controlled by utilizing different
concentrations of the individual Sp-dNTPaS or a combination thereof. The mutation
rate can be tuned by varying the concentration of NaCl and/or NaOH in the DNA
melting step. Second-generation SeSaM-Tv constitutes an improvement because it
bypasses the requirement of ssDNA template and utilizes Vent(exo-) and Deep Vent
(exo-) polymerases, which increase PCR yields and allow for transversion-enriched
sequences (Figure 5.2) [33].
A very different whole gene mutagenesis method is random insertion and deletion
(RID), which allows for deletion of an arbitrary number of consecutive bases (up to
16) at random positions while enabling insertion of a specific sequence or random
sequences of an arbitrary number into the same position [34]. Figure 5.3 summarizes
the eight-step sequence of events necessary in RID. Random substitutions are
Figure 5.2 SeSaM and SeSaM-Tv þ TdT-catalyzed elongation of DNA fragments

methods [33]. SeSaM-Tv þ consists of four with degenerate base dPTP; step 3: full-length
steps: step 1: generation of a DNA fragment gene synthesis; and step 4: degenerate base
pool with random length distribution; step 2: replacement by standard nt.
possible, as is the introduction of unnatural amino acids, but unwanted secondary

mutations occurring in the PCR process also occur (which is not unusual in any PCR-
based technique). The method was illustrated in the genetic manipulation of the
GFPUV gene by replacing three randomly chosen consecutive bases by the Bg/II
recognition sequence AGATCT and by replacing the randomly selected bases by a
mixture of 20 codons. Following expression in E. coli, six mutants were selected,
several of which showed different fluorescence properties. For example, a yellow
fluorescent protein and an enhanced green fluorescent variant were obtained, neither
126
of which could be generated by epPCR [34]. Other deletion approaches have been
described, claiming to involve somewhat simpler experimental protocols [35].
Yet another way to create insertions, deletions, and point mutations is a procedure
called INSULT (so dubbed because it uses uncomplementary primers), a method that
avoids subcloning and obviates the need for special ultra-competent cells [36].
Several cycles of linear amplification with a thermophilic polymerase are transversed,
with nick repair after each cycle with a thermophilic ligase being ensured. Following
production of multiple single-stranded copies of circular mutation-bearing plasmid
DNA, a generic primer is added followed by the generation of double-stranded
circular DNA with the desired mutations by one or more polymerase reaction cycles.
An alternative method for addressing whole gene mutagenesis is random insertio-
nal–deletional strand exchange (RAISE) mutagenesis, which consists of only three
steps, including gene shuffling [37] (see discussion regarding DNA shuffling in
Section 5.4.3). In the first step the DNA under study is fragmented randomly by
DNase I to be followed in the second step by attachment of several random
nucleotides to the 30 terminus of the fragment using TdT, and finally by the
reconstruction of each fragment with a tail of random nucleotides into full-length
sequence by self-priming PCR (Figure 5.4). The method was applied to TEM
b-lactamase in the quest to evolve mutants with improved activity against the
b-lactam antibiotic ceftazidime. It was discovered that indels and substitutions are
often combined; the replacement of IPNDERD at amino acids 173–179 by KKMRA
consisted of two amino acid deletions and five amino acid substitutions. In the
RAISE-study, novel mutations were discovered showing higher activities than
variants having only point mutations. It was concluded that such mutations are not
accessible by epPCR [37].
All of these advanced methods are characterized by certain advantages relative to
traditional epPCR, although at the expense of additional laboratory effort. Never-
theless, users should be aware of all facets that are delineated here as well as the
possibility of applying higher-tech methods based on the more sophisticated labo-
ratory techniques if so desired. In summary, conventional epPCR is easy to apply, and
3
Figure 5.3 A general scheme of random with CeIV-EDTA complex. Step 5: the linear
insertion/deletion (RID) mutagenesis for the ssDNAs, which have unknown sequences at
construction of a library of mutant genes [34]. both ends, are ligated to the 50 -anchor and the
The procedure consists of eight major steps. 30 -anchor, respectively. Step 6: the DNAs that
Step 1: (1) the fragment obtained by digesting are linked to the two anchors at both ends are
the original gene with EcoRI and HindIII is amplified by PCR. Step 7: the PCR products are
ligated to a linker; (2) the product is then treated with BciVI, leaving several bases from
digested with HindIII to make a linear dsDNA the 50 -anchor, at the 50 -end. The BciVI treatment
with a nick in the antisense chain. Step 2: the also deletes a specific number of bases at the 30 -
gene fragment is cyclized with T4 DNA ligase to end. Step 8: the digested products are treated
make a circular dsDNA with a nick in the with Klenow fragment to make blunt ends and
antisense chain. Step 3: the circular dsDNA is cyclized again with T4 DNA ligase. The products
treated with T4 DNA polymerase to produce a are treated with EcoRI and HindIII, and the
circular ssDNA. Step 4: the circular ssDNA is fragments are cloned into an EcoRI-HindIII site
randomly cleaved at single positions by treating of modified pUC18 (pUM).
128
Figure 5.4 Schematic diagram of the RAISE method [37].
it is certainly the method of choice when no structural information regarding the

enzyme under study is available (no direct X-ray structure nor homology model).
Choosing the mutation rate and deciding on the number of cycles constitute a
strategic question (Section 5.5). Notably, the process of several cycles of epPCR
provides point mutations apparently necessary for the evolutionary process, but also
leads to the accumulation of superfluous amino acid substitutions that cause
unnecessary screening work and may even result in stability deterioration.
5.4.2
Saturation Mutagenesis
Saturation mutagenesis, sometimes also called cassette mutagenesis, combinatorial

saturation mutagenesis, or site saturation mutagenesis, involves amino acid ran-
domization at a predetermined position or site in the enzyme with the introduction of
all of the other 19 canonical amino acids thereat leading to focused libraries [2]. This is
quite different from blind directed evolution based on epPCR because it requires
some structural knowledge to make the right decision as to where in the enzyme
amino acid randomization should take place, which in turn depends upon the protein
property under study. In the simplest case a single residue is targeted, but it is also
possible to randomize sites composed of two or more amino acid positions. Since in
the latter process the individual point mutations at a given site may influence
not only each other but also interact with other mutations introduced earlier or
subsequently, two different types of cooperative effects become possible, not just
additivity [2b] (Section 5.8.2.2).
Many methods for saturation mutagenesis have been developed since the mid-
1980s, usually as variations of the general approach based on the use of appropriate
oligodeoxynucleotides. Accordingly, primers need to be designed and prepared that
carry the genetic information encoding the desired mutational changes. The most
widely used procedure is the so-called QuikChangeÒ protocol of Stratagene [38],
which is based on previous work [39]. It was originally developed for traditional site-
specific introduction of a given amino acid, but can also be used for randomization at
a defined residue or at up to five amino acid positions in the enzyme (Stratagenes
QuikChange Multi Site-Directed Mutagenesis Kit). Figure 5.5 illustrates the Quik-
Change protocol. Two complementary oligonucleotide primers flanking the desired
mutated nucleotide on both the sense and anti-sense strands are required, and each
primer needs to contain one to several base-pair changes in the region of interest. User-
friendly computer aids for designing libraries are available, generally free of charge,
for example, CASTER [40], B-FITTER [40], GLUE [41], GLUE-IT/PEDEL-AA [42],
and MAX [43]. Success in generating a given library can be checked experimentally by
isolating and sequencing as many mutant genes as possible, and also by analyzing the
degree of amino acid bias. Alternatively for this purpose is the proposed use of the
computer aid SiteFind, which is a tool for introducing a restriction site as a marker for
successful site-directed mutagenesis [44]. However, since the cost of sequencing has
decreased dramatically in recent years, the empirical approach constitutes the more
reliable quality check in most cases [2b].
The extended QuikChange protocol cannot be used in all cases because problems
related to the primer length and design may occur. Consequently, various improve-
ments were proposed, for example, the use of partially overlapping or even non-
overlapping oligonucleotides, with the resulting amplicon being employed as a
Figure 5.5 Schematic illustration of saturation mutagenesis using QuikChange (Stratagene) [38].
130
megaprimer that completes the synthesis of the plasmid in a second PCR [45].
Despite these improvements, difficulties were still encountered in the case of
recalcitrant targets such as large plasmids, as in the case of P450-BM3 from Bacillus
megaterium [46]. Extending the idea of using non-overlapping oligonucleotides [47], a
highly improved two-stage PCR-based method for the creation of saturation muta-
genesis libraries was developed recently, specifically for difficult-to-amplify templates
(Figure 5.6) [46]. In the first stage, both the mutagenic primer and the anti-primer that
are not complementary anneal to the template. The amplified sequence is then used
in the second stage as a megaprimer. In this straightforward process, sites composed
of one or more residues can be randomized in a single PCR reaction, irrespective of
Figure 5.6 Improved method for PCR-based the amplified sequence is used as a megaprimer
saturation mutagenesis that is useful in the case in the second stage. Finally, the template
of difficult-to-amplify templates [46]. The gene is plasmids are digested using DpnI and the
represented by the dotted section, the vector resulting library is transformed in bacteria.
backbone is shown in light gray, and the formed The scheme on the left illustrates the three
megaprimer in black. In the first stage of the possible options in the choice of the
PCR both the mutagenic primer (positions megaprimer size for a single site randomization
randomized represented by a white square) and experiment. The scheme to the right represents
the anti-primer (or another mutagenic primer, an experiment with two sites simultaneously
shown to the right) anneal to the template and randomized.
their location in the gene sequence. In a carefully performed comparative investi-
gation, the virtues of the new method relative to QuikChange and related protocols
were demonstrated using four different enzymes [46].
Notably, some degree of amino acid bias occurs with these saturation mutagenesis
methods. When applying them, the question of oversampling should not be
neglected [40, 42, 48–50]. Fortunately, algorithms have been developed that are
useful in the design of libraries, assuming the absence of amino acid bias [42, 50].
They have been applied in the form of the computer aids CASTER [40] (stereo-
selectivity) and B-FITTER [40] (thermostability) for assessing the degree of over-
sampling as a function of the nature of the codon degeneracy and of the %-coverage of
the respective protein sequence space. When designing saturation mutagenesis
libraries, the following derivations are useful in a practical way [49].
The algorithm [50a] for estimating completeness as a function of the number of
transformants (clones) actually screened, T, can be transformed into Eq. (5.2), where
Pi denotes the probability that a particular sequence occurs in the library, and Fi is the
frequency [49a]:
T ¼ lnð1Pi Þ=Fi ð5:2Þ
Upon substituting for Fi, the relationship reduces to Eq. (5.3), where V is the
number of gene mutants comprising a given library:
T ¼ V ln ð1Pi Þ ð5:3Þ
which defines the correlation between the number of mutants V of a given library and
the number of transformants T that have to be screened for a specified degree of
completeness. The oversampling factor (Of), defined by Eq. (5.4), provides the
experimenter with a useful parameter when designing saturation mutagenesis
libraries [49a]:
Of ¼ T=V ¼ lnð1Pi Þ ð5:4Þ
Upon calculating Of as a function of the library %-coverage, the curve shown in

Figure 5.7 results [49a]. It can be seen that to ensure 95% coverage, for example, the
10
9
Oversampling factor Of
8
7
6
5
4
3
2
1
0
60 70 80 90 100
Coverage [%]
Figure 5.7 Correlation between library coverage and oversampling factor (Of ) [49a].
132
oversampling factor Of amounts to about 3, which means that a threefold excess of

transformants needs to be screened. Owing to the exponential character of the
relationship, degrees of coverage beyond 95% require vastly higher screening efforts.
Lower degrees of library coverage may suffice in a given experiment, but decisions
regarding this important aspect should be viewed in the light of Figure 5.7.
To reduce amino acid bias, hand-mixing of designed oligonucleotides [51] or the
process of Sloning [52] can be applied, although these techniques are labor-intensive
and therefore relatively expensive. In most saturation mutagenesis experiments,
NNK codon degeneracy is employed (N: adenine/cytosine/guanine/thymine; K:
guanine/thymine), encoding all 20 canonical amino acids. However, it has been
shown that the use of a reduced amino acid alphabet can have dramatic advantages,
for example, when employing NDT codon degeneracy (D: adenine/guanine/thy-
mine, T: thymine) encoding 12 amino acids (Phe, Leu, Ile, Val, Tyr, His, Asn, Asp,
Cys, Arg, Ser, and Gly) [49]. When performing simultaneous randomization at sites
composed of more than one amino acid position, the potential screening effort is very
different for the NNK versus NDTsystems. For example, for 95% coverage in the case
of a site composed of three amino acid positions, NNK requires almost 100 000
clones, whereas NDT needs only about 5000. In a comparative study in which
randomization was performed at a three-residue site, NNK and NDT codons were
employed [49]. Upon screening a 5000 membered library in each case, the NDT
library (95% coverage) proved to have a higher density of improved hits than the NNK
library (15% coverage). Thus, consideration of reduced amino acid alphabets is a
powerful tool in beating the numbers problem in directed evolution. This technique
has been used successfully in several recent studies regarding the enhancement of
stereoselectivity. It was concluded that it is best to choose systems in which the
number of codons is equal to the number of amino acids, which helps to reduce the
inherent bias and overrepresentation of certain amino acids [49, 53, 54]. More than
one codon degeneracy in a given saturation mutagenesis experiment is also possible.
Since amino acid bias is neglected, these numbers are only approximations, but
they serve as a useful guide. Calculations have been performed over the whole range
of coverage (0–95%) for sites composed of 1, 2, 3, 4, and 5 amino acid positions as a
function of NNK versus NDT (Figures 5.8 and 5.9). The CASTER computer aid can be
employed to derive analogous curves for other codon degeneracies as well [49].
The generation and screening of focused libraries by saturation mutagenesis has
been practiced successfully many times for various purposes [2, 39, 55, 56]. The first
case of a focused library for enhancing the enantioselectivity of an enzyme concerns a
lipase, in which a site composed of four amino acid positions aligning the binding
pocket was randomized [56], although oversampling was not considered. In a later
statistical study, it was suggested that whenever substrate acceptance or enantios-
electivity need to be improved [57] sites closer to the active center are liable to be more
important than remote positions, a conclusion that was corroborated in a more recent
analysis [58].
It was not until systematization and especially iterativity were implemented that
saturation mutagenesis was really exploited to its full potential, specifically in the
form of iterative saturation mutagenesis (ISM) [59]. The experimenter systematically
identifies all relevant hot sites with the help of structural data (X-ray data or
5 aa
10000
4 aa
9000
8000
7000
Transformants
6000
3 aa
5000
4000
3000
2 aa
2000
1000
1 aa
0
0 10 20 30 40 50 60 70 80 90
Coverage [%]
Figure 5.8 Library coverage calculated for NNK codon degeneracy at sites consisting of 1, 2, 3, 4,
and 5 amino acid positions (aa ¼ amino acids) [49].
homology model), with a given site being composed of one or more amino acid
positions. Following the generation and screening of the respective saturation
mutagenesis libraries, the genes of the respective hits are then used as templates
for performing further rounds of randomization at the other sites. For illustrative
purposes the case of four sites is considered here (A, B, C, and D, Figure 5.10). If each
site is visited only once in a given upward pathway, convergency is reached after
preparing and screening a total of 64 libraries. As will be seen in Section 5.8.2.2,
complete scanning of such a defined and limited section of protein sequence space is
not necessary, that is, any one of the 24 pathways can be chosen [60]. If a dead end
as a consequence of a local minimum is encountered, i.e., if a given library contains
no improved mutants, backtracking is possible. Recent work has shown that in such
cases it is also possible to use non-improved or even inferior mutants as templates in
the subsequent round of saturation mutagenesis [60b]. In all applications of ISM, it is
crucial to develop reliable criteria for choosing the appropriate randomization sites.
10000
5 aa 4 aa
9000
8000
Transformants
7000
6000
5000
4000
3 aa
3000
2000
1000
2 aa
1 aa
0
0 10 20 30 40 50 60 70 80 90
Coverage [%]
Figure 5.9 Library coverage calculated for NDT degeneracy at sites consisting of 1, 2, 3, 4, and 5
amino acid positions (aa ¼ amino acids) [49].
134
Figure 5.10 Iterative saturation mutagenesis (ISM) employing four sites (A, B, C, and D), with each
site in a given upward pathway in the fitness landscape being visited only once [59].
The most important enzyme properties in practical biocatalysis have been

addressed when applying ISM, including enantioselectivity/substrate acceptance [54,
59, 60] and thermostability [40, 61]. In doing so it is necessary to develop reliable
criteria for choosing the appropriate randomization sites, specifically by turning to
structural biology as a guide. In the case of enantioselectivity and/or substrate
acceptance (rate), the criterion is defined by the combinatorial active-site saturation
test (CAST) [62]. All sites next to the binding pocket are systematically identified on
the basis of X-ray data or a homology model, with a site being composed of one or
more amino acid residues (Figure 5.11). CAST is a practical acronym that distin-
guishes the randomization process at such sites from saturation mutagenesis at sites
in different parts of the enzyme.
Rather than focusing only on a select single site as was generally done in most
previous saturation mutagenesis experiments [2, 39, 56], CASTing constitutes
systematization, thereby setting the stage for iterativity (ISM) [59]. Iterative CASTing
as a form of ISM is a relatively new approach, but it has already been applied to many
different enzymes in the quest to broaden substrate scope and to enhance enantios-
electivity, but also in attempts to alter the binding specificity of cofactors, introducing
enzyme promiscuity, and modulating biosynthetic pathways (see Table 5.2, Sec-
tion 5.5). Such a systematic method is a practical way to reduce the screening
effort [20], which is the bottleneck of directed evolution [2]. When applying ISM, a
decision has to be made regarding the question of how to group the CAST-identified
A B C
D
G binding E
pocket
H
F etc.
Figure 5.11 General scheme for CASTing [2b, g, h, 62]. The sites A, B, C, and so on align the binding
pocket and can be composed of one or more amino acid positions.
single residues into randomization sites. For example, if ten single residues
surrounding the binding pocket have been identified, ten single-residue saturation
mutagenesis libraries can be generated and screened, followed by iterative steps
according to an extended form of Figure 5.10. Alternatively they can be grouped into
five two-residue sites, among other possibilities. This strategic decision can have far-
reaching consequences – experience so far points to the use of two- or three-residue
sites as the preferred choice [59, 60]. It is also possible to randomize multiple residue
(5–10) sites, provided reduced amino acid alphabets are used (Section 5.8.3.2).
A different criterion for choosing randomization sites needs to be applied when
applying ISM to the thermostabilization of proteins [40, 61]. Since it was well known
that hyperthermophilic enzymes are more rigid than the mesophilic analogs [63] it
appeared reasonable to introduce appropriate mutations at sites displaying high
degrees of flexibility. As a rough but reliable guide for identifying such sites, atomic
displacement parameters can be used, namely, B-factors available from X-ray data,
which reflect smearing of atomic electron densities with respect to equilibrium
positions as a result of thermal motion and positional disorder. Consequently, the B-
factor iterative test (B-FIT) was developed, according to which only those sites
displaying the highest B-factors are considered for saturation mutagenesis, with
the process being performed iteratively (Figure 5.10) (see also Section 5.7) [40, 61].
Criteria for choosing randomization sites other than the two standard possibilities
described above are also possible, such as homology-based approaches utilizing
information from sequence alignments as one of several data-driven strategies [64].
For example, consensus engineering is a way to increase the thermostability of a
protein by modifying a protein sequence so that it more closely resembles a
consensus defined from the alignment of members of a particular family [65a,b].
Related to this is combinatorial consensus mutagenesis (CCM), in which sequence
alignment serves as a guide for finding mutagenesis sites [65c,d]. In principle these
methods can be been used as a basis for choosing randomization sites for ISM
processes, which can be expected to be promising.
Along a different line, cell-free protein synthesis has been combined with
saturation mutagenesis at sites near the binding pocket in a process called single-
molecule-PCR-linked in vitro expression (SIMPLEX) [66a]. Accordingly, DNA mole-
cules are diluted to such an extent that one molecule per well is reached and PCR-
amplified, then the cloned nucleic acid library is directly transformed into a protein
library by an in vitro coupled transcription/translation system. This approach was
applied in the quest to invert the stereoselectivity of Burkholderia cepacia lipase in a
process in which four residues were chosen at which only hydrophobic amino acids
(Gly, Ala, Val, Leu. Ile, Met, and Phe) were introduced combinatorially [66b].
5.4.3
Recombinant Methods
Rather than introducing point mutations, as in the above methods, recombination

can also be considered, meaning the breaking and rejoining of DNA in new
combinations [2, 14, 18, 67–69]. DNA shuffling is the most prominent method [14].
136
In general, one or more genes are first digested with a DNase to yield double-stranded
oligonucleotide fragments of 10–50 base pairs, which are then amplified in a PCR-
like process. Accordingly, a series of cycles of strand separation and re-annealing in
the presence of a DNA polymerase followed by a final PCR-amplification result in the
reassembly of full-length mutant genes. The size of the fragments needs to be
optimized in addition to the temperature cycle during reassembly. It has been noted
that the amount of assembly reaction as well as the number of cycles in the
amplification are critical variables that likewise need to be optimized [67]. Point
mutations can occur in the PCR process. The method was first applied in the activity
enhancement of TEM b-lactamase [14]. DNA shuffling can be performed with one
gene, with two or more natural genes, or with mutant genes (Figure 5.12). In general a
relatively high degree of homology is necessary (at least 70%). A particularly efficient
version is family shuffling [68], in which homologous genes from different species
are chosen, such an approach provides high catalyst diversity. In these recombinant
methods a certain degree of self-hybridization of parental genes occurs, especially if
homology is low, which lowers the quality of the mutant libraries. To assess the
efficiency of recombination and to optimize shuffling protocols, probe hybridization
has been used in macroarray format, allowing the analysis of chimeric DNA
libraries [69]. Accordingly, several hundred shuffled genes encoding dioxygenases
were characterized, revealing biases in the shuffling reaction.
Various improvements and variations of recombinant methods have been
reported [2, 18], including the so-called staggered extension process (StEP), which
is based on cross hybridization of growing gene fragments as the DNA polymerase-
catalyzed primer extension process [70]. Subsequent to denaturation the primers
anneal and extend under conditions that limit extension, allowing the primers to re-
anneal to different parent sequences throughout the multiple cycles randomly.
Finally, the recombinant full-length gene products are amplified by PCR. Another
Figure 5.12 DNA shuffling [14, 68] for the case in which the parental genes originate from the WT
by some other sort of mutagenesis.
shuffling variation is biased mutation-assembly (BMA), in which a mutant library is
generated by employing the overlap extension polymerase chain reaction technique
with DNA fragments from WTand phenotypically improved mutant genes (e.g., from
epPCR) [65d]. By mixing the ratio of the DNA fragments to WT fragments, the
number of mutations assembled in the WT gene can be controlled stochastically.
BMA was applied to the thermostabilization of prolyl endopeptidase from Flavo-
bacterium meningosepticum, with the proportion of thermostable mutants increasing
as the mixing ratio was raised.
Various other homology-dependent and independent in vitro recombination
methods have been developed and summarized in reviews [2, 18] – some examples
are incremental truncation for the creation of hybrid enzymes (ITCHY) [71a],
Thio-ITCHY [71b], (SCRATCHY, a combination of ITCHY and DNA shuffling) [72],
sequence homology-independent protein recombination (SHIPREC) [73], sequence-
independent site-directed chimeragenesis (SISDC) [74], recombined extension on
truncated templates (RETT) [75], recombination-dependent exponential amplifica-
tion PCR (RDA-PCR) [76], and SCOPE [77].
Only a few of the advancements are highlighted here. A prominent method is
random chimeragenesis on transient templates (RACHITT) [78], which is a con-
ceptually distinct alternative to sexual PCR for gene family shuffling. Accordingly,
thermocycling, strand switching, or staggered extension are not necessary; instead,
the method relies on the trimming, gap filling, and ligation of parental gene
fragments hybridized on a transient DNA template. This elegant approach was
applied successfully to the directed evolution of dibenzothiophene monooxygenase,
which catalyzes the first step of the dszABCD diesel biodesulfurization pathway.
Significantly increased reaction rate and a broadened scope of substrate acceptance
were achieved [78]. Another approach to shuffling of low-homology genes is
degenerate oligonucleotide gene shuffling (DOGS), which requires the design of
strictly complementary pairs of primers [79]. A non-degenerate core flanked by both
50 and 30 degenerate ends characterizes each primer. The overall process ends up in
the formation of chimeric fragments that maintain parental sequence at the points of
segment overlap. It does not require the use of endonucleases for gene fragmen-
tation, while allowing for random mutagenesis of selected segments of the gene.
These advantages have been combined with random drift mutagenesis (RNDM),
which enables a wider exploration of the sequence space of shuffled genes. A diverse
family of b-xylanase genes having significantly different G and C contents was
successfully subjected to DOGS and RNDM [80].
Another noteworthy development is random strand transfer recombination
(RSTR), which is based on the ability of reverse transcriptases to undergo homol-
ogy-independent template switches during the DNA synthesis [81]. The method
appears to be fairly general, involving spontaneous base-pairing dependent recom-
bination at high frequency between genes having low or high sequence homology.
A different recombinant method is biased mutation-assembly, in which a library is
generated by overlap extension PCR with DNA fragments from a WT enzyme and
phenotypically advantageous mutant genes [65d]. The number of mutations assem-
bled in the WT gene is controlled stochastically by the mixing ratio of the WT
138
Figure 5.13 Schematic representation of ISOR [83]. The use of biotinylated DNA and purification
by capture onto streptavidin-coated beads is optional.
fragments to the mutant DNA fragments. In yet another approach, synthetic oligo-
nucleotides are added to a mixture of gene fragments prior to reassembly [82]. This
method was later optimized and dubbed incorporating synthetic oligonucleotides via
gene reassembly (ISOR) (Figure 5.13) [83]. A biotinylated PCR product of the target
gene is subjected to DNase-I-mediated fragmentation, the fragments then being
mixed with a set of synthetic nucleotides. Following reassembly by self-primed
extension catalyzed by Taq-polymerase, the genes are enriched by capture on strepta-
vidin-coated magnetic beads. This step maintains the diversity in the assembly process
by minimizing mispriming and reducing amplification of short products. ISOR was
applied to a cytosine-C5 methyltransferase, with 45 individual positions being
randomized, and also to serum paraoxonase PON1, with insertions and deletions
(indels) at different sites surrounding the binding pocket being the goal. In both cases
libraries were obtained that harbored mutants with altered substrate specificities [83].
As an alternative to starting from genes and subjecting them to fragmentation/
reassembly, it is also possible to perform artificial shuffling by exploiting the
information regarding their sequences in designing appropriate DNA fragments that
are then assembled, as three independent studies have shown [84–86]. When
Case I Gene A
Gene B
Case II Gene A
Gene B
Figure 5.14 General concept of ADO [86], with denote the synthetic oligonucleotide fragment
two strategies for the linking of fragments being with degenerate nucleotides. The gray blocks
possible (case I and II). In case I the two genes A denote conserved regions of sequence that can
and B to be virtually shuffled are aligned; the be used as the linking part with homologous
different colored stars refer to information that recombination. Case II shows no homology
encoded different amino acids, while between flanking oligos, which can be
oligonucleotide fragments with both colored assembled by ligation between ssDNA with an
stars in the same position of the parent gene unknown terminal sequence.
considering two or more genes, overlapping oligonucleotides encoding all of the

degeneracy, as determined by gene alignment and computer analysis, are used as
substrates in a polymerase-mediated assembly process. The three methods, termed
degenerate homoduplex recombination [84], synthetic shuffling [85], and
assembly of designed oligonucleotides (ADO) [86], differ in some respects but
they basically utilize the same principle. In all cases, libraries of gene mutants are
generated in which all shuffled variants are equally likely, regardless of how tightly
parental diversity is linked. Low homology is tolerated, and undesired self-hybrid-
ization of parental genes leading to the appearance of WTenzyme is minimized. This
means higher-quality libraries. The advantage of ADO (Figure 5.14), which includes a
certain degree of new non-parental diversity due to the oligonucleotide design
strategy, was analyzed on a statistical basis [18]. Two strategies for linking fragments
are possible [86]. The three approaches to synthetic DNA shuffling offer interesting
perspectives, with ADO not depending on matters relating to intellectual property.
5.4.4
Other Methods
A very different approach to generating enzymes with different catalytic profiles

utilizes domain swapping, which involves the replacement of a secondary or tertiary
element of a given protein by the same element of another (related) protein [87]. It has
140
been used to trace evolutionary relationships of enzymes, but also to some extent in
the endeavor to create hybrids with different catalytic profiles. This approach has not
been used very often in directed evolution [87], but it appears to offer interesting
perspectives, as, for example, in an effort to manipulate the substrate scope of
glycosyltransferases [88], an area of great practical importance. Another example is
the generation of domain-swapped chimeras of the glutamate dehydrogenase from
Clostridium symbiosum and from E. coli (EcGDH), catalyzing the reversible oxidative
deamination of L-glutamate to 2-oxoglutarate with release of ammonia [89].
Yet another technique recently applied to directed evolution is circular permuta-
tion of proteins, which is defined as the intramolecular relocation of a proteins C and
N termini [90]. Natural evolution makes use of this trick, and in the laboratory it is
best accomplished by gene manipulation. This novel technique has been used to
enhance the catalytic activity of the lipase B from Candida antarctica (CALB, Candida
antarctica lipase B) [91]. In the case of the xylanase from Bacillus circulans circular
permutants were obtained that showed altered catalytic profiles; these enzymes can
in principle be used as starting points directed evolution using conventional
techniques [92]. An efficient database search tool (CPSARST) has been developed
which serves as an aid when performing circular permutation [93].
5.5
Strategies for Applying Gene Mutagenesis Methods
5.5.1
General Guidelines
The number of gene mutagenesis methods is almost bewildering (Section 5.4),

making it difficult for experimenters to make appropriate choices. It is unlikely that a
single method or strategy can be expected to be universal, yet trends are beginning to
emerge. The choice depends upon the immediate goal at hand. Many laboratories
utilize directed evolution for mechanistic purposes, and in these cases the choice may
not be crucial since all that is needed is structural perturbation followed by
interpretation leading to a gain in information concerning the question of how
enzymes function on a molecular level. EpPCR and/or saturation mutagenesis at
predetermined positions in the enzyme have proven to be successful in these
endeavors. Other academic groups interested in exploring the effects of laboratory
evolution of enzymes not previously subjected to protein engineering may likewise
utilize such easy-to-use methods. However, researchers desiring rapid and efficient
directed evolution, be it for reasons inherent in basic research or in industrial
applications, need to make the right decisions regarding the type of gene mutagen-
esis method as well as the optimal strategy for applying a given technique.
In any laboratory evolution project, evaluation of the mutants is crucial, be it by a
screening or a selection procedure (Section 5.6). Screening assays can typically
handle 800–8000 transformants per day, while selection may well entail 107 clones or
more on the DNA level. However, universal screening assays or selection systems
5.5 Strategies for Applying Gene Mutagenesis Methods j141
have not been developed. Thus, the decision regarding the mutagenesis method and
the size of the libraries depends upon the availability of an appropriate screening
assay [20] or selection system [19] (Section 5.6). This also makes the comparative
assessment of different mutagenesis methods difficult, inter alia because one
technique may have been used in conjunction with a screening procedure and
the other utilized in a selection system. Unfortunately, very few comparative studies
have been reported based on a given enzyme and using one and the same screening
system [17, 18, 60a]. When proposing new mutagenesis methods and/or novel
strategies as part of methodology development in laboratory evolution, sooner or later
comprehensive comparisons need to be made.
At this point a few general comments are in order. When applying epPCR, the
question of mutation rate arises. In the 1990s it was customary to utilize low-error
epPCR averaging one amino acid exchange per enzyme molecule, as, for example, in
the original study concerning directed evolution of an enantioselective lipase in which
four cycles of epPCR were performed with the accumulation of four point mutations
(Section 5.5.2) [15]. However, a subsequent study using the same enzyme and the
identical chiral substrate showed that epPCR at higher mutation rate resulting in the
simultaneous introduction of three amino acid substitutions provides considerably
better results [56], as in other reports [2]. Later it was noted that libraries generated by
high-error-rate mutagenesis are enriched in improved sequences because they
contain more unique functional clones (but fewer with retained function) [94].
Nevertheless, it is advisable in any new study to test different mutation rates.
Sometimes, contradictory statements regarding optimal evolutionary pathways
adorn the literature. For example, it has been claimed that climbing the hill in
a fitness landscape is best accomplished if single mutations accumulate one by
one [95]. Indeed, such strategies have been shown to be successful [2], a recent example
being the application of ISM for increasing rate and enantioselectivity of an enoate-
reductase [54]. On the other hand, the simultaneous introduction of two or more
mutations in saturation mutagenesis makes possible cooperative effects operating
between the point mutations within a given site in addition to possible synergism with
the mutations introduced in previous or subsequent cycles [49b]. The advantages
arising from such a strategy have been demonstrated experimentally (Section 5.8.2.2).
Experimental and theoretical studies have shown that the accumulation of too
many point mutations when attempting to increase rate, manipulate substrate
scope, and enhance or reverse stereoselectivity may lead to a destabilization of the
protein [2, 94]. This needs to be kept in mind, especially when aiming for practical
applications. One way to avoid this is to incorporate in the screening system
thermostability as one of the parameters. Another guideline in this respect originates
from the conclusion that evolvability becomes easier when starting from more robust
enzymes [96]. This means that it may be wise to evolve the respective parameters
successively, specifically by first increasing thermostability, and then proceeding by
utilizing the best mutant as a template for targeting other catalytic parameters such as
stereoselectivity [2b,h]. A theoretical and practical question arises when evolving two
parameters simultaneously [2]. In a recent study involving the evolution of rate
(broadening substrate scope) and stereoselectivity of an enoate-reductase using
142
iterative CASTing, it was demonstrated that it is important not to apply constraints that
are too stringent, that is, not to pick the best mutant displaying maximum improve-
ment of one catalytic parameter [54]. Rather, it is better to accumulate a panel of
medium- and higher-quality hits in terms of activity, which are then screened for the
second parameter relating to stereoselectivity, before going into the next round of
mutagenesis/screening. These kinds of non-discarded variants (lateral hits) [54] can
be related to the neutral drift theory proposed in other systems [97]. A relationship
with the Eigen–Schuster notion of quasi-species [98] was also noted [54], which
had been proposed in other directed evolution work [99]. The strategy regarding
relaxed constraints is generalized in Figure 5.15 – the upper right section of the
cartoon pictures the desired variants with improved properties of both catalytic
parameters A and B [54]. Most recently it has been shown that even inferior mutants
in a given library constitute superior starting points in further mutagenesis rounds, as
in ISM [60b].
With the increasing emphasis on efficacy in directed evolution [2, 17, 18, 58, 59,
100–102] it worth considering quality control of libraries, especially with respect
to amino acid bias. Owing to the relatively high experimental effort, such controls are
rarely made, yet without such checks screening efforts may be wasted, that is,
Figure 5.15 Preferred strategy when which are not used in further mutagenesis;
optimizing two catalyst properties A and B red-crossed blue and green circles:
simultaneously [54]. The black star indicates the variants with improved property A or B;
desired variant; blue and green dashed lines: red-crossed black circles: mutants with
stringent thresholds; blue and green rectangles: improved A and B property. Black
relaxed thresholds; blue and green filled circles: dashed arrows: second round of
best mutant for property A and B, respectively, mutagenesis.
You should not search for something that does not exist [54]! For example, saturation
mutagenesis libraries in which the circular template has not been efficiently
eliminated will require considerably more oversampling. Even worse, in those cases
in which certain codons are under-represented or even missing from the library,
higher degrees of oversampling will not solve the problem [54]. Extracting and
analyzing all plasmids in a given library entails an insurmountable amount of work.
Therefore, a short-cut control was developed, in which the quality of libraries was
checked by performing sequence analyses of pooled plasmids for each library prior to
transformation into the expression strain [54]. The fact that the costs of sequencing is
continuing to decrease makes this approach viable and highly advisable for preventing
wrong conclusions.
5.5.2
Rare but Helpful Comparative Studies
This section highlights several comparative studies that likewise serve as guides
when faced with the problem of choosing an appropriate mutagenesis strategy in
future laboratory evolution studies [58–60, 100, 102]. A revealing case pertains to
a study regarding the targeted activity switch from natural b-galactosidase to b-
fucosidase behavior [100]. The purpose was to compare the virtues of DNA shuffling
as applied to the model reaction in an earlier study [103] with saturation mutagenesis
in the new attempt. In both cases, the E. coli b-galactosidase (BGAL) was used as the
enzyme in the hydrolysis of p-nitrophenyl-b-D-fucopyranoside (pNP-fuc), a substrate
showing very low activity with WT BGAL. The successful earlier study had used seven
cycles of DNA shuffling, requiring high-throughput screening of 10 000 transfor-
mants in each round (totaling 70 000), which was achieved by a color-based pre-
screen followed by kinetic characterization. The best variant contained eight amino
acid substitutions, with only two being at the active site. It showed a tenfold increase
(kcat/KM) in reactivity toward pNP-fuc and a 39-fold decrease in reactivity toward the
native substrate p-nitrophenyl-b-D-galactopyranoside (pNP-gal), which amounts to
a 1000-fold shift in selectivity [103]. Notably, the two substrates are essentially
identical, except that pNP-gal lacks the hydroxyl-function at C6. In the comparative
study [109], saturation mutagenesis was focused on residues thought to be critical for
binding the galactose substrate as revealed by an earlier X-ray study [104], namely,
Asp201, His540, and Asn604 (Figure 5.16).
In the comparative study the authors grouped the three residues 201, 540, and 604
into one site and randomized all three components simultaneously using NNK codon
degeneracy encoding all 20 canonical amino acids [100]. This is reminiscent of an
earlier study of saturation mutagenesis at a 4-residue site of a lipase [56]. Thereafter
10 000 transformants were screened using the previously described assay. Several
active mutants with switched selectivity were identified, with the best variant being a
double mutant His540Val/Asn604Thr with retained amino acid at Asp201. The
evolved b-fucosidase showed a 180-fold increase in kcat/KM in reaction with pNP-
fuc and a 700 000-fold inversion of selectivity. As noted in this comparative study, only
a small portion of the total library was actually screened, yet superb results were
observed. Assuming the absence of amino acid bias, about 100 000 transformants
144
Figure 5.16 Structure of the E. coli here), except that it lacks the C6 hydroxyl group.
b-galactosidase active site [104] used to choose Dotted lines represent hydrogen bonds. The
randomization sites [100]. The p-nitrophenyl- Asp201, His540, and Asn604 residues were
b-D-fucopyranoside (novel substrate) is randomized in this study. The sodium ion
identical with p-nitrophenyl-b-D- (filled sphere) is reduced in scale so as not to
galactopyranoside (native substrate, shown obscure these amino acid residues.
would have had to be screened for 95% coverage (Table 5.1). Upon comparing the two
approaches, it is obvious that the molecular biological work (a single round of
saturation mutagenesis versus seven rounds of DNA shuffling) as well as the
screening effort (10 000 versus 70 000 transformants) differ vastly, allowing the
authors to conclude that the semi-rational approach based on focused library gener-
ation is clearly more efficient [100]. However, they were careful not to generalize.
Table 5.1 Oversampling necessary for 95% coverage as a function of NNK and NDT codon
degeneracy assuming the absence of amino acid bias [40, 49].
NNK NDT
No.a) Codons Transformants Codons Transformants

needed needed
1 32 94 12 34
2 1 028 3 066 144 430
3 32 768 98 163 1 728 5 175
4 1 048 576 3 141 251 20 736 62 118
5 33 554 432 100 520 093 248 832 745 433
6 >1.0 109 >3.2 109 >2.9 106 >8.9 106
7 >3.4 1010 >1.0 1011 >3.5 107 >1.1 108
8 >1.0 1012 >3.3 1012 >4.2 108 >1.3 109
9 >3.5 1013 >1.0 1014 >5.1 109 >1.5 1010
10 >1.1 1015 >3.4 1015 >6.1 1010 >1.9 1011
a) Number of amino acid positions at a given site.

Another study using an epoxide hydrolase as the catalyst in the hydrolytic kinetic
resolution of a chiral epoxide allowed for the first time the comparison of epPCR with
ISM (Section 5.8.2.2). In the original epPCR study [105], about 20 000 transformants
had to be screened to boost the selectivity factor E from 4.6 to 11. In the ISM approach
involving five iterative cycles of saturation mutagenesis, the same number of trans-
formants were screened, yet the final mutant showed dramatically higher stereo-
selectivity (E ¼ 115) [59]. Although systematization was not strived for, this initial
example of ISM served as a hint that this kind of protein engineering is characterized
by high efficacy, although again generality could not be claimed at the time. Later a
method for quality control was devised and applied to this particular system, based on
exhaustive deconvolution and the construction of a fitness landscape featuring 120
pathways from the WT enzyme to the best mutant (Section 5.8.2.2) [106]. It was
discovered, inter alia, that ISM is characterized by pronounced cooperative effects
operating between point mutations and sets of mutations along a given evolutionary
pathway, and that superfluous sets of mutations are not accumulated.
The most systematically studied enzyme in terms of comparing different muta-
genesis methods and strategies is the lipase from Pseudomonas aeruginosa (PAL) as a
catalyst in the hydrolytic kinetic resolution of rac-1 (Scheme 5.1) [15, 17, 56, 60, 107].
WT PAL shows a selectivity factor of only E ¼ 1.1 (S). Following four cycles of epPCR at
low error rate, an improved variant (E ¼ 11.3) with four point mutations Val47Gly/
Ser149Gly/Ser155Leu/Phe259Leu was identified [15]. This study also served as proof-
of-principle of the concept of directed evolution of stereoselective enzymes, and
marked the beginning of a fundamentally new approach to asymmetric catalysis
[15, 17].
O O O
R H2O R R
O NO2 OH + O NO2 + -O NO2
lipase
CH3 CH3 CH3
rac-1 (R = n-C8H17 ) (S)-2 (R)-1 3
Scheme 5.1 Hydrolytic kinetic resolution of rac-1 catalyzed by PAL mutants [15, 17, 56, 60, 107].
A fifth epPCR round improved enantioselectivity only marginally (E ¼ 13), indi-

cating that better strategies had to be developed. Saturation mutagenesis experiments
at the four hot spots identified earlier by the epPCR process [15] provided in one case
an improved mutant, that is, when randomizing position 155 remote from the active
site (E ¼ 20), but in others no hits could be detected [107]. The problem with this
strategy, which is used often in directed evolution [2], is the fact that epPCR generally
leads to the accumulation of some superfluous point mutations, which means that
saturation mutagenesis thereat may not result in a positive response.
In an important experiment, a focused library was generated by saturation
mutagenesis at a site aligning the binding pocket of PAL that consisted of positions
160–163 as indicated by the PAL X-ray structure [108], which led to the identification
of a mutant Glu160Ala/Ser161Asp/Leu162Gly/Asn163Phe with an E-value of 30
following the screening of 5000 transformants, although oversampling was not
146
Figure 5.17 Binding pocket of PAL for the acid usual catalytic triad Asp/His/Ser, serine at
part of rac-1, showing the geometric position of position 82 attacks the carbonyl function
amino acids 160–163, which were randomized nucleophilically with rate- and stereochemistry-
simultaneously by saturation mutagenesis to determining formation of a short-lived
enhance enantioselectivity [56]. As part of the oxyanion [108].
considered (Figure 5.17) [56]. This was the first case of a focused library at a site
aligning the binding pocket of an enzyme with the purpose of enhancing stereo-
selectivity (the acronym CASTas defined in Figure 5.11 did not exist at the time.). The
positive result served as a signal that mutations near the active center may have a
greater influence on stereoselectivity than mutations at remote sites, a preliminary
hint that was later corroborated by a statistical analysis [57] and recently once again
substantiated experimentally [2a, 58].
Other exploratory experiments regarding PAL began to suggest that randomiza-
tion at one residue and then turning to another position, applying once again
saturation mutagenesis (or epPCR), could constitute a useful way to probe protein
sequence space [17, 56, 107], as in several other reports [39, 109]. However, this
approach was not systematized until a few years later with the emergence of iterative
saturation mutagenesis (ISM) (Section 5.4.2) [59].
The most stereoselective PAL-variant was obtained by applying a strategy consisting
of high error-rate epPCR and DNA shuffling with simultaneous randomization again
at positions 155/162 in a modified combinatorial multiple-cassette mutagenesis
(CMCM) process [110], leading to a selectivity factor of E ¼ 51 [56]. This (S)-selective
mutant is characterized by six point mutations, with only one (Leu162Gly) being near
the binding pocket, which came as a surprise. A QM/MM study not only unveiled the
source of enhanced enantioselectivity as being caused by a relay mechanism, it also
predicted that only two of the six point mutations are actually necessary, namely,
Ser53Pro and Leu162Gly [111]. The double mutant Ser53Pro/Leu162Gly was subse-
quently prepared by site-directed mutagenesis and was found to be even more
Figure 5.18 Summary of early work on directed evolution of enantioselective PAL variants as
catalysts in the hydrolytic kinetic resolution of rac-1 [15, 17, 56, 107].
enantioselective (E ¼ 63 in favor of (S)-2) [111]. This was a triumph of theory, but the
fact that four superfluous mutations had accumulated clearly demonstrated that the
chosen strategy was far from efficient. The accumulation of such mutations means
unnecessary laboratory work, especially with regard to screening. The total effort in
obtaining the best mutant involved the screening of more than 50 000 transfor-
mants [15, 17, 56, 107]. Figure 5.18 summarizes the extensive exploration made of
protein sequence space in the quest to enhance the stereoselectivity of PAL.
Recently, this experimental platform was revisited, this time applying saturation
mutagenesis iteratively according to ISM. Six single residues aligning the acid-part of
the binding pocket were considered, raising the question of how to group them [60a].
Rather than choosing six single-residue sites for ISM, three double-residue CAST
sites were defined. After screening only 10 000 transformants, a highly active PAL
variant characterized by three point mutations aligning the binding pocket was
evolved, showing a selectivity factor of E ¼ 594 in the hydrolytic kinetic resolution of
rac-1 (Scheme 5.1). Deconvolution demonstrated that none of the three point
mutations are superfluous, and that strong cooperative effects are acting between
the mutations [60]. This comparative study shows that ISM is faster and more
efficient than all previous attempts based on epPCR at different mutation rates,
saturation mutagenesis at hot spots, DNA shuffling, or combinations thereof. The
study also revealed that the process of grouping can be crucial – randomization at
single-residue sites was not successful in this particular system. Finally, it was shown
that the triple mutant accepts various different chiral esters with good to excellent
148
Table 5.2 Examples of iterative saturation mutagenesis (ISM).
Enzyme Property evolved Reference
Epoxide hydrolase from Aspergillus Increasing enantioselectivity [59]

niger
Enoate reductase YqjM Increasing rate and [54]
enantioselectivity
Lipase from Pseudomonas aeruginosa Increasing rate and [60]
enantioselectivity
Lipase from Candida antarctica A Increasing rate and [112]
enantioselectivity
Arylmalonate decarboxylase from Broadening substrate scope and [113]
Bordetella bronchiseptica enhancing stereoselectivity
Amidohydrolase from Deinococcus Introducing promiscuity by [114]
radiodurans inducing phosphotriesterase activity
Esterase form Burkholderia gladioli Inverting enantioselectivity [115]
Cellobiose phosphorylase from Introducing promiscuity by tuning a [116]
Cellulomonas uda cellobiose into a lactose-
phosphorylase
Xylose reductase from Pichia stipitis Enforcing a switch from NADPH to [117]
NADH binding specificity useful in
xylose fermentation
Cyclodextrin glucanotransferase Introducing promiscuity by conver- [118]
sion into an a-amylase
Isoeugenol 4-O-methyltransferase Modulating lignin biosynthesis for [119]
from Clarkia breweri better utilization of plants in paper-
making, biofuel production, and
agriculture
Streptavidin Enhancing enantioselectivity of a [120]
hybrid catalyst based on a biotiny-
lated diphosphine/Rh-complex
anchored non-covalently to
streptavidin
enantioselectivity, with most of the substrates being sterically demanding com-

pounds not accepted by WT PAL [60].
The use of ISM for improving and altering various different protein properties,
including tasks in metabolic pathway engineering, is rapidly expanding, as sum-
marized in Table 5.2 (Section 5.5.2). Generally, in these studies comparisons with
epPCR, DNA shuffling, or other mutagenesis methods or strategies were not made,
which means that general conclusions regarding relative efficacy cannot be made.
Nevertheless, the method is efficient, because in all cases small libraries requiring a
minimum of screening effort were involved leading to excellent results.
More comparative studies of this type are needed in the rapidly growing area of
directed evolution, hopefully providing reliable guidelines regarding the optimal
choice of mutagenesis methods and strategies. Another possibility for maximizing
efficacy in directed evolution concerns computational guides (Section 5.5.3).
5.5.3
Computational Guides
Various very different computational aids have been devised for directed evolution
and protein design, including user-friendly computer programs helpful in library
construction, such as GLUE-IT/PEDEL-AA [42], CASTER [40], B-FITTER [40], and
HotSpotWizard [121]. Computational protein design and other types of in silico
platforms including QM/MM methods have been reviewed [122, 123]. Strategies
relying on bioinformatics data can be quite efficient [73,124].
A structure-guided directed evolution method utilizing recombination processes
is SCHEMA (Figure 5.19) [125]. It is based on identifying blocks of sequences that
minimize structural disruption when recombination into chimeric proteins occurs.

Pairs of interacting amino acid residues within 4.5 A of each other are identified and
used as a basis for contact matrices. SCHEMA provides an optimization algorithm
that selects crossovers that minimize the average disruption of the library. Interac-
tions that are broken upon recombination contribute to a so-called disruption score
needed in the design of shuffling experiments. To calculate the average disruption, a
high-resolution X-ray structural data of at least one of the proteins is necessary.
This computational guide has been applied to cytochrome P450 enzymes [125a],
b-lactamases [125a], and cellulases [125b].
Several alternative approaches have been proposed. The algorithm HybNat, which
likewise utilizes structural data and partitions residues into mutually exclusive
clusters of interacting amino acids, is another way to minimize disruption when
applying recombination methods [126], as is the evolutionary information inherent
in natural multiple sequence alignment used in the computer aid FamClash [127].
A hybrid of SCHEMA and FamClash has been proposed that seems to be particularly
Figure 5.19 SCHEMA disruption is based simplified model) [125]. (a) Disruptions in a
upon a contact matrix representing interactions simplified model; (b) Contact matrix to be
between amino acids in the three-dimensional adjusted for the sequence identity of the parent
structure of a protein (illustrated here with a enzymes.
150
effective [128]. Genetic algorithms have also been used to model the process of
directed evolution in silico [129]. The robustness and ease of application of these
computational approaches need to be tested on a broad scale.
In a different development, an algorithm relating to protein sequence–activity
relationships (ProSARs) was devised [130], analogous to quantitative structure–
activity relationships (QSARs) used in therapeutic drug discovery. Accordingly,
shuffling-based directed evolution is augmented by a strategy for statistical analysis
of protein sequence activity relationships, that is, additional information provided by
sequence–activity data as evolutionary cycles are transversed is exploited in mutation-
oriented enzyme optimization. Following each round of mutagenesis/screening, the
best hit is selected to serve as a template for programming diversity in the next round
by inferring the contributions of mutational effects on enzyme function (Figure 5.20).
About 50 mutations (variables) are evaluated in the combinatorial libraries
(in hopper) at any point. Then the characterized hits are sequenced, as are a fraction
of less improved variants. Following ProSAR analysis, individual mutations are
parsed into four classes: beneficial, which are fixed into the population by retention
in the next round parental enzyme, potentially beneficial, which are sent back into
the hopper for retesting, deleterious, which are discarded and neutral, which
have little or no effect on protein function and are discarded. The extent of diversity is
maintained by addition of further diversity discovered, for example, through rational
design, homologous sequences, saturation or PCR mutagenic libraries, or other
evolution programs [130].
The method was applied to the evolution of mutants of halohydrin dehalogenase
(HHDH) from Agrobacterium radiobacter as a catalyst in the production of a chiral
Figure 5.20 Formal representation of ProSAR [130].

intermediate 6 needed in the production of the cholesterol-lowering drug
LipitorÒ [130] (Scheme 5.2). WT HHDH is not active enough for practical application.
The goal was a production procedure characterized by 100% conversion of at least
100 g per liter substrate, a volumetric productivity of >20 g product per liter per hour
per gram of enzyme. Moreover, a simple isolation procedure and an easy enzyme
formulation process obviating extensive enzyme purification was strived for. All of
these prerequisites for industrial production were fulfilled, although with a great deal
of experimental effort. Following 18 rounds of the ProSAR procedure, each involving a
1.5-fold improvement in activity, a significantly more active final mutant was identified
that displayed a 4000-fold increase in volumetric productivity. Enantiomeric purity of
the product 6 starting from pure (S)-4 was maintained at >99.9% (R) [130].
OH O O OH O
HHDH O HHDH
Cl NC
OEt pH7.3 OEt pH7.3 OEt
4 5 6
HO O-
O
OH
F CH3
N
CH 3
O
HN
7
R
Lipitor
Scheme 5.2 Halohydrin dehalogenase (HHDH)-catalyzed formation of chiral intermediate 6

needed in the production of LipitorÒ (7) [130].
Another approach to the synthesis of LipitorÒ is based on the enhancement of

stereoselectivity of a keto-reductase, again using ProSAR, in which an appropriate
prochiral ketone was reduced with high enantioselectivity [131] although details are
lacking. Other companies are likewise poised to enter the generic drug market
regarding this multi-blockbuster when the original patents expire, with some
likewise using directed evolution.
Yet another approach to computational protein library redesign is a method called
IPRO (iterative protein redesign and optimization) [132]. Using known energy-based
scoring functions as part of a protein-docking algorithm [133], residue and rotamer
design choices are provided on a globally convergent mixed-integer linear program-
ming formulation (Figure 5.21). Rotamer–backbone and rotamer–rotamer energies
are computed for all of the selected rotamers using appropriate energy functions, and
152
Figure 5.21 Steps necessary in iterative protein redesign and optimization (IPRO) [132].
a mixed-integer linear programming formulation is applied to select the optimal

rotamer at each of the positions with energy minimization. IPRO as a protein
redesign software was illustrated by increasing the average binding energy of a
dihydrofolate reductase through targeted mutations in the parental sequences.
Subsequently, it was employed in the computational design of cofactor switching
in Candida boidinii xylose reductase by changing from NADPH to NADH binding
selectivity [134]. Out of 8000 possible combinations of mutations, ten were chosen on
computational grounds, and of these seven were shown to induce the predicted
switch, the best mutant being 27-times more active on NADH. Generalization of
IPRO needs to be achieved, especially in enzyme catalysis. One of the proposed
perspectives in another sense, namely, that grafting binding sites from one protein to
another should be possible, has been put into practice by the development and
experimental application of OptGraft [135].
Finally, the web server HotSpotWizard has been designed so that hot spots in
protein engineering can be identified rapidly, the aim being to support protein
engineering of activity, substrate scope, and stereoselectivity [121]. The sensitive
residues are selected by integration of structural, functional, and evolutionary
information obtained from databases and protein engineering tools. It remains to
be seen how this computation-driven approach performs when applied to such
directed evolution methods as ISM (Section 5.4.2).
5.6
Screening Versus Selection
Screening in laboratory evolution [20] needs to be distinguished from selection [19].

When screening for rate, stereoselectivity, or thermostability, every mutant enzyme is
evaluated individually in a high- or medium-throughput manner using some kind of
5.6 Screening Versus Selection j153
automated analytical technique based on UV–vis spectroscopy, fluorescence, MS, IR
spectroscopy, or even NMR techniques. Depending upon the specific enzyme
and substrate, throughput typically ranges between 500 and 800 transformants per
day, as in e.e.-determinations based on UV–vis plate readers, automated GC [136a],
or HPLC [136b]. Multiplexing MS allows 5000–8000 e.e.-determinations per day,
but it requires expensive instrumentation [136c]. This assay utilizes isotope-labeled
meso-type substrates in desymmetrization reactions or pseudo-enantiomers in
kinetic resolution of racemates. None of these systems are general; for
example, such processes as the asymmetric reduction of prochiral ketones are not
covered by the MS-based e.e.-assay. Multiplexing GC and HPLC as devised by
Trapp [137] may provide in the near future a fairly general solution to the problem
of screening [2b, 138]. Pre-tests that roughly assess activity, preferably in an on-plate
manner, are extremely useful and have been developed for several enzyme types [20],
the tributyrin test for lipase activity being an example [139]. The currently
available screening assays have been reviewed [20]. A pitfall in terms of practical
application of an enzyme, once evolved, needs to be considered when using a
surrogate substrate, namely, that a compound containing a covalently attached
UV–vis chromophore or a fluorophore. In such cases the evolutionary process may
well be specific for that particular model compound, which in real applications will
never be used.
To increase throughput in screening processes, the concept of pooling has been
introduced, which involves the combining of multiple cells in one and the same assay
[20c]. In the Bommarius approach, a Monte Carlo simulation model was developed
and applied to a library of b-galactosidase mutants obtained by saturation mutagen-
esis at residues around the binding pocket, the goal being to increase activity toward
fucosides [20d]. The results of pooling proved to be significantly better those of the
non-pooled protocol, with the overall process requiring less screening labor. In
another approach, the so-called Phizicky concept of screening predefined pools of
proteins, originally derived from genome sequencing for linking biological activities
with previously undefined proteins, was adapted to suit the requirements of directed
evolution [54]. When applying it to the laboratory evolution of a stereoselective
enoate-reductase, in which 20 single residue sites around the binding pocket were
initially targeted by saturation mutagenesis to be followed by iterative steps (ISM), the
screening effort was reduced by more than 50%. These two studies demonstrate that
the application of pooling strategies definitely pays off.
In the case of selection [19], the experimental platform is designed so that the host
organism experiences a growth advantage because it harbors an improved mutant.
Theoretically, only colonies having positively evolved enzymes will then appear on the
agar plates, meaning that the usual vast numbers of junk mutants never appear.
Thus, if a suitable selection system can be devised, large libraries on the DNA (gene)
level can be considered (107–1010), with such scenarios having great advantages.
Selection systems based on antibiotic resistance are well known, for example, when
evolving b-lactamases. However, it is not a trivial task to develop reliable selection
systems in a general way to suit all needs in biotechnology, as, for example, in the case
of laboratory evolution of stereoselectivity. Why should the survival rate of a host
154j 5 Directed Evolution of Enzymes
organism be higher just because it harbors an evolved enantioselective enzyme
variant?
Two different approaches to address this question have been described, both
making use of lipases in the kinetic resolution of chiral esters. In the first report the
researchers used a mutant library in an aspartate auxotroph E. coli that was
supplemented with an L-aspartate ester of a chiral alcohol, (S)-isopropylidene
glycerol (IPG), assisted by an inhibitory compound, namely, a phosphonate ester
incorporating the enantiomeric (R)-IPG [140]. It was known that WT LipA from
Bacillus subtilis shows low (S)-selectivity in the kinetic resolution of IPG-butyrate, and
the goal of the selection study was to invert stereoselectivity. Three selection rounds
led to the identification of a mutant showing noticeable diastereoselectivity in the
desired sense, corresponding to a selectivity factor of E ¼ 8.5 (R).
In the second approach, a general scheme for genetic selection for laboratory
evolution was proposed based on the use of isosteric pseudo-enantiomers in
mixtures [141]. The basic idea is to mimic kinetic resolution in such a way that
positive and negative components are placed in a single system according to the
absolute configuration of the chiral compounds of interest, thereby ensuring
simultaneous selection for activity and enantioselectivity. For example, to evolve
(R)-selectivity, the (R)-substrate needs to contain a positive component serving as a
potential energy source for the host organism, thereby promoting growth following
the desired enantioselective cleavage. At the same time the pseudo-enantiomeric (S)-
substrate is designed so as to harbor a negative component, with the respective
cleavage generating a toxic compound as a poison for the organism (Figure 5.22).
The concept was tested by implementing the reactions shown in Scheme 5.3 in
which acetic acid (10) and fluoroacetic acid (12) are the cleavage products acting as an
energy source and poison, respectively [141].
Using the lipase from Candida antarctica B (CALB) as the enzyme and the yeast
Pichia pastoris as the host, proof-of-principle was achieved, the goal being reversal of
enantioselectivity observed for WT CALB as a catalyst in the hydrolytic kinetic
resolution of rac-8 [141]. Saturation mutagenesis at appropriate sites aligning the
binding pocket of CALB was applied (CASTing), and after appropriate follow-up steps
the first plate harboring 70–80 colonies was considered. Ten of the largest colonies
were picked and analyzed, eight proving to have the desired selectivity favoring the
reaction of (S)-8 with formation of (R)-9, and one being inactive. The best variant
displayed a selectivity factor of E ¼ 10 when carrying out a control hydrolytic kinetic
resolution using a 1 : 1 mixture (racemate) of the acetate (R)- and (S)-8, favoring the
energy enzyme energy

(R)-component (R )-product + growth
source source
toxic enzyme toxic growth

(S)-component (S)-product +
analog analog inhibition
Figure 5.22 Genetic selection system for laboratory evolution of enantioselectivity in a kinetic
resolution [141].
5.6 Screening Versus Selection j155
O
O O O O
H2O
O + OH
lipase
O OH
(S)-8 (R)-9 10
O
O O O O
H2O F
O + OH
lipase
F
O OH
(R)-11 (S)-9 12
Scheme 5.3 Model system for genetic selection based on a mixture consisting of enantiomer (S)-8,
which provides acetic acid (10) as an energy source for the host organism, and a pseudo-enantiomer
(R)-11, which generates fluoroacetic acid (12) as a poison [141].
formation of (R)-9. Since WT CALB favors the hydrolysis of (R)-8 with preferential
formation of (S)-9, it can be seen that the goal defined in terms of reversal of
enantioselectivity was indeed reached. It remains to be seen if selection systems of
this kind can be used in bacterial hosts such as E. coli, and whether larger libraries can
be targeted. It should be noted critically that even if such perspectives become reality,
the problem of devising selection systems for stereoselectivity is far from being
solved in a general way.
Other approaches to selection need to be viewed in a different sense, because these
are based on specific genotype–phenotype linkages and do not involve survival of
organisms. They make use of various display systems, as, for example, in ribozyme
display, phage display, bacterial surface display, and yeast display [19, 20]. In nature,
compartmentalization of genes in cells ensures the genotype–phenotype linkage.
Inspired by this phenomenon, droplet-based strategies for in vitro compartmentaliza-
tion have been developed [142]. Using oil, detergents, and emulsifiers, emulsions with
droplets having diameters of about 2 mm are easily constructed, mimicking natural
cells. The members of a gene library are then partitioned into microscopic compart-
ments in such a way that one copy comes to exist in each droplet, and in vitro protein
expression is used to synthesize multiple copies of the encoded protein. Since the
droplets are so small, large libraries (106–108 members) can be generated and handled.
Several variations of this technique have been reported, including systems in which the
droplets are the sole connection between genotype and phenotype and DNA display in
such droplets using covalent or non-covalent links or even beads [142]. It remains to be
seen how this technique will develop in the future, such as, for example, in the directed
evolution of enantioselective enzymes, and how it compares with other approaches.
Sometimes the desired mutants are isolated or enriched directly from a suspen-
sion of the corresponding display-species, while in other systems analytical methods
such as fluorescence activated cell sorting (FACS) need to be invoked. The numbers
describing the size of the libraries in all of these systems are high (106–1010)
and, indeed, successful examples of directed evolution of proteins have been
156
reported [143]. Nevertheless, the question of generality and ease of performance

needs to be addressed. For example, phage display is primarily suited to handling the
binding properties of proteins. This is probably why attempts to apply phage display
in the directed evolution of enantioselective enzymes (as in the case of a lipase) have
not been exceedingly rewarding so far [144]. In contrast, a promising approach to
yeast surface display for selecting horseradish peroxidase (HRP) variants as catalysts
in the stereoselective dimerization of tyrosinol was successfully developed, with the
synthesis of fluorescence-labeled enantiomeric tyrosinal substrates being neces-
sary [145]. One was immobilized on the surface of live yeast cells (together with the
library of enzyme variants), while the other was supplied in solution and used by the
active variants to label those cells that express the active enzyme variants. A library of
about 2 106 HRP variants was generated by saturation mutagenesis at five positions
next to the binding pocket. The library was subjected to FACS analysis, once for
enantioselectivity favoring D-tyrosinol over the L-enantiomer, and once favoring the
reverse. Variants with up to eightfold altered enantioselectivity toward L/D-tyrosinol
were identified, including those with reversed stereoselectivity [145].
A system for single-cell super high-throughput screening to identify enantiose-
lective esterase-mutants as catalysts in the hydrolytic kinetic resolution of chiral
esters also deserves mention [146]. The FACS-based concept requires the two
enantiomeric esters derived from rac-2-methyldecanoic acid to be labeled, each with
a different fluorescent dye. Appropriately labeled (R)- and (S)-tyramide esters as
substrates were subjected to hydrolytic kinetic resolution; peroxidase-mediated radical
formation ensured the immediate covalent attachment of the reaction products to the
surface of an esterase-proficient bacterial cell (E. coli). The system allows 108 cells, and
thus this number of clones, to be screened within a few hours. The goal was to reverse
the sense of enantioselectivity. In a proof-of-principle study, epPCR was applied at a
mutation rate corresponding to the introduction of 2-4 amino acid substitutions per
enzyme molecule. The best hit was found to display a selectivity factor of E ¼ 16 in
favor of the (R)-enantiomer (WT: E ¼ 1.2 in favor of the (S)-substrate) [146]. More
work is required here, especially to see how far enantioselectivity can be boosted with
epPCR or with other gene mutagenesis methods such as ISM using saturation sites
defined by 5–6 amino acid positions that otherwise pose insurmountable screening
problems if 95% library coverage is desired. Reduced amino acid alphabets should be
considered in such case [2b, 40, 49] (Table 5.1 and Section 5.4.2).
5.7
Engineering Enzyme Stability
Sufficient thermostability of enzymes is a prerequisite for their application as

catalysts in such areas as synthetic organic chemistry, polymer production, deter-
gents, pollution cleanup, and paper manufacture [147a], but also when manufactur-
ing biosensors and other bionanotechnological devices [147b]. For these and other
reasons, hundreds of reports have appeared describing various different ways to
enhance the kinetic or thermodynamic thermostability of proteins, including de novo
5.7 Engineering Enzyme Stability j157
design using site-specific mutagenesis [63a,b] and directed evolution [63b–d]. In the
1990s, numerous studies appeared in which multiple rounds of epPCR, saturation
mutagenesis at hot spots, and/or DNA shuffling were employed successfully for
enhancing thermostability and/or robustness toward hostile organic solvents [72c,d].
In many but not all cases thermostability increased without compromising enzyme
activity. It is difficult to compare the results systematically because different ther-
mostability indices were used. Most authors reported changes in the melting
temperature (Tm), in the so-called T50-value (temperature at which 50% of enzyme
activity is lost after heating for a given period of time), or in the half-life at a defined
temperature. Sometimes an increase in half-life is not accompanied by a shift in Tm
in other cases a tendency to aggregate plays a role and therefore concentration may
influence the measured thermostability [148]. DTm- or DT50-values ranging between
5 and 20 C were described as being typical improvements in thermostability as a
result of directed evolution [63b–d]. Notably, the thermostability of proteins mea-
sured in various systems depends upon the environment due to several different
interactions with solvent, other proteins, and various biomolecules [63, 147, 148].
Thus, different assessments may arise, depending upon whether whole cells,
supernatants, or partially purified or fully purified enzymes are involved.
In addition to the use of epPCR and DNA shuffling in conjunction with appropriate
screening assays [63b,c], the so-called PROSIDE method in which large mutant
libraries are evaluated by Q7 selection procedures based on phage- or ribosome-
display systems also deserves mention [149], as does a procedure based on terminal
truncation [150], with the B-FITmethod based on ISM being yet another option [40, 61].
Some representative examples from the recent literature are highlighted here. The
first involves the esterase from Burkholderia gladioli (EstB) [151]. This 392 amino acid
enzyme catalyzes the mild hydrolysis of the acetyl group at the 3-position of
cephalosporins, for example, 13 ! 14 (Scheme 5.4), which is of considerable
industrial interest [152]. A more robust EstB variant was therefore desired.
CO2H CO2H
O O
N OAc H2O N OH
O O + CH3CO2H
NH2 S EstB NH2
N N S
H H
HO2C HO2C
13 (Cephalosporin C) 14 (Deacetylcephalosporin C)
Scheme 5.4 Esterase-catalyzed hydrolysis of cephalosporin C [151].
A crude but efficient colony filter assay based on the pH-change allowed large
numbers of clones to be evaluated for activity. In the initial round of epPCR
generating up to five amino acid substitutions per enzyme, about 1 million clones
were isolated and screened [151]. The transformants were plated on solid LB/Kan
medium at a density of about 500 colonies per plate. This means that 2000 plates were
prepared and assayed. The process led to the identification of eleven moderately
improved mutants (DTm up to 7.6 C), which were sequenced. To enforce further
158
improvement in thermostability, DNA shuffling of the mutants was attempted.

Unfortunately, even after optimizing every stage of the multistep process (fragment
sizes, elution protocols, DNA concentrations, and PCR conditions), no product was
obtained following the final reassembling PCR step. It was postulated that the high
GC content of EstB gene (74%) imposes a barrier for successful gene reassem-
bly [151]. An alternative strategy was then tested by combining the mutations of three
of the best previous mutants in various permutational ways. One of the mutants
showing highest expression level and activity as well as some improvement in
thermal stability was then used as a template for another round of epPCR in which
500 000 colonies were grown and assayed. As a result of these efforts, a variant
harboring 17 mutations was obtained. It shows an increase in thermostability of
DTm ¼ 13 C, and displays a slightly higher activity than the WT in the hydrolysis of
cephalosporin C (13) at room temperature. All of the mutations of EstB were found to
be located on or near the protein surface, as in several previous studies of other
enzymes utilizing the same or similar strategies [16, 63, 149].
In yet another epPCR-based approach, the thermostability of the NADPH-depen-
dent aldo-keto reductase from Penicillium citrinum was improved (62% retention of
activity after heat treatment versus 15% for WT), with saturation mutagenesis at a hot
spot showing that a different amino acid at that position leads to the best mutant [153].
Thus, this kind of check is simple and therefore advisable. Enantioselectivity in the
reduction of methyl 4-bromo-3-oxobutyrate as the model reaction was not compro-
mised – in fact a slight increase from 97.1% to 99.0% e.e. was observed.
Two studies targeting the evolution of thermostable variants of the recombinant
xylanase A from Bacillus subtilis (XylA) allow some insight concerning different
mutagenesis strategies. Xylanases (EC 3.2.18) catalyze the hydrolytic endo cleavage of
b-1,4 bonds of xylan polymers present in plant-derived xylans. Engineering ther-
mostable variants is of considerable industrial interest because their application as
biocatalysts can potentially improve the economics of processing lignocellulosic
materials for the production of chemicals and liquid fuels, among other industrial
uses. The initial study utilized systematic saturation mutagenesis at every amino acid
position of the 20-kDa protein [154a]. Following this massive search, positive point
mutations found in several improved mutants were combined using a special
technique with generation of a variant having nine point mutations and enhanced
thermostability amounting to DTm ¼ 35 C, which is an extremely high score. The
concept of systematic saturation mutagenesis at every residue of a protein, although
labor-intensive, was also applied to the evolution of hyperthermostability of a xylanase
[155a] and in the quest to obtain enantioselective Bacillus subtilis lipase mutants
[155b]. In the newer study, likewise focusing on xylanase A, a different approach was
taken that also proved to be successful [154b]. The first round of epPCR at an error
frequency of 3–4 base substitutions per kb per copy, and evaluating 10 644 colonies
using an on-plate color-based pre-screen, furnished eight single, four double, one
triple, and one quintuple mutant. The plasmids of five single mutants were then
mixed together and used as a template for performing the second round of epPCR,
with a total of 3413 being screened. Fifteen hits were identified and sequenced, four
of them being parent clones. All of the remaining seven variants in addition to the
first generation hits were then used in DNA shuffling, with evaluation of 717
5.7 Engineering Enzyme Stability j159
transformants leading to several variants that contained additional mutations.
A quadruple mutant Gln7His/Gly13Arg/Ser22Pro/Ser179Cys showing an increase
in thermostability by DTm ¼ 19 C was characterized by thermodynamic parameters
[154b]. A correlation was found between enhanced thermostability and decreasing
DCp. In the two studies, only one common point mutation was found, namely,
Ser179Cys, which appears to form a stabilizing disulfide bond. The older study
required considerably more experimental effort [153], but led to a more stable Xyl
variant. The interpretation of the data appears to be difficult, and effects relating to the
protein–solvent interface may play a crucial role.
In other newer studies different recombinant methods and strategies were used.
For example, in the case of the thermostabilization of the maltogenic amylase from
Bacillus thermoalkalophilus ET2, three rounds of DNA shuffling followed by recom-
bination of positive mutations provided several improved variants, the best one
showing a half-life of 568 min at 80 C compared to <1 min for the WT [156]. In the
thermostabilization of L-asparaginases from Erwinia chrysanthemi and Erwinia car-
otovora, StEP (Section 5.4.3) was applied followed by saturation mutagenesis at a hot
spot, leading to a variant showing DTm ¼ 9 C [157].
In a very different approach to thermostabilization, ISM was applied in a process
called B-FIT [40, 61]. To make the appropriate choice regarding the putative sites for
saturation mutagenesis, B-factors available from X-ray data were utilized that
indicate sites of maximal flexibility. In the case of the lipase from Bacillus subtilis
(LipA), five residues having the highest B-factors were employed. Five iterative
steps, in which thermostability was measured in supernatants following heat
treatment to eliminate other proteins, provided a markedly stabilized variant
(increase by 40 C). An MD study suggested that a communicating network of
amino acids connected by H-bonds and salt bridges on the surface of the protein is
responsible for the enhanced thermostability or thermoresistance to aggregation
[158]. In another B-FIT study, the thermostability of the epoxide hydrolase from
Aspergillus niger (ANEH) was also notably increased (DT5060 ¼ 21 C) [60b, 138].
B-FITcan also be used in the controlled thermodestabilization of proteins, which in
rare cases is of practical importance [159]. Table 5.3 highlights some recent directed
evolution studies of protein thermostabilization.
In conclusion, epPCR, DNA shuffling, PROSIDE, and B-FIT are reliable methods
in the quest to enhance thermostabilization of enzymes (see Table 5.3 for recent
examples). Since thermostability often correlates with robustness toward hostile
organic solvents [16], thermostabilization studies may fulfill a dual purpose [63].
However, in most studies regarding robustness toward hostile organic solvents, high
concentrations were not actually used. Indeed, directed evolution studies pertaining
to enzymatic reactions in pure organic solvents, as practiced, for example, in the
lipase-catalyzed acylation of chiral alcohols, have not been reported to date. Since
such reactions are known to be slower than normal enzymatic transformations
in water [164], directed evolution offers practical opportunities in the future. In a
systematic study of a fungal laccase, various cosolvents, including ethanol
and acetonitrile, at fairly high concentrations were tested [165]. In this extensive
investigation, a combination of DNA shuffling, epPCR using polymerases with
different biases, and saturation mutagenesis proved to be particularly successful.
160
Table 5.3 Recent examples of enzyme thermostabilization based on laboratory evolution.
Enzyme Method Reference
Esterase (Burkholderia gladioli) epPCR [151]

Aldo-keto reductase (Penicillium citrinum) epPCR, saturation mutagenesis [153]
Xylanase (Bacillus subtilis) Saturation mutagenesis at all [154a]
residues
Xylanase (Bacillus subtilis) epPCR, DNA shuffling [154b]
Amylase (Bacillus thermoalkalophilus) DNA shuffling [156]
L-Asparaginase (Erwinia chrysanthemi) StEP [157]
Lipase (Bacillus subtilis) B-FIT [40, 61]
Epoxide hydrolase (Aspergillus niger) B-FIT [60b]
Phosphite dehydrogenase epPCR, saturation mutagenesis [160]
Pseudomonas stutzeri)
Cellulase (CBHII) SCHEMA [161]
Lactate oxidase (Aerococcus viridans) epPCR, DNA shuffling [162]
Pyruvate decarboxylase (Pdc1) epPCR [163]
After five rounds of mutagenesis, a laccase variant was evolved that was capable of
resisting a wide range of cosolvents at concentrations as high as 50 vol.%. It was
found that the intrinsic laccase characteristics, including the redox potential and
geometry of the catalytic coppers, changed only to a small extent as the result of the
protein engineering. This basic research is important in view of the use of laccases in
remediation of environmental contaminants such as polychlorinated biphenyls, in
pulp-kraft bleaching, and in the textile industries [165].
Interpretation of thermostabilization on a molecular level is not a trivial task
because many different kinds of effects play a role, some being subtle. A general
empirical trend relates to the observation that most mutations occur on the surface of
the proteins, often in somewhat flexible loops. The formation of new H-bonds, salt
bridges, or disulfide bonds were invoked in many cases, but in most studies
numerous point mutations have evaded interpretation [63]. This may have two
causes: (i) such point mutations are superfluous, not actually contributing to
thermostabilization; (ii) a given mutation induces a remote effect. Indeed, a recent
study has shown that such remote effects may result from the evolution of a
communicating amino acid network [158]. In a rare case epPCR was found to
induce point mutations in the interior of an enzyme (xylanase) with the introduction
of two cysteines that do not form a disulfide bond [166].
5.8
Engineering Enzyme Stereoselectivity
5.8.1
General Remarks
The catalytic synthesis of chiral intermediates for the production of pharmaceuti-

cals [167], fragrances, and other compounds is of fundamental importance in
5.8 Engineering Enzyme Stereoselectivity j161
synthetic organic chemistry and biotechnology [1, 2b,g,h], with the major options
being synthetic catalysts, enzymes, or combinations thereof [168]. Biocatalysis is
already playing a significant role in this endeavor, but enzymes have suffered
traditionally from limited substrate scope and often from poor stereoselectivity. As
delineated in Section 5.5.2, the most extensively studied case of directed evolution of
enantioselective enzymes as a new approach to asymmetric catalysis [15, 17]
pertains to the lipase from P. aeruginosa (PAL) as a catalyst in the hydrolytic kinetic
resolution of the chiral ester rac-1. The combination of epPCR, saturation muta-
genesis at hot spots, and DNA shuffling led to an improved variant (E ¼ 51 versus
E ¼ 1.2 for WT PAL) [15, 17, 56, 107]; recent work based on ISM led to far better
results (E ¼ 594), while requiring considerably less screening (10 000 versus 50 000
transformants [60a]. The reason why ISM is so efficient was traced to two
phenomena [49, 60a, 106, 158]: none of the mutations are superfluous, and
pronounced cooperative effects are operating between the point mutations within
a randomization site on the one hand and between sets of mutations along the
evolutionary pathway. Moreover, in the case of the ISM-based PAL study [60], the
best mutant proved to be an excellent biocatalyst in the hydrolytic kinetic resolution
of eight structurally different chiral esters, most of which are not accepted by
WT PAL, without the need to perform additional mutagenesis experiments.
Organic chemists require stereoselective catalysts not just for a single compound
but for a reasonable range of substrates. Table 5.4 contains recent cases of directed
evolution of stereoselective enzymes utilizing various methods and strategies;
Sections 5.8.2–5.8.5 feature illustrative examples.
5.8.2
Hydrolases
5.8.2.1 Nitrilase from an Environmental Sample

In an industrial project, a nitrilase identified in a genomic library as a catalyst in the
hydrolytic desymmetrization of dinitrile 15 with formation of the acid (R)-16 (87.5%
e.e.) needed in the synthesis of the cholesterol-lowering therapeutic drug LipitorÒ (7)
showed unsatisfactory enantioselectivity (94.5% e.e.) as well as product inhibition
(Scheme 5.5) [176]. At the required substrate concentration, the rate and enantios-
electivity (87.5% e.e.) decreased. To establish an industrial process with stringent
requirements, systematic saturation mutagenesis experiments at all 330 positions in
the enzyme were performed. The total number of clones amounted to 31 585 [176],
which were screened by the MS-based assay using multiplexing mass spectrometry
[136c]. This provided a mutant Ala190His that showed no product inhibition even at
3-M substrate concentration and high enantioselectivity (98.5% e.e.) [176].
5.8.2.2 Epoxide Hydrolase from Aspergillus niger

The concept of ISM in the form of CASTing for enhanced stereoselectivity was first
applied to the epoxide hydrolase from Aspergillus niger (ANEH) as a catalyst in the
hydrolytic kinetic resolution of rac-17 (Scheme 5.6) [59]. WT ANEH favors the
formation of (S)-18, but enantioselectivity is poor (E ¼ 4.6) [59, 105]. Although
Jacobsens chiral salen complexes solved the synthetic problem years ago [195], this
162
Table 5.4 Recent examples of directed evolution of enzymes having enhanced stereoselectivity
and/or altered substrate scope.
Lipase Pseudomonas aeruginosa epPCR, DNA shuffling, saturation [17]

mutagenesis
Lipase Pseudomonas aeruginosa Saturation mutagenesis (ISM) [60a]
Lipase Pseudomonas aeruginosa Saturation mutagenesis (CASTing) [169]
Lipase Burkholderia cepacia Saturation mutagenesis (SIMPLEX) [66b]
Lipase Candida antartica, B Family shuffling [170]
Lipase Candida antartica, B Circular permutation [91a]
Lipase Candida antartica, A Saturation mutagenesis (ISM) [112]
Esterase Pseudomonas fluorescens Saturation mutagenesis [57b]
Esterase Pseudomonas fluorescens epPCR [171]
Esterase Bacillus subtilis Saturation mutagenesis (CASTing) [172]
Esterase (Burkholderia gladioli) Saturation mutagenesis (ISM) [115]
Epoxide hydrolase Aspergillus niger epPCR [105]
Epoxide hydrolase Aspergillus niger Saturation mutagenesis (ISM) [59]
Epoxide hydrolase Aspergillus niger Saturation mutagenesis [173]
Epoxide hydrolase Agrobacterium epPCR, DNA shuffling [174]
radiobacter
Epoxide hydrolase Aspergillus niger ISM, saturation mutagenesis [175]
Nitrilase (genomic library) Saturation mutagenesis (at every [176]
residue)
Phosphotriesterase Saturation mutagenesis [177]
Aldolase (D-2-keto-3-deoxy-6- epPCR, DNA shuffling [178]
phosphor-gluconate aldolase
Aldolase (N-acetylneuraminate epPCR, DNA shuffling [179]
lyase)
Aldolase (N-acetylneuraminate epPCR, saturation mutagenesis [180]
lyase)
Aldolase (tagatose-1,6-biphosphate DNA shuffling, saturation mutagenesis [181]
aldolase)
Aldolase (2-deoxy-D-ribose-5-phos- epPCR [182]
phate aldolase)
Aldolase (2-keto-3-deoxy-6- epPCR, DNA shuffling [183]
phosphogalactonate)
Aldolase (2-keto-3-deoxy-6- Saturation mutagenesis [184]
phosphogalactonate)
Benzoylformate decarboxylase epPCR [185]
(Pseudomonas putida)
b-Keto ester reductase (Penicillium epPCR, saturation mutagenesis [153]
citrinum)
Ketoreductase (environmental ProSAR [186]
sample)
Enoate-reductase (YgjM) Saturation mutagenesis (ISM) [54]
Baeyer–Villiger monooxygenase epPCR [187]
(CHMO)
Table 5.4 (Continued)
Baeyer–Villiger monooxygenase Saturation mutagenesis (CASTing) [53]

(CPMO)
Baeyer–Villiger monooxygenase Saturation mutagenesis (CASTing) [124]
(PAMO)
Baeyer–Villiger monooxygenase Saturation mutagenesis at second- [188]
(PAMO) sphere residuum
Baeyer–Villiger monooxygenase Saturation mutagenesis at allostery- [189]
(PAMO) sensitive sites
Baeyer–Villiger monooxygenase epPCR, saturation mutagenesis [190]
(Pseudomonas fluorescens)
P450 monooxygenases epPCR, saturation mutagenesis, DNA [191]
shuffling
Horse radish peroxidase epPCR (FACS) [192]
Monoamine oxidase (Aspergillus Mutator strain [193]
niger)
Arylmalonate decarboxylase (Borde- Saturation mutagenesis (ISM) [113]
tella bronchiseptica)
Oxynitrilase (Hevea brasiliensis) epPCR [194]
HO O-
O
OH
OH OH F CH3
H2O several N
NC CN nitrilase NC CO 2H steps CH 3
15 (R)-16 O
HN
Scheme 5.5 Formation of the chiral compound (R)-16 as an intermediate in the synthesis of the
cholesterol-lowering therapeutic drug 7(LipitorÒ ), catalyzed by a mutant nitrilase [176].
O O HO OH
H2O
+
PhO ANEH PhO PhO
r ac-17 (R)-17 (S)-18
Scheme 5.6 Hydrolytic kinetic resolution of rac-17 catalyzed by ANEH [59].

164
transformation was employed as a model reaction to test and compare different

strategies in directed evolution. In a preceding study, the traditional use of epPCR had
proven to be only marginally successful, with the E-value increasing to only 11,
although 20 000 transformants had to be screened [105]. The X-ray structure of WT
ANEH reveals the presence of a narrow tunnel as the binding pocket [196], which may
be the reason why this enzyme appears to be a difficult case.
Six CASTsites A–F were identified with the help of the X-ray structure (Figure 5.23),
with each one being composed of two or three amino acid positions [59]. It was
already known that Asp192 initiates rate- and stereoselectivity-determining nucle-
ophilic attack at the less substituted C-atom of epoxide rac-17 [196].
The variant with the highest enantioselectivity originated from the library at site B
(E ¼ 14), which was then used as a template to randomize site C. Successive visits at
sites D, F, and E provided the best mutant LW202, which showed unusually
high enantioselectivity in favor of (S)-18 (E ¼ 115), the order being arbitrarily cho-
sen [59]. As summarized in Figure 5.24, five sets of mutations accumulated stepwise,
leading to LW202 characterized by five sets of mutations adding up to a total of nine point
mutations. Since enantioselectivity was already so high, site A was not considered, nor
were other pathways explored. Thermostability was not impaired by the mutational
changes. A total of about 20 000 transformants were screened, which happens to be the
same number needed in the earlier study based on epPCR [105]. Since the results are
dramatically superior, E ¼ 115 versus E ¼ 11, it was concluded that ISM constitutes an
exceptionally efficient way to generate smart mutant libraries.
The source of enhanced enantioselectivity was uncovered by a mechanistic and
structural study consisting of kinetics, induced docking, MD simulations, inhibition
experiments, and the crystal structure of the best mutant LW202 [197]. This showed
Figure 5.23 Binding pocket and CAST sites A–F of the epoxide hydrolase from Aspergillus niger
(ANEH) [59] based on its X-ray structure [196].
120 E = 115 (Mutant LW202:
9 mutations)
110
100
90
80 E (Thr317Trp/Thr318Val)
Selectivity factor (E )
70
60
50
40 E = 35
30 F (Leu249Tyr; 244/245 stay)

E = 24
E = 21
D (Cys350Val; 349 stays)
20 E = 14 C (Met329Pro/Leu330Tyr)
B (Leu215Phe/Ala217Asp/Arg219Ser)
10
WT-ANEH
Figure 5.24 Iterative CASTing in the evolution of enantioselective ANEH mutants as catalysts in
the hydrolytic kinetic resolution of rac-17 [59].
that the reaction of the disfavored (R)-substrate is essentially shut down, as in an ideal
kinetic resolution. The X-ray data of LW202 is the first and thus far only case of
structural elucidation of an evolved enantioselective enzyme, and it proved to be
revealing when comparing it to the WTANEH structure. The folds of the two enzymes
are essentially identical, but the shapes of the respective binding pockets take on
dramatically different forms, which allowed interpretation on a molecular level [197].
The respective model led to the prediction that the best mutant LW202 should be a
good catalyst for the hydrolytic kinetic resolution of essentially any mono-substituted
epoxide, which was substantiated by studying several different substrates, all of them
showing reasonable rates and good to excellent degrees of enantioselectivity.
To reveal the reason for efficacy when applying ISM, the fitness landscape relevant
on going from WT ANEH to the final mutant LW202 in the five-step evolutionary
pathway was constructed experimentally [106]. Of the 5! ¼ 120 experimentally
determined pathways, 55 proved to be energetically favored, lacking any local minima
(Figure 5.25). This is a high score, especially given that new mutational amino acid
exchange events are not involved. In the case of disfavored pathways, backtracking by
one step puts the evolutionary process back on a favored trajectory [106].
Analysis of the empirical results based on DGz values in kinetic resolution showed
that in each of the 120 trajectories strong cooperative effects operating between the
sets of mutations are involved, none of which are superfluous. Figure 5.26 features
the positive epistatic effects in the original pathway B ! C ! D ! F ! E. Clearly, the
interactions between the sets of mutations are more than additive. This type of
166
Figure 5.25 Fitness landscape featuring the indicate examples of energetically disfavored
120 pathways leading from WT ANEH to the trajectories due to the presence of local
best mutant LW202 [106]. Green pathways are minima. For a schematic view of all 120
examples of energetically favored trajectories trajectories at each evolutionary stage see
lacking local minima, while red pathways Reference [106].
–2.5
∆∆G‡i (first set of mutations)
∆∆G‡j (second set of mutations)
∆∆G‡k (third set of mutations)
–2.0 ∆∆G‡l (fourth set of mutations)
∆∆G‡m (fifth set of mutations)
Expected additive incr. (∆∆G‡i + ∆∆G‡j + ∆∆G‡k +∆∆G‡l + ∆∆G‡m)
Experimentally found increment (∆∆G‡exp)
∆∆G‡ (kcal·mol–1)
–1.5
–1.0
–0.5
0.0
B BC BCD BCDF BCDFE

combinations of sets of mutations
Figure 5.26 Epistatic interactions operating between sets of mutations accumulated in the
evolutionary pathway B ! C ! D ! F ! E [106].
analysis constitutes a kind of quality control, which was also applied to one of the ISM
studies regarding thermostability [40, 158].
5.8.2.3 Esterase from Bacillus subtilis

Kinetic resolution of acetates derived from tertiary alcohols represents a particular
challenge. In a study concerning the esterase from Bacillus subtilis (BS2), saturation
mutagenesis was applied at a three amino acid CAST site Glu188, Ala190, and
Met193 [172]. Although only a small section of the protein sequence was screened
(1100 transformants following NNK-based randomization), remarkable results were
observed for substrates of the type rac-19 (Scheme 5.7). One and the same library
harbored mutants with enhanced (R)-selectivity and reversed (S)-selectivity. Under-
going iterative steps can be expected to provide further improvements.
BS2 E188D E R > 100

BS2 WT E R = 42
HO CF3 AcO CF3
AcO CF3
(R)-20a (S)-20
Buffer, DMSO
HO CF3 AcO CF3
(R,S)-19
(S)-20a (R)-20
BS2 E188W/M193C E S = 64
Scheme 5.7 Hydrolytic kinetic resolution of rac-19 using WT and mutants of the esterase from
Bacillus subtilis (BS2) [172].
5.8.3
Oxidases
5.8.3.1 Monoamine Oxidase from Aspergillus niger

Directed evolution of the monoamine oxidase from Aspergillus niger constitutes
another impressive example of protein engineering of enantioselective
enzymes [193]. In nature such enzymes catalyze the racemization of amino acids.
Both (R)- and (S)-selective monoamine oxidases are known, catalyzing selectively the
reaction of either the (R)- or (S)-enantiomer. This property was exploited in an
ingenious deracemization scheme with the formation of enantiomerically highly
enriched amines (Scheme 5.8) [193a]. Accordingly, achiral reducing agents such as
NaBH4, NaB(CN)H3, or H3NBH3 were employed in the presence of either an (R)- or
(S)-selective monoamine oxidase. Since the activity of the WT enzymes is very low
with substrates of the type phenylethylamine (rac-21) and enantioselectivity is only
moderate to good, directed evolution was applied. Following several cycles of
mutagenesis using the E. coli XL1-Red mutator strain and transformation of the
plasmid library in E. coli, a total of 150 000 bacterial colonies were assayed for activity
using a colorimetric pre-screen followed by conventional e.e.-characterization of the
best hits. This procedure led to the identification of mutants displaying excellent
activity and enantioselectivity (>98% e.e.) [193a]. Later it was shown that some of the
mutants display a remarkably broad substrate range, encompassing primary and
secondary amines, and even the difficult class of tertiary amines [193b]. This is yet
168
NH2
(S)-selective
MAO
Ph CH3
(S)-21
NH
Ph CH3
H3N . BH3
22
NH2
Ph CH3
(R)-21
Scheme 5.8 Ingenza system for deracemizing amines [193].
another example of fairly broad substrate acceptance of mutants generated by directed

evolution without the need to perform new mutagenesis/screening experiments. The
simple two-step, one-pot process has been commercialized by Ingenza. The experi-
mental platform has been extended to include other types of mutagenesis protocols.
5.8.3.2 Baeyer–Villiger Monooxygenases

Baeyer–Villiger monooxygenases (BVMOs) such as cyclohexanone monooxygenase
(CHMO) have long been used as catalysts in the oxidative desymmetrization of
prochiral ketones or kinetic resolution of racemic substrates with formation of
enantiomerically enriched esters or lactones [198]. They are flavin-dependent and
require NADPH for reducing the oxidized form of flavin, which is the reason that
whole cells are generally employed, with oxygen from air being the oxidant. The use
of isolated CHMO in conjunction with an NADPH in vitro regeneration system,
based on a secondary alcohol dehydrogenase and isopropanol as the reductant, is not
feasible due to the relative instability of these BVMOs [198]. In cases of poor
enantioselectivity, directed evolution utilizing conventional strategies based on
epPCR have proven to be quite successful [187a], including the use of a single
mutant as a catalyst for a wide range of structurally different prochiral ketones in a
whole-cell process (Table 5.5) [187b]. Stereoselective CHMO mutants have also been
evolved for the sulfoxide-forming oxidation of thio-ethers [136b].
Industrial processes based on CHMO or other BVMOs have not been announced
to date, possibly due to the instability of these enzymes. Therefore, the report of the
first thermostable BVMO, phenylacetone monooxygenase (PAMO), attracted a great
deal of attention [200], as did its crystal structure [201], the first of a BVMO.
Unfortunately, PAMO accepts essentially only phenylacetone and a few related
linear ketones. Rational design led to an improved mutant, but only 2-phenyl-
substituted cyclohexanone derivatives were accepted [199]. In contrast, saturation
mutagenesis at two second-sphere CAST residues (positions 437 and 440 separately)
(Figure 5.27) provided an active mutant Pro440Phe that accepts a remarkably wide
range of 2-alkyl substituted cyclohexanone derivatives, with kinetic resolution
leading to high E-values [188]. Thermostability was not compromised. In another
study concerning the same enzyme (PAMO), a completely new approach to directed
evolution utilizing saturation mutagenesis was reported [189]. Based on rational
Table 5.5 Oxidative desymmetrization catalyzed by CHMO mutant 1-K2-F5 (Phe432Ser) using air
as the oxidant in a whole cell process [187b].
Substrate e.e. (%)
O
O 94
O
O
Cl O Cl 99
O
O
O 91
O
O
O 97
O
O O 78
O
O
O 96
O
O O
O
>99
O O
O
>99
O O
O
>99
OH OH
O O
O
99a)
H3C OH HO CH3
a) Unpublished data of C. Clouthier, M.M. Kayser, and M.T. Reetz.
design utilizing the X-ray structure of PAMO, a two-residue site was chosen for
saturation mutagenesis that was predicted to induce allostery-based rearrangement
of the enzyme with concomitant re-shaping of the binding pocket. This novel strategy
proved to be successful – particularly broad substrate scope and high enantioselec-
tivities were evolved [189]. An informatics-based form of CASTing was also applied
to PAMO, a reduced amino acid alphabet being used in the randomization of a
4-residue site [124]. This approach is also amenable to sites composed of 5–10
residues. ISM in the form of iterative CASTing at a loop aligning the binding pocket
still needs to be explored.
170
Figure 5.27 Illustration of the putative binding pocket of PAMO based on the induced-fit docking
model, featuring second-sphere residues Pro437 and Pro440 (blue) [188]. Phenylacetone and the
loop segment 441–444 are shown in cyan and red, respectively.
5.8.3.3 Cytochrome P450 Monooxygenases

Cytochrome P450 enzymes catalyze the oxidative hydroxylation of a wide range of
substrates, the respective CH-activation being of considerable synthetic potential [1].
Since the late 1940s certain steroids have been regio- and stereoselectively hydrox-
ylated using various different strains, a process that requires extensive screening;
unfortunately, these searches are not always successful. Moreover, similar problems
arise when considering other substrates such as terpenes or simple synthetic
compounds. Directed evolution of P450 enzymes constitutes a way to solve these
problems [191]. Early efforts were directed to the control of regioselectivity in simple
alkanes, with P450-BM3 being the enzyme under study. Surrogate substrates contain-
ing p-nitrophenoxy moieties for easy screening using a UV–vis plate reader were often
employed. Subsequently, some of these mutants were tested for enantioselectivity in
other substrates such as phenylacetic acid esters, with stereoselective hydroxylation at
the methylene group being observed [191b]. Directed evolution of P450 monoox-
ygenases with the purpose of controlling regio- and enantioselectivity directly in
mutagenesis experiments has not been reported until recently. Using the CAST/ISM
approach and testosterone as the model compound, 2b- and 15b-selective mutants of
P450-BM3 were evolved on an optional basis with diastereoselectivity being 100%
[191f]. In contrast, the starting enzyme delivers a 1:1mixture of the two regioisomers.
5.8.4
Reductases
5.8.4.1 b-Keto Ester Reductase from Penicillium citrinum

The b-keto ester reductase (KER) from Penicillium citrinum was known to convert,
with 97% e.e., 4-bromo-3-oxobutyrate (23) into (S)-24, a useful chiral intermediate
needed in the synthesis of several therapeutic drugs, an example being an inhibitor of
3-hydroxy-3-methylglutaryl coenzyme A [153]. Unfortunately, WT KER is not ther-
mostable enough for industrial applications, and the observed enantioselectivity also
had to be improved. Two rounds of epPCR followed by saturation mutagenesis at one
of the hot spots provided a mutant showing 99% e.e. as well as sufficiently enhanced
thermostability (Scheme 5.9) [153].
O OH
KER-
Br CO2Me Br CO2Me
mutant
23 (S)-24 (99% ee)
Scheme 5.9 Enantioselective reduction catalyzed by a KER-mutant [153].
5.8.4.2 Ketoreductase from an Environmental Sample

Directed evolution of an enantioselective ketoreductase (KRED) was performed to
access a chiral alcohol needed in the synthesis of the leukotriene receptor antagonist
SingulairÒ (27) (Merck/USA) for treating asthma and allergies [186a]. The conven-
tional industrial chemical route involves asymmetric reduction of ketone 25 with
formation of the key intermediate (S)-26 using stoichiometric amounts of a chiral
boron reagent followed by several appropriate steps. To develop a greener industrial
process, a biocatalytic reduction was sought using a KRED. Such an enzyme was
found, but it was not robust enough to tolerate the presence of an organic solvent, and
the enantioselectivity, although respectable, was nevertheless insufficient for prac-
tical purposes. Utilizing three rounds of evolution based on the computational aid
ProSAR (Section 5.5.3) [130] in the presence of THF or toluene for substrate
solubilization and isopropanol for NADPH regeneration, a mutant was isolated
that fulfilled all prerequisites for an industrial production on a large scale [186a]. At
high substrate loading (100 g l1) in the presence of 70% organic solvent, a mutant
showing sufficiently high enzyme activity and 99.9% e.e. was evolved, allowing the
biocatalytic process to be run at >200 kg scale (Scheme 5.10). Stringent comparison
with the alternative chemical route based on organic synthetic steps proved the
superiority of biocatalysis in the industrial production [186a]. A similar approach was
used in the evolution of an enantioselective KRED needed in the asymmetric
reduction of tetrahydrothiophene-3-one, the chiral alcohol being an intermediate
in the synthesis of an antibacterial pharmaceutical [186b].
5.8.4.3 Enoate-Reductase YqjM

In a study concerning methodology development, ISM in the form of CASTing was
applied to the enoate-reductase YqjM, a flavin-dependent protein from the family of
Old Yellow Enzymes, which was used as a catalyst in the enantioselective reduction
of the olefinic bonds in 3-substituted cyclohexenone-derivatives [54]. Both (R)- and
(S)-selective variants (>97% e.e.) were evolved for several prochiral substrates.
Substrate acceptance was also broadened by the ISM approach. The significance of
this study not only touches on the actual results in terms of substrate acceptance and
enantioselectivity but on the fact that strategies were developed regarding the
172
O OMe
O
Cl N
25
KRED-mutant
O OMe
OH
Cl N
(S)- 26
(99.9% ee)
Na O
OH
O
S
Cl N
27
R
Singulair
Scheme 5.10 Biocatalytic route to intermediate (S)-26 on the >200 kg scale needed in the
preparation of SingulairÒ (27) [186a].
question of grouping randomization residues, pooling protocols in screening, and

implementing a fast quality control of mutant libraries (amino acid bias) [54].
5.8.5
CC Bond-Forming Enzymes
5.8.5.1 Aldolases
Aldolases have been shown to be of particular utility in synthetic organic chemistry
because of exquisite control of stereoselectivity leading to relatively complex products
without the need to engage in labor-intensive protective group technology [202].
However, the fact that many substrates of practical interest are not accepted or react
with low stereoselectivity constitutes a serious limitation. Directed evolution has
made major strides in solving this problem, as summarized in a recent review [203].
Seminal studies based on epPCR and DNA shuffling addressed the problem of
limited substrate acceptance of D-2-keto-3-deoxy-6-phospho-gluconate (KDPG) aldol-
ase, leading to mutants with essentially complete diastereoselectivity in the reaction
of chiral aldehydes not accepted by WT KDPG [178]. In an extension of this work, the
N-acetylneuraminate lyase (Neu5Ac aldolase) from E. coli was subjected to directed
evolution to expand its catalytic activity for enantiomeric forms of the usual substrates
to include N-acetyl-L-mannosamine and L-arabinose with formation of the synthet-
ically valuable products L-sialic acid and L-3-deoxy-L-manno-oct-2-ulosonic acid [179].
Rather than evolving aldolase mutants that selectively accept stereoisomers of
substrates, in a different conceptual approach the configuration of the chiral starting
aldehyde is maintained, the goal being to evolve opposite diastereoselectivity. In an
initial study, three rounds of DNA shuffling using tagatose-1,6-biphosphate aldolase
provided a mutant that indeed showed opposite stereoselectivity in the aldol addition
of 28 to 30 [181] (Scheme 5.11). The concept has been generalized to include Neu5Ac
aldolase [180]. In a likewise synthetically valuable endeavor, saturation mutagenesis
was applied to 2-keto-3-deoxy-6-phosphogalactonate (KDPGal) aldolase, with the
evolved mutant allowing the replacement of 3-deoxy-D-arabino-heptulosonic acid
7-phosphate synthase [184]. This enabled fermentation-based production of valuable
3-dehydro-shikimic acid at 13 g l1 in 6.5% molar yield from glucose.
OH O
wild-type 2-O PO
OPO32-
3
O
OH OH
HO OPO32-
28 29
directed evolution
OHC
OPO32-
OH O
OH
mutant 2-
OPO 32-
30 O3PO
OH OH
31
Scheme 5.11 Reversal of diastereoselectivity of tagatose-1,6-biphosphate aldolase [181].
5.8.5.2 Benzoylformate Decarboxylase from Pseudomonas putida

Thiamine diphosphate-dependent benzoylformate decarboxylase (BFD) from
Pseudomonas putida has been shown to be an efficient catalyst in the acyloin-like
reaction of several aromatic aldehydes 32 with acetaldehyde (33) leading to acyloins
34 (Scheme 5.12) [185]. However, enantioselectivity is not always optimal, and ortho-
substituted benzaldehyde derivatives are not accepted. Therefore, epPCR
was applied, and after screening 40 000 transformants using benzaldehyde and
acetaldehyde as reaction partners and a colorimetric on-plate assay, two mutants
R O R OH
CH3CHO
H
32 33 34
Scheme 5.12 Stereoselective acyloin-like reactions catalyzed by the benzoylformate decarboxylase

(BFD) [185].
174
showing high activity and enhanced enantioselectivity (>95% e.e.) were identified,
namely, Leu476Gln and Met365Leu/Leu461Ser. These mutants accept ortho-substi-
tuted benzaldehyde derivatives in the reaction with acetaldehyde, generally with high
enantioselectivity (95–99% e.e.), but are also excellent catalysts in the analogous
reactions with meta- and para-substituted benzaldehyde derivatives, with the enan-
tioselectivity being higher than in the case of WT BFD [185].
5.9
Summary and Outlook
Directed evolution of enzymes has emerged as a general and reliable way to engineer
many properties of enzymes, including substrate acceptance (rate), stereoselectivity,
robustness toward hostile organic solvents, and thermostability, which are particularly
important parameters in white biotechnology [2]. Thus, it seems that the traditional
limitations of enzymes as catalysts in synthetic organic chemistry and biotechnology
no longer exist, because the first 15 years of research have shown that some degree of
catalyst improvement is always possible, irrespective of the applied gene mutagenesis
method, such as epPCR, saturation mutagenesis at hot spots or at all amino acid
positions in an enzyme, or recombinant protocols such as DNA family shuffling.
Moreover, other problems of classical biocatalysis that may arise can also be solved,
including elimination of product inhibition and reduction of undesired side-reactions.
This assessment reflects the power of laboratory evolution as a protein engineering
method, yet the crucial current issue concerns efficacy [2, 18, 26, 49, 58, 69, 100–102].
Especially, industrial biotechnology requires methods and strategies for rapid
directed evolution that enable reliable timeframes for the discovery and production
of effective biocatalysts needed for new tasks. Several such methods and strategies
enabling the generation of high-quality libraries, generally flanked by computational
aids, have been reported since 2005. Thus far the most practical approach is structure-
guided saturation mutagenesis, especially in an iterative manner (ISM) [2b,g,h, 40,
54, 59, 120]. It is a straightforward and labor-saving concept for generating high-
quality mutant libraries for enhancing stereoselectivity, broadening substrate
acceptance (rate), and increasing thermostability, the method being devoid of any
patent (IP) restrictions. The question of how to optimally group single residues
identified on the basis of structural information into putative randomization sites
has not been answered in final form, but thus far sites consisting of two or more
amino acid positions are recommended due to the increased probability of
encountering cooperative effects. Combined with the possibility of using reduced
amino acid alphabets based on appropriate codon degeneracies and the concept of
pooling, a major step toward solving the screening and therefore the numbers
problem in laboratory evolution has been taken. It is advisable to perform short-cut
quality controls when using saturation mutagenesis as delineated in a detailed
protocol [54]. A limitation of this knowledge-driven semi-rational approach
becomes apparent when structural data or homology models are lacking. Especially
in these cases, such approaches as epPCR and recombinant methods retain their
References j175
importance in laboratory evolution. Along a different line, it has been shown that
chaperonin overexpression promotes genetic variation and enzyme evolution, as in
the ability of E. coli GroEL/GroES chaperonins to buffer destabilizing and adaptive
mutations [204]. Lifting the thermodestabilization constraint may make directed
evolution more efficient, although at the expense of added laboratory input.
Parallel to increasing the efficacy of probing protein sequence space, developing
improved screening assays or selection systems is also a crucial endeavor [19, 20].
Hopefully, the Trapp method [137] for multiplexing GC and HPLC will be adapted to
the needs of directed evolution in the near future, allowing the routine screening of
2000–4000 transformants per day, which in combination with the generation of
small but smart mutant libraries would solve the screening problem in a general and
cheap way [2b, 138].
Yet another challenge concerns the observation that in some cases enzymes cannot
be expressed efficiently in the usual workhorses of molecular biology such as E. coli,
which means that other expression systems need to be developed that are suitable for
directed evolution studies and for enzyme production. Examples are horse radish
peroxidase (yeast/FACS) [192] and hydroxynitrilases (Pichia pastoris) [205].
Finally, the methodsand strategiesdescribed hereincan be applied to areas other than
the usual enzyme catalysis used in synthetic organic chemistry and white biotechnol-
ogy, including metabolic pathway engineering inwhite biotechnology (designer bugs)
and in green biotechnology, production of therapeutic peptides and proteins (red
biotechnology), and microbial-based degradation of environmentally harmful organic
compounds (gray biotechnology). They can also be used to tune computationally
designed enzymes which thus far have proven to be of low activity [123g, 206].
References
1 (a) Tao, J., Lin, G.-Q., and Liese, A. (2009) prolific source of catalysts for asymmetric
Biocatalysis for the Pharmaceutical reactions. Angew. Chem. Int. Ed., 50,
Industry: Discovery, Development, and 138–174; (c) Romero, P.A. and Arnold,
Manufacturing, John Wiley & Sons, Ltd., F.H. (2009) Exploring protein fitness
Chichester; (b) Gotor, V., Alfonso, I. and landscapes by directed evolution. Nat.
Garcıa-Urdiales, E. (eds) (2008) Rev. Mol. Cell Biol., 10, 866–876; (d)
Asymmetric Organic Synthesis with Turner, N.J. (2009) Directed evolution
Enzymes, Wiley-VCH Verlag GmbH, drives the next generation of biocatalysts.
Weinheim; (c) Faber, K. (2004) Nat. Chem. Biol., 5, 567–573; (e) J€ackel,
Biotransformations in Organic Chemistry, C., Kast, P. and Hilvert, D. (2008) Protein
5th edn, Springer, Berlin; (d) design by directed evolution. Annu. Rev.
Bommarius, A.S. and Riebel- Biophys. Biomol. Struct., 37, 153–173;
Bommarius, B.R. (2004) Biocatalysis: (f) Bershtein, S. and Tawfik, D.S. (2008)
Fundamentals and Applications, Wiley- Advances in laboratory evolution of
VCH Verlag GmbH, Weinheim. enzymes. Curr. Opin. Chem. Biol., 12,
2 (a) Lutz, S. and Bornscheuer, U.T. (2009) 151–158; (g) Reetz, M.T. (2008) Directed
Protein Engineering Handbook, vols 1–2, evolution as a means to engineer
Wiley-VCH Verlag GmbH, Weinheim; enantioselective enzymes, in Asymmetric
(b) Reetz, M.T. (2011) Laboratory Organic Synthesis with Enzymes (eds V.
evolution of stereoselective enzymes: A Gotor, I. Alfonso, and E. Garcıa-Urdiales),
176
Wiley-VCH Verlag GmbH, Weinheim, random and site-directed mutagenesis.

pp. 21–63; (h) Reetz, M.T. (2010) Enzyme Biochem. Biophys. Res. Commun., 192,
engineering by directed evolution, in 15–21.
Manual of Industrial Microbiology and 12 Liao, H., Mckenzie, T., and Hageman, R.
Biotechnology, 3rd edn (eds H. Zhao et al..) (1986) Isolation of a thermostable
ASM Press, Washington D.C., pp. enzyme variant by cloning and selection
466–479. in a thermophile. Proc. Natl. Acad. Sci.
3 Mills, D.R., Peterson, R.L., and U.S.A., 83, 576–580.
Spiegelman, S. (1967) An extracellular 13 Chen, K. and Arnold, F.H. (1993) Tuning
Darwinian experiment with a the activity of an enzyme for unusual
self-duplicating nucleic acid molecule. environments: sequential random
Proc. Natl. Acad. Sci. U.S.A., 58, 217–224. mutagenesis of subtilisin E for catalysis
4 Joyce, G.F. (2007) Forty years of in vitro in dimethylformamide. Proc. Natl. Acad.
evolution. Angew. Chem., 119, Sci. U.S.A., 90, 5618–5622.
6540–6557; Angew. Chem. Int. Ed., 46, 14 Stemmer, W.P.C. (1994) Rapid evolution
6420–6436. of a protein in vitro by DNA shuffling.
5 Francis, J.C. and Hansche, P.E. (1972) Nature, 370, 389–391.
Directed evolution of metabolic pathways 15 (a) Reetz, M.T., Zonta, A.,
in microbial populations. I. Modification Schimossek, K., Liebeton, K., and
of the acid phosphatase ph optimum in S. Jaeger, K.-E. (1997) Creation of
Cerevisiae. Genetics, 70, 59–73. enantioselective biocatalysts for organic
6 Hall, B.G. (1977) Number of mutations chemistry by in vitro evolution. Angew.
required to evolve a new lactase function Chem., 109, 2961–2963; Angew. Chem. Int.
in Escherichia coli. J. Bacteriol., 129, Ed. Engl., 36, 2830–2832; (b) Reetz, M.T.
540–543. (1999) Strategies for the development of
7 Eigen, M. and Gardiner, W. (1984) enantioselective catalysts. Pure Appl.
Evolutionary molecular engineering Chem., 71, 1503–1509.
based on RNA replication. Pure Appl. 16 Arnold, F.H. (1998) Design by directed
Chem., 56, 967–978. evolution. Acc. Chem. Res., 31, 125–131.
8 Matsumura, M. and Aiba, S. (1985) 17 Reetz, M.T. (2004) Controlling the
Screening for thermostable mutant of enantioselectivity of enzymes by directed
kanamycin nucleotidyltransferase by the evolution: practical and theoretical
use of a transformation system for a ramifications. Proc. Natl. Acad. Sci.
thermophile Bacillus stearothermophilus. U.S.A., 101, 5716–5722.
J. Biol. Chem., 260, 15298–15303. 18 Lutz, S. and Patrick, W.M. (2004) Novel
9 Bryan, P.N., Rollence, M.L., methods for directed evolution of
Pantoliano, M.W., Wood, J., enzymes: quality, not quantity. Curr.
Finzel, B.C., Gilliland, G.L., Opin. Biotechnol., 15, 291–297.
Howard, A.J., and Poulos, T.L. (1986) 19 (a) Taylor, S.V., Kast, P., and Hilvert, D.
Proteases of enhanced stability: (2001) Investigating and engineering
characterization of a thermostable enzymes by genetic selection. Angew.
variant of subtilisin. Proteins, 1, 326–334. Chem., 113, 3408–3436; Angew. Chem.
10 (a) Leung, D.W., Chen, E., and Int. Ed., 40, 3310–3335; (b) Lin, H. and
Goeddel, D.V. (1989) A method for Cornish, V.W. (2002) Screening and
random mutagenesis of a defined DNA selection methods for large-scale analysis
segment using a modified polymerase of protein function. Angew. Chem., 114,
chain reaction. Technique, 1, 11–15; 4580–4606; Angew. Chem. Int. Ed., 41,
(b) Cadwell, R.C. and Joyce, G.F. (1994) 4402–4425; (c) Boersma, Y.L., Dr€oge,
Mutagenic PCR. PCR Methods Appl., 3, M.J., and Quax, W.J. (2007) Selection
S136–S140. strategies for improved biocatalysts.
11 Zhang, H.Y., Zhang, J., Lin, L., Du, W.Y., FEBS J., 274, 2181–2195.
and Lu, J. (1993) Enhancement of the 20 (a) Reymond, J.-L. (2005) Enzyme Assays -
stability and activity of aspartase by High-Throughput Screening, Genetic
References j177
Selection and Fingerprinting, Wiley-VCH Comb. Chem. High Throughput Screen.,
Verlag GmbH, Weinheim; (b) Reetz, M.T. 10, 197–217.
(2004) Screening for enantioselective 27 Neylon, C. (2004) Chemical and
enzymes, in Enzyme functionality: Design, biochemical strategies for the
Engineering and Screening (ed. A. randomization of protein encoding DNA
Svendsen), Marcel Dekker, New York, sequences: library construction methods
pp. 559–598; (c) Reymond, J.-L., Fluxa, for directed evolution. Nucleic Acids Res.,
V.S., and Maillard, N. (2009) Enzyme 32, 1448–1459.
assays. Chem. Commun., 34–46; (d) 28 (a) Sylvestre, J., Chautard, H.,
Polizzi, K.M., Parikh, M., Spencer, C.U., Cedrone, F., and Delcourt, M. (2006)
Matsumura, I., Lee, J.H., Realff, M.J., and Directed evolution of biocatalysts. Org.
Bommarius, A.S. (2006) Pooling for Process Res. Dev., 10, 562–571;
improved screening of combinatorial (b) Cirino, P.C., Mayer, K.M. and
libraries for directed evolution. Umeno, D. (2003) Generating mutant
Biotechnol. Prog., 22, 961–967. libraries using error-prone PCR, in
21 Selifonova, O. and Schellenberger, V. Directed Evolution Library Creation:
(2003) Evolution of microorganisms Methods and Protocols (eds F.H. Arnold
using mutator plasmids, in Directed and G. Georgiou), Humana Press,
Evolution Library Creation: Methods and Totowa, NJ, pp. 3–10.
Protocols (eds F.H. Arnold and G. 29 (a) Fujii, R., Kitaoka, M., and Hayashi, K.
Georgiou), Methods in Molecular (2006) Error-prone rolling circle
Biology, vol. 231, Humana Press Inc., amplification: the simplest random
Totowa, NJ, pp. 45–52. mutagenesis protocol. Nat. Protocols, 1,
22 Wong, T.S., Roccatano, D., and 2493–2497; (b) Fujii, R., Kitaoka, M., and
Schwaneberg, U. (2007) Challenges Hayashi, K. (2004) One-step random
of the genetic code for exploring sequence mutagenesis by error-prone rolling
space in directed protein evolution. circle amplification. Nucleic Acids Res.,
Biocatal. Biotransform., 25, 229–241. 32, e145.
23 Zaccolo, M., Williams, D.M., Brown, 30 Fromant, M., Blanquet, S., and Plateau, P.
D.M., and Gherardi, E. (1996) An (1995) Direct random mutagenesis of
approach to random mutagenesis of gene-sized DNA fragments using
DNA using mixtures of triphosphate polymerase chain reaction. Anal.
derivatives of nucleoside analogous. Biochem., 224, 347–353.
J. Mol. Biol., 255, 589–603. 31 Vanhercke, T., Ampe, C., Tirry, L., and
24 Claveau, S., Sasseville, M., and Denolf, P. (2005) Reducing mutational
Beauregard, M. (2004) Alcohol-mediated bias in random protein libraries. Anal.
error prone PCR. DNA Cell Biol., 23, Biochem., 339, 9–14.
789–795. 32 Wong, T.S., Tee, K.L., Hauer, B., and
25 Stratagene (2003) GeneMorph II Schwaneberg, U. (2004) Sequence
Random Mutagenesis Kit: Instruction saturation mutagenesis (SeSaM): a novel
Manual, Stratagene, La Jolla, CA, U.S.A., method for directed evolution. Nucleic
http://www.bioc.rice.edu/bios576/ Acids Res., 32, e26.
proteng/200550.pdf. 33 (a) Wong, T.S., Roccatano, D., Loakes, D.,
26 (a) Eggert, T., Reetz, M.T., and Tee, K.L., Schenk, A., Hauer, B., and
Jaeger, K.-E. (2004) Directed evolution by Schwaneberg, U. (2008) Transversion-
random mutagenesis: a critical enriched sequence saturation
evaluation, in Enzyme Functionality – mutagenesis (SeSaM-Tv þ ): a random
Design, Engineering, and Screening mutagenesis method with consecutive
(ed. A. Svendsen), Marcel Dekker, nucleotide exchanges that complements
New York, pp. 375–390; (b) Tee, K.L. and the bias of error-prone PCR. Biotechnol. J.,
Schwaneberg, U. (2007) Directed 3, 74–82; (b) Shivange, A.V.,
evolution of oxygenases: screening Marienhagen, J., Mundhada, H.,
system, success stories and challenges. Schenk, A. and Schwaneberg, U. (2009)
178
Advances in generating functional of a serine protease can generate

diversity for directed protein evolution. enzymes with increased activities and
Curr. Opin. Chem. Biol., 13, 19–25. altered primary specificities.
34 Murakami, H., Hohsaka, T., and Sisido, Biochemistry, 32, 6250–6258; (d) Warren,
M. (2002) Random insertion and deletion M.S. and Benkovic, S.J. (1997)
of arbitrary number of bases for codon- Combinatorial manipulation of three key
based random mutation of DNAs. Nat. active site residues in glycinamide
Biotechnol., 20, 76–81. ribonucleotide transformylase. Protein
35 (a) Pikkemaat, M.G. and Janssen, D.B. Eng., 10, 63–68.
(2002) Generating segmental mutations 40 (a) Reetz, M.T. and Carballeira, J.D.
in haloalkane dehalogenase: a novel part (2007) Iterative saturation mutagenesis
in the directed evolution toolbox. Nucleic (ISM) for rapid directed evolution of
Acids Res., 30, e35; (b) Jones, D.D. (2005) functional enzymes. Nat. Protocols, 2,
Triplet nucleotide removal at random 891–903; (b) CASTER and B-FITTER are
positions in a target gene: the tolerance of available free of charge on the author's
TEM-1 b-lactamase to an amino acid homepage (http://www.kofo.mpg.de/en/
deletion. Nucleic Acids Res., 33, e80. research/organic-synthesis).
36 Erdogan, E., Jones, R.J., Matzlin, P., 41 Firth, A.E. and Patrick, W.M. (2005)
Hanna, M.H., Smith, S.M.E., and Statistics of protein library construction.
Salerno, J.C. (2005) A novel Bioinformatics, 21, 3314–3315.
mutagenesis method generating high 42 Firth, A.E. and Patrick, W.M. (2008)
yields of closed circular mutant DNA with GLUE-IT and PEDEL-AA: new
one primer per mutant. Mol. Biotechnol., programmes for analyzing protein
30, 21–30. diversity in randomized libraries. Nucleic
37 Fujii, R., Kitaoka, M., and Hayashi, K. Acids Res., 36, W281–W285.
(2006) RAISE: a simple and novel method 43 Hughes, M.D., Nagel, D.A., Santos, A.F.,
of generating random insertion and Sutherland, A.J., and Hine, A.V. (2003)
deletion mutations. Nucleic Acids Res., 34, Removing the redundancy from
e30. randomised gene libraries. J. Mol. Biol.,
38 Hogrefe, H.H., Cline, J., Youngblood, 331, 973–979.
G.L., and Allen, R.M. (2002) Creating 44 Evans, P.M. and Liu, C. (2005) SiteFind: a
randomized amino acid libraries with the software tool for introducing a restriction
QuikChangeÒ multi site-directed site as a marker for successful site-
mutagenesis kit. BioTechniques, 33, directed mutagenesis. BMC Mol. Biol.,
1158–1165. 6, 22.
39 (a) Dube, D.K. and Loeb, L.A. (1989) 45 Miyazaki, K. and Takenouchi, M. (2002)
Mutants generated by the insertion of Creating random mutagenesis libraries
random oligonucleotides into the active using megaprimer PCR of whole
site of the beta-lactamase gene. plasmid. BioTechniques, 33, 1033–1038.
Biochemistry, 28, 5703–5707; 46 Sanchis, J., Fernandez, L.,
(b) Teplyakov, A.V., Van Der Laan, J.M., Carballeira, J.D., Drone, J.,
Lammers, A.A., Kelders, H., Kalk, K.H., Gumulya, Y., H€obenreich, H.,
Misset, O., Mulleners, L.J.S.M., and Kahakeaw, D., Kille, S., Lohmer, R.,
Dijkstra, B.W. (1992) Protein engineering Peyralans, J.J.-P., Podtetenieff, J.,
of the high-alkaline serine protease PB92 Prasad, S., Soni, P., Taglieber, A., Wu, S.,
from Bacillus alcalophilus: functional and Zilly, F.E., and Reetz, M.T. (2008)
structural consequences of mutation at Improved PCR method for the creation
the S4 substrate binding pocket. Protein of saturation mutagenesis libraries in
Eng., 5, 413–420; (c) Graham, L.D., directed evolution: application to
Haggett, K.D., Jennings, P.A., Le difficult-to-amplify templates. Appl.
Brocque, D.S., Whittaker, R.G., and Microbiol. Biotechnol., 81, 387–397.
Schober, P.A. (1993) Random 47 (a) Kirsch, R.D. and Joly, E. (1998) An
mutagenesis of the substrate-binding site improved PCR-mutagenesis strategy for
References j179
two-site mutagenesis or sequence Baeyer–Villiger monooxygenases using
swapping between related genes. restricted CASTing. J. Org. Chem., 71,
Nucleic Acids Res., 26, 1848–1850; 8431–8437.
(b) Zheng, L., Baumann, U., and 54 Bougioukou, D.J., Kille, S., Taglieber, A.,
Reymond, J.-L. (2004) An efficient one- and Reetz, M.T. (2009) Directed evolution
step site-directed and site-saturation of an enantioselective enoate-reductase:
mutagenesis protocol. Nucleic Acids Res., Testing the utility of iterative saturation
32, e115. mutagenesis. Adv. Synth. Catal., 351,
48 Rui, L., Kwon, Y.M., Fishman, A., 3287–3305.
Reardon, K.F., and Wood, T.K. (2004) 55 (a) Climie, S., Ruiz-Perez, L.,
Saturation mutagenesis of toluene ortho- Gonzalez-Pacanowska, D.,
monooxygenase of Burkholderia cepacia Prapunwattana, P., Cho, S.-W.,
G4 for enhanced 1-naphthol synthesis Stroud, R. and Santi, D.V. (1990)
and chloroform degradation. Appl. Saturation site-directed mutagenesis of
Environ. Microbiol., 70, 3246–3252. thymidylate synthase. J. Biol. Chem., 265,
49 (a) Reetz, M.T., Kahakeaw, D. and 18776–18779; (b) Gabor, E.M. and
Lohmer, R. (2008) Addressing the Janssen, D.B. (2004) Increasing the
numbers problem in directed evolution. synthetic performance of penicillin
ChemBioChem, 9, 1797–1804; acylase PAS2 by structure-inspired semi-
(b) Reetz, M.T., Kahakeaw, D., and random mutagenesis. Protein Eng.,
Sanchis, J. (2009) Shedding light on the Design & Select., 17, 571–579; (c) Rui, L.,
efficacy of laboratory evolution based on Cao, L., Chen, W., Reardon, K.F. and
iterative saturation mutagenesis. Wood, T.K. (2004) Active site engineering
Molecular BioSystems, 5, 115–122. of the epoxide hydrolase from
50 (a) Patrick, W.M. and Firth, A.E. (2005) Agrobacterium radiobacter AD1 to
Strategies and computational tools for enhance aerobic mineralization of cis-
improving randomized protein libraries. 1,2-dichloroethylene in cells expressing
Biomol. Eng., 22, 105–112; (b) Bosley, A.D. an evolved toluene ortho-
and Ostermeier, M. (2005) Mathematical monooxygenase. J. Biol. Chem., 279,
expressions useful in the construction, 46810–46817; (d) Schmitzer, A.R.,
description and evaluation of protein Lepine, F., and Pelletier, J.N. (2004)
libraries. Biomol. Eng., 22, 57–61; Combinatorial exploration of the catalytic
(c) Denault, M. and Pelletier, J.N. (2007) site of a drug-resistant dihydrofolate
Protein Library Design and Screening: reductase: creating alternative functional
Working Out the Probabilities, in Protein configurations. Protein Eng., Design &
Engineering Protocols, vol. 352 (eds K.M. Select., 17, 809–819; (e) Antikainen, N.M.,
Arndt and K.M. M€ uller), Humana Press, Hergenrother, P.J., Harris, M.M.,
Totowa, NJ, pp. 127–154. Corbett, W. and Martin, S.F. (2003)
51 Iyidogan, P. and Lutz, S. (2008) Altering substrate specificity of
Systematic exploration of active site phosphatidylcholine-preferring
mutations on human deoxycytidine phospholipase c of Bacillus cereus by
kinase substrate specificity. Biochemistry, random mutagenesis of the headgroup
47, 4711–4720. binding site. Biochemistry, 42, 1603–1610.
52 Van den Brulle, J., Fischer, M., 56 Reetz, M.T., Wilensek, S., Zha, D., and
Langmann, T., Horn, G., Waldmann, T., Jaeger, K.-E. (2001) Directed evolution of
Arnold, S., Fuhrmann, M., Schatz, O., an enantioselective enzyme through
OConnell, T., OConnell, D., combinatorial multiple cassette
Auckenthaler, A., and Schwer, H. (2008) mutagenesis. Angew. Chem., 113,
A novel solid phase technology for high- 3701–3703; Angew. Chem. Int. Ed., 40,
throughput gene synthesis. 3589–3591.
BioTechniques, 45, 340–343. 57 (a) Horsman, G.P., Liu, A.M.F.,
53 Clouthier, C.M., Kayser, M.M., and Reetz, Henke, E., Bornscheuer, U.T. and
M.T. (2006) Designing new Kazlauskas, R.J. (2003) Mutations in
180
distant residues moderately increase the the range of substrate acceptance of

enantioselectivity of Pseudomonas enzymes: combinatorial active-site
fluorescens esterase towards methyl saturation test. Angew. Chem., 117,
3-bromo-2-methylpropanoate and ethyl 4264–4268; Angew. Chem. Int. Ed., 44,
3-phenylbutyrate. Chem. Eur. J., 9, 4192–4196.
1933–1939; (b) Park, K.L., Morley, K.L., 63 (a) Oshima, T. (1994) Stabilization of
Horsman, G.P., Holmquist, M., Hult, K., proteins by evolutionary molecular
and Kazlauskas, R.J. (2005) Focusing engineering techniques. Curr. Opin.
mutations into the P. fluorescens esterase Struct. Biol., 4, 623–628; (b) ÓFagain, C.
binding site increases enantioselectivity (2003) Enzyme stabilization - recent
more effectively than distant mutations. experimental progress. Enzyme Microb.
Chem. Biol., 12, 45–54; (c) Kazlauskas, R.J. Technol., 33, 137–149; (c) Eijsink, V.G.H.,
(2003) Creating enantioselective Gaseidnes, S., Borchert, T.V., and Van
hydrolases for organic synthesis: Den Burg, B. (2005) Directed evolution of
combining rational and random methods. enzyme stability. Biomol. Eng., 22, 21–30;
Chem. Listy, 97, 329–337. (d) Bommarius, A.S. and Broering, J.M.
58 Paramesvaran, J., Hibbert, E.G., (2005) Established and novel tools to
Russell, A.J., and Dalby, P.A. (2009) investigate biocatalyst stability. Biocatal.
Distributions of enzyme residues Biotransform., 23, 125–139.
yielding mutants with improved 64 (a) Chaparro-Riggers, J.F., Polizzi, K.M.,
substrate specificities from two different and Bommarius, A.S. (2007) Better
directed evolution strategies. Protein library design: data-driven protein
Eng., Design & Select., 22, 401–411. engineering. Biotechnol. J., 2, 180–191;
59 Reetz, M.T., Wang, L.-W., and in part (b) Gustafsson, C., Govindarajan, S., and
Bocola, M. (2006) Directed evolution of Minshull, J. (2003) Putting engineering
enantioselective enzymes: iterative cycles back into protein engineering:
of CASTing for probing protein- bioinformatic approaches to catalyst
sequence space. Angew. Chem., 118, design. Curr. Opin. Biotechnol., 14,
1258–1263, erratum, p. 2556; Angew. 366–370.
Chem. Int. Ed. Engl., 45, 1236–1241, 65 (a) Steipe, B., Schiller, B., Pl€
uckthun, A.,
erratum, p. 2494. and Steinbacher, S. (1994) Sequence
60 (a) Reetz, M.T., Prasad, S., Carballeira, statistics reliably predict stabilizing
J.D., Gumulya, Y., and Bocola, M. (2010) mutations in a protein domain. J. Mol.
Iterative saturation mutagenesis Biol., 240, 188–192; (b) Dai, M.,
accelerates laboratory evolution of Fisher, H.E., Temirov, J., Kiss, C.,
enzyme stereoselectivity: rigorous Phipps, M.E., Pavlik, P., Werner, J.H.,
comparison with traditional approaches. and Bradbury, A.R.M. (2007) The
J. Am. Chem. Soc., 132, 9144–9152; creation of a novel fluorescent protein by
(b) Gumulya, Y. and Reetz, M.T. (2011) guided consensus engineering. Protein
Enhancing the thermal robustness of an Eng., Design & Select., 20, 69–79;
enzyme by directed evolution: Least (c) Amin, N., Liu, A.D., Ramer, S.,
favorable starting points and inferior Aihle, W., Meijer, D., Metin, M.,
mutants can map superior evolutionary Wong, S., Gualfetti, P., and
pathways. ChemBioChem., 12, 2502–2510. Schellenberger, V. (2004) Construction
61 Reetz, M.T., Carballeira, J.D., and of stabilized proteins by combinatorial
Vogel, A. (2006) Iterative saturation consensus mutagenesis. Protein Eng.,
mutagenesis on the basis of B factors Design & Select., 17, 787–793;
as a strategy for increasing protein (d) Hamamatsu, N., Aita, T., Nomiya, Y.,
thermostability. Angew. Chem., 118, Uchiyama, H., Nakajima, M., Husimi, Y.,
7909–7915; Angew. Chem. Int. Ed., 45, and Shibanaka, Y. (2005) Biased
7745–7751. mutation-assembling: an efficient
62 Reetz, M.T., Bocola, M., Carballeira, J.D., method for rapid directed evolution
Zha, D., and Vogel, A. (2005) Expanding through simultaneous mutation
References j181
accumulation. Protein Eng., Design & 73 Sieber, V., Martinez, C.A., and
Select., 18, 265–271. Arnold, F.H. (2001) Libraries of
66 (a) Rungpragayphan, S., Kawarasaki, Y., hybrid proteins from distantly related
Imaeda, T., Kohda, K., Nakano, H., and sequences. Nat. Biotechnol., 19,
Yamane, T. (2002) High-throughput, 456–460.
cloning-independent protein library 74 Higara, K. and Arnold, F.H. (2003)
construction by combining single- General method for sequence-
molecule DNA amplification with independent site-directed
in vitro expression. J. Mol. Biol., 318, chimeragenesis. J. Mol. Biol., 330,
395–405; (b) Koga, Y., Kato, K., Nakano, 287–296.
H., and Yamane, T. (2003) Inverting 75 Lee, S.H., Ryu, E.J., Kang, M.J.,
enantioselectivity of Burkholderia Wang, E.-S., Piao, Z., Choi, Y.J.,
cepacia KWI-56 lipase by combinatorial Jung, K.H., Jeon, J.Y.J., and Shin, Y.C.
mutation and high-throughput (2003) A new approach to directed gene
screening using single-molecule PCR evolution by recombined extension on
and in vitro expression. J. Mol. Biol., 331, truncated templates (RETT). J. Mol.
585–592. Catal., B: Enzym., 26, 119–129.
67 Joern, J.M. (2003) DNA shuffling, in 76 Ikeuchi, A., Kawarasaki, Y., Shinbata, T.,
Directed Evolution Library Creation (eds and Yamane, T. (2003) Chimeric gene
F.H. Arnold and G. Georgiou), Humana library construction by a simple and
Press Inc., Totowa, NJ, pp. 85–89. highly versatile method using
68 Powell, K.A., Ramer, S.W., del Cardayre, recombination-dependent exponential
S.B., Stemmer, W.P.C., Tobin, B., amplification. Biotechnol. Prog., 19,
Longchamp, P.F., and Huismann, G.W. 1460–1467.
(2001) Directed evolution and biocatalysis. 77 OMaille, P.E., Bakhtina, M., and
Angew. Chem., 113, 4068–4080; Angew. Tsai, M.D. (2002) Structure-based
Chem. Int. Ed., 44, 3948–3959. combinatorial protein engineering
69 Joern, J.M., Meinhold, P., and (SCOPE). J. Mol. Biol., 321, 677–691.
Arnold, F.H. (2002) Analysis of shuffled 78 Coco, W.M., Levinson, W.E.,
gene libraries. J. Mol. Biol., 316, 643–656. Crist, M.J., Hektor, H.J., Darzins, A.,
70 Zhao, H., Giver, L., Shao, Z., Affholter, Pienkos, P.T., Squires, C.H., and
J.A., and Arnold, F.H. (1998) Molecular Monticello, D.J. (2001) DNA shuffling
evolution by staggered extension process method for generating highly
(StEP) in vitro recombination. Nat. recombined genes and evolved enzymes.
Biotechnol., 16, 258–261. Nat. Biotechnol., 19, 354–359.
71 (a) Ostermeier, M., Shim, J.H., and 79 Gibbs, M.D., Nevalainen, K.M.H., and
Benkovic, S.J. (1999) A combinatorial Bergquist, P.L. (2001) Degenerate
approach to hybrid enzymes oligonucleotide gene shuffling (DOGS): a
independent of DNA homology. Nat. method for enhancing the frequency of
Biotechnol., 17, 1205–1209; (b) Lutz, S., recombination with family shuffling.
Ostermeier, M., and Benkovic, S.J. (2001) Gene, 271, 13–20.
Rapid generation of incremental 80 Bergquist, P.L., Reeves, R.A., and
truncation libraries for protein Gibbs, M.D. (2005) Degenerate
engineering using a-phosphothioate oligonucleotide gene shuffling (DOGS)
nucleotides. Nucleic Acids Res., 29, e16. and random drift mutagenesis (RNDM):
72 Kawarasaki, Y., Griswold, K.E., two complementary techniques for
Stevenson, J.D., Selzer, T., Benkovic, S.J., enzyme evolution. Biomol. Eng., 22,
Iverson, B.L., and Georgiou, G. (2003) 63–72.
Enhanced crossover SCRATCHY: 81 Reiter, B., Faschinger, A., Glieder, A.,
construction and high-throughput and Schwab, H. (2007) Random strand
screening of a combinatorial library transfer recombination (RSTR) for
containing multiple non-homologous homology-independent nucleic acid
crossovers. Nucleic Acids Res., 31, e126. recombination. J. Biotechnol., 129, 39–49.
182
82 (a) Stemmer, W.P.C. (1994) DNA glycosyltransferases of Arabidopsis

shuffling by random fragmentation and thaliana. J. Biol. Chem., 282,
reassembly: in vitro recombination for 15724–15731; (c) Hansen, E.H.,
molecular evolution. Proc. Natl. Acad. Sci. Osmani, S.A., Kristensen, C.,
U.S.A., 91, 10747–10751; (b) Stutzman- Møller, B.L., and Hansen, J. (2009)
Engwall, K., Conlon, S., Fedechko, R., Substrate specificities of family
Mcarthur, H., Pekrun, K., Chen, Y., 1 UGTs gained by domain swapping.
Jenne, S., La, C., Trinh, N., Kim, S., Phytochemistry, 70, 473–482.
Zhang, Y.-X., Fox, R., Gustafsson, C., and 89 Sharkey, M.A. and Engel, P.C. (2009)
Krebber, A. (2005) Semi-synthetic DNA Modular coenzyme specificity; a domain-
shuffling of aveC leads to improved swopped chimera of glutamate
industrial scale production of doramectin dehydrogenase. Proteins, 77, 268–278.
by Streptomyces avermitilis. Metab. Eng., 90 Meister, G.E., Kanwar, M., and
7, 27–37. Ostermeier, M. (2009) Circular
83 Herman, A. and Tawfik, D.S. (2007) permutation pf proteins, in Protein
Incorporating synthetic oligonucleotides Engineering Handbook, vol. 2 (eds S. Lutz
via gene reassembly (ISOR): a versatile and U.T. Bornscheuer), Wiley-VCH
tool for generating targeted libraries. Verlag GmbH, Weinheim, pp. 454–471.
Protein Eng., Design & Select., 20, 91 (a) Qian, Z. and Lutz, S. (2005) Improving
219–226. the catalytic activity of Candida antarctica
84 Coco, W.M., Encell, L.P., Levinson, W.E., lipase b by circular permutation. J. Am.
Crist, M.J., Loomis, A.K., Licato, L.L., Chem. Soc., 127, 13466–13467; (b) Yu, Y.
Arensdorf, J.J., Sica, N., Pienkos, P.T., and Lutz, S. (2010) Improved triglyceride
and Monticello, D.J. (2002) Growth factor transesterification by circular permuted
engineering by degenerate homoduplex Candida antarctica lipase B. Biotechnol.
gene family recombination. Nat. Bioeng., 105, 44–50.
Biotechnol., 20, 1246–1250. 92 Reitinger, S., Yu, Y., Wicki, J.,
85 Ness, J.E., Kim, S., Gottman, A., Ludwiczek, M., DAngelo, I., Baturin, S.,
Pak, R., Krebber, A., Borchert, T.V., Okon, M., Strynadka, N.C.J., Lutz, S.,
Govindarajan, S., Mundorff, E.C., and Withers, S.G., and Mcintosh, L.P. (2010)
Minshull, J. (2002) Synthetic shuffling Circular permutation of Bacillus
expands functional protein diversity by circulans xylanase: a kinetic and
allowing amino acids to recombine structural study. Biochemistry. 49 (11),
independently. Nat. Biotechnol., 20, 2464–2474.
1251–1255. 93 Lo, W.-C. and Lyu, P.-C. (2008) CPSARST:
86 Zha, D., Eipper, A., and Reetz, M.T. an efficient circular permutation search
(2003) Assembly of designed tool applied to the detection of novel
oligonucleotides as an efficient method protein structural relationships.
for gene recombination: a new tool in GenomeBiology, 9, R11.
directed evolution. ChemBioChem, 4, 94 Drummond, D.A., Iverson, B.L.,
34–39. Georgiou, G., and Arnold, F.H. (2005)
87 Li, Y. and Palmer, A.G. III (2009) Domain Why high-error-rate random
swapping in the kinase superfamily: mutagenesis libraries are enriched in
OSR1 joins the mix. Protein Sci., 18, functional and improved proteins. J. Mol.
678–681. Biol., 350, 806–816.
88 (a) Park, S.-H., Park, H.-Y., Sohng, J.K., 95 Tracewell, C.A. and Arnold, F.H. (2009)
Lee, H.C., Liou, K., Yoon, Y.J., and Directed enzyme evolution: climbing
Kim, B.-G. (2009) Expanding substrate fitness peaks one amino acid at a time.
specificity of gt-b fold glycosyltransferase Curr. Opin. Chem. Biol., 13, 3–9.
via domain swapping and high- 96 Bloom, J.D., Labthavikul, S.T., Otey, C.R.,
throughput screening. Biotechnol. and Arnold, F.H. (2006) Protein stability
Bioeng., 102, 988–994; (b) Cartwright, promotes evolvability. Proc. Natl. Acad.
A.M., Lim, E.-K., Kleanthous, C., and Sci. U.S.A., 103, 5869–5874.
Bowles, D.J. (2008) A kinetic analysis of 97 (a) Peisajovich, S.G. and Tawfik, D.S.
regiospecific glucosylation by two (2007) Protein engineers turned
References j183
evolutionists. Nat. Methods, 4, 991–994; S.G., and Matthews, B.W. (2001) A
(b) Bloom, J.D. and Arnold, F.H. (2009) structural view of the action of Escherichia
In the light of directed evolution: coli (lacZ) b-galactosidase. Biochemistry,
pathways of adaptive protein evolution. 40, 14781–14794.
Proc. Natl. Acad. Sci. U.S.A., 106, 105 Reetz, M.T., Torre, C., Eipper, A.,
9995–10000; (c) Depristo, M.A. (2007) Lohmer, R., Hermes, M., Brunner, B.,
The subtle benefits of being Maichele, A., Bocola, M., Arand, M.,
promiscuous: adaptive evolution Cronin, A., Genzel, Y., Archelas, A.,
potentiated by enzyme promiscuity. and Furstoss, R. (2004) Enhancing the
HFSP J., 1, 94–98.; (d) Kryazhimskiy, S., enantioselectivity of an epoxide hydrolase
Dushof, J., Bazykin, G.A., and Plotkin, J. by directed evolution. Org. Lett., 6,
B. (2011) Prevalence of epistasis in the 177–180.
evolution of influenza A surface proteins. 106 Reetz, M.T. and Sanchis, J. (2008)
PLoS Genetics, 7, e1001301. Constructing and analyzing the fitness
98 Eigen, M., McCaskill, J., and Schuster, P. landscape of an experimental
(1988) Molecular quasi-species. J. Phys. evolutionary process. ChemBioChem, 9,
Chem., 92, 6881–6891. 2260–2267.
99 (a) Kurtovic, S. and Mannervik, B. (2009) 107 Liebeton, K., Zonta, A., Schimossek, K.,
Identification of emerging quasi-species Nardini, M., Lang, D., Dijkstra, B.W.,
in directed enzyme evolution, Reetz, M.T., and Jaeger, K.-E. (2000)
Biochemistry, 48, 9330–9339; Directed evolution of an enantioselective
(b) Runarsdottir, A. and Mannervik, B. lipase. Chem. Biol., 7, 709–718.
(2010) A novel quasi-species of 108 Nardini, M., Lang, D.A., Liebeton, K.,
glutathione transferase with high activity Jaeger, K.-E., and Dijkstra, B.W. (2000)
towards naturally occurring Crystal structure of Pseudomonas
isothiocyanates evolves from aeruginosa lipase in the open
promiscuous low-activity variants. J. Mol. conformation. J. Biol. Chem., 275,
Biol., 401, 451–464. 31219–31225.
100 Parikh, M.R. and Matsumura, I. (2005) 109 (a) Hughes, M.D., Zhang, Z.-R.,
Site-saturation mutagenesis is more Sutherland, A.J., Santos, A.F., and
efficient than DNA shuffling for the Hine, A.V. (2005) Discovery of active
directed evolution of b-fucosidase from proteins directly from combinatorial
b-galactosidase. J. Mol. Biol., 352, randomized protein libraries without
621–628. display, purification or sequencing:
101 Pavlova, M., Klvana, M., Prokop, Z., identification of novel zinc finger
Chaloupkova, R., Banas, P., Otyepka, M., proteins. Nucleic Acids Res., 33, e32;
Wade, R.C., Tsuda, M., Nagata, Y., and (b) Chockalingam, K., Chen, Z.,
Damborsky, J. (2009) Redesigning Katzenellenbogen, J.A., and Zhao, H.
dehalogenase access tunnels as a (2005) Directed evolution of specific
strategy for degrading an receptor-ligand pairs for use in the
anthropogenic substrate. Nat. Chem. creation of gene switches. Proc. Natl.
Biol., 5, 727–733. Acad. Sci. U.S.A., 102, 5691–5696.
102 Wang, T.-W., Zhu, H., Ma, X.-Y., 110 Crameri, A., and Stemmer, W.P.C. (1995)
Zhang, T., Ma, Y.-S., and Wei, D.-Z. (2006) Combinatorial multiple cassette
Mutant library construction in directed mutagenesis creates all the permutations
molecular evolution. Mol. Biotechnol., 34, of mutant and wild-type sequences.
55–68. BioTechniques, 18, 194–196.
103 Zhang, J.-H., Dawes, G., and 111 Reetz, M.T., Puls, M., Carballeira, J.D.,
Stemmer, W.P.C. (1997) Directed Vogel, A., Jaeger, K.-E., Eggert, T.,
evolution of a fucosidase from a Thiel, W., Bocola, M., and Otte, N.
galactosidase by DNA shuffling and (2007) Learning from directed evolution:
screening. Proc. Natl. Acad. Sci. U.S.A., further lessons from theoretical
94, 4504–4509. investigations into cooperative mutations
104 Juers, D.H., Heightman, T.D., Vasella, A., in lipase enantioselectivity.
McCarter, J.D., Mackenzie, L., Withers, ChemBioChem, 8, 106–112.
184
112 (a) Sandstr€

om, A.G., Engstr€om, K., 120 Reetz, M.T., Peyralans, J.J.-P.,
Nyhlen, J., Kasrayan, A. and Maichele, A., Fu, Y., and Maywald,
B€ackvall, J.-E. (2009) Directed evolution M. (2006) Directed evolution of
of Candida antarctica lipase A using an hybrid enzymes: evolving
episomaly replicating yeast plasmid. enantioselectivity of an achiral Rh-
Protein Eng., Design & Select., 22, 413–420; complex anchored to a protein. Chem.
(b) Engstr€om, K., Nyhlen, J., Sandtr€om, Commun., 4318–4320.
A.G., and B€ackvall, J.-E. (2010) Directed 121 Pavelka, A., Chovancova, E., and
evolution of an enantioselective lipase Damborsky, J. (2009) HotSpot Wizard: a
with broad substrate scope for hydrolysis web-server for identification of hot spots
of a-substituted esters. J. Am. Chem. Soc., in protein engineering. Nucleic Acids Res.,
132, 7038–7042. 37, W376–W383.
113 Okrasa, K., Levy, C., Wilding, M., Goodall, 122 Saven, J.G. (2009) Computational protein
M., Baudendistel, N., Hauer, B., Leys, D., design, in Protein Engineering Handbook,
and Micklefield, J. (2009) Structure- vol. 1 (eds S. Lutz and U.T. Bornscheuer),
guided directed evolution of alkenyl and Wiley-VCH Verlag GmbH, Weinheim,
arylmalonate decarboxylases. Angew. pp. 325–342.
Chem., 121, 7827–7830; Angew. Chem. 123 (a) Damborsky, J. and Brezovsky, J. (2009)
Int. Ed., 48, 7691–7694. Computational tools for designing end
114 Hawwa, R., Larsen, S.D., Ratia, K., and engineering biocatalysts. Curr. Opin.
Mesecar, A.D. (2009) Structure-based and Chem. Biol., 13, 26–34; (b) Braiuca, P.,
random mutagenesis approaches Lorena, K., Ferrario, V., Ebert, C., and
increase the organophosphate-degrading Gardossi, L. (2009) A three-dimensional
activity of a phosphotriesterase quantitative structure-activity
homologue from Deinococcus relationship (3D-QSAR) model for
radiodurans. J. Mol. Biol., 393, 36–57. predicting the enantioselectivity of
115 Ivancic, M., Valinger, G., Gruber, K., and Candida antarctica lipase B. Adv. Synth.
Schwab, H. (2007) Inverting Catal., 351, 1293–1302; (c) Martı, S.,
enantioselectivity of Burkholderia gladioli Andres, J., Moliner, V., Silla, E., and
esterase EstB by directed and designed Tu~non, I. (2008) Computational design of
evolution. J. Biotechnol., 129, 109–122. biological catalysts. Chem. Soc. Rev., 37,
116 De Groeve, M.R.M., De Baere, M., 2634–2643; (d) Sterner, R., Merkl, R., and
Hoflack, L., Desmet, T., Vandamme, E.J., Raushel, F.M. (2008) Computational
and Soetaert, W. (2009) Creating lactose design of enzymes. Chem. Biol., 15,
phosphorylase enzymes by directed 421–423; (e) Chiappori, F., Dursi, P.,
evolution of cellobiose phosphorylase. Merelli, I., Milanesi, L. and Rovida, E.
Protein Eng., Design & Select., 22, 393–399. (2009) In silico saturation mutagenesis
117 Liang, L., Zhang, J., and Lin, Z. (2007) and docking screening for the analysis of
Altering coenzyme specificity of Pichia protein-ligand interaction: the
stipitis xylose reductase by the semi- endothelial protein C receptor case study.
rational approach CASTing. Microbial Cell BMC Bioinformatics, 10, S3; (f) Senn,
Factories, 6, 36. H.M. and Thiel, W. (2009) QM/MM
118 Kelly, R.M., Leemhuis, H., and methods for biomolecular systems.
Dijkhuizen, L. (2007) Conversion of a Angew. Chem., 121, 1220–1254; Angew.
cyclodextrin glucanotransferase into an Chem. Int. Ed., 48, 1198–1229 ; (g) Siegel,
a-amylase: assessment of directed J.B., Zanghellini, A., Lovick, H.M., Kiss,
evolution strategies. Biochemistry, 46, G., Lambert, A.R., St.Clair, J.L., Gallaher,
11216–11222. J.L., Hilvert, D., Gelb, M.H., Stoddard,
119 Bhuiya M.-W. and Liu C.-J. (2010) B.L., Houk, K.N., Michael, F.E., and
engineering monolignol 4-O- Baker, D. (2010) Computational design of
methyltransferase to modulate lignin an enzyme catalyst for a stereoselective
biosynthesis. J. Biol. Chem., 285, biomolecular Diels-Alder reaction.
277–313. Science, 329, 309–313.
References j185
124 Reetz, M.T. and Wu, S. (2008) Kochrekar, D.A., Cherat, R.N., and
Greatly reduced amino acid alphabets Pai, G.G. (2010) Development of a
in directed evolution: making the right biocatalytic process as an alternative
choice for saturation at homologous to the ()-DIP-CI-mediated asymmetric
enzyme positions. Chem. Commun., reduction of a key intermediate of
5499–5501. montelukast. Org. Process Res. Dev., 14,
125 (a) Saab-Rincon, G., Li, Y., Meyer, M., 193–198.
Carbone, M., Landwehr, M., and 132 Saraf, M.C., Moore, G.L., Goodey, N.M.,
Arnold, F.H. (2009) Protein engineering Cao, V.Y., and Benkovic, S.J. (2006) IPRO:
by structure-guided SCHEMA an iterative computational protein library
recombination, in Protein Engineering redesign and optimization procedure.
Handbook (eds S. Lutz and U.T. Biophys. J., 90, 4167–4180.
Bornscheuer), Wiley-VCH Verlag 133 Chen, R., Li, L., and Weng, Z. (2003)
GmbH, Weinheim, pp. 481–492 ZDOCK: an initial-stage protein-docking
(b) Heinzelman, P., Snow, C.D., Smith, algorithm. Proteins, 52, 80–87.
M.A., Yu, X., Kannan, A., Boulware, K., 134 Khoury, G.A., Fazelinia, H., Chin, J.W.,
Villalobos, A., Govindarajan, S., Pantazes, R.J., Cirino, P.C., and
Minshull, J., and Arnold, F.H. (2009) Maranas, C.D. (2009) Computational
SCHEMA recombination of a fungal design of Candida boidinii xylose
cellulase uncovers a single mutation that reductase for altered cofactor specificity.
contributes markedly to stability. J. Biol. Protein Sci., 18, 2125–2138.
Chem., 284, 26229–26233. 135 Fazelinia, H., Cirino, P.C., and
126 Hernandez, G. and Lemaster, D.M. Maranas, C.D. (2009) OptGraft: a
(2005) Hybrid native partitioning of computational procedure for transferring
interactions among nonconserved a binding site onto an existing protein
residues in chimeric proteins. Proteins, scaffold. Protein Sci., 18, 180–195.
60, 723–731. 136 (a) Reetz, M.T., K€uhling, K.M.,
127 Saraf, M.C., Horswill, A.R., Benkovic, Wilensek, S., Husmann, H.,
S.J., and Maranas, C.D. (2004) FamClash: H€ausig, U.W., and Hermes, M. (2001)
a method for ranking the activity of A GC-based method for high-throughput
engineered enzymes. Proc. Natl. Acad. screening of enantioselective catalysts.
Sci. U.S.A., 101, 4142–4147. Catal. Today, 67, 389–396; (b) Reetz, M.T.,
128 Pantazes, R.J., Saraf, M.C., and Daligault, F., Brunner, B., Hinrichs, H.,
Maranas, C.D. (2007) Optimal protein and Deege, A. (2004) Directed evolution
library design using recombination or of cyclohexanone monooxygenases:
point mutations based on sequence- enantioselective biocatalysts for the
based scoring functions. Protein Eng., oxidation of prochiral thioethers. Angew.
Design Select., 20, 361–373. Chem., 116, 4170–4173; Angew. Chem.
129 Wedge, D.C., Rowe, W., Kell, D.B., and Int. Ed., 43, 4078–4081; (c) Reetz, M.T.,
Knowles, J. (2009) In silico modelling of Becker, M.H., Klein, H.-W., and
directed evolution: implications for St€ockigt, D. (1999) A method for high-
experimental design and stepwise throughput screening of enantioselective
evolution. J. Theor. Biol., 257, 131–141. catalysts. Angew. Chem., 111, 1872–1875;
130 Fox, R.J., Davis, S.C., Mundorff, E.C., Angew. Chem. Int. Ed., 38, 1758–1761.
Newman, L.M., Gavrilovic, V., Ma, S.K., 137 Trapp, O. (2007) Boosting the throughput
Chung, L.M., Ching, C., Tam, S., of separation techniques by
Muley, S., Grate, J., Gruber, J., Whitman, multiplexing. Angew. Chem., 119,
J.C., Sheldon, R.A., and Huisman, G.W. 5706–5710; Angew. Chem. Int. Ed., 46,
(2007) Improving catalytic function by 5609–5613.
ProSAR-driven enzyme evolution. Nat. 138 Gumulya, Y. (2010) Directed evolution of
Biotechnol., 25, 338–344. enantioselective epoxide hydrolases,
131 Liang, J., Lalonde, J., Borup, B., Dissertation, Ruhr-Universit€at Bochum,
Mitchell, V., Mundorff, E., Trinh, N., 2010.
186
139 Reetz, M.T. (2001) Combinatorial and horseradish peroxidase variants with
evolution-based methods in the creation enhanced enantioselectivity by yeast
of enantioselective catalysts. Angew. surface display. Chem. Biol., 14,
Chem., 113, 292–320; Angew. Chem. Int. 1176–1185.
Ed., 40, 284–310. 146 Becker, S., H€obenreich, H., Vogel, A.,
140 Boersma, Y.L., Dr€oge, M.J., Van Der Knorr, J., Wilhelm, S., Rosenau, F.,
Sloot, A.M., Pijning, T., Cool, R.H., Jaeger, K.-E., Reetz, M.T., and Kolmar, H.
Dijkstra, B.W., and Quax, W.J. (2008) (2008) Single-cell high-throughput
A novel genetic selection system for screening to identify enantioselective
improved enantioselectivity of Bacillus hydrolytic enzymes. Angew. Chem., 120,
subtilis lipase A. ChemBioChem, 9, 5163–5166; Angew. Chem. Int. Ed. Engl.,
1110–1115. 47, 5085–5088.
141 Reetz, M.T., H€obenreich, H., Soni, P., 147 (a) Liese, A., Seelbach, K., and Wandrey,
and Fernandez, L. (2008) A genetic C. (2006) Industrial Biotransformations,
selection system for evolving Wiley-VCH Verlag GmbH, Weinheim;
enantioselectivity of enzymes. Chem. (b) Renugopalakrishnan, V., Gardu~ no-
Commun., 5502–5504. Juarez, R., Narasimhan, G., Verma, C.S.,
142 (a) Griffiths, A.D. and Tawfik, D.S. (2006) Wei, X., and Li, P. (2005) Rational design
Miniaturising the laboratory in emulsion of thermally stable proteins: relevance to
droplets. Trends in Biotechnol., 24, bionanotechnology. J. Nanosci.
395–402; (b) Aharoni, A. and Tawfik, D.S. Nanotechnol., 5, 1759–1767.
(2009) In vitro compartmentalization 148 (a) Gatti-Lafranconi, P., Caldarazzo, S.M.,
(IVC) and other high-throughput screens Villa, A., Alberghina, L., and Lotti, M.
of enzyme libraries, in Protein Engineering (2008) Unscrambling thermal stability
Handbook, vol. 2 (eds S. Lutz and U.T. and temperature adaptation in evolved
Bornscheuer), Wiley-VCH Verlag variants of a cold-active lipase. FEBS Lett.,
GmbH, Weinheim, pp. 649–667. 582, 2313–2318; (b) Frieden, C. (2007)
143 (a) Becker, F.S., Schmoldt, H.-U., Protein aggregation processes: in search
Adams, T.M., Wilhelm, S., and of the mechanism. Protein Sci., 16,
Kolmar, H. (2004) Ultra-high-throughput 2334–2344; (c) Ahmad, S. and Rao, N.M.
screening based on cell-surface display (2009) Thermally denatured state
and fluorescence-activated cell sorting or determines refolding in lipase: mutational
the identification of novel biocatalysts. analysis. Protein Sci., 18, 1183–1196.
Curr. Opin. Biotechnol., 15, 323–329; 149 Martin, A., Schmid, F.X., and Sieber, V.
(b) Olsen, M.J., Gam, J., Iverson, B.L., and (2003) PROSIDE - a phage-based method
Georgiou, G., High-throughput FACS for selecting thermostable proteins, in
method for directed evolution of Directed Enzyme Evolution: Screening
substrate specificity, in Methods in and Selection Methods, vol. 230
Molecular Biology, vol. 230 (eds F.H. (eds F.H. Arnold and G. Georgiou),
Arnold and G. Georgiou), Humana Press Humana Press, Totowa, NJ, pp. 57–70.
Inc., Totowa, NJ, pp. 329–342. 150 Hecky, J., Mason, J.M., Arndt, K.M., and
144 Dr€oge, M.J., Boersma, Y.L., Van M€ uller, K.M. (2007) A general method of
Pouderoyen, G., Vrenken, T.E., terminal truncation, evolution, and re-
R€uggeberg, C.J., Reetz, M.T., elongation to generate enzymes of
Dijkstra, B.W., and Quax, W.J. (2006) enhanced stability, in Protein Engineering
Directed evolution of Bacillus subtilis Protocols, vol. 352 (eds K.M. Arndt and
lipase a by use of enantiomeric K.M. M€ uller), Humana Press, Totowa, NJ,
phosphonate inhibitors: crystal pp. 275–304.
structures and phage display selection. 151 Valinger, G., Hermann, M.,
ChemBioChem, 7, 149–157. Wagner, U.G., and Schwab, H. (2007)
145 Lipovsek, D., Antipov, E., Armstrong, Stability and activity improvement of
K.A., Olsen, M.J., Klibanov, A.M., Tidor, cephalosporin esterase EstB from
B., and Wittrup, K.D. (2007) Selection of Burkholderia gladioli by directed evolution
References j187
and structural interpretation of muteins. 157 Kotzia, G.A. and Labrou, N.E. (2009)
J. Biotechnol., 129, 98–108. Engineering thermal stability of L-
152 Petersen, E.I., Valinger, G., S€olkner, B., asparaginase by in vitro directed
Stubenrauch, G., and Schwab, H. (2001) evolution. FEBS J., 276, 1750–1761.
A novel esterase from Burkholderia 158 Reetz, M.T., Soni, P., Acevedo, J.P., and
gladioli which shows high deacetylation Sanchis, J. (2009) Creation of an amino
activity on cephalosporins is related acid network of structurally coupled
to b-lactamases and DD-peptidases. residues in the directed evolution of a
J. Biotechnol., 89, 11–25. thermostable enzyme. Angew. Chem.,
153 Asako, H., Shimizu, M., and Itoh, N. 121, 8418–8422; Angew. Chem. Int. Ed.
(2008) Engineering of NADPH- Engl., 48, 8268–8272.
dependent aldo-keto reductase from 159 Reetz, M.T., Soni, P., and Fernandez, L.
Penicillium citrinum by directed evolution (2009) Knowledge-guided laboratory
to improve thermostability and evolution of protein thermolability.
enantioselectivity. Appl. Microbiol. Biotechnol. Bioeng., 102, 1712–1717.
Biotechnol., 80, 805–812. 160 Mclachlan, M.J., Johannes, T.W., and
154 (a) Palackal, N., Brennan, Y., Callen, W.N., Zhao, H. (2008) Further improvement of
Dupree, P., Frey, G., Goubet, F., phosphite dehydrogenase thermostability
Hazlewood, G.P., Healey, S., Kang, Y.E., by saturation mutagenesis. Biotechnol.
Kretz, K.A., Lee, E., Tan, X., Bioeng., 99, 268–274.
Tomlinson, G.L., Verruto, J., 161 Heinzelman, P., Snow, C.D., Wu, I.,
Wong, V.W.K., Mathur, E.J., Short, J.M., Nguyen, C., Villalobos, A.,
Robertson, D.E., and Steer, B.A. (2004) Govindarajan, S., Minshull, J., and
An evolutionary route to xylanase Arnold, F.H. (2009) A family of
process fitness. Protein Sci., 13, 494–503; thermostable fungal cellulases created
(b) Ruller, R., Deliberto, L., by structure-guided recombination.
Lopes Ferreira, T., and Ward, R.J. (2007) Proc. Natl. Acad. Sci. U.S.A., 106,
Thermostable variants of the 5610–5615.
recombinant xylanase A from Bacillus 162 Minagawa, H., Yoshida, Y.,
subtilis produced by directed evolution Kenmochi, N., Furuichi, M., Shimada, J.,
show reduced heat capacity changes. and Kaneko, H. (2007) Improving the
Proteins, 70, 1280–1293. thermal stability of lactate oxidase by
155 (a) Dumon, C., Varvak, A., Wall, M.A., directed evolution. Cell. Mol. Life Sci., 64,
Flint, J.E., Lewis, R.J., Lakey, J.H., 77–81.
Morland, C., Luginb€ uhl, P., Healey, S., 163 Stevenson, B.J., Liu, J.-W., and Ollis, D.L.
Todaro, T., Desantis, G., Sun, M., (2008) Directed evolution of yeast
Parra-Gessert, L., Tan, X., Weiner, D.P., pyruvate decarboxylase 1 for attenuated
and Gilbert, H.J. (2008) Engineering regulation and increased stability.
hyperthermostability into a GH11 Biochemistry, 47, 3013–3025.
xylanase is mediated by subtle changes to 164 (a) Klibanov, A.M. (2001) Improving
protein structure. J. Biol. Chem., 283, enzymes by using them in organic
22557–22564; (b) Funke, S.A., Eipper, A., solvents. Nature, 409, 241–246;
Reetz, M.T., Otte, N., Thiel, W., Van (b) Carrea, G. and Riva, S. (2008) Organic
Pouderoyen, G., Dijkstra, B.W., Synthesis with Enzymes in Non-aqueous
Jaeger, K.-E., and Eggert, T. (2003) Media, Wiley-VCH Verlag GmbH,
Directed evolution of an enantioselective Weinheim.
Bacillus subtilis lipase. Biocatal. 165 Zumarraga, M., Bulter, T., Shleev, S.,
Biotransform., 21, 67–73. Polaina, J., Martınez-Arias, A., Plou, F.J.,
156 Tang, S.-Y., Le, Q.-T., Shim, J.-H., Ballesteros, A., and Alcalde, M. (2007)
Yang, S.-J., Auh, J.-H., Park, C., and In vitro evolution of a fungal laccase in
Park, K.-H. (2006) Enhancing high concentrations of organic
thermostability of maltogenic amylase cosolvents. Chem. Biol., 14, 1052–1064.
from Bacillus thermoalkalophilus ET2 by 166 You, C., Huang, Q., Xue, H., Xu, Y., and
DNA shuffling. FEBS J., 273, 3335–3345. Lu, H. (2009) Potential hydrophobic
188
interaction between two cysteines in enantioselectivity by error-prone PCR and

interior hydrophobic region improves DNA shuffling. Chem. Biol., 11, 981–990.
thermostability of a family 11 xylanase 175 Kotik, M., Archelas, A., Famerová, V.,
from Neocallimastix patriciarum. Qubrechtová, P., and Kren V., (2011)
Biotechnol. Bioeng., 105, 861–870. Laboratory evolution of an epoxide
167 Francotte, E. and Lindner, W. (2006) hydrolase – towards an enantioconvergent
Chirality in Drug Research, Wiley-VCH biocatalyst J. Biotechnol., 156, 1–10.
Verlag GmbH, Weinheim. 176 DeSantis, G., Wong, K., Farwell, B.,
168 Martın-Matute, B. and B€ackvall, J.-E. Chatman, K., Zhu, Z., Tomlinson, G.,
(2007) Dynamic kinetic resolution Huang, H., Tan, X., Bibbs, L., Chen, P.,
catalyzed by enzymes and metals. Curr. Kretz, K., and Burk, M.J. (2003) Creation
Opin. Chem. Biol., 11, 226–232. of a productive, highly enantioselective
169 Carballeira, J.D., Krumlinde, P., nitrilase through gene site saturation
Bocola, M., Vogel, A., Reetz, M.T., and mutagenesis (GSSM). J. Am. Chem. Soc.,
B€ackvall, J.-E. (2007) Directed evolution 125, 11476–11477.
and axial chirality: optimization of the 177 Hill, C.M., Li, W.-S., Thoden, J.B., Holden,
enantioselectivity of Pseudomonas H.M., and Raushel, F.M. (2003) Enhanced
aeruginosa lipase towards the kinetic degradation of chemical warfare agents
resolution of a racemic allene. Chem. through molecular engineering of the
Commun., 1913–1915. phosphotriesterase active site. J. Am.
170 Suen, W.-C., Zhang, N., Xiao, L., Chem. Soc., 125, 8990–8991.
Madison, V., and Zaks, A. (2004) 178 Fong, S., Machajewski, T.D., Mak, C.C.,
Improved activity and thermostability of and Wong, C.-H. (2000) Directed
Candida antarctica lipase B by DNA evolution of D-2-keto-3-deoxy-6-
family shuffling. Protein Eng., Design & phosphogluconate aldolase to new
Select., 17, 133–140. variants for the efficient synthesis of
171 Schmidt, M., Hasenpusch, D., K€ahler, D- and L-Sugars. Chem. Biol., 7, 873–883.
M., Kirchner, U., Wiggenhorn, K., Langel, 179 Wada, M., Hsu, C.-C., Franke, D.,
W., and Bornscheuer, U.T. (2006) Directed Mitchell, M., Heine, A., Wilson, I., and
evolution of an esterase from Pseudomonas Wong, C.-H. (2003) Directed evolution of
fluorescens yields a mutant with excellent N-acetylneuraminic acid aldolase to
enantioselectivity and activity for the catalyze enantiomeric aldol reactions.
kinetic resolution of a chiral building Bioorg. Med. Chem., 11, 2091–2098.
block. ChemBioChem, 7, 805–809. 180 Williams, G.J., Woodhall, T., Farnsworth,
172 Bartsch, S., Kourist, R. and L.M., Nelson, A., and Berrry, A. (2006)
Bornscheuer, U.T. (2008) Complete Creation of a pair of stereochemically
inversion of enantioselectivity towards complementary biocatalysts. J. Am.
acetylated tertiary alcohols by a double Chem. Soc., 128, 16238–16247.
mutant of a Bacillus subtilis esterase. 181 Williams, G.J., Domann, S., Nelson, A.,
Angew. Chem., 120, 1531–1534; Angew. and Berry, A. (2003) Modifying the
Chem. Int. Ed. Engl., 47, 1508–1511. stereochemistry of an enzyme-catalyzed
173 epanek, V., Kyslık, P., and
Kotik, M., St reaction by directed evolution. Proc. Natl.
Maresova, H. (2007) Cloning of an epoxide Acad. Sci. U.S.A., 100, 3143–3148.
hydrolase-encoding gene from Aspergillus 182 Jennewein, S., Sch€ urmann, M.,
niger M200, overexpression in E. coli, and Wolberg, M., Hilker, I., Luiten, R.,
modification of activity and enantio- Wubbolts, M., and Mink, D. (2006)
selectivity of the enzyme by protein Directed evolution of an industrial
engineering. J. Biotechnol., 132, 8–15. biocatalyst: 2-deoxy-D-ribose 5-phosphate
174 Van Loo, B., Spelberg, J.H.L., aldolase. Biotechnol. J., 1, 537–548.
Kingma, J., Sonke, T., Wubbolts, M.G., 183 Ran, N., Draths, K.M., and Frost, J.W.
and Janssen, D.B. (2004) Directed (2004) Creation of a shikimate pathway
evolution of epoxide hydrolase from A. variant. J. Am. Chem. Soc., 126,
radiobacter toward higher 6856–6857.
References j189
184 Ran, N. and Frost, J.W. (2007) Directed 189 Wu, S., Acevedo, J.P., and Reetz, M.T.
evolution of 2-keto-3-deoxy-6- (2010) Induced allostery in the directed
phosphogalactonate aldolase to replace 3- evolution of an enantioselective
deoxy-D-arabino-heptulosonic acid 7- Baeyer–Villiger monooxygenase. Proc.
phosphate synthase. J. Am. Chem. Soc., Natl. Acad. Sci. U.S.A., 107, 2775–2780.
129, 6130–6139. 190 Kirschner, A. and Bornscheuer, U.T.
185 Lingen, B., Kolter-Jung, D., D€ unkelmann, (2008) Directed evolution of a
P., Feldmann, R., Gr€otzinger, J., Pohl, M., Baeyer–Villiger monooxygenase to
and M€ uller, M. (2003) Alteration of the enhance enantioselectivity. Appl.
substrate specificity of benzoylformate Microbiol. Biotechnol., 81, 465–472.
decarboxylase from Pseudomonas putida 191 (a) Peters, M.W., Meinhold, P.,
by directed evolution. ChemBioChem, 4, Glieder, A., and Arnold, F.H. (2003)
721–726. Regio- and enantioselective alkane
186 (a) Liang, J., Lalonde, J., Borup, B., hydroxylation with engineered
Mitchell, V., Mundorff, E., Trinh, N., cytochromes P450 BM-3. J. Am. Chem.
Kochrekar, D.A., Cherat, R.N. and Soc., 125, 13442–13450; (b) Landwehr,
Pai, G.G. (2010) Development of M., Hochrein, L., Otey, C.R., Kasrayan, A.,
biocatalytic process as an alternative to the B€ackvall, J.-E., and Arnold, F.H. (2006)
()-DIP-Cl-mediated asymmetric Enantioselective a-hydroxylation of 2-
reduction of a key intermediate of arylacetic acid derivatives and Buspirone
montelukast. Org. Process Res. Dev., 14, catalyzed by engineering cytochrome
193–198; (b) Liang, J., Mundorff, E., P450 BM-3. J. Am. Chem. Soc., 128,
Voladri, R., Jenne, S., Gilson, L., 6058–6059; (c) Urlacher, V.B.,
Conway, A., Krebber, A., Wong, J., Lutz-Wahl, S. and Schmid, R.D. (2004)
Huismann, G., Truesdale, S., and Microbial P450 enzymes in
Lalonde, J. (2010) Highly enantioselective biotechnology. Appl. Microbiol.
reduction of a small heterocyclic ketone: Biotechnol., 64, 317–325; (d) Gillam,
biocatalytic reduction of E.M.J. (2008) Engineering cytochrome
tetrahydrothiophene-3-one to the P450 enzymes. Chem. Res. Toxicol., 21,
corresponding (R)-alcohol. Org. Process 220–231; (e) M€ unzer, D.F., Meinhold, P.,
Res. Dev., 14, 188–192. Peters, M.W., Feichtenhofer, S., Griengl,
187 (a) Reetz, M.T., Brunner, B., H., Arnold, F.H., Glieder, A., and De
Schneider, T., Schulz, F., Clouthier, C.M., Raadt, A. (2005) Stereoselective
and Kayser, M.M. (2004) Directed hydroxylation of an achiral
evolution as a method to create cyclopentanecarboxylic acid derivative
enantioselective cyclohexanone using engineered P450s BM-3. Chem.
monooxygenases for catalysis in Commun., 2597–2599; (f) Kille, S., Zilly,
Baeyer–Villiger reactions. Angew. Chem., F.E., Acevedo, J.P., and Reetz, M.T. (2011)
116, 4167–4170; Angew. Chem. Int. Ed., 43, Regio- and stereoselectivity of P450-
4075–4078; (b) Mihovilovic, M.D., catalysed hydroxylation of steroids
Rudroff, F., Winninger, A., Schneider, controlled by laboratory evolution. Nat.
T., Schulz, F., and Reetz, M.T. (2006) Chem., 3, 738–743.
Microbial Baeyer–Villiger oxidation: 192 Antipov, E., Cho, A.E., Wittrup, K.D., and
stereopreference and substrate Klibanov, A.M. (2008) Highly L and D
acceptance of cyclohexanone enantioselective variants of horseradish
monooxygenase mutants prepared by peroxidase discovered by an ultrahigh-
directed evolution. Org. Lett., 8, throughput selection method. Proc. Natl.
1221–1224. Acad. Sci. U.S.A., 105, 17694–17699.
188 Reetz, M.T. and Wu, S. (2009) Laboratory 193 (a) Alexeeva, M., Enright, A., Dawson,
evolution of robust and enantioselective M.J., Mahmoudian, M. and Turner, N.J.
Baeyer–Villiger monooxygenases for (2002) Deracemization of
asymmetric catalysis. J. Am. Chem. Soc., a-methylbenzylamine using an enzyme
131, 15424–15432. obtained by in vitro evolution. Angew.
190
Chem., 114, 3309–3312; Angew. Chem. Discovery of a thermostable

Int. Ed., 41, 3177–3180; (b) Carr, R., Baeyer–Villiger monooxygenase by
Alexeeva, M., Enright, A., Eve, T.S.C., genome mining. Appl. Microbiol.
Dawson, M.J., and Turner, N.J. (2003) Biotechnol., 66, 393–400; (b) Zambianchi,
Directed evolution of an amine oxidase F., Fraaije, M.W., Carrea, G., De Gonzalo,
possessing both broad substrate G., Rodrıguez, C., Gotor, V., and Ottolina,
specificity and high enantioselectivity. G. (2007) Titration and assignment of
Angew. Chem., 115, 4955–4958; Angew. residues that regulate the
Chem. Int. Ed., 42, 4807–4810. enantioselectivity of phenylacetone
194 Avi, M., Wiedner, R.M., Griengl, H., and monooxygenase. Adv. Synth. Catal., 349,
Schwab, H. (2008) Improvement of a 1327–1331; (c) Rodrıgues, C., De Gonzalo,
stereoselective biocatalytic synthesis G., Torres Pazmi~ no, D.E., Fraaije, M.W.,
by substrate and enzyme engineering: and Gotor, V. (2008) Selective Baeyer–
2-hydroxy-(40 -oxocyclohexyl)acetonitrile Villiger oxidation of racemic ketones in
as the model. Chem. Eur. J., 14, aqueous-organic media catalyzed by
11415–11422. phenylacetone monooxygenase.
195 Jacobsen, E.N. (2000) Asymmetric Tetrahedron: Asymmetry, 19, 197–203.
catalysis of epoxide ring-opening 201 Malito, E., Alfieri, A., Fraaije, M.W., and
reactions. Acc. Chem. Res., 33, 421–431. Mattevi, A. (2004) Crystal structure of a
196 Zou, J.Y., Hallberg, B.M., Bergfors, T., Baeyer–Villiger monooxygenase. Proc.
Oesch, F., Arand, M., Mowbray, S.L., and Natl. Acad. Sci. U.S.A., 101,
Jones, T.A. (2000) Structure of Aspergillus 13157–13162.

niger epoxide hydrolase at 1.8 A 202 Koeller, K.M. and Wong, C.-H. (2001)
resolution: implications for the structure Enzymes for chemical synthesis. Nature,
and function of the mammalian 409, 232–240.
microsomal class of epoxide hydrolases. 203 Bolt, A., Berry, A., and Nelson, A. (2008)
Structure (London), 8, 111–122. Directed evolution of aldolases for
197 Reetz, M.T., Bocola, M., Wang, L.-W., exploitation in synthetic organic
Sanchis, J., Cronin, A., Arand, M., Zou, J., chemistry. Arch. Biochem. Biophys., 474,
Archelas, A., Bottalla, A.-L., Naworyta, A., 318–330.
and Mowbray, S.L. (2009) Directed 204 (a) Tokuriki, N. and Tawfik, D.S. (2009)
evolution of an enantioselective epoxide Chaperonin overexpression promotes
hydrolase: uncovering the source of genetic variation and enzyme evolution.
enantioselectivity at each evolutionary Nature, 459, 668–673;(b) Matsumura, I.
stage. J. Am. Chem. Soc., 131, 7334–7343. and Ivanov, A.A. (2009) How evolving
198 (a) Kayser, M.M. (2009) Designer enzymes can beat the heat and avoid
reagents recombinant microorganisms: defeat. Nat. Chem. Biol., 5, 538–539.
new and powerful tools for organic 205 Glieder, A., Weis, R., Skranc, W.,
synthesis. Tetrahedron, 65, 947–974; Poechlauer, P., Dreveny, I., Majer, S.,
(b) Mihovilovic, M.D. (2006) Enzyme Wubbolts, M., Schwab, H., and Gruber, K.
mediated Baeyer–Villiger oxidations. (2003) Comprehensive step-by-step
Curr. Org. Chem., 10, 1265–1287. engineering of an (R)-hydroxynitrile lyase
199 Bocola, M., Schulz, F., Leca, F., Vogel, A., for large-scale asymmetric synthesis.
Fraaije, M.W., and Reetz, M.T. (2005) Angew. Chem., 115, 4963–4966; Angew.
Converting phenylacetone Chem. Int. Ed., 42, 4815–4818.
monooxygenase into 206 Khersonsky, O., Röthlisberger, D.,
phenylcyclohexanone monooxygenase by Wollacott, A.M., Murphy, P., Dym, O.,
rational design: towards practical Albeck, S., Kiss, G., Houk, K.N., Baker,
Baeyer–Villiger monooxygenases. Adv. D., and Tawfik, D.S. (2011) Optimization
Synth. Catal., 347, 979–986. of the in-silico-designed Kemp eliminase
200 (a) Fraaije, M.W., Wu, J., Heuts, KE70 by computational design and
D.P.H.M., Van Hellemond, E.W., directed evolution. J. Mol. Biol., 407,
Spelberg, J.H.L., and Janssen, D.B. (2005) 391–412.
j191
6
Production and Isolation of Enzymes
Yoshihiko Hirose
6.1
Introduction
This chapter gives a brief review of the isolation and production of enzymes. More
detailed informationcanbe obtainedfromvarious publishedtextbooksand reviews [1–5].
Mostindustrialenzymesusedforchemicalsynthesisaresuppliedinacrudeformwithan
active-enzyme content of only a few percent. The other constituents are inorganic salts,
polysaccharides, and diatomaceous earth used as stabilizers and excipients. Purified
enzymes for biotransformation are supplied by some manufacturers in a crystal or
immobilized form. These enzymes, though expensive, are easy to apply for biotrans-
formations in organic media. The use of more purified enzymes is increasing.
Barriers to the production of industrial enzymes include economic factors, the
availability of optimal enzymes, and safety issues. Figures 6.1–6.3 illustrate common
fermentation and purification processes. The process differs for extracellular and
intracellular enzymes, liquid and solid culture, and enzyme application. Fermenta-
tion conditions such as temperature, pH, agitation speed, aeration, demand oxygen,
and so on are computer-controlled for optimization.
There are no internationally standard assay methods for industrial enzymes and
the definition of enzyme activity unit is also different for each enzyme. The activity of
industrial enzymes is shown by various methods depending on the manufacturers.
For instance, commercial lipase activities are measured by the hydrolysis of olive oil
under the various conditions and these figures are not comparable with each other.
When customers apply these biocatalysts for chemical synthesis in organic solvents,
these figure are sometimes reliable, and sometimes not. Users should not judge
commercial enzymes based only on price and the activity shown in the table the
manufacturer provides. Enzymes should be evaluated based on their practical
performance under the conditions used. Most users of biotransformation are not
experts in measuring enzyme activity, so the establishment of an assay method and
practice are essential if one is to optimize the performance of enzymes.
Several commercial enzymes are powders that include diatomaceous earth or
dextrin. These enzymes should be used after immobilization on a suitable carrier.
The activity of an immobilized enzyme usually is enhanced up to tenfold.
and Oliver May.
j 6 Production and Isolation of Enzymes
192
Preparation of fermentation medium ----- Starch, sugar, soybean powder,

yeast extracts, minerals, inducer, etc.
Dissolution
Sterilization
Inoculation
Fermentation in flask (100–500 ml)
Partial inoculation
Fermentation in jar (3–30 l)
Partial inoculation
Fermentation in seed tank (1000–3000 l)
Partial inoculation
Fermentation in main tank (> 3000 l)
Broth-out
Filtration
Concentration by evaporation or ultrafiltration
Precise filtration
Solvent precipitation
Filtration or centrifugation
Drying
Product (crude product)
Dissolution
Ion-exchange chromatography
Salting-out
Desalting
Hydrophobic chromatography
Gel filtration
Desalting
Freeze-drying
Product
Figure 6.1 Common production process for industrial extracellular enzymes.

Preparation of fermentation medium ----- Starch, sugar, soybean powder,
yeast extracts, minerals, inducer, etc.
Dissolution
Sterilization
Inoculation
Fermentation in flask to in main tank
Centrifugation or filtration
Disruption and extraction
Filtration
Concentration by evaporation or ultrafiltration
Further purification
(Salting-out, chromatography, etc)
Precise filtration
Crystallization or freeze-drying
Product
Figure 6.2 Common production process for special intracellular enzymes.
Solid culture
Sterilization
Inoculation
Extraction of enzyme
Filtration
Concentration by ultrafiltration
Precipitation
Drying
Product (crude product)
Further purification
(Salting-out, chromatography, etc)
Product
Figure 6.3 Common production process for special enzymes by solid fermentation.
194
Table 6.1 AMFEP recommended specification.
Component Recommended specification
Arsenic 3 ppm
Lead 10 ppm
Heavy metals < 40 ppm
Mycotoxins Negative
Antibacterial activity Negative
Coliforms < 30 g1
Escherichia coli Negative in 25 g
Salmonella Negative in 25 g
Total viable count < 50 000 g1
Regulatory assessments for enzymes used in biotransformation are not clearly

stipulated. At present, food assessments of microbial enzymes are provided by
AMFEP (Association of Manufacturers and Formulators of Enzyme Products), which
has suggested microbial enzyme purity and immobilization as given below:
Purity: a chemical and microbial specification must be given. Based on FCC
(Food Chemicals Codex) recommendations, AMFEP recommended the speci-
fication shown in Table 6.1.
Immobilization: the immobilization system should be described in detail. Tests
to indicate the physicochemical stability of both the system and its carrier and
enzyme are essential.
These regulatory aspects would be acceptable for biocatalysts.
6.2
Enzyme Suppliers for Biotransformation
Worldwide, over 400 companies deal with enzymes and there are approximately 12 major
producers with an increasingly distinct separation of product ranges. About 60 compa-
nies produce substantial amounts of a small range and about 400 companies produce a
very limited range of industrial enzymes. Japanese enzyme producers have a special
range for industrial or in-house use and contribute 12–15% of world production. There
are 24 companies that supply special enzymes for biotransformation (Table 6.2).
6.3
Origins of Enzymes
6.3.1
Microbial Enzymes
More than 90% of enzymes are produced by fermentation by microorganisms, which

are used to prepare industrial and special use enzymes. Prokaryotic cells and
6.3 Origins of Enzymes j195
Table 6.2 Main enzyme suppliers for biotransformation.
Company Country
Amano Pharmaceutical Co., Ltd. Japan

Asahi Chemical Co. Japan
Biocatalysts Ltd. UK
Biozyme Labs Ltd. UK
Calbiochem Corp. USA
Christian Hansen AS DK
Codexis, Inc. USA
Fluka Chemicals Ltd UK
Genencor Int. Finland
Genzyme Ltd UK
DSM (Gist) Holland
Meito Sangyo Co. Japan
Merck Germany
Nagase Biochemicals Japan
Novo Nordisk AS Denmark
Oriental Yeast Co. Japan
Osaka Saiken KK Japan
Roche Diagnostics GmbH Germany
Rohm GmbH Germany
Shin Nihon Chem Co. Japan
Sigma Chemical Co. USA
Toyobo Co. Japan
eukaryotic cells can be easily grown in culture, and the technology of scale-up is well
established on an industrial scale. Various kinds of fungi, bacteria, and yeast have
been screened for the production of special enzymes. Extracellular enzymes, for
instance hydrolytic enzymes, are secreted into liquid and solid culture and are
relatively stable in cultivation media.
The production by genetically modified organisms (GMOs) is becoming popular
in this field, and several kinds of GMO have been used to increase the productivity of
biocatalysts. When one employs GMO enzymes for industrial use, one should know
the origin of the microorganisms used and how the production method was changed.
The new enzyme preparation is likely to have a different compositional spectrum of
enzymes and side activities. The regulations covering biocatalysts are not severe at
present, but are likely to become more stringent.
6.3.2
Plant Enzymes
Some proteases, such as papain, bromelain, and ficin, lipoxygenases from soy bean
and white germ, and peroxidase from horseradish, are typical plant enzymes. Plant
196
proteases are extracted and partially purified to give a powder extract. Some are
supplied as digestive enzymes or nutraceutical enzymes. These have the character-
istics of an SH-enzyme (thiol protease) and work in the hydrolysis of racemic esters as
a protease. Lipoxygenases are only available from soybean, but the activity is not high
and the regiospecificity for unsaturated fatty acids is not severe. Lipoxygenases from
other plants are relatively unstable and used in-house only.
6.3.3
Animal Enzymes
Porcine liver esterase (PLE), porcine pancreas lipase (PPL), and arginase are well
known as biocatalysts among industrial animal enzymes. PLE catalyzes very well the
hydrolysis of certain kinds of prochiral diesters and is supplied as a suspension with
ammonium sulfate or in liquid form. The substrate specificity of PLE is not wide, but
this is a well-investigated enzyme. PPL is very cheap and a useful biocatalyst in
industry. Commercial PPL is a mixture of many kinds of pancreas enzymes, and the
name pancreatin is well known as a digestive enzyme. Pregastric esterase is applied
for transesterification of triglycerides. Arginase from calf liver is used to produce L-
ornithine from the proteinogenic amino acid-arginine. The use of animal enzymes
seems to be gradually decreasing because of disease and a variable supply. In the
future, animal enzymes will no doubt be replaced by microbial enzymes of equivalent
performance.
6.4
Fermentation of Enzymes
6.4.1
Liquid Fermentation
Liquid fermentation is useful for the production of enzymes as well as antibiotics. It is

good for scale-up and reproduction. Two types of enzyme produced: intracellular and
extracellular enzymes. With advances in genetic engineering, Escherichia coli is now
being used to produce enzymes. When E. coli is used, the enzyme is accumulated
inside the cell. This method is very popular.
6.4.2
Solid Fermentation
In Japan, solid fermentation is still used to produce many kinds of enzymes,

including lipases, proteases, and acylases. Some glycotransferases are also
produced by solid fermentation. In the production of proteases, solid fermen-
tation is often used to increase the productivity in solids. On changing to liquid
fermentation, the protease is not produced and its properties change. Solid
fermentation is old-fashioned and difficult to scale up because of the expensive
facilities needed.
6.5 Extraction of Enzymes j197
6.4.3
Extraction of Enzymes
To improve the extraction of enzymes, organic solvents and surfactants are some-
times used.
6.5
Extraction of Enzymes
6.5.1
Microbial Enzymes
Extraction methods depend on the fermentation conditions and the microorganism,

for example, liquid or solid culture, intracellular or extracellular enzyme, laboratory
or production scale preparation, and so on.
Because enzymes are more soluble in buffer solution than water, enzymes in solid
fermentation are extracted by stirring several times in a suitable buffer solution. The
pH of the buffer solution is adjusted based on the stability and the pI (isoelectric
point) of the enzyme. The solid medium is removed by filtration after extraction and
the crude enzyme solution is concentrated at the next step.
Liquid fermentation produces two types of product, intracellular and extracellular
enzymes. In the case of extracellular enzymes, the crude enzyme solution is collected by
filtration or centrifugation of the microorganism. For intracellular enzymes, collection of
the microorganism and extraction of enzymes after disruption is required. On a
laboratory scale, ultrasound equipment or a French press is used for disruption of
microorganisms. On a large or industrial scale, a mechanical grinding mill with glass
beads (e.g., a Dynamill) is used. It can treatmicroorganisms suspended in buffer solution
at a rate of 100 l h1. Another method is enzymatic disruption of microorganisms with
lysozyme or YL (i.e., a commercial cell wall lytic enzyme like YL-15 provided by Amano
Enzyme Inc.). This method is easy to apply on an industrial scale because it does not
require special apparatus. The microorganism suspended in buffer is stirred for several
hours in the presence of a suitable amount of lysozyme at room temperature. During the
enzymatic treatment, freezing and thawing of the microorganism is effective for
disruption. In this case, lysozyme should be added before freezing the microorganism.
The combination of mechanical and enzymatic treatment is much more effective.
After disruption of the cell wall, enzymes should be extracted with buffer solution.
Sometimes enzymes are adsorbed by or adhere to the cell wall and must be extracted
by adding a small amount of surfactant, such as Triton X-100. The cell wall is then
filtered off or centrifuged to obtain the crude enzyme solution.
Recombinant heterogeneous enzyme produced by GMOs such as E. coli are
sometimes obtained as inclusion bodies that form insoluble aggregates and show
less than the original activity. To regenerate the activity from the inclusion body, it is
dissolved in the presence of denaturing agents, usually a highly concentrated
guanidinium salt and urea and reducing agents, usually thiol compounds. Protein
is allowed to refold into its original active conformation after removal of the agents.
198
6.5.2
Plant Enzymes
Some proteases, papain and bromelain, are derived from plants and are extracted
from fruits. The fruits are ground by a grinder or cutter and the proteases are
extracted by buffer solution. A diluted cooled buffer is more effective than water for
extraction and the extracted solution including desired enzymes should be cooled
during all treatments. The content varies depending on the time and place as well as
the plant.
6.5.3
Animal Enzymes
Animal organs containing desired enzymes are stored frozen and then ground and
crushed by a homogenizer. Enzyme stabilizers or protease inhibitors are sometimes
added on homogenizing the organs, and a buffer is preferable for extraction. To
improve the extraction, the residue is removed by filtration or by centrifugation to
obtain crude extract. Animal organs contain various kinds of enzymes and a large
volume of protein is extracted. Even relatively unstable enzymes retain their activity
in crude extracts, but it is necessary to purify the enzymes step by step.
6.6
Concentration
After the extraction of extracellular enzymes of solid culture or the fermentation of

extracellular enzymes of liquid culture, the fermentation medium including desired
enzymes is centrifuged or filtered to remove the microorganism and so on. The next
step is concentration by evaporation under reduced pressure or ultrafiltration to
reduce the volume of the enzyme solution. It is not easy to evaporate large amounts of
water under reduced pressure. Evaporation should be carried out at < 30 C, except in
the case of thermostable proteins, which can be evaporated at higher temperatures.
The concentrations of the salt and other soluble materials of the concentrated
solution are increased by evaporation.
The most convenient and simple method for production is ultrafiltration. The

method uses membrane tubes with pore sizes from about 6000 to 50 000 A. Small
molecules like salt ions as well as water pass through the pores of the membrane tube
while large molecules like proteins remain inside the tube. The concentration of the
buffer solution is the same before and after ultrafiltration. The leaking of desired
proteins in permeates should be checked during the concentration stage. Regular
maintenance is carried out by using a standard protein. The membrane tube is made
of polyethylene, polypropylene, and so on; the irreversible adsorption of desired
proteins should be avoided. The materials for the membrane should be selected
before use. The flow rate of ultrafiltration depends on the facility and the protein
solution applied. The final protein concentration is up to about 100 g l1.
6.7 Purification of Enzymes j199
Ultrafiltration is also used for desalting. Smaller membrane tubes are used for this
purpose.
Another concentration method is precipitation using organic solvents or salts
(Section 6.7.2). Ethanol is especially useful for this purpose. Despite the volume of
organic solvents, it is still frequently used as a first step in the purification.
6.7
Purification of Enzymes
6.7.1
Chromatography
Chromatography is the major purification method. The choice of technique is

determined by the overall yield, efficiency, speed, and convenience. To purify a
desired enzyme in the broth, a combination of different types of chromatography is
an effective approach.
6.7.1.1 Ion-Exchange Chromatography (IEX)

Ion-exchange chromatography (IEX) [7] is the most typical and frequently used
method for separating enzymes. Some of the advantages of IEX are high resolution
power, applicability, and ease of control and scale-up. There are two types of
exchanger in IEX (Figures 6.4 and 6.5). Positively charged exchangers have negatively
charged counter-ions (anions) available for exchange and are called anion exchan-
gers. Negatively charged exchangers have positively charged counter-ions (cations)
and are called cation exchangers.
The basic principle of separation in IEX is the reversible adsorption of charged
protein molecules dissolved in a buffer solution by oppositely charged ion-exchanged
groups on the matrix (Figure 6.6).
Anion exchanger with counter-ions Cation exchanger with counter-ions
Figure 6.4 Two types of ion exchanger [7].

200
Figure 6.5 Principle of ion-exchange equilibrium [4].
Figure 6.6 Principle of anion-exchange chromatography [7].

Figure 6.7 Net charge of protein as a function of pH [7].
Proteins dissolved in buffer solution have different charges depending on the

solution. When the pH of the buffer solution is below the pI, the protein has a positive
charge, and when the pH is above the pI the protein has a negative charge (Figures 6.7
and 6.8).
An ion exchanger consists of a solid matrix covalently bound to a charged group.
The matrix is made of an organic compound, synthetic resin, or polysaccharide, such
as Sepharose and SephadexÔ. A typical matrix is a round microbead. The char-
Figure 6.8 Relationship between the charge of proteins and the pH [6].
202
Table 6.3 Functional groups used on ion exchangers and its structure.
Anion exchangers Structure
Dimethylaminoethyl (DE) -CH2CH2N(CH3)2

Diethylaminoethyl (DEAE) -CH2CH2N(C2H5)2
Quaternary ammonium (QA) -CH2N þ (CH3)3
Quaternary ammonium (QAE) -CH2CH2N þ (C2H5)2CH2CH(OH)CH3
Quaternary ammonium (QMA) -N þ (CH3)3
Cation exchangers Structure
Carboxymethyl (CM) -CH2COOH

Phosphate (P) -PO4H2
Sulfonic ethyl (SE) -CH2CH2SO3H
Sulfonic propyl (SP) -CH2CH2CH2SO3H
acteristics of the matrix determine its chromatographic properties, such as efficiency,

capacity, recovery, chemical stability, mechanical strength, and flow properties. The
properties of the matrix affect its behavior towards biological substances and the
maintenance of biological activity.
The charged group determines the basic property of IEX, such as the type and the
strength of the ion exchanger. The number of charged substituent groups per gram of
dry ion exchanger or per ml of swollen gel affects its total ionic capacity. This number
can be measured by titration with a strong base or acid and is shown as m mol per ml of
gel. Table 6.3 shows typical functional groups, classified as two types. The functional
groups of anion exchangers are substituted ammonium groups. Cation exchangers
have sulfoxyl or carboxyl groups. Ion exchangers with sulfonic and quaternary
ammonium groups are called strong ion exchangers. Those with carboxyl and
diethylaminoethyl groups form weak ion exchangers. The terms strong and weak
refer to the extent of the variation of ionization with pH and not the strength of
binding. A strong exchanger is ionized over a broad pH range, a weak one over a
narrower range.
Experimental Design The choice of matrix and functional group depends on the pH
stability, molecular size, and isoelectric points (pI) of the protein, and on the
requirements of the application. The pI can be measured by electrophoresis or can
be checked in the comprehensive lists of pI for proteins.
The starting pH of buffer is chosen so that proteins to be bound to the exchanger
are charged. Thus, the starting pH is at least more than 1 unit above the pI for anion
exchangers or at least less than 1 unit below the pI for cation exchangers to facilitate
adequate binding. Proteins begin to dissociate from ion exchangers at about 0.5 pH
units from their pI at 0.1 M ionic strength.
Most proteins have their pI within the acidic range, so they are usually negatively
charged in neutral buffer solution and show the properties of an anion.
Ion-exchange separation can be carried out using the following three procedures:
column chromatography, a batch method, and an expanded bed adsorption. Indus-
trial-scale preparation is used.
Column Separation Ion exchangers are available for laboratory-scale separations,

and factors such as cost and reproducibility and so on are not very important. For
industrial separation, however, it is necessary to optimize the purification conditions.
The DEAE exchanger is the most useful in terms of the pI and stability of most
proteins.
Choice of Exchanger Group and Buffer The choice of ion exchanger and buffer
solution is limited by the stability of the proteins. Because most proteins have their pI
in the acidic range, they have a slightly positive charge below the pI and can be easily
absorbed on a cation exchanger (e.g., CM). In contrast, they have a negative charge
above the pI and an anion exchanger (e.g., DEAE) is used.
Choice of pH and Ionic Strength The pH of the buffer depends on the pI of the
proteins, and the ionic strength causes absorption on the ion exchanger and
desorption from it. The required concentration of starting buffer depends on the
nature of the buffering substance. It should be at least 10 mM. Suitable ion salts
stabilize the proteins in solution and excess salts cause the denaturation and
precipitation of the protein.
Batch Separation Batch separation is conducted in the same way as column

development by a stepwise elution method. Batch separation is a better method for
large sample volumes with low concentration protein. Large volumes take a long
time. The initial conditions for batch separation are almost the same as for column
chromatography, for instance, buffer pH, ionic strength, and so on.
The conditions must be quite strong so that the proteins completely bind to the
adsorbent. To increase recovery, the pH of the buffer is maintained at a couple of units
from the pI of the protein. Batch separation is simple and useful to concentrate a low
protein solution, but the resolution is not high.
Experimental Procedure Batch separation is a simple method whereby the protein

solution is stirred together with the ion exchanger in an appropriate buffer for 1 h.
The slurry is collected by filtration and the ion exchanger is washed with fresh buffer
solution. When no desired protein is observed in the filtrate, a couple of bed volumes
of the elution buffer are added and stirred for 1 h to desorb the desired protein. The
solution including the desired protein is collected by filtration. The change of pH or
ionic strength of the elution is determined by gradient or stepwise chromatography.
6.7.1.2 Hydrophobic Interaction Chromatography (HIC)

Hydrophobic interaction chromatography (HIC) [8] is based on the hydrophobic
properties of proteins and the hydrophobic ligands covalently bonded to the matrix.
The chromatography has three forms: (i) hydrophobic chromatography, where both
absorption and desorption are based only on hydrophobic binding, (ii) hydropho-
204
R-
C2H5- Ethyl
C4H9- Butyl
C6H13- Hexyl
C8H17- Octyl
C10H21- Decyl
C6H5- Phenyl
Figure 6.9 Hydrophobic ligands attached to a matrix.
bic–ionic chromatography, where absorption is based on hydrophobic binding and

desorption on ionic exclusion by changing the pH of the buffer, and (iii) mixed
function chromatography, based on hydrophobic, ionic, and hydrogen binding.
HIC and reversed-phase chromatography (RPC) operate by very similar principles
based on hydrophobic interaction. Adsorbents for RPC are much substituted with
hydrophobic ligands, such as octadecyl, octyl, or phenyl groups (Figures 6.9 and 6.10).
Protein binding to RPC adsorbents is usually very strong despite the concentration of
salt, and some kinds of organic solvents, such as acetonitrile and isopropyl alcohol,
are used to desorb the protein. Consequently, RPC is carried out for low molecular
weight molecules such as peptides because most proteins are unstable in highly
concentrated organic solvents. In contrast, adsorbents for HIC are less substituted
with similar groups, mainly butyl groups. Protein binding to HIC adsorbents
(mg protein / ml gel)
Binding capacity
Length of the n-alkyl chain
Figure 6.10 Effect of alkyl chain length on binding capacity in HIC [8].
Figure 6.11 Principle of hydrophobic chromatography [8]. P: polymer matrix; S: soluble molecule;
L: ligand attached to polymer matrix; H: hydrophobic patch on surface of soluble molecule; W: water
molecules in the bulk solution; S: salt (ammonium sulfate).
requires a neutral salt like ammonium sulfate in the mobile phase, and the protein is
desorbed by decreasing the concentration of salt. The slope of ionic strength in HIC is
opposite to that of IEX.
The surface of a protein is relatively hydrophilic in the lower concentration buffer
solution, but hydrophobic interaction increases at high ionic strength (Figure 6.11). It
is estimated that 40–50% of the surface area of a protein is non-polar.
The HIC parameters are type of ligand, degree of substitution, concentration of
salt, and effect of temperature and pH.
The immobilized ligands used are hydrocarbon groups like butyl and octyl groups
and phenyl groups. The polarity of the ligand increases with alkyl chain length and its
degree of substitution. The interaction of the phenyl group is not simple because of
an aromatic effect as well as hydrophobicity.
The most typical salt in HIC is ammonium sulfate. As the concentration of
ammonium sulfate is increased, the amount of protein adsorbed on the ligand
increases linearly up to the precipitation point. Table 6.4 shows the effect of the ion
used in HIC on the precipitation of proteins. Sodium, potassium, or ammonium
sulfates have a relatively high salting-out effect and the molar surface tension of water
an increasing effect.
Ammonium (1 M) sulfate is a good starting point for experiments. If the protein
does not retain the ligand, a more hydrophobic ligand should be selected. The
recovery of protein in HIC should be > 80%. When a small amount of miscible
organic solvent is needed, the ligand should be changed to a less hydrophobic one.
Hydrophobic interaction in HIC is diminished by increasing the pH and increased
by decreasing the pH. The pI of protein is in the acidic range and the hydrophilicity
206
Table 6.4 Effect of some anions and cations in precipitating proteins [4].
Protein Antichaotropic effect $ chaotropic effect
Ribonuclease SO42< CH3COO< Cl< Br< NO3< CIO4 NHþ þ

4, K ,
þ þ
Na < Li < Ca 2þ
Collagen zelatin SO42< CH3COO< Cl< Br< NO3< CIO4< I< SCN NHþ þ
4 < Rb ,
þ þ þ þ þ
K , Na , Cs < Li < Mg2 < Ca < Ba
2þ 2þ
increases in the basic range. The hydrophobic interaction strength changes strongly
at a pH below 5 or above 8.5. In addition, on increasing the temperature of HIC, the
hydrophobicity slightly increases. A small amount of miscible organic solvent affects
a decrease in hydrophobicity of protein and facilitates elution in the buffer solution.
6.7.1.3 Gel Filtration (GF)

Gel filtration (GF) [9] is a key method in the purification of enzymes as well as
biological macromolecules. It is a reliable and simple separation technique that does
not involve adsorption or interaction on the GF media. In gel filtration, the principle
of separation is very simple: macromolecules in solution are separated based on
differences in their size as they pass through a column (Figure 6.12). Large molecules
Figure 6.12 Principle of gel filtration.

pass through the stationary phase first while smaller molecules diffuse further into
pores of the GF medium and, thereby, take longer to elute. Gel filtration is also called
molecular-sieve chromatography. Molecules are eluted in the order large to small. Gel
filtration is usually deployed at the final or latter stages for changing buffer and
concentration.
Gel filtration is carried out using a single buffer solution of appropriate pH and
ionic strength. Some GF media have a small number of ionic charged groups, such as
carboxyl and sulfonic groups, which sometimes cause non-specific adsorption of
basic proteins at low ionic strengths. To avoid such adsorption, gel filtration should be
carried out at an ionic strength above 0.15 M. Non-ionic interactions between
proteins and gel filtration media are negligible at buffer concentrations between
0.15 and 1.5 M. An ionic strength below 0.15 M causes a slight retardation of basic
proteins and exclusion of acidic proteins.
The first GF medium Sephadex, provided by Pharmacia, was a bead-formed gel
prepared by crosslinking dextran with epichlorohydrin. GF media with various
particle size grades are now available and globular proteins of sizes between 700

and 4 107 A can be separated. The fractionation range of the medium determines
the porosity of the gel and is measured using typical globular biological molecules or
dextrans. The shape (its diameter and length) of biological molecules affects the
theoretical separation. When the molecule is not globular but a linear string, the
separation is quite different.
Choice of Column, Sample Volume, and Flow Rate Several factors affect the choice of
column equipment in order to obtain a good separation. The length of the column
must more than 30 times the diameter, because the resolution increases at the square
root of column length. This is why a longer column is used for gel filtration, especially
for analytical fractionations. A bed length of more than 1 m is not useful and effective
for industrial separation.
The dead volume at the inlet and outlet should be less than 0.1%. A sample volume
of 0.5–5% of the bed volume is recommended for good resolution and depends on the
GF medium. The relationship between sample volume, medium, and resolution has
been described; however, the actual sample volume should be determined by
experiment. A smaller sample size is not good for resolution. Up to 30% of the
total bed volume can be applied for changing the buffer and salting out.
An effective flow rate for resolution of the order of 5 ml cm2 h1 and up to about
25 ml cm2 h1 is allowed for industrial preparations. The length of the column and
the flow rate are basically in an inverse relation.
6.7.1.4 Reversed-Phase Chromatography

Except for a few specific applications, reversed-phase chromatography (RPC) [10] is
rarely used in biological purification. RPC is commonly used for the purification of
organic compounds and low molecular weight peptides. The principle of RPC is
similar to that of HIC. The ligands are stronger in RPC than in HIC, and include
octadecyl, octyl, butyl, and phenyl groups (Scheme 6.1). The remaining silanols are
quenched with trimethylsilyl groups.
208
Scheme 6.1 Reaction of a silanol with an octadecyldimethylsilyl group.
The RPC medium consists of hydrophobic ligands chemically bonded to porous

microbeads. The microbeads are made of silica gel or a synthetic organic polymer
such as polystyrene.
It is necessary to use water-miscible organic solvents to elute proteins (Table 6.5).
Most proteins are, consequently, apt to be denatured. RPC is useful for the
purification of small samples or peptides and is usually carried out as high-
performance liquid chromatography (HPLC).
6.7.1.5 Hydrogen Bond Chromatography

There are three types of chemical interaction between ligands and proteins: ionic,
hydrophobic, and hydrogen bond interaction. Hydrogen bond chromatography is not
as popular as ionic and hydrophobic chromatography. Precipitation of proteins is
sometimes observed in the presence of a water-soluble polysaccharide and poly
(ethylene glycol). The complex with proteins easily forms via hydrogen bonding at
high ionic strength. Ionic cellulose, such as DEAE-cellulose and CM-cellulose, as well
as cellulose, is used as a matrix for this purpose. Hydrogen bond–ion chromatog-
raphy is complicated because the ionic strength of the buffer solution used for each of
the two methods is opposite to elute proteins.
Table 6.5 Solvents used in reversed-phase chromatography.
Solvent Dielectric constant (20 C) Viscosity (cP at 20 C) Bp ( C)
Acetonitrile 38.8 0.36 82

Ethanol 24.3 1.20 78
Methanol 33.6 0.06 65
Propanol 20.1 2.26 98
Isopropanol 18.3 2.30 82
Water 80.4 1.00 100
Protein is adsorbed on DEAD-cellulose in buffer solution with 3 M ammonium
sulfate and is eluted by decreasing the concentration of ammonium sulfate or by
adding releasing reagents, such as urea and sucrose. Sodium formate and sodium
acetate are used instead of ammonium sulfate. On the other hand, small amounts of
ethanol, glycerol, and ethylene glycol are available for elution.
6.7.1.6 Affinity Chromatography

Figure 6.13 shows the principle of affinity chromatography.
Affinity chromatography [11] is used for a biologically specific ligand bound to
the matrix. The protein binds with ligand specifically in an active form and the rest
of the material passes through without adsorption upon washing with a buffer
solution.
The ligand should have specificity and reversibility for the protein and release it on
affinity elution or change of ionic strength and pH. Interactions between proteins and
ligands include ionic binding, hydrogen binding, and hydrophobic binding. The
Figure 6.13 Principle of affinity chromatography.

210
factors necessary for ionic binding have been listed above. The effect of ionic strength
on ionic binding is opposite to that on hydrophobic binding and the recovery of
protein is sometimes not good. The use of a 1–3 M urea solution or 5–20% sucrose is a
good idea in such a case.
Affinity chromatography is carried out by both batch and column methods. The
procedure involves (i) equilibration of the adsorbent, (ii) preparation of sample, (iii)
application of the sample, (iv) washing away of unbound materials, (v) elution, and
(vi) regeneration of adsorbent.
When the ligand has a simple specificity for protein, about 90% purified protein is
obtained by one-step purification. Consequently, affinity chromatography is a rev-
olutionary purification method. Adsorbents are relatively expensive and affinity
chromatography is useful for small-scale purification.
There are several types of affinity chromatography. Two typical types, immobi-
lized dye chromatography and metal chelate affinity chromatography, are described
below.
Immobilized Dye Chromatography Immobilized dye chromatography is the most

useful affinity chromatography. Its ligands are synthetic polycyclic dyes. These
structures are very similar to the cofactors NADþ and NADPþ as a dinucleotide
analog, which are apt to bind strongly with a protein like kinases, dehydrogenases,
and so on (Figure 6.14). Some of the proteins bind biospecifically with the
dye because of its structural similarity to NAD and NADP. Some proteins like
lipoproteins and albumin bind in a less specific manner by electrostatic and
hydrophobic interactions with the aromatic anionic ligand. Bound proteins can be
eluted by affinity elution using low concentration free cofactors. In contrast, non-
specifically bound proteins need a much higher concentrate of cofactors or ionic
strength.
Metal Chelate Affinity Chromatography Metal chelate affinity chromatography is a

kind of separation method that has, as a ligand, a metal ion. Some proteins and
peptides are purified on the basis of affinity for metal ions immobilized by chelation
on the adsorbents. Histidine and cysteine form complexes with the chelated
Figure 6.14 Example of a ligand employed in immobilized dye chromatography [11].

metals around neutral pH. Biological proteins include many histidines as well as
recombinant proteins as polyhistidine fusions; for instance, His-tag proteins have a
specific metal chelate affinity. The adsorbent is prepared by coupling a metal chelate
ligand with an iminodiacetic acid group, which forms a chelate with divalent metal
ions such as Zn2þ , Cu2þ , Cd2þ , Hg2þ , Co2þ , Ni2þ , Fe2þ , and so on.
Elution is carried out by reducing the pH and increasing the ionic strength of
the buffer or by adding EDTA to the buffer. The most typical method is to gradually
add sodium chloride (0.5–1.0 M) or imidazole (0–0.05 M). This ligand is very
expensive, so metal chelate affinity chromatography is used only for small-scale
purification.
His-tag proteins produced by a recombinant are easily purified by metal chelating
chromatography. They have about six histidines at the N- or C-terminal site and
the His-tag easily forms a chelate with Ni2þ , Zn2þ , and Cu2þ ; elution of His-tag
proteins is carried out by increasing the concentration of imidazole in the buffer
solution.
6.7.1.7 Salting-Out Chromatography

Salting out is popular for the purification of proteins. Salting-out chromatography is a
precise method based on the same principle; however, it is not popular. Both positive
and negative salting-out chromatography are carried out. The former is a combination of
molecular-sieving chromatography and salting out. Proteins in buffer solution are
applied to a concentration gradient column of salts. With the latter method, the
precipitation of proteins salted out in the presence of Celite fills the column and
proteins are eluted with a flowing buffer solution by decreasing the concentration of salt.
6.7.2
Precipitation
Among the methods of purifying protein, precipitation is the most useful and typical
for both small- and large-scale procedures. Precipitation methods are classified into
four types, salting-out, organic solvent precipitation, pH changing precipitation, and
water-soluble precipitation. The precipitation is usually carried out at early stage and
the total protein concentration should be > 0.1 mg ml1
6.7.2.1 Precipitation by Salting Out

The solubility of macromolecules such as proteins in water generally increases in the
presence of a suitable concentration of salt, so-called salting in. Furthermore,
increasing the concentration of salt further leads to a decrease in the solubility of
the proteins and their precipitation (salting out).
Salting out depends greatly on the pH and temperature of the solution. Proteins
show minimum solubility around their isoelectric point (pI) in water and a little lower
solubility in buffer solution, in other words in the presence of salt. Regarding
temperature, the solubility of proteins generally decreases at higher temperature in
buffer solution with higher ionic strength.
212
Many salts are used for salting out, including ammonium sulfate, sodium sulfate,
potassium phosphate, magnesium sulfate, sodium citrate, and sodium chloride. The
solubility of these salts is independent of temperature, and the salts do not affect the
denaturation of the proteins. Ammonium sulfate is the most effective salt for salting
out because of its high solubility at any temperature and its low cost; it is also a useful
stabilizer for proteins.
6.7.2.2 Precipitation by Organic Solvents

Water-miscible organic solvents, such as ethanol, isopropanol, and acetone, reduce
the solubility of proteins by decreasing the dielectric constant of an aqueous solution
and taking away water from around the proteins. In addition, these organic solvents
can remove the lipids bound to a protein. Precipitation by organic solvents is affected
by temperature, ionic strength, and the pH of the buffer solution.
Common proteins are precipitated at about 40% in ethanol, but proteins with hydro-
phobic surface are soluble, like lipases, under such conditions. Alcohol concentrations of
80–90% are necessary to precipitate lipases. The concentration of protein should be > 1
mg ml1 and that of the buffer solution < 50 mM. The solution and organic solvent should
be cooled and the mixture kept at below 0 C during the addition of organic solvents.
6.7.2.3 Precipitation by Changing pH

There are two types of precipitation by pH change, isoelectric point (pI) precipitation
and acidic precipitation.
The use of pI precipitation is suitable for a protein with very low solubility and is
more effective in combination with salting-out and organic solvent precipitation.
Anions bind with proteins more easily than cations, so the pI of proteins shifts a little
to the acidic range.
On the other hand, acidic precipitation is good when the protein is stable, but
impure proteins are unstable in the acidic range.
6.7.2.4 Precipitation by Water-Soluble Polymer

Precipitation by water-soluble polymer is a simple method for the purification and
crystallization of proteins. Many proteins are easily precipitated in the presence of
water-soluble non-ionic polymers such as poly(ethylene glycol)s (PEG 2000, 4000,
6000), methyl cellulose, poly(vinyl alcohol) (PVA), and dextrans (DEX).
These water-soluble polymers remove water from around proteins. The proteins
bind with these polymers via hydrogen bonds, and then the complex precipitates as a
solid or sometimes becomes a viscous liquid. Hydrogen bond chromatography is
based on this principle.
The complex contains water-soluble polymers, which must be removed by ionic
chromatography, salting out, ethanol precipitation, electrophoresis, and so on.
6.7.3
Crystallization
Relatively purified proteins are easily crystallized at > 1%, usually 5–10%, of the protein
concentrationinbuffer. Consequently, crystallizationisthe final stage of purification, and
useful for storage of proteins and X-ray crystal structure analysis. In protein chemistry,
crystallizationdoesnotmeantheproteinis100%pureeventhoughitisincrystallineform.
As described for salting out, a crystallized protein is in a solid state together with
precipitation aids such as salts, organic solvents, water-soluble polymers, and so on.
Freeze drying is one crystallization method; however, denaturation, deactivation,
or a slight change in the three-dimensional structure of a protein is sometimes
observed. It is necessary to check the stability before freeze drying.
6.7.4
Stabilization During Purification
Care must be taken not to lose the activity during purification of the enzyme after
fermentation. Enzymes are macromolecules influenced by changes in pH, temper-
ature, concentration of buffer and salts, metal ions, detergents, organic solvents, and
so on. To preserve their activity, enzymes should be kept under natural physiological
conditions such as a low temperature of about 4 C, natural pH for the enzyme,
physiological buffer solution and concentration, and so on. Some additives for
enzyme stabilization are used during purification. Mercaptoethanol and dithiothrei-
tol work as antioxidants and EDTA works as a chelating agent to prevent inactivation
by heavy metal ions and metalloproteases.
Polysaccharides like dextrin, sugars, sugar-alcohols like sorbitol and mannitol, glycerol,
and ethylene glycol are sometimes used as stabilizers. Some peptides and amino acids are
useful excipients for purification. Compounds with a similar structure to that of the
substrate are generally effective as stabilizers and are used as fillers for storage.
Degradation by proteases derived from the same microorganism or from con-
tamination during purification must be avoided. Once a protease contaminates an
enzyme solution, the desired enzyme is degraded during purification and might
disappear. To prevent degradation by proteases, it is helpful to add protease inhibitors
like PMSF (SH protease) and EDTA (metal protease).
6.7.5
Storage of Enzymes
6.7.5.1 Storage in Liquids

Common enzymes in liquid form should be stored below 4 C in a refrigerator and
kept with a stabilizer. Most enzymes retain their activity for several years under
suitable conditions, especially thermostable enzymes.
Ammonium sulfate (2 M) is a popular storage solution for commercial porcine
liver esterase (PLE). It prevents microbial growth on the solution. Storage in 50%
glycerol is also useful and such glycerol stock can be stored below 0 C.
6.7.5.2 Storage in Solids

Solid forms for storage are preferred for commercial enzymes. Generally, an enzyme
is much more stable in solid form than in liquid form even without a stabilizer. A
solid form for storage is prepared by precipitation with organic solvents and freeze-
214
drying or spray-drying depending on the purification stage. Precipitation using an

organic solvent is convenient, but the purity is not so high. Freeze drying is very
useful but expensive. Spray drying is preferable for commercial enzymes. Spray
drying is commonly carried out at about 120–140 C, for which the enzyme needs
moderate thermostability.
Stabilizers are effective in avoiding loss of activity and the typical stabilizers
described above are used during precipitation and crystallization.
Some enzymes in solid form are very stable and can be stored at room temperature
for several years without loss of activity.
6.8
Commercial Biocatalysts
Among biocatalysts, hydrolases like lipases and proteases are the most popular. There
are several types of biocatalysts in commercial products. Immobilized lipases and
crosslinking enzymes are briefly described in this section.
The most popular immobilization method is adsorption on a carrier such as
diatomaceous earth or a synthetic polymer. The advantage of this method is that the
original activity of the enzyme is maintained, but the disadvantage is that the enzyme
cannot be used in an aqueous solution.
Lipases immobilized on ceramics modified with a chemical silyl reagent adsorb
strongly and can be used in aqueous solutions as well as organic solvents. The activity
is sometimes ten times that of the original and the thermostability is also increased.
These products can be reused more than ten times, depending on conditions.
Crosslinked enzymes are commercial biocatalysts and can be reused in organic
solvent and aqueous solution. They are purchased as crystals derived from a single
crosslinked enzyme.
Some screening kits are provided for user convenience. The main suppliers are
listed in Chapter 46.
References
1 Deutscher, M.P. (ed.) (1990) Methods in 5 Drauz, K. (ed.) (1995) Enzyme Catalysis in
Enzymology, vol. 182, Academic Press, San Organic Synthesis, VCH, Weinheim.
Diego. 6 Amersham Pharmacia Biotech (1999)
2 Jakoby, W.B. (ed.) (1984) Methods in Purification for Proteins: Principles and
Enzymology, vol. 104, Academic Press, San Methods, APB, Uppsala.
Diego. 7 Amersham Pharmacia Biotech (1999) Ion
3 Godfrey, T. (ed.) (1996) Industrial Exchange Chromatography: Principles
Enzymology, Macmillan Press Ltd, and Methods, APB, Uppsala.
London. 8 Amersham Pharmacia Biotech (1999)
4 Horio, T. (ed.) (1994) Theory and Practice Hydrophobic Interaction
on Enzymes and Other Proteins, Nankodo, Chromatography: Principles and
Tokyo. Methods, APB, Uppsala.
References j215
9 Amersham Pharmacia Biotech (1998) Gel Chromatography: Principles and
Filtration Chromatography: Principles Methods, APB, Uppsala.
and Methods, APB, Uppsala. 11 Amersham Pharmacia Biotech (1999)
10 Amersham Pharmacia Biotech Affinity Chromatography: Principles and
(1999) Reversed Phase Methods, APB, Uppsala.
j217
7
Reaction and Process Engineering
John M. Woodley
7.1
Introduction
7.1.1
Scope and Background
In recent years, the use of biocatalysis for the synthesis of chemical products has
become increasingly widespread in the chemical industry. To date several hundred
biocatalytic processes have been reported that operate at a commercial scale and the
excellent examples given in other chapters are a clear testament to the fact that many
chemical products made today involve one or more enzymatic steps. The techno-
logical advances in molecular biology and developments in recombinant DNA
technology in particular have reduced the costs of enzyme production. In many
cases this has led to the development of commercial processes that previously were
considered infeasible. In addition, the drive for higher selectivity and more sustain-
able processes means that the trend of implementing enzymatic processes will
increase in the coming decades [1]. Biocatalysis addresses many of the sustainability
requirements of tomorrows chemical industry [2, 3] since they frequently operate at
moderate temperatures and pressures, using renewable feedstocks and catalysts and
usually no toxic chemicals in the process.
This chapter addresses the issues required to design and scale-up a biocatalytic
process. The scope is restricted to the use of enzymes (biocatalysis – use of either
isolated or immobilized enzymes as catalysts for the synthesis of chemical products),
although complementary approaches using growing microbial cells (fermentation)
or resting microbial cells (biotransformation) also have important roles in present
and future chemical production. Many other texts deal with fermentation and whole-
cell based reactions, which find use in particular types of reaction, especially where
multiprotein complexes or the use of membrane-bound enzymes makes isolation
difficult, along with many of the cofactor-requiring enzymes [4]. Furthermore, for
integration of biocatalysis into organic synthesis it is perhaps the isolated enzymes
that have attracted most interest in recent years in synthetic routes since it is here that
the benefits of selectivity and clean product streams are most easily found. For higher
and Oliver May.
j 7 Reaction and Process Engineering
218
value products (such as pharmaceuticals) it is particularly important to have a clean

product stream. For this reason it is common to immobilize the enzyme (as discussed
in other chapters) and this also enables recycle and reuse (often essential for
economic use). Several methods for immobilization are available for use at scale.
Frequently, the enzymatic step is a relatively small part of the overall synthetic process
but it nevertheless requires adequate attention along with suitable integration with
the other synthetic and product recovery steps.
Biocatalysts are frequently the preferred choice of catalyst when high selectivity is
required. Potentially, the introduction of biocatalysis can reduce the total number of
processing steps and in particular avoid protection and deprotection steps, leading to
higher atom efficiency [5]. Based on published reports reviewing the application of
biocatalysis in industry it is clear that most products from industrially implemented
biocatalytic processes to date are chiral compounds [6–9]. However, biocatalysis is not
only interesting for high-value, low-volume products like chiral pharmaceuticals but
also for specialty and effect chemicals, like surfactants, as well as for bulk chemicals
and biofuels. Indeed the impact of biocatalysis in other industry segments
is increasing. A consequence of the use of biocatalysis to produce lower value
products is that the scale of operations will increase. As the scale of biocatalytic
processes increases, more emphasis will be required on the chemical engineering
and bioprocess engineering considerations, alongside the necessary biotechnolog-
ical developments. In particular, the requirements regarding process intensity and
cost reduction are more demanding for high-volume chemicals and biofuels.
7.1.2
Role of Reaction Engineering
While it is obvious that the implementation of an enzymatic reaction into an

industrial process requires adequate reaction and process engineering, for many
processes it is hard to justify the considerable effort in process optimization required
until the product and suitable reaction scheme is fixed or at least limited to a few
alternatives. Nevertheless appropriate reaction and process engineering at an early
stage in process development may give useful information about the economic
feasibility of a process at an early stage. Establishing the right balance between early
stage process studies and sufficiently detailed work is a matter of judgment that is
critical for effective implementation of many processes. Most processes are still
designed on a pragmatic case-by-case approach although several efforts have been
made towards more systematic methods (e.g., Reference [10]). Likewise the type of
process studies required depends on the industry sector. For example, pharmaceu-
tical processes that are characterized by a high product attrition rate and the need to
rapidly implement a process to get the product on the market as quickly as possible
after patent approval clearly require minimal process studies. In contrast, the
production of bulk chemicals using biocatalysis requires far greater optimization
to be economically viable and the development time is not so critical.
A further issue concerns the type of plant. In some cases the main aim of process
engineering is to fit the enzymatic reaction into existing multipurpose plant
7.2 Reactor Options and Characteristics j219
(usually highly flexible and operated in batch or semi-batch mode). The alternative
can be the design of a dedicated plant, with specifically designed reactors and
separation tasks to make optimal use of the enzyme.
7.1.3
Applications
The application of enzymes to assist organic synthesis on an industrial scale has been
an active area of research for several decades [11, 12]. Since the late 1960s,
immobilized enzymes have been used in amino acid production in continuous
processes on a large scale [13, 14]. In the late 1970s, the use of soluble enzymes,
especially in membrane reactors, broadened the scope of enzyme technology [15, 16]
and opened the way to simultaneous use of more than one enzyme for complex
conversions – especially coenzyme-dependent biotransformations [17–19]. In the
early 1980s the use of enzymes was extended further to involve organic solvents [20,
21] and biphasic mixtures [22]. Multi-enzyme conversion [23] and chemoenzymatic
conversions [24] have again become areas of particular research interest in recent
years, driven in part by new technological developments and new tools for evaluation.
Enzymes are also used as catalysts for large scale bioconversions such as glucose
isomerase in the high fructose corn syrup (HFCS) process [25], penicillinase in the
synthesis of semi-synthetic penicillins [26], as well as amino amidase [27] and
aminoacylase [28] in the production of L-amino acids. Additionally, various processes
for fine chemical synthesis have been developed, for example, for amino acids,
peptides, and a broad spectrum of other optically active substances [5, 8, 29–35]. More
recent studies have examined other classes of enzyme such as transaminases to
prepare them too for process evaluation and scale-up [36].
Based on the classical methods of enzyme isolation and characterization and the
screening for appropriate microorganisms, about 3200 different enzymes are known
today and are listed with EC numbers. Modern methods of genetic engineering give
access to sufficient quantities of enzymes by overexpression in microorganisms, thus
reducing the cost of enzymes [37–40]. Enzyme reaction engineering allows further
reduction of the amount of enzyme consumed per kilogram of product. Therefore,
the cost of enzymes does not necessarily dominate the overall cost of production, but
it is often, still, a major factor.
7.2
Reactor Options and Characteristics
7.2.1
Introduction
In many cases the scale of production means that the opportunity to choose a
dedicated reactor for a specific reaction is not always available and, therefore, it is
often necessary to match the specific enzymatic reaction to a given reactor. Here the
220
principle features of the main reactor types are described. There are several reactor
types available for enzymatic reactions, all of which offer specific benefits and
drawbacks [6, 41]. The aspects that need to be considered are cost, space, mass-
transfer, kinetics, heating and cooling, ease of operation, and reusability of the
catalyst [42, 43]. The criteria for selection of an appropriate reactor for a particular
enzymatic reaction were first reviewed in the 1970s [44, 45] and the 1990s [43]. The
basic principles remain the same and this early work guided several large-scale
processes, although technological developments have subsequently widened
the options for biocatalytic reactors and their operation. This section discusses the
ideal reactor types, followed by an explanation of alternatives.
7.2.2
Ideal Reactor Types
Three ideal types of reactor exist: the batch stirred tank reactor (BSTR), the
continuous stirred tank reactor (CSTR), and the continuous plug flow reactor (CPFR).
The definitions are based on the chemical engineering principles of mode of
operation (batch or continuous) and the flow pattern inside the reactor (well mixed
or plug flow). Several excellent and well-established texts describe the basic princi-
ples [46, 47]; Figure 7.1 shows schematic representations of the three ideal reactor
types.
The batch stirred tank reactor is well mixed, implying no concentration gradient or
alteration of concentration spatially in the reactor. At the start of the reaction substrate
and enzyme is loaded and the reaction is allowed to proceed at a rate determined by
the kinetics of the reaction. Most reactions obey the so-called Michaelis–Menten
Figure 7.1 Schematic representation of the and product, respectively. E represents the
three ideal reactor types: (a) BSTR (batch stirred enzyme, which may be in soluble or
tank reactor), (b) CSTR (continuous stirred tank immobilized form (on a solid or porous
reactor), and (c) CPFR (continuous plug flow support).
reactor). S and P represent the flow of substrate
Figure 7.2 Reaction profiles for reaction time and for a continuous reactor
Michaelis–Menten kinetics: (a) BSTR, (b) CSTR, (CSTR, CPFR) the volumetric flow rate/reactor
and (c) CPFR. Tr represents the reactor volume. [S] ¼ substrate concentration and
residence time – for a batch reactor (BSTR) the [P] ¼ product concentration.
kinetics with mixed order kinetics meaning that the reaction is fast initially and slows
towards the end of the reaction (Figure 7.2). In a BSTR it is possible to achieve
complete conversion provided the reaction time is sufficient. Achieving complete
conversion is very important for reactions where high selectivity is required or where
the yield from the reactant is important for economic reasons. The batch stirred tank
is the most commonly used reactor set-up, owing to its simplicity, equipment
flexibility, and ease of operation. A drawback at large scale is the low volumetric
productivity and that immobilized enzymes will be exposed to mechanical stress
222
from the stirring, which could lead to the physical loss of the enzyme preparation and
thereby contamination of the product along with significantly decreased catalytic
activity [48–50]. Another difficulty is how to deal with the inevitable gradual decline in
enzyme activity after many cycles [51]; either the reaction conditions (e.g., residence
time or temperature) have to be adjusted or more enzyme must be added to the
reactor as the activity decreases. Both of these strategies will naturally only be feasible
to a limited extent. Feeding to a batch reactor is a common variant on the theme of a
BSTR to overcome toxicity or inhibition problems [52].
The continuous stirred tank reactor (CSTR) operates in the same way as the BSTR
except with a continuous input and output stream. Since the reactor is well mixed the
output stream is at the same concentration of components as that in the reactor and
since the reaction must proceed at an adequate rate it is necessary that there is some
substrate in the reactor (and therefore the output stream). This has two very
important implications. First it means that the kinetics will always be governed by
the reaction rate at the leaving substrate concentration. Inevitably, the reaction is slow
and therefore to achieve a given productivity more enzyme is required. This is
emphasized when higher conversions are required since the reactor operates at lower
substrate concentrations. The second implication is that complete conversion cannot
be achieved in this type of reactor (Figure 7.2).
The continuous plug flow reactor (CPFR) is fundamentally different from the
CSTR and BSTR in that the material in the reactor is not mixed, implying a
concentration gradient from the start to the end of the reactor length. To achieve
this no material must mix with anything before or after it in the reactor, implying that
the material flows as a slug or a plug down the reactor (Figure 7.2). The kinetic
profile is identical to the BSTR except that we replace the time variable with the length
variable. For simplicity in Figure 7.2 this is shown as the residence time (volumetric
flow-rate/volume). Hence, provided the reactor is long enough it is possible to obtain
complete conversion. The fixed-bed reactor, behaving as a plug flow reactor, is most
often used for immobilized enzyme reactions. Typically, the reactor is used with an
upward direction of the flow to avoid compression of the bed and to release any gas
bubbles generated during the reaction. The flow rate through the bed will determine
not only the extent of conversion for a given enzyme activity but also the pressure
drop over the bed. Some work is required to optimize the system. The fixed or packed-
bed reactor is essentially an alternative set-up for running enzymatic conversions
using immobilized biocatalysts [53]. The benefits over the stirred tank are, generally,
the lower investment cost and higher volumetric productivity and that it can be run in
a continuous mode. Possible problems of mechanical shear forces are eliminated [54],
and separation of the enzyme from the product is simplified. The shorter residence
time in the reactor can also lead to fewer side reactions. On the other hand, drawbacks
of using a PBR (packed-bed reactor) could be internal and external mass transfer
limitations, channeling over the bed and high pressure drops over the bed (depen-
dent on the support). Further, adjustments of pH or in situ product removal or
addition of substrates become quite complicated. One way to deal with this problem
is by running several PBRs sequentially with intermittent product removal or
addition of substrate [51]. Alternatively the packed-bed holding the biocatalyst could
Table 7.1 Reactor selection criteria based on metrics.
Criteria BSTR CSTR CPFR
Conversion High Limited High

[E] (enzyme concentration) Limited to 10% Limited to 10% Up to 66%
Yb ) High requires RBP High requires low Xa) High
[P] High Limited High
a) Yb ¼ yield on biocatalyst (kg-product per kg-biocatalyst).

b) X ¼ fractional conversion of substrate (dimensionless).
be attached to the reactor via a loop [53]. The use of a continuous reactor (whether
CSTR or CPFR) requires the retention of the enzyme in the reactor. Usually, this is
done by immobilization of the enzyme onto a support, which means it can be
filtered like a normal heterogeneous catalyst. In well-mixed reactors up to 10% by
volume can be the solid catalyst without damage from the agitator or other particles
leading to abrasion. In the CPFR if the solid catalyst is packed in the form of a bed
then up to 66% by volume can be catalyst. This can have significant implications on
the space–time yield (STY) of these reactors [43]. Table 7.1 compares the reactor
types.
7.2.3
Use of Multiple Reactors
In some cases it may be beneficial to use more than one reactor. For example, two or
more sequential CSTRs can be used to operate each reactor at different reactant
leaving concentrations, thus saving enzyme. Nevertheless some extra considerations
are required, such as the increased capital cost per volume of reactor. In theory, at
around 30 sequential reactors plug flow conditions can be approximated. Clearly,
such a large number of reactors could never be justified but up to four reactors can
find application in cases where high conversion is required and the cost of the
enzyme is relatively high [55]. A second example concerns the use of a CSTR followed
by a CPFR to polish the product and convert all the remaining reactant to provide an
adequate level of conversion for the downstream process. Where separation of
reactant from product is difficult this may be justified. Figure 7.3 illustrates these
alternatives.
7.2.4
Addition of Reagents
In some reactions mixing is required due to the need to add reagents in a controlled
manner during the reaction (feeding). In many cases reactants need to be fed so as to
avoid toxic or inhibitory concentrations in the reactor from the start. Likewise it is
often necessary to control pH, given that many reactions produce or consume an acid
or base, and critically enzyme activity and stability is greatly affected by pH. At an
224
Figure 7.3 Multiple-reactor configurations: In the exit concentrations from each reactor. In
configuration (a) plug flow operation is configuration (b) a plug flow reactor (reactor 2)
simulated by operation with three CSTRs in is used to polish the stream coming from a
series. Such a scheme benefits cases where the CSTR (reactor 1), thereby overcoming the
enzyme cost is dominant. The reactor sizes are inevitable low conversion in a CSTR for cases
not necessarily equal and depend upon the where the enzyme cost is significant.
kinetics in each reactor, which in turn is set by
industrial scale a small amount of buffer may be used to stabilize the system but most
of the pH control will be achieved by titration of acid or base as necessary. This
requires a well-mixed system (BSTR or CSTR) so as to avoid regions in the reactor of
high or low pH. In a CPFR such a system can only be implemented via a recycle loop
where the pH is adjusted in an adjacent stirred tank. Mixing is also required for
multiphasic systems (those with solid catalyst, insoluble substrate, organic solvents,
and/or gases). In fact most processes operated in a scalable process exhibit one or
more of these situations and, therefore, require some mixing to adequately suspend
catalyst or create enough interfacial area for good mass transfer. Experimental studies
focused on simulating large-scale mixing at a smaller scale using the so-called scale-
down approach [56] may prove useful to help test systems in the future. Likewise, new
mixing methods such as rotary jet head mixers [57] and more conventional impellers
such as marine propellers (for axial flow) or Rushton turbines (for radial flow) can be
applied using expertise from fermentation studies [58]. Power inputs should
not exceed 2 kW m3 for scale-up. At intermediate scales and where intense mixing
is not necessary, the use of gas can be used in some cases to provide adequate mixing.
7.2.5
Alternative Reactors for Insoluble Enzymes
In addition to the three ideal reactor types a range of other reactors can be used,
including membrane reactors, bubble columns, fluidized beds, and expanded beds.
The mode of operation is also important and variations of the ideal reactors also exist
for fed-batch operation and intermittent feed and bleed situations.
Fluidized-bed reactors are advantageous if small particles that would give high flow
resistance in a fixed-bed reactor are used to minimize internal and external mass
transfer limitations or if a further phase is present (such as a gas, insoluble reactant,
or second organic phase). Often it is useful to install nets at different heights in the
reactor to approach plug flow characteristics. If the immobilized enzyme particle size
is so small (and/or the density too low) that effective retention is not possible in
fluidized bed reactors a slurry reactor may be used instead, where the catalyst passes
through the bed. For larger particle sizes, the use of a stirred tank reactor is not
advantageous because the energy input necessary to give an optimal suspension of
the particles is much higher than in a fluidized-bed reactor. Aside from the
immobilization of enzymes on solid (or more normally porous) particles, enzymes
may also be immobilized on the inner or outer surface of tubular supports such as
hollow fibers or flat membranes. The enclosure of enzymes by the use of an
ultrafiltration or dialysis membrane may therefore also be regarded as a form of
immobilization. Figure 7.4 shows some alternative reactor configurations.
The same reactors can be used to deal with immobilized enzymes in organic
solvents or with single-phase organic systems in the same way as dealing with
enzymes in aqueous solutions. For single-phase systems, the enzyme may be
recovered from the solution by means of membrane filtration. Suspended enzyme
particles may be retained in a slurry reactor by microfiltration membranes or
stainless steel sieves. In recent years ultrafiltration membranes that are stable toward
organic solvents have become available (e.g., from polyaramide or cellulose). In these
cases the enzyme membrane reactor (EMR), as described earlier for the pure aqueous
system, may be used without modifications, if all materials (sealing rings, tubes, etc.)
are stable toward the solvent used.
7.2.6
Alternative Reactors for Soluble Enzymes
If soluble enzymes exhibit sufficient operational stability, their use is advantageous,

as the cost and effort of immobilization and resulting mass transfer limitations can be
avoided. Different techniques have been developed to retain soluble enzymes. For
226
Figure 7.4 Alternative reactor configurations: (a) fluidized bed, (b) slurry, and (c) immobilized
enzyme membrane reactors.
synthesis at a large scale, repetitive batch processing (RBP) has proven to be an

effective and easy-to-handle method [59]. The repeated use of the enzyme is possible
after concentration of the solution by means of commercially available ultrafiltration
equipment and adding fresh substrate solution (Figure 7.5). Up to 10 cycles may be
required, depending on the maximum allowable cost contribution of the enzyme to
the product. (For an immobilized enzyme up to 700 cycles would be required.)
Continuous processes often show a higher space–time yield than batch processes.
Reaction conditions may be kept within certain limits more easily. For easier scale-up
Figure 7.5 Schematic representation of repetitive batch processing (RBP).

of some enzyme-catalyzed reactions, the enzyme membrane reactor (EMR) has been
developed. The difference in size between an enzyme (which as a protein will have a
very high molecular weight) and the substrates, enables continuous homogeneous
catalysis to be achieved while retaining the catalyst in the vessel. For this purpose,
commercially available ultrafiltration membranes are used. When continuously
operated, the EMR behaves as a continuous stirred tank reactor (CSTR) with
complete backmixing (well mixed). For large-scale membrane reactors, hollow-fiber
membranes or stacked flat membranes may be used [60]. To prevent concentration
polarization on the membrane, the reaction mixture is circulated along the mem-
brane surface by a low-shear recirculation pump. Some of the implications of EMR
and RBP implementation are listed below:
. Advantages:
– working under sterile conditions is easy to achieve;
– no loss of enzyme activity by immobilization;
– no mass transport limitation;
– use of multi-enzyme systems with easy replacement;
– use of coenzymes without mass transport limitation;
– simple addition of fresh enzyme to compensate for enzyme deactivation;
. Disadvantages:
– enzyme stability limited by solution deactivation;
– well-mixed operation may not suit high product concentration requirements.
7.2.7
Reactors for use with Multiphasic Systems
Two common types of multiphasic system exist in addition to the presence of a solid-
supported biocatalyst in an aqueous solution. The first concerns the use of oxygen as a
substrate and the second the introduction of a water-immiscible organic solvent. In
principle, solids may also be present as resins for the supply of substrate or removal of
product.
Biocatalytic reactions carried out in a two-liquid phase dispersion have been used
to facilitate substrate supply or product removal from the site of enzymatic reaction
(see Section 7.5.2.2), using the second phase as a substrate or product reservoir,
respectively. In such cases the mixing of the two liquid phases is essential to promote
mass transfer, which occurs at the interface. Use of fixed-bed reactors (CPFR) usually
leads to channeling and poor phase contact. The reaction may occur at the interface
(as with lipases) or in the bulk phase. The use of immobilized enzymes is common
since the emulsification properties of the enzyme may otherwise lead to dispersions
that are very hard to separate afterwards. Immobilization normally prevents this
although evaluation on a case-by-case basis is required. Likewise (dependent upon
the density difference and viscosity difference between the two liquid phases) the
reactor is followed by a settler, which might be a clarification tank (decanter) at
the simplest level or a liquid–liquid centrifuge in more difficult cases. In all cases the
degree of mixing needs to be optimized. See Figure 7.6.
228
Figure 7.6 Two-liquid-phase stirred tank either soluble or immobilized as a solid or on a

reactors. In (a) the enzyme is either soluble or porous support in the dispersed phase. The
immobilized as a solid or on a porous support in reactor is followed by liquid–liquid phase
the continuous phase. In (b) the enzyme is separation in a decanter or a centrifuge.
It is rare to add gaseous substrates to isolated (or even immobilized) enzymes since
in general proteins do not fold correctly after they have been exposed to a gas–liquid
interface. More normal is the requirement of whole-cell biocatalysts for oxygen [102].
While oxygenases are in general unstable outside the cell environment, the use of
oxidases in isolated enzyme form is more common [61]. In these cases air is the usual
way to supply oxygen (since it is cheap). Potentially, the air can be enriched with
oxygen (if it is essential to exceed the normally scalable oxygen supply rate of
100 mmol l1 h1), although this comes at a cost. Air is normally supplied via a
sparger at the bottom of a stirred tank. The agitation breaks up the bubbles to give
the maximum surface area for mass transfer. Several correlations exist to calculate
the mass transfer rate but these are not generally accurate.
A bubble column or air-lift reactor is a reactor in which the reaction medium is
kept mixed and aerated by introduction of air into the bottom of the reactor.
This reactor type is mainly applied to facilitate the contact and/or reaction of a
liquid and a gaseous phase, but it can also serve a purpose when reactants with a high
viscosity make the use of a packed bed impractical [48]. The air serves both as a
non-abrasive mixer and in some cases as a medium for removing the water formed in
the reaction.
7.2.8
Reactor Scale-Up
The scale-up of all reactors, whether biocatalytic or otherwise, is often associated with
reduced process performance. However, provided certain considerations are taken
into account this should not be a problem. One of the major challenges for future
biocatalytic processes targeted at bulk chemicals and biofuels will be the development
of processes on a large scale and, thus far, as discussed previously, only a handful of
processes have been operated at a truly large scale (greater than 10 000 tons per year).
7.3 Downstream Processing and Product Recovery j229
Successful scale-up of a biocatalytic reactor requires a good understanding of the
interactions between the biocatalyst and the chemical and physical environment in
the reactor. The objective in reactor selection and operation is to control this
environment at all scales, such that accurate predictions can be made as scales are
changed. However, it is more difficult to control the physical environment during
scale-up. Individual reactor types will provide their own challenges as they are scaled-
up. For example, as a stirred tank is scaled the mixing time is increased. Indeed in
large reactors mixing times of 1–2 min can be expected. In fact the time constants of
all processes are increased with scale and therefore an evaluation of the effect of time-
dependent variables on performance is an important pre-requisite to scale-up. Much
can be learnt for the extensive scientific literature on fermentation (for an overview
see Reference [62]). In addition, the surface area/unit volume is also reduced such
that heat transfer via a jacket becomes limiting. Temperature control of biocatalytic
reactions at large scale is not normally a problem. However, highly exothermic
reactions require efficient cooling. This can be achieved with a cooling jacket on
small-scale reactors but usually requires, additionally, the insertion of cooling coils at
a larger scale. Likewise, for packed-bed reactors this will also be a limitation. In the
past, the bed height employed used has been limited by the pressure drop across the
bed. However, improvements in particle properties are now allowing faster flow rates
to be used such that they can be deployed as differential reactors with rapid recycle. To
avoid the problem of pressure drop across the bed, a fluidized bed may be used as an
alternative. However, until recently these have not been operated successfully on a
large scale with immobilized biocatalysts. The major problem has been the difficulty
in achieving constant linear liquid velocities across the bed, which is essential to
maintain plug flow. In some cases, companies are now accepting the loss in efficiency
in catalyst utilization caused by deviations from plug-flow in large fluidized beds,
because of other advantages. Bubble reactors have been built to around 500 m3 and
stirred tanks to around 200 m3.
7.3
Downstream Processing and Product Recovery
7.3.1
Downstream Schemes
Recovery of the product from a biocatalytic reaction is of critical importance for the
development of a suitable process and is often overlooked. Clearly, the extent of
purification required depends on the subsequent use, but in most cases the removal
of by-products and unreacted substrate (for recycle) is more than desirable, it is a
necessity for an effective economic process. To reduce the cost of these separation
steps it is necessary to concentrate the product-rich stream. This means that the
biocatalyst should be separated prior to concentration. The concentration step is
more normally carried out by either evaporation of water (may be expensive) or
extraction into a lower boiling solvent, which can then be evaporated. If at all possible
230
the use of solvents should be avoided, in line with the development of a green and
sustainable process, and therefore the product should already be removed from the
reactor at a high enough concentration. This places particular demands upon the
reaction. Enzymes usually operate effectively at low concentrations of substrate and
therefore product. Interestingly, at higher concentrations of product, enzymes
frequently show product inhibition where the rate of reaction is lowered. For this
reason it can be beneficial to remove the product during the course of the reaction –
so-called in situ product removal (ISPR). For reactions with two substrates it is
inevitable that one will be in excess and therefore a separation of product from the
excess substrate will be required. The excess substrate can then be recycled. In
considering the choice of which substrate will be in excess, the ease of separation and
recycle should be considered. Since, in general, enzymes operate under mild
conditions the separation process may also need to be under mild conditions, such
that the process is in line with the development of a green process but also
critically because the highly functionalized and complex molecules that are frequent-
ly the product of an enzyme-catalyzed reaction are usually unstable under
extreme conditions. For this reason distillation under reduced pressure, solvent
extraction, chromatography, or crystallization are the most common downstream
operations.
7.3.2
Biocatalyst Recovery
Biocatalysts can be recovered by filtration. The type of filtration chosen will be

dependent upon whether the enzyme is in soluble form, as a cross-linked enzyme
aggregate (CLEA), or immobilized on a support. For pharmaceutical operations in
particular it is very important to ensure no protein leaves the biocatalytic step and so
an extra filtration is often required.
While an isolated enzyme-catalyzed reaction is easier to implement than a whole-
cell catalyzed reaction due to simplicity, the trade-off is higher upstream cost and
therefore reuse of the enzyme is often necessary (potentially via immobilization) to
keep costs down. As a rule-of-thumb the crudest possible form of the enzyme
acceptable, to maintain product quality, should be used [9].
Immobilization is often the key to improving operational performance of an
enzyme [63]. Especially for use in a dry media, such as an organic solvent, it is
difficult to use biocatalysts without immobilization or crosslinking since the
enzyme molecules are prone to aggregation. Other potential benefits are enhanced
stability, the possibility of use in a packed-bed or fluidized-bed, and prevention of
protein contamination in the product stream [64]. Fast and easy separation of the
biocatalyst from the reaction medium is sometimes a key factor for enzymatic
resolution reactions where the reaction has to be stopped at a certain conversion to
preserve the high enantiomeric excess (e.e.) of the product. Simpler downstream
processing and easy biocatalyst recovery for immobilized enzymes also lead to an
improved process economy, which may be essential for developing a competitive
process.
7.4 Process Operation j231
7.4
Process Operation
7.4.1
Control of Operating Parameters
Biocatalytic processes may require control of oxygen supply (for oxidations), tem-
perature, substrate and product concentrations, as well as pH within critical limits.
Enzymes are sensitive to all these parameters and, therefore, studies should be made
of the effect of these parameters on reaction rate (expressed via enzyme activity or the
reaction rate at a given substrate and product concentration), enzyme stability, and
thermodynamics. The temperature is controlled via a conventional jacketed reactor
vessel. Few enzymatic reactions are exothermic (an exception being phenol poly-
merization). Substrate and product concentration are controlled by substrate feeding
strategy (SFS) (either in the same phase or from an auxiliary phase) and via in situ
product removal (ISPR), respectively (these topics are discussed in Section 7.5.2).
Many biocatalytic reactions are associated with a pH change (acid production/
consumption or alkali production/consumption or combinations thereof). Since the
biocatalyst and often also the substrates and products need to be controlled in a tight
window of pH, then the control of pH in the reactor is critical. As discussed in
Section 7.2.4, on a laboratory scale buffers can be used, but this is not possible at
larger scales due to the need to use higher concentration of reagents and the
inevitable extra cost. Hence, at scale, acid or alkali will be titrated into the reactor
to neutralize the pH change that occurs. Surprisingly perhaps, many plants lack the
equipment for such metered addition of acid or alkali. The concentration of titrant
used depends upon the mixing and resultant dilution, but to ensure control good
mixing is required to avoid hot spots of high or low pH. Poor mixing may cause
problems unless some kind of distribution system is used instead of the normal
single point addition. For example, in the synthesis of lactobionic acid from lactose
via oxidation using a carbohydrate oxidase (from Microdochium nivale) the pH drops
through the reaction and in this case alkali (NaOH) was titrated into the reactor [65].
The operational stability was found to be heavily dependent on the strength of alkali
used for titration. The effect of mixing was also found to be very important and the
process was ultimately scaled-up using a rotary jet head system [66].
7.4.2
Reaction Control
As a consequence of enzyme deactivation, conversion may drop during the contin-

uous operation of enzyme reactors or over repeated batches. To maintain a constant
degree of conversion, two methods can be employed:
. addition of fresh enzyme to the reactor,
. increased reactor residence time (via a reduction of flow rate for continuous
processes or increase in reaction time for (repeated) batch processes).
232
Both can be carried out very effectively by using on-line analytical methods
combined with an appropriate feedback using a control algorithm. Useful methods
for on-line analysis of enzymatic processes are:
. polarimetry (useful for reactions where chiral reactants are involved),
. UV spectrometry,
. on-line HPLC (may be used effectively for controlling complex reactions (e.g., in
peptide or carbohydrate synthesis)).
Addition of fresh enzyme to the reactor may in some cases require replacement
or bleeding of used enzyme such that the concentration of catalyst in the reactor
does not exceed operational limits. This is particularly important in agitated
reactors (operating as BSTR or CSTR) with immobilized enzymes. For fixed-bed
reactors (operating as a CPFR) spent beds (which no longer have adequate
enzyme activity) can be taken out of line and replaced with fresh beds at a time
dependent on the economics of the process. Such a strategy is used effectively in
the large-scale glucose isomerase process for the synthesis of high fructose corn
syrup (HFCS).
7.5
Process Intensification
7.5.1
Process Metrics Required for an Effective Process
Clearly, for many reactions it is essential to intensify a reaction or process in order that it
can be applied in an industrial context. This is not surprising given that the conditions
required in a biocatalytic reactor are very different to those in nature.
The space–time yield (often referred to as STY) gives an indication of the size of
reactors required for a given productivity. For enzyme based processes values around
100 kg1 m3 day1 are required. However, by addition of more enzyme it is possible
to increase the STY. Nevertheless this comes at the expense of increased enzyme cost
and for that reason an even more useful parameter is the catalyst yield (Yb) (kg-
product per kg-enzyme). Together with the value of the enzyme this will set the cost
contribution of the enzyme to the process [67]. The exact figure required is therefore
somewhat dependent upon the economic value of the product. For low-value
products a Yb of 1000–5000 kg-product per kg-immobilized-enzyme would be
required. For higher value pharmaceutical products around 100 kg-product per
kg-immobilized-enzyme is adequate. Such a metric is not often reported in the
scientific literature and it requires effort to develop fed-batch, continuous, or recycle
based processes, all of which take significant experimental work. The final metric of
importance is the product concentration leaving the reactor. This sets the cost of
downstream processing in as much as the cost of concentration operations will be a
significant part. Again the exact figure required to achieve a commercial operation
depends upon the value of the product. For pharmaceutical products around
7.5 Process Intensification j233
Table 7.2 Reactor selection criteria based flexibility and possibility for intensification.
Criteria BSTR CSTR CPFR
Use of IE Limited Limited Possible

pH control/additions Possible Possible Possible,
requires recycle
Use of two-liquid phase Possible Possible Limited
biocatalysis (TLPB)
Flexibility High Limited Limited
20–100 g l1 is required and for lower value products around 300–400 g l1 is
necessary. Achieving these metrics requires process intensification, prior to scale-
up, and indeed this has been the subject of considerable bioprocess engineering
efforts in the last 15–20 years. Substrates will most normally be added in batch mode
at the start of the reaction. However, in cases where the solubility is limited,
substrates can be used above saturation concentration in the form of a slurry reactor
or two-liquid phase reactor. In other cases, if the substrates are toxic or inhibitory it
can be essential to feed the substrate to the reactor, which can be done directly (if the
solubility in the reaction medium is high enough) or via another phase (organic
solvent or resin). Table 7.2 gives some of the implications of intensification for reactor
selection.
The approach discussed here can be complemented by protein and enzyme
engineering efforts. In many cases this is an essential requirement alongside the
process engineering. Enormous developments have taken place in enzyme engi-
neering in recent years [68–71]. Table 7.3 outlines some of the different process
intensification techniques.
Table 7.3 Process intensification techniques.a).
Techniques Process intensification methodology
SFS ISPR EI
TLPB In reactor In reactor Frequently required

or external loop to avoid emulsification
APR In reactor In reactor Need to separate enzyme
or external loop or external loop from APR beads
Feeding Requires sufficient Not applicable Frequently required
solubility to avoid hot spots
a) SFS – substrate feeding strategy, ISPR – in situ product removal, EI – enzyme immobilization,
TLPB – two-liquid phase biocatalysis, APR auxiliary phase resins.
234
7.5.2
Intensification Methods
7.5.2.1 Enzyme Immobilization

Enzyme immobilization is a particularly important intensification method for
obtaining adequate productivity based on the biocatalyst. Improvements in
activity, stability, and selectivity are potentially possible [72], although improved
stability and separation are the usual arguments for immobilization. A general
method for immobilization that can be applied to any enzyme is not available
and, therefore, the typical approach used is trial and error. Efficient immobili-
zation protocols should take into account the physicochemical properties of the
carrier (or matrix) as well as the enzyme, to obtain the best compromise between
stability, activity, handling, and cost (see References [73] and [74] for compre-
hensive reviews). For large-scale production of the biocatalyst the procedures
should be quick, robust, scalable, and reproducible and the enzyme should be
stable during each step. In addition, working environment issues such as
the handling of crosslinking chemicals and dust-producing materials should be
considered [75].
Some examples of industrially used immobilized enzymes used at the multi-tonne
scale are glucose isomerase (on an inorganic carrier) for production of high fructose
corn syrup, penicillin G acylase (covalently attached to polyacrylate) for the produc-
tion of semi-synthetic penicillins, lactase (on an ion-exchange resin) for producing
low-lactose milk, TL lipase (on silica) for fat modification, and lipase B from Candida
antarctica (NZ435) (on polyacrylate) for use in resolutions, for example, in the
manufacture of pharmaceutical intermediates.
By selecting the appropriate technique, savings can be made in terms of the added
cost of the biocatalyst to the process. A key factor is the selection of an appropriate
enzyme loading on the resin [76]. Optimal cost effectiveness will depend on the cost
of both enzyme and immobilization matrix and the proportions in which they are
mixed. To make 1 kg of immobilized enzyme a loading of about 5–10% of protein is
normally used. Carrier prices vary between 20 and 200 D kg1 and the cost of the
enzyme lies in the range 300–2000 D kg1 for standard enzymes such as lipases and
proteases (special Enzymes can cost 10–50 kD kg1). Because the immobilization
process takes up to a few days to finish, a final pricing is between 200 and
1500 D kg1.
It is important to keep in mind that the cost of an immobilized enzyme does not say
much on its own. For an effective industrial process a productivity of around 10 tons
of product per kg of immobilized enzyme is required. This normally requires
considerable stability. For example, immobilized glucose isomerase used to produce
high fructose corn syrup has an operating life time of about one year. Depending on
the number of recycles of the enzyme (up to 200 is usually required) the cost
contribution to the produced product varies between a few hundred Euro per kg
(for pharmaceuticals) down to a few cents per kg (for bulk chemicals), but is most
normally in the range 10–0.1 D kg1 [67, 77].
7.5.2.2 Use of Organic Solvents
Poor water-solubility of substrates can be ameliorated via the addition of a
water-miscible organic solvent. However, the solubility increases are limited and
the water-miscible solvents are often harmful to the enzymes. In other cases
two-liquid phase biocatalysis (TLPB) may be used to enhance solubility.
In addition, an established way to run some batch processes with biocatalysts
is to avoid the toxic and inhibitory concentrations by supplying the substrate via a
second liquid phase (TLPB). In this case, the solvent acts as a reservoir. Such a
scheme is outlined in Figure 7.6. Order of magnitude improvements in
product concentrations can be achieved. In general, conventional equipment
can be used although the extent of mixing and subsequent phase separation is
critical. Extra parameters to be evaluated are the phase ratio of the two phases,
mass transfer, and the partition of the substrate and product between the
phases. It still remains difficult to select the organic solvent and this is the
subject of ongoing research, although much can be learnt for whole-cell
biotransformation studies [78]. Several processes have been scaled-up using this
technology [79].
7.5.2.3 Use of Resins

An alternative to organic solvents is to use a resin to supply the substrate or
remove the product [103]. Such an approach has several advantages since it avoids
the difficulties of organic solvent selection (both from an environmental perspec-
tive and the problems frequently faced by enzyme contact with organic solvents).
Similar parameters need to be investigated to organic solvent TLPB methods, but it
is clear that the capacity of the resin is particularly important. Indeed, the option of a
recycle stream containing the resin to either supply substrate or remove product
looks particularly attractive. Few processes have been scaled-up using this
technology [80].
7.5.2.4 In Situ Product Removal

In process development it is essential to examine not only the enzymatic reaction
but also the preceding and following reaction steps as well as the necessary
separation steps. Changes to pH, temperature, concentration, and solvent should
be minimized. For many enzymatic processes this is a challenge since they are
carried out under benign and mild conditions, frequently different to non-
enzymatic reactions. Hence the integration of the enzymatic step into the
complete process needs to be considered at an early stage of route selection
and flow-sheet design. Process integration is also required to consider hybrid
operations where reaction and separation take place simultaneously. This is
referred to as in situ product removal (ISPR) [81]. Internal and external ISPR
are possible using methods other than just TLPB and resins. Much can be learnt
from the extensive scientific literature on ISPR for whole-cell biotransforma-
tions [82]. Order of magnitude improvements have been found. Several processes
have been scaled up using this technology [80].
236
7.6
Process Intensification
7.6.1
Introduction
In comparison with the development of a conventional chemical process, bioprocess

design problems are frequently challenging due to the lack of precedent and, to some
extent, the additional complexity arising from intra-process interactions. Conse-
quently, there is a particular need for a systematic design framework that can handle a
large range of different design problems and guide the engineer through the whole
bioprocess development. Although significant progress has been made, such a
framework has yet to be developed; here some useful engineering tools are suggested
to assist in bioprocess development so as to facilitate knowledge-based decision-
making.
In the various sectors of the chemical industry there are different process
objectives. For pharmaceutical processes, rapid process development is required to
enable effective implementation [9]. For fine chemicals, developing a process that
ensures a high yield is crucial [83]. Alternative technologies, such as process
intensification techniques and process integration, may be applied to increase the
yield, reduce the number of process steps, and finally reduce the process cost. In
biorefining, numerous possible process configurations and feedstock and product
combinations result in a highly complex process synthesis problem [84]. At the very
early stage of process development, a lot of potential catalysts, starting material, or
technologies result in a large number of possible process configurations. The
challenge is thus to use the limited information at hand to eliminate the least
promising configurations and provide a focus for the investment of research and
development effort. Table 7.4 outlines some of the potential tools that can be used to
deal with this task.
7.6.2
Process Simulation
Since carrying out experiments is expensive and time consuming, especially when
scaling up, process modeling and simulations can allow an efficient evaluation of
various process options. To model the process, some data are needed, including, in
particular, thermodynamic and other property data for the various components. As
many compounds involved in biocatalytic processes are relatively new compared to
conventional chemicals, data (including pure property and thermodynamic data) are
generally not available. One solution is to use available software models to estimate
the required parameters. Many packages are available that can estimate the pure
property data as well as thermodynamic data of a compound through its chemical
structure, for example, the ThermoData Engine from NIST [93] and ICAS developed
by Gani and coworkers [94].
Table 7.4 Engineering toolbox [36, 67].
Tool and example reference Application
Process modeling [85] Mass and energy balance; prediction; evaluation of process
options
Sensitivity analysis [86] Evaluation of sensitivity to parameters and option selection
Uncertainty analysis [87] Evaluation of model robustness
Economic evaluation [88] Evaluation of bottlenecks and feasibility from an economic
perspective
Environmental evaluation [89–92] Evaluation of environmental impact of a process
Once the required data has been generated, chemical process design software such
as ProII, Aspen, or other packages can be used to simulate the alternative process
configurations, although depending on the complexity of the process a simple
spreadsheet could be sufficient. The cost of the materials and energy for producing
the same amount of product can then be used to eliminate unattractive process
options and identify promising options for further studies. Cost analysis can also be
used to identify bottlenecks in the flow sheets [95]. Information on the cost
contribution from the different parts of the process can be used to identify the most
costly parts that need improvement.
7.6.3
Environmental Assessment Tools
Design metrics for evaluating process options should not only include profit-
ability measures but also environmental metrics, and other techno-economic
metrics [96]. One method available is life cycle assessment (LCA), a standardized
methodology [97] used for assessing the environmental impact of a product,
including the full life cycle from cradle-to-grave as well as the impact during its
use-phase. One important lesson from the LCA work to date is that it is crucial to
identify the step in a products life cycle that has the highest impact on the
environment so that efforts for improvement can be focused there and to avoid
shifting the environmental burden of one step into another. Recently, more
research has been directed towards supporting the general view of biocatalysis
being a green technology [83, 89–92]. For example, Adlercreutz and coworkers [92]
found that the main contribution to energy consumption in the enzymatic
production of wood coating was not in the actual manufacturing step but rather
in the production of the raw materials (crop cultivation). This means that the
benefit of biocatalysis does not come from a lower process temperature, but from
an improved utilization of the raw materials, which can be achieved with a high
process yield. Although the LCA methodology is straightforward in principle,
limited availability of data and decisions, for example, regarding allocation of
environmental impact between products and side-products, can make it a time
consuming task.
238
7.6.4
Operating Windows
Another tool recently studied in our laboratory and by others is operating windows,
which graphically illustrate how process constraints impact the performance of a
process [98]. Briefly, the windows of operation are found by evaluating how various
process variables, for example, catalyst concentration or stability, impact key process
metrics, for example, the reaction rate and productivity (Figure 7.7). Defining hurdle
(or threshold) values for the process metrics allows identification of process condi-
tions that fulfill these constraints [96]. In this way windows of operation may be used
to help understand and optimize biocatalytic processes. The method has been
developed and applied in chemoenzymatic process design [99], pharmaceutical
process design, and other biocatalytic processes [96].
7.6.5
Sensitivity and Uncertainty Analysis
Sensitivity and uncertainty analysis are tools that can be applied to investigate the
robustness of the process models, quantify the expected extent of variation in the
process outcome, and identify the source of variations in process performance [86].
Sensitivity analysis can help guide the development of biocatalytic processes, by
relating sources of variation to process performance and evaluating different process
scenarios. For example, Schmid and coworkers [88] have used sensitivity analysis to
determine that mass transfer rates in a two-liquid phase system have little effect on
Figure 7.7 Operating windows. Identification of operational windows is a systematic process

development approach to help in understanding and optimizing the bioprocess design.
the overall productivity in a biocatalytic oxidation process. Uncertainty analysis, on
the other hand, analyses the propagation of various sources of uncertainties for
quantifying the overall uncertainty on the modeled process output [85]. Such
information is of great value to engineers in making decisions, in terms of needing
to ensure a consistent production and evaluation of potential process improvements.
7.6.6
Parameter Estimation
The use of modeling and the introduction of mathematical tools reflect in part the
confidence now felt by engineers to describe enzymatic processes. A very important
implication of this work is that the model can be used to predict different reactor
configurations, different flow-sheets via mass and energy balance. Above all, pre-
dictions can be made to examine the performance of a reaction and flow-sheet under
conditions where no experimental data exist. It is this quality of modeling that makes
this work of particular value (i.e., saving experimental time and building confidence).
Of course such methods also require adequate experimental validation.
Kinetic measurements [100] have to be carried out to examine the dependence of
the reaction rate on the concentrations of all relevant components. As described in a
previous chapter, to measure enzyme kinetics the initial reaction rates are deter-
mined at optimal reaction conditions. The initial reaction rates are measured by
varying the concentration of only one component and keeping all other concentra-
tions (e.g., of cosubstrates and inhibitors) at a constant level. The rate of conversion
has to be smaller than 5–10%, essentially to keep all initial concentration values
constant.
The parameters of the kinetic model can be identified by fitting the kinetic data
using methods of nonlinear regression. Methods of linear regression that are often
used need a rearrangement of rate equations into a linear form (e.g., a double
reciprocal plot according to Lineweaver–Burk). This gives different weight to the data
points measured at different concentration levels. For the correct calculation of the
regression line this point must be considered, otherwise the Lineweaver–Burk
double reciprocal plot is not acceptable [101].
Initial rates are not significant in large-scale processes where high conversion of
the substrate is desired. With rising conversion, the simultaneous effects of both
substrate decrease and product increase on the reaction rate have to be described. In
the case of equilibrium reactions, the forward reaction and the back reaction have to
be described by one rate equation: they can only be treated separately under initial
rate conditions. The overall rate equation has to describe the reaction rate as a
function of all relevant components at all relevant concentration levels. A correct fit
of all initial reaction rate data gives no guarantee that the kinetic model will fit the
overall reaction data. A proper fit of the time courses of some batch reactor
experiments at different starting concentrations represents an appropriate test of
the rate equation. This implies that numerical integration of the rate equation (e.g.,
by the Runge–Kutta method), yielding a simulated time-course, has to fit the data of
the measured time-course over the whole range of conversion. A combination of the
240
Table 7.5 Cost limitations and potential solution strategies.
Cost limitation Metric Potential solutions
High substrate Minimize Tr Avoid side reactions

cost (reactor residence time)
Increase Ysa) Avoid CSTR (low X);
recycle substrate;
high [E]
High biocatalyst cost Increase Ybb Recycle biocatalyst; avoid toxic
[S], [P]; consider SF, ISPR
High DSP cost Increase [P] High [S]; consider SFS, ISPR
Increase Ysa) Avoid CSTR (low X); consider high [E]
High reactor cost Increase STY Consider high [E]; consider continuous;
avoid toxic [S], [P]; consider SFS, ISPR
a) Ys ¼ yield on substrate (kg-product per kg-substrate).

b) Yb ¼ yield on biocatalyst (kg-product per kg-biocatalyst).
Runge–Kutta method and methods of nonlinear regression allows parameter

identification from the time-course data. This technique starts with a given set of
parameters, performs the numerical integration of the rate equation, and compares
the simulated with the measured time-course. Then the parameters are changed
and the same steps are repeated until the simulation fits the measured data. This
method requires specially designed computer software, which is commercially
available (MatlabÔ, MapleÔ). Use of mathematical tools should include design of
experiments (DOE) or optimal experimental design approaches, which are fre-
quently discussed and of high interest (also in the context of quality by design as
promoted by the US FDA).
Figure 7.8 Biocatalytic reactor scale-up.

7.7 Summary and Outlook j241
Figure 7.9 Bioprocess engineering strategy.
7.7
Summary and Outlook
There is increasing interest in the use and application of biocatalytic processes at an

industrial scale, driven by the need for renewable feed-stocks and also green
chemistry requirements. Most industrial biocatalytic processes today use single-step
conversions, but future biocatalytic processes will incorporate some new complex-
ities. For example, multi-enzymatic reactors are already starting to provide solutions
to several chemical challenges. Likewise chemoenzymatic reactions will also become
242
more common and bring new challenges for process development. Future processes
will also incorporate process intensification techniques that will complicate
the process. In particular, in situ product removal will require recycle streams
(the time-constants for which are highly dependent on scale).
In comparison with conventional industrial chemistry, the use of bio-processes
and biocatalysis is a rather young technology. Although enormous progress has
been made in the implementation of new processes (especially in the pharma-
ceutical industry) no fixed methods for process design have been established to
date. This chapter has presented some of the considerations (Table 7.5) required to
scale-up a biocatalytic process and some of the engineering tools available to assist
in this procedure. The tools will have a decisive role in helping to identify bottle-
necks in the biocatalytic development process and to justify where to put effort and
resources. These analyses should be performed from project inception and
continue throughout the life-time of the project and involve environmental as
well as economic indicators to achieve a solution where resources are used
efficiently. Biocatalytic reactor scale-up should follow a two-stage approach as
illustrated in Figure 7.8. It seems likely that all processes should first be intensified
such that the required metrics are met, prior to increasing scale. Figure 7.9
illustrates the start of an algorithm for such a bioprocess engineering strategy.
With the development of more standardized protocols and technology platforms,
the development process will be simplified, making it easier to implement
new processes in a collaborative manner between chemists, biochemists, micro-
biologists, and biochemical and chemical engineers.
References
1 Centi, G. and Perathoner, S. (2008) 6 Straathof, A.J.J., Panke, S., and Schmid,
Catalysis, a driver for sustainability and A. (2002) The production of fine
societal challenges. Catal. Today, 138, chemicals by biotransformations. Curr.
69–76. Opin. Biotechnol., 13, 548–556.
2 Jaeger, K.-E. (2004) Protein technologies 7 Schmid, A., Hollmann, F., Park, J.B., and
and commercial enzymes: white is the B€uhler, B. (2002) The use of enzymes in
hype – biocatalysts on the move. Curr. the chemical industry in Europe. Curr.
Opin. Biotechnol., 15, 269–271. Opin. Biotechnol., 13, 359–366.
3 Sheldon, R.A. and van Rantwijk, F. 8 Liese, A., Seelbach, K., and Wandrey, C.
(2004) Biocatalysis for sustainable (2006) Industrial Biotransformations,
organic synthesis. Aust. J. Chem., 57, Extended Edition, Wiley-VCH Verlag
281–189. GmbH, Weinheim.
4 Woodley, J.M. (2006) Microbial 9 Pollard, D.J. and Woodley, J.M. (2007)
biocatalytic processes and their Biocatalysis for pharmaceutical
development. Adv. Appl. Microbiol., 60, intermediates: the future is now. Trends
1–15. Biotechnol., 25, 66–73.
5 Schmid, A., Dordick, J.S., Hauer, B., 10 Lilly, M.D. and Woodley, J.M. (1996)
Kiener, A., Wubbolts, M., and Witholt, B. A structured approach to design
(2001) Industrial biocatalysis today and and operation of biotransformation
tomorrow. Nature, 409, 258–268. processes. J. Ind. Microbiol., 17, 24–29.
References j243
11 Kirk, O., Borchert, T.V., and Fuglsang, 21 Zaks, A. and Klibanov, A.M. (1984)
C.C. (2002) Industrial enzyme Enzymatic catalysis in organic media at
applications. Curr. Opin. Biotechnol., 13, 100 degrees C. Science, 224, 1249–1251.
345–351. 22 Woodley, J.M. and Lilly, M.D. (1992)
12 Sch€afer, T., Borchert, T.W., Nielsen, V.S., Process engineering of two-liquid phase
Skagerlind, P., Gibson, K., Wenger, K., biocatalysis. Prog. Biotechnol., 8, 147–154.
Hatzack, F., Nilsson, L.D., Salmon, S., 23 Santacoloma, P.A., Sin, G., Gernaey, K.V.,
Pedersen, S., Heldt-Hansen, H.P., and Woodley, J.M. (2011) Multi-enzyme
Poulsen, P.B., Lund, H., Oxenbøll, K.M., catalyzed processes: next generation
Wu, G.F., Pedersen, H.H., and Xu., H. biocatalysis. Org. Process Res. Dev., 15,
(2007) Industrial enzymes. 266–274.
Adv. Biochem. Eng. Biotechnol., 24 Vennestrøm, P.N.R., Christensen, C.H.,
105, 59–131. Pedersen, S., Grunwaldt, J.-D., and
13 Chibata, I., Tosa, T., Sato, T., and Mori, T. Woodley, J.M. (2011) Next generation
(1976) Production of L- amino acids by catalysis for renewables: combining
aminoacylase adsorbed on DEAE- enzymatic with inorganic heterogeneous
Sephadex. Methods Enzymol., 44, catalysis for bulk chemical production.
746–759. ChemCatChem, 2, 249–258.
14 Sato, T. and Tosa, T. (1993) in Industrial 25 Jensen, V.J. and Rugh, S. (1987)
Application of Immobilized Biocatalysts Industrial-scale production and
(eds A. Tanaka, T. Tosa, and T. application of immobilized glucose
Kobayashi,), Marcel Dekker, Inc., isomerase. Methods Enzymol., 136,
New York, pp. 3–24. 365–70.
15 Wandrey, C. and Flaschel, E. (1979) 26 Shewale, J.G. and Sivaraman, H. (1989)
Process development and economic Penicillin acylase enzyme production its
aspects in enzyme engineering. Acylase applications in the manufacture of 6-
L-methionine system. Adv. Biochem. Eng., APA. Process Biochem., 24, 146–154.
12, 147–218. 27 Kamphuis, J., Boesten, W.H.J.,
16 Wandrey, C., Wichmann, R., Broxterman, Q.B., Hermes, H.F.M., van
Leuchtenberger, W., and Kula, M.R. Balken, J.A.M., Mejer, E.M., and
(1981) Degussa AG/GBF Schoemaker, H.E. (1990) New
US Pat. 4 304 858. developments in the chemoenzymatic
17 Chenault, H.K., Simon, E.S., and production of amino acids. Adv. Biochem.
Whitesides, G.M. (1988) Cofactor Eng. Biotechnol., 42, 134–182.
regeneration for enzyme-catalysed 28 Leuchtenberger, W., Karrenbauer, M.,
synthesis. Biotechnol. Genet. Eng. Rev., 6, and Pl€ocker, U. (1984) Scale-up of an
221–267. enzyme membrane reactor process for
18 Kula, M.R. and Wandrey, C. (1987) the manufacture of L-enantiomeric
Continuous enzymatic transformation in compounds. Ann. N. Y. Acad. Sci.,
an enzyme-membrane reactor with 434, 78.
simultaneous NADH regeneration. 29 Faber, K. (2000) Biotransformations in
Methods Enzymol., 136, 9–21. Organic Chemistry, Springer, Berlin.
19 Wichmann, R., Wandrey, C., B€ uckmann, 30 Roberts, S.M. (1992) Preparative
A., and Kula, M.R. (1981) Continuous Biotransformations, John Wiley & Sons,
enzymatic transformation in an enzyme Ltd., Chichester.
reactor with simultaneous NAD(H) 31 Collins, A.N., Sheldrake, G.N., and
regeneration. Biotechnol. Bioeng., 23, Crosby, J. (1992) Chirality in Industry,
2789–2802. John Wiley & Sons, Ltd., Chichester.
20 Martinek, K., Semenov, A.N., and 32 Gerhartz, W. (1990) Enzymes in Industry,
Berezin, I. (1981) Enzymatic synthesis in VCH, Weinheim.
biphasic aqueous-organic systems. I. 33 Schneider, M.P. (1986) Enzymes as
Chemical equilibrium shift. Biochim. Catalysts in Organic Synthesis, D. Reidel
Biophys. Acta, 658, 76–89. Publishing, Dordrecht.
244
34 Patel, R.N. (2000) Stereoselective 48 Hilterhaus, L., Thum, O., and Liese, A.
Biocatalysis, Marcel Dekker, New York. (2008) Reactor concept for lipase-
35 Griengl, H. (2000) Biocatalysis, Springer, catalyzed solvent-free conversion of
Berlin. highly viscous reactants forming two-
36 Tufvesson, P., Lima Ramos, J., Jensen, phase systems. Org. Process Res. Dev., 12,
J.S., Al-Haque, N., and Woodley, J.M. 618–625.
(2011) Process considerations for the 49 Shimada, Y., Watanabe, Y., Sugihara, A.,
asymmetric synthesis of chiral amines and Tominaga, Y. (2002) Enzymatic
using transaminases. Biotechnol. Bioeng., alcoholysis for biodiesel fuel production
108, 1479–1493. and application of the reaction to oil
37 Bornscheuer, U.T. and Pohl, M. (2001) processing. J. Mol. Catal., B: Enzym., 17,
Improved biocatalysts by directed 133–142.
evolution and rational protein design. 50 Watanabe, T., Sugiura, M., Sato, M.,
Curr. Opin. Chem. Biol., 5, 137–143. Yamada, N., and Nakanishi, K.
38 Pohl, M. (2000) Optimierung von (2005) Diacylglycerol production in
biokatalysatoren f€ur technische prozesse. packed bed bioreactor. Process Biochem.,
Chem.-Ing.-Tech., 72, 883. 40, 637–643.
39 Arnold, F. (2001) Combinatorial and 51 Nielsen, P.M., Brask, J., and Fjerbaek, L.
computational challenges for biocatalyst (2008) Enzymatic biodiesel production:
design. Nature, 409, 253. technical and economical considerations.
40 Reetz, M.T. and J€ager, K.-E. (2000) Eur. J. Lipid Sci. Technol., 110, 692–700.
Enantioselective enzymes for organic 52 Durany, O., de Mas, C., and Lopez-Santin,
synthesis created by directed evolution. J. (2005) Fed-batch production of
Chem. Eur. J., 6, 407. recombinant fuculose-1-phosphate
41 Balc~ao, V.M., Paiva, A.L., and Malcata, aldolase in E. coli. Process Biochem., 40,
F.X. (1996) Bioreactors with immobilized 707–716.
lipases: State of the art. Enzyme Microb. 53 Hills, G. (2003) Industrial use of lipases
Technol., 18, 392–416. to produce fatty acid ester. Eur. J. Lipid Sci.
42 Fernandes, J.F.A., McAlpine, M., and Technol., 105, 601–607.
Halling, P.J. (2005) Operational stability of 54 Xu, X. (2003) Engineering of enzymatic
subtilisin CLECs in organic solvents in reactions and reactors for lipid
repeated batch and in continuous modification and synthesis. Eur. J. Lipid
operation. Biochem. Eng. J., 24, 11–15. Sci. Technol., 105, 289–304.
43 Woodley, J.M. and Lilly, M.D. (1994) 55 Carleysmith, S.W. and Lilly, M.D. (1979)
Biotransformation reactor selection and Deacylation of benzylpenicillin by
operation, in Applied Biocatalysis (eds immobilized penicillin acylase in a
J.M.S. Cabral, D. Best, and J. Tramper), continuous four-stage stirred-tank
Harwood Academic, Chur, Switzerland, reactor. Biotechnol. Bioeng., 21,
pp. 371–393. 1057–1073.
44 Lilly, M.D. and Dunnill, P. (1971) 56 Amanullah, A., McFarlane, C.M., Emery,
Biochemical reactors. Process Biochem., A.N., and Nienow, A.W. (2001) Scale-
August, 29–32. down model to simulate spatial pH
45 Lilly, M.D. and Dunnill, P. (1976) variations in large-scale bioreactors.
Immobilized-enzyme reactors, in Biotechnol. Bioeng., 73, 390–399.
Methods in Enzymology (ed. K. Mosbach), 57 Nordkvist, M., Grotkjær, T., Hummer,
Academic Press, New York, vol. XLIV, pp. J.S., and Villadsen, J. (2003) Applying
717–738. rotary jet heads for mixing and mass
46 Levenspeil, O. (1998) Chemical Reaction transfer in a forced recirculation tank
Engineering, 3rd edn, John Wiley and reactor system. Chem. Eng. Sci., 58,
Sons, Inc., New York. 3877–3890.
47 Fogler, H.S. (2006) Elements of Chemical 58 Albaek, M.O., Gernaey, K.V., Hansen,
Reaction Engineering, 4th edn, Pearson M.S., and Stocks, S.M. (2011) Modelling
Education, Inc., Upper Saddle River, NJ. enzyme production with Aspergillus
References j245
oryzae in pilot scale vessels with different 68 Burton, S.G., Cowan, D.A., and Woodley,
agitation, aeration, and agitator types. J.M. (2002) The search for the ideal
Biotechnol. Bioeng., 108, 1828–1840. biocatalyst. Nature, 20, 37–45.
59 Kragl, U., G€odde, A., Wandrey, C., Kinzy, 69 Fox, R.J. and Huisman, G.W. (2008)
W., Cappon, J.J., and Lugtenburg, J. Enzyme optimization: moving from
(1993) Repetitive batch as an efficient blind evolution to statistical exploration
method for preparative scale enzymic of sequence-function space. Trends
synthesis of 5-azido-neuroaminic acid Biotechnol., 26, 132–137.
and 15N-L-glutamic acid. Tetrahedron: 70 Turner, N.J. (2009) Directed evolution
Asymmetry, 4, 1193–1202. drives the next generation of biocatalysts.
60 Salagnad, C., G€odde, A., Ernst, B., and Nat. Chem. Biol., 5, 567–573.
Kragl, U. (1997) Enzymatic large scale 71 Tracewell, C.A. and Arnold, F.H. (2009)
production of 2-keto-3-deoxy-D-glycero- Directed evolution: climbing fitness
D-galactononulopyranosonic acid (KDN) peaks one amino acid at a time. Curr.
in enzyme membrane reactors. Opin. Chem. Biol., 13, 3–9.
Biotechnol. Progr., 13, 810–813. 72 Mateo, C., Palomo, J.M., Fernandez-
61 Fernandez-Lafuente, R., Rodriguez, V., Lorente, G., Guisan, J.M., and
and Guisan, J.M. (1998) The Fernandez-Lafuente, R. (2007)
coimmobilization of D-amino acid Improvement of enzyme activity, stability
oxidase and catalase enables the and selectivity via immobilization
quantitative transformation of D-amino techniques. Enzyme Microb. Technol., 40,
acids (D-phenylalanine) into a-keto acids 1451–1463.
(phenylpyruvic acid). Enzyme Microb. 73 Hanefeld, U., Gardossi, L., and Magner,
Technol., 23, 28–33. E. (2009) Understanding enzyme
62 Nielsen, J., Villadsen, J., and Liden, G. immobilisation. Chem. Soc. Rev., 38,
(2011) Bioreaction Engineering Principles, 453–468.
3rd edn, Kluwer Academic/Plenum 74 Cao, L., van Langen, L., and Sheldon, R.
Publishing, New York. (2003) Immobilised enzymes: carrier-
63 Sheldon, R.A. (2007) Enzyme bound or carrier-free? Curr. Opin.
immobilization: the quest for optimum Biotechnol., 14, 387–394.
performance. Adv. Synth. Catal., 349, 75 Kirk, O. and Christensen, M.W.
1289–1307. (2002) Lipases from Candida
64 Bornscheuer, U.T. (2003) Immobilizing antarctica: unique biocatalysts from a
enzymes: how to create more suitable unique origin. Org. Process Res. Dev., 6,
biocatalysts. Angew Chem. Int. Ed., 43, 446–451.
3336–3337. 76 Bosley, J.A. and Peilow, A.D. (1997)
65 Nordkvist, M., Nielsen, P.M., and Immobilization of lipases on porous
Villadsen, J. (2007) Oxidation of lactose to polypropylene: reduction in esterification
lactobionic acid by a Microdochium nivale efficiency at low loading. J. Am. Oil.
carbohydrate oxidase: kinetics and Chem. Soc., 74, 107–111.
operational stability. Biotechnol. Bioeng., 77 Rozzell, D.J. (1999) Commercial scale
97, 694–707. biocatalysis: myths and realities. Bioorg.
66 Hua, L., Nordkvist, M., Nielsen, Med. Chem., 7, 2253–2261.
P.M., and Villadsen, J. (2007) Scale-up 78 Straathof, A.J.J. (2003) Auxiliary phase
of enzymatic production of lactobionic guidelines for microbial
acid using the rotary jet head biotransformations of toxic substrate into
system. Biotechnol. Bioeng., 97, toxic product. Biotechnol. Prog., 19,
842–849. 755–762.
67 Tufvesson, P., Lima Ramos, J., Nordblad, 79 Lye, G.J. and Woodley, J.M. (2001)
M., and Woodley, J.M. (2011) Guidelines Advances in the selection and design
and cost analysis for catalyst production of two-liquid phase biocatalytic
in biocatalytic processes. Org. Process Res. reactors, in Multiphase Bioreactor Design
Dev., 15, 266–274. (eds J.M.S. Cabral, M. Mota, and J.
246
Tramper), Taylor & Francis, London, pp. production of cosmetic ingredients.

115–134. €
SOFW J.: Int. J. Appl. Sci. (Engl. Ed.), 134,
80 Vicenzi, J.T., Zmijewski, M.J., Reinhard, 44–47.
M.R., Landen, B.E., Muth, W., and 90 Kim, S., Jimenez-Gonzalez, C., and Dale,
Marler, P.G. (1997) Large-scale B.E. (2009) Enzymes for pharmaceutical
stereoselective enzymatic ketone applications – a cradle-to-gate life cycle
reduction with in situ product removal assessment. Int. J. LCA, 14, 392–400.
via polymeric adsorbent resins. Enzyme 91 Henderson, R.K., Jimenez-Gonzalez, C.,
Microb. Technol., 20, 494–499. Preston, C., Constable, D.J.C., and
81 Woodley, J.M., Bisschops, M., Straathof, Woodley, J.M. (2008) Comparison of
A.J.J., and Ottens, M. (2008) Future biocatalytic and chemical synthesis: EHS
directions for in-situ product removal and LCA comparison for 7-ACA
(ISPR). J. Chem. Tech. Biotechnol., 83, synthesis. Ind. Biotechnol., 4, 180–192.
121–123. 92 Petersson, A.E.V., Gustafsson, L.M.,
82 Stark, D. and von Stockar, U. (2003) In Nordblad, M., Borjesson, P., Mattiasson,
situ product removal (ISPR) in whole cell B., and Adlercreutz, P. (2005) Wax esters
biotechnology during the past 20 years. produced by solvent-free energy-efficient
Adv. Biochem. Eng./Biotechnol., 80, enzymatic synthesis and their
49–175. applicability as wood coatings. Green
83 Hatti-Kaul, R., T€ornvall, U., Gustafsson, Chem., 7, 837–843.
L., and B€orjesson, P. (2007) Industrial 93 Diky, V., Muzny, C.D., Lemmon, E.W.,
biotechnology for the production of bio- Chirico, R.D., and Frenkel, M. (2007)
based chemicals – a cradle-to-grave ThermoData engine (TDE): software
perspective. Trends Biotechnol., 25, implementation of the dynamic data
119–124. evaluation concept. 2: Equations of state
84 Sammons, N., Eden, M., Cullinan, H., on-demand and dynamic updates over
Perine, L., and Connor, E. (2007) A the web. J. Chem. Inf. Model, 47,
flexible framework for optional 1713–1725.
biorefinery product allocation. Environ. 94 Gani, R., Hytoft, G., Jaksland, C., and
Prog., 26, 349–354. Jensen, A.K. (1997) An integrated
85 Sin, G., Woodley, J.M., and Gernaey, K.V. computer aided system for integrated
(2009) Application of modeling and design of chemical processes. Computers
simulation tools for the evaluation of Chem. Eng., 21, 1135–1146.
biocatalytic processes: a future 95 Alvarado-Morales, M., Terra, J., Gernaey,
perspective. Biotechnol. Prog., 25, K.V., Woodley, J.M., and Gani, R. (2008)
1529–1538. Biorefining: computer aided tools for
86 Saltelli, A., Ratto, M., Tarantola, S., and sustainable design and analysis of
Campolongo, F. (2005) Sensitivity bioethanol production. Chem. Eng. Res.
analysis for chemical models. Chem. Rev., Des., 87, 1171–1183.
105, 2811–2827. 96 Law, H.E.M., Lewis, D.J., McRobbie, I.,
87 Helton, JC. and Davis, FJ. (2003) Latin and Woodley, J. (2008) Model
hypercure sampling and the visualization for evaluation of biocatalytic
propagation of uncertainty in analyses of processes. Food Bioprod. Process, 86,
complex systems. Eng. Syst. Safety, 81, 96–103.
23–69. 97 ISO (1998) ISO 14040-43. Life Cycle
88 B€uhler, B., Straathof, A.J., Witholt, B., Assessment. International Organization
and Schmid, A. (2006) Analysis of two- for Standardization, Geneva.
liquid-phase multistep biooxidation 98 Woodley, J.M. and Titchener-Hooker, N.J.
based on a process model: indications for (1996) The use of windows of operation
biological energy shortage. Org. Process as a bioprocess design tool. Bioprocess
Res. Dev., 10, 628–643. Eng., 14, 263–268.
89 Thum, O. and Oxenbøl, K.M. (2008) 99 Blayer, S., Woodley, J.M., and Lilly, M.D.
Biocatalysis: a sustainable process for (1996) Characterization of the
References j247
chemoenzymatic synthesis of N-acetyl-d- 102 Law, H.E.M., Baldwin, C.V.F., Chen,
neuraminic acid (Neu5Ac). Biotechnol. B.H., and Woodley, J. (2006) Process
Prog., 12, 758–763. limitations in a whole-cell catalysed
100 Eisenthal, R. and Danson, M.J. (1992) oxidation: sensitivity analysis. Chem. Eng.
Enzyme Assays, A Practical Approach, IRL Sci., 61, 6646–6652.
Press, Oxford. 103 Kim, P.-Y., Pollard, D.J., and Woodley,
101 Henderson, P.J.F. (1992) in Enzyme J.M. (2001) Substrate supply for
Assays, A Practical Approach (eds R. effective biocatalysis. Biotechnol. Prog., 23,
Eisenthal and M.J. Danson), IRL Press, 74–82.
Oxford.
j249
Part II
Hydrolysis and Formation of CO Bonds
and Oliver May.
j251
8
Hydrolysis and Formation of Carboxylic Acid Esters
Monica Paravidino, Philipp B€ohm, Harald Gr€oger, and Ulf Hanefeld
8.1
Introduction
Esters are of central importance in nature. Nature, consequently, has developed the
catalytic machinery to synthesize and hydrolyze the ester bond. Typically this is not
carried out by the same enzymes. While the synthesis of esters is part of the anabolic
pathway [1] and commonly starts from thioesters, the activated acids of nature [2],
ester hydrolysis is often part of the catabolism, that is, the digestion of compounds.
The enzymes responsible for this, esterases and the subfamily of esterases known as
lipases, are proteins with one single activity – the hydrolysis of esters [3]. They are
selective for esters as functional groups, but they are unselective concerning
the general structure of the substrate. This makes them ideal catalysts in the hands
of the organic chemists, because they can be used either the way nature intended, to
hydrolyze, or in reverse mode: to synthesize an ester [3–6].
8.1.1
How Do Esterases (Lipases) Work?
The hydrolases evolved by nature for the hydrolysis of esters are versatile enzymes
called esterases. The subfamily of enzymes that nature specially evolved for the
hydrolysis of fats and oils are known as lipases. Esterases and the lipase subfamily
share a common mechanism but their structure varies slightly since they are adapted
to the environment they have to work in. As catalysts they lower the energy barrier of a
reaction but do not influence the equilibrium of the reaction. Consequently,
esterases, including the lipase subfamily, should also catalyze the formation of
esters (Scheme 8.1). As early as 1900 it was demonstrated that this is indeed the case.
In a landmark experiment Kastle and Loevenhart demonstrated that the action of
these enzymes is indeed reversible [7]. Utilizing porcine pancreas lipase (PPL) as
catalyst they could show that butyric acid and ethanol did react to give ethyl butanoate
(Scheme 8.2). The distinct smell of the ester easily allowed its identification. It was
and Oliver May.
j 8 Hydrolysis and Formation of Carboxylic Acid Esters
252
O O
esterase
2
R 1
OR 2 + H 2O R 1
OH + R OH
(lipase)
Scheme 8.1 Esterases (lipases) catalyze the hydrolysis and synthesis of esters.
O O
PPL
OH + HO O + H2 O
Scheme 8.2 The first lipase catalyzed ester synthesis was described in 1900 [7].
distilled from the reaction mixture and hydrolyzed chemically to again yield butyric
acid.
Chemically, the synthesis and hydrolysis of esters and amides is related. The main
difference between these classes of compounds is the higher stability of amides. As
might be expected from a chemical point of view, the enzymes that were evolved to
hydrolyze amides can often also hydrolyze esters [3, 8–10]. The hydrolysis of amides
by esterases, however, is the exception [11]. This seems to be due to the different
binding of the substrates. Amides offer the possibility of forming an additional
hydrogen bond via NH. Bacillus subtilis esterase BS2, which is known to hydrolyze
amides, was shown to form the hydrogen bond with the amide nitrogen. It lost its
catalytic activity towards amides when a point mutation was made, which led to a loss
of this hydrogen bond [12].
While esterases often cannot hydrolyze amides, amidases and proteases can
frequently hydrolyze esters. This means that, next to the esterases/lipases, all the
amide hydrolyzing enzymes are potentially available for the synthesis and hydrolysis
of esters, too [3, 8, 13]. Interestingly, esterases and proteases such as subtilisin even
have, in essence, the same mechanism. Delocalization is the key to generation of the
alcoholate ion of these serine hydrolases (Scheme 8.3).
His His
O O
Glu O Ser Glu O Ser
H N N H N N H
O OH
O
Glu
O
Scheme 8.3 Charge delocalization in the catalytic triad is the key to serine hydrolase activity.
Esterases (lipases) are cofactor and metal-free catalysts. They can be used as
supplied and no special care needs to be taken. The hydrolysis of esters is based on a
catalytic triad and the oxyanion hole (Scheme 8.4). Key to catalytic success is the
charge delocalization achieved by these two features of the enzyme. The catalytic triad
O O
hydrolase
1 2 + H 2O 1 + R 2 OH
R OR R OH
catalytic triad
His
oxyanion
N H hole
N
H O
O
Ser
O
Glu
-R 1COOH +R 1COOR 2
+R1 COOH -R 1 COOR2
H 1
R2 1
His His
OR OR
N H O oxyanion N H O
N hole N
H O H O
O O
Ser Ser
O O oxyanion
Glu Glu
hole
2
-H2 O +R OH
+H 2O -R 2OH
His
R1
N O oxyanion
N hole
H O
O
Ser
O
Glu
Scheme 8.4 Catalytic cycle of esterases (lipases) proceeds via a covalently bound acid moiety.
enables deprotonation of the serine hydroxyl group at neutral and even acidic pH
values rather than pH 14–15, the pKa of alcohols. Subsequently, the serine alcoholate
attacks the ester, generating a tetrahedral intermediate. The charge density on the
oxyanion of this intermediate is stabilized by delocalization in the oxyanion hole, thus
allowing the occurrence of this charged species under neutral reaction conditions.
Elimination of the alcohol then generates the acyl enzyme complex. Here the acid
moiety is covalently bound to the enzyme [1, 3]. Attack of water, activated by the
imidazole, then yields the second tetrahedral intermediate. This again is stabilized
via the oxyanion hole. Expulsion of the acid regenerates the enzyme for the next
catalytic cycle. As mentioned above, the catalytic cycle can also proceed the other way
round, generating esters from alcohols and acids. Consequently, it is also possible to
substitute one alcohol moiety of an ester against another or to replace one acid group
by another [13].
In addition to the mechanistic considerations there are physicochemical proper-
ties to consider. Esterases in general are the enzymes that can hydrolyze all kinds of
esters. The subfamily that was evolved to hydrolyze especially lipids is called lipases
254
and in many cases the literature discusses them as two types of enzymes. Mech-
anistically lipases and the other esterases do not differ. The physicochemical
properties do, however, differ very distinctively [3, 14, 15]. As early as 1930 it was
demonstrated by Sym that PPL is active at the interface of the aqueous and the organic
layer of a biphasic reaction mixture [16]. A few years later it was shown that lipases are
indeed activated in biphasic mixtures, while they are distinctly less active in mono-
phasic reactions [17, 18]. The second layer can be an organic solvent or simply the
substrate [3, 19, 20]. This interfacial activation is a key feature for identifying these
esterases as lipases. From here onwards esterase will be used for all those esterases
that do not show this interfacial activation and lipases will be used for enzymes
that do.
What are the crucial differences between lipases and other esterases? Lipases have
a hydrophilic surface and the active site is in many cases covered by a lid. This lid has a
lipophilic side, covering the lipophilic active site [14]. Once the lipase comes into
contact with a lipophilic surface the lid opens, revealing a large lipophilic area that
includes the active site (Figures 8.1 and 8.2). This is oriented towards the organic
layer; thus the lipid can easily enter the active site and the more hydrophilic products
of the hydrolysis reaction will quickly diffuse away, into the aqueous layer [14, 15].
Some lipases – such as for instance the very popular Candida antarctica lipase B
(CALB, more recently also called Pseudozyma antarctica lipase B: PALB) – do not have
lipase lipase
lid closed lid open, active site accessible
hydrophilic surface
hydrophobic surface
hydrophobic solvent (fat or oil)
lipase
aqueous layer
Figure 8.1 Many lipases show interfacial activation. This is ascribed to the lid that covers the active
site when the lipase is not catalytically active. This lid needs to open to give access to the active site.
Figure 8.2 The lipase from Thermomyces right-hand side the lipophilic surface is
lanuginosus (TLL, formerly known as Humicola indicated in green. When the lid is open the
lanuginosa lipase) both in open-active lipophilic active site is accessible. Reproduced
conformation (a) and closed-inactive by permission of the Royal Society of Chemistry
conformation (b). On the left-hand side the from Reference [44].
active site amino acids are highlighted, on the
a lid and, therefore, no interfacial activation as described above can occur [21].
Nonetheless, the polarity of the surface of CALB evolved in such a way that the active
site will always be oriented towards the lipophilic phase of a biphasic mixture
(Figure 8.3) [22].
The question of how the water gets into the active site remains unanswered.
Diffusion through the organic layer is very unlikely. Indeed, it was recently dem-
onstrated that lipases have tunnels that allow water molecules to diffuse from the
aqueous layer through the protein into the active site [23].
8.1.2
Ester Synthesis versus Ester Hydrolysis
The equilibrium constant for the ester hydrolysis and formation (Scheme 8.1) is
dependent on the substrate, but in general complete conversions can only be
achieved with an excess of one of the reagents. In the hydrolysis reaction this is
256
Figure 8.3 Not all lipases have a lid. The most prominent lipase, CALB, does not have one.
However, it does have a lipophilic and a hydrophilic surface area and the active site is in the lipophilic
area. Reproduced with permission from Reference [22]. Copyright 2007, Wiley-VCH Verlag GmbH.
readily achieved by performing the reaction in aqueous media. To avoid acidification

of the reaction medium due to acid formation high buffer concentrations can be
used. It is, however, more recommendable to work with an automatic burette (pH
stat) that continuously neutralizes the acid formed. This also allows online moni-
toring of the reaction.
Since most substrates do not dissolve well in water a biphasic approach is favorable
in hydrolysis reactions. This is recommended in particular for lipases since most
lipases are interfacially activated and display significantly higher activity in a two-
phase mixture. Again, a pH stat should be utilized to monitor the reaction.
For the synthesis of esters one could now assume that replacing water by an alcohol
might solve all equilibrium problems. However, most esterases and lipases are not
very stable in polar and water-miscible solvents [3]. Furthermore, studies on the
esterification of butyric acid with butanol (ratio 1 : 2) in hexane in the presence of salt
pairs (a salt and its hydrated form, to control the water activity) can lead to very high
yields but are relatively slow [24]. Overall, it has proven to be superior to use activated
acids in combination with an alcohol in water-immiscible solvents.
8.1.2.1 Ester Synthesis – Reactions in Organic Solvents

The first enzyme-catalyzed reactions in organic solvents were described in the first half
of the last century [25]. In particular, Sym investigated the PPL-catalyzed reaction of
alcohols and acids in organic solvents in detail [16, 26–31]. The application of enzymes
inorganic solvents was rediscoveredsome 30 years ago,pioneered by Klibanov [32–34],
and since then much work has been done [3, 5, 13]. A significant advantage of organic
solvents is the much better solubility of the substrates and products, enabling a high
productivity; in the context of this chapter it is the key to ester synthesis. To date,
hydrolases and in particular lipases are the only enzymes that can be applied in dry
solvents, which makes it possible to obtain very high yields in lipase-catalyzed
esterification reactions [4–6]. Here, dry refers to commercially available dry solvents.
All other enzymes need at least water-saturated solvents [35–38]. The organic solvents
have been further expanded by ionic liquids (IL) as solvents for enzymes. Excellent
reviews and books on both topics can be recommended [3, 39–43].
Enzymes are not soluble in organic solvents. In principle they are heterogeneous
when utilized under these conditions. However, this also means that enzyme
molecules might stick together, reducing the surface area and causing diffusion
limitations. To improve accessibility of the catalytic centers it is better to employ
immobilized enzymes [44]. In many cases the immobilisates also display an
improved stability under reaction conditions. Indeed, many commercial enzymes
are immobilized already, easing their application. Detailed reviews and books on all
aspects of enzyme immobilization are available [44–47].
The core problem of ester synthesis in organic solvents is the equilibrium
(Schemes 8.1 and 8.5). Removal of water from the reaction system is possible, for
instance via salt pairs to control the water activity, as has been described in early work
from Halling and Anthonsen [24]. However, many hydrolases function less well at
very low water activities [42, 48]. Therefore, the problem is commonly addressed by
altering the reaction by employing an ester rather than an acid as starting material
(Scheme 8.5) [3–6]. Strictly speaking this is no longer an ester synthesis but a
transesterification [13]. When starting with a thioester the resulting thiol can easily be
evaporated [49], but due to the accompanying stench this is not widely applied.
Another approach is to utilize esters that release non-nucleophilic alcohols, thus
ensuring that the reaction is almost irreversible. In particular, trifluoroethanol esters
are popular in this class of activated esters [8, 13]. Trichloroethanol-, oxime-, and
cyanohydrin-esters are rarely used, the latter due to the toxicity of the HCN released
from the leaving group [4–6].
O O
enzyme
R1 OH + HO R2 R1 OR2 + H2O
O O
enzyme
R1 SR2 + HO R3 R1 OR3 + HS R2
NO2
O O
enzyme
R1 O + HO R2 R1 OR2 + O NO2
O O
1 2 enzyme 1
R X + HO R R OR2 + H X
O
X = Cl3C O F3C O N C O
N
Scheme 8.5 Application of activated esters and esters that release non-nucleophilic alcohols helps
overcome the equilibrium problem in ester synthesis.
258
A truly irreversible approach is the application of acid anhydrides introduced in

1988 by the group of Cesti (Scheme 8.6) [50]. The resulting ester is obtained in a
mixture with the acid, or if a cyclic anhydride is used the product is an acid. In this
case the work-up is greatly eased, since a straightforward extraction is sufficient to
obtain the pure product [51]. A drawback of this approach is that one equivalent of
acid is released, which might influence the enzyme activity negatively. The most
versatile and most often applied approach is the use of enol esters [4–6, 52]. These
esters are activated and the side product, the enol of an aldehyde, ketone, or ester,
immediately tautomerizes into its stable form. Thus no nucleophile is existent. In
particular, vinyl acetate and isopropenyl acetate, introduced in 1987 by Maillard [53,
54], are widely employed since both compounds are readily available on a large scale.
Most other enol esters have to be synthesized, which in practice limited this approach
to the commercially available esters.
O O
1 2
enzyme 1
R OH + HO R R OR2 + H 2O
O O
O O + enzyme
HO R HO R
O
O
O R2 O R2 R2
+ HO R 3 enzyme +
R1 O R1 OR3 HO O
R2 = H, CH3 , OEt
Scheme 8.6 Irreversible ester synthesis with acid anhydrides or enol esters.
Careful investigation of ester synthesis with reagents that should lead to an

irreversible ester formation revealed that even under dry conditions and when
applying enol esters as starting materials some hydrolysis takes place [55, 56].
Obviously, lipases can use even tiny quantities of water, such as that bound to the
enzyme or the carrier, for hydrolysis of the reactive esters and also the product.
Care therefore has to be taken when ester syntheses are performed to stop the
reaction on time and to ensure really dry conditions, including possibly drying the
lipase.
8.1.3
Stereochemistry
Hydrolases are extremely stereoselective [3, 57]. It is this stereoselectivity combined

with the low substrate specificity that has made hydrolases such popular tools in
organic chemistry. Hydrolases are mainly utilized to prepare enantiopure com-
pounds from racemic or prochiral materials. In many different ways they have
replaced classical resolutions by crystallization.
O
H OH H O R
O
+ Lipases
L M HO R L M
Lipases: react
Subtilisin: does not react
O
HO H R O OH
O
+ Subtilisin
L M HO R L M
Lipases: do not react

Subtilisin: reacts
Figure 8.4 The rule of Kazlauskas describes with high reliability the enantioselectivity of lipases for
secondary alcohols. Subtilisin tends to display the opposite enantioselectivity. L: large substituent;
M: medium-sized substituent; hydrogen is the small substituent.
The high enantioselectivity of hydrolases can in particular be found for secondary

alcohols, as described by Kazlauskas in the rule later named after him (Figure 8.4)
[3, 5, 58]. Lipases in general catalyze the conversion of only one of the two
enantiomers of secondary alcohols or the hydrolysis of the corresponding ester.
Subtilisin displays the opposite enantioselectivity to lipases [59, 60]. Consequently,
either enantiomer of a secondary alcohol can be converted into an ester and either
enantiomer of the corresponding ester can be hydrolyzed enantioselectively.
This stereodifferentiation is due to the stereodifferentiating site or pocket in the
enzymes [3]. This only accepts medium-sized groups and no large groups
(Figure 8.5). Conversions of secondary alcohols with two relatively large groups are
consequently slow and display low selectivity, while secondary alcohols with two
relatively small substituents are converted rapidly, again with relatively low enantios-
electivity [61, 62]. All other secondary alcohols – the vast majority – are converted with
excellent enantioselectivities.
In general the enantioselectivity of lipases and esterases for primary alcohols with a
defined stereochemistry in b-position is not very high; a few enzymes, however, are
very selective. For such primary alcohols, again Kazlauskas observed that – in
particular Burkholderia cepacia lipase (BCL, formerly known as Pseudomonas cepacia
lipase, PCL) – follows a general rule (Figure 8.6) [63]. Enantiomers with medium and
large substituents on the b-carbon are distinguished by the enzyme. Elegant
hydrolysis experiments (Scheme 8.7) combined with modeling studies revealed that
the medium and large substituent of both enantiomers dock in the same way. The
stereodifferentiation takes place due to a pocket that easily accommodates
the hydrogen atom of the chiral carbon but can only give a fit with the a-carbon
of the primary alcohol with a large loss of activity (Figure 8.6) [64].
260
oxyanion
hole
O O enzyme
H O R
O L M stereodifferentiating
site
H O R
L M
oxyanion
O hole
O O enzyme
H O R
H O R
M L
M L stereodifferentiating
site - too small
Figure 8.5 Enantioselectivity for secondary alcohols is due to the low reactivity of the enantiomer
that does not match the orientation of the stereodifferentiating site.
Chiral tertiary alcohols are potentially very interesting, since it is a major challenge
in chemistry to prepare optically pure quaternary carbon centers. Chiral tertiary
alcohols have such a carbon atom, but this also means that the difference in size
between the three substituents is normally smaller than in secondary alcohols, where
one of the substituents is a hydrogen atom. Moreover, hydrolases tend to be not very
OH OH
H H
M L L M
reacts does not react

enzyme enzyme
oxyanion oxyanion
hole hole
O O O O
R R
O O
H
H H
H H
H
M M
L L
Figure 8.6 BCL generally catalyzes the acylation of the left-hand enantiomer of the chiral primary
alcohol, in particular if no oxygen atom is bound to the chiral carbon. L: large substituent; M:
medium-sized substituent.
O
C 6H 13 O
O
BCL, phosphate buffer,
C6 H13 O pH 7, n-propanol +
HO
Scheme 8.7 Enantioselective hydrolysis of the rac-primary alcohol ester was performed to
elucidate the mechanism of stereodifferentiation.
active towards these bulky substrates and indeed not many tertiary alcohols or their
esters exist in nature. Nonetheless, the topic has attracted considerable attention in
recent years and the pioneering work of Bornscheuer has yielded a large number of
esterases that hydrolyze various tertiary alcohol esters enantioselectively [65]. Thus,
kinetic resolutions for manufacture of enantiopure quaternary carbon atoms have
become available. Indeed recent work demonstrated that an esterase could be found
for every problem within a systematic study. A general rule, however, does not
exist [66].
The stereocenter that esterases and lipases can distinguish can also be in the acid
moiety. Acids with a stereocenter in that a-position have successfully been resolved.
For Candida rugosa lipase (CRL, formerly Candida cylindracea lipase) a general rule
could be established by Kazlauskas and confirmed by Franssen (Figure 8.7) [67, 68].
Other lipases and esterases are also selective for this type of substrate. In particular,
proteases have a strong preference for L-amino acid esters, that is, acids with a
stereocenter in a-position [3].
In addition to the common cases described here much more is known and
described in detail in the excellent book by Bornscheuer and Kazlauskas [3]. It is
of great importance to know how reliable these general rules are. In general they hold
true. However, if a switch of reaction medium is performed, in particular from water
to organic solvents, then reversals of enantioselectivity have been reported, in
particular for proteases [69, 70]. But in addition to that a switch of substrate polarity
can also lead to reversals of enantioselectivity [60]. It is, therefore, recommendable to
always provide independent proof of absolute stereochemistry.
H COOH HOOC H
M M
L L
reacts does not react
Figure 8.7 CRL catalyzes the enantioselective ester synthesis of the left-hand enantiomer of chiral
acids or in reversal the hydrolysis of esters of the left-hand enantiomer.
262
8.1.4
Reaction Concepts
Hydrolases were evolved to hydrolyze. In reversal of their natural function they can
also be utilized to form the bond they normally hydrolyze. However, esterases and
lipases are commonly not used to make or hydrolyze esters. Because they are
extremely stereoselective their application in most cases replaces classical resolutions
by crystallization. For these reactions several approaches are used, which are
discussed in detail in Chapter 2 [71].
In a kinetic resolution (KR) the enantioselectivity of the hydrolase in the hydrolysis
(commonly in water or biphasic reaction mixtures) or synthesis reaction (dry organic
solvents) is employed. Hydrolases with an excellent enantioselectivity, E, will convert
50% of the racemic starting material, leaving a mixture of enantiopure ester and
alcohol (KR of rac-alcohol) or acid and ester (KR of rac-acid). Even in this ideal case the
theoretical yield is limited to 50% and recycling via racemization of the undesired
enantiomer is necessary. Essentially, this is a clean-up operation after the unselective
synthesis of the racemic material (Scheme 8.8) [3, 5].
(a) in water X = O, NH; Y = C or heteroatom

R1 R1
enzyme-catalyzed
(S) Y C Xacyl hydrolysis, fast (S) Y C XH
R2 R2
R1 R1
not catalyzed,
(R) Y C Xacyl very slow (R) Y C XH
R2 R2
(b) in dry organic solvent X = O, NH; Y = C or heteroatom

R1 R1
enzyme-catalyzed
(S) Y C XH hydrolysis, fast (S) Y C Xacyl
R2 R2
R1 R1
not catalyzed,
(R) Y C XH (R) Y C Xacyl
very slow
R2 R2
Scheme 8.8 Kinetic resolution of rac acids or alcohols can be performed either by enantioselective
hydrolysis of the corresponding esters (a) or by their enantioselective synthesis (b). Here this
concept is depicted for the rac alcohols.
The low yields of a KR can be doubled by combining the KR with the reversible and
dynamic racemization of the starting material (Scheme 8.9) [71–73]. In this manner
the hydrolase that catalyzes the enantioselective hydrolysis or synthesis is always
confronted with rac alcohol or rac ester (depending whether it is a KR of an alcohol or
ester). Numerous heterogeneous and homogeneous catalysts have been developed to
O
OH subtilisin O
R1 R2 R1 R2
O
O O O
homogeneous and R3
R1 R2 R 3
O O
heterogeneous catalysts
OH lipases O
1 2 1
R R R R2
Scheme 8.9 In a DKR a rac mixture is Kazlauskas rule for secondary alcohols
dynamically racemized. Here an alcohol is (Figure 8.4) 100% product that is 100%
racemized via a redox process. The enzyme enantiopure can be obtained for either
enantioselectively converts one enantiomer into enantiomer, depending on whether a lipase or
a stable unracemizable ester. According to the subtilisin is utilized.
catalyze racemization in the presence of hydrolases [74]. In addition, the rather

limited number of racemases has been utilized for this purpose [75]. By racemizing
the starting material the yield can now be increased to 100% with 100% enantiopurity.
The overall process is called dynamic kinetic resolution (DKR). However, notably, in
principle no new material is formed, it is just the cleaning up of a racemate.
A further development of DKR is not to start with a racemate but with a prochiral
starting material. If an aldehyde or ketone is employed in the reversible formation of a
new chiral center then the necessary racemization step is also a bond forming step
(Scheme 8.10). This can be the reduction of a ketone [76], the addition of cyanide to
form a cyanohydrin [72, 77, 78], or any other bond-forming reversible reaction in
R1 enzyme-catalyzed R1
acylation, fast
(S) Y C XH (S) Y C Xacyl
+YH R2 R2
R1
C X dynamic dynamic racemization X = O, NH
R2
R1 not catalyzed, R1
+YH very slow
(R) Y C XH (R) Y C Xacyl
R2 R2
Scheme 8.10 A synthetic DKR begins with a prochiral starting material and the reversible
formation of a new bond is combined with the irreversible enantioselective esterification of the rac
material.
264
which a chiral alcohol is formed. If this dynamic reaction is combined with the
irreversible KR then a DKR results that is the specific synthesis of a new chiral center,
and the reaction can be utilized as a step in a synthesis sequence rather than an
undesired additional step to solve the problem created by stereo-unselective
reactions.
Another approach to improve the yield of a KR to 100% is to combine the KR of an
ester (R-COOR ) with a subsequent Mitsunobu reaction (Scheme 8.11) [79–81]. After
KR of the ester of a secondary alcohol, a mixture of enantiopure ester and enantiopure
alcohol is obtained. The Mitsunobu reaction, the SN2 inversion of the alcohol, can be
performed with this mixture, since the Mitsunobu reagents only convert the alcohol
and leave the ester unchanged; 100% starting material is, ideally, recovered but now
enantiopure instead of racemic. This approach has the disadvantage that it generates
large amounts of waste and that acetate is not a very good nucleophile.
OAc
OAc lipase catalyzed R1 R2 Mitsunobu OAc

kinetic resolution reaction
R1 R2 OH R1 R2
R1 R2
Scheme 8.11 KR combined with a Mitsunobu reaction for 100% yield of an enantiopure product.
Another approach based on prochiral substrates is to start from symmetric diols,

diesters of diols, or diesters of diacids (Scheme 8.12). Even cyclic anhydrides can be
utilized. In all these cases the hydrolase can be used to break the symmetry by the
enantioselective conversion of just one of the two, chemically equivalent, functional
groups. This approach theoretically always leads to 100% yield with 100% enantios-
electivity [3, 5, 71].
X Y
n hydrolase, water n
R1 R1 100 %
n n
X X
n = 0,1,2,3...
X = COOR 2, OCOR2
Y = COOH, OH
Scheme 8.12 In the desymmetrization of prochiral bifunctional molecules the hydrolase is utilized
to break the symmetry.
An extension of the approach is the meso procedure (Scheme 8.13). It is in essence

the same strategy but starting with a meso compound, that is, again a prochiral
molecule. Again, theoretically 100% yield with 100% stereoselectivity can be
attained [3, 5, 71].
8.2 Enantioselective Hydrolysis of Racemic Acyclic Carboxylates (Resolutions) j265
R1 X R1 Y
hydrolase, water
100 %
R1 X R1 X
X = COOR 2 , OCOR 2
Y = COOH, OH
Scheme 8.13 A meso compound is a prochiral material that can be desymmetrized by the selective
hydrolysis or esterification of one of its functional groups.
8.2
Enantioselective Hydrolysis of Racemic Acyclic Carboxylates (Resolutions)
8.2.1
Overview
A classic biocatalytic approach towards chiral carboxylic acids or alcohols consists of

the hydrolysis of racemic esters in enzymatic hydrolysis reactions [3]. Thus, the
stereogenic center(s) can be integrated in either the alcohol or carboxylic acid moiety.
As enzyme components lipases, esterases, and proteases are preferably used. For
such an enzymatic resolution of chiral esters a broad range of different types of
racemic substrates has been successfully applied, typically leading to the desired
chiral carboxylates and alcohols in high enantiomeric excess.
8.2.2
Carboxylates with a Chiral Acid Moiety
The field of enzymatic hydrolytic reactions for the synthesis of chiral carboxylic acids
has been comprehensively reviewed [82]. Therefore, the focus here is on more
recently published contributions as well as on selected synthetically highly useful
examples leading to high yields, enantioselectivities, and volumetric productivities.
Notably, to date, a broad spectrum of racemic carboxylates has been resolved highly
successfully by hydrolases, and certainly this type of biotransformation belongs to the
most frequently applied enzymatic reaction types in organic chemistry. The substrate
spectrum consists of non-functionalized carboxylates, a-, b-, and c-heteroatom
substituted carboxylates, carboxylates with so-called remote stereogenic centers, and
many other (diverse) types of chiral carboxylates. Scheme 8.14 gives a graphical
overview of the substrate spectrum. Many of those compounds exhibit interesting
biological activity, thus representing important target molecules for pharmaceutical
purposes.
An example of a pharmaceutically important non-heteroatom functionalized
carboxylic acid is (S)-naproxen, which has been prepared by hydrolase-catalyzed
enantioselective hydrolysis of corresponding esters. Among important pharmaceu-
tical intermediates in the field of heteroatom-functionalized carboxylic acids, a-,
b-, and c-amino acids are particularly noteworthy. A recent trend can be seen in the
resolution of carboxylates bearing a remote stereogenic center.
266j 8 Hydrolysis and Formation of Carboxylic Acid Esters
R1
* CO H
2
R2
non-heteroatom NH2 OH
NH2 OH functionalized acids * CO2 H * CO2 H
* * R1 R1
R CO 2H R CO 2H R2 R2
α-heteroatom β -heteroatom
substituted acids substituted acids
enzymatic
resolution of
racemic carboxylates
R3
2 1
R R
R1 OH
S
R3 CO2H
R2 O
special applications acids with remote
stereogenic centers
Scheme 8.14 Overview of resolutions based on the hydrolysis of acyclic carboxylates with a chiral
acid moiety (selected examples).
8.2.2.1 Resolution of Carboxylates with a Non-functionalized Stereogenic Center

at the a-Position
To start with the resolution of a carboxylate, which is not functionalized with a
heteroatom, racemate rac-1 has been resolved successfully in the presence of a
recombinant esterase [83–85]. By means of this enzyme the resolution has been
carried out at a high substrate loading of 150 g l1, leading to the desired (S)-
enantiomer of naproxen in high enantiomeric excess of >98%. Notably, this process
for the synthesis of (S)-naproxen [(S)-2], an important anti-inflammatory drug, has
already been scaled up to pilot production scale (Scheme 8.15). For (S)-2 a range of
other resolutions based on enantioselective ester hydrolysis have also been reported.
An enzyme that has been studied in detail with respect to its suitability to hydrolyze
CH3 CH3 CH3
CO2Et hydrolase CO2Et CO2H

+
H3CO H3CO H3CO
rac-1 (R)-1 (S)-2
Scheme 8.15 Enzymatic synthesis of (S)-naproxen, (S)-2, based on a hydrolytic resolution process.
selectively a-substituted carboxylates is the lipase from Candida rugosa [86, 87].
Whereas the crude lipase from C. rugosa exhibits a low to moderate enantioselectivity
(E ¼ 10), the isomeric form B (obtained by protein chromatography) shows a high
enantioselectivity with an excellent E-value of >100 for the hydrolytic enantioselec-
tive cleavage of racemic methyl 2-phenylpropanoate. The Kazlauskas group found
that simple treatment of the crude lipase from C. rugosa with 50% aqueous
isopropanol and subsequent centrifugation and dialysis represents an attractive and
scalable alternative to extensive protein chromatography [88]. For example, resolution
of 2-chloroethyl 2-phenylpropionate proceeded with a high enantioselectivity of E >
100 when using the isopropanol-treated enzyme formulation in contrast to an E-value
of 10 when using the crude lipase from C. rugosa.
Very recently, the B€ackvall group developed a variant of a lipase from Candida
antarctica A by directed evolution, which turned out to be a highly efficient biocatalyst
for enantioselective hydrolysis of various a-substituted nitrophenyl esters [89].
Compared with E-values of 2–20 for the wild-type enzyme, the optimized enzyme
showed a tremendous improvement of the enantioselectivity, with E-values in the
range E ¼ 45–276. An increased activity was also found for most substrates and the
enantiospecificity is the opposite in comparison with the wild-type enzyme.
Owing to the great importance of (S)-naproxen and related drugs such as, for
example, ibuprofen many efforts have been made to realize a DKR. Although so far a
DKR based on the use of (alkyl) esters of naproxen still represents a challenge, DKRs
using corresponding thioesters have been carried out successfully. Based on pio-
neering work by Drueckhammer et al. [90] with related carboxylate derivatives, Tsai
et al. demonstrated the proof-of-concept for such a DKR process when starting from a
racemic thioester as substrate component [91–94]. An organic amine, for example,
trioctylamine, was used as a racemization catalyst and the reaction was carried out in
isooctane. However, an excess of the amine base is required, and this process requires
long reaction times while running at low substrate concentration. For example, when
using the racemic trifluoroethyl thioester rac-3 and a high enzyme loading of 30 g l1
in combination with a reaction time of 294 h, a yield of 67% and an enantioselectivity
in the range 92–98% e.e. was achieved (Scheme 8.16) [91].
Furthermore, a DKR for the synthesis of naproxen has also been carried out
starting from the corresponding racemic trifluoroethyl ester [94]. A DKR starting
from racemic ibuprofen-derived 2-ethoxyethyl ester has been reported, too [95].
An interesting DKR using the methyl ester rac-4 as starting material has been
reported by Pietruszka et al. recently [96]. This DKR consists of an enantioselective
enzymatic ester hydrolysis in combination with an in situ racemization of the ester
substrate in the presence of 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) as a base. Using
Candida antarctica lipase B (CAL-B) together with 1.2 equivalents of TBD in buffer/n-
heptane as a two-phase reaction medium furnished the desired (R)-2,3-dihydro-1H-
indene-1-carboxylic acid (R)-5 in a high yield of 95% and with >96% e.e.
(Scheme 8.17).
A highly efficient synthesis of an indole ethyl ester (R)-6, an intermediate for a
prostaglandin D2 receptor antagonist, by applying a lipase-catalyzed hydrolysis of the
corresponding racemate has been reported by Merck researchers [97, 98]. The desired
CH3 CH3
lipase
COSCH2CF3 CO2H
water
H3CO H3CO
(S)-3
(S)-2
67% yield
Oct3N 92-98% ee
CH3
COSCH2CF3
H3CO
(R)-3
Scheme 8.16 DKR using a racemic thioester.
CO2Me lipase from CO2H

C. antarctica B
TBD (1.2 eq.)

buffer, pH > 8.5
rac- 4 n-heptane (R)-5
95% yield
>96%ee
Scheme 8.17 Synthesis of (R)-2,3-dihydro-1H-indene-1-carboxylic acid via DKR.
(R)-ester, (R)-6, was obtained with an excellent enantiomeric excess of 99.7% e.e. as
remaining ester after hydrolysis of rac-6 with a perfect conversion of 50%
(Scheme 8.18). As catalyst, a lipase from Pseudomonas fluorescens was used. Notably,
the reaction runs at a high substrate concentration of 100 g l1. In addition, the
process turned out to be technically feasible and was applied successfully on a 40 kg
scale [98].
Another pharmaceutically interesting resolution is the enantioselective hydrolysis
of rac-3-(acetylthio)-2-methylpropanoate (rac-8) in the presence of a hydrolase to give
lipase from
Pseudomonas
f luorescens
+
N buffer/DMF(3:1), N CO 2Et N CO 2H
H CO2 Et H H
pH 8.0,
rac- 6 50% conversion (R)-6 (S)- 7
(100 g/l) >99% ee
Scheme 8.18 Enzymatic resolution via enantioselective hydrolysis of a racemic indole-derived

ester.
the desired acid (S)-9. This is an intermediate in the synthesis of the antihypertensive
drug captopril [99]. When using an esterase from a Pseudomonas sp. strain, acid (S)-9
was successfully obtained with both excellent conversion of 49% and enantiomeric
excess of 99.9% (Scheme 8.19). Alternatively, lipases have also been applied suc-
cessfully as biocatalysts.
esterase
CH3 from CH3
H3C S OCH3 Pseudomonas sp. H3C S OH
O O 49% conversion O O
rac-8 (S)-9
99.9%ee
Scheme 8.19 Enzymatic resolution of rac-3-(acetylthio)-2-methylpropanoate.
A different approach towards the drug captopril is based on the synthesis of the
racemic b-chloropropionate (rac-10) via hydrolytic resolution in the presence of a
lipase from Candida rugosa (Scheme 8.20). This process applied at DSM is based on
the hydrolysis of the undesired (R)-enantiomer, while the remaining desired (S)-
ester (S)-11, which serves as the intermediate for the synthesis of captopril, was
obtained with a high enantiomeric excess of 98% at a conversion of 64% [100, 101].
lipase from
CH3 CH3 CH3
C. rugosa
Cl OCH3 Cl OH Cl OCH3
64% conversion
O O O
rac- 10 (R)-11 (S)-10
98% ee
Scheme 8.20 Enzymatic resolution of rac-b-chloropropionate.
A novel recombinant isoform of pig liver esterase (PLE), namely, the so-called
alternative pig liver esterase (APLE), has been successfully expressed and directed for
secretion in Pichia pastoris by Pichler et al. recently [102]. This enzyme was
subsequently used for a highly enantioselective hydrolysis of methyl (2R,4E)-5-
chloro-2-isopropyl-4-pentenoate as an industrially highly valuable substrate. This
resolution proceeds with an excellent enantioselectivity as indicated by the high
E-value of >200.
When using diester substrates with an a-non-heteroatom, functionalized stereo-
genic carbon center, regio- and enantioselective enzymatic resolutions are conceiv-
able. A synthetic example has been reported by Cohen et al. for the synthesis of the
chiral acid (S)-13 starting from the racemic diester rac-12 [103]. When using
a-chymotrypsin, the (S)-acid (S)-13 was obtained regioselectively as the favored
enantiomer with an enantioselectivity of E ¼ 13 (Scheme 8.21). The opposite regio-
selectivity was observed when using a lipase from porcine pancreas, which then gave
the corresponding (S)-acid with an enantioselectivity of E ¼ 23 [104].
CH3 α -chymotrypsin CH3 CH3
H3CO2C CO2CH3 buffer H3CO2C CO2H H3CO2C CO2CH3

E =13
rac-12 (S)-13 (R)-12
70% ee 76% ee
Scheme 8.21 Regio- and enantioselective enzymatic resolution of a racemic diester.
A synthetic example for an enzymatic hydrolytic resolution of a racemic ester

bearing a quaternary carbon center without any heteroatom substituent at the
a-position has been reported by Zwanenburg and coworkers [105, 106]. Transfor-
mation of the tricyclodecadienone ester rac-14 in the presence of PLE led to the
corresponding acid 15 in 40% yield and with >99% e.e. (Scheme 8.22). Notably, the
type of cosolvent has a significant impact on both reactivity and enantioselectivity and
a buffer/acetonitrile mixture turned out to be the reaction medium of choice.
CO 2Et HO2C CO 2Et

PLE
buffer, pH 8,
acetonitrile
O O O
rac- 14 15 14
(only 1 enantiomer 40% yield 48% yield
shown) >99% ee 83% ee
Scheme 8.22 Hydrolytic resolution of a racemic ester bearing a quaternary carbon center.
8.2.2.2 Resolution of Carboxylates with an Amino-Functionalized Stereogenic Center

at the a-Position
A widely applied lipase-catalyzed resolution is the enantioselective hydrolysis of
racemic a-amino acid esters, which is especially useful for the synthesis of non-
proteinogenic a-amino acids [107]. A representative example is the efficient kinetic
resolution achieved by Kazlauskas et al. of racemic octyl pipecolate rac-16 using a
lipase from Aspergillus niger [108]. The desired (S)-2-piperidinecarboxylic acid, (S)-
17, was obtained with an enantioselectivity of E > 100 (Scheme 8.23); (S)-17 is the
essential intermediate for enantiopure local anesthetics.
lipase from
A. niger
N CO2-n-C8H17 pH 5 N CO2H N CO2-n-C8H17

H H H
E > 100
rac- 16 (S)-17 (R)-16
Scheme 8.23 Lipase-catalyzed enantioselective hydrolysis of racemic octyl pipecolate.

A further example is the efficient enantioselective resolution for the kg-scale
production of N-benzyl (S)-4,4-difluoro-3,3-dimethylproline, (S)-19, which is a key
intermediate in the synthesis of HIV protease inhibitors [109]. This process,
developed by Pfizer researchers, runs at a substrate level of 100 g l1. By applying
PLE for this resolution process (S)-19 was obtained with >99% e.e., an enantiomeric
selectivity of E >200, and a total conversion of 44% within 24 h (Scheme 8.24).
Ph CO2 CH 3 Ph CO2 H
PLE
N N
CH 3 CH 3
CH 3 buffer, pH 8.0, CH 3
acetonitrile (10%)
F F F F
rac- 18 (S)- 19
substrate loading 44% yield
100 g/l 99% ee
Scheme 8.24 Enzymatic synthesis of N-benzyl-(S)-4,4-difluoro-3,3-dimethylproline [(S)-19].
Engineering of the reaction medium for lipase-catalyzed resolution via ester

hydrolysis at high substrate input has been reported jointly by Landfester, Gr€ oger,
and coworkers. In the presence of porcine pancreas lipase, the hydrolytic resolution
of racemic phenylalanine n-propyl ester reaction proceeds at substrate concentra-
tions of up to 827 g l1 of solvent [110].
A further example of an unusual nitrogen-containing ester moiety as substrate in
an enzymatic enantioselective hydrolytic resolution has been reported by Pfizer
researchers [111, 112]. The desired acid (S)-21 is a key intermediate in the synthesis
of a human rhinovirus protein inhibitor. When using an alkaline protease from
Bacillus lentus and racemate 20 as substrate enzymatic resolution gave the desired
product (S)-21 in 49% yield and 98% e.e. (Scheme 8.25). This enzymatic process was
conducted in the presence of 40% of acetonitrile as cosolvent and runs at a substrate
O
N N CO2H
protease from O N
O H
B. lentus O
N N CO2Et
O N
H water, pH 8, (S)-21
O
acetonitrile (40%) 49% yield
98% ee
rac-20
100 g/l
O
N N CO2Et
O N
H
O
(R)-20
Scheme 8.25 Protease-catalyzed synthesis of an a-nitrogen-substituted acid.

272
loading of 100 g l1. Furthermore, it was possible to racemize subsequently the

undesired enantiomer of the ester, (R)-20, when using the base DBU under mild
conditions. The protease from Bacillus lentus, which had never before been reported
to be used as a biocatalyst, was identified by screening of a comprehensive library of
hydrolases. Other efficient resolutions of pharma intermediates via enzymatic
hydrolysis have also been reported by the same group [113, 114].
The a-amino acid ester rac-22, which can be also regarded as an aniline derivative,
has been successfully resolved by researchers from LG Chem and LG Life
Sciences [115, 116] in the presence of an immobilized lipase from Pseudomonas
cepacia. The acid (R)-23 was formed with 98.5% e.e. at a conversion of 47%
(Scheme 8.26). This desired product (R)-23 is a key intermediate in the synthesis
of the fungicide mefenoxam. The immobilized enzyme was successfully re-used with
an overall recycling of >20 times without loss of activity.
O immobilized O
H lipase from H
N O Pseudomonas cepacia N
O OH
pH 7.0
47% conversion
rac- 22 (R)-23
98.5% ee
Scheme 8.26 Enzymatic hydrolytic resolution for the synthesis of an intermediate of the fungicide
mefenoxam.
Furthermore, a DKR based on a protease-catalyzed ester hydrolysis in combination

with amino ester racemization using an aromatic aldehyde has been developed by the
Beller group. For example, L-tyrosine, (S)-26, is formed in 92% yield and with 97%
e.e. from the corresponding racemic benzyl ester rac-24 when using AlcalaseÒ in
combination with 3,5-dichlorosalicylaldehyde, 25 (Scheme 8.27) [117]. Based on a
modeling study of the reaction course, this type of DKR process (exemplified for
racemic phenylalanine ethyl ester as a substrate) has been optimized by Beller and
Kragl et al., leading to a 10% higher yield and a decrease of the reaction time by factor
Alcalase
H2O, pH 8.5,
HO HO
NH2 acetonitrile NH2
O OH
O
O Cl
O
rac- 24 (S)-26
OH 92% yield
Cl 97% ee
25
(5 mol%)
Scheme 8.27 Chemoenzymatic dynamic kinetic resolution of a-amino esters.

2 [118]. Pyridoxal 5-phosphate as aldehyde was successfully used in combination with
a-chymotrypsin and Alcalase, respectively [119, 120].
8.2.2.3 Resolution of Carboxylates with a Hydroxy-(or oxo-)Functionalized

Stereogenic Center at the a-Position
In addition, hydroxy-substituted carboxylates also turned out to be versatile
substrates in the hydrolase-catalyzed kinetic resolution. As an early example,
rac-dimethyl malate 27 has been enantioselectively hydrolyzed to form the corre-
sponding monocarboxylic acid 28 (Scheme 8.28), an intermediate in the total
synthesis of (–)-tulipalin B [121]. Pig liver esterase (PLE) was a suitable biocatalyst
for this transformation as well as for the hydrolysis of other a-hydroxy carboxy-
lates [122, 123] and the synthesis of b-glucuronic acid [124].
CO2Me PLE CO2H

MeO2C MeO2C
OH OH
rac-27 (S)-28
Scheme 8.28 Enzymatic resolution of racemic dimethyl malate.
Lipases are also suitable for the resolution of more complex molecules bearing
more than one additional functional group. This is exemplified by the enzymatic
hydrolysis of rac-29, a key building block of epothilones. The Wessjohann and
Bornscheuer groups found that in presence of Burkholderia cepacia lipase (Amano
PS) the acyloin acetate rac-29 was resolved with an E value of >300, leading to the
corresponding diol (3S,10R)-30 in >99% e.e. (Scheme 8.29) [125].
Me Me Me
lipase from
O O O
Burkholderia 10 10
OH cepacia 3 OH 3 O
Me Me Me Me + Me Me H
O Me buffer/toluene, OH O Me
50% conversion
O (3S,10R)-30 O
rac-29 >99% ee (3R,10R)-29
E>300
>98% ee
Scheme 8.29 Enzymatic hydrolysis of a key building block of epothilones.
Further examples in this field are resolutions of racemic mandelic acid esters and
related derivatives, in particular those bearing a quaternary stereogenic carbon
center [126]. A direct resolution of free (non-O-acylated) racemic methyl man-
delate has been reported by Wong and coworkers, leading to a resolution process
with a moderate enantioselectivity of 40% e.e. [127]. A highly enantioselective
resolution has been achieved by Fuganti and Rosell et al. when using racemic
O-formylated methyl mandelate (rac-31) as substrate in combination with a
274
penicillin amidase [128]. This enantioselective hydrolysis affords the desired O-

protected acid (S)-32 with >98% e.e. at a conversion of 50% (Scheme 8.30).
O penicillin O
amidase
O H O H
CO2CH3 50% conversion CO2H
rac-31 (S)-32
>98% ee
Scheme 8.30 Enzymatic resolution of racemic O-formylated methyl mandelate with a penicillin
amidase.
Racemic mandelate derivatives bearing a quaternary stereogenic carbon center

have been successfully resolved as well. For example, this has been demonstrated by
Kellogg and Kloosterman et al. for the synthesis of (S)-carboxylic acid (S)-34 starting
from the corresponding racemic ethyl ester [129, 130]. When applying PLE as a
biocatalyst, the desired carboxylic acid (S)-34 was obtained in 39% yield and with 80%
e.e. (Scheme 8.31). The enantiomeric excess of (S)-34 has been further increased by
one crystallization step, leading to an enantiomerically pure product. The remaining
(R)-ester (R)-33 was isolated in 44% yield and 86% e.e.
OH PLE OH OH
CO2Et CO2H CO2Et
rac-33 (S)-34 (R)-33

39% yield 44% yield
80% ee 86% ee
Scheme 8.31 Enzymatic resolution of a racemic mandelate bearing a quaternary stereogenic

carbon center.
Related a-hydroxy carboxylates of type rac-35, bearing a trifluoromethyl group

at the a-carbon center, have been resolved successfully by Faber and Griengl et al.
[131, 132]. In the presence of a protease from Aspergillus oryzae hydrolytic resolution
of rac-35 proceeds enantioselectively with a conversion of 40%, leading to the (R)-acid
(R)-36 with 88% e.e. (Scheme 8.32). A single recrystallization step led to an
enantiomerically pure product. The protease from Aspergillus oryzae also turned out
to be a suitable catalyst for enantioselective hydrolysis of several other sterically
hindered racemic carboxylates.
protease from
F3C OH Aspergillus oryzae F3C OH
CO2Me buffer CO2H
40% conversion
rac-35 (R)-36
88% ee
Scheme 8.32 Enzymatic resolution of a-hydroxy carboxylates bearing a trifluoromethyl group at

the a-carbon center.
Furthermore, a range of pharmaceutically interesting a-aryloxypropanoic acids

have been prepared enantioselectively by enzymatic resolution. Candida cylindracea
lipase turned out to be a suitable biocatalyst, hydrolyzing, for example, rac-37, to
furnish the corresponding (S)-acids, for example, (S)-38, with enantioselectivities of
up to E > 100 (Scheme 8.33) [133, 134]. Notably, additives such as levomethorphan or
dextrometorphan (as enantioselective inhibitors) were required to obtain such high
enantioselectivities. For example, in the presence of these bases the Sih group
obtained high enantioselectivities (E > 100) for the resolution of racemic methyl
a-(4-chlorophenyl)propanoate (rac-37), whereas in the absence of such bases a
significantly decreased enantioselectivity of E ¼ 17 has been obtained (Scheme 8.33).
This positive effect of such additives is based on a non-competitive inhibition of the
lipase. The drawback of using morphinan alkaloids as additives, in particular with
respect to large scale processes, has been successfully addressed by using simpler
amine bases such as N,N-dimethyl-4-methoxyphenylethylamine (DMPA) [135].
lipase from
Cl C. cylindracea Cl Cl
CH3 CH3 CH3
O CO2CH3 buffer O CO2H O CO2CH3

additives:
rac-37 (R)-38 (S)-37
none: 50% conversion
E = 17
DM: dextrometorphan DM: 41% conversion
LM: levometorphan E = >100
LM: 31% conversion
E = >100
Scheme 8.33 Enantioselective synthesis of an a-aryloxypropanoic acid via enzymatic resolution.
8.2.2.4 Resolution of Carboxylates with Two Heteroatom-Substituted Stereogenic

Centers at the a,b-Positions
A range of racemic carboxylates bearing two heteroatom-substituted stereogenic
centers at the a,b-positions has also been successfully resolved by a biocatalytic
hydrolytic reaction. A pharmaceutically important example is the enantioselective
synthesis of the enantio- and diastereomerically pure a,b-epoxy ester (2R,3S)-39, an
intermediate in the preparation of the calcium channel blocker drug diltiazem [136].
276
This desired ester has been synthesized by, for example, Tanabe Seiyaku researchers
in a hydrolytic resolution process using a lipase from Serratia marcescens starting
from the diastereomerically pure racemate rac-39, leading to the remaining enan-
tiomer (2R,3S)-39 in 40–43% yield and with an excellent enantiomeric excess of
>99% (Scheme 8.34) [137]. Furthermore, a highly elegant downstream processing
has been developed. The phenylacetaldehyde 41 formed through spontaneous
decarboxylation from the undesired acid (2S,3R)-40 suppresses crystallization of
the desired ester (2R,3S)-39 but has been removed from the organic phase via
formation of the water soluble bisulfite adduct 42.
H3CO lipase from H3CO H3CO

S. marcescens
H2O
O CO2Me O CO2Me O CO2H
rac-39 (2R,3S)-39 (2S,3R)-40

40-43% yield
>99% ee
- CO2
H3CO + NaHSO3 H3CO

OSO2Na
CHO
OH
42 41
Scheme 8.34 Enzymatic enantioselective synthesis of an enantio- and diastereomerically pure

a,b-epoxy ester.
A further resolution is based on the use of related racemic 1,3-dioxolane-4-

carboxylates as substrates [138]. These reactions were conducted in the presence
of a lipase from Candida rugosa, which also tolerates racemates bearing a quaternary
stereogenic center. For example, the racemate rac-43a has been enantioselectively
hydrolyzed to afford the acid (S,R)-44a with an enantioselectivity of E ¼ 91
(Scheme 8.35a). In addition, racemic C2-symmetric diesters with two stereogenic
centers are also suitable substrates for enzymatic hydrolytic resolution. For example,
in the presence of porcine liver esterase (PLE), the resolution of racemic diester rac-
43b proceeds with a high enantioselectivity of E > 145, leading to both the acid
product (S,S)-44b and remaining ester substrate (R,R)-43b with >95% e.e.
(Scheme 8.35b) [139]. Resolutions of C2-symmetric racemic diesters that are non-
functionalized at the a,b-position have also been reported [140].
8.2.2.5 Resolution of Carboxylates with an Amino-Functionalized Stereogenic

Center at the b-Position
Enzymatic hydrolysis is also very suitable for the lipase-catalyzed resolution of a
broad range of b-amino acid esters. Pioneering work in this field is the resolution of
(a)
lipase from
O O C. rugosa O O O O
CO2H CH3
H3C CO2n-Bu buffer H3C CH3 H3C CO2n-Bu
E = 91
rac- 43a
(S,R)-44a (R,S)-43a
(only 1 enantiomer
shown)
(b)
porcine liver
O esterase (PLE) O O
H3CO2C CO2CH3 buffer H3CO2C CO2H H3CO2C CO2CH3
rac-43b E > 145 (S,S)-44b (R,R)-43b

>95% ee >95%ee
Scheme 8.35 Enzymatic resolution of (a) a 1,3-dioxolane-4-carboxylate with two stereogenic

centers and (b) a C2-symmetric diester.
the b3-amino acid ester rac-45 in the presence of a hydrolase, reported by researchers
from Sumitomo Pharmaceuticals [141]. In an initial screening the lipase from
Candida antarctica B was found to be the preferred biocatalyst. After reaction medium
engineering an efficient resolution process with this biocatalyst was developed that
gave the remaining substrate (S)-45 at 50% conversion in 95% e.e. when using THF
with 5% water content as solvent. The long reaction time of 96 h was successfully
optimized by switching to acetone containing 10% water as a solvent. Methyl (S)-
1,2,3,4-tetrahydroquinoline-2-acetate, (S)-45, was obtained as a remaining substrate
after 20 h at 50% conversion and with 94% e.e., which corresponds to a high
enantioselectivity of the resolution process of E ¼ 115 (Scheme 8.36).
lipase from
C. antarctica B
CO2Me CO2H CO2Me
N H2O (10%), N N
H H H
acetone (90%)
rac-45 50% conversion (R)-46 (S)-45
94% ee
E=115
Scheme 8.36 Enzymatic resolution of a racemic 1,2,3,4-tetrahydroquinoline-2-acetate.
Enantioselective synthesis of 3-arylated b-amino acids of type (S)-48 in a lipase-

catalyzed resolution of racemic esters has been reported by Celltech Chiroscience and
Chirotech Technology researchers [142]. In the presence of a lipase from Pseudo-
monas cepacia resolution proceeds with a broad range of substrates and the desired
b-amino acids (S)-48 were obtained by simple filtration from the reaction mixture in
yields of 13–46% and enantiomeric excesses of 77–99%. Scheme 8.37 gives an
278
lipase from
NH2 P. cepacia NH2 NH2
CO2Et CO2H CO2Et
Ar Ar Ar
rac-47 (S)-48 (R)-47
a: Ar = Ph a: 44% yield, 99% ee a: 36% yield, 98% ee
b: Ar = 3-Br-C6H4 b: 44% yield, 77% ee b: 42% yield, 74% ee
c: Ar = 4-Br-C6H4 c: 46% yield, 99% ee c: 18% yield, 99% ee
d: Ar = 4-F-C6H4 d: 13% yield, 91% ee d: 46% yield, 90% ee
Scheme 8.37 Enzymatic resolution of 3-arylated b-amino acid esters.
overview of different synthesis examples. The biotransformation is volume efficient

and runs with a substrate input of 200 g l1.
Substrate as well as reaction medium engineering of this type of resolution process
has been carried out by Degussa researchers [143, 144]. As reaction media a
classical biphasic solvent mixture of MTBE (methyl tert-butyl ether) and water, as
well as miniemulsions, have been highly suitable for the biotransformation of the
propyl ester rac-49 as preferred ester. A miniemulsion, so far a rarely explored type of
reaction medium for biotransformations, was prepared by treatment of an aqueous
mixture consisting of rac-49, a surfactant, and hexadecane, using ultrasound. The
enzymatic resolution proceeds at a high substrate input of 484 g l1 with 45%
conversion, leading to the desired b-amino acid (S)-50 in 37% yield and >99.4% e.e.
after isolation (Scheme 8.38) [143]. The use of n-propyl esters as a preferred ester
group in b-amino ester substrates in this type of resolution has also been reported by
Ube Industries researchers [145].
NH2 O lipase from NH2 O

P. cepacia
CH3
O OH
miniemulsion
as a reaction
rac- 49 medium (S)-50
484 g/l H2O, surfactant, 37% yield
hexadecane, >99.4%ee
ultrasound
pH 8.2
45% conversion
Scheme 8.38 Enzymatic resolution of a racemic b-amino acid ester in a miniemulsion.
A further b-amino acid as pharmaceutical building block is the benzodiazepine-

derived acid (S)-53, which represents a key intermediate in the synthesis of lotrafiban.
This compound has been synthesized by researchers from GlaxoSmithKline Phar-
maceuticals via enzymatic hydrolytic resolution as a key step running in water–tert-
butanol reaction media [146]. Using Candida antarctica B lipase, resolution proceeded
successfully and gave in combination with a subsequent regioselective iodination the
desired b-amino acid (S)-53 with an average yield of 37.5% (over 13 batches in a
production environment) and with excellent enantiomeric excess of >99.9%
pyridine
iodine
Me lipase from Me monochloride Me
N C. antartica B N complex I N
O O O
H2O, pH 7.0 H2O, NaOH,
N N N
H CO2Me t-BuOH H CO2H pH 7.0 H CO2H
rac-51 (S)-52 (S)-53
not isolated 37.5% average yield
(over 13 batches)
>99.9% ee
Scheme 8.39 Enantioselective synthesis of a b-amino acid as a key intermediate in the synthesis of
lotrafiban.
(Scheme 8.39). The resolution process has also been carried out successfully in a
water–ionic liquid mixture as a reaction medium [147].
A highly enantioselective hydrolysis of racemic alicyclic cis- and trans-b-amino
esters rac-54 in presence of lipase from C. antarctica B has been reported by the F€
ul€
op
group [148]. The hydrolysis was performed in diisopropyl ether containing only 0.5
equivalents of water. The resulting cis-b-amino acids, for example, (1S,2R)-55, were
obtained in high yields of 42–47% and with excellent enantiomeric excess of 96–99%
e.e. The opposite enantiomers have been isolated after hydrolysis as hydrochloric acid
salts of the resulting amino acids (1R,2S)-55 in high enantiomeric excess, too.
Scheme 8.40 gives a representative example.
lipase from
CO2Et C. antarctica B CO2H CO2Et
+
NH2 water, i-Pr2O, 65°C NH2 NH2
49% conversion
rac-cis-54 (1S,2R)-55 (1R,2S)-55
E>200 47% yield
98% ee HCl
CO2H
NH3 Cl
(1R,2S)-55
46% yield
99% ee
Scheme 8.40 Enzymatic resolution of a racemic alicyclic cis-b-amino ester.
Transformation of a mixture of the meso cis- and racemic trans-diastereomers (d.r.

ratio of cis/trans ¼ 15 : 85) of the cyclic b-amino acid ester cis(meso)/rac-trans-56 into
the corresponding (trans) (R,R)- and (S,S)-enantiomers (R,R)-57 and (S,S)-57,
which are intermediates for potential inhibitors of aspartyl proteases, has been
reported by F. Hoffmann-La-Roche researchers [149]. The first step of this conse-
cutive enzymatic synthesis consists of a selective hydrolysis of the cis-diastereomer
(meso-form) in the presence of a lipase from Candida rugosa (lipase AY), leading to a
nearly diastereomerically pure racemic trans-product rac-trans-56 showing a very low
content of only 0.2% of the undesired cis-diastereomer in 84% yield (Scheme 8.41).
Using rac-trans-56 as starting material for a subsequent resolution in the presence of
a Candida antarctica lipase B, Iding et al. succeeded in synthesizing the diastereomer
(R,R)-56 in 41.2% yield and >99% e.e. After identifying an enzyme with the opposite
enantiopreference, a route for the synthesis of the opposite (S,S)-enantiomer with a
high enantiomeric excess of 97.7% e.e. has been developed as well. Subsequently, the
diesters were monohydrolyzed smoothly in the presence of a pig liver esterase (PLE),
leading to the (R,R)-monoester (R,R)-57 in 88.4% yield and >99% e.e. Scheme 8.41
shows the synthetic sequence for the preparation of the monoester (R,R)-57 starting
from the cis/trans-mixture cis(meso)/rac-trans-56.
HO2C CO2Me
N
Boc
MeO2C CO2Me lipase from MeO2C CO2Me lipase from
C. rugosa C. antarctica B (S,S)-57
N buffer, pH 7.5, N buffer, pH 7.5,

Boc cyclohexane Boc cyclohexane MeO2C CO2Me
cis(meso)/rac-trans- 56 rac-trans- 56
(only 1 enantiomer N
shown) Boc
(R,R)-56
41.2%yield
HO2C CO2Me >99%ee
PLE
N
Boc buffer, pH 7.5,
MeOH
(R,R)-57
88.4% yield
>99% ee
Scheme 8.41 Enzymatic transformation of a meso cis- and rac-trans-b-amino diester mixture into an
enantio- and diastereomerically pure monoester.
8.2.2.6 Resolution of Carboxylates with a Hydroxy-(or oxo-)Functionalized Stereogenic

Center at the b-Position
The enantioselective synthesis of b-hydroxy-substituted carboxylic acids is also of
pharmaceutical interest. An elegant enzymatic approach towards (S)-3-hydroxy-3-
phenylpropanoic acid (S)-59, an intermediate in the synthesis of the (S)-enantiomers
of the antidepressants tomoxetine and fluoxetine, has been reported by several
groups [150–152]. The hydrolytic resolution starts from racemate rac-58; an enan-
tioselectivity of 93% e.e. at a conversion of 39% has been obtained for the b-hydroxy
acid (S)-59 when using a commercial lipase PS-30 from Pseudomonas sp. (Amano) as
a biocatalyst (Scheme 8.42) [151]. A further hydrolase used for this reaction is porcine
liver esterase, although resolution proceeds with lower enantioselectivity [150–152].
OH lipase from OH
CO2Et Pseudomonas sp. CO2H
buffer, pH 7
39% conversion
rac-58 (S)-59
93% ee
Scheme 8.42 Enzymatic resolution of an aromatic b-hydroxy carboxylic ester.
A completely different and untypical racemization strategy for a-non-heteroatom

functionalized acids has been applied successfully by Bristol-Myers Squibb research-
ers in a DKR for the enantioselective synthesis of a roxifiban intermediate
(Scheme 8.43) [153, 154]. In this process the in situ racemization is based on a
(reversible) base-catalyzed retro-Michael addition of the racemic starting material rac-
60. When using a lipase from Pseudomonas cepacia this type of DKR led to the desired
acid (R)-61 in 80% yield and 94% e.e. The thioester was used as an activated ester
substrate on a 44.9 kg scale, and trimethylamine was chosen as a base. After
recrystallization the desired product (R)-61 was obtained in enantiomerically pure
form (>99.9% e.e.).
O O O
lipase from
O O Pseudomonas O
N Sn-Pr N Sn-Pr N O
cepacia H
NC NC NC
(S)-60 (R)-60 (R)-61
80% yield
NEt3 NEt3 94% ee
O O
O O
N Sn-Pr N Sn-Pr
NC NC
O
O
N Sn-Pr
NC
Scheme 8.43 Chemoenzymatic DKR for the synthesis of a roxifiban intermediate.

282
8.2.2.7 Resolution of Carboxylates with a Stereogenic Heteroatom Center

at the b-Position
Hydrolases are not only capable of recognizing stereogenic carbon centers but also
stereogenic heteroatom centers. An interesting class of chiral molecules with a
stereogenic heteroatom center are sulfoxides. The resolution of racemic esters bearing
a chiral sulfoxide moiety at the b-position such as rac-62 has been reported by Burgess
and a coworker [155]. In the presence of a lipase from Pseudomonas sp., the hydrolytic
resolution afforded the corresponding acids in yields of 17–38% and enantiomeric
excesses from 88 to >98%, as demonstrated for various synthetic examples. When
starting from rac-62 as substrate the desired acid (S)-63 was formed with an
enantioselectivity of E > 200 (Scheme 8.44). In all cases the remaining substrate
was isolated with a high enantiomeric excess of >98% in yields in the range 33–49%.
lipase from
O O O O O O
Pseudomonas sp.
S S S
OCH3 OH OCH3
buffer/
O2N toluene O2N O2N
E > 200
rac-62 (S)-63 (R)-62
97% ee >98% ee
Scheme 8.44 Enzymatic resolution of an ester with a chiral sulfoxide moiety.
An enzymatic hydrolytic resolution has been also applied by Mikolajczyk and

Kielbasinski et al. for the resolution of racemic phosphine oxides of the type rac-64,
bearing a chiral phosphorus center as well as an ester group for hydrolytic clea-
vages [156]. Scheme 8.45 shows a representative example. In the presence of a
porcine liver esterase, the resolution proceeds under formation of the remaining
ester (R)-64 in 40% yield and with about 95% e.e., whereas the acid (S)-65 is formed
in 44% yield and with 64% e.e.
O PLE O O
Ph
P CH2CO2Me Ph P CO2Me MeO P CO2H
MeO buffer,
NaOH, H2O MeO Ph
rac-64 (R)-64 (S)-65
40% yield
ca. 95% ee
Scheme 8.45 Enzymatic resolution of an ester with a chiral phosphine oxide moiety.
8.2.2.8 Resolution of Carboxylates with a Remote Stereogenic Center

The capability of hydrolases to also recognize remote chiral centers has been
demonstrated by several groups. Resolutions of this type have been reported recently
in particular in the field of synthesis of drug intermediates. A selected example of
remote stereogenic center recognition is the hydrolase-catalyzed synthesis of an
intermediate of the novel GARFT inhibitor pelitrexol, which has been reported by
Pfizer researchers [157]. In the substrate of type rac-66, the stereogenic center in the
tetrahydropterin moiety is located far from the ester moiety that has to be hydrolyzed
in the enzymatic resolution. Screening a range of hydrolases resulted in low
enantioselectivities with E values of 1–2 only. Introduction of an oxamic ester adjacent
to the stereogenic center increased the enantioselectivity of the enzymatic hydrolytic
reaction. When using oxamic ester rac-66 as substrate in the presence of C. antarctica
lipase B, enzymatic hydrolysis proceeds with a significantly improved enantioselec-
tivity of E ¼ 59 and conversion of 45% after optimization (Scheme 8.46). After
work-up the desired (S)-carboxylic acid was obtained with >38% yield and high
enantiomeric excess (>98% e.e.).
H 3C H 3C
O O
H lipase from H
O CH3 Candida antarctica B OH
HN S HN S
R O buffer, pH 4.2, R O
N N N N N N
H DMF (30%(v/v)) H
O O
O E=59 O
OEt OEt
r ac-66 45% conversion (S)-67
95% ee
Scheme 8.46 Enzymatic resolution of an intermediate of pelitrexol.
8.2.2.9 Resolution of Carboxylates with Axial and Planar Chirality

A further interesting class of chiral molecules are compounds that do not have a
stereogenic carbon (or heteroatom) center. Typical members in this field show axial or
planar chirality. The suitability of enzymes in differentiating between the two
enantiomers of an axial chiral racemate has been demonstrated by Jones and
coworkers in the synthesis of a range of chiral allenes of type 69 (Scheme 8.47) [158].
In presence of porcine liver esterase, resolutions proceed at conversions of up to 54%
to furnish the chiral acids with up to 93% e.e. Scheme 8.47 shows a selected example.
The advantageous use of an immobilized form of a PLE has been demonstrated by
Pietzsch et al. for the resolution of a racemic allene [159]. Compared to biotrans-
formations in presence of the free biocatalyst, the immobilized PLE led to a fourfold
enhancement of enantioselectivity (E ¼ 60). Under optimized conditions, consisting
of the use of acetone as cosolvent and the emulsifier Triton X-100, the desired acid was
obtained in 41% yield and with 96% e.e.
porcine liver
H 3C CO 2CH3 esterase (PLE) H3 C CH 3 H3 C CO2 CH 3
Ph CH 3 buffer Ph CO2 H Ph CH 3
E = 35
r ac-68 69 68
90% ee 61% ee
Scheme 8.47 Enzymatic resolution of a chiral allene containing an ester moiety.

284
In a recent joint contribution by the Reetz and B€ackvall groups an evolved lipase
from Pseudomonas aeruginosa has been reported as a highly enantioselective biocat-
alyst for the resolution of the racemic allene rac-70, leading to enantioselectivities of
up to E ¼ 111 (Scheme 8.48) [160]. The high enantioselectivity of the preferred
Leu162Phe mutant was rationalized by a p–p stacking between the phenyl moiety of
phenylalanine in position 162 and the p-system of the allene.
lipase mutant
Me (P. aeruginosa Me
• O mutant Leu162Phe) •
CO2H
O 44% conversion
NO2
E = 111
r ac- 70 71
96% ee
Scheme 8.48 Resolution of an allenic ester with an evolved lipase mutant.
Furthermore, racemates with planar chirality have also been successfully bioca-
talytically resolved. Crout and coworkers reported an enantioselective hydrolysis of
the racemic iron carbonyl complex rac-72 in the presence of PLE, which produced the
acid 73 in 40% yield and with an enantiomeric excess of 85% (Scheme 8.49) [161].
(CO) 3Fe porcine liver (CO)3 Fe

CO2 Et esterase (PLE) CO2 H CO 2Et
buffer
E = 33 (CO)3 Fe
rac-72 72
73 85% ee
Scheme 8.49 Enzymatic resolution of a racemate possessing planar chirality.
8.2.3
Carboxylates with a Chiral Alcohol Moiety
A second option for enzymatic hydrolytic resolution of chiral esters is to start from
racemates bearing a chiral alcohol moiety. Typically, an acetyl moiety serves as acyl
group that is cleaved enantioselectively in the enzymatic resolution step. Compared
to the corresponding reverse reaction, the enantioselective acetylation using, for
example, vinyl acetate as acylating agent, this type of enzymatic ester hydrolysis is less
widely applied. Among other reasons this might be due to the difficulty in extending a
kinetic resolution based on ester hydrolysis under alcohol formation towards a
dynamic kinetic resolution process since alcohols can be more easily racemized than
their O-acylated counterparts. Nevertheless, enzymatic resolution of O-acylated
esters with a chiral alcohol moiety is of great importance in the synthesis of chiral
alcohols. Major advantages of this process technology are the high enantioselectivity
and activity of many enzymes for these resolutions and the robustness and scalability
of such processes. Selected examples for some typical biotransformations in this
field are given here. The substrate spectrum consists of esters including those with
heteroatom functionalized chiral alcohol moieties, with so-called remote stereogenic
centers in the alcohol moiety as well as esters with axial chirality in the alcohol moiety.
Scheme 8.50 gives a graphical overview about the substrate spectrum.
R1
* OH
R2
secondary alcohols R3
R 2 * OH
R1
R1
tertiary alcohols
* OH
R2
primary alcohols
enzymatic resolution of
racemic carboxylates
with a chiral alcohol moiety
R1
R2 R H
•
H OH
R´ R´
HO
alcohols with a special applications
remote stereogenic center
Scheme 8.50 Overview of resolutions of esters with a chiral alcohol moiety via enzymatic
hydrolysis (selected examples).
8.2.3.1 Resolution of Esters with a Chiral Alcohol Moiety (Non-heteroatom

Functionalized)
The hydrolysis of esters bearing a chiral alcohol moiety without any heteroatom
functionalization typically serves as model reaction when initially evaluating hydro-
lases. These esters have often been used with esterases [3, 82], whereas lipases are
more frequently used for the reverse acylation reaction (also giving an option to
extend this kinetic resolution towards a dynamic kinetic resolution). The most
popular and widely used esterase in organic synthesis is the so-called pig liver
esterase (PLE) [162]. Recent work has been performed by the Bornscheuer group as
286
well as the Pichler and Schwab groups together with their industrial partners to get
access to the isoenzymes in recombinant form [102, 163, 164]. Notably, a recent
comparison of the synthetic properties of different isoenzymes in recombinant form
revealed significant differences in terms of enantioselectivity [164]. When using
racemic (1-phenyl)ethyl acetate (rac-74) as a substrate enantioselectivities, biotrans-
formations with the five recombinant isoenzymes led to the hydrolyzed product with
E-values in a broad range of 2–94, thus indicating the importance of using iso-
enzymes in purified form. Scheme 8.51 shows a selected example leading to an
enantioselectivity of E ¼ 94.
O O
isoenzyme
O CH 3 PLE 5 OH O CH3
CH3 45% conversion CH 3 CH 3

E = 94
r ac- 74 (R)- 75 (S)- 74
95% ee 79% ee
Scheme 8.51 Resolution of (1-phenyl)ethyl acetate using isoenzyme PLE 5.
Furthermore, a chiral tertiary alcohol bearing no heteroatom functionalization has

also been prepared successfully by the Bornscheuer and Hult groups [165]. When
using a mutated esterase from Bacillus subtilis (mutant Gly105Ala) hydrolysis of
racemic 3-phenylbut-1-yn-3-yl acetate proceeds enantioselectivity (E ¼ 56) when
carrying out the resolution in the presence of 20 vol.% of DMSO. In the absence
of this cosolvent a decreased enantioselectivity (E ¼ 28) was observed.
A further successful application of esterases in the enantioselective synthesis of
non-heteroatom functionalized alcohols bearing more than one stereogenic center
has been reported by Xu et al., who resolved racemic O-acylated menthol [166].
Bacillus subtilis esterase showed a high enantioselectivity with E > 200 for the
resolution of racemic menthyl acetate, yielding L-menthol with 98% e.e. at a
conversion of 49%.
8.2.3.2 Resolution of Esters with a Heteroatom Functionalized Chiral Alcohol Moiety

An enzymatic resolution of 6-acetoxybuspirone, rac-76, has been reported by
Bristol-Myers Squibb researchers [167, 168]. An amino acid acylase from
Aspergillus melleus (which typically accepts N-acetylated amino acids) turned out
to be a very efficient enzyme for this ester hydrolysis, leading to the desired (S)-6-
hydroxybuspirone, (S)-77, with 95% e.e. at a conversion of 48%, which corre-
sponds to an enantioselectivity of E ¼ 121 (Scheme 8.52). Since both enantiomers
are of interest, attempts to synthesize the (R)-enantiomer in high enantiomeric
excess have also been made. When stopping the biotransformation at a conver-
sion of 53%, the desired remaining (R)-ester (R)-76 has been obtained with
98% e.e.
N
O N N
N
N
N HO O
(S)-77
O N N acylase from 95% ee
N A. melleus
N
buffer, pH 6, N
toluene O N N
O O
48% conversion N
O
E = 121 N
H 3C
r ac-76
O O
O
H 3C
(R)- 76
Scheme 8.52 Enzymatic resolution of 6-acetoxybuspirone.
An example of an efficient resolution of an ester bearing an alcohol moiety with two

stereogenic centers is the enantioselective lipase-catalyzed hydrolysis of the racemic
O-acetylated cis-azetidinone rac-cis-78 (Scheme 8.53) [167, 169]. The resolution
proceeds with a conversion of 52%, leading to the desired (i.e., remaining)
(3R,4S)-enantiomer (3R,4S)-78 in a reaction yield of 48% and with an excellent
enantiomeric excess of >99.5%. Finally, downstream-processing gave the isolated
product (3R,4S)-79 in a yield of 56% and with an enantiomeric excess of 99.5%. The
corresponding (3R,4S)-alcohol (3R,4S)-79, which was subsequently obtained via
classic chemical hydrolysis at pH 9.4, is a valuable intermediate in the synthesis
of the paclitaxel side-chain.
H 3C O
lipase from
O Ph Pseudomonas sp. HO Ph
NH 52% conversion NH
O O
r ac-cis- 78 (3S,4R)-79
H3 C O
NaHCO3
O Ph (pH 9.4) HO Ph
NH CH 3OH NH
O H 2O O
(3R,4S)- 78 (3R,4S)-79
48% reaction yield
>99.5% ee
after isolation: 45% yield, 99% ee
Scheme 8.53 Enzymatic resolution of O-acetylated cis-azetidinone 78.

288
The related 4-tert-butyl-substituted (3R,4S)-3-hydroxyazetidinone, an intermediate

in the semi-synthesis of taxane, was synthesized in an analogous way. Also for this
resolution, high reaction yields (48–49%) and enantiomeric excesses of >99% have
been obtained for the desired O-acylated (3R,4S)-enantiomer in the presence of
lipases from Pseudomonas sp. and Pseudomonas cepacia [170].
Furthermore, lipase-catalyzed hydrolysis has been used in a resolution for the
synthesis of (S)-O-acetyl 5-hydroxyhexanenitrile [(S)-80] [167, 171]. This chiral O-
acetyl alcohol is an intermediate for the synthesis of an anti-Alzheimer drug. When
starting from the racemic O-acetylated derivative rac-80, enantioselective hydrolysis
using a lipase from Candida antarctica furnished the desired remaining (S)-
enantiomer (S)-80 in a reaction yield of 42% and with an excellent enantiomeric
excess of >99% (Scheme 8.54).
O O
lipase from
O CH3 Candidaant arctica OH O CH 3
CN CN + CN
H 3C H2 O H 3C H 3C
rac-80 (R)-81 (S)-80
42% reaction yield
>99% ee
Scheme 8.54 Enzymatic resolution of O-acetyl 5-hydroxyhexanenitrile.
A lipase-catalyzed process has also been successfully developed for the resolution
of racemic trans-4-phenyl-3-buten-2-yl acetate [172]. In addition, various examples of
the resolution of racemic esters leading enantioselectively to (functionalized) cyclic
alcohols have been developed by several groups [173–175].
This process technology of enzymatic resolution via hydrolysis of racemic esters
also offers attractive access to chiral tertiary alcohols. Such compounds bearing a
quaternary stereogenic center are still challenging molecules for enzymatic resolu-
tion processes in general. At the same time a range of efficient transformations have
already been reported and selected examples are given in the following [65].
Resolution of quinuclidine ester rac-82 has been successfully accomplished by
Coope and Main by means of a porcine liver esterase, leading to the corresponding
alcohol (R)-83 in 36% yield and with 97% e.e. (Scheme 8.55) [176]. A high
enantioselectivity of E > 100 has been reported by the Bornscheuer group for the
n -Pr
pig liver
O esterase OH
O
buffer, MeOH
N N
rac -82 (R )- 83
36% yield
97% ee
Scheme 8.55 Enzymatic resolution of a quinuclidine ester.

resolution of racemic 1,1,1-trifluoro-2-phenyl-3-yn-2yl acetate [177]. The suitable
biocatalyst was found by metagenome screening.
This type of resolution has also been applied by the Hanefeld group for the
enantioselective synthesis of chiral O-acylated cyanohydrins bearing a quaternary
stereogenic center [178]. When starting from the racemic a,a-disubstituted
cyanohydrin acetate rac-84, hydrolytic resolution with subtilisin A protease pro-
ceeds with an enantioselectivity of E ¼ 58 to furnish the desired O-acylated
cyanohydrin (R)-84 as the remaining enantiomer with 90% e.e. at a conversion
of 50% (Scheme 8.56). Notably, the opposite enantiomer is preferably hydrolyzed
when using a lipase from Candida rugosa, albeit the enantioselectivity was
somewhat lower in this case.
protease
AcO CN subtilisin A HO CN AcO CN
CH 3 CH3 CH 3
E = 58
H3 CO H3 CO H3 CO
50% conversion
r ac-84 (S)-85 (R)-84
90% ee
Scheme 8.56 Enzymatic resolution of an a,a-disubstituted cyanohydrin acetate.
The enantioselective synthesis of chiral tertiary alcohols bearing a trifluoro-

methyl group has been demonstrated by the Bornscheuer group using racemic
4,4,4-trifluoro-3-phenylbut-1-yn-3-yl acetate as a substrate [179]. Mutants of an
esterase from Bacillus subtilis have been found that show the opposite
enantiopreference.
8.2.3.3 Resolution of Esters with a Remote Stereogenic Center at the Alcohol Moiety
The recognition of remote chiral centers in esters with a stereogenic center at
the alcohol moiety has also been studied. Liu et al. have developed a synthesis of
lasofoxifene, which represents a potent and selective estrogen receptor modula-
tor [180]. The Pfizer researchers found that in particular a cholesterol esterase
from porcine pancreas is capable of this type of resolution. Although in substrate
cis-rac-86 the functional group for enzymatic hydrolysis (ester group) is separated
from the stereogenic center by an aromatic moiety, enzymatic resolution proceeds
with a high enantioselectivity, as indicated by the E value of 60. The desired
product lasofoxifene (cis-87) is obtained at 35% conversion and enantiomeric
excess of 96% (Scheme 8.57).
Very recently, resolution of an O-acylated derivative of racemic monastrol, a
current lead structure in anticancer research bearing a remote stereogenic center,
enabled the first enantioselective biocatalytic synthesis of (S)-monastrol [(S)-89]
[181]. This enantiomer displays a 15 times higher activity. Whereas attempts for a
direct hydrolysis of racemic monastrol have not been successful, the formation
of racemic O-butanoyl monastrol followed by enantioselective hydrolysis furnished
290
N N N
O O O
cholesterol
esterase
+
O 35% conversion O
Me O E=60 HO Me O
cis-rac-86 cis-87 cis-86
96% ee 51% ee
Scheme 8.57 Enzymatic resolution for the synthesis of lasofoxifene.
O-butanoyl (S)-monastrol, (S)-88, in 31% yield and with 97% e.e. Subsequent
enzymatic cleavage of the O-butanoyl moiety then gave the desired (S)-89 with
96% e.e. [181]. Scheme 8.58 shows the overall synthesis of (S)-monastrol, (S)-89.
OH
O NH
O n-Pr N S
H
lipase from
O (R)-89
O C. antarctica B
(CAL-B)
O NH +
water-CH 2Cl2 O n-Pr OH
N S 59% conversion
H O lipase from
O O
rac-88 C. rugosa
O NH O NH
buffer-CH 2Cl2
N S >95% conversion N S
H H
(S)-88 (S)-89
31% yield 98% yield
97% ee 96% ee
Scheme 8.58 Enzymatic resolution of O-butanoyl (S)-monastrol.
A further interesting example is the hydrolase-catalyzed enantioselective synthesis

of (R)-a-tocopherol, which is of interest as vitamin and antioxidant [182, 183]. The
corresponding acetate and various oxalates turned out to be suitable esters.
Scheme 8.59 shows a representative example. In the presence of a lipase from
Candida cylindracea, the stereogenic center far from the reaction site in the a-tocoph-
erol oxalyl amide rac-90 is still recognized and the ester enantioselectively hydrolyzed,
leading to the remaining ester (R)-90 as a precursor for the natural enantiomer of
a-tocopherol in 41% yield and with 80% e.e.
CH3
HO
R
H3C O CH3
CH3
O CH3 (S)-91
lipase from 48% yield,55% ee
O
H2N C.cylindracea
O CH3
H3C O H2O, IPE
R
CH3 O CH3
rac-90 O
H2N
O CH3
H3C O R
CH3
CH3 CH3 CH3
R= (R)-90
CH3 41% yield
80%ee
Scheme 8.59 Enzymatic enantioselective step towards the synthesis of (R)-a-tocopherol.
8.2.3.4 Resolution of Esters with Axial Chirality at the Alcohol Moiety

The suitability of enzymes to differentiate between the two enantiomers of an axial
chiral racemate by an enantioselective hydrolysis has been described in Section
8.2.2.9 for the resolution of esters bearing a chiral acid moiety. Such a type of
resolution has also been applied for corresponding racemates with an axial chirality
bearing a chiral alcohol moiety. For example, the resolution of racemic allenic esters
of type rac-92, to afford the corresponding primary allenic alcohols, has been reported
by the Cipiciani group [184]. An isopropanol-treated commercial lipase from Candida
rugosa turned out to be suitable, leading to the desired primary allenic alcohol (R)-93
with >99% e.e. at a conversion of 36%, which corresponds to a high enantioselectivity
of E ¼ 100 for this resolution (Scheme 8.60). Under optimized conditions this
resolution was carried out at a low temperature of 4 C in a water–n-hexane reaction
medium.
lipase from
O O
H H C. rugosa H3C H H H
• • •
H3C O CH3 H OH H3C O CH3
H3C CH3 water, pH 7.2, H3C CH3 H3C CH3
n-hexane
36% conversion
rac-92 (R)-93 (S)-92
E = 100 >99% ee 57% ee
Scheme 8.60 Enzymatic resolution of an allenic ester bearing a chiral alcohol moiety.
8.3
Enantioselective Hydrolysis of Prochiral and meso-Carboxylates (Desymmetrization)
8.3.1
Overview
The desymmetrization of prochiral and meso-carboxylates via enantioselective hydro-

lysis is a powerful tool for organic chemists to access chiral carboxylic acids and
alcohols. Compared to other enantioselective hydrolysis, this desymmetrization has
the advantage of theoretically yielding 100% of the desired chiral compound, bearing
the chiral information either in the alcoholic or in the acidic residue of the diesters;
hydroxy-esters or carboxy-esters are the products described in this chapter
(Scheme 8.61). For the enantioselective formation of carboxylates from prochiral
and meso-diols see Section 8.7.
R1 R2
R(O)CO OH
monoacylated diols
CO2 H CO2 H
R1 R2 or
RO2 C CO2 H CO2 R CO2 R
acyclic monoesters cyclic monoesters
products from enzymatic

desymmetrization of
prochiral and meso-carboxylates
R R2 OH OH
or
OC(O)R OC(O)R
3
R1 R
cyclic monoacylated diols
biarylic systems
(axial chirality)
Scheme 8.61 Products from enzymatic desymmetrization of prochiral and meso-carboxylates.
This section gives an overview about more recent and selected examples of
enzymatic desymmetrization of prochiral and meso-carboxylates via enantioselective
hydrolysis. For further reading a selection of excellent publications is available
[3, 167, 185–189].
8.3 Enantioselective Hydrolysis of Prochiral and meso-Carboxylates (Desymmetrization) j293
8.3.2
Hydrolysis of Prochiral Carboxylates
Numerous optically active compounds can be obtained by the enzymatic desymme-

trization of prochiral carboxylates. Common substrates are a,a-disubstituted mal-
onates, glutarates, or larger dicarboxylates and examples of sulfinyl dicarboxylates or
phosphine oxides are described in the literature as well [185]. In the pioneering work
of Schneider et al., prochiral malonates 94 were hydrolyzed enzymatically for the first
time [190]. Good yields and enantioselectivities could be achieved by applying pig
liver esterase (PLE) as catalyst (Scheme 8.62).
CO2R CO2H R, R2 = alkyl

R1 70 U PLE R1
R1 = alkyl, phenyl
R2 R2
CO2R buffer pH8, rt, CO2R
94 95
10 - 50 mM substrate conc. ee = up to 86%
Scheme 8.62 First enzymatic desymmetrization of prochiral malonates.
In general, prochiral malonates are easily accessible substrates and yield useful
chiral intermediates for subsequent synthetic applications. Since the pioneering
work of Schneider et al., the substrate concentrations have been enhanced consid-
erably, making many transformations interesting for industrial applications; scores
of malonate-derivatives have been transformed into mono-acid-mono-carboxylates
with excellent yields and enantioselectivities, in most cases with PLE
(Scheme 8.63) [185, 191–194].
R1 R2 PLE R1 R2 R1 R2
or
MeO2C CO2Me MeO2C CO2H HO2C CO2Me
96 (R)-97 (S)-97
Scheme 8.63
Selected examples of dimethyl malonates, which usually show better enantio-

selectivities than their respective diethyl-analogs, are depicted here to demonstrate
the scope of possible malonate derivatives (Scheme 8.63; Table 8.1) [192, 195, 196].
The drawbacks of any enzyme originating from mammals, like PLE, make their
application, especially for the pharmaceutical industry, unattractive [197, 198].
Notably, Bornscheuer et al. succeeded in cloning and recombinant expression of
functional PLE in Pichia pastoris [199, 200] and in Escherichia coli [201], leading to
recombinant PLE (c-PLE now PLE1). Further PLE isoenzymes (PLE2 to PLE6) with
increased enantioselectivities or even reversed enantiopreferences were identified
and expressed in E. coli [202, 203]. Recently, Schwab et al. cloned and expressed
another isoenzyme of PLE in Pichia pastoris named APLE (alternative PLE) [204]. This
progress gives access to pure PLE consisting of only one isoenzyme in reproducible
294
Table 8.1 PLE-catalyzed desymmetrizations of malonates (according to Scheme 8.63).
Entry R1 R2 Yield (%) E.e. (%) Configuration
1 [195] CH3 CH2OH 37 6 (S)

2 [195] CH3 CH2OCH3 86 21 (S)
3 [195] CH3 CH2O(t-Bu)(CH3)2Si 49 95 (R)
4 [195] CH3 PhCH2 90 67 (R)
5 [195] CH3 CH2O(t-Bu) 90 96 (R)
6 [196] CH3 2-Nitrophenoxy 80 77 (R)
7 [196] CH3 5-Fluoro-2-nitrophenoxy 84 69 (R)
8 [192] CH3(CH2)5 CH3 85 67 (R)
9 [192] CH3CHCH3 CH3 82 42 (R)
10 [192] CH3 HCCCH2 85 63 (S)
11 [192] Cy CH3 94 >98 (R)
12 [192] CyCH2 C2H5 84 >98 (R)
13 [192] Ph CH3 90 >98 (R)
quality from microbial origin and in addition facilitates the development of further
tailor made enzymes by mutagenesis or directed evolution [205].
Notably, cosolvent effects have a tremendous influence on conversion rates and
enantioselectivities in many cases, as can be seen in the conversion of the phenyl-
methyl-substituted malonate 98 (Scheme 8.64) [193, 206–211].
240 mg PLE
CO2Et pH 7.0, 30°C, 25 h CO2Et
H3C H3C
Ph Ph
CO2Et buffer/i-PrOH/t-BuOH CO2H
(8:1:1)
98 (R)-99
200 mM substrate conc. yield= 93%(11.2 g),
(50 g/l), conversion >95%
300 ml scale ee = 96%
Scheme 8.64 Improved conversion rates and enantioselectivities by cosolvent effects.
The original enantioselectivity (80–85% e.e.) and conversion (>90%) could be

raised to 96% e.e. and >95%, respectively, by adding a mixture of i-PrOH and t-BuOH
to the buffer. These conditions allow high substrate concentrations (e.g., 200 mM)
at a multi-gram scale.
Likewise, the addition of different ionic liquids as additives to the reaction media
was described in enzyme-catalyzed syntheses. Pioneering work in this field was
performed by Erbeldinger et al., who observed an increased enzyme stability in a two-
phase system ([BMIM]PF6 : water ¼ 95 : 5) [212]. In 2005 Bolm and Drauz et al.
demonstrated an enhanced enzyme activity and enantioselectivity by the addition of
isopropanol and catalytic amounts of ionic liquids (<1%, Ammoeng 100) as additives
to the reaction media in the desymmetrization of prochiral malonates [213].
As a representative example for the hydrolytic desymmetrization of glutarates, the
transformation of diethyl 3-hydroxyglutarate 100 into the corresponding mono-acid
101 in the first step of the synthesis of (þ)-peloruside A, a marine natural product, is
described in Scheme 8.65 [214].
O OH O CAL-B(354 U/mmol) O OH O
EtO OEt EtO OH

buffer pH 7, rt, 45 min
100 101
714 mM substrateconc. yield = 95%,
(144 g/l), ee = 90%
280 ml scale
Scheme 8.65 First step of the synthesis of ( þ )-peloruside A.
The reaction was performed on a multi-gram scale with immobilized lipase from
Candida antarctica (CAL-B, NovozymÒ 435), which could be easily used several times.
Further examples of prochiral glutarates as substrates in enzymatic desymmetriza-
tions are described in the literature [185, 215, 216].
The facile and highly enantioselective synthesis of axially chiral, tetra-ortho-
substituted biphenyl-derivatives was developed by Matsumoto et al. [217]. The
described method gives access to both corresponding enantiomers in high yields
by suitable choice of enzyme (Scheme 8.66).
While the (S)-enantiomer was attained with porcine pancreas lipase (PPL) and
i-Pr2O as cosolvent, the best result for the corresponding (R)-enantiomer was
OMe
O
MeO
AcO OAc
102
PPL (35 wt%) ROL (50 wt%)

0.1 M phosphate buffer (pH 7.0) 0.1 M phosphate buffer (pH 7.0)
i-Pr2O heptane
35°C, 20 h 35°C, 48 h
OMe OMe
O O
MeO MeO
AcO OH AcO OH
(S)-(+)-103 (R)-(-)-103
quant., ee = >99% 98%, ee = >99%
Scheme 8.66 Enantioselective synthesis of axially chiral, tetra-ortho-substituted biphenyl-

derivatives.
296
O O O O O O
S α -chymotrypsin S
H3CO OCH3 H3CO OH
104 buffer, pH 7.5, 105
NaOH 63% yield
92% ee
Scheme 8.67 Enzymatic enantioselective synthesis of a chiral sulfoxide.
obtained with a lipase from Rhizopus oryzae (ROL) in heptane. Both reactions were
performed in phosphate buffer at 35 C.
The concept of a desymmetrization reaction by means of enzymatic hydrolytic
reactions has also been applied successfully by Mikolajczyk et al. towards the
enantioselective synthesis of stereogenic heteroatom centers, namely, chiral sulf-
oxides, for example, of type (R)-105 (Scheme 8.67) [218]. When methoxycarbonyl-
methyl carboxymethyl sulfoxide (104) was used as substrate in combination with
a-chymotrypsin at pH 7.5 the desired chiral sulfoxide (R)-105 was obtained in 63%
yield and with an enantioselectivity of 92% e.e. In addition, PLE turned out to be a
suitable biocatalyst for this type of biotransformation, with enantioselectivities of up
to 79% e.e.
8.3.3
Hydrolysis of meso-Carboxylates
Enzymatic hydrolysis of meso-carboxylates complements existing methods to gain

access to enantiopure monoesters that can be used as versatile chiral building blocks.
The desymmetrization of a meso-compound affords in principle two possible
enantiomers, both of which are often synthesized selectively. A common approach
is to hydrolyze the cis-diester with a selective enzyme to afford one enantiomer and
the other enantiomer results from mono-acylation of the corresponding diol with the
same enzyme. With some exceptions [219–222] this strategy yields the mirror
images [223, 224]. In this way both cis-enantiomers of the pyrroline-derivative 107
could be accessed enantioselectively with high yields (Scheme 8.68) [223]. Hydrolysis
of the meso-diester 106 yielded the mono-ester (2S,5R)-107. Esterification of the
corresponding meso-diol with lipoprotein lipase from Pseudomonas species led to the
monoester (2R,5S)-107 on a 30 mmol scale. These results are in line with the (R)-
specific tendency of the enzyme as described in the literature [224, 225]. In addition,
analogous desymmetrizations have been reported [167, 226–229].
37°C, 15 h,
1.95 ml buffer pH 7,
AcO OAc 2.5% i-PrOH HO OAc
N N
Boc 10 mg lipoprotein lipase Boc
from Pseudomonas species
meso-106 (2S,5R)-107
yield = 76%, ee = >98%
Scheme 8.68 Hydrolysis of the pyrroline-derivative meso-106.

Scheme 8.69 depicts an example given by Chênevert et al. in which both
enantiomers were available with high yields and selectivities by applying two
different enzymes with opposing selectivities [222]. The meso-diacetate 108 is
transformed into (1S,3R)-109 with CAL-B, whereas (1R,3S)-109 could be obtained
using BCL. The reactions took place in phosphate buffer and hexanes to increase the
solubility of the diacetate (Scheme 8.69).
AcO OH AcO OAc HO OAc

pH 7.2, rt, 39 h pH 7.2, rt, 30 h
BCL CAL-B
(1S,3R)-109 108 (1R,3S)-109
isolated yield = 99%, 23 mM isolated yield = 94%,
ee = >98% ee = >98%
Scheme 8.69 Access to both enantiomers of the half-ester 109.
The meso-type diacetate meso-110, bearing three stereogenic centers, has been
enantioselectively converted into the monoacetate by the Danishefsky group
(Scheme 8.70) [226, 227], and by Bristol-Myers Squibb researchers [167, 230]. In
the presence of a lipase from Pseudomonas cepacia (lipase PS-30) a product-related
conversion of 85% and an enantioselectivity of 98% e.e. was obtained for the desired
monoacetate 111, which serves as an intermediate for entecavir, an approved drug for
treatment of hepatitis B virus infection [167, 230].
OBn lipase from OBn

P. cepacia
AcO OAc HO OAc
buffer/toluene
(90:10)
meso-110 85% product- 111
related conversion 98% ee
Scheme 8.70 Enzymatic desymmetrization of a meso-diacetate with three stereogenic centers.
Another example, in which both enantiomers of a mono-ester could be afforded by

enzymatic hydrolysis of the corresponding meso-diester, was published in 2009 by
Goswami et al. (Scheme 8.71) [231].
CO2Me CO2H
pH 8.5, 40°C, 27 h
CO2Me 150 g CAL-B CO2Me
cis-112 (1S,2R)-113
8.53 mol, yield = 99.8%,
568.7 mM ee = >99%
Scheme 8.71 Enzymatic hydrolysis of meso-diester 112.

298
Though the synthesis of one enantiomer, (1R,2S)-113, had been described

earlier [231–239], a broad screening was necessary to find an efficient catalyst for
the other enantiomer (1S,2R)-113, which is a key intermediate for the synthesis of a
promising agent in the field of chemokine receptors [231, 240, 241]. Successful
screening and scale up resulted in a kg-preparation of (1S,2R)-113 with excellent yield
and enantioselectivity using CAL-B (Novozym 435) as catalyst.
An alternative enantioselective route to the synthesis of the pharmacologically
relevant oseltamivir phosphate (Tamiflu) (117) by Zutter et al. starts from the cheap
2,6-dimethoxyphenol (114) and involves desymmetrization of the all-cis meso-diester
115 as a key step (Scheme 8.72) [242]. The reaction took place in TRIS buffer pH 8
without addition of organic solvents, resulting in high yield and enantioselectivity.
Notably, the enzyme tolerated a 10% substrate concentration.
O OH OH
HO 5 steps O CO2Et pH 8.0, 35°C, 46 h O CO2H
PLE
O HO HO
CO2Et CO2Et
114 meso-115 116
yield (crude product) = 98%,
ee = 96-98%
4 steps O CO2Et
AcHN
NH2 H3PO4
117
Oseltamivir phosphate
Scheme 8.72 An alternative enantioselective route to oseltamivir phosphate (Tamiflu) (117).
8.4
Other Stereoselective and Non-stereoselective Hydrolysis of Acyclic Carboxylates
Most enzymatic hydrolytic reactions using esters as a substrate for either a resolution
or desymmetrization process belong to the class of enantioselective biotransforma-
tions. Nonetheless, some other types of stereoselective transformations have been
reported as well, in particular regioselective and diastereoselective hydrolysis reac-
tions. In addition, non-stereoselective hydrolytic reactions starting from chiral
carboxylates have also been reported. An overview of such types of transformations
is given here.
A classic and also typical example for a regioselective hydrolysis is the transfor-
mation of racemic dimethyl malate (rac-118) into its monoester rac-119.
8.5 Enantioselective Hydrolysis of Cyclic Esters (Lactones) and Derivatives Thereof j299
This transformation, which is difficult to carry out by means of chemical
hydrolytic reactions, can be conducted efficiently when using a pig liver esterase
(PLE) as a catalyst (Scheme 8.73) [243]. According to the consumption of base
required to neutralize the acid formed in the hydrolytic process, a complete
reaction can be assumed. Notably, this process is highly regioselective but not
enantioselective.
CO2CH3 PLE CO2H

H3CO2C H3CO2C
OH buffer, pH8, OH
NaOH
rac-118 quantitative conversion rac-119
Scheme 8.73 Regioselective and non-enantioselective enzymatic hydrolysis of racemic dimethyl

malate.
Furthermore, enzymatic hydrolytic reactions have been widely used for the non-
stereoselective cleavage of ester bonds under mild conditions [244]. This method
is a particularly valuable strategy when other functional groups, which are
sensitive to classic chemical hydrolyses conditions, are present in the substrate
molecule.
A further interesting application is the hydrolysis of diastereotopic ester groups in
non-chiral alkenes of, for example, type 120 (Scheme 8.74). The Otto group found in
pioneering work that PLE can differentiate (in a diastereoselective fashion) between
the two ester moieties in such (E/Z)-diastereotopic esters [245]. For example, using
PLE in combination with 120 as a substrate furnished exclusively the diastereomeric
ester (Z)-121 with an excellent diastereoselectivity of >99% de and a quantitative
conversion.
PLE
CO2Et CO2Et
H CO2Et buffer, pH 7 H CO2H
100% conversion
120 (Z)-121
>99% de
Scheme 8.74 Enzymatic diastereoselective hydrolysis of diastereotopic ester groups in a non-

chiral alkene.
8.5
Enantioselective Hydrolysis of Cyclic Esters (Lactones) and Derivatives Thereof
In addition to the classic type of ester hydrolysis starting from acyclic esters, several
enzymatic resolutions based on the use of cyclic esters (lactones) and derivatives
thereof have been developed. In particular, substituted lactones of type rac-122,
oxazolin-5-ones (azlactones) and thiazolin-5-ones turned out to be suitable
substrates.
300
8.5.1
Resolution of Lactones
A prominent synthetic example in the field of enzymatic enantioselective lactone

hydrolysis is the resolution of racemic pantolactone to form the industrially valuable
D-pantothenic acid ((R)-pantothenic acid), (R)-123 [246, 247]. An interesting finding
in a screening study by the Shimizu group was that fungal strains show a preference
for hydrolysis of the D-enantiomer, whereas bacterial strains have a tendency to
preferably hydrolyze the L-enantiomer of racemic pantolactone. When using a fungal
lactonohydrolase from Fusarium oxysporum as a biocatalyst, resolution proceeds to
give the desired (R)-pantothenic acid, (R)-123, with 96% e.e. at a conversion of 46%.
Notably, this process shown in Scheme 8.75, runs at a high substrate input of 700 g
l1 [246, 248]. Furthermore, an efficient immobilized biocatalyst has also been
developed by the Shimizu group [246, 249]. By means of F. oxysporum cells entrapped
in calcium alginate gels the synthesis of (R)-pantothenic acid has been carried out
over 180 resolution cycles, with an impressive remaining activity of 60% after these
numerous recycling cycles. The process runs at a substrate input of 350 g l1, leading
to high conversion and enantiomeric excess (91–95% e.e.). Notably, this enzymatic
resolution has been in commercial operation since 1999 [246].
lactonohydrolase
HO O from HO O
H3C CH3O
F. oxysporum
H3C HO H3C
O OH O
H3C H2O, pH 7.0, H3C
OH
46% conversion
rac-122 (R)-123 (S)-122
substrate input: 96% ee
700 g/l
Scheme 8.75 Enzymatic synthesis of D-pantothenic acid [(R)-123] via enantioselective hydrolysis of
racemic pantolactone.
In addition, the porcine pancreas lipase turned out to be suitable for the
resolution of c- and d-lactones under enantioselective formation to give the
corresponding hydroxy acids [250]. Furthermore, the resolution of racemic a-ami-
nobutyrolactones representing precursors for a-amino acids bearing one or two
stereogenic centers has been reported by the Gutman and Guibe-Jampel
groups [251]. When using a-aminobutyrolactones of type rac-124, bearing one
stereogenic center as substrate, resolution proceeds in dependency on the type of
acyl group to furnish the remaining substrate 124 with 62% e.e. at 50% conversion
(Scheme 8.76). In this case the hydrolyzed product (125) is an enantiomerically
enriched N-acyl protected homoserine, which has been re-converted into the lactone
124 with an enantiomeric excess of 71%. Furthermore, related diastereomerically
pure or enriched racemic cis-lactones with a substituent in the c-position have been
used as substrates and gave in one example the remaining substrate with 95% e.e. at
50% conversion.
8.5 Enantioselective Hydrolysis of Cyclic Esters (Lactones) and Derivatives Thereof j301
O
O lipase from O
HN HN
Porcine pancreas HN R
R R
O H2O O HO CO2Na
O O
r ac-124 124 125
a: R = Ph: 40% conversion R = Ph: 43% ee,
b: R = OMe: 50% conversion R = OMe: 62% ee H
O
HN
R
O O
124
R = Ph: 59% ee,
R = OMe: 71% ee
Scheme 8.76 Enzymatic resolution of a racemic a-aminobutyrolactone.
8.5.2
Resolution of Azlactones
As a further cyclic substrate, racemic azlactones (oxazolin-5-ones) of type rac-126

turned out to be very interesting starting materials. This is, in particular, due to their
tendency to racemize easily even under neutral or weakly basic pH conditions, thus
allowing a combination with an enzymatic hydrolysis reaction towards a dynamic
kinetic resolution (DKR) process. A general limitation of this process technology,
however, is the high tendency of azlactones to undergo a spontaneous hydrolytic ring-
opening reaction forming the racemic acid as a product [252]. Such phenomena have
been the reason for the low enantioselectivities observed when using proteases as a
catalyst due to the comparable reaction rates for both the enzymatic and spontaneous
ring-opening reaction [252, 253]. However, lipases turned out to be more active and
highly suitable biocatalysts for enantioselective ring-opening reactions. When using
a lipase from porcine pancreas the corresponding N-benzoyl L-amino acids L-127 have
been formed with enantiomeric excesses in the range of 20 to >99% e.e. depending
on the type of substituent at the azlactone ring [253–255]. Scheme 8.77 shows selected
examples of this dynamic kinetic resolution process. The opposite D-enantiomers
were formed when using a lipase from Aspergillus sp.
R O lipase from R CO2H
porcine pancreas
N O
HN Ph
H2O
Ph O
rac-126 L-127
a: R = Mes-(CH2)2-, a: 80% ee
b: R = Ph-CH2-, b: >99% ee
c: R = Me2CH-CH2-, c: 87% ee
d: R = Ph-CH2-S-CH2- d: 20% ee
Scheme 8.77 Dynamic kinetic resolution of azlactones (oxazolin-5-ones).

302
8.5.3
Resolution of Thiazolin-5-ones
Furthermore, thiazolin-5-ones have been used as substrates for enantioselective

enzymatic hydrolysis processes. After an early study by Previero and coworkers using
chymotrypsin as a biocatalyst [256], the Sih group reported an enzyme and substrate
screening of this reaction using racemic 4-substituted 2-phenylthiazolin-5-ones rac-
128 as substrate [254]. Scheme 8.78 shows selected synthesis examples of this
enzymatic process using proteases. In the presence of the lipase prozyme 6 and
acetonitrile (10 vol.%) as a cosolvent, resolution of rac-128b affords the L-methionine
precursor (S)-129b in a high yield of 94% and with a high enantiomeric excess of 98%.
R O protease R CO2H
"prozyme 6"
N S
HN Ph
buffer, pH 7.5,
Ph acetonitrile (10%(v/v)) S
rac-128 L-129
a: R = Me2CH-, a: 39% yield, 94% ee
b: R = MeS-(CH2)2-, b: 94% yield, 98% ee
c: R = Me-(CH2)3-, c: 86% yield, 99% ee
d: R = PhCH2- d: 30% yield, 78% ee
Scheme 8.78 Dynamic kinetic resolution of thiazolin-5-ones.
8.6
Enantioselective Formation of Carboxylates via Esterification
8.6.1
Overview
The enzymatic formation of carboxylic acid esters relies on the use of hydrolases in
organic solvents. In addition to its synthetic utility the hydrolase-catalyzed (trans)
esterification has mainly gained impressive popularity in synthetic organic chemistry
as an efficient resolution technique for racemic alcohols and acids. As such the
esterification reaction represents, usually, the additional step required at the end of a
non-selective synthesis to obtain enantiopure products. Relevant examples of this
important application are reviewed in Sections 8.6.2 and 8.6.3. Section 8.6.3 covers the
desymmetrization of either prochiral or meso acids and esters via transesterification.
8.6.2
Resolution of rac-Alcohols
The resolution of racemic alcohols via enzymatic esterification represents one of the
main applications for lipases. The acylation is usually performed in organic solvents,
which can affect both enzyme activity and enantioselectivity, and has been inves-
tigated in several studies [257–259]. The acylating agents used in the esterification are
usually classified as non-activated (e.g., methyl or ethyl acetate), activated (e.g.,
trichloromethyl esters), or irreversible acyl donors (e.g., anhydrides [50], vinyl-,
8.6 Enantioselective Formation of Carboxylates via Esterification j303
isopropenyl esters). In the latter case, the reverse reaction is prevented by the
tautomerization of the leaving enol to the corresponding carbonyl compound [260].
Section 8.2.1 gives a more detailed discussion of the different acylating agents and
their use in the formation of carboxylates.
The following subsections illustrate the resolution of monoalcohols (Scheme 8.79),
followed by the resolution of diols in Section 8.6.2.4.
OH OH
1 2
R R
acyclic
cyclic
secondary alcohols
rac-monoalcohols OH
1
R R2
R3
enzymatic resolution
enzymatic resolution of tertiary alcohols
of primary alcohols
X
OH
R1
R2
X
quaternary stereocenter
OH (hindered substrates)
R
tertiary stereocenter
R1 X
R R2 OH
Ar OH R2
OH OH X = C, N
Ar R R1
2-aryl-substituted 3-aryl-substituted 2,2-dialkyl-substituted aziridinyl- and
substrates substrates substrates cyclopropylmethanol
Scheme 8.79 Overview of the resolution of monoalcohols.

304
8.6.2.1 Enzymatic Resolution of Primary Alcohols
Primary Alcohols Having a Tertiary Stereocenter at the b-Position Chiral primary

alcohols are versatile building blocks, often employed as key intermediates in the
synthesis of biologically relevant compounds [261]. For this reason, their enantiopure
preparation via enzymatic (dynamic) kinetic resolution of racemates has been
extensively studied. The lipase-mediated resolution of primary alcohols suffers from
a lower enantioselectivity than that of secondary alcohols, and high enantiopurity is
often more difficult to attain. One exception is the resolution mediated by Burkhol-
deria cepacia lipase (BCL) [262]. This lipase showed high enantioselectivity towards a
wide range of primary 2-methylalcohols possessing a b-tertiary stereocenter
(Scheme 8.80) [263, 264].
vinyl acetate
R OH R OAc
+ R OH
BCL
O S S
R=
E = 18 E = 105 E = 200 E = 300

O
S S S
E = 144 E = 108 E = 75 E = 18
BnO
E = 172 E >>100 E = 67
E = 150 E = 45
Scheme 8.80 BCL-mediated kinetic resolution of 3-substituted-2-methylpropanols.
3-Aryl-2-methylpropanols gave the most satisfactory results in terms of enantio-

selectivity, with E values often above 100. On moving the aryl substituent closer to the
stereocenter, the enantioselectivity dropped dramatically, as shown in Scheme 8.81
for some 2-aryl-2-methylethanols.
To enhance the enantioselectivity of lipases towards primary alcohols, adoption of
low temperatures has been proposed [265, 266]. Lowering the temperature in the
vinyl acetate
OH OAc OH
R R + R
BCL
S S
R=
E = 2.3 E = 1.2 E = 1.8 E = 9.0
Scheme 8.81 BCL-mediated kinetic resolution of 2-substituted-2-methylethanols.
kinetic resolution of 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal) catalyzed by

Pseudomonas fluorescens lipase (PFL) immobilized on Celite (lipase Amano AK)
resulted in an enhancement of E from 16 at 23 C [267] up to 55 at 40 C
(Scheme 8.82).
PFL
O O O
vinyl butyrate O +
O iPr2 O O C 3H 7 O
OH O OH
E = 55 at -40°C
Scheme 8.82 Kinetic resolution of solketal.
The same strategy has been applied to the resolution of trans- and cis-(3-methyl-3-
phenyl-2-aziridinyl)methanol [268]. Using BCL immobilized on a porous ceramic
support (Toyonite) (PS-C II) at low temperatures, both diastereoisomers of these
interesting classes of aziridine alcohols were resolved with good enantioselectivities
(Scheme 8.83).
H H H
Ph N H Ph N H H3 C N
BCL + OAc
H3 C OH vinyl acetate H 3C OH Ph H
acetone
E = 55 at -40°C
H H
H3C N H N H
BCL H3 C H N
Ph OH
OH +
Ph vinyl acetate Ph OAc
acetone H3 C H
E = 73 at -20°C
Scheme 8.83 Kinetic resolution of (3-methyl-3-phenyl-2-aziridinyl)methanol.

306
Ph BCL (Amano PS) Ph Ph

HO HO + RO
F vinyl ester F F
toluene or MTBE E = 4.3-13
R = CH 3CO, CH 3CH2 CO, ClCH 2CO
HO Ph BCL (Amano PS) HO Ph RO Ph

+
F vinyl ester F F
toluene or MTBE E = 7.5-200
R = CH 3CO, CH 3 CH 2CO, ClCH2 CO
Scheme 8.84 Kinetic resolution of (2-fluoro-2-phenylcyclopropyl)methanol.
Rosen et al. [269] reported the kinetic resolution of cis- and trans-(2-fluoro-2-
phenylcyclopropyl)methanol using Pseudomonas cepacia lipase and various vinyl
esters as donors (Scheme 8.84). Interestingly, whereas the enzymatic acylation gave
low selectivities for trans-alcohols (E ¼ 13), the corresponding cis-diastereoisomers
were obtained with very high optical purity (E > 200).
The substituent effect on enantioselectivity in lipase-catalyzed transesterification
of C2-symmetric trans-2,5-disubstituted pyrrolidines was analyzed by Kawanami
et al. [270]. A significant dependence of the enantioselectivity on the substituents at
the phenyl ring was observed; the 3,5-dimethyl substitution pattern gave the best
results (E ¼ 108) with PFL immobilized in sol–gel (Scheme 8.85).
OH
R1
N
R2
OH
R1 R3
AcO
N 2 vinyl acetate R4
R
+
immobilized PFL
3
AcO R OAc
R4
R1
N
R2
AcO R3
R4
E = 108 (R 1 = R 3 = H; R 2 = R4 = Me)
Scheme 8.85 Kinetic resolution of trans-2,5-disubstituted pyrrolidines.

OH OAc OH
O vinyl acetate O O
O O + O
BCL (Amano PS)
Br Br Br
Br Br Br
E = 36
Scheme 8.86 Kinetic resolution of glycerol derivatives.
In subsequent work [271] the same group analyzed the effect of substitution in the
resolution of protected glycerol derivatives. Under optimized conditions (BCL as
Amano PS and vinyl acetate), the bis(4-bromophenyl)ketal was the best resolved
substrate (Scheme 8.86).
A highly enantioselective resolution was achieved by the group of Kawasaki [272]
for primary alcohol 130, which was used as intermediate for the synthesis of
norsesquiterpene 5,6-dehydrosenedigitalene (Scheme 8.87). In previous work [273],
it was envisaged that primary alcohols might be efficiently resolved by using bulky
R O
HO
O
+ BCL
O R + E = 331
hexane, rt HO
130
R=
CF3
5,6-dehydrosenedigitalene
Scheme 8.87 Application of enantioselective transesterification to the synthesis of 5,6-

dehydrosenedigitalene.
308
acyl donors, or by acyl donors having specific stereoelectronic effects. Indeed, the acyl
group exerts both steric and electronic effects on the acyl transfer process from the
acylated lipase onto the alcohol. Thus, acyl donors exerting specific stereoelectronic
effects should ensure high enantioselectivity [274]. In the resolution of 130 mediated
by BCL, replacing conventional vinyl acetate with vinyl 3-(4-trifluoromethylphenyl)
propanoate increased E from 26 to 331.
Since optically active primary alcohols are often useful building blocks for the
synthesis of biologically relevant compounds, their enzymatic resolution enables
interesting applications, especially in the pharmaceutical field. For example, the
resolution of (R,S)-N-(tert-butoxycarbonyl)-3-hydroxymethylpiperidine with succi-
nic anhydride and BCL has been applied for the preparation of alcohol 131, a key
intermediate for the synthesis of a potent tryptase inhibitor (Scheme 8.88) [275].
The (S)-hemisuccinic ester could be easily separated and subsequently hydrolyzed
to 131. Reiteration of the process allowed 131 to be obtained in 32% yield and 98.9%
e.e. [275, 276].
O
OH OH
OH O
BCL (Amano PS-30) O
N N + N
O O O
O O O O O O
base
OH
O O
131
yield = 32%, ee = 98.9%
Scheme 8.88 Preparation of 131, a chiral synthon for a tryptase inhibitor.
Piperidine derivative 132 was viewed by Gotor and coworkers as a convenient

intermediate for the chemoenzymatic synthesis of (–)-paroxetine, a selective serotonin
re-uptake inhibitor that acts as an antidepressant drug [277]. It was prepared by
enantioselective acylation of the N-substituted trans-4-(4-fluorophenyl)-3-hydroxy-
methylpiperidine catalyzed by lipases [278]. Under optimized experimental conditions,
CAL-AandCAL-Bgavethehighestselectivities(E> 100intoluene)andshowedopposite
stereochemical preference, with CAL-B leaving the desired (3S,4R)-alcohol unchanged
(Scheme 8.89).
F F
O O OH
N N
H H
(-)-paroxetine 132
F F F
CAL-A, vinyl acetate

or O
OH CAL-B, vinyl benzoate OH O R2
toluene +
N N N
R1 R1 R1
trans-rac E = 103-131 (CAL-A)
R1 = Cbz, Boc, Alloc, Poc E = 114-142 (CAL-B)
Scheme 8.89 Lipase-catalyzed synthesis of a (–)-paroxetine precursor.
Using CAL-B and glutaric anhydride as the acyl donor, effective separation of the
alcohol from the monoester could be achieved by simple extraction [279].
Other relevant applications of lipase-mediated esterification in the pharmaceutical
field are the synthesis of lobucavir analog 133 [280] and ribavirin prodrug 134 [281],
both potentially useful for the treatment of hepatitis. In both cases, biocatalysis offers
a more convenient approach compared to the chemical esterification, while ensuring
the required regioselectivity. Compound 133 could be obtained by regioselective
acylation of only one hydroxyl group using either crystalline subtilisin as
ChiroCLECÔ BL (61% yield) or BCL [282] (Scheme 8.90).
In the case of ribavirin prodrug 134, esterification with the oxime ester of Cbz-
alanine in the presence of immobilized CAL-B (Chirazime L-2) resulted in exclusive
acylation of the primary alcohol (Scheme 8.91) [283].
Primary Alcohols Having a Quaternary Stereocenter at the b-Position Primary alcohols

with a quaternary stereocenter at the b-position are quite readily accepted by
enzymes; however, the presence of a methylene unit between the chiral center and
the site of reaction causes in most cases a quite poor stereodifferentiation [284].
310
O O
N N NH
NH (R)-valine-p-nitrophenylester
N N NH 2 BCL (AmanoPS-30 IME) HO N N NH2
HO
or
lobucavir subtilisin (ChiroCLEC TM BL) O

OH O
R = Cbz, Boc
RHN
R = H ( 133 )
Scheme 8.90 Regioselective acylation of lobucavir.
O
H2N
N
N
O
H N
N N HO
Cbz O +
O
OH OH
ribavirin
O
H 2N
CAL-B (Chirazyme L-2) N
N
THF N
Cbz O
N O
H
O
OH OH
O
H2 N
N
N
N
+H N O
3 O
SO3 - O
134 OH OH
Scheme 8.91 Regioselective acylation of ribavirin.

Nevertheless, examples of effective resolution of such substrates are also reported in
the literature.
Resolution of four amino-alcohols of type 135 was studied as an approach to set
enantiomerically pure L-a-vinyl amino acids, used as intermediates for synthetic
peptides [285] due to their high biostability. The best conditions for acylation were
found for Ala- and Phe-derivatives, which could be resolved by Pseudomonas
fluorescens (lipase AK) in wet benzene; E values up to 200 and 25, respectively, could
be obtained (Scheme 8.92).
OH PFL OAc OH
vinyl acetate
+
R NHBz benzene R NHBz R NHBz
135
ee = 86% (R = CH 2Ph)
R = Me, CH 2Ph ee = 98% (R = Me)
Scheme 8.92 Kinetic resolution of a-vinyl amino acids.
2-Substituted oxiranemethanols, a class of compounds in which the quaternary

carbon is part of a three-membered ring, show interesting applications
(Scheme 8.93) [286]. Their resolution by enzymatic acetylation catalyzed by PFL was
investigatedbythegroupofSantaniello asavalidalternative tothe Sharpless asymmetric
epoxidation of allylic alcohols [284]. Transesterification with vinyl acetate was performed
on two oxiranemethanols (136) bearing different side chains and resulted in good to
excellent enantioselectivity. The substrate specificity of Pseudomonas fluorescens lipase
(PFL) was confirmed by the observation that 3-methyl-3-oxetanemethanol 137 was not
accepted as a substrate by this enzyme under identical conditions.
O PFL O O
OH R OH + AcO
R vinyl acetate R
136a, R = CH 2Ph E > 200, R = CH 2 Ph

136b, R = n-C9 H 19 E = 89, R = n -C 9H 19
O
OH
137
Scheme 8.93 Kinetic resolution of oxiranemethanols.
Resolution of epoxy-1,4-butanediol 138 represents a special case of enantio- and

regioselectivity at the same time, due to the presence of two primary alcohols at
different distances from the stereogenic carbon (Scheme 8.94). Acylation with BCL as
312
O vinyl acetate, CHCl 3 O

O +
HO OH BCL HO OH HO OAc
138 E = 9.7
Scheme 8.94 Kinetic resolution of epoxy-1,4-butanediol 138.
Amano PS proved to be highly regioselective, favoring C1 over C4 (97 : 3). Enantio-

selectivity, however, was only moderate (E ¼ 9.7) [287].
Resolution of 2-cyano-2-phenyl-1-hexanol (139) was attempted by Cheong and
coworkers [288]. This primary alcohol was designed as a precursor of SysthaneÒ , a
systemic fungicide inhibiting ergosterol biosynthesis. Using lipases from Candida
rugosa (CRL) and Pseudomonas fluorescens, both enantiomers were obtained with
E ¼ 5.5 and > 200, respectively (Scheme 8.95). The modest e.e. (62%) obtained for the
(R)-acetate with CRL could be increased to up to 99% by further resolution of the
enantio-enriched product.
CN CN
+
OH OAc
L E = 5.5
CN CR
n- hexane, vinyl acetate
OH
139
PF
L
CN CN
+
OAc OH
E > 200
Scheme 8.95 Resolution of 2-cyano-2-phenyl-1-hexanol.
A non-obvious example of chiral primary alcohols is represented by allenic alcohols

showing axial chirality along the 1,2-cumulated diene system (140, Scheme 8.96).
During recent years, the synthetic utility of these molecules as versatile intermediates
has been recognized, as proven by an increasing number of applications [289].
However, their synthesis in optically active form is still limited to the use of
enantiopure precursors. Enzymatic kinetic resolution suffers from poor enantio-
selectivity and a narrow substrate scope [290]. A simple but efficient access to
enantiomerically pure a-allenols via enzyme resolution has been provided recently
by B€ackvalls group (Scheme 8.96) [290]. PPL (porcine pancreatic lipase) in diiso-
O
R1 C PPL, vinyl butyrate R1 C R1 C
OH OH + O C 3 H7
R2 iPr 2O R2 R2
140 E = 9 - 200
HOOC
O O
(S)-(-)-striatisporolide A
(R 1 = pentyl, R 2 = Me)
Scheme 8.96 Kinetic resolution of allenols.
propyl ether using vinyl butyrate demonstrated very high selectivity (E > 200) towards
these substrates. Moreover, a scope study proved that substitution at the 2-position of
the allene is crucial for the enantioselectivity. As an application of this elegant
method, one of the resolved a-allenols was used as precursor for (–)-striatisporolide
A [291], a natural product showing interesting antifungal properties.
Primary Alcohols Having a Remote Stereocenter Primary alcohols bearing a remote

stereocenter are more easily accepted by hydrolases than primary alcohols having the
stereocenter at the b-position. On the other hand, moving the center of reaction away
from the stereocenter usually results in a loss of chiral recognition by the enzyme.
However, a few examples can be found in literature in which substrates possessing
remote stereocenters are efficiently resolved [284].
Compound 141 is the alcoholic derivative of the analgesic etodolac, and presents a
CH2CH2-spacer between the center of reaction and the hindered quaternary stereo-
center. It was resolved (Scheme 8.97) with reasonable enantioselectivity (E ¼ 17) by
O
O
O CRL, MTBE + N
N H OAc
N H OH
vinyl acetate
H OH E = 17
141
O
N COOH
H
Etodolac
Scheme 8.97 Kinetic resolution of an etodolac derivative.

314
using CRL and vinyl acetate as a donor in polar solvents (MTBE, acetonitrile) [292].
Nonpolar solvents, such as n-hexane or cyclohexane, were found to enhance the
reaction rate while causing a drastic decrease of selectivity.
Citalopram is a selective inhibitor of serotonin applied as an antidepressant [277].
Interestingly, it has been demonstrated that all of the inhibitory activity resides in the
(S)-( þ )-enantiomer and, therefore, a selective synthesis of this enantiomer is
needed. An effective strategy is the resolution of the corresponding cyanodiol
142, which has a quaternary stereocenter quite far from the primary alcohol
(Scheme 8.98). After screening different enzymes and conditions, E ¼ 70 was
obtained with CAL-B and vinyl acetate in acetonitrile [293].
The dependence of chiral recognition on the distance between the reaction site and
the stereocenter can be illustrated by the resolution of perillyl alcohol
(Scheme 8.99) [294], a hydroxylated metabolite of d-limonene that shows very
interesting chemopreventive and chemotherapeutic activity against some malignan-
cies [295]. To increase its lipophilicity and hence its biological properties, acylation
with fatty acids was chosen as a promising strategy. Eleven lipases were screened in
the resolution with decanoic acid [296], and only a very modest enantioselectivity was
observed (Eapp ¼ 0.6–3.8).
8.6.2.2 Enzymatic Resolution of Secondary Alcohols

Secondary alcohols are by far the most common substrates in hydrolase-catalyzed
resolution. The reason is twofold: not only are secondary alcohols crucial building
blocks in organic synthesis but they also can be resolved with very high enantio-
selectivity, especially in comparison to their primary and tertiary counterparts [261].
As a further consequence, sec-alcohols are frequently used as model substrates to test
the activity and/or selectivity of enzymes, and also to screen experimental conditions
for enzymatic resolutions.
Enzymatic Resolution of Acyclic Secondary Alcohols 3-Aryloxy-1-nitrooxypropan-2-

ols 144 are expected to show cardiovascular activity similar to that of the family of
nitrovasodilators, including also nitroglycerin (Scheme 8.100). Such compounds
can be prepared in enantiomerically pure form by resolution of the racemic
alcohols using BCL (Amano PS) and PFL (Amano AK). The catalytic activity and
enantioselectivity of the enzymes are affected by the polarity of the solvent; among
all media screened, n-hexane is the best in terms of both reaction rate and
selectivity. A dependence of the enantioselectivity on the substitution pattern of
the aryl group is also observed; as a consequence the E values vary in the range
31–111 [297].
Rivastigmine and related miotine (Scheme 8.101) are known for the treatment of
mild to moderate forms of Alzheimers disease. Since the therapeutic activity is
limited to their (S)-enantiomers, enantioselective syntheses are highly desired. A
chemoenzymatic approach based on a CAL-B mediated stereoselective acylation
proved to be very convenient [298]. The key step is represented by the resolution of
rac-145 using vinyl acetate. The obtained (R)-acetate was subsequently converted into
rivastigmine.
OH OH O O
NC NC NC
CAL-B
OH vinyl acetate OH + OH
N N N
F F F
142
NC
O
F
(S)-(+)-Citalopram
Scheme 8.98 Kinetic resolution as a key step in the synthesis of (S)-( þ )-citalopram.
8.6 Enantioselective Formation of Carboxylates via Esterification
j315
316
O C 9 H19
HO
O
Various lipases
143
E app = 0.6-3.8
Scheme 8.99 Resolution of perillyl alcohol.
OH BCL or OH OAc
O ONO2 PFL O ONO 2 + O ONO2
Ar Ar Ar
144 n-hexane,
vinyl acetate
E = 31-111
Scheme 8.100 Resolution of aryloxynitropropanols.
OH OAc OH
MeO vinyl acetate, MeO MeO
CAL-B, MTBE
+
145
NMe2
N O
rivastigmine
Scheme 8.101 Resolution of an intermediate for rivastigmine.
Theuse of activated estersis acommon strategy to shift theequilibrium of enzymatic

acylations to the desired direction [5]. The most popular irreversible acyl donors are
enol esters such as vinyl or isopropenyl acetate. The simple acetyl moiety is in most
cases removed or replaced at a later stage of the synthesis. However, the introduction
and retention of functionalized acyl moieties as part of the final structure would be
highly desirable from the atom efficiency perspective. Several different vinyl esters
were recently screened in the kinetic resolution of 1-phenylethanol to this aim
(Scheme 8.102) [299]. The substrate tolerance of three lipases (CAL-B, CRL, and BCL)
with respect to these acyl donors was tested. The stereochemical outcome of all
O
O
OH R O OH
O R
+
lipase
Et2 O
a: R = CH3 d: R = g: R = Ph
b: R = t-Bu H
e: R = h: R = N Ot -Bu
c: R =
f: R = Ph O
i: R = S
Ph
Scheme 8.102 Screening of functionalized acyl donors.
reactions was in agreement with Kazlauskas rule, and in each case (R)-1-phenyletha-
nol was the fastest reacting enantiomer. While CRL showed low activity or low
enantioselectivity and BCL showed varying degrees of enantioselectivity, CAL-B
accepted all the acyl donors with high enantioselectivity. In most cases, E > 200 were
obtained. Reactions with vinyl 4-pentenoate, cinnamate, and N-Boc glycinate
were slow.
The resolution of secondary alcohols bearing another functional group plays an
important role in the synthesis of some pharmaceutically valuable compounds [277].
The resolution of halogeno-alcohols [300] and cyanoalcohols [301] through BCL-
catalyzed acylation has been applied to the synthesis of (S)-propranolol, a b-adren-
ergic blocking agent. Scheme 8.103 depicts the general strategy.
CN CN CN
O O O
OH BCL OAc OH
+
vinyl acetate, DCM
Cl O Cl O Cl
O
BCL
OH OAc + OH
vinyl acetate, DCM
Scheme 8.103 Resolution of intermediates in the synthesis of (S)-propranolol.
A further example of resolving bifunctional alcohols applied for the synthesis of a

chiral drug is shown in the synthesis of (R,R)-formoterol, a potent b2-receptor
agonist (Scheme 8.104) [302]. The desired diastereoisomer was assembled by
coupling the intermediates 146 and 149, which previously were prepared via kinetic
resolution processes. The desired (R)-146 and (R)-149 were obtained in 46% and
42% yield using BCL (Amano PS-30) and CAL-B, respectively.
318
O OAc
- N+ Br
O
O OH
-O N+ Br BnO
( S)-147
BCL (Amano PS-30)
BnO +
146 O OH
- N+ Br
O
BnO
(R)-146
NHAc
NH 2
H 3CO
(R)-149
H 3CO CAL-B
+
148
NH 2
H 3CO
(S)-148
O OH
H
-
N+ N
(R)-146 + (R)-149 O
BnO OCH3
(R,R)- formoterol
Scheme 8.104 Enzymatic synthesis of formoterol.
On an industrial scale, lipase-catalyzed resolution of b-chlorostyrenes was shown

to be an effective strategy for the synthesis of enantiopure styrene oxides [303].
b-Hydroxynitriles are extensively employed in organic chemistry for the prepara-
tion of various intermediates for bioactive compounds. An efficient synthetic method
is the kinetic resolution using BCL [304], yielding the acetate in >99% e.e. CAL-B
(Novozym 435) was instead used for the DKR of several alkyl-, aryl-, aryloxymethyl
substituted b-hydroxynitriles, in the presence of Ru-complex 150 to ensure constant
racemization of the substrate [305]. The same approach was also applied to azido-
alcohols (Scheme 8.105), precursors of optically active aziridines and b-amino
alcohols [306].
Optically active c-hydroxy acid derivatives are well known as chiral building blocks
in the synthesis of natural products and enantiomerically enriched lactones. Unlike
a- and b-hydroxy acids, their preparation is challenging due to their tendency to
lactonization. An efficient approach towards this valuable class of compounds is a
DKR combining a lipase-catalyzed resolution and a Ru-promoted racemization
OH CAL-B, 150, OAc
X X
R R
p-Cl-C6 H 4-OAc
ee = 85-99% (X = N3)
ee = 94-97% (X = CN)
R = phenyl, substituted phenyl, benzyl, naphthyl; X = N3
R = phenyl, substituted phenyl, naphthyl; X = CHN
O O Ph
Ph H
Ph Ph
H H
Ph Ph Ru Ru H
OC CO COCO
150
Scheme 8.105 DKR of b-azido- and b-cyano-alcohols.
(Scheme 8.106) [307]. This lactonization could be reduced using sterically hindered
tert-butyl 4-hydroxypentanoate and N,N-diisopropyl-4-hydroxypentanamide. The
best results are obtained using toluene as a solvent and BCL as enzyme (E ¼ 68
and 400, respectively). The racemization, caused by a hydrogen transfer, is optimal in
the presence of 2,4-dimethyl-3-pentanol as reducing reagent in the hydrogen transfer
reaction.
OH p-Cl-C6 H4 -OAc OAc

R BCL, 150 R
toluene
O OH O
R = Ot Bu R = OtBu, E = 68
R = N(iPr) 2 R = N(iPr)2 , E = 400
60°C or 70°C
Scheme 8.106 DKR of c-hydroxy acids.
Enantiomerically pure allylic alcohols are an important structural motif, with a

range of useful applications in the synthesis of both natural and non-natural
compounds. The first kinetic resolution of allylic alcohols was reported by Burgess
and Jennings [308], and subsequently applied to the synthesis of a statin analog [309].
The synthesis of optically pure allylic alcohols starts often from the corresponding
a,b-unsaturated ketones, combining the chemical reduction of the carbonyl group
with the lipase-catalyzed resolution of the alcohol. Chemical reductions using either
NaBH4 [310] or chiral diamines (ether-phosphine)Ru(II) complexes [311] have been
reported (Scheme 8.107). In the latter case, an enantiomerically enriched mixture of
unsaturated alcohols was obtained. Resolution with BCL immobilized on either
ceramics (PS-C) or diatomaceous earth (PS-D) afforded the alcohol with high e.e.
320
O OH OAc
a or b
Ph Ar + Ar
(a) (i) NaBH 4 , moist alumina; (ii) BCL (PS-C),

isopropenylacetate, E = 357
(b) 151, H2 ; BCL (PS-D), isopropenylacetate, E = 230
Ph Ph Ph Ph
Cl P O
O P
Ru
H 2N Cl NH 2
151
Scheme 8.107 Resolution of allylic alcohols.
An interesting domino transformation has been developed [312] for allylic

alcohols of structure 152 in the presence of a suitable acyl donor (Scheme 8.108).
Optimization of the experimental conditions led to the combination of a CAL-B
mediated acylation, a Ru-catalyzed racemization, and an intramolecular Diels–Alder
reaction in one pot. The outcome of this tandem process is the tricyclic compound
153. Recently, the first application of enzymes for the synthesis of optically pure
boron-containing compounds was described [313]. A set of various alcohols contain-
ing a boronic group or a boronate ester was resolved using lipases (Scheme 8.109).
A screening of different enzymes revealed very poor activity for lipases from
Aspergillus niger, C. rugosa, Mucor javanicus, PPL, and Penicillium camemberti. On
the other hand, CAL-B and BCL exhibited E values above 200.
O CO2 Et
CO2 Et
O
OH EtO O O
O
CAL-B, Et3N, MS CO2 Et
Cl
152 Ru 2 153
Cl
81%, ee = 97%
MeCN
Scheme 8.108 Domino transformation involving DKR of allylic sec-alcohols.
A common problem of the resolution of alcohols via acylation is the separation

between the unreacted alcohol and the formed ester. A recently proposed solution
consists of the use of highly fluorinated acyl donors (Scheme 8.110). In this case the
OH OH OAc
CAL-B or BCL
RO RO RO
B n-hexane, B + B
OR vinyl acetate
OR OR
E > 200
Scheme 8.109 Resolution of boron-containing alcohols.
O
OH OH O (CF2 )CF3
Ph CAL-B, MeCN
Ph + Ph
O
ee > 99% ee = 98%
F3 C O (CF2 )7 CF3
154
Scheme 8.110 Use of fluorinated acyl donors.
acylation is accompanied by a simultaneous enantiomer-selective fluorous phase

labeling [314]. The fluorinated ester is separated from the remaining alcohol by
simple liquid–liquid extraction using perfluoro-n-hexane and methanol. This
method has been applied to the kinetic resolution of low-molecular weight sec-
alcohols with good results in terms of conversion and optical purity; moreover, the
fluorous tag 154 could be quantitatively recovered by hydrolysis of the ester. On the
other hand, fluorinated acyl donors are not competitive with respect to the inexpen-
sive conventional reagents for lipase-catalyzed acylation. Moreover, the high envi-
ronmental impact of perfluoro-n-hexane does not allow the application of this
separation technology on an industrial scale.
Enzymatic Resolution of Cyclic Secondary Alcohols Ghanem and Schurig studied the
asymmetric acylation of secondary alcohols catalyzed by BCL immobilized on
ceramic particles (Amano PSL-C); best performances were observed in toluene and
with isopropenyl acetate as acyl donor (Scheme 8.111) [315]. The cyclic alcohols
155–157 were acylated under these conditions with moderate to good enantioselec-
tivity (E ¼ 9, 10, and 42, respectively).
A much more efficient resolution (E > 1500) of compound 156 and of other
alcohols is obtained in sc-CO2 with CAL-B (Novozym 435, Scheme 8.112) [316]. The
positive effect of sc-CO2 as a reaction medium for lipase-mediated transesterifica-
tions [317] has been studied extensively by Matsuda and coworkers [318].
Resolution by acylation was also successfully achieved for four-membered rings.
As an example, the readily available 2-hydroxycyclobutanone and the corresponding
acetals were enantioselectively acylated using various lipases in organic solvents
[317, 319]. The best enantioselectivities were achieved using CAL-B in n-hexane
(Scheme 8.113).
322
HO O HO
PSL-C
+
toluene
isopropenylacetate (R)-ester (S)-alcohol
155: R = E =9
156: R = E = 10
157: R = E = 42
Scheme 8.111 Resolution of cyclic sec-alcohols using BCL.
OH OAc OH
CAL-B (Novozym 435),
vinyl acetate
+
sc-CO 2, 13 MPa
156 E > 1500
Scheme 8.112 Enzymatic resolution in sc-CO2.
OR OR OR
OR OR OR
CAL-B
+
hexane O
OH OH
O
(R/ S)-158 (S)-158 (R)-159
a: R = Me a: ee = 99.9% a: ee = 97.6%
b: R = Et b: ee = 64.5% b: ee = 98%
Scheme 8.113 Resolution of cyclobutanols.
An interesting example of cyclic sec-alcohols is given by N-acyl hemiaminal 160

(Scheme 8.114) [74]. The enzymatic DKR using various lipases and isopropenyl
acetate was investigated by Sharfuddin et al. [320]. In all cases, the resolution afforded
exclusively the (R)-acetates in quantitative yield.
OH OAc
lipase, n-hexane
NCOR NCOR
isopropenyl acetate
160 O 161 O
R = Me, lipase = PFL: ee = 63%

R = Et, lipase = PFL: ee > 99%
R = n-Pr, lipase = BCL: ee > 99%
R = i-Pr, lipase = BCL: ee > 99%
R = t-Bu, lipase = BCL: ee > 99%
R = Ph, lipase = BCL: ee > 99%
Scheme 8.114 Resolution of hemiaminals.
Simvastatin (164), a semi-synthetic derivative of the fungal polyketide lovastatin,

shows excellent cholesterol-lowering activity. Marketed by Merck under the trade
name ZocorÒ, it is one of the best-selling drugs in the USA. As an alternative to the
two semi-synthetic processes used in current production, Xie et al. developed a one-
step synthesis starting from monacolin J (Scheme 8.115) [321]. The enzyme of choice
is LovD, a dedicated acyltransferase from the lovastatin biosynthetic gene cluster.
Interestingly, this enzyme regioselectively transfers the a-dimethylbutyryl group
onto the C8 hydroxyl group of monacolin J to afford simvastatin. Conversions >99%
and purity >98% were obtained using whole cells and the membrane-permeable
substrate a-dimethylbutyryl-S-methyl-mercaptopropionate.
O O O
OH OH OH
O OH OH O OH
O OH O
Hydrolysis LovD
DMB-S-MMP
Lovastatin, 162 Monacolin J, 163 Simvastatin, 164
O O
DMB-S-MMP = S O
Scheme 8.115 Enzymatic synthesis of simvastatin.
The (4S,6S)-stereoisomer of dorzolamide 166 is used as active principle in

ophthalmic solutions (TrusoptÒ ) to reduce elevated intraocular pressure. Recently,
lipase-mediated kinetic resolution of rac-165 was investigated [322], as a viable
approach to an optically active intermediate for dorzolamide (Scheme 8.116). After
screening several different acyl donors, vinyl butanoate was chosen. A rac-cis- 165
324
OH
NHEt
O
S NH 2
S S
S S O
O O
r ac-cis 165 Dorzolamide, 166
OH
OCOPr OH
S S
PrCO 2CH=CH2
+ +
OH BCL (lipase PS-D) S S S S
(4R,6S)-167 (4S,6R)-165
S S
BuOH H 3O +
r ac-cis 165 BCL
OH
OH
S S
S S
(4R,6S)-165
(4R,6R)-165
de = 58%
H3 O+
OH
S S
(4S,6S)-165
de = 51%
Scheme 8.116 Resolution of an intermediate for dorzolamide (I).
mixture was acylated, followed by alcoholysis of the (4R,6S)-ester. Both enzymatic

steps were catalyzed by BCL as Amano PS-D. As previously reported, epimerization
of the alcohol involving the 4-position takes place under slightly acidic conditions. As
a consequence, mixtures of cis and trans alcohols were obtained, with the more stable
trans isomers slightly prevailing.
Both (4R,6S), (4S,6S), and (4S,6R), (4R,6R) mixtures were further elaborated to
obtain (4S,6S)-165 and (4R,6R)-165, respectively. In the first case, (4S,6S)-165 was
easily prepared via diastereoselective acylation of the unepimerized counterpart in
the mixture (Scheme 8.117). In the second case (Scheme 8.118) both diastereo-
isomers were first acylated to a mixture of butanoates, and then a highly diastereo-
selective alcoholysis of (4R,6R)-167 yielded (4R,6R)-165.
OH
S S
OH
(4 R,6S)-165
PrCO2 CH=CH 2
BCL, MTBE
+ S S
OH (4S,6S)- 165
ee = 99%
de = 96%
S S
OCOPr
(4S,6S)- 165
de = 51%
S S
(4 R,6 S)-167
Scheme 8.117 Resolution of an intermediate for dorzolamide (II).
OH OCOPr
S S S
S
(4S,6R)-165 (4S,6R)-167 OH
ee = 99% PrCO2 CH=CH 2 BuOH
BCL, MTBE BCL, MTBE
+ + S
S
OH OCOPr (4R,6R)-165
ee = 97%
de = 93%
S S S
S
(4R,6R)-165 (4R,6R)-167 OCOPr
ee = 98%
de = 58%
S S
(4S,6R)-167
Scheme 8.118 Resolution of an intermediate for dorzolamide (III).

326
A major drawback of the hydrolase-catalyzed enantioselective acylation of

alcohols is the product isolation. Separation of the remaining alcohol from the
ester is indeed often laborious, and column chromatography is frequently needed.
Methods that avoid the purification step are therefore especially appreciated. One
strategy was recently developed and applied to the kinetic resolution of pyridyl
alcohols 168 and 169 [323]. The method is based on the fact that many monoesters
of phthalic acid are known to be crystalline solids (Scheme 8.119). The result is a
two-step procedure, in which enzymatic kinetic resolution of 168 and 169 is
combined with derivatization of the unreacted alcohol with phthalic anhydride.
Acetates 172 and 173 and phthalic acid monoesters 170 and 171 could be easily
separated by filtration; hydrolysis of the esters with sodium hydroxide and
subsequent recrystallization from cyclohexane afforded enantiopure alcohols
168 and 169.
1) CAL-B, vinyl acetate, iPr 2O, 30 h

( )n 2) phthalic anhydride, Et 3N, 6 h
N 3) sat. NH4 Cl
HO 4) separation by filtration
168: n = 1
169: n = 2
( )n
N
O ( )n
O + N
O AcO
OH
( S)-170: n = 1, 49% (R )-172: n = 1, 49%

( S)-171: n = 2, 43% (R )-173: n = 2, 39%
2M NaOH,
MeOH, 6 h
( )n + ( )n
N N
HO HO
( S)-168: n = 1, 88%, ee > 99% (R )-168: n = 1, 91%, ee > 99%

( S)-169 : n = 2, 87%, ee > 99% (R )-169: n = 2, 72%, ee > 99%
Scheme 8.119 Resolution of pyridyl alcohols.

8.6.2.3 Enzymatic Resolution of Tertiary Alcohols
The availability of chiral building blocks containing quaternary carbon centers is
essential for several applications, including the synthesis of pharmaceuticals. The
synthesis of enantiopure tertiary alcohols and their esters represents therefore an
important task [65]. The application of enzyme-catalyzed transesterification to
tertiary alcohols is not very common, because tertiary alcohols and their derivatives
are indeed difficult substrates for most hydrolases when compared to primary and
secondary alcohols. Only a few hydrolases able to accept tertiary alcohols have been
identified, mainly in microorganisms [65]; in addition, metagenomic resources
represent a viable alternative. A crucial result in this field is the discovery of a
common GGG(A)X motif in the so-called oxyanion hole of most hydrolases active
towards tertiary alcohols [324, 325]. CAL-A, CRL, PLE, acetyl choline esterases, and
esterase II from Bacillus subtilis share all this feature, and exhibit activity towards
esters of tertiary alcohols. CAL-A efficiently resolves 2-phenylbut-3-yn-2-ol in a
transesterification reaction in organic solvent, with E value as high as 65
(Scheme 8.120) [326].
OH OH
OAc
CAL-A
+
vinyl acetate
solvent E = 65
Scheme 8.120 Resolution of 2-phenylbut-3-yn-2-ol.
The ability of CAL-A to accept sterically hindered tertiary alcohols is also dem-
onstrated by the resolution of racemic 1-methyl-2,3-dihydro-1H-inden-1-ol and 1-
methyl-1,2,3,4-tetrahydronaphthalen-1-ol [327]. Both aromatic fused cyclic tertiary
alcohols can be resolved using vinyl acetate as the acyl donor with moderate to good
enantioselectivity (Scheme 8.121).
OH OH OAc
vinyl acetate
+
n CAL-A or
n n
CAL-A-CLEA
n = 1, 2 (S)-(+)-alcohol
n = 1, E = 4 (CAL-A-CLEA)
n = 2, E = 253 (CAL-A)
Scheme 8.121 Resolution of aromatic fused tertiary alcohols.
8.6.2.4 Enzymatic Resolution of rac-Diols

The resolution of a diol under transesterification conditions can lead in principle
to a mixture containing unchanged diol and mono- and diacetate (Scheme 8.122).
A simpler analysis of this problem is, however, possible in two cases, which can be
328 j 8 Hydrolysis and Formation of Carboxylic Acid Esters
OH
OH OH
R1 OH
R2
R1 R2
OH 1,2- 1,3- OH
n OH
OH
cyclic 1,2-diols
R2 OH
R1 HO
1,4- OH acyclic diols
bicyclic diols
HO
of r ac-diols
R1
OH
HO
R2
OH
1,1-disubstituted 1,2-diols
OH
aromatic (Binol)
Scheme 8.122 Overview of the resolution of diols.
regarded as acylations of monohydroxy compounds. In the first case, the reaction

stops after the first acylation step. In the second case, the first acylation is a fast
step with low enantioselectivity, which completely converts the diol into the
corresponding racemic monoacetate. The latter is then resolved with high
enantioselectivity to give a monoacylated derivative and the corresponding dia-
cylated derivative [328].
Enzymatic Resolution of rac-Acyclic Diols When diastereoisomeric mixtures of

unsymmetrical diols are converted into enantiopure derivatives, the term kinetic
asymmetric transformation (KAT) should be preferred to the older kinetic
resolution. Indeed, the latter strictly refers to the separation of the enantiomers
from a racemic mixture, with recovery of at least one enantiomer [329], and is
therefore not appropriate in this case [330].
The dynamic kinetic asymmetric transformation (DYKAT) of unsymmetrical 1,5-,
1,4-, and 1,3-diols by combining a lipase-mediated acylation and a Ru-catalyzed
epimerization has been effectively developed by B€ackvall.
The transformation of unsymmetrical 1,3-diols containing one large and one small
group into enantiomerically pure syn-1,3-diacetates took advantage of the tendency of
syn-1,3-diol monoacetates to undergo syn-1,3-acyl migration (Scheme 8.123 and
Table 8.2). The integration of such spontaneous rearrangements in an established
chemoenzymatic sequence (a CAL-B catalyzed acetylation and a Ru-mediated
OH OH isopropenyl acetate, OAc OAc
CAL-B, 150
R R
(S,S), (R,R) toluene
(S,R), (R,S)
Scheme 8.123 DYKAT of syn-1,3-diacetates.
Table 8.2 DYKAT of syn-1,3-diacetates (see Scheme 8.123).
Entry R Yield (%) D.r. (%) E.e. (%)
1 CH3(CH2)4- 69 86 : 14 >99
2 Ph 73 93 : 7 >99
3 PhCH2 63 88 : 12 >99
4 2-Naphthyl 53 96 : 4 >99
5 p-Br-C6H4 62 92 : 8 >99
6 p-Cl-C6H4 59 92 : 8 >99
epimerization) resulted in the preparation of a set of syn-1,3-diacetates with >99%

e.e. and syn:anti ratio >90% [330].
In a similar approach, anti-1,5-diacetates [331] and syn-1,4-diacetates [332] were
obtained via DYKAT from the corresponding diols in excellent enantioselectivity and
moderate to very good diastereoselectivity. In both cases, either CAL-B or BCL was
isopropenyl acetate,
OH CAL-B or BCL, OAc
174 , tBuOK, Na 2 CO 3, 1 R2
R R
175 OH 176 OAc
toluene
a R = CH3 a R = CH3 , dr = 95:5, ee > 99%
b R = Cl b R = Cl, dr = 84:16, ee > 99%
isopropenyl acetate, CAL-B

and BCL,
OH OH OAc OAc
174 , tBuOK, Na 2CO3 ,
R R
177 178
toluene
a R = C(O)OCH 3 a R = C(O)OCH 3 , dr = 80:20, ee = 98%
b R = CH2CN Ph b R = CH2 CN, dr = 94:6, ee > 99%
Ph
Ph
Cl
Ph Ph Ru
OC CO
174
Scheme 8.124 DYKAT of syn-1,4- and 1,5-diacetates.

330
OAc
TsO PPh 2
OAc BCL R OAc
DEAD, AcOH
TsO TsO
R buffer-iPr 2O toluene R
OH
TsO
R
R E value Total yield (%) ee (%)

Bu >200 91 98
Me 113 85 93
CH 2 OPMB 145 83 96
CH2 Cl >200 73 95
Scheme 8.125 Deracemization of 1,2-diol monotosylate derivatives.
used in combination with the Ru-based racemization/epimerization catalyst

(Scheme 8.124). Interestingly, with substrate 177a CAL-B did not react in the second
acylation, thus affording the monoacetate (2RS,6R) as the only product. Addition of
BCL after the first acylation helped overcome this limitation.
1,2-Diol monotosylate derivatives were effectively deracemized through the com-
bination of a lipase-mediated hydrolysis and a Mitsunobu inversion without inter-
mediate purification (Scheme 8.125) [333]. As a result of such a protocol, racemic 2-
acetoxyhexyl tosylate was converted into the corresponding (S)-enantiomer as the
only reaction product. The method was successfully extended to a small group of
different 1,2-diols, which were all converted into the corresponding optically active
monoacetates with very high total yield and e.e.ps. The use of polymer-supported
PPh3 for the Mitsunobu reaction allows separation of the products from the reagent-
derived by-products by simple filtration.
Enzymatic Resolution of Acyclic 1,1-Disubstituted 1,2-Diols 1,1-Disubstituted 1,2-

diols are important building blocks for the preparation of bioactive compounds [284].
Their resolution is usually achieved via lipase-catalyzed acyl transfer rather than
hydrolysis; in all cases present in the literature, acylation takes place exclusively at the
primary alcohol group with high efficiency, while the bias of the tertiary alcohol
provides the enantioselectivity. By analogy to the conventional model for lipase-
catalyzed reactions on primary and secondary alcohols (Kazlauskas rule [334]), Fang
and coworker [335] proposed a model to describe and predict enantioselectivity in
lipase-catalyzed resolution of these substrates. In this model, the tertiary OH is
considered as the smallest group and believed to serve only as the directing
stereocenter. High enantioselectivity could be expected for substrates having suffi-
cient differences between the large, medium, and small substituent. To test this
model, enzymatic resolution by acylation of several 1,1-disubstituted 1,2-diols was
attempted (Scheme 8.126). A preliminary survey indicated a good tolerance towards
these hindered substrates by most lipases, with PFL (lipase AK) showing the highest
OH OH OH
X OH PFL X OH X OAc
+
isopropenyl acetate
X = H, E = 2.0
X = Cl, E = 11.5
X = Br, E = 13a
X = I, E = 69
OH OH OH
X OH PPL X OH X OAc
vinyl acetate
MTBE
R R R
R = Ph, X = Cl, E = 72 b
R = Ph, X = I, E > 1000
R = PhCH 2OCH 2, X = I, E = 24
R = Me 3Si, X = I, E = 368
R = CH3 (CH 2)3 , X = I, E = 47
R = CH3 (CH 2)5 , X = I, E = 10
R = Me 3C, X = I, E = 751
a
vinyl acetate was used as the acyl donor
b No MTBE was used
Scheme 8.126 Resolution of 1,1-disubstituted 1,2-diols.
enantioselectivity with isopropenyl acetate as both acyl donor and solvent. The
influence of the X group was also observed, and enantioselectivity was found to
increase with the size of this substituent. To further broaden the scope of the method,
the same group investigated the enzymatic resolution of racemic 2-alkynyl-3-halo-
1,2-propanediols (Scheme 8.126). The enzyme of choice was this time PPL; very high
enantioselectivities were obtained using vinyl acetate as the donor in MTBE as the
solvent. Furthermore, the highest E values were observed for R ¼ Ph, Me3C and
Me3Si, thus showing a clear correlation between the bulkiness of the R group and the
efficiency of the chiral recognition process.
In the case of compound 179, the racemic diol was resolved by conversion into the
mono- and diacetate (Scheme 8.127); e.e.s as high as 98% and 95%, respectively, were
obtained in the BCL-catalyzed acylation with vinyl acetate [336].
Enzymatic Resolution of Cyclic and Bicyclic Diols Figure 8.8 and Table 8.3 summarize
some significant examples of attempted resolution of cyclic diols via enzymatic
acylation. Resolution of diol 180 (Table 8.3, entry 1) afforded a complex reaction
mixture consisting of unreacted diol, two regioisomeric monoacetates, and traces of
the corresponding diacetate with very modest e.e. for all products. Acetylation of
332
OH
O OAc
OH MeO
ee = 98%
O OH
BCL
+
MeO vinyl acetate
179 OAc
O OAc
MeO
ee = 95%
Scheme 8.127 Resolution of propanediol 179.
racemic 181 with PFL resulted in the efficient formation of a mixture of the two
regioisomeric monoacetates. Interestingly, acylation of both enantiomers of 181 took
place, but with completely different regioselectivity as the result of the bulky trityl
group. Each monoacetate was then converted separately in one of the enantiomers of
the diacetate. Symmetric diols 182–184 (entries 3–5) all gave mixtures of unreacted
diol, monoacetate, and diacetate. For the five-membered diol 182 very high e.e.s were
observed for both mono- and diacetate. In contrast, diol 183 yielded a mixture of the
three possible compounds, of which only the diacetate was enantiomerically pure. A
mixture of the three possible compounds was obtained also for the seven-membered
OTr OTr
OH OH
OH
OH OH
OH OH
OH
180 181 182 183
OH HO H
OH OH
OH
H OH
184 185 186
O O Br
OH
HO OH
OH
HO OH
OH OH
187 188
189
Figure 8.8 Resolution of cyclic diols.

Table 8.3 Resolution of cyclic diols (Figure 8.8).
Entry Substrate Enzyme Acyl donor E Reference
1 180 BCL VAa) — [337]

2 181 PFL VA — [338]
3 182 PFL VA — [339]
4 183 PFL VA — [339]
5 184 PFL VA — [339]
6 185 PPL TCAb) — [340]
7 186 PFL AcOPh 4 [341]
8 187 PFL n-PrCO2Ph 11 [341]
9 188 PFL AcOPh 2 [341]
10 189 BCL VA — [342]
a) VA ¼ vinyl acetate.
b) TCA ¼ 2,2,2-trichloroethyl acetate.
diol 184 as well, with very high e.e.s for the diol and the diacetate. The C2-symmetric
compound 185, a building block for prostaglandins, could be efficiently resolved to
give one enantiomer as diacetate with >99% e.e. and the other enantiomer as
remaining diol with moderate enantioselectivity. The e.e. of the latter could be
enhanced up to 90% by reiterating the acylation on the enriched diol fraction.
Resolution of the series of the bicyclic diols 186–188 (entries 7–9) was successful only
for 188 using PFL as Amano YS. The racemic tetrol 189 (entry 10), representing a
calicheamicenone intermediate, was resolved by repeated lipase-catalyzed acetylation
of the primary hydroxy group.
Enzymatic Resolution of Aromatic Diols Enzymatic resolution also finds an inter-

esting application in the preparation of optically active 1,10 -binaphthyl-2,20 -diols
(BINOL), which have been used as chiral auxiliaries for asymmetric transformations.
An example [343] is given by the resolution via acetylation of 6,60 -disubstituted
BINOL derivatives using BCL and PFL immobilized on Celite. Depending on the
substituent at the 6,60 -positions, 78–96% e.e. for the monoacetate and 55–80% e.e. for
the diol were obtained (Scheme 8.128).
X X X
O
OH vinyl acetate O OH
OH OH + OH
iPr 2O/acetone
PFL or BCL
X X X
190
ee 78-96% ee 55-80%
X=H
X = Br
x = OMe
Scheme 8.128 Resolution of BINOLS.

334
Similarly, the enzymatic resolution of racemic 2,20 -dihydroxy-6,60 -dimethoxy-1,10 -

biphenyl was investigated (Scheme 8.129). Among all the enzymes screened, only
BCL was able to resolve the racemic diol in MTBE [344]. Differences in enantio-
selectivity and reaction rate were observed for different enzyme preparations; the best
compromise between catalytic efficiency and enantiodiscrimination was achieved
with BCL adsorbed on ceramics (PSL-C).
MeO OH MeO OH MeO OH

vinyl acetate
MeO OH MeO OAc + MeO OH
BCL
1 day
191 E = 74
Scheme 8.129 Resolution of a biphenyl compound.
8.6.2.5 Enzymatic Resolution of rac-Acids and rac-Esters

The kinetic resolution of rac-acids and esters is usually carried out in hydrolytic mode
starting from the corresponding esters or lactones, because the hydrolysis in water is
generally faster than the esterification in organic media. Lipases from C. rugosa and
porcine pancreas are the most used enzymes for such hydrolyses [345]. However,
successful examples of lipase-catalyzed resolution of chiral acids via esterification or
transesterification can be found in the literature (Scheme 8.130).
of rac -acids and
esters
R3
R1
R1 CO 2R 4
CO2 R4 2
R
R2
stereocenter in α-position stereocenter in β-position
Scheme 8.130 Overview of the resolution of rac-acids and esters.

at the a Position
Rhizomucor miehei lipase (RML) proved to be effective in the resolution of racemic
acid and esters on several occasions. This allowed the resolution of methyl trans-
b-phenylglycidate 192, a possible precursor for the taxol side-chain [346]. The use of
the more hindered isobutanol instead of n-butanol resulted in higher yields and
enantioselectivity (Scheme 8.131).
RML
O O CO 2Me + Ph O
CO2 Me
iBuOH
Ph hexane Ph CO2 iBu
192
ee = 77% ee = 95%
conv. = 45%
Scheme 8.131 Resolution of a precursor of the taxol side-chain.
Resolution via transesterification is often applied to 2-arylpropionic acids (pro-

fens), which constitute an important class of non-steroidal anti-inflammatory drugs
(NSAIDs). Since only the (S)-enantiomers are responsible for their pharmacological
activity, several strategies have been implemented for their selective preparation, and
many of them exploit hydrolase-catalyzed esterification reactions in organic solvents.
Figure 8.9 summarizes some successful examples for the most prominent profens.
Excellent enantioselectivity (E > 1000) was observed for naproxen [347] using CRL
and 1-butanol at controlled water activity (aw ¼ 0.8). Ketoprofen [348] and flurbipro-
fen [349] were resolved with CAL-B as Novozym 435 through (R)-selective esteri-
fication; in both cases the desired unreacted (S)-acid was recovered in >98% e.e.
RML (Lipozyme IM20) catalyzed the esterification of ibuprofen [350] with butanol
with E ¼ 113 [351].
Recently, mycelia of different moulds proved to be a viable alternative for the
resolution of 2-arylpropionic acids [352]. As an example, the resolution of rac-2-
phenylpropanoic acid in organic solvent was studied by comparing dry mycelia of A.
oryzae MIM and R. oryzae CBS 112.07. The best results were obtained using ethanol
in either heptane or pentadecane; interestingly, the two strains were enantiocom-
CO2H Ph CO2 H
H3 CO
Naproxen Ketoprofen
S-selective esterification R-selective esterification
CRL,1-butanol CAL-B,ethanol
E > 1000 Remaining acid: ee = 98%
F
CO 2H CO 2H
Ph
Flurbiprofen Ibuprofen
R-selective interesterification S-selective esterification
CAL-B,tri-n-propylorthoformate RML,1-butanol
remaining acid:ee > 98% E= 113
Figure 8.9 Resolution of 2-arylpropionic acids (profens).

336
plementary. An efficient protocol for the resolution of flurbiprofen was developed

based on the dry mycelia-catalyzed transesterification in toluene at 50 C
(Scheme 8.132); at 55% of conversion, the unreacted (S)-flurbiprofen was recovered
in 84% e.e. [352].
F O
OH
dry mycelia
+ CH 3(CH 2) 7OH
CH 3
toluene
rac-flurbiprofen
F O F O
OH O(CH2 )7 CH 3
+
CH3 CH 3
(S)-flurbiprofen
Scheme 8.132 Mycelia-catalyzed resolution of flurbiprofen.
Captopril (1-[(2S)-3-mercapto-2-methylpropionyl]-l-proline) is a potent ACE inhib-

itor. By preventing the conversion of angiotensin I into angiotensin II, it is
responsible for the suppression of the renin–angiotensin–aldosterone system, thus
acting as a potent antihypertensive agent. The inhibiting capacity of captopril
depends on the configuration of the mercapto alkanoyl moiety; indeed, the (2S)-
stereoisomer was found to be about 100 times more active than the corresponding
(2R)-stereoisomer [353]. Compound (S)-193, a key intermediate for the synthesis of
the chiral side chain of captopril and of the analog zofenopril (Scheme 8.133), was
prepared by lipase-catalyzed enantioselective esterification of the corresponding
racemic mixture in organic solvent. Using BCL in toluene and methanol as the
Ph S
CO 2Me
Ph S
CO2 H BCL O
O toluene, MeOH +
193 Ph S
CO2 H
O
( S)- 193 , ee = 97%
Ph Ph
S S
Ph S N HS N
O O CO2 H O CO2 H
zofenopril captopril
Scheme 8.133 Enzyme-mediated synthesis of the side-chain of zofenopril and captopril.

Table 8.4 Resolution of rac-acid and esters via transesterification.
Entry Substrate Enzyme E Transformation Reference
1 OH CRL >50 Esterification with n-butanol [354]

COOH
2 OH CAL-B 7-10 Esterification with i-butanol [355]

i COOH
BuOOC
3 Ph O CAL-B 23 Transesterification with n-hexanol [356]
nucleophile, the desired (S)-193 was obtained in 37% yield and 97% e.e. Comparable
e.e. (97.7%) but higher yield (45%) were observed using BCL immobilized on Accurel
polypropylene (PP).

at the b-Position
Despite the distance between the site of reaction and the stereocenter, the enantio-
selective resolution of this class of acids and esters has been observed for several
lipases. The resolution is usually performed in hydrolytic mode; however, successful
examples of both esterification and transesterification have been reported (Table 8.4).
Effective resolution of b-amino acid derivative 194 was performed by alcoholysis
with 1-butanol and CAL-B (Scheme 8.134) [357]. The transformation was highly
enantioselective and provided a separable mixture of the remaining ethyl ester and
the butyl ester.
O O O
1-butanol n-C3 H7 NH + n-C3 H7 NH
n-C 3 H7 NH
CAL-B
CO2 Et CO 2n-Bu CO 2Et
194 E > 100
Scheme 8.134 Alcoholysis of a b-amino acid derivative.
8.6.3
Desymmetrization of Prochiral and meso-Carboxylates via Transesterification
In the enantioselective formation of carboxylates starting from prochiral or meso

compounds, the desymmetrization of acids and esters has received considerably less
attention than the complementary desymmetrization of prochiral or meso diols [3].
Therefore, only a limited number of examples have been reported to date.
338
Table 8.5 Alcoholysis of dibenzoates (Scheme 8.135).
Substrate R E
195a CH3 14.0

195b C2H5 7.8
195c C4H9 7.0
195d C8H17 4.0
195e Ph —a)
195f Cl 1.7
a) No reaction occurred.
Dibenzoates 195a–f were desymmetrized by immobilized BCL (PSL-C)-mediated

octanolysis in DIPE (diisopropyl ether) (Scheme 8.135 and Table 8.5). Despite the
modest enantioselectivity displayed, this resolution represents the first enzymatic
route towards 1,2-diols protected as benzoate at the secondary alcohol [358].
OCOPh BCL OCOPh

OCOPh 1-octanol, DIPE OH
R R
195a-f 196a-f
Scheme 8.135 Alcoholysis of dibenzoates.
Recently, a new approach for the enzymatic desymmetrization of 3-substituted

glutaric acids has been reported (Scheme 8.136) [359]. The Novozym 435 catalyzed
monoesterification in iso-octane has been studied with the aid of molecular docking
to predict the correct size for the alcohol employed. The experimental results were all
in agreement with the outcome of the docking: small and medium-sized alcohols
gave the corresponding esters in high yield and e.e., while bulky alcohols did not react
at all.
TBS TBS
O O O O O O
CAL-B
HO OH + ROH HO OR
isooctane
ROH Yield/ee (%)

MeOH 81/92
EtOH 84/94
n-PrOH 93/98
i-PrOH 89/98
n-BuOH 85/97
i-BuOH 82/95
t -BuOH 81/96
n-pentanol 77/95
i-pentanol 78/94
Scheme 8.136 Desymmetrization of 3-substituted glutaric acids.

8.7 Enantioselective Formation of Carboxylates from Prochiral and meso-Diols j339
Although less common, the desymmetrization of anhydrides to optically active
esters and acids has also been reported. The alcoholysis of 3-arylglutaric anhydrides
197a–d with ethanol and CAL-B has been described (Scheme 8.137) [360]. Reaction
times and enantioselectivities were found to depend on the solvent used; DIPE
ensured optimal rates and good selectivity for substrates 197a–c, while MTBE proved
to be the best solvent for 197d. Esterification of the anhydrides with different esters
led to markedly lower e.e. values.
R1
R2
R1
CAL-B R2
ethanol
iPr 2 O or MTBE O O
O O O EtO OH
197a-d 198a-d
a: R 1 = R 2 = H a: R1 = R 2 = H, 78% ee
b: R1 = Cl, R 2 = H b: R1 = Cl, R 2 = H, 68% ee
c: R1 = R2 = Cl c: R1 = R2 = Cl, 60% ee
d: R1 = OMe, R 2 = H d: R1 = OMe, R2 = H, 69% ee
Scheme 8.137 Alcoholysis of glutaric acid anhydrides.
8.7
Enantioselective Formation of Carboxylates from Prochiral and meso-Diols
(Desymmetrization via Acylation)
8.7.1
Overview
Desymmetrization of meso- and prochiral diols via acylation has become a useful tool
to access chiral compounds, even on an industrial scale. This section discusses the
most relevant examples of desymmetrization reported in the literature, starting from
prochiral diols as substrates.
8.7.2
Desymmetrization of Prochiral Diols
The desymmetrization of prochiral diols has proven to be an efficient tool for the
generation of chiral intermediates, and also industrially. 2-Substituted and 2,2-
disubstituted 1,3-propanediols and, to a lesser extent, 1,3,5-triol derivatives
(Scheme 8.138) are very common substrates for such desymmetrizations, due to
the prominent role played by their corresponding optically pure forms in the
synthesis of valuable compounds.
340
desymmetrization
of prochiral diols
R' O OH HO O R
R H O
OR O
2-substituted 1,3-diols HO O R'
1,3,5-triol derivatives
R R'
2,2-disubstituted 1,3-diols
Scheme 8.138 Overview of the structures accessible via desymmetrization of prochiral diols.
8.7.2.1 Desymmetrization of 2-Substituted 1,3-Diols

An important example of industrially applied desymmetrization of 2-substituted 1,3-
diols is offered by Schering Plough in applying this strategy to the synthesis of 200
(Scheme 8.139). This monoacetate, a key intermediate for the preparation of the
antifungal azole derivative 201 [303], was obtained in 97% e.e. using CAL-B and vinyl
acetate in acetonitrile. Lowering the reaction temperature to 0 C allowed minimi-
zation of the undesired diacylation. The enzyme could be recovered and reused for six
cycles without significant loss of activity.
OAc
OH F
F
OH
OH CAL-B
vinyl acetate F
F 199 200 , ee = 97%
MeCN
O
F
O
N N
N
F
N N
201
O
N
N
N
HO
Scheme 8.139 Desymmetrization of an antifungal precursor at Schering-Plough.

BCL is often employed in the desymmetrization of 1,3-propandiols, and several
successful examples can be found in the literature.
This enzyme proved to be very efficient in the desymmetrization of 2-(6-benzyloxy-
4-methyl-4-hexenyl)propane-1,3-diol 202, a key step in the chemoenzymatic synthe-
sis of diol 204, discovered in very small amounts in the hair-pencils of Danaus
chrysippus (Scheme 8.140) [361]. Using BCL in a mixture of 1,4-dioxane and THF, a
good compromise between chemical yield (75%) and e.e. (90%) could be achieved.
OH OH
OH BCL, vinyl acetate OAc

BnO BnO
202 1,4-dioxane/THF 5:1 203, 75% (+20% diacetate),
ee = 90%
OH
HO
204
Scheme 8.140 Desymmetrization of 2-(6-benzyloxy-4-methyl-4-hexenyl)propane-1,3-diol.
In the chemoenzymatic synthesis of (S)-imperanene [362], a phenolic compound

possessing interesting platelet aggregation inhibitory activity (Scheme 8.141),
desymmetrization using BCL (PS-30) allowed the preparation of intermediate 205
with a maximum yield of 62%. The latter could be increased up to 90% by recovering
the unreacted diol and subsequent acetylation with recycled enzyme.
O O
TsO OH TsO OAc
H
OH BCL, vinyl acetate OH
THF
205 , 62%
O
HO
HO OH
imperanene
Scheme 8.141 Desymmetrization as a key step in the synthesis of imperanene.
An enzymatic desymmetrization strategy was successfully applied also to the

synthesisofenantioenrichedEvansauxiliaries(2-oxazolidinones,Scheme8.142)[363].
The enzyme of choice, after screening different catalysts, was in this case PPL.
342
O O O
PPL, 30°C
O NH O NH O NH
+
HO OH vinyl acetate HO OAc AcO OAc
206 207 208
69% isolated yield traces
ee = 99%
O H
N OAc
O
Scheme 8.142 Desymmetrization in the synthesis of Evans auxiliaries.
The maximum conversion (>99%) and selectivity (>99% e.e.) were obtained upon
performing the reaction in vinyl acetate at 30 C. The desired monoacetate 207 could be
isolated with a maximum yield of only 69%, however, because of some material loss
during the workup procedure.
Compound 210 (Scheme 8.143) is a 1,2-diacylglycerol (DAG) surrogate and a
protein kinase C (PKC) ligand. Its preparation in enantiopure form via enzymatic
desymmetrization of the corresponding diol 209 was recently reported by Chênevert
et al. [299]. The best results were obtained in this case using the lipase from Aspergillus
niger (ANL). No reaction was observed at all with CRL and BCL, while PFL and CAL-B
gave very poor enantioselectivity.
OBn OBn
Aspergillus niger
vinyl pivalate, Et2O
OH OH OH O
209 210
O
96%, ee = 98%
Scheme 8.143 Desymmetrization of protein kinase C ligand 209.
8.7.2.2 Desymmetrization of 2,2-Disubstituted 1,3-Diols

2,2-Disubstituted 1,3-diols are less readily accepted by enzymes than the mono-
substituted counterparts because of their bulkiness; therefore, their desymmetriza-
tion is often challenging.
Desymmetrizationofdisubstitutedpropanediol211(Scheme8.144)wasenvisagedas
a convenient strategy for the synthesis of gem-difluorinated cyclopropane-1-carboxylic
F F
F OH F OH
BCL,vinyl acetate
benzene/iPr 2O
HO 20:1 AcO
2 11 21 2
91%, e.e. = 96%
Scheme 8.144 Desymmetrization of a gem-difluorinated cyclopropane-diol.
acids, which represent attractive synthetic targets [364]. BCL allowed the conversion of
211 into the (R)-monoacetate 212 in96% yield and 91% e.e. The best performances were
observed in a 20 : 1 mixture of benzene–diisopropyl ether.
The enzymatic transesterification was also found to allow the construction of chiral
benzylic quaternary centers in good to excellent enantioselectivity [365]. The resulting
monoacetates 213 and 214, key intermediates for the synthesis of ()-aphanorphine
and ( þ )-eptazocine, respectively, were obtained (Scheme 8.145). The monoacylation
was catalyzed by BCL immobilized on Hyflo Super Cell (PSL-HSC) using isopropenyl
and vinyl acetate as acyl donors.
BCL, MTBE
O OH O OAc
isopropenyl acetate
OH OH
213, 85%, ee = 71%
BCL, Et 3 N
O O
vinyl acetate, MTBE
OH OH OAc OH
214, 89%, ee = 93.2%
Scheme 8.145 Introducing chirality into quaternary benzylic centers.
Acylation of the chromanedimethanol derivative 215, an intermediate for the

synthesis of a-tocotrienol, with vinyl acetate and CAL-B afforded the corresponding
(S)-monoacetate 216 in high enantiomeric purity (Scheme 8.146) [366]. Significant
amounts of diacetate were isolated as well.
The enzymatic desymmetrization of a 2-isopropenylpropane-1,2,3-triol represents
the key step in the formal synthesis of phosphonotrixin (Scheme 8.147), an herbicidal
antibiotic characterized by a unique structure bearing a CP bond on an isoprene
unit [367]. The monoacylation was performed in the presence of PPL using vinyl
acetate and afforded the desired product in 88% yield and 93% e.e., while giving only
a small percentage of diacetate.
344
HO
OAc
O
OH
HO 216, 60%, ee = 98%
CAL-B, Et2 O
+
OH
O vinyl acetate
HO
OH
215 OAc
O
OAc
27%
Scheme 8.146 Desymmetrization as a key step in the synthesis of a-tocotrienol.
OH OH OH
PPL
+
vinyl acetate
OH OH OAc OH OAc OAc
88%, ee = 93%
7%
O
OH
OH P O
HO OH
phosphonotrixin
Scheme 8.147 Desymmetrization of an intermediate for phosphonotrixin.
Similarly, a stereoselective acylation represented the crucial step for the chemoen-
zymatic synthesis of the pheromone ()-frontalin (Scheme 8.148) [368]. The desired
monoacetate was recovered in 45% yield and 90% e.e. after acylation mediated by
PFL, together with 55% of the corresponding achiral diacetate.
Vinyl and isopropenyl acetate are the most frequently used acyl donors for kinetic
resolutions and desymmetrizations. However, they sometimes display poor reactivity
towards sterically demanding substrates. A viable alternative in such cases is 1-
ethoxyvinyl 2-furoate [369]. This reagent was tested in the acylation of a series of
prochiral 1,3-diols having a chiral quaternary carbon (Scheme 8.149 and Table 8.6). In
the presence of CRL (MY and immobilized on Hyflo Super Cell), all substrates were
effectively converted in 82–97% e.e. and no tendency to acyl group migration and
subsequent racemization was observed, not even under acidic conditions.
8.7.2.3 Desymmetrization of 1,3,5-Triol Derivatives

In searching for an effective approach to chiral building blocks, K€
ohler et al. [370]
investigated the enzymatic desymmetrization of a prochiral pentane-1,3,5-triol
OH
O O
OAc
OH
OH
218 , 45%, ee = 90%
O O PFL, vinyl acetate
OH +
benzene
217 OH OAc
O O
OAc
OH
219, 55%
Scheme 8.148 Chemoenzymatic step in the synthesis of (–)-frontalin.
O
O O
EtO O
R2 R2
O
R1 OH R1 O
lipase, wet iPr2 O
OH OH
Scheme 8.149 Desymmetrization using 1-ethoxyvinyl 2-furoate as acyl donor.
derivative (Scheme 8.150). Among the different enzymes screened, PFL, BCL, and
RML gave the best enantioselectivities, while CAL-B promoted a fast but not selective
acylation with large amounts of diacetate formed. Subsequent optimization with BCL
immobilized on ceramic particles (Amano PS-CII) allowed the preparation of (S)-221
with >99.9% e.e. at 10 C.
8.7.3
Desymmetrization of meso-Diols
This section discusses the desymmetrization of primary and secondary cyclic meso-
diols (Scheme 8.151).
8.7.3.1 Desymmetrization of Primary Cyclic meso-Diols

Many natural products, often showing biological properties, bear a tetrahydropyranyl
moiety. Among the different strategies available for the preparation of such com-
pounds, the desymmetrization of meso-diols is a very effective one, as it allows us to
reveal multiple stereogenic centers in one operation.
Candy et al. have investigated the scope of the RML-catalyzed desymmetrization of
the diols depicted in Scheme 8.152 [371].
Interestingly, the C4 position was found to exert little or no effect on the yield and
the selectivity: comparable e.e. values were obtained for differently protected OH
groups (entries 1–4, Table 8.7), and even inversion of the configuration at this
346
Table 8.6 Desymmetrization using 1-ethoxyvinyl 2-furoate (Scheme 8.149).
Diol Lipase Time (h) E.e. (%) Yield (%)
OH CRL (lipase MY) 5 81 69

OMe
OH
CRL (lipase MY) 30 96 21

OH
OH

Cl
OH
OH

OH
MeO
Ph
R OH
OH
R ¼ Me CRL (HSC) 5 >99 66

R ¼ Et CRL (HSC) 2 89 87
R ¼ allyl CRL (HSC) 2 82 72
HO OH BCL, MTBE HO OAc

iso propenyl acetate
OSiMe2 Ph OSiMe2 Ph
-10°C
220 221, 52%, ee > 99.9%
Scheme 8.150 Desymmetrization of a 1,3,5-triol derivative.

desymmetrization
of meso- diols
OH R'
O
O R' O
n OH
n
O
primary cyclic secondary cyclic
meso- diols meso- diols
Scheme 8.151 Overview of the desymmetrization of meso-diols.
R4 R3 R4 R3
R2 R2 R2 R2
RML
HO OH AcO OH
O vinyl acetate O
R1 R1 R1 R1
iPr 2O
Scheme 8.152 Desymmetrization of tetrahydropyranyl diols.
carbon (entry 5), or the total reduction of this position (entry 6), had no significant
effect on the e.e. Replacing the methyl group at C3 for an ethyl (entry 7) resulted
only in a slight decrease of the enantioselectivity, but substitution by a phenyl group
caused a dramatic drop of reactivity and selectivity (entry 8). Satisfactory yield (72%)
and good e.e. (94%) were observed for the fully substituted tetrahydropyranyl diol in
entry 9.
Table 8.7 Scope study for the desymmetrization of tetrahydropyranyl diols (see Scheme 8.152).
Entry R1 R2 R3 R4 Yield (%) E.e. (%)
1 H Me H OMe 90 >98
2 H Me H OTBS 89 >98
3 H Me H OBn 88 >98
4 H Me H OBz 89 >98
5 H Me OMe H 58 95
6 H Me H H 67 96
7 H Et H OMe 82 96
8 H Ph H OMe 21 13
9 Me Me Me OH 72 94
348
8.7.4
Desymmetrization of Secondary Cyclic meso-Diols
meso-Diols having secondary hydroxyl groups, such as cyclopentane, cyclohexane,

and cycloheptane diols, are widely used as substrates for enzyme-mediated desym-
metrizations. Indeed, the corresponding enantiomerically pure monoesters repre-
sent valuable building blocks, often used for the synthesis of natural compounds,
chiral transition metal complexes, and pharmaceuticals.
One such example is given by monoacetate 223, key intermediate in the total
synthesis of entecavir, which has been approved for the treatment of hepatitis B viral
infection (Scheme 8.153) [372]. This monoacetate was obtained in 80% yield and 98%
e.e. from the corresponding diol by using immobilized BCL (Lipase PS-30).
BnO BnO
HO OH AcO OH
BCL
222 isopropenyl acetate 223

heptane/MTBE 10:1 80%, ee = 98%
N
N O
HO
N NH
HO
NH 2
entecavir
Scheme 8.153 Desymmetrization of an intermediate for entecavir.
Monoacetates of diol 224 (Figure 8.10) are also attractive starting materials for the
synthesis of bioactive products, for example, prostaglandins. For this reason, many
efforts have been devoted to the enzymatic desymmetrization of 224.
HO HO HO HO
O O
HO HO HO HO
224 225 226 227
HO HO OH
O
O
HO HO OH
228 229 230
Figure 8.10 Secondary cyclic meso-diols.

Table 8.8 Desymmetrization of diols 224–230 (Figure 8.10).
Entry Substrate Acyl donor/solvent Lipase Yield (%) E.e. (%) Configuration
1 224 TCAa)/THF–NEt3 PPL 48 >99 (S)-OAc

2 224 TCBb)/THF–NEt3 PPL 51 99 (S)-OAcyl
3 224 TCOc)/THF–NEt3 PPL 53 80 (S)-OAcyl
4 224 VAd)/THF–NEt3 PPL 65 >99 (S)-OAc
5 224 VBe)/THF–NEt3 PPL 55 93 (S)-OAcyl
6 224 VA/THF–NEt3 Mucor sp. 85 94 (S)-OAc
7 224 TCA/Py PPL 50 >99 (S)-OAc
8 224 IPAf)/t-BuOMe CAL-B 48 >99 (S)-OAc
9 225 VA/THF CAL-A/CAL-B 82 >99 (S)-OAc
10 226 VA/THF Mucor sp. 60 82 (S)-OAc
11 227 VA/THF PPL 75 86 (S)-OAc
12 228 VA/THF PPL 94 >99 (S)-OAc
13 229 VA/THF–NEt3 Various — — —
14 230 VA CRL 81 98 (R)-OAc
a) TCA ¼ 2,2,2-trichloroethyl acetate.

b) TCB ¼ 2,2,2-trichloroethyl butanoate.
c) TCO ¼ 2,2,2-trichloroethyl octanoate.
d) VA ¼ vinyl acetate.
e) VB ¼ vinyl butanoate.
f) IPA ¼ isopropenyl acetate.
Table 8.8 summarizes the results of the screening of different enzymes for the
monoacylation of 224 and related compounds (Figure 8.10) [328].
In most cases, the monoacylated products were obtained with high enantiomeric
purity. Interestingly, the isosteric compounds 226 and 228 gave remarkably different
e.e.s under otherwise identical conditions, thus suggesting a prominent role of the
electronic effects over the steric interactions. All tested lipases failed in converting the
bicyclic compound 229, while tricyclic 230 proved to be a very good substrate for CRL.
OH OH OH
O OH
O
OH
OH OH OH
231 232 233 234
OH OH OH OH
O O TBDMSO TBDMSO
OH OH OH OH
235 236 237 238
Figure 8.11 Six- and seven-membered ring diols.

350
Table 8.9 Desymmetrization of six- and seven-membered ring compounds (Figure 8.11).
Entry Substrate Acyl donor/solvent Lipase Yield (%) E.e. (%) Configuration
1 231 IPA BCL 51 95 (S)-OAc

2 231 IPA/t-BuOMe CAL-B 25 59 (R)-OAc
3 232 IPA BCL 90 >95 (R)-OAc
4 233 AcOMe PPL Not determined 84 Not determined
5 234 IPA/t-BuOMe CAL-B 81 >99 (R)-OAc
6 235 IPA/t-BuOMe BCL 46 >98 (S)-OAc
7 236 IPA BCL 92 84 (R)-OAc
8 237 IPA BCL 95 >95 (S)-OAc
9 238 IPA BCL 95 >99 (S)-OAc
The desymmetrization of the six- and seven-membered ring compounds depicted

in Figure 8.11 is listed in Table 8.9 [328].
Compounds 235 and 236 showed opposite stereoselectivities, giving the (S)- and
the (R)-monoester, respectively. In contrast, the two diastereoisomers 237 and 238
were acylated by BCL with the same stereochemical preference.
8.8
Non-stereoselective Formation of (Fatty Acid-Based) Esters
The non-stereoselective synthesis of fatty acid-based esters is of interest in the

chemical industry due to, for example, the application of such types of compounds as
soaps, skin creams, lubricants, and greases [373]. The synthetic concept is based on
the use of the corresponding long-chain fatty acid and alcohol, in the presence of a
lipase under neat conditions. The avoidance of additional organic solvents is a
particular advantage of this process technology, which has already found technical
applications. An early example was reported by Unichema Chemie researchers with
the synthesis of isopropyl palmitate (241) [373–375]. A lipase from Candida antarctica
was used as a biocatalyst, which furnished the desired product in an excellent yield of
99% and with a high chemical purity of >99% (Scheme 8.154).
lipase from
O OH Candida antarctica O CH 3
H 3C H 3C
14
OH H 3C CH 3 - H 2O 14
O CH 3
239 240 241

800 g/l 99% yield
>99% purity
Scheme 8.154 Enzymatic synthesis of isopropyl palmitate.

References j351
O lipase from O
Candida antarctica
H3 C CH 3 H 3C CH3
12
OH HO 12
O 12
12
- H2 O
2 42 243 24 4
STY: 6731 g(l*d)
Scheme 8.155 Enzymatic synthesis of myristyl myristate.
A recent example in this field has been reported by the Liese group jointly with an
Evonik Goldschmidt researcher, demonstrating the usefulness of a new reactor
concept for the synthesis of myristyl myristate (244,Scheme 8.155) [376]. By means of
a reactor that includes a bubble column that prevents mechanical erosion of the
immobilized lipase from Candida antarctica B caused by mechanical stirring of the
reaction mixture, myristyl myristate has been efficiently prepared with an impressive
space–time yield of 6731 g l1 day1. Furthermore, this reactor concept turned out to
be suitable for the solvent-free esterification of polyglycerol-3 with lauric acid to yield
polyglycerol-3 laurate at a high space–time yield of 3042 g l1 day1. The high
viscosity of this reaction medium has been a particular challenge. A further
successfully synthesized product was PEG-55-propylene glycol dioleate with a
space–time yield of 738 g l1 day1.
References
1 Nelson, D.L. and Cox, M.M. (2008) 10 Busto, E., Gotor-Fernandez, V., and
Lehninger Principles of Biochemistry, 5th Gotor, V. (2010) Chem. Soc. Rev., 39,
edn, W. H. Freeman and Company, 4504–4523.
New York. 11 Simons, C., van Leeuwen, J.G.E.,
2 Dewick, P.M. (2006) Essentials of Organic Stemmer, R., Arends, I.W.C.E.,
Chemistry, John Wiley & Sons, Ltd, Maschmeyer, T., Sheldon, R.A., and
Chichester. Hanefeld, U. (2008) J. Mol. Catal. B:
3 Bornscheuer, U.T. and Kazlauskas, R.J. Enzym., 54, 67–71.
(2006) Hydrolases in Organic Synthesis, 12 Kourist, R., Bartsch, S., Fransson, L.,
2nd edn, Wiley-VCH Verlag GmbH, Hult, K., and Bornscheuer, U.T. (2008)
Weinheim. ChemBioChem, 9, 67–69.
4 Faber, K. and Riva, S. (1992) Synthesis, 13 Kanerva, L.T. and Liljeblad, A. (2010) in
895–910. Encyclopedia of Catalysis (editor in chief
5 Hanefeld, U. (2003) Org. Biomol. Chem., I.T. Horvath), John Wiley and Sons, Ltd,
1, 2405–2415. pp. 1–25. Doi: 10.1002/
6 Chenevert, R., Pelchat, N., and Jacques, 0471227617.eoc197.
F. (2006) Curr. Org. Chem., 10, 14 Schmid, R.D. and Verger, R. (1998)
1067–1094. Angew. Chem. Int. Ed., 37, 1608–1633.
7 Kastle, J.H. and Loevenhart, A.S. (1900) 15 Verger, R. (1997) Trends Biotechnol., 15,
Am. Chem. J., 24, 491–525. 32–38.
8 Liljeblad, A. and Kanerva, L.T. (2006) 16 Sym, E.A. (1930) Biochem. J., 24,
Tetrahedron, 62, 5831–5854. 1265–1281.
9 Gotor-Fernandez, V. and Gotor, V. (2009) 17 Holwerda, K., Verkade, P.E., and de
Curr. Opin. Drug Discovery Dev., 12, Willigen, A.H.A. (1936) Rec. Trav. Chim.
784–797. Pays-Bas, 55, 43–57.
352
18 Schønheyder, F. and Volqvartz, K. (1945) 39 Cantone, S., Hanefeld, U., and Basso, A.
Acta Physiol. Scand., 9, 57–67. (2007) Green Chem., 9, 954–971.
19 Chapus, C., Semeriva, M., Bovier- 40 van Rantwijk, F. and Sheldon, R.A. (2007)
Lapierre, C., and Desnuelle, P. (1976) Chem. Rev., 107, 2757–2785.
Biochemistry, 15, 4980–4987. 41 Carrea, G. and Riva, S. (eds) (2008)
20 Louwrier, A., Drtina, G.J., and Organic Synthesis with Enzymes in
Klibanov, A.M. (1996) Biotechnol. Bioeng., Non-Aqueous Media, Wiley-VCH Verlag
50, 1–5. GmbH, Weinheim.
21 Salis, A., Svensson, I., Monduzzi, M., 42 Secundo, F. and Carrea, G. (2003) Chem.
Solinas, V., and Adlercreutz, P. (2003) Eur. J., 9, 3194–3199.
Biochim. Biophys. Acta, 1646, 145–151. 43 Perez, M., Sinisterra, J.V., and
22 Basso, A., Braiuca, P., Cantone, S., Hernaiz, M.J. (2010) Curr. Org. Chem., 14,
Ebert, C., Linda, P., Spizzo, P., Caimi, P., 2366–2383.
Hanefeld, U., Degrassi, G., and 44 Hanefeld, U., Gardossi, L., and
Gardossi, L. (2007) Adv. Synth. Catal., Magner, E. (2009) Chem. Soc. Rev., 38,
349, 877–886. 453–468.
23 Larsen, M.W., Zielinska, D.F., 45 Palomo, J.M. (2009) Curr. Org. Synth., 6,
Martinelle, M., Hidalgo, A., Jensen, L.J., 1–14.
Bornscheuer, U.T., and Hult, K. (2010) 46 Hartmann, M. and Jung, D. (2010)
ChemBioChem, 11, 796–801. J. Mater. Chem., 20, 844–857.
24 Kvittingen, L., Sjursnes, B., 47 Cao, L. (2005) Carrier-Bound Immobilized
Anthonsen, T., and Halling, P. (1992) Enzymes, Wiley-VCH Verlag GmbH,
Tetrahedron, 48, 2793–2802. Weinheim.
25 Bourquelot, E. and Bridel, M. (1913) Ann. 48 Persson, M., Wehtje, E., and
Chim. Phys., 29, 145–218. Adlercreutz, P. (2002) ChemBioChem, 3,
26 Sym, E.A. (1931) Biochem. Z., 230, 19–50. 566–571.
27 Sym, E.A. (1933) Biochem. Z., 258, 49 €
Frykman, H., Ohrner, N., Norin, T., and
304–324. Hult, K. (1993) Tetrahedron Lett., 34,
28 Sym, E.A. (1933) Biochem. Z., 262, 1367–1370.
406–424. 50 Bianchi, D., Cesti, P., and Battistel, E.
29 Sym, E.A. (1936) Biochem. J., 30, 609–617. (1988) J. Org. Chem., 53, 5531–5534.
30 Sym, E.A. and Swiatkowska, W. (1937) 51 Schmid, A., Dordick, J.S., Hauer, B.,
Enzymologia, 2, 79–80. Kiener, A., Wubbolts, M., and Witholt, B.
31 Sym, E.A. (1937) Enzymologia, 2, (2001) Nature, 409, 258–268.
107–109. 52 Schudok, M. and Kretzschmar, G. (1997)
32 Klibanov, A.M. (2001) Nature, 409, Tetrahedron Lett., 38, 387–388.
241–246. 53 Degueil-Castaing, M., De Jeso, B.,
33 Klibanov, A.M. (1989) Trends Biochem. Drouillard, S., and Maillard, B. (1987)
Sci., 14, 141–144. Tetrahedron Lett., 28, 953–954.
34 Zaks, A. and Klibanov, A.M. (1984) 54 Sweers, H.M. and Wong, C.-H. (1986)
Science, 224, 1249–1251. J. Am. Chem. Soc., 108, 6421–6422.
35 Cambou, B. and Klibanov, A.M. (1984) 55 Li, Y.-X., Straathof, A.J.J., and
J. Am. Chem. Soc., 106, 2687–2692. Hanefeld, U. (2002) Tetrahedron:
36 Effenberger, F., Ziegler, T., and F€orster, S. Asymmetry, 13, 739–743.
(1987) Angew. Chem. Int. Ed. Engl., 26, 56 Hara, P., Hanefeld, U., and Kanerva, L.T.
458–460. (2009) Green Chem., 11, 250–256.
37 Musa, M.M., Ziegelmann-Fjeld, K.I., 57 Ema, T. (2004) Curr. Org. Chem., 8,
Vieille, C., Zeikus, J.G., and Phillips, R.S. 1009–1025.
(2007) Angew. Chem. Int. Ed., 46, 58 Kazlauskas, R.J., Weissfloch, A.N.E.,
3091–3094. Rappaport, A.T., and Cuccia, L.A. (1991)
38 Paravidino, M., Sorgedrager, M.J., J. Org. Chem., 56, 2656–2665.
Orru, R.V.A., and Hanefeld, U. (2010) 59 Kazlauskas, R.J. and Weissfloch, A.N.E.
Chem. Eur. J., 16, 7596–7604. (1997) J. Mol. Catal. B: Enzym., 3, 65–72.
References j353
60 Savile, C.K. and Kazlauskas, R.J. (2005) 77 Veum, L., Kanerva, L.T., Halling, P.J.,
J. Am. Chem. Soc., 127, 12228–12229. Maschmeyer, T., and Hanefeld, U. (2005)
61 Magnusson, A.O., Rotticci-Mulder, J.C., Adv. Synth. Catal., 347, 1015–1021.
Santagostino, A., and Hult, K. (2005) 78 Hietanen, A., Ekholm, F.S., Leino, R.,
ChemBioChem, 6, 1051–1056. and Kanerva, L.T. (2010) Eur. J. Org.
62 Leonard, V., Fransson, L., Lamare, S., Chem., 6974–6980.
Hult, K., and Graber, M. (2007) 79 Mitsuda, S., Umemura, T., and
ChemBioChem, 8, 662–667. Hirohara, H. (1988) Appl. Microbiol.
63 Tuomi, W.V. and Kazauskas, R.J. (1999) Biotechnol., 29, 310–315.
J. Org. Chem., 64, 2638–2647. 80 Danda, H., Furukawa, Y., and
64 Mezzetti, A., Schrag, J.D., Cheong, C.S., Umemura, T. (1991) Synlett, 441–442.
and Kazlauskas, R.J. (2005) Chem. Biol., 81 Danda, H., Nagatomi, T., Maehara, A.,
12, 427–437. and Umemura, T. (1991) Tetrahedron, 47,
65 For a review of the enzymatic synthesis of 8701–8716.
chiral tertiary alcohols, see: Kourist, R., 82 (a) For comprehensive overviews, see:
Dominguez de Maria, P., and Gais, H.-J. and Theil, F. (2002) in Enzymes
Bornscheuer, U.T. (2008) ChemBioChem, in Organic Synthesis, 2nd edn, vol. 2 (eds
9, 491–498. K. Drauz and H. Waldmann), Wiley-VCH
66 Nguyen, G.-S., Kourist, R., Paravidino, Verlag GmbH, Weinheim, pp. 335–578;
M., Hummel, A., Rehdorf, J., Orru, (b) Koskinen, A.M.P. and Klibanov, A.M.
R.V.A., Hanefeld, U., and Bornscheuer, (eds) (1996) Enzymatic Reactions in
U.T. (2010) Eur. J. Org. Chem., Organic Media, Blackie Academic &
2753–2758. Professional, Glasgow.
67 Ahmed, S.N., Kazlauskas, R.J., 83 Smeets, J.W.H. and Kieboom, A.P.G.
Morinville, A.H., Grochulski, P., (1992) Recl. Trav. Chim. Pays-Bas, 111,
Schrag, J.D., and Cygler, M. (1994) 490–495.
Biocatalysis, 9, 209–225. 84 Stahly, G.P. and Starrett, R.M. (1997) in
68 Franssen, M.C.R., Jongejan, H., Chirality in Industry II (eds A.N. Collins,
Kooijman, H., Spek, A.L., Camacho G.N. Sheldrake, and J. Crosby), John
Mondril, N.L.F.L., Boavida dos Santos, Wiley & Sons, Ltd, Chichester, pp. 19–40.
P.M.A.C., and de Groot, A. (1996) 85 Rasor, J.P. and Voss, E. (2001) Appl. Catal.
Tetrahedron: Asymmetry, 7, 497–510. A: General, 221, 145–158.
69 Tawaki, S. and Klibanov, A.M. (1992) 86 Wu, S.-H., Guo, Z.-W., and Sih, C.J.
J. Am. Chem. Soc., 114, 1882–1884. (1990) J. Am. Chem. Soc., 112, 1990–1995.
70 Sharma, A. and Chattopadhyay, S. (2000) 87 Allenmark, S. and Ohlsson, A. (1992)
J. Mol. Catal. B: Enzym., 10, 531–534. Biocatalysis, 6, 211–221.
71 Steinreiber, J., Faber, K., and Griengl, H. 88 Colton, I.J., Ahmed, S.N., and
(2008) Chem. Eur. J., 14, 8060–8072. Kazlauskas, R.J. (1995) J. Org. Chem., 60,
72 Inagaki, M., Hiratake, J., Nishioka, T., 212–217.
and Oda, J. (1991) J. Am. Chem. Soc., 113, 89 Engstr€om, K., Nyhlen, J., Sandstr€om,
9360–9361. A.G., and B€ackvall, J.-E. (2010) J. Am.
73 Dinh, P.M., Howarth, J.A., Hudnott, Chem. Soc., 132, 7038–7042.
A.R., Williams, J.M.J., and Harris, W. 90 Tan, D.S., G€ unter, M.M., and
(1996) Tetrahedron Lett., 37, Drueckhammer, D.G. (1995) J. Am.
7623–7626. Chem. Soc., 117, 9093–9094.
74 Pellissier, H. (2008) Tetrahedron, 64, 91 Chang, C.-S., Tsai, S.-W., and
1563–1601. Kuo, J. (1999) Biotechnol. Bioeng., 64,
75 Glueck, S.M., Larissegger-Schnell, B., 120–126.
Csar, K., Kroutil, W., and Faber, K. (2005) 92 Lu, C.-H., Cheng, Y.-C., and Tsai, S.-W.
Chem. Commun., 1904–1905. (2002) Biotechnol. Bioeng., 79, 200–210.
76 Kim, M.-J., Choi, M.Y., Han, M.Y., Choi, 93 Chen, C.-H., Cheng, Y.-C., and Tsai, S.-W.
Y.K., Lee, J.K., and Park, J. (2002) J. Org. (2002) J. Chem. Technol. Biotechnol., 77,
Chem., 67, 9481–9483. 699–705.
354
94 Lin, H.-Y. and Tsai, S.-W. (2003) J. Mol. Landfester, K. (2006) Angew. Chem., 118,
Catal. B: Enzym., 24–25, 111–120. 1676–1679; (2006) Angew. Chem. Int. Ed.,
95 Fazlena, H., Kamaruddin, A.H., and 45, 1645–1648.
Zulkali, M.M.D. (2006) Bioprocess Biosyst. 111 Martinez, C.A., Yazbeck, D.R., and Tao, J.
Eng., 28, 227–233. (2004) Tetrahedron, 60, 759–764.
96 Pietruszka, J., Simon, R.C., Kruska, F., 112 Review: Tao, J., Zhao, L., and Ran, N.
and Braun, M. (2010) Eur. J. Org. Chem., (2007) Org. Process Res. Dev., 11, 259–267.
6217–6224. 113 Yazbeck, D., Derrick, A., Panesar, M.,
97 Shafiee, A., Upadhyay, V., Corley, E., Deese, A., Gujral, A., and Tao, J. (2006)
Biba, M., Zhao, D., Marcoux, J.-F., Org. Process Res. Dev., 10, 655–660.
Campos, K., Journet, M., King, A., 114 Hu, S., Martinez, C., Kline, B.,
Larsen, R., Grabowski, E., Volante, R., Yazbeck, D., Tao, J., and Kucera, D. (2006)
and Tillyer, R. (2005) Tetrahedron: Org. Process Res. Dev., 10, 650–654.
Asymmetry, 16, 3094–3098. 115 Arends, I.W.C.E., Sheldon R.A., and
98 Truppo, M.D., Journet, M., Shafiee, A., Hanefeld, U. (2007) Green Chemistry and
and Moore, J.C. (2006) Org. Process Res. Catalysis, Wiley-VCH Verlag GmbH,
Dev., 10, 592–598. Weinheim, ch. 6.
99 Patel, R.N. (2008) Coord. Chem. Rev., 252, 116 Park, O.-J., Lee, S.-H., Park, T.-Y.,
659–701. Chung, W.-G., and Lee, S.-W. (2006) Org.
100 Elferink, V.H.M., Kierkels, J.G.T., Process Res. Dev., 10, 588–591.
Kloosterman, M., and Roskam, J.H. 117 Riermeier, T., Dingerdissen, U.,
(1990) EP 369553. Gross, P., Holla, W., Beller, M., and
101 Arends, I.W.C.E., Sheldon, R.A., and Schichl, D. (2001) DE19955283.
Hanefeld, U. (2007) Green Chemistry and 118 Zimmermann, V., Beller, M., and
Catalysis, Wiley-VCH Verlag GmbH, Kragl, U. (2006) Org. Process Res. Dev., 10,
Weinheim, ch. 6. 622–627.
102 Hermann, M., Kietzmann, M.U., 119 Pugniere, M., Commeyras, A., and
Ivancic, M., Zenzmaier, C., Luiten, Previero, A. (1983) Biotechnol. Lett., 5,
R.G.M., Skranc, W., Wubbolts, M., 447–452.
Winkler, M., Birner-Gruenberger, R., 120 Chen, S.-T., Huang, W.-H., and
Pichler, H., and Schwab, H. (2008) Wang, K.-T. (1994) J. Org. Chem., 59,
J. Biotechnol., 133, 301–310. 7580–7581.
103 Cohen, S.G. and Milovanovic, A. (1968) 121 Papageorgiou, C. and Benezra, C. (1985)
J. Am. Chem. Soc., 90, 3495–3502. J. Org. Chem., 50, 1144–1145.
104 Guibe-Jampel, E., Rousseau, G., and 122 Hof, R.P. and Kellogg, R.M. (1995) J.
Salaun, J. (1987) J. Chem. Soc., Chem. Chem. Soc., Perkin Trans 1, 1247–1249.
Commun., 1080–1081. 123 Ribeiro, C.M., Passaroto, E.N., and
105 Klunder, A.J.H., van Gastel, F.J.C., and Brenelli, E.C.S. (2001) Tetrahedron Lett.,
Zwanenburg, B. (1988) Tetrahedron Lett., 42, 6477–6479.
29, 2697–2700. 124 Rawale, S., Hrihorczuk, L.M., Wei, W.Z.,
106 Klunder, A.J.H., Huizinga, W.B., Sessink, and Zemlicka, J. (2002) J. Med. Chem., 45,
P.J.M., and Zwanenburg, B. (1987) 937–943.
Tetrahedron Lett., 28, 357–360. 125 Scheid, G., Ruijter, E., Bessler, M.K.,
107 Review: Miyazawa,T. (1999) Amino Acids, Bornscheuer, U.T., and Wessjohann, L.A.
16, 191–213. (2004) Tetrahedron: Asymmetry, 15,
108 Ng-Youn-Chen, M.C., Serreqi, A.N., 2861–2869.
Huang, Q., and Kazlauskas, R.J. (1994) J. 126 Review of biocatalytic approaches
Org. Chem., 59, 2075–2081. towards mandelic acid and derivatives:
109 Hu, S., Martinez, C.A., Kline, B., Gr€oger, H. (2001) Adv. Synth. Catal., 343,
Yazbeck, D., Tao, J., and Kucera, D.J. 547–558.
(2006) Org. Process Res. Dev., 10, 650–654. 127 Wong, C.H., Chen, S.-T., Hennen, W.J.,
110 Gr€oger, H., May, O., H€ usken, H., Bibbs, J.A., Wang, Y.-E., Liu, J.L.-C.,
Georgeon, S., Drauz, K., and Pantoliano, M.W., Whitlow, M., and
References j355
Bryan, P.N. (1990) J. Am. Chem. Soc., 112, 144 Gr€
oger, H. and Werner, H. (2005)
945–953. (Degussa AG), US Patent
128 Fuganti, C., Rosell, C.M., Servi, S., 6869781.
Tagliani, A., and Terreni, M. (1992) 145 Yamamoto, Y., Miyata, H., Konegawa, T.,
Tetrahedron: Asymmetry, 3, 383–386. and Sakata, K. (2006) (Ube Industries,
129 Moorlag, H., Kellogg, R.M., Ltd), US Pat. Appl.
Kloosterman, M., Kaptein, B., and 20060178433.
Schoemaker, H.E. (1990) J. Org. Chem., 146 Atkins, R.J., Banks, A., Bellingham, R.K.,
55, 5878–5881. Breen, G.F., Carey, J.S., Etridge, S.K.,
130 Moorlag, H. and Kellogg, R.M. (1991) Hayes, J.F., Hussain, N., Morgan, D.O.,
Tetrahedron: Asymmetry, 2, 705–720. Oxley, P., Passey, S.C., Walsgrove, T.C.,
131 Berger, B., de Raadt, A., Griengl, H., and Wells, A.S. (2003) Org. Process Res.
Hayden, W., Hechtberger, P., Dev., 7, 663–675.
Klempier, N., and Faber, K. (1992) Pure 147 Roberts, N.J., Seago, A., Carey, J.S.,
Appl. Chem., 64, 1085–1088. Freer, R., Preston, C., and Lye, G.J. (2004)
132 Feichter, C., Faber, K., and Griengl, H. Green Chem., 6, 475–482.
(1991) J. Chem. Soc., Perkin Trans. 1, 148 Forro, E. and F€
ul€op, F. (2007) Chem. Eur.
653–654. J., 13, 6397–6401.
133 Guo, Z.-W. and Sih, C.J. (1989) J. Am. 149 Iding, H., Wirz, B., and Sarmiento, R.-
Chem. Soc., 111, 6836–6841. M.R. (2003) Tetrahedron: Asymmetry, 14,
134 Faber, K. (2004) Biotransformations in 1541–1545.
Organic Chemistry, 5th edn, Springer, 150 Santaniello, E., Ferrabosh, P., Grisenti,
Berlin, pp. 121–122. P., Manzocchi, A., and Trave, S. (1989)
135 Sih, C.J., Gu, Q.-M., Holdgr€ un, X., and Gazz. Chim. Ital., 119, 581–584.
Harris, K. (1992) Chirality, 4, 91. 151 Boaz, N.W. (1992) J. Org. Chem., 57,
136 Liese, A., Seelbach, K., and Wandrey, C. 4289–4292.
(2006) Industrial Biotransformations, 2nd 152 Ribeiro, C.M.R., Passaroto, E.N., and
edn, Wiley-VCH Verlag GmbH, Brenelli, E.C.S. (2001) J. Braz. Chem. Soc.,
Weinheim, pp. 306–310. 12, 742–746.
137 Furutani, T., Furui, M., Mori, T., and 153 Pesti, J.A., Yin, J., Zhang, L.-H., and
Shibatani, T. (1996) Appl. Biochem. Anzalone, L. (2001) J. Am. Chem. Soc.,
Biotechnol., 59, 319–328. 123, 11075–11076.
138 Pottie, M., van der Eycken, J., 154 Pesti, J.A., Yin, J., Zhang, L.-H.,
Vandevalle, M., Dewanckele, J.M., and Anzalone, L., Waltermire, R.E., Ma, P.,
R€oper, H. (1989) Tetrahedron Lett., 30, Gorko, E., Confalone, P.N., Fortunak, J.,
5319–5322. Silverman, C., Blackwell, J., Chung, J.C.,
139 Mohr, P., R€ osslein, L., and Tamm, C. Hrytsak, M.D., Cooke, M., Powell, L., and
(1987) Helv. Chim. Acta, 70, 142–152. Ray, C. (2004) Org. Process Res. Dev., 8
140 Suemune, H., Tanaka, M., Ohaishi, H., 22–27.
and Sakai, K. (1988) Chem. Pharm. Bull., 155 Burgess, K. and Henderson, I. (1989)
36, 15–21. Tetrahedron Lett., 30, 3633–3636.
141 Katayama, S., Ae, N., and Nagata, R. 156 Kielbasinski, P., Goralczyk, P.,
(1998) Tetrahedron: Asymmetry, 9, Mikolajczyk, M., Wieczorek, M.W., and
4295–4299. Wajzner, W.R. (1998) Tetrahedron:
142 Faulconbridge, S.J., Holt, K.E., Asymmetry, 9, 2641–2650.
Sevillano, L.G., Lock, C.J., Tiffin, P.D., 157 Hu, S., Kelly, S., Lee, S., Tao, J., and
Tremayne, N., and Winter, S. (2000) Flahive, E. (2006) Org. Lett., 8, 1653–1655.
Tetrahedron Lett., 41, 2679–2681. 158 Ramaswamy, S., Hui, R.A.H.F., and
143 Gr€oger, H., May, O., H€ usken, H., Jones, J.B. (1986) J. Chem. Soc., Chem.
Georgeon, S., Drauz, K., and Commun., 1545–1546.
Landfester, K. (2006) Angew. Chem., 118, 159 Pietzsch, M., Vielhauer, O.,
1676–1679; (2006) Angew. Chem. Int. Ed., Pamperin, D., Ohse, B., and Hopf, H.
45, 1645–1648. (1999) J. Mol. Cat. B: Enzym., 6, 51–57.
356
160 Carballeira, J.D., Krumlinde, P., 175 Oguzkaya, F., Sahin, E., and Tanyeli, C.
Bocola, M., Vogel, A., Reetz, M.T., and (2006) Tetrahedron: Asymmetry, 17,
B€ackvall, J.-E. (2007) Chem. Commun., 3004–3009.
1913–1915. 176 Coope, F. and Main, B.G. (1995)
161 Alcock, N.W., Crout, D.H.G., Tetrahedron: Asymmetry, 6, 1393–1398.
Henderson, C.M., and Thomas, S.E. 177 Kourist, R., Krishna, S.H., Patel, J.S.,
(1988) J. Chem. Soc., Chem. Commun., Bartnek, F., Hitchman, T., Weiner, D.P.,
746–747. and Bornscheuer, U.T. (2007) Org.
162 Dominguez de Maria, P., Biomol. Chem., 5, 3310–3313.
Garcia-Burgos, C.A., Bargeman, G., and 178 Holt, J., Arends, I.W.C.E., Minnaard, A.J.,
van Gemert, R.W. (2007) Synthesis, and Hanefeld, U. (2007) Adv. Synth.
1439–1452. Catal., 349, 1341–1344.
163 Musidlowska, A., Lange, S., and 179 Bartsch, S., Kourist, R., and Bornscheuer,
Bornscheuer, U.T. (2001) Angew. Chem., U.T. (2008) Angew. Chem., 120,
113, 2934–2936; (2001) Angew. Chem. Int. 1531–1534; (2008) Angew. Chem. Int. Ed.,
Ed., 40, 2851–2853. 47, 1508–1511.
164 Hummel, A., Br€ usehaber, E., B€ottcher, 180 Yang, X., Reinhold, A.R., Rosati, R.L., and
D., Trauthwein, H., Doderer, K., and Liu, K.K.-C. (2000) Org. Lett., 2,
Bornscheuer, U.T. (2007) Angew. Chem., 4025–4027.
119, 8644–8646. 181 Alfaro Blasco, M., Thumann, S.,
165 Heinze, B., Kourist, R., Fransson, L., Wittmann, J., Giannis, A., and Gr€oger, H.
Hult, K., and Bornscheuer, U.T. (2010) Bioorg. Med. Chem. Lett., 20,
(2007) Protein Eng. Design Select., 20, 4679–4682.
125–131. 182 Mizuguchi, E., Takemoto, M., and
166 Zheng, G.-W., Yu, H.-L., Zhang, J.-D., and Achiwa, K. (1993) Tetrahedron: Asymmetry,
Xu, J.-H. (2009) Adv. Synth. Catal., 351, 4, 1961–1964.
405–414. 183 Mizuguchi, E. and Achiwa, K. (1993)
167 For a review, see: Patel, R.N. (2008) Coord. Tetrahedron: Asymmetry, 4, 2303–2306.
Chem. Rev., 252, 659–701. 184 Cipiciani, A. and Belleza, F.J. (2002) J.
168 Hanson, R.L., Parker, W.L., Mol. Cat. B: Enzym., 17, 261–266.
Brzozowski, D.B., Tully, T.P., Kotnis, A., 185 Dominguez de Maria, P.,
and Patel, R.N. (2005) Tetrahedron: Garcia-Burgos, C.A., Bargeman, G., and
Asymmetry, 16, 2711–2716. van Gemert, R.W. (2007) Synthesis,
169 Patel, R., Banerjee, A., Ko, R., Howell, J., 1439–1452.
Li, W.-S., Comezoglu, F., Partyka, R., and 186 Molinari, F., Romano, D., Gandolfi, R.,
Szarka, L. (1994) Biotechnol. Appl. Gardossi, L., Hanefeld, U., Converti, A.,
Biochem., 20, 23–33. and Spizzo, P. (2009) in Modern
170 Patel, R., Howell, J., Chidambaram, R., Biocatalysis (eds W.-D. Fessner and T.
Benoit, S., and Kant, J. (2003) Tetrahedron: Anthonsen), Wiley-VCH Verlag GmbH,
Asymmetry, 14, 3673–3677. Weinheim.
171 Nanduri, V.B., Banerjee, A., Howell, J., 187 Garcia-Urdiales, E., Alfonso, I., and
Brzozowski, D., Eiring, R., and Patel, R.N. Gotor, V. (2005) Chem. Rev., 105, 313–354.
(2000) J. Ind. Microbiol. Biotechnol., 25, 188 Faber, K. (2004) Biotransformation in
171–175. Organic Chemistry, 5th edn, Springer,
172 Ghanem, A. and Schurig, V. (2003) Heidelberg.
Tetrahedron: Asymmetry, 14, 57–62. 189 Sheldon, R.A., Arends, I., and
173 Crout, D.H.G., Gaudet, V.S.B., Hanefeld, U. (2006) Green Chemistry and
Laumen, K., and Schneider, M.P. (1986) Catalysis, Wiley-VCH Verlag GmbH,
J. Chem. Soc., Chem. Commun., Weinheim.
808–810. 190 Schneider, M., Engel, N., and
174 Ganey, M.V., Padykula, R.E., and Boensmann, H. (1984) Angew. Chem., 96,
Berchtold, G.A. (1989) J. Org. Chem., 54, 54–55;(1984) Angew. Chem. Int. Ed. Engl.,
2787–2793. 23, 66.
References j357
191 Noguchi, N. and Nakada, M. (2006) Org. Szarka, L.J. (1998) J. Am. Oil Chem. Soc.,
Lett., 8, 2039–2042. 75, 1473–1482.
192 Iosub, V., Haberl, A.R., Leung, J., 207 Bjoerkling, F., Boutelje, J., Gatenbeck, S.,
Tang, M., Vembaiyan, K., Parvez, M., and Hult, K., and Norin, T. (1985) Tetrahedron
Back, T.G. (2010) J. Org. Chem., 75, Lett., 26, 4957–4958.
1612–1619. 208 Guanti, G., Banfi, L., Narisano, E.,
193 Dominguez de Maria, P., Kossmann, B., Riva, R., and Thea, S. (1986) Tetrahedron
Potgrave, N., Buchholz, S., Trauthwein, Lett., 27, 4639–4642.
H., May, O., and Gr€oger, H. (2005) Synlett, 209 Guanti, G., Banfi, L., and Narisano, E.
1746–1748. (1989) Tetrahedron Lett., 30, 2697–2698.
194 Buchholz, K., Kasche, V., and 210 Lam, L.K.P., Hui, R.A.H.F., and Jones,
Bornscheuer, U.T. (2005) Biocatalysts and J.B. (1986) J. Org. Chem., 51, 2047–2050.
Enzyme Technology, Wiley-VCH Verlag 211 Chartrain, M., Maligres, P., Cohen, D.,
GmbH, Weinheim. Upadhyay, V., Pecore, V., Askin, D., and
195 Luyten, M., M€ uller, S., Herzog, B., and Greasham, R. (1999) J. Biosci. Bioeng., 87,
Keese, R. (1987) Helv. Chim. Acta, 70, 386–389.
1250–1254. 212 Erbeldinger, M., Mesiano, A.J., and
196 Breznik, M. and Kikelj, D. (1997) Russell, A.J. (2000) Biotechnol. Progr., 16,
Tetrahedron: Asymmetry, 8, 425–434. 1129–1131.
197 Gaede, B.J. and Nardelli, C.A. (2005) Org. 213 Wallert, S., Drauz, K., Grayson, I., Gr€oger,
Process Res. Dev., 9, 23–29. H., Dominguez de Maria, P., and Bolm,
198 Farb, D. and Jencks, W.P. (1980) Arch. C. (2005) Green Chem., 7, 602–605.
Biochem. Biophys., 203, 214–226. 214 Sch€onherr, H., Mollitor, J., and
199 Lange, S., Musidlowska, A., Schmidt- Schneider, C. (2010) Eur. J. Org. Chem.,
Dannert, C., Schmitt, J., and 20, 3908–3918.
Bornscheuer, U.T. (2001) ChemBioChem, 215 Huang, F.-C., Lee, L.F.H., Mittal, R.S.D.,
2, 576–582. Ravikumar, P.R., Chan, J.A., Sih, C.J.,
200 Musidlowska, A., Lange, S., and Caspi, E., and Eck, C.R. (1975) J. Am.
Bornscheuer, U.T. (2001) Angew. Chem. Chem. Soc., 97, 4144–4145.
Int. Ed., 40, 2851–2853. 216 Kosmol, H., Kieslich, K., and Gibian, H.
201 B€ottcher, D., Br€
usehaber, E., Doderer, K., (1968) Justus Liebigs Ann. Chem., 711,
and Bornscheuer, U. (2007) Appl. 38–41.
Microbiol. Biotechnol., 73, 1282–1289. 217 Okuyama, K., Shingubara, K., Tsujiyama,
202 Hummel, A., Br€ usehaber, E., B€ottcher, S.-i., Suzuki, K., and Matsumoto, T.
D., Trauthwein, H., Doderer, K., and (2009) Synlett, 941–944.
Bornscheuer, U.T. (2007) Angew. Chem., 218 Mikolajczyk, M., Kielbasinski, P.,
119, 8644–8646; (2007) Angew. Chem. Int. Zurawinski, R., Wieczorek, M.W., and
Ed., 46, 8492–8494. Blaszczyk, J. (1994) Synlett, 127–129.
203 Br€usehaber, E., Schwiebs, A., Schmidt, 219 Chênevert, R., Jacques, F., Giguere, P.,
M., B€ ottcher, D., and Bornscheuer, U. and Dasser, M. (2008) Tetrahedron:
(2010) Appl. Microbiol. Biotechnol., 86, Asymmetry, 19, 1333–1338.
1337–1344. 220 Izquierdo, I., Plaza, M.T., Rodrıguez, M.,
204 Hermann, M., Kietzmann, M.U., Ivancic, and Tamayo, J. (1999) Tetrahedron:
M., Zenzmaier, C., Luiten, R.G.M., Asymmetry, 10, 449–455.
Skranc, W., Wubbolts, M., Winkler, M., 221 Morgan, B., Dodds, D.R., Zaks, A.,
Birner-Gruenberger, R., Pichler, H., and Andrews, D.R., and Klesse, R. (1997) J.
Schwab, H. (2008) J. Biotechnol., 133, Org. Chem., 62, 7736–7743.
301–310. 222 Chênevert, R., Levesque, C., and Morin,
205 Musidlowska-Persson, A. and P. (2008) J. Org. Chem., 73, 9501–9503.
Bornscheuer, U.T. (2003) Tetrahedron: 223 Donohoe, T.J., Rigby, C.L., Thomas, R.E.,
Asymmetry, 14, 1341–1344. Nieuwenhuys, W.F., Bhatti, F.L., Cowley,
206 Patel, R.N., Banerjee, A., Chu, L., A.R., Bhalay, G., and Linney, I.D. (2006) J.
Brozozowski, D., Nanduri, V., and Org. Chem., 71, 6298–6301.
358
224 Bianchi, D., Cesti, P., and Battistel, E. Srivastava, A.S., and Yang, M.G. (2005)
(1988) J. Org. Chem., 53, 5531–5534. PCT Int. Appl. WO/2005/021500 A1.
225 Wallace, J.S., Baldwin, B.W., and 242 Zutter, U., Iding, H., Spurr, P., and
Morrow, C. (1992) J. Org. Chem., 57, Wirz, B. (2008) J. Org. Chem., 73,
5231–5239. 4895–4902.
226 Griffith, D.A. and Danishefsky, S.J. (1996) 243 Papageorgiou, C. and Benezra, C. (1985)
J. Am. Chem. Soc., 118, 9526–9538. J. Org. Chem., 50, 1144–1145.
227 Danishefsky, S.J., Paz Cabal, M., and 244 Faber, K. (2004) Biotransformations in
Chow, K. (1989) J. Am. Chem. Soc., 111, Organic Chemistry, 5th edn, Springer,
3456–3457. Berlin, pp. 66–67.
228 Patel, R.N., Banerjee, A., Pendri, Y.R., 245 Schirmeister, T. and Otto, H.-H. (1993)
Liang, J., Chen, C.-P., and Mueller, R. J. Org. Chem., 58, 4819–4822.
(2006) Tetrahedron: Asymmetry, 17, 246 Review: Shimizu, S., Kataoka, M.,
175–178. Honda, K., and Sakamoto, K. (2001)
229 Rodrıguez-Rodrıguez, J.A., Brieva, R., J. Biotechnol., 92, 187–194.
and Gotor, V. (2010) Tetrahedron, 66, 247 Review: Shimizu, S. and Kataoka, M.
6789–6796. (1996) Chimia, 50, 409–410.
230 Patel, R.N., Banerjee, A., Pendri, Y.R., 248 Kataoka, M., Shimizu, K., Sakamoto, K.,
Liang, J., Chen, C.-P., and Mueller, R. Yamada, H., and Shimizu, S. (1995) Appl.
(2006) Tetrahedron: Asymmetry, 17, Microbiol. Biotechnol., 44, 333–338.
175–178. 249 Shimizu, S. and Kataoka, M. (1999) in
231 Goswami, A. and Kissick, T.P. (2009) Org. Encyclopedia of Bioprocess Technology:
Process Res. Dev., 13, 483–488. Fermentation, Biocatalysis, and
232 Gais, H.J., Lukas, K.L., Ball, W.A., Bioseparation (eds M.C. Flickinger and
Braun, S., and Lindner, H.J. (1986) Liebigs S.W. Drews), John Wiley & Sons, Inc.,
Ann. Chem., 687–716. New York, pp. 1571–1577.
233 Schneider, M., Engel, N., H€onicke, P., 250 Blanco, L., Guibe-Jampel, E., and
Heinemann, G., and G€orisch, H. (1984) Rouseeau, G. (1988) Tetrahedron Lett., 29,
Angew. Chem., 96, 55–56; (1984) Angew. 1915–1918.
Chem. Int. Ed. Engl., 23, 67–68. 251 Gutman, A.L., Zuobi, K., and Guibe-
234 Kobayashi, S., Kamiyama, K., Iimori, T., Jampel, K. (1990) Tetrahedron Lett., 31,
and Ohno, M. (1984) Tetrahedron Lett., 25, 2037–2038.
2557–2560. 252 Benvinakatti, H.S., Newadkar, R.V., and
235 Kurihara, M., Kamiyama, K., Banerji, A.A. (1990) J. Chem. Soc., Chem.
Kobayashi, S., and Ohno, M. (1985) Commun., 1091–1092.
Tetrahedron Lett., 26, 5831–5834. 253 Faber, K. (2004) Biotransformation in
236 Kobayashi, S., Kamiyama, K., and Ohno, Organic Chemistry, 5th edn, Springer,
M. (1990) Chem. Pharm. Bull., 38, Heidelberg, pp. 104–105.
350–354. 254 Crich, J.Z., Brieva, R., Marquart, P., Gu,
237 Sousa, H.A., Afonso, C.A.M., Mota, R.-L., Flemming, S., and Sih, C.J. (1993) J.
J.P.B., and Crespo, J.G. (2003) Ind. Eng. Org. Chem., 58, 3252.
Chem. Res., 42, 5516–5525. 255 Gu, R.-L., Lee, I.-S., and Sih, C.J. (1992)
238 Heiss, L. and Gais, H.J. (1995) Tetrahedron Tetrahedron Lett., 33, 1953–1956.
Lett., 36, 3833–3836. 256 Coletti-Previero, M.A., Kraicsovits, P.,
239 Cope, A.C. and Herrick, E.C. (1963) and Previero, A. (1973) FEBS Lett., 37,
Organic Syntheses, John Wiley 93–96.
and Sons, Inc., New York, Collective Vol. 257 Valivety, R.H., Johnston, G.A., Suckling,
4, p. 304. C.J., and Halling, P.J. (1991) Biotechnol.
240 Cherney, R.J., Carter, P., Duncia, J.V., Bioeng., 38, 1137–1143.
Gardner, D.S., and Santella, J.B. (2004) 258 Zaks, A. and Klibanov, A.M. (1988) J. Biol.
PCT Int. Appl. WO/2004/071460 A2. Chem., 263, 3194–3201.
241 Carter, P.H., Cherney, R.J., Batt, D.G., 259 Zaks, A. and Klibanov, A.M. (1988) J. Biol.
Duncia, J.V., Gardner, D.S., Ko, S., Chem., 263, 8017–8021.
References j359
260 Kamal, A., Ameruddin-Azhar, M., 278 de Gonzalo, G., Brieva, R., Sanchez, V.M.,
Krishnaji, T., Shaheer-Malik, M., and Bayod, M., and Gotor, V. (2001) J. Org.
Azeeza, S. (2008) Coord. Chem. Rev., 252, Chem., 66, 8947–8953.
569–592. 279 de Gonzalo, G., Brieva, R., Sanchez, V.M.,
261 Ghanem, A. and Aboul-Enein, H.Y. Bayod, M., and Gotor, V. (2003) J. Org.
(2004) Tetrahedron: Asymmetry, 15, Chem., 68, 3333–3336.
3331–3351. 280 Patel, R.N. (2008) Coord. Chem. Rev., 252,
262 Gentner, C., Schmid, R.D., and Pleiss, J. 659–701.
(2002) Colloids Surf., B, 26, 57–66. 281 Patel, R.N. (2004) Food Technol.
263 Ghanem, A. and Aboul-Enein, H.Y. Biotechnol., 42, 305–325.
(2005) Chirality, 17, 1–15. 282 Hanson, R.L., Shi, Z., Brzozowski, D.B.,
264 Nordin, O., Nguyen, B.V., V€orde, C., Banerjee, A., Kissick, T.P., Singh, J.,
Hedenstr€ om, E., and H€ogberg, H.E. Pullockaran, A.J., North, J.T., Fan, J.,
(2000) J. Chem. Soc., Perkin Trans. 1, Howell, J., Durand, S.C., Montana, M.A.,
367–376. Kronenthal, D.R., Mueller, R.H., and
265 Sakai, T. (2004) Tetrahedron: Asymmetry, Patel, R.N. (2000) Bioorg. Med. Chem., 8,
15, 2749–2756. 2681–2687.
266 Sakai, T., Kawabata, I., Kishimoto, T., 283 Tamarez, M., Morgan, B., Wong, G.S.K.,
Ema, T., and Utaka, M. (1997) J. Org. Tong, W., Bennett, F., Lovey, R.,
Chem., 62, 4906–4907. McCormick, J.L., and Zaks, A. (2003) Org.
267 V€anttinen, E. and Kanerva, L.T. (1997) Process Res. Dev., 7, 951–953.
Tetrahedron: Asymmetry, 8, 923–933. 284 Pogorevc, M. and Faber, K. (2000) J. Mol.
268 Sakai, T., Liu, Y., Ohta, H., Korenaga, T., Catal. B: Enzym., 10, 357–376.
and Ema, T. (2005) J. Org. Chem., 70, 285 Berkowitz, D.B., Pumphrey, J.A., and
1369–1375. Shen, Q. (1994) Tetrahedron Lett., 35,
269 Rosen, T.C. and Haufe, G. (2002) 8743–8746.
Tetrahedron: Asymmetry, 13, 1397–1405. 286 Ferraboschi, P., Brembilla, D., Grisenti,
270 Kawanami, Y., Iizuna, N., Maekawa, K., P., and Santaniello, E. (1991) J. Org.
Maekawa, K., Takahashi, N., and Kawada, Chem., 56, 5478–5480.
T. (2001) Tetrahedron, 57, 3349–3353. 287 Ferraboschi, P., Grisenti, P., Manzocchi,
271 Kawanami, Y., Honnma, A., Ohta, K., and A., and Santaniello, E. (1994) Tetrahedron:
Matsumoto, N. (2005) Tetrahedron, 61, Asymmetry, 5, 691–698.
693–697. 288 Im, D.S., Cheong, C.S., Lee, S.H.,
272 Kawasaki, M., Hayashi, Y., Kakuda, H., Youn, B.H., and Kim, S.C. (2000)
Toyooka, N., Tanaka, A., Goto, M., Tetrahedron, 56, 1309–1314.
Kawabata, S., and Kometani, T. (2005) 289 Ma, S. (2005) Chem. Rev., 105,
273 Kawasaki, M., Goto, M., Kawabata, S., and 290 Deska, J. and B€ackvall, J.-E. (2009) Org.
Kometani, T. (2001) Tetrahedron: Biomol. Chem., 7, 3379–3381.
Asymmetry, 12, 585–596. 291 Stewart, M., Capon, R.J., Lacey, E.,
274 Chen, C.-S. and Sih, C.J. (1989) Angew. Tennant, S., and Gill, J.H. (2005) J. Nat.
Chem. Int. Ed., 28, 695–707. Prod., 68, 581–584.
275 Patel, R.N. (2006) Curr. Org. Chem., 10, 292 Brenna, E., Fuganti, C., Fuganti, D.,
1289–1321. Grasselli, P., Malpezzi, L., and Pedrocchi-
276 Goswami, A., Howell, J.M., Hua, E.Y., Fantoni, G. (1997) Tetrahedron, 53,
Mirfakhrae, K.D., Soumeillant, M.C., 17769–17780.
Swaminathan, S., Qian, X., Quiroz, F.A., 293 Solares, L.F., Brieva, R., Quiros, M.,
Vu, T.C., Wang, X., Zheng, B., Llorente, I., Bayod, M., and Gotor, V.
Kronenthal, D.R., and Patel, R.N. (2001) (2004) Tetrahedron: Asymmetry, 15,
Org. Process Res. Dev., 5, 415–420. 341–345.
277 Gotor-Fernandez, V., Brieva, R., and 294 Boon, P.J.M., Boon, D.V.D., and Mulder,
Gotor, V. (2006) J. Mol. Catal. B: Enzym., G.J. (2000) Toxicol. Appl. Pharmacol., 167,
40, 111–120. 55–62.
360
295 Skouridou, V., Stamatis, H., and 313 Andrade, L.H. and Barcellos, T. (2009)
Kolisis, F.N. (2003) Eur. J. Lipid Sci. Org. Lett., 11, 3052–3055.
Technol., 103, 115–120. 314 Hungerhoff, B., Sonnenschein, H., and
296 Skouridou, V., Chrysina, E.D., Theil, F. (2002) J. Org. Chem., 67,
Stamatis, H., Oikonomakos, N.G., and 1781–1785.
Kolisis, F.N. (2004) J. Mol. Catal. B: 315 Ghanem, A. and Schurig, V. (2003)
Enzym., 29, 9–12. Monatsh. Chem., 134, 1151–1157.
297 Pchelka, B.K., Loupy, A., Plenkiewiez, J., 316 Matsuda, T., Harada, T., Nakamura, K.,
Petit, A., and Blanco, L. (2001) and Ikariya, T. (2005) Tetrahedron:
Tetrahedron: Asymmetry, 12, 2109–2119. Asymmetry, 16, 909–915.
298 Mangas-Sanchez, J., Rodriguez-Mata, M., 317 Jurcek, O., Wimmerova, M., and
Busto, E., Gotor-Fernandez, V., and Wimmer, Z. (2008) Coord. Chem. Rev.,
Gotor, V. (2009) J. Org. Chem., 74, 252, 767–781.
5304–5310. 318 Matsuda, T., Watanabe, K., Harada, T.,
299 Chênevert, R., Pelchat, N., and Morin, P. and Nakamura, K. (2004) Catal. Today, 96,
(2009) Tetrahedron: Asymmetry, 20, 103–111.
1191–1196. 319 Hazelard, D., Fadel, A., and Morgant, G.
300 Wang, Y.-F., Chen, S.-T., Liu, K.K.C., and (2004) Tetrahedron: Asymmetry, 15,
Wong, C.-H. (1989) Tetrahedron Lett., 30, 1711–1718.
1917–1920. 320 Sharfuddin, M., Narumi, A., Iwai, Y.,
301 Bevinakatti, H.S. and Banerji, A.A. (1991) Miyazawa, K., Yamada, S., Kakuchi, T.,
J. Org. Chem., 56, 5372–5375. and Kaga, H. (2003) Tetrahedron:
302 Patel, R.N. (2001) Curr. Opin. Biotechnol., Asymmetry, 14, 1581–1585.
12, 587–604. 321 Xie, X. and Tang, Y. (2007) Appl. Environ.
303 Pollard, D. and Kosjek, B. (2008) in Microbiol., 73, 2054–2060.
Organic Synthesis with Enzymes in Non 322 Turcu, M.C., Rantapaju, M., and Kanerva,
Aqueous Media (eds G. Carrea and S. L.T. (2009) Eur. J. Org. Chem., 5594–5600.
Riva), Wiley-VCH Verlag GmbH, 323 Maywald, M. and Pfalz, A. (2009)
Weinheim, pp. 169–188. Synthesis, 21, 3654–3660.
304 Kamal, A. and Ramesh Khanna, G.B. 324 Henke, E., Bornscheuer, U.T., Schmid,
(2001) Tetrahedron: Asymmetry, 12, R.D., and Pleiss, J. (2003) ChemBioChem,
405–410. 4, 485–493.
305 Pamies, O. and B€ackvall, J.-E. (2001) Adv. 325 Henke, E., Pleiss, J., and Bornscheuer,
Synth. Catal., 343, 726–731. U.T. (2002) Angew. Chem. Int. Ed., 41,
306 Pamies, O. and B€ackvall, J.-E. (2001) J. 3211–3213.
Org. Chem., 66, 4022–4026. 326 Hari Krishna, S., Persson, M., and
307 Runmo, A.-B.L., Pamies, O., Faber, K., Bornscheuer, U.T. (2002) Tetrahedron:
and B€ackvall, J.-E. (2002) Tetrahedron Lett., Asymmetry, 13, 2693–2696.
43, 2983–2986. 327 €
Ozdemirhan, D., Sezer, S., and
308 Burgess, K. and Jennings, L.D. (1991) J. S€onmez, Y. (2008) Tetrahedron:
Am. Chem. Soc., 113, 6129–6139. Asymmetry, 19, 2717–2720.
309 Burgess, K., Cassidy, J., and 328 Theil, F. (1994) Catal. Today, 22,
Henderson, I. (1991) J. Org. Chem., 56, 517–536.
2050–2058. 329 Eliel, E.L., Wilen, S.H., and Mander, L.N.
310 Kamal, A., Sandbhor, M., Shaik, A.A., and (1994) in Stereochemistry of Organic
Sravanthi, V. (2003) Tetrahedron: Compounds, John Wiley & Sons, Inc.,
Asymmetry, 14, 2839–2844. New York, p. 1206.
311 Lindner, E., Ghanem, A., Warad, I., 330 Edin, M., Steinreiber, J., and B€ackvall, J.E.
Eichele, K., Mayer, H.A., and Schurig, V. (2004) Proc. Natl. Acad. Sci. USA, 101,
(2003) Tetrahedron: Asymmetry, 14, 5761–5766.
1045–1053. 331 Leijondahl, K., Boren, L., Braun, R., and
312 Akai, S., Tanimoto, K., and Kita, Y. (2004) B€ackvall, J.E. (2009) J. Org. Chem., 74,
Angew. Chem. Int. Ed., 43, 1407–1410. 1988–1993.
References j361
332 Boren, L., Leijondahl, K., and Media (eds G. Carrea and S. Riva),
B€ackvall, J.E. (2009) Tetrahedron Lett., 50, Wiley-VCH Verlag GmbH, Weinheim,
3237–3240. pp. 75–112.
333 Shimada, Y., Usuda, K., Okabe, H., 352 Molinari, F., Romano, D., Gandolfi, R.,
Suzuki, T., and Matsumoto, K. (2009) Gardossi, L., Hanefeld, U., Converti, A.,
Tetrahedron: Asymmetry, 20, 2802–2808. and Spizzo, P. (2009) in Modern
334 Cygler, M., Grochulski, P.l., Biocatalysis (eds W.-D. Fessner and T.
Kazlauskas, R.J., Schrag, J.D., Anthonsen), Wiley-VCH Verlag GmbH,
Bouthillier, F., Rubin, B., Serreqi, A.N., Weinheim, pp. 79–92.
and Gupta, A.K. (1994) J. Am. Chem. Soc., 353 Patel, R.N. (2002) Enz. Microb. Technol.,
116, 3180–3186. 31, 804–826.
335 Chen, S.-T. and Fang, J.-M. (1997) J. Org. 354 Chattopadhyay, S. and Mamdapur, V.R.
Chem., 62, 4349–4357. (1993) Biotechnol. Lett., 15, 245–250.
336 Theil, F. (2001) Methods Biotechnol., 15, 355 Ozegowski, R., Kunath, A., and Schick,
277. H. (1995) Liebigs Ann., 1699–1702.
337 Baschang, G. and Inderbitzin, W. (1992) 356 Yang, H., Henke, E., and
Tetrahedron: Asymmetry, 3, 193–196. Bornscheuer, U.T. (1999) J. Org. Chem.,
338 Henly, R., Elie, C.J.J., Buser, H.P., 64, 1709–1712.
Ramos, G., and Moser, H.E. (1993) 357 Gedey, S., Liljeblad, A., Lazar, L., F€
ul€op,
Tetrahedron Lett., 34, 2923–2926. F., and Kanerva, L.T. (2002) Can. J. Chem.,
339 Seemayer, R. and Schneider, M.P. (1991) 80, 565–570.
J. Chem. Soc., Chem. Commun., 49–50. 358 Santaniello, E., Ciuffreda, P., Casati, S.,
340 Djadchenko, M.A., Pivnitsky, K.K., Alessandrini, L., and Repetto, A. (2006) J.
Theil, F., and Schick, H. (1989) J. Chem. Mol. Catal. B: Enzym., 40, 81–85.
Soc., Perkin Trans. 1, 2001–2002. 359 Wang, B., Liu, J., Tang, X., Chen, C., Gu,
341 Naemura, K., Ida, H., and Fukuda, R. J., Dai, L., and Yu, H. (2010) Tetrahedron
(1993) Bull. Chem. Soc. Jpn., 66, 573–577. Lett., 51, 309–312.
342 Rocco, V.P., Danishefsky, S.J., and 360 Fryszkowska, A., Komar, M.,
Schulte, G.K. (1991) Tetrahedron Lett., 32, Koszelewski, D., and Ostaszewski, R.
6671–6674. (2006) Tetrahedron: Asymmetry, 17,
343 Juarez-Hernandez, M., Johnson, D.V., 961–966.
Holland, H.L., McNulty, J., and 361 Takabe, K., Mase, N., Hashimoto, H.,
Capretta, A. (2003) Tetrahedron: Tsuchiya, A., Ohbayashi, T., and Yoda, H.
Asymmetry, 14, 289–291. (2003) Bioorg. Med. Chem. Lett., 13,
344 Sanfilippo, C., Nicolosi, G., Delogu, G., 1967–1969.
Fabbri, D., and Dettori, M.A. (2003) 362 Carr, J.A. and Bisht, K.S. (2004) Org. Lett.,
Tetrahedron: Asymmetry, 14, 3267–3270. 6, 3297–3300.
345 Krishna, S.H. and Karanth, N.G. (2002) 363 Neri, C. and Williams, J.M.J. (2003) Adv.
Catal. Rev., 44, 499–591. Synth. Catal., 345, 835–848.
346 Gou, D.-M., Liu, Y.-C., and Chen, C.-S. 364 Kirihara, M., Kawasaki, M., Takuwa, T.,
(1993) J. Org. Chem., 58, 1287–1289. Kakuda, H., Wakikawa, T., Takeuchi, Y.,
347 Wu, J.-Y. and Liu, S.-W. (2000) Enz. and Kirk, K.L. (2003) Tetrahedron:
Microb. Technol., 26, 124–130. Asymmetry, 14, 1753–1761.
348 Park, H.J., Choi, W.J., Huh, E.C., Lee, 365 Fadel, A. and Arzel, P. (1997) Tetrahedron:
E.Y., and Choi, C.Y. (1999) J. Biosci. Asymmetry, 8, 283–291.
Bioeng., 87, 545–547. 366 Chênevert, R., Courchesne, G., and
349 Morrone, R., Piattelli, M., and Nicolosi, Pelchat, N. (2006) Bioorg. Med. Chem., 14,
G. (2001) Eur. J. Org. Chem., 1441–1443. 5389–5396.
350 Sanchez, A., Valero, F., Lafuente, J., and 367 Chênevert, R., Simard, M., Bergeron, J.,
Sola, C. (2000) Enz. Microb. Technol., 27, and Dasser, M. (2004) Tetrahedron:
157–166. Asymmetry, 15, 1889–1892.
351 H€ ogberg, H.-E. (2008) in Organic 368 Chênevert, R. and Caron, D. (2002)
Synthesis with Enzymes in Non Aqueous Tetrahedron: Asymmetry, 13, 339–342.
362
369 Akai, S., Naka, T., Fujita, T., Takebe, Y., 373 Liese, A., Seelbach, K., and Wandrey, C.
Tsujino, T., and Kita, Y. (2002) J. Org. (2006) Industrial Biotransformations, 2nd
Chem., 67, 411–419. edn, Wiley-VCH Verlag GmbH,
370 K€ohler, J. and W€
unsch, B. (2006) Weinheim, pp. 313–315.
Tetrahedron: Asymmetry, 17, 3091–3099. 374 Hills, G.A., MsCrae, A.R., and Poulina,
371 Candy, M., Audran, G., Bienaym e, H., R.R. (1990) (Unichema Chemie BV), Eur.
Bressy, C., and Pons, J.M. (2009) Org. Pat. EP0383405.
Lett., 11, 4950–4953. 375 Kemp, R.A. and Macrae, A.R. (1992)
372 Patel, R.N., Banerjee, A., Pendri, Y.R., (Unichema Chemie BV), Eur. Pat. Appl.
Liang, J., Chen, C.-P., and Mueller, R. EP0506159.
(2006) Tetrahedron: Asymmetry, 17, 376 Hilterhaus, L., Thum, O., and Liese, A.
175–178. (2008) Org. Process Res. Dev., 12, 618–625.
j363
9
Hydrolysis and Formation of Epoxides
Jeffrey H. Lutje Spelberg and Erik J. de Vries
9.1
Introduction
Because of their reactivity, epoxides are considered valuable multifunctional inter-

mediates in synthetic strategies [1–3]. This chapter discusses the enzyme-catalyzed
formation of chiral epoxides, hydrolysis, and the formation of b-substituted alcohols
from epoxides.
An epoxide has at least one chiral center and ring opening can occur at either of the
two carbon atoms. This regioselectivity is determined by a combination of chemical
reactivity based on electronic and steric factors and enzymatic selectivity. A mono-
substituted epoxide can react at the stereocenter with inversion of configuration, or at
the adjacent achiral center with retention of configuration. Depending on the
enzymes enantioselectivity either the ring-opening product or the remaining
epoxide will be the final product of the process. Should the enzyme have selectivity
for the undesired enantiomer, a chemical follow-up step can convert the remaining
epoxide into the product, or, for example when employing an epoxide hydrolase, the
diol back to the epoxide. In this chapter we demonstrate the power of state-of-the-art
enzymatic technology and the incredible flexibility for process design it offers in
sometimes complicated but elegant synthetic strategies involving epoxides.
Because of the reactivity of epoxides, this functional group does not often remain in
the final target but rather serves as a conduit to more stable final products. For
instance, very few drugs are epoxides but the Becker Conceptus Pharmaceutical
Intermediates Database lists many epoxides as pharmaceutical intermediates. Rather
than focusing strictly on enzymes that can hydrolyze epoxides, as was done in the
previous edition of this book, we have expanded the scope to include enzymes that
can serve to make precursors for epoxides, as well as a range of different derivatives.
Most epoxides suffer from some degree of chemical hydrolysis under the reaction
conditions typically used for enzymatic reactions. This is a major drawback and has
proven to be one of the hurdles for industrial-scale application – we are not aware of
any multi-ton scale epoxide hydrolase processes. To circumvent this instability
problem one can omit isolating the epoxide and have it react further to a more
and Oliver May.
j 9 Hydrolysis and Formation of Epoxides
364
stable compound. Thus, this chapter will also describe recent developments towards
process intensification by combining multiple enzymatic steps shown in Scheme 9.1,
either simultaneously or in tandem.
OH
OAc
NO2
Cl
OH OH
N3 NCO
Hydrolase O O
HHDH
Epox EpCarb OH
OH OH
HHDH O
EH OH
Cl
GST
other HHDH OH
KRED O
R enzyme
SG
OH OH
O
OCHO NCS
Cl
OH
CN
Scheme 9.1 Overview of biocatalytic reactions dehydrogenases (known as ADH or KRED),

involving epoxides, precursors, and products. epoxidizing enzymes (Epox), epoxide
Reactions (in bold) catalyzed by halohydrin carboxylase (EpCarb), glutathione-S-transferase
dehalogenases (HHDHs) and epoxide (GST), and diverse other enzymes reacting
hydrolases (EHs) are the topic of this chapter. on an adjacent functional group (other
Other abbreviated enzymes are alcohol enzyme).
Epoxide hydrolases of various origins have been studied by many research groups
for many years and the field is still very much alive. For instance, a Scopus literature
search over the period 1997–2009 showed that more than 150 individual researchers
each published more than five research papers on epoxide hydrolases. Almost 500
book titles cover an aspect of epoxide hydrolases and the primary research has been
reported in close to 6000 papers! These numbers indicate the huge interest in
academia and industry for converting an epoxide into a diol, which, from some
viewpoints, can be considered a rather narrow reaction scope. Furthermore, in the
past ten years, only 14% of these epoxide hydrolase papers dealt with chirality, whilst
this is the main focus of the synthetic organic or medicinal chemist.
It is desirable to have well-understood multifunctional tools that can be used for
different purposes by simply changing reaction conditions or reactants. Owing to
their promiscuity, halohydrin dehalogenases can be considered the Swiss army
knife of enzymes. The full potential of this enzyme class has only been started to be
evaluated in the past twelve years, resulting in 56 research papers by a total of 37
authors. A strong focus on chirality resulted in half of these reports discussing chiral
products. We demonstrate in this chapter that this promising enzyme class, which
has not been reviewed before, deserves more attention.
9.1.1
Biocatalytic Strategies Towards Optically Pure Epoxides and Derivatives
Scheme 9.1 shows an overview of enzymatic reactions involving epoxides or

derivatives and precursors of epoxides. The key reactions shown in bold, which are
catalyzed by halohydrin dehalogenases and epoxide hydrolases, are within the scope
of this chapter. Non-highlighted reactions are either discussed elsewhere in this book
(hydrolytic reactions in Chapters 8–12, reductions in Chapters 2–39, and oxidations
in Chapters 30–38) or, like epoxide conjugation, have only a limited synthetic
applicability [110]. Since the purpose of this chapter is to describe enzymatic
reactions that can be performed on at least a practical gram scale by the organic
chemist, such reactions will be discussed only briefly.
9.1.1.1 Epoxide Conjugation

Mammalian metabolizing enzymes like glutathione-S-transferases (GSTs) can
perform an epoxide ring opening through conjugation with a cofactor
(Scheme 9.2) [4]. Some reports describe enantioselectivity and optical enrichment
of the remaining epoxide, which was reviewed in 2003 [5]. GST enzymes have a
broad substrate range and require a glutathione cofactor that is incorporated in the
product conjugate and is therefore not recyclable. Only a few reports on protein
engineering and improvement of catalytic performance for epoxide substrates have
appeared. Mannervik and coworkers reported site-directed mutagenesis of human
GST M1-1 in combination with an iodo-alkane chemical modification in the active
site that resulted in a tenfold increase in enantioselectivity for epoxides like styrene
oxide, trans-b-methyl styrene, and trans-stilbene oxide [6]. It is possible to change
the catalytic properties of these enzymes by traditional mutagenesis and by
replacement with unnatural amino acids [7, 8]. The reported E-values of around
30 are very promising and are in a similar range as E-values of epoxide hydrolases.
However, the current state-of-the-art in this field is that enantioselective GSTs do
not provide a viable alternative to more established cofactor-independent enzymes
unless the cofactor GSH can be replaced by cysteine such as in fosfomycin
conjugating GSTs or with even simpler thiols.
A different class of enzymes conjugating epoxides is the epoxyalkane Co-
enzyme M transferases, which use the unusual cofactor mercaptoethane sulfonic
acid for epoxide ring opening [9]. The mechanism of this enzyme has been studied
Epixide conjugation
O OH
O
+ Conj
R R R
Scheme 9.2 Conjugation of the epoxide by metabolizing enzymes can be used to isolate the
optically enriched remaining epoxide (stereochemistry chosen arbitrarily).
366
and there is some promiscuity with respect to the cofactor [10]. Thiols like
mercapto-ethanol, 3-mercaptopropionate, and cysteine can be used, albeit at a
much decreased reaction rate. There is no physiological requirement for the
enzyme being enantioselective and indeed both enantiomers of the natural
substrate epoxypropane are converted without preference. Surprisingly, there is
some degree of enantioselectivity for small aliphatic substrates like 2,3-
epoxybutane [11].
9.1.1.2 Oxidation of Alkenes

There are several oxidative enzyme classes known to epoxidize alkenes with a certain
degree of enantioselectivity (Scheme 9.3). These enzymes will be discussed in
Chapters 30–38 and only one illustrative reference for each enzyme class will be
given here.
Oxidation of alkenes
Oxidation O
R R
Scheme 9.3 Oxidation of a pro-chiral alkene to yield an epoxide (stereochemistry chosen

arbitrarily).
Cytochrome P450s catalyze an unparalleled breadth of oxidation reactions using

either molecular oxygen or hydrogen peroxide [12]. Other enzymes able to use
hydrogen peroxide are the haloperoxidases, such as bromo peroxidase and chloro
peroxidase and horseradish peroxidase [13, 14]. Monooxygenases like alkene mono-
oxygenase [15], styrene monooxygenase [16], methane monooxygenase [17], and
Baeyer–Villiger monooxygenases [18] can epoxidize certain alkenes, too.
9.1.1.3 Alcohol Dehydrogenases

Prochiral haloketone reduction using alcohol dehydrogenase (ADH) enzymes is an
economical method for manufacture of optically pure halohydrins (Scheme 9.4).
Once the halohydrin is obtained, a simple alkaline treatment or a halohydrin
dehalogenase (HHDH)-catalyzed ring closure step under mild conditions often
gives access to the desired enantiomer of the epoxide in close to 100% overall yield.
ADHs are easy to use under conditions that are favorable for organic synthesis. High
substrate loadings of up to 50% have been used. Engineered ADHs are often tolerant
to high fractions of many different cosolvents, with the use of isopropanol for cofactor
recycling employed in quantities of 50% up to 90%. ADHs have a broad pH range and
Alcohol dehydrogenases
O OH
Hal Reduction Hal
R R
Scheme 9.4 Reduction of pro-chiral haloketones offers a high-yielding entry into epoxide
precursors (stereochemistry chosen arbitrarily).
can be optimized for use under acidic pH (pH 5) to very basic (pH 11) conditions,
and at elevated temperatures. One can choose from many different cofactor recycling
systems and the cost contribution of cofactor is no longer an issue. Since redox
processes are essential for life, sources of ADHs are plentiful and it seems one only
needs to find the one with the desired selectivity. A state-of-the-art paper was
published in 2009 by Zhu et al. describing an ADH from the hyperthermophile
Pyrococcus furiosus. The paper shows that this enzyme operates at up to 70 C, with
different co-solvents and cofactor recycles, and lists a few examples on a preparative
scale [19]. For an overview of ADH catalyzed reductions, we refer to Chapter 26 by
Gr€oger et al.
9.1.1.4 Hydrolases and Other Enzymes Acting on an Ancillary Functional Group

When an ester, acid, alcohol, or amine functionality is present in the epoxide, a
hydrolase or acylase can be used to resolve the molecule without touching the epoxide
(Scheme 9.5). As an example, recent papers describe a lipase resolution of glycidyl
esters and optimization of a Bacillus subtilis lipase strain L2 to resolve glycidyl butyrate
with an E-value of 108 [20, 21]. Acylation of glycidol with vinyl butyrate catalyzed by
Novozyme 435 (Candida antarctica lipase B) to give (S)-glycidyl butyrate occurs with
good selectivity (E ¼ 69) [22].
The hydrolysis of alkyl 3,4-epoxybutyrates with mostly (S)-selective esterases or
(R)-selective proteases affords isolation of the remaining epoxide-ester in good
optical purities [23]. Aminolysis of ethyl-(3,4)-epoxybutyrate to the N-benzyl-epoxy-
amide with Candida antarctica lipase has been reported to occur with (S)-
selectivity [24].
Using hydrolytic enzymes such as lipases and esterases to resolve halohydrins is
well established and we refer to Chapter 8 by Gr€oger and Hanefeld for a review of the
field. Since secondary vicinal halohydrins are accessible in 100% using an ADH
enzyme, to achieve competitive high yields of halohydrin the group of B€ackvall has
developed dynamic kinetic resolution (DKR) processes of a whole range of aromatic
alcohols using ruthenium-based racemization catalysts in combination with, for
instance, Pseudomonas cepacia lipase [25]. The results for the substituted styrene
oxide precursors approach those of a perfect DKR: 98–99% conversion with >99%
e.e. for the product (Scheme 9.6).
O O
O O O
+ OH
O O
lipase/esterase
O O O
O O O
O O + OH/NHR
Scheme 9.5 Ester resolution using the nearby chiral epoxide functionality only as a selector.
368
OH 5 mol% Ru catalyst OAc

Cl 10 mol% tBuOK Cl
PS-C II (Amano)
R isopropenyl-OAc R
Na2CO3, toluene, rt
Scheme 9.6 Chemoenzymatic DKR for obtaining optically pure (protected) halohydrins.
As described above, alcohol dehydrogenases are very powerful in terms of

producing optically pure halohydrins, but come with the drawback that they do
not provide access to tertiary alcohols. Hydrolytic or acylating enzymes that
convert tertiary alcohols represent an important element in synthetic strategies
towards epoxides with multiple substituents. However, hydrolysis of protected
tertiary alcohols is less common and there are only a few biocatalysts known to
catalyze this reaction. The wild-type esterase from Bacillus subtilis BS2 was not so
enantioselective but could be improved by mutagenesis and medium engineer-
ing, and Candida antarctica lipase A was very selective but suffered from low
activity [26, 27].
The chloroperoxidase from Caldariomyces fumago has been reported to selectively
oxidize one enantiomer of (substituted) glycidol to the aldehyde, leaving the optically
enriched epoxide behind (Scheme 9.7) [28].
O Oxidation O O
OH OH O
+
R R=H, cis-nPr R R
"(R)"-epoxide "(S)"-aldehyde
Scheme 9.7 Chloroperoxidase oxidation of an ancillary alcohol group to effect kinetic resolution of
the epoxide.
9.1.2
Scope and Outline of this Chapter
This chapter covers the formation and conversion of epoxides using two classes of
enzymes: halohydrin dehalogenases and epoxide hydrolases (highlighted reactions
in Scheme 9.1). Epoxide hydrolases catalyze the addition of a water molecule to an
epoxide, to yield a vicinal diol. Halohydrin dehalogenases catalyze the reversible ring
closure of halohydrins and the irreversible ring opening with nucleophiles such as
azide, cyanide, and nitrite.
Table 9.1 gives an overview of the characteristics of both enzyme classes. Many of
these enzymes have been overexpressed in Escherichia coli, making them abundantly
available for reactions on preparative scale. Should the enzyme characteristics not be
good enough for commercial purposes, often the crystal structure and catalytic
mechanism have been elucidated, making optimization by mutagenesis easier.
9.2 Conversion and Formation of Epoxides Catalyzed by Halohydrin Dehalogenases j369
Table 9.1 Comparison of epoxide hydrolases and halohydrin dehalogenases.
Epoxide hydrolases Halohydrin dehalogenases
Provides synthetic Epoxides, diols Epoxides and chloro- and

access to bromo-alcohols.
Azido-, cyano-, and
nitro-alcohols.
Diols, thiiranes and
oxazolidinones
Nucleophile Water Various anionic nucleophiles
(ring-opening reaction)
Requirements Cofactor independent Cofactor independent
Activity Up to 50 U mg1 Up to 40 U mg1 for ring-closure
reaction.
From < 5 (cyanate, cyanide) to
500 (azide) U mg1 for ring-
opening reaction
Enantioselectivity Mainly (R)-selective. (R)-Selective enzyme, moderate
to high E-values.
Many examples with high (S)-Selective enzymes, low
E-values E-values
Regioselectivity Mainly terminal, 90–98%. Strictly terminal, >95%.
Few non-terminal selectivities In many cases >99%
Occurrence Mammals, plants, insects, Bacteria only
fungi, yeast, bacteria
Number variants Several hundred enzymes < 10 enzymes in three closely
related families.
Industrial application — C3 synthons by Daiso
(Japan) [58].
Atorvastatin intermediate by
Codexis (USA) [88]
9.2
Conversion and Formation of Epoxides Catalyzed by Halohydrin Dehalogenases
9.2.1
Classification, Structure, and Mechanism of Halohydrin Dehalogenases
Halohydrin dehalogenases catalyze the interconversion of halohydrins and epoxides.

Various alternative names for the enzyme have been used in the literature such as
haloalcohol dehalogenase, halohydrin epoxidase, and halohydrin hydrogen halide
lyase (E.C. 4.5.1).
Although halohydrin dehalogenase catalyzed conversions have been studied for at
least 40 years, the number and variety of the enzymes is limited. Only halohydrin
dehalogenases from bacterial origin have been identified. Based on their specific
activities towards a range of substrates they have been divided into three distinct
370
groups. Group A halohydrin dehalogenases are characterized by the very high activity
towards 1,3-dibromo-2-propanol (>20-fold higher compared to the chlorinated
variant). Members of this group are HheA (halohydrin dehalogenase from Arthro-
bacter sp. AD2), H-lyase A (halohydrin dehalogenase from Corynebacterium sp. N-
1074), and DehC (halohydrin dehalogenase from Arthrobacter erithii H10a). Mem-
bers of group B are H-lyase B (halohydrin dehalogenase from Corynebacterium sp. N-
1074), HheB (halohydrin dehalogenase from Mycobacterium sp. GP1), and DehA
(halohydrin dehalogenase from Arthrobacter erithii H10a). Besides the lower activity
towards 1,3,-dibromo-2-propanol, the substrate range of this group of enzymes is
similar to that of the members of group A. The enzyme in group C, HheC (halohydrin
dehalogenase from Agrobacterium radiobacter AD1) is distinctly different compared
to groups A and B in terms of the high enantioselectivity towards aromatic substrates.
Because there is not a big difference in substrate range between groups A and B, there
was a need for an alternative classification system.
Analysis of the known DNA sequences confirmed that they can be divided into
three different phylogenetic groups [60]. Members of each group: Arthrobacter sp
AD2 (HheA), Mycobacterium sp. GP1 (HheB), and Agrobacterium radiobacter AD1
(HheC) were cloned and expressed in E. coli. The sequence identity between the
groups is only as high as 25.5%. A sequence homology and structure prediction
investigation showed that halohydrin dehalogenases are structurally similar to short-
chain dehydrogenases/reductases (SDR proteins). The members from group A
(HheA and H-lyase A) show 97.1% sequence similarity and the members from
group B (H-lyase B and HheB) show 98.2% sequence similarity. The halohydrin
dehalogenase Agrobacterium radiobacter AD1 (HheC) is the only member of group C.
A recently cloned halohydrin dehalogenase from Agrobacterium sp. NHG3 was
shown to be identical to HheC [29, 30]. Another enzyme obtained from Agrobacterium
tumefaciens HK7 (HalB) was 91% identical. The substrate range and enantioselec-
tivity of the latter two enzymes has not been studied in detail.
The halohydrin dehalogenase from Agrobacterium radiobacter AD1 (HheC) is the
best studied enzyme and distinguishes itself from HheA and HheB because of the
high enantioselectivity towards a range of substrates, most notably aromatic sub-
strates. The enzymes catalytic mechanism, employing a triad of Ser132/Tyr145/
Arg149 that is conserved in all halohydrin dehalogenases, was proposed based on
sequence alignments with structurally related short-chain dehydrogenases
(Scheme 9.8) [60]. Mutation of either one of these positions yielded an inactive
enzyme. The hydroxyl group of the bromoalcohol is deprotonated by the Tyr145
(of which the pKa is lowered by Arg149) thereby forming an oxyanion that attacks the
carbon atom bearing the halogen, resulting in formation of the epoxide and
hydrobromic acid.
Protein crystallography on HheC has been carried out by de Jong et al. [31]. The
study confirmed the amino acids involved in catalysis, identified a halide binding site
that was predicted by kinetic studies, and showed that HheC is a tetrameric protein
consisting of two tightly bound dimers. Each monomer is bound to its opposite in the
dimer via its C-terminal end and, surprisingly, contributes a tryptophan (Trp249)
residue to the active site of the opposite monomer.
Arg149 Arg149
H 2N N H Tyr145 H2 N N H Tyr145
H H
O O
H
H
O
H O H
Br O Br O
Ser132 Ser132
Halide
R R
bindingsite
Scheme 9.8 Reaction mechanism of the reversible ring closure of a bromoalcohol catalyzed by the
halohydrin dehalogenase from Agrobacterium radiobacter AD1 (HheC). Hydrogen bond interactions
between substrate and Tyr145 and Ser132 are critical to enzyme activity.
9.2.2
Discovery of Halohydrin Dehalogenases
The first activity of a halohydrin dehalogenase was reported as early as 1968 [32].
Castro and Bartnicki studied the growth of a Flavobacterium sp. on various halide-
containing compounds. Cell-free extracts of this organism converted 2,3-dibromo-
propanol in four steps into glycerol (Scheme 9.9). In step 1, the substrate 2,3-
dibromo-1-propanol is converted into epibromohydrin, which is hydrolyzed (likely by
an epoxide hydrolase) in step 2 to the corresponding 1-bromo-2,3-propanediol. Ring
closure in step 3 yields glycidol, which is subsequently hydrolyzed to the end product
glycerol. Both ring-closure steps were shown to be catalyzed by a halohydrin
dehalogenase. With partially purified halohydrin dehalogenase the degradation route
stopped at the epoxide stage, proving that the enzyme does not catalyze the hydrolysis
of epoxides. Investigation of the substrate spectrum of this Flavobacterium showed
that the best substrate is in fact 1,3-dibromo-2-propanol, which is converted approx-
imately ten times faster than 2,3-dibromo-1-propanol and 1-bromo-2,3-propanediol.
The reversibility of the reaction, the ring opening of the epihalohydrins with
chloride and bromide ion, was studied in more detail [33]. This demonstrated that the
ring opening of the epoxides occurred exclusively at the terminal carbon atom,
yielding the corresponding 1,3-dihalo-2-propanols instead of 2,3-dihalo-1-propanol.
Br
1 O
Br OH Br
2
Halohydrin
OH
dehalogenase
Br OH
3
OH 4
O
HO OH OH Halohydrin
dehalogenase
Scheme 9.9 Biodegradation of 2,3-dibromopropanol to glycerol by Flavobacterium sp.

372
Some 15 years later, in 1983 the halohydrin dehalogenase from Flavobacterium was
further investigated by the Cetus Corporation1) as part of a multi-enzyme conversion
of propylene into propylene oxide (Section 9.2.8) [34]. Conversions with whole cell
preparations showed that besides 1-bromo-2-propanol (taken as 100% activity) and 1-
chloro-2-propanol (26%) 1-iodo-2-propanol was also accepted as a substrate. A regio-
isomeric substrate with the bromine on the 2-position such as 2-bromo-1-propanol
(5%) was converted at a much lower rate. Since fluoride is a poor leaving group, no
conversion was observed with fluoroalcohols. Only vicinal halohydrins are substrates
for all halohydrin dehalogenases and oxetanes have never been found from 1,3
bromoalcohols [35].
Studies towards the biodegradation of xenobiotic halogenated compounds have
been a great source of industrially interesting enzymes such as halohydrin dehalo-
genases, haloalkane dehalogenases, haloacid dehalogenases, and epoxide hydro-
lases [36]. Before problems with toxicity and persistence came to light, halogenated
compounds found frequent application in industry as solvents, agrochemicals
(nematicides, fumigants), flame retardants, and chemical intermediates. Especially,
C3-chloroalcohols and their corresponding epoxides glycidol and epihalohydrin have
a high toxicity [37]. Epichlorohydrin is a versatile building block with various
applications in, for example, polymers, resins, and fine chemicals. It is prepared
via chloropropanols as intermediates and the global production approached 1000000
metric tons per year in 2006. A common food contaminant in, for example, starches,
salami, and soy sauce (100–800 mg kg1) is 3-chloro-1,2-propanol, which is formed
by the reaction of hydrochloric acid with lipids [38]. The enzymatic degradation of 3-
chloro-1,2-propanol was described using a preparation of Bakers yeast [39]. The
slightly enantioselective conversion of 3-chloro-1,2-propanediol is accompanied by
an equal increase in chloride ion, indicating an action catalyzed by a dehalogenating
enzyme such as a haloalkane dehalogenase or a halohydrin dehalogenase.
Neutral curing poly(aminoamide)-epichlorohydrin chemicals are applied for the
preparation of paper products to increase their wet-strength. They are prepared via a
reaction with epichlorohydrin, yielding various halohydrin side products that must
be removed for any application in consumer products. In research collaboration
between Carbury Herne, Ltd. (GB) and Hercules, Inc. (US) an enzymatic dehalo-
genation process was developed [40]. A consortium of two microorganisms was able
to degrade the halogenated compounds. It was shown that the biotransformations
were catalyzed by halohydrin dehalogenases. The process was implemented on a
3000 liter scale in an established production plant. A more detailed study showed that
one of the two strains contained two halohydrin dehalogenases degrading the
halohydrins while the other non-dehalogenating bacterium in the consortium used
the formed products glycidol and glycerol as sole carbon sources [41]. The two
dehalogenating enzymes were from the strain Arthrobacter erithii H10a (DehA and
DehC) and were characterized in more detail [42]. A later version of the process uses a
1) Cetus was founded in Berkeley, California, and in 1991 was acquired by Chiron (USA) which itself was
acquired in 2006 by Novartis (CH).
consortium of Arthrobacter histidinolovorans and Agrobacterium radiobacter and
achieves complete mineralization to CO2 [43].
9.2.3
Ring-Closure Reactions
9.2.3.1 Production of Chiral C3 Building Blocks Through Ring Closure

An enantioselective ring-closure catalyzed by a halohydrin dehalogenase can be used
to prepare optically active halohydrins and epoxides and is the basis of several
interesting synthetic strategies. For instance, starting from a racemic mixture of
chloroalcohols, a highly enantioselective kinetic resolution will yield the epoxide in
high optical purity (Scheme 9.10). If the reaction is stopped after the faster reacting
enantiomer is completely consumed, the remaining chloroalcohol enantiomer can
also be obtained in high optical purity. After isolation and purification, this enan-
tiomer can easily be converted into the other epoxide enantiomer under basic
conditions. In this way, both enantiomers of the epoxide are accessible and this
strategy is an advantage when the enzyme might have the wrong enantiopreference.
OH OH Enzymatic O
Optically pure epoxide
R
Cl + R
Cl R
- HCl
+
OH
Chemical O
Cl
R R
Isolated - HCl
remaining
enantiomer
Scheme 9.10 Access to both enantiomers of an epoxide via a two-step reaction, combining a
halohydrin dehalogenase catalyzed reaction with a chemical step.
Around 1985, several research groups started studying the microbial biodegra-
dation pathways of pollutants such as 1,3-dichloropropanol, 2,3-dichloropropanol, 3-
chloro-1,2-propanediol, and epichlorohydrin [44]. Van den Wijngaard reported the
degradation of epichlorohydrin and chloropropanols [45]. The organism Pseudomo-
nas sp. strain AD1 (now renamed to Agrobacterium radiobacter AD1) was able to
degrade epichlorohydrin through the action of an epoxide hydrolase (EchA) and a
subsequent halohydrin dehalogenase (HheC) step. In Arthrobacter sp. AD2, a slow
chemical hydrolysis step proceeds the enzyme-catalyzed ring closure of 3-chloro-1,2-
propanol. A similar degradation pathway was described by Nakamura [46]. The strain
Corynebacterium sp. N-1074 was able to convert 1,3-dichloro-2-propanol into glycerol
and contained two halohydrin dehalogenases (H-lyase A and H-lyase B). The
halohydrin dehalogenases were purified, brought to overexpression in E. coli, and
studied extensively [47–50].
374 j 9 Hydrolysis and Formation of Epoxides
The group of Kasai and Suzuki at Daiso (Japan) studied halohydrin dehalogenase
activity in various organisms [51, 52]. Their goal was to identify organisms that were
able to enantioselectively degrade the racemic 2,3-dichloro-1-propanol, allowing the
isolation of the remaining enantiomer in high optical purity. They identified two
organisms, Alcaligenes sp. DS-K-S38 and Pseudomonas sp. OS-K-29, showing
opposite enantiopreference, thus giving access to either enantiomer (Scheme 9.11).
Although one enantioselective halohydrin dehalogenase would give access to both
enantiomers of an epoxide as mentioned earlier (Scheme 9.10), such an approach is
not possible in the case of epichlorohydrin due to racemization under the process
conditions.
Cl base
Alcaligenes sp. DS-K-S38 O
Cl OH Cl
Cl (S)-epichlorohydrin
99.5% e.e.
Cl OH or
Cl
base O
Cl OH Cl
Pseudomonas sp. OS-K-29 (R)-epichlorohydrin
99.5% e.e.
Scheme 9.11 Access to either enantiomer of epichlorohydrin via resolution of racemic 2,3-
dichloropropanol by either an Alcaligenes or a Pseudomonas species.
By immobilizing the cells using calcium alginate, 19 repetitive batches in 50 days

were performed without noticeable reduction in activity [53]. Attempts to synthesize
optically pure glycidol from optically pure epichlorohydrin were not successful [54].
Therefore, a similar approach was followed and two organisms were identified that
selectively degraded 3-chloro-1,2-propanediol with opposite enantiopreference, giv-
ing access to either enantiomer of glycidol [55]. A more detailed study of the
degradation route in an (R)-3-chloro-1,2-propanediol assimilating Alcaligenes sp.
DS-S-7G showed that a novel enzyme was responsible for the dehalogenation
reaction. The product was hydroxyacetone instead of glycidol [56]. Since the scope
of this chapter is on epoxide forming and converting enzymes, these halohydrin
dehydro-dehalogenases will not be discussed here.
The above-described strategies towards chiral C3 synthons were applied at Daiso
on an industrial scale in a 33 000-liter reactor, giving the products in yields of several
hundreds of kilos to metric tons [57].
Nakamura et al. studied the production of optically pure epichlorohydrin starting
from the 1,3-dichloropropanol. The obvious advantage of this prochiral substrate is a
maximum yield of 100% instead of 50% in a kinetic resolution strategy. The
conversion of 1,3-dichloropropanol by a halohydrin dehalogenase from Corynebac-
terium sp. N-1074 (H-lyase B) yielded (R)-epichlorohydrin in >90% yield in the initial
stages of the reaction. However, due to enzyme-catalyzed racemization, the e.e.
reduced drastically in the course of the reaction.
When whole cells of Corynebacterium sp. N-1074 were employed the follow up
reaction step decreased the extent of racemization (see Section 9.2.7.2).
A similar racemization behavior was observed by Assis et al. when studying the
ring closure of 1,3-dichloro-2-propanol by the halohydrin dehalogenase from Arthro-
bacter erithii H10a (DehA). The decrease in optical purity of (R)-epichlorohydrin was
circumvented through the addition of 200 mM of KBr. The enzyme catalyzed the ring
opening of epichlorohydrin by bromide ion with a preference towards the (S)-
enantiomer, thereby enriching (R)-epichlorohydrin towards >95% e.e. but in only
11% yield [58]. The enzyme-catalyzed racemization was applied later in a DKR
strategy (Section 9.2.6).
9.2.3.2 Production of Aromatic Building Blocks Through Ring Closure

Almost all reports exclusively discuss the ring closure of small aliphatic halohydrins.
Wandel et al. investigated the ring closure with a broader range of substrates, using
resting cells from the Flavobacterium sp. (obtained from the Cetus Corporation) [59].
Substrates such as 1-chloro-2-octanol, 1-bromo-2-cyclohexanol, and halohydrins
carrying the halide on the secondary carbon atom were not converted by the
microorganism. Apart from a range of aliphatic 1-bromo-2-alkanols, it was shown
for the first time that halohydrins containing an aromatic moiety were also accepted.
The highest observed enantioselectivity (expressed as the E-value) was only low, with
a value of around E ¼ 3 for 1-bromo-2-phenylethanol.
The halohydrin dehalogenase obtained from Agrobacterium radiobacter AD1
(HheC) was cloned and expressed in E. coli [60]. The biocatalytic potential of this
enzyme was studied in detail using purified enzyme [61, 62]. Along with known
substrates such as 2,3-dichloro-1-propanol, halohydrins containing an aromatic
moiety were also converted with high enantioselectivity (Figure 9.1). The styrene
oxide precursor 1-chloro-2-phenylethanol was converted with an E-value of 73 and the
related regio-isomer 2-chloro-1-phenylethanol was also converted with good enan-
tioselectivity. Note that this finding is contrary to the Flavobacterium enzyme that was
not active when the halide was on the secondary carbon atom. The chemical
hydrolysis (42% per hour) of this compound limits a practical application. With the
exception of 2,3-dihalo-1-propanol no enzyme-catalyzed racemization of the formed
epoxide was observed, allowing the isolation of the aromatic epoxides in up to 90% e.e.
HO Cl
HO Cl Cl OH HO Br
Cl
Cl OH
O2N
Cl
Enantioselectivity: E >100 E = 73 E > 50 E = 23 E = 124
Specific activity: 0.9 U/mg 14.1 U/mg 8.7 U/mg 4.5 U/mg 35 U/mg
Figure 9.1 Enantioselectivity (E values) and specific activities of the ring-closure reactions
catalyzed by purified halohydrin dehalogenase from Agrobacterium radiobacter AD1 (HheC).
376
The reversibility of the reaction could be overcome by in situ removal of the epoxide
(see cascade reactions in Section 9.2.7).
The kinetic resolution of a range of 3-alkenyl and heteroaryl chlorohydrins was
investigated using the HheC enzyme [63]. Thiophene chlorohydrins were converted
with a specific activity of 47 U mg1 with very high enantioselectivity (E > 200). After
enzymatic formation, rapid hydrolysis of the formed epoxides occurred, hampering
in certain cases isolation of the epoxides. This instability was an advantage for the
enzymatic kinetic resolution by essentially removing the equilibrium. This in situ
product removal resulted in a complete kinetic resolution, allowing isolation of the
(S)-chloroalcohol in >99% e.e. The enzymatic resolution of 2-chloro-1-thiophen-2-yl-
ethanol (Scheme 9.12) was demonstrated on practical scale. Using a reaction solution
containing toluene (10 vol.%) as a second phase, 20.9 g (117 mM) of the halohydrin
was resolved using 9.3 mg l1 of purified HheC. The remaining (S)-halohydrin was
isolated in >99% e.e. and 47% yield, while a non-regioselective hydrolysis of the
formed epoxide yielded the diol almost racemic.
HO Cl HO Cl O chemical HO OH
Halohydrin
S dehalogenase S S hydrolysis S
+
E > 200 low regio-
>99% e.e. selectivity "almost racemic"
47% yield
Scheme 9.12 Halohydrin dehalogenase (Agrobacterium radiobacter AD1) catalyzed kinetic

resolution of an aromatic halohydrin.
9.2.4
Ring-Opening Reactions
From the viewpoint of the organic chemist, the enantioselective and regioselective
ring opening of epoxides is highly interesting since it gives access to a wide range of
1,2-difunctionalized compounds [64]. The origin of such studies lie in the observa-
tion of the reversibility of the halohydrin dehalogenase catalyzed ring-opening
reactions. In the first halohydrin dehalogenase study by Castro, the complete
conversion of epichlorohydrin via ring opening with bromide is described [33].
Similar transhalogenation (halogenation combined with dehalogenase) studies were
performed using the halohydrin dehalogenase (DehA) from Arthrobacter erithii H10a
(Scheme 9.13) [58]. The kinetic resolution of racemic epibromohydrin via ring
opening with chloride can yield the remaining (S)-epibromohydrin in >95% e.e.
(14.5% yield). Through ring closure of the intermediate 1-bromo-3-chloro-2-propa-
nol, the formed (S)-epichlorohydrin can be obtained with a maximum enantiomeric
excess of 87.9% (37% yield).
The first ring-opening reaction with a non-halogen nucleophile was described by
Nakamura et al. using halohydrin dehalogenases from Corynebacterium sp. N-
1074 [65, 66]. H-lyase A catalyzed the ring opening of various epoxides with KCN,
yielding the corresponding cyanoalcohols with low optical purity. H-Lyase B proved to
O OH O
DehA DehA
Br Cl Br Cl
excess KCl
(R)-epibromohydrin - HBr (S)-epichlorohydrin
O
Br
(S)-epibromohydrin
Scheme 9.13 Transhalogenation catalyzed by the halohydrin dehalogenase DehA, giving access to
optically enriched (S)-epibromohydrin and (S)-epichlorohydrin.
be more enantioselective in the ring opening of epichlorohydrin, but the optical

purity decreased due to the non-enzymatic ring opening. The product was of interest
since it represents an intermediate in the synthesis of L-carnitine.
A new investigation in the nucleophile range was performed using the
halohydrin dehalogenase from Agrobacterium radiobacter AD1 (HheC) [67]. Inter-
estingly, this research was initiated after the discovery that the presence of azide
originating from the enzyme solution to prevent microbial growth in the buffer
caused side reactions. A range of nucleophiles were screened using the chro-
mogenic substrate para-nitrostyrene oxide (PNSO). This epoxide substrate gives a
decrease in absorbance at 310 nm upon ring opening, independent of the nature
of the nucleophile. A screening of a set of nucleophiles showed that besides I,
Br, Cl, and CN, the enzyme also catalyzed the ring opening by NO2
(0.55 U mg1) and N3 (0.18 U mg1).
Ring openings with either azide or nitrite are novel enzymatic reactions with
considerable synthetic potential. Ring opening by nitrite gives access to either amino
alcohols by reduction of the nitro alcohol formed after N-attack or diols through O-
attack followed by hydrolysis of the nitrite ester. Ring opening with azide gives azido
alcohols, which are direct precursors of amino alcohols.
All nucleophiles that were accepted in the ring-opening reaction (besides cyanide)
were good inhibitors of the ring-closure reaction. To try to find other potential
nucleophiles, competitive inhibition of the nucleophile on the ring closure of the
related para-nitro-2-bromo-1-phenylethanol (PNSHH – para-nitrostyrene halohy-
drin) was studied. Certain nucleophiles that were very good inhibitors of the ring-
closure reactions such as cyanate (OCN) and thiocyanate (SCN) did not appear to
give a productive ring-opening reaction with PNSO.
HheC is not very active on aromatic substrates and since ring opening requires
both the epoxide and the nucleophile to bind into the enzyme simultaneously in a
reactive conformation, the reaction rate can become very slow. This could explain the
lack of detectable reactivity of cyanate and thiocyanate with PNSO. Therefore, a
detailed study was carried out with the much more active aliphatic epoxybutane as
model substrate. With this modification, HheC catalyzed ring opening was dem-
onstrated with cyanate and with thiocyanate (1.9 U mg1). The specific activity of
1.9 U mg1 is considerably lower than reactions with azide (160 U mg1) and nitrite
(18 U mg1) [68]. The enzyme also catalyzed the ring opening by formate (HCOO),
378
showing that nucleophiles accommodating an additional hydrogen atom are accept-

ed. The nucleophile binding site (halide binding site) is not large and rather specific
and no conversion was observed with non-ionic nucleophiles such as water and
alcohols (methanol, ethanol), with amines (methylamine, ethylamine, ethanol-
amine), and with monovalent anions such as (substituted) acetic acids, fluoride,
and nitrate. In addition, no activity was detectable with various bivalent anions. All
possible reactions with the nine accepted nucleophiles and their products are shown
in Scheme 9.14. All accepted nucleophiles are either halides or a selection of small
negatively charged nucleophiles sometimes referred to as pseudohalides. In follow-
up studies the characteristics of ring opening with the various nucleophiles were
studied in more detail.
OH
OH Cl OH
I Br
iodide bromide
178% 52%
OH OH
OCHO NCO O
chloride O
100% NH
OH formate O cyanate
chemical 34% 106% spontaneous
OH hydrolysis
cyclization S
azide
OH 8888% OH
nitrite thiocyanate
ONO 1000% SCN
106%
nitrite cyanide
1000% 139%
OH OH
NO2 CN
OH
N3
Scheme 9.14 Overview of (R)-selective ring opening of racemic epoxybutane with various anions,
catalyzed by the halohydrin dehalogenase HheC. The relative activity of the ring opening is depicted
as a percentage of the reaction with the chloride ion (kcat ¼ 1.8 s1) [68].
The HheC catalyzed ring opening with azide was studied with a range of styrene
oxide derivatives [69]. The azidolysis of para-nitrostyrene oxide was highly enantio-
selective (E > 200), yielding the remaining (S)-epoxide (>99% e.e.) and formed azido
alcohol (96% e.e.) in high optical purity. A second remarkable feature of this reaction
is the high regioselectivity of ring opening. The regioselectivity of the non-enzymatic
azidolysis of terminal styrene oxides is mainly determined by electronic instead of
steric factors. The phenyl group stabilizes the formation of a positive charge at the
benzylic carbon atom (Ca) in the transition state, thereby favoring an attack on this
position over a reaction at the terminal and least substituted carbon atom (Cb). The
chemical reaction between styrene oxide and sodium azide (NaN3) under identical
reaction conditions (room temperature, neutral pH) but in the absence of enzymes
yielded the azido alcohols with a Ca: Cb selectivity of 98: 2. The enzymatic reaction
(with excess HheC to rule out interfering chemical background reaction) showed an
exactly opposite behavior with a Ca: Cb selectivity of 2: 98. The actual observed
selectivity under practical reaction conditions with a catalytic amount of enzyme is
much lower (Ca: Cb about 21: 79) because of the concurrent chemical reaction. In a
preparative-scale reaction, the chemical reaction could be minimized through dosing
of the azide solution to keep the concentration of azide as low as possible. The
halohydrin dehalogenases HheA and HheB also catalyzed the azidolysis reaction but
the enantioselectivity of these reactions was very low.
An explanation for the high enantioselectivity of HheC in the ring opening of
styrene oxides was given in a study of de Jong et al. [70]. Purified HheC enzyme
was crystallized in the presence of either (R)- or (S)-PNSO and the crystal
structures of the complexes were solved. This showed that both enantiomers
were able to bind in the active site of HheC. However, with (S)-PNSO the oxygen
of the epoxide atom is in the wrong position in terms of interacting with Tyr145
and the distance between the nucleophile binding site and Cb is too long,
resulting in an unproductive binding mode. The regioselectivity of ring opening
was studied by Hopmann et al. using density functional theory and quantum
chemical models based on the crystal structure of HheC [71]. Based on the results
of in silico mutations, several residues are proposed to play a pivotal role in
directing the regioselectivity towards Cb.
The copper-catalyzed 1,3-dipolar cycloaddition of azides and alkynes (a click
reaction) affords 1,4-disubstituted triazoles [72]. In a one-pot reaction (Scheme 9.15),
a tandem enantioselective biocatalytic epoxide opening, catalyzed by a mutant of
HheC (C153S), was combined with a [3 þ 2] azide alkyne cycloaddition, yielding the
corresponding triazole in high optical purity [73].
Ph
CuSO4.5H2O N
O HheC (W153S) HO N3 HO N N
Sodium ascorbate
NaN3 MonoPhos
99% e.e.
O2N O2N O2N
Ph
Scheme 9.15 Chemoenzymatic one-pot conversion of an epoxide into a triazole.
Faber et al. have employed a crude enzyme preparation derived from Rhodococcus
sp. for ring opening of an epoxide with azide. This particular Rhodococcus species
contains an epoxide hydrolase and it was postulated that the epoxide hydrolase
catalyzed the ring opening, acting as if it were a halohydrin dehalogenase [74]. It later
became evident that, for mechanistic reasons (the formation and hydrolysis of the
acyl intermediate), it is unlikely to be an epoxide hydrolase activity (Section 9.3.3).
Since the simultaneous expression of epoxide hydrolases and halohydrin dehalo-
genases in one organism is found to be quite common it is likely to be activity of a
true, unidentified, halohydrin dehalogenase. On the other hand, a chiral protein
surface-catalyzed reaction has also been proposed similarly to what is observed in a
lipase-catalyzed ring opening [75, 76].
Hasnaoui et al. studied the HheC-catalyzed ring opening of styrene oxides
employing nitrite as the nucleophile [77]. The product analysis was complex since,
besides enantioselectivity and regioselectivity of the ring opening, the attack of the
nitrite anion can also occur either with the oxygen or the nitrogen atom. Ring
opening with nitrogen yields relatively stable nitro alcohols, while attack with the
oxygen yields nitrite esters that spontaneously hydrolyze to the corresponding
diols. Thus, the total number of possible compounds in the mixture is twelve! The
ring opening of para-nitrostyrene oxide catalyzed by HheC was again highly
enantioselective and regioselective. The nucleophilic opening occurred with an
oxygen: nitrogen regioselectivity of 80: 20. The main product was the nitrite ester,
which was hydrolyzed to the diol. In a kinetic resolution the remaining (S)-
epoxide can be recovered in >99% e.e. (48% yield) while para-nitrophenyl-1,2-
ethanediol was formed with 91% e.e. The nitrite that is released during the
hydrolysis can react again and, in principle, could be used catalytically. These
results show that the halohydrin dehalogenase catalyzed ring opening with nitrite
gives a similar overall result to an enantioselective hydrolysis of an epoxide.
Therefore, a halohydrin dehalogenase can act as an epoxide hydrolase in the
presence of a catalytic amount of nitrite ions. One such application was demon-
strated in the chemoenzymatic optically pure synthesis of the two enantiomers of
chromanemethanol [78]. Enantioselective ring opening (E > 200) with nitrite of
racemic 2-benzyloxymethyl-2-methyl-oxirane yielded the corresponding diol in
96% e.e. The remaining optically pure (S)-epoxide (31% isolated yield) was
employed for the synthesis of the final product (Scheme 9.16). The (R)-epoxide
was made using the epoxide hydrolase from Rhodococcus ruber.
HHDH HheC,
O O
OBn Rhodococcus ruber EH O NO -
2 OBn
OBn
40%, >99% e.e. 31%, >99% e.e.
Scheme 9.16 EH and HHDH routes to the two enantiomers of the epoxide precursor for
chromanemethanol.
Majeric-Elenkov et al. studied the ring opening of various aliphatic epoxides with
NaCN catalyzed by the halohydrin dehalogenases HheA, HheB, and HheC [79]. The
regioselectivity (Ca:Cb) of ring opening was in all cases completely towards Cb. With
the exception of the ring opening of an epoxide bearing a cyclohexyl substituent
(E ¼ 109 with HheA) all reactions catalyzed by HheA and HheB proceeded with an E-
value up to a maximum of 10. With HheC the enantioselectivity varied from E ¼ 5 to
200, depending on the substrate structure. In particular, epoxides bearing two
substituents on Ca were ring opened with very high enantioselectivity. Non-terminal
epoxides such as cyclohexene oxide and 2,3-epoxybutane were not accepted by any of
the three enzymes.
The ring opening by cyanate was investigated using HheC [80]. Literature reports
describing the chemical ring opening of epoxides by cyanate anions and the
subsequent formation of the oxazolidinone are limited. This is likely due to the
poor nucleophilicity of the cyanate anion. Similar to the regioselectivity of attack by
the nitrite anion, cyanate can attack either with the oxygen, yielding the b-hydroxy-
cyanate, or with the nitrogen atom, yielding the b-hydroxy-isocyanate. The cyanate
isomerizes to the isocyanate, which spontaneously cyclizes to oxazolidinones.
Surprisingly, the range of accepted epoxides was far more restricted compared to
the ring-opening reaction with, for example, cyanide or azide. Only linear aliphatic
epoxides of up to four carbon atoms were accepted. Again, epoxides containing a
methyl group on Ca, such as 1,2-epoxy-2-methylbutane, reacted with the highest
enantioselectivity (E > 200), yielding the 2-oxazolidinone in 97% e.e. (Scheme 9.17).
HheC O
O OH NCO
E>200 cyclization
O NH
NaOCN
fast
racemic 97% e.e.

44% yield
Scheme 9.17 Formation of 5-substituted oxazolidinones catalyzed by halohydrin dehalogenase

HheC.
Since the best results were obtained in the ring opening of epoxides bearing a
substituent on Ca, the HheC-catalyzed ring opening was studied in more detail using
cyanide and azide as nucleophile [81]. Unlike other substrates, the non-enzymatic
ring opening with the Ca substituted epoxides is insignificant. This low chemical
activity and the very high enantioselectivity (E > 200) with all tested substrates
allowed the isolation of the b-substituted tertiary alcohols in >99% e.e.
9.2.5
Improving Halohydrin Dehalogenases by Mutagenesis and Evolution
The X-ray structure of the halohydrin dehalogenases HheA and HheC have been
solved by de Jong et al. [31, 82]. Detailed studies towards understanding the catalytic
mechanism, kinetics, and stability yielded mutant enzymes with improved capabil-
ities for their use in organic chemistry [83–86]. The halohydrin dehalogenase HheC is
present in solution both as the catalytically active tetramer and the inactive monomer
form. In the monomeric state the cysteine residues can form intramolecular disulfide
bonds, thereby preventing reversion to the active tetrameric state. Mutations of
selected cysteine residues (C153S and C30A) gave enzymes with drastically improved
stability without affecting the high activity and enantioselectivity [83].
Based on structural data of HheC, four tryptophan residues were identified that
were located near the active site. One tryptophan residue (W249) is positioned close to
the halide binding site. A study of the protein fluorescence of mutants having these
residues changed to phenylalanine provided more insight into the steady state
kinetics and substrate interactions [85]. One of these mutants, W249F, demonstrated
remarkably improved properties. In many other mutagenesis or evolution studies,
improving the enantioselectivity leads to a decreased catalytic activity. However, the
single mutation W249F in HheC resulted in an enzyme with a sixfold improvement
of enantioselectivity and 40% increase in activity towards the enantioselective ring
closure of PNSHH. The superiority of the use of W249F was also demonstrated in
ring-opening reactions. Compared to wild-type HheC, the enantioselectivity of the
ring opening of substituted styrene oxide derivatives with nitrite increased in the
range twofold (ortho-chloro) to 39-fold (para-methyl) [77].
Codexis (USA) optimized HheC for the two-step conversion of ethyl (S)-4-chloro-
3-hydroxybutyrate, synthesized in a preceding step using an evolved alcohol dehy-
drogenase, into ethyl (R)-4-cyano-3-hydroxybutyrate (HN, ATS-5, Scheme 9.18a) [87].
HN is of commercial interest since it is used as an intermediate for the production of
the cholesterol-lowering drug atorvastatin (Lipitor/Pfizer). The volumetric produc-
tivity of HheC could be improved by a factor of 4000 using a combination of
mutagenesis techniques such as error-prone PCR, site-directed mutagenesis,
focused evolution, and shuffling, all in combination with ProSAR analysis. The
final variant of HheC carried at least 35 mutations (15% of total), including only four
residues in the active site.
(a) O OH Mutant O Mutant O OH

HheC O HheC
Cl CN
O O O
NaCN 'HN'
- HCl (R)-ethyl 4-cyano-3-
hydroxybutanoate
(b) Mutant Mutant
O OH O O OH
HheC O HheC
Cl CN
O O O
NaCN
(S)-methyl 4-cyano-3-
- HCl
hydroxybutanoate
Scheme 9.18 Two-step halohydrin dehalogenase catalyzed synthesis of HN (a) and an analogous
sequential kinetic resolution (b), leading to ester products with opposite configuration.
A similar approach was described starting with methyl 4-chloro-3-hydroxybutyrate

(Scheme 9.18b) [88]. In a sequential resolution process, this racemic chloroalcohol is
converted into methyl (S)-4-cyano-3-hydroxybutyrate in 96.8% e.e. (40% yield). The
unreacted enantiomer methyl (S)-4-chloro-3-hydroxybutyrate could be obtained in
95.2% e.e. (41% yield). The kinetic resolution of the intermediate racemic epoxide
with wild-type HheC had a moderate E-value of 15, yielding the product in a low
optical purity. Again the HheC-W249F mutant proved to be superior, with an E-value
of 60.
Note that although Codexis and Majeric-Elenkov et al. use mutants of the same
HheC, the end products of the two routes are of opposite stereochemistry. In the
sequential kinetic resolution of Majeric-Elenkov, the selectivity of the halohydrin
dehalogenase determines the stereochemical outcome whereas in Codexis approach
the chirality of the halohydrin is set in the preceding alcohol dehydrogenase step and
the halohydrin dehalogenase actually converts the unpreferred enantiomer.
9.2.6
Towards 100% Yield
The advantage of a kinetic resolution strategy is the ability to isolate the remaining
enantiomer optically pure, even if the enantioselectivity of the enzymatic reaction is
only moderate (E > 10). In some cases, having access to the molecule in >99% e.e. is
more relevant than the yield, but more often than not there is severe pressure to find
the most cost-effective synthetic route. The advantage of an asymmetric synthesis
approach to preparing optically pure epoxides, such as reduction of haloketones
followed by a base-catalyzed ring closure, is the ability to obtain the target molecule in
close to 100% yield. Since there are so many different alcohol dehydrogenases it is
often possible to find one that has the desired enantioselectivity. However, finding the
optimal enzyme often requires a significant screening effort.
One approach for halohydrin dehalogenases is to start from a prochiral substrate.
For example, Nakamura et al. described the two-step conversion of the prochiral 1,3-
dichloro-2-propanol, via optically enriched epichlorohydrin, to (R)-3-chloro-1,2-
propanediol (84% e.e. and 97% yield) using whole cells of Corynebacterium sp. N-
1074 [89]. In a similar approach starting from the same substrate, the formed
epichlorohydrin is ring opened again with cyanide to yield the b-hydroxynitrile in
95.2% e.e. and 65.3% yield. Isolation of optically pure epoxide is not possible due to
racemization [66].
The above-described HheC catalyzed racemization of epihalohydrins was
exploited for a DKR approach by combining it with an azidolysis reaction catalyzed
by the same enzyme (Scheme 9.19) [90]. During the reaction the e.e. of the substrate
epibromohydrin remained below 20%, indicating an efficient racemization.
O fast OH slow O
Br N3
N3 Br
Br - N3- Br - N3-
OH >99% e.e. OH
Br Br
racemisation kinetic 77% yield kinetic N3 N3
of epoxide resolution resolution
N3 -
Br - O
OH
fast O
Br slow N3 Br N3
N3- Br -
Scheme 9.19 Overall reaction scheme of the combined DKR and kinetic resolution of racemic
epibromohydrin yielding (S)-1-azido-3-bromo-2-propanol in >99% e.e. and 77% yield. All reactions
are catalyzed by HheC.
384
The E-value of the kinetic resolution was only moderate, yielding the product (S)-1-
azido-3-bromo-2-propanol in 94% e.e. However, on prolonged incubation the (R)-
1-azido-3-bromo-2-propanol was selectively ring closed to the epoxide. As a result of
this one-pot DKR and follow-up kinetic resolution, (S)-1-azido-3-bromo-2-propa-
nol could be obtained in >99% e.e. and 77% yield.
The above-described approach was also tested using cyanate as nucleophile, but
this reaction was hampered by the instability of the substrate and product [80].
A DKR approach is an in situ racemization of the substrate combined with a KR.
The transition metal catalyzed racemization of secondary alcohols such as b-chlor-
oalcohols has been described [91]. Haak et al. [99] combined such a halohydrin
racemization with the enantioselective ring closure catalyzed by HheC in a water–-
toluene biphasic reaction system. In this case, an activated ruthenium catalyst,
present in the organic phase, racemizes the halohydrin. The optimal halohydrin
dehalogenase was HheC carrying mutations C153S for improved stability and
W249F for improved activity and enantioselectivity. The combination of these two
catalysts resulted in an effective DKR yielding epoxides such as styrene oxide in
98% e.e. at 90% conversion.
9.2.7
Cascade Reactions Using Multiple Enzymes
In the preceding paragraphs, we have discussed several enzyme classes that can be
applied in synthetic strategies towards epoxides or derivatives. The chemistries that
these enzymes catalyze are often compatible with each other and when combined
give the potential for uni-directional reaction cascades. A cascade can be practised in
separate reaction vessels, with or without isolation of the products, or simultaneously
in one pot.
9.2.7.1 Haloperoxidase and Halohydrin Dehalogenase

At the end of the 1970s, the halohydrin dehalogenase from Flavobacterium sp. was
used in an industrial process to make propylene oxide from propylene [92]. This
Cetus Process is known to be the first bioconversion process for an application in
the petrochemical industry [93]. The process was developed as an alternative to
chemical processes such as direct oxidation with an expensive and toxic catalyst, or via
the intermediate halohydrin employing the use of gaseous halogen and generating
side products. The biocatalytic alternative shown in Scheme 9.20 used three different
enzymes. In the first step propylene is converted into the corresponding halohydrin,
catalyzed by the chloroperoxidase from Caldariomyces fumago. In the second step the
Halohydrin
Chloroperoxidase OH dehalogenase O
Cl
H2O2
Scheme 9.20 Outline of the conversion of propylene into propylene oxide according to the Cetus
Process.
1-chloro-2-propanol is converted into the epoxide, catalyzed by the halohydrin
dehalogenase from Flavobacterium sp. [94]. To improve the stability of the biocatalyst
the whole cell preparation of Flavobacterium sp. was entrapped within crosslinked
polyacrylamide gel. This allowed a column packed with the immobilized catalyst to
run continuously for at least three months [35]. The hydrogen peroxide needed for the
first step was generated in situ via the conversion of glucose catalyzed by a glucose-
2-oxidase. The latter has as advantage that the product of this oxidation could easily be
converted by an additional chemical step into D-fructose, which was considered as a
valuable side product.
The Cetus Process was not operated on an industrial scale for economic reasons
such as availability of the biocatalysts. In addition, technical issues such as the
difference in pH optimum between the two enzymes and operational stability needed
to be overcome.
9.2.7.2 Halohydrin Dehalogenase and Epoxide Hydrolase

The conversion of 1,3-dichloropropanol by a halohydrin dehalogenase from Cory-
nebacterium sp. N-1074 (H-lyase B) yielded (R)-epichlorohydrin in higher than 90%
e.e. in the initial stages of the reaction [49]. However, due to racemization of
epichlorohydrin catalyzed by the same enzyme, the optical purity drastically
decreased during the reaction. When applying resting cells of Corynebacterium,
1,3-dichloro-2-propanol was converted into epichlorohydrin, which was hydrolyzed
to 3-chloro-1,2-propanediol with catalysis by an epoxide hydrolase [48]. The optical
purity of the final product decreased with increasing substrate concentrations. This
loss of optical purity was caused by an increase in the transient epichlorohydrin
concentration causing a higher extent of racemization. Through optimizing the pH
and employing a substrate feeding strategy that involved adding five subsequent
doses of 15.5 mM, (R)-3-chloro-1,2-propanediol could be obtained in 84% e.e. and
97% yield (Scheme 9.21) [95].
Microorganisms harboring a combination of halohydrin dehalogenase and epox-
ide hydrolase are often found. Thus, a similar cascade reaction was developed for the
kinetic resolution of 1-chloro-2-phenylethanol catalyzed by the recombinant enzymes
HheC and EchA originating from Agrobacterium radiobacter AD1. Because of the
reversibility of the ring-closure reaction, conversion of the preferred (R)-1-chloro-2-
phenylethanol stalled at around 95% conversion. By in situ removal of the epoxide
through the addition of an epoxide hydrolase, the remaining halohydrin (S)-1-chloro-
2-phenylethanol could be obtained in >99% e.e. [61].
OH halohydrin epoxide OH
dehalogenase O hydrolase
Cl Cl Cl Cl OH
prochiral (R)-3-chloro-1,2-propanediol
1,3-dichloro-2-propanol 97% yield, 84% e.e.
Scheme 9.21 Cascade conversion of 1,3-dichloropropanol with substrate dosing to give (R)-
3-chloro-1,2-propanediol catalyzed by resting cells of Corynebacterium sp. N-1074.
386
9.2.7.3 Alcohol Dehydrogenase and Halohydrin Dehalogenase

Two examples of cascade reactions with an alcohol dehydrogenase (ADH) and a
halohydrin dehalogenase have been published by the group of Kroutil
(Scheme 9.22). Optically pure chlorohydrins were obtained via stereoselective
reduction of prochiral chloroketones. Both enantiomers could be prepared by
selecting the alcohol dehydrogenase with the desired enantiopreference (ADH-LB
or ADH-A). The cofactor is recycled in situ by the use of isopropanol (IPA) as
sacrificial cosubstrate. Through the addition of the non-enantioselective halohy-
drin dehalogenase HheB, the formed halohydrin was converted into the corre-
sponding epoxide with complete conservation of optical purity (>99% e.e.) [96]. The
equilibrium of the reaction could be shifted to the epoxide side by addition of
hydroxide-loaded anion exchange resins to remove the chloride ions that were
liberated during epoxide formation [97]. The removal of chloride also alleviated any
halide product inhibition. In another approach to overcome the reversibility of the
ring-closure reaction, the formed epoxide was ring opened using the nucleophiles
azide and cyanide. In such a three-step, two-enzyme, one-pot reaction cascade
optically pure b-azido-alcohols and b-hydroxynitriles could be obtained in >99%
(analytical) yield and >99% e.e. (Scheme 9.22) [98].
O OH OH
ADH A HheC O HheC
Cl Cl N3
R R R R
NADP NaN3
IPA - HCl
>99% e.e. >99% e.e. >99% e.e.
Scheme 9.22 One-pot cascade reaction towards optically pure azido-alcohols starting from
prochiral chloroketones.
9.2.8
Outlook on Halohydrin Dehalogenases
Halohydrin dehalogenase catalyzed ring opening with alternative nucleophiles such

as azide and cyanide was discussed in Section 9.2.4. Alternative leaving groups in the
ring-closure reaction have only cursorily been studied. Wandel et al. tested the
halohydrin dehalogenase from Flavobacterium sp. towards a range of b-substituted
alcohols and Haak et al. carried out similar experiments using formate as a leaving
group [56, 99]. As with fluorohydrins, in both studies no conversion was observed.
Thus, although the extent of these studies is limited, it is likely that the scope of the
leaving group is limited to the halides chloride and bromide. Iodohydrins are too
chemically unstable for practical applications and often give elimination to
the alkene.
The use of azide and the manufacture of, especially, small aliphatic azides are very
challenging for the chemical industry because of toxicity and risk of explosion,
respectively. To direct the regioselectivity of ring opening one usually needs Lewis
acid assistance or acidic pH, leading to explosive metal-azides or explosive hydrazoic
acid HN3. In comparison, the corrosive nature of azide is only a minor nuisance,
requiring glass-lined stainless steel reactors. Typically, an azide salt is used in excess
relative to the epoxide. Enzymatic ring opening does not have any of these dis-
advantages: a 10% excess of azide reacts smoothly with the epoxide, in water at
neutral pH. The regioselectivity is excellent and product purification is relatively
straightforward. The produced azido-alcohol can also be converted in situ into, for
example, a triazole or an amino alcohol [73, 100].
What is the limit of the scope of accepted nucleophiles? Most nucleophiles that are
accepted in the ring-opening reaction also inhibit the ring-closure reaction. However,
some organic acids such as acetate, chloroacetic acid, iodoacetic acid, 2-chloropro-
pionic acid, and 3 chloropropionic acid are inhibitors of the forward reaction. Since
the related formate inhibits but is also accepted as a nucleophile it is likely that these
other acids also bind in the halide binding site. A thorough screening with various
epoxides and excess enzyme to increase reaction rate or a possible directed evolution
study might allow identification of appropriate enzyme–substrate combinations. In
particular, chiral nucleophiles such as 2-chloropropionic acid can open up a whole
new application field.
A curious side activity is the conversion of chloroacetone into hydroxyacetone
(Scheme 9.23). The mechanism has not yet been elucidated, but might involve a ring
closure and rearrangement after the initial chemical hydration to the gem-diol.
Understanding this reaction and the substrate range will be of value for further
investigations into alcohol dehydrogenase/halohydrin dehalogenase tandem reac-
tions since that strategy starts from similar substrates.
Halohydrin
O dehalogenase O
Cl OH
- HCl
Scheme 9.23 Halohydrin dehalogenase catalyzed dehalogenation of chloroacetone.
The enantioselective ring-closure reaction is a valuable tool for accessing optically

pure halohydrins and their corresponding epoxides. The enzymatic ring closure can
also serve as a very mild alternative for ring closure of base-sensitive substrates.
Although the pH-optimum of the ring closure is around pH 8.0, the enzymatic
activity at pH 6.5 is sufficient for synthetic applications. A disadvantage of HheC for
such applications is the high enantioselectivity, resulting in long reaction times if the
unpreferred enantiomer needs to react. The HheC mutant W139F is an optimal
alternative since this enzyme lost enantioselectivity for all tested substrates, while
retaining high activity.
Release of halide in the ring-closure reaction can have an inhibiting effect on
overall conversion. The use of anion exchange resins is a relatively straightforward
technological solution to this halide inhibition. Halohydrin dehalogenase catalyzed
ring-opening reactions do not suffer from substrate or product inhibition, unlike in
the case of epoxide hydrolases where the produced diol often acts as an inhibitor.
388
9.3
Hydrolysis of Epoxides Catalyzed by Epoxide Hydrolases
9.3.1
Epoxide Hydrolases in Nature
The role of epoxide hydrolases in Nature is threefold: detoxification, catabolism, and

regulation of signaling molecules [101, 102]. In eukaryotes, microsomal and cytosolic
epoxide hydrolases play a key role in the detoxification of mutagenic, poisonous, and
carcinogenic epoxides, which are formed by the action of cytochrome P450 mono-
oxygenases [103, 104]. In addition, epoxide hydrolases are involved in the biosyn-
thesis of hormones (e.g., leukotrienes) and play a role in the control of blood pressure,
inflammation, and cell proliferation [105]. Five different types of mammalian epoxide
hydrolases have been identified: microsomal epoxide hydrolase (mEH), soluble
epoxide hydrolase (sEH), cholesterol epoxide hydrolase, leukotriene A4 epoxide
hydrolase, and hepoxilin hydrolase [106].
The substrate specificity and regulatory behavior of the plant soluble epoxide
hydrolases (EHs) point to a physiological role of these enzymes in host-defense and
growth. The defensive functions of these enzymes can be related to both passive
(cutin biosynthesis) and active (antifungal chemical synthesis) roles [107]. Epoxide
hydrolases (EHs) from Solanum (potato) and Arabidopsis are the most extensively
described, but a study has shown that EHs are rather ubiquitous [108].
Insect epoxide hydrolases degrade juvenile hormones and pheromones bearing an
oxirane moiety [102]. A recent case describes an EH that is involved in the
biosynthesis of an epoxide-containing pheromone by optical enrichment of a fatty
acid epoxide, converting one enantiomer into the diol [109].
In higher organisms, epoxide hydrolases often perform a very specific task and
have a very limited substrate range such as converting only leukotriene A4 into
leukotriene B4. In contrast, in microorganisms these enzymes are multifunctional:
(i) they can function as detoxifying agents, (ii) they can play a role in biosynthetic
routes of complex (secondary) metabolites, or (iii) they may be crucial for the
degradation of epoxides during the metabolism of alkenes and aromatics [110].
Although early epoxide hydrolase activities had been detected in bacteria or fun-
gi [147, 148] quite some years ago [111–114], for a long time it was assumed that
epoxide hydrolases are predominantly found in mammals [115]. Nowadays, it is
generally accepted that epoxide hydrolases occur in all Kingdoms of Life.
While epoxide hydrolases from mammalian, plant, and insect sources were mainly
studied to gain insight into their physiological roles, newly discovered microbial
enzymes were applied for the synthesis of optically active epoxides and diols. Over the
last 15 years many bacterial, fungal, and yeast epoxide hydrolases have been
characterized [116, 117]. The epoxide hydrolase from Agrobacterium radiobacter AD1
was the first bacterial enzyme to be cloned, overexpressed, and characterized [247].
Since then many enzymes have been cloned and subjected to detailed study, such as
the epoxide hydrolase from Corynebacterium sp. C12 [118], Rhodotorula glutinis ATCC
9.3 Hydrolysis of Epoxides Catalyzed by Epoxide Hydrolases j389
201718 [119], Rhodococcus erythropolis DCL 14 [120], Aspergillus niger LCP 521 [121],
and various other yeast [122–126], fungi [127, 236], and bacteria [128–132, 141].
9.3.2
Discovery of Novel Microbial Epoxide Hydrolase Activity
Novel epoxide hydrolases can be obtained using various strategies. New enzymes
were found in biodegradation studies, where organisms were identified that
could degrade environmental pollutants such as epoxides [110]. For example, the
organism Agrobacterium radiobacter expressing a highly active and enantioselective
epoxide hydrolase was identified by its ability to grow on epichlorohydrin as its sole
carbon source [133]. Kotik et al. screened 270 microbial isolates from biofilters and
petroleum-polluted bioremediation sites and identified various strains that were able
to enantioselectively degrade epoxides [146]. The strain Rhodococcus erythropolis
DCL14 was isolated from a sediment sample because of its ability to grow on
limonene as sole source of carbon and energy. The epoxide hydrolase converting
limonene-1,2-epoxide into the corresponding diol belongs to a novel class of epoxide
hydrolases that acts through non-covalent catalysis (Section 9.3.3) [134, 135]. A more
general, and the most common, method of finding enzyme activity is by screening of
culture collections for the desired activity [136]. The most intensively studied fungal
epoxide hydrolases come from two distinct groups of microorganisms: (i) the
filamentous fungus Aspergillus niger and (ii) the basidiomycetous red yeasts belong-
ing to the genera Rhodotorula and Rhodosporidium [116, 137, 138]. Initial screening of
bacterial strains by the group of Faber yielded hits from genera such as Rhodococcus,
Mycobacterium, and Nocardia [139].
Research towards identifying new epoxide hydrolases yielded insight into their
structure and mechanism as well as their amino acid sequences. Most of the
sequenced epoxide hydrolases are members of the a/b hydrolase fold family, to
which certain enzymes from other classes like haloalkane dehalogenases, lipases,
and esterases also belong. Now that many microbial genome sequences are available
in databases, in silico screening has all but replaced laboratory-based screening. The
databases can be screened for enzymes using known epoxide hydrolase (partial)
sequences as a query. Putative epoxide hydrolases can be distinguished from
structurally related classes of enzymes using conserved epoxide hydrolase sequence
motifs that define the active site. A database compiling sequence and structure
information of epoxide hydrolases has been created and can be accessed [140]. Van
Loo et al. analyzed various genomic databases and showed that around 20% of all
sequenced organisms contain one or more putative epoxide hydrolase genes [141].
Expression of a number of these genes and testing them on model substrates showed
that 60% of them were indeed functional epoxide hydrolases. The activity and
enantioselectivity of the active enzymes were tested with a diverse set of model
substrates. Enzymes could be identified that had an opposite enantiopreference ((R)-
instead of (S)-specific) compared to the best described epoxides hydrolases such as
mammalian epoxide hydrolase and ones obtained from A. radiobacter AD1, A. niger,
390
(a) Tyr152 (b) Tyr152

Tyr215 Tyr215
O H O-
H O H O
OH
O
O O
-O O O O
O- H H O- H H
N N
N H O N H O
Asp246 H N Asp246 H N
Phe108 Phe108
Asp107 Asp107
His275 His275
Figure 9.2 Reaction mechanism of the epoxide hydrolase from Agrobacterium radiobacter AD1: (a)
alkylation reaction and (b) hydrolysis of covalent intermediate.
and Rhodotorula glutinis. In addition, unusually high enantiopreference towards

meso-epoxides was demonstrated. Similar strategies allowed the identification of new
epoxide hydrolases from various microbial sequence data [142–144]. Diversa (USA)
discovered over 50 novel microbial epoxide hydrolases by screening environmental
DNA libraries, which were created from numerous global habitats.2) The enzymes
showed a diverse substrate scope and, especially, meso-epoxides were hydrolyzed with
high selectivity [145]. Kotik et al. amplified a novel epoxide hydrolase encoding gene
directly from the metagenome using genome-walking PCR [146].
9.3.3
Structure and Mechanism of Microbial Epoxide Hydrolases
Most known epoxide hydrolases are a/b-hydrolase fold enzymes [147]. The topology
of this class of enzymes shows two domains. The main domain consists of a central
b-sheet surrounded by a-helices. The first crystal structure of an epoxide hydrolase
was solved for the enzyme from Agrobacterium radiobacter AD1 (EchA) in 1999 [148].
Prior to that, the sequence similarity to haloalkane dehalogenases, of which the
structure and mechanism had already been studied, allowed a strong hypothesis on
the structure and mechanism of epoxide hydrolases [149, 150]. The most prominent
data that the structure of EchA provided was that two tyrosine residues are positioned
in such a way that they can serve as substrate activators and proton donors. One of the
tyrosine residues (Tyr215) was shown to be conserved in all related epoxide hydro-
lases. All the common bacterial epoxide hydrolases share the same overall fold, the
conserved tyrosine residues, and active site with a catalytic triad consisting of one
histidine and two aspartate residues. The mechanism shown in Figure 9.2 is based on
the crystal structure of EchA, but is analogous for all other a/b-hydrolase fold epoxide
hydrolases. Upon binding of the epoxide, the two tyrosine residues form hydrogen
bonds to the epoxide, thereby activating the epoxide ring for nucleophilic attack. In
the first half-reaction Asp107 attacks the epoxide ring, thereby forming a covalent
2) Diversa merged with Celunol (USA) in 2007 and continued activities as Verenium (USA).
intermediate. In the second half-reaction, facilitated by the charge relay system of
Asp246, the histidine residue (His275) abstracts a proton from the water molecule
and hydrolysis occurs.
Two exceptions that do not follow this general mechanism are the epoxide
hydrolases obtained from Rhodococcus erythropolis (also known as limonene
epoxide hydrolase) and Mycobacterium tuberculosis [151–154]. Although the struc-
ture and catalytic triad of these two enzymes are different, they have in common
that the epoxide hydrolysis occurs without the formation of a covalent
intermediate.
9.3.4
Practical Application of Epoxide Hydrolases to the Synthesis of Chiral
Epoxides and Diols
Table 9.2 outlines general concepts for the use of epoxide hydrolases in the
synthesis of chiral epoxides. Reaction A, the resolution of a racemic mixture of
epoxides with an enantioselective epoxide hydrolase, is the most simple and
frequently used method. After complete reaction of the fast reacting enantiomer
the reaction is stopped and the remaining epoxide is obtained optically pure. Even
if the E-value is moderate, the epoxide can be isolated in high optical purity
(>99%), at the cost of a somewhat lower yield. An E-value of >200 gives the
epoxide in a theoretical yield of close to 50%, while a moderate E-value of 25 still
gives a yield of close to 40%. For many applications, a kinetic resolution strategy
giving >99% e.e. with a 30% yield is more desirable than an asymmetric reaction
yielding the product in 100% yield, but only 95% e.e.
In reactions B and C, the kinetic resolution yields the optically active diol.
Depending on whether the attack occurs on either the a or b position, the diol is
formed with inversion or retention of configuration. In many applications the
obtained diol is not optically pure (>99%) due to a low regioselectivity of ring
opening. The regioselectivity is defined as b/a and this ratio should preferably be
above 200 (meaning less than 0.5% of the other regioisomer). Low regioselectivity is
either a characteristic of the enzyme or is caused by a background chemical reaction.
The concepts depicted as D, E, and F are examples of enantioconvergent reactions
where the theoretical yield can be 100%. These concepts will be discussed in
Section 9.3.7. Concept G shows conversion of a meso-epoxide with an inherent
theoretical yield of 100%.
The enantioselectivity of a kinetic resolution is most commonly defined by the
enantioselectivity ratio, the E-value. In general, the E-value can be calculated from
either the conversion and optical purity of the substrate (ees) or the conversion and
e.e. of the product (eep). However, whereas the ees at a certain degree of conversion is
defined by the E-value alone, the eep depends on the enantioselectivity and regios-
electivity [155]. Therefore, the enantioselectivity of an enzymatic hydrolysis of an
epoxide is most accurately calculated using conversion and ees. Alternatively, the E-
value can be calculated by taking regioselectivity into account, by the use of a
retention–inversion ratio [156].
392
Table 9.2 Concepts for synthesis of optically pure diol or epoxide using epoxide hydrolases.
Reaction scheme Best possible goal Desired selectivity Possible bottlenecks
A (R) (R)-Epoxide. E > 50. If E < 50: decrease in yield.

(R) O O
50% yield. % a/b not relevant If chem. conversion: decrease in yield
α >99% e.e.
β
(S) O
B (R) O (S)-Diol: E > 200. If E < 200: decrease in e.e.

50% yield. % b > 99.5 If b/a < 200: decrease in e.e.
>99% e.e. If chem. conversion decrease in e.e.
OH
(S) O retention
OH
(S)
C (R) O (R)-Diol: E > 200. If E < 200: decrease in e.e.

50% yield. % a > 99.5 If a/b < 200: decrease in e.e.
>99% e.e. If chem. conv.: decrease in e.e.
OH
(S) O inversion
OH
(R)
D OH (S)-Diol: E < 2. If E > 10: long reaction time.
(R) O inversion
OH 100% yield. % a > 99.5. If a/b < 200: decrease in e.e.
(S) >99% e.e. % b > 99.5 If b/a < 200: decrease in e.e.
If chem. conversion: decrease in e.e.
OH
(S) O retention
OH
(S)
OH
E (R) O inversion (S)-Diol. A and B: If E < 200: decrease in e.e.
OH 100% yield. E > 200. If % a/b < 200: decrease in e.e.
Epoxide hydrolase A
(S) >99% e.e. % a > 99.5 If % b/a < 200: decrease in e.e.
OH % b > 99.5 If chem. conversion decrease in e.e.
(S) O retention
OH
Epoxide hydrolase B
(S)
F (R) O (R) O (S)-Diol. If E < 200: decrease in e.e.

100% yield. If a/b < 200: decrease in e.e.
>99% e.e. If b/a < 200: decrease in e.e.
inversion acid If chem. conversion: decrease in e.e.
OH
(S) O retention
OH
(S)
G OH (R,R)-Diol. If chem. conversion: decrease in e.e.

O 100% yield.
>99% e.e.
OH >99% d.e.
9.3 Hydrolysis of Epoxides Catalyzed by Epoxide Hydrolases
meso (R,R)
j393
394
During the last 15 years many applications have been described that were aimed at
the preparation of optically pure epoxides by means of kinetic resolution. Table 9.3
gives an overview of kinetic resolutions yielding the epoxide in at least 98% e.e. and
20% yield. These arbitrary values were set as the lower limit for practical application
in the synthetic laboratory, not as an indication of economic feasibility. In reports
where several closely related compounds are tested, only one representation is
entered in the table.
The pioneering work (1990–2000) towards identifying enantioselective epoxide
hydrolases in various organisms and applying them for practical purposes was
undertaken by researchers such as Belluci, Faber, Furstoss, Janssen, de Bont, and
Weijers [157–159]. Two recent reviews give an overview of the acceptance of various
substrates by epoxide hydrolases [160, 161].
The enantioselectivity of ring opening using mammalian epoxide hydrolases
(mEH and sEH) has been investigated since 1970. Belluci et al. studied the
enantioselective hydrolysis of various cycloalkane oxides and epoxides of hetero-
cycles. The enantioselectivity of the hydrolysis of substituted styrene oxides by sEH
and mEH is usually low to moderate [162, 163]. An exception is the mEH-catalyzed
hydrolysis of cis-b-substituted styrene oxides. Rabbit liver mEH catalyzed the
hydrolysis of cis-b-methyl styrene oxide and cis-b-ethyl styrene oxide with very high
enantioselectivity, yielding nearly optically pure (1S,2R)-epoxides and (1R,2R)-
diols [164]. The same trend can be observed with aliphatic epoxides as mostly cis-
b-substituted aliphatic epoxides are hydrolyzed by mEH with high enantioselec-
tivity [165–167].
In one of their first studies, Faber et al. discovered that a commercially available
immobilized enzyme preparation from Rhodococcus sp. displayed epoxide hydrolase
activity [168]. Although the enantioselectivity of this strain was low, it prompted a
comprehensive screening for enantioselective epoxide hydrolases in bacteria [169].
The highest enantioselectivity was observed with aliphatic 2,2-disubstituted epox-
ides [170, 171]. The group of de Bont showed that limonene-1,2-oxide hydrolase from
Rhodococcus erythropolis DCL14 is a novel type of enzyme. It has a rather narrow
substrate specificity. Of the compounds tested, only the natural substrate limonene-
1,2-oxide and several highly substituted (alicyclic) epoxides were substrates for the
enzyme. The enantioselectivities were usually low, except for the (4R)- or the (4S)-
limonene-1,2-epoxide diastereomers (Table 9.3, 40) and a spiroepoxide (Table 9.3,
39) [172, 211].
The first application of a recombinant microbial epoxide hydrolase was reported by
the group of Janssen. The epoxide hydrolase obtained from Agrobacterium radiobacter
AD1 (EchA) was cloned and brought to overexpression in E. coli [247]. A range of
substituted styrene oxides was hydrolyzed with moderate enantioselectivity to give
the (S)-styrene oxides in 27 to 36% yield [173]. Aliphatic epoxides were converted with
low enantioselectivity (E < 5). Remarkable behavior was observed with the enantio-
selective hydrolysis of styrene oxide (E ¼ 16). In experiments with separate enantio-
mers, the (S)-enantiomer was hydrolyzed at a close to threefold higher rate than the
(R)-enantiomer. However, in a kinetic resolution experiment, the (R)-enantiomer is
hydrolyzed initially, due to a 45 times lower Km of this enantiomer [174].
Table 9.3 Overview of wild-type epoxide hydrolase catalyzed kinetic resolution of racemic epoxides, yielding the epoxide in at least 98% e.e. and 20% yield.
O O O O O
1 (S) Rhodotorula arau- 2 (S) Rhodotorula glutinis 3 (S) Rhodotorula glutinis 4 (S) Rhodotorula glutinis 5 (S) Rhodotorula glutinis
cariae CBS 6031; R. glu- ATCC 201718; Rhodospori- ATCC 201718; Rhodospori- ATCC 201718 [178] ATCC 201718 [178]
tinis ATCC 201718; Rho- dium toruloides UOFS Y- dium toruloides UOFS Y-
dosporidium toruloides 0471 [178, 179] 0471 [178, 179]
UOFS Y-0471; Chryseo-
monas luteola [137, 178–
180]
O O O O O
Br Br Br Cl
6 (S) Aspergillus niger LCP 7 (R) Nocardia H8 [182] 8 (S) Yarrowia lipolytica [183] 9 (S) Yarrowia lipolytica [183] 10 (S) Aspergillus niger;
521 and other versions; (R) Rhodotorula glutinis in
(R) Nocardia TB1 [139, pichia [184, 185]
181]
O O O
O O O O
O
11 (R,R) Rhodotorula 12 (R,S) Rhodotorula glutinis 13 (R,R) Rhodotorula glutinis 14 (R) Acinetobacter 15 (R) Aspergillus niger
glutinis ATCC 201718; ATCC 201718 [175] ATCC 201718 [175] baumannii [187] M200 [188]
Xanthobacter Py2 [175,
186]
(Continued )
j395
396
Table 9.3 (Continued )

O O O O O
R
R R
16 (S) Rhodotorula gluti- 17 (S) Agrobacterium radio- 18 Cl(S) Agrobacterium 20 Cl(S) Agrobacterium 21 Cl (S) Agrobacterium
nis ATCC 201718; Asper- bacter AD1 [173] radiobacter AD1; Sphingomo- radiobacter AD1; Sphingomo- radiobacter AD1; Sphin-
gillus niger LCP 521 and nas sp. HXN-200 [173, 195] nas sp. HXN-200 [173, 195] gomonas sp. HXN-
other versions; Agrobac- 19 NO2(S) Aspergillus niger 200 [173, 195]
terium radiobacter AD1; CGMCC0496 [196] 22 NO2(S) Aspergillus
Rhodosporidium kratoch- niger LCP 521 and other
vilovae SYU-08; Danio versions; Yarrowia lipoly-
rerio; Sphingomonas echi- tica; (R)Yarrowia lipoly-

noides EH-983 [173, 178, tica [183, 197, 198]
181, 189–194] (R) Beau- 23 SCF3(S) Aspergillus
veria sulfurescens ATCC niger LCP 521 and other
7159; Caulobacter crescen- versions [199]
tus; Sphingomonas sp.
HXN-200 [173]
Cl
O O O O
O
N Br
F F
24 (S) Aspergillus niger 25 (S,R): rabbit mEH [164] 26 (R,R) Rhodotorula glutinis 27 (S) Aspergillus niger LCP 28 (S) Aspergillus niger
LCP 521 and other ver- ATCC 201718; Rhodotorula 521 and other versions [203] LCP 521 and other
sions; Agrobacterium glutinis UOFS Y-0123 [175, versions [204]
radiobacter AD1 [200, 201] 202]
O O
O O
O O
O O
29 (R) Aspergillus niger; 30 (R) Aspergillus niger [184] 31 (R) Aspergillus niger [184] 32 (R) Rhodococcus ruber CBS
Agrobacterium radiobacter 717.73; Bacillus subtilis JCM
AD1 [173, 184] 10629 [78, 205]
O O O
O O
O
33 (S) Rhodotorula gluti- 34 (R,S) Rhodotorula glutinis 35 (R,S) Beauveria sulfurescens 36 (2S,3R): Phaseolus
nis SC 16293 [206] ATCC 201718; Beauveria sul- ATCC 7159 [207] radiatus [208]
furescens ATCC 7159 [175,
207]
O O O
O
O
O
37 (3S,4R) Rhodotorula 38 (R) Aspergillus niger LCP 39 (R,R) Aspergillus niger LCP 40 (1S, 2R, 4S) Rhodotorula
glutinis ATCC 521 and other versions [210] 521 and other versions; (S,S): glutinis ATCC 201718; Rho-
201718 [209] Rhodococcus erythropolis [211] dococcus erythropolis [172, 175]
j397
398j 9 Hydrolysis and Formation of Epoxides
The group of Weijers demonstrated the epoxide hydrolase activity of Rhodotorula
glutinis on several aryl, alicyclic, and aliphatic epoxides [175]. In follow-up studies,
additional yeast strains exhibiting good activities and sufficient enantioselectivities
were found – the application of these biocatalysts has great potential [176].
The first studies towards identifying enantioselective fungal epoxide hydrolases
were performed by the group of Furstoss (Scheme 9.24). Styrene oxide was hydro-
lyzed by Aspergillus niger LCP521 with moderate enantioselectivity, affording the (S)-
epoxide in 99% e.e. [177]. In contrast, the fungus Beauveria bassiana (formerly
Beauveria sulfurescens) showed opposite enantioselectivity, leading to the (R)-epoxide
in 98% e.e. The epoxide hydrolase from Aspergillus niger has proven to be a very
versatile catalyst over the past 15 years. Many preparative-scale applications have been
demonstrated and the enzyme is now commercially available via Fluka (CH). The
drawback that the wild-type enzyme is only enantioselective towards aromatic
epoxides has been overcome by directed evolution (Section 9.3.6) [245]. See also
Chapters 4 and 5 of this book for a more in-depth discussion on the principles and
application of mutagenesis.
O O O
Resting cells of Resting cells of
B. sulfurescens A. niger
(R)-styrene oxide Racemic (S)-styrene oxide

98% e.e. styrene oxide >99% e.e.
34% yield 28% yield
Scheme 9.24 Initial studies demonstrating access to both enantiomers of styrene oxide using
resting cells containing epoxide hydrolases with opposite enantiopreferences [177].
9.3.5
Reaction Engineering
Common issues when scaling up enzyme-catalyzed reactions are substrate and

product inhibition, enzyme inactivation, and low substrate solubility. Since epoxides
are very reactive molecules, there is an inherent risk of aspecific protein modification
resulting in loss of activity and a certain degree of background hydrolysis competing
with the enzyme-catalyzed reaction. Epoxide hydrolases are prone to denaturation on
solvent–solvent and solvent–air interfaces. This presents problems because most
epoxides are not very soluble in water and often form suspensions or emulsions.
Therefore, unsurprisingly, most synthetic procedures require a certain degree of
reaction engineering.
The general enzyme stability issues can partly be addressed through immobi-
lization of the enzyme on supports such as Eupergit [212], DEAE-cellulose [213–
215], or silica gel [216]. Some good results have been obtained by imprinting the
enzyme with substrate during immobilization [217]. Alternative approaches are
encapsulation within a gelatin gel (geloenzymes) [218] and formation of crosslinked
enzyme crystals (CLECs) [219].
Some more specific problems for epoxide hydrolase catalyzed reactions are the
aforementioned high chemical reactivity of the epoxide and the difficult extractability
of the polar diol reaction product from the aqueous media. Problems such as low
solubility and chemical hydrolysis of the epoxide, substrate and product inhibition,
and enzyme inactivation can be overcome by engineering the reaction medium. In
the simplest reaction setup the epoxide is added as a second phase to the medium. If
the reaction is a kinetic resolution it is important that the rate of dissolving in the
water phase is faster than the enzymatic reaction, thus avoiding conversion of the
desired epoxide enantiomer which would occur when the fast-reacting enantiomer
gets depleted. This is a common problem with solid substrates with a low solubility
(<1 mM). Liquid substrates can form an emulsion and, thereby, increase the rate of
solubilization but also the contact area with the enzyme. The latter generally causes
an increased enzyme deactivation. An example of a preparative-scale reaction with
the epoxide as a second phase is the kinetic resolution of para-chlorostyrene oxide (0

C, 306 g l1) by the enzyme from Aspergillus niger [220]. The use of water-miscible
additives can increase the substrate solubility and improve enzyme stability. Exam-
ples are the use of detergents to stabilize membrane-associated enzymes [221], a 20-
fold increase in activity using 20% of a deep eutectic solvent (DES) [222], and a
sevenfold increase in substrate solubility using cyclodextrins [223].
A second non-miscible organic phase can act as substrate reservoir, thereby
reducing the contact of the enzyme with the insoluble substrate and minimizing
the non-enzymatic hydrolysis. The selected solvent should allow a fast partitioning of
the epoxide between the two phases but minimize a deleterious effect on the enzyme.
Such solvent compatibility with enzymes can in general be described using the log P
value, which for Agrobacterium radiobacter AD1 should be above 4 [224]. The kinetic
resolution of styrene oxide in two-phase systems, catalyzed by the epoxide hydrolase
from Agrobacterium radiobacter AD1 (EchA), was studied in detail. These studies
demonstrated various bottlenecks such as molecular and interfacial inactivation,
mass transfer limitations [225], and inhibition and inactivation by the formed
diols [226]. A recent study showed an epoxide enantiomer-specific inactivation
process. Once in about every 119 catalytic cycles, the nucleophilic active site aspartic
acid is converted into iso-aspartate [227]. As a result the most optimal reaction setup
with EchA is shown in Table 9.4 and, in general, one should expect to optimize at least
these three parameters.
Table 9.4 Reaction engineering of Agrobacterium radiobacter EchA reactions.
Parameter Reason
Enzyme variant: Y215F mutant of EchA Improved enantioselectivity and activity

Minimal inactivation by diol
Reactor type: slowly stirred batch reactor Interfacial inactivation minimized
Solvent: iso-octane as second organic phase Appropriate log P and mass transfer
400
The use of iso-octane seems to give a general benefit since wild-type epoxide
hydrolase from Aspergillus niger also efficiently catalyzed the hydrolysis of a series of
trifluoromethyl-substituted styrene oxides in the presence of iso-octane as second
phase [199]. The enantioselective hydrolysis of such epoxides present at 1.8 M
resulted in complete kinetic resolutions within a few hours reaction time [228].
Alternative second phases are ionic liquids. Epoxide hydrolases from cress and
mouse were able to catalyze the hydrolysis of racemic and meso epoxides in several
ionic liquids in the presence of only around 10% water [229].
Contact of the biocatalyst with the organic phase can be minimized through the use
of a membrane reactor. The kinetic resolution of 1,2-epoxyhexane was achieved using
such a membrane bioreactor, either in a batch-wise mode or as a continuous cascade
where, after separation of the optically pure epoxide, the inhibitory diol is removed in
a separate membrane unit [230, 231].
9.3.6
Improving Epoxide Hydrolases by Mutagenesis and Evolution
After determining the crystal structure of the enzyme from Agrobacterium radio-
bacter AD1 (EchA) the influence of several active site residues on the activity and
enantioselectivity were investigated. Two tyrosine residues (Y152 and Y215) present
in the so-called cap domain play an important role in catalysis (Section 9.3.3) [232].
Mutation of either tyrosine residue to phenylalanine resulted in mutants with a two-
to fourfold increase in enantioselectivity with (substituted) styrene oxides [233]. At
least one of the two tyrosine residues is necessary since simultaneous mutation of
both tyrosines abolished all enzymatic activity.
With EchA-Y215F the specific activity of the slow reacting enantiomer decreased a
factor of 150, thereby obviating close monitoring of the reaction around the 50%
conversion mark [234]. A similar increase in enantioselectivity was observed in the
kinetic resolution of 2-, 3-, and 4-pyridyl oxirane catalyzed by the EchA-215F
mutant [235]. However, this increased enantioselectivity is not a general phenom-
enon and seems limited to aromatic substrates because hydrolysis of aliphatic
epoxides such as epichlorohydrin and 1,2-epoxyhexane catalyzed by EchA-Y215F
still occurred with low enantioselectivity (E < 2). Based on the above-described role of
the two tyrosine residues, the corresponding conserved tyrosine residues present in
the homologous epoxide hydrolase from Aspergillus niger M200 were mutated to
histidine and phenylalanine. However, this yielded catalytically inactive enzymes and
it seems that here both tyrosines are a requirement [236].
Mutation of residue F108 to alanine in EchA resulted in large enhancements in
enantioselectivity and activity for various substrates. Mutant F108A showed a 150-
fold increase in activity of the hydrolysis of cyclohexene oxide to (R,R)-1,2-cyclohex-
anediol. The practical applicability was demonstrated by a reaction on a 13.5-gram
scale. EchA-F108C showed a 17-fold increase in E-value for the hydrolysis of cis-2,3-
epoxybutane, while with F108T the enantioselectivity towards para-nitrophenyl
glycidyl ether (pNPGE) increased by a factor of seven. The enantioselectivity of
EchA was further improved using error-prone PCR and DNA shuffling. Mutants with
increased enantioselectivity towards aliphatic epoxides were identified using the
chromogenic pNPGE as an epoxyalkane mimic [237]. A selection of mutants with the
best results towards the hydrolysis of pNPGE indeed showed a most prominent
increase in enantioselectivity towards aliphatic substrates. The enantioselectivity of
an EchA variant with four mutations towards epichlorohydrin increased from <2 to
40. Remarkably, in this mutagenesis program, the eight selected best variants all
contained a mutation of either one of the two tyrosine residues.
DNA shuffling and saturation mutagenesis of positions F108, L190, I219, D235,
and C248 were used to generate variants of the epoxide hydrolase of EchA with
enhanced enantioselectivity and activity for styrene oxide and enhanced activity for
1,2-epoxyhexane and epoxypropane [238]. EchA-I219F has more than fivefold
enhanced enantioselectivity towards racemic styrene oxide and the E value for the
production of (R)-1-phenylethane-1,2-diol increased from 17 for the wild-type
enzyme to 91, as well as showing a twofold improved activity for the production
of the (R)-diol. EchA was also tuned to accept cis-1,2-dichloroepoxyethane as a
substrate by accumulating beneficial mutations from three rounds of saturation
mutagenesis at three selected active site residues [239].
The enantioselectivity of the epoxide hydrolase from Aspergillus niger has been
improved using error-prone PCR and iterative combinatorial active site saturation
testing (CASTing) using the model substrate phenyl glycidyl ether (PGE). The results
of these studies and the principles of CASTing are discussed in more detail in
Chapter 5. Prior to these studies, a more stable (and suitable for directed evolution)
expression system was developed [240]. In the initial study using error-prone PCR,
the best variant containing three mutations (A217V, K332E, and A390E) gave an
increase in enantioselectivity from 4.6 to 10.8 [241]. The enzyme was further evolved
using CASTing. On the basis of the X-ray structure, several sites (consisting of one or
more amino acid positions) were identified and subjected to saturation mutagenesis.
The gene of the best variant obtained from each focused library was used as template
for the next. The best mutant (LW202) harboring a total of nine mutations was highly
enantioselective (E ¼ 115) towards PGE [242]. In a follow-up study, the origin of the
enantioselectivity of mutant LW202 and the intermediate stages of the evolutionary
pathway were studied in detail [243]. A study of the substrate range showed that,
similar to the results with EchA using pNPGE, the enantioselectivity was increased
over a range of substrates. Recently, the epoxide hydrolase of a marine fish, Mugil
cephalus, was engineered to improve the catalytic activity based on comparative
homology modeling using the crystal structure of the homologous EH from
Aspergillus niger as a template. By introducing triple point mutations, the initial
hydrolysis rate of racemic styrene oxide was enhanced from 0.01 U mg1 cells for the
wild type to 0.35 U mg1 of cells [244].
9.3.6.1 Epoxide Hydrolase Assays

Every enzyme study starts with finding active enzymes and various assays have been
developed for measuring epoxide hydrolase activity: chromatographic analysis, mass
402
spectrometric analysis, derivatization of the remaining epoxide or formed diol to

produce a color, measuring the difference of extinction coefficient between substrate
and product, release of a chromophore upon hydrolysis, or using a fluorescent probe.
For every specific application a different assay is most appropriate. Chromogenic
substrates allowing online measurement of conversion have shown to be suitable for
mechanistic and kinetic studies, while endpoint assays are more suitable for high-
throughput screening of epoxide hydrolase libraries.
Enzyme selection can be performed in a three-tiered approach that starts with a
pre-selection of variants to eliminate non-active clones in epoxide hydrolase libraries.
This has been carried out using agar plate assays. Oxidation of the formed diol due to
epoxide hydrolysis caused an increased uptake of safranin O within the cell, resulting
in the formation of a pink color in the colonies on the agar plate [237]. In another
approach, bacterial growth on agar plates was shown to be directly related to the
presence of active epoxide hydrolases that catalyze the hydrolysis of the epoxide
substrate in the medium that is purposefully made toxic for growth [245].
Secondly, the activity of enzymes must be quantified. Epoxide hydrolase activity
can be determined by allowing the remaining epoxide to react with 4-nitrobenzylpyr-
idine to furnish a blue dye [246]. The assay was adapted for use in 96-well microtiter
plates in the screening of mutants of the epoxide hydrolase from Agrobacterium
radiobacter AD1 [247]. The assay was also optimized for screening using filter
paper [248]. The formation of low concentrations of diol can be measured more
precisely than a small decrease in substrate concentration due to a low conversion of
the epoxide. Several groups developed epoxide hydrolase activity assays based on the
periodate cleavage of diols yielding aldehydes and ketones. One version of the assay is
based on the quantitative colorimetric detection of ketones and aldehydes with
Schiffs reagent containing fuchsin and sulfurous acid [249]. In other versions of the
assay, the remaining periodate is allowed to react with adrenaline (Scheme 9.25) or
carboxyfluorescein, resulting in a red dye or a decrease in fluorescence, respective-
ly [250, 251]. In a third adaption, the formed aldehyde product is eliminated (catalyzed
by albumin), resulting in the release of the fluorescent compound umbellifer-
one [252]. By measuring both enantiomers separately, with or without a competing
substrate, the enantioselectivity can be estimated [253].
Online assays measuring the progress curve beyond 50% conversion allow for the
determination of the enantioselectivity and kinetic constants such as initial activity
in one run. Chromogenic substrates having a difference in extinction coefficient
between the epoxide and diol, such as para-nitrostyrene oxide and para-nitrophenyl
glycidyl ether, have been used for screening and characterization of epoxide
hydrolase libraries [141, 233, 237, 254]. Substrates containing an ester moiety and
releasing a chromophore such as para-nitrophenol give highly sensitive assays, but
suffer from background reaction by esterases and lipases that are often present
when using whole cells or crude enzyme preparations [255]. A high-throughput
assay for determining the % e.e. has been developed by the group of Reetz [256].
Two pseudo-enantiomers, (S)-phenyl glycidyl ether and the isotopically labeled (R)-
d5-phenyl glycidyl ether, have been used in a kinetic resolution. The outcome of the
reaction was determined using high-throughput multichannel mass spectrometry.
Epoxide OH
O hydrolase O
R OH
R R
NaIO4 NaIO3
+ water
MeHN
HO
OH O
HO
OH N O
Adrenaline Adrenochrome
Red dye
Scheme 9.25 Adrenaline test, used for determining epoxide hydrolase activity.
Integration of peaks of the two distinct masses, M and M þ 5, allowed for

determination of enantiomeric excess.
9.3.7
Towards 100% Yield
In a perfect kinetic resolution the remaining epoxide and the diol are obtained in a
high optical purity. However, in most cases the enantioselectivity and regioselectivity
are not absolute, some chemical hydrolysis occurs, and the reaction is allowed to
proceed past 50%. In these cases, the produced diol will only have a moderate e.e.
With some diols, the e.e. can be upgraded via recrystallization. The optically pure diol
can be converted back into the epoxide, giving access to both enantiomers of the
epoxide [257]. Hydrolysis of an optically pure terminal epoxide can, depending on
the regioselectivity of ring opening, yield either enantiomer of the diol. This enables
the development of enantioconvergent processes. For such a process the ring
opening of one enantiomer of the epoxide should occur at the a carbon atom, while
ring opening of the other enantiomer occurs at the b carbon atom. An optimal
enantioconvergent process will yield the diol optically pure in 100% yield. Three
concepts of such one-pot reactions are shown in Table 9.2, using one enzyme (D), two
enzymes (E), or a chemoenzymatic process (F).
9.3.7.1 Enantioconvergent Reactions Catalyzed by a Single Enzyme

An important requirement in many processes is to obtain the target compound in
high optical purity. If the optical purity of the diol cannot easily be improved by
crystallization, the process requirements will have to be very strict in order to obtain
404
Table 9.5 Enantioconvergent reactions of substituted styrene oxides catalyzed by recombinant

epoxide hydrolases.
Entry Substituent Regioselectivity E-value Chemical E.e. (%) Source of

per enantiomer hydrolysis epoxide
fast (%)/slow (%) (% h1) hydrolase
1 m-Chloro 97/94 6 0.5 91 Solanum tuberosum

2 p-Chloro 97/92 70 1.2 74 Solanum tuberosum
3 p-Chloro 99/94 30 Not 96 Caulobacter
determined. crescentus [259]
the diol with the desired specifications. The regioselectivity of the enzymatic reaction
must very high and no (non-regioselective) chemical hydrolysis should occur. A high
enantioselectivity will result in long reaction times because the reaction rate slows
down in the final stage of the reaction, especially for enzymes having high Km values.
This later stage is also the typical moment where any chemical hydrolysis will start to
compete with the enzymatic reaction. Various examples have been published where
enantioconvergent reactions were catalyzed by a single enzyme or whole cell
biocatalyst. With the latter, it should be taken into account that possibly more than
one enzyme is accountable for the observed reaction. Table 9.5 shows some examples
of enantioconvergent processes catalyzed by a single (recombinant) enzyme. The
recombinant epoxide hydrolase from Solanum tuberosum (potato) hydrolyzed meta-
and para-chlorostyrene oxide with close-to equal regioselectivities [258]. However,
with the para substituted epoxide, the combination of a higher enantioselectivity
(longer reaction times) and faster chemical hydrolysis (1.2% h1 versus 0.5% h1)
resulted in the diol having lower optical purity.
An enantioconvergent synthesis catalyzed by an epoxide hydrolase from Nocar-
dia EH1 starting from cis-2,3-epoxyheptene yielded the product in 91% e.e. and
79% isolated yield. Resolution occurred with an E-value of 10 while the regios-
electivities were 99.7% (first enantiomer) and 77% (slow enantiomer) [260].
Almost perfect reaction characteristics were observed in the hydrolysis of a
trisubstituted epoxide. Whole cells of Streptomyces lavendulae ATCC 55209 showed
opposite regioselectivities of 99% and >99%, respectively. Although the E-value is
only 1.3, the long reaction times (>100 h) would hamper a practical applica-
tion [261]. Similar enantioconvergent reactions were observed in the hydrolysis of
cis-b-methylstyrene oxide by the fungus Beauveria bassiana (85% yield, 98% e.e.)
and the conversion of limonene oxide diastereomers by limonene epoxide
hydrolase [172, 262].
9.3.7.2 Enantioconvergent Reactions by Employing Two Enzymes

The first examples of enantioconvergent reactions with two different bacterial
epoxide hydrolases with opposite enantioselectivity and regioselectivity were
described by Furstoss et al. (Scheme 9.26) [263, 265]. Similar approaches have since
O
A.niger
HO OH
(R) 100% conversion
O 92% yield
89% e.e.
B. sulfurescens (R)
(S)
Scheme 9.26 One-pot enantioconvergent hydrolysis of styrene oxide deploying two enzymes with
opposite enantioselectivity and regioselectivity.
been published using the same substrate but with recombinant enzymes, ensuring
shorter reaction times and thereby less chemical background reaction. The combi-
nation of enzymes from Aspergillus niger NK and Rhodotorula glutinis yielded (R)-
phenylethane diol in 95% yield and 90% e.e. while the combination of the epoxide
hydrolase from Solanum tuberosum and a mutant of Agrobacterium radiobacter AD1
(I219F) resulted in >99% yield and 98% e.e. [265, 266]. Yet another report deploys the
recombinant enzymes from Caulobacter crescentus, and the striped mullet (Mugil
cephalus, a fish), giving (R)-phenylethane diol in 90% e.e. and 94% yield [267].
Remarkably, the two enzymes from the one-enzyme enantioconvergent reactions
described above (EH from Solanum tuberosum and Caulobacter crescentus) were also
used in combination in a two-enzyme approach. By using the second enzyme,
product inhibition and the long reaction times due to the moderate enantioselectivity
causing low activity towards the (R)-enantiomer could be overcome. Mung bean
(Phaseolus radiatus) harbors two enantioconvergent epoxide hydrolases with opposite
enantioselectivity. By using a crude enzyme powder of mung bean (containing both
EHs) para-nitrostyrene oxide was hydrolyzed to the corresponding diol in 82% e.e.
and 84% yield [268].
9.3.7.3 Enantioconvergent Chemoenzymatic Reactions

The chemoenzymatic variant of the enantioconvergent reaction is mostly per-
formed in two sequential steps. After the kinetic resolution catalyzed by an
epoxide hydrolase, the remaining epoxide is chemically converted into the desired
diol. The hydrolysis of para-nitrostyrene oxide by Aspergillus niger yielded the
corresponding diol in 66% e.e. The remaining close-to optically pure epoxide was
converted into the same enantiomer of the diol via an acid-catalyzed hydrolysis,
with attack mainly at the a position, yielding the diol in 90% yield and 83% e.e.
Similar approaches were performed in the chemoenzymatic hydrolysis of 1-
methyl-1,2-epoxycyclohexane employing Corynebacterium C12 cells and dilute
HClO4 and of 2,2-disubstituted epoxides using Nocardia EH1 cells and concen-
trated sulfuric acid [269, 270].
406
9.3.7.4 Conversion of Meso-Epoxides

The regioselective ring opening of meso-epoxides yields the diol in either the (R,R) or
(S,S) configuration in up to 100% yield (Table 9.2, entry G). Owing to the inversion of
configurationof oneof theepoxidecarbon atoms, thediolis alwaysformed in >99%d.e.
However, the enantiomeric excess is determined by the regioselectivity of ring opening.
A range of cycloalkane oxides is accepted by mammalian epoxide hydrolases. The
selective ring opening of cyclohexene oxide catalyzed by a mammalian epoxide
hydrolase yielded (R,R)-cyclohexane-1,2-diol in 76% e.e., while that of rabbit origin
gave the diol in up to 94% e.e. [271, 272]. The epoxide hydrolase from Rhodotorula
glutinis catalyzed the selective hydrolysis of cyclohexene and cyclopentene oxide,
yielding the (R,R)-diol in 90% e.e. and 98% e.e. respectively [175]. The hydrolysis of
the meso N-benzyloxycarbonyl-3,4-epoxy-pyrrolidine and cyclohexene oxide catalyzed
by an enzyme from Sphingomonas HXN-200, yielding the diols in 95% and 86% e.e.
respectively, were the first reported examples using bacterial enzymes [273]. The
most extensive study towards hydrolyzing meso epoxides has been performed using a
range of novel microbial epoxide hydrolases, selected by screening environmental
DNA libraries [145]. Both the (R,R) and (S,S) diols were accessible. Besides aliphatic
diols such as (R,R)- (96% e.e.) and (S,S)-cyclohexane-1,2-diol (56% e.e.), substituted
stilbene oxides and dipyridyl analogues were also hydrolyzed with high regioselec-
tivity (Scheme 9.27).
O Epoxide N
hydrolase HO
BD8877
N N OH
31 U/mg N
meso epoxide (R,R)-diol

>99% e.e., >99% d.e.
Scheme 9.27 Epoxide hydrolase catalyzed hydrolysis of a meso epoxide yielding the optically and
diastereomerically pure diol [145].
The epoxide hydrolase from Agrobacterium radiobacter AD1 hydrolyses cyclohex-

ene oxide with high selectivity to give the diol in >99% e.e. [274]. However, the low
initial activity (and non-enzymatic conversion) prevents a practical application. By
introducing the mutation F108A, the activity was increased by a factor 150, while
retaining the high specificity. In a preparative-scale reaction, using 13.5 gram epoxide
(138 mM) and 60 mg l1 F108A mutant enzyme, the (R,R)-cyclohexane-1,2-diol
could be obtained in 98% e.e. and 61% isolated yield.
9.3.8
Outlook on Epoxide Hydrolases
At the beginning of this century, in the second edition of this book, Professor Faber
painted a rather bright outlook for epoxide hydrolases. There was a lot of optimism
based on crystal structures of epoxide hydrolases becoming available and advances in
directed evolution technology resulting in significantly improved enzymes. Now,
almost ten years later, in this chapter we have shown that the prediction has come true
with respect to having access to multiple wild-type and improved enzymes that allow
effective epoxide hydrolysis in laboratory-scale practical applications.
However, there is still an ill-defined barrier to application at large scale. In our
opinion this is not due to lack of knowledge collected in thousands of papers or lack of
good enzymes and methods of improving processes using them. It seems more likely
that epoxides are not considered to be viable intermediates in cost-effective synthetic
routes and, consequently, alternatives such as the reduction of prochiral haloketones
using alcohol dehydrogenases and resolutions of various precursors have become
industry standards.
An intrinsic drawback of epoxides is the chemical instability resulting in hydrolysis
of the epoxide. One approach to improve the ratio of enzymatic conversion over
chemical conversion is to add more enzyme to the reaction mixture. It is generally
accepted that for an industrial application the substrate loading should be 10% or
higher. This high concentration combined with the typically low aqueous solubility of
the epoxide will result in mass transfer limitations at high enzyme loading. Thus, the
solubilization of solid epoxide or phase transfer in biphasic mixtures becomes rate
limiting and lowers the enantioselectivity of the overall reaction. Therefore, in
addition to enzymes with a high enantioselectivity, which is a major goal in most
mutagenesis studies described in this book, enzymes with increased specific activity,
decreased Km for the preferred enantiomer, and better solvent tolerance are also
desired.
The regioselectivity of most epoxide hydrolases is not optimal and the optical purity
of the resulting diols is often of too low. In addition, diols are more difficult to extract
from the reaction broth due to their, in general, higher water solubility. Therefore,
only the resolution of the racemic epoxides with isolation of the remaining optically
pure epoxide appears to be a viable strategy, but this means that the overall yield will
be less than 50%.
One of the most remarkable reactions catalyzed by a single epoxide hydrolase is the
enantioconvergent reaction leading to an optically pure diol in 100% yield starting
from a racemic epoxide. This unique feature of epoxides is worth exploiting because
in industry there is a strong push towards designing processes that yield 100%
conversion to only one enantiomer or diastereomer. However, besides the problem of
chemical hydrolysis leading to a decrease in optical purity of diol, the occurrence of
such a reaction is rather rare and most concepts use two different enzymes dosed in
just the right amounts to effect the enantioconvergent reaction. Clearly, it would be
preferable from an economic and process technological point of view to have only one
catalyst present but it will be a challenge to optimize an epoxide hydrolase in such a
way that it has opposite regioselectivity for each substrate enantiomer, while retaining
a high activity for both enantiomers.
Epoxide hydrolases only accept water as nucleophile whereas halohydrin dehalo-
genases have been shown to accept ten different anions but are not capable of
hydrolysis. In this sense, these two enzyme classes are synthetically complementary.
408
Because of the shape of the halide binding site (and possibly other factors),
halohydrin dehalogenases are not active with bulky nucleophiles, or nucleophiles
that do not carry a negative charge. From a synthetic perspective, a holy grail is to
identify (or make) an enzyme that can accept other nucleophiles besides water, such
as alcohols or amines. Apart from one short report using cell-free extract containing
epoxide hydrolase and azide as nucleophile, no evidence for epoxide hydrolase
catalyzed ring opening with other nucleophiles has been found [74]. This makes
sense because in the most common form of epoxide hydrolases, the a/b-hydrolase
fold enzymes, the formation and hydrolysis of the alkyl-enzyme intermediate limits
the use of non-water nucleophiles (Figure 9.2). Since the hydrolysis of the alkyl-
enzyme intermediate results in the re-formation of the catalytic aspartate, reaction
with another nucleophile would afford the diol and a chemically modified aspartate.
This modification immediately inactivates the enzyme. A more likely approach
would be the study and directed evolution of the class of epoxide hydrolases that do
not follow this two-step mechanism but relies on non-covalent catalysis such as the
epoxide hydrolase obtained from Rhodococcus erythropolis (limonene epoxide
hydrolase) [151].
References
1 Breuer, M., Ditrich, K., Habicher, T., 11 Weijers, C.A.G.M., de Haan, A., and de
Hauer, B., Keßeler, M., St€ urmer, R., and Bont, J.A.M. (1988) Appl. Microbiol.
Zelinski, T. (2004) Angew. Chem. Int. Ed., Biotechnol., 27, 337–340.
43, 788–824. 12 Kubo, T., Peters, M.W., Meinhold, P., and
2 Kumar, P., Naidu, V., and Gupta, P. Arnold, F.H. (2006) Chem. Eur. J., 12,
(2007) Tetrahedron, 63, 2745–2785. 1216–1220.
3 Kumar, P. and Gupta, P. (2009) Synlett, 13 Dembitsky, V.M. (2003) Tetrahedron, 59,
1367–1382. 4701–4720.
4 Ketterer, B. (2001) Chem. Biol. Interact., 14 Ozaki, S.I. and Ortiz De Montellano, P.R.
138, 27–42. (1995) J. Am. Chem. Soc., 117, 7056–7064.
5 de Vries, E.J. and Janssen, D.B. (2003) 15 Champreda, V., Choi, Y.J., Zhou, N.Y.,
Curr. Opin. Biotechnol., 14, 414–420. and Leak, D.J. (2006) Appl. Microbiol.
6 Ivarsson, Y., Norrgård, M.A., Hellman, Biotechnol., 71, 840–847.
U., and Mannervik, B. (2007) BBA-Gen. 16 Gursky, L.J., Nikodinovic-Runic, J.,
Subjects, 1770, 1374–1381. Feenstra, K.A., and OConnor, K.E. (2010)
7 Balogh, L.M., Le Trong, I., Kripps, K.A., Appl. Microbiol. Biotechnol., 85, 995–1004.
Tars, K., Stenkamp, R.E., Mannervik, B., 17 Zhang, Y., Xin, J., Chen, L., and Xia, C.
and Atkins, W.M. (2009) Biochemistry, 48, (2009) Appl. Biochem. Biotechnol., 157,
7698–7704. 431–441.
8 Hegazy, U.M., Tars, K., Hellman, U., and 18 Rial, D.V., Bianchi, D.A., Kapitanova, P.,
Mannervik, B. (2008) J. Mol. Biol., 376, Lengar, A., van Beilen, J.B., and
811–826. Mihovilovic, M.D. (2008) Eur. J. Org.
9 Krishnakumar, A.M., Sliwa, D., Endrizzi, Chem., 1203–1213.
J.A., Boyd, E.S., Ensign, S.A., and Peters, 19 Zhu, D., Hyatt, B.A., and Hua, L.J. (2009)
J.W. (2008) Microbiol. Mol. Biol. Rev., 72, J. Mol. Catal. B: Enzym., 56, 272–276.
445–456. 20 Pregnolato, M., Terreni, M., De Fuentes,
10 Boyd, J.M. and Ensign, S.A. (2005) I.E., Alcantara Leon, A.R., Sabuquillo, P.,
Biochemistry, 44, 13151–13162. Fernandez-Lafuente, R., and Guisan,
References j409
J.M. (2001) J. Mol. Catal. B: Enzym., 11, 37 van den Wijngaard, A.J.,
757–763. Reuvekamp, P.T.W., and Janssen, D.B.
21 Li, C., Wang, P., Zhao, D., Cheng, Y., (1991) J. Bacteriol., 173, 124–129.
Wang, L., Wang, L., and Wang, Z. (2008) 38 Velısek, J., Dolezal, M., Crews, C., and
J. Mol. Catal. B: Enzym., 55, 152–156. Dvorak, T. (2002) Czech. J. Food Sci., 20,
22 Yu, D., Wang, L., Gu, Q., Chen, P., Li, Y., 161–170.
Wang, Z., and Cao, S. (2007) Process 39 Bel-Rhlid, R., Talmon, J.P., Fay, L.B., and
Biochem., 42, 1319–1325. Juillerat, M.A. (2004) J. Agric. Food Chem.,
23 Bianchi, D., Cabri, W., Cesti, P., 52, 6165–6169.
Francalanci, F., and Ricci, M. (1988) J. 40 Hardman, D.J., Huxley, M., Bull, A.T.,
Org. Chem., 53, 104–107. Slater, J.H., and Bates, R. (1997) J. Chem.
24 Garcia, M.J., Rebolledo, F., and Gotor, V. Technol. Biotechnol., 70, 60–66.
(1993) Tetrahedron: Asymmetry, 4, 41 Fauzi, A.M., Hardman, D.J., and
2199–2210. Bull, A.T. (1996) Appl. Microbiol.
25 Tr€aff, A., Bogar, K., Warner, M., and Biotechnol., 46, 660–666.
B€ackvall, J.E. (2008) Org. Lett., 10, 42 Assis, H.M.S., Sallis, P.J., Bull, A.T., and
4807–4810. Hardman, D.J. (1998) Enz. Microb.
26 Domınguez De Marıa, P., Carboni- Technol., 22, 568–574.
Oerlemans, C., Tuin, B., Bargeman, G., 43 Riehle, R.J. (2005) in Polymer Biocatalysis
van der Meer, A., and van Gemert, R. and Biomaterials (eds H.N. Cheng and
(2005) J. Mol. Catal. B: Enzym., 37, 36–46. R.A. Gross), ACS Symposium Series, vol.
27 Heinze, B., Kourist, R., Fransson, L., 900, American Chemical Society,
Hult, K., and Bornscheuer, U.T. (2007) Washington, D.C., pp. 302–316.
Protein Eng., Design Select., 20, 125–131. 44 van Agteren, M.H., Keuning, S., and
28 Kiljunen, E. and Kanerva, L.T. (1999) Janssen, D.B. (1998) Chapter 3 Halogenated
Tetrahedron: Asymmetry, 10, 3529–3535. Aliphatics, Handbook on Biodegradation and
29 Higgins, T.P., Hope, S.J., Effendi, A.J., Biological Treatment of Hazardous Organic
Dawson, S., and Dancer, B.N. (2005) Compounds, Kluwer Academic Publishers,
Biodegradation, 16, 485–492. ISBN: 0-7923-4989-X.
30 Effendi, A.J., Greenaway, S.D., and 45 van den Wijngaard, A.J., Janssen, D.B.,
Dancer, B.N. (2000) Appl. Environ. and Witholt, B. (1989) J. Gen. Microbiol.,
Microbiol., 66, 2882–2887. 135, 2199–2208.
31 de Jong, R.M., Tiesinga, J.J.W., 46 Nakamura, T., Yu, F., Mizunashi, W., and
Rozeboom, H.J., Kalk, K.H., Tang, L., Watanabe, I. (1991) Agric. Biol. Chem.,
Janssen, D.B., and Dijkstra, B.W. (2003) 1931–1933.
EMBO J., 22, 4933–4944. 47 Nagasawa, T., Nakamura, T., Yu, F.,
32 Castro, C.E. and Bartnicki, E.W. (1968) Watanabe, I., and Yamada, H. (1992)
Biochemistry, 7, 3213–3218. Appl. Microbiol. Biotechnol., 36, 478–482.
33 Bartnicki, E.W. and Castro, C.E. (1969) 48 Nakamura, T., Nagasawa, T., Yu, F.,
Biochemistry, 8, 4677–4680. Watanabe, I., and Yamada, H. (1992)
34 Geigert, J., Neidleman, S.L., Liu, T.E., J. Bacteriol., 174, 7613–7619.
DeWitt, S.K., Panschar, B.M., Dalietos, 49 Nakamura, T., Nagasawa, T., Yu, F.,
D.J., and Siegel, E.R. (1983) Appl. Environ. Watanabe, I., and Yamada, H. (1994)
Microbiol., 45, 1148–1149. Appl. Environ. Microbiol., 60, 1297–1301.
35 Geigert, J. and Neidleman, S.L. (1985) 50 Yu, F., Nakamura, T., Mizunashi, W., and
Enzymic synthesis of halohydrins and Watanabe, I. (1994) Biosci. Biotechnol.
their conversion to epoxides, in Enzymes Biochem., 58, 1451–1457.
and Immobilized Cells in Biotechnology, 51 Kasai, N., Tsujimura, K., Unoura, K., and
Biotechnology Series 5 (ed. A.I. Laskin), Suzuki, T. (1990) Agric. Biol. Chem., 54,
Butterworth–Heinemann, pp. 283–296. 3185–3190.
36 Janssen, D.B., Oppentocht, J.E., and 52 Kasai, N., Tsujimura, K., Unoura, K., and
Poelarends, G.J. (2001) Curr. Opin. Suzuki, T. (1992) J. Ind. Microbiol., 10,
Biotechnol., 12, 254–258. 37–43.
410
53 Kasai, N., Tsujimura, K., Unoura, K., and 69 Lutje Spelberg, J.H.,
Suzuki, T. (1992) J. Ind. Microbiol., 9, van Hylckema Vlieg, J.E.T., Tang, L.,
97–101. Janssen, D.B., and Kellogg, R.M. (2001)
54 Kasai, N. and Suzuki, T. (2003) Adv. Synth. Org. Lett., 3, 41–43.
Catal., 345, 437–455. 70 de Jong, R.M., Tiesinga, J.J.W., Villa, A.,
55 Suzuki, T. and Kasai, N. (1991) Bioorg. Tang, L., Janssen, D.B., and Dijkstra, B.W.
Med. Chem. Lett., 1, 343–346. (2005) J. Am. Chem. Soc., 127,
56 Suzuki, T., Kasai, N., Yamamoto, R., and 13338–13343.
Minamiura, N. (1994) Appl. Microbiol. 71 Hopmann, K.H. and Himo, F. (2008)
Biotechnol., 42, 270–279. Biochemistry, 47, 4973–4982.
57 Suzuki, T. and Kasai, N. (2003) Trends 72 Tron, G.C., Pirali, T., Billington, R.A.,
Glycosci. Glycotechnol., 15, 329–349. Canonico, P.L., Sorba, G., and
58 Assis, H.M.S., Bull, A.T., and Genazzani, A.A. (2007) Med. Res. Rev., 28,
Hardman, D.J. (1998) Enz. Microb. 278–308.
Technol., 22, 545–551. 73 Campbell-Verduy, L.S., Szymanski, W.,
59 Wandel, U., K€onigsberger, K., Postema, C.P., Dierckx, R.A.,
and Griengl, H. (1994) Biocatalysis, 10, Elsinga, P.H., Janssen, D.B., and
159. Feringa, B.L. (2010) Chem. Commun.,
60 van Hylckama Vlieg, J.E.T., Tang, L., 898–900.
Lutje Spelberg, J.H., Smilda, T., 74 Mischitz, M. and Faber, K. (1994)
Poelarends, G.J., Bosma, T., Tetrahedron Lett., 35, 81–84.
van Merode, A.E.J., Fraaije, M.W., and 75 Faber, K., Mischitz, M., and Kroutil, W.
Janssen, D.B. (2001) J. Bacteriol., 183, (1996) Acta Chem. Scand., 50, 249–258.
5058–5066. 76 Kamal, A., Damayanthi, Y., and Rao, M.V.
61 Lutje Spelberg, J.H., (1992) Tetrahedron: Asymmetry, 3,
van Hylckama Vlieg, J.E.T., Bosma, T., 1361–1364.
Kellogg, R.M., and Janssen, D.B. (1999) 77 Hasnaoui, G., Lutje Spelberg, J.H., de
Tetrahedron: Asymmetry, 10, 2863–2870. Vries, E.J., Tang, L., Hauer, B., and
62 Lutje Spelberg, J.H., Tang, L., Janssen, D.B. (2005) Tetrahedron:
van Gelder, M., Kellogg, R.M., and Asymmetry, 16, 1685–1692.
Janssen, D.B. (2002) Tetrahedron: 78 Fuchs, M., Simeo, Y., Ueberbacher, B.T.,
Asymmetry, 13, 1083–1089. Mautner, B., Netscher, T., and Faber, K.
63 Haak, R.M., Tarabiono, C., Janssen, D.B., (2009) Eur. J. Org. Chem., 833–840.
Minnaard, A.J., de Vries, J.G., and 79 Majeric-Elenkov, M., Hauer, B., and
Feringa, B.L. (2007) Org. Biomol. Chem., Janssen, D.B. (2006) Adv. Synth. Catal.,
5, 318–323. 348, 579–585.
64 Schneider, C. (2006) Synthesis, 80 Majeric-Elenkov, M., Tang, L.,
3919–3944. Meetsma, A., Hauer, B., and Janssen,
65 Nakamura, T., Nagasawa, T., Yu, F., D.B. (2008) Org. Lett., 10, 2417–2420.
Watanabe, I., and Yamada, H. (1991) 81 Majeric-Elenkov, M., Hoeffken, H.W.,
Biochem. Biophys. Res. Commun., 180, Tang, L., Hauer, B., and Janssen, D.B.
124–130. (2007) Adv. Synth. Catal., 349,
66 Nakamura, T., Nagasawa, T., Yu, F., 2279–2285.
Watanabe, I., and Yamada, H. (1994) 82 de Jong, R.M., Kalk, K.H., Tang, L.,
Tetrahedron, 50, 11821–11826. Janssen, D.B., and Dijkstra, B.W. (2006)
67 Lutje Spelberg, J.H., Tang, L., J. Bacteriol., 188, 4051–4056.
van Gelder, M., Kellogg, R.M., and 83 Tang, L., van Hylckama Vlieg, J.E.T., Lutje
Janssen, D.B. (2002) Tetrahedron: Spelberg, J.H., Fraaije, M.W., and
Asymmetry, 13, 1083–1089. Janssen, D.B. (2002) Enz. Microb.
68 Hasnaoui-Dijoux, G., Technol., 30, 251–258.
Majeric-Elenkov, M., Lutje Spelberg, J.H., 84 Tang, L., Lutje Spelberg, J.H.,
Hauer, B., and Janssen, D.B. (2008) Fraaije, M.W., and Janssen, D.B. (2003)
ChemBioChem, 9, 1048–1051. Biochemistry, 42, 5378–5386.
References j411
85 Tang, L., van Merode, A.E.J., haloalcohols and epoxides. Available for
Lutje Spelberg, J.H., Fraaije, M.W., and download at http://
Janssen, D.B. (2003) Biochemistry, 42, dissertations.ub.rug.nl/FILES/faculties/
14057–14065. science/2008/r.m.haak/thesis.pdf
86 Tang, L., Torres Pazmi~ no, D.E., (accessed 20 June 2011).
Fraaije, M.W., de Jong, R.M., 100 Rolla, F. (1982) J. Org. Chem., 47,
Dijkstra, B.W., and Janssen, D.B. (2005) 4327–4329.
Biochemistry, 44, 6609–6618. 101 Morisseau, C. and Hammock, B.D.
87 Fox, R.J., Davis, S.C., Mundorff, E.C., (2005) Ann. Rev. Pharmacol. Toxicol., 45,
Newman, L.M., Gavrilovic, V., Ma, S.K., 311–333.
Chung, L.M., Ching, C., Tam, S., 102 Morisseau, C. and Hammock, B.D.
Muley, S., Grate, J., Gruber, J., (2008) Pest Manag. Sci., 64, 594–609.
Whitman, J.C., Sheldon, R.A., and 103 Lu, A.Y.H. and Miwa, G.T. (1980) Ann.
Huisman, G.W. (2007) Nat. Biotechnol., Rev. Pharmacol. Toxicol., 20, 513–531.
25, 338–344. 104 Morisseau, C., Du, G., Newman, J.W.,
88 Majeric-Elenkov, M., Tang, L., Hauer, B., and Hammock, B.D. (1998) Arch.
and Janssen, D.B. (2006) Org. Lett., 8, Biochem. Biophys., 356, 214–228.
4227–4229. 105 Decker, M., Arand, M., and Cronin, A.
89 Nakamura, T., Yu, F., Mizunashi, W., and (2009) Arch. Toxicol., 83, 297–318.
Watanabe, I. (1993) Appl. Environ. 106 Sandberg, M., Hasset, C., Adman, E.T.,
Microbiol., 59, 227–230. Meijer, J., and Omiecinski, C.J. (2000)
90 Lutje Spelberg, J.H., Tang, L., Kellogg, J. Biol. Chem., 275, 28873.
R.M., and Janssen, D.B. (2004) 107 Newman, J.W., Morisseau, C., and
Tetrahedron: Asymmetry, 15, 1095–1102. Hammock, B.D. (2005) Prog. Lipid Res.,
91 Pamies, O. and B€ackvall, J.E. (2002) 44, 1–51.
J. Org. Chem., 67, 9006–9010. 108 Stark, A., Lundholm, A., and Meijer, J.
92 Neidleman, S.L., Amon, W.F., and (1995) Phytochemistry, 38, 31–33.
Geigert, J. (1981) in Preparation of 109 Abdel-Latief, M., Garbe, L.A., Koch, M.,
epoxides and glycols from gaseous and Ruther, J. (2008) Proc. Natl. Acad. Sci.
alkenes, US 4284723, to Cetus U.S.A., 105, 8914–8919.
Corporation. 110 Swaving, J. and de Bont, J.A.M. (1998)
93 Nagasawa, T. and Yamada, H. (1995) Pure Enzyme Microb. Technol., 22, 19–26.
Appl. Chem., 67, 1241–1256. 111 Niehaus, W.G., Kisic, A., Torkelson, A.,
94 Geigert, J., Neidleman, S.L., Liu, T.E., Bednarczyk, D.J., and Schroepfer, G.J.
DeWitt, S.K., Panschar, B.M., Dalietos, (1970) J. Biol. Chem., 245, 3802–3890.
D.J., and Siegel, E.R. (1983) Appl. Environ. 112 Michaels, B.C., Ruettinger, R.T., and
Microbiol., 45, 1148–1149. Fulco, A. (1980) Biochem. Biophys. Res.
95 Nakamura, T., Yu, F., Mizunashi, W., and Commun., 92, 1189–1195.
Watanabe, I. (1993) Appl. Environ. 113 Hartmann, G.R. and Frear, D.S. (1963)
Microbiol., 59, 227–230. Biochem. Biophys. Res. Commun., 10,
96 Seisser, B., Lavandera, I., Faber, K., 366–372.
Lutje Spelberg, J.H., and Kroutil, W. 114 Wackett, L.P. and Gibson, D.T. (1982)
(2007) Adv. Synth. Catal., 349, 1399–1404. Biochem. J., 205, 117–122.
97 Schrittwieser, J.H., Lavandera, I., 115 Oesch, F., Jerina, D.M., and Daly, J.W.
Seisser, B., Mautner, B., (1971) Arch. Biochem. Biophys., 144,
Lutje Spelberg, J.H., and Kroutil, W. 253–261.
(2009) Tetrahedron: Asymmetry, 20, 116 Smit, M.S. (2004) Trends Biotechnol., 22,
483–488. 123–129.
98 Schrittwieser, J.H., Lavandera, I., 117 Choi, W.J. and Choi, C.Y. (2005)
Seisser, B., Mautner, B., and Kroutil, W. Biotechnol. Bioprocess. Eng., 10, 167–179.
(2009) Eur. J. Org. Chem., 2293–2298. 118 Misawa, E., Chan Kwo Chion, C.K.,
99 Thesis Haak, R.M. (2008) Archer, I.V.J., Woodland, M.P., Zhou,
Chemo-enzymatic routes to enantiopure N.Y., Carter, S.F., Widdowson, D.A., and
412
Leak, D.J. (1998) Eur. J. Biochem., 253, 136 Weijers, C.A.G.M. and de Bont, J.A.M.
173–183. (1999) J. Mol. Catal. B: Enzym., 6,
119 Visser, H., Vreugdenhil, S., de Bont, 199–214.
J.A.M., and Verdoes, J.C. (2000) Appl. 137 Botes, A.L., Weijers, C.A.G.M., and van
Microbiol. Biotechnol., 53, 415–419. Dyk, M.S. (1998) Biotechnol. Lett., 20,
120 Barbirato, F., Verdoes, J.C., de Bont, 421–426.
J.A.M., and van der Werf, M.J. (1998) 138 Moussou, P., Archelas, A., and Furstoss, R.
FEBS Lett., 438, 293–296. (1998) J. Mol. Catal. B: Enzym., 5, 447–458.
121 Arand, M., Hemmer, H., Durk, H., 139 Osprian, I., Kroutil, W., Mischitz, M., and
Baratti, J.C., Archelas, A., Furstoss, R., Faber, K. (1997) Tetrahedron: Asymmetry,
and Oesch, F. (1999) Biochem. J., 344, 8, 65–71.
273–280. 140 Barth, S., Fischer, M., Schmid, R.D., and
122 Visser, H., Weijers, C.A.G.M., Pleiss, J. (2004) Bioinformatics, 20,
van Ooyen, A.J.J., and Verdoes, J.C. 2845–2847. Database URL: www.led.uni-
(2002) Biotechnol. Lett., 24, 1687–1694. stuttgart.de, last updated Dec 10, 2009,
123 Visser, H., Villela Filho, M.D.O., Liese, A., last accessed Aug 28, 2011.
Weijers, C.A.G.M., and Verdoes, J.C. 141 van Loo, B., Kingma, J., Arand, M.,
(2003) Biocatal. Biotransform., 21, 33–40. Wubbolts, M.G., and Janssen, D.B. (2006)
124 Kim, H.S., Lee, J.H., Park, S., and Lee, Appl. Environ. Microbiol., 72, 2905–2917.
E.Y. (2004) Biotechnol. Bioprocess Eng., 9, 142 Hwang, S., Hyun, H., Lee, B., Park, Y.,
62–64. Choi, C., Han, J., and Joo, H. (2006) J.
125 Labuschagne, M., Botes, A.L., and Microbiol. Biotechnol., 16, 32–36.
Albertyn, J. (2004) DNA Seq., 15, 202–205. 143 Li, N., Zhang, Y., and Feng, H. (2009) Acta
126 Labuschagne, M. and Albertyn, J. (2007) Biochim. Biophys. Sin., 41, 638–647.
Yeast, 24, 69–78. 144 Woo, J.H., Kang, J.H., Kang, S., Hwang,
127 Liu, Y., Wu, S., Wang, J., Yang, L., and Y.O., and Kim, S.J. (2009) Appl. Microbiol.
Sun, W. (2007) Protein Expr. Purif., 53, Biotechnol., 82, 873–881.
239–246. 145 Zhao, L., Han, B., Huang, Z., Miller, M.,
128 Johansson, P., Unge, T., Cronin, A., Huang, H., Malashock, D.S., Zhu, Z.,
Arand, M., Bergfors, T., Jones, T.A., and Milan, A., Robertson, D.E., Weiner, D.P.,
Mowbray, S.L. (2005) J. Mol. Biol., 351, and Burk, M.J. (2004) J. Am. Chem. Soc.,
1048–1056. 126, 11156–11157.
129 Biswal, B.K., Garen, G., Cherney, M.M., 146 epanek, V., Maresova, H.,
Kotik, M., St
Garen, C., and James, M.N.G. (2006) Acta Kyslık, P., and Archelas, A. (2009) J. Mol.
Crystallogr. Sect. F, 62, 136–138. Catal. B: Enzym., 56, 288–293.
130 Liu, Z., Li, Y., Xu, Y., Ping, L., and Zheng, 147 Beetham, J.K., Grant, D., Arand, M.,
Y. (2007) Appl. Microbiol. Biotechnol., 74, Garbarino, J., Kiyosue, T., Pinot, F.,
99–106. Oesch, F., Belknap, W.R., Shinozaki, K.,
131 Woo, J.H., Hwang, Y.O., Kang, S.G., and Hammock, B.D. (1995) DNA Cell
Lee, H.S., Cho, J.C., and Kim, S.J. (2007) Biol., 14, 61–71.
Appl. Microbiol. Biotechnol., 76, 365–375. 148 Nardini, M., Ridder, I.S., Rozeboom, H.J.,
132 Woo, J.H., Kang, J.H., Kang, S.G., Kalk, K.H., Rink, R., Janssen, D.B., and
Hwang, Y.O., and Kim, S.J. (2009) Appl. Dijkstra, B.W. (1999) J. Biol. Chem., 274,
Microbiol. Biotechnol., 82, 873–881. 14579–14586.
133 Jacobs, M.H., van den Wijngaard, A.J., 149 Janssen, D.B., Pries, F., van der Ploeg, J.,
Pentenga, M., and Janssen, D.B. (1991) Kazemier, B., Terpstra, P., and Witholt, B.
Eur. J. Biochem., 202, 1217–1222. (1989) J. Bacteriol., 171, 6791–6799.
134 van der Werf, M.J., Swarts, H.J., and de 150 Arand, M., Grant, D.F., Beetham, J.K.,
Bont, J.A.M. (1999) Appl. Environ. Friedberg, T., Oesch, F., and Hammock,
Microbiol., 65, 2092–2102. B.D. (1994) FEBS Lett., 338, 251–256.
135 van der Werf, M.J., Overkamp, K.M., and 151 van der Werf, M.J., Overkamp, K.M., and
de Bont, J.A.M. (1998) J. Bacteriol., 180, de Bont, J.A.M. (1998) J. Bacteriol., 180,
5052–5057. 5052–5057.
References j413
152 Johansson, P., Unge, T., Cronin, A., 171 Faber, K., Mischitz, M., and Kroutil, W.
Arand, M., Bergfors, T., Jones, T.A., and (1996) Acta Chim. Scand., 50, 249.
Mowbray, S.L. (2005) J. Mol. Biol., 351, 172 van der Werf, M.J., Orru, R.V.A.,
1048–1056. Overkamp, K.M., Swarts, H.J.,
153 Hopmann, K.H., Hallberg, B.M., and Osprian, I., Steinreiber, A.,
Himo, F. (2005) J. Am. Chem. Soc., 127, de Bont, J.A.M., and Faber, K. (1999) Appl.
14339–14347. Microbiol. Biotechnol., 52, 380–385.
154 Arand, M., Hallberg, B.M., Zou, J., 173 Lutje Spelberg, J.H., Rink, R.,
Bergfors, T., Oesch, F., van der Werf, M., Kellogg, R.M., and Janssen, D.B. (1998)
de Bont, J.A.M., Jones, T.A., and Tetrahedron: Asymmetry, 9, 459–466.
Mowbray, S.L. (2003) EMBO J., 22, 174 Rink, R. and Janssen, D.B. (1998)
2583–2592. Biochemistry, 37, 18119–18127.
155 Moussou, P., Archelas, A., Baratti, J.C., 175 Weijers, C.A.G.M. (1997) Tetrahedron:
and Furstoss, R. (1998) Tetrahedron: Asymmetry, 8, 639–647.
Asymmetry, 9, 1539–1547. 176 Botes, A.L., Weijers, C.A.G.M., and van
156 Faber, K. and Kroutil, W. (2002) Dyk, M.S. (1998) Biotechnol. Lett., 20,
157 Archelas, A. and Furstoss, R. (2001) Curr. 177 Pedragosa-Moreau, S., Archelas, A., and
Opin. Chem. Biol., 5, 112–119. Furstoss, R. (1993) J. Org. Chem., 58,
158 Archer, I.V.J. (1997) Tetrahedron, 53, 5533–5536.
15617–15662. 178 Weijers, C.A.G.M., Botes, A.K.,
159 Smit, M.S. and Labuschagne, M. (2006) van Dyk, M.S., and de Bont, J.A.M. (1998)
Curr. Org. Chem., 10, 1145–1161. Tetrahedron: Asymmetry, 9, 467–473.
160 Choi, W.J. (2009) Appl. Microbiol. 179 Botes, A.L., Weijers, C.A.G.M., Botes,
Biotechnol., 84, 239–247. P.J., and van Dyk, M.S. (1999)
161 Choi, W.J. and Choi, C.Y. (2005) Tetrahedron: Asymmetry, 10, 3327–3336.
Biotechnol. Bioprocess Eng., 10, 167–179. 180 Botes, A.L., Steenkamp, J.A.,
162 Williamson, K.C., Morisseau, C., Letloenyane, M.Z., and van Dyk, M.S.
Maxwell, J.E., and Hammock, B.D. (2000) (1998) Biotechnol. Lett., 20, 427–430.
Tetrahedron: Asymmetry, 11, 4451. 181 Moussou, P., Archelas, A., and Furstoss,
163 Belluci, G., Chiappe, C., Cordoni, A., and R. (1998) Tetrahedron, 54, 1563–1572.
Marioni, F. (1993) Tetrahedron: 182 Orru, R.V.A., Mayer, S.F., Kroutil, W., and
Asymmetry, 4, 1153. Faber, K. (1998) Tetrahedron, 54, 859–874.
164 Belluci, G., Chiappe, C., and Cordoni, A. 183 Pienaar, D.P., Mitra, R.K., van Deventer,
(1996) Tetrahedron: Asymmetry, 7, 197. T.I., and Botes, A.L. (2008) Tetrahedron
165 Belluci, G., Chiappe, C., Marioni, F., and Lett., 49, 6752–6755.
Benetti, M. (1991) J. Chem. Soc., Perkin 184 Choi, W.J., Huh, E.C., Park, H.J.,
Trans. 1, 361. Lee, E.Y., and Choi, C.Y. (1998) Biotechnol.
166 Belluci, G., Chiappe, C., Conti, L., Tech., 12, 225–228.
Marioni, F., and Pierini, G. (1989) J. Org. 185 Kim, H.S., Lee, J.H., Park, S., and
Chem., 54, 5978. Lee, E.Y. (2004) Biotechnol. Bioprocess Eng.,
167 Wistuba, D. and Schurig, V. (1986) Angew. 9, 62–64.
Chem. Int. Ed. Engl., 25, 1032. 186 Weijers, C.A.G.M., de Haan, A., and de
168 Hechtberger, P., Wirnsberger, G., Bont, J.A.M. (1988) Appl. Microbiol.
Mischitz, M., Klempier, N., and Faber, K. Biotechnol., 27, 337–340.
(1993) Tetrahedron: Asymmetry, 4, 187 Choi, W.J., Puah, S.M., Tan, L.L., and Ng,
1161–1164. S.S. (2008) Appl. Microbiol. Biotechnol., 79,
169 Mischitz, M., Kroutil, W., Wandel, U., and 61–67.
Faber, K. (1995) Tetrahedron: Asymmetry, 188 Kotik, M., Brichac, J., and Kyslik, P. (2005)
7, 1261. J. Biotechnol., 120, 364–375.
170 Wandel, U., Mischitz, M., Kroutil, W., and 189 Lee, J.W., Lee, E.J., Yoo, S.S., Park, S.H.,
Faber, K. (1995) J. Chem. Soc., Perkin Kim, H.S., and Lee, E.Y. (2003) Biotechnol.
Trans. 1, 735. Bioprocess Eng., 8, 306–308.
414
190 Kim, H.S., Lee, S.J., Lee, E.J., 206 Goswami, A., Totleben, M.J., Singh, A.K.,
Hwang, J.W., Park, S., Kim, S.J., and and Patel, R.N. (1999) Tetrahedron:
Lee, E.Y. (2005) J. Mol. Catal. B: Enzym., Asymmetry, 10, 3167–3175.
37, 30–35. 207 Pedragosa-Moreau, S., Archelas, A., and
191 Kim, H.S., Lee, O.K., Lee, S.J., Hwang, S., Furstoss, R. (1996) Tetrahedron, 52,
Kim, S.J., Yang, S., Park, S., and Lee, E.Y. 4593–4606.
(2006) J. Mol. Catal. B: Enzym., 41, 208 Devi, A.V., Lahari, C., Swarnalatha, L.,
130–135. and Fadnavis, N.W. (2008) Tetrahedron:
192 Pedragosa-Moreau, S., Morisseau, C., Asymmetry, 19, 1139–1144.
Zylber, J., Archelas, A., Baratti, J.C., and 209 Weijers, C.A.G.M., Meeuwse, P.,
Furstoss, R. (1996) J. Org. Chem., 61, Herpers, R.L.J.M., Franssen, M.C.R., and
7402–7407. Sudh€olter, E.J.R. (2005) J. Org. Chem., 70,
193 Hwang, S., Hyun, H., Lee, B., Park, Y., 6639–6646.
Choi, C., Han, J., and Joo, H. (2006) J. 210 Doumeche, B., Archelas, A., and
Microbiol. Biotechnol., 16, 32–36. Furstoss, R. (2006) Adv. Synth. Catal., 348,
194 Liu, Z., Michel, J., Wang, Z., Witholt, B., 1948–1957.
and Li, Z. (2006) Tetrahedron: Asymmetry, 211 Bottalla, A., Ibrahim-Ouali, M., Santelli,
17, 47–52. M., Furstoss, R., and Archelas, A. (2007)
195 Jia, X., Wang, Z., and Li, Z. (2008) Adv. Synth. Catal., 349, 1102–1110.
Tetrahedron: Asymmetry, 19, 407–415. 212 Mateo, C., Archelas, A., Fernandez-
196 Hao, J. and Li, Z.-Y. (2004) Org. Biomol. Lafuente, R., Guisan, J.M., and Furstoss,
Chem., 2, 408–414. R. (2003) Org. Biomol. Chem., 1,
197 Pedragosa-Moreau, S., Morisseau, C., 2739–2743.
Baratti, J.C., Zylber, J., Archelas, A., and 213 Karboune, S., Archelas, A., Furstoss, R.,
Furstoss, R. (1997) Tetrahedron, 53, and Baratti, J.C. (2005) J. Mol. Catal. B:
9707–9714. Enzym., 32, 175–183.
198 Nellaiah, H., Morisseau, C., Archelas, A., 214 Karboune, S., Archelas, A., Furstoss, R.,
Furstoss, R., and Baratti, J.C. (1996) and Baratti, J.C. (2005) Biocatal.
Biotechnol. Bioeng., 49, 70–77. Biotransform., 23, 397–405.
199 Deregnaucourt, J., Archelas, A., 215 Kroutil, W., Orru, R.V.A., and Faber, K.
Barbirato, F., Paris, J.M., and (1998) Biotechnol. Lett., 20, 373–377.
Furstoss, R. (2006) Adv. Synth. Catal., 348, 216 Petri, A., Marconcini, P., and Salvadori, P.
200 Genzel, Y., Archelas, A., 219–224.
Lutje Spelberg, J.H., Janssen, D.B., and 217 Kronenburg, N.A.E., de Bont, J.A.M., and
Furstoss, R. (2001) Tetrahedron, 57, Fischer, L. (2001) J. Mol. Catal. B: Enzym.,
2775–2779. 16, 121–129.
201 Genzel, Y., Archelas, A., Broxterman, 218 Devi, A.V., Lahari, C., Swarnalatha, L.,
Q.B., Schulze, B., and Furstoss, R. (2001) and Fadnavis, N.W. (2008) Tetrahedron:
J. Org. Chem., 66, 538–543. Asymmetry, 19, 1139–1144.
202 Lotter, J., Botes, A.L., van Dyk, M.S., and 219 Thesis Baldascini, H.G. (2004)
Breytenbach, J.C. (2004) Biotechnol. Lett., Bioreaction engineering for the kinetic
26, 1197–1200. resolution of racemic epoxides
203 Cleij, M., Archelas, A., and Furstoss, R. by epoxide hydrolase. Available for
(1998) Tetrahedron: Asymmetry, 9, download at http://
1839–1842. dissertations.ub.rug.nl/FILES/faculties/
204 Monfort, N., Archelas, A., and science/2004/h.g.baldascini/thesis.pdf
Furstoss, R. (2004) Tetrahedron, 60, (accessed 20 June 2011)
601–605. 220 Manoj, K.M., Archelas, A., Baratti, J.C.,
205 Fujino, A., Asano, M., Yamaguchi, H., and Furstoss, R. (2001) Tetrahedron, 57,
Shirasaka, N., Sakoda, A., Ikunaka, M., 695–701.
Obata, R., Nishiyama, S., and Sugai, T. 221 Kronenburg, N.A.E. and de Bont, J.A.M.
(2007) Tetrahedron Lett., 48, 979–983. (2001) Enz. Microb. Technol., 28, 210–217.
References j415
222 Gorke, J.T., Srienc, F., and Kaslauskas, R. 238 Rui, L., Cao, L., Chen, W., Reardon, K.F.,
(2008) Chem. Commun., 1235–1237. and Wood, T.K. (2005) Appl. Environ.
223 Yeates, C.A., Krieg, H.M., and Microbiol., 71, 3995–4003.
Breytenbach, J.C. (2007) Enz. Microb. 239 Rui, L., Cao, L., Chen, W., Reardon, K.F.,
Technol., 40, 228–323. and Wood, T.K. (2004) J. Biol. Chem., 279,
224 Laane, C., Boeren, S., Vos, K., and 46810–46817.
Veeger, C. (1987) Biotechnol. Bioeng., 30, 240 Cedrone, F., Niel, S., Roca, S.,
81–87. Bhatnagar, T., Ait-Abdelkader, N.,
225 Baldascini, H., Ganzeveld, K.J., Torre, C., Krumm, H., Maichele, A.,
Janssen, D.B., and Beenackers, A.A. Reetz, M.T., and Baratti, J.C. (2003)
(2001) Biotechnol. Bioeng., 73, Biocatal. Biotransform., 21, 357–364.
44–54. 241 Reetz, M.T., Torre, C., Eipper, A.,
226 Baldascini, H. and Janssen, D.B. (2005) Lohmer, R., Hermes, M., Brunner, B.,
Enz. Microb. Technol., 36, Maichele, A., Bocola, M., Arand, M.,
285–293. Cronin, A., Genzel, Y., Archelas, A., and
227 van Loo, B., Permentier, H.P., Kingma, J., Furstoss, R. (2004) Org. Lett., 6, 177–180.
Baldascini, H., and Janssen, D.B. (2008) 242 Reetz, M.T., Wang, L.W., and Bocola, M.
FEBS Lett., 582, 1581–1586. (2006) Angew. Chem., 118, 1258–1263.
228 Deregnacourt, J., Archelas, A., 243 Reetz, M.T., Bocola, M., Wang, L.W.,
Barbirato, F., Paris, J.M., and Furstoss, R. Sanchis, J., Cronin, A., Arand, M., Zou, J.,
(2007) Adv. Synth. Catal., 349, 1405–1417. Archelas, A., Bottalla, A.L., Naworyta, A.,
229 Chiappe, C., Leandri, E., Hammock, and Mowbray, S.L. (2009) J. Am. Chem.
B.D., and Morisseau, C. (2007) Green Soc., 131, 7334–7343.
Chem., 9, 162–168. 244 Choi, S.H., Kim, H.S., and Lee, E.Y.
230 Choi, W.J., Choi, C.Y., de Bont, J.A., and (2009) Biotechnol. Lett., 31, 1617–1624.
Weijers, C.A.G.M. (1999) Appl. Microbiol. 245 Reetz, M.T. and Wang, L.W. (2006) Comb.
Biotechnol., 53, 7–11. Chem. High Throughput Screen., 9,
231 Choi, W.J., Choi, C.Y., de Bont, J.A., and 295–299.
Weijers, C.A.G.M. (2000) Appl. Microbiol. 246 Miller, R.E. and Guengerich, F.P. (1982)
Biotechnol., 54, 641–646. Biochemistry, 21, 1090–1097.
232 Rink, R., Kingma, J., Lutje Spelberg, J.H., 247 Rink, R., Fennema, M., Smids, M.,
and Janssen, D.B. (2000) Biochemistry, 39, Dehmel, U., and Janssen, D.B. (1997)
5600–5613. J. Biol. Chem., 272, 14650–14657.
233 Rink, R., Lutje Spelberg, J.H., 248 Zocher, F., Enzelberger, M.M.,
Pieters, R.J., Kingma, J., Nardini, M., Bornscheuer, U.T., Hauer, B., and
Kellogg, R.M., Dijkstra, B.W., and Schmid, R.D. (1999) Anal. Chim. Acta,
Janssen, D.B. (1999) J. Am. Chem. Soc., 391, 345–351.
121, 7417–7418. 249 Doderer, K., Lutz-Wahl, S., Hauer, B., and
234 Lutje Spelberg, J.H., Rink, R., Schmid, R.D. (2003) Anal. Biochem., 321,
Archelas, A., Furstoss, R., and 131–134.
Janssen, D.B. (2002) Adv. Synth. Catal., 250 Wahler, D. and Reymond, J.-L. (2002)
344, 980–985. Angew. Chem. Int. Ed., 41, 1229–1232.
235 Genzel, Y., Archelas, A., 251 Doderer, K. and Schmid, R.D. (2004)
Lutje Spelberg, J.H., Janssen, D.B., and Biotechnol. Lett., 26, 835–839.
Furstoss, R. (2001) Tetrahedron, 57, 252 Badalassi, F., Wahler, D., Klein, G.,
2775–2779. Crotti, P., and Reymond, J.-L. (2000)
236 Kotik, M., Stepanek, V., Kyslik, P., and Angew. Chem. Int. Ed., 39, 4067–4070.
Maresova, H. (2007) J. Biotechnol., 132, 253 Mantovanni, S.M., de Oliveira, L.G., and
8–15. Marsaioli, A.J. (2008) J. Mol. Catal. B:
237 van Loo, B., Lutje Spelberg, J.H., Kingma, Enzym., 53, 173–177.
J., Sonke, T., Wubbolts, M.G., and 254 Westkaemper, R.B. and Hanzlik, R.P.
Janssen, D.B. (2004) Chem. Biol., 11, (1981) Arch. Biochem. Biophys., 208,
981–990. 195–204.
416
255 Dietze, E.C., Kuwano, E., and 265 Hwang, S., Choi, C.Y., and Lee, E.Y.
Hammock, B.D. (1994) Anal. Biochem., (2008) Biotechnol. Bioprocess Eng., 13,
216, 176–187. 453–457.
256 Schrader, W., Eipper, A., Pugh, D.J., and 266 Cao, L., Lee, J., Chen, W., and Wood, T.K.
Reetz, M.T. (2002) Can. J. Chem., 80, (2006) Biotechnol. Bioeng., 94,
626–636. 522–529.
257 Pedragosa-Moreau, S., Morisseau, C., 267 Kim, H.S., Lee, O.K., Hwang, S.,
Baratti, J.C., Zylber, J., Archelas, A., and Kim, B.J., and Lee, E.Y. (2008)
Furstoss, R. (1997) Tetrahedron, 53, Biotechnol. Lett., 30,
9707–9714. 127–133.
258 Monterde, M.I., Lombard, M., 268 Xu, W., Xu, J.H., Pan, J., Gu, Q.,
Archelas, A., Cronin, A., Arand, M., and and Wu, X.Y. (2006) Org. Lett., 8,
Furstoss, R. (2004) Tetrahedron: 1737–1740.
Asymmetry, 15, 2801–2805. 269 Archer, I.V.J., Leak, D.J., and
259 Hwang, S., Choi, C.Y., and Lee, E.Y. Widdowson, D.A. (1996) Tetrahedron
(2008) Biotechnol. Lett., 30, Lett., 37, 8819–8822.
1219–1225. 270 Orru, R.V.A., Kroutil, W., and Faber, K.
260 Kroutil, W., Mischitz, M., Plachota, P., (1997) Tetrahedron Lett., 38,
and Faber, K. (1996) Tetrahedron Lett., 37, 1753–1754.
8379–8382. 271 Jerina, D.M., Ziffer, H., and Daly, J.W.
261 Steinreiber, A., Mayer, S.F., Saf, R., and (1970) J. Am. Chem. Soc., 92, 1056–1061.
Faber, K. (2001) Tetrahedron: Asymmetry, 272 Belluci, G., Capitani, I., and Chiappe, C.
12, 1519–1528. (1989) J. Chem. Soc., Chem. Commun.,
262 Pedragosa Moreau, S., Archelas, A., and 1170.
Furstoss, R. (1996) Tetrahedron, 52, 273 Chang, D., Wang, Z., Heringa, M.F.,
4593–4606. Wirthner, R., Witholt, B., and Li, Z. (2003)
263 Manoj, K.M., Archelas, A., Baratti, J.C., Chem. Commun., 960–961.
and Furstoss, R. (2001) Tetrahedron, 57, 274 van Loo, B., Kingma, J., Heyman, G.,
695–701. Wittenaar, A., Lutje Spelberg, J.H.,
264 Pedragosa-Moreau, S., Archelas, A., and Sonke, T., and Janssen, D.B. (2009)
Furstoss, R. (1993) J. Org. Chem., 58, Enz. Microb. Technol., 44,
5533–5536. 145–153.
j417
10
Hydrolysis and Formation of Glycosidic Bonds
Daniela Monti and Sergio Riva
10.1
Introduction
Carbohydrates play an important role in many biological events, such as cellular

communication and physiological responses. In fact, oligosaccharides and
various glycoconjugates have been shown to modulate different molecular
recognition processes, including viral and bacterial pathogens recognition [1–5],
tumor-associated cell adhesion and metastasis events [6–8], immune response [9],
fertilization [10], and neuronal development [11]. This significant role, together with
advances in the glycomics field and the recent developments of new synthetic
methodologies, for example, the glycorandomization of biologically active natural
products [12, 13], makes glycoconjugates extremely attractive therapeutic targets
nowadays [14–17].
However, the presence of multiple functional groups and stereocenters in carbo-
hydrates makes them challenging targets for organic chemists. Time-consuming
protecting group manipulations and complex synthetic schemes are often required to
suitably discriminate among hydroxyl groups with similar reactivities and/or to
obtain the desired glycosidic linkage by using specifically activated donors and
acceptors. Both regioselectivity and stereospecificity need to be strictly controlled
and, as the number of different monosaccharide units as well as of glycan structures
of interest is very large, a general methodology for oligosaccharide synthesis is still far
from being reached. Moreover, despite recent progress in the solid-phase synthesis of
oligosaccharides [14, 16, 18], an automated oligosaccharide synthesizer for all
purposes is not yet commercially available and not all the common glycosidic
linkages are accessible. Even the structural characterization of the saccharidic part
of novel glycoconjugates from natural sources is still a demanding task – no
automated sequencing techniques are fully developed, unlike the case of nucleic
acids and peptides.
Finally, the natural complexity of glycoconjugates is amplified by the extreme
variety of aglycon structures coupled to the glycan chains, which can range from
and Oliver May.
418 j 10 Hydrolysis and Formation of Glycosidic Bonds
HO NH2
H3C O
CH3 OH O
O O OH
O OH
OH
O OH
Cl OH
O O
OMe O OH O
HO Cl OH
O O H3C
H H H O
O N N N
O N N N CH3 NH2
H H H
NH O O CH3 OH
HO
O Doxorubicin
NH2 CH3
(anticancer)
HO OH OH Vancomycin (antibiotic)
OH NHAc OH
HO O OH OH
HO O C13H27
O O O OH H3 C
HO HO O
OH OH OH O O O C17H35
O O HN
OH HO
O
HO OH O
H3C
O HO O
Ganglioside GM1a (cholera toxin receptor)
HO O
OH OH
HO H3C
NHAc O O
HO HO
H3C
HO O O
HN OH
HO
OMe
OH
Bacterial spore surface tetrasaccharide
O O ( anthrax toxin)
HO
HO NHCOCH3
OH
MeO
HO O
OMe
H3CSSS NHAc
O
Colchicoside OMe O H
(anti-inflammatory) I O O O
S N
H HO
O
O OMe OH O
OMe H O
HO O N
MeO MeO
OH
Calicheamicin γ1I (anticancer)
Figure 10.1 Some representative biologically active glycoconjugates.
peptides and proteins to lipids, nucleic acids, antibiotics, steroids, and so on (see
Figure 10.1 for some representative structures).
As an alternative to the organic chemistry approach, carbohydrate-active biocata-
lysts have been investigated in depth in recent years both for the synthesis and for the
characterization of oligosaccharides and glycoconjugates.
10.2 Glycosidases j419
Specifically, glycosidases (EC 3.2.1.-) and glycosyltransferases (EC 2.4.-) are the two
classes of carbohydrate-active enzymes mostly involved in the synthesis, degradation,
and modification of glycoconjugates. Beside the EC classification – according to the
recommendations of the Nomenclature Committee of the International Union of
Biochemistry and Molecular Biology (IUBMB) – that is mainly related to the substrate
specificity of the enzymes and occasionally to their molecular mechanism, glycosi-
dases and glycosyltransferases have been recently classified on the basis of their
structural features and sequence similarities in the Carbohydrate-Active EnZyme
database (CAZy, www.cazy.org) [19].
10.2
Glycosidases
Glycosidases (or glycoside hydrolases) are a widespread group of enzymes that

hydrolyze the glycosidic bond between two or more carbohydrates or between a
carbohydrate and a non-carbohydrate moiety. At present, they form 115 different
sequence-based families and, according to the stereochemical outcome of the
hydrolysis reaction, they can be classified as either retaining or inverting
enzymes. Moreover, they can also be defined as exo- or endo-acting, dependent
upon whether they act at the (usually non-reducing) end or in the middle, respec-
tively, of an oligo/polysaccharidic chain. Their main characteristic is the strict
stereoselectivity towards the glycosidic bond to be broken (a-glycosidases hydrolyze
only a-glycosidic bonds, whereas b-glycosidases act on b-glycosides) and the pro-
nounced selectivity for the remaining part of the sugar structure to be cleaved.
A combination of these two properties is normally used to unequivocally identify
the different exoglycosidases (i.e., a-mannosidases, b-galactosidases, b-xylanases,
and so on).
As the breakdown of glycan structures is of primary importance in many
processes in the agricultural, pulp and paper, textile, and food industries, glycosi-
dases have been extensively investigated in recent years to improve or alter their
catalytic activity and substrate specificity [20–22]. This includes the application of
cellulases and xylanases in the biotransformation of lignocellulosic biomass for the
production of biofuels, a process that has been described in several recent
reviews [23–25].
Glycosidases have been also used to synthesize oligosaccharides by transglycosy-
lation or reverse hydrolysis reactions using either monosaccharides, oligosacchar-
ides, or activated glycosides as glycosyl donors [26]. This synthetic approach is
attractive for the general availability of cheap substrates and enzymes and for the
robustness shown by glycosidases acting in various reaction conditions, but it has
often been limited by the scarce regioselectivity of these biocatalysts and/or by
unsatisfactory reaction yields. However, different engineering strategies, involving
both the substrates and the enzymes, have allowed new interesting perspectives to
open up for valuable exploitation of glycosidases for complex carbohydrates
synthesis [27].
10.2.1
Catalytic Mechanism
The enzymatic hydrolysis of glycosidic bonds is mechanistically similar to their acid-

catalyzed breaking. According to their mechanism, these enzymes can be classified
as inverting or retaining glycosidases, with some variants (i.e., the retaining mech-
anism of b-N-acetylhexosaminidases) [27, 28].
The presence of two proximal carboxylates in the active site is a common structural
motif of these biocatalysts; additionally, all of them are thought to catalyze the
formation of a transition state having oxonium character [29, 30]. It was long debated
whether an oxonium ion intermediate is also formed as a stabilized ion pair, or if the
hydrolysis proceeds via a glycosidic ester intermediate. Eventually, an a-glycosyl
enzyme intermediate could be observed by 19F NMR in a b-glucosidase-catalyzed
hydrolysis of 2-deoxy-2-fluoro-D-glucosyl fluoride, and it was shown to be a catalyt-
ically competent species [31, 32].
10.2.1.1 Inverting Glycosidases

These enzymes hydrolyze glycosidic bonds by a single-step mechanism driven
by acid–base catalysis. As shown in Scheme 10.1a, the aglycon leaving group is
directly displaced by a nucleophilic water molecule with a single inversion at the
anomeric center. Examples of inverting glycosidases are trehalases [33] and
b-amylases [34].
(a) Acid-base
O O _O O
X H X
O O
HO OR HO
HO H HO
HO HO
O OH
H ROH
_
O O O OH
Nucleophile
(b) Acid-base
O O O_ O O O
X H H X H
O X O O
HO OR O HO OH
HO H
HO HO
HO HO HO
_ HO _
O O ROH O O O O
Nucleophile
Scheme 10.1 Mechanisms of (a) inverting glycosidases and (b) retaining glycosidases.
10.2.1.2 Retaining Glycosidases
These enzymes hydrolyze glycosidic bonds by a double-displacement mechanism
with a double inversion at the anomeric center (Scheme 10.1b). In the first step the
enzyme is glycosylated by the concerted action of the two carboxylates in the active
site; while the acid residue protonates the aglycon, making it a good leaving group,
the second carboxylate acts as a nucleophile, leading to the formation of a glycosyl-
enzyme intermediate. In the second step the first carboxylate residue acts as a base
and deprotonates a water molecule (or another nucleophile, see below), which attacks
the glycosyl-enzyme intermediate, yielding to the hydrolyzed products. Examples of
retaining glycosidases are lysozymes and b-galactosidases.
10.2.2
Glycosidases Inhibitors
Carbohydrate-cleaving processes are present in all living systems. Therefore, unsur-

prisingly, glycosidases are implicated in several pathologies, and it sounds reasonable
that the selective inhibition of these enzymes might offer promising therapeutic
routes. Non-covalent, reversibly binding glycosidases inhibitors have great potential
in human medicine for the therapy of different illnesses, such as viral infections,
cancer, Alzheimers disease, and genetic disorders [35, 36]. Applications in agricul-
tural pest control, for example, as antifungal agents and insecticides, have also been
proposed [37].
The three-dimensional structure of glycosidases and the study of molecules that
mimic the oxonium ion character of the glycosides in the transition state enabled the
identification of inhibitors with Ki values from 109 to 1010 M [38, 39]. For instance,
well-known examples of glycosidases inhibitors are given by the alkaloids deoxyno-
jirimycin, swainsonine, and castanospermine (Figure 10.2) [40], all of which can be
classified as amino-polyols.
10.2.3
Synthetic Applications of Glycosidases
Glycosidases found applications long before being classified as enzymes [41].

Thousands of patents and papers have been published reporting on their exploitation
in synthesis or for technological applications. Just a few representative examples will
be discussed here to offer a clear picture of the potential and also the drawbacks
related to the use of these biocatalysts.
OH HO OH HO OH
HO OH HO
OH
OH N N
N HO
H
Deoxynojirimycin Swainsonine Castanospermine
Figure 10.2 Examples of glycosidases inhibitors.

j 10 Hydrolysis and Formation of Glycosidic Bonds
422
10.2.3.1 Glycosidase-Catalyzed Hydrolysis of Glycosidic Bonds

Belonging to class 3 hydrolases, glycosidases have obviously found applications based
on their natural activity. The use of these biocatalysts is well known for the selective
and mild trimming of glycoproteins, both for producing the naked peptide chains
and for obtaining structural information on the oligosaccharidic substituents [42].
In this respect, several glycosidases kits are now commercially available.
Preparative-scale synthesis of di-, tri-, and higher oligosaccharides fragments can
be performed using endoglycosidases on larger oligosaccharides or polysaccharides.
For instance, digalacturonic acid has been prepared on the kg scale from pectin by
action of a pectinase [43].
The use of these enzymes has been suggested as a simple and efficient way to
resolve anomerically impure glycosides obtained from classical acid-catalyzed
Fischer glycosylation [44].
In another example related to natural compounds, glycosidases have been
exploited to unequivocally establish the structure of kenyaloside (1, Figure 10.3),
a metabolite isolated from the exudate of a Kenyan Aloe species. Its structure was
established by the combined use of a b-D-glucosidase, an a-L-rhamnosidase, and
spectral and chemical methods [45].
If needed, it can be relatively easy to generate a library of these enzymes. For
instance, this can be achieved by screening different fungal strains grown under
different cultivation conditions and in the presence of different specific inductors.
This approach has been exemplified in the generation of an a-L-rhamnosidase library
that was applied for the selective derhamnosylation of natural glycosides [46].
Specifically, derivatives of the complex saponin asiaticoside (2a, Figure 10.3) were
generated by exploiting an a-L-rhamnosidase isolated from a culture of Fusarium
oxysporum to produce the desrhamno-desgluco-asiaticoside (2c) and also the
OR OR' O
R = α-L-rhamnopyranosyl
R' = β-D-glucopyranosyl
OR"
R" = β-D-xylopyranosyl
Kenyaloside (1)
HO Me
OH O
HO OH
OH HO OH
(2a), R = O
HO
O O O
OR OH OH
H HO
OH HO
HO (2b), R = O
HO
O OH
O O
OH OH
HO HO
(2c), R = O OH
HO
OH
Figure 10.3 Chemical structures of kenyaloside (1) and asiaticoside derivatives (2a–c).
desrhamno-asiaticoside (2b) by in situ glucose-inhibition of contaminating
b-D-glucosidases [47].
10.2.3.2 Glycosidases-Catalyzed Formation of Glycosidic Bonds

Analogously to what can be achieved with other hydrolases, the catalytic activity of
glycosidases can be reversed and their striking selectivity can be exploited to form
glycosidic bonds. However, contrary to lipases and proteases, these enzymes are not
active when suspended in pure organic solvents, even the more polar ones that can
easily dissolve sugars and their derivatives. As a possible explanation to this behavior,
it was suggested that the affinity of sugars for the polar active site of glycosidases is so
strong that only water is able to displace them, making the enzyme competent again
for a new catalytic cycle. Instead, in organic solvents the products stick into the active
site, acting as competitive inhibitors [48].
In contrast, glycosylation reactions can be achieved upon working in water
solutions (with or without water-miscible organic cosolvents) or in biphasic systems
using two different approaches: equilibrium-controlled synthesis (the so-called
reverse hydrolysis) and kinetically-controlled synthesis (transglycosylation reac-
tions, Scheme 10.2).
H2O ROH
Glycosyl-OR Glycosyl-OH
Hydrolysis
R'OH R'OH
Transglycosylation Reverse hydrolysis
ROH H2O
Glycosyl-OR'
Scheme 10.2 Reactions catalyzed by glycosidases.
Both approaches take advantage of the great versatility of glycosidases towards the
aglycon moiety. In fact these enzymes can glycosylate many xeno-substrates
carrying primary or secondary OH groups.
A major advantage of glycosidase-catalyzed glycosylation is the minimal need of
preliminary protection steps of reactive functionalities in the donor and in the sugar
acceptor molecules. Additionally, the stereochemistry at the anomeric center can be
completely controlled through choice of the appropriate enzyme, that is, an a- or a
b-glycosidase. The main drawback is related to the scarce regioselectivity of these
enzymes: when the acceptor is a sugar usually a mixture of isomeric di- and
oligosaccharides is formed.
Equilibrium-controlled syntheses are based on the thermodynamic approach: a
free non-activated monosaccharide is used as substrate and the reaction equilibrium
is shifted towards the synthesis of glycosides (the so-called reverse hydrolysis)
by using high concentrations of both the monosaccharide and the nucleophilic
component (carbohydrate or a non-sugar alcohol). Though quite simple in theory,
this approach generally provides poor yields and affords side products that cannot be
easily prevented. To shift the equilibrium towards product formation, the addition of
water-miscible cosolvents (that is working at low water activity) or the use of biphasic
systems (an approach based on the removal of the product) in which the alcoholic
nucleophile can even constitute a separate phase by itself have been investigated.
Generally speaking, this approach is less important for oligosaccharide
synthesis but it has been successfully applied for the preparation of simple alkyl
derivatives, such as allyl or benzyl glycosides (Scheme 10.3) [49–51] or for amino acid
glycosides [52]. Glycosides bearing a spacer moiety, which can be used to make
glycoconjugates or be coupled to solid supports, as well as building blocks for
glycopolymers can also be obtained using functionalized alcohols [51, 53].
HO OH
HO OH HO HO O
OH OH
NHCOCF3
O O
HO
HO O HO HO
OH
HO β-galactosidase HO α-galactosidase O
NHCOCF3
Scheme 10.3 Examples of functionalized alkyl galactosides obtained using the reverse hydrolysis
approach.
Owing to the importance of alkyl glycosides as surfactants or as controlled-release

precursors of flavor and fragrances, a significant amount of work has been performed
to scale up these reactions [54] and scientists are still quite active in this area of applied
research.
Kinetically-controlled synthesis relies on transglycosylation reactions. As shown
in Scheme 10.4, a glycosyl-enzyme intermediate is formed from a suitable sugar
donor (aryl glycosides, glycosyl fluorides, di- or oligosaccharides) then the sugar
HO OH
O OH
O
HO O OH HO OH
HO HO
HO O
HO OH
HO
β-galactosidase
glucose H2O
Hydrolysis Hydrolysis
(primary) H2O (secondary)
HO OH
O
HO R'OH
HO
O Enzyme HO OH
Synthesis O
HO OR'
HO
Scheme 10.4 Kinetically-controlled synthesis of b-D-galactopyranosides from lactose catalyzed by a

b-galactosidase.
moiety is subsequently transferred to a suitable nucleophilic acceptor (an alcohol or
another sugar moiety). As both the glycosyl-enzyme intermediate and the product
can suffer enzymatic hydrolysis by action of water, the success of this procedure is
based on the following crucial parameters:
. The transglycosylation reaction must be faster than the hydrolysis of the glycosyl
donor.
. The rate of hydrolysis of the product must be less than that of the glycosyl donor.
In addition, the reactions must be carefully controlled and arrested when the
glycosyl donor has been mostly consumed so as to minimize subsequent product
hydrolysis.
Generally speaking, a transglycosylation reaction gives higher yields and is more
rapid than the reverse hydrolysis, even though yields remain quite low and rarely
exceed 30%. For instance, using phenyl b-glucopyranoside as a donor and simple
alcohols as acceptors in reactions catalyzed by a lactase from Kluyveromyces lactis,
the corresponding alkyl b-glucopyranosides were isolated in 5 to 67% yields, with the
best results being obtained with methanol and n-butanol [55]. Later, allyl, benzyl,
and trimethylsilylethyl b-galactopyranosides from the corresponding alcohols were
prepared on a 1–20 g scale using a b-galactosidase and lactose as a donor [56]. In early
studies, transglycosylation reactions were also studied for the kinetic resolutions of
racemic alcohols [57] or for the desymmetrization of meso compounds [58]. However,
this approach proved to be much less efficient than the one based on esterase- and
protease-catalyzed acylations or deacylations reactions.
In other examples dealing with bioactive natural compounds, the same protocol
was used for the galactosylation of antibiotics (i.e., 3, Figure 10.4, 12% yield) [59], for
the synthesis of alkaloid glycosides (i.e., 4, 2–24% yield) [60], and for cardiac
glycosides (i.e., 5) [61]. In the latter report the yields were quite low (3%) and
these unattractive results emphasize the problems often observed when trying to
enzymatically glycosylate a secondary alcohol.
Glycosylation of amino acids and peptides bearing hydroxyl groups have also been
investigated, starting from serine and its protected derivatives. Depending on
reactions conditions isolated yields were in the range 13–25% [62, 63]. Glycosylated
serine derivatives have been also used as acceptors for additional sugar moieties. For
HO OH
O
HO O O
HO
O
HO OH OH
O
HO O
H H OH
HO NH N
Cl NO2 Me HO
H OH H OH
O O
3 HO O
HO H 5
N 4
H
Figure 10.4 Chemical structures of glycosylated compounds 3–5.

426
HO HO OH
HO OH OH
O O
O HO OR OH
O HO
HO NH2 HO 6' O O
HO AcHN O O
O OH
HO HO
COOH HO
6 7 (R = H)
Figure 10.5 Chemical structures of glycosylated compounds 6 and 7.
instance, the 3-O-galactosylated derivative 6 (Figure 10.5) could be isolated in 68%

yield, exploiting a b-galactosidase from Bacillus circulans [64]. The complexity of
the donor and acceptor molecules has been growing significantly, and finally it
could be successfully applied to the synthesis of glycoproteins. For instance, by using
the enzyme ribonuclease as a model system, it was shown that an endo-b-N-
acetylglucosaminidase from Arthrobacter protophormiae (Endo-A) could efficiently
attach a preassembled oligosaccharide, activated as a sugar oxazoline, to a GlcNAc-
containing protein in a regio- and stereospecific manner. The corresponding
products (containing a penta- or an heptasaccharide) could be isolated in 82–96%
yields [65].
The results described above are really quite remarkable. In fact, the strict stereo-
selectivity of glycosidases for the configuration of the anomeric carbon is usually
matched by a significantly low regioselectivity. In general, the primary OH of the
acceptor reacts preferentially compared to its secondary hydroxyl groups, but the
difference is not absolute and, moreover, no significant preference among different
secondary OH groups is usually observed. As a consequence, when another sugar is
used as an acceptor, usually complex mixtures of isomeric products are formed and
this drawback, always accompanied by competitive enzymatic hydrolysis, signifi-
cantly lowers product yields.
In an attempt to optimize glycosidases regioselectivity, several parameters have
been investigated, such as the temperature, concentration of organic cosolvent,
reactivity of the activated donor, nature of the aglycon, and the anomeric configu-
ration of the acceptor glycoside (so-called anomeric control) [66].
The influence on glycosidase regioselectivity towards functional groups intro-
duced at different positions of sugar acceptors has also been investigated. For
instance, 60 -O-acyl-lactose derivatives were prepared by subtilisin-catalyzed acylation.
These compounds were used as acceptors for a transglycosylation reaction catalyzed
by an a-galactosidase from Talaromyces flavus to provide, after deacylation, the
trisaccharide iso-globotriose (7, Figure 10.5) in 20–30% yield [67].
Probably the best approach to improve the regioselectivity of these reactions is to
look for new enzymes with improved performance. For instance, it was found that a
b-galactosidase from testis is more selective for the secondary C3 OH of the sugar
acceptor. In this way the preferential formation of Galb1,3GlcNAc from lactose and
GlcNAc was observed [68]. This approach is facilitated by the fact that, as discussed
previously, it can be relatively easy to produce a library of glycosidases. A further
example is the auto-condensation of p-nitrophenyl a-D-galactopyranoside catalyzed
by a-galactosidases. By screening a panel of 33 different enzymatic activities it was
possible to identify a galactosidase from an Aspergillus terreus strain selective for the
formation of an a-(1 ! 6) bond, another enzyme (from a Talaromyces flavus strain)
with a strong preference for the formation of an a-(1 ! 3) bond, and a third one (from
a Circinella muscae strain) with a preference for the formation of an a-(1 ! 2)
bond [69].
10.2.4
Glycosynthases
Asdiscussed above,one of the major drawbacks inthe glycosidase-catalyzed synthesisof

oligosaccharides is the concomitant competitive hydrolysis of the products
(Scheme 10.4). To solve this problem the use of artificial glycosidases obtained by
site-directed mutagenesis, the so-called glycosynthases, has been suggested.
This approach was in some way anticipated in a seminal paper published in
1994 [70]. This report focused on glycosidase mechanisms and, specifically, it was
shown how to convert a retaining enzyme into an inverting one by a single site-
specific mutation. This goal was achieved by mutating the nucleophilic residue in the
active site of a retaining b-glucosidase from Agrobacterium faecalis (a Glu in this
specific case, see Scheme 10.1) into a neutral alanine. The mutant Glu358Ala was, as
expected, virtually inactive (kcat was 107-fold lower than that of the wild-type enzyme)
and direct attack of water to p-nitrophenyl-glucopyranoside was extremely inefficient.
However, addition of formate or azide ions increased kcat 105-fold, almost back to
wild-type levels. An a-glucosyl azide intermediate could be identified by 1H NMR,
thus confirming the inverting mechanism of the mutant enzyme. To further support
the inverting mechanism, it was shown that fluoro a-glucopyranoside could be
hydrolyzed by the mutant enzyme, while it was recovered unaffected when it was
treated with the wild-type glucosidase. Incidentally, in the presence of the mutant
catalyst a cellobiose derivative was formed, a disaccharidic product that the mutated
enzyme was not able to hydrolyze (Scheme 10.5). The conclusion of this paper that
The potential use of such mutant glycosidases in oligosaccharides synthesis is
currently being explored gave origin to the investigations on glycosynthase (for
reviews see References [27, 71, 72]).
Acid-base Acid-base
O_ O HO O
OH OH OH
H OH
O O
O O HO O
HO O
HO HO
HO HO HO HO
HO F _ HO
F F F
CH3 CH3
Scheme 10.5 Glycosidic bond formation catalyzed by a glycosynthase.

428
The single amino acid site-directed mutation approach has been used to generate
inverting glycosynthases from retaining wild-type enzymes [71, 73] and retaining
glycosynthases from inverting wild-type glycosidases [74, 75]. Recent reports
have described new glycosynthases able to transfer glucoronyl [76, 77] or N-acetyl-
glucosamine residues [78]. Other examples are related to the development
of thioglycoligases (single mutants) [79] and thioglycosynthases (double
mutants) [80].
The noteworthy performances of these rationally designed artificial enzymes have
been documented by high isolated yields (up to 90% with a glycosynthase derived
from a Thermus thermophilus b-glycosidase selective for the formation of b-1,3
bonds [81]) and by the ability to catalyze new reactions. For instance, the mentioned
thioglycoligases and thioglycosynthases represent the only way to obtain thioglyco-
sides using glycosyl hydrolases.
In fact, although some natural glycosidases were shown to cleave thioglycosides
with efficiency comparable to O-glycosides, they have never been shown to be able to
reverse their catalytic hydrolyzing activity and to synthesize thioligosaccharides.
10.3
Glycosyltransferases
Glycosyltransferases are responsible for the biosynthesis of oligosaccharides and

other glycoconjugates by transferring sugar moieties from donor molecules to
acceptors whose possible structural diversity can be as large as that of natural
glycoconjugates [82, 83]. Glycosyltransferases (GTs) constitute 91 different protein
families and are traditionally divided into two main groups, Leloir pathway GTs and
non-Leloir GTs, according to the type of glycosyl donor used in the glycosylation
reaction. In fact, Leloir pathway GTs exclusively use nucleotide sugars whereas non-
Leloir GTs can use phosphorylated glycosyl donors or even non-activated di- or
oligosaccharides, that is, sucrose or starch-derived dextrins.
Both GT groups have been widely investigated for synthetic applications, showing
different advantages and disadvantages. Non-Leloir GTs are particularly attractive
for their use of cheap glycosyl donors, but the reversibility of the catalyzed
reactions and the limited array of monosaccharides transferable restrict the biotech-
nological potential of this group of enzymes. On the other hand, Leloir pathway
GTs use various nucleotide sugar donors and show absolute regio- and stereose-
lectivity in the catalyzed reactions, thus allowing quantitative recovery of the newly
glycosylated products. The major limitations to the synthetic application of Leloir
pathway GTs have been related to the enzyme availability and to the cost of nucleotide
sugars. Both issues have been tackled recently and in some cases solved by
using heterologous expression systems and sugar nucleotide regeneration systems,
respectively. Moreover, interest in Leloir pathway GTs has been renewed recently by
the discovery of numerous bacterial enzymes, both from actinomycetes and path-
ogenic bacteria strains, whose synthetic potential has been only partially investigated
up to now.
10.3 Glycosyltransferases j429
10.3.1
Glycosyltransferases of the Leloir Pathway
Glycosyltransferases (GTs) of the so-called Leloir pathway catalyze the transfer of a

sugar residue from an activated nucleotide sugar donor to specific acceptor mole-
cules, forming new glycosidic bonds with an almost absolute regioselectivity and
stereospecificity.
Leloir pathway GTs are ubiquitous enzymes present in both prokaryotes and
eukaryotes. In mammalian cells they are mostly expressed in the endoplasmic
reticulum and Golgi apparatus as membrane-bound enzymes and are responsible
for the assembly process of glycoprotein oligosaccharides and for the synthesis of
other glycoconjugates. Only nine different nucleotide sugar donors (UDP-Glc, UDP-
Gal, UDP-GlcNAc, UDP-GalNAc, UDP-GlcUA, UDP-Xyl, GDP-Fuc, GDP-Man, and
CMP-Neu5Ac, Figure 10.6) are usually accepted by mammalian GTs, but, thanks to
the variety of possible glycoside acceptors and of formed glycosidic linkages, a large
number of different oligosaccharides can be generated for glycoproteins having post-
translational modification.
OH HO OH OH
O O O
HO
HO HO HO
HO
HO HO AcHN
OUDP OUDP OUDP
α-UDP-D-glucose α-UDP-D-galactose α-UDP-N-acetyl-D-glucosamine

(UDP-Glc) (UDP-Gal) (UDP-GlcNAc)
HO OH CO2H
O O O
HO HO
HO HO
HO
AcHN HO HO
OUDP OUDP OUDP
α-UDP-N-acetyl-D-galactosamine α-UDP-D-glucuronic acid α-UDP-D-xylose

(UDP-GalNAc) (UDP-GlcUA) (UDP-Xyl)
HO OH
HO HO OCMP
Me O
O OGDP HO
OH HO HO O CO2H
OH AcHN
OH OH
OGDP
β-GDP-L-fucose α-GDP-D-mannose β-CMP-N-acetyl-neuraminic acid

(GDP-Fuc) (GDP-Man) (CMP-Neu5Ac)
Figure 10.6 Nucleoside diphosphate (NDP) and nucleoside monophosphate (NMP) sugars used
as donors by mammalian Leloir GTs.
430
As proteins are usually not glycosylated in microorganisms, GTs play different

roles in prokaryotes. Specifically, they can be involved in cell-wall polysaccharides
synthesis, as, for example, the enzyme MurG from Escherichia coli, which acts in
peptidoglycan biosynthesis using UDP-GlcNAc as a glycosyl donor [84], or SpsA from
Bacillus subtilis, implicated in spore coat formation [85]. Moreover, bacterial GTs are
involved in the biosynthesis of specific glycoconjugates that have important functions
in the virulence of pathogenic microorganisms such as Neisseria gonorrhoeae, N.
meningitidis, E. coli, Campylobacter jejuni, Streptococcus sp., and Haemophilus influ-
enzae. In fact, these bacteria may display mimics of human glycan structures (e.g.,
sialylated oligosaccharides) on their cell surfaces, being consequently not recognized
by the immune system of the host as foreign [86]. Finally, a large group of GTs are
expressed by polyketide-producing microorganisms, mainly actinomycetes belong-
ing to the Streptomyces genus, for the biosynthesis of various glycosylated natural
compounds with antibiotic, antitumor, cholesterol-lowering, immunosuppressive,
and antifungal activity [83, 87]. When compared to the other Leloir GTs, these
antibiotic GTs show a different specificity towards the sugar donors by accepting a
large array of nucleoside diphosphate (NDP) sugars (mainly TDP-6-deoxyhexoses)
produced by the same microorganism through an extremely various set of biochem-
ical pathways (Scheme 10.6).
Analog to glycosidases, GTs can be mechanistically classified as inverting or
retaining enzymes depending on whether the stereochemistry of the anomeric
OH
O
HO D-Glucose-1-phosphate
HO
HO
OPO3H2
TTP
thymidylyl
PPi transferase
OH
O
HO TDP-D-Glucose
HO
HO
C-4 OTDP
Deoxygenation
O-methylation
Transamination 4,6-dehydratase
N-methylation H2O
Keto-reduction
C-5 Epimerization, C-methylation
O Me O
C-3
HO
Epimerization
HO
Deoxygenation OTDP TDP-4-keto-6-deoxy-D-Glucose
O-methylation
C-methylation
Transamination C-2
N-methylation Deoxygenation
Keto-reduction O-methylation
Scheme 10.6 Possible biosynthetic pathways of modified TDP-deoxysugars in actinomycetes.

carbon is retained or inverted in the product relative to that in the donor substrate [88].
Inverting glycosyltransferases most likely follow a single displacement SN2-like
mechanism in which the acceptor performs a nucleophilic attack at carbon C1 of
the sugar donor, which is to some extent similar to the mechanism of inverting
glycosidases. Participation of an acidic amino acid (either Asp or Glu) in the active site
is usually required for activation of the acceptor hydroxyl group by deprotonation.
The mechanism of retaining glycosyltransferases is less clear. The proposed double-
displacement mechanism – in analogy with the one observed with retaining glyco-
sidases – has not been supported by crystal structure analyses of GTs. Therefore, an
internal return SNi-like mechanism, involving a short-lived oxocarbenium interme-
diate, has been proposed, in which leaving group departure and nucleophilic attack
occur in a concerted but asynchronous manner on the same face of the glycoside [89].
Structurally, in most cases GTs adopt one of two common folds, termed GT-A and
GT-B (Figure 10.7) [88].
Enzymes of the GT-A fold, that is, the major part of the mammalian GTs and
various bacterial enzymes, contain a Rossmann-type N-terminal domain, which
recognizes the nucleotide sugar donor, and a C-terminal domain consisting mainly of
mixed-b-sheets and dedicated to the recognition of the acceptor. Moreover, a general
feature of all the enzymes of the GT-A fold that utilize a nucleoside-diphospho-sugar
is the presence of a DxD motif and their requirement of a divalent cation for activity
(usually Mn2 þ or Mg2 þ ), both of which are involved in activated donor coordination
and in the catalytic mechanism of sugar transfer reaction [90]. The GT-B fold consists
instead of two similar Rossmann domains connected by a linker region and a catalytic
site located between the domains. This overall structure is highly conserved among
members of GT-B family, whereas there is no strong evidence to conserve specific
residues or to bind metal cations for catalysis. This subfamily includes the previously
described actinomycete GTs as well as various insect and plant GTs. It is remarkable
Figure 10.7 Ribbon diagrams of different GTs folds: (a) GT-A fold, bovine b-1,4-
galactosyltransferase (b4GalT I) catalytic domain complexed with UDP-Gal (PDB code 1FR8);
(b) GT-B fold, E. coli b-1,4-GlcNAc transferase (MurG) complexed with UDP-GlcNAc
(PDB code 1NLM).
432
that members of GT-A and GT-B subfamilies can be found in prokaryotes as well as in
eukaryotes and, according to sequence comparison of members of subfamilies, these
two groups of enzymes appear to be unrelated despite the similarities among domain
structures.
The three-dimensional structures of over 200 GTs are currently available at the
Protein Data Bank (PDB) database (www.rcsb.org/pdb), either free or bound to
substrates. A resume of prokaryotic and eukaryotic GTs with available crystal
structures is presented in Tables 10.1 and 10.2, respectively (data taken from the
Glyco3D database, accessible at the website www.cermav.cnrs.fr/glyco3d, whereas
references are limited to the original structure determination work).
10.3.2
Synthesis of Sugar Nucleoside Phosphates
GT-catalyzed oligosaccharide synthesis requires stoichiometric amounts of sugar

nucleotides as activated donor molecules and their availability can be critical for the
scaling-up of the reactions.
As large amounts of these compounds are still not commercially available, both
chemical and enzymatic synthetic strategies have been pursued in recent years. The
chemical synthesis of natural and unnatural nucleoside diphosphate sugars (NDP-
sugars), such as UDP-a-D-galactose and GDP-a-D-mannose, and nucleoside mono-
phosphate sugars (NMP-sugars), for instance CMP-sialic acid, has been thoroughly
reviewed recently [136]. Concerning NDP-sugars, most methods are based on
pyrophosphate bond formation as a result of the reaction between two monopho-
sphate precursors, that is, a nucleoside monophosphate protected with a suitable
leaving group and a glycosyl phosphate. The use of nucleoside phosphoramidates
and phosphoromorpholidates in the presence or in the absence of various catalysts
has been investigated as well. As an alternative, NDP-sugars have been synthesized
by glycosylation of nucleoside diphosphates, but, in this case, adequate control of the
stereoselectivity of the reaction is needed as a new anomeric linkage is formed.
Analogously, NMP-sugars can be synthesized by reaction of a nucleoside 50 -mono-
phosphate with a properly activated monosaccharide. Another approach relies on PIII
chemistry to form the new O–P–O interconnection, followed by subsequent oxida-
tion to achieve the final PV phosphate oxidation level.
Despite the recent progress in chemical methods, enzymatic synthesis of sugar
nucleotides is by far the most promising approach for GT-catalyzed large-scale
preparation of oligosaccharides. In fact, regeneration of sugar nucleotides can be
coupled in situ with the glycosylation reaction to allow not only a reduction of
substrates costs but also the removal of feedback inhibition effects by the generated
nucleoside phosphate towards GTs [137].
NDP-sugars can be regenerated by the subsequent action of a kinase and a
suitable sugar nucleotidyltransferase, also known as sugar pyrophosphorylase, at
the expenses of a phosphate donor, for example, phosphoenolpyruvate (PEP) in the
case of pyruvate kinase (Scheme 10.7a). Concerning CMP-sugars, such as, for
example, CMP-sialic acid, a two-step phosphorylation of CMP to CTP can be achieved
Table 10.1 Viral and prokaryotic GTs with available 3D-structures.
Organism Function Name Mechanism CAZy Fold PDB Reference

family structures
Virus
Bacteriophage T4 a-Glucosyltransferase AGT Retaining GT72 GT-B 5 [91]
b-Glucosyltransferase BGT Inverting GT63 GT-B 20 [92]
Prokaryotes
Agrobacterium tumefaciens Glycogen synthase 1 AtGS Retaining GT5 GT-B 2 [93]
Amycolatopsis orientalis b-Epi-vancosaminyltransferase GtfA Inverting GT1 GT-B 2 [94]
b-Glucosyltransferase GtfB Inverting GT1 GT-B 1 [95]
b-Vancosaminyltransferase GtfD Inverting GT1 GT-B 1 [96]
Aquifex aeolicus Peptidoglycan glycosyltransferase MrcA Inverting GT51 GT-A 1 [97]
Bacillus subtilis Putative glycosyltransferase SpsA Inverting GT2 GT-A 5 [85]
Bradyrhizobium sp. a-1,6-Fucosyltransferase NodZ Inverting GT23 GT-B 1 [98]
Campylobacter jejuni a-2,3-Sialyltransferase CstI Inverting GT42 GT-A like 2 [86]
a-2,3/2,8-Sialyltransferase CstII Inverting GT42 GT-A like 3 [99]
Clostridium difficile a-Glucosyltransferase Toxin B Retaining GT44 GT-A 2 [100]
Escherichia coli b-1,4-GlcNAc transferase MurG Inverting GT28 GT-B 2 [101]
Trehalose-6-phosphate synthase OtsA Retaining GT20 GT-B 3 [102]
Heptosyltransferase I WaaC Inverting GT9 GT-B 3 [103]
(Continued )
10.3 Glycosyltransferases
j433
434

family structures
Heptosyltransferase II RfaF Inverting GT9 GT-B 1 —

L-glycero-D-manno-heptose II-1, WaaG Retaining GT4 GT-B 2 [104]
3-glucosyltransferase I
Helicobacter pylori a-1,3-Fucosyltransferase FucT Inverting GT10 GT-B 3 [105]
Mycobacterium smegmatis Phosphatidylinositol PimA Retaining GT4 GT-B 2 [106]
mannosyltransferase
Neisseria meningitidis a-1,4-Galactosyltransferase LgtC Retaining GT8 GT-A 3 [107]
Pasteurella multocida a-2,3-Sialyltransferase PmST(1) Inverting GT80 GT-B 6 [108]
Pyrococcus abyssi Glycogen synthase PaGS Retaining GT5 GT-B 2 [109]
Rhodothermus marinus Mannosylglycerate synthase MGS Retaining GT78 GT-A 4 [110]
Staphylococcus aureus Penicillin-binding protein 2 PBP2 Inverting GT51 GT-A 2 [111]

Streptococcus pneumoniae Penicillin-binding protein 1b PBP1B Inverting GT51 GT-A 1 [112]
Streptomyces antibioticus Oleandomycin glycosyltransferase OleD/OleI Inverting GT1 GT-B 2 [113]
Streptomyces fradiae dTDP-D-Olivose-transferase UrdGT2 Inverting GT1 GT-B 1 [114]
Streptomyces viridochromogenes Eurekanate-attachment enzyme AviGT4 Retaining GT4 GT-B 2 [115]
Table 10.2 Eukaryotic GTs with available 3D-structures.

family structures
Bos taurus (bovine) a-1,3-Galactosyltransferase a3GalT Retaining GT6 GT-A 18 [116]

b-1,4-Galactosyltransferase I b4GalT I Inverting GT7 GT-A 21 [117]
Homo sapiens (human) b-1,4-Galactosyltransferase I b4GalT I Inverting GT7 GT-A 6 [118]
a-1,6-Fucosyltransferase Fut8 Inverting GT23 GT-B 1 [119]
Polypeptide GalNAc transferase GalNAc-T10 Retaining GT27 GT-A 2 [120]
b-1,3-Glucuronyltransferase GlcAT I Inverting GT43 GT-A 2 [122]
b-1,3-Glucuronyltransferase GlcAT P Inverting GT43 GT-A 3 [123]
a-1,3-GalNAc transferase A GTA Retaining GT6 GT-A 19 [124]
b-1,3-Gal transferase B GTB Retaining GT6 GT-A 19 [124]
UDP-glucuronosyltransferase Ugt2b7 Inverting GT1 GT-B 1 [125]
Medicago truncatula b-Glucosyltransferase UGT71G1 Inverting GT1 GT-B 2 [126]
Flavonoid glycosyltransferase UGT85H2 Inverting GT1 GT-B 1 [127]
Mus musculus (mouse) b-1,6-GlcNAc transferase C2GNT Inverting GT14 GT-A 2 [128]
a-1,4-N-Acetylhexosaminyl transferase Extl2 Retaining GT64 GT-A 4 [129]
O-Fucosylpeptide b-1,3-GlcNAc transferase Mfng Inverting GT31 GT-A 2 [131]
Oryctolagus cuniculus (rabbit) a-Glucosyltransferase Glycogenin Retaining GT8 GT-A 3 [132]
b-1,2-GlcNAc transferase I GnT I/Mgat 1 Inverting GT13 GT-A 7 [133]
Saccharomyces cerevisiae (yeast) a-1,2-Mannosyltransferase Kre2P/Mnt1P Retaining GT15 GT-A 3 [134]
Vitis vinifera Anthocyanidin-3-O-Glc transferase VvGT1 Inverting GT1 GT-B 3 [135]
10.3 Glycosyltransferases
j435
(a) Sugar2 Sugar1- Sugar2
(b)
Sugar2 Sugar1- Sugar2
E1 E1
CMP-Sugar1 CMP
NDP-Sugar1 NDP E4 ATP Pyr
Pi PPi E5
E4 E2
Pi PPi PEP E6
ADP PEP
E3 E2 Sugar1 CTP CDP
NTP
Sugar1-P E2
Pyr
Pyr PEP
Sugar2 Sugar1- Sugar2 Sugar2 Sugar1- Sugar2

(c) (d)
E1 E1
CMP-Sugar1 CMP
NDP-Sugar1 NDP E4
Pi PPi E9
E4
Pi PPi E6
ATP
E3 E7 Sugar1 CTP
CDP
NTP ADP polyPn-1
Sugar1-P E8 E8
polyPn-1 polyPn
polyPn
Legend:
E1, glycosyltransferase; E2, pyruvate kinase; E3, sugar nucleotidyltransferase; E4, pyrophosphatase; E5, myokinase;
E6, CMP-sugar synthetase; E7, nucleoside diphosphate kinase; E8, polyphosphate kinase; E9, cytidilate kinase
PEP, phosphoenol pyruvate; Pyr, pyruvate; polyP, inorganic polyphosphate
Scheme 10.7 Enzymatic regeneration of NDP-sugars (a and c) and CMP-sugars (b and d), using
phosphoenolpyruvate or inorganic polyphosphate as phosphate donors, respectively.
by subsequent use of myokinase and pyruvate kinase, in both cases with PEP
as a sacrificial cosubstrate, followed by the action of a CMP-sugar synthetase
(Scheme 10.7b). As PEP is still a quite expensive reagent to be used in stoichiometric
amounts, cheaper alternative systems have been suggested such as, for example, the
one using creatine phosphate/creatine kinase, which has been employed for regen-
eration of both NDP- and NMP-sugars with a remarkable cost reduction [138].
Different sets of enzymes using inorganic polyphosphate as a phosphate donor
have been coupled to GTs for oligosaccharides synthesis (Scheme 10.7c and d).
NDPs can be phosphorylated thanks to the concerted action of a nucleoside
diphosphate kinase (NDPK) and a polyphosphate kinase (PPK) through an ATP/
ADP cycle, then coupling to the sugar donor is achieved by action of a sugar
nucleotidyltransferase as in the previous approach [139]. Suitable NDPKs and
PPKs have been cloned from E. coli as well as from other sources. In recent
work, overexpressed enzymes for UDP-Gal regeneration have been purified
and co-immobilized on Ni2 þ -NTA agarose beads together with various galactosyl-
transferases, thus giving the so-called Superbeads, to be used for the synthesis of
different oligosaccharides even on the gram-scale [140]. Polyphosphate kinases have
been also exploited for regeneration of CMP-sugars. Synthesis of CMP-sialic acid
catalyzed by a CMP-NeuAc synthetase has been coupled with an enzymatic CTP-
generating system consisting of a PPK and a cytidilate kinase using CMP and
inorganic polyphosphate as substrates (Scheme 10.7d) [141]. Similar enzymatic
activities have also been used as active inclusion bodies in combination with whole
cells expressing sialic acid aldolase and CMP-sialic acid synthetase, yielding
30 -sialyllactose as final product [142].
As an alternative to isolated enzymes-coupled systems, bacterial coupling, that is,
the combined used of microorganism strains with different biochemical pathways
acting in a coordinate way, has been suggested. For instance, the in vitro production of
30 -sialyllactose has been achieved by using resting cells of a Corynebacterium
ammoniagenes strain, showing a CMP kinase activity and able to convert orotic acid
into UTP, with three metabolically engineered E. coli strains, showing CTP synthe-
tase, CMP-Neu5Ac synthetase, and a-2,3-sialyltransferase activities, respectively
(Scheme 10.8) [143].
E. coli
HO
OH NM522/p
-
COO TA23
E. coli NM522/pYP3
AcHN O OH
HO HO
E. coli Neu5Ac OH
OH
HO O
MM294/ O
HO
pMW6 O HO OH
CMP-Neu5Ac OH
OH
CTP
Lactose
O
OH
HO -
COO
NH
OH
HO O OH
UTP CDP CMP AcHN O O
N O O
H HO HO HO OH
O HO
O OH
3'-Sialyllactose OH
Orotic acid
C. ammoniagenes DN510
Scheme 10.8 CMP-Neu5Ac regeneration by bacterial coupling.
Further improvements in this field are expected to come from the recent finding of
new microbial sugar nucleotidyltransferases with unusual substrate specificity and
thermal stability. Recently, two new enzymes, a UDP-Glc pyrophosphorylase and a
phosphomannose isomerase/GDP-mannose pyrophosphorylase, have been isolated
and cloned from a Pyrococcus furiosus strain, both showing a very broad substrate
tolerance by accepting various NTPs and sugar phosphates as substrates, and
keeping, at the same time, the typical thermostability of archaeal enzymes
[144, 145]. Other bacterial enzymes, namely, the CMP-sialic acid synthetase from
Neisseria meningitides [146] and the thymidylyltransferase Cps2L from Streptococcus
438
pneumoniae [147, 148], have been exploited successfully for the regeneration of CMP-
and dTDP-sugars, respectively.
10.3.3
Substrate Specificity and Synthetic Applications
Being generally highly selective in vivo for given glycosyl donors and acceptors as well
as for the type (a or b) of the newly formed glycosidic linkage and its position (e.g.,
1 ! 3 or 1 ! 4, etc.), GTs have long been considered one of the best examples of
enzyme specificity developed by nature. Actually, the one-enzyme-one-linkage
paradigm, together with the controlled space distribution of GTs in cellular orga-
nelles such as the endoplasmic reticulum and Golgi apparatus, can still be considered
as the basis of oligosaccharide sequence fidelity for glycan chain synthesis in
eukaryotes. Contrary to the enzymes involved in nucleic acids replication, GTs do
not need a template to catalyze a reaction.
Mammalian GTs, for example, the bovine b-1,4-galactosyltransferase (b-1,4-GalT),
porcine a-1,3-galactosyltransferase (a-1,3-GalT), human a-1,3- and a-1,4-fucosyl-
transferases (a-1,3-FucT and a-1,4-FucT), and murine a-2,3- and a-2,6-sialyltrans-
ferases (a-2,3-SiaTand a-2,6-SiaT), were the first GTs whose substrate specificity was
extensively investigated [149]. This is due both to their natural availability, not only as
membrane proteins but also in soluble form from body fluids like blood and milk,
and to the interest in exploitation of these enzymes for the synthesis of various
bioactive oligosaccharides, for example, sialyl-Lewis X and sialyl-Lewis A tetrasac-
charides, by mimicking their original biosynthetic pathways [137, 150]. The b-1,4-
GalT from bovine milk is the enzyme for which most information on substrate
specificity and synthetic performance is available; it was also the first mammalian GT
with an available 3D structure [117], both free and in the presence of the specifier
protein a-lactalbumin (a-LA), in the so-called lactose synthase complex.
Concerning donor specificity, mammalian GTs usually show a strong, but not
absolute, preference for a defined UDP-sugar. For example, concerning bovine b-1,4-
GalT, it has been reported that transfer of Glc, GlcNAc, 2-deoxy-Glc, arabinose, and
GalNAc from the corresponding UDP-sugars is also possible, albeit with reduced
rates (0.3–5%) in comparison with Gal transfer [149, 151]. Moreover, it has been
shown that the donor specificity can significantly differ among members of the same
GT family when dealing with unnatural substrates. For instance, Elling and cow-
orkers showed that UDP-6-biotinyl-Gal was accepted as a activated donor by diverse
members of the GT7 family, that is, the human isoenzymes b-1,4-GalT1, b-1,4-GalT2,
b-1,4-GalT3, b-1,4-GalT4, and b-1,4-GalT5, but not by bovine b-1,4-GalT [152].
Analogously, differences have been observed between two a-1,3-GalT and an
a-1,4-GalT when using various UDP-deoxy-Gal analogues as donor substrates
(Table 10.3) [153], suggesting that GTs with different (or more relaxed) specificity
might be found or generated from existing enzymes by site-directed mutagenesis.
In this regard, in some cases it has been demonstrated already that this exquisite
specificity for donor substrates can be related to a few amino acid substitutions. For
example, the blood group transferases A (a-1,3-GalNAcT-A) and B (a-1,3-GalT-B)
Table 10.3 Relative rates of GalT-catalyzed galactosylation reactions with UDP-Gal analogues.
(a: b) Calf thymus Blood group Neisseria meningitides

(a-1,3-GalTa)) B (a-1,3-GalTb)) (a-1,4-GalTa))
UDP-Gal (Sigma) a 100 100 100

UDP-Gal (1 : 1) 105 65 78
UDP-2-deoxy-Gal (3 : 1) 342 173 28
UDP-3-deoxy-Gal (1 : 1) 0.20 0.12 0.0
UDP-4-deoxy-Gal (1 : 3) 0.61 0.21 1.8
UDP-6-deoxy-Gal (1 : 1) 1.53 18 42
UDP-L-Ara (4 : 7) 0.77 0.11 6.9
a)
Acceptor: tetramethyrhodamine (TMR)-labeled lactose.
Acceptor: TMR-labeled a-L-Fuc-1,2-b-D-Gal.
b)
transfer GalNAc and Gal, respectively, to a galactose moiety of the fucosylated LacNAc
acceptor (a-L-Fuc-1,2-b-D-Gal-OR, where R is glycolipid or glycoprotein). This quite
absolute specificity toward different donors and acceptors is due to the identity of just
four critical amino acids out a total of 354, that is, Arg/Gly176, Gly/Ser235, Leu/
Met266, and Gly/Ala268 in GTA and GTB, respectively. Specifically, the Leu/Met266
proved to be crucial in the A/B donor substrate specificity, so that a significant change
in the corresponding specificity constants could be observed by a single amino acidic
mutation [154]. These findings have been applied recently also to the design of novel
GTs with broader donor specificities to transfer sugar residues with a chemically
reactive handle, such as a keto or azido group, from the corresponding UDP-sugar
analogues [155, 156]. In the case of bovine a-1,3-GalT, mutants 280 SGG282 and
280
AGG282 with the highest GalNAcT activity (about 10–20% of the initial GalT
activity) have been exploited for transferring 2-keto-Gal or GalNAz (Gal-2-NH-CO-
CH2-N3) to LacNAc (Gal-b-1,4-GlcNAc) terminal moieties of glycoconjugates
(Scheme 10.9).
Regarding the specificity of Leloir GTs towards the acceptors, a broad tolerance in
the respect of non-natural substrates has been generally observed, which is possibly
related both to the observed conformational flexibility of loop regions surrounding
the acceptor binding site and its accessibility to the solvent [157].
In the case of bovine b-1,4-GalT, many possible acceptor modifications are
tolerated, giving preparatively useful yields (Figure 10.8), while the OH group at
C4 seems to be essential both for acceptor binding and catalysis. This requirement
seems to be not so stringent for other GTs. For instance, thiooligosaccharides
syntheses catalyzed by recombinant bovine a-1,3-GalT and by N. meningitides
b-1,3-GlcNAcT in the presence of 30 -thiolactose as an acceptor have been
described [158].
The highest variability is reported for the aglycon moiety, with almost any
derivative being accepted as long as it is linked in the b-conformation and
sufficiently soluble in the reaction mixture. This fact has opened up several synthetic
applications of bovine b-1,4-GalT for the preparation of oligosaccharides containing
OH
440
O
HO R = chitotriose or glycoprotein
HO OR
NHAc
GlcNAc-OR
UDP-Gal
β-1,4-GalT
UDP
OH OH
OH
O
O
HO O OR
HO
OH
NHAc
LacNAc-OR
UDP-2-keto-Gal UDP-GalNAz
280SGG282 (or 280AGG282)

UDP
α-1,3-GalT UDP
OH OH
OH OH
O
OH OH O
HO OH OH OH
O HO OH
H2C O O
O O HN O
OR O O OR
O HO HO
OH OH
NHAc O NHAc
N3
Scheme 10.9 Transfer of C2-modified galactose catalyzed by bovine a-1,3-GalT mutants.

Galβ,
Fucα,
Neu5Acα,
CO2CH3 S,CH2
OH
O
HO OR H,
HO oligosaccharide,
NHAc polymer,
H, peptide...
CH3CH(COOH)-O,
CH3COO, OH,
H,
CH2=CH-CH2-O epimer,
N-acyl
Figure 10.8 Acceptor modifications tolerated by bovine b-1,4-GalT.
the Gal-b-1,4-GlcNAc unit as well as galactosylated derivatives of glucosamides with

various b-linked aglycons [149]. In most cases, small-scale synthetic applications of
bovine b-1,4-GalT as well as investigations on its acceptor substrate specificity have
been carried out using a simplified three-enzymatic system with in situ formation of
UDP-Gal from the cheaper UDP-Glc by a UDP-glucose epimerase and UDP removal
catalyzed by an alkaline phosphatase (Scheme 10.10).
β-1,4-GalT
UDP-Gal + Acceptor Gal - Acceptor + UDP
(α-LA)
(organic cosolvent)
UDP-Glc alkaline
epimerase phosphatase
UDP-Glc Uridine + Pi
Scheme 10.10 Coupling of b-1,4-GalT-catalyzed galactosylation with UDP-Glc epimerase and

alkaline phosphatase activities.
Regarding the use of glucose derivatives as acceptors, it was quite recently observed
that the strict requirement of the modifier protein a-lactalbumin (a-LA) for the
lactose synthase activity (Km for Glc is 2 M in the absence of a-LA, but is reduced by
1000-fold if a-LA is present) does not occur in the galactosylation of complex
natural glucosides such as, for example, ginsenoside Rg1 (8, Scheme 10.11) or the
alkaloid colchicoside [151]. Conversely, a-LA showed a strong inhibitory effect in
the galactosylation of b-glucopyranosides with bulky hydrophobic aglycons, the
possible rationale being a competition between a-LA and the acceptor substrate
for the same hydrophobic binding site on b-1,4-GalT, as also suggested by crystal-
lographic analyses of the lactose synthase complex [159].
Another hurdle that has been overcome in recent years is related to the possibility
of using organic cosolvents in GTs-catalyzed reactions to improve acceptor substrates
solubility. In fact, in the case of b-1,4-GalT it was shown that some of the tested
442
OH
O
HO O
HO
HO OH
Ginsenoside Rg1
(8)
HO
OH
O O OH
HO
OH
Epimerase
UDP-Glc UDP-Gal
Alkaline Bovine β-1,4-GalT
phosphatase 10 % (v/v) DMSO
Uridine + Pi UDP
OH OH
OH
O
O O
HO O
HO HO
HO OH
Monogalactosylated
+ Ginsenoside Rg1 derivatives
HO
OH OH
O O O O
HO HO
HO HO OH
Scheme 10.11 Galactosylation of ginsenoside Rg1 (8) catalyzed by bovine b-1,4-GalT.
cosolvents were very well tolerated even at concentrations up to 20% (v/v)

(Table 10.4) [151]. These findings have been subsequently exploited for the prepa-
ration of selectively galactosylated derivatives of complex natural compounds with an
extremely low solubility in aqueous medium (e.g., the partially deglycosylated
derivatives of the saponin asiaticoside 2b and 2c, see Figure 10.3 [47]).
Large-scale synthesis of glyconjugates by GTs-catalyzed reactions has become
much more viable in recent years thanks also to the cloning and heterologous
expression of various enzymes belonging to this class [160]. High-level expression of
human b-1,4-GalT and mouse a-1,3-GalT in recombinant form has been achieved in
E. coli, while human a-1,3-FucT VI was recovered as inclusion body from Pichia
pastoris cells and refolded to the active enzyme.
Cloning of genes coding for GTs allows also their co-expression with other
enzymes needed for the complete biosynthetic pathway of oligosaccharides in a
unique engineered E. coli strain: the so-called superbugs or microbial cell
factories. In an interesting example, gram-scale synthesis of globotriose derivatives
Table 10.4 Effect of organic cosolvents [20% (v/v)] on b-1,4-GalT activity.
Cosolvent Relative initial rates
Acceptor: Glc ( þ a-LA) Acceptor: GlcNAc ( a-LA)
Blank 100 100

Ethanol 96 58
Methanol 117 103
Acetone 41 63
Acetonitrile 0 0
Tetrahydrofuran 0 0
N, N-Dimethylformamide 10 16
Dimethyl sulfoxide 57 38
Dioxane 12 91
N-Methyl pyrrolidone 0 2
has been achieved by co-expression of the a-1,4-GalT (LgtC) from N. meningitidis and
four enzymes for UDP-Gal regeneration, namely, galactokinase (GalK), galactose-1-
phosphate uridyltransferase (GalPUT), glucose-1-phosphate uridyltransferase
(GalU), and pyruvate kinase (PK) (Scheme 10.12) [161].
HO OH
O
HO OH OH HO
O O HO
HO O O OH OH
HO OH O O
OH OH HO O
HO OH
Lactose OH
LgtC OH
Globotriose
PEP ATP Gal
UDP
PK GalK UDP-Gal
Pyr ADP PEP
GalPUT
Gal-1-P PK
Glc-1-P
Pyr
UDP-Glc UTP
GalU
PPi
Scheme 10.12 Enzymatic synthesis of globotriose through an engineered E. coli strain expressing
the a-1,4-GalT (LgtC) from N. meningitidis and enzymes for UDP-Gal regeneration.
Moreover, permeabilized resting cells expressing the five enzymes showed similar
performances in the preparation of various non-natural globotriose derivatives when
compared with the corresponding purified biocatalysts (Table 10.5).
444
Table 10.5 Isolated yields of globotriose derivatives obtained by sugar acceptor galactosylation in
the presence of superbug whole cells or purified enzymes.
Acceptor Yield (%)
Whole cells Enzymes
HO OH OH 75 92
O O
HO O
HO OH
OH OH
HO OH OH 85 66
O O
HO O
HO OBn
OH OH
OH 60 77
O
HO OH OH
O HO
HO O
OH
50 84
HO OH OH
O O
HO O OMe
HO
OH OH
50 81
HO OH OH
O O
HO O SPh
HO
OH OH
HO OH 45 45
O
O O OH
HO
HO
OH OH
OH
HO OH OH 20 10
O OH
HO O
OH
HO
OH OH
HO OH 10 5
O
HO OMe
OH
Several examples of GTs-catalyzed oligosaccharide synthesis by metabolically

engineered bacterial strains have been reported. Successful co-expression of
N. meningitidis b-1,4-GalT (LgtB) and bovine a-1,3-GalT with the chitin oligosaccha-
ride synthase NodC in a E. coli strain allowed production of the Gal-a-1,3-Gal-b-1,4-
GlcNAc epitope, the major porcine antigen responsible for xenograft rejection [162].
This strategy has been applied also to the conversion of lactose into different human
milk oligosaccharides, such as lacto-N-tetraose and sialyllactose, using E. coli strains
overexpressing metabolic pathways with suitable bacterial GTs [163]. Specific sugar
permeases can be as well expressed by the recombinant strain to improve acceptor
up-take into the cells. For the synthesis of the oligosaccharide moiety of ganglioside
GM1a, that is, the cholera toxin receptor (Figure 10.1), expression of the LacY and
NanT permeases for internalization of lactose and sialyc acid, respectively, together
with genes coding for the needed GTs and enzymes for sugar nucleotides synthesis,
has allowed the in vivo production of the desired product in gram-quantities [164].
Other gangliosides oligosaccharides, for example, GM2 [GalNAc-b-1,4-(Neu5Ac-
a-1,3)-Gal-b-1,4-Glc], GM3 (Neu5Ac-a-1,3-Gal-b-1,4-Glc), and GD3 (Neu5Ac-
a-1,8-Neu5Ac-a-1,3-Gal-b-1,4-Glc), have been synthesized by engineered E. coli
strains using analogous strategies [165].
Synthetic applications arising from the observed tolerance of GTs towards non-
natural acceptors have been mostly related to the preparation of oligosaccharides
derivatives and other glycoconjugates as well as to the glycosylation of natural
compounds that were already glycosylated themselves. However, the applicability
of this class of enzymes might be further extended by exploiting unusual types of
glycosylation reactions or by using a substrate engineering approach, both
advances possibly coming from a deeper understanding of the recognition binding
sites of GTs for acceptor molecules.
In the first case, the possibility of extending a standard glycosylation reaction to
non-natural glycosidic bonds or even to non-saccharidic acceptor substrates has
been reported for the bovine b-1,4-GalT. It has been observed that key enzyme–sub-
strate interactions can be satisfied by preserving a correctly oriented hydroxyl group,
that is, the one subsequently involved in the formation of the new glycosidic bond,
and a N-acetyl group of the acceptor substrate for binding at the A-1 and N loci of the
protein, respectively, or, alternatively, the anomeric OH group for interaction with the
A-2 locus of the lactose synthase (b-1,4-GalT þ a-lactalbumin) complex (Figure 10.9).
(a) (b)
OH OH
O O
HO OR HO
HO HO O
H A-2
A-1 N A-1 OH
O
H
H3C
Acceptor
Acceptor
binding site
binding site
N N
β-1,4-GalT Lactose synthase

Figure 10.9 Key enzyme–substrate interactions between sugar acceptors and bovine b-1,4-GalT
(a) or the lactose synthase complex (b).
According to these minimal requirements, various novel b-1,4-GalT-catalyzed
reactions have been described. A b-1,1 transfer of galactose has been first described
using N-acetyl kanosamine (3-acetamido-3-deoxy-Glc) or N-acetyl gentosamine
(3-acetamido-3-deoxy-Xyl) as acceptors (Scheme 10.13a) [166], whereas subsequently
it was demonstrated that a L-sugar could be used as an acceptor by this enzyme,
leading to b-1,3-Gal transfer (Scheme 10.13b) [167]. Finally, it has been shown that
3-acetamido-1,2-propanediol, an acyclic compound, could be utilized as an acceptor
substrate by b-1,4-GalT as well (Scheme 10.13c); the catalyzed reaction is enantio-
selective with exclusive galactosylation of the (R)-isomer [168].
(a)
HO OH
1
O HO
X β-1,4-GalT O 1 X
HO X O OH O O
OH = HO HO 1 OH
3 UDP-Gal HO 3
HN 1 HN
3 HO OH HN
O O O
CH3 H3C H3C
Glc3NAc, X = CH(CH2OH)
Gal-β-1,1-Glc3NAc (or Xyl3NAc)
Xyl3NAc, X = CH2
(b)
OH HO OH OH
3 3
HO R β-1,4-GalT O
O R
HO 1
HO 1 O HO O
UDP-Gal 1
HN OH HN
O O
H3C H3C
L-Glc1NAc, R = CH2OH Gal-β-1,3-L-Glc1NAc (or L-Xyl1NAc)

L-Xyl1NAc, R = H
(c)
1 3 HO
2
OH
HO NHAc β-1,4-GalT O (R)
UDP-Gal
HO O NHAc + Unreacted
(S)-isomer
OH
OH HO H
rac-3-Acetamido-1,2-propanediol
Scheme 10.13 Novel b-1,4-GalT-catalyzed reactions: (a) and (b), synthesis of b-1,1- and
b-1,3-linked disaccharides, respectively; (c), synthesis of 3-O-b-D-galactopyranosyl-sn-glycerol.
Alternative productive binding modes of acceptor substrates are also exploited in

the substrate engineering approach. In this case, productive binding of potential
GTs substrates is enhanced by a readily removable functional group, for example,
an aromatic or an alkyl substituent, thus expanding the array of novel products that
can be produced by wild-type enzymes. A nice example of substrate engineering has
been recently reported for the a-1,4-GalT (LgtC) from N. meningitidis [169]. UDP-Gal
is usually transferred by this enzyme to disaccharides such as lactose with the
formation of an a-1,4 linkage (Scheme 10.14a). If free galactose is used as an acceptor
instead of lactose, a very low catalytic activity (about 1000 times lower kcat/Km values)
has been observed, with a consequent detrimental effect also on the reaction
specificity. However, p-nitrophenyl-b-galactoside (pNPbGal), but not the correspond-
ing a-galactoside, is a good acceptor substrate and gives rise to a single product
(Scheme 10.14b), presumably thanks to a strong interaction of the aromatic chain
with hydrophobic residues of the disaccharide binding site. Interestingly, the
presence of hydrophobic substituents has also been shown to influence enzyme
regioselectivity, with a-1,3 and a-1,2 linkages being synthesized by LgtC when using
(a)
HO OH α-1,4
O
HO OH OH HO
O O LgtC
HO O HO OH
HO O OH
OH
OH OH UDP-Gal O O
HO O
HO OH
Lactose OH OH
(b)
HO OH α-1,4
O
HO OH NO2 LgtC HO
O HO OH
O O NO2
HO UDP-Gal O
OH HO O
pNPβGal OH
(c)
α-1,3
OH
HO LgtC HO OH
O =
HO O HO O
HO ( )7 HO O ( )6 HO
O UDP-Gal OH
OH OH
HO O
HO O O
Alkyl glucoside ( )7
OH
(d)
BzO HO
OH OH LgtC HO OH
HO
O = OH O α-1,2
HO OH O HO
UDP-Gal
HO OBz
BzO HO
O
O
6-OBz-Mannose HO
HO OH
Scheme 10.14 Substrate engineering of a-1,4-GalT-catalyzed reactions.

448
an alkyl glucoside and 6-OBz-mannose as acceptors, respectively (Scheme 10.14c

and d). The basis of molecular recognition of engineered acceptor substrates has
been investigated recently also by crystallographic analyses on another Leloir GT, the
retaining bovine a-1,3-GalT [170].
This approach represents therefore an alternative strategy for expanding the
synthetic applications of GTs. Moreover, further developments in defining GTs
specificity in the presence of natural and non-natural substrates are expected to
come from exploitation of high-throughput tools, such as glycoarrays, as recently
reviewed elsewhere [171].
10.3.4
New Glycosyltransferases from Microbial Sources
Thanks also to the extensive genome sequencing carried out in recent years, the
occurrence of wide array of genes coding for GTs in the microbial world has been
clearly demonstrated and novel enzymes, both from bacteria and yeasts, have been
both isolated and characterized.
As anticipated in Section 10.3.1, GTs from microbial sources can be divided into
two main groups according to their biological role. A large group of GTs has been in
fact identified from pathogenic microorganisms, both bacteria and yeasts, where they
can provide an improved virulence by synthesizing surface lipooligosaccharides
(LOS) or lipopolysaccharides (LPS) that mimic the host self-expressed carbohydrate
structures and help the pathogens to evade immune targeting. The other group of
microbial enzymes consists of GTs involved in natural products glycosylation and is
produced mainly by actinomycetes. Such enzymes are often denominated
antibiotic GTs, although recent examples suggest that their substrate specificity
is presumably not restricted to this class of compounds.
Table 10.6 presents an overview of GTs isolated from pathogenic microorganisms.
A wide array of different glycosidic bonds can be synthesized by these enzymes,
which, besides retaining the typical high degree of regio- and stereoselectivity of
mammalian GTs, are usually efficiently expressed in recombinant form in E. coli as
soluble proteins. The heterologous expression of synthetic genes coding for these
enzymes, overcoming the need to work directly with pathogenic microorganisms,
has made bacterial GTs available for numerous synthetic applications. Moreover,
these enzymes usually show good stability even in the presence of organic cosolvents.
In a recent example, three different GTs (an a-2,3-Neu5AcT, a b-1,4-GalNAcT, and a
b-1,3-GalT), all originated from the same microorganism, Campylobacter jejuni, were
successfully expressed in E. coli and subsequently used for the synthesis of gang-
liosides mimics (Figure 10.10) [196]. Owing to the presence of long alkyl chains on
acceptor glycosides, GTs-catalyzed reactions have been performed with preparatively
useful yields in the presence of suitable amounts of methanol, which – in all cases –
has been well tolerated.
In addition to the observed high-level expression of bacterial GTs in recombinant
form, preliminary investigations suggest that these enzymes might also show more
pronounced substrate promiscuity than their mammalian counterparts. For exam-
Table 10.6 New GTs isolated from microbial pathogens.
Organism Activity Reference
Azorhizobium caulinodans b-1,4-GlcNAcT [172]

Campylobacter jejuni a-1,4-GalT, b-1,3-GalNAcT, b-1,3-GalT, b-1, [173–176]
4-GalNAcT, a-2,3-Neu5AcT, a-2,3/8-Neu5AcT,
Candida albicans a-1,2-ManT, a-1,6-ManT [177, 178]
Cryptococcus neoformans a-1,3-ManT [179]
Haemophilus ducreyi b-1,3-GlcNAcT [180]
Haemophilus influenzae b-1,2-GalT, b-1,3-GalNAcT [181, 182]
Helicobacter pylori b-1,4-GalT, a-1,2-FucT, a-1,3-FucT, a-1,4-FucT, [143, 183–187]
a-1,3/4-FucT, a-1,6-GlcT, b-1,3-GlcNAcT
Neisseria gonorrhoeae a-1,2-GlcNAcT [188]
Neisseria meningitidis a-1,4-GalT, b-1,4-/a-1,3-GalT [107, 189]
Pasteurella multocida a-2,3/6-Neu5AcT, a-1,3-GalNAcT [174, 190]
Photobacterium sp. a-2,3-Neu5AcT, a-2,6-Neu5AcT [191]
Pseudomonas aeruginosa a-1,3-RhaT [192]
Salmonella enterica a-1,3-GalT [193]
Sphingomonas paucimobilis b-1,4-GlcUAT [194]
Streptococcus thermophilus a-1,6-GalT [195]
ple, the b-1,4-GalT from Helicobacter pylori catalyzes the quantitative synthesis of the
thiosaccharide Gal-b-S-1,4-GlcNAc-pNP as well as Gal-b-1,4-Man-pNP
(Scheme 10.15) [197]. Therefore, it might be foreseen that further investigations
on the synthetic potential of GTs from bacterial pathogens will broaden their
application as efficient and versatile biocatalysts even for the formation of non-
natural products.
OH OH OH OH
O H. pylori β-1,4-GalT
HS O O O
HO S O
UDP-Gal HO HO
AcHN HO AcHN
NO2
NO2
pNP-4S-β-GlcNAc Gal-β-S-1,4-GlcNAc-pNP
HO HO OH OH HO
O HO
HO H. pylori β-1,4-GalT O O
HO O
O O
HO HO
UDP-Gal HO
NO2
NO2
pNP-β-Man Gal-β-1,4-Man-pNP
Scheme 10.15 Substrate promiscuity of the b-1,4-GalT from Helicobacter pylori.
Actinomyces GTs have been investigated in depth in recent years and insights
concerning their structural and functional properties as well as their possible
450
HO OH OH
O HO R R= CH CH2
HO O O ( )8
O or C CH
OH OH
or CH2CH2N3
C. jejuni α-2,3-Neu5AcT
25 % (v/v) MeOH
HO
OH HO OH
OH OH
HOOC O HO R
O O O ( )8
O
AcHN O OH OH
HO GM3 mimic
C. jejuni β-1,4-GalNAcT
10 % (v/v) MeOH
HO OH
O
HO O
OH OH
NHAc O HO R
O O ( )8
HOOC O O
OH OH OH
OH
HO O
GM2 mimic
OH
AcHN
C. jejuni β-1,3-GalT
aqueous buffer
HO OH HO OH
O O
HO O O
OH OH
OH NHAc O HO R
O O ( )8
HOOC O O
HO OH OH
OH
HO O
GM1a mimic
OH
AcHN
Figure 10.10 Exploitation of GTs from bacterial pathogens for the synthesis of ganglioside GM1a
mimics.
applications in the modification of glycosylated natural compounds have been largely

reviewed [12, 13, 83, 87].
In many cases, these enzymes show a remarkable degree of promiscuity
toward both the glycosyl donor (usually a deoxygenated sugar derivative, see
Scheme 10.6) and acceptor. For instance, a surprisingly significant flexibility toward
(a)
H Me O
HO N Me H
N O N
VinC
H
O OH O
Me O
Me
N
H
Vicenilactam OH O-dTDP
Vicenistatin
dTDP-Vicenisamine
(b)
OH
OH
OH O
O H
H
O H H
HO H H
OH O
HO
H β-Estradiol 3-O-acetyl-β-estradiol
β-Zearalenol
OH H
O H N
O
O HO NHMe
OH
O
H
HO
Brefeldin A
Figure 10.11 (a) Glycosylation of vicenilactam catalyzed by the Streptomyces halstedii VinC GT;
(b) VinC-tolerated non-natural acceptors.
glycosyl acceptors has been demonstrated for the glycosyltransferase VinC from
Streptomyces halstedii HC-34, which can catalyze the transfer of the deoxy sugar
vicenisamine not only to the natural substrate vicenilactam (Figure 10.11a), but also
to a wide array of diverse hydrophobic aglycons (see Figure 10.11b for some
representative structures; hydroxyl groups involved in the formation of new glyco-
sidic bonds are circled) [198].
Some antibiotic GTs that catalyze C- or N-glycosylation reactions in vivo, instead
of the more common O-glycosylation, have been characterized. UrdGT2, a glycosyl-
transferase produced by a S. fradiae strain and involved in the synthesis of the
antitumor drug urdamycin, catalyzes the transfer of D-olivose (2,6-dideoxy-D-glucose)
from its NDP-donor to an aromatic polyketide acceptor to yield the b-C-glycoside
aquayamycin [199], while another Streptomyces GT, the enzyme StaG, catalyzes
the first N-glycosylation step in the synthesis of the alkaloid antibiotic staurosporine,
452
H
N O
O CH3
C-C bond
O OH
HO
N N
H3C O
O OH H3C H
HO
HO C-N bond
OH O MeO
NHCH3
Aquayamycin
Staurosporine
Figure 10.12 C- and N-glycoside formation catalyzed by actinomycete GTs.
the second C–N linkage being formed by action of the P450 oxidase StaN
(Figure 10.12) [200].
The promiscuity of actinomycete GTs toward donors and acceptors can be
significantly broadened by protein engineering approaches. The substrate specificity
of the oleandomycin GT OleD in respect of both donor sugar and aglycon acceptor
has been altered by means of directed evolution experiments, providing a triple
mutant variant with the ability to glucosylate diverse drug-like scaffolds, such as
steroids, b-lactams, alkaloids, macrolides, flavonoids, anthraquinones, and poly-
enes [201]. Moreover, this engineered GTcan also form S- and N-glycosidic bonds, as
shown by the isolation of glucosylation products obtained using UDP-Glc as donor
and thiophenol or aniline as acceptor substrates, respectively [202]. As an alternative
to directed evolution, novel chimeric GTs variants can be obtained by domain
swapping as it has been shown that enzymes belonging to the GT-B superfamily
are composed of N-terminal and C-terminal domains that contain the acceptor and
NDP-sugar donor sites, respectively [203, 204].
Synthetic applications of actinomycete GTs are mostly related to the generation of
biologically active compounds with non-natural glycosylation patterns. This goal
has been pursued by two different strategies called combinatorial biosynthesis and
glycorandomization.
Combinatorial biosynthesis consists of in vivo pathway engineering of both sugar
donor synthesis and glycosyl transfer reactions with addition of other suitable
enzymes or complete pathways for the generation of the desired products. Because
bacterial genes involved in NDP-sugar synthesis are usually clustered in operons that
can be easily transferred, and because bacterial GTs are naturally promiscuous, this
metabolic engineering approach has been successfully exploited for the generation of
new natural products derivatives. For example, a library of more than 30 different
indolocarbazole compounds has been prepared by co-expressing different combina-
tions of genes isolated from rebeccamycin- and staurosporine-producing microor-
ganisms as well as genes coding for other modifying enzymes, for example,
halogenases [205].
The glycorandomization strategy involves two distinct steps, that is, the generation
of a library of NDP-sugars by either chemical synthesis or nucleotidyltransferase-
catalyzed reactions and the coupling of the donors to the aglycon by means of GTs
with relaxed substrate specificity. In one example, this process has been successfully
applied to the preparation of a library of 21 known and 39 novel vancomycin
derivatives by dTDP-sugar pool generation, GtfE-catalyzed glycosylation of the
vancomycin aglycon, and further chemical modification of the carbohydrates
moieties [206].
10.3.5
Non-Leloir Glycosyltransferases
The denomination of non-Leloir GTs has been traditionally used to group those
enzymes that catalyze the transfer of sugar units without using nucleotide sugars as
donors. This has led to the creation of a quite heterogeneous family of enzymes that
show similarities from an applicative point of view, but have less related structural
and functional properties.
Two main subgroups of non-Leloir GTs can be distinguished on the basis of the
sugar donor. They can be activated as a sugar phosphate in the case of the so-called
phosphorylases, or non-activated for those enzymes using sucrose or starch-derived
oligosaccharides as donors.
From a structural and mechanistic point of view, phosphorylases belonging to
different classes can be divided into two distinct groups, the glycosidase-like, for
example, sucrose and maltose phosphorylases, and the transferase-like, for exam-
ple, glycogen and starch phosphorylases [207].
Most of the enzymes described up to now use glucose-1-phosphate as glycosyl
donor and release inorganic phosphate after the formation of the new glycosidic
bond. Only few examples of phosphorylases that accept Gal-1-P or GlcNAc-1-P as
donors have being reported as well. Depending on the stereochemical outcome of the
catalyzed reaction, phosphorylases can be further classified as inverting or retaining
enzymes. In both cases, a high specificity towards the donor substrate is usually
observed, whereas a much more relaxed behavior is described for the acceptors, so
that different glucosylated derivatives can be prepared using the same phosphorylase.
In a recent example, sucrose phosphorylase, an enzyme whose physiological reaction
is glucosyl transfer from sucrose to phosphate with formation of a-D-glucopyranosyl
phosphate (phosphorolysis), has been used in reverse transglucosylation reactions
for the regioselective preparation of glucosyl glycerol [208] and glucosyl glyce-
rates [209] (Figure 10.13).
Other enzymes using sucrose as a non-activated donor have been evolved in the
selective transfer of either the glucose or fructose moiety and, accordingly, are divided
in two distinct groups, that is, glucosyltransferases and fructosyltransferases [210].
Glucosyltransferases, also named glucansucrases, are extremely selective for
sucrose as a donor; sucrose analogs carrying different sugars, for example, galactose
or mannose, are usually not utilized for the transfer reaction. Nevertheless, thanks to
the availability of such a cheap donor, these enzymes have some technological
454
OH OH
OH
O O
O HO HO
HO HO
HO HO O
OH O
HO O HO O
HO O
O O
HO HO HO
Glucosyl glycerol Glucosyl (R)-glycerate Glucosyl (S)-glycerate
Figure 10.13 Different glucosides prepared using a sucrose phosphorylase.
applications for the preparation of glucosylated oligosaccharides of interest in the

food industry as probiotic compounds [211]. Fructosyltransferases catalyze the
synthesis of fructose-based oligosaccharides (fructans) with release of glucose. In
agreement with the proposed reaction mechanism, which hypothesizes a fructosyl-
enzyme intermediate, these enzymes show a less specific recognition of the glucose
unit and accept some sucrose analogs, for example, the galactosyl fructoside, as donor
substrates [212].
Finally, starch-derived di- and oligosaccharides can be used as either donors and
acceptors by cyclodextrin glucanotransferases (CGTases) for the synthesis of glucose-
based a-1,4-linked cyclic oligosaccharides (a-, b-, or c-cyclodextrins, consisting of 6,
7, or 8 glucose units, respectively) [213]. As revealed by sequence comparison of
CGTases from different sources, diversification of cyclodextrin product specificity
arises from specific incorporation and/or substitution of amino acids at the substrate
binding sites [214]. Besides cyclodextrin preparation, these enzymes have been
recently applied to the preparation of glycoconjugate derivatives of some natural
compounds, such as stevioside [215] and taxol [216].
References
1 Hecht, M.-L., Stallforth, P., 5 Kallenius, G., Pawlowski, A.,

Varon Silva, D., Adibekian, A., and Hamasur, B., and Svenson, S.B. (2008)
Seeberger, P.H. (2009) Recent advances Mycobacterial glycoconjugates as vaccine
in carbohydrate-based vaccines. Curr. candidates against tuberculosis. Trends
Opin. Chem. Biol., 13, 354–359. Microbiol., 16, 456–462.
2 Schriver, Z., Raman, R., Viswanathan, K., 6 Lau, K.S. and Dennis, J.W. (2008)
and Sasisekharan, R. (2009) Context- N-Glycans in cancer progression.
specific target definition in influenza Glycobiology, 18, 750–760.
A virus hemagglutinin-glycan receptor 7 Buskas, T., Thompson, P., and
interactions. Chem. Biol., 16, 803–814. Boons, G.J. (2009) Immunotherapy for
3 Kobayashi, M., Lee, H., Nakayama, J., cancer: synthetic carbohydrate-based
and Fukuda, M. (2009) Carbohydrate- vaccine. Chem. Commun., 5335–4349.
dependent defense mechanism against 8 Cipolla, L., Peri, F., and Airoldi, C. (2008)
Helicobacter pylori infection. Curr. Drug Glycoconjugates in cancer therapy.
Metab., 10, 29–40. Anti-Cancer Agents Med. Chem., 8 (1),
4 Vasta, G.R. (2009) Roles of galectins in 92–121.
infection. Nat. Rev. Microbiol., 7, 9 Sperandio, M., Gleissner, C.A., and
424–438. Ley, K. (2009) Glycosylation in immune
References j455
cell trafficking. Immunol. Rev., 230, enzymes in food processing. Crit. Rev.
97–113. Food Sci., 46, 197–205.
10 Diekman, A.B. (2003) Glycoconjugates in 22 Gupta, R., Gigras, P., Mohapatra, H.,
sperm function and gamete interactions: Goswami, V.K., and Chauhan, B. (2003)
How much sugar does it take to sweet- Microbial a-amylases: a biotechnological
talk the egg? Cell. Mol. Life Sci., 60, perspective. Process Biochem., 38,
298–308. 1599–1616.
11 Murrey, H.E. and Hsieh-Wilson, L.C. 23 Wilson, D.B. (2009) Cellulases and
(2008) The chemical neurobiology of biofuels. Curr. Opin. Biotechnol., 20,
carbohydrates. Chem. Rev., 108, 295–299.
1708–1731. 24 Wackett, L.P. (2008) Biomass to fuels via
12 Griffith, B.R., Langenhan, J.M., and microbial transformations. Curr. Opin.
Thorson, J.S. (2005) Sweetening Chem. Biol., 12, 187–193.
natural products via glycorandomization. 25 Hess, M. (2008) Thermoacidophilic
Curr. Opin. Biotechnol., 16, 622–630. proteins for biofuel production. Trends
13
Kren, V. and Rezanka, T. (2008) Sweet Microbiol., 16, 414–419.
antibiotics – the role of glycosidic 26 Murata, T. and Usui, T. (2006) Enzymatic
residues in antibiotic and antitumor synthesis of oligosaccharides and
activity and their randomization. FEMS neoglycoconjugates. Biosci. Biotechnol.
Microbiol. Rev., 32, 858–889. Biochem., 70, 1049–1059.
14 Stallforth, P., Lepenies, B., Adibekian, A., 27 Bojarova, P. and Kren, V. (2009)
and Seeberger, P.H. (2009) Glycosidases: a key to tailored
Carbohydrates: a frontier in medicinal carbohydrates. Trends Biotechnol., 27,
chemistry. J. Med. Chem., 52, 5561–5577. 199–209.
15 Ernst, B. and Magnani, J.L. (2009) From 28 Rye, C.S. and Whiters, S.G. (2000)
carbohydrate leads to glycomimetic Glycosidase mechanisms. Curr. Opin.
drugs. Nat. Rev. Drug. Discov., 8, 661–677. Chem. Biol., 4, 573–580.
16 Seeberger, P.H. and Werz, D.B. (2007) 29 Vasella, A., Davis, G.J., and B€ohm, M.
Synthesis and medical applications of (2002) Glycosidase mechanisms. Curr.
oligosaccharides. Nature, 446, Opin. Chem. Biol., 6, 619–629.
1046–1051. 30 Sinnott, M.L. (1990) Catalytic
17 Kovensky, J. (2009) Sulfated mechanisms of enzymic glycosyl
oligosaccharides: new targets for drug transfer. Chem. Rev., 90, 1171–1202.
development? Curr. Med. Chem., 16, 31 Whiters, S.G. and Street, I.P.
2338–2344. (1988) Identification of a covalent
18 Bernardes, G.J.L., Castagner, B., and alpha-D-glucopyranosyl enzyme
Seeberger, P.H. (2009) Combined intermediate formed on a beta-
approaches to the synthesis and study glucosidase. J. Am. Chem. Soc., 110,
of glycoproteins. ACS Chem. Biol., 4, 8551–8553.
703–713. 32 Whiters, S.G., Warren, R.A., Street, I.P.,
19 Cantarel, B.L., Coutinho, P.M., Rupitz, K., Kempton, J.B., and
Rancurel, C., Bernard, T., Lombard, V., Aebersol, R. (1990) Unequivocal
and Henrissat, B. (2009) The demonstration of the involvement
Carbohydrate-Active EnZyme database of a glutamate residue as a nucleophile in
(CAZy): an expert resource for the mechanism of a retaining
glycogenomics. Nucleic Acid Res., glycosidase. J. Am. Chem. Soc., 112,
37 (Database issue), D233–D238. 5887–5889.
20 Aehle, W. (ed.) (2007) Enzymes in 33 Defaye, J., Driguez, H., Henrissat, B.,
Industry, Wiley-VCH Verlag GmbH, and Bar-Guilloux, E. (1983)
Weinheim. Stereochemistry of the hydrolysis of
21 Synowiecki, J., Grzybowska, B., and a,a-trehalose by trehalase, determined by
Zdzieblo, A. (2006) Sources, properties
/
using a labelled substrate. Carbohydr.
and suitability of new thermostable Res., 124, 265–273.
456
34 Hehre, E.J., Brewer, C.F., and 44 Juarez Ruiz, J.M., Oswald, G.,
Genghof, D.S. (1979) Scope and Petersen, M., and Fessner, W.D. (2001)
mechanism of carbohydrase action - The natural strategy for the glycosidase-
Hydrolytic and non-hydrolytic actions assisted synthesis of simple glycosides.
of beta-amylase on alpha-maltosyl and J. Mol. Catal. B-Enzym., 11, 189–197.
beta-maltosyl fluoride. J. Biol. Chem., 254, 45 Speranza, G., Monti, D., Crippa, S.,
5942–5950. Cairoli, P., Morelli, C.F., and Manitto, P.
35 Gerber-Lemaire, S. and Juillerat- (2006) Kenyaloside, a novel O,O,O-
Jeanneret, L. (2006) Glycosylation triglycosylated naphthalene derivative
pathways as drug targets for cancer: from the exudate of Kenya Aloe species.
glycosidases inhibitors. Mini Rev. Med. Nat. Prod. Commun., 1, 1085–1088.
Chem., 6, 1043–1052. 46 Monti, D., Pisvejcova, A., Kren, V.,
36 Yuzwa, S.A., Macauley, M.S., Lama, M., and Riva, S. (2004) Generation
Heinonen, J.E., Shan, X., Dennis, R.J., of an a-L-rhamnosidase library and its
He, Y., Whitworth, G.E., Stubbs, K.A., application for the selective
McEachern, E.J., Davies, G.J., and derhamnosylation of natural products.
Vocadlo, D.J. (2008) A potent Biotechnol. Bioeng., 87, 763–771.
mechanism-inspired O-GlcNAcase 47 Monti, D., Candido, A., Cruz Silva, M.M.,
inhibitor that blocks phosphorylation Kren, V., Riva, S., and Danieli, B. (2005)
of tau in vivo. Nat. Chem. Biol., 4, 483–490. Biocatalyzed generation of molecular
37 Asano, N. (2003) Glycosidase inhibitors: diversity: selective modification of the
update and perspectives on practical use. saponine asiaticoside. Adv. Synth. Catal.,
Glycobiology, 13, 93R–104R. 347, 1168–1174.
38 Asano, N., Nash, R.J., Molyneux, R.J., and 48 Kieboom, A.P.G. (1988) Enzymes that do
Fleet, G.W.J. (2000) Sugar-mimic not work in organic-solvents - too polar
glycosidase inhibitors: natural substrates give too tight enzyme-product
occurrence, biological activity and complexes. Rec. Trav. Chim. Pays-Bas,
prospects for therapeutic application. 107, 347–348.
Tetrahedron Asymmetry, 11, 1645–1680. 49 Vic, G., Thomas, D., and Crout, D.H.G.
39 Lillelund, V.H., Jensen, H.H., Liang, X.F., (1997) Solvent effect on enzyme-
and Bols, M. (2002) Recent developments catalyzed synthesis of b-D-glucosides
of transition-state analogue glycosidase using the reverse hydrolysis method:
inhibitors of non-natural product origin. application to the preparative-scale
Chem. Rev., 102, 515–553. synthesis of 2-hydroxybenzyl and octyl
40 Asano, N. (2000) Alkaloidal sugar b-D-glucopyranosides. Enz. Microb.
mimetics: biological activities and Technol., 20, 597–603.
therapeutic applications. J. Enzym. Inhib., 50 Vic, G. and Crout, D.H.G. (1995)
15, 215–234. Synthesis of allyl and benzyl
41 Buchholz, K.B. and Poulsen, P.B. (2000) b-D-glucopyranosides, and allyl
Introduction, in Applied Biocatalysis, 2nd b-D-galactopyranoside from D-glucose
edn (eds A.J. Straathof and or D-galactose and the corresponding
P. Adlercreutz), Harwood Academic alcohol using almond b-D-glucosidase.
Publisher, Amsterdam, pp. 1–15. Carbohydr. Res., 279, 315–319.
42 Brooks, S.A. (2009) Strategies for analysis 51 Vic, G. and Crout, D.H.G. (1994)
of the glycosylation of proteins: current Synthesis of glucosidic derivatives with a
status and future perspectives. Mol. spacer arm by reverse hydrolysis using
Biotechnol., 43, 76–88. almond b-D-glucosidase. Tetrahedron
43 Dahmen, J., Trejd, T., Lave, F., Lindh, F., Asymmetry, 5, 2513–2516.
Magnusson, G., Noori, G., and 52 Joahnsonn, E., Hedbys, L., and
Palsson, K. (1983) 2-Bromoethyl Larsson, P.O. (1991) Enzymatic-
glycosides. 3. Applications in the synthesis of monosaccharide
synthesis of spacer-arm glycosides. amino-acid conjugates. Enzyme Microb.
Carbohydr. Res., 113, 219–224. Technol., 13, 781–787.
References j457
53 Casali, M., Tarantini, L., Riva, S., 63 Nilsson, K.G.I., Ljunger, G., and
Hunkowa, Z., Weignerova, L., and Melin, M. (1997) Glycosidase-catalysed
Kren, V. (2002) Exploitation of a library synthesis of glycosylated amino acids:
of a-galactosidases for the synthesis of synthesis of GalNAca-Ser and
building blocks for glycopolymers. GlcNAcb-Ser derivatives. Biotechnol.
Biotechnol. Bioeng., 77, 105–110. Lett., 19, 889–892.
54 De Roode, B.M., Franssen, M.C.R., 64 Suzuki, K., Fujimoto, H., Ito, Y.,
van der Padt, A., and Boom, R.M. (2003) Sasaki, T., and Ajisara, K. (1997) An
Perspectives for the industrial enzymatic efficient synthesis of Galb1-3GalNAc-
production of glycosides. Biotechnol. serine derivative using b-galactosidase.
Prog., 19, 1391–1402. Tetrahedron Lett., 38, 1211–1214.
55 Mitsuo, N., Takeichi, H., and Satoh, T. 65 Li, B., Song, H., Hauser, S., and Wang,
(1984) Synthesis of b-alkyl glucosides by L.-X. (2006) A highly efficient
enzymatic transglycosylation. Chem. chemoenzymatic approach toward
Pharm. Bull., 32, 1183–1187. glycoprotein synthesis. Org. Lett., 8,
56 Nillsson, K.G.I. (1988) A simple strategy 3081–3084.
for changing the regioselectivity of 66 Nilsson, K.G.I. (1987) A simple strategy
glycosidase-catalysed formation of for changing the regioselectivity of
disaccharides: part II, enzymic synthesis glycosidase-catalysed formation of
in situ of various acceptor glycosides. disaccharides. Carbohydr. Res., 167,
Carbohydr. Res., 180, 53–59. 95–103.
57 Trincone, A., Nicolaus, B., Lama, L., and 67 Weignerova, L., Sedmera, P.,
Gambacorta, A. (1991) Stereochemical Hunkova, Z., Halada, P., Kren, V.,
studies of enzymatic transglycosylation Casali, M., and Riva, S. (1999) Enzymatic
using Sulfolobus solfataricus. J. Chem. Soc., synthesis of iso-globotriose from partially
Perkin Trans. 1, 2841–2844. protected lactose. Tetrahedron Lett., 40,
58 Gais, H.J., Zeissler, A., and Maidonis, P. 9297–9299.
(1988) Diastereoselective D- 68 Hedbys, L., Johansson, E., Mosbach, K.,
galactopyranosyl transfer to meso diols Larsson, P., Gunnarsson, A.,
catalyzed by b-galactosidases. Tetrahedron Svensson, S., and L€onn, A. (1989)
Lett., 29, 5743–5744. Synthesis of Galb1-3GlcNAc and
59 Scheckermann, C., Wagner, F., and Galb1-3GlcNAcb-SEt by an enzymatic
Fischer, L. (1997) Galactosylation of method comprising the sequential use of
antibiotics using the beta-galactosidase b-galactosidases from bovine testes and
from Aspergillus oryzae. Enzym. Microb. Escherichia coli. Glyconjugate J., 6,
Technol., 20, 629–634. 161–168.
60 Scigelova, M., Sedmera, P., Havliceck, V., 69 Weignerova, L., Hunkova, Z., Kuzma, M.,
Prikrylova, V., and Kren, V. (1998) and Kren, V. (2001) Enzymatic synthesis
Glycosidase-catalyzed synthesis of ergot of three pNP-a-galactobiopyranosides:
alkaloids a-glycosides. J. Carbohydr. application of the library of fungal
Chem., 17, 981–986. a-galactosidases. J. Mol. Catal. B: Enzym.,
61 Ooi, T., Hashimoto, T., Mitsuo, N., 11, 219–224.
and Satoh, T. (1984) Enzymatic 70 Wang, Q., Graham, R.W., Trimbur, D.,
synthesis of chemically unstable Warren, R.A.J., and Whiters, S.G. (1994)
cardiac glycosides by b-galactosidase Changing enzymatic reaction
from Aspergillus oryzae. Tetrahedron Lett., mechanisms by mutagenesis: conversion
25, 2241–2244. of a retaining glucosidase to an inverting
62 Baker, A., Turner, N.J., and enzyme. J. Am. Chem. Soc., 116,
Webberley, M.C. (1994) An improved 11594–11595.
strategy for the stereoselective synthesis 71 Hancock, S.M., Vaughan, M., and
of glycosides using glycosidases as Withers, S.G. (2006) Engineering of
catalysts. Tetrahedron Asymmetry, 5, glycosidases and glycosyltransferases.
2517–2522. Curr. Opin. Chem. Biol., 10, 509–519.
458
72 Perugino, G., Cobucci-Ponzano, B., double mutant glycosidases that serve as

Rossi, M., and Moracci, M. (2005) Recent scaffolds for thioglycoside synthesis.
advances in the oligosaccharide synthesis Chem. Commun., 274–275.
promoted by catalytically engineered 81 Drone, J., Feng, H.Y., Tellier, C.,
glycosidases. Adv. Synth. Catal., 347, Hoffmann, L., Tran, V., Rabiller, C., and
941–950. Dion, M. (2005) Thermus thermophilus
73 Moracci, M., Trincone, A., Perugino, G., glycosynthases for the efficient synthesis
Ciaramella, M., and Rossi, M. (1998) of galactosyl and glucosyl b-(1-3)-
Restoration of the activity of active-site glycosides. Eur. J. Org. Chem., 10,
mutants of the hyperthermophilic 1977–1983.
beta-glycosidase from Sulfolobus 82 Weijers, C.A.G.M., Franssen, M.C.R.,
solfataricus: Dependence of the and Visser, G.M. (2008)
mechanism on the action of external Glycosyltransferase-catalyzed synthesis
nucleophiles. Biochemistry, 37, of bioactive oligosaccharides. Biotechnol.
17262–17270. Adv., 26, 436–456.
74 Honda, Y., Fushinobu, S., Hidaka, M., 83 Salas, J.A. and Mendez, C. (2007)
Wakagi, T., Shoun, H., Taniguchi, H., and Engineering the glycosylation of natural
Kitaoka, M. (2008) Alternative strategy for products in actinomycetes. Trends
converting an inverting glycoside Microbiol., 15, 219–232.
hydrolase into a glycosynthase. 84 Hu, Y., Chen, L., Ha, S., Gross, B.,
Glycobiology, 18, 325–330. Falcone, B., Walker, D.,
75 Honda, Y. and Kitaoka, M. (2006) The first Mokhtarzadeh, M., and Walker, S. (2003)
glycosynthase derived from an inverting Crystal structure of the MurG:UDP-
glycoside hydrolase. J. Biol. Chem., 281, GlcNAc complex reveals common
1426–1431. structural principles of a superfamily of
76 Wilkinson, S.M., Liew, C.W., glycosyltransferases. Proc. Natl. Acad. Sci.
Mackay, J.P., Salleh, H.M., Withers, S.G., U.S.A., 100, 845–849.
and McLeod, M.D. (2008) Escherichia coli 85 Charnock, S.J. and Davies, G.J. (1999)
glucuronylsynthase: an engineered Structure of the nucleotide-diphospho-
enzyme for the synthesis of sugar transferase, SpsA from Bacillus
b-glucuronides. Org. Lett., 10, 1585–1588. subtilis, in native and nucleotide-
77 M€ ullegger, J., Chen, H.-M., Chan, W.Y., complexed forms. Biochemistry, 38,
Reid, S.P., Jahn, M., Warren, A.J., 6380–6385.
Salleh, H.M., and Withers, S.G. (2006) 86 Chiu, C.P.C., Lairson, L.L., Gilbert, M.,
Thermostable glycosynthases and Wakarchuk, W.W., Withers, S.G., and
thioglycologases derived from Strynadka, N.C.J. (2007) Structural
Thermotoga maritima b-glucuronidase. analysis of the alpha-2,3-sialyltransferase
ChemBioChem, 7, 1028–1030. Cst-i from Campylobacter jejuni in apo and
78 Umekawa, M., Huang, W., Li, B., substrate-analogue bound forms.
Fujita, K., Ashida, H., Wang, L.-X., and Biochemistry, 46, 7196–7204.
Yamamoto, K. (2008) Mutants of Mucor 87 Thibodeaux, C.J., Melan:̧8con, C.E., and
hiemalis endo-b-N- Liu, H.W. (2008) Natural-product sugar
acetylglucosaminidase show enhanced biosynthesis and enzymatic
transglycosylation and glycosynthase-like glycodiversification. Angew. Chem. Int.
activities. J. Biol. Chem., 283, 4469–4479. Ed., 47, 9814–9859.
79 Jahn, M., Marles, J., Warren, R.A., and 88
Breton, C., Snajdrov a, L., Jeanneau, C.,
Withers, S.G. (2003) Thioglycoligases: Koca, J., and Imberty, A. (2006)
mutant glycosidases for thioglycoside Structures and mechanism of
synthesis. Angew. Chem. Int. Ed., 42, glycosyltransferases. Glycobiology, 16,
352–354. 29R–37R.
80 Jahn, M., Chen, H.M., Mullegger, J., 89 Lairson, L.L., Henrissat, B., Davies, G.J.,
Marles, J., Warren, R.A., and and Withers, S.G. (2008)
Withers, S.G. (2004) Thioglycosynthases: Glycosyltransferases: structures,
References j459
functions, and mechanisms. Annu. Rev. (2007) High-resolution structure of
Biochem., 77, 512–555. NodZ fucosyltransferase involved in the
90 Ramakrishnan, B., Boeggeman, E., biosynthesis of the nodulation factor.
Ramasamy, V., and Qasba, P.K. (2004) Acta Biochim. Pol., 54, 537–549.
Structure and catalytic cycle of b-1,4- 99 Chiu, C.P., Watts, A.G., Lairson, L.L.,
galactosyltransferase. Curr. Opin. Struct. Gilbert, M., Lim, D., Wakarchuk, W.W.,
Biol., 14, 593–600. Withers, S.G., and Strynadka, N.C. (2004)
91 Lariviere, L., Sommer, N., and Morera, S. Structural analysis of the sialyltransferase
(2005) Structural evidence of a passive CstII from Campylobacter jejuni in
base-flipping mechanism for AGT, an complex with a substrate analog. Nat.
unusual GT-B glycosyltransferase. J. Mol. Struct. Mol. Biol., 11, 163–170.
Biol., 352, 139–150. 100 Reinert, D.J., Jank, T., Aktories, K., and
92 Vrielink, A., Ruger, W., Driessen, H.P., Schulz, G.E. (2005) Structural basis for
and Freemont, P.S. (1994) Crystal the function of Clostridium difficile toxin
structure of the DNA modifying enzyme B. J. Mol. Biol., 351, 973–981.
b-glucosyltransferase in the presence and 101 Ha, S., Walker, D., Shi, Y., and Walker, S.

absence of the substrate uridine (2000) The 1.9 A crystal structure of
diphosphoglucose. EMBO J., 13, Escherichia coli MurG, a membrane-
3413–3422. associated glycosyltransferase involved in
93 Buschiazzo, A., Ugalde, J.E., peptidoglycan biosynthesis. Protein Sci.,
Guerin, M.E., Shepard, W., Ugalde, R.A., 9, 1045–1052.
and Alzari, P.M. (2004) Crystal structure 102 Gibson, R.P., Turkenburg, J.P.,
of glycogen synthase: homologous Charnock, S.J., Lloyd, R., and Davies, G.J.
enzymes catalyze glycogen synthesis and (2002) Insights into trehalose synthesis
degradation. EMBO J., 23, 3196–3205. provided by the structure of the retaining
94 Mulichak, A.M., Losey, H.C., Lu, W., glycosyltransferase OtsA. Chem. Biol., 9,
Wawrzak, Z., Walsh, C.T., and 1337–1346.
Garavito, R.M. (2003) Structure of the 103 Grizot, S., Salem, M., Vongsouthi, V.,
TDP-epi-vancosaminyltransferase GtfA Durand, L., Moreau, F., Dohi, H.,
from the chloroeremomycin biosynthetic Vincent, S., Escaich, S., and Ducruix, A.
pathway. Proc. Nat. Acad. Sci. U.S.A., 100, (2006) Structure of the Escherichia coli
9238–9243. heptosyltransferase WaaC: binary
95 Mulichak, A.M., Losey, H.C., Walsh, C.T., complexes with ADP and
and Garavito, R.M. (2001) Structure of ADP-2-deoxy-2-fluoro heptose. J. Mol.
the UDP-glucosyltransferase Gtfb that Biol., 363, 383–394.
modifies the heptapeptide aglycone in 104 Martinez-Fleites, C., Proctor, M.,
biosynthesis of the vancomycin group Roberts, S., Bolam, D.N., Gilbert, H.J.,
antibiotic. Structure, 9, 547–557. and Davies, G.J. (2006) Insights into
96 Mulichak, A.M., Lu, W., Losey, H.C., the synthesis of lipopolysaccharide and
Walsh, C.T., and Garavito, R.M. (2004) antibiotics through the structures of two
Crystal structure of retaining glycosyltransferases from
vancosaminyltransferase GtfD from the family GT4. Chem. Biol., 13, 1143–1152.
vancomycin biosynthetic pathway: 105 Sun, H.-Y., Lin, S.-W., Ko, T.-P., Pan, J.-F.,
interactions with acceptor and nucleotide Liu, C.-L., Lin, C.-N., Wang, A.H.-J., and
ligands. Biochemistry, 43, 5170–5180. Lin, C.-H. (2007) Structure and
97 Yuan, Y., Barrett, D., Zhang, Y., mechanism of Helicobacter pylori
Kahne, D., Sliz, P., and Walker, S. (2007) fucosyltransferase: a basis for
Crystal structure of a peptidoglycan lipopolysaccharide variation and
glycosyltransferase suggests a model for inhibitor design. J. Biol. Chem., 282,
processive glycan chain synthesis. Proc. 9973–9982.
Nat. Acad. Sci. U.S.A., 104, 5348–5353. 106 Guerin, M.E., Buschiazzo, A.,
98 Brzezinski, K., Stepkowski, T., Kordulakova, J., Jackson, M., and
Panjikar, S., Bujacz, G., and Jaskolski, M. Alzaria, P.M. (2005) Crystallization and
460
preliminary crystallographic analysis of glycosyltransferase UrdGT2 involved in

PimA, an essential mannosyltransferase the biosynthesis of the antibiotic
from Mycobacterium smegmatis. Acta urdamycin. J. Mol. Biol., 372, 67–76.
Crystallogr. Sect. F, 61, 518–520. 115 Martinez-Fleites, C., Proctor, M.,
107 Persson, K., Hoa, D.L., Diekelmann, M., Roberts, S., Bolam, D.N., Gilbert, H.J.,
Wakarchuk, W.W., Withers, S.G., and and Davies, G.J. (2006) Insights into the
Strynadka, N.C.J. (2001) Crystal structure synthesis of lipopolysaccharide and
of the retaining galactosyltransferase antibiotics through the structures of two
LgtC from Neisseria meningitidis in retaining glycosyltransferases from
complex with donor and acceptor sugar family Gt4. Chem. Biol., 13, 1143–1152.
analogs. Nat. Struct. Biol., 8, 166–175. 116 Gastinel, L.N., Bignon, C., Misra, A.K.,
108 Ni, L., Chokhawala, H.A., Cao, H., Hindsgaul, O., Shaper, J.H., and
Henning, R., Ng, L., Huang, S., Yu, H., Joziasse, D.H. (2001) Bovine
Chen, X., and Fisher, A.J. (2007) Crystal a-1,3-galactosyltransferase catalytic
structures of Pasteurella multocida domain structure and its relationship
sialyltransferase complexes with acceptor with ABO histo-blood group and
and donor analogues reveal substrate glycosphingolipid glycosyltransferase.
binding sites and catalytic mechanism. EMBO J., 20, 638–649.
Biochemistry, 46, 6288–6298. 117 Gastinel, L.N., Cambillau, C., and
109 Horcajada, C., Guinovart, J.J., Fita, I., and Bourne, Y. (1999) Crystal structure of the
Ferrer, J.C. (2006) Crystal structure of an bovine b-1,4-galactosyltransferase
archaeal glycogen synthase: insights into catalytic domain and its complex with
oligomerization and substrate binding of uridine diphosphogalactose. EMBO J.,
eukaryotic glycogen synthases. J. Biol. 18, 3546–3557.
Chem., 281, 2923–2931. 118 Ramasamy, V., Ramakrishnan, B.,
110 Flint, J., Taylor, E., Yang, M., Bolam, D.N., Boeggeman, E., Ratner, D.M.,
Tailford, L.E., Martinez-Fleites, C., Seeberger, P.H., and Qasba, P.K. (2005)
Dodson, E.J., Davis, B.G., Gilbert, H.J., Oligosaccharide preferences of beta1,
and Davies, G.J. (2005) Structural 4-galactosyltransferase-I: crystal
dissection and high-throughput structures of Met340His mutant of
screening of mannosylglycerate synthase. human beta1,4-galactosyltransferase-I
Nat. Struct. Mol. Biol., 12, 608–614. with a pentasaccharide and trisaccharides
111 Lovering, A.L., De Castro, L.H., Lim, D., of the N-glycan moiety. J. Mol. Biol., 353,
and Strynadka, N.C. (2007) Structural 53–67.
insight into the transglycosylation step of 119 Ihara, H., Ikeda, Y., Toma, S.,
bacterial cell wall biosynthesis: Wang, X., Suzuki, T., Gu, J., Miyoshi, E.,
apoenzyme. Science, 315, 1402–1405. Tsukihara, T., Honke, K., and Matsum, A.
112 Lovering, A.L., De Castro, L.H., Lim, D., (2006) Crystal structure of mammalian
and Strynadka, N.C. (2006) Structural alpha-1,6-fucosyltransferase, FUT8.
analysis of an open form of PBP1B Glycobiology, 17, 455–466.
from Streptococcus pneumoniae. Protein 120 Kubota, T., Shiba, T., Sugioka, S.,
Sci., 15, 1701–1709. Furukawa, S., Sawaki, H., Kato, R.,
113 Bolam, D.N., Roberts, S., Proctor, M.R., Wakatsuki, S., and Narimatsu, H. (2006)
Turkenburg, J.P., Dodson, E.J., Structural basis of carbohydrate transfer
Martinez-Fleites, C., Yang, M., Davis, B.J., activity by human UDP-GalNAc:
Davies, G.J., and Gilbert, H.J. (2007) The polypeptide alpha-N-
crystal structure of two macrolide acetylgalactosaminyltransferase
glycosyltransferases provides a blueprint (pp-GalNAc-T10). J. Mol. Biol., 359,
for host cell antibiotic immunity. Proc. 706–727.
Nat. Acad. Sci. U.S.A., 104, 5336–5341. 121 Fritz, T.A., Raman, J., and Tabak, L.A.
114 Mittler, M., Bechthold, A., and (2006) Dynamic association between
Schulz, G.E. (2007) Structure and the catalytic and lectin domains of
action of the C–C bond-forming human UDP-GalNAc:polypeptide
References j461
alpha -N-acetylgalactosaminyl- evidence for a covergence of metal ion
transferase-2. J. Biol. Chem., 281, independent glycosyltransferase
8613–8619. mechanism. J. Biol. Chem., 281,
122 Pedersen, L.C., Tsuchida, K., 26693–26701.
Kitagawa, H., Sugahara, K., Darden, T.A., 129 Pedersen, L.C., Dong, J., Taniguchi, F.,
and Negishi, M. (2000) Heparan/ Kitagawa, H., Krahn, J.M., Pedersen,
chondroitin sulfate biosynthesis: L.G., Sugahara, S., and Negishi, M. (2003)
structure and mechanism of human Crystal structure of an alpha-1,4-N-
glucuronyltransferase I. J. Biol. Chem., acetylhexosaminyltransferase (Extl2), a
275, 34580–34585. member of the exostosin gene family
123 Kakuda, S., Shiba, T., Ishiguro, M., involved in heparan sulfate biosynthesis.
Tagawa, H., Oka, S., Kajihara, Y., J. Biol. Chem., 278, 14420–14428.
Kawasaki, T., Wakatsuki, S., and Kato, R. 130 Fritz, T.A., Hurley, J.H., Trinh, L.B.,
(2004) Structural basis for acceptor Shiloach, J., and Tabak, L.A. (2004)
substrate recognition of a human The beginnings of mucin biosynthesis:
glucuronyltransferase, GlcAT-P, an the crystal structure of UDP-GalNAc:
enzyme critical in the biosynthesis polypeptide alpha-N-acetylgalactosa-
of the carbohydrate epitope HNK-1. J. minyltransferase-T1. Proc. Nat. Acad. Sci.
Biol. Chem., 279, 22693–22703. U.S.A., 101, 15307–15312.
124 Patenaude, S.I., Seto, N.O.L., 131 Jinek, M., Chen, Y.-W., Clausen, H.,
Borisova, S.N., Szpacenko, A., Cohen, S.M., and Conti, E. (2006)
Marcus, S.L., Palcic, M.M., and Structural insights into the Notch-
Evans, S.V. (2002) The structural basis for modifying glycosyltransferase Fringe.
specificity in human AB0(H) blood group Nat. Struct. Mol. Biol., 13, 945–946.
biosynthesis. Nat. Struct. Biol., 9, 132 Gibbons, B.J., Roach, P.J., and
685–690. Hurley, T.D. (2002) Crystal structure
125 Miley, M.J., Zielinska, A.K., Keenan, J.E., of the autocatalytic initiator of glycogen
Bratton, S.M., Radominska-Pandya, A., biosynthesis, glycogenin. J. Mol. Biol.,
and Redinbo, M.R. (2007) Crystal 319, 463–477.
structure of the UDP-glucuronic acid 133 Unligil, U.M., Zhou, S., Yuwaraj, S.,
binding domain of the human drug Sarkar, M., Schachter, H., and Rini, J.M.
metabolizing UDP- (2000) X-ray crystal structure of rabbit
glucuronosyltransferase 2B7. J. Mol. Biol., N-acetylglucosaminyltransferase I:
369, 498–511. catalytic mechanism and a new protein
126 Shao, H., He, X., Achnine, L., superfamily. EMBO J., 19, 5269–5280.
Blount, J.W., Dixon, R.A., and Wang, X. 134 Lobsanov, Y.D., Romero, P.A., Sleno, B.,
(2005) Crystal structures of a Yu, B., Yip, P., Herscovics, A., and
multifunctional triterpene/flavonoid Howell, P.L. (2004) Structure of
glycosyltransferase from Medicago Kre2p/Mnt1p: a yeast alpha1,2-
truncatula. Plant Cell., 17, 3141–3154. mannosyltransferase involved in
127 Li, L., Modolo, L.V., mannoprotein biosynthesis. J. Biol.
Escamilla-Trevino, L.L., Achnine, L., Chem., 279, 17921–17931.
Dixon, R.A., and Wang, X. (2007) Crystal 135 Offen, W., Martinez-Fleites, C., Yang, M.,
structure of Medicago truncatula Kiat-Lim, E., Davis, B.G., Tarling, C.A.,
UGT85H2 - Insights into the structural Ford, C.M., Bowles, D.J., and Davies, G.J.
basis of a multifunctional (iso)flavonoid (2006) Structure of a flavonoid
glycosyltransferase. J. Mol. Biol., 370, glucosyltransferase reveals the basis for
951–963. plant natural product modification.
128 Pak, J.E., Arnoux, P., Zhou, S., EMBO J., 25, 1396–1405.
Sivarajah, P., Satkunarajah, M., Xing, X., 136 Wagner, G.K., Pesnot, T., and Field, R.A.
and Rini, J.M. (2006) X-ray crystal (2009) A survey of chemical methods for
structure of leukocyte type core 2 beta1, sugar-nucleotide synthesis. Nat. Prod.
6-N-acetylglucosaminyltransferase: Rep., 26, 1172–1194.
462
137 Hanson, H., Best, M., Bryan, M.C., and production, properties, and applications.
Wong, C.-H. (2004) Chemoenzymatic Appl. Microbiol. Biotechnol., 80, 757–765.
synthesis of oligosaccharides and 147 Timmons, S.C., Hui, J.P.M.,
glycoproteins. Trends Biochem. Sci., 29, Pearson, J.L., Peltier, P., Daniellou, R.,
656–663. Nugier-Chauvin, C., Soo, E.C.,
138 Zhang, J., Wu, B., Zhang, Y., Kowal, P., Syvitski, R.T., Ferrieres, V., and
and Wang, P.G. (2003) Creatine Jakeman, D.L. (2008) Enzyme-catalyzed
phosphate – creatine kinase in enzymatic synthesis of furanosyl nucleotides. Org.
synthesis of glycoconjugates. Org. Lett., 5, Lett., 10, 161–163.
2583–2586. 148 Jakeman, D.L., Young, J.L., Huestis, M.P.,
139 Kameda, A., Shiba, T., Kawazoe, Y., Peltier, P., Daniellou, R., Nugier-
Satoh, Y., Ihara, Y., Munekata, M., Chauvin, C., and Ferrieres, V. (2008)
Ishige, K., and Noguchi, T. (2001) A novel Engineering ribonucleoside triphosphate
ATP regeneration system using specificity in a thymidylyltransferase.
polyphosphate-AMP phosphotransferase Biochemistry, 47, 8719–8725.
and polyphosphate kinase. J. Biosci. 149 €
Ohrlein, R. (1999) Glycosyltransferase-
Bioeng., 91, 557–563. catalyzed synthesis of non-natural
140 Nahalka, J., Liu, Z., Chen, X., and oligosaccharides. Top. Curr. Chem., 200,
Wang, P.G. (2003) Superbeads: 227–240.
immobilization in sweet chemistry. 150 Sears, P. and Wong, C.-H. (2001) Toward
Chem. Eur. J., 9, 372–377. automated synthesis of oligosaccharides
141 Ishige, K., Hamamoto, T., Shiba, T., and and glycoproteins. Science, 291,
Noguchi, T. (2001) Novel method for 2344–2350.
enzymatic synthesis of CMP-NeuAc. 151 Pisvejcova, A., Rossi, C., Husakova, L.,
Biosci. Biotechnol. Biochem., 65, Kren, V., Riva, S., and Monti, D. (2006)
1736–1740. b-1,4-Galactosyltransferase-catalyzed
142 Nahalka, J. and P€atoprsty, V. (2009) glycosylation of sugar derivatives:
Enzymatic synthesis of sialylation modulation of the enzyme activity by
substrates powered by a novel a-lactalbumin, immobilization and
polyphosphate kinase (PPK3). Org. solvent tolerance. J. Mol. Cat. B: Enzym.,
Biomol. Chem., 7, 1778–1780. 39, 98–104.
143 Endo, T., Koizumi, S., Tabata, K., and 152 B€ulter, T., Schumacher, T., Namdjou, D.-J.,
Ozaki, A. (2000) Large-scale production Gutierrez Gallego, R., Clausen, H., and
of CMP-NeuAc and sialylated Elling, L. (2001) Chemoenzymatic
oligosaccharides through bacterial synthesis of biotinylated nucleotide sugars
coupling. Appl. Microbiol. Biotechnol., 53, as substrates for glycosyltransferases.
257–261. ChemBioChem, 2, 884–894.
144 Mizanur, R.M., Zea, C.J., and Pohl, N.L. 153 Sujino, K., Uchiyama, T., Hindsgaul, O.,
(2004) Unusually broad substrate Seto, N.O.L., Wakarchuk, W.W., and
tolerance of a heat-stable archaeal sugar Palcic, M.M. (2000) Enzymatic synthesis
nucleotidyltransferase for the synthesis of of oligosaccharide analogues: evaluation
sugar nucleotides. J. Am. Chem. Soc., 126, of UDP-Gal analogues as donors for three
15993–15998. retaining a-galactosyltransferases. J. Am.
145 Mizanur, R.M. and Pohl, N.L.B. (2009) Chem. Soc., 122, 1261–1269.
Phosphomannose isomerase/GDP- 154 Patenaude, S.I., Seto, N.O.L.,
mannose pyrophosphorylase from Borisova, S.N., Szpacenko, A.,
Pyrococcus furiosus: a thermostable Marcus, S.L., Palcic, M.M., and
biocatalyst for the synthesis of Evans, S.V. (2002) The structural basis
guanidinediphosphate-activated and for specificity in human ABO(H) blood
mannose-containing sugar nucleotides. group biosynthesis. Nat. Struct. Biol., 9,
Org. Biomol. Chem., 7, 2135–2139. 685–690.
146 Mizanur, R.M. and Pohl, N.L. (2008) 155 Boeggeman, E., Ramakrishnan, B.,
Bacterial CMP-sialic acid synthetases: Kilgore, C., Khidekel, N.,
References j463
Hsieh-Wilson, L.C., Simpson, J.T., and 164 Antoine, T., Priem, B., Heyraud, A.,
Qasba, P.K. (2007) Direct identification of Greffe, L., Gilbert, M., Wakarchuk, W.W.,
nonreducing GlcNAc residues on N- Lam, J.S., and Samain, E. (2003)
glycans of glycoproteins using a novel Large-scale in vivo synthesis of the
chemoenzymatic method. Bioconjugate carbohydrate moieties of gangliosides
Chem., 18, 806–814. GM1 and GM2 by metabolically
156 Pasek, M., Ramakrishnan, B., engineered Escherichia coli.
Boeggeman, E., Manzoni, M., ChemBioChem, 4, 406–412.
Waybright, T.J., and Qasba, P.K. (2009) 165 Antoine, T., Heyraud, A., Bosso, C., and
Bioconjugation and detection of Samain, E. (2005) Highly efficient
lactosamine moiety using a1,3- biosynthesis of the oligosaccharide
galactosyltransferase mutants that moiety of the GD3 ganglioside by
transfer C2-modified galactose with a using metabolically engineered
chemical handle. Bioconjugate Chem., 20, Escherichia coli. Angew. Chem. Int. Ed., 44,
608–618. 1350–1352.
157 Qasba, P.K., Ramakrishnan, B., and 166 Nishida, Y., Wiemann, T., and Thiem, J.
Boeggeman, E. (2005) Substrate-induced (1992) Transfer of galactose to the
conformational changes in anomeric position of N-acetyl
glycosyltransferases. Trends. Biochem. gentosamine. Tetrahedron Lett., 33,
Sci., 30, 53–62. 8043–8046.
158 Rich, J.R., Szpacenko, A., Palcic, M.M., 167 Nishida, Y., Tamakoshi, H., Kitagawa, Y.,
and Bundle, D.R. (2004) Kobayashi, K., and Thiem, J. (2000)
Glycosyltransferase-catalyzed synthesis A novel bovine b-1,4-
of thiooligosaccharides. Angew. Chem. galactosyltransferase reaction to yield b-D-
Int. Ed., 43, 613–615. galactopyranosyl-(1-3)-linked
159 Ramakrishnan, B., Balaji, P.V., and disaccharides from L-sugars. Angew.
Qasba, P.K. (2002) Crystal structure of Chem. Int. Ed., 39, 2000–2003.
b1,4-galactosyltransferase complex with 168 Nishida, Y., Tamakoshi, H.,
UDP-Gal reveals an oligosaccharide Kobayashi, K., and Thiem, J. (2001)
acceptor binding site. J. Mol. Biol., 318, The first bovine b1,4-
491–502. galactosyltransferase reaction with an
160 Wohlgemuth, R. (2008) Tools and acyclic acceptor substrate, 3-acetamido-
ingredients for the biocatalytic synthesis 1,2-propanediol, to yield a 3-O-b-D-
of carbohydrates and glycoconjugates. galactopyranosyl-sn-glycerol skeleton.
Biocatal. Biotransform., 26, 42–48. Org. Lett., 3, 1–3.
161 Zhang, J., Kowal, P., Chen, X., and Wang, 169 Lairson, L.L., Watts, A.G.,
P.G. (2003) Large-scale synthesis of Wakarchuk, W.W., and Withers, S.G.
globotriose derivatives through (2006) Using substrate engineering to
recombinant E. coli. Org. Biomol. Chem., harness enzymatic promiscuity and
1, 3048–3053. expand biological catalysis. Nat. Chem.
162 Bettler, E., Imberty, A., Priem, B., Biol., 2, 724–728.
Chazalet, V., Heyraud, A., Joziasse, D.H., 170 Jamaluddin, H., Tumbale, P., Ferns, T.A.,
and Geremia, R.A. (2003) Production of Thiyagarajan, N., Brew, K., and
recombinant xenotransplantation Acharya, K.R. (2009) Crystal structure
antigen in Escherichia coli. Biochem. of a-1,3-galactosyltransferase (a3GT) in a
Biophys. Res. Commun., 302, 620–624. complex with p-nitrophenyl-b-galactoside
163 Priem, B., Gilbert, M., Wakarchuk, W.W., (pNPbGal). Biochem. Biophys. Res.
Heyraud, A., and Samain, E. (2002) A new Comm., 385, 601–604.
fermentation process allows large-scale 171 Laurent, N., Voglmeir, J., and Flitsch, S.L.
production of human milk (2008) Glycoarrays – tools for
oligosaccharides by metabolically determining protein-carbohydrate
engineered bacteria. Glycobiology, 12, interactions and glycoenzyme specificity.
235–240. Chem. Commun., 4400–4412.
464
172 Dumon, C., Bosso, C., Utille, J.P., Gow, N.A.R. (2006) Outer chain
Heyraud, A., and Samain, E. (2006) N -glycans are required for cell wall
Production of Lewis x tetrasaccharides by integrity and virulence of Candida
metabolically engineered Escherichia coli. albicans. J. Biol. Chem., 281, 90–98.
ChemBioChem, 7, 359–365. 179 Sommer, U., Liu, H., and Doering, T.L.
173 Scott Houliston, R., Bernatchez, S., (2003) An a-1,3-mannosyltransferase of
Karwaski, M.-F., Mandrell, R.E., Cryptococcus neoformans. J. Biol. Chem.,
Jarrell, H.C., Wakarchuk, W.W., and 278, 47724–47730.
Gilbert, M. (2009) Complete 180 Sun, S., Scheffer, N.K., Gibson, B.W.,
chemoenzymatic synthesis of the Wang, J., and Munson, R.S. (2002)
Forssman antigen using novel Identification and characterization of the
glycosyltransferases identified in N-acetylglucosamine
Campylobacter jejuni and Pasteurella glycosyltransferases gene of Haemophilus
multocida. Glycobiology, 19, 153–159. ducreyi. Infect. Immun., 70, 5887–5892.
174 Bernatchez, S., Gilbert, M., 181 Deadman, M.E., Lundstr€om, S.L.,
Blanchard, M.C., Karwaski, M.-F., Li, J., Schweda, E.K.H., Moxon, E.R., and
DeFrees, S., and Wakarchuk, W.W. (2007) Hood, D.W. (2006) Specific amino acids
Variants of the b1,3-galactosyltransferase of the glycosyltransferases LpsA direct the
CgtB from the bacterium Campylobacter addition of glucose or galactose to the
jejuni have distinct acceptor specificities. terminal inner core heptose of
Glycobiology, 17, 1333–1343. Haemophilus influenzae
175 Gilbert, M., Kawarski, M.F., lipopolysaccharide via alternative
Bernatchez, S., Young, N.M., linkages. J. Biol. Chem., 281,
Taboada, E., Michniewicz, J., 29455–29467.
Cunningham, A.M., and 182 Shao, J., Zhang, J., Kowal, P., Lu, Y.,
Wakarchuk, W.W. (2002) The genetic and Wang, P.G. (2002) Overexpression
bases for the variation in the lipo- and biochemical characterization of
oligosaccharide of the mucosal pathogen, b-1,3-N-acetylgalactosaminyltransferase
Campylobacter jejuni. J. Biol. Chem., 277, LgtD from Haemophilus influenzae
327–337. strain Rd. Biochem. Biophys. Res.
176 Gilbert, M., Brisson, J.R., Kawarski, M.F., Commun., 295, 1–8.
Michniewicz, J., Cunningham, A.M., 183 Wang, G., Rasko, D.A., Sherburne, R.,
Wu, Y., Young, N.M., and and Taylor, D.E. (1999) Molecular genetic
Wakarchuk, W.W. (2000) Biosynthesis basis for the variable expression of Lewis
of ganglioside mimics in Campylobacter Yantigen in Helicobacter pylori: analysis of
jejuni OH4384. J. Biol. Chem., 275, the a(1,2)fucosyltransferase gene. Mol.
3896–3906. Microbiol., 31, 1265–1274.
177 Munro, C.A., Bates, S., Buurman, E.T., 184 Sun, H.Y., Lin, S.W., Ko, T.P., Pan, J.F.,
Bleddyn Hughes, H., MacCallum, D.M., Liu, C.L., Lin, C.N., Wang, A.H.J., and
Bertram, G., Atrih, A., Ferguson, M.A.J., Lin, C.H. (2007) Structure and
Bain, J.M., Brand, A., Hamilton, S., mechanism of Helicobacter pylori
Westwater, C., Thomson, L.M., fucosyltransferase - A basis for
Brown, A.J.P., Odds, F.C., and lipopolysaccharide variation and
Gow, N.A.R. (2005) Mnt1p and Mnt2p of inhibitor design. J. Biol. Chem., 282,
Candida albicans are partially redundant 9973–9982.
a-1,2-mannosyltransferases that 185 Rabbani, S., Miska, V., Wipf, B., and
participate in O-linked mannosylation Ernst, B. (2005) Molecular cloning and
and are required for adhesion and functional expression of a novel
virulence. J. Biol. Chem., 280, 1051–1060. Helicobacter pylori
178 Bates, S., Hughes, H.B., Munro, C.A., a-1,4fucosyltransferase. Glycobiology, 15,
Thomas, W.P.H., MacCallum, D.M., 1076–1083.
Bertram, G., Atrih, A., Ferguson, M.A.J., 186 Rasko, D.A., Wang, G., Palcic, M.M.,
Brown, A.J.P., Odds, F.C., and and Taylor, D.E. (2000) Cloning and
References j465
characterization of the a(1,3/4) Escherichia coli is hybrid of those found in
fucosyltransferase of Helicobacter pylori. J. Escherichia coli K-12 and Salmonella
Biol. Chem., 275, 4988–4994. enterica. J. Biol. Chem., 273, 8849–8859.
187 Logan, S.M., Altman, E., Mykytczuk, O., 194 Videira, P., Fialho, A., Geremia, R.A.,
Brisson, J.-R., Chandan, V., Breton, C., and Sa-Correia, I. (2001)
St. Michael, F., Masson, A., Leclerc, S., Biochemical characterization of the b-1,4-
Hiratsuka, K., Smirnova, N., Li, J., Wu, Y., glucuronosyltransferase GelK in the
and Wakarchuk, W.W. (2005) Novel gellan gum-producing strain
biosynthetic functions of Sphingomonas paucimobilis ATCC 31461.
lipopolysaccharide rfaJ homologs from Biochem. J., 358, 457–464.
Helicobacter pylori. Glycobiology, 15, 195 Stingele, F., Neeser, J.R., and Mollet, B.
721–733. (1996) Identification and characterization
188 Wakarchuk, W.W., Schur, M.J., St. of the eps (exopolysaccharide) gene
Michael, F., Li, J., Eichler, E., and cluster from Streptococcus thermophilus
Whitfield, D. (2004) Characterization of Sfi6. J. Bacteriol., 178, 1680–1690.
the a-1,2-N 196 Pukin, A.V., Weijers, C.A.G.M.,
-acetylglucosaminyltransferase of van Lagen, B., Wechselberger, R.,
Neisseria gonorrhoeae, a key control point Sun, B., Gilbert, M., Karwaski, M.-F.,
in lipooligosaccharide biosynthesis. Florack, D.E.A., Jacobs, B.C.,
Glycobiology, 14, 537–546. Tio-Gillen, A.P., van Belkum, A.,
189 Jennings, M.P., Virji, M., Evans, D., Endtz, H.P., Visser, G.M., and Zuilhof, H.
Foster, V., Srikhanta, Y.N., Steeghs, L., (2008) GM3, GM2 and GM1 mimics
van der Ley, P., and Moxon, E.R. (1998) designed for biosensing:
Identification of a novel gene involved in chemoenzymatic synthesis, target
pilin glycosylation in Neisseria affinities and 900MHz NMR analysis.
meningitidis. Mol. Microbiol., 29, 975–984. Carbohydr. Res., 343, 636–650.
190 Yu, H., Chokhawala, H., Karpel, R., 197 Namdjou, D.-J., Chen, H.-M.,
Yu, H., Wu, B., Zhang, J.B., Zhang, Y.X., Vinogradov, E., Brochu, D., Withers, S.G.,
Jia, Q., and Chen, X. (2005) A and Wakarchuk, W.W. (2008) A b-1,4-
multifunctional Pasteurella multocida galactosyltransferase from Helicobacter
sialyltransferase: a powerful tool for the pylori is an efficient and versatile
synthesis of sialoside libraries. J. Am. biocatalyst displaying a novel activity for
Chem. Soc., 127, 17618–17619. thioglycoside synthesis. ChemBioChem,
191 Tsukamoto, H., Takakura, Y. Mine, T., and 9, 1632–1640.
Yamamoto, Y. (2008) Photobacterium sp. 198 Minami, A., Uchida, R., Eguchi, T., and
JT-ISH-224 produces two Kakinuma, K. (2005) Enzymatic approach
sialyltransferases, a/b-galactoside a2,3- to unnatural glycosides with diverse
sialyltransferase and b-galactoside aglycon scaffolds using
a2,6-sialyltransferase. J. Biochem., 143, glycosyltransferase VinC. J. Am. Chem.
187–197. Soc., 127, 6148–6149.
192 Rocchetta, H.L., Burrows, L.L., Pacan, 199 D€urr, C., Hoffmeister, D., Wohlert, S.-E.,
J.C., and Lam, J.S. (1998) Three Ichinose, K., Weber, M., von Mulert, U.,
rhamnosyltransferases responsible for Thorson, J.S., and Bechthold, A. (2004)
assembly of the A-band D-rhamnan The glycosyltransferase UrdGT2
polysaccharide in Pseudomonas catalyzes both C- and O-glycosidic sugar
aeruginosa: a fourth transferase, WbpL, is transfers. Angew. Chem. Int. Ed., 43,
required for the initiation of both A-band 2962–2965.
and B-band lipopolysaccharide synthesis. 200 Salas, A.P., Zhu, L., Sanchez, C.,
Mol. Microbiol., 28, 1103–1119. Brana, A.F., Rohr, J., Mendez, C., and
193 Heinrichs, D.E., Monteiro, M.A., Salas, J.A. (2005) Deciphering the late
Perry, M.B., and Whifield, C. (1998) steps in the biosynthesis of the antitumor
The assembly system for the indolocarbazole staurosporine: sugar
lipopolysaccharide R2 core-type of donor substrate flexibility of the StaG
466
glycosyltransferase. Mol. Microbiol., 209 Sawangwan, T., Goedl, C., and

58, 17–27. Nidetzky, B. (2009) Single-step enzymatic
201 Williams, G.J., Zhang, C., and synthesis of (R)-2-O-a-D-glucopyranosyl
Thorson, J.S. (2007) Expanding the glycerate, a compatible solute from
promiscuity of a natural-product micro-organisms that functions as a
glycosyltransferase by directed protein stabiliser. Org. Biomol. Chem., 7,
evolution. Nat. Chem. Biol., 3, 4267–4270.
657–662. 210 Seibel, J., J€ordening, H.-J., and
202 Gantt, R.W., Goff, R.D., Williams, G.J., Buchholz, K. (2006) Glycosylation with
and Thorson, J.S. (2008) Probing the activated sugars using
aglycon promiscuity of an engineered glycosyltransferases and
glycosyltransferase. Angew. Chem. Int. transglycosidases. Biocatal. Biotransform.,
Ed., 47, 8889–8892. 24, 311–342.
203 Park, S.-H., Park, H.-Y., Sohng, J.K., 211 Plou, F.J., Martin, M.T.,
Lee, H.C., Liou, K., Yoon, Y.J., and Gomez de Segura, A., Alcalde, M.,
Kim, B.-G. (2009) Expanding substrate and Ballesteros, A. (2002)
specificity of GT-B fold Glucosyltransferases acting on starch
glycosyltransferase via domain swapping or sucrose for the synthesis of
and high-throughput screening. oligosaccharides. Can. J. Chem., 80,
Biotechnol. Bioeng., 102, 988–994. 743–752.
204 Truman, A.W., Dias, M.V.B., Wu, S., 212 Baciu, I.E., J€ordening, H.-J., Seibel, J.,
Blundell, T.L., Huang, F., and and Buchholz, K. (2005) Investigations
Spencer, J.B. (2009) Chimeric of the transfructosylation reaction by
glycosyltransferases for the generation fructosyltransferase from B. subtilis
of hybrid glycopeptides. Chem. Biol., 16, NCIMB 11871 for the synthesis of
676–685. the sucrose analogue galactosyl-
205 Sanchez, C., Zhu, L., Brana, A.F., fructoside. J. Biotechnol., 116,
Salas, A.P., Rohr, J., Mendez, C., and 347–357.
Salas, J.A. (2005) Combinatorial 213 Biwer, A., Antranikian, G., and
biosynthesis of antitumor Heinzle, E. (2002) Enzymatic production
indolocarbazole compounds. Proc. Natl. of cyclodextrins. Appl. Microbiol.
Acad. Sci. U.S.A., 102, 461–466. Biotechnol., 59, 609–617.
206 Fu, X., Albermann, C., Jiang, J.Q., 214 Kelly, R.M., Dijkhuizen, L., and
Liao, J.C., Zhang, C.S., and Thorson, J.S. Leemhuis, H. (2009) The evolution of
(2003) Antibiotic optimization via in vitro cyclodextrin glucanotransferase product
glycorandomization. Nat. Biotechnol., 21, specificity. Appl. Microbiol. Biotechnol., 84,
1467–1469. 119–133.
207 Lairson, L.L. and Withers, S.G. (2004) 215 Kochikyan, V.T., Markosyan, A.A.,
Mechanistic analogies amongst Abelyan, L.A., Balayan, A.M., and
carbohydrate modifying enzymes. Chem. Abelyan, V.A. (2006) Combined
Commun., 20, 2243–2248. enzymatic modification of stevioside and
208 Goedl, C., Sawangwan, T., Mueller, M., rebaudioside A. Appl. Biochem. Microbiol.,
Schwarz, A., and Nidetzky, B. (2008) 42, 31–37.
A high-yielding biocatalytic process 216 Shimoda, K., Hamada, H., and
for the production of 2-O-(a-D- Hamada, H. (2008) Chemo-enzymatic
glucopyranosyl)-sn-glycerol, a natural synthesis of ester-linked
osmolyte and useful moisturizing taxol–oligosaccharide conjugates as
ingredient. Angew. Chem. Int. Ed., 47, potential prodrugs. Tetrahedron Lett.,
10086–10089. 49, 601–604.
j467
11
Addition of Water to C¼C Bonds and its Elimination
Jianfeng Jin, Isabel W.C.E. Arends, and Ulf Hanefeld
11.1
Introduction
The addition of water to a C¼C bond and its elimination is at the core of
undergraduate teaching. Nonetheless, very few examples of the successful chemical
addition of water to an isolated or a conjugated C¼C bond exist on a preparative scale,
nor have many selective water eliminations been reported yielding defined products.
Essentially, the reports on preparative examples for addition reactions are limited to
the production of tert-butanol and similar alcohols [1, 2]. Enzymatically, the addition
of water to a C¼C bond and the reverse reaction, the elimination of water, are a
mainstay of our metabolism. Fumarase and aconitase (Acn) are vital enzymes in the
citric acid cycle. Equally, the synthesis and degradation of fatty acids involves an
indispensable water elimination/addition step. Thus, our energy generation and
storage systems rely on water addition to C¼C bonds. Consequently, nature provides
us with very efficient enzymes that catalyze this reaction with a selectivity that is well
beyond current chemical standards.
The enzymes that catalyze the reversible addition of water to C¼C bonds are called
hydro-lyases or hydratases (E.C. 4.2.1-) (hydro for the water that is added, lyase since
water is added to a double bond). Given their essential character they are ubiquitous
in nature and a large variety of enzymes from different sources have been biochem-
ically characterized. In addition to the many hydro-lyases, some tautomerases and
dehalogenases also catalyze the addition of water to a C¼C bond to form an
alcohol [3].
Although hydro-lyases are widespread in nature, their substrate spectrum is
somewhat limited. Since all the reactions they catalyze in nature are highly specific,
hydro-lyases usually display high selectivity for one compound and even very small
changes in the structure thereof are not accepted. This has, to date, limited their
appreciation to several well-developed examples that are even industrially
employed [3]. On the other hand, generally applicable, reliable hydro-lyases with a
broad substrate range are still missing.
and Oliver May.
j 11 Addition of Water to C¼C Bonds and its Elimination
468
Almost a 100 hydration/dehydration reactions catalyzed by hydro-lyases have

been discovered [3]. These reactions can be grouped into two types of reactions
based on the mechanism of hydration (Scheme 11.1). One type of reaction is the
hydration of isolated double bonds, that is, unpolarized double bonds. The other
type is the addition of water to double bonds conjugated with a carbonyl function –
in chemical terms it is a Michael addition of water [4]. As catalysts the hydro-lyases
catalyze both the addition of water and its elimination. The industrially relevant
equilibrium between fumarate and malate, but also maleate and malate as well as
citraconate and citramalate, was established to be in favor of the hydration
product. However, quantitative yields can only be obtained by in situ product
removal [3].
R1 R1
+ H2 O HO
R2 R3 - H 2O R2 R3
HO
+ H2 O
R1 O - H 2O R1 O
R2 R2
Scheme 11.1 Hydro-lyases catalyze the addition of water to isolated double bonds and the Michael
addition to conjugated double bonds.
The stereochemistry of enzymatic addition and b-elimination reactions can be

either syn or anti (Scheme 11.2) [5]. For example, enol-CoA hydratase catalyzes
dehydration via a syn manner, while dehydration catalyzed by enolase has an anti
stereochemistry. Full details of the mechanistic aspects and the mechanisms of the
different enzymes, if they are elucidated, can be found in excellent biochemical
reviews and are well beyond the scope of this chapter [6].
H
H OH
+ H 2O
OH syn
OH
+ H 2O
H OH anti
H
Scheme 11.2 Addition and elimination are known to occur in both syn and anti fashion.
11.2 Addition of Water to Isolated Double Bonds j469
11.2
Addition of Water to Isolated Double Bonds
11.2.1
Oleate Hydratase
Oleate hydratase (EC 4.2.1.53) has been known for almost 50 years [7]. However,
the enzyme has been isolated from Elizabethkingia meningoseptica (formerly
Pseudomonas sp. Strain 3266) and characterized only very recently [8]. It is a
monomeric enzyme with a molecular mass of 73 kDa that contains a catalytically
non-essential calcium ion. The mechanism remains to be elucidated. It has been
identified in several other organisms, though never characterized. But the enzyme
has been applied successfully. It catalyzes the reversible, (R)-selective addition of
water and has been utilized to produce (R)-10-hydroxystearate. Catalyzed by bakers
yeast this has then been converted into (R)-c-dodecalactone, an essential flavor in
whiskey (Scheme 11.3) [9, 10].
OH
oleic acid
oleate hydratase
OH
OH
(R)-10-hydroxystearate
yeast
(R)-dodecalactone; ee = 87 %
O O
Scheme 11.3 Oleate hydratase catalyzed synthesis of a whiskey flavor.
11.2.2
Carotenoid Hydratases
In the biosynthesis of terpenes, functionalization of the highly lipophilic chains built

up from isopentenyl pyrophosphate is of great importance. In the biosynthesis
of the different carotenoids several hydratase were identified [11–13]. They are
commonly called 1,2-hydratases as they add water to the C¼C double bond at the
1,2 position according to carotenoid nomenclature. Carotenoid1,2-hydratase CrtC of
the photosynthetic bacteria Rubrivivax gelatinosus is a membrane-bound protein with
470
2 H2 O neurosporene
CrtC
OH
1-(OH)-neurosporene
OH
OH
1,1'-(OH) 2-neurosporene
lycopene
2 H2 O
CrtC
OH
1-(OH)-lycopene
OH
OH
1,1'-(OH)2-lycopene
Scheme 11.4 Carotenoid hydratases add water selectively and exclusively to the non-conjugated
double bonds of lycopene and similar compounds.
a molecular mass of 44 kDa. It catalyzes the hydration of neurosporene and lycopene

to 1-OH-neurosporene and 1,10 -(OH)2-neurosporene and 1-OH-lycopene and 1,10 -
(OH)2-lycopene respectively (Scheme 11.4). Recently, a structurally unrelated carot-
enoid 1,2-hydratase from non-photosynthetic bacterium Deinococcus was
described [14]. Although extremely selective for the double bond these hydro-lyases
add water to, no synthetic applications have been reported yet.
11.2 Addition of Water to Isolated Double Bonds j471
11.2.3
Kievitone Hydratase
Kievitone hydratase of Fusarium solani f. sp. phaseoli (Khase, EC 4.2.1.95) catalyzes the
conversion of kievitone into kievitone hydrate, which is less fungitoxic
(Scheme 11.5) [15, 16]. In this manner the fungus protects itself against the toxin
produced by its host plant, the French bean. The secreted glycoprotein is a homo-
dimer with a subunit molecular mass of 47 to 49 kDa [17]. There have been no
mechanistic studies on the enzyme, and the scope of the enzyme has not been
studied.
HO OH HO OH
OH O OH O
kievitone hydratase
HO O HO O
H2O
OH
H3C CH3 H3 C CH3
kievitone
Scheme 11.5 Kievitone hydratase detoxifies kievitone.
11.2.4
Acetylene Hydratase
Acetylene hydratase (AH, EC 4.2.1.112) catalyzes the addition of water to acetylene,

yielding acetaldehyde [18]. The enzyme purified from Pelobacter acetylenicus contains
one [4Fe-4S] cluster and a tungsten center, which is coordinated by two molybdop-
terin-guanine dinucleotides [19, 20]. Unlike other tungstoenzymes, the reaction
catalyzed by AH does not involve electron transfer and the oxidation states of W and
the [4Fe:4S] cluster does not change during catalysis. AH is extremely sensitive to
oxygen and activation by a strong reductant is required for AH activity. Interestingly, it
is not acetylene that coordinates to the tungsten, but rather water. The tungsten and a
close-by acid group (with an unusually high pKa) activate the water. This activation
can occur through a hydroxo ligand, implying a nucleophilic attack on the acetylene
and protonation of the vinyl anion by the acid. Tautomerization of the enol would then
yield acetaldehyde. However, the hydroxo ligand of tungsten is only stable in the
oxidized form of tungsten – the inactive form of the enzyme. This points to a
Markovnikov-type addition of water as an electrophile. The latter mechanism would
imply a vinyl cation, highly susceptible to the attack of a nucleophile, which here
would have to be a hydride in order to obtain the desired acetaldehyde enol
(Scheme 11.6).
472
H OH O
acetylene hydratase
H H
H
H H
Asp 13
H H O
H O
O
S S
W
S S
S
Scheme 11.6 Acetylene hydratase contains a tungsten atom in its active site that binds water,
thereby prearranging it for the reaction with acetylene.
11.2.5
Diol Dehydratase/Glycerol Dehydratase
Diol dehydratase (EC 4.2.1.28) and more specifically glycerol dehydratase belongs to
the subclass of isomerases of the B12 enzymes. This group of enzymes requires
vitamin B12 (adenosylcobalamin, AdoCbl) as the coenzyme and normally catalyzes
carbon-skeleton rearrangements, heteroatom eliminations, and intramolecular ami-
no group migrations [21]. Diol dehydratase catalyzes the dehydration of 1,2-diols
(e.g., 1,2-ethanediol, 1,2-propanediol, 1,2-butanediol) and glycerol to the correspond-
ing aldehydes [22, 23]. Scheme 11.7 depicts the generally accepted mechanism for the
transformation of glycerol into 3-hydroxypropionaldehyde by glycerol dehydratase
(GDH). The first step is homolytic cleavage of the CoC bond of AdoCbl to form cob
(II)alamin and the 50 -deoxyadenosyl radical (Ado ). H-abstraction from the C1 atom
.
of glycerol by Ado generates a radical and 50 -deoxyadenosine (Ado-H). Migration

.
of the C2 hydroxy group onto the C1 atom generates the product-related radical.
Adenine
HO O
Ado =
OH OH OH
HO OH HO CH 2 HO OH HO OH
Ado
HO O
CoIII CoII Ado Ado H
HO OH HO O HO OH HO OH
3-hydroxy H 2O OH OH
propionaldehyde
Scheme 11.7 General reaction mechanism for the glycerol dehydratase-catalyzed transformation
of glycerol into 3-hydroxypropionaldehyde. This is a key step in the industrial synthesis of 1,3-
propanediol.
11.3 Addition of Water to Conjugated Double Bonds j473
Re-abstraction of the hydrogen atom from Ado-H by the C2 radical affords 3-
.
hydroxypropanal hydrate and regeneration of Ado . Then water is expelled from
the unstable hydrate to yield the 3-hydroxypropanal [24–26].
The glycerol dehydratase is a hydro-lyase that is not used to add water but rather, as
the name glycerol dehydratase indicates, to eliminate water. It is a key part of the
industrial synthesis of 1,3-propanediol from glycerol by fermentation. DuPont
produces 500 t year-1 of this diol; initially this was performed from glycerol but now
modified microorganisms are utilized that even accept starch rather than glycerol as
feedstock [27]. The diol is used for the production of the polymer SORONA (poly
(trimethylene terephthalate)) [28].
11.3
Addition of Water to Conjugated Double Bonds
Most hydro-lyases add water to a double bond that is conjugated to a carbonyl

group. This can be an acid group, a ketone, or a thioester. The distinct difference
between these three activating groups is the pKa of the proton that needs to be
removed in the elimination reaction. While the a-CH in an acid has a pKa of
around 30 that of a ketone is around 24 and that of a thioester is much lower, at
approx. 20 [6]. It is not always entirely clear how the hydro-lyases manage to
deprotonate the a-CH of acids. In many cases an iron sulfur cluster acts as a
complexing and activating reagent. However, fumarase C does not require any
cofactor and can achieve the elimination of water as a pure organocatalyst
(Scheme 11.9 below). A detailed discussion of the mechanisms of the different
hydratases/hydro-lyases is given below. In addition the thorough treatment thereof
by P. A. Frey and A. D. Hegeman can be recommended [6].
11.3.1
The Activating Group is an Acid Group
11.3.1.1 Fumarase
The interconversion of fumarate and malate catalyzed by fumarase is a long studied
example of a hydro-lyase catalyzed reaction (Scheme 11.8) [29]. Fumarase (EC 4.2.1.2)
is part of the vital citric acid cycle, thus ensuring the oxidation of acetate to CO2. As an
essential part of the primary metabolism it catalyzes the reversible hydration/
dehydration of fumarate to (S)-malate (in the older literature written as L-malate).
fumarase
H COO + H2O H COO
H H
OOC H OOC OH
fumarate (S)-malate
Scheme 11.8 All fumarases catalyze the addition of water to fumarate to yield (S)-malate.
474
Escherichia coli contains three fumarases that are categorized in two classes.
Fumarase A (Fum A) and Fumarase B (Fum B) belong to class I fumarases, which
are Fe2 þ dependent, dimeric, and heat-labile proteins. Fum A and Fum B have the
same number of amino acids and share 90% sequence homology [30]. The active Fum
A contains a [4Fe-4S] cluster that can be inactivated upon oxidation, resulting in a
[3Fe-4S] cluster. The activity can be restored by anaerobic incubation with iron and
thiol [31, 32]. In addition to catalyzing the hydration/dehydration of fumarate to
malate, Fum A can also catalyze the isomerization of enol to keto oxalacetic acid
(OAA) [33]. Although class I fumarases are rather flexible in the substrates they accept
they are not commonly used in the laboratory due to their low stability.
Fumarase C (Fum C) from Escherichia coli, yeast fumarase, and pig heart fumarase
belong to class II fumarases, which are tetrameric, heat stable, and independent of
Fe2 þ [34–37]. Indeed, virtually all fumarases that are commonly used and just called
fumarase are class II fumarases. Although they have very different origins, they are
structurally related and all display a very limited substrate spectrum. The crystal

structure of Fum C revealed that there are two dicarboxylate-binding sites located 12 A
from each other [38, 39]. Site A is located in a relatively deep pit at the interface of three
subunits while site B is near the molecular surface and formed by a single
subunit [40]. It has been proven that site A is the active site and site B can be
regarded as the stretch of site A, which is gated by a histidine residue. The reversible
protonation/deprotonation of His129 induces the conformational change and thus
controls access to the B site [41–43]. This complex mechanism is necessary to achieve
the deprotonation of malate. Thus the catalytic cycle of fumarase includes the
reaction steps one would expect in any catalyst, and an additional conformational
change that takes place after product release (Scheme 11.9). Overall this is the Iso-
mechanism of fumarase. The class II fumarases family also includes aspartase,
adenylosuccinate lyase, arginosuccinate lyase, and d-crystallin [34]. The catalytically
active class II fumarase family members catalyze the addition of water, NH3, or an
organo-nitrogen-containing compound to the olefinic bond of fumarate.
Fumarases have a very narrow substrate spectrum. Next to fumarate, only
chloro-, fluoro-, and difluoro-fumarate are converted at a useful rate. Fumarate is
isomerization
E1 A H of fumarase E2 A
O2 C H B: BH O2 C
(S)-malate
H CO2 HO CO2
HR
O2C H O 2C HS
E1 A H E2 A HO CO2
H CO 2
H
B: BH
H OH
Scheme 11.9 Iso-mechanism in the reaction with fumarase. HR is added/eliminated during this
trans-oriented reaction.
industrially converted into (S)-malate on a multi-ton scale but other applications
are scarce [44]. Chloro-fumarate was converted into L-threo-chloromalic acid and
this was utilized for the synthesis of 2-deoxy-D-ribose and of trans-D-erythro-
sphingosine (Scheme 11.10) [45].
pig heart OH OH
CO2 fumarase CO2 HO O
O2 C O2 C
Cl Cl OH H
chlorofumarate 2-deoxy-D-ribose
OH
H3C(H2C)12 OH
NH2
trans-D-erythro-sphingosine
Scheme 11.10 Application of fumarase in organic synthesis.
11.3.1.2 Malease and Citraconase

While fumaric acid is an integral part of our primary metabolism, maleic acid is of
much less importance. The interconversion of fumarate and malate catalyzed by
fumarase is a long studied example of a hydro-lyase catalyzed reaction. Malease only
occurs in few organisms and structural information is still scarce [46]. It is speculated
that it might act in a similar way to fumarase since it yields the opposite enantiomer of
malate, (R)-malate. It also does not contain a cofactor and it is relatively stable.
Malease has a very limited substrate spectrum; it essentially only accepts maleic acid
and citraconic acid, converting both of them into the (R)-alcohol (Scheme 11.11)
[46–48]. More recently an enzyme that was induced by citraconic acid and not by
malease/
citraconase
OOC COO + H 2O H COO
H
OH
H H OOC H
maleic acid (R)-malate

malease/
citraconase
OOC COO + H 2O H COO
H
OH
H CH 3 OOC CH 3
citraconate (R)-citramalate
Scheme 11.11 Malease and citraconase display similar stability and the same enantioselectivity in
the conversion of their two substrates.
476
maleic acid was described as a citraconase. As malease it accepts maleic acid and
citraconic acid as substrates and displays similar stability [49].
11.3.1.3 Aconitase
Aconitase (Acn) or citrate (isocitrate) hydro-lyase (EC 4.2.1.3) catalyzes the isomer-
ization of citrate and isocitrate via cis-aconitate in the citric acid cycle (Scheme 11.12).
The elimination of water from citrate yields cis-aconitate, which then flips inside the
active site to undergo addition of water to yield isocitrate [50]. Small quantities of cis-
aconitate can reversibly leak out of the active site; however, the enzyme was not
evolved to produce it but rather to isomerize citrate [51, 52]. Aconitases are
monomeric proteins that generally consist of four domains and an active [4Fe-4S]
cluster that reacts directly with the substrate. Domains 1–3 surround the [4Fe-4S]
cluster and domain 4 is connected by a long linker, thus forming a deep active-site
cleft together with the other three domains. When there is sufficient iron in the cell,
these aconitases assemble [4Fe-4S] clusters and function as an enzyme. During iron
starvation or oxidative stress, these aconitases dissemble the Fe-S cluster and the
resulting apo-aconitases function as site-specific mRNA binding proteins to enhance
transcription stability or block translation [53]. As [4Fe-4S] cluster containing
enzymes aconitases are very sensitive, they have not been employed very much.
citrate cis-aconitate
OH
aconitase OOC
OOC COO
COO -H2O
H COO H
COO
H*
flip in active site
OH
H
H COO COO
aconitase
+H2O OOC
COO
OOC COO
H*
(2R,3S)-isocitrate cis-aconitate
Scheme 11.12 Iron-sulfur-containing aconitase catalyzes the elimination and addition of water,
thus ensuring the isomerization of citrate to isocitrate.
11.3.1.4 Urocanase
The degradation of histidine to glutamate proceeds via urocanate [54]. Urocanase
(urocanate hydratase EC 4.2.1.49) catalyzes the conversion of urocanate into
imidazolone-5-propionate. The imidazole ring might bear a limited number of
substituents (Scheme 11.13) [55]. The reaction at first glance appears to be a
vinylogous-Michael addition of water with subsequent isomerization of the double
O uroconase OH O
+ H2 O
N O N O
NH NH
R R
R = H: urocanate R = H: imidazolone-5-propionate
R = F, Me, NH2 R = F, Me, NH 2
R CONH 2 Asp R CONH2 Asp

N H N H
O O
O O O O
H H
H H
H O H O
N O N O
NH N
H
R CONH2 Asp R CONH2 Asp

N N H
O
HO O O O
H H
OH O O
N O N O
N N
H H
Scheme 11.13 Urocanase catalyzes the addition of water in the histidine degradation pathway. The
reaction is catalyzed by NAD þ .
bond. Instead an intriguing NAD þ (required as the cofactor; NAD ¼ nicotinamide

adenine dinucleotide) catalyzed reaction proceeds [56, 57]. C4 of the imidazole ring
acts as nucleophile, attacking NAD þ and forming a transient carbon–carbon bond.
Double bond isomerization leads to an electron depleted imidazole ring susceptible
to attack of water. Finally, the product is liberated from NAD þ and released from the
active site.
11.3.1.5 Dihydroxy Acid Dehydratase

Dihydroxy acid dehydratase (DHAD, EC 4.2.1.9) catalyzes the dehydration and
tautomerization of 2,3-dihydroxycarboxylic acids to the corresponding 2-keto
acids, which is the third step in branched-chain amino acid biosynthesis [58, 59].
Dihydroxy-acid dehydratase from E. coli is a homodimer with a subunit molecular
mass of 66 kDa containing a [4Fe-4S] cluster, and it is part of the ILVD/EDD
superfamily (EDD or ILVD, 6-phosphogluconate dehydratase) [60]. Another dihy-
droxy-acid dehydratase from spinach is somewhat different: it is a homodimer with a
subunit of 63 kDa containing a [2Fe-2S] cluster [61]. In all cases the activity of DHAD
is dependent on Mg2 þ (for mechanism see below). The enzyme has been applied
successfully for the synthesis of 2-keto-3-deoxy sugar analogues (Scheme 11.14)
478
HO OH R OH R O
DHAD, -H2 O L-valine
R H H L-isoleucine
H 3C COO H3C COO H 3C COO
R = CH 3 or CH 2 CH 3
COO COO
OH OH COO OMe COO
DHAD, -H2 O
HO or HO O or O
MeO OH OH
CH2 OH CH2 OH
Scheme 11.14 DHAD catalyzes the elimination of water and the subsequent tautomerization. It
has been employed in the preparation of deoxy-sugar derivatives.
[62, 63]. Indeed this enzyme is also related to the 6-phosphogluconate dehydratase
discussed below. Along with DHAD, other dehydroxy acid dehydrates have been
identified that can convert tartrate and similar dihydroxy-acids [64].
11.3.1.6 Sugar Dehydratases

Dihydroxy-acids are also part of sugar metabolism. Oxidation of aldoses yields this
dihydroxy-acid moiety. It can be converted into an a-ketoacid via the action of a
dehydratase (Scheme 11.15a and b) [65, 66]. The resulting deoxysugar is now
ideally set up for a retro aldol reaction, splitting it in two. Thus, the breakdown is
preceded and induced by dehydration. In accordance with the large number of
sugars a vast array of dehydratases exists that catalyze this reaction. Even two
sequential dehydration reactions are possible and lead to the dicarbonyl
compound.
Most of these water eliminations are catalyzed by enzymes that are members of the
enolase superfamily [67]. These enzymes contain a divalent metal such as magne-
sium in their active site. Chelation of the acid group and the a-hydroxy function
allows deprotonation and enolization of the acid [68]. Water is expelled and tauto-
merization yields the final product (Scheme 11.16) [69]. In addition to the enolase
superfamily other enzyme families, such as the NAL (N-acetylneuraminate lyase)
superfamily and the ILVD/EDD superfamily, also catalyze this or similar steps in the
sugar metabolism [70–72]. A member of the NAL superfamily that utilizes an
enamine mechanism catalyzes the decarboxylation and dehydration step in the
breakdown of 5-keto-4-deoxy-D-glucarate (5-KDG) (Scheme 11.15b) [73].
11.3.1.7 2-Hydroxy-4-Dienoate Hydratases

In the degradation of aromatic compounds several very challenging steps have to be
catalyzed. Since the phenols, catechols and biphenyls do not contain many hydroxy-
functions these need to be added to increase solubility. Once the aromatic ring has
been broken via oxidative addition of hydroxyl groups, extra hydroxy functions are
CHO CHO CHO
Glucose oxidase H OH H OH H OH
EC 1.1.3.4/ Glucokinase
Glucose HO H EC 2.7.1.2 HO H HO H
1-dehydrogenase H OH H OH HO H
EC 1.1.1.118 ATP ADP H OH
+ H OH H OH
NADP
H C OH H C OPO3H H C OH
NADPH
H H H
D-Glucose Glucose 6-phosphate D-Galactose
OH
O O NAD(P)+ Glucose Glucose-6-phosphate NAD(P)
+
Galactose
dehydrogenase dehydrogenase 1-dehydrogenase
HO OH NAD(P)H EC 1.1.99.10 EC 1.1.149 NAD(P)H EC 1.1.1.48
OH
COO COO
D-Glucono-1,5-lactone H OH H OH OH
O O
HO H HO H OH
Glucono
lactonase H OH H OH
EC 3.1.1.17 H OH H OH HO OH
H C OH H C OPO3H
H H
Gluconate 6-Phosphogluconate D-Galactonic acid-γ-lactone
D-Gluconate 6-Phosphogluconate 1,4-Lactonase
dehydratase (EDD) dehydratase (EDD) EC 3.1.1.25
EC 4.2.1.39 EC 4.2.1.12
H2O H2O
COO COO COO

O O H OH
H H H H HO H
H OH H OH HO H
H OH H OH H OH
H C OH H C OPO3H H C OH
H H H
2-Keto-3-deoxy KDG kinase 2-Keto-3-deoxy D-Galactonate
gluconate (KDG) phosphogluconate
ATP ADP (KDPG) Galactonate
dehydratase
KDG aldolase KDPG aldolase (enolase)
EC 4.1.2.20 EC 4.1.2.14 EC 4.2.1.6
H2O
CHO CHO COO
H C OH H C OH O
H C OH H C OPO3H H H
H H HO H
Glyceraldehyde Glyceraldehyde-3-phosphate H OH
CH3 CH3 H C OH
C O C O H
CO2 CO2 2-Keto-3-deoxy-D-galaconate
(KDGal)
Pyruvate Pyruvate
KDGal aldolase
4.1.2.20
(a) Glyceraldehyde + Pyruvate
Scheme 11.15 (a) and (b) The dehydratases in sugar metabolism convert diols into ketones. These
ketones are susceptible to an aldol reaction.
480
COO COO CHO

H OH H OH L-Arabinose H OH
HO H HO H 1-dehydrogenase HO H L-Arabinose
EC 1.1.1.46 isomerase
H OH HO H NAD(P)+ HO H EC 5.3.1.4
H OH H OH H OH
NAD(P)H
COO COO H H2C OH
D-Glucarate D-Galactarate H O L-Arabinose O
O HO H
D-Galactarate OH H
dehydratase (enolase) HO H
EC 4.2.1.42/ H OH
D-Glucarate H OH
dehydratase (enolase) CH2OH H
EC 4.2.1.40 L-Arabinolactone L-Ribulose
H2O
L-Arabino Ribulokinase
lactonase EC 2.7.1.16
COO EC 3.1.1.15
OH
COO H2C OH
HO H
H OH O
H H
HO H HO H
O
HO H HO H
COO
H OH H OPO3H
5-keto-4-deoxy-D-glucarate (5-KDG)
H H
5-KDGGluc KDG dehydratase L-Arabonate L-Ribulose-phosphate
aldolase (NAL) EC 4.2.1.41
EC 4.1.2.20 L-Arabonate L-Ribulose-
dehydratase (EDD) phosphate
H2O EC 4.2.1.25 4-epimerase
EC 5.1.3.4
CO2 H2O
COO COO L-KDA COO H2C OH
dehydratase
H OH O (NAL) O O
CHO H H EC 4.2.1.43 H H H OH
Tartronate H H HO H HO H
semialdehyde H2O
O H OH H OPO3H
CH3 H H H
O α-Ketoglutaric L-2-Keto-3- D-Xylulose 5-phosphate
COO semialdehyde deoxyarabonate
(α-KGSA) (L-KDA)
Pyruvate
NAD(P)H α-KGSA 2-Keto-3-deoxy-
dehydrogenase L-arabonate aldolase
EC 1.2.1.26 EC 4.1.2.18 CH3COO-
NAD(P)+ Acetate
COO CHO
CHO
O CH2OH
H OH
H H Glycolaldehyde H OPO3H
H H
CH3 H
COO Glyceraldehyde
O
α-Ketoglutarate -3-phosphate
COO
Pyruvate
(b)
Scheme 11.15 (continued)
incorporated by addition of water. 2-Oxo-hept-3-ene-1,7-dioic acid hydratase (HpcG)

participates in the meta-fission pathway of homoprotocatechuic acid, the first
degradation product of 4-hydroxyphenylacetic acid [74]. HpcG catalyzes the addition
of water across the double bond to form 4-hydroxy-2-ketoheptane-1,7-dioate
Lys220 Lys 220
NH2 NH 3 2 O O
2 O O 2 O O O O
Mg Mg -OH Mg Mg
-H
O H O O O
H H H
H OH H OH H H H
R R R R
Scheme 11.16 Elimination of water from the dihydroxy-acid moiety of a sugar derivative catalyzed
by an enolase.
COO COO
HpcG BphH/MhpD
OH OH
+ H2 O +H 2O
HO COO O COO HO COO O COO
homoprotocatechuic 4-hydroxy-2-ketoheptane- 2-hydroxypent- 2-keto-4-hydroxypentanoate
acid 1,7-dioate 2,4-dienoate
Asp 79 Asp 79
Ala166 Ala166
O O O O
HN O HN O
H H
H O H O
O OH O
H
COO H COO COO
OOC OOC OOC
OH O OH O
H H
COO COO
H
OOC OOC H 2O
Scheme 11.17 2-Hydroxy-4-dienoate hydratases isomerize the double bond of their substrate and
then catalyze a subsequent Michael addition of water.
(Scheme 11.17). At first glance it might be assumed that the reaction should take
place in a vinylogous Michael addition relative to the acid group. However, then water
should add at a different position and this is not the case. Instead, kinetic analysis has
demonstrated that the mechanism proceeds via isomerization of 2-oxo-hept-4-ene-
1,7-dioate to its a,b-unsaturated ketone form followed by Michael addition of water to
this unsaturated ketone. Strictly speaking the ketone is here the activating group. The
enzymes thus catalyze an isomerization and the water addition. HpcG is metal-
dependent and a broad range of metal ions proved to be effective cofactors. HpcG
shows a slight preference for Mn2 þ , but Mg2 þ gives the best activity. Although the
metal is essential for the enzyme activity it is assumed not to be involved in the
482
mechanism of the enzyme (Scheme 11.17) [75, 76]. In light of the mechanism of
enolases this is somewhat surprising, as the metals there play a crucial role in
substrate activation (Scheme 11.16).
2-Hydroxypent-2,4-dienoate hydratase (HPDH or BphH) and 2-oxopent-4-dieno-
ate hydratase (2-hydroxypentadienoic acid hydratase) (OEH or MhpD, EC 4.2.1.80)
are homologous to HpcG [77]. MhpD catalyzes the hydroxylation of 2-oxopent-4-
dienoate to yield 4-hydroxy-2-oxovalerate via meta-cleavage in the phenylproprio-
nate degradation pathway and exhibits the highest activity with Mn2 þ as the
cofactor [78–80]. BphH catalyzes the formation of 2-keto-4-hydroxypentanoate from
2-hydroxypent-2,4-dienoate, which is part of the catabolic pathway of polychlorinated
biphenyls (PCBs), an environment pollutant. Like HpcG, BphH shows the highest
activity when using Mg2 þ as the cofactor [80].
11.3.1.8 Serine and Threonine Dehydratases

The elimination of water from an a,b-difunctionalized acid can not only occur
with diols but also with amino alcohols, such as in serine and threonine. In the
primary metabolism this is commonly catalyzed by pyridoxal 50 -phosphate (PLP)
enzymes and the enzymes are selective for the L-amino acids (Scheme 11.18).
However, D-selective PLP dependent enzymes also exist [81–85]. In addition to
these PLP dependent enzymes [4Fe-4S]-dependent L-serine/L-threonine dehydra-
tases have been identified [86–88]. All these enzymes have only been utilized
with their natural substrates and applications in synthesis have not been
described yet.
H OH
OOC OH OH
R R R
H3N H Lys Enz H OOC
OOC H H
L-serine/L-threonine NH NH H NH
H
2 O 2 O 2 O
O3PO O3PO O3PO
N CH3 N CH3 N CH3

H H H
H2O
R = H, Me
R
Lys Enz OOC
H
NH NH3 H2O NH
R
H
OOC 2 O 2 O
H + O3PO O3PO
O
N CH3 N CH3
H H
Scheme 11.18 Mechanism for the PLP-dependent deamination of L-serine (R ¼ H) and L-threonine
(R ¼ CH3) catalyzed by L-serine dehydratase [82].
11.3.1.9 Hydratase-Tautomerase Bifunctionality
As early as 1958 an enzyme from a Pseudomonas strain was identified that could add
water to acetylene dicarboxylate, followed by tautomerization and elimination of CO2,
yielding pyruvate. This enzyme was studied in more detail with acetylene mono-
carboxylate and malonic semialdehyde was isolated, clearly proving that water added
in Michael fashion [89, 90]. Almost 50 years later this study was repeated, employing
Pseudomonas putida (DSMZ ID 99–842) and utilizing several acetylenes. Unfortu-
nately, the enzyme was not isolated but clear evidence for the Michael addition of
water was provided (Scheme 11.19) [91].
Independent of this research, tautomerases were studied and two distinct enzymes
could be isolated from Pseudomonas strains that both add water to 3-chloroacrylic
acid, one to the cis compound the other to the trans isomer [92–95]. The structures of
both have been elucidated and they are closely related [96, 97]. Even more interest-
ingly both enzymes could add water to 2-oxo-3-pentynoate to yield acetopyruvate. This
is again a Michael addition of water to an acetylene with subsequent tautomerization
(Scheme 11.19) [98].
H 2O O
HO CO 2
O2C CO 2 +CO 2
O2 C H CO 2
O
H 2O HO CO 2
CO 2 CO 2
H
H
αGlu 52 α Glu52 OH O
O O O OH Cl O
H αArg8 αArg 8 A -HCl
O
O OH O
H O O
αArg 11 αArg 11
Cl O Cl O
H O
3-chloroacrylic H
H H H
N βPro 1 N βPro 1 -HCl
acid OH O
B
H O
αGlu52 αGlu 52
O O O OH
H OH O
O αArg 8 αArg 8
H O HO O H 3C CO 2
αArg 11 αArg11
H3 C
2-oxo-3-pentynoate CO 2 H3 C CO 2
H H βPro 1
H H
βPro1
O O
N N
H 3C CO 2
acetopyruvate
Scheme 11.19 Hydratases from Pseudomonas add water to acetylenic acids and the related
chloroacrylates.
484
11.3.2
The Activating Group is a Ketone
11.3.2.1 Dehydroquinase
In the primary metabolism of aromatic compounds shikimate is a key intermediate.
Thorough understanding of the shikimate pathway has allowed the fermentative
production of natural and unnatural compounds [99]. In particular Frost has
pioneered the synthesis of compounds such as phenol, vanillin, and catechol directly
from sugars via the shikimate pathway [100–103]. A key step in the biosynthesis is the
dehydration of dehydroquinate catalyzed by dehydroquinase (EC 4.2.1.10), leading to
dehydroshikimate [104]. Interestingly, two structurally completely different dehy-
droquinases exist [105, 106]. Type I dehydroquinase catalyzes syn elimination
whereas type II dehydroquinase catalyzes anti elimination [107–110]. It is thought
that type I, which acts via an enamine intermediate, evolved first and has only an
anabolic function [111]. Type II might have evolved as part of a catabolic pathway and
only later was utilized in the biosynthesis of aromatic compounds [112]. It depro-
tonates HS in alpha position to the ketone (Scheme 11.20) [113, 114].
Enz Enz
HS Lys 170 Lys 170
HO HS HO
HO NH NH
OH TypeI
O2 C 1
2
3 O2 C O2 C
HR O
OH HO OH H R HO OH
H
B
B
HO COO CO 2 Enz
H Lys 170
TypeII HO NH
O OH O OH O2 C
OH OH OH
dehydroquinate dehydroshikimate B
B
CO 2
HO HS HO H 1
OH OH 2
O2 C O2 C
HR O
OH HR O
3
OH O OH
H OH
A
Scheme 11.20 Dehydroquinase types I and II have very different mechanisms.
11.3.2.2 Scytalone Dehydratase

Scytalone is an intermediate in melanin biosynthesis of the fungus Magnaporthe
grisea. This fungus causes the blast disease in rice and interruption of the fungal
metabolism is therefore a prime target in the development of efficient fungicides to
fight this blight. Scytalone dehydratase (EC. 4.2.1.94) is a target in this research [115].
Scytalone dehydratase catalyzes syn elimination via two sequential steps, most likely
by an E1cb mechanism (Scheme 11.21) [116–119]. Along with scytalone the enzyme
accepts vermelone and 2,3-dihydro-2,5-dihydroxy-4H-benzopyran-4-one [120]. The
latter is structurally related to the 2-hydroxyisoflavanones that are intermediates in
the biosynthesis of daidzein and genistein [121]. The dehydratase catalyzing this
biosynthesis, 2-hydroxy-isoflavanone dehydratase, has a similar mechanism to that of
scytalone dehydratase.
OH
OH O OH O OH OH OH O
8 1 1 8
7 2 2 7
6 3 3 6
5 4 4 5
HO OH HO HO HO O OH
H H
scytalone 1,3,8-trihydroxynaphthalene 2-hydroxy-isoflavanone
H H H
N N N
His 110 His 110 His110
N N N
H H H
HO O His 85 HO O His 85 HO O His85
N HN H N
H N N N
H H H H
H
O O O O O O
O O O O O O
H H H H H H
O Asp32 O Asp32 O Asp 32
H H H H H H
Tyr 50 O O Tyr 50 O O Tyr50 O O
H Tyr30 H Tyr 30 H Tyr 30
Scheme 11.21 Scytalone dehydratase also accepts vermelone and flavanone related compounds;
2-hydroxyisoflavanone dehydratase displays a similar mechanism to that of scytalone dehydrase in
the elimination of water from 2-hydroxyisoflavanone.
11.3.2.3 1,5-Anhydro-D-Fructose Dehydratase and Aldos-2-Ulose Dehydratase

1,5-Anhydro-D-fructose dehydratase (AFDH, EC 4.2.1.111) from the fungal species
Anthracobia melaloma is found on the anhydrofructose pathway of the degradation of
glycogen and starch [122–125]. AFDH catalyzes the dehydration of 1,5-anhydro-D-
fructose (1,5AnFru) to 1,5-anhydro-D-glycero-hex-3-en-2-ulose (ascopyrone M, APM),
which is further isomerized to 1,5-anhydro-4-deoxy-D-glycerohex-1-en-3-ulose (asco-
pyrone P, APP) by ascopyrone tautomerase (APTM) [126].
The same reaction is also catalyzed by aldos-2-ulose dehydratase (AUDH, EC
4.2.1.110) from the fungus Phanerochaete chrysosporium. This is, however, a
bifunctional enzyme that in addition to the dehydration of 1,5-AnFru also
catalyzed isomerization to microthecin [127]. Microthecin has antimicrobial
properties and this enzymatic route has been explored for production
(Scheme 11.22) [125].
486
Glycogen/ α-1,4-Glucon lyase
α-1,4-Glucan lyase EC 4.2.1.13
CH2 OH
O
OH
OH O
anhydrofructose(1,5AnFru)
AF dehydratase (AFDH) EC 4.2.1.111

or
H 2O Aldos-2-ulose dehydratase (AUDH)
EC 4.2.1.110
CH2 OH
O
HO O
ascopyrone M (APM)
Dehydratase APM tautomerase AUDH
H 2O EC 4.2.1.- EC 5.3.3.15 EC 4.2.1.110
H CH 2OH
OH O
OH O
OH
CH2 OH O CH 2OH
O O OH
5-epipentenomycin ascopyrone P (APP) microthecin
Scheme 11.22 Two different dehydratases catalyze the elimination of water from anhydrofructose.
11.3.3
The Activating Group is a Thioester
11.3.3.1 Fatty Acid Biosynthesis

The prime example for the elimination of water from a thioester is the biosynthesis of
fatty acids (Scheme 11.23) [128]. In this vital part of our primary metabolism acetyl
CoA is activated to malonyl CoA and transferred onto the acyl carrier protein (ACP).
All the enzymes involved in fatty acid biosynthesis recognize their substrate when
linked to this ACP. After a Claisen condensation and selective reduction, water is
eliminated via a syn-elimination from the (R)-3-hydroxy-acid [129]. Owing to the
thioester the pKa of the a-CH is about 20 and the elimination occurs without the help
of a cofactor. Subsequent reduction then closes the cycle. The product can either
undergo a further chain extension or, once the correct chain length is reached, can be
cleaved from ACP by a thioesterase. The enzymes that catalyze this overall process
together form the fatty acid synthase (FAS) [130]. FAS type I is one huge protein that
accommodates all enzyme activities and weight approximately 250 kDa. FAS I is only
O O trans O O
acylase
HO SCoA HO ACP
malonyl CoA O O
NADPH
CO2 , biotin R ACP
acetyl-CoA ketoacyl ACP synthase NADP
carboxylase 3-ketoacyl ACP
O
reductase
O
R SCys 161
SCoA
acetyl CoA OH O
R ACP
H
O sy n-elimination
(R)-hydroxyacetyl
R ACP ACP dehydrase
thioesterase
enoylacyl ACP
reductase O
H2 O
fatty acid NADP R ACP
NADPH
Scheme 11.23 Biosynthesis of fatty acids proceeds via a Claisen-type condensation, reduction, an
essential syn-elimination of water, and a second reduction. This sequence is performed repetitively
until the right chain length is obtained.
active as a dimer of these huge proteins and is present in mammalian and higher
organisms except plants. Plant and bacteria have FAS type II, which consists of
separate enzymes. However, mammalian mitochondria also contain FAS II; thus
animals have both FAS I in the cytosol and FAS II in the mitochondria. Both types of
FAS act similarly and the active sites resemble each other. Since the final product in
each case is a fatty acid one might even conclude that the stereochemistry is of no
great interest. However, polyketides are produced via modified FA synthases, the
polyketide synthases [131, 132]. Since polyketides do contain a vast variety of
stereocenters the biosynthesis of fatty acids and its stereochemistry has gained great
interest in this field of secondary metabolites research. Several excellent reviews can
be recommended [133–138].
In the FAS II system, the genes fabA and fabZ both encode b-hydroxyacyl-acyl
carrier protein (ACP) dehydratases [139]. FabA is a bifunctional enzyme only found in
Gram-negative bacteria catalyzing the reversible dehydration of b-hydroxyacyl-ACP
and the isomerization of trans-2-decenoyl-ACP to cis-3-decenolyl-ACP [140]. FabZ is
ubiquitous, catalyzing the reversible dehydration of (3R)-b-hydroxyacyl-ACP to trans-
2-acyl-ACP but lacks the ability to catalyze the isomerization reaction. FabA is
essentially inactive in the elongation cycles of unsaturated fatty acid biosynthesis
while FabZ is involved in both the unsaturated and saturated fatty acid biosynthesis
and more active on long-chain saturated acyl-ACP than FabA [141]. The structures of
488
FabA and FabZ are very similar and both are homodimeric proteins adopting a b þ a
hot dog-fold in which the antiparallel b-sheet wraps the a-helix like a bun wrapping
a sausage [142]. Two active sites are formed along the dimer interface, with the critical
active site residues contributed by opposite monomers. The catalytic residues in the
active site of FabA are a histidine and an aspartate whereas in the active site of FabZ
they are a histidine and a glutamate [143]. The minor difference in the shapes of
the active site tunnels prevents the adoption of a cis-3 conformation in FabZ. At the
bottom of the tunnel of FabZ, the conformational change of a flexible phenylalanine
can either close or open the tunnel to accommodate short or long substrates.
However, this lid does not exist in FabA and thus only short substrates can be
R β α OH
Acyl-CoA ligase
EC 6.2.1.3
O
O O
α SCoA FAD
R SCoA R β α SCoA
FADH 2
CoA
3-ketoacyl-CoA Acyl-CoA
thiolase dehydrogenase
EC 2.3.1.16 EC 1.3.99.-
O O O
R β SCoA R β SCoA
α α
(3S)-Hydroxyacyl- Enoyl-CoA
CoA dehydrogenase H2 O hydratase 1
NADH+H+ EC 1.1.1.35 EC 4.2.1.17/
∆3,∆2-enoyl-CoA
OH O isomerase
EC 5.3.3.8
NAD+ R β SCoA
α
Epimerase H 2O
EC 5.1.2.-
(3R)-Hydroxyacyl- Enoyl-CoA
CoA dehydrogenase OH O hydratase 2
EC 1.1.1.36 EC 4.2.1.B3
R β α SCoA
PHA synthase
CoASH EC 2.3.1.-
Polyhydroxyalkanoates (PHAs)
Scheme 11.24 Hydration reactions in the b-oxidation of fatty acids occur enantioselectively to yield
either the (S) or the (R) enantiomer.
Ala98
N H CoA
S
O H
N H H R
Gly 141 H O
H
Ala98 O O Ala98
N H CoA O O N H CoA
S Glu164 Glu144 S
O H concerted transition-state O H
N H H R N H H R
O H O
Gly 141 Ala98 Gly 141
H H N H CoA H
O O O O
O O S O O
Glu164 Glu144 O H Glu 164 Glu144
N H H R
O (S)-3-hydroxy-carboxylic CoA
Gly 141
H H
O O
O O
Glu 164 Glu144
carbanion intermediate
Scheme 11.25 Reaction mechanism of (S)-specific enoyl-CoA hydratase (enol-CoA hydratase 1).
bound [144, 145]. In FAS I, the dehydratase subdomain (DH) is also a hotdog-fold but,
unlike FAS II dehydratases, two DHs interact to form a single active site [139].
11.3.3.2 Fatty Acid Degradation, b-Oxidation

An enzyme is a catalyst and thus it helps to enable rapid establishment of the reaction
equilibrium. It might therefore be assumed that the breakdown of fatty acids occurs
by reversal of their biosynthesis (Scheme 11.23). This is, however, not the case.
Instead, nature has developed a second set of enzymes that are specialized in the
breakdown of fatty acids; this is also known as b-oxidation (Scheme 11.24). Unlike the
enzymes involved in the biosynthesis of fatty acids two different sets of enzymes,
multifunctional enzyme type 1 (MFE-1) and multifunctional enzyme type 2 (MFE-2)
are involved in the fatty acid degradation [146]. MFE-1 degrades straight-chain fatty
acids and MFE-2 degrades very long chain and 2-methyl-branched fatty acids, thus
linking the b-oxidation and the biosynthesis of bile acids and polyhydroxyalkanoates
(PHAs) [147–149]. In MFE-1, there is a bifunctional D3, D2-enoyl-CoA isomerase/2-
enoyl-CoA hydratase 1, which is (S)-selective, and in MFE-2 an (R)-selective 2-enoyl-
CoA hydratase 2 is found [150].
The (S)-specific 2-enoyl-CoA hydratase 1 (ECH 1, EC 4.2.1.17), also known as
crotonase, is the second enzyme in the b-oxidation pathway of fatty acids and
catalyzes syn-addition of a water molecule to the double bond of a trans-2-enoyl-CoA
thioester, forming (S)-3-hydroxyacyl-CoA [151–156]. ECH 1 belongs to the low-
similarity isomerase/hydratase superfamily [157–159]. It forms a homohexamer
built up from two trimeric disks. The molecular mass of one subunit is approximately
29 kDa. ECH 1 has very broad substrate specificity, from C4 to C20 [160]. A flexible
loop (residues 113–120 in rat liver mitochondrial crotonase, PDB entry 1DUB) in the
490j 11 Addition of Water to C¼C Bonds and its Elimination
binding pocket can move and open a tunnel to accommodate longer fatty acid
tails [161].
Two structurally conserved backbone NH groups (e.g., Ala98 and Gly141 in rat liver
mitochondrial crotonase) form an oxyanion hole that stabilizes an enolate anion
intermediate of acyl-CoA substrates. In addition, G141 is located at the N-terminus of
an a-helix. The positive end of the helix dipole moment contributes to the polari-
zation of the carbonyl group of the thioester [162, 163]. During the hydration reaction,
Glu164 protonates C2 of the substrate while Glu144 activates water so that it adds at
C3 to the double bond of the substrate. Both a concerted and a sequential mechanism
have been suggested (Scheme 11.25).
(R)-Specific enoyl-CoA hydratase 2 catalyzes the reversible syn-addition of water to
the C¼C double bond of trans-2-enoyl-CoA thioester with the opposite chiral
specificity to 2-enoyl-CoA hydratase 1 [164]. Bacterial hydratases 2 are about 15 kDa
in weight and are specific for short chain length enoyl-CoA. Eukaryotes enzymes, in
contrast, are about twice as big as bacterial enzymes and specific for very long chain
Ile828 Ile 828

O O
H H
His 813 N His 813 N
N N
SCoA SCoA
H H
H O H O
H δ H
Asp808 O N δ O
N
COO H Gly831 COO H
R H R H
Asp808 Gly831
H H
Asn810 N Asn810 N
H H
O O
Ile 828
O
H
His813 N
N
SCoA
H
H O
H
Asp808 O N
COO H R H
Gly 831
H
Asn 810 N
H
O
Scheme 11.26 Reaction mechanism of (R)-specific enoyl-CoA hydratase (enol-CoA hydratase 2).
OOC Glu 143
Tyr239 Tyr239 H H
OH OH O
O SCoA O SCoA
OH OH
Tyr75 MeO O Tyr 75 MeO O
OOC Glu143 HOOC Glu 143

Tyr239 H Tyr239
OH O OH OH
O SCoA O SCoA
OH OH
Tyr75 MeO O Tyr 75 MeO O
HOOC Glu143
HO CHO SCoA HO CHO +acetylSCoA
MeO O MeO
Scheme 11.27 Hydroxycinnamoyl-CoA hydratase lyase (HCHL, formerly feruloyl-CoA hydratase)

catalyzes the addition of water and the subsequent retro-Perkin reaction, breaking down feruloyl-
CoA to vanillin and acetyl-CoA.
lengths [152]. Similar to FabA and FabZ, the structure of hydratase 2 monomer is a
hotdog fold. Two monomers associate side by side to form a functional dimer [151].
However, the two catalytically important residues that form the dyad are from the
same monomer of hydratase 2. Two active sites are formed at the interface of the two
monomers. Each active site is located deep in the substrate-binding tunnel which is
constituted by both monomers. The hydration/dehydration reactions are suggested
to proceed via a concerted transition state that is similar to that found in the
mechanism of hydratase 1 (Scheme 11.26). Because the active site geometry of
hydratase 2 is opposite to that of hydratase 1, the stereochemistry of these two
enzymes is opposite [164].
11.3.3.3 Hydroxycinnamoyl-CoA Hydratase Lyase (HCHL)

An enzyme that has attracted particular attention in recent years is feruloyl-CoA
hydratase. This enzyme is involved in ferulic acid degradation and hydrates feruloyl-
CoA to form 4-hydroxy-3-methoxyphenyl-b-hydroxypropionyl–CoA followed by
cleavage to produce vanillin and acetyl-CoA [165, 166]. These two steps are of great
importance since natural ferulic acid is abundantly available from corn hulls while
the product of this sequence would be natural vanillin – even though it was never
close to a vanilla pod [167]. In Vanilla planifolia the same reaction steps occur as part of
492
the biosynthesis of this precious natural fragrance compound. Recently, the mech-
anism of feruloyl CoA hydratase was elucidated and this led to the new name:
hydroxycinnamoyl-CoA hydratase lyase (HCHL) [168]. The enzyme does catalyze the
addition of water to an a,b-unsaturated thioester; however, this occurs via a quinine-
methide enolate. The thioester then is essential in the retro-Perkin condensation that
splits the molecule into the desired vanillin and acetyl-CoA (Scheme 11.27).
11.4
Outlook
Based on current knowledge the application of hydro-lyases has to be judged as

limited. The enzymes are still seen as curiosities rather than workhorses for the
organic chemist. This is mainly due to the limited accessibility for the chemist and to
the often limited substrate spectrum. Nonetheless, a bright future is on the horizon.
With recent advances in biochemistry and enzyme production the hydro-lyases will
come into reach of all chemists (Chapter 3). The huge advances made in enzyme
modification, both to improve stability and substrate range are well described in
Chapters 4 and 5, which to date have not been applied to hydro-lyases but to the well
investigated hydrolases. However, now that the methodologies for enzyme produc-
tion and modification have advanced so far the time is ripe for organic chemists to
step in and utilize the power of these enzymes to solve synthetic problems. Indeed,
the potential of these enzymes is already obvious. Industrial processes based on
selected hydro-lyases are already well established and running.
References
1 Hahn, H.-D., D€
ambkes, G., 5 Babbitt, P.C., Hasson, M.S.,
Rupprich, N., and Bahl, H. (2010) Wedekind, J.E., Palmer, D.R.,
Butanols, in Ullmanns Encyclopedia Barrett, W.C., Reed, G.H., Rayment, I.,
of Industrial Chemistry, Electronic Ringe, D., Kenyon, G.L., and Gerlt, J.A.
version, Wiley-VCH Verlag GmbH, (1996) The enolase superfamily: a general
Weinheim. strategy for enzyme-catalyzed abstraction
2 Davey, P.N., Earle, M.J., Hamill, J.T., of the alpha-protons of carboxylic acids.
Katdare, S.P., Rooney, D.W., and Biochemistry, 35, 16489–16501.
Seddon, K.R. (2010) Selective hydration 6 Frey, P.A. and Hegeman, A.D. (2007)
of dihydromyrcene in ionic liquids. Green Addition and elimination, in Enzymatic
Chem., 12, 628–631. reaction mechanisms, Oxford University
3 van der Werf, M.J., van den Tweel, W.J.J., Press, New York, pp. 433–475.
Kamphuis, J., Hartmans, S., and 7 Kisic, A., Miura, Y., and SchroepherJr.,
de Bont, J.A.M. (1994) The potential of G.J. (1971) Oleate hydratase: studies of
lyases for the industrial production of substrate specificity. Lipids, 6, 541–545.
optically active compounds. Trends 8 Bevers, L.E., Pinkse, M.W.H., Verhaert,
Biotechnol., 12, 95–103. P.D.E.M., and Hagen, W.R. (2009) Oleate
4 Tokoroyama, T. (2010) Discovery of the hydratase catalyzes the hydration of a
Michael reaction. Eur. J. Org. Chem., nonactivated carbon-carbon bond.
2009–2016. J. Bacteriol., 191, 5010–5012.
References j493
9 Gocho, S., Tabogami, N., Inagaki, M., acetylene hydratase of Pelobacter
Kawabata, C., and Komai, T. (1995) acetylenicus, a tungsten iron-sulfur
Biotransformation of oleic acid to protein. J. Bacteriol., 177, 5767–5772.
optically active c-dodecalactone. Biosci. 19 Meckenstock, R.U., Krieger, R.,
Biotechnol. Biochem., 59, 1571–1572. Ensign, S., Kroneck, P.M.H. and
10 Wanikawa, A., Hosoi, K., Takise, I., and Schink, B. (1999) Acetylene hydratase of
Kato, T. (2000) Detection of c-lactones in Pelobacter acetylenicus: molecular and
malt whisky. J. Inst. Brew., 106, 39–43. spectroscopic properties of the tungsten
11 Umeno, D., Tobias, A.V., and Arnold, iron-sulfur enzyme. Eur. J. Biochem., 264,
F.H. (2005) Diversifying carotenoid 176–182.
biosynthetic pathways by directed 20 Seiffert, G.B., Ullmann, G.M.,
evolution. Microbiol. Mol. Biol. Rev., 69, Messerschmidt, A., Schink, B.,
51–78. Kroneck, P.M.H., and Einsle, O. (2007)
12 Steiger, S., Takaichi, S., and Structure of the non-redox-active
Sandmann, G. (2002) Heterologous tungsten/[4Fe:4S] enzyme acetylene
production of two unusual acyclic hydratase. Proc. Natl. Acad. Sci. U.S.A.,
carotenoids, 1,10 -dihydroxy-3,4- 104, 3073–3077.
didehydrolycopene and 1-hydroxy- 21 Banerjee, R. and Ragsdale, S.W. (2003)
3,4,30 ,40 -tetradehydrolycopene by The many faces of vitamin B12: catalysis
combination of the crtC and crtD genes by cobalamin-dependent enzymes. Annu.
from Rhodobacter and Rubrivivax. J. Rev. Biochem., 72, 209–247.
Biotechnol., 97, 51–58. 22 Bachovchin, W.W., Eagar R.G. Jr.,
13 Steiger, S., Mazet, A., and Sandmann, G. Moore, K.W., and Richards, J.H. (1977)
(2003) Heterologous expression, Mechanism of action of
purification and enzymatic adenosylcobalamin: glycerol and other
characterization of the acyclic carotenoid substrate analogues as substrates and
1,2-hydratase from Rubrivivax. Arch. inactivators for propanediol dehydratase-
Biochem. Biophys., 414, 51–58. kinetics, stereospecificity, and
14 Sun, Z., Shen, S., Wang, C., Wang, H., mechanism. Biochemistry, 16, 1082–1092.
Hu, Y., Jiao, J., Ma, T., Tian, B., and 23 Hartmanis, M.G.N. and Stadtman, T.C.
Hua, Y. (2009) A novel carotenoid (1986) Diol metabolism and diol
1,2-hydratase (CruF) from two species of dehydratase in Clostridium glycolicum.
the non-photosynthetic bacterium Arch. Biochem. Biophys., 245, 144–152.
Deinococcus. Microbiology, 155, 24 Daniel, R., Bobik, T.A., and Gottschalk, G.
2775–2783. (1999) Biochemistry of coenzyme B12-
15 Li, D., Chung, D.-R., Smith, D.A., and dependent glycerol and diol dehydratases
Schardl, C.L. (1995) The Fusarium solani and organization of the encoding genes.
gene encoding kievitone hydratase, a FEMS Microbiol. Rev., 22, 553–566.
secreted enzyme that catalyzes 25 Toraya, T. (2002) Enzymatic radical
detoxification of a bean phytoalexin. Mol. catalysis: coenzyme B12-dependent diol
Plant-Microbe Interact., 8, 388–397. dehydratase. Chem. Rec., 2, 352–366.
16 Turbek, C.S., Smith, D.A., and 26 Sandala, G.M., Kovacevic, B., Bric, D.,
Schardl, C.L. (1992) An extracellular Smith, D.M. and Radom, L. (2009) On the
enzyme from Fusarium solani f. sp. reaction of glycerol dehydratase with but-
phytoalexin, phaseollidin. FEMS 3-ene1,2-diol. Chem. Eur. J., 15,
Microbiol. Lett., 94, 187–190. 4865–4873.
17 Turbek, C.S., Li, D.G., Choi, H.C., 27 Celi
nska, E. (2010) Debottlenecking the
Schardl, C.L., and Smith, D.A. (1990) 1,3-propanediol pathway by metabolic
Induction and purification of kievitione engineering. Biotechnol. Adv., 28,
hydratase from Fusarium solani f. sp. 519–530.
phaseoli. Phytochemistry, 29, 2841–2846. 28 Behr, A., Eilting, J., Irawadi, K.,
18 Rosner, B.M. and Schink, B. (1995) Leschinski, J. and Lindner, F. (2008)
Purification and characterization of Improved utilisation of renewable
494
resources: new important derivatives of Escherichia coli. Biochemistry, 35,

glycerol. Green Chem., 10, 13–30. 13955–13965.
29 Lamartiniere, C.A., Braymer, H.D., and 39 Weaver, T., Lees, M., and Banaszak, L.
Larson, A.D. (1970) Purification and (1997) Mutations of fumarase that
characterization of fumarase from distinguish between the active site and a
Pseudomonas putida. Arch. Biochem. nearby dicarboxylic acid binding site.
Biophys., 141, 293–302. Protein Sci., 6, 834–842.
30 Flint, D.H. and Allen, R.M. (1996) 40 Rose, I.A. and Weaver, T.M. (2004) The
Iron-sulfur proteins with nonredox role of the allosteric B site in the fumarase
functions. Chem. Rev., 96, 2315–2334. reaction. Proc. Natl. Acad. Sci. U.S.A.,
31 Yuji, U., Yumoto, N., Tokushige, M., 101, 3393–3397.
Fukui, K., and Ohya-Nishiguchi, H. 41 Weaver, T. (2005) Structure of free
(1991) Purification and characterization fumarase C from Escherichia coli. Acta
of two types of fumarases from Crystallogr. Sect. D, 61, 1395–1401.
Escherichia coli. J. Biochem., 109, 728–733. 42 Yumoto, N. and Tokushige, M. (1988)
32 Flint, D.H. (1994) Initial kinetic and Characterization of multiple fumarase
mechanistic characterization of proteins in Escherichia coli. Biochim.
Escherichia coli fumarase A. Arch. Biophys. Res. Commun., 153, 1236–1243.
Biochem. Biophys., 311, 509–516. 43 Ueda, Y., Yumoto, N., Tokushige, M.,
33 Flint, D.H. (1993) Escherichia coli Fukui, K., and Ohya-Nishiguchi, H.J.
fumarase A catalyzes the isomerization of (1991) Purification and characterization
enol and keto oxalacetic acid. of two types of fumarase from Escherichia
Biochemistry, 32, 799–805. coli. J. Biochem., 109, 728–733.
34 Woods, S.A., Schwartzbach, S.D., and 44 Genda, T., Watabe, S., and Ozaki, H.
Guest, J.R. (1988) Two biochemically (2006) Purification and characterization
distinct classes of fumarases in of fumarase from Corynebacterium
Escherichia coli. Biochim. Biophys. Acta, glutamicum. Biosci. Biotechnol. Biochem.,
954, 14–26. 70, 1102–1109.
35 Flint, D.H., Emptage, M.H., and Guest, 45 Findeis, M.A. and Whitesides, G.M.
J.R. (1992) Fumarase A from Escherichia (1987) Fumarase-catalyzed synthesis of
coli: purification and characterization as L-threo-chloromalic acid and its
an iron-sulfur cluster containing conversion to 2-deoxy-D-ribose and
enzyme. Biochemistry, 31, 10331–10337. D-erythro-sphingosine. J. Org. Chem., 52,
36 Weaver, T., Lees, M., Zaitsev, V., 2838–2848.
Zaitseva, I., Duke, E., Lindley, P., 46 van der Werf, M.J., van den Tweel, W.J.J.,
McSweeny, S., Svensson, A., and Hartmans, S. (1992) Screening for
Keruchenko, J., Keruchenko, I., microorganisms producing D-malate
Gladilin, K., and Banaszak, L. (1998) from maleate. Appl. Environ. Microbiol.,
Crystal structures of native and 58, 2854–2860.
recombinant yeast fumarase. J. Mol. Biol., 47 van der Werf, M.J., van den Tweel, W.J.J.,
280, 431–442. and Hartmans, S. (1993) Purification and
37 Yang, J., Wang, Y., Woolridge, E.M., characterization of maleate hydratase
Arora, V., Petsko, G.A., Kozarich, J.W., from Pseudomonas pseudoalcaligenes.
and Ringe, D. (2004) Crystal Appl. Environ. Microbiol., 59, 2823–2829.
structure of 3-carboxy-cis,cis-muconate 48 van der Werf, M.J., van den Tweel, W.J.J.,
lactonizing enzyme from Pseudomonas and Hartmans, S. (1993)
putida, a fumarase class II type Thermodynamics of the maleate and
cycloisomerase: enzyme evolution in citraconate hydration reactions catalysed
parallel pathways. Biochemistry, 43, by malease from Pseudomonas
10424–10434. pseudoalcaligenes. Eur. J. Biochem., 217,
38 Weaver, T. and Banaszak, L. (1996) 1011–1017.
Crystallographic studies of the catalytic 49 He, B.-F., Nakajima-Kambe, T., Ozawa,
and a second site in fumarase C from T., and Nakahara, T. (2000) Production of
References j495
D-malate and D-citratemalate by a,b-dihydroxyacid dehydratase. J. Am.
Arthrobacter pascens DMDC12 having Chem. Soc., 113, 1020–1025.
stable citraconase. Process Biochem., 36, 59 Kim, S. and Lee, S.B. (2006) Catalytic
407–414. promiscuity in dihydroxy-acid
50 Beinert, H., Kennedy, M.C., and dehydratase from the thermoacidophilic
Stout, C.D. (1996) Aconitase as archaeon Sulfolobus solfataricus. J.
iron-sulfur protein, enzyme, and Biochem., 139, 591–596.
iron-regulatory protein. Chem. Rev., 60 Flint, D.H., Emptage, M.H.,
96, 2335–2373. Finnegan, M.G., Fu, W., and
51 Kennedy, M.C., Mende-Mueller, L., Johnson, M.K. (1993) The role and
Blondin, G.A., and Beinert, H. (1992) properties of the iron-sulfur cluster in
Purification and characterization of Escherichia coli dihydroxy-acid
cytosolic aconitase from beef liver and its dehydratase. J. Biol. Chem., 268,
relationship to the iron-responsive 14732–14742.
element binding protein. Proc. Natl. 61 Flint, D.H. and Emptage, M.H. (1988)
Acad. Sci. U.S.A., 89, 11730–11734. Dihydroxy acid dehydratase from spinach
52 Haile, D.J., Ronault, T.A., Harford, J.B., contains a [2Fe-2S] cluster. J. Biol. Chem.,
Kennedy, M.C., Blondin, G.A., 263, 3558–3564.
Beinert, H., and Klausner, R.D. (1992) 62 Limberg, G., Klaffke, W., and
Cellular regulation of the iron-responsive Thiem, J. (1995) Conversion of aldonic
element binding protein: disassembly of acids to their corresponding 2-keto-3-
the cubane iron-sulfur cluster results in deoxy-analogs by the non-carbohydrate
high-affinity RNA binding. Proc. Natl. enzyme dihydroxy acid dehydratase
Acad. Sci. U.S.A., 89, 11735–11739. (DHAD). Bioorg. Med. Chem., 3, 487–494.
53 Williams, C.H., Stillman, T.J., 63 Limberg, G. and Thiem, J. (1996)
Barynin, V.V., Sedelnikova, S.E., Tang, Y., Synthesis of modified aldonic acids and
Green, J., Guest, J.R., and Artymiuk, P.J. studies of their substrate efficiency for
(2002) E. coli aconitase B structure reveals dihydroxy acid dehydratase (DHAD).
a HEAT-like domain with implications Aust. J. Chem., 49, 349–356.
for protein-protein recognition. Nat. 64 Reney, S.K., Begg, C., Bungrad, S.J., and
Struct. Biol., 9, 447–452. Guest, J.R. (1993) Identification of the
54 Tyagi, R., Eswaramoorthy, S., Burley, L-tartrate dehydratase genes (ttdA and
S.K., Raushel, F.M., and Swaminathan, S. ttdB) of Escherichia coli and evolutionary
(2008) A common catalytic mechanism relationship with the class I fumarase
for proteins of the HutI family. genes. J. Gen. Microbiol., 139, 1523–1530.
Biochemistry, 47, 5608–5615. 65 Ahmed, H., Ettema, T.J.G., Tjaden, B.,
55 Gerlinger, E., Hull, W.E. and Retey, J. Geerling, A.C.M., van der Oost, J., and
(1983) Mechanistic studies of the Siebers, B. (2005) The semi-
urocanase reaction using 1 H- and phosphorylative Entner–Doudoroff
31
P-NMR spectroscopy and the substrate pathway in hyperthermophilic archaea - a
analogue 2-methylurocanate. re-evaluation. Biochem. J., 390, 529–540.
Tetrahedron, 39, 3523–3528. 66 Watanabe, S., Shimada, N., Tajima, K.,
56 Grueninger, D., Treiber, N., Ziegler, Kodaki, T. and Makino, K. (2006)
M.O.P., Koetter, J.W.A., Schulze, M.-S., Identification and characterization of
and Schulz, G.E. (2008) Designed L-arabonate dehydratase, L-2-keto-3-
protein-protein association. Science, 319, deoxyarabonate dehydratase, and
206–209. L-arabinolactonase involved in an
57 Kessler, D., Retey, J. and Schulz, G.E. alternative pathway of L-arabinose
(2004) Structure and action of urocanase. metabolism: Novel evolutionary insight
J. Mol. Biol., 342, 183–194. into sugar metabolism. J. Biol. Chem.,
58 Pirrung, M.C., Holmes, C.P., 281, 33521–33536.
Horowitz, D.M., and Nunn, D.S. (1991) 67 Sakai, A., Fedorov, A.A., Fedorov, E.V.,
Mechanism and stereochemistry of Schnoes, A.M., Glasner, M.E., Brown, S.,
496
Rutter, M.E., Bain, K., Chang, S., Gheyi, 75 Izumi, A., Rea, D., Adachi, T., Unzai, S.,
T., Sauder, J.M., Burley, S.K., Babbitt, Park, S.Y., Roper, D.I., and Tame, J.R.H.
P.C., Almo, S.C., and Gerlt, J.A. (2009) (2007) Structure and mechanism of
Evolution of enzymatic activities in the HpcG, a hydratase in the
enolase superfamily: stereochemically homoprotocatechuate degradation
distinct mechanisms in two families of pathway of Escherichia coli. J. Mol. Biol.,
cis,cis-muconate lactonizing enzymes. 370, 899–911.
Biochemistry, 48, 1445–1453. 76 Horsman, G.P., Ke, J., Dai, S.,
68 Sims, P.A., Larsen, T.M., Poyner, R.R., Seah, S.Y.K., Bolin, J.T., and Eltis, L.D.
Cleland, W.W., and Reed, G.H. (2003) (2006) Kinetic and structural insight into
Reverse protonation is the key to general the mechanism of BphD, a C-C bond
acid-base catalysis in enolase. hydrolase from the biphenyl degradation
Biochemistry, 42, 8298–8306. pathway. Biochemistry, 45, 11071–11086.
69 Yew, W.S., Fedorov, A.A., Fedorov, E.V., 77 Harayama, S., Rekik, M., Ngai, K.-L., and
Rakus, J.F., Pierce, R.W., Almo, S.C., and Ornston, L.N. (1989) Physically
Gerlt, J.A. (2006) Evolution of enzymatic associated enzymes produce and
activities in the enolase superfamily: metabolize 2-hydroxy-2,4-dienoate, a
L-fuconate dehydratase from chemically unstable intermediate formed
Xanthomonas campestris. Biochemistry, 45, in catechol metabolism via meta cleavage
14582–14597. in Pseudomonas putida. J. Bacteriol., 171,
70 Jeffcoat, R. (1975) Studies on the subunit 6251–6258.
structure of 4-deoxy-5-oxoglucarate 78 Pollard, J.R. and Bugg, T.D.H. (1998)
hydro-lyase (decarboxylating) from Purification, characterization and
Pseudomonas acidovorans. Biochem. J., reaction mechanism of monofunctional
145, 305–309. 2-hydroxypentadienoic acid hydratase
71 Dobson, R.C.J., Gerrard, J.A., and Pearce, from Escherichia coli. Eur. J. Biochem., 251,
F.G. (2004) Dihydrodipicolinate synthase 98–106.
is not inhibited by its substrate (S)- 79 Masai, E., Sugiyama, K., Iwashita, N.,
aspartate b-semialdehyde. Biochem. J., Shimizu, S., Hauschild, J.E., Hatta, T.,
377, 757–762. Kimbara, K., Yano, K. and Fukuda, M.
72 Brouns, S.J.J., Barends, T.R.M., Worm, (1997) The bphDEF meta-cleavage
P., Akerboom, J., Turnbull, A.P., Salmon, pathway genes involved in biphenyl/
L., and van der Oost, J. (2008) Structural polychlorinated biphenyl degradation are
insight into substrate binding and located on a linear plasmid and separated
catalysis of a novel 2-keto-3-deoxy-D- from the initial bphACB genes in
arabinonate dehydratase illustrates Rhodococcus sp. Strani RHA1. Gene, 187,
common mechanistic features of the 141–149.
FAH superfamily. J. Mol. Biol., 379, 80 Wang, P. and Seah, S.Y.K. (2005)
357–371. Determination of the metal ion
73 Lawrence, M.C., Barbosa, J.A.R.G., dependence and substrate specificity of a
Smith, B.J., Hall, N.E., Pilling, P.A., Ooi, hydratase involved in the degradation
H.C., and Marcuccio, S.M. (1997) pathway of biphenyl/chlorobiphenyl.
Structure and mechanism of a sub-family FEBS J., 272, 966–974.
of enzymes related to N- 81 Husain, A., Jeelani, G., Sato, D., Ali, V.,
acetylneuraminate lyase. J. Mol. Biol., and Nozaki, T. (2010) Characterization of
266, 381–399. two isotypes of L-threonine dehydratase
74 Roper, D.I., Stringfellow, J.M. and from Entamoeba histolytica. Mol. Biochem.
Cooper, R.A. (1995) Sequence of the hpcC Parasitol., 170, 100–104.
and hpcG genes of the meta-fission 82 Grabowski, R., Hofmeister, A.E.M., and
homoprotocatechuic acid pathway of Buckel, W. (1993) Bacterial L-serine
Escherichia coli C: nearly 40% amino-acid dehydratases: a new family of enzymes
identity with the analogous enzymes of containing iron-sulfur clusters. Trends
the catechol pathway. Gene, 156, 47–51. Biochem. Sci., 18, 297–300.
References j497
83 Christen, P. and Mehta, P.K. (2001) From 93 Almrud, J.J., Poelarends, G.J.,
cofactor to enzymes: The molecular Johnson, W.H. Jr., Serrano, H.,
evolution of pyridoxal-50 -phosphate- Hackert, M.L., and Whitman, C.P. (2005)
dependent enzymes. Chem. Rec., 1, Crystal structure of the wild-type, P1A
436–447. mutant, and inactivated malonate
84 Eliot, A.C. and Kirsch, J.F. (2004) semialdehyde decarboxylate: a structural
Pyridoxal phosphate enzymes: basis for the decarboxylase and hydratase
mechanistic, structural, and evolutionary activities. Biochemistry, 44, 14818–14827.
considerations. Annu. Rev. Biochem., 73, 94 Wang, S.C., Johnson, W.H. Jr., and
383–415. Whitman, C.P. (2003) The 4-
85 Tanaka, H., Yamamoto, A., Ishida, T., and oxalocrotonate tautomerase- and
Horiike, K. (2008) D-Serine dehydratase YwhB-catalyzed hydration of
from chicken kidney: A vertebral 3E-halocrylates: implications for the
homologue of the cryptic enzyme from evolution of new enzymatic activities.
Burkholderia cepacia. J. Biochem., 143, J. Am. Chem. Soc., 125, 14282–14283.
49–57. 95 Poelarends, G.J., Serrano, H.,
86 Hofmeister, A.E.M., Berger, S., and Johnson, W.H. Jr., Hoffman, D.W., and
Buckel, W. (1992) The iron-sulfur-cluster- Whitman, C.P. (2004) The hydratase
containing L-serine dehydratase from activity of malonate semialdehyde
Peptostreptococcus asaccharolyticus: decarboxylase: mechanistic and
Stereochemistry of the deamination of L- evolutionary implications. J. Am. Chem.
threonine. Eur. J. Biochem., 205, 743–749. Soc., 126, 15658–15659.
87 Marceau, M., McFall, E., Lewis, S.D., and 96 de Jong, R.M., Brugman, W.,
Shafer, J.A. (1988) D-Serine dehydratase Poelarends, G.J., Whitman, C.P., and
from Escherichia coli: DNA sequence and Dijkstra, B.W. (2004) The X-ray structure
identification of catalytically inactive of trans-3-chloroacrylic acid
glycine to aspartic acid variants. J. Biol. dehalogenase reveals a novel hydration
Chem., 263, 16926–16933. mechanism in the tautomerase
88 Ito, T., Hemmi, H., Kataoka, K., superfamily. J. Biol. Chem., 279,
Mukai, Y., and Yoshimura, T. (2008) A 11546–11552.
novel zinc-dependent D-serine 97 Poelarends, G.J., Veetil, V.P., and
dehydratase from Saccharomyces Whitman, C.P. (2008) The chemical
cerevisiae. Biochem. J., 409, 399–406. versatility of the b-a-b fold: catalytic
89 Yamada, E.W. and Jakoby, W. (1958) promiscuity and divergent evolution in
Enzymatic utilization of an acetylenic the tautomerase superfamily. Cell. Mol.
compound. J. Am. Chem. Soc., 80, Life. Sci., 65, 3606–3618.
2343–2344. 98 Poelarends, G.J. and Whitman, C.P.
90 Yamada, E.W. and Jakoby, W. (1959) (2004) Evolution of enzymatic activity in
Enzymatic utilization of the tautomerase superfamily:
acetylenic compounds. II. mechanistic and structural studies of the
Acetylenemonocarboxylic acid 1,3-dichloropropene catabolic enzymes.
hydratase. J. Biol. Chem., 234, 941–945. Bioorg. Chem., 32, 376–392.
91 Brecker, L., Petschnigg, J., Depine, N., 99 Woodard, R.W. (2004) Unique
Weber, H., and Ribbons, D.W. (2003) biosynthesis of dehydroquinic acid?
In situ proton NMR analysis of Biorg. Chem., 32, 309–315.
a-alkynoate biotransformations. Eur. J. 100 Kr€amer, M., Bongaerts, J., Bovenberg, R.,
Biochem., 270, 1393–1398. Kremer, S., M€ uller, U., Orf, S.,
92 Poelarends, G.J., Serrano, H., Wubbolts, M., and Raeven, L. (2003)
Johnson, W.H. Jr., and Whitman, C.P. Metabolic engineering for microbial
(2005) Inactivation of malonate production of shikimic acid. Metab. Eng.,
semialdehyde decarboxylase by 3- 5, 277–283.
halopropiolates: evidence for hydratase 101 Gibson, J.M., Thomas, P.S.,
activity. Biochemistry, 44, 9375–9381. Thomas, J.D., Barker, J.L.,
498
Chandran, S.S., Harrup, M.K., II dehydroquinase from Streptomyces

Draths, K.M., and Frost, J.W. (2001) coelicolor. Structure, 10, 493–503.
Benzene-free synthesis of phenol. Angew. 111 Moore, J.D., Hawkins, A.R., Charles, I.G.,
Chem. Int. Ed., 40, 1945–1948. Deka, R., Coggins, J.R., Cooper, A.,
102 Li, W., Xie, D., and Frost, J.W. (2005) Kelly, S.M., and Price, N.C. (1993)
Benzene-free synthesis of catechol: Characterization of the type I
interfacing microbial and chemical dehydroquinase from Salmonella typhi.
catalysis. J. Am. Chem. Soc., 127, Biochem. J., 295, 277–285.
2874–2882. 112 White, P.J., Young, J., Hunter, I.S.,
103 Li, K. and Frost, J.W. (1998) Synthesis of Nimmo, H.G., and Coggins, J.R. (1990)
vanillin from glucose. J. Am. Chem. Soc., The purification and characterization of
120, 10545–10546. 3-dehydroquinase from Streptomyces
104 Deka, R.K., Anton, I.A., Dunbar, B., and coelicolor. Biochem. J., 265, 735–738.
Coggins, J.R. (1994) The characterisation 113 Bottomley, J.R., Clayton, C.L.,
of the shikimate pathway enzyme Chalk, P.A., and Kleanthous, C. (1996)
dehydroquinase from Pisum sativum. Cloning, sequencing, expression,
FEBS Lett., 349, 397–402. purification and preliminary
105 Kleanthous, C., Deka, R., Davis, K., Kelly, characterization of a type II
S.M., Cooper, A., Harding, S.E., Price, dehydroquinase from Helicobacter pylori.
N.C., Hawkins, A.R., and Coggins, J.R. Biochem. J., 319, 559–565.
(1992) A comparison of the 114 Bottomley, J.R., Hawkins, A.R., and
enzymological and biophysical properties Kleanthous, C. (1996) Conformational
of two distinct classes of dehydroquinase changes and the role of metals in the
enzymes. Biochem. J., 282, 687–695. mechanism of type II dehydroquinase
106 Chaudhuri, S., Duncan, K., from Aspergillus nidulans. Biochem. J.,
Graham, L.D., and Coggins, J.R. (1991) 319, 269–278.
Identification of the active-site lysine 115 Kubo, Y., Takano, Y., Endo, N., Yasuda, N.,
residues of two biosynthetic 3- Tajima, S., and Furusawa, I. (1996)
dehydroquinases. Biochem. J., 275, 1–6. Cloning and structural analysis of the
107 Harris, J.M., Watkins, W.J., Hawkins, melanin biosynthesis gene SCD1
A.R., Coggins, J.R., and Abell, C. (1996) encoding sytalone dehydratase in
Comparison of the substrate specificity of Colletotrichum lagenarium. Appl. Environ.
type I and type II dehydroquinases with 5- Microbiol., 62, 4340–4344.
deoxy- and 4,5-dideoxy-dehydroquinic 116 Basarab, G.S., Steffens, J.J., Wawrzak, Z.,
acid. J. Chem. Soc. Perkin. Trans. 1, Schwartz, R.S., Lundqvist, T., and Jordan,
2371–2377. D.B. (1999) Catalytic mechanism of
108 Gourley, D.G., Shrive, A.K., Polikarpov, I., scytalone dehydratase: site-directed
Krell, T., Coggins, J.R., Hawkins, A.R., mutagenesis, kinetic isotope effects, and
Isaacs, N.W., and Sawyer, L. (1999) The alternate substrates. Biochemistry, 38,
two types of 3-dehydroquinase have 6012–6024.
distinct structures but catalyze the same 117 Jordan, D., Zheng, Y.-J., Lockett, B.A., and
overall reaction. Nat. Struct. Biol., 6, Basarab, G.S. (2000) Stereochemistry of
521–525. the enolization of scytalone by scytalone
109 Parker, E.J., Bello, C.G., Coggins, J.R., dehydratase. Biochemistry, 39, 2276–2282.
Hawkins, A.R., and Abell, C. (2000) 118 Okimoto, N., Nakamura, T., Suenaga, A.,
Mechanistic studies on type I and type II Futatsugi, N., Hirano, Y., Yamaguchi, I.,
dehydroquinase with (6R)- and (6S)-6- and Ebisuzaki, T. (2004) Cooperative
fluoro-3-dehydroquinic acids. Bioorg. motions of protein and hydration water
Med. Chem. Lett., 10, 231–234. molecules: molecular dynamics study of
110 Roszak, A.W., Robinson, D.A., Krell, T., scytalone dehydratase. J. Am. Chem. Soc.,
Hunter, I.S., Fredrickson, M., Abell, C., 126, 13132–13139.
Coggins, J.R., and Lapthorn, A.J. (2002) 119 Zheng, Y.-J. and Bruice, T. (1998)
The structure and mechanism of the type Role of a critical water in scytalone
References j499
dehydratase-catalyzed reaction. Proc. 128 Rawlings, B.J. (1997) Biosynthesis of fatty
Natl. Acad. Sci. U.S.A., 95, 4158–4163. acids and related metabolites. Nat. Prod.
120 Zheng, Y.-J., Basarab, G.S., and Rep., 14, 335–358.
Jordan, D.B. (2002) Roles of substrate 129 Sedgwick, B., Morris, C.A., and French, S.
distortion and intramolecular hydrogen (1978) Stereochemical course of
bonding in enzymatic catalysis by dehydration catalysed by the yeast fatty
scytalone dehydratase. Biochemistry, 41, acid synthetase. J. Chem. Soc., Chem.
820–826. Commun., 193–194.
121 Akashi, T., Aoki, T., and Ayabe, S.-I. 130 Smith, S., Witkowski, A., and Joshi, A.K.
(2005) Molecular and biochemical (2003) Structural and functional
characterization of 2- organization of the animal fatty acid
hydroxyisoflavanone dehydratase. synthase. Prog. Lipid Res., 42, 289–317.
Involvement of carboxylesterase-like 131 Gokhale, R.S., Sankaranarayanan, R., and
proteins in leguminous isoflavone Mohanty, D. (2007) Versatility of
biosynthesis. Plant Phys., 137, 882–891. polyketide synthases in generating
122 Yu, S., Andreassen, M., and Lundt, I. metabolic diversity. Curr. Opin. Struct.
(2008) Enzymatic production of Biol., 17, 736–743.
microthecin by aldos-2-ulose dehydratase 132 Smith, S. and Tsai, S.-C. (2007) The type I
from 1,5-anhydro-D-fructose and stability fatty acid and polyketide synthases: a tale
studies of microthecin. Biocatal. of two megasynthases. Nat. Prod. Rep., 24,
Biotransform., 26, 169–176. 1041–1072.
123 Lundt, I. and Yu, S. (2010) 1,5-Anhydro-D- 133 Das, A. and Khosla, C. (2009)
fructose: biocatalytic and chemical Biosynthesis of aromatic polyketides in
synthetic methods for the preparation, bacteria. Acc. Chem. Res., 42, 631–639.
transformation and derivatization. 134 Gallimore, A.R. (2009) The biosynthesis
Carbohydr. Res., 345, 181–190. of polyketide-derived polycyclic ethers.
124 Yu, S., Refdahl, C. and Lundt, I. (2004) Nat. Prod. Rep., 26, 266–280.
Enzymatic description of the 135 Koglin, A. and Walsh, C.T. (2009)
anhydrofructose pathway of glycogen Structural insights into nonribosomal
degradation. I. Identification and peptide enzymatic assembly lines. Nat.
purification of anhydrofructose Prod. Rep., 26, 987–1000.
dehydratase, ascopyrone tautomerase 136 Kopp, F. and Marahiel, M.A. (2007)
and a-1,4-glucan lyase in the fungus Where chemistry meets biology: the
Anthracobia melaloma. Biochim. Biophys. chemoenzymatic synthesis of
Acta, 1672, 120–129. nonribosomal peptides and polyketides.
125 Yu, S. and Fiskesund, R. (2006) The Curr. Opin. Biotechnol., 18, 513–520.
anhydrofructose pathway and its possible 137 Abe, I. (2008) Engineering of plant
role in stress response and signaling. polyketide biosynthesis. Chem. Pharm.
Biochim. Biophys. Acta, 1760, 1314–1322. Bull., 56, 1505–1514.
126 Andersen, S.M., Lundt, I., Marcussen, J., 138 Sattely, E.S., Fischbach, M.A., and Walsh,
and Yu, S. (2002) 1,5-Anhydro-D-fructose: C.T. (2008) Total biosynthesis: in vitro
a versatile chiral building block: reconstitution of polyketide and
biochemistry and chemistry. Carbohydr. nonribosomal peptide pathways. Nat.
Res., 337, 873–890. Prod. Rep., 25, 757–793.
127 Yu, S. (2005) Enzymatic description of the 139 Heath, R.J. and Rock, C.O. (1996) Roles of
anhydrofructose pathway of glycogen the FabA and FabZ b-hydroxyacyl-acyl
degradation. II. Gene identification and carrier protein dehydratases in
characterization of the reactions catalyzed Escherichia coli fatty acid biosynthesis.
by aldos-2-ulose dehydratase that J. Biol. Chem., 271, 27795–27801.
converts 1,5-anhydro-D-fructose to 140 Kimber, M.S., Martin, F., Lu, Y.,
microthecin with ascopyrone M as the Houston, S., Vedadi, M., Dharamsi, A.,
intermediate. Biochim. Biophys. Acta, Fiebig, K.M., Schmid, M., and Rock, C.O.
1723, 63–73. (2004) The structure of (3R)-hydroxyl-acyl
500
carrier protein dehydratase (FabZ) from recombinant Escherichia coli. J. Bacteriol.,

Pseudomonas aeruginosa. J. Biol. Chem., 185, 5391–5397.
279, 52593–52602. 149 Qin, Y.-M., Haapalainen, A.M., Conry, D.,
141 Kirkpatrick, A.S., Yokoyama, T., Choi, Cuebas, D., Hiltunen, J.K., and Novikov,
K.-J., and Yeo, H.-J. (2009) Campylobacter D.K. (1997) Recombinant 2-enoyl-CoA
jejuni fatty acid synthase II: structural hydratase derived from rat peroxisomal
and functional analysis of b-hydroxyacyl- multifunctional enzyme 2: role of the
ACP dehydratase (FabZ). Biochem. hydratase reaction in bile acid synthesis.
Biophys. Res. Commun., 380, Biochem. J., 328, 377–382.
407–412. 150 Hiltunen, J.K and Qin, Y.-M. (2000)
142 Dillon, S.C. and Bateman, A. (2004) The b-Oxidation – strategies for the
hotdog fold: wrapping up a superfamily of metabolism of a wide variety of acyl-CoA
thioesterases and dehydratases. BMC esters. Biochim. Biophys. Acta, 1484,
Bioinformatics, 5, 109–122. 117–128.
143 Leesong, M., Henderson, B.S., 151 Koski, M.K., Haapalainen, A.M.,
Gillig, J.R., Schwab, J.M., and Smith, J.L. Hiltunen, J.K., and Glumoff, T. (2005)
(1996) Structure of a dehydratase- Crystal structure of 2-enoyl-CoA
isomerase from the bacterial pathway for hydratase 2 from human peroxisomal
biosynthesis of unsaturated fatty acids: multifunctional enzyme type 2. J. Mol.
two catalytic activities in on active site. Biol., 345, 1157–1169.
Structure, 4, 253–264. 152 Koski, M.K., Haapalainen, A.M.,
144 Zhang, L., Liu, W., Hu, T., Du, L., Luo, C., Hiltunen, J.K., and Glumoff, T. (2004) A
Chen, K., Shen, X., and Jiang, H. (2008) two-domain structure of one subunit
Structural basis for catalytic and explains unique features of eukaryotic
inhibitory mechanisms of hydratase 2. J. Biol. Chem., 279,
b-hydroxyacyl-acyl carrier protein 24666–24672.
dehydratase (FabZ). 153 Bell, A.F., Feng, Y., Hofstein, H.A.,
J. Biol. Chem., 283, 5370–5379. Parikh, S., Wu, J., Rudolph, M.J.,
145 Kostrewa, D., Winkler, F.K., Folkers, G., Kisker, C., Whitty, A., and Tonge, P.J.
Scapozza, L., and Perozzo, R. (2005) (2002) Stereoselectivity of enoyl-CoA
The crystal structure of PfFabZ, the hydratase results from preferential
unique b-hydroxyacyl-ACP dehydratase activation of one of two bound
involved in fatty acid biosynthesis of substrate conformers. Chem. Biol., 9,
Plasmodium falciparum. Protein Sci., 14, 1247–1255.
1570–1580. 154 Bahnson, B.J., Anderson, V.E., and
146 Bhaumik, P., Koski, M.K., Glumoff, T., Petsko, G.A. (2002) Structural
Hiltunen, J.K., and Wierenga, R.K. (2005) mechanism of enoyl-CoA hydratase:
Structural biology of the thioester- three atoms from a single water are added
dependent degradation and synthesis of in either an E1cb stepwise or concerted
fatty acids. Curr. Opin. Struct. Biol., 15, fashion. Biochemistry, 41, 2621–2629.
621–628. 155 Holden, H.M., Benning, M.M., Haller, T.,
147 Dieuaide-Noubhani, M., Asselberghs, S., and Gerlt, J.A. (2001) The crotonase
Mannaerts, G.P., and van Veldhoven, P.P. superfamily: divergently related enzymes
(1997) Evidence that multifunctional that catalyze different reactions involving
protein 2, and not multifunctional protein acyl coenzyme A thioesters. Acc. Chem.
1, is involved in the peroxisomal Res., 34, 145–157.
b-oxidation of pristanic acid. Biochem. J., 156 Agnihotri, G. and Liu, H.-W. (2003)
325, 367–373. Enoyl-CoA hydratase: reaction,
148 Park, S.J. and Lee, S.Y. (2003) mechanism, and inhibition. Bioorg. Med.
Identification and characterization of a Chem., 11, 9–20.
new enoyl coenzyme A hydratase 157 M€uller-Newen, G., Janssen, U., and
involved in biosynthesis of medium- Stoffel, W. (1995) Enoyl-CoA hydratase
chain-length polyhydroxyalkanoates in and isomerase form a superfamily with a
References j501
common active-site glutamate residue. (CoA) hydratase at 2.5 angstroms
Eur. J. Biochem., 228, 68–73. resolution: a spiral fold defines the
158 Gerlt, J.A. and Babbitt, P.C. (1998) CoA-binding pocket. EMBO J., 15,
Mechanistically diverse enzyme 5135–5145.
superfamilies: the importance of 164 Hisano, T., Tsuge, T., Fukui, T., Iwata, T.,
chemistry in the evolution of catalysis. Miki, K., and Doi, Y. (2003) Crystal
Curr. Opin. Chem. Biol., 2, 607–612. structure of the (R)-specific enoyl-CoA
159 Babbitt, P.C. and Gerlt, J.A. (1997) hydratase from Aeromonas caviae involved
Understanding enzyme superfamilies. in polyhydroxyalkanoate biosynthesis.
Chemistry as the fundamental J. Biol. Chem., 278, 617–624.
determinant in the evolution of new 165 Masai, E., Harada, K., Peng, X.,
catalytic activities. J. Biol. Chem., 272, Kitayama, H., Katayama, Y., and
30591–30594. Fukada, M. (2002) Cloning and
160 Bhaumik, P., Koski, M.K., Glumoff, T., characterization of the ferulic acid
Hiltunen, J.K., and Wierenga, R.K. (2005) catabolic genes of Sphingomonas
Structural biology of the thioester- paucimobilis SYK-6. Appl. Environ.
dependent degradation and synthesis of Microbiol., 68, 4416–4424.
fatty acids. Curr. Opin. Struct. Biol., 15, 166 Calisti, C., Ficca, A.G., Barghini, P., and
621–628. Ruzzi, M. (2008) Regulation of ferulic
161 Bahnson, B., Anderson, V.E., and catabolic genes in Pseudomonas fluorescens
Petsko, G.A. (2002) Structural BF13: involvement of a MarR family
mechanism of enoyl-CoA hydratase: regulator. Appl. Microbiol. Biotechnol., 80,
three atoms from a single water are added 475–483.
in either an E1cb stepwise or concerted 167 Negishi, O., Sugiura, K., and Negishi, Y.
fashion. Biochemistry, 41, 2621–2629. (2009) Biosynthesis of vanillin via ferulic
162 Engel, C.K., Kiema, T.R., Hiltunen, J.K., acid in Vanilla planifolia. J. Agric. Food
and Wierenga, R.K. (1998) The crystal Chem., 57, 9956–9961.
structure of enoyl-CoA hydratase 168 Bennett, J.P., Bertin, L., Moulton, B.,
complexed with octanoyl-CoA reveals the Fairlamb, I.J.S., Brzozowski, A.M.,
structural adaptations required for Walton, N.J., and Grogan, G. (2008) A
binding of a long chain fatty acid-CoA ternary complex of hydroxycinnamoyl-
molecule. J. Mol. Biol., 275, 847–859. CoA hydratase-lyase (HCHL) with acetyl-
163 Engel, C.K., Mathieu, M., Zeelen, J.P., CoA and vanillin gives insights into
Hiltunen, J.K., and Wierenga, R.K. (1996) substrate specificity and mechanism.
Crystal structure of enoyl-coenzyme A Biochem. J., 414, 281–289.
j503
12
Industrial Application and Processes Forming CO Bonds
12.1
Processes Using Lipases
The hydrolysis and acylation of different C–O and C–N substrates are catalyzed by
hydrolases. Lipases belonging to EC 3.1 are the most used enzyme within such
reactions. Here one can differentiate the carboxylic ester hydrolases (EC 3.1.1), the
thiol ester hydrolases (EC 3.1.2), and the so-called phosphatases (phosphohydrolases;
EC 3.1.3). In the following, different process examples will be illustrated using
carboxylic ester hydrolases in the direction of hydrolysis or in the direction of
acylation (Scheme 12.1). Most of the processes use immobilized enzymes; however,
two examples will illustrate the application of whole cells or suspended enzymes, too.
Some lipases of EC 3.1.1 also catalyze the hydrolysis and formation of amides, which
is the natural reaction of the amidases (EC 3.5). Here one example will be given.
12.1.1
Processes Using Lipases in Hydrolytic Reactions
Compared to the classical chemical hydrolysis using acids, lipases have besides their
prominent high stability a certain enantioselectivity, which in organic solvents makes
them attractive catalysts in the resolution of racemates. Therefore, a racemic ester can
be hydrolyzed most often enantio- or regioselectively and the products are separated
afterwards to yield a chiral alcohol and an enantiopure ester in the optimal case
(Scheme 12.2).
Table 12.1 summarizes examples of industrial biotransformations that apply this
strategy. The different biotransformations use water for hydrolysis or alcohol for
alcoholysis to convert an ester enantioselectively. The desired product can be the
alcohol or the ester. Depending on the target product, which can be the converted or
non-converted enantiomer, different strategies of downstream processing need to be
applied.
Table 12.1 summarizes some hydrolytic processes used to resolve racemates and
two processes where the hydrolase converted an already chiral substrate. As exam-
and Oliver May.
2012 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2012 by Wiley-VCH Verlag GmbH & Co. KGaA.
j 12 Industrial Application and Processes Forming CO Bonds
504
Scheme 12.1 Reactions catalyzed by carboxylic ester hydrolases.
Scheme 12.2 Enantioselective hydrolysis of esters using lipases as a catalyst, yielding an alcohol
and the non-converted ester.
ples, the processes of Pfizer Inc., Tanabe Seiyaku Co., Ltd, Fuji Chemical Industries
Co., Ltd, Sumitomo Chemical Co., and Sanofi Aventis are explained. However, the
principles of reaction engineering can be transferred also to other reactions. Here the
focus is on the separation of the desired or the non-desired enantiomer from the
reaction medium. Different strategies will be illustrated in the following examples.
Ibuprofen is an important nonsteroidal anti-inflammatory drug. The in vivo activity
of the (S)-enantiomer is 100 times that of the (R)-enantiomer. Therefore, the
biocatalytic resolution catalyzed by immobilized lipase from Candida cylindracea is
highly attractive (Figure 12.1). Although the lipase shows good activity over a broad
pH range, a low pH value has to be employed because the enzyme is deactivated by
the ibuprofen ester. At low pH values the solubility of ibuprofen ester is very low. In
this case the solubility of the ester is below 1 mM and prevents deactivation. However,
this is also the main problem of the enzymatic synthesis. To circumvent problems of
handling large volumes of water, a membrane reactor concept is realized. Here, a
hollow fiber membrane is used, where the lipase is immobilized in the pores of the
membrane by entrapment. The hydrophobic ibuprofen methoxyethyl ester is deliv-
ered solubilized in the organic phase to the outside of the asymmetric membrane.
After conversion, the ibuprofen is extracted by the aqueous phase into the lumen of
the hollow fibers. The advantage of this reactor setup is the stabilization of the
aqueous/organic interphase by the membrane, which provides a high surface area
for contact between the organic and aqueous phases without dispersing one phase
into the other. In combination with another membrane module adjusted to a high
pH, the product can be easily separated from the non-converted ester, which can be
easily recycled to the first membrane system. These techniques allow a low ibuprofen
concentration at low pH, leading to high catalyst stability [16–22]. Present research
covers the fields of kinetic modeling, molecular modeling, and the separation of
substrates and products [23–25].
Table 12.1 Industrial hydrolysis processes making use of lipases enantioselectivity; the desired product is printed in italics.
Company Racemic ester (substrate) Yield (%) Converted component (alcohol) E.e. Non-converted component E.e.
(%) (ester) (%)
Bristol-Myers Squibb [1–4] cis-Azetidinone acetate 96 (3S,4R)-cis-Azetidinone (3R,4S)-cis-Azetidinone acetate 99.5

Celltech Group plc [5, 6] (R,S)-Azlactone of tert-leucine 90 (S)-Ester amide 97 (R)-Azlactone of tert-leucine
DSM [7, 8] (R,S)-Glycidate 85 (S)-Oxiranyl-methanol (R)-Glycidate 99.9
Celltech Group plc [9–12] ( þ / )-4-Hydroxy-2-oxabicyclo 22 ( þ )-4-endo-Hydroxy-2-oxabicy- 92 ( )-4-Hydroxy-2-oxabicyclo[3.3.0]
[3.3.0]oct-7-en-3-one butyrate ester clo[3.3.0]oct-7-en-3-one oct-7-en-3-one butyrate ester &
( þ / )-4-exo-hydroxy-2-oxabicyclo
[3.3.0]oct-7-en-3-one butyrate ester
Pfizer Inc.a) (R,S)-Ibuprofen methoxyethyl (S)-Ibuprofen 96 (R)-Ibuprofen methoxyethyl ester
ester
GlaxoSmithKline Pharma- ( þ / )-SB-215346 84 (S)-SB-214857 (Lotrafiban) 99 SB-215346
ceuticals plc [13]
Sumitomo Chemical Co.a) (R,S)-Acetic acid 2-methyl-4-oxo-3- (R)-4-Hydroxy-3-methyl-2-prop-2- (S)-Acetic acid 2-methyl-4-oxo-3-prop- 99.2
prop-2-ynyl-cyclopent-2-enyl ester ynyl-cyclopent-2-enone 2-ynyl-cyclopent-2-enyl ester
Tanabe Seiyaku Co., Ltda) 3-(4-Methoxyphenyl)glycidic acid 45 (2S,3R)-3-(4-Methoxyphenyl)gly- (2R,3S)-3-(4-methoxyphenyl) glycidic 99.9
methyl ester cidic acid acid methyl ester
Fuji Chemical Industries D,L-Pantolactone D-Pantoic acid 97 L-Pantolactone
Co., Ltda)
Bristol-Myers Squibbb) exo,exo-7-Oxabicyclo[2.2.1]hep- 90 [(1R,2R,3S,4S)-3-(Hydroxy- 98
[14, 15] tane-2,3-dimethanol diacetate ester methyl)-7-oxabicyclo[2.2.1]hep-
tan-2-yl]methyl acetate
Sanofi Aventisa,b) Glutaryl-7-aminocephalosporanic 7-Aminocephalosporanic acid
acid
a) These processes are explained in detail in this chapter.

b) These processes use no racemates as starting material but represent hydrolytic enzymatic reactions.
12.1 Processes Using Lipases
j505
506
Figure 12.1 Flow scheme for the lipase-catalyzed resolution yielding (S)-ibuprofen.
The (2R,3S)-3-(4-methoxyphenyl) glycidic acid methyl ester is an intermediate in

the synthesis of diltiazem. Diltiazem hydrochloride is a coronary vasodilator and a
calcium channel blocker that is produced worldwide in excess of 100 t a1. In
comparison to the chemical route only five, instead of nine, steps are necessary.
The kinetic resolution (Figure 12.2) is carried out in an early step during the
synthesis, resulting in reduction of waste. The lipase from Serratia marcescens Sr41
8000 is immobilized onto a spongy layer of a hollow fiber membrane by pressurized
adsorption. Thus the hydrophilic polyacrylonitrile hollow fiber is used as reactor unit.
The enzyme loading on the membrane is 1.6 105 U m2 and the apparent vmax for
the hydrolysis of 3-(4-methoxyphenyl)glycidic acid methyl ester is 1.7 U mg1
protein, while an enantiomeric excess of 99.9% and a yield of 45% can be obtained.
The lipase does not attack (2R,3S)-3-(4-methoxyphenyl)glycidic acid methyl ester.
However, the ester does act as a competitive inhibitor of the enzyme. The formed acid
is unstable and undergoes decarboxylation to yield 4-methoxyphenylacetaldehyde.
Figure 12.2 Flow scheme for the lipase-catalyzed resolution yielding (2R,3S)-3-(4-methoxyphenyl)
glycidic acid methyl ester.
12.1 Processes Using Lipases j507
This aldehyde strongly inhibits and deactivates the enzyme. Here, a removal by
continuous filtration as bisulfite adduct is possible. This bisulfite also acts as buffer to
maintain constant pH during synthesis. Along with the mentioned inhibitory effects,
the lipase is inhibited by Co2 þ , Ni2 þ , Fe2 þ , Fe3 þ , and EDTA, but can be activated by
Ca2 þ and Li þ [8, 26–32]. Present research is focused on strain improvement, lipase
secretion, and fermentation optimization [33–35].
Pantoic acid is used as part of the vitamin B2-complex and both D- and L-
pantolactone are used as chiral intermediates in chemical synthesis. The biotrans-
formation that yields these enantiopure compounds skips several steps that are
necessary in the chemical resolution process. Using lactonase from Brevibacterium
protophormiae the L-lactones can be obtained and using the lactonase from Fusarium
oxysporum furnishes the D-lactones. The Fusarium lactonase has a very broad
substrate spectrum covering different isomers of D,L-galactono-c-lactone, D,L-glu-
cono-d-lactone, and dihydrocoumarin and other aromatic substrates. Figure 12.3
illustrates the process yielding D-pantolactone. For the synthesis whole cells immo-
bilized in calcium alginate beads are used, retaining >90% of their initial activity even
after 180 days of continuous use. At the end of the reaction L-pantolactone is extracted
or re-racemized to D,L-pantolactone that is recycled into the reactor. The D-pantoic acid
is chemically lactonized to D-pantolactone and subsequently extracted and crystal-
lized [36, 37].
(S)-4-Hydroxy-3-methyl-2-prop-2-ynyl-cyclopent-2-enone is used as an intermedi-
ate in the synthesis of pyrethroids, which are used as insecticides that show excellent
insecticidal activity and a low toxicity in mammals. The lipase-catalyzed resolution of
the enone enantiomers is carried out with a conversion of 49.9% (Scheme 12.3).
Subsequent to the lipase-catalyzed hydrolysis, the released alcohol is sulfonated with
methanesulfonyl chloride in the presence of the non-converted acylated compound.
Hydrolysis of the sulfonated enantiomer in the presence of small amounts of calcium
carbonate takes place under inversion of the chiral center in contrast to the hydrolysis
Figure 12.3 Flow scheme for the lactonase-catalyzed resolution yielding D-pantoic acid and its
work up to yield D-pantolactone [38].
508
Scheme 12.3 Reaction cascade for the lipase-catalyzed resolution of an intermediate in the
synthesis of pyrethroids, yielding the (R)-alcohol, and the non-converted (S)-ester, and the reactive
work up to yield only the (S)-alcohol [38].
of the acylated enantiomer, which is carried out with retention of the chiral center. By
this means, an enantiomeric excess of 99.2% and a very high yield is achieved for the
(R)-alcohol. For this resolution an E-value of 1300 was determined [39].
12.1.2
Processes Using Lipases in Esterifications
In comparison to the classical chemical esterification using acids or heavy metal

catalysts, esterifications catalyzed by lipases can be carried out under very mild
conditions. This makes them attractive catalysts in the resolution of racemates as well
as in the synthesis of commodity esters (Table 12.2).
(1S,2S)-2-Methoxycyclohexanol is used as a building block for trinems, a new class
of totally synthetic b-lactam antibiotics containing a tricyclic skeleton. The immo-
bilized lipase B from Candida antarctica is used in a repetitive batch process, with
cyclohexane as a solvent, to obtain this building block (Figure 12.4). After nine cycles
the immobilized enzyme retained more than half of its activity. Free Pseudomonas
fluorescens lipase lost 75% of activity after four cycles. However, by immobilizing this
lipase on Celite powder the same stability as for the commercially available Novozym
435 could be obtained. The biotransformation is followed by a filtration step, an
extraction, and an evaporation step. This enzymatic resolution allows a product with
an enantiomeric excess of more than 98% [40].
(S)-N-(tert-Butoxycarbonyl)-3-hydroxymethylpiperidine is a key intermediate in
the synthesis of a potent tryptase inhibitor. Amano Lipase PS was found to be the best
Table 12.2 Industrial esterification processes using lipases.
Company Substrate Yield (%) Ester E.e. (%) Non-converted substrate E.e. (%)
BASF AG [41–43] (R,S)-1-Phenylethylamine 90 (R)-Phenylethyl-methoxyamide 93 (S)-1-Phenylethylamine 99

Bristol-Myers {4-[4a,6b(E)]}-6-[4,4-bis(4- 48 (S)-Acetic acid ester (R)-Alcohol 98.5
Squibb [44, 45] Fluorophenyl])-3-(1-methyl-
1H-terazol-5-yl)-1,3-butadie-
nyl]-tetrahydro-4-hydroxy-2-
pyran-2-one
GlaxoSmithKline ( þ / )-trans-2- 36 Acetic acid trans-2-methoxycyclo- ( þ )-(1S,2S)-2-Methoxycyclohexanol 98
plca) Methoxycyclohexanol hexanyl ester
Bristol-Myers (R,S)-N-(tert-Butoxycarbo- 32 (S)-Hemisuccinate 98.9 (R)-N-(tert-Butoxycarbonyl)-3-
Squibba) nyl)-3- hydroxymethylpiperidine
hydroxymethylpiperidine
Schering 2-[(2,4-Difluoro-phenyl)allyl] 74 Acetic acid 4-(2,4-difluorophenyl)- 99 —
Plow [46–51] propane-1,3-diol 2-hydroxymethyl-pent-4-enyl ester
& acetic acid 2-acetoxymethyl-4-
(2,4-difluoro-phenyl)-pent-4-enyl
ester
Unichema Che- Isopropanol 99 Isopropyl palmitate — —
mie BVa,b)
Evonik Indus- Polyglycerol-3 99 Polyglycerol-3 laurate — —
triesa,b)
a) These processes are explained in detail in this chapter.

b) These processes represent non-specific acylation.
12.1 Processes Using Lipases
j509
510
Figure 12.4 Flow scheme for the lipase-catalyzed resolution yielding enantiopure 2-
methoxycyclohexanol [38].
enzyme for the stereospecific hydrolysis of the respective (R,S)-esters. The enzymatic
resolution of the (R,S)-esters by lipase P from Pseudomonas fluorescens was reported
for the large-scale production as a process with low substrate concentrations and
chromatographic separation. Such a process will have limited practical use in large-
scale industrial application. The process for enzymatic resolution followed by easy
separation involved lipase PS catalyzed esterification of the (R,S)-alcohol with
succinic anhydride (Figure 12.5). Subsequently, the (S)-hemisuccinate is extracted
from the organic phase with base (5% NaHCO3). The non-converted (R)-alcohol
remains in the organic phase. Hydrolysis of the (S)-hemisuccinate with NaOH then
provides the desired (S)-alcohol. By using toluene as solvent, the (S)-alcohol was
isolated in 23% yield (maximum theoretical yield of 50%) with an enantiomeric
excess of >95%. The E-value was found to be 65–70 [52, 53].
Figure 12.5 Flow scheme for the lipase-catalyzed resolution yielding enantiopure the (S)-
hemisuccinate from racemic N-(tert-butoxycarbonyl)-3-hydroxymethylpiperidine and its subsequent
chemical hydrolysis to yield (S)-N-(tert-butoxycarbonyl)-3-hydroxymethylpiperidine [38].
12.1 Processes Using Lipases j511
The above-mentioned examples illustrate the use of lipases for the resolution of
racemic alcohols via enantioselective esterification. In addition, lipases are used in
the industrial synthesis of achiral esters, too. Isopropyl palmitate and isopropyl
myristate are used in the preparation of soaps, skin creams, lubricants, and greases.
The synthesis is carried out without any solvent, directly in the mixture of substrates,
using the immobilized lipase from Candida antarctica. The problem during ester
synthesis is the side-product water, which leads to equilibrium conditions, which
means that forward and backward reaction have the same rates. Two possible process
layouts have been published for these reaction systems (Figure 12.6). In the first
Figure 12.6 Two possible process layouts for solvent-free esterification using immobilized lipase.
Water is removed by azeotropic distillation (a) or by pervaporation (b) [38].
512
process, the reaction water is removed by azeotropic distillation (alcohol–water) at

0.26 bar. Isopropanol is continuously fed into the reactor (58 g h1) to replace that
which is distilled. The immobilized enzyme can be easily removed by filtration
(Figure 12.6a). Alternatively, the reaction water is removed during esterification by
pervaporation at 80 C (Figure 12.6b). The reaction solution with the lower water
content is cooled to 65 C and passed to the second reactor unit. After a second
pervaporation step the water content is lowered to 0.2 wt%. The process can be
adapted to other alcohols and acid, limited only by the boiling point of the alcohol or
by the pressure drop over the fixed bed reactor [54–56].
The enzymatic synthesis of fatty acid esters can be carried out as a solvent-free reaction
in a plug-flow reactor. However, the produced water has to be removed from the reaction
medium to obtain total conversion by shifting the thermodynamic equilibrium. This is
also possible by the sequence of a plug-flow reactor containing the biocatalyst and an
evacuated stirred tank reactor for water removal. In this case the reaction progress is
limited by the liquid–gas interface. Additionally, a plug-flow reactor is restricted to raw
materials and products having a viscosity that is low enough to allow its pumping through
the bed of immobilized enzyme, resulting in a low differential pressure drop. The
production of surfactants from high-viscous or high-melting reactants, such as, for
example, diglycerol, polyglycerol, or other polyols, illustrates therefore further challenges
besides overcoming limitations caused by thermodynamic equilibrium.
The enzymatic synthesis of polyglycerol-3-esters starting from polyglycerol-3 and
lauric acid using an immobilized lipase B from Candida antarctica (Novozym 435)
leads to an emulsifier for personal care products. Both substrate and the product
show a very high viscosity and, therefore, removal of the formed water by vacuum
only is not sufficient. Furthermore, the pressure drop in a fixed bed reactor is too
high. These limitations can be overcome by using a bubble column reactor. The side-
product removal can be enhanced in this reactor type but is strongly dependent on the
aeration rate and size of the formed gas bubbles (Figure 12.7). Additionally, aeration
enables mixing of the two liquid reactants as well as of the immobilized biocatalyst. A
detailed description of the benefits of the bubble column reactor regarding reaching
minimal reaction time and the preferable use also for low viscous products like
myristyl myristate has been published recently [57, 58]. Comparing the effective
reaction times ERn (derived from the experiment) with the theoretical reaction times
TRn and calculating the efficiency quotients EQn revealed the high efficiency of this
reactor setup in the case of enzymatic esterifications for a given conversion n:

cS; 0 cS Km cS; 0
TRn ¼ þ ln
vmax vmax cS
ERn
EQn ¼
TRn
The efficiency quotient EQ99.6 of 8.3 when using a plug-flow reactor is significantly
reduced to 3.3 when applying the bubble column reactor. This means that at a
constant enzyme concentration the space–time yield can be approximately doubled.
At a conversion at 99.6%, a space–time yield of 6.7 kg l1 day1 and a biocatalyst
consumption of less than 4 mg g1 product were obtained.
12.2 Processes Using Glycosyltransferases, Glycosidases, and Carbon–Oxygen Lyases j513
Figure 12.7 Solvent-free esterification in a bubble column using immobilized lipase. Water is
removed by stripping.
The mechanical and leaching stability of enzymes adsorbed on macroporous

carriers is an important issue for the technical applicability of such biocatalysts in the
production of surfactants. Both can considerably benefit from the deposition of
silicone coating on the carrier surface. Deposition of silicone on the poly(methyl
methacrylate) (PMMA) carrier was found to form an interpenetrating network
composite rather than the anticipated core–shell structure. The silicone precursors
homogeneously wet the carrier surface, including all inner pores, and gradually fill
the complete carrier. A visible layer of silicone on the outer surface of the carrier was
only observed at a silicone concentration of 54 wt% and more. Maximum leaching
stability corresponds to the formation of this layer. The mechanical stability increases
with the amount of deposited silicone [59].
12.2
Processes Using Glycosyltransferases, Glycosidases, and Carbon–Oxygen Lyases
Glycosyltransferases, glycosidases, and carbon–oxygen lyases are enzymes that belong

to the enzyme classes EC 2.4, EC 3.2.1, and EC 4.2, respectively. Glycosyltransferases
have been widely used in the synthesis of glycoconjugates. Suitable enzymes can be
isolated from natural sources or produced recombinantly. As an alternative, whole cell-
based systems utilizing either endogenous glycosyl donors or cell-based systems
containing cloned and expressed systems for synthesis of glycosyl donors have been
developed. Glycosidases have various uses, including degradation of plant materials in
food and paper and pulp industries. In organic chemistry, glycosidases can be used as
catalysts to synthesize glycosidic bonds through either reverse hydrolysis where the
equilibrium position is reversed or by transglycosylation whereby glycoside hydrolases
514
can catalyze the transfer of a glycosyl moiety from an activated glycoside to an acceptor
alcohol to afford a new glycoside [60, 61]. Carbon–oxygen lyases are an attractive group
of catalysts as demonstrated by their use in many industrial processes. The reaction
catalyzed is the cleavage of the CO bond. Importantly, this bond cleavage is different
from hydrolysis, often leaving unsaturated products with double bonds that may be
subjected to further reactions. In industrial processes these enzymes are most
commonly used in the synthetic mode, meaning that the reverse reaction – addition
of a molecule to an unsaturated substrate – is of interest. To shift the equilibrium these
reactions are often carried out at very high substrate concentrations, or in situ product
removal methods are applied. Along with the examples of processes forming CO
bonds discussed in this chapter, several processes use oxidative enzymes to epoxidize
alkenes. These are discussed in Chapter 31 of this book. Furthermore, an interesting
process using a halohydrin dehalogenase (such enzymes are described in detail in
Chapter 9) can be found in Chapter 45. Here the enzyme is used in combination with a
glucose dehydrogenase to yield ethyl-(R)-4-cyano-3-hydroxybutyrate.
12.2.1
Processes Applying Glycosyltransferases
Cyclodextrins serve as molecular hosts and are used in the food industry for capturing
and retaining flavors. They are used in the formulation of pharmaceuticals and
produced by Mercian Co., Ltd from liquefied starch by cyclodextrin glycosyltransfer-
ase (EC 2.4.1.19) as a mixture of a-, b-, and c-cyclic oligosaccharides (Figure 12.8).
Figure 12.8 Process for the production of a-cyclodextrin from starch by cyclodextrin
glycosyltransferase [38].
The main problem that had to be overcome to establish an economic cyclodextrin
production was to separate the cyclodextrins from the aqueous reaction media.
This is important because the reaction mixture contains many by-products and
increasing cyclodextrin concentrations will inhibit the enzyme. The separation is
established by selective adsorption of a- and b-cyclodextrins on chitosan beads with
appropriate ligands [62]. a-Cyclodextrins selectively interact with stearic acid and
b-cyclodextrins with cyclohexanepropanamide-n-caproic acid [6-(3-cyclohexylpropa-
namido)hexanoic acid]. The adsorption selectivity is almost 100%. For b-cyclodex-
trins a capacity of 240 g l1 chitosan gel bed is reached. The reaction is carried out at
55 C to keep the viscosity low and to obtain a higher solubility of the reactants. Before
entering the adsorption column the temperature is lowered from 55 to 30 C for
effective adsorption. At this temperature almost no cyclodextrins are formed during
circulation. Before re-entering the main reactor the temperature of the solution is
again adjusted to 55 C by using the energy of the reaction solution leaving the
reactor. To prevent adsorption of the cyclodextrin glycosyltransferase in the a-cyclo-
dextrin adsorbent NaCl is added [62–64].
Novartis has developed a process based on uridine diphosphate glucuronic
acid (UDPGA) transferase (EC 2.4.1.17) for the production of acyl-glucuronide
of mycophenolic (MPA) acid (Scheme 12.4). This substance is biologically
active and requested for pharmaceutical studies in transplantation research.
The preparative scale synthesis of the acyl-glucuronide was therefore a great
challenge.
Scheme 12.4 Enzymatic glucuronidation of mycophenolic acid.

516
The acyl-glucuronide of mycophenolic acid was enzymatically synthesized on a

preparative scale under optimized reaction conditions with 51% conversion. By
screening nine liver homogenates from eight vertebrate species it was shown that
only with the liver homogenate from horse as catalyst were the acyl- and the O-
glucuronide formed in an approximately 1 : 1 ratio. With homogenates from other
sources, the O-glucuronide was produced in high yields. By optimizing the con-
centration of the cosubstrate UDP-glucuronic acid and the reaction temperature, the
conversion to the acyl-glucuronide was increased from initially 34% to 55% and the
ratio of acyl- to O-glucuronide from 1.5 : 1 to 3.9 : 1. In human metabolism, the ratio of
the acyl-glucuronide to the 7-O-glucuronide is reported to be 1 : 80 in favor of the 7-O-
glucuronide. The reaction was also performed continuously in an enzyme mem-
brane reactor, with, however, lower conversion yield and, therefore, higher specific
UDP-glucuronic acid consumption [65, 66].
The following process is part of the production of high fructose corn syrup (HFCS)
and is carried out by several companies (Scheme 12.5). After a number of improve-
ments this process (Scheme 12.5) provides an effective way for the most important,
low-cost sugar substitute derived from grain. At various stages enzymes are applied
in this process [56, 57]. The corn kernels are softened to separate oil, fiber,
and proteins by centrifugation. The enzymatic steps are cascaded to yield the
source product for the invertase process after liquefaction in continuous cookers,
Scheme 12.5 Hydrolysis of starch to glucose by a-amylase (EC 3.2.1.1) and glucoamylase
(EC 3.2.1.3) [38].
Figure 12.9 Flow scheme for the hydrolysis of starch to glucose applying a-amylase (E1) and
glucoamylase (E2) [38].
debranching, and filtration (Figure 12.9). Since starches from different natural
sources have different compositions the procedure varies. The process ends if all
starch is completely broken down to limit the amount of oligomers of glucose and
dextrins. Additionally, reglycosylation of hydrosylate molecules has to be prevented.
The thermostable a-amylase can be used up to 115 C. The enzymes need Ca2 þ ions
for stabilization and activation. Since several substances in corn can complex cations,
the cation concentration has to be increased, which requires further product
purification by refining the product. There is no alternative industrial chemical
process for starch liquefaction. The worldwide production of HFCS is about
10 000 000 t a1 [67–69].
12.2.2
Syntheses Using Carbon–Oxygen Lyases
About 40 000 t a1 of L-malic acid are used worldwide as a supplement in food,
cosmetics, and pharmaceutical industries. The synthesis of L-malic acid from
fumaric acid carried out by Amino GmbH is catalyzed by suspended whole cells
of Corynebacterium glutamicum containing the fumarase (Scheme 12.6). This bio-
transformation produces only L-malate; D-malate is not detectable. Microbial
fumarases lead to a mixture of approx. 85% malate and 15% fumarate, but according
to German drug regulations the fumaric acid content of malic acid has to be less than
Scheme 12.6 Production of malic acid from fumaric acid by fumarase (EC 4.2.1.2) [38].
518
Scheme 12.7 Production of malate from fumarate by fumarase [38].
0.15%. Fumaric acid separation is circumvented by forcing a quantitative transfor-

mation in a slurry reaction. The reaction is carried out in a slurry of crystalline
calcium fumarate and crystalline calcium malate (Scheme 12.7). The precipitation of
the product shifts the equilibrium towards calcium malate. The biotransformation is
carried out under non-sterile conditions, necessitating the addition of preservatives
to prevent microbial growth. As preservatives p-hydroxybenzoic acid esters are used
because of their low effective concentrations and their biodegradability. They are
permitted as food preservatives. Enzyme stabilization is achieved by the addition of
soy bean protein or bovine serum albumin. The surplus of foreign protein adsorbs to
interfaces at walls, stirrers, and liquid surfaces and protects the enzyme from
interface denaturation. The biocatalyst is easily separated by filtration of the slurry.
Downstream processing is carried out by acidification with sulfuric acid, yielding L-
malic acid and gypsum. The latter is separated by filtration and L-malic acid is
subsequently purified by ion-exchange chromatography [70–74].
A very similar process is carried out by Tanabe Seiyaku Co., Ltd with an
immobilized fumarase. Here cells of Brevibacterium flavum containing the fumarase
are immobilized on k-carrageenan gel. The formation of succinic acid is a side
reaction and can be eliminated by treatment of immobilized cells with bile extracts.
Additionally, the activity and stability can be improved by immobilization in the
presence of Chinese gallotannin [75–80]. Alternatively, the addition of NH3 instead of
H2O as nucleophile is catalyzed by aspartase, as reported in Chapter 20.
The biotransformation carried out by Kanegafuchi Chemical Industries Co., Ltd to
synthesize b-hydroxy-n-butyric acid from butyric acid occurs in three steps
(Scheme 12.8). Initially, the aliphatic acid is dehydrogenated to the a,b-unsaturated
acid. In a subsequent step, enantioselective hydration takes place. The catalysts used
are whole cells from Candida rugosa containing an enoyl-CoA hydratase and are
applied in aqueous solution at pH 7.2–7.5. (R)-b-Hydroxybutyric acid is yielded with
an enantiomeric excess of >98% and is used in the synthesis of a carbapenem
intermediate [79, 81].
The same catalyst can be also used for the conversion of isobutyric acid into
b-hydroxy-isobutyric acid (3-HBA). This biotransformation occurs in the same three
steps, as mentioned above. The product (R)-b-hydroxy-isobutyric acid is used as a
chiral synthon in the synthesis of captopril, an ACE-inhibitor [79–82].
Scheme 12.8 Production of b-hydroxy-n-butyric acid from butyric acid and of b-hydroxy-isobutyric
acid from isobutyric acid by enoyl-CoA hydratase (EC 4.2.1.17) [38].
The synthesis of L-tryptophan from L-serine and indole is possible by applying

suspended whole cells containing L-tryptophan synthase (Figure 12.10). The estab-
lished process of the Amino GmbH is dedicated to the production of L-tryptophan as a
pharmaceutically active ingredient. The applied enzyme uses pyridoxal phosphate as
a cofactor and works enantiospecifically for a-L-amino acid substrates. The starting
material L-serine is separated from molasses – the best separation is performed with
ion-exchange chromatography close to the isoelectric point of serine (pI 5.68). By
concentration of the serine fraction to 35% dry mass the main fraction of D-serine can
be separated by filtration, leaving a L-serine stock solution. The process is carried out
Figure 12.10 Process for the production of L-tryptophan from L-serine by tryptophan synthase
(E ¼ EC 4.2.1.20) [38].
520
Figure 12.11 Process for the production of malic acid from maleic anhydride by malease
(E ¼ EC 4.2.1.31) [38].
as a fed-batch whereas the indole dosage is directed via online HPLC analysis of the
product/starting material ratio. The L-tryptophan is produced in such high concen-
trations that it crystallizes instantaneously and is isolated together with the cells at the
end of the fed-batch. The crude tryptophan is solubilized in hot water and the cells are
separated after addition of charcoal. The L-tryptophan yield is >95% based on indole
and used in parenteral nutrition and as a pharmaceutical active ingredient in
sedatives, neuroleptics, antidepressants, and food additives as well as an interme-
diate for the production of other pharmaceutical compounds [83–86].
Instead of maleic acid the cheaper maleic anhydride, which hydrolyses in situ to
maleate, is used by DSM in the following process (Figure 12.11). Here immobilized
whole cells of Pseudomonas pseudoalcaligenes containing a malease are used in a
reaction sequence that starts from maleic anhydride hydrolysis to yield maleic acid,
which is then enzymatically converted into malic acid. The enzyme does not need
cofactors and although D-malate is a competitive inhibitor it stabilizes the enzyme.
The chosen strain is not able to grow on maleate as the sole carbon and energy source,
because it is probably not capable of synthesizing a transport mechanism for maleate.
Therefore, to overcome transport problems of substrate and product across the cell
membrane, the cells are permeabilized with Triton X-100. During growth the
enzymatic activity is constant in the logarithmic phase; the cells must be harvested
before the substrate for growth is completely consumed, because otherwise malease
activity drops rapidly. The yield of the reaction is >99% and the enantiomeric excess
is >99.99%. The D-malate can be used as chiral synthon or as resolving agent in the
resolution of racemic compounds [87–90].
L-Carnitine is used in infant, health sport, and geriatric nutrition. The biotrans-
formation of Lonza AG is catalyzed by carnitine dehydratase in whole cells
(Scheme 12.9). (R)-Carnitine is produced with >99.5% conversion of 4-butyrobe-
taine and >99.5% e.e. On the mutant strain the L-carnitine dehydrogenase is blocked
and excretes the accumulated product. The purified enzyme could not be used for the
biotransformation due to its high instability. Apart from the usual batch fermenta-
tions, continuous production is also feasible since the cells go into a maintenance
state with high metabolic activity and low growth rate. The cells can be recycled after
Scheme 12.9 Synthesis of carnitine: comparison of chemical and biocatalytic – catalyzed by

carnitine dehydratase in whole cells (E ¼ EC 4.2.1.89) – routes.
separation from the fermentation broth by filtration. The chemical resolution process
with L-tartaric acid developed at Lonza was no longer competitive with the biotech-
nological route. A more attractive route would be the Ru-BINAP catalyzed asym-
metric hydrogenation of 4-chloroacetoacetate (Scheme 12.9). Here an e.e. of 97% is
obtained [91–95].
5-Cyanovaleramide is used as intermediate for the synthesis of the DuPont
herbicide azafenidine (Scheme 12.10). Whole cells from Pseudomonas chlororaphis
are immobilized in calcium alginate beads. The biotransformation carried out by
DuPont is catalyzed by a nitrile hydratase that converts a nitrile into the corre-
sponding amide by addition of water [96]. Nitrile hydratases belonging to the
enzyme class of lyases (EC 4) are not to be confused with the nitrilases belonging to
the class of hydrolases (EC 3) that hydrolyze nitriles to the corresponding carbon
acids. For strain selection it was important that the cells did not show any amidase
activity that would further hydrolyze the amide to the carboxylic acid. The bio-
transformation is carried out in a two-phase system with pure adiponitrile forming
the organic phase. A reaction temperature of 5 C was chosen, since the solubility of
the by-product adipodiamide is only 37–42 mM in 1–1.5 M 5-cyanovaleramide. A
batch reactor is preferred over a fixed-bed reactor, because of the lower selectivity to
5-cyanovaleramide that was observed and the possibility of precipitation of adipo-
diamide, which would plug the column. Excess water is removed at the end of the
522
Scheme 12.10 Synthesis of 5-cyanovaleramide: comparison of chemical and biocatalytic –

catalyzed by nitrile hydratase from Pseudomonas chlororaphis B23 (EC 4.2.1.84) – routes [38].
reaction by distillation. The process has a selectivity of 96% and a yield of 93%. The
by-product adipodiamide is precipitated by dissolution of the resulting oil in
methanol at >65 C. The raw product solution is transferred directly to the
herbicide synthesis.
By this method 13.6 metric tons have been produced in 58 repetitive batch cycles
with 97% conversion and 96% selectivity. This biotransformation was chosen over
the chemical transformation due to a higher conversion and selectivity, production of
more product per catalyst weight [3150 kg kg1 (dry cell weight)], and less waste. The
catalyst consumption is 0.006 kg per kg product [97, 98].
Acrylamide (Scheme 12.11) is an important commodity monomer used for fibers,
coagulators, soil conditioners and stock additives for paper treatment and paper
sizing, and for adhesives, paints, and petroleum recovering agents. Since acryloni-
trile is the most poisonous nitrile, screening for microorganisms was conducted with
low-molecular weight nitriles instead.
Scheme 12.11 Synthesis of acrylamide catalyzed by nitrile hydratase from Rhodococcus rhodochrous
(EC 4.2.1.84) [38].
Acrylamide is unstable and polymerizes easily, therefore the process is carried out
at a low temperature (5 C). Although the cells, immobilized on polyacrylamide gel,
and the inside enzyme are very stable towards acrylonitrile, the starting material has
to be fed continuously to the reaction mixture due to inhibition effects at higher
concentrations. The biotransformation is started with an acrylonitrile concentration
of 0.11 M and is stopped at an acrylamide concentration of 5.6 M. The process is
operated at a capacity of 30 000 t a1.
This nitrile hydratase acts also on other nitriles, yielding 100% of the correspond-
ing amides. The most impressive example is the conversion of 3-cyanopyridine
into nicotinamide. The product concentration is about 1465 g l1. This conversion
(1.17 g l1 dry cell mass) can be termed pseudo-crystal since at the start of the
reaction the starting material is solid and with ongoing reaction it is solubilized.
The chemical synthesis uses copper salt as catalyst for the hydration of acrylonitrile
and has several disadvantages:
1) The rate of acrylamide formation is lower than that of acrylic acid formation;
2) the double bond of the starting material and the product causes the formation of
by-product such as ethylene, cyanohydrin, and nitrilotrispropionamide;
3) polymerization occurs;
4) copper needs to be separated from the product (an additional step within the
chemical synthesis).
The biotransformation of Nitto Chemical Industry has the advantages that no
recovery of unreacted nitrile is necessary since the conversion is 100% and no
removal of copper is needed. This is also the first case of a biocatalytic conversion of a
bulk fiber monomer [99–103].
Nicotinamide (vitamin B3) is used as vitamin supplement for food and animal
feed. It is the same strain that is also used in the industrial production of acrylamide.
The biotransformation is carried out at a scale of 3000 t a1.
In contrast to the chemical alkaline hydrolysis of 3-cyanopyridine with 4% by-
product of nicotinic acid (96% yield) the biotransformation works with absolute
selectivity and no acid or base are required. The continuously carried out biotrans-
formation is operated at low temperature and atmospheric pressure. In contrast to
the old synthesis route of nicotinamide at Lonza the new one is environmentally
friendly and safe (Scheme 12.12). Only one organic solvent is used throughout the
whole process in four highly selective continuous and catalytic reactions. The process
water, NH3, and H2 are recycled [104, 105].
12.2.3
Outlook
The discussed examples demonstrate successful complementation of molecular

engineering and reaction engineering. The beneficial combination of technologies
will become even more important for future processes in the area of biorefineries that
utilize high concentrations of oligo- and polysaccharides as starting materials.
524
Scheme 12.12 Synthesis of nicotinamide: comparison of chemical and biocatalytic – catalyzed by

nitrile hydratase (EC 4.2.1.84) – routes.
References j525
Alternatively, this will also become of high importance if polymers are not only
formed as side but as target products. In view of these applications enzymes like
glycosyltransferases, glycosidases, and in general carbon–oxygen lyases are asked for
that show a high regioselectivity.
References
1 Holton, R.A. (1992) Method for 9 Evans, C.T., Roberts, S.M., Shoberu,
preparation of taxol using b-lactam, K.A., and Sutherland, A.G. (1992)
Florida State University, US 5,175,315. Potential use of carbocyclic nucleosides
2 Patel, R.N., Szarka, L.J., and Partyka, R.A. for the treatment of AIDS: chemo-
(1993) Enzymatic process for resolution enzymatic syntheses of the enantiomers
of enantiomeric mixtures of compounds of carbovir. J. Chem. Soc., Perkin Trans. 1,
useful as intermediates in the 589–592.
preparation of taxanes, E.R. Squibb Sons 10 MacKeith, R.A., McCague, R.,
& Inc., EP 552041A2. Olivo, H.F., Palmer, C.F., and
3 Patel, R.N., Banerjee, A., Ko, R.Y., Roberts, S.M. (1993) Conversion of ( – )-4-
Howell, J.M., Li, W.-S., Comezoglu, hydroxy-2-oxabicyclo[3.3.0]oct-7-en-3-
F.T., Partyka, R.A., and Szarka, L.J. one into the anti-HIV agent carbovir.
(1994) Enzymic preparation of J. Chem. Soc., Perkin Trans. 1, 313–314.
(3R-cis)-3-(acetyloxy)-4-phenyl-2- 11 MacKeith, R.A., McCague, R.,
azetidinone: a taxol side-chain Olivo, H.F., Roberts, S.M., Taylor, S.J.C.,
synthon. Biotechnol. Appl. Biochem., and Xiong, H. (1994) Enzyme-catalysed
20, 23–33. kinetic resolution of 4-endo-hydroxy-2-
4 Zaks, A. and Dodds, D.R. (1997) oxabicyclo[3.3.0]oct-7-en-3-one and
Application of biocatalysis and employment of the pure enantiomers
biotransformations to the synthesis of for the synthesis of anti-viral and
pharmaceuticals. Drug Discovery Today, 2, hypocholesteremic agents. Bioorg. Med.
513–531. Chem., 2, 387–394.
5 McCague, R. and Taylor, S.J.C. (1997) 12 Taylor, S.J.C. and McCague, R. (1997)
Dynamic resolution of an oxazolinone by Resolution of a versatile hydroxylactone
lipase biocatalysis: Synthesis of (S)-tert- synthon 4-endo-hydroxy-2-oxabicyclo
leucine, in Chirality In Industry II [3.30]oct-7-en-3-one by lipase
(eds A.N. Collins, G.N. Sheldrake, and J. deesterification, in Chirality In Industry II
Crosby), John Wiley & Sons, Inc., (eds A.N. Collins, G.N. Sheldrake, and
New York, pp. 201–203. J. Crosby), John Wiley & Sons, Inc.,
6 Turner, N.J., Winterman, J.R., New York, pp. 190–193.
McCague, R., Parratt, J.S., and 13 Walsgrove, T.C., Powell, L., and Wells, A.
Taylor, S.J.C. (1995) Synthesis of (2002) A practical and robust process to
homochiral L-(S)-tert-leucine via a lipase produce SB-2124857, Lotrafiban, [(2S)-7-
catalyzed dynamic resolution process. (4,40 -bipiperidinylcarbonyl)-2,3,4,5,-
Tetrahedron Lett., 1113–1116. tetrahydro-4-methyl-4-oxo-1H-1,4-
7 Sheldon, R.A. (1993) Chirotechnology, benzodiazepine-2-acetic acid]
Marcel Dekker Inc., New York. utilising an enzymatic resolution as the
8 Elferink, V.H.M. (1995) Progress in the final step. Org. Process Res. Dev., 6,
application of biocatalysis in the 488–491.
industrial scale manufacture of chiral 14 Patel, R.N. (2001) Enzymatic synthesis
molecules. Chiral USA 96, 11th of chiral intermediates for drug
International Spring Innovations development. Adv. Synth. Catal.,
Chirality Symposium, Boston, pp. 79–80. 343 (6–7), 527–546.
526
15 Patel, R.N. (2001) Enzymatic preparation resolution of 1-phenylethanol integrated

of chiral pharmaceutical intermediates with separation of substrates and
by lipases. J. Liposome Res., 11 (4), products by a supported ionic liquid
355–393. membrane. J. Chem. Tech. Biotechnol.,
16 Adams, S.S., Bresloff, P., and 84 (3), 337–342.
Mason, C.G. (1976) Pharmacologica. 26 Kierkels, J.G.T. and Peeters, W.P.H.
Differences between the optical (1994) Process for the enzymatic
isomers of ibuprofen: evidence preparation of optically active
for metabolic inversion of the ()- transglycidic acid esters, DSM NV, EP
isomer. J. Pharm. Pharmacol., 28, 602740 A1.
256–257. 27 Lopez, J.L., Matson, S.L., Stanley, T.J., and
17 Cesti, P. and Piccardi, P. (1986) Quinn, J.A. (1991) Liquid-liquid
Process for the biotechnological extractive membrane reactors. Bioprocess
preparation of optically active Technol., 11, 27–66.
a-arylalkanoic acids, Montedison S.p.A., 28 Matson, S.L. (1987) Method and
Italy, EP 195717 A2. apparatus for catalyst containment in
18 Lopez, J.L., Wald, S.A., Matson, S.L., and multiphase membrane reactor systems,
Quinn, J.A. (1990) Multiphase PCT WO 87/02381, PCT US 86/02089.
membrane reactors for separating 29 Matsumane, H., Furui, M., Shibatani, T.,
stereoisomers. Ann. N. Y. Acad. Sci., 613, and Tosa, T. (1994) Production of optically
155–166. active 3-phenylglycidic acid ester by the
19 McConville, F.X., Lopez, J.L., and lipase from Serratia marcescens on a
Wald, S.A. (1990) Enzymatic resolution hollow-fiber membrane reactor.
of ibuprofen in a multiphase membrane J. Ferment. Bioeng., 78, 59–63.
reactor, in Biocatalysis (ed. D.A. 30 Matsumane, H. and Shibatani, T. (1994)
Abramowicz), van Nostrand Reinhold, Purification and characterization of the
New York, pp. 167–177. lipase from Serratia marcescens Sr41 8000
20 Sheldon, R.A. (1993) Chirotechnology, responsible for asymmetric hydrolysis
Marcel Dekker Inc., New York. of 3-phenylglycidic acid esters. J. Ferment.
21 Sih, C.J. (1987) Process for preparing Bioeng., 77, 152–158.
(S)-a-methylarylacetic acids, Wisconsin 31 Tosa, T. and Shibatani, T. (1995)
Alumni Research Foundation, USA, EP Industrial application of immobilized
227078 A1. biocatalysts in Japan. Ann. N. Y. Acad. Sci.,
22 Yamamoto, K., Ueno, Y., Otsubo, K., 750, 364–375.
Kawakami, K., and Komatsu, K.I. (1990) 32 Zaks, A. and Dodds, D.R. (1997)
Production of (S)-( þ )-ibuprofen from Application of biocatalysis and
a nitrile compound by Acinetobacter sp. biotransformations to the synthesis of
strain AK226. Appl. Environ. Microbiol., pharmaceuticals. Drug Discovery Today, 2,
56, 3125–3129. 513–531.
23 Zhang, Y.Y. and Liu, J.H. (2011) Kinetic 33 Zhao, L.L., Chen, X.X., and Xu, J.H.
study of enantioselective hydrolysis of (2010) Strain improvement of Serratia
(R, S)-ketoprofen ethyl ester using marcescens ECU1010 and medium cost
immobilized T. laibacchii lipase. reduction for economic production of
Biochem. Eng. J., 54 (1), 40–46. lipase. World J. Microbiol. Biotechnol.,
24 Foresti, M.L., Galle, M., and 26 (3), 537–543.
Ferreira, M.L. (2009) Enantioselective 34 Long, Z.D., Xu, J.H., and Pan, J. (2007)
esterification of ibuprofen with ethanol as Significant improvement of Serratia
reactant and solvent catalyzed by marcescens lipase fermentation, by
immobilized lipase: experimental and optimizing medium, induction, and
molecular modeling aspects. J. Chem. oxygen supply. Appl. Biochem. Biotechnol.,
Technol. Biotechnol., 84 (10), 1461–1473. 142 (2), 148–157.
25 Hernandez-Fernandez, F.J., Rios, A.P.D., 35 Shibatani, T., Omori, K., and
and Tomas-Alonso, F. (2009) Kinetic Akatsuka, H. (2000) Enzymatic
References j527
resolution of diltiazem intermediate by palladium as the racemization catalyst.
Serratia marcescens lipase: molecular Chimia, 50, 668–669.
mechanism of lipase secretion and its 44 Patel, R.N. (1992) Enantioselective
industrial application. J. Mol. Cat. B: enzymatic acetylation of racemic
Enzym., 10 (1–3), 141–149. [4-[4a,6bE]]-6-[4,4-bis(4-fluorophenyl)-3-
36 Shimizu, H., Ogawa, J., Kataoka, M., and (1-methyl-1H-tetrazol-5-yl)-1,3-
Kobayashi, M. (1997) Screening of novel butadienyl]-tetrahydro-4-hydroxy-2H-
microbial enzymes for the production of pyran-2-one. Appl. Microbiol. Biotechnol.,
biologically and chemically useful 38, 56–60.
compounds, in Advances in Biochemical 45 Patel, R.N. (1997) Stereoselective
Engineering Biotechnology, New biotransformations in synthesis of some
Enzymes for Organic Synthesis, vol. 58 pharmaceutical intermediates. Adv. Appl.
(ed. T. Scheper), Springer, New York, Microbiol., 43, 91–140.
pp. 45–88. 46 Morgan, B., Dodds, D.R., Zaks, A.,
37 Bonrath, W., Karge, R., and Netscher, T. Andrews, D.R., and Klesse, R. (1997)
(2002) Lipase-catalyzed transformations Enzymatic desymmetrization of
as key-steps in the large-scale preparation prochiral 2-substituted-1,3-propanediols:
of vitamins. J. Mol. Cat. B: Enzym., 19, a practical chemoenzymatic synthesis
67–72. of a key precursor of SCH51048, a
38 Liese, A., Seelbach, K., Buchholz, A., and broad-spectrum orally active antifungal
Haberland, J. (2006) Processes, in agent. J. Org. Chem., 62, 7736–7743.
Industrial Biotranformations (eds. A. 47 Morgan, B., Stockwell, B.R., Dodds, D.R.,
Liese, K. Seelbach, and C. Wandrey), Andrews, D.R., Sudhakar, A.R.,
Wiley-VCH, Weinheim. pp. 147–513. Nielsen, C.M., Mergelsberg, I., and
39 Hirohara, H. and Nishizawa, M. (1998) Zumbach, A. (1997) Chemoenzymatic
Biochemical synthesis of several approaches to SCH 56592, a new azole
chemical insecticide intermediates and antifungal. J. Am. Oil Chem. Soc., 74,
mechanism of action of relevant 1361–1370.
enzymes. Biosci. Biotechnol. Biochem., 48 Pantaleone, D.P. (1999)
62, 1–9. Biotransformations: green processes
40 Stead, P., Marley, H., Mahmoudian, M., for the synthesis of chiral fine chemicals,
Webb, G., Noble, D., Ip, Y.T., Piga, E., in Handbook of Chiral Chemicals
Rossi, T., Roberts, S., and Dawson, M.J. (ed. D.J. Ager), Marcel Dekker Inc.,
(1996) Efficient procedures for the New York, pp. 245–286.
large-scale preparation of (1S,2S)-trans-2- 49 Saksena, A.K., Girijavallabhan, V.M.,
methoxycyclohexanol, a key chiral Pike, R.E., Wang, H., Lovey, R.G.,
intermediate in the synthesis of tricyclic Liu, Y.-T., Ganguly, A.K., Morgan, W.B.,
b-lactam antibiotics. Tetrahedron: and Zaks, A. (1995) Process for preparing
Asymmetry, 7 (8), 2247–2250. intermediates for the synthesis of
41 Balkenhohl, F., Hauer, B., Ladner, W., antifungal agents, Schering Corporation,
Schnell, U., Pressler, U., and US 5,403,937.
Staudenmaier, H.R. (1995) Lipase 50 Saksena, A.K., Girijavallabhan, V.M.,
katalysierte acylierung von alkoholen Wang, H., Liu, Y.-I., Pike, R.R., and
mit diketenen, BASF AG, Ganguly, A.K. (1996) Concise
DE 4329293 A1. asymmetric routes to 2,2,4-trisubstituted
42 Balkenhohl, F., Ditrich, K., Hauer, B., and tetrahydrofurans via chiral titanium
Ladner, W. (1997) Optisch aktive amine imide enolates: key intermediates
durch lipase-katalysierte towards synthesis of highly active azole
methoxyacetylierung. J. Prakt. Chem., antifungals SCH 51048 and SCH 56592.
339, 381–384. Tetrahedron Lett., 37, 5657–5660.
43 Reetz, M.T. and Schimossek, K. (1996) 51 Zaks, A. and Dodds, D.R. (1997)
Lipase-catalyzed dynamic kinetic Application of biocatalysis and
resolution of chiral amines: use of biotransformations to the synthesis of
528
pharmaceuticals. Drug Discovery Today, cyclodextrin production process using

2 (6), 513–531. specific adsorbents, in Industrial
52 Goswami, A., Howell, J.M., Hua, E.Y., Application of Immobilized Biocatalysts
Mirfakhrae, K.D., Sourmeillant, M.C., (eds A. Tanaka, T. Tosa, and T. Kobayashi),
Swaminathan, S., Qian, X., Quiroz, F.A., Marcel Dekker Inc., New York,
Vu, T.C., Wang, X., Zheng, B., pp. 109–129.
Kronenthal, D.R., and Patel, R.N. (2001) 63 Tsuchiyama, Y., Yamamoto, K.-I.,
Chemical and enzymatic resolution of Asou, T., Okabe, M., Yagi, Y., and
(R,S)-N-(tert-butoxycarbonyl)-3- Okamoto, R. (1991) A novel process of
hydroxymethylpiperidine. Org. Process cyclodextrin production by use of specific
Res. Dev., 5, 415–420. adsorbents, Part I, screening of specific
53 Patel, R.N. (2008) Chemo-enzymatic adsorbents. J. Ferment. Bioeng., 71,
synthesis of pharmaceutical 407–412.
intermediates. Expert Opin. Drug 64 Yang, C.-P. and Su, C.-S. (1998) Study
Discovery, 3 (2), 187–245. of cyclodextrin production using
54 Hills, G.A., McCrae, A.R., and cyclodextrin glycosyltransferase
Poulina, R.R. (1990) Ester preparation, immobilized on chitosan. J. Chem.
Unichema Chemie BV, EP 0383405 B1. Technol. Biotechnol., 46, 283–294.
55 Kemp, R.A. and Macrae, A.R. (1992) 65 Kittelmann, M., Rheinegger, U.,
Esterification process, Unichema Espigat, A., Oberer, L., Aichholz, R.,
Chemie BV, EP 0506159 A1. Francotte, E., and Ghisalba, O. (2003)
56 McCrae, A.R., Roehl, E.-L., and Brand, Preparative enzymatic synthesis of the
H.M. (1990) Bio-ester – bio-esters. acylglucuronide of mycophenolic acid.
€
Seifen-Ole-Fette-Wachse, 116, (6), 201–205. Adv. Synth. Cat., 345, 825–829.
57 Hilterhaus, L., Thum, O., and Liese, A. 66 Ghisalba, O., Meyer, H.P., and
(2008) Reactor concept for lipase- Wohlgemuth, R. (2010) Industrial
catalyzed solvent-free conversion of biotransformations, in Encyclopedia
highly viscous reactants forming of Industrial Biotechnology. Bioprocess,
two-phase systems. Org. Process Res. Dev., Bioseparation, and Cell Technology
12 (4), 618–625. (ed. M.C. Flickinger), John Wiley & Sons,
58 Thum, O., Hilterhaus, L., and Liese, A. Inc., Hoboken.
(2009) Method for heterogeneously 67 Holm, J., Bjoerck, I., Ostrowska, S.,
catalysed manufacture of carboxylic acid Eliasson, A.C., Asp, N.G., Larsson, K.,
derivatives, Evonik Goldschmidt GmbH, and Lundquist, I. (1983) Digestibility of
EP 2080806A3. amylose-lipid complexes in vitro and
59 Wiemann, L.O., Nieguth, R., in-vivo. St€arke, 35, 294–297.
Eckstein, M., Naumann, M., Thum, O., 68 Kainuma, K. (1998) Applied glycoscience
and Ansorge-Schumacher, M.B. (2009) - past, present and future. Food Ingredients
Composite particles of Novozyme 435 J. Jpn., 178, 4–10.
and silicone: advancing technical 69 Labout, J.J.M. (1985) Conversion of
applicability of macroporous liquefied starch into glucose using a novel
enzyme carriers. ChemCatChem, 1 (4), glucoamylase system. St€arke, 37,
455–462. 157–161.
60 Kim, D., Kang, H.K., and Seo, E.S. (2006) 70 Daneel, H.J. and Geiger, R. (1994)
Potential industrial application of Verfahren zur herstellung von L-
glycosyltransferases and their evolutions. €apfels€aure aus fumars€aure, AMINO
J. Biomed. Nanotechnol., 2, 120–124. GmbH, DE 4424664 C1.
61 Buchini, S. and Withers, S.G. (2007) 71 Daneel, H.J. and Geiger, R. (1994)
Engineered glycosidases for glycoside Verfahren zur abtrennung von
synthesis. Chim. Oggi-Chem. Today, fumars€aure, maleins€aure und/oder
25, 4–6. bernsteins€aure von einem
62 Okabe, M., Tsuchiyama, Y., and hauptbestandteil €apfels€aure, AMINO
Okamoto, R. (1993) Development of a GmbH, DE 4430010 C1.
References j529
72 Daneel, H.J., Busse, M., and Faurie, R. 84 Plischke, H. and Steinmetzer, W. (1988)
(1995) Pharmaceutical grade L-malic acid Verfahren zur herstellung von
from fumaric acid – development of an L-tryptophan und D,L-serin, AMINO
integrated biotransformation and GmbH, DE 3630878 C1.
product purification process. Med. Fac. 85 Wagner, F., Klein, J., Bang, W.-G.,
Landbouww. Univ. Gent, 60/4a, Lang, S., and Sahm, H. (1980)
2093–2096. Verfahren zur mikrobiellen herstellung
73 Daneel, H.J., Busse, M., and Faurie, R. von L-tryptophan, DE 2841642 C2.
(1996) Fumarate hydratase from 86 Faurie, R. and Fries, G. (1999) From sugar
Corynebacterium glutamicum – process beet molasses to Lyphan (R) - integrated
related optimization of enzyme quality management from the raw
productivity for biotechnical L-malic acid material to the drug, in Tryptophan,
synthesis. Med. Fac. Landbouww. Univ. Serotonin, and Melatonin: Basic Aspects
Gent, 61/4a, 1333–1340. and Applications (eds G. Huether, W.
74 Mattey, M. (1992) The production of Kochen, T.J., Simat, and H. Steinhart),
organic acids. Crit. Rev. Biotechnol., 12, Kluwer Academic/Plenum Publishers,
87–132. New York, pp. 443–452.
75 Tosa, T. and Shibatani, T. (1995) 87 Chibata, I., Tosa, T., and Shibatani, T.
Industrial applications of immobilized (1992) The industrial production of
biocatalysts in Japan. Ann. N. Y. Acad. Sci., optically active compounds by
750, 364–375. immobilized biocatalysts, in Chirality
76 Tanaka, A., Tosa, T., and Kobayashi, T. in Industry (eds A.N. Collins, G.
(1993) Industrial Application of Sheldrake, and J. Crosby), John Wiley &
Immobilized Biocatalysts, Marcel Dekker Sons, Inc., New York, pp. 351–370.
Inc., New York. 88 Subramanian, S.S. and Raghavendra
77 Lilly, M.D. (1994) Advances in Rao, M.R. (1968) Purification and
biotransformation processes. Eighth P. V. properties of citraconase. J. Biol. Chem.,
Danckwerts memorial lecture presented 243, 2367–2372.
at Glaziers Hall, London, U.K. 13 May 89 van der Werf, M.J., van den Tweel, W.J.J.,
1993. Chem. Eng. Sci., 49, 151–159. and Hartmans, S. (1992) Screening for
78 Wiseman, A. (1995) Handbook of Enzyme microorganisms producing D-malate
and Biotechnology, Ellis Horwood, from maleate. Appl. Environ. Microbiol.,
Chichester. 58, 2854–2860.
79 Sheldon, R.A. (1993) Chirotechnology, 90 van der Werf, M.J., van den Tweel, W.J.J.,
Marcel Dekker Inc., New York. Kamphuis, J., Hartmans, S., and
80 Crosby, J. (1991) Synthesis of optically de Bont, J.A.M. (1994) D-Malate and
active compounds: a large scale D-citramalate production with maleate
perspective. Tetrahedron, 47, hydro-lyase from Pseudomonas
4789–4846. pseudoalcaligenes, in Proceedings of the
81 Kieslich, K. (1991) Biotransformations of 6th European Congress on Biotechnology
industrial use, 5th Leipzig Biotechnology (eds L. Alberghina, L. Frontali, and
Symposium 1990. Acta Biotechnol., 11 (6), P. Sensi), Elsevier Science B. V.,
559–570. Amsterdam, pp. 471–474.
82 Hasegawa, J., Ogura, M., Kanema, H., 91 Sheldon, R.A. (1993) Chirotechnology:
Noda, N., Kawaharada, H., and Industrial Synthesis of Optically Active
Watanabe, K. (1982) Production of D- Compounds, Marcel Decker, New York.
b-hydroxyisobutyric acid from isobutyric 92 Kulla, H.G. (1991) Enzymatic
acid by Candida rugosa and its mutant. hydroxylations in industrial application.
J. Ferment. Technol., 60, 501–508. Chimia, 45, 81–85.
83 Bang, W.-G., Lang, S., Sahm, H., and 93 Kitamura, M., Ohkuma, T., Takaya, H.,
Wagner, F. (1983) Production of and Noyori, R. (1988) A practical
L-tryptophan by Escherichia coli cells. asymmetric synthesis of carnitine.
Biotechnol. Bioeng., 25, 999–1011. Tetrahedron Lett., 29, 1555–1556.
530
94 Macy, J., Kulla, H., and Gottschalk, G. 100 Shimizu, H., Fujita, C., Endo, T., and
(1976) H2-dependent anaerobic growth Watanabe, I. (1993) Process for
of Escherichia coli on L-malate: succinate preparing glycine from glycinonitrile,
formation. J. Bacteriol., 125, 423–428. Nitto Chemical Industry Co., Ltd, US
95 Zimmermann, Th.P., Robins, K.T., 5238827.
Werlen, J., and Hoeks, F.W. (1997) 101 Shimizu, H., Ogawa, J., Kataoka, M.,
Bio-transformation in the production and Kobayashi, M. (1997) Screening of
of L-carnitine, in Chirality in Industry novel microbial enzymes for the
(eds A.N. Collins, G.N. Sheldrake, and J. production of biologically and chemically
Crosby), John Wiley and Sons, Inc., useful compounds, in New Enzymes for
New York, 287–305. Organic Synthesis, Advances in Biochemical
96 Banerjee, A., Sharma, R., and Engineering/Biotechnology, vol. 58 (eds T.K.
Banerjee, U.C. (2002) The nitrile-degrading Ghose, A. Fiechter, and N. Blakebrough),
enzymes: current status and future Springer Verlag GmbH, Berlin,
prospects. Appl. Microbiol. Biotechnol., 60, pp. 56–59.
33–44. 102 Yamada, H. and Kobayashi, M. (1996)
97 Hann, E.C., Eisenberg, A., Fager, S.K., Nitrile hydratase and its application to
Perkins, N.E., Gallagher, F.G., industrial production of acrylamide.
Cooper, S.M., Gavagan, J.E., Stieglitz, B., Biosci. Biotechnol. Biochem., 60 (9),
Hennesey, S.M., and DiCosimo, R. 1391–1400.
(1999) 5-Cyanovaleramide production 103 Yamada, H. and Tani, Y. (1987)
using immobilized Pseudomonas Process for biological preparation of
chlororaphis B23. Bioorg. Med. Chem., 7, amides, Nitto Chemical Industry Co., Ltd,
2239–2245. US 4637982.
98 Yamada, H., Ryuno, K., Nagasawa, T., 104 Petersen, M. and Kiener, A. (1999)
Enomoto, K., and Watanabe, I. (1986) Biocatalysis – preparation and
Agric. Biol. Chem., 50, 2859–2865. functionalization of N-heterocycles,
99 Nagasawa, T., Shimizu, H., and Green Chem., 2, 99–106.
Yamada, H. (1993) The superiority of the 105 Heveling, J. (1996) Catalysis at Lonza:
third-generation catalyst, Rhodococcus from metallic glasses to fine chemicals.
rhodochrous J1 nitrile hydratase, for Chimia, 50, 114–118.
industrial production of acrylamide.
Appl. Microb. Biotechnol., 40, 189–195.
j531
Part III
Hydrolysis and Formation of CN Bonds
and Oliver May.
j533
13
Hydrolysis of Nitriles to Amides
Alexander Yanenko and Steffen Osswald
13.1
Nitrile Hydratases
Under acid and alkaline condition nitriles undergo hydrolysis and are converted into
amides that, in turn, under such conditions behave as intermediates and are further
converted into carbonic acids. In the strict sense the first step is a hydration – an
addition of water to the CN triple bond – rather than a hydrolysis but both terms
have been used in literature.
The discovery of the nitrile hydratase enzyme (NHases, EC 4.2.1.84), which
catalyzes the hydrolysis of nitriles to amides, has created a new approach to amide
synthesis under mild conditions (neutral pH and room temperature). The application
of NHases for the hydrolysis of acrylonitrile to acrylamide proved to be the first
successful example of the use of enzymatic processes for the production of com-
modity chemicals. The acrylamide history and other examples of NHase application
made this enzyme one of the most intensively used enzymes in terms of production
volume in the chemical industry.
13.1.1
Occurrence and Classification of Nitrile Hydratases
NHase was first discovered in the cells of Rhodococcus (formerly Arthrobacter sp.) J-1
some 30 years ago [1], and Rhodococcus and other Actinomycetes are still regarded as
the most common source for novel NHases. Rhodococcus strains that contain NHase
enzymes are the dominant group among nitrile-utilizing microorganisms recovered
from marine sediments and terrestrial samples [2]. Moreover, Rhodococcus sp.
expressing nitrile hydratase activity can be selectively isolated through enrichment
with various nitriles, serving as a sole source of nitrogen (e.g., with substituted
2-phenylpropionitriles) [3]. Such enrichment strategies have enabled the discovery of
NHase in the cells of other genera, such as Agrobacterium, Alcaligenes, Bacillus, and
Pseudomonas [4, 5]. Enrichments with soda lake sediments resulted in the isolation of
novel species of microorganisms possessing NHase, such as haloalkaliphilic
and Oliver May.
j 13 Hydrolysis of Nitriles to Amides
534
actinobacterium – Nitriliruptor alkaliphilus and Natronocella acetonitrilicum [6]. The

application of strains capable of catalyzing enzymatic nitrile hydration under highly
alkaline conditions (pH 10) might have certain advantages, particularly when cyanide
is involved in the reaction process.
Recently, the sequence homology based screening approach has been used to
investigate the distribution of NHases in the environment. The application of PCR
screening protocols to environmental samples shows a wide distribution of NHase
genes in bacterial communities [7–9]. However, the use of specific primers (that
correspond precisely to highly conserved sequences, for example, metal-binding site)
for screening leads to the isolation of NHases with high homology (>90%) compared
to previously described [8]. A larger variety of NHases has been identified using
consensus-degenerate hybrid oligonucleotide primers [9] and real-time PCR
assay [10].
The molecular screening of NHases has allowed the identification of homological
sequences in all samples of the Global Ocean Sampling Expedition and for most
samples of the North Pacific Subtropical Gyre [11]. In addition, it enabled the
identification of the first eukaryotic NHase in the choanoflagellate Monosiga brevi-
collis [11]. Unlike bacterial NHase, two subunits of eukaryotic NHase are fused in one
protein, connected by a histidine-rich stretch. The newly discovered eukaryotic
NHase might be of biotechnological relevance due to its unconventional structure.
Using proteomics and genomic sequence analyses a new NHase (denoted ANHase)
possessing no significant amino acid sequence similarity with characterized NHases
was discovered [12]. Apparently this NHase is the first recognized member of a novel
class of NHases. In addition to its sequence, the ANHase of RHA1 differs from
characterized NHases in at least three additional respects: size, metal content, and
substrate specificity.
Depending on the metal ion in the catalytic center, well-characterized NHases are
classified into two groups – Fe-type and Co-type. Fe-type NHases have a non-heme
iron atom [13] and that Co-type NHases have a non-corrinoid cobalt atom [14] in the
catalytic center. Both types of NHases differ in their substrate specificity and activity
level, although their amino acid sequences exhibit significant homology, especially in
the active center. Several enzymes containing other ions (Zn or Cu) are known, but
there are no data for their participation in catalytic activity [12].
13.1.2
Protein Structure, Metal Cofactors, and Posttranslational Modifications
Nitrile hydratases share a homologous protein sequence and a distinctive three-

dimensional basic protein structure composed of two different subunits, designated
a and b. In most cases these subunits form larger heteromultimers but the minimal
functional unit, the ab dimmer, can be detected in solution [15]. The size of the a
subunits is in the range 23–29 kDa while the b subunits of 24–34.5 kDa are slightly
larger [16]. All known crystal structures of nitrile hydratases show a unique metal
center on the a subunit close to the a/b interface, containing either a Co(III)- or a Fe
(III)-ion (Figure 13.1). The metal is coordinated by two nitrogens from the protein
backbone and three cysteine sulfur atoms – two of the sulfur atoms are oxidized to
13.1 Nitrile Hydratases j535
Peptide
O
OH
O N S
H OH
Me
H OH
R N S
S
O
Cys Peptide
Peptide
Me: Fe3+, Co3+
Figure 13.1 Structure of the nitrile hydratase metal center.
sulfenic and sulfinic acid, respectively. One coordination site is occupied by a molecule
of water or OH. It has been shown that in vitro both cysteines are oxidized by
atmospheric oxygen in an autocatalytic manner after reconstitution of the unmodified
a and b subunits in the presence of 10 mM ferric citrate. Activity can only be detected
after oxidation but the maximum level of specific activity is limited to 15% compared to
the wild-type enzyme [17]. Coordination of other ligands to the free binding site at the
metal center has also been reported: nitric oxide (NO) inactivates the nitrile hydratase
from Rhodococcus sp. N-771 while light irradiation removes NO from the iron center
and converts the inactive form into the active one. Since the wild-type cells are able to
form NO the whole process can be regarded as photoregulation of the activity [18]. In
addition, cyanide is also known to be a strong inhibitor but the sensitivity of nitrile
hydratases varies vastly. The enzyme from Rhodococcus rhodochrous J1 shows a residual
activity of only 38% in the presence of 0.01 mM KCN [19] while the nitrile hydratase
from Rhodococcus opacus is highly active even at 100 mM cyanide [20]. The inhibition
by cyanide is technically relevant, for example, for the hydration of acrylonitrile and
amino nitriles which involves substantial amounts of HCN.
13.1.3
Reaction Mechanism
Depending on the mode of nitrile activation, the suggested reaction mechanisms

could be related to three different models [21]. According to the first model nitriles
directly bind to a metal ion in the active center and the activated nitrile carbon
undergoes nucleophilic attack of a water molecule. In the second and third models, a
water molecule activated by the metal directly or indirectly attacks nitriles detained
near the metal ion.
There are multiple facts in favor of the first mechanism, that is, the direct
interaction of nitrile with the metal ion in the active center. Direct binding of nitriles
with the metal is confirmed by the absorption spectra data [13]. It was shown that
iodoacetonitrile (substrate analog) binds directly to the metal ion [21]. X-Ray
structural studies for the complex of the NHase from Pseudonocardia thermophila
with butyric acid (weak competitive inhibitor) proved that the carboxylic oxygen binds
directly to the metal [22]. However, the exact mechanism remains unclear because of
a lack of detailed information on the reaction intermediates.
536
Recently, Hashimoto et al. [23] presented the first structural evidence for the catalytic
mechanism of a Fe-type NHase from Rhodococcus sp. N771. In addition to nitrile
hydration this enzyme also catalyzes the conversion of isonitriles [tert-butylisonitrile
(tBuNC)] into the corresponding amines (tert-butylamine). Owing to slow reaction of
tBuNC, the authors were able to investigate the time course of the tBuNC catalysis with
X-ray crystallography. The results confirmed that the substrate binds to metal directly.
In addition, the authors propose a new reaction mechanism in which the sulfenate
group of aCys114-SO plays a key role in catalysis. They revealed that aCys114-SOH
activates a water molecule, which further attacks the substrate–metal complex. The
above reaction mechanism is generally supported by theoretical calculations that
confirm that the oxygen of aCys114-SO could behave as a catalytic base [24].
13.1.4
Substrate Specificity
NHases usually show very broad substrate specificity and cannot be divided into
groups based on their activity for specific types of substrates like nitrilases. While
many NHases show a preference for aliphatic nitriles most of them are also able to
convert aromatic and heterocyclic ones even with bulky substituents [16, 25]. The
universalism of NHases is demonstrated by the fact that the same enzyme from
Rhodococcus rhodochrous J1 is used for the industrial-scale production of nicotinamide
as well as acrylamide (Section 13.1.3). Cyanohydrins and 2-aminonitriles are effi-
ciently hydrated only by a limited number of NHases because while decomposing in
aqueous solution they form cyanide, which is a potent inhibitor for most NHases
(Section 13.1.2).
The specific activities of most NHases are extremely high, for example, > 2000 U
mg1 protein for the NHase from Rhodococcus rhodochrous J1 [19]. Based on the data
available it can be estimated that specific activities of many other NHases are at least
in the same range.
13.1.5
Enantioselectivity
In accordance with the broad substrate specificity most NHases do not show efficient
discrimination of enantiomers. Numerous publications deal with the enantioselec-
tive biotransformations using whole cells of the wild-type bacterial strains. In these
cases it is sometimes difficult to find out which type of enzymes are involved and
which enzyme provides the enantioselectivity since NHases, amidases and also
nitrilases occur together in many different bacterial species.
Matsumoto et al. have investigated the two-step NHase-/amidase-catalyzed hydro-
lysis of 2-isopropyl-2-(4-chlorophenyl)acetonitrile and found that the amide is formed
with a very low e.e. at the beginning, which then increases to >99% e.e. while the (R)-
amide is converted into the carboxylic acid (Scheme 13.1). Therefore, it has been
proved that the NHase is almost non-selective while the amidase shows excellent
enantioselectivity [26].
13.2 Biocatalysts Containing Nitrile Hydratase j537
Pseudomonas
sp. B21C9
CN resting cells COOH CONH2
+
Cl Cl Cl
> 99% ee (S) >99% ee (R)
Scheme 13.1
Song et al. have purified the Fe-type NHase from Rhodococcus sp AJ270 and tested
its enantioselectivity for seven substrates [27]. The highest selectivity of >80% e.e.
was found for cyclopropane-carbonitriles – 95% e.e. for cis-2,2-dimethyl-3-phenylcy-
lopropancarbonitrile though activity was quite low and <5 mM of the substrate was
converted.
Prepechalova et al. have tested the NHase from Rhodococcus equi A4 for the
conversion of Profen-like compounds and determined enantioselectivities (ES) of
5–15 [28].
Tucker et al. have developed a NHase-based resolution process of 2-(2-oxopyrro-
lidin-1-yl)butanamide to produce levetiracetam [29]. Six NHases were expressed in
Escherichia coli and tested. The best of these enzymes was optimized by a semi-
rational mutagenesis approach: candidate residues were identified by computer
modeling and a mutant library was constructed by saturation mutagenesis at the
identified positions. The best mutant found produced the product with 92% e.e.
(Scheme 13.2).
13.2
Biocatalysts Containing Nitrile Hydratase
13.2.1
Whole-Cell Biocatalysts – Native Strains
Native microbial strains in the free or immobilized state are, so far, the most used
whole-cell biocatalysts for the hydration of nitriles to amides. Native strains with
3.3 g of NHase
O from NH33 per O
N N
100 g of substrate
CN CONH2
92% ee (S)
43-46% conversion
Scheme 13.2
538
different characteristics suitable for industrial application are available by screening

and optimization of the culture conditions rather than by genetic manipulation.
The important factor is that native strains can produce an enormous amount
of nitrile hydratase (NHase) when they are cultivated under optimal conditions.
The amount of NHase in microbial cells can exceed 40% or even 50% of the total
soluble protein [30, 31]. Another reason for the widespread use of whole cells as
biocatalysts is the low stability of most of the isolated enzymes under operating
conditions [19, 27].
An indispensable condition for the formation of NHase activity is the presence of
Fe ions (for Fe-type NHases) or Co ions (for Co-type NHases) in the growth
medium [32]. This is because Fe or Co ions are part of the active sites of NHases.
In addition, Bacillus smithii SC-J05-1 required the non-coordinated ions of manganese
for the production of active NHase [33]. No other ions (Zn, Cu, Ni) are obligatory for
NHase activity. Moreover, the cobalt ions were indispensable not only for formation of
activity but also for the synthesis of the NHase protein in R. rhodochrous J1 induced by
crotonamide [34]. It was shown by Northern blot analysis that cobalt ions regulate
expression of NHase gene in Rhodococcus rhodochrous M8 [35]. The addition of cobalt
ions leads to a significant increase of the NHase gene transcription. However, the
cobalt ions can act not only at the transcription level but also at other levels,
for example, enhancing the enzyme folding. Recently, it was found a unique
mechanism of posttranslation maturation of NHase by cobalt incorporation into
enzyme [36, 37].
Another important precondition for the formation of NHase activity at high levels
is the addition of an inducer in the medium. Unlike many enzymes for which a
substrate acts as inducer, NHase is induced by the reaction product and their
analogues. Acetamide, crotonamide, methacrylamide, urea, and e-caprolactam are
the most effective inducers of NHase [30, 34, 38–40]. However, NHases in some
native strains are constitutive and do not require any inducers (Table 13.1). Notably,
strains showing a constitutive synthesis of NHase can easily be selected
experimentally [41].
The level of NHase activity depends mostly upon carbon and nitrogen sources,
used for the cultivation of the strains. Usually, the synthesis of NHase in Rhodococcus
frequently used for transformations is not sensitive to glucose [38, 42]. However, if
the concentration of glucose in the cultivation medium is rather high (30–40 g l1)
NHase activity can be reduced. To avoid catabolic repression during the production of
high biomass concentrations at high specific NHase activity a fed-batch cultivation
strategy or a cultivation with a change of substrate is usually employed [43, 44].
Nitrile metabolism provides the cell a nitrogen source. It is known [45] for different
bacteria that utilization of nitrogen-containing compounds is regulated by nitrogen
catabolite control. Probably, NHases involved in the nitrile metabolism are also
subject to such a nitrogen catabolite control. It was shown that the production of
inducible NHase by Pseudomonas chlororaphis B23 was decreased approximately
twofold after the addition of NH4SO4 as nitrogen source [46]. In the case of
R. rhodochrous M8 the same reduction of NHase activity after ammonium addition
was due to the decrease of NHase gene expression [43]. The production of NHase by
Table 13.1 Characteristics of most used NHases from different sources.
Type of NHase Source Preferred substrates NHase synthesis Light activation Reference
Fe-type Rhodococcus sp. R312 Aliphatic nitriles Constitutive Yes [66]

Rhodococcus sp. N774 Aliphatic nitriles Constitutive Yes [67]
Pseudomonas chlororaphis B23 Aliphatic nitriles; adiponitrile Inducible by methacrylamide No [30]
R. equi A4 Aromatic, heterocyclic, and aryl- Inducible by methacrylamide Not determined [68]
aliphatic nitriles; enantioselective for
a-arylpropionitrile
Rhodococcus sp. AJ270 Aliphatic, heterocyclic, and aromatic Inducible by acetamide No [69]
nitriles; enantioselective for 3-phe-
nylcyclo-propanecarbo-nitriles
Rhodococcus sp. JCM6823 Aliphatic, heterocyclic, and aromatic Inducible e-caprolactam Not determined [40]
nitriles
Co-type R. rhodochrous J1 (H-NHase) Aliphatic, heterocyclic, and aromatic Inducible by urea No [34]
nitriles
Rhodococcus rhodochrous M33 Aliphatic, heterocyclic, and aromatic Constitutive No [50]
nitriles
Pseudomonas putida NRRL-18668 Aliphatic aryl-2-substituted nitriles; Constitutive No [70]
enantioselective
Bacillus pallidus RAPc8 Linear, cyclic, and branched aliphatic Constitutive No [71]
nitrile, dinitrile
Nocardia sp. RS-1 Acrylonitrile [57]
Rhodococcus ruberTH Acrylonitrile Inducible by urea Not determined [51]
Co, Cu, Zn Rhodococcus jostii RHA1 Aliphatic nitriles Inducible by acetonitrile Not determined [12]
13.2 Biocatalysts Containing Nitrile Hydratase
j539
540
B. smithii SC-J05-1 is strictly repressed by ammonium and is initiated in the absence

of ammonium [33].
Thus, control of the concentration of ammonium and glucose is essential for
improvement of enzyme production in different microorganisms that can utilize
nitriles.
The disadvantages of native strains as biocatalysts are determined by versatility of
nitrile metabolism in microbial cells. Any substitution in the medium composition
can lead to a changed enzyme expression profile and, as a result, to the appearance of
new activities in cells. For example, R. rhodochrous J1 produces two types of NHases, a
nitrilase, and several amidases [31, 47, 48] if cultivation conditions are varied. In
cobalt-containing medium urea induces the production H-NHase (520 kDa) and
cyclohexanecarboxamide the L-NHase (130 kDa), while both H and L-NHases were
induced by crotonamide. It should be considered that both NHases exhibit different
physicochemical properties and substrate specificities. An increase of NHase induc-
er concentration (methacrylamide) leads to the appearance of new activities (nitrilase
and amidase) in Rhodococcus sp. CGMCC0497, which significantly reduces the
enantioselectivity of the biocatalyst [49].
The amidase activity is one of the main disadvantages of natural strains as
biocatalysts. Owing to the residual amidase activity of biocatalysts, amides obtained
by means of such biocatalyst usually contain organic acids as by-products. In many
strains that utilize nitriles amidases are included in the NHase operons (or clusters)
and are induced by the same inducers as NHases [16].
One of the approaches to decrease the amidase activity is inactivation or deletion
of the amidase gene. Strain M33 is a mutant of the strain M8 that has a deleted
amidase gene. As a result the amidase activity of M33 in comparison with the parent
strain M8 is reduced by a factor of 20, from 0.3 to 0.015 U mg1 [50]. In the strain
R. ruber TH amidase activity was inactivated by an insertion into the amidase gene.
However, even in this case amidase activity was not completely abolished [51].
Genome analyzes of Rhodococcus sp. RHA1 [52] revealed that 24 genes encoding
different functional amidases probably exist in this bacterium. It is possible that
Rhodococcus rhodochrous M33 and R. ruber TH contain additional genes coding for
proteins that exhibit some degree of amidase activity. The alternative approach to
reducing amidase activity is the usage of amidase inhibitors such as aldehydes,
2-methyl-1-propaneboronic acid, and so on, which are added to the reaction
medium [20, 53, 54].
However, in this case it is necessary to eliminate the inhibitors from the final
product. Reduction of the transformation temperature is one of the most frequently
employed approaches to reducing amidase activity [55].
Thus, optimization of the cultivation medium and conditions of transformation
allow for effective application of biocatalysts on the basis of native strains. At present
all the industrial biocatalysts for acrylamide and nicotinamide production are native
strains or their mutants selected by traditional methods [50, 56–58]. Information
about the usage of native strains for different transformations has been reviewed in
detail [5, 59]. All the above-mentioned points allow us to confirm that the potential of
native strains as biocatalyst for nitrile hydrolysis is not exhausted.
13.2 Biocatalysts Containing Nitrile Hydratase j541
13.2.2
Whole-Cell Biocatalysts – Recombinant Strains
In principle, heterologous expression of NHases provides several advantages over

wild-type strains:
. No side-activities caused by other NHases, amidases or nitrilases have to be
considered.
. The use of standard expression hosts facilitates the efficient induction at high
biomass concentrations without the limitation of media composition.
. Expression levels are usually high also in cases in which the amount of NHase in
the wild-type cells is low.
. Enzyme optimization via directed evolution is more straightforward with plasmid
coded enzymes.
Heterologous expression of NHases has been shown in several hosts, such as
E. coli [60], Rhodococcus [61], and Pichia [62]. In addition to the two structure genes,
a subunit and b subunit, the expression of an activator protein, sometimes also
designated as a metallochaperone, which is usually part of the native gene clusters, is
required for efficient expression of Co- and Fe-dependent NHases (Figure 13.2) [60].
Only for the thermostable NHase from Bacillus BR449 has an expression without
activator protein and an activation of the apoenzyme in the presence of Co2 þ at 50 C
been reported [63]. The activator protein is assumed to be involved in cobalt/iron
insertion rather than cysteine oxidation [60]. For Co-dependent NHases, in some
cases also a fourth gene – besides amidases – was present on the gene cluster, which
was highly homologous to Co/Ni transporters and has been shown to increase the
Co2+-Ions
Co-Transporter
Metallochaperone
“P15K”
α β α β
Nitrile hydratase
structure genes
Figure 13.2 Suggested mechanism of NHase formation involving a metallochaperone and a cobalt
transporter.
542
activity of the NHase of R. rhodochrous J1 while co-expressed in the presence of

Co2 þ [64].
In most cases the recombinant strains show excellent activity. The NHase from
Commamonas testosteroni 5-MGAM-4D expressed in E. coli using the T7 expression
system reached a level of 7752 U g1 dry cell weight compared to 2577 U g1 dry cell
weight of the wild-type strain [65]. Surprisingly, enzyme stability within the E. coli
whole-cell catalyst was also higher than that of the wild-type cells. The catalyst showed
no significant loss of activity after three recycle reactions at 35 C at acrylonitrile
concentrations up to 160 g l1.
13.3
Summary and Outlook
Nitrile hydratase catalyzed processes have replaced several chemical production

methods at the multi-1000 tons scale. Their main advantages are an exceptionally
high specific activity and a virtually perfect chemoselectivity, which permits access to
extremely pure amides without the formation of the respective acids. Despite their
complex structure and the essential posttranslational oxidation of cysteine residues,
nitrile hydratases can be produced efficiently in wild-type bacteria as well as in
recombinant expression hosts. For most applications whole-cell catalysts are applied
at laboratory scale and also at production scale since the stability of the isolated
enzymes is substantially lower. The very broad substrate spectrum of most nitrile
hydratases facilitates their use in organic synthesis and based on recent publications
of regio- and enantioselective conversions an increasing number of applications can
be assumed in the coming years.
References
1 Asano, Y., Tani, Y., and Yamada, H. (1980) 7 Liebeton, K. and Eck, J. (2004) Eng. Life
Agric. Biol. Chem., 44, 2251–2252. Sci., 4, 557–562.
2 Brandao, P., Clapp, J.P., and Bull, A.T. 8 Precigou, S., Goulas, P., and Duran, R.
(2002) Environ. Microbiol., 4, 262–327 (2001) FEMS Microbiol. Lett., 204,
3 Layh, N., Hirrlinger, B., Stolz, A., and 155–161.
Knackmuss, H.-J. (1997) Appl. Microbiol. 9 Lourenco, P., Almeida, T., Mendonca, D.,
Biotechnol., 47, 668–674. Simoes, F., and Novo, C. (2004) J. Basic
4 Bunch, A.W. (1998), Nitriles in Microbiol., 44, 203–214.
Biotechnology, Volume 8a, 10 Coffey, L., Owens, E., Tambling, K.,
Biotransformations I (eds H.J. Rehm and ONeill, D., OConnor, L., and OReilly, C.
G. Reed), Wiley-VCH Verlag GmbH, (2010) Antonie van Leeuwenhoek, 98 (4),
Weinheim, ch. 6, pp. 277–324. 455–463.
5 Mylerova, V. and Martinkova, L. (2003) 11 Foerstner, K., Doerks, T., Muller, J.,
Curr. Org. Chem., 7, 1–17. Raes, J., and Bork, P. (2008) PLoS One, 3,
6 Sorokin, D., van Pelt, S., Tourova, T.P., and e3976.
Evtushenko, L.I. (2009) Int. J. Syst. Evol. 12 Okamoto, S. and Eltis, L. (2007)
Microbiol., 59, 248–253. Mol. Microbiol., 65, 828–838.
References j543
13 Sugiura, Y., Kuwahara, J., Nagasawa, T., 29 Tucker, J., Xu, L., Yu, W., Williams, S.,
and Yamada, H. (1987) J. Am. Chem. Soc., Zhao, L., and Ran, N. (2008) Patent
109, 5848–5850. application, WO 2009/009117.
14 Brennan, B., Alms, G., Nelson, M., 30 Nagasawa, T., Ryuno, K., and Yamada, H.
Durney, L., and Scarrow, R. (1996) (1989) Experientia., 45, 1066–1070.
J. Am. Chem. Soc., 118, 9194–9195. 31 Kobayashi, M., Nagasawa, T., and
15 Nakasako, M., Odaka, M., Yohda, M., Yamada, H. (1992) Trends Biotechnol.,
Dohmae, N., Takio, K., Kamiya, N., and 10, 402–408.
Endo, I. (1999) Biochemistry, 38, 32 Kobayashi, M. and Shimizu, S. (1998)
9887–9898. Nat. Biotechnol., 16, 733–736.
16 Cowan, D., Cameron, R., and Tsekoa, T. 33 Takashima, Y., Kawabe, T., and Mitsuda, S.
(2003) Adv. Appl. Microbiol., 52, 123–158. (2000) J. Biosci. Bioeng., 89, 282–284.
17 Murakami, T., Taku, N., Nojiri, M., 34 Nagasawa, T., Takeuchi, K., and
Nakayama, H., Odaka, M., Yohda, M., Yamada, H. (1988) Biochem. Biophys.
Dohmae, N., Takio, K., Nagamune, T., and Res. Commun., 155, 1008–1016.
Endo, I. (2000) Protein Sci., 9 (5), 35 Pogorelova, T., Ryabchenko, L.,
1024–1030. Sunzov, N., and Yanenko, A. (1996)
18 Endo, I. and Odaka, M. (2000) FEMS Microbiol. Lett., 144, 191–195.
J. Mol. Catal., B: Enzym., 10 (1–3), 81–86. 36 Zhou, Z., Hashimoto, Y., Shiraki, K., and
19 Nagasawa, T., Takeuchi, K., and Yamada, Kobayashi, M. (2008) Proc. Natl. Acad. Sci.
H. (1991) Eur. J. Biochem., 196 (3), 581–589 U.S.A., 105, 14849–14854.
20 Osswald, S., Weckbecker, C., 37 Zhou, Z., Hashimoto, Y., and
Huthmacher, K., Gerasimova, T., Kobayashi, M. (2009) J. Biol. Chem., 284,
Novikov, A., Ryabchenko, L., Yanenko, A., 14930–14938.
and Egorova, K. (2005) PCT Int. Appl., WO 38 OMahony, R., Doran, J., Coffey, L.,
2005090394. Cahill, O.J., Black, G.W., and OReilly, C.
21 Huang, W., Jia, J., Cummings, J., (2005) Antonie Van Leeuwenhoek., 87,
Nelson, M., Schneider, G., and 221–232.
Lindqvist, Y. (1997) Structure, 5, 691–699. 39 Precigou, S., Wieser, M., Pommares, P.,
22 Miyanaga, A., Fushinoba, S., Ito, K., and Goulas, Ph., and Duran, R. (2004)
Shoun, H., and Wakagi, T. (2004) Biotechnol. Lett., 26, 1379–1384.
Eur. J. Biochem., 271, 429. 40 Duran, R., Nishiyama, M., Horinouchi, S.,
23 Hashimoto, K., Suzuki, H., Taniguchi, K., and Beppu, T. (1993) Biosci. Biotechnol.
Noguchi, T., Yohda, M., and Odaka, M. Biochem., 57, 1323–1328.
(2008) J. Biol. Chem., 283, 36617. 41 Yanenko, A., Astaurova, O., Gerasimova, T.,
24 Yono, T., Wasada-Tsatsui, Y., Ari, H., Polyakova, I., Pogorelova, T., and Paukov, V.
Yamaguchi, S., Funahashi, Y., Osawa, T., (1994) Regulation of nitrile utilization in
and Masada, H. (2007) Inorg. Chem., 46, Rhodococcus in Proceeding of the 9th
10345. International Symposium on the Biology of
25 Brady, D., Beeton, A., Zeevaart, J., Actinomycetes July 10–15, 1994, Moscow,
Kgaje, C., Van Rantaijk, F., and Russia (ed. V.G. Debabov), Allerton Press,
Sheldon, R. (2004) Appl. Microbiol. pp. 139–144.
Biotechnol., 64, 76–85. 42 Bernet, N., Arnaud, A., and Galzy, P.
26 Matsumoto, S., Inue, A., Kumagai, K., (1990) Biocatalysis, 3, 259–267.
Murai, R., and Mitsuda, S. (1995) 43 Leonova, T., Astaurova, O.,
Biosci. Biotechnol. Biochem., 59 (4), Ryabchenko, L., and Yanenko, A.
720–722. (2000) Appl. Biochem. Biotechnol., 88,
27 Song, L., Wang, M., Yang, X., and Qian, S. 231–241.
(2007) Biotechnol. J., 2, 717–724. 44 Lorenz, J., Voronin, S., Kosulin, S.,
28 Prepechalova, I., Martinkova, L., Stolz, A., Singirzev, I., Debabov, V., and Yanenko, A.
Ovesna, M., Bezouska, K., Kopecky, J., and (2004) EP1737946.
Kren, V. (2001) Appl. Microbiol. Biotechnol., 45 Merrick, M. and Edwards, R. (1995)
55, 150–156. Microbiol. Rev., 59, 604–622.
544
46 Yamada, H., Ryuno, K., Nagasawa, T., 59 Chen, J., Zheng, R.-C., Zheng, Y.-G., and
Enomoto, K., and Watanabe, I. (1986) Shen, Y.-C. (2009) Adv. Biochem. Eng./
Agric. Biol. Chem., 50, 2859–2865. Biotechnol., 113, 33–77.
47 Kobayashi, M., Nagasawa, T., and 60 Nojori, M., Yohda, M., Okada, M.,
Yamada, H. (1989) Eur. J. Biochem., 182, Matsushita, Y., Tsujimura, M., Yoshida, T.,
349–356. Dohmae, N., Takio, K., and Endo, I. (1999)
48 Kobayashi, M., Komeda, H., Nagasawa, T., J. Biochem., 125, 696–704.
Yamada, H., and Shimizu, S. (1993) 61 Mizunashi, W., Nishiyama, M.,
Biosci. Biotechnol. Biochem., 11, Morinouchi, S., and Beppu, T. (1998) Appl.
1949–1950. Microbiol. Biotechnol., 49, 568–572.
49 Wu, Zh.-L. and Li, Z.-Y. (2002) Biotechnol. 62 Wu, S., Fallon, R., and Payne, M. (1999)
Appl. Biochem., 35, 61–67. Appl. Microbiol. Biotechnol., 52, 186–190.
50 Yanenko, A., Astaurova, O., Voronin, S., 63 Kim, S., Padmakumar, R., and Oriel, P.
Gerasimova, T., Kirsanov, N., Paukov, V., (2001) Appl. Biochem. Biotechnol., 91–93,
Polyakova, I., and Debabov, V. (1998) US 597–603.
Patent, 5827699. 64 Komeda, H., Kobayashi, M., and Shimizu,
51 Ma, Yu., Yu, H., Pan, W., Liu, Ch., S. (1997) Proc. Natl. Acad. Sci. U.S.A., 94,
Zhang, Sh., and Shen, Zh. (2010) 36–41.
Bioresour. Technol., 101 285–291. 65 Petrello, K., Wu, S., Hann, E., Cooling, F.,
52 McLeod, M. et al. (2006) Proc. Natl. Acad. Ben-Basset, A., Gavagan, J., DiCosimo, R.,
Sci. U.S.A., 103, 15582–15587; http:// and Payne, M. (2005) Appl. Microbiol.
www.rhodococcus.ca. Biotechnol., 67, 664–670.
53 Amarant, T., Vered, Y., and Bohak, Z. 66 Bui, K., Arnaud, A., and Galzy, P. (1982)
(1989) Biotechnol. Appl. Biochem., 11, Enzyme Microbiol. Technol., 4, 195–197.
49–59. 67 Watanabe, I., Satoh, Y., Enomoto, K., Seki,
54 Graham, D., Pereira, R., Barfield, D., and S., and Sakashita, K. (1987) Agric. Biol.
Cowan, D. (2000) Enzyme Microb. Technol., Chem., 51, 3201–3206.
26, 368–373. 68 Martinkova, L., Stolz, A., and Knackmuss,
55 Raj, J., Sharma, N.N., Prasad, Sh., and H. (1996) Biotechnol. Lett., 18, 1073–1076.
Bhalla, T.Ch. (2008) J. Ind. Microbiol. 69 Blakey, A., Colby, J., Williams, E., and
Biotechnol., 35, 35–40. OReilly, C. (1995) FEMS Microbiol. Lett.,
56 Yamada, H. and Nagasawa, T. (1990) 129, 57–62.
EP0362829. 70 Fallon, R., Stieglitz, B., and Turner, I.
57 Liu, M., Li, C., Gao, Y., Huang, Y., and (1997) Appl. Microbiol. Biotechnol., 47,
Cao, Z. (2004) Huagong Xuebao, 55, 156–161.
1678. 71 Pereira, R., Graham, D., Reiney, F., and
58 Chassin, C. (1996) Spec. Chem., 16, Cowan, D. (1998) Extremophiles, 2,
102–103. 347–357.
j545
14
Hydrolysis of Nitriles to Carboxylic Acids
Steffen Oßwald and Alexander Yanenko
14.1
Introduction
The one-step hydrolysis of nitriles to carboxylic acids is an unusual process for

conventional organic synthesis. As a general rule, the conventional process of nitrile
hydrolysis to carboxylic acids occurs through the formation of the amide as an
intermediate and requires harsh acidic or basic reaction conditions. The main
disadvantages of such a process are the formation of undesirable by-products, low
yields, and large quantities of waste, for example, inorganic salts. In contrast, the direct
conversion of nitriles into organic acids is widely observed in living organisms and it is
catalyzed by nitrilases – a group of enzymes, first discovered in plants [1]. The
enzymatic conversion of nitriles takes place under mild conditions (low temperatures
and neutral pH) with high conversion and selectivity, and is, in certain cases, enantio-
and regioselective [2, 3]. The capability of nitrilases to act selectively only on CN groups,
while not affecting other labile groups of the molecule, represents a major benefit of
the enzymatic process. Owing to these advantages nitrilases have found wide
application in organic synthesis for the production of pharmaceuticals and chemicals.
More than five nitrilase-based processes have already been introduced in industry.
14.2
Nitrilases
According to the Nomenclature Committee of IUBMB, a nitrilase (E.C. 3.5.5.1.) is an

enzyme catalyzing the hydrolysis of organic cyanides or nitriles into their corre-
sponding carboxylic acids and ammonia, without formation of a free amide as an
intermediate. Based on sequence homology considerations, nitrilases, together with
other CN bond cleaving enzymes, belong to a larger group of related proteins
known as CN-hydrolases [4] – more recently designated as the nitrilase super-
family [5]. Besides nitrilases, the nitrilase branch of the nitrilase superfamily con-
sists of cyanide dihydratases (CDH), also designated as cyanidases, and cyanide
hydratases (CH), which only use HCN, not the nitriles, as the efficient substrate [6].
and Oliver May.
j 14 Hydrolysis of Nitriles to Carboxylic Acids
546
The nitrilase- related enzymes have similar structures, with an a-b-b-a- sandwich
fold, they utilize a thiol mechanism, and contain conserved a Glu-Lys-Cys catalytic
triad at the catalytic site [7, 8].
14.2.1
Occurrence and Classification of Nitrilases
For a long time, supported by traditional enrichment strategies, it seemed that

nitrilases occurred only in several bacterial species that either belong to the Proteo-
bacteria (such as Acinetobacter, Alcaligenes, or Pseudomonas strains) or to the high-GC
Gram-positive bacteria (mostly rhodococci) [9–11]. However, the application of cul-
ture-independent screening methods, and also bulk genome sequencing, demon-
strated that nitrilases are largely presented in various taxons, including plants, fungi,
archaea, and different bacteria [12–14]. It appears that a wide distribution of nitrilases
results from their key role in metabolism of organocyanides in plants, animals, and
microorganisms [15].
According to the substrate specificity, nitrilases isolated from different sources
were subdivided into three groups (Table 14.1) [10, 11]: (i) aromatic nitrilases that
hydrolyze predominantly aromatic or heterocyclic nitriles like benzonitrile (e.g., the
nitrilase from Rhodococcus rhodochrous J1or fungal nitrilases); (ii) aliphatic nitrilases
that preferentially hydrolyze aliphatic nitriles (e.g., the nitrilase from Rhodococcus
rhodochrous K22 or Acidovorax facilis 72W); (iii) nitrilases that hydrolyze preferentially
arylacetonitriles (e.g., arylacetonitrilases from Alcaligenes faecalis JM3). This classi-
fication has been commonly in use up to the present day, despite the fact that some
enzymes fit simultaneously in several groups, such as nitrilases from Synechocystis
sp. PCC6803 and Bradyrhizobium japonicum USDA110.
14.2.2
Protein Structure and Oligomerization
Nitrilases are homooligomeric proteins [9]. Nitrilase subunits, similar to other

members of the nitrilase subfamily share a characteristic a-b-b-a fold [7]. Most
nitrilases exist in solution in the form of non-active dimers, which are able to
associate to the active oligomers [11]. The association of the subunits can be initiated
by the presence of a substrate (substrate activation), and, also, through the addition
of certain salts in high concentration [(NH4)SO4, NH4Cl, MgSO4, KCl], or with
organic solvents, heat treatment, or at high enzyme concentrations [16–19]. The
phenomenon of substrate activation is typical for nitrilases isolated from the genus
Rhodococcus [17–19]. However, other nitrilases demonstrate substrate-dependent
activation as well, for example, Alcaligenes faecalis ATCC8750 [20].
Using a combination of negative stain electron microscopy and the docking of
homology models it was shown that certain microbial nitrilases can form regular
structures – homo-oligomeric spirals or helices of variable length [11, 21].
The nitrilase from Rhodococcus rhodochrous J1, one of the best characterized
enzymes, exists in three homo-oligomeric structural forms – a dimer, a 480 kDa
Table 14.1 Characteristics of the most used nitrilases from different microorganisms.
Type of nitrilase Source Preferred substrates Selectivity Amide Reference

formation
Aliphatic nitrilase Rhodococcus rhodochrous K22 Acrylonitrile, glutaronitrile, Not determined No [36]
crotononitrile
Acidovorax facilis 72W Fumaronitrile, 2-methylglutaronitrile Regioselective for aliphatic dinitrile No [38]
Acinetobacter sp. (strain Acrylonitrile, 2-(4-isobutylphenyl)- Enantioselective for (S)-aryl No [46]
AK226) propionitrile, benzonitrile propionitriles
Pyrococcus abyssi Fumaronitrile malononitrile Regioselective for aliphatic dinitriles No [14]
Arabidopsis thaliana, NIT1 3-Phenylpropionitrile, octanenitrile, Regioselective, (E)-selective, Yes, 0–95% [28, 30]
butyronitrile enantioselective for R-a-fluoroaceto-
nitrile
Bradyrhizobium japonicum Hydrocinnamonitrile, heptanenitrile, Moderate enantioselectivity, No [64]
USDA 110, blr3397 phenylacetonitrile regioselective
Synechocystis sp.PCC6803 Fumarodinitrile, naphthalenecarboni- Regioselective for dinitrile Not [59]
trile, adipinic acid dinitrile, determined
3-butenenitrile
Aromatic nitrilase Rhodococcus rhodochrous J1 Benzonitrile, 2-furonitrile, 4-tolunitrile Regioselective for dinitrile No [53]
Bacillus pallidus Dac521 4-Cyanopyridine, benzonitrile, Not determined No [58]
4-chlorobutyronitrile
Aspergillus niger K10 4-Cyanopyridine, benzonitrile, Not determined Yes, 0–84% [55]
1,4-dicyanobenzene
Fusarium solani O1 4-Cyanopyridine, benzonitrile, Not determined 1–3% [56]
3-hydroxybenzonitrile
Arylaceto-nitrilase Alcaligenes faecalis JM3 p-Fluorobenzyl cyanide, 2-thiophenea- Not determined No [54]
cetonitrile, 3-pyridineacetonitrile
Pseudomonas fluorescens 2-Phenylvaleronitrile, 2-thiopheneace- Enantioconservative for Yes, 0–43% [31]
EBC191 tonitrile, phenylacetonitrile (S)-and (R)-nitriles
14.2 Nitrilases
Pseudomonas putida Amino- or hydroxyl – phenylacetonitrile, Enantioselective for (R)- No [51]

j547
MTCC5110 mandelonitrile mandelonitrile

548
complex, and a regular helix of variable length. In solution this enzyme exists as an
inactive dimer that oligomerizes to form an active 480 kDa complex in the presence of
benzonitrile [19]. Active helical oligomers were found only for recombinant enzymes,
isolated from E. coli [22]. Thuku et al. have shown that such a recombinant enzyme
undergoes posttranslational cleavage at approximately residue 327. Upon the loss of
39 C-terminal amino acid a protein acquires the ability to form active, helical homo-
oligomers [22]. Similar helical filaments were demonstrated for cyanide dihydratase
from Bacillus pumilis under certain pH conditions [23].
To date, the crystal structure of a microbial nitrilase is not available. A high-
resolution structure determination will provide more insight into the mechanism of
enzyme activation under oligomerization processes.
Although the role and functional significance of helical assembly in vivo remains
unclear, there are promising perspectives for biotechnological applications of arti-
ficially made helices. Long helices are likely to be effective biocatalysts because they
have a high concentration of active sites, and they can easily be purified and stored,
probably, for a long time.
14.2.3
Reaction Mechanism
The catalytic mechanism of nitrilases is as yet hypothetical due to the lack of crystal
structures. It is generally accepted, nevertheless, that nitrilases contain a Glu-Lys-Cys
catalytic triad and utilize a thiol mechanism via a thioimidate intermediate as
originally proposed by Stevenson et al. [17].
The mechanism suggests a nucleophilic attack on the nitrile carbon atom by
a conserved cysteine residue of the nitrilase, with the formation of a thioimidate
that subsequently affords a tetrahedral intermediate with the addition of water
(Figure 14.1). The glutamate acts as a general base and the lysine residues are
involved in the stabilization of a tetrahedral transition state [7, 8]. It is assumed that
the tetrahedral intermediate can be broken down by two pathways, which lead to
different products [24]. The first (pathway 1) includes NH3 elimination from this
intermediate to give a thioester, which reacts with a second water molecule to give the
carboxylic acid, which is the usual route for nitrilases and cyanide dihydratases. The
second pathway leads to thiol elimination and release of the amide, as happens to
some nitrilases with certain substrates, as well as with cyanide hydratases. Currently,
the factors that rule the selection of the pathways are unknown. It is suggested that
this choice depends on the charge distribution in the tetrahedral intermediate, which
can act as a mechanistic switch [25].
14.2.4
Side Activities
Early work with nitrilases already showed that the hydrolysis of some nitriles afforded
small amounts of the corresponding amides, in addition to acid – from 1% up to 6%
of total products [17, 26, 27]. This fact was long ignored and it was considered that
14.2 Nitrilases j549
Figure 14.1 Proposed catalytic mechanism for the nitrilase. Adapted from References [24, 25]
amide formation was conditioned by nitrile hydratase contamination or by anom-

alous process termination. However, it was found that the purified, recombinantly
expressed nitrilases from Arabidopsis thaliana produced amides as the major product
under certain conditions. More than 60% of amide was formed during the hydrolysis
of 3-cyano-L-alanine, catalyzed by NIT4 [28], while the other nitrilase from A. thaliana,
NIT1, preferentially converted 2-fluoroarylacetonitriles into amides with a selectivity
of more than 95% [29]. The enhanced formation of amides was caused by the
presence of highly electron-withdrawing substituents, such as fluorine or nitro
groups [30].
Kiziak et al. have shown that the nitrilase from Pseudomonas fluorescens EBC191
produced significant amounts of amides from different substituted phenylacetoni-
triles [31]. The highest proportions of amides were obtained with racemic
O-acetylmandelonitrile and the optically pure (R)-O-acetylmandelonitrile as sub-
strates (43% and 28% of the substrates converted into the amides, respectively). The
acids and amides formed from chiral nitriles demonstrated, in most cases, opposite
enantiomeric excesses. Thus, the nitrilase preferentially converted mandelonitrile
into (R)-mandelic acid and (S)-mandelic acid amide.
Recently [32], mutants of nitrilase from Pseudomonas fluorescens EBC191 with
increased amide formation (C-terminal deleted mutants and point mutants with
amino acid exchanges at the position H296) have been isolated. In addition, it was
found that the point mutation of Trp164Ala converted the (R)-specific nitrilase from
A. faecalis ATCC8750 into an enzyme that hydrolyzed (R,S)-mandelonitrile prefer-
entially to (S)-mandeloamide [33].
Thus, towards certain substrates, nitrilases behave as nitrile hydratases, producing
stoichiometric amounts of the corresponding amides. As the known nitrile hydra-
tases usually show low enantioselectivities, nitrilases provide a new synthetic route to
enantiomerically pure amides. It could be stated that the nitrilase-mediated hydro-
550
lysis of nitriles to the carboxylic acids and into the amides represent two branches of a
single nitrilase mechanism, as Hook and Robinson originally assumed [26].
14.2.5
Substrate Specificity
The classification of nitrilases into the three subgroups aromatic nitrilases, arylace-
tonitrilases, and aliphatic nitrilases [9] is of substantial practical use even though
some of the enzymes show a broader substrate spectrum that is not limited to only
one of the aforementioned groups of nitriles. Aromatic nitrilases include most of the
bacterial nitrilases and virtually all of the fungal nitrilases [34] so far characterized.
They are highly specific for benzonitriles and heterocyclic nitriles with the cyano
group directly attached to the heterocycle and they usually accept a broad range of
substituents at the ring, with some limitations for those in ortho position. Since there
is no stereocenter (or other isomeric structures) at the a- or b-position to the cyano
group enantioselective conversions have not been reported and aromatic nitrilases
are only of limited value for organic synthesis.
Arylacetonitrilases do not show any activity towards benzonitriles but aliphatic
nitriles, especially unsaturated ones, are hydrolyzed with lower activities in some
cases. Substituents at the a-position to the cyano group are usually tolerated but
decrease the reaction velocity. Nevertheless, the enantioselective hydrolysis of, for
example, mandelonitriles is an important application for arylacetonitrilases
(Section 14.2.8).
Few aliphatic nitrilases were available before nitrilases from plants were charac-
terized. The first bacterial one was isolated from Rhodococcus rhodochrous K22 [35].
Plant nitrilases generally belong to this group. The substrate spectrum is somewhat
broader than that of aromatic nitrilases and arylacetonitrilases. As with arylacetoni-
trilases, the presence of substituents at the a-position typically decrease the reaction
velocity – only a-hydroxy- and a-fluoro compounds are hydrolyzed with high specific
activities.
14.2.6
Regioselectivity/Monohydrolysis of Dinitriles
As mentioned in the previous section, nitrilases clearly differ in their substrate

specificity and therefore a specific cyano group can be targeted within a dinitrile, for
example, an aromatic nitrile can be hydrolyzed in the presence of an aliphatic one.
Additionally, in most cases nitrile groups with a non-substituted a-CH2 group
are preferentially hydrolyzed compared to branched nitriles. This feature can be
exploited to differentiate between several cyano groups, as in the case of the
pregabalin synthesis described in Section 14.2.8.
A widespread characteristic of nitrilases is the selective conversion of dinitriles
with identical nitrile groups into the cyanocarboxylic acids (Scheme 14.1). In the strict
sense, this is not a case of regioselectivity but of chemoselectivity since a dinitrile is
accepted as the substrate and the cyanocarboxylic acid is not further converted. This
type of selectivity has been demonstrated for aromatic nitrilases [36], arylacetoni-
trilases [37], and aliphatic nitrilases [3, 37–39]. The nitrilase from Arabidopsis thaliana
shows a particularly good mono selectivity. Up to a chain length of five carbon
atoms (adipinonitrile) [39] the selectivity is excellent while at a chain length of six
carbon atoms (pimelonitrile) the selectivity is acceptable.
CN Nitrilase from COOH Nitrilase from COOH

A.thaliana A. thaliana
CH 2 CH 2 CH 2
n n n
NC NC HOOC
n = 1-3 < 1% dicarboxylic acid at complete conversion
n=5 1.1% dicarboxylic acid at 83% conversion
n=6 12.5% dicarboxylic acid at 66% conversion
Scheme 14.1
The mono-hydrolysis of dinitriles has been used for the synthesis of lactams in
80–94% yield over two steps (Scheme 14.2). In the first step the ammonium salt of the
cyanocarboxylic acid is produced using whole cells of Acidovorax facilis 72W. This
intermediate is then converted directly into the corresponding lactam without
isolation by hydrogenation in aqueous solution.
H2
CN Nitrilase from COO -NH 4+ H 2, NH3 R
C
A. facilis 72W Raney-Ni n
R CH 2 R CH 2 O
n n
N
H
CN CN
R = Me, Et
n = 1, 2
Scheme 14.2
14.2.7
(E)-/(Z)-Selectivity
A broad range of nitrilases can hydrolyze a,b-unsaturated nitriles. In most cases the
pure (E)-isomer (or a nitrile without possible isomerism such as acrylonitrile) was
used as the substrate. In those cases where an isomeric mixture or the pure (E)- and
(Z)-isomer were applied, a distinct discrimination of the (E)-isomer could be
detected for the nitrilases from Arabidopsis thaliana [40], Acidovorax facilis
72W [41], and Rhodococcus rhodochrous ATCC 29484 [42]. The enzymes from
A. thaliana and R. rhodochrous were not able to hydrolyze the (Z)-isomer of
b-substituted acrylonitriles while several (E)-a,b-unsaturated nitriles without sub-
stituents at the a-position were converted with high activities (Scheme 14.3).
552
Nitrilase from
R + CN A. thaliana COOH
R +
R R
CN CN
R = CH3- E : Z = 40 : 60
R= Ph- E : Z = 60 : 40
R= CH2=CH- E : Z = 68 : 32
R= CH3-O- E : Z = 37 : 63
Scheme 14.3
In contrast, the enzyme from A. facilis did not show any selectivity with 2-
pentenenitrile but was able to hydrolyze only (E)-2-methyl-2-butenenitrile from the
isomeric mixture (Scheme 14.4).
Nitrilase from
CN A. facilis COOH
+ +
CN CN
28 : 72
Scheme 14.4
Since completely (E/Z)-selective synthesis of a,b-unsaturated nitriles or carboxylic

acids, for example, by the Wittig, Knoevenagel, or related reactions, is extremely
difficult the nitrilase catalyzed separation of isomers could be of high synthetic value
for the manufacture of (E)-acids or the (Z)-nitriles, respectively.
14.2.8
Enantioselectivity
Numerous enantioselective conversions catalyzed by nitrilases have been

reported but, in contrast to other hydrolases like esterases, resolutions reaching
>98% e.e. are known only for certain nitrilase/substrate combinations. The most
intensely investigated process is the dynamic kinetic resolution of (substituted)
benzaldehyde cyanohydrins to yield (R)-mandelic acids. Since benzaldehyde cyano-
hydrins racemize even under nearly neutral conditions at pH 7.5 in aqeous solution
through reversible loss of HCN, a theoretical yield of 100% is possible (Scheme 14.5).
O OH OH OH
Nitrilase
+ HCN +
R H R CN R CN R COOH
Scheme 14.5
The process was first described by Yamamoto et al. using the nitrilase from
Alcaligenes feacalis ATCC 8750 [43] starting at a concentration 50 mM mandelonitrile.
No (S)-mandelic acid could be detected at full conversion of the nitrile. Hauer et al.
reported the identification of a mutant of a nitrilase from Alcaligenes feacalis with a
fourfold increased activity for 2-chloromandelonitrile [44]. The synthesis was expand-
ed to several substituted mandelic acids and also to aryllactic acids by DeSantis
et al. [45]. For the substituted mandelic acids the e.e. was in the range 95–99%, while
for the lactic acid derivatives it was 91–99%.
The resolution of nitriles with a-substituents other than hydroxy is also possible.
Yamamoto et al. reached up to 95% e.e. for (S)-ibuprofen [2-(p-isobutlyphenyl)
propionic acid] with Acinetobacter sp. AK 226 whole cells [46]. a-Fluoroacetonitriles
have been resolved using the nitrilase from A. thaliana (Scheme 14.6) [29]. As
described in Section 14.2.4, the main product of the reaction is not the acid (approx.
15% of the product) but the amide (approx. 85% of the product). Surprisingly, the
enzyme is (R)-selective with respect to the amide whereas the acid is slightly enriched
in the (S)-enantiomer. The (R)-a-fluorocarboxylic acids can be obtained from the
amides in 88–92% yield with 97–99% e.e. after hydrolysis with sulfuric acid and a
single recrystallization.
F F F
Nitrilase from
CN A. thaliana COOH CONH2
+
21 - 41% conv.
R R < 5 - 14% ee (S )
R
75 - 82 ee (R )
Scheme 14.6
There are also examples of nitrilase-catalyzed resolutions of products bearing

stereocenters at the b-position of the nitrile group. The active pharmaceutical
ingredient Pregabalin can be produced by resolution of the respective dinitrile and
subsequent hydrogenation of the nitrile (Scheme 14.7) [47]. The wrong enantiomer
can be racemized with strong bases such as sodium ethanolate. In this way the
original chemical route can be shortened by two steps. Nitrilase1 from A. thaliana has
been shown to be the best enzyme among the ten nitrilases tested and, furthermore,
the specific activity has been improved by the factor of 2.7 by directed evolution. In
addition to the enzymes enantioselectivity, the regioselectivity and also the selectivity
regarding the mono-hydrolysis are exploited in this synthesis.
Another synthetic application of a monohydrolysis of a dinitrile bearing a stereo-
center in the b-position utilizes the meso trick to overcome the limitation of 50% yield
of a racemic resolution: approx 385 g l1 of the statin side chain intermediate shown
in Scheme 14.8 can be produced in 16 h in 81% yield and 99% e.e. using 60 g l1 of
nitrilase BD9570 [48].
554
CHO CO 2Et 1. Piperidine CN

2. KCN
+ rac.
CN
CN
Nitrilase from NaOEt
A. thaliana
NH2 H 2 / Ni CN CN
+
COOH
Pregabalin COOH CN
overall 40% yield (w/o recycle) 45% conversion, 98% ee
Scheme 14.7
O OH OH
1. HCN Nitrilase
Cl NC CN NC COOH
2. NaCN
16 h, complete conv.
Scheme 14.8
14.3
Nitrilase-Containing Biocatalysts for Hydrolysis of Nitriles to Acids
14.3.1
Whole Cell Biocatalysts
The application of wild-type bacterial cells is the most straightforward approach to

obtaining a nitrilase biocatalyst but there are several drawbacks. First of all a
fermentation protocol has to be established that combines a fast growth with a
strong induction of the nitrilase. In many cases the screening substrate is used as the
inducer, which can be a significant cost factor, and culture media have to be adapted to
keep up the induction. For some bacteria, for example, Rhodococci and Pseudomonads,
this is possible but the efficiency of a recombinant expression system based on E. coli
can normally not be reached. For example, DiCosimo et al. reached a threefold higher
expression level for the nitrilase from Acidovorax facilis 72W expressed in E. coli
compared to the wild-type strain at a significantly higher biomass concentration [49].
The occurrence of additional nitrilases and/or nitrile hydratases/amidases, which is
common for many bacterial strains, can be a severe problem for reaction selectivity
and the development of selective induction methods or the construction of knock-out
mutants is laborious. Fungal nitrilases usually do not show a sufficient enzyme
14.3 Nitrilase-Containing Biocatalysts for Hydrolysis of Nitriles to Acids j555
production and a heterologous expression is required [34]. The same is apparently the
case for nitrilases from plants.
Generally, the application of both whole cells (wild type or recombinant) and isolated
enzymes is possible. The performance of the A. facilis 72W as well as recombinant
E. coli whole cell catalyst is excellent, showing that diffusion of substrate and product
across the cell wall is not a big issue. Whole cells can also be used in immobilized form
just like the isolated enzymes [49, 50]. Kaul et al. reported a dramatically increased
stability for whole cells immobilized in alginate or polyacrylamide [50].
14.3.2
Enzyme Preparations
The use of purified enzymes or of the corresponding enzyme preparations as

biocatalysts, in contrast to the whole cells, is still limited. The major reason for this
is the commercial unavailability of nitrilases at bulk scale due to their rather low
stability in combination with their relatively high price.
Purified nitrilases are quite labile and require a reducing environment for their
activity. But even in the presence of a reducing agent (e.g., 1 mM DTT) they rapidly
lose activity when stored in phosphate buffer (0.01 M, pH 7.5) at 4 C. Under these
conditions the half-life period for most nitrilases ranges from several hours up to
several days [31, 51, 52]. In addition, only the nitrilases from R. rhodochrous J1 [53],
R. rhodochrous K22 [35], and A. faecalis JM3 [54] remained stable for over two months.
However, such compounds as glycerol [51, 54], D-sorbitol [55], sucrose [56], NaCl, and
ammonium sulfate [31] remarkably enhance the stability of nitrilases under long-
term storage at low temperature. In the presence of 20–50% of glycerol, the purified
enzymes can be stored at 20 C from several months up to a year without significant
loss of activity [52, 53, 57].
One reason for the lack of nitrilase stability is their oxygen sensitivity, which, in
turn, is presumed to be related to the autoxidation of the cysteine residue, which is
part of the catalytic triad. Mateo et al. have shown that the replacement of oxygen with
an argon atmosphere results in a substantial stabilization of isolated enzymes [57].
The most suitable working environment for nitrilases appeared to be an aqueous
media with a neutral pH range (pH 6–8.5), and moderate temperature (20–37 C) –
although most of nitrilases from mesophilic microorganisms reveal their maximum
activity at 45–50 C, they rapidly lose the activity at such temperatures. For instance,
the nitrilase from P. putida was quite labile at high temperatures (its half-life values
for 40 and 50 C were 76 and 9 min, respectively). At the same time, at 30 C the
enzyme retains 87% of its initial activity even after the 6-h incubation in phosphate
buffer (0.01 M, pH 7.5) [51]. Nitrilases from thermophilic microorganisms have been
isolated and have demonstrated high thermostability. The nitrilase from termophilic
bacteria Bacillus pallidus Dac521 has a half-life of 8.5 h at 50 C [58]. The nitrilase from
the hyperthermophilic archaeon Pyrococcus abyssi with half-lives of 25 h at 70 C, 9 h
at 80 C, and 6 h at 90 C represents the most thermostable enzyme [14].
To enhance the bioavailability of the hydrophobic substrates, organic solvents are
typically added to reaction media. Most nitrilases are able to work in the presence of
556
low (<5%) concentrations of organic solvent and their activity increases with the
availability of substrates. However, in the presence of higher (>10%) concentrations
of organic solvent the activity, as a rule, decreases considerably due to enzyme
denaturation. In recent years, nonetheless, a few enzymes have been discovered that
can work in the presence of higher concentrations of water-soluble and water-
immiscible organic solvents. Heinemann et al. have shown that the nitrilase from
Synechocystis sp. PCC6803 shows preference for hydrophobic substrates and demon-
strates high tolerance against organic solvents. Hydrolysis rates for dodecanoic acid
nitrile and naphthalenecarbonitrile increased in the presence of 20% methanol or
40% n-hexadecane [59]. The nitrilase from Pseudomonas sp. DSM 11387 exhibits
pronounced activity in the presence of hydrocarbons (66% in 75% n-octane, 97% in
95% n-hexadecane, and 25–58% in buffer-saturated primary alcohols) [60]. Nitrilase
from Fusarium solani O1 is also suitable for the use in selected organo-aqueous
media. More than a half of its initial activity was retained in the presence of 5–50% of
n-hexane or n-heptane or 5–15% of xylene or ethanol [56].
Enzyme immobilization is one approach to increasing the stability and to improv-
ing enzyme performance. One successful application is the purified nitrilase from
ATC8750 immobilized on alumina. This immobilized catalyst was used for the large-
scale production of a hydroxy-analog of methionine, with an excellent turnover [61].
The most encouraging results have been obtained via crosslinking of enzyme
aggregates (CLEAs) with a bifunctional crosslinking agent, most typically with
glutaraldehyde. Sheldon et al. have elaborated a new approach to obtaining stable
and effective catalysts on the basis of nitrilases that involves simple precipitation of
the enzyme from the aqueous solution, using standard techniques, and crosslinking
of the resulting physical aggregates of enzyme molecules [62]. CLEAs derived from
nitrilases appear more stable than intact enzymes, making possible their recyclable
use. The CLEAs can be composed of two or more enzymes (combi-CLEA), which
favors catalytic cascade processes. Combi-CLEA containing the (S)-selective oxyni-
trilase and a non-selective nitrilase were utilized for the one-pot conversion of
benzaldehyde into (S)-mandelic acid with a high yield and enantioselectivity [63].
In recent years more and more nitrilases have become commercially available on
account of the mature recombination technique. Today, a whole range of nitrilases
can be readily purchased from several manufactures at least in laboratory-scale
quantities.
14.4
Summary and Outlook
Nitrilases have been shown to be highly versatile biocatalysts for the production of
chiral pharmaceutical intermediates as well as bulk products of several thousands
tons per year. The advantage of the general characteristic of enzymes to convert their
substrates very selectively without the formation of considerable amounts of
by-products is even more significant for the hydrolysis of nitriles since their chemical
hydrolysis needs even harsher conditions than ester hydrolysis. Most notably,
References j557
though, nitrilases show unique regio- (E/Z)- and enantioselectivities that can also be
exploited in combination to facilitate the construction of complex molecules from
basic compounds.
A large diversity of nitrilases can be found in bacteria, archaea, fungi, and plants,
giving access to biocatalysts with a broad scale of specificities and selectivitities. Their
usually straightforward recombinant expression in several hosts allows large-scale
production and also improvement of specific enzyme characteristics by directed
evolution.
Based on these facts, it is most likely that nitrilases will play an even more
important role in organic synthesis in the near future.
References
1 Thimann, K.V. and Mahadevan, S. (1964) Burk, M.J., and Short, J.M. (2004)
Arch. Biochem. Biophys., 105, 133–141. Appl. Environ. Microbiol., 70
2 Mylerova, V. and Martinkova, L. (2003) 2429–2436.
Curr. Org. Chem., 7, 1–17. 13 Seffernick, J.L., Samanta, S.K., Louie,
3 Chen, J., Zheng, R.-C., Zheng, Y.-G., and T.M., Wackett, L.P., and Subramanian, M.
Shen, Y.-C. (2009) Adv. Biochem. Eng./ (2009) J. Biotechnol., 143, 17–26.
Biotechnol., 113, 33–77. 14 Mueller, P., Egorova, K., Vorgias, C.E.,
4 Bork, P. and Koonin, E.V. (1994) Protein Boutou, E., Trauthwein, H., Verseck, S.,
Sci., 3, 1344–1346. and Antranikian, G. (2006)
5 Pace, H.C. and Brenner, C. (2001) Genome Cloning, Protein Expression Purific.,
Biol., 2, 1–9. 47, 672–681.
6 OReilly, C. and Turner, P.D. (2003) 15 Janowitz, T., Trompetter, I., and
J. Appl. Microbiol., 95, 1161–1174. Piotrowski, M. (2009) Phytochemistry,
7 Nakai, T., Hasegawa, T., Yamashita, E., 70, 1680–1686.
Yamamoto, M., Kumasaka, T., Ueki, T., 16 Harper, D.B. (1977) Biochem. J., 165,
Nanba, H., Ikenaka, Y., Takahashi, S., 309–319.
Sato, M., and Tsukihara, T. (2000) 17 Stevenson, D.E., Feng, R., Dumas, F.,
Structure, 8, 729–737. Groleau, D., Mihoc, A., and Storer, A.C.
8 Wang, W.-C., Hsu, W.H., Chien, F.T., and (1992) Biotechnol. Appl. Biochem., 15,
Chen, C.-Y. (2001) J. Mol. Biol., 306, 283–302.
251–261. 18 Hoyle, A.J., Bunch, A.W., and Knowles,
9 Kobayashi, M. and Shimizu, Y. (1994) C.J. (1998) Enzyme Microb. Technol., 23,
FEMS Microbiol. Lett., 120, 217–224. 475–482.
10 Bunch, A.W. (1998) Nitriles, in 19 Nagasawa, T., Wieser, M., Nakamura, T.,
Biotechnology: Biotransformations I vol. 8a Iwahara, H., Yoshida, T., and Gekko, K.
(eds H.J. Rehm and G. Reed), Wiley-VCH (2000) Eur. J. Biochem., 267, 138–144.
Verlag GmbH, Weinheim, ch. 6, 20 Yamamoto, K., Fujimatsu, I., and
pp. 277–324. Komatsu, K.-I. (1992) J. Ferment. Bioeng.,
11 Thuku, F.R.N., Brady, D., Benedik, M.J., 73, 425–430.
and Sewell, B.T. (2009) J. Appl. Microbiol., 21 Sewell, B.T., Berman, M.N., Meyers, P.R.,
106, 703–727. Jandhyala, D., and Benedik, M.J. (2003)
12 Robertson, D.E., Chaplin, J.A., DeSantis, Structure, 11, 1–20.
G., Podar, M., Madden, M., Chi, E., 22 Thuku, R.N., Weber, B.W., Varsani, A., and
Richardson, T., Milan, A., Miller, M., Sewell, B.T. (2007) FEBS J., 274,
Weiner, D.P., Wong, K., McQuaid, J., 2099–2108.
Farwell, B., Preston, L.A., Tan, X., Snead, 23 Jandhyala, D.M., Berman, M.,
M.A., Keller, M., Mathur, E., Kretz, P.L., Meyers, P.R., Sewell, B.T., Willson, R.C.,
558
and Benedik, M.J. (2003) Appl. Environ. 42 Stevenson, D., Feng, R., Dumas, F.,
Microbiol., 69, 4794–4805. Groleau, D., Mihoc, A., and Storer, A.
24 Kobayashi, M., Goda, M., and Shimizu, S. (1992) Biotechnol. Appl. Biochem., 15,
(1998) Biochem. Biophys. Res. Commun., 283–302.
253, 662–666. 43 Yamamoto, K., Fujimatsu, I., and
25 Fernandes, B.C.M., Mateo, C., Kiziak, C., Komatsu, K. (1992) J. Ferment Bioeng.,
Chmura, A., Wacker, W., Rantwijk, F., 73, 425–430.
Stolz, A., and Sheldon, R.A. (2006) 44 Zelinski, T., Kesseler, M., Hauer, B.,
Adv. Synth. Catal., 348, 2597–2603. and Friedrich, T. (2004) WO 2004076655.
26 Hook, R.H. and Robinson, W.G. (1964) 45 DeSantis, G., Zhu, Z., Greenberg, W.,
J. Biol. Chem., 239, 4263–4267. Wong, K., Chaplin, J., Hanson, S.,
27 Goldlust, A. and Bohak, Z. (1989) Farwell, B., Nicholson, L., Rand, C.,
Biotechnol. Appl. Biochem., 11, Weiner, D., Robertson, D., and Burk, M.
581–601. (2002) J. Am. Chem. Soc., 124,
28 Piotrowski, M., Sch€onfelder, S., and 9024–9025.
Weiler, E.W. (2001) J. Biol. Chem., 276, 46 Yamamoto, K. and Komatsu, K. (1991)
2616–2621. Agric. Biol. Chem., 55, 1459–1466.
29 Effenberger, F. and Osswald, S. (2001) 47 Xie, Z., Feng, J., Garcia, E., Bernett, M.,
Tetrahedron: Asymmetry, 12, 279–285. Yazbeck, D., and Tao, J. (2006) J. Mol.
30 Osswald, S., Wajant, H., and Cat. B, 41, 75–80.
Effenberger, F. (2002) Eur. J. Biochem., 48 Bergeron, S., Chaplin, D., Edwards, J.,
269, 680–687. Ellis, B., Hill, C., Holt-Tiffin, K., Knight, J.,
31 Kiziak, C., Conradt, D., Stolz, A., Mahoney, T., Osborne, A., and
Mattes, R., and Klein, J. (2006) Ruecroft, G. (2006) Org. Proc. Res. Dev.,
Microbiology, 151, 3639–3648. 10, 661–665.
32 Kiziak, C., Klein, J., and Stolz, A. (2007) 49 Chauhan, S., Wu, S., Wu, S.,
Protein Eng. Des. Select., 20, 385–396. Blumerman, S., Fallon, R., Gavagan, J.,
33 Kiziak, C. and Stolz, A. (2009) Appl. DiCosimo, R., and Payne, M. (2003)
Environ. Microbiol., 75, 5592–5599. Appl. Microbiol. Botechnol., 61,
34 Martinkova, L., Vejvoda, V., Kaplan, O., 118–122.
Kubac, D., Malandra, A., Cantarella, M., 50 Kaul, P., Banerjee, A., and Banerjee, U.
Bezouska, K., and Kren, V. (2009) (2006) Biomacromolecules, 7, 1536–1541.
Biotechnol. Adv., 27, 661–670. 51 Banerjee, A., Kaul, P., and Banerjee, U.
35 Kobayashi, M., Yanaka, N., Nagasawa, T., (2006) Arch. Microbiol., 184, 407–418.
and Yamada, H. (1990) J. Bacteriol., 172, 52 Harper, D. (1985) Int. J. Biochem., 17,
4807–4815. 677–683.
36 Kobayashi, M., Yanaka, N., Nagasawa, T., 53 Kobayashi, M., Nagasawa, T., and
and Yamada, H. (1988) Appl. Microbiol. Yamada, H. (1989) Eur. J. Biochem.,
Biotechnol., 29, 231–235. 182, 349–356.
37 Bengis-Garber, C. and Gutman, A. (1989) 54 Nagasawa, T., Mauger, J., and Yamada, H.
Appl. Microbiol. Biotechnol., 32, 11–16. (1990) Eur. J. Biochem., 194, 765–772.
38 Gavagan, J., Fager, S., Fallon, R., 55 Kaplan, O., Vejvoda, V., Plıhal, O.,
Folsom, P., Herkes, F., Eisenberg, A., Pompach, P., Kavan, D., Bojarova, P.,
Hann, E., and DiCosimo, R. (1998) Bezouska, K., Mackova, M., Cantarella, M.,
J. Org. Chem., 63, 4792–4801. Jirku , V., Kren, V., and Martınkova, L.
39 Effenberger, F. and Osswald, S. (2001) (2006) Appl. Microbiol. Biotechnol., 73,
Synthesis, 12, 1866–1872. 567–575.
40 Effenberger, F. and Osswald, S. (2001) 56 Vejvoda, V., Kaplan, O., Bezouska, K.,
Tetrahedron: Asymmetry, 12, 2581–2587. Pompach, P., Sulcab, M., Cantarella, M.,
41 Hann, E., Sigmund, A., Fager, S., Benada, O., Uhnakova, Br., Rinagelova, A.,
Cooling, F., Gavagan, J., Bramucci, M., Lutz-Wahl, S., Fischer, L., Kren, V., and
Chauhan, S., Payne, M., and DiCosimo, R. Martinkova, L. (2008) J. Mol. Cat. B:
(2004) Tetrahedron, 60, 577–581. Enzym., 50, 99–106.
References j559
57 Mateo, C., Fernandes, B., van Rantwijk, F., 61 Rey, P., Rossi, J.-Ch., Taillades, J., Gros, G.,
Stolz, A., and Sheldon, R.A. (2006) and Nore, O. (2001) Aventis Animal
J. Mol. Cat. B: Enzym., 38, 154–157. Nutrition, U.S. Patent 6180359.
58 Almatawah, Q.A., Cramp, R., and 62 Mateo, C., Palomo, J.M., van Langen, L.M.,
Cowan, D.A. (1999) Extremophiles, Van Rantwijk, F., and Sheldon, R.A. (2004)
3, 283–291. Biotechnol. Bioeng., 86, 273–276.
59 Heinemann, U., Engels, D., B€ urger, S., 63 Sheldon, R.A. (2007) Biochem. Soc. Trans.,
Kiziak, C., Mattes, R., and Stolz, A. 35, 1583–1587.
(2003) Appl. Environ. Microbiol., 69, 64 Zhu, D., Mukherjee, C., Yang, Y., Rios, B.,
4359–4366. Gallagher, D., Smith, N., Biehl, E.,
60 Layh, N. and Willetts, A. (1998) Biotechnol. and Hua, L. (2008) J. Biotechnol., 133,
Lett., 20, 329–331. 327–333.
j561
15
Hydrolysis of Amides
Theo Sonke and Bernard Kaptein
15.1
Introduction
Organic carboxylic acid amides and carboxylic acids find widespread use in industry
with applications in the production of commodity chemicals, pharmaceuticals,
agrochemicals, and compounds used in the food and feed industry. Although
especially in the commodities industry most of these compounds are still produced
chemically, biocatalytic production processes have gained sharply in importance in
the last 10–15 years. Besides environmental considerations, the ever increasing
demand by the pharmaceutical and agrochemical industries for enantiomerically
pure building blocks in combination with strongly improved methodologies to
design tailor-made biocatalysts has driven this interest in biocatalytic processes.
Because of the regio-, chemo-, and enantioselectivity of enzymes, biocatalytic
processes are the ideal method to produce enantiopure compounds [1–4], including
amides and acids.
This chapter gives a detailed overview of the potential of microorganisms and
enzymes to catalyze the regio- and enantioselective hydrolysis of carboxylic acid
amides. Especially, the properties of amide hydrolyzing enzymes and their use for the
resolution of chiral carboxylic acid amides are discussed, with emphasis on the
resolution of amino acid amides, hydroxy acid amides, and azido acid amides.
Moreover, the properties of peptide amidases that specifically hydrolyze the
C-terminal amide bonds in peptides and their potential role in chemoenzymatic
peptide synthesis are described.
15.2
Enantioselective Hydrolysis of Carboxylic Acid Amides
Amidases (amidohydrolases, EC 3.5.1.4) are a class of enzymes that hydrolyze amides

into carboxylic acids and amines and are especially suited for the preparation of chiral
carboxylic acids [5–8]. Although acylases, penicillin acylase, amino acid acylase,
peptide deformylase, peptidases, proteases (such as a-chymotrypsin, trypsin,
and Oliver May.
j 15 Hydrolysis of Amides
562
subtilisin and thermolysin) as well as certain lipases also hydrolyze amide and peptide
bonds [9–15], strictly speaking the family of amidases only hydrolyze primary
carboxylic acid amides 1 into carboxylic acids 2 and ammonia (Scheme 15.1). Many
amidases are active for a broad range of substrates, varying from aliphatic substrates,
such as butyramide (3) and acrylamide (4), aromatic substrates, such as benzamide (5)
and nicotinamide (6), and chiral substrates (see below). Amidases varying in subunit
size from 37–66 kDa are reported and exist as monomeric, as well as dimeric, trimeric,
tetrameric, hexameric, or octameric, enzymes. Most amidases can be divided into two
subgroups: small amidases with subunits of approximately 38 kDa and large amidase
with subunits of approximately 55 kDa [8, 16]. Short-chain aliphatic amides are the
most common substrates for small amidases, while hydrolysis of mid-chain aliphatic
and aromatic amides is frequently reported for large amidases. In addition, these
amidases also hydrolyze stereoselectively various chiral and prochiral amide
substrates as well as cyclic and acyclic a-amino acid amides (see below).
R NH2 R O-
+ H2 O + NH 4+
O O
1 2
3 R = -(CH2 )2 CH 3 5R= 6R= N

4 R = -CH=CH 2
Scheme 15.1 Basic reaction catalyzed by amidases (EC 3.5.1.4) and a few typical aliphatic and
aromatic substrates.
In nature the large amidases frequently appear in combination with nitrile

hydratases [17, 18]. In this way CN groups can be hydrolyzed to ammonia as nitrogen
source for growth. Whereas the nitrile hydratases in general are not stereospecific
most amidases are (see below).
For various large amidases the catalytic role of a cysteine residue has been proven.
These amidases belong to the nitrilase superfamily (EC 3.5.5.1) that have a conserved
cysteine, glutamate, and lysine residue in the catalytic site for the direct hydrolysis of a
nitrile group to the corresponding carboxylate [19–22]. The free thiol group of the
cysteine amidases is assumed to act as a nucleophile, resulting in a covalent
enzyme–acyl complex. For example, the enantioselective and thermostable amidase
from Pseudonocardia thermophila belongs to this class of enzymes [23]. The purified
enzyme (108 kDa) consists of two identical subunits and can hydrolyze a wide range
of aliphatic amides, aromatic amides, and amino acid amides. Remarkably, the
amidase is fully (S)-selective for 2-phenylpropionamide (7) [E-ratio > 100], but does
not show any stereoselectivity for the amino acid amides tested.
Many of the small amidases belong to a widespread group of so-called
amidase signature (AS) enzymes, found all over nature in prokaryotic and eukaryotic
cells [6, 24, 25]. This family of enzymes shows a highly conserved sequence of about
130 amino acids known as the AS sequence (Pfam family: PF01425). The enzymes
15.2 Enantioselective Hydrolysis of Carboxylic Acid Amides j563
are furthermore characterized by a unique Ser-cisSer-Lys catalytic triad, or occasion-
ally a Cys-cisSer-Lys triad [26], with the serine hydroxyl or cysteine sulfhydryl group
functioning as the catalytic nucleophile in the hydrolysis reaction [27]. Although the
enzymes belonging to this family all catalyze the hydrolysis of an amide bond
(CONH), their biochemical roles and substrate specificities vary widely. As an
example, the recently characterized amidase from the nitrile-degrading Rhodococcus
species N-771 by Ohtaki et al. belongs to the amidase signature family [28]. From the
crystal structure of this dimeric amidase (107 kDa) three domains can be identified: a
unique N-terminal a-helical domain involved in the formation of the dimeric
structure, and a large and small domain. The large domain shows high structural
similarity to other amidase signature family enzymes and contains the catalytic Ser-
cisSer-Lys triad, while the small domain together with the dimeric structure forms a
narrow substrate-binding tunnel for aliphatic and aromatic amides. In addition, for
some of the signature amidases a catalytic role of an aspartate residue has also been
identified, although it is not clear if this residue acts directly as the nucleophile.
Metallo amidases with an active site metal ion, as seen in nitrile hydratase [17] or
proteases like a-chymotrypsin and thermolysin [11], are not described. Although
Nawaz et al. reported that the amidases from a Rhodococcus species and from Klebsiella
pneumoniae NCTR1 are iron and cobalt depending enzymes, it is unlikely that these
metal ions play a catalytic role [29, 30].
Note that of the different amidases described in Table 15.1 the amidase from
Brevibacterium sp. R312 is identical to the amidase from Rhodococcus sp. strain N-774
and strain N-771 [28, 45] and also that both microorganisms are closely related or
similar as described by Mayaux et al. [25]. In addition, the amidase from Pseudomonas
chlororaphis B23 shows a significant similarity to these amidases [77].
Various other amidases described in Table 15.1 have been cloned and overexpressed
in Escherichia coli or Brevibacterium lactofermentum strains [16, 25, 65, 90, 98, 106].
In addition to standard amidase catalyzed hydrolysis, aliphatic amides can be
converted with hydroxylamine or hydrazine as a nucleophile, following a ping pong
bi-bi mechanistic mode [8, 37, 50, 59, 69, 102, 110]. The stereoselective conversion of
(RS)-2-phenylpropionamide (7) into the (S)-hydroxamic acid (8) with hydroxylamine
is also described (Scheme 15.2) [7, 91].
CH3 CH 3 NH2 -OH CH 3

NH 2 Amidase S enzyme NHOH
O O O
(S)-7 NH3 E-S complex (S)-hydroxamic acid
+ (8)
CH3
NH 2
O
(R)-7
Scheme 15.2 Amidase transfer reaction with hydroxylamine as nucleophile.

564
Table 15.1 Overview of different amidases with some of their specifications.
Microorganism Mass (kDa) Subunits, # (kDa) Optimum Amide specificity Reference
# (kDa) pH T ( C)
Achromobacter xylosoxidans — – (50) 7.0–8.5 37–48 Aliphatic (N-aryl) [31]

Agrobacterium tumefaciens d3 490 8 (63) 7.4 37 Chiral (S) [32, 33]
Arthrobacter sp. J1 320 8 (39) 7.0–9.0 30–45 Aliphatic [34]
Arthrobacter sp. RC100 100 2 (51) 9.0 50 Broad spectrum, [35, 36]
carbamates
Aspergillus nidulans Various — 7.2 37 Broad spectrum [37–40]
amidases
Bacillus stearothermophilus BR388 — – (40) 7.0 55 Aliphatic [41]
Bacillus sp. BR449 — – (37) 7.6 50 Acrylamide [42]
Blastobacter sp. A17p-4 48 1 7.0 30 Aliphatic, aromatic [43]
Bradyrhizobium japonicum Malonamide [27]
Brevibacterium sp. R312 120 2 (49) 7.0 — Chiral (S) [44–51]
Burkholderia sp. DSM 9925 — — 8.0 40–47 Piperazine carboxamide, [52]
(R)-selective
Comamonas acidovorans A:18 DSM 6351 7.0 37 Cyclopropyl-carboxa- [53, 54]
mide, (R)-selective
Comamonas acidovorans KPO-2771-4 55 1 8.5–10.0 35 Broad spectrum, (R)- [55, 56]
ketoprofen
Corynebacterium sp. C5 — — 6.5–10.0 50 Aliphatic [57]
Corynebacterium sp. N-771 — — 7.0 25 Chiral (S) [47]
Delftia tsuruhatensis ZJB-05174 — — — 30–56 Cyclopropyl-carboxa- [58]
mide; T dependent
stereoselectivity
Geobacillus pallidus RAPc8 228 6 (38) 7.0 50 Aliphatic [59]
Helicobacter pylori — – (38) 7.4 20 Aliphatic [60]
Klebsiella oxytoca PRS1 — – (37) 9–10 70 2-CF3-lactamide, (R)- [61–63]
selective
Klebsiella pneumonia NCTR 1 64 1 5.0–8.5 30–65 Aliphatic [29]
Klebsiella terrigena DSM 9174 — — 8.0 40–47 Piperazine carboxamide, [52]
chiral (S)
Mycobacterium smegmatis NCTC 8159 7.0–9.0 37 Aliphatic [64]
Mycobacterium sp. AJ115 — (55) 7.2 50 Broad spectrum, chiral [65]
(S)
Nocardia farcinica 190 4 8–11 50 Aliphatic [66]
Pseudocardia thermophila DSM 43832 108 2 (52) 5.0–10.0 70 Broad spectrum, chiral [23]
(S)
Pseudomonas aeruginosa 8602 — — 7.2 37 Aliphatic [67–70]
Pseudomonas aeruginosa þ mutants 220 6 (38) 7.0–7.5 55 Aliphatic [37, 71–74]
Pseudomonas azotoformans IAM1603 (AAA) 34.5 1 6.0–9.5 45 (S)-Piperazine [75]
carboxamide
Pseudomonas chlororaphis B23 105 2 (54) 5.9–9.9 25–50 Aliphatic, chiral (S) [76, 77]
Pseudomonas fluorescens DSM 9924 — — 7.0 30 (S)-Piperazine [52, 63]
carboxamide
Pseudomonas thermophila 147 3 7.0 70 Broad spectrum, chiral [78]
(S)
Pseudomonas sp. B21C9 — — 7.7 30 Chiral (S) [79]
Pseudomonas sp. GDI 211 — — 5.5–8.5 45 Aromatic [80]
Pseudomonas sp. MCI3434 36 1 (30) 8.0 45 (R)-Piperazine [81]
carboxamides
150 4 (27 (a) and 9.0–10.0 60 b-Amino acid amides [82]
13 (b))
Pseudomonas putida NRRL-18668 — — 7.1 28 Chiral (S) [83]
(Continued )
15.2 Enantioselective Hydrolysis of Carboxylic Acid Amides
j565
566
Microorganism Mass (kDa) Subunits, # (kDa) Optimum Amide specificity Reference
# (kDa) pH T ( C)
Rat liver 126 2 (63) Fatty acid amides [84, 85]

Rhodococcus butanica — — 7.2 30 Chiral (S) [86]
Rhodococcus equi A4 — — 7.0 30 Lysergamide [7, 87]

Rhodococcus equi TG328 — — 7.0 60–65 Chiral (S) [88]
Rhodococcus erythropolis MP50 480 8 (61) 6.0–9.0 40–60 Aromatic, chiral (S) [89–93]
Rhodococcus erythropolis No. 7 — (55) 8.5 55 Broad spectrum, chiral [16]
(S)
Rhodococcus globerulus A-4 — (55?) — — Aliphatic [94]
Rhodococcus rhodochrous ATCC 21197 — — 6.0 30 Malonamide, chiral (R) [95]
Rhodococcus rhodochrous IFO15564 — — — — Chiral (S) [96]
Rhodococcus rhodochrous J1 110 2 (55) 7.5 30 Broad spectrum, chiral [97, 98]
(S)
Rhodococcus rhodochrous M8 150 4 (42) 5.0–8.0 55–60 Broad spectrum [99]
Rhodococcus sp. N-771 107 2 6.0–8.0 55 Aliphatic, aromatic [28]
Rhodococcus sp. N-774 — (55) — — Aliphatic [100, 101]
Rhodococcus sp. NCTR4 360 8 (45) 8.5 40 Aliphatic [30]
Rhodococcus sp. R312 — — 7.0–8.0 30 Broad spectrum [24, 102]
Rhodococcus sp. AJ270 — — 7.0 30 Broad spectrum, chiral [103, 104]
(S)
Rhodococcus sp. 118 2 (49) — — Chiral (S) [25]
Sulfolobus solfataricus MT4 56 1 6.5–8.5 60–95 Broad spectrum, chiral [26, 105, 106]
(S), (R)-lactamide, chiral
c-lactam
Microorganism DSM 6320 125 2(66) 7.0–9.5 45 (R)-Carnitine amide [107–109]
The amidase CahA from the Gram-positive bacterium Arthrobacter sp. RC100 is
also able to hydrolyze aryl N-methylcarbamates, such as the insecticides Carbaryl (9),
Xylylcarb (10), and Metolcarb (11), and therefore are also classified as carbaryl
hydrolases (Scheme 15.3) [35, 36]. N-Methylcarbamate hydrolyzing enzymes have
also been isolated from the Gram-negative bacteria Achromobacter sp. WM111 [111,
112], Blastobacter sp. M501 [113], Pseudomonas aeruginosa 50 581 [114], and Pseudo-
monas strain CRL-OK [115].
O
CH3 OH
O N Amidase
H + +
CO2 H 3C NH 2
Carbaryl (9)
O O
CH 3 CH 3
O N O N
H H
Me Me
Me
Xylylcarb (10) Metolcarb (11)
Scheme 15.3 Amidase catalyzed hydrolysis of carbamates.
Stereospecific hydrolysis is often tested with 2-phenylpropionamide (7) or 40 -

substituted 2-phenylpropionamides (12) as model substrates (Scheme 15.4). Most
small and large amidases that are active for this class of substrates are (S)-selective
[23, 25, 32, 33, 45, 47–49, 65, 76, 88–90, 93, 96, 98, 103, 104, 106, 116] (see also
Table 15.1 and review by Weiser and Nagasawa [5], and Martınkova and Kren [7]).
These substrates mimic a class of NSAID drugs of which the amides of ibuprofen
(13) [83, 86, 93, 104], naproxen (14) [33, 76, 83, 89], and ketoprofen (15) [33, 47–49, 89]
have also been described as substrates. Remarkably, the amidase from Comamonas
acidovorans KPO2771-4 shows (R)-selectivity with ketoprofen amide (15) as sub-
strate [55, 56], while all other amidases are (S)-selective for the proven substrates.
Other chiral substrates used in amidase resolutions are 2-(4-chlorophenyl)pro-
pionamide (12b) [86], 2-(4-methoxyphenyl)propionamide (12c) [86, 96], 2-(4-chlor-
ophenyl)-3-methylbutyramide (16) [76, 79, 83], 2-phenyl-4-pentenoic acid amide (17c
R ¼ allyl) [116], 4-cyano-2-phenylbutyramide (18) [116], 2-(4-hydroxyphenoxy)propio-
CH 3 CH3 CH 3
NH 2 Amidase OH NH2
+
O O O
r ac-7 (S)-acid (R)-7
Scheme 15.4 (S)-Stereoselective amidase resolution of rac-2-phenylpropionamide (7).

568
namide (19) [25, 45, 48, 49] [(R)-selective hydrolysis; absolute selectivity in line with
the (S)-selectivity for 2-arylpropionamides], atrolactic acid amide (20) [93], lyserga-
mide (21) [7, 87], 2-methyl-3-phenylpropionamide (22) [32, 90], 2-hydroxy-4-phenyl-
butyramide (23 R ¼ H) and 2-hydroxy-2-methyl-4-phenylbutyramide (24 R ¼
Me)) [93], 2-chloro-phenylacetic acid amide (25) [32, 90], mandelic acid amide
(26) [90, 93], O-methyl-mandelic acid amide (27) [32, 90], lactic acid amide
(28) [59], 4-chloro-3-hydroxybutyramide (29) [16], carnitine amide (30) [107–109],
3-benzoyloxypentanoic acid amide (31) [96], 2-benzyl-2-methylmalondiamide
(32) [96], and 2-butyl-2-methylmalondiamide (33) [95]. The hydrolysis of a- and
b-amino acid amides, cyclic amides (lactams), and related compounds are not
considered in this section, but will be extensively discussed further on. The molecular
structures of the substrates described above as well as other compounds used in
enantioselective amidase hydrolysis are shown here.
CH 3
CH3 O CH3
NH2
NH2 NH2
O
O H 3CO O
ibuprofen amide naproxen amide ketoprofen amide
(13) (14) (15)
CH3 R HO
CH 3
NH 2 NH 2 NH 2
NH 2
O
O O O
X Cl O
4'-substituted 2-substituted 2-(4-chlorophenyl)- 2-(4-hydroxyphenoxy)-
2-phenylpropionamides phenylacetamides 3-methylbutyramide propionamide
(X = H, Me, Cl, OMe) (R = Et, Pr, allyl) (16) (19)
(7,12a-c) (17a-c)
H 3 C OH O R OH
NH2 NH2
NH 2 O NH 2
O CH 3 O
2-methyl-3-phenyl 4-phenylbutyramides
atrolactamide
propionamide (R = H, Me) N
(20) (22) (23,24) CH 3
H
Cl OH OMe
HN
NH2 NH2 NH 2
lysergamide
O O O (21)
2-chloro-phenylacetamide mandelamide O-methyl-mandelamide
(25) (26) (27)
O
OH OH O
NH2 Cl OH O Ph O O
NH 2 +
Me3 N
O 4-chloro-3-hydroxy- NH 2 NH2
butyramide 3-(benzoyloxy)-
lactamide carnitine amide
(29) pentanamide
(28) (30) (31)
Ph CH 3 CH3
H2 N NH 2 H 2N NH 2
O O O O
2-benzyl-2-methyl- 2-butyl-2-methyl-
malondiamide malondiamide
(32) (33)
H H H
N N N
H
NH 2 N NH 2 NH 2 NH 2
N N N
H H H
O O O O O
piperazine- N-t er t-butyl piperidine- nipecotic acid 2,2-dimethylcyclo-
carboxamide piperazinecarboxamide carboxamide carboxamide propylcarboxamide
(34) (35) (36) (37) (38)
Despite the many amidases that have been described for the stereoselective
hydrolysis of chiral amides, only a few processes have been scaled up to pilot plant
volume. Lonza AG has used microorganisms containing (S)- and (R)-selective
amidases for the resolution of 2-piperazinecarboxamide (34) on a multi-kg scale
(Section 15.4.1.1) [52, 63]. In another process Lonza AG applied the amidase from
Comamonas acidovorans A:18 (DSM 6351), expressed in E. coli XL1Blue/pCAR6, for
the resolution of 2,2-dimethylcyclopropane carboxamide (38) (Scheme 15.5). The
process with suspended whole cells containing an (R)-selective amidase has been
performed under aqueous conditions at pH 7.0 and 37 C and run on 15 m3 scale [53,
54, 62, 63]. The (S)-enantiomer of the remaining amide 38 was isolated in 35% yield
and >98% e.e. and used as an intermediate in the synthesis of cilastatin, a renal
dehydroxypeptidase inhibitor used to prevent deactivation of penem antibiotics in the
kidneys.
Amidase
NH 2 OH + NH 2
Comamonas
O acidovor ans A:18 O O
(RS)-38 (R)-acid (S)-38
Scheme 15.5 Resolution of rac-2,2-dimethylcyclopropane carboxamide (38) by the amidase from

Comamonas acidovorans A:18.
Recently, a similar (R)-selective amidase was identified in the newly isolated

Brevibacterium epidermidis strain ZJB-07021, which was obtained via a continuous
enrichment strategy with (RS)-2,2-dimethylcyclopropane carboxamide (38) as sole
nitrogen source. Interestingly, the enantioselectivity of the B. epidermidis ZJB-07021
catalyzed resolution reaction of (RS)-38 was strongly influenced by temperature.
Whereas the E-ratio for this resolution reaction was 12.6 at 45 C, it increased to 65.9
at 14 C [117].
570
15.3
Enantioselective Hydrolysis of Cyclic Amides
Amidases can also be applied for the enantioselective hydrolysis of chiral

cyclic amides (lactams). A well-known example is the L-selective hydrolysis of DL-
a-amino-e-caprolactam (DL-ACL, 39) by the L-ACL lactamase from the yeast Crypto-
coccus laurantii, which is at the heart of the L-lysine process commercialized by Toray
Industries in 1970 [118]. By combining this L-selective lactamase with the ACL
racemase from the bacterium Achromobacter obae in one reaction vessel, DL-ACL is
converted into L-lysine 40 in near quantitative yield (Scheme 15.6) [119]. The L-ACL
lactamase containing C. laurantii strain was isolated from soil in a screening program
for microorganisms that were able to grow on L- and DL-ACL but not on D-ACL as
nitrogen source [120]. Owing to this enzymes high activity for L-ACL (240 mmol
min1 mg)1) [121]) and near absolute stereoselectivity, DL-ACL (100 g l1) was
completely resolved by 500 mg of the lyophilized yeast cells in only 7 h. The e.e. of
both the L-lysine and D-ACL obtained after crystallization was at least 99% [122]. As far
as we know, sequence information on this L-ACL lactamase has surprisingly not been
disclosed thus far.
O O
NH2 L-ACL
HN NH2
lactamase HO
NH2
L-39 L-40
ACL racemase
O
NH2
HN
D- 39
Scheme 15.6 Two-enzyme DKR process to L-lysine 40 operated by Toray Industries, Inc.
The ACL specific racemase has been studied in much greater detail [123–129].
Recently, the crystal structure of this racemase in its native form and in complex with
e-caprolactam has been solved [130].
Meanwhile commercial operation of this dynamic kinetic resolution type of
lysine process has been discontinued in favor of highly efficient fermentation
processes [118].
In addition, Chirotech Technology Limited, a UK based company that has been
part of Dr. Reddys Custom Pharma Services (CPS) since 2008, applies enantiose-
lective lactamases on a commercial scale. For instance, both enantiomers of the
15.3 Enantioselective Hydrolysis of Cyclic Amides j571
carbocyclic nucleoside precursor 2-azabicyclo[2.2.1]hept-5-en-3-one (41) are pro-
duced by applying c-lactamases with opposite stereoselectivity (Scheme 15.7). These
chiral lactams are versatile synthons for a growing several existing drugs as well as
new chemical entities [131, 132].
O
NH
+
H 3N CO 2-
+
(+)-Lactamase
(1R ,4S)-(-)-41 (1 R ,3 S)-(+)- γ-amino acid

O
NH
(±)-41
O
(-)-Lactamase HN
-
O2C NH 3 +
+
(1 S,4 R)-(+)-41 (1 S,3 R)-(- )-γ-amino acid
Scheme 15.7 Resolution reaction of rac-2-azabicyclo[2.2.1]hept-5-en-3-one (c-lactam 41) by

enantiocomplementary c-lactamases.
The development of this bio-resolution process dates back to the 1980s, when two
strains containing enantiocomplementary c-lactamases were isolated from the
environment using enrichment on a range of N-acyl compounds as the sole source
of carbon and nitrogen [133, 134]. Whereas Rhodococcus equi NCIB 40213
selectively hydrolyzed ()-2-azabicyclo[2.2.1]hept-5-en-3-one (41), yielding the
( þ )-c-lactam with >98% e.e. (45% yield), Pseudomonas solanacearum NCIB
40249 produced the ()-c-lactam with similar yield and e.e. due to the presence
of a ( þ )-c-lactamase [133]. These first-generation processes, which were executed
with whole-cell biocatalysts, were already characterized by a relatively low concen-
tration of biocatalyst (6 g-dry-mass l1), the high concentration of substrate
(50 g l1), and the speed of the reaction (completed in 3 h). Over the years the
processes to both lactam enantiomers have steadily been improved, especially
through isolation of more stable and more selective biocatalysts. An important step
forward was the isolation of a highly efficient ()-c-lactamase in an Aureobacterium
species in 1993 [135]. Besides this enzymes excellent enantioselectivity (E 7000),
its outstanding stability is especially noteworthy, allowing its preparative purifica-
tion and subsequent immobilization onto a glutaraldehyde-containing polymeric
support with 60% recovery. The immobilized lactamase appeared to be very stable,
as only little loss of activity was noted during 8 months of continuous operation in a
small reactor [136]. These features enable a highly efficient resolution process that
is currently running at 400 g l1 substrate concentration and gives the ( þ )-lactam
572
in 34% isolated yield and >99% e.e. [131]. The gene encoding this ()-c-lactamase
appeared to encode a protein of Mw 30978 with 68% sequence identity with two
Streptomyces aureofaciens cofactor-free haloperoxidases. The crystal structure of this
lactamase was solved in its native form as well as in complex with a covalently bound
ligand originating from the E. coli host cell [137, 138]. This enzyme is a homotrimer
and belongs to the a/b hydrolase fold family with a typical Ser-His-Asp catalytic
triad (residues Ser98, Asp230, and His259) and a positively charged oxyanion hole
formed by the main-chain nitrogen atoms of Tyr32 and Met99. Modeling of the
tetrahedral complexes between the active site Ser and both the ( þ )- and
()-c-lactam 41 also revealed why this c-lactamase exhibits exquisite stereospec-
ificity towards the ()-enantiomer [138]. Although this Aureobacterium
()-c-lactamase could be fermented using an E. coli expression system, over
five-times more efficient expression was obtained upon applying a Pseudomonas
fluorescens based protein production system. Because this strain is cultivated in a
simple defined medium, the enzyme produced can be certified animal free, which
is another advantage of the use of P. fluorescens as protein production host [131].
For a long time a ( þ )-c-lactamase with similar efficiency to the Aureobacterium
()-c-lactamase could not be isolated. Although promising ( þ )-c-lactamases were
identified in different Pseudomonas strains [135, 139] and in Kluyvera citrophila (now
Kluyvera cryocrescens) ATCC 21285 [140], lack of stability prevented immobilization
of these lactamases in isolated form. Therefore, a biotransformation process was
developed using a whole-cell biocatalyst. While scale-up of this process to tonne-
scale was possible, larger product volumes were difficult to produce with this
process due to several drawbacks of which the more complex downstream proces-
sing of the ()-c-lactam 41 due to cell lysis was the most serious [131, 141]. This
situation was finally changed by the isolation of a Comamonas acidovorans strain that
contained a considerably more stable lactamase [141]. The gene encoding this 575-
residue long ( þ )-c-lactamase was cloned and highly efficiently expressed in E. coli
in a medium without animal derived medium components, leading to a biocatalyst
that is certified animal free [131, 141]. This enzyme shows high sequence identity to
the formamidase from Methylophilus methylotrophus (63%) [142, 143] and the
acetamidase from Mycobacterium smegmatis (56%) [144], and thus seems to belong
to the same amidase family as the L-amidase from Ochrobactrum anthropi NCIMB
40321 (Section 15.4.2). This lactamase does not contain the amidase signature
sequence. Although the crystallization and preliminary X-ray analysis of this
( þ )-c-lactamase has been reported [145], its crystal structure has not been reported
yet. An efficient process with a substrate concentration of 500 g l1 and E > 400 was
developed applying this enzyme in semi-purified form. Because a clean concen-
trated enzyme is used in this process, the desired ()-c-lactam 41 can be isolated via
a much simpler work-up than when whole cells are used as biocatalyst, leading to
significant cost savings [131, 141].
In 2004, an even more thermostable ( þ )-c-lactamase with activity toward 2-
azabicyclo[2.2.1]hept-5-en-3-one (41) was isolated from the thermophilic archaeon
Sulfolobus solfataricus MT4 [105]. This enzyme, with a calculated molecular mass of
55.7 kDa, belongs to the signature amidase family. Its highest activity was observed at
15.3 Enantioselective Hydrolysis of Cyclic Amides j573
pH 7.0 and 85 C for the substrate ( þ )-c-lactam 41. Other substrates were (R)-
( þ )-lactamide 28 and a range of aliphatic and aromatic amides that are also
hydrolyzed by other signature amidases [105]. This enzyme in immobilized form
has recently been used as model enzyme to prove the suitability of a microreactor
system for enzyme substrate screening [146].
Besides the c-lactamase-based processes described above, b-lactamases have been
applied for the synthesis of enantiomerically pure b-lactams and b-amino acids. In
1991, Evans and colleagues described the use of Rhodococcus equi NCIMB 40213 cells
for the resolution of the isomeric b-lactam ()-6-azabicyclo[3.2.0]hept-3-en-7-one
(42), providing (1R,5S)-lactam 42, which is a precursor for the antifungal agent
cispentacin, in >99% e.e. and 40% yield (Scheme 15.8). The b-amino acid product
(1S,2R)-43 was obtained in 38% yield and 96% e.e. (determined as its corresponding
methyl ester acetamide) [147, 148]. Owing to the very low activity of this biocatalyst
(980 mg cell paste was needed to resolve 340 mg of rac-lactam 42 in 312 h),
commercialization of this process was not feasible.
NH3 +
NH R. equi NH
+
pH 7, RT
O O CO2-
(± )-42 (1R,5S)-42 (1S,2R)-43
Scheme 15.8 Rhodococcus equi catalyzed kinetic resolution of 6-azabicyclo[3.2.0]hept-3-en-7-

one (42).
In a screening of over 400 microbial strains for hydrolysis activity against the
lactams 44 and 45, scientists at Chirotech Technology Ltd. identified two micro-
organisms with much higher b-lactamase activity than R. equi [149, 150]. One of
these novel biocatalysts, Rhodococcus globerulus NCIMB 41042, appeared to be
highly enantioselective, and was used for the efficient resolution of the bi-/tricyclic
lactams 44–47. Both the residual b-lactam enantiomers and the corresponding
b-amino acid products 48–51 were obtained in acceptable to excellent isolated
yield and greater than 98% e.e. (E-ratio > 1000) (Scheme 15.9) [149–151].
Optimization of the bio-resolution of ()-6-azabicyclo[3.2.0]heptane-7-one (46)
led to a process that can be operated at 60 g l1 substrate concentration with
a 20% wt/wt cell paste loading, which was successfully applied for the synthesis of
multigram quantities of the enantiomerically pure b-lactam (1R,5S)-46 and
b-amino acid (1S,2R)-50 [149].
Enantioselective hydrolysis of substituted monocyclic lactams has been described
by BASF. Microorganisms with enantiocomplementary lactamase activities were
obtained from soil samples by an enrichment strategy applying racemic 5-vinylpyr-
rolidone (52) and 3-methylpyrrolidone (53) as the sole nitrogen source [152]. In
subsequent application research, (S)-5-vinylpyrrolidone with an e.e. of 98.6% was
prepared from the racemic lactam on applying whole cells of Rhodococcus erythropolis
DSMZ 9002. (R)-5-vinylpyrrolidone, on the other hand, was obtained when Pseudo-
monas aeruginosa DSMZ 9001 was used [152].
574
NH NH NH3+
R. globerulus
+
O pH 7, 37°C, 24 h O CO 2-
(±)-44 (1R,6S)-44 (1S,2R)-48
e.e. 98%; y 40% e.e. 99.3%; y 44%
O R. globerulus O -
+ O2C
+
NH pH 7, 37°C, 18 h NH H 3N
(±)-45 (1S,2S,5R,6R)-45 (1S,2S,3R,4R)-49
e.e. 99%; y 50% e.e. 99.4%; y 35%
NH NH NH3 +
R. globerulus
+
O pH 7, 37°C, 7 h O CO 2-
(±)-46 (1R,5S)-46 (1S,2R)-50
e.e. 99.3%; y 34% e.e. 99%; y 28%
NH NH NH3+
R. globerulus
+
O pH 7, 37°C, 5 h O CO 2-
(±)-47 (1R,6S)-47 (1S,2R)-51
e.e. 99%; y 38% e.e. 99%; y 31%
Scheme 15.9 Resolutions of bi- and tricyclic azabicyclo[3.2.0]heptan-7-one (46), and

b-lactams (1R,6S)-7-azabicyclo[4.2.0]oct-4-en- (1R,6S)-7-azabicyclo[4.2.0]octan-8-one (47)
8-one (44), (1S,2S,5R,6R)-3-azatricyclo catalyzed by whole cells of Rhodococcus
[4.2.1.02,5]non-7-en-4-one (45), (1R,5S)-6- globerulus NCIMB 41042 [149].
N O N O
H H
5-vinylpyrrolidone 3-methylpyrrolidone
(52) (53)
15.4
Enantioselective Hydrolysis of Amino Acid Amides
Enantiomerically pure amino acids, natural as well as synthetic ones, are extensively
used in the food, feed, agrochemical, cosmetics, and pharmaceutical industries. In
2004, the total world market for amino acids was estimated at approximately US$4.5
billion [153]. The largest share thereof is made up of the amino acids that are applied
15.4 Enantioselective Hydrolysis of Amino Acid Amides j575
as feed additives, that is, L-lysine, DL-methionine, and L-threonine. Monosodium
L-glutamate (MSG), which is used as a taste enhancer/seasoning agent, is another
proteinogenic amino acid produced in large volumes. In recent decades, enantio-
merically pure amino acids are also increasingly utilized for synthetic applications.
D-Phenylglycine (D-Phg) (54) and D-p-hydroxyphenylglycine (D-HPG) (55) are exam-
ples of important D-a-H-a-amino acids (Figure 15.1). These are produced in several
thousands of tons per year as side chains for the manufacturing of semi-synthetic
b-lactam antibiotics such as ampicillin (56) and amoxicillin (57) [154, 155]. D-Valine
(58), an intermediate for the pyrethroid insecticide fluvalinate (59) [156], is another
example of an industrially relevant D-a-H-a-amino acid. An L-a-H-a-amino acid
frequently used is L-tert-leucine (60); it is applied as a building block for various
antiviral (e.g., anti-human immunodeficiency virus), antiarthritic, and anticancer
drugs under development and as a chiral auxiliary in chemical asymmetric synthe-
sis [157, 158]. (S)-6-Heptenylglycine (61), (S)-4-hydroproline (62), and (2R,3S)-3-
vinyl-2-amino-2-cyclopropylcarboxylic acid (63) are further examples of a-H-a-amino
acids of commercial interest. The three non-natural amino acids are building blocks
of a precursor for BILN 2061 (64), a novel NS3 protease inhibitor with antiviral activity
in humans [159]. Besides a-H-a-amino acids, a,a-disubstituted a-amino acids also
constitute a group of compounds of increasing importance, as exemplified by the use
of L-a-methyl-3,4-dihydroxyphenylalanine (L-a-methylDOPA) (65) as antihyperten-
sive drug [160, 161], a-methylvaline (66) as an intermediate for the herbicide Arsenal
(67) and related herbicides [162, 163], and L-a-methylphenylglycine (68) as a building
block for the new fungicide fenamidone (69) [164].
The enzymatic kinetic resolution of racemic amino acid amides is one of the
chemoenzymatic processes for the production of enantiomerically pure amino
acids [165]. This section gives information on the enzymes reported for the stereo-
selective hydrolysis of a-H-a- and a-alkyl-a-amino acid amides, as well as on their
application.
15.4.1
Synthesis of Enantiopure a-H-a-Amino Acids
At DSM an efficient and universally applicable industrial process for the production
of enantiomerically pure a-H-a-amino acids was developed in the mid-1970s. This
chemoenzymatic process, which is based on the enzymatic kinetic resolution of
racemic a-H-a-amino acid amides 70 with L-selective amide hydrolases, was com-
mercialized by DSM in the mid-1980s for the production of several L- and D-amino
acids [166–168]. As a rule, unsubstituted amide substrates 70 are used as substrates,
which are readily available from simple raw materials by Strecker synthesis on the
corresponding aldehydes followed by hydrolysis under mild basic conditions (room
temperature, pH 10–12) in the presence of catalytic amounts of an aldehyde or
ketone [169]. The amide substrate is thus a precursor of the amino acid to be
prepared, which sets this process apart from the acylase process, another frequently
applied chemoenzymatic process for the production of enantiomerically pure amino
acids [170, 171].
576
NH 2 NH 2
NH S
CO2 H Me
O N Me
R R O
CO 2H
D-Phg (R = H) (54) Ampicillin (R = H) (56)
D-HPG (R = OH) (55 ) Amoxicillin (R = OH) (57)
NH 2 Cl O CN
Me NH O
CO2 H O
Me CF3 Me Me
D-Valine ( 58) Fluvalinate (59)
NH2
CO 2H H
N S
(S)-6-heptenylglycine (61) Me
N
Me N
HO
O
Me
N CO 2H O
H
H
(S)-4-Hydroxyproline (62 ) N CO 2H
N
H
O N O
H 2N CO2 H O
O
(2R,3S)-3-vinyl-2-amino- BILN 2061 (64)

2-cyclopropylcarboxylic acid
( 63 )
CO 2H·NEt 3
Me O NH
CO2 H
Me Me N N
Me NH 2 Me
α -methylvaline (66 ) Me
Arsenal (67)
Me
S
CO 2H N
N N
Me NH 2 Me
H
L-α-methylphenylglycine (68) O
Fenamidone ( 69)
NH 2
Me HO CO2H
CO2 H
Me Me NH 2
Me
HO
L-tert -leucine (60) L- α-MethylDOPA (65)
Figure 15.1 Examples of applications of a-H-a-amino and a,a-disubstituted a-amino acids.

If needed, the racemic amino acid amides 70 can also be prepared on a laboratory
scale by the alkylation of N-acetamidomalonate esters followed by hydrolysis and
amination of the amino esters [172]. Another method is based on the a-alkylation of
iminoacetic acid amides followed by an acidic hydrolysis. These imines can be easily
prepared by the reaction of glyoxylic acid esters with branched primary amines
followed by a reaction with methanolic ammonia. This method was successfully used
for the preparation of a wide array of racemic a-H-a-amino acid amides [173].
The amidase process at DSM has long been operated with permeabilized whole
cells of Pseudomonas putida ATCC 12633. This biocatalyst combines a nearly 100%
enantioselectivity for the hydrolysis of L-amides (enantiomeric ratio E > 200) with
good to excellent activity for a broad range of substrates (vide infra). In combination
with the fact that the amide hydrolysis is not thermodynamically limited, which
implies that the substrate conversion can, in principle, be quantitative at every
substrate concentration applied [170], both the L-amino acid and the D-amino acid
amide can be obtained in almost 100% e.e. at 50% conversion. The amidase process
can thus also be used for the production of enantiomerically pure D-amino acids. The
resolution reaction further furnishes one equivalent of ammonia (compared to the L-
amino acid amide) as side product. An elegant and scalable alternative to the use of a
basic ion-exchange resin for the separation of the L-amino acid and the D-amino acid
amide has been developed by DSM. After completion of the hydrolysis reaction one
equivalent of benzaldehyde is added to the reaction mixture to form the water-
insoluble Schiff base of the unreacted D-amino acid amide, which precipitates from
the solution (Scheme 15.10).
The amidase resolution process is hampered by a maximum yield of 50% of the
desired amino acid enantiomer per pass. Because the competitiveness of chemoen-
zymatic production methods for enantiomerically pure amino acids primarily
depends on the costs of substrate manufacturing [174], recycling of the unwanted
enantiomer by racemization will greatly improve the feasibility of amidase processes.
Ex situ racemization of the remaining D-amino acid amide can be easily carried out via
formation of the benzaldehyde Schiff base under basic conditions [175]. Because
formation of this Schiff base is also the basis for the separation of the L-acid and D-
amide, recycling can be performed without any additional step. Should the D-amino
acid (amide) be the required product, racemization and recycling of the L-amino acid
is also possible. In this case, the L-amino acid is first converted into its amide which is
subsequently racemized in its N-benzylidene form (Scheme 15.10).
An important improvement was made by the development of different
a-H-a-amino acid amide racemases that could be combined with the enantioselective
amidase resolution to obtain more cost-efficient fully enzymatic dynamic kinetic
resolution (DKR) processes. Already in the 1970s, such a process was operated by
Toray Industries for the production of L-lysine (Section 15.3) [118, 119]. In this
process, racemic a-amino-e-caprolactam (ACL, 39) is L-selectively hydrolyzed to L-
lysine 40 by the action of the L-ACL lactamase from Cryptococcus laurantii, whereas the
remaining D-ACL is in situ racemized by the ACL racemase from Achromobacter obae
(Scheme 15.6). More recently, Asano and Yamaguchi showed that this ACL racemase,
in contrast to earlier reports [125], is also able to racemize linear a-H-a-amino acid
578
1) NH 3
2) PhCHO
3) OH – (racemization)
4) H+/H 2O
R R
R Pseudomonas putida
NH 2 NH 2 + OH
H2 N H2 N H 2N
pH ~ 8.5 O
O O
37 °C
70 D-amide L-acid
PhCHO
R H+ R
1) OH – (racemization) NH 2 OH
Ph N H2N
2) H+/H 2O ∆
O O
D-acid
Scheme 15.10 Enantioselective hydrolysis of a-H-a-amino acid amides 70 by P. putida ATCC

12633.
amides [176], although with at least 35-fold lower specific activity than for ACL [177].
The A. obae ACL racemase has been combined with the D-aminopeptidase from
Ochrobactrum anthropi C1-38 (Section 15.4.1.3) for the synthesis of D-alanine from L-
and DL-alanine amide in near stoichiometric amounts [177, 178]. Similarly, complete
conversion into L-alanine was obtained when the ACL racemase was combined with
the L-amino acid amidase from Pseudomonas azotoformans IAM 1603 (Section
15.4.1.1) [178].
A more generic a-H-a-amino acid amide racemase has been isolated at DSM from
Ochrobactrum anthropi NCIMB 41129 [179]. The gene (amaR) that is responsible for
the racemase activity was identified by screening a pZErO–2 based E. coli
expression library [180]. The amaR gene appeared to encode a protein of 439 amino
acids with a calculated molecular weight of 46 810. Analysis of the amino acid
sequence revealed that AmaR belongs to the aminotransferase class-III pyridoxal-
phosphate-dependent family of proteins, and that it has 52% sequence identity to the
ACL racemase from A. obae.
AmaR combines a good thermostability with a broad pH optimum (6–10) and is
able to racemize a range of linear a-H-a-amino acid amides, although with at least
sevenfold lower activity than for ACL. Of the amino acid amides tested, highest
racemase activity was observed for aminobutyric acid amide, alanine amide, and
norvaline amide. Amino acid amides with a Cb branched side-chain, like valine
amide and isoleucine amide, as well as amino acid amides with an aromatic
side-chain, like phenylglycine amide and phenylalanine amide, are racemized with
low activity. Upon combining the O. anthropi AmaR with the P. putida PepA
aminopeptidase (Section 15.4.1.1), L-aminobutyric acid was formed in 95% conver-
sion and over 99% e.e. [181].
15.4.1.1 L-Selective a-H-a-Amino Acid Amide Hydrolase

Since the pioneering work of Greenstein and Winitz (Reference [182] and references
therein) demonstrating the absolute L-enantioselectivity of hog kidney amidase for
leucine amide, countless amino acid amide hydrolases have been described in open
and patent literature. Through the advent of recombinant DNA technologies robust
amino acid amide hydrolases from, for instance, thermophilic bacteria have become
available. Examples are an intracellular aminopeptidase from Sulfolobus solfatari-
cus [183] and leucine aminopeptidases from Thermotoga maritima [184], Bacillus
stearothermophilus [185], and Bacillus kaustophilus [186]. This last enzyme, which
belongs to the family of M17 leucine aminopeptidases (LAPs), has been studied in
great detail, including several investigations to identify the roles of conserved
residues that are not directly involved in catalysis, metal ion binding, and/or substrate
binding [187–189]. The thermostability of the B. kaustophilus LAP could be further
enhanced by immobilization in Ca-alginate/k-carrageenan beads (ID 1270), whereas
its resistance against oxidative damage was improved by replacement of the critical
methionine residues with leucine [190]. This engineering study resulted in the
generation of more active mutants as well. A very thermostable leucine aminopep-
tidase has been identified in Geobacillus thermoleovorans (formerly Bacillus thermo-
leovorans). This enzyme, which was developed as a thermotolerant debittering
enzyme for use in food protein processing, retains its full activity after a 1 h
incubation up to 90 C [191].
In addition, solvent resistant aminopeptidases have been described in the liter-
ature. A leucine aminopeptidase from the hyperthermophilic bacterium Aquifex
aeolicus, for instance, combines an extreme thermostability (retention of 80% of its
activity after 30 min at 100 C) with resistance to organic solvents like methanol,
ethanol, acetonitrile, tetrahydrofuran, dimethylformamide, dimethyl sulfoxide, diox-
ane, and isopropanol [192]. An even more solvent resistant extracellular leucine
aminopeptidase was recently purified from a solvent tolerant Pseudomonas aeruginosa
strain. This enzyme exhibited remarkable stability in both polar and nonpolar
solvents (1.22 < log P < 7), which was explained by the high content of hydrophobic
amino acid residues on the surface of the protein [193]. Interestingly, this amino-
peptidase appeared to be similar to the leucine aminopeptidase that is co-expressed
with several known virulence factors (e.g., elastase encoded by lasB) by P. aeruginosa
PAO1 at high cell density [194]. Recently, the processing of this aminopeptidase,
which is secreted as a proenzyme, has been elucidated. The two-step maturation
process starts with the proteolytic cleavage (e.g., by elastase) of the aminopeptidase
proenzyme at its C terminus, followed by an intramolecular autocatalytic removal of
the 12-amino acid propeptide from its N-terminus [195]. Many leucine aminopepti-
dases from parasites have been isolated and characterized in the last decade, because
these enzymes are seen as a potential target for new drugs. Examples here are the
580
characterization of the M17 leucine aminopeptidases from Plasmodium falciparum,

the causative malaria parasite [196, 197], and from three pathogenic Leishmania
species, protozoan parasites that cause visceral, cutaneous, and mucosal diseases in
human [198]. Surprisingly, the substrate specificity of these protozoan LAPs is much
more restricted than that of their bacterial counterparts.
In contrast to the nearly infinite number of amino acid amide hydrolyzing
enzymes described, reports detailing the enantioselectivity of such enzymes have
appeared much less frequently. One of the first papers dealing with an L-selective
a-amino amidase was published in 1981 by Kieny-LHomme et al. [199]. This paper
describes the kinetic resolution of DL-alanine amide (alternatively called a-amino-
propionamide) with an L-amino amidase that was isolated from Brevibacterium
A4 [200], a mutant strain of Brevibacterium R312 that had lost its wide-spectrum
acylamide amidohydrolase (EC 3.5.1.4), also named acetamidase [46, 201]. This L-
amino amidase displays activity for a wide range of L-a-amino acid amides including
the amides of the non-proteinogenic amino acids phenylglycine and aminobutyric
acid. Whereas the Km for the different substrates varies to a limited extent only, the
Vmax greatly changes with the nature and size of the side chain, with leucine amide
being hydrolyzed most efficiently by far [199].
About a decade later another enantioselective amidase (AmdA) was purified from
Brevibacterium R312. This enzyme, which was identified by its ability to hydrolyze
racemic 2-aryl and 2-aryloxy propionamides in an enantioselective manner [47], was
later found to catalyze the enantioselective acyl transfer reaction to hydroxylamine
from a wide range of amides including a-hydroxy and a-amino acid amides
(Scheme 15.2) [102]. With respect to the a-amino acid amides, the highest affinity
was observed for substrates containing a hydrophobic moiety, such as L-leucine
amide and L-phenylalanine amide, whereas L-methionine amide turned out to be
transformed most rapidly. All amino acid amides tested were converted L-selectively,
albeit to a different extend. The gene encoding this enantioselective 521-amino-acid
amidase (Mr 54 671) was cloned and efficiently overexpressed in E. coli applying the
T7 promoter system [45]. This secured easy access to large amounts of enzyme or
cells as these can be used directly in amide conversion reactions [202]. Recently, the
crystal structures of this dimeric enantioselective amidase, which belongs to the
amidase signature (AS) family, and its inactive mutant S195A in complex with
benzamide were solved [28]. These studies revealed that the N-terminal a-helical
domain, which is not found in other AS family enzymes like the Stenotrophomonas
maltophilia peptide amidase (Section 15.7.2), participates in dimer formation. The
residues forming the Ser-cisSer-Lys catalytic triad (Ser171, Ser195, and Lys96) typical
for the amidase signature family enzymes as well as the AS sequence (a conserved
stretch of approximately 130 amino acids (residues 95–204)), are located in the large
domain, which shows high structural similarity to the other family members. In this
catalytic triad, Ser195 acts as the nucleophile that forms the covalent substrate–en-
zyme intermediate in the course of the reaction. AmdA has a narrow active site that is
very hydrophobic as it does not contain any hydrophilic amino acids except for the

catalytic residues, and which is accessible through an entrance of 8 A in width,
explaining its strong preference for amide substrates with a hydrophobic moiety of
intermediate size [102]. Unfortunately, Ohtaki et al. did not give a structural basis for
this enzymes enantioselectivity [28].
Over the years similar enantioselective amidases have been identified in other
microorganisms as well. Examples are the amidases from Rhodococcus rhodochrous J1
(full length identity: 61%) [98], Pseudomonas chlororaphis B23 (full length identity:
49%) [76, 77], and Rhodococcus erythropolis MP50 (full length identity: 36%) [77, 89,
90]. Although these enzymes exhibit (S)-enantioselectivity towards several 2-aryl-
propionamides (Section 15.2) no, or only a marginal, enantioselectivity towards
a-amino acid amides and a-hydroxy acid amides has been reported.
L-Selective a-H-a-amino acid amide hydrolases have successfully been applied for
thesynthesisofthenon-proteinogenic aminoacidpiperazine-2-carboxylic acid(76)and
derivatives thereof. These enantiopure amino acids have found widespread application
as building blocks for different pharmacologically active agents. As alternative for
the existing synthesis procedures via, for instance, fractional crystallization of the
diastereomeric menthyl ester of N,N0 -dibenzylpiperazine-2-carboxylic acid [203],
asymmetrictransformation[204],andpalladium-catalyzedasymmetricsynthesis[205],
scientists at Bristol-Myers Squibb developed a biocatalytic route that was based on the
kinetic resolution of 4-(tert-butoxycarbonyl)piperazine-2-carboxamide (71) by
the cytosolic leucine aminopeptidase from porcine kidney (Scheme 15.11) [206]. The
substrate needed to contain the Boc group because the unsubstituted piperazine-2-
carboxamide (34) did not prove to be a useful substrate for the resolution reaction. The
(S)-acid formed in this reaction was subsequently separated from the remaining (R)-
amide using anion exchange chromatography and converted in a multistep procedure
into the nucleoside transport blocker ()-draflazine (72) [206].
O O O O O O
LAP
O N NH 2 O N O- + O N NH 2
NH NH NH
H2 O NH 4+
(RS)-71 (S)-acid (R)-71
Scheme 15.11 L-Leucine aminopeptidase (LAP) catalyzed resolution of racemic N-t-Boc

piperazine-2-carboxamide [(RS)-71].
N N OH OH
Cl NH2 N NH
N O
N O
F N O NH
H
Cl
O NH2
Draflazine (72) Indinavir (73)

582
At Lonza AG an alternative version of this enzymatic resolution process was

developed. By screening for microorganisms using an enrichment strategy applying
racemic piperazine-2-carboxamide (34) or piperidine-2-carboxamide (36) as sole
nitrogen source, strains were isolated that convert these amides in an enantiose-
lective way without the need for the additional Boc group [207, 208]. Thus (S)-
amidase-containing Klebsiella terrigena DSM 9174 cells were used for the synthesis of
(S)-piperazine-2-carboxylic acid ((S)-76) with 99.4% e.e. and 41% isolated yield
(isolated as the dihydrochloride salts). The (R)-isomer of this compound, on the other
hand, could be synthesized with 99.0% e.e. and 22% isolated yield using Burkholderia
sp. DSM 9925 cells containing an (R)-selective amidase [209]. Finally, (S)-piperidine-
2-carboxylic acid (pipecolic acid, (S)-77) of 97.3% e.e. was synthesized from the
corresponding racemic amide in 20% isolated yield on applying Pseudomonas
fluorescens DSM 9924. Besides the fact that the resolution reaction proceeds on the
free (so non-Boc protected) carboxamides with near absolute stereoselectivity, these
substrates can be easily prepared from 2-cyanopyrazine (74) and 2-cyanopyridine (75,
Scheme 15.12). A further advantage of this process is that the microorganisms can be
grown on the racemic carboxamides at the same time as the resolution reactions are
taking place [207, 210].
K.terrigena O O
DSM9174 -
HN O + HN NH 2 + NH4 +
NH NH
(S)-76 (R)- 34
O
N
N HN NH 2
N NH
74 (RS)-34 O O
HN O- + HN NH 2 + NH4 +
Burkholderia sp.
NH NH
DSM9174
(R)-76 (S)-34
O O O
N
NH 2 O- + NH 2 + NH4 +
N NH P.fluorescens
NH NH
DSM9924
75 (RS)-36 (S)- 77 (R)- 36
Scheme 15.12 Amidase catalyzed resolution reactions of piperazine-2-carboxamide (34) and

piperidine-2-carboxamide (36) developed by Lonza AG.
Since the N-tert-butyl amide of (S)-piperazine-2-carboxylic acid is applied as a

building block for indinavir (73), a HIV protease inhibitor marketed by Merck & Co.
under the name Crixivan , the resolution directly on the piperazine-2-tert-butylcar-
boxamide (35) is an attractive shortcut route. Asano and coworkers were the first to
report on the kinetic resolution of this amide, with a very bulky leaving group, in
2004. They succeeded in purification of an amidase from Pseudomonas azotoformans
IAM 1603 that hydrolyzed the (RS)-piperazine-2-tert-butylcarboxamide 35 completely
(S)-stereoselectively (Scheme 15.13) [75]. The gene for this hydrolase (laaAPa) was
cloned and efficiently expressed in E. coli under control of the lac promoter and with
an optimized ribosome-binding site (RBS). LaaAPa contains 310 amino acid residues
(Mr 34 514) and has significant homology to proline iminopeptidases (EC 3.4.11.5).
Because the enzyme could not hydrolyze peptidic substrates like L-prolyl-L-alanine
and L-prolyl-glycine, but displays activity towards L-Pro-NH2, L-Ala-NH2, and
L-Met-NH2, it was named L-amino acid amidase. Besides the amides of these
proteinogenic amino acids, LaaA also acted (S)-stereoselectively on (RS)-piperi-
dine-2-carboxamide (36), (RS)-piperazine-2-carboxamide (34) and (RS)-piperazine-
2-tert-butylcarboxamide (35). The presence of the bulky N-tert-butyl group in the
amide leaving group, however, led to an approximately 20-fold reduction in specific
activity. Table 15.2 gives more characteristics of LaaAPa. The same group also published
work on an amidase that hydrolyses (RS)-piperazine-2-tert-butylcarboxamide (35) with
opposite stereoselectivity [81]. This (R)-selective amidase (RamA) was identified in
Pseudomonas sp. MCI3434. Given that (S)-35 is the desired building block and both
enzymes have a rather similar specific activity towards this bulky substrate, the (R)-
selective amidase RamA is preferred over the L-amino acid amidase from P. azotofor-
mans IAM 1603. More information on RamA is given in Section 15.4.1.3.
O O
LaaA
HN O- + HN N + NH 2
H
NH NH
O H2 O
(S)-76 (R)- 35
HN N
H
NH
H2 O
O O
(RS)- 35
HN O- + HN N + NH 2
RamA H
NH NH
(R)- 76 (S)-35
Scheme 15.13 Resolution of (RS)-piperazine-2-tert-butylcarboxamide 35 using the L-amino acid

amidase from Pseudomonas azotoformans IAM 1603 (LaaAPa) and the (R)-selective amidase from
Pseudomonas sp. MCI3434 (RamA).
An L-stereoselective amino acid amidase with very broad substrate specificity has
been identified in the bacterium Brevundimonas diminuta TPU 5720 (Table 15.2) [211].
Of the amino acid amides tested as substrate, highest activity was observed for L-Phe-
NH2, followed by L-2-aminobutyric acid amide, L-Gln-NH2, L-Leu-NH2, L-Met-NH2,
584
Table 15.2 Properties of microbial L-selective a-H-a-amino acid amide hydrolases.
Pseudomonas putida ATCC Pseudomonas azotoformans IAM Brevundimonas

12633 L-aminopeptidase 1603 L-amino acid amidase diminuta TPU 5720 L-aminopeptidase
Reference [219–221] [75] [211]

Name ppLAP; PepA LaaAPa LaaABd

Subunit molecular weight 52 468 Da (calculated). 34 514 Da (calculated). 51 127 Da (calculated).
52 kDa (SDS-PAGE) 34 kDa (SDS-PAGE) 53 kDa (SDS-PAGE)
No. of subunits 6 (X-ray crystallography) 1 (gel filtration – 32 kDa) 6 (gel filtration – 288 kDa)
No. of amino acids 497 310 491
Homologous to Leucine aminopeptidases Proline iminopeptidases Leucine aminopeptidases
Isoelectric point (pH) 10.5 (IEFa))
Optimum pH 9.5 9.0 7.5
pH stability 6.0–9.5
Optimum temp. ( C) 40 45 50
Temp. stability ( C) (preincubation time) 30 (1 h; in presence of substrate) 40 (5 min) 50 (20 min)
Activation by Mg2 þ , Co2 þ , Mn2 þ , DTTa) No Co2 þ , Ni2 þ , Mn2 þ , Mg2 þ
Inhibitors Partly: Cu2 þ , Ca2 þ , Zn2 þ , o- Partly: Co2 þ , Ni2 þ , Pb2 þ , Completely: EDTA
phenanthroline, PMSFa), pCMBa), iodoacetate, N-
iodoacetamide. ethylmaleimide.
Completely: pCMBa), DFPa), Completely: Zn2 þ , Ag þ , Cd2 þ ,
EDTAa) Hg2 þ , phenylhydrazine
Substrate specificityb)
Gly-NH2 0.55 ND ND
L-Ala-NH2 0.55 11 3.3
L-Abu-NH2 4.4 — 96
L-Val-NH2 3.0 ND 17
L-Leu-NH2 100 0.46 89
L-Ile-NH2 5.5 0.17 29
L-Met-NH2 27 4.2 82
L-Phg-NH2 47 — —
L-Phe-NH2 6.6 0.97 100
L-Tyr-NH2 — 0.086 37
L-Trp-NH2 6.6 0.20 41
L-Pro-NH2 ND 100 2.2
L-Arg-NH2 — ND 78
L-Glu-NH2 ND — 11
L-Ser-NH2 0.38 0.43 2.8
L-Thr-NH2 — 0.12 14
L-Gln-NH2 — ND 91
(RS)-Piperidine-2-carboxamide (36) — 32 —
(RS)-Piperazine-2-carboxamide (34) — 3.7 —
Peptidase activity Yes; L-Phe-L-Phe, L-Phe-L-Leu, L- No Yes; L-Phe-L-Phe, L-Ala-Gly, D-Phe-L-Phe
Leu-L-Phe, L-Leu-L-Leu
a) IEF: isoelectric focusing; DTT: dithiothreitol, PMSF: phenylmethanesulfonyl fluoride; pCMB: p-chloromercuribenzoic acid; DFP: diisopropylfluorophosphate; EDTA:
ethylenediaminetetraacetic acid.
b) Substrate specificity is given relative to the activity for the substrate that is converted most efficiently. ND: not detected; –: not measured.
15.4 Enantioselective Hydrolysis of Amino Acid Amides
j585
586
and L-Arg-NH2. Because a high activity was also obtained with dipeptides, the enzyme
(LaaABd) may thus be called an aminopeptidase. LaaABd converts amino acid amides
(S)-stereoselectively without exception, but to what extent depends on the type of side
chain. The gene for this enzyme appeared to code for a 491 amino acid protein with a
calculated molecular weight of 51 127; because the molecular mass of the native
enzyme was estimated by gel filtration to be about 288 000 Da, LaaABd is most likely
active as a homohexamer. Based on its primary structure LaaABd should be catego-
rized as an M17 leucine aminopeptidase. The positive effect of Co2 þ on the activity of
this enzyme and its complete inhibition by EDTA (ethylenediaminetetraacetic acid)
further support this classification (Section 15.4.1.1).
In 2009 Tishinov and coworkers applied the major aminopeptidase from sun-
flower seed (Helianthus annuus L.) for the kinetic resolution of different a-amino acid
amides [212]. Both racemic Phe-NH2 and Phg-NH2 were completely resolved by this
enzyme into the (S)-acid and (R)-amide due to the absolute stereoselectivity of this
enzyme for these two amides (E > 300). This aminopeptidase also displayed high (S)-
selectivity for the aliphatic amides Ala-NH2 and Leu-NH2. A structured investigation
of the substrate specificity showed that this enzyme has a hydrophobic S1-sub-
site [213] of limited size; the S0 1 -subsite, on the other hand, allowed a great variety of
leaving groups. Furthermore, the enzyme has an absolute requirement for a free
amino group at the acyl part of the substrate. This aminopeptidase was purified to
homogeneity from the sunflower seed water extracts and characterized. It is an
80 kDa (SDS-PAGE) enzyme with isoelectric point of 4.6 and optimal activity between
pH 7.5 and 8.0 and 45 and 50 C. The enzyme is strongly inhibited by thiol-modifying
reagents (e.g., p-hydroxymercuribenzoate and 5,50 -dithio-bis(2-nitrobenzoate)),
which implies a crucial role of a sulfhydryl group in catalysis. Because chelating
agents did not influence this protein, metal ions are not involved in enzyme function
or stability [214]. Cloning of the gene encoding this interesting aminopeptidase has
not been reported yet.
Although several enantioselective amidases with activity towards a-H-a-amino
acid amides have been described since the identification of P. putida ATCC
12633 [215], this strains aminopeptidase is still one of the preferred catalysts for
industrial application due to a set of unique features. As indicated above, P. putida
ATCC 12633 cells combine an exquisite enantioselectivity with a broad substrate
specificity. At DSM, already over 100 different a-H-a-amino acid amides have been
successfully resolved, furnishing both the L-amino acid and D-amino acid amide in
high enantiomeric excess. The size of the side chain may range from the small methyl
group in alanine amide to the very bulky group in b-naphthylglycine amide or lupinic
acid amide [166, 168, 216]. Furthermore, heteroatoms like sulfur, nitrogen, and
oxygen are accepted in alkyl or (hetero)aryl side chains. In addition, cyclic amino acid
amides like proline amide and piperidine-2-carboxyamide (36), and amino acid
amides with alkenyl and alkynyl substituents can be resolved by this biocatalyst [217,
218]. When applying whole cells, only for methionine amide and homomethionine
amide was a somewhat lower enantioselectivity was observed, which is probably
caused by enzymatic racemization of the L-amino acid under the basic conditions
applied (vide infra). The only prerequisite for activity of the P. putida catalyst is the
presence of a hydrogen atom on the a-carbon; thus, a,a-disubstituted amino acid
amides cannot be hydrolyzed.
Protein purification experiments identified an L-aminopeptidase that contributes
to a considerable extent to the broad substrate specificity of P. putida ATCC
12633 [219, 220]. This enzyme, which has a subunit molecular mass of 52 kDa and
an isoelectric point of 10.5, exhibits activity at pH 7–11 with an optimum at pH
9.0–9.5 and 40 C. Divalent metal ions have a marked effect on the activity of this
enzyme: whereas Cu2 þ and Ca2 þ significantly inhibit the aminopeptidase, Mg2 þ ,
Co2 þ (two- to threefold), and especially Mn2 þ (12-fold) stimulate its activity. The
enzyme displays activity for a broad range of a-H-a-amino acid amides, which are
hydrolyzed L-selectively without exception, and for dipeptides. Simple amides (e.g.,
acetamide and butyramide) and a,a-disubstituted amino acid amides, on the other
hand, are not converted. Based on its substrate range, the enzyme was classified as an
L-aminopeptidase [219].
The gene encoding this aminopeptidase (pepA) was cloned by reversed genet-
ics [220]. It encodes a protein of 497 amino acids with a calculated molecular weight of
52 468. Protein database searches revealed that this enzyme belongs to the M17
peptidase family (leucine aminopeptidase (LAP) family – MEROPS Accession:
MER001235), which contains zinc- and manganese-dependent exopeptidases (EC
3.4.11.1). More information on M17 LAPs can be found in Section 15.4.1.1. Efficient
overexpression of P. putida LAP in E. coli was realized by placing the pepA gene under
control of the trp promoter. This resulted in a highly active E. coli based whole-cell
aminopeptidase biocatalyst [220].
An important advantage of this novel whole-cell biocatalyst became manifest in
the preparation of a set of enantiopure unsaturated a-H-a-amino acids (Figure
15.2) [222]. In general, the resolution of these unsaturated amino acid amides with
the recombinant E. coli based system as well as with the wild-type P. putida cells
proceeded smoothly, yielding the L-acid and the D-amide in high enantiomeric
excess (>95%) at 50% conversion. Owing to the over 25-fold higher expression of
the P. putidaL-aminopeptidase in the E. coli cells, they could be applied in a
cell : substrate ratio of only 1 : 500, whereas this ratio needed to be 1 : 10 with the
wild-type P. putida cells. More interestingly, however, L-3-butenylglycine (79a), L-3-
butynylglycine (82a), and the methylated homolog of L-3-butynylglycine (84a) were
obtained in a moderate e.e. only (97, 91, and 70%, respectively) using the wild-type
P. putida cells, whereas these L-amino acids were obtained in a superior e.e. (>99,
97, and 99%, respectively) when the resolution reactions were performed with the
recombinant E. coli biocatalyst. Further experiments revealed that this unsatisfactory
low e.e. of these L-acids using P. putida cells as biocatalyst originated from the
presence of an amino acid racemase that is absent in the recombinant E. coli whole-
cell biocatalyst. This racemase has a narrow substrate specificity and also recog-
nizes, besides the three unsaturated amino acids mentioned above, the proteino-
genic amino acid methionine 86 [222], which is in line with an earlier study that
showed that these unsaturated amino acids are structurally and electronically related
to methionine, and consequently are excellent methionine analogues in enzymatic
reactions [223].
588
H 2N COX H 2N COX H 2N COX

78 81 84

79 82 85

80 83 86
a: X = OH ; b: X = NH2
Figure 15.2 Unsaturated amino acids resolved by Wolf et al. using a recombinant E. coli based
whole-cell biocatalyst expressing the P. putida LAP [222].
The recombinant E. coli biocatalyst was subsequently used to prepare the L- and D-
enantiomers of amino acids 78a–85a on a multi-gram scale using the procedure
depicted in Scheme 15.14. After standard work-up of the unreacted D-amides, these
were hydrolyzed under mild conditions by the non-selective amidase present in
Rhodococcus erythropolis NCIMB 11540 cells. Chemical hydrolysis was not an option
for this type of unsaturated amide, since the harsh acidic conditions needed would
lead to decomposition of some of the side chains. All L-acids were obtained in above
98% e.e. except for 85a, which was obtained in an e.e. of 96%. The e.e.s of the D-acids
appeared to be excellent without exception [222].
15.4.1.2 Leucine Aminopeptidases of the M17 Family

LAPs catalyze the hydrolytic removal of amino acids from the amino terminus of
polypeptides. They are widely distributed in nature, being ubiquitously present in
all kingdoms of life. They play a key role in the processing and regular turnover
of intracellular as well as in the utilization of exogenous proteins and peptides [224–
229]. LAPs hydrolyze peptides and amino acyl substrates with an N-terminal leucine
residue most efficiently, although substantial cleavage rates are also observed with
various other amino acids in the P1 position (e.g., Ala, Ile, Arg, and Met) [224, 230,
231]. Peptides with an amino terminal Gly and Asp, on the other hand, are
inefficiently cleaved [232]. Comparison of the specificity of the M17 LAPs from
tomato, E. coli, and porcine showed that the penultimate (P10 ) residue also strongly
influences the rate of peptide hydrolysis. An Asp and Lys residue in this position
decreases the activity of these enzymes markedly, whereas dipeptides with a P10 Pro
residue are almost resistant to cleavage. The P20 residue, in contrast, appears to have a
much less pronounced influence, although arginine in this position leads to a
R E.coli / pTrpLAP R R
NH2 NH 2 + OH
H 2N pH 9.2, 40°C H2 N H 2N
O O O
D-78b - 85b L-78a - 85a
PhCHO
R
NH 2
Ph N
O
HCl
Acetone
R
- NH 2
Cl+H3 N
O
R.erythropolis R
NCIMB 11540 OH
H 2N
pH 8, 37°C O
D-78a - 85a
Scheme 15.14 Optimized amidase-based process for the multi-gram synthesis of enantiomerically
pure L- and D-unsaturated amino acids [222].
significant (three- to sixfold) reduction in activity of all three enzymes [232]. Earlier
papers reported that LAPs also catalyze the hydrolysis of amino acid amides,
alkylamides, arylamides, and hydrazides [227, 233], as well as several amino acid
esters [225]. LAPs are thus characterized by a broad substrate specificity.
At present the leucine aminopeptidases from bovine lens (blLAP) and E. coli
(ecLAP) are clearly the best studied representatives of this class of enzymes. X-Ray
crystallographic studies of blLAP and ecLAP, which share 31% and 53% sequence
identity with P. putida L-aminopeptidase PepA (ppLAP), respectively, have provided
important insights into the structure and catalytic mechanisms of the M17 LAPs.
Comprehensive overviews of the structural features of both enzymes have been
written by Str€ater and Lipscomb [234] and Colloms [235]. Recently, also the crystal

structures of ppLAP in its native form (2.2 A) and in complex with the inhibitor

bestatin (1.5 A) have been reported [221]. These crystallographic studies revealed a
common architecture for the three M17 LAPs. Their monomers consist of two mixed
a/b-type globular domains of different size, which are linked by a long a-helix
(Figure 15.3b). The active sites of the LAPs are entirely located in the larger and well-
conserved C-terminal domain, which is consequently also referred to as catalytic
domain. The smaller N-terminal domain, in contrast, is more variable between
different LAPs. In the native enzyme, six of these monomers assemble into a homo-
hexameric protein with two layers of trimers stacked on top of each other, which gives
these enzymes their characteristic triangular shape when viewing along the threefold
590
Figure 15.3 X-ray structure of the P. putida cavity containing the six active sites. (b) Ribbon
leucine aminopeptidase (ppLAP) at high representation of the ppLAP monomer. The long
pH [221]: (a) Ribbon representation of the a-helix that connects the N-terminal domain
ppLAP hexamer viewed along its threefold (N-domain) at the top to the catalytic
symmetry axis. The long a-helices that connect C-terminal domain (C-domain) at the bottom is
the N- and C-terminal domains in each indicated in dark gray. The two catalytic metal
monomer are indicated with darker shades. The ions (M1 and M2: site-1 and site-2 metal ion,
six Mn2 þ and six Zn2 þ ions are in the interior of respectively) are indicated as spheres.
the protein at the edge of the central solvent
symmetry axis (Figure 15.3a). Whereas the N-terminal domains extend outwards to
the corners of the triangular shaped proteins where they provide most of the
interactions that are required for trimer-trimer formation, the six active sites in the
C-terminal domains are located in the interior of the hexamer at the edge of a disk-
shaped solvent cavity [233]. Substrates enter this central cavity through six channels
that are located between the N-terminal domains and the hexamer core [236].
Each M17 LAP subunit has two non-equivalent metal-binding sites in the C-

terminal domain (M1 and M2), which are approximately 3.0 A apart [237]. The M1 site
is the more accessible of the two sites and the metal ion in this site can be readily
exchanged. In contrast, the M2 site is more deeply embedded in the protein and,
consequently, exchanges metal ions only slowly. In the native blLAP and ecLAP, both
sites are occupied by Zn2 þ ions. It has been shown that the Zn2 þ ion in site M1 can
be replaced by Mn2 þ , Mg2 þ , and Co2 þ by simply incubating the native Zn/Zn
blLAP with these metal ions in high concentrations [238, 239]. The Zn2 þ in binding
site M2, on the other hand, can be exchanged for Co2 þ only, and this is preferably
done via the metal-free apoenzyme [240, 241]. Kinetic studies have shown that
substitution of the metal ions in the two binding sites exert significant effects on both
Km and kcat of blLAP, suggesting a role in substrate binding and transition state
stabilization [240]. Recently, it has been found that exchange of Zn2 þ in site M1
against Mn2 þ can even lead to a totally new enzymatic activity. Whereas the dipeptide
CysGly cannot be hydrolyzed by Zn/Zn blLAP, and actually acts as a competitive
inhibitor for the hydrolysis of LeuGly, the Mn/Zn enzyme readily converts this
cysteinyl containing substrate [242, 243]. Interestingly, the metal binding site M1 in
ppLAP that has been heterologously produced in E. coli is already occupied by Mn2 þ ,
which leads to a much more active enzyme than when this site contains Zn2 þ
(ID 1313).
The LAP residues coordinating the two metal ions are fully conserved in the active
sites of blLAP, ecLAP, and ppLAP. Both metal ions are pentacoordinated in an
arrangement best described as a distorted octahedron from which one ligand is
missing [233, 244, 245]. In addition to a metal-bound water molecule, which was
observed in the high-resolution structure of these three M17 LAPs in unliganded
form [221, 237, 246], the two metal ions are bridged by the two side-chain carboxylate
oxygen atoms of (ppLAP numbering) Glu-351 (so bidentately) and one side-chain
carboxylate oxygen atom of Asp-272 (so monodentately). Further ligands are one side-
chain carboxylate and the main-chain carbonyl oxygen atoms of Asp-349 to the metal
ion in site M1, and one side-chain carboxylate oxygen atom of Asp-290 and the side-
chain amino group of Lys-267 to the site M2 metal ion [221].
Based on a large number of structural studies, which were mainly executed with
blLAP, and other biochemical data, a reaction mechanism has been proposed in
which the di-metal ion bridging water molecule acts as the nucleophile Figure
15.4 [237, 247]. The catalytic cycle starts off with the binding of the peptidic substrate
with the terminal amino group coordinated to the site-2 Zn2 þ and the carbonyl group
coordinated to the site-1 Mn2 þ /Zn2 þ . Thus, substrate binding leads to an increase of
the coordination number of both metal ions from five to six. The carbonyl group is
then polarized by coordination to the site-1 Mn2 þ /Zn2 þ and hydrogen bond
592
Figure 15.4 Reaction mechanism of peptide bond hydrolysis by leucine aminopeptidases as

proposed by Str€ater et al. Residue numbering according to ppLAP. Figure adapted from
References [237, 247].
formation with the side-chain amino group of Lys-279. Nucleophilic attack of the
bridging water molecule, which is activated by proton transfer to the bicarbonate ion
adjacent to Arg-353 acting as a general base [237, 246], then affords the gem-diolate
tetrahedral intermediate that is stabilized by a hydrogen bond from the protonated
side-chain of Lys-279 and coordination of both gem-diolate oxygen atoms to the di-
metal reaction center. Finally, this intermediate collapses, which is facilitated by
transfer of a proton from the bicarbonate to the NH-leaving group.
By modeling different amino acid amides into the active site of ppLAP in
energetically favorable conformations and under consideration of crucial binding
interactions and the reaction mechanism described above, the structural basis for the
broad substrate specificity and the exquisite enantioselectivity of the P. putida LAP
was established [221]. While amino acid amides with the (S)-configuration at their
a-carbon atom can bind in a productive mode, their optical antipodes are excluded
from the active site either due to steric hindrance of their Ca side chains or due to
unfavorable interactions with their Ca-linked amino and carbonyl groups. Similarly,
this approach showed that the distance between the Ca proton of the a-H-amino acid

amide substrates and the nearest residue of ppLAP is <3 A, which explains why this
enzyme is inactive with a,a-disubstituted amino acid amides. Finally, this structural
study also confirmed steric hindrance as the major reason why amino acid amides
with an additional methyl group at the Cb atom (e.g., L-valine amide and L-isoleucine
amide) are poor substrates for ppLAP [219]. Likewise, intolerance of their polar or
charged side chains by the hydrophobic S1 subsite of the enzyme was established as
the main reason why the amides of L-aspartic acid, L-glutamic acid, and L-serine are
inefficiently hydrolyzed by ppLAP.
Protease Independent Functions of Bacterial Leucine Aminopeptidases A fascinating

feature of many bacterial LAPs is their ability to bind DNA, as a result of which they
can exert accessory functions as transcriptional regulators and mediators of site-
specific DNA-recombination. Interestingly, at least some of these functions are
independent of their proteolytic activity [235].
One of such intriguing functions is the role E. coli PepA (alias XerB and ecLAP)
plays in Xer-mediated site-specific recombination [248]. This type of recombination
acts among other things at the cer and psi sites found in the multicopy plasmids ColE1
and pSC101, respectively, effecting the conversion of multimers into multiple
monomers, thereby ensuring a more uniform segregation of the plasmids to the
daughter cells and thus an improved plasmid stability [249]. To resolve ColE1
multimers, E. coli PepA in conjunction with ArgR bind to the accessory sequences
of two cer sites, resulting in formation of a protein–DNA complex in which the two
recombination sites are interwrapped around the PepA/ArgR complex. In this way
the two cross-over sites are brought together for strand exchange, which is catalyzed
by the heterotetrameric XerCD recombinase, consisting of XerC and XerD [246, 250,
251]. The fact that the peptidase activity of PepA is not required for this role in the Xer-
specific recombination confirms such architectural role [252]. Based on the E. coli
PepA structure and a comparison with blLAP, Str€ater et al. identified a groove at the
molecular surface of E. coli PepA running from one trimer face across the twofold
symmetry axis to the other trimer face, which might be the DNA-binding site.
Moreover, the authors proposed a model for the cer synaptic complex in which the two
cer sites wrap around two hexamers of PepA and one hexamer of ArgR aligned along
their threefold symmetry axes in a molecular sandwich [246]. The role of this groove
and the decisive role of the N-terminal domain for DNA binding was later corrob-
orated in an extensive mutagenesis study [253]. This work also led to the proposal of a
new model for the synaptic cer complex in which the two cer sites are wrapped around
a single PepA hexamer. In contrast with this, recent studies applying atomic force
microscopy to analyze the architecture of this synaptic complex showed that both the
shape and the volume of the PepAcer complex suggested the presence of two PepA
hexamers in the interwrapped synaptic complex instead of one [254]. More research,
594
for instance to determine the stoichiometry of the synaptic complex, is needed to

conclusively establish this complexs architecture.
Escherichia coli PepA, again with ArgR, is also involved in the regulation of the
expression of the E. coli carAB operon, which encodes the two subunits of the sole
carbamoyl phosphate synthetase (EC 6.3.5.5), that catalyzes the formation of carba-
moyl phosphate, a precursor in the metabolic pathways to arginine and pyrimidine
residues [255]. In this regulatory process, PepA induces an extreme remodeling of
one of the two car control regions (carP1), ensuring the cross-talk between various
upstream bound transcription regulators (IHF, PyrH, PurR, and RutR) and the RNA
polymerase [254]. Similar to its role in the cer/Xer-mediated resolution of ColE1
plasmid multimers, the aminopeptidase activity of E. coli PepA is dispensable for this
regulatory role in the expression of the carAB operon [255]. E. coli PepA also
negatively regulates its own expression by binding to a region overlapping the
outermost of the three promoters in the pepA control region, which interferes with
transcription initiation [255].
Homologues of E. coli PepA play an important role in the regulation of protein
expression in other microorganisms, too. In Vibrio cholera, for instance, PepA (81%
sequence identity to E. coli PepA) is involved as negative regulator in the formation of
virulence factors such as the cholera toxin and the toxin-co-regulated pilus
(TCP) [256]. The Pseudomonas aeruginosa PhpA (56% sequence identity to E. coli
PepA) was identified as a negative regulator of the algD transcription and thus of the
alginate biosynthesis in this strain [257]. In contrast to the alternative functions of
M17 LAPs mentioned above, the aminopeptidase activity of PhpA is essential for this
regulatory function, which almost certainly excludes the idea that this protein
operates as a DNA-binding transcriptional regulator in alginate biosynthesis [258].
More information on these alternative functions of bacterial M17 LAPs can be found
in recent reviews [181, 259].
Whether the P. putidaL-aminopeptidase also plays any of the additional roles
described above remains to be demonstrated.
15.4.1.3 D-Selective a-H-a-Amino Acid Amide Hydrolase

Although enantiomerically pure D-amino acids are accessible through the amidase
process while employing L-selective amide hydrolyzing enzymes (Scheme 15.10),
this requires an extra chemical or enzymatic hydrolysis step in which the non-reacted
D-amino acid amide is converted into the corresponding D-amino acid. Access to D-
selective amino acid amide hydrolases thus allows a one-step shorter production
route to D-amino acids. More importantly, however, these enzymes are essential when
a dynamic kinetic resolution type of amidase process for a D-a-H-a-amino acid is
desired (Section 15.4.1).
For a long time the number of available D-selective amino acid amide hydrolyzing
enzymes was limited, but this situation has changed for the better in the last two
decades through, among others, the work of Asano and colleagues [260–262]. To
obtain biocatalysts for the synthesis of D-amino acid derivatives, they screened for
suitable microorganisms using an enrichment culture technique that was based on
growth in media with D-alanine amide and D-valine amide as sole nitrogen source.
This led to the isolation of Ochrobactrum anthropi SCRC C1-38 and Ochrobactrum
anthropi SV3, respectively, from which D-stereospecific hydrolases with rather
different features were purified and characterized (Table 15.3) [263, 264].
The enzyme from O. anthropi SCRC C1-38 showed strict D-stereospecificity for low
molecular weight D-amino acid amides such as D-alanine amide, D-a-aminobutyric
acid amide, D-threonine amide, and D-serine amide, for D-alanine N-alkylamides such
as D-alanine-p-nitroanilide, D-alanine benzylamide, and D-alanine n-butylamide, and
for peptides with a D-alanine at their N-terminus such as D-alanylglycine, D-alanyl-L-
alanine, and D-alanine oligomers [263]. In addition, the methyl esters of glycine,
alanine, and serine were effectively hydrolyzed, albeit with low stereospecificity for
the latter two. Simple amides such as acetamide and propionamide as well as
N-protected amino acid amides were not hydrolyzed by this enzyme, which further-
more did not show endopeptidase and carboxypeptidase activity. Because of the range
of substrates hydrolyzed by this enzyme, it was classified as D-aminopeptidase (DAP
or DmpB) (EC 3.4.11.19). The gene encoding this D-aminopeptidase (dap) was cloned
and efficiently expressed in E. coli under control of the lac promoter to about 30% of
the total extractable cellular protein. This correlated to a 3600-fold overexpression as
compared with the activity in O. anthropi SCRC C1-38 [265]. Based on its low but
significant sequence identity with the Streptomyces R61 DD-carboxypeptidase and the
E. coli class C b-lactamase, DAP was assigned to the family of penicillin-recognizing
proteins. This classification was further supported by the presence of the three motifs
that are conserved in all members of this family of enzymes (S-X-X-K, Y/S-X-N, and K/
H/R-T/S-G), and its inhibition by b-lactam compounds without being a substrate for
the enzyme. DAP is thus a serine peptidase rather than a cysteine one as was
originally reported because of its inhibition by p-chloromercuribenzoic acid (pCMB)
and other thiol-blocking reagents [263]. It has been postulated that this inhibition
may have been caused by steric hindrance resulting from reaction of pCMB with Cys-
60, the residue adjacent to the catalytic nucleophile Ser-61. The crucial role of Ser-61
(and Lys-64, which are both part of the S-X-X-K motif) for catalysis has been
confirmed by mutagenesis studies [265]. The crystal structure of DAP has been

determined to 1.9 A resolution [266]. This aminopeptidase folds into three domains,
A, B, and C. The N-terminal domain (domain A) contains all catalytic residues and is
structurally highly analogous R61 DD-carboxypeptidase and serine b-lactamases.
Domains B and C, on the other hand, are both antiparallel eight-stranded b barrels.
The function of domain B is probably to maintain domain C in a good position to
protrude its c-loop into the active site of the enzyme, thereby conferring DAP its
unique substrate- and inhibitor-specificity [266]. The importance of this c-loop for the
specificity of DAP was recently confirmed in a directed mutagenesis study [267].
Removal of this 13-residue loop in combination with the Asn275Arg mutation
completely modified DAPs specificity from a D-aminopeptidase to a DD-carboxypep-
tidase. A DAP mutant with increased thermostability was obtained by random
mutagenesis [268].
Over the years the O. anthropi SCRC C1-38 D-aminopeptidase has been tested for
different applications. Intact cells as well as cell-free extracts of E. coli expressing this
enzyme were used for the resolution of several racemic amino acid amides [269].
596
Table 15.3 Properties of microbial D-selective a-H-a-amino acid amide hydrolases.
Ochrobactrum anthropi SCRC C1-38 Ochrobactrum anthropi SV3 Arthrobacter sp. NJ-26
Reference [263, 265] [264, 273] [284]

Name D-Aminopeptidase (DAP, DmpB) D-Amino acid amidase (DaaA) D-Alanine amidase
Subunit molecular weight 57 257 Da (calculated). 40 082 Da (calculated). 51 kDa (SDS-PAGE)

59 kDa (SDS-PAGE) 40 kDa (SDS-PAGE)
No. of subunits 2 (gel filtration – 122 kDa) 1 (gel filtration – 38 kDa) 1 (gel filtration – 67 kDa)
No. of amino acids 520 (Nb. Met-1 is posttranslationally 363
removed)
Homologous to DD-carboxypeptidases, b-lactamases DD-carboxypeptidases, penicillin-
(class C), penicillin-recognizing recognizing enzymes
enzymes
Isoelectric point (pH) 4.2 5.3 5.2
Optimum pH 8.5 9.0 7.5
pH stability 7.0–10.0 6.5–9.5
Optimum temp ( C) 45 45 45
Temp. stability ( C) (preincuba- 45 (10) 35 (5)
tion time, min)
Activation by
Inhibitors Partly: Ca2 þ , Cu2 þ , Zn2 þ , Cd2 þ , Partly: Cd2 þ , Pb2 þ , Ni2 þ , Co2 þ , Partly: Ba2 þ , Co2 þ , Mn2 þ .
Ni2 þ , DTNBa), N-ethylmaleimide, Hg2 þ , Ag þ , Cu2 þ
hydroxylamine, b-lactams (e.g., 6-
APA, 7-ACA, ampicillin).
Strongly: Ag þ , pCMB Completely: PMSF, Zn2 þ Strongly: Zn2 þ , Cu2 þ , pCMB
b)
Substrate specificity
Gly-NH2 44 ND 97
D-Ala-NH2 100 23 100
D-Abu-NH2 30 — —
D-Val-NH2 ND 0.50 —
D-Leu-NH2 ND 37 0.37
D-Met-NH2 2.0 17 —
D-Norleu-NH2 0.80 16 —
D-Norval-NH2 1.8 12 —
D-Phg-NH2 0.70 15 —
D-Phe-NH2 ND 100 0.64
D-Tyr-NH2 ND 98 —
D-Trp-NH2 ND 79 —
D-Pro-NH2 ND 6.2 —
D-Glu-NH2 ND ND —
D-Asp-NH2 ND 0.16 —
D-Ser-NH2 29 0.92 —
D-Thr-NH2 9.0 0.46 —
D-Gln-NH2 ND 0.39 —
D-Lys-NH2 ND ND —
D-Lactic acid amide — — —
D-Ala-pNa 96 0.15 —
Peptidase activity Yes, for example, D-Ala-Gly, D-Ala- No No
Gly2, D-Ala2, D-Ala3, D-Ala4, D-Ala-L-
Ala, DL-Ala-DL-Ser, DL-Ala-DL-Met, DL-
Ala-DL-Phe, DL-Ala-DL-Asn
a) DTNB: 5,5-dithiobis(2-nitrobenzoic acid).

j597
598
Table 15.4 Properties of microbial D-selective a-H-a-amino acid amide hydrolases.
Variovorax paradoxus DSM Delftia acidovorans Brevibacillus borstelensis Brevibacterium iodinum

14468 BCS-1 TPU 5850
Reference [285, 287] [289] [292] [290]

Name D-Amidase D-Amino acid amidase D-Alanine amidase D-Amino acid amidase
(DamA)
Subunit molecular weight 49 605 Da (calculated) 50 kDa (SDS-PAGE) 29 045 Da (calculated) 30 035 Da (calculated)
50 kDa (SDS-PAGE) 30 kDa (SDS-PAGE) 30 kDa (SDS-PAGE)
No. of subunits 2 (gel filtration – 101 kDa) ?a) 6 (gel filtration – 199 kDa) 10 (gel filtration –
290 kDa)
No. of amino acids 465 (Nb. Met-1 is posttran- 466 264 266
slationally removed)
Homologous to Amidase signature family Amidase signature family D-Aminopeptidase DppA b-Lactamases (class A)
2þ
(Zn -dependent self-com-
partmentalizing protease)
Isoelectric point (pH) — — — —
Optimum pH 7.0–9.5 8.5 9.0 7.2
Optimum temp. ( C) 47–49 40 85 35
Temp. stability ( C) 40 (30) 70 (30) 40 (10)
(preincubation time, min)
Activation by Co2 þ , Mn2 þ
Inhibitors Partly: Fe2 þ , DTNB. Partly: Fe3 þ , Sn2 þ , Cu2 þ , DTT, b-mercaptoethanol, Partly: Ag þ , Ni2 þ , Cu2 þ ,
Al3 þ , Ni2 þ . EDTA Pb2 þ .
Strongly: Hg2 þ , Ag þ , Cu2 þ , Strongly: As3 þ , Zn2 þ , Strongly: Hg2 þ
pCMB, PMSF Hg2 þ , Ag þ , Cd2, pCMB,
PMSF
Substrate specificityb)
Gly-NH2 — — — —
D-Ala-NH2 4.9 ND 100 20
D-Abu-NH2 — — — 21
D-Val-NH2 1.5 ND 15 0.99
D-Leu-NH2 45 46 38 26
D-Met-NH2 — 54 6.3 100
D-Norleu-NH2 — — 22 —
D-Norval-NH2 — 71 19 —
D-Phg-NH2 — — 17 2.3
D-Phe-NH2 100 100 17 50
D-Tyr-NH2 — 42 16 33
D-Trp-NH2 — 48 13 7.5
D-Pro-NH2 2.7 6.2 0.40 20
D-Glu-NH2 — ND — 6.4
D-Asp-NH2 — 0.40 0.20 0.99
D-Ser-NH2 — 4.1 — 1.7
D-Thr-NH2 — 23 — —
D-Gln-NH2 — ND 10 62
D-Lys-NH2 — 2.5 15 96
D-Lactic acid amide 6.6 — — —
D-Ala-pNa — — — —
Peptidase activity — Yes, D-Phe4, D-Phe3, D-Phe2, No No
D-Phe-L-Phe, L-Phe-D-Phe
a) Owing to its strong hydrophobicity, the native molecular mass could not be determined via gel permeation chromatography.
j599
600j 15 Hydrolysis of Amides
Starting from 5 M DL-alanine amide HCl, 2.5 M (220 g l1) D-alanine was obtained in
4.5 h on applying whole cells. D-a-Aminobutyric acid, D-methionine, D-norvaline, and
D-norleucine were prepared in a similar manner using whole cells or their cell-free
extracts. The synthesis of D-norvaline and D-norleucine of sufficient enantiomeric
purity, however, required pretreatment of the biocatalyst with EDTA to inhibit
L-specific or unspecific amide hydrolases endogenous to E. coli. The aminopeptidase
also catalyzed the stereoselective synthesis of D-alanine N-alkylamides 88 from an
amine and the amide or methyl ester of DL-alanine (Scheme 15.15) [270, 271]. Best
results were obtained in non-aqueous organic media with urethane pre-polymer PU-
6 immobilized enzyme. Using this system, 100 mM DL-alanine methyl ester 87 (acyl
donor) and 5 equiv. of 3-aminopentane (acyl acceptor), D-alanine-3-aminopentane
could be obtained in 45.2% yield and over 99% e.e. [271]. The PU-6 immobilized
D-aminopeptidase was also used for the synthesis of D-alanine oligopeptides in non-
aqueous media [272]. Starting from 250 mM D-alanine methyl ester HCl, (D-Ala)2 and
(D-Ala)3 were obtained in 58% and 6%, respectively, in water-saturated toluene with 3
equiv. of triethylamine. This organic base was essential for the peptide bond
formation, to deprotonate the D-alanine methyl ester to serve as nucleophile.
DAP, 30°C H
O + N + O
H 2N R NH 2 H 2N R H 2N
butyl acetate
O O O
DL-87 D- 88 L-87
(>99% e.e.)
R = 3-pentyl-, neopentyl-, benzyl-, n-butyl-
Scheme 15.15 D-Aminopeptidase catalyzed aminolysis reaction of DL-alanine methyl ester (87) as
acyl donor with different amines.
In comparison with the D-aminopeptidase from O. anthropi SCRC C1-38, the D-

stereospecific amino acid amidase (DAA) from O. anthropi SV3 displayed activity
towards a much broader range of amino acid amides [264, 273]. This enzyme
hydrolyzed the aromatic amino acid amides D-phenylalanine amide, D-tyrosine
amide, and D-tryptophan amide most efficiently, but also the amides of D-leucine,
D-alanine, D-methionine, D-norleucine, D-norvaline, and D-phenylglycine were con-
verted with moderate to good activity. L-Amino acid amides, aliphatic and aromatic
amides, peptides, and b-lactams, in contrast, were not hydrolyzed. In addition, Boc-
D-alanine amide was not a substrate, indicating the importance of a free a-NH2 group
for activity. This D-amino acid amidase was completely inhibited by Zn2 þ and
phenylmethanesulfonyl fluoride (PMSF), pointing to a catalytic serine residue in the
active site. Chelating compounds, thiol reagents, as well as b-lactams had no
significant effect on this enzyme. The daa gene was cloned and overexpressed in
E. coli under control of the lac promoter [273]. The primary structure of DAA showed
homology to that of the alkaline D-peptidase from Bacillus cereus DF4-B, the
DD-carboxypeptidase from Streptomyces R61, and other penicillin-recognizing
proteins. The active site motifs typical for this family of enzymes (S-X-X-K, Y-X-N
and H/K-X-G) are also perfectly conserved in the O. anthropi DAA. To increase this
enzymes industrial attractiveness, its thermostability was improved by a directed
evolution approach. In two rounds of error-prone PCR a double mutant (Lys278Met,
Glu303Val) was obtained that combined a 5 C increase in thermostability with a
threefold enhanced Vmax. E. coli cells expressing this improved DAA mutant
successfully resolved 1.0 M of DL-phenylalanine amide HCl in 2 h, whereas the same
amount of E. coli cells expressing the wild-type DAA produced only 300 mM
D-phenylalanine in the same reaction time [274].
Recently, the crystal structures of this D-amino acid amidase in native form and in

complex with D-phenylalanine were determined at 2.1 and 2.4 A resolution, respec-
tively [275]. As expected, the structure is very similar to that of the penicillin-
recognizing proteins, especially of the DD-carboxypeptidase from Streptomyces R61
and the class C b-lactamase from Enterobacter cloacae GC1, but also to that of the
D-aminopeptidase (DAP) from O. anthropi SCRC C1-38. Notwithstanding this
structural similarity, DAA does not display the transpeptidase and carboxypeptidase
activities typical for DD-carboxypeptidases, which was attributed to steric hindrance of
the hydrophobic substrate-binding pocket by a loop consisting of residues 278–290
and the V-loop, respectively [275]. Based on the geometry of the hydrogen bonds
between the active site residues Tyr149 Og, Ser60 Oc, and Lys63 Nf in the substrate-
bound as well as in the ground state, a catalytic mechanism of DAA activity has been
proposed [276]. In 2008, the crystal structures of DAA in complex with L-phenylal-

anine and with L-phenylalanine amide were also solved, at 2.3 and 2.2 A, respective-
ly [277]. Comparison of these crystal structures with that of DAA in complex with
D-phenylalanine [275] revealed three structural features that together produce an
environment in which only D-phenylalanine amide can bind in a productive orien-
tation (i.e., with its carboxyl group near the catalytic nucleophile Ser60 Oc), which
explains this enzymes high D-stereospecificity [277].
About a decade after the first D-selective amino acid amide hydrolyzing enzymes
were described by Asano [263, 264], a search for novel D-alanine-p-nitroanilide
hydrolyzing enzymes led to the isolation of two other intracellular D-aminopeptidases
from O. anthropi LMG7991 [278]. The first enzyme (DmpB) was purified to 90%
homogeneity from the wild-type strain, and its NH2-terminal amino acid sequence
was determined. This exhibited nearly 60% identity with the N-terminus of DAP
from O. anthropi SCRC C1-38, suggesting similar properties, which was supported by
the cross reactivity of DmpB with rabbit anti-DAP antibodies [278]. The second
enzyme, DmpA [279], hydrolyzed the p-nitroanilide, amide, and ester derivatives of
glycine and D-alanine more efficiently than that of L-alanine [280]. The p-nitroanilide
derivatives of the larger amino acids phenylalanine and leucine, on the other hand,
were hydrolyzed L-selectively. This reversal of stereoselectivity was also observed
when regular peptides were used as substrate – with these substrates DmpA behaves
as an aminopeptidase with a preference for N-terminal residues in an L-configura-
tion. The best peptide substrate tested was the tripeptide L-Ala-Gly-Gly, which was
first hydrolyzed into L-Ala and the dipeptide Gly-Gly [280]. Acylation of the free
a-NH2-group of the substrates severely impaired DmpAs activity. This unique
enzyme was thus described as an L-aminopeptidase D-alanine-esterase/amidase [279].
602
The O. anthropi dmpA gene was cloned and sequenced and actively expressed in
E. coli under the control of the vector borne lac promoter [278]. It encodes a 375-
residue inactive polypeptide that partly accumulated in E. coli in inclusion bodies.
This inactive precursor is turned into the active protein by a two-step maturation
process, that is, the removal of the N-terminal methionine residue and the cleavage of
the Gly249-Ser250 peptide bond. The active DmpA is thus a heterodimer composed
of an a-chain (residues 2–249) and a b-chain (residues 250–375). By mutagenesis of
Gly249 and Ser250 it was shown that both residues were essential for maturation of
this enzyme; all mutants tested were produced in E. coli as an inactive, non-cleaved
precursor only. Because the processing of the precursor occurred both in O. anthropi
and E. coli it has been hypothesized that this is an autocatalytic splicing process [280].
Computer-assisted secondary structure prediction showed that the N-terminal serine
residue of the b-chain (Ser250) is located at the N-terminus of a b-sheet. In
combination with the highly hydrophobic nature of the Ser250-Lys267 peptide, this
pointed to DmpA as the first representative of a novel subfamily of the N-terminal
nucleophile (Ntn) hydrolase superfamily [280]. The crystal structure of this hydrolase

at 1.82 A later confirmed this hypothesis [281, 282]. DmpA consists of four identical
heterodimeric subunits that are grouped into a donut-shaped molecule. Each of these
four heterodimers contains a central motif consisting of an abba sandwich in which
two stacked mixed b sheets are flanked on both sides by two a helices. This spatial
arrangement of DmpA shows clear homology to the structure of other members of
the Ntn hydrolase family, although the direction and connectivity of DmpAs
secondary structure elements differ significantly from the consensus Ntn hydrolase
fold [281]. Because DmpAs active-site residues are all functionally equivalent to the
corresponding residues in other Ntn hydrolases, it likely employs the same catalytic
mechanism. This includes the bifunctional role of the active site serine residue
(Ser250), which not only acts as the catalytic nucleophile but also enhances the
nucleophilicity of its hydroxyl group via its a-amino group [281].
Because the specific activity of DmpA for the substrates mentioned above was
moderate to low without exception (the best peptide substrate L-Ala-Gly-Gly, for
instance, was hydrolyzed with an initial rate of only 1.25 mmol per min per mg of
protein) it has been questioned whether these compounds are the natural substrates
of this enzyme. A broad evaluation of the substrate spectrum of a homologous
enzyme from Pseudomonas sp. MCI3434 (BapA, 43% sequence identity), a few years
later, revealed that dipeptides with a b-alanine residue at their amino terminus as well
as b-alanine amide were hydrolyzed much more efficiently than D-alanine-p-nitroa-
nilide [82]. BapA was thus named b-Ala-Xaa dipeptidase instead of L-aminopeptidase
D-alanine-esterase/amidase. More about this enzyme can be found in Section 15.4.3.
In line with this finding, DmpA was later found to cleave peptides with an N-terminal
b-hGly and L-b3-hAla residue much more efficiently than a-peptides
(Section 15.4.3) [283].
Over the years, D-a-H-a-amino acid amide hydrolyzing enzymes have also been
isolated from microorganisms other than O. anthropi. One of the first of these
enzymes was a D-alanine amidase, which was isolated form an Arthrobacter sp. by
Kyowa Hakko Kogyo scientists via an enrichment strategy based on a medium with
DL-alanine amide as sole nitrogen source and D-cycloserine to inhibit alanine
racemase [284]. This enzyme, which was highly active toward D-alanine amide
[specific activity 1380 mmol min1 mg1] was characterized by an extremely narrow
substrate specificity. Next to D-alanine amide, only glycine amide was hydrolyzed with
high activity. The enzyme, furthermore, hydrolyzed seven other amides with poor
activity (<2% relative to D-alanine amide), and was completely inactive toward D-and
L-alanine-containing dipeptides and esters. Although the cloning and heterologous
expression of the gene for this enzyme has not been reported to date, an efficient
process for the production of D-alanine employing wild-type whole cells could be
developed. By optimizing the pH and temperature of the reaction, cellular alanine
racemase activity could be sufficiently reduced. Because the D-alanine amidase did
not suffer from substrate nor product inhibition, high substrate concentrations were
possible that efficiently inhibited the cellular L-amidase. Using 10 g l1 of wet cells,
210 g l1 (2.4 M) of DL-alanine amide was completely resolved in 5 h, giving 1.2 M of
D-alanine with a more than 99% e.e. [284].
Another interesting D-amidase was identified by Kula and coworkers. Applying an
enrichment procedure with DL-tert-leucine amide as sole nitrogen source, three
strains were isolated that hydrolyzed this very bulky racemic amide D-stereoselec-
tively [285]. The amidase from Variovorax paradoxus DSM 14468 was further studied
because the crude extract of this strain displayed the highest specific activity. This
enzyme was purified by three chromatographic steps to near homogeneity in 25%
yield. DL-tert-Leucine amide was converted by this amidase with standard Michae-
lis–Menten kinetics with a Km of 0.74 mM and a Vmax of 1.4 U mg1. Furthermore,
slight substrate inhibition was observed (Ki of 640 mM). The V. paradoxus D-amidase
not only hydrolyzed a broad range of a-amino acid amides but also carboxamides and
a-hydroxy acid amides. Highest activity was observed for D-phenylalanine amide,
which was converted with 890-fold higher activity than DL-tert-leucine amide. The
relative low activity for this latter substrate was most likely caused by its extensive
branching at the Cb atom. Most substrates were hydrolyzed with a preference for
the D-enantiomer, but the degree of D-stereospecificity was highly dependent on the
substrate. DL-tert-Leucine amide, for instance, was hydrolyzed almost completely
D-selectively, leading to an e.e. of over 99% at 47% conversion (E > 200). Similar
results were obtained in the resolution of the amides of DL-leucine and DL-valine [285,
286]. DL-Phenylalanine amide, on the other hand, was converted with an E-ratio of 5
only, resulting in an e.e. of 50.9% at 51% conversion, whereas both enantiomers of
lactic acid amide were hydrolyzed with near equal rates, and L-proline amide was
converted even faster than its optical antipode [285].
The gene encoding this D-amidase was cloned and functionally expressed in
E. coli [287]. Although approximately 80% of the amidase protein resided in the
insoluble fraction, the expression of this enzyme in E. coli was 120-fold higher than in
V. paradoxus. Comparison of the primary structure of this enzyme with the protein
database showed that this D-amidase belongs to the amidase signature (AS) family of
enzymes [25]. In line with the assignment of the V. paradoxus D-amidase in this
enzyme family, the residues forming the active site catalytic triad (Ser178, Ser154,
and Lys81) were found to be fully conserved.
604 j 15 Hydrolysis of Amides
Escherichia coli cells expressing this D-amidase under control of the rhamnose
inducible promoter were applied in a biocatalytic cascade [288]. By combining these
cells with E. coli cells expressing the non-selective NHase from Rhodococcus erythro-
polis 870-AN019, DL-tert-leucine nitrile 89 was converted in two sequential conver-
sions into D-tert-leucine 60 (Scheme 15.16). After optimization of the reaction
conditions, including the ratio of the two E. coli cells to compensate for the different
specific activities of the two biocatalysts, 700 mg of D-tert-leucine was produced from
the corresponding nitrile in a fed-batch process in 4.5 h, along with unconverted
nitrile 89 and L-amide 90. The accumulation of tert-leucine nitrile 89 was caused by a
reduction of its conversion due to the inhibition of the NHase by the build-up of the
L-tert-leucine amide 90 in the reactor [288].
NHase D-amidase OH + NH2

NH2 H2 N H 2N
H 2N H 2N
N H2 O H2 O NH3 O O
O
DL-89 DL-90 D -60 L-90
Scheme 15.16 Nitrile hydratase/D-amidase catalyzed cascade for the production of D-tert-leucine
(60) from the corresponding racemic nitrile 89 [288].
A novel D-stereospecific amino acid amidase with high homology to the

V. paradoxus D-amidase (67.9% identity) was reported [289]. This amidase was
purified from the bacterium Delftia acidovorans, which was isolated from soil by
applying an enrichment strategy with D-phenylalanine amide as sole nitrogen source.
The enzyme catalyzed the hydrolysis of a broad range of D-amino acid amides with a
preference for substrates with a long-chain aliphatic or aromatic side-chain, like
D-phenylalanine amide, D-tryptophan amide, D-norvaline amide, and D-leucine
amide. Small aliphatic amino acid amides, like D-alanine amide and glycine amide,
and oligopeptides were not hydrolyzed. Whereas strict D-stereoselectivity was
observed toward D-tryptophan amide, D-serine amide, D-proline amide, and D-lysine
amide, the amides of phenylalanine, norvaline, methionine, leucine, tyrosine, and
isoleucine were converted with moderate to low stereoselectivity only. The gene
encoding this enzyme was cloned through an activity-based screening of a
D. acidovorans expression library in E. coli. It coded for a 466 amino acid protein,
with a molecular mass of 49 860 Da, which also belonged to the amidase signature
(AS) family [25]. Table 15.4 gives more features of this D-amino acid amidase.
A few years later a much more D-selective amino acid amidase (DaaABi) was
identified in the bacterium Brevibacterium iodinum TPU 5850 [290]. The amidase was
purified form an E. coli transformant expressing the daaABi gene, and characterized.
This amidase displayed activity toward a broad range of D-amino acid amides
including D-phenylalanine amide, D-methionine amide, D-lysine amide, D-glutamine
amide, D-tyrosine amide, D-arginine amide, D-leucine amide, D-a-aminobutyric acid
amide, D-proline amide, and D-alanine amide. Interestingly, all L-amino acid amides
tested were not converted by this enzyme, demonstrating its near absolute
D-stereoselectivity. Dipeptides, tripeptides, and b-lactam compounds also were not
hydrolyzed. The practicability of DaaABi for the production of D-amino acid amides
was demonstrated by the kinetic resolution of DL-phenylalanine amide (180 mM).
After 3 h of reaction when the conversion of the amide had reached 49%, still no
L-phenylalanine could be detected, and D-phenylalanine of over 99.5% e.e. was
formed [290]. Homology searches revealed that this B. iodinum D-amino acid amidase
belonged to the class A b-lactamases. This relationship, however, is quite remote,
because DaaABi neither exhibited b-lactamase activity toward b-lactam compounds
like ampicillin and benzylpenicillin nor suffered inhibition by clavulanic acid, an
inhibitor specific for class A b-lactamases.
Two highly thermostable D-amino acid amidases have been identified in the
thermophile Brevibacillus borstelensis BCS-1. The first, a D-methionine amidase,
efficiently hydrolyzed a broad range of D-amino acid amides with highest activity
toward the amides of D-methionine (100%), D-norvaline (78%), D-norleucine (77%),
D-lysine (60%), and D-leucine (59%) [291]. The enzyme also displayed activity toward
different D-amino acid esters [e.g., D-alanine methyl ester (25%) and D-alanine benzyl
ester (22%)] and arylamides [e.g., D-alanine-b-naphthylamide (33%) and D-leucine-p-
nitroanilide (40%)]. These substrates were hydrolyzed D-stereoselectively without
exception. D-Amino acid containing di- and oligopeptides, N-acetyl-D-amino acids,
and simple amides (like acetamide, propionamide and benzamide) were not con-
verted. Using this enzyme, D-phenylalanine with 97.1% e.e. was produced from the
racemic amide (E ¼ 196). The gene encoding this D-methionine amidase has not
been reported yet, although the NH2-terminal amino acid sequence was already
published in 2003 [291].
The second D-amino acid amidase isolated from B. borstelensis BCS-1 was a
thermostable D-alanine amidase (BDA). Although the characteristics of this BDA,
like molecular mass, subunit structure, and effects of metal ions and inhibitors, were
generally similar to those of the D-methionine amidase mentioned above it is clearly
another enzyme based on differences in, for instance, pH optimum, thermostability,
and, most pronounced, substrate preference. This D-alanine amidase displayed
highest activity towards D-alanine amide (activity for D-methionine amide was only
6.3% of that) but other aromatic, aliphatic, and branched chain amino acid amides
were also efficiently hydrolyzed D-stereoselectively [292]. Most L-amino acid amides,
N-terminally protected amides and esters, simple amides, and D-amino acid contain-
ing peptides were not substrates for this enzyme. The optimum pH and temperature
for this amidase were 9.0 and 85 C, respectively. The enzyme retained full activity
after 30 min incubation up to 70 C. BDA was inhibited by dithiothreitol,
b-mercaptoethanol, and the metal chelator EDTA, and strongly activated by Co2 þ
and Mn2 þ , pointing to a critical role for a divalent metal ion in the catalytic
mechanism. The BDA encoding gene has been cloned and expressed in E. coli
under control of the high-level constitutive expression (HCE) promoter. Using the
cell extract of this recombinant E. coli strain, 0.2 M DL-phenylalanine amide was
successfully resolved, yielding D-phenylalanine with more than 99.0% e.e. at 97%
conversion (E > 500) [292].
606
Homology studies showed that the primary structure of the BDA from
B. borstelensis BCS-1 exhibited strong similarity with bacterial dipeptide ABC
transporter proteins. Of the other D-amino acid-specific enzymes known, BDA
only displayed sequence similarity with DppA from Bacillus subtilis (26% identity
over 264 amino acids), including conservation of the N-terminal motif SXDXEG
(vide infra) [292]. B. subtilis dppA, which belongs to the dipeptide ABC transport
(dpp) operon expressed early during sporulation, encodes a binuclear zinc-dependent
aminopeptidase that hydrolyzes D-alanine-p-nitroanilide with high activity [293].
The catalytic efficiency (kcat/Km) of this self-compartmentalized D-aminopeptidase
for its best peptide substrates, (D-Ala)2 and D-Ala-Gly-Gly, is significantly lower

than for D-alanine-p-nitroanilide. Elucidation of the X-ray structure at 2.4 A resolution
revealed that the B. subtilis DppA has a unique tertiary structure that is organized as a

barrel-shaped decamer with identical 30 kDa subunits [294]. A 20 A wide channel runs
through this complex, giving access to a central cavity holding the ten active sites, a
spatial organization that serves as a molecular sieve that protects larger potential
substrates from unwanted hydrolysis. The near N-terminal motif SXDXEG,
which is fully conserved in almost all DppA homologs, encompasses two of the
Zn2 þ -coordinating residues. Because hydrolysis of (D-)amino acid amides by DppA
has not been reported, more information on this enzyme is not given here.
A last amidase worth mentioning in this section is the (R)-stereoselective amidase
from Pseudomonas sp. MCI3434. This strain was identified by Asano and coworkers
in a screening for microorganisms that can hydrolyze piperazine-2-tert-butylcarbox-
amide (35,Scheme 15.13) [81]. (S)-Piperazine-2-tert-carboxamide is an important
chiral building block for pharmacologically active compounds, like the HIV protease
inhibitor indinavir (73) (Section 15.4.1.1). Although the Pseudomonas sp. MCI3434
cells hydrolyzed both the (R)- and the (S)-piperazine-2-tert-butylcarboxamide (35),
they displayed a clear preference for the (R)-carboxamide. Purification of the
main amidase from this strain showed that this enzyme is strictly (R)-selective,
suggesting the presence in these cells of another amidase with activity for (S)-
piperazine-2-tert-carboxamide.
Based on the N-terminal amino acid sequence, the gene encoding this (R)-amidase
(ramA) was cloned from the genomic DNA of Pseudomonas sp. MCI3434. It encodes a
protein of 274 amino acids with a calculated molecular weight of 30 128 and
significant homology to the carbon-nitrogen hydrolase family of enzymes (nitrilase
superfamily), including the typical catalytic triad consisting of Glu40, Lys108, and
Cys140. The ramA gene was efficiently expressed in E. coli JM109 under control of
the lac promoter and with an optimized ribosome-binding site (RBS), leading
to a 30 000 times higher activity than in Pseudomonas sp. MCI3434. These recom-
binant E. coli cells were successfully applied in the resolution of racemic piperazine-2-
tert-butylcarboxamide (35); (R)-piperazine-2-carboxylic acid of e.e. >99.5% was
produced throughout the reaction.
RamA displayed rather narrow substrate specificity, with activity toward carbox-
amide compounds with an amino or imino group connected to a a- or b-carbon
only. Examples of substrates converted are piperazine-2-tert-butylcarboxamide (35)
(relative activity 9.0%), piperazine-2-carboxamide (34) (rel. act. 100%), piperidine-
3-carboxamide (37, nipecotic acid amide) (rel. act. 68.9%), b-alaninamide (rel. act.
108%), and D-glutaminamide (rel. act. 27.0%), which is converted into D-glutamic
acid amide instead of D-glutamine. Other a-amino acid amides, peptides, aliphatic
amides, aromatic amides, and nitriles were not converted by RamA [81].
15.4.2
Synthesis of Enantiopure a,a-Disubstituted Amino Acids
Inspired by the advantages of the amidase process for a-H-a-amino acids, DSM
scientists developed a similar enzymatic kinetic resolution process for the production
of enantiopure a,a-disubstituted amino acids 92. Although alternative routes to the
racemic disubstituted amino acid amides 91 have been described [295–297], Strecker
synthesis is also in this case the most direct way to prepare these amide substrates.
The hydrolysis of the aminonitrile intermediate, however, needs harsher conditions
in this case (e.g., benzaldehyde/pH 14, conc. H2SO4 or HCl-saturated formic acid)
because of the increased steric hindrance [298]. Because the P. putida L-aminopep-
tidase requires substrates with an a-hydrogen atom for activity, a novel amidase
biocatalysts was identified in a screening program. This biocatalyst, Mycobacterium
neoaurum ATCC 25795, affords the (S)-a,a-disubstituted amino acids 92 and the
corresponding (R)-amides 91 in almost 100% e.e. at 50% conversion for most
a-methyl-substituted compounds tested (E > 200) (Scheme 15.17) [299, 300]. Only
for glycine amides with two small substituents at the chiral center is the enantios-
electivity moderate (E 15). Generally, a-methyl-substituted amino acid amides are
hydrolyzed with high activity, but increasing the size of the smallest substituent to
ethyl, propyl, or allyl dramatically reduces the activity, especially if the larger
substituent does not contain a -CH2- spacer at the chiral carbon atom [299]. In
addition to a,a-disubstituted amino acid amides 91, their a-hydrogen containing
counterparts are also good substrates that are hydrolyzed enantioselectively. Dipep-
tides, in contrast, are not hydrolyzed. Based on the resolution of numerous sub-
strates, a schematic model of the M. neoaurum amidase active site has been
proposed [299].
R1 R2 Mycobacterium neoaurum R2 R1 R2 R1
NH 2 OH NH 2
H2 N H 2N H2 N
pH ~ 8.0–8.5
O O O
91 (S)-92 (R)-91
R 1 = broad
R 2 = H, CH3 , CH2 CH3 , CH2 CH=CH 2
Scheme 15.17 Resolution of a,a-disubstituted amino acid amides 91 by whole cells of

Mycobacterium neoaurum ATCC 25795.
608
H 2N Me H2 N Me H2 N Me H 2N Me
Me NH2 NH 2 NH 2 NH 2
Me O O Me O O
Me Me
95 96 97 98
Me
H 2N Me H2 N H2 N Me H 2N Me
NH2 NH 2 NH 2 NH 2
O O O O
99 100 101 102
Figure 15.5 Examples of a,a-disubstituted a-amino acid amides that were successfully resolved
using the L-amino amidase from M. neoaurum ATCC 25795 and/or the L-amidase from O. anthropi
NCIMB 40321.
The M. neoaurum biocatalyst can be used on preparative scale as freeze-dried whole

cells without further purification of the fermented cells. After completion of the
resolution reaction, which is typically performed at 2–10 wt% amide concentration
(depending on the solubility) in an aqueous solution at pH 8–8.5 and 37 C [299], and
removal of the biomass by centrifugation, the (S)-amino acid product and (R)-amide
substrate can be separated by, for instance, extraction or by use of an ion-exchange
column [299, 300].
The enzyme from M. neoaurum ATCC 25795 responsible for the enantioselective
hydrolysis of DL-a-methyl valine amide 95 (Figure 15.5) has been purified from the
crude extract, by a procedure including ammonium sulfate fractionation, gel
filtration, and anion exchange chromatography, prior to its characterization [301].
This enzyme, which was classified as an amino amidase based on its substrate
spectrum, is active toward a broad range of a-H- and a-alkyl-substituted amino acid
amides, with the highest activity for the cyclic amino acid amide DL-proline amide
and the a-alkyl-substituted amino acid amides DL-isovaline amide 97 and DL-
a-allylalanine amide 98. The amino acid amides are hydrolyzed with moderate to
high L-selectivity, the lowest enantioselectivity being obtained toward alanine amide
(E 25). This amino amidase also converted the aliphatic amides acetamide and
propionamide, albeit with modest activity. Dialkyl amino acid amides with very
bulky substituents, a-hydroxy acid amides, and dipeptides, in contrast, were not
hydrolyzed by this amidase [301]. Cloning of the gene encoding this L-selective
amino amidase from M. neoaurum has not been reported yet. Table 15.5 gives more
features of this enzyme.
To further broaden the range of compounds that can be resolved by the amidase
technology to very bulky a,a-dialkyl-substituted amino acid amides 91, a-hydroxy
acid amides 93, and N-hydroxyamino acid amides 94, a novel amidase biocatalyst was
identified applying an enrichment strategy with DL-mandelic acid amide (26) as sole
Table 15.5 Properties of microbial L-selective a-alkyl-a-amino acid amide hydrolases.
Mycobacterium neoaurum ATCC 25795 Ochrobactrum anthropi NCIMB 40321 Xanthobacter flavus NR303
Reference [301] [309, 310] [315]

Name L-Amino amidase L-Amidase (LamA) L-Amidase (McaA)
Subunit molecular weight 40 kDa (SDS-PAGE) 33 870 (calculated) 38 555 (calculated)
36 kDa (SDS-PAGE) 40 kDa (SDS-PAGE)
No. of subunits 3–4 (gel filtration – 136 kDa) 2 (native electrophoresis – 65 kDa)
No. of amino acids 313 (Nb. Met-1 is posttranslationally 355
removed)
Homologous to Acetamidase/formamidase family Acetamidase/formamidase family
Isoelectric point (pH) 4.2 (IEF) 5.4 (IEF)
Optimum pH 8.0–9.5 6.0–8.5 7.0–8.0
pH stability
Optimum temp. ( C) 50 70 55
Temp. stability ( C) (preincuba- 55 (30) 60 (60)
tion time, min)
Activation by DFP Mn2 þ , Mg2 þ , Zn2 þ Mn2 þ
Inhibitors Completely: DTT. Completely: EDTA.
Partly: iodoacetamide, 1,10- Partly: 1,10-phenanthroline
phenanthroline
Substrate specificitya)
Acetamide 2.3 ND No quantitative analysis
publishedb)
Propionamide 3.5 ND
Gly-NH2 — 40
DL-Ala-NH2 — 76
DL-Val-NH2 37 10
(Continued )
j609
610
Mycobacterium neoaurum ATCC 25795 Ochrobactrum anthropi NCIMB 40321 Xanthobacter flavus NR303
DL-Leu-NH2 — 29
DL-Met-NH2 — 92
DL-Phg-NH2 — 71
DL-Phe-NH2 14 98
DL-Pro-NH2 85 100
DL-Glu-NH2 — 0.73
DL-tert-Leu-NH2 — 0.10
DL-a-Me-Val-NH2 (95) 27 0.4
DL-a-Me-Leu-NH2 (96) 52 —
DL-a-Me-Abu-NH2 (97) 67 —
DL-a-Allyl-Ala-NH2 (98) 100 —
DL-a-Me-Phg-NH2 (99) 10 1.4
DL-a-Et-Phg-NH2 (100) ND 2.3
DL-a-Me-Phe-NH2 (101) 25 9.8
DL-a-Me-homo-Phe-NH2 (102) 8.3 —
DL-Mandelic acid amide (26) ND 1.9
Peptidase activity No No
a) Substrate specificity is given relative to the activity for the substrate that is converted most efficiently. ND: not detected; –: not measured.
b) X. flavusL-amidase has activity toward valine amide, phenylglycine amide, tert-leucine amide, phenylalanine amide, a-aminobutyric acid amide, and a-methylcysteine
amide; it was inactive toward propionamide and butyramide.
R1 R2 Ochrobactrum anthropi R2 R1 R2 R1
NH 2 OH NH 2
X X X
pH 5.5–8.5
O O O
(S)-acid (R)-amide
91 X = NH 2
93 X = OH R 1 = alkyl, aryl
94 X = NH-OH R 2 = H, alkyl
Scheme 15.18 Resolution of different types of a-substituted amides by Ochrobactrum anthropi

NCIMB 40321.
nitrogen source (Scheme 15.18) [302]. This procedure resulted in the isolation of an
Ochrobactrum anthropi strain that was deposited at the NCIMB culture collection
(NCIMB 40321). Besides its extremely broad substrate specificity, this novel whole-
cell biocatalyst is characterized by an excellent enantioselectivity, good temperature,
salt and solvent stability [303], and especially a relaxed pH profile. Although the
amidase displays its highest activity at pH 8.5, 55% of this activity is retained at pH
5.0. This property makes the O. anthropi amidase biocatalyst very useful for the
resolution of hydrophobic amino acid amides that are only very poorly soluble at the
weakly alkaline conditions needed for the M. neoaurum amino amidase to be active.
Simply by performing the hydrolysis reaction at slightly acidic conditions the
solubility of the amide substrates increases due to the presence of the protonizable
amino group. This feature has been employed in the resolution of the racemic threo-
phenylserine amides 103a/b, which are intermediates in a novel route to
the antibiotics thiamphenicol (104a) and florfenicol (104b), of which only the
(1R,2R)-enantiomers display the required biological activity [304]. In this case, the
O. anthropi amidase-catalyzed resolution reaction was performed at pH 5.6–6.0 to
ensure a fair solubility of the otherwise insoluble amides. The (2S,3R)-phenylserines
were obtained in >99% e.e. and chemical yields up to 50% [304]. Another example
can be found in the preparation of (S)-a-methyl-(3,4-dichlorophenyl)alanine (105)
from the corresponding racemic amide. This is a useful precursor for the synthesis of
cericlamineHCl (106), a potent and selective synaptosomal 5-hydroxytryptamine
(serotonin) uptake inhibitor [305, 306]. Because this amide is nearly insoluble at
slightly alkaline and neutral conditions, the activity of the O. anthropi amidase at low
pH (in this case pH 5.3) was also in this case of decisive importance [307]. Yet another
example was the application of the O. anthropiL-amidase in the resolution of racemic
1-naphthylglycine amide (107) [308]. Although this substrate could be successfully
resolved with the L-aminopeptidase from P. putida ATCC 12633 at pH 8.3 and 37 C,
the low amide-solubility under these conditions caused the resolution to proceed very
slowly. The solubility problem was overcome by turning to O. anthropiL-amidase as
biocatalyst, which enabled execution of this resolution reaction at pH 6.5 and 50 C.
By using this amidase both the (R)-amide and the (S)-acid were obtained in over 90%
isolated yield and greater than 98% e.e. The (R)-amide 107 was subsequently
hydrolyzed under mild conditions into the corresponding (R)-acid with the non-
selective amidase from Rhodococcus erythropolis NCIMB 11540 (see Scheme 15.14 for
612
similar approach). In total, the O. anthropiL-amidase has a unique set of properties for
application in the fine-chemicals industry.
OH
OH O
Y
O
NH 2 HN CHCl2
S
NH 2 Me
X O O
103a X = SMe 104a Y = OH : Thiamphenicol
103b X = SO 2Me 104b Y = F : Florfenicol
Cl Cl
Cl Cl
Me Me
OH Me OH NH2
H 2N N H 2N
O Me O
105 106 Cericlamine 107
The most important L-amidase in O. anthropi NCIMB 40321 was purified from the
cell-free extract by ammonium sulfate fractionation, anion-exchange chromatogra-
phy, gel filtration, and hydrophobic interaction chromatography [309]. This L-
amidase, which was named LamA, converts a broad range of a-hydrogen- and
(bulky) a,a-disubstituted a-amino acid amides. Moreover, also mandelic acid amide
(26) (an a-hydroxy acid amide) and N-hydroxyphenylalanine amide (an a-N-hydro-
xyamino acid amide) are hydrolyzed by LamA, which is thus responsible on its own
for the extremely broad substrate specificity of the O. anthropi whole cells. Simple
aliphatic amides, b-amino and b-hydroxy acid amides, and dipeptides are not
substrates for LamA. This enzymes broad substrate specificity does not come at
the expense of its enantioselectivity; of all racemic substrates tested, only the L-
enantiomer was hydrolyzed (E > 150). O. anthropi LamA is a metalloenzyme as it is
strongly inhibited by the metal-chelating compounds EDTA and 1,10-phenanthro-
line. The activity of the EDTA-treated enzyme could be restored by the addition of
Zn2 þ (to 80%), Mn2 þ (to 400%), and Mg2 þ (to 560%). The gene encoding this L-
amidase was cloned via reverse genetics and efficiently expressed in E. coli under
control of the trc or aro promoter [309, 310]. It encodes a polypeptide of 314 amino
acids with clear homology to the acetamidase/formamidase family of proteins,
including the stereoselective amidases from Enterobacter cloacae N-7901 [311], Ther-
mus sp. 0-3-1 [312], and Klebsiella oxytoca PRS1 [61]. The Enterobacter and Thermus
amidase genes, which were isolated by Mitsubishi researchers, encode amidases of
315 and 309 amino acids, respectively, which are 67 and 53% identical to O. anthropi
LamA. Both amidases are highly L-selective toward DL-tert-leucine amide and are also
active toward lactate amide (28), implying that these amidases can also convert
a-hydroxy acid amides. The amidase from K. oxytoca (328 amino acid residues and
28% identity to O. anthropi LamA) has been developed by Lonza AG for the (R)-
selective hydrolysis of racemic 3,3,3-trifluoro-2-hydroxy-2-methylpropionamide
(108) [313]. More information on this enzyme and biocatalytic process can be found
in Section 15.5. The O. anthropiL-amidase also shares moderate but significant
sequence identity (25–26%) with the formamidases from Methylophilus methylotro-
phus [142, 143] and Aspergillus nidulans [314] and the acetamidase from Mycobacterium
smegmatis [64, 144]. These amidases are characterized by a very narrow substrate
specificity that is restricted to short-chain aliphatic amides like formamide, acetamide,
and propionamide, which are exactly the substrates O. anthropi LamA cannot convert.
Asano and coworkers reported on the identification of a novel L-amidase for the
resolution of DL-a-methylcysteine amide in Xanthomonas flavus NR303 [315]. This
intracellular enzyme (named McaA) was purified to near homogeneity and its gene
was cloned by reverse genetics. The mcaA gene encodes a protein of 355 amino acids
with a calculated molecular mass of 38 555. Like O. anthropi LamA, this L-amidase has
clear homology to the acetamidase/formamidase protein family, although the
homology with LamA is only moderate (33% sequence identity over 297 amino
acids). X. flavus McaA was expressed in E. coli JM109 under the control of the lac
promoter, and the whole cells were tested in the resolution of different carboxamides.
McaA displayed activity toward different a-H-a-amino acid amides, which were
resolved L-selectively, but not for the aliphatic amides propionamide and butyramide.
Using the recombinant E. coli cells, L-a-methylcysteine was obtained from the
corresponding amide on a gram scale in 40% isolated yield and >98% e.e. [315].
15.4.3
Synthesis of Enantiopure b-Amino Acids by b-Aminopeptidases
One of the interesting features of b-peptides (peptides built from b-amino acids with
an additional methylene group in their backbone relative to a-amino acids, Fig-
ure 15.6) is their much better resistance towards proteolytic cleavage in the digestive
system as compared to their a-amino acid containing analogs. This limited biode-
gradability, which translates into an increased half-life and improved bioavailability,
combined with a higher structural diversity and folding into well-defined secondary
structures means that b-peptides have great potential as peptidomimetics for
pharmaceutical applications [316–319].
The interest in b-peptides has triggered the development of numerous approaches
for the synthesis of optically pure b-amino acids. Most of these methods are chemical
in nature [319–321], but enzymatic methods applying, for instance, a lipase [322, 323],
R R
OH H 2N OH H 2N OH
H 2N
O R O O
α-amino acid β3 -amino acid β2-amino acid
Figure 15.6 Structures of mono-substituted a- and b-amino acids.

614
aminoacylase [324], penicillin G acylase [325], peptide deformylase [9], and amino-
mutase [326] have also been described (for a review see Reference [327]).
An enzyme with activity toward a b-amino acid amide was described in 2005 by
Komeda and Asano [82]. This enzyme was purified from Pseudomonas sp. MCI3434, the
bacterium from which also an amidase acting (R)-stereoselectively on piperazine-2-
tert-butylcarboxamide (35) and nipecotic acid amide (37) (RamA) had been isolated [81].
In the region upstream of ramA an ORF designated bapA was found that encoded a
protein of 366 amino acids sharing 43% sequence identity to the L-aminopeptidase
D-Ala-esterase/amidase DmpA from O. anthropi LMG7991 (Section 15.4.1.3). BapA
was heterologously produced in E. coli JM109 cells by expressing the bapA gene under
control of the lac promoter on vector pUC19, and purified to near homogeneity. Like O.
anthropi DmpA, this enzyme was formed as a propeptide that was converted into the
mature heterodimeric enzyme by autocatalytic cleavage of the Gly238-Ser239 peptide
bond. Investigation of the substrate specificity of BapA revealed that this enzyme
hydrolyzed D-Ala-pNA more efficiently than L-Ala-pNA, but still with moderate activity
only (7.7 U mg1). Much higher activities were observed for dipeptides containing
b-alanine (also called b-homoglycine (H-b-hGly-OH)) at their N-terminus, including
b-Ala-L-Ala, b-Ala-Gly, b-Ala-L-His (L-carnosine), b-Ala-L-Leu and (b-Ala)2, and b-Ala-
NH2. BapA was therefore called b-Ala-Xaa dipeptidase (EC 3.4.13.-) [82]. The enantios-
electivity of BapA for b-amino acid amides has not been reported yet. Table 15.6 gives
more features of BapA.
Recently, the enzymatic kinetic resolution of b-amino acid amides was described
for the first time by Heck and colleagues [328]. Next to the O. anthropi DmpA they
employed a b-peptidyl aminopeptidase from two different Sphingomonadaceae for the
resolution of four aliphatic b3-amino acid amides, that is, BapA from Sphingosinicella
xenopeptidilytica 3-2W4 and BapA from Sphingosinicella microcystinivorans Y2 (3-2W4
BapA and Y2 BapA, respectively) (Scheme 15.19). Strain 3-2W4 was originally
isolated from material from a wastewater treatment plant via enrichment on the
b-tripeptide H-b-hVal-b-hAla-b-hLeu-OH and b-dipeptide H-b-hAla-b-hLeu-OH as
the sole carbon and nitrogen source [329]. Later it was shown that the closest
phylogenetic homolog of strain 3-2W4, S. microcystinivorans Y2, is also able to utilize
both b-peptides as the sole carbon, energy, and nitrogen source [330]. The b-peptidyl
aminopeptidases from both strains were purified and characterized, and their genes
H2 N NH 2 β -aminopeptidase H2 N NH2 H 2N OH
+
R O pH 8.0, 37 °C R O R O
109b R= CH 3 D-109b - 112b L-109a - 112a
110b R= CH 2CH2 CH3
111b R= C6 H11
112b R= C(CH 3) 3
Scheme 15.19 b-Aminopeptidase catalyzed kinetic resolution of racemic b3-amino acid amides
109b–112b [328].
Table 15.6 Properties of microbial b-peptidyl aminopeptidase (b-aminopeptidases).
Ochrobactrum anthropi Pseudomonas sp. Sphingosinicella xenopep- Sphingosinicella micro-

LMG7991 MCI3434 tidilytica 3-2W4 cystinivorans Y2
Reference [279–281] [82] [329] [330]

Name L-Aminopeptidase D-ala- b-Ala-Xaa dipeptidase b-Aminopeptidase 3-2W4 b-Aminopeptidase Y2
nine esterase/amidase (BapA) BapA BapA
(DmpA)
Subunit molecular weight a-Chain: 30 kDa (SDS- a-Chain: 27 kDa (SDS- a-Chain: 27 kDa (SDS- a-Chain: 27 kDa (SDS-
PAGE) 26 564.3 2.6 PAGE). PAGE). PAGE) 25 465 (MALDI-
(ESI-MS) 26 565 TOF-MS) 25 333
(calculated). (calculated).
b-Chain: 13 kDa (SDS- b-Chain: 13 kDa (SDS- b-Chain: 13 kDa (SDS-
PAGE) PAGE) PAGE) 13 168 (MALDI-
TOF-MS) 13 144
(calculated)
b-Chain: 15 kDa (SDS-
PAGE) 13 736.8 0.6
(ESI-MS) 13 737
(calculated)
No. of subunits 4 [a4b4] (X-ray 4 (gel filtration – 150 kDa) 4 [a4b4] (gel filtration – 4 [a4b4] (gel filtration –
crystallography) 130 kDa) (ESI-MS – 150 130 kDa) (ESI-MS – 155
230) 805)
No. of amino acids Precursor: 375. Precursor: 366. Precursor: 402. Precursor: 399.
a-Chain: 248 (2–249). Signal peptide: -. Signal peptide: 29 (1–29). Signal peptide: 26 (1–26).
b-Chain: 126 (250–375) a-Chain: 238 (1–238). a-Chain: 249 (30–278). a-Chain: 249 (27–275).
b-Chain: 128 (239–366) b-Chain: 124 (279–402) b-Chain: 124 (276–399)
(Continued )
j615
616
Ochrobactrum anthropi Pseudomonas sp. Sphingosinicella xenopep- Sphingosinicella micro-

LMG7991 MCI3434 tidilytica 3-2W4 cystinivorans Y2
Homologous to Ntn hydrolases (structur- Ntn hydrolases (structur- Ntn hydrolases (structur- Ntn hydrolases (structur-
al fold) al fold) al fold) al fold)
Isoelectric point (pH) 5.0
Optimum pH 7.5–8.5 9.0–10.0 8–9 10

Optimum temp. ( C) 60
Temp. stability ( C) (preincuba- 45 (10 min) 60 (24 h)
tion time)
Activation by
Inhibitors Completely: pCMB, Completely: Pefablock SC Completely: Pefablock SC
HgCl2, ZnSO4, ZnCl2,
AgNO3.
Partly: N-ethylmaleimide,
DTT, CdCl2
Substrate specificitya)
D-Ala-pNA 6.4 16 — —
L-Ala-pNA 1.1 2.7 — —
L-Lys-pNA 0.21 — — —
L-Arg-pNA 0.27 — — —
H-bhGly-pNAb) 59 — 2.5 2.7
H-L-b3hAla-pNA 100 — 44 62
H-L-b3hPhe-pNA 0.06 — 100 12
Gly-NH2 1.8 ND — —
D-Ala-NH2 0.30 0.65 — —
L-Ala-NH2 0.12 ND — —
H-bhGly-NH2b) — 58 — —
H-DL-b3hAla-NH2 (109b) 44 — 29 100
DL-110b 0.10 — 35 59
DL-111b 0.00 — 19 5.2
DL-112b 0.06 — 5.5 10
H-bhGly-L-Ala-OHb) — 100 — —
H-bhGly-L-His-OHb),c) 13 57 0.06 0.86
H-b3hAla-b3hLeu-OH 7.2 — 2.4 42
Activity for peptides with
N-term. L-a-amino acids Gly, Ala, Leu, Phe, Ser, No No No
Lys, Arg
N-term. L-b-amino acids bhGlyb), b3hAla, b3hVal, bhGlyb) bhGlyb), b3hAla, b3hVal, bhGlyb), b3hAla, b3hVal,
b3hLeu, b3hSer b3hLeu, b3hPhe, b3hTyr, b3hLeu, b3hPhe, b3hTyr,
b3hTrp, b3hSer, b3hThr, b3hTrp, b3hSer, b3hThr,
b3hGln, b3hHis, b3hLys, b3hGln, b3hHis, b3hLys,
b3hArg b3hArg
a) Substrate specificity is given relative to the activity for the substrate that is converted most efficiently. ND: not detected; –: not measured.
b) bhGly and b-alanine are synonyms.
c) H-bhGly-L-His is the systematic name for carnosine.
j617
618
have been cloned. Their most important features are given in Table 15.6 (for a review,
see also Reference [331]). DmpA, 3-2W4 BapA, and Y2 BapA have been used for
the hydrolysis [283] as well as for the synthesis of different b- and mixed
b/a-peptides [332, 333].
The three b-aminopeptidases DmpA, 3-2W4 BapA, and Y2 BapA converted the b3-
amino acid amides 109b–112b L-stereoselectively, forming the L-b3-amino acids
L-109a–112a in high enantiomeric excesses without exception. DmpA was especially
suited for the resolution of rac-109b, which was converted with both high activity
(34 U mg1 protein) and near absolute enantioselectivity (E > 400). Although DmpA
also displayed the highest enantioselectivity (E > 100) for the other three b3-amino
acid amides tested (110b–112b), these sterically more demanding substrates were
converted by this enzyme with very low activities only, confirming that DmpAs
substrate specificity is confined to H-b-hGly- and H-b3-hAla-containing peptides and
amides [283]. The two Sphingosinicella BapA aminopeptidases, in contrast, showed
much broader substrate specificity; all four of the tested b3-amino acid amides were
converted with acceptable rates (0.38–16 U mg1 protein) and moderate to excellent
enantioselectivities (E > 53). Given that very recently the application of DmpA and
3-2W4 BapA for the synthesis of enantiopure b2-amino acids has also been pub-
lished [334], these b-aminopeptidases form a new and promising enzyme platform
for the production of a wide array of enantiopure b-amino acids under mild
conditions.
15.5
Enantioselective Hydrolysis of Hydroxy Acid Amides
As well as enantiomerically pure amino acids, (a-alkylated) a-hydroxy acids are also
important synthons for application in the pharmaceutical industry. Examples are (R)-
and (S)-3,3,3-trifluoro-2-hydroxy-2-methylpropionic acid (113), which are intermedi-
ates for the synthesis of several potential pharmaceuticals, including drugs for the
treatment of incontinence and diabetes (see Reference [63] and primary references
therein). At Lonza AG an efficient chemoenzymatic process was developed for the
large-scale production of both enantiomers of this a-hydroxy acid in high optical
purity [61, 62, 313]. The key step in this process is the amidase-catalyzed kinetic
resolution of racemic 3,3,3-trifluoro-2-hydroxy-2-methylpropionamide (108,
Scheme 15.20). Strains with a suitable amidase were obtained by an enrichment
strategy employing (RS)-amide 108 as sole nitrogen source followed by assessing
their enantiospecificity using a chiral GC analysis. This approach resulted in the
identification of Klebsiella oxytoca PRS1, which contained an amidase specific for the
(R)-amide. The amidase was purified from the cell-free extract from the wild-type
strain by heat treatment, followed by chromatography on an anion-exchange,
hydroxyapatite, and gel filtration column. Characterization showed that the amidase
from K. oxytoca is robust, stable, and does not require cofactors. Assessment of its
substrate specificity revealed that substitution of the methyl group in 108 with an
ethyl group resulted in a 3.7-fold lower activity, whereas substrates with a propyl and
15.5 Enantioselective Hydrolysis of Hydroxy Acid Amides j619
F3C CH 3 Klebsiella oxytoca F3C CH 3 F3C CH 3
amidase
NH2 OH + NH2
HO HO HO
pH 8.0, 37 °C
O O O
r ac 108 (R)-113 (S)-108
Scheme 15.20 Kinetic resolution step in the Lonza process for the production of (R)- and (S)-3,3,3-
trifluoro-2-hydroxy-2-methylpropionic acid (113) [61, 313].
phenyl group were not converted [335]. Substitution of the CF3 group with CCl3 also
resulted in no activity, as was the case when the hydroxyl group was substituted with a
methoxy group. Finally, substituting the hydroxyl group with an amino group, giving
to 3,3,3-trifluoro-2-amino-2-methylpropanamide as substrate, led to a 28-fold higher
hydrolysis rate than for the corresponding a-hydroxy compound 108 [335]. A similar
positive effect of an a-amino over an a-hydroxy group was observed for LamA from
O. anthropi [309].
Escherichia coli clones expressing the K. oxytoca amidase were isolated from an
expression library by their ability to grow on a medium with (RS)-amide 108 as sole
source of nitrogen. Sequencing revealed that the gene encoding this amidase (sad)
coded for a polypeptide of 328 amino acid residues with a calculated molecular
weight of 36 344 [313] and homology to the acetamidase/formamidase family of
proteins, including the O. anthropi NCIMB 40321 LamA (28% full-length sequence
identity) [309]. Efficient overexpression of this amidase in E. coli was possible by
placing the gene downstream of the strong lac promoter. Biotransformations of
(RS)-amide 108 were performed with washed whole cells that had been heat treated
(70 C, 10 min) to stabilize the amidase activity. The production of 113 was
successfully scaled up to the 100-kg scale (volume 1500 l). The substrate concen-
tration was 10% (w/v) and heat treated E. coli cells were used at OD650nm ¼ 1.0–5.0.
After completion the pH was lowered to 4.0 to stop the reaction. Then cells were
removed by microfiltration, followed by removal of traces of protein from lysed cells
by ultrafiltration. The (R)-acid 113 obtained had a chemical purity of >98% and an
e.e. of essentially 100% [61].
As already indicated in Section 15.4.2 the O. anthropi L-amidase LamA is active
towards the a-hydroxy acid amide DL-mandelic acid amide (26). This substrate was
resolved L-selectively by this enzyme (E > 300), furnishing the reaction product in an
e.e. of 98.4% [336]. In addition, the D-amidase from V. paradoxus (Section 15.4.1.3)
displayed activity toward an a-hydroxy acid amide, that is, lactic acid amide (28);
however, the L- and D-enantiomers of this amide were hydrolyzed with nearly equal
rates [285, 286], most likely precluding its application for resolving this substrate.
Another enzyme catalyzing the hydrolysis of a hydroxy acid amide is mandelamide
hydrolase from P. putida ATCC 12633 (MAH, encoded by mdlY) [337]. This enzyme,
which is part of the mandelate pathway, converts (R)- and (S)-mandelamide into
mandelic acid and ammonia with almost equal catalytic efficiency, and is, thus, non-
enantioselective [338]. This enzyme belongs to the amidase signature family and thus
contains the Ser-cisSer-Lys catalytic triad. Studies to determine the substrate
620
specificity of MAH showed that phenylacetamide is the optimal substrate for this
enzyme with an approximately tenfold higher kcat/Km than that for (R)- and
(S)-mandelamide, which is mainly caused by a lowered binding affinity for these
a-hydroxy acid amides [338]. Besides aromatic substrates, also aliphatic substrates
are converted by MAH, albeit with a much lower efficiency. This reduced efficiency is
mainly caused by a decreased affinity. Compared to phenylacetamide, for instance,
(R)- and (S)-lactamide showed a largely unaffected kcat but a more than five orders of
magnitude increased Km [339]. Determination of the substrate specificity of MAH
also revealed that substituents at the a-carbon atom have only a relative minor effect
on the kinetics of this enzyme, leading to rather insignificant enantioselectivity with
aromatic and aliphatic substrates. In a recent study aimed at converting the MAH into
an lactamide hydrolase by combining random mutagenesis with a selection method
for variants with an enhanced ability to utilize lactamide as sole carbon source, a
mutant (I437N) was identified that increased MAHs enantioselectivity toward
(S)-lactamide [339]. In addition, this study showed that Gly202 played a role in the
specificity for aromatic versus aliphatic substrates. Whereas mutant G202A had
drastically increased Km values for all aromatic substrates tested (250–650-fold), its
Kms for the aliphatic substrates changed only marginally (0.4–2.6-fold). Interestingly,
introduction of the mutation G202A did not result in major changes in kcat [339].
Optimization of the tools developed in this study in combination with increased
knowledge of MAHs structure–function relationship will enable the future con-
struction of more enantioselective a-hydroxy acid amide hydrolyzing amidases.
15.6
Enantioselective Hydrolysis of Azido Acid Amides
a-Azido carboxylic acids may be used as synthetic precursors for natural and non-
natural amino acids. Their use can solve one of the last difficulties in solid-phase
(and solution-phase) peptide synthesis, that is, the problem of sterically hindered
couplings (e.g., a,a-disubstituted amino acids) [340]. In this approach, the azido
group of the incoming monomer acts effectively as a protected amino group, which
allows high activation of its carboxyl group as the acid chloride without by-product
formation or detectable racemization. This activated azido acid can then be coupled
to the N-terminus of the growing peptide and reduced in high yield on the solid-
phase [341–343].
Because many applications of peptides require products of high enantiomeric
purity, especially in pharmaceutical use, and an efficient and generally applicable
chemical method for the synthesis of enantiomerically pure a-azido acids is not
available yet, Meldal and coworkers investigated whether the aminopeptidase based
kinetic resolution of a-H-a-azido acid amides offers an attractive alternative
approach. As biocatalyst they used the recombinant E. coli cells heterologously
expressing the L-aminopeptidase gene pepA from P. putida ATCC 12633 (Section
15.4.1.1); 2-azidohexanoic acid amide (114) and 2-azidophenylacetic acid amide (115)
were tested as racemic substrates (Scheme 15.21) [344].
15.6 Enantioselective Hydrolysis of Azido Acid Amides j621
R R R
E.coli DH5 /pTrpLAP OH
NH 2 NH2 +
N3 N3 N3
pH 9.0, 40 °C O
O O
1 mM Mn2+
114 R = (CH2)3CH3 (R)-114-115 (S)-114a-115a
115 R = Ph
Scheme 15.21 Pseudomonas putida aminopeptidase catalyzed resolution of a-H-a-azido

carboxamides.
The recombinant E. coli cells (cell : substrate ratio 1 : 10) displayed low but
significant activity toward the two azido carboxamides tested, which could be solely
attributed to the P. putida aminopeptidase. Hydrolysis of the racemic 2-azidohex-
anoic acid amide (114) progressed as a typical kinetic resolution with an L-enantio-
selective enzyme, affording the L-2-azidohexanoic acid (114a) with >99.8% e.e. at
50% conversion (20 h).
Unexpectedly, the course of the hydrolysis reaction of the racemic 2-azidopheny-
lacetic acid amide (115) was quite different. Besides a fourfold higher activity, this
reaction continued when 50% conversion had been reached, albeit with an approx-
imately 100-fold reduced rate. Because the e.e. of the L-2-azidophenylacetic acid
(115a) remained above 98% throughout the whole reaction, the second phase is
caused by racemization of the remaining D-2-azidophenylacetic acid amide (115b) in
combination with hydrolysis of the formed L-amide by the P. putida PepA. Thus, the
conversion of aromatic azido acid amide 115 with the recombinant E. coli system
proceeds as a dynamic kinetic resolution, and, therefore, has a theoretical maximum
yield of the L-azido acid 115a of 100%. It has been hypothesized that the 2-
azidophenylacetic acid amide 115 racemizes in situ because its three electron-
withdrawing substituents render the a-hydrogen atom more acidic than in the
corresponding a-amino and a-hydroxy analogues, phenylglycine amide, and man-
delic acid amide, which are optically stable under similar conditions [344].
More recently Sewald and coworkers described the enzymatic resolution of two
a,a-dialkylated a-azido carboxamides, 2-azido-2,4-dimethylpentanamide (116) and
2-azido-2-methyl-3-phenylpropanamide (117) (Scheme 15.22) [345]. Because these
substrates did not contain an a-hydrogen atom, the L-amidase from O. anthropi
NCIMB 40321 was used for the resolution reactions. This amidase converted both
H 3C R O.anthropi L-amidase H 3C R H3 C R
NH 2 NH 2 + OH
N3 N3 N3
pH 8.0, 55 °C
O 1 mM Zn 2+ O O
116 (R =i-Bu) (R)-116-117 (S)-116a-117a
117 (R = Bn)
Scheme 15.22 Amidase catalyze resolution of a,a-dialkylated a-azido carboxamides.

622
substrates, albeit with a very low activity. Whereas 117 was converted with a moderate
stereoselectivity only (33% e.e. at 45% conversion), the resolution of 116 proceeded
with very high stereoselectivity leading to an e.e. of the a-azido acid 116a product of
96% (30% conversion). Resolution of amide 116 was subsequently performed on a
preparative scale employing the same biocatalyst. The (S)-2-azido-2,4-dimethylpen-
tanoic acid (116a) formed and the remaining (R)-carboxamide were obtained in e.e.s
of 96% and 95%, and yields of 48% and 50%, respectively (E-ratio 150). The (S)-
2-azido-2,4-dimethylpentanoic acid (116a), which is a synthetic precursor of a-
methylleucine, was subsequently incorporated into an analog of the peptide antibi-
otic efrapeptin C [345].
15.7
Selective Cleavage of a C-Terminal Amide Bond
The use of protecting groups is inextricably bound up with the in vitro synthesis of
peptides from its component amino acids. It prevents the formation of side products
by, for instance, uncontrolled polymerization or reaction to side chains. Owing to the
selectivity of enzymes, biocatalytic peptide synthesis methods often require no or
only limited protection of side chains. Furthermore, no additional activating reagents
and less organic solvents are needed. The complete absence of racemization during
the coupling step is another advantage of the use of enzymes in peptide synthesis, as
this leads to purer products and easier product isolation. Thus, biocatalytic peptide
synthesis is a more environmentally friendly and more cost-effective alternative to
chemical peptide synthesis [346].
The carboxamide as C-terminal protecting group offers some important advan-
tages in peptide synthesis. These include a good chemical stability and increased
peptide solubility in water, the solvent mainly used in enzymatic peptide synthe-
sis [347]. Unfortunately, the selective cleavage of the amide bond in the C-terminal
position of peptide amides, which is essential in certain methods for stepwise chain
elongation and in case the final peptide product contains a free carboxylic acid
function, has long been impossible, because both chemical and enzymatic means
resulted in concomitant hydrolysis of the internal peptide bonds and/or side chain
amide groups. Consequently, for a long time the carboxamide group found limited
use in peptide synthesis. The isolation of a novel type of amidase, called peptide
amidase, about two decades ago, however, has changed this situation.
15.7.1
Peptide Amidase from the Flavedo of Oranges
During a search for a carboxypeptidase C in the flavedo (the outer colored layer of the
exocarp of citrus fruit) of oranges, Steinke and Kula serendipitously isolated an
enzyme with a novel kind of peptide amidase activity (peptide amidase from the
flavedo of oranges – PAF) [347, 348]. PAF was partially purified from the extract of the
orange flavedo by a simple procedure based on ammonium sulfate fractionation, gel
15.7 Selective Cleavage of a C-Terminal Amide Bond j623
Table 15.7 Properties of the peptide amidases from the flavedo of oranges (Citrus sinensis L.) and
Stenotrophomonas maltophilia.
Source Citrus sinensis L. Stenotrophomonas maltophilia
Reference [347–350] [359]

Subunit molecular mass (kDa) 23 3 (SDS-PAGE) 53.5 (calculated)a)
Native molecular mass (kDa) Not detected 50 (gel filtration)
Isoelectric point (pH) 9.5 5.8a)
Optimal temperature ( C) 30–35 46–54b)
Optimal pH 7.6 0.8 7.6 0.6 and 11.0b)
Inhibitors PMSF (weak) Chymostatin, Pefabloc SC (weak)
Cofactor/metal ion requirement No No
Specific activity at pH 7.5 (U mg1) 0.59 58a)
[10 mM Z-Gly-Tyr-NH2, 5% DMF]
a) Data of PAM variant 10 [359].

b) Data of PAM variant 6 [359].
filtration, and DEAE ion exchange chromatography [347]. Table 15.7 gives some of the
physicochemical characteristics of this enzyme. Typical of this peptide amidase is its
activity toward the C-terminal amide bond in peptide amides without any concom-
itant hydrolysis of internal peptide bonds or side chain amide bonds in substrate or
product (Scheme 15.23). PAF accepts a broad range of substrates – the C-terminal
amide bond is hydrolyzed from N-protected and unprotected peptide amides of
apparently any length. In addition, the amino acid composition of the peptide
amides, including the side-chain of the C-terminal residue, is only of minor
importance. Exceptions with regard to the C-terminal position reported so far are
L-Pro-NH2 and D-amino acid amides, which is a consequence of the absolute
L-stereoselectivity of the enzyme [347]. However, D-amino acids in the penultimate
position are tolerated.
O Peptide Amidase, H2 O O
H H
R2 N R2 N +
NH 2 O NH 4
O R1 O R1
Scheme 15.23 Cleavage reaction catalyzed by the peptide amidase from the flavedo of oranges
(Citrus sinensis). R1: side chain of C-terminal amino acid residue; R2: amino acid residue, peptide
residue, or N-terminal protecting group.
Besides C-terminal peptide amides, PAF also hydrolyzes N-terminal protected

amino acid amides in an L-selective manner, which is the basis of a process for
the enzymatic resolution of racemic N-acyl amino acid amides. Resolution of Z-
DL-Ala-NH2, Ac-DL-Met-NH2, and Ac-DL-neopentylglycine amide (Ac-DL-Npg-NH2) with
PAF furnished the N-acyl-L-amino acids in over 99% e.e. at half conversion [349].
624 j 15 Hydrolysis of Amides
Interestingly, PAF has no activity towards amino acid amides with a free a-amino
group [347, 350]. This feature makes PAF especially suited for a two-step enzymatic
synthesis method of dipeptides [351]. In a typical example (Scheme 15.24), carboxy-
peptidase Y (CPD-Y) catalyzed the kinetically driven coupling of H-Tyr-OEt (118)
and H-Arg-NH2 (119). The dipeptide amide formed (120) was then deamidated to H-
Tyr-Arg-OH (121) by PAF without concurrent hydrolysis of the non-reacted Arg-NH2
or the dipeptide, which significantly simplified downstream processing. Because of
PAFs broad pH range (approx. 6–9) and good operational stability at pH 9, both
enzymatic reactions were operated at pH 9, the optimal pH for the CPD-Y catalyzed
reaction. Modeling clearly demonstrated that the two reactions had to be carried out
in a cascade of two reactors to prevent hydrolysis of the peptide by CPD-Y and to
obtain good yields. Continuous H-Tyr-Arg-OH (121) production in such a cascade
was possible with a space–time yield of 239 g l1 d1 [351].
H 2N
NH
HN
HO HO
CPD-Y O
+ H
OC 2H 5 NH2 N
H 2N H 2N pH 9 H 2N NH 2
O O 20 °C O
118 119
120
NH
HN
HO NH2
O PAF
H
N
H 2N OH pH 9
O 20 °C
121
NH
HN
NH2
Scheme 15.24 CPD-Y/PAF catalyzed synthesis of dipeptide H-Tyr-Arg-OH (121).
Recently, it has been found that PAF can also be applied for the direct conversion of
an N-terminal protected peptide amide into the corresponding peptide methyl
ester [346]. By tuning of the water and methanol concentrations, the hydrolysis
reaction furnishing the C-terminal carboxylic acid side product instead of the desired
C-terminal methyl ester could be minimized. With Z-Gly-Tyr-NH2 (122) as substrate,
methanol as solvent, and a water concentration increasing from 7 to 22 wt% due to the
addition of fresh amounts of PAF, Z-Gly-Tyr-OMe (123) and Z-Gly-Tyr-OH (124) were
obtained in a ratio of 4 to 1 at 50% substrate conversion (Scheme 15.25).
OH
O
H
O N OCH 3
N
OH H
O O
PAF 123
O
H
O N NH 2 CH3 OH/H2 O, + OH
N
H MgHPO 4
O O
O
122 H
O N OH
N
H
O O
124
Scheme 15.25 PAF catalyzed C-terminal activation of Z-Gly-Tyr-NH2 (122) to Z-Gly-Tyr-OMe

(123).
rovsky and Kula have shown that PAF can also catalyze the reverse reaction, that
Ce
is, the C-terminal amidation of peptides. Using a thermodynamically controlled
reaction in acetonitrile containing 5 vol.% of water and a 1.4 molar excess of
ammonium hydrogen carbonate (NH4HCO3), Z-Gly-Phe-OH was amidated in yields
of up to 35% [352]. This reaction is of importance because the presence of a
C-terminal carboxamide group is essential for the biological activity of many peptide
hormones. No less than 50% of the mammalian and >80% of the insect peptide
hormones are amidated, which makes C-terminal a-amidation the most important
posttranslational enzymatic modification by far [353]. Peptides produced by fermen-
tation applying recombinant DNA technology, however, lack such C-terminal amide
group. Because its chemistry-based introduction requires laborious protection and
deprotection of certain side chain functional groups, a mild enzymatic method is of
great value.
Thus, although PAF catalyzed C-terminal amidation of peptides is in principle
possible, its application has been hampered by the precipitation of some of the
peptide substrates as insoluble ammonium salts. Ce rovsky and Kula have partly
solved this problem by optimization of the solvent mixture with regard to substrate
solubility and PAF stability. A reaction medium of acetonitrile with 20–25 vol.% of
dimethylformamide and 3 vol.% of water led to the maximal amidation yield of most
peptides [354]. Under these conditions, the substrate specificity of PAF was deter-
mined applying a broad range of N-protected di-, tri-, tetra-, and pentapeptides as
model substrates. This surprisingly showed that this was much more restricted than
for amide hydrolysis. Peptides with a hydrophilic or charged amino acid residue at
their C-terminus, for example, were not amidated or with low yields only. In addition,
626
the presence of a charged amino acid residue in the penultimate position severely
hindered amidation. Furthermore, this study showed that the yields of peptide
amidation are influenced by the length of the peptide chain, dropping dramatically
when the peptide is longer than four residues. A clear rationale for this much
narrower substrate specificity in the direction of amidation could not be given, but an
influence of the different reaction medium on the secondary structure of the
substrate, the structure of the PAF active site, and/or the peptide solubility have
been mentioned as potential reasons [354].
Finally, it was demonstrated that PAF can catalyze the C-terminal amidation of
peptides in nearly anhydrous ionic liquids. Maximum yields of the same order of
magnitude as in conventional organic reaction media were obtained [355].
15.7.2
Peptide Amidase from Microbial Sources
Although PAF can also be isolated from orange peel, which is a waste product of the
juice industry and thus cheaply available at large scale [356], its supply for commercial
applications is severely hampered because its concentration in orange peel is highly
dependent on seasonal influences and other uncontrollable factors [357]. A more
secure source of supply was thus highly desirable. However, all attempts to purify
PAF to homogeneity have failed so far, which has been attributed to its varied
glycosylation [357]. Therefore, the gene encoding PAF could not be identified to date,
which prevents its efficient heterologous production in a microorganism [358].
As an alternative, Kula and coworkers looked for microbial sources of peptide
amidase, and this was identified in different strains of the bacteria Stenotrophomonas
maltophilia (originally named Xanthomonas maltophilia) and Ochrobactrum
anthropi [357, 358]. One of the S. maltophilia strains displayed the highest peptide
amidase activity, and the amidase from this strain (PAM) was purified to near
homogeneity by a three-step procedure (anion-exchange chromatography, gel filtra-
tion, and isoelectric focusing) [357]. Analysis of PAMs substrate spectrum showed
that this is nearly identical to that of PAF [349]. PAM also L-stereoselectively
deamidates the C-terminal amide group in peptide amides and N-protected amino
acid amides, without hydrolyzing internal peptide bonds or amide functions of the
glutamine and asparagine side chains [357]. Furthermore, substrates with a bulky
b-branched side chain (e.g., valine, isoleucine, and tert-leucine) at their C-terminus
are hardly converted [349].
A few years after the identification of this microbial peptide amidase, its gene (pam)
was isolated using a probe based on PAMs N-terminal amino acid sequence.
Analysis of this gene showed that PAM belongs to the amidase signature (AS)
family [25]. Comparison of the gene sequence and N-terminal amino acid sequence
of the purified protein revealed that PAM is formed with a 37 amino acid N-terminal
signal sequence that is cleaved off during its translocation to the periplasm. The
processed protein is 503 amino acids long and has a molecular weight of 53.5 kDa.
Table 15.7 gives some other characteristic properties of this microbial peptide
amidase. Efficient formation of PAM in the cytoplasm of E. coli Origami (DE3) cells
was established by expressing the gene without signal peptide encoding region. By
optimization of the IPTG concentration and growth temperature, PAM represented
31% of the soluble cellular protein [359]. Because the protein was formed with a
C-terminal His6 tag, it could be purified to near homogeneity in a final yield close to
100% using a Ni-NTA column. Interestingly, the specific activity of the recombinantly
produced PAM was found to be much higher than that of the PAM isolated
from S. maltophilia: 194 [359] and 4.6 U mg1 [357], respectively (substrate 10 mM
Ala-Phe-NH2). Although this higher activity of the recombinant PAM partly stemmed
from optimization of the assay conditions, it was also hypothesized that the PAM
isolated from S. maltophilia was C-terminally truncated during the purification process
with a much lower enzyme activity as a result [359]. The fact that the molecular mass of
the wild-type PAM isolated from S. maltophilia was determined by gel filtration to be
38 kDa [357], as compared to a native molecular mass of 50 kDa (gel filtration) and
53.5 kDa (DNA sequence), respectively, for the recombinant PAM [359], supported this
observation. In addition, other physicochemical properties like pH and temperature
optima of the recombinant and wt PAM appeared to be quite different.
The improved availability of PAM through its efficient recombinant production
enabled a more detailed assessment of its substrate spectrum. In contradiction to
earlier reports, PAM appeared to hydrolyze free amino acid amides, too, with a
strong preference for the L-amino acid amide, albeit with rather low activity. As
expected, N-acylation or addition of a further amino acid residue resulted in a
dramatically increased hydrolytic activity [359]. It was also found that the C-terminal
and the penultimate amino acid residue have a much larger effect on the activity than
earlier thought. A glycyl residue in the ultimate and penultimate position, for
example, had a clear negative impact on the activity. Thus, interactions of central
importance between the PAM and the substrate extend to parts of the substrate
molecule beyond the ultimate amino acid amide [359].
The reason why di- and tripeptide amides and N-protected amino acid amides are
much better substrates than free amino acid amides has become clear from the

crystal structure of the native PAM (1.4 A resolution) and, especially, of PAM in

complex with the competitive inhibitor chymostatin (1.8 A resolution) [360, 361]. The
latter structure showed that the enzyme forms hydrogen bonds between Nd of
Asn172 and the carbonyl O of the 2nd and 3d residue of the substrate, a directed
interaction that cannot take place in case of a free amino acid amide. Undirected
interactions of van der Waals type are another major driving force for substrate
binding, explaining the broad substrate spectrum favoring large hydrophobic side-
chains. The fact that the X-ray structure revealed that b-branched side-chains for the
two C-terminal amino acid residues are sterically hindered is also in line with earlier
experimental results [349, 361].
Like other members of the amidase signature family of enzymes, PAM employs a
unique, highly conserved Ser-Ser-Lys catalytic triad for amide hydrolysis [361]. The
catalytic triad residues, Ser226, Ser202, and Lys123, form a hydrogen-bonding
network, where Ser226 acts as the primary nucleophile and Ser202 bridges Ser226
and Lys123. In line with their essential function, mutagenesis of these residues
greatly impacted the enzymatic activity. Whereas the mutation Ser202Ala resulted in
628
a 140-fold reduced activity, the mutants Ser226Ala and Lys123Ala were completely
inactive [361]. Molecular dynamics (MD) simulations not only supported the pres-
ence of the hydrogen bonding network mentioned above, but also showed that an
oxyanion hole is created by the backbone amide nitrogens of Asp224 and Thr223 by
hydrogen bonding to the terminal carboxamide oxygen of the substrate, which
stabilizes the oxyanion tetrahedral intermediate. This intermediate is further stabi-
lized by another hydrogen bond, that is, that with the backbone amide NH group of
Ser226 [362]. Based on these MD simulations and the bimodal pH profile with
maxima at pH 7.6 and pH 10.4, two rather similar mechanisms for PAM catalysis
have been proposed. At higher pH, Lys123-NH2 functions as a general base catalyst
facilitating proton transfer from Ser226-OH via the bridging Ser202-OH. At lower
pH involving Lys123-NH3 þ , in the transition state the proton on Ser226-OH is being
removed to the proton transfer channel consisting of ordered water molecules [362].
Because S. maltophilia is found in many different environments, the existence of
more than one pH-dependent reaction pathways is not surprising, as this will give the
PAM a broader pH range of activity.
15.8
Summary and Outlook
In this chapter the hydrolysis of primary amides by amidases, a-amino amidases, and
related enzymes is discussed, with special emphasis on stereoselective conversions. In
general, these enzymes are (S)-selective and possess a high regio- and enantioselec-
tivity under mild aqueous reaction conditions, but also (R)-selective amidases are
known. For thermodynamic reasons the conversion of these enzymes is limited to
hydrolysis and transamination reactions. A few of the amidase catalyzed reactions
have been commercialized and used for the preparation of enantiomerically pure
carboxylic acids on a multi-ton scale for pharmaceutical applications. In addition,
amino amidases have been used on a production scale for the preparation of
enantiomerically pure unnatural a-amino acids. Some of these amino amidases have
considerable substrate flexibility and can hydrolyze a-hydroxy and a-azido amides in
addition to a-amino amides. A specific class of amidases is able to hydrolyze cyclic
amides (lactams), including cyclic a-amino amides like a-amino caprolactam. Besides
amidase, other enzymes, such as acylases and deformylases, are also able to hydrolyze
amide bonds and give access to chiral amines, while alcalase and various lipases have
been used in the stereoselective aminolysis of carboxylic acid esters to (primary)
amides. These enzymatic reactions will be discussed in more detail in other chapters.
An important disadvantage of the amidase catalyzed reactions for further com-
mercialization is their limited maximum yield of 50% yield, which is typical of a
resolution process. For the enzymatic hydrolysis of a-amino acid amides, combi-
nation with chemical and enzymatic racemization of the remaining substrate
enantiomer resulted in a considerable process improvement, making commercial
application more attractive. Finally, the use of peptide amidases holds good prospects
for the enzymatic synthesis of peptides from all kinds of natural and unnatural amino
References j629
acid esters and amides. As indicated in the preceding sections, amides and their
derivatives are important versatile building blocks for the (agro)chemical and
pharmaceutical industry. Owing to the selectivity of amidases (both regio- and
enantioselectivity) and the fact that these conversions can be achieved under very
mild conditions, several biocatalytic processes based on amidases and amino
amidase have recently been commercialized. The use of these biocatalysts in the
chemical industry is expected to increase in importance in the near future as
environmental restrictions become more pronounced.
References
1 Woodley, J.M. (2008) New opportunities J.H.M., and Broxterman, Q.B. (2004)
for biocatalysis: making pharmaceutical Peptide deformylase as biocatalyst for the
processes greener. Trends Biotechnol., 26, synthesis of enantiomerically pure
321–327. amino acid derivatives. J. Mol. Catal. B:
2 Tao, J. and Xu, J.-H. (2009) Biocatalysis in Enzym., 29, 265–277.
development of green pharmaceutical 10 Laumen, K., Kittelmann, M., and
processes. Curr. Opin. Chem. Biol., 13, Ghisalba, O. (2002) Chemo-enzymatic
43–50. approaches for the creation of novel chiral
3 Pollard, D.J. and Woodley, J.M. (2007) building blocks and reagents for
Biocatalysis for pharmaceutical pharmaceutical applications. J. Mol.
intermediates: the future is now. Trends Catal. B: Enzym., 19–20, 55–66.
Biotechnol., 25, 66–73. 11 Bornscheuer, U.T. and Kazlauskas, R.J.
4 May, O. (2011) Sustainable processes (2006) Hydrolases in Organic Synthesis,
based on enzymes enabling 100% yield Wiley-VCH Verlag GmbH, Weinheim.
and 100% ee concepts, in Pharmaceutical 12 Wahl, P., Walser-Volken, P., Laumen, K.,
Process Chemistry, 1st edn (eds T. Shioiri, Kittelmann, M., and Ghisalba, O. (1999)
K. Izawa, and T. Konoike), Wiley-VCH Microbial production, purification, and
Verlag GmbH, Weinheim, pp. 321–344. characterization of (S)-specific N-acetyl-2-
5 Wieser, M. and Nagasawa, T. (2000) amino-1-phenyl-4-pentene
Stereoselective nitrile-converting amidohydrolase from Rhodococcus
enzymes, in Stereoselective Biocatalysis globerulus K1/1. Appl. Microbiol.
(ed. R.N. Patel), Marcel Dekker, Inc., Biotechnol., 53, 12–18.
New York, pp. 461–486. 13 Graf, M., Brunella, A., Kittelmann, M.,
6 Sharma,M., Sharma, N., and Bhalla, T. Laumen, K., and Ghisalba, O. (1997)
(2009) Amidases: versatile enzymes in Isolation and characterization of highly
nature. Rev. Environ. Sci. Biotechnol., 8, (R)-specific N-acetyl-1-phenylethylamine
343–366. amidohydrolase, a new enzyme from
7 Martınkova, L. and Kren, V. (2002) Nitrile- Arthrobacter aurescens AcR5b. Appl.
and amide-converting microbial Microbiol. Biotechnol., 47, 650–657.
enzymes: stereo-, regio- and 14 Brunella, A., Graf, M., Kittelmann, M.,
chemoselectivity. Biocatal. Biotransform., Laumen, K., and Ghisalba, O. (1997)
20, 73–93. Production, purification, and
8 Fournand, D. and Arnaud, A. (2001) characterization of a highly
Aliphatic and enantioselective amidases: enantioselective (S)-N-acetyl-1-
from hydrolysis to acyl transfer activity. phenylethylamine amidohydrolase from
J. Appl. Microbiol., 91, 381–393. Rhodococcus equi Ac6. Appl. Microbiol.
9 Sonke, T., Kaptein, B., Wagner, A.F.V., Biotechnol., 47, 515–520.
Quaedflieg, P.J.L.M., Schultz, S., 15 Forro, E. and F€ul€op, F. (2004) Advanced
Ernste, S., Schepers, A., Mommers, procedure for the enzymatic ring
630
opening of unsaturated alicyclic structural evidence for a conserved

b-lactams. Tetrahedron: Asymmetry, 15, genetic coupling with nitrile hydratase. J.
2875–2880. Bacteriol., 173, 6694–6704.
16 Park, H.-J., Uhm, K.-N., and Kim, H.-K. 26 Cilia, E., Fabbri, A., Uriani, M.,
(2008) R-stereoselective amidase from Scialdone, G.G., and Ammendola, S.
Rhodococcus erythropolis No. 7 (2005) The signature amidase from
acting on 4-chloro-3-hydroxybutyramide. Sulfolobus solfataricus belongs to the
J. Microbiol. Biotechnol., 18, 552–559. CX3C subgroup of enzymes
17 Zhou, Z., Hashimoto, Y., and cleaving both amides and nitriles:
Kobayashi, M. (2005) Nitrile Ser195 and Cys145 are predicted to be the
degradation by Rhodococcus: useful active site nucleophiles. Eur. J. Biochem.,
microbial metabolism for industrial 272, 4716–4724.
productions. Actinomycetologica, 19, 27 Shin, S., Lee, T.-H., Ha, N.-C., Koo, H.M.,
18–26. Kim, S., Lee, H.-S., Kim, Y.S., and
18 Banerjee, A., Sharma, R., and Oh, B.-H. (2002) Structure of
Banerjee, U.C. (2002) The nitrile- malonamidase E2 reveals a novel
degrading enzymes: current status and Ser-cisSer-Lys catalytic triad in a new
future prospects. Appl. Microbiol. serine hydrolase fold that is
Biotechnol., 60, 33–44. prevalent in nature. EMBO J., 21,
19 Ramakrishna, C., Dave, H., and 2509–2516.
Ravindranathan, M. (1999) Microbial 28 Ohtaki, A., Murata, K., Sato, Y.,
metabolism of nitriles and Its Noguchi, K., Miyatake, H., Dohmae, N.,
biotechnological potential. J. Sci. Ind. Yamada, K., Yohda, M., and Odaka, M.
Res., 58, 925–947. (2010) Structure and characterization of
20 Thuku, R.N., Brady, D., Benedik, M.J., amidase from Rhodococcus sp. N-771:
and Sewell, B.T. (2009) Microbial insight into the molecular mechanism of
nitrilases: Versatile, spiral forming, substrate recognition. Biochim. Biophys.
industrial enzymes. J. Appl. Microbiol., Acta, 1804, 184–192.
106, 703–727. 29 Nawaz, M.S., Khan, A.A.,
21 Wang, M.-X. (2005) Enantioselective Bhattacharayya, D., Siitonen, P.H., and
biotransformations of nitriles in Cerniglia, C.E. (1996) Physical,
organic synthesis. Top. Catal., 35, biochemical, and immunological
117–130. characterization of a thermostable
22 Pace, H.C. and Brenner, C. (2001) amidase from Klebsiella
The nitrilase superfamily: classification, pneumoniae NCTR 1. J. Bacteriol., 178,
structure and function. Genome Biol., 2, 2397–2401.
reviews0001.1– reviews000.9. 30 Nawaz, M.S., Khan, A.A., Seng, J.E.,
23 Egorova, K., Trauthwein, H., Verseck, S., Leakey, J.E., Siitonen, P.H., and
and Antranikian, G. (2004) Purification Cerniglia, C.E. (1994) Purification and
and properties of an enantioselective and characterization of an amidase from an
thermoactive amidase from the acrylamide-degrading Rhodococcus sp.
thermophilic actinomycete Appl. Environ. Microbiol., 60,
Pseudonocardia thermophila. Appl. 3343–3348.
Microbiol. Biotechnol., 65, 38–45. 31 Cai, G., Zhu, S., Wang, X., and Jiang, W.
24 Chebrou, H., Bigey, F., Arnaud, A., and (2005) Cloning, sequence analysis and
Galzy, P. (1996) Study of the amidase expression of the gene encoding a novel
signature group. BBA-Protein Struct. M., wide-spectrum amidase belonging to the
1298, 285–293. amidase signature superfamily from
25 Mayaux, J.-F., Cerbelaud, E., Soubrier, F., Achromobacter xylosoxidans. FEMS
Yeh, P., Blanche, F., and Petre, D. (1991) Microbiol. Lett., 249, 15–21.
Purification, cloning, and primary 32 Trott, S., Bauer, R., Knackmuss, H.-J., and
structure of a new enantiomer-selective Stolz, A. (2001) Genetic and biochemical
amidase from a Rhodococcus strain: characterization of an enantioselective
References j631
amidase from Agrobacterium tumefaciens 42 Kim, S.-H. and Oriel, P. (2000)
strain d3. Microbiology, 147, 1815–1824. Cloning and expression of the nitrile
33 Bauer, R., Hirrlinger, B., Layh, N., hydratase and amidase genes from
Stolz, A., and Knackmuss, H.-J. (1994) Bacillus sp. BR449 into Escherichia coli.
Enantioselective hydrolysis of Enzyme Microb. Technol., 27,
racemic 2-phenylpropionitrile and other 492–501.
(R,S)-2 arylpropionitriles by a new 43 Soong, C.-L., Ogawa, J., and Shimizu, S.
bacterial isolate, Agrobacterium (2000) A novel amidase (half-amidase) for
tumefaciens strain d3. Appl. Microbiol. half-amide hydrolysis involved in the
Biotechnol., 42, 1–7. bacterial metabolism of cyclic
34 Asano, Y., Tachibana, M., Tani, Y., and imides. Appl. Environ. Microbiol., 66,
Yamada, H. (1982) Purification and 1947–1952.
characterization of amidase which 44 Maestracci, M., Thiery, A., Arnaud, A.,
participates in nitrile degradation. Agric. and Galzy, P. (1986) A study of the
Biol. Chem., 46, 1175–1181. mechanism of the reactions
35 Hashimoto, M., Mizutani, A., Tago, K., catalyzed by the amidase Brevibacterium
Ohnishi-Kameyama, M., Shimojo, T., sp. R312. Agric. Biol. Chem., 50,
and Hayatsu, M. (2006) Cloning and 2237–2241.
nucleotide sequence of carbaryl 45 Mayaux, J.-F., Cerbelaud, E., Soubrier, F.,
hydrolase gene (cahA) from Arthrobacter Faucher, D., and Petre, D. (1990)
sp. RC100. J. Biosci. Bioeng., 101, Purification, cloning, and primary
410–414. structure of an enantiomer-selective
36 Hayatsu, M., Mizutani, A., amidase from Brevibacterium sp. strain
Hashimoto, M., Sato, K., and Hayano, K. R312: structural evidence for genetic
(2001) Purification and characterization coupling with nitrile hydratase.
of carbaryl hydrolase from Arthrobacter J. Bacteriol., 172, 6764–6773.
sp. RC100. FEMS Microbiol. Lett., 201, 46 Maestracci, M., Thiery, A., Bui, K.,
99–103. Arnaud, A., and Galzy, P. (1984) Activity
37 Maestracci, M., Bui, K., Thiery, A., and regulation of an amidase (acylamide
Arnaud, A., and Galzy, P. (1988) amidohydrolase, EC 3.5.1.4)
The amidases from a Brevibacterium with a wide substrate spectrum from a
strain: study and applications. Brevibacterium sp. Arch. Microbiol., 138,
Adv. Biochem. Eng. Biotechnol., 36, 315–320.
67–115. 47 Cerbelaud, E. and Petre, D. (1989)
38 Hynes, M.J. (1975) Amide utilization in Procede de preparation dacides aryl-2
Aspergillus nidulans: evidence for a third propioniques optiquement
amidase enzyme. J. Gen. Microbiol., 91, actifs. EP 0.330.529 to Rhone-Poulenc
99–109. Sante.
39 Hynes, M.J. and Pateman, J.A. (1970) 48 Petre, D., Cerbelaud, E., Mayaux, J.-F.,
The use of amides as nitrogen sources by and Yeh, P. (1993) Enantioselective
Aspergillus nidulans. J. Gen. Microbiol., 63, amidases, DNA sequences encoding
317–324. them, method of preparation and
40 Hynes, M.J. (1970) Induction and utilization. US 5,260,208 to Rhone-
repression of amidase enzymes in Poulenc Sante.
Aspergillus nidulans. J. Bacteriol., 103, 49 Petre, D., Cerbelaud, E., Mayaux, J.-F.,
482–487. and Yeh, P. (1991) Novel polypeptides, the
41 Cheong, T.K. and Oriel, P.J. (2000) DNA sequences allowing their
Cloning of a wide-spectrum amidase expression, method of preparation, and
from Bacillus stearothermophilus BR388 in utilization. EP 0,433,117 to Rhone-
Escherichia coli and marked enhancement Poulenc Sante.
of amidase expression using directed 50 Thiery, A., Maestracci, M., Arnaud, A.,
evolution. Enzyme Microb. Technol., 26, and Galzy, P. (1986) Acyltransferase
152–158. activity of the wide spectrum amidase of
632
Brevibacterium sp. R312. J. Gen. 58 Zheng, R.-C., Wang, Y.-S., Liu, Z.-Q.,
Microbiol., 132, 2205–2208. Xing, L.-Y., Zheng, Y.-G., and Shen, Y.-C.
51 Chan Kwo Chion, C.K.N., Duran, R., (2007) Isolation and characterization of
Arnaud, A., and Galzy, P. (1991) N- Delftia tsuruhatensis ZJB-05174, capable
Terminal amino acid sequence of of R-enantioselective degradation of
Brevibacterium sp. R312 wide-spectrum 2,2-dimethylcyclopropanecarboxamide.
amidase. Appl. Microbiol. Biotechnol., 36, Res. Microbiol., 158, 258–264.
205–207. 59 Makhongela, H.S., Glowacka, A.E.,
52 Eichhorn, E., Roduit, J.-P., Shaw, N.M., Agarkar, V.B., Sewell, B.T., Weber, B.,
Heinzmann, K., and Kiener, A. (1997) Cameron, R.A., Cowan, D.A., and
Preparation of (S)-piperazine-2- Burton, S.G. (2007) A novel
carboxylic acid (R)-piperazine-2- thermostable nitrilase superfamily
carboxylic acid, and (S)-pipecolic amidase from Geobacillus pallidus
acid by kinetic resolution of the showing acyl transfer activity.
corresponding racemic carboxamides Appl. Microbiol. Biotechnol., 75,
with stereoselective amidase in whole 801–811.
bacterial cells. Tetrahedron: Asymmetry, 8, 60 Skouloubris, S., Labigne, A., and De
2533–2536. Reuse, H. (1997) Identification and
53 Robins, K. and Gilligan, T. (1992) characterization of an aliphatic amidase
Biotechnological process for the in Helicobacter pylori. Mol. Microbiol., 25,
production of S-( þ )-2,2- 989–998.
dimethylcyclopropanecarboxamide and 61 Shaw, N.M., Naughton, A., Robins, K.,
R-()-2,2- Tinschert, A., Schmid, E., Hischier, M.-
dimethylcyclopropanecarboxylic acid. EP L., Venetz, V., Werlen, J., Zimmermann,
0,502,525 to Lonza AG. T., Brieden, W., de Riedmatten, P.,
54 Zimmermann, T., Robins, K., Roduit, J.-P., Zimmermann, B., and
Birch, O.M., and B€ohlen, E. (1993) Neum€ uller, R. (2002) Selection,
Process for the preparations of S-( þ )-2,2- purification, characterisation, and
dimethyl-cyclopropane carboxamid by cloning of a novel heat-stable
genetically engineered microorganisms. stereo-specific amidase from Klebsiella
EP 0,524,604 to Lonza AG. oxytoca, and its application in the
55 Hayashi, T., Yamamoto, K., Matsuo, A., synthesis of enantiomerically pure (R)-
Otsubo, K., Muramatsu, S., Matsuda, A., and (S)-3,3,3-trifluoro-2-hydroxy-2-
and Komatsu, K.I. (1997) methylpropionic acids and (S)-3,3,3-
Characterization and cloning of an trifluoro-2-hydroxy-2-
enantioselective amidase from methylpropionamide. Org. Process Res.
Comamonas acidovorans Dev., 6, 497–504.
KPO-2771-4. J. Ferment. Bioeng., 83, 62 Liese, A., Seelbach, K., and Wandrey, C.
139–145. (2006) Industrial Biotransformations,
56 Yamamoto, K., Otsubo, K., Matsuo, A., Wiley-VCH Verlag GmbH, Weinheim,
Hayashi, T., Fujimatsu, I., and pp. 377–383.
Komatsu, K.-I. (1996) Production of 63 Shaw, N.M., Robins, K.T., and Kiener, A.
R-(-)-ketoprofen from an amide (2003) Lonza: 20 years of
compound by Comamonas acidovorans biotransformations. Adv. Synth. Catal.,
KPO-2771-4. Appl. Environ. Microbiol., 62, 345, 425–435.
152–155. 64 Draper, P. (1967) The aliphatic acylamide
57 Tani, Y., Kurihara, M., and Nishise, H. amidohydrolase of Mycobacterium
(1989) Characterization of smegmatis: its inducible nature and
nitrile hydratase and amidase, relation to acyl-transfer to
which are responsible for the conversion hydroxylamine. J. Gen. Microbiol., 46,
of dinitriles to mononitriles, from 111–123.
Corynebacterium sp. Agric. Biol. Chem., 65 Doran, J.P., Duggan, P., Masterson, M.,
53, 3151–3158. Turner, P.D., and OReilly, C. (2005)
References j633
Expression and purification of a active site nucleophile of the
recombinant enantioselective amidase. catalytic mechanism. FEBS Lett., 367,
Protein Expr. Purif., 40, 190–196. 275–279.
66 Heumann, S., Eberl, A., 75 Komeda, H., Harada, H., Washika, S.,
Fischer-Colbrie, G., Pobeheim, H., Sakamoto, T., Ueda, M., and Asano, Y.
Kaufmann, F., Ribitsch, D., Cavaco- (2004) S-Stereoselective piperazine-2-
Paulo, A., and Guebitz, G.M. (2009) A tert-butylcarboxamide hydrolase from
novel aryl acylamidase from Nocardia Pseudomonas azotoformans IAM 1603 is a
farcinica hydrolyses novel L-amino acid amidase. Eur. J.
polyamide. Biotechnol. Bioeng., 102, Biochem., 271, 1465–1475.
1003–1011. 76 Ciskanik, L.M., Wilczek, J.M., and
67 Kelly, M. and Kornberg, H.L. (1962) Fallon, R.D. (1995) Purification and
Amidase from Pseudomonas aeruginosa: a characterization of an enantioselective
multi-headed enzyme. Biochim. Biophys. amidase from Pseudomonas chlororaphis
Acta, 64, 190–191. B23. Appl. Environ. Microbiol., 61,
68 Woods, M.J., Findlater, J.D., and 998–1003.
Orsi, B.A. (1979) Kinetic mechanism of 77 Nishiyama, M., Horinouchi, S.,
the aliphatic amidase from Pseudomonas Kobayashi, M., Nagasawa, T.,
aeruginosa. Biochim. Biophys. Acta, 567, Yamada, H., and Beppu, T. (1991)
225–237. Cloning and characterization of genes
69 Kelly, M. and Kornberg, H.L. (1964) responsible for metabolism of nitrile
Purification and properties of compounds from Pseudomonas
acyltransferases from Pseudomonas chlororaphis B23. J. Bacteriol., 173,
aeruginosa. Biochem. J., 93, 2465–2472.
557–566. 78 Egorova, K., Antranikian, G.,
70 Brown, J.E., Brown, P.R., and Clarke, Trauthwein, H., Verseck, S., and
P.H. (1969) Butyramide-utilizing Dingerdissen, U. (2004) Thermisch
mutants of Pseudomonas aeruginosa 8602 stabile Amidasen. DE 103,12,842 to
which produce an amidase with altered Degussa AG.
substrate specificity. J. Gen. Microbiol., 57, 79 Masutomo, S., Inoue, A., Kumagai, K.,
273–285. Murai, R., and Mitsuda, S. (1995)
71 Karmali, A., Pacheco, R., Tata, R., and Enantioselective hydrolysis of (RS)-2-
Brown, P. (2001) Substitutions of Thr- isopropyl-40 -chlorophenylacetonitrile by
103-Ile and Trp-138-Gly in amidase from Pseudomonas sp. B21C9. Biosci.
Pseudomonas aeruginosa are responsible Biotechnol. Biochem., 59, 720–722.
for altered kinetic properties and 80 Kagayama, T. and Ohe, T. (1990)
enzyme instability. Mol. Biotechnol., 17, Purification and properties of an
201–212. aromatic amidase from Pseudomonas sp.
72 Karmali, A., Tata, R., and Brown, P.R. GDI 211. Agric. Biol. Chem., 54,
(2000) Substitution of Glu-59 by Val in 2565–2571.
amidase from Pseudomonas aeruginosa 81 Komeda, H., Harada, H., Washika, S.,
results in a catalytically inactive enzyme. Sakamoto, T., Ueda, M., and Asano, Y.
Mol. Biotechnol., 16, 5–16. (2004) A novel R-stereoselective amidase
73 Ambler, R.P., Auffret, A.D., and from Pseudomonas sp. MCI3434 acting
Clarke, P.H. (1987) The amino acid on piperazine-2-tert-butylcarboxamide.
sequence of the aliphatic amidase from Eur. J. Biochem., 271, 1580–1590.
Pseudomonas aeruginosa. FEBS Lett., 215, 82 Komeda, H. and Asano, Y. (2005) A
285–290. DmpA-homologous protein from
74 Novo, C., Tata, R., Clemente, A., and Pseudomonas sp. is a dipeptidase specific
Brown, P.R. (1995) Pseudomonas for b-alanyl dipeptides. Eur. J. Biochem.,
aeruginosa aliphatic amidase is related to 272, 3075–3084.
the nitrilase/cyanide hydratase enzyme 83 Fallon, R.D., Stieglitz, B., and Turner JJr.,
family and Cys166 is predicted to be the I. (1997) A Pseudomonas putida capable of
634
stereoselective hydrolysis of nitriles. 92 Stolz, A., Trott, S., Binder, M., Bauer, R.,
Appl. Microbiol. Biotechnol., 47, 156–161. Hirrlinger, B., Layh, N., and Knackmuss,
84 Bracey, M.H., Hanson, M.A., Masuda, H.-J. (1998) Enantioselective nitrile
K.R., Stevens, R.C., and Cravatt, B.F. hydratases and amidases from different
(2002) Structural adaptations in a bacterial isolates. J. Mol. Catal. B: Enzym.,
membrane enzyme that terminates 5, 137–141.
endocannabinoid signaling. Science, 298, 93 Effenberger, F. and Graef, B.W. (1998)
1793–1796. Chemo- and enantioselective
85 Cravatt, B.F., Giang, D.K., Mayfield, S.P., hydrolysis of nitriles and acid amides,
Boger, D.L., Lerner, R.A., and Gilula, N.B. respectively, with resting cells of
(1996) Molecular characterization of an Rhodococcus sp. C3II and Rhodococcus
enzyme that degrades neuromodulatory erythropolis MP50. J. Biotechnol., 60,
fatty-acid amides. Nature, 384, 83–87. 165–174.
86 Kakeya, H., Sakai, N., Sugai, T., and 94 Xie, S.-X., Kato, Y., Komeda, H.,
Ohta, H. (1991) Microbial hydrolysis as a Yoshida, S., and Asano, Y. (2003) A gene
potent method for the preparation of cluster responsible for alkylaldoxime
optically active nitriles, amides and metabolism coexisting with nitrile
carboxylic acids. Tetrahedron Lett., 32, hydratase and amidase in Rhodococcus
1343–1346. globerulus A-4. Biochemistry, 42,
87 Martınkova, L., Kren, V., Cvak, L., 12056–12066.
Ovesna, M., and Prepechalova, I. (2000) 95 Yokoyama, M., Sugai, T., and Ohta, H.
Hydrolysis of lysergamide to lysergic acid (1993) Asymmetric hydrolysis of a
by Rhodococcus equi A4. J. Biotechnol., 84, disubstituted malononitrile by the aid of a
63–66. microorganism. Tetrahedron: Asymmetry,
88 Gilligan, T., Yamada, H., and 4, 1081–1084.
Nagasawa, T. (1993) Production of S- 96 Ohta, H. (1996) Stereochemistry of
( þ )-2-phenylpropionic acid from (R,S)- enzymatic hydrolysis of nitriles. Chimia,
2-phenylpropionitrile by the combination 50, 434–436.
of nitrile hydratase and stereoselective 97 Kobayashi, M., Fujiwara, Y., Goda, M.,
amidase in Rhodococcus equi TG328. Komeda, H., and Shimizu, S. (1997)
Appl. Microbiol. Biotechnol., Identification of active sites in amidase:
39, 720–725. evolutionary relationship between
89 Hirrlinger, B. and Knackmuss, H.-J. amide bond- and peptide
(1996) Purification and properties of an bond-cleaving enzymes. Proc. Natl. Acad.
amidase from Rhodococcus erythropolis Sci. U.S.A., 94, 11986–11991.
MP50 which enantioselectively 98 Kobayashi, M., Komeda, H., Nagasawa,
hydrolyzes 2-arylpropionamides. J. T., Nishiyama, M., Horinouchi, S.,
Bacteriol., 178, 3501–3507. Beppu, T., Yamada, H., and Shimizu, S.
90 Trott, S., B€
urger, S., Calaminus, C., and (1993) Amidase coupled with low-
Stolz, A. (2002) Cloning and molecular-mass nitrile hydratase from
heterologous expression of an Rhodococcus rhodochrous J1. Eur. J.
enantioselective amidase from Biochem., 217, 327–336.
Rhodococcus erythropolis strain MP50. 99 Kotlova, E.K., Chestukhina, G.G.,
Appl. Environ. Microbiol., Astaurova, O.B., Leonova, T.E.,
68, 3279–3286. Yanenko, A.S., and Debabov, V.G. (1999)
91 Hirrlinger, B., Stolz, A., and Isolation and primary characterization of
Knackmuss, H.-J. (1997) Enzymatic an amidase from Rhodococcus
synthesis of optically active hydroxamic rhodochrous. Biochemistry (Moscow), 64,
acids and their conversion into optically 384–389.
active primary amines by Lossen 100 Han, W.-W., Wang, Y., Zhou, Y.-H.,
transposition. WO 97/044480 to Yao, Y., Li, Z.-S., and Feng, Y. (2009)
Fraunhofer-Gesellschaft zur F€orderung Understanding structural/functional
der angewandten Forschung E.V. properties of amidase from Rhodococcus
References j635
erythropolis by computational approaches. 109 Joeres, U. and Kula, M.R. (1994)
J. Mol. Model., 15, 481–487. Purification and characterisation of a
101 Hashimoto, Y., Nishiyama, M., Ikehata, microbial L-carnitine amidase. Appl.
O., Horinouchi, S., and Beppu, T. (1991) Microbiol. Biotechnol., 40, 606–610.
Cloning and characterization of an 110 Kobayashi, M., Goda, M., and Shimizu, S.
amidase gene from Rhodococcus species (1999) Hydrazide synthesis: novel
N-774 and its expression in Escherichia substrate specificity of amidase.
coli. Biochim. Biophys. Acta, 1088, Biochem. Biophys. Res. Commun., 256,
225–233. 415–418.
102 Fournand, D., Bigey, F., and Arnaud, A. 111 Derbyshire, M.K., Karns, J.S.,
(1998) Acyl transfer activity of an amidase Kearney, P.C., and Nelson, J.O. (1987)
from Rhodococcus sp. strain R312: Purification and characterization of an
formation of a wide range of hydroxamic N -methylcarbamate pesticide
acids. Appl. Environ. Microbiol., 64, hydrolyzing enzyme. J. Agric. Food Chem.,
2844–2852. 35, 871–877.
103 Wang, M.-X., Lu, G., Ji, G.-J., 112 Karns, J.S. and Tomasek, P.H. (1991)
Huang, Z.-T., Meth-Cohn, O., and Carbofuran hydrolase - purification and
Colby, J. (2000) Enantioselective properties. J. Agric. Food Chem., 39,
biotransformations of racemic 1004–1008.
a-substituted phenylacetonitriles and 113 Hayatsu, M. and Nagata, T. (1993)
phenylacetamides using Rhodococcus sp. Purification and characterization of
AJ270. Tetrahedron: Asymmetry, 11, carbaryl hydrolase from Blastobacter sp.
1123–1135. strain M501. Appl. Environ. Microbiol., 59,
104 Snell, D., and Colby, J. (1999) 2121–2125.
Enantioselective hydrolysis of racemic 114 Chapalmadugu, S. and Chaudhry, G.R.
ibuprofen amide to S-( þ )-ibuprofen by (1993) Isolation of a constitutively
Rhodococcus AJ270. Enzyme Microb. expressed enzyme for hydrolysis of
Technol., 24, 160–163. carbaryl in Pseudomonas aeruginosa.
105 Toogood, H.S., Brown, R.C., Line, K., J. Bacteriol., 175, 6711–6716.
Keene, P.A., Taylor, S.J.C., McCague, R., 115 Mulbry, W.W. and Eaton, R.W.
and Littlechild, J.A. (2004) (1991) Purification and characterization
The use of a thermostable signature of the N-methylcarbamate hydrolase
amidase in the resolution of the bicyclic from Pseudomonas strain CRL-OK.
synthon (rac)-c-lactam. Tetrahedron, 60, Appl. Environ. Microbiol., 57,
711–716. 3679–3682.
106 DAbusco, A.S., Ammendola, S., 116 Martınkova, L., Vejvoda, V., and Kren, V.
Scandurra, R., and Politi, L. (2001) (2008) Selection and screening for
Molecular and biochemical enzymes of nitrile metabolism.
characterization of the recombinant J. Biotechnol., 133, 318–326.
amidase from hyperthermophilic 117 Jin, S.-J., Zheng, R.-C., Zheng, Y.-G., and
archaeon Sulfolobus solfataricus. Shen, Y.-C. (2008) R-enantioselective
Extremophiles, 5, 183–192. hydrolysis of 2,2-
107 Joeres, U. and Kula, M.R. (1994) dimethylcyclopropanecarboxamide by
Screening for a novel enzyme amidase from a newly isolated
hydrolysing L-carnitine amide. strain Brevibacterium epidermidis ZJB-
Appl. Microbiol. Biotechnol., 07021. J. Appl. Microbiol., 105,
40, 599–605. 1150–1157.
108 Joeres, U., Bommarius, A.S., 118 Liese, A., Seelbach, K., and Wandrey, C.
Th€ommes, J., and Kula, M.-R. (1995) (2006) Industrial Biotransformations,
Studies on the kinetics and application of Wiley-VCH Verlag GmbH, Weinheim,
L-carnitine amidase for the production of pp. 424–426.
L-carnitine. Biocatal. Biotransform., 12, 119 Fukumura, T. (1977) Conversion of D- and
27–36. Dl-a-amino-e-caprolactam into L-lysine
636
using both yeast cells and bacterial cells. fold-type I racemase, a-amino-
Agric. Biol. Chem., 41, 1327–1330. e-caprolactam racemase from
120 Fukumura, T. (1976) Screening, Achromobacter obae. Biochemistry, 48,
classification and distribution of L- 941–950.
a-amino-e-caprolactam-hydrolyzing 131 Holt-Tiffin, K.E. (2009) ( þ )- and ()-2-
yeasts. Agric. Biol. Chem., 40, azabicyclo[2.2.1]hept-5-en-3-one.
1687–1693. Extremely useful synthons. Chim. Oggi,
121 Fukumura, T., Talbot, G., Misono, H., 27, 23–25.
Teramura, Y., Kato, K., and Soda, K. (1978) 132 Liese, A., Seelbach, K., and Wandrey, C.
Purification and properties of a novel (2006) Industrial Biotransformations,
enzyme, L-a-amino-e-caprolactamase Wiley-VCH Verlag GmbH, Weinheim,
from Cryptococcus laurentii. FEBS Lett., pp. 420–423.
89, 298–300. 133 Taylor, S.J.C., Sutherland, A.G., Lee, C.,
122 Fukumura, T. (1976) Hydrolysis of Wisdom, R., Thomas, S., Roberts, S.M.,
L-a-amino-e-caprolactam by yeasts. Agric. and Evans, C. (1990) Chemoenzymatic
Biol. Chem., 40, 1695–1698. synthesis of ()-carbovir utilizing a whole
123 Ahmed, S.A., Esaki, N., and Soda, K. cell catalysed resolution of 2-azabicyclo
(1982) Purification and properties of [2.2.1]hept-5-en-3-one. J. Chem. Soc.,
a-amino-e-caprolactam racemase from Chem. Commun., 1120–1121.
Achromobacter obae. FEBS Lett., 150, 134 Evans, C.T. and Roberts, S.M. (1991)
370–374. Chiral compounds. EP 0,424,064 to
124 Ahmed, S.A., Esaki, N., Tanaka, H., and Enzymatix Ltd.
Soda, K. (1982) Production and 135 Taylor, S.J., McCague, R., Wisdom, R.,
stabilization of a-amino-e-caprolactam Lee, C., Dickson, K., Ruecroft, G.,
racemase from Achromobacter OBrien, F., Littlechild, J., Bevan, J.,
obae. Bull. Inst. Chem. Res., Kyoto Univ., Roberts, S.M., and Evans, C.T. (1993)
60, 342–346. Development of the biocatalytic
125 Ahmed, S.A., Esaki, N., Tanaka, H., and resolution of 2-azabicyclo[2.2.1]hept-
Soda, K. (1983) Properties of a-amino- 5-en-3-one as an entry to
e-caprolactam racemase from single-enantiomer carbocyclic
Achromobacter obae. Agric. Biol. Chem., 47, nucleosides. Tetrahedron: Asymmetry, 4,
1887–1893. 1117–1128.
126 Ahmed, S.A., Esaki, N., Tanaka, H., and 136 McCague, R. and Taylor, S.J.C. (1997)
Soda, K. (1986) Mechanism of a-amino- Four case studies in the development of
e-caprolactam racemase reaction. biotransformation-based processes, in
Biochemistry, 25, 385–388. Chirality in Industry II (eds A.N. Collins,
127 Fukumura, T. (1977) Partial purification G.N. Sheldrake, and J. Crosby), John
and some properties of a-amino- Wiley & Sons Ltd., Chichester, pp.
e-caprolactam-racemizing enzyme from 184–206.
Achromobacter obae. Agric. Biol. Chem., 41, 137 Connelly, F.S., Line, K., Isupov, M.N., and
1509–1510. Littlechild, J.A. (2005) Synthesis and
128 Fukumura, T. (1977) Bacterial characterisation of a ligand that forms a
racemization of a-amino-e-caprolactam. stable tetrahedral intermediate in the
Agric. Biol. Chem., 41, 1321–1325. active site of the Aureobacterium species
129 Naoko, N., Oshihara, W., and Yanai, A. () c-lactamase. Org. Biomol. Chem., 3,
(1987) a-Amino-e-caprolactam racemase 3260–3262.
for L-lysine production, in Biochemistry of 138 Line, K., Isupov, M.N., and
vitamin B6, Birkh€auser Verlag, Basel, pp. Littlechild, J.A. (2004) The crystal
449–452. structure of a () c-lactamase from an
130 Okazaki, S., Suzuki, A., Mizushima, T., Aureobacterium species reveals a
Kawano, T., Komeda, H., Asano, Y., and tetrahedral intermediate in the
Yamane, T. (2009) The novel structure of a active site. J. Mol. Biol.,
pyridoxal 5-phosphate-dependent 338, 519–532.
References j637
139 Brabban, A.D., Littlechild, J., and Azabicycloheptanone. WO9218477 to
Wisdom, R. (1996) Stereospecific Chiros Ltd.
c-lactamase activity in a Pseudomonas 149 Lloyd, M., Lloyd, R., Keene, P., and
fluorescens species. J. Ind. Microbiol. Osborne, A. (2007) A concise synthesis of
Biotechnol., 16, 8–14. single-enantiomer b-lactams and
140 Joshi, R.R., Prabhune, A.A., Joshi, R.A., b-amino acids using Rhodococcus
and Gurjar, M.K. (2003) Process for the globerulus. J. Chem. Technol. Biotechnol.,
preparation of optically active azabicyclo 82, 1099–1106.
heptanone derivatives. EP 1,348,765 to 150 Taylor, S.J.C. and Keene, P.A. (2000)
Council of Scientific and Industrial Biocatalyst and its use in enzymatic
Research. resolution of racemic beta-lactams. WO
141 Taylor, S.J.C., Brown, R.C., Keene, P.A., 2000/58283 to Chirotech Technology
and Taylor, I.N. (1999) Novel screening Limited.
methods - The key to cloning 151 Lloyd, R.C., Lloyd, M.C., Smith, M.E.B.,
commercially successful biocatalysts. Holt, K.E., Swift, J.P., Keene, P.A.,
Bioorg. Med. Chem., 7, 2163–2168. Taylor, S.J.C., and McCague, R. (2004)
142 Wyborn, N.R., Scherr, D.J., and Use of hydrolases for the synthesis of
Jones, C.W. (1994) Purification, cyclic amino acids. Tetrahedron, 60,
properties and heterologous expression 717–728.
of formamidase from Methylophilus 152 Staudenmaier, H.R., Hauer, B.,
methylotrophus. Microbiology, 140, Balkenhohl, F., Ladner, W., Schnell, U.,
191–195. and Pressler, U. (1995) Verfahren zur
143 Wyborn, N.R., Mills, J., Williams, S.G., Herstellung von enantiomerenreinen
and Jones, C.W. (1996) Molecular Lactamen. EP 0,687,736 to BASF
characterisation of formamidase from Aktiengesellschaft.
Methylophilus methylotrophus. 153 Leuchtenberger, W., Huthmacher, K., and
Eur. J. Biochem., 240, 314–322. Drauz, K. (2005) Biotechnological
144 Mahenthiralingam, E., Draper, P., production of amino acids and
Davis, E.O., and Colston, M.J. (1993) derivatives: current status and
Cloning and sequencing of the gene prospects. Appl. Microbiol. Biotechnol.,
which encodes the highly inducible 69, 1–8.
acetamidase of Mycobacterium 154 Bruggink, A. and Roy, P.D. (2001)
smegmatis. J. Gen. Microbiol., 139 (Pt 3), Industrial synthesis of semisynthetic
575–583. antibiotics, in Synthesis of b-Lactam
145 Gonsalvez, I.S., Isupov, M.N., and Antibiotics - Chemistry, Biocatalysis &
Littlechild, J.A. (2001) Crystallization and Process Integration (ed. A. Bruggink),
preliminary X-ray analysis of a Kluwer Academic Publishers, Dordrecht,
c-lactamase. Acta Crystallogr. Sect. D, 57, pp. 12–54.
284–286. 155 Wegman, F.M.A., Janssen, M.H.A.,
146 Hickey, A.M., Ngamsom, B., Wiles, C., van Rantwijk, F., and Sheldon, R.A.
Greenway, G.M., Watts, P., and (2001) Towards biocatalytic synthesis of
Littlechild, J.A. (2009) A microreactor for b-lactam antibiotics. Adv. Synth. Catal.,
the study of biotransformations by a 343, 559–576.
cross-linked c-lactamase enzyme. 156 Henrick, C.A. and Garcia, B.A. (1981)
Biotechnol. J., 4, 510–516. Esters and thiolesters of amino acids,
147 Evans, C., McCague, R., Roberts, S.M., processes for their production, and
Sutherland, A.G., and Wisdom, R. (1991) compositions including them. GB
Whole cell catalysed kinetic resolution of 1,588,111 to Zoecon Corporation.
6-azabicyclo[3.2.0]hept-3-en-7-one: 157 Baxter, A.D., Bhogal, R., Bird, J., Keily,
synthesis of ()-cispentacin (FR 109615). J.F., Manallack, D.T., Montana, J.G.,
J. Chem. Soc., Perkin Trans. 1, 2276–2277. Owen, D.A., Pitt, W.R., Watson, R.J., and
148 Evans, C.T., Roberts, S.M., and Wills, R.E. (2001) Arylsulphonyl
Sutherland, A.G. (1992) hydroxamic acids: potent and selective
638
matrix metalloproteinase inhibitors. Wiley & Sons Ltd., Chichester, pp.

Bioorg. Med. Chem. Lett., 11, 1465–1468. 187–208.
158 Bommarius, A.S., Schwarm, M., 167 Kamphuis, F.J., Boesten, W.H.J.,
Stingl, K., Kottenhahn, M., Huthmacher, Broxterman, Q.B., Hermes, H.F.M.,
K., and Drauz, K. (1995) Synthesis and van Balken, J.A.M., Meijer, E.M., and
use of enantiomerically pure Schoemaker, H.E. (1990)
tert-leucine. Tetrahedron: Asymmetry, 6, New developments in the chemo-
2851–2888. enzymatic production of amino acids.
159 Faucher, A.-M., Bailey, M.D., Adv. Biochem. Eng./Biotechnol., 42,
Beaulieu, P.L., Brochu, C., Duceppe, J.-S., 133–186.
Ferland, J.-M., Ghiro, E., Gorys, V., 168 Kamphuis, J., Meijer, E.M., Boesten,
Halmos, T., Kawai, S.H., Poirier, M., W.H.J., Broxterman, Q.B., Kaptein, B.,
Simoneau, B., Tsantrizos, Y.S., and Hermes, H.F.M., and Schoemaker, H.E.
Llinas-Brunet, M. (2004) Synthesis of (1992) Production of natural and
BILN 2061, an HCV NS3 protease synthetic L- and D-amino acids by
inhibitor with proven antiviral effect in aminopeptidases and amino amidases, in
humans. Org. Lett., 6, 2901–2904. Biocatalytic Production of Amino Acids and
160 Kleemann, A., Engel, J., Reichert, D., and Derivatives (eds J.D. Rozzell and F.
Kutscher, B. (1999) Pharmaceutical Wagner), Hanser Publishers, M€ unchen,
Substances: Syntheses, Patents, pp. 177–206.
Applications, Thieme, Stuttgart. 169 Boesten, F.W.H.J. (1977) Process for
161 Reinhold, D.F., Firestone, R.A., preparing a-amino-acid amides. GB
Gaines, W.A., Chemerda, J.M., and 1,548,032 to DSM/Stamicarbon B.V.
Sletzinger, M. (1968) Synthesis of L- 170 Bommarius, A.S., Schwarm, M., and
a-methyldopa from asymmetric Drauz, K. (2001) Comparison of different
intermediates. J. Org. Chem., 33, chemoenzymatic process routes to
1209–1213. enantiomerically pure amino acids.
162 Stepek, W.J. and Nigro, M.M. (1988) Chimia, 55, 50–59.
Novel process for the preparation of 171 Chibata, I., Tosa, T., and Shibatani, T.
aminonitriles useful for the preparation (1992) The industrial production of
of herbicides. EP 0,123,830 to American optically active compounds by
Cyanamid Company. immobilized biocatalysts, in
163 Tomlin, C.D.S. (2003) The Pesticide Chirality in Industry (eds A.N. Collins,
Manual, British Crop Protection Council G.N. Sheldrake, and J. Crosby),
(BCPC), Alton, Hampshire, UK. John Wiley & Sons, Chichester, pp.
164 Genix, P., Guesnet, J.-L., and Lacroix, G. 351–370.
(2003) Chemistry and stereo-chemistry of 172 Kjær, A. and Wagner, S. (1955) A
fenamidone. Pflanz.-Nachrichten Bayer, convenient synthesis of Dl-
56, 421–434. homomethionine (5-
165 Sonke, T., Kaptein, B., and Schoemaker, methylthionorvaline). Acta Chem. Scand.,
H.E. (2009) Use of enzymes in the 9, 721–726.
synthesis of amino acids, in Amino Acids, 173 Hyett, D.J., Didone, M., Milcent, T.J.A.,
Peptides and Proteins in Organic Chemistry, Broxterman, Q.B., and Kaptein, B. (2006)
1st edn (ed. A.B. Hughes), Wiley-VCH A new method for the preparation of
Verlag GmbH, Weinheim, pp. 79–117. functionalized unnatural a-H-a-amino
166 Kamphuis, F.J., Boesten, W.H.J., acid derivatives. Tetrahedron Lett., 47,
Kaptein, B., Hermes, H.F.M., Sonke, T., 7771–7774.
Broxterman, Q.B., Van den Tweel, W.J.J., 174 Ikeda, M. (2003) Amino acid production
and Schoemaker, H.E. (1992) processes. Adv. Biochem. Eng./Biotechnol.,
The production and uses of optically pure 79, 1–35.
natural and unnatural amino acids, in 175 Boesten, W.H.J., Schoemaker, H.E., and
Chirality in Industry (eds A.N. Collins, Dassen, B.H.N. (1986) Process for
G.N. Sheldrake, and J. Crosby), John racemizing an optically active
References j639
N-benzylidene amino-acid amide. EP extreme thermophilic archaebacterium
0,199,407 to Stamicarbon B.V. Sulfolobus solfataricus. BBA-Gen. Subjects,
176 Asano, Y. and Yamaguchi, S. (2005) 1033, 148–153.
Dynamic kinetic resolution of 184 Ratnayake, S., Selvarkumar, P., and
amino acid amide catalyzed by Hayashi, K. (2003) A putative proline
D-aminopeptidase and a-amino- iminopeptidase of Thermotoga maritima
e-caprolactam racemase. J. Am. Chem. is a leucine aminopeptidase with lysine-p-
Soc., 127, 7696–7697. nitroanilide hydrolyzing activity. Enzyme
177 Asano, Y. and Yamaguchi, S. (2005) Microb. Technol., 32, 414–421.
Discovery of amino acid amides as 185 Kuo, L.-Y., Hwang, G.-Y., Lai, Y.-J.,
new substrates for a-amino- Yang, S.-L., and Lin, L.-L. (2003)
e-caprolactam racemase from Overexpression, purification, and
Achromobacter obae. J. Mol. Catal. B: characterization of the recombinant
Enzym., 36, 22–29. leucine aminopeptidase II of Bacillus
178 Yamaguchi, S., Komeda, H., and stearothermophilus. Curr. Microbiol., 47,
Asano, Y. (2007) New enzymatic method 40–45.
of chiral amino acid synthesis by dynamic 186 Lin, L.-L., Hsu, W.-H., Wu, C.-P.,
kinetic resolution of amino acid amides: Chi, M.-C., Chou, W.-M., and Hu, H.-Y.
use of stereoselective amino acid (2004) A thermostable leucine
amidases in the presence of a-amino- aminopeptidase from Bacillus
e-caprolactam racemase. Appl. Environ. kaustophilus CCRC 11223. Extremophiles,
Microbiol., 73, 5370–5373. 8, 79–87.
179 Boesten, W.H.J., Raemakers-Franken, 187 Chi, M.-C., Ong, P.-L., Hsu, W.-H.,
P.C., Sonke, T., Euverink, G.J.W., and Chen, Y.-H., Huang, H.-B., and Lin, L.-L.
Grijpstra, P. (2003) Polypeptides having (2008) Role of the invariant Asn345 and
a-H-a-amino acid amide racemase Asn435 residues in a leucine
activity and nucleic acids encoding the aminopeptidase from Bacillus
same. WO 03/106691 to DSM IP Assets kaustophilus as evaluated by site-directed
B.V. mutagenesis. Int. J. Biol. Macromol., 43,
180 Sonke, T., Duchateau, A.L.L., Schipper, 481–487.
D., Euverink, G.J.W., van der Wal, Sj., 188 Chi, M.-C., Liu, J.-S., Wang, W.-C.,
Henderickx, H.J.W., Bezemer, R., and Lin, L.-L., and Huang, H.-B. (2008)
Vollebregt, A. (2006) Industrial Site-directed mutagenesis of the
perspectives on assays, in Enzyme Assays - conserved Ala348 and Gly350 residues at
High-Throughput Screening, Genetic the putative active site of Bacillus
Selection and Fingerprinting, 1st edn kaustophilus leucine aminopeptidase.
(ed. J.-L. Reymond), Wiley-VCH Verlag Biochimie, 90, 811–819.
GmbH, Weinheim, pp. 95–135. 189 Chi, M.-C., Huang, H.-B., Liu, J.-S.,
181 Sonke, F.T. (2008) Novel developments in Wang, W.-C., Liang, W.-C., and Lin, L.-L.
the chemo-enzymatic synthesis of (2006) Residues threonine 346 and
enantiopure a-hydrogen- and leucine 352 are critical for the proper
a,a-disubstituted a-amino acids and function of Bacillus kaustophilus leucine
derivatives, Ph.D. thesis, University of aminopeptidase. FEMS Microbiol. Lett.,
Amsterdam, The Netherlands. 260, 156–161.
182 Greenstein, J.P. and Winitz, M. (1961) 190 Chi, M.-C., Chou, W.-M., Wang, C.-H.,
Enzymes involved in the determination, Chen, W., Hsu, W.-H., and Lin, L.-L.
characterization, and preparation of the (2004) Generating oxidation-resistant
amino acids, in Chemistry of the Amino variants of Bacillus kaustophilus leucine
Acids, John Wiley & Sons, Inc., New York, aminopeptidase by substitution of the
pp. 1753–1816. critical methionine residues with leucine.
183 Hanner, M., Redl, B., and St€offler, G. Antonie Van Leeuwen., 86, 355–362.
(1990) Isolation and characterization of 191 Deejing, S., Yoshimune, K., Lumyong, S.,
an intracellular aminopeptidase from the and Moriguchi, M. (2005) Purification
640
and characterization of 198 Morty, R.E. and Morehead, J. (2002)

hyperthermotolerant leucine Cloning and characterization of a leucyl
aminopeptidase from Geobacillus aminopeptidase from three pathogenic
thermoleovorans 47b. J. Ind. Microbiol. Leishmania species. J. Biol. Chem., 277,
Biotechnol., 32, 269–276. 26057–26065.
192 Khan, A.R., Nirasawa, S., Kaneko, S., 199 Kieny-LHomme, M.-P., Arnaud, A., and
Shimonishi, T., and Hayashi, K. (2000) Galzy, P. (1981) Etude dune L-
Characterization of a solvent resistant a-aminoamidase particulaire de
and thermostable aminopeptidase from Brevibacterium sp. en vue de lobtention
the hyperthermophilic bacterium, dacides a-amines optiquement actifs.
Aquifex aeolicus. Enzyme Microb. Technol., J. Gen. Appl. Microbiol., 27, 307–325.
27, 83–88. 200 Jallageas, J.C., Arnaud, A., and Galzy, P.
193 Gaur, R., Grover, T., Sharma, R., (1979) Remarques sur le spectre dactivite
Kapoor, S., and Khare, S.K. (2010) amidasique dun mutant de
Purification and characterization of a Brevibacterium. C. R. Acad. Sci. Paris D.,
solvent stable aminopeptidase from 288, 655–658.
Pseudomonas aeruginosa: cloning and 201 Thiery, A., Maestracci, M., Arnaud, A.,
analysis of aminopeptidase gene Galzy, P., and Nicolas, M. (1986)
conferring solvent stability. Process Purification and properties of an
Biochem., 45, 757–764. acylamide amidohydrolase (E.C. 3.5.1.4)
194 Cahan, R., Axelrad, I., Safrin, M., with a wide activity spectrum from
Ohman, D.E., and Kessler, E. (2001) A Brevibacterium sp. R 312. J. Basic
secreted aminopeptidase of Pseudomonas Microbiol., 26, 299–311.
aeruginosa. J. Biol. Chem., 276, 202 Bigey, F., Chebrou, H., Fournand, D., and
43645–43652. Arnaud, A. (1999) Transcriptional
195 Sarnovsky, R., Rea, J., Makowski, M., analysis of the nitrile-degrading
Hertle, R., Kelly, C., Antignani, A., operon from Rhodococcus sp. ACV2 and
Pastrana, D.V., and FitzGerald, D.J. high level production of recombinant
(2009) Proteolytic cleavage of a C- amidase with an Escherichia coli – T7
terminal prosequence, leading to expression system. J. Appl. Microbiol., 86,
autoprocessing at the N terminus, 752–760.
activates leucine aminopeptidase from 203 Aebischer, B., Frey, P., Haerter, H.-P.,
Pseudomonas aeruginosa. J. Biol. Chem., Herrling, P.L., Mueller, W.,
284, 10243–10253. Olverman, H.J., and Watkins, J.C. (1989)
196 Maric, S., Donnelly, S.M., Robinson, Synthesis and NMD A antagonistic
M.W., Skinner-Adams, T., Trenholme, properties of the enantiomers of
K.R., Gardiner, D.L., Dalton, J.P., 4-(3-phosphonopropyl)piperazine-2-
Stack, C.M., and Lowther, J. (2009) carboxylic acid (CPP) and of the
The M17 leucine aminopeptidase of the unsaturated analogue
malaria parasite Plasmodium (E)-4-(3-phosphonoprop-2-enyl)
falciparum: importance of active site piperazine-2-carboxylic acid (CPP-ene).
metal ions in the binding of substrates Helv. Chim. Acta, 72, 1043–1051.
and inhibitors. Biochemistry, 48, 204 Shiraiwa, T., Shinjo, K., and
5435–5439. Kurokawa, H. (1991) Asymmetric
197 Stack, C.M., Lowther, J., Cunningham, E., transformations of proline and 2-
Donnelly, S., Gardiner, D.L., piperidinecarboxylic acid via
Trenholme, K.R., Skinner-Adams, T.S., formation of salts with optically active
Teuscher, F., Grembecka, J., Mucha, A., tartaric acid. Bull. Chem. Soc. Jpn., 64,
Kafarski, P., Lua, L., Bell, A., and 3251–3255.
Dalton, J.P. (2007) Characterization of the 205 Uozumi, Y., Tanahashi, A., and
Plasmodium falciparum M17 leucyl Hayashi, T. (1993) Catalytic asymmetric
aminopeptidase. J. Biol. Chem., 282, construction of morpholines and
2069–2080. piperazines by palladium-catalyzed
References j641
tandem allylic substitution reactions. biocatalytic transformations. J. Mol.
J. Org. Chem., 58, 6826–6832. Catal. B: Enzym., 59, 106–110.
206 Bruce, M.A., St. Laurent, D.R., 213 Schechter, I. and Berger, A. (1967) On the
Poindexter, G.S., Monkovic, I., Huang, S., size of the active site in proteases. I.
and Balasubramanian, N. (1995) Kinetic Papain. Biochem. Biophys. Res. Commun.,
resolution of piperazine-2-carboxamide 27, 157–162.
by leucine aminopeptidase. An 214 Tishinov, K., Stambolieva, N., Petrova, S.,
application in the synthesis of the Galunsky, B., and Nedkov, P. (2009)
nucleoside transport blocker Purification and characterization of the
()-draflazine. Synth. Commun., 25, sunflower seed (Helianthus annuus L.)
2673–2684. major aminopeptidase. Acta Physiol.
207 Eichhorn, E., Roduit, J.-P., Shaw, N., Plant., 31, 199–205.
Heinzmann, K., and Kiener, A. 215 Boesten, W.H.J. and Meyer-Hoffman,
(1997) Preparation of (S)-piperazine- L.R.M. (1978) Process of preparing L- and
2-carboxylic acid (R)-piperazine- D-a-amino acids by enzyme treatment of
2-carboxylic acid, and (S)-piperidine- DL-a-amino acid amide. US 4,080,259 to
2-carboxylic acid by kinetic resolution of Novo Industri A/S.
the corresponding racemic 216 Shadid, B., van der Plas, H.C.,
carboxamides with stereoselective Boesten, W.H.J., Kamphuis, J.,
amidases in whole bacterial Meijer, E.M., and Schoemaker, H.E.
cells. Tetrahedron: Asymmetry, 8, (1990) The synthesis of L () and
2533–2536. D ( þ ) lupinic acid. Tetrahedron, 46,
208 Kiener, A., Roduit, J.-P., Kohr, J., and 913–920.
Shaw, N. (1995) Biotechnologisches 217 Rutjes, F.P.J.T. and Schoemaker, H.E.
verfahren zur herstellung von cyclischen (1997) Ruthenium-catalyzed ring closing
(S)-a-aminocarbons€auren und (R)- olefin metathesis of non-natural a -amino
a-aminocarbons€aureamiden. EP acids. Tetrahedron Lett., 38, 677–680.
0,686,698 to Lonza AG. 218 Wolf, L.B., Tjen, K.C.M.F., Rutjes,
209 Kiener, A., Roduit, J.-P., and F.P.J.T., Hiemstra, H., and Schoemaker,
Heinzmann, K. (1996) H.E. (1998) Pd-Catalyzed cyclization
Biotechnologisches verfahren zur reactions of acetylene-containing
herstellung von (R)- a-amino acids. Tetrahedron Lett., 39,
a-piperazincarbons€aure und (S)- 5081–5084.
a-piperazincarbons€aureamid. WO 96/ 219 Hermes, H.F.M., Sonke, T., Peters,
35775 to Lonza AG. P.J.H., van Balken, J.A.M., Kamphuis, J.,
210 Petersen, M. and Kiener, A. (1999) Dijkhuizen, L., and Meijer, E.M. (1993)
Biocatalysis: preparation and Purification and characterization of an L-
functionalization of N-heterocycles. aminopeptidase from Pseudomonas
Green Chem., 1, 99–106. putida ATCC 12633. Appl. Environ.
211 Komeda, H., Hariyama, N., and Microbiol., 59, 4330–4334.
Asano, Y. (2006) L-Stereoselective 220 Sonke, T., Kaptein, B., Boesten, W.H.J.,
amino acid amidase with broad Broxterman, Q.B., Kamphuis, J.,
substrate specificity from Brevundimonas Formaggio, F., Toniolo, C., Rutjes,
diminuta: characterization of a new F.P.J.T., and Schoemaker, H.E. (2000)
member of the leucine aminopeptidase Aminoamidase-catalyzed preparation
family. Applied Microbiol. Biotechnol., 70, and further transformations of
412–421. enantiopure a-hydrogen- and
212 Tishinov, K., Bayryamov, S., a,a-disubstituted a-amino acids, in
Nedkov, P., Stambolieva, N., and Stereoselective Biocatalysis
Galunsky, B. (2009) A highly (ed. R.N. Patel), Marcel Dekker, Inc.,
enantioselective aminopeptidase from New York, pp. 23–58.
sunflower seed-Kinetic studies, 221 Kale, F.A., Pijning, T., Sonke, T., Dijkstra,
substrate mapping and application to B.W., and Thunnissen, A.-M.W.H. (2010)
642
Crystal structure of the leucine 232 Gu, Y.-Q. and Walling, L.L. (2000)
aminopeptidase from Pseudomonas Specificity of the wound-induced leucine
putida reveals the molecular aminopeptidase (LAP-A) of tomato:
basis for its enantioselectivity and broad activity on dipeptide and tripeptide
substrate specificity. J. Mol. Biol., 398, substrates. Eur. J. Biochem., 267,
703–714. 1178–1187.
222 Wolf, L.B., Sonke, T., Tjen, K.C.M.F., 233 Burley, S.K., David, P.R., Sweet, R.M.,
Kaptein, B., Broxterman, Q.B., Taylor, A., and Lipscomb, W.N. (1992)
Schoemaker, H.E., and Rutjes, F.P.J.T. Structure determination and refinement
(2001) A biocatalytic route to of bovine lens leucine aminopeptidase
enantiomerically pure unsaturated a-H- and its complex with bestatin. J. Mol.
a-amino acids. Adv. Synth. Catal., 343, Biol., 224, 113–140.
662–674. 234 Str€ater, N. and Lipscomb, W.N. (2004)
223 van Hest, J.C., Kiick, K.L., and Leucyl aminopeptidase (animal), in
Tirrell, D.A. (2000) Efficient Handbook of Proteolytic Enzymes, 2nd edn
incorporation of unsaturated methionine (eds A.J. Barrett, N.D. Rawlings, and J.F.
analogues into proteins in vivo. J. Am. Woessner), Elsevier Academic Press,
Chem. Soc., 122, 1282–1288. London, pp. 896–901.
224 Taylor, A. (1993) Aminopeptidases: 235 Colloms, S.D. (2004) Leucyl
structure and function. FASEB J., 7, aminopeptidase PepA, in Handbook of
290–298. Proteolytic Enzymes, 2nd edn (eds A.J.
225 Smith, E.L. and Hill, R.L. (1960) Leucine Barrett, N.D. Rawlings, and J.F.
aminopeptidase, in The Enzymes, 2nd edn Woessner), Elsevier Academic Press,
(eds P.D. Boyer, H.A. Lardy, and K. London, pp. 905–910.
Myrback), Academic Press, vol. 4, Part A, 236 Kraft, M., Schleberger, C., Weckesser, J.,
pp. 37–62. and Schulz, G.E. (2006) Binding structure
226 Delange, R.J. and Smith, E.L. (1971) of the leucine aminopeptidase inhibitor
Leucine aminopeptidase and other N- microginin FR1. FEBS Lett., 580,
terminal exopeptidases, in The Enzymes, 6943–6947.
(ed. P.D. Boyer), Academic Press, New 237 Str€ater, N. and Lipscomb, W.N. (1995)
York, vol. III, pp. 81–118. Two-metal ion mechanism of bovine lens
227 Hanson, H. and Frohne, M. (1976) leucine aminopeptidase: active site
Crystalline leucine aminopeptidase solvent structure and binding mode of
from lens (a-aminoacyl-peptide L-leucinal, a gem-diolate transition state
hydrolase; EC 3.4.11.1). Methods analog, by X-ray crystallography.
Enzymol., 45, 504–521. Biochemistry, 34, 14792–14800.
228 Taylor, A. (1993) Aminopeptidases: 238 Carpenter, F.H. and Vahl, J.M. (1973)
towards a mechanism of action. Trends Leucine aminopeptidase (bovine lens).
Biochem. Sci., 18, 167–171. Mechanism of activation by Mg2 þ and
229 Gonzales, T. and Robert-Baudouy, J. Mn2 þ of the zinc metalloenzyme,
(1996) Bacterial aminopeptidases: amino acid composition, and sulfhydryl
properties and functions. FEMS content. J. Biol. Chem., 248, 294–304.
Microbiol. Rev., 18, 319–344. 239 Thompson, G.A. and Carpenter, F.H.
230 Gu, Y.-Q., Holzer, F.M., and Walling, L.L. (1976) Leucine aminopeptidase
(1999) Overexpression, purification and (bovine lens). Effect of pH on the relative
biochemical characterization of the binding of Zn2 þ and Mg2 þ to and on
wound-induced leucine activation of the enzyme. J. Biol. Chem.,
aminopeptidase of tomato. Eur. J. 251, 53–60.
Biochem., 263, 726–735. 240 Allen, M.P., Yamada, A.H., and
231 Herbers, K., Prat, S., and Willmitzer, L. Carpenter, F.H. (1983) Kinetic
(1994) Functional analysis of a leucine parameters of metal-substituted leucine
aminopeptidase from Solanum tuberosum aminopeptidase from bovine lens.
L. Planta, 194, 230–240. Biochemistry, 22, 3778–3783.
References j643
241 Thompson, G.A. and Carpenter, F.H. leucine aminopeptidase. EMBO J., 8,
(1976) Leucine aminopeptidase 1623–1627.
(bovine lens). The relative binding of 249 Summers, D.K. and Sherratt, D.J.
cobalt and zinc to leucine aminopeptidase (1984) Multimerization of high copy
and the effect of cobalt substitution on number plasmids causes instability:
specific activity. J. Biol. Chem., 251, ColE1 encodes a determinant
1618–1624. essential for plasmid
242 Cappiello, M., Alterio, V., Amodeo, P., monomerization and stability. Cell, 36,
Del Corso, A., Scaloni, A., Pedone, C., 1097–1103.
Moschini, R., De Donatis, G.M., De 250 Colloms, S.D., McCulloch, R., Grant, K.,
Simone, G., and Mura, U. (2006) Metal Neilson, L., and Sherratt, D.J. (1996)
ion substitution in the catalytic site greatly Xer-mediated site-specific recombination
affects the binding of sulfhydryl- in vitro. EMBO J., 15, 1172–1181.
containing compounds to leucyl 251 Alen, C., Sherratt, D.J., and Colloms, S.D.
aminopeptidase. Biochemistry, 45, (1997) Direct interaction of
3226–3234. aminopeptidase A with recombination
243 Cappiello, M., Lazzarotti, A., Buono, F., site DNA in Xer site-specific
Scaloni, A., DAmbrosio, C., Amodeo, P., recombination. EMBO J., 16,
Mendez, B.L., Pelosi, P., Del Corso, A., 5188–5197.
and Mura, U. (2004) New role for leucyl 252 McCulloch, R., Burke, M.E., and
aminopeptidase in glutathione turnover. Sherratt, D.J. (1994) Peptidase activity of
Biochem. J., 378, 35–44. Escherichia coli aminopeptidase A is not
244 Str€ater, N. and Lipscomb, W.N. (1995) required for its role in Xer site-specific
Transition state analogue L- recombination. Mol. Microbiol., 12,
leucinephosphonic acid bound to bovine 241–251.
lens leucine aminopeptidase: X-ray 253 Reijns, M., Lu, Y., Leach, S., and Colloms,

structure at 1.65 A resolution in a S.D. (2005) Mutagenesis of PepA
new crystal form. Biochemistry, 34, suggests a new model for the Xer/cer
9200–9210. synaptic compl. Mol. Microbiol., 57,
245 Str€ater, N., Lipscomb, W.N., Klabunde, T., 927–941.
and Krebs, B. (1996) Two-metal ion 254 Minh, P.N.L., Devroede, N., Massant, J.,
catalysis in enzymatic acyl- and Maes, D., and Charlier, D. (2009) Insights
phosphoryl-transfer reactions. Angew. into the architecture and stoichiometry of
Chem. Int. Ed. Engl., 35, 2024–2055. Escherichia coli PepADNA complexes
246 Str€ater, N., Sherratt, D.J., and Colloms, involved in transcriptional control and
S.D. (1999) X-ray structure of site-specific DNA recombination by
aminopeptidase A from Escherichia coli atomic force microscopy. Nucleic Acids
and a model for the nucleoprotein Res., 37, 1463–1476.
complex in Xer site-specific 255 Charlier, D., Hassanzadeh Gh., G.,
recombination. EMBO J., 18, 4513–4522. Kholti, A., Gigot, D., Pierard, A., and
247 Str€ater, N., Sun, L., Kantrowitz, E.R., and Glansdorff, N. (1995) carP, involved in
Lipscomb, W.N. (1999) A bicarbonate ion pyrimidine regulation of the
as a general base in the mechanism of Escherichia coli carbamoylphosphate
peptide hydrolysis by dizinc leucine synthetase operon encodes a sequence-
aminopeptidase. Proc. Natl. Acad. Sci. specific DNA-binding protein identical to
U.S.A., 96, 11151–11155. XerB and PepA, also required for
248 Stirling, C.J., Colloms, S.D., Collins, J.F., resolution of ColEl multimers. J. Mol.
Szatmari, G., and Sherratt, D.J. (1989) Biol., 250, 392–406.
xerB, an Escherichia coli gene required for 256 Behari, J., Stagon, L., and
plasmid ColE1 site-specific Calderwood, S.B. (2001) pepA, a gene
recombination, is identical to pepA, mediating pH regulation of virulence
encoding aminopeptidase A, a protein genes in Vibrio cholerae. J. Bacteriol., 183,
with substantial similarity to bovine lens 178–188.
644
257 Woolwine, S.C. and Wozniak, D.J. Ochrobactrum anthropi, a new member of
(1999) Identification of an Escherichia coli the penicillin-recognizing enzyme
pepA homolog and its involvement in family. Structure, 8, 971–980.
suppression of the algB phenotype in 267 Delmarcelle, M., Boursoit, M.-C., Filee,
mucoid Pseudomonas aeruginosa. J. P., Baurin, S.L., Frere, J.-M., and Joris, B.
Bacteriol., 181, 107–116. (2005) Specificity inversion of
258 Woolwine, S.C., Sprinkle, A.B., and Ochrobactrum anthropiD-aminopeptidase
Wozniak, D.J. (2001) Loss of Pseudomonas to a Dd-carboxypeptidase with new
aeruginosa PhpA aminopeptidase penicillin binding activity by directed
activity results in increased algD mutagenesis. Protein Sci.,
transcription. J. Bacteriol., 183, 14, 2296–2303.
4674–4679. 268 Asano, Y. and Yamaguchi, K. (1995)
259 Matsui, M., Fowler, J.H., and Walling, L.L. Mutants of D-aminopeptidase with
(2006) Leucine aminopeptidases: increased thermal stability. J. Ferment.
diversity in structure and function. Bioeng., 79, 614–616.
Biol. Chem., 387, 1535–1544. 269 Asano, Y., Kishino, K., Yamada, A.,
260 Asano, Y. and L€ ubbeh€usen, T.L. (2000) Hanamoto, S., and Kondo, K. (1991)
Enzymes acting on peptides Plasmid-based, D-aminopeptidase-
containing D-amino acid. J. Biosci. Bioeng., catalysed synthesis of (R)-amino
89, 295–306. acids. Recl. Trav. Chim. Pays-Bas, 110,
261 Asano, Y. (2007) Enzymes acting on 206–208.
D-amino acid amides, in d-Amino Acids: 270 Asano, Y., Nakazawa, A., Kato, Y., and
A New Frontier in Amino Acid and Kondo, K. (1989) Isolierung einer
Protein Research (eds R. Konno, H. D-stereospezifischen Aminopeptidase
Br€uckner, A. DAniello, G. Fischer, N. und ihre anwendung als katalysator in der
Fujii, and H. Homma), organische synthese. Angew. Chem., 101,
Nova Science Publishers, Inc., New York, 511–512.
pp. 579–589. 271 Kato, Y., Asano, Y., Nakazawa, A., and
262 Asano, Y. (2000) New enzymes acting on Kondo, K. (1989) First stereoselective
peptides containing D-amino acids: their synthesis of D-amino acid N-alkyl amide
properties and application. J. Microbiol. catalyzed by D-aminopeptidase.
Biotechnol., 10, 573–579. Tetrahedron, 45, 5743–5754.
263 Asano, Y., Nakazawa, A., Kato, Y., and 272 Kato, Y., Asano, Y., Nakazawa, A., and
Kondo, K. (1989) Properties of a novel D- Kondo, K. (1990) Synthesis of D-alanine
stereospecific aminopeptidase from oligopeptides catalyzed by
Ochrobactrum anthropi. J. Biol. Chem., D-aminopeptidase in non-aqueous media.
264, 14233–14239. Biocatalysis, 3, 207–215.
264 Asano, Y., Mori, T., Hanamoto, S., 273 Komeda, H. and Asano, Y. (2000) Gene
Kato, Y., and Nakazawa, A. (1989) cloning, nucleotide sequencing, and
A new D-stereospecific amino acid purification and characterization of the
amidase from Ochrobactrum anthropi. D-stereospecific amino-acid amidase
Biochem. Biophys. Res. Commun., 162, from Ochrobactrum anthropi SV3. Eur. J.
470–474. Biochem., 267, 2028–2035.
265 Asano, Y., Kato, Y., Yamada, A., and 274 Komeda, H., Ishikawa, N., and Asano, Y.
Kondo, K. (1992) Structural similarity of (2003) Enhancement of the
D-aminopeptidase to carboxypeptidase Dd thermostability and catalytic activity of
and beta-lactamases. Biochemistry, 31, D-stereospecific amino-acid amidase
2316–2328. from Ochrobactrum anthropi SV3 by
266 Bompard-Gilles, C., Remaut, H., directed evolution. J. Mol. Catal. B:
Villeret, V., Prange, T., Fanuel, L., Enzym., 21, 283–290.
Delmarcelle, M., Joris, B., Frere, J., and 275 Okazaki, S., Suzuki, A., Komeda, H.,
Van Beeumen, J. (2000) Crystal structure Yamaguchi, S., Asano, Y., and Yamane, T.
of a D-aminopeptidase from (2007) Crystal structure and functional
References j645
characterization of a D-stereospecific preliminary X-ray analysis of a new L-
amino acid amidase from Ochrobactrum aminopeptidase-D-amidase/D-esterase
anthropi SV3, a new member of the activated by a Gly-Ser peptide bond
penicillin-recognizing proteins. J. Mol. hydrolysis. Acta Crystallogr., Sect. D, 55
Biol., 368, 79–91. (Pt 3), 699–701.
276 Okazaki, S., Suzuki, A., Komeda, H., 283 Heck, T., Limbach, M., Geueke, B.,
Asano, Y., and Yamane, T. (2008) Zacharias, M., Gardiner, J.,
Deduced catalytic mechanism of Kohler, H.-P.E., and Seebach, D. (2006)
D-amino acid amidase from Enzymatic degradation of b- and mixed
Ochrobactrum anthropi SV3. J. a,b-oligopeptides. Chem. Biodivers., 3,
Synchrotron Radiat., 15, 250–253. 1325–1348.
277 Okazaki, S., Suzuki, A., Mizushima, T., 284 Ozaki, A., Kamasaki, H., Yagasaki, M.,
Komeda, H., Asano, Y., and Yamane, T. and Hashimoto, Y. (1992) Enzymatic
(2008) Structures of D-amino-acid production of D-alanine from Dl-
amidase complexed with L-phenylalanine alaninamide by novel D-alaninamide
and with L-phenylalanine amide: insight specific amide hydrolase.
into the D-stereospecificity of D-amino- Biosci. Biotechnol. Biochem., 56,
acid amidase from Ochrobactrum anthropi 1980–1984.
SV3. Acta Crystallogr., Sect. D, 64, 285 Krieg, L., Ansorge-Schumacher, M.B.,
331–334. and Kula, M.-R. (2002) Screening for
278 Fanuel, L., Thamm, I., Kostanjevecki, V., amidases: isolation and characterization
Samyn, B., Joris, B., Goffin, C., of a novel D-amidase from Variovorax
Brannigan, J., Van Beeumen, J., and paradoxus. Adv. Synth. Catal., 344,
Frere, J.M. (1999) Two new 965–973.
aminopeptidases from Ochrobactrum 286 Verseck, S., Drauz, K., Bommarius, A.,
anthropi active on D-alanyl-p-nitroanilide. Kula, M.-R., Krieg, L., Slusarczyk, H., and
Cell Mol. Life Sci., 55, 812–818. Ansorge, M. (2003) D-Amidase aus
279 Frere, J.-M. and Van Beeumen, J. (2004) Variovorax. EP 1,318,193 to Degussa AG.
DmpA L-aminopeptidase D-Ala esterase/ 287 Krieg, L., Slusarczyk, H., Verseck, S., and
amidase of Ochrobactrum anthropi, in Kula, M.R. (2005) Identification and
Handbook of Proteolytic Enzymes, 2nd edn characterization of a novel D-amidase
(eds A.J. Barrett, N.D. Rawlings, and J.F. gene from Variovorax paradoxus and
Woessner), Elsevier Academic Press, its expression in Escherichia coli.
London, pp. 2055–2057. Appl. Microbiol. Biotechnol.,
280 Fanuel, F.L., Goffin, C., Cheggour, A., 66, 542–550.
Devreese, B., Van Driessche, G., Joris, B., 288 Brand~ao, P.F.B., Verseck, S., and
Van Beeumen, J., and Frere, J.M. (1999) Syldatk, C. (2004) Bioconversion of
The DmpA aminopeptidase from Dl-tert-leucine nitrile to D-tert-leucine by
Ochrobactrum anthropi LMG7991 is the recombinant cells expressing nitrile
prototype of a new terminal nucleophile hydratase and D-selective amidase. Eng.
hydrolase family. Biochem. J., 341Pt (1), Life Sci., 4, 547–556.
147–155. 289 Hongpattarakere, T., Komeda, H., and
281 Bompard-Gilles, C., Villeret, V., Asano, Y. (2005) Purification,
Davies, G.J., Fanuel, L., Joris, B., characterization, gene cloning and
Frere, J.M., and Van Beeumen, J. (2000) A nucleotide sequencing of D-stereospecific
new variant of the Ntn hydrolase fold amino acid amidase from soil bacterium:
revealed by the crystal structure of L- Delftia acidovorans. J. Ind. Microbiol.
aminopeptidase D-ala-esterase/amidase Biotechnol., 32, 567–576.
from Ochrobactrum anthropi. Struct. Fold. 290 Komeda, H. and Asano, Y. (2008) A novel
Des., 8, 153–162. D-stereoselective amino acid amidase
282 Bompard-Gilles, C., Villeret, V., Fanuel, from Brevibacterium iodinum: gene
L., Joris, B., Frere, J.M., and Van cloning, expression and characterization.
Beeumen, J. (1999) Crystallization and Enzyme Microb. Technol., 43, 276–283.
646
291 Baek, D.H., Song, J.J., Lee, S.-G., 299 Kaptein, B., Boesten, W.H.J.,
Kwon, S.J., Asano, Y., and Sung, M.-H. Broxterman, Q.B., Peters, P.J.H.,
(2003) New thermostable D-methionine Schoemaker, H.E., and Kamphuis, J.
amidase from Brevibacillus borstelensis (1993) Enzymatic resolution of
BCS-1 and its application for D- a,a-disubstituted a-amino acid esters
phenylalanine production. Enzyme and amides. Tetrahedron: Asymmetry, 4,
Microb. Technol., 32, 131–139. 1113–1116.
292 Baek, D.H., Kwon, S.J., Hong, S.P., 300 Kruizinga, W.H., Bolster, J.,
Kwak, M.S., Lee, M.H., Song, J.J., Kellogg, R.M., Kamphuis, J.,
Lee, S.G., Yoon, K.H., and Sung, M.H. Boesten, W.H.J., Meijer, E.M., and
(2003) Characterization of a thermostable Schoemaker, H.E. (1988) Synthesis of
D-stereospecific alanine amidase from optically pure a-alkylated a-amino acids
Brevibacillus borstelensis BCS-1. Appl. and a single-step method for
Environ. Microbiol., 69, 980–986. enantiomeric excess determination.
293 Cheggour, A., Fanuel, L., Owingz, C., J. Org. Chem., 53, 1826–1827.
Joris, B., Bouillenne, F., Devreese, B., 301 Hermes, H.F.M., Tandler, R.F.,
Van Driessche, G., Van Beeumen, J., Sonke, T., Dijkhuizen, L., and
Frere, J.-M., and Goffin, C. (2000) Meijer, E.M. (1994) Purification and
The dppA gene of Bacillus subtilis encodes characterization of an L-amino
a new D-aminopeptidase. Mol. Microbiol., amidase from Mycobacterium
38, 504–513. neoaurum ATCC 25795. Appl. Environ.
294 Remaut, H., Bompard-Gilles, C., Microbiol., 60, 153–159.
Goffin, C., Frere, J.-M., and Van 302 Van den Tweel, W.J.J.,
Beeumen, J. (2001) Structure of the van Dooren, T.J.G.M., de Jonge, P.H.,
Bacillus subtilisD-aminopeptidase Kaptein, B., Duchateau, A.L.L., and
DppA reveals a novel self- Kamphuis, J. (1993) Ochrobactrum
compartmentalizing protease. Nat. anthropi NCIMB 40321: a new biocatalyst
Struct. Biol., 8, 674–678. with broad-spectrum L-specific amidase
295 Kaptein, B., Boesten, W.H.J., activity. Appl. Microbiol. Biotechnol., 39,
Broxterman, Q.B., Schoemaker, H.E., 296–300.
and Kamphuis, J. (1992) Synthesis of a, 303 de Vries, E.J. (1997) Enzymatic
a-disubstituted a-amino acid amides by preparation of optically active carboxylic
phase-transfer catalyzed alkylation. acid derivatives using Ochrobactrum
Tetrahedron Lett., 33, 6007–6010. anthropi. Ph.D. thesis, University of
296 Roos, E.C., Lopez, M.C., Brook, M.A., Groningen, The Netherlands.
Hiemstra, H., Speckamp, W.N., 304 Kaptein, B., Dooren, T.J.G.M.V.,
Kaptein, B., Kamphuis, J., and Boesten, W.H.J., Sonke, T.,
Schoemaker, H.E. (1993) Synthesis of Duchateau, A.L.L., Broxterman, Q.B., and
a-substituted a-amino acids via cationic Kamphuis, J. (1998) Synthesis of 4-sulfur-
intermediates. J. Org. Chem., 58, substituted (2S,3R)-3-phenylserines by
3259–3268. enzymatic resolution. Enantiopure
297 Roos, E.C., Hiemstra, H., precursors for thiamphenicol and
Speckamp, W.N., Kaptein, B., florfenicol. Org. Process Res. Dev., 2,
Kamphuis, J., and Schoemaker, H.E. 10–17.
(1992) Synthesis of c,d-unsaturated 305 Gouret, C.J., Wettstein, J.G.,
a-methyl-a-amino acids via coupling of Porsolt, R.D., Puech, A., and Junien, J.L.
allylsilanes with an a-methylglycine (1990) Neuropsychopharmacologial
cation equivalent. Synlett, 451–452. profile of JO 1017, a new antidepressant
298 Becke, F., Fleig, H., and P€assler, P. (1971) and selective serotonin uptake inhibitor.
General method for the preparation of Eur. J. Pharmacol., 183, 1478.
amides from their corresponding nitriles. 306 Gouret, C.J., Porsolt, R.D.,
II. Justus Liebigs Ann. Chem., 749, Wettstein, J.G., Puech, A., Soulard, C.,
198–201. Pascaud, X., and Junien, J.L. (1990)
References j647
Biochemical and pharmacological 315 Inoue, A., Komeda, H., and Asano, Y.
evaluation of the novel antidepressant (2005) Asymmetric synthesis of L-
and serotonin uptake inhibitor a-methylcysteine with the amidase from
(2-(3,4-dichlorobenzyl)-2- Xanthobacter flavus NR303. Adv. Synth.
dimethylamino-1-propanol Catal., 347, 1132–1138.
hydrochloride. Arzneim-Forsch/Drug Res., 316 Steer, D.L., Lew, R.A., Perlmutter, P.,
40, 633–640. Smith, A.I., and Aguilar, M.-I. (2002)
307 Kaptein, B., Moody, H.M., b-Amino acids: versatile
Broxterman, Q.B., and Kamphuis, J. peptidomimetics. Curr. Med. Chem., 9,
(1994) Chemo-enzymatic synthesis of (S)- 811–822.
( þ )-cericlamine and related 317 Cheng, R.P., Gellman, S.H., and
enantiomerically pure 2,2-disubstituted- DeGrado, W.F. (2001) b-Peptides: from
2-aminoethanols. J. Chem. Soc., Perkin structure to function. Chem. Rev., 101,
Trans. 1, 1495–1498. 3219–3232.
308 van Lingen, H.L., van de Mortel, J.K.W., 318 Steer, D.L., Lew, R.A., Perlmutter, P.,
Hekking, K.F.W., van Delft, F.L., Smith, A.I., and Aguilar, M.-I. (2002)
Sonke, T., and Rutjes, F.P.J.T. (2003) An The use of b-amino acids in the design of
efficient synthesis of 1-naphthylbis protease and peptidase inhibitors. Lett.
(oxazoline) and exploration of the scope in Pept. Sci., 8, 241–246.
asymmetric catalysis. Eur. J. Org. Chem., 319 Seebach, D. and Gardiner, J. (2008)
317–324. b-Peptidic peptidomimetics. Acc. Chem.
309 Sonke, T., Ernste, S., Tandler, R.F., Res., 41, 1366–1375.
Kaptein, B., Peeters, W.P.H., 320 Lelais, G. and Seebach, D. (2004) b2-
Assema, F.B.J.V., Wubbolts, M.G., and Amino acids - syntheses, occurrence in
Schoemaker, H.E. (2005) L-Selective natural products, and components of
amidase with extremely broad substrate b-peptides. Biopolym. - Pept. Sci. Sect., 76,
specificity from Ochrobactrum anthropi 206–243.
NCIMB 40321. Appl. Environ. Microbiol., 321 Seebach, D., Beck, A.K., Capone, S.,
71, 7961–7973. Deniau, G., Groselj, U., and Zass, E.
310 Sonke, T., Tandler, R.F., Korevaar, C.G.N., (2009) Enantioselective preparation of b2-
van Assema, F.B.J., and van der Pol, R. amino acid derivatives for b-peptide
(2003) Nucleic acid sequences encoding synthesis. Synthesis, 1–32.
enantioselective amidases. WO 03/ 322 Tasnadi, G., Forro, E., and F€
ul€op, F. (2008)
010312 to DSM N.V. An efficient new enzymatic method for
311 Nakamura, T. and Yu, F. (2003) Amidase the preparation of b-aryl-b-amino acid
gene. US 6,617,139 to Mitsubishi Rayon enantiomers. Tetrahedron: Asymmetry, 19,
Co., Ltd. 2072–2077.
312 Katoh, O., Akiyama, T., and Nakamura, T. 323 Tasnadi, G., Forro, E., and F€
ul€op, F. (2009)
(2004) Novel amide hydrolase gene. EP Burkholderia cepacia lipase is an excellent
1,428,876 to Mitsubishi Rayon Co., Ltd. enzyme for the enantioselective
313 Brieden, W., Naughton, A., Robins, K., hydrolysis of b-heteroaryl-b-amino
Shaw, N.M., Tinschert, A., and esters. Tetrahedron: Asymmetry, 20,
Zimmermann, T. (2004) Method of 1771–1777.
preparing (S)- or (R)-3,3,3-trifluoro-2- 324 Gr€oger, H., Trauthwein, H., Buchholz, S.,
hydroxy-2-methylpropionic acid. US Drauz, K., Sacherer, C., Godfrin, S., and
6,773,910 to Lonza AG. Werner, H. (2004) The first aminoacylase-
314 Fraser, J.A., Davis, M.A., and Hynes, M.J. catalyzed enantioselective synthesis of
(2001) The formamidase gene of aromatic b-amino acids. Org. Biomol.
Aspergillus nidulans: regulation by Chem., 2, 1977–1978.
nitrogen metabolite repression and 325 Li, D., Cheng, S., Wei, D., Ren, Y., and
transcriptional interference by an Zhang, D. (2007) Production of
overlapping upstream gene. Genetics, 157, enantiomerically pure (S)-
119–131. b-phenylalanine and (R)-b-phenylalanine
648
by penicillin G acylase from Escherichia 334 Heck, T., Reimer, A., Seebach, D.,
coli in aqueous medium. Biotechnol. Lett., Gardiner, J., Deniau, G., Lukaszuk, A.,
29, 1825–1830. Kohler, H.-P.E., and Geueke, B. (2010)
326 Szymanski, W., Wu, B., Weiner, B., b-aminopeptidase-catalyzed
De Wildeman, S., Feringa, B.L., and biotransformations of b2-dipeptides:
Janssen, D.B. (2009) Phenylalanine kinetic resolution and enzymatic
aminomutase-catalyzed addition of coupling. ChemBioChem, 11, 1129–1136.
ammonia to substituted cinnamic 335 Shaw, N.M. and Naughton, A.B. (2004)
acids: a route to enantiopure a- and The substrate specificity of the
b-amino acids. J. Org. Chem., 74, heat-stable stereospecific amidase from
9152–9157. Klebsiella oxytoca. Tetrahedron, 60,
327 Liljeblad, A. and Kanerva, L.T. (2006) 747–752.
Biocatalysis as a profound tool in the 336 van Dooren, T.J.G.M. and van den Tweel,
preparation of highly enantiopure W.J.J. (1992) Enzyme-catalysed
b-amino acids. Tetrahedron, 62, preparation of optically active
5831–5854. carboxylic acids. EP0494716 to DSM N.V.
328 Heck, T., Seebach, D., Osswald, S., 337 McLeish, M.J., Kneen, M.M.,
ter Wiel, M.K.J., Kohler, H.-P.E., and Gopalakrishna, K.N., Koo, C.W.,
Geueke, B. (2009) Kinetic resolution of Babbitt, P.C., Gerlt, J.A., and Kenyon,
aliphatic b-amino acid amides by G.L. (2003) Identification and
b-aminopeptidases. ChemBioChem, 10, characterization of a mandelamide
1558–1561. hydrolase and an NAD(P)
329 Geueke, B., Namoto, K., Seebach, D., and þ -dependent benzaldehyde
Kohler, H.P. (2005) A novel dehydrogenase from Pseudomonas
b-peptidyl aminopeptidase (BapA) from putida ATCC 12633. J. Bacteriol., 185,
strain 3-2W4 cleaves peptide bonds of 2451–2456.
synthetic b-tri- and b-dipeptides. 338 Gopalakrishna, K.N., Stewart, B.H.,
J. Bacteriol., 187, 5910–5917. Kneen, M.M., Andricopulo, A.D.,
330 Geueke, B., Heck, T., Limbach, M., Kenyon, G.L., and McLeish, M.J. (2004)
Nesatyy, V., Seebach, D., and Mandelamide hydrolase from
Kohler, H.P.E. (2006) Bacterial Pseudomonas putida: characterization of a
b-peptidyl aminopeptidases with new member of the amidase
unique substrate specificities for signature family. Biochemistry, 43,
b-oligopeptides and mixed b, 7725–7735.
a-oligopeptides. Eur. J. Biochem., 273, 339 Wang, P.-F., Yep, A., Kenyon, G.L., and
5261–5272. McLeish, M.J. (2009) Using directed
331 Geueke, B. and Kohler, H.P. (2007) evolution to probe the substrate
Bacterial b-peptidyl aminopeptidases: on specificity of mandelamide hydrolase.
the hydrolytic degradation of b-peptides. Protein Eng., 22, 103–110.
Appl. Microbiol. Biotechnol., 74, 1197–1204. 340 Kent, S.B.H. (1988) Chemical synthesis
332 Heck, T., Kohler, H.-P.E., Limbach, M., of peptides and proteins. Annu. Rev.
Fl€ogel, O., Seebach, D., and Geueke, B. Biochem., 57, 957–989.
(2007) Enzyme-catalyzed formation of 341 Meldal, M., Juliano, M.A., and
b-peptides: b-peptidyl aminopeptidases Jansson, A.M. (1997) Azido acids in a
BapA and DmpA acting as b-peptide- novel method of solid-phase peptide
synthesizing enzymes. Chem. Biodivers., synthesis. Tetrahedron Lett., 38,
4, 2016–2030. 2531–2534.
333 Heck, T., Makam, V.S., Lutz, J., 342 Tornøe, C.W., Sengeløv, H., and Meldal,
Blank, L.M., Schmid, A., Seebach, D., M. (2000) Solid-phase synthesis of
Kohler, H.P.E., and Geuekea, B. (2010) chemotactic peptides using a-azido acids.
Kinetic analysis of L-carnosine formation J. Pept. Sci., 6, 314–320.
by b-aminopeptidases. Adv. Synth. Catal., 343 Tornøe, C.W., Davis, P., Porreca, F., and
352, 407–415. Meldal, M. (2000) a-Azido acids for direct
References j649
use in solid-phase peptide synthesis. J. 353 Kim, K.-H. and Seong, B.L. (2001)
Pept. Sci., 6, 594–602. Peptide amidation: production of
344 Tornøe, C.W., Sonke, T., Maes, I., peptide hormones in vivo and
Schoemaker, H.E., and Meldal, M. (2000) in vitro. Biotechnol. Bioprocess Eng., 6,
Enzymatic and chiral HPLC 244–251.
resolution of a-azido acids and 354 Cerovsky, V. and Kula, M.-R. (2001)
amides. Tetrahedron: Asymmetry, 11, Studies on peptide amidase-catalysed C-
1239–1248. terminal peptide amidation in organic
345 Jost, M., Sonke, T., Kaptein, B., media with respect to its substrate
Broxterman, Q.B., and Sewald, N. (2005) specificity. Biotechnol. Appl. Biochem., 33,
Synthesis and enzymatic resolution of 183–187.
Ca-dialkylated a-azido carboxamides: 355 Kaftzik, N., Neumann, S., Kula, M.-R.,
new enantiopure a-azido acids as and Kragl, U. (2003) Enzymatic
building blocks in peptide synthesis. condensation reactions in
Synthesis, 272–278. ionic liquids. ACS Symp. Ser., 856,
346 Quaedflieg, P.J.L.M., Sonke, T., 206–211.
Verzijl, G.K.M., and Wiertz, R.W. (2007) 356 Stelkes-Ritter, U. (1994) Reinigung,
Enzymatic conversion of oligopeptide charakterisierung und anwendung der
amides to oligopeptide peptidamidasen aus Citrus sinensis L.
alkylesters. WO 2007/045470 to DSM IP und Stenotrophomonas maltophilia.
Assets B.V. Ph.D. thesis, Heinrich-Heine-Universit€at
347 Kammermeier-Steinke, D., Schwarz, A., D€usseldorf.
Wandrey, C., and Kula, M.-R. (1993) 357 Stelkes-Ritter, U., Wyzgol, K., and
Studies on the substrate specificity of a Kula, M.-R. (1995) Purification and
peptide amidase partially purified from characterization of a newly screened
orange flavedo. Enzyme Microb. Technol., microbial peptide amidase.
15, 764–769. Appl. Microbiol. Biotechnol.,
348 Steinke, D., Kula, M.-R., Schwarz, A., and 44, 393–398.
Wandrey, C. (1991) Peptidamidase 358 Stelkes-Ritter, U., Kula, M.-R.,
und deren Verwendung. DE 4014564 to Wyzgol, K., Bommarius, A.,
Forschungszentrum Juelich GmbH and Schwarm, M., and Drauz, K. (1995)
Degussa Aktiengesellschaft. Verfahren zur Gewinnung von
349 Stelkes-Ritter, U., Beckers, G., peptidamidase enthaltenden
Bommarius, A., Drauz, K., mikroorganismen, damit gewonnene
G€unther, K., Kottenhahn, M., mikroorganismen, darin enthaltene
Schwarm, M., and Kula, M.-R. (1997) peptidamidasen und deren verwendung.
Kinetics of peptide amidase and its DE 19516018 to Degussa
application for the resolution of Aktiengesellschaft.
racemates. Biocatal. Biotransform., 15, 359 Neumann, S. and Kula, M.-R. (2002)
205–219. Gene cloning, overexpression and
350 Steinke, D. and Kula, M.-R. (1990) biochemical characterization of the
Selective deamidation of peptide amides. peptide amidase from Stenotrophomonas
Angew. Chem. Int. Ed. Engl., 29, maltophilia. Appl. Microbiol. Biotechnol.,
1139–1140. 58, 772–780.
351 Schwarz, A., Wandrey, C., Steinke, D., 360 Neumann, S., Granzin, J., Kula, M.-R.,
and Kula, M.-R. (1992) A two-step and Labahn, J. (2002) Crystallization and
enzymatic synthesis of dipeptides. preliminary X-ray data of the
Biotechnol. Bioeng., 39, 132–140. recombinant peptide amidase
352 rovsky, V. and Kula, M.-R.
Ce from Stenotrophomonas maltophilia.
(1998) C-terminal peptide amidation Acta Crystallogr., Sect. D,
catalyzed by orange flavedo peptide 58, 333–335.
amidase. Angew. Chem. Int. Ed., 37, 361 Labahn, J., Neumann, S., B€ uldt, G., Kula,
1885–1887. M.-R., and Granzin, J. (2002) An
650
alternative mechanism for amidase Ser-Ser-Lys catalytic triad mechanism of

signature enzymes. J. Mol. Biol., 322, peptide amidase: computational studies of
1053–1064. the ground state, transition state, and
362 Vali~
na, A.L.B., Mazumder-Shivakumar, D., intermediate. Biochemistry, 43,
and Bruice, T.C. (2004) Probing the 15657–15672.
j651
16
Hydrolysis and Formation of Hydantoins
Jun Ogawa, Nobuyuki Horinouchi, and Sakayu Shimizu
16.1
Overview of Microbial Hydantoin Metabolism and its Application to Biotechnology
5-Monosubstituted hydantoins are chemical intermediates in the organic synthesis

of amino acids by the Bucherer method. The systematic term for hydantoins is
imidazole-2,4-diones or 2,4-diketotetrahydroimidazoles, of which 5-monosub-
stituted hydantoins may be regarded as cyclic ureides of a-amino acid.
DL-5-Monosubstituted hydantoins are used as substrates for enzymatic asymmet-
ric synthesis of optically active amino acids. Based on the investigations of Dudley
et al. [1, 2], who studied the metabolism of N-substituted DL-hydantoins that were
postulated to be D-stereoselectively hydrolyzed, Cecere et al. [3] found in 1975 that
dihydropyrimidinase from calf liver could be used to produce several N-carbamoyl-D-
amino acids from the corresponding DL-5-monosubstituted hydantoins. In 1978,
Yamada et al. showed that microbial cells are good catalysts for amide hydrolysis of
hydantoin derivatives [4]. Thereafter, several bacteria that could hydrolyze N-carba-
moyl-D-amino acids to corresponding D-amino acids were found [5].
In mammal, plants, and microorganisms, hydantoin derivatives are metabolized
to an amino acid through two-step hydrolysis via an N-carbamoyl amino acid
(Figure 16.1b). The enzyme catalyzing the first step, hydrolysis of hydantoin to N-
carbamoyl amino acid, is called hydantoinase. As indicated in Figure 16.2, three
classes of hydantoinases with different stereospecificities towards D-, L-, or DL-5-
monosubstituted hydantoins named D-hydantoinase, L-hydantoinase, and DL-hydan-
toinase, respectively, have been reported [6–9]. Takahashi et al. [10] revealed that, in
Pseudomonas putida (same as P. striata) IFO 12996, D-hydantoinase is identical to
dihydropyrimidinase, which catalyzes the cyclic ureide-hydrolyzing step of the
reductive degradation of pyrimidine bases (Figure 16.1a). The same results were
obtained for other Pseudomonas species [11, 12], Comamonas species [12], Bacillus
species [13], Arthrobacter species [14], Agrobacterium species [15], and rat liver [16].
From these results, it is proposed that D-amino acid production from DL-5-mono-
substituted hydantoins involves the action of the series of enzymes involved in the
pyrimidine degradation pathway [14, 15, 17]. However, this contention has remained
and Oliver May.
j 16 Hydrolysis and Formation of Hydantoins
652
(a) (b) (c)
Pyrimidine Hydantoin Cyclic imide

metabolism metabolism metabolism
O
Oxidative Reductive
NH
1 HN 4
O
O O O O
O NH NH R NH NH
HN HN
*
HN
O O O O
2 5 7 9
COOH COOH COOH COOH
O NH 2 NH2 R NH 2
HN HN
O
*
HN
NH2
O O O
3 6 8 10
COOH
COOH COOH COOH
HOOC
+ R
H2N
NH 2
NH 2 *NH COOH
2
O
11
HOOC
COOH
12
OH
HOOC
COOH
13
O
HOOC
CH 3
Figure 16.1 Pyrimidine and related 6: b-ureidopropionase, 7: hydantoinase, 8:

cyclic-amide metabolism. 1: uracil/thymine N-carbamoylase, 9: imidase, 10: half-amidase,
dehydrogenase, 2: barbiturase, 3: and 11–13: TCA cycle like reactions. (a)
ureidomalonase, 4: dihydrouracil pyrimidine metabolism, (b) hydantoin
dehydrogenase, 5: dihydropyrimidinase, metabolism, and (c) cyclic imide metabolism.
moot because of the lack of systematic studies on the enzymes involved in these
transformations [18].
The L- and DL-hydantoinases might be rarer in nature than D-hydantoinase
(dihydropyrimidinase). These enzymes can be divided into two groups: one needs
16.1 Overview of Microbial Hydantoin Metabolism and its Application to Biotechnology j653
Dihydroorotase
D-Hydantoinase Pseudomonas putida IFO12996
Pseudomonas putida IFO 12996 O
Blastobacter sp. A17p-4 HN
H 2O HOOC
H2N
O O N COOH O COOH
N
R H 2O R COOH H L-specific H
NH NH2
H H
HN HN
O D-specific O
ATP
O +
2 H 2O
O COOH
H 2O COOH NH NH2
NH NH2 HN ADP HN
HN HN O + O
O O H3C Pi H3 C
ATP
O +
2 H 2O
H H COOH
O NH NH2
H 2O COOH R R
HN ADP HN
NH NH2 O + O
Pi
O O L-specific
Imidase
Blastobacer sp. A17p-4 N-Methylhydantoin amidohydrolase
Pseudomonas putida 77
Figure 16.2 Substrate specificities of typical cyclic amide-hydrolyzing enzymes.
ATP for the activity and the other not. The ATP-requiring enzyme from Pseudomonas
putida 77, which functions in creatinine metabolism, showed L-hydantoinase
activity [19].
The hydantoinases can often be found in microorganisms together with highly
stereoselective N-carbamoyl-a-amino acid amidohydrolases (N-carbamoylase;
E.C.3.5.1.77 or 87), which catalyze the further hydrolysis of N-carbamoyl-a-amino
acid, a hydantoic acid derivative, to the free a-amino acid in an irreversible manner.
Two typical N-carbamoylases with stereospecificity to N-carbamoyl-D- and N-carba-
moyl-L-amino acids are named D-N-carbamoylase and L-N-carbamoylase, respectively
[6] (Figure 16.3). D-N-Carbamoylase generally shows a wide substrate specificity to
both aromatic and aliphatic N-carbamoyl-D-amino acids [20]. L-N-Carbamoylase
shows rather limited specificity to aromatic or aliphatic N-carbamoyl-L-amino acids,
while L-N-carbamoylase with a relatively broad substrate specificity has been found in
Alcaligenes xylosoxidans [21].
D-N-Carbamoylase was previously proposed to be identical with b-ureidopropio-
nase, which catalyzes N-carbamoyl-b-alanine hydrolysis in reductive pyrimidine
degradation. This is, however, no longer valid since the investigation of Ogawa
et al. on the distribution of hydantoin- and dihydropyrimidine-transforming
enzymes in various aerobic bacteria [22]. They reported the occurrence of a D-N-
carbamoylase activity independent from D-hydantoinase activity in a microorgan-
ism [20]. In this context, it may be of interest that Runser and Meyer described a
D-hydantoinase with no dihydropyrimidinase activity [23]. Ogawa et al. also con-
ducted detailed analysis of b-ureidopropionase (E.C. 3.5.1.6) from Pseudomonas
654 j 16 Hydrolysis and Formation of Hydantoins
Hydantoin transformation pathway Pyrimidine transformation pathway
O NH3 O NH3
+ H2O +
H2O COOH H2O CO COOH COOH H2O CO2 COOH
2
R NH R NH2 R NH NH2
HN HN NH2 HN HN NH2
O O O O
D-N-Carbamoylase
β-Ureidopropionase
Comamonas sp. E222c
Blastobacter sp. A17p-4 Pseudomonas putida IFO 12996
R COOH R COOH COOH COOH

NH2 + H2O + NH3 + CO2 NH2 + H2O + NH3 + CO2
H HN H NH2 HN NH2
O O
COOH COOH
L-N-Carbamoylase NH2 + H2O + NH3 + CO2

Alcaligenes xylosoxidans AKU110 HN NH2
O
H COOH H COOH H COOH H COOH
NH2 + H2O + NH3 + CO2 NH2 + H 2O + NH3 + CO2
R HN R NH2
O R HN R NH2
O
Figure 16.3 Substrate specificities of typical N-carbamoyl amino acid-hydrolyzing enzymes.
putida IFO 12996 [24]. b-Ureidopropionase (E.C. 3.5.1.6) from P. putida IFO 12996
showed broad substrate specificity not only toward N-carbamoyl-b-amino acids but
also N-carbamoyl-c-amino acids and several N-carbamoyl-a-amino acids [24]. The
enzyme showed strict stereospecificity to the L-isomers in the hydrolysis of
N-carbamoyl-a-amino acids and N-formyl- and N-acetyl-alanine [24]. These results
clearly showed the opposite stereospecificity of D-N-carbamoylase and b-ureidopro-
pionase in N-carbamoyl-a-amino acids hydrolysis.
Different combinations of hydantoin-metabolizing enzymes, namely, hydantoi-
nases and N-carbamoylases, provide various processes for the production of optically
pure amino acids (Figure 16.4) [6]. The broad substrate range of the processes is
Hydantoin N-Carbamoyl amino acid

hydrolysis hydrolysis
D-Hydantoinase
identical with
5-Monosubstituted dihydropyrimidinase N-Carbamoyl-D- D-N-Carbamoylase
D-Amino acid
D-hydantoin Pseudomonas amino acid Comamonas
Blastobacter etc. Blastobacter
Agrobacterium etc.
specific for D-hydantoin
Agrobacterium N -Carbamoylsarcosine
amidohydrolase
DL-Hydantoinase
Racemization
D-specific for N-carbamoyl-

R ATP-independent R
Base-catalysis α-amino acids
Bacillus
HN HN Pseudomonas
Hydantoin H Arthrobacter H H NH2
ATP-dependent L-N-Carbamoylase
racemase COOH R
O Pseudomonas O Alcaligenes C COOH
Pseudomonas N O NH2 Flavobacterium
Arthrobacter L-Hydantoinase
H ATP-independent Arthrobacter
Arthrobacter Pseudomonas
ATP-dependent Bacillus
Bacillus β-Ureidopropinase
Dihydroorotase L-specific for N-carbamoyl-
Pseudomonas α-amino acids
N -Methylhydantoin Pseudomonas
5-Monosubstituted amidohydrolase N-Carbamoyl-L- Chemical
ATP-dependent and decabamoylation
L-Amino acid
L-hydantoin amino acid
specific for L-hydantoin
Pseudomonas
Figure 16.4 Processes for production of optically active a-amino acids with combinations of
various hydantoin-transforming enzymes.
16.1 Overview of Microbial Hydantoin Metabolism and its Application to Biotechnology j655
valuable, especially for the production of D-amino acids and unnatural L-amino
acids [25–27].
A practical representative is the production of D-p-hydroxyphenylglycine, a
building block for semisynthetic penicillins and cephalosporins, from racemic
5-(p-hydroxyphenyl)hydantoin (Scheme 16.1). The process called the D-hydantoinase
process – involving one chemical step for racemic 5-(p-hydroxyphenyl)hydantoin
synthesis from phenol, glyoxylic acid, and urea, and two enzymatic steps with
immobilized D-hydantoinase and D-N-carbamoylase [28, 29] – has been used in
commercial production since 1995. In this process, the L-isomer of the remaining
5-(p-hydroxyphenyl)hydantoin is racemized through base catalysis under alkaline
conditions. Therefore, racemic hydantoin can be converted quantitatively into
D-p-hydroxyphenylglycine through the total process.
NH2
CHO OH
+ CO +
COOH
NH2
Amidoalkylation
CO HO
H Spontaneous
CO
HO NH racemization
HN CO NH
H HN CO
D-Hydantoinase
HO
COOH
H NHCONH2
Chemical
D-N-carbamoylase
decarbamoylation
HO
COOH
H NH2
Scheme 16.1 Reaction scheme for the D-hydantoinase process for the production of D-p-
hydroxyphenylglycine using D-hydantoinase and D-N-carbamoylase.
656
In some cases an enzyme, hydantoin racemase (E.C. 5.1.99.5), which is respon-

sible for racemizing hydantoins, is involved in hydantoin metabolism [30–32]. These
three enzymes together, hydantoinase, N-carbamoylase, and hydantoin racemase,
accomplish the total conversion of racemic DL-5-monosubstituted hydantoin deri-
vatives into the corresponding optically pure a-amino acids in a dynamic kinetic
resolution process.
16.2
D-Hydantoinase
In 1970 and 1973, Dudley et al. were the first to publish on the D-selective cleavage of
5-phenylhydantoin to N-carbamoyl-D-phenylglycine by a mammalian enzyme and on
the spontaneous in vivo racemization of the residual L-isomer [1, 2]. In 1975, Cecere
et al. [3] published work on the enzymatic production of other N-carbamoyl-D-amino
acids starting from chemically synthesized DL-5-monosubstituted hydantoin deriva-
tives using a partially purified fraction of the dihydropyrimidinase from calf liver.
They were the first to stress that this enzyme might find an industrial application for
the preparation of optically active D-amino acids. In 1978, the same group published
on the production of various N-carbamoyl-D-amino acids using an immobilized calf
liver dihydropyrimidinase preparation [33, 34]. The occurrence of the enzyme in
plant cell cultures has also been reported [35]. Rai and Taneja published work on the
use of a plant enzyme from Lens esculenta immobilized onto DEAE-cellulose for the
same purpose [36].
In the late 1970s the group of Yamada et al. in Japan found that the ability to
hydrolyze DL-5-monosubstituted hydantoin with strict D-enantiomer selectivity is
widely distributed in microorganisms and called the corresponding enzyme for
hydantoin hydrolysis as D-hydantoinase [4, 10]. With the increasing interest in the
production of D-phenylglycine and D-p-hydroxyphenylglycine, several publications
have described D-hydantoinases isolated from various microorganisms such as
Pseudomonas striata [10], Pseudomonas fluorescens DSM 84 [37], Pseudomonas sp.
AJ-11220 [17], Arthrobacter crystallopoietes AM2 [38], Agrobacterium sp. IP-I 671 [23], in
anaerobic microorganisms [39], Pseudomonas sp. KBEL 101 [40], Agrobacterium
tumefaciens [41], thermophilic microorganisms [42], Pseudomonas desmolyticum [43],
Bacillus sp. [44], Bacillus stearothermophilus SD-1 [45, 46], and Bacillus circulans [47].
Runser and coworkers described a D-hydantoinase of an Agrobacterium sp. with
remarkably high temperature and pH stability but no dihydropyrimidinase activity
[48]. Soong et al. were able to show that D-hydantoinase from Blastobacter sp. A17p-4
also is able to hydrolyze cyclic imides with bulky substituents to the corresponding
half-amides and postulated that this enzyme may also function in cyclic imide
metabolism in addition to pyrimidine metabolism [49].
In 1985 the first gene sequence of a D-hydantoinase derived from thermophilic
Bacillus sp. LU 1220 and its overproduction in Escherichia coli HB 101 ¼ was
published [50]. Subsequently, various examples of cloning, sequencing, and expres-
sion of D-hydantoinase genes from Pseudomonas putida DSM 84 [12], Bacillus
16.3 L-Hydantoinase j657
stearothermophilus NS 1122A [51], Bacillus stearothermophilus SD-1 [52, 53], and
Pseudomonas putida CCRC 12857 [54] were reported. Molecular cloning and sequenc-
ing of a cDNA encoding dihydropyrimidinase from rat liver was reported by Matsuda
et al. [55], and the complete sequencing of a 24.6 kb segment of yeast chromosome XI
including homologies to D-hydantoinases was published by Tzermia et al. [56]. Based
on this genetic information, new screening methods for the isolation of D-hydantoi-
nase-producing microorganisms were described by LaPointe et al. using a polymer-
ase-chain-reaction-amplified DNA probe to detect D-hydantoinase-producing micro-
organisms by direct colony hybridization [57].
16.3
L-Hydantoinase
In 1988, Yamashiro et al. [58, 59] reported on an L-hydantoinase from Bacillus brevis AJ
12299. This Bacillus L-hydantoinase requires ATP and Mg2 þ , Mn2 þ , or K þ as
cofactors and acts selectively on L-configured substrates. An ATP-dependent ami-
dohydrolase, N-methylhydantoin amidohydrolase, which catalyzes the reaction pre-
sented in Scheme 16.2, was first found in Pseudomonas putida 77 [60, 61]. The enzyme
catalyzes the second step reaction in the degradation route from creatinine to glycine,
via N-methylhydantoin, N-carbamoylsarcosine, and sarcosine as successive inter-
mediates [60–68]. The hydrolysis of amide compounds and coupled hydrolysis of
ATP were observed with hydantoin, L-5-methylhydantoin, glutarimide, and succini-
mide besides N-methylhydantoin. Some naturally-occurring pyrimidine compounds
such as dihydrouracil, dihydrothymine, uracil, and thymine, effectively stimulate
ATP hydrolysis by the enzyme without undergoing detectable hydrolysis themselves.
The ATP-dependent hydrolysis of 5-monosubstituted hydantoins, for example,
5-methylhydantoin, by the enzyme proved to be L-isomer specific [19]. Watabe
et al. [69] reported that an ATP-dependent hydantoin-hydrolyzing enzyme is involved
in L-methionine production from DL-5-(2-methylthioethyl)hydantoin by Pseudomonas
sp. NS671. This enzyme is different from N-methylhydantoin amidohydrolase in that
it shows no stereospecificity (DL-hydantoinase). Furthermore, it dose not hydrolyze
N-methylhydantoin, but shows a significant sequence similarity with N-methylhy-
dantoin amidohydrolase, especially in the N-terminal region [70]. Production of
L-methionine from DL-5-(2-methylthioethyl)hydantoin was also described by Ishi-
kawa et al. for resting cells of Bacillus stearothermophilus NS1122A [71] after growth of
this strain on a medium containing DL-5-(2-methylthioethyl)hydantoin as an inducer.
O COOH
ATP, Mg2+, K+
NH R * NH 2
R * + ADP + Pi
2H2O HN
HN O
O
Scheme 16.2 N-Methylhydantoin amidohydrolase catalyzed reaction.

658
The resting cells were reported to be stimulated by addition of cobalt and manganese
ions, while copper and zinc ions caused a strong inhibition of the enzymatic
activities. B. stearothermophilus NS1122A was found to have ATP-independent
DL-hydantoinase [71]. This ATP-independent DL-hydantoinase seems to have a
preference for hydantoin derivatives containing aliphatic side chains and, therefore,
differs distinctly from those enzymes found in Arthrobacter sp. by Cotoras et al. [72],
Yokozeki et al. [73–75], and Syldatk et al. [76] as well as in Flavobacterium sp. by
Nishida et al. [77]. These L-hydantoinases with preference for hydantoins derivatives
containing aromatic side chains are called L-5-arylalkylhydantoinases. They could
be used for L-stereospecific production of aromatic amino acids such as L-tryptophan
and L-naphthylalanine.
Other examples of the enzyme showing L-stereospecificity are dihydroorotase,
carboxymethylhydantoinase, and carboxyethylhydantoinase. Dihydroorotase (EC
3.5.2.3), which functions in pyrimidine biosynthesis, catalyzes reversible cyclization
of N-carbamoyl-L-aspartic acid (L-ureidosuccinate) to dihydro-L-orotate. Dihydroor-
otase from P. putida IFO 12996 was purified to homogeneity and characterized [78].
The enzyme only hydrolyzed dihydro-L-orotate and its methyl ester, and the reactions
were reversible. 5-Carboxymethylhydantoin, which is described as the product of a
non-enzymatic cyclization of N-carbamoyl-L-aspartic acid [79, 80] and occurs as a side-
product in the metabolism of dihydro-L-orotate [81], is hydrolyzed L-specifically by
carboxymethylhydantoinase (E.C. 3.5.2.2). Tsugawa et al. [82] reported L-glutamic acid
production from DL-5-carboxyethylhydantoin by microorganisms carrying L-5-car-
boxyethylhydantoinase activity.
16.4
D-N-Carbamoylase
The enzymes hydrolyzing N-carbamoyl-D-amino acids were first purified homo-

geneously from Comamonas sp. E222c [20] and Blastobacter sp. A17p-4 [83] and
characterized. This enzyme is one of the key enzymes in D-amino acids
production from DL-5-monosubstituted hydantoins. Both purified enzymes hydro-
lyzed various N-carbamoyl-D-amino acids to D-amino acids, ammonia, and carbon
dioxide. N-Carbamoyl-D-amino acids having hydrophobic groups served as good
substrates for these enzymes. These enzymes are highly stereoselective and
hydrolyze only the D-enantiomer of the N-carbamoyl-a-amino acid. This property
can be applied for the optical resolution of racemic N-carbamoyl-a-amino acids.
These enzymes did not hydrolyze b-ureidopropionate, suggesting that they are
different from the enzymes involved in the pyrimidine degradation pathway, that
is, b-ureidopropionase. Both enzymes did not require metal ions for the activity
and were sensitive to thiol reagents. The D-N-carbamoylases of the various
Agrobacterium sp. have a wide substrate specificity in common and hydrolyze
only the D-enantiomers of aliphatic and aromatic N-carbamoyl-a-amino acids [35].
The N-carbamoylsarcosine amidohydrolase from Pseudomonas putida 77 is
reported to have its biological function in creatinine metabolism and hydrolyzed
16.5 L-N-Carbamoylase j659
N-carbamoyl derivatives of D-tryptophan, D-phenylalanine, D-phenylglycine, and D-
p-hydroxyphenylglycine [64].
The enzyme from Blastobacter sp. was stable at high temperature (50 C), and
suitable for the practical production of D-amino acids. The main problems of D-N-
carbamoylases, however, still seem to be (i) their instability and rapid inactivation in
the absence of a reducing agent [84], which is probably caused by oxidation of an SH
group [28], and (ii) their inhibition by ammonium ions [85]. Grifantini et al. were able
to prove the role of cysteine-172 out of five cysteines for enzyme activity by site-
directed mutagenesis [86], while Nanba et al. were able to obtain a more thermo-
tolerant D-N-carbamoylase by substitution of Pro203 by Leu in the gene from
Agrobacterium sp. KNK712 [87]. For stabilization, the same group immobilized the
enzyme by glutaraldehyde coupling to Duolite A-568, a macroporous phenol form-
aldehyde resin [28]. Kim and Kim tried to overcome limitations in the production of D-
p-hydroxyphenylglycine with resting cells of Agrobacterium sp. I-671 by adsorptive
removal of the ammonium ions with a silicate complex [88].
16.5
L-N-Carbamoylase
N-Carbamoyl-L-amino acid-hydrolyzing activities have been found in microorgan-

isms [58, 71, 72], and applied for L-amino acid production from DL-5-monosubstituted
hydantoins. To reveal whether these activities were derived from b-ureidopropionase
activity, Ogawa et al. purified and characterized the enzyme hydrolyzing N-carba-
moyl-L-amino acid from Alcaligenes xylosoxidans [21]. The enzyme from A. xylosox-
idans resembles b-ureidopropionase in structure and in metal ion dependency
toward Mn2 þ , Ni2 þ , or Co2 þ for activity. However, the substrate specificity of the
enzyme was different from that of b-ureidopropionase. b-Ureidopropionase from P.
putida IFO 12996 hydrolyzed short-chain N-carbamoyl-L-amino acids, but not long-
chain aliphatic and aromatic N-carbamoyl-L-amino acids [24]. On the other hand, the
enzyme from A. xylosoxidans showed broad substrate specificity not only for short
chain but also for long chain and aromatic N-carbamoyl-L-amino acids. Furthermore,
the enzyme did not hydrolyze b-ureidopropionate, suggesting that its function is
distinct from b-ureidopropionase.
Aliphatic N-carbamoyl-L-amino acids are preferentially hydrolyzed by L-N-carba-
moylase from the genera Alcaligenes, Bacillus, and Pseudomonas. The L-N-carbamoy-
lase from Pseudomonas strain NS 671 [89] accepts aromatic amino acids as well as
aliphatic ones. Aromatic N-carbamoyl-L-amino acids are preferentially hydrolyzed by
the enzymes from the genera Arthrobacter and Flavobacterium.
The enzymes from Arthrobacter, Bacillus stearothermophilus NCIB 8224 and NS
1122A, and Pseudomonas sp. NS 671 have been cloned and expressed in E. coli. The
enzymes from Bacillus and Pseudomonas share approximately 38% sequence identity
with the Arthrobacter enzyme whereas the 20 amino acids known from the N-termini
of the enzymes from Alcaligenes and Pseudomonas putida IFO 12996 are completely
different.
660
16.6
Hydantoin Rasemase
The stereoselectivities of the hydantoinase and/or N-carbamoylases play an essential

role in determining the optical purity of the resulting amino acid. The racemization of
hydantoin can, however, be a rate-limiting step, resulting in unacceptably high levels of
residual substrate. To avoid this disadvantage, it is necessary to improve the effec-
tiveness of hydantoin racemization during the process. Hydantoin racemases are the
enzymes responsible for catalyzing the racemization of hydantoins, and can be used in
addition to hydantoinase and N-carbamoylases to improve production efficiency.
Hydantoin racemases occur in several microorganisms, including Pseudomonas
sp. NS761 [30], Arthrobacter aurescens DSM 3747 [31], Agrobacterium tumefaciens C58
¼ [32], and Sinorhizobium meliloti CECT 4114 [90]. The first hydantoin racemase to be
described in detail was a 5-arylalkylhydantoin racemase, which was isolated and
purified from Arthrobacter sp. DSM 3747 [91]. Only some aliphatic and aromatic
hydantoin derivatives are accepted by the enzyme out of various substrates. The
enzyme was cloned and heterologously expressed in E. coli [31]. The gene encoding
the hydantoin racemase, designated hyuA, was identified upstream of an L-N-
carbamoylase gene in the plasmid pAW16 containing genomic DNA of Arthrobacter
aurescens.
The hydantoin racemase from Pseudomonas sp. NS 671 can racemize both
enantiomers of 5-(2-methylthioethyl)hydantoin, 5-isopropylhydantoin, 5-isobutylhy-
dantoin, and 5-benzylhydantoin [92]. Suzuki et al. isolated a new hydantoin racemase
from Microbacterium liquefaciens [93]. This hydantoin racemase catalyzes the race-
mization of 5-benzylhydantoin. The purified enzyme showed a chiral preference for
L-5-benzylhydantoin rather than D-5-benzylhydantoin.
Taken together, the presence of hydantoin racemases used in industrial processes
can be of great importance for a fast and total conversion of hydantoins that racemize
chemically very slowly.
16.7
Biotechnology of Hydantoin-Transforming Enzymes
16.7.1
D-Amino Acid Production
D-p-Hydroxyphenylglycine and its derivatives are important as side-chain precursors

for semisynthetic penicillins and cephalosporines. In 1978, Yamada and coworkers
found that these amino acids can be efficiently prepared from the corresponding
5-monosubstituted hydantoins using a microbial enzyme, D-hydantoinase [4]. Inter-
estingly, the enzyme attacked various aliphatic and aromatic D-5-monosubstituted
hydantoins, yielding the corresponding D-form of N-carbamoyl-a-amino acids. Thus,
the enzyme can be used for the preparation of various D-amino acids, and the process
using D-hydantoinase is called as D-hydantoinase process.
16.7 Biotechnology of Hydantoin-Transforming Enzymes j661
Initially, the synthetic process for D-p-hydroxyphenylglycine involves two chemical
steps and one enzymatic step [94] (Scheme 16.1). The substrate, DL-5-(p-hydroxy-
phenyl)hydantoin, is synthesized through an efficient chemical method involving the
amidoalkylation reaction of phenol with glyoxylic acid and urea under acidic con-
ditions. Then, the D-5-(p-hydroxyphenyl)hydantoin is hydrolyzed enzymatically to N-
carbamoyl-D-p-hydroxyphenylglycine. Under the conditions used for the enzymatic
hydrolysis of hydantoin at pH 8–10, the L-isomer of the remaining 5-(p-hydroxyphe-
nyl)hydantoin is racemized through base catalysis. Therefore, the racemic hydantoin
can be converted quantitatively into N-carbamoyl-D-p-hydroxyphenylglycine through
this step. Decarbamoylation to D-p-hydroxyphenylglycine was performed by treating
the N-carbamoyl-D-amino acid with equimolar nitrite under acidic conditions [95].
This step can also be carried out enzymatically by using D-N-carbamoylase [20, 83].
Therefore, a sequence of two enzyme-catalyzed reactions, the D-stereospecific
hydrolysis of DL-5-(p-hydroxyphenyl)hydantoin and subsequent hydrolysis of the
D-carbamoyl derivative to D-p-hydroxyphenylglycine, is possible. Based on these
results, a new commercial process for the production of D-p-hydroxyphenylglycine
has been developed. Key developments in the D-hydantoinase process were made by
Nanba et al. at the Kaneka Corporation [96]:
1) Highly stable and active D-hydantoinases of thermophilic microorganisms were
isolated from soil samples and used in the production of D-p-hydroxyphenylgly-
cine. Because of the high stability of the enzyme, it can be reused many times in
repetitive batch reactions as an immobilized enzyme.
2) Thermostable D-N-carbamoylase was developed by means of directed revolution [29,
97, 98] using D-N-carbamoylase from Agrobacterium sp. strain KNK712 as the
parental protein. The improved D-N-carbamoylase was highly expressed in E. coli,
and immobilized by adsorption onto a macroporous phenol formaldehyde resin
with tertiary amine as a functional group, followed by crosslinking with glutaral-
dehyde [28]. The immobilized thermostable D-N-carbamoylase showed improved
stability and was industrially applied for the production of D-amino acids [96].
Since 1995, immobilized D-N-carbamoylase has been used together with immo-
bilized D-hydantoinase for the commercial production of D-p-hydroxyphenylglycine
(more than 700 repeated batch reactions have been realized) [96]. This process using
immobilized enzymes improved the reaction and purification yield, and led to
reduced by-product and waste production. As a result, both production costs and
the environmental burden have been reduced.
16.7.2
L-Amino Acid Production
Enzymatic hydantoin cleavage is already important in the production of not only

D- but also L-amino acids. Several microorganisms were found to have L-specific
hydantoin-transforming activity. These microorganisms have been applied for the
production of L-amino acids such as L-tryptophan and L-phenylalanine derivatives.
Resting cell L-hydantoinase processes were first developed for the industrial
662
production of L-tryptophan by the companies Ajinomoto and Tanabe [73–75, 77]. In

1992 the R€uttgers company tried to enter the amino acid market with a resting cell
process for the production of unnatural aromatic L-amino acids using Arthrobacter sp.
DSM 3745 or DSM 3747, which both contain an L-hydantoinase, hydantoin racemase,
and L-N-carbamoylase. Further developments have been conducted by Syldatk et al. in
collaboration with Degussa (now Evonik Industries) to improve the process:
1) L-N-Carbamoylase from Arthrobacter aurescens 3747 could be produced as
recombinant enzymes in high cell density culture in E. coli [99].
2) Purification of the recombinant L-N-carbamoylases could be optimized by expres-
sion of enzymes carrying different tags, making the purification protocols much
easier [100], and the hydantoin-cleaving enzymes from Arthrobacter aurescens
DSM 3747 could be stabilized significantly by immobilization [101, 102].
3) A more L-selective hydantoinase with higher activity was developed by directed
evolution [103]; additionally, an E. coli whole cell biocatalyst has been constructed
containing the genes of hydantoinase, hydantoin racemase, and L-N-carbamoy-
lase from Arthrobacter aurescens in optimal proportions to avoid L-N-carbamoyl
amino acid accumulation during the reaction [104].
16.7.3
Recent Application of Hydantoin Racemase
A similar process using hydantoinase, N-carbamoylase and hydantoin racemase was

developed by Suzuki et al. at Ajinomoto Co. [105]. A DNA fragment from Micro-
bacterium liquefaciens AJ 3912, containing the genes responsible for the conversion of
5-substituted-hydantoins into a-amino acids, was cloned in E. coli and sequenced.
Seven open reading frames (hyuP, hyuA, hyuH, hyuC, ORF1, ORF2, and ORF3) were
identified on the 7.5 kb fragment. The deduced amino acid sequence encoded by the
hyuA gene included the N-terminal amino acid sequence of the hydantoin racemase
from M. liquefaciens AJ 3912. The hyuA, hyuH, and hyuC genes were heterologously
expressed in E. coli. The deduced amino acid sequence of hyuP was similar to the
allantoin (5-ureido-hydantoin) permease from Saccharomyces cerevisiae, suggesting
that the hyuP protein might function as a hydantoin transporter [105]. The recom-
binant E. coli JM 109 cells were successfully used for the production of L-amino acids.
The hydantoin racemase activity was most dramatically increased in the recombinant
E. coli compared with the crude extract of M. liquefaciens AJ 3912.
16.7.4
Recent Applications of Hydantoinase
Ohishi et al. applied the cyclizing function of D-hydantoinase to the synthesis of D-

a-methylcysteine [106]. In this process, hydantoinase catalyzed D-stereoselective cycli-
zation of N-carbamoyl-S-tert-butyl-DL-a-methylcysteine. The reaction gave access to
D-5-tert-butylthiomethyl-5-methylhydantoin and N-carbamoyl-S-tert-butyl-L-a-methylcys-
teine with high optical purity and excellent yield, which could be easily separated by
16.8 Structural Analysis and Protein Engineering of Hydantoin-Transforming Enzymes j663
filtration. After hydrolysis and cleavage of the tert-butyl group, D- and L-a-methylcysteine
hydrochloride were obtained.
Optically active silyl amino acid production is another example of recent
application. Several biocatalysts (isolated enzymes as well as whole cells) have been
compared with respect to stereoselectivity for the hydrolysis of DL-5-
trimethylsilylhydantoin [107].
16.7.5
Recent Applications of N-Carbamoylase
N-Carbamylases were used independently from hydantoinase in some cases. Cell-

free extracts of Blastobacter sp. A17p-4 were used for the preparation of optically active
D-p-trimethylsilylalanine from the corresponding DL-carbamoyl amino acid [108].
Some N-carbamoylases recognize multi-chiral centers other than the a-carbon of
amino acids, which enables simultaneous resolution of multi-chiral amino acids such
as b-methylphenylalanine. Whole cells of Alcaligenes xylosoxidans having L-N-carba-
moylase activity were able to distinguish not only the configuration of the a- but also
that of the b-carbon of N-carbamoyl-b-methylphenylalanine: from the mixture of the
four diastereoisomers only threo-L-b-methylphenylalanine was produced [109, 110].
Cell-free extracts of Blastobacter sp. A17p-4 with D-N-carbamoylase activity were shown
to have stereoselectivity not only at the a-carbon but also at the b-carbon of N-carbamoyl
amino acids such as threonine, isoleucine, and b-methylphenylalanine [111].
16.7.6
Recent Application for b-Amino Acid Production
Syldatk et al.demonstrated b-aminoacid productionby the hydantoinase process [112].

Using the L-hydantoinase from Arthrobacter aurescens [113], 6-phenyl-(5,6)-dihydrour-
acil was converted into N-carbamoyl-b-phenylalanine with good enantioselectivity
towards the (S)-enantiomer. b-Ureidopropionase (N-carbamoyl-b-alanine amidohy-
drolase) was also studied as a candidate for application in the b-amino acid production.
N-Carbamoyl-b-alanine amidohydrolase from Agrobacterium tumefaciens C58 has a
broad substrate spectrum and hydrolyzes nonsubstituted N-carbamoyl-a-, -b-, -c-, and
-d-amino acids, with the greatest catalytic efficiency for N-carbamoyl-b-alanine. The
enzyme also recognizes substrate analogues substituted with sulfonic and phospho-
nic acid groups to produce the b-amino acids taurine and ciliatine, respectively. The
enzyme is able to produce monosubstituted b2- and b3-amino acids [114].
16.8
Structural Analysis and Protein Engineering of Hydantoin-Transforming Enzymes
Cyclic amidohydrolases belong to a superfamily of enzymes that are mostly found in

nucleotide metabolism. Alignment of the amino acid sequences of the amidohy-
drolase enzymes revealed that microbial hydantoinases share about 40% identity
664
with mammalian dihydropyrimidinases. The conserved residues, including the four

histidine residues, play an essential role in the metal coordination, substrate binding,
and catalysis of these functionally related cyclic amidohydrolases. Three-dimensional
structural studies together with biochemical studies have suggested that cyclic ami-
dohydrolase enzymes have the following common characteristics. They have the most
prevalent TIM barrel fold loaded with two divalent metal ions, such as Zn2 þ or Mn2 þ .
These metal ions play a critical catalytic role in the deprotonation of water molecules for
hydrophilic attack on the substrate. In these superfamily enzymes, four histidines and
one carbamoylated lysine are strictly conserved as ligand residues for the coordination
of catalytic metal ions. In addition, aspartic acid is conserved, acting as a catalytic base in
their active site. They have functional similarity in terms of the catalytic reaction they
perform [115], that is, most of them are directly involved in nucleotide metabolisms and
catalyze metal-assisted hydrolysis of cyclic amide bonds [116].
Based on the results of these structural analyses, the substrate specificity of
hydantoinase was rationally manipulated towards a commercially important aro-
matic substrate [117]. The mode of substrate binding was simulated by fitting D-(p-
hydroxyphenyl)hydantoin as a target substrate into the active site pocket and the
structural loops at the active site. Site directed and saturation mutagenesis toward
these residues were performed and induced remarkable changes in substrate
specificity towards D-(p-hydroxyphenyl)hydantoin [118].
Directed evolution was also applied to improve the enantioselectivity of a hydan-
toinase [103]. The optical preference towards the L-enantiomer was increased along
with catalytic activity by error-prone PCR and saturation mutagenesis. The produc-
tion of L-methionine was improved fivefold, and the accumulation of unwanted
intermediates was decreased fourfold, by using a whole-cell biocatalyst expressing
the mutant created by the directed evolution technique.
The crystal structure of D-N-carbamoylase was also determined [119, 120]. The
results explained well the properties of a thermostable mutant created by Nanba et al.
that had three amino acid substitutions, His57Tyr/Pro203Glu/Val236Ala, showing
an increase of approximately 19 C in the thermostable temperature [29, 96]. Most
important for stabilization were hydrophobic interactions, relaxation of the strain
energy of the peptide backbone, and the release in the strain of the side-chain
conformation.
For efficient synthesis of D-amino acid in a concerted fashion, a bifunctional
enzyme composed of D-N-carbamoylase from Agrobacterium radiobacter NRRL and
D-hydantoinase from Bacillus stearothermophilus SD1 or Bacillus thermocatenulatus
GH2 was created. The functional fusion of two consecutive enzymes offers several
potential advantages in an enzymatic process with respect to reaction kinetics and
enzyme production [121]. The resulting fusion enzyme displayed a distinct bifunc-
tional activity, and converted monosubstituted hydantoins directly into correspond-
ing D-amino acids. However, it exhibited structural instability resulting in extensive
proteolysis in vivo, probably due to the low stability of the N-terminal fusion partner
N-carbamoylase. This unfavorable property of a fusion enzyme was improved by
using DNA shuffling [122]. Through three rounds of directed evolution, an evolved
fusion enzyme with nine amino acid substitutions was obtained. This variant was
16.9 Diversity and Versatility of Cyclic Amide Transforming Enzymes and its Application j665
found to possess enhanced structural stability, leading to a sixfold increase in
performance in the synthesis of D-amino acid.
16.9
Diversity and Versatility of Cyclic Amide Transforming Enzymes and its Application
Various metabolisms of cyclic amide compounds such as cyclic ureides and cyclic
imides were analyzed in detail and applied for biotransformations. The metabolism
of nucleobases such as pyrimidines and purines involves various cyclic amide
hydrolases (EC 3.5.2.-), such as dihydropyrimidinase in reductive pyrimidine metab-
olism (Figure 16.5c) [14], barbiturase in oxidative pyrimidine metabolism [123]
(Figure 16.5b), dihydroorotase in pyrimidine biosynthesis [78], and allantoinase in
Nucleoside metabolism
Base
O
HO
Nucleobase-related-compound metabolism
HO Pyrimidine Purine
metabolism metabolism
Pi H2O
nucleoside H
nucleosidase
phosphorylase O N O
Base Base
OP O HN
NH
O Sulfur-
HO
NH containing HN
Oxidative O
Reductive Hydantoin Cyclic imide cyclic amide
uracil/thymine HN metabolism metabolism
HO O metabolism
dehydrogenase NH2
O O O O O O
phosphopentomutase O
OH
O NH NH R NH NH R NH HN NH
PO
O
HN HN
*
HN
* *
O O O S HN
O O O
HO dihydro- hydantoinase imidase allantoinase
barbiturase pyrimidinase
deoxyriboaldolase NH2
COOH COOH COOH COOH COOH O COOH
H2COH CHO O NH 2 R NH 2 R NH 2 HN NH 2
CO HCOH + CH3CHO
HN
NH 2
HN
* NH 2
*
S
*
HN
O O HN O O O
H 2CO P H2CO P O
ureidomalonase β-ureido-
propionase N-carbamoylase half-amidase
FDP COOH
COOH COOH COOH COOH NH2
(a) HOOC
+ R R O
HN
NH 2
NH 2 *NH COOH
*SH NH2
Glucose 2 2
O +
HOOC COOH
(b) (c) (d) (f) HO NH2
COOH
*
HN
OH O
HOOC (g)
COOH
O
HOOC
CH3
(e)
Figure 16.5 Overview of microbial nucleic acid pyrimidine metabolism, (d) hydantoin
and related cyclic amide metabolisms. metabolism, (e) cyclic imide metabolism,
(a) nucleoside metabolism, (b) oxidative (f) sulfur-containing cyclic amide metabolism,
pyrimidine metabolism, (c) reductive and (g) purine metabolism.
666
purine metabolism [124] (Figure 16.5g). DL-5-Monosubstituted hydantoins are

members of cyclic ureide compounds. Ogawa et al. investigated the pyrimidine-
transforming activity in a typical hydantoin-transforming bacterium, Blastobacter sp.
A17p-4, which was screened from soil as a hydantoin-assimilating bacterium for the
purpose of D-amino acid production [83]. During studies on hydantoin metabolism in
this bacterium (Figure 16.5d) it was found that it showed not only hydantoin- but also
cyclic imide-metabolizing activity [125]. A recent study revealed that the strain has not
only hydantoin-metabolizing enzymes but also enzymes specific to cyclic imide
derivatives, as described below (Figure 16.5e).
Based on the finding of cyclic imide-hydrolyzing activity in Blastobacter sp., [125]
the metabolism of various cyclic imides by microorganisms has been investigated.
Blastobacter sp. can metabolize various cyclic imides such as succinimide, malei-
mide, 2-methylsuccinimide, and glutarimide, and sulfur-containing cyclic imides
such as 2,4-thiazolidinedione and rhodanine [126]. Further investigation of the
metabolic fate of these cyclic imides showed that they were metabolized through
a novel metabolic pathway (Figure 16.5e). This pathway consists of the hydrolytic
ring-opening of cyclic imides into half-amides, hydrolytic deamidation of the half-
amides to dicarboxylates, and dicarboxylate transformation similar to that in the
tricarboxylic acid (TCA) cycle.
Cyclic imide metabolism has been applied to the production of a high-value
organic acid, pyruvate [127]. Commercial demand for pyruvate has been increasing
due to its use as an effective precursor in the synthesis of various drugs and
agrochemicals in addition to its use as a component of mammalian-cell culture
media. Pseudomonas putida s52 isolated with succinimide as the sole carbon source
exhibits highly active cyclic imide metabolism. This activity has been used for
pyruvate production from fumarate, an inexpensive cyclic imide metabolism inter-
mediate (Figure 16.6a). Bromopyruvate-resistant mutants derived from P. putida s52
produced 770 mM pyruvate from 1000 mM fumarate in 96 h.
Two novel enzymes, imidase and half-amidase, and D-hydantoinase were found to
function in this pathway. Three types of imidases with different substrate specificities
were found (Figure 16.7). An imidase with specificity toward simple cyclic imides was
O O
(a) COOH (b)
H3C NH NH2
O
N N COOH
O
COOH
HOOC OH O
(c) COOH COOH
COOH NH R NH2 R
R *
HOOC * *
S S SH
O O
Figure 16.6 Application of cyclic imide- 2,3-pyridinedicarboxyimide to 3-carbamoyl-

metabolizing enzymes to the production of a-picolinic acid, and (c) stereoselective
useful compounds. (a) pyruvate production hydrolysis of 5-substituted 2,4-
from fumarate, (b) regioselective hydrolysis of thiazolidinedione to a-mercapto acid.
Blastobacter sp. imidase Blastobacter sp. D-hydantoinase Arthrobacter ureafaciens O-86 phthalimidase
O
NH
O O
H3C H2N
NH O O N OH
O O H3C O NH
O O O
NH NH NH O
O N NH2
S O HN NH O
O O O
NH O N S
N O
HN O
O O O O
O R O O
O NH N Br
NH NH H NH O
S NH NH O
S HN N O O
O HN O O N
O
O CH3
Sulfur-containing Simple Simple cyclic ureides & Substituted Bulky Phthalimide N CH3
cyclic imides cyclic imides 2-methylsuccinimide cyclic ureides cyclic imides derivatives
O
Figure 16.7 Substrate specificities of cyclic imide-hydrolyzing enzymes.

16.9 Diversity and Versatility of Cyclic Amide Transforming Enzymes and its Application
j667
668
purified from Blastobacter sp. [128]. This enzyme is also active toward sulfur-contain-
ing cyclic imides such as 2,4-thiazolidinedione and rhodanine. Bulky cyclic imides
are hydrolyzed by the D-hydantoinase of Blastobacter sp. and mammalian dihydro-
pyrimidinases [49]. Another imidase, phthalimidase, with specificity toward phtha-
limide derivatives was found in Alcaligenes ureafaciens [129]. Enzymatic hydrolysis of
the N-iminylamide was investigated with N-iminylamidase from pig liver, which
catalyzed the hydrolysis of 3-iminoisoindolinone bearing an N-iminylamide func-
tional group [130]. This enzyme was active with typical substrates of mammalian
imidase, such as phthalimide, dihydrouracil, and maleimide. Typical substrates of
some cyclic amidases and imidases were inactive for pig liver N-iminylamidase.
Half-amides, the products of imidase, were further metabolized to dicarboxylates
by half-imidase. The enzyme was purified from Blastobacter sp. and found to be
specific toward half-amides [131]. These enzyme activities are widely distributed
among bacteria, yeasts, and molds [132].
Based on these findings, potential imidases that are applicable to the regiospecific
hydrolysis of 2,3-pyridinedicarboxyimide to 3-carbamoyl-a-picolinic acid were
screened (Figure 16.6b). 3-Carbamoyl-a-picolinic acid is a promising intermediate
for modern insecticide synthesis, but there is a synthetic difficulty in selective
amidation at one of two equivalent carboxyl groups. Phthalimide-assimilating
Arthrobacter ureafaciens O-86 was selected as the best strain and applied to the
cyclohexanone–water two-phase reaction system at pH 5.5, in which the spontaneous
non-selective hydrolysis of 2,3-pyridinedicarboxyimide was avoided while the
enzyme maintained its activity. Under optimized conditions, with the periodical
addition of 2,3-pyridinedicarboxyimide (in total, 40 mM), 36.6 mM 3-carbamoyl-
a-picolinic acid accumulated in the water phase with a molar conversion yield of
91.5% and a regioisomeric purity of 94.5% in 2 h) [129].
Based on the finding that imidase hydrolyzes sulfur-containing cyclic imides,
enzymatic stereoselective conversion of thiazolidinedione derivatives into optically
active a-mercapto acids was established in a fashion similar to the hydantoinase
process for optically active a-amino acid production (Figure 16.6c). Similar to
a-amino acids, a-mercapto acids, which contain a chiral center at a-carbon, have
received increasing attention as a novel chiral building block for the synthesis of
pharmaceuticals. Brevibacterium linens C-1 and Pseudomonas sp. Y7 were found to
produce (S)- and (R)-3-phenyl-2-mercaptopropionic acid respectively from racemic
5-benzyl-2,4-thiazolidinedione. The cyclic imide hydrolase purified from B. linens
C-1 showed allantoinase activity, indicating a metabolic relation with purine base
metabolism (Figure 16.5f and g).
While reductive pyrimidine metabolism attracted much attention, oxidative
pyrimidine metabolism was not studied in detail. The enzymes in the oxidative
pathway were investigated in Rhodococcus erythropolis JCM 3132, and the involvement
of uracil/thymine dehydrogenase, barbiturase, and ureidomalonase was revealed
(Figure 16.5b) [133]. Uracil/thymine dehydrogenase is a molybdenum-containing
iron-sulfur flavoprotein, and transformed uracil into barbiturate with methylene blue
as an electron acceptor. The enzyme activity was enhanced by cerium (Ce), a
rare-earth metal. Barbiturase, a zinc-containing cyclic amide hydrolase, transformed
References j669
barbiturate into ureidomalonate [133], and ureidomalonase successively trans-
formed ureidomalonate into malonate and urea. The structural difference between
barbituric acid and the original substrate of D-hydantoinase (dihydropyrimidinase),
dihydrouracil, is the presence of a keto-group instead of a hydrogen-group in the 6-
position of the ring. The characteristics and gene organization of barbiturase from
Rhodococcus erythropolis were revealed [123]. The amino acid sequences of D-hydan-
toinase (dihydropyrimidinase) and barbiturase do not show considerable similarity,
and the positions of these two enzymes on the dendrogram of cyclic amide
amidohydrolase family are far apart. These observations imply that the amidohy-
drolases in reductive and oxidative pyrimidine degradation pathways have developed
in different evolutionary directions. The oxidative pathway is promising for control of
the reaction equilibrium of the nucleoside phosphorylase-catalyzing base-exchange
reaction, which is useful for anti-viral nucleoside analog synthesis [134, 135].
16.10
Conclusion
The hydantoinase process has already become a common method for the preparation
of optically pure amino acids, especially for the production of D-p-hydroxyphenylgly-
cine. Nowadays, most of the annual production of D-p-hydroxyphenylglycine (about
10 000 t) is carried out by the hydantoinase process.
Recently, research on hydantoin-transforming enzymes has moved from screen-
ing in natural sources to screening in sequence databases [136] and to development of
active and robust catalyst by means of directed evolution techniques and rational
design based on known crystal structures [103, 118, 137, 138]. Process development,
including enzyme immobilization, has been a major subject in the development of
the hydantoinase process [139–142].
Future research not only on the hydantoin-transforming enzymes but also on
related cyclic amide-transforming enzymes will reveal the natural metabolic func-
tions of these enzymes and also open up new applications of these enzymes in the
chemical industries besides the production of amino acids [135].
References
1 Dudley, K.H., Bius, D.L., and Butler, T.C. 5 Olivier, R., Fascetti, E., Angelini, L., and
(1970) J. Pharmoacol. Exp. Ther., 175, Degen, L. (1979) Enzyme Microb. Technol.,
27–37. 1, 201–204.
2 Dudley, K.H. and Bius, D.L. (1973) J. 6 Ogawa, J. and Shimizu, S. (2000)
Heterocycl. Chem., 10, 173–182. Stereoselective synthesis using
3 Cecere, F., Galli, G., and Morisi, F. (1975) hydantoinase and carbamoylase in
FEBS Lett., 57, 192–194. Stereoselective Biocatalysis (ed. R.N. Patel),
4 Yamada, H., Takahashi, S., Kii, Y., and Marcel Dekker, New York, pp. 1–21.
Kumagai, H. (1978) J. Ferment. Technol., 7 Kim, G.J. and Kim, H.S. (1998) Biochem.
56, 484–491. J., 330, 295–302.
670
8 May, O., Habenicht, A., Mattes, R., 26 Syldatk, C., Müller, R., Pietzsch, M., and
Syldatk, C., and Siemann, M. Wagner, F. (1992) Microbial and
(1998) Biol. Chem., 379, enzymatic production of L-amino acids
743–747. from D,L-monosubstituted hydantoins in
9 Syldatk, C., May, O., Altenbucher, J., Biocatalytic Production of Amino Acids and
Mattes, R., and Siemann, M. (1999) Derivatives (eds D. Rozzell and F.
Appl. Microbiol. Biotechnol., 51, Wagner), Hanser Publishers, Munich,
293–309. pp. 129–176.
10 Takahashi, S., Kii, Y., Kumagai, H., and 27 Altenbuchner, J., Siemann, M., and
Yamada, H. (1978) J. Ferment Technol., 56, Syldatk, C. (2001) Curr. Opin. Biotechnol.,
492–498. 12, 559–563.
11 Morin, A., Hummel, W., and Kula, M.-R. 28 Nanba, H., Ikenaka, Y., Yamada, Y.,
(1986) Appl. Microbiol. Biotechnol., 25, Yajima, K., Takano, M., Ohkubo, K.,
91–96. Hiraishi, Y., Yamada, K., and
12 LaPointe, G., Viau, S., Leblanc, D., Takahashi, S. (1998) Biosci. Biotechnol.
Robert, N., and Morin, A. (1994) Appl. Biochem., 62, 1839–1844.
Environ. Microbiol., 60, 888–895. 29 Ikenaka, Y., Nanba, H., Yajima, K.,
13 Yamada, H., Shimizu, S., Shimada, H., Yamada, Y., Takano, M., and
Tani, Y., Takahashi, S., and Ohashi, T. Takahashi, S. (1999) Biosci. Biotechnol.
(1980) Biochimie, 62, 395–399. Biochem., 63, 91–95.
14 Vogels, G.D. and Van der Drift, C. (1976) 30 Watanabe, K., Ishikawa, T., Mukohara, Y.,
Bacteriol. Rev., 40, 403–468. and Nakamura, H. (1992) J. Bacteriol.,
15 Runser, S., Chinski, N., and Ohleyer, E. 174, 7989–7995.
(1990) Appl. Microbiol. Biotechnol., 33, 31 Wiese, A., Pietzsch, M., Syldatk, C.,
382–388. Mattes, R., and Altenbuchner, J. (2000)
16 Dudley, K.H., Butler, T.C., and Bius, D.L. J. Bacteriol., 80, 217–230.
(1974) Drug Metab. Dispos., 2, 103–112. 32 Las Heras-Vazquez, F.J., Martinez-
17 Yokozeki, K. and Kubota, K. (1987) Agric. Rodrıguez, S., Mingorance-Cazorla, L.,
Biol. Chem., 51, 721–728. Clemente- Jimenez, J.M., and
18 Maguire, J.H. and Dudley, K.H. (1978) Rodrıguez-Vico, F. (2003) Biochem.
Drug Metab. Dispos., 6, 140–145. Biophys. Res. Commun., 303, 541–547.
19 Ogawa, J., Kim, J.M., Nirdnoy, W., 33 Dinelli, D., Marconi, W., Cecere, F.,
Amano, Y., Yamada, H., and Shimizu, S. Galli, G., and Morisi, F. (1978) Enzyme
(1995) Eur. J. Biochem., 229, 284–290. Engineering, Plenum Publishing
20 Ogawa, J., Shimizu, S., and Yamada, H. Corporation, vol. III, pp. 477–481.
(1993) Eur. J. Biochem., 212, 685–691. 34 Cecere, F., Marconui, W., Morisi, F., and
21 Ogawa, J., Miyake, H., and Shimizu, S. Rappuoli, B. (1978) German patent DE
(1995) Appl. Microbiol. Biotechnol., 43, 2615594 AI.
1039–1043. 35 Morin, A. (1993) Enzyme Microb. Technol.,
22 Ogawa, J. and Shimizu, S. (1997) J. Mol. 15, 208–214.
Catal. B: Enzym., 2, 163–176. 36 Rai, R. and Taneja, V. (1998) Appl. Microb.
23 Runser, S.M. and Meyer, P.C. (1993) Eur. Biotechnol., 50, 658–662.
J. Biochem., 213, 1315–1324. 37 Morin, A., Hummel, W., Sch€ utte, H., and
24 Ogawa, J. and Shimizu, S. (1994) Eur. J. Kula, M.-R. (1986) Biotechnol. Appl.
Biochem., 223, 625–630. Biochem., 8, 564–574.
25 Syldatk, C., Müller, R., Siemann, M., and 38 M€oller, A., Syldatk, C., Schulze, M., and
Wagner, F. (1992) Microbial and Wagner, F. (1988) Enzyme Microb.
enzymatic production of D-amino acids Technol., 10, 618–625.
from D,L-monosubstituted hydantoins in 39 Morin, A., Touzel, J.-P., Lafond, A., and
Biocatalytic Production of Amino Acids and Leblanc, D. (1991) Appl. Microbiol.
Derivatives (eds D. Rozzell and F. Biotechnol., 35, 536–540.
Wagner), Hanser Publishers, Munich, 40 Kim, D.-M., Kim, G.-J., and Kim, H.-S.
pp. 75–128. (1994) Biotechnol. Lett., 16, 11–16.
References j671
41 Durham, D.R. and Weber, J.E. (1995) 58 Yamashiro, A., Yokozeki, K., Kano, H.,
Biochem. Biophys. Res. Commun., 216, and Kubota, K. (1988) Agric. Biol. Chem.,
1095–1100. 52, 2851–2856.
42 Keil, O., Schneider, M., and Rasor, J.P. 59 Yamashiro, A., Kubota, K., and Yokozeki,
(1995) Tetrahedron: Asymmetry, 6, K. (1988) Agric. Biol. Chem., 52,
1257–1260. 2857–2863.
43 Gokhale, D.V., Bastawde, K.B., Patil, S.G., 60 Kim, J.-M., Shimizu, S., and Yamada, H.
Kalkote, U.R., Joshi, R.R., Joshi, R.A., (1987) Biochem. Biophys. Res. Commun.,
Ravindranathan, T., Gaikwad, B.G., 142, 1006–1012.
Jogadand, V.V., and Nene, S. 61 Shimizu, S., Kim, J.-M., and
(1996) Enzyme Microb. Technol., 18, Yamada, H. (1989) Clin. Chim. Acta, 185,
353–357. 241–252.
44 Sharma, A. and Vohra, R.M. (1997) 62 Yamada, H., Shimizu, S., Kim, J.-M.,
Biochem. Biophys. Res. Commun., 234, Shinmen, Y., and Sakai, T. (1985) FEMS
485–488. Microbiol. Lett., 30, 337–340.
45 Lee, S.-G., Lee, D.-C., Hong, S.-P., 63 Shimizu, S., Kim, J.-M., Shinmen, Y., and
Sung, M.-H., and Kim, H.-S. (1995) Yamada, H. (1986) Arch. Microbiol., 145,
Appl. Microbiol. Biotechnol., 322–328.
43, 270–276. 64 Kim, J.-M., Shimizu, S., and
46 Lee, S.-G., Lee, D.-C., and Kim, H.-S. Yamada, H. (1986) J. Biol. Chem., 261,
(1997) Appl. Biochem. Biotechnol., 62, 11832–11839.
251–266. 65 Kim, J.-M., Shimizu, S., and Yamada, H.
47 Luksa, V., Starkuviene, V., Starkuviene, (1987) Arch. Microbiol., 147, 58–63.
V., and Dagys, R. (1997) Appl. Biochem. 66 Kim, J.-M., Shimizu, S., and Yamada, H.
Biotechnol., 62, 219–231. (1987) FEBS Lett., 210, 77–80.
48 Runser, S. and Ohleyer, E. (1990) 67 Kim, J.-M., Shimizu, S., and
Biotechnol. Lett., 12, 259–264. Yamada, H. (1986) Agric. Biol. Chem., 50,
49 Soong, C.-L., Ogawa, J., Honda, M., and 2811–2816.
Shimizu, S. (1999) Appl. Environ. 68 Kim, J.-M., Shimizu, S., and
Microbiol., 65, 1459–1462. Yamada, H. (1987) Agric. Biol. Chem., 51,
50 Jacob, E., Henco, K., Marcinowski, S., and 1167–1168.
Schenk, G. (1985) German patent DE 69 Ishikawa, T., Watabe, K., Mukohara, Y.,
3535987. and Nakamura, H. (1997) Biosci.
51 Mukohara, Y., Ishikawa, T., Watabe, K., Biotechnol. Biochem., 61, 185–187.
and Nakamura, H. (1994) Biosci. 70 Watabe, K., Ishikawa, T., Mukohara, Y.,
Biotechnol. Biochem., 58, 1621–1626. and Nakamura, H. (1992) J. Bacteriol.,
52 Lee, D.-C., Kim, G.-J., Cha, Y.-K., Lee, C.- 174, 962–969.
Y., and Kim, H.-S. (1997) Biotechnol. 71 Ishikawa, T., Mukohara, Y., Watabe, K.,
Bioeng., 56, 449–455. Kobayashi, S., and Nakamura, H.
53 Kim, G.-J. and Kim, H.-S. (1998) Biochem. (1994) Biosci. Biotechnol. Biochem., 58,
Biophys. Res. Commun., 243, 96–100. 265–270.
54 Chien, H.R., Jih, Y.-L., Yang, W.-Y., and 72 Syldatk, C., Cotoras, D., Dombach, G.,
Hsu, W.-H. (1998) Biochim. Biophys. Acta, Groß, C., Kallwaß, H., and
1395, 68–77. Wagner, F. (1987) Biotechnol. Lett., 9,
55 Matsuda, K., Sakata, S., Kaneko, M., 25–30.
Hamajima, N., Nonaka, M., Sasaki, M., 73 Yokozeki, K., Sano, K., Eguchi, C.,
and Tamaki, N. (1996) Biochim. Biophys. Yamada, K., and Mitsugi, K. (1987)
Acta, 1307, 140–144. Agric. Biol. Chem., 51, 363–369.
56 Tzermia, M., Horaitis, O., and 74 Yokozeki, K., Sano, K., Eguchi, C.,
Alexandraki, D. (1994) Yeast, 10, 663–679. Iwagami, K., and Mitsugi, K. (1987) Agric.
57 LaPointe, G., Leblanc, D., and Morin, A. Biol. Chem., 51, 729–736.
(1995) Appl. Microbiol. Technol., 42, 75 Yokozeki, K., Hirose, Y., and Kubota, K.
895–900. (1987) Agric. Biol. Chem., 51, 737–746.
672
76 Syldatk, C. and Wagner, F. (1990) J. Food 95 Shimizu, S., Shimada, H., Takahashi, S.,
Biotechnol., 4, 87–95. Ohashi, T., Tani, Y., and Yamada, H.
77 Nishida, Y., Nakamichi, K., Nabe, K., and (1980) Agric. Biol. Chem., 44,
Tosa, T. (1987) Enzyme Microb. Technol., 9, 2233–2234.
721–725. 96 Nanba, H., Yasohara, Y., Hasegawa, J.,
78 Ogawa, J. and Shimizu, S. (1995) Arch. and Takahashi, S. (2007) Org. Proc. Res.,
Microbiol., 164, 353–357. 11, 503–508.
79 Akamatsu, N. (1960) J. Biochem., 47, 97 Ikenaka, Y., Nanba, H., Yajima, K.,
809–819. Yamada, Y., Takano, M., and Takahashi, S.
80 Daniel, R., Kokel, B., Caminade, E., (1998) Biosci. Biotechnol. Biochem., 62,
Martel, A., and Le Goffic, F. (1996) Anal. 1668–1671.
Biochem., 239, 130–135. 98 Ikenaka, Y., Nanba, H., Yajima, K.,
81 Lieberman, I. and Kornberg, A. (1954) Yamada, Y., Takano, M., and Takahashi, S.
J. Biol. Chem., 207, 911–924. (1998) Biosci. Biotechnol. Biochem., 62,
82 Tsugawa, R., Okumura, S., Ito, T., and 1672–1675.
Katsuga, N. (1966) Agric. Biol. Chem., 30, 99 Wilms, B., Hauck, A., Reuss, M., Syldatk,
27–34. C., Mattes, R., Siemann, M., and
83 Ogawa, J., Chung, M.C.-M., Hida, S., Altenbuchner, J. (2001) Biotechnol.
Yamada, H., and Shimizu, S. (1994) Bioeng., 73, 95–103.
J. Biotechnol., 38, 11–19. 100 Pietzsch, M., Wiese, A., Ragnitz, K.,
84 Louwrier, A. and Knowles, C.J. Wilms, B., Altenbuchner, J., Mattes, R.,
(1997) Biotechnol. Appl. Biochem., 25, and Syldatk, C. (2000) J. Chromatogr. B,
143–149. 737, 179–186.
85 Kim, G.-J. and Kim, H.-S. (1994) 101 Pietzsch, M., Oberreuter, H., Petrovska,
Biotechnol. Lett., 16, 17–22. B., Ragnitz K., and Syldatk, C. (1998)
86 Grifantini, R., Pratesi, C., Galli, G., and Immobilization of hydantoin cleaving
Grandi, G. (1996) J. Biol. Chem., 271, enzymes from Arthrobacter aurescens
9326–9331. DSM 3747—Effect of the coupling
87 Nanba, H., Ikenaka, Y., Yamada, Y., method on the stability of the L-N-
Yajima, K., Takano, M., and Takahashi, S. carbamoylase in Stability and
(1998) Biosci. Biotechnol. Biochem., 62, Stabilization of Biocatalysts (eds A.
875–881. Ballesteros, F.J. Plou, J.L. Iborra, and P.
88 Kim, G.-J. and Kim, H.-S. (1995) Enzyme Halling), Elsevier Science, Amsterdam,
Microb. Technol., 17, 63–67. pp. 517–522.
89 Ishikawa, T., Watabe, K., Mukohara, Y., 102 Ragnitz, K., Pietzsch, M., and Syldatk, C.
and Nakamura, H. (1996) Biosci. (2001) J. Biotechnol., 92, 179–186.
Biotechnol. Biochem., 60, 612–615. 103 May, O., Nguyen, P.T., and Arnold, F.H.
90 Martinez- Rodrıguez, S., Las Heras- (2000) Nat. Biotechnol., 18, 317–320.
Vazquez, F.J., Mingorance-Cazorla, L., 104 Wilms, B., Wiese, A., Syldatk, C., Mattes,
Clemente- Jimenez, J.M., and Rodrıguez- R., and Altenbuchner, J. (2001)
Vico, F. (2004) Appl., Environ., Microbiol., J. Biotechnol., 86, 19–30.
70, 625–630. 105 Suzuki, S., Takenaka, Y., Onishi, N., and
91 Pietzsch, M., Syldatk, C., and Wagner, F. Yokozeki, K. (2005) Biosci. Biotechnol.
(1992) Ann. N.Y. Acad. Sci., 672, 478–483. Biochem., 69, 1473–1482.
92 Watabe, K., Ishikawa, T., Mukohara, Y., 106 Ohishi, T., Nanba, H., Sugawara, M.,
and Nakamura, H. (1992) J. Bacteriol., Izumida, M., Honda, T., Mori, K.,
174, 3461–3466. Yanagisawa, S., Ueda, M., Nagashima, N.,
93 Suzuki, S., Onishi, N., and Yokozeki, K. and Inoue, K. (2007) Tetrahedron Lett., 48,
(2005) Biosci. Biotechnol. Biochem., 69, 3437–3440.
530–536. 107 Pietzsch, M., Waniek, T., Smith, R.J.,
94 Takahashi, S., Kii, Y., Kumagai, H., and Bratovanov, S., Bienz, S., and
Yamada, H. (1979) J. Ferment. Technol., Syldatk, C. (2000) Chem. Month., 131,
57, 328–332. 645–653.
References j673
108 Yamanaka, H., Kawamoto, T., and 122 Kim, G.-J., Cheon, Y.-H., and
Tanaka, A. (1997) J. Ferment. Bioeng., 84, Kim, H.-S. (2000) Biotechnol Bioeng.,
181–184. 68, 211–217.
109 Ogawa, J., Ryouno, A., Xie, S.-X., Rakesh, 123 Soong, C.-L., Ogawa, J., Sakuradani, E.,
V.M., Indrati, R., Miyakawa, H., Ueno, T., and Shimizu, S. (2002) J. Biol. Chem.,
and Shimizu, S. (1999) Biotechnol. Lett., 277, 7051–7058.
21, 711–713. 124 Bongaerts, G.P.A. and Vogels,
110 Ogawa, J., Ryono, A., Xie, S.-X., G.D. (1976) J. Bacteriol., 125,
Vohra, R.M., Indrati, R., Miyakawa, H., 689–697.
Ueno, T., and Shimizu, S. (2001) 125 Ogawa, J., Honda, M., Soong, C.-L., and
J. Mol. Catal. B: Enzym., 12, 71–75. Shimizu, S. (1995) Biosci. Biotechnol.
111 Ogawa, J., Ryouno, A., Xie, S.-X., Biochem., 59, 1960–1962.
Rakesh, V.M., Indrati, R., Miyakawa, H., 126 Ogawa, J., Soong, C.-L., Honda, M., and
Ueno, T., Ikenaka, Y., Nanba, H., Shimizu, S. (1996) Appl. Environ.
Takahashi, S., and Shimizu, S. (1999) Microbiol., 62, 3814–3817.
Appl. Microbiol. Biotechnol., 55, 797–801. 127 Ogawa, J., Soong, C.-L., Ito, M., and
112 Bretschneider, U., Syldatk, C., and Shimizu, S. (2001) J. Mol. Catal. B:
Rudat, J. (2010) J. Biotechnol., 150, 123. Enzym., 11, 355–359.
113 May, O., Siemann, M., Pietzsch, M., 128 Ogawa, J., Soong, C.-L., Honda, M., and
Mattes, R., and Syldatk, C. (1998) Shimizu, S. (1997) Eur. J. Biochem., 243,
J. Biotechnol., 61, 1–13. 322–327.
114 Martınez-G omez, A.I., 129 Ogawa, J., Soong, C.-L., Ito, M.,
Martınez-Rodrıguez, S., Pozo-Dengra, J., Segawa, T., Prana, T., Prana, M.S., and
Tessaro, D., Servi, S., Clemente- Jimenez, Shimizu, S. (2000) Appl. Microbiol.
J.M., Rodrıguez-Vico, F., and Heras- Biotechnol., 54, 331–334.
Vazquez, F.J.L. (2009) Appl. Environ. 130 Huang, C.H. and Yang, Y.S. (2005)
Microbiol., 75, 514–520. Protein Expression Purif., 40,
115 Abendroth, J., Niefind, K., and 203–211.
Schomburg, D. (2002) J. Mol. Biol., 320, 131 Soong, C.-L., Ogawa, J., and Shimizu, S.
143–156. (2000) Appl. Environ. Microbiol., 66,
116 Nam, S.-H., Park, H.-S., and Kim, H.-S. 1947–1953.
(2005) Chem. Rec., 5, 298–307. 132 Soong, C.-L., Ogawa, J., Sukiman, H.,
117 Cheon, Y.-H., Park, H.-S., Kim, J.-H., Prana, T., Prana, M.S., and Shimizu, S.
Kim, Y., and Kim, H.-S. (2004) (1998) FEMS Microbiol. Lett.,
Biochemistry, 43, 7413–7420. 158, 51–55.
118 Lee, S.-C., Chang, Y.-J., Shin, D.-M., 133 Soong, C.-L., Ogawa, J., and Shimizu, S.
Han, J., Seo, M.-H., Fazelinia, H., (2001) Biochem. Biophys. Res. Commun.,
Maranas, C.-D., and Kim, H.-S. (2009) 286, 222–226.
Enzyme Microb. Technol., 44, 170–175. 134 Yokozeki, K., Shirae, H., and Kubota, K.
119 Nakai, T., Hasegawa, T., Yamashita, E., (1990) Ann. N.Y. Acad. Sci.,
Yamamoto, M., Kumasaka, T., Ueki, T., 613, 757–759.
Nanba, H., Ikenaka, Y., Takahashi, S., 135 Ogawa, J., Soong, C.-L., Kishino, S., Li,
Sato, M., and Tsukihara, T. (2000) Q.-S., Horinouchi, N., and Shimizu, S.
Structure, 8, 729–739. (2006) Biosci. Biotechnol. Biochem., 70,
120 Martınez-Rodrıguez, S., Garcıa-Pino, A., 574–582.
Las Heras-Vazquez, F.J., Clemente- 136 Cai, Y., Trodler, P., Jiang, S., Zhang, W.,
Jimenez, J.M., Rodrıguez-Vico, F., Loris, Wu, Y., Lu, Y., Yang, S., and
R., Garcıa-Ruiz, J.M., and Gavira, J.A. Jiang, W. (2009) FEBS J., 276,
(2008) Acta Crystallogr. Sect. F, 64, 3575–3588.
1135–1138. 137 Lo, C.-K., Kao, C.-H., Wang, W.-C., Wu,
121 Kim, G.-J., Lee, D.-E., and Kim, H.-S. H.-M., Hsu, W.-H., Lin, L.-L., and
(2000) Appl. Environ. Microbiol., 66, Hu, H.-Y. (2009) Process Biochem., 44,
2133–2138. 309–315.
674
138 Martınez-Rodrıguez, S., Martınez- 140 Chiang, C.-J., Chern, J.-T.,

omez, A.I., Rodrıguez-Vico, F.,
G Wang, J.-Y., and Chao, Y.-P. (2008)
Clemente-Jimenez, J.M., and Las J. Agric. Food Chem.,
Heras-Vazquez, F.J. (2010) 56, 6348–6354.
Appl. Microbiol. Biotechnol., 85, 141 Aranaz, I., Acosta, N., and Heras, A.
441–458. (2009) J. Mol. Catal. B: Enzym.,
139 Andrade, J.M., Peres, R.C.D., 58, 54–64.
Oestreicher, E.G., Antunes, O.A.C., and 142 Yen, M.-C., Hsu, W.-H., and Lin, S.-C.
Dariva, C.J. (2008) J. Mol. Catal. B: (2010) Process Biochem., 45,
Enzym., 55, 185–188. 667–674.
j675
17
Hydrolysis and Synthesis of Peptides
Timo Nuijens, Peter J.L.M. Quaedflieg, and Hans-Dieter Jakubke
17.1
Introduction
Peptides and proteins play a fundamental role in the formation and maintenance of
the structure and function of living systems. Peptides comprise various biologically
active linear and cyclic compounds with diverse functions. The different classes of
peptides include, for instance, hormones and other signaling or regulatory factors,
antibiotics, alkaloids, toxins, enzyme inhibitors, and sweeteners. There is perma-
nently great interest in pharmaceutically active peptides and proteins since they have
many applications and great potential in medicine, such as in cardiovascular
diseases, mental illness, connective tissue diseases, cancer, regulation of fertility
and growth, and the control of pain. Furthermore, peptides find applications in
human and animal nutrition and are being used as cosmetic ingredients. Therefore,
the demand for peptides and proteins is enormous, and continuously increasing.
In a peptide chain, amino acids are linked together by bonds between the carboxyl
group of one and the amino group of another amino acid, known as peptide bonds.
This amide or peptide bond has some characteristics of a double bond: it does not
rotate freely and is shorter than other CN bonds. Nature provides a wide range of
special enzymes, the proteolytic enzymes or correctly designated as peptidases, that
can cleave these bonds in peptide and protein substrates. In contrast, for catalyzing
the formation of peptide bonds the number of efficient enzymes is rather low.
Peptidases catalyze a single reaction, the hydrolysis of a peptide bond. The ubiquitous
distribution among all life forms and their enormous diversity of function makes
peptidases one of the most fascinating families of enzymes. As a result of complete
analysis of several genomes it has been shown that about 2% of all gene products are
proteolytic enzymes. In biological and biochemical research proteolytic enzymes play
a contrary role: researchers either hate them or love them. In the first case, the only
good peptidase is a dead one, no longer capable of degrading the desired protein
during isolation and purification. Irreversible inhibition of any contaminating
proteolytic enzyme is the best way to solve this problem. However, for most purposes
proteolytic enzymes are of great importance. Owing to their special physiological
and Oliver May.
j 17 Hydrolysis and Synthesis of Peptides
676
functions, some proteolytic enzymes are active in degrading proteins for digestive
and nutritional purposes. These enzymes act both extracellularly (e.g., in the intestine
of animals) and intracellularly (in the hydrolytic subcellular organelles, preferentially
in liver and kidney cells). Other peptidases are responsible for controlling processes,
for example, they can act to cause limited proteolysis of peptide and protein
substrates. In limited proteolytic processes a single susceptible peptide bond may
be cleaved followed by a dramatic biological response due to the formed products.
Physiological functions are often the result of proteolytic conversion of inactive
precursors into biologically active proteins, for example, in blood coagulation,
prohormone, or proenzyme activation. Pancreatic peptidases frequently exist as
zymogens, a special inactive proenzyme arrangement that ensures that the pancreas
does not digest itself. These enzymes have their function outside cells and will be
activated by another peptidase at the place of action. The number of peptidases within
the cell is more numerous but much more difficult to investigate in comparison with
the extracellular enzymes [1]. A much smaller group is the cell-surface peptidases,
which are specialized in the hydrolysis of relatively simple peptides rather than
proteins. This group of peptidases does not need activation. Usually the biological
function consists of the inactivation of signaling peptides to terminate a hormonal or
neuropeptide signal but sometimes they activate peptide substrates, for example, in
the conversion of angiotensin I into angiotensin II [2, 3].
Contrary to the well-known native function of peptidases the reverse reaction, the
peptidase-catalyzed peptide bond formation, can only be successfully carried out by
manipulating the reaction conditions, the enzyme, or the substrate. Besides enzy-
matic techniques, classical chemical synthesis in solution or on the solid-phase and
recombinant techniques belongs to the most important methods of peptide
synthesis.
This chapter gives an overview of the present state-of-the-art of the use of proteases
in the hydrolysis and synthesis of peptides.
17.2
Hydrolysis of Peptides
17.2.1
Peptide-Cleaving Enzymes
17.2.1.1 Introduction and Terminology

More than 500 proteolytic enzymes are known and, in a general sense, they all
catalyze the same reaction: hydrolysis of peptide bonds. An excellent handbook [4]
provides a ready reference to over 500 proteolytic enzymes known up to 2004. These
enzymes are classified as peptidases or proteases. In the past there has been
widespread confusion about the exact meaning of the terms proteases, peptidases,
and proteinases, as well as proteolytic enzymes. There is no doubt that the term
proteolytic enzymes is the most generally understood in current usage. However,
this is ambiguous since many of the enzymes that are capable of hydrolyzing peptide
17.2 Hydrolysis of Peptides j677
Figure 17.1 Proposed terms for the major types of peptidases.
bonds do not accept proteins as substrates. The Nomenclature Committee of the

International Union of Biochemistry and Molecular Biology (NC-IUBMB) recom-
mends the term peptidase as the general term for all peptide bond-hydrolyzing
enzymes. The EC List can be found in its revised version on the World Wide Web
(www) at http://www.chem.qmw.ac.uk/iubmb/enzyme/index.html.
Figure 17.1 shows the accepted terms for the major types of peptidases. Herein, the
meanings of the terms are described by the italicized semi-systematic descriptions.
The terms in bold type are preferred, whereas the terms in parentheses have
historical precedence and are satisfactory when used in the correct context. Most
of the peptidases fall into one of two categories, depending on the positional
specificity of the peptide bond cleavage process. An enzyme is said to be an
endopeptidase when the susceptible peptide bond is an internal one in a peptide
or protein. In contrast, an enzyme is termed an exopeptidase when the susceptible
peptide linkage is at the carboxyl terminus or at the amino terminus of the substrate.
In the EC List there are also terms for subtypes of exopeptidases and endopeptidases.
Exopeptidases acting at the free N-terminus liberate a single amino acid residue
(aminopeptidases) or a dipeptide or a tripeptide (dipeptidyl-peptidases and tripepti-
dyl-peptidases), whereas those acting at the free C-terminus liberate a single residue
(carboxypeptidases) or a dipeptide (peptidyl-dipeptidases). Furthermore, other exo-
peptidases are specific for dipeptides (dipeptidases) or remove terminal residues that
are substituted, cyclized, or linked by isopeptide bonds (omega peptidases). Endo-
peptidases act on bonds in the middle of the peptide chain (Figure 17.2). The term
oligopeptidase is used to refer to endopeptidases that act optimally on oligopeptide
substrates rather than on proteins.
Peptidases differ in the specificities that they display in a hydrolytic reaction. It is
somewhat simplistic to designate a peptidase on the basis of a single amino acid
Figure 17.2 Action of endo-peptidases and exopeptidases.

678
residue at the active site. Near the active site of the peptidase there is a pocket in the
surface of the enzyme molecule that is specific for amino acid side chains of the
substrate. Owing to different interactions in this region there are great differences in
the so-called primary specificity of the peptidases. Trypsin, for example, cleaves only
those peptide bonds adjacent to the amino acids lysine or arginine that carry a positive
charge and are hydrophilic. In the binding pocket of trypsin a negatively charged
aspartic acid unit is at the back, holding the positively charged lysine or arginine side
chain in the pocket by electrostatic forces. Despite the fact that this pocket for specific
side chains is obviously important for binding, it is not the only binding site. It has
been concluded from kinetic studies that the binding of substrates (and inhibitors)
involves interactions at several subsites on either side of the pair of residues
containing the peptide bond to be hydrolyzed. The enzyme and substrate must be
fixed at several points, so that the susceptible bond is oriented at the active site in
optimal configuration.
In 1967, a system of nomenclature to describe the interaction of peptidases with
their substrates was introduced by Schechter and Berger [5]. According to this system
the binding site for a peptide substrate in the active site of a peptidase is envisioned as
a series of subsites S that interact with the amino acid building blocks P of the peptide
or protein substrate (Figure 17.3). The amino acid residues of the substrate are
denoted by P and P0 , respectively, which interact with the corresponding S and S0
subsites within the active site of the peptidase. The sites are numbered from the
catalytic site, S1 . . . Sn towards the N-terminus of the peptide substrate, and S01 . . . S0n
towards the C-terminus. In analogy, the residues that they accommodate are
numbered P1 . . . Pn, and P01 . . . P0n , respectively. The arrow indicates the site of
enzymatic cleavage of the substrate between the residues P1 P01 . With the increasing
knowledge of the amino acid sequences of peptidases and particularly when the 3D
structure of enzymes began to emerge, a functional division of peptidases became
possible. Detailed mapping of the active sites has provided a better understanding of
the interaction of substrate and peptidase and has permitted both the design and
synthesis of highly specific inhibitors as well as a useful prediction of the outcome of
the reverse peptidase action in peptide synthesis (Section 17.3.3).
Figure 17.3 Simplified representation of the They interact with the corresponding S and S0
peptidase specificity according to Schechter and subsites of the enzyme active site, respectively.
Berger [5]. The amino acid residues of the The arrow indicates the site of hydrolytic
substrate are denoted by P and P0 , respectively. cleavage.
The general mechanism for the hydrolysis of a peptide bond is shown in
Scheme 17.1. Water attacks the electron-deficient carbonyl atom, generating first a
tetrahedral adduct, which then eliminates the amine fragment and produces the acid.
The process is characterized by transferring the aminoacyl moiety of the peptide to
water. In this type of group-transfer reaction the nucleophilic cosubstrate is water;
55.5 M water is the most ubiquitous weak nucleophile in degradative enzymatic
processes in the cell. Under physiological conditions the hydrolysis of peptide bonds
will proceed in the absence of peptidases, but only at an exceedingly low rate, since the
reactants only rarely attain the high internal energy required for the hydrolytic process.
Scheme 17.1 General mechanism for the hydrolysis of a peptide bond.
In contrast, enzymes allow the reaction to follow a different pathway from the
substrate to the products and, therefore, reduce the energy barriers. During the
reaction new intermediate states of highest energy appear that lower the internal
energy barriers, that is, the high-energy transitions between one intermediate and the
following one. Proteolysis is functionally irreversible, since energy is liberated in the
hydrolysis of peptide bonds because the ionized hydrolysis products are thermody-
namically more stable. On the other hand, aminoacyl-group transfer is involved in
protein biosynthesis. As a result of the ionized state of amino acids at physiological
pH, the attack by the amino group of another amino acid to form a peptide bond
would involve formal expulsion of O2. This species is very instable and, therefore,
the reaction would not proceed to any reasonable extent. In protein biosynthesis the
carboxylate must be chemically modified so that an oxygen atom can be eliminated
with low activation energy. The key concept in protein biosynthesis is that an activated
C-terminal acyl group of a growing peptide is transferred to the amino group of an
acyl activated amino acid catalyzed by the ribosomal peptidyltransferase, resulting in
a newly formed peptide bond. The resulting C-terminal acyl activated peptide is again
reacted with an amino group of an acyl activated amino acid and is thus elongated in a
stepwise manner. The reaction takes place via the transfer of a peptidyl residue from
peptidyl-tRNA in the ribosomal P site to the amino group of the aminoacyl-tRNA in
the A site. Extensive research has been carried out in recent decades on the structure
and function of the ribosomal peptidyltransferase. Recently, the Nobel Prize in
Chemistry (2009) was awarded to V. Ramakrishnan, T.A. Steitz and A.E. Yonath for
their contributions to this field.
Zhang and Cech [6] demonstrated that an in vitro-selected ribozyme can catalyze
the same type of peptide bond formation as a ribosome. The ribozyme resembles
the ribosome in such a way that a very specific RNA structure is necessary for
substrate binding and catalysis, and both amino acids to be coupled are attached to
680
nucleotides. These results provide evidence that RNA itself can generate peptides and
support the RNA world hypothesis in biological evolution.
Since the ribosomal peptidyltransferase is not suitable for practical use as a simple
C-N ligase and, in addition, the multienzyme complexes involved in bacterial peptide
synthesis [7] do not seem to possess a general applicability, only the reverse catalytic
potential of peptidases can be considered as valuable supplement to chemical
coupling methods (cf. Section 17.3). In addition, peptidases have been used suc-
cessfully for enzymatic manipulation of protecting groups in peptide synthesis and
the C-terminal modification of peptides (cf. Section 17.3.8) [8–10].
17.2.1.2 Catalytic Mechanism [11, 12]

The overall process of peptide bond scission is identical in all classes of peptidases
and differences between the catalytic mechanisms are rather subtle. The attack on the
carbonyl group of the peptide bond requires a nucleophilic moiety, either oxygen or
sulfur, to approach the slightly electrophilic carbonyl carbon atom. To remove a
proton from the attacking nucleophile, general base catalysis will assist this process.
Furthermore, some type of electrophilic action on the carbonyl oxygen increases the
polarization of the CO-bond.
Generally, the four classes of peptidases (serine, cysteine, aspartic, and metallo-
peptidases) differ in the groups that perform nucleophilic attack, general base
catalysis, and electrophilic assistance. In addition, different groups are involved in
the breakdown of the tetrahedral intermediate that is formed after the initial nucle-
ophilic attack, requiring general acid catalysis to promote the departure of the amine
fragment. The four types of peptidases are based on the different catalytic mechan-
isms, which were first recognized by the use of some group-specific inhibitors.
The reactive serine residue in the active site of serine peptidases (but also in other
serine hydrolases, such as acetylcholine esterase) reacts in an irreversible step with
organophosphate compounds, for example, diisopropyl phosphofluoridate (DFP or
DipF), resulting in the death of the appropriate enzyme. Owing to the high toxicity of
DFP other reagents, for example, phenylmethylsulfonyl fluoride (PMSF) and 3,4-
dichloroisocoumarin (3,4-DCl), have been used instead. The reactive cysteine residue
of cysteine peptidases is susceptible to oxidation and can react with various reagents:
iodoacetate, N-ethyl-maleimide, heavy metals (e.g., Hg), and with the highly selective
inhibitor N-[L-3-trans-carboxyoxiran-2-carbonyl-L-leucyl-amido(4-guanidino)butane]
(E-64). The highly acidic pH optima led to the first recognition of aspartic peptidases.
Later, with pepstatin A from a strain of Streptomyces, a specific inhibitor was found.
Chelating agents, for example, EDTA (ethylenediaminetetraacetic acid and 1,10-
phenanthroline are prone to inhibit metallopeptidases.
Serine Peptidases [13] These form the most studied class of peptidases. They
possess a reactive serine residue, that is, the hydrolysis of a peptide substrate involves
an acyl-enzyme intermediate in which the active site hydroxyl group of Ser196 (from
the chymotrypsin numbering system) is acylated by the acyl moiety of the substrate,
releasing the amine fragment of the substrate as the first product. The formation of
the acyl-enzyme complex is the rate-determining step in peptide bond hydrolysis, but
the acyl-enzyme intermediate often accumulates in the hydrolysis of ester substrates.
The acyl-enzyme complex thus formed will be the same for a series of substrates that
differ in their leaving group.
The catalytic mechanism of serine peptidases will be given in terms of chymo-
trypsin (Scheme 17.2). After chymotrypsin has bound the substrate to form the
Michaelis complex, nucleophilic attack of Ser196 on the peptide bond of the substrate
forms a high energy tetrahedral intermediate. At the same time the proton of the
serine hydroxyl function is transferred to the nearby His57, the serine hydroxyl
Scheme 17.2 Catalytic mechanism of serine proteases (chymotrypsin numbering).

682
moiety forms a covalent bond with the carbonyl carbon atom of the peptide bond to be
cleaved. The liberated proton is taken by the imidazole ring of His57, thereby forming
an imidazolium ion (general base catalysis). This process is supported by the
polarizing effect of the unsolvated carboxylate anion of Asp102, which is hydrogen
bonded to His57 in the sense of electrostatic catalysis. Mutagenic replacement of
Asp104 by Asn in trypsin, for example, did not change the KM substantially at neutral
pH. On the other hand, kcat was reduced to <0.05% of its wild-type value. Further-
more, neutron diffraction studies have shown that Asp104 in trypsin remains as a
carboxylate anion rather than that it abstracting a proton from the imidazolium ion of
His57 to form an uncharged carboxylic moiety. The active site of serine peptidases is
complementary in structure to the transition state of the reaction, a structure that is
very close to the tetrahedral adduct of Ser196 and the carbonyl carbon of the peptide
substrate. Indeed, transition state binding catalysis provides the catalytic power of the
appropriate serine peptidase.
During the formation of the tetrahedral intermediate a conformational distortion
causes the carbonyl oxygen of the scissile peptide bond to move deeper into the active
site to occupy the oxyanion hole. The resulting oxyanion is hydrogen-bonded to the
backbone NH groups of Gly193 and Ser196, whereas the NH group of the peptide
bond preceding the scissile bond forms a hydrogen bond to the backbone carbonyl of
Gly193. The decomposition of the tetrahedral intermediate forming the acyl-enzyme
complex and the amine product proceeds under the driving force of proton donation
from the N3-atom of His57 through general acid catalysis. The N-terminal part of the
cleaved peptide chain (amine product) will be released in the next step and replaced by
a water molecule forming a second tetrahedral intermediate. The latter decomposes
to the carboxyl product (C-terminal portion of the cleaved peptide chain) and the
active enzyme. Generally, all the serine peptidases employ the same catalytic triad
of Ser, His, and Asp to hydrolyze peptide bonds. The diversity of serine peptidases
results entirely from the way they accommodate their specific substrates.
Cysteine Peptidases [14] Other terms for cysteine peptidases are cysteine-type
peptidases, thiol peptidases, or sulfhydryl peptidases. They are peptidases in which
the attacking nucleophile is the sulfhydryl group of a cysteine residue (Cys25 in the
papain numbering system). The mechanism of catalysis is similar to that of serine
peptidases because a covalent intermediate is formed. Beside the cysteine nucleo-
phile a proton donor/general base is required, which in most cysteine peptidases is a
His residue (His159, in the papain numbering system). Despite the fact that in some
families of cysteine peptidases a third amino acid residue is required to orient the
imidazolium ring of the histidine moiety in the course of the catalytic process, in
general only a catalytic dyad is necessary.
The archetypal cysteine peptidase is papain, which was isolated from the latex of the
tropical papaya fruit (Carica papaya) [15, 16]. It is a single protein of 212 amino acid

residues containing three disulfide bonds and the 3D structure is known with 1.65 A
resolution [17]. The catalytic amino acid residues have been identified as Cys25, His159,
and Asn175, whereas Gln19 helps to stabilize the oxyanion hole. A second category of
cysteine peptidases that is very diverse in sequence is the group of papain-like
endopeptidases of RNA viruses containing only the catalytic dyad Cys/His without any
additional residues being involved in the catalytic mechanism. The same holds for
caspases, a group of ten cytosolic endopeptidases with strict specificity for cleavage of
aspartyl bonds. Members of this family transmit the events leading to apoptosis of
animal cells. Clostripain from the anaerobic bacterium Clostridium histolyticum is a
heterodimericproteinof526aminoacidresidues.Theheavychain(Mr 43 000 Da)and
the light chain (Mr 15 398 Da) are held together by strong noncovalent forces rather
than by disulfide bridges. Cys41 of the heavy chain was identified as the catalytic residue
of the active site. This peptidase is well known for the selective cleavage of arginyl bonds,
whereas lysyl bonds are hydrolyzed at a lower rate. The catalytic mechanism of the
adenovirus endopeptidase is similar to that of papain, the difference being that the four
amino acids His, Glu (or Asp), Gln, and Cys are involved.
Aspartic Peptidases [18] The aspartic peptidases catalyze the hydrolysis of peptide
bonds without the use of nucleophilic attack by a functional group of the enzyme. The
nucleophile attacking the scissile peptide bond in this case is an activated water
molecule and no covalent intermediate will be formed between the enzyme and a
fragment of the substrate. The name of this group of peptidases is based on the
catalytic domain that consists of two aspartic acid side chains (Asp32 and Asp215 of
the porcine pepsin numbering system) activating the water molecule directly. These
two side chain carboxyl groups are close enough to share a hydrogen bond between
two of their oxygens holding the water in place. However, not in all members of the
group of aspartic peptidases are two Asp residues present in the catalytic dyad. An
endopeptidase from nodavirus has an Asp and an Asn as catalytic residues, and in a
related tetravirus endopeptidase one of the Asp residues is replaced by Glu. Inter-
estingly, all the enzymes so far described are endopeptidases.
Metallopeptidases [19] As with the aspartic peptidases, metallopeptidases do not

form covalent intermediates and the nucleophilic attack on the peptide bond to be
cleaved is performed by a water molecule. The latter is activated by a divalent metal
cation, usually Zn2 þ but sometimes also Co2 þ or Mg2 þ . To assist the water molecule
in its nucleophilic attack the metal ion provides a strong electrophilic pull.
The metallopeptidase has a water molecule coordinated to the fourth tetrahedral
site. In thermolysin and carboxypeptidase A, beside the metal ion the other ligands
there are two histidine residues and a glutamic acid residue. The enzymes of this
family can be divided in two groups depending on the number of metal ions involved
in catalysis. In many cases only one zinc ion is required, but often two metal ions act
co-catalytically. All metallopeptidases that contain cobalt or manganese require two
metal ions, but there are also zinc-dependent enzymes in which two zinc ions act in a
co-catalytic manner. Metallopeptidases known to date containing co-catalytical metal
ions are exopeptidases, whereas those with one catalytic metal ion are either
exopeptidases or endopeptidases. His, Glu, Asp, or Lys are known metal ligands
in metallopeptidases. Together with the metal ligand very often a Glu residue
is engaged in the catalytic process. In the leucyl aminopeptidase Lys or Arg fulfills
this function.
684
Table 17.1 Principles of peptidase classification according to the Enzyme Commission (EC) of the
International Union of Biochemistry and Molecular Biology [20].
EC number Type of peptidase Type of cleavage
Exopeptidases
3.4.11.- Aminopeptidase N-terminal residue
3.4.14.- Dipeptidase Dipeptides only
3.4.14.- Dipeptidyl peptidase N-terminal dipeptide
Tripeptidyl peptidase N-terminal tripeptide
3.4.15.- Peptidyl dipeptidase C-terminal dipeptide
3.4.16.- Carboxypeptidase (serine) C-terminal residue
3.4.17.- Carboxypeptidase (metallo) C-terminal residue
3.4.18.- Carboxypeptidase (cysteine) C-terminal residue
3.4.19.- Omega peptidase Terminal modified residue
Endopeptidases
3.4.21.- Serine endopeptidase
3.4.22.- Cysteine endopeptidase
3.4.23.- Aspartic endopeptidase
3.4.24.- Metalloendopeptidase
3.4.99.- Endopeptidase with, unknown mechanism
17.2.1.3 EC Classification
As shown above, based on the chemical moieties that are responsible for their
catalytic activity, peptidases have been classified into four distinct groups. As
recommended by the International Union of Biochemistry and Molecular Biology
(1992) [20] all hydrolases are designated as EC 3., and the peptidases as EC 3.4.,
defining the main classes of peptidases by a third numeral (11–24) as indicated in
Table 17.1. The sub-subclasses are not further divided. Unfortunately, the molecular
structures and evolutionary relationships are not taken into account in the EC
classification. In this EC list the exopeptidases are mainly classified based on their
action. Generally, only peptides with an unprotected terminus are hydrolyzed. The
only exception is so-called omega peptidases which comprise a very small number of
peptidases that are capable of releasing certain modified terminal residues. To this
group belong, for example, acylaminoacyl peptidases that release acetyl or formyl
moieties from the N-terminus, and pyroglutamyl peptidase, capable of releasing the
cyclic residue. An isopeptide bond can be cleaved by b-aspartyl peptidase. Other
omega peptidases are directed to the substituted C-terminus, for example, peptidyl
glycinamidase releasing a C-terminal glycine amide, and c-glutamyl carboxypepti-
dase cleaving a C-terminal glutamic acid linked by an isopeptide bond.
17.2.1.4 Peptidase Families and Clans

Starting with the earlier work of Rawlings and Barrett [21] and improved in the
handbook [4] another level of sophistication to the classification of peptidases has
been developed. Evolutionary considerations can be taken into account due to the
relative ease by which cDNA-derived sequences can currently be obtained. According
to this principle of classification a family of peptidases is defined as a group in which
every member indicates a statistically significant relationship in the amino acid
sequence to at least one other member of the family in the part of the molecule that is
responsible for peptidase activity. Applying strict statistical criteria implies confi-
dence that any two peptidases that are placed in the same family have evolved from a
common ancestor and thus are homologous according to the definition of Reeck
et al. [22]. Each peptidase family is named with a letter that denotes the catalytic type
(S, T, C, A, M, or U, for serine, threonine, cysteine, aspartic acid, metallo-, or
unknown, respectively), followed by an arbitrarily assigned number (Table 17.2).
The term clan is used for defining a group of families the members of which have
Table 17.2 Evolutionary classification of peptidases into families and clans based on primary and
tertiary structure.
Class (EC list) Families Clans (families) Catalytic residues
Serine S1–S44 SA (S1–3,6,7,29–32,35,43) His, Asp, Ser

(EC 3.4.21.)
SB (S8) Asp, His, Ser
SC (S9,10,15,28.33,37) Ser, Asp, His
SE (S11–13) Ser, Lys
SF (S24,26,41,44) Ser, Lys, (His)
SH (S21) His, Ser, His
TA (S42) Thr
SX (14,16,18,19,34,38,39,43)
Cysteine C1–C47 CA(C1,2,10,12,19) Cys, His, Asp (Asn)
(EC 3.4.22.)
CB (C3,4,24,30,37,3S) His, Cys
CC (C6–9,16,21,23, Cys, His
27–29,31–36,41–43)
CD (C14) His, Cys
CE (C5) His, Glu(Asp), Gln, Cys
CX (C11,13,15,22,25,26,39,40)
Aspartic A1–A21 AA (A1–3,9,10–18) Asp, Asp
(EC 3.4.23)
AE (A6, 21) Asp, Asn
Metallo M1–M51 MA (M1,2,4,5,9,13,30,36,48) His, Glu. His (HEXXH)
(EC 3.4.24)
MB (M6–8,10–12) His, His/Asp
(HGXXHXXGXXH/D)
MC (M14) His, Glu, His (HXXE/H)
MD (M15) His, His, Asp
(HMYGHAAD)
ME(M16,44) His, Glu, His (HXXEH)
MF (M17) Lys, Asp3, Glu
(NTDAEGRL)
MG (M24) Asp2, His, Glu2
MH (M18,20,25,28,40,42) His, Asp3, Glu
MX (M3,19,22,23,26,27,29,32,
34–38,41,43,45,47)
686
evolved from a single ancestral protein, but have diverged so far that their relationship
can no longer be proven by comparison of the primary structures. Clan-level
relationships between families can at best be made evident by similarities in 3D
structures. The name of the clan is formed from the letter for the catalytic type (in
analogy to families) followed by an arbitrary second capital letter.
About 40 families of serine- and threonine-type peptidases can be distinguished on
the basis of sequence comparison. However, only a few known families of threonine-
dependent peptidases are included. By comparing the tertiary structures and the
order of the catalytic residues in the sequence most of these families can be grouped
into seven clans (cf. Table 17.2).
The serine peptidases and their clans can be used to demonstrate this type of
classification in more detail. In clan SA with the order of the catalytic triad His, Asp,
Ser the tertiary structure is characterized by a b sheet-based two-domain structure.
Each domain contains a b barrel and between the domains the active site cleft is
located. The largest family S1 of trypsin consists of more than 70 sequenced proteins.
Well-known members of the family S2 are, for instance, streptogrisin A, glutamyl
endopeptidase, and lysyl endopeptidase (from Achromobacter). Togavirin (S3), IgA1-
specific serine-type prolyl endopeptidase (S6), flavivirin (S7), hepatitis C polyprotein
peptidase (S29), helper component proteinase (S30), pestivirus NS2-3/NS3 serine
peptidase, and arterivirus serine endopeptidase (S32) complete the families of clan
SA. The order of the catalytic triad of clan SB is Asp, His, Ser and the tertiary structure
contains both b sheets and a helices. This clan contains only the subtilisin family
(S8), including peptidases from archaea, bacteria, and eukaryotes.
Clan SC contains peptidases with the a/b hydrolase fold bearing the catalytic triad
in the order Ser, Asp, His. This clan includes the families (characteristic member in
parentheses) S9 (prolyl oligopeptidase), S10 (carboxypeptidase C), S15 (Xaa-Pro
dipeptidyl-peptidase), S28 (lysosomal Pro-Xaa carboxypeptidase), S33 (prolyl ami-
nopeptidase), and S37 (Streptomyces PS-10 peptidase). The characteristic catalytic
dyad Ser, Lys of clan SE is represented by the motif Ser-Xaa-Xbb-Lys, and the fold
consists of helices and an a þ b sandwich. The families of this clan, S11 (penicillin-
binding protein 5), S12 (Streptomyces R61 D-Ala-D-Ala carboxypeptidase), and S13
(penicillin-binding protein 4) are involved in the biosynthesis, turnover, and lysis of
bacterial cell walls.
The catalytic residues in clan SF (catalytic dyad Ser, Lys or Ser, His) are more widely
spaced in comparison with clan SE. The families of this clan include only endo-
peptidases from bacteriophages, bacteria, archaea, and eukaryotes with the members
S24 (Lex A repressor), S26 (signal peptidase I), S41 (TSP protease), and S44 (tricorn
protease). All known members of clan SH (catalytic triad: His, Ser, His) are
endopeptidases from DNA viruses that are involved in virus prohead assembly. The
clan includes only the family S21 (Cytomegalovirus assemblin). Clan TA with the
catalytic residue Thr, Ser or Cys, and an a,b,a,b sandwich fold includes several
peptidases whose only proteolytic activity is self-activation. Important families of this
clan are T1 (proteasome) and S42 (c-glutamyl transpeptidase).
Other families (clan SX) of serine peptidases, including S14 (endopeptidase Clp),
S16 (endopeptidase La), S18 (omptin), S19 (cell wall-associated endopeptidase of
Trichophyton), S34 (HflA endopeptidase), S38 (Treponema chymotrypsin-like endo-
peptidase), S39 (cocksfoot mottle virus endopeptidase), and S43 (porin), cannot yet be
assigned to clans, since neither the tertiary structure nor the order of catalytic
residues is known.
The cysteine peptidases consist of the clans CA, CB, CC, CD, CE, and CX. The last
includes several other families of cysteine peptidases for which tertiary structures are
unknown and virtually nothing is known about the specificity of the catalytic
machinery.
The clan CA contains papain and its relatives. Papain was the first well-studied
cysteine peptidase. From the crystal structure of papain and a few closely related
peptidases of the family C1, it could be concluded that the catalytic residues are Cys,
His, and Asn. Further members of C1 are the cathepsine B, H, K, L, and O, the
dipeptidyl peptidase I, and glycyl endopeptidase. The C2 family contains various
calpains, whereas streptopain belongs to C10, ubiquitin C-terminal hydrolase PGP 9,
5 to C12, and the isopeptidase T to C19.
Clan CB contains viral chymotrypsin-like cysteine peptidases that process the
viral polyproteins, and in clan CC are listed viral papain-like endopeptidases. The
only family of clan CD (C14) consists of several cytosolic endopeptidases that cleave
aspartyl bonds with high specificity. This family of caspases consists of ten members,
of which caspase-1 and caspase-3 are best known. The mature caspase-1, processed
from a single-chain precursor presumably by autocatalytic cleavage of four aspartyl
bonds, is a heterodimer of a 22 kDa heavy chain and a 10 kDa light chain [23]. This
peptidase was formerly known as interleukin 1b-converting enzyme (ICE) since it
mediates, among other things, the processing of interleukin 1b at aspartyl bonds.
Human caspase-3 is also a heterodimer consisting of the subunit p12 (11 896 Da) and
the subunit p17 (16 617 Da) with a tertiary and quaternary structure similar to
caspase-1 [24]. This peptidase appears to proteolytically inactivate proteins that are
involved in cellular repair and homeostasis during the effector phase of apoptosis.
Clan CE contains only the adenovirus endopeptidase [25]. A catch-all clan CX
contains all other families of cysteine peptidases that could not be classified up to now
due to the lack of necessary data of structure and catalytic machinery.
For aspartic peptidases, unfortunately, the tertiary structure has only been elucidated
of four families. Endopeptidases of the family A1 consist of two lobes, with the active
site between them. One lobe has been derived from the other by gene duplication. In
the active site each lobe, with very similar 3D structures, bears one Asp residue of the
catalytic dyad. Interestingly, the crystal structure of retropepsin from family A2 of
clan AA showed a single lobe with one catalytic Asp residue with structural similarity
to one lobe of the pepsin from family A1. Retropepsin is only active as a homodimer
forming the catalytic site between the two monomeric molecules. There is evidence
that the peptidases of families A1 and A2 have evolved from a common ancestor.
Unfortunately, several other families could not yet be assigned to any clan.
Metallopeptidases are allocated to eight clans. A couple of families could not yet be
assigned to these clans since, in particular, the metal ligands have not been
biochemically characterized. Zinc-dependent metallopeptidases, both exopeptidases
and endopeptidases, with the HEXXH motif are listed in the clan MA. The family M4
688
contains, along with thermolysin and elastase (Staphylococcus), well-known pepti-

dases. The tertiary structure has been determined for members of this family
showing a two-domain structure with the active site between the domains. The
N-terminal domain contains the HEXXH motif and includes both a helices and b
sheets as dominating structural elements and shows some similarities to the domain
structure of clan MB. The C-terminal domain consists of five helices in a closed
bundle. This characteristic fold is typical for thermolysin-like peptidases. Clan MC
contains metallocarboxypeptidases that belong to only one family (M14), which is
divided into the subfamilies A, B, and C. Typical for this clan is that one zinc ion is
tetrahedrally coordinated by a water molecule, two histidine residues, and one
glutamate residue. Only one catalytic zinc ion is required for peptidases of clans
MA, MB, MC, MD, and ME. Clan MF includes aminopeptidases that require co-
catalytic metal ions for their enzymatic activity (zinc or manganese). The well-known
leucyl aminopeptidase has a two-domain structure bearing the active site in the C-
terminal domain. Whereas exopeptidases of clan MG require co-catalytic ions of
cobalt or manganese, the peptidases of clan MH also require co-catalytic metal ions,
but these are all zinc ions.
17.2.2
Importance of Proteolysis
Historically, enzymatic proteolysis has generally been associated with protein

digestion. Therefore, the digestive peptidases of the pancreatic and gastric secretions
are among the best characterized peptidases and much of the current knowledge of
structure and function has been derived from investigations of those proteolytic
enzymes. Activation of the pancreatic digestive enzymes is initiated by enterokinase,
an enzyme secreted by the mucous membrane of the stomach. It converts some
trypsinogen into active trypsin, which then activates all the proenzymes, including
more trypsinogen. The function of the digestive proteases is merely to break down all
the proteins they encounter.
Later, it became evident that peptidases play regulatory roles in a great variety of
physiological processes [26]. These include processing and molecular assembly of
nascent polypeptide chains, processing of protein hormones, developing enzyme
precursors to mature enzymes, fertilization, and the regulation of programmed cell
death (apoptosis). The last is a mechanism that regulates cell number and is vital
throughout the life of all animals. Apart from various biochemical events involved in
apoptosis, the most fundamental one is the participation of members of the caspase
family in both the initiation and execution phases of cell death. The mechanism of
activation of caspases constituting the different apoptosis-signaling complexes can be
explained by an unusual capability of the caspase zymogen to autopress to an active
enzyme [23].
Proteolytic processing occurs in many different ways and is triggered by different
proteases. Limited proteolysis is the key to this selectivity, which depends on the
accessibility of the scissile peptide bond to the acting peptidase and on its specificity.
In this case proteolysis is directed and limited to the cleavage of specific bonds in the
target protein. A wide variety of prokaryotic and eukaryotic proteins are synthesized as
larger pre- or pre-proforms. Some of these are biologically inactive and become activated
upon limited proteolysis. Lysosomal enzymes, mitochondrial proteins, membrane
proteins, and secreted proteins all undergo intracellular proteolytic maturation.
Various viruses code for specialized peptidases that are essential for virus assem-
bly [27]. A couple of viral peptidases are interesting therapeutic targets. Numerous
publications have been dedicated to the aspartic peptidases, especially to the enzyme of
the human immunodeficiency virus (HIV), which is a key target in the treatment of
AIDS. HIV-1 protease (HIV-1 PR), more exactly named as human immunodeficiency
virus 1 retropepsin (HIV-1 retropepsin; EC 3.4.23.16), has become the most thor-
oughly investigated system in the history of peptidases. The biological function of the
retroviral peptidase is to cleave the polyprotein precursor into its constituent func-
tional units such as the matrix, capsid, and nucleocapsid structural proteins of HIV to
permit assembly. For this reason, the great interest in HIV-1 retropepsin has focused
on the development of compounds that selectively inhibit the viral enzyme and not the
related human aspartic peptidases. Useful principles of inhibition have been com-
bined by several companies to produce antiviral compounds that have achieved
approval from the Food and Drug Administration (FDA) in the USA (cf. Reference [28],
a review). Despite the development of extremely strong and selective inhibitors that
have been demonstrated to be effective in human trials one major problem remains:
the extremely rapid development of forms of the virus that are resistant to the drugs
containing the inhibitors.
Secretory proteins are usually synthesized as precursors bearing an amino-
terminal extension. The signal peptide is removed co-translationally by signal
peptidases during translocation across the membrane. In the next step precursors
of protein hormones, growth factors, and certain polycistronic proteins are processed
by specific enzymes. In contrast to consecutive zymogen activation, consecutive pre-
pro-cleavage reactions are regulated independently. The pathway of processing of
many pre-proteins is known but many of the maturation peptidases cannot yet be
characterized. For this reason, the application of molecular cloning techniques will
be helpful in the near future for the sequence elucidation of pro-proteins as well as the
cDNA and genomic sequences for maturation enzymes. The structure changes range
from the relatively simple alterations in zymogen activation to more complex
processing events in multidomain peptidase precursors, such as prothrombin or
plasminogen. Generally, proteolytic processing induces intramolecular rearrange-
ments required for the expression of biological response. Like blood coagulation, the
complementary system is triggered by a signal that activates several consecutive
zymogen activation reactions. This latter system takes part in the immune reaction
directed against foreign organisms of tissues. Several components of the comple-
mentary system are serine peptidases.
Peptidases as integral components of cells have only been partly explored, for
example, lysosomal peptidases, granulocyte serine peptidases, membrane-bound
peptidases, and enzymes of specialized tissues, such as the reproductive tract, skin,
lens, muscle, pituitary, adrenals, and so on. Various ATP-dependent peptidases have
been isolated.
690
The proteasome is a large multifunctional protease complex that degrades intra-

cellular proteins. The name is derived from protease (protea-) and large particle
(-some) [29]. This complex is an exception among peptidases with regard to the
nucleophilic residue and the general structure. Both in eukaryotes and archaea the
proteasome is a multisubunit complex that consist of four stacked rings each
containing seven subunits (Mr 20–30 kDa). In eukaryotes the 20S proteasome (EC
3.4.99.46; also named multicatalytic proteinase, macropain, and prosome) has the

form of a hollow cylinder (length 148 A, diameter 113 A). It shows several different
catalytic activities and contains 14 different but homologous subunits, whereas in
archaea there are just two different kinds of subunits, and the enzyme complex
possesses only one catalytic activity.
In bacteria the proteasome is built up of two rings of six subunits. One of the two
different subunits is related to the eukaryote and archaea proteasome subunits, the
other is an ATPase. The 26S proteasome [30, 31] (Mr 2100 kDa) consists of the 20S
proteasome and at least one other multisubunit regulatory protein known as PA700,
19S cap, m-particle, ball, and ATPase complex. It was first found in extracts of rabbit
reticulocytes by its capability to degrade ubiquitinated proteins in an ATP-dependent
manner. Since this complex can also degrade various non-ubiquitinated proteins the
older designation ubiquitin-conjugate degrading enzyme (UCDEN) is probably
inappropriate. The 20S proteasome subcomplex of the 26S proteasome containing
multiple catalytic sites with distinct specificities is responsible for the whole
proteolytic activity. In addition, the PA700 regulatory complex displays further
enzymatic activities, such as ATPase activity and isopeptidase activity, and seems
to contain a substrate protein unfolding activity. The ATPase activity is necessary for
assembly of the 26S proteasome from the 20S proteasome and PA700 subcomplexes
and also for the degradation process. Since peptide bond hydrolysis is not energy
dependent, the hydrolysis of ATP might be required for unfolding protein substrates
and/or for translocation of the unfolded peptide substrate into the central channel of
the proteasome. The proteasome is responsible for the turnover of most cellular
proteins in mammalian cells and for the selective degradation of proteins with
abnormal structures. Last but not least, the proteasome is involved in the production
of antigenic peptides for presentation by MHC class I complexes. The generation of
antigenic peptides seems to be performed by a specific subpopulation of proteasomes
containing two or three subunits encoded in the major histocompatibility complex.
Considerable attention has been paid to a group of intracellular serine peptidases
associated with granulocytes as well as leukocytes and mast cells as mentioned above.
These peptidases are stored in granulas and released in response to inflammatory or
allergic stimuli. Many of the peptidases are relevant to human health and disease [32],
some as natural components of the human body and others because they are
important in species that provide us with food, or cause diseases.
To understand proteolytic activity in biological processes, knowledge of the
contribution of the natural peptidase inhibitors to the regulation of the activity is
essential [33]. Inhibitors are as diversified as the proteases themselves. Generally,
they can be divided into two main classes: (i) active site-specific low-molecular-mass
inhibitors and (ii) naturally occurring protein peptidase inhibitors. Examples of the
first group are the serine peptidase inhibitors diisopropyl phosphofluoridate (DFP)
and phenylmethanesulfonyl fluoride (PMSF). Both react with the serine moiety in the
active site. Many of the naturally occurring peptidase inhibitors, isolated from
animal, plant, and bacterial organisms, behave as pseudo-substrates. They combine
essentially irreversibly with the active site upon cleavage of an amide bond, yielding a
stable acyl-enzyme complex.
Of special physiological interest are inhibitors that react with mammalian plasma
serine peptidases, especially those involved in blood coagulation. In principle, such
inhibitors have both protective and regulatory functions. Approximately 10% of the
nearly 200 proteins in blood serum are peptidase inhibitors. The a1-proteinase
inhibitor secreted by the liver, for example, inhibits leukocyte elastase, which is
thought to be part of the inflammatory process. Furthermore, special variants of this
inhibitor with reduced inhibiting potency are associated with pulmonary emphyse-
ma. The latter is a degenerative disease of the lungs that results from the hydrolysis of
its elastic fibers. Interestingly, certain plants release peptidase inhibitors in response
to insect bites to inactivate the digestive enzymes of the attacking insect.
Peptidases are valuable tools in the study of the primary and higher-order structure
of proteins [34]. Proteolysis of proteins for sequence analysis and peptide mapping
can be carried out according to different strategies [35]. Based on the extent of
proteolytic reaction, it is allowed to reach completion or it is prevented from reaching
completion. In the first case the products constitute an equimolar set of peptides
whose composition will not be influenced by further digestion with the same
enzyme. Depending on the restriction imposed by the primary specificity of the
peptidase used, a protein will be fragmented to varying degrees. The fragments can
subsequently be separated and sequenced. Combining these data with sequence data
of other overlapping sets that are generated with different peptidases allows the
reconstruction of the sequence. If proteolysis is prevented from reaching completion
a different set of data is obtained. Inhibition or removal of the peptidase are desirable
interventions to determine the initial cleavage products.
Furthermore, peptidases are also structural probes of the conformation of soluble
proteins [35]. Although X-ray crystallography [36] and 2D NMR [37] are currently the
methods of choice for the determination of the 3D structure of globular proteins,
some weaknesses of these techniques demand alternative methods even if these will
provide structural information at a lower level of resolution. For example, limited
proteolysis can be used to probe the structure and the dynamics of proteins in
solution, providing experimental data that are easy to obtain and complement well the
results derived from the techniques mentioned above. The targets of these investiga-
tions are soluble proteins in their native or near-native states. Limited proteolysis
occurs in this case at the level of only one or very few peptide bonds, which leads to the
formation of nicked proteins. This species of proteins consists of rather large
fragments that remain associated in a stable and often also functional complex.
Usually, a nicked protein is much more labile than the native form. Therefore, the
unfolding leads to a substrate suitable for extensive proteolytic degradation to small
peptides. This further proteolysis is much faster than the initial peptide bond
cleavage at the level of the native protein. Consequently, in this case, during
692
proteolysis the intact protein and small proteins are present in the incubation
mixture, without intermediate sized products. In the case where nicked proteins
are sufficiently stable, they may resist further extensive proteolytic degradation and
can be isolated and characterized.
It is assumed that the limited proteolysis phenomenon derives from the fact that a
specific polypeptide chain segment of the compact, folded protein substrate is
exposed and flexible so that it can fit the active site of the appropriate peptidase for
an efficient and selective limited hydrolysis. There is no doubt that enhanced chain
flexibility or segment mobility is the key feature of the site of peptide bond hydrolysis,
as demonstrated by a clear-cut correlation between sites of proteolytic attack and sites
of enhanced chain flexibility. The present availability of automated, efficient, and
sensitive techniques of protein sequencing and, particularly, the recent dramatic
advances of mass spectrometry [38] in the analysis of peptides and proteins allows a
more systematic use of the limited proteolysis approach as a simple first step in the
elucidation of structure–dynamics–function relationships for novel proteins that are
only available in minute amounts.
Since a growing number of newly discovered peptidases are specifically expressed
in single tissues, especially at low expression levels or often only at certain devel-
opment stages, it is very complicated to isolate the enzymes in sufficient quantities
using classical biochemical procedures. Therefore, the only alternative is the cloning
and expression of these peptidases. In addition, recombinant techniques allow
directed structural alterations to program mechanistic or functional features. Pepti-
dases can be expressed in most of the developed expression systems (yeast, viral,
bacterial, insect cells, and mammalian). It is not usually easy to predict which
expression system is the method of choice. For functional expression of recombinant
peptidases various examples have been presented [37].
Last but not least, it should be mentioned that a couple of peptidases have
industrial importance; in particular, subtilisins, since they have a broad substrate
specificity and are highly stable at neutral and alkaline pH, are of considerable
industrial interest as protein-degrading additives to detergents. These reasons
combined with their large database make subtilisins attractive for protein engineer-
ing. Extensive engineering studies have been carried out on the Bacillus subtilisins
and more than 500 site-directed mutants have been produced to alter specific enzyme
properties, such as pH profile, thermal stability, or substrate specificity [39].
17.3
Synthesis of Peptides
17.3.1
Tools for Peptide Synthesis
Although the origins of peptide chemistry are usually traced back to the early twentieth
century when Emil Fischer obtained the most simple dipeptide glycyl-glycine by
cleavage of the appropriate diketopiperazine, the first peptide bond in a chemical
17.3 Synthesis of Peptides j693
laboratory was synthesized by the young Theodor Curtius in the laboratory of
Hermann Kolbe at Leipzig University in 1881. Despite the fact that Emil Fischer
with coworkers in Berlin made basic contributions to peptide synthesis, the productive
epoch of peptide chemistry began some decades later in the 1950s. Wieland and
Bodanszky [40] have written an excellent account of the history of peptide synthesis.
Peptides are an increasingly important class of bioactive molecules in physiology,
biochemistry, medicinal chemistry, and pharmacology [41, 42] They act as hormones,
neurotransmitters, cytokines, growth factors, and so on. However, it is not only
naturally occurring physiologically relevant peptides that are the subjects of interest.
Peptide analogs possessing agonist or antagonist activity are also useful tools in
investigations to identify suitable drugs. Radiolabeled analogs and molecules bearing
affinity labels have been applied for the characterization and isolation of receptors.
Furthermore, peptides are useful as substrates of peptidases, kinases, phosphatases,
and special transferases in investigations on enzyme kinetics and mechanisms of
action. In the preparation of polyclonal and monoclonal antibodies, peptides play an
important role as synthetic antigen. Epitope mapping using synthetic peptides has
been developed as a valuable approach for the identification of specific antigenic
peptides for the preparation of synthetic vaccines and also for the determination of
protein sequence regions that are important for biological function. In addition, the
design of small peptide mimetics of protein function or structure and the develop-
ment of various peptidomimetics in drug development are further goals in peptide
chemistry. In particular, in the last ten years the number of marketed peptide
pharmaceuticals has significantly increased and the numbers of peptides in clinical
development is growing almost exponentially. Besides the development of efficient
chemicals for peptide synthesis methods, the field of peptide and protein chemistry
has been opened up to molecular biology and genetic engineering.
The classical chemical peptide synthesis is a synthesis [43–47] in a homogeneous
solution, which, in the 1950s, started to gain industrial importance. This was
followed by the solid-phase technique in the early 1960s, invented by the Nobel
laureate Bruce Merrifield [48–51]. The most fundamental time-consuming opera-
tions in chemical solution phase peptide synthesis (sometimes not free from
undesirable side reactions) are the selective protection and deprotection of the
a-amino function, the carboxyl group and the various side chain functionalities of
the trifunctional amino acids. Despite the development of numerous efficient
protection methods based on chemical techniques, the whole process is rather slow
as all intermediate products have to be purified and characterized after each reaction
step. The formation of each peptide bond requires the activation of the carboxylic acid
function of the carboxyl moiety.
An important point in selecting a coupling method is the degree of racemization of
the C-terminal amino acid residue, since chemical activation of the C-terminal
carboxylic acid function has a permanent risk to racemize this amino acid residue.
Therefore, the synthesis of peptides with a multitude of chiral centers continues to be
a formidable chemical effort. The existence of more than 150 chemical variations for
peptide bond formation indicates that an ideal universal coupling method does not
exist, that is, a fast procedure without racemization or other side reactions to realize
694
quantitative coupling of equimolar amounts of the carboxyl and amine components.

There is no doubt that the use of well-known strategies and the application of
activation methods with well established safety steps, to protect against racemization
in simple model systems, does not assure the loss of optical purity during a multitude
of coupling steps in the synthesis of medium-sized and long peptides. For the
chemical peptide synthesis in homogeneous solution, which still plays an important
role in the production of large quantities of peptides for pharmaceutical use, highly
skilled personnel is required. In this manner multi-kilogram quantities or even tons
of peptides consisting of 2–30 amino acid residues can be produced. However,
especially for longer peptides (>5 amino acids), purification of the intermediates
becomes more troublesome and time consuming.
For the synthesis of longer peptides, the time-saving solid-phase peptide synthesis
method [48–51] can be used. The strategy is in principle similar to that in solution,
with the difference that there is no need for isolation of the intermediate products. As
the growing peptide chain is synthesized on a suitable resin the whole procedure
lends itself to automation. The drawback is that every reaction step at the resin has to
be forced to give an almost 100% yield. In practice, this cannot be accomplished, with
the consequence that the desired product must be isolated from a mixture of side-
products by the final, normally HPLC, purification procedure, which is sometimes
difficult to perform and also expensive, especially on large scale. Furthermore, many
peptide sequences, generally longer than ten amino acid residues, tend to form
secondary structures. Owing to this hydrophobic collapse, N-terminal deprotection
and elongation become very troublesome. Therefore, longer peptides of up to 100
amino acid residues are usually synthesized using special dipeptide building blocks,
that is, pseudoprolines, or using longer reaction times and larger excesses of reagents
and amino acid building blocks, which is not cost-efficient in large scale production.
Despite the expenses and difficulties, large-scale stepwise solid-phase peptide
synthesis has been applied up to the multi-ton level [52, 53].
An alternative for the preparation of larger polypeptides and proteins is the
biotechnological production (genetic engineering, recombinant DNA technology)
in bacteria, yeast, or cultured mammalian cells [54–56]. In principle, this is an
economic way to produce peptides of more than 50 amino acid residues and even
small proteins with complicated glycosyl or other groups attached to amino acid side
chains. Compared with the problems connected to the chemical synthesis strategies,
recombinant techniques provide quite a different set of problems. Whereas the
principle of the expression of a gene in host cells through the normal biosynthetic and
genetic machinery of the host cell using a suitable expression vector is relatively
simple, putting this technique into practice poses some problems: for each individual
peptide or protein one has to develop the appropriate expression strategy, the host cell
system as well as the optimal vector system, the control of the stability of mRNA and
also of the translated protein, isolation and purification of the product, scale-up,
downstream processing, and so on.
The development of cloning vectors that propagate in eukaryotic hosts, for
example, yeast or cultured animal cells, has in particular eliminated many of the
problems associated with the synthesis of eukaryotic proteins. Notably, posttrans-
lational processing may also vary among different eukaryotes. It is an advantage that
shuttle vectors are available that are capable of propagating in both yeast and
Escherichia coli and thus transfer genes between these two cell types. Recombinant
protein production is of great medical, agricultural, and industrial importance [57].
Human insulin, human growth factor, erythropoietin, various types of colony-
stimulating factors, and blood clotting factors are typical examples of recombinant
proteins that are in routine clinical use.
Despite the fact that heterologous expression of recombinantly cloned genes is by
far the most commonly employed method of engineering proteins, this approach is
only applicable to naturally occurring amino acids. This limitation is in principle
overcome by unnatural amino acid mutagenesis [57] and some other chemistry-
driven approaches. Among the various chemical ligation methods the so-called
native chemical ligation [58] has proven to be a useful route to fully synthetic
proteins [58–63], although it has not been applied on large scale. As shown in
Scheme 17.3 this procedure relies on the reaction between a peptide fragment
possessing an essential N-terminal cysteine residue (peptide 2; Scheme 17.3), which
can be expressed in principle using recombinant DNA procedures, and a
second peptide fragment possessing an a-thioester group (peptide 1; Scheme 17.3).
In an initial intermolecular, chemoselective reaction a thioester-linked intermediate
is formed (step 1) that spontaneously rearranges via S ! N acyl transfer to
the final amide-linked product (step 2). The rearrangement step corresponds
mechanistically to an intramolecular S ! N acyl transfer reaction described by
Wieland et al. [64] in 1953.
Scheme 17.3 Principle of native chemical ligation according to Dawson et al. [58].
696
Pulling together protein splicing (for a review see Reference [65]) and native
chemical ligation led to expressed protein ligation (EPL) [66], which is also termed
intein-mediated protein ligation (IPL) [67]. As shown in Scheme 17.4 the protein
Clone gene into intein vector
Express in
E. coli
HS
O
N Recomb. protein NH Cys Intein CBD
Affinity
purification Contaminants
HS
O
N Recomb. protein NH Cys Intein CBD Bead
N to S Acyl SLOW
transfer
O
N Recomb. protein S
H2N Cys Intein CBD Bead
Synthetic peptide
Trans thioesterification + thiophenol
(both in large excess)
O
N Recomb. protein S
HS
H2N Cys Synthetic peptide
Native chemical
ligation QUICK
N Recomb. protein CONH Cys Synthetic peptide
Semi-synthetic protein
Scheme 17.4 Principle of expressed protein ligation according to Muir et al. [66].
fragment of interest is expressed in E. coli as an intein-CBD (chitin binding domain)
fusion protein. The CBD allows protein affinity purification using chitin beads. The
required expression vector is commercially available. The N ! S acyl transfer results
in a thioester-linked intermediate. The unfavorable equilibrium is drawn forwards by
the addition of a large excess of a suitable thiol agent (e.g., thiophenol) generating, by
trans-thioesterification in situ, the protein a-thioester, which reacts quickly with the
simultaneously added synthetic amine component bearing an N-terminal cysteine
residue, and the desired semi-synthetic protein is formed by a second N ! S acyl
transfer. Customized peptides containing N-terminal cysteine residues are available
from various sources. Expressed enzymatic ligation [68] combines the advantages of
EPL and the substrate mimetic approach (Section 17.3.6.2) of protease-catalyzed
ligation. In this procedure the requirement of a Cys residue at the ligation site is
lacking. Sortase-mediated ligation [69] is an enzyme-based variant of native protein
ligation. The first protease-catalyzed ligation of cleavage-sensitive fragments in ionic
liquid containing solvents was published by Bordusas group [70].
17.3.2
Identification of the Ideal Enzyme
Enzymes have become valuable tools in medium to large-scale synthetic organic

chemistry [71–74]. Because hydrolases possess a wide substrate spectrum and
usually do not need cofactors for their catalytic function they are at present the
enzymes most widely used as biocatalysts in preparative organic chemistry. Among
the hydrolases the huge family of peptidases plays an important role in various
processes of proteolysis as demonstrated in Section 17.2.
Unfortunately, a universal CN ligase with a high catalytic efficiency for all
possible combinations of the 21 proteinogenic amino acids both as C- and N-terminal
amino acid residues, respectively, in fragments to be coupled could not be developed
during evolution. Such heavy demands on specificity could not even be solved by
nature. Therefore, protein biosynthesis has been developed as a step-wise strategy
starting with the N-terminus of the growing peptide chain and catalyzed by the
ribosomal peptidyltransferase. Limited proteolysis of the biosynthesis precursor
molecules and posttranslational modifications provide the bioactive peptides and
proteins. In nature the peptide bond formation is accomplished on the ribosome and
takes place via the transfer of a peptidyl residue from the peptidyl-tRNA in the
ribosomal P site to the amino group of aminoacyl-tRNA in the A site. Owing to
intensive investigations in recent years, the nature and the basic mechanism of the
peptidyltransferase reaction within the ribosome is partially elucidated [75]. Accord-
ing to studies of the Nobel laureate Thomas R. Cech and coworkers [76] an in vitro
selected ribozyme is capable of catalyzing the same type of peptide bond formation as
a ribosome. Its sequence and secondary structure seem to be strikingly similar to the
helical wheel portion of 23S rRNA implicated in the activity of the ribosomal
peptidyltransferase. These results provide evidence for the feasibility of the RNA
world hypothesis by demonstrating that RNA itself is capable of catalyzing peptide
bond formation.
698
Even in the case in which it would be possible to separate ribozyme activity from
the ribosome or to isolate an in vitro selected ribozyme that can catalyze the same type
of peptide bond formation as a ribosome, such a biocatalyst seems not to be suitable
for simple practical use. This conclusion also holds for the nonribosomal poly- or
multienzymes that are involved in the biosynthesis of peptide antibiotics [77]. Up to
now, they have only found application in the synthesis field of cyclosporine,
gramicidin S, and special b-lactam antibiotics and analogs.
At the end of this short assessment only those enzymes that usually act as
hydrolases catalyzing the cleavage of peptide bonds remain potential candidates for
the practical enzymatic synthesis of peptides. The fundamental suitability of pepti-
dases for catalyzing the formation of peptide bonds is based on the principle of
microscopic reversibility that was predicted by vant Hoff in 1898 [78]. The concept of
vant Hoff of the equilibrium constant of a reversible chemical reaction, along with
the function of a catalyst (including biocatalysts) for accelerated achievement of the
equilibrium according to Ostwald [79], is the theoretical background of enzyme-
catalyzed peptide synthesis. However, about 40 years elapsed before the first
experimental proof of vant Hoffs prediction became evident through the first
clear-cut peptidase-catalyzed synthesis of an amide bond carried out by Bergmann
and Fraenkel-Conrat [80]. Before this approach gained any practical importance
another 40 years elapsed, and in recent decades considerable efforts have been made
to find the optimum conditions for peptidase-catalyzed peptide synthesis, as can be
seen in various reviews [81–101].
17.3.3
Principles of Enzymatic Peptide Synthesis
As shown in Scheme 17.5 the equilibrium of a peptidase-catalyzed reaction is

normally situated on the side of the thermodynamically more stable cleavage
Scheme 17.5 Peptidases function in vivo as hydrolases rather than as ligases.

Figure 17.4 Comparison of the equilibrium (a) and the kinetically controlled approach (b) of
peptidase-catalyzed peptide synthesis.
products. In contrast to proteolysis, peptide bond formation is a two-substrate

reaction and requires not only a specificity-dependent insertion of the carboxyl
component into the S-subsites of the active site but also an optimal binding of the
amine component in the S0 region. To shift the equilibrium in favor of product
formation various manipulations may be employed that differ mechanistically. The
approaches to peptidase-catalyzed peptide bond formation are generally divided into
two basic strategies (see below) according to the type of carboxyl component used. In
the equilibrium-controlled approach the carboxyl component bears a free carboxyl
group as shown in Figure 17.4a, whereas in the kinetically controlled approach the
carboxyl component is employed in an activated form, such as an alkyl ester or a more
reactive ester. Both strategies are fundamentally different due to the energy required
for the conversion of the starting components into the peptide products. Before
interpreting the two mechanisms in more detail some general considerations of
reversing proteolysis must be discussed.
17.3.3.1 General Manipulations in Favoring Synthesis

Looking at the equilibrium for the reversal of proteolysis, under normal conditions
the equilibrium is situated on the side of the hydrolysis products. For example, a
synthesis of a dipeptide from its constituent free amino acids is, from the energetic
point of view, a very unfavorable process because of considerable increase in the free
enthalpy involved. Under these circumstances it is not possible to accomplish peptide
bond formation by simple reversal of hydrolysis, even using high concentrations of
the starting amino acid zwitterions. Energetically more favorable is the reaction of an
anion and a cation using an Na-protected amino acid as a carboxyl component and a Ca-
protected amino acid as an amine component, respectively. According to the under-
lying thermodynamic principles, the outcome of peptide synthesis in aqueous
solution depends on (i) the value of the equilibrium constant, (ii) the ionization
constants of the selectively protected starting compounds, and (iii) the initial
concentrations of the ionized and non-ionized forms of the carboxyl and amine
700
component. The thermodynamic parameters only allow statements relating to the

free enthalpy change between the start and the end of the reaction, that is, the
equilibrium of the reaction. Only the velocity with which the equilibrium is reached
depends on the catalytic action of the enzyme used.
According to the law of mass action the product yield is proportional to the starting
component concentration. Using the least expensive starting component in excess,
manipulations described in the following subsections make it possible to transform
the other starting component almost quantitatively into product. Notably, however,
using an excess of one of the starting compounds is highly unfavorable in industrial
processes, especially if peptide fragments are enzymatically fused.
The formation of insoluble products is a useful way of shifting the equilibrium
towards synthesis. The reaction medium must be designed so that both starting
components are soluble in the medium while the peptide product is insoluble.
Under these conditions the product is continuously removed from the reaction
medium by precipitation and sometimes an almost quantitative product yield can be
obtained. A second way of reversing the proteolysis reaction can be performed by
product extraction, a concept quite close to the solubility-controlled process of
precipitation. The reaction is carried out in a biphasic system where the
product is much more soluble in the organic phase and is continuously removed
from the aqueous phase where the starting components and the enzyme are soluble.
In both approaches to product removal the benefits of the appropriate organic
solvent must be taken into consideration, which will be discussed later. In special
cases the formed product can be separated from the equilibrium by molecular traps,
where the desired product is removed by specific complex formation, as demon-
strated, for example, in the course of clostripain-catalyzed fragment condensation of
the ribonuclease (RNase) fragments 1–10 with 11–15 using RNase S(21–124) as a
trap [102]. The major drawback of all methodologies discussed above is that they lack
general applicability, especially if enzymatic peptide fragment condensations are
concerned.
17.3.3.2 Equilibrium-Controlled Synthesis

This equilibrium-controlled or thermodynamic approach represents the direct
reversal of proteolysis. Consequently, all peptidases, independent of their mechan-
isms, can be used. Major drawbacks of this approach are that (i) the yield is not
quantitative despite the fact that the equilibrium can be manipulated by various
techniques, which usually leads to a complicated work-up procedure; (ii) the reaction
rate is usually very low; (iii) large amounts of peptidase are required.
Preceding the conversion, determined by Kcon, is an ionization equilibrium Kion:
þ Kion Kion
RCOO þ H3 N R0 Ð RCOOH þ H2 NR0 Ð RCONHR0 þ H2 O ð17:1Þ
When the water concentration is taken into the equilibrium constant, Eq. (17.2) is
obtained:
1
þ
0 0
Ksyn ¼ Kion Kcon ¼ ½RCONHR ½RCOO ½H3 N R ð17:2Þ
The reaction conditions, especially the pH, determine the constants for a given pair
of reactants. To obtain an equilibrium that is shifted in favor of peptide product
formation the ionization equilibrium must be manipulated. One efficient method is
the addition of water-miscible organic solvents to the aqueous reaction mixture,
thereby decreasing the dielectric constant of the medium, reducing the acidity of the
carboxyl group, and to a lesser extent reducing the basicity of the amino group of the
nucleophilic amine component [103, 104]. The use of biphasic systems (for a review
see Reference [105]), that is, solvent systems consisting of an aqueous phase and a
nonmiscible phase (apolar organic solvents), does not damage the enzyme since it is
localized in the aqueous phase. Under ideal conditions the reactants diffuse from the
organic phase into the aqueous phase and after the peptide bond forming step
the product diffuses back into the organic phase. Only the insufficient solubility of the
reactants in nonpolar organic solvents limits the general applicability of the biphasic
approach, particularly for the condensation of longer segments.
For the direct reversal of catalytic hydrolysis of peptides, discussed in this chapter,
the term equilibrium-controlled approach is preferred. Because of the thermody-
namic control of both equilibria in (17.1) the reversal of proteolysis is often denoted as
a thermodynamic approach. To increase the product yield of this endergonic process
various manipulations are required. In addition to those mentioned above, reverse
micelles [106], anhydrous media containing minimal water concentrations [107,
108], water mimics [109], ionic liquids [110], and reaction conditions promoting
product precipitation as discussed in Section 17.3.3.1 are often employed.
17.3.3.3 Kinetically Controlled Synthesis

In contrast to the equilibrium-controlled approach the peptidase-catalyzed kinetically
controlled peptide synthesis (for a review see Reference [84]) needs much less
enzyme, the reaction time to reach maximal product yield is significantly shorter,
and the product yield depends both on the properties of the enzyme used and
the substrate specificity. Kinetic control means that the product appearing
with the highest rate and disappearing with the lowest rate would accumulate.
Whereas the equilibrium-controlled approach ends with a true equilibrium, in the
kinetic approach the product concentration goes through a maximum before the
slower hydrolysis of the product becomes significant. If the reaction is not stopped
after the acyl donor ester is consumed, the same equilibrium position would be
obtained as with the equilibrium-controlled approach.
Figure 17.4 compares both approaches schematically. The kinetic approach
(Figure 17.4b) requires the use of an acyl donor ester as the carboxyl component
and is limited to peptidases that rapidly form an acyl-enzyme intermediate, for
example, serine and cysteine peptidases. The peptidase acts as a transferase cata-
lyzing the transfer of the acyl moiety to the amino acid- or peptide-derived amine
component. Specifically, the acyl-enzyme reacts, in competition with water, with the
nucleophilic amine component to form the peptide bond. The ratio of formation of
aminolysis and hydrolysis products is of decisive importance for successful prepar-
ative peptide synthesis. Scheme 17.6 shows the kinetics of the peptidase-catalyzed
acyl transfer reaction. First, the acyl-enzyme is formed via the Michaelis–Menten
702
Scheme 17.6 Kinetics of peptidase-catalyzed ester, Ac-OH ¼ hydrolysis product,

acyl-transfer reaction. EH ¼ enzyme, Ac- Ac-N ¼ peptide product; E. . .Ac-
X ¼ acyl donor ester (carboxyl component), X ¼ enzyme–substrate complex (Michaelis
HN ¼ nucleophile (amine component), complex), Ac-E. . .HN ¼ acyl-enzyme
HX ¼ leaving group of the acyl donor nucleophile complex.
complex, which binds the amine component to the acyl-enzyme. The resulting acyl-
enzyme–nucleophile complex can undergo aminolysis as well as hydrolysis. The acyl
transfer efficiency of the peptidase for the corresponding substrates is determined by
the ratio of the aminolysis and hydrolysis product formed, which is also denoted as
selectivity or as synthesis/hydrolysis ratio.
17.3.3.4 Prediction of Synthesis by S0 Subsite Mapping

Serine and cysteine peptidases are not perfect acyltransferases. Therefore, it is useful
to have a method for the prediction of the outcome of a kinetically controlled peptide
synthesis. To obtain a simple efficiency parameter the partition value p [111] was
introduced, analogous with the definition of the Michaelis constant according to
Eq. (17.3), where P2 ¼ Ac-OH, P3 ¼ Ac-N, and N ¼ HN:
nH d½P2 p
¼ ¼ ð17:3Þ
nA d½P3 ½N
The p value corresponds to the nucleophile concentration at which hydrolysis and

aminolysis of the acyl-enzyme proceeds with the same rate. The advantage of p is that
the definition is not based on a particular kinetic scheme. Furthermore, p allows a
rapid estimation of the yield of any acyl transfer reaction. A concentration of the
nucleophilic amine component [N] p is necessary for peptide formation in high
yield. Assuming equilibrium between the acyl-enzyme and the acyl-enzyme-nucle-
ophile complex, Eqs. (17.4) and (17.5) can be derived from Scheme 17.6 for the rates
of hydrolysis and aminolysis of the acyl-enzyme, where E ¼ EH, EA ¼ Ac-E, A ¼ Ac-
X, and EAN ¼ Ac-E. . .HN:
k5
nH ¼ ½EAk3 þ ½EAN ð17:4Þ
KN
k4
nA ¼ ½EAN ð17:5Þ
KN
Equation (17.6) results from combining Eqs. (17.4) and (17.5):
½Nk5 KN k3
p¼ þ ð17:6Þ
k4 k4
It follows from Eq. (17.6) that a linear correlation between the partition value p and the
nucleophile concentration is obtained. The quotient k5/k4 corresponds to the ratio of
hydrolysis and aminolysis of the EAN complex whereas the term kNk3/k4 is a measure
of the nucleophile efficiency.
The partition value p can be determined by different methods [112–114]. In the
presence of a large excess of nucleophile ([N] [A]0) the decrease in the nucleophile
concentration during the reaction course can be ignored. Under these conditions vH/
vA ¼ [P2]/[P3]. The determination of p can be established from the product ratio
obtained by HPLC analysis according to Eq. (17.7):
½P2 ½N
p¼ ð17:7Þ
½P3
In the preparative application of acyl transfer reactions, however, a large excess of the
nucleophile is not desired. For this reason, p is calculated from the integrated rate
equation [114] according to Eq. (17.8):

½P2 k5 k3 ln ½N0 =ð½N0 ½P3 Þ
¼ þ KN ð17:8Þ
½P3 k4 k4 ½P3
A plot of [P2]/[P3] versus ln{[N]0/([N]0 [P3])}/[P3] gives a straight line with the slope
KN(k3/k4) and an intercept with the y axis at k5/k4. Since this method permits the
determination of p under the conditions employed in preparative peptide synthesis it
should be useful for the optimization of the reaction conditions. An understanding of
the molecular interactions between the acyl-enzyme and the attacking nucleophilic
amine component allows an optimization of the acyl transfer efficiency. The
efficiency of the nucleophilic attack of the amine component depends essentially
on an optimal binding within the active site by S0 P0 interactions (Figure 17.5).
Consequently, more information on the specificity of the S0 subsites of serine and
cysteine peptidases is useful, and can be obtained by systematic acyl transfer studies
using libraries of nucleophilic amine components. According to the definition of the
p value, small values of p indicate high S0 subsite specificity for the appropriate amine
component in peptidase-catalyzed acyl transfer reactions.
A couple of different serine peptidases were studied (for a review see Refer-
ence [84]), that is, the cysteine peptidases papain [115] and clostripain [116, 117], and
the prolyl endopeptidase from Flavobacterium meningoseptum [118] and the p values
for various nucleophilic amine components were determined. Apart from clostripain
none of the enzymes under investigation catalyzed acyl transfer to nucleophilic
amine components with P01 ¼ Pro or D-amino acids. The efficiency of chymotrypsin-
catalyzed acyl-transfer decreases in the order of positively charged > aliphatic >
aromatic > negatively charged P01 side chains. The specificity of chymotrypsin for
P01 ¼ Arg and Lys is attributed to electrostatic interactions between these side chain
704
Figure 17.5 Schematic representation of subsite–substrate interactions in the course of the

acyl transfer from the acyl-enzyme to the nucleophilic amine component catalyzed by a serine
peptidase.
moieties and Asp35 and Asp36 in the active site. A statistical analysis of proteolysis
data confirmed that chymotrypsin possesses specificity for peptide bonds
bearing Arg or Lys at the P01 position, whereas Leu-Asp bonds of proteins were
cleaved by this enzyme considerably less frequently than one expects from the
frequency of occurrence of this peptide bond [119]. These results confirm this
statistical evaluation exactly. Remarkably, chymotrypsin prefers arginine residues at
the P01 and P03 positions, which offers an interesting option for using chymotrypsin as
a restriction peptidase for peptide-catalyzed processing of recombinant proteins
(cf. Figure 17.14 below).
The selectivity of the S0 subsites of different peptidases is reflected by the
broad range of data obtained as shown for simple nucleophilic amino acid
amides in Table 17.3. The values demonstrate the preference for basic and hydro-
phobic P01 residues for chymotrypsin and also for papain. In the case of chymotrypsin
the strongly basic side chain of arginine amide gives rise to a higher efficiency
than all other nucleophiles. Despite the difficulties in catalyzing Xaa-Pro bonds, we
have studied the clostripain-catalyzed acyl-transfer using a large number of proline-
containing peptides as well as Ala-Xaa dipeptides and amino acid amides [116, 117].
The efficiency of clostripain-catalyzed acyl-transfer, using Bz-Arg-OEt as the acyl
donor, to amino acid amides decreases in the order Leu > Lys > Gly > Arg > Gln >
Ser > Pro > Thr > Ala > Asn > Asp > Glu. S0 subsite mapping using an Ala-
Xaa library led to the result that clostripain prefers P02 residues with positively
charged side chains, followed by proline, whereas negatively charged side chains
of Asp and Glu are weak nucleophilic acceptors. In a pentapeptide series,
containing only one proline residue, the efficiency decreases in the order
Pro-P03 > Pro-P02 > Pro-P01 .
Table 17.3 Comparison of p values of selected amino acid amides H-Xaa-NH2 in acyl transfer
reactions catalyzed by various serine and cysteine peptidases according to Schellenberger and
Jakubke [84].
Enzyme p (mM)
Xaa Arg Leu Val Met
Endoproteinase Glu-C V8 >500 16 117 64

Endoproteinase Glu/Asp-C 30 132 Not determined 382
Chymotrypsin 0.11 4.2 6.7 3.3
Trypsin 66 72 130 12
Elastase 16 62 69 34
Papain 1.3 0.41 3.9 1.5
17.3.3.5 What Approach should be Preferred?

As mentioned above, the equilibrium-controlled approach has the advantage that all
peptidases can be used. However, the generally low product yield, the large amount of
enzyme required, and the low reaction rate are serious drawbacks. Owing to the
endergonic process the reaction conditions must be manipulated to increase
the product yield, which should be repeated for every individual peptide coupling.
The use of high concentrations of water-miscible organic solvents to decrease the pK
value of the carboxyl component very often decreases the catalytic activity of the
peptidases. Furthermore, by carrying out equilibrium-controlled synthesis in aque-
ous media using fragments with unprotected side chain functions, the specificity-
determining amino acid residue should not be present on other positions within the
peptide fragments.
In the kinetic approach, the serine or cysteine peptidase rapidly reacts with a
suitable acyl donor ester to form the acyl-enzyme intermediate, which is deacylated
competitively by the nucleophilic amine component and water. The ratio between
aminolysis and hydrolysis of the acyl donor ester is of great importance for the
outcome of the peptide coupling. This selectivity is essentially determined by the S0
subsite specificity of the enzyme as shown above. To establish an optimal synthetic
strategy, it is useful to know the basic kinetic parameters for the reaction course.
Depending on the specificity of the peptidase used, pH and solvent conditions, the
peptide product formed in the kinetic approach is quite stable since the amidase
activity of most peptidases is lower than the esterase activity. In addition, the esterase
activity can be positively manipulated by varying the type of leaving group, as shown
later. For preparative peptide synthesis such a manipulation is very important as it
may allow complete conversion of the acyl donor ester before product hydrolysis
becomes significant. There is no doubt that the course of kinetically controlled
protease-catalyzed peptide synthesis can be influenced more efficiently and generally
than the equilibrium approach.
Although the kinetic approach is preferable, the decision must depend on the
overriding total synthesis concept. The largest industrial-scale application of
706
the equilibrium approach is probably the enzymatic synthesis of Z-Asp-Phe-OMe,

the precursor of the peptide sweetener aspartame [120]. The best known use of
transpeptidation technology is the large-scale conversion of porcine insulin into
human insulin by trypsin [121] or Achromobacter lyticus protease [122].
17.3.4
Manipulations to Suppress Competitive Reactions
The most important factors that limit the widespread routine application of pepti-
dases in kinetically controlled peptide synthesis are the undesired hydrolysis of the
acyl donor ester and the proteolysis of both the starting fragments to be coupled and
the final peptide product (Scheme 17.7). An elimination or minimization of these
undesired reactions can be performed by various manipulations concerning the
reaction medium, the enzyme, and the substrate as well as on mechanistic features of
the process. In particular, an efficient leaving group of the acyl donor ester can
provide high reaction rates in combination with a decreasing extent of proteolysis of
the starting fragments and the final product.
Scheme 17.7 General course of the kinetic approach to fragment condensation catalyzed by serine
or cysteine peptidases.
17.3.4.1 Medium Engineering with Organic Solvents

In peptidase-catalyzed peptide synthesis the solubility of the starting components
dramatically influences the course of the synthesis. The spectrum of solvent systems
that can be used ranges from water, the ideal medium, to water/water-miscible
organic solvents, aqueous-organic biphasic systems, and monophasic organic sol-
vents with trace amounts of water necessary for the catalytic activity of the enzyme
(Table 17.4).
Solubilizing Protecting Groups These offer the only alternative way of bypassing the
poor solubility of most amino acid-derived starting components when reactions are
performed in a fully aqueous environment, and synthesis of peptides can only be
performed if one or both reactants bear such a solubility-promoting group. A
Table 17.4 Influence of the reaction medium on peptidase-catalyzed peptide synthesis.
Reaction Advantages Drawbacks Alternatives

medium
Water
Ideal medium for Poor solubility for Use of solubilizing
enzymes partially protected protecting groups
reactants
Optimal ecological Kinetic approach
conditions only for promotion
of hydrolysis
Water/water-
miscible
organic
solvents
Increased reactant Reduced enzyme Use of chemically or
solubility activity genetically modified
enzymes
Promoting equilibrium Difficult product
controlled approach isolation
Water/water- Biphasic
non-miscible systems
organic solvents
Prevention of enzyme Higher enzyme Use of chemically or
activity requirement genetically modified
enzymes
Easy product isolation Limitation of reactant
solubility lowering of
velocity
Monophasic Prevention of Reduced enzyme Use of chemically or

organic hydrolysis activity genetically modified
solvents enzymes
No solubility problems Change of stereo- and
of partially protected regiospecificity
reactants
Adjusting media Higher enzyme
between chemical requirement
and enzymatic
strategies
successful synthesis of kyotorphin (Tyr-Arg) in a continuous large-scale procedure

using highly solubilizing Na-protecting groups was carried out by Fischer et al. [123].
They used maleyl (Mal-, 3-carboxyacryloyl-), a group that increases both the solubility
of the tyrosine ethyl ester and the activity of chymotrypsin. The coupling was
performed with concentrations of Mal-Tyr-OEt of up to 1.5 mol l1 and an equimolar
concentration of H-Arg-OEt. An overall yield of 50% was obtained including
708
protecting group removal, purification by ion exchange chromatography, and final

product isolation by spray drying. Further large-scale procedures using solubilizing
protecting groups were carried out by Fl€ orsheimer et al. [124] and Hermann
et al. [125]. It was also reported that carboxypeptidases can couple N-terminally
unprotected amino acid esters (50 mM) to unprotected amino acids as well as
amino acid derivatives (0.2–1.5 M) in one step at room temperature in aqueous
solution, but the product isolation was very laborious [126]. This synthesis
principle is more generally applicable to other esterolytic endopeptidases or
lipases [127–130]. The chymotrypsin-catalyzed coupling of H-Phe-OMe with nucle-
ophilic amine components in a frozen aqueous state [131] starting from lower acyl
donor/nucleophile ratios should be mentioned as an interesting alternative; enzyme-
catalyzed synthesis in frozen-aqueous systems will be discussed later in more detail
(Section 17.3.4.2).
Water/Water-Miscible Organic Solvent Systems These systems promote the

solubility of N- and C-terminally protected starting compounds and increase the
pK value of the carboxyl component in equilibrium-controlled processes thereby
promoting this synthesis course. However, reduced enzyme activity in the presence
of high portions of (polar) organic solvents and difficulties in product isolation are
sometimes serious drawbacks. Despite these limitations such media with a small
organic solvent content are preferred in chemoenzymatic peptide synthesis. The
application of more stable immobilized enzymes as well as chemically or genetically
engineered enzymes offers advantages in the case of high contents of organic
solvents, as will be discussed below.
Biphasic Aqueous/Organic Systems [132, 133] These have been employed as an

alternative to water/water-miscible organic solvents systems. This approach leads to
preservation of enzyme activity and allows simple product separation, an advantage
that is counteracted by prolonged reaction times where additional mass transport
between the layers is most likely the rate-determining step. The general use of
biphasic systems is mostly limited by the solubility of the starting components in the
nonpolar organic phase. This alternative to the use of water-miscible organic solvents
has been employed with various peptidases and good yields were obtained using no
more than two equivalents of the nucleophilic amine component (for a review see
Reference [105]).
Synthesis in Reversed Micelles [134] This technique is in principle very similar to the
approach discussed above. After adding small amounts of water and a surfactant to a
hydrocarbon, the polar ends of the surfactant form a sphere that contains the water.
Since the lipophilic group of the surfactant is facing outside into the surrounding
hydrocarbon, the reverse structure of a normal micelle is formed. Liposome-assisted
dipeptide synthesis and selective polycondensation of amino acid and peptides shows
an interesting continuation along this line [135, 136].
Monophasic Organic Solvents [137] The ultimate way of preventing undesired

hydrolytic side reactions in the course of peptide synthesis is offered by these
solvents. The coupling reactions can be driven to completion without hydrolysis of
the peptide products or starting compounds, resulting not only in higher yields but
also in easier purification. Often, trace amounts of water, between approximately 0.1
to about 1 vol.%, are necessary to maintain the catalytic activity of the enzyme.
Although it has been generally assumed that higher concentrations of water-miscible
organic solvents significantly reduce the catalytic activity of the peptidases, few
papers have demonstrated successful enzymatic peptide synthesis performed in
some hydrophilic organic solvents, such as aliphatic alcohols, dimethylformamide,
and acetonitrile [138–142]. Recently, the enzymatic synthesis of peptides in super-
critical CO2 and in ionic liquids has also been described [143–146]. Generally, enzyme
specificities change dramatically in organic solvents. Higher enzyme requirement
and reduced rates should be noted. Interestingly, peptidases also catalyze esterifi-
cation and transesterification reactions in organic solvents when the appropriate
alcohol is added [10].
Chemically or Genetically Modified Peptidases These provide a useful alternative

for peptide synthesis using high concentrations of organic solvents since they
are more stable than the native enzymes. Various possibilities for modification are
known.
Immobilized Enzymes Such enzymes can be used in a very simple way for enzymatic
peptide synthesis as first reported by Jakubke and coworkers [146–148] in the early
1980s. The effort involved in immobilizing an enzyme is mostly compensated by the
possibility of its repeated use and by easier work-up of the reaction mixture.
Immobilized biocatalysts have almost the same efficiency as the (non-immobilized)
free enzymes. The peptidase is covalently linked or physically adsorbed to an
insoluble gel or resin, or a combination of both. The water content in these systems
plays an important role in modulating the catalytic properties of the immobilized
peptidase. The presence of water molecules on the enzyme is required to retain the
catalytic activity. The measurement and control of the thermodynamic water activity
is necessary to quantify the water effect on enzyme activity and the intrinsic influence
of other variables such as support, solvent, and educts [149, 150]. The advantage of
immobilization has been demonstrated by the synthesis of various biologically active
peptides [150, 151]. Of special technical interest are the continuous synthesis of the
aspartame precursor Z-Asp-Phe-OMe with thermolysin immobilized on Amberlite
XAD-7 in a plug flow type reactor [152] and the conversion of porcine insulin into
human insulin catalyzed by Achromobacter lyticus protease I immobilized on SiO2-
polyglutamic acid [153].
Solvent-Modified Enzymes These enzymes are modified, for example, with poly
(ethylene glycol) (PEG), allowing synthesis in monophasic organic solvents as
described, for example, for chymotrypsin [154, 155], papain [156], thermolysin [157],
and subtilisin [158]. Using PEG-modified enzymes in monophasic organic solvents
undesired proteolytic reactions can be almost completely eliminated. However, due
to the solubility properties the use of hydrophobic organic solvents makes the
710
application to the synthesis of longer peptides very complicated and often impossible.
Another important class of solvent-modified enzymes are insoluble crosslinked
enzyme aggregates or crystals (CLEAs or CLECs, respectively), which can be obtained
using glutaraldehyde [159]. These crosslinked enzyme systems are applicable to
various enzymes and generally are much more stable to, for example, organic
solvents, high temperature, and pH variation. For instance, crosslinked chymotryp-
sin [160] was used in a medium with 60% (v/v) dimethylformamide (DMF) for the
successful synthesis of short peptides, and subtilisin A was applied to the synthesis of
peptides [161–163], peptide C-terminal esters [10], and peptide C-terminal carbox-
amides [164] in anhydrous organic solvents.
Chemically Modified Enzymes [165] Enzymes are often modified with the aim of
reducing the peptidase activity with some of the esterase activity remaining, thus
preventing the hydrolytic cleavage of peptide bonds [166]. Methyl-chymotrypsin
(MeCT) obtained by N-methylation of His57 shows a significant change in the
enzymatic catalysis. MeCT is less active than native chymotrypsin by a factor 104 to
105 but it is virtually without any hydrolytic activity [167]. To compensate for the low
activity the more activated cyanomethyl ester is used instead of the methyl ester.
Subtilisin can also be changed to an acyltransferase via modification of the active site
serine to cysteine (thiol subtilisin with low amidase activity [168]) or seleno subtil-
isin [169]. Successful synthesis of various L,D-dipeptides using [Met(O)192]chymo-
trypsin [170] was carried out as well as the synthesis of Ac-Tyr-OEt from Ac-Tyr-OH
and ethanol catalyzed by hexyl-chymotrypsin in a biphasic system [171].
Genetically Engineered Enzymes Enzyme engineering consists of a range of tech-

niques from deliberate chemical modification as shown above to remodeling a wild-
type enzyme by gene technology. The aim of engineering peptidases to generate
peptide ligases by conversion of serine and cysteine peptidases via site-directed
mutagenesis is to make enzymes more stable (e.g., towards organic solvents) and
favor aminolysis rather than hydrolysis [172]. For instance, using multiple site-
directed mutagenesis, subtilisin can be converted into a mutant that allows kinetically
controlled synthesis to be performed in the presence of high concentrations of DMF.
An ingenious combination of chemical and enzymatic steps accelerated the progress
in peptide and protein synthesis, as was demonstrated with subtiligase, a double
mutant of subtilisin BNP0 . This variant was prepared by protein design and used in a
total synthesis of ribonuclease A (RNase A) [173] by combining solid-phase synthesis
for building the fragments and enzymatic coupling of these fragments to form the
protein (Section 17.3.7.2). The selection for improved subtiligases by phage display
resulted in the identification of two new mutants that increased the activity of
subtiligase [174].
17.3.4.2 Medium Engineering by Reducing Water Content

Competitive reactions in enzymatic peptide synthesis are, as mentioned above,
mainly undesired hydrolysis of the acyl donor ester in the kinetic approach, and
undesired proteolytic side reactions in both the starting components in fragment
Figure 17.6 Extended approaches to medium engineering in enzymatic peptide synthesis [93].
condensation as well as the final product. It can be demonstrated that these hydrolytic
side reactions can be largely, and sometimes even completely, avoided by synthesis in
organic solvents of controlled water activity. The main drawbacks are enzyme
deactivation and changes in specificity caused by organic solvents, hence limiting
the number of enzymes that can be used in organic solvents. Therefore, new
strategies have been developed (Figure 17.6) based on reducing the water concen-
tration without substitution by organic solvents (for a review see Reference [88]).
Enzymatic Peptide Synthesis in Frozen Aqueous Systems This technology is based on

observations by Grant and Alburn [175] that trypsin-catalyzed hydroxylaminolysis of
amino acid esters was favored over hydrolysis in frozen aqueous reaction mixtures
(for a review see Reference [176]). In 1990 Schuster et al. [177] first reported on the
influence of freezing on peptidase-catalyzed kinetically controlled peptide synthesis.
The peptidase was added to the reactants precooled to 0 C in a polypropylene tube
and immediately inserted into liquid nitrogen. After 20 s the tube was transferred
into a cryostat at 15 C. Amino components that are considered to be inefficient
nucleophiles in enzymatic synthesis at room temperature gave substantially higher
yields in frozen reaction mixtures. Later these results could be explained on the basis
of the so-called freeze-concentration model [178] and were confirmed by other
investigators [179].
712
Table 17.5 Comparative model peptide synthesis catalyzed by chymotrypsin in frozen aqueous
systems and at room temperature.
Acyl donor Amino component Peptide ice Yield (%) Reference

(% yield) 25 C
Mal-Tyr-OMe H-b-Ala-Gly-OH 79 <2 [165]

Mal-Tyr-OMe H-D-Leu-NH2 73 10 [166]
Mal-Phe-OMe H-Lys-OH 60 0 [168]
H-Tyr-OEt H-Lys-OH 71 0 [169]
H-Phe-OMe H-Leu-NH2 94 52 [170]
H-4-fluoro-PheOMe H-Leu-NH2 90 47
H-2-naphthyl-Ala-OMe H-Leu-NH2 93 55
H-Leu-Phe-OMe H-Ala-Ala-OH 88 5 [171]
H-Asp-Phe-OMe H-Ala-Ile-OH 91 23
H-Gly-Phe-OMe H-Ala-Ile-OH 85 23
In frozen aqueous systems the endopeptidase chymotrypsin is capable of acting

as a reverse carboxypeptidase, catalyzing coupling of free amino acids as the amino
component [180]. Various amino acids were acylated under catalysis of chymotrypsin
starting from 2 mM Mal-Phe-OMe and 50 mM (50% as free base) of the appropriate
amino acids at 25 C in unexpectedly high yields (% given in parentheses): Met (75),
Val (58), Ser (52), Ile (35), Thr (30), Asn (29), Leu (26), and Lys (60). Tougu et al. [181]
described similar results on the coupling of Mal-Tyr-OEt with free amino acids.
Table 17.5 demonstrates the surprising catalytic behavior of chymotrypsin under
frozen state conditions. Na-unprotected amino acid esters as well as dipeptide esters,
even containing non-proteinogenic amino acids, can be coupled in frozen aqueous
systems in high yields, indicating both reverse aminopeptidase and dipeptidyl
peptidase activities. Furthermore, cysteine proteases, with the exception of clostri-
pain, were capable of catalyzing peptide synthesis in frozen reaction mixtures in high
yields using amine components with low efficiency at room temperature. The
specific properties of peptide synthesis in frozen solutions such as changes in
specificity observed in serine and cysteine peptidase-catalyzed reactions strongly
suggest that factors other than concentration of the reactants are probably involved in
yield enhancement by freezing. This assumption is supported by investigations
reported by Jakubke et al. [182, 183] who determined the amount of unfrozen water in
frozen samples at 15 C using the 1 H NMR-relaxation time techniques and
obtained an apparent concentration factor of 50. Synthesis experiments carried out
under these concentration conditions at room temperature gave substantially lower
yields than with frozen reaction mixtures and, therefore, could not simulate the
reaction conditions in ice.
In addition to the freeze-concentration effect, a catalytic role for ice crystals,
a favorable orientation of substrate and biocatalyst, the markedly lower dielectric
constant of ice compared with water, and the high proton mobility in ice have been
considered as factors that possibly influence reactions in frozen systems [176].
Figure 17.7 General principle of application of equilibrium shift towards the product by solid-
phase substrate pools (b) compared with synthesis starting from solution (a) [184].
Peptidase-Catalyzed Synthesis in Solvent-Free Micro-aqueous Systems Such systems

show a second route to reducing water concentration without substitution by organic
solvents. This technology allows the application of reaction systems with partly
undissolved reactants and is based on an extensive theoretical treatment of the
equilibrium position described by Halling et al. [184]. The principle of a solid-to-solid
conversion is illustrated graphically in Figure 17.7 and selected examples of its
experimental implementation are shown in Table 17.6. The application of solid-phase
substrate pools combines the equimolar (or nearly equimolar) supply of starting
components with high obtainable yields, easy work-up procedures, and compatibility
with chemical standard procedures. The key parameter for obtaining high product
yields via acyl-transfer reactions is the ratio of aminolysis and hydrolysis favored by
high nucleophile concentrations. In combination with solid-phase acyl donor pools,
this approach allows an equimolar supply of starting materials without any addition
Table 17.6 Selected examples of peptide synthesis in water-based solid–liquid systems according to
Eichhorn et al. [185] and Jakubke et al. [88].
Carboxyl component Amine Peptide Time (h) Enzyme

component yield (%)
Z-Ala-OH H-Leu-NH2 95 0.5 Thermolysin

Z-Asp-OH H-Phe-OMe 90 7 Thermolysin
Z-Gln-OH H-Leu-NH2 94 4 Thermolysin
Z-Phg-OH H-Leu-NH2 89 2 Thermolysin
Z-Ser-OH H-Leu-NH2 89 2.5 Thermolysin
Z-His-OH H-Leu-NH2 95 3 Thermolysin
Z-Phe-OH H-Met-NH2 88 24 Thermolysin
Ac-Tyr-OEt H-Arg-NH2 90 1 Chymotrypsin
Ac-Tyr-OEt H-GIy-Gly-OH 63 2 Chymotrypsin
Z-His-Phe-OBzl H-Arg-Tip-NH2 95 2.5 Chymotrypsin
Z-Ser-OCam H-His-ONb 85 2.5 Papain
Z-Gly-His-ONb H-Lys-NH2 90 1.5 Chymotrypsin
Z-Arg-His-ONb H-Gly-NH2 55 6 Subtilisin
714
of organic solvents. The synthetic potential of systems with partly unsolved reactants
was proven by pilot-scale synthesis of Z-His-Phe-OMe and the low calorie sweetener
precursor of Z-aspartame using the thermodynamic approach [185] and by kinetically
controlled synthesis of enkephalin derivatives [186]. Furthermore, Halling and
coworkers have studied the effect of water and enzyme concentration of thermo-
lysin-catalyzed solid-to-solid peptide synthesis in detail [187] and reviewed the
approach to enzymatic synthesis with mainly undissolved substrates at very high
concentrations [188].
17.3.4.3 Substrate Engineering

In the case where undesired subsequent reactions occur during kinetically controlled
synthesis, the specificity of the peptidase for the acyl donor ester does not lie
sufficiently above its specificity for the peptide product or the starting fragments.
Since the sequence of the starting components cannot be changed, the only practical
alternative, besides using anhydrous reaction conditions, to suppress such compet-
itive reactions is to use a highly specific leaving group such as the acyl donor ester. As
a simple model peptide with a highly sensitive cleavage site for chymotrypsin
Schellenberger et al. [189] used the chromogenic chymotrypsin substrate Mal-
Leu-Phe-pNA (Mal ¼ maleyl; pNA ¼ p-nitro-anilino), which is formed by chymotryp-
sin-catalyzed coupling of Mal-Leu-OY with H-Phe-pNA. Table 17.7 demonstrates that
the leaving group moiety Y is a major influence on the reaction course. When Mal-
Leu-OMe is employed as the carboxyl component, the rate of product cleavage
reaches the rate of its formation after a short time. By using the more specific acyl
donor esters (higher specificity constants kcat/KM) a clear product accumulation is
attained. With Mal-Leu-ONb (Nb ¼ p-nitro-benzyl) or an ester of similar or higher
specificity, starting components containing highly protease-labile cleavage sites can
be coupled even in a homogeneous phase in high yields. According to this finding
Bongers et al. [190] published a two-step enzymatic semi-synthesis of the superpotent
analog of human growth hormone releasing factor [deNH2Tyr1, D-Ala2, Ala15] GFR
(1-29)-NH2 from the amine component [Ala15]GRF(4-29)-NH2 and the carboxyl
component deNH2Tyr-D-Ala-Asp-OY (Y ¼ Et or Nb) catalyzed by V8 protease and
Glu/Asp-specific endopeptidase (GSE) from Bacillus licheniformis, respectively.
Using the Nb leaving group compared with the ethyl moiety results in a higher
yield without the undesired proteolytic side reactions. The state-of-the-art of substrate
Table 17.7 Influence of the specificity constants of acyl donor esters on the yield of the
chymotrypsin-catalyzed synthesis of Mal-Leu-Phe-pNAa) starting from Mal-Leu-OY with varying
leaving groups Y and H-Phe-pNA according to Schellenberger [189].
Leaving group Y KM (mM) kcat (s1) kcat/KM Reaction Yield (%)

(M1 s1) time (min)
Me (methyl) 120 30 5.6 0.4 4.7 10 10 3

Bzl (benzyl) 1.7 0.4 5.4 0.4 3.2 103 5 65
Nb (4-nitrobenyl) 0.38 0.1 5.9 0.4 1.5 104 5 80.5
a) kcat/KM of Mal-Leu-Phe-pNA: 1.2 10 M1 s1 (Mal ¼ maleyl).

engineering is without doubt the substrate mimetic-mediated C–N ligation strategy,
which allows irreversible peptide bond formation and will be presented separately
(Section 17.3.6).
17.3.5
Approaches to Irreversible Formation of Peptide Bond
Despite the development of various options to suppress competitive reactions, as

shown in the preceding section, absolute avoidance of proteolytic cleavage of the
peptide bond formed cannot be guaranteed. The only alternative seems to be the use
of biocatalysts that do not have the catalytic potential to hydrolyze peptide bonds.
17.3.5.1 Use of Nonpeptidases

Nonpeptidases are supposed to possess favorable prerequisites for the formation of
peptide bonds because undesired proteolytic cleavages in the starting components
and the product can be ruled out.
Enzymes involved in protein synthesis also possess potential, for example,
aminoacyl-tRNA-synthetase-aminoadenylate complex [191], the arginyl-tRNA:pro-
tein arginyltransferase [192], and nonribosomal poly- or multienzyme com-
plexes [193], which require ATP or GTP to activate the carboxyl group of an amino
acid, and seem to accept various amino acid nucleophiles for peptide bond formation.
However, the application of these enzyme systems to universal practical peptide bond
formation is limited.
Furthermore, lipases [194–199] containing the catalytic triad typical for serine
peptidases have been used for peptide synthesis as well as pig liver esterase [200, 201].
These enzymes accept both D- and L-amino acid derivatives as weak acyl donors or
nucleophilic acceptors, but when using peptide fragments as the acyl donor the
activities become very low.
Particularly promising from the theoretical point of view seems to be the devel-
opments in catalytic antibodies. It could be established that antibodies raised against
suitable transition state analogs are capable of catalyzing the formation of peptide
bonds [202, 203]. At present the practical importance is low, but the development of
more tailor-made catalytic antibodies for peptide bond formation could change the
situation.
17.3.5.2 Use of Proteolytically Inactive Zymogens

In 1994, it was first established that zymogens, the catalytically inactive precursors of
various peptidases, can be used as biocatalysts for practically irreversible peptide
bond formation [88, 93, 97, 204]. The capability of reacting slowly with site-specific
reagents indicated that such reactions proceed via the formation of an acyl-zymogen
intermediate [205, 206]. Although, the second-order rate of ester hydrolysis is 106–107
times slower than with the appropriate active enzyme, the deacylation rates of
zymogen and active enzyme do not differ significantly. Therefore, it can be concluded
that the conversion of a zymogen into an enzyme should not be the activation of an
inert zymogen but the potentiation of catalytic activity intrinsic to the zymogen.
716
Based on this feature Jakubkes group used the zymogens of the well-studied
serine peptidases trypsin and chymotrypsin, respectively, in peptide synthesis
experiments and, surprisingly, observed catalysis of peptide bond formation by the
zymogens trypsinogen and chymotrypsinogen. In several cases S0 subsite mapping
studies showed significant differences in the deacylation of the acyl enzymes
compared with the corresponding acyl zymogens, based on acyl transfer to various
peptide derivatives. Although the zymogens possess the same catalytic triad, which is
necessary for the formation of the appropriate covalent acyl intermediate, the non-
optimal formed substrate binding cleft prevents proteolysis. In particular, Gly193 is
distorted and is not capable of forming a hydrogen bond to the carbonyl oxygen of the
substrate, which is necessary for the stabilization of the oxyanion hole [207].
However, because of the high flexibility in this region, principal oxyanion stabiliza-
tion takes place, although not in an ideal manner. To confirm true zymogen catalysis it
was essential to prove that the zymogen preparations were not contaminated with
traces of the appropriate active enzyme. Based on the significantly different affinity of
enzyme and zymogen to the basic pancreatic trypsin inhibitor (BPTI) it was possible
to analyze the esterase activity of zymogens, which is an efficiency parameter used in
estimating their peptide bond forming potential. Since the differences in kcat/KM
cover a range of about five orders of magnitude, for general use of zymogen catalysis it
is essential to improve the acylation rate.
The application of zymogens to irreversible fragment condensations was studied
by coupling a synthetic tetrapeptide methyl ester with a recombinant 24-peptide. Two
coupling reactions (Scheme 17.8) were carried out, that is, by dropping the acyl donor
ester 1 into a solution of the amine component 2 and, alternatively, using batch
conditions. This gave the desired product 3 in 60% and 52% yield, respectively. To
avoid any undesired zymogen activation by limited proteolysis, for example, of the
Lys15-Ile16 peptide bond in the case of trypsinogen, it appeared to be useful to
prevent this reaction by chemical means. The guanylation of trypsinogen by 1-guanyl-
3,5-dimethylpyrazole caused a stable zymogen because of the conversion of all lysine
residues into homoarginine (Har), including the crucial Lys15. Peptide synthesis
with the guanylated zymogens led to very surprising results. Dipeptides with a free
carboxyl group were much more effectively accepted as amine component by the
guanylated species [88]. From molecular modeling studies it could be concluded that
there is an interaction between the carboxyl group of the dipeptide and the only lysine
within the active site (Lys61). The conversion of Lys61 into homoarginine increases
the pK of the side chain and therefore the basic character.
Scheme 17.8 (4 þ 24)-Fragment condensation catalyzed by chymotrypsinogen according to

y et al. (cf. Reference [95]).
Cerovsk
17.3.6
Irreversible CN Ligations by Mimicking Enzyme Specificity [108–110][208–210]
The synthetic importance of peptidases as biocatalysts for peptide synthesis is

undisputed due to a couple of advantages over pure chemical coupling methods.
The mild activation and the high regio- and stereospecificity guarantee that absolutely
no racemization occurs and that temporary protection of side-chain functionalities is
not required. On the other hand, there are some serious drawbacks of the classical
peptidase approach that have been discussed above in detail. Most important is the
fact that the formed peptide bond can be cleaved again in the course of the catalytic
process by the same enzyme. Since the specificity is manifested by the side-chain of
the P1 amino acid residue, for example, Arg or Lys in the case of trypsin, an
irreversible peptide bond formation seems not to be possible within the classical
concept of reversal of proteolysis.
In ribosomal peptide bond formation the mechanism is based on an acyl transfer
of the acyl moiety from the peptidyl-tRNA (or fMet-tRNA at the start of the prokaryotic
biosynthesis) located at the P site of the ribosome to the amino group of the
aminoacyl-tRNA in the A site, catalyzed by the side-chain unspecific ribozyme
peptidyltransferase. Learning from nature Jakubke and coworkers envisioned that
mimicking specificity is the only way to make the peptidase-catalyzed peptide bond
forming step irreversible [211–213]. Since, from the mechanistic point of view the
kinetic approach with serine and cysteine peptidases is also an acyl transfer process
the idea arose of transferring the specificity-directing moiety of the P1 amino acid side
chain to the leaving group of the acyl donor ester. In this manner the enzyme should
recognize the acyl donor ester. Hence, after the acylation of the enzyme, the leaving
group with the specificity determinant would be released from the enzyme with the
consequence that the formed peptide bond cannot be cleaved by the enzyme due to
the lack of specificity for recognizing this bond. In 1991 this hypothesis was
confirmed by model peptide synthesis catalyzed by trypsin using various nonspecific
Na-protected amino acid 4-guanidinophenyl esters (OGp) as acyl donors and various
amino acid and peptide derivatives as nucleophilic acyl acceptors [211, 212], and was
later confirmed by further examples [213–220]. At that time this type of acyl donor
ester was named an inverse substrate according to time-dependent irreversible
inhibitors of trypsin and trypsin-like peptidases, such as 4-amidino- and 4-guanidi-
nophenyl esters, which were found to be hydrolyzed by these peptidases virtually
independently of their acyl moieties [213, 221]. Although this fact was first published
in 1973 by Wagner and Horn [221], very little was known about the basic mechanism
of the hydrolysis of these inverse esters. In 1997 an extension of this new approach to
irreversible peptide segment condensation with other peptidases was described and
the term substrate mimetics was introduced by Bordusa et al. [213].
17.3.6.1 Mechanism of Substrate Mimetic Hydrolysis

The most striking structural differences of Na-protected amino acid or peptide C-
terminal 4-guanidinophenyl esters compared with common peptide substrates are
the nonspecific acyl residue and the highly specific leaving group. It was established
718
Table 17.8 Steady-state kinetic parameters for the hydrolysis of Boc-Xaa-OGp by trypsina) according
to Thormann et al. [222].
Ester (Xaa) KM (mM) kcat (s1) kcat/KM (M1 s1)
L-Ala 0.206 32.4 1.6 105

D-Ala 0.161 0.61 3.8 103
Gly 0.087 23.5 2.7 105
L-Leu 0.146 38.8 2.7 105
D-Leu 0.035 0.85 2.5 104
L-Gln 0.239 35.2 1.5 105
D-Gln 0.071 0.68 9.6 103
L-Phe 0.211 66.1 3.1 105
D-Phe 0.249 9.0 3.6 104
L-Glu 0.071 5.5 7.6 104
D-Glu 0.039 0.43 1.1 104
L-Lys 0.107 270 2.5 106
D-Lys 0.314 15.7 5.0 104
a) Conditions: 25 mM Mops, pH 7.6, 100 mM NaCl, 5 mM CaCl2 25 C; errors less than 15%.
by Bordusas group [222] that all 4-guanidinophenyl esters, independently of struc-

ture and chirality of the acyl moiety, are hydrolyzed despite the lack of trypsin-specific
acyl moieties, with the exception of the lysine derivatives (Table 17.8). This behavior is
in contrast to common trypsin substrates. According to the familiar model, con-
ventional trypsin substrates bind with their acyl residue to the S-binding site of the
enzyme, having the leaving group at the S0 -subsite and the scissile bond between
attacked by Ser195.
As shown schematically in Figure 17.8 applying the same binding principles to the
acyl moiety of the substrate mimetics leads to a catalytically unproductive binding.
The acyl residue would bind at the S-subsite of trypsin, but the scissile bond would be
far from the active site and, therefore, cannot be attacked by Ser195. However,
docking calculations show that the specificity-bearing OGp group binds to the S1-
binding pocket such as the side chain of L-arginine of common peptide substrates.
Surprisingly, this holds even for the substrates Boc-L-Arg-OGp and Boc-L-Lys-OGp
despite the presence of the S1 specific arginine and lysine residues, thus indicating a
higher S1-specificity for the 4-guanidinophenyl moiety. Indeed, all L- and D-substrate
mimetics realize an arrangement in such a way that the scissile bond is very close to
the hydroxyl group of the active Ser195. Furthermore, the carbonyl group of the
scissile ester bond of the appropriate substrate mimetic is located at exactly the same
position as the carbonyl group of the scissile peptide bond, between P1-Lys15 and
P01 -Ala16 in the trypsinogen-BPTI complex. This implies a possible attack by trypsin,
which was confirmed by the hydrolysis studies.
How does it work from the mechanistic point of view? Contrary to common trypsin
substrates, the acyl residues of these enzyme–substrate mimetic arrangements bind
to the S0 -subsite of trypsin (Figure 17.9). For this reason, all binding sites beyond S1
are only of minor importance for the substrate mimetics. Furthermore, the acyl
residues of the substrate mimetics do not reflect the specificity of the S-binding site of
cleavage site
Protease
195
Ser
S3 S2 S1 S’1 S’2 S’3
H2N + NH2
C OH
NH
H2N + NH2
C
NH
(CH2 )3
N C C N
H H O H
Figure 17.8 Schematic comparison of the binding of a peptide 4-guanidinophenyl ester and a
common trypsin substrate to the active site of the enzyme according to the conventional binding
model.
the enzyme. Since the direction of the peptide backbone chain is reversed, the S0 -
subsite specificity is also not reflected. Therefore, substrate mimetics show unique
specificity behavior.
The deacylation step, however, requires an unoccupied S0 -subsite since water can
only attack the acyl enzyme from this site without hindrance. Hence, the flipping acyl
moiety acts like a sliding window within the active site, spanning the primed and
unprimed subsite regions. The extended kinetic model requires a rearrangement
step between the two arrangements (E-Ac and Ac-E) of the acyl enzyme, which is
described by the equilibrium constant KR (Figure 17.9). From the experimental data
of Table 17.8 it follows that D-configured substrates exhibit lower kcat values, which
might be related to lower KR values. Exploring the dynamic behavior by molecular
dynamics simulations of Boc-L-Ala-trypsin and Boc-D-Ala-trypsin indicated that the
flip of the D-Ala complex to the S-subsite takes about 1.5 ns – much longer than in the
L-Ala complex (300 ps).
For an experimental study of the S0 -subsite accessibility, S0 mapping studies
(cf. Section 17.3.3.4) are suitable. By their specific S0 -binding capacity, peptide
KS k2 E-Ac KR Ac-E k3 EH
EH E..Ac-X
S3 S2 S1 S'1 S'2 S'3 S3 S2 S1 S'1 S'2 S'3 S3 S2 S1 S'1 S'2 S'3 S3 S2 S1 S'1 S'2 S'3 S3 S2 S1 S'1 S'2 S'3
H2O COOH
Ac-X HX Ac-OH
Figure 17.9 Schematic representation complex; HX, leaving group; E-Ac,

of the new extended kinetic model of acyl enzyme intermediate located in
peptidase-catalyzed hydrolysis of substrate S0 -region; Ac-E, acyl enzyme intermediate
mimetics according to Thormann et al. [222]. located in S-region; KR, rearrangement
EH, free enzyme; Ac-X, substrate (substrate equilibrium constant; Ac-OH, hydrolysis
mimetic); [E. . .Ac-X], Michaelis–Menten product.
720
nucleophiles should be capable of pushing aside the acyl moiety from the S0 region
more efficiently than water. Therefore, the aminolysis of acyl enzymes bearing the
acyl moiety in S0 should proceed at higher rates compared with their hydrolysis.
Indeed, from the mapping studies it follows that the p-values for the deacylation of
Bz-D-Ala-trypsin are dramatically lower than for Bz-L-Ala-trypsin. Consequently, the
experimental data of aminolysis also support this unique catalytic mechanism for
substrate mimetics.
17.3.6.2 Cationic Substrate Mimetics

The Na-protected amino acid 4-guanidinophenyl ester was the first example of
substrate mimetics for Arg-specific peptidases used for irreversible peptide bond
formation [211, 212]. Apart from the guanidino group linked to various aromatic and
aliphatic spacers, the amidino moiety is also suitable as a specificity-determining
residue in the leaving group of cationic substrate mimetics [213–220]. After the basic
studies with trypsin it could also be established that other Arg-specific peptidases
such as thrombin and clostripain are suitable enzymes for peptide synthesis using
cationic substrate mimetics [213]. In particular, clostripain has appeared to be very
useful in substrate mimetic-mediated fragment condensation. As shown in
Scheme 17.9 the (3 þ 5) fragment condensation provided a product yield of over
90% within a few minutes and the formed product remains unchanged after 72 h.
The course of this synthesis clearly proves the irreversibility of this model CN
ligation.
For synthesis planning, clostripain has an additional decisive advantage due to the
extremely low P01 specificity for the N-terminal amino acid residue of the amine
component. Firstly, Bordusa and coworkers [223] demonstrated the capability of the
cysteine peptidase clostripain as a biocatalyst for the synthesis of peptide isosteres.
The authors investigated the application of clostripain to acylating aliphatic acyclic
and cyclic amines varying in chain length and ring size using the trypsin standard acyl
donor ester Bz-Arg-OEt. Furthermore, using a model substrate mimetic, clostripain
was able to catalyze the reaction with non-proteinogenic amino acids and non-amino
acid-derived amines. The results of these investigations indicate that the substrate
mimetic approach may extend beyond peptide synthesis.
Scheme 17.9 Clostripain-catalyzed (3 þ 5) fragment condensation of Boc-Phe-Gly-Gly-OGp and

H-Ala-Phe-Ala-Ala-Gly-OH [213]. Conditions: 50 mM HEPES-buffer, pH 8, 100 mM NaCl, 10 mM
CaCl2, 25 C, [clostripain]: 1.6 mM
Table 17.9 Clostripain-catalyzed coupling of 4-guanidinophenyl esters of 4-phenylbutyric acid (Pbu-
OGp) and benzoyl-b-alanine (Bz-b-Ala-OGp), respectively, with various amino acid amides and
peptides according to G€unther et al. [223].
Acyl donor ester Acyl acceptor Product Yield (%)
Pbu-OGp H-Leu-NH2 Pbu-Leu-NH2 98

Pbu-OGp H-Lys-NH2 Pbu-Lys-NH2 96
Pbu-OGp H-Ala-Pro-OH Pbu-Ala-Pro-OH 93
Pbu-OGp H-AFAAG-OH Pbu-AFAAG-OH 92
Bz-b-Ala-OGp H-Leu-NH2 Bz-b-Aia-Leu-NH2 98
Bz-b-Ala-OGp H-Lys-NH2 Bz-b-Ala-Lys-NH2 93
Bz-b-Ala-OGp H-Ala-Pro-OH Bz-b-Ala-Ala-Pro-OH 91
Bz-b-Ala-OGp H-AFAAG-OH Bz-b-Ala-AFAAG-OH 93
More recently, Bordusas group disclosed a novel enzymatic synthesis of car-

boxylic acid amides using substrate mimetics and clostripain as a biocatalyst [224].
This unexpected peptidase-mediated approach to the coupling of non-proteino-
genic amino acids and non-amino-acid-derived amines with organic esters could
only be realized by the combination of the substrate mimetic strategy with the use of
clostripain that possesses a broad tolerance towards amines. Table 17.9 sum-
marizes selected examples of the clostripain-catalyzed coupling of Bz-b-Ala-OGp
and the 4-guanidinophenyl ester of 4-phenylbutyric acid (Pbu-OGp) with various
amino acid amides and peptides. Furthermore, the broad tolerance of clostripain
toward non-proteinogenic amino acids and even simple amines, such as aliphatic,
aromatic, or substituted amines and diamines as acyl acceptors is demonstrated by
the results of appropriate syntheses compiled in Table 17.10. The substrate mimetic
approach has opened up a new range of synthetic applications beyond peptide
synthesis, offering efficient and selective organic amide bond formation under
mild reaction conditions.
In a recent paper, Bordusas group combined the substrate mimetic approach with
all-D-specific proteases for the synthesis of all L-peptides [225]. The D-specific
proteases can ligate the peptides in an irreversible way because they are not capable
of hydrolyzing the peptide products.
17.3.6.3 Anionic Substrate Mimetics

Owing to the general validity of the concept of substrate mimetics, Wehofsky and
Bordusa [226] have expanded this strategy to anionic leaving groups in the appro-
priate mimetic structures based on the specificity determinants of Glu-specific
endopeptidases. Since the leaving group moiety of a substrate mimetic binds in
place of the specificity-determining amino acid side chain, for the strong Glu-
preferred V8 protease from Staphylococcus aureus a carboxylate function linked to
a suitable spacer was chosen as the ester leaving group. Unfortunately, the so far
unknown 3D structure of this enzyme allows the design of suitable mimetic
722
Table 17.10 Clostripain-catalyzed coupling of non-amino acid-derived carboxyl and amine

componentsa) according to G€ unther and Bordusa [224].
Acyl donor Acyl acceptor Product Yield (%)
Pbu-OGp H2N Pbu-NH 81
Pbu-OGp H2N OH Pbu-NH OH 65
H2N OH Pbu-NH OH
Pbu-OGp 92
OH OH
NH2 NH2
Pbu-OGp 95
H2N Pbu-NH
Pbu-OGp HO Pbu–O n.s.
Bz-OGp H2N Bz-NH 82
Bz-OGp H2N Bz-NH 76
Bz-OGp H2N Bz-NH 56
Bz-OGp OH OH 57
H2N Bz-NH
Bz-OGp H2N OH Bz-NH OH 84
Bz-OGp H2N OH Bz-NH OH 70
H2N OH Bz-NH OH
Bz-OGp 82
OH OH
NH2 NH2
Bz-OGp 94
H2N Bz-NH
Bz-OGp HO Bz–O n.s.b)
a) Conditions: 0.2 M HEPES-buffer (pH 8.0), 0.1 M NaCl, 0.01 M CaCL2, 5% DMF, 25 C, (acyl
donor): 2 mM, (acyl acceptor): 12 mM;
b) n. s., no synthesis.
structures only by empirical structure–function relationship studies. Apart from

other structures, the carboxymethyl thioester moiety in particular was selected as a
potentially suitable leaving group for imitating the Glu residue in P01 position. The
capability of the carboxymethyl thioester group to act as an artificial recognition site
for the V8 protease was initially studied by steady-state hydrolysis kinetic studies.
Indeed, the carboxymethyl thioester moiety was found to mediate specific hydrolysis
of all carboxymethyl thioester substrate mimetics independently of the P01 amino acid
Scheme 17.10 V8 protease-catalyzed (3 þ 10) fragment condensation of Z-Pro-Leu-Gly-Leu-Ala-

Phe-Ala-Lys-Ala-Asp-Ala-Phe-Gly-OH [226]. Conditions: 0.2 M HEPES buffer, pH 8.0, 37 C, [V8
protease]: 4.9 mM.
residue. This also held for Pro and even for D-Ala, which caused only a slight decrease
in specificity compared with the L-enantiomer. Generally, a one to four orders of
magnitude lower specificity compared with the common substrate Z-Glu-SMe (kcat/
KM ¼ 1.12 104 M1 s1) was found.
In contrast to the non-specificity of the V8 protease for the acyl part, the negative
charge of the leaving group is essential for substrate mimicry. Lacking this charge in
Z-Phe-SCam (-S-CH2-CONH2 instead of -S-CH2-COOH) a complete loss of speci-
ficity resulted since no hydrolysis of Z-Phe-SCam could be observed. The utility of
carboxymethyl thioesters for V8 protease-catalyzed peptide synthesis could be
demonstrated both by model acyl transfer reactions using amino acids and dipeptides
as acyl acceptors and by fragment condensation, respectively. Scheme 17.10 demon-
strates a semi-preparative (3 þ 10) model fragment condensation. After 2 h the
enzymatic coupling reaction of 7 with an excess of 8 in HEPES-buffer containing 5%
DMSO was stopped by the addition of diluted trifluoroacetic acid (TFA) and resulted
in a yield of 55%.
In addition to carboxymethyl thioesters, Bordusas group [227] investigated
additional types of thioesters and phenylesters bearing a free carboxyl group, for
example, the carboxyethyl thioester, 2-carboxyphenyl thioester, and 3- and 4-carbox-
yphenyl ester, which also mediate acceptance by V8 protease. Surprisingly, despite
the lower degree of structural similarities, the aromatic part of the leaving group led to
even higher specificity constants than found for the aliphatic counterparts. In
addition, these studies have been expanded to the use of the cost-efficient but equally
Glu-specific endopeptidase from Bacillus licheniformis (BL-GSE), which can easily be
purified from alcalase in good yield.
17.3.6.4 Hydrophobic Substrate Mimetics

In addition to the enzymes mentioned above with a high specificity for positively and
negatively charged P01 amino acid residues, a third important class of enzymes is
represented by peptidases with specificity for aromatic and hydrophobic function-
alities. Well-known representatives of this family are the serine peptidases chymo-
trypsin and subtilisin, which have found application in classical enzymatic peptide
synthesis. Both enzymes primarily prefer bulky hydrophobic and aromatic P01 amino
acid residues. In addition, the S1 binding pocket of subtilisin contains a carboxylic
724
Figure 17.10 Arrangements of Boc-Ala-OGp at the active site of chymotrypsin (a) and trypsin (b),
respectively, according to Bordusa et al. (see, for example, Reference [208]).
acid moiety (Glu156) that causes additional activity towards Arg and Lys [228]. For this
reason, aromatic leaving groups with additional positively charged substitutions, for
example, the 4-guanidinophenyl ester should fit the natural specificity of these
peptidases.
Parallel to an empirical design of specific mimetic structures, the well-known 3D
structures of the two enzymes allow the use of rational approaches such as computer-
assisted protein–ligand docking. Using the latter to predict the function of the 4-
guanidinophenyl ester functionality, Bordusa selected Boc-Ala-OGp as a model
ligand and docked it towards the enzyme [208]. Figure 17.10 shows the arrangement
of the ligand Boc-Ala-OGp in the active site of chymotrypsin in the lowest energy
complex (a) in comparison with that found for trypsin (b) [222]. In analogy to the
natural specificity of chymotrypsin, hydrophobic contacts between the phenyl moiety
of the ester group and the residues Cys191 and Val213 of the enzyme predominate.
Interestingly, the guanidino functionality favors this binding mode by formation of
additional hydrogen bonds with three serine residues that are located at the bottom of
the S1 binding pocket. This specific binding pattern, the orientation of the carbonyl
oxygen to Gly193 (oxyanion hole), the distance between the carbonyl C-atom of the
scissile ester bond and the active Ser195, and the reversed binding of the acyl moiety
fulfill the conditions for the binding and catalytic mechanism of substrate mimetics.
Indeed, acyl 4-guanidinophenyl esters were hydrolyzed by chymotrypsin, and also
peptide bond formation using various 4-guanidinophenyl esters with nonspecific
proteinogenic and non-proteinogenic acyl residues could be successfully performed
(Figure 17.11) [208]. The yields obtained are in the same range as those obtained
using the normal-type acyl donor Bz-Phe-OMe.
100
100 80
60
Yield [%]
80
40
60
Yield [%]
20
40 0
H-Arg-NH
20 H-Met-NH
0 H-Leu-NH
Bz-OGp H-Ala-Ala-NH
Bz-D-Leu-OGp
Bz-D-Ala-OGp
Bz-L-Pro-OGp H-Gly-Leu-NH2
Bz-L-Glu-OGp
Bz-L-Ala-OGp H-Leu-Ala-NH2
Bz-L-Phe-OMe
Figure 17.11 Chymotrypsin-catalyzed peptide synthesis using 4-guanidinophenyl esters of various

non-specific and non-proteinogenic acyl moieties [208].
Furthermore, unsubstituted phenyl esters are also suitable substrate mimetics

for chymotrypsin-catalyzed peptide synthesis, as was established by Bordusas
group and will be demonstrated by sophisticated fragment condensations in
Section 17.3.7.
17.3.6.5 Chemoenzymatic Substrate Mimetic Approach

To synthesize longer polypeptides and proteins the condensation of the initial
fragments is an essential prerequisite. Despite different chemical ligation techni-
ques in the field of protein semi-synthesis (cf. Section 17.3.1) enzymatic CN
ligation seems to be the only way to avoid partial epimerization, which cannot be
completely eliminated in the course of a chemical fragment coupling reaction.
Consequently, the application of the substrate mimetic strategy for the peptidase-
mediated condensation of peptide fragments indisputably needs to be combined
with solid-phase peptide synthesis (SPPS) of the fragments. Since a peptide ester

can be obtained using the oxime resin strategy [229, 230] Cerovsk
y and Bordusa [231]
developed a procedure for the synthesis of peptide fragments in the form of
substrate mimetics esterified as 4-guanidinophenyl-, phenyl-, and mercaptopropio-
nic acid esters. The synthesis protocol involves covalent attachment of the first Na-
Boc-protected amino acid to the oxime resin, protecting free hydroxylic groups by
acetic anhydride, deprotection of the Na-amino group of the attached amino
acid, followed by successive chain elongation according to the well-known SPPS
726
(a) NO2
O N C P
(b)
+ Protease
protecting group individual amino acid specific leaving group
Figure 17.12 General approach to fragment substrate mimetics via the oxime resin strategy (a) and
substrate mimetic-supported peptide fragment condensation (b) catalyzed by specific peptidases

according to Cerovsk
y and Bordusa [231].
methodology. Finally, the peptide fragment substrate mimetic is generated by

aminolysis of the oxime ester linkage between the peptide and resin, with the
appropriate free amino acid substrate mimetic ester, as shown schematically in
Figure 17.12a. After deprotection of the side-chain functions of the amino acid
residues and, if necessary, also those of the ester leaving group, the Na-protected
peptide ester can be coupled with an amine component using the suitable peptidase
(Figure 17.12b). Some examples of model fragment condensations using this
approach with three different peptidases are given in Scheme 17.11. The coupling
reactions were performed on a preparative scale using 1 : 2 ratios of acyl donor ester
to the nucleophilic acyl acceptors (in the case of trypsin 1 : 2.5) resulting in product
yields between 60% and 70%.
17.3.6.6 Highly Activated Acyl Donors

Another way of circumventing the S1 acyl donor site specificity of a peptidase and
thereby broadening the scope of peptide synthesis is to use highly activated acyl donor
esters with respect to the leaving group. It was shown for several peptidase catalyzed
reactions that esters containing strong electron-withdrawing groups are superior to
the regularly used methyl, ethyl or benzyl esters. For instance, particularly good
results have been obtained with the highly activated 2,2,2-trifluoroethyl (Tfe) or 2,2,2-
trichloroethyl (Tce) [232], acetonoxime (Ax) [233], and cyanomethyl (Cm) [234] esters
(Figure 17.13).
The effect of various activated esters (R) on the chymotrypsin catalyzed coupling of
Z-Ala-OR to H-Leu-NH2 was investigated by the group of Yamada [235], as shown in
Table 17.11. Clearly, the yields were drastically increased compared to the regular
unsubstituted alkyl esters when Tfe, Tce, and Cm esters (entries 6, 8, and 15,
respectively) were used. Another remarkable result was the improvement in peptide
yield using the carbamoylmethyl (Cam, entry 20) ester.
Since the Cam-ester is chemically less activated, that is, electron withdrawing, than
for instance the Tfe, Tce, and Cm esters it was argued that this ester may have specific
(a) Boc-Tyr(Bzl)-Pro-Ser(Bzl)-Ala-Leu-O-P + H-Ala-OGp(Z)2
Boc-Tyr(Bzl)-Pro-Ser(Bzl)-Ala-Leu-Ala-OGp(Z)2
10
H2/Pd
Boc-Tyr-Pro-Ser-Ala-Leu-Ala-OGp + H-Met-Ala-Ala-Ala-Gly-OH
11 12
trypsin
Boc-Tyr-Pro-Ser-Ala-Leu-Ala-Met-Ala-Ala-Ala-Gly-OH
13
(b) Boc-Trp-Ile-Ile-Leu-O-P + H-Gly-SCe
Boc-Trp-Ile-Ile-Leu-Gly-SCe + H-Leu-Ala-Ala-Ala-Gly-OH
14 15
V8 protease
Boc-Trp-Ile-Ile-Leu-Gly-Leu-Ala-Ala-Ala-Gly-OH
16
(c) Boc-Leu-Asn-Lys(Z)-Ile-O-P + H-Val-OPh
Boc-Leu-Asn-Lys(Z)-Ile-Val-OPh
17
H2/Pd
Boc-Leu-Asn-Lys-Ile-Val-OPh + H-Arg-Ala-Ala-Ala-Gly-OH
18 19
chymotrypsin
Boc-Leu-Asn-Lys-Ile-Val-Arg-Ala-Ala-Ala-Gly-OH
20
Scheme 17.11 Combination of solid-phase peptide synthesis and substrate mimetic-supported

segment condensations with different peptidases and substrate mimetics according to Cerovsk
y
and Bordusa [231].
hydrogen bonding interactions between its amide group and the enzyme. The Cam-
ester has been successfully employed in the enzymatic synthesis of peptides using
several peptidases, including subtilisin [236], Amano peptidase P and A [237],
a-chymotrypsin, and papain [238, 239].
As for most substrate mimetics, the difficult synthesis of amino acid and peptide C-
terminal highly activated esters is a major drawback. Very recently, however, the mild
and efficient enzymatic synthesis of amino acid and peptide Tfe- and Cam-esters
using Candida antarctica lipase B (Cal-B) was described by Quaedflieg and cow-
orkers [163]. These esters could be used as starting materials for peptide synthesis
728
Figure 17.13 Highly activated acyl donor esters.
catalyzed by the peptidase subtilisin A. Furthermore, when both enzymes were

combined in one pot, the esterification and subsequent peptide coupling could be
performed simultaneously (Scheme 17.12).
Enzymatic synthesis of activated esters and subsequent peptide coupling using
these esters is an elegant solution for the synthesis of small peptides as compared to
chemical methods, which lack selectivity and are fraught with racemization.
Table 17.11 Effect of various highly activated acyl donors on the chymotrypsin-catalyzed coupling of
Z-Ala-OR to H-Leu-NH2 [235].
Entry R Vrela) Peptideb) (%) Z-Ala-OH (%)
1 CH3 1c) 6.7 1.0

2 CH2CH3 0.54 4.2 0.7
3 (CH2)2CH3 0.39 2.5 0.5
4 CH2CH(CH3)2 0.76 10.2 1.0
5 CH2C6H11c) 0.68 4.9 0.7
6 CH2CF3 17.5 82.4 8.6
7 CH2CF2CF3 8.8 63.2 6.8
8 CH2CCl3 14.1 87.8 6.9
9 CH2Ph 4.2 28.6 4.2
10 CH2C6H4NO2-4 18.7 79.3 9.2
11 CH2C6H4CN-4 11.4 72.7 9.6
12 CH2C6H4Cl-4 5.3 33.6 4.0
13 CH2C6H4OMe-4 2.8 19.2 2.2
14 CH2C5H4N 29.1 89.9 10.0
15 CH2CN 29.6 88.3 6.4
16 CH2OCH3 6.7 45.6 5.6
17 CH2COCH3 6.9 43.0 7.0
18 CH2COPh 0.51 7.5 4.1
19 CH2CO2Et 3.0 24.0 4.1
20 CH2CONH2 133 88.4 10.9
21 CH2CONHCH3 110 89.1 10.8
22 CH2CON(CH3)2 4.8 34.4 5.2
a) Ester consumption was monitored for 8 h.

b) Coupling conditions: Z-Ala-OR (0.05 mmol), H-Leu-NH2HCl (0.2 mmol), TEA (0.2 mmol),
immobilized a-chymotrypsin on Celite (150 mg), acetonitrile (2 ml), 0.05 M Tris buffer (pH 7.8,
80 ml), 30 C, 48 h.
c) 2.79 102 mM h1 mg1.
Scheme 17.12 Activated ester synthesis (by Cal-B) and simultaneous peptide coupling (by
subtilisin A-CLEA) in one pot.
Another example of this approach was given by Kuhl and coworkers, who synthesized
amino acid glyceryl esters using papain for the use of trypsin-catalyzed peptide
couplings [240].
17.3.7
Planning and Process Development of Enzymatic Peptide Synthesis
The high enantio- and diastereoselectivity in peptidase-catalyzed peptide synthesis

allows, in contrast to most chemical coupling methods, the formation of peptide
bonds with complete absence of epimerization of the C-terminal amino acid residue
of the carboxyl component. Furthermore, owing to the regiospecificity of peptidases,
tedious protection/deprotection steps are not required in the enzymatic approach.
Using serine and cysteine peptidases a further point needs to be decided, namely,
should the carboxyl component be used as the free carboxylic acid or should an ester
be used to favor acylation of the enzyme? Enzymatic synthesis using peptide esters or
amino acid esters as substrates has the clear advantage of proceeding at a high rate,
thereby demanding a low concentration of enzyme and, furthermore, being
completely independent of the solubility of the starting materials and product.
Although the kinetically controlled synthesis will be preferable in most cases, the
decision should depend on the total synthetic concept. An unfavorable nucleophile
specificity may be better taken care of in an equilibrium-controlled reaction with the
necessary manipulations of conditions. Despite some limitations the equilibrium-
controlled approach has proven worthwhile in the trypsin-catalyzed semi-synthesis of
human insulin as well as industrial aspartame synthesis using thermolysin. To
overcome poor solubility of the starting components the introduction of solubilizing
protecting groups is frequently necessary.
17.3.7.1 Stepwise Chain Elongation

Contrary to chemical synthesis, enzymatic stepwise chain building may start from
either the N-terminus or the C-terminus. In chemical synthesis, stepwise chain
730
elongation from the N-terminus, as performed in ribosomal protein synthesis, is

normally not recommended under preparative conditions, since the effort needed
to avoid the permanent risk of partial epimerization outweigh the potential
gain. Despite these principal limitations, investigations on solid-phase peptide
synthesis in an N-to-C direction, called inverse synthesis, have been performed
using HOBt salts of the amino acid 9-fluorenylmethyl esters [241]. Unfortunately,
the racemization problem could not be excluded. Furthermore, Mitin and Ryad-
nov [242] have described chemical inverse peptide synthesis to exclude deprotec-
tion reactions at every solution synthesis stage. This could be realized using the
high solubility of free amino acids in dimethylformamide containing Ba(ClO4)2, Ca
(ClO4)2, or Ca(NO3)2. An attempt to solve the extensive racemization was made
using copper(II) ions (CuCl2) during activation of the carboxyl group with ethyl
(dimethylaminopropyl)carbodiimide (EDC) as the coupling reagent in the presence
of HOBt [243].
If the stepwise chain elongation cannot be performed under anhydrous condi-
tions, exopeptidases should be the enzymes of choice since once formed internal
peptide bonds of the growing chain can no longer be proteolytically cleaved by this
type of peptidases. Carboxypeptidases exhibit superior properties for stepwise
synthesis, especially carboxypeptidase Y (CPD-Y) [244] or other serine peptidases
of this type. In principle, aminopeptidases can also be used starting from the C-
terminus. However, since under these conditions not only the carboxyl component
but also the amine component has a free a-amino function, product isolation is more
difficult, particularly if one component is used in excess.
A classical example of a kinetically controlled synthesis starting from the N-
terminus using CPD-Y as an enzyme for all coupling steps was described by
Widmer et al. [245] for [Met]enkephalin (Tyr-Gly-Gly-Phe-Met). Bz-Arg-OEt was
coupled with H-Tyr-NH2 at pH 9.6 to give the Bz-dipeptide amide in 85% yield. The
CPD-Y-catalyzed deamidation at pH 9.6 provided Bz-Arg-Tyr-OH in 90% yield.
After chemical esterification with EtOH–HCl, the resulting Bz-Arg-Tyr-OEt was
coupled with H-Gly-OEt at pH 9 to give the protected tripeptide derivative (yield:
60%), followed by successive enzymatic coupling of the other amino acid deriva-
tives in the same manner. Amino acid amides were preferred as the amine
components, since free amino acids (except Met) only give low yields and amino
acid esters give rise to side reactions that are difficult to control. Finally, the
protecting group for the Na-amino function of Tyr, the Bz-Arg moiety, was easily
removed with trypsin. The disadvantage of this synthetic strategy seems to be the
complicated selective removal of the C-terminal amide grouping by means of CDP-
Y. This step followed by chemical esterification of the peptide had to be resorted to
before it was possible to use the intermediate in the next coupling reaction as the
carboxyl component.
A second step-by-step peptide synthesis from the N- to C-terminus was described
by Bordusa et al. [246] for the model tetrapeptide H-Lys-Tyr-Arg-Ser-OH but using the
endopeptidases clostripain and chymotrypsin as biocatalysts (Figure 17.14). The
synthesis could be performed without side chain protection for all trifunctional
Figure 17.14 Synthesis of H-Lys-Tyr-Arg- (a) and (c) clostripain; (b) chymotrypsin;
Ser-OH from N- to C-terminus using (d) catalytic hydrogenation using
clostripain and chymotrypsin, respectively, 10% Pd/C; -OPr, propyl ester; -SBzl,
as biocatalysts according to Bordusa et al. [246]; thiobenzyl ester.
building blocks and the only non-enzymatic reaction was the final catalytic hydro-
genation for cleavage of the terminal blocking groups.
A drawback of fully enzymatic step-by-step peptide synthesis is that enzymatic N-
or C-terminal deprotection is often accompanied by hydrolytic side reactions.
Furthermore, when synthesizing a peptide in the N- to C-terminal direction, often
a chemical esterification is needed after deprotection to obtain the acyl donor for the
next enzymatic peptide elongation step (Scheme 17.13).
Recently, Quaedflieg and coworkers demonstrated that the enzymatic C-terminal
protecting group hydrolysis and subsequent esterification could be performed in
one single step by enzymatic protecting group interconversion using subtilisin
A [161, 162]. It was shown that protected peptide C-terminal tert-butyl esters
Scheme 17.13 Enzymatic peptide synthesis via C-terminal protecting group interconversion.
732
Figure 17.15 Enzymatic synthesis of Z-Phe-Leu-Ala-OH [247] via (a) t Bu-ester interconversion
using subtilisin A in all steps and (b) amide to ester interconversion using subtilisin A in all steps
except in the last step, where peptide amidase from the flavedo of oranges was used.
and peptide C-terminal primary amides could be interconverted into their corre-
sponding methyl, ethyl, or benzyl esters, which in turn could be used as the acyl
donor for enzymatic peptide elongation. Besides the fact that enzymatic deprotec-
tion and activation are performed in one step, another advantage is that the
interconversion is performed under anhydrous reaction conditions and thus
hydrolytic side reactions are completely avoided. Figure 17.15 shows two typical
examples for the synthesis of the thermolysine assay substrate Z-Phe-Leu-Ala-
OH [247] using the t Bu-ester (Figure 17.15a) and amide interconversion strategies
(Figure 17.15b).
The enzymatic esterification, peptide coupling, interconversion, and hydrolysis
steps were all performed with high yields. The two interconversion strategies were
further demonstrated by the enzymatic stepwise synthesis of several biologically
active and functionalized peptides [161, 162].
As a rule, peptidases can only make a meaningful contribution to a synthetic
strategy if the full advantages of the enzymatic reactions can be utilized. It seems
unrealistic to compare a stepwise peptidase-catalyzed assembly of a long peptide
chain on laboratory-scale with the chemical automatic solid-phase technique. On the
other hand, selected di- and tripeptides are cost-efficiently produced enzymatically on
a large scale, even in a continuous process (cf. Section 17.3.4.1) [248–250].
17.3.7.2 Fragment Condensation

This approach has some advantages over the stepwise strategy. First, if small
fragments are combined to make one that is larger its isolation is easier than for
a stepwise synthesis. Second, the fragment condensation approach is convergent and
Figure 17.16 Fully enzymatic synthesis of [Leu]enkephalin tert-butyl ester [243]. PA, penicillin
acylase; CT, chymotrypsin; P, papain; PhAc, phenylacetyl.
the yields based on the individual amino acid building blocks are higher. Chemical
fragment condensation is limited to positions where a Pro or Gly residue is located,
since C-terminal coupling at the other residues would give rise to racemization.
With enzymatic fragment condensation there is much more freedom of choice for
the coupling positions. In principle it is possible to utilize enzymes both for
protection/deprotection procedures as well as for the formation of the peptide bonds.
Figure 17.16 shows the fully enzymatic synthesis of the tert-butyl ester of Leu-
enkephalin [251] using both equilibrium and kinetically controlled coupling steps. To
obtain the unprotected Leu-enkephalin, the C-terminal protecting group must be
cleaved by chemical means.
Although in principle it is possible to perform totally enzymatic synthesis of
peptides, in practice combined chemical and enzymatic steps are preferred. For the
classical chemoenzymatic synthesis of peptides, and even small proteins, the optimal
approach is usually synthesis of fragments using the SPPS methodology for
enzymatic conjunction in an overall convergent strategy. In a given synthesis project,
initially it is necessary to separate the total sequence into segments containing
favorable combinations of amino acids that permit peptidase-catalyzed segment
coupling according to the elucidated S0 -subsite specificity. Since the kinetic para-
meters of the enzymatic synthesis course are often not available, they can be
estimated from the data for similar acyl donors and nucleophilic amine components.
Based on such estimates an optimal synthetic strategy can be established. In
Table 17.12 selected examples of enzymatically synthesized peptides are compiled.
Many efforts have been made to find optimal conditions for peptidase-catalyzed
peptide synthesis, including the development of new reaction conditions and new
biocatalysts. Once the optimal synthesis conditions have been established, kg
734
Table 17.12 Selected examples of enzymatically synthesized peptides.
Peptide/protein Synthesis routea) Reference
Angiotensin II (analog) E (part) [252]

Aspartame E (total) [107, 120, 185]
Calcitonin (salmon) K (part) [253]
Calcitonin (dicarba analogs) K (part) [254]
Caerulein E (part) [248]
Caerulein (analog) E, K (total) [249]
Cholecystokinin-8 E, K (total) [249]
K (part) [250]
Cholecystokinin-8 (analogs) K (part) [255, 256]
Delta sleep inducing peptide (DSIP) E, K (total) [257]
Dynorphin-(1–8) K, E (total) [258]
EGF (3–14, 21–31, 33–42) K (part) [259]
EGF (29–44) K, E (part) [260]
Eledoisin (6–11) E, K (part) [261, 262]
Eledoisin K (part) [263]
[Met/Leu]enkephalin E, K (part) [264]
[Met]enkephalin K (total) [245]
[Leu]enkephalin K (total) [251]
[Leu]enkephalin derivatives K (part) [265]
Growth hormone releasing K (part) [190]
factor (human) analog
Hepatitis B S antigen (122–137) K (part) [266]
Ht31(493–515) peptide K (part) [267]
Insulin (human) E (part) [121, 122, 268]
Kyotorphin K (total) [130, 248, 249]
LH-RH K (part) [263, 269, 270]
[D-Phe6]LH-RH K (part) [271]
MSH (5–8, 9–12, 13–16) K (total) [272]
Oxytocin (1–9) K (part) [273]
Ribonuclease A K (part) [173]
Somatostatin K, E (part) [274]
Substance P (6–11, 7–11) K (part) [261, 275]
Vasopressin (1–6) K, E (part) [276]
a) E, equilibrium approach; K, kinetic approach; total, totally enzymatic coupling; part, partly
enzymatic coupling.
amounts of biologically active peptides can be produced. The enzymatic condensa-

tions can normally be achieved with commercially available enzymes that are easy to
handle, especially when they are immobilized on a solid support. In addition, owing
to the requirement of only catalytic amounts the higher costs of the enzymes used are
usually insignificant in comparison with the costs of the highly sophisticated
stoichiometrically used chemical coupling reagents plus the financial expense of
the reagents necessary for protection/deprotection procedures in chemical synthesis.
Scheme 17.14 Peptidase-catalyzed modification of an artificial substance P precursor protein

according to Schellenberger et al. [278].
As a model system for peptidase-catalyzed modification of peptides produced by

recombinant DNA technology, Schellenberger et al. [277, 278] developed a new
approach to the production of peptides based on ribosomal fermentation and
peptidase-catalyzed processing (Scheme 17.14). First, they produced an artificial
substance P precursor as a b-galactosidase (1–459) fusion protein containing nine
copies of the sequence H-Arg-Leu-Arg-Arg1-Pro-Lys-Pro-Gln-Gln-Phe7-OH. The
sequence of the peptide precursor was designed to meet the specific requirements
of chymotrypsin and papain, used in conversion reactions as the complete amino acid
sequence should be generated by addition of the appropriate dipeptide derivatives.
After isolation and purification of the fusion protein, which was accumulated in E. coli
as inclusion bodies, the dodecapeptide ester H-Arg-Leu-Arg-Arg1-Pro-Lys-Pro-Gln-
Gln-Phe-Phe-Gly9-OMe was formed by chymotrypsin-catalyzed transpeptidation in
the presence of H-Phe-Gly-OMe. In a papain-catalyzed acyl transfer reaction and
subsequent tryptic cleavage, the dodecapeptide ester was converted into substance P.
These results indicate that peptides can be readily produced by a combination of
recombinant DNA technology and peptidase-catalyzed conversion with the advan-
tage of potential incorporation of groups other than proteinogenic amino acids into
the recombinant product.
The chemoenzymatic synthesis of RNase A [173] using a mutant of subtilisin
BNP0 , called subtiligase, underlines the progress of enzyme-catalyzed fragment
condensations in the course of the synthesis of a small protein. The fragments
(98–124, 77–97, 64–76, 52–63, 21–51, and 1–20) were individually synthesized
by standard SPPS methodology. The fragments were chosen in such a way that the
C-terminal residues of the appropriate fragments (Tyr97, Tyr76, Val63, and Ala20)
were the closest to matching the substrate specificity of the subtilisin mutant. Using a
considerable excess of the fragments bearing a Phe-NH2-modified carboxamido
methyl ester ensured that most of the side reactions could be suppressed. Starting
736
with the C-terminal fragment (98–124) the total yield after five fragment condensa-
tions was 15%, and after folding the final protein could be obtained in 8% yield. In a
similar manner three analogs of RNase A were synthesized in which the two residues
His12 and His119 of the active center were exchanged individually and simulta-
neously for L-4-fluorohistidine.
Despite this impressive example of five successful enzyme-catalyzed fragment
condensations with average yields of roughly 75% in the course of the synthesis of
RNase A, all the peptide bond forming steps could not be performed irreversibly.
Even though the new CN ligation strategy based on the substrate mimetic concept
(cf. Section 17.3.6) has not yet been proven for the synthesis of a similar protein
target, it guarantees the irreversibility of the enzymatic coupling reaction, as can be
demonstrated by the chymotrypsin-catalyzed (8 þ 16) fragment condensation of the
Ht 31(493–515) peptide derived from the human protein kinase A anchoring protein
(sequence 493–515) [267]. The 24-peptide sequence was synthesized by chymotryp-
sin-catalyzed fragment condensation at a nonspecific Ser-Arg peptide bond via the
substrate mimetic strategy (Scheme 17.15). The fully protected carboxyl component
21 was synthesized on Kaisers oxime resin and was released from the support by
aminolysis with H-Ser(Bzl)-OPh. After side-chain deprotection by catalytic
Scheme 17.15 Chymotrypsin-catalyzed (8 þ 16) segment synthesis of the Ht 31 (493–515)

peptide by the substrate mimetic strategy [277].
hydrogenation, 22 was coupled with the unprotected amine segment 23 catalyzed by
chymotrypsin, leading to the complete conversion of both peptide segments. Finally,
the N-terminal Boc group was cleaved by TFA to give the desired Ht 31 (493–515)
peptide 25.
17.3.8
Enzymatic Modification of Peptides
As well as in peptide bond formation, the application of enzymes in protection,

deprotection, and (C-terminal) modification reactions of amino acids and peptides
has similar advantages over chemical conversions, such as the lack of racemization or
hydrolytic side reactions and the high specificity. For instance, chemical synthesis of
activated ester acyl donors requires the same toxic coupling reagents as for chemical
peptide bond formation. Another example is the deprotection of tert-butyl groups
using trifluoroacetic acid. Furthermore, biologically active peptides are often deco-
rated at specific positions with, for example, fatty acids, poly(ethylene glycol) chains,
or arylamide groups; their chemical introduction usually requires complicated
protection/deprotection strategies.
The selective enzymatic cleavage of N- and C-terminal protecting groups is one of
the oldest applications in this respect, and the easiest to perform, although one
should be aware of hydrolytic side reactions, that is, hydrolysis of the peptide bonds.
Enzymatic hydrolysis of amino acid and peptide N-terminal protecting groups
includes among others the deprotection of acetyl, formyl, phenylacetyl, phthaloyl,
tert-butoxycarbonyl, and carboxybenzyl groups; for a review see Reference [279].
Enzymatic hydrolysis of amino acid and peptide C-terminal protecting groups [279]
includes the deprotection of methyl, benzyl, allyl, trimethylsilyl, and primary
amide [280] groups. Recently, the enzymatic hydrolysis of peptide C-terminal tert-
butyl esters was disclosed [281] using the serine protease subtilisin A, as shown in
Scheme 17.16. The selective enzymatic hydrolysis of a-protecting groups with
Scheme 17.16 Subtilisin A catalyzed C-terminal modification reactions of amino acids and
peptides; X ¼ alkyl or highly activated ester.
738
subtilisin A has also been utilized to synthesize b-Asp and c-Glu side-chain protected
building blocks [10].
Various N- and C-terminal protecting groups [279] can also be enzymatically
introduced. Recently, the synthesis of peptide C-terminal amides from the corre-
sponding peptide C-terminal methyl esters using subtilisin A was described by
Boeriu and coworkers [164]. This C-terminal modification is important for the
biological activity of many peptides, but also increases their stability against exo-
proteases. Another recent example, using the same enzyme, is the synthesis of amino
acid and peptide C-terminal tert-butyl esters, which are chemically hard to obtain [10].
The reversible synthesis of C-terminal amides and tert-butyl esters was used to
develop two fully enzymatic peptide elongation strategies using protecting group
interconversion as discussed in Section 17.3.7.1 [161, 162]. Furthermore, the
selectivity of subtilisin A was used to obtain a-protected Asp and Glu building
blocks [10].
Additionally, a large number of important enzymatic C-terminal modification
reactions of amino acids and peptides have been described. For instance, the
recently disclosed synthesis of amino acid and peptide C-terminal arylamides by
subtilisin A in organic solvents proceeds smoothly and without racemization
(Scheme 17.16), in contrast to known chemical methods [282]. Peptide C-terminal
arylamides are often used in chromogenic, fluorogenic, or amperogenic enzymatic
assays, for instance to quantify enzymes involved in blood coagulation [283].
Another recent discovery is the enzymatic synthesis of peptide C-terminal thioacids
and thioesters by Liu and coworkers using subtiligase [287, 288]. Even more
recently, the enzymatic C-terminal thioesterification of amino acids and peptides
in high yields and with commercially available enzymes was demonstrated by
Quaedflieg et al. [289]. Peptide C-terminal thioesters are the pivotal building
blocks for protein synthesis using native chemical ligation, as described is
Section 17.3.1.
Furthermore, many enzymes can PEGylate [284], glycosylate [285], or (fatty acid)
acylate [286] peptides on specific sites, giving them the desired biological activity.
Additionally, there are many modifications of peptides that stabilize them against
peptidase activity. One of the most important ones is via head to tail cyclization, which
can be performed mildly by enzymes [290]. Other stabilizing modifications include
methylation, halogenation, and oxidation; for a review see Reference [291].
17.4
Conclusion and Outlook
Despite the fact that chemical methods are popular for the synthesis of peptides, a
huge number of papers have been published in recent decades dealing with
enzymatic formation of peptide bonds, enzymatic manipulation of protecting
groups, and enzymatic modification of peptides. However, at present more peptides
are synthesized by chemical synthesis than by peptidase-catalyzed processes. The use
17.4 Conclusion and Outlook j739
of peptide synthesizers, in addition to recent new developments in the field of
chemical ligation procedures, still favors chemical methods over the enzymatic
approach. However, there is no doubt that enzymatic methods have advantages,
including the prevention of racemization, no need for time-consuming and expen-
sive protection/deprotection procedures of side-chain functions, the reduced use of
problematic (toxic) solvents and stoichiometrically required reagents, and possible
reuse of the biocatalysts. The question should not be whether to use a chemical or an
enzymatic approach; an ingenious combination of chemical and enzymatic steps
should promote the general progress in peptide synthesis.
It could be demonstrated that, after establishing the optimal synthesis conditions,
kg amounts of biologically active peptides and analogs can be obtained using
enzymatic coupling methods. The semi-synthetic synthesis of human insulin
and the production of aspartame on the ton-scale underline the industrial importance
of the enzymatic approach. However, the enzymatic approach does not have the
versatility of chemical synthesis methods and suffers from some limitations. The
main reason seems to be the lack of a universal enzyme that is capable of catalyzing
peptide bond formation for all possible combinations of the 21 proteinogenic
amino acid residues at the C- and N-terminal positions in the peptide fragments
to be coupled. Such an enzyme was not developed during evolution due to the
extremely high specificity requirements. In ribosomal protein synthesis nature
prefers the stepwise synthesis from the N- to C-terminus followed by maturation
procedures based on limited proteolysis and further modifications. The only bio-
catalyst involved in ribosomal synthesis, the peptidyl transferase, seems to be an old
ribozyme without any specificity for the P1 side chain functions of the amino acids,
only catalyzing the acyl transfer reaction of the selected aminoacyl-tRNAs. Since such
a biocatalyst has no practical importance in peptide synthesis in a peptide laboratory,
the only alternative for this purpose is the reverse catalytic hydrolysis potential of
peptidases.
The advantages and drawbacks of peptidases used for catalyzing peptide bond
formation have been demonstrated in this chapter. An ingenious combination of
chemical and enzymatic strategies was demonstrated in a new synthesis of RNase A.
Also, two fully enzymatic N- to C-directed peptide synthesis strategies based on
protecting group interconversion were described wherein all reactions are performed
in anhydrous organic solvents, thereby fully eliminating hydrolytic side reactions.
Furthermore, using the CN ligation strategy based on the substrate mimetic
concept, irreversible peptide bond formation catalyzed by highly specific peptidases
can be performed. In combination with peptidase mutants, which lack amidase
activity, this new CN ligation approach will contribute to significant progress in
enzymatic peptide synthesis, especially in clear-cut fragment condensations using
recombinant polypeptide thioesters as the substrate mimetics with chemically
synthesized or recombinant fragments. This specific programming of enzyme
specificity by molecular mimicry corresponds in practice to a conversion of a
peptidase into a CN ligase, a biocatalyst that could not be developed by nature
during evolution.
740
References
1 Coopland, R.A. (2000) Enzymes, a 13 Hedstrom, L. (2002) Chem. Rev., 102,

Practical Introduction to Structure, 4501–4523.
Mechanism and Data Analysis, 2nd edn, 14 McKerrow, J.H. and James, M.N.G. (eds)
John Wiley & Sons, Inc., New York. (1996) Cysteine Proteases: Evolution,
2 Unger, T. and Sch€olkens, B.A. (2004) Function, And Inhibitor Design, ESCOM
Angiotensin, vol. 1, Springer, Berlin. Science Publishers, Leiden.
3 Kenny, A.J. and Boustead, C.M. (eds) 15 Kimmel, J.R. and Smith, E.L. (1954)
(1997) Cell-Surface Peptidases in Health J. Biol. Chem., 207, 515–531.
and Disease, Bios Scientific Publishers, 16 Shindo, T. and Van der Hoorn, R.A.
Oxford, p. 384. (2008) Mol. Plant Pathol., 9,
4 Barret, A.J., Rawlings, N.D., and 119–125.
Woessner, J.F. (eds) (2004) Handbook 17 Kamphuis, I.G., Kalk, K.H.,
of Proteolytic Enzymes, Elsevier Science. Swarte, M.B.A., and Drenth, J. (1984)
5 Schechter, I. and Berger, A. (1967) J. Mol. Biol., 179, 233–256.
Biochem. Biophys. Res. Commun., 27, 18 Horimoto, Y., Dee, D.R., and Yada, R.Y.
157–162. (2009) New Biotechnol., 25, 319–323.
6 Zhang, B. and Cech, T. (1998) 19 Turner, A.J. (2003) Biochem. Soc. Trans.,
Chem. Biol., 5, 539–553. 31, 723–727.
7 Schwarzer, D., Finking, R., and 20 International Union of Biochemistry and
Marahiel, M.A. (2003) Nat. Prod. Rep., 20, Molecular Biology Nomenclature
275–287. Committee (1992) Enzyme Nomenclature,
8 Kadereit, D. and Waldmann, H. (2001) Academic Press Inc., New York.
Chem. Rev., 101, 3367–3396. 21 Rawlings, N.D. and Barrett, A.J. (1993)
9 (a)Schmidt, M., Barbayianni, E., Biochem. J., 290, 205–218.
Fotakopoulou, I., H€ohne, M., 22 Reeck, G.R., de Haen, C., Teller, D.C.,
Kokotou, V.C., Bornscheuer, U.T., and Doolittle, R.F., and Fitch, W.M. (1987)
Kokotos, George (2005) J. Org. Chem., 70, Cell, 50, 667.
3737–3740; (b) Schmidt, M., 23 Chowdhury, I., Tharakan, B., and
Barbayianni, E., Fotakopoulou, I., Bhat, G.K. (2008) Comp. Biochem. Phys.,
H€ ohne, M., Kokotou, V.C., Bornscheuer, B, 151, 10–27.
U.T., and Kokotos, George (2005) J. Org. 24 Rotonda, J., Nicholson, D.W., Fazil, K.M.,
Chem., 70, 8730–8733; (c) Schmidt, M., Gallant, M., Gareau, Y., Labelle, M.,
Barbayianni, E., Fotakopoulou, I., Peterson, E.P., Rasper, D.M., Ruel, R.,
H€ ohne, M., Kokotou, V.C., Bornscheuer, Vaillancourt, J.P., Thornberry, N.A., and
U.T., and Kokotos, George (2007) J. Org. Becker, J.W. (1996) Nat. Struct. Biol., 3,
Chem, 72 782–786. 619–625.
10 (a)Nuijens, T., Cusan, C., Kruijtzer, 25 Weber, J.M. (1995) in Molecular Repertoire
J.A.W., Rijkers, D.T.S., Liskamp, R.M.J., of Adenoviruses (eds W. Doerfler and
and Quaedflieg, P.J.L.M. (2009) Synthesis, P. Bohm), Current Topics in
5, 809–814; (b) Nuijens, T., Cusan, C., Microbiology and Immunology, vol. 199,
Kruijtzer, J.A.W., Rijkers, D.T.S., Springer, Berlin, pp. 227–235.
Liskamp, R.M.J., and Quaedflieg, 26 Dolinar, M., Stoka, V., and Turk, B. (2008)
P.J.L.M. (2009) Tetrahedron Lett., 50, Proteinase Inhibitors and Biological
2719–2721. Control, Jozef Stefan Institute.
11 Frey, P.A. and Hegeman, A.D. (2007) 27 Lendeckel, U. (2009) Viral Proteases and
Enzymatic Reaction Mechanism, IRL, Antiviral Protease Inhibitor Therapy:
Oxford. Proteases in Biology and Disease, Springer,
12 Beynon, R.J. and Bond, J.S. (2001) Berlin.
Proteolytic Enzymes: A Practical Approach, 28 Rodrıguez-Barrios, F. and Gago, F. (2004)
IRL, Oxford. Curr. Top. Med. Chem., 4, 991–1007.
References j741
29 Arrigo, A.P., Tanaka, K., Goldberg, A.L., the Synthesis of Peptides and Proteins,
and Welch, W.J. (1988) Nature, 331, CRC Press, Boca Raton.
192–194. 47 Howl, J. (ed.) (2005) Peptide Synthesis
30 Gallastegui, N. and Groll, M. (2010) and Applications, Humana Press.
Trends Biochem. Sci., 35, 634–642. 48 Merrifield, R.B. (1963) J. Am. Chem. Soc.,
31 Voges, D., Zwickl, P., and Baumeister, W. 85, 2149.
(1999) Annu. Rev. Biochem., 68, 49 Fields, G.B. (ed.) (1997) Solid-Phase
1015–1068. Peptide Synthesis, Methods in Enzymology,
32 Ansorge, S. (2000) Cellular Peptidases in vol. 289, Academic Press, San Diego.
Immune Functions and Diseases 2, 50 Kates, S.A. and Albericio, F. (2000) Solid-
Springer, Berlin. Phase Synthesis: A Practical Guide, Marcel
33 Barret, A.J. and Salvesen, G. (eds) (1986) Dekker.
Proteinase Inhibitors, Elsevier, 51 Chan, W.C. and White, P.D. (2000) Fmoc
Amsterdam. Solid Phase Peptide Synthesis: A Practical
34 Beynon, R.J. and Bond, J.S. (2001) Approach, IRL, Oxford.
Proteolytic Enzymes: A Practical Approach, 52 Bray, B.L. (2003) Nat. Rev. Drug Discovery,
IRL, Oxford, pp. 145–162. 2, 587–593.
35 Sterchi, E.E. and St€ocker, W. (1999) 53 Andersson, L., Blomberg, L., Flegel, M.,
Proteolytic Enzymes: Tools and Targets, Lepsa, L., Nilsson, B., and Verlander, M.
Springer, Berlin. (2000) Biopolymers, 55, 227–250.
36 McPherson, A. and Eisenberg, D.S. 54 Berger, S.L. and Kimmel, A.R. (eds)
(2010) Advances in Protein Chemistry (1987) Guide to Molecular Cloning
and Structural Biology, Academic Press. Techniques, Methods Enzymology, vol. 152,
37 Cavanagh, J. (2007) Protein NMR Academic Press, San Diego.
Spectroscopy: Principles and Practice, 55 Fersht, A. and Winter, G. (1992) Trends
Academic Press. Biochem. Sci., 17, 292–294.
38 Domon, B. and Aebersold, R. (2006) 56 El-Mansi, M. and Bryce, C.F.A. (2007)
Science, 312, 212–217. Fermentation Microbiology and
39 Polaina, J. and MacCabe, A.P. (2007) Biotechnology, CRC/Taylor & Francis.
Industrial Enzymes: Structure, Function 57 Gellissen, G. (ed.) (2004) Production of
and Applications, Springer, Berlin. Recombinant Proteins, Wiley-VCH Verlag,
40 Wieland, T. and Bodanszky, M. (1991) GmbH, Weinheim.
The World of Peptides: A Brief History of 58 Dawson, P.E., Muir, T.W., Clark-Lewis, I.,
Peptide Chemistry, Springer, Berlin. and Kent, S.B.H. (1994) Science, 266,
41 Sewald, N. and Jakubke, H.-D. (2009) 776–779.
Peptides: Chemistry and Biology, 2nd edn, 59 Tam, J.P., Lu, Y.-A., Liu, C.-F., and Shao, J.
Wiley-VCH Verlag, GmbH, Weinheim. (1995) Proc. Natl. Acad. Sci. USA, 92,
42 Jakubke, H.-D. and Sewald, N. (2008) 12485–12489.
Peptides from A to Z, Wiley-VCH Verlag, 60 Dawson, P.E. and Kent, S.B.H. (2000)
GmbH, Weinheim. Annu. Rev. Biochem., 69, 923–960.
43 Goodman, M., Felix, A., Moroder, L., and 61 Bordusa, F. (2002) Chem. Rev., 102,
Toniolo, C. (2002) Synthesis of Peptides and 4817–4867.
Peptidomimetics Houben-Weyl-Methoden 62 Hermanson, G.T. (2008) Bioconjugate
der Organischen Chemie, vol. E22 (ed. K.H. Techniques, Academic Press.
B€uchel), Thieme, Stuttgart. 63 Haase, C. and Seitz, O. (2009) Eur. J. Org.
44 Bodanszky, M. and Bodanszky, A. (1994) Chem., 2096–2101.
The Practice of Peptide Synthesis, Springer, 64 Wieland, T., Bokelmann, E., Bauer, L.,
Berlin. and Lang, H.U. (1953) Annalen, 383,
45 Jakubke, H.-D. (1996) Peptide: Chemie 129–149.
und Biologie, Spektrum Akademischer 65 Saleh, L. and Perler, F.B. (2006) Chem.
Verlag, Heidelberg. Rec., 6, 183–193.
46 Lloyd-Williams, P., Alberico, F., and 66 Muir, T., Sondhi, D., and Cole, P.A. (1998)
Giralt, E. (1997) Chemical Approaches to Proc. Natl. Acad. Sci. USA, 95, 6705–6710.
742
67 Evans, T.C. Jr., Benner, J., and Xu, M.-Q. ASI Series, vol. 381, Kluwer Academic
(1998) Prot. Sci., 7, 2256–2264. Publ., pp. 189–198.
68 Machova, Z., von Eggelkraut-Gottanka, 86 Jakubke, H.-D. (1994) J. Chin. Chem. Soc.,
R., Wehofsky, N., Bordusa, F., and 41, 355–370.
Beck-Sickinger, G. (2003) Angew. Chem. 87 Bongers, J. and Heimer, E.P. (1994)
Int. Ed., 42, 4916–4918. Peptides, 15, 183–193.
69 Mao, H.Y., Hart, S.A., Schink, A., and 88 Jakubke, H.-D., Eichhorn, U.,
Pollok, B.A. (2004) J. Am. Chem. Soc., 126, H€ansler, M., and Ullmann, D. (1996)
2670–2671. Biol. Chem., 377, 455–464.
70 Wehofsky, N., Wespe, C., Ce rovsky, V., 89 Jakubke, H.-D. (1995) Angew. Chem., Int.
Pech, A., Hoess, E., Rudolph, R., and Ed. Engl., 34, 175–177.
Bordusa, F. (2008) ChemBioChem., 9, 90 Jakubke, H.-D., Eichhorn, U., Gerisch, S.,
1493–1499. H€ansler, M., Salchert, K., Ullmann, G.,
71 Copping, L.G., Martin, R., Pickett, J.A., and Ullmann, D. (1996) Molecular Design
Bucke, C., and Bunch, A.W. (eds) (1990) and Bioorganic Catalysis (eds C.S.
Opportunities in Biotransformations, Wilcox, and A.D. Hamilton), NATO ASI
Elsevier, London. Series, vol. 478, Kluwer Academic
72 Elleuche, S. and P€oggeler, S. (2010) Publishers, pp. 53–70.
Appl. Microbiol. Biotechnol., 87, 479–489. 91 H€ansler, M. and Jakubke, H.-D. (1996)
73 Aehle, W. (2007) Enzymes in Industry: Amino Acids, 11, 379–395.
Production and Applications, Wiley-VCH 92 Ullmann, D. and Jakubke, H.-D. (1999)
Verlag GmbH, Weinheim. in Proteolytic Enzymes: Tools and Targets
74 Gotor, V., Alfonso, I., and Garcıa- (eds E.E. Sterchi, and W. St€ocker),
Urdiales, E. (2008) Asymmetric Organic Springer-Verlag, Berlin, ch. 18.
Synthesis with Enzymes, Wiley-VCH 93 Jakubke, H.-D. (1999) in Future Aspects
Verlag GmbH, Weinheim. in Peptide Chemistry, vol. 1 (eds W. Voelter,
75 Ramakrishnan, V. (2002) Cell, 108, and G. Fischer), Collection Symposium
5575–5672. Series, Academy of Sciences of the Czech
76 Zhang, B. and Cech, T. (1998) Chem. Biol., Republic, Prague, pp. 47–66.
5, 539–553. 94 Ullmann, D., Kurth, T., Grahn, S.,
77 Marahiel, M.A. (2009) J. Pept. Sci., 12, Elsner, C., Bordusa, F., and Jakubke, H.-
799–807. D. (1999) in Future Aspects in
78 vant Hoff, J.H. (1898) Z. Anorg. Chem., Peptide Chemistry, vol. 1 (eds W.
18, 1–13. Voelter, and G. Fischer), Collection
79 Ostwald, W. (1901) Z. Elektrochem., 7, Symposium Series, Academy of Sciences
995–1004. of the Czech Republic, Prague,
80 Bergmann, M. and Fraenkel-Conrat, H. pp. 194–210.
(1937) J. Biol. Chem., 119, 707–720. 95 Bordusa, F. and Jakubke, H.-D. (2002)
81 Jakubke, H.-D., Kuhl, P., and Enzymatic peptide synthesis, in Synthesis
K€onnecke, A. (1985) Angew. Chem., Int. of Peptides and Peptidomimetics (eds M.
Ed. Engl., 24, 85–93. Goodman, A. Felix, L. Moroder, and C.
82 Kullmann, W. (1987) Enzymatic Peptide Toniolo), Houben-Weyl, vol. E22a, Georg
Synthesis, CRC Press, Boca Raton. Thieme Verlag, Stuttgart.
83 Jakubke, H.-D. (1987) Enzymatic peptide 96 Tai, D.-F. (2003) Curr. Org. Chem., 7,
synthesis, in The Peptides: Analysis, 515–554.
Synthesis, Biology, vol. 9 (eds S. 97 Lombard, C., Saulnier, J., and
Udenfriend and J. Meienhofer), Wallach, J.M. (2005) Prot. Pep. Lett., 12,
Academic Press, New York, pp. 103–165. 621–629.
84 Schellenberger, V. and Jakubke, H.-D. 98 Kumar, D. and Bhalla, T.C. (2005)
(1991) Angew. Chem., Int. Ed. Engl., 30, Appl. Microbiol. Biotechnol., 68, 726–736.
1437–1449. 99 Hagen, J. (2006) Industrial Catalysis: A
85 Jakubke, H.-D. (1992) Microbial Reagents Practical Approach, 2nd edn, Wiley-VCH
in Organic Synthesis, (ed. S. Servi), NATO Verlag GmbH, Weinheim.
References j743
100 Guzm
an, F., Barberis, S., and Illanes, A. 119 Keil, B. (1987) Protein Sequences Data
(2007) Electron. J. Biotechnol., 10, Anal., 1, 13–20.
279–314. 120 Isowa, Y., Ohmori, M., Ichikawa, T.,
101 Chen, F., Zhang, F., Wang, A., Li, H., Mori, K., Nonaka, Y., Kihara, K.,
Wang, Q., Zeng, Z., Wang, S., and Xie, T. Oyama, K., Satoh, H., and Nishimura, S.
(2010) Afric. J. Pharm. Pharmacol., 4, (1979) Tetrahedron Lett., 28, 2611–2614.
721–730. 121 Markussen, J. (1987) Human Insulin by
102 Homandberg, G.A., Komoriya, A., and Tryptic Transpeptidation of Porcine Insulin
Chaiken, I.M. (1982) Biochemistry, 21, and Biosynthetic Precursors, MTP Press,
3385–3389. Lancaster.
103 Mizutani, M. (1925) Z. Phys. Chem., 116, 122 Morihara, K. and Oka, T. (1983) Peptide
350–358. Chemistry 1982 (ed. S. Sakakibara,),
104 Homandberg, G.A., Mattis, J.A., and Protein Research Foundation, Osaka,
Laskowski, M. Jr. (1978) Biochemistry, 17, p. 231.
5220–5227. 123 Fischer, A., Bommarius, A., Drauz, K.,
105 Kuhl, P. and Jakubke, H.-D. (1990) and Wandrey, C. (1994) Biocatalysis, 8,
Pharmazie, 45, 393–400. 289–307.
106 L€uthi, P. and Luisi, P.D. (1984) J. Am. 124 Fl€orsheimer, A., Kula, M.-R., Sch€utz,
Chem. Soc., 106, 7285–7286. H.J., and Wandrey, C. (1989) Biotechnol.
107 Oyama, K., Nishimura, S., Sonaka, Y., Bioeng., 33, 1400–1405.
Kihara, K., and Hashimoto, T. (1981) 125 Hermann, G., Schwarz, A., Wandrey, C.,
J. Org. Chem., 46, 5242–5244. Knaup, G., Drauz, K., and Berndt, H.
108 Carrea, G. and Riva, S. (eds) (2008) (1991) Biotechnol. Appl. Biochem., 13,
Organic Synthesis with Enzymes in Non- 346–353.
Aqueous Media, Wiley-VCH Verlag 126 Thorbek, P., Aasmul-Olsen, S., and
GmbH, Weinheim. Widmer, F. (1988) Patent Appl.
109 Kitaguchi, H. and Klibanov, A.M. WO8806187(A1).
(1989) J. Am. Chem. Soc., 111, 127 Aasmul-Olsen, S., Thorbek, P., Hansen,
5242–5244. S., and Widmer, F. (1990) Proceedings
110 van Rantwijk, F. and Sheldon, R.A. (2007) 5th European Congress on Biotechnology
Chem. Rev., 107, 2757–2785. (eds C. Christiansen, L. Munk, and
111 K€onnecke, A., Schellenberger, V., J. Villadsen), Munsgaard,
Hofmann, H.-J., and Jakubke, H.-D. Copenhagen, pp. 429–432.
(1984) Pharmazie, 39, 785–786. 128 Aasmul-Olsen, S., Thorbek, P.,
112 Schellenberger, V. and Jakubke, H.-D. Hansen, S., and Widmer, F. (1991)
(1986) Biochim. Biophys. Acta, 869, Peptides 1990 (eds E. Giralt and D.
54–60. Andreu), ESCOM Science Publishers,
113 Schellenberger, V., Schellenberger, U., Leiden, pp. 299–300.
Mitin, Y.V., and Jakubke, H.-D. (1987) 129 Aasmul-Olsen, S., Thorbek, P.,
Anal. Biochim., 165, 327–333. Hansen, S., and Widmer, F. (1990) Patent
114 Schellenberger, V., Schuster, M., and Appl. WO9001555(A1).
Jakubke, H.-D. (1990) Biocatalysis, 4, 130 Thorbek, P., Houen, G., Aasmul-Olsen,
105–111. S., and Widmer, F. (1989) Peptides 1988
115 Schuster, M., Kasche, V., and (eds G. Jung and E. Bayer), de Gruyter,
Jakubke, H.-D. (1992) Biochim. Biophys. Berlin, pp. 271–273.
Acta, 1121, 207–212. 131 Ullmann, G. and Jakubke, H.-D. (1993)
116 Ullmann, D. and Jakubke, H.-D. (1994) Peptides 1992 (eds C.-H. Schneider, and
Eur. J. Biochem., 223, 865–872. A. Eberle), ESCOM Science Publishers,
117 Bordusa, F., Ullmann, D., and Leiden, pp. 36–37.
Jakubke, H.-D. (1997) Biol. Chem., 378, 132 (a)Martinek, K., Semenov, A.N., and
1193–1198. Beresin, A.N. (1980) Dokl. Akad.
118 Bordusa, F. and Jakubke, H.-D. (1998) Nauk, SSSR, 254, 121–123; (b) Martinek,
Bioorg. Med. Chem., 6, 1775–1780. K., Semenov, A.N., and Beresin, A.N.
744
(1981) Biochim. Biophys. Acta, 658, Gonzalez, G., Dolor Benaiges, M.,
76–89. Lopez-Santin, J., and Adlercreutz, P.
133 Barberis, S., Quiroga, E., Arribere, M.C., (1999) Future Aspects in Peptide Chemistry,
and Priolo, N. (2002) J. Mol. Cat. B: vol. 1 (eds W. Voelter and G. Fischer),
Enzym., 1, 39–47. Collection Symposium Series, Academy
134 L€uthi, P. and Luisi, P.D. (1984) of Sciences of the Czech Republic,
J. Am. Chem. Soc., 106, 7285–7286. Prague, pp. 203–210.
135 Dias, A., Feliciano, A., Cabral, J., and 152 Nakanishi, K., Kamikubo, A.T., and
Prazeres, D. (2007) J. Colloid Interface Sci., Matsuno, R. (1985) Biotechnology, 3,
305, 198–201. 459–464.
136 Blocher, M., Liu, D., Walde, P., and 153 Oka, T., Muneyuki, R., and Morihara, K.
Luisi, P.L. (1999) Macromolecules, 32, (1984) Proceedings 8th American Peptide
7332–7334. Symposium (eds V.J. Hruby and
137 Dordick, J.S. (1989) Enz. Microb. Technol., D.H. Rich), Pierce Chemical Company,
11, 194–211. Rockford, IL, pp. 199–201.
138 Margolin, A.L., Tai, D.-F., and 154 Matsushima, A., Okada, M., and Inada, Y.
Klibanov, A.M. (1987) J. Am. Chem. Soc., (1984) FEBS Lett., 178, 275–277.
109, 7885–7887. 155 Gaertner, H.F. and Puigserver, A.J. (1988)
139 Kise, H., Hayakawa, A., and Noritomi, H. Proteins, 3, 130–137.
(1990) J. Biotechnol., 14, 239–254. 156 Lee, H., Takahashi, K., Kodera, Y.,
140 Li, S., Wang, J., Xu, L., Zhang, X., Li, J., Owada, T., Tsuzuki, T., Matsushima, A.,
and Suo, D. (2008) Prep. Biochem. and Inada, Y. (1988) Biotechnol. Lett., 10,
Biotechnol., 38, 334–347. 403–407.
141 Yeş iloglu, Y. and Sagiroglu, A. (2002) 157 Ferjancis, A., Puigserver, A.J., and
Turk. J. Chem., 26, 529–534. Gaertner, H.F. (1988) Biotechnol. Lett., 10,
142 Filippova, I.Y. and Lysogorskaya, E.N. 101–106.
(2003) Rus. J. Bioorg. Chem., 29, 158 Castillo, B., Sola, R.J., Ferrer, A.,
496–501. Barletta, G., and Griebenow, K. (2008)
143 Mishima, K., Matsuyama, K., Baba, M., Biotechnol. Bioeng., 99, 9–17.
and Chidori, M. (2003) Biotechnol. Prog., 159 Sheldon, R.A. (2007) Biochem. Soc. Trans.,
19, 281–284. 35, 1583–1587.
144 Miao, W. and Chan, T.J. (2005) 160 Riva, S., Chopineau, J., Kieboom, A.P.G.,
Org. Chem., 70, 3251–3255. and Klibanov, A.M. (1988) J. Am. Chem.
145 Plaquevent, J., Levillain, J., Guillen, F., Soc., 110, 584–589.
Malhiac, C., and Gaumont, A. (2008) 161 Nuijens, T., Cusan, C., van Dooren,
Chem. Rev., 108, 5035–5060. T.J.G.M., Moody, H.M., Merkx, N.S.M.,
146 Erbeldinger, M., Mesiano, A., and Kruijtzer, J.A.W., Rijkers, D.T.S.,
Russell, A. (2008) Biotechnol. Prog., 16, Liskamp, R.M.J., and Quaedflieg,
1129–1131. P.J.L.M. (2010) Adv. Syn. Catal., 352,
147 Jakubke, H.-D., Bullerjahn, R., 2399–2404.
H€ansler, M., and K€onnecke, A. (1982) 162 Nuijens, T., Piva, E., Kruijtzer, J.A.W.,
Affinity Chromatography and Related Rijkers, D.T.S., Liskamp, R.M.J., and
Techniques, Elsevier, Amsterdam, Quaedflieg, P.J.L.M. (2011) Adv. Syn.
pp. 529–536. Catal., 353, 1039–1044.
148 Jakubke, H.-D. and K€onnecke, A. (1987) 163 Nuijens, T., Cusan, C., Schepers, C.H.M.,
Methods Enzymol., 136, 178–188. Kruijtzer, J.A.W., Rijkers, D.T.S.,
149 Gill, I. and Vulfson, E.N. (1992) J. Chem. Liskamp, R.M.J., and Quaedflieg,
Soc., Perkin Trans. I, 667–779. P.J.L.M. (2011) J. Mol. Cat. B: Enzym., 71,
150 Clapes, P., Torres, J.L., and 79–84.
Adlercreutz, P. (1995) Bioorg. Med. Chem., 164 Boeriu, C.G., Frissen, A.E., Boer, E.,
3, 245–257. van Kekem, K., van Zoelen, D.-J., and
151 Clapes, P., Capellas, M., Pera, E., Calvet, Eggen, I.F. (2010) J. Mol. Cat. B: Enzym.,
S., Torres, J.-L., Fite, M., Caminal, G., 66, 33–38.
References j745
165 Davis, B.G. (2003) Curr. Opin. Biotechnol., and bioorganic catalysis, in Proceedings
14, 379–386. of NATO Advanced Workshop of Bioorganic
166 Wong, C.-H. and Wang, K.T. (1991) Catalysis (eds C.S. Wilcox and A.D.
Experientia, 47, 1123–1129. Hamilton), Kluwer Academic Publishers,
167 West, J.B., Scholten, J., Stolowich, N.J., Deventer, pp. 53–70.
Hogg, J.L., Scott, A.I., and Wong, C.-H. 184 Halling, P.J., Eichhorn, U., Kuhl, P., and
(1988) J. Am. Chem. Soc., 110, 3709–3710. Jakubke, H.-D. (1995) Enzyme Microb.
168 Nakasuka, T., Sasaki, T., and Kaiser, E.T. Technol., 17, 601–606.
(1987) J. Am. Chem. Soc., 109, 3808–3810. 185 Eichhorn, U., Bommarius, A.S.,
169 Wu, Z.-P. and Hilvert, D. (1989) Drauz, K., and Jakubke, H.-D. (1997)
J. Am. Chem. Soc., 111, 4513–4514. J. Pept. Sci., 3, 245–266.
170 West, J.B. and Wong, C.-H. (1986) 186 Richards, A.O., Gill, I.S., and
J. Chem. Soc., Chem. Commun., 417–418. Vulfson, E.N. (1993) Enzyme Microb.
171 Remy, H.M., Bacon, P., Bourdillon, C., Technol., 15, 928–935.
and Thomas, D. (1987) Biochim. Biophys. 187 Erbeldinger, M., Ni, X., and Halling, P.J.
Acta, 911, 252–255. (1998) Biotechnol. Bioeng., 59, 68–72.
172 (a)Wong, C.-H. (1992) TIBTECH, 10, 188 Erbeldinger, M., Ni, X., and Halling, P.J.
378–381; (b) Wong, C.-H. (1993) Chimia, (1998) Microb. Technol., 23, 141–148.
47, 127–132. 189 Schellenberger, V., G€orner, A.,
173 Jackson, D.Y., Vurnier, J., Quan, C., K€onnecke, A., and Jakubke, H.-D. (1991)
Stanley, M., Tom, J., and Wells, J.A. (1994) Peptide Res., 4, 265–269.
Science, 266, 243–247. 190 Bongers, J., Lambros, T., Liu, W.,
174 Atwell, S. and Wells, J.A. (1999) Ahmad, M., Campbell, R.M., Felix, A.M.,
Proc. Natl. Acad. Sci. USA, 96, 9497–9502. and Heimer, E.P. (1992) J. Med. Chem.,
175 Grant, N.H. and Alburn, H.E. (1966) 35, 3934–3941.
Nature, 212, 194. 191 Nakajima, I., Kitabake, S., Tsurutani, R.,
176 (a)H€ansler, M. and Jakubke, H.-D. (1996) Tomiaka, I., and Yamamoto, K. (1984)
J. Pept. Sci., 2, 279–289; (b) H€ansler, M. Biochim. Biophys. Acta, 790, 197–199.
and Jakubke, H.-D. (1996) Amino Acids, 192 Kato, M. and Takanishi, T. (1984)
11, 379–395. Agric. Biol. Chem., 48, 2577–2578.
177 Schuster, M., Aaviksaar, A., and 193 Kleinkauf, H., van Liempt, H.,
Jakubke, H.-D. (1990) Tetrahedron, 46, Palissa, H., and D€ohren, H.v. (1992)
8093–8102. Naturwissenschaften, 79, 153–162.
178 Schuster, M., Aaviksaar, A., Haga, M., 194 Winkler, F.K., DAcry, A., and
Ullmann, U., and Jakubke, H.-D. (1991) Hunzicker, W. (1990) Nature, 343,
Biomed. Biochim. Acta, 50, 84–89. 771–778.
179 Littlemore, L., Schober, P., and 195 Brady, L., Brzozowski, A.M.,
Widmer, F. (1993) Peptide Chemistry 1992 Derewenda, Z.S., Dodsen, E.,
(ed. N. Yanaihara), ESCOM Science Dodsen, G., Tolley, S., Turkenerg, J.P.,
Publishers, Leiden, pp. 185–187. Christiansen, L., Huge-Jensen, B.,
180 Ullmann, G. and Jakubke, H.-D. (1993) Norskov, L., Thim, L., and Menge, U.
Peptides 1992 (eds C.H. Schneider and (1990) Nature, 343, 767–770.
A.N. Eberle), ESCOM Science 196 Maruyama, T., Nakajima, M., Kondo, H.,
Publishers, Leiden, pp. 36–37. Kawasaki, K., Seki, M., and Gato, M.
181 Tougu, V., Meos, H., Haga, M., (2003) Enz. Microb. Technol., 32,
Aaviksaar, A., and Jakubke, H.-D. (1993) 655–657.
FEBS Lett., 329, 40–42. 197 So, J.-E., Kang, S.-H., and Ki, B.-G. (1998)
182 Gerisch, S., Jakubke, H.-D., and Enz. Microb. Technol., 23, 211–215.
Kreuzfeld, H.-J. (1995) Tetrahedron: 198 Goujard, L., Figueroa, M.C., and
Asymmetry, 6, 3039–3049. Villeneuve, P. (2004) Biotechnol. Lett., 26,
183 Jakubke, H.-D., Eichhorn, U., Gerisch, S., 1211–1216.
H€ansler, M., Salchert, K., Ullmann, G., 199 Hanefeld, U. (2003) Org. Biomol. Chem.,
and Ullmann, D. (1996) Molecular design 1, 2405–2415.
746
200 West, J.B. and Wong, C.-H. (1987) 218 G€

unther, R., Elsner, C., Schmidt, S.,
Tetrahedron Lett., 28, 1629–1632. Hofmann, H.J., and Bordusa, F. (2004)
201 Margolin, A.L. and Klibanov, A.M. (1987) Org. Biomol. Chem., 2, 1442–1446.
J. Am. Chem. Soc., 109, 3802–3804. 219 Rall, K. and Bordusa, F. (2002)
202 Hirschmann, R., Smith, A.B. III, J. Org. Chem., 67, 9103–9106.
Taylor, C.M., Benkovic, P.A., Taylor, S.D., 220 Thust, S. and Koksch, B. (2003)
Yager, K.M., Sprengeler, P.A., and J. Org. Chem., 68, 2290–2296.
Benkovic, S.J. (1994) Science, 265, 221 (a)Wagner, G. and Horn, H. (1973)
234–237. Pharmazie, 28, 428–431; (b) Tanizawa, K.,
203 Jacobsen, J.R. and Schulz, P.G. (1994) Kasaba, Y., and Kanaoka, Y. (1977)
Proc. Natl. Acad. Sci. USA, 91, 5888–5892. J. Am. Chem. Soc., 99, 4484–4488.
204 Ullmann, D., Salchert, K., Bordusa, F., 222 Thormann, M., Thust, S.,
Schaaf, R., and Jakubke, H.-D. (1994) Hofmann, H.-J., and Bordusa, F. (1999)
Proceedings 5th Akabori Conference (ed. E. Biochemistry, 38, 6056–6062.
W€ unsch) Max-Planck Gesellschaft, 223 G€unther, R., Stein, A., and Bordusa, F.
Munich, pp. 70–75. (2000) J. Org. Chem., 65, 1672–1679.
205 Kay, J. and Kassell, B. (1971) 224 G€unther, R. and Bordusa, F. (2000)
J. Biol. Chem., 246, 6661–6665. Chem. Eur. J., 6, 463–467.
206 Gertler, A., Walsh, K.A., and Neurath, H. 225 Wehofsky, N., Pech, A., Liebscher, S.,
(1974) Biochemistry, 13, 1302–1310. Schmidt, S., Komeda, H., Asano, Y., and
207 Huber, R. and Bode, W. (1978) Chem. Res., Bordusa, F. (2008) Angew. Chem. Int. Ed.,
11, 114–118. 47, 5456–5460.
208 Bordusa, F. (2000) Braz. J. Med. Biol. Res., 226 Wehofsky, N. and Bordusa, F. (1998)
33, 469–485. FEBS Lett., 443, 220–224.
209 G€unther, R. and Bordusa, F. (2000) 227 Wehofsky, N., Wissmann, J.-D.,
Chemistry, 6, 463–467. Alisch, M., and Bordusa, F. (2000)
210 Bordusa, F. (2002) Curr. Prot. Pep. Sci., 3, Biochim. Biophys. Acta, 1479, 114–122.
159–180. 228 Morree, W.J., Sears, P., Kawashiro, K.,
211 (a)Schellenberger, V., Jakubke, H.-D., Witte, K., and Wong, C.-H. (1997)
Zapevalova, N.P., and Mitin, Y.V. (1991) J. Am. Chem. Soc., 119, 3942–3947.
Biotechnol. Bioeng., 38, 104–108; (b) 229 Degrado, W.F. and Kaiser, E.T. (1980)
Schellenberger, V., Jakubke, H.-D., J. Org. Chem., 45, 1295–1300.
Zapevalova, N.P., and Mitin, Y.V. (1991) 230 Mihara, H., Chmielewski, J.A., and
Biotechnol. Bioeng., 38, 319–321. Kaiser, E.T. (1993) J. Org. Chem., 58,
212 Mitin, Y.V., Schellenberger, V., 2209–2215.
Schellenberger, U., Jakubke, H.-D., and 231
Cerovsk y, V. and Bordusa, F. (2000)
Zapevalova, N.P. (1991) Peptides 1990 J. Peptide Res., 55, 325–329.
(eds E. Giralt and D. Andreu), ESCOM 232 Yan, A.-X., Chan, R.Y.K., Lau, W.-S.,
Science Publishers, Leiden, Lee, K.-S., Wong, M.-S., Xing, G.-W.,
pp. 287–288. Tiana, G.-L., and Ye, Y.-H. (2005)
213 Bordusa, F., Ullmann, D., Elsner, C., and Tetrahedron, 61, 5933–5941.
Jakubke, H.-D. (1997) Angew. Chem., Int. 233 Park, O.-J., Jeon, G.-J., and Yang, J.-W.
Ed. Engl., 36, 2473–2475. (1999) Enz. Microb. Technol., 25, 455–462.
214 Sekizaki, H., Itoh, K., Toyota, E., and 234 Babonneau, M.T., Jacquier, R., Lazaro, R.,
Tanizawa, K. (1996) Chem. Pharm. Bull. and Viallefont, P. (1989) Tetrahedron Lett.,
Jpn., 44, 1577–1579. 30, 2787–2790.
215 Sekizaki, H., Itoh, K., Toyota, E., and 235 Miyazawa, T., Tanaka, K., Ensatsu, E.,
Tanizawa, K. (1997) Tetrahedron Lett., 38, Yanagihara, R., and Yamada, T. (2001)
1777–1780. J. Chem. Soc., Perkin Trans. 1, 87–93.
216 Xu, S., Rall, K., and Bordusa, F. (2001) 236 Belyaeva, A.V., Bacheva, A.V.,
J. Org. Chem., 66, 1627–1632. Oksenoit, E.S., Lysogorskaya, E.N.,
217
Cerovsk y, V. and Bordusa, F. (2000) Lozinskii, V.I., and Filippova, I.Yu. (2005)
J. Pept. Res., 55, 325–329. Rus. J. Bioorg. Chem., 31, 6529–6534.
References j747
237 Miyazawa, T., Hiramatsu, M.,
255 Cerovsk
y, V., Pirkova, J., Majer, P.,
Murashima, T., and Yamada, T. (2003) Slaninova, J., and Hlavacek, J. (1989)
Biocatal. Biotransform., 21, 93–100. Peptides 1988 (eds G. Jung and E. Bayer),
238 Kuhl, P., Zacharias, U., Burckhardt, H., de Gruyter Berlin, New York,
and Jakubke, H.-D. (1986) Monatsh. pp. 265–267.
Chem., 117, 1195–1204. 256 Capellas, M., Caminal, G., Gonzalez, G.,
239 Salam, S.M.A., Kagawa, K.-I., and Lopezsantin, J., and Clapes, P. (1997)
Kawashiro, K. (2008) Enz. Microb. Biotechnol. Bioeng., 56, 456–463.
Technol., 43, 537–543. 257 Sakina, K., Kawazura, K., and
240 Mitin, Y.V., Braun, K., and Kuhl, P. (1997) Morihara, K. (1988) Int. J. Peptide Prot.
Biotechnol. Bioeng., 54, 287–290. Res., 31, 245–251.
241 Henkel, B., Zhang, L., and Bayer, E. 258 Kullmann, W. (1982) J. Org. Chem., 47,
(1997) Liebigs Ann./Recueil, 2161–2168. 5300–5303.
242 Mitin, Y.V. and Ryadnov, M.G. (1999) 259 Widmer, F., Bayne, S., Houen, G.,
Protein Peptide Lett., 6, 87–90. Moss, B.A., Rigby, R.D., Whittaker, R.G.,
243 Ryadnov, M.G., Klimenko, L.V., and and Johansen, J.P. (1984) Peptides 1984
Mitin, Y.V. (1999) J. Peptide Res., 53, (ed. U. Ragnarson), Alqvist Wiksell,
322–328. Stockholm, pp. 193–200.
244 Breddam, K. (1986) Carlsberg Res. 260 Widmer, F., Bayne, S., Houen, G.,
Commun., 51, 83–128. Moss, B.A., Rigby, R.D., Whittaker, R.G.,
245 Widmer, F., Breddam, K., and and Johansen, J.T. (1985) Forum
Johansen, J.T. (1981) Peptides 1980 Peptides Le Cap dAgde 1984 (eds J.B.
(ed. K. Brunfeldt), Scriptor, Copenhagen, Castro, and J. Martinez), Groupe Francais
pp. 46–55. de Peptides.
246 Bordusa, F., Ullmann, D., and 261 Jakubke, H.-D., Kuhl, P., K€onnecke, A.,
Jakubke, H.-D. (1997) Angew. Chem., Int. D€oring, G., Walpuski, J., Wilsdorf, A., and
Ed. Engl., 36, 1099–1101. Zapevalova, N.P. (1983) Peptides 1982
247 Kunugi, S., Yokoyama, M., Kuroda, Y., (eds K. Blaha, and P. Malon),
Yoshida, M., Koyasu, A., Yamada, T., de Gruyter, Berlin, pp. 43–45.
and Sakamoto, A. (1996) Bull. Chem. Soc. 262 Kuhl, P., D€oring, G., Neubert, K., and
Jpn., 69, 1747–1754. Jakubke, H.-D. (1984) Monatsh. Chem.,
248 Takai, H., Sakato, K., Nakamizo, N., and 115, 423–430.
Isowa, Y. (1981) Peptide Chemistry 1980 263 Bjorup, P., Torres, J.L., Adlercreutz, P.,
(ed. K. Okawa), Protein Research and Clapes, P. (1998) Bioorgan. Med.
Foundation, Osaka, pp. 213–218. Chem., 6, 891–901.
249 Kullmann, W. (1982) Proc. Natl. Acad. Sci. 264 Kullmann, W. (1984) Biochem. J., 220,
USA, 79, 2840–2844. 405–416.
250
Cerovsk y, V., Hlavacek, J., Slaninova, J., 265 Ye, Y.H., Tian, G.L., Dai, D.C., Chen, G.,
and Jost, K. (1980) Coll. Czech. Chem. and Li, C.X. (1998) Tetrahedron, 54,
Commun., 53, 2766–2772. 12585–12596.
251 Didziapeptris, R.J., Drabnig, B., 266 Aasmul-Olsen, S., Andersen, A.J.,
Schellenberger, V., Jakubke, H.-D., Thorbek, P., and Widmer, F. (1989) 8th
and Svedas, V. (1991) FEBS Lett., 287, International Conference on Tetanus (eds
31–33. G. Nistica, B. Bizzini, B. Butchenko, and
252 Isowa, Y., Ohmori, M., Sato, M., and R. Trian), Pythagora Press, Rome, Milan,
Mori, K. (1977) Bull. Chem. Soc. Jpn., 50, pp. 191–208.
2766–2772. 267
Cerovsk y, V., Kockskaemper, J., Glitsch,
253 Gondo, M., Yamashita, H., Sakakibara, H.G., and Bordusa, F. (2000)
K., and Isowa, Y. (1982) Peptide Chemistry ChemBioChem., 1, 126–129.
1981 (ed. T. Shiori), Protein Research 268 Obermeier, R. and Seipke, G. (1984)
Foundation, Osaka, pp. 93–98. Process Biochem., 19, 29–34.
254
Cerovsk y, V., W€
unsch, E., and Brass, J. 269 Andersen, A.J., Widmer, F., and
(1997) Eur. J. Biochem., 247, 231–237. Johansen, J.T. (1987) Peptides 1986
748
(ed. D. Theodoropoulos), de Gruyter, 279 Kadereit, D. and Waldmann, H. (2001)

Berlin, pp. 183–188. Chem. Rev., 101, 3367–3396.
270 Schuster, M., Aaviksaar, A., and Jakubke, 280 Steinke, D. and Kula, M.R. (1990) Angew.
H.-D. (1992) Tetrahedron Lett., 33, Chem. Int. Ed. Engl., 29, 1139–1140.
2799–2802. 281 Gini, F., Eggen, I.F., van Zoelen, D.-J.,
271 Schellenberger, V., Schellenberger, U., and Boeriu, C. (2009) Chem. Today, 27,
Jakubke, H.-D., H€ansicke, A., Bienert, 24–26.
M., and Krause, E. (1990) Tetrahedron 282 Nuijens, T., Cusan, C., Kruijtzer, J.A.W.,
Lett., 31, 7305–7308. Rijkers, D.T.S., Liskamp, R.M.J., and
272 Kullmann, W. (1983) J. Prot. Chem., 2, Quaedflieg, P.J.L.M. (2009) J. Org. Chem.,
289–301. 74, 5145–5150.
273 Thorbek, P., Lauridsen, J., and Widmer, 283 Budzynski, A.Z. (2001) Lab. Med., 32,
F. (1988) Peptides: chemistry and biology, 365–370.
in Proceedings 10th American Peptide 284 Sato, A. (2002) Adv. Drug Deliv. Rev., 54,
Symposium (ed. G.R. Marshall), ESCOM 487–504.
Science Publishers, Leiden, pp. 279–281. 285 Bourgeaux, V., Cadene, M., Piller, F., and
274 Bille, V., Ripak, C., van Assche, I., Forni, Piller, V. (2007) ChemBioChem, 8, 37–40.
L., Degelaen, J., and Scarso, A. (1991) 286 Gardossi, L., Bianchi, D., and Klibanov,
Peptides 1990 (eds E. Giralt and D. A.M. (1991) J. Am. Chem. Soc., 113,
Andreu), ESCOM Science Publishers, 6328–6329.
Leiden, pp. 253–254. 287 Tan, X.H., Wirjo, A., and Liu, C.F. (2007)
275 Kuhl, P., D€oring, G., Neubert, K., and ChemBioChem, 8, 1512–1515.
Jakubke, H.-D. (1984) Pharmazie, 39, 288 Tan, X.H., Yang, R., Wirjo, A., and Liu,
814–816. C.F. (2008) Tetrahedron Lett., 49,
276
Cerovsk y, V. (1986) Collect. Czech. Chem. 2891–2894.
Commun., 51, 1352–1360. 289 Quaedflieg, P.J.L.M. and Merkx, N.S.M.,
277 Schellenberger, V., Tegge, W., and Frank, (2009) Patent Appl. WO2009047354 (A1).
R. (1992) Int. J. Pept. Protein Res., 39, 290 Kohli, R.M., Burke, M.D., Tao, J., and
472–476. Walsh, C.T. (2003) J. Am. Chem. Soc., 125,
278 Schellenberger, V., Pompejus, M., and 7160–7161.
Fritz, H.-J. (1993) Int. J. Pept. Protein Res., 291 Gr€unewald, J. and Marahiel, M.A. (2006)
41, 326–332. Microbiol. Mol. Biol. Rev., 70, 121–146.
j749
18
CN Lyases Catalyzing Addition of Ammonia, Amines,
and Amides to C¼C and C¼O Bonds
Bian Wu, Wiktor Szymanski, Ciprian G. Crismaru, Ben L. Feringa, and Dick B. Janssen
18.1
Introduction
The addition of ammonia and amines to double bonds is catalyzed by various

microbial and plant lyases belonging to different EC classification groups and
phylogenetic families. Several ammonia lyases are classified in group EC 4.3.1,
which encompasses enzymes that add ammonia to C¼C double bonds, with L-
aspartase and L-histidine ammonia lyase as classical examples. Enzymes that are
phylogenetically related to the L-aspartases in group EC 4.3.1 can be found in group
EC 4.3.2, which encompasses CN lyases that add amines and amidines to C¼C or
C¼O bonds. Lyases in group EC 4.3.3 include enzymes that add amines to C¼C or
C¼O double bonds. Clearly, the EC classification according to reaction type does not
correlate with a phylogenetic classification and reaction mechanism. In this chapter,
different classes of CN lyases are grouped according to their phylogenetic classi-
fication. Other reviews discussing ammonia lyases have appeared [1–5].
A very common ammonia lyase is L-aspartase, which occurs in many bacteria and
plants, where it is involved in the catabolism of L-aspartate to produce ammonia and
fumarate, the latter being channeled into the tricarboxylic acid cycle or, in some
bacteria, used as an electron acceptor for fermentative growth. The biological
function of microbial L-histidine and L-phenylalanine ammonia lyases is also
catabolic, and the same holds for methylaspartate ammonia lyase, an enzyme that
is not phylogenetically related to L-aspartase, but catalyses a very similar reaction.
Other CN lyases, such as plant L-phenylalanine ammonia lyases and tyrosine
ammonia lyases, play a role in biosynthesis by generating the corresponding
unsaturated carboxylic acids, which are used for lignin synthesis. Tyrosine ammonia
lyase is also involved in the synthesis of coumarate, which can serve as a precursor for
cofactors or as a building block for some antibiotics.
The biotechnological relevance of ammonia lyases is mainly based on the
reverse reactions: addition of ammonia to double bonds, with fumarate and
and Oliver May.
j 18 CN Lyases Catalyzing Addition of Ammonia, Amines, and Amides to C¼C and C¼O Bonds
750
aromatic a,b-unsaturated carboxylic acids as the most common acceptors. Thus,

L-aspartase has found application since the 1970s in the form of a whole-cell
biocatalyst in the production of L-aspartate from fumarate and ammonia [5].
L-Phenylalanine ammonia lyase can be used for the production of L-phenylalanine
from cinnamic acid, and is active with several substituted analogs [6]. As the
progress of these reactions is dependent on the thermodynamic equilibrium, high
concentrations of ammonia are required. Several other amines can possibly be
synthesized with these enzymes as well. Further applications will depend on the
discovery and/or engineering of enzyme variants that show improved conversion
kinetics, accept a broader range of substrates, or that are better compatible with
the required reaction conditions.
18.2
Addition of Ammonia and Amines to Fumaric Acid: L-Aspartase-Fumarase
Superfamily
18.2.1
General Properties
Aspartases (L-aspartate ammonia lyases, EC 4.3.1.1) catalyze the reversible elimina-

tion of ammonia from L-aspartic acid to produce fumaric acid [5]. The enzymes are
very specific. For example, the Escherichia coli aspartase, the gene of which was cloned
in 1985 [7], has no activity with D-aspartic acid, L-asparagine, other L-amino acids, or
substituted aspartic acids like (2S,3S)-3-methylaspartic acid. In the reverse reaction,
there is no activity with unsaturated acids other than fumaric acid, and activity with
nucleophiles other than ammonia is very restricted. In aerobic bacteria, aspartases
are involved in the degradation of L-aspartate, and in facultative or obligate anaerobes
they play a role in the formation of fumarate that can serve as an electron acceptor
during fermentative growth.
Well-studied aspartases originate from E. coli, Bacillus sp. YM55-1 and Hafnia
alvei [5]. These enzymes are homotetrameric proteins with subunits of 52–55 kDa.
The E. coli enzyme binds Mg2 þ at high pH and is activated allosterically by aspartate if
Mg2 þ is present [8]. The metal-bound form is more active at higher pH. In contrast,
the thermostable enzyme from Bacillus sp. binds no metal and lacks allosteric
regulation [9]. The aspartase from H. alvei is important for industrial L-aspartate
synthesis from fumarate and ammonia and has also been studied kinetically [10].
Based on sequence similarities and structural comparison, aspartases can be
classified as members of the aspartase–fumarase superfamily. Other members of this
superfamily that catalyze CN lyase reactions are argininosuccinate lyase, adeny-
losuccinate lyase, and (S,S)-ethylenediamine disuccinate (EDDS) lyase (see below).
The sequence similarity between these enzymes is low, but there are three highly
conserved regions, called C1, C2, and C3, (residues 96–110, 183–194, and 317–328 in
Bacillus sp. YM55-1 enzyme numbering), that in the structure contribute to shaping
the active site.
18.2 Addition of Ammonia and Amines to Fumaric Acid: L-Aspartase-Fumarase Superfamily j751
18.2.2
Structure and Catalytic Mechanism
The X-ray structures of two different aspartases have been solved, namely, the
enzymes from E. coli (called AspA, PDB 1JSW [11]) and Bacillus sp. YM55-1 (called
AspB, PDB 1J3U [12]).Three domains (N-terminal large domain, central helix
domain, C-terminal small domain) can be distinguished in each subunit. In the
quaternary structure, the helices from the four subunits form a 20- or 24-helix
cluster, mainly through central helix domain interactions. The active sites of the
tetrameric enzymes are formed by three subunits that come together and the active
site groups are donated by residues in the conserved C motifs that form loops.
Often, these loops occur in disordered form in the X-ray structure of different
members of the aspartase–fumarase superfamily. Further proteins with similar
structures are 3-carboxy-cis,cis-muconate lactonizing enzyme and d-crystallin, an
eye lens protein.
Details about the catalytic mechanism of aspartases are still lacking since there are
no structures of enzyme with substrate or product bound. It is also difficult to identify
catalytic residues by comparison since the active site structures of aspartases (and
fumarases) diverge from that of the phylogenetically related adenylosuccinate lyase
and argininosuccinate lyase, for which there are structures with substrate bound. In
addition, the fact that the protonation state of the leaving group (neutral or protonated
ammonia) is unknown makes matters complicated.
Ser318
Ser318
-
O
HO
H -O CO2-
Lys327 NH3+ -O CO2- Lys327 NH3+
O- NH3+
O NH3+
Scheme 18.1 Mechanism of the aspartase reaction. Residue numbering for aspartase from Bacillus
sp. YM55-1 (PDB 1J3U).
Nevertheless, a reaction mechanism for aspartases can be proposed

(Scheme 18.1) [12, 13]. A general base of the enzyme is expected to be involved in
abstraction of the proR proton from the C3 position of aspartate, with stabilization of
the resulting carbanion by delocalization of the negative charge in the aci-form of the
b-carboxylate group. Two carboxylate binding sites have been proposed, which may
include (Bacillus sp. YM55-1 enzyme numbering) aLys324, bThr187, and aAsn326
for binding of the a-carboxylate and cSer140 and cThr141 for the b-carboxylate that
transiently accepts the negative charge during catalysis. If the mechanism of
adenylosuccinate lyase and aspartase is conserved, the catalytic base abstracting the
proton from C3 could be a Ser (e.g., the first Ser of the conserved sequence motif
GSSIMP). After formation of the enedioate, the CN bond is cleaved, releasing
ammonia, which may leave as a neutral species or may become protonated during
752
bond cleavage by an acidic group in the active site. The slowest step in the catalytic
cycle is this cleavage of the CN bond [10]. Assuming that the ammonia becomes
protonated, three hydrogen bond acceptors are expected, and the most likely
candidates are cThr101 OG, cAsn142 OD1, and bHis188 NE2, which may donate
a proton.
18.2.3
Diversity
As mentioned above, aspartases have been purified and/or cloned and sequenced
from several bacterial sources, including E. coli [7], H. alvei [10], Pseudomonas
fluorescens [14], B. subtilis [15], and Bacillus sp. YM55-1 [16]. These enzymes share
high sequence similarity (45–72%) and thus belong to the same superfamily. The
Bacillus enzyme is interesting because of its higher activity, enantioselectivity,
thermostability, and lack of allosteric activation at high pH by aspartate or metal
ions. A surprisingly thermostable variant was obtained from the bacterium
Cytophaga sp. KUC-1, an organism that is a psychrophile and yet is a source of
several thermostable enzymes [17]. The enzyme has the usual tetrameric structure,
and it is cooperatively activated by substrate and Mg2 þ ions. The Vmax value was
higher than that of comparable enzymes: 99 and 326 U mg1 at pH 7 and 8.5,
respectively. Database searches show that many highly aspartase-similar genes (up to
95% identity) can be found in sequenced bacterial genomes, indicating the existence
of a wide variety of unexplored aspartases.
18.2.4
Biocatalytic Scope and Applications
The aspartase reaction is reversible with a Keq ([ fumarate][NH4 þ ]/[aspartate]) at pH

7.0 and 25 C of 6.0 0.2 103 M. This makes it possible to use the enzyme for the
synthesis of L-aspartate from fumarate and ammonia, for example, with the enzyme
from E. coli or H. alvei [18, 19]. Indeed, aspartase is applied in a classical whole-cell
industrial process for the production of L-aspartate that is used for preparing
pharmaceuticals and food additives, such as the low-calorie sweetener aspartame.
Continuous production is possible with cells immobilized by polyacrylamide entrap-
ment, carrageenan entrapment with glutaraldehyde, and hexamethylenediamine
crosslinking, or with cells adsorbed to a phenolformaldehyde resin [20–23].
It has been proposed to use aspartase in a tandem reaction for L-phenylalanine
production [24]. This system is based on aspartase-mediated formation of L-aspartic
acid from fumaric acid and transaminase-mediated transfer of the amino group to
phenylpyruvic acid. The reaction was catalyzed by whole cells of an E. coli strain that
produces both transaminase and aspartase. Phenylalanine accumulated up to 0.24 M
with 97% conversion [25]. The two-enzyme system was suggested to be economically
attractive since the relatively expensive amine donor L-aspartate was replaced by
fumaric acid and ammonia, which is cheaper. Aspartase has also been co-immobilized
18.2 Addition of Ammonia and Amines to Fumaric Acid: L-Aspartase-Fumarase Superfamily j753
with carboxypeptidase A for use in an electrode for the determination of aspartame in
food products or for aspartate assays [26].
Another application would be in production of recombinant proteins. It has been
described that co-expression of aspartase in E. coli can improve the expression of
heterologous proteins by reducing imbalance in the tricarboxylic acid cycle [27].
A restriction concerning other applications in biocatalysis is the narrow substrate
range of the enzyme. No carboxylic acids other than fumarate are known to be
converted. Weiner et al. studied the possibility of using substituted amines as
nucleophiles instead of ammonia, thus producing N-substituted aspartic acid [28].
A decent conversion was found with hydroxylamine and hydrazine, but rates with
methoxylamine and methylamine were lower (Scheme 18.2). All reactions proceeded
with very high enantioselectivity, and the expected N-substituted L-aspartic acids were
formed. Addition of hydroxylamine and hydrazine to fumarate occurred with
a similar kcat as with ammonia, but Km values for the alternative nucleophiles were
2–4-fold higher. The highly enantiopure methoxylamine thus obtained could be
isolated after esterification. Common amino acids did not react. The alternative
product, N-hydroxyaspartic acid, proved difficult to isolate because of its instability.
OH NH2
HN COOH HN COOH
H H H H N
N
HOOC H HOOC H
N
N
Me R
OMe HN COOH
HN COOH HN COOH 1 2 H
H
H H H HOOC H
H
HOOC H HOOC H
COOH R= ribose monophosphate
1 HOOC 3
H2N COOH NH2
H
H H2N N
H COOH
HOOC H
HN COOH
H H
L-aspartate
HOOC H
Scheme 18.2 Reactions catalyzed by ammonia lyases of the aspartase/fumarase superfamily. 1,

L-aspartase; 2, adenylosuccinate lyase; and 3, argininosuccinate lyase.
18.2.5
Enzyme Engineering
Engineering attempts have been aimed at improving the stability, substrate range,
and activity of aspartases. For example, Wang et al. [29] used error-prone PCR and
gene shuffling to obtain aspartase variants with enhanced activities (28-fold higher
kcat/Km) that were also more thermostable. The same group reported studies on fused
754
variants. Kong et al. [30] isolated monomeric aspartase variants by fusing together
domains of different subunits with randomized hexapeptide linkers. After expres-
sion in E. coli and growth selection on plates containing aspartate as sole nitrogen
source, variants that were active as monomers (mimicking a native dimer) were
obtained. Most variants had a reduced activity compared to wild type, with a kcat that
was about fivefold lower and a Km that was up to 50% higher compared to the wild-
type tetrameric enzyme.
Hybrid enzymes combining a-aspartyl dipeptidase (PepE) and E. coli L-aspartase
(AspA) activity in a single polypeptide chain have also been reported [31]. AspA
catalyzes addition of ammonia to fumaric acid to obtain L-aspartic acid, which can
serve as a substrate for PepE to synthesize N-terminal L-Asp dipeptides. This
cooperation can be improved by using a hybrid enzyme. After selection on plates
with L-aspartate an evolved variant was obtained. The fusion had a deletion of eleven
aa at the PepE terminus and a 24 aa deletion at the AspA N-terminus.
Asano et al. [32] have used directed evolution to modify the substrate specificity of
E. coli aspartase. The wild-type enzyme does not accept L-aspartic acid amide, but a
Lys327Asn mutant, which was found after error-prone PCR and screening for
product formation, could convert this substrate, albeit with rather low kcat and very
high Km. The lysine-327 is proposed to participate in a hydrogen bonding interaction
with the a-carboxylate of aspartate.
Such engineering studies may well have great potential in view of the flexibility of
the aspartase–fumarase scaffold, although at the same time the aspartase itself is
hardly promiscuous when it comes to acceptance of alternative substrates.
18.3
Other Aspartase/Fumarase Family Members: Adenylosuccinate Lyase,
Argininosuccinate Lyase, and EDDS Lyase
18.3.1
Adenylosuccinate Lyase
In purine biosynthesis, adenylosuccinate lyase (EC 4.3.2.2) catalyzes the conversion

of succinyladenosine monophosphate into adenosine monophosphate (AMP) and
fumarate (Scheme 18.2). The reaction is important for the transfer of an amino group
from aspartate to IMP to produce AMP. The sequence and structure of adenylo-
succinate lyases are similar to those of other members of the aspartase/fumarase
superfamily. In view of this, the catalytic mechanism is also expected to be similar to
that of aspartase, with CN cleavage again being the rate-limiting step [33]. The
reaction type suggests a catalytic base, an amine-site, and two carboxylate binding
sites. A conserved serine in the GSS motif was suggested to be the catalytic base and
His171 (corresponding to His188 in Bacillus aspartase mentioned above) could
protonate the nitrogen of the leaving group. In a structure with a closed active site
(PSB 3GZH), Glu308 makes a salt bridge with this histidine, contributing to its
position and basicity [34].
18.3 Other Aspartase/Fumarase Family Members j755
Little work has been done on the biotechnological application of this enzyme. The
substrate specificity is rather narrow, but alternative substrates that are accepted by
the enzyme from Leishmania donovani include 20 -dIMP, arabinosyl-IMP, 8-aza-IMP,
and allopurinol ribonucleotide, whereas activity is inhibited by GMP and some GMP
analogs [33]. These side activities are important for activation of antiprotozoal agent
allopurinol, which after conversion into allopurinol ribonucleoside 50 -monopho-
sphate (HPPR-MP) is aminated to form 4-aminopyrazolo[3,4-d]-pyrimidine ribonu-
cleotide (APPR-MP). The antiprotozoal activity of this AMP-analog is due to its
conversion to the triphosphate followed by incorporation into RNA.
18.3.2
Argininosuccinate Lyase
Another aspartase–fumarase superfamily member that has been studied is argini-

nosuccinate lyase (ASL) (EC 4.3.2.1). The enzyme plays a role in arginine metabolism
by catalyzing the reversible breakdown of argininosuccinate to arginine and fumarate
(Scheme 18.2), which is part of the urea cycle in liver. It is also important for arginine
synthesis in organisms that do not use the urea cycle for ammonia excretion. A close
homolog occurs as a lens protein (dII-crystallin).
The structures of bacterial, avian, and mammalian argininosuccinate lyases have
been solved. In two active sites of a recent structure of E. coli argininosuccinate lyase
(PDB 1TJ7), the C1, C2, and C3 loops, contributed by subunits a, c, and b, are well
ordered in a closed form due to binding of phosphate ions. This indicates how the
succinate group may be bound and thus provides insight in the catalytic mechanism
of aspartase superfamily enzymes, especially suggesting that the proximal Ser in the
SSAMP motif of the C3 loop is involved in proton abstraction during the first step of
the reaction [35].
Practical interest in argininosuccinate lyases stems mainly from the enhanced
serum levels of the enzyme that occur after liver damage. Furthermore, serious disease
is caused by ASL deficiency, which can be due to several mutations. Use of the enzyme
in cancer therapy to reduce arginine availability to arginine-dependent tumors has
been suggested [36]. Of several alternative substrates tested, only L-canavanine, which
has a CH2 in the side chain of the arginino moiety replaced by an oxygen atom, gave
similar conversion as L-arginine, indicating that the enzyme is highly specific [37].
18.3.3
EDDS Lyase
Further members of the aspartase–fumarase superfamily were detected in Agrobac-

terium tumefaciens BY6 and Ralstonia sp. strain SLRS7, which are bacteria growing on
EDDS ((S,S)-ethylenediamine-N,N0 -disuccinic acid, a structural isomer of
EDTA) [38]. A lyase from strain BY6 converts EDDS into fumarate and
N-(2-aminoethyl) aspartate. The reaction is reversible, with an equilibrium constant
of approximately 4.3 M. Both (S,S)-EDDS and (R,S)-EDDS can be converted by the
enzyme, but (R,R)-EDDS is not a substrate. The iminodisuccinate (IDS) lyase from
756j 18 CN Lyases Catalyzing Addition of Ammonia, Amines, and Amides to C¼C and C¼O Bonds
the Ralstonia strain is a similar tetrameric enzyme [38]. Both lyase genes have been
cloned and could be expressed in E. coli, and the sequence indicates that the
conserved fumarase superfamily regions (see above) are also present in these lyases.
Patents have been published describing the production of EDDS from fumarate
and ethylenediamine using microorganisms that produce such lyases, including a
Brevundimonas strain (Scheme 18.3 [39]). A process for preparing (S)-2-hydroxypro-
pylenediamine-N-succinate and (S,S)-2-hydroxypropylenediamine-N,N0 -disuccinate
was also described [40]. In general, diaminoalkylene-N,N0 -disuccinates are chelating
agents and intermediates for the production of pharmaceuticals and agrochemicals.
If maleic acid isomerase was included in the reaction mixtures containing the lyase,
the production process could start with maleic acid, which can be formed from the
corresponding anhydride. Directed evolution studies on ethylenediamine lyase
aimed at improving this type of conversion by enhancing the thermostability of the
enzyme were reported in the patent literature [41].
HOOC H
H H H2 N
HN COOH
HO
HO +
1 HN COOH
H H
HOOC H HN COOH
H H HOOC H
H COOH H
HN COOH HOOC H
1
HOOC
HN COOH +
2
H H H2N HOOC H
NH2
HOOC H H H
HN COOH
H COOH
HOOC H
H
Scheme 18.3 Amine elimination and amine addition reactions catalyzed by members of the
aspartase superfamily; (1) EDDS lyase and (2) IDS lyase.
18.4
Addition of Ammonia to Mesaconic Acid: L-Methylaspartase
18.4.1
General Properties
Methylaspartate ammonia lyases (EC 4.3.1.2) are involved in the anaerobic degra-
dation of glutamate in Clostridium tetanomorphum and several other (facultative)
anaerobic organisms that initiate glutamate degradation via a mutase reaction that
yields L-threo-3-methylaspartic acid [(2S,3R)-3-methylaspartatic acid]. This product is
18.4 Addition of Ammonia to Mesaconic Acid: L-Methylaspartase j757
converted by the lyase into mesaconic acid. The enzymes are homodimeric proteins
with subunits of 45 kDa the structure and sequence of which indicate that they are
members of the enolase superfamily [42]. Enzymes of this class catalyze a wide
diversity of reactions, including lyase-, (de)hydratase-, racemase-, and decarboxylase-
type conversions. Like other members of the enolase superfamily, methylaspartate
ammonia lyase requires Mg2 þ and K þ for activity, and an Mg2 þ ion located in the
active site is involved in substrate binding. Similar methylaspartate ammonia lyases
have been obtained from Clostridium tetanomorphum [43], Citrobacter amalonaticus
[44], E. coli [45], and other organisms (see references in the cited papers).
18.4.2
Structure and Mechanism
The structures of enolase family members are characterized by a large a/b-barrel that
is decorated with an a þ b cap domain. The catalytic residues, including three acidic
amino acids that bind the catalytically important Mg2 þ , are located at the C-terminal
ends of consecutive strands that form the a/b barrel and in a part of the cap domain
that covers the active site (PDB 1KKR, [46]). Early mechanistic studies on these
enzymes were flawed by the assumption that a dehydroalanine group was involved in
activity, which is probably incorrect.
The structure of the C. amalonaticus MAL suggests that (2S,3S)-3-methylaspartic
acid, which is the natural and preferred stereoisomer formed from glutamate, is
bound with its b-carboxylate interacting with the Mg2 þ ion, and that the reaction
starts with abstraction of the Cb proton by a basic group in the active site (PDB 1KCZ,
Scheme 18.4, [47]). The aci-carboxylate that is formed and stabilized by the metal
undergoes ammonia elimination, yielding the final product. This mechanism
suggests, besides binding sites for the b-carboxylate group, a base for proton
abstraction, an ammonia binding site that provides two or three hydrogen bond
donors (dependent on the protonation state of the leaving group), and possibly a
binding site for the a-carboxylate. Binding of the amino group probably involves
Gln73 and Gln172. The b-carboxylate is stabilized by the aforementioned Mg2 þ and
a conserved histidine (His194 in C. amalonaticus MAL), and also seems to interact
with Gln329. The base abstracting a proton could be a lysine (Lys331).
Mutation of Lys331 and Gln329 in this enzyme influences the diastereoselectivity,
indicating altered interactions between the substrate and the active site in these
mutants [48]. Since mutation of His194 abolishes activity with the non-preferred
Lys331 Lys331
H2N H2 N +
H
O- CO2- O- CO2-
Mg2+ O NH3+ Mg2+ O- NH3+
Scheme 18.4 Mechanism of the methylaspartase reaction [47, 48].

758
(2S,3R)-isomer of 3-methylaspartic acid, it was proposed that this residue is involved

in the deprotonation step with this substrate enantiomer.
18.4.3
Substrate Scope and Biocatalytic Application
Methylaspartate ammonia lyases have a much broader substrate range than aspar-
tases. These enzymes catalyze the stereoselective addition of ammonia and other
nitrogen compounds not only to mesaconic acid but also to several analogs.
Compounds that were used include chloro- and bromo-fumaric acid and, at a very
low rate, n- and i-propylfumaric acid, giving 3-substituted (2R,3R)-aspartates [49–51].
In all cases the same stereochemistry prevailed, with Si-face attack.
Alternative N-nucleophiles that can be used include hydrazine, methylamine,
hydroxylamine, methoxylamine, and to some extent ethylamine and dimethylamine,
and in the presence of excess N-nucleophile at high pH (9.0) good conversion of
mesaconate (40–90%) was achieved (Scheme 18.5 [52]). In all cases only a single
diastereomer with the expected stereo-configuration was obtained. When substrate
analogs were employed – ethyl-, isopropyl-, and propyl-fumaric acid were used with
hydrazine – the 2-hydrozino-3-alkyl substituted succinates were also found. Most of
these reactions occurred very slowly. Even though some alternative substrates are
accepted, the active site appears quite small, as indicated by the fact that a combi-
nation of a larger 3-substituent (instead of methyl) and a larger N-nucleophile
abolishes the reaction.
R'
H 2N COOH H COOH
HN COOH
H Me
H R
HOOC H HOOC R
HOOC H
R = Me, Et, iPr, Pr, Cl
R = NH2, Me, OH, OMe, Et
Scheme 18.5 Reactions catalyzed by methylaspartate ammonia lyase [52].
MAL has also been used to convert stereospecifically deuterated mesaconic acids
into the corresponding mono-, di-, and tri-deuterated 3-methylaspartic acids, which
could be used for mechanistic studies with glutamate mutases, for example, to
measure kinetic isotope effects [53].
18.5
Aromatic Amino Acid Ammonia Lyases
18.5.1
General Properties
Histidine ammonia lyase (HAL, EC 4.3.1.3) and related enzymes that act on aromatic
amino acids (phenylalanine ammonia lyase (PAL, EC 4.3.1.5), tyrosine ammonia
18.5 Aromatic Amino Acid Ammonia Lyases j759
Enz-B Enz-B Enz-BH+
- - O-
H H O H H O
O
O O
H2N H +
H2N H NH2
MIO N N
N
-O
-O N
O N N
Scheme 18.6 Proposed mechanism of a MIO-type ammonia lyase.
lyase (TAL, EC 4.3.1.23)) form a group of proteins that possess the unusual cofactor 4-
methylidene-imidazole-5-one (MIO, Scheme 18.6). The enzymes catalyze reversible
amine elimination from histidine, tyrosine, or phenylalanine, and some synthetic
derivatives thereof. Their main role is catabolic, but phenylalanine ammonia lyase is
also involved in producing cinnamic acid for lignin biosynthesis. For some time it
was thought that the active site of HAL from Pseudomonas putida, which is the first
MIO enzyme that was well studied, consisted of a dehydroalanine group, but the
X-ray structure showed that it contains a MIO cofactor, which is derived from an
internal Ala-Ser-Gly tripeptide [54]. The cofactor is formed autocatalytically via
cyclization and dehydration, and accordingly the tripeptide sequence is conserved
in all members of the aromatic ammonia lyases. The MIO cofactor has an electro-
philic ethylene group that was proposed to carry out a Friedel–Crafts like attack on the
aromatic ring, but according to more recent evidence it reacts with the substrates
amine group during amine elimination [3].
Another group of MIO enzymes is formed by phenylalanine and tyrosine ami-
nomutases. These enzymes shift the a-amino group to the b-position of their
substrates, and the reactions may proceed with high or low enantioselectivity. Since
these reactions do not occur with equilibrium exchange of ammonia with solvent, it
seems that the active site is more closed, which is supported by crystallographic
studies on tyrosine aminomutase [55].
18.5.2
Structure and Mechanism
The MIO enzymes are tetramers that are composed of a dimer of tail-to-head
dimers [3]. The N-terminal part of the subunits form a globular domain composed
of helices and short b-strands, whereas the C-terminal domain entirely consists of
a-helices. The aligned dipoles of helices create an electropositive region at the site
where the electrophilic MIO group is located and the residues flanking MIO are
hydrogen bonded to these helices. Plant and yeast PALs have an additional multi-
helix domain close to the C-terminus that may play a regulatory role [56].
Differences between the active site structures of MIO enzymes are mainly due to
variations in the disorder of two loop regions. The first loop is called the inner active
760
site loop, which is close to the N-terminus, donated by the same subunit as the MIO
group, and harbors the tyrosine supposed to act as a catalytic base. The outer active
site loop is located further down the sequence and consists of residues from the
C-terminal domain of a dyad-related subunit in the homotetramer [57]. The inner
loop contributes to formation of the complete active center and the outer loop further
covers and protects the active center. The two loops are proposed to be involved in
substrate binding and catalysis [57]. In early crystal structures of aromatic ammonia
lyases the two loops were disordered (P. putida HAL (PDB 1B8F) [54], Petroselinum
crispum PAL (PDB 1W27) [58], Rhodosporidium toruloides PAL (PDB 1T6J) [59],
Anabaena variabilis PAL (PDB 2NYN) [60], and Nostoc punctiforme PAL (PDB
2NYF) [60]). Structures of an ammonia lyase with ordered active site loops were
obtained with the Rhodobacter sphaeroides TAL (PDB 2O7F) [61] and a mutant A.
variabilis PAL in which two surface cysteines were replaced by serines (PDB
3CZO) [62]. A recently solved TAM structure also exhibits a closed active center
(PDB 2RJS) [55]. If and how this flexibility of the loops influences the catalytic
functions is not clear at present.
The substrate binding site is composed of residues from three subunits. Based on
the reaction type, one expects a proton abstracting group (a conserved tyrosine in the
inner active site loop, close to the N-terminus), one or two hydrogen bond donors that
interact with the amino group, a basic group that can accept a proton from the
substrates NH2 group, a carboxylate binding pocket, and a pocket for accommo-
dating the aromatic group. There is a conserved arginine from an adjacent subunit
that seems to bind the carboxylate of the substrate.
A likely reaction mechanism is an E1cb-like elimination, involving abstraction of a
proton from the b-carbon, which is activated by c-carboxyl group, and subsequent
removal of the amino group (Scheme 18.6). Most information about the mechanisms
of the MIO enzymes points to an electrophilic attack of the MIO group on the amino
group of the substrate. The evidence can be summarized as follows: a covalent adduct
has been observed by X-ray crystallography of the closely related MIO-enzyme
tyrosine aminomutase [55], computational studies indicate that the alternative
Friedel–Crafts mechanism is energetically unlikely, and the MIO cofactor is required
for the re-addition of amine in aminomutase reaction [63]. Earlier, the MIO group was
proposed to carry out a Friedel–Crafts alkylation of the aromatic ring, a suggestion
supported in several papers [1, 54, 64] that seemed attractive because delocalization of
negative charge to the aromatic ring is in accordance with the a-regioselectivity of a
1,4-Michael addition [78]. Nevertheless, most recent evidence points to the MIO-
amine adduct mechanism.
18.5.3
Distribution and Diversity
MIO-type ammonia lyases occur in bacteria, fungi, plants, and animals. Histidine
ammonia lyase catalyses the reversible elimination of ammonia from L-histidine to
produce urocanic acid. This is the first step in the metabolism of histidine in many
organisms. Therefore, HAL is widespread, occurring both in eukaryotes and in
prokaryotes. However, such pathways initiated with an ammonia lyase are not
prevalent in the catabolism of most amino acids; instead, aminotransferases yielding
a-keto acids appear more common for nitrogen utilization and biodegradation of
amino acids [1]. PALs are common in terrestrial plants where they catalyze the
formation of trans-cinnamic acid, which is a precursor of plant metabolites, including
lignin and flavonoids. A few PALs were discovered in bacteria, where they are
involved in the production of cinnamic acid or coumaric acid, which are intermedi-
ates for the biosynthesis of antibiotics or antifungal compounds [65]. In bacteria, TAL
is involved in the production of the chromophore of photoactive yellow protein and
caffeic acid occurs in the biosynthesis of secondary metabolites in certain
actinomycetes.
To date, only a few MIO-based aminomutases have been discovered. PAM from
Taxus chinensis plays a role in secondary metabolite formation by catalyzing a 2,3-
amine shift leading to (R)-b-phenylalanine, which is an important step in biosyn-
thesis of the anticancer drug taxol. Tyrosine aminomutase (TAM) from Streptomyces
globisporus and TAM from Cupriavidus crocatus have different kinetically preferred
products, (S)-b-tyrosine and (R)-b-tyrosine, which are subsequently incorporated
into antibiotics enedine C-1027 and cytostatic actin-targeting chondramides,
respectively [66, 67]. Analysis of gene clusters shows that similar MIO enzymes
may occur in other pathways leading to secondary metabolites [3]. However, since the
way to differentiate aminomutases from ammonia lyases is presently unclear, it is
difficult to predict the real function of the annotated MIO enzyme in genome
sequences.
18.5.4
Biocatalytic Relevance and Applications
As mentioned above, the natural role of most MIO-dependent ammonia lyases is the
elimination of ammonia from amino acids, and the resulting carboxylic acid is used
either as a growth substrate or as a phenylpropanoid building block in plants.
Phenylpropanoid synthesis has been reconstituted in yeast by expression of plant
phenylalanine ammonia lyase [68]. Under extreme conditions (>4 M NH4OH,
pH 10) these ammonia lyases also catalyze the reverse reaction, that is, addition
of ammonia to the double bond of a series of arylacrylates (Scheme 18.7), making
these enzymes attractive for the preparation of L-amino acids by biotransforma-
tion [23, 69–71]. In the 1980s, Genex Corp. developed a process for producing L-
phenylalanine from cinnamic acid and ammonia in a bioreactor with immobilized
enzyme. In a single pass, up to 90% of cinnamic acid in the feed was converted and
ammonia concentrations up to 7.85 M were used. Using yeast cells overexpressing
the enzyme, product concentrations up to 0.35 M were obtained [23].
Application of phenylalanine ammonia lyase for the preparation of other
L-phenylalanine derivatives has been described in several papers and patents
(Scheme 18.7). Conversions of 37–99% were obtained in purified PAL-catalyzed
addition of ammonia to pyridine-acrylic acids and cinnamic acids with various
halogen substituents on the ring (Table 18.1 [70]). PAL can also be used in whole-cell
762 j 18 CN Lyases Catalyzing Addition of Ammonia, Amines, and Amides to C¼C and C¼O Bonds
1 1
R O R O
2 PAL 2
R R
OH + NH3 OH
3 3 NH2
R R
R1-R3 = F, Br, NO2, CN, OH
O O
PAL
OH + NH3 OH
N N
NH2
1 1
R O R NH2 O
2 PAM 2
R R
OH OH
3 NH2 3
R R R1-R3 = Me, F, MeO
O NH2 O
PAM
X X
OH OH
NH2 X=O,S
1 1 1
R O R O R NH2 O
2 PAM 2 2
R R R
OH + NH3 OH + OH
3 3 NH2 3
R R R
R1-R3 = Me, F, Cl, Br

R3 = MeO, Et, Pr, iPr, NO2, CN
Scheme 18.7 Reactions catalyzed by MIO-type aromatic ammonia lyases and mutases.
preparations and in this form it has been shown to catalyze the conversion of
cinnamic acids with various substituents, such as halogens, nitro, cyano, and
hydroxyl [72]. For example, DSM uses phenylalanine ammonia lyase from Rhodotor-
ula glutinis produced in recombinant form in E. coli or an enzyme from the halophilic
bacterium Idiomarina loihiensis for converting 2-chloro or 2-bromocinnamic acid into
the corresponding ortho-substituted (R)-phenylalanine. This is an intermediate for
the preparation of enantioenriched (R)-indoline-2-carboxylic acid [73], which is used
for the synthesis of an antihypertensive agent.
Phenylalanine ammonia lyase has potential therapeutic application in the removal
of phenylalanine that builds up in the body in case of phenylketonuria. In a rodent
model, the enzyme could indeed lower levels of L-Phe in the blood, but application
requires a longer enzyme lifetime (reduced clearance rate), and reduced immuno-
genicity, which may be achieved with covalent modification of the enzyme with poly
(ethylene glycol) (PEG) [75]. PEG-modification of surface lysine, replacement of
chymotrypsin sensitive sites, and the use of a sol–gel matrix improved the intestinal
stability of Anabaena variabilis phenylalanine ammonia lyase [76]. Another potential
Table 18.1 Conversion of substituted cinnamic acids into L-phenylalanine derivatives.
Entry Ar Yield/conversion (%) Biocatalyst Reference
1 Ph 70 Whole cells [74]

2 Ph 52 Isolated enzyme [70]
3 2-F-Ph 50 Isolated enzyme [70]
6 2-Cl-Ph 37 Isolated enzyme [70]
9 2,2-di-F-Ph 88 Isolated enzyme [70]
10 3,3-di-F-Ph 69 Isolated enzyme [70]
11 Perfluorophenyl 51 Isolated enzyme [70]
12 2-Pyridyl 63 Isolated enzyme [69]
15 2,2-Di-Cl-Ph 24 Whole cells [72]
16 3-Br-Ph 80 Whole cells [72]
17 2-NO2-Ph 45 Whole cells [72]
18 3- NO2-Ph 45 Whole cells [72]
19 3-CN-Ph 80 Whole cells [72]
20 4-CN-Ph 20 Whole cells [72]
21 3-OH-Ph 80 Whole cells [72]
22 4-OH-Ph 30 Whole cells [72]
23 a-Naphthyl 10 Whole cells [72]
24 2-Thienyl 80 Whole cells [72]
medical application is in cancer treatment since the lyase may help to limit the supply
of phenylalanine to tumors.
Phenylalanine aminomutase (PAM) has been used in a process that involves
isomerization between aromatic a- and b-amino acids [77]. Another application of
PAM is the addition of ammonia to substituted cinnamic acids, which yields a
mixture of a- and b-amino acids (Scheme 18.7) [63, 78]. It has been shown that the
enzyme converts substrates with various substituents (alkyl, alkoxy, halogen, nitro,
cyano) and that the regioselectivity of the reaction depends on the electronic
properties of the substituents: substrates with electron-donating groups are prefer-
ably converted into b-amino acids, while cinnamic acids with electron-withdrawing
groups yield mainly the a-isomers.
18.5.5
Engineering Studies
Attempts to engineer properties of MIO ammonia lyase have aimed at improving

PAL for application in phenylketonuria treatment (see above) and at modification of
764
the substrate specificity. The use of PAL as an enzyme therapy is limited by

immunogenicity and proteolytic sensitivity. Wang et. al. [78] have constructed several
mutants of R. toruloides PAL to remove the protease sensitive sites, modify the
PEGylation sites, or minimize the size of functional PAL. Most mutants were less
active, with the exception of an Arg91Lys mutant, which showed a 30% increase in
specific activity compared to wild-type PAL. Residue Arg91 is proposed to be involved
in stabilization of the inner loop. The same group also reported an engineering study
of A. variabilis PAL. Wild-type A. variabilis PAL is unstable and easily aggregates. By
mutating two surface cysteine residues (Cys503 and Cys565) to serine, protein
aggregation is reduced and the catalytic behavior is not affected [62].
The aromatic ammonia lyases are specific for their natural substrate: histidine,
phenylalanine, or tyrosine. Two groups reported that the substrate selectivity of the
lyases is determined by a single semi-conserved His/Phe residue [61, 80]. By
mutating this His89 to Phe in R. sphaeroides TAL, the selectivity of the enzyme
switched to that of a phenylalanine ammonia lyase. Similarly, replacement of the
corresponding Phe144 with His in A. thaliana PAL resulted in an 18-fold increase
in catalytic efficiency with tyrosine and an 80-fold reduction of the kcat/Km for
phenylalanine [80]. The R. sphaeroides TAL has a closed structure, which indicates
that the histidine is responsible for forming a hydrogen bond with the phenolic
hydroxyl group of tyrosine [61].
Phenylalanine aminomutase exhibits high enantioselectivity towards b-amino
acids, whereas TAM has only low enantioselectivity with b-tyrosine [3]. The
molecular basis for the control of product stereoselectivity in the MIO aminomu-
tases has not been elucidated. Krug and M€ uller have identified one semi-conserved
residue that influences the enantioselectivity of TAM. Replacement of Glu399 by
Lys increased the enantiomeric excess (e.e.) of the obtained (R)-b-tyrosine from
69% to 97%, but the activity of this mutant is reduced when compared to wild-type
TAM [67].
18.5.6
b-Alanyl CoA Ammonia Lyase
Anaerobic bacterial biodegradation of uracil and L-aspartate can proceed via

b-alanine, which becomes conjugated to CoA, after which it is deaminated in a
lyase-type reaction that produces acryloyl CoA (Scheme 18.8). A highly active
b-alanyl-CoA ammonia lyase (EC 4.3.1.6) is induced to very high level in Clostridium
propionicum during growth on b-alanine, and it was purified from this same
organism [81]. The enzyme is a homopentamer composed of 16 kDa subunits. Two
homologous genes encoding these enzymes in C. propionicum were cloned and
expressed in E. coli. Homologous sequences are also found in the genomes of various
other anaerobes, for example, bacteria that anaerobically ferment lysine and possess
b-aminobutyryl-CoA ammonia lyase [82]. The equilibrium of the b-alanyl-CoA
ammonia lyase reaction is in the direction of the amine adduct, and amine
elimination only proceeds rapidly in the presence of a dehydrogenase that converts
acryloyl-CoA into propionyl-CoA. It was suggested that the broad substrate range of
O β-alanyl CoA lyase
SCoA + NH3
NH2 SCoA
O
O β-alanyl CoA lyase O
+ NH3
NH2 SCoA SCoA
O β-alanyl CoA lyase SCoA

+ NH3
NH2 SCoA
O
Scheme 18.8 Reactions catalyzed by b-alanyl CoA ammonia lyase.
the enzyme with respect to the enoyl-CoA derivatives (crotonyl-CoA, methacryloyl-

CoA are accepted) can be used for production of amino compounds such as
3-aminobutyrate or 3-aminoisobutyrate [81]. Another biotechnological application
would be the production of acryloyl-CoA from alanine in a pathway leading to
3-hydroxypyruvate, which has been considered as a bio-based building block for
polymers [83]. Besides ammonia, methylamine can also be accepted by b-amino-
butyryl-CoA ammonia lyase, but detailed kinetic data have not been reported.
As no structures of (putative) b-alanyl-CoA lyases or sequence related enzymes are
known, mechanistic details are lacking. There is slight sequence similarity with a
broad-specificity acyl-CoA thioesterase from Haemophilus influenzae (PDB 3BJK). In
fact, low sequence similarity can be detected with several members of the HotDog
protein family, which encompasses enoyl-CoA hydratases, b-hydroxyacyl-ACP dehy-
dratases, and acyl-CoA thioesterases [84], but apart from the notion that these
proteins possess a tight-binding CoA site at the dimer interface the mechanistic
implications of these similarities are not clear.
18.5.7
Serine Dehydratase, Threonine Dehydratase, and Other Class IIPLP-Dependent
Enzymes
Several phylogenetically related pyridoxal phosphate (PLP)-dependent enzymes

catalyze CN bond formation on carbon atoms flanking carbonyl (C¼O) groups,
or lyase reactions in which CN bonds are cleaved and C¼O bonds are formed
[4, 85]. The class II PLP enzymes that allow such reactions include catabolic L- and
D-serine dehydratase and L-threonine dehydratase (Scheme 18.9). The conversions
result in a net elimination of ammonia, producing a ketoacid, and therefore the
enzymes are often called deaminases. Mechanistically they may be dehydratases
since the amino group of amino acids forms an internal aldimine in PLP enzymes
and, probably, is not directly released. Besides various catabolic enzymes, this
family also includes biosynthetic serine dehydratases and threonine dehydratases
(see also Chapter 11).
766
L-Serine
OH O ammonia lyase HO O
+ NH3
OH
O
NH2
D-Serine
OH O ammonia lyase HO O
+ NH3
OH
O
NH2
threonine
O deaminase HO O
+ NH3
HO OH
O
NH2
threo-3-hydroxy-
HO O L-aspartate
O ammonia lyase HO O
O
+ NH3
HO OH
O OH
NH2
diaminopropionate
NH2 ammonia lyase O
H2N OH OH
O O
Scheme 18.9 Reactions catalyzed by class II PLP-dependent ammonia lyases.
18.5.8
L-Serine Dehydratase/Deaminase
Catabolic L-serine dehydratase (SDH, EC 4.3.1.17) is a group II PLP enzyme that

catalyzes the elimination of ammonia from L-serine and L-threonine to yield pyruvate
and 2-oxobutyrate, respectively (Scheme 18.9 [86]). It occurs in mammals mainly in
the liver and has a catabolic role or a function in gluconeogenesis. Structures have
been solved for several L-serine dehydratases, including the homodimeric rat (PDB
1PWE, [87]) and human (PDB 1P5J, [88]) enzymes. The conserved lysine, which
forms the internal aldimine with the PLP cofactor, is present in the conserved
sequence motif Ser-Xaa-Lys-Ile-Arg-Gly. The amine elimination reactions are mech-
anistically considered to be a water elimination from the b-position of the substrate,
since dehydration of, for example, serine will produce an enamine that is unstable
and will release ammonia due to hydrolysis (Scheme 18.10).
Biotechnological applications have not been reported, although variants with
interesting substrate scope may exist, such as the phenylserine dehydratase from
Ralstonia pickettii PS22 [89].
Lys Lys
O R H+
Lys H
O C OH NH2 NH2
O O R
NH2 H R O O
N O H CH H
H O
H CH O O
N N H
O O O H H
H
P O O O O O O O
N O O P O
N O O
N
(1) (2) (3)
Lys H2O NH4+

O R
O R
CH
O CH2
O
N NH3
H CH O
O O O
P O
N O
(4)
Scheme 18.10 Proposed mechanism for serine (R ¼ H) and threonine (R ¼ CH3) dehydratase.
18.5.9
D-Serine Dehydratase/Deaminase
Another PLP-dependent enzyme that catalyzes elimination of ammonia is D-serine

ammonia-lyase (EC 4.3.1.18). The monomeric E. coli enzyme participates in D-serine
catabolism and has been the subject of various biochemical and kinetic studies. The
structure of the enzyme from Burkholderia xenovorans LB400 was recently solved
(PDB 3GWQ, unpublished). The properties of the enzyme are similar to those of
other class II PLP enzymes.
Biocatalytic applications have, to our knowledge, not yet been reported. An
exception is the use a D-serine hydratase in an enzymatic assay, but in this a case
a novel type of PLP- and zinc-dependent D-serine dehydratase was used that has a
higher enantioselectivity and was obtained from Saccharomyces cerevisiae [90].
18.5.10
L-Threonine Dehydratase/Deaminase
PLP-dependent threonine dehydratases (EC 4.3.1.19) catalyze the deamination of

L-threonine to produce 2-oxobutanoate. This is the first committed step of the
isoleucine biosynthesis pathway, and also an initial step in one of the possible routes
for L-threonine catabolism. There are catabolic and biosynthetic versions of this
enzyme, and the substrate range overlaps with that of biosynthetic L-serine
hydratases.
768
The activity of the tetrameric E. coli biosynthetic L-threonine dehydratases is

regulated allosterically. It is subject to feedback inhibition by isoleucine and
activated by valine. The X-ray structure of the E. coli enzyme (PDB 1TDJ [91])
showed that it is a homotetramer with each subunit being composed of a regulatory
domain and a catalytic domain. The regulatory domains can bind valine and
isoleucine.
The catabolic enzymes, which catalyze the first step in the degradation of
L-threonine to propionate, are usually not subject to feedback regulation and
therefore would be more suitable for biocatalytic production of amino acids.
Structures have been solved for the dimeric catabolic enzyme from Salmonella
typhimurium (PDB 2GN0 [92]), and for the enzymes from Solanum lycopersicum
(3IAU, unpublished) and Thermus thermophilus (PDB 1VE5, unpublished).
18.5.11
Threo-3-Hydroxy-L-Aspartate Ammonia-Lyase
The enzyme threo-3-hydroxy-L-aspartate ammonia-lyase (systematic name threo-

3-hydroxy-L-aspartate ammonia-lyase oxaloacetate-forming, EC 4.3.1.16) has been
obtained from Pseudomonas sp. [93, 94] and from Saccharomyces cerevisiae [95]. The
function of this enzyme may be catabolic since the substrate occurs in antibiotics
such as syringomycins and cormycin A.
18.5.12
Diaminopropionate Ammonia-Lyase
Another bacterial deaminase of the class II of PLP-dependent enzymes is diami-

nopropionate ammonia-lyase (DAP ammonia lyase, EC 4.3.1.15), which catalyzes
an a,b-elimination reaction with L- and D-a,b-diaminopropionate, producing pyru-
vate and ammonia. The substrate of the enzyme is an intermediate in the formation
of the Lathyrus sativus (a grain legume) neurotoxins 3-oxalyl and 2,3-dioxalyl
DAP [96, 97]. The enzymes were isolated and the genes were cloned from E. coli
and S. typhimurium, which revealed that the sequence is homologous to that of
L-threonine dehydratase from Clostridium tetani, but much less so to the E. coli
L-threonine dehydratase. The enzymes were also active with D-serine, but with
L-serine activity was very low. As with other enzymes of the class II PLP enzymes,
biotechnological application, such as use in the synthesis of amino acids, seems to
be hardly explored.
18.5.13
D-Glucosaminate Dehydratase
A glucosaminic acid deaminase has been discovered in Pseudomonas fluorescence [98].

It replaces an amino group with oxygen (Scheme 18.11). The enzyme also has
aldolase activity, and contains both PLP and Mg2 þ [99]. Biochemical details are
scarce, but the protein has been used for determination of D-glucosaminate [100].
COOH 2-amino-2-deoxy- COOH
H NH2 D-gluconate O
ammonia lyase
HO H H H
H OH H OH
H OH H OH
CH2OH CH2OH
Scheme 18.11 Reaction catalyzed by the PLP-enzyme 2-amino-2-deoxy-D-gluconate ammonia-

lyase from Pseudomonas fluorescence.
18.5.14
Fe-S-Dependent Serine Hydratases
Catabolic bacterial serine hydratases in (facultatively) anaerobic bacteria possess an

iron-sulfur cluster with serine hydratase activity [101]. They are highly specific and
unstable, and contain an iron-sulfur cluster instead of PLP in the active site. The
enzymes have a catabolic function, for example, in the fermentative (anaerobic)
breakdown of L-serine in Peptostreptococcus asaccharolyticus and Clostridium
propionicum. Similar iron-sulfur cluster serine dehydratases have been reported in
E. coli and Campylobacter jejuni [102], and a structure is available of the homologous
enzyme from Legionella pneumophila (PDB 2IQQ, unpublished). An iron-sulfur
cluster containing enantioselective L-serine hydratase was isolated from the marine
bacterium Paracoccus seriniphilus and used in constructing an L-selective elec-
trode [103], but we are not aware of other biotechnological applications.
18.5.15
Miscellaneous Lyases Adding Amines to C¼C Bonds
3-Ketovalidoxylamine A CN lyase catalyzes a lyase reaction that cleaves a CN bond
in validoxylamine A (Scheme 18.12). Such enzymes (EC 4.3.3.1) were purified from
Flavobacterium saccharophilum [104] and Stenotrophomonas maltrophilia [105]; both
are monomers of about 35 kDa and convert 4-nitrophenyl-3-ketovalidamine, which
can be formed by biocatalytic oxidation and hydrolysis from validamycin A. The latter
is an aminoglucoside antibiotic produced by Streptomyces hygroscopicus; it is effective
against the plant pathogenic fungus Rhizoctonia solani. The product of the lyase
reaction is valienamine, which is a component of acarbose and validamycin family of
antibiotics. Structural or mechanistic properties of the enzyme have not been
reported.
O 3-ketovalidoxylamine A O NO2
HO OH NO2 C-N lyase
HO O +
HO H 2N
N HO
H
Scheme 18.12 Conversion of validoxylamine A by 3-ketovalidoxylamine A CN lyase.

770 j 18 CN Lyases Catalyzing Addition of Ammonia, Amines, and Amides to C¼C and C¼O Bonds
O strictosidine
NH2 synthase N NH
+ OGlc H H
OGlc
N O
H MeOOC O
MeOOC
Scheme 18.13 Reaction catalyzed by strictosidine synthase.
Strictosidine synthase (EC 4.3.3.2, Scheme 18.13), sometimes called a Pictet–

Spenglerase, catalyzes the Pictet–Spengler reaction of tryptamine with secologanin,
yielding the glucoalkaloid 3a-(S)-strictosidine. This is the first committed step in the
synthesis of the monoterpenoid indole alkaloids in the Indian medicinal plant
Rauvolfia serpentina. Such alkaloids have a wide range of bioactivities, including
antitumor, antimalaria, and antiarrhythmic effects. The crystal structure of the
enzyme was solved [106]. It is a six-bladed beta propeller, and structures with
substrate bound provided insight in the mechanism of the enzyme. The
amino group of tryptamine forms a Schiff base with the aldehyde group of
secologanin, after which the C2 of tryptamine is added to this Schiff base, and a
glutamate residue was proposed to be involved in amine deprotonation. Various
homologs of strictosidine synthase have been discovered by sequencing and com-
parison of structures.
An engineered variant was made by modifying the so-called indole sandwich,
which is a site where the tryptamine is stacked between two aromatic groups. When a
valine that lines this site was replaced by an alanine, making more space, improved
conversion of 5- and 6-substituted tryptamines was obtained, and especially
5-hydroxytryptamine was rapidly converted [107]. Such modified variants are
expected to broaden the range of accessible strictosidine variants, which can be
used for further chemoenzymatic modification.
Deacetylisoipecoside dopamine-lyase (EC 4.3.3.3) and deacetylipecoside synthase
(EC 4.3.3.4) catalyze the condensation of dopamine with secologanin to produce
deacetylisoipecoside, an isoquinoline alkaloid of the plant Alangium lamarckii
(Scheme 18.14). No structures or sequences have been reported. The enzymes
catalyze the same reaction, but give opposite product enantiomers.
NH2 HO HO
O lyase
NH + NH
OGlc HO HO
+ H H
OGlc OGlc
HO O
MeOOC O O
OH MeOOC MeOOC
Scheme 18.14 Reactions catalyzed by deacetylisoipecoside dopamine-lyase (left-hand product)

and deacetylipecoside synthase (right-hand product).
18.6 Conclusions and Outlook j771
ethanolamine
ammonia lyase
NH2 + NH3
OH O
Scheme 18.15 Reaction of ethanolamine ammonia lyase from Clostridium sp. or E. coli.
Ethanolamine ammonia-lyase (EC 4.3.1.7, Scheme 18.15) is a cobalamin (vitamin

B12)-dependent enzyme that has been purified from E. coli and Clostridium sp. It is a
dodecameric protein of composition a6b6 with subunits of 55 kDa and 35 kDa. In
E. coli, it occurs in a microcompartment surrounded by shell proteins that form a
semipermeable membrane [108].
Structures of the complete enzyme are lacking, but detailed mechanistic studies
have been performed, which led to the conclusion that the enzyme uses a radical
mechanism. During conversion of ethanolamine into acetaldehyde the hydrogen is
transferred from the carbinol carbon to the amino carbon without exchange with
water. The enzyme also deaminates (S)- and (R)-2-aminopropanol and to propio-
naldehyde, and using isotopically labeled substrates details about stereochemical
events in the active site were unraveled [109].
18.6
Conclusions and Outlook
Amino-lyases that add amines and/or ammonia to double bonds are a phylogenet-
ically diverse group of enzymes that can be classified in different mechanistic and
structural classes. Biocatalytic applications are somewhat scarce at present, mainly
because the thermodynamic equilibrium of the reactions is strongly in the direction
of cleavage, making it necessary to use extremely high concentrations of ammonia for
applying the enzymes in the synthesis of amine-substituted products. Furthermore,
the well-studied enzymes aspartate ammonia lyase, methylaspartate ammonia lyase,
and phenylalanine ammonia lyase are quite restricted concerning the range of amine
nucleophiles that is accepted. Nevertheless, full-scale application of aspartase is
a classical example of biocatalytic amino acid production in industry, and also the
MIO-dependent aromatic ammonia lyase is used in industrial processes.
Especially with aspartases, one would expect possibilities to discover or engineer
mutants with altered nucleophile range because other members of the fumarase/
aspartase superfamily do have such activity (e.g., adenylosuccinate lyase, EDDS
lyase, fumarase). With the MIO enzymes, mechanistic reasons seem to limit the
substrate range, and it has been reported that these enzymes are reluctant to have
their substrate range modified by protein engineering. A large diversity of reactions
is also encountered in the family of PLP-dependent ammonia lyases. We expect that
further biotechnological applications will emerge as a result of the growing
availability of a diversity of ammonia lyases and the increasing possibilities for
obtaining variants with new activities or enhanced process performance by protein
engineering.
772
References
1 Poppe, L. and R
etey, J. (2005) 11 Shi, W., Dunbar, J., Jayasekera, M.M.,
Friedel–Crafts-type mechanism for the Viola, R.E., and Farber, G.K. (1997)
enzymatic elimination of ammonia from The structure of L-aspartate ammonia-
histidine and phenylalanine. lyase from Escherichia coli. Biochemistry,
Angew. Chem. Int. Ed., 44 (24), 36 (30), 9136–9144.
3668–3688. 12 Fujii, T., Sakai, H., Kawata, Y., and Hata,
2 MacDonald, M.J. and DCunha, G.B. Y. (2003) Crystal structure of
(2007) A modern view of phenylalanine thermostable aspartase from Bacillus sp.
ammonia lyase. Biochem. Cell. Biol., YM55-1: structure-based exploration of
85 (3), 273–282. functional sites in the aspartase family.
3 Cooke, H.A., Christianson, C.V., and J. Mol. Biol., 328 (3), 635–654.
Bruner, S.D. (2009) Structure and 13 Puthan Veetil, V., Raj, H., Quax, W.J.,
chemistry of 4-methylideneimidazole-5- Janssen, D.B., and Poelarends, G.J.
one containing enzymes. Curr. Opin. (2009) Site-directed mutagenesis,
Chem. Biol., 13 (4), 453–461. kinetic and inhibition studies of
4 Asano, Y., Kato, Y., Levy, C., Baker, P., and aspartate ammonia lyase from
Rice, D. (2004) Structure and Bacillus sp. YM55-1. FEBS J., 276 (11),
function of amino acid ammonia-lyases. 2994–3007.
Biocatal. Biotransform., 22 (2), 14 Takagi, J.S., Tokushige, M., and
131–138. Shimura, Y. (1986) Cloning and
5 Mizobata, T. and Kawata, Y. (2007) nucleotide sequence of the aspartase
Aspartases, in Industrial Enzymes - gene of Pseudomonas fluorescens.
Structure, Functions and Applications J. Biochem., 100 (3), 697–705.
(eds J. Polaina and A.P. MacCabe), 15 Sun, D.X. and Setlow, P. (1991)
Springer, Dordrecht, pp. 549–565. Cloning, nucleotide sequence, and
6 Poechlauer, P., Braune, S., de Vries, A., expression of the Bacillus subtilis ans
and May, O. (2010) Sustainable route operon, which codes for L-asparaginase
design for pharmaceuticals. Why, how and L-aspartase. J. Bacteriol., 173 (12),
and when. Chim. Oggi, 28 (4), 14–17. 3831–3845.
7 Takagi, J.S., Ida, N., Tokushige, M., 16 Kawata, Y., Tamura, K., Kawamura, M.,
Sakamoto, H., and Shimura, Y. (1985) Ikei, K., Mizobata, T., Nagai, J., Fujita, M.,
Cloning and nucleotide sequence of the Yano, S., Tokushige, M., and Yumoto, N.
aspartase gene of Escherichia coli W. (2000) Cloning and over-expression of
Nucleic Acids Res., 13 (6), 2063–2074. thermostable Bacillus sp. YM55-1
8 Karsten, W.E., Gates, R.B., and Viola, R.E. aspartase and site-directed mutagenesis
(1986) Kinetic studies of L-aspartase for probing a catalytic residue.
from Escherichia coli: substrate activation. Eur. J. Biochem., 267 (6),
Biochemistry, 25 (6), 1299–1303. 1847–1857.
9 Kawata, Y., Tamura, K., Yano, S., 17 Kazuoka, T., Masuda, Y., Oikawa, T., and
Mizobata, T., Nagai, J., Esaki, N., Soda, K., Soda, K. (2003) Thermostable aspartase
Tokushige, M., and Yumoto, N. (1999) from a marine psychrophile, Cytophaga
Purification and characterization of sp. KUC-1: molecular characterization
thermostable aspartase from Bacillus sp and primary structure. J. Biochem.,
YM55-1. Arch. Biochem. Biophys., 366 (1), 133 (1), 51–58.
40–46. 18 Sato, T., Nishida, Y., Tosa, T., and
10 Nuiry, I.I., Hermes, J.D., Weiss, P.M., Chibata, I. (1979) Immobilization of
Chen, C.Y., and Cook, P.F. (1984) Kinetic Escherichia coli cells containing aspartase
mechanism and location of rate- activity with k-carrageenan: enzymic
determining steps for aspartase from properties and application for l-aspartic
Hafnia alvei. Biochemistry, 23 (22), acid production. Biochim. Biophys. Acta,
5168–5175. 570 (1), 179–186.
References j773
19 Tokushige, M. (1985) Aspartate ammonia lyase. Chemistry, 14 (32),
ammonia-lyase. Meth. Enzymol., 113, 10094–10100.
618–627. 29 Wang, L.J., Kong, X.D., Zhang, H.Y.,
20 Chibata, I., Tosa, T., and Sato, T. (1974) Wang, X.P., and Zhang, J. (2000)
Immobilized aspartase-containing Enhancement of the activity of L-
microbial cells: preparation and aspartase from Escherichia coli W by
enzymatic properties. Appl. Microbiol., directed evolution. Biochem. Biophys. Res.
27 (5), 878–885. Commun., 276 (1), 346–349.
21 Tosa, T., Sato, T., Nishida, Y., and 30 Kong, X., Li, Z., Gou, X., Zhu, S.,
Chibata, I. (1977) Reason for higher Zhang, H., Wang, X., and Zhang, J.
stability of aspartase activity of (2002) A monomeric L-aspartase
immobilized Escherichia coli cells. obtained by in vitro selection. J. Biol.
Biochim. Biophys. Acta, 483 (1), Chem., 277 (27), 24289–24293.
193–202. 31 Sheng, Y., Li, S., Gou, X., Kong, X.,
22 Nishida, Y., Sato, T., Tosa, T., and Wang, X., Sun, Y., and Zhang, J. (2005)
Chibata, I. (1979) Immobilization of The hybrid enzymes from alpha-aspartyl
Escherichia coli cells having aspartase dipeptidase and L-aspartase.
activity with carrageenan and locust bean Biochem. Biophys. Res. Commun.,
gum. Enzyme Microb. Technol., 1 (2), 331 (1), 107–112.
95–99. 32 Asano, Y., Kira, I., and Yokozeki, K. (2005)
23 Hamilton, B.K., Hsiao, H.Y., Alteration of substrate specificity of
Swann, W.E., Anderson, D.M., and aspartase by directed evolution. Biomol.
Delente, J.J. (1985) Manufacture of L- Eng., 22 (1–3), 95–101.
amino acids with bioreactors. Trends 33 Bulusu, V., Srinivasan, B.,
Biotechnol., 3 (3), 64–68. Bopanna, M.P., and Balaram, H. (2009)
24 Chao, Y.P., Lo, T.E., and Luo, N.S. (2000) Elucidation of the substrate specificity,
Selective production of L-aspartic acid kinetic and catalytic mechanism of
and L-phenylalanine by coupling adenylosuccinate lyase from Plasmodium
reactions of aspartase and falciparum. Biochim. Biophys. Acta,
aminotransferase in Escherichia coli. 1794 (4), 642–654.
Enzyme Microb. Technol., 27 (1–2), 34 Kozlov, G., Nguyen, L., Pearsall, J., and
19–25. Gehring, K. (2009) The structure of
25 Xu, H., Wei, P., Zhou, H., Fan, W., and phosphate-bound Escherichia coli
Ouyang, P. (2003) Efficient production of adenylosuccinate lyase identifies His171
L-phenylalanine catalyzed by a coupled as a catalytic acid. Acta Crystallogr., Sect. F,
enzymatic system of transaminase and 65 (9), 857–861.
aspartase. Enzyme Microb. Technol., 33 (5), 35 Bhaumik, P., Koski, M.K., Bergmann, U.,
537–543. and Wierenga, R.K. (2004) Structure
26 Fatibello-Filho, O., Suleiman, A.A., determination and refinement at 2.44 A
Guilbault, G.G., and Lubrano, G.J. (1988) resolution of argininosuccinate lyase
Bienzymatic electrode for the from Escherichia coli. Acta. Crystallogr.,
determination of aspartame in dietary Sect. D, 60 (11), 1964–1970.
products. Anal. Chem., 60 (21), 36 Lam, T.L., Wong, G.K., Chong, H.C.,
2397–2399. Cheng, P.N., Choi, S.C., Chow, T.L.,
27 Wang, Z.W., Chen, Y., and Chao, Y.P. Kwok, S.Y., Poon, R.T., Wheatley, D.N.,
(2006) Enhancement of recombinant Lo, W.H., and Leung, Y.C. (2009)
protein production in Escherichia coli by Recombinant human arginase inhibits
coproduction of aspartase. J. Biotechnol., proliferation of human hepatocellular
124 (2), 403–411. carcinoma by inducing cell cycle arrest.
28 Weiner, B., Poelarends, G.J., Cancer Lett., 277 (1), 91–100.
Janssen, D.B., and Feringa, B.L. (2008) 37 Wu, C.Y., Lee, H.J., Wu, S.H., Chen, S.T.,
Biocatalytic enantioselective synthesis of Chiou, S.H., and Chang, G.G. (1998)
N-substituted aspartic acids by aspartate Chemical mechanism of the endogenous
774
argininosuccinate lyase activity of duck 47 Asuncion, M., Blankenfeldt, W., Barlow,

lens delta2-crystallin. Biochem J., 333 (2), J.N., Gani, D., and Naismith, J.H. (2002)
327–334. The structure of 3-methylaspartase from
38 B€auerle, B., Cokesa, Z., Hofmann, S., and Clostridium tetanomorphum functions via
Rieger, P.G. (2006) Sequencing and the common enolase chemical step.
heterologous expression of an J. Biol. Chem., 277 (10), 8306–8311.
epimerase and two lyases from 48 Raj, H., Weiner, B., Veetil, V.P.,
iminodisuccinate-degrading bacteria. Reis, C.R., Quax, W.J., Janssen, D.B.,
Appl. Environ. Microbiol., 72 (4), Feringa, B.L., and Poelarends, G.J. (2009)
2824–2828. Alteration of the diastereoselectivity of
39 Wataru, M. (1998) Protein having 3-methylaspartate ammonia lyase by
ethylenediamine-N,N0 -disuccinic acid: using structure-based mutagenesis.
ethylenediamine lyase activity and gene ChemBioChem, 10 (13), 2236–2245.
encoding the same. European Patent 49 Akhtar, M., Botting, N.P., Cohen, M.A.,
EP0845536. and Gani, D. (1987) Enantiospecific
40 Kaneko, M. (2003) Processes for synthesis of 3-substituted aspartic acids
producing S,S-2-hydroxypropylene via enzymic amination of substituted
diamine-N,N0 -disuccinic acid. US Patent fumaric acids. Tetrahedron, 43 (24),
6503739. 5899–5908.
41 Mizunashi, W. and Nakamura, T. (2005) 50 Akhtar, M., Cohen, M.A., and Gani, D.
Modified ethylenediamine-N,N0 (1987) Stereochemical course of the
disuccinic acid: ethylenediamine lyase. enzymic amination of chloro- and bromo
European Patent EP1757685. fumaric acid by 3-methylaspartate
42 Gerlt, J.A., Babbitt, P.C., and Rayment, I. ammonia-lyase. Tetrahedron Lett., 28 (21),
(2005) Divergent evolution in the enolase 2413–2416.
superfamily: the interplay of mechanism 51 Botting, N.P., Akhtar, M., Cohen, M., and
and specificity. Arch. Biochem. Biophys., Gani, D. (1988) Substrate-specificity of
433 (1), 59–70. the 3-methylaspartate ammonia-lyase
43 Goda, S.K., Minton, N.P., Botting, N.P., reaction: observation of differential
and Gani, D. (1992) Cloning, sequencing, relative reaction-rates for substrate
and expression in Escherichia coli of the product pairs. Biochemistry, 27 (8),
Clostridium tetanomorphum gene 2953–2955.
encoding beta-methylaspartase and 52 Gulzar, M.S., Akhtar, M., and Gani, D.
characterization of the recombinant (1997) Preparation of N-substituted
protein. Biochemistry, 31 (44), aspartic acids via enantiospecific
10747–10756. conjugate addition of N-nucleophiles to
44 Kato, Y. and Asano, Y. (1988) Cloning, fumaric acids using methylaspartase:
nucleotide sequencing, and expression of synthetic utility and mechanistic
the 3-methylaspartate ammonia-lyase implications. J. Chem. Soc. Perkin Trans.
gene from Citrobacter amalonaticus strain 1, 649–655.
YG-1002. Appl. Microbiol. Biotechnol., 53 Lee, H.Y., Yoon, M., and Marsh, E.N.
50 (4), 468–474. (2007) Synthesis of mono- and
45 Kato, Y. and Asano, Y. (1995) di-deuterated (2S, 3S)-3-methylaspartic
3-Methylaspartate ammonia-lyase from a acids to facilitate measurement of
facultative anaerobe, strain YG-1002. intrinsic kinetic isotope effects in
Appl. Microbiol. Biotechnol., 43 (5), enzymes. Tetrahedron, 63 (22),
901–907. 4663–4668.
46 Levy, C.W., Buckley, P.A., Sedelnikova, S., 54 Schwede, T.F., Retey, J., and Schulz, G.E.
Kato, Y., Asano, Y., Rice, D.W., and (1999) Crystal structure of histidine
Baker, P.J. (2002) Insights into enzyme ammonia-lyase revealing a novel
evolution revealed by the structure of polypeptide modification as the catalytic
methylaspartate ammonia lyase. electrophile. Biochemistry, 38 (17),
Structure, 10 (1), 105–113. 5355–5361.
References j775
55 Christianson, C.V., Montavon, T.J., Structural and biochemical
Van Lanen, S.G., Shen, B., and characterization of the therapeutic
Bruner, S.D. (2007) The structure of Anabaena variabilis phenylalanine
L-tyrosine 2,3-aminomutase from the ammonia lyase. J. Mol. Biol., 380 (4),
C-1027 enediyne antitumor antibiotic 623–635.
biosynthetic pathway. Biochemistry, 63 Wu, B., Szymanski, W., Wietzes, P., de
46 (24), 7205–7214. Wildeman, S., Poelarends, G.J.,
56 Pilbak, S., Tomin, A., Retey, J., and Feringa, B.L., and Janssen, D.B. (2009)
Poppe, L. (2006) The essential tyrosine- Enzymatic synthesis of enantiopure
containing loop conformation and the alpha- and beta-amino acids by
role of the C-terminal multi-helix region phenylalanine aminomutase-catalysed
in eukaryotic phenylalanine ammonia- amination of cinnamic acid derivatives.
lyases. FEBS J., 273 (5), 1004–1019. ChemBioChem, 10 (2), 338–344.
57 Wang, L., Gamez, A., Archer, H., 64 Bartsch, S. and Bornscheuer, U.T. (2009)
Abola, E.E., Sarkissian, C.N., A single residue influences the reaction
Fitzpatrick, P., Wendt, D., Zhang, Y., mechanism of ammonia lyases and
Vellard, M., Bliesath, J., Bell, S.M., mutases. Angew. Chem. Int. Ed., 48 (18),
Lemontt, J.F., Scriver, C.R., and 3362–3365.
Stevens, R.C. (2008) Structural and 65 Xiang, L. and Moore, B.S. (2002)
biochemical characterization of the Inactivation, complementation, and
therapeutic Anabaena variabilis heterologous expression of encP, a novel
phenylalanine ammonia lyase. J. Mol. bacterial phenylalanine ammonia-lyase
Biol., 380 (4), 623–635. gene. J. Biol. Chem., 277 (36),
58 Ritter, H. and Schulz, G.E. (2004) 32505–32509.
Structural basis for the entrance into the 66 Christenson, S.D., Liu, W., Toney, M.D.,
phenylpropanoid metabolism catalyzed and Shen, B. (2003) A novel 4-
by phenylalanine ammonia-lyase. Plant methylideneimidazole-5-one-containing
Cell, 16 (12), 3426–3436. tyrosine aminomutase in enediyne
59 Calabrese, J.C., Jordan, D.B., antitumor antibiotic C-1027
Boodhoo, A., Sariaslani, S., and biosynthesis. J. Am. Chem. Soc., 125 (20),
Vannelli, T. (2004) Crystal structure of 6062–6063.
phenylalanine ammonia lyase: multiple 67 Krug, D. and M€ uller, R. (2009) Discovery
helix dipoles implicated in catalysis. of additional members of the tyrosine
Biochemistry, 43 (36), 11403–11416. aminomutase enzyme family and the
60 Moffitt, M.C., Louie, G.V., mutational analysis of CmdF.
Bowman, M.E., Pence, J., Noel, J.P., and ChemBioChem, 10 (4), 741–750.
Moore, B.S. (2007) Discovery of two 68 Ro, D.K. and Douglas, C.J. (2004)
cyanobacterial phenylalanine Reconstitution of the entry point of plant
ammonia lyases: kinetic and structural phenylpropanoid metabolism in yeast
characterization. Biochemistry, 46 (4), (Saccharomyces cerevisiae): implications
1004–1012. for control of metabolic flux into the
61 Louie, G.V., Bowman, M.E., phenylpropanoid pathway. J. Biol. Chem.,
Moffitt, M.C., Baiga, T.J., Moore, B.S., 279 (4), 2600–2607.
and Noel, J.P. (2006) Structural 69 Gloge, A., Langer, B., Poppe, L., and
determinants and modulation of Retey, J. (1998) The behavior of substrate
substrate specificity in phenylalanine- analogues and secondary deuterium
tyrosine ammonia-lyases. Chem. Biol., 13, isotope effects in the phenylalanine
1327–1338. ammonia-lyase reaction. Arch. Biochem.
62 Wang, L., Gamez, A., Archer, H., Biophys., 359 (1), 1–7.
Abola, E.E., Sarkissian, C.N., Fitzpatrick, 70 Gloge, A., Zo n, J., K€ovari, A., Poppe, L.,
P., Wendt, D., Zhang, Y., Vellard, M., and Retey, J. (2000) Phenylalanine
Bliesath, J., Bell, S.M., Lemontth, J.F., ammonia-lyase: the use of its broad
Scriver, C.R., and Stevens, R.C. (2008) substrate specificity for mechanistic
776
investigations and biocatalysis-synthesis enzyme replacement treatment of

of L-arylalanines. Chem. Eur. J., 6 (18), phenylketonuria. Mol. Genet. Metab.,
3386–3390. 86 (1–2), 134–140.
71 Paizs, C., Katona, A., and Retey, J. (2006) 79 Watts, K.T., Mijts, B.N., Lee, P.C.,
The interaction of heteroaryl-acrylates Manning, A.J., and Schmidt-Dannert, C.
and alanines with phenylalanine (2006) Discovery of a substrate
ammonia-lyase from parsley. Chem. Eur. selectivity switch in tyrosine ammonia-
J., 12 (10), 2739–2744. lyase, a member of the aromatic amino
72 Liu, W. (1999) Synthesis of optically active acid lyase family. Chem. Biol., 13 (12),
phenylalanine analogs using Rhodotorula 1317–1326.
graminis. US Patent 5981239. 80 Herrmann, G., Selmer, T., Jessen, H.J.,
73 Van Assema, F.B.J. and Sereinig, N. Gokarn, R.R., Selifonova, O., Gort, S.J.,
(2008) Method for producing optically and Buckel, W. (2005) Two beta-alanyl-
active phenylalanine compounds from CoA: ammonia lyases in Clostridium
cinnamic acid derivatives employing a propionicum. FEBS J., 272 (3), 813–821.
phenylalanine ammonia lyase derived 81 Kreimeyer, A., Perret, A., Lechaplais, C.,
from Idiomarina loihiensis. Intern. Patent Vallenet, D., Medigue, C., Salanoubat,
Appl. No. PCT/EP2007/007945. M., and Weissenbach, J. (2007)
74 Gamez, A., Sarkissian, C.N., Wang, L., Identification of the last unknown genes
Kim, W., Straub, M., Patch, M.G., in the fermentation pathway of lysine.
Chen, L., Striepeke, S., Fitzpatrick, P., J. Biol. Chem., 282 (10), 7191–7197.
Lemontt, J.F., ONeill, C., Scriver, C.R., 82 Jiang, X., Meng, X., and Xian, M. (2009)
and Stevens, R.C. (2005) Development of Biosynthetic pathways for 3-
pegylated forms of recombinant hydroxypropionic acid production.
Rhodosporidium toruloides phenylalanine Appl. Microbiol. Biotechnol., 82 (6),
ammonia-lyase for the treatment of 995–1003.
classical phenylketonuria. Mol. Therapy, 83 Dillon, S.C. and Bateman, A. (2004)
11 (6), 986–989. The hotdog fold: wrapping up a
75 Kang, T.S., Wang, L., Sarkissian, C.N., superfamily of thioesterases and
Gamez, A., Scriver, C.R., and dehydratases. BMC Bioinformatics, 12 (5),
Stevens, R.C. (2010) Converting an 109.
injectable protein therapeutic into an oral 84 Eliot, A.C. and Kirsch, J.F. (2004)
form: phenylalanine ammonia lyase for Pyridoxal phosphate enzymes:
phenylketonuria. Mol. Genet. Metab., mechanistic, structural, and evolutionary
99 (1), 4–9. considerations. Annu. Rev. Biochem., 73,
76 Klettke, K.L., Sanyal, S., Mutatu, W., and 383–415.
Walker, K.D. (2007) Beta-aryl-beta- 85 John, R.A. (1995) Pyridoxal
alanine products of phenylalanine phosphate-dependent enzymes. Biochim.
aminomutase catalysis. J. Am. Chem. Biophys. Acta, 1248 (2), 81–96.
Soc., 129 (22), 6988–6989. 86 Yamada, T., Komoto, J., Takata, Y.,
77 Szymanski, W., Wu, B., Weiner, B., Ogawa, H., Pitot, H.C., and
de Wildeman, S., Feringa, B.L., and Takusagawa, F. (2003) Crystal structure of
Janssen, D.B. (2009) Phenylalanine serine dehydratase from rat liver.
aminomutase-catalyzed addition of Biochemistry, 42 (44), 12854–12865.
ammonia to substituted cinnamic acids: a 87 Sun, L., Bartlam, M., Liu, Y., Pang, H.,
route to enantiopure alpha- and beta- and Rao, Z. (2005) Crystal structure of the
amino acids. J. Org. Chem., 74 (23), pyridoxal-50 -phosphate-dependent serine
9152–9157. dehydratase from human liver. Protein
78 Wang, L., Gamez, A., Sarkissian, C.N., Sci., 14 (3), 791–798.
Straub, M., Patch, M.G., Han, G.W., 88 Okuda, H., Nagata, S., and Misono, H.
Striepeke, S., Fitzpatrick, P., Scriver, (2002) Cloning, sequencing, and
C.R., and Stevens, R.C. (2005) Structure- overexpression in Escherichia coli of a
based chemical modification strategy for phenylserine dehydratase gene from
References j777
Ralstonia pickettii PS22. Biosci. Biotechnol. Salmonella typhimurium. J. Biol. Chem.,
Biochem., 66 (12), 2755–2758. 263 (2), 958–964.
89 Ito, T., Takahashi, K., Naka, T., 97 Imanaga, Y. (1958) Metabolism of D-
Hemmi, H., and Yoshimura, T. (2007) glucosamine. III. Enzymic degradation
Enzymatic assay of D-serine using D- of D-glucosaminic acid. J. Biochem.,
serine dehydratase from Saccharomyces 45 (8), 647–650.
cerevisiae. Anal. Biochem., 371 (2), 98 Iwamoto, R., Taniki, H., Koishi, J., and
167–172. Nakura, S. (1995) D-glucosaminate
90 Gallagher, D.T., Gilliland, G.L., Xiao, G., aldolase activity of D-glucosaminate
Zondlo, J., Fisher, K.E., Chinchilla, D., dehydratase from Pseudomonas
and Eisenstein, E. (1998) Structure and fluorescens and its requirement for Mn2 þ
control of pyridoxal phosphate dependent ion. Biosci. Biotechnol. Biochem., 59,
allosteric threonine deaminase. 408–411.
Structure, 6 (4), 465–475. 99 Iwamoto, R. and Imanga, Y. (1982)
91 Simanshu, D.K., Savithri, H.S., and Enzymatic microdetermination of D-
Murthy, M.R. (2006) Crystal structures of glucosaminate with D-glucosaminate
Salmonella typhimurium biodegradative dehydratase. Anal. Lett., 15 (2),
threonine deaminase and its complex 161–169.
with CMP provide structural insights into 100 Grabowski, R., Hofmeister, A.E., and
ligand-induced oligomerization and Buckel, W. (1993) Bacterial L-serine
enzyme activation. J. Biol. Chem., dehydratases: a new family of enzymes
281 (51), 39630–39641. containing iron-sulfur clusters. Trends
92 Wada, M., Matsumoto, T., Nakamori, S., Biochem. Sci., 18 (8), 297–300.
Sakamoto, M., Kataoka, M., Liu, J.Q., 101 Velayudhan, J., Jones, M.A., Barrow, P.A.,
Itoh, N., Yamada, H., and Shimizu, S. and Kelly, D.J. (2004) L-serine catabolism
(1999) Purification and characterization via an oxygen-labile L-serine dehydratase
of a novel enzyme, L-threo-3- is essential for colonization of the avian
hydroxyaspartate dehydratase, from gut by Campylobacter jejuni. Infect.
Pseudomonas sp. T62. FEMS Microbiol. Immun., 72 (1), 260–268.
Lett., 179 (1), 147–151. 102 Laroche, M., Pukall, R., and Ulber, R.
93 Murakami, T., Maeda, T., Yokota, A., and (2003) Recovery and characterization of a
Wada, M. (2009) Gene cloning and L-serine dehydratase from the marine
expression of pyridoxal 5-phosphate- bacterium Paracoccus seriniphilus for the
dependent L-threo-3-hydroxyaspartate construction of bioanalytical systems.
dehydratase from Pseudomonas sp. T62, Chem. Ing. Tech., 75, 146–149.
and characterization of the recombinant 103 Takeuchi, M., Asano, N., Kameda, Y., and
enzyme. J. Biochem., 145 (5), 661–668. Matsui, K. (1985) Purification and
94 Wada, M., Nakamori, S., and Takagi, H. properties of 3-ketovalidoxylamine A
(2003) Serine racemase homologue of CN lyase from Flavobacterium
Saccharomyces cerevisiae has L-threo-3- saccharophilum. J. Biochem., 98 (6),
hydroxyaspartate dehydratase 1631–1638.
activity. FEMS Microbiol. Lett., 225 (2), 104 Zhang, J.F., Zheng, Y.G., and Shen, Y.C.
189–193. (2010) Purification and characterization
95 Khan, F., Jala, V.R., Rao, N.A., and of 3-ketovalidoxylamine A CN lyase
Savithri, H.S. (2003) Characterization of produced by Stenotrophomonas
recombinant diaminopropionate maltrophilia. Appl. Biochem. Biotechnol.,
ammonia-lyase from Escherichia coli and 162 (4), 966–974.
Salmonella typhimurium. Biochem. 105 St€ockigt, J., Barleben, L., Panjikar, S., and
Biophys. Res. Commun., 306 (4), Loris, E.A. (2008) 3D-Structure and
1083–1088. function of strictosidine synthase - the
96 Nagasawa, T., Tanizawa, K., Satoda, T., key enzyme of monoterpenoid indole
and Yamada, H. (1988) alkaloid biosynthesis. Plant Physiol.
Diaminopropionate ammonia-lyase from Biochem., 46 (3), 340–355.
778
106 Loris, E.A., Panjikar, S., Ruppert, M., 108 Gani, D., Wallis, O.C., and Young, D.W.
Barleben, L., Unger, M., Sch€ ubel, H., and (1983) Stereochemistry of the
St€
ockigt, J. (2007) Structure-based rearrangement of 2-aminoethanol by
engineering of strictosidine synthase: ethanolamine ammonia-lyase.
auxiliary for alkaloid libraries. Chem. Eur. J. Biochem., 136 (2), 303–311.
Biol., 14 (9), 979–985. 109 Yamada, S., Nabe, K., Izuo, N.,
107 Sagermann, M., Ohtaki, A., and Nakamichi, K., and Chibata, I. (1981)
Nikolakakis, K. (2009) Crystal structure of Production of L-phenylalanine from
the EutL shell protein of the trans-cinnamic acid with Rhodotorula
ethanolamine ammonia lyase glutinis containing L-phenylalanine
microcompartment. Proc. Natl. Acad. Sci. ammonia-lyase activity. Appl. Environ.
U.S.A., 106 (22), 8883–8887. Microbiol., 42 (5), 773–778.
j779
19
Application of Transaminases
Matthias H€ohne and Uwe T. Bornscheuer
19.1
Introduction
Optically pure amines and a- and b-amino acids play a key role in living organisms.
These compounds are also highly important in pharmaceutical applications
(Table 19.1). For instance, glutamate, c-aminobutyrate, and derivatized biogenic
amines of tyrosine and tryptophan act as neurotransmitters [1]. Peptide-
based immunomodulators [2, 3], hormones, and enzymatic inhibitors influence
cell-to-cell communication [4] and thus control several important functions in
complex organisms. Various non-proteinogenic amino acids have been discovered
that play an important role in secondary metabolism such as peptide-based
antibiotics [5].
Thus it is not surprising that numerous pharmaceutical applications based on
modified chiral amino acids and amines have been developed. Incorporation of non-
natural amino acids, D-amino acids, or b-amino acids in peptidomimetics can also
lead to several benefits such as improved in vivo stability in these peptides and confer
higher resistance against proteases [6–8]. Furthermore, better bioavailability and
enhanced potency and selectivity of a peptide drug can be obtained as non-natural
amino acids allow a fine tuning of biochemical properties: a restriction of the
flexibility of the peptide or an altered modification of hydrophobicity and dipole
moment of the side chains result in different binding properties to the target
molecules [5, 9].
Consequently, there is a high demand not only for non-natural amino acids but also
amines as building blocks. One of the most successful drugs is Enalapril, which has
achieved annual sales of >US$1 billion [11]. Other ACE inhibitors include Ramipril,
Benazapril, Lisinopril, Zestril, Trandolopril, and Quinipril. A key component in all
these compounds is L-4-phenyl-3-amino-n-butanoic acid, or L-homophenylalanine.
Further examples of pharmacologically active compounds, which are composed of
non-natural amino acids or amines, are given in Table 19.1.
Beside transaminases, many other enzymes can be used to obtain optically pure
amines or amino acids. This includes hydrolases [12, 13], monoamine oxidases
and Oliver May.
j 19 Application of Transaminases
780
Table 19.1 Pharmacologically active compounds containing non-natural amino acids [5] or amines [10].
a-Amino acids b-Amino acids
R Cl
O O
H O H
N N O NH2 O N
N N
H OH H
OH O N O
O
O
b-Homophenylalanine
L-tert-Leucine, (1-aminoindan-2-ol)
Antidiabetic, DPPIV-Inhibitor
HIV-protease inhibitor
Phe
O
NH2 H H
N N N
O N COOH
N H
H NH O
O O F OH
OH
O
L-Homophenylalanine b-Homoserine
Elanapril, Antihypertensive Cytostatic
HOOC
NH2 O
H H
N S
N NH
O N H
HN
R O O
COOH NH2
D-(4-Hydroxy)-phenylglycine b-Lysine
R¼H: Ampicillin R¼OH: Amoxicillin, Antibiotic Platelet GPIIb/IIIa-Antagonist
O
O O N
H H O
N Trp Tyr
O Ser Leu Pro NH2 N N
His N Arg Gly N
H N
O COOH
D-2-Naphthylalanine, 5-oxoproline 3-Amino-3-(6-methoxypyridine-3-yl)-

Nafarelin, gonadotropin-releasing hormone agonist propionic acid; Osteopathic
Amines
F
OEt F
H NH2 O N
N
O N N O
N
OCH3 N
F N O
SO2NH2
CF3
Flomax/Tamsulosin Sitagliptin Rivastigmine

Prostate drug Treatment of diabetes Treatment of Alzheimers
O
N O R3 NH2
R1 N R2 H
HN O N O N
N R3
O N N R1
R2 NH2
CF3
Oxazolone derivative Xanthine derivative Mexiletine Amphetamines

Treatment of diabetes Treatment of diabetes Antiarrhytmic Psychostimulant
19.2 Occurrence and Properties of Transaminases j781
(MAO) [14–16], amino acid dehydrogenases [17], ammonia lyases [18], amino
mutases [19], hydantoinases, [20], aldolases [21], and a few others. All these alter-
native enzymatic approaches have their individual pros and cons and differ for
instance by reaction strategy (i.e., (dynamic) kinetic resolution versus asymmetric
synthesis), cofactor dependencies, enzyme availability, substrate spectra, stereose-
lectivities and -preferences. For more information, the reader is referred to recent
reviews [10, 12, 22].
19.2
Occurrence and Properties of Transaminases
Transaminases (TAs) or aminotransferases (ATs) (EC 2.6.1.X) are probably the most
important and ubiquitous enzymes for the synthesis and degradation of chiral amino
acids and amines in nature. In the overall reaction, the amino group of an amino
donor is transferred to a carbonyl carbon atom of an a-keto acid, ketone, or aldehyde,
the amino acceptor (Scheme 19.1).
NH3 O
R 1 R2 R 1 R2
amino donor keto by-product
Transaminase
O NH3
R 3 R4 R 3 R4
amino acceptor amino product
Scheme 19.1 Principle of transamination catalyzed by a transaminase.
Transaminases are key enzymes in several metabolic pathways, and as a result

are widely distributed in nature [23]. The first evidence for the presence of an
enzyme catalyzing a transamination reaction was published by Needham who
noticed a relationship between the L-glutamic acid, L-aspartic acid, and oxaloacetic
acid levels in pigeon breast muscle [24]. Annau et al. demonstrated the revers-
ibility of glutamic-pyruvic transaminase (E.C. 2.6.1.2, alanine aminotransferase)
by chemically isolating the amino acid products L-glutamate and L-alanine [25].
Since that time, numerous different aminotransferases have been discovered and
characterized, including aminotransferases, capable of catalyzing the transami-
nation of all naturally-occurring amino acids. Of special interest are transami-
nases that also convert chiral primary amines or b-amino acids. As of August
2011, there are more than 86,000 amino acid sequences of aminotransferases
stored in the UniProt database, compared with 51 sequences published in 1993.
Additionally, >450 structures of aminotransferases are available in the PDB
databank.
782j 19 Application of Transaminases
19.2.1
Classification Based on Substrate Specificity
Aminotransferases (ATs) are a very diverse group of enzymes in terms of the

substrate specificity: about 81 distinct enzyme activities are currently listed in the
BRENDA-database (http://www.brenda-enzymes.org/). In principle, amino donor
substrates can be divided in three chemically different groups, which differ in the
presence and position of a negatively charged group (usually a carboxylate), with
respect to the amino group being transferred (Scheme 19.2). Most known enzymes
that are involved in proteinogenic amino acid metabolism, for example, aspartate TA
and aromatic TA, convert only substrates having a carboxyl group at the a-position.
These enzymes are referred to as a-TA and are useful for the synthesis of non-natural
amino acids (non-proteinogenic and/or in D-configuration). A second, smaller group
of enzymes, transfers a terminal amino group attached to a primary carbon atom at
least one carbon atom away from a carboxyl group. Examples include transamination
reactions of the e-amino group of lysine and conversion of so called v-amino acids
like b-alanine, c-aminobutyrate, and (acetyl)-ornithine. As the carbon atom bearing
the amino group being transferred is achiral, this type of v-TA did not find many
applications in synthesis.
Transaminases
α-TA ω-Amino acid TA Amine-TA
NH2 R2 NH2
Amino donor
NH2
R 1
COOH HOOC n R R4
3
3
R : COOH, alkyl, aryl
R1: side chain of amino acid R2: NHR or H
R4: H, CH3, C2H5, CH2OH
Amino acceptor α-keto glutarate α-keto glutarate pyruvate
Equilibrium constant ≈1 >> 1 (alanine formation favored)
Scheme 19.2 Classification of transaminases based on substrate specificity.
A third relatively small group of enzymes can act on substrates lacking any carboxyl
group, such as amines and ketones. This offers the possibility for the synthesis of
chiral primary amines and, thus, these amine transaminases (ATAs) are especially
synthetically useful. In the literature, they sometimes are denoted as v-TA as some of
them additionally convert b-amino [27–29] or v-amino acids, but to avoid confusion
with the above-described enzymes the designation amine transaminase is preferred
throughout this chapter. Compared to a-TA and v-amino acid TA, amine transa-
minases use pyruvate as universal amino acceptor in contrast to a-ketoglutarate,
which is the preferred amino acceptor for most a-TA. A second important difference
between amine-TA and a-TA is the equilibrium of the reaction: whereas
the equilibrium constant in a-amino acid transaminations is close to unity, the
production of alanine is strongly favored in reactions with amine transaminases and
hence additional efforts are required to shift the equilibrium towards synthesis of the
desired chiral amine (see Section 19.3.3 for details).
19.2.2
Classification Based on Sequence Similarities and Three-Dimensional Structures
The different substrate specificities correlate somewhat to sequence similarities [26,

30]. Based on sequence homology, TAs were divided into six subgroups (Table 19.2,
classification according to PFAM: http://pfam.sanger.ac.uk/). Despite the great
diversity of different substrate specificities, transaminases are found in only two
different protein folds (Table 19.2, Scheme 19.2) [31, 32], suggesting a high flexibility
of the active site to explain the very broad substrate specificities.
Most of the enzymes can be assigned to Fold Class I, with aspartate aminotrans-
ferase as the most prominent and best characterized member. At present, the
only known members of Fold Class IV are branched chain amino acid transaminases
(BCATs), D-amino acid transaminases (DATA), and (R)-selective amine transami-
nases [31–33]. The enzymes of both folds are homodimers or higher-order
oligomers. Each monomer consists of a large and a small domain, and the active
site is located at the monomer interface; thus, residues of both monomers contribute
to the active site. A significant difference of both folds is that the PLP-cofactor is
bound mirror-inverted in the active site: in Fold Class I or IV, the Si- or the Re-face of
PLP is exposed to the solvent, respectively.
Table 19.2 Subdivision of aminotransferases based on sequence/structural similarity.a)
Sub-group Fold Enzyme Substrates Amino

acceptor
I/II I Aspartate AT L-Aspartate 2-KG

Alanine AT L-Alanine
Aromatic AAAT L-Aromatic amino acid
Histidinol-phosphate AT L-Histidinol phosphate
III I (Acyl)-ornithine-AT N-Acetyl-L-ornithine 2-KG
c-Aminobutyrate-AT c-Aminobutyrate
Lysine-e-AT L-Lysine
Diaminopelargonate AT 7,8-Diaminopelargonate SAM
(S)-Selective amine-AT (S)-2-Phenylethylamine Pyruvate
IV IV D-Amino acid-AT D-Amino acids 2-KG
Branched chain AA-AT L-Valine, (iso)leucine
(R)-selective amine-AT (R)-2-Phenylethylamine Pyruvate
V I Serine-AT L-Serine Pyruvate
Phosphoserine-AT 3-phospho-L-serine 2-KG
VI I Sugar-AT UDP/TDP linked 3- or 2-KG
4-aminosugars scyllo-inosose
AT: aminotransferase, AA: amino acid, KG: keto glutarate, SAM: S-adenosyl-L-methionine.
784
19.2.3
Mechanism
The mechanism of the transamination reaction is well understood and was investi-
gated in great detail for aspartate aminotransferase [34–36]. The amino group transfer
is mediated by the cofactor pyridoxal phosphate (PLP), which is reversibly bound to the
enzyme through a Schiff-base linkage to the e-amino group of an active-site lysine.
Mechanistically, the reaction catalyzed by an aminotransferase can be thought of as the
result of two discrete steps. The first step is the transfer of an amino group from the
amino donor, e.g. alanine, to pyridoxal phosphate. This generates an enzyme-bound
pyridoxamine phosphate intermediate and a keto by-product, which subsequently
dissociates from the enzyme (Scheme 19.3, step 1). The second step involves the
transfer of the amino group from the enzyme-bound pyridoxamine phosphate to
the amino acceptor, producing the corresponding amino product, and regenerating
the pyridoxal phosphate cofactor for another catalytic cycle (Scheme 19.3, step 2). As a
result, aminotransferases characteristically exhibit ping-pong-bi-bi kinetics [37].
Unfortunately, substrate and product inhibition of transaminases arise as a
consequence of the reaction mechanism (Scheme 19.3): On the one hand, the
substrate may bind to the free enzyme, forming abortive dead-end complexes [38],
e.g. pyruvate and E-PLP (other examples are shown in Scheme 19.3, dark grey boxes).
Note that in kinetic resolutions using amine-TA, both amino donor enantiomers
can act as inhibitor. On the other hand, product inhibition is caused by the formation
of the Michaelis complex of the product with the correct free enzyme, for example,
if the generated amine binds to the PLP-form of the amine-transaminase. If, thus, the
seat is already taken a fast conversion of the substrate is prevented.
A second general problem when using TA is the equilibrium of the reaction. Since
all steps in the mechanism are reversible, different methods have to be applied to
shift the equilibrium towards products to ensure high yields of the reaction. Possible
solutions for circumventing product/substrate inhibition and different approaches
for equilibrium shift will be discussed in detail below.
19.2.4
Methods to Assay Transaminase Activity and Enantioselectivity
Methods for characterization of enzyme properties and determination of substrate

spectra, conversion, and especially optical purity of substrate and product are important
requirements in enzymology and, especially, biocatalysis. The outcome of a biocatalytic
reaction is usually determined by classical analytical methods such as HPLC or gas
chromatography; for transaminases, capillary electrophoresis [39] was also applied as
an analytical method [40, 41], as charged compounds occur during the reaction.
Nevertheless, all of these methods are time consuming and usually require
extraction and derivatization steps and hence are not useful for fast measurement
of enzyme rate, conversion, and optical purity. In addition, fast and reliable assays are
very important if protein engineering of an enzyme by rational design and, especially,
directed evolution is performed. Scheme 19.4 and Table 19.3 show a summary of
available assays, which are described in the next paragraph. For a-TA, the situation is
Lys
N
O O
Pi
(S)-amine L-alanine
N
H
E-PLP 1
internal aldimine
R R' H
COO
+
N H N+
H ketone H
O O pyruvate (R)-amine O O
Pi Pi
N N
H H
E-PLP-amine E-PLP-alanine
external aldimine
R R' COO
N+ H
H N+
H H
(S)-amine alanine
O O O O
Pi (R)-amine Pi
N N
H 2 H
E-PMP-pyruvate
E-PMP-ketone = ketimine
ketone pyruvate
NH3
O O
Pi
N
H
E-PMP
Scheme 19.3 Reaction cycle of transaminases released from the enzyme. During the
exemplified by the asymmetric synthesis of an reaction cycle, two forms of the free enzyme
amine with amine-TA. Although all reactions are (E-PLP and E-PMP) occur. Substrate and
fully reversible, only simple reaction arrows are product inhibition may be caused by binding
shown to indicate the direction of the desired of substrates (shaded in dark gray) to the
asymmetric synthesis and the chronological wrong free enzyme, forming abortive
order in which the substrates and products complexes (see text), which results in
(shaded in light gray) have to be bound or inhibition of the enzyme.
786
relatively easy since most enzymes use glutamate or aspartate as amino-donor. Thus,
the formed a-ketoglutarate or oxaloacetate can be monitored by glutamate dehydro-
genase or malate dehydrogenase [42] (Scheme 19.4a). Alternatively, the keto-acids
corresponding to a range of a-amino acids can be reduced by, for example, NADH-
dependent Lactobacillus delbrueckii hydroxyisocaproate dehydrogenase [43]
(Scheme 19.4b), which does not convert a-ketoglutarate. For amine-TA, however,
the situation is more complicated and the CuSO4/MeOH assay (Scheme 19.4c) was
the first spectrophotometric method, described in 2004 [44]. In recent years, however,
several alternative assays were developed to determine amine transaminase activity
(Scheme 19.4, Table 19.3).
The CuSO4/MeOH assay (Scheme 19.4c) is based on the formation during the
reaction of a blue copper-complex with alanine [44]. Any primary amine or b-amino
acid can be used as amino donor and, thus, the substrate specificity can be
investigated. Since the staining solution inhibits the enzyme, the assay can only be
used as an end point measurement. A further disadvantage is the very low sensitivity
(e 0.1 M1 cm1), and that the most commonly used buffers (e.g., phosphate and
the Goods buffer) and cell extract form blue Cu-complexes, too. Thus it was
recommended to use whole cells in the assay reaction. Beside its disadvantages
this assay is the only spectrophotometric assay that can be used for b-amino acids.
In the acetophenone assay (Scheme 19.4d) the high UVabsorbance of acetophenone
(lmax ¼ 245 nm, e245 ¼ 12 mM1 cm1) allows a sensitive spectrophotometric mea-
surement ofa transaminationreaction, if(R,S)-1-phenylethylamine(a-MBA)isusedas
amino donor [45]. Since most amine-TAconvert a-MBAvery well, this assay can be used
asa standardmethod for thedeterminationof enzyme activity during purificationsteps,
for biocatalysis reactions, and furthermore for the determination of pH and temper-
ature profiles. Additionally, it was shown that the assay can be used for investigating the
amino acceptor specificity, since as well as pyruvate, all non-absorbing substrates (e.g.,
aliphatic ketones, aldehydes, keto acids) can be used as cosubstrates in the assay.
On the other hand, the amino donor specificity can easily be investigated with a
conductometric assay (Scheme 19.4e) [46]. During the reaction, charged reactants
(amine and keto acid) are converted into non-charged species (the ketone and a
zwitterionic amino acid) and, thus, the conductivity of the reaction solution decreases.
Any primary amine that does not contain an additional negatively charged group can
be used together with a keto acid as substrate [46]. Since both enantiomers of a given
amine can be used separately, information about the enantioselectivity of the enzyme
can be obtained. For high sensitivity it is important to use a zwitterionic buffer, for
example, CHES or Tricine, to keep the background conductivity to a minimum.
The acetophenone and the conductometric assays complement each other and
allow rapid characterization of the complete substrate specificity of an amine-TA.
Furthermore, they offer the advantage of a kinetic measurement, and handling is very
easy, since no additional enzyme or staining solution is involved in the assay reaction.
Purified proteins as well as crude cell extract can be used as enzyme sources.
In contrast to the assays described above, in the pH-shift assay (Scheme 19.4f) the
transamination reaction is performed as an asymmetric synthesis [47]: any ketone
can be reacted with alanine as cosubstrate, and the unfavorable equilibrium of
(a) Glutamate- / malate dehydrogenase assay (b) Hydroxyisocaproate dehydrogenase assay
O NH2 NH2 O
α-TA
OH OH OH OH
R R R R
O α-TA O O O
amino acid α-KG Glu
NADH
HICDH
Glu or Asp α-KG or oxaloacetate NAD+
NADH
NADH OH
MDH NAD+ OH
GlDH R
H2O, NH3, NAD+
O
NAD+ NADH malate
(c) CuSO4 /MeOH assay (d) Acetophenone assay

NH2 O O NH2
R1 R2 R1 R2 R1 R2 R1 R2
amino donor Amine-TA ketone amino acceptor amino product
Amine-TA
O NH2 NH2 O
R3 COOH R3 COOH
α-keto acid α-amino acid
Cu2+ 1-phenylethylamine acetophenone

copper amino acid complex λmax=245 nm, ε=12 mM-1cm-1
λmax = 595 nm, ε=0,1 mM-1cm-1
(e) Conductivity assay (f) pH Indicator assay
O O NH2
NH3 Amine-TA
R2 R3 R1 R2 R1 R2
R2 R3
amino donor ketone amino acceptor amine product
Amine-TA
alanine pyruvate lactate
O NH3 LDH
R1 COO R1 COO
keto acid amino acid NADH NAD+
GDH
charged substrates zwitterionic/uncharged
high conductivity products gluconic acid glucose
low conductivity δ-lactone
H2O
(g) Selection assay
gluconic acid
NH2 O pH-indicator
Amine-TA
phenol red
R1 R2 R1 R2 color change,
amine as sole pyruvate alanine ketone detection: 560 nm
nitrogen source
usable N-source
cell growth
Scheme 19.4 Principle of assay methods used to determine transaminase activity. GlDH –
glutamate dehydrogenase; MDH – malate dehydrogenase; HICDH – hydroxyisocaproate
dehydrogenase; LDH – lactate dehydrogenase; GDH – glucose dehydrogenase. See text and table
19.3 for details.
788
Table 19.3 Overview of assay methods used to determine transaminase activity.a)
Principle Reaction Substrates/reference Application

mode
a-TA
(a) Reduction of AS Any a-keto acid [42] Screening of a-keto acids for
a-KG, oxaloacetate asymmetric synthesis
(b) Reduction of KR L-Phe, L-Tyr, L-Trp, L-Met, Screening of L-amino acids/
a-keto acid L-Leu [43] screening of enantioselectivity
Amine-TA
(c) Cu-amino acid KR Any primary amine/ Amino donor specificity, espe
complex b-amino acid and any cially b-amino acids, HTS
a-keto acid [44] possible
(d) Photometric KR (R,S)-1-phenylethyl Rapid kinetic determination of
determination amine, any non- enzyme activity; characteriza
of acetophenone absorbing amino tion of amino acceptor
acceptor [45] specificity
(e) Conductivity KR Any primary amine/any Characterization of amino donor
keto acid [46] specificity and apparent
enantioselectivity
(f) pH-shift AS Alanine and any Screening of ketones for
ketone [47] asymmetric synthesis
(g) Growth assay KR Any non-toxic primary Selection
amine [48]
a)
KR: kinetic resolution, AS: asymmetric synthesis, HTS: high-throughput.
the reaction is shifted by reduction of the produced pyruvate to lactate. The NADH
consumed is regenerated with glucose dehydrogenase, whereby gluconic acid forms,
causing a decrease in pH of the solution. Therefore, the reaction can be monitored
either by a pH-indicator or by titration with a base. Although no information about
enantioselectivity can be obtained with this assay, the feasibility of an asymmetric
synthesis with a given ketone can be investigated easily since after the analytical-scale
biocatalysis the reaction can be scaled up easily.
In a growth assay (Scheme 19.4g) it is possible to select for a microorganism
possessing an amine-TA that is active toward a desired amine [48]. In the growth
medium, the amine must represent the sole nitrogen source. If it is converted by
amine-TA, the amino group is transferred to pyruvate and thus a nitrogen source for
cell growth is generated. The limitation of this approach lies in the potential toxicity of
the amine and corresponding ketone.
19.3
Strategies for Using Transaminases in Biocatalysis
Aminotransferases (transaminases) have been studied as biocatalysts for the pro-

duction of a wide range of different amino acids and amines. As described above in
19.3 Strategies for Using Transaminases in Biocatalysis j789
Section 19.2.3, pyridoxal-50 -phosphate is involved in the catalysis. The cofactor,
which is only required in concentrations of 50–100 mM, is reversibly bound to
the enzyme through a Schiff-base linkage to the e-amino group of an active-site
lysine [35, 37].
The use of aminotransferases offers certain advantages:
. Especially a-TAs often show high stereoselectivity for a given enantiomer.
. a-TA and amine-TA are potentially applicable to the production of a wide
range of amino acids and amines because enzymes with opposite enantioprefer-
ences are available. In addition, a wide range of aminotransferases with side-
chain specificity are known, including enzymes for the production of amino
acids with aromatic side chains, acidic side chains, branched alkyl side chains,
and so on.
. The catalytic rates of these enzyme-catalyzed reactions are generally relatively
high. For a-TA, specific activities are often higher compared to amine-TA (50–500
versus 1–20 U mg1). This might be related to the fact that the a-carboxyl group
increases the acidity of the a-proton and thus facilitates its abstraction by the
enzyme during the reaction.
The transamination reaction can be carried out as kinetic resolution
(Scheme 19.5a) or asymmetric synthesis (Scheme 19.5b) starting with a prostereo-
genic keto substrate. The asymmetric synthesis is often the preferred route because
of the higher yield that can be obtained. Furthermore, the enantiomeric purity of the
product is not dependent on conversion, in contrast to a kinetic resolution in which
50% conversion must be achieved for high enantiomeric excess of the substrate.
Nevertheless, a strategy for equilibrium shift has to be employed to achieve high
yields. This is usually afforded by the removal of the keto coproduct, and several
examples are illustrated below.
In some cases, however, the prostereogenic keto substrate is not available or
unstable. By combining kinetic resolution and asymmetric synthesis, a deracemiza-
tion can be carried out in a multi-enzyme process [49, 50]. Deracemization of an
amine racemate starts with a kinetic resolution, yielding one amine enantiomer and
the ketone. After removal of the first amine-TA by heat treatment, the ketone is used
in an asymmetric synthesis reaction in the second step. An amine-TA with opposite
enantiopreference has to be used, which leads to a theoretical yield of 100% for the
desired enantiomer after completion of the reaction (Scheme 19.5c). By exchanging
the order of transaminases employed in the two steps, the amine with inverted
configuration will be generated during the reaction.
Deracemization of amino acids can be achieved by a similar process
(Scheme 19.5d) [51] except that an amino acid oxidase or dehydrogenase is used
in the kinetic resolution step. Hence, the asymmetric synthesis step carried out by an
a-TA can be performed simultaneously as a one-pot reaction. Again, amino acid
oxidase/dehydrogenase and a-TA must have opposite enantiopreference. By these
processes, either the D- or L-amino acid can be obtained in optically pure form,
depending on the enantiopreference of the enzymes used.
790 j 19 Application of Transaminases
(a) Kinetic resolution with amine-TA (c) One-pot, two-step deracemization with amine-TA
NH2 O NH2 2nd step:

+
R1 R2 R1 R2 R1 R2
NH2 (R)−amine-TA
amine-TA 1st step: pyruvate
+ + ≤ 50 % NH2 R1 R2
(S)−amine-TA co-product
O NH2 + removal
R1 R2 O
R3 COOH R3 COOH pyruvate alanine alanine
amino acceptor, amino acid by-product R1 R2
e.g. pyruvate, 2-KG
+ equilibrium constant >> 1 if pyruvate is used as amino + 100 % yield possible

acceptor No equilibrium shift is neccessary + method of choice if ketone is not available/unstable
- ≤ 50 % yield, ≥ 50 % ketone by-product - two enantiocomplementary enzymes with same substrate
- high %eeS can only be reached after complete specificity needed
conversion of fast reacting enantiomer - inactivation of TA after first step needed
(b) Asymmetric synthesis with α-TA or amine-TA (d) Deracemization with α-TA
O NH2 D-amino acid NH2

α−TA or oxidase L-α-TA
R1 R2 amine-TA R1 R2 1 2-KG
NH2 O2 (catalase) R COOH
+ ≤ 100 % or +
NH2 O R1 COOH
D-amino acid O
dehydrogenase glutamate
R3 COOH R3 COOH
R1 COOH
amino donor keto acid co-product
NAD+ NADH
+ prochiral substrate
+ 100 % yield possible + 100 % yield possible
+ %eeP not depending on conversion + simultanous one-pot reaction
- TA with excellent enantioselectivity needed + method of choice if keto acid is not available/unstable
- for high conversion, equilibrium has to be shifted - two enantiocomplementary enzymes with same substrate
by e.g. co-product removal specificity needed
Scheme 19.5 General strategies for transaminase-catalyzed reactions. See text for details.
19.3.1
Kinetic Resolution with Amine-TA
Kinetic resolution was studied in detail with amine transaminases. Apart from the
lower yield, kinetic resolution using amine-TA shows some advantages. The
equilibrium strongly favors product formation, if pyruvate is used as amino
acceptor. Therefore, no equilibrium shift has to be applied and thus the trans-
aminase reaction is very easy in terms of practical handling. For example, the
amine-TA catalyst can be expressed in Escherichia coli and subsequently used as
whole cell system [52]. Furthermore, the enantiomer with opposite configuration
can easily be obtained compared to asymmetric synthesis. Since most amine-TAs
discovered in the last decade show (S)-enantiopreference, kinetic resolution was
an attractive method for preparation of (R)-amines. Since a dozen (R)-selective
amine-TAs were identified very recently, the method can also be used for the
preparation of (S)-amines [33]. Even if the enantioselectivity of the amine-TA is not
perfect, high enantiomeric excess can be obtained, although at the expense of a
decreased yield.
However, kinetic resolution suffers from two main disadvantages. During the
reaction, stoichiometric amounts of the two by-products alanine and ketone are
formed. The enantiomeric excess of the remaining amine is dependent on the
conversion of the fast reacting enantiomer. This is especially important in
upscaling to increased substrate concentrations: owing to substrate/product
inhibition, the reaction slows down if a limiting concentration of product is achieved,
and consequently can yield the amine in only low or moderate enantiomeric
purity [38, 53].
Different solutions were developed to deal with these problems. The amino acid
by-product D- or L-alanine can be converted in situ by a D- or L-amino acid oxidase and
molecular oxygen into pyruvate [54]. Thus, only a catalytic amount of pyruvate is
needed (Scheme 19.6). In a kinetic resolution of 100 mM 1-phenylethylamine, 2 mM
pyruvate was sufficient for a fast resolution. Lowering the amount of pyruvate
increased the reaction time significantly. The ketone by-product might be isolated
and recycled after the reaction to the racemic amine by, for example, reductive
amination.
NH2 NH2 O
amine-TA +
R1 R2 R1 R2 R1 R2
O NH2
COOH COOH
H2O2 O2
alanine-oxidase
Scheme 19.6 An efficient kinetic resolution is achieved by recycling the amino acceptor by the
oxidation of the formed alanine with molecular oxygen in the presence of amino acid oxidase [54].
Product inhibition is an intrinsic property of the transaminase that arises as a

consequence of the mechanism (Scheme 19.3). In kinetic resolutions, the highest
inhibition often results from the presence of the ketone by-product. One example is
the kinetic resolution of sec-butylamine with Vibrio fluvialis amine-TA [55]. Increasing
the substrate concentration from 20 to 200 mM amine led to a dramatic decrease of
enantiomeric purity as the reaction did not proceed further, even if a higher enzyme
concentration was used (Table 19.4, Scheme 19.7). If the formed 2-butanone was
removed from the reaction under reduced pressure, 400 mM sec-butylamine could be
resolved to 98%e.e.S [55]. To circumvent a loss of the similarly volatile substrate
amine, the reaction had to be run at pH 7 (pH-optimum of the transaminase was pH
9) where the amine substrate is protonated and less volatile. The impact of product
inhibition may differ dramatically from enzyme to enzyme: When the amine-TA
from Bacillus megaterium was used, kinetic resolution of approx. 700 mM isobuty-
lamine was possible without any need to remove 2-butanone [52].
792
Table 19.4 Product inhibition in the kinetic resolution of sec-butylamine [55].
sec-Butylamine (mM) Optical purity (% e.e.S)
20 99
80 94
200 32
400a) 98
a)
At reduced pressure.
NH2 amine-TA NH2 O

+
pyruvate L-alanine
Scheme 19.7 Severe product inhibition occurs in the kinetic resolution of sec-butylamine.
Another example is the kinetic resolution of 1-phenylethylamine with Bacillus

thuringensis amine-TA [56]. The acetophenone formed during the reaction caused
severe product inhibition, since at concentrations >20 mM the enzyme was virtually
inactive. Nevertheless, the kinetic resolution of 500 mM 1-phenylethylamine could
be carried out successfully to 95% e.e.S by the application of a biphasic system with
cyclohexanone as organic solvent. Thus, the acetophenone formed was extracted and
did not disturb the reaction. Alternatively to a biphasic system, a membrane reactor
can be used with similar success [53].
In summary, evaporation or extraction represent straightforward strategies for the
removal of inhibiting ketone products, which may be chosen according to the
properties of the ketone. Alternatively, an amine-TA showing less inhibition might
be identified, or a given enzyme can be optimized by means of protein engineering to
reduce inhibition. One example is described in Section 19.3.4.
19.3.2
Asymmetric Synthesis with a-TA
Because the transamination reaction involves an amino acid reacting with a 2-keto
acid to generate products that consist of a 2-keto acid and an amino acid, the
equilibrium constant is often close to unity. As a result, the net conversion of
substrates into products is thermodynamically limited. The key to the development of
an efficient transamination technology lies in overcoming the problem of incomplete
conversion of the 2-keto acid precursor into the desired amino acid product. One
option is to use a large excess of amino donor. Aside from higher costs, there are two
main reasons why this strategy usually is avoided: a too high amino donor concen-
tration might lead to substrate inhibition, and, secondly, a high amount of non-
converted amino donor may complicate the purification of the desired amino acid.
Several solutions for equilibrium shift were developed, which are described in the
following subsections.
19.3.2.1 Product Precipitation
Some amino acids like naphthylalanine [57] and homophenylalanine [58] show a very
low solubility, contrary to their corresponding keto acids. Thus, during an asym-
metric synthesis reaction the amino acid product precipitates (Scheme 19.8). Its
concentration in solution remains fairly low compared to the substrate a-keto acid
and, consequently, high yields can be obtained. This represents the easiest case since
no additional equilibrium shift has to be applied. In most other applications, the
reaction must be driven towards completion by removing the arising coproduct. This
is usually achieved by coupling the transamination reaction to a second reaction that
consumes the keto acid by-product in an essentially irreversible step.
O NH2
R OH α-TA R OH R = naphthyl, benzyl
O O
α-keto acid Glu α-KG amino acid
- medium solubility - - low solubility -
Scheme 19.8 Equilibrium shift by product precipitation. Glu – glutamate, a-KG – a-ketoglutarate.
19.3.2.2 Decomposition of the Keto Acid By-Product

By using an aminotransferase that can utilize aspartic acid efficiently as the amino
group donor (instead of glutamic acid), the corresponding 2-keto acid by-product is
oxaloacetate (rather than 2-ketoglutarate). Oxaloacetate is a b-ketoacid and can be
easily decarboxylated in the presence of various metal ions to pyruvate [59], which is
often a poor substrate for many a-TA (Scheme 19.9a). If only glutamate is accepted as
amino donor, aspartate TA can be included in the reaction and thus the equilibrium
shift by decomposition of oxaloacetate is linked to the desired asymmetric synthesis
reaction. In the net-reaction, the desired amino acid is formed by amination of the
keto acid with aspartate as amino donor, and decarboxylation of oxaloacetate will drive
the reaction to completion (Scheme 19.9a).
The efficiency of the equilibrium shift can be further enhanced by accelerating the
decarboxylation enzymatically using the enzyme oxaloacetate decarboxylase, an
Mg2 þ -requiring enzyme (Scheme 19.9a) [60].
The effectiveness of this strategy was demonstrated in a coupled enzymatic process
by using phenylpyruvate as the starting 2-keto acid. In this experiment, phenylpyr-
uvate sodium salt and L-aspartate were incubated with E. coli broad-range transam-
inase at room temperature and pH 7.5. When oxaloacetate decarboxylase from
Pseudomonas putida was included in the mixture, the reaction preceded to completion
much more rapidly (Figure 19.1) [60].
A similar but more cost-intensive approach is the use of cysteine sulfinic acid as
aspartate analog (Scheme 19.9b). The formed 2-oxo-3-sulfinopropanoic acid is highly
instable and decomposes rapidly to pyruvate by the cleavage of sulfur dioxide, thus
rendering the reaction irreversible [61, 62].
794
O α-TA NH2
OH OH
R R
O O
Glu α-KG
GOT
spontanous or
O NH2 O O decarboxylase O
(a)
HO COOH HO COOH COOH
CO2
OR
O NH2 O O spontanous O
(b) S S
HO COOH HO COOH COOH
cysteine 2-oxo-3-sulfino- SO2
sulfinic acid propanoic acid
O α-TA NH2
(c) OH OH
R R
O O
alanine pyruvate
PDC
acetaldehyde + CO2
Scheme 19.9 Driving the reaction by decarboxylase. (b) Alternatively, cysteine

decomposition of the keto acid coproduct: sulfinic acid can be used as amino-donor,
(a) With aspartate as amino donor; if the desired since its corresponding keto acids
a-TA does not convert aspartate well, aspartate decompose readily to pyruvate and SO2.
conversion can be linked by coupling with (c) If alanine is accepted by the a-TA,
glutamate:oxaloacetate-TA (GOT). pyruvate can be decarboxylated by pyruvate
Decarboxylation of oxaloacetate is significantly decarboxylase to yield the volatile
promoted by including oxaloacetate acetaldehyde as product.
If alanine is accepted as amino donor, pyruvate decarboxylase can be used for

removal of the keto acid by-product pyruvate (Scheme 19.9c) [40]. The acetaldehyde
formed in this reaction is highly volatile and may be removed from the reaction
medium by evaporation. If the transaminase does not convert alanine well, glutamate
may be used as primary amino donor in catalytic amounts, and the two reactions
might be linked by addition of alanine-TA.
19.3.2.3 Recycling of the Amino Donor via Reductive Amination

In a very elegant solution the transamination reaction was coupled with amino acid
dehydrogenase (Scheme 19.10). The amino donor can be recycled by reductive
amination of the keto acid by-product with ammonia and NADH. Cofactor regen-
eration can be performed, for example, with formate dehydrogenase (FDH). In the
200
TA alone
TA + ODC broad range
Phenylpyruvate [mM]
150
α-TA (E. coli)
O NH2
100 HOOC HOOC

Asp Oxac
50 ODC CO2
Pyruvate
0
0 40 80
Reaction time [min]
Figure 19.1 Oxaloacetate decarboxylase (ODC) speeds up the equilibrium shift. Oxac: oxaloacetate.
O α-TA NH2
OH OH
R R
O Glu or α-KG or O
Leu MOPA
AADH
NH4+
+
NAD NADH
FDH
HCOOH CO2
Scheme 19.10 Driving the reaction by recycling of the amino donor via reductive amination. AADH:
amino acid dehydrogenase, FDH: formate dehydrogenase, MOPA: 4-methyl-2-oxopentanoic acid.
net reaction, the desired keto acid substrate is reductively aminated with ammonium
formate, which simultaneously serves as amino and hydrogen donor [63–65].
19.3.2.4 Coupling with v-Amino Acid TA

The asymmetric synthesis reaction can also be coupled to lysine-e-aminotransferase.
Thereby, the amino donor glutamate is regenerated from a-ketoglutarate using the
e-amino group of lysine, thus generating lysine-6-semialdehyde. Intramolecular
cyclization of the latter is highly favored and leads to D1-piperidine-6-carboxylic acid,
which cannot act as a substrate in the transamination and, thus, the reverse reaction
is prevented (Scheme 19.11) [66].
19.3.2.5 Synthesis of D-Amino Acids

The synthesis of D-amino acids requires the more expensive D-glutamate or
D-aspartate as amino donors. Furthermore, if the equilibrium shift reaction involves
the coupling of aspartate-TA or amino acid dehydrogenase for recycling of amino
donor, the L-enantiomer is formed instead of the necessary D-glutamate. In these
796
O NH2
R OH α-TA R OH
O O
Glu α-KG
Lys-ε-TA
NH2 NH2
HOOC HOOC
N COOH
NH2 O
Scheme 19.11 By including Lys-e-TA, the reverse reaction is prevented as the semi-aldehyde by-
product undergoes intramolecular cyclization.
cases, including an amino acid racemase solves both problems, since the D-amino
acid is generated in situ from its L-enantiomer [65].
19.3.2.6 Equilibrium Shift in Action

The coupled system with aspartate TA and subsequent decarboxylation of oxaloacetate
was applied in the asymmetric synthesis of L-phosphinothricin (Scheme 19.12a), a
glutamate analog that is used as the active ingredient of the broad-spectrum herbicide
BastaÒ . 4-Aminobutyrate:a-ketoglutarate transaminase from E. coli and aspartate
TA from Bacillus stearothermophilus were used as immobilized enzymes for its
synthesis. Beside an increased yield (85%), one benefit of the coupled reaction with
glutamic:oxaloacetic transaminase (GOT) was a simplified purification: since only a
catalytic amount of glutamate was necessary in the coupled reaction, the glutamate
concentration at the end of the reaction was significantly lower, thus facilitating the
separation of phosphinothricin from the structurally similar glutamate [67].
One disadvantage of the GOT-coupled process is that the pyruvate formed in the
decarboxylation step can lead to a side reaction where alanine is formed by
transamination, which limits the maximum conversion and might complicate the
purification. This was examined in detail in the synthesis of 2-aminobutyrate,
providing an excellent example how genetic engineering can be used to
improve the efficiency of biocatalytic application of TA (see Scheme 19.12.b) [68,
69]. At the end of the transamination of 2-ketobutyrate with E. coli whole cells
overexpressing tyrosine aminotransferase, the reaction medium consisted of a
mixture of 2-aminobutyrate and approx. 0.5 equivalents of alanine. The co-expression
of acetolactate synthase from Bacillus subtilis resulted in a decrease of alanine to only
6 mol % of aminobutyrate, as pyruvate was converted into acetolactate, which
spontaneously decarboxylated to acetoin. To further decrease the cost of the process,
the gene of threonine deaminase was introduced into the expression host. Thus, the
cheaper threonine could be used as starting material and 2-ketobutyrate was
generated in situ by the action of threonine deaminase.
(a) Asymmetric synthesis of L-phosphinothricin
O α-TA NH2 NH2

O O
P COOH P COOH COOH
HO HO
Glu α-KG L-phosphinothricin
GOT side-
reaction
NH2 O O
HOOC HOOC
COOH COOH COOH
CO2
(b) Asymmetric synthesis of 2-aminobutyrate
NH2 O NH2
TDA TAT
O O O
OH OH NH3 OH OH
threonine COOH COOH 2-aminobutyrate
NH2 O
COOH COOH
Asp Oxac
CO2
O COOH
ALS COOH
O OH
O
OH CO2 CO2
acetoin acetolactate pyruvate
Scheme 19.12 Asymmetricsynthesis ofamino acids by multi-enzymecascade reactions.a)Synthesis

of phosphinothricin, b) Synthesis of 2-aminobutyrate. GOT: glutamic:oxaloacetic transaminase,
TDA: threonine deaminase, TAT: tyrosine aminotransferase, ALS: acetolactate synthase.
The utility of the diversity of strategies for equilibrium shift was also demonstrated
in the synthesis of several glutamate analogues with branched chain aminotrans-
ferase. Depending on the nature of the substrate, the choice of amino-donor and
strategy for equilibrium shift was dictated by purification constraints. For the
synthesis of 3-methyl- or 3-ethylglutamate, leucine had to be used as amino donor
since the products were difficult to separate from glutamate impurities by ion-
exchange chromatography. Thus, leucine dehydrogenase was the preferred strategy
for equilibrium shift. In contrast, for preparing the 3-propyl- or 3-phenyl-derivatives
separation of glutamate was easily achieved, and the equilibrium shift was performed
by coupling the reaction to aspartate-TA with cysteine sulfinic acid (CSA) as amino
donor [64].
798
19.3.3
Asymmetric Synthesis with Amine-TA
The first groundbreaking work in this field was done in the late-1980s by the US
company Celgene [70]. In the last decade, amine-transaminases (ATA) have been
studied extensively and they have been identified in a mere dozen organism [38, 52,
71–77] – they have been biochemically characterized and also overexpressed in
microbial hosts such as E. coli with the enzyme from Vibrio fluvialis as probably the
most intensively studied ATA. Most ATA exhibit (S)-selectivity, but a few examples of
(R)-selective enzymes were also discovered [72, 75]. Some of the ATA can also convert
b-amino acids [27, 28, 29, 78]. Because most ATA show excellent stereoselectivity,
they offer presently the unique possibility of synthesizing optically active amines or
b-amino acids directly from the prostereogenic ketone with a theoretically quanti-
tative yield. Compared to deracemization with monoamine oxidase, which requires
prior synthesis of racemic amines, ATA can directly use the more readily available
ketone. Although the great potential of ATA – especially for the asymmetric synthesis
of chiral amines from ketones – was recognized many years ago, only more recently
developed strategies have allowed their efficient use.
The major limitation in asymmetric synthesis starting from prostereogenic
ketones is the unfavorable equilibrium. Over ten years ago, Kim and coworkers
reported [79] that only 0.5% a-MBA is formed from acetophenone even if a tenfold
excess of alanine serving as amine donor was used. Hence a powerful method is
needed to shift the reaction equilibrium. Thus, the initial focus of research with ATA
was on the kinetic resolution of racemic amines [41, 53, 55, 56] as here 50%
conversion can be easily achieved.
Various strategies have been developed recently to drive the reaction to completion
and to optimize the reaction by circumventing substrate and product inhibition,
which are described next.
19.3.3.1 Shifting the Equilibrium by Cyclization of the Amine Product

The equilibrium is shifted very easily in the synthesis of chiral lactams
(Scheme 19.13) [80]. After transamination of the substrate, a d-keto acid ester, the
amino group of the product displaces the ester alcohol in an intramolecular
nucleophilic substitution and a chiral lactam is formed. As the amino group is
protected in the product it is not available for the reverse reaction and, consequently,
quantitative consumption of the starting keto acid ester is observed.
O O Amine-TA NH2 O
O O - EtOH O N
H
δ-keto acid ester NH2 O δ-amino acid ester lactam
Scheme 19.13 Product cyclization drives the reaction to completion.

19.3.3.2 Shifting the Equilibrium by Removal of Coproduct
Lactate Dehydrogenase Kim and coworkers were the first to study the removal of
pyruvate with lactate dehydrogenase (LDH): Combining LDH with ATA increased
conversion in the asymmetric synthesis of a-methylbenzylamine from 0.5 to 5% [79].
Instead of using isolated enzymes and cofactor recycling, Kim et al. tried a whole-cell
approach since E. coli produces lactate dehydrogenase and other enzymes, which
consume pyruvate in additional pathways. This enabled up to 90% conversion at
27 mM product concentration. As a disadvantage, large amounts of cells had to be
used and in general the desired chiral amine might either be metabolized or be toxic
to the whole cell.
Kroutils group reinvestigated the equilibrium shift using lactate dehydrogenase
more systematically and employed it in combination with a glucose dehydrogenase
for cofactor recycling (Scheme 19.14) [75]. A range of ketones (50 mM) were
efficiently converted into the respective amines at high to quantitative conversions
and with excellent enantiomeric purities (>98% e.e.). Interestingly, aryl alkyl ketones
like acetophenone gave significantly lower conversion than other ketones with one or
two carbon atoms between the aryl substituent and the carbonyl group. This study
also showed that the addition of cosolvents may increase conversion significantly, but
(a) Equilibrium shift by enzymatic removal of pyruvate
O NH2
amine-TA
R1 R2 R1 R2
ketone amine
acetaldehyde
alanine pyruvate PDC
CO2
AADH H2O2
NAD+, H2O LDH cell acetate
NADH, NH3
CO2
NADH
NAD+ metabolites
lactate
(b) Cofactor recycling used with AADH or LDH
NAD+ NADH
HCOOH CO2
FDH
glucose gluconolactone gluconate + H+
GDH
Scheme 19.14 Equilibrium shift for asymmetric synthesis with amine-TA: (a) different methods are
based upon removal of the coproduct pyruvate; (b) cofactor recycling used for amino acid
dehydrogenase (AADH) or lactate dehydrogenase (LDH). PDC: pyruvate decarboxylase, FDH:
formate dehydrogenase, GDH: glucose dehydrogenase.
800
in some cases leads to reduced enantioselectivity. For a (R)-selective ATA (ATA-117),

high amounts (1 mM) of PLP were required to achieve high conversion. The gluconic
acid formed during cofactor regeneration from glucose also causes a decrease of pH
in the reaction medium. Hence, the reaction progress on a large scale can be easily
monitored by titration with a base.
Alanine Dehydrogenase Two other possibilities were developed recently in which

isolated enzymes were used for cofactor recycling (Scheme 19.14). Similar to a-TA
(Scheme 19.10), an elegant way is the recycling of pyruvate with ammonia, NADH,
and amino acid dehydrogenase (AADH), to alanine [74]. The formed NAD þ can be
regenerated with the well-established formate dehydrogenase (FDH) cofactor recy-
cling system. In the overall reaction, a ketone is converted with ammonium formate,
yielding optically pure amine, water, and carbon dioxide – which essentially resem-
bles an asymmetric reductive amination catalyzed by an amine dehydrogenase. In
most cases, high yields of 90–99% were obtained – only a-methylbenzylamine gave
unsatisfactory 6% conversion. Although theoretically only a small amount of alanine
should be sufficient for efficient regeneration, in practice 0.5 equivalents were
needed to reach a conversion of 56% (even after a prolonged reaction time of
68 h) and only a fivefold excess of alanine allowed high conversion within 24 h.
Pyruvate Decarboxylase Compared to the above-mentioned examples, decarboxyl-

ation of pyruvate to yield acetaldehyde and CO2 by readily available pyruvate
decarboxylase (PDC) [81, 82] is a very simple process since only one additional
enzyme and no cofactor recycling of NADH is needed and the reaction equilibrium is
irreversibly shifted due to carbon dioxide formation. Furthermore, PDCs are
commercially available or a crude extract from Zymomonas mobilis can be used.
To reach high conversions, a high alanine excess is necessary since the acetaldehyde
formed as by-product from the pyruvate decarboxylation is aminated by the amine
transaminase from Vibrio fluvialis in a side reaction and yields ethanamine. Hence,
either alternative (or engineered) transaminases not reacting with acetaldehyde have
to be developed or methods for efficient acetaldehyde removal are needed to make
this approach technically feasible.
Yeast Alcohol Dehydrogenase All the above-described approaches used alanine as

amino donor, leading to an equilibrium that favors ketone formation. In a totally
different approach other amino donors were investigated that lead to a more product-
favored equilibrium compared to alanine. In principle, chiral amines like aminoin-
dane or aminotetralin or the achiral benzylamine are excellent amino donors for
V. fluvialis ATA; although it has been shown that they can be applied in asymmetric
synthesis [83], upscaling would involve higher costs. In contrast, the inexpensive
isopropylamine [84] can be used as alternative amino donor.
The acetone so-produced can be removed by yeast alcohol dehydrogenase (YADH)
in combination with a cofactor recycling by, for example, formate dehydrogenase
(FDH, Scheme 19.15). As YADH has a very narrow substrate specificity for small
ketones, this strategy can be applied for the asymmetric synthesis of larger sized
aliphatic amines or aryl-alkylamines.
O amine-TA NH3
R1 R2 R1 R2
ketone amine
NH3 O
1) Evaporation
NADH CO2
2) Reduction FDH
YADH NAD+ HCOOH
OH
Scheme 19.15 If isopropylamine is used as amino donor, the equilibrium can be shifted either by
reduction of acetone or by evaporation. YADH: yeast alcohol dehydrogenase, FDH: formate
dehydrogenase.
Non-enzymatic Processes If isopropylamine is used as amino donor, the reaction

can alternatively be driven to completion by stripping out the formed coproduct
acetone under reduced pressure [84] since under the final process conditions,
at pH 7, acetone has a lower boiling point than the other reactants (Scheme 19.15).
Unfortunately, not all transaminases accept isopropylamine well as amino
donor.
Compared to the enzymatic decarboxylation of pyruvate, Fotheringham showed
that pyruvate or other sterically simple a-keto acids can also be efficiently decar-
boxylated by treatment with hydrogen peroxide and thus an equilibrium shift is
possible without the addition of any auxiliary enzyme (Scheme 19.14) [85]. One
problem is the reduced stability of enzymes in the presence of hydrogen peroxide,
so that either the peroxide treatment has to be carried out in a separate reaction vessel
or fresh enzyme has to be added after a certain period of time.
In summary, most strategies for shifting the equilibrium rely on the removal of
the coproduct pyruvate, since alanine is the universal amino donor for virtually all
amine-transaminases known so far. Success depends strongly on which amine is
used for asymmetric synthesis, as the equilibrium constant may vary for different
substrates. In particular, the asymmetric synthesis of aryl-alkyl amines seems to be
most difficult. Whole cells were investigated as an alternative for shifting efficiently
the equilibrium in the asymmetric synthesis with the most challenging model
substrate acetophenone. A future development may be the combination of the
described processes: whole cells that simultaneously overexpress the transaminase
and enzymes for pyruvate removal and cofactor regeneration would be the most
practical means for biocatalytic asymmetric synthesis of amines on an industrial
scale.
With respect to a favored equilibrium, simple amines like isopropylamine appear
to be the better alternative, but this strategy needs protein engineering of trans-
aminases to make it a broadly applicable concept.
802
19.3.4
Amine-TA in Action: Optimization of Reactions for Industrial Scale
Three main factors determine efficient production of fine chemicals on an industrial

scale using transaminases: final product concentration (preferred >1 M) and isolated
yield, reaction time, and cost of the biocatalyst(s).
As shown above, choosing a suitable amino donor or method for equilibrium
shift can solve the first problem. Factors influencing reaction time are on the
one hand the specific activity of the amine-TA towards the desired substrate. On
the other hand, product and substrate inhibition may significantly slow down
the reaction rate and thus diminish product yield, especially if the reaction is
upscaled to higher substrate/product concentrations desired for efficient indus-
trial processes. Therefore, two different strategies were developed to speed up the
reaction: (i) the amine-product can be bound in situ to decrease its concentration
to a value below the critical value where product inhibition becomes an issue; (ii)
protein engineering may be used for optimization of the amine-TA since enzyme
inhibition is an intrinsic property of the biocatalyst itself and hence can be altered
by modification of the biocatalyst. Protein engineering is further a powerful tool to
increase the specific activity for a desired amine and to generate highly thermo-
stable biocatalysts. Both factors will contribute significantly to keeping the cost for
biocatalysts as low as possible. Some examples of these strategies are presented
next.
19.3.4.1 In Situ Product Removal

As mentioned above, aryl-alkyl amines are especially difficult to prepare by asym-
metric synthesis, even if a powerful strategy for equilibrium shift is applied. Turner
and coworker discovered that in cases where coproduct removal is not sufficient for
an efficient equilibrium shift, absorbing the produced amine on a resin can
significantly increase reaction yield [80].
As shown in Figure 19.2, the addition of Amberlite XAD 1180 resin to the reaction
medium reduces the concentration of free 1-phenylethylamine up to about 20-fold.
Amberlite XAD 1180 is a polymeric adsorbent resin that is used primarily in the
purification and preparation of hydrophobic compounds. At a substrate concentra-
tion of 50 g l1 acetophenone, 200 g l1 resin is sufficient to decrease the product
inhibition level to below 50%. This dramatically increases the product yield
(Figure 19.2c): Without any equilibrium shift, conversion remains <1%. By com-
bination of amine-TA with lactate dehydrogenase and cofactor recycling, the con-
version does not exceed about 10% after 20 h. In contrast, acetophenone is quan-
titatively converted into 1-phenylethylamine if the resin is included. After the reaction
is finished, the amine can be retrieved from the resin by simply rinsing it with an
alkaline buffer. This demonstrates the strength of the transaminase technology, if
appropriate tricks for shifting the equilibrium as well as circumventing product
inhibition are applied.
(a) O NH2
amine-TA CH CH2 CH CH2 NH2
Amberlite
alanine pyruvate XAD 1180
glucono-
NADH CH2 CH CH2
lactone LDH
glucose NAD+ n
GDH
lactate
(b) (c) 100

50
1-phenylethylamine [g/l]
aqueous concentration
40 80
conversion [%]
30 60
TA+LDH+resin
20 40
TA+LDH
80 % inhibition TA alone
10 50 % inhibition 20
0 0
0 10 25 50 100 200 400 0 5 10 15 20
resin concentration [g/l] time [h]
Figure 19.2 Optimization of the asymmetric synthesis of 1-phenylethylamine for an industrial

scale: (a) reaction scheme; (b) impact of different amounts of resin on 1-phenylethylamine
concentration in the aqueous phase; and (c) effect of in situ product removal on yield. LDH: lactate
dehydrogenase, GDH: glucose dehydrogenase.
19.3.4.2 Protein Engineering for Increasing Activity and Thermostability

Unfortunately, there is still no crystal structure of an ATA available and hence rational
protein design is very difficult. However, directed evolution (random mutagenesis
followed by screening of mutant libraries) enabled the identification of ATA with
strongly increased specific activity or reduced substrate and product inhibition. In a
recent study, an (S)-selective ATA from Arthrobacter species was used to produce
optically pure substituted aminotetralin [86]. Although isopropylamine was used as
amino donor, the equilibrium disfavored product formation. To allow for an efficient
asymmetric synthesis, the catalyst had to be optimized to withstand high amounts of
isopropylamine and reaction temperatures of 55 C to apply acetone distillation for
equilibrium shift, and, secondly, specific activity for tetralin had to be increased to
reduce reaction time. Five rounds of random mutagenesis afforded a biocatalyst with
a temperature optimum at 55 C and a 280-fold increased turnover frequency while
kM stayed constant at about 10 mM. Substrate and product inhibition was not
investigated explicitly in this study, possibly as this phenomenon was not yet
significant at the substrate concentrations used (100–130 mM).
In a second example, the substrate specificity of a (R)-selective ATA was intensively
expanded to enable asymmetric synthesis of sitagliptin (Scheme 19.16) by protein
804
engineering [87]. Sitagliptin is used as dipeptidyl peptidase-4 inhibitor in treatment

of diabetes mellitus type 2 [88]. As described above, all amine-transaminases studied
in the literature so far posses a small and a large binding pocket for accommodation of
a large and a rather small substituent of the substrate. Thus, substrates with a small
substituent exceeding two carbon atoms are virtually not converted. In the structure
of sitagliptin, however, the small substituent consists of a trifluorobenzyl group
rather than a methyl or ethyl group (Scheme 19.16). Thus, unsurprisingly, no
conversion could be obtained by the only available (R)-selective ATA-117. To optimize
the enzyme, the company Codexis applied a substrate walking approach: In the first
step, the ATA was optimized to convert a truncated version of the substrate in which
F F Reaction conditions:
evolved
F (R)-amine-TA F
O O NH2 O
200 g/L ketone (≈ 0.5 M)
N N N N 1 M ispropylamine
N NH2 O N 50 % DMSO
F N F N
pH 8, 40 °C
CF3 92 % y, > 99.95 % eeP CF3 6 g/L catalyst
Scheme 19.16 Asymmetric synthesis of sitagliptin with an evolved (R)-selective amine-TA.
the large trifluorobenzyl group was exchanged by a methyl group (Scheme 19.16). In
an asymmetric synthesis reaction with the wild-type transaminase, conversion of the
truncated ketone was still very modest: 4% yield was obtained after 24 h for 2 g l1
substrate ketone and 10 g l1 catalyst. After a first round of mutagenesis, an eleven-
fold improved variant was obtained. With this mutant, already a conversion of 0.5% of
the complete ketone could be achieved under the same conditions. This variant was
the starting point for an extensive ProSAR [89] driven protein engineering study,
where variants showing better substrate recognition, increased substrate concentra-
tions, high organic solvent tolerance, and optimized thermostability were selected in
successive cycles of mutagenesis and screening. After the 11th round of screening, a
mutant suitable for an industrial-scale process was obtained that was able to convert
200 g l1 substrate in the presence of 50% DMSO as cosolvent. At an elevated reaction
temperature of 45 C, which allows shifting of the equilibrium through distillation
of the coproduct acetone (see example below), the (R)-enantiomer of sitagliptin was
obtained in 84% yield and perfect enantiomeric purity (as required for pharmaceu-
tical applications). This example demonstrates that it is possible to overcome
limitations of substrate specificity and to provide transaminases suitable for indus-
trial-scale processes. Compared to a known already efficient process for the prep-
aration of sitagliptin by asymmetric hydrogenation [90], the transaminase technology
has the clear advantage of a superior enantioselectivity and the avoidance of transition
metals and, hence, low cost reagents and equipment can be used [91].
19.3.4.3 Protein Engineering for Decreasing Substrate and Product Inhibition

In a very interesting directed evolution study, Matcham and coworkers focused
directly on avoiding product inhibition in the preparation of 2-amino-1-methoxy-
propane [84]. The wild-type transaminase showed significant product inhibition at
1 M methoxyacetone using 1.5 equivalents isopropylamine, which led to a conversion
of only 12%. Given an equilibrium constant of 8 for the reaction this low conversion
was not caused by an unfavorable equilibrium but rather by an inhibition or
inactivation. Thus, the transaminase was improved by three rounds of random
mutagenesis and screening and, finally, an enzyme was obtained where, owing to a
higher Ki-value for substrates and products (Table 19.5), a conversion of 65% could be
reached without any need to shift the equilibrium. An explanation for the lowered
inhibition constants may be the observed increased KM values for both substrates
(Table 19.5). As a next step, five rounds of directed evolution were carried out, yielding
a variant with enhanced thermal and chemical stability that enabled above-described
removal of acetone at 50 C under reduced pressure. With the best mutant, >99%
optically pure (S)-2-amino-1-methoxypropane was obtained from 2 M methoxyace-
tone and 2.5 M isopropylamine at 93% yield after 7 h. Since the cost of the biocatalyst
was rather low (5% of the overall cost), it could be discarded after each production
batch.
Thus protein engineering by directed evolution turned these reactions into useful
processes.
19.3.5
Scope and Limitations of Amine-TA
Various structurally different amines and amino acids can be prepared by the
transaminase technologies described above and selected examples are summarized
in Table 19.6. A more detailed summary for the substrate scope of individual enzymes
(ATA only) can be found in a recent review [92].
Table 19.5 Kinetic parameters of wild-type transaminase and an improved mutant with significantly
lower product inhibition [84].
Kinetic parameter Wild-type Mutant
Substrate inhibition Ki (M)

Methoxyacetone 0.34 1.35
Isopropylamine 1.11 >2.0
Product inhibition Ki (M)
2-Amino-1-methoxypropane 0.17 0.65
Acetone 0.70 >2.5
Substrate Km (M)
Methoxyacetone 0.023 0.21
Isopropyl amine 0.24 0.80
Vmax (mmol min1mg1) 13 73
Product yielda) (%) 12 65
Reaction conditions: 30 C, pH 8, 0.2 mM PLP, 1.0 M methoxyacetone, 1.5 M isopropylamine.

a)
806
Table 19.6 Examples of applications of transaminases.
Products Enzyme/synthesis modea) Reference
Small aliphatic amines KR, AS [55, 74, 75, 93]

NH2 NH2 NH2 Vibrio fluvialis ATA, Codexis (S)-/(R)-ATA
O COOEt
R
1-5
R=H, Me, Ph
1 2 3
Sterically demanding aliphatic amines 4: B. megaterium ATA; KR. [52, 72, 74, 75]
NH2
NH2 MeO
NH2
5: Codexis (S)-/(R)-ATA; KR.
MeO 6: Arthrobacter/Pseudomonas strains having (R)- or (S)-selective
ATA activity; AS (whole cells)
4 5 6
NH2 Codexis (S)-/(R)-ATA; DR. [50]

O
Both enantiomers accessible, dependent of the order of the applied
ATA during DR.
7
F
F
NH2 O
N N Codexis (R)-ATA-variant; AS [87]

N
F N
8 CF3
Aryl-alkyl amines (Substituted) 1-phenylethylamine is the model substrate of most [27, 33, 53, 56, 73]
(S)-ATA; some (R)-ATA convert it only very slowly. KR, AS
NH2 mutants of the (R)-selective ATA-117 also accept substrates
bearing bulkier R2-substituents. Reported are -CF3, n-Pr, n-Bu,
R1
R2 iso-Pr, -CH2-CH2-Ph. However, enantioselectivity was not
investigated in this study (88).
R1 = Me, halogen, NO2, OH
R2 = Me, Et
9
NH2 NH2 10: Codexis (S)-/(R)-ATA; AS [94]

11: Codexis (S)-ATA; KR, AS was not efficient
R
Se Se
R = Me, Et, Bu
10 11
Cyclic amines 12, 13: Celgene, Codexis ATA; KR. [40, 41, 70]
NH2
H2N
NH2 14: V. fluvialis/Alcaligenes denitrificans ATA; AS/KR; protection
N 0-1 group influences reaction rate and enhances enantioselectivity
0-1 R
R = H, Boc, Cbz, Bz
12 13 14
Amino alcohols 15: Chromobacterium violaceum ATA. [83, 95–99]
NH2 NH2
16: Vibrio fluvialis ATA, AS, KR.
OH
R No enantioselectivity towards hydroxyl group
19.3 Strategies for Using Transaminases in Biocatalysis
OH
OH (Continued )
j807
R = Me, OH, Ph
15 16
808
b amino acids A. denitrificans ATA, Mesorhizobium ATA, b-amino acids are often [27–29]
NH2 O only prepared by KR, since the substrate b-keto acid is unstable; for
NH2 O AS, in situ substrate delivery was employed by enzymatic
OH
OH hydrolysis of b-keto ester
17 18
NH2 O NH2 O
OH OH
19 20
Aromatic L-a acids Broad range TA; AS [100]

NH2 OH
NH2
COOH
F COOH
21 22
NH2 NH2 Thirty-two different thermophilic a-TA, AS of various [101]

R R
phenylglycine derivates on an analytical scale.
COOH COOH
R
mono-, di-, trisubstituted:

R = CH3, OH, OMe, F
23 24
HOOC NH2 Single mutant (Y66L) of E. coli aromatic amino acid TA; AS [102]
25
NH2
NH2
26: Enterobacter aromatic TA; AS. [57, 58]
COOH
COOH 27: Thermococcus aromatic TA; AS
26 27
Aliphatic L-a-amino acids E. coli BCAT; AS. [64]
NH2
NH2 R
HOOC COOH
HOOC COOH 28: Diastereoselectivity in position only with R ¼ Pr, Ph.
R
29: No stereoselectivity in position 4
R=Me, Et, Pr, Ph, R=Me, Et, Pr, Bz
28 29
NH2 NH2 E. coli BCAT; AS. [64]

R
HOOC COOH COOH
R=Me, OH
30 31
R NH2 E. coli Aspartate-TA; AS. [62, 63, 103, 104]

NH2
COOH 32: High enantioselectivity towards position 4
HOOC
HOOC COOH
R=H, Me, Et, Ph, OH,
OPh, COOH, COOMe,
CONH2, CONH-alkyl
32 33
D-amino acids E. coli L-aromatic amino acid TA; KR led to D-configuration; [105]
19.3 Strategies for Using Transaminases in Biocatalysis
NH2 N N equilibrium shift by excess amino acceptor

R N N N
R=
j809
COOH N N
(Continued )
34
810
NH2 S NH2 AS with D-amino acid TA from Bacillus sphaericus and Bacillus sp. [106–108]
R
YM-1. B. sphaericus TA shows the broader substrate specificity.
COOH COOH
R = Me, Et, Pr
35 36
O
D-Met, D-Ala, D-Asp,
NH2 D-Asn, D-Phe, D-Leu

N
H D-His, D-Arg
COOH
37 38
NH2
Cbz Transamination of 39 with Sphingomonas paucimobilis e-lysine-TA [109]
Cbz O 3 HN
O leads to the aldehyde, which undergoes spontaneous cyclization to
HN COOH 40, an omapatrilat precursor
N COOH
H N
α-KG Glu S
SH
39 40
15
Isotopically labeled amino acids N-Labeled amino acids are accessible through AS with 15N- [110]
15 labeled glutamate as amino donor, which can be produced from 2-
NH2
ketoglutarate and 15N-labeled ammonium sulfate, catalyzed by
COOH glutamate dehydrogenase
HO
41
a) KR: kinetic resolution, AS: asymmetric synthesis, ATA: amine transaminase, DR: deracemization, and BCAT: branched chain amino acid transaminase.
19.3.5.1 Enantioselectivity
Usually, high enantioselectivity is observed for the chiral carbon atom bearing the
amino group involved in transamination. In cases where only a low enantioselectivity
is observed, protein engineering might be useful for optimization of the enzyme. For
amine transaminases, a selection strategy based on the substrate 1-phenyl-n-propy-
lamine (PPA) as the sole source of nitrogen in a chemostat was applied with a
recombinant Pseudomonas putida strain carrying an (R)-ATA gene [70]. A single
amino acid change, Y112F, presumably at or near the active site, improved enantios-
electivity of the reaction of racemic 1-phenyl-n-propylamine to (S)-1-phenyl-n-pro-
pylamine and propiophenone to 37.8% e.e. from 6.5% e.e. in the wild type. Further
site-directed mutagenesis of position 112 yielded a mutant which allowed the
preparation of the desired (S)-amine with 99.4 % e.e.
In contrast, no or little enantioselectivity is observed for additional stereogenic
centers in beta-position, as in the case of amino alcohols [83, 99]. To prepare amino
alcohols with high diastereomeric excess, the problem of the low enantioselectivity
for the b-hydroxy substituent can be circumvented by the application of a two-enzyme
cascade reaction. In the first step, transketolase generates the a-keto-alcohol with
high enantioselectivity, which then serves as substrate for amine-TA in the second
step (Scheme 19.17a) [97, 98]. Alternatively, the enantiomerically pure a-hydroxy-
ketone may be prepared by kinetic resolution with lipase, as exemplified on an
analytical scale for 1-amino-2-indanol (Scheme 19.17b) [84].
Only one example has been described recently in which significant enantioselec-
tivity was observed in a transamination where the chiral center is at the b-position of
the transferred amino group (Scheme 19.17c) [111]. The aldehyde function of 3-phenyl
substituted succinate semi-aldehyde ester is aminated by an amine-TA, and the
generated amino group displaces the ester alcohol by an intramolecular nucleophilic
substitution leading to the cyclic 4-phenylpyrrolidine-2-one. The semi-aldehyde
substrate racemizes spontaneously, thus providing the opportunity for a dynamic
kinetic resolution. Several transaminases were investigated for enantioselectivity at 3-
position bearing the phenyl substituent. After 24 h, conversions >95% were reached
in most cases and the enantiomeric excess of the formed product varied between 6 and
68% e.e. Additionally, it was investigated whether the enantiopreference of an (S)-ATA
can be reversed by means of protein engineering to yield an (R)-selective enzyme.
Indeed, one variant carrying a single point mutation could be identified by rational
design, which shows (R)-selectivity towards 4-fluorophenylacetone. Interestingly,
other ketones were still converted with (S)-selectivity [112]. Hence, a more complex
approach is needed for a general change of the enantiopreference.
19.3.5.2 Substrate Scope

Most wild-type ATA can convert efficiently only ketones with one small and one larger
substituent next to the carbonyl group. A methyl group is usually the best accepted
small substituent, but often the reaction rate drops significantly when the methyl
group is replaced by ethyl or hydroxymethyl. Furthermore, substrates that can act as
mechanism-based inhibitors of transaminases cannot be used. These include, for
example, b-halogene, propargyl-, or vinyl ketones [113].
812 j 19 Application of Transaminases
(a)
O O NH2
O transketolase amine-TA
+ OH OH
R LiO OH R R
O OH OH
R = Me, OH i-propyl- acetone
amine
(b)
AS with
O lipase O NH2
(S)-ATA
trans-(1R,2R)-1-amino-2-indanol
OAc OH OH
reductive
amination
KR with
NH2 NH2
(S)-ATA
cis-(1S,2R)-1-amino-2-indanol
OH OH
(c)
O NH
(R)-amine TA O
CO2Et
D-Ala pyruvate
spont. NADH- 92 % yield
LDH 68 %ee
recycling
O lactate
CO2Et
Scheme 19.17 Use of amine-TA for synthesizing amine compounds with stereogenic centers at the
b-position in respect to the amino group. An enantioselective transketolase (a) or lipase (b) might be
combined with an ATA to provide amino alcohols with high diastereomeric excess. (c) The moderate
enantioselectivity of ATA allowed the preparation of an enantioenriched lactam by dynamic kinetic
resolution. See text for details.
To broaden substrate specificity, protein engineering might be a valuable tool, but

there are only a few examples (Section 19.3.4). Furthermore, changing substrate
specificity might not always be as straightforward as in other cases. This can be seen
from some studies dealing with alteration of the substrate specificity of a-TA, which
are described below.
Using directed evolution, the substrate specificity of an aspartate-TA has been
changed to one favoring branched amino acids and their respective oxoacids,
effectively converting aspartate-TA into a branched-chain aminotransferase (BCAT).
The evolved TA had 17 amino acid substitutions [114]. Interestingly, only one
mutated amino acid residue is located at a distance from the substrate that would
allow direct interactions; the other mutated residues were far from the active site.
Aspartate-TA had been transformed into an L-tyrosine aminotransferase (TAT) by
site-specific mutation of up to six amino acid residues lining the active site, which are,
19.4 Conclusions j813
however, not in direct contact with the substrate of wild-type TAT. The hextuple
aspartate-TA-mutant achieved kinetic data (vmax/kM) towards the transamination
of aromatic substrates such as L-phenylalanine within an order of magnitude of wild-
type TAT [115, 116]. One conclusion drawn from such investigations is that although
modification of substrate specificity is possible, it is not trivial. Furthermore, it seems
that combinatorial or evolutionary methods are probably superior to rational design
methods when changing substrate specificity, and remote residues and their inter-
actions with the active site environment are important determinants of enzyme
activity and specificity. Such remote residues act cumulatively, possibly by remodel-
ing the active site, by altering the subunit interfaces, or by a slight movement of
different enzyme domains leading to an altered orientation of crucial residues.
19.3.5.3 Enzyme Availability

Since L-amino acid aminotransferases are ubiquitous enzymes required for many basic
metabolic functions, there is a huge source of enzymes that were and can be further
explored for applications in organic chemistry. Therein, a considerable number of
thermostable enzymes exist, which are interesting for industrial applications because
of their usually high stability. Recently, 32 novel thermostable a-TAs were cloned [101].
Conversely, only a handful of D-amino acid aminotransferase were purified and
investigated. The reason for the lower occurrence of D-amino acid aminotransferases
might be because there are fewer metabolic functions for these enzymes, for example,
providing D-amino acids for cell-wall synthesis and secondary metabolites.
In addition, ATAs are found relatively seldom compared to L-amino acid TA, and
their physiological role is not yet understood. It might be speculated that they play a role
in degradation of biogenic amines, for example, histamine, tryptamine, and the like.
Since these compounds are not chiral, in principle there is no need for enantioselec-
tivity. Furthermore, this would explain why most enzymes found in nature only convert
amines with one small substituent, as found in biogenic amines, where one hydrogen
atom represents the small substituent (!). In the scientific literature, about 12, mainly
(S)-selective, amine-TAs were described, and about 20 more (S)-selective enzymes
were recently cloned and characterized [117] (although further enzymes were discov-
ered by industry that are not available to the public). The recent discovery of about
17 novel (R)-selective amine-TAs [33] closes an important gap, since these enzymes
will facilitate asymmetric synthesis of the (R)-enantiomer of a given amine [118].
19.4
Conclusions
Transaminase technology represents a powerful approach for efficient preparation of

optically pure amines and amino acids. The previously important limitation of the
unfavored equilibrium in asymmetric synthesis has been solved by various strate-
gies, for example, by application of multiple-enzyme reactions. However, for the
development of industrial-scale processes, protein engineering using directed evo-
lution is often necessary. Recent examples given in this chapter demonstrate that by
814
means of protein engineering a range of important limitations can be overcome: The

activity towards desired substrates, for example, amines with bulky substituents, can
be dramatically increased. Enhanced thermostability, solvent tolerance, and
decreased product and substrate inhibition mean that efficient processes with high
product concentrations, in the molar range, can be achieved. The 3D-structures of
amine-transaminases are currently being solved [119] and will certainly facilitate
further protein engineering.
References
1 von Bohlen und Halbach, O. and 12 Breuer, M., Ditrich, K., Habicher, T.,
Dermietzel, R. (2006) Neurotransmitters Hauer, B., Keßeler, M., St€ urmer, R., and
and Neuromodulators: Handbook of Zelinski, T. (2004) Industrial methods for
receptors and biological effects, Wiley-VCH the production of optically active
Verlag GmbH, Weinheim, Germany. intermediates. Angew. Chem. Int. Ed., 43,
2 Bielekova, B. and Martin, R. (2001) 788–824.
Antigen-specific immunomodulation via 13 Bornscheuer, U.T. and Kazlauskas, R.J.
altered peptide ligands. J. Mol. Med., 79, (2006) Hydrolases in Organic Synthesis,
552–565. 2nd edn, Wiley-VCH Verlag GmbH,
3 Katsara, M., Minigo, G., Plebanski, M., Weinheim.
and Apostolopoulos, V. (2008) The good, 14 Turner, N., Fotheringham, I., and
the bad and the ugly: how altered peptide Speight, R. (2004) Novel biocatalyst
ligands modulate immunity. Exp. Opin. technology for the preparation of chiral
Biol. Therapy, 8, 1873–1884. amines. Innov. Pharm. Technol., 4,
4 Hancock, J.T. (2010) Cell Signalling, 114–116.
Oxford University Press, New York. 15 K€ohler, V., Bailey, K.R., Znabet, A.,
5 Ma, J.S. (2003) Unnatural amino acids in Raftery, J., Helliwell, M., and Turner, N.J.
drug discovery. Chim. Oggi, 21, 65–68. (2010) Enantioselective biocatalytic
6 Gracia, S.R., Gaus, K., and Sewald, N. oxidative desymmetrization of
(2009) Synthesis of chemically modified substituted pyrrolidines. Angew. Chem.
bioactive peptides: recent advances, Int. Ed., 49, 2182–2184.
challenges and developments for 16 Carr, R., Alexeeva, M., Enright, A.,
medicinal chemistry. Fut. Med. Chem., 1, Eve, T.S.C., Dawson, M.J., and
1289–1310. Turner, N.J. (2003) Directed evolution
7 Nestor, J.J. (2009) The medicinal of an amine oxidase possessing both
chemistry of peptides. Curr. Med. Chem., broad substrate specificity and high
16, 4399–4418. enantioselectivity. Angew. Chem. Int. Ed.,
8 Seebach, D. and Gardiner, J. (2008) 42, 4807–4810.
Beta-peptidic peptidomimetics. Acc. 17 De Wildeman, S.M.A., Sonke, T.,
Chem. Res., 41, 1366–1375. Schoemaker, H.E., and May, O. (2007)
9 Witt, K.A., Gillespie, T.J., Huber, J.D., Biocatalytic reductions: From lab
Egleton, R.D., and Davis, T.P. (2001) curiosity to first choice. Acc. Chem. Res.,
Peptide drug modifications to enhance 40, 1260–1266.
bioavailability and blood-brain barrier 18 Weiner, B., Poelarends, G.J., Janssen,
permeability. Peptides, 22, 2329–2343. D.B., and Feringa, B.L. (2008) Biocatalytic
10 H€ohne, M. and Bornscheuer, U.T. (2009) enantioselective synthesis of N-
Biocatalytic routes to optically active substituted aspartic acids by aspartate
amines. ChemCatChem, 1, 42–51. ammonia lyase. Chem. Eur. J., 14,
11 Downton, C. and Clark, I. (2003) 10094–10100.
Statins – the heart of the matter. Nat. Rev. 19 Wu, B., Szymanski, W., Wietzes, P.,
Drug Discov., 2, 343–344. de Wildeman, S., Poelarends, G.J.,
References j815
Feringa, B.L., and Janssen, D.B. (2009) Identification of v-aminotransferase
Enzymatic synthesis of enantiopure from Caulobacter crescentus and
alpha- and beta-amino acids by site-directed mutagenesis to broaden
phenylalanine aminomutase-catalysed substrate specificity. J. Microbiol.
amination of cinnamic acid derivatives. Biotechnol., 18, 48–54.
ChemBioChem, 10, 338–344. 30 Jansonius, J.N. (1998) Structure,
20 Burton, S.G. and Dorrington, R.A. (2004) evolution and action of vitamin B6-
Hydantoin-hydrolysing enzymes for the dependent enzymes. Curr. Opin. Struct.
enantioselective production of amino Biol., 8, 759–769.
acids: new insights and applications. 31 Percudani, R. and Peracchi, A. (2009)
Tetrahedron: Asymmetry, 15, 2737–2741. The B6 database: a tool for the description
21 Fesko, K., Uhl, M., Steinreiber, J., and classification of vitamin B6-
Gruber, K., and Griengl, H. (2010) dependent enzymatic activities and of the
Biocatalytic access to a,a-dialkyl-a-amino corresponding protein families. BMC
acids by a mechanism-based approach. Bioinformatics, 10, 273.
Angew. Chem. Int. Ed., 49, 121–124. 32 Schneider, G., Kack, H., and Lindqvist, Y.
22 Gotor-Fernandez, V. and Gotor, V. (2009) (2000) The manifold of vitamin B-6
Biocatalytic routes to chiral amines and dependent enzymes. Structure, 8, R1–R6.
amino acids. Curr. Opin. Drug Discov. 33 H€ohne, M., Sch€atzle, S., Jochens, H.,
Dev., 12, 784–797. Robins, K., and Bornscheuer, U.T. (2010)
23 Moore, J. and Langley, R. (2008) Rational assignment of key motifs
Biochemistry for Dummies, John Wiley & for function guides in silico enzyme
Sons, Inc., Hoboken, NJ. identification, Nature Chem. Biol.,
24 Needham, D.M. (1930) A quantitative 6, 807–813
study of succinic acid in muscle. 34 Goldberg, J.M. and Kirsch, J.F. (1996)
Glutamic and aspartic acids as The reaction catalyzed by Escherichia coli
precursors. Biochem. J., 24, 208–227. aspartate aminotransferase has multiple
25 Annau, E., Banga, I., Blazso, S., partially rate-determining steps, while
Bruckner, V., Laki, K., Straub, F.B., and that catalyzed by the Y225F mutant is
Gyorgyi, A. (1936) The significance of dominated by ketimine hydrolysis.
fumaric acid for the respiration of animal Biochemistry, 35, 5280–5291.
tissue. Z. Physiol. Chem., 244, 105–152. 35 Metzler, D.E. (2001) Biochemistry,
26 Metha, P.K., Hale, T.I., and Christen, P. Academic Press, San Diego.
(1993) Aminotransferases: 36 Mizuguchi, H., Hayashi, H., Okada, K.,
demonstration of homology and Miyahara, I., Hirotsu, K., and
division into evolutionary subgroups. Kagamiyama, H. (2001) Strain is more
Eur. J. Biochem., 214, 549–561. important than electrostatic interaction
27 Yun, H., Lim, S., Cho,B.-K.,and Kim, B.-G. in controlling the pK(a) of the catalytic
(2004) v-Amino acid:pyruvate group in aspartate aminotransferase.
transaminase fromAlcaligenes denitrificans Biochemistry, 40, 353–360.
Y2k-2: a new catalyst for kinetic resolution 37 Frey, P.A. and Hegemann, P. (2007)
of b-amino acids and amines. Appl. Enzymatic Reaction Mechanism, Oxford
Environ. Microbiol., 70, 2529–2534. University Press, New York.
28 Kim, J., Kyung, D., Yun, H., Cho, B.K., 38 Shin, J. and Kim, B. (2002) Substrate
Seo, J.H., Cha, M., and Kim, B.G. (2007) inhibition mode of v-transaminase from
Cloning and characterization of a novel Vibrio fluvialis JS17 is dependent on the
b-transaminase from Mesorhizobium sp. chirality of substrate. Biotechnol. Bioeng.,
strain LUK: a new biocatalyst for the 77, 832–837.
synthesis of enantiomerically pure 39 Weinberger, R. (2000) Practical Capillary
b-amino acids. Appl. Environ. Microbiol., Electrophoresis, 2nd edn, Academic Press,
73, 1772–1782. San Diego.
29 Bum-Yeol, H., Ko, S.H., Park, H.Y., 40 H€ohne, M., K€ uhl, S., Robins, K., and
Seo, J.H., Lee, B.S., and Kim, B.G. (2008) Bornscheuer, U.T. (2008) Efficient
816
asymmetric synthesis of chiral amines by v-transaminases. Eur. J. Org. Chem.,

combining transaminase and pyruvate 2289–2292.
decarboxylase. ChemBioChem, 9, 50 Koszelewski, D., Pressnitz, D., Clay, D.,
363–365. and Kroutil, W. (2009) Deracemization of
41 H€ohne, M., Robins, K., and mexiletine biocatalyzed by omega-
Bornscheuer, U.T. (2008) A protection transaminases. Org. Lett., 11, 4810–4812.
strategy substantially enhances rate and 51 Servi, S., Tessaro, D., and Pedrocchi-
enantioselectivity in transaminase- Fantoni, G. (2008) Chemo-enzymatic
catalyzed kinetic resolutions. Adv. Synth. deracemization methods for the
Catal., 350, 807–812. preparation of enantiopure non-natural
42 Schadewaldt, P. and Adelmeyer, F. (1996) alpha-amino acids. Coord. Chem. Rev.,
Coupled enzymatic assay for estimation 252, 715–726.
of branched-chain L-amino acid 52 Hanson, R.L., Davis, B.L., Chen, Y., and
aminotransferase activity with 2-oxo acid Goldberg, A.L. (2008) Preparation of (R)-
substrates. Anal. Biochem., 238, 65–71. amines from racemic amines with an (S)-
43 Luong, T.N. and Kirsch, J.F. (1997) A amine transaminase from Bacillus
continuous coupled spectrophotometric megaterium. Adv. Synth. Catal., 350,
assay for tyrosine aminotransferase 1367–1375.
activity with aromatic and other 53 Shin, J.-S., Kim, B.-G., Liese, A., and
nonpolar amino acids. Anal. Biochem., Wandrey, C. (2001) Kinetic resolution of
253, 46–49. chiral amines with v-transaminase using
44 Hwang, B.-Y. and Kim, B.-G. (2004) High- an enzyme-membrane reactor.
throughput screening method for the Biotechnol. Bioeng., 73, 179–187.
identification of active and 54 Truppo, M.D., Turner, N.J., and
enantioselective v-transaminases. Rozzell, J.D. (2009) Efficient kinetic
Enzyme Microb. Technol., 34, 429–436. resolution of racemic amines using a
45 Sch€atzle, S., H€ohne, M., Redestad, E., transaminase in combination with an
Robins, K., and Bornscheuer, U.T. (2009) amino acid oxidase. Chem. Commun.,
Rapid and sensitive kinetic assay for 2127–2129.
characterization of v-transaminases. 55 Yun, H., Cho, B.-K., and Kim, B.-G.
Anal. Chem., 81, 8244–8248. (2004) Kinetic resolution of
46 Sch€atzle, S., H€ohne, M., Robins, K., (R,S)-sec-butylamine using omega-
and Bornscheuer, U.T. (2010) transaminase from Vibrio fluvialis JS17
Conductometric method for the rapid under reduced pressure. Biotechnol.
characterization of the substrate Bioeng., 87, 772–778.
specificity of amine-transaminases. Anal. 56 Shin, J.-S. and Kim, B.-G. (1997) Kinetic
Chem., 82, 2082–2086. resolution of a-methylbenzylamine with
47 Truppo, M.D., Rozzell, J.D., Moore, J.C., v-transaminase screened from soil
and Turner, N.J. (2009) Rapid screening microorganisms: application of a
and scale-up of transaminase catalysed biphasic system to overcome product
reactions. Org. Biomol. Chem., 7, inhibition. Biotechnol. Bioeng., 55,
395–398. 348–358.
48 Yun, H., Hwang, B.Y., Lee, J.H., and 57 Hanzawa, S., Oe, S., Tokuhisa, K.,
Kim, B.G. (2005) Use of enrichment Kawano, K., Kobayashi, T., Kudo, T., and
culture for directed evolution of the Kakidani, H. (2001) Chemo-enzymatic
Vibrio fluvialis JS17 omega-transaminase, synthesis of 3-(2-naphthyl)-L-alanine by
which is resistant to product inhibition by an aminotransferase from the extreme
aliphatic ketones. Appl. Environ. thermophile, Thermococcus profundus.
Microbiol., 71, 4220–4224. Biotechnol. Lett., 23, 589–591.
49 Koszelewski, D., Clay, D., Rozzell, D., 58 Cho, B.-K., Seo, J.-H., Kang, T.-W., and
and Kroutil, W. (2009) Deracemisation Kim, B.-G. (2003) Asymmetric synthesis
of a-chiral primary amines by a one-pot, of L-homophenylalanine by equilibrium-
two-step cascade reaction catalysed by shift using recombinant aromatic
References j817
L-amino acid transaminase. Biotechnol. herbicide phosphinothricin
Bioeng., 83, 226–234. (glufosinate): purification of aspartate
59 Rozzell, D. (1987) Immobilized transaminase from Bacillus
aminotransferases for amino acid stearothermophilus, cloning of the
production. Methods Enzymol., 136, corresponding gene, aspC, and
479–497. application in a coupled transaminase
60 Crump, S.P., Heier, J.S., and Rozzell, J.D. process. Appl. Environ. Microbiol., 62,
(1990) The production of amino acids by 3794–3799.
transamination, in Biocatalysis (ed. D.A. 68 Ager, D.J., Li, T., Pantaleone, D.P.,
Abramowicz), Van Nostrand Reinhold, Senkpeil, R.F., Taylor, P.P., and
New York, pp. 133–155. Fotheringham, I.G. (2001) Novel
61 Furumo, N.C. and Kirsch, J.F. (1995) biosynthetic routes to non-proteinogenic
Accumulation of the quinonoid amino acids as chiral pharmaceutical
intermediate in the reaction catalyzed by intermediates. J. Mol. Catal. B, 11,
aspartate-aminotransferase with cysteine 199–205.
sulfinic acid. Arch. Biochem. Biophys., 69 Fotheringham, I.G., Grinter, N.,
319, 49–54. Pantaleone, D.P., Senkpeil, R.F., and
62 Helaine, V., Rossi, J., Gefflaut, T., Alaux, Taylor, P.P. (1999) Engineering of a novel
S., and Bolte, J. (2001) Synthesis of 4,4- biochemical pathway for the biosynthesis
disubstituted L-glutamic acids by of L-2-aminobutyric acid in Escherichia
enzymatic transamination. Adv. Synth. coli K12. Bioorg. Med. Chem., 7,
Catal., 343, 692–697. 2209–2213.
63 Faure, S., Jensen, A.A., Maurat, V., Gu, X., 70 Matcham, G.W. and Bowen, A.R.S.
Sagot, E., Aitken, D.J., Bolte, J., (1996) Biocatalysis for chiral
Gefflaut, T., and Bunch, L. (2006) intermediates: meeting commercial
Stereoselective chemoenzymatic and technical challenges. Chim. Oggi,
synthesis of the four stereoisomers of 14, 20–24.
L-2-(2-carboxycyclobutyl)glycine and 71 Chen, D., Wang, Z., Zhang, Y., Sun, Z.,
pharmacological characterization at and Zhu, Q. (2008) An amine:
human excitatory amino acid transporter hydroxyacetone aminotransferase from
subtypes 1, 2, and 3. J. Med. Chem., 49, Moraxella lacunata WZ34 for alaninol
6532–6538. synthesis. Bioprocess. Biosyst. Eng., 31,
64 Xian, M., Alaux, S., Sagot, E., and 283–289.
Gefflaut, T. (2007) Chemoenzymatic 72 Iwasaki, A., Yamada, Y., Ikenaka, Y., and
synthesis of glutamic acid analogues: Hasegawa, J. (2003) Microbial synthesis
Substrate specificity and synthetic of (R)- and (S)-3,4-
applications of branched chain dimethoxyamphetamines through
aminotransferase from Escherichia coli. stereoselective transamination.
J. Org. Chem., 72, 7560–7566. Biotechnol. Lett., 25, 1843–1846.
65 Bae, H.S., Lee, S.G., Hong, S.P., 73 Kaulmann, U., Smithies, K.,
Kwak, M.S., Esaki, N., Soda, K., and Smith, M.E.B., Hailes, H.C., and
Sung, M.H. (1999) Production of Ward, J.M. (2007) Substrate spectrum
aromatic D-amino acids from of v-transaminase from
alpha-keto acids and ammonia by Chromobacterium violaceum DSM30191
coupling of four enzyme reactions. J. Mol. and its potential for biocatalysis.
Catal. B, 6, 241–247. Enzyme Microb. Technol.,
66 Fotheringham, I.G., Pantalone, D.P., 41, 628–637.
and Taylor, P.F. (1997) Biocatalytic 74 Koszelewski, D., Lavandera, I.,
production of unnatural amino acids, Clay, D., Guebitz, G.M., Rozzell, D., and
mono esters, and N-protected derivatives. Kroutil, W. (2008) Formal asymmetric
Chim. Oggi, 15, 33–36. biocatalytic reductive amination,
67 Bartsch, K., Schneider, R., and Schulz, A. Angew. Chem. Int. Ed., 47,
(1996) Stereospecific production of the 9337–9340.
818
75 Koszelewski, D., Lavandera, I., Clay, D., engineering in biocatalysis. Synthesis of

Rozzell, D., and Kroutil, W. (2008) (S)-methoxyisopropylamine (¼ (S)-1-
Asymmetric synthesis of optically pure methoxypropan-2-amine). Chimia, 53,
pharmacologically relevant amines 584–589.
employing v-transaminases. Adv. Synth. 85 Fotheringham, I. (2008). A method to
Catal., 350, 2761–2766. increase the yield and improve
76 Shin, J.-S. and Kim, B.-G. (2001) purification of products from
Comparison of the v-transaminases transaminase reactions Patent
from different microorganisms and application WO2008108998.
application to production of chiral 86 Martin, A.R., DiSanto, R., Plotnikov, I.,
amines. Biosc. Biotechnol. Biochem., 65, Kamat, S., Shonnard, D., and Pannuri, S.
1782–1788. (2007) Improved activity and
77 Yonaha, K., Toyama, S., Yasuda, M., and thermostability of (S)-aminotransferase
Soda, K. (1976) Purification and by error-prone polymerase chain reaction
crystallization of bacterial omega-amino for the production of a chiral amine.
acid-pyruvate aminotransferase. FEBS Biochem. Eng. J., 37, 246–255.
Lett., 71, 21–24. 87 Postlethwaite, S. (2010) Evolving a novel
78 Banerjee, A., Chase, M., Calyton, A., and transaminase to enable a commercially
Landis, B. (2005) Patent application viable biocatalytic route to sitagliptin.
WO2008108998. Presented at Novel Enzymes Conference
79 Shin, J.-S. and Kim, B.-G. (1999) 2010, Exeter.
Asymmetric synthesis of chiral amines 88 Dhillon, S. (2010) Sitagliptin a
with v-transaminase. Biotechnol. Bioeng., review of its use in the management of
65, 206–211. type 2 diabetes mellitus. Drugs, 70,
80 Truppo, M.D., Rozzell, J.D., and 489–512.
Turner, N.J. (2010) Efficient production 89 Fox, R.J., Davis, S.C., Mundorff, E.C.,
of enantiomerically pure chiral amines at Newman, L.M., Gavrilovic, V., Ma, S.K.,
concentrations of 50g/L using Chung, L.M., Ching, C., Tam, S.,
transaminases. Org. Process Res. Dev., 14, Muley, S., Grate, J., Gruber, J.,
234–237. Whitman, J.C., Sheldon, R.A.,
81 Raj, K.C., Talarico, L.A., Ingram, L.O., and Huisman, G.W. (2007) Improving
and Maupin-Furlow, J.A. (2002) catalytic function by ProSAR-driven
Cloning and characterization of the enzyme evolution. Nat. Biotechnol., 25,
Zymobacter palmae pyruvate 338–344.
decarboxylase gene (pdc) and 90 Hansen, K.B., Yi, H., Xu, F., Rivera, N.,
comparison to bacterial homologues. Clausen, A., Kubryk, M., Krska, S.,
Appl. Environ. Microbiol., 68, 2869–2876. Rosner, T., Simmons, B., Balsells,
82 Goetz, G., Iwan, P., Hauer, B., Breuer, M., J., Ikemoto, N., Sun, Y., Spindler, F.,
and Pohl, M. (2001) Continuous Malan, C., Grabowski, E.J.J., and
production of (R)-phenylacetylcarbinol in Armstrong, J.D. (2009) Highly
an enzyme-membrane reactor using a efficient asymmetric synthesis of
potent mutant of pyruvate decarboxylase sitagliptin. J. Am. Chem. Soc., 131,
from Zymomonas mobilis. Biotechnol. 8798–8804.
Bioeng., 74, 317–325. 91 Desai, A.A. (2011), Sitagliptin
83 Yun, H., Kim, J., Kinnera, K., and manufacture: A compelling tale of green
Kim, B.G. (2005) Synthesis of chemistry, process intensification, and
enantiomerically pure trans-(1R,2R)- industrial asymmetric catalysis, Angew.
and cis-(1S,2R)-1-amino-2-indanol by Chem. Int. Ed., 50, 1974–1976.
lipase and v-transaminase. Biotechnol. 92 Koszelewski, D., Tauber, K., Faber, K., and
Bioeng., 93, 391–395. Kroutil, W. (2010) v-Transaminases for
84 Matcham, G., Bhatia, M., Lang, W., the synthesis of non-racemic a-chiral
Lewis, C., Nelson, R., Wang, A., and primary amines. Trends. Biotechnol., 28,
Wu, W. (1999) Enzyme and reaction 324–332.
References j819
93 Shin, J.-S. and Kim, B.-G. (2002) of Amino Acids and Derivatives: New
Exploring the active site of amine: Developments and Process Considerations
pyruvate aminotransferase on the (eds J.D. Rozzell and F. Wagner), Hanser
basis of the substrate structure-reactivity Publishers, Munich, pp. 43–58.
relationship: how the enzyme controls 101 Koma, F.D., Sawai, T., Hara, R.,
substrate specificity and stereoselectivity. Harayama, S., and Kino, K. (2008) Two
J. Org. Chem., 67, 2848–2853. groups of thermophilic amino acid
94 Andrade, L.H., Silva, A.V., Milani, P., aminotransferases exhibiting broad
Koszelewski, D., and Kroutil, W. (2010) substrate specificities for the synthesis of
omega-Transaminases as efficient phenylglycine derivatives. Appl. Environ.
biocatalysts to obtain novel chiral Microbiol., 79, 775–784.
selenium-amine ligands for 102 Cho, B.K., Seo, J.H., Kang, T.J., Kim, J.,
Pd-catalysis. Org. Biomol. Chem., Park, H.Y., Lee, B.S., and Kim, B.G. (2006)
8, 2043–2051. Engineering aromatic L-amino acid
95 Dalby, P.A., Baganz, F., Lye, G.J., and transaminase for the asymmetric
Ward, J.M. (2009) Protein and pathway synthesis of constrained analogs of
engineering in biocatalysis. Chim. Oggi, L-phenylalanine. Biotechnol. Bioeng., 94,
27, 18–29. 842–850.
96 Hailes, H.C., Dalby, P.A., Lye, G.J., and 103 Sagot, E., Pickering, D.S., Pu, X.,
Ward, J.M. (2009) Biocatalytic approaches Umberti, M., Stensbol, T.B., Nielsen, B.,
to ketodiols and aminodiols. Chim. Oggi, Chapelet, M., Bolte, J., Gefflaut, T., and
27, 28–31. Bunch, L. (2008) Chemo-enzymatic
97 Ingram, C.U., Bommer, M., Smith, synthesis of a series of 2,4-syn-
M.E.B., Dalby, P.A., Ward, J.M., functionalized (S)-glutamate analogues:
Hailes, H.C., and Lye, G.J. (2007) new insight into the structure-activity
One-pot synthesis of amino-alcohols relation of ionotropic glutamate receptor
using a de-novo transketolase and subtypes 5, 6, and 7. J. Med. Chem., 51,
beta-alanine:pyruvate transaminase 4093–4103.
pathway in Escherichia coli. Biotechnol. 104 Gu, X., Xian, M., Roy-Faure, S., Bolte, J.,
Bioeng., 96, 559–569. Aitken, D.J., and Gefflaut, T. (2006)
98 Smith, M.E.B., Chen, B.H., Synthesis of the constrained glutamate
Hibbert, E.G., Kaulmann, U., analogues (2S,10 R,20 R)- and (2S,10 S,20 S)-
Smithies, K., Galman, J.L., Baganz, F., 2-(20 -carboxycyclobutyl)glycines L-CBG-
Dalby, P.A., Hailes, H.C., Lye, G.J., II and L-CBG-I by enzymatic
Ward, J.M., Woodley, J.M., and transamination. Tetrahedron Lett., 47,
Micheletti, M. (2010) A multidisciplinary 193–196.
approach toward the rapid and 105 Cho, B.-K., Park, H.Y., Seo, J.-H.,
preparative-scale biocatalytic synthesis of Kinnera, K., Lee, B.S., and Kim, B.G.
chiral amino alcohols: a concise (2004) Enzymatic resolution for the
transketolase-/omega-transaminase- preparation of enantiomerically enriched
mediated synthesis of (2S,3S)-2- D-b-heterocyclic alanine derivatives
aminopentane-1,3-diol. Org. Process Res. using Escherichia coli aromatic L-amino
Dev., 14, 99–107. acid transaminase Biotechnol. Bioeng., 88,
99 Smithies, K., Smith, M.E.B., 512–519.
Kaulmann, U., Galman, J.L., Ward, J.M., 106 Yonaha, K., Misono, H., Yamamoto, T.,
and Hailes, H.C. (2009) Stereoselectivity and Soda, K. (1975) D-Amino acid
of an omega-transaminase-mediated aminotransferase of Bacillus sphaericus.
amination of 1,3-dihydroxy-1- J. Biol. Chem., 250, 6983–6989.
phenylpropane-2-one. Tetrahedron: 107 Tanizawa, K., Masu, Y., Asano, S.,
Asymmetry, 20, 570–574. Tanaka, H., and Soda, K. (1989)
100 Crump, S.P. and Rozzell, J.D. (1992) Thermostable D-amino-acid
The production of amino acids by aminotransferase from a thermophilic
transamination. in Biocatalytic Production Bacillus species - purification,
820
characterization, and active-site sequence 113 Nanavati, S. and Silverman, R.B.

determination. J. Biol. Chem., 264, (1989) Design of potential
2445–2449. anticonvulsant agents: mechanistic
108 Gutierrez, A., Yoshimura, T., classification of GABA
Fuchikami, Y., and Esaki, N. (2000) aminotransferase inactivators.
Modulation of activity and substrate J. Med. Chem., 32, 2413–2421.
specificity by modifying the backbone 114 Oue, S., Okamoto, A., Yano, T., and
length of the distant interdomain loop Kagamiyama, H. (1999) Redesigning the
of D-amino acid aminotransferase. substrate specificity of an enzyme by
Eur. J. Biochem., 267, 7218–7223. cumulative effects of the mutations of
109 Patel, R.N., Banerjee, A., Nanduri, V.B., non-active site residues. J. Biol. Chem.,
Goldberg, S.L., Johnston, R.M., 274, 2344–2349.
Hanson, R.L., McNamee, C.G., 115 Shaffer, W.A., Luong, T.N.,
Brzozowski, D.B., Tully, T.P., Ko, R.Y., Rothman, S.C., and Kirsch, J.F. (2002)
LaPorte, T.L., Cazzulino, D.L., Quantitative chimeric analysis of six
Swaminathan, S., Chen, C.K., specificity determinants that differentiate
Parker, L.W., and Venit, J.J. (2000) Escherichia coli aspartate from tyrosine
Biocatalytic preparation of a chiral aminotransferase. Protein Sci., 11,
synthon for a vasopeptidase inhibitor: 2848–2859.
enzymatic conversion of N-2-[N- 116 Malashkevich, V.N., Onuffer, J.J.,
phenylmethoxy)carbonyl] L- Kirsch, J.F., and Jansonius, J.N. (1995)
homocysteinyl]-L-lysine (1-> 10 )- Alternating arginine-modulated substrate
disulfide to [4S-(4I,7I,10aJ)] 1-octahydro- specificity in an engineered tyrosine
5-oxo-4-[phenylmethoxy)carbonyl] aminotransferase. Nat. Struct. Biol., 2,
amino]-7H-pyrido-[2,1-b] [1,3] 548–553.
thiazepine-7-carboxylic acid methyl ester 117 Ward, J.M. (2010) Developments in
by a novel L-lysine epsilon- screening for novel enzymes:
aminotransferase. Enzyme Microb. transketolase and transaminase -
Technol., 27, 376–389. colorimetric screening, bioinformatics
110 Chanatry, J.A., Schafer, P.H., Kim, M.S., and metagenomics. Presented at Novel
and LeMaster, D.M. (1993) Synthesis of Enzymes 2010, Conference, Exeter.
a,b-deuterated 15N amino acids using a 118 Schätzle, S., Steffen-Munsberg, F.,
coupled glutamate dehydrogenase- Thontowi, A., Höhne, M., Robins, K. and
branched chain amino acid Bornscheuer, U.T. (2011) Enzymatic
aminotransferase system. Anal. Biochem., asymmetric synthesis of
213, 147–151. enantiomerically pure aliphatic,
111 Koszelewski, D., Clay, D., Faber, K., and aromatic and arylaliphatic amines with
Kroutil, W. (2009) Synthesis of 4- (R)-selective amine transaminases. Adv.
phenylpyrrolidin-2-one via dynamic Synth. Catal., DOI: 10.1002/
kinetic resolution catalyzed by omega- adsc.201100435
transaminases. J. Mol. Catal. B, 60, 119 Sayer, C., Isupov, M.N., and
191–194. Littlechild, J.A. (2007) Crystallization
112 Svedendahl, M., Branneby, C., Lindberg, and preliminary X-ray diffraction
L., Berglund, P. (2010) Reversed analysis of omega-amino acid:pyruvate
enantiopreference of an v-transaminase transaminase from Chromobacterium
by a single-point mutation. violaceum. Acta Crystallogr., Sect. F, 63,
ChemCatChem, 2, 976–980 117–119.
j821
20
Industrial Applications and Processes Using Enzymes Acting
on C–N Bonds
Ruslan Yuryev, Lutz Hilterhaus, and Andreas Liese
20.1
Introduction
Nitrogen-containing organic compounds were, are, and will be always of great

interest to the chemical industry. They are found in everyday use commodities like
food, cosmetics, and polymers, and are also important fine chemicals like pharma-
ceuticals and agrochemicals, whose market capacities vary from several kilos to
>10 000 tons per year. Organic nitrogen is also an indispensable part of any living
system: nitrogen compounds are building blocks for two vital classes of biopolymers
– nucleic acids and proteins – as well as many other biomolecules, for example,
hormones, neurotransmitters, cytokines, and antibiotics. Therefore, nature has
developed a wide arsenal of enzymes, for instance, proteases, amidases, C–N lyases,
transaminases and many more, which can act on C–N bonds in very efficient
manners. These enzymes are relatively robust and in general do not require
expensive cofactors, like NAD(P)H or ATP, which makes them very attractive for
practical applications. Unsurprisingly, people started to use them in prehistoric
times. The first application of C–N bond acting enzymes was probably in cheese
making, dated as early as 3000 BCE when ancient Arabs discovered that milk is
turned into cheese if it is stored in a pouch made from the animal stomach excreting
rennet, a cocktail of proteases. Much later, in 1913, these enzymes became fore-
runners of industrial enzymology when Otto R€ ohm patented the use of a crude
protease mixture extracted from pancreases in laundry detergents. Since then the
chemical industry has tried to employ enzymes acting on C–N bonds as biocatalysts
for the production and transformation of nitrogen-containing compounds. Processes
based on these enzymes belong to the oldest established industrial biotransforma-
tions: the biocatalytic production of L-aspartic acid by Tanabe Seiyaku Co. (Japan) was
launched at 1958, and in 1970 another Japanese company, Asahi Chemical Industry,
succeeded in applying penicillin amidase for the production of 7-ADCA (7-amino-
deacetoxycephalosporanic acid). Since then the number of industrial processes
involving C–N acting enzymes has drastically increased, and in many cases they
have even replaced the respective classical chemical routes. This chapter gives an
and Oliver May.
j 20 Industrial Applications and Processes Using Enzymes Acting on C–N Bonds
822
overview of application of C–N acting enzymes on an industrial scale. The following

examples represent not only heavyweight processes used for multi-ton production
of bulk chemicals but also those performed by industrial companies to synthesize
compounds in kg amounts for research or clinical trials [1].
20.2
Hydration of Nitriles to Amides
The hydration of nitriles to amides is a classic functional group transformation in

organic synthesis. Despite widespread use of this reaction on both laboratory and
industrial scales, it is often a challenging task to perform it in efficient manner,
because it is difficult to avoid the subsequent hydrolysis of formed amides to their
corresponding acids due to the rather harsh reaction conditions. Although nowadays
many chemical catalysts can hydrate selectively nitriles to amides, from an industrial
point of view nitrile hydratases (EC 4.2.1.) offer benign green alternatives to the
chemical routes because the biotransformations proceed at ambient conditions,
generate less salt and organic waste, and do not require complex downstream
processing to remove traces of chemocatalysts from the end products.
DuPont (USA) produces 5-cyanovaleramide, an intermediate of DuPonts herbi-
cide azafenidin, on a multi-ton per annum scale by selective hydration of adipodini-
trile catalyzed by nitrile hydratase from Pseudomonas chlororaphis B23 (Scheme 20.1).
The process is performed in repetitive batches in a 2300-l reactor containing the
biocatalyst in a form of whole cells immobilized in calcium alginate beads. During
strain selection it was important that the cells did not show amidase activity that
would further hydrolyze the amide to the carboxylic acid. The product is obtained in
93% yield at 97% conversion and 96% selectivity, while the catalyst productivity is
3150 kg of product per kg of dry cell weight and the catalyst consumption is 0.006 kg
per kg product. During the biotransformation adipodiamide is formed as a main by-
product, which, however, is easily separated from 5-cyanovaleramide by precipitation
when dissolving the crude product in methanol at >65 C. Owing to higher
conversion, selectivity, and less waste production the biocatalytic hydration of
adipodinitrile is economically more attractive than the respective chemical
transformation [2].
O
E
CN
NC NC NH2
+ H2O
Scheme 20.1 Reaction scheme for DuPonts process for the production of 5-cyanovaleramide by
hydration of adipodinitrile catalyzed by nitrile hydratase from Pseudomonas chlororaphis B23 (E) [1].
Nitrile hydratase from Rhodococcus rhodochrous J1 is utilized by Lonza AG

(Switzerland) in the production of nicotinamide from 3-cyanopyridine (Scheme 20.2).
In contrast to the chemical alkaline hydrolysis with formation of nicotinic acid (4%
yield) as a by-product, the process catalyzed by immobilized whole cells works with
20.2 Hydration of Nitriles to Amides j823
CN CONH2
E
+ H2O
N N
Scheme 20.2 Reaction scheme for Lonza AGs process for the production of nicotinamide
by hydration of 3-cyanopyridine catalyzed by nitrile hydratase from Rhodococcus rhodochrous
J1 (E) [1].
absolute chemoselectivity, allowing a yield of 100% to be reached. Using this

biotransformation the company produces 6000 t a1 of nicotinamide, which is
mainly used as a vitamin supplement for food and animal feed [3].
Another company, Nitto Chemical Industry Co., Ltd. (now part of Mitsubishi
Rayon, Japan), applies a nitrile hydratase from Rhodococcus rhodochrous J1 in the
biocatalytic hydration of acrylonitrile to produce >30 000 t a1 of acrylamide, which is
an important bulk chemical used in coagulators, soil conditioners and stock additives
for paper treatment and sizing, and for adhesives, paints, and petroleum recovering
agents. This biotransformation is the first example of applying enzymes in the
petrochemical industry to manufacture bulk chemicals. The reaction is catalyzed by
immobilized whole cells in a polyacrylamide gel, showing catalyst productivity of
>7000 kg of acrylamide per kg of cell dry weight (Figure 20.1). To prevent possible
polymerization of acrylamide the process is carried out at about 5 C. Although the
nutrients
inducing agent
CN
H2O
fermentation
medium
analytics
cells
E
immobilization
NH2 spent
decoloring cells
O
Figure 20.1 Flow scheme for the process of Nitto Chemical Industry Co., Ltd. used to produce
acrylamide by hydrolysis of acrylonitrile catalyzed by nitrile hydratase from Rhodococcus rhodochrous
J1 (E) [1].
824
cells and the enzyme are very stable towards acrylonitrile, the educt has to be fed
continuously to the reaction mixture due to inhibition effects at higher concentra-
tions. The biocatalytic hydration of acrylonitrile proceeds with >99.99% conversion,
selectivity, and yield, and, therefore, has replaced the corresponding chemical
process involving copper salts as a hydration catalyst [4].
20.3
Hydrolysis of Nitriles to Acids
Hydrolysis of nitriles is often considered as one of the most attractive ways to obtain
carboxylic acids. When the hydrolysis is carried out chemically, harsh conditions
(extreme pH and high temperature) are usually required for the reaction to take place.
However, in this case the process performance is restricted by formation of unwanted
by-products and by troublesome product contamination with salts, which are formed
in large quantities after neutralization of either acid or base used to attain a complete
reaction. These drawbacks can be partially overcome by applying metal-based
chemocatalysts, but the most elegant and environmentally friendly solution would
be to employ nitrilases (EC 3.5.5.1) catalyzing the hydrolysis of nitriles to acids in one
step under mild conditions and with high chemoselectivity. Furthermore, if nitriles
have chiral centers, the biocatalytic hydrolysis often runs with high enantioselectivity
and, therefore, it could be a convenient industrially relevant route to enantiopure
carboxylic acid.
DuPont have developed a two-step chemoenzymatic process for the production of
1,5-dimethyl-2-piperidone (Xolvone) – a precious cleaning solvent widely applied in
electronics and coatings industries (Scheme 20.3). In the first biocatalytic step
2-methylglutaronitrile is selectively hydrolyzed to 4-cyanopentanonic acid ammoni-
um salt by immobilized, in alginate, whole cells of Escherichia coli expressing
Acidovorax facilis 72 W nitrilase. Finally, the acid is chemically hydrogenated over
Pd/C in the presence of methylamine to yield 1,5-dimethyl-2-piperidone. During the
biocatalytic hydrolysis of 2-methylglutaronitrile 100% conversion, 98% selectivity,
and 98.7% yield are achieved, while the catalyst productivity is 3500 gproduct gcatalyst1.
The chemoenzymatic route to Xolvone replaced the solely chemical process, which
employed direct hydrogenation of 2-methylglutaronitrile in the presence of methyl-
amine and which produced a mixture of 1,3- and 1,5-dimethyl-2-piperidones [5].
Nitrilase from the strain E. coli JM (pDHE19.2) was used by BASF AG (Germany)
for enantioselective hydrolysis of racemic mandelonitrile (Scheme 20.4) to produce,
O
CN E CO2-NH4+
N
CN + 2 H2O CN
Scheme 20.3 Hydrolysis of 2-methylglutaronitrile catalyzed by Acidovorax facilis 72 W nitrilase (E)

as the first step of DuPonts process for the production of 1,5-dimethyl-2-piperidone (Xolvone) [1].
20.3 Hydrolysis of Nitriles to Acids j825
OH OH OH
CN E CO2H CN
+
+ 2 H2O
- NH3
Scheme 20.4 Reaction scheme for BASF AGs production of (R)-mandelic acid by kinetic
resolution of racemic mandelonitrile with nitrilase from E. coli JM (E) [1].
on a multi-ton scale, enantiomerically pure (>99% e.e.) (R)-mandelic acid – a

valuable synthetic intermediate commonly used for pharmaceutical synthesis [6].
Hydrolysis of 2-cyanopyrazine to pyrazine-2-carboxylic acid catalyzed by a nitrilase
from the strain Agrobacterium sp. DSM 6336 is the first reaction in a two-step in vivo
reaction cascade route to 5-hydroxypyrazine-2-carboxylic acid (Scheme 20.5), which is
a versatile building block for the synthesis of new antitubercular agents, for example,
5-chloropyrazine-2-carboxylic acid esters. The process was commercialized by Lonza
AG and is conducted on a multi-kilogram scale. The reaction sequence is catalyzed by
suspended living whole cells grown on a different substrate, 2-cyanopyridine,
because in this case the cells are much more active due to an optimized expression
of the second enzyme hydroxylase catalyzing selective hydroxylation of pyrazine-2-
carboxylic acid. After the biotransformation the biomass is separated by ultrafiltration
(cutoff 10 kDa) and 5-hydroxypyrazine-2-carboxylic acid is precipitated from the
permeate by acidification with sulfuric acid to pH 2.5. The lower practical yield of
80% in comparison to the analytical yield of 95% is due to re-precipitation of the
product during downstream processing. In contrast to the biotransformation the
chemical synthesis of 5-substituted pyrazine-2-carboxylic acid leads to a mixture of 5-
and 6-substituted isomers and requires multiple steps [7].
Lonza AG also utilizes another microbial strain, Alcaligenes faecalis DSM 6335, to
produce 1 t a1 of 6-hydroxypicolinic acid – an intermediate for pharmaceuticals, for
example, 2-oxypyrimidine, and herbicides – using 2-cyanopyridine as starting
material (Scheme 20.6). The strain contains two enzymes, a nitrilase that catalyzes
hydrolysis of 2-cyanopyridine to picolinic acid and a hydroxylase that catalyzes
selective hydroxylation of the intermediate picolinic acid to 6-hydroxypicolinic acid.
The biotransformation is carried out under aerobic conditions with a resting whole-
cell biocatalyst. Since the intermediate picolinic acid inhibits the second reaction of
the cascade, the educt 2-cyanopyridine has to be maintained at a low concentration
N CN N COOH N COOH
E1 E2 (EC 1.5.1.13)
N + 2 H2O N +½ O2 HO N
- NH3
Scheme 20.5 Reaction scheme for Lonza AGs production of 5-hydroxypyrazine-2-carboxylic acid
from 2-cyanopyrazine by a two-step biotransformation catalyzed by nitrilase (E1) and hydroxylase
(E2) from Agrobacterium sp. DSM 6336 [1].
826
E1 E2
+ 2 H2O +½ O2
N CN N COOH HO N COOH
- NH3
Scheme 20.6 Reaction scheme for Lonza AGs production of 6-hydroxypicolinic acid from 2-
cyanopyridine by a two-step biotransformation catalyzed by nitrilase (E1) and hydroxylase (E2) from
Alcaligenes faecalis DSM 6335 [1].
level. Therefore, 2-cyanopyridine is continuously fed to the reaction solution and its
feed rate is controlled by on-line analysis of the picolinic acid concentration. To
precipitate the product, the cells are removed from the reaction solution and the pH is
adjusted to 2.5 using sulfuric acid at 60 C. During the process 100% conversion and
87% yield are achieved [8].
20.4
Hydrolysis and Formation of Amides
The technical application of amidohydrolases (EC 3.5.) in hydrolysis or formation of

amides is mainly predetermined by two intrinsic properties of these biocatalysts –
high chemo- and stereoselectivity. Owing to high chemoselectivity, enzymatic
hydrolysis or formation of amides proceeds in a very controlled manner even in
the presence of other labile functional groups. To achieve such a selectivity in
chemical routes tedious protection/deprotection chemistry is often required, which
has a negative impact on the overall process economy. Industrial production of
7-aminocephalosporanic (7-ACA) and 6-aminopenicillanic (6-APA) acids by biocat-
alytic hydrolysis of the respective natural b-lactam antibiotics is a classic example
illustrating the exceptional chemoselectivity of hydrolases during their action on C–N
bonds: during hydrolysis the enzymes cleave only one, the more stable, of two present
amide groups, whereby the labile b-lactam rings remain intact.
Sanofi (France) employs glutaryl amidase from Escherichia coli in the production of
200 t a1 of 7-ACA, which is an intermediate for semisynthetic cephalosporins. Natural
antibiotic cephalosporin C obtained by fermentation of the filamentous fungus Acre-
monium chrysogenum is used as a starting material for the process. First, this antibiotic is
deaminated by D-amino-acid oxidase from Trigonopsis variabilis to N-a-ketoadipinyl-7-
aminocephalosporanic acid. In the second stage the acid is decarboxylated by
the reaction with H2O2 to glutaryl-7-aminocephalosporanic acid, which is finally
hydrolyzed by glutaryl amidase to 7-ACA and glutaric acid (Figure 20.2). The glutaryl
HO
H
N S E H2N S
O O N O N O
O O
COOH O COOH O
Figure 20.2 Flow scheme of the process of Sanofi for the production of 7-ACA by hydrolysis of
glutaryl-7-aminocephalosporanic acid catalyzed by glutaryl amidase from Escherichia coli (E) [1].
20.4 Hydrolysis and Formation of Amides j827
NaOH H
pH HO N S
O O N O
O
E COOH O
H2N S
N O
O
COOH O
Figure 20.3 Flow scheme of the process of Asahi Kasei Corporation and Toyo Jozo for the
production of 7-ACA by hydrolysis of glutaryl-7-aminocephalosporanic acid catalyzed by glutaryl
amidase from Pseudomonas GK-16 (E) [1].
amidase is immobilized on a spherical carrier and can be reused many times.

Owing to environmental considerations, the enzymatic production of 7-ACA has
replaced the multistep chemical process involving the use of heavy-metal salts
(ZnCl2), chlorinated hydrocarbons, and highly flammable compounds. For every
metric ton of 7-ACA produced enzymatically, the waste-gas emission has been
reduced from 7.5 to 1.0 kg. Mother liquors requiring incineration have been
reduced from 29 to 0.3 t, while the consumption of zinc (1.8 t) has been eliminated
completely [9].
Enzymatic hydrolysis of glutaryl-7-aminocephalosporanic acid has also been
established by two Japanese companies, Asahi Kasei Corporation and Toyo Jozo
(now part of Asahi Kasei), for the production of 90 t a1 of 7-ACA. In their process the
glutaryl amidase from Pseudomonas GK-16 strain is used (Figure 20.3). The enzyme is
immobilized by adsorption onto a porous styrene anion-exchange resin and subse-
quently crosslinked with 1% glutaraldehyde. The process is performed in repetitive
batches in a set-up consisting of a stirred-tank reactor filled with the aqueous reaction
medium and a fixed-bed reactor packed with the immobilized enzyme. The glutaric
acid liberated during hydrolysis is an inhibitor of the glutaryl amidase and addi-
tionally lowers the pH. Therefore, the pH of the reaction medium is adjusted by
feeding the reactor with NaOH and is controlled by an autotitrator. The process is
started at 15 C, but to compensate for enzyme deactivation during the reaction the
temperature is gradually increased to 25 C. After 70 cycles the enzyme has to be
replaced. The product yield is 95% [10].
The Indian company Dr. Vig Medicaments manufactures 300 t a1 of 7-amino-
deacetoxycephalosporanic acid (7-ADCA), a derivative of 7-ACA, by selective hydro-
lysis of cephalosporin G catalyzed by E. coli penicillin acylase (Figure 20.4). The
828
NH3
pH
H
N S
E
O N CH2Cl2
O
COOH
extraction crystallization
COOH
centrifugation
H2N S
N
O
COOH
Figure 20.4 Flow scheme of the process of Dr. Vig Medicaments for the production of 7-ADCA
by hydrolysis of cephalosporin G catalyzed by E. coli penicillin acylase (E) [1].
process is performed in repetitive batches using the immobilized enzyme, which has
a specific activity of 1000 U g1. The biocatalyst is retained in the reaction vessel by a
filter sieve installed at the bottom. During the biotransformation the pH in the reactor
is controlled by a pH-stat and is adjusted by feeding aqueous NH3. After conversion
reaches 99%, the formed by-product phenylacetic acid is extracted from the reaction
mixture with dichloromethane; the product is then crystallized by acidification and is
separated from the mother liquor by centrifugation. The process runs with 93% yield,
94% selectivity, and 450 U kg1 enzyme consumption (C.B. Vig, personal commu-
nication, 1999).
A similar process using E. coli penicillin acylase was launched by Unifar (Turkey)
for the production of 300 t a1 of 6-APA from penicillin G (Figure 20.5). Production
is carried out in a repetitive batch mode. The enzyme is immobilized on a polymeric
carrier, Eupergit -C (R€ohm, Germany), and is retained by a sieve with a mesh size
of 400. The initial specific activity of the biocatalyst is 800 U g1 of the dry carrier,
but after 800 batch cycles, which is one production campaign, it decreases by about
50%. The biotransformation proceeds to 98% conversion with >99% selectivity. At
the end of the reaction phenylacetic acid is removed by extraction and 6-APA is
crystallized and isolated in 86% yield with 99% purity. The enzyme consumption in
this process is 345 U per kg of the product (D. Kr€amer, and C. Boller, personal
communication, 1998).
NH3
pH
H
N
E
S
O N
O
COOH
extraction crystallization
H2N S
N
O
COOH
Figure 20.5 Flow scheme of the process of Unifar for the production of 6-APA by hydrolysis of
penicillin G catalyzed by penicillin acylase from E. coli (E) [1].
Asahi Kasei Corporation designed another process layout for the manufacture of 6-
APA (Figure 20.6). Production is carried out batchwise in a recirculation reactor
consisting of 18 parallel 30-l columns packed with Bacillus megaterium penicillin
H
N S
O N
O
COOH
NaOH
pH
E E E E
18 x
H2N S
6-APA N
O
COOH
Figure 20.6 Flow scheme of the process of Asahi Kasei Corporation for the production of 6-APA by
hydrolysis of penicillin G catalyzed by penicillin acylase from Bacillus megaterium (E) [1].
830
NH2
fermentation
medium H2N S NH2
N O
O
COOH
7-ADCA PGA
cells
E
preparation of bioreactor with
immobilized enzyme (O) special sieve
NH2
H
N S
recycling of
D-(-)-PGA
O N
O
COOH
PhCHO
cefalexin
Figure 20.7 Flow scheme of the DSM process for the production of cefalexin by acylation of
7-ADCA with D-phenylglycine amide (PGA) catalyzed by penicillin acylase (E) [1].
acylase immobilized on aminated porous polyacrylonitrile fibers. The reaction

solution is recirculated through the columns with a flow rate of 6000 l h1 and its
pH is maintained at 8.4 by feeding aqueous NaOH during the process. The lifetime of
each column is 360 batches. The obtained 6-APA is purified by isoelectric precip-
itation at pH 4.2 with subsequent filtration and washing with methanol, resulting in
86% yield and >98% purity [11].
The synthetic potential of penicillin acylase is also realized in the production of
b-lactam antibiotics operated worldwide by DSM on 2000 t a1 scale. The biocatalytic
process replaced the established ten-step chemical synthesis, which started from
benzaldehyde and penicillin G and generated a waste stream of 30–40 kg waste per kg
of product. The waste contained organic solvents, silylating agents, and many by-
products from side-chain protection and acylating promoters. In comparison, the
chemoenzymatic route needs only six steps (including three biocatalytic ones). The
production of cefalexin was the first successful application (Figure 20.7). To reach a
non-equilibrium concentration of the antibiotic during the enzyme-catalyzed acyl-
ation of 7-ADCA, the acylation agent D-phenylglycine is activated as an amide and
used in excess. For the process the enzyme is immobilized on particles with defined
diameter, which facilitates separation of the biocatalyst from the solid reaction
products. After biotransformation the reaction solution and solid substances are
removed from the reactor using a special sieve that is not permeable to the
immobilized enzyme. Separation and recycling of the non-converted D-phenylglycine
fermentation NH2
medium
H2N S NH2
N O
O
COOH
6-APA
cells base
E
preparation of
immobilized enzyme (O) pH
acid
bioreactor with
special sieve
NH2
H
N S
ampicillin O N
O
COOH
Figure 20.8 Flow scheme of the DSM process for the production of ampicillin by acylation
of 6-APA with D-phenylglycine amide catalyzed by penicillin acylase (E) [1].
is achieved by addition of benzaldehyde and formation of the poorly soluble Schiff

base, which is subsequently filtered off.
For ampicillin synthesis (Figure 20.8), the acylation of 6-APA has to be complete
and the product has to be recovered rapidly by crystallization, because the penicillanic
acid derivatives are more prone to degradation than cephalosporanic acid derivatives
at almost all pH values. In the ampicillin process the biocatalyst can be retained in the
reactor by the sieve method analogous to the cephalexin procedure. At the end of the
reaction the precipitated product and unreacted D-phenylglycine crystals are dis-
solved at acidic pH, and then pure ampicillin is precipitated by adjusting the pH to its
isoelectric point.
In a similar manner amoxicillin is produced (Figure 20.9). The advantage in this
case is the low solubility of the antibiotic under reaction conditions, so that hydrolysis
of the product is suppressed since it precipitates first. Constant removal of the
product by filtration and feeding of the substrates allows the production of amox-
icillin to be performed in a semicontinuous mode. Owing to the intrinsically high
stereo- and chemoselectivity of penicillin acylase the biocatalytic synthesis of anti-
biotics proceeds with >95% selectivity, >99% e.e., and >90% yield [12].
Besides the high chemoselectivity, the second industrially relevant property of
amidohydrolases is their high enantioselectivity during conversion of C–N bonds.
This property is widely exploited by many companies in the kinetic resolution of
racemates to produce several valuable enantiomerically pure nitrogen-containing
compounds.
832
fermentation
medium NH2
H2N S
NH2
N
O O
COOH HO
cells
E
preparation of
immobilized enzyme (O)
bioreactor with
special sieve
NH2
H
N S
amoxicillin
O N
HO O
COOH
Figure 20.9 Flow scheme of the DSM process for the production of amoxicillin by acylation
of 6-APA with D-p-hydroxyphenylglycine amide catalyzed by penicillin acylase (E) [1].
Kinetic resolution of racemic amino acid amides catalyzed by L-aminopeptidase

from Pseudomonas putida ATCC 12633 is a platform process established by DSM to
produce various enantiomerically pure natural and non-natural amino acids (>99%
e.e.) on the tons scale (Figure 20.10). The substrates for this biotransformation can be
readily obtained from the appropriate aldehydes via Strecker synthesis and subse-
quent hydrolysis of the corresponding a-aminonitriles under alkaline conditions in
the presence of a catalytic amount of ketone with yields >90%. The kinetic resolution
is achieved by permeabilized whole cells. At the end of the process the cells are
separated from the reaction mixture and benzaldehyde is added, so that the Schiff
base of the D-amide precipitates and can be easily isolated by filtration and trans-
formed into the corresponding D-amino acid by hydrolysis. The remaining L-amino
acid can be converted into the respective amide, racemized, and then used again as
substrate in the next batch. By applying such a resolution–racemization–recycling
process it is possible to obtain a theoretical yield of 100%. The same process can be
used for the synthesis of L-amino acids by racemizing the Schiff base of the D-amide
in a short time using catalytic amounts of base in organic solvents. The produced
amino acids are sold as intermediates for the synthesis of antibiotics, injectables, and
insecticides or as food and feed additives [13].
Hydrolysis of amides catalyzed by enantioselective amidases is one of the core
technologies of the Swiss company Lonza AG used in the production of valuable fine
chemicals. The company produces >1 t a1 of (S)-2,2-dimethylcyclopropanecarbox-
fermentation NH2
medium
NH2
R
O racemization
conversion into
E amide precursor
cells
NH2
PhCHO
NaOH OH
R
O
spent
cells
NH2
NH2
R
O
Figure 20.10 Flow scheme of the DSM process for the production of enantiomerically pure
natural and non-natural amino acids by hydrolysis of amides catalyzed by L-aminopeptidase from
Pseudomonas putida ATCC 12633 (E) [1].
amide, an intermediate for the synthesis of the dehydropeptidase-inhibitor cilastatin,

by kinetic resolution of the racemate catalyzed by the amidase from Comamonas
acidovorans A18 DSM:6351, expressed in Escherichia coli XL1Blue/pCAR6
(Scheme 20.7). The process is performed with suspended whole cells in the presence
of 5 vol.% ethanol, which is added to increase the reaction rate and selectivity. The
product is purified by extraction, electrodialysis, or drying and isolated in 44% yield
and 98.6% e.e. [14].
In another Lonza process the amidase from the strain Klebsiella terrigena DSM 9174
is used as a catalyst for the kinetic resolution of piperazine-2-carboxamide to prepare
(S)-piperazine-2-carboxylic acid (Scheme 20.8), which is used as an intermediate for
pharmaceuticals, for example, the orally active HIV protease inhibitor crixivan from
2 E
NH2 OH + NH2 + NH3
+ H2O
O O O
Scheme 20.7 Reaction scheme of the Lonza AG process employed for the production of
(S)-2,2-dimethylcyclopropanecarboxamide by kinetic resolution of racemate catalyzed by amidase
from Comamonas acidovorans A18 (E) [1].
834
H H H
N N N
E
2 + + NH3
NH2 + H2O OH NH2
N N N
H H H
O O O
Scheme 20.8 Reaction scheme of the Lonza AG process for the production of (S)-piperazine-2-
carboxylic acid by kinetic resolution of piperazine-2-carboxamide catalyzed by amidase from
Klebsiella terrigena (E) [1].
Merck, and is also a precursor of numerous bioactive compounds. This kinetic

resolution is attractive because the starting material can be easily prepared by catalytic
hydrogenation of pyrazine-2-carboxamide. The process is conducted batchwise on a
multi-kg scale with suspended whole cells. After completion of conversion the cells
are removed by centrifugation and the supernatant is concentrated tenfold at 60 C
under reduced pressure. The product is then precipitated by acidifying with conc.
HCl (pH 1). At the end of the process 41% yield and 99.4% e.e. are achieved [15].
Enantioselective hydrolysis of racemic 2-(trifluoromethyl)-2-hydroxypropanamide
by recombinant Klebsiella oxytoca PRS1 amidase expressed in Escherichia coli is
utilized by Lonza in the production of (R)-3,3,3-trifluoro-2-hydroxy-2-methylpropa-
noic acid (Scheme 20.9). The amidase gene was cloned into E. coli to improve safety
and productivity of the biotransformation, and additionally to avoid the slime capsule
problem associated with the original bacterial strain. In the process, suspended
whole cells are employed as a biocatalyst. After the conversion the cells are removed
by microfiltration and the residual proteins by ultrafiltration through a 70-kDa
membrane. The filtrate is concentrated by thin-film evaporation, the (S)-amide is
extracted with ethyl acetate, and the (R)-acid with methyl tert-butyl ether (MTBE). If
necessary, the (S)-amide could then be hydrolyzed chemically to (S)-3,3,3-trifluoro-2-
hydroxy-2-methylpropionic acid. With the help of this process the company produces
the (R)-acid on a 100-kg scale with practically theoretical yield and almost 100%
enantiopurity. The product is used as an intermediate for the synthesis of several
potential pharmaceuticals, including ATP-sensitive potassium channel openers for
treatment of incontinence and inhibitors of pyruvate dehydrogenase kinase for
treatment of diabetes [16].
N-Acetyl-L-amino-acid amidohydrolase from Aspergillus niger was used by Celltech
Group plc (UK; now part of UCB, Belgium) as a biocatalyst for the kinetic resolution
of N-acetyl-D,L-3-(4-thiazolyl)alanine (Scheme 20.10). The L-enantiomer of this
unnatural amino acid mimics histidine and is used as a component of antihyper-
OH OH OH
E
2 + + NH3
F3C CONH2 + H2O F3C COOH F3C CONH2
Scheme 20.9 Reaction scheme of the Lonza AG process for the production of (R)-3,3,3-trifluoro-
2-hydroxy-2-methylpropanoic acid by kinetic resolution of 2-(trifluoromethyl)-2-
hydroxypropanamide catalyzed by amidase from Klebsiella oxytoca PRS1 (E) [1].
COOH COOH COOH
O
E
2 N NHAc N NH2 + + N NHAc
+ H2O OH
S S S
Scheme 20.10 Reaction scheme of the Celltech Group plc process for the production of
L-3-(4-thiazolyl)alanine by kinetic resolution of N-acetyl-D,L-3-(4-thiazolyl)alanine catalyzed by
N-acetyl-L-amino-acid amidohydrolase from Aspergillus niger (E) [1].
tensive renin inhibitors. The racemic substrate is obtained chemically from

4-chloromethylthiazole by reaction with diethyl 2-(acetylamino)malonate followed
by partial hydrolysis and decarboxylation. The kinetic resolution is carried out on a
several-kg scale using a packed bed reactor with the immobilized enzyme. Finally, the
product is extracted directly from the aqueous reaction mixture with MTBE in high
enantiomeric purity (>99% e.e.). The remaining D-isomer can be recycled via an
oxazolinone that tautomerizes to the respective enol. The immobilized enzyme is re-
used several times [17].
Evonik (formerly Degussa AG, Germany) produces >300 t a1 of L-methionine
(80% yield, 99.5% e.e.) by means of enantioselective hydrolysis of racemic N-acetyl-
methionine catalyzed by N-acyl-L-amino-acid amidohydrolase from Aspergillus oryzae
(Figure 20.11). The kinetic resolution is performed in a continuously operated
enzyme membrane reactor loaded with the solubilized enzyme, which is retained
by a polyamide ultrafiltration membrane having a cutoff of 10 kDa. To increase the
operational stability of the aminohydrolase, Co2 þ is added as an effector. The product
L-methionine is recovered by crystallization, because it is much less soluble than the
substrate. The unconverted acetyl-D-methionine is racemized under alkaline condi-
tions and then recycled. Several other proteinogenic and non-proteinogenic amino
acids are produced in the same way by the company, and are marketed as additives for
parenteral nutrition (infusion solutions), feed and food, cosmetics, agrochemicals,
S COOH
E
crystallization racemization
HN
Figure 20.11 Flow scheme of the Evonik process for the production of L-methionine by
enantioselective hydrolysis of racemic N-acetyl-methionine catalyzed by N-acyl-L-amino-acid
amidohydrolase from Aspergillus oryzae (E) [1].
836
acetone
E
crystallization
Figure 20.12 Flow scheme of the process of Celltech Group plc for the production of ()-4-amino-
cyclopent-2-enecarboxylic acid by kinetic resolution of cyclic lactam 2-azabicylo[2.2.1]hept-5-en-3-
one using b-lactamhydrolase from Aureobacterium sp. (E) [1].
and as intermediates for pharmaceuticals as well as chiral synthons for organic

synthesis [18].
Enantiopure ()-4-aminocyclopent-2-enecarboxylic acid, a building block in the
synthesis of carbocyclic nucleosides in natural configuration, was produced by
Celltech Group plc from the racemic cyclic lactam 2-azabicylo[2.2.1]hept-5-en-3-one
using b-lactamhydrolase from Aureobacterium sp. (Figure 20.12). Kinetic resolution
of the amide is realized in a batch mode wherein an aqueous solution of the substrate
is circulated through the fixed bed of the enzyme immobilized on a glutaraldehyde-
activated solid support. Control of pH during the process is not required, since the
product 2-amino-cyclopent-2-ene carboxylic acid acts as its own buffer. The reaction is
complete when the ()-enantiomer is hydrolyzed completely (E-value >7000). The
stability of the enzyme is improved by immobilization so that a nearly steady-state
production can be achieved for more than six months. To separate the amino acid
from the unconverted ( þ )-lactam the solution at the end of hydrolysis is slurried with
acetone. The ( þ )-lactam remains in the solution while the amino acid crystallizes
and is removed by filtration [19].
In the analog process for the ()-lactam the company applied (þ)-specific
b-lactamhydrolase from Pseudomonas solanacearum (Scheme 20.11). Herein enan-
tioselective hydrolysis of the racemic lactam is catalyzed by whole cells, since the
isolated microbial enzyme, contrary to the enzyme from Aureobacterium sp., is of
limited stability. The reaction rate is increased by partial liberation of enzyme during
some cell lysis, which is caused by the freeze–thaw process. After the resolution the
remaining ()-lactam is extracted with dichloromethane, and the amino acid formed
O O
NH NH
E +H N
COO-
2 3
+
+ H2O
Scheme 20.11 Reaction scheme for the Celltech Group plc process for the production of the
cyclic lactam (-)-2-azabicylo[2.2.1]hept-5-en-3-one, involving kinetic resolution of racemate using
( þ )-specific b-lactamhydrolase from Pseudomonas solanacearum (E) [1].
NOCl NH3 HCl
O O
N N NOH
Cl Cl NH2
NOH
NH3+Cl-
hydrolase
COOH
racemase H2SO4
H2N NH2
O
NH2
HN
E1 E2
crystallization
Beckmann
rearrangement
cells
Figure 20.13 Flow scheme of the process of Toray Industries Inc. for the production of L-lysine by
dynamic kinetic resolution of a-amino-e-caprolactam catalyzed by lactamase from Cryptococcus
laurentii (E1) and racemase from Achromobacter obae (E2) [1].
is recovered from the aqueous phase as hydrochloride after acidification with HCl
and evaporation. The ()-lactam is produced batchwise on a ton scale in 45% yield,
85% selectivity and >98% e.e., and is mainly used as a direct precursor of carbovir, a
potent and selective inhibitor of HIV-1 [20].
The lactamase from Cryptococcus laurentii together with the racemase from
Achromobacter obae were utilized by Toray Industries Inc. (Japan) in a tandem
dynamic kinetic resolution of a-amino-e-caprolactam for the production of L-lysine
(Figure 20.13), an important nutrient and food supplement. The Torray process
started from cyclohexane, which in several chemical steps was converted into the
racemic caprolactam. In the final enzymatic step the L-enantiomer of the lactam was
enzymatically hydrolyzed to the amino acid, and the remaining D-caprolactam was
racemized by the racemase and thus recycled in situ. In contrast to the classical kinetic
resolution, this reaction scheme gives a possibility of reaching >50% yield of the
desired enantiomer. After the biotransformation the cells were harvested and the
838
product (99.5% e.e.) was isolated by crystallization. The Torray process for L-lysine
operated from 1970 onwards on 4000 t a1 scale but nowadays it has been totally
replaced by much more effective fermentation methods [21].
In a process of Eli Lilly (USA) a substrate promiscuity of penicillin acylase from
Escherichia coli is exploited: the enzyme was found to be highly enantioselective in the
acylation of cis-3-amino-azetidinone with methyl phenoxyacetate, and thus was applied
in the kinetic resolution of the racemate to obtain acylated (2R,3S)-azetidinone
(Scheme 20.12). The product is a key intermediate in the synthesis of loracarbef –
a carbacephalosporin antibiotic, which is a stable analog of the clinically important
antibiotic cefaclor. The resolution is achieved in an aqueous medium containing the
enzyme immobilized on Eupergit. At the end of the process the acylated (2R,3S)-
azetidinone is isolated with 45% yield and >99.9% e.e. [22].
O H O
H2N O N
O COOMe
E O
N + N
O -MeOH O
COOH COOH
Scheme 20.12 Reaction scheme for the process of Eli Lilly for the production of acylated (2R,3S)-
azetidinone by enantioselective acylation of cis-3-amino-azetidinone with methyl phenoxyacetate
catalyzed by penicillin acylase from Escherichia coli (E) [1].
Pfizer Inc. (USA) applied penicillin acylase from E. coli for kinetic resolution of
racemic ethyl 3-amino-5-(trimethylsilyl)-4-pentynoate by enantioselective acylation
with phenylacetic acid (Scheme 20.13). In this biotransformation the enzyme is (R)-
selective. However, the product of interest is the (S)-enantiomer of the b-amino acid,
a chiral synthon for the synthesis of the anti-platelet agent xemilofiban hydrochloride.
Bioconversions were performed in batches with an immobilized enzyme preparation
PGA-450 (Roche Diagnostics GmbH, Mannheim) on a 70-l scale in a stirred tank
reactor equipped with a bottom filter screen serving to recycle the biocatalyst.
Approximately 25 reaction cycles are possible before the enzyme loses 50% of its
initial activity. During the resolution 98% conversion and 99.5% selectivity were
achieved. After the biotransformation the formed (R)-amide and the unreacted (S)-
amine were extracted from the aqueous phase with MTBE, and then were separated
by extraction of the MTBE extract with aqueous HCl. Using this procedure the
(S)-amine was recovered with 43–46% yield and 96–96% e.e. [23].
O O O
O O E O O O O
2 + +
+ H2O
N H2N OH N
H H
TMS TMS TMS
Scheme 20.13 Reaction scheme for the Pfizer Inc. process for the production of (S)-ethyl 3-amino-
5-(trimethylsilyl)-4-pentynoate by kinetic resolution of racemate with penicillin acylase from
Escherichia coli (E) (TMS ¼ trimethylsilyl) [1].
20.5 Processes Using Hydantoinases j839
NH2
NH2
O
+ O
O
Figure 20.14 Flow scheme of the process of BASF AG (Germany) for the kinetic resolution of
racemic phenylethylamine by enantioselective acylation with ethyl methoxyacetate catalyzed by
lipase from Burkholderia plantarii (E) [1].
Not only amidohydrolases can hydrolyze or form amides. Some other hydrolyses
like lipases or esterases often reveal promiscuous activity and enantioselectivity in
this biotransformation, and thus are interesting for industrial application. Kinetic
resolution of racemic phenylethylamine by enantioselective acylation with ethyl
methoxyacetate catalyzed by the lipase from Burkholderia plantarii has been com-
mercialized by BASF AG (Germany) on >100 t a1 scale (Figure 20.14). The
resolution is carried out in an organic solvent mixture of MTBE and ethyl methox-
yacetate using the enzyme immobilized on polyacrylate. The lowering in lipase
activity caused by the use of organic solvent can be offset (about 1000 times and more)
by freeze-drying a solution of the enzyme together with fatty acids (e.g., oleic acid).
The E-value of the lipase in this reaction is >500, which allows the process to reach
93% e.e. for formation of the (R)-amide and >99% e.e. for the remaining (S)-amine
at 50% conversion. Both products are separated by distillation with >90% yield. The
(R)-phenylethyl methoxy amide can be easily hydrolyzed to give the (R)-phenyleth-
ylamine, which together with the (S)-enantiomer is an intermediate for pharma-
ceuticals and pesticides, and can also be used as a chiral synthon in asymmetric
synthesis [24].
20.5
Processes Using Hydantoinases
For many years the hydrolysis of 5-substituted hydantoins was considered an

attractive synthetic route to racemic a-amino acids, because hydantoins can be
840
HNO2
crystallization
Figure 20.15 Flow scheme of the process of Kaneka Corporation for the production of D-(p-
hydroxyphenyl)glycine by enantioselective hydrolysis of racemic 5-(p-hydroxyphenyl)hydantoin
catalyzed by D-specific hydantoinase from Bacillus brevis (E) [1].
conveniently obtained from cheap starting materials, for instance from carbonyl
compounds by the Bucherer–Bergs reaction. After the discovery of hydantoinases
(EC 3.5.2.2), which can hydrolyze racemic hydantoins enantioselectively, this syn-
thetic route also becomes appealing for industrial production of enantiopure natural
and unnatural amino acids.
Kaneka Corporation (formerly Kanegafuchi Chemical Industries Co., Ltd., Japan)
pioneered the application of hydantoinases in the large-scale synthesis of enantio-
pure D-amino acids. In their process for D-(p-hydroxyphenyl)glycine on a 300–700 t
a1 scale immobilized whole cells of Bacillus brevis expressing D-specific hydantoi-
nase are used as biocatalyst for enantioselective hydrolysis of racemic 5-(p-hydro-
xyphenyl)hydantoin (Figure 20.15). The unhydrolyzed L-hydantoin is readily race-
mized in situ under the conditions of enzymatic hydrolysis (pH 8.0), enabling 100%
conversion to be reached. In the second reaction the carbamoyl group of the
hydrolyzed D-hydantoin is removed by chemical treatment with sodium nitrite. The
racemic hydantoin as a starting material for the biotransformation is synthesized
from phenol, glyoxylic acid, and urea via the Mannich condensation. The final
product D-(p-hydroxyphenyl)glycine is a key raw material for several semisynthetic
penicillins, such as ampicillin and amoxicillin, and it is also used in photographic
developers [25].
D-(p-Hydroxyphenyl)glycine is also produced on a pilot scale by the Indian
company Dr. Vig Medicaments. Unlike the process of Kaneka, the carbamoyl group
of the intermediate is removed biocatalytically by a carbamoylase, and not by the
reaction with sodium nitrite (Scheme 20.14). The strain of Pseudomonas sp. used in
this process contains both enzymes – hydantoinase and carbamoylase. After the
biotransformation, which is carried out in batches with 95% conversion and 84%
selectivity using suspended whole cells in a 15 m3 reactor, the product is isolated with
80% yield and 98.5% chemical purity (C. Vig, personal communication, 1997).
20.6 Hydrolysis and Formation of Peptides j841
HO HO HO HO
O O
E1 COOH E2 COOH
NH NH + H O
HN HN 2 HN NH2 + H2O NH2
- CO2
O O O
Scheme 20.14 Reaction scheme for the process of Dr. Vig Medicaments for the production of D-(p-
hydroxyphenyl)glycine using hydantoinase (E1) and carbamoylase (E2) from Pseudomonas sp. [1].
Evonik has extended the scope of the hydantoinase process to cover also the
production of optically pure natural and non-natural L-amino acids (Scheme 20.15).
The company developed a tailor-made recombinant Escherichia coli strain overex-
pressing a L-hydantoinase, a carbamoylase, and a hydantoin racemase from Arthro-
bacter sp. DSM 9771. This highly active recombinant whole-cell biocatalyst was
produced in high-cell density fermentation on a m3-scale at concentrations above
50 g l1 dry cell weight. Although the enantioselectivity of the designed L-hydantoi-
nase is not impressive, but technically significant – only 20% e.e. is achieved at 40%
conversion – the feasibility of the process has been confirmed on the m3-scale using a
simple batch reactor coupled to a continuous centrifuge for cell separation [26].
R O R O R COOH
E1 E2 E3 R COOH
HN NH HN NH HN NH2
+ H2O + H2O NH2
O O O - NH3
Scheme 20.15 Reaction scheme for the process of Evonik for the production of natural and non-
natural L-amino acids using racemase (E1), hydantoinase (E2), and carbamoylase (E3) from
Arthrobacter sp. [1].
20.6
Hydrolysis and Formation of Peptides
Proteases or peptidases (EC 3.4.) belong to the class of hydrolases acting on C–N bonds
in peptides. In hydrolysis or formation of peptides proteases show clearly their benefits
over chemocatalysts: C–N bonds in peptides scarcely differ in their chemical reactivity
and chemocatalysts usually fail to distinguish between them; in contrast, proteases are
by their nature very specific towards peptides and can with high precision cut or form
C–N bonds even in complex polypeptides built up of more than 50 amino residues.
Biocatalytic production of insulin is an illustrative example of an industrial
processes relying on selective proteases. Historically, insulin has been isolated and
purified from animal tissues, but today it is mainly produced by fermentation with a
capacity of about 5–6 t a1 worldwide and it belongs to the first mammal proteins that
were synthesized with an identical amino acid sequence using recombinant DNA-
technology. Eli Lilly established a multistep process for insulin production starting
from the precursor Trp-LE-met -pro-insulin, which is directly obtained by fermentation
842
E. coli
plasmid
separation purification
gene for pro-insulin
oxidative sulfitolysis CNBr-cleavage

separation pro-insulin trypLE-met-pro-insulin
E
-
pro-insulin-(S-SO 3)8 purification human insulin
Figure 20.16 Flow scheme of the process of Eli Lilly for the production of human insulin from
Trp-LE-met-pro-insulin using tryptase and carboxypeptidase from pig pancreas (E) [1].
of recombinant E. coli (Figure 20.16). The signal peptide sequence Trp-LE-met is

removed from the peptide chemically by treatment with CNBr to yield a denatured pro-
insulin. Then oxidative sulfitolysis is performed to unfold the peptide, to break any
disulfide bonds, and to protect all cysteine side chains. The pro-insulin-S-sulfonate
formed is purified by chromatography and afterwards is treated with mercaptoethanol.
During the last step all cysteine residues and disulfide bonds are restored, and the pro-
insulin is correctly refolded to its native form. In the first enzymatic step catalyzed by a
tryptase from pig pancreas the C-chain of the native pro-insulin is cut off and the
peptide is converted with >70% yield into a mono/di-arg-insulin. In the second
enzymatic step the enzyme carboxypeptidase from pig pancreas removes, finally, two
C-terminal arginines from the mono/di-arg-insulin with >99% conversion, >99%
selectivity, and >95% yield to give the end product human insulin [27].
In the process of Sanofi insulin is produced on a >0.5 t a1 scale from another
precursor, pre-pro-insulin. The precursor is expressed in recombinant E. coli and is
further transformed into insulin in the same way as in the Eli Lillys process: first, pre-
pro-insulin is hydrolyzed to mono/di-arg-insulin by the tryptase from pig pancreas in
a 10 m3 reactor with >99.9% conversion, >65% selectivity, and >65% yield
(Figure 20.17a); second, the enzyme carboxypeptidase from pig pancreas converts
mono/di-arg-insulin into human insulin (Figure 20.17b). The second biotransfor-
mation is conducted in a 7.5 m3 reactor and it also proceeds practically to complete-
ness, though with >90% selectivity and >90% yield [28].
Novo Nordisk (Denmark) produces insulin by a process involving only one
biocatalytic step, in which the precursor pro-insulin is converted into a threonine
ester of insulin by a transpeptidation catalyzed by the tryptase from pig pancreas
(Figure 20.18). This is what differentiates this process from the synthetic routes of
C-chain
PRO GLN LEU SER GLY ALA GLY PRO GLY GLY GLY LEU
SER GLY GLU LEU ALA LEU
(a) LEU
GLU
VAL
GLN
GLN
GLY
LYS VAL
ARG S S
GLN
GLY A-chain
PRO
ILE VAL
GLU GLN CYS CYS THR SER ILE CYS SER ASN
LEU THR GLN LEU GLU ASN TYR CYS ASP
COOH
NH2 GLU
S ALA
PHE S GLU
VAL
S S
ARG
ASN
GLN ARG
HIS THR
LEU CYS GLY SER LYS
HIS LEU VAL GLU ALA LEU TYR
LEU VAL CYS GLY GLU ARG GLY PHE TYR THR PRO
B-chain
E - C-chain
S S
H2N GLY
ILE VAL
LEU THR GLN LEU GLU ASN TYR CYS
COOH
H2N
S
PHE S COOH
VAL
S S
ARG
ASN
GLN ARG
HIS THR
LEU CYS LYS
GLY SER HIS LEU
VAL GLU ALA LEU TYR LEU VAL CYS GLY GLU ARG PRO
GLY PHE TYR THR
(b) S S
H2N GLY
ILE VAL
COOH
H2N
S
PHE S COOH
VAL
S S
ARG
ASN
GLN ARG
HIS THR
LEU CYS GLY SER LYS
HIS LEU VAL GLU ALA
LEU TYR LEU VAL CYS GLY GLU ARG GLY PHE TYR THR PRO
E - Arg
S S
H2N GLY
ILE VAL
COOH
H2N
S
PHE S
VAL
S S
ASN COOH
GLN
HIS THR
LEU CYS GLY SER LYS
HIS LEU VAL GLU ALA
LEU TYR LEU VAL CYS GLY GLU ARG GLY PHE TYR THR PRO
Figure 20.17 (a) Hydrolysis of pre-pro-insulin (b) hydrolysis of mono/di-arg-insulin to human

to mono/di-arg-insulin by tryptase from pig insulin by carboxypeptidase from pig pancreas
pancreas (E) as the first step in the process of (E) as the second step of the Sanofi process for
Sanofi for the production of human insulin [1]; the production of human insulin [1].
844
supernatant
medium
centrifuge
cells concentrate
crystallization
stripped microfiltration
fermentation
liquid
H 2O mother-
liquid
human insulin
threonine ester
E
purification of insulin
insulin precursor crystals
by HPLC
Figure 20.18 Flow scheme of the process of Novo Nordisk for the production of human insulin
from pro-insulin using tryptase from pig pancreas (E) [1].
Sanofi and Eli Lilly, where the tryptase catalyzes the hydrolysis of pro-insulin to
insulin. The precursor is directly produced by fermenting a recombinant Saccha-
romyces cerevisiae in 80-m3 fermenters, which are operated continuously for 3–4
weeks. During the fermentation the medium is added at the same speed as the broth
is drawn. Pro-insulin is purified by crystallization and then subjected to transpepti-
dation, affording >99.9% conversion, >97% selectivity, and >97% yield. To prevent
the possible trypsin-catalyzed cleavage of the B-chain at position 22, the biotrans-
formation is performed at low water concentration in the presence of organic
solvents, a surplus of threonine ester, low temperature (6 C), and low pH (<7.0).
Finally, the formed threonine ester is converted into human insulin by simple
hydrolysis with subsequent purification steps [29].
Beside insulin production, proteolytic enzymes have been applied in the synthesis
of the dipeptide aspartame, a multi-thousand-tons product, used as a low-calorie
sweetener in food and beverages, table-top sweeteners, dairy products, instant mixes,
dressings, jams, confectionery, and toppings and in pharmaceuticals. The worldwide
capacity of aspartame production is >20 000 t a1 and at present this substance is
predominantly synthesized chemically. The main problem in the chemical synthesis
is the formation of a by-product, b-aspartame, which tastes bitter and, therefore, has
to be completely removed from the a-isomer. DSM together with Tosoh (formerly
Toyo Soda, Japan) developed an alternative process for production of 2500 t a1 of
aspartame based on stereoselective enzymatic coupling of phenylalanine methyl
ester and protected aspartic acid catalyzed by a thermolysin from Bacillus proteolicus
20.7 Processes Using C–N Lyases j845
racemization
fermentation
medium
NH2
COOMe
cells
HCl
E
enzyme
recovery
Figure 20.19 Flow scheme of the process of DSM and Tosoh for the production of aspartame
by stereoselective enzymatic coupling of phenylalanine methyl ester and protected aspartic acid
catalyzed by thermolysin from Bacillus proteolicus (E) [1].
(Figure 20.19). The enzymatic route offers the advantages that no b-isomer is
produced and that no racemization occurs during synthesis. Moreover, the enzyme
is completely stereoselective towards phenylalanine, so that the racemic mixture of
the ester can be used as the acylation agent. The biotransformation takes place with
the solubilized enzyme in aqueous media under mild conditions. Since the reaction
is limited by the equilibrium, the products have to be removed from the reaction
mixture to obtain high yields. This is achieved by adding an excess of phenylalanine
methyl ester, which forms a poorly soluble adduct with the carboxylic anion of the
protected aspartame. The adduct precipitates from the reaction mixture and is
removed easily by filtration. Final steps of the process are deprotection of aspartame
and racemization and recycling of the remaining L-phenylalanine methyl ester [30].
20.7
Processes Using C–N Lyases
Carbon–nitrogen lyases (EC 4.3.) catalyze cleavage of the C–N bond. Importantly, this
bond cleavage is different from hydrolysis, often leaving unsaturated products with
double bonds that may be subjected to further reactions. In industrial processes these
enzymes are most commonly used in the synthetic mode, meaning that the reverse
reaction – addition of a molecule to an unsaturated substrate – is of interest. To shift
the equilibrium these reactions are carried out at very high substrate concentrations,
which results in very high conversions of the desired products.
846
H2SO4
precipitation
pH 2.8
Figure 20.20 Flow scheme of the process of BioCatalytics Inc. for the production of L-aspartic
acid by amination of fumaric acid catalyzed by aspartase from Escherichia coli (E) [1].
Several companies use aspartase for the production of L-aspartic acid from fumaric
acid. The enzyme is expressed in different host microorganisms and can be applied
either in isolated form or as a whole-cell catalyst. L-Aspartic acid is a precursor for the
synthesis of aspartame, but it is also used as an acidulant, as a food additive, in
parenteral nutrition, and as a chiral synthon in organic synthesis.
The production of L-aspartic acid by BioCatalytics Inc. (now part of Codexis Inc.,
USA) was carried out with the help of the aspartase from Escherichia coli
(Figure 20.20). An aqueous solution of substrates was pumped through a 75-l plug
flow reactor packed with the isolated enzyme immobilized on silica support. The
presence of MgCl2 enhanced the activity of the enzyme and prolonged its half-life up
to six months. Downstream processing was performed by acidifying the product
solution to pH 2.8, and precipitating the product by chilling. This process achieved a
conversion of 99%, selectivity of 96%, and a higher productivity than the process
using immobilized whole cells. The acid was isolated in 95% yield with an optical
purity of >99.9% (D. Rozzell, BioCatalytics, personal communication, 1998).
Kyowa Hakko Kirin Co., Ltd. (formerly Kyowa Hakko Kogyo Co., Ltd., Japan) uses
the immobilized aspartase from Escherichia coli for the synthesis, too. Here the
enzyme is immobilized on Duolite A-7, a weakly basic anion-exchange resin, and
packed into the column reactor, which is fed with the aqueous medium containing
2 M fumaric acid and 4 M NH4OH as amine source. The reactor is operated for over
three months at >99% conversion and >99.9% e.e. [31].
Suspended whole cells of Brevibacterium flavum containing the aspartase are
applied in the process of Mitsubishi Chemical Corporation (Japan). The biotrans-
formation is performed in repetitive batches with a yield of >99.99% and a selectivity
of >99.99%, and the bacterial cells are retained by ultrafiltration (Figure 20.21); also
in this process, 4 M NH4OH is added as amine source. To achieve stoichiometric
conversion of fumaric acid into L-aspartic acid it was necessary to suppress complete-
ly a side reaction, catalyzed by an intracellular fumarase, that afforded L-malic acid.
The suppression is achieved by thermal deactivation of the fumarase, which takes
20.7 Processes Using C–N Lyases j847
cells
for repetitive batch
Figure 20.21 Flow scheme of the process of Mitsubishi Chemical Corporation for the production of
L-aspartic acid by amination of fumaric acid catalyzed by aspartase from Brevibacterium flavum (E) [1].
place when the cells are incubated at 45 C for 5 h in the presence of 2 M NH4OH,
0.75 M L-aspartic acid, 0.0075 M CaCl2, and 0.08% (w/v) of the nonionic detergent
Tween 20. During the thermal treatment L-aspartic acid and CaCl2 act as protectors
against the unwanted thermal inactivation of the aspartase, and by the addition of
Tween 20 the production of L-aspartic acid is increased by 40% [32].
Tanabe Seiyaku Co., Ltd. (now part of Mitsubishi Pharma Corporation, Japan)
produces 700 t a1 of L-aspartic acid using immobilized whole cells of Escherichia coli
B ATCC 11303 (Figure 20.22). The company decided to use the whole cells as a
biocatalyst because the stability of the isolated aspartase in free or immobilized form
was not satisfactory. The cells are immobilized in polyacrylamide or, preferably, in
k-carrageenan gel and the catalyst is packed in a plug flow reactor. The costs of the
continuous process are reduced to two-thirds of those of a batchwise operation. This
pyridoxal phosphate
NH3 +
pyruvate
CO2
NaOH P
HOOC evaporation
COOH
pH
crystallization
E1 E2
NH2
HOOC
crystallization
NH2
COOH
HOOC
Figure 20.22 Flow scheme of the process of Tanabe Seiyaku Co., Ltd. for the production of
L-aspartic acid and L-alanine by a two-step biotransformation catalyzed by aspartase from Escherichia
coli (E1) and L-aspartate b-decarboxylase from Pseudomonas dacunhae (E2) [1].
848
process is the first example of the application of immobilized whole cells and it is one
of the rare examples where the synthesis of an amino acid via an enzymatic route is
economically more attractive than the usual fermentation methods. The activity of
the cells is increased tenfold by immobilization and the half-life of the cells is about 12
days. By addition of about 1 mM Mg2 þ , Mn2 þ , or Ca2 þ the half-life can be extended
to more than 120 days. During downstream processing the product is isolated with
>95% overall yield by titration to the isoelectric point (pH 2.8) with H2SO4 and
filtration of the precipitate. Beside the production of L-aspartic acid, the company also
employs this aspartase catalyzed biotransformation as the first step in a two-step
synthesis of L-alanine, the second step of which is decarboxylation of L-aspartic acid by
L-aspartate b-decarboxylase from Pseudomonas dacunhae [33]. The tandem production
of L-aspartic acid and L-alanine from fumaric acid by the two-step biotransformation is
also established by Evonik on multi-hundred-tons scale.
Another enzyme, L-phenylalanine ammonia-lyase (PAL), was commercialized by
Genex Corporation (now part of Enzon Inc., USA) in the production of the amino acid
L-phenylalanine (Scheme 20.16), which is used as a building block for the syntheses
of the artificial sweetener aspartame and the macrolide antibiotic rutamycin B, as well
as an ingredient in parenteral nutrition. The biotransformation takes place at 25 C in
a bioreactor loaded with whole cells of the PAL-producing microorganism Rhodotor-
ula rubra suspended in aqueous medium. The bioreactor is operated in a fed-batch
mode by periodic feeding with a concentrated ammonia trans-cinnamate solution
obtained by mixing of an aqueous solution of trans-cinnamic acid with 29% aqueous
ammonia and by adjusting the pH to 10.6 with carbon dioxide. The cells are initially
cultivated under aerobic, growth-promoting conditions, but the biotransformation
itself is performed under anaerobic, static conditions due to the instability of the
enzyme towards oxygen and agitation. Therefore, before addition of the cells and
after each addition of the substrate solution the bioreactor content is sparged with
nitrogen. At the end of the biotransformation the cells are harvested by centrifuga-
tion, the supernatant is evaporated, and the amino acid is isolated in 85.7% yield by
crystallization. Prior to this and related processes, L-phenylalanine was mainly
obtained from hydrolysates of human hair, feathers, and other waste proteins [34].
COOH COOH
E
+ NH3
NH2
Scheme 20.16 Reaction scheme of the Genex Corporation process for the production of L-
phenylalanine from trans-cinnamic acid and ammonia using L-phenylalanine ammonia-lyase from
Rhodotorula rubra (E) [1].
20.8
Processes Using Transaminases
Although the application of transaminases (EC 2.6.1.) for the industrial production of
valuable chiral amines and amino acids from cheap carbonyl- and amino-donors is
20.8 Processes Using Transaminases j849
very encouraging, the large-scale usage of transaminases is still limited because of the
equilibrium of transamination reaction. This means that one of the substrates should
be added in excess or one of the products should be removed in situ to reach high
conversion levels during the biotransformation, but these efforts are often not
compatible with the enzymes, causing either inhibition or deactivation of the
biocatalysts. Nevertheless, some companies have overcome these challenges and
developed economically viable transaminase-based biocatalytic processes.
D-Aspartate transaminase from Bacillus sp. is one of the enzymes that catalyzes a
reaction network, which has been exploited by NSC Technologies (now part of
Chemtura Corporation, USA) for the production of unnatural D-amino acids on a
multi-tons scale (Scheme 20.17). The reaction network is designed to overcome the
main drawback of transaminases – the equilibrium conversion of about 50%. The
process starts from a cheap racemic amino acid, with racemic aspartate playing the
role of an amino donor. In the first reaction of the network the enzyme L-amino acid
deaminase catalyzes enantioselective deamination of the L-enantiomer of the amino
acid, yielding a corresponding a-keto acid, which with the help of the D-transaminase
is converted into the respective D-enantiomer. The net effect of such an asymmetric
NH3 + H2O2
R R R
+ O2 + H2O +
H2N COOH L-amino acid O COOH H2N COOH
deaminase
D,L-amino acid α-keto acid D-amino acid
HOOC
COOH
NH2 L-amino acid
id D,L-aspartate
transferase
o ac
amin ase
m
race
R1
HOOC HOOC
COOH COOH
+ H2N COOH
NH2 O
D-amino acid
L-aspartate 2-oxo-succinic acid
CO2
acetolactate HOOC
HOOC synthase
HO HO
O O O
pyruvic acid acetolactate acetoin
CO2 CO2
Scheme 20.17 Reaction scheme of the NSC Technologies process for the production of unnatural
D-amino acids by asymmetric transformation of racemates using L-amino acid deaminase,
transferase, racemase, and aceto lactate synthase expressed in Escherichia coli [1].
850
transformation is the total conversion of the racemate into the pure D-amino acid with
100% e.e. Amination of the a-keto acid catalyzed by the D-transaminase is conjugated
with deamination of D-aspartate to 2-oxosuccinic acid, which further decarboxylates
to pyruvic acid and carbon dioxide. The unconverted L-aspartate is racemized in situ
by an aspartate racemase. The equilibrium of this reaction system is shifted to the
product side by the action of another enzyme, acetolactate synthase, which catalyzes
dimerization of pyruvic acid to acetolactate. The latter undergoes spontaneous
decarboxylation to acetoin that can be easily removed and does not participate in
other reactions. The process is performed in one pot with suspended whole cells of
recombinant E. coli strains expressing separately all four enzymes involved in the
synthesis. After the conversion, the formed D-amino acid is isolated by
crystallization [35].
Celgene Corporation (USA) synthesizes (S)-2-amino-1-methoxypropane with
>99% e.e. by enantioselective transamination of 1-methoxy-2-propanone with iso-
propylamine catalyzed by suspended whole cells of Bacillus megaterium containing an
(S)-selective transaminase (Scheme 20.18). The product is used as an intermediate
for the synthesis of several agrochemicals, for example, that herbicide (S)-metola-
chlor introduced by the former company Ciba-Geigy (now Novartis, Switzerland) and
produced on 20 000 t a1 scale. Isopropylamine is chosen as the amino donor for the
transamination because it is cheap and attractive from the point of view of reaction
kinetics. When the wild-type enzyme was used as the catalyst, conversion during the
reaction was limited due to product inhibition. This drawback was overcome by a
single mutation in the gene coding the transaminase that allowed an increase in the
possible product concentration from 0.16 to 0.45 M. The biotransformation runs with
94% selectivity and catalyst consumption of 0.15 kg of dried cell mass per kg of
product. Since no recycling of the catalyst is integrated, the residual activity after one
batch run is of no interest. The Celgene process is competitive with that of Ciba-
Geigy, who took about ten years to find a chemical catalyst for the production of the
enantioenriched (S)-metolachlor, but only achieved an e.e. of 79%. Other companies
also developed alternative approaches to the herbicide, albeit the associated produc-
tion costs appeared to be higher than for the enzymatic route [36].
O NH2 E NH2 O
O + O +
Scheme 20.18 Reaction scheme of the Celgene Corporation process for the production of (S)-2-
amino-1-methoxypropane by enantioselective transamination of 1-methoxy-2-propanone with
isopropylamine catalyzed by transaminase from Bacillus megaterium (E) [1].
20.9
Summary and Outlook
The industrial biotransformations presented in this section demonstrate that the

enzymes acting on C–N bond are widely applied in the chemical industry and,
References j851
thus, might be regarded as one of the main pillars of industrial biocatalysis. In
many cases companies have preferred to use the enzymes for production or
transformation of nitrogen-containing compounds, because the enzymatic pro-
cesses were found to be more sustainable, green, and hence more competitive
than the corresponding chemical syntheses. However, quite often the application
of the enzymes acting on C–N bonds in organic synthesis on a large scale is limited
by the drawbacks that are common for every enzyme: low biocatalyst stability, low
space–time yield, or narrow substrate spectrum. In many mentioned processes the
problem of biocatalyst stability could be adequately solved by employing whole
cells as biocatalysts. If stability was not a problem, free or immobilized enzymes
were used to reach higher space–time yields. Several companies also overcame
another drawback, the narrow substrate spectrum originating from high speci-
ficity of enzymes, and established flexible platform processes, for example,
L-aminopeptidase process by DSM or amidase-catalyzed kinetic resolution of
amides by Lonza AG, accepting a range of starting materials and yielding a range
of related products. But, in general, to overcome all these challenges is still not a
trivial task. For this reason the industry is constantly looking for superior enzymes
able to act on various substrates at higher temperatures, pressures, substrate
concentrations, and in the presence of organic solvents, and to provide these better
enzymes is a major task of applied biocatalysis. In this respect, screening for new
C–N acting enzymes from extremophile microorganisms or improving existing
ones by methods of directed evolution or rational protein design are two possible
ways to find attractive enzyme candidates suitable for a scale-up. On the other
hand, it might be foreseen that in future the paradigm of industrial biocatalysis will
migrate from one enzyme–one reaction to a multiple enzymes–coupled
reactions concept that would give rise to in vitro multi-enzymatic processes, like
the one of NSC Technologies with D-aspartate transaminase and the three-enzyme
hydantoinase process of Evonik. Such in pot biotransformations involving
coupled reactions would be, for certain, economically attractive since they offer
the advantage of avoiding isolation of intermediates and shifting reaction equi-
librium to the desired direction. In view of recent findings in the field of synthetic
and systems biology, one may also anticipate that in future the industry would be
also interested in in vivo multistep biotransformations performed by so-called
designer bugs – whole-cell biocatalysts optimized by metabolic engineering to
increase a product titer or to produce a totally new compound via an incorporated
artificial pathway.
References
1 Liese, A., Seelbach, K., and Wandrey, C. S.M., Gavagan, J.E., Stieglitz, B.,
(2006) Industrial Biotransformations, Hennesey, S.M., and DiCosimo, R. (1999)
Wiley-VCH Verlag GmbH, Weinheim. 5-Cyanovaleramide production using
2 Hann, E.C., Eisenberg, A., Fager, S.K., immobilized Pseudomonas chlororaphis
Perkins, N.E., Gallagher, F.G., Cooper, B23. Bioorg. Med. Chem., 7, 2239–2245.
852
3 Petersen, M. and Kiener, A. (1999) 13 Crosby, J. (1991) Synthesis of optically

Biocatalysis – preparation and active compounds: a large scale
functionalization of N-heterocycles. Green perspective. Tetrahedron, 47, 4789–4846.
Chem., 4, 99–106. 14 Robins, K. and Gilligan, T. (1992)
4 Yamada, H. and Kobayashi, M. (1996) Biotechnologisches verfahren zur
Nitrile hydratase and its application to herstellung von (S)-( þ )-2,2-
industrial production of acrylamide. dimethylcyclopropancarboxamid und (R)-
Biosci. Biotechnol. Biochem., 60, ()-2,2-dimethylcyclopropancarbons€aure,
1391–1400. Lonza AG, EP 0502525 A1.
5 Hann, E.C., Sigmund, A.E., Hennessey, 15 Eichhorn, E., Roduit, J.-P., Shaw, N.,
S.M., Gavagan, J.E., Short, D.R., Bassat, Heinzmann, K., and Kiener, A. (1997)
A.B., Chauhan, S., Fallon, R.D., Payne, Preparation of (S)-piperazine-2-
M.S., and DiCosimo, R. (2002) carboxylicacid, (R)-piperazine-2-
Optimization of an immobilized-cell carboxylic acid, and (S)-piperidine-2-
biocatalyst for production of 4- carboxylic acid by kinetic resolution of the
cyanopentanoic acid. Org. Process Res. corresponding racemic carboxamides
Dev., 6, 492–496. with stereoselective amidase in whole
6 Gr€oger, H. (2001) Enzymatic routes to bacterial cells. Tetrahedron: Asymmetry, 8,
enantiomerically pure aromatic a-hydroxy 2533–2536.
carboxylic acids: a further example for the 16 Shaw, N.M., Naughton, A., Robins, K.,
diversity of biocatalysis. Adv. Synth. Catal., Tinschert, A., Schmid, E., Hischier, M.-L.,
343, 547–558. Venetz, V., Werlen, J., Zimmermann, T.,
7 Wieser, M., Heinzmann, K., and Kiener, Brieden, W., de Riedmatten, P., Roduit, J.-
A. (1997) Bioconversion of 2- P., Zimmermann, B., and Neum€ uller, R.
cyanopyrazine to 5-hydroxypyrazine-2- (2002) Selection, purification,
carboxylic acid with Agrobacterium sp. characterization, and cloning of a novel
DSM 6336. Appl. Microbiol. Biotechnol., 48, heat-stable stereo-specific amidase from
174–180. Klebsiella oxytoca, and its application in the
8 Gl€ockler, R. and Roduit, J.-P. (1996) synthesis of enantiomerically pure (R)-
Industrial bioprocesses for the production and (S)-3,3,3-trifluoro-2-hydroxy-2-
of substituted aromatic heterocycles. methylpropionic acids and (S)-3,3,3-
Chimia, 50, 413–415. trifluoro-2-hydroxy-2-methylpropionamide.
9 Christ, C. (1995) Biochemical production Org. Process Res. Dev., 6, 497–504.
of 7-aminocephalosporanic acid, in 17 McCague, R. and Taylor, S.J.C. (1997)
Ullmanns Encyclopedia of Industrial Integration of an acylase
Chemistry (ed. H.-J. Arpe), VCH, biotransformation with process
Weinheim, B8, 240–241. chemistry: a one-pot synthesis of NtBoc-L-
10 Tsuzuki, K., Komatsu, K., Ichikawa, S., 3-(4-thiazolyl)alanine and related amino
and Shibuya, Y. (1989) Enzymatic acids, in Chirality In Industry II (eds
synthesis of 7-aminocephalosporanic acid S.A.N. Collin, G.N. Sheldrake, and J.
(7-ACA). Nippon Nogei Kagaku Kaishi, 63, Crosby), John Wiley & Sons, Inc.,
1847. New York, pp. 194–200.
11 Matsumoto, K. (1993) Production of 6- 18 Bommarius, F.A.S., Drauz, K., Klenk, H.,
APA, 7-ACA, and 7-ADCA by and Wandrey, C. (1992) Operational
immobilized penicillin and cephalosporin stability of enzymes – acylase-catalyzed
amidases, in Industrial Application of resolution of N-acetyl amino acids to
Immobilized Biocatalysts (eds A. Tanaka, T. enantiomerically pure L-amino acids. Ann.
Tosa, and T. Kobayashi), Marcel Dekker N. Y. Acad. Sci., 672, 126–136.
Inc., New York, pp. 67–88. 19 McCague, R. and Taylor, S.J.C. (1997)
12 Bruggink, A., Roos, E.C., and de Vroom, E. Development of an immobilised
(1998) Penicillin acylase in the industrial lactamase resolution process for the
production of b-lactam antibiotics. Org. ( þ )-c-lactam and the ring-opened
Process Res. Dev., 2, 128–133. ()-amino acid, in Chirality In Industry II
References j853
(eds S.A.N. Collin, G.N. Sheldrake, and J. 28 Frank, B.H. and Chance, R.E. (1983) Two
Crosby), John Wiley & Sons, Inc., New routes for producing human insulin
York, pp. 187–188. utilizing recombinant DNA technology.
20 Taylor, S.J.C. and McCague, R. (1997) M€unch. Med. Wschr., 125, 14–20.
Resolution of the carbocyclic nucleoside 29 Jørgensen, L.N., Rasmussen, E., and
synthon 2-azabicyclo[2.2.1]hept-5-en-3- Thomsen, B. (1989) HM(ge), Novos
one with lactamases, in Chirality In biosynthetic insulin. Med. View., III (4),
Industry II (eds S.A.N. Collin, G.N. 1–7.
Sheldrake, and J. Crosby), John Wiley & 30 Harada, T., Irino, S., Kunisawa, Y., and
Sons, Inc., New York, pp. 184–190. Oyama, K. (1996) Improved enzymatic
21 Schmidt-Kastner, G. and Egerer, P. (1984) coupling reaction of N-protected-L-
Amino acids and peptides, in aspartic acid and phenylalanine methyl
Biotechnology, Vol. 6a (ed. K. Kieslich), ester, Holland Sweetener Company, The
Verlag Chemie, Weinheim, pp. 387–419. Netherlands, EP 0768384 A1.
22 Zmijewski, M.J., Briggs, B.S., Thompson, 31 Tanaka, A., Tosa, T., and Kobayashi, T.
A.R., and Wright, I.G. (1991) (1993) Industrial Application of
Enantioselective acylation of a beta-lactam Immobilized Biocatalysts, Marcel Dekker
intermediate in the synthesis of Inc., New York.
Loracarbef using penicillin G amidase. 32 Yamagata, H., Terasawa, M., and Yukawa,
Tetrahedron Lett., 32, 1621–1622. H. (1994) A novel industrial process for L-
23 Landis, B.H., Mullins, P.B., Mullins, K.E., aspartic acid production using an
and Wang, T. (2002) Kinetic resolution of ultrafiltration membrane. Catal. Today, 22,
b-amino esters by acylation using 621–627.
immobilized penicillin amidohydrolase. 33 Takamatsu, S., Umemura, I., Yamamoto,
Org. Process Res. Dev., 6, 539–546. K., Sato, T., Tosa, T., and Chibata, I. (1982)
24 Balkenhohl, F., Ditrich, K., Hauer, B., and Production of L-alanine from ammonium
Ladner, W. (1997) Optisch aktive amine fumarate using two immobilized
durch lipasekatalysierte microorganisms: elimination of side
methoxyacetylierung. J. Prakt. Chem., 339, reactions. Eur. J. Appl. Microbiol.
381–384. Biotechnol., 15, 147–152.
25 Cheetham, P.S.J. (1994) Case studies in 34 Vollmer, P.J., Schruben, J.J., Montgomery,
applied biocatalysis, in Applied Biocatalysis J.P., and Yang, H.-H. (1986) Method for
(eds J.M.S. Cabral, D. Best, L., Boross, and stabilizing the enzymatic activity of
J. Tramper), Harwood Academic phenylalanine ammonia lyase during L-
Publishers, Chur, Switzerland, pp. 68–70. phenylalanine production, Genex
26 May, O., Verseck, S., Bommarius, A., and Corporation, US 4584269.
Drauz, K. (2002) Development of dynamic 35 Ager, D.J., Fotheringham, I.G., Laneman,
kinetic resolution processes for S.A., Pentaleone, D.P., and Taylor,
biocatalytic production of natural and P.P. (1997) The large scale synthesis of
nonnatural l-amino acids. Org. Process Res. unnaturalacids. Chim. Oggi, 15 (3/4), 11–14.
Dev., 6, 452–457. 36 Matcham, G.W. and Lee, S. (1994) Process
27 Ladisch, M.R. and Kohlmann, K.L. (1992) for the preparation of chiral 1-aryl-2-
Recombinant human insulin. Biotechnol. aminopropanes, Celgene Corporation, US
Prog., 8, 469–478. 5360724.
j855
Part IV
Formation and Cleavage of CC Bonds
and Oliver May.
j857
21
Aldol Reactions
Wolf-Dieter Fessner
21.1
Aldol Reactions
Stereoselective carboligation is a pivotal process in the asymmetric construction of

the skeletal framework of complex molecular targets from smaller and simpler
building blocks. Among the tools used to perform this transformation, the catalytic
asymmetric aldol reaction constitutes one of the most powerful methodologies for
the stereocontrolled formation of carbon–carbon bonds in the synthesis of enantio-
pure oxygenated compounds [1]. The aldol reaction can create, in a single operation, a
new carbon–carbon bond and two new stereogenic centers on the a- and b-carbons of
the aldol adduct. The ability to control the absolute configuration of both newly
formed stereogenic centers is of fundamental importance.
Complementary to conventional chiral auxiliary-mediated processes and chemical
catalysis using chiral Lewis acids or organocatalysis, biocatalytic transformations by
means of carboligating enzymes offer a unique tool to perform such preparative
transformations in a sustainable, environmentally benign fashion. Aldolases are a
specific group of lyases that catalyze the reversible stereoselective addition of an aldol
donor component (nucleophile) onto an acceptor component (electrophile). Products
are usually typified as 3-hydroxy carbonyl compounds, a structural element that is
frequently incorporated in the framework of complex natural products. Indeed,
aldolases have evolved to catalyze the formation and degradation of oxygenated
metabolites, and are thus found in various metabolic pathways of carbohydrates, keto
acids, and some amino acids. Owing to their high selectivity and catalytic efficiency,
aldolases offer considerable synthetic utility, and have found increasing acceptance,
as chiral catalysts for the in vitro synthesis of chiral compounds [2–13]. Thereby,
molecular complexity can be rapidly built up under mild conditions, without a need
for tedious and time-consuming iterative steps for protection and deprotection of
sensitive or reactive functional groups, yet with high chemical efficiency and often
uncompromised stereochemical fidelity. Although chemical methods in asymmetric
synthesis have reached extraordinary levels of sophistication, the corresponding
development of asymmetric catalysis in water is still in progress.
and Oliver May.
j 21 Aldol Reactions
858
Addition reactions are intrinsically atom-economic, because they do not produce

unwanted side products as a waste material. The hallmark of the aldol strategy in
synthesis, however, is its combinatorial value to produce groups of new structure
types (i.e., generate molecular diversity for investigations in drug discovery) by an
individual variation of the nucleophilic and the electrophilic reagent parts
(Scheme 21.1). A further combinatorial dimension of the aldol strategy arises by
the fact that with formation of the new carbon–carbon bond up to two new
stereogenic centers are concomitantly established, offering the potential for stereo-
divergent product generation [2]. By this technique multiple, stereoisomeric products
can be derived from a common pair of synthetic precursors by chiral catalyst control,
sidestepping the need for a subsequent asymmetric functionalization reaction [2, 14].
Obviously, transliteration of this synthetic opportunity to the biocatalytic world
depends on the existence of related, stereo-complementary enzymes that would
need to have a similarly broad substrate tolerance.
O
O
R H
- R'
X
OH O OH O OH O OH O
R R' R R' R R' R R'

X X X X
Scheme 21.1 Generation of stereo-diversity by the aldol addition.
21.1.1
Classes of Aldolases
To date, several dozen distinct aldolases are known (classified by enzyme numbers
EC 4.1.2.x) [15, 16], and many of these enzymes are commercially available at a scale
sufficient for preparative applications. Aldolase catalysis is most attractive for the
synthesis and modification of biologically relevant classes of organic compounds that
are typically complex, multifunctional, and water soluble. Typical examples are those
structurally related to amino acids or carbohydrates [17–20], which are difficult to
prepare and to handle by conventional methods of chemical synthesis and thus
mandate the laborious manipulation of protective groups.
Most of the enzymes known to facilitate carbon–carbon bond formation and
cleavage (lyases) catalyze a crossed aldol reaction as a reversible, stereocontrolled
addition of a nucleophilic ketone donor (enolate or analog) onto an electrophilic
aldehyde acceptor. Whereas lyases are typically quite flexible in using a broad range of
21.1 Aldol Reactions j859
O O O O
H H HO OPO32- H2N
H CO2H OH
HH 1 HH 2 HH 3 HH 4
acetaldehyde pyruvate dihydroxyacetone glycine
phosphate
Figure 21.1 Nucleophilic donor substrates of synthetically useful aldolases.
aldehydes as acceptors, owing to mechanistic requirements they are quite specific for
the nucleophilic donor component, which usually is a prochiral two- or three-carbon
fragment. Hence, the enzymes are conveniently categorized based on their func-
tional requirement for a specific nucleophile (Figure 21.1). From a synthetic
perspective, the most useful and most extensively studied enzymes are (i) one
acetaldehyde dependent aldolase, (ii) pyruvate/phosphoenolpyruvate dependent
aldolases, (iii) dihydroxyacetone phosphate/dihydroxyacetone dependent aldolases,
and (iv) glycine dependent enzymes. However, the discovery of novel, but less
abundant, aldolases using other donors continues.
Members of the first two types produce a-methylene carbonyl compounds and
thereby generate only a single aldol stereocenter, while members of the latter two
types form a,b-disubstituted carbonyl derivatives containing two new vicinal chiral
centers at the new CC bond, a fact that makes them particularly appealing for
asymmetric synthesis. Following deprotonation of the enzyme bound nucleophile
the approach of the aldehyde acceptor to form the new carbon–carbon bond usually
occurs stereospecifically, following an overall retention mechanism. The relative
positioning and differentiation of the appropriate face of the aldehyde carbonyl is
responsible for the enantio- or diastereoselectivity during the aldol attack. In this
manner, the stereochemistry of the CC bond formation is strictly controlled by the
enzymes, in general irrespective of the constitution or chirality of the substrate,
which renders the configuration of the product highly predictable. Whereas most
aldolases are quite specific for their donor compound in the aldolization reaction,
they often tolerate a wide range of aldehydes as acceptor components. This feature
makes possible a powerful combinatorial-type generation of a structural diversity,
varying in the acceptor part but keeping the common donor motif and the enzyme-
induced chirality. Using an aldolase of identical donor specificity but distinct
stereoselectivity can be used to selectively generate a complementary array of
stereoisomeric compounds; such situations are found with enzymes involved in
the carbohydrate and amino acid metabolism (Scheme 21.2). In cases where wild-
type enzymes give rise to practical limitations associated with the permissible
substrate repertoire or stereoselectivity, directed evolution has been used successfully
to improve the catalytic abilities of aldol formation.
The aldolase family of enzymes is further divided into two classes based on the
different enzyme mechanisms employed to activate the nucleophilic component.
Activation of the enzyme bound aldol donor substrates occurs by stereospecific
deprotonation along two distinct pathways (Figure 21.2) [21, 22]: class I aldolases
exhibit a strictly conserved lysine residue in the active site, which forms a covalent
Schiff base intermediate with the donor compound to generate an enamine
860 j 21 Aldol Reactions
OH O OH O OH O OH O
OPO3H2 OPO3H2 OPO3H2 OPO3H2
R R R R
OH OH OH OH
FruA RhuA FucA TagA

si,si re,re re,si si,re
DHAP 3
O
R H
glycine 4
L-ThrA D-ThrA D-allo-ThrA L-allo-ThrA

si,si re,re re,si si,re
OH O OH O OH O OH O
R OH R OH R OH R OH
NH2 NH2 NH2 NH2
Scheme 21.2 Combinatorial explosion originating from a single aldehyde substrate, using two
families of aldolases distinct in specificity for nucleophile and stereo-configuration.
nucleophile [23]. (As a variation, glycine-utilizing aldolases activate their substrate

with the aid of pyridoxal phosphate as an imine forming cofactor.) In contrast, class II
aldolases utilize a divalent metal ion cofactor to promote enolization of the donor
substrate via bidentate Lewis acid complexation [24]. Usually, this effect is achieved by
a tightly bound Zn2 þ ion but some other divalent cations can act in its place [21]. The
nucleophilic enamine or enolate then attacks the carbonyl carbon of the acceptor
substrate to initiate CC bond formation (or reverse for cleavage). Class II aldolases
are often more stable than class I aldolases, which is important for synthetic
purposes, and occur only in prokaryotes or lower eukaryotes.
For most aldolases that are attractive for synthetic applications, the CC bond
forming processes are favored by thermodynamic relations [25]. In less favorable
cases, the product fraction at equilibrium may be increased by driving the reaction
with a higher concentration of one of the reactants. Individual choice will certainly
enzyme enzyme His His His His
NH2 NH Zn His Zn His

X O X O
O X OH OH
R
X
R H
H R' O R
O 2C O HO2C
H
H
(a) enzyme—O (b) enzyme O—enzyme
Figure 21.2 Schematic mechanism of class I aldolases (a) and of class II aldolases (b).
depend primarily on the cost of starting material and ease of its separation from
product when used in large excess. However, enzyme inhibition by substrate(s) or
product may be a critical factor independent of the class types because for most lyases
both the donor and acceptor components contain strong electrophilic sites such as
aldehyde or ketone carbonyl groups, which may covalently modify an enzyme in and
out of the active site and thereby compromise its catalytic activity.
21.1.2
2-Deoxyribose 5-Phosphate Aldolase (EC 4.1.2.4)
In the group of acetaldehyde-dependent aldolases, only one enzyme is known so far,

namely, the 2-deoxy-D-ribose 5-phosphate aldolase (RibA or DERA; EC 4.1.2.4).
RibA is a class I enzyme found in the DNA salvage pathway of many microorganisms.
In vivo, it catalyzes the reversible aldol addition of ethanal (1) to D-glyceraldehyde
3-phosphate (5; Scheme 21.3) to furnish 2-deoxy-D-ribose 5-phosphate (6). Hence, it is
unique among the aldolases in that it catalyzes the aldol reaction between two
aldehydes, rather than using a ketone as the natural aldol donor.
O RibA OH O
2-O
CHO + 2-O
3PO H 3PO H
OH OH
5 1 6
Scheme 21.3 Metabolic reaction catalyzed by 2-deoxyribose 5-phosphate aldolase (RibA).
RibA has a strong preference for its phosphorylated acceptor substrate but also
tolerates uncharged aldehydes up to a chain length of four non-hydrogen atoms
instead. 2-Hydroxyaldehydes are relatively good acceptors, and the D-isomers are
preferred over the L-isomers [26]. Reactions that lead to thermodynamically unfa-
vorable structures may proceed with low stereoselectivity at the reaction center [27].
Recently, a single-point mutant aldolase was found to be 2.5 times more effective than
the wild type in accepting unphosphorylated glyceraldehyde [28, 29]. Interestingly,
the enzyme from Escherichia coli shows a somewhat relaxed specificity for its donor
substrate, where propanal (10), propanone (7a), or fluoropropanone (7b) can replace
1, albeit at strongly reduced (<1% of vmax) catalytic rates. Thus, the enzyme can
catalyze chain elongation by two or three carbon atoms to form, for example, variously
substituted 3-hydroxyketones (or 3-hydroxyaldehydes) such as 8a,b or 11
(Scheme 21.4) [26, 30].
Inspired by its natural function, RibA has been applied in a multienzymatic
commercial process for the production of different purine- or pyrimidine-containing
deoxyribonucleosides such as 13 in good yield (Scheme 21.5) [31, 32].
Synthetic applications of RibA include the preparation of pyranose building
blocks, such as 15, that are useful as chiral key intermediates for the synthesis of
anticancer agents epothilone A and C (Scheme 21.6) [33]. Various sugar derivatives,
such as 2-deoxy, thio-, and deoxyaza-sugars have also been prepared by using RibA
catalysis [26, 30]. Starting from azidoaldehydes, several azasugars containing a low
862
O O RibA OH O
H3 C + X H3C X
H
H3 C 7a X=H H3C 8a,b
b X=F
N3
O O RibA O
OH
R
+
N3 H H
OH HO CH3
D-9 10 11
O
O N3
H O NH
OH H2 HO
N3 H
RibA Pd/C
OH HO HO
D-9 12
Scheme 21.4 Utilization of non-natural aldol donors, and azasugar precursors prepared by
stereoselective RibA catalysis.
DHAP stage 1
FruA
FBP TPI
O
O OH O stage 2
H
2- 2-
O3PO H O3PO H
RibA
OH OH
6
5
NH2
N
PPM adenine
N N
PNP
HO O OPO32– HO O N
HO Pi HO 13
Scheme 21.5 Two-stage aldolase-based process that is useful for the synthesis of
deoxyribonucleosides.
O
S
O O RibA O OH steps HO N
HO +
O
14 1 OH 15 O OH O
epothilone A
Scheme 21.6 Chemoenzymatic synthesis of a chiral epothilone A building block based on an

aldolization reaction.
degree of functionalization (e.g., 12; Scheme 21.4) have been prepared by sequential
aldolization–hydrogenation [26].
Preparatively, the most attractive feature of this enzyme derives from the fact that
both its substrates, and thus the products, are aldehydes. Consequently, RibA has a
unique ability to catalyze cascade processes because the first reaction furnishes an
aldehyde that can be employed as a new acceptor for a subsequent aldol addition.
When 1 is used as the only substrate, the sequential addition of two donor molecules
yields (3R,5R)-2,4,6-trideoxyhexose (16) (Scheme 21.7) [34, 35]. Intramolecular
cyclization of 16 to give a stable hemiacetal masks the aldehyde function and thereby
effectively precludes formation of higher order adducts. When the first acceptor is an
a-substituted acetaldehyde, related aldol products from twofold donor additions can
be prepared that are very useful chiral synthons. By minimizing the number of
synthetic operations while maximizing the buildup of structural and functional
complexity, such highly step-economical tandem aldol reactions are particularly
appealing in the context of target-oriented synthesis. Indeed, the reaction with
chloroethanal (17), which is a particularly good acceptor substrate, leading to
(3R,5S)-6-chloro-2,4,6-trideoxyhexose (18) has most recently been developed into
an industrial process for the large-scale manufacturing of the chiral skipped polyol
side chain in statins, cholesterol-lowering drugs related to compactin
(Scheme 21.7) [36–38]. From 18, further variously substituted c-lactones (19–21)
are accessible that are useful as versatile chiral building blocks [39]. One problem was
that RibA is inactivated by the aldehydes at the high concentrations necessary for
O O
O 1 OH O 1 OH OH O
O
OH
R R R
RibA RibA
1 OH 16
O O O RibA OH OH O HO OH
+ +
Cl Cl O
R
H
17 1 1 S configuration 18
Cl
oxidation
HO O HO O O Nu O
H+ Nu–
O O O O
O
19 20 21
O Cl Cl Cl
compactin
Scheme 21.7 RibA catalyzed sequential aldol additions, yielding a key chiral building blocks, e.g. 18
as a precursor for cholesterol lowering drugs.
864
economical large-scale synthesis. To overcome the inhibition, one solution was to

develop a fed-batch process wherein the substrates are added to the reaction mixture
at the rate at which they are consumed so that they do not reach inhibitory
concentrations [37]. Another solution was to apply directed evolution in vitro to
increase the wild-type enzymes resistance to 17 as well as its volumetric productivity,
which resulted in identifying a variant with about tenfold improved productivity
under industrially relevant conditions [38]. As a third alternative, RibA enzymes from
hyperthermophilic organisms, such as from Pyrobaculum aerophilum and Thermo-
toga maritima, were screened and found to show much higher stability than wild-type
RibA from E. coli against inactivating agents [40].
Combination of two aldolases for sequential aldol additions, such as RibA for the
initial and FruA (D-fructose 1,6-bisphosphate aldolase) or NeuA (N-acetylneuraminic
acid aldolase) (vide infra) for a consecutive reaction, has also been studied towards
the synthesis of unnatural sugars [27, 34].
21.1.3
Pyruvate/Phosphoenolpyruvate-Utilizing Aldolases
Pyruvate (2) dependent lyases are a large family of enzymes that serve catabolic
functions in vivo, such as in the degradation of sialic acids and 2-keto-3-deoxy
aldonic acid intermediates from hexose or pentose catabolism. Important represen-
tatives in this group are aldolases that degrade N-acetylneuraminic acid (Neu5Ac; 23),
2-keto-3-deoxy-D-manno-octosonate (KDO; 26), 2-keto-3-deoxygluconate (KDG), 2-
keto-3-deoxy-6-phosphogluconate (KDPG; 63), or its 4-epimer 2-keto-3-deoxy-6-phos-
phogalactonate (KDPGal; 64). Pyruvate-dependent aldolases are usually class I aldo-
lases that reversibly bind their substrates via Schiff base/enamine formation.
These freely reversible aldol additions often have less favorable equilibrium
constants [21, 25], which often means that synthetic reactions have to be driven by
an excess of one substrate to achieve satisfactory conversions – for economic reasons
this usually is 2. A few related enzymes have been identified that utilize phospho-
enolpyruvate (PEP) instead of 2, which upon CC bond formation releases
inorganic phosphate, and thus renders the aldol addition essentially irreversible
(Scheme 21.8) [8]. Although attractive from a synthetic point of view, the latter types of
enzymes have been less studied yet for preparative applications [12, 41].
21.1.3.1 N-Acetylneuraminate (NeuNAc) Aldolase (EC 4.1.3.3) and NeuNAc Synthetase

(EC 4.1.3.19)
N-Acetylneuraminic acid aldolase (or sialic acid aldolase, NeuA; EC 4.1.3.3) catalyzes
the reversible addition of 2 to N-acetyl-D-mannosamine (ManNAc; 22) in the degra-
dation of the parent sialic acid 23 (NeuNAc; Scheme 21.8). The NeuA lyases promote a
Si-face attack of 2 to the aldehyde carbonyl group with formation of a (4S) configured
stereocenter. The enzyme from E. coli is commercially available; it has a broad pH
optimum around 7.5 and useful stability in solution at ambient temperature [42].
NeuA-catalyzed reactions are among the most studied and have been extensively
exploited for the synthesis of 23 and its analogues, because sialo-conjugates are
O
+ NeuA
OH CO2H OH OH OH O
NHAc 2
O HO
HO S CO2H
HO OH OPO32- OH NHAc
22 + NeuS
CO2H
Pi
phosphoenol
pyruvate
OH
HO OH
O CO2H
AcHN
HO OH
23
Scheme 21.8 Natural substrates of the N-acetylneuraminic acid aldolase (NeuA) and
N-acetylneuraminic acid synthetase (NeuS).
involved in several important physiological functions and pathological processes, such

as cellular recognition and communication, bacterial and viral infection, and tumor
metastasis [43]. The natural substrate of the aldolase has been an especially important
target for synthesis. Zanamivir was introduced in 1999 for the treatment of influenza
infection and represented the first medicinally used sialic acid derivative with antiviral
activity. As the precursor 23 is a rare and expensive natural product, this required a
cost-effective synthesis. Thus, NeuA was the first aldolase to find attention for the
development of an industrial bioconversion process that is now operated at multi-ton
scale (Scheme 21.9) [44–47]. In this equilibrium-controlled reaction (equilibrium
constant 12.7 M1 in favor of the retro-aldolization), the expensive ManNAc substrate
22 can be produced by an integrated enzymatic in situ isomerization from inexpensive
N-acetylglucosamine (24) by a combination of NeuA and an N-acylglucosamine 2-
epimerase catalyst (EC 5.1.3.8) in an enzyme membrane reactor [48–50]. Product
isolation, however, is complicated by the excess of 2 needed to achieve a preparatively
OH
HO
O O CO2H
HO HO
HO OH OH
NHAc
24 AcHN
HN NH2
Zanamivir
NH
epimerase
steps
OH
HO OH
HO NHAc NeuA
HO O
pyruvate O CO2H
HO OH AcHN
22 HO OH
23
Scheme 21.9 Industrial process for the production of N-acetylneuraminic acid as a precursor to an
influenza inhibitor.
866j 21 Aldol Reactions
useful level of conversion [46]. To facilitate product recovery, excess 2 can be removed
by formation of a separable bisulfite adduct, or by decomposition with yeast pyruvate
decarboxylase (PDC) into volatiles [44, 51]. An alternative method to reduce inhibition
effects from 2 is its in situ formation from inexpensive and innocuous lactate by the
action of an oxidase, as was reported for a coupled whole-cell biocatalytic synthesis of
23 from 22 on a large scale [52]. Another approach utilized a coupling of bacteria
expressing independently the 2-epimerase activity and PEP-dependent NeuS (Neu-
NAc synthetase; EC 4.1.3.19) activities, respectively [53].
As an example for continuous process design, KDN (25) has been produced on a
100-gram scale from D-mannose and 2 using a pilot-scale enzyme membrane reactor
(EMR) at a space–time yield of 375 g l1 d1 and an overall crystallized yield of 75%
(Scheme 21.10) [54]. Similarly, L-KDO (L-26) can be synthesized from L-arabinose [55].
Directed evolution of the wild-type NeuA from E. coli was employed to engineer a
mutant variant with improved preference for L-arabinose for L-KDO synthesis [56].
OH
HO OH HO OH
NeuA NeuA HO O CO2H
O CO2H
D-mannose HO L-arabinose
pyruvate S pyruvate
HO OH OH
HO
25 D-KDN 26 L-KDO
Scheme 21.10 Sialic acids prepared on larger scale.
Isotope labeled sialic acids can be prepared enzymatically for determination of

kinetic isotope effects or for spectroscopic tracing experiments. Starting from [3-
18
O]-labeled ManNAc (27), the corresponding [6-18 O]-NeuNAc (28) carrying the
label at the ring oxygen was prepared for mechanistic studies (Scheme 21.11) [57].
Similarly, [3-13 C] labeled NeuNAc (29) was prepared from unlabeled 23 by a one-pot
retro-aldol/forward aldol strategy in the presence of [3-13 C] labeled pyruvate (30),
exploiting the reversible nature of the aldolase reaction (Scheme 21.11) [58]. To
ensure complete conversion into the desired direction, the first step was promoted
by a cofactor dependent reduction of 2; the second, synthetic step, required
destruction of the nucleotide cofactor before addition of labeled 30. Sialic acid
29 and several chemically prepared derivatives were obtained with high label
incorporation (>87%) in good yields (46–76%). Double [3,9-13 C]-labeling of 23
and 25 from appropriately labeled D-ManNAc and D-Man precursors has also been
reported [59].
Extensive studies have indicated that only 2 is acceptable as the NeuA donor
substrate, with the exception of 3-fluoropyruvate (32) being the only permissible
variation (Scheme 21.12) [60–62]. Based on this observation, the synthesis of 3-
fluorinated NeuNAc diastereomeric derivatives 33/34 was reported in acceptable
yield (44%) for the preparation of fluoro-labeled sialo-conjugates 35/36. In addition, a
series of related 3-deoxy-3-fluoro-ulosonic acids became accessible from pentoses or
hexose derivatives, which are of interest as mechanistic probes for kinetic and crystal
structural studies of sialic acid processing enzymes or for non-invasive in vivo
pharmacokinetic studies by NMR tomography (19 F derivatives) or positron emission
spectroscopy (18 F derivatives).
OH
HO OH
HO NHAc
O NeuA
HO O CO2H
H OH AcHN
pyruvate HO OH
27 28
OH OH
HO OH HO OH
HO NHAc 1. NPP'ase
NeuA O 2. NeuA
O CO2H HO O CO2H
AcHN
HO OH
HO OH AcHN
HO OH
♦
23 22 29
OH O O
L-LDH
CO2H
2
CO2H ♦ CO2H 30 stage 2
NAD+ NADH
EtOH acetaldehyde
ADH stage 1
Scheme 21.11 Synthesis of isotope-labeled N-acetylneuraminic acids by direct aldol synthesis or

by controlled reversible aldolization (. and ¤ denote [18 O]- and [13 C]-label, respectively).
O
F OH OH
CO2H HO HO
OH OH
HO HO 32
HO O O F CO2H
+ O CO2H
HO OH NeuA HO HO
HO OH HO OH
31 33 F 34
CTP, CSS CSS, CTP

Galβ-OR, 2,3SiaT 2,6SiaT, Galβ-OR
OH HO OH OH
HO HO2C HO CO2H
O
O O OR O F O
HO F HO
HO OH OH HO OH HO
O
35 36 HO OR
OH
Scheme 21.12 Use of fluoropyruvate as non-natural donor substrate for the synthesis of fluoro-
labeled sialo-conjugates.
More importantly, the NeuA enzyme displays a fairly broad tolerance for various
aldehyde substrates stereochemically related to ManNAc as alternative aldol accep-
tors, such as several sugars and their derivatives larger or equal to pentoses [42, 55, 63,
64]. Permissible variations include replacement of the natural D-manno configured
substrate 22 with derivatives containing modifications such as epimerization,
substitution, or deletion at positions C2, -4, or -6 (e.g., 37) [8, 19]. Epimerization
OH OH
X NHCOR X OH X OH
NeuA
HO O
O CO2H O CO2H
HO OH pyruvate RCOHN AcHN
HO OH HO OH
37 38 39
NHBoc
X = OH a X = OCH3, OMOM,
R = OtBu, CH2OH, CH2Ph b X = OAc, OBz
O
c X = OCH2OCH2CH2(CF2)5CF3
O
NH
OH OH HO OH
O OH HO OH HO
O O H O CO2H
AcHN CO2H O N N
Bn HO
HO OH HO OH OH
40 O 41 42
HO OH
HO NHAc NeuA HO2C NeuA
HO O ll O CO2H ll
X OH pyruvate AcHN
HO HO
43 44
X = N3, NH2, NHBoc
Scheme 21.13 Neuraminic acid derivatives accessible by NeuA catalysis, including an

intermediate for alkaloid synthesis; also shown are the limits of substrate tolerance of NeuA in the
direction of synthesis or cleavage.
at C2, however, is restricted to small polar substituents due to strongly decreasing

reaction rates [65, 66]. The broad substrate tolerance of the catalyst for sugar
precursors has been exploited in the equilibrium generation of sialic acid and
analogs for an in situ screening of a dynamic combinatorial library [67, 68].
On account of the importance of sialic acids in a wide range of biological
recognition events, the aldolase has become instrumental for the chemoenzymatic
synthesis of a multitude of other natural and unnatural derivatives or analogs of 23
(Scheme 21.13). Many examples have been reported for sialic acid modifications at
C5/C9 [8, 19] such as differently N-acylated derivatives 38 [69–72], including amino
acid conjugates [73], or 9-modified analogs 39 [74–76] from suitable mannosamine
precursors 37 in the search for new neuraminidase (influenza) inhibitors. Most
notably, the N-acetyl group in 22 may be either omitted [65, 66] or replaced by
sterically demanding substituents such as N-Cbz (41) [77, 78] or even a nonpolar
phenyl group [65] without destroying activity. Large acyl substituents are also
tolerated at C6, as shown by the conversion of Boc-glycyl derivative 40 as a precursor
to a fluorescent sialic acid conjugate [79]. Similarly, C6 derivatives containing a
2-(perfluorohexyl)ethoxymethyl tag were shown to be acceptable for the NeuA
enzyme to serve both as a protecting group and efficient purification aid for the
aldol products by using fluorous separation technology [80]. The 7-fluoro [81] and
8-O-methyl [82] analogs of NeuNAc also became accessible from the corresponding
ManNAc derivatives. In contrast, no 3-azido, 3-amino, or Boc-protected mannosa-
mine analogues 43 were accepted by the enzyme (Scheme 21.13) [83], which suggests
Enz Enz
O NeuA OH OH
HO H HO H HO
3S 3R
α-4C1 OH
Enz Enz
O O
H Enz H Enz
H
si-face re-face
normal attack inverted attack
OH
OH
HO OH OH
HO
NeuA O AcHN O NeuA
N-acetyl-D- AcHN CO2H HO CO2H N-acetyl-L-
mannosamine pyruvate S pyruvate mannosamine
HO OH R
OH
23 D-NeuAc ent-23 L-NeuAc
OH
OH
HO OH OH
HO
NeuA HO O NeuA
O CO2H
D-mannose HO HO CO2H L-mannose
pyruvate S pyruvate
HO OH R
OH
25 D-KDN ent-25 L-KDN
HO
HO OH OH
HO
NeuA HO O O NeuA
CO2H HO CO2H D-arabinose
L-arabinose
pyruvate OH R pyruvate
HO OH
26 L-KDO ent-26 D-KDO
Figure 21.3 Three-point binding model for prediction of NeuA stereoselectivity based on
conformational analysis, and the unusual formation of mirror-image products with inverted (4R)-
configuration.
that the presence of a 3-hydroxyl group is a specific precondition for substrates of the
aldolase. Likewise, conformationally inflexible acrylate 44 was not accepted in the
cleavage direction.
In most cases investigated so far, a high level of asymmetric induction by NeuA for
the (4S)-configuration is retained. However, some carbohydrates were also found to
be converted with random or even inverse stereoselectivity for the C4 configuration,
such as for ent-23, ent-25, or ent-26 (Figure 21.3) [55, 64, 84–87]. A critical and
distinctive factor seems to be recognition of the configuration at C3 in the aldehydic
substrate by the enzymic catalyst, which essentially means that the stereochemical
outcome of the aldol reaction is unexpectedly determined by the substrate [55, 64].
A three-point binding model has been proposed to account for the inverse confor-
mational preference in direction of synthesis (Figure 21.3) [2]. On the basis of the
(3S)-a-4 C1 structure of the natural substrate and a conformational analysis of its
analogs, this model can predict the occasionally observed compromise in, or even
total inversion of, the facial stereoselectivity of CC bond formation.
870
Starting from the N-Cbz-protected aldolase product 41, iminocyclitol 42 has been
obtained stereoselectively by intramolecular reductive amination as an analog of the
bicyclic, indolizidine-type glycosidase inhibitor castanospermine (Scheme 21.13) [78].
In addition, it had been recognized that the C12–C20 sequence of the macrolide
antibiotic amphotericin B resembles the b-pyranose tautomer of 46 (Scheme 21.14).
Thus, the branched-chain manno-configured substrate 45 was successfully chain-
extended under NeuA catalysis to yield the potential amphotericin B synthon 46 in
good yield [88, 89].
OH OH
HO OH HO2C OH
OH
HO NeuA
HO O O CO2H O OH
HO OH pyruvate HO HO OH
45 46 HO OH
OH OH
12 OH
O OH
HO O OH OH OH OH O 16
CO2H
20
amphotericin B OR
Scheme 21.14 NeuA-catalyzed preparation of a synthetic precursor to the macrolide antibiotic

amphotericin B.
Recently, efforts have been directed towards an evolution of NeuA mutants that
tolerate substrate modifications in an effort to facilitate the synthesis of novel
neuraminidase inhibitors [90–92]. Screening efforts identified mutant E192N to have
a 50-fold higher kcat/Km for the chiral tartaric N,N-dipropylamide semialdehyde (47)
than the wild-type enzyme [90, 91]. To improve the rather low diastereoselectivity of
this mutant, subsequent focused mutagenesis of active site residues resulted in a pair
of stereochemically complementary (S)-selective (E192N, T167G) and (R)-selective
(E192N, T167V, S208V) NeuA variants, each useful for the synthesis of (4S)- and (4R)-
configured diastereoisomeric NeuNAc mimetics 48/49 (Scheme 21.15) [92]. An in-
depth discussion of this and further examples of the improvement of pyruvate
aldolases by directed evolution are contained in recent reviews [93, 94].
Whereas the inefficient equilibrium constant of the NeuA reaction usually
requires an excess of 2 to drive product formation, this complication may be
circumvented altogether by coupling of the aldol synthesis (e.g., 50 , 51) to a
thermodynamically more favored process, for example, by combination with a
practically irreversible formation of sialo-conjugates (e.g., 52) via CMP-sialate
synthase (CSS) catalyzed nucleotide activation followed by sialyl transfer
(Scheme 21.16). This principle has been utilized early on for the one-pot preparation
of complex sialylated oligosaccharides including in situ cofactor regeneration [95, 96].
Recently, this methodology has been applied to the synthesis of structurally diverse
NeuA OH
E192N/T167G N
O CO2H
HO
si attack O OH 4S 48
OH
N pyruvate
O
O OH NeuA OH
47 OH
E192N/T167V/S208V N
O CO2H
HO
re attack O 4R 49
Scheme 21.15 Generation of a pair of stereochemically complementary NeuA mutants by directed

evolution for the synthesis of sialic acid mimetics.
O OH
R2 OH
R1 NeuA
R2 HN
O H O CO2H
HO N
HO OH pyruvate
R1 HO OH
50 O 51
CTP
OH CSS
HO
O PPi
HO OR3
OH HO OH
R2 CO2H R2 O CMP
OH 2,6SiaT
H O O H O CO2H
N O N
R1 HO OH HO OR 3 R1 HO OH
O 52 HO CMP O 53
Scheme 21.16 Overcoming yield limitations from less favorable aldol equilibrium by coupling to
thermodynamically favorable in situ activation/sialyl transfer cascade.
p-nitrophenol-tagged 2,3- and 2,6-linked sialoglycoside libraries for substrate

specificity studies of sialidases [97]. In a related study, sialyl-Tn derivatives were
generated by one-pot enzymatic synthesis for the preparation of sialoside-protein
conjugates [98].
Using an unprecedented ability of NeuA to accept disaccharide b-D-Galp-(1,6)-D-
Man (54) as an acceptor, the efficient synthesis of 9-glycosylated KDN (55) was
accomplished containing a sialic acid in non-terminal position, which represents an
unusual disaccharide component of the cell wall of Streptomyces sp. MB-8
(Scheme 21.17) [99]. A subsequent study showed that various other disaccharides
carrying a reducing D-Man or D-ManNAc residue are tolerated as NeuA acceptors
even when containing a sterically more demanding 1,4-linkage (e.g., 56) [100, 101].
The tolerance of the PEP-dependent neuraminic acid synthetase (NeuS) in higher
organisms to accept acyl-modified ManNAc analogs as acceptors to produce cell
surface oligosaccharides modified in their neuraminic acid constituents has allowed
the development of novel methods for metabolic labeling of living organisms
HO OH HO OH
O NeuA O OH
HO O HO O OH
HO
HO HO O pyruvate HO
OH O CO2H
HO 85% HO
HO OH
Galβ1,6Man 54 Galβ1,9KDN 55
HO OH OH
HO OH
O HO HO NeuA
HO O O HO OH O CO2H
HO OH pyruvate HO
HO O
38% HO O OH
Galβ1,4Man 56 HO Galβ1,7KDN 57
Scheme 21.17 Disaccharides with reducing D-Man as NeuA substrates for the synthesis of non-
terminal sialo-glycosides.
(Scheme 21.18) [102, 103]. The reactive ketone group in the N-acyl chain of the non-
natural N-levulinoyl D-mannosamine (58), thus displayed by the cellular machinery
on the cell surface in vivo, could be utilized in a versatile fashion for covalent cell
redecoration under physiological conditions by attaching functional nucleophiles
(Nu ) such as fluorescent hydrazine markers or toxin conjugates. More advanced
studies have employed azidoacetyl (59) or alkinoyl mannosamine derivatives that are
incorporated into cellular oligosaccharides with improved efficiency and that can be
utilized by click chemistry-based ligation for advanced non-invasive cell imaging
and for dynamic investigations of glycan processing using cross-reactive fluorescent
probes [104].
O cell Nu* OH
HO CO2H
HO HN O
HO O O O cell
O NH
HO OH [NeuS, PEP] HO OH
58 O
O cell OH
"click" HO
N3 CO2H
HO HN
HO O O O cell
HO OH NH
[NeuS, PEP] N3 HO OH
59 O
Scheme 21.18 Cellular synthesis of modified sialic acids by exposure of human cells to
D-mannosamine derivatives, generating opportunities for bio-orthogonal labeling of cell surface
oligosaccharides.
21.1.3.2 3-Deoxy-D-manno-2-octulosonate (Kdo) Aldolase (EC 4.1.2.23)

2-Keto-3-deoxy-manno-octosonate (KDO) aldolase (KdoA, EC 4.1.2.23) is involved
in the catabolism of the eight-carbon sugar KDO D-26, which is reversibly degraded to
O KdoA OH OH O
OH +
O HO CO2H
OH CO2H R
OH OH OH
HO 60 2
OH
HO HO
OH OH F OH
HO HO HO
O O O
HO CO2H HO CO2H HO CO2H
61 OH 62 OH D-26 OH
Scheme 21.19 Natural substrates of the 2-keto-3-deoxy-manno-octosonic acid aldolase (KdoA),

and non-natural sialic acids obtained by KdoA catalysis.
D-arabinose 60 and 2 (Scheme 21.19). The enzyme has been partially purified from
bacterial sources and studied for synthetic applications [87, 105]. It seems that the
KdoA, similar to NeuA, has broad substrate specificity for aldoses while pyruvate was
found to be irreplaceable. As a notable distinction, KdoA was also active on smaller
acceptors such as glyceraldehyde. Preparative applications, for example, for the
synthesis of KDO (D-26) and its homologs or analogs 61/62, suffer from an
unfavorable equilibrium constant of 13 M1 in the direction of synthesis [25]. The
stereochemical course of aldol additions generally seems to adhere to a Re-face attack
on the aldehyde carbonyl, which is complementary to the stereoselectivity of NeuA.
On the basis of the results published so far it may be concluded that a (3R)-
configuration is necessary (but not sufficient), and that stereochemical requirements
at C2 are less stringent [87].
21.1.3.3 2-Keto-3-deoxy-6-phosphogluconate (KDPG) Aldolase (EC 4.1.2.14)

and 2-Keto-3-deoxy-6-phosphogalactonate Aldolase (EC 4.1.2.21)
A class I aldolase specific for cleavage of 2-keto-3-deoxy-6-phospho-D-gluconate (63)
(KDPGlc aldolase or GlcA; EC 4.1.2.14) is central to the Entner–Doudoroff glycolytic
pathway to give 2 and 5 (Scheme 21.20). The equilibrium constant favors synthesis
(103 M1) [106]. Comparable to the situation for the NeuA and KdoA enzyme pair
(vide supra), a related class I lyase is also known that acts on 2-keto-3-deoxy-6-phospho-
D-galactonate (64) (KDPGal aldolase or GalA; EC 4.1.2.21) and thus has a comple-
mentary stereopreference for the (4S)-configuration (Scheme 21.20).
GlcA enzyme preparations from liver or microbial sources were reported to show
rather high substrate specificity for the natural phosphorylated acceptor D-63 [107].
Enzymes isolated from E. coli, Pseudomonas putida, and Zymomonas mobilis by dye-
ligand chromatography [108] were shown to offer a rather broad substrate tolerance
for polar, short-chain aldehydes, at, however, much reduced reaction rates [109–112].
Simple aliphatic or aromatic aldehydes are not converted. Recombinant GlcA from
Thermotoga maritima was shown to provide products of the same absolute config-
uration as mesophilic enzymes but with diminished stereoselectivity [113]. GalA
from Pseudomonas cepacia [112] or E. coli [114] was also studied and found to share the
2-
OH O O3PO OH
GlcA O
2-
O3PO S CO2H S CO2H
O OH HO
63
2-
pyruvate
O3PO H
OH O 2-
O3PO OH
OH O
5 2-
HO
O3PO R CO2H CO2H
GalA OH R
64
O OH O OH
GlcA GlcA O
O CO2H
HO H S
H S CO2H
pyruvate pyruvate
OH OH OH HO
D-65
HO 66 67
GlcA OH
O
KA3-L1
O CO2H
HO H HO
pyruvate
OH HO
L-65 68
Scheme 21.20 Aldol reactions catalyzed in vivo by the stereo-complementary 2-keto-3-deoxy-6-

phospho-D-gluconate aldolase (GlcA) and 2-keto-3-deoxy-6-phospho-D-galactonate aldolase (GalA);
synthesis of hexulosonic acids catalyzed by GlcA and evolved L-selective mutant.
high affinity for phosphorylated and D-configurated substrates. Therefore, the GlcA
from E. coli has been mutated for improved acceptance of non-phosphorylated
substrates and for L-configured aldehydes to facilitate the development of enzymatic
syntheses of both D- and L-sugars [115, 116]. The E. coli GlcA was applied to prepare
both enantiomers of 2-keto-4-hydroxyglutarate (70) by direct aldol formation to yield
the (S)-enantiomer (L-70) and by racemate resolution (i.e., (S)-selective retro-aldol
cleavage) to leave the antipode D-70 (Scheme 21.21) [107]. Bacterial in vivo selection
against 2-keto-4-hydroxyoctonate, a pentanal-derived non-substrate for wild-type
aldolase, produced GlcA variants capable of rescuing pyruvate-auxotrophic cells
[117]. Mutated variants created to perturb the phosphate-binding pocket of GlcA were
identified to show up to 2000-fold improved selectivity for unnatural substrates and
40-fold improved catalytic efficiency [118].
Wild-type GlcA enzymes have been used to prepare deoxysugar acids 66/67 from
D-glyceraldehyde (D-65) and D-lactaldehyde, respectively [119], whereas conversion of
L-glyceraldehyde (L-65) to give the diastereomeric 68 was made possible when using
an adapted enzyme variant [115]. High stereoselectivity and activity of GlcA enzymes
towards pyridine 2-carbaldehyde (72) has been utilized in a two-step enzymatic
synthesis of 73 (via 71), the unbranched N-terminal amino acid portion of nikko-
mycin antibiotics (vide infra). The latter are a group of potent chitin synthase
inhibitors regarded as promising fungicidal agents in agriculture and human therapy
(Scheme 21.22) [119, 120]. The mirror image precursor ent-71 could be generated by
O O GlcA OH O
+
HO2C H CO2H HO2C CO2H
69 2 L-70
70%, >95% ee
OH O GlcA OH O O O
+ +
HO2C CO2H HO2C CO2H HO2C H CO2H
DL-70 D-70 69 2
78%, 60% ee
CO2 NADH
FDH LDH
HCO2H NAD+
lactate
Scheme 21.21 Preparation of both enantiomers of 4-hydroxyketoglutarate by direct GlcA-catalyzed

synthesis or racemate resolution.
GlcA GalA
CO2H H CO2H
N pyruvate(2) N pyruvate N
71 OH O 72 O ent-71 OH O
ee >99.7%
NH3 NADH CO2
FDH
NAD+ HCO2H O
O CO2H NH
CO2H N
N N N O
H O
73 OH NH2 OH NH2
HO OH
Nikkomycin Kz
Scheme 21.22 Stereoselective synthesis of the amino acid portion of nikkomycin antibiotics using
enantiocomplementary aldolases.
using the corresponding GalA activity [112, 114]. A GalA variant, obtained by directed
evolution, exhibits a 60-fold improved activity in catalyzing the addition of pyruvate
(2) to D-erythrose 4-phosphate 74 to form 3-deoxy-D-arabino-heptulosonic acid 7-
phosphate (DAHP; 75). As the latter is the entry metabolite to the biosynthesis of
aromatic amino acids, this GalA variant allowed the production of 3-dehydroshiki-
mate in a strain deficient of the essential DAHP synthase (Scheme 21.23) [121, 122].
Hyperthermophilic archaea Sulfolobus are assumed to metabolize glucose via a
non-phosphorylated Entner–Doudoroff pathway. However, the aldolase was found to
contain a novel phosphate binding site and to be more active with 5 than with non-
phosphorylated substrates, equivalent to bacterial GlcA enzymes. Analysis of
enzymes from several Sulfolobus subtypes revealed that the enzymes readily accept
876
CO2H
GalA HO CO2H
OH
2–O NR8.276-2 O steps
3PO
O pyruvate
OH OH O OH
2–O OH
3PO OH
74 75 DAHP 3-dehydroshikimate
Scheme 21.23 Utilization of a mutant GalA obtained by directed evolution for the synthesis of
DAHP as a metabolic bypass entry into the shikimic acid pathway.
GlcA
O (Sulfolobus) OH O OH O
+
HO H HO CO2H HO CO2H
pyruvate
OH 50°C OH OH
D-65 (4S,5R)-76 50 : 50 (4R,5R)-77
GlcA
O (Sulfolobus) OH O OH O
+
O H O CO2H O CO2H
pyruvate
O 50°C O O
D-78 79 96 : 4 80
O OH O OH O
MPS
+
O H O CO2H O CO2H
oxaloacetate
O O O
D-78 79 8 : 1 80
Scheme 21.24 Substrate engineering to improve the stereocontrol of a promiscuous GlcA, and
application of macrophomate synthase (MPS) for stereoselective aldol synthesis.
various polar aldehydes with two to four carbon atoms [123, 124]. Surprisingly, this
GlcA exhibits no diastereocontrol in the aldol addition of its natural substrates and
furnishes D-KDGlc (76) and D-KDGal (77) in approximately equal amounts from 2
and D-glyceraldehyde (D-65, Scheme 21.24) [123]. A similar lack of stereoselectivity
was observed for additions to L-glyceraldehyde (L-65) [125], as well as to the four
aldotetroses [126]. Substrate engineering by way of more rigid D- and L-glyceraldehyde
acetonides (e.g., D-78) resulted in the greatly improved stereoselective formation of
the corresponding anti-(4S,5R)-adduct (79) and syn-(4S,5S)-adduct (80) with >92%
and >94% d.e., respectively [125].
Macrophomate synthase (MPS) from Macrophoma commelinae catalyzes the
synthesis of macrophomate via formation of two CC bonds in a multistep reaction
cascade from oxalacetate and 2-pyrone. Although long considered a rare case of
enzymatic Diels–Alderase reactivity, it has been discovered recently that this enzyme
can form pyruvate enolate from oxaloacetate, followed by stereoselective aldol
addition to various aldehydes, as exemplified by the addition to D-78
(Scheme 21.24) [127]. These findings strongly corroborate an alternative two-step
Michael-aldol sequence instead of the suggested Diels-Alder pathway as the most
plausible mechanism of macrophomate synthesis.
21.1.3.4 SanM and 4-Hydroxy-3-methyl-2-keto-pentanoate Aldolase (EC 4.1.3.39)

Quite recently, novel enzymes have been discovered that utilize 2-oxobutanoic acid
(81) as aldol donor component. Unlike the pyruvate-dependent aldolases, the
catalyzed addition in these cases results in the creation of two stereogenic centers
at the newly developed CC bond. Also at variance to the pyruvate aldolases, the
2-oxobutanoate aldolases seem to belong to class II aldolase types, that is, they require
divalent metal ions for activity.
The 4-hydroxy-3-methyl-2-keto-pentanoate aldolase (HkpA) from Arthrobacter
simplex AKU 626, which is specific for the (3R)-configuration, catalyzed the formation
of the (3R,4S)-stereoisomer 82 with a sixfold diastereomeric preference
(Scheme 21.25). By coupling to a stereospecific branched-chain amino acid amino-
transferase (BcaT), a one-pot bi-enzymatic transformation could be realized for the
preparation of (2S,3R,4S)-4-hydroxyisoleucine (83). The latter compound possesses
insulinotropic bioactivity and therefore is of interest for the treatment of type II
diabetes [128, 129].
L-glutamate α-ketoglutarate
O O OH O OH NH2
HkpA BcaT
+
H CO2H CO2H CO2H
1 81 (3R,4S)-82 (2S,3R,4S)-83
Scheme 21.25 Coupled enzymatic synthesis of insulinotropic 4-hydroxyisoleucine (83) using a

novel diastereoselective aldolase activity.
Likewise, the aldolase SanM from a gene cluster of Streptomyces ansochromogenes,

coding for the biosynthesis of the nikkomycin peptidyl nucleoside antibiotics, was
found to catalyze a (3R,4S)-diastereoselective aldol addition between 81 and picoli-
naldehyde (72). In this transformation, 84 is formed as a precursor to the branched-
chain hydroxypyridyl-homothreonine part of peptidyl antibiotics (e.g., nikkomycin X,
Scheme 21.26). Interestingly, the aldolase was shown to be active only in the presence
of a dehydrogenase SanN coded for by a second biosynthetic gene [130].
21.1.4
DHA/DHAP-Utilizing Aldolases
Dihydroxyacetone phosphate (DHAP, 3) dependent aldolases constitute a family of

lyases, which are involved in glycolytic cleavage of glucose and related pathways of
hexose metabolism. Whereas pyruvate aldolases form only a single stereogenic
center, aldolases specific for 3 as a nucleophile create two new asymmetric centers at
878
SanN SanM
SCoA H CO2H
N N CO2H N
O O OH O
72 84
O 81
O CO2H NH
N
N N O
H O
OH NH2
HO OH
Nikkomycin X
Scheme 21.26 Diastereoselective synthesis of a branched keto acid by a novel class II aldolase as
part of the biosynthesis of nikkomycin antibiotics.
the termini of the new CC bond. Particularly useful for synthetic applications is the
fact that Nature has evolved a full set of four unique aldolases (Scheme 21.27) to
cleave all possible stereochemical permutations of the vicinal diol at C3/C4 of ketose
1-phosphates 85–88 during the retro-aldol cleavage [8, 14]. A plethora of studies has
been reported that demonstrate their synthetic usefulness. These aldolases have
proved to be exceptionally powerful tools for asymmetric synthesis, particularly for
the stereocontrolled synthesis of polyoxygenated compounds, because of their
relaxed substrate specificity, high level of stereocontrol, and commercial availability.
In the direction of synthesis this situation formally allows us to generate all four
O
X
H O
OH OPO32-
OH 3
2-O
3PO OH O OH O
FruA RhuA
OPO32- H3C OPO32-
3S,4R 3R,4S
HO HO 85 HO HO 86
D-fructose 1,6-bisphosphate L-rhamnulose 1-phosphate
2-O
3PO OH O OH O
TagA FucA
OPO32- H3C OPO32-
3S,4S 3R,4R
HO HO 87 HO HO 88
D-tagatose 1,6-bisphosphate L-fuculose 1-phosphate
Scheme 21.27 Aldol reactions catalyzed in vivo by the four stereo-complementary

dihydroxyacetone phosphate dependent aldolases.
possible stereoisomers of a desired product in a building block fashion [8]. In this
manner, the deliberate preparation of a specific target molecule can be addressed by
simply choosing the corresponding enzyme and suitable starting material, thereby
offering full control over constitution as well as absolute and relative configuration of
the desired product. All DHAP aldolases are quite specific for the phosphorylated
nucleophile 3, which therefore must be prepared independently or generated in situ.
A kinetic model consisting of an ordered two-substrate mechanism and possible
inhibition modes has been developed as a step towards applications on an industrial
scale, and validated for batch and fed-batch synthesis [131].
21.1.4.1 Fructose 1,6-Bisphosphate Aldolase (EC 4.1.2.13)

The D-fructose 1,6-bisphosphate aldolase (FruA; EC 4.1.2.13) catalyzes in vivo the
equilibrium addition of 3 to D-glyceraldehyde 3-phosphate (5) to give D-fructose 1,6-
bisphosphate (85, Scheme 21.28). The equilibrium constant for this reaction of
104 M1 strongly favors synthesis [25]. The enzyme occurs ubiquitously and has been
isolated from various prokaryotic and eukaryotic sources, both as class I and class II
forms [21]. Typically, class I FruA enzymes are tetrameric, while the class II FruA are
dimers. As a rule, the microbial class II aldolases are much more stable in solution
(half-lives of several weeks to months) than their mammalian counterparts of class I
(a few days) [132–134].
OH O 2-O PO OH
3 O
2-
O3PO O + HO OPO3 2- FruA HO
OPO32-
5 3 HO 85
Scheme 21.28 Natural glycolytic substrate of the fructose 1,6-bisphosphate aldolase (FruA).
Traditionally, the class I FruA isolated from rabbit muscle (RAMA) is the aldolase
employed for preparative synthesis in the widest sense owing to its commercial
availability and useful specific activity of 20 U mg1. Its operative stability in
solution is limiting, but the more robust homologous enzyme from Staphylococcus
carnosus has been cloned for overexpression [135], which offers unusual stability for
synthetic purposes.
Literally hundreds of aldehydes have so far been tested successfully by enzymatic
assay and preparative experiments as a replacement for 5 in rabbit muscle FruA
catalyzed aldol additions [8, 17], and most of the corresponding aldol products have
been isolated and characterized. In vitro, unhindered aliphatic aldehydes and a--
heteroatom-substituted aldehydes, including different aldoses (C3–C5), are generally
suitable substrates. Phosphorylated substrates are usually preferred over nonpho-
sphorylated ones. The aldol reaction thus leads to an elongation by three carbon
atoms with formation of the respective a-keto-sugar 1-phosphates. Aromatic alde-
hydes, sterically hindered aldehydes, and a,b-unsaturated aldehydes are usually not
substrates. It was shown that less polar substrates may be converted as highly
concentrated water-in-oil emulsions [136]. The rabbit FruA can discriminate racemic
DL-5,its natural substrate, with high preference for the D-antipode, but kinetic
enantioselectivity for nonionic chiral aldehydes is rather low [132, 137].
21.1.4.2 Fuculose 1-Phosphate Aldolase (EC 4.1.2.17), Rhamnulose 1-Phosphate

Aldolase (EC 4.1.2.19) and Tagatose 1,6-Bisphosphate Aldolase (EC 4.1.2.40)
In catabolic pathways for some rare sugars, aldolases are involved that also utilize 3 as
a donor (Scheme 21.28) but result in the other three possible stereo-configurations at
carbon atoms C3 and C4 of the respective aldol products.
The D-tagatose 1,6-bisphosphate aldolase (TagA; EC 4.1.2.40) is central to the
catabolism of D-galacto-configured carbohydrates; it catalyzes the reversible cleavage
of D-tagatose 1,6-bisphosphate (87) to 3 and D-5 (Scheme 21.29). Enzymes of class I
seem to have apparently no stereochemical selectivity with regard to distinction of 4-
epimeric 85/87 [138], while class II TagA types show a high stereoselectivity for the
natural substrate in both cleavage and synthesis directions [139, 140].
O glycerol O
kinase
HO OH HO OPO32-
89 3 2-O OH
ATP ADP 3PO O
TagA HO HO
TPI OPO32
-
pyruvate 87
OH
pyruvate kinase PEP
O OPO32-
H 5
Scheme 21.29 Enzymatic one-pot synthesis of tagatose 1,6-bisphosphate based on the

stereoselective TagA from E. coli.
Utilizing the synthetic capacity of a TagA purified from E. coli the all-cis (3S,4S)-
configured D-tagatose 1,6-bisphosphate 87 has been prepared from 89 by an expe-
ditious multienzymatic system (Scheme 21.29) [139, 140]. The aldolase also accepts a
range of unphosphorylated aldehydes as substrates but produces diastereomeric
mixtures only. This lack of stereoselectivity with generic substrate analogs, which
makes native TagA enzymes synthetically less useful, has stimulated protein engi-
neering studies to improve its properties [141].
L-Fuculose 1-phosphate aldolase (FucA; EC 4.1.2.17) and L-rhamnulose 1-phos-
phate aldolase (RhuA; EC 4.1.2.19) are found in many microorganisms, where they
are responsible for the degradation of deoxysugars L-fucose and L-rhamnose to give 3
and L-lactaldehyde (L-90) (Scheme 21.30). FucA is specific for cleavage and synthesis
of a D-erythro diol unit while RhuA recognizes the corresponding L-threo configura-
tion. Both enzymes are active as Zn2 þ -dependent homotetramers [134, 142]. Like
several other aldolases, both the RhuA and FucA enzymes are commercially available
or can be efficiently overproduced [143, 144].
Overall practical features make the FucA and RhuA enzymes quite similar for
synthetic applications. Both metalloproteins are quite robust under conditions of
OPO32- RhuA O FucA OPO32-
HO O DHAP DHAP O
H3C
OH H OH
H 3C H 3C
OH OH HO OH
86 L-90 88
Scheme 21.30 Natural substrates of microbial deoxysugar phosphate aldolases.
organic synthesis and show a very high stability in the presence of low Zn2 þ
concentrations with half-lives in the range of months at room temperature. The
enzymes even tolerate the presence of large fractions of organic cosolvents
(30%) [134], and they are active in highly concentrated water-in-oil emulsion
systems [145, 146]. Both offer a very broad substrate tolerance for variously substi-
tuted aldehydes, which is very similar to that of the FruA enzymes. Characteristically,
the RhuA has the greatest tolerance for sterically congested acceptor substrates,
as exemplified in the conversion of the tertiary aldehyde 2,2-dimethyl-3-
hydroxypropanal [8].
The stereospecificity of both enzymes for an absolute (3R)-configuration is
mechanism-based (vide supra). FucA generally directs an attack of the DHAP enolate
to the Si-face of an approaching aldehyde carbonyl and thereby is specific for
synthesis of a (3R,4R)-cis diol unit [134, 147], while RhuA controls a Re-face attack
to create the corresponding (3R,4S)-trans configuration [134]. However, this spec-
ificity for a vicinal configuration is somewhat substrate dependent, in that simple
aliphatic aldehydes can give rise to a certain fraction of the opposite diastereomer
[8, 134, 146]. Stereocontrol in general is usually highly effective with aldehydes
carrying a 2- or 3-hydroxyl group. In addition, both aldolases offer a powerful kinetic
preference for L-configured enantiomers of 2-hydroxyaldehydes (91, Scheme 21.31),
which facilitates racemate resolutions [148, 149]. Essentially, this feature allows the
concurrent determination of three contiguous chiral centers in final products 92 or
93, having an L-configuration (d.e. 95) even when starting from the more readily
accessible racemic material.
OPO32- O
HO O
RhuA R
OH + H
R
O OH OH
R 92 de ≥ 90% D-91
H + DHAP
2-
OH OPO3 O
O
D,L-91 FucA R
+ H
R OH
HO OH OH
93 de ≥ 90% D-91
R = H3C–, H5C2–, H2C=CH–, H2C=CH–CH2–, FH2C–, N3CH2–, H3COCH2–
Scheme 21.31 Kinetic enantiopreference of class II DHAP aldolases useful for resolution of
racemic a-hydroxyaldehydes.
21.1.4.3 Synthetic Strategies, Stereoselectivity, and Product Diversity Using
DHAP-Dependent Aldolases
The synthetic utility of the DHAP-dependent aldolases has been thoroughly dem-
onstrated with a wide array of novel acceptor aldehydes. Typical applications of the
DHAP aldolases concern the synthesis of monosaccharides and derivatives of sugars
from suitable functionalized aldehyde precursors. Complex eight- and nine-carbon
monosaccharide derivatives (such as 94; Figure 21.4) could be obtained from pentose
and hexose monophosphates by stereospecific chain extension using FruA from
rabbit muscle [150]. High conversion rates and yields are generally achieved with 2- or
3-hydroxyaldehydes because in such cases reaction equilibria profit from the fact that,
in aqueous solution, the products will cyclize to give more stable furanose or pyranose
isomers. For example, enantiomers of glyceraldehyde (65) are good substrates, and
stereoselective addition of 3 produces enantiomerically pure ketohexose 1-phos-
phates in high yield [132, 134, 148], from which the free keto-sugars are obtained by
OH O HO
OPO32- P'ase O
HO HO
(a) OH
RhuA OH OH HO
O OH
DHAP L-fructose
HO H
FucA OH O OH
OH P'ase OH
OPO32- O
HO HO
DL-65
OH OH HO
OH
L-tagatose
(b)
OH OH OH
2-O
3PO
OPO32- OPO32- H 3C O
OH
O O HO
OH H3C OH H3 C OPO32-
HO OH
OH HO OH HO
94 95 96
(CH3)2N
H
N O OCH3 O O OH
SO2 OPO32-
HO HO
OH O
OH OH
CH3
HO HO HO
97 98 99
N
HO OH N
OH OH HO NH2
O O
HO F17C8 O HO
HO OH N N
HO
OH
OH HO HO
100 HO 101 102
Figure 21.4 (a) Aldolase-catalyzed asymmetric synthesis of uncommon L-configured sugars;

(b) selected examples of carbohydrate-related product structures accessible by enzymatic
aldolization using FruA (94–102).
enzymatic dephosphorylation. For example, the less common L-configured L-fructose
and L-tagatose can be prepared directly from racemic 65 by RhuA or FucA catalysis
making use of the high kinetic enantioselectivity of Zn2 þ -dependent aldolases
(Figure 21.4) [148, 151].
The general approach has been followed for the de novo synthesis of a multitude of
differently substituted, unsaturated [152, 153], or regiospecifically 13 C-labeled
sugars [154, 155]. Unusual branched-chain (95,96) and spiro-annulated sugars
(98,99) have been synthesized from the corresponding aldehyde precursors
(Figure 21.4) [156, 157]. 6-Substituted D-fructofuranoside derivatives such as aro-
matic sulfonamide 97 (a low nanomolar Trypanosoma brucei inhibitor) [158] are
accessible via 6-azido-6-deoxyfructose from 3-azido-2(R)-hydroxypropanal (9) by
FruA catalysis [133, 159]. In an approach resembling the inversion strategy (vide
infra) a-C-mannoside 100 has been prepared from D-ribose 5-phosphate [160]. The
synthesis of 6-C-perfluoroalkyl-D-fructose (101) met the challenges of the strong
hydrophobicity and electron-withdrawing capacity of a fluorous chain, as well as the
products potential surfactant properties [161]. The L-sorbo-configured homo-C-
nucleoside analog 102 has been synthesized as a structural analog to adenosine
from an enantiopure (S)-aldehyde precursor [162]. On the basis of FruA catalyzed
aldol reactions, DAHP (75) has been synthesized from N-acetylaspartic semialde-
hyde (103) via adducts 104 (Scheme 21.32) [163]. Precursors 106 to KDO and its
4-deoxy analog have been prepared by FruA catalysis from aldehydes 105 that
incorporate an acrylic moiety for further functionalization [164].
2-
FruA O3PO OH OPO32-
H CO2CH3 DHAP COOH
HO O
HO CO2H
O NHAc (AcO)3BH– OH OH AcHN
OH
103 104 75 DAHP
OH
FruA HO HO
X OH OH X
H DHAP O
X CO2H
O P'ase O HO
COOH COOH OH
X = OH D-KDO
105 106 X=H 4-deoxy-KDO
Scheme 21.32 Synthetic approaches to DAHP and KDO by a backbone inversion strategy using
FruA catalysis.
Fluorogenic compound 108 for transketolase assays has been prepared making use
of FruA specificity [165]. Pendant anionically charged chains have been extended
from O- or C-glycosidic aldehydes to furnish low molecular weight mimics of the
sialyl LewisX tetrasaccharide such as 107 (Figure 21.5) [166]. Other higher carbon
sugar derivatives such as the bicyclic sugar 111 have been prepared by diastereo-
selective chain extension of simple alkyl galactosides (110) after their terminal
oxidation in situ by using a galactose oxidase (GalO). The whole scheme can be
884
HO
OH OH O
OH
O OH O O O O OPO32-
2- OH
O3PO OH
107 108
OH
OH O
HO OPO32- GPO HO OPO32-
2-
O3PO
OH
109 3 HO
Cat O
O2 H2 O 2 RhuA OH
O
HO OH HO HO OCH3
CHO
O O OH
HO OCH3 GalO HO OCH3
OH OH
110 OH
HO H O OCH3
HO O H OH
2-
O3PO OH
111
X
Figure 21.5 Sialyl Lewis -related selectin inhibitor 107 and fluorogenic screening compound 108 for
transketolase prepared using enzymatic aldolization, and a multienzymatic oxidation–aldolization
strategy for the synthesis of bicyclic higher carbon sugars.
conveniently effected as a one-pot operation including the parallel generation of 3 by

the GPO (glycerol phosphate oxidase) method (vide infra) [167]. Further bicyclic
carbohydrate structures similar to 111 have also been achieved by unidirectional [135]
and bidirectional extension of dialdehyde substrates (Scheme 21.42 below) [168].
Class II aldolases are effective in the kinetic resolution of racemic 2-hydroxyalde-
hydes (Scheme 21.31). Under fully equilibrating conditions, however, diastereos-
electivity of aldolase reactions can be steered also by thermodynamic control to favor
the energetically most stable product [132, 156, 169, 170]. Particularly strong
discrimination results from utilization of 3-hydroxylated aldehydes such as 112
owing to the cyclization of products in water to form a pyranoid ring (Scheme 21.33).
The pronounced conformational destabilization by diaxial repulsions (114) strongly
supports those diastereoisomers having a maximum of equatorial substituents
[168, 170]. Thus, in FruA-catalyzed reactions (3S)-configured hydroxyaldehydes are
the preferred substrates to give the most stable all-equatorial substitution in the
product (such as 113) with a d.e. up to 95%. Similarly, 2-alkylated aldehydes can be
resolved because of the high steric preference of an alkyl group for an equatorial
position [169]. Owing to the enantio-complementary nature of the FruA–RhuA
biocatalyst pair, under conditions of thermodynamic control this enables the con-
struction of mirror imaged products 113 (FruA) and ent-113 (RhuA) from racemic 3-
hydroxybutanal 112 with similar selectivity, but preference for opposite enantio-
mers [2]. The all-equatorial substitution in the predominant product can facilitate its
separation by crystallization so that the remaining mixture may be re-subjected to
OH OH O OH
S OPO3=
H 3C H3C O OH
OH O OH OPO3= HO 113
FruA
H 3C H + DHAP
OH OH O H3C OH
112 OPO3=
R O
H3 C OH
OH OPO3= 114
HO
1. FBP, FruA, TPI HO OH

OH O NaIO4
2. P'ase OH
X 3C O X3C O
X 3C H OH OH
NaBH4
OH OH
115 116 X = H, F 117 X = H, F
Scheme 21.33 Diastereoselectivity in FruA-catalyzed aldol additions to 3-hydroxyaldehydes under

thermodynamic control, and the synthesis of L-fucose derivatives based on thermodynamic
preference.
further equilibration to maximize the yield of the preferred isomer 113 [154]. This
general technique was applied in a novel approach for the de novo synthesis of 4,6-
dideoxy sugars such as 4-deoxy-L-fucose or its trifluoromethylated analog (117;
Scheme 21.33) via stable ketose intermediates 116 [2].
Because of the structure of nucleophile 3, the enzymatic aldolization technique is
ideal for the direct synthesis of ketose monosaccharides and related derivatives or
analogs. However, the product invariably is a ketone, while some of the most
desired products would be the corresponding aldehydes. This problem has been
addressed synthetically by two different strategies, namely (i) by incorporation of a
masked aldehyde function into the electrophilic substrate to be released after
aldolization and (ii) by employing enzymatic ketol isomerization after product
dephosphorylation.
As a first entry to aldoses, the inversion strategy has been developed
(Scheme 21.34), which utilizes monoprotected dialdehydes (e.g., 118/120) for
aldolization and, after stereoselective ketone reduction (e.g., in 119), provides free
aldoses upon deprotection of the remaining masked aldehyde function [171]. In
addition, this method when using the phosphorothioate analog 121 (vide infra) also
makes terminally deoxygenated sugars accessible via a sequence of enzymatic
aldolization followed by chemical reductive desulfurization (e.g., of 122), as illus-
trated by the FruA-catalyzed preparation of D-olivose along the inversion
strategy [172]. Otherwise, deoxy sugars are usually only attained when the deoxy
functionality is introduced by the aldehyde.
A more general access to biologically important and structurally more diverse
aldose isomers makes use of ketol isomerases for the enzymatic interconversion of
ketoses into aldoses. For a full realization of the concept of enzymatic stereodivergent
carbohydrate synthesis, the stereochemically complementary L-rhamnose isomerase
aldolase
O OR DHAP O OH OR reduction OH OH O
P'ase HO deprotection HO
H OR OR H
OH OH
118 119
2-deoxyaldose
O
2-O
3PS OH 1. NaBH 4
121 2. H3O+ H3C
O O O OH O O
HO
2-
O3PS HO OH
H O FruA O 3. H2/Ni
4. HCl D-olivose
OH
120 122
Scheme 21.34 Inverted approach for aldose synthesis using FruA catalysis, and application of
the strategy for deoxysugar synthesis based on a phosphorothioate analog.
(RhaI) and L-fucose isomerase (FucI) from E. coli have been shown to display a relaxed
substrate tolerance [8, 148, 153, 173]. Both enzymes convert sugars and their
derivatives that have a common (3R)-OH configuration but may deviate in stereo-
chemistry or substitution pattern at subsequent positions of the chain [8, 14]. Because
ketose products from RhuA and FucA catalyzed aldol reactions share the (3R)
specificity they can both be converted by the isomerases into corresponding aldose
isomers, which provides access to a broad segment of aldose configurational space in
a stereospecific, building block manner [14, 174]. This strategy has been illustrated by
tandem FucA–FucI catalysis in the synthesis of new L-fucose analogues 124 having
tails with increased hydrophobicity and reactivity (Scheme 21.35), starting from
simple higher homologues and unsaturated analogs 91 of lactaldehyde (90), as well as
by the synthesis of L-rhamnose (126) and other L-configured aldohexoses using
different enzyme combinations [153, 173]. Similar results have been realized by
utilizing a glucose isomerase (GlcI), which is an industrially important enzyme for
the isomerization of D-glucose to D-fructose. The latter enzyme has a narrower
specificity for D-fructose modifications but could be used in combined enzymatic
syntheses, particularly of 6-modified D-glucose derivatives [175].
OH 1. FucA, DHAP OH FucI

2. P'ase O OH
O R O
OH
R R OH
91 HO OH HO OH
123 124
R = CH3 : L-fucose
C2H5, CH=CH2, C≡CH
OH 1. RhuA, DHAP OH RhaI

2. P'ase O OH
O HO O
OH HO
OH OH
90 OH 126
125 L-rhamnose
Scheme 21.35 Short enzymatic synthesis of L-fucose and hydrophobic analogs (124) and of
L-rhamnose (126) by coupled aldolization–ketol isomerization, including kinetic resolution of
racemic hydroxyaldehyde precursors.
Phosphonate analogs to phosphate esters, in which the PO bond is formally
replaced by a PC bond, have attracted attention due to their stability towards the
hydrolytic action of phosphatases, which renders them potential inhibitors or
regulators of metabolic processes. Two alternative pathways, in fact, may achieve
introduction of the phosphonate moiety by enzyme catalysis. The first employs the
bioisosteric methylene phosphonate analog 128, which yields products related to
sugar 1-phosphates such as 129/130 (Scheme 21.36) [154, 176]. This strategy is rather
effective because of the inherent stability of 128 as a replacement for 3 but depends on
the individual tolerance of the aldolase for structural modification close to the reactive
center. The second option is the suitable choice of a phosphonylated aldehyde such as
131, which gives rise to analogs of sugar v-phosphates (132) [177, 178].
OH
O
HO
PO32-
FruA 129
HO
O O O PO32-
RhuA HO
HO + HO
H PO32- OH
130
OH
127 128 FucA
1. FruA/TPI O OH
O OH H FBP (EtO)2P OH
(EtO)2P ( )n O OH
( )n O 2. P'ase
HO
131 132 n = 1,2
Scheme 21.36 Complementary routes for the stereoselective synthesis of hydrolytically stable
sugar phosphonates, from either the bio-isosteric phosphonate analog of DHAP or phosphonylated
aldehydes.
Several cyclitol derivatives of varying ring size, for example, 134/137–140, have
been prepared via a chemoenzymatic carboligation cascade based on an enzymatic
aldolization as the initial step, which is followed by an intramolecular chemical
cyclization step, making use of the electrophilic carbonyl unit introduced upon
addition of 3. One strategy utilized halogen substitution in the aldehyde, allowing a
subsequent reductive cyclization via radical intermediate [179]. Aldehydes carrying a
suitably installed C,H-acidic functional group such as a nitro, ester, or phosphonate
functionality allow a facile base-catalyzed nucleophilic cyclization to occur with, or
subsequent to, the enzyme-catalyzed aldol addition (Scheme 21.37) [180–182]. As an
example, the synthesis of aminocyclitols has been achieved, initiated by the FruA
catalyzed aldol addition of 3 to nitrobutanal (135), followed by a rapid intramolecular
Henry-type cyclization occurring within the hydroxynitroketone intermediates
136 [183]. This twofold CC bond forming reaction cascade delivered, after nitro
888
1. FruA, AcO OH
DHAP O2N OH
OH 2. P'ase O AcO NO2
HO steps
O2 N OH
CHO
HO AcO
133 134
OH
NO2
OH
HO 1. FruA, HO HO
NO2 DHAP OH
CHO OPO22- HO NH2

OH O OH
135 136
valiolamine
2. P'ase
OH HO OH HO
OH HO OH HO
HO NO2 HO NO2 HO NH2 HO NH2
H2
+ +
HO OH HO OH PtO2 HO OH HO OH
137 138 139 140
Scheme 21.37 Preparation of aminocyclitol precursors by chemoenzymatic tandem aldolase-

Henry reactions.
group reduction, aminocyclitol analogues 139 and 140 of valiolamine, which is a

specific inhibitor of intestinal glycosidases.
When placing a thiol substitution in the aldehyde component for aldolase
catalysis, the reaction products can cyclize to form rather stable cyclic thiohemia-
cetal structures. Such thiosugars are a structural variation of carbohydrates that
possess interesting biological properties such as glycosidase inhibition. From 2-
and 3-thiolated aldehydes (e.g., 141), stereochemical sets of furanoid or pyranoid
thiosugars such as 142–144 have been prepared using different DHAP aldolases
(Scheme 21.38) [184–186]. Notably, the observed unbiased stereoselectivity indi-
cates the full equivalence of OH and SH substituents for correct substrate
recognition. As an alternative, 5-thio-D-xylopyranose 147, a xylosidase inhibitor
and useful chiral building block for the synthesis of antithrombotic D-xylopyrano-
side derivatives, was synthesized using FruA-catalyzed aldol addition of 3 to 2-
chloro (17) or 2-bromo acetaldehyde as a key step followed by sulfide substitution
(Scheme 21.38). The 5-thio-D-xylulofuranose (146) thus obtained was converted
into 147 by treatment with glucose isomerase [187].
The structural resemblance of azasugars (1-deoxy sugars in which an imino
group replaces the ring oxygen) to transition states or intermediates of glyco-
processing enzymes has made these compounds an attractive research object
because of their potential value as enzyme inhibitors for therapeutic applications.
An important and flexible synthetic strategy has been developed that consists of a
stereoselective enzymatic aldol addition to an azido aldehyde followed by azide
OH
HS
CHO
141
1. FruA, 1. FucA, 1. RhuA,
DHAP DHAP DHAP
2. P'ase 2. P'ase 2. P'ase
OH OH S
OH OH HO
S HO
S HO
HO OH HO OH
HO OH
HO OH
142 143 144
1. FruA, OH
O DHAP OH O S
2. P'ase NaSH HO
Cl Cl OH OH
H
OH 145 HO
17 146
40 : 60 GlcI
HO S
HO OH
OH
147
Scheme 21.38 Different strategies for the synthesis of thiosugars based on stereoselective
enzymatic aldolizations.
hydrogenation with intramolecular reductive amination. Numerous examples of the

application of DHAP aldolases have been reported for the synthesis of iminocyclitols
as specific glycosidase inhibitors [188–190].
Particularly noteworthy are the stereodivergent chemoenzymatic syntheses of
diastereomers of the nojirimycin type from 3-azidoglyceraldehyde (9) that have been
developed independently by several groups (Scheme 21.39) [133, 159, 191–194].
Because of the low kinetic selectivity of the rabbit FruA for 2-hydroxyaldehydes, use
of enantiopure aldehyde proved superior to the racemate for preparation of the
parent 1-deoxy-D-nojirimycin.
An extensive array of further five-, six-, and seven-membered ring alkaloid analogs
have since been made accordingly by following the same general strategy. For
structural variation, as exemplified by 150,152, or 155, differently substituted azido
aldehydes (149,9,156) or N-Cbz-protected amino aldehydes (153) of suitable chain
length were converted by the distinct DHAP aldolases (Scheme 21.40) [133, 145, 192,
195–199]. Stereocontrol during the reductive cyclization seems to be effected best by
Pd-catalyzed hydrogenation. The preparation of structurally diverse polyhydroxylated
pyrrolidine derivatives was achieved based on RhuA and FucA catalyzed processing
of N-Cbz-aminoaldehyde derivatives (e.g., 153), some of which proved inhibitory for
a-L-fucosidase, a-L-rhamnosidase and a-D-mannosidase activities [200]. RhuA toler-
ated a large structural diversity of acceptors while FucA did not convert Ca-branched
OH HO
NH FruA RhuA HN
HO OH
HO a a OH
HO HO
1-deoxynojirimycin
HO OH
HO RhuA FruA OH
HN OH HO NH
OH a a HO
CHO OHC
1-deoxymannojirimycin
R S
OH OH HO HO
HO OH
FucA N3 N3 TagA
NH HN
HO a D-9 L-9 OH
a
HO OH
HO OH HO OH
HO TagA FucA OH
HN NH
OH a a HO
Scheme 21.39 Stereodivergent synthesis of 1-deoxy azasugars of the nojirimycin type by two-step
enzymatic aldolization/catalytic reductive amination; a: (i) aldolase and DHAP, (ii) phosphatase,
(iii) H2, Pd/C.
H 1. FruA, 1. FucA, H
S N O S N
DHAP DHAP
HO OH S HO OH
HO H
2. P'ase 2. P'ase
HO OH 3. H2, Pd/C N3 3. H2, Pd/C HO OH
148 149 150
1. FruA, DHAP HO OH
OH 2. P'ase N3
N3 3. GlcI O H2 HO
O HO OH
HO OH
Pd/C
H OH N
9 151 H 152
H
O 1. RhuA, OH O N
DHAP H2
OH OH
H
2. P'ase Pd/C
NHCbz HO OH
CbzHN OH
153 154 155
O 2-O
3P
HO H2
O PO32- OH O N+
128 H2 OH
H PO32-
Pd/C
N3 FruA N3 OH HO
156 157 158
Scheme 21.40 Stereoselective synthesis of five- and seven-membered ring azasugars and of novel
azasugar phosphonates.
N-Cbz-aminoaldehyde derivatives; on the other hand, FucA was generally more
diastereoselective, whereas RhuA was selective only for the (S)-configurated accep-
tors. Interestingly, the type and steric bulk of N-protecting groups may crucially
influence the stereoselectivity of enzymatic aldol additions [199]. The technique has
been extended to the bifunctional class of azasugar phosphonic acids such as 158 by
exploiting the tolerance of the rabbit FruA for the bioisosteric phosphonate nucle-
ophile 128 [201]. In a strategy inverse to that employed for compound 158, FucA and
FruA were employed in the chemoenzymic synthesis of six-membered iminocyclitol
phosphonic acids [202].
Another illustrative example for the azasugar synthetic strategy concerns the
chemoenzymatic synthesis of the bioactive natural products australine, 3-epiaustra-
line (Scheme 21.41) and 7-epialexin [203]. The bicyclic pyrrolizidine core structure
resulted from twofold reductive amination of a linear precursor 160 in which the
asymmetric hydroxylation sites had been installed during an aldolase-catalyzed chain
extension from aminoaldehyde 159. Related glycosidase inhibitors of the hyacintha-
cine type, such as ()-hyacinthacine A2, have been prepared by a RhuA-catalyzed
aldol addition of 3 (DHAP) to N-Cbz-protected prolinal (161), followed by reductive
amination [204].
1. FruA, DHAP OH
OH O OH OH O
2. P'ase O
OH O H
H OH
OH
OHCHN OHCHN OH
159 160 O3 NHCHO
NaCNBH3 H2, Pd/C
HO OH HO OH
OH OH
N N
australine OH 3-epiaustraline OH
OH
Cbz O 1. RhuA, Cbz OH O H
N DHAP N OH H2
H OH
2. P'ase Pd/C N
OH
161 162 (–)-hyacinthacine A2 OH
1. O3 H
HO OH 2. FruA, DHAP HO N OH
3. P'ase HO OH
N3 O OH
4. H2, Pd/C HO HO
HO
(±) 163 164
Scheme 21.41 Synthetic routes to oxygenated pyrrolizidine alkaloids, and to an aza-C-disaccharide

as glycosidase inhibitors.
892
A bidirectional aldolization approach furnished the C-glycosidically linked azadi-

saccharide 164 as an example of a disaccharide mimic (Scheme 21.41). Ozonolysis of
a racemic azido-substituted cyclohexenediol precursor 163 was followed by tandem
additions of 3 to both aldehydic termini to yield an intermediate azido-substituted
dipyranoid 2,11-diulose which, when hydrogenated over Pd catalyst, highly selec-
tively gave the aza-C-disaccharide 164 as a single diastereomer [2].
Such aldolase catalyzed bidirectional chain elongation (tandem aldolization) of
simple, readily available dialdehydes has been developed into an efficient method for
the generation of higher-carbon sugars (e.g., 166/168) by simple one-pot operations
(Scheme 21.42) [168, 205]. The choice of furanoid (166) or pyranoid (168) nature of
the products can be determined by a suitable hydroxyl substitution pattern in a
corresponding cyclo-olefinic precursor (165 versus 167). The overall specific substi-
tution pattern in the carbon-linked disaccharide mimetics is deliberately addressable
by the relative hydroxyl configuration and choice of the aldolase. Single diastereo-
mers may be obtained in good overall yield from racemic precursors, if the tandem
aldolizations are conducted under thermodynamic control (see Scheme 21.33).
Similarly, highly complex structures like annulated (170) and spirocyclic (172)
1. O3 HO
O R
2. FruA, DHAP HO
3. P'ase HO OH
HO OH O
HO R HO
(±) 165 166 OH
HO
1. O3 OH
HO OH 2. FruA, DHAP OH
3. P'ase HO O R
OH
HO O
OH R OH
(±) 167 168 HO
OH
OH 1. FruA, DHAP OH
HO O
2. P'ase OH
CHO
OHC HO
O OH
(±) OH HO
OH
169 170
single
OH
diastereomer
1. O3 HO
2. FruA, DHAP
3. P'ase HO
HO OH OH
O O
HO OH
171 172
OH
OH
Scheme 21.42 Applications of bidirectional chain extension for the synthesis of disaccharide
mimetics and of annulated and spirocyclic oligosaccharide mimetics using tandem enzymatic aldol
additions, including racemate resolution under thermodynamic control.
carbohydrate mimics may be obtained from appropriately customized precursors
(Scheme 21.42) [168].
DHAP aldolases typically yield carbohydrates or carbohydrate-derived materials
according to the nature of the reactive components, but they may also be advanta-
geous in the construction of stereochemically homogenous fragments of non-
carbohydrate natural products. An impressive illustration is the FruA-based che-
moenzymatic syntheses of ( þ )-exo-brevicomin (Scheme 21.43), the aggregation
pheromone of the Western pine bark beetle Dendroctonus brevicomis [206]. Addition of
3 to 5-oxohexanal (173) generated an enantiopure vicinal syn-diol structure 174 which
includes the only independent stereogenic centers of brevicomin. A backbone-
inverting approach towards the same target made use of 5,6-dideoxyketose precursor
176 [207], which is easily generated by FruA catalysis from propanal (10) [132]. In a
related approach, transketolase has also been utilized for the stereo-differentiating
key step in the chemoenzymatic syntheses of ( þ )-exo-brevicomin from 2-hydro-
xybutyraldehyde via the intermediate 176 [207]. This enzyme only creates a single
chiral center but is highly efficient in the resolution of racemic 2-hydroxyaldehydes,
while the conversion can be driven thermodynamically by the choice of donor
substrate.
O OH O
1. P'ase O O
FruA OPO32- 2. H+
H OH
OH O
+ DHAP
173 174 175
O O O
O steps
{+)-exo-brevicomin
1. FruA,
DHAP acetone
O 2. P'ase O OH O
ZnI2
HO HO
H
10 OH 176 177 O O
Scheme 21.43 Complementary, backbone inverting approaches for the asymmetric synthesis of
the insect pheromone ( þ )-exo-brevicomin.
Application of an aldolase to the synthesis of the tricyclic microbial elicitor

()-syringolide (Scheme 21.44) is another excellent example of how enzyme-cata-
lyzed aldolizations can be used to generate sufficient quantities of enantiopure
material in multistep syntheses of complex natural and unnatural products [208].
Remarkably, the aldolase reaction established absolute and relative configuration of
the only chiral centers that needed to be externally induced in the adduct 179 from
achiral precursor 178; during the subsequent cyclization events, all other chiral
centers seemed to follow by kinetic preference.
894
O O
O O
HO OPO32–
HO OH
3 1. FruA 4 steps CH3(CH2)6
+ O
2. P'ase OMPM O
OMPM HO
O
O OH
178 179
OH O H+ 55%
CH3(CH2)6
O O
O
HO (–)-syringolide
Scheme 21.44 Aldolase-based creation of two independent chiral centers in the total synthesis of
the complex microbial plant defense elicitor ()-syringolide.
Using FruA catalysis and protected 4-hydroxybutanal 180, compound 181 has been
stereoselectively prepared as a synthetic equivalent to the C3–C9 fragment of
( þ )-aspicillin, a lichen macrolactone (Scheme 21.45) [209]. Similarly, FruA-mediated
stereoselective addition of 3 to a suitably crafted aldehyde precursor (182) served as
the key step in the synthesis of the non-carbohydrate, skipped polyol C9–C16 chain
fragment 183 of the macrolide antibiotic pentamycin [210, 211].
O 1. FruA OH OH
DHAP OH 9 OH
H
2. P'ase
BnO 180 BnO 181 O
O OH
3
OH O
(+)-aspicillin
MeO OBn MeO OBn HO HO OH OH OH OH

11 9 11 9
C5H11
1. FruA 13 13
O DHAP HO HO
OMe 14 OMe 14
OH O O
H 2. P'ase 16 O 16
182 HO 15 183 HO 15 CH3
H3 C pentamycin OH
OH
Scheme 21.45 Stereoselective generation of chiral precursors for the synthesis of the lichen
macrolactone ( þ )-aspicillin and of the macrolide antibiotic pentamycin, using FruA catalysis.
A two-stage enzymatic sequence of arene dihydroxylation, using a naphthalene

dioxygenase (NDO) from Pseudomonas putida, followed by ozonolytic ring cleavage to
yield dialdehyde 184 and RhuA-catalyzed aldolization has been developed for the
synthesis of novel analogs of the cytotoxic pancratistatin pharmacophore such as 186
(Scheme 21.46) [212]. This strategy converts a simple naphthalene core into a
complex hybrid arene–carbohydrate structure, with simultaneous creation of four
contiguous stereocenters, in just three steps.
O
1. NDO
2. O3
OH OH OH O
O CHO O OPO32-
RhuA
OH OH OH
O DHAP O
CHO CHO
184 185
P'ase,
Br2/CaCO3
HO
OH
OH OH
HO OH
HO OH OH O
O
+
O O O H OH
H OH H OH
O
NH O
O O O
O
OH O pancratistatin O 186 187
Scheme 21.46 Enzyme-catalyzed asymmetric synthesis of a pancratistatin analog using a

naphthalene dioxygenase (NDO) and RhuA-catalyzed aldolization for the rapid creation of four
contiguous stereocenters.
21.1.4.4 Synthesis of Dihydroxyacetone Phosphate (DHAP)

Apparently, all DHAP aldolases are highly specific for 3 as the donor component for
mechanistic reasons [21–24], which means that economical access to this compound
is needed for synthetic applications. While 3 is a regular metabolite of glycolysis, the
expense for commercial sources is prohibitive for preparative applications beyond
academic interest and continues to be a major drawback of these enzymes for
applications on a larger scale. Moreover, in most instances the phosphate group is
undesired in final products and must be removed in a separate reaction disfavoring
the atom economy of the process. In contrast, the phosphate ester moiety in aldol
products may facilitate their isolation, for example, by barium salt precipitation or by
using ion-exchange techniques. Subsequent dephosphorylation can be achieved by
mild enzymatic hydrolysis using an inexpensive alkaline phosphatase at pH
8–9 [184], or at pH 5–6 using a more expensive acid phosphatase for those
compounds that are base-labile [132].
Densely functionalized 3 can be prepared by several chemical or enzymatic routes
from various precursors (for an extensive review see Reference [213]). Enzymatic
methods usually are advantageous because they are often more step-economical and
can be interfaced directly to subsequent enzymatic applications. For example,
dihydroxyacetone (89) can be enzymatically phosphorylated using glycerol
kinase [214] or DHA kinase [215, 216] with ATP regeneration, or by transpho-
sphorylation from pyrophosphate [217]. Alternatively, 3 can be formed from glycerol
by successive phosphorylation and oxidation effected by a combination of glycerol
896
kinase and glycerol phosphate dehydrogenase, with an integrated double ATP/

NAD þ cofactor recycling system [184].
However, owing to the limited stability of 3 in solution, particularly at alkaline pH,
it is preferentially generated in situ for immediate consumption to avoid high
stationary concentrations. The most convenient method for in situ generation is
the retro-aldol cleavage of commercially available 85 with formation of two equiva-
lents of 3 by the combined action of FruA and triose phosphate isomerase
(Scheme 21.47 inset) [132]. This scheme has been further extended into a highly
integrated, artificial metabolism for the efficacious in situ preparation of 3 from
inexpensive feedstock such as glucose and fructose (two equivalents of 3 each), or
sucrose (four equivalents of 3) by a combination of up to seven inexpensive enzymes
in vitro (Scheme 21.47) [156]. When employing the class II FruA of E. coli for aldol
cleavage, which displays high substrate specificity for 5 and thus is inactive with most
aldehyde substrates for synthesis, this system can be metabolically engineered by
adding another aldolase to furnish products having a different stereospecificity [2].
A related multistep reaction cascade was assembled from the FruA equilibrium
to generate the labile 5 in situ for the transketolase-catalyzed synthesis of D-xylulose
5-phosphate [218].
sucrose or glucose or fructose
5 enzymes
ATP, PEP
2-
O3PO
O OH
HO
D-threo OPO32-
HO 85
FruAeco
O
OH O OH O
TPI R
O OPO32- HO OPO32- OPO32-
R
FucA
H 5 3 OH D-erythro
Scheme 21.47 Enzymatic in situ generation glycolysis cascade (top), and utilization
of dihydroxyacetone phosphate from fructose for subsequent stereoselective
1,6-bisphosphate (box), with extension to an in carbon–carbon bond formation using an
vitro artificial metabolism for its preparation aldolase with distinct stereoselectivity
from inexpensive sugars along the (bottom, right).
An advanced technique for the clean generation of 3 in situ is based on the oxidation
of L-glycerol 3-phosphate (109) catalyzed by microbial flavine-dependent glycerol
phosphate oxidases (GPOs; Scheme 21.48, box) [154]. This method generates 3
practically quantitatively and with high chemical purity without a need for separate
cofactor regeneration. Both oxygen from air or from a H2O2/catalase system can be
used to sustain oxygenation [154, 219]. Since DHAP aldolases were found to be
insensitive to oxygenated solutions, the oxidative generation of 3 can be smoothly
coupled to synthetic aldol reactions [154]. This method has been extended to include a
reversible glycerol phosphorylation by phytase, an inexpensive acid phosphatase,
from inexpensive pyrophosphate [220], or by controlled ring opening of glycidol by
inorganic phosphate [221].
Furthermore, the GPO procedure can also be used for a preparative synthesis of the
corresponding phosphorothioate (121), phosphoramidate (190), and methylene
phosphonate (128) analogs of 3 (Scheme 21.48) from suitable diol precursors [222]
to be used as aldolase substrates [154]. In fact, such isosteric replacements of the
phosphate ester oxygen were found to be tolerable by several class I and class II
aldolases, and only some specific enzymes failed to accept the less polar phosphonate
121 [176]. Thus, sugar phosphonates (e.g., 129/130) that mimic metabolic inter-
mediates but are hydrolytically stable to phosphatase degradation can be rapidly
synthesized (Scheme 21.36).
Decomposition of 3, which hampers the yield of enzymatic aldol additions, can be
considerably reduced by lowering the reaction temperature to 4 C, which constitutes
OH GPO O FruA OH O
2-
pH 7.5 2-
pH 7.5
HO OPO3 HO OPO3 OPO32-
butanal
+ 109 O2 H2O2 3 OH 188
1/2
OH
HO OPO32- H2O
Pi phytase phytase
pH 4.0 pH 4.0
PPi Pi
OH OH O
HO OH OH
4 steps, 1 pot
glycerol OH 189
O O O H O
HO O HO S HO N HO
PO32- PO32- PO32- PO32-
3 121 190 128
Scheme 21.48 Oxidative enzymatic controlled, integrated precursor preparation

generation of dihydroxyacetone phosphate in and product liberation; substrate analogs of
situ for stereoselective aldol reactions using dihydroxyacetone phosphate (3) accessible by
DHAP aldolases (box), and extension by pH- the GPO oxidation method (121, 128, and 190).
898
an optimum between residual aldolase activity and minimum rate for loss of 3 [223].
Various other reaction engineering approaches have been explored, in particular
means to allow the use of 89 as a donor for DHAP-dependent aldolases via transient
formation of DHAP mimics in situ. Interestingly, 89 in the presence of higher
concentrations of inorganic arsenate reversibly reacts to form the corresponding
arsenate ester 191a in situ, which mimics the natural phosphate ester of 3 as a donor
in enzyme-catalyzed aldol reactions (Scheme 21.49) [224, 225]. However, this
procedure suffers from rather low reaction rates and the high toxicity of arsenates.
Inorganic vanadate also spontaneously forms the corresponding vanadate ester
analog under conditions that reduce the unwanted redox activity of vanadate but
so far only RhuA could be shown to accept the vanadate mimic 191b for preparative
conversions [2]. As a complementary development, it was observed that the RhuA
enzyme converts 89 at reasonable rates when reactions are conducted in 200 mM
borate buffer (Scheme 21.49). Apparently, the DHA-borate ester 191c is formed in
situ, which is accepted by the enzyme as a DHAP mimic [226]. Control experiments
suggest that under the reaction conditions the product is further trapped by ensuing
formation of stable borate diesters (e.g., 192), thereby effectively shifting the reaction
equilibrium in the direction of synthesis.
O
HO OH
O
89 HX HO OX
191 a X = AsO32–
b VO 42–
DHAP
aldolase c BO32–
ketose aldehyde
ketose
X-ester
OH
HO O
65
HO OH
O O O
B(OH)3 O– RhuA
HO OH HO O O
B O–
–
89 191c OH HO O B O
–
192 O
Scheme 21.49 Spontaneous reversible formation of analogs of dihydroxyacetone phosphate in situ

for enzymatic aldol additions, and subsequent product trapping in the presence of borate buffer.
21.1.4.5 Transaldolase (EC 2.2.1.2) and Fructose 6-Phosphate Aldolase (EC 4.1.2.n)
The transaldolase (Tal, EC 2.2.1.2) is a class I aldolase enzyme that is involved in the
pentose phosphate pathway where it transfers a dihydroxyacetone unit between
several phosphorylated metabolites (Scheme 21.50) [21]. However, native transaldo-
lases cannot utilize free 89 as a nucleophile source (only at unacceptably low rates) but
require an v-phosphorylated ketose (such as fructose 6-phosphate, 193) as a source
(a)
D-fructose 6-phosphate D-glyceraldehyde 3-phosphate
transaldolase
+ +
O
D-erythrose 4-phosphate D-sedoheptulose 7-phosphate

OH OH
(b)
2-O OH OH O
3PO O
HO FSA 2-
O3PO O + HO OH
OH or
HO TalB F178Y
193 5 89
(c)
OH transketolase OH O
O OH
O O O O O O
O CO2 OH
194 196
HO transaldolase
CO2H
195
O base O O
+ OH
O O OH O O O +
197 OH
Scheme 21.50 (a) Metabolic function of 6-phosphate aldolase (FSA) and an evolved
transaldolase to shuffle a dihydroxyacetone transaldolase mutant; (c) fluorogenic assay
unit among sugar phosphates; (b) aldol principle for the screening of transaldolase-
cleavage reaction catalyzed by fructose type catalysts.
for the DHA moiety with concomitant release of a phosphorylated aldehyde (5).
Functionally related to transaldolase is the novel class I fructose 6-phosphate aldolase
(FSA) from E. coli, which unlike transaldolase catalyzes the fully reversible cleavage of
193, and thus its formation from free 89 and D-5 (Scheme 21.50) [228]. Recently, a
transaldolase B (TalB) variant of E. coli, in which Phe178 was replaced by Tyr, was
reported to show activity as a DHA dependent aldolase [230]. This single amino acid
substitution, which corresponds to the catalytic ensemble of FSA, changes the
enzyme from aldol transfer to a freely dissociating aldolase activity. In fact, the
TalBF178Y mutant has a capacity for the formation of 193 from 89 and D-5 at rates
similar to the FSA activity. This variant opens new possibilities in biocatalysis as
TalBF178Y and FSA differ in their tolerable spectrum for donor and acceptor sub-
strates [227, 231]. A set of fluorogenic substrates has been prepared, for example, 196
by transketolase catalysis, to establish a highly sensitive screening assay for deter-
mination of transaldolase stereoselectivity [229].
(a)
O O OH O OH
FSA O
HO + HO HO
H
OH OH HO
127 198
1-deoxy-D-xylulose
(b)
O O OH O O
FSA –H2O
+
H
O OH O OH HO O
199 198
furaneol
H
N
OH
(c)
H2, Pd/C OH
OH
CbzHN O O CbzHN OH O
FSA D-fagomine
+ OH OH
H
R
OH OH
200 89 201 R–CHO N
H2, Pd/C OH
OH
202 OH
Scheme 21.51 (a) Synthesis of 1-deoxysugars by FSA-catalyzed addition of hydroxyacetone;

(b) application to the preparation of furaneol, an industrial aroma compound; (c) chemoenzymatic
synthesis of fagomine and N-alkylated derivatives.
Remarkably, besides 89 FSA also utilizes hydroxyacetone (198) and 1-hydroxybu-

tanone as alternative donors with regiospecificity for the oxygenated carbon
nucleophile to generate 1-deoxysugars (such as 1-deoxy-D-xylulose) directly
(Scheme 21.51) [232–235]. A first industrial application of FSA catalysis concerns
the preparation of furaneol, an important compound used in the flavor and perfume
industry because of its sweet strawberry aroma, from 198 and pyruvaldehyde 199 with
in situ rearrangement/dehydration (Scheme 21.51) [236].
Yeast transaldolase was early on demonstrated to accept unphosphorylated alde-
hydes as the acceptor component in preparative studies [237–239]. More recently, a
multitude of synthetic applications using FSA were reported, demonstrating that it is
a robust and synthetically useful catalyst with a great potential for highly stereo-
selective aldol additions towards a large variety of aldehydes (Scheme 21.51)
[232–235]. A straightforward one-pot, two-step synthesis of D-fagomine, a naturally
occurring imino sugar that inhibits glycosidases, and N-alkylated derivatives 202 has
been realized by using FSA-catalyzed addition of 89 to N-Cbz-protected aminopro-
panal 200, followed by reductive cyclization and ensuing N-alkylation [240]. Fur-
thermore, it has recently been discovered that FSA can accept glycolaldehyde (127) as
an alternative donor substrate (Scheme 21.52) for enzymatic homo-aldol and cross-
O O FSA OH O O
+ HO OH
HO HO
H H H
OH OH HO
127 127
D-threose
O O FSA O
+
H H H HO H
OH OH
127
L-glyceraldehyde L-65
O O FSA OH O O
+ HO OH
H H H
OH OH OH OH HO
L-65 127
5-deoxy-L-xylose
Scheme 21.52 Stereoselective FSA-catalyzed self- and crossed-aldol reactions using

glycolaldehyde (127) as a novel donor substrate.
aldol addition reactions of 127 to other aldehydes. This activity offers the unprec-
edented opportunity of biocatalytic strategies for the immediate and stereoselective
synthesis of aldoses (instead of ketoses derived from typical ketone donors) and
related complex analogues or derivatives, such as L-65, D-threose, or 5-deoxy-L-
xylose [241].
21.1.5
Glycine-Utilizing Aldolases
Glycine-dependent aldolases catalyze the reversible formation of the b-hydroxylated

a-amino acids serine as well as D- and L-threonine (203 and 204) [8, 242]. Metabolism
of the latter involves pyridoxal phosphate dependent enzymes, classified as serine
hydroxymethyltransferase (SHMT; EC 2.1.2.1) or threonine aldolases (ThrAs; L-
threonine selective ¼ EC 4.1.2.5, L-allo-threonine selective ¼ EC 4.1.2.6). Both L- and
D-specific ThrA enzymes are known. All enzymes catalyze reversible aldol-type
cleavage reactions yielding glycine (4) and the corresponding aldehyde
(Scheme 21.53) [242]. For biocatalytic applications, the known enzymes show broad
substrate tolerance for various acceptor aldehydes, notably including aromatic
aldehydes [243–247]; however, a,b-unsaturated aldehydes are not accepted [248].
Because the primary aldol products from ThrA catalysis usually cannot engage in
subsequent constitutional isomerization – such as cyclization to more stable ring
forms typical for sugar-type products, which removes the product from the aldol-
ization equilibrium – reactions generally suffer from an unfavorable equilibrium
constant [25]. Thus, for preparative reactions usually an excess of 4 is applied to
achieve economical yields. As an example, the synthesis of L-serine from formalde-
hyde has been developed at multi-molar scale using an SHMT from Klebsiella
902
R H O
OH
NH2 4
OH O OH O
L-ThrA D-ThrA
R OH 3S,4R 3R,4S R OH
H2N H2N
L-threonine (203) D-threonine (204)
OH O OH O
L-allo-ThrA D-allo-ThrA
R OH 3S,4S 3R,4R R OH
H2N H2N
L-allo-threonine D-allo-threonine
Scheme 21.53 Stereo-complementary aldol additions catalyzed by the four subtypes of glycine
dependent aldolases.
aerogenes or E. coli to furnish the product at a high final concentration of >450 g

l1 [249, 250].
ThrA biocatalysts are powerful tools for enzyme-catalyzed syntheses due to the
introduction of both multifunctionality and chirality in a single step. Since two new
stereogenic centers are formed, four different products with complementary ste-
reochemistry can be obtained formally from addition of 4 to a single aldehyde using
either L-ThrA and D-ThrA or the corresponding allo-threonine selective aldolases,
analogous to the situation with the DHAP-dependent aldolases (Scheme 21.27).
Typically, SHMT and ThrA enzymes show complete enantiopreference for their
natural a-L-amino (or a-D-amino) acid configuration but, with few exceptions, have
only limited kinetic selectivity for the relative threo/erythro-configuration (e.g., 205/
206) [243, 244, 251]. It must be pointed out that the stereochemical outcome of
individual aldol additions not only depends on the kinetic selectivity of the individual
enzyme but also on thermodynamic relations due to the reversible nature of the aldol
reactions, particularly at advanced levels of conversion (Scheme 21.54). For example,
the addition of 4 to benzaldehyde, when catalyzed by different types of ThrA
enzymes, furnished a mixture of phenylserine diastereomers (e.g., L-205/L-206) with
high enantiospecificity for the L-configuration but a thermodynamically controlled
d.e. of only about 20% in favor of the syn-product at longer reaction times for all tested
enzymes when targeting higher yields [251].
b-Hydroxy-a-amino acids constitute an important class of compounds, either by
themselves as metabolic intermediates or as valuable chiral building blocks for the
synthesis of various bioactive drugs. Long-chain aliphatic aldehydes 207 were
converted into higher homologs of L- and D-threonine as intermediates towards
aminodiol analogs 208 of sphingosine [252]. Interestingly, for reactions catalyzed by
OH
CO2H
L-ThrA Ph
fast NH2
kinetic
O CO2H L-threo-205 product
+ thermodynamic
Ph H mixture
NH2 OH
slow CO2H
4 L-ThrA Ph
NH2
L-erythro-206
Scheme 21.54 Superposition of enzyme-specific kinetic stereoselectivity and thermodynamic

equilibration for glycine-dependent aldolases, causing a time-dependent challenge for
diastereoselective product formation.
L-ThrA a switch from syn- to anti-stereoselectivity was observed when aldehydes

larger than hexanal were employed. Better syn-selectivity resulted with both L-ThrA
and D-ThrA for the sterically more demanding isobutyraldehyde as well as for more
electrophilic halogenated acetaldehyde derivatives.
The erythro-selective L-ThrA from the yeast Candida humicola has been used for the
synthesis of (S,S,R)- and (S,S,S)-3,4-dihydroxyprolines 212/213 from acetonide-pro-
tected glyceraldehyde D-78 (Scheme 21.55) [253]. The same enzyme has also been used
for the preparationof a chiral building block 215 from 4-benzyloxybutanal(180) towards
the synthesis of the immunosuppressive lipids mycestericin D/F [254, 255].
The product 214 [248] derived from O-benzylglycolaldehyde (211) under kinetic
control was applied to the stereoselective synthesis of novel sialyl LewisX mimics [256].
Glyoxylic acid (216) and succinic semialdehyde were converted using L-ThrA
from E. coli for the synthesis of L-b-hydroxy-aspartates 217/218 and L-b-hydroxy-
a-aminoadipic acids (e.g., 220), respectively (Scheme 21.56) [257]. Because the
aldolase was lacking b-selectivity, product isolation from the resulting mixtures
proved difficult. Starting from the methyl ester of succinic semialdehyde 219, but
employing SHMT instead, pure L-erythro-diastereomer 220 has been prepared as a
potential precursor for carbocyclic b-lactams and nucleotides [258]. b-Hydroxy-
a,v-diamino acids are interesting as intermediates for the synthesis of statine
derivatives, protease inhibitors, antivirals, and peptide mimetics, among other
bioactive compounds. Upon lowering the reaction temperature, with N-Cbz pro-
tected aminoaldehydes (221) a novel SHMT from Streptococcus thermophilus and
L-ThrA from E coli turned out to be stereo-complementary biocatalysts useful for the
concise synthesis of L-anti- and L-syn-b-hydroxy-a,v-diamino acid derivatives (e.g.,
222), respectively, from which stable 2-oxazolidinones 223/224 could be easily
obtained (Scheme 21.56) [259].
The Candida L-ThrA enzyme has been utilized to prepare nucleobase-modified
allo-threonine derivatives of type 228 (both adenine and guanine) for the assembly to
peptide-based RNA analogs (Scheme 21.57) [260]. Using the same technology,
peptide-based mimetics of CMP-sialic acid 229 [261, 262] and GDP-fucose
230 [263] were prepared as potential inhibitors of corresponding glycosyltrans-
ferases. The protected chiral aldehyde 225 derived from D-ribose was converted
O OH 1) SOCl2 OH OH
D-ThrA 2) KBH4
CO2H
Alkyl Alkyl Alkyl OH R1 OH
glycine
207 NH2 208 NH2 NH2
syn
sphingosine
OH OH
CHO ThrA CO2H CO2H
O O O
O glycine O NH2
+ O NH2
D-78 209 210

steps steps
HO OH HO OH
O
BnO N CO2H N CO2H
H
H H
211 212 213
L-ThrA
glycine
OH O OH
L-ThrA
BnO CO2H BnO BnO CO2H
H
glycine
214 NH2 180 215 NH2
OH
C6H13 CO2H
CH2OH
O NH2
mycestericin D
Scheme 21.55 Application of ThrA catalysis to the preparation of sphingosine mimetics (208), for
the stereoselective synthesis of dihydroxyprolines (212,213), and for precursors to sialyl LewisX
mimetics and the immunosuppressive lipid mycestericin (214,215).
using the yeast L-ThrA to give the erythro-configurated amino acid building block 226
in acceptable yield, which could be further elaborated to complete a formal synthesis
of antifungal thymine polyoxin C (also referred to as deoxypolyoxin C) [264] as well
as imino-analogs of (deoxy)digitoxose [265].
Phenylserine derivative 232, precursor to the opposite enantiomer of the antibiotic
thiamphenicol, has been prepared with 92% d.e. and >99% e.e. using a recombinant
D-ThrA from Alcaligenes xylosoxidans, whereas the correctly L-configurated
isomer was obtained by L-ThrA catalysis with low diastereoselectivity only
(Scheme 21.58) [247]. Owing to the fully reversible equilibrium nature of the aldol
addition process, enzymes with low diastereoselectivity will typically lead to a
thermodynamically controlled mixture of erythro/threo-isomers that are difficult to
separate. One way to avoid limitations arising from low selectivity is by assembling a
cascade reaction in which one stereoisomer is further transformed selectively in situ
by the action of another enzyme. The thermodynamic origin of poor threo/erythro
O OH OH
L-ThrA
CO2H + CO2H
HO2C HO2C HO2C
glycine
NH2 NH2
216 217 218
O OH
SHMT CO2H
MeO2C HO2C
glycine O
NH2
219 220 O
HN
CO2H
O OH KOH
SHMT
CbzHN CbzHN CO2H 223 NH2
glycine O
NH2 COCl2
221 222 O
NH
CbzHN
224 CO2H
Scheme 21.56 Synthetic applications of ThrA, and of SHMT, catalysis to the preparation of
hydroxyamino diacids or hydroxydiamino acids.
HO NH
(+)-imino-deoxy-
digitoxose
O HO2C OH OH
L-ThrA
H O
steps
glycine H2N
O O O O CO2H NH
225 226 N
H2N O
O
NH2 NH2 HO OH deoxypolyoxin C
N N N N
N N N N NH2
H L-ThrA
227 228 CO2H
glycine
O OH
NH2 O
N N NH
OH
HO CO2H
O CO2H N O O CO2H N N NH2
O O O O
AcHN N OH N
HO OH H H
229 OH OH 230 OH
HO
Scheme 21.57 ThrA-catalyzed synthesis of a precursor to deoxypolyoxin C and an iminocyclitol;

nucleobase modified amino acids for the synthesis of RNA mimics or inhibitors of glycosyl
transferases (CMP-sialic acid and GDP-fucose analogs).
906
O OH OH
D-ThrA
CO2H
H glycine OH
NH2 HN CHCl2
H3CO2S H3CO2S H3CO2S
231 232 thiamphenicol O
OH OH
CO2H
L-TyrD
L-ThrA
NH2 fast NH2
CHO CO2 (R)-233
L-threo-205
+4 fast
OH OH
CO2H
L-TyrD
L-ThrA
NH2 slow NH2
L-erythro-206 CO2 (S)-233
CO2 OH
L-TyrD
L-threo-205
NH2
racemate +
D-ThrA L-ThrA (R)-233
D-threo-205 L-threo-205
Scheme 21.58 ThrA-based preparation of irreversible enantioselective

the antibiotic thiamphenicol; dynamic kinetic decarboxylation, for either arresting the
asymmetric transformations (DYKATs) kinetic selectivity of aldol synthesis or
by coupling of the ThrA equilibrium to racemate resolution.
selectivity of ThrA enzymes has most recently been turned to an asset by the design of
a diastereoselective dynamic kinetic resolution process by coupling of L-ThrA and a
diastereoselective L-tyrosine decarboxylase (L-TyrD) (Scheme 21.58) [266]. By this
concept, a reversible enzymatic aldol reaction generates a mixture of L-threo/erythro
aldol diastereomers 205/206 from which the L-threo isomer 205 is preferentially
decomposed by an irreversible decarboxylation to furnish aromatic aminoalcohol
(R)-233 with 78% e.e. in high isolated yield. In a closely related strategy, a three-
enzyme combination of L-ThrA, D-ThrA, and L-TyrD was used to effect a dynamic
kinetic asymmetric transformation (DYKAT, see Reference [267]) via stereo-random-
ization/resolution of chemically produced racemic syn-phenylserine DL-205 to pro-
duce (R)-233 in 58% isolated yield and >99% e.e. [266, 268].
In practice, the kinetic specificity of ThrA enzymes can generally be exploited in a
straightforward manner for kinetic resolutions of stereoisomer mixtures such as
those produced by chemical synthesis (Scheme 21.59). This is particularly promising
for aryl analogs of threonine that are of interest as building blocks of pharmaceuticals,
including vancomycin antibiotics. Thus, an L-ThrA from Streptomyces amakusaensis
has been shown to be particularly useful for the resolution of racemic threo-aryl
OH OH
CHO
CO2H R CO2H
L-ThrA S + + 4
NH2 NH2 X
X X
DL-threo-234 D-threo-234
OH OH OH
O CO2H D-ThrA O CO2H HO CO2H
O NH2 O NH2 NH2

HO
DL-threo-235 piperonal L-threo-235 DOPS
+
120
Scheme 21.59 Kinetic resolution of diastereomer mixtures by retro-aldolization for the preparation
of enantiopure arylserines and for a synthetic precursor to DOPS, an anti-parkinsonism drug.
serines 234 by retro-aldolization under kinetic control to furnish enantiomerically

pure D-threo-amino acids [243, 258, 269, 270]. As a complementary example, the
recombinant low-specificity D-ThrA from Alcaligenes xyloxidans has been used for the
resolution of DL-threo-b-(3,4-methylenedioxyphenyl)serine (DL-threo-235) by retro-
aldol cleavage to furnish the desired L-threo isomer with a molar yield of 50% and
almost 100% e.e. [271]. The latter compound serves as a synthetic intermediate
en route to L-threo-3,4-dihydroxyphenylserine (DOPS), a drug used for treatment of
Parkinsons disease. An immediate access to DOPS at industrial scale has been
recently reported by employing L-ThrA from Streptomyces avermitilis MA-4680 in a
whole-cell, high-density bioreactor to achieve productivities of around 8 g l1 [272].
Taking advantage of the common evolutionary origin and structurally and mech-
anistically close relationship of PLP-dependent alanine racemase and threonine
aldolase [273], a novel aldolase has been created from the alanine racemase from
Geobacillus stearothermophilus by a single active site mutation. The Tyr265Ala point
mutant was found to exhibit a kcat/Km for the retro-aldol reaction with b-phenylserine
>105-fold higher than the wild-type alanine racemase [274]. Various aromatic
aldehydes tested positive as acceptors, affording D-amino acids exclusively (>99%
e.e.), in moderate to high b-selectivity in favor of the syn-diastereomer (40–97%
d.e.) [275]. This engineered ThrA has been shown recently to also cleave a-alkyl
substituted b-phenylserines, which are not acceptable to natural ThrA enzymes, with
high D-stereoselectivity if only low b-preference [276]. Interestingly, both the (3R)- and
(3S)-diastereomers of a-methyl-b-phenylserine were accepted with even greater
efficiency than b-phenylserine [277]. In the direction of synthesis, this catalyst is
able to use D-alanine as an alternative aldol donor with an electron-deficient
acceptor (236), thereby generating an a-branched D-amino acid (237/238)
(Scheme 21.60) [275]. In common with an improved Tyr265Lys mutant, both variants
show a 8 : 1 preference for the D-syn-stereoisomer of b-phenylserine, which renders
them more diastereoselective than many natural PLP-dependent aldolases. Similar to
the engineered ThrA enzyme variants, a PLP-dependent a-methylserine aldolase
was recently discovered among different microorganisms [278, 279]. The enzyme
908
alr-ThrA*
O (Y265A) OH OH
O2 N O2N CO2H O2N CO2H
H
D-alanine +
H3C NH2 H3C NH2
236 syn 237 anti 238
α-methylserine
O H CO2H aldolase CO2H
+ HO
H H R NH2 R NH2
239 a R=CH3 240 a,b
b R=C2H5
Scheme 21.60 Synthesis of a-alkylated phenylserine derivatives and an a-alkylated serine by using
an engineered alanine racemase variant and a novel methylserine aldolase, respectively.
catalyzes the addition of L-alanine (239a) and L-2-aminobutyric acid (239b) to

formaldehyde as an acceptor, which results in the rare case of a branched carbon
chain generated by a natural aldolase. The enzyme from Ralstonia sp. was applied
in a whole-cell catalyzed stereospecific synthesis of a-methyl-L-serine 240a on a
30-mmol scale.
21.1.6
Development of Novel Catalysts
Natural aldolase enzymes are somewhat limited in that they are usually highly
selective towards their donor substrate and, in some instances, in that the product
stereoselectivity does not always fulfill high expectations, especially when dealing
with non-natural acceptors that are structurally very distant to the natural substrate.
Some of the limitations may be overcome by most recent technological develop-
ments. Advanced knowledge of the catalytic function of proteins, as well as advanced
strategies towards the in vitro evolution of a wild-type biocatalyst increasingly foster
more effective, accelerated pathways to improve an enzymes substrate tolerance,
stereoselectivity, and other functional properties to broaden its window of applica-
bility [280]. Rational re-design of the reaction mechanism of a promiscuous enzyme
is another opportunity [281], as demonstrated by the aldolase reactivity engineered
into a lipase mutant variant [282].
Apart from the discovery, development, and engineering or evolution of novel
naturally occurring aldolases, many efforts have been devoted to design biologically
compatible artificial catalysts that mimic the effectiveness and selectivity of natural
aldolases, while also potentially widening their synthetic scope. The underlying
concepts and classes of catalysts cover a broad range, from organocatalysis using
small molecules as simple as L-proline that act via covalent enamine intermediates
similar to class I aldolases [283] through RNA-based ribozymes capable of catalyzing
the aldol reaction in the presence of Zn2 þ ions and therefore plausibly stabilizing an
enediolate nucleophile in a manner similar to class II aldolases [284], up to huge
References j909
oligomeric protein complexes such as catalytic antibodies, which have significantly
extended the scope of aldol reactions that can be catalyzed chemically or enzymat-
ically [285]. More recent developments include studies on medium-sized designer
oligopeptide amides (consisting of 24–35 amino acids) [286], the combinatorial
construction of peptide dendrimers [287], or designed 14-helical b-peptide folda-
mers [288], all of which are experimental systems that were found to catalyze a retro-
aldol reaction with a rate acceleration of several orders of magnitude despite the
absence of a well-structured active site.
Notwithstanding the progress in developing artificial non-protein catalysts,
advances in the understanding of protein function and in computational chemistry
have most recently opened up new possibilities for designing aldolase biocatalysts for
chemical reactions from scratch. Very recently, from the amalgamation of computed
transition state structures with appropriate side chain orientations from binding
pockets of known protein scaffolds, the de novo computational construction of a
large array of potential active sites for a specific retro-aldol reaction was reported, for
which many active ensembles could be experimentally determined to catalyze
detectable rate accelerations of up to four orders of magnitude [289]. Although there
is still a gap to natural enzymes that have been optimized by natural evolution, it is
reassuring to note that our understanding of enzyme catalysis, even in case of highly
complex multi-substrate reactions such as the stereoselective aldol addition, has
advanced to a level where theory and modern computer algorithms, paired with
clever screening technologies, may finally allow us to approach the holy grail of
creative protein design for tailor-made enzyme catalysts, ultimately to predictably suit
any desired function.
References
1 Mahrwald, R. (ed.) (2004) Modern Aldol 8 Fessner, W.-D. (2004) Enzyme-catalyzed

Reactions, Wiley-VCH Verlag GmbH, Aldol Additions in Modern Aldol Reactions,
Weinheim. vol. 1 (ed. R. Mahrwald) Wiley-VCH
2 Fessner, W.-D. and Walter, C. (1996) Top. Verlag GmbH, Weinheim, pp. 201–272.
Curr. Chem., 184, 97–194. 9 Sukumaran, J. and Hanefeld, U. (2005)
3 Fessner, W.-D. (1998) Curr. Opin. Chem. Chem. Soc. Rev., 34, 530–542.
Biol., 2, 85–97. 10 Samland, A.K. and Sprenger, G.A. (2006)
4 Machajewski, T.D. and Wong, C.-H. Appl. Microbiol. Biotechnol., 71, 253–264.
(2000) Angew. Chem. Int. Ed., 39, 11 Fessner, W.-D. and Jennewein, S. (2007)
1352–1374. Biotechnological Applications of
5 Fessner, W.-D. (2000) Enzymatic Aldolases in Biocatalysis in the
Asymmetric Synthesis Using Aldolases Pharmaceutical and Biotechnology
in Stereoselective Biocatalysis (ed. R.N. Industries (ed. R.N. Patel) CRC Press,
Patel) Marcel Dekker, New York, pp. Boca Raton, pp. 363–400.
239–265. 12 Fessner, W.-D. (2007) Biocatalytic C–C
6 Fessner, W.-D. and Helaine, V. (2001) Bond Formation in Asymmetric
Curr. Opin. Biotechnol., 12, 574–586. Synthesis in Asymmetric Synthesis with
7 Silvestri, M.G., Desantis, G., Chemical and Biological Methods (eds D.
Mitchell, M., and Wong, C.-H. (2003) Enders and K.-E. Jaeger), Wiley-VCH
Top. Stereochem., 23, 267–342. Verlag GmbH, Weinheim, pp. 351–375.
910
13 Fessner, W.-D. (2008) Aldolases: 30 Barbas, C.F., Wang, Y.F., and Wong, C.-H.
Enzymes for Making and Breaking C–C (1990) J. Am. Chem. Soc., 112, 2013–2014.
Bonds in Asymmetric Organic Synthesis 31 Tischer, W., Ihlenfeldt, H.-G., Barzu, O.,
with Enzymes (eds V. Gotor, I. Alfonso, Sakamoto, H., Pistotnik, E., Marliere, P.,
and E. Garcia-Urdiales), Wiley-VCH and Pochet, S., WO 0114566;Chem. Abstr.
Verlag GmbH, Weinheim, pp. 275–318. 134 (2001) 208062.
14 Fessner, W.-D. (1992) A building block 32 Ouwerkerk, N., Van Boom, J.H.,
strategy for asymmetric synthesis: the Lugtenburg, J., and Raap, J. (2000)
DHAP-Aldolases in Microbial Reagents in Eur. J. Org. Chem., 861–866.
Organic Synthesis, vol. 381 33 Liu, J. and Wong, C.-H. (2002) Angew.
(ed. S. Servi), Kluwer Academic Chem. Int. Ed., 41, 1404–1407.
Publishers, Dordrecht, pp. 43–55. 34 Gijsen, H.J.M. and Wong, C.-H. (1995)
15 Schomburg, D. and Salzmann, M. J. Am. Chem. Soc., 117, 7585–7591.
(1990–1995) Enzyme Handbook, 35 Wong, C.-H., Garcia-Junceda, E.,
Springer, Berlin. Chen, L.R., Blanco, O., Gijsen, H.J.M.,
16 Henderson, D.F. and Toone, E.J. (1999) and Steensma, D.H. (1995) J. Am. Chem.
Aldolases in Comprehensive Natural Soc., 117, 3333–3339.
Product Chemistry, vol. 3 (ed. B.M. Pinto) 36 Liu, J., Hsu, C.-C., and Wong, C.-H.
Elsevier Science, Amsterdam, pp. (2004) Tetrahedron Lett., 45, 2439–2441.
367–440. 37 Greenberg, W.A., Varvak, A.,
17 Toone, E.J., Simon, E.S., Bednarski, Hanson, S.R., Wong, K., Huang, H.,
M.D., and Whitesides, G.M. (1989) Chen, P., and Burk, M.J. (2004) Proc. Natl.
Tetrahedron, 45, 5365–5422. Acad. Sci. USA, 101, 5788–5793.
18 David, S., Auge, C., and Gautheron, C. 38 Jennewein, S., Sch€ urmann, M.,
(1991) Adv. Carbohydr. Chem. Biochem., Wolberg, M., Hilker, I., Luiten, R.,
49, 175–237. Wubbolts, M., and Mink, D. (2006)
19 Gijsen, H.J.M., Qiao, L., Fitz, W., and Biotechnol. J., 1, 537–548.
Wong, C.-H. (1996) Chem. Rev., 96, 39 Wolberg, M., Dassen, B.H.N.,
443–473. Sch€urmann, M., Jennewein, S.,
20 Wymer, N. and Toone, E.J. (2000) Curr. Wubbolts, M.G., Schoemaker, H.E., and
Opin. Chem. Biol., 4, 110–119. Mink, D. (2008) Adv. Synth. Catal., 350,
21 Horecker, B.L., Tsolas, O., and 1751–1759.
Lai, C.Y. (1972) Aldolases in The Enzymes, 40 Sakuraba, H., Yoneda, K., Yoshihara, K.,
vol. VII (ed. P.D. Boyer), Academic Press, Satoh, K., Kawakami, R., Uto, Y.,
New York, pp. 213–258. Tsuge, H., Takahashi, K., Hori, H., and
22 Kluger, R. (1990) Chem. Rev., 90, 1151–1169. Ohshima, T. (2007) Appl. Environ.
23 Kuo, D.J. and Rose, I.A. (1985) Microbiol., 73, 7427–7434.
Biochemistry, 24, 3947–3952. 41 Fessner, W.-D. and Knorst, M., DE
24 Fessner, W.-D., Schneider, A., Held, H., 10034586;Chem. Abstr. 136 (2002)
Sinerius, G., Walter, C., Hixon, M., and 166160.
Schloss, J.V. (1996) Angew. Chem. Int. Ed. 42 Kim, M.J., Hennen, W.J., Sweers, H.M.,
Engl., 35, 2219–2221. and Wong, C.-H. (1988) J. Am. Chem.
25 Goldberg, R.N. and Tewari, Y.B. (1995) Soc., 110, 6481–6486.
J. Phys. Chem. Ref. Data, 24, 1669–1698. 43 Angata, T. and Varki, A. (2002) Chem.
26 Chen, L., Dumas, D.P., and Wong, C.-H. Rev., 102, 439–469.
(1992) J. Am. Chem. Soc., 114, 741–748. 44 Sugai, T., Kuboki, A., Hiramatsu, S.,
27 Gijsen, H.J.M. and Wong, C.-H. (1995) Okazaki, H., and Ohta, H. (1995)
J. Am. Chem. Soc., 117, 2947–2948. Bull. Chem. Soc. Jpn., 68, 3581–3589.
28 Liu, J., DeSantis, G., and Wong, C.-H. 45 Blayer, S., Woodley, J.M., Lilly, M.D., and
(2002) Can. J. Chem., 80, 643–645. Dawson, M.J. (1996) Biotechnol. Prog., 12,
29 DeSantis, G., Liu, J., Clark, D.P., 758–763.
Heine, A., Wilson, I.A., and Wong, C.-H. 46 Mahmoudian, M., Noble, D., Drake, C.S.,
(2003) Bioorg. Med. Chem., 11, 43–52. Middleton, R.F., Montgomery, D.S.,
References j911
Piercey, J.E., Ramlakhan, D., Todd, M., 64 Fitz, W., Schwark, J.R., and Wong, C.-H.
and Dawson, M.J. (1997) Enzyme Microb. (1995) J. Org. Chem., 60, 3663–3670.
Technol., 20, 393–400. 65 Aug
e, C., Bouxom, B., Cavaye, B., and
47 Blayer, S., Woodley, J.M., Dawson, M.J., Gautheron, C. (1989) Tetrahedron Lett.,
and Lilly, M.D. (1999) Biotechnol. Bioeng., 30, 2217–2220.
66, 131–136. 66 Liu, J.L.C., Shen, G.J., Ichikawa, Y.,
48 Kragl, U., Gygax, D., Ghisalba, O., and Rutan, J.F., Zapata, G., Vann, W.F., and
Wandrey, C. (1991) Angew. Chem. Int. Ed. Wong, C.-H. (1992) J. Am. Chem. Soc.,
Engl., 30, 827–828. 114, 3901–3910.
49 Maru, I., Ohnishi, J., Ohta, Y., and Tsukada, 67 Lins, R.J., Flitsch, S.L., Turner, N.J.,
Y. (1998) Carbohydr. Res., 306, 575–578. Irving, E., and Brown, S.A. (2002) Angew.
50 Lee, J.-O., Yi, J.-K., Lee, S.-G., Chem. Int. Ed., 41, 3405–3407.
Takahashi, S., and Kim, B.G. (2004) 68 Lins, R.J., Flitsch, S.L., Turner, N.J.,
Enzyme Microb. Technol., 35, 121–125. Irving, E., and Brown, S.A. (2004)
51 Lin, C.H., Sugai, T., Halcomb, R.L., Tetrahedron, 60, 771–780.
Ichikawa, Y., and Wong, C.-H. (1992) 69 Wu, W.-Y., Jin, B., Kong, D.C.M., and
J. Am. Chem. Soc., 114, 10138–10145. von Itzstein, M. (1997) Carbohydr. Res.,
52 Xu, P., Qiu, J.H., Zhang, Y.N., Chen, J., 300, 171–174.
Wang, P.G., Yan, B., Song, J., Xi, R.M., 70 Kuboki, A., Okazaki, H., Sugai, T., and
Deng, Z.X., and Ma, C.Q. (2007) Ohta, H. (1997) Tetrahedron, 53,
Adv. Synth. Catal., 349, 1614–1618. 2387–2400.
53 Tabata, K., Koizumi, S., Endo, T., and 71 Humphrey, A.J., Fremann, C.,
Ozaki, A. (2002) Enzyme Microb. Technol., Critchley, P., Malykh, Y., Schauer, R., and
30, 327–333. Bugg, T.D.H. (2002) Bioorg. Med. Chem.,
54 Salagnad, C., Godde, A., Ernst, B., and 10, 3175–3185.
Kragl, U. (1997) Biotechnol. Prog., 13, 72 Pan, Y., Ayani, T., Nadas, J., Wen, S., and
810–813. Guo, Z. (2004) Carbohydr. Res., 339,
55 Kragl, U., G€odde, A., Wandrey, C., 2091–2100.
Lubin, N., and Auge, C. (1994) J. Chem. 73 Lin, C.-C., Lin, C.-H., and Wong, C.-H.
Soc., Perkin Trans. 1, 119–124. (1997) Tetrahedron Lett., 38, 2649–2652.
56 Hsu, C.-C., Hong, Z., Wada, M., 74 Murakami, M., Ikeda, K., and Achiwa, K.
Franke, D., and Wong, C.-H. (2005) Proc. (1996) Carbohydr. Res., 280, 101–110.
Natl. Acad. Sci. USA, 102, 9122–9126. 75 Kong, D.C.M. and von Itzstein, M. (1997)
57 Indurugalla, D. and Bennet, A.J. (2008) Carbohydr. Res., 305, 323–329.
Can. J. Chem., 86, 1005–1009. 76 Kiefel, M.J., Wilson, J.C., Bennett, S.,
58 Miyazaki, T., Sato, H., Sakakibara, T., and Gredley, M., and von Itzstein, M. (2000)
Kajihara, Y. (2000) J. Am. Chem. Soc., 122, Bioorg. Med. Chem., 8, 657–664.
5678–5694. 77 Sparks, M.A., Williams, K.W., Lukacs, C.,
59 Sato, K., Akai, S., Hiroshima, T., Aoki, H., Schrell, A., Priebe, G., Spaltenstein, A.,
Sakuma, M., and Suzuki, K. (2003) and Whitesides, G.M. (1993) Tetrahedron,
Tetrahedron Lett., 44, 3513–3516. 49, 1–12.
60 Beliczey, J., Kragl, U., Liese, A., 78 Zhou, P.Z., Salleh, H.M., and Honek, J.F.
Wandrey, C., Hamacher, K., (1993) J. Org. Chem., 58, 264–266.
Coenen, H.H., and Tierling, T., US 79 Fitz, W. and Wong, C.-H. (1994)
6355453; Chem. Abstr. 136 (2002) 231340. J. Org. Chem., 59, 8279–8280.
61 Watts, A.G. and Withers, S.G. (2004) Can. 80 Ikeda, K., Mori, H., and Sato, M. (2006)
J. Chem., 82, 1581–1588. Chem. Commun., 3093–3094.
62 Chokhawala, H.A., Cao, H., Yu, H., and 81 Hartlieb, S., Guenzel, A., Gerardy-
Chen, X. (2007) J. Am. Chem. Soc., 129, Schahn, R., Muenster-Kuehnel, A.K.,
10630–10631. Kirschning, A., and Draeger, G. (2008)
63 Auge, C., David, S., Gautheron, C., Carbohydr. Res., 343, 2075–2082.
Malleron, A., and Cavaye, B. (1988) 82 Li, Y.H., Yu, H., Cao, H.Z., Lau, K.,
New. J. Chem., 12, 733–744. Muthana, S., Tiwari, V.K., Son, B., and
912
Chen, X. (2008) Appl. Microbiol. 103 Yarema, K.J., Mahal, L.K., Bruehl, R.E.,
Biotechnol., 79, 963–970. Rodriguez, E.C., and Bertozzi, C.R.
83 Kok, G.B., Campbell, M., Mackey, B.L., (1998) J. Biol. Chem., 273, 31168–31179.
and von Itzstein, M. (2001) Carbohydr. 104 Laughlin, S.T. and Bertozzi, C.R. (2009)
Res., 332, 133–139. Proc. Natl. Acad. Sci. USA, 106, 12–17.
84 Lin, C.H., Sugai, T., Halcomb, R.L., 105 Ghalambor, M.A. and Heath, E.C. (1966)
Ichikawa, Y., and Wong, C.-H. (1992) J. Biol. Chem., 241, 3222–3227.
J. Am. Chem. Soc., 114, 10138–10145. 106 Wood, W.A. (1972) 2-Keto-3-deoxy-6-
85 David, S., Malleron, A., and Cavaye, B. phosphogluconic and related aldolases in
(1992) New J. Chem, 16, 751–755. The Enzymes, 3rd edn, vol. 7 (ed P.D.
86 Auge, C., Gautheron, C., David, S., Boyer), Academic, New York, pp.
Malleron, A., Cavaye, B., and Bouxom, B. 281–302.
(1990) Tetrahedron, 46, 201–214. 107 Floyd, N.C., Liebster, M.H., and
87 Sugai, T., Shen, G.J., Ichikawa, Y., and Turner, N.J. (1992) J. Chem. Soc., Perkin
Wong, C.-H. (1993) J. Am. Chem. Soc., Trans. 1, 1085–1086.
115, 413–421. 108 Shelton, M.C. and Toone, E.J. (1995)
88 Koppert, K. and Brossmer, R. (1992) Tetrahedron: Asymmetry, 6, 207–211.
Tetrahedron Lett., 33, 8031–8034. 109 Allen, S.T., Heintzelman, G.R., and
89 Malleron, A. and David, S. (1996) Toone, E.J. (1992) J. Org. Chem., 57,
New J. Chem., 20, 153–159. 426–427.
90 Woodhall, T., Williams, G., Berry, A., and 110 Shelton, M.C., Cotterill, I.C.,
Nelson, A. (2005) Angew. Chem. Int. Ed., Novak, S.T.A., Poonawala, R.M.,
44, 2109–2112. Sudarshan, S., and Toone, E.J. (1996)
91 Williams, G.J., Woodhall, T., Nelson, A., J. Am. Chem. Soc., 118, 2117–2125.
and Berry, A. (2005) Prot. Eng. Design 111 Cotterill, I.C., Shelton, M.C.,
Select., 18, 239–246. Machemer, D.E.W., Henderson, D.P.,
92 Williams, G.J., Woodhall, T., Farnsworth, and Toone, E.J. (1998) J. Chem. Soc.,
L.M., Nelson, A., and Berry, A. (2006) Perkin Trans. 1, 1335–1341.
J. Am. Chem. Soc., 128, 16238–16247. 112 Henderson, D.P., Cotterill, I.C.,
93 Bolt, A., Berry, A., and Nelson, A. (2008) Shelton, M.C., and Toone, E.J. (1998)
Arch. Biochem. Biophys., 474, 318–330. J. Org. Chem., 63, 906–907.
94 Dean, S.M., Greenberg, W.A., and 113 Griffiths, J.S., Wymer, N.J., Njolito, E.,
Wong, C.-H. (2007) Adv. Synth. Catal., Niranjanakumari, S., Fierkeb, C.A., and
349, 1308–1320. Toone, E.J. (2002) Bioorg. Med. Chem., 10,
95 Ichikawa, Y., Liu, J.L.C., Shen, G.J., and 545–550.
Wong, C.-H. (1991) J. Am. Chem. Soc., 114 Walters, M.J., Srikannathasan, V.,
113, 6300–6302. McEwan, A.R., Naismith, J.H.,
96 Blixt, O. and Paulson, J.C. (2003) Adv. Fierke, C.A., and Toone, E.J. (2008)
Synth. Catal., 345, 687–690. Bioorg. Med. Chem., 16, 710–720.
97 Chokhawala, H.A., Yu, H., and Chen, X. 115 Fong, S., Machajewski, T.D., Mak, C.C.,
(2007) ChemBioChem, 8, 194–201. and Wong, C.-H. (2000) Chem. Biol., 7,
98 Yu, H., Chokhawala, H.A., Varki, A., and 873–883.
Chen, X. (2007) Org. Biomol. Chem., 5, 116 Wymer, N., Buchanan, L.V., Henderson,
2458–2463. D., Mehta, N., Botting, C.H., Pocivavsek,
99 Yu, H. and Chen, X. (2006) Org. Lett., 8, L., Fierke, C.A., Toone, E.J., and
2393–2396. Naismith, J.H. (2001) Structure, 9, 1–10.
100 Huang, S., Yu, H., and Chen, X. (2007) 117 Griffiths, J.S., Cheriyan, M., Corbell, J.B.,
Angew. Chem. Int. Ed., 46, 2249–2253. Pocivavsek, L., Fierkeb, C.A., and
101 Yu, H., Huang, S., Chokhawala, H., Toone, E.J. (2004) Bioorg. Med. Chem., 12,
Sun, M., Zheng, H., and Chen, X. (2006) 4067–4074.
Angew. Chem. Int. Ed., 45, 3938–3944. 118 Cheriyan, M., Toone, E.J., and
102 Mahal, L.K., Yarema, K.J., and Bertozzi, Fierke, C.A. (2007) Protein Sci., 16,
C.R. (1997) Science, 276, 1125–1128. 2368–2377.
References j913
119 Henderson, D.P., Shelton, M.C., Aguilar, J. (1991) Angew. Chem. Int. Ed.
Cotterill, I.C., and Toone, E.J. (1997) Engl., 30, 555–558.
J. Org. Chem., 62, 7910–7911. 135 Zannetti, M.T., Walter, C., Knorst, M.,
120 Walters, M.J. and Toone, E.J. (2007) Nat. and Fessner, W.-D. (1999) Chem. Eur. J., 5,
Protoc., 2, 1825–1830. 1882–1890.
121 Ran, N., Draths, K.M., and Frost, J.W. 136 Espelt, L., Clapes, P., Esquena, J.,
(2004) J. Am. Chem. Soc., 126, 6856–6857. Manich, A., and Solans, C. (2003)
122 Ran, N. and Frost, J.W. (2007) Langmuir, 19, 1337–1346.
J. Am. Chem. Soc., 129, 6130–6139. 137 Lees, W.J. and Whitesides, G.M. (1993)
123 Lamble, H.J., Heyer, N.I., Bull, S.D., J. Org. Chem., 58, 1887–1894.
Hough, D.W., and Danson, M.J. (2003) 138 Bissett, D.L. and Anderson, R.L. (1980)
J. Biol. Chem., 278, 34066–34072. J. Biol. Chem., 255, 8750–8755.
124 Wolterink-van Loo, S., van Eerde, A., 139 Fessner, W.-D. and Eyrisch, O. (1992)
Siemerink, M.A.J., Akerboom, J., Angew. Chem. Int Ed. Engl., 31, 56–58.
Dijkstra, B.W., and van der Oost, J. (2007) 140 Eyrisch, O., Sinerius, G., and
Biochem. J., 403, 421–430. Fessner, W.-D. (1993) Carbohydr. Res.,
125 Lamble, H.J., Danson, M.J., 238, 287–306.
Hough, D.W., and Bull, S.D. (2005) 141 Williams, G.J., Domann, S., Nelson, A.,
Chem. Commun., 124–126. and Berry, A. (2003) Proc. Natl. Acad. Sci.
126 Lamble, H.J., Royer, S.F., Hough, D.W., USA, 100, 3143–3148.
Danson, M.J., Taylor, G.L., and Bull, S.D. 142 Jimenez, A., Clapes, P., and Crehuet, R.
(2007) Adv. Synth. Catal., 349, 817–821. (2009) Chem. Eur. J., 15, 1422–1428.
127 Serafimov, J.M., Gillingham, D., 143 Vidal, L., Durany, O., Suau, T., Ferrer, P.,
Kuster, S., and Hilvert, D. (2008) Benaiges, M.D., and Caminal, G. (2003)
J. Am. Chem. Soc., 130, 7798–7799. J. Chem. Technol. Biotechnol., 78,
128 Ogawa, J., Yamanaka, H., Mano, J., 1171–1179.
Doi, Y., Horinouchi, N., Kodera, T., 144 Ardao, I., Benaiges, M.D., Caminal, G.,
Nio, N., Smirnov, S.V., Samsonova, N.N., and Alvaro, G. (2006) Enzyme Microb.
Kozlov, Y.I., and Shimizu, S. (2007) Biosci. Technol., 39, 22–27.
Biotechnol. Biochem., 71, 1607–1615. 145 Espelt, L., Parella, T., Bujons, J.,
129 Smirnov, S.V., Samsonova, N.N., Solans, C., Joglar, J., Delgado, A., and
Novikova, A.E., Matrosov, N.G., Clapes, P. (2003) Chem. Eur. J., 9,
Rushkevich, N.Y., Kodera, T., Ogawa, J., 4887–4899.
Yamanaka, H., and Shimizu, S. (2007) 146 Espelt, L., Bujons, J., Parella, T.,
FEMS Microbiol. Lett., 273, 70–77. Calveras, J., Joglar, J., Delgado, A., and
130 Ling, H., Wang, G., Tian, Y., Liu, G., and Clapes, P. (2005) Chem. Eur. J., 11,
Tan, H. (2007) Biochem. Biophys. Res. 1392–1401.
Commun., 361, 196–201. 147 Ozaki, A., Toone, E.J., von der
131 Suau, T., Alvaro, G., Benaiges, M.D., and Osten, C.H., Sinskey, A.J., and
Lopez-Santin, J. (2008) Biochem. Eng. J., Whitesides, G.M. (1990) J. Am. Chem.
41, 95–103. Soc., 112, 4970–4971.
132 Bednarski, M.D., Simon, E.S., 148 Fessner, W.-D., Badia, J., Eyrisch, O.,
Bischofberger, N., Fessner, W.-D., Schneider, A., and Sinerius, G. (1992)
Kim, M.J., Lees, W., Saito, T., Tetrahedron Lett., 33, 5231–5234.
Waldmann, H., and Whitesides, G.M. 149 Fessner, W.-D., Schneider, A., Eyrisch, O.,
(1989) J. Am. Chem. Soc., 111, Sinerius, G., and Badia, J. (1993)
627–635. Tetrahedron: Asymmetry, 4, 1183–1192.
133 von der Osten, C.H., Sinskey, A.J., 150 Bednarski, M.D., Waldmann, H.J., and
Barbas, C.F., Pederson, R.L., Wang, Y.F., Whitesides, G.M. (1986) Tetrahedron Lett.,
and Wong, C.-H. (1989) J. Am. Chem. 27, 5807–5810.
Soc., 111, 3924–3927. 151 Franke, D., Machajewski, T., Hsu, C.-C.,
134 Fessner, W.-D., Sinerius, G., Schneider, and Wong, C.-H. (2003) J. Org. Chem., 68,
A., Dreyer, M., Schulz, G.E., Badia, J., and 6828–6831.
914
152 Kim, M.-J. and Lim, I.T. (1996) Synlett, 172 Duncan, R. and Drueckhammer, D.G.
138–140. (1996) J. Org. Chem., 61, 438–439.
153 Fessner, W.-D., Gosse, C., Jaeschke, G., 173 Wong, C.-H., Alajarin, R.,
and Eyrisch, O. (2000) Eur. J. Org. Chem., Moris-Varas, F., Blanco, O., and
125–132. Garcia-Junceda, E. (1995) J. Org. Chem.,
154 Fessner, W.-D. and Sinerius, G. (1994) 60, 7360–7363.
Angew. Chem. Int. Ed. Engl., 33, 209–212. 174 Fessner, W.-D. (1993) GIT Fachz. Lab., 37,
155 Wong, C.-H. and Whitesides, G.M. (1983) 951–956.
J. Org. Chem., 48, 3199–3205. 175 Durrwachter, J.R., Drueckhammer, D.G.,
156 Fessner, W.-D. and Walter, C. (1992) Nozaki, K., Sweers, H.M., and
Angew. Chem. Int. Ed. Engl., 31, Wong, C.-H. (1986) J. Am. Chem. Soc.,
614–616. 108, 7812–7818.
157 Maliakel, B.P. and Schmid, W. (1993) 176 Arth, H.L. and Fessner, W.-D. (1997)
J. Carbohydr. Chem, 12, 415–424. Carbohydr. Res., 305, 313–321.
158 Azema, L., Bringaud, F., Blonski, C., and 177 Guanti, G., Banfi, L., and Zannetti, M.T.
Perie, J. (2000) Bioorg. Med. Chem., 8, (2000) Tetrahedron Lett., 41,
717–722. 3181–3185.
159 Ziegler, T., Straub, A., and Effenberger, F. 178 Guanti, G., Zannetti, M.T., Banfi, L., and
(1988) Angew. Chem. Int. Ed. Engl., 27, Riva, R. (2001) Adv. Synth. Catal., 343,
716–717. 682–691.
160 Nicotra, F., Panza, L., Russo, G., and 179 Schmid, W. and Whitesides, G.M. (1990)
Verani, A. (1993) Tetrahedron: Asymmetry, J. Am. Chem. Soc., 112, 9670–9671.
4, 1203–1204. 180 Chou, W.-C., Fotsch, C., and Wong, C.-H.
161 Zhu, W. and Li, Z. (2000) J. Chem. Soc., (1995) J. Org. Chem., 60, 2916–2917.
Perkin Trans. 1, 1105–1108. 181 Gijsen, H.J.M. and Wong, C.-H. (1995)
162 Liu, K.K.C. and Wong, C.-H. (1992) Tetrahedron Lett., 36, 7057–7060.
J. Org. Chem., 57, 4789–4791. 182 El Blidi, L., Crestia, D., Gallienne, E.,
163 Turner, N.J. and Whitesides, G.M. (1989) Demuynck, C., Bolte, J., and Lemaire, M.
J. Am. Chem. Soc., 111, 624–627. (2004) Tetrahedron: Asymmetry, 15,
164 Crestia, D., Guerard, C., Bolte, J., and 2951–2954.
Demuynck, C. (2001) J. Mol. Catal. B: 183 El Blidi, L., Ahbala, M., Bolte, J., and
Enzym., 11, 207–212. Lemaire, M. (2006) Tetrahedron:
165 Sevestre, A., Helaine, V., Guyot, G., Asymmetry, 17, 2684–2688.
Martin, C., and Hecquet, L. (2003) 184 Fessner, W.-D. and Sinerius, G. (1994)
Tetrahedron Lett., 44, 827–830. Bioorg. Med. Chem., 2, 639–645.
166 Wong, C.-H., Moris-Varas, F., 185 Effenberger, F., Straub, A., and Null, V.
Hung, S.-C., Marron, T.G., Lin, C.-C., (1992) Liebigs Ann., 1297–1301.
Gong, K.W., and Weitz-Schmidt, G. 186 Chou, W.C., Chen, L.H., Fang, J.M., and
(1997) J. Am. Chem. Soc., 119, 8152–8158. Wong, C.-H. (1994) J. Am. Chem. Soc.,
167 Eyrisch, O., Keller, M., and 116, 6191–6194.
Fessner, W.-D. (1994) Tetrahedron Lett., 187 Charmantray, F., Dellis, P., Helaine, V.,
35, 9013–9016. Samreth, S., and Hecquet, L. (2006) Eur. J.
168 Petersen, M., Zannetti, M.T., and Org. Chem., 5526–5532.
Fessner, W.-D. (1997) Top. Curr. Chem., 188 Look, G.C., Fotsch, C.H., and Wong, C.-H.
186, 87–117. (1993) Acc. Chem. Res., 26, 182–190.
169 Durrwachter, J.R. and Wong, C.-H. (1988) 189 Wong, C.-H., Halcomb, R.L., Ichikawa, Y.,
J. Org. Chem., 53, 4175–4181. and Kajimoto, T. (1995) Angew. Chem. Int.
170 Liu, K.K.C., Pederson, R.L., and Ed. Engl., 34, 412–432.
Wong, C.-H. (1991) J. Chem. Soc., Perkin 190 Whalen, L.J. and Wong, C.-H. (2006)
Trans. 1, 2669–2673. Aldrichim. Acta, 39, 63–71.
171 Borysenko, C.W., Spaltenstein, A., 191 Pederson, R.L., Kim, M.-J., and
Straub, J.A., and Whitesides, G.M. Wong, C.-H. (1988) Tetrahedron Lett., 29,
(1989) J. Am. Chem. Soc., 111, 9275–9276. 4645–4648.
References j915
192 Liu, K.K.C., Kajimoto, T., Chen, L., 211 Shimagaki, M., Muneshima, H.,
Zhong, Z., Ichikawa, Y., and Wong, C.-H. Kubota, M., and Oishi, T. (1993) Chem.
(1991) J. Org. Chem., 56, 6280–6289. Pharm. Bull., 41, 282–286.
193 Zhou, P.Z., Salleh, H.M., Chan, P.C.M., 212 Phung, A.N., Zannetti, M.T., Whited, G.,
Lajoie, G., Honek, J.F., Nambiar, P.T.C., and Fessner, W.-D. (2003) Angew. Chem.
and Ward, O.P. (1993) Carbohydr. Res., Int. Ed., 42, 4821–4824.
239, 155–166. 213 Schumperli, M., Pellaux, R., and
194 Lees, W.J. and Whitesides, G.M. (1992) Panke, S. (2007) Appl. Microbiol.
Bioorg. Chem., 20, 173–179. Biotechnol., 75, 33–45.
195 Straub, A., Effenberger, F., and Fischer, P. 214 Crans, D.C. and Whitesides, G.M. (1985)
(1990) J. Org. Chem., 55, 3926–3932. J. Am. Chem. Soc., 107, 7019–7027.
196 Hung, R.R., Straub, J.A., and 215 Sanchez-Moreno, I., Garcia-Garcia, J.F.,
Whitesides, G.M. (1991) J. Org. Chem., Bastida, A., and Garcia-Junceda, E. (2004)
56, 3849–3855. Chem. Commun., 1634–1635.
197 Moris-Varas, F., Qian, X.-H., and 216 Iturrate, L., Sanchez-Moreno, I.,
Wong, C.-H. (1996) J. Am. Chem. Soc., Doyaguez, E.G., and Garcia-Junceda, E.
118, 7647–7652. (2009) Chem. Commun., 1721–1723.
198 Mitchell, M.L., Lee, L.V., and Wong, C.-H. 217 van Herk, T., Hartog, A.F.,
(2002) Tetrahedron Lett., 43, 5691–5693. Schoemaker, H.E., and Wever, R. (2006)
199 Calveras, J., Bujons, J., Parella, T., J. Org. Chem., 71, 6244–6247.
Crehuet, R., Espelt, L., Joglar, J., and 218 Zimmermann, F.T., Schneider, A.,
Clapes, P. (2006) Tetrahedron, 62, Sch€orken, U., Sprenger, G.A., and
2648–2656. Fessner, W.-D. (1999) Tetrahedron:
200 Calveras, J., Egido-Gabas, M., Gomez, L., Asymmetry, 10, 1643–1646.
Casas, J., Parella, T., Joglar, J., Bujons, J., 219 Hettwer, J., Oldenburg, H., and
and Clapes, P. (2009) Chem. Eur. J., 15, Flaschel, E. (2002) J. Mol. Catal. B:
7310–7328. Enzym., 19–20, 215–222.
201 Schuster, M., He, W.F., and 220 Schoevaart, R., van Rantwijk, F., and
Blechert, S. (2001) Tetrahedron Lett., 42, Sheldon, R.A. (2000) J. Org. Chem., 65,
2289–2291. 6940–6943.
202 Mitchell, M., Qaio, L., and Wong, C.-H. 221 Charmantray, F., Dellis, P., Samreth, S.,
(2001) Adv. Synth. Catal., 343, 596–599. and Hequet, L. (2006) Tetrahedron Lett.,
203 Romero, A. and Wong, C.-H. (2000) 47, 3261–3263.
J. Org. Chem., 65, 8264–8268. 222 Arth, H.L., Sinerius, G., and Fessner, W.-D.
204 Calveras, J., Casas, J., Parella, T., Joglar, J., (1995) Liebigs Ann., 2037–2042.
and Clapes, P. (2007) Adv. Synth. Catal., 223 Suau, T., Alvaro, G., Benaiges, M.D., and
349, 1661–1666. Lopez-Santin, J. (2006) Biotechnol.
205 Eyrisch, O. and Fessner, W.-D. (1995) Bioeng., 93, 48–55.
Angew. Chem. Int. Ed. Engl., 34, 224 Drueckhammer, D.G., Durrwachter, J.R.,
1639–1641. Pederson, R.L., Crans, D.C., Daniels, L.,
206 Schultz, M., Waldmann, H., Kunz, H., and Wong, C.-H. (1989) J. Org. Chem., 54,
and Vogt, W. (1990) Liebigs Ann., 70–77.
1019–1024. 225 Schoevaart, R., van Rantwijk, F., and
207 Myles, D.C., Andrulis, P.J.I., and Sheldon, R.A. (2001) J. Org. Chem., 66,
Whitesides, G.M. (1991) Tetrahedron Lett., 4559–4562.
32, 4835–4838. 226 Sugiyama, M., Hong, Z.Y., Whalen, L.J.,
208 Chenevert, R. and Dasser, M. (2000) Greenberg, W.A., and Wong, C.-H. (2006)
J. Org. Chem., 65, 4529–4531. Adv. Synth. Catal., 348, 2555–2559.
209 Chenevert, R., Lavoie, M., and Dasser, M. 227 Samland, A.K., Rale, M., Sprenger, G.A.,
(1997) Can. J. Chem., 75, 68–73. and Fessner, W.-D. (2011) ChemBioChem,
210 Matsumoto, K., Shimagaki, M., 12, 1454–1474.
Nakata, T., and Oishi, T. (1993) 228 Sch€urmann, M. and Sprenger, G.A.
Tetrahedron Lett., 34, 4935–4938. (2001) J. Biol. Chem., 276, 11055–11061.
916
229 Gonzalez-Garcia, E., Helaine, V., 245 Lotz, B.T., Gasparski, C.M., Peterson, K.,
Klein, G., Schuermann, M., and Miller, M.J. (1990) J. Chem. Soc.,
Sprenger, G.A., Fessner, W.-D., and Chem. Commun., 16, 1107–1109.
Reymond, J.-L. (2003) Chem. Eur. J., 9, 246 Bycroft, M., Herbert, R.B., and
893–899. Ellames, G.J. (1996) J. Chem. Soc., Perkin
230 Schneider, S., Sandalova, T., Trans. 1, 2439–2442.
Schneider, G., Sprenger, G.A., and 247 Steinreiber, J., Fesko, K., Reisinger, C.,
Samland, A.K. (2008) J. Biol. Chem., 283, Sch€urmann, M., van Assema, F.,
30064–30072. Wolberg, M., Mink, D., and Griengl, H.
231 Rale, M., Schneider, S., Sprenger, G.A., (2007) Tetrahedron, 63, 918–926.
Samland, A.K., and Fessner, W.-D. (2011) 248 Vassilev, V.P., Uchiyama, T., Kajimoto, T.,
Chem. Eur. J., 17, 2623–2632. and Wong, C.-H. (1995) Tetrahedron Lett.,
232 Sch€urmann, M., Sch€ urmann, M., and 36, 4081–4084.
Sprenger, G.A. (2002) J. Mol. Catal. B: 249 Ura, D., Hashimukai, T., Matsumoto, T.,
Enzym., 19–20, 247–252. and Fukuhara, N. (1991) EP 421477;
233 Castillo, J.A., Calveras, J., Casas, J., Chem. Abstr. 115 (1991) 90657.
Mitjans, M., Vinardell, M.P., Parella, T., 250 Anderson, D.M. and Hsiao, H.-H. (1992)
Inoue, T., Sprenger, G.A., Joglar, J., and Enzymatic Method for L-Serine
Clapes, P. (2006) Org. Lett., 8, 6067–6070. Production in Biocatalytic Production of
234 Sugiyama, M., Hong, Z., Liang, P.H., Amino Acids & Derivatives (eds D. Rozzell
Dean, S.M., Whalen, L.J., and F. Wagner), Hanser, Munich, pp.
Greenberg, W.A., and Wong, C.-H. (2007) 23–41.
J. Am. Chem. Soc., 129, 14811–14817. 251 Fesko, K., Reisinger, C., Steinreiber, J.,
235 Concia, A.L., Lozano, C., Castillo, J.A., Weber, H., Sch€ urmann, M., and
Parella, T., Joglar, J., and Clapes, P. (2009) Griengl, H. (2008) J. Mol. Catal. B:
Chem. Eur. J., 15, 3808–3816. Enzym., 52–3, 19–26.
236 Sch€urmann, M., Mink, D., and Hyett, 252 Steinreiber, J., Fesko, K., Mayer, C.,
D.J., WO 067997;Chem. Abstr. 149 (2008) Reisinger, C., Sch€ urmann, M., and
52069. Griengl, H. (2007) Tetrahedron, 63,
237 Grazi, E., Mangiarotti, M., and 8088–8093.
Pontremoli, S. (1962) Biochemistry, 1, 253 Fujii, M., Miura, T., Kajimoto, T., and
628–631. Ida, Y. (2000) Synlett, 1046–1048.
238 Williams, J.F. and Blackmore, P.F. (1983) 254 Shibata, K., Shingu, K., Vassilev, V.P.,
Int. J. Biochem., 15, 797–816. Nishide, K., Fujita, T., Node, M.,
239 Arora, K.K., Collins, J.G., MacLeod, J.K., Kajimoto, T., and Wong, C.-H.
and Williams, J.F. (1988) Biol. Chem. (1996) Tetrahedron Lett., 37,
Hoppe Seyler, 369, 549–557. 2791–2794.
240 Castillo, J.A., Calveras, J., Casas, J., 255 Nishide, K., Shibata, K., Fujita, T.,
Mitjans, M., Vinardell, M.P., Parella, T., Kajimoto, T., Wong, C.-H., and
Inoue, T., Sprenger, G.A., Joglar, J., and Node, M. (2000) Heterocycles, 52,
Clapes, P. (2006) Org. Lett., 8, 6067–6070. 1191–1201.
241 Garrabou, X., Castillo, J.A., 256 Wu, S.H., Shimazaki, M., Lin, C.C.,
Guerard-Helaine, C., Parella, T., Joglar, J., Qiao, L., Moree, W.J., Weitz-Schmidt, G.,
Marielle, L., and Clapes, P. (2009) Angew. and Wong, C.-H. (1996) Angew. Chem. Int.
Chem. Int. Ed., 48, 5521–5525. Ed. Engl., 35, 88–90.
242 Liu, J.Q., Dairi, T., Itoh, N., Kataoka, M., 257 Sagui, F., Conti, P., Roda, G.,
Shimizu, S., and Yamada, H. (2000) Contestabile, R., and Riva, S. (2008)
J. Mol. Catal. B: Enzym, 10, 107–115. Tetrahedron, 64, 5079–5084.
243 Kimura, T., Vassilev, V.P., Shen, G.J., and 258 Miller, M.J. and Richardson, S.K.
Wong, C.-H. (1997) J. Am. Chem. Soc., (1996) J. Mol. Catal. B: Enzym., 1,
119, 11734–11742. 161–164.
244 Saeed, A. and Young, D.W. (1992) 259 Gutierrez, M.L., Garrabou, X., Agosta, E.,
Tetrahedron, 48, 2507–2514. Servi, S., Parella, T., Joglar, J., and
References j917
Clapes, P. (2008) Chem. Eur. J., 14, 274 Seebeck, F.P. and Hilvert, D. (2003)
4647–4656. J. Am. Chem. Soc., 125, 10158–10159.
260 Miura, T., Fujii, M., Shingu, K., 275 Fesko, K., Giger, L., and Hilvert, D.
Koshimizu, I., Naganoma, J., (2008) Bioorg. Med. Chem. Lett., 18,
Kajimoto, T., and Ida, Y. (1998) 5987–5990.
Tetrahedron Lett., 39, 7313–7316. 276 Seebeck, F.P., Guainazzi, A.,
261 Tanaka, T., Ozawa, M., Miura, T., Amoreira, C., Baldridge, K.K., and
Inazu, T., Tsuji, S., and Kajimoto, T. Hilvert, D. (2006) Angew. Chem. Int. Ed.,
(2002) Synlett, 1487–1490. 45, 6824–6826.
262 Nakahara, S., Tanaka, T., Noguchi, K., 277 Toscano, M.D., M€ uller, M.M., and
Nozaki, K., Tsuji, S., Miura, T., and Hilvert, D. (2007) Angew. Chem. Int. Ed.,
Kajimoto, T. (2004) Heterocycles, 63, 46, 4468–4470.
779–784. 278 Nozaki, H., Kuroda, S., Watanabe, K., and
263 Tanaka, T., Tsuda, C., Miura, T., Inazu, T., Yokozeki, K. (2008) Appl. Environ.
Tsuji, S., Nishihara, S., Hisamatsu, M., Microbiol., 74, 7596–7599.
and Kajimoto, T. (2004) Synlett, 243–246. 279 Nozaki, H., Kuroda, S., Watanabe, K., and
264 Nishiyama, T., Kajimoto, T., Mohile, S.S., Yokozeki, K. (2008) Biosci. Biotechnol.
Hayama, N., Otsuda, T., Ozeki, M., and Biochem., 72, 2580–2588.
Node, M. (2009) Tetrahedron: Asymmetry, 280 Williams, G.J., Nelson, A.S., and Berry, A.
20, 230–234. (2004) Cell. Mol. Life Sci., 61,
265 Nishiyama, T., Mohile, S.S., Kajimoto, T., 3034–3046.
and Node, M. (2007) Heterocycles, 71, 281 Hult, K. and Berglund, P. (2003) Curr.
1397–1405. Opin. Chem. Biol., 14, 395–400.
266 Steinreiber, J., Sch€urmann, M., 282 Branneby, C., Carlqvist, P., Hult, K.,
Wolberg, M., van Assema, F., Brinck, T., and Berglund, P. (2004)
Reisinger, C., Fesko, K., Mink, D., and J. Mol. Catal. B: Enzym., 31, 123–128.
Griengl, H. (2007) Angew. Chem. Int. Ed., 283 Enders, D. and Narine, A.A. (2008)
46, 1624–1626. J. Org. Chem., 73, 7857–7870.
267 Steinreiber, J., Faber, K., and Griengl, H. 284 Fusz, S., Eisenf€uhr, A., Srivatsan, S.G.,
(2008) Chem. Eur. J., 14, 8060–8072. Heckel, A., and Famulok, M. (2005)
268 Steinreiber, J., Sch€urmann, M., van Chem. Biol., 12, 941–950.
Assema, F., Wolberg, M., Fesko, K., 285 Tanaka, F. and Barbas, C.F. (2004)
Reisinger, C., Mink, D., and Griengl, H. Antibody-catalyzed Aldol Reactions in
(2007) Adv. Synth. Catal., 349, 1379–1386. Modern Aldol Reactions, vol. 1 (ed
269 Herbert, R.B., Wilkinson, B., and R. Mahrwald), Wiley-VCH Verlag GmbH,
Ellames, G.J. (1994) Can. J. Chem., 72, Weinheim, pp. 273–310.
114–117. 286 Tanaka, F., Fuller, R., and Barbas, C.F.
270 Liu, J.Q., Odani, M., Dairi, T., Itoh, N., (2005) Biochemistry, 44, 7583–7592.
Shimizu, S., and Yamada, H. (1999) Appl. 287 Kofoed, J., Darbre, T., and Reymond, J.-L.
Microbiol. Biotechnol., 51, 586–591. (2006) Org. Biomol. Chem., 4, 3268–3281.
271 Liu, J.Q., Odani, M., Yasuoka, T., Dairi, T., 288 M€ uller, M.M., Windsor, M.A.,
Itoh, N., Kataoka, M., Shimizu, S., and Pomerantz, W.C., Gellman, S.H., and
Yamada, H. (2000) Appl. Microbiol. Hilvert, D. (2009) Angew. Chem. Int. Ed.,
Biotechnol., 54, 44–51. 48, 922–925.
272 Baik, S.-H. and Yoshioka, H. (2009) 289 Jiang, L., Althoff, E.A., Clemente, F.R.,
Biotechnol. Lett., 31, 443–448. Doyle, L., Rothlisberger, D.,
273 Paiardini, A., Contestabile, R., Zanghellini, A., Gallaher, J.L., Betker, J.L.,
DAguanno, S., Pascarella, S., and Tanaka, F., Barbas, C.F., Hilvert, D.,
Bossa, F. (2003) Biochim. Biophys. Acta, Houk, K.N., Stoddard, B.L., and Baker, D.
1647, 214–219. (2008) Science, 319, 1387–1391.
j919
22
Acyloin and Benzoin Condensations
Martina Pohl, Carola Dresen, Maryam Beigi, and Michael M€
uller
22.1
Umpolung Reactions in Chemistry and Biology
The benzoin condensation, the homo-coupling of two benzaldehyde molecules in the

presence of cyanide, was initially published by Liebig and W€ ohler in the journal
Annalen der Pharmacie in 1832 [1]. This reaction is now attributed as the first name
reaction in synthetic chemistry. It was then 50 years later before Emil Fischer used
this transformation for the cross-coupling of two different aldehydes, benzaldehyde
with furfural, resulting in a mixed benzoin [2]. Since then many applications of this
useful umpolung reaction have been introduced. The demonstrated usefulness of
the classical cyanide-catalyzed benzoin condensation and the versatility of the
thiazolium organocatalyst resulted in an unprecedented rise of new organocatalytic
transformations. In a review by Enders et al. covering more than 200 publications,
most of which were published between 1990 and 2007, a comprehensive in-depth
overview is given about the progress of organocatalysis by N-heterocyclic
precatalysts [3].
Apart of the development of organocatalytic transformations thiamine diphos-
phate (ThDP)-dependent enzymatic transformations have been explored. As early as
1921, the first modern biotechnological process on an industrial scale, based on a
bakers yeast whole cell biotransformation with a ThDP-dependent key transforma-
tion, was invented [4]. In a recent review by Kluger and Tittmann the state-of-the-art
with respect to enzymic and non-enzymic covalent heterozolium intermediates is
summarized [5].
The discovery of ThDP (Figure 22.1) as a cofactor and the elucidation of its catalytic
function are landmarks in both organic chemistry and biochemistry. The parallel
discovery of enzymatic and non-enzymatic reactivity in this system is a role model for
cofactor-dependent enzymes as well as organocatalysts.
Thiamine in its active form, ThDP, serves as a key cofactor in all forms of life. The
early evolutionary emergence of thiamine is suggested by its essential role in most, if
not all, organisms and its requirement at several central points in metabolism. It is
likely that thiamines existence even predates that of functional proteins. Among
and Oliver May.
j 22 Acyloin and Benzoin Condensations
920
CH3
1'
N N
OP2O63-
H 3C N 4' NH2 H 2 S
Figure 22.1 Structure of thiamine diphosphate (ThDP).
other attributes, ThDP-dependent enzymes have key functions in the energy-yielding

metabolism of carbohydrates of muscle and brain, in the respiratory chain, the Krebs
cycle, the pentose phosphate pathway, and in anaerobic fermentation. In addition,
they play a role in the biosynthesis of amino acids, neurotransmitters, and those
pentoses used as nucleic acid precursors. The last role causes increased thiamine
utilization in tumor cells. Moreover, these enzymes are important for the production
of reducing equivalents used in oxidative stress defenses.
Thiamine, produced by plants, fungi, and bacteria, has to be ingested by humans,
i.e. it is an essential part of the diet. Thiamine deficiency results in serious
neurological disorders, known as Beriberi and the Wernicke–Korsakoff syndrome.
In contrast, the absence of some ThDP-dependent enzymes like acetohydroxyacid
synthase or SEPHCHC synthase (MenD) in mammals suggests they are targets for
specific inhibitors, for example, for the development of herbicides or antibiotics.
ThDP-dependent enzymes are exceedingly multifunctional biocatalysts involved
in numerous metabolic pathways and catalyze a broad range of reactions. In their
varied metabolic roles they are involved in the making and breaking of bonds between
carbon and hydrogen, oxygen, sulfur, or nitrogen, and, most importantly, between
two carbon atoms. Any enzymatic reaction involving a carbon–carbon bond that links
two vicinal carbonyl groups or the carbinol and carbonyl groups of a 2-hydroxy ketone
is almost certain to be ThDP-dependent. This chapter deals with the versatility of
asymmetric thiamine catalysis, with special emphasis on enzymatic carbon–carbon
bond ligations leading to acyloins and benzoins.
22.2
Acyloin Condensations
The potential of ThDP-dependent enzymes to catalyze the cross acyloin condensation

(In this review we will use ‘acyloin condensation’ instead of ‘benzoin condensation’
to point out the general applicability) was already described in 1921 by Neuberg
and Hirsch [6], who investigated the formation of (R)-phenylacetylcarbinol (PAC), a
precursor of (1R,2S)-(–)-ephedrine, by fermenting yeast. This reaction was one of
the first commercialized biotechnological processes (Scheme 22.1) [7].
A basic result of these early studies was the enzymatic promiscuity of pyruvate
decarboxylase (PDC) in yeast: this enzyme catalyzes both the decarboxylation of
pyruvate to acetaldehyde and its subsequent carboligation with benzaldehyde to form
(R)-PAC. Based on these early findings, a detailed investigation of this reaction type
was started in the 1990s, when the first crystal structures of ThDP-dependent
enzymes became available and detailed studies of the reaction mechanism were
22.2 Acyloin Condensations j921
O O OH H2NCH3 OH
OH yeast Pt/H2
+
O -CO2 O HN
pyruvate (R)-PAC (1R,2S)-ephedrine
glucose
Scheme 22.1 Chemoenzymatic synthesis of (1R,2S)-ephedrine [7].
started [8–10]. The reaction mechanism of 2-ketoacid decarboxylases, shown in

Figure 22.2, shows that lyase and carboligase activity occur at the same active site,
which is constituted by enzyme bound ThDP.
The reactions start with a chemically challenging breaking of a carbon–carbon or
carbon–hydrogen bond adjacent to a carboxyl group in the substrate (Figure 22.2b,c).
To mediate this activity the C2 carbon of the ThDP thiazolium ring first needs to be
activated by deprotonation to form a potent nucleophilic ylide. Pioneering work in this
field has been published by, among others, Breslow (1957) [11] and Kern et al.
(1997) [10]. For structural insights into the mechanism of enzymatic decarboxylation
see References [12–14]. A recent survey about the known mechanism of ThDP-
enzymes is given in Reference [15].
This deprotonation is supported by a conserved glutamatic acid residue. This
residue is present in all ThDP-dependent enzymes, except glyoxylate carboligase
R= OP2O63-
R'
O N R' = N
O O
S R H2N N
OH
R1 HO R1 CO2
O
B
R' -H+ R' R' R'

N N N N
HO HO
H R +H+ S R R S R
S S
R1 C R1
ThDP A ylide
'active aldehyde' O
H R2
-H+
O
R' +H+
R1 H N O
HO
H S R 1 R2
R
R1
OH
D E
Figure 22.2 Reaction mechanism of the lyase and ligase activity of 2-ketoacid decarboxylases.
922
[16, 17], which stabilizes the formation of an 10 ,40 -imino tautomer in the pyrimidine
ring. The constrained V-conformation of the cofactor is essential for the formation of
the ylide (Figure 22.2a) since the C2 proton is positioned at a reactive distance to the
40 -imino group of the neighboring pyrimidine ring (Figure 22.1).
In the first step of a decarboxylase reaction (Figure 22.2a-d) the carbonyl group of
the 2-ketoacid substrate reacts with the ylide form of ThDP (a) to yield the
corresponding 2-hydroxy acid adduct (b). After CO2 is released, a carbanion-
enamine, the so-called active aldehyde, is formed as a reactive intermediate (c).
Subsequently, this intermediate is protonated (d) and the corresponding aldehyde
is released, thereby reconstituting the ylide (a).
Carboligation reactions result from the binding of a carbonyl compound, such as an
aldehyde, as the second substrate (acceptor aldehyde), leading to the formation of 2-
hydroxy ketones (e). In this case again a proton donor is required to neutralize the
resulting negative charge. Therefore, all ThDP-dependent enzymes need a proton
acceptor in the first half of the reaction cycle and a proton donor in the second half.
While the proton acceptor for the first step is the 10 ,40 -imino tautomer of ThDP in
combination with an almost invariant glutamate residue (see above), the proton relay
systems for the second step are different in these enzymes [15]. As was shown for PDCs,
branched-chain ketoacid decarboxylase from Lactococcus lactis (KdcA) and phenylpyr-
uvate decarboxylase from Azospirillum brasiliense (PPDC) [20,21], two adjacent
thiazolium-proximal histidine residues are sequentially and positionally conserved.
Whereas the respective histidine residues in benzoylformate decarboxylase (BFD) from
Pseudomonas putida [18, 19] and benzaldehyde lyase (BAL) from Pseudomonas fluor-
escens [22, 23] originate from different positions in the protein sequence (see also
Reference [24]). Since binding of an aldehyde as the donor substrate without a
preliminary decarboxylation step is also possible in many cases, the decarboxylation
ofa2-ketoacidisnotstrictlymandatoryforcarboligation.Exceptionsarediscussedbelow.
In summary, this reaction mechanism explains the two possible products that can
occur: if the acceptor is a proton a simple decarboxylase reaction takes place and an
aldehyde is released. If the acceptor is an aldehyde acyloin condensation takes place.
As both reactions occur at the same active site, the steric and electronic properties of
the active center influence both reactions similarly. Further, the mechanism explains
the principles of chemoselectivity, meaning the reaction sequence of two aldehydes
forming the resulting acyloin: the donor aldehyde is activated by ThDP and con-
stitutes the carbonyl part in the final acyloin, whereas the acceptor aldehyde forms the
carbinol part (see Figure 22.3 below).
22.2.1
Chemoselectivity of Enzymatic Acyloin Condensations
In principle four different chiral acyloins can be expected from a mixed carboligation
employing two different aldehydes (Scheme 22.2). The scope of products depends on
the properties of the enzymes active sites, the substrates, and the reaction conditions
(e.g., stoichiometric ratio of substrates).
O OH
OH OH
CH 3 CH3
O donor O donor
OH O
(R)-2-HPP (R)-PAC
O
OH
CH3 OH OH CH3
O d
donor O d
donor
OH O
(S)-2-HPP S-pocket
acceptor S-pocket (S)-PAC
Figure 22.3 Chemoselectivity of ThDP-dependent enzymes depends on the binding sequence of

donor and acceptor to the active site, whereas the relative orientation of donor and acceptor directs
the stereoselectivity of the carboligation reaction.
O O
acetoin
OH OH
O O
phenylacetyl-
carbinol
O OH OH (PAC)
O
O O
2-hydroxypropio-phenone
(2-HPP)
OH OH
O O
benzoin
OH OH
(R) (S)
Scheme 22.2 Possible products resulting from the carboligation of benzaldehyde and
acetaldehyde.
22.2.2
Stereoselectivity of Enzymatic Acyloin Condensations
One significant advantage of enzyme-catalyzed acyloin condensations is the high

stereoselectivity of such reactions. Investigation of the PAC synthesis by different
ThDP-dependent enzymes demonstrated that the (R)-enantiomer is the highly
preferred product in almost all cases. The high preference of the (R)-enantiomer can
again be explained based on the architecture of the enzymes active sites; at least for
the 2-ketoacid decarboxylases and BAL. Owing to steric constrains defined by the
924
orientation of the ThDP-bound donor aldehyde and the location of the mechanis-
tically relevant proton donors/acceptors (see above), there is only limited space
available for the binding of the acceptor aldehyde. As shown in Figure 22.3, the
acceptor may approach the ThDP-bound donor aldehyde in a parallel or antiparallel
mode, resulting in the formation of an (R)- or (S)-2-hydroxy ketone, respectively.
An antiparallel approach of the acceptor aldehyde is only possible if sufficient space
is available in the enzyme to accomodate the aldehydes side chain. This space the so-
called S-pocket, has been identified in only a few enzymes. One of them is BFD from
Pseudomonas putida, which can bind acetaldehyde in this way, yielding (S)-2-hydro-
xypropiophenone ((S)-2-HPP) with benzaldehyde as the donor (see below) [25].
An overview about those ThDP-dependent enzymes, whose carboligation activity
has been studied intensively, is given in References [26–29]. In this review we will
focus on the different formation of the various acyloins.
22.2.3
Aliphatic–Aromatic Acyloins
22.2.3.1 Acyloin Condensations with Aliphatic Donor Aldehydes and Aromatic

Acceptors
Based on the fermentative production of PAC (Scheme 22.1), the cross acyloin
condensation of benzaldehyde and acetaldehyde has been intensively studied
with various decarboxylases (such as a variety of PDCs, BFD, and KdcA). In addition,
other ThDP-dependent enzymes, such as acetohydroxyacid synthase (AHAS) and
(1R,6R)-2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexadiene-1-carboxylate (SEPHCHC)
synthase (MenD), have also been studied (Scheme 22.3).
(a) O PDC or
HO O MenD
H
+
- CO2
H3 C O
OH
O CH3
KdcA or
O H PDC
O
+
H3 C H (R)-PAC
- CO2
(b)
O OH
HO O MenD
H CO2H
HO2C +
- CO2 O
O
Scheme 22.3 Different enzyme-catalyzed (R)-PAC syntheses. (a) PDCs, KdcA and MenD catalyze
the synthesis of (R)-PAC either using (decarboxylated) pyruvate or acetaldehyde, respectively.
(b) Alternatively, (R)-PAC is also accessable by MenD starting from 2-ketoglutarate as the donor.
All PDCs prefer small aliphatic donor aldehydes (in the form of the respective 2-
ketoacid) and aliphatic or aromatic acceptor aldehydes. Thus, acetoin and PAC
(Scheme 22.2) are typical ligation products of PDCs, whereas 2-HPP and benzoin are
observed only in trace amounts, if at all [30–32]. While PDCs from yeasts are usually
applied in whole-cell biotransformation either with resting cells or in a fermentative
process, the PDC from Zymomonas mobilis (ZmPDC) can be used as an isolated
enzyme, due to its higher stability. The carboligase activity of ZmPDC was improved
by mutagenesis of tryptophan-392, which is a key residue limiting access to the active
center for sterically demanding aromatic substrates [33, 34]. The highly potent variant
ZmPDC-W392M was tested in a continuous enzyme-membrane reactor, giving
space–time yields of 81 g l1 d1 [32, 35, 36]. Recently, a further potent carboligating
variant was described, which was obtained by substitution of Glu473 by gluta-
mine [37]. ZmPDC is also unique due its high stereoselectivity forming amost
enantiopure (R)-PAC (98% e.e.) [38]. This is in contrast to Acetobacter pasteurianus
PDC (ApPDC) [31], which yields (R)-PAC in 91% e.e.. Similar results were reported
for several whole-cell biotransformations employing various yeast strains for which
enantiomeric excesses in the range 90–94% (R)-PAC were obtained [30, 39]. How-
ever, almost enantiomerically pure (R)-PAC was obtained with two variants of yeast
PDC (ScPDC-E477Q, -D28A) [40]. In addition, the synthesis of ring-substituted PAC
derivatives has been reported using ScPDC and ZmPDC [41, 42]. The investigation
and optimization of the PAC synthesis by different yeast strains is ongoing, and has
resulted in PAC production with space–time yields of >100 g l1 d1 [30, 43–47].
Recent approaches involve the use of non-conventional media in aqueous–organic
two-phase systems [44, 48], supercritical CO2 as well as liquefied gases [49], and PEG-
induced cloud point systems [47].
Most recently, ApPDC variants with tailor-made catalytic activities were designed
and generated. Whereas the exchange of Trp388 by smaller amino acids yields
variants with higher carboligase activity (such as ZmPDC-W392M), the replacement
of Glu469 (analogous to ZmPDC-E473) by glycine opens the S-pocket in ApPDC for
aromatic aldehydes and thus alters the stereoselectivity. The variant ApPDC-E469G
provides access to (S)-PAC derivatives by enzymatic carboligation with enantioselec-
tivity of up to 89% e.e. [50].
The branched-chain keto acid decarboxylase (KdcA) from Lactococcus lactis shows
almost no chemoselectivity in its carboligase reaction with acetaldehyde and benz-
aldehyde. As demonstrated in Scheme 22.4 both aldehydes may act either as donor or
acceptor, yielding almost equimolar amounts of (R)-PAC and (R)-2-HPP, which is
probably a consequence of the steric properties in the active site. With larger aliphatic
aldehydes such as propanal, cyclopropylcarbaldehyde, and isovaleraldehyde KdcA
catalyzes exclusively the formation of the respective PAC-derivatives (1–3). The 2-
HPP-derivative 4 is exclusively formed, if 3,5-dichlorobenzaldehyde is employed
together with acetaldehyde (Scheme 22.4) [51]. Thus, KdcA is an impressive example
of how the chemoselectivity of the carboligation reaction can be influenced by an
appropriate combination of donor and acceptor aldehydes. These observations can be
explained based on the 3D-structure of the enzyme [52].
An alternative way to catalyze the synthesis of (R)-PAC has been described for
SEPHCHC synthase, an enzyme that is part of the menaquinone biosynthesis
926
O O OH O
R' KdcA
+
[ThDP] O OH
R
40 : 60
(R)-PAC (R)-2-HPP
92% ee 93% ee
OH
O
1, >98% ee (R)
O
OH Cl
OH
O
Cl
2, >98% ee (R)
4, 96.5% ee (R)
OH
O
3, 88% ee (R)
Scheme 22.4 Chemoselectivity of KcdA from Lactococcus lactis is influenced by the combination of
substrates [51].
pathway. It is dencoded by the gene menD and is also termed MenD [53, 54]. The
physiological donor of MenD, 2-ketoglutarate, can be replaced by oxaloacetate; but it
results in reduced enzymatic activity. When incubated with the acceptor 2-fluoro-
benzaldehyde, oxaloacetate reacts to 2-fluoro-PAC (24% yield) (cf. Scheme 22.3). The
(R)-configuration was determined for the product by chiral phase HPLC. Thus,
oxaloacetate can be used as a mimic for pyruvate or acetaldehyde, depending on the
enzymes preference [29]. MenD accepts a wide range of aldehydes as acceptor
substrates to produce chiral 2-hydroxy ketones with conserved regioselectivity, with
the active succinyl-semialdehyde serving as selective donor (Scheme 22.5b) [29].
Two isoenzymes of acetohydroxyacid synthase from Escherichia coli (AHAS) have
been intensively studied concerning their potential to catalyze the formation of chiral
2-hydroxy ketones, such as acetoin and PAC and derivatives thereof [55–59]. As was
observed with most other ThDP-dependent enzymes, both isoenzymes are strictly
(R)-specific concerning the formation of PAC derivatives [59]. AHAS I proved to be
especially useful for 2-hydroxy ketone formation and the substrate range concerning
the formation of chiral mixed 2-hydroxy ketones from aromatic or heteroaromatic
aldehydes and acetaldehyde (generated by decarboxylation of pyruvate) was studied in
detail [57]. Although several AHASs produce PAC in the presence of pyruvate and
benzaldehyde, none of the wild-type isozymes in bacteria are particularly efficient at
this. They all preferentially produce acetolactate by carboligation of decarboxylated
(a) CO2 MenD O CO2-
O H
OH [ThDP] - OH
CO2- O2C
HO - CO2
O CO2- O O CO2-
isochorismate SEPHCHC
O OH
(b) MenD
O [ThDP] CO2-
H
CO2-
HO - CO2 O
R3 R1 R3 R1
O 2
R2 R
26-87% yield
>94% ee
MenD OH
(c) O O [ThDP] CO2-
CO2-
H HO - CO2
O
O 83% yield
25% ee
(d) CO2
MenD
OH O O CO2-
[ThDP] H
CO2- - OH
HO - CO2 O2C
OH
O OH
2,3-CHD
Scheme 22.5 MenD-catalyzed physiological (a) and non-physiological [(b)–(d)]

transformations [29].
pyruvate (active acetaldehyde) and pyruvate, even at benzaldehyde concentrations

high enough to significantly inhibit the enzyme. However, replacement of the highly
conserved arginine residue Arg276 in AHAS II by methionine, glutamine, or even
lysine leads to enzyme variants with reduced acetolactate synthase activity. These
variants catalyze the formation of PAC at least one order of magnitude faster than the
formation of acetolactate (in the presence of a standard reaction mixture with 30 mM
of pyruvate and benzaldehyde) [58, 60].
22.2.4
Carboligation of Aromatic Donors and Aliphatic Acceptors
Up to now the only biochemically characterized benzaldehyde lyase (BAL) was found
in Pseudomonas fluorescens [61, 62]. Besides the synthesis of benzoins and acyloins
(see below) BAL is the most efficient catalyst for the mixed carboligation of aromatic
donors and aliphatic acceptor aldehydes, due to its broad substrate range and its strict
(R)-selectivity. Further, BAL is the most active ThDP-dependent carboligase currently
928
known, with specific activities of up to >400 U mg1 reported for the synthesis of
benzoins and mixed araliphatic 2-hydroxy ketones. Various benzaldehyde derivatives
can be used as donors in carboligations with a broad range of aliphatic acceptor
aldehyde, including formaldehyde [63–66]. BAL is currently also the only known
ThDP enzyme with carbolyase activity on benzoins, which enables the application of
this enzyme in kinetic resolution of racemic benzoins (see below) [65, 67, 68]. BAL
has been applied in several bioreactors for the synthesis of various 2-hydroxy ketones
with good productivities [63, 64, 69–76]. Owing to the limited solubility in aqueous
buffer of the aromatic aldehydes and reaction products, the influence of several
organic cosolvents in monophasic and biphasic systems on the isolated enzyme as
well as the whole cell biocatalyst (recombinant E. coli strain) has been probed, with the
finding that DMSO, methyl tert-butyl ether (MTBE) [77], and 2-methyltetrahydro-
furan (2-MTHF) [78] are appropriate cosolvents [69, 72]. Although organic solvents
increase the solubility of aromatic compounds they influence the enzymes activity
and stability, as well as the reaction kinetics [79].
Benzoylformate decarboxylase (BFD) from Pseudomonas putida prefers aromatic
donor aldehydes and aliphatic aldehydes as acceptors. Although BFD does not reach
the high reaction rates obtained with BAL it is well established: BFD is one of the few
enzymes so far that catalyze the carboligation of benzaldehyde derivatives with
acetaldehyde (S)-selectively. This had already been reported by Wilcocks et al. in 1992
for the synthesis of different 2-HPP derivatives [80, 81]. Investigation of the substrate
range demonstrated that meta- and para-substituted benzaldehyde derivatives as well
as heteroaromatic, olefinic, and cyclic aliphatic aldehydes are transformed with good
to excellent conversions and enantioselectivities [82–84]. The activity towards ortho-
substituted benzaldehydes has significantly been improved by directed evolution of
BFD, yielding variants with generally improved stereoselectivity, ligase activity, and
higher stability in the presence of organic cosolvents [85, 86].
The structural reason for the (S)-selectivity of BFD has recently been elucidated by
site-directed mutagenesis [25, 87] based on its X-ray crystal structure analysis.
Presence of a small S-pocket allows the antiparallel approach of acetaldehyde towards
the ThDP-benzaldehyde-adduct (Figure 22.3). Structure-based engineering of BFD
yielded variants with an increased S-pocket, which form (S)-2-HPP derivatives with
propanal and butanal, respectively, as the acceptor substrates [87].
Depending on the substitution pattern of the aromatic ring diverse 2-HPP analo-
gues are accessible in high yields and with good to high optical purity. Selectivity,
activity, and stability of BFD have been optimized using reaction engineering. Best
results have been obtained by adjusting very low benzaldehyde concentrations in a
continuous reactor [82]. As with BAL, BFD has also been applied in different types of
bioreactors, either as purified enzyme or as whole cells [69, 76, 82, 83, 88, 89].
22.2.5
Araliphatic–Aliphatic Acyloins
CH-acidic aldehydes such as phenylacetaldehyde and indole-3-acetaldehyde are

prone to enolization and thus are difficult substrates for enzymatic as well as
chemical transformations [90]. Moreover indole-3-acetaldehyde is very unstable
and decomposes rapidly, which makes its direct application in biotransformations
very difficult [91]. Phenylacetaldehyde is commercially available; however, in
aqueous buffer aldol reaction occurs spontaneously. Yet, these problems can
easily be overcome by in situ production of these aldehydes through enzymatic
decarboxylation of the corresponding 2-ketoacids. Referring to the reaction
mechanism (Figure 22.2), the decarboxylation of a 2-ketoacid results in a reaction
intermediate (active aldehyde) with the corresponding aldehyde bound to the
cofactor ThDP, which can further react with a suitable electrophile/acceptor
aldehyde.
Carboligations employing phenylacetaldehyde have been reported with BAL [92]
and phenylpyruvate decarboxylase (PPDC) from Achromobacter eurydice [93, 94].
KdcA [51, 95, 96] and PPDCs from both S. cerevisiae [24, 97] and Azospirillum
brasilense (AbPPDC) [98] are able to decarboxylate phenyl pyruvate. KdcA and
AbPPDC were also shown to decarboxylate indole-3-pyruvate with low activity [51,
96, 98]. Further, KdcA can catalyze the formation of the 2-HPP analog from
acetaldehyde and indole-3-pyruvate as substrates with high chemoselectivity [51].
22.2.6
Aliphatic Acyloins
The self-ligation of aliphatic acyloins has recently been described using BAL, BFD,
and KdcA as biocatalysts. Highest stereoselectivities have been observed with
branched-chain aldehydes, for example, isovaleraldehyde, yielding enantio-comple-
mentary products with BAL/BFD [70] and KdcA [51]. Nevertheless, the enantio-
selectivity of this reaction is in all cases only low to moderate – most probably a
consequence of less stabilization of small aldehydes in the active site [27]. Likewise,
the enantioselectivity of the enzymatic acetoin formation catalyzed by PDCs, for
example, from S. cerevisiae, (ScPDC) Z. mobilis, (ZmPDC) A. pasteurianus, (ApPDC)
and Zymobacter palmae, is only moderate [31]. Similar results were obtained with
KdcA [51] and BFD [70]. In addition, an enantiomeric excess of 94% was reported for a
variant of ScPDC (ScPDC-E477Q) [40]. A solid/gas bioreactor system was investi-
gated with BAL and BFD as catalysts in the acyloin condensation of propanal [99].
Kinetic data for the enzyme-catalyzed enantioselective propion formation were
determined in aqueous solution as well [100].
22.2.7
Olefinic Aliphatic and Araliphatic Acyloins
Both BFD and BAL catalyze the carboligation (1,2-addition) of aliphatic or aromatic
a/b-unsaturated donor aldehydes with acetaldehyde, acetaldehyde derivatives, or
formaldehyde as acceptors with good yields and usually high stereoselectivity. The
isomeric products can be obtained using yeast PDC, which catalyzes the 1,2-addition
of pyruvate (active acetaldehyde) and several aliphatic a/b-unsaturated acceptor
aldehydes with good conversions and very good stereoselectivity [101].
930
MenD is known to catalyze the 1,4-addition of a ThDP adduct onto the b-carbon
atom of a a,b-unsaturated carboxylate; this reaction therefore constitutes a Michael-
type addition similar to the Stetter reaction [29]. The physiological reaction catalyzed
by MenD is the decarboxylation of 2-keto glutarate and the concomitant (decarboxy-
lase activity with release of succinyl semialdehyde has not been identified for MenD
so far) addition of the resulting succinyl-semialdehyde-ThDP to isochorismate,
formed itself from chorismate (Scheme 22.5a). Addition of 2-ketoglutarate after
decarboxylation to a broad range of aldehydes gave 2-hydroxy ketones with isolated
yields of 26–87% and 94–97.5% e.e. by (Scheme 22.5b and c). The physiological 1,4-
addition of 2-ketoglutarate to isochorismate was enlarged to 2,3-dihydroxy-2,3-
dihydrobenzoate (2,3-CHD) [102] as a substrate, which lacks the pyruvyl moiety
group found in isochorismate (Scheme 22.5d). Hence, a wide variety of new chiral
building blocks are available through effective asymmetric enzymatic synthesis with
MenD [29].
22.2.8
2-Acyl-2-Hydroxy Acids
The main activity of AHAS is the synthesis of acetolactate and acetohydroxybutyrate

by addition of pyruvate (active acetaldehyde) to pyruvate or 2-ketobutyrate as
acceptors, respectively. AHAS prefers pyruvate by more than tenfold over any other
ketoacid as the first substrate and steric hindrance seems to be the major factor for
this specificity. A variant of AHAS II from E. coli with a V375A exchange shows a
preference for 3-ketobutyrate as the first substrate, yielding 2-propio-2-hydroxybu-
tyrate [60]. Carboligation reactions catalyzed by AHAS II from E. coli were shown to
give (S)-2-acyl-2-hydroxy acids preferentially [103].
22.2.9
Sugar Derivatives
Transketolase (TK) from yeast and E. coli is catalyzes the reversible transfer of two-
carbon (dihydroxyethyl) fragments between ketose and aldose substrates [104].
Because of its innate carboligation competence, this enzyme has attracted consid-
erable interest as a tool for chemoenzymatic synthesis in particular for two-carbon
chain extension resulting in 1,3-dihydroxy ketones with (3S) product configura-
tion [105]. Although carbohydrates such as D-erythrulose were shown to be viable
substrates for TK-catalyzed carboligation reactions, the utilization of hydroxypyru-
vate as an alternative donor substrate has found widespread application [106]. A clear
advantage of employing hydroxypyruvate rather than sugar substrates is due to the
quasi-irreversibility of the decarboxylation step. In addition, sugar donors will be
processed in a reversible manner and mechanistic NMR-based analysis further
revealed that the initial tetrahedral donor-ThDP adduct is thermodynamically sta-
bilized under equilibrium conditions with negligible amounts (<5%) of the reactive
carbanion/enamine intermediate (see Figure 22.2) being present in the rapid
equilibrium [107].
22.3 Benzoin Condensations j931
With hydroxypyruvate as a donor, TK could be successfully exploited to catalyze
two-carbon chain elongations of many phosphorylated and non-phosphorylated
hydroxylated aldehyde acceptor substrates (carbohydrates and carbohydrate-like
compounds) to yield among others deoxy-sugars (5-deoxy-xylulose, 6-deoxy-sorbose),
benzyl sugar derivates, and sugars of defined length such as xylulose 5-phosphate,
sedoheptulose 7-phosphate, or octulose 8-phosphate [105].
Directed evolution of TK from E. coli yielded variants with improved activities
towards non-native aldehydes and non-phosphorylated aldose acceptors [108–110].
Although the substrate range of TK is quite broad, caution should be taken with
respect to some substrates that might be prone to non-enzymatic side reaction. As we
could show, mixing hydroxypyruvate and 2-fluorobenzaldehyde in phosphate buffer
results in a clean conversion towards 1,3-dihydroxy fluoroacetophenone in the
absence of TK or other ThDP-dependent enzymes (P. Lehwald and A. M€ uller,
unpublished results). This is nicely matched by a biomimetic TK reaction as published
by Hailes et al. [111]. Most recently, the same group published that for many putative
TK-catalyzed transformations no data corresponding to the TK products have ever
been reported [112]. Instead of product formation, the decrease of hydroxyl pyruvate
has often been used to determine supposed TK activity.
ThDP-dependent 1-deoxy-D-xylulose 5-phosphate synthase (DXPS) is the starting
point of the non-mevalonate pathway of terpene biosynthesis (MEP pathway). The
enzyme from E. coli has been used in the synthesis of 1-deoxysugar phosphates and
derivatives thereof [113, 114].
22.3
Benzoin Condensations
22.3.1
Benzoin Condensations
BFD and BAL catalyze the benzoin condensation, acting on a broad range of aromatic
aldehydes to give (R)-benzoins in high yields and excellent enantiomeric excess [115, 116].
The physiological function of BFD is the non-oxidative conversion of benzoylfor-
mate into benzaldehyde and CO2 [80, 81, 117]. In 1998 it was mentioned for the first
time that BFD from P. putida is able to catalyze the condensation of two benzaldehyde
molecules to (R)-benzoin (Scheme 22.6) [82, 83, 118]. Further studies showed that
this BFD has an activity concerning the formation of (R)-benzoin (99% e.e.) of
0.25 U mg1 and can catalyze the ligation of a broad range of different aromatic and
heteroaromatic aldehydes [82, 119].
Pseudomonas fluorescens is able to grow on lignin-derived compounds like anisoin
and benzoin as the sole carbon and energy sources [120]. In 1989 BAL from this
organism was identified as catalyzing the cleavage of benzoin to yield benzaldehydes
(Scheme 22.7) [61].
The lyase reaction is reversible: BAL possesses high relative activity (up to >400
U mg1) for the ligase reaction, affording enantioselective (R)-benzoins (>99% e.e.)
932
O O
R
BFD or BAL
2 H
[ThDP] OH
R R
Scheme 22.6 Asymmetric, enzymatic synthesis of symmetric (R)-benzoins via BFD [115] or
BAL [116] catalysis.
O O
R
BAL 2 H
OH [ThDP, Me2+]
R R
Scheme 22.7 Carbon–carbon bond cleavage reaction catalyzed by benzaldehyde lyase (BAL) [61].
(Scheme 22.6). Benzaldehyde derivatives substituted at the ortho-, meta-, or para-

position are suitable substrates for BAL [116]. BAL, in comparison to BFD, has a
considerably higher ligase activity and is able to act on ortho-substituted benzalde-
hydes. To enable the conversion of hydrophobic substrates, BAL was entrapped in poly
(vinyl alcohol) and suspended in hexane. Compared to the reported application of the
biocatalyst in an aqueous phase containing 20 vol.% DMSO, the productivity of the
resulting gel-stabilized two-phase system was threefold increased [121]. The entrap-
ment in polymer beads of poly(vinyl alcohol) and continuous operation in a fluidized
bed reactor has been shown to enhance the stability of BAL. This resulted in a half-life
of more than 100 h under operation conditions, as well as superior enzyme utilization
in terms of productivity [122]. As an interesting tool for more volatile products in
carboligation, the solid/gas system, was investigated with BAL as catalyst in the
benzoin condensation [123]. To contribute to the empirically-derived kinetic models
for this stereoselective carbon–carbon coupling a mechanistic kinetic model with
accurate parameter estimates and an excellent prediction capability was derived [124].
The model suggests the release of benzoin as the rate-limiting step in BAL, which was
corroborated by NMR-based determination of reaction intermediates [79].
In contrast to BFD and BAL, PDCs prefer small aliphatic donor aldehydes as
substrates. The benzoin condensation is rarely observed with PDCs: these enzymes
catalyze the non-oxidative decarboxylation of pyruvate to acetaldehyde as the phys-
iological reaction; thus, acetoin and PAC (Scheme 22.2) are typical ligation products
of PDCs [125]. For ApPDC a very low benzoin forming activity was determined
(2.4 103 U mg1) if benzaldehyde is used as the only substrate [126]. Furthermore,
ZmPDC-I472A catalyzes the benzoin formation [34] and proved to be a real chimera
between PDC and BFD [38]. Similar behavior was observed with KdcA via self-ligation
of benzaldehyde: even with high enzyme loading only small amounts of (R)-benzoin
(>98% e.e.) could be gained [51].
22.3.2
Cross Benzoin Condensations
As shown above, enantiopure benzoins can be efficiently obtained by the use

BAL [116] or BFD [82, 115]. M€
uller et al. have presented results leading to a
22.4 Miscellaneous Acyloin Condensations j933
donor-acceptor concept for the asymmetric synthesis of mixed benzoins (benzoins
with non-identical aromatic moieties) using an enzyme-catalyzed benzoin conden-
sation (Scheme 22.8) [119].
H O H O BFD-H281A OH R2 OH
or O O R1
R2
+ BAL
+
[ThDP]
R1
R1 5 R1
donor and/or selective
acceptor acceptor
Scheme 22.8 Combined enzyme–substrate screening for the catalytic asymmetric cross-benzoin
condensation [119].
A variant of BFD from P. putida (PpBFD-H281A) and BAL were identified as

potent catalysts for asymmetric cross-carboligation. Mixed benzoins (compound 5)
were synthesized regio- and stereoselectively on a preparative scale and the
absolute configuration of the mixed benzoins was determined to be (R) (up to
>99% e.e.) [119]. Both enzymes were also compared in kinetic investigations,
demonstarting that the rate-limiting step in BAL is the product release whereas the
binding of substrates is the slowest step in PpBFD-H281A [79].
22.4
Miscellaneous Acyloin Condensations
22.4.1
Stetter-Type Reactions
The conjugate addition of aldehydes to a,b-unsaturated carbonyl compounds is

commonly known as the Stetter reaction [127]. The mechanism of the 1,4-addition
catalyzed by ThDP-dependent enzymes should be in accordance with the general
mechanism of the 1,2-addition [15]. Whereas the electrophilic acceptor in case of the
1,2-addition is the carbonyl function of another aldehyde, it is the electron-poor
double bond of the Michael system in case of a 1,4-addition (Scheme 22.9).
So far, two ThDP-dependent enzymes are known to catalyze a physiological conjugate
addition of an active aldehyde to yield 1,4-diketones. Bacterial MenD, for example, from E.
coli, catalyzes the second step in the biosynthesis of menaquinones (vitamin K2
derivatives) [128]. MenD accepts 2-ketoglutarate as a donor to form a ThDP-succinyl-
semialdehyde adduct, which is added to the b-carbon of the a,b-unsaturated carboxylic
O R'
N R4 O
+ O
R1 R3 HO
R
R2 S R1 R3
R4
R2
Scheme 22.9 ThDP-dependent intermolecular 1,4-addition.

934
acid isochorismate to produce 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-

carboxylate (SEPHCHC) selectively (Scheme 22.10) [129–131].
The acceptor substrate variability was investigated for the 1,2- and 1,4-addition
(Scheme 22.5). For the latter, the only substrate accepted by MenD apart from
isochorismate was (2S,3S)-dihydroxy-2,3-dihydrobenzoate (2,3-CHD), which lacks
the pyruvyl function at position 3 [29].
PigD has been postulated to catalyze the first step in the biosynthesis of the red
antibiotic prodigiosin in Serratia marcescens [132]. According to this work, PigD
ketoglutarate
CO2 CO2 O CO2-
MenD H
EntC OH OH
-O
2C
[ThDP]
O CO2- O CO2- -CO2 O CO2
OH
chorismate isochorismate SEPHCHC
SHCHC
synthase pyruvate
O CO2
-O C OH
menaquinones 2
SHCHC
Scheme 22.10 MenD catalyzes the second step in the biosynthesis of menaquinones starting from
chorismate [29].
decarboxylates pyruvate and catalyzes the Stetter addition of active acetaldehyde to C3

of 2-octenal to give 3-acetyloctanal (Scheme 22.11).
HN
PigD
H3 C O
O O O [ThDP] O HN
+ H3 C
C5H11 H N
C5H11 H H3 C O - CO2 OCH3
C5H11
octenal pyruvate 3-acetyloctanal prodigiosin
Scheme 22.11 Part of the prodigiosin biosynthetic pathway as postulated by Williamson et al. [132].
ThDP-dependent PigD was applied for enzymatic transformations in vitro. Alde-

hydes as substrates resulted in a selective acyloin condensation, giving 2-hydroxy
ketones as products. When a,b-unsaturated ketones instead of aldehydes were used
as acceptor substrates chemo- and stereoselective conjugate addition of pyruvate
could successfully be shown (Figure 22.4) [133].
H3C O H 3C O H3 C O
O O O
H3 C CH3 CH3
>99% ee 95% ee >99% ee
H3 C O H 3C O
H3 C O O O
O
H3 C CH3
CH3
Cl OH
>99% ee >99% ee* 94% ee
Figure 22.4 1,4-Diketones obtained by PigD-catalyzed 1,4-addition of active acetaldehyde to

a,b-unsaturated ketones [133].
A broad range of different ketone substrates were accepted by PigD, yielding

Stetter products with excellent enantiomeric excess. The absolute configuration of
the aromatic 1,4-diketones was determined as (R) [133].
22.4.2
Acyloin Condensations with Ketones and Imines
The exchange of the acceptor aldehyde by a ketone in the general mechanism of the
1,2-addition catalyzed by ThDP-dependent enzymes offers the opportunity for the
catalytic asymmetric formation of chiral tertiary alcohols. In the biosynthesis of
yersiniose A, which is a two-carbon branched-chain 3,6-dideoxyhexose found in
the O-antigen of Yersinia pseudotuberculosis, the ThDP-dependent flavoenzyme
YerE catalyzes the decarboxylation of pyruvate and the addition of the active
acetaldehyde to the carbonyl function of CDP-3,6-dideoxy-4-keto-D-glucose
(Scheme 22.12) [134, 135].
HO O CH3 HO CH HO CH
YerE 3 3
O O H 3C O reductase H 3C O
HO
HO [ThDP] NAD(P)H HO H
HO OPO 2- HO OCDP pyruvate O HO OCDP HO OCDP
3
(CDP)-3,6-dideoxy- yersiniose A
4-keto-D-glucose
Scheme 22.12 Part of the biosynthesis pathway toward yersiniose A in Yersinia

pseudotuberculosis [134].
An extended examination of the acceptor substrate range of YerE showed that cyclic
and open-chain ketones, 1,2-diketones, and a- and b-ketoesters can act as acceptor
substrates (Figure 22.5).
The enantiomeric excess of the products spans from high to moderate (96% e.e.).
Determination of the absolute configuration via single-crystal X-ray diffraction
analysis or vibrational circular dichroism showed the (R)-configuration of the
synthesized tertiary alcohols [136].
936
(a) YerE O
O O
+ [ThDP] R2
CO2H 1 2
R R
HO R1
pyruvate ketone -CO2 tertiary alcohol
(b)
O O O
O O
O HO O
Br
O O
O S
O
O O
O
O O
Br
O O O
O
O
O O
Figure 22.5 (a) YerE-catalyzed addition of active acetaldehyde to ketones and (b) the substrate
range.
Most recently, a similar transformation using acyclic 1,2-diketones as acceptor

substrate has been published by application of acetylacetoin synthase in the form of
the crude extract of Bacillus stearothermophilus [137]. However, the gene and enzyme
sequences have not been determined so far.
The enzymatic 1,2-addition of an active aldehyde as a donor and an imine as an
acceptor generating chiral 2-amino ketones has not been identified yet. Nevertheless, TK
from Amycolatopsis mediterranei is proposed to catalyze the formation of iminoerythrose
4-phosphate from ketosamine 6-phosphate after isomerization to aminofructose
6-phosphate (Scheme 22.13) [138–140]. Considering that many TKs also catalyze the
reverse reaction, this could become a role model for acyloin condensations with imines.
22.4.3
Acyloin Condensations with Formaldehyde and Formaldehyde Synthons
Elongation of aldehydes through selective carboligation with formaldehyde (or

formaldehyde synthons like gyloxylate) would be highly desirable; however, it is a
difficult task using enzymatic [93] as well as non-enzymatic chemical methods [141].
O
R
H
H2O3PO R5P OH transketolase H2O3PO
O (orf15= tktA) OH
OH HO
HO
OH HN
H2 N O OH
aminoF6P HO iminoE4P
R
S7P OH
Scheme 22.13 TK from Amycolatopsis erythrose 4-phosphate; aminoF6P, 3-amino-3-

mediterranei catalyzes the last step in the deoxy-D-fructose 6-phosphate; R5P, D-ribose-5-
formation of iminoerythrose 4-phosphate [139]. phosphate; S7P, D-sedoheptulose 7-
(Abbreviations: iminoE4P, 1-deoxy-1-imino-D- phosphate.)
Krampitz and coworkers [142–144] have detected a radioactive a,b-dihydroxyethyl-

ThDP when ThDP was incubated with [14 C]-formaldehyde at pH 8.5. Here, hydro-
xymethyl substitution is expected as the first adduct that undergoes the addition to
formaldehyde. Further incubation of this product with ribose 5-phosphate in the
presence of crystalline yeast TK led to the formation of sedoheptulose 7-phosphate,
proving that formaldehyde has played a role as the donor.
Glyoxylate is accepted by pyruvate dehydrogenase (PDH) from E. coli as the
acceptor substrate when incubated with pyruvate [145]. As shown in Scheme 22.14,
2-hydroxy-3-ketobutyric acid (6) is formed as initial product, which undergoes non-
enzymatic decarboxylation to yield the corresponding acetol 7 as the major product.
Moreover, 2,3-dihydroxy-4-ketovaleric acid (8) was identified as a minor product.
O
H 7
- CO2
O PDH O OH
O
[ThDP] CO2H
HO H
HO - CO2 O OH
O O OH
glyoxylate 8
COOH
6 - CO2 OH
Scheme 22.14 Pyruvate-glyoxylate carboligation activity of the E. coli pyruvate dehydrogenase

(PDH) complex [145].
Similar results were reported when a-ketoglutarate-glyoxylate carboligase activity

was investigated using the a-ketoglutarate dehydrogenase complex from E. coli, beef
heart [146], or an enzyme from Rhodopseudomonas spheroides [147].
Lactaldehyde formation was confirmed in the PDC-catalyzed decarboxylation of
glyoxylate in the presence of acetaldehyde, indicating that glyoxylate is accepted as
donor [148, 149]. (R)-Lactaldehyde was generated with yeast PDC while the enzymes
from Z. mobilis and from wheat germ formed predominantly the (S)-enantiomer [150,
151].
938
Dihydroxyacetone synthase (DHAS, also named as formaldehyde transketolase), is

known to catalyze the reaction between formaldehyde as a natural acceptor and
xylulose 5-phosphate as a donor [152–157]. Besides xylulose 5-phosphate as the
physiological substrate, hydroxypyruvate, fructose 6-phosphate, and ribulose 5-
phosphate were also accepted as donor substrates by some of the enzymes. Moreover,
DHAS has been used for the synthesis of [1,3-13 C]-dihydroxyacetone from
13
C-labeled formaldehyde and hydroxypyruvate [158].
ThDP-dependent glyoxylate carboligase (GCL) from E. coli uses glyoxylate as the
only substrate to give tartronic semialdehyde, which is a key intermediate in
glyoxylate catabolism [159–161]. GCL has the highest level of sequence similarity
to the AHAS and the pyruvate oxidase from E. coli (30% and 26%, respectively) [162].
Nevertheless, the CD spectrum of tartronate semialdehyde suggests a (R)-configu-
ration, which is opposite to the product obtained with the E. coli AHAS [103]. The
recently solved 3D structure of this enzyme features valine replacement for the
glutamate that is conserved in most other ThDP-dependent enzymes, known for its
interaction with the N10 of the ThDP moiety [16].
Demir and coworkers proved the direct hydroxymethylation of aromatic aldehydes
with formaldehyde catalyzed by BAL resulting in the corresponding 2-hydroxy-1-
arylethan-1-one (9) in high yields [65] (Scheme 22.15). BAL accepts a broad range of
aromatic and heteroaromatic aldehydes as donor when formaldehyde acts as the
acceptor. To avoid the formation of (R)-benzoins as a side product, an excess of
formaldehyde is required.
O O
O BAL
OH
[ThDP]
+ H H
R R 9
R = 2-OCH3 68%
R = 3-OCH3 92%
R = 4-OCH3 91%
Scheme 22.15 Synthesis of 2-hydroxyacetophenone derivatives catalyzed by BAL [65].
BAL proved to have a broad donor substrate range, since condensation of different
a,b-unsaturated aldehydes with formaldehyde as the acceptor led to very high
product yields as well (Table 22.1) [101]. Compounds 14–16 were synthesized on
a preparative scale and were isolated in 51%, 82%, and 56% yield, respectively.
22.4.4
Racemic Resolution via Lyase/Ligase Reactions
Kinetic racemic resolution offers the advantage of using less expensive racemic
starting material for the synthesis of enantiopure products. The main drawback
of the method is the maximum conversion of 50% only for the reactive
enantiomer and a maximum yield of 50% for the desired product. Enzymatic
Table 22.1 Carboligation of a,b-unsaturated aromatic and aliphatic aldehydes with formaldehyde
catalyzed by BAL [101].
O BAL O
O
[ThDP] OH
R H H H R
R' R'
10-13 14-17
Substrate R R0 Product Conv. % (% yield)
10 Ph H 14 92 (51)
11 Ph CH3 15 95 (82)
12 -(CH2)4- 16 >99 (56)
13 H CH2CH3 17 79
kinetic resolution via CC bond cleavage adjacent to a carbonyl group has been
applied in biotransformations using isolated ThDP-dependent enzymes.
BAL has been applied for the synthesis of enantiopure 2-hydroxy ketones using
enzymatic kinetic resolution of racemates by CC bond cleavage and concomitant
CC bond formation. (R)-Benzoin, in contrast to its enantiomer, is accepted as a
substrate by BAL and yields (R)-2-HPP, when acetaldehyde is present in the reaction
medium [67]. BAL-catalyzed reactions using rac-benzoin afforded (R)-2-HPP (>99%
e.e.) and (S)-benzoin (>99% e.e.) after separation of the products by column
chromatography (Scheme 22.16). This method was further applied to access both
enantiomers of mixed benzoins [119]. Similarly, the kinetic resolution of different
substituted rac-benzoins with formaldehyde afforded 2-hydroxy-1-arylethan-1-ones
and unreacted (S)-benzoins (>93% e.e.) [65].
BAL O O
O CH3CHO
[ThDP] OH OH
OH
rac-benzoin conv. >49% (S)-benzoin (R)-2-HPP

>99% ee >99% ee
Scheme 22.16 BAL-catalyzed kinetic resolution of rac-benzoin [67].
Yeast TK is an efficient catalyst for kinetic resolution of racemic 2-hydroxy

aldehydes: only the (R)-enantiomer is accepted as a donor substrate [163]. Therefore,
the TK-catalyzed reaction of racemic 2-hydroxy aldehydes with hydroxypyruvate
delivered enantiopure L-2-hydroxy aldehydes bearing different substituents at the
3-position (Scheme 22.17). Similar results were observed using TK from spin-
ach [164, 165] while TK from E. coli showed no conversion when aromatic aldehydes
such as benzaldehyde or hydroxybenzaldehydes were used as substrates [166, 167].
940
O yeast TK OH OH O
OH
R O HO OH R O R OH
- CO2
rac O OH
R = -OCH2Ph, -OCH3,
-CH3, -SH, -SEt,
-F, -CN
Scheme 22.17 TK-catalyzed reaction of 2-hydroxy aldehydes with hydroxypyruvate [163].
References
1 W€
ohler, F. and Liebig, J. (1832) Ann. D.I., Petsko, G.A., Kenyon, G.L., McLeish,
Pharm., 3, 249–287. M.J., Jordan, F., and Ringe, D. (2009)
2 Fischer, E. (1882) Justus Liebigs Ann. Biochemistry, 48, 3247–3257.
Chem., 211, 214–232. 14 Bruning, M., Berheide, M., Meyer, D.,
3 Enders, D., Niemeier, O., and Henseler, Golbik, R., Bartunik, H., Liese, A., and
A. (2007) Chem. Rev., 107, 5606–5655. Tittmann, K. (2009) Biochemistry, 48,
4 Neuberg, C. and Ohle, H. (1922) 3258–3268.
Biochem. Z., 128, 610–618. 15 Frank, R.A.W., Leeper, F.J., and Luisi,
5 Kluger, R. and Tittmann, K. (2008) Chem. B.F. (2007) Cell. Mol. Life Sci., 64,
Rev., 108, 1797–1833. 892–905.
6 Neuberg, C. and Hirsch, J., Biochem. Z 16 Kaplun, A., Binshtein, E., Vyazmensky,
(1921) 115, 282–310. M., Steinmetz, A., Barak, Z., Chipman,
7 Hildebrandt, G. and Klavehn, W. (1934) D.M., Tittmann, K., and Shaanan, B.
Manufacture of Laevo-1-phenyl-2- (2008) Nat. Chem. Biol., 4, 113–118.
methylaminopropanol, US 1956950. 17 Shaanan, B. and Chipman, D.M. (2009)
8 Dyda, F., Furey, W., Swaminathan, S., FEBS J., 276, 2447–2453.
Sax, M., Farrenkopf, B., and Jordan, F. 18 Hasson, M.S., Muscate, A., McLeish,
(1993) Biochemistry, 32, 6165–6170. M.J., Polovnikova, L.S., Gerlt, J.A.,
9 Tittmann, K., Golbik, R., Uhlemann, K., Kenyon, G.L., Petsko, G.A., and Ringe, D.
Khailova, L., Patel, M.S., Jordan, F., (1998) Biochemistry, 37, 9918–9930.
Chipman, D.M., Duggleby, R.G., 19 Polovnikova, E.S., McLeish, M.J.,
H€ubner, G., and Schneider, G. (2004) Sergienko, E.A., Burgner, J.T., Anderson,
How Thiamine Works in Enzymes: N.L., Bera, A.K., Jordan, F., Kenyon, G.L.,
Time-Resolved NMR Snapshots of and Hasson, M.S. (2003) Biochemistry, 42,
ThDP-Dependent Enzymes in Action in 1820–1830.
Thiamine: Catalytic Mechanism in Normal 20 Versees, W., Spaepen, S., Vanderleyden,
and Disease States (eds F. Jordan and M.S. J., and Steyaert, J. (2007) FEBS J., 274,
Patel), Marcel Dekker, Inc., New York, 2363–2375.
pp. 57–76. 21 Versees, W., Spaepen, S., Wood, M.D.,
10 Kern, F.D., Kern, G., Neef, H., Tittmann, Leeper, F.J., Vanderleyden, J., and
K., Killenberg-Jabs, M., Wikner, C., Steyaert, J. (2007) J. Biol. Chem., 282,
Schneider, G., and H€ ubner, G. (1997) 35269–35278.
Science, 275, 67–70. 22 Mosbacher, T.G., M€ uller, M., and Schulz,
11 Breslow, R. (1957) J. Am. Chem. Soc., 79, G.E. (2005) FEBS J., 272, 6067–6076.
1762–1763. 23 Kneen, M.M., Pogozheva, I.D., Kenyon,
12 Pei, X.Y., Erixon, K.M., Luisi, B.F., and G.L., and McLeish, M.J. (2005) Biochim.
Leeper, F.J. (2010) Biochemistry, 49, Biophys. Acta, Proteins Proteomics, 1753,
1727–1736. 263–271.
13 Brandt, G.S., Kneen, M.M., Chakraborty, S., 24 Kneen, M.M., Stan, R., Yep, A., Tyler,
Baykal, A.T., Nemeria, N., Yep, A., Ruby, R.P., Saehuan, C., and McLeish, M.J.
References j941
(2011) FEBS J., 278, 1842–1853. For 40 Baykal, A., Chakraborty, S., Dodoo, A.,
KdcA, see Reference [52]. and Jordan, F. (2006) Bioorg. Chem., 34,
25 Knoll, M., M€ uller, M., Pleiss, J., and Pohl, M. 380–393.
(2006) ChemBioChem, 7, 1928–1934. 41 Bornemann, S., Crout, D.H.G., Dalton,
26 M€ uller, M., Gocke, D., and Pohl, M. H., Kren, V., Lobell, M., Dean, G.,
(2009) FEBS J., 276, 2894–2940. Thomson, N., and Turner, M.M. (1996)
27 Pohl, M., Gocke, D., and M€ uller, M. J. Chem. Soc., Perkin Trans. 1, 425–430.
(2008) in Handbook of Green Chemistry, 42 Kren, V., Crout, D.H.G., Dalton, H.,
Green Catalysis, vol. 3, Biocatalysis (ed. Hutchinson, D.W., K€onig, W., Turner,
P.T. Anastas), Wiley-VCH Verlag GmbH, M.M., Dean, G., and Thomson, N. (1993)
Weinheim, pp. 75–114. J. Chem. Soc., Chem. Comm., 341–343.
28 Demir, A.S., Ayhan, P., and Sopaci, S.B. 43 Gunawan, C., Breuer, M., Hauer, B.,
(2007) Clean, 35, 406–412. Rogers, P., and Rosche, B. (2008)
29 Kurutsch, A., Richter, M., Brecht, V., Biotechnol. Lett., 30, 281–286.
Sprenger, G.A., and M€ uller, M. (2009) 44 Gunawan, C., Satianegara, G., Chen,
J. Mol. Catal. B: Enzym., 61, 56–66. A.K., Breuer, M., Hauer, B., Rogers, P.L.,
30 Satianegara, G., Breuer, M., Hauer, B., and Rosche, B. (2007) FEMS Yeast Res., 7,
Rogers, P., and Rosche, B. (2006) 33–39.
Appl. Microbiol. Biotechnol., 70, 170–175. 45 Satianegara, G., Rogers, P.L., and
31 Gocke, D., Graf, T., Brosi, H., Frindi- Rosche, B. (2006) Biotechnol. Bioeng., 94,
Wosch, I., Walter, L., M€ uller, M., and 1189–1195.
Pohl, M. (2009) J. Mol. Cat. B: Enzym., 61, 46 Xue, Y., Qian, C., Wang, Z., Xu, J.-H.,
30–35. Yang, R., and Qi, H. (2010) Appl.
32 Goetz, G., Iwan, P., Hauer, B., Breuer, M., Microbiol. Biotechnol., 85, 517–524.
and Pohl, M. (2001) Biotechnol. Bioeng., 47 Zhang, W., Wang, Z., Li, W., Zhuang, B.,
74, 317–325. and Qi, H. (2008) Appl. Microbiol.
33 Bruhn, H., Pohl, M., Mesch, K., and Kula, Biotechnol., 78, 233–239.
M.R. (1999) Process for obtaining 48 Rosche, B., Breuer, M., Hauer, B., and
acyloins, pyruvate decarboxylases Rogers, P.L. (2004) Biotechnol. Bioeng., 86,
suitable therefore and their production 788–794.
and DNA-sequence of the PDC gene 49 Smallridge, A.J., Trewhella, M.A., and
coding them, US 6004789. Wilkinson, K.A. (2006) Yeast-based
34 Pohl, M., Siegert, P., Mesch, K., Bruhn, process for the production of L-PAC,
H., and Gr€ otzinger, J. (1998) Eur. J. US 7063969.
Biochem., 257, 538–546. 50 Rother, D., Kolter, G., Gerhards, T.,
35 Iwan, P., Goetz, G., Schmitz, S., Hauer, Berthold, C.L., Gauchenova, E., Knoll,
B., Breuer, M., and Pohl, M. (2001) J. Mol. M., Pleiss, J., M€uller, M., Schneider, G.,
Catal. B: Enzym., 11, 387–396. and Pohl, M. (2011) ChemCatChem, 3,
36 Breuer, M., Hauer, B., Mesch, K., Iding, 1587–1596.
H., Goetz, G., Pohl, M., and Kula, M.R. 51 Gocke, D., Nguyen, C.L., Pohl, M.,
(1999) Preparation of enantiomerically Stillger, T., Walter, L., and M€uller, M.
pure phenylacetycarbinol derivatives, DE (2007) Adv. Synth. Catal., 349,
19736104. 1425–1435.
37 Meyer, D., Walter, L., Kolter, G., Pohl, M., 52 Berthold, C.L., Gocke, D., Wood, M.D.,
M€ uller, M., and Tittmann, K. (2011) J. Leeper, F.J., Pohl, M., and Schneider, G.
Am. Chem. Soc., 133, 3609–3616. (2007) Acta Crystallogr., Sect. D, 63,
38 Siegert, P., McLeish, M.J., Baumann, M., 1217–1224.
Iding, H., Kneen, M.M., Kenyon, G.L., 53 Priyadarshi, A., Kim, E.E., and Hwang,
and Pohl, M. (2005) Protein Eng. Des. K.Y. (2009) Biochem. Biophys. Res.
Select., 18, 345–357. Commun., 388, 748–751.
39 Rosche, B., Sandford, V., Breuer, M., 54 Bhasin, M., Billinsky, J.L., and Palmer,
Hauer, B., and Rogers, P. (2001) Appl. D.R. (2003) Biochemistry, 42,
942
55 Vinogradov, V., Vyazmensky, M., Engel, 70 Domınguez de Marıa, P., Pohl, M.,
S., Belenky, I., Kaplun, A., Kryukov, O., Gocke, D., Gr€oger, H., Trauthwein, H.,
Barak, Z., and Chipman, D.M. (2006) Stillger, T., and M€ uller, M. (2007) Eur. J.
Biochim. Biophys. Acta, 1760, 356–363. Org. Chem., 2940–2944.
56 Engel, S., Vyazmensky, M., Berkovich, 71 Domınguez de Marıa, P., Stillger, T., Pohl,
D., Barak, Z., Merchuk, J., and Chipman, M., Wallert, S., Drauz, K., Gr€oger, H.,
D.M. (2005) Biotechnol. Bioeng., 89, Trauthwein, H., and Liese, A. (2006)
733–740. J. Mol. Catal. B: Enzym., 38, 43–47.
57 Engel, S., Vyazmensky, M., Berkovich, 72 Stillger, T., Pohl, M., Wandrey, C., and
D., Barak, Z., and Chipman, D.M. (2004) Liese, A. (2006) Org. Proc. Res. Dev., 10,
Biotechnol. Bioeng., 88, 825–831. 1172–1177.
58 Engel, S., Vyazmensky, M., Vinogradov, 73 Dr€ager, G., Kiss, C., Kunz, U., and
M., Berkovich, D., Bar-Ilan, A., Qimron, Kirschning, A. (2007) Org. Biomol. Chem.,
U., Rosiansky, Y., Barak, Z., and 5, 3657–3664.
Chipman, D.M. (2004) J. Biol. Chem., 279, 74 Kurlemann, N., Lara, M., Pohl, M.,
24803–24812. Kroutil, W., and Liese, A. (2009) J. Mol.
59 Engel, S., Vyazmensky, M., Geresh, S., Catal. B: Enzym., 61, 111–116.
Barak, Z., and Chipman, D.M. (2003) 75 Kurlemann, N. and Liese, A. (2004)
Biotechnol. Bioeng., 83, 833–840. Tetrahedron: Asymmetry, 15,
60 Chipman, D.M., Barak, Z., Shaanan, B., 2955–2958.
Vyazmensky, M., Binshtein, E., Belenky, 76 Kihumbu, D., Stillger, T., Hummel, W.,
I., Temam, V., Steinmetz, A., Golbik, R., and Liese, A. (2002) Tetrahedron:
and Tittmann, K. (2009) J. Mol. Catal. B: Asymmetry, 13, 1069–1072.
Enzym., 61, 50–55. 77 Villela, M., Stillger, T., M€ uller, M., Liese,
61 Gonzalez, B. and Vicuna, R. (1989) J. A., and Wandrey, C. (2003) Angew. Chem.
Bacteriol., 171, 2401–2405. Int. Ed., 42, 2993–2996.
62 Janzen, E., M€ uller, M., Kolter-Jung, D., 78 Shanmuganathan, S., Natalia, D., van
Kneen, M.M., McLeish, M.J., and den Wittenboer, A., Kohlmann, C.,
Pohl, M. (2006) Bioorg. Chem., 34, Greiner, L., and Dominguez de Marıa, P.
345–361. (2010) Green Chem., 12, 2240–2245.
63 Hildebrand, F., K€ uhl, S., Pohl, M., 79 Kokova, M., Zavrel, M., Tittmann, K.,
Vasic-Racki, D., M€ uller, M., Wandrey, C., Spiess, A.C., and Pohl, M. (2009) J. Mol.
and L€ utz, S. (2007) Biotechnol. Bioeng., 96, Cat. B: Enzym., 61, 73–79.
835–843. 80 Wilcocks, R. and Ward, O.P. (1992)
64 K€uhl, S., Zehentgruber, D., Pohl, M., Biotechnol. Bioeng., 39, 1058–1063.
M€ uller, M., and L€utz, S. (2007) Chem. Eng. 81 Wilcocks, R., Ward, O.P., Collins, S.,
Sci., 62, 5201–5205. Dewdney, N.J., Hong, Y., and Prosen, E.
65 Demir, A.S., Ayhan, P., Igdir, A.C., and (1992) Appl. Environ. Microbiol., 58,
Duygu, A.N. (2004) Tetrahedron, 60, 1699–1704.
6509–6512. 82 Iding, H., D€ unnwald, T., Greiner, L.,
66 Demir, A.S., Sesenoglu, O., Liese, A., M€ uller, M., Siegert, P.,
D€ unkelmann, P., and M€ uller, M. (2003) Gr€otzinger, J., Demir, A.S., and
Org. Lett., 5, 2047–2050. Pohl, M. (2000) Chem. Eur. J., 6,
67 Demir, A.S., Pohl, M., Janzen, E., and 1483–1495.
M€ uller, M. (2001) J. Chem. Soc., Perkin 83 D€unnwald, T., Demir, A.S., Siegert, P.,
Trans 1, 633–635. Pohl, M., and M€ uller, M. (2000) Eur. J.
68 Chakraborty, S., Nemeria, N., Yep, A., Org. Chem., 2161–2170.
McLeish, M.J., Kenyon, G.L., and Jordan, 84 D€unnwald, T. and M€ uller, M. (2000)
F. (2008) Biochemistry, 47, 3800–3809. J. Org. Chem., 65, 8608–8612.
69 Domınguez de Marıa, P., Stillger, T., Pohl, 85 Lingen, B., Kolter-Jung, D.,
M., Kiesel, M., Liese, A., Gr€oger, H., and D€unkelmann, P., Feldmann, R.,
Trauthwein, H. (2008) Adv. Synth. Catal., Gr€otzinger, J., Pohl, M., and M€ uller, M.
350, 165–173. (2003) ChemBioChem, 4, 721–726.
References j943
86 Lingen, B., Gr€
otzinger, J., Kolter, D., Kula, 101 Cosp, A., Dresen, C., Pohl, M., Walter, L.,
M.R., and Pohl, M. (2002) Protein Eng., R€ohr, C., and M€ uller, M. (2008) Adv.
15, 585–593. Synth. Catal., 350, 759–771.
87 Gocke, D., Walter, L., Gauchenova, E., 102 Franke, D., Sprenger, G.A., and M€ uller,
Kolter, G., Knoll, M., Berthold, C.L., M. (2001) Angew. Chem. Int. Ed., 40,
Schneider, G., Pleiss, J., M€ uller, M., and 555–557.
Pohl, M. (2008) ChemBioChem, 9, 103 Vinogradov, M., Kaplun, A., Vyazmensky,
406–412. M., Engel, S., Golbik, R., Tittmann, K.,
88 Hilterhaus, L., Minow, B., M€ uller, J., Uhlemann, K., Meshalkina, L., Barak, Z.,
Berheide, M., Quitmann, H., Katzer, M., H€ ubner, G., and Chipman, D.M. (2005)
Thum, O., Antranikian, G., Zeng, A.P., Anal. Biochem., 342, 126–133.
and Liese, A. (2008) Bioprocess Biosyst. 104 Schneider, G. and Lindqvist, Y. (1998)
Eng., 31, 163–171. Biochim. Biophys. Acta, 1385, 387–398.
89 Berheide, M., Peper, S., Kara, S., Long, 105 Wohlgemuth, R. (2009) J. Mol. Catal. B:
W.S., Schenkel, S., Pohl, M., Niemeyer, Enzym., 61, 23–29.
B., and Liese, A. (2010) Biotechnol. 106 Hecquet, L., Demuynck, C., Schneider,
Bioeng., 106, 18–26. G., and Bolte, J. (2001) J. Mol. Catal. B:
90 Sch€utz, A., Golbik, R., Tittmann, K., Enzym., 11, 771–776.
Svergun, D.I., Koch, M.H., H€ ubner, G., 107 Asztalos, P., Parthier, C., Golbik, R.,
and K€onig, S. (2003) Eur. J. Biochem., 270, Kleinschmidt, M., H€ ubner, G., Weiss,
2322–2331. M.S., Friedemann, R., Wille, G., and
91 Koga, J., Adachi, T., and Hidaka, H. Tittmann, K. (2007) Biochemistry, 46,
(1992) J. Biol. Chem., 267, 15823–15828. 12037–12052.
92 Sanchez-Gonzalez, M. and Rosazza, 108 Smith, M.E.B., Hibbert, E.G., Jones, A.B.,
J.P.N. (2003) Adv. Synth. Catal., 345, Dalby, P.A., and Hailes, H.C. (2008) Adv.
819–824. Synth. Catal., 350, 2631–2638.
93 Guo, Z., Goswami, A., Nanduri, V.B., and 109 Hibbert, E.G., Tarik, S.A., Mark, E.B.S.B.,
Patel, R.N. (2001) Tetrahedron: Costelloe, S.J., Ward, J.M., Hailes, H.C.,
Asymmetry, 12, 571–577. and Dalby, P.A. (2008) J. Biotechnol., 134,
94 Guo, Z., Goswami, A., Mirfakhrae, K.D., 240–245.
and Patel, R.N. (1999) Tetrahedron: 110 Cazares, A., Galman, J.L., Crago, L.G.,
Asymmetry, 10, 4667–4675. Smith, M.E.B., Strafford, J., Rios-Solis, L.,
95 Yep, A., Kenyon, G.L., and McLeish, M.J. Lye, G.J., Dalby, P.A., and Hailes, H.C.
(2006) Bioorg. Chem., 34, 325–336. (2010) Org. Biomol. Chem., 8, 1301–1309.
96 Smit, B.A., Vlieg, J., Engels, W.J.M., 111 Smith, M.E.B., Smithies, K., Senussi, T.,
Meijer, L., Wouters, J.T.M., and Smit, G. Dalby, P.A., and Hailes, H.C. (2006) Eur.
(2005) Appl. Environ. Microbiol., 71, J. Org. Chem., 1121–1123.
303–311. 112 Galman, J.L., Steadman, D., Bacon, S.,
97 Vuralhan, Z., Luttik, M.A., Tai, S.L., Boer, Morris, P., Smith, M.E.B., Ward, J.M.,
V.M., Morais, M.A., Schipper, D., Dalby, P.A., and Hailes, H.C. (2010)
Almering, M.J., Kotter, P., Dickinson, Chem. Commun., 7608–7610.
J.R., Daran, J.M., and Pronk, J.T. (2005) 113 Sprenger, G.A., Sch€orken, U., Wiegert, T.,
Appl. Environ. Microbiol., 71, Grolle, S., deGraaf, A.A., Taylor, S.V.,
3276–3284. Begley, T.P., Bringer-Meyer, S., and
98 Spaepen, S., Versees, W., Gocke, D., Pohl, Sahm, H. (1997) Proc. Natl. Acad. Sci.
M., Steyaert, J., and Vanderleyden, J. USA, 94, 12857–12862.
(2007) J. Bacteriol., 189, 7626–7633. 114 Querol, J., Grosdemange-Billiard, C.,
99 Mikolajek, R., Spiess, A.C., Pohl, M., Rohmer, M., Boronat, A., and
Lamare, S., and B€ uchs, J. (2007) Imperial, S. (2002) Tetrahedron Lett., 43,
ChemBioChem, 8, 1063–1070. 8265–8268.
100 Mikolajek, R., Spiess, A.C., Pohl, M., and 115 Demir, A.S., D€ unnwald, T., Iding, H.,
B€uchs, J. (2009) Biotechnol. Prog., 25, Pohl, M., and M€ uller, M. (1999)
944
116 Demir, A.S., Sesenoglu, O., Eren, E., 132 Williamson, N.R., Simonsen, H.T.,
Hosrik, B., Pohl, M., Janzen, E., Kolter, Ahmed, R.A.A., Goldet, G., Slater, H.,
D., Feldmann, R., D€ unkelmann, P., and Woodley, L., Leeper, F.J., and Salmond,
M€ uller, M. (2002) Adv. Synth. Catal., 344, G.P.C. (2005) Mol. Microbiol., 56,
96–103. 971–989.
117 Prosen, E. and Ward, O.P. (1994) J. Ind. 133 Dresen, C., Richter, M., Pohl, M., L€udeke,
Microbiol., 13, 287–291. S., and M€ uller, M. (2010) Angew. Chem.
118 Iding, H., Siegert, P., Mesch, K., and Int. Ed., 49, 6600–6603.
Pohl, M. (1998) Biochim. Biophys. Acta, 134 Chen, H.W., Guo, Z.H., and Liu, H.-w.
1385, 307–322. (1998) J. Am. Chem. Soc., 120,
119 D€unkelmann, P., Kolter-Jung, D., 11796–11797.
Nitsche, A., Demir, A.S., Siegert, P., 135 Mansoorabadi, S.O., Thibodeaux, C.J.,
Lingen, B., Baumann, M., Pohl, M., and and Liu, H.-w. (2007) J. Org. Chem., 72,
M€ uller, M. (2002) J. Am. Chem. Soc., 124, 6329–6342.
12084–12085. 136 Lehwald, P., Richter, M., R€ohr, C., Liu,
120 Stanier, R.Y. and Ornston, L.N. (1973) H.-w., and M€ uller, M. (2010) Angew.
Adv. Microb. Physiol., 9, 89–151. Chem. Int. Ed., 49, 2389–2392.
121 Hischer, T., Gocke, D., Fernandez, M., 137 Giovannini, P.P., Pedrini, P., Venturi, V.,
Hoyos, P., Alcantara, A.R., Sinisterra, Fantin, G., and Medici, A. (2010) J. Mol.
J.V., Hartmeier, W., and Ansorge- Cat. B: Enzym., 64, 113–117.
Schumacher, M.B. (2005) Tetrahedron, 61, 138 Guo, J.T. and Frost, J.W. (2002) J. Am.
7378–7383. Chem. Soc., 124, 528–529.
122 Ansorge-Schumacher, M.B., Greiner, L., 139 Guo, J.T. and Frost, J.W. (2002) J. Am.
Schroeper, F., Mirtschin, S., and Chem. Soc., 124, 10642–10643.
Hischer, T. (2006) Biotechnol. J., 1, 140 Guo, J.T. and Frost, J.W. (2004) Org. Lett.,
564–568. 6, 1585–1588.
123 Mikolajek, R., Spiess, A.C., and B€ uchs, J. 141 Teles, J.H., Melder, J.P., Ebel, K.,
(2007) J. Biotechnol., 129, 723–725. Schneider, R., Gehrer, E., Harder, W.,
124 Zavrel, M., Schmidt, T., Michalik, C., Brode, S., Enders, D., Breuer, K., and
Ansorge-Schumacher, M., Raabe, G. (1996) Helv. Chim. Acta, 79,
Marquardt, W., B€ uchs, J., and 61–83.
Spiess, A.C. (2008) Biotechnol. Bioeng., 142 Krampitz, L.O., Greull, G., Miller, C.S.,
101, 27–38. Bicking, J.B., Skeggs, H.R., and Sprague,
125 Pohl, M., Lingen, B., and J.M. (1958) J. Am. Chem. Soc., 80,
M€ uller, M. (2002) Chem. Eur. J., 8, 5893–5894.
5289–5295. 143 Krampitz, L.O. (1969) Annu. Rev.
126 Gocke, D., Graf, T., Brosi, H., Frindi- Biochem., 38, 213–240.
Wosch, I., Walter, L., M€ uller, M., and 144 Krampitz, L.O. (1982) Ann. N. Y. Acad.
Pohl, M. (2009) J. Mol. Cat. B: Enzym., 61, Sci., 378, 1–6.
30–35. 145 Kubasik, N.P., Richert, D.A., Bloom, R.J.,
127 Stetter, H. (1976) Angew. Chem. Int. Ed. Hsu, R.Y., and Westerfeld, W.W. (1972)
Engl., 11, 639–647. Biochemistry, 11, 2225–2229.
128 Jiang, M., Cao, Y., Guo, Z.F., Chen, M.J., 146 Schlossberg, M.A., Bloom, R.J., Richert,
Chen, X.L., and Guo, Z.H. (2007) D.A., and Westerfeld, W.W. (1970)
Biochemistry, 46, 10979–10989. Biochemistry, 9, 1148–1153.
129 Simantiras, M. and Leistner, E. (1991) Z. 147 Saito, T., Tuboi, S., Nichemura, Y., and
Naturforsch. C, 46, 364–370. Kikucki, G. (1971) J. Biochem., 69,
130 Emmons, G.T., Campbell, I.M., and 265–273.
Bentley, R. (1985) Biochem. Biophys. Res. 148 Bornemann, S., Crout, D.H.G., Dalton,
Commun., 131, 956–960. H., Hutchinson, D.W., Dean, G.,
131 Jiang, M., Chen, X.L., Guo, Z.F., Cao, Y., Thomson, N., and Turner, M.M.
Chen, M.J., and Guo, Z.H. (2008) (1993) J. Chem. Soc., Perkin Trans. 1,
Biochemistry, 47, 3426–3434. 309–311.
References j945
149 Lobell, M. and Crout, D.H.G. (1996) 158 Yanase, H., Okuda, M., Kita, K., Sato, Y.,
J. Chem. Soc., Perkin Trans. 1, Shibata, K., Sakai, Y., and Kato, N. (1995)
1577–1581. Appl. Microbiol. Biotechnol., 43, 228–234.
150 Singer, T.P. and Pensky, J. (1952) J. Biol. 159 Gupta, N.K. and Vennesland, B. (1964) J.
Chem, 196, 375–388. Biol. Chem., 239, 3787–3789.
151 Crout, D.H.G., Littlechild, J., and Morrey, 160 Gupta, N.K. and Vennesland, B. (1966)
S.M. (1986) J. Chem. Soc., Perkin Trans. 1, Arch. Biochem. Biophys., 113, 255–264.
105–108. 161 Cromartie, T.H. and Walsh, C.T. (1976)
152 van Dijken, J.P., Harder, W., Beardsmore, J. Biol. Chem., 251, 329–333.
A.J., and Quayle, J.R. (1978) FEMS 162 Chang, Y.Y., Wang, A.Y., and
Microbiol. Lett., 4, 97–102. Cronan, J.E., Jr. (1993) J. Biol. Chem., 268,
153 Kato, N., Higuchi, T., Sakazawa, C., 3911–3919.
Nishizawa, T., Tani, Y., and Yamada, H. 163 Effenberger, F., Null, V., and Ziegler,
(1982) Biochim. Biophys. Acta, 715, T. (1992) Tetrahedron Lett., 33, 5157–5160.
143–150. 164 Kobori, Y., Myles, D.C., and
154 Waites, M.J. and Quayle, J.R. (1983) Whitesides, G.M. (1992) J. Org. Chem.,
J. Gen. Microbiol., 129, 935–944. 57, 5899–5907.
155 Goodman, J.M. (1985) J. Biol. Chem., 260, 165 Demuynck, C., Bolte, J., Hecquet, L., and
7108–7113. Dalmas, V. (1991) Tetrahedron Lett., 32,
156 Bystrykh, L.V., De Koning, W., and 5085–5088.
Harder, W. (1990) Methods Enzymol., 188, 166 Sprenger, G.A., Sch€orken, U., Sprenger,
435–445. G., and Sahm, H. (1995) Eur. J. Biochem.,
157 Ro, Y.T., Eom, C.Y., Song, T., Cho, J.W., 230, 525–532.
and Kim, Y.M. (1997) J. Bacteriol., 179, 167 Sprenger, G.A. and Pohl, M. (1999) J. Mol.
6041–6047. Catal. B: Enzym., 6, 145–159.
j947
23
Cleavage and Formation of Cyanohydrins
Mandana Gruber-Khadjawi, Martin H. Fechter, and Herfried Griengl
23.1
Introduction
In the last 30 years enantiopure cyanohydrins (a-hydroxynitriles) have become a

versatile source for the synthesis of various chiral building blocks. Diverse methods
for the enantioselective synthesis of cyanohydrins have been published and
reviewed [1–3].
Besides enzyme catalyzed methods [4–7] hydrocyanation or silylcyanation of
aldehydes or ketones controlled by chiral metal complexes or cyclic dipeptides, as
well as diastereoselective hydrocyanation of chiral carbonyl compounds, have been
applied [2, 8–14].
However, the most advantageous preparations of cyanohydrins, with respect to the
enantioselectivities obtained, are the enzymatically controlled approaches discussed
in the present chapter. Two common enzyme systems are described and reviewed:
first, esterases or lipases, which have been employed for the resolution of racemic
cyanohydrins or alkoxynitriles, and second hydroxynitrile lyases (HNLs), which
catalyze the reversible formation of cyanohydrins (Scheme 23.1), from HCN and
aldehydes or ketones.
Cyanohydrins are used by many plants for defense by release of HCN and also as
storage compounds [12, 13]. Storage compounds are cyanohydrins where the hydroxy
function is glycosylated to a carbohydrate or protected as a fatty acid ester. The plant
defense mechanism in the case of sugar compounds is a two-step reaction. Initially, a
glycosidase liberates the cyanohydrin moiety, which is cleaved either spontaneously
by base catalysis or enzymatically by the action of hydroxynitrile lyases to release the
corresponding carbonyl compound and HCN [14].
The application of an HNL was the subject of one of the earliest reports in the field
of biocatalysis, namely, the synthesis of mandelonitrile from benzaldehyde and
hydrocyanic acid using a crude enzyme preparation obtained from almonds (termed
emulsin) [15]. However, little attention was paid to this discovery [16–18] until the
1960s, when this enzyme (EC 4.1.2.10) was isolated, characterized [19, 20], and used
for the preparation of enantiomerically enriched (R)-cyanohydrins from aromatic
and Oliver May.
j 23 Cleavage and Formation of Cyanohydrins
948
O OH
CN
+ HCN
R1 R2 R1
R2
Scheme 23.1 Cyanohydrin formation: R1 ¼ alkyl, cycloalkyl, aryl, heteroaryl; R2 ¼ H, alkyl.
and aliphatic aldehydes [21–23]. The first (S)-selective hydroxynitrile lyase was
detected in 1960 in millet seedlings [24–27].
Today, a broad spectrum of both (R)- and (S)-selective hydroxynitrile lyases is
available. A wide range of substrates is accepted and by overexpression the enzymes
can be obtained in large quantities. This also made possible an application for
syntheses on an industrial scale.
Within this chapter the literature is covered up to early 2009 and 323 references
are cited.
23.2
Hydroxynitrile Lyases Commonly Used for Preparative Application
About 3000 plant species and a few non-plant sources exhibit the ability to release
HCN from their tissues, a process called cyanogenesis [28–30]. Hydroxynitrile
lyases, also known as oxynitrilases, are the enzymes that catalyze the decomposition
of cyanohydrins and the reverse reaction, the stereoselective addition of hydrocyanic
acid to aldehydes and ketones [31–38]. Nearly a dozen of these enzymes have been
isolated, purified, and characterized from cyanogenic plants [39]. The main plant
families are Rosaceae (e.g., Prunus amygdalus, PaHNL), Poaceae (e.g., Sorghum
bicolor, SbHNL), Euphorbiaceae (e.g., Hevea brasiliensis, HbHNL, and Manihot
esculenta, MeHNL), and Linaceae (e.g., Linum usitatissimum, LuHNL). HNLs are
enantiocomplementary enzymes as (R)- and (S)-selective HNLs are found in
nature [40]. The enzymes can be classified according to the different enantioselec-
tivities as (R)- and (S)-HNLs or due to their biochemical specification as FAD- or
non-FAD-containing enzymes [41]. While FAD-containing HNLs were exclusively
found in Rosaceae, the non-FAD-containing enzymes are more heterogeneous
regarding protein structure and origin. Table 23.1 outlines the properties of a
selection of HNLs.
23.2.1
(R)-Selective HNLs
The hydroxynitrile lyase (EC 4.1.2.10) from Rosaceae (e.g., Prunus sp.) contains
the cofactor FAD. However, the latter is not involved in redox reactions. Instead, it
seems to have a structure-stabilizing effect, and its presence might be explained
on evolutionary grounds [42–44]. The enzymes are to a certain extent highly
glycosylated single chain proteins with (R)-mandelonitrile as their natural sub-
strate [41, 45].
23.2 Hydroxynitrile Lyases Commonly Used for Preparative Application j949
Table 23.1 Selected hydroxynitrile lyases (HNLs) for organic synthesis.
Source Enzyme Natural Substrate Stereo

availability substrate acceptance for selectivity
synthesisa)
Prunus Bitter almond (R)-Mandelonitrile All R1 and R2 (R)

amygdalus overexpression
Linum Flax seedlings Acetone cyanohydrin All R1 and R2 (R)
usitatissimum overexpression (R)-2-butanone
cyanohydrin
Arabidopsis Mouse ear cress All R1 and R2 (R)
thaliana overexpression
Sorghum bicolor Millet seedlings (S)-4-Hydroxyman- Aromatic aldehydes (S)
del-onitrile methyl ketones
Hevea Rubber tree leaves Acetone All R1 and R2 (S)
brasiliensis overexpression cyanohydrin (R)-2-
butanone
cyanohydrin
Manihot Manioc leaves Acetone All R1 and R2 (S)
esculenta overexpression cyanohydrin
a) See artwork for definition of R1 and R2.
Apple, apricot, cherry, and plum meals were prepared from the seeds or kernels of
mature garden fruits. These preparations and almond meal were used as the source
of (R)-HNL for the synthesis of cyanohydrins from aliphatic, unsaturated, aromatic,
and heteroaromatic aldehydes and ketones [46–49]. Apple seed meal, the most
favorable of the crude enzyme preparations, accepts sterically hindered aldehydes
(e.g., pivalaldehyde) as substrates, leading to (R)-cyanohydrins with high enantio-
meric purity (e.e.s >90%) [50]. Subsequently, the hydroxynitrile lyase from apple
meal was found to also accept methyl ketones as substrates and when a direct
comparison with almond meal was carried out the apple enzyme gave a slightly
higher e.e. [51]. Recently, a new (R)-hydroxynitrile lyase was reported for catalysis of
the asymmetric synthesis of d,e-unsaturated cyanohydrins with yields 70% and e.e.s
up to 98% [52]. This new (R)-HNL from seeds of the ripened fruit Prunus armeniaca
(shakarpara apricot) was also reported to be active for sterically demanding aromatic
aldehydes like 3-phenoxybenzaldehyde, leading to good yields and very good selectiv-
ities [53]. Li and coworkers compared the (R)-HNL activity of peach and loquat
preparations with that of almond meal. The enzyme extracted from loquat had a
rather narrow substrate range, being restricted to aromatic and heteroaromatic
aldehydes, and gave lower e.e.s than those obtained with almond HNL. In contrast,
peach meal had a substrate range similar to that of almond meal and in some cases
gave products with superior e.e.s. Thus, cinnamaldehyde was converted into its
(R)-cyanohydrin with 69% e.e. by peach meal, while under the same conditions
almond meal gave a product with only 51% e.e. [54]. Enzymes from natural sources
as well as enzyme preparations from different batches contain different propor-
tions of isoenzymes. This might be the main reason for the varying enantioselec-
950
tivity and conversion rates observed for the same substrate in different process-
es [39, 55]. Generally, HNL from almond, which is the most widely applied enzyme
for the synthesis of (R)-cyanohydrins, shows low substrate specificity combined
with high enantioselectivity and is therefore an ideal biocatalyst. Nowadays, PaHNL
(Prunus amygdalus HNL) is available not only from natural sources like almonds but
also from a fermentation process that involves the gene for the PaHNL isoenzyme 5
cloned into and overexpressed in the methylotrophic yeast Pichia pastoris [56, 57]. A
big step forward, towards further applications of the Prunus amygdalus HNL, was
achieved by the Kratky group in elucidating the crystal structure of this enzyme [58].
Both the knowledge about crystal structure and the expression of recombinant
PaHNL opened the way for the preparation of optimal muteins for specific
applications by enzyme engineering [59].
Hydroxynitrile lyase activity has been found in crude enzyme preparations from
the leaves of mamey (Pouteria sapota), cherry (Prunus avium), plum (Prunus domes-
tica), peach (Prunus persica), capulin (Prunus serotina), and the seeds of quince
(Cydonia oblonga) where e.e.s of synthesized (R)-mandelonitrile were over 90%. For
melon seeds (Cucumis melo) the e.e. was only 48%, and for sugar-apple and cherimoya
seeds it was only 18% and 16% [(S)-enantiomer], respectively. In the case of catalysis
with leaf or seed extracts of sweet acacia, bonete, pomegranate, clover, and canistel,
the product of the addition of HCN to benzaldehyde was racemic [60]. High
enantiomeric purities (e.e. up to 98%) were achieved with the defatted meal from
capulin seeds and leaves as well as mamey leaves as catalysts for the synthesis of
cyanohydrins of aromatic and aliphatic aldehydes [61]. There were differences in the
reactivities and enantioselectivities of both meals. The enzyme from mamey showed
higher enantioselectivities [62–64]. The HNL from mamey catalyzed the addition of
cyanide to imines, prepared from substituted aromatic aldehydes and aniline, to yield
a-amino nitriles with moderate enantioselectivity (23% e.e.) [65]. In a related manner,
optically active a-amino nitriles were attained by the addition of acetone cyanohydrin
to chiral imines. The reaction was catalyzed by the (R)-hydroxynitrile lyase in almond
meal with moderate yields and selectivities [66].
Presently, (S)- and (R)-mandelic acids derived from cyanohydrin precursors and
subsequent acidic hydrolysis are produced on an industrial scale. The chiral acids are
mainly used for racemate resolution. Both (R)-2-chloromandelic acid (250 g l1
day1, 95% e.e.) [67] and (R)-2-hydroxy-4-phenylbutyronitrile are further large-scale
products of improved HNLs [68].
Besides the FAD-dependent HNLs, a (R)-selective hydroxynitrile lyase from
Linum usitatissimum (flax) (LuHNL) has been recognized and isolated [69–71].
Using this enzyme, it is possible to synthesize (R)-butan-2-one cyanohydrin with an
e.e. of up to 88%, this being noteworthy due to the relatively small steric difference
between the methyl and ethyl groups in the neighborhood of the carbonyl functional
group of the substrate, 2-butanone. This LuHNL (Linum usitatissimum HNL) (EC
4.1.2.37) has a completely different substrate specificity from that of the Prunus
enzyme. It catalyses the addition of HCN to various aliphatic ketones and aldehydes,
while aromatic ketones were reported to be not converted [72]. More recently,
Roberge and coworkers reported the conversion of aromatic ketones into optically
23.2 Hydroxynitrile Lyases Commonly Used for Preparative Application j951
active cyanohydrins by LuHNL with inverted stereoselectivity [(S)-products were
obtained] [73].
Initially, cloning of LuHNL was hampered by low expression levels of the
recombinant enzyme in Escherichia coli. To overcome this problem Wajant and
coworkers cloned the LuHNL-cDNA into Pichia pastoris for overexpression. With
aromatic aldehydes and this recombinant HNL the conversion into cyanohydrins did
not come to completion and the enantioselectivity was low [74]. Kula and coworkers
expressed an active enzyme in E. coli as an N-terminal hexa-histidine fusion protein,
allowing the purification of homogeneous protein in one step. The formation of
inclusion bodies was reduced by using a thioreductase deficient E. coli strain as the
host. Under these conditions, recombinant LuHNL was obtained with a specific
activity of 76 U mg1 [75].
In 1995, Wajant described the purification of a novel (R)-HNL from the fern
Phlebodium aureum, which contains no FAD. This PhaHNL has no properties in
common with the flavoprotein lyases from Rosaceae, except that it has the same
natural substrate, (R)-mandelonitrile, which is released from prunasin in Prunus
species and from vicianin in Phlebodium aureum. PhaHNL is a multimer of 20 kDa
subunits and is suitable for the synthesis of (R)-cyanohydrins in organic media [76].
In 2006, Han and coworkers reported a new (R)-HNL found in the defatted seed
meal of vetch (Vicia sativa a Fabaceae). Under micro-aqueous conditions a quanti-
tative yield of mandelonitrile with 99% e.e. was achieved. With some other aromatic
aldehydes 52–97% yield and 3–97% e.e. were obtained, while an aliphatic aldehyde
tested was not converted [77].
More recently a (R)-selective HNL was found in Arabidopsis thaliana (AtHNL),
which belongs to the a/b-hydrolase fold superfamily [78]. Interestingly, the
structure of this HNL is very similar to the (S)-selective MeHNL (Manihot
esculenta HNL) and HbHNL (Hevea brasiliensis HNL) (K. Gruber, unpublished
results), it also shows a highly similar substrate range and stability as these
enzymes but the reversed enantioselectivity [79]. Inhibition studies regarding
acetate and the inverted enantioselectivity of AtHNL (Arabidopsis thaliana HNL)
compared to the (S)-selective HNLs from Hevea brasiliensis and Manihot esculenta
indicate a different mechanism of substrate binding. Notably, this enzyme was not
found by using the traditional approach of screening tissue extracts but by
following a sequence-based approach of database screening for sequences similar
to known enzymes. Among the promising sequences, which were cloned from
genomic DNA or mRNA and expressed to corresponding proteins in a heterol-
ogous host, AtHNL showed the desired activity.
23.2.2
(S)-Selective HNLs
Sorghum bicolor HNL (SbHNL) (EC 4.1.2.11) was purified from seedlings of Sorghum
vulgare [25]. HNL from Sorghum bicolor was the first (S)-HNL used in an organic
solvent for the preparation of (S)-cyanohydrins. The natural substrate is (S)-4-
hydroxymandelonitrile. Its major drawback is the limited substrate tolerance – only
952
aromatic and heteroaromatic aldehydes are accepted, while aliphatic aldehydes or

ketones are not converted. For a wide range of 3- and 4-substituted aromatic
aldehydes excellent selectivities were obtained [80, 81]. Overexpression of this
enzyme has not been successful yet due to the complex posttranslational processing
of the native enzyme.
Hevea brasiliensis HNL (HbHNL) (EC 4.1.2.39) was isolated from the leaves of the
tropical rubber tree. It is an unglycosylated protein with 257 amino acids and a
subunit molecular mass of 29.2 kDa [82, 83]. The gene has been cloned and over-
expressed in several microorganisms and the enzyme was isolated in an active form,
enabling the production of recombinant purified biocatalyst. The most efficient
expression system was developed using methanol-inducible Pichia pastoris. The
protein is produced in high levels and exhibits high specific activity (40 U mg1).
High-cell-density cultivation yields more than 20 g per liter culture volume pure
HNL [84]. The natural substrate of HbHNL is acetone cyanohydrin. Many different
starting materials such as aromatic, heteroaromatic, aliphatic, unsaturated and
branched aldehydes, and even unusual substrates such as formylferrocene (99%
e.e.) as well as methyl and heterocyclic ketones have been transformed into the
corresponding cyanohydrins [85–91]. The structure of HbHNL has been deter-
mined [92–96]. The enzyme was found to contain a large b-sheet that is surrounded
by a-helices and a cap region on both sides. The active site is buried deep inside the
protein and connected to the surface by a narrow channel. Here again, knowledge of
the crystal structure and the expression of recombinant HbHNL opened the way for
the preparation of optimal muteins for specific applications by enzyme engineer-
ing [97]. Furthermore, the enzyme has been used encapsulated in a sol–gel
matrix [98], as crosslinked enzyme aggregates (CLEAs) [99], CLECs (crosslinked
enzyme colloids), and in ionic liquids [100].
Presently, the (S)-cyanohydrin of 3-phenoxybenzaldehyde used as an intermediate
for various pyrethroid type insecticides is the largest commercial HNL product. This
reaction is catalyzed by overexpressed (S)-HNL from Hevea brasiliensis and the
cyanohydrin is produced on the hundred ton per year scale. More recently, a two-step
chemoenzymatic synthesis of (R)-2-amino-1-(2-furyl)ethanol was scaled up to kilo-
gram scale. The asymmetric center was generated by a cyanohydrin reaction
catalyzed by HNL from Hevea brasiliensis [101].
Similar discoveries were published for the (S)-HNL from Manihot esculenta (EC
4.1.2.39) [102, 103], which is highly homologous (77% sequence identity) to HbHNL.
MeHNL has also been successfully overexpressed in microorganisms [104–107].
Several different carbonyl substrates could be converted into the corresponding
cyanohydrins catalyzed by MeHNL. While aldehydes yield cyanohydrins with high
enantioselectivities the ketones provide the products less selectively. The best results
were obtained in organic solvent with the biocatalyst supported on nitrocellu-
lose [108–111]. MeHNL is striking for its stability under various reaction conditions
as well as for its long time storage [31, 112]. The stability is presumed to be due to
formation of a tetramer in solution [74]. The active site includes the catalytic triad of
serine, histidine, and asparagine that is characteristic for the a/b-hydrolase family
members.
23.3 Hydroxynitrile Lyase Catalyzed Addition of HCN to Aldehydes j953
The search for new hydroxynitrile lyases is ongoing. Asano and coworkers
screened a total of 163 plant species among 74 families for HNL activity for the
stereoselective synthesis of mandelonitrile and methyl propyl ketone cyanohy-
drins. They found (S)-selectivity in Baliospermum montanum (Euphorbiaceae)
with rather low selectivity (37.8% e.e.) and (R)-selectivity in Passiflora edulis
(Passifloraceae), Eriobotyra japonica, Prunus mume, Prunus persica, Chaenomeles
sinensis, and Sorbus aucuparia (Rosaceae). The highest found HNL activity (in
Passiflora edulis) was about 20% compared to the activity of Prunus amygdalus. The
homogenate from the leaves showed activity towards benzaldehyde and 2-penta-
none substrates with 69 and 87% e.e., respectively [113]. Comparable results with
Prunus mume HNL (PmHNL), isolated from seeds of the ripe fruit [114], to PaHNL
for saturated and unsaturated aliphatic aldehydes were obtained [115]. Subse-
quently, the FAD-containing (R)-selective HNL from the seeds of Eriobotyra
japonica (loquat) was isolated, purified [116], and characterized with respect to
its application properties in organic synthesis [117]. With acetyltrimethylsilane in
a transhydrocyanation reaction in a biphasic system 95% conversion and 98% e.e.
were obtained; acetyltrimethylsilane was a more appropriate substrate than tert-
butyl methyl ketone [118]. Guanabana (Annona muricata) seed meal has been
reported as a source of (S)-HNL for the synthesis of (S)-aromatic, heteroaromatic,
and a,b-unsaturated cyanohydrins [119].
While HPLC methods are established for HNL activity assays in tissues from
natural sources, high-throughput screening methods are needed when HNL libraries
have to be screened. The large number of samples requires faster methods with (very)
small volumes.
23.3
Hydroxynitrile Lyase Catalyzed Addition of HCN to Aldehydes
23.3.1
(R)-Selective HNLs
For preparative applications, (R)-HNL from almonds has been extensively inves-
tigated. Brussee et al. [4, 120, 121] showed that without enzyme purification a
crude extract from almond meal in aqueous methanol using in situ HCN
generation from a solution of KCN in an acetate buffer affords cyanohydrins in
up to 93% e.e. By performing the reaction with a minimum amount of water and
slow addition of reactants (R)-o-chloro-mandelonitrile was obtained in high yield
(98%) and e.e. of 90%; this is worth noticing as o-chlorobenzaldehyde is not a good
substrate [122]. Apple meal, in the form of unpurified enzyme preparations,
accepts sterically hindered aldehydes (e.g., pivalaldehyde) as substrates, leading to
(R)-cyanohydrins with high enantiomeric purity (usually e.e. >90%) [50, 51]. A
purified enzyme from Prunus amygdalus supported on cellulose using non-
aqueous systems was employed for the first time by Effenberger and cowor-
kers [123]. Optimal results were obtained by almost completely suppressing the
954
non-enzymatic HCN addition using ethyl acetate as solvent. In this manner the
enantiomeric purity could be improved. Besides crystalline cellulose (AvicelÒ ),
other hydrophobic enzyme immobilization systems such as Celite were used
[124, 125]. Utilizing the natural support, unpurified almond meal in organic
solvents with small amounts of aqueous phase (4%) provides products with e.e.s
of up to 99% [51, 126–130]. Similar results were achieved with so-called micro-
aqueous systems in batch [131] and continuous processes [132]. In 2001 Lin and
coworkers examined the PaHNL catalyzed cyanohydrin reaction for fluorinated
benzaldehydes [133] and N-heteroaryl carboxaldehydes [134] under micro-aqueous
conditions and could not achieve better selectivities for the latter as described
before, which once more makes it obvious that the substrate nature is the first
parameter to address for selectivity. N-heteroaryl carboxaldehydes are not appro-
priate substrates for the known HNLs regarding stereoselectivity. Here the
selectivity could be increased by the concept of substrate engineering. N-substi-
tuted pyrrole-2- and -3-carboxaldehydes gave moderate to good enantiopurities;
91% absolute configuration with both PaHNL and HbHNL was achieved with N-
benzylpyrrole-3-carboxaldehyde [135].
To reduce the amount of racemic cyanohydrin produced by chemical conversion,
low concentrations of HCN were used by employing a relatively safe and convenient
source of this reagent: acetone cyanohydrin [127, 128, 136–138]. Kanerva has
developed a method in which HCN diffuses into the reaction mixture from a second
flask [129]. Wandrey used an enzyme membrane reactor for the continuous pro-
duction of product employing an (R)-HNL. In a production run the volumetric yield
was increased to 2400 g (R)-mandelonitrile per liter per day with a residence time of
just 3.8 min. The enzyme consumption was 17 000 U per kg of product [139]. By
applying a biphasic system a second industrial-scale procedure was developed [140].
Based on these findings, four parameters (pH, temperature, and concentration of
HCN and benzaldehyde) were optimized to obtain a throughput of 6700 g (R)-
mandelonitrile per liter per day.
A novel synthesis of (R)-cyanohydrins was described based on the use of cross-
linked and subsequently poly(vinyl alcohol)-entrapped (R)-hydroxynitrile lyases.
These immobilized lens-shaped biocatalysts have a well-defined macroscopic size
in the mm range, show no catalyst leaching, and can also be efficiently recycled.
Furthermore, this immobilization method is cheap, and the entrapped (R)- hydro-
xynitrile lyases gave similar results to those using free enzymes. Accordingly, (R)-
cyanohydrins were obtained in good yields and with high enantioselectivities of up to
>99% e.e. [141]. Some substrates, for example, acrolein, gave only low optical purity
with the PaHNL.
The catalytic capability of (R)-specific HNL from L. usitatissimum for the prepa-
ration of aliphatic cyanohydrins was investigated [72, 74, 141] and gave encouraging
results (e.e. up to 99%).
(R)-HNL from Arabidopsis thaliana shows high activity towards mandelonitrile
and the substrate range is similar to the (S)-selective HNLs from Hevea brasiliensis
and Manihot esculenta, including for aromatic and aliphatic aldehydes. The selectivity
of AtHNL is high [78, 79].
23.4 HNL-Catalyzed Addition of Hydrogen Cyanide to Ketones j955
23.3.2
(S)-Selective HNLs
As already mentioned, the (S)-hydroxynitrile lyase from Sorghum bicolor adds HCN only
to aromatic and heteroaromatic aldehydes. Initial investigations were performed on the
natural substrate 4-hydroxybenzaldehyde, and rather promising results concerning the
enantiomeric excess were found [81]. These results were confirmed and extended using a
suspension of enzyme immobilized on Avicel cellulose [143] or etiolated shoots of
S. bicolor [144] in diisopropyl ether. The Sorghum enzyme was one of the first recombinant
hydroxynitrile lyases [105], overexpressed in E. coli. In parallel to this work HbHNL was
also overexpressed [82], giving access to sufficient quantities of this enzyme both on a
preparative scale and for industrial use. To date only a few preparative applications for
Sorghum HNL [81] are known because of the narrow substrate range.
A similarly broad substrate range to that for the (R)-HNL from Prunus amygdalus is
revealed by the (S)-HNLs from Manihot esculenta and Hevea brasiliensis (EC 4.1.2.39).
Detailed sequence studies have revealed high homologies between both enzymes
(Manihot esculenta [106, 145], Hevea brasiliensis [44, 87]), as already mentioned. This
result was confirmed by the crystal structures. The latter was solved for Hevea brasiliensis
by the group of Kratky in Graz [92] and for the Manihot esculenta enzyme by the group
of Lauble in Stuttgart [103]. Expectations that these enzymes would be similar with
respect to substrate specificity were realized by experimental data from both groups.
The cyanoglycoside linamarin was found in 1965 in the seeds of the rubber tree
(Hevea brasiliensis) [146]. Two decades later the corresponding hydroxynitrile lyase
was described [147, 148]. Studies regarding the synthetic potential of this enzyme
with respect to the preparation of optically pure cyanohydrins started with the wild
type [83, 85, 86, 149]. As already mentioned, groundbreaking results were obtained
with the synthesis of the (S)-cyanohydrin of 3-phenoxybenzaldehyde, which is a
precursor for some important synthetic pyrethroids [150–152].
HNL from Manihot esculenta Crantz (termed EC 4.1.2.37 at the time because EC
4.1.2.39 was not created before 1999 [153] meanwhile termed as EC 4.1.2.47) was
purified to homogeneity from young leaves of the cyanogenic tropical crop plant
cassava in 1994 [106]. Initial experiments demonstrated a broad substrate range, but
only unsatisfactory optical purities were obtained [154]. Overexpression of the cloned
Manihot esculenta HNL gene in E. coli increased the accessibility and specific activity
of the biocatalyst [105].
Table 23.2 shows a selection of substrates with typical enantioselectivities of the
obtained cyanohydrins from the respective HNLs.
23.4
HNL-Catalyzed Addition of Hydrogen Cyanide to Ketones
Preparative elaboration of the (R)-cyanohydrins of ketones employing the hydro-

xynitrile lyase from Prunus amygdalus was first investigated in organic solvents [47].
Alkyl methyl ketones were obtained in moderate yields and in high optical purity,
956
Table 23.2 Aldehydes R-CHO as substrates for hydroxynitrile lyase-catalyzed cyanohydrin formation.
R HNL Source Conv. (%) E.e. (%) Reference
Ph (R) Prunus amygdalus Quan. 99 [126]

(R) Arabidopsis thaliana >99 >99 [78]
(S) Sorghum bicolor 97 97 [311]
(S) Hevea brasiliensis 96 99 [83]
(S) Manihot esculenta Quan. 98 [105]
(S) W128A Manihot esculenta 97 97 [202]
(E)-PhCH¼CH (R) Prunus amygdalus 54 87 [312]
(S) Manihot esculenta 80 95 [202]

3-PhO(C6H4) (R) Prunus amygdalus 99 98 [292]
(R) Prunus mume 42 >99 [48]
(R) Prunus armeniaca 82 99 [52]
(R) Arabidopsis thaliana 83 >95 [78]
PhCH2OCH2 (S) Hevea brasiliensis 92 12 [83]
PhCH2 (R) Prunus amygdalus 83 88 [136]
(R) Arabidopsis thaliana 97 96 [78]
PhCH2CH2 (R) Linum usitatissimum 10 10 [142]

4-CH3(C6H4) (S) Manihot esculenta 50 99 [202]
2-CH3O(C6H4) (R) Prunus amygdalus 65 96 [136]
(R) Prunus mume 17 97 [48]
4-CH3S(C6H4) (R) Prunus communis 98 96 [314]
3,4-CH2O2(C6H3) (S) Manihot esculenta 84 86 [34]
2-F(C6H4) (R) Prunus amygdalus 96 84 [133]
(R) Arabidopsis thaliana >99 99 [78]
3-F(C6H4) (R) Arabidopsis thaliana >99 >99 [78]
4-F(C6H4) (R) Prunus amygdalus 92 84 [132]
3,4-F2(C6H3) (R) Prunus amygdalus 71 84 [133]
2,3-F2(C6H3) (R) P. amygdalus 92 46 [133]
2,6-F2(C6H3) (R) P. amygdalus 70 41 [133]
2-Cl(C6H4) (R) Arabidopsis thaliana >99 99 [78]
23.4 HNL-Catalyzed Addition of Hydrogen Cyanide to Ketones

(Continued )
j957
958
3-Cl(C6H4) (R) Arabidopsis thaliana 99 >99 [78]

4-Cl(C6H4) (R) Prunus amygdalus 94 97 [25]
(R) Prunus mume 21 99 [48]
2-Br(C6H4) (R) Arabidopsis thaliana 99 98 [78]
s Manihot esculenta 96 96 [202]
s W128A Manihot esculenta 98 93 [202]
3-Br(C6H4) (R) Arabidopsis thaliana 99 95 [78]
4-Br(C6H4) (R) Prunus mume 22 99 [48]
2-I(C6H4) (R) A. thaliana >99 >95 [78]
3-I(C6H4) (R) A. thaliana 98 93 [78]

4-I(C6H4) (R) A. thaliana 99 92 [78]
2-HO(C6H4) (S) Manihot esculenta 47 91 [202]
3-HO(C6H4) (S) Manihot esculenta 88 97 [202]
4-HO(C6H4) (R) Arabidopsis thaliana 96 97 [78]
3-NO2(C6H4) (R) Prunus amygdalus 89 89 [112]
2-Ph(C6H4) (R) P. armeniaca 68 32 [52]
4-Ph(C6H4) (R) P. armeniaca 72 96 [52]
1-Naphthyl (R) P. amygdalus 89 90 [248]
(R) P. mume 60 93 [48]
2-Naphthyl (R) Prunus amygdalus >99 95 [248]
(R) P. mume 58 98 [48]

2-Furyl (R) Prunus amygdalus 96 99 (S)a) [315]
(S) Sorghum bicolor 80 80 (R)a) [315]
(S) Hevea brasiliensis 95 98 (R)a) [159]
(S) Annona muricata 95 87 [119]
2-(5-Methylfuryl) (R) Prunus amygdalus 60 97 (S)a) [134]
3-Furyl (R) P. amygdalus 96 99 [315]
3-(2-Methylfuryl) (R) Prunus amygdalus 78 24 (S)a) [134]
2-Thiophenyl (R) Prunus amygdalus 71 99 (S)a) [315]
(S) Sorghum bicolor 64 91 (R)a) [315]
(S) Hevea brasiliensis 98 99 (R)a) [159]
(S) Manihot esculenta 85 96 (R)a) [105]
2-(5-Bromothiophenyl) (R) Prunus amygdalus 72 86 [134]
2-(3-Methylthiophenyl) (R) P. amygdalus 50 65 [134]
3-Thiophenyl (R) P. amygdalus 95 99 [315]
2-(1-Tosylpyrrolyl) (R) Prunus amygdalus 4 1 (S)a) [134]
2-(1-Boc-pyrrolul) (R) P. amygdalus 5 2 (S)a) [134]
2-(1-MOM-pyrrolyl) (R) P. amygdalus 33 81 (S)a) [134]
2-(1-Methylpyrrolyl) (R) P. amygdalus 17 40 (S)a) [134]
(S) Hevea brasiliensis 17 (yield) 5 (R)a) [135]
2-(1-Benzylpyrrolyl) (R) Prunus amygdalus 2 (yield) 29 (S)a) [135]
(Continued )
j959
960
(S) Hevea brasiliensis 7 (yield) 51 (R)a) [135]

2-(5-Acetyl-1-methylpyrrolyl) (R) Prunus amygdalus 99 34 [134]
3-(5-Cyano-1-methylpyrrolyl) (R) P. amygdalus 84 67 [134]
3-(1-Boc-tryptophanyl) (R) P. amygdalus 2 8 [134]
2-Pyridyl (R) P. amygdalus 42 0 [134]
(R) P. mume 89 22 [48]
2-(6-Methylpyridyl) (R) P. amygdalus 38 0 [134]
2-(6-Bromopyridyl) (R) P. amygdalus 92 65 [134]
3-Pyridyl (R) P. amygdalus 99 50 [134]
(R) P. mume 90 75 [48]
4-Pyridyl (R) P. mume 65 41 [48]
4-Quinolyl (R) P. mume 73 28 [48]

2-(1-Methylimidazolyl) (R) P. amygdalus 94 5 (S)a) [134]
2-(Thiazolyl) (R) P. amygdalus 97 67 (S)a) [134]
CH2¼C(CH3) (R) P. mume 88 (yield) 32 [48]
CH2¼CH(CH2)2 (R) P. armeniaca 80 98 [52]
CH2¼CHC(CH3)2CH2 (R) P. armeniaca 82 98 [52]
CH2¼CHC(Et)2CH2 (R) P. armeniaca 78 96 [52]
CH2¼CHC(Ph)2CH2 (R) P. armeniaca 70 97 [52]
(CH3)2C¼CH(CH2)2 (R) P. armeniaca 85 98 [52]
CH2¼CHC(cyclopentyl)CH2 (R) P. armeniaca 82 98 [52]
CH2¼CHC(cyclohexyl)CH2 (R) P. armeniaca 76 96 [52]
CH2¼CHC(cycloheptyl)CH2 (R) P. armeniaca 72 95 [52]
CH2¼CH (R) P. mume 90 (yield) 42 [48]
(R) Linum usitatissimum Quan. 74 [74]
(E)-PhCH¼CH (R) Prunus amygdalus 99 54 [112]
(S) Annona muricata 11 82 [119]
(E)-CH3CH¼CH (R) Prunus amygdalus 99 98 [312]

(R) P. mume 70 (yield) 96 [48]
(E)-CH3CH¼C(CH3) (R) Prunus mume 78 (yield) 90 [48]
(E)-CH3CH¼C(CH2CH3) (R) P. mume 50 (yield) 92 [48]
(CH3)2C¼CH (R) P. mume 72 (yield) 92 [48]
(E)-CH3(CH2)2CH¼CH (R) P. mume 62 (yield) 94 [48]
(E)-CH3CH2CH¼C(CH3) (R) Prunus mume 58 (yield) 96 [48]
(E)-CH3(CH2)2CH¼C(CH2CH3) (R) P. mume 48 (yield) 92 [48]
(E)-CH3(CH2)5CH¼CH (R) Prunus mume 40 (yield) 12 [48]
(Z)-CH3(CH2)2CH¼CH (S) Hevea brasiliensis 35 80 [149]
(E,E)-CH3CH¼CHCH¼CH (R) Prunus amygdalus 36 95 [136]
(R) Prunus mume 32 (yield) 96 [48]
(E,E)-CH3CH2CH¼CHCH¼CH (R) P. mume 38 (yield) 97 [48]
(E,E)-(CH3)2C¼CH(CH2)2C(CH3)¼CH (R) P. mume 48 (yield) 98 [48]
CH3(CH2)2C:C (S) Hevea brasiliensis 88 80 [149]
1-Cyclohexenyl (R) P. mume 54 (yield) 90 [48]
3-Cyclohexenyl (R) P. amygdalus 86 55 [121]
(R) P. mume 70 (yield) 96 [48]
Cyclohexanyl (R) Prunus amygdalus 90 99 [255]
(Continued )
j961
962
(R) P. mume 72 (yield) 93 [48]

Cyclopentanyl (R) Prunus mume 70 (yield) 94 [48]
Cyclobutanyl (R) P. mume 78 (yield) 92 [48]
CH3(CH2)10 (S) Manihot esculenta 80 71 [202]
CH3(CH2)8 (R) Prunus mume 38 (yield) 12 [48]
CH3(CH2)7 (R) Prunus amygdalus 82 96 [316]

(R) P. mume 42 (yield) 20 [48]
CH3(CH2)6 (R) Prunus mume 58 (yield) 52 [48]
BrCH2(CH2)6 (R) Prunus amygdalus 40 (yield) 97 [137]
CH3(CH2)5 (R) P. mume 56 (yield) 82 [48]
BrCH2(CH2)5 (R) P. amygdalus 41 (yield) 92 [137]
CH3(CH2)4 (R) P. amygdalus 72 97 [144]
(R) P. mume 22 (yield) 90 [48]
BrCH2(CH2)4 (R) Prunus amygdalus 65 (yield) 90 [137]
CH3(CH2)3 (R) P. mume 57 (yield) 88 [48]
HOCH2(CH2)3 (R) P. amygdalus Quan. 46 [55]

(R) P. mume 58 (yield) 90 [48]
(R) Linum usitatissimum 91 98 [74]
HOCH2(CH2)2 (R) Prunus amygdalus 85 27 [55]
HSCH2CH2 (R) P. amygdalus 75 96 [255]
CH3CH2 (R) P. mume 68 (yield) 94 [48]
(CH3)2CH (R) Prunus amygdalus 99 83 [127]
(R) P. mume 62 (yield) 94 [48]
(CH3)2CHCH2 (R) Prunus mume 65 (yield) 95 [48]
(CH3)3C (R) P. amygdalus 58 92 [136]
(R) P. mume 6 (yield) 96 [48]
2-Bicyco[2.2.1]hept-5-enyl (R) Prunus mume 62 (yield) 89 [48]
Myrtenyl (R) P. amygdalus 71 >99 [248]
(R) P. mume 82 (yield) 99 [48]
(S) Hevea brasiliensis 63 >99 [248]
a) Change of product configuration owing to a priority replacement according CIP rules.

j963
964
whereas with alkyl ethyl ketones the chemical and optical yields were reported to be
lower [155]. Working with almond meal instead of purified enzyme resulted in
an astonishingly high enantiomeric excess [51]. Similar results were obtained with
98% e.e. for the (R)-cyanohydrin of butyl methyl ketone [156]. The substrate scope is
not limited to acyclic aliphatic ketones and a few examples of methyl phenyl ketones
but covers also cyclic, bicyclic, heterocyclic, and silicon-containing [157] com-
pounds [2, 43, 44, 81, 95, 98, 158].
(R)-Hydroxynitrile lyase from Linum usitatissimum (flax) has been used for the
synthesis of (R)-butan-2-one cyanohydrin on a preparative scale [72].
Concerning (S)-ketone cyanohydrins, impressive results were achieved with
aliphatic and aromatic ketones, for example, acetophenone cyanohydrin. The latter
was obtained using the hydroxynitrile lyase from either Hevea brasiliensis
(40% conversion, 99% e.e.) [159] or Manihot esculenta HNL (87% conversion,
98% e.e.) [160].
4-Substituted cyclohexanones were subjected to enzymatic cyanohydrin synthesis
with PaHNL and MeHNL to obtain access to starting materials for substituted
tetronic acids and also for comparing the results with the Prelog/Ringold model,
which was developed for HLADH (horse liver alcohol dehydrogenase). While PaHNL
catalyzed almost completely the formation of the trans isomers with all the tested
ketones, MeHNL favors the cis isomers. The rate of conversion appeared to be faster
for MeHNL than for PaHNL.
In 2004 five- and six-membered cyclic ketones, namely, tetrahydrofuran-3-one and
tetrahydro-2H-3-pyranone, were subjected to hydroxynitrile lyase catalyzed cyano-
hydrin syntheses. Both substrates were accepted by PaHNL and HbHNL, yielding
moderate e.e.s (up to 81%). Racemic mixtures of methyl substituted tetrahydrofuran-
3-one and tetrahydrothiophen-3-one were also substrates for the above-mentioned
HNLs. The diastereomeric distribution was analyzed taking the reaction conditions
into account [90]. Both enzymes led to a nearly racemic mixture of the corresponding
cis- and trans-cyanohydrins bearing a methyl substituent at C2. Molecular modeling
calculations confirmed the experimental data regarding the steric outcome of the
transformation.
Table 23.3 shows the results gained by HNL-catalyzed conversions of selected
methyl ketones into the corresponding cyanohydrins.
23.5
Transhydrocyanation
Transhydrocyanation of aromatic and aliphatic aldehydes with acetone cyanohydrin

catalyzed by (R)-hydroxynitrile lyase to give cyanohydrins (Scheme 23.2) was first
performed by Kyler and coworker [136]. This innovative method avoids the use of free
HCN as the cyanide source and is mostly accompanied by a slight decrease in e.e.
compared to standard conditions. The procedure was optimized in the group of
Kanerva [127] by comparing the feasibility of powdered almond meal as a catalyst to
that of a purified enzyme preparation in an organic solvent.
23.5 Transhydrocyanation j965
Table 23.3 Methyl ketones R-CO-Me as substrates for hydroxynitrile lyase-catalyzed cyanohydrin
formation.
R HNL Source Conversion (%) E.e. (%) Reference
CH3CH2 (R) Prunus amygdalus 80 76 [318]

(R) P. mume 48 72 [48]
CH3(CH2)2 (R) Prunus amygdalus 70 97 [318]
ClCH2(CH2)2 (R) Prunus amygdalus 87 84 [318]
(CH3)2CH (R) Prunus amygdalus 54 90 [318]
(CH3)2CHCH2 (R) Prunus amygdalus 57 98 [318]
(CH3)3C (S) Hevea brasiliensis 49 78 [83]
CH2¼CH(CH2)2 (R) Prunus amygdalus 80 97 [318]
(CH3)2C¼CH(CH2)2 (S) Manihot esculenta 78 61 [202]
Ph (R) Prunus amygdalus 14 90 [144]
PhCH2 (S) Hevea brasiliensis 74 95 [83]
3-Br(C6H4)CH2 (R) Prunus amygdalus 1 (yield) 86 [73]
(S) Manihot esculenta 44 (yield) 97 [73]
3-F(C6H4)CH2 (R) Prunus amygdalus 2 (yield) 20 [73]
(R) Linum usitatissimum 20 (yield) 83 (S) [73]
3-Cl(C6H4)CH2 (R) Prunus amygdalus 1 (yield) 44 [73]
3-Br(C6H4)CH2 (R) Linum usitatissimum 37 (yield) 99 (S) [73]
3-CH3(C6H4)CH2 (R) Linum usitatissimum 40 (yield) 97 (S) [73]
3-CF3(C6H4)CH2 (R) Linum usitatissimum 19 (yield) 96 (S) [73]
3-CH3O(C6H4)CH2 (R) Linum usitatissimum 31 (yield) 99 (S) [73]
(Continued )
966
R HNL Source Conversion (%) E.e. (%) Reference
4-Br(C6H4)CH2 (R) Prunus amygdalus 3 (yield) 29 [73]

4-CH3O(C6H4)CH2 (R) Linum usitatissimum 30 (yield) 84 (S) [73]
PhCH2CH2 (R) Prunus amygdalus 2 (yield) 55 [73]
4-HO(C6H4)CH2CH2 (R) Linum usitatissimum 13 (yield) 75 (S) [73]
4-CH3O(C6H4)CH2CH2 (R) Linum usitatissimum 1 (yield) 99 (S) [73]
The attempt to use racemic 2-methyl-2-hydroxyhexanenitrile as the cyanide donor

was rewarded by obtaining aliphatic v-bromo cyanohydrins from the corresponding
aldehydes in 90–97% e.e. [137]. In 2002 the concept was applied to v-hydroxyalk-
anals [55] and the e.e.s could be improved slightly on performing the transformation
in a micro-aqueous medium.
In 2003 transhydrocyanation was applied to silicon-containing aliphatic ketones
with (R)-hydroxynitrile lyase from apple seed meal showing better activity and
selectivity than almond meal [161]. Later, additionally, (R)-HNLs from plum, loquat,
and peach as well as (S)-MeHNL were applied for the above-mentioned reaction and
compared to the transhydrocyanation of the carbon counterpart 3,3-dimethyl-2-
butanone. In all cases substrate conversion and product e.e. were much higher for
the silicon-containing substrate than the carbon-containing derivative [162, 163].
The transhydrocyanation concept was further modified by the application of ethyl
cyanoformate as cyanide donor. In a chemoenzymatic one-pot reaction of ethyl
cyanoformate with benzaldehyde, catalyzed by PaHNL, ethoxycarbonylated (R)-
mandelonitrile was formed. Investigations revealed a two-step procedure consisting
of an enzyme-catalyzed addition of HCN, which was generated by hydrolysis of ethyl
cyanoformate to the aldehyde, followed by protection of the free cyanohydrin in the
second step [164]. Recently, transhydrocyanation was also applied to ketones. As a
biocatalyst (R)-hydroxynitrile lyase was used [165].
O OH O
OH CN
+ +
R1 R2 CN R1
R2
Scheme 23.2 Transhydrocyanation approach: R1 ¼ alkyl, cycloalkyl, aryl, heteroaryl; R2 ¼ H, alkyl.

23.6 Mechanistic Aspects and Enzymatic Promiscuity j967
Hanefeld and coworkers applied the concept of transhydrocyanation to HbHNL
and improved this thermodynamically unfavored reaction by coupling the cyano-
hydrin formation with a lipase-catalyzed acylation [166].
23.6
Mechanistic Aspects and Enzymatic Promiscuity
Detailed mechanistic studies concerning PaHNL, HbHNL, and MeHNL have been
reported. The results of these investigations are summarized here.
General acid–base catalysis is the mechanism of the hydroxynitrile lyase catalyzed
reaction involving all types of (R)- and (S)-selective HNLs, which differ regarding
details for each enzyme [45]. In the following we summarize the reported mechan-
isms of the some HNLs.
23.6.1
(R)-PaHNL (EC 4.1.2.10)
The HNLs from Prunus species (Rosaceae) are FAD-containing enzymes [167].
Binding of competitive inhibitors affects the absorption spectrum of the flavin [42],
and FAD in the reduced state leads to an inactive enzyme. Experimental data
confirmed that the redox properties of the flavin are required for enzymatic activity,
even though FAD does not have a redox role in HNL. The crystal structure of the

61 kDa PaHNL isoenzyme has been solved to 1.5 A resolution. It is a member of the
GMC-oxidoreductase family and has four glycosylation sites. A hydrophobic tunnel
leads to the active site, which has a positive electrostatic potential and is assumed to be
responsible for the stabilization of the negatively charged cyanide ion. The FAD is
deeply buried with no contact with solvent and is close to the active site [168]. Docking
calculations with the natural substrate were used to locate the active site and identify
His497 as the general base in catalysis [169]. These simulations could be confirmed
by 3D structural data of this lyase with benzaldehyde bound within the active site. A
second histidine (His459) within the active site could also function as a proton donor
for the cleaved cyanide ion [170].
Kinetic data yield an ordered Uni Bi mechanism in which the aldehyde is the first
substrate bound (for the synthesis direction) [171]. Blanch et al. investigated the
PaHNL catalyzed cyanohydrin reaction in a biphasic system [172, 173]. Experimental
data and modeling confirmed the assumption that the reaction takes place at the
interface [174, 175]. By performing dynamic interfacial measurements is was
possible to study the adsorption behavior at the liquid–liquid interface. For five
hydrophobic solvents large changes in the interfacial pressure were observed,
whereas no changes were found for the non-hydrophobic solvents ethyl acetate and
diisopropyl ether. The interpretation of this result was that the structure of the native
enzyme was not destroyed by adsorption at the interface and that the adsorption is
reversible [176]. Straathof and coworkers describe the PaHNL catalyzed cyanohydrin
reaction to take place in the aqueous phase [177].
968
23.6.2
(R)-LuHNL (EC 4.1.2.46)
The 46kDa protein from Linum usitatissimum is a zinc-dependent hydroxynitrile lyase

with homology to alcohol dehydrogenase family members [142]; the cysteine and
histidine residues responsible for coordination of an active site Zn2 þ and a second
structurally important Zn2 þ as in ADHs are conserved.
23.6.3
(S)-HbHNL (EC 4.1.2.47)
(S)-HbHNL and (S)-MeHNL (see below) are highly homologous (77% sequence
identity), have no cofactor, are non-glycosylated, and belong to the a/b-hydrolase
superfamily.
HbHNL exists in neutral aqueous solution as a homodimer [178]. The crystal

structure of the HbHNL, resolved to 1.9 A, shows an active site that is buried deep
within the protein and connected with the outside by a narrow tunnel [92]. Subse-
quently, structural parameters were reported for the same enzyme, refined against

crystallographic data collected to 1.1 A resolution [95]. Crystallographic data were also
measured and solved for HbHNL complexed with the natural substrate acetone as
well as with various inhibitors, including trichloroacetaldehyde, hexafluoroacetone,

and rhodanide [94]. Further X-ray crystal structures at 1.54 and 1.76 A of HbHNL
complexes with the two chiral substrates mandelonitrile and 2,3-dimethyl-2-hydro-
xybutyronitrile obtained by soaking and rapid freeze quenching techniques were
determined [179]; this was the first observation of the complex of a HNL and a chiral
substrate. As expected only the (S)-enantiomer was bound to the active site in the
same mode as the natural substrate acetone cyanohydrin. In this enzyme, the
catalytic triad Ser80-His236-Asp207 acts as general acid/base for deprotonation of
the cyanohydrin hydroxyl group and an active site lysine (Lys236) provides the
positive charge to stabilize the cyanide ion [94]. The mutein K236L is inactive in
the cyanohydrin cleavage/formation reaction although the 3D structure is similar to
the wild-type enzyme, which is further evidence for the crucial role of Lys236 for the
enzyme activity [180]. A large hydrophobic pocket was identified in the active site. The
current view of the molecular reaction mechanism of HbHNL-catalyzed cyanohydrin
cleavage and synthesis was deduced from crystallographic experiments [83, 92, 94,
95, 170], NMR [181], molecular modeling [96], and ab initio quantum chemical
calculations [182] and involves the following four key steps in the cleavage direction:
(i) the substrate cyanohydrin is attached to the active site by hydrophobic interactions
and by hydrogen bonding between its hydroxy group and the OH groups of Thr11 and
Ser80; (ii) after the substrate binds, the OH-Ser80 is deprotonated by His235, which
induces the simultaneous deprotonation of the substrate hydroxyl by Ser80; (iii)
subsequent cleavage of the cyanohydrin is assisted by stabilization of the charge of
the nascent cyanide through interaction with the positive charge of Lys236; (iv) the
cyanide ion formed is protonated by His235 [96]. All these conclusions are confirmed
by enzyme-kinetic data [183]. The inhibition pattern observed for benzaldehyde and
HCN corresponds well to an ordered Uni Bi mechanism including the formation of a
23.6 Mechanistic Aspects and Enzymatic Promiscuity j969
dead-end complex of the enzyme, (S)-mandelonitrile, and HCN. In the degradation
of cyanohydrins the latter is the first product released from the enzyme followed by
benzaldehyde, while in the synthesis reaction benzaldehyde is the first substrate
bound to the enzyme followed by HCN. Steiner and Griengl used a Lewis cell to
investigate the interaction between mass transfer and the biocatalytic reaction of
HbHNL in a two-phase system. Their results show that the enzymatic reaction takes
place in the bulk of the aqueous phase and in the thin film close to the interface and/
or directly at the interface. Mass transfer of benzaldehyde from the organic to the
aqueous phase is enhanced by the biocatalytic reaction [184].
A major surprise was a report on the biocatalytic nitroaldol (Henry) reaction
catalyzed by HbHNL [185], which represents impressive proof of the possible
promiscuity in enzymes. A nitroaldol reaction has never been detected with enzymes
as catalysts before. A broad range of aromatic, heteroaromatic, and aliphatic
aldehydes were transformed into the corresponding nitro alcohols [186].
23.6.4
(S)-MeHNL (EC 4.1.2.47)
Initial mechanistic studies proposed that the enzyme-catalyzed reaction proceeded

through the hydrogen bonding of Gly78 and Ser80 to the carbonyl compound and
nucleophilic attack of cyanide after deprotonation of the latter by the catalytic triad,
the anion was then protonated by the catalytic triad to build the cyanohydrin [145]. For
analysis of the structure and function of the MeHNL enzyme X-ray crystallography
and site-directed mutagenesis were applied some years later. The 3D structure of the

MeHNL-S80A-acetone cyanohydrin complex was determined at 2.2 A resolution. The
mechanism of cyanogenesis was proposed [102]. Later, the crystal structures of
MeHNL complexed with acetone and the product analog chloroacetone were deter-

mined and refined at 2.2 A resolution [103]. In the enzyme–substrate complex with
acetone the carbonyl unit is hydrogen bonded to Thr11 and Ser80 and to a lesser
extent to Cys81. The proposed mechanism still expresses the base-catalyzed chemical
reaction of HCN and oxo-compounds. In a previous study, it was shown that the active
site is accessible through a narrow channel and consists of a smaller and a larger
binding pocket. Trp128 was believed to cover a significant part of the hydrophobic
channel leading to the active site of MeHNL. The mutein MeHNL-W129A was
prepared to enable study of the effect of the amino acid Trp and its exchange. Wild-
type MeHNL and MeHNL-W129A showed comparable activity towards the natural
substrate acetone cyanohydrin, but the specific activities towards unnatural sub-
strates mandelonitrile and 4-hydroxymandelonitrile were increased 9- and 450-
fold, respectively, in comparison to the wild-type enzyme. Obviously the W128A
mutein has a significantly larger channel at the entrance of the active site [187].
23.6.5
(S)-SbHNL (EC 4.1.2.11)
This enzyme has a molecular weight of 95 kDa with 510 amino acids and contains
two different subunits a and b, and is a member of the a/b hydrolase family.
970
The active enzyme form is a a2b2 heterotetramer assembled as a dimer of ab

units [188, 189]. The mature enzyme is N-glycosylated and exists in three different
isoforms. The SbHNL amino acid sequence is not similar to that of any other
known HNL, but it is 60% identical and 73% similar to that of wheat serine
carboxypeptidase [190]. SbHNL has been crystallized in the presence of the

inhibitor benzoic acid and the 3D structure has been determined at 2.3 A
resolution [191]. The suggested mechanism for the cyanohydrin cleavage predicts
that the substrate forms a hydrogen bond between the cyanohydrin hydroxyl
group and Ser158, while the nitrile group is directed towards Leu190 and Gly159.
The phenyl ring is stacked between the side chain of His160 and the nonpolar face

of Pro64, and the additional hydroxyl group at C4 is within 2.7 A of Asp126 and
thereby assists in binding the substrate in the active site. A docking simulation
further revealed that the carboxylate moiety of the C-terminal residue Trp270
undergoes movement to form a good hydrogen bond contact between the Trp270
carboxylate and the cyanohydrins hydroxyl group, while an active-site water
adjusts to maintain a hydrogen bond contact between the Trp270 carboxylate
and the substrate nitrile group. Cyclohexane carbaldehyde differs from benzal-
dehyde in that it has sp3-hybridized C atoms with additional axial protons, rather
than the in-plane protons of the phenyl ring of benzaldehyde, and is not accepted
as aliphatic substrate. Additionally, carbonyl substrates with small aliphatic side
chains such as butyraldehyde and isobutyraldehyde are not accepted as substrates.
Models of complexes of SbHNL with cyclohexane carbaldehyde reveal several
closer contacts of the axial protons, in particular at C2–C4 of cyclohexane
carbaldehyde with Pro64 and His160. Similar disfavored interactions are observed
for C4 of butyraldehyde.
23.7
Improvement of HNLs by Enzyme Engineering, Enzyme Stabilization
Although many HNLs are well-characterized enzymes and have already made
their way into industrial applications, there is still room for improvement. Not all
substrates can be converted in sufficient amount and enantiomeric purity.
Enzyme and/or substrate engineering are widely used approaches to decrease
such shortcomings. Directed evolution [192] and rational design [193] are, mainly,
two well-established approaches for enzyme improvement regarding activity,
selectivity, and even stability. During the last few years also a combination of
these approaches – a semi-rational approach – is coming into prominence [194].
Substrate engineering attempts were reported by Wang and Withers with glyco-
sidases [195] and by Griengl and coworkers with hydroxylating enzymes [196–198]
and HNLs [199]. An example of a coupled approach of substrate and enzyme
engineering published recently showed impressive results regarding both activity
(10–20 times less enzyme amount) and selectivity (e.e. increased from 10% to
about 90%) [97].
23.7 Improvement of HNLs by Enzyme Engineering, Enzyme Stabilization j971
Glieder and coworkers have improved the HNL from Prunus amygdalus
starting from (R)-HNL isoenzyme 5 for synthesizing (R)-pantolactone, which
is used in vitamin B5 synthesis. (R)-Pantolactone can be synthesized from
hydroxypivalaldehyde and HCN catalyzed by PaHNL. The e.e. and the amount
of enzyme needed for the reaction was not satisfying. Several preparations of
natural and recombinant PaHNL isoenzymes and also other Rosaceae HNLs
were screened. Enzymes with improved properties regarding activity and selec-
tivity were not found. At this point, the best enzyme was subjected to saturation
mutagenesis at several positions identified by molecular modeling. The e.e. could
be increased from 89% to 97% [59]. Another success story regarding PaHNL
improvement is mutein PaHNL5-L1Q-A111G. Large-scale production of (R)-2-
chloromandelic acid – the chiral building block for the drug ClopidogrelÒ – via
(R)-2-chlorobenzaldehyde cyanohydrin was hindered by low turnover rates and
moderate e.e. both in the enzymatic and metal-catalyzed reaction. Rationally
designed mutation of alanine to glycine at position 111 raised the yield enor-
mously [200, 201].
Another example of improved HNL is the tunnel-variant W128A of MeHNL [187].
Based on the crystal structure and reaction mechanism of MeHNL, a tryptophan
residue at the entrance to the active site was supposed to play a crucial role regarding
enzyme activity. Exchange of the bulky amino acid by site-directed mutagenesis to the
smaller amino acid alanine leads to the enhanced activity [202].
In two other examples the HNLs were highly improved by single point mutations
in terms of both converting sterically demanding substrates [57] and also regarding
the enantioselectivity [203]. Another example was reported for recombinant MeHNL
in E. coli, where a single replacement improved the folding and stability of the
enzyme [204].
The very first step to finding improved HNLs is the establishment of high-
throughput screening methods. Eggert and coworkers have developed a spec-
troscopic assay based on HCN, which makes it independent of substrate
nature [205]. Two high-throughput screening assays for the cleavage direction
of the cyanohydrin reaction were developed by the group of Schwab. One is based
on HCN detection. The librated gaseous HCN from the cyanohydrin cleavage is
detected by a colorimetric reaction semi-quantitatively and is not restricted to the
substrate [206], while the second screening assay [207] is a coupled assay with
dehydrogenases capable of oxidizing or reducing the reaction product (aldehyde)
released from the bacterial colony (filter assay). The release or consumption of
NADH in the area of colonies was monitored by its fluorescence at 450 nm.
Assays for the synthesis direction of the cyanohydrin reaction are also avail-
able [208], excluding all possible cleavage scenarios and aiming at both activity
and selectivity of the HNL.
The choice of expression conditions greatly influences the production of stable
enzymes. Semba et al. investigated in detail the expression conditions of MeHNL and
could improve the enzyme activity and yield 850-fold by employing the expression
at 17 C (instead of 37 C) in E. coli [209]. The expression of MeHNL in a
972
multi-auxotrophic mutant of Saccharomyces cerevisiae cells was also reported by

Semba [210]. Glieder and coworkers achieved higher expression levels of PaHNL5
through a new PAOX1 promoter variant and the helper protein PDI (protein disulfide
isomerase) [208].
HNLs have been immobilized to obtain the known benefits of immobilized
enzymes like improved stability and recycling. Several techniques have been
employed, such as sol–gels, immobilization on solid supports, CLEAs, and CLECs,
as well as polymer-entrapped, lens-shaped hydrogels[141, 211–213]. Adsorptive
immobilization on Celite was the first method investigated [124]. A similar
procedure is immobilization on nitrocellulose [108–111]. For the production of
active enzymes entrapped in gels a two-step procedure was applied. First the
enzymes were crosslinked and subsequently entrapped in a hydrogel matrix based
on poly(vinyl alcohol) [141]. Encapsulation techniques have been reported as
well [214]. MeHNL was the first HNL for which crosslinked enzyme crystals were
used in the asymmetric cyanohydrin reaction, showing higher stability than
immobilized enzymes. The drawback of this technique was a significant loss of
specific activity. Crosslinked enzyme aggregates (CLEAs) of HNLs show improve-
ments in terms of stability and activity using dextran polyaldehyde for crosslinking
the precipitation instead of the frequently used glutaraldehyde [215]. Meanwhile,
CLEAs from almost every purchasable HNL have been prepared. PaCLEAs proved
to be highly effective catalysts under micro-aqueous conditions for slow-reacting
aldehydes and could be recycled ten times without loss of activity [216]. LuCLEAs
with a high specific activity (300 U g1) were prepared and could partially be
recycled for the synthesis of (R)-2-butanone cyanohydrin on a preparative scale over
two batches [217]. MeCLEAs were prepared that showed high preservation of the
synthetic activity (93%) and were examined in micro-aqueous and anhydrous media
with various aldehydes and ketones as substrates [218]. Sheldon and coworkers
compared HNLs immobilized in sol–gels and in CLEAs. They found that sol–gels
were highly efficient at low catalyst loading and were particularly stable towards
organic solvents and substrate/product inactivation. For CLEAs the results were
ambiguous. Commercial MeHNL-CLEA proved to be a very robust and efficient
catalyst, while CLEAs from Pa- and HbHNL were obviously inactivated [219].
Roberge and coworkers could synthesize (R)- and (S)-3-pyridine-carboxaldehyde
cyanohydrins with >93% e.e. and >65% yield with PaCLEA and MeCLEA,
respectively. These aldehydes are difficult substrates for HNLs. Applying CLEA
technology and fine tuning of the reaction conditions led to improvements in the
enantioselectivity [220].
A major drawback of enzyme-catalyzed reactions is the stability under reaction
conditions. For the synthesis of enantiomerically pure cyanohydrins the unselective
chemical reaction can be suppressed by performing the reaction at low pH. The HNL
stability decreases with decreased pH values. In recombinant PaHNL5 produced by
Pichia pastoris N-glycosylation results in enhanced protein stability through protec-
tion by the attached sugar moieties and shows extraordinary stability at low pH. To
identify the important glycosylation sites, so-called serine scanning was
developed [221].
23.8 Resolution of Racemates j973
23.8
Resolution of Racemates
23.8.1
Hydroxynitrile Lyase as Catalyst
It is possible to treat a racemic cyanohydrin with a (R)- or (S)-HNL to decompose

selectively one enantiomer of this mixture (exemplified in Scheme 23.3). The (R)-
HNL from Prunus amygdalus was used for the resolution of racemic cyanohydrins.
Employing a biphasic system, namely, citrate buffer/diisopropyl ether (40 : 1) at 39

C, catalytic amounts of PhNH2 and semicarbazide were added for aldehyde capture.
In this manner the (S)-cyanohydrin of 3-phenoxybenzaldehyde was obtained with
91% e.e. at 50% conversion [222].
HO CN OH O
HNL CN
2 + + HCN
R1 R2 R1 R1 R2
R2
Scheme 23.3 Enantioselective HNL-catalyzed decomposition of racemic cyanohydrins: R1 ¼ alkyl,

cycloalkyl, aryl, heteroaryl; R2 ¼ H, alkyl.
More recently, almond meal was used for the resolution of rac-2-hydroxy-2-
phenylpropanenitrile. Under optimized conditions, (S)-2-hydroxy-2-phenylpropa-
nenitrile, as the less reactive enantiomer, was obtained in 98–99% e.e. at approx-
imately 50% conversion [165]. In a similar way the (S)-cyanohydrin was afforded
from racemic 2-methyl-2-hydroxyhexanenitrile with P. amygdalus HNL in more than
90% e.e. [128, 137].
23.8.2
Esterase or Lipase as Catalyst
Application of hydrolytic enzymes is realized in three different systems: enzymatic

hydrolysis or transesterification of racemic cyanohydrin esters (Scheme 23.4) and
enzymatic acylation of racemic cyanohydrins. The latter types of biotransformations
incorporate one of the very first reports of enzymatic dynamic kinetic resolution,
because of the easy in situ racemization of unreacted cyanohydrins under alkaline
conditions [223–226].
R2O CN OR2 OR3

HNL CN
2 + R3OH + NC + R2OH
R1 H R1 H H R1
Scheme 23.4 Lipase-catalyzed formation of optically enriched cyanohydrins: R1 ¼ alkyl, cycloalkyl,

aryl, heteroaryl; R2 ¼ acyl; R3 ¼ H, acyl.
A series of cyanohydrin acetates with an e.e. up to 98% has been prepared by

enzymatic hydrolysis of their racemic acetates in the presence of an esterase from
974
Pseudomonas sp. [227]. Lipoprotein lipase from Pseudomonas sp. catalyzed irreversible
transesterification using enol esters in the resolution of different aromatic cyanohy-
drins [228–230].
The enantioselective hydrolysis of the racemic acetate by Arthrobacter lipase gave
the optically pure (S)-3-phenoxybenzaldehyde cyanohydrin. The unhydrolyzed (R)-
acetate was re-racemized by heating with triethylamine and submitted again to
enzymic hydrolysis [231]. In addition, the resolution of the racemic acetate ester of
the cyanohydrin of 3-phenoxybenzaldehyde using a highly enantioselective lipase
from Pseudomonas sp. was carried out with an e.e. of >96% [232]. Both the
cyanohydrin esters and the free cyanohydrins (which are prone to racemization)
can be isolated as enantiomers with high optical purity (e.e. 97%) on a preparative
scale by the hydrolysis of the racemic butyrates with Candida cylindracea lipase and
Pseudomonas sp. lipase [233].
For the kinetic resolution of a,a-disubstituted cyanohydrin acetates, to obtain
tertiary alcohols, the protease subtilisin A ((S)-selective) and Candida rugosa lipase
((R)-selective) were utilized. The enantiomeric ratio E is moderate [234] also in the
case of esterase BS2 from Bacillus subtilis [235]. Recently, this limitation could be
overcome by the availability of enzymes from the metagenome with an amino acid
motif within the active site, which is more suitable for bulky substrates through
directed evolution or by rational protein design [236].
A one-pot synthesis of optically active cyanohydrin acetates from aldehydes has
been accomplished by lipase-catalyzed kinetic resolution coupled with in situ
formation and racemization of cyanohydrins in an organic solvent. Racemic cyano-
hydrins, generated from aldehydes and acetone cyanohydrin in diisopropyl ether
under the catalysis of a basic anion-exchange resin, were acetylated enantioselectively
by a lipase from Pseudomonas cepacia (Amano) with isopropenyl acetate as the
acylating reagent. The (S)-cyanohydrin was preferentially acetylated by the lipase,
while the unreacted (R)-isomer was continuously racemized through reversible
transhydrocyanation catalyzed by the resin. These processes consequently led to a
one-pot conversion with up to 94% e.e. in 63–100% conversion yields [237, 238].
The Pseudomonas aeruginosa lipase (immobilized on Hyflo Super-Cel) catalyzed the
kinetic resolution of rac-2-(acetyloxy)-2-(pentafluorophenyl)acetonitrile, yielding
enantiomerically pure cyanohydrin and its antipodal ester [239–241]. By immobili-
zation of lipase B from Candida antarctica (CalB) on Celite the enantioselective
synthesis of aromatic and heteroaromatic cyanohydrin esters could be improved in
terms of enantiopurity and reaction time for the dynamic kinetic resolution [242].
The diastereoselectivity of HNLs from almond and Hevea brasiliensis was inves-
tigated by Riva and Griengl for a-alkoxy and a,b-dialkoxy substituted aldehydes.
Thereby, the syn diastereomers could be enriched [243]. The diastereoselectivity of
MeHNL was investigated by Effenberger in 2005 for chiral 4-alkylcyclohexanones. In
this case the syn diastereomers were formed almost quantitatively [244], while in the
case of 2- and 3-substituted cyclohexanones the selectivities were not that high [245].
With PaHNL, 2-alkylcyclohexanone cyanohydrins show high (R)-selectivity, when
alkyl is larger than C1 (chain or branched), whereas 2-methylcyclohexanone yields the
(S)-product, while the cis/trans ratio is almost 1 : 1. With MeHNL the products formed
23.9 Follow-Up Chemistry of Enantiomerically Pure Cyanohydrins j975
show all (S)-selectivity and also here the diastereoselectivity is only moderate. The
catalytic activity of both enzymes decreases with increasing size of the alkyl sub-
stituents. The diastereoselectivity for the formation of 2- and 3-alkoxy-cyclohexanone
cyanohydrins was only moderate as well. The investigations were extended to 2-
substituted cyclopentanone substrates [246]. Effenberger and coworkers also
reported interesting results with aldehydes bearing stereogenic centers adjacent to
the CHO group with MeHNL mutein W128A. An inversion of stereoselectivity was
obtained with a d.e. 96% for (2S,3R)- and a d.e. of 80% for the (2S,3S)-cyanohydrin
from (R)-2-phenylpropanal. The experimental data were explained and rationalized
with crystal-structure-based molecular modeling [247]. Furthermore, the enzymatic
preparation of enantiomerically or diastereomerically enriched aromatic and non-
aromatic polycyclic cyanohydrins has been investigated. While HCN addition
catalyzed by HNLs of Prunus amygdalus and Hevea brasiliensis gave good results
with bicyclic aldehydes, the biocatalytic enantio- or diastereoselective acylation of
racemic cyanohydrins by hydrolases (lipases and proteases) proved to be a more
versatile methodology for aromatic and non-aromatic polycyclic aldehydes to obtain
the corresponding cyanohydrins [248].
23.9
Follow-Up Chemistry of Enantiomerically Pure Cyanohydrins
Enantiopure cyanohydrins are important synthetic building blocks for a numerous

follow-up-products [1, 4–6, 31, 34, 37, 38, 137, 138, 249–254], as can be seen from the
selected examples shown in Scheme 23.5. Both functional groups, the hydroxy and
the cyanide moiety, can easily be converted into a large number of chiral inter-
mediates such as a-hydroxy acids and esters, a-hydroxy aldehydes and ketones,
b-amino alcohols, and a-fluorocyanides to give value-added products like drugs,
agrochemicals, flavors, and fragrances.
All the transformations described in the following are (with almost no exception)
stereoselective and racemization free transformations so that the absolute config-
uration of the chiral cyanohydrins is conserved in the products of the follow-up
chemistry. Chemical [255] and enzymatic [256, 257] hydrolysis of the nitrile group
gives access to a-hydroxy carboxylic acids and amides. The O-protected a-hydroxy
aldehydes can be prepared by hydrogenation of O-protected cyanohydrins with a mild
reducing agent like diisobutylaluminum hydride (DIBAlH) followed by a mild
hydrolysis [258] or by reaction with Grignard reagents. These compounds can be
further converted into amino alcohols [259] and polyols [109]. The reaction of O-
protected cyanohydrins with Reformatsky reagents and subsequent mild hydrolysis
of the latter yields enamines, which can further be converted into tetronic acids by
treatment with strong acids [260, 261]. The stereoselective synthesis of b-amino-
c-butyrolactones was performed by addition of allyl Grignard reagents to O-protected
chiral cyanohydrins and subsequent reduction [262]. b-Amino alcohols can be
synthesized starting from cyanohydrins or O-protected derivatives with LiAlH4 [255].
The hydroxyl functionality of a cyanohydrin can be substituted by a great number of
976
H H
TBDMSO HO * CH2R2
R * R
* H3
H O NHR
HO * R2
R c d
* H b
NH2
OTBDMS OR3
r
* H * H
F R R
CN e CHO
* H
R OH
CN a
q *H i
R
OTMS OH NH2
* H
p * H OR3
R R *H
CN CN R
COOR2
f
j OH h
N3 * H
R
*R OSO2R2 COOH g
H k OH
CN * H * H
R o R
CN COOR2
l n
H
R m OAc
* *R
N H
CN
H NPht
*R
H
CN
Scheme 23.5 Selected follow-up reactions of 78 C, conc. HCl, MeOH [321]; (j) R2SO2Cl/
optically pure cyanohydrins: (a) TBDMSCl/ pyr [80]; (k) KN3/DMF [263]; (l) LAH/ether/
imidazole [120, 121]; (b) R2MgX/ether, NaBH4, 80 C, phosphate buffer pH 7.0/ 70
H3O þ [143, 319]; (c) CH3MgI/ether,
C [263]; (m) potassium phthalimide/
H3O þ [120]; (d) R2CH2MgI/ether, MeOH, DMF [263]; (n) KOAc/DMF [80, 263, 322]; (o)
R3NH2, NaBH4 [320]; (e) LAH (lithium conc. HCl or lipase [80, 263, 322]; (p) Me3SiCl
aluminium hydride) [143]; (f) H3O þ [47]; (g) (TMSCl)/pyr/ether/0 to 25 C [272]; (q) DAST/
R2OH/CHCl3/wolfatite [321]; (h) R3Cl/NaI/ CH2Cl2/–80 to 25 C [272]; (r) DIBAlH/CH2Cl2/
CH3CN/pyr/0 C [321]; (i) DIBAlH/hexane/ 78 C, 0.5M H2SO4 [323].
nucleophiles after activation (e.g., sulfonylation) [263, 264]. The Mitsunobu reaction
is another possibility for HO substitution [265]. 2-Azidonitriles can be hydrogenated
selectively to a-aminonitriles and 1,2-diamines [266]. The substitution of O-activated
cyanohydrins with K-phthalimide gives access to a-amino acids after deprotec-
tion [263]. Sulfur nucleophiles yield a-mercaptonitriles and b-amino thiols after
hydrogenation, which can be used as complexing agents for metal ions in chiral
catalysts or as starting materials for S-containing heterocycles [267]. Gotor and
coworkers performed an enzymatic cyanohydrin synthesis of v-bromoaldehydes to
obtain precursors for the synthesis of chiral 2- and 2,3-substituted piperidines [138,
268, 269], azepan-3-ol, and azocan-3-ol [137] as well as chiral 2-cyano-tetrahydrofuran
23.10 Experimental Techniques for HNL-Catalyzed Biotransformations and Safe Handling of Cyanides j977
and -tetrahydropyran [270]. The authors extended the substrate range to v-alkox-
yaldehydes [55]. 5-Hydroxypiperidin-2-one derivatives were prepared starting from
chiral cyanohydrins [271]. Trimethylsilyl derivatives of cyanohydrins are precursors
for the introduction of fluorine by (diethylamino)sulfur trifluoride (DAST) [272]. A
chemoenzymatic synthesis was developed for Fmoc-protected (2S,3S)-2-hydroxy-3-
amino acids, starting from 2-furaldehyde [273], that were used for solid-phase
synthesis of a-hydroxylated b-oligopeptides without protection at the hydroxyl
function [274], where the key step to chirality is the (R)-HNL catalyzed cyanohydrin
synthesis. The same concept was applied for the stereoselective synthesis of (2R,5R)-
and (2S,5R)-5-hydroxylysine [275]. The ferrocenyl-containing amino alcohols derived
from HbHNL-catalyzed synthesis of the (S)-cyanohydrin from formylferrocene were
converted into ferrocenyl-oxazolidinones, which proved to be effective chiral aux-
iliaries for asymmetric alkylations and aldol reactions [276]. Some selected examples
are shown in Scheme 23.5.
Chemoenzymatic syntheses of D- and L-sphingosines as well as L-2-deoxypentono-
1,4-lactones and L-2-deoxypentoses with a HNL-catalyzed step at the start of the
synthetic strategy have been reported [277, 278], and recently a de novo synthesis of
D- and L-pentoses via a cyanohydrin intermediate as the key step was established [279].
The presence of unsaturation in the cyanohydrin side chain was shown to make
these compounds potential starting materials in intramolecular Diels–Alder reac-
tions, especially when a furan ring is present [280].
A new approach to convert cyanohydrins into follow-up products is to combine the
HNL with other enzymes in a one-step process. This approach can be realized by
using a one-pot bi-enzymatic cascade of immobilized enzymes [281], whole cell
systems with co-expressed [282] enzymes or as (combi-)CLEAs [283–287].
23.10
Experimental Techniques for HNL-Catalyzed Biotransformations and Safe Handling
of Cyanides
Reaction conditions have a great influence on the outcome of the enzymatic

cyanohydrin reactions. The ratio of the enzymatic to the non-enzymatic cyanohy-
drin synthesis is responsible for the enantiopurity of the product. Performing the
reaction at low-pH values suppresses the non-enzymatic reaction to a higher extent
than the enzymatic reaction and therefore improves e.e.s [288]. In contrast,
decreasing pH normally decreases the enzyme activity due to its inactivation.
Thus, the possibility of enhancing the enzymatic reaction and suppressing the
non-enzymatic reaction is the application of a small pH range in the process.
Temperature is another important factor in process optimization. The non-enzy-
matic reaction is diffusion limited and therefore directly affected by temperature
variations. At lower temperatures the non-enzymatic reaction is suppressed in a
larger extent than the enzymatic reaction [289]. The choice of buffer system is
another important factor for the cyanohydrin reaction, for example, acetate is an
inhibitor for MeHNL [79].
978
23.10.1
HNL Catalysis in Aqueous Medium
Reaction in aqueous solution is performed with an appropriate acidic component and

alkali cyanide for in situ development of the required HCN. The greatest disadvantage
of the aqueous system for the cyanohydrin reaction is the undesired chemical
addition of HCN to the carbonyl compound, which results in lower e.e. values of
the product. This drawback can be reduced by lowering the pH below 4.5, where
almost no spontaneous chemical cyanohydrin reaction occurs [139]. The following
procedure is a typical example [85].
A crude cytosolic extract (1 ml) of (S)-HNL from Hevea brasiliensis (100 IU ml1)
was added to a stirred solution of aldehyde (1 mmol) in 1.7 ml of 0.1 M sodium citrate
buffer (pH 4.0) and the resultant mixture was cooled to 0 C. Subsequently, 2 mmol of
potassium cyanide adjusted to pH 4.0 with cold 0.1 M citric acid (17 ml) was added in
one portion. After stirring for 1 h at 0–5 C, the reaction mixture was extracted with
methylene chloride (3 50 ml). The combined organic layers were dried over
anhydrous sodium sulfate and the solvent was removed to give the crude cyanohy-
drin. This was then purified by column chromatography on silica gel using petro-
leum ether–ethyl acetate acidified with trace amounts of anhydrous HCl as the eluent.
The limitation of performing the reaction in aqueous medium is the low water
solubility of most aldehydes and ketones. The addition of water-miscible organic
solvents to the reaction mixture leads to increased reactant solubility, but may also
affect the enzyme activity and stability in a negative manner [290].
23.10.2
HNL Catalysis in Organic Medium
An important breakthrough for enzymatic cyanohydrin production was achieved by

performing the transformation in water-immiscible organic solvents. It has been
observed that there is virtually no spontaneous chemical addition of HCN to the
carbonyl moiety [47, 124–126, 175, 255, 291–295], which is responsible for a reduced
e.e. in the product. The enzymes show no loss of activity in, for example, diisopropyl
ether. The application of organic solvents under micro-aqueous conditions has been
reported [77, 131]. For reactions carried out in organic solvents, it is advantageous to
apply the enzyme adsorbed on a suitable support, which facilitates the work-up
procedure and the catalyst can be reused. Aldercreutz and coworkers performed
detailed investigations of the effects of solvent, water activity, and temperature on the
enantioselectivity of HNLs. They showed that the enantioselectivity increased with
decreasing reaction temperature. In addition, increasing water content up to water
saturation of the organic solvent showed a positive impact on the enantioselectivity,
while the log P value of the solvent did not influence it [296]. Faber and coworkers
investigated the PaHNL activity in ten different organic solvent with log Ps from 1.5
to þ 6.5 [297]. A representative protocol for cyanohydrin formation in organic
solvents with immobilized hydroxynitrile lyase is the following [80].
23.10 Experimental Techniques for HNL-Catalyzed Biotransformations and Safe Handling of Cyanides j979
A suspension of Avicel cellulose (0.5 g) in 0.05 mM phosphate buffer (pH 4.5,
10 ml) containing ammonium sulfate (4.72 g) was stirred for 1 h, and a solution of
(S)-HNL from Sorghum bicolor (50 ml, 1000 IU ml1, specific activity 70 IU mg1) was
added. The mixture was stirred at room temperature for 10 min and filtered, and the
immobilized enzyme was suspended in diisopropyl ether (10 ml). After addition of
aldehyde (2 mmol) and dry liquid HCN (300 ml, 7.5 mmol), the mixture was stirred
until all aldehyde had reacted. After removal of the immobilized enzyme, the filtrate
was concentrated to yield the crude cyanohydrin.
23.10.3
HNL Catalysis in Biphasic Medium
Biphasic solvent mixtures were reported employing (R)-hydroxynitrile lyase [140,

222], (S)-HNL from Hevea brasiliensis [83, 159], the application of micro-channels for
the synthesis of enantiopure cyanohydrins [298, 299], and organic solvents under
micro-aqueous conditions [77, 131, 300]. The biphasic system has also been adapted
to a continuous flow reactor. A pre-mixed solution of aldehyde and HCN in wet
diisopropyl ether was pumped through a column filled with almond meal [301]. In an
organic–aqueous biphasic system the substrate is supposed to diffuse from the
organic phase where it is dissolved into the aqueous phase, where the enzyme is
dissolved and the conversion takes place. After the transformation the product will
diffuse back into the organic phase. Such a multiphase system shares many
advantages: first the accumulation of organic compounds in the organic phase
leading to an easier separation procedure, and second lower reactant concentrations
in the aqueous phase leading to reduced inhibition effects by substrate and/or
product. A disadvantage might be the low substrate concentration in the aqueous
phase due to high partition coefficients, leading to low reaction rates. At the same
time an increased water solubility of substrates and products favors the back reaction
(decomposition of cyanohydrins), leading to reduced enantiopurity. For optimal
reaction the phases must be mixed vigorously to increase the interface [174]; tert-butyl
methyl ether is the most suitable solvent [140]. Lowering pH and temperature also
decreases the non-enzymatic reaction to a higher extent than the enzyme-catalyzed
transformation (for PaHNL pH 5.5 and 5 C) [177]. By applying the mass transfer
limitation principle in a two-phase system improved results were gained for difficult
substrates [302]. In principle biphasic systems need only a simple experimental set-
up and show general applicability [303]. The successful application of ionic liquids
(ILs) in the HNL-catalyzed cyanohydrin formation was reported for both Pa- and
HbHNL. The enantioselectivity remained untouched but the reaction rate was
significantly increased when comparing organic–aqueous with IL-aqueous biphasic
systems [304], while Zong and coworkers reported the increase of both activity and
selectivity for PaHNL but inactivation of MeHNL in ILs compared to biphasic
systems [305].
A typical procedure is as follows [140]: Freshly distilled benzaldehyde (37.1 g,
0.35 mmol), HCN (12.2 g, 0.45 mmol), and (R)-hydroxynitrile lyase (78 mg) were
980
dissolved in 225 ml of methyl t-butyl ether (MTBE) and 250 ml of citrate buffer
(50 mM, pH 5.5) at 22 C. After stirring for 20 min the MTBE layer was separated and
the aqueous layer was extracted once with 25 ml of MTBE. The combined organic
layers were dried over MgSO4, filtered, and concentrated under reduced pressure;
yield: 45.2 g (97%), purity 98%, e.e. The aqueous layer was reused in a series of four
consecutive experiments using the same amounts of reagents in the organic phase. A
total of 185.5 g of benzaldehyde was converted into 226 g of (R)-mandelonitrile using
78 mg of (R)-hydroxynitrile lyase (0.035 wt.%).
23.10.4
Transhydrocyanation for HCN Generation
An alternative method of employing organic solvents that allows the safe use of HCN
is transhydrocyanation [127, 128, 136–138, 166, 306]. An example of cyanohydrin
formation using acetone cyanohydrin as the cyanide source is given in the following
procedure [136].
(R)-Hydroxynitrile lyase buffer solution (0.5 ml) (10 mg ml1, 0.4 mol l1 acetate
buffer, pH 5.0) was added to a solution of 120 mg (1 mmol) of phenylacetaldehyde and
110 mg (1.3 mmol) of acetone cyanohydrin in 11 ml of diethyl ether at 23 C. The
mixture was stirred for 18 h at 23 C and then diluted with 50 ml of ether. The aqueous
phase was extracted with 2 10 ml of ether and the combined organic phases were
dried over anhydrous magnesium sulfate. Evaporation of solvent gave a pale amber
liquid that was purified by flash chromatography on a silica gel column in ethyl
acetate–benzene–dichloromethane (1: 30: 50) to afford 122 mg (83%) of cyanohydrin,
88% e.e.
Hydrogen cyanide smells like bitter almonds, although many people cannot smell
it at all. Cyanide is a fast-acting poison in the human body; the ability to block the
intracellular respiratory chain is the main reason for the high toxicity of hydrogen
cyanide. Severe breathing difficulties develop very rapidly when cyanide is swal-
lowed, inhaled, or absorbed through the skin. Cyanide poisoning symptoms in the
early stages include general weakness, breathing difficulty, headache, nausea,
giddiness, vomiting, the victims breath smell like bitter almonds, and irritation of
the nose, mouth, and throat occurs. Hydrogen cyanide is liberated by the addition of
acid to cyanide compounds.
The TLV (threshold limit value) for HCN is 11 mg m3 or 10 ppm [307]. This limit
includes the potential contribution of skin absorption to the overall exposure.
Proper gloves should be worn when handling dry sodium cyanide. Rubber gloves
and splash-proof goggles should also be worn when substantial amounts of sodium
cyanide solution are used. All reaction equipment in which cyanides are used or
produced should be placed in well-ventilated hoods, and it should be determined
immediately whether anyone has been exposed to cyanide vapors or liquid splash-
ing [308–310].
Vapor-detector tubes sensitive to 1 ppm of HCN are available commercially. The
presence of free cyanide ion in aqueous solution may be detected by treating an
aliquot of the sample with ferrous sulfate and an excess of sulfuric acid. A precipitate
References j981
of Prussian blue indicates that free cyanide ion is present. More sophisticated for
continuous warning is the use of electrochemical sensors for HCN detection.
Waste solutions containing cyanides treated with sodium hypochlorite are con-
verted into harmless cyanate, which can be further processed to ammonia and carbon
dioxide by addition of dilute sulfuric acid to pH 7. Surplus HCN gas can be
neutralized by aqueous sodium hydroxide and then oxidized. Caution has to be
advised with liquid hydrogen cyanide because bases, including sodium hydroxide
and sodium cyanide, may initiate a violent polymerization [307].
Explosive hazards can occur on exposure of HCN to air in the presence of sources
of ignition (flammable limits in air: 5.6–40 vol.%), including heat (polymerizes
explosively at 50–60 C), and when HCN is stored for long periods of time.
23.10.5
Technical Applications
For technical applications the readers is referred to Chapter 24 (entitled Industrial

Applications).
23.11
Summary and Outlook
The enzymatic synthesis of enantiopure cyanohydrins has been brought to a high

stage of development. Both (R)- and (S)-cyanohydrins are accessible for a broad
variety of substrates in, as a rule, excellent yield and enantiopurity. Following recent
progress in overexpression, HNLs are also available in quantities needed for
industrial production. The procedures for safe handling of cyanides are well
established so that they do not restrict the exploitation of HNLs.
References
1 Gregory, R.J.H. (1999) Chem. Rev., 99, Stereoselective Biocatalysis (ed. R.N. Patel)
3649–3682. Marcel Dekker Inc., New York/Basel,
2 North, M. (2003) Tetrahedron: Asymmetry, pp. 321–342.
14, 147–176. 6 van der Gen, A. and Brussee, J. (2000)
3 Chen, F.-X. and Feng, X. (2006) Curr. Org. Stereoselective biocatalytic formation of
Synth., 3, 77–97. cyanohydrins, versatile building blocks
4 Brussee, J. and van der Gen, A. (2000) for organic synthesis, in Enzymes in
Biocatalysis in the enantioselective Action (eds B. Zwanenburg, M.
formation of chiral cyanohydrins, Mikolajczyk, and P. Kielbasinski),
valuable building blocks in organic Kluwer Academic Publishers, The
synthesis, in Stereoselective Biocatalysis Netherlands, pp. 365–396.
(ed. R.N. Patel), Marcel Dekker Inc., New 7 Holt, J. and Hanefeld, U. (2009) Curr.
York/Basel, pp. 289–320. Org. Synth., 6, 15–37.
5 Effenberger, F. (2000) Hydroxynitrile 8 Brunel, J.-M. and Holmes, I.P. (2004)
lyases in stereoselective synthesis, in Angew. Chem., Int. Ed., 43, 2752–2778.
982
9 Achard, T.R.J., Clutterbuck, L.A., Rosenthal and M.R. Berenbaum),

and North, M. (2005) Synlett, 12, Academic Press, New York, pp. 35–77.
1828–1847. 30 Lechtenberg, M. and Nahrstedt, A. (1999)
10 Khan, N.H., Kureshy, R.I., Abdi, S.H.R., Cyanogenic glycosides, in Naturally
Agrawal, S., and Jasra, R.V. (2008) Coord. Occurring Glycosides, (ed. R. Ikan),
Chem. Rev., 252, 593–623. John Wiley & Sons, Ltd., Chichester,
11 North, M., Usanov, D.L., and Young, C. pp. 147–191.
(2008) Chem. Rev., 108, 5146–5226. 31 Effenberger, F., Foerster, S., and
12 Jones, P.R., Andersen, M.D., Nielsen, Kobler, C. (2007) State of the art and
J.S., Høj, P.B., and Møller, B.L. (2000) applications in stereoselective synthesis
Recent Adv. Phytochem., 34, 191–247. of chiral cyanohydrins, in Biocatalysis in
13 Dewick, P.M. (1984) Nat. Prod. Rep., 1, the Pharmaceutical and Biotechnology
545–549. Industries (ed. R. Patel), CRC Press, Boca
14 Poulton, J.E. (1990) Plant Physiol., 94, Raton, pp. 677–698.
401–405. 32 Sharma, M., Sharma, N.N., and Bhalla,
15 Rosenthaler, L. (1908) Biochem. Z., 14, T.C. (2005) Enzyme Microb. Technol., 37,
238–253. 279–294.
16 Krieble, V.K. and Wieland, W.A. (1921) 33 Fechter, M.H. and Griengl, H. (2004)
J. Am. Chem. Soc., 43, 164–175. Food Technol. Biotechnol., 42, 287–294.
17 Albers, H. and Hamann, K. (1932) 34 Effenberger, F., F€orster, S., and Wajant,
Biochem. Z., 255, 44–65. H. (2000) Curr. Opin. Biotechnol., 11,
18 Albers, H. and Hamann, K. (1934) 532–539.
Biochem. Z., 269, 14–25. 35 Johnson, D.V., Zabelinskaja-Mackova,
19 Becker, W., Benthin, U., Eschenhof, E., A.A., and Griengl, H. (2000) Curr. Opin.
and Pfeil, E. (1963) Biochem. Z., 337, Chem. Biol., 4, 103–109.
156–166. 36 Seoane, G. (2000) Curr. Org. Chem., 4,
20 Becker, W. and Pfeil, E. (1964) 283–304.
Naturwissenschaften, 51, 193. 37 Johnson, D.V. and Griengl, H. (1999) Adv.
21 Becker, W., Freund, H., and Pfeil, E. Biochem. Eng./Biotechnol., 63, 31–55.
(1965) Angew. Chem., Int. Ed. Engl., 38 Sukumaran, J. and Hanefeld, U. (2005)
4, 1079. Chem. Soc. Rev., 34, 530–542.
22 Becker, W. and Pfeil, E. (1966) Biochem. 39 Liu, Z., Weis, R., and Glieder, A. (2004)
Z., 346, 301–321. Food Technol. Biotechnol., 42, 237–249.
23 Becker, W. and Pfeil, E. (1966) J. Am. 40 Mugford, P.F., Wagner, U.G., Jiang, Y.,
Chem. Soc., 88, 4299–4300. Faber, K., and Kazlauskas, R.J. (2008)
24 Akazawa, T., Miljanich, P., and Conn, Angew. Chem., Int. Ed., 47, 8782–8793.
E.E. (1960) Plant Physiol., 35, 535–538. 41 Hickel, A., Hasslacher, M., and Griengl,
25 Bove, C. and Conn, E.E. (1961) J. Biol. H. (1996) Physiol. Plant., 98, 891–898.
Chem., 236, 207–210. 42 Jorns, M.S. (1979) J. Biol. Chem., 254,
26 Seely, M.K., Criddle, R.S., and Conn, E.E. 12145–12152.
(1966) J. Biol. Chem., 241, 4457–4462. 43 Wajant, H. and Effenberger, F. (1996)
27 Mao, C.-H. and Anderson, L. (1967) Biol. Chem., 377, 611–617.
Phytochemistry, 6, 473–483. 44 Hasslacher, M., Kratky, C., Griengl, H.,
28 Conn, E.E. (1981) Cyanogenic glycosides, Schwab, H., and Kohlwein, S.D. (1997)
in The Biochemistry of Plants: a Proteins: Struct., Funct., Genet., 27,
Comprehensive Treatise, vol. 7 (eds P.K. 438–449.
Stumpf and E.E. Conn), Academic Press, 45 Gruber, K. and Kratky, C. (2004) J. Polym.
New York, pp. 479–500. Sci. Pol. Chem., 42, 479–486.
29 Seigler, D.S. (1991) Cyanide and 46 Effenberger, F. (1994) Angew. Chem., Int.
cyanogenic glycosides, in Hervivores: Ed. Engl., 33, 1555–1564.
Their Interactions with Secondary Plant 47 Effenberger, F., H€orsch, B., Weingart, F.,
Metabolites: Volume 1 The Chemical Ziegler, T., and K€ uhner, S. (1991)
Participants, 2nd edn, (eds G.A. Tetrahedron Lett., 32, 2605–2608.
References j983
48 Nanda, S., Kato, Y., and Asano, Y. (2005) 65 Solıs, A., Luna, H., P
erez, H.I.,
Tetrahedron, 61, 10908–10916. Manjarrez, N., and Sanchez, R. (2002)
49 Gregory, R.J.H., Roberts, S.M., Barkley, Rev. Mex. Cienc. Farm., 33, 27–31.
J.V., Coles, S.J., Hursthouse, M.B., and 66 Lee, T. and Ahn, Y. (2002) Bull. Korean
Hibbs, D.E. (1999) Tetrahedron Lett., 40, Chem. Soc., 23, 1490–1492.
7407–7411. 67 Poechlauer, P., Skranc, W., and
50 Kiljunen, E. and Kanerva, L.T. (1997) Wubbolts, M. (2004) The large-scale
Tetrahedron: Asymmetry, 8, biocatalytic synthesis of enantiopure
1225–1234. cyanohydrins, in Asymmetric Catalysis on
51 Kiljunen, E. and Kanerva, L.T. (1997) Industrial Scale: Challenges, Approaches
Tetrahedron: Asymmetry, 8, 1551–1557. and Solutions (eds H.U. Blaser and E.
52 Bhunya, R., Jana, N., Das, T., and Schmidt), Wiley-VCH Verlag GmbH,
Nanda, S. (2009) Synlett, 1237–1240. Weinheim, pp. 151–164.
53 Bhunja, R., Mahapatra, T., and Nanda, S. 68 Purkarthofer, T., Skranc, W., Schuster, C.,
(2009) Tetrahedron: Asymmetry, 20, and Griengl, H. (2007) Appl. Microbiol.
1526–1530. Biotechnol., 76, 309–320.
54 Lin, G., Han, S., and Li, Z. (1999) 69 Rosenthaler, L. (1913) Arch. Pharm., 251,
Tetrahedron, 55, 3531–3540. 56–84.
55 de Gonzalo, G., Brieva, R., and Gotor, V. 70 Xu, L.L., Singh, B.K., and Conn, E.E.
(2002) J. Mol. Catal. B: Enzym., 19–20, (1988) Arch. Biochem. Biophys., 263,
223–230. 256–263.
56 Weis, R., Poechlauer, P., Bona, R., 71 Cutler, A.J., Sternberg, M., and
Skranc, W., Luiten, R., Wubbolts, M.G., Conn, E.E. (1985) Arch. Biochem.
Schwab, H., and Glieder, A. (2004) J. Mol. Biophys., 238, 272–279.
Catal. B: Enzym., 29, 211–218. 72 Albrecht, J., Jansen, I., and Kula, M.R.
57 Glieder, A., Weis, R., Skranc, W., (1993) Biotechnol. Appl. Biochem., 17,
Poechlauer, P., Dreveny, I., Majer, S., 191–203.
Wubbolts, M., Schwab, H., and 73 Roberge, C., Fleitz, F., Pollard, D., and
Gruber, K. (2003) Angew. Chem., Int. Ed., Devine, P. (2007) Tetrahedron: Asymmetry,
42, 4815–4818. 18, 208–214.
58 Dreveny, I., Gruber, K., Glieder, A., 74 Trummler, K., Roos, J., Schwaneberg, U.,
Thompson, A., and Kratky, C. (2001) Effenberger, F., F€orster, S.,
Structure (Cambridge), 9, 803–815. Pfizenmaier, K., and Wajant, H. (1998)
59 Pscheidt, B., Avi, M., Gaisberger, R., Plant Sci., 139, 19–27.
Hartner, F.S., Skranc, W., and Glieder, A. 75 Breithaupt, H., Pohl, M., B€onigk, W.,
(2008) J. Mol. Catal. B: Enzym., 52–53, Heim, P., Schimz, K.-L., and Kula, M.-R.
60 Hernandez, L., Luna, H., Ruiz-Teran, F., 315–332.
and Vazquez, A. (2004) J. Mol. Catal. B: 76 Wajant, H., F€orster, S., Selmar, D.,
Enzym., 30, 105–108. Effenberger, F., and Pfizenmaier, K.
61 Hernandez, L., Luna, H., Solıs, A., and (1995) Plant Physiol., 109, 1231–1238.
Vazquez, A. (2006) Tetrahedron: 77 Han, S.-Q., Ouyang, P.-K., Wei, P., and
Asymmetry, 17, 2813–2816. Hu, Y. (2006) Biotechnol. Lett., 28,
62 Solıs, A., Luna, H., Manjarrez, N., and 1909–1912.
Perez, H. (2004) Tetrahedron, 60, 78 Andexer, J., von Langermann, J., Mell, A.,
10427–10431. Bocola, M., Kragl, U., Eggert, T., and
63 Solıs, A., Luna, H., Perez, H.I., Pohl, M. (2007) Angew. Chem., Int. Ed.,
Manjarrez, N., and Sanchez, R. (1998) 46, 8679–8681.
Rev. Soc. Quim. Mex., 42, 214–216. 79 Guterl, J.-K., Andexer, J.N., Sehl, T., von
64 Solıs, A., Luna, H., Perez, H.I., Langermann, J., Frindi-Wosch, I.,
Manjarrez, N., Sanchez, R., Albores- Rosenkranz, T., Fitter, J., Gruber, K.,
Velasco, M., and Castillo, R. (1998) Kragl, U., Eggert, T., and Pohl, M. (2009)
Biotechnol. Lett., 20, 1183–1185. J. Biotechnol., 141, 166–173.
984
80 Effenberger, F., H€
orsch, B., F€orster, S., 97 Avi, M., Wiedner, R.M., Griengl, H., and
and Ziegler, T. (1990) Tetrahedron Lett., 31, Schwab, H. (2008) Chem.–Eur. J., 14,
1249–1252. 11415–11422.
81 Niedermeyer, U. and Kula, M.R. (1990) 98 Veum, L., Hanefeld, U., and Pierre, A.
Angew. Chem., Int. Ed. Engl., 29, 386–387. (2004) Tetrahedron, 60, 10419–10425.
82 Hasslacher, M., Schall, M., Hayn, M., 99 Sheldon, R.A., Schoevaart, R., and
Griengl, H., Kohlwein, S.D., and L.M van. Langen, (2006) Cross-linked
Schwab, H. (1996) J. Biol. Chem., 271, enzyme aggregates, in Methods in
5884–5891. Biotechnology, Immobilization of
83 Griengl, H., Hickel, A., Johnson, D.V., Enzymes and Cells, 2nd edn (ed. J.M.
Kratky, C., Schmidt, M., and Schwab, H. Guisan) Humana Press, Totowa, NJ,
(1997) Chem. Commun., 1933–1940. pp. 31–45.
84 Hasslacher, H., Schall, M., Hayn, M., 100 Gaisberger, R.P., Fechter, M.H., and
Bona, R., Rumbold, K., Luckl, J., Griengl, H. (2004) Tetrahedron:
Griengl, H., Kohlwein, S.D., and Schwab, Asymmetry, 15, 2959–2963.
H. (1997) Protein Expression Purif., 11, 101 Purkarthofer, T., Papst, T., van, C.,
61–71. Broek, d., Griengl, H., Maurer, O., and
85 Schmidt, M., Herve, S., Klempier, N., and Skranc, W. (2006) Org. Process Res. Dev.,
Griengl, H. (1996) Tetrahedron, 52, 10, 618–621.
7833–7840. 102 Lauble, H., Miehlich, B., F€orster, S.,
86 Klempier, N., Griengl, H., and Hayn, M. Wajant, H., and Effenberger, F. (2001)
(1993) Tetrahedron Lett., 34, 4769–4772. Protein Sci., 10, 1015–1022.
87 Danieli, B., Frattini, S., Roda, G., Carrea, 103 Lauble, H., F€orster, S., Miehlich, B.,
G., and Riva, S. (1998) J. Mol. Catal. B: Wajant, H., and Effenberger, F. (2001)
Enzym., 5, 223–228. Acta Cryst., Sect. D, 57, 194–200.
88 Roda, G., Riva, S., and Danieli, B. (1999) 104 Daußmann, T., Rosen, T.C., and
Tetrahedron: Asymmetry, 10, 3939–3949. D€unkelmann, P. (2006) Eng. Life Sci., 6,
89 Roda, G., Riva, S., Danieli, B., Griengl, 125–129.
H., Rinner, U., Schmidt, M., and 105 F€orster, S., Roos, J., Effenberger, F.,
Zabelinskaja-Mackova, A.A. (2002) Wajant, H., and Sprauer, A. (1996) Angew.
Tetrahedron, 58, 2979–2983. Chem., Int. Ed. Engl., 35, 437–439.
90 Avi, M., Fechter, M.H., Gruber, K., Belaj, 106 Hughes, J., Carvalho, F.J.P.D.C., and
F., P€
ochlauer, P., and Griengl, H. (2004) Hughes, M.A. (1994) Arch. Biochem.
Tetrahedron, 60, 10411–10418. Biophys., 311, 496–502.
91 Fr€
ohlich, R.F.G., Zabelinskaja-Mackova, 107 Hughes, J., Lakey, J.H., and
A.A., Fechter, M.H., and Griengl, H. Hughes, M.A. (1997) Biotechnol. Bioeng.,
(2003) Tetrahedron: Asymmetry, 14, 53, 332–338.
355–362. 108 Effenberger, F., Roos, J., Kobler, C., and
92 Wagner, U.G., Hasslacher, M., Griengl, B€uhler, H. (2002) Can. J. Chem., 80,
H., Schwab, H., and Kratky, C. (1996) 671–679.
Structure (London), 4, 811–822. 109 Roos, J. and Effenberger, F. (1999)
93 Wagner, U.G., Schall, M., Hasslacher, M., Tetrahedron: Asymmetry, 10, 2817–2828.
Hayn, M., Griengl, H., Schwab, H., and 110 Effenberger, F., Roos, J., and Kobler, C.
Kratky, C. (1996) Acta Crystallogr., Sect. D, (2002) Angew. Chem., Int. Ed., 41,
52, 591–593. 1876–1879.
94 Zuegg, J., Gruber, K., Gugganig, M., 111 Kobler, C., Bohrer, A., and Effenberger, F.
Wagner, U.G., and Kratky, C. (1999) (2004) Tetrahedron, 60, 10397–10410.
Protein Sci., 8, 1990–2000. 112 Schmidt, M. and Griengl, H. (1999) Top.
95 Gruber, K., Gugganig, M., Wagner, U.G., Curr. Chem., 200, 193–226.
and Kratky, C. (1999) Biol. Chem., 380, 113 Asano, Y., Tamura, K., Doi, N.,
993–1000. Ueatrongchit, T., H-Kittikun, A., and
96 Gruber, K. (2001) Protein Struct., Funct., Ohmiya, T. (2005) Biosci., Biotechnol.,
Genet., 44, 26–31. Biochem., 69, 2349–2357.
References j985
114 Nanda, S., Kato, Y., and Asano, Y. (2005) 133 Han, S., Chen, P., Lin, G., Huang, H., and
Tetrahedron, 61, 10908–10916. Li, Z. (2001) Tetrahedron: Asymmetry, 12,
115 Nanda, S., Kato, Y., and Asano, Y. (2006) 843–846.
Tetrahedron: Asymmetry, 17, 735–741. 134 Chen, P., Han, S., Lin, G., Huang, H., and
116 Ueatrongchit, T., Kayo, A., Komeda, H., Li, Z. (2001) Tetrahedron: Asymmetry, 12,
Asano, Y., and H-Kittikun, A. (2008) 3273–3279.
Biosci. Biotechnol. Biochem., 72, 135 Purkarthofer, T., Gruber, K., Fechter, M.,
1513–1522. and Griengl, H. (2005) Tetrahedron, 61,
117 Ueatrongchit, T., Komeda, H., Asano, Y., 7661–7668.
and H-Kittikun, A. (2009) J. Mol. Catal. B: 136 Ognyanov, V.I., Datcheva, V.K., and
Enzym., 56, 208–214. Kyler, K.S. (1991) J. Am. Chem. Soc., 113,
118 Huang, S.-R., Liu, S.-L., Zong, M.-H., and 6992–6996.
Xu, R. (2005) Biotechnol. Lett., 27, 79–82. 137 Monterde, M.I., Nazabadioko, S.,
119 Solıs, A., Luna, H., Perez, H.I., and Rebolledo, F., Brieva, R., and Gotor, V.
Manjarrez, N. (2003) Tetrahedron: (1999) Tetrahedron: Asymmetry, 10,
Asymmetry, 14, 2351–2353. 3449–3455.
120 Brussee, J., Roos, E.C., and 138 Monterde, M.I., Brieva, R., and Gotor, V.
van der Gen, A. (1988) Tetrahedron Lett., (2001) Tetrahedron: Asymmetry, 12,
29, 4485–4488. 525–528.
121 Brussee, J., Loos, W.T., Kruse, C.G., and 139 Kragl, U., Niedermeyer, U., Kula, M.-R.,
van der Gen, A. (1990) Tetrahedron, 46, and Wandrey, C. (1990) Ann. N. Y. Acad.
979–986. Sci., 613, 167–175.
122 van Langen, L.M., van Rantwijk, F., and 140 Loos, W.T., Geluk, H.W., Ruijken,
Sheldon, R.A. (2003) Org. Process Res. M.M.A., Kruse, C.G., Brussee, J., and van
Dev., 7, 828–831. der Gen, A. (1995) Biocatal. Biotransform.,
123 Effenberger, F., Ziegler, T., and F€orster, S. 12, 255–266.
(1987) Angew. Chem., Int. Ed. Engl., 26, 141 Gr€oger, H., Capan, E., Barthuber, A., and
458–460. Vorlop, K.-D. (2001) Org. Lett., 3,
124 Wehtje, E., Adlercreutz, P., and 1969–1972.
Mattiasson, B. (1990) Biotechnol. Bioeng., 142 Trummler, K. and Wajant, H. (1997)
36, 39–46. J. Biol. Chem., 272, 4770–4774.
125 Wehtje, E., Adlercreutz, P., and 143 Brussee, J., Dofferhoff, F., Kruse, C.G.,
Mattiasson, B. (1993) Biotechnol. Bioeng., and van der Gen, A. (1990) Tetrahedron,
41, 171–178. 46, 1653–1658.
126 Zandbergen, P., van der Linden, J., 144 Kiljunen, E. and Kanerva, L.T. (1997)
Brussee, J., and van der Gen, A. (1991) Tetrahedron: Asymmetry, 8,
Synth. Commun., 21, 1387–1391. 1551–1557.
127 Huuhtanen, T.T. and Kanerva, L.T. (1992) 145 Wajant, H. and Pfizenmaier, K.
Tetrahedron: Asymmetry, 3, 1223–1226. (1996) J. Biol. Chem., 271,
128 Menendez, E., Brieva, R., Rebolledo, F., 25830–25834.
and Gotor, V. (1995) J. Chem. Soc., Chem. 146 Butler, G.W. (1965) Phytochemistry, 4,
Commun., 989–990. 127–131.
129 Kiljunen, E. and Kanerva, L.T. (1996) 147 Lieberei, R. (1988) J. Phytopathol., 122,
130 Hulsbos, E., Marcus, J., Brussee, J., and 148 Selmar, D., Lieberei, R., Biehl, B., and
van der Gen, A. (1997) Tetrahedron: Conn, E.E. (1989) Physiol. Plant., 75,
Asymmetry, 8, 1061–1067. 97–101.
131 Han, S., Lin, G., and Li, Z. (1998) 149 Klempier, N., Pichler, U., and Griengl, H.
Tetrahedron: Asymmetry, 9, (1995) Tetrahedron: Asymmetry, 6,
1835–1838. 845–848.
132 Chen, P., Han, S., Lin, G., Huang, H., and 150 Elliott, M., Farnham, A.W., Janes, N.F.,
Li, Z. (2002) J. Org. Chem., 67, Needham, P.H., and Pulman,
8251–8253. D.A. (1974) ACS Symp. Ser., 2, 80–91.
986
151 Briggs, G.G., Elliott, M., Farnham, A.W., 169 Dreveny, I., Kratky, C., and Gruber, K.
and Janes, N.F. (1974) Pestic. Sci., 5, (2002) Protein Sci., 11, 292–300.
643–649. 170 Dreveny, I., Andryushkova, A.,
152 Elliott, M., Farnham, A.W., Janes, N.F., Glieder, A., Gruber, K., and Kratky, C.
Needham, P.H., and Pulman, D.A. (1974) (2009) Biochemistry, 48, 3370–3377.
Nature, 248, 710–711. 171 Jorns, M.S. (1980) Biochim. Biophys. Acta,
153 Moss, G.P. (1999) IUBMB Enzyme 613, 203–209.
Nomenclature, EC 4.1.2.39, International 172 Casc~ao-Pereira, L.G., Hickel, A.,
Union of Biochemistry and Molecular Radke, C.J., and Blanch, H.W. (2003)
Biology, http://www.chem.qmw.ac.uk/ Biotechnol. Bioeng., 78, 595–605.
iubmb/enzyme/EC4/1/2/39.html. 173 Casc~ao-Pereira, L.G., Hickel, A.,
(accessed in 1999) Radke, C.J., and Blanch, H.W. (2003)
154 Wajant, H., F€orster, S., B€ottinger, H., Biotechnol. Bioeng., 83, 498–501.
Effenberger, F., and Pfizenmaier, K. (1995) 174 Hickel, A., Radke, C.J., and Blanch,
Plant Sci. (Shannon, Irel.), 108, 1–11. H.W. (1999) Biotechnol. Bioeng., 65,
155 Effenberger, F. and Heid, S. (1995) 425–436.
Tetrahedron: Asymmetry, 6, 2945–2952. 175 Hickel, A., Radke, C.J., and Blanch, H.W.
156 Effenberger, F. and Gaupp, S. (1999) (2001) Biotechnol. Bioeng., 74, 18–28.
Tetrahedron: Asymmetry, 10, 1765–1775. 176 Hickel, A., Radke, C.J., and Blanch, H.W.
157 Li, N., Zong, M.-H., Liu, C., Peng, H.-S., (1998) J. Mol. Catal. B: Enzym., 5,
and Wu, H.C. (2003) Biotechnol. Lett., 25, 349–354.
219–222. 177 Gerrits, P.J., Willeman, W.F., Straathof,
158 Effenberger, F., Roos, J., and Kobler, C. A.J.J., Heijnen, J.J., Brussee, J., and van
(2002) Angew. Chem., Int. Ed., 41, der Gen, A. (2001) J. Mol. Catal. B:
1876–1879. Enzym., 15, 111–121.
159 Griengl, H., Klempier, N., P€ochlauer, P., 178 Hanefeld, U., Stranzl, G., Straathof,
Schmidt, M., Shi, N., and Zabelinskaja- A.J.J., Heijnen, J.J., Bergmann, A.,
Mackova, A.A. (1998) Tetrahedron, 54, Mittelbach, R., Glatter, O., and Kratky, C.
14477–14486. (2001) Biochim. Biophys. Acta, 1544,
160 B€uhler, H., Bayer, A., and Effenberger, F. 133–142.
(2000) Chem.–Eur. J., 6, 2564–2571. 179 Gartler, G., Kratky, C., and Gruber, K.
161 Li, N., Zong, M.-H., Peng, H.-S., Wu, H.- (2007) J. Biotechnol., 129, 87–97.
C., and Liu, C. (2003) J. Mol. Catal. B: 180 Gruber, K., Gartler, G., Krammer, B.,
Enzym., 22, 7–12. Schwab, H., and Kratky, C. (2004) J. Biol.
162 Xu, R., Zong, M.-H., Liu, Y.-Y., He, J., Chem., 279, 20501–20510.
Zhang, Y.-Y., and Lou, W.-Y. (2004) Appl. 181 Stranzl, G.R., Gruber, K., Steinkellner, G.,
Microbiol. Biotechnol., 66, 27–33. Zangger, K., Schwab, H., and Kratky, C.
163 Liu, S.L., Zong, M.H., and Huang, S.R. (2004) J. Biol. Chem., 279, 3699–3707.
(2005) Chin. Chem. Lett., 16, 1317–1320. 182 Schmidt, A., Gruber, K., Kratky, C., and
164 Purkarthofer, T., Skranc, W., Weber, H., Lamzin, V.S. (2008) J. Biol. Chem., 283,
Griengl, H., Wubbolts, M., Scholz, G., 21827–21836.
and P€ochlaner, P. (2004) Tetrahedron, 60, 183 Bauer, M., Griengl, H., and Steiner, W.
735–739. (1999) Biotechnol. Bioeng., 62, 20–29.
165 Rotcenkovs, G. and Kanerva, L.T. (2000) J. 184 Bauer, M., Geyer, R., Griengl, H., and
Mol. Catal. B: Enzym., 11, 37–43. Steiner, W. (2002) Food Technol.
166 Hanefeld, U., Straathof, A.J.J., and Biotechnol., 40, 9–19.
Heijnen, J.J. (2001) J. Mol. Catal. B: 185 Purkarthofer, T., Gruber, K., Gruber-
Enzym., 11, 213–218. Khadjawi, M., Waich, K., Skranc, W.,
167 Bornemann, S. (2002) Nat. Prod. Rep., 19, Mink, D., and Griengl, H. (2006) Angew.
761–772. Chem., Int. Ed., 45, 3454–3456.
168 Dreveny, I., Gruber, K., Glieder, A., 186 Gruber-Khadjawi, M., Purkarthofer, T.,
Thompson, A., and Kratky, C. (2001) Skranc, W., and Griengl, H. (2007) Adv.
Structure (Cambridge, MA), 9, 803–815. Synth. Catal., 349, 1445–1450.
References j987
187 Lauble, H., Miehlich, B., F€
orster, S., 205 Andexer, J., Guterl, J.-K., Pohl, M., and
Kobler, C., Wajant, H., and Eggert, T. (2006) Chem. Commun.,
Effenberger, F. (2002) Protein Sci., 11, 4201–4203.
65–71. 206 Krammer, B., Rumbold, K.,
188 Wajant, H., Boettinger, H., and Tschemmernegg, M., P€ochlauer, P., and
Mundry, K.W. (1993) Biotechnol. Appl. Schwab, H. (2007) J. Biotechnol., 129,
Biochem., 18, 75–82. 151–161.
189 Wajant, H. and Mundry, K.W. (1993) 207 Reisinger, C., van Assema, F.,
Plant Sci., 89, 127–133. Sch€urmann, M., Hussain, Z., Remler, P.,
190 Wajant, H., Mundry, K.W., and and Schwab, H. (2006) J. Mol. Catal. B:
Pfizenmaier, K. (1994) Plant Mol. Biol., Enzym., 39, 149–155.
26, 735–746. 208 Pscheidt, B., Liu, Z., Gaisberger, R.,
191 Lauble, H., Miehlich, B., F€orster, S., Avi, M., Skranc, W., Gruber, K.,
Wajant, H., and Effenberger, F. (2002) Griengl, H., and Glieder, A. (2008) Adv.
Biochemistry, 41, 12043–12050. Synth. Catal., 350, 1943–1948.
192 Jaeger, K.-E., Eggert, T., Eipper, A., and 209 Semba, H., Ichige, E., Imanaka, T.,
Reetz, M.T. (2001) Appl. Microbiol. Atomi, H., and Aoyagi, H. (2008) Appl.
Biotechnol., 55, 519–530. Microbiol. Biotechnol., 79, 563–569.
193 Berglund, P. and Park, S. (2005) Curr. 210 Semba, H., Dobashi, Y., and Matsui, T.
Org. Chem., 9, 325–336. (2008) Biosci. Biotechnol. Biochem., 72,
194 Fox, R.J. and Huisman, G.W. (2008) 1457–1463.
Trends Biotechnol., 26, 132–138. 211 Hanefeld, U., Gardossi, L., and
195 Wang, Q. and Withers, S.G. (1995) J. Am. Magner, E. (2009) Chem. Soc. Rev., 38,
Chem. Soc., 117, 10137–10138. 453–468.
196 de Raadt, A., Griengl, H., and Weber, H. 212 Sheldon, R.A. (2007) Adv. Synth. Catal.,
(2001) Chem.–Eur. J., 7, 27–31. 349, 1289–1307.
197 de Raadt, A. and Griengl, H. (2002) Curr. 213 Costes, D., Wehtje, E., and Aldercreutz, P.
Opin. Biotechnol., 13, 537–542. (2001) J. Mol. Catal. B: Enzym., 11,
198 Li, Z., van Beilen, J.B., Duetz, W.A., 607–612.
Schmid, A., de Raadt, A., Griengl, H., and 214 Veum, L., Hanefeld, U., and Pierre, A.
Witholt, B. (2002) Curr. Opin. Chem. Biol., (2004) Tetrahedron, 60, 10419–10425.
6, 136–144. 215 Mateo, C., Palomo, J.M., van Langen,
199 Fechter, M.H., Gruber, K., Avi, M., L.M., van Rantwijk, F., and Sheldon, R.A.
Skranc, W., Schuster, C., P€ochlauer, P., (2004) Biotechnol. Bioeng., 86, 273–276.
Klepp, K.O., and Griengl, H. (2007) 216 van Langen, L.M., Selassa, R.P., van
Chem.–Eur. J., 13, 3369–3376. Rantwijk, F., and Sheldon, R.A. (2005)
200 Gaisberger, R., Weis, R., Luiten, R., Org. Lett., 7, 327–329.
Skranc, W., Wubbolts, M., Griengl, H., 217 Cabriol, F.L., Tan, P.L., Tay, B., Cheng, S.,
and Glieder, A. (2007) J. Biotechnol., 129, Hanefeld, U., and Sheldon, R.A. (2008)
30–38. Adv. Synth. Catal., 350, 2329–2338.
201 Liu, Z., Pscheidt, B., Avi, M., Gaisberger, 218 Chmura, A., van der Kraan, G.M., Kielar,
R., Hartner, F.S., Schuster, C., Skranc, W., F., van Langen, L.M., van Rantwijk, F.,
Gruber, K., and Glieder, A. (2008) and Sheldon, R.A. (2006) Adv. Synth.
ChemBioChem, 9, 58–61. Catal., 348, 1655–1661.
202 B€uhler, H., Effenberger, F., F€orster, S., 219 Cabriol, F.L., Hanefeld, U., and Sheldon,
Roos, J., and Wajant, H. (2003) R.A. (2006) Adv. Synth. Catal., 348,
ChemBioChem, 4, 211–216. 1645–1654.
203 Weis, R., Gaisberger, R., Skranc, W., 220 Roberge, C., Fleitz, F., Pollard, D., and
Gruber, K., and Glieder, A. (2005) Angew. Devine, P. (2007) Tetrahedron Lett., 48,
Chem., Int. Ed., 44, 4700–4704. 1473–1477.
204 Yan, G., Cheng, S., Zhao, G., Wu, S., Liu, 221 Weis, R., Gaisberger, R., Gruber, K., and
Y., and Sun, W. (2003) Biotechnol. Lett., 25, Glieder, A. (2007) J. Biotechnol., 129,
1041–1047. 50–61.
988
222 Effenberger, F. and Schw€

ammle, A. 239 Sakai, T., Miki, Y., Tsuboi, M.,
(1997) Biocatal. Biotransform., 14, Takeuchi, H., Ema, T., Uneyama, K., and
167–179. Utaka, M. (2000) J. Org. Chem., 65,
223 Inagaki, M., Hiratake, J., Nishioka, T., and 2740–3274.
Oda, J. (1989) Bull. Inst. Chem. Res., Kyoto 240 Sakai, T., Miki, Y., Nakatani, M., Ema, T.,
Univ., 67, 132–135. Uneyama, K., and Utaka, M. (1998)
224 Inagaki, M., Hiratake, J., Nishioka, T., and Tetrahedron Lett., 39, 5233–5236.
Oda, J. (1991) J. Am. Chem. Soc., 113, 241 Sakai, T., Takayama, T., Ohkawa, T.,
9360–9361. Yoshio, O., Ema, T., and Utaka, M. (1997)
225 Paizs, C., T€ahtinen, P., Lundell, K., Tetrahedron Lett., 38, 1987–1990.
Poppe, L., Irimie, F.-D., and Kanerva, L.T. 242 Veum, L., Kanerva, L.T., Halling, P.J.,
(2003) Tetrahedron: Asymmetry, 14, Maschmeyer, T., and Hanefeld, U. (2005)
1895–1904. Adv. Synth. Catal., 347, 1015–1021.
226 Sakai, T., Wang, K., and Ema, T. (2008) 243 Bianchi, P., Roda, G., Riva, S., Danieli, B.,
Tetrahedron, 64, 2178–2183. Zabelinskaja-Mackova, A., and
227 van Almsick, A., Buddrus, J., H€onicke- Griengl, H. (2001) Tetrahedron, 57,
Schmidt, P., Laumen, K., and 2213–2220.
Schneider, M.P. (1989) J. Chem. Soc., 244 Kobler, C. and Effenberger, F. (2005)
Chem. Commun., 1391–1393. Chem.–Eur. J., 11, 2783–2787.
228 Hsu, S.H., Wu, S.S., Wang, Y.F., and 245 Kobler, C., Bohrer, A., and Effenberger, F.
Wong, C.H. (1990) Tetrahedron Lett., 31, (2004) Tetrahedron, 60,
6403–6406. 10397–10410.
229 Wang, Y.F., Chen, S.T., Liu, K.K.C., and 246 Kobler, C. and Effenberger, F. (2004)
Wong, C.H. (1989) Tetrahedron Lett., 30, Tetrahedron: Asymmetry, 15,
1917–1920. 3731–3742.
230 Xu, Q., Geng, X., and Chen, P. (2008) 247 B€uhler, H., Miehlich, B., and Effenberger,
Tetrahedron Lett., 49, 6440–6441. F. (2005) ChemBioChem, 6, 711–717.
231 Mitsuda, S., Yamamoto, H., 248 Cruz Silva, M.M., Sa e Melo, M.L.,
Umemura, T., Hirohara, H., and Parolin, M., Tessaro, D., Riva, S., and
Nabeshima, S., (1990) Agric. Biol. Chem., Danieli, B. (2004) Tetrahedron: Asymmetry,
54, 2907–2912. 15, 21–27.
232 Fishman, A. and Zviely, M. (1998) 249 Schmidt, M. and Griengl, H. (1999) Top.
Tetrahedron: Asymmetry, 9, 107–118. Curr. Chem., 200, 193–226.
233 Effenberger, F., Gutterer, B., Ziegler, T., 250 Effenberger, F.X. (1992) (R)- and (S)-
Eckhardt, E., and Aichholz, R. (1991) Cyanohydrins – Their enzymatic
Liebigs Ann. Chem., 47–54. synthesis and their reactions, in Microbial
234 Holt, J., Arends, I.W.C.E., Minnaard, A.J., Reagents in Organic Synthesis (ed. S. Servi),
and Hanefeld, U. (2007) Adv. Synth. Kluwer, Dordrecht, pp. 25–33.
Catal., 349, 1341–1344. 251 Kruse, C.G., Geluk, H.W., and van
235 Wiggers, M., Holt, J., Kourist, R., Bartsch, Scharrenburg, G.J.M. (1992) Chim. Oggi,
S., Arends, I.W.C.E., Minnaard, A.J., 10, 59–63.
Bornscheuer, U.T., and Hanefeld, U. 252 Breuer, M. and Hauer, B. (2003) Curr.
(2009) J. Mol. Catal. B: Enzym., 60, 82–86. Opin. Biotechnol., 14, 570–576.
236 Kourist, R., Domınguez de Marıa, P., and 253 Johnson, D.V. and Griengl, H. (1997)
Bornscheuer, U.T. (2008) ChemBioChem, Chim. Oggi., 15, 9–13.
9, 491–498. 254 Deardorff, D.R., Taniguchi, C.M., Nelson,
237 Inagaki, M., Hiratake, J., Nishioka, T., and A.C., Pace, A.P., Kim, A.J., Pace, A.K.,
Oda, J. (1992) J. Org. Chem., 57, Jones, R.A., Tafti, S.A., Nguyen, C.,
5643–5649. OConnor, C., Tang, J., and Chen, J.
238 Inagaki, M., Hatanaka, A., Mimura, M., (2005) Tetrahedron: Asymmetry, 16,
Hiratake, J., Nishioka, T., and Oda, J. 1655–1661.
(1992) Bull. Chem. Soc. Jpn., 65, 255 Ziegler, T., H€orsch, B., and Effenberger,
111–120. F. (1990) Synthesis, 575–478.
References j989
256 Osprian, I., Fechter, M.H., and van der Marel, G.A., and van der Gen, A.
Griengl, H. (2003) J. Mol. Catal. B: (2001) Tetrahedron: Asymmetry, 12,
Enzym., 24–25, 89–98. 1109–1112.
257 Reisinger, C., Osprian, I., Glieder, A., 275 van der Nieuwendijk, A.M.C.H., Kriek,
Schoemaker, H.E., Griengl, H., and N.M.A.J., Brussee, J., van Boom, J.H., and
Schwab, H. (2004) Biotechnol. Lett., 26, van der Gen, A. (2000) Eur. J. Org. Chem.,
1675–1680. 3683–3691.
258 Zandbergen, P., Brussee, J., van der Gen, 276 Ueberbacher, B.J., Griengl, H., and
A., and Kruse, C.G. (1999) Tetrahedron: Weber, H. (2008) Tetrahedron: Asymmetry,
Asymmetry, 10, 2817–2828. 19, 838–846.
259 Marcus, J., Vandermeulen, G.W.M., 277 Johnson, D.V., Felfer, U., and Griengl, H.
Brussee, J., and van der Gen, A. (1999) (2000) Tetrahedron, 56, 781–790.
Tetrahedron: Asymmetry, 10, 1617–1622. 278 Johnson, D.V., Fischer, R., and Griengl,
260 Duffield, J.J. and Regan, A.C. (1996) H. (2000) Tetrahedron, 56, 9289–9295.
Tetrahedron: Asymmetry, 7, 663–666. 279 Avi, M., Gaisberger, R., Feichtenhofer, S.,
261 B€uhler, H., Bayer, A., and Effenberger, and Griengl, H. (2009) Tetrahedron, 65,
F. (2000) Chem.–Eur. J., 6, 2564–2571. 5418–5426.
262 Roos, J. and Effenberger, F. (2002) 280 Tromp, R.A., Brussee, J., and van der Gen,
Tetrahedron: Asymmetry, 13, 1855–1862. A. (2003) Org. Biomol. Chem., 1,
263 Effenberger, F. and Stelzer, U. (1991) 3592–3599.
Angew. Chem., Int. Ed. Engl., 30, 873–874. 281 van Pelt, S., van Rantwijk, F., and
264 Effenberger, F. and Roos, J. Sheldon, R.A. (2009) Adv. Synth. Catal.,
(2000) Tetrahedron: Asymmetry, 11, 351, 397–404.
1085–1095. 282 Rustler, S., Motejadded, H.,
265 Warmerdam, E.G.J.C., Brussee, J., Altenbuchner, J., and Stolz, A. (2008)
Kruse, C.G., and van der Gen, A. (1993) Appl. Microbiol. Biotechnol., 80, 87–97.
Tetrahedron, 49, 1063–1070. 283 Sheldon, R.A., Schoevaart, R., and van
266 Effenberger, F., Kremser, A., and Stelzer, Langen, L.M. (2005) Biocatal.
U. (1996) Tetrahedron: Asymmetry, 7, Biotransform., 23, 141–147.
607–618. 284 Sheldon, R.A. (2005) Green Chem., 7,
267 Gaupp, S. and Effenberger, F. (1999) 267–278.
Tetrahedron: Asymmetry, 10, 1777–1786. 285 Mateo, C., Chmura, A., Rustler, S.,
268 Nazabadioko, S., Perez, R.J., Brieva, R., van Rantwijk, F., Stolz, A., and
and Gotor, V. (1998) Tetrahedron: Sheldon, R.A. (2006) Tetrahedron:
Asymmetry, 9, 1597–1604. Asymmetry, 17, 320–323.
269 Gotor, V. (2002) Org. Process Res. Dev., 6, 286 Sheldon, R.A. (2008) Chem. Commun.,
420–426. 3352–3365.
270 Gotor, V. (2002) J. Biotechnol., 96, 35–42. 287 van Pelt, S., van Rantwijk, F., and
271 Vink, M.K.S., Schortinghuis, C.A., Sheldon, R.A. (2009) Adv. Synth. Catal.,
Mackova-Zabelinskaja, A., Fechter, M., 351, 397–404.
P€ochlauer, P., Castelijns, A.M.C.F., van 288 von Langermann, J., Guterl, J.-K., Pohl,
Maarseveen, J.H., Hiemstra, H., Griengl, M., Wajant, H., and Kragl, U. (2008)
H., Schoemaker, H.E., and Rutjes, Bioprocess Biosyst. Eng., 31, 155–161.
F.P.J.T. (2003) Adv. Synth. Catal., 345, 289 Willemann, W.F., Gerrits, P.J., Hanefeld,
483–487. U., Brussee, J., Straathof, A.J.J., van der
272 Stelzer, U. and Effenberger, F. (1993) Gen, A., and Heijnen, J.J. (2002)
Tetrahedron: Asymmetry, 4, 161–164. Biotechnol. Bioeng., 77, 239–247.
273 Tromp, R.A., van der Hoeven, M., Amore, 290 von Langermann, J., Mell, A., Paetzold,
A., Brussee, J., Overhand, M., van der E., Daußmann, T., and Kragl, U. (2007)
Marel, G.A., and van der Gen, A. (2003) Adv. Synth. Catal., 349, 1418–1424.
Tetrahedron: Asymmetry, 14, 1645–1652. 291 Wajant, H., F€orster, S., Sprauer, A.,
274 Tromp, R.A., van der Hoeven, M., Amore, Effenberger, F., and Pfizenmaier, K.
A., Brussee, J., Overhand, M., (1996) Ann. N.Y. Acad. Sci., 799, 771–776.
990
292 Effenberger, F., Ziegler, T., and F€

orster, S. 306 Hickel, A., Gradnig, G., Griengl, H.,
(1987) Angew. Chem., Int. Ed., 26, Schall, M., and Sterk, H. (1996)
458–460. Spectrochim Acta, Part A, 52, 93–96.
293 van den Nieuwendijk, A.M.C.H., 307 National Research Council (1995)
Warmerdam, E.G.J.C., Brussee, J., and Prudent Practices in the Laboratory:
van der Gen, A. (1995) Tetrahedron: Handling and Disposal of Chemicals,
Asymmetry, 6, 801–806. National Academic Press,
294 Costes, D., Wehtje, E., and Adlercreutz, P. Washington, D.C.
(1999) Enzyme Microb. Technol., 25, 308 Lewis, R.J. (1999) Saxs Dangerous
384–391. Properties of Industrial Materials, 10th edn,
295 Gotor, V. (2000) Molecules, 5, 290–292. John Wiley & Sons, Inc., New York.
296 Persson, M., Costes, D., Wehtje, E., and 309 Somerville, R.L. (1990) Chem. Eng. Prog.,
Aldercreutz, P. (2002) Enzyme Microb. 86, 64–68.
Technol., 30, 916–923. 310 Baxter, R. (1977) Ind. Finish. (Wheaton,
297 Pogorevc, M., Stecher, H., and Faber, K. IL), 53, 38–41.
(2002) Biotechnol. Lett., 24, 857–860. 311 Kiljunen, E. and Kanerva, L.T. (1996)
298 Koch, K., van den Berg, R.J.F., Tetrahedron: Asymmetry, 7, 1105–1116.
Nieuwland, P.J., Wijtmans, R., 312 Warmerdam, E.G.J.C., van den
Wubbolts, M.G., Schoemaker, H.E., Nieuwendijk, A.M.C.H., Kruse, C.G.,
Rutjes, F.P.J.T., and van Hest, J.C.M. Brussee, J., and van der Gen, A. (1996)
(2008) Chem. Eng. J. (Amsterdam), 135, Recl. Trav. Chim. Pays-Bas, 115, 20–24.
89–92. 313 Tellitu, I., Badia, D., Dominguez, E., and
299 Koch, K., van den Berg, R.J.F., Garcia, F.J. (1994) Tetrahedron:
Nieuwland, P.J., Wijtmans, R., Asymmetry, 5, 1567–1578.
Schoemaker, H.E., van Hest, J.C.M., and 314 Lu, W., Chen, P., and Lin, G. (2008)
Rutjes, F.P.J.T. (2008) Biotechnol. Bioeng., Tetrahedron, 64, 7822–7827.
99, 1028–1033. 315 Effenberger, F. and Eichhorn, J. (1997)
300 Chen, P.R., Gu, J.-X., Wie, Z.-L., Tetrahedron: Asymmetry, 8, 469–476.
Han, S.Q., Li, Z.-Y., and Lin, G. (2003) 316 Syed, J., F€orster, S., and Effenberger, F.
Chin. J. Chem., 21, 983–993. (1998) Tetrahedron: Asymmetry, 9,
301 Chen, P.R., Han, S.Q., Lin, G., and 805–815.
Li, Z.-Y. (2002) J. Org. Chem., 67, 317 Effenberger, F. (1994) Angew. Chem., Int.
8251–8253. Ed. Engl., 33, 1555–1564.
302 Gerrits, P.J., Marcus, J., Birikaki, L., and 318 Effenberger, F., H€orsch, B., Weingart, F.,
van der Gen, A. (2001) Tetrahedron: Ziegler, T., and K€ uhner, S. (1991)
Asymmetry, 12, 971–974. Tetrahedron Lett., 32, 2605–2608.
303 Avi, M. and Griengl, H. (2008) 319 Effenberger, F., Gutterer, B., and Syed, J.
Biocatalysis in biphasic systems: (1995) Tetrahedron: Asymmetry, 6,
oxynitrilases, in Organic Synthesis with 2933–2943.
Enzymes in Non-Aqueous 320 Brussee, J. and van der Gen, A. (1991)
Media (eds G. Carrea and S. Riva), Recl. Trav. Chim. Pays-Bas, 110, 25–26.
Wiley-VCH Verlag GmbH, Weinheim, 321 Effenberger, F., Hopf, M., Ziegler, T., and
pp. 211–226. Hudelmayer, J. (1991) Chem. Ber., 124,
304 Gaisberger, R.P., Fechter, M.H., and 1651–1659.
Griengl, H. (2004) Tetrahedron: 322 Effenberger, F. and Stelzer, U. (1993)
Asymmetry, 15, 2959–2963. Chem. Ber., 126, 779–786.
305 Lou, W.-Y., Xu, R., and Zong, M.-H. 323 Jackson, W.R., Jacobs, H.A., Jayatilake,
(2005) Biotechnol. Lett., 27, G.S., Matthews, B.R., and Watson, K.G.
1387–1390. (1990) Aust. J. Chem., 43, 2045–2062.
j991
24
Industrial Application and Processes Using
Carbon–Carbon Lyases
24.1
Processes Using Carbon–Carbon Lyases
Carbon–carbon lyases (EC 4.1) are an attractive group of catalysts as demonstrated by

their use in many industrial processes. The reaction catalyzed is the cleavage of the
CC bond. Notably, this bond cleavage is different from hydrolysis, often leaving
unsaturated products with double bonds that may be subjected to further reactions.
In industrial processes these enzymes are most commonly used in the synthetic
mode, meaning that the reverse reaction – addition of a molecule to an unsaturated
substrate – is of interest. If the substrate is a carboxylic acid, one of the products will
be carbon dioxide. If the substrate is an aldehyde, carbon monoxide could be a
product. To shift the equilibrium these reactions are carried out at very high substrate
concentrations, which results in high reaction yields of the desired products.
24.2
Syntheses Using Carboxy-Lyases
The carboligation of acetaldehyde and benzaldehyde is carried out by Krebs Bio-

chemicals & Industries Ltd. using whole cells of Saccharomyces cerevisiae containing
pyruvate decarboxylase (EC 4.1.1.1). The yeast grows on molasses. The product of this
conversion is phenylacetylcarbinol ((R)-PAC) (Scheme 24.1). PAC production from
benzaldehyde by yeast is also operated on a large scale by Knoll (BASF) and Malladi
Drugs. Subsequently, (R)-PAC is converted chemically in a two-step process into
D-pseudoephedrine. D-Ephedrine and D-pseudoephedrine are used for the treatment
of asthma, hay fever, and as a bronchodilating agent and decongestant (R.T. Ravi,
Krebs Biochemicals Ltd., personal communication) [1].
During beer fermentation diacetyl is formed as a by-product by non-enzymatic
oxidative decarboxylation of a-acetolactate (Scheme 24.2). This by-product has to be
eliminated; therefore, fast processes are required. The problem is the slow reaction
rate for the conversion of a-acetolactate into diacetyl with subsequent conversion of
and Oliver May.
j 24 Industrial Application and Processes Using Carbon–Carbon Lyases
992
Scheme 24.1 Synthesis of (R)-phenylacetylcarbinol from acetaldehyde and benzaldehyde catalyzed

by whole cells of Saccharomyces cerevisiae [2].
Scheme 24.2 Acetolactate decarboxylase process for fast degradation of a-acetolactate to acetoin [2].
diacetyl into acetoin. Diacetyl has a very low flavor threshold, compared to acetoin.
The addition of acetolactate decarboxylase (EC 4.1.1.5) from Bacillus brevis allows the
bypassing of the slow oxidation step. The enzyme can be activated at low pH values by
addition of glutardialdehyde, which intermolecularly crosslinks the active dimer that
otherwise dissociates under acidic conditions [3].
Aspartate b-decarboxylase (EC 4.1.1.12) is used for the production of alanine from
aspartic acid. L-Alanine is produced industrially by Tanabe Seiyaku Co., Ltd. by a batch
process with L-aspartate b-decarboxylase from Pseudomonas dacunhae. To improve the
productivity a continuous process was established. Here the formation of carbon
dioxide was the main problem in comparison to the catalyst stability and the microbial
enzyme activity. The production of carbon dioxide occurs stoichiometrically, which
means that 50 l of carbon dioxide per liter reaction mixture (2 M aspartate) are
generated. The consequence is difficulties in obtaining plug-flow conditions in fixed
bed reactors and the pH shift that takes place due to formation carbon dioxide.
Therefore, a pressurized (10 bar) fixed bed reactor was designed, in which the enzyme
stability is not affected by the elevated pressure. The main side reaction, the formation
of L-malic acid, can be avoided completely. The aspartate b-decarboxylase activity is
24.3 Syntheses Using Aldehyde Lyases j993
Figure 24.1 Production of L-alanine and D-aspartic acid catalyzed by aspartate b-decarboxylase (E).
stabilized by addition of pyruvate and pyridoxal phosphate whereas the alanine

racemase and fumarase activities can be destroyed by acid treatment of the micro-
organisms. This treatment improves the yield of L-alanine. The process is often
combined with the aspartase catalyzed synthesis of L-aspartic acid from fumarate (see
also Chapters 18 and 20). In this combination L-alanine can efficiently be produced by
co-immobilization of Escherichia coli and Pseudomonas dacunhae cells. If D,L-aspartic
acid is used as a substrate for the reaction, L-aspartic acid is converted into L-alanine and
D-aspartic acid remains unchanged in the resolution step. Both products can be
separated after crystallization by addition of sulfuric acid (Figure 24.1) [4–8].
24.3
Syntheses Using Aldehyde Lyases
The (R)-specific hydroxynitrile lyase (HNL) from almond (Prunus amygdalus,

PaHNL) [9] and the (S)-specific HNL from rubber tree (Hevea brasiliensis, HbHNL)
are nowadays available by overexpression in Pichia pastoris, Saccharomyces cerevisiae,
and E. coli [10, 11]. This has paved the way for the introduction of HNL-based
industrial processes for the production of chiral hydroxynitriles and 2-hydroxycar-
boxylic acid [12, 13].
Hydroxynitrilase (EC 4.1.2.39) also known as oxynitrilase catalyzes the reaction of
m-phenoxybenzaldehyde and HCN to yield (S)-m-phenoxybenzaldehyde cyanohy-
drin (Scheme 24.3a). Whole cells containing the cloned and overexpressed enzyme
from Hevea brasiliensis are used in a biphasic solvent system consisting of aqueous
buffer and methyl tert-butyl ether. The enzyme for this HCN addition can be reused
994
Scheme 24.3 Asymmetric HCN addition to aldehydes and ketones catalyzed by (a) HbHNL (Hevea
brasiliensis hydroxynitrile lyase) and (b)/(c) PaHNL (Prunus amygdalus hydroxynitrile lyase) [2].
more than five times without loss of activity. The product is obtained with a yield of
98% as well as an enantiomeric excess of 98% and is used as an intermediate for the
manufacture of pyrethroids [14–16].
PaHNL has been implemented in industrial syntheses of some chiral aromatic
2-hydroxycarboxylic acids, such as (R)-2-chloromandelic acid by DSM Fine Chemi-
cals Austria, Nippon Shokubai, and Clarinat (Scheme 24.3b) [12, 17–20]. This product
is an intermediate for the synthesis of the antidepressant and platelet-aggregation
inhibitor clopidogrel. PaHNL is applied in the form of an almond-flour extract or
immobilized on Avicel microcrystalline cellulose for the enantioselective addition of
HCN to 2-chlorobenzaldehyde. The enzyme can be used for several months
depending on the solvent employed. The (R)-2-choromandelonitrile formed
(Scheme 24.3b) is converted into the corresponding carboxylic acid by hydrolysis
with concentrated HCl without racemization. Thus 100% theoretical yield is possible.
The active site cavity of almond (R)-HNL has been customized by means of site-
directed mutations for increased enantioselectivity in respect to (R)-2-hydroxy-4-
phenylbutyronitrile (Scheme 24.3c). This nitrile is a key intermediate in the synthesis
of different angiotensin-converting enzyme inhibitors [21].
Aliphatic nitriles are produced using halohydrin dehalogenase (EC 3.8.X.X), a
process developed by Codexis Inc. This enzyme is mentioned here because a CC-
bond is formed although it is not a carbon–carbon lyase but instead it is a hydrolase
24.4 Syntheses Using Oxo-Acid Lyases j995
acting on halide bonds. Here NaCN is used in the cyanation reaction using a soluble
recombinant protein. The process is described in detail in Chapter 38.
Wong et al. first described the potential of deoxyribose-5-phosphate aldolase
(E.C. 4.1.2.4) (synonym: DERA) to catalyze the aldol condensation of chloroacetal-
dehyde with two molecules of acetaldehyde yielding (3R,5S)-6-chloro-3,5-di-
hydroxyhexanal (Scheme 24.4). This chiral compound is an important precursor in
the syntheses of 3-hydroxy-3-methylglutaryl CoA (HMG-CoA) reductase inhibitors
(statins), hypolipidemic agents, which are multibillion-dollar drugs [22]. The need for
high enantioselectivity at both stereo centers has led to the development of at least six
different routes involving biocatalysis [23, 24]. The product (97% d.e.) of the DERA
biotransformation is stabilized as a hemiacetal under optimized process conditions
at DSM. Independently, a similar approach also utilizing a DERA aldolase was
developed by the Wong group together with Diversa Corp. (USA) [25]. To broaden the
range of accepted substrates they isolated a DERA variant (S238A) that accepts
azidopropionaldehyde, thereby enabling access to 7-azido-(3R,5S)-dihydroxyhepta-
nal, the key intermediate for Atorvastatin (Scheme 24.4) [26].
Scheme 24.4 Deoxyribose-5-phosphate aldolase (DERA) as biocatalyst in the synthesis for HMG-
CoA reductase inhibitors (statins) [2].
24.4
Syntheses Using Oxo-Acid Lyases
N-Acetyl-D-neuraminic acid aldolase (EC 4.1.3.3) is an oxo-acid lyase that can be

applied in the synthesis of N-acetyl-D-neuraminic acid (Neu5Ac). Here the enzyme is
used covalently immobilized on Eupergit C. Neu5Ac is the major representative of
amino sugars (sialic acids) that are incorporated at the terminal positions of
glycoproteins and glycolipids and plays an important role in a wide range of biological
recognition processes. The synthesis of Neu5Ac analogues and Neu5Ac-containing
oligosaccharides is of interest in studies towards inhibitors of neuraminidase,
996
hemagglutinin, and selectin-mediated leukocyte adhesion. The chemical synthesis of

Neu5Ac is costly since it requires complex protection and deprotection steps. Since N-
acetyl-D-mannosamine is very expensive, it is synthesized from N-acetyl-D-glucos-
amine by chemical epimerization at C2 (Scheme 24.5). The equilibrium of the
epimerization is on the side of N-acetyl-D-glucosamine (GlcNAc: ManNAc ¼ 4 : 1).
After neutralization and addition of isopropanol GlcNAc precipitates. In the remain-
ing solution a ratio of GlcNAc: ManNAc ¼ 1 : 4 is reached. Alternatively, an epimerase
can be applied (see below).
Scheme 24.5 Synthesis of N-acetyl-D-neuraminic acid catalyzed by N-acetylneuraminate pyruvate

lyase [2].
N-Acetyl-D-glucosamine is not a substrate for the aldolase, but is an inhibitor and

thereby limits the applied maximal concentration of ManNAc. Non-converted
GlcNAc can be recycled after downstream processing by epimerization to ManNAc.
The natural direction of the aldolase catalyzed reaction is the cleavage of Neu5Ac to
pyruvate a ManNAc. The KM for ManNAc is 700 mM. Therefore, a very high ManNAc
concentration of up to 20% (w/v) is used. By this means ManNAc itself drives the
equilibrium. Pyruvate is used in a 1.5 molar ratio. Neu5Ac can be crystallized directly
from the reaction mixture simply by the addition of acetic acid. In the repetitive batch
mode the immobilized enzyme can be reused for at least nine cycles without any
significant loss of activity [27–29].
Another possibility to synthesize Neu5Ac is the application of N-acetyl-d-neur-
aminic acid aldolase from E. coli K12 and N-acetyl-D-glucosamine epimerase (EC 5.1)
from porcine kidney, which have been cloned and overexpressed in E. coli.
The enzyme-catalyzed aldol condensation to Neu5Ac is combined in a one-pot
24.4 Syntheses Using Oxo-Acid Lyases j997
synthesis with the enzyme-catalyzed epimerization of GlcNAc. The production of
Neu5Ac can be carried out using the free solubilized enzymes at a pH of 7.2. Since
excess amounts of pyruvate inhibit the epimerase, a fed batch in regard to pyruvate is
performed. After the start of the reaction with a ratio of pyruvate to GlcNAc of 1 : 0.6,
pyruvate is added twice up to a total amount of 251 mol (ratio of pyruvate to GlcNAc
after first addition of 1 : 1.5; after second addition of GlcNAc of 1 : 2). Before the
product is purified by crystallization initiated by the addition of 5 volumes of glacial
acetic acid, the enzymes are denatured by heating to 80 C for 5 min. The reaction
solution is then filtered.
Another process layout has been realized by the Research Center J€ ulich, Germany,
which already established in 1991 the one-pot synthesis of Neu5Ac with the
combined use of epimerase and aldolase. Here a continuously operated membrane
reactor is used, where the enzymes are retained by an ultrafiltration membrane. By
this technology kg quantities of Neu5Ac are produced. The advantage of this
approach is a simplified downstream processing, since the catalysts are already
separated. The product is additionally pyrogen free [30–34].
24.4.1
Synthesis of L-DOPA Catalyzed by Tyrosine Phenol Lyase from Erwinia herbicola
The product is applied for the treatment of Parkinsonism that is caused by a lack of
L-dopamine and its receptors in the brain. L-Dopamine is synthesized in organisms by
decarboxylation of L-3,4-dihydroxyphenylalanine (L-DOPA). Since L-dopamine cannot
pass the blood–brain barrier L-DOPA is applied in combination with DOPA-decar-
boxylase inhibitors to avoid formation of L-dopamine outside the brain. Ajinomoto
produces L-DOPA by this lyase-biotransformation with suspended whole cells in a fed
batch reactor on a scale of 250 t a1 (Scheme 24.6). The catalyzing enzyme is tyrosine
phenol lyase (EC 4.1.99.2). Much earlier Monsanto successfully scaled up the
chemical synthesis of L-DOPA (Scheme 24.7).
Scheme 24.6 Synthesis of L-3,4-dihydroxyphenylalanine (L-DOPA) from catechol catalyzed by

whole cells of Erwinia herbicola [2].
The enantioselective hydrogenation of 3,4-dihydroxy-N-acetylaminocinnamic acid

is catalyzed by the cationic Rh-bisphosphine complex DIPAMP, in which the
enantioselectivity is introduced by the chiral phosphine. The hydrogenation proceeds
quantitatively with 94% e.e. The optically pure L-DOPA is separated from the catalyst
by crystallization [35–40].
998
Scheme 24.7 Chemical synthesis of L-DOPA (Monsanto) [2].
24.5
Outlook
Carbon–carbon bond forming reactions are of key importance to organic chemistry,

and a broad range of reactions can be catalyzed by enzymes. Enzymatic CC bond
forming reactions are advantageous over chemical methods since they are highly
chemo-, regio-, and stereoselective. Predominantly, lyases have been used so far to
synthesize complex, highly functionalized target molecules and several biologically
important natural compounds for CC bond formations [41]. Nevertheless, there is
great demand for cyanation reactions of aliphatic compounds. Here, novel lyase
catalyzed processes are sought after. Additionally, cascade reactions to synthesize
polyfunctional molecules, as demonstrated by the synthesis of an intermediate for
Atorvastatin catalyzed by DERA, will be of great interest in future.
References
1 Cheetham, P.S.J. (1994) Case studies in Academic Publishers, Chur, Switzerland,

applied biocatalysis, in Applied pp. 77–78.
Biocatalysis (eds J.M.S. Cabral, D. Best, L. 2 Liese, A., Seelbach, K., Buchholz, A., and
Boross, and J. Tramper), Harwood Haberland, J. (2006) Processes, in
References j999
Industrial Biotranformations (eds. Chemicals Austria GmbH); Chem. Abstr.,
A. Liese, K. Seelbach, C. Wandrey), 132 (2000) 191223.
Wiley-VCH, Weinheim, pp. 147–513. 12 Breuer, M., Ditrich, K., Habicher, T. et al.
3 Pedersen, S., Lange, N.K., and Nissen, (2004) Industrial methods for the
A.M. (1995) Novel industrial enzyme production of optically active
applications. Ann. N. Y. Acad. Sci., 750, intermediates. Angew. Chem. Int. Ed., 43,
376–390. 788–824.
4 Chibata, L., Tosa, T., and Shibatani, T. 13 Sharma, M., Sharma, N.N., and Bhalla,
(1992) The industrial production of T.C. (2005) Hydroxynitrile lyases: at the
optically active compounds by interface of biology and chemistry. Enzyme
immobilized biocatalysts, in Chirality in Microb. Technol., 37, 279–294.
Industry (eds A.N. Collins, G.N. 14 Gr€oger, H. (2001) Enzymatic routes to
Sheldrake, and J. Crosby), John Wiley & enantiomerically pure aromatic a-hydroxy
Sons, Inc., New York, pp. 351–370. carboxylic acids: a further example for the
5 Furui, M. and Yamashita, K. (1983) diversity of biocatalysis. Adv. Synth. Catal.,
Pressurized reaction method for 343 (6–7), 547–558.
continuous production of 15 P€ochlauer, P., Schmidt, M., Wirth, I.,
L-alanine by immobilized Pseudomonas Neuhofer, R., Zabelinskaja-Mackova, A.,
dacunhae cells. J. Ferment. Technol., 61, Griengl, H., van den Broek, C.,
587–591. Reintjens, R., and W€ories, J. (1999) EP
6 Schmidt-Kastner, G. and Egerer, P. (1984) 927766.
Amino acids and peptides, in 16 Griengl, H., Schwab, H., and Fechter, M.
Biotechnology, vol. 6a (ed. K. Kieslich), (2000) The synthesis of chiral
Verlag Chemie, Weinheim, cyanohydrins by oxynitrilases. Opht. Gen.,
pp. 387–419. 18, 252.
7 Takamatsu F.S., Umemura, I., 17 Bousquet, A. and Musolino, A. (1998)
Yanamoto, K., Sato, T., Tosa, T., and hydroxyacetic ester derivatives,
Chibata, I. (1982) Production of L-alanine preparation method and
from ammonium fumarate using two use as synthesis intermediates.
immobilized microorganisms: EP1021449A1 (Sanofi-Synthelabo);
elimination of side reactions. Eur. J. Appl. Chem. Abstr., 130 (1999) 296510.
Microbiol. Biotechnol., 15, 147–152. 18 P€ochlauer, P. and Mayrhofer, H. (2001)
8 Takamatsu, S. and Tosa, T. (1993) Method for producing optically and
Production of L-alanine, in Industrial chemically, highly pure (R)- or (S)-alpha-
Application of Immobilized Biocatalysts hydroxycarboxylic acids. EP1148042A2
(eds A. Tanaka, T. Tosa, and T. Kobayashi), (DSM Fine Chemicals Austria); Chem.
Marcel Dekker Inc., New York. Abstr., 135 (2001) 303603.
9 Glieder, F.A., Weis, R., Skranc, W. et al. 19 Okuda, N., Semba, H., and Dobashi, Y.
(2003) Comprehensive step-by-step (2001) Method for producing alpha-
engineering of an (R)-hydroxynitrile lyase hydroxycarboxylic acid. EP 1160235A2
for large-scale asymmetric (Nippon Catalytic Chem. Ind.); Chem.
synthesis. Angew. Chem. Int. Ed., 42, Abstr., 136 (2002) 5721.
4815–4818. 20 Semba, H. and Dobashi, Y. (2001)
10 Chemie Linz (Deutschland) GmbH (1997) An enzyme reaction method and a method
DNA encoding Hevea brasiliensis (S)- for enzymatically producing an optically
hydroxynitrilase. DE19529116A1. Chem. active cyanohydrin. EP 1160329 (Nippon
Abstr., 126 (1997) 208953. Catalytic Chem. Ind.); Chem. Abstr., 136
11 Effenberger, F., Lauble, P., Buehler, H., (2002) 4802.
Wajant, H., Foerster, S., Schwab, H., 21 P€ochlauer, P., Schmidt, M.,
Kratky, C., Wagner, U., and Steiner, E. Wirth, I., Neuhofer, R., Zabelinskaja-
(1999) (S)-Hydroxynitrile lyases with Mackova, A., Griengl, H., van den Broek,
improved substrate acceptance and uses C., Reintjens, R.W. E. G., and Wories, H. J.
thereof. EP969095A2 (DSM Fine (1999) Enzymatic process for the
1000
preparation of (S)-cyanhydrines. Acetylneuraminic acid: from a rare

EP 927766 (DSM Fine Chemicals Austria); chemical isolated from natural sources to
Chem. Abstr., 131 (2000) 72766. a multi-kilogram enzymatic synthesis for
22 Panke, S. and Wubbolts, M. (2005) industrial application, Ann. N. Y. Acad. Sci.
Advances in biocatalytic synthesis of 750, 300–305.
pharmaceutical intermediates. Curr. Opin. 33 Maru, I., Ohnishi, J., and Tsukada, Y.
Chem. Biol., 9, 188–194. (1998) Simple and large-scale production
23 M€uller, M. (2005) Chemoenzymatic of N-acetylneuraminic acid from
synthesis of building blocks for statine N-acetyl-D-glucosamine and pyruvate
side chains. Angew. Chem. Int. Ed. Engl., using N-acyl-D-glucosamine 2-epimerase
44, 362–365. and N-acetylneuraminate lyase.
24 Kierkels, J.G.T., Mink, D., Panke, S. et al. Carbohydr. Res., 306, 575–578.
(2003) Process for the preparation of 2, 34 Ohta, Y., Tsukada, Y., Sugimori, T.,
4-dideoxyhexoses and therapeutic uses Murata, K., and Kimura, A. (1989)
thereof WO 2003/006656 (DSM). Isolation of a constitutive N-
25 DeSantis, G., Liu, J., Clark, D.P. et al. (2003) acetylneuraminate lyase-producing
Structure based mutagenesis approaches mutant of Escherichia coli and its use for
towards expanding the substrate specificity NPL production. Agric. Biol. Chem., 53,
of D-2-deoxyribose-5-phosphate aldolase. 477–481.
Bioorg. Med. Chem., 11, 43–52. 35 Tsuchida, T., Nishimoto, Y., Kotani, T., and
26 Liu, J., Hsu, C.-C., and Wong, C.-H. (2004) Iiizumi, K. (1993) Production of L-3,4-
Sequential aldol condensation catalyzed Dihydroxyphenylalanine. JP5123177A
by DERA mutant Ser238Asp and a formal (Ajinomoto Co., Ltd.).
total synthesis of atorvastatin. Tetrahedron 36 Ager, D.J. (1999) Handbook of Chiral
Lett., 45, 2439–2441. Chemicals, Marcel Dekker, New York.
27 Dawson, M., Noble, D., and 37 Yamamoto, A., Yokozeki, K., and
Mahmoudian, M. (1994) Process for the Kubota, K. (1989) Production of aromatic
preparation of N-acetyl-neuraminic acid, amino acids. JP1010995A (Ajinomoto Co.,
PCT WO 94/29476. Ltd.).
28 Mahmoudian, M., Noble, D., Drake, C.S., 38 Yamada, H. (1998) Screening of novel
Middleton, R.F., Montgomery, D.S., enzymes for the production of useful
Piercey, J.E., Ramlakhan, D., Todd, M., compounds, in New Frontiers in Screening
and Dawson, M.J. (1997) An efficient for Microbial Biocatalysis, Studies in Organic
process for production of N- Chemistry, vol. 53 (eds K. Kieslich, C.P. van
acetylneuraminic acid using N- der Beek, J.A.M. de Bont, and W.J.J.
acetylneuraminic acid aldolase. Enzyme van den Tweel), Elsevier, Amsterdam,
Microb. Technol., 20, 393–400. pp. 13–17.
29 Zak, A. and Dodds, D.R. (1997) 39 Cornils, B., Herrmann, W.A., Schl€ogl, R.,
Application of biocatalysis and and Wong, C.-H. (2000) Catalysis
biotransformations to the synthesis of from A to Z, Wiley-VCH Verlag GmbH,
pharmaceuticals. Drug Discov. Today, 2, Weinheim.
513–531. 40 Knowles, W., Sabacky, M., Vineyard, B.,
30 Ghisalba, O., Gygax, D., Kragl, U., and and Weinkauff, D. (1975) Asymmetric
Wandrey, C. (1991) Enzymatic method for hydrogenation with a complex of rhodium
N-acetylneuraminic acid production, and a chiral bisphosphine. J. Am. Chem.
Novartis EP 0428947. Soc., 97, 2567–2568.
31 Kragl, U., Gygax, D., Ghisalba, O., and 41 Kara, S. and Liese, A. (2010) Enzymatic
Wandrey, C. (1991) Enzymatic process for CC bond formation, asymmetric, in
preparing N-acetylneuraminic acid. Encyclopedia of Industrial Biotechnology:
Angew. Chem. Int. Ed. Engl., Bioprocess, Bioseparations, and Cell
30, 827–828. Technology, vol. 3 (ed. M.C. Flickinger),
32 Kragl, U., Kittelmann, M., Ghisalba, O., John Wiley & Sons, Inc., Hoboken, NJ,
and Wandrey, C. (1995) N- pp. 2034–2049.
j 1001
Part V
Hydrolysis and Formation of PO Bonds
and Oliver May.
j 1003
25
Hydrolysis and formation of PO Bonds
Ron Wever and Teunie van Herk
25.1
Introduction
The role of phosphate esters is vitally important for all cell processes [1–3]. These
esters play an essential part in photosynthesis, lipid metabolism and glycolysis,
nitrogen cycle, immune response, host–pathogen interactions, transmembrane
signaling, activation of metabolites, cellular control by protein phosphorylation and
dephosphorylation, and in numerous other biochemical reactions. Furthermore,
phosphor is part of the backbone of both DNA and RNA and phospholipids are the
main structural components of all cellular membranes. Several essential cofactors or
cosubstrates for enzyme-catalyzed reactions of significant synthetic importance
involve phosphate esters. For instance, nicotinamide adenine dinucleotide phos-
phate, in the oxidized (NADP þ ) or the reduced (NADPH) form, is an essential
cofactor for some enzymatic redox reactions [4]. Adenosine triphosphate (ATP)
represents the energy-rich phosphate donor for most biological and synthetic
phosphorylation reactions [5–7].
Phosphate ester containing compounds have also found applications as drugs,
are applied as seasoning or taste enhancer in the food industry, and are used as
active ingredients in cosmetics, for example, shampoo and shower gels [8]. Phos-
phate-containing prodrugs have been successfully utilized to overcome various
drug delivery problems that might otherwise have compromised the therapeutic
value of the parent drug [9]. Several glycosidase inhibitors that are orally admin-
istered as antiviral agent cause gastrointestinal problems since also the glycosidases
in the gastrointestinal tract are inhibited. By phosphorylation of a free hydroxyl
group of the drug the inhibition of glycosidases in the gastrointestinal tract is
substantially reduced and after uptake in the circulation phosphatases easily
remove the labile phosphate group [10]. The ionic nature of the phosphate group
in these prodrugs significantly improves the solubility and dissolution rate of poorly
soluble drugs, thereby increasing the bioavailability. However, since the high
polarity of monophosphate esters precludes their cellular uptake, various prodrugs
have also been devised [11, 12] in which lipophilic phosphate masking groups are
present. Phosphate esters are also valuable synthetic intermediates that can be used
and Oliver May.
1004 j 25 Hydrolysis and formation of PO Bonds
as a source of organolithium compounds [13], be dehydrated to yield alkenes [14],
be used as substrates for stereoselective displacement with Grignard reagents
[15, 16], or are activated building blocks in the synthesis of many carbohydrates. An
example is dihydroxyacetone phosphate (DHAP) that is used by DHAP-dependent
aldolases that catalyze the highly stereoselective synthesis of a wide variety of
natural and non-natural carbohydrates [17–19]. These sugar mimics inhibit a wide
range of carbohydrate-degrading enzymes and have an enormous therapeutic
potential in many diseases such as diabetes, viral infections, and lysosomal storage
disorders [20].
Unsurprisingly, given the importance of the phosphate group many chemical and
biochemical methods have been developed for the introduction of phosphate
groups into compounds. The chemical methods that currently exist for the
introduction of a phosphate group into a substrate molecule largely depend on
the substrate itself, since functional group tolerance is the key to facilitating efficient
phosphorylation. The most widely used phosphorylating reagent for alcohols is
phosphoryl chloride (POCl3) [21] but many other phosphorylating reagents such as
phosphorochloridates and N-phosphoryl oxazolidinones exist [22, 23]. All reagents
have their advantages and disadvantages but in general they are very reactive and
precautions have to be taken to prevent side reactions. In particular, for large
polyhydroxy compounds where alternative sites for chemical phosphorylation exist
various protection and deprotection steps have to be used during synthesis.
Furthermore, undesired side products such as oligophosphate esters are easily
formed [7]. Many of these problems can be eliminated when enzymatic phosphor-
ylation procedures are used. In addition, the enzyme-catalyzed introduction of
phosphoryl groups may be enantio- or regiospecific. Moreover, from an environ-
mental point of view the use of enzymes also has advantages since a considerable
amount of waste may be eliminated [24]. This chapter gives an overview of
structural and catalytic properties of the different dephosphorylating and phos-
phorylating enzymes and the phosphotransferases that have been used in synthetic
conversions. Since a perplexing array of enzymes exists, for clarity the enzyme class
is occasionally mentioned. For further details concerning the enzyme nomenclature
the reader is referred to the website of the International Union of Biochemistry and
Molecular Biology [25] and to Table 25.1. Further, considering the diversity and
complexity of reactions catalyzed, the phosphodiesterases and endo- and exonu-
cleases will not be discussed here.
25.2
Biological Phosphorylating Agents, Phosphate Esters, and Thermodynamic
Considerations
The possible transfer of a phosphate group from a phosphorylated donor to a

suitable acceptor by an enzyme is determined by thermodynamic rules. Choosing a
proper phosphate donor is essential since not all phosphorylated compounds will
acts as good phosphate donors. The choice is determined by the free energies of
25.2 Biological Phosphorylating Agents, Phosphate Esters, and Thermodynamic Considerations j 1005
Table 25.1 Reactions catalyzed by key enzymes used in phosphorylation, dephosphorylation, and
phosphate transfer reactions.
EC Enzymes Reaction catalyzed
EC 2.4.1 Phosphorylases Hexosyl transfer

EC 2.4.1.1 Phosphorylase [(1 ! 4)-a-D-glucosyl]n þ Pi ! [(1 ! 4)-
a-D-glucosyl]n1 þ a-D-glucose 1-phosphate
EC 2.4.1.7 Sucrose phosphorylase Sucrose þ Pi ! D-fructose þ a-D-glucose
1-phosphate
EC 2.4.1.8 Maltose phosphorylase Maltose þ Pi ! D-glucose þ b-D-glucose
1-phosphate
EC 2.4.1.64 a,a-Trehalose a,a-Trehalose þ Pi ! D-glucose þ b-
phosphorylase D-glucose 1-phosphate
EC 2.4.1.230 Kojibiose phosphorylase 2-a-D-Glucosyl-D-glucose þ Pi ! D-glucose
þ b-D-glucose 1-phosphate
EC 2.7 Phosphotransferases Transfer of phosphate from ATP to an acceptor
(kinases)
EC 2.7.1 Alcohol group as acceptor ATP þ R-OH ! ADP þ R-OP
EC 2.7.2 Carboxy group as ATP þ R-COO ! ADP þ R-COOP
acceptor
EC 2.7.3 Nitrogenous group as For example, creatine kinase: ATP þ creatine
acceptor ! ADP þ phosphocreatine
EC 2.7.4 Phosphate group as For example, polyphosphate kinase: ATP þ (P)n
acceptor ! ADP þ (P)n þ 1
EC 3.1.3 Hydrolases Hydrolysis of phosphoric monoester
EC 3.1.3.1 Alkaline phosphatase R-O-P þ H2O ! R-OH þ Pi
EC 3.1.3.2 Acid phosphatase R-O-P þ H2O ! R-OH þ Pi
EC 3.1.3.5 5-Nucleotidase 5-Ribonucleotide þ H2O !
ribonucleoside þ Pi
EC 3.1.3.16 Phosphoprotein Phosphoprotein þ H2O ! protein þ Pi
phosphatase
EC 3.1.3.26 Phytase myo-Inositol hexakisphosphate þ H2O ! 1-D-
myo-inositol-1,2,3,5,6-pentakisphosphate þ Pi
EC 3.6 Hydrolases Reacting on acid anhydrides
EC 3.6.1.5 Apyrase ATP þ 2H2O ! AMP þ 2Pi
EC 4.1.2 Lyases Reacting on CC bonds
EC 4.1.2.4 Deoxyribose-phosphate 2-Deoxy-D-ribose 5-phosphate ! D-glyceralde-
aldolase hyde 3-phosphate þ acetaldehyde
EC 4.1.2.13 Fructose-bisphosphate D-Fructose 1,6-bisphosphate ! glycerol-phos-
aldolase phate þ D-glyceraldehyde 3-phosphate
EC 5.4.2 Mutases Phosphotransferases
EC 5.4.2.2 Phosphoglucomutase a-D-Glucose 1-phosphate ! D-glucose
6-phosphate
hydrolysis (DG0 hydro) of both phosphate donor and acceptor. This so-called phos-
phorylating potential is used to compare the ability of different compounds to
effectively transfer a phosphoryl group. Table 25.2 summarizes the phosphorylating
potentials of several important biological compounds having phosphoryl donor
Table 25.2 Standard free energies of hydrolysis for common metabolites [26].
Metabolite DG0 hydro (kJ mol1)
Phosphoenolpyruvate 62
Carbamoyl phosphate 51
1,3-Bisphosphoglycerate 49
Acetyl phosphate 43
Phosphocreatine 43 High-energy compounds
Pyrophosphate (PPi) 33
Phosphoarginine 32
ATP ! AMP þ PPi 32
Acetyl CoA 32
ATP ! ADP þ Pi 30
Glucose 1-phosphate 21 Low-energy compounds
Glucose 6-phosphate 14
Glycerol 3-phosphate 9
AMP ! adenosine þ Pi 14
abilities [26]. Each compound is capable of driving the phosphorylation of com-

pounds that are lower on the scale of Table 25.2. Adenosine 50 -triphosphate (ATP) is
by far the most important phosphorylating agent in biological systems and it is used
by numerous kinases. Phosphoenolpyruvate, an intermediate in the glycolytic
pathway, has the highest phosphorylating potential known. Phosphorylation of
substrates by low-potential phosphorylating agents is thermodynamically not
favored and the equilibrium is in the direction of the donor. By coupling to another
reaction that is thermodynamically more favored such a reaction is enabled [5] in
biological systems.
The standard free energy of phosphocreatine is higher than that of nucleoside
triphosphates such as ATP; therefore, adenosine diphosphate and cytidine diphos-
phate are readily phosphorylated by phosphocreatine to their respective tripho-
sphates in the presence of an appropriate enzyme [27]. The standard free energy
is higher because phosphocreatine is a phosphoamide. The difference in phosphor-
ylating potential between glucose-1- and glucose-6-phosphate explains why in an
enzyme-catalyzed interconversion the equilibrium is in the direction of glucose-6-
phosphate.
From Table 25.2 it can be seen that the standard free energy of hydrolysis of
pyrophosphate (PPi) is even somewhat higher than that of ATP. Pyrophosphate
(H4P2O7) is a particular case of a phosphate monoester that unlike the other
phosphorylated compounds can be simply prepared, is very stable, and is an
inexpensive bulk chemical. This also holds for polyphosphates. Table 25.2 also
shows that hydrolysis of phosphate esters is always favored thermodynamically
but this process is kinetically restricted. Phosphatases in fact lift the kinetic
restrictions and are able to hydrolyze these compounds in several ways and heat
is liberated.
25.3 Enzymatic Phosphoryl Transfer Reactions and Phosphorylated Intermediates j 1007
25.3
Enzymatic Phosphoryl Transfer Reactions and Phosphorylated Intermediates
25.3.1
Phosphorylation by Kinases
In biological systems, phosphate esters are usually produced by phosphorylating

enzymes belonging to the class of kinases, which catalyze the transfer of a phosphate
moiety (or a di- or triphosphate moiety in certain cases) from ATP [5–7] to an acceptor
molecule. Other nucleoside triphosphates have similar phosphorylating potentials
but they are rarely used as phosphoryl group donors [28, 29]. Not all these kinases will
be discussed here but only those that have found application in the synthesis of
phosphorylated compounds. They all belong to the class EC 2.7.1–2.7.4 [25] and use
an alcohol group, a carboxy group, a nitrogenous group, or a phosphate group as an
acceptor.
Hexokinase (EC 2.7.1.1) is an enzyme that uses ATP to phosphorylate D-glucose
to D-glucose-6-phosphate, a useful reagent for the regeneration of nicotinamide
cofactors, in a one-step reaction [30–32]. The enzyme has broad substrate specificity
since a wide range of pyranoid and furanoid analogues of glucose, including
fluorinated analogues, are selectively phosphorylated on the primary alcohol moiety
located at position 6 [33, 34]. D-Arabinose, a pentose, is also a substrate for
hexokinase [35]. Ribokinase (EC 2.7.1.17) can phosphorylate D-ribose to D-ribose-
5-phosphate [36]. Glycerol kinase [37] (EC 2.7.1.30) has also a broad specificity since
it not only accepts its natural substrate glycerol to form sn-glycerol-3-phosphate [38]
and close analogues such as dihydroxyacetone [39, 40] (Section 25.4.2.3) but it also
phosphorylates prochiral or racemic primary alcohols into chiral phosphates with
very high enantiomeric excesses (e.e.) (>90–95%) and yields of 75–95% [41, 42].
Adenylate kinase (EC 2.7.4.3) has been used in the synthesis of several nucleotide
phosphate analogues. For example, ribavarin triphosphate, a compound with
antiviral properties, was prepared from ribavarin monophosphate with adenylate
kinase [43]. Other nucleotide analogues, for example, ATP-a-S and ATP-c-S have
been synthesized using immobilized enzymes [44, 45]. NAD kinase (EC 2.7.1.23)
phosphorylates nicotinamide adenine dinucleotide (NAD þ ) into its more expensive
phosphate NADP þ [46]. It was possible to synthesize 8-azido-[20 -32 P]NADP(H) as a
photo-affinity label for NADP(H)-specific enzymes using [c-32 P]ATP and a NAD
kinase [47].
25.3.2
Enzymes Used in the Regeneration of ATP
The use of kinases in the synthesis of phosphorylated compounds has the disad-
vantage that they are in general specific for their substrates. In addition, the formed
ADP has to be recycled for economical reasons and also the concentration of ADP has
to be controlled to prevent accumulation of ADP since this may inhibit the reactions.
Thus, ATP must be used in catalytic amounts and continuously regenerated [40, 48, 49].
There are several kinases available for this purpose that use a cheaper phosphate donor
to convert ADP back into ATP (Scheme 25.1).
Scheme 25.1 Phosphorylation of alcohols by ATP consuming kinases and enzymatic ATP
recycling systems.
These methods have in common that phosphoryl groups are transferred from a
high-energy donor (cf. Table 25.2) to ADP. For most synthetic applications, either
phosphoenolpyruvate (PEP)/pyruvate kinase (PK) or acetyl phosphate (AcP)/acetyl
kinase (AcK) are used to regenerate ATP. However, the phosphor donor compounds
are either unstable or difficult to synthesize and consequently they are expensive. For
example, acetyl phosphate is easily prepared [50] but it is unstable in solution and its
application is limited to fast phosphorylation reactions. Further, the regeneration
system is quite sensitive to pH changes. Phosphoenolpyruvate is a strong phos-
phorylating agent (Table 25.2) but its synthesis is more elaborate [51] and both
phosphoenolpyruvate and the product pyruvate may inhibit kinases [34]. Therefore,
the reaction can only be carried out in diluted solutions. For a more detailed
discussion on the use of the ATP regenerating kinases and their substrates see
Reference [7].
Interestingly, several Gram-positive bacteria such as Arthrobacter sp. and Myco-
bacterium tuberculosis posses inorganic polyphosphate (poly(P))-dependent kinases
that use poly(P) instead of ATP as the phosphoryl donor [52]. Poly(P) is a biopolymer
of several orthophosphate residues linked by a high-energy phospho-anhydride bond
in energy approximately equivalent to that of ATP. This biological high-energy
compound is presumed to be an ancient energy carrier preceding ATP [52]. Several
of these poly(P) dependent kinases that use poly(P) are known to function in bacteria.
Some of these enzymes also use ATP as donor. Examples are the poly(P)/ATP-
glucomannokinase [53], poly(P)/ATP-NAD kinases [54], and poly(P)/ATP glucoki-
nase [55]. A specific poly(P)-glucokinase has also been described [55] and the crystal
structure of poly(P)/ATP glucomannokinase has been determined [56]. Comparison
of this structure to the structure of hexokinase has allowed the conclusion that some
ATP specific proteins have evolved from a primordial poly(P) glucomannokinase and
have lost the ability to use poly(P) during the evolution [56].
The poly(P)-glucose-6-phosphotransferase from Mycobacterium phlei was also used
to study the phosphorylation of glucose to glucose-6-phosphate [57] with poly(P) as a
phosphate donor. The immobilized enzyme was used in a continuous enzyme
reactor to generate glucose-6-phosphate. Recently a method was published that
described the preparation of a wide variety of C6 phosphorylated-aldohexoses and C6
25.4 Phosphate Hydrolyzing Enzymes: The Phosphatases j 1009
phosphorylated D-aldohexose derivatives using poly(P)-glucose phosphotransferases
from Mycobacterium species and poly(P) as sole and cheap source of phosphate
donor [8]. A poly(P)/AMP phosphotransferase has been described [58] that catalyzes
the phosphorylation of AMP to ADP. It has been used in an ATP regenerating system;
however, an adenylate kinase also has to be present to convert ADP into ATP. Poly(P)
as a direct phosphate donor takes away the disadvantage of the use of ATP and its
associated regeneration [32, 59]. It would be interesting to see whether kinases can be
transformed by directed evolution to accept polyphosphate rather than ATP.
25.4
Phosphate Hydrolyzing Enzymes: The Phosphatases
Most dephosphorylations in vivo are catalyzed by one class of enzymes: the phos-
phatases (EC 3.1.3). These enzymes hydrolyze organic phospho-esters and are crucial
to life. Many different mechanisms have evolved in nature for cleaving the PO bond
and a stunning variety of very different phosphatases can be found [60]. Many
phosphatases employ metal ions to lower the activation energy for PO bond fission.
Some phosphatases have evolved specialized functions relevant to microbial viru-
lence [61], signal transduction [62], and energy conversion and metabolism [63]. An
example of the latter is the glucose-6-P phosphatase found in mammalian liver which
regulates the glucose concentration in the blood stream [64]. Each of these enzymes
has its own mechanism to deal with the PO bond. A cursory overview of the
different enzymes present in nature will be given here and the properties and catalytic
mechanism of only those phosphatases that have been used in biotransformations
will be discussed in some detail below. The rationale behind this is that when one
wants to use enzymes successfully in biotransformations at least some details of the
enzyme mechanism should be known.
Alkaline phosphatases (EC 3.1.3.1) belong to the non-specific phospho-mono-
esterases, which have maximal activity at pH 9–10 (hence their name). The Zn- and
Mg-containing enzymes are found in both prokaryotes and eukaryotes and are the
most well-studied phosphatases. For a more detailed overview on alkaline phospha-
tase the reader is referred to References [65, 66] and Section 25.4.1. Purple acid
phosphatases (EC 3.1.3.2) also belong to the group of metal-containing phospha-
tases [67]. The enzymes occur in bacteria, plants, and animals and contain Fe ions in
the active site. They hydrolyze phospho-monoesters and the phosphoserine residue
of phosphoproteins via direct attack by a metal coordinated hydroxide nucleophile.
Transfer of the phosphoryl group to water takes place without formation of a
phospho-enzyme intermediate. The optimum activity is at low pH (pH 4–7). The
use of these enzymes in synthetic reactions has not been reported.
Non-specific enzymes that do not require metal ion cofactors also exist. These
enzymes have a molecular mass of 40–60 kDa and a very acidic optimum pH of
around 2.5. They are found in various sources, including plants, yeast, bacteria, and
human and animal material. These enzymes belong to the histidine superfamily [68],
which covers a large functionally diverse group of proteins including the phytases
from bacterial or fungal origin. Crystal structures are available for the rat acid
phosphatase (E.C.3.1.3.2) [69] and phytase (EC 3.1.3.26) [70]. A histidine residue is
present in the active site that functions as a nucleophile and positively charged
residues bind and position the phosphate monoester. During catalysis a phospho-
histidine intermediate is formed that is hydrolyzed, but in the presence of a suitable
acceptor transphosphorylation occurs. It was shown already in 1958 [71] for the
prostate enzyme that certain hydroxyl containing compounds could compete with
water and that glucose, propanediol, glyceraldehyde, and dihydroxyacetone were
phosphorylated using creatine phosphate as a phosphate donor. Phytases have found
interesting applications. The enzyme hydrolyses in several steps phytate, the primary
storage of phosphorus in plant seeds, into inositol and inorganic phosphate [72].
Addition of phytases to the diet of animals increases the availability of plant
phosphorus, which reduces the need for addition of phosphate to the diet. As a
result the phosphorus load in the environment decreases [73]. Inositol monopho-
sphatase and fructose-1,6-bisphosphate 1-phosphatase also belong to this histidine
superfamily. The phytases and the histidine phosphatases should not be confused
with another group of acid phosphates that up to now have only been found in
bacteria. As will be discussed in Section 25.4.2 the active site of these enzymes is very
different in its architecture though nature has used again histidine [68] and basic
residues to bind phosphate and to split the PO bond.
A very diverse group of enzymes responsible for the dephosphorylation of a range
of phosphoproteins are the protein phosphatases (EC 3.1.3.16). Most of them are
involved in control of cellular processes [74]. They do not require metals for their
action and the group can be further split by function into a group that depho-
sphorylates phosphotyrosine residues and a panoply of enzymes that dephosphor-
ylate phosphoserine or phosphothreonine residues within proteins [75]. Some
phosphotyrosine enzymes have been crystallized and the X-ray structures show
[76, 77] that the nucleophile in the active site is a cysteine residue. During catalysis a
phosphocysteine intermediate is formed that normally transfers its phosphate group
to water but it can also phosphorylate alcohols [78]. As far as we know this has not
been explored further.
25.4.1
Structural and Mechanistic Description of Alkaline Phosphatase
Alkaline phosphatases are widely spread in nature and are found in both eukaryotes
and prokaryotes. They appear to act strictly as non-specific phosphomonoesterases.
The enzymes are dimeric metalloproteins with two Zn2 þ ions and one Mg2 þ ion in
each active site region. All three metal ions are involved in catalysis. Many X-ray
structures are available and the reaction of phosphate monoesters with the enzyme
is known in detail [65, 66, 79]. Scheme 25.2 illustrates the sequence of events
during catalysis.
The Mg2 þ ion in the active site activates a bound water molecule, making it a better
nucleophile. Upon binding of the phospho-monoester the OH group of the serine
residue (Ser102 in the enzyme from E. coli) becomes fully deprotonated for
Scheme 25.2 Overall scheme of the reaction in which both hydrolysis and transferase
activities are given. E-P is the covalent serine-phosphate intermediate. After hydrolysis E.P is formed,
a species in which Pi is still bound to the active site. Modified according to Reference [65].
nucleophilic attack on the phosphorus center. When the phospho-monoester reacts

with the serine in the active site a covalent phosphoserine intermediate is formed and
the alcohol group is released. This intermediate corresponds to the covalent species
E-P in Scheme 25.2. In the next step a nucleophilic hydroxide ion coordinating to one
of the Zn2 þ ions attacks the phosphor atom, hydrolyzing the covalent serine
phosphate intermediate, and the E.P intermediate is formed in which phosphate
is still bound to the active site.
As illustrated in Figure 25.1 the phosphate group in this intermediate (EP) is very
closely associated with the Zn2 þ ions. It bridges both Zn2 þ ions and is hydrogen
bonded to a water molecule coordinated to the Mg2 þ ion. The other two oxygen
atoms are tightly held by the amino functions of the guanidine group of Arg160.
At acidic pH the hydrolysis of the phospho-enzyme (E-P) is rate limiting while at
alkaline pH the dissociation of the non-covalent enzyme-phosphate complex (E.P) is
the rate-determining step [65].
As was already discovered a very long time ago, the alkaline phosphatases are also
able to catalyze the transfer of the phosphate from E-P to various acceptor alcohols
and glucose (Scheme 25.2) [80–82]. In particular, tris(hydroxymethyl)aminomethane
(Tris) is a very good acceptor and the phosphate monoester of Tris is formed [65, 82].
Many other compounds participate as active acceptors in the phosphotransferase
reaction, including ethanolamine, diethanolamine, serine, and glycerol [83]. The
formation of glycerol phosphate and Tris-phosphate has been studied by 31 P NMR as
a function of pH [84]. The transferase activity of the amino alcohol shows a bell-
shaped pH dependency. Aliphatic alcohol acceptors show small increases in acceptor
activity between pH 6 and 8, with fivefold increases from pH 8 to 10. The rapid
increase in the phosphate transfer to aliphatic alcohols is compatible with a reaction
in which the alcohol group is in its deprotonated form when bound to the active site.
This is probably due to binding of the alcohol group to one of the Zn2 þ ions that
probably lowers the pKa of the alcohol group sufficiently.
Figure 25.1 Schematic drawing of the interactions of the phosphate group with the zinc ions
and the guanidinium group in alkaline phosphatase. This structure corresponds to the species E.P in
Scheme 25.2. Reproduced with permission from Reference [79].
25.4.1.1 Application of Alkaline Phosphatases in Dephosphorylation

Alkaline phosphatases are used as non-specific phosphatases, for example, in the
hydrolysis of polyprenol phosphates like 6,7-epoxygeranyl diphosphate and 6,7-epoxy
bis-homogeranyl diphosphate [85] and in sphingoside base 1-phosphate analysis of
biological samples [86]. A regioselective dephosphorylation of 20 -carboxyl-D-arabini-
tol 1,5-bisphosphate was used in the synthesis of 20 -carboxy-D-arabinitol 1-phosphate,
a natural inhibitor of ribulose 1,5-bisphosphate carboxylase [87]. The hydrolysis of the
5-phosphoryl group by alkaline phosphatase gave a 4 : 1 mixture of the 1- and 5-
phosphate derivatives. In contrast, acid phosphatase converted 20 -carboxyl-D-arabi-
nitol 1,5-bisphosphate almost quantitatively into the 1-phosphate derivative [87].
Alkaline phosphatases were also used in the hydrolysis of nucleotides [88] and
aromatic phosphate esters [89] as potential chemiluminescent 1,2-dioxetane based
compounds [90]. Alkaline phosphatases are also a very useful tool in molecular
biology. DNA normally possesses phosphate groups on the 50 end and the enzymes
are able to remove these phosphates to prevent the DNA from ligating the 50 end to
the 30 end [91].
25.4.1.2 Transphosphorylation by Alkaline Phosphatases

Pradines et al. [92, 93] were the first to use alkaline phosphatases from E. coli, calf, and
chicken intestine for synthetic purposes. Many phosphorylated alcohols, diols, and
polyols were synthesized on a preparative scale using PPi and other phosphate
donors. To prevent hydrolysis by water and to shift the equilibrium towards transpho-
sphorylation a very high concentration of acceptor was required. Interestingly, the pH
optimum for the hydrolysis (pH 6) of the phosphorylated compounds by alkaline
phosphatase was reported to be very different compared to the transphosphorylation
reaction (pH 8) using PPi as phosphate donor, in line with the earlier results by
Gettins et al. [84]. Glycerol phosphorylation by alkaline phosphatase and PPi was
optimized and glycerol-3-phosphate could be prepared. Only a minor fraction of
glycerol-2-phosphate was formed. The reaction is therefore regioselective but not
stereoselective since, according to the investigators, a racemic mixture of DL-glycerol-
3-phosphate was formed. Polyphosphate could also be used as a phosphate donor in
the transphosphorylation but was less effective than PPi. Surprisingly, it was also
possible to use plain phosphate as a donor and considerable amounts (35 g l1) of
glycerol phosphate were synthesized using immobilized alkaline phosphatase in a
packed bed reactor with a very high concentration of glycerol (80%) and 0.4 M
phosphate [93]. This reaction corresponds to formation of the EP intermediate in
Scheme 25.2 and subsequent phosphotransferase step to glycerol. It is surprising
that this process to produce phosphorylated compounds has received so little
attention given its simplicity. The very high concentration of glycerol is easily
removed from the product by precipitating phosphate at high pH with magnesium
chloride and binding the ester to ion-exchange material (AmberliteÔ). After sub-
sequent elution with sodium hydroxide the solution is neutralized with sulfuric
acid [93].
25.4.2
Structural and Mechanistic Description of Acid Phosphatases
A family of enzymes that has been used more recently in phosphorylation and
dephosphorylation processes belongs to non-specific acid phosphatases (NSAPs) [94]
from enteric bacteria. These are non-metal soluble periplasmic proteins or mem-
brane-bound lipoproteins able to hydrolyze a broad range of structurally unrelated
organic phospho-monoesters and are therefore called non-specific. The optimal pH
for this class of enzymes is at acidic to neutral pH values. NSAPs are monomeric or
oligomeric enzymes containing a subunit with a Mr of 25–30 kDa. On basis of amino
acid sequences three different families of NSAPs were identified: molecular class
A [95], B [96], and C [97], which are completely unrelated at the sequence level. Class A
NSAPs possess a conserved sequence motif, K-(X6)-R-P-(X12–54)-P-S-G-H-(X31–54)-S-
R-(X5)-H-(X2)-D [98], with three domains that are also found in several lipid
phosphatases, mammalian glucose-6-phosphatase and vanadium haloperoxi-
dases [99–103]. From the X-ray structures available it is known that these conserved
residues form the active site in these enzymes and the architecture of the active site is
essentially identical.
The detailed X-ray structures of the acid phosphatases from Escherichia blattae [104]
and Salmonella typhimurium [105] and mutational analysis [106] of active site residues
have given detailed insight into the mechanism of phospho-ester hydrolysis. The
active site scaffold (Figure 25.2) of these enzymes consists of Lys123 (numbering as
in Salmonella typhimurium [105]), Arg130, Ser156, Gly157, His158, Arg191, and
His197. It is generally accepted that the His197 carries out a nucleophilic attack at
Figure 25.2 Active site of the acid phosphatase and Arg191 residues interact with the
from Salmonella enterica ser. typhimurium phosphate oxygen atoms, keeping the
MD6001 co-crystallized with phosphate phosphate group of the substrate close to
[105, 106]. The substrate-binding site consists of His197. Val78 is located at the entrance of the
invariant Lys123, Arg130, Ser156, Gly157, active site. The figure was created with YASARA
His158, Arg191, and His197 residues. The side- (www.yasara.org) and POVRay (www.povray.
chain atoms of Lys123, Arg130, Ser156, Gly157, org).
the phosphor center, affording a phosphohistidine intermediate, and that His158 acts
as a general acid/base catalyst.
The other residues bind and position the phosphate group in the active site.
Scheme 25.3 illustrates the reaction mechanism and the various steps during
catalysis. Once the phosphohistidine intermediate is formed a water molecule enters
the active site and His158 accepts a proton from the water and turns this into a strong
nucleophile, leading to the release of inorganic phosphate. The active site of these
enzymes is exposed and this may explain the broad enzyme specificity.
The active sites of the acid phosphatases and of the apo-chloroperoxidase, from
which vanadate, the prosthetic group, is removed, are nearly superimposable
[104, 107]. The vanadium apo-enzyme also possesses phosphatase activity, although
the turnover with p-nitrophenyl phosphate (pNPP) as a substrate is very slow
(1.2 min1) [108]. As in the alkaline phosphatase, hydrolysis of the phospho-inter-
mediate is the rate-determining step in the phosphatase activity of the apo-chlor-
operoxidase. By incubating apo-chloroperoxidase crystals with pNPP and subsequent
flash cooling of the crystals it was possible to trap this intermediate and to obtain a
high-resolution X-ray structure [109]. The intermediate formed in the hydrolysis of a
phosphorylated substrate consists of a metaphosphate anion PO3 covalently bound
via its phosphorous atom to the Ne2 atom of a histidine with a water molecule in the
position for a nucleophilic attack on the phosphorus. It is very likely that such an
intermediate is also formed during catalysis of the acid phosphatases.
Scheme 25.3 Reaction mechanism of the hydrolysis of phosphorylated compounds according to Renirie et al. [108] and Makde et al. [105, 106].
The first step involves the formation of the phospho-enzyme intermediate and release of alcohol. In the second step, water (a nucleophile
activated by His158) attacks the phosphorus center of the phospho-enzyme intermediate, leading to the release of the inorganic phosphate moiety.
25.4 Phosphate Hydrolyzing Enzymes: The Phosphatases
j 1015
Class A NSAPs are further classified into classes A1, A2, and A3 depending upon
the amino acid sequences, substrate specificities, and inhibition effects [94, 97]. The
acid phosphatases from Shigella flexneri (PhoN-Sf) and Escherichia blattae (EB-NSAP)
that have been studied in detail belong to class A1 NSAPs [110] and these enzymes
exhibit broad substrate specificity. They can hydrolyze 50 - and 30 -nucleotide mono-
phosphates (NMPs), hexose-, pentose-, and aryl phosphates, such as pNPP and
phenolphthalein phosphate, but not diesters [94].
The prototype of class A2 NSAPs is the non-specific acid phosphatase from
Salmonella enterica ser. typhimurium (PhoN-Se) [111, 112]. The enzyme is active
against a very broad array of substrates, showing even wider substrate specificity than
that of class A1 enzymes.
An enzyme belonging to the class A3 group is the apyrase [113] from Shigella
flexneri (Apy-Sf). The enzyme shows a distinctive activity on nucleotide tripho-
sphates (NTPs), which are hydrolyzed to corresponding nucleotide diphosphates
(NDPs). The enzyme hydrolyses PPi, but pNPP is a poor substrate. Because of the
nucleotide triphosphates hydrolyzing activity and its optimum pH (7–7.5) Apy-Sf
can be considered as an ATP diphosphohydrolase or an apyrase (EC 3.6.1.5.).
Despite the functional dissimilarity with other NSAPs, it shows a striking similarity
in sequence with the other class A enzymes [94].
Class A1 NSAPs show higher phosphatase activity towards 50 -NMPs (primary
alcohol) rather than 30 -NMPs (secondary alcohol) whereas class A2 NSAPs can
hydrolyze both 50 - and 30 -NMPs equally well. Class A3 NSAPs hardly hydrolyze NMPs
but they catalyze the hydrolysis of NTPs.
25.4.2.1 Dephosphorylation by Acid Phosphatases and 50 Ribonucleotide

Phosphohydrolases
Dephosphorylation under mild conditions by acid phosphatases, without isolation of
the intermediate phosphate species, is a method frequently used to obtain chiral
polyol products [114–118]. Similarly, hydrolysis of polyprenyl diphosphate catalyzed
by acid phosphatase readily produced the corresponding dephosphorylated products
in acceptable yields and side reactions that normally occur during chemical hydro-
lysis were absent [119, 120].
A wide variety of chiral alcohols can be obtained when carboxyl esters are
hydrolyzed by lipases. In principle, phosphatases may also be able to hydrolyze
phosphate esters enantioselectively. However, only a few reactions have been
reported concerning enantioselective dephosphorylations. Scollar et al. [121]
reported that rac-threonine could be resolved into its enantiomers via hydrolysis
of its O-phosphate by an acid phosphatase. In addition, D-allo-threonine and
D-threonine have been resolved by acid phosphatases [122]. The potential applica-
tion of the acid phosphatases in this area was further explored more recently [123].
It was shown that the acid phosphatase from Salmonella enterica showed very high
enantioselectivity (E > 200) in the hydrolysis of O-phospho-DL-threonine, with a
preference for the L-isomer. This means that this dephosphorylation reaction can be
used for preparative synthesis of L-threonine from racemic phosphothreonine. By
random mutagenesis two variants were obtained from a library of 9600 mutants
that showed an increased selectivity in the hydrolysis of O-phospho-DL-serine with
preference for the D-isomer.
A very good example of a specific enzymatic dephosphorylation is the hydrolysis of
()-50 -phosphorylated aristeromycin by 50 -ribonucleotide phosphohydrolase (EC
3.1.3.5) or shortened 50 -nucleotidase from Crutalus atrox venom. The ()-enantiomer
of aristeromycin shows some cytostatic and antiviral activity, while the ( þ )-enan-
tiomer is inactive [124]. Only the ()-enantiomer is selectively hydrolyzed by the
phosphohydrolase. The ()-alcohol and the ( þ )-50 -phosphate derivative were sep-
arated easily on a silica gel column. Subsequent hydrolysis of the ( þ )-enantiomer
with a non-specific alkaline phosphatase yielded pure ( þ )-alcohol. Similarly, after
phosphorylation with a thymidine kinase and ATP the racemic fluorinated analogues
of guanosine were resolved with the 50 -ribonucleotide phosphohydrolase and after
separation the non-accepted enantiomer was hydrolyzed by alkaline phospha-
tase [125]. The 50 nucleotidase is also present in bacteria, plants, and mammals [126]
but little is known about the mechanism of hydrolysis.
25.4.2.2 Transphosphorylation by Acid Phosphatases

It has been known for some time that acid phosphatases can carry out transpho-
sphorylation reactions [127], that is, the transfer of phosphate from one molecule
(donor phosphate, e.g., phospho-monoesters or PPi) to another different molecule
(acceptor alcohol). Axelrod [128] and Appleyard [129] were the first to demonstrate that
organic phosphates can be synthesized by enzymatic transfer of the phosphate group
from organic donor compounds to a suitable alcohol function. Axelrod [128] used a
phosphatase from orange juice and pNPP and showed that primary alcohols are better
substrates than secondary alcohols. Appleyard [129] using prostatic extracts and
phenolphthalein diphosphate as phosphate donor showed similar results. Later it
was shown that the bacterial enzyme from Salmonella sp. [112] is also able to
phosphorylate simple alcohols, polyalcohols, and nucleosides with pNPP or ribonu-
cleoside monophosphates as donors, showing the broad substrate specificity of the
enzyme. Interestingly, this phosphatase is able to catalyze the transfer of the phosphate
group of 30 -AMP from the 30 -position both to the 20 - and 50 -positions. Thus the enzyme
also has phosphomutase activity.
Transphosphorylation of alcoholic substrates is thought to be a two-step reaction.
Scheme 25.4 shows the transphosphorylation from PPi and competing hydrolysis by
water catalyzed by the bacterial acid phosphatases [130]. First the enzyme binds to the
phosphate donor to form a phosphoryl intermediate. This intermediate is probably a
metaphosphate anion covalently bound to a nitrogen atom of a histidine in the active
site [109].
In the second step the phospho-enzyme intermediate is either attacked by water
(hydrolysis) or by an alcoholic acceptor, resulting in (trans)phosphorylation. The Km
for the alcohol, therefore, is a very important factor that determines whether effective
phosphorylation occurs. When the affinity for the substrate is low the phosphorylated
enzyme intermediate prefers to react with water, resulting in hydrolysis.
The group headed by Asano showed in pioneering studies that NSAPs transfer a
phosphate group from PPi to nucleosides [127]. PPi is a very simple compound that
1018
j 25 Hydrolysis and formation of PO Bonds
Scheme 25.4 Overall mechanism of phosphorylation and dephosphorylation catalyzed by acid phosphatases. The enzyme reacts with PPi to produce a binary
PPi–enzyme complex (1). This complex dissociates (2) to yield an activated phosphorylated enzyme intermediate (E.Pi). A reaction (3) with water may occur, resulting in
dissociation of the intermediate as well as hydrolysis of PPi. The intermediate may also transfer (4) the phosphate to a bound acceptor (R-OH), which dissociates
(5) to form a phospho-monoester and the free enzyme. Hydrolysis of phospho-monoesters also proceeds via the E.Pi intermediate. Modified from Reference [130].
can easily be synthesized from phosphate at low cost [131]. It is a safe compound and
is used as a food additive. However, PPi has a chelating effect and binds multivalent
metals such as Ca2 þ , Mg2 þ , and Fe2 þ and care should be taken to use it in
combination with phosphatases that require metal ions because their activity may be
inhibited. Class A acid phosphatases do not require metal ions, and therefore PPi can
be used as cheap phosphate donor. Nucleotides are often used as food additives and as
pharmaceutical intermediates. Their biological activity is related to the position of the
phosphate group. Inosine 50 -monophosphate (50 -IMP) or guanosine 50 -monopho-
sphate (50 -GMP) are used as a flavor potentiator (umami) in various foods whereas
the 20 -monophosphates are tasteless [127]. Considering the worldwide production of
nucleotides of approximately 16 000 t year1 [132] it is no surprise that considerable
effort has been put into optimizing the nucleoside phosphorylation process by the
phosphatase. The advantages of this new process are the simplicity, low cost, and mild
reaction conditions. Initially the enzyme from Morganella morgani was selected as a
50 -IMP producer [133]. However, there were several problems to be solved. Firstly, the
solubility of nucleosides is often below the Km value of the enzyme and, secondly, all
of the synthesized 50 -NMP is hydrolyzed again to the nucleoside as the reaction time
is prolonged and the PPi is consumed. To suppress the dephosphorylation reaction a
random mutagenesis approach was used. A mutant acid phosphatase was obtained
that was able to produce a considerable amount of 50 -IMP from inosine. Many other
Enterobacteriaceae produce acid phosphatases and their phosphotransferase activity
was also investigated [130, 134]. Since the X-ray structure of the Escherichia blattae
phosphatase was available the Japanese group carried out a rational site-directed
mutagenesis study [135]. The final triple mutant obtained had a productivity that
amounted to 140 g l1 with a molar yield of 71% from inosine. The increased
productivity probably relates to a decrease in the Km value for inosine or alternatively
the hydrolysis by water of the phospho-enzyme intermediate is suppressed.
The enzymatic procedure based upon the use of inosine kinase from Escherichia
coli as a phosphorylating enzyme using ATP [136] probably cannot compete with the
above process. The kinase requires ATP, which needs to be regenerated, making the
process more complex. Similarly, 50 -nucleotides can be obtained by a chemical
method [137] but this is not acceptable due to the complexity and toxicity of the
reagents during chemical phosphorylation.
Since the acid phosphatases were able to (regiospecifically) phosphorylate the
ribose group in inosine, it came as no surprise that many simple carbohydrates
were phosphorylated as well by PhoN-Sf (acid phosphatase from Shigella flexneri)
and PhoN-Se [130, 138]. Both phosphatases can phosphorylate D-glucose to
D-glucose-6-phosphate using PPi as phosphate donor in a very efficient manner.
Under optimized conditions 80 mM of glucose-6-phosphate could be obtained from
100 mM PPi and 400 mM glucose. The utility of the enzymatic method to produce
glucose-6-phosphate in a preparative manner was also shown. Simple alcohols,
polyalcohols, and cyclic and aromatic alcohols could be phosphorylated by PhoN-Sf
and PhoN-Se [112, 139], showing the broad substrate specificity of these enzymes.
Enantioselectivity in the phosphorylation of a secondary alcohol group could not be
demonstrated.
25.4.2.3 Formation of DHAP
Dihydroxyacetone (DHA) is among the substrates that are phosphorylated by both
PhoN-Sf and PhoN-Se [140]. The formed dihydroxyacetone phosphate (DHAP) is the
key compound in aldol condensations using DHAP-dependent aldolases, resulting
in a CC coupled product with two new stereocenters with high optical purity
[17, 18, 118]. For a detailed discussion of aldolases and aldol reactions please see
Chapter 21. While the substrate specificity for the aldehydes is rather relaxed, the
aldolases show only very limited tolerance for substituting the donor. Access to
DHAP is therefore very important in the use of these aldolases. DHAP is unstable in
particular at neutral to alkaline conditions and is generally produced as a stable
precursor. The chemical synthesis requires multistep procedures, including protec-
tion and deprotection steps [17, 118]. Commercially it is available on a small scale for
approximately 3000 D g1 (http://www.sigmaaldrich.com), which hampers its use in
large-scale synthesis (even if lower prices can be expected at larger scale) of
compounds by these aldolases.
Alternatively, DHAP may be enzymatically produced by retro-aldol cleavage of
fructose-1,6-biphosphate using the fructose-1,6-biphosphate aldolase and triose
isomerase [116] or transphosphorylation using other phosphate donors. Based on
this a highly integrated multienzyme system has been developed that converts
sucrose, fructose, or glucose into DHAP [141]. It is also possible to phosphorylate
dihydroxyacetone by dihydroxyacetone kinases [142, 143] or by glycerol kinase. This
enzyme has a broad substrate specificity and phosphorylates also dihydroxyace-
tone [39, 41, 144]. Most of these methods rely on ATP as a donor and regeneration of
ATP is required.
The possibility to use a cheap phosphate donor and circumventing the ATP
regeneration issue is very attractive. This has been explored by van Herk
et al. [140], and by proper choice of pH and reaction conditions about 50 mM
of DHAP was generated by 2 mM PHoN-Sf from 500 mM DHA and 240 mM PPi
at pH 4.5 in about 100 min. After this time PPi becomes exhausted and the
DHAP concentration slowly drops because of the hydrolytic phosphatase activity.
Considering the small Km value for DHAP of most aldolases this paves the way
to an acid phosphatase aldolase cascade system in one pot, which will be
discussed in Section 25.6. A disadvantage of the method is the large Km for
DHA of 3.6 M.
Another option to obtain DHAP is to oxidize L-glycerol-3-phosphate enzymatically
to DHAP [145]. Pradines et al. [93] had already explored in 1991 the preparative
phosphorylation of glycerol by using alkaline phosphatase and PPi but this yielded a
racemic mixture. Alternatively, phytase and PPi can carry out the phosphorylation of
glycerol but concentrations of up to 95% glycerol have to be used and phytases are
only active at low pH values (pH 4) and inactive at pH 7. Thus, after formation of
glycerol phosphate the pH must be increased and the glycerol concentration
decreased to maximize glycerol phosphate oxidase activity [146].
Glycerol phosphate oxidase has a rather broad substrate specificity. The oxidase
converts suitable diol precursors, which has allowed the synthesis of phosphorothio-
ate, phosphoramidate, and methylene phosphonate analogs of DHAP [145, 147].
25.5 Phosphorylases j 1021
Rhamnulose-1-phosphate aldolase accepts these compounds as substrates [145] to
give, for example, sugar phosphonates. Glycerol kinase has also been used to produce
L-glycerol-3-phosphate exclusively on a preparative scale [38, 39, 148], but an ATP
regenerating system and a suitable non-expensive phosphate donor are also needed.
To overcome the problem of ATP regeneration and production of L-glycerol-3-phos-
phate by whole cells the yeast Saccharomyces cerevisiae has been engineered to increase
the biosynthesis of L-glycerol-3-phosphate as well as to hamper further metabolization.
The glycerol-3-phosphate dehydrogenase was overexpressed and the gene coding for
cytosolic glycerol-3-phosphatase was deleted [149]. An alternative possibility is to
expose rac-glycidol, a common bulk chemical, to phosphate [150]. After ring opening
racemic glycerol phosphate is formed. The authors used this method to synthesize
monosaccharides by a sequential step procedure. This procedure consisted of pH
adjustment after the opening of glycidol and the addition of catalase and glycerol
phosphate oxidase, followed by addition of an aldolase, and finally dephosphorylation
of the aldol product by an acid phosphatase.
Instead of oxidation of glycerol phosphate to DHAP by an oxidase another
possibility is to use glycerol phosphate dehydrogenase in the presence of NAD þ .
Though the equilibrium of this reaction is in the direction of glycerol phosphate,
regeneration of NAD þ would drive the equilibrium towards DHAP.
DHAP may also be synthesized using phosphatidylcholine as starting material and
substituting the choline polar head by DHA using phospholipase D. The formed ester
is subsequently hydrolyzed by phospholipase C, affording DHAP and a 1,2-diacyl-
glycerol [151]. Similarly, using phosphatidylcholine new phospholipids may be
synthesized by transphosphatidylation catalyzed by phospholipase D in the presence
of appropriate alcohols and phosphatidylcholine [152, 153].
25.5
Phosphorylases
Phosphorylases belong to the class of hexosyltransferases (EC 2.4.1) and catalyze

glycosyl transfer to and from phosphate. Polysaccharide phosphorylases are able to
catalyze the removal of glucose-1-P from starch (starch phosphorylase in plants) or
from glycogen in many other organisms (glycogen phosphorylase). X-Ray structures
are available and details of the catalytic reaction are known [154–156]. This enzyme
cleaves the bond between two glucose residues at the non-reducing end of a glycogen
chain in which inorganic phosphate in the active site of the phosphorylase donates a
proton to the oxygen atom of the terminal glycosidic bond. The nucleophilic oxygen
atom of the phosphate then attacks at the C1 position and glucose-1-phosphate
is formed:
Polysaccharide þ Pi ! Polysaccharide þ glucose-1-phosphate ð25:1Þ

ðn residuesÞ ðn-1 residuesÞ
The formed glucose-1-phosphate can easily be converted into glucose-6-

phosphate by phosphoglucomutase (EC 5.4.2.2). This is an easy way of obtaining
glucose-6-phosphate that may be used in the enzymatic production of hydrogen
(Section 25.6). The authors are unaware of other applications in biotransformations.
In contrast, sugar-1-phosphates are widely used in the synthesis of oligosacchar-
ides. For example, as first shown by Haynie and Whitesides [157], trehalose
phosphorylase (EC 2.4.1.64) can be used to synthesize trehalose from glucose and
glucose-1-phosphate as an activated donor. Similarly, the enzyme from Pleurotus
osteatus accepts several structural analogues of D-glucose as glycosyl acceptor for the
enzymatic reaction with a-D-glucose-1-phosphate [158]. A coupled enzyme system
consisting of a sucrose phosphorylase, glucose isomerase, and the trehalose phos-
phorylase was used to synthesize a,a-trehalose from sucrose. In addition, novel
disaccharides were synthesized in a two-step procedure [159]. First maltose phos-
phate was converted into b-D-glucose-1-phosphate and D-glucose by a maltose
phosphorylase (EC 2.4.1.8). The D-glucose was removed with glucose consuming
yeast cells and in the second step the b-D-glucose-1-phosphate was condensed with an
appropriate carbohydrate by the synthetic activity of maltose phosphorylase or
trehalose phosphorylase. Several novel disaccharides were synthesized in this
way [159]. Sucrose phosphorylase (EC 2.4.1.7) normally splits the disaccharide
sucrose into D-fructose and a-D-glucose-1-phosphate. The enzyme also shows
transglycosylation activity, for example, it can accept benzoic acid using sucrose as
the donor molecule [160] to furnish glycosylated benzoic acids.
The kojibiose phosphorylase (EC 2.4.1.230) was used to synthesize various
oligosaccharides by glycosyl transfer from b-D-glucose-1-phosphate to isokestose
and nystose [161]. It is also possible to produce trisaccharides using the enzyme from
Thermoanaerobacter brockii, b-D-glucose-1-phosphate as a glucosyl donor, and disac-
charide as an acceptor [162]. Further details concerning the formation of glycosidic
linkages by glycosyl transferases following the more complex Leloir pathway in which
activated nucleoside diphosphate sugars are involved are found in the review by
Koeller and Wong [163]. Concerning the use of nucleoside phosphorylases in, for
example, the enzymatic production of Ribavirin and its derivatives, which are
important virus inhibitors, the reader is referred to Hanrahan and Hutchinson [164]
and Borthwick et al. [125].
25.6
Enzyme-Cascade Reactions in One Pot Using Phosphorylated Intermediates
Most cascade reactions in one pot using phosphorylated intermediates are focused on
the generation of DHAP in situ and coupling the intermediate in one pot to an
aldehyde in an aldolase-catalyzed condensation reaction.
The research groups of Fessner, Wong, and Whitesides have pioneered the use of
several enzymes in one pot to arrive at the formation of natural and non-natural
phosphorylated carbohydrates. These procedures start from glycerol [39], glycerol
phosphate [145], DHA [165], or sucrose via DHAP [141] and recycling of ATP is
required using PEP or acetyl phosphate and appropriate kinases. A sequential one-
pot aldol reaction has also been reported using fructose-1,6-diphosphate aldolase
25.6 Enzyme-Cascade Reactions in One Pot Using Phosphorylated Intermediates j 1023
(RAMA, EC 4.1.2.13) and the 2-deoxyribose-5-phosphate aldolase (DERA,
EC 4.1.2.4) and acetaldehyde, various aldehydes, and DHAP [166, 167] to afford
several ketoses. At the end an acid phosphatase was added to dephosphorylate these
carbohydrates. This integration of catalytic steps into one-pot cascade reactions is
the ultimate in green chemistry when sufficient space–time yields and chemical
yields are achieved, which is sometimes not the case [24]. A practical inexpensive
one-pot synthesis of L-fructose from DHAP and racemic glyceraldehyde using
rhamnulose-1-phosphate aldolase is also possible [168]. During the incubation acid
phosphatase was present to hydrolyze the phosphorylated carbohydrate without
prior isolation. This enzymatic method was further extended by enzymatic oxida-
tion of glycerol to L-glyceraldehyde by a galactose oxidase and by subsequent
coupling of the L-glyceraldehyde formed to DHAP present using the rhamnu-
lose-1-phosphate aldolase. However, the yield was low due to a reaction of DHAP
with galactose oxidase. By inactivating the oxidase and subsequent condensation of
the glyceraldehyde to DHAP and dephosphorylation further optimization of the
yield was obtained. The multienzyme system used by Sanchez-Moreno et al. [143] is
more versatile. The combination of dihydroxyacetone kinase from Citrobacter
freundii, fuculose-1-phosphate aldolase and acetate kinase allows the formation of
several phosphorylated carbohydrates. As Scheme 25.5 shows, the formation of
DHAP is central and is coupled with the aldol condensation catalyzed by a DHAP-
dependent aldolase. The system is completed with the in situ generation of ATP that
is present in only catalytic amounts.
Scheme 25.5 Multienzyme system in which dihydroxyacetone (DHA) is phosphorylated by a

kinase and ATP, with in situ ATP regeneration by acetyl phosphate and acetate kinase (AK). Modified
according to Reference [143].
Recently, this group reported [169] the expression of a bifunctional aldolase-kinase

that showed a 20-fold increase in the reaction rate over the parent enzyme. This has
been explained by channeling of the DHAP produced by the kinase to the active site of
the aldolase.
As already discussed glycerol may also be phosphorylated using PPi by phytase at
low pH and DL-glycerol phosphate is produced [146]. After oxidation to DHAP at
higher pH the aldolase couples the DHAP generated to butanal. After formation of
the phosphorylated product the pH was decreased to 4 and the phytase hydrolyses the
phosphorylated product to 5-deoxy-5-ethyl-D-xylulose. Drawbacks of this method are
the high concentration of glycerol needed, the pH shifts, and a dilution step to
decrease the concentration of glycerol.
Van Herk et al. [140] used cheap PPi as phosphate donor and the bacterial acid
phosphatase from Shigella flexneri to generate DHAP and a one-pot two-enzyme
system was developed in which the aldolase coupled the DHAP to an aldehyde to
produce the phosphorylated sugar and once PPi was consumed the product was
dephosphorylated by the phosphatase. Thus, both a phosphorylation and a dephos-
phorylation reaction are catalyzed by the phosphatase (Scheme 25.6). Using pro-
pionaldehyde and subsequent additions of PPi it was possible to convert more than
95% of the aldehyde into the aldol adduct (5,6-dideoxy-D-threo-2-hexulose), which
was obtained in an isolated yield of 53%.
Scheme 25.6 One-pot cascade reaction involving acid phosphatase and aldolase with
substrates DHA, PPi, and an aldehyde, leading to enantiopure carbohydrates according to
Reference [140].
This one-pot method is versatile since also 3-hydroxybutyraldehyde and (RS)-3-

azido-2-hydroxypropanal could be used as donors in the aldol reaction. Since RAMA
has a very broad substrate specificity [116] this method can lead to a wide range of
non-natural carbohydrates. It was also demonstrated that upon dephosphorylation of
the phosphorylated carbohydrate the formed enzyme-phosphate intermediate reacts
again with another DHA molecule to regenerate DHAP that is re-used in the aldol
coupling, resulting in higher yield [140]. Thus phosphate cycling occurs and the
amount of PPi needed is lowered. Importantly, the reaction is driven to completion
because once the phosphate group from the phosphorylated carbohydrate is hydro-
lyzed by the phosphatase present the product does not participate anymore in the
retro-aldol reacton. This method, however, has the disadvantage that rather high
concentrations of DHA should be present since the Km of the acid phosphatase from
Shigella for the DHA is 3.6 M, making it less convenient for large-scale applications.
To address this problem the gene coding for the acid phosphatase from Salmonella
enterica was exposed to random mutagenesis and the resulting mutant enzymes were
screened for a higher affinity for DHA. It was shown [170] that the enzyme from
Salmonella enterica has a higher DHA phosphorylating activity at a more alkaline
pH (7–8) than the Shigella enzyme used previously. Furthermore, the directed
evolution approach resulted in a mutant that was far more effective in the cascade
reaction. As shown in Figure 25.3 it produced more of the aldol product in much
less time. Interestingly, the mutation (V78L) occurred in a residue that is situated
at the entrance of the active site and is exposed to the solvent (Figure 25.2). This
residue also has a role in steering the enantioselective hydrolysis of racemic
O-phosphoserine [123].
25.6 Enzyme-Cascade Reactions in One Pot Using Phosphorylated Intermediates j 1025
Figure 25.3 Formation of the aldol adduct (250 mM), propionaldehyde (100 mM),
catalyzed by the wild-type Phon-Se [*] PhoN (1 mM), and RAMA (3 units) in 0.5 ml at
and PhoN-Sf [&] and Phon-Se mutant V78L [], pH 6. The concentrations of the
modified according to Reference [170]. The dephosphorylate end-product were based on
incubations contained DHA (500 mM), PPi HPLC measurements.
The scope of the cascade reactions in which (unstable) phosphorylated inter-

mediates are formed by the acid phosphatase and pyrophosphate is probably broad.
It was demonstrated [170] that this enzyme system also phosphorylates DL-glycer-
aldehyde to glyceraldehyde-3-phosphate. D-Glyceraldehyde-3-phosphate is the donor
for DERA and in a one-pot pilot experiment it was shown that 2-deoxy-D-ribose was
formed in this cascade.
An impressive one-pot six-step enzymatic synthesis of riboflavin from glucose
using eight enzymes has been reported [171]. Glucose is converted into ribulose-5-
phosphate in three enzymatic steps requiring ATP and NADP þ . These cofactors
were regenerated in situ by pyruvate kinase using phosphoenolpyruvate as high-
energy phosphate donor and glutamate dehydrogenase using 2-ketoglutarate as
oxidant. The ribulose-5-phosphate (1) (Scheme 25.7) was subsequently converted by a
3,4-dihydroxy-2-butanone-4-phosphate synthase into 3,4-dihydroxy-2-butanone-4-
phosphate (2) in a thermodynamically irreversible step.
A combined action of lumazine synthase and riboflavin synthase furnishes
riboflavin (5). The only products that are generated in this cascade are riboflavin,
glutamate, pyruvate, formate, and inorganic phosphate.
An interesting cascade of a series of bioreactors employing immobilized enzymes
has been described for the conversion of 3-phospho-D-glycerate into D-ribulose-1,5
biphosphate [172]. In total eleven enzymes, including a transketolase, were immo-
bilized. Cofactor regeneration was carried out by separation and using separate
devices. The authors also proposed to use this in a continuous biocatalytic carbon
dioxide fixation process. However, given the complexity and the need for separate
recycling, it is unlikely that this process will have synthetic applications. A more
Scheme 25.7 Enzymatic synthesis of riboflavin 4-phosphate (2); (2) condenses with 5-amino-6-
according to Reference [171]. Glucose is ribitylamino-2,4-pyrimidinedione (4) by 7-
converted into ribulose 5-phosphate (1) in three dimethyl-8-ribityllumazine synthase (B) to yield
enzymatic steps (not shown) requiring ATP and 6,7-dimethyl-8-ribityllumazine (3), and
NADP þ that have to be recycled. Next, 3,4- riboflavin synthase (C) dismutates two
dihydroxy-2-butanone 4-phosphate synthase molecules of (3) into riboflavin (5) and (4).
(A) converts (1) into 3,4-dihydroxy-2-butanone
viable process was developed by Zhang et al. [173] for producing hydrogen from
starch and water according to:
C6 H10 O5 þ 7H2 O ! 12H2 þ 6CO2 ð25:2Þ

It consists of 13 reversible enzymatic reactions: (i) a chain-shortening phosphor-
ylation reaction catalyzed by starch phosphorylase yielding glucose-1-phosphate
(Eq. (25.3)), (ii) the conversion of glucose-1-phosphate (G-1-P) into glucose-6-phos-
phate (G-6-P) catalyzed by phosphoglucomutase (Eq. (25.4)), (iii) a pentose phosphate
pathway containing ten enzymes (Eq. (25.5)), and (iv) hydrogen generation from
NADPH catalyzed by hydrogenase (Eq. (25.6)):
ðC6 H10 O5 Þn þ H2 O þ Pi ! ðC6 H10 O5 Þn1 þ G-1-P ð25:3Þ

G-1-P ! G-6-P ð25:4Þ
G-6-P þ 12NADP þ þ 6H2 O ! 12NADPH þ 12H þ þ 6CO2 þ Pi ð25:5Þ
12NADPH þ 12H þ ! 12H2 þ 12NADP þ ð25:6Þ
Although this method converts food into fuel it may be an efficient way to obtain
hydrogen from carbohydrates provided enzyme costs can be kept low enough.
25.7 Outlook j 1027
25.7
Outlook
Cheap and convenient enzymatic phosphorylation methods are available and are
useful transformations in organic synthesis. The future will tell whether ATP
dependent kinases will be a better choice in terms of scalability and overall costs
than processes using cheap phosphate donors and phosphatases working in the
synthetic mode. A drawback of the latter is the large excess of starting phosphate
donor that still has to be used and which may interfere with the subsequent workup of
the final product. However, modern biology techniques may be used to engineer a
phosphatase enzyme that, like kinases, is able to transfer the phosphate group from
PPi in a 1: 1 reaction to a substrate. An enzyme that only functions in the synthetic
mode would thus overcome undesired hydrolysis. This may be achieved by increas-
ing the enzyme affinity towards the substrates in order to suppress the competing
hydrolytic reaction by water. In addition, poly(P)-dependent kinases may replace the
classical ATP dependent enzymes in the future. Alternatively, one may look for
enzymes that do not require a phosphorylated substrate.
Indeed, in some DHAP requiring aldolases the phosphate ester of DHA may be
replaced by arsenate, vanadate, or even borate [174–176]. However, their application
is limited because of toxicity and difficult purifications. An aldolase that does not
require DHAP but is able to use dihydroxyacetone derivatives in the one pot-synthesis
of iminocyclitols is the D-fructose-6-phosphate aldolase [177–180]. A drawback of this
enzyme is that up to now only one of the four possible stereoisomers can be
produced. Given the potency of enzyme engineering techniques it is no surprise
that efforts are being made to design other DHAP-dependent aldolases to produce
variants of the aldolase that also accept DHA. As a step in this direction an in vivo
selection system was reported that allowed screening of libraries of mutated
L-rhamnulose-1-phosphate aldolase to develop a DHA dependent aldolase [180]. In
addition, the 2-keto-3-deoxy-D-phosphogluconate aldolase, which is highly specific
for D-glyceraldehyde-3-phosphate, has been subjected to directed evolution to create a
new enzyme that lacks the requirement for phosphate and that is able to synthesize
both D- and L-sugars [182].
Enzymatic dephosphorylation reactions are widely used in conversions since the
mild conditions avoid side reactions that occur during chemical hydrolysis. Fur-
thermore, dephosphorylation may occur regioselectively when more phosphate
groups are present, but to date only one report [87] has appeared. Enantioselective
dephosphorylation reactions may have potential applications that up to now have
hardly been explored. Phosphatases are indeed able to discriminate [121–123]
between chiral enantiomers in the highly enantioselective hydrolysis of O-phos-
pho-D,L-threonine, with a preference for the L-isomer. Thus in principle it should be
possible to hydrolyze racemic phosphate esters enantioselectively.
Numerous biobased cascade reactions already exist [183] and it is likely that with
the increased availability of activated phosphorylated substrates, the increasing
number of recombinant enzymes, and the progress in enzyme engineering methods
more of these efficient multienzyme cascades will be developed. In the future, these
versatile cascade procedures without intermediate recovery steps are likely to find
their way into more sustainable and safer syntheses of fine chemicals.
References
1 Westheimer, F.H. (1987) Science, 235, 18 Schoevaart, R., van Rantwijk, F., and
1173–1178. Sheldon, R.A. (2000) Biotechnol. Bioeng.,
2 Berridge, M.J. and Irvine, R.F. (1984) 70, 349–352.
Nature, 312, 315–321. 19 Calveras, J., Egido-Gabas, M., Gomez, L.,
3 Dzeja, P.P. and Terzic, A. (2003) Casas, J., Parella, T., Joglar, J., Bujons, J.,
J. Exp. Biol., 206, 2039–2047. and Clapes, P. (2009) Chem. Eur. J., 15,
4 Kroutil, W., Mang, H., Edegger, K., and 7310–7328.
Faber, K. (2004) Adv. Synth. Catal., 346, 20 Asano, N. (2009) Cell. Mol. Life Sci., 66,
125–142. 1479–1492.
5 Stryer, L. (1995) Biochemistry, 4th edn, W. 21 Sowa, T. and Ouchi, S. (1975) Bull. Chem.
H. Freeman, New York, pp. 441–462. Soc. Jpn., 48, 2084–2090.
6 Koeller, K.M. and Wong, C.-H. (2001) 22 Ireland, R.E., Muchmore, D.C., and
Nature, 409, 232–240. Hengartner, U. (1972) J. Am. Chem. Soc.,
7 Faber, K. (2004) Biotransformations in 94, 5098–5100.
Organic Chemistry, 5th revised edn, 23 Jones, S. and Smanmoo, C. (2004)
Springer, Berlin. Tetrahedron Lett., 45, 1585–1588.
8 Auriol, D., Nalin, R., Lefevre, F., 24 Sheldon, R.A. (2008) Chem. Commun.,
Ginolhac, A., De Guembecker, D., and 3352–3365.
Zago, C. (2008) EP 1995323 (A1). 25 Moss, G.P. (2011) Enzyme
9 Heimbach, T., Oh, D.-M., Li, L.Y., Nomenclature: Recommendations of the
Forsberg, M., Savolainen, J., nomenclature committee of the
Lepp€anen, J., Matsunaga, Y., Flynn, G., International Union of Biochemistry and
and Fleisher, D. (2003) Pharm. Res., 20, Molecular Biology on the nomenclature
848–856. and classification of enzymes by the
10 Scudder, P.R., Dwek, R.A., reactions they catalyse. Available at
Rademacher, T.W., and Jacob, G.T. (1991) http://www.chem.qmul.ac.uk/iubmb/
EP 0413674 (A2). enzyme/ (accessed on 15 July 2011).
11 Nicolaou, M.G., Yuan, C.-S., and 26 Mathews, C.K. and Van Holde K.E. (1996)
Borchardt, R.T. (1996) J. Org. Chem., 61, Biochemistry, 2nd edn, The Benjamin/
8636–8641. Cummings Publishing Company, Inc.,
12 Ariza, M.E. (2005) Drug Design Rev., 2, pp. 55–83.
373–387. 27 Zhang, J.B., Wu, B.Y., Zhang, Y.X.,
13 Gruijarro, D., Mancheno, B., and Yus, M. Kowal, P., and Wang, P.G. (2003) Org.
(1994) Tetrahedron, 50, 8551–8558. Lett., 5, 2583–2586.
14 Quast, H. and Dietz, T. (1995) Synthesis, 28 Kessel, D. (1968) J. Biol. Chem., 243,
1300–1304. 4739–4744.
15 Yanagisawa, A., Hibino, H., 29 Hiles, R.A. and Henderson, L.M. (1972) J.
Nomura, N., and Yamamoto, H. (1992) Biol. Chem., 247, 646–651.
J. Am. Chem. Soc., 115, 30 Pollak, A., Baughn, R.L., and Whitesides,
5879–5880. G.M. (1977) J. Am. Chem. Soc., 99,
16 Yanagisawa, A., Nomura, N., and 2366–2367.
Yamamoto, H. (1994) Tetrahedron, 50, 31 Wong, C.-H. and Whitesides, G.M.
6017–6028. (1981) J. Am. Chem. Soc., 103, 4890–4899.
17 Sch€umperli, M., Pellaux, R., and 32 Ishikawa, H., Takase, S., Tanaka, T., and
Panke, S. (2007) Appl. Microbiol. Hikita, H. (1989) Biotechnol. Bioeng., 34,
Biotechnol., 75, 33–45. 369–379.
References j 1029
33 Drueckhammer, D.G. and 50 Crans, D.C. and Whitesides, G.M. (1983)
Wong, C.-H. (1985) J. Org. Chem., 50, J. Org. Chem., 48, 3130–3132.
5912–5913. 51 Hirschbein, B.L., Mazenod, F.P., and
34 Chenault, H.K., Mandes, R.F., and Whitesides, G.M. (1982) J. Org. Chem.,
Hornberger, K.R. (1997) J. Org. Chem., 47, 3765–3766.
62, 331–336. 52 Kornberg, A., Rao, N.N., and
35 Bednarski, M.D., Crans, D.C., DiCosimo, Ault-Riche, D. (1999) Ann. Rev. Biochem.,
R., Simon, E.S., Stein, P.D., and 68, 89–125.
Whitesides, G.M. (1988) Tetrahedron 53 Mukai, T., Kawai, S., Matsukawa, H.,
Lett., 29, 427–430. Matuo, Y., and Murata, K. (2003) Appl.
36 Gross, A., Abril, O., Lewis, J.M., Environ. Microbiol., 69, 3849–3857.
Geresh, S., and Whitesides, G.M. (1983) 54 Kawai, S., Mori, S., Mukai, T., Suzuki, S.,
J. Am. Chem. Soc., 105, 7428–7435. Yamada, T., Hashimoto, W., and Murata,
37 Thorner, J.W. and Paulus, H. (1973) in K. (2000) Biochem. Biophys. Res.
The Enzymes, vol. 8 (ed. P.D. Boyer), Commun., 276, 57–63.
Academic Press, New York, 55 Tanaka, S., Lee, S.O., Hamaoka, K.,
pp. 487–508. Kato, J., Takiguchi, N., Nakamura, K.,
38 Rios-Mercadillo, V.M. and Othake, H., and Kuroda, A. (2003) J.
Whitesides, G.M. (1979) J. Am. Chem. Bacteriol., 185, 5654–5665.
Soc., 101, 5828–5829. 56 Mukai, T., Kawai, S., Mori, S., Mikami, B.,
39 Crans, D.C. and Whitesides, G.M. (1985) and Murata, K. (2004) J. Biol. Chem., 279,
J. Am. Chem. Soc., 107, 7019–7026. 50591–50600.
40 Crans, D.C., Kazlauskas, R.J., 57 Girbal, E., Binot, R.A., and Monsan, P.F.
Hirschbein, B.L., Wong, C.-H., Abril, O., (1989) Enzyme Microb. Technol., 11,
and Whitesides, G.M. (1987) Immobilized 518–527.
Enzymes and Cells, Methods Enzymology, 58 Resnick, S.M. and Zehnder, A.J.B.
part C, vol. 136 (ed. K. Mosbach), (2000) Appl. Environ. Microbiol., 66,
Academic Press Inc., New York, 2045–2051.
pp. 263–280. 59 Berke, W., Schuz, H.J., Wandrey, C.,
41 Crans, D.C. and Whitesides, G.M. (1985) Morrr, M., Denda, G., and Kula, M.R.
J. Am. Chem. Soc., 107, 7008–7018. (1988) Biotechnol. Bioeng., 32, 130–139.
42 Chenault, H.K., Chafin, L.F., and Liehr, S. 60 Str€ater, N., Lipscomb, W.N., Klabunde, T.,
(1998) J. Org. Chem., 63, 4039–4045. and Krebs, B. (1996) Angew. Chem. Int.
43 Kim, M.J. and Whitesides, G.M. (1988) Ed. Engl., 35, 2024–2055.
Appl. Biochem. Biotechnol., 16, 95–108. 61 Dowling, J.N., Saha, A.K., and Glew, R.H.
44 Moran, J.R. and Whitesides, G.M. (1984) (1992) Microbiol. Rev., 56, 32–60.
J. Org. Chem., 49, 704–706. 62 Neel, B.G. and Tonks, N.K. (1997) Curr.
45 Abril, O., Crans, D.C., and Opin. Cell Biol., 9, 193–204.
Whitesides, G.M. (1984) J. Org. Chem., 63 Zhang, C.-C., Jang, J., Sakr, S., and
49, 1360–1364. Wang, L. (2005) J. Mol. Microbiol.
46 Walt, D.R., Findeis, M.A., Rios- Biotechnol., 9, 154–166.
Mercadillo, V.M., Auge, J., and 64 Van Schaftingen, E. and Gerin, E. (2002)
Whitesides, G.M. (1984) J. Am. Chem. Biochem. J., 362, 512–532.
Soc., 106, 243–239. 65 Coleman, J.E. (1992) Annu. Rev. Biophys.
47 Hartog, A.F. and Berden, J.A. (1990) Biomol. Struct., 21, 441–483.
FEBS Lett., 261, 161–164. 66 Stec, B., Holtz, K.M., and
48 Chenault, H.K. and Whitesides, G.M. Kantrowitz, E.R. (2000) J. Mol. Biol.,
(1987) Appl. Biochem. Biotechnol., 14, 299, 1303–1311.
147–157. 67 Strater, N., Klabunde, T., Tucker, P.,
49 Langer, R.S., Hamilton, B.C., Witzel, H., and Krebs, B. (1995) Science,
Gardner, C.R., Archer, M.D., and 268, 1489–1492.
Colton, C.C. (1976) AIChE J., 22, 68 Rigden, D.J. (2008) Biochem. J., 409,
1079–1090. 333–348.
69 Lindquist, Y., Schneider, G., and 90 Edwards, B., Sparks, A., Voyta, J.C.,
Vihko, P. (1994) Eur. J. Biochem., 221, Strong, R., Murphy, O., and
139–142. Bronstein, I. (1990) J. Org. Chem., 55,
70 Kostrewa, D., Leitch, F.G., DArcy, A., 6225–6229.
Broger, C., Mitchell, D., and 91 Sambrook, J. and Russell, D.W. (2001)
van Loon, A.P.G.M. (1997) Nat. Struct. Molecular Cloning, Laboratory Manuals,
Biol, 4, 185–190. CSHL Press, New York.
71 Morton, R.K. (1958) Biochem. J., 70, 92 Pradines, A., Klaebe, A., Perie, J., Paul, F.,
150–155. and Monsan, P. (1988) Tetrahedron, 44,
72 Lei, X.G. and Stahl, C.H. (2001) Appl. 6373–6386.
Microbiol. Biotechnol., 57, 474–481. 93 Pradines, A., Klaebe, A., Perie, J., Paul, F.,
73 Harper, A.F., Kornegay, E.T., and and Monsan, P. (1991) Enzyme Microb.
Schell, T.C. (1997) J. Anim. Sci., 57, Technol., 13, 19–23.
3174–3186. 94 Rossolini, G.M., Schippa, S.,
74 Kinoshita, N., Ohkura, H., and Riccio, M.L., Berlutti, F., Macaskie, L.E.,
Yanagida., M. (1990) Cell., 63, 405–415. and Thaller, M.C. (1998) Cell. Mol. Life.
75 Virshup, D.M. (2000) Curr. Opin. Cell Sci., 54, 833–850.
Biol., 12, 180–185. 95 Thaller, M.C., Berlutti, F., Schippa, S.,
76 Wang, S., Tabernero, L., Zhang, M., Lombardi, G., and Rossolini, G.M. (1994)
Harms, E., Van Etten, R.L., and Microbiology, 140, 1341–1350.
Stauff, C.V. (2000) Biochemistry, 39, 96 Thaller, M.C., Lombardi, G., Berlutti, F.,
1903–1914. Schippa, S., and Rossolini, G.M. (1995)
77 Pannifer, A.D.B., Flint, A.J., Tonks, N.K., Microbiology, 141, 147–154.
and Barford, D. (1998) J. Biol. Chem., 273, 97 Thaller, M.C., Schippa, S., and
10454–10462. Rossolini, G.M. (1998) Protein Sci., 7,
78 Zhao, Y. and Zhang, Z.-Y. (1996) 1647–1652.
Biochemistry, 35, 11797–11804. 98 Hemrika, W., Renirie, R., Dekker, H.L.,
79 Kim, E.E. and Wyckoff, H.W. (1991) Barnett, P., and Wever, R. (1997)
J. Mol. Biol., 218, 449–464. Proc. Natl. Acad. Sci. USA, 94,
80 Meyerhof, O. and Green, H. (1949) 2145–2149.
Science, 11, 503–504. 99 Hemrika, W. and Wever, R. (1997) FEBS
81 Morton, R.K. (1958) Biochem. J., 70, Lett., 409, 317–319.
139–150. 100 Pan, C.-J. and Lei, K.-J. (1998) J. Biol.
82 Ostrowski, W. and Barnard, E.A. (1973) Chem., 273, 6144–6148.
Biochemistry, 12, 3893–3898. 101 Stukey, J. and Carman, G.M. (1997)
83 Reid, T.W. and Wilson, I.B. (1971) Protein Sci., 6, 469–472.
Escherichia coli alkaline phosphatases, in 102 Neuwald, A.F. (1997) Protein Sci., 6,
The Enzymes, vol. 4 (ed. P.D. Boyer), 1764–1767.
Academic Press, pp. 373–415. 103 Brindley, D.N. and Waggoner, D.W.
84 Gettins, P., Metzler, M., and Coleman, E. (1998) J. Biol. Chem., 273,
(1985) J. Biol Chem., 260, 2875–288. 24281–24284.
85 Koyama, T., Inoue, H., Ohnuma, S., and 104 Ishikawa, K., Mihara, Y., Gondoh, K.,
Ogura, K. (1990) Tetrahedron Lett., 31, Suzuki, E.-I., and Asano, Y. (2000) EMBO
4189–4190. J., 19, 2412–2423.
86 Min, J.-K., Yoo, H.-S., Lee, E.-Y., Lee, W.- 105 Makde, R.D., Dikshit, K., and
J., and Lee, Y.-M. (2002) Anal. Biochem., Kumar, V. (2006) Biomol. Eng., 23,
303, 167–175. 247–251.
87 Gutteridge, S., Reddy, G.S., and Lorimer, 106 Makde, R.D., Mahajan, S.K., and
G. (1989) Biochem. J., 260, 711–716. Kumar, V. (2007) Biochemistry, 46,
88 Billich, A., Stockhove, U., and Witzel, H. 2079–2090.
(1983) Nucleic Acids Res., 11, 7611–7624. 107 Littlechild, J., Garcia-Rodriguez, E.,
89 Breslow, R. and Katz, I. (1968) J. Am. Dalby, A., and Isupov, M. (2002) J. Mol.
Chem. Soc., 90, 7376–7377. Recognit., 15, 291–296.
References j 1031
108 Renirie, R., Hemrika, W., and Wever, R. 124 Herdewijn, P., Balzarini, J., De Clerq, E.,
(2000) J. Biol. Chem., 275, 11650–11657. and Vanderhaege, H. (1985) J. Med.
109 Macedo-Ribeiro, S., Renirie, R., Wever, Chem., 28, 1385–1386.
R., and Messerschmidt, A. (2008) 125 Borthwick, A.D., Butt, S., Biggadike, K.,
Biochemistry, 47, 929–934. Exall, A.M., Roberts, S.M., Youds, P.M.,
110 Uchiya, K.-I., Tohsuji, M., Nikai, T., Kirke, B.E., Booth, B.R., Cameron, J.M.,
Sugihara, H., and Sasakawa, C. (1996) Cox, S.W., Marr, C.L.P., and Shill, M.D.
J. Bacteriol., 178, 4548–4554. (1988) Chem. Commun., 656–658.
111 Uerkvitz, W. (1981) J. Biol. Chem., 256, 126 Hunsucker, S.A., Mitchell, B.S., and
382–389. Spychala, J. (2005) Pharmacol. Ther., 107,
112 Uerkvitz, W. (1988) J. Biol. Chem., 263, 1–30.
15823–15830. 127 Asano, Y., Mihara, Y., and Yamada, H.
113 Babu, M.M., Kamalakkannan, S., (1999) J. Mol. Catal. B: Enzym., 6,
Subrahmanyam, Y.V.B.K., and 271–277.
Sankaran, K. (2002) FEBS Lett., 128 Axelrod, B. (1948) J. Biol. Chem., 172,
512, 8–12. 1–13.
114 Durrwachter, J.R. and Wong, C.-H. (1988) 129 Appleyard, J. (1948) Biochem. J., 42,
J. Org. Chem., 53, 4175–4181. 596–597.
115 Straub, A., Effenberger, F., and Fisher, P. 130 Tanaka, N., Hasan, Z., Hartog, A.F.,
(1990) J. Org. Chem., 55, 3926–3932. van Herk, T., and Wever, R. (2003) Org.
116 Bednarski, M.D., Simon, E.S., Biomol. Chem., 1, 2833–2839.
Bishofberger, N., Fessner, W.-D., 131 Schr€odter, K., Betterman, G., Staffel, T.,
Kim, M.-J., Lees, W., Saito, T., and Hofmann, T. (2006) Phosphoric acid
Waldmann, H., and Whitesides, G.M. and phosphates, in Ulmanns
(1989) J. Am. Chem. Soc., 111, Encyclopedia of Industrial Chemistry,
627–635. vol. A19, Wiley-VCH Verlag GmbH,
117 Schultz, M., Waldmann, H., Kunz, H., Weinheim.
and Vogt, W. (1990) Liebigs Ann. Chem., 132 Suzuki, E.-I., Ishikawa, K., Mihara, Y.,
1019–1024. Shimba, N., and Asano, Y. (2007) Bull.
118 Silvestri, M.G., Desantis, G., Mitchel, M., Chem. Soc. Jpn., 80, 276–286.
and Wong, C.-H. (2003) Asymmetric 133 Mihara, Y., Utagawa, T., Yamada, H., and
aldol reactions using aldolases, in Topics Asono., Y. (2000) Appl. Environ. Microbiol.,
in Stereochemistry, vol. 23 (ed. S.E. 66, 2811–2816.
Denmark), pp. 267–342. 134 Mihara, Y., Utagawa, T., Yamada, H., and
119 Fujii, H., Koyama, T., and Ogura, K. Asano, Y. (2001) J. Biosci. Bioeng., 92,
(1982) Biochim. Biophys. Acta, 712, 50–54.
716–718. 135 Mihara, Y., Ishikawa, K., Suzuki, E.-I., and
120 Davies, H.G., Green, R.H., Kelly, D.R., Asano, Y. (2004) Biosci. Biotechnol.
and Roberts, S.M. (1989) Biochem., 68, 1046–1050.
Biotransformations in Preparative Organic 136 Mori, H., Iita, A., Jujio, F., and Teshiba, S.
Chemistry, Academic Press, London, (1997) Appl. Microbiol. Biotechnol., 48,
pp. 75–86. 693–698.
121 Scollar, M.P., Sigal, G., and Klibanov, 137 Yoshikawa, M., Kato, T., and Takenishi, T.
A.M. (1985) Biotechnol. Bioeng., 27, (1969) Bull. Chem. Soc. Jpn., 42,
247–252. 3505–3508.
122 Kimura, T., Vassilev, V.P., Shen, G.-J., and 138 van Herk, T., Hartog, A.F.,
Wong, C.-H. (1997) J. Am. Chem. Soc., van der Burg, A.M., and Wever, R.
119, 11734–11742. (2005) Adv. Synth. Catal., 347, 1155–1162.
123 van Herk, T., Hartog, A.F., 139 Dissing, K. and Uerkvitz, W. (2006) Enz.
Ruijssenaars, H.J., Kerkman, R., Microbiol. Technol., 38, 683–688.
Schoemaker, H.E., and Wever, R. 140 van Herk, T., Hartog, A.F., Schoemaker,
(2007) Adv. Synth. Catal., 349, H.E., and Wever, R. (2006) J. Org. Chem.,
1349–1352. 71, 6244–6247.
141 Fessner, W.-D. and Walter, C. (1992) 159 Aisaka, K., Masuda-Kato, T.,
Angew. Chem. Int. Ed. Engl., Chikamune, T., Kamitori, K., Uosaki, Y.,
31, 614–616. and Saito, Y. (2000) J. Biosc. Bioeng., 90,
142 Itoh, N., Tujibata, Y., and Liu., J.Q. (1999) 208–213.
J. Appl. Microbiol. Biotechnol., 51, 160 Sugimoto, K., Nomura, K., Nishiura, H.,
193–200. Ohdan, K., Hayashi, H., and
143 Sanchez-Moreno, I., Garcia-Garcia, J.F., Kuriki, T. (2007) J. Biosc. Bioeng.,
Bastida, A., and Garcia-Junceda, E. (2004) 104, 22–29.
Chem. Commun., 1634–1635. 161 Okada, H., Fukushi, E., Onodera, S.,
144 Wong, C.-H. and Whitesides, G.M. (1983) Nishimoto, T., Kawabata, J., Kikuchi, M.,
J. Org. Chem., 48, 3199–3205. and Shiomi, N. (2003) Carbohydr. Res.,
145 Fessner, W.-D. and Sinerius, G. (1994) 338, 879–885.
Angew. Chem. Int. Ed. Engl., 162 Maruta, K., Watanabe, H., Nishimoto, T.,
33, 209–212. Kubota, M., Chaen, H., Fukuda, S.,
146 Arth, H.L., Sinerius, G., and Fessner, Kurimoto, M., and Tsujisaka, Y. (2006)
W.-D. (1995) Liebigs. Ann., 11, 2037–2042. J. Biosc. Bioeng., 101, 385–390.
147 Schoevaart, R., van Rantwijk, F., and 163 Koeller, K.M. and Wong, C.-H. (2000)
Sheldon, R.A. (2000) J. Org. Chem., 65, Chem. Rev., 100, 4465–4493.
6940–6943. 164 Hanrahan, J.R. and Hutchinson, D.W.
148 Hettwer, J., Oldenburg, H., and (1992) J. Biotechnol., 23, 193–210.
Flaschel, E. (2002) J. Mol. Catal. B Enzym., 165 Eyrisch, O., Sinerius, G., and
19, 215–222. Fessner, W.-D. (1993) Carbohydr. Res.,
149 Nguyen, H.T.T., Dieterich, A., 238, 287–306.
Athenstaedt, K., Truong, N.H., Stahl, U., 166 Gijssen, H.J.M. and Wong, C.-H. (1995)
and Nevoigt, E. (2004) Metabol. Eng., 6, J. Am. Chem. Soc., 117, 2947–2948.
155–163. 167 Gijssen, H.J.M. and Wong, C.-H. (1995)
150 Charmantray, F., Dellis, P., Samreth, S., J. Am. Chem. Soc., 117, 7585–759.
and Hecquet, L. (2006) Tetrahedron Lett., 168 Franke, D., Machajewski, T., Hsu, C.C.,
47, 3261–3263. and Wong, C.-H. (2003) J. Org. Chem., 68,
151 DArrigo, P., Piergianni, V., 6828–6831.
Pedrocchi-Fantoni, G., and Servi, S. 169 Iturrate, L., Sanchez-Moreno, I.,
(1995) Chem. Commun., 2505–2506. Doyaguez, E.G., and Garcia- Junceda, E.
152 Ulbrich-Hofmann, R. (2000) (2009) Chem. Commun., 1721–1723.
Phospholipases used in lipid 170 van Herk, T., Hartog, A.F., Babich, L.,
transformations, in Enzymes in Lipid Schoemaker, H.E., and Wever, R. (2009)
Modification (ed. U.T. Bornscheuer), ChemBioChem., 10, 2230–2235.
Wiley-VCH Verlag GmbH, Weinheim, 171 R€omisch, W., Eisenreich, W., Richter, G.,
pp. 219–262. and Bacher, A. (2002) J. Org. Chem., 67,
153 Dippe, M., Mrestani-Klaus, C., 8890–8894.
Schierhorn, A., and 172 Bhattacharya, S., Schiavone, M., Gomes,
Ulbrich-Hofmann, R. (2008) Chem. Phys. J., and Bhattacharya, S.K. (2004)
Lipids, 152, 71–78. J. Biotechnol., 111, 203–217.
154 Barford, D. and Johnson, L.N. (1989) 173 Zhang, Y.-H.P., Evans, B.R.,
Nature, 340, 609–616. Mielenz, J.R., Hopkins, R.C., and
155 Johnson, L.N., Hu, S.H., and Barford, D. Adams, M.W.W. (2007) PLoS ONE, 2 (5),
(1992) Faraday Discuss., 93, 131–142. e456.
156 Johnson, L.N. and Barford, D. (1990) 174 Drueckhammer, D.G., Durrwachter, J.R.,
J. Biol. Chem., 265, 2409–2412. Pederson, R.L., Crans, D.C., Daniels, L.,
157 Haynie, S.L. and Whitesides, G.M. (1990) and Wong, C.-H. (1989) J. Org. Chem., 54,
Appl. Biochem. Biotechnol., 23, 155–170. 70–77.
158 Schwartz, A., Goedl, C., Minani, A., and 175 Schoevaart, R., van Rantwijk, F., and
Nidetzky, B. (2007) J. Biotechnol., 129, Sheldon, R.A. (2001) J. Org. Chem., 66,
140–150. 4559–4562.
References j 1033
176 Sugiyama, M., Hong, Z., Whalen, L.J., 180 Castillo, J.A., Calveras, J., Casas, J.,
Greenberg, W.A., and Wong, C.-H. (2006) Mitjans, M., Vinaedell, M.P., Parella, T.,
Adv. Synth. Catal., 348, 255–259. Inoue, T., Sprenger, G.A., Joglar, J., and
177 Sch€urmann, M. and Sprenger, G.A. Calper, P. (2006) Org. Lett., 8, 6067–6070.
(2001) J. Biol. Chem., 276, 181 Sugiyama, M., Hong, Z., Greenberg,
11055–11061. W.A., and Wong, C.-H. (2007) Bioorg.
178 Sch€urmann, M., Sch€ urmann, M., and Med. Chem., 15, 5905–5911.
Sprenger, G.A. (2002) J. Mol. Catal. B 182 Fong, S., Machajewski, T.D., Mak, C.C.,
Enzym., 19, 247–252. and Wong, C.-H. (2000) Chem. Biol., 7,
179 Sugiyama, M., Hong, Z., Liang, P.-H., 873–883.
Dean, C.M., Wahlen, L.J., 183 Bruggink, A., Schoevaart, R., and
Greenberg, W.A., and Wong, C.-H. (2007) Kieboom, T. (2003) Org. Process Res. Dev.,
J. Am. Chem. Soc., 129, 14811–14817. 7, 622–640.
j 1035
Part VI
Reductions
and Oliver May.
j1037
26
Reduction of Ketones and Aldehydes to Alcohols
Harald Gr€oger, Werner Hummel, Sonja Borchert, and Marina Kraußer
26.1
Introduction
The transformation of a C¼O double bond into the corresponding reduced CHOH
functionality represents a straightforward and atom-economical approach towards
the synthesis of alcohols. When starting from prochiral compounds bearing a C¼O
double bond, such a transformation can be carried out in an enantioselective fashion
when using a chiral catalyst. Owing to the importance of a broad range of resulting
chiral alcohol products in the field of chiral drug synthesis, enantioselective catalytic
reduction of ketones additionally gained tremendous industrial interest. Notably,
numerous efficient catalytic routes have already been developed to date for enantio-
selective ketone reductions. As outstanding technologies based on the use of
synthetic catalysts, the metal-catalyzed asymmetric hydrogenation of ketones [1]
and borane reduction [2] are widely applied on an industrial scale. Both technologies
represent landmarks also in industrial asymmetric catalysis in general.
These chemocatalytic technologies for the production of enantiomerically pure
alcohols are complemented by biocatalytic methodologies for the reduction of
ketones [3] and aldehydes, which turned out to be a highly efficient and competitive
alternative (Scheme 26.1). Notably, technologies for the biocatalytic reduction of
carbonyl compounds have already made the jump from an interesting academic
synthetic tool towards an industrially feasible technology platform [4]. Besides
robustness and industrial large-scale feasibility, high catalytic efficiency and excellent
enantioselectivities are further key features of enantioselective ketone reduction
processes with biocatalysts. Thus, it is no surprise that the suitability for industrial
purposes, in particular for large-scale manufacture of enantiomerically pure alcohols
as drug intermediates [3], has been already demonstrated by a broad range of
technical applications in the chemical and pharmaceutical industry. The enzymatic
reduction of aldehydes is industrially applied in the production of aroma chemicals.
Research in the field of enantioselective reduction of ketones using alcohol
dehydrogenases has been already comprehensively reviewed [3, 5]. This chapter
and Oliver May.
1038 j 26 Reduction of Ketones and Aldehydes to Alcohols
enantioselective
biocatalytic
O reduction OH OH
or
R1 R2 1
R R 2 R1 R2
1 (S)- or (R)-2
Scheme 26.1
gives an overview of alcohol dehydrogenases suitable for synthetic purposes, syn-

thetic concepts in the field of biocatalytic reduction in organic synthesis (in particular
with respect to the different cofactor-regeneration concepts), and selected, in
particular recent, synthetic applications for specific target molecules in this field,
including technically feasible processes that turned out to be highly efficient and
fulfill criteria such as high productivity, substrate concentrations, conversions, and
selectivities, particularly enantioselectivities.
26.2
Alcohol Dehydrogenases as Biocatalysts
Alcohol dehydrogenases (ADHs) constitute a large family of enzymes responsible for

the reversible reduction of aldehydes, ketones, and a-, b-, or v-keto esters to the
corresponding hydroxy compounds. This reduction is coupled with the stoichio-
metric consumption of NADH or NADPH, which needs to be regenerated
simultaneously.
ADHs have been identified in both prokaryotic and eukaryotic microorganisms
and plant and mammalian cells. Most of them may be classified into subgroups. They
represent a large database to give information about phylogenetic and evolutionary
relationships and to allow comparative biochemical studies on the function of
conserved sequences and structure–function relationships. So far only few of these
enzymes are important for preparative applications, particularly those that are
commercially available or easy to prepare (preferably in a recombinant form through
overexpression in a host organism).
An important feature for the classification of dehydrogenases is the occurrence of a
Rossmann-fold motif to bind the nicotinamide-based coenzymes. Initially, this
structure motif was described in lactate, alcohol, malate, and glyceraldehyde 3-
phosphate dehydrogenases [6–8]. Based on distinct sequence motifs, protein chain
length, mechanistic features, and structural comparisons, a system of aldo-keto
reductases, short-, medium-, and long-chain dehydrogenases/reductases has now
been established [9–11]. Table 26.1 gives typical key features of ADHs belonging to the
classes of aldo-keto reductases, short-, and medium chain dehydrogenases/reductases.
A typical member of the short-chain dehydrogenases/reductases (SDRs) is
Drosophila ADH. The Zn-containing yeast and liver ADHs were the first members
of the medium-chain dehydrogenases (MDRs), while prokaryotic polyol dehydro-
genases and eukaryotic glucose 6-phosphate dehydrogenases or UDP-glucose
Table 26.1 Structural properties and preparatively useful examples of enzymes of the most important superfamilies of alcohol dehydrogenases (AKR ¼ aldo-keto
reductase; MDR ¼ medium-chain dehydrogenase/reductase; SDR ¼ short-chain dehydrogenase/reductase).
Parameter AKR MDR SDR
Amino acid chain length of 325 350 250

monomer
Subunit structure Monomeric Di- or tetrameric Di- or tetrameric
Characteristic structural feature (a/ß)8-Barrel
Coenzyme binding motif No Rossmann fold Rossmann fold Rossmann fold
Reference enzyme Rat liver steroid DH Yeast ADH; (horse) liver ADH Drosophila ADH
Structural catalytic feature Catalytic tetrad:a) Tyr55, Lys84, Catalytic tetrad:b) Asn111, Ser138,
Asp50, His117 Tyr151, Lys155
Examples of preparatively useful Glycerol-DH from Gluconobacter (S)-ADH from Thermoanaero- (R)-ADH from Lactobacillus kefir/
enzymes available as recombinant oxydans bacter sp.(S)-ADH from Rhodo- Lactobacillus brevis; (R)-ADH from
proteins coccus erythropolis Candida magnoliae; (S)-ADH from
Saccharomyces cerevisiae (Gre2p)
a) Numbered according to reference enzyme rat liver 3a-hydroxysteroid dehydrogenase.

b) Numbered according to reference enzyme 3b/17b-hydroxysteroid dehydrogenase from Comamonas testosteroni.
26.2 Alcohol Dehydrogenases as Biocatalysts
j1039
dehydrogenases are now classified into the heterogeneous group of long-chain
dehydrogenases/reductases (LDRs) [10, 12]. Besides these enzymes possessing a
Rossmann-fold, aldo-keto reductases (AKRs) form a further superfamily with the
common feature of lacking this structural motif.
26.2.1
Overview of the Types of Alcohol Dehydrogenases
26.2.1.1 Aldo-keto Reductases (AKRs)

Aldo-keto reductases (AKRs) are a superfamily of enzymes [13] present in a wide
range of organisms, ranging from bacteria to mammals. Most AKRs are monomeric
proteins with a characteristic feature of an (a/b)8-barrel structure. The chain length
lies in the range of approximately 325 amino acids, corresponding to a molecular
mass of about 37 kDa. A further characteristic feature of AKRs is the lack of a
Rossmann-fold motif for binding of the coenzyme NAD(P)H [14, 15]. Based on
amino acid sequence comparisons, the enzymes have been divided into fourteen
families (AKR1 to AKR14) [16, 17]. Members of an AKR family should have <40%
amino acid identity with any other family. AKRs catalyze a 4-pro-R-hydride transfer of
NAD(P)H to the ketone. For substrate binding and conversion, AKRs use a catalytic
tetrad at the active site consisting of Tyr55, Lys84, Asp50, and His117 (numbered
according to rat liver 3a-hydroxysteroid dehydrogenase).
Despite their overall structural similarity, these enzymes show considerable
diversity in terms of substrate specificity, acting on numerous aliphatic and aromatic
aldehydes and ketones, including sugars, steroids, and xenobiotics. Members of
AKRs are, for example, the well-known aldehyde reductase (EC 1.1.1.2) from human
liver [13, 18], microbial enzymes such as xylose reductase from yeasts like Pichia
stipitis [19], Kluyveromyces lactis [20], or Candida tropicalis [21], arabinose dehydro-
genase (Ara1p) [22], 2-methylbutyraldehyde reductase (Ypr1p) [23], or aryl-alcohol
dehydrogenases (Aad14p; Aad3p; Aad4p; Aad10p) from Saccharomyces cerevisiae [23],
or 2,5-diketo-D-gluconic acid reductase from Escherichia coli [24]. Furthermore, two
classes of steroid transforming enzymes are included in this super-group: the
hydroxysteroid dehydrogenases (HSDs), acting on carbonyl groups, and D4-3-keto-
steroid-5b-reductases (5b-reductases), which reduce C ¼ C double bonds of steroid
molecules.
26.2.1.2 Medium-Chain Dehydrogenases/Reductases (MDR)

The MDR superfamily [9, 25–27] with a subunit size of about 350 amino acid residues
consists of over 10 000 members in the UniProt database [28]. The first well-
characterized member was the class I type of mammalian ADH. Its primary structure
was solved already in 1970 [29]. The primary structure of another MDR member,
sorbitol dehydrogenase, was published in 1984 [30]. Based on sequence similarities,
the MDR superfamily can be divided into separate families. The largest family
constitutes of zinc-dependent ADHs, with approximately 1000 members, followed
by, for example, cinnamyl ADH (CAD), leukotriene B4 dehydrogenases (LTDs),
tetrameric ADHs (TADHs), or polyol dehydrogenases (PDHs). Liver ADHs as
26.2 Alcohol Dehydrogenases as Biocatalysts j1041
members of the zinc-dependent ADH family are dimeric proteins, whereas the
common yeast ADHs have a tetrameric structure, now divided from the zinc-
dependent ADH as a separate subfamily, TADH. The yeast ADH enzymes are
active with primary, unbranched, aliphatic alcohols [31], but allyl alcohol and
cinnamyl alcohol are good substrates, too [32]. The PDH family contains sorbitol
DH, which catalyzes the interconversion of L-iditol and L-sorbose and that of
D-glucitol and D-fructose. The enzyme also acts on related sugar alcohol substrates.
The 3D structure of sorbitol DH reveals a tetrameric structure [33], but binding of
only one zinc ion per subunit, since it does not possess the structural zinc binding
loop of the ADH family. Sorbitol DHs are present in bacteria, yeast, and animals,
suggesting an important function in a wide range of life forms. Cinnamyl ADHs
(CAD) are important in lignin biosynthesis in plant cell walls [34]. Each monomer
binds two zinc ions and uses NADPH as cofactor. Members of the CAD family are
not limited to plants. The ADHs from Saccharomyces cerevisiae (ADH6 and ADH7)
have high activity towards cinnamaldehyde [35, 36], although yeast does not syn-
thesize lignin.
Several MDRs have been found in thermophilic organisms, especially among the
TADH family. The 3D structure of ADH from the archaeon Sulfolobus solfataricus is
available, revealing interesting structural properties [37]. This ADH has both the
catalytic and the structural zinc ions. One of the structural zinc ligands is glutamic
acid instead of cysteine. Removal of the structural zinc in Sulfolobus ADH reduces the
structural stability [38, 39].
Structurally, the MDR superfamily shows various quaternary structures, ranging
from monomers, via dimers and trimers, to tetramers. Furthermore, the number of
zinc ions is also variable, between 0, 1, and 2 ions per subunit in different families.
Typically, the MDR proteins consist of two domains, the C-terminal domain contains
the coenzyme-binding Rossmann-fold motif, an often six-stranded parallel b-sheet
sandwiched between a-helices on each side. The N-terminal domain is the substrate
binding region with a core of antiparallel b-strands and a-helices positioned at the
surface. Both domains are separated by a cleft containing a deep pocket.
26.2.1.3 Short-Chain Dehydrogenases/Reductases (SDR)

The SDR superfamily [26, 40–42] constitutes one of the largest enzyme super-
families, with presently over 46 000 members and over 300 crystal structures.
Common to all members of SDR is the occurrence of a Rossmann-fold motif for
binding of NAD(P)H. A typical SDR is Drosophila ADH. Two main types of SDR
enzymes can be distinguished, the classical with a chain length of about 250 amino
acid residues and the extended ones with an additional 100-residue domain in the
C-terminal region proteins [43, 44]. Numerous studies show that the central
acid–base catalyst in SDRs is the acidic OH group at the aromatic Tyr moiety that
donates or abstracts a proton to or from the substrate [45–48]. The ability of the Tyr
residue to act as a catalytic acid/base is enhanced by an adjacent Lys residue that
together with an oxidized, positively charged nicotinamide-based cofactor lowers the
pKa of the OH functionality at the Tyr moiety [45, 49]. The lysine e-amino group is also
involved in nicotinamide ribose binding, whereas the role of the active site Ser
residue is to stabilize and polarize the carbonyl substrate group [47, 50]. The
conserved sequence regions include a variable N-terminal TGxxxGxG motif (around
position 12, all numbering refers to 3b/17b-hydroxysteroid dehydrogenase from
Comamonas testosteroni) as part of the nucleotide binding region, and the active site
with a triad of catalytically important Ser138, Tyr151, and Lys155 residues with Tyr as
the most conserved residue within the whole family [47, 51–53]. For most SDR
enzymes, a fourth amino acid, Asn11, can be regarded as essential part for catalysis,
extending the previously recognized catalytic triad of Ser-Tyr-Lys residues to form a
tetrad of Asn-Ser-Tyr-Lys in most characterized SDRs [54, 55]. Interestingly, a pair-
wise comparison of SDR proteins illustrates that the 3D folding pattern of SDRs is
more conserved than their underlying sequence motifs.
26.2.2
Sources of Alcohol Dehydrogenases Useful for Biocatalysis
Even though a large number of ADH catalyzed asymmetric reactions are described in
the literature, it may still be necessary to develop novel catalysts applicable for
technical applications. Only a small amount of enzyme is commercially available, for
example, yeast or horse liver ADH or the ADH from Thermoanaerobacter brockii.
Major suppliers of ADHs are, for example, the companies Sigma-Aldrich (Buchs,
Switzerland) and evocatal GmbH (D€ usseldorf, Germany), which offers a screening
kit with 12 ADHs. Furthermore, for a many other ADHs found to be suitable as a
biocatalyst sequence information and protocols for the efficient overexpression of
recombinant strains are published. Some selected ADHs that have already proved
their worth in organic synthetic applications as a recombinant enzyme are summa-
rized below.
26.2.2.1 (S)-Specific NADH-Dependent ADH from Horse Liver

This ADH represents a commercially available enzyme, which can be used for the
reduction of a broad variety of cyclic ketones and a- or b-keto esters on a laboratory
scale. Scheme 26.2 shows selected examples. Other functional groups like alkynes,
alkenes, (thio)ethers, nitriles, and esters are tolerated, and even the bioconversion of
organosilicon [56, 57] and sensitive organometallic [58] compounds has been
described. In the oxidative mode, the enzyme has found broadest application
for the regio- and enantioselective oxidation of primary 1,4-diols to chiral c-lac-
tones [59–62].
26.2.2.2 (S)-Specific NADPH-Dependent ADH from Thermoanaerobacter sp

Open-chain methyl and ethyl ketones are the preferred substrates for T. brockii ADH
(TBADH) [63–66]. It reveals a high tolerance towards increased concentrations of
organic solvents such as 20 vol.% of isopropanol. Moreover, TBADH presents a
remarkable capacity to withstand high temperature, up to 85 C. It shows broad
substrate specificity that has attracted much technological interest. The alcohol (R)-5-
hydroxyhept-6-enonoate, which is a useful intermediate for the synthesis of various
arachidonic acid metabolites, was prepared by reduction of ethyl 5-oxo-6-heptenoate
(S)-ADH from
O horse liver OH
R1 R2 R1 R2
1 2
Selected examples
OH OH
Cl O
O
Ar F
32% conversion 96% ee

>99% ee
H CH3
O OH
CH2OH
H Cr
OC CO
CO
89% conversion
>98% ee (1R)= >99% ee
Scheme 26.2 Examples of HLADH-catalyzed reductions.
with NADPH and isopropanol catalyzed by Thermoanaerobacter sp. ADH expressed

in E. coli. Scheme 26.3 gives an overview of the substrate spectrum.
26.2.2.3 (R)-Specific NADPH-Dependent ADH from Lactobacillus kefir and L. brevis

These ADHs, which represent two quite homologues enzymes, turned out to be
valuable biocatalysts for the preparation of (R)-hydroxy compounds. From Lactoba-
cillus kefir DSM 20 587, the adh gene encoding a protein of 252 amino acids (Lk-ADH)
could be isolated. The deduced amino acid sequence indicated that this enzyme
belongs to the family of short-chain dehydrogenases. The enzyme was cloned and
(S)-ADH from
Thermoanaerobacter brockii
O OH
2
R1 R R1 R2
1 2
Selected examples
OH OH OH
Cl
H 3C H3C CH3
99% ee >99% ee
Scheme 26.3
(R)-ADH from
O Lactobacillus sp. OH
R1 R2 R1 R2
1 2
Selected examples
OH OH
OH O
CH3 Cl
OEt
OH
>99% conversion >99% conversion >99% conversion
>99.9% ee >99.9% ee >99.9% ee
OH O O OH O O
Cl
OtBu OtBu
72% conversion 77% conversion
>99.5% ee >99.4% ee
OH OH
Cl OH
Me
OMe
Me3Si O
94% conversion 70% conversion 15% conversion
>99% ee >99% ee 97% ee
Scheme 26.4
expressed in E. coli, purified, and characterized biochemically [67]. For the reduction
of acetophenone the specific activity of the homogeneous recombinant ADH was
558 U mg1. The enzyme shows its maximum activity at 50 C. The 3D structure was

solved by X-ray analysis with a resolution of at least 1.8 A [68]. Lk-ADH was applied
meanwhile for the reduction of a broad range of aliphatic and aromatic ketones as
well as b-keto esters [69–76]. All prochiral ketones were stereoselectively reduced to
the corresponding alcohols with >99% e.e. and in the case of diketones >99% d.e.
Scheme 26.4 shows selected examples.
26.2.2.4 (S)-Specific NADH-Dependent ADH from Rhodococcus erythropolis

An (S)-specific ADH from Rhodococcus erythropolis (Re-ADH) reduces a broad range
of carbonyl compounds with high enantiomeric excess. The 1047 bp gene coding for
348 amino acids was cloned in E. coli cells and successfully expressed [77]. The
recombinant enzyme exhibits high thermostability, which facilitated its purification
by heat treatment, followed by two column-chromatography steps. Re-ADH shows
high similarity to several zinc-containing medium-chain ADHs. All zinc ligands
seem to be conserved except one of the catalytic zinc ligands, where Cys is probably
replaced by Asp. Preparative examples concerning the reduction of hydrophobic
(S)-ADH from
O Rhodococcus erythropolis OH
R1 R2 R1 R2
1 2
Selected examples
OH
O OH OH
CH3
H3 C O CH3 H3C CH3
Cl
>99% conversion >95% conversion >95% conversion
>99% ee >99% ee >99% ee
Scheme 26.5
ketones at high concentrations [78–81] demonstrate the broad applicability of this

biocatalyst. Scheme 26.5 gives some selected examples.
26.2.2.5 (S)-Specific NADH-Dependent ADH from Rhodococcus ruber

An (S)-specific ADH from Rhodococcus ruber (Rr-ADH), which has been successfully
overexpressed in E. coli, also represents a versatile, highly enantioselective biocatalyst
for the reduction of a broad range of hydrophobic ketones [82]. Scheme 26.6 shows
some selected examples. The substrate range of this recombinant enzyme Rr-ADH
(used as lyophilized E. coli cells) turned out to be the same as the one when using the
(S)-ADH from
O Rhodococcus ruber OH
R1 R2 R1 R2
1 2
Selected examples
OH
OH OH Cl
CH3
n-C5H11
76% conversion 89% conversion >99% conversion
>99% ee >99% ee >99% ee
OH
N3 OH O OH
Cl
OiPr n-C4H9 CH3
>99% conversion >99% conversion 66% conversion

>99% ee >99% ee >99% ee
Scheme 26.6
wild-type strain Rhodococcus ruber DSM44541 (which has already been used suc-
cessfully for many biotransformations) [83]. The reaction time, however, could be
significantly reduced when using lyophilized E. coli cells with the recombinant ADH
compared to the wild-type cells, indicating a high overexpression of the ADH.
Notably, isopropanol is accepted as a cosubstrate and high concentrations of
isopropanol of up to 80 vol.% are tolerated, thus making substrate-coupled cofac-
tor-regeneration an interesting (and widely applied) option for this enzyme.
26.2.2.6 (S)-Specific NADPH-Dependent ADH Gre2p from Saccharomyces cerevisiae

Gre2p serves as a versatile biocatalyst that exhibits high enantioselectivity for various
ketones, including a-chloro ketones, a-acetoxy ketones, a- and b-keto esters, and
diketones (Scheme 26.7) [83–86]. Most of the chiral products were obtained with a
high enantiomeric purities of >98% e.e. Furthermore, Gre2p is responsible for the in
vivo reduction of 2,5-hexanedione to (2S,5S)-hexanediol [87, 88]. Besides the reduc-
tion of hexanedione, Gre2p catalyzes reduction of a broad range of a-, b- and
c-diketones. Bioreduction using the recombinant enzyme afforded the (2S,5S)-
hexanediol with >99% conversion and in >99.9% d.e. and e.e. [88].
(S)-ADH from
O Saccharomyces cerevisiae OH
R1 R2 R1 R2
1 2
Selected examples
OH O OH
Cl OH
CH3
OMe OMe H3C
OH
O
76% conversion 68% conversion >99% conversion
>99% ee >98% ee >99.9% ee
Scheme 26.7
26.2.2.7 (R)-Specific NADH-Dependent ADH from Nocardia globerula

A highly active ADH was isolated from Nocardia globerula that shows a unique substrate
spectrum towards different prochiral aliphatic ketones and bulky keto esters as well as
thioesters (Scheme 26.8) [89]. For example the enzyme reduces ethyl 4-chloro-3-oxo
butanoate to form (S)-4-chloro-3-hydroxy butanoate with >99% e.e. Interestingly, 3-
oxobutanoic acid tert-butylthioester is reduced with 110 U mg1, while the correspond-
ing tert-butyloxyester is not reduced at all. The specific activity with ethyl-2-oxo-4-
phenylbutanoate as the substrate was 220 U mg1. The corresponding 915 bp-long
gene was determined, cloned, expressed in E. coli, and applied in biotransformations.
The N. globerula ADH is a tetramer about 135 kDa in size as determined from
gel filtration. Its sequence is related to several hypothetical 3-hydroxyacyl-CoA
ADH from
Nocardia globerula
O OH
R1 R2 R 1 R2
1 2
Selected examples
OH
OH O
OEt
Cl
OEt
O
>99% ee
>99% conversion
>96% ee
Scheme 26.8
dehydrogenases whose sequences were derived by whole-genome sequencing from

bacterial sources as well as known mammalian 3-hydroxyacyl-CoA dehydrogenases
and b-hydroxyacyl-CoA dehydrogenases from different Clostridia.
26.2.2.8 (R)-Specific NADPH-Dependent ADH from Candida magnolia

An NADPH-dependent carbonyl reductase catalyzing the reduction of ethyl 4-chloro-
3-oxobutanoate (COBE) to enantiomerically pure ethyl (S)-4-chloro-3-hydroxybu-
tanoate (CHBE) was isolated from Candida magnoliae by the group of Shimizu
[90–93]. The gene consists of 849 bp, which corresponds to a subunit size of 30 420
Da; it was overexpressed in E. coli under the control of the lac promoter. The deduced
amino acid sequence revealed that this ADH belongs to the SDR superfamily.
Acetophenone derivatives and a- and b-keto esters were reduced with activities in
the range of 0.5–6 U mg1, with best activity against COBE (Scheme 26.9) [94].
(R)-ADH from
O Candida magnoliae OH
R1 R2 R1 R2
1 2
Selected example
OH O
Cl
OEt
85% conversion
>99% ee
Scheme 26.9
26.2.2.9 (S)-Specific NADH-Dependent ADH from Sporobolomyces salmonicolor

From the yeast Sporobolomyces salmonicolor an NADPH-dependent carbonyl reduc-
tase (SSCR) was isolated that reduces diaryl ketones [95–100]. Such substrates with a
para-substituent were converted with high enantioselectivity up to 99% e.e. Mutation
of SSCR at amino acid position Q245 led to a higher yield of the (S)-enantiomer [101].
For reduction of para-substituted diaryl ketones, the enantioselectivity was increased
in the order F (44% e.e.) < Cl (88%) < CH3 (92%) < CH3O (96%). For most diaryl
ketones, methanol as cosolvent increased conversion significantly. Scheme 26.10
gives an example for the reduction of a diaryl ketone and of a b-keto ester.
(S)-ADH from
O Sporobolomyces salmonicolor OH
R1 R2 R 1 R2
1 2
selected examples
OH
OH O
Cl
OEt
MeO
92.7% ee
26% conversion
>96% ee
Scheme 26.10
26.2.2.10 NADPH-Dependent Glycerol Dehydrogenase (Gox1615) from

Gluconobacter oxydans
The acetic acid bacterium Gluconobacter oxydans has a high potential for finding
oxidoreductases with various different catalytic abilities. The NADP-dependent
glycerol dehydrogenase (Gox1615; EC 1.1.1.72) is a member of family 11 of the
aldo-keto reductase (AKR) enzyme superfamily. It has been cloned, overexpressed in
E. coli, purified, and characterized [102]. Gox1615 has an apparent native molecular
mass of 39 kDa, which corresponds to the mass of 37.213 kDa calculated from the
primary structure. From HPLC measurements, a monomeric structure can be
deduced. GlyDH is a biotechnologically attractive enzyme because of its broad
substrate spectrum focused on different aliphatic, branched, and aromatic aldehydes
combined with a distinctive regio- and stereoselectivity (Scheme 26.11). Additionally,
the enzyme has been shown to oxidize various different alcohols. The highest
activities were observed for the conversion of D-glyceraldehyde in the reductive
direction and for L-arabitol in the oxidative direction. Since high enantioselectivities
were observed for the reduction of glyceraldehyde, a kinetic resolution of glyceral-
dehyde has been investigated, yielding enantiopure L-glyceraldehyde on a preparative
scale. A whole-cell catalyst, using E. coli BL21(DE3) as a host, co-expressing glycerol
dehydrogenase (GlyDH) from Gluconobacter oxydans, and glucose dehydrogenase
(GDH) from Bacillus subtilis for cofactor-regeneration, has been successfully con-
structed and used for the production of L-glyceraldehyde in high enantioselectivity at
54% conversion [103]. This whole-cell catalyst shows several advantages over the
26.3 Concepts of Biocatalytic Ketone and Aldehyde Reduction j1049
ADH from
O Gluconobacter oxydans OH
1 1
R H R H
1 2
Selected example
OH OH OH
GlyDH
HO O HO O + HO OH
glyceraldehyde L-glyceraldehyde glycerol
NADPH NADP
Scheme 26.11
cell-free system, such as a higher thermal and operational stability and the ability to
recycle the catalyst without any loss of activity.
26.2.3
Screening Methods to Obtain Novel ADHs
Besides the above-mentioned ADHs that are commercially available or easily

prepared, there is a high demand for further enzymes reducing carbonyl compounds
efficiently and showing new substrate specificities. New or improved catalysts can be
obtained by different methods such as, for example, screening for enzymes among
naturally available sources, in silico-based screening methods, or mutation of known
structures. Several new enzymes that have proved to be useful as technical catalysts
were discovered by screening natural habitats. Essentially, natural sources for new
enzymes means microorganisms that can be obtained by special enrichment or
selection methods. However, also cells or gene libraries from plant or animal tissues
can be used to screen for novel activities.
A very promising powerful method makes use of metagenome samples, a
collection or library of DNA extracted directly from environmental samples. This
screening method enables us to find enzymes from organisms that must not be
propagated in laboratory. In fact, novel NAD(P)H-dependent ADHs converting short-
chain polyols were found by this method [104, 105].
26.3
Concepts of Biocatalytic Ketone and Aldehyde Reduction
26.3.1
Overview of Process Concepts
The enzymatic reduction of C¼O double bonds in ketones and aldehydes is based on
the use of an ADH as a catalyst component and a so-called cofactor (coenzyme) as a
reducing agent. The most preferred cofactors (also for technical applications) are
either NADH or NADPH. Since cofactors are expensive reducing agents, being too
costly to be applied in stoichiometric amount, a common key feature of all preparative
(and technical) biocatalytic reductions is the use of cofactors in catalytic amounts and
their recycling in situ by coupling the ketone reduction process with a second process,
in which the cofactor is regenerated in situ. In the following, different concepts for in
situ cofactor-regeneration are exemplified for the enzymatic reduction of prochiral
ketones 1 as a carbonyl component, leading to the formation of (chiral) alcohols (R)-
or (S)-2.Scheme 26.12 shows the general process concept.
OH OH
reducing NAD(P)+ 1 2 or
agent R R R1 R2
(S)- or (R)-2
alcohol
oxidized O
reducing agent NAD(P)H
R1 R2
1
Scheme 26.12
For in situ cofactor-regeneration two major approaches have been developed,

namely, the so-called substrate-coupled and enzyme-coupled cofactor-regeneration.
In the substrate-coupled cofactor-recycling one ADH catalyzes both the reduction of
the ketone substrate and the dehydrogenation of a cheap alcohol with formation of
the corresponding ketone as a side-product. Typically, isopropanol serves as such a
desired easily available and cheap reducing agent, and is oxidized in this process to
give acetone. The substrate-coupled cofactor-regeneration concept is illustrated in
Scheme 26.13 (and is described in detail in Section 26.3.2).
OH OH OH
NAD(P)+ or
H 3C CH3 R1 R2 R1 R2
(S)- or (R)-2
alcohol alcohol
O
O
H3C CH3
NAD(P)H
R1 R2
1
Scheme 26.13
An advantage of this method is the requirement for only one enzyme, which
catalyzes both required reactions [namely, the desired reduction of the (prochiral)
ketone and the oxidation of isopropanol]. However, two limitations have to be
considered: First, each ADH that is suitable for the reduction of the target ketone
also needs to accept isopropanol as a substrate. Although so far numerous enzymes
have been identified that fulfill this criteria, in specific cases ADHs might be found in
an enzyme screening that are preferred enzymes for the desired ketone reduction but
not for the oxidation of isopropanol. A second drawback is the lack of an irreversible
step in the reaction concept shown in Scheme 26.13. Both reactions are reversible,
thus leading to equilibrium between the corresponding ketone and alcohol mixtures
of the substrate and product side. To overcome this limitation, strategies to shift the
equilibrium have been successfully developed. These strategies are based on removal
of the formed acetone as the most volatile component in the reaction mixture during
the reaction and the use of excess isopropanol. The use of excess isopropanol also
ensures an increase in solubility of hydrophobic ketones, which are typically water
immiscible.
In the alternative approach, the so-called enzyme-coupled cofactor-regeneration, the
in situ regeneration of the cofactor is carried out by means of a second enzyme. This
second enzyme utilizes a (preferably) cheap compound (e.g., formate, D-glucose) as
reducing agent for the reduction of the oxidized form of the cofactor, namely, NADþ or
NADPþ , thus regenerating the required reduced form of the cofactor, NAD(P)H. This
concept, exemplified for the use of formate and D-glucose, respectively, is shown in
Scheme 26.14. In addition, the different enzyme-coupled cofactor-regeneration
methodologies are described in more detail in Sections 26.3.3–26.3.6.
When using formate as a reducing agent a formate dehydrogenase is the required
enzyme component, which oxidizes formate to carbon dioxide (Scheme 26.14a).
Since this step is irreversible under reaction conditions applied in general, the
equilibrium of the desired ketone reduction is shifted towards the product side.
This typically enables the formation of the desired alcohols with excellent
conversions. Another enzyme-coupled cofactor-recycling method is based on the
use of D-glucose in the presence of a glucose dehydrogenase (GDH). The GDH
catalyzes the oxidative transformation of D-glucose into D-gluconolactone, thus
regenerating the oxidized cofactor form, NADþ or NADPþ , into its reduced form,
NADH or NADPH (Scheme 26.14b). The irreversibility of the process is achieved by
hydrolytic ring-opening of D-gluconolactone to produce D-gluconic acid, which
is subsequently neutralized by addition of a base to form the corresponding
D-gluconate salt.
In general, advantages of these concepts of enzyme-coupled cofactor-recycling are
the irreversibility of the cofactor-regeneration process and the use of a very cheap
basic chemical such as formic acid or D-glucose as reducing agent in stoichiometric
amount, which makes this method economically highly attractive. Furthermore,
downstream-processing is simplified by the fact that both side-products, carbon
dioxide and (neutralized) D-gluconic acid (obtained from oxidation of formate and
D-glucose, respectively, as described above), can be easily separated from the resulting
alcohol product. Potential general drawbacks of this concept of enzyme-coupled
(a) Concept based on the use of a formate dehydrogenase
OH OH
formate NAD(P)+ 1 2 or 1
R R R R2
(S)- or (R)-2
formate alcohol
O
CO2
NAD(P)H 1
R R2
1
(b) Concept based on the use of a glucose dehydrogenase
OH OH
D-glucose NAD(P)+ or
R1 R 2 R1 R2
(S)- or (R)-2
glucose alcohol
D-glucono- O
lactone NAD(P)H
R1 R2
1
Scheme 26.14
cofactor-recycling are the need for a second enzyme, which ideally should be
compatible with the ADH in terms of pH and temperature optimum. Furthermore,
substrate or product inhibition effects of the enzyme chosen for cofactor-recycling
might occur, as well as destabilization effects caused by the reaction medium (in
particular when using organic solvents as cosolvents). However, a range of reduction
processes based on enzyme-coupled cofactor-recycling, with either a formate dehy-
drogenase or GDH, that address these criteria and that turned out to be highly
efficient, in particular when using directly recombinant whole-cell catalysts, have
been successfully developed. These processes are discussed in more detail in
Sections 26.3.3 and 26.3.4.
It should be added that enzyme-coupled cofactor-regeneration methodologies are
not restricted to the use of a formate dehydrogenase or a GDH. For example, a further
method reported for enzyme-coupled cofactor-recycling is based on the use of a
glucose-6-phosphate dehydrogenase. A more detailed description of this method is
given in Section 26.3.5. Owing to the high price of D-glucose-6-phosphate the direct
use of this compound is not suitable for commercial processes. A valuable synthetic
option for the application of a glucose-6-phosphate dehydrogenase, however, is based
on the use of a (recombinant) whole-cell catalyst bearing this enzyme besides an
ADH in overexpressed form. Under fermentation-like conditions and the use of
these recombinant strains as non-permeabilized living cells, D-glucose can serve as
substrate. After the transport of glucose into the cells and its (metabolic) transfor-
mation into D-glucose-6-phosphate, this intermediate then serves as substrate for
cofactor-regeneration. A limitation of this process, however, is the need for non-
permeabilized living cells in general. This might be disadvantageous with respect to a
(desirable) high substrate loading, since permeabilization can be expected under
those conditions.
Further enzyme-coupled cofactor-regenerations, for example, based on the use of a
phosphite dehydrogenase able to oxidize phosphite to phosphate, have been devel-
oped as well, but are (so far) rarely applied in organic synthesis. A more detailed
description of the method based on the use of a phosphite dehydrogenase is given in
Section 26.3.6.
In general the ADHs can be used as isolated enzymes (in purified free form or as
a crude extract or in immobilized form) or incorporated in whole-cells. With respect
to the latter approach, options are the use of wild-type cells or recombinant whole-cell
organisms. Recently, tailor-made recombinant whole-cell catalysts, bearing the
desired ADH and (in the case of the enzyme-coupled cofactor-regeneration) an
additional enzyme in overexpressed form, have elicited tremendous interest for
technical applications due to their beneficial properties in the resulting biotransfor-
mations. Owing to overexpression of the desired enzyme(s) the impact of undesired
side-reactions by other (competing) dehydrogenases is suppressed (since a lower
amount of biomass is required for the biotransformation). This enables an econom-
ically attractive access, in particular when using high-cell density fermentation for
biocatalyst production.
A further option for cofactor-regeneration is the use of wild-type whole cells based
on fermentation-like conditions. Once again glucose serves as a cheap substrate,
which is consumed in the cell, thus contributing to regenerate the cofactors by the
cell-internal metabolism. This concept is visualized in Scheme 26.15, and described
in more detail in Section 26.3.7. As mentioned above for the glucose-6-phosphate
wild-type whole-cells
containing
alcohol dehydrogenase(s)
enzymes for cofactor-regeneration,
O NAD(P)H OH OH
or
R1 R2 R1 R 2 R1 R2
D-glucose,
1 aqueous medium (S)- or (R)-2
Scheme 26.15
dehydrogenase-based approach, this type of cofactor-recycling requires the use of
living cells. Typically, such processes have some limitations: They do not run at high
substrate input, selectivity is reduced by interfering ADHs with contrary stereo-,
regio-, or chemoselectivity, and additionally they require a large amount of biomass
due to the lack of overexpression of the desired ADH and cofactor-regenerating
enzyme. A further potential drawback is the tedious work-up in many cases.
This route, however, can be attractive, in particular when easily accessible micro-
organisms are available and it turns out to be suitable for the desired biotransfor-
mation. Well-known examples in this field are, for example, processes based on the
use of bakers yeast.
Furthermore, chemocatalytic as well as electrochemical cofactor-regeneration
methodologies have been developed. These methodologies are described in more
detail in Section 26.3.8.
26.3.2
Ketone Reduction Based on Substrate-Coupled Cofactor-Regeneration with
Isopropanol
26.3.2.1 Use of Isolated Enzymes

For substrate-coupled cofactor-recycling in enantioselective enzymatic processes a
broad range of ADHs already turned out to be suitable. This section discusses
examples that are based on the use of isolated enzymes. In a pioneering work in the
field of substrate-coupled cofactor-regeneration for enantioselective ketone reduction
Wong and coworkers reported such a process with an isolated NADH-dependent
ADH from a Pseudomonas sp. strain [106]. Notably, a broad substrate spectrum has
been observed when carrying out the reactions in a two-phase solvent system with n-
hexane as an organic phase. A selected example is shown in Scheme 26.16.
Furthermore, the recombinant ADH from Leifsonia sp., which was developed by
O O ADH from OH O
Candida parapsilosis
H 3C OEt H3C OEt
3 (S)-4
75% without removal of acetone
>95 / >97% with removal of acetone
(via different removal stategies)
NADH + H+ NAD+ >99.5% ee
O OH
H3C CH3 H3C CH3

ADH from
Candida parapsilosis
strategies for in situ-
removal of acetone
to increase conversion
Scheme 26.16
Itoh and coworkers, was applied as a purified enzyme in reductions and turned out to
also have a broad substrate spectrum [107, 108]. Using this enzyme, trifluoroace-
tophenone was transformed successfully into the corresponding (S)-alcohol with
quantitative conversion and excellent >99% e.e. [108]. In addition, ADHs from the
thermophilic glycolytic anaerobes Thermoanaerobium and Thermoanaerobacterium
can catalyze enantioselectively the reduction of a broad range of ketones using
isopropanol as a cosubstrate for cofactor regeneration [109, 110]. A mutant of an ADH
from Thermoanaerobacter ethanolicus turned out to be useful for the enantioselective
reduction of prochiral ketones bearing an aromatic moiety [111]. These reactions
were conducted at an isopropanol concentration of 30 vol.%. Further ADHs that have
been widely used in asymmetric ketone reductions under isopropanol-based cofac-
tor-regeneration are from Lactobacillus kefir and L. brevis, which were found and
successfully produced in recombinant E. coli strains by the Hummel group [69–76].
Among many synthetic applications (which are described in more detail below), for
example, the ADH from L. brevis also turned out to be suitable for the asymmetric
reduction of perfluorinated ketones [112]. For example, 2,2,2-trifluoroacetophenone
was transformed into the corresponding chiral alcohol with 87% conversion and an
excellent enantiomeric excess of >99% e.e. The recombinant ADH from L. brevis has
also been very successfully applied for the enantioselective reduction of alkynones
and a-halogenated derivatives thereof, leading to the desired propargylic ketones in
enantiomerically pure form (>99% e.e.) [71, 73].
In addition to the search for suitable enzymes, alternative reaction media have
been investigated. For example, ionic liquids have been identified successfully as
suitable reaction media for the biocatalytic reduction of ketones under cosubstrate
cofactor-recycling by Kragl and Liese and coworkers [113].
To overcome the limitation of incomplete conversions due to the lack of an
irreversible step in the substrate-coupled cofactor-regeneration, intensive process
development has been carried out. The Liese group reported interesting process
concepts applied for the reduction of ethyl 5-oxohexanoate (3) with substrate-coupled
in situ cofactor-recycling [114], demonstrating that by pervaporation or stripping off
the acetone the conversion can be increased significantly. This is due to shifting the
equilibrium in the favored direction by removing the by-product acetone from the
reaction mixture. As a biocatalyst the ADH from Candida parapsilosis was used,
catalyzing the desired reaction with a high enantioselectivity of >99.5% e.e.
(Scheme 26.16).
Without removal of acetone during the reaction, a lower conversion of 75% was
found when using a substrate concentration of 30 mM. By means of pervaporation
for the separation of acetone from the reaction mixture, Liese and coworkers
succeeded in increasing the conversion to >95%. This pervaporation concept is
based on the use of an external membrane module in addition to the reactor for the
biotransformation. This membrane module separates the product and acetone from
isopropanol and water. Isopropanol and water run back into the reactor and can be
reused. Thus acetone is constantly removed from the reaction reactor and the
equilibrium is shifted towards the desired alcohol (S)-4. The process concept is
graphically shown in Figure 26.1. As an alternative, the removal of acetone during the
substrate
reactor
membrane
module
pump pump
product
condenser
Figure 26.1 Pervaporation based on the use of an external membrane module in addition to the
reactor for the biotransformation.
reaction by insertion of water vapor-saturated air was studied, and a high conversion
of >97% was obtained by this methodology.
An efficient enzymatic reduction on a technical scale, which is based on substrate-
coupled cofactor-recycling, has been applied at Wacker Fine Chemicals for a multi-
ton production of different optically active alcohols [115, 116]. An efficient enzyme
used in this process technology is the ADH from L. brevis, which has been reported to
be routinely used for asymmetric ketone reduction by Julich Chiral Solutions and
Wacker Fine Chemicals. An example of a technical-scale application is the production
of (R)-methyl 3-hydroxybutanoate [(R)-6] through an asymmetric reduction of
methyl 3-oxobutanoate (5) with an ADH from L. brevis at Wacker Fine Chemicals
(Scheme 26.17) [115, 116].
ADH from
O O Lactobacillus brevis OH O
28 - 32°C, pH 6.5
H 3C OCH3 H3 C OCH3
5 (R)-6
94% yield
>99.8% ee
NADPH + H+ NADP+
O OH
H3C CH3 H3C CH3

ADH from
Lactobacillus brevis
28 - 32°C, pH 6.5
Scheme 26.17
continuous extraction with MTBE
product
reactant removal
(R)-6
of
acetone
reactant product product

+ + + enzyme
enzyme enzyme enzyme
product:
OH O
H3C OCH3
(R)-6
Figure 26.2 Process scheme for the production of (R)-6.
An impressive substrate concentration of 3.7 mol l1, corresponding to 430 g l1

has been achieved [115, 116]. A key feature of this process is removal of acetone
during the reaction through reduced pressure, thus shifting the reaction equilibrium
in the direction of the desired (R)-alcohol product (R)-6. In addition, the aqueous
phase can be re-used in a simple stirred-tank reactor. After continuous extraction and
purification via distillation the desired product (R)-6 was obtained in 94% yield and
with an excellent enantiomeric excess of >99.8% e.e. Figure 26.2 shows the process
scheme for the production of the desired (R)-alcohol (R)-6 [115, 116].
Almac Sciences researchers have reported the diastereoselective reduction of an
N-protected (S)-3-amino-1-chloro-4-phenyl-butan-2-one in combination with a sub-
strate-coupled cofactor-regeneration using isopropanol [117]. The resulting alcohol,
which is an intermediate in the synthesis of the pharmaceutically important
atazanavir, was formed with 98.5% conversion and was isolated in 93% yield with
a diastereoselectivity of 100% d.e.
Extension of this biocatalytic reduction concept towards an enantio- and diaster-
eoselective reduction process has been reported recently by the Gotor group [118].
When starting from 1,2-diketones the ADH-catalyzed reduction with substrate-
coupled cofactor-regeneration delivers the corresponding diols. A limitation of this
method, however, is that, typically, mixtures containing the corresponding hydroxy
ketones (which are formed at first and serve as an intermediate) are obtained and
in several cases these hydroxy ketones are obtained as the major reaction
products. Scheme 26.18 shows an example of a reaction leading to the diol
(2S,3R)-10 in high yield.
The use of (racemic) alcohols other than isopropanol proved to be an alternative
way to achieve high conversion. When replacing isopropanol by a racemic alcohol a
O ADH from O OH
H3 C Thermoanaerobacter sp. H3C H3C
CH3 CH3 + CH3
O i-PrOH, OH O
cofactor
7 (R)-8 (S)-9
3%, 99% ee 3%, 99% ee
OH
H3C
+ CH3
OH
(2S,3R)-10
91%, 99% de, >99% ee
Scheme 26.18
one-pot tandem-reaction was realized by the Gotor group [119], leading to simul-
taneous formation of two enantiomerically pure sec-alcohols through enantioselec-
tive reduction of the ketone substrate and enantioselective oxidation of one enan-
tiomer of the used racemic alcohol substrate in the cofactor-regeneration step.
26.3.2.2 Use of Whole Cells

Besides isolated enzymes, whole-cell catalysts have also been successfully used in
enantioselective enzymatic ketone reductions under substrate-coupled cofactor-
recycling. This was demonstrated, for example, in early work by Matsumura and
coworkers using wild-type cells of Candida boidinii [120], and later by Itoh et al. [121]
using recombinant E. coli whole-cell catalyst overexpressing an ADH from a
Corynebacterium strain. Furthermore, the development of an efficient recombinant
whole-cell E. coli biocatalyst overexpressing an ADH from Candida parapsilosis and its
use in asymmetric ketone reduction was achieved by Daicel researchers [122]. The
recombinant biocatalyst is very suitable for the asymmetric reduction of ethyl
4-chloro-3-oxobutanoate (11) using isopropanol as reducing agent and without
addition of external cofactor. At 36.6 g l1 substrate input, the desired ethyl (R)-
4-chloro-3-hydroxybutanoate [(R)-12] was obtained in 95.2% yield and with excellent
enantioselectivity of 99% e.e. (Scheme 26.19).
E. coli
whole-cell catalyst,
containing
(S)-ADH from
O O Candida parapsilosis, OH O
Cl NAD(P)H Cl
OEt OEt
11 + i-PrOH (R)-12
(36.6 g/l - acetone 95.2% yield
substrate input) 99% ee
Scheme 26.19
An analogous synthesis of the corresponding (S)-enantiomer has been developed
by the Weuster-Botz group [123] using Lactobacillus kefir wild-type cells. When using
5 vol.% of isopropanol as cosubstrate a final product concentration of 1.2 M was
obtained in combination in 97% yield and with 99.5% e.e. The Schmid and Buehler
group reported a recombinant ADH from Thermus sp., which is suitable for
enzymatic reduction of ketones under substrate-coupled cofactor-recycling, leading
to several chiral alcohols with >99% e.e. [124].
The M€uller group used recombinant E. coli cells with an overexpressed ADH from
Lactobacillus brevis in combination with isopropanol as cosubstrate for an impressive
regio- and enantioselective reduction of 3,5-dioxocarboxylates [69]. For example, the
3,5-diketo ester 13 was transformed into the resulting alcohol (S)-14 with 72%
conversion and >99.5% e.e. (Scheme 26.20).
recombinant ADH from

O O O Lactobacillus brevis OH O O
Cl Cl
Ot-Bu Ot-Bu
13 (S)-14
72% yield
NADPH NADP+ >99.5% ee
O OH
H3 C CH3 recombinant ADH from H3C CH3

Lactobacillus brevis
Scheme 26.20
A very efficient whole-cell biocatalyst in combination with a substrate-coupled

cofactor-regeneration has also been reported by the Faber and Kroutil group using a
Rhodococcus ruber strain. This microorganism turned out to be suitable for the
enantioselective transformation of a broad range of ketones into the corresponding
(S)-alcoholsathighisopropanolconcentrationsofupto50vol.%[125,126].Theutilityof
high isopropanol concentrations is also attractive for commercial applications.
Scheme 26.21 gives selected synthetic applications. A broad range of aliphatic and
aromatic ketones are reduced with high enantioselectivities of >99% e.e. in most cases.
Lyophilized cells of Rhodococcus ruber have also been suitable for the enantioselective
reduction of a,b-unsaturated ketones, yielding the corresponding allylic alcohols with
up to >99% e.e. [127]. In addition, recombinant expression of the R. ruber ADH was
reported, thus expanding further the scope of this versatile enzyme [82, 128]. Immo-
bilization of this enzyme with different amino-functionalized carrier materials after
their activation with glutaraldehyde was reported by Liese and coworkers [129]. Besides
mono- and biphasic aqueous–organic solvent media, substrate-coupled regeneration
with isopropanol in the presence of ADH from R. ruber has been successfully applied in
micro-aqueous organic systems with 99 vol.% of an organic solvent. Notably, by means
ofthismethodologyhighsubstrateconcentrationsofuptoabout2 Mwererealized[130].
Rhodococcus ruber whole-cell catalyst
containing
(S)-ADH,
O NAD(P)H OH
R1 R2 R1 R2
15 + i-PrOH (S)-16
- acetone
Selected examples
OH OH CH3 OH
CH3 CH3 H3C CH3
(S)-16a (S)-16b (S)-16c

81% yield 92% yield 70% yield
>99% ee >99% ee >99% ee
Scheme 26.21
The Kroutil group also reported an elegant enantioselective route towards aliphatic
epoxides based on enzymatic a-halo ketone reduction with substrate-coupled cofac-
tor-regeneration as a key step [131]. In a first reaction, a-chloro ketones were
converted into the corresponding halohydrins with enantioselectivities of up to
>99% e.e. when using lyophilized whole-cells of R. ruber. For example, (R)-1-chloro-
2-octanol, (R)-18, was formed with >99% conversion and 99% e.e. (Scheme 26.22).
Subsequently, the synthesized halohydrin was transformed into the corresponding
epoxide under basic reaction conditions. As well as a two-step sequence it is possible
to combine both steps in a single-step synthesis, leading to the desired product (R)-
19 with 90% conversion and >99% e.e. Scheme 26.22 shows both synthetic process
options.
step 1: step 2:
R. ruber + KOH
O OH pH > 12 O
lyophilized cells
Cl Cl C6H13
C6H13 i-PrOH (16% (v/v)) C6H13
pH 7.5, buffer,
17 24h, 30°C (R)-18 (R)-19
single step:
R. ruber lyophilized cells,
i-PrOH (10% (v/v)),
buffer, KOH, pH ~13
24h, 30°C
Scheme 26.22
Furthermore, highly enantio- and diastereoselective reduction of diketones to
furnish the corresponding diols with >99% e.e. and >99% d.e. has been reported by
the Kroutil group [132]. Since this synthesis proceeds in a stepwise-fashion, in most
cases the desired diol was obtained as a mixture with the hydroxyketone formed in the
first reduction step. The conversions were good to excellent (65–97%). Wild-type
strains from R. ruber in combination with isopropanol as cosubstrate are also suitable
for the enantioselective reduction of heteroaryl methyl ketones, leading to the desired
products with excellent enantioselectivity at high substrate concentrations of up to
0.4 mol l1 [133].
Kroutil et al. also showed that an ADH from Paracoccus pantotrophus, which has
been overexpressed in E. coli, is exceptionally DMSO-tolerant and can reduce several
ketones by means of a substrate-coupled cofactor-regeneration with isopropanol
(5 vol.%) [134]. The authors also reported the reduction of ketones with two sterically
demanding substituents at a substrate input of 10 g l1 in the presence of wild-type
strains from Ralstonia sp. and Sphingobium yanoikuyae sp. using substrate-coupled
cofactor-regeneration with isopropanol or, alternatively, ethanol [135]. Notably, in the
presence of the wild-type cells from Rhodococcus ruber the reduction of those bulky
ketones only took place when using ethanol as cosubstrate, which indicates that
another enzyme is involved compared to the ADH in this strain for the above-
mentioned reduction reactions.
As an alternative reaction medium, a biphasic system consisting of water and
supercritical carbon dioxide has been used by the Matsuda group [136]. Applying
whole-cells from Geotrichum candidum dried by acetone led to the enantioselective
reduction of a range of ketones, forming the desired alcohols with up to 82% yield
and >99% e.e. Notably, the use of sodium bicarbonate improved the reactivity
significantly.
The use of recombinant whole-cells also proceeded efficiently when operating in a
continuous mode [137]. This has been demonstrated by L€ utz et al. for the reduction of
methyl acetoacetate in the presence of E. coli cells overexpressing an ADH from
Lactobacillus brevis. This process runs at a high substrate concentration of 2.5 mol l1,
and gave the desired alcohol product with a space–time-yield of 700 g l1 d1 and an
enantioselectivity of >99% e.e.
Besides process development of biocatalytic reductions, the combination of
enzymatic reduction processes with other types of reactions towards multistep
one-pot syntheses of chiral building blocks in aqueous media represents a further
interesting extension of biocatalytic reductions. The Kroutil group recently reported
the successful combination of an enantioselective biocatalytic reduction of a-chloro-
ketones under substrate-coupled in situ-cofactor-regeneration with subsequent enzy-
matic ring closure to the epoxide followed by ring opening with an azide or cyanide
nucleophile [138]. The latter steps are catalyzed by a halohydrin dehalogenase, and
proceed under complete retention of the absolute configuration. This three-step one-
pot synthesis furnished the corresponding b-azido-alcohols and b-hydroxynitriles
with conversions of up to >99% and with excellent enantioselectivity of >99% e.e.
Scheme 26.23 shows selected examples.
halohydrin halohydrin
O ADH OH dehalogenase O dehalogenase OH
Cl * Cl * * Nu
R NAD(P)H, R R NaNu R
20 i-PrOH 21 22 (Nu: N3 23
or CN)
Selected examples
OH OH OH
O N3 N3 O CN
Ph n-C6H13 Ph
(S)-23a (R)-23b (S)-23c
98% conversion >99% conversion >92% conversion
>99% ee >99% ee >99% ee
Scheme 26.23
As an example of combining a biocatalytic reduction with a chemocatalytic process

the Gr€oger group recently reported the combination of an ADH-catalyzed reduction
of ketones under substrate-coupled cofactor-regeneration with a palladium-catalyzed
Suzuki-cross coupling reaction in a one-pot synthesis in aqueous reaction media [75].
When carrying out the Suzuki cross-coupling reaction in the initial step starting from
aromatic boronic acids and a halogenated acetophenone, subsequent biocatalytic
reduction gave enantiomerically pure biaryl-containing alcohols with conversions of
up to 91% and high enantioselectivities of >99% e.e. Scheme 26.24 shows a selected
example. This type of synthesis has also been expanded towards the diastereo- and
enantioselective synthesis of C2-symmetric biaryl-containing diols [139]. In addition,
the combination of a Wittig reaction with an asymmetric biocatalytic reduction
towards a two-step one-pot process in aqueous reaction media has been reported by
the same group [76].
O O OH
[Pd(PPh3)2Cl2] (S)-ADH from
CH3 (2 mol%), Rhodococcus sp.,
CH3 CH3
water, 70°C pH 7, r.t.
Br
24
+ 26 (S)-27
B(OH)2 in situ formation, NADH NAD+ 91% conversion
not isolated >99% ee
OH O
25
(1 equiv.) H3C CH3 H3C CH3
(S)-ADH from
Rhodococcus sp.
Scheme 26.24
A combination of an enantioselective organocatalytic aldol reaction and a subse-

quent biocatalytic reduction without isolation of the intermediate resulting from the
O
Ph
N Ph
N H OH
H OH OH
O O
1. organocatalyst (S,S)-30
CH3
H + H3C CH3 (5 mol%), r.t., 20h
2. (S)-ADH, buffer, NADH, Cl
Cl 29 i-PrOH (25% (v/v)), (1R,3S)-31
28 (4 equiv.) r.t., 18h >99% ee
d.r.(syn/anti) = 1:10
80% product-related conversion
(>95% overall conversion)
Scheme 26.25
aldol reaction was also developed by the Gr€oger group [140]. The aldol reaction was
carried out as a solvent-free synthesis and the resulting reaction mixture was directly
passed into an aqueous–isopropanol solution of the enzyme. In such a process the
desired 1,3-diol (1R,3S)-31 was obtained with a product-related conversion of 80%
over two steps (at an overall conversion of >95%), a high diastereomeric ratio of d.r.
(syn/anti) ¼ 1: 10, and an excellent enantiomeric excess of >99% e.e. (Scheme 26.25).
In summary, substrate-coupled cofactor-recycling is a highly efficient methodol-
ogy. The resulting enantioselective enzymatic reductions have been performed very
successfully with a broad range of ketones, and industrial applications based on this
biocatalytic synthetic concept have also already been reported. Further examples of
the substrate-coupled cofactor-regeneration with isolated enzymes and whole-cells,
in particular with respect to the enzymatic reduction of specific types of ketones,
which are, for example, of pharmaceutical importance, are discussed in Section 26.4.
26.3.3
Enzyme-Coupled Cofactor-Regeneration Using a Formate Dehydrogenase

When applying an enzyme-coupled cofactor-regeneration for asymmetric biocata-
lytic reduction processes, the use of a formate dehydrogenase (FDH) is one of
the most popular approaches. This concept, which has been described in detail
in Section 26.2.1, is based on the FDH-catalyzed oxidation of formate into
carbon dioxide, while reducing the oxidized form of the cofactor into its reduced
form, NAD(P)H. The most frequently applied formate dehydrogenase is most likely
the FDH from Candida boidinii and optimized mutants thereof [141]. These
recombinant enzymes were developed in the Kula group who are – jointly with
the Wandrey and Hummel group – pioneers in the field of FDH-based applica-
tions [142–144] in addition to the Whitesides group [145]. A major advantage of the
FDH-based cofactor-regeneration is the irreversible step of carbon dioxide formation
and removal, thus shifting the equilibrium towards (complete) product formation.
A further advantage is the simple work-up, since (ideally) no organic by-product
remains in the reaction mixture.
The initial work on enzymatic reduction of ketones with a FDH for cofactor-
recycling has been carried out based on the use of isolated enzymes in homogeneous
aqueous media. Owing to the low solubility of the hydrophobic ketones in water, the
reactions were long carried out at low substrate concentrations, typically in the range
of 5–20 mM or below. The suitability of different types of ADHs in combination with
a FDH for the enantioselective reduction of a broad range of ketones has been studied
in detail by Hummel et al. as well as the Kula group [144, 146–149]. For example,
ADHs from R. erythropolis and C. parapsilosis were used in combination with the FDH
from C. boidinii. When carrying out reductions of several keto esters and a keto dialkyl
acetal at a substrate concentration of 100 mM the desired alcohols were typically
obtained with high conversion (up to 100%), and high enantioselectivities of >99%
e.e. Scheme 26.26 gives a selected example [148].
OH O
HCO2- NAD+
H3C OCH3
(S)-33
90% conversion
>99% ee
FDH from (S)-ADH from

Candida boidinii Rhodococcus erythropolis
O O
CO2 NADH H3C OCH3

32
Scheme 26.26
The issue of high space–time yields despite the limitation by low ketone solubility
has been addressed by the Wandrey group, who developed elegant engineering
solutions by means of continuously operating processes with an enzyme-membrane
reactor. The efficient three-loops-concept consists of the following steps: (i)
enzymatic reaction in pure aqueous medium, (ii) separation of the aqueous phase
from the enzyme via ultrafiltration, and (iii) subsequent continuous extraction of the
aqueous phase with an organic solvent. The organic and aqueous phases are
separated by a hydrophobic membrane [150–153]. Albeit the reaction in this
enzyme-membrane reactor is limited by the low solubility of the ketone in water
(9–12 mM), good space–time yields in the range 60–104 g l1 d1 have been
obtained. This has been demonstrated for the synthesis of, for example, (S)-1-
phenylpropan-2-ol and (S)-4-phenylbutan-2-ol, which were obtained in enantiomeri-
cally pure form. A representative example, including the process scheme, is shown in
Scheme 26.27. An extended emulsion membrane reactor concept has also been
applied by Wandrey et al. for the asymmetric reduction of 2-octanone [154].
A conversion of 97% has been achieved at a residence time of 1 h, corresponding
to a space–time-yield of 21.1 g l1 d1. Notably, this emulsion membrane reactor has
been operated over a period of >4 months. Further processes in a continuous mode,
ADH,
O FDH, OH
NAD+
H3C H3C
34 CSTR (S)-35
(enzyme 72% conversion
bimembrane space-time-yield:
reactor) 64 g/(l*d)
Scheme 26.27
which were reported by Kula et al., are based on the use of cyclodextrin-containing
buffers. In such an aqueous reaction medium, a high stability of the enzymes,
namely, an (S)-ADH from C. parapsilosis and a FDH from C. boidinii, was observed
and for the synthesis of (S)-1-naphthylethan-1-ol a high space–time-yield of
120 g l1 d1 was obtained [155]. The cyclodextrin-containing buffers have also
been successfully used by the same group for batch-mode reactions.
The use of a biphasic reaction medium represents an alternative solvent system,
which is also suitable in particular for processes in a batch mode. Although the use of
organic solvents could improve the solubility of water-immiscible ketones the known
instability of the FDH from C. boidinii towards many organic solvents makes this type
of reaction media engineering difficult. Addressing this issue, Gr€ oger and Hummel
et al. developed a suitable aqueous–organic two-phase solvent reaction medium
based on the use of n-heptane and n-hexane as organic phases [80, 81]. In this reaction
medium a recombinant (S)-ADH and a mutant of the FDH from C. boidinii remained
stable. The reductions proceed with good conversions and high enantioselectivities
with various aromatic ketones as substrates. Although reactions proceed at substrate
concentrations of up to 200 mM, at higher concentrations conversions decrease and
prolonged reaction times are required. A further improvement of the substrate
concentrations up to 500 mM has been realized when using an emulsion system for
the synthesis of the corresponding alcohols [156, 157]. For example, the reduction of
4-chloroacetophenone as a model substrate on a 6-l scale, leading to the desired (S)-
alcohol with >98% conversion and >99.4% e.e., was reported. As enzymes, the ADH
from R. erythropolis and the FDH from C. boidinii have been used.
In the presence of a biphasic reaction medium consisting of an aqueous phase and
toluene (20%) and by means of a so-called diketoreductase from Acinetobacter baylyi
and a formate dehydrogenase, an efficient enantioselective reduction of ethyl 2-oxo-4-
phenylbutyrate has been achieved by Chen and coworkers [158]. When operating at a
high substrate loading of 164.8 g l1 the desired product ethyl (S)-2-hydroxy-4-
phenylbutyrate, (S)-37, was obtained with a conversion of 91.8% and an excellent
enantioselectivity of 99.5% e.e. (Scheme 26.28). Subsequent work-up furnished the
product (S)-37 in 88.7% yield. Other types of ketones have been also reduced
successfully by means of this recombinant whole-cell catalyst, although conversion
and/or enantioselectivity were lower in these cases.
Enantioselective enzymatic reductions of ketones based on a FDH as isolated
enzyme for cofactor-recycling have also been applied by the Patel group in the
synthesis of (S)-2-pentanol [(S)-39] using an ADH from Gluconobacter oxydans (SC
recombinant ADH
O from Acinetobacter baylyi, OH
OEt FDH OEt
aqueous buffer (pH 6.0),
O O
toluene (20% (v/v)),
36 r.t., 9 h, NaHCO2, NAD+ (S)-37
(0.8 M, 164.8 g/l) 91.8% conversion
99.5% ee
88.7% yield
Scheme 26.28
13851) [159]. As biocatalyst G. oxydans cells, pretreated with Triton X-100, were used
in combination with the formate dehydrogenase from C. boidinii. This process was
carried out at a 1500-l scale with a substrate input of 3.2 kg (2.13 g l1), and the
desired (S)-2-pentanol [(S)-39] was formed with a conversion of 32.2% and an
enantioselectivity of >99% e.e. (Scheme 26.29; see also Chapter 29).
(S)-ADH
(Gluconobacter oxydans cells,
pre-treated with Triton X-100),
FDH from
O OH
Candida boidinii
H3C CH3 H3C CH3
NAD+, NaHCO2
38 (S)-39
32.2% conversion
>99% ee
Scheme 26.29
For a long time, a major limitation for applications using the FDH from C. boidinii
was its inability to regenerate NADPþ as a cofactor, thus limiting it to the regeneration
of NADþ only. An elegant solution of this problem has been found by the Hummel
group, expanding the application range of FDH-based cofactor-regeneration also to
NADPþ -dependent ADHs [160]. As suitable enzyme the highly efficient ADH from L.
kefir [161, 162] was chosen. The key step is the integration of an additional enzymatic
step within the cofactor-regeneration cycle, namely the pyridine nucleotide transhy-
drogenase (PNT)-catalyzed regeneration of NADPH from NADPþ under consump-
tion of NADH and formation of NADþ [160]. The concept is shown in Scheme 26.30,
and exemplified for the synthesis of (R)-phenylethanol [(R)-41].

An interesting alternative process option for enantioselective reductions of ketones is
the direct use of recombinant whole-cells, overexpressing a suitable ADH and a
formate dehydrogenase, as biocatalysts. Such an efficient FDH-based whole-cell
catalyst for synthetic applications has been developed by researchers from Daicel
Chemical Industries Ltd., who constructed a recombinant E. coli W3110 strain
that co-expresses an ADH from Pichia finlandica and a FDH from Mycobacteri-
OH
- H3C
HCO2 NAD+ NADP+
(R)-41
pyridine
FDH nucleotide (R)-ADH
(NADH-dependent) transhydrogenase (NADPH-dependent)
CO2 NADH NADPH H3C
40
Scheme 26.30
um [163]. This tailor-made whole-cell catalyst has been successfully applied, for
example, in the enantioselective reduction of ethyl 4-chloro-3-oxobutanoate (42)
to give the corresponding (S)-alcohol, (S)-43, at 32.2 g l1, with 98.5% yield and 99%
e.e. (Scheme 26.31).
OH O
- Cl
HCO2 NAD+ OEt
(S)-43
98.5% yield
99% ee
tailor-
FDH made ADH
whole-cell
catalyst
O O
Cl
CO2 NADH OEt
42
substrate input:
32.2 g/l
Scheme 26.31
Analogous whole-cell-catalyzed biotransformations in a continuous mode have

been studied in detail by the L€utz group [137]. When using methyl acetoacetate as a
substrate at a concentration of 30 mM in buffer and a recombinant whole-cell catalyst
overexpressing an ADH from L. brevis and a FDH from C. boidinii, a maximal
space–time-yield of 56 g l1 d1 and a maximal conversion of 66% was obtained.
However, a drastic drop in conversion has been observed during a 24 h reaction time,
indicating a high deactivation rate of this catalyst.
As alternative biphasic reaction media for biocatalytic reduction with whole-cell
catalysts, overexpressing an ADH from L. brevis and a FDH from C. boidinii, ionic
liquids as a water-immiscible phase have been used successfully by the Weuster-Botz
group [164]. For example, a space–time yield of 180 g l1 d1 in combination with a
high product-related conversion of 95% and 97% e.e. have been obtained for the
synthesis of (R)-2-octanol in this kind of biphasic reaction medium.
26.3.4
Enzyme-Coupled Cofactor-Regeneration Using a Glucose Dehydrogenase

A further efficient option to regenerate the cofactor NAD(P)H is based on the use of a
GDH, which oxidizes D-glucose to D-gluconolactone under transformation of the
oxidized cofactor NAD(P) þ into the required reduced form, NAD(P)H. Since
D-gluconolactone is subsequently hydrolyzed and neutralized to afford a D-gluconate
salt this process also contains an irreversible step, thus shifting the whole reaction
towards formation of the desired alcohol product. Synthetic applications of GDH-
coupled cofactor-regeneration in asymmetric ketone reduction have been reported by
means of isolated enzymes [165–167] as well as recombinant whole-cell systems.
Representative examples of this technology are given below.
Pioneering work for an enzymatic reduction using a GDH-coupled cofactor-
regeneration process has been carried out by Wong and coworkers [165, 166]. The
corresponding enzymatic reduction of ketones in the presence of different types of
ADHs such as ADHs from horse liver, yeast, and Thermoanaerobium brockii gave the
desired alcohols, for example, (S)-45, with good to high enantioselectivities. Both
ADH and GDH were used in an immobilized form. The yields of these enzymatic
biotransformations were in most cases about 90%. Albeit enantioselectivities varied,
they exceeded 90% e.e. in many cases. A selected example is shown in Scheme 26.32.
OH
CF3
D-gluconic irreversible D-glucono-
NAD(P)+
acid lactone (R)-45
ca. 90% yield
94% ee
GDH from ADH from

Bacillus cereus Thermoanaerobium brockii
CF3
D-glucose NAD(P)H
44
Scheme 26.32
The recombinant NADPH-dependent and (R)-enantioselective ADH from Lac-
tobacillus kefir overexpressed in E. coli by the Hummel group has been very
successfully used as an isolated enzyme in combination with a GDH [67]. In addition
to substrate-coupled cofactor regeneration, the enzyme-coupled cofactor-regenera-
tion with a GDH also turned out to be very useful, leading to a broad range of chiral
alcohols [(R)-47] with high conversions and excellent enantioselectivities of typically
>99% e.e. Scheme 26.33 gives selected synthesis examples.
O (R)-ADH OH
from Lactobacillus kefir
2
R1 R R 1 R2
GDH,
46 D-glucose, NADP+, MgCl2, (R)-47
aqueous buffer (pH 7.5),
20-120 min, 37°C
Selected examples
OH OH
OH
H3C
CH3 H3C CH3
CH3
OH
(R)-47a (R)-47b (R)-47c
>99% ee >99% ee >99% ee
O OH O OH O O OH
Cl Cl
EtO CH3 EtO EtO
(R)-47d (S)-47e (S)-47f
>99% ee >99% ee >99% ee
Scheme 26.33
The Hua group has developed an enantioselective reduction of a-chlorinated

ketones in the presence of isolated ADHs, regenerating the cofactor with a
GDH [168]. A range of a-chlorinated alcohols were formed in high yields of
72–99%, and with excellent enantioselectivities of typically >99% e.e.
Enantioselective reduction of, for example, benzoyl hydroxyacetone and a-tetra-
lone in the presence of an isolated ADH (ketoreductase; KRED) and a GDH has been
described by BioCatalytics researchers [169]. Using the isolated ADHs (1–7 wt%
compared to the amount of substrate) and a catalytic amount of cofactor led to the
synthesis of the optically active alcohols, for example, (S)-49, in high yields. The
reductions have been carried out at high substrate concentrations of up to
0.75–1.4 M. A selected example is shown in Scheme 26.34.
The combination of an ADH-catalyzed reduction of ethyl 4-chloro-oxo-butanoate
with a subsequent halohydrin dehalogenase-catalyzed replacement of the chloro
O ADH KRED117, OH
O GDH O
H 3C H3C
O glucose, NADPH O
buffer / DMSO
48 (S)-49
(249 g/l) 92% yield
>98% ee
Scheme 26.34
substituent by a cyano group has been successfully demonstrated by Huisman and

Sheldon et al., thus realizing a sustainable two-step three-enzyme process for the
synthesis of (R)-4-cyano-3-hydroxybutanoate as a key intermediate in the manufac-
ture of atorvastatin [170]. In the first enzymatic step, which proceeds under in situ-
regeneration of the cofactor with a GDH, the desired ethyl (S)-4-chloro-3-hydro-
xybutanoate is obtained in 96% yield and with >99.5% e.e.

The design of recombinant whole-cells, which not only contain the required cofactor
but also both of the desired enzymes, ADH and GDH, in overexpressed form is an
elegant approach towards tailor-made (bio-)catalysts for the reduction of ketones.
Advantages of such a recombinant whole-cell over a wild-type cell are the higher
amount of the desired (overexpressed) enzymes within the cell, their cost-effective
access, and excellent performance in synthetic applications. Technical applications of
the recombinant whole-cell technology based on an ADH and a GDH have also
already been reported, for example, by the Kaneka Corporation and Degussa AG (now
Evonik Degussa GmbH). The development of such a recombinant whole-cell
technology and applications thereof are described in the following.
The pioneering work in design and application of highly efficient recombinant
whole-cell biocatalysts, consisting of an ADH and a GDH, has been carried out by
Shimizu and coworkers [171]. As enzyme the GDH from Bacillus megaterium was
used, accepting both NADH and NADPH as a cofactor. Already in the 1990s, Shimizu
et al. developed an effective E. coli catalyst, as well as a highly efficient reaction system
for the reduction of 4-chloro-3-oxobutanoate [91, 92, 172–174]. The use of these
efficient recombinant whole-cell catalysts in the enantioselective reduction of ethyl
4-chloro-3-oxobutanoate (50) to form the corresponding pharmaceutically important
alcohol (R)-51 has been intensively investigated and optimized by the Shimizu
group. As reaction media, an n-butyl acetate–water two-phase solvent system turned
out to be beneficial [175]. When using the E. coli host organism overexpressing an
NADP þ -dependent ADH from Sporobolomyces salmonicolor and a GDH as biocata-
lysts, the desired optically active (R)-alcohol (R)-51 was formed with up to 255 g l1 in
the organic phase [176, 177]. A conversion of 91% and an enantioselectivity of 91%
e.e. was observed [177]. As well as glucose as a cosubstrate, a low amount of NADP þ
is required. A further improvement was achieved on using E. coli, co-expressing both
the ADH from S. salmonicolor and the GDH from B. megaterium, resulting in the
formation of the desired alcohol (R)-51 with 94.1% conversion and 91.7% e.e. when
NAD(P)+ OH O
D-glucono- Cl
lactone OEt
(R)-51
94.1% conversion
91.7% ee
GDH ADH
E. coli cells
O O
Cl
D-glucose NAD(P)H OEt
50
(300 g/l substrate input)
aqueous phase organic phase

(n-butyl acetate)
Scheme 26.35
operating at a substrate concentration of 300 g l1 and adding a catalytic amount of

the NADP þ cofactor [178]. Scheme 26.35 shows a result of this application using a
tailor-made whole-cell biocatalyst in a two-phase reaction medium.
Notably, the Shimizu group also designed a whole-cell catalyst for the synthesis
of the analog (S)-enantiomeric form of ethyl 4-chloro-3-hydroxybutanoate [179].
A further improved, highly efficient process for this enantiomer has been
obtained by using a recombinant whole-cell catalyst overexpressing an ADH
from Candida magnoliae as ADH-type catalyst for the asymmetric reduction. In
the presence of such a whole-cell biocatalyst bearing an ADH and GDH a
conversion of 96% and an enantioselectivity of >99% e.e. has been obtained [179].
In addition, Kaneka researchers jointly with the Shimizu group reported the
extension of this reduction technology to the reduction of other types of functio-
nalized b-keto ester substrates, for example, 4-bromo-3-oxobutanoate [93], and a
range of other substrates [171]. This impressive biocatalytic reduction technology
developed by the Shimizu group has already been commercialized. Since 2000,
Kaneka Corporation has applied this methodology for the manufacture of ethyl
(S)-4-chloro-3-hydroxybutanoate on an industrial scale (see also Chapter 29,
Industrial Applications) [171].
Furthermore, an E. coli whole-cell catalyst, harboring an (R)-selective ADH from L.
kefir and a GDH from Bacillus subtilis, has been successfully constructed by the
Hummel group [180].
In addition, Degussa researchers jointly with the Hummel group reported the
application of recombinant whole-cell biocatalysts in asymmetric reductions of a
range of ketones at high substrate input, exceeding 150 g l1, in pure aqueous media,
and in general without the need of addition of an external amount of cofactor [79]. By
means of this process technology both (R)- and (S)-enantiomers of alcohols can be
E. coli whole-cell catalyst
containing
(S)- or (R)-ADH,
GDH,
O OH OH
NAD(P)+
or
R1 R2 R1 R2 R1 R2
D-glucose
52 (S)- or (R)-53
Selected examples
OH OH OH
Br
CH3 CH3
Cl O Br
(S)-53a (R)-53b (S)-53c

94% conversion >95% conversion 94% conversion
>99.8% ee >99.4% ee 97% ee
(156 g/l substrate input) (212 g/l substrate input) (140 g/l substrate input)
Scheme 26.36
synthesized when using the corresponding (R)- and (S)-selective recombinant

whole-cell catalysts. This methodology, which is economical and simple to carry
out, has been used for the synthesis of a broad range of (R)- and (S)-alcohols, (R)- and
(S)-53. Typical substrate concentrations are in the range of 1 M, thus exceeding
100 g l1. The reduction proceeds with high conversions of up to >95%, and with
high enantioselectivities of up to >99.8% e.e. Scheme 26.36 shows selected exam-
ples. The synthesis of a fluorinated 4-phenylethan-1-ol as well as aliphatic halohy-
drins also turned out to proceed efficiently and with high enantioselectivity of >99%
e.e. with recombinant whole-cell catalysts [181, 182]. After further optimization, this
recombinant whole-cell reduction technology platform has already been applied on
an industrial scale at Degussa AG (see also Chapter 29).
A further efficient process based on the use of recombinant E. coli cells over-
expressing ADH and GDH has been reported by Ema and Sakai et al. with the
efficient enantioselective reduction of a broad range of ketones [85, 183]. An ADH
from Saccharomyces cerevisiae has been identified to be highly suitable. For co-
expression with the GDH different types of plasmids have been constructed. Under
optimized conditions, a broad range of chiral alcohols (55) have been successfully
prepared in enantiomerically pure form. Selected examples are shown in
Scheme 26.37. Notably, this reduction technology was also successfully extended
to the preparation of an intermediate for clopidogrel (Section 26.4.7).
An elegant alternative whole-cell concept for the biocatalytic reduction of ketones
under enzymatic in situ cofactor-regeneration has been reported by the Li group [184].
For the reduction process, two different types of permeabilized whole-cells have been
used simultaneously, namely, an ADH-containing microorganism and in parallel a
GDH-containing microorganism. This is outlined in Scheme 26.38. It has been
recombinant E. coli whole-cells
containing
ADH from
Saccharomyces cerevisiae,
O GDH, NADPH O
R1 R2 R1 R2
54 55
Selected examples
OH O OH OH
CH3 CO2CH3 OAc
H 3C H 3C
(S)-55a (S)-55b (S)-55c
>99% ee >99% ee 99% ee
OH OH
CO2CH3 OAc
(S)-55d (S)-55e
68% yield 78% yield
>98% ee 98% ee
Scheme 26.37
successfully demonstrated for the enantioselective reduction of ethyl 3-oxo-4,4,4-

trifluorobutyrate, leading to the desired (R)-alcohol with 91% e.e. For this reaction,
Bacillus pumilus (bearing an ADH) was chosen as the microorganism for the
reduction process, and Bacillus subtilis (which contains a GDH) was used for
cofactor-regeneration.
O OH
∗
glucono-
1 2 1 glucose
R R R R2 lactone
aqueous
phase
O OH
∗
glucono-
glucose
R 1 R2 R 1 R2 lactone
ADH GDH
NAD(P)H NAD(P)+ NAD(P)+ NAD(P)H
permeabilized permeabilized
microorganism A microorganism B
Scheme 26.38
26.3.5
Enzyme-Coupled Cofactor-Regeneration Using a Glucose-6-Phosphate
Dehydrogenase

A further type of enzymatic cofactor-regeneration, which is related to the one
with a GDH, is based on the use of a glucose-6-phosphate dehydrogenase (G-6-
PDH). As substrate glucose-6-phosphate is needed, which can either be formed
by the cells starting from glucose (in the case of a whole-cell approach) or used
directly (in the case of isolated enzymes). The latter option, however, is less attractive
since glucose-6-phosphate is quite expensive. Accordingly, syntheses based on this
type of cofactor-regeneration with isolated enzymes have only been reported
for small-scale applications and are typically performed, for example, for
characterization of the biochemical properties of ADHs with respect to their
substrate spectra.
Pioneering work in the synthesis of chiral alcohols by means of this methodology
was carried out by Whitesides and Wong [185, 186]. Combination of the G-6-PDH
with the ADH from L. kefir for the synthesis of optically active (R)-phenylethan-1-ol
has been additionally reported by Hummel [187], and the Stewart group applied their
impressive set of 19 recombinant ADHs from Saccharomyces cerevisiae in a screening
of numerous a- and b-keto ester substrates by means of G-6-PDH-based reduc-
tions [86, 188]. Scheme 26.39 shows selected examples.

In contrast to the use of isolated enzymes in G-6-PDH-coupled cofactor-regeneration
processes, the whole-cell-based approach has potential for large-scale applications.
Such a process has been reported by Hanson and Patel et al. [189, 190] for the
synthesis of (S)-2-chloro-1-(30 -chloro-40 -fluorophenyl)ethanol ((S)-59). This product
serves as an intermediate for an IGF-1 receptor antagonist, which is a leading drug
candidate in an anticancer program. The applied whole-cell biocatalyst reduced 2,30 -
dichloro-40 -fluoroacetophenone (58) enantioselectively to furnish the desired alcohol
(S)-59 in 89% yield and with >99% e.e. (Scheme 26.40) [84]. The substrate input of
this reduction was 20 g l1, and the intact E. coli cells, which contain a ketoreductase
and a G-6-PDH from S. cerevisiae in overexpressed form, were provided with glucose.
The cofactor was then regenerated by G-6-PDH from S. cerevisiae.
In addition, the Li group reported the use of permeabilized cells of Bacillus
pumilus containing an ADH and a G-6-PDH in the reduction of ethyl 3-oxo-4,4,4-
trifluorobutanoate [191]. These permeabilized cells turned out to be stable and active
over a long period of time. In the presence of 1 mM of externally added cofactor
NADP þ and at a substrate concentration of 60 mM, the reduction afforded the
corresponding (R)-alcohol with 98% conversion and 95% e.e. A disadvantage
of this strategy, however, is the need to use the expensive cosubstrate glucose-6-
phosphate (instead of the economically more favorable glucose when using non-
permeabilized cells).
(a) reduction of α-keto esters
ADH from yeast (YOL151w),

glucose-6-phosphate dehydrogenase,
O glucose-6-phophate, O
R NADP+ R
OEt OEt
O OH
56
Selected examples
O O
H3C
OEt OEt
OH OH
56a 56b
(b) reduction of β-keto esters
ADH from yeast (YDR368w),

glucose-6-phosphate dehydrogenase,
O O glucose-6-phosphate, OH O
NADP+
R1 OEt R1 OEt
2
R R2
57
Selected examples
OH O OH O OH O
H3C OEt H3C OEt H3C OEt

CH3
CH3
57a 57b
57c
Scheme 26.39
26.3.6
Enzyme-Coupled Cofactor-Regeneration Using a Phosphite Dehydrogenase
A further, more recently developed enzymatic cofactor-regeneration technology is

based on the use of a phosphite dehydrogenase [192, 193]. This enzyme catalyzes the
regeneration of NAD(P)H from NAD(P)þ by oxidizing of phosphite into phosphate.
The use of phosphite as a cheap and readily available cosubstrate as well as favorable
kinetic parameters (such as low Km values for both NADþ and phosphite) represent
advantages of this method, thus making it an interesting alternative to other
enzymatic cofactor-regeneration technologies. The application of a phosphite dehy-
drogenase in combination with an ADH for the reduction of ketones has been
reported by Zhao et al. recently [194]. The reduction of acetophenone (60) in the
recombinant E. coli whole-cells
containing
ADH from Hansenula polymorpha,
O glucose-6-phosphate dehydrogenase OH
from Saccharomyces cerevisiae,
Cl NADPH Cl
F D-glucose, NADP+ F
Cl Cl
58 (S)-59
(substrate input: 89% yield
ca. 20 g/l) >99% ee
Scheme 26.40
presence of an ADH from L. brevis under cofactor-recycling with an engineered

phosphite dehydrogenase from Pseudomonas stutzeri in a batch mode furnished
the corresponding (R)-phenylethanol, (R)-61, with quantitative conversion and a
space–time yield of 88 g l1 d1 (Scheme 26.41). This reduction has also been carried
out successfully in a continuous mode. Furthermore, an efficient enzymatic reduc-
tion of D-xylose in combination with this phosphite dehydrogenase leading to xylitol
with an impressive space–time yield of 230 g l1 d1 has been reported by the same
authors [194].
26.3.7
Ketone Reduction Based on Wild-Type Microorganism and Glucose in a
Fermentation-Like Processes
A further approach is based on the use of wild-type microorganism and D-glucose as a

cosubstrate, which is consumed by the (living) wild-type microorganism, delivering
the required cofactor form NADH or NADPH from cell-internal metabolic process-
es. As potential drawbacks of this approach one can regard the large amount of
biomass required due to typically low enzyme expression, potential side-reactions
O OH
ADH from
CH3 L. brevis CH3
60 (R)-61
NADPH + H+ NADP+
O O
P P
H O phosphite dehydrogenase HO O
O O
from
Pseudomonas stutzeri
Scheme 26.41
from enzymes being expressed in a similar (or higher) manner, and typically low
volumetric productivity. However, an advantage of this approach using wild-type
whole-cells, for example, bakers yeast, is the easy access to the biocatalyst for groups
lacking the ability to produce recombinant whole-cells. Thus, without the need for
isolation and cloning of the desired ADH direct fermentation of the wild-type
microorganisms can represent a fast route to biomass suitable for the required
biotransformation. The use of bakers yeast as biocatalyst in organic synthesis is
widely known and synthetic applications have been reviewed extensively [195].
More recent examples of this method include, for example, reduction of aryl-
substituted acetone [196], 1,3-diketones [197–200], a 1,4-diketone [201], 1-heteroaryl
ketones [202], b-hydroxyketones [203], a-ketoesters [204], and b-ketoesters [205, 206].
Besides bakers yeast a broad range of other wild-type cells such as, for example,
Geotrichum candidum [207], Lactobacillus kefir [123], Rhizopus arrhizus [208], Phaseolus
aureus L [209], and Pisum sativa [210] have also been used successfully as biocatalysts
in enantioselective ketone reductions.
This section will focus on – considering the existing reviews for transformations
with bakers yeast and other wild-type microorganisms – selected examples of
fermentation-like processes, in particular with respect to process development and
applications for the synthesis of fine chemicals and pharmaceutical ingredients.
The commercial viability of asymmetric microbial ketone reduction based on the
use of bakers yeast for industrial-scale applications has been reported by researchers
of Rohner Ltd. [211, 212]. The large-scale synthesis of (S)-3-hydroxybutyric acid ethyl
ester [(S)-63] at Rohner Ltd. via reduction of the corresponding b-ketoester 62 was
carried out in water at ambient temperature using bakers yeast as biocatalyst and
sugar. The desired product (S)-63, which is an intermediate in the synthesis of chiral
drugs such as carbapenem ( þ )-PS 5, thienamycin, daunosamine, benzothiazepin,
and carumonam, was obtained in 60–75% yield and with an enantiomeric excess of
>98% e.e. (Scheme 26.42¸ see also Chapter 29) [211]. Another example of a
commercial-scale application of microbial reduction at the same company is the
synthesis of (1R,2S)-cis-2-hydroxycyclohexane carboxylic acid ethyl ester [211]. Using
bakers yeast as a biocatalyst afforded this alcohol in 70% yield, with a diastereos-
electivity of >97% d.e. and an enantioselectivity of >93% e.e.
A further microbial reduction based on wild-type whole-cell microorganisms as
biocatalysts is the synthesis of an intermediate for Paclitaxel, which is the active
pharmaceutical ingredient of Taxol, in the presence of strains of Hansenula [213].
Taxol has been developed commercially by Bristol-Myers Squibb, and is an antimi-
totic agent used for the treatment of various types of cancer. For the synthesis of the
O O baker´s yeast, OH O
sugar
H3 C OEt H3C OEt
water, 20-50h
62 (S)-63
60–75% yield
>98% ee
Scheme 26.42
chiral C13 side chain of this complex compound, in a single-stage bioreduction
process cells of Hansenula fabianii SC 13894 were grown in a 15-l fermenter for 48 h.
When using 2-keto-3-(N-benzoylamino)-3-phenylpropionic acid ethyl ester (64) as a
substrate and glucose as a cosubstrate, the product (2R,3S)-N-benzoyl-3-phenyl
isoserine ethyl ester ((2R,3S)-65) was obtained with a reaction yield of 88% and
95% e.e. (Scheme 26.43).
O O
NH NH
Hansenula fabianii SC13894
CO2Et CO2Et
glucose, 72h
O OH
64 (2R,3S)-65
88% reaction yield
95% ee
Scheme 26.43
Suspended whole-cells of Candida sorbophila were used as biocatalyst in a microbial

asymmetric ketone reduction process established at Merck & Co, Inc. for the
synthesis of the N-acylated aminoalcohol (R)-67 [89, 90]. This whole-cell catalyzed
reduction process, which runs batchwise on a multi-kg scale using a reactor volume
of 280 l (see also Chapter 29), led to the formation of the desired product (R)-67 with
excellent (>99%) conversion (Scheme 26.44) [214–216]. Furthermore, the product
(R)-67 was obtained in 82.5% yield. The enantioselectivity was >98% e.e. and has
been further increased to 99.8% e.e. after purification.
O OH
H Candida sorbophila H
N whole-cells N
O D-glucose, pH 5.6, 34°C

O
N NO2 N NO2
two-phase (aqueous/solid)
66 (R)-67
(0.2 M; added as >99% conversion
ethanol slurry) 82.5% yield
>98% ee; 99.8% ee after purification
Scheme 26.44
A further application of a whole-cells biocatalyst is the microbial asymmetric

reduction of 3,4-methylenedioxyphenylacetone (68) at Eli Lilly (Scheme 26.45)
[217–219]. The resulting alcohol (S)-69 is an intermediate in the synthesis of
LY300164. As a biocatalyst suspended whole-cells of Zygosaccharomyces rouxii were
used. The limit of the substrate is 6 g l1 due to its toxicity to the microorganism, and
therefore the substrate is adsorbed on a XAD-7 resin, enabling a ketone loading of
80 g l1. By means of this adsorption technology desorption of the substrate is
limited, thus keeping the permanent substrate and product concentration in the
aqueous phase at the desired low level [of less than about 2 g l1 of 68 and (S)-69].
Zygosaccharomyces rouxii CH3
O CH3 whole cells O CH3 O O
N
O O O OH O
pH 7.0, 33 - 35°C NH CH3
aqueous medium
68 (S)-69
(40 g/l; 96% yield
adsorbed on XAD-7 resin) >99.9% ee
H2N
LY300164
Scheme 26.45
After separation of the yeast cells, the product is liberated from the resin by acetone
washing. Notably, the resin can be re-used three times without loss of performance.
Using this type of microbial reduction via in situ product removal, the corresponding
(S)-alcohol (S)-69 is obtained in 96% yield with an excellent enantiomeric excess of
>99.9% e.e. [217, 219]. This process also has been carried out on a kg scale based on a
batch process with a reactor volume of 300 l (see also Chapter 29).
Further examples applying a fermentation-like microbial reduction process using
glucose are biocatalytic reductions of the C¼O double bonds in steroid molecules.
These reductions as well as further microbial reductions of specific ketones will be
described subsequently in Section 26.4 on specific applications of enzymatic
reductions.
26.3.8
Cofactor Regeneration Using Chemocatalytic and Electrochemical Methods
In addition to the enzymatic cofactor-regeneration processes described in

Sections 26.3.1–26.3.7, chemocatalytic as well as electrochemical cofactor-regener-
ation methodologies have been developed. A chemocatalytic cofactor-regeneration
for enantioselective ADH-catalyzed ketone reductions, which is based on the use of a
Rh(III)-metal complex, has been reported by the Steckhan group [220]. Such a Rh
(III)-catalyst acts like a formate dehydrogenase, thus oxidizing formate to carbon
dioxide while reducing the oxidized cofactor NAD(P) þ to NAD(P)H. In the presence
of [Cp Rh(bpy)H2O]Cl2 (70) as chemocatalyst and an (S)-ADH from Rhodococcus sp.
as biocatalyst, enantioselective reduction of 4-phenyl-2-butanone, 72, proceeds with
89% conversion and an excellent enantioselectivity of >99% e.e. (Scheme 26.46). The
opposite (R)-enantiomer was obtained with 81% conversion and 96% e.e. when
applying an (R)-enantioselective ADH from Lactobacillus kefir.
A further alternative for cofactor recycling is electrochemical methods [221].
However, in contrast to the electrochemical oxidation of reduced cofactors
NAD(P)H (which has been widely used in combination with oxidative biotransfor-
mations), analogous reduction of oxidized cofactors NAD(P) þ (resulting from
enzymatic reduction reactions) turned out to be more challenging. This is due to
the formation of cofactor radicals by a one-electron transfer reduction, and subse-
quent formation of the cofactor dimers as an undesired side-reaction. This can be
suppressed by an indirect electrochemical cofactor-regeneration, which requires
OH
∗
CO2 [Cp*Rh(bpy)H]+ NAD(P)+ CH3

71
73
H2O H2O
ADH
- CH3
HCO2 [Cp*Rh(bpy)H2O]2+ NAD(P)H
70 72
Scheme 26.46
the presence of a mediator and a second enzyme. In such a process, the mediator is
electrochemically reduced in a one-electron transfer process and the enzyme is
capable of accepting two electrons from two reduced mediator moieties. In addition,
the reduced form of the enzyme itself transfers an electron pair to the oxidized
form NAD(P) þ . Based on this concept, a suitable process for a range of ketone
reductions has been developed by Yoneyama and coworkers when using a combi-
nation of an ADH with a diaphorase (for NAD þ ) or ferredoxin-NADP þ reductase
(for NADP þ ) as second enzyme and methyl viologen as a mediator [222].
For example, in the presence of a NAD(P)H-dependent ADH from Thermoanaer-
obacter brockii reduction of acetophenone (74) proceeds with 61% yield and led to
the formation of the corresponding (R)-1-phenylethanol, (R)-75, with 98% e.e.
(Scheme 26.47).
OH
CH3
MV NAD(P)+
(R)-75
61% yield
e- 98% ee
ferredoxin-
ADH from
NADP+ T. brockii
reductase
CH3
MV2+ NAD(P)H
74
Scheme 26.47
26.4 Specific Synthetic Applications of Enzymatic Reductions j1081
26.4
Specific Synthetic Applications of Enzymatic Reductions
26.4.1
Introduction and General Remarks
The typical substrate spectrum of many ADHs currently used in organic synthesis
includes acetophenone and substituted derivatives thereof, 2-alkanones and simple
a- and b-keto esters such as, for example, ethyl acetoacetate and ethyl 4-chloro-3-oxo-
butanoate. In general, ketones are well tolerated by numerous ADHs when bearing a
large and a small substituent (such as the aceto-type substrates mentioned above).
Numerous efficient biocatalytic processes based on different types of cofactor-
regeneration methodologies have been developed for the reduction of these types
of molecules, as demonstrated by the biotransformation processes described in
Section 26.3. However, the analogous reduction of ketones with two small sub-
stituents, multi-substituted and hydroxy-substituted acetophenone derivatives, bulky
ketones with two large substituents, more complex cyclic ketones, sterically demand-
ing prochiral as well as racemic keto esters, and steroid-type ketones often remains a
challenge. Nonetheless, for the reduction of a range of these structurally highly
challenging molecules, which often play a role as ketone substrates for reductions
applied in drug synthesis, ADHs have already been identified as suitable catalysts by
numerous groups. The following subsection describes selected highlights of such
biocatalytic enantioselective reduction processes (focusing in particular on those
syntheses that have not been described in Section 26.3 on process development
achievements with respect to in situ cofactor recycling).
26.4.2
Reduction of Ketones with Two Small Substituents
Ketones bearing two substituents of comparable small size are challenging substrates
for asymmetric reduction processes due to a low degree of stereodifferentiation.
Nevertheless, ADHs turned out to be also suitable as enantioselective catalysts for
the reduction of such substrates. An impressive example is the enantioselective
reduction of methyl ethyl ketone (76, butanone). This reduction in the presence of the
(R)-enantioselective ADH from Lactobacillus brevis gave the desired (R)-2-butanol
with an enantioselectivity of 91% e.e. [223]. This process, which is carried out in a two-
phase system under substrate-coupled cofactor-regeneration with isopropanol, has
been conducted in a continuous mode, achieving a steady state of 71% conversion
after four residence times. The reduction of butanone has also attracted the interest of
industrial chemists, as underlined by biocatalytic enantioselective reduction pro-
cesses reported in the patent literature [224, 225]. For example, an enantioselectivity
of 98.4% e.e. and a conversion of 68% have been obtained by IEP researchers when
using an ADH from Candida parapsilosis [225]. Reduction of this challenging ketone
76 in a highly enantioselective fashion has also been reported by Nakamura et al.
when using whole-cells of Geotrichum candidum as a biocatalyst [226]. By applying a
dried
O Geotrichum candidum OH
H3 C whole-cells H3C
CH3 CH3
isopropanol
76 (S)-77
73% conversion
94% ee
Scheme 26.48
substrate-coupled cofactor-regeneration with isopropanol the corresponding (S)-

butan-2-ol, (S)-77, was formed with 73% conversion and an enantioselectivity of 94%
e.e. (Scheme 26.48). When using other enzymes lower enantioselectivities were
observed (e.g., 48% e.e. was obtained in the presence of an ADH from Thermo-
anaerobium brockii) [227], thus underlining that such a simple ketone is a quite
challenging molecule for ADHs.
A further successful example for the reduction of another small ketone is the
enantioselective biotransformation of 2-pentanone into (S)-pentan-2-ol. A highly
enantioselective process based on the use of ADH-containing whole-cells of Gluco-
nobacter oxydans and a cofactor-regeneration with a formate dehydrogenase, which
gives (S)-2-pentanol with >99% e.e., was reported by Patel and coworkers [159]. This
process, which runs on a 1500-l scale, has been summarized in more detail in
Section 26.3.3.1 (see also Chapter 29).
A further challenging ketone with small substituents for enantioselective reduction
is 1,1,1-trifluoroacetone (78). Only a few reports are available for the asymmetric
reduction of 78. The use of a (–)-B-chlorodiisopinocampheylborane as a chiral
reducing agent has been reported by the Brown group, leading to the corresponding
alcohol with 89% e.e. [228]. One drawback of this process is the need for stoichiometric
amounts of the chiral camphor-based organoboron compound. An alternative syn-
thesis of enantiomerically enriched 1,1,1-trifluoroisopropanol is based on bakers
yeast reduction but enantioselectivity did not exceed 80.3% e.e., which is significantly
below that necessary for industrial applications (preferably being >99% e.e.) [229]. A
highly enantioselective reduction of 78 has been reported by Degussa research-
ers [230]. When using a recombinant whole-cell catalyst bearing an ADH from
Rhodococcus erythropolis overexpressed together with a GDH, reduction of 78 afforded
(S)-1,1,1-trifluoroisopropanol (S)-79 with 94% conversion and an excellent enantio-
selectivity of >99% e.e. (Scheme 26.49). In contrast, a decreased enantioselectivity of
90% e.e. was observed in the presence of the (R)-enantioselective ADH of L. kefir as
recombinant E. coli
whole-cell catalyst
O (containing an ADH from OH
R. erythropolis, GDH)
F3C CH3 F3 C CH3
D-glucose,
78 phosphate buffer, (S)-79
pH ~ 6.5, 24h, r.t. 94% conversion
>99% ee
Scheme 26.49
biocatalyst. The reduction of 1,1,1-trifluoroacetophenone with several ADHs has been
also studied by Julich Chiral Solutions researchers, leading to the formation of the
resulting (R)-alcohol with 100% conversion and 98% e.e. when using an ADH from L.
brevis [112]. When using 3,3,4,4,4-pentafluorobutanone as a substrate, increased
enantioselectivities of >99% e.e. have been obtained for the (R)- and (S)-enantiomers
of the resulting alcohol in the presence of an (R)-enantioselective ADH from L. kefir
and an (S)-enantioselective ADH from Rhodococcus sp., respectively.
26.4.3
Reduction of Multisubstituted and Hydroxy-Substituted Acetophenone Derivatives
In recent years several reductions of substituted acetophenone derivatives, which

bear more than one substituent, leading to chiral alcohols of pharmaceutical interest
have been reported. In the following, selected examples thereof will be discussed as
well as the reduction of non-protected hydroxy-substituted acetophenone derivatives.
The successful use of several microorganisms in the enantioselective reduction
of 2-bromo-4-fluoro-acetophenone (80) has been reported by Bristol-Myers Squibb
researchers [231]. The resulting (S)-alcohol (S) -81 has been obtained with 99–100%
conversion and an excellent enantioselectivity of >99% e.e. when using yeasts from
the genera Candida, Hansenula, and Pichia. The use of commercially available bakers
yeast gave a conversion of 90% and also an excellent enantioselectivity of 99.9% e.e.
Applying the latter biocatalyst on a 200-l biotransformation scale gave (S)-alcohol (S)-
81 in 70% yield and with an enantiomeric excess of 99.9% e.e. (Scheme 26.50). The
downstream-processing consists of product adsorption on XAD-16 resin after the
end of the reaction, subsequent filtration and separation of the product from the resin
by extraction, and isolation of the desired (S)-alcohol (S)-81 by silica gel chroma-
tography (see also Chapter 29) [231].
Br O Br OH
baker's yeast
CH3 CH3
phosphate buffer,
F pH 7.0, glucose F
80 (S)-81
90% conversion
70% yield
99.9% ee
Scheme 26.50
The same authors also studied the enantioselective reduction of multisubstituted

acetophenone derivatives of type 82, which bear a carboxylate ester moiety [231].
Numerous strains and their suitability towards different types of esters were
screened. A suitable ADH from Pichia methanolica, which was isolated, cloned, and
overexpressed in E. coli, catalyzes the reduction of the keto methyl ester 82 to give the
corresponding (S)-hydroxy methyl ester (S)-83 in quantitative overall conversion, a
product-related conversion of 95%, and an excellent enantioselectivity of 99.9% e.e.
(Scheme 26.51) [231].
O recombinant ADH OH
from P. methanolica
CH3 (overexpressed in E. coli) CH3
CO2CH3 glucose dehydrogenase, CO2CH3
F F
glucose, pH 6.8–7.0,
82 quantitative overall (S)-83
conversion 95% product-related
conversion
99.9% ee
Scheme 26.51
The microbial enantioselective reduction of 30 ,50 -bis(trifluoromethyl)acetophe-

none (84) in the presence of a Lactobacillus kefir strain, which gives the corresponding
(R)-alcohol (R)-85 with excellent enantioselectivity (>99% e.e.), has been reported by
Rhodia-CRTL researchers (Scheme 26.52) [232]. Quantitative conversion has been
also achieved, albeit at a low substrate loading of 1.2 g l1. In contrast, higher
substrate loadings led to decreased conversions. The (R)-alcohol (R)-85 is of
pharmaceutical interest since it is a substructure of the tachykinin NK1 receptor
antagonist L-754030, which is a potent antidepressant.
O OH
F3C F3 C
CH3 L. kefir whole-cells CH3
isopropanol,
phosphate buffer,
CF3 CF3
pH 7, 30°C, 16h
84 (R)-85
(1.2 g/l) quantitative conversion
>99% ee
Scheme 26.52
The (S)-enantioselective reduction of ketone 84 to yield (S)-3,5-bistrifluoromethyl-

phenyl-ethanol ((S)-85) has also been reported by Merck researchers [233]. In the
presence of the isolated ADH from Rhodococcus erythropolis reduction of 84 at a
substrate input of 100 g l1 proceeds with 96% conversion and gave the desired (S)-
alcohol in 93% yield and with an excellent enantiomeric excess of 99% e.e.
(Scheme 26.53). A GDH was used as enzyme for cofactor regeneration in this
O OH
F3C ADH from F3C
CH3 R. erythropolis CH3
GDH,
glucose, phosphate buffer,
CF3 CF3
pH 6.5, 45°C
84 (S)-85
(100 g/l) 96% conversion
93% yield
99% ee
Scheme 26.53
process. Notably, after further process development an increased substrate concen-
tration of 580 mM and an impressive space–time yield of 260 g l1 d1 have been
achieved [233].
The biocatalytic enantioselective reduction of 2,30 -dichloro-40 -fluoroacetophenone
has been reported by Bristol-Myers Squibb researchers [189, 190]. This process,
which has been described in Section 26.3.5.2 as an example for a whole-cell catalyzed
reduction based on a G-6-PDH-coupled cofactor-regeneration, gave the desired
product (S)-2-chloro-1-(3-chloro-4-fluorophenyl)ethanol, which serves as an inter-
mediate for an IGF-1 receptor antagonist, in 89% yield and with >99% e.e.
(Scheme 26.40) [189, 190].
Further interesting acetophenone-derived substrate types for enzymatic reductions
are hydroxy-substituted acetophenone derivatives. Their biocatalytic reduction with-
out the need to protect the hydroxy moiety prior to the reduction of the C¼O double
bond appears to be superior in comparison to a typical chemocatalytic synthesis
strategy consisting of the three steps (i) protection of the hydroxy group, (ii) asym-
metric reduction of the ketone, and (iii) subsequent removal of the protecting
group [234]. However, although biocatalytic reduction is recognized as a highly
efficient approach towards enantiomerically pure alcohols, surprisingly studies on
the enzymatic enantioselective reduction of unprotected hydroxyacetophenones to
afford hydroxy-substituted 1-phenylethan-1-ols, which represent substructures of
various drugs [235], are rare so far. The microbial enantioselective reduction of para-
hydroxyacetophenone has been reported by Kumaraswamy and a coworker, obtaining
the corresponding (S)-alcohol with 23% yield and an enantioselectivity of 72% e.e. in
the presence of whole-cells from Phaseolus aureus L as a biocatalyst [236]. When
starting from the same substrate, an increased enantioselectivity of 93% e.e. and yield
(60%) have been obtained by Rao et al. when using Pisum sativa whole-cells [237].
A further improved, high enantioselectivity of at least >95% e.e. has been achieved by
Gr€oger, Hummel, and coworkers for the reduction of 2-, 3-, and 4-hydroxyacetophe-
none (86a–c) by means of (R)-enantioselective ADHs from L. kefir and L. brevis and an
(S)-enantioselective ADH from Rhodococcus sp. (Scheme 26.54). Whereas reduction
of ortho-hydroxyacetophenone (86a) only gave low to medium conversions, direct
reduction of meta- and para-hydroxyacetophenone (86b,c) proceeds efficiently without
the need to protect the hydroxy moiety. The resulting products 1-(3-hydroxyphenyl)
ethanol (87b) and 1-(4-hydroxyphenyl)ethanol (87c) were formed with high conversion
(up to >95%) and excellent enantioselectivity (up to >99% e.e.) [238].
Very recently BASF researchers have reported the enantioselective reduction of an
analogous a-halogenated ketone, namely, meta-hydroxy-a-chloroacetophenone, in the
presence of an ADH [239]. The resulting (R)-alcohol, which is formed with enantio-
selectivities of >98% e.e., is a versatile intermediate for producing phenylephrine.
26.4.4
Reduction of Bulky Ketones with Two Large Substituents
Despite many efficient biocatalytic reductions, which in part also run on an industrial
scale, the enzymatic reduction of so-called bulky ketones bearing two sterically
O (S)- or (R)-ADH, OH OH
NAD(P)+
CH3 CH3 or CH3
HO phosphate buffer, pH 7, HO HO
i-PrOH or GDH, glucose
86 (S)-87 (R)-87
Selected examples
OH OH OH OH
HO
CH3 CH3 CH3
HO
(R)-87a (R)-87b (R)-87c
5% conversion >95% conversion >95% conversion
>98% ee >99% ee >95% ee
OH OH OH OH
HO
CH3 CH3 CH3
HO
(S)-87a (S)-87b (S)-87c
>99% ee >99% ee >99% ee
Scheme 26.54
demanding substituents at the carbonyl moiety remained a major challenge. This is

because many ADHs accept a broad range of ketones bearing one bulky and one
small group as substituents (in particular aceto-type ketones like acetophenone and
2-alkanones) but do not tolerate ketones with bulky substituents on both sides. For
example, acetophenone and a-halogenated derivatives thereof represent excellent
substrates (see also the many examples in Section 26.3), whereas already propio-
phenone (bearing an ethyl instead of a methyl group as substituent attached to the
carbon of the C¼O moiety) has been regarded as a difficult substrate for many ADHs.
Other bulky ketones also typically turned out not to be reduced with many standard
ADHs. However, several screening efforts revealed that there are also ADHs that are
also suitable to reduce such bulky ketones as, for example, propiophenone,
efficiently.
A highly enantioselective biocatalytic reduction of propiophenone and a range of
other bulky ketones has been reported by the Kroutil group recently using a novel
recombinant ADH from Ralstonia sp., which was overexpressed in E. coli [240]. In the
presence of this enzyme the reduction of propiophenone, butyrophenone, and
related homolog derivatives furnished the corresponding (S)-alcohols with excellent
enantioselectivities of >99% e.e. The same group also reported reduction of ketones
with two sterically demanding substituents in the presence of wild-type strains from
Ralstonia sp. and Sphingobium yanoikuyae as biocatalysts [135].
A further biocatalyst suitable for the reduction of propiophenone is an ADH from
Rhodotorula [241]. In the presence of wild-type cells of Rhodotorula sp., the Xu group
obtained the desired (S)-alcohol with 100% conversion, 62% yield, and with an
excellent enantiomeric excess of >99% e.e. The Hua group reported the enantio-
selective reduction of a range of aryl alkyl ketones in the presence of an isolated ADH
from Sporobolomyces salmonicolor [99]. Aryl alkyl ketones bearing a sterically bulky
alkyl group such as isopropyl and tert-butyl have been reduced with a high enantio-
selectivity of 98% e.e., too.
The enantioselective biocatalytic reduction of a broad range of prochiral diaryl
ketones has been reported by Merck researchers [242]. As a biocatalyst an isolated
ADH has been used in combination with a GDH for cofactor regeneration. Starting
from prochiral substituted benzophenones as well as benzoylpyridines the desired
diarylmethanols were afforded with high yields of >90% and impressive enantios-
electivities of up to >99% e.e. Excellent synthetic results were obtained independent
of the substitution pattern, thus allowing the use of ortho-, meta-, and para-substituted
diaryl ketones in the presence of both electron-donating and -withdrawing substi-
tuents. This is in contrast to most conventional chemocatalytic reductions such as
the CBS reduction, which require an ortho-substituent for high selectivity.
Scheme 26.55 shows some selected examples of this efficient reduction method.
O (S)- or (R)-ADH OH
∗
Ar1 Ar2 Ar1 Ar2

88 89
NADPH NADP+
glucose gluconolactone
GDH
Selected examples
Cl OH OH NH2 OH
O2N
(S)-89a (S)-89b (R)-89c

95% ee >99% ee 91% ee
OH OH
Cl
N
(S)-89d (S)-89e
95% yield 98% yield
>99% ee >99% ee
Scheme 26.55
A further impressive biocatalytic reduction process for bulky ketones has been
developed at Codexis for the manufacture of the (S)-alcohol (S)-91, which serves as a
key intermediate in the synthesis of Montelukast [243, 244]. Montelukast is the active
pharmaceutical ingredient of Mercks drug Singulair, developed for the treatment of
asthma. The chiral key intermediate (S)-91 is obtained via reduction from the
corresponding ketone 90. For the reduction of this sterically demanding ketone 90
b-chlorodiisopinocampheylborane (()-DIP-chloride) was found to be a suitable
reducing agent to obtain the desired hydroxy ester (S)-91 [244, 245]. An alternative
enzymatic process based on the use of an improved ketoreductase mutant for this
reduction has been developed by Codexis [243, 244]. This economically and ecolog-
ically attractive process is characterized by high conversion of >99%, an excellent
enantiomeric excess of >99.9% e.e. (Scheme 26.56), and a simpler manufacturing
O OCH3
O
Cl N
90
KRED
cofactor
i-PrOH, water, toluene
45°C
O OCH3
OH
Cl N
(S)-91
99.3% conversion
>99.9% ee
COO-Na+
OH
S
Cl N
Montelukast
(Singulair)
Scheme 26.56
process than the above-mentioned chemical benchmark process [243, 244]. The
enzyme-catalyzed process is carried out at 45 C with a high substrate loading of 100 g
l1 in a mixture of isopropanol, water, and toluene (which was used as a cosolvent for
enhanced substrate solubility) [243, 244]. The downstream processing consists of
isolation of the product (S)-91 by filtration. The process has been scaled to 100 þ kg
batches at Arch Pharmlabs, thus underlining the technical feasibility of this biocat-
alytic reduction process (see also Chapter 29) [244].
A process for the enantioselective reduction of the sterically demanding
a,b-unsaturated ketone 92 under formation of the desired (R)-allylic alcohol (R)-
93 in >90% yield and with an enantiomeric excess of >95% e.e. was reported by
Merck researchers (Scheme 26.57) [246]. As a biocatalyst whole-cells of Candida
chilensis were used.
CH3 CH3
N N N N
whole-cells of
H Candida H
N N N N
chilensis
O glucose OH
92 (R)-93
>90% yield
>95% ee
Scheme 26.57
The enantioselective synthesis of 2-azido-1-arylethanols by enzymatic reduction of

the corresponding ketones was reported by the Hua group [247]. Both types of
enantiomers have been obtained with 100% conversion and >99% e.e. As suitable
(S)-enantioselective catalyst, an ADH from Candida magnoliae has been identified,
whereas an ADH from Saccharomyces cerevisiae can serve as an (R)-enantioselective
catalyst for this reduction.
The enzymatic diastereoselective reduction of the structurally demanding and
stereochemically challenging keto-salinosporamide 94, bearing four stereogenic
centers as well as a relatively labile lactone moiety, has been described by researchers
from Nereus Pharmaceuticals (Scheme 26.58) [248]. When using an ADH in
combination with a GDH for cofactor recycling, the desired ()-salinosporamide
A (95, NPI-0052), which is a highly potent 20S proteasome inhibitor, has been
obtained with excellent diastereoselectivity. Formation of the undesired opposite
diastereomer has not been observed and decomposition of the product turned out to
be only 2–5%. After process optimization, conversion of 95% has been achieved as
well as a yield of 90% for the product 95 after work-up.
Synthesis of the pharmaceutically interesting a-hydroxy-b-lactam (3R,4R)-97
having a functionalized side-chain in the 4-position by means of a diastereoselective
bakers yeast reduction of the corresponding 3-oxo-b-lactam 96 has been developed by
O OH
H O H O
N ADH N
O O
O GDH, O
CH3 glucose, pH 6.9 CH3
32–39°C, NAD+
Cl Cl
94 95
90% yield
Scheme 26.58
the Kayser group [249]. The desired trans-(3R,4R)-stereoisomer (3R,4R)-97 has been
obtained in 90% yield and with an excellent enantioselectivity of >99% e.e.
(Scheme 26.59).
H3C CH3 H3C CH3

O OBn baker´s yeast HO OBn
N N
O PMP O PMP
(4R)-96 (3R,4R)-97
90% yield
>99% ee
Scheme 26.59
26.4.5
Reduction of More Complex Cyclic Ketones
The reduction of cyclic ketones is still challenging although numerous examples have
been developed. However, enzymes often show unusual behavior in the reduction
of cyclic ketones, which makes a prediction difficult to some extent. An interesting
study in this direction by Faber and Kroutil investigated different types of aromatic
cyclic ketones. While 1-tetranone and 1-indanone were not reduced by the ADH from
Rhodococcus ruber, 2-tetralone was reduced although enantioselectivity was moderate
and activity was strongly decreased compared to substrates bearing an aceto
subunit [126].
The enantioselective reduction of 6-bromo-b-tetralone (98) is of interest due to the
use of the resulting (S)-bromo-b-tetralol [(S)-99] as a pharmaceutical intermediate in
the synthesis of the antiarrhythmia drug candidate MK-0499 [250]. Jointly with Merck
researchers the Lye group reported that in the presence of an ADH from Rhodococcus
erythropolis in combination with a GDH-catalyzed cofactor regeneration the reaction
proceeds with high enantioselectivity, leading to the desired (S)-alcohol (S)-99 in
88% overall yield and >99% e.e. (Scheme 26.60). A beneficial effect of both water
ADH from
O OH
Rhodococcus erythropolis
Br GDH, Br
glucose, buffer, pH ~ 6.8,
98 ionic liquid ([BMP][NTf2]), (S)-99
(50 g/l) 30°C, NAD+ quantitative conversion
88% yield
>99% ee
Scheme 26.60
miscible and immiscible ionic liquid cosolvents, which turned out to be more
favorable than several organic solvents in terms of enzyme stability, was also found.
The use of an isolated ADH and GDH for a cyclic ketone reduction at Merck has
been reported for the production of (R)-4,4-dimethoxytetrahydro-2H-pyran-3-ol [(R)-
101], which is an intermediate in the synthesis of a chemokine receptor inhibi-
tor [251]. This biocatalytic reduction with a GDH-coupled in situ cofactor regener-
ation of NADPH was carried out at a high substrate loading of 100 g l1, leading to the
desired (R)-alcohol (R)-101 in 96–98% yield and with excellent enantioselectivity of
>99% e.e. (Scheme 26.61). This process has been applied on an 80-kg pilot-plant
scale to produce (R)-101 (see also Chapter 29) [251].
CH3 CH3 CH3 CH3

O O O O
O ADH KRED101 HO
O O
100 (R)-101
(substrate input: 96-98% yield
100 g/l) >99% ee
NADPH NADP+
D-glucose D-gluconic acid

GDH (GDH 101),
H2O, pH 6.5, 35°C
Scheme 26.61
An elegant example of the integration of an (diastereo- and enantioselective)

enzymatic reduction of a cyclic ketone into a dynamic kinetic resolution process has
been reported by Merck researchers [252]. When starting from the racemic
a,b-unsaturated cyclic ketone 102, ADH-catalyzed carbonyl reduction and in situ
racemization at pH 6.5 afforded the desired allylic alcohol 103 in 94% yield and high
diastereo- and enantioselectivities of 99% d.e. (favoring the cis-diastereomer) and
95% e.e. (Scheme 26.62) [252].
O OH
ADHs
(KRED108, KRED104)
CO2CH3 buffer, pH 6.5, CO2CH3

CO2CH3 i-PrOH, NADP+, CO2CH3
35°C, 12h,
rac-102 ~20% (v/v) DMSO 103
94% yield
99% de (cis)
95% ee
Scheme 26.62
Application of an ADH-catalyzed reduction for the resolution of the pharmaceu-

tically important bridged ketone rac-104 has also been reported by Merck research-
ers [253]. The (6S,9R)-enantiomer of this ketone ((6S,9R)-104), which has been
shown to be an important intermediate in the synthesis of pharmaceutically active
substances, was obtained in 44% yield and with excellent enantiomeric excess of
>99% e.e. (Scheme 26.63). As a reaction medium for this resolution process, which
was carried out on a 1-kg scale, a combination of cyclodextrin and DMSO as cosolvent
in the aqueous reaction system turned out to be beneficial for increased substrate
solubility.
OH
HO
OH 105
ADH KRED 101
O +
GDH, OH
glucose, buffer, pH 7.0
rac-104 O
(10 g/l) cyclodextrin (70 g/l)
DMSO (10% (v/v)),
NADP+, 10°C (6S,9R)-104
44% yield
>99% ee
Scheme 26.63
26.4.6
Reduction of Steroid Ketones
Among reduction of cyclic ketones, steroids have an exceptional position due to their
fused ring-system (sterane core) and various functional groups. Notably, (regiose-
lective and diastereoselective) biocatalytic reduction towards the formation of
complex steroid type-alcohols turned to be a promising synthetic approach
although the solubility of sterols and steroids in water is low. The latter issue has
also been addressed in process development work, and selected examples thereof are
described below.
One of the first examples in the literature of a biochemical reduction of C¼O
double bonds in steroids was described by Mamoli and Vercellone in 1937, namely,
the reduction of D5-dehydro-androsterone to give D5-androstendiol in a fermentation
process with yeast [254]. In further studies the same authors found that this method
could also be used for the selective reduction of D5-androstendione to isoandros-
tandiol and D4-testosterone [255] and of D4-androstendione (106) to D4-testosterone,
107 [256]. The latter fermentation-like reduction process, which is based on the use
of yeast and glucose, leading in a stereospecific reaction course to the formation
of D4-testosterone, 107, has been also described in a more recent review by Schering
researchers (Scheme 26.64) [257].
O OH
Saccharomyces sp.
glucose
O O
106 107
80% yield
Scheme 26.64
Furthermore, Vercellone et al. described the reduction of the carbonyl group of

steroids in the pregnane series using a yeast [258]. The reduction of pregnan-3,11,20-
trione (108) to pregnan-3a-ol-11,20-dione (109) proceeds within 6 days with 60%
yield (Scheme 26.65). The steroid pregnan-3a-ol-11,20-dione is an important inter-
mediate in the synthesis of cortisone [259].
O O
CH3 CH3
O O
H yeast H
H H 6d H H
O HO
108 109
60% yield
Scheme 26.65
A remarkably short reaction sequence for the total synthesis of the steroid
hormone D-estradiol, 114, has been developed by Torgov and coworkers
(Scheme 26.66) [260]. The total synthesis is based on an enantioselective reduction
of intermediate 110 using Saccharomyces cerevisiae as a key step in this sequence. The
resulting alcohol 111 can be subsequently transformed via intermediates 112 and 113
into the biologically active steroid hormone D-estradiol, 114 [260].
Process development through reaction media engineering, especially for improv-
ing the low solubility of steroid-type molecules in aqueous media, has been addressed
O OH
Saccharomyces cerevisiae
O O
H3CO H3CO
110 111
OAc OAc OH
O H H
H3CO H3CO HO
112 113 114
Scheme 26.66
by several groups. Hirakawa et al. employed a reversed micellar system by means of

the use of sodium dioctyl sulfosuccinate (AOT) and isooctane as additives for the
enzymatic reduction of androstandione, 115 [261]. This reaction medium offers
advantages such as a high concentration for a hydrophobic substrate and larger
interfacial areas for better mass transfer. In this reaction medium androstandione
(115) was reduced in the presence of a 3a-hydroxysteroid dehydrogenase (HSDH)
from Pseudomonas testosteroni, producing androsterone (116) continuously for up to
24 h (Scheme 26.67). Cofactor-regeneration was carried out via oxidation of ethanol
by a yeast ADH.
A further reaction media for the application of a 3a-hydroxysteroid dehydrogenase
consists of a biphasic aqueous–organic system with an ionic liquid as cosolvent.
Honda et al. used 1-butyl-3-methylimidazolium (L)-lactate [bmim][lactate] as
O HSDH from O
Pseudomonas testosteroni
Tris-HCl buffer, pH 9
O HO
115 116
+
100% conversion
NADH NAD
acetaldehyde ethanol
ADH from
baker´s yeast
Scheme 26.67
cosolvent (5 vol.%) to enhance the activity of a HSDH for the reduction of andros-
tandione (115) in a biphasic reaction medium consisting of a buffer as the aqueous
phase (pH 7.6) and octane [262]. In the presence of this steroid dehydrogenase and a
formate dehydrogenase for regeneration of NADH, complete conversion of andros-
tandione (115) and a twofold increase in production rate of androsterone (116) was
obtained in this reaction medium (Scheme 26.68).
HSDH from
O Pseudomonas testosteroni O
pH 7.6 / octane,
5% (v/v) [bmim][lactate]
25°C, 8h
O HO
115 116
+
100% conversion
NADH NAD
CO2 formate
FDH from
Candida boidinii
Scheme 26.68
26.4.7
Reduction of Keto Esters
Among the most prominent representatives of b-ketoesters used as substrate for

enzymatic reduction are certainly ethyl or methyl acetoacetate as well as ethyl
4-chloro-3-oxobutanoate. The importance of the latter ketone is due to the use of
the resulting b-hydroxyester as intermediate in the synthesis of the statin side chain.
Since these reductions have also been studied extensively in process development,
the corresponding synthetic biotransformations have been described in Sec-
tions 26.3.2 and 26.3.4.
An efficient biocatalytic keto ester reduction for the synthesis of methyl (R)-o-
chloromandelate, which serves as an intermediate for the drug clopidogrel, has been
reported by Ema and Sakai [263]. When using the a-keto ester 117 as a substrate in the
presence of a recombinant whole-cell catalyst, which contains an overexpressed ADH
from bakers yeast and a GDH, the biocatalytic reduction proceeds at an impressive
substrate input of 198 g l1 with 86% conversion. The desired product (R)-118 was
obtained in 82% yield and with an excellent enantioselectivity of >99%
An elegant way to synthesize diastereo- and enantiomerically pure b-hydroxy esters
bearing two stereogenic centers (in a- and b-positions) has been reported by the
Smonou group jointly with BioCatalytics researchers [264]. The concept is based on a
dynamic kinetic resolution starting from easily available racemic b-keto esters 119.
containing
Cl O ADH, Cl OH
GDH,
CO2Me NAD(P)H CO2Me
NAD(P)+,
117 D-glucose (R)-118
(198 g/l 86% conversion 82% yield
substrate input) >99% ee
Scheme 26.69
The applied ketoreductases recognize only one of the two substrate enantiomers, and
convert the accepted enantiomer in a highly diastereo- and enantioselective reduction
into the b-hydroxyesters 120. Since the diastereoselectivity of this process is typically
also excellent, the desired products of type 120 are obtained with high enantio- and
diastereomeric excess. Since the substrate is permanently racemized under the
applied reaction conditions due to the CH-acidic stereogenic a-carbon center, a
dynamic kinetic process has been realized, thus leading to formation of the desired
products of type 120 in excellent conversions (of 100% in most cases) and with
high diastereoselectivities (d.r. up to >99:1) and excellent enantioselectivities of
>99% e.e. [264]. Scheme 26.70 shows selected examples.
The extension of this bioreduction concept to a regio- and stereoselective reduction
of a-substituted diketones, to obtain the corresponding b-keto alcohols or 1,3-diols
with high diastereo- and enantioselectivities as well as excellent conversions, has also
been reported by the Smonou group jointly with BioCatalytics researchers
(Scheme 26.70) [265]. The same authors also developed a synthesis of sitophilate,
the aggregation pheromone of the granary weevil Sitophilus granarius, based on a
highly diastereo- and enantioselective enzymatic reduction of methyl 3-oxopentano-
ate as key step. The resulting b-hydroxy ester has been obtained in 90% yield and with
90% d.e. and >99% e.e. [266].
The M€ uller group used recombinant E. coli cells containing an overexpressed ADH
from Lactobacillus brevis for an impressive regio- and enantioselective reduction of
3,5-dioxocarboxylates [69, 267]. With isopropanol as reducing agent the 3,5-diketo
ester was transformed into (S)-6-chloro-5-hydroxy-3-oxohexanoate, which serves as a
valuable intermediate in the synthesis of HMG-CoA reductase inhibitors, in 72%
yield and with an excellent enantiomeric excess of >99.5% e.e. [69]. This process is
also described in Section 26.3.2. The use of a 4-substituted 3,5-dioxoester, namely,
tert-butyl 4-methyl-3,5-dioxohexanoate, has also been reported by the M€ uller
group [70, 268, 269]. Regio- and enantioselective reduction under dynamic kinetic
resolution conditions were carried out in the presence of ADHs from Lactobacillus
brevis, Rhodococcus erythropolis, and Saccharomyces cerevisiae, leading to the corre-
sponding syn-(4S,5R)-, syn-(4R,5S)-, and anti-(4S,5S)-diastereomers of tert-butyl
5-hydroxy-4-methyl-3-oxohexanoate with both high diastereo- and enantioselectivity.
The ADH-catalyzed enantio- and diastereoselective reduction of both keto moieties
in a 3,5-dioxoester has been demonstrated by the Patel group starting from ethyl
6-benzyloxy-3,5-dioxohexanoate (121) as a substrate [270, 271]. When using cell
O O O OH
ADH ∗
2 ∗
R1 R R 1 R2
R3 R3
119 120
NADPH NADP+
gluconic acid glucose

GDH
Selected examples
O OH O OH O OH
H 3C CH3 H3C CH3 H3C CH3
H 3C H 3C H3C
(3R,4S)-120a (3S,4R)-120a (3S,4S)-120a
dr >99:1 dr=90:10 dr >99:1
>99% ee >99% ee >99% ee
O OH O OH O OH
H3C
H 3C O CH3 H3C O CH3 CH3
CH3 CH3
H 3C
(2R,3S)-120b (2R,3S)-120c (4R,5S)-120d
dr >99:1 dr >99:1 dr >99:1
>99% ee >99% ee >99% ee
Scheme 26.70
extracts of Acinetobacter calcoaceticus in combination with a GDH and glucose, the

desired product ethyl (3R,5S)-6-benzyloxy-3,5-dihydroxyhexanoate ((3R,5S)-122) was
formed with 92% conversion, a diastereoselectivity of 95% d.e., and an enantio-
selectivity of >99% e.e. (Scheme 26.71). After product isolation, (3R,5S)-122 was
obtained in 72% yield and with an enantiomeric excess of 99.5% e.e.
b-Keto nitriles can be considered as derivatives of b-keto esters, in which the
ester group is replaced by the cyano group (which can be transformed back into the
O O O ADH from
OH OH O
Acinetobacter calcoaceticus
O O
O Me O Me
GDH,
121 (3R,5S)-122
glucose, NAD+, 72% yield
92% conversion 99.5% ee
Scheme 26.71
ester group by hydrolysis and esterification of the resulting carboxylic acid).
The enantioselective reduction of 3-oxo-3-phenylpropanenitrile in the presence of
different types of recombinant ADHs from bakers yeast overexpressed in E. coli
has been reported by Feske and coworkers [272]. Notably, enzymes have been found
for the synthesis of both enantiomers of the resulting alcohol, which is of pharma-
ceutical relevance since it serves (dependent on the absolute configuration) as an
intermediate in the synthesis of both enantiomers of the drug fluoxetine, as well as
the enantiomerically pure drugs atomoxetine and nisoxetine. Enantioselectivities
of up to 97% e.e. for the (R)-enantiomer, and 99% e.e. for the (S)-enantiomer have
been achieved. Furthermore, enantioselective reduction of a range of aromatic
b-ketonitriles also has been reported by the Hua group, achieving high yields
of up to 92% and enantioselectivities of 97–99% e.e. for the resulting b-hydroxy
nitriles [273].
26.4.8
Reduction of Aldehydes
Although most reported enzymatic reductions of C¼O double bonds are related to
the (preferably highly enantioselective) transformation of ketones into secondary
alcohols, biocatalytic reduction of aldehydes has also attracted the interest of
academic and industrial researchers. Selected recent examples are summarized in
more detail in the following.
As a representative example, the use of alcohol dehydrogenases as catalysts for
reduction of aldehydes plays a role in the reduction of achiral aldehydes to give
primary alcohols. Such primary alcohols are valuable compounds for, for example,
the flavor and fragrance industry [274]. Advantages of such an enzymatic approach to
primary alcohols are the simple operational set up of the reaction as well as the high
(chemo-)selectivity of the reduction processes (when other reducible bonds are
present in the molecule such as C¼C double bonds).
The enzymatic reduction of green note aldehydes has been studied by Faucon-
nier and coworkers by means of different yeasts [275]. In the presence of Pichia
anomala as the preferred biocatalyst, (Z)-3-hexenal was converted into the so-called
leaf alcohol (Z)-3-hexenol with >90% conversion.
The synthesis of 2-phenylethanol via reduction using yeast strains as biocatalysts in
a molasses-based reaction medium has been reported by Schrader and cowor-
kers [276]. When carrying out the reaction under in situ product removal with oleyl
alcohol as a second phase, 3 g l1 of 2-phenylethanol was obtained. A strain from
Kluyveromyces marxianus turned out to be the most productive among the fourteen
yeast strains tested.
A biocatalytic aldehyde reduction process based on the use of a recombinant whole-
cell catalyst, which proceeds with high conversion and additionally run at a high
substrate loading, has been reported jointly by researchers from Degussa AG and
Cargill Flavor Systems [277]. As a substrate cinnamyl aldehyde (123) was used, and
cofactor regeneration was carried out with D-glucose as a cosubstrate in the presence
of an (overexpressed) GDH in the whole-cell catalyst. The reduction ran at a substrate
recombinant E. coli
whole-cell catalyst
(containing:
ADH from L. kefir,
GDH,
O NADP+) OH
D-glucose,
123 water, 124
(substrate input: pH 6.5–7.0, 24h, r.t. 98% conversion
166 g/l) 77% yield
Scheme 26.72
input of 166 g l1 of 123 and led to a conversion of 98% (Scheme 26.72). After work-
up the desired cinnamyl alcohol (124) was obtained in 77% yield.
Interestingly, reduction of aldehyde functionalities is also used within the field of
enantioselective synthesis of chiral compounds. Although (in contrast to ketone
reductions) a stereogenic center is not formed when reducing an aldehyde, racemic
aldehydes might serve as a substrate in an enzymatic resolution process via reduction
of the aldehyde moiety of one enantiomer. Such a resolution has been developed
recently for the synthesis of the versatile building block L-glyceraldehyde by the
Hummel and Liese groups jointly with researchers from Evocatal as well as Sigma-
Aldrich Chemie [102]. As a substrate racemic glyceraldehyde (rac-125) was used; the
D-enantiomer has been reduced selectively by an ADH from Gluconobacter oxydans to
afford (achiral) glycerol, 126. This resolution led to a 50% conversion and gave the
desired L-glyceraldehyde, L-125, as the remaining enantiomer with an excellent
enantiomeric excess of >99% e.e. (Scheme 26.73). The in situ cofactor regeneration
was carried out with a GDH and glucose as a cosubstrate.
Further extensions of this type of biocatalytic resolution via aldehyde reduction
towards dynamic kinetic resolution processes have been successfully developed. The
concept of such a dynamic kinetic resolution is based on the use of enolizable
aldehydes as substrates that have a stereogenic center in the a-position (e.g. rac-127).
This kind of a dynamic kinetic resolution has been recently reported by the Giacomini
group for the synthesis of (S)-2-phenylpropanol [(S)-128] (Scheme 26.74) [278].
Within such a dynamic kinetic resolution process, starting from racemic 2-phenyl-
propanal in a aqueous buffer with acetonitrile as a cosolvent (16 vol.%), reduction
ADH from
OH G. oxydans OH OH
HO O HO O + HO OH
rac
rac-125 (S)-125 126
>99% ee 50% conversion
NADPH NADP+
D-glucose D-gluconolactone
GDH
Scheme 26.73
CH3 ADH from CH3
O horse liver OH
rac
rac-127 (S)-128
NADH NAD+ 90% yield
88% ee
ethanol acetaldehyde
ADH from
horse liver
Scheme 26.74
proceeds efficiently in the presence of the ADH from horse liver to give (S)-128 in
90% yield and with 88% e.e. This type of dynamic kinetic resolution was also
successfully applied for the synthesis of (S)-2-(4-isobutylphenyl)propanol, which was
obtained in 93% yield and with excellent enantiomeric excess of >99% e.e. This
compound serves as an intermediate in the synthesis of (S)-ibuprofen.
An impressive extension of this technology towards a highly efficient dynamic
kinetic resolution of a broad spectrum of 2-arylpropanals (rac-129) has been reported
very recently by the Berkowitz group, using an alcohol dehydrogenase from the
archael hyperthermophile Sulfolobus sulfataricus (Scheme 26.75) [279]. Under
CH3 CH3 ADH from CH3

O fast O S. sulfataricus
OH
Ar Ar Ar
buffer, pH 9,
ethanol (5%(v/v)), (S)-130
NADH, 80 °C
rac-129
(Ar=aryl)
Selected examples
CH3 CH3 CH3
OH OH OH
H3CO
(S)-130a (S)-130b (S)-130c
98% ee 94% ee 98% ee
CH3 CH3 O CH3

OH O OH OH
CH3
H3C
(S)-130d (S)-130e (S)-130f
99% ee 95% ee 95% ee
Scheme 26.75
References j1101
optimized reaction conditions, the desired (S)-2-arylpropanols [(S)-130] were
obtained in yields of up to 99% and with enantioselectivities of up to 99% e.e.
Selected examples are shown in Scheme 26.75. The (S)-2-arylpropanols (S)-130c,
(S)-130d, (S)-130e, and (S)-130f are intermediates for the synthesis of the nonste-
roidal anti-inflammatory drugs naproxen, ibuprofen, fenoprofen, and ketoprofen,
respectively. An interesting aspect from the point of view of downstream-processing
is the low amount of 5% ethanol, which serves as cosolvent and cosubstrate for
cofactor regeneration. Thus, after carrying out the reduction at 80 C until comple-
tion of the reaction, cooling the reaction mixture led to precipitation of the desired
product, which then can be easily isolated by filtration. This elegant work-up certainly
represents a further main advantage of this reduction technology. In addition, the
enzyme has been recycled successfully over five reaction cycles.
26.5
Summary and Outlook
In summary, numerous efficient methodologies for the asymmetric biocatalytic

reduction of ketones and aldehydes have been developed. These methods consist of
syntheses based on the use of isolated enzymes as well as whole-cell catalysts. In this
connection, the use of recombinant enzymes and whole-cell catalysts turned out to be
particularly attractive for synthetic purposes. Enzymatic reductions of prochiral
ketones into optically active alcohols in general proceed with excellent enantioselec-
tivity, typically exceeding 99% e.e. Furthermore, enzymatic reductions of ketones and
aldehydes have been proven to be highly efficient also with respect to conversion
(which often exceeds 95%) and substrate loading (which exceeds >100 g l1 in many
reactions), thus enabling the development of biocatalytic synthetic processes show-
ing economically and ecologically favorable data. Notably, already a broad range of
applications of biocatalytic reductions of C¼O double bonds have been realized also
on an industrial scale. Thus, this technology today already represents a highly
competitive and attractive industrial process technology for the enantioselective
manufacture of optically active alcohols. Among remaining challenges for the future
are, for example, further extension of the substrate range in general, and in particular
the development of recombinant enzymes suitable for the reduction of those bulky
ketones that have proven to be difficult to reduce up to now.
References
1 Noyori, R. (2002) Angew. Chem., 114, 3 (a) Reviews: Patel, R.N. (2008) Coord.
2108–2123; (2002) Angew. Chem. Int. Ed., Chem. Rev., 252, 659–701;(b)De
41, 2008–2022. Wildeman, S.M.A., Sonke, T.,
2 Corey, E.J. and Helal, C.J. (1998) Schoemaker, H.E., and May, O. (2007)
Angew. Chem., 110, 2092–2118; Acc. Chem. Res., 40, 1260–1266;
(1998) Angew. Chem. Int. Ed., 37, (c)Moore, J.C., Pollard, D., Kosjek, B.,
1986–2012. and Devine, P.N. (2007) Acc. Chem. Res.,
40, 1412–1419;(d)Buchholz, S. and 17 Jez, J.M. and Penning, T.M. (2001) Chem.-
Gr€oger, H. (2006) in Biocatalysis in the Biol. Interact., 130, 499–525.
Pharmaceutical and Biotechnology 18 Wermuth, B., Omar, A., Forster, A.,
Industries (ed. R.N. Patel), CRC Press, Di Francesco, C., Wolf, M.,
New York, ch. 32, pp. 757–790. Wartburg, J.-P., Bullock, B., and
4 Liese, A., Seelbach, K., and Wandrey, C. Gabbay, K.H. (1987) Prog. Clin. Biol. Res.,
(eds) (2006) Industrial Biotransformations, 232, 297–307.
2nd edn, Wiley-VCH Verlag GmbH, 19 Amore, R., Kotter, P., Kuster, C.,
Weinheim. Ciriacy, M., and Hollenberg, C.P. (1991)
5 For a very comprehensive summary of Gene, 109, 89–97.
biocatalysis in organic synthesis in 20 Billard, P., Menart, S., Fleer, R., and
general see Drauz, K. and Bolotin-Fukuhara, M. (1995) Gene, 162,
Waldmann, H. (eds) (2002) in Enzymes 93–97.
in Organic Synthesis, vols 1–3, 21 Yokoyama, S., Suzuki, T., Kawai, K.,
2nd edn, Wiley-VCH Verlag GmbH, Horitsu, H., and Takamizawa, K. (1995)
Weinheim. J. Ferment. Bioeng., 79, 217–223.
6 Adams, M.J., Ford, G.C., Koekoek, R., 22 Kim, S.-T., Huh, W.-K., Lee, B.-H., and
Lentz, P.J., McPherson, A.J., Kang, S.-O. (1998) Biochim. Biophys. Acta,
Rossmann, M.G., Smiley, I.E., 1429, 29–39.
Schevitz, R.W., and Wonacott, A.J. (1970) 23 Goffeau, A., Barrell, B.G., Bussey, H.,
Nature, 227, 1098–1103. Davis, R.W., Dujon, B., Feldmann, H.,
7 Rossmann, M., Liljas, A., Br€anden, C.I., Galibert, F., Hoheisel, J.D., Jacq, C.,
and Banaszak, L. (1975) The Enzymes, Johnston, M., Louis, E.J., Mewes, H.W.,
3rd edn, vol. 2 (ed. P.D. Boyer), Murakami, Y., Philippsen, P., Tettelin,
Academic Press, New York, H., and Oliver, S.G. (1996) Science, 274,
pp. 61–102. 546–567.
8 Rossmann, M.G., Moras, D., and 24 Habrych, M., Rodriguez, S., and
Olsen, K.W. (1974) Nature, 250, 194–199. Stewart, J.D. (2002) Biotechnol. Prog., 18,
9 Nordling, E., Jornvall, H., and 257–261.
Persson, B. (2002) Eur. J. Biochem., 269, 25 Persson, B., Hedlund, J., and Jornvall, H.
4267–4276. (2008) Cell. Mol. Life Sci., 65, 3879–3894.
10 Persson, B., Jeffery, J., and Jornvall, 26 J€ornvall, H. (2008) Cell. Mol. Life Sci., 65,
H. (1991) Biochem. Biophys. Res. 3875–3878.
Commun., 177, 218–223. 27 Hjelmqvist, L., Norin, A., El-Ahmad, M.,
11 Persson, B., Krook, M., and Griffiths, W., and Jornvall, H. (2003) Cell.
Jornvall, H. (1995) Adv. Exp. Med. Biol., Mol. Life Sci., 60, 2009–2016.
372, 383–395. 28 Bairoch, A., Apweiler, R., Wu, C.H.,
12 Klimacek, M. and Nidetzky, B. (2002) Barker, W.C., Boeckmann, B., Ferro, S.,
Biochem. J., 367, 13–18. Gasteiger, E., Huang, H.Z., Lopez, R.,
13 Bohren, K., Bullock, B., Wermuth, B., and Magrane, M., Martin, M.J., Natale, D.A.,
Gabbay, K. (1989) J. Biol. Chem., 264, ODonovan, C., Redaschi, N., and
9547–9551. Yeh, L.S.L. (2005) Nucleic Acids Res., 33,
14 Hoog, S.S., Pawlowski, J.E., Alzari, P.M., D154–D159.
Penning, T.M., and Lewis, M. (1994) 29 J€ornvall, H. (1970) Eur. J. Biochem., 16,
Proc. Natl. Acad. Sci. USA, 91, 25–40.
2517–2521. 30 Jeffery, J., Cederlund, E., and Jornvall, H.
15 Rondeau, J.M., Tete-Favier, F., (1984) Eur. J. Biochem., 140, 7–16.
Podjarny, A., Reymann, J.M., Barth, P., 31 Leskovac, V., Trivic, S., and Pericin, D.
Biellman, J.F., and Moras, D. (1992) (2002) FEMS Yeast Res., 2, 481–494.
Nature, 355, 469–472. 32 Green, D.W., Sun, H.W., and Plapp, B.V.
16 Jez, J.M., Flynn, T.G., and (1993) J. Biol. Chem., 268, 7792–7798.
Penning, T.M. (1997) Biochem. 33 Pauly, T.A., Ekstrom, J.L., Beebe, D.A.,
Pharmacol., 54, 639–647. Chrunyk, B., Cunningham, D.,
References j1103
Griffor, M., Kamath, A., Lee, S.E., Ladenstein, R., and J€ornvall, H. (1997)
Madura, R., and McGuire, D. (2003) Biochemistry, 36, 34–40.
Structure, 11, 1071–1085. 48 Thoden, J.B., Wohlers, T.M.,
34 Sibout, R., Eudes, A., Mouille, G., Fridovich-Keil, J.L., and Holden, H.M.
Pollet, B., Lapierre, C., Jouanin, L., and (2000) Biochemistry, 39, 5691–5701.
Seguin, A. (2005) Plant Cell, 17, 49 Benach, J., Atrian, S., Gonzalez-
2059–2076. Duarte, R., and Ladenstein, R. (1999)
35 Larroy, C., Fernandez, M.R., J. Mol. Biol., 289, 335–355.
Gonzalez, E., Pares, X., and Biosca, J.A. 50 J€
ornvall, H., Danielsson, O.,
(2002) Biochem. J., 361, 163–172. Hjelmqvist, L., Persson, B., and
36 Larroy, C., Pares, X., and Biosca, J.A. Shafqat, J. (1995) Adv. Exp. Med. Biol.,
(2002) Eur. J. Biochem., 269, 5738–5745. 372, 281–294.
37 Esposito, L., Sica, F., Raia, C.A., 51 J€
ornvall, H., Persson, B., Krook, M.,
Giordano, A., Rossi, M., Mazzarella, L., Atrian, S., Gonzalezduarte, R., Jeffery, J.,
and Zagari, A. (2002) J. Mol. Biol., 318, and Ghosh, D. (1995) Biochemistry, 34,
463–477. 6003–6013.
38 Ammendola, S., Raia, C.A., Caruso, C., 52 Nakamura, S., Oda, M., Kataoka, S.,
Camardella, L., DAuria, S., De Rosa, M., Ueda, S., Uchiyama, S., Yoshida, T.,
and Rossi, M. (1992) Biochemistry, 31, Kobayashi, Y., and Ohkubo, T. (2006)
12514–12523. J. Biol.Chem., 281, 31876–31884.
39 Raia, C.A., DAuria, S., and Rossi, M. 53 Ghosh, D., Pletnev, V.Z., Zhu, D.W.,
(1994) Biocatalysis, 11, 143–150. Wawrzak, Z., and Duax, W.L. (1995)
40 Kallberg, Y., Oppermann, U., Jornvall, H., Structure, 3, 503–513.
and Persson, B. (2002) Eur. J. Biochem., 54 Filling, C., Berndt, K.D., Benach, J.,
269, 4409–4417. Knapp, S., Prozorovski, T., Nordling, E.,
41 Kavanagh, K.L., Jornvall, H., Persson, B., Ladenstein, R., J€ornvall, H., and
and Oppermann, U. (2008) Cell. Mol. Life Oppermann, U. (2002) J. Biol. Chem.,
Sci., 65, 3895–3906. 277, 25677–25684.
42 Persson, B., Kallberg, Y., Bray, J.E., 55 Oppermann, U., Filling, C., Hult, M.,
Bruford, E., Dellaporta, S.L., Favia, A.D., Shafqat, N., Wu, X.Q., Lindh, M.,
Duarte, R.G., Jornvall, H., Shafqat, J., Nordling, E., Kallberg, Y.,
Kavanagh, K.L., Kedishvili, N., Persson, B., and Jornvall, H. (2003)
Kisiela, M., Maserk, E., Mindnich, R., Chem.-Biol. Interact., 143, 247–253.
Orchard, S., Penning, T.M., Thornton, 56 Fukui, T., Zong, M.H., Kawamoto, T., and
J.M., Adamski, J., and Oppermann, U. Tanaka, A. (1992) Appl. Microbiol.
(2009) Chem.-Biol. Interact., 178 94–98. Biotechnol., 38, 209–213.
43 Oppermann, U.C., Maser, E., 57 Tsuji, Y., Fukui, T., Kawamoto, T., and
Hermans, J.J., Koolman, J., and Tanaka, A. (1994) Appl. Microbiol.
Netter, K.J. (1992) J. Steroid Biochem. Mol. Biotechnol., 41, 219–224.
Biol., 43, 665–675. 58 Yamazaki, Y., Uebayasi, M., and
44 Persson, B., Krook, M., and Hosono, K. (1989) Eur. J. Biochem., 184,
Jornvall, H. (1991) Eur. J. Biochem., 200, 671–680.
537–543. 59 Jones, J.B. and Schwartz, H.M. (1981)
45 Koumanov, A., Benach, J., Atrian, S., Can. J. Chem., 59, 1574–1579.
Gonzalez-Duarte, R., Karshikoff, A., and 60 Jakovac, J.J., Goodbrand, H.B., Lok, K.P.,
Ladenstein, R. (2003) Proteins, 51, and Jones, J.B. (1982) J. Am. Chem. Soc.,
289–298. 104, 4659–4665.
46 Liu, Y., Thoden, J.B., Kim, J., Berger, E., 61 Jones, J.B. and Takemura, T. (1984) Can.
Gulick, A.M., Ruzicka, F.J., Holden, J. Chem., 62, 77–80.
H.M., and Frey, P.A. (1997) Biochemistry, 62 Jones, J.B. (1985) Enzymes in Organic
36, 10675–10684. Synthesis, Pitman, London, pp. 3–21.
47 Oppermann, U.C.T., Filling, C., 63 Keinan, E., Seth, K.K., and Lamed, R.
Berndt, K.D., Persson, B., Benach, J., (1987) Ann. N.Y. Acad. Sci., 501, 130–149.
64 Belan, A., Bolte, J., Faure, A., H€ usken, H., and Abokitse, K. (2003)
Gourcy, J.G., and Veschambre, H. (1987) Org. Lett., 5, 173–176.
J. Org. Chem., 52, 256–260. 81 Gr€oger, H., Hummel, W., Rollmann, C.,
65 Rothing, T.R., Kulbe, K.D., Chamouleau, F., H€ usken, H.,
B€uckmann, F., and Carrea, G. (1990) Werner, H., Wunderlich, C., Abokitse, K.,
Biotechnol. Lett., 12, 353–356. Drauz, K., and Buchholz, S. (2004)
66 Yang, H., Jonsson, A., Wehtje, E., Tetrahedron, 60, 633–640.
Adlercreutz, P., and Mattiasson, B. (1997) 82 Edegger, K., Gruber, C.C., Poessl, T.M.,
Biochim. Biophys. Acta, 1336, 51–58. Wallner, S.R., Lavandera, I., Faber, K.,
67 Weckbecker, A. and Hummel, W. (2006) Niehaus, F., Eck, J., Oehrlein, R.,
Biocatal. Biotrans., 24, 380–389. Hafner, A., and Kroutil, W. (2006) Chem.
68 Niefind, K., Riebel, B., M€ uller, J., Commun., 2402–2404.
Hummel, W., and Schomburg, D. (2000) 83 Ema, T., Sugiyama, Y., Fukumoto, M.,
Acta Crystallogr., 56, 1696–1698. Moriya, H., Cui, J.N., Sakai, T., and
69 Wolberg, M., Hummel, W., Wandrey, C., Utaka, M. (1998) J. Org. Chem., 63,
and M€ uller, M. (2000) Angew. Chem., 112, 4996–5000.
4476–4478; (2000) Angew. Chem. Int. Ed., 84 Ema, T., Yagasaki, H., Okita, N.,
39, 4306–4308. Nishikawa, K.T., and Sakai, T. (2005)
70 Wolberg, M., Ji, A.G., Hummel, W., and Tetrahedron: Asymmetry, 16, 1075–1078.
M€ uller, M. (2001) Synthesis, 937–942. 85 Ema, T., Yagasaki, H., Okita, N.,
71 Schubert, T., Hummel, W., Kula, M.R., Takeda, M., and Sakai, T. (2006)
and M€ uller, M. (2001) Eur. J. Org. Chem., Tetrahedron, 62, 6143–6149.
4181–4187. 86 Kaluzna, I.A., Matsuda, T., Sewell, A.K.,
72 Wolberg, M., Hummel, W., and and Stewart, J.D. (2004) J. Am. Chem.
M€ uller, M. (2001) Chem. Eur. J., 7, Soc., 126, 12827–12832.
4652–4571. 87 Lieser, J.K. (1983) Synth. Commun., 13,
73 Schubert, T., Hummel, W., and 765–767.
M€ uller, M. (2002) Angew. Chem., 114, 88 M€ uller, M., Katzberg, M., Bertau, M., and
656–659; (2002) Angew. Chem. Int. Ed., Hummel, W. (2010) Org. Biomol. Chem.,
41, 634–637. 8, 1540–1550.
74 M€ uller, M., Wolberg, M., Schubert, T., 89 Parkot, J., Gr€oger, H., and Hummel, W.
and Hummel, W. (2005) Adv. Biochem. (2010) Appl. Microbiol. Biotechnol., 86,
Eng./Biotechnol., 92, 261–287. 1813–1820.
75 Burda, E., Hummel, W., and Gr€oger, H. 90 Morikawa, S., Nakai, T., Yasohara, Y.,
(2008) Angew. Chem., 120, 9693–9696; Nanba, H., Kizaki, N., and Hasegawa, J.
(2008) Angew. Chem. Int. Ed., 47, (2005) Biosci. Biotechnol. Biochem., 69,
9551–9554. 544–552.
76 Kraußer, M., Hummel, W., and 91 Wada, M., Kataoka, M., Kawabata, H.,
Gr€oger, H. (2007) Eur. J. Org. Chem., Yasohara, Y., Kizaki, N., Hasegawa, J., and
5175–5179. Shimizu, S. (1998) Biosci. Biotechnol.
77 Abokitse, K. and Hummel, W. (2003) Biochem., 62, 280–285.
Appl. Microbiol. Biotechnol., 62, 380–386. 92 Yasohara, Y., Kizaki, N., Hasegawa, J.,
78 Hummel, W., Abokitse, K., Drauz, K., Wada, M., Kataoka, M., and Shimizu, S.
Rollmann, C., and Gr€oger, H. (2003) Adv. (2000) Biosci. Biotechnol. Biochem., 64,
Synth. Catal., 345, 153–159. 1430–1436.
79 Gr€oger, H., Chamouleau, F., 93 Yasohara, Y., Kizaki, N., Hasegawa, J.,
Orologas, N., Rollmann, C., Drauz, K., Wada, M., Kataoka, M., and Shimizu, S.
Hummel, W., Weckbecker, A., and (2001) Tetrahedron: Asymmetry, 12,
May, O. (2006) Angew. Chem., 118, 1713–1718.
5806–5809; (2006) Angew. Chem. Int. Ed., 94 Zhu, D.M., Yang, Y., and Hua, L. (2006)
45, 5677–5681. J. Org. Chem., 71, 4202–4205.
80 Gr€oger, H., Hummel, W., Buchholz, S., 95 Kataoka, M., Yamamoto, K., Shimizu, S.,
Drauz, K., Nguyen, T.V., Rollmann, C., Kita, K., Hashimoto, T., Yanase, H.,
References j1105
Kita, A., and Miki, K. (1996) Acta 112 Rosen, T.C., Feldmann, R.,
Crystallogr. Sect. D, 52, 405–406. D€unkelmann, P., and Daußmann, T.
96 Kataoka, M., Sakai, H., Morikawa, T., (2006) Tetrahedron Lett., 47, 4803–4806.
Katoh, M., Miyoshi, T., Shimizu, S., and 113 Eckstein, M., Villela Filho, M., Liese, A.,
Yamada, H. (1992) Biochim. Biophys. Acta, and Kragl, U. (2004) Chem. Commun.,
1122, 57–62. 1084–1085.
97 Yamada, H., Shimizu, S., Kataoka, M., 114 Stillger, T., B€onitz, M., Vilella Filho, M.,
Sakai, H., and Miyoshi, T. (1990) FEMS and Liese, A. (2002) Chem. Ing. Techn., 74,
Microbiol. Lett., 70, 45–48. 1035–1039.
98 Shimizu, S. and Yamada, H. (1990) Ann. 115 Daußmann, T., Hennemann, H.-G.,
N.Y. Acad. Sci., 613 (Enzyme Engineering Rosen, T.C., and D€ unkelmann, P. (2006)
10), 628–632. Chem. Ing. Techn., 78, 249–255.
99 Zhu, D. and Hua, L. (2006) J. Org. Chem., 116 Daußmann, T., Rosen, T.C., and
71, 9484–9486. D€unkelmann, P. (2006) Life Sci. Eng., 6,
100 Zhu, D., Yang, Y., Buynak, J.D., and 125–129.
Hua, L. (2006) Org. Biomol. Chem., 4, 117 Alanvert, E., Doherty, C., Moody, T.S.,
2690–2695. Nesbit, N., Rowan, A.S., Taylor, S.J.C.,
101 Li, H.M., Zhu, D.M., Hua, L., and Vaughan, F., Vaughan, T., Wiffen, J., and
Biehl, E.R. (2009) Adv. Synth. Catal., 351, Wilson, I. (2009) Tetrahedron: Asymmetry,
583–588. 20, 2462–2466.
102 Richter, N., Neumann, M., Liese, A., 118 Kurina-Sanz, M., Bisogno, F.R.,
Wohlgemuth, R., Eggert, T., and Lavandera, I., Orden, A.A., and
Hummel, W. (2009) ChemBioChem, 10, Gotor, V. (2009) Adv. Synth. Catal., 351,
1888–1896. 1842–1848.
103 Richter, N., Neumann, M., Liese, A., 119 Bisogno, F.R., Lavandera, I., Kroutil, W.,
Wohlgemuth, R., Eggert, T., and and Gotor, V. (2009) J. Org. Chem., 74,
Hummel, W. (2010) Biotechnol. Bioeng., 1730–1732.
106, 541–552. 120 Matsumura, S., Imafuku, H.,
104 Knietsch, A., Waschkowitz, T., Takahashi, Y., and Toshima, K. (1993)
Bowien, S., Henne, A., and Daniel, R. Chem. Lett., 251–254.
(2003) Appl. Environ. Microbiol., 69, 121 Itoh, N., Matsuda, M., Mabuchi, M.,
1408–1416. Dairi, T., and Wang, J. (2002)
105 Knietsch, A., Waschkowitz, T., Eur. J. Biochem., 269, 2394–2402.
Bowien, S., Henne, A., and Daniel, R. 122 Matsuyama, A., Yamamoto, H., and
(2003) J. Mol. Microbiol. Biotechnol., 5, Kobayashi, Y. (2002) Org. Proc. Res. Dev.,
46–56. 6, 558–561.
106 Bradshaw, C.W., Fu, H., Shen, G.-J., and 123 Amidjojo, M. and Weuster-Botz, D.
Wong, C.-H. (1992) J. Org. Chem., 57, (2005) Tetrahedron: Asymmetry, 16,
1526–1532. 899–901.
107 Inoue, K., Makino, Y., and 124 H€ollrigl, V., Hollmann, F., Kleeb, A.C.,
Itoh, N. (2005) Appl. Environ. Microbiol., Buehler, K., and Schmid, A. (2008) Appl.
71, 3633–3641. Microb. Biotechnol., 81, 263–273.
108 Inoue, K., Makino, Y., and Itoh, N. 125 Stampfer, W., Kosjek, B., Moitzi, C.,
(2005) Tetrahedron: Asymmetry, 16, Kroutil, W., and Faber, K. (2002) Angew.
2539–2549. Chem., 114, 1056–1059; (2002) Angew.
109 Kelly, D.R. and Lewis, J.D. (1991) J. Chem. Chem. Int. Ed., 41, 1014–1017.
Soc., Chem. Commun., 19, 1330–1332. 126 Stampfer, W., Kosjek, B., Faber, K., and
110 Graham, D.L., Simpson, H.D., and Kroutil, W. (2003) J. Org. Chem., 68,
Cowan, D.A. (1996) Ann. N.Y. Acad. Sci., 402–406.
799, 244–250. 127 van Deursen, R., Stampfer, W.,
111 Musa, M.M., Ziegelmann-Fjeld, K.I., Edegger, K., Faber, K., and Kroutil, W.
Vieille, C., Zeikus, J.G., and Phillips, R.S. (2004) J. Mol. Catal. B Enzym., 31,
(2007) J. Org. Chem., 72, 30–34. 159–163.
128 Goldberg, K., Eddeger, K., Kroutil, W., 142 Kroner, K.H., Sch€
utte, H., Stach, W.,
and Liese, A. (2006) Biotechnol. Bioeng., and Kula, M.R. (1982) J. Chem. Technol.
95, 192–198. Biotechnol., 32, 130–137.
129 Goldberg, K., Krueger, A., Meinhardt, T., 143 Wandrey, C., Wichmann, R.,
Kroutil, W., Mautner, B., and Liese, A. B€uckmann, A.F., and Kula, M.-R. (1980)
(2008) Tetrahedron: Asymmetry, 19, Enzyme Eng., 5, 453–456.
1171–1173. 144 Hummel, W. and Gottwald, C. (1992)
130 de Gonzalo, G., Lavandera, I., Faber, K., (Bayer AG) DE4209022.
and Kroutil, W. (2007) Org. Lett., 9, 145 Shaked, Z. and Whitesides, G.M.
2163–2166. (1980) J. Am. Chem. Soc., 102,
131 Poessl, T.M., Kosjek, B., Ellmer, U., 7104–7105.
Gruber, C.C., Edegger, K., Faber, K., 146 Zelinski, T., Peters, J., and Kula, M.-R.
Hildebrandt, P., Bornscheuer, U.T., and (1994) J. Biotechnol., 33, 283–292.
Kroutil, W. (2005) Adv. Synth. Catal., 347, 147 Peters, J., Minuth, T., and Kula, M.-R.
1827–1834. (1993) Enzyme Microb. Technol., 15,
132 Edegger, K., Stampfer, W., Seisser, B., 950–958.
Faber, K., Mayer, S.F., Oehrlein, R., 148 Peters, J., Zelinski, T., Minuth, T., and
Hafner, A., and Kroutil, W. (2006) Kula, M.-R. (1993) Tetrahedron:
Eur. J. Org. Chem., 1904–1909. Asymmetry, 4, 1683–1692.
133 Stampfer, W., Edegger, K., Kosjek, B., 149 Peters, J., Zelinski, T., and Kula, M.-R.
Faber, K., and Kroutil, W. (2004) Adv. (1992) Appl. Microb. Biotechnol., 38,
Synth. Catal., 346, 57–62. 334–340.
134 Lavandera, I., Kern, A., 150 Kruse, W., Hummel, W., and Kragl, U.
Schaffenberger, M., Gross, J., Glieder, A., (1996) Recl. Trav. Chim. Pays-Bas, 115,
De Wildeman, S., and Kroutil, W. (2008) 239–243.
ChemSusChem, 1, 431–436. 151 Kragl, U., Kruse, W., Hummel, W., and
135 Lavandera, I., Oberdorfer, G., Gross, J., Wandrey, C. (1996) Biotechnol. Bioeng., 52,
de Wildeman, S., and Kroutil, W. (2008) 309–319.
Eur. J. Org. Chem., 2539–2543. 152 Liese, A., Seelbach, K., and Wandrey, C.
136 Harada, T., Kubota, Y., Kamitanka, T., (eds) (2006) Industrial Biotransformations,
Nakamura, K., and Matsuda, T. (2009) 2nd edn, Wiley-VCH Verlag GmbH,
Tetrahedron Lett., 50, 4934–4936. Weinheim, pp. 157–160.
137 Schroer, K., Mackfeld, U., Tan, I.A.W., 153 Rissom, S., Beliczey, J., Giffels, G.,
Wandrey, C., Heuser, F., Kragl, U., and Wandrey, C. (1999)
Bringer-Meyer, S., Weckbecker, A., Tetrahedron: Asymmetry, 10, 923–928.
Hummel, W., Daußmann, T., Pfaller, R., 154 Liese, A., Zelinski, T., Kula, M.-R.,
Liese, A., and L€
utz, S. (2007) J. Biotechnol., Kierkels, H., Karutz, M., Kragl, U., and
132, 438–444. Wandrey, C. (1998) J. Mol. Cat. B: Enzym.,
138 Schrittwieser, J.H., Lavandera, I., 4, 91–99.
Seisser, B., Mautner, B., and 155 Zelinski, T., Liese, A., Wandrey, C., and
Kroutil, W. (2009) Eur. J. Org. Chem., Kula, M.-R. (1999) Tetrahedron:
2293–2298. Asymmetry, 10, 1681–1687.
139 Burda, E., Bauer, W., Hummel, W., and 156 Gr€oger, H., Rollmann, C., H€ usken, H.,
Gr€oger, H. (2010) ChemCatChem, 2, Werner, H., Chamouleau, F.,
67–72. Hagedorn, C., Drauz, K., and
140 Baer, K., Kraußer, M., Burda, E., Hummel, W. (2004) (Degussa AG)
Hummel, W., Berkessel, A., and WO2004085662.
Gr€oger, H. (2009) Angew. Chem., 121, 157 Gr€oger, H. (2004) Chiral Europe
9519–9522; (2009) Angew. Chem. Int. Ed., Conference Proceedings, Mainz,
48, 9355–9358. Germany, June 14–16, 2004.
141 Slusarczyk, H., Felber, S., Kula, M.-R., 158 Wu, X., Wang, Y., Ju, J., Chen, C., Liu, N.,
and Pohl, M. (2000) Eur. J. Biochem., 267, and Chen, Y. (2009) Tetrahedron:
1280–1289. Asymmetry, 20, 2504–2509.
References j1107
159 Nanduri, V.B., Banerjee, A., Howell, J.M., Kataoka, M., and Shimizu, S. (1996) Appl.
Brzozowski, B.B., Eiring, R.F., and Environ. Microbiol., 62, 2303–2310.
Patel, R.N. (2000) J. Ind. Microbiol. 175 Shimizu, S., Kataoka, M., Katoh, M.,
Biotechnol., 25, 171–175. Morikawa, T., Miyoshi, T., and
160 Weckbecker, A. and Hummel, W. (2004) Yamada, H. (1990) Appl. Environ.
Biotechnol. Lett., 26, 1739–1744. Microbiol., 56, 2374–2377.
161 Hummel, W. and Kula, M.-R. (1991) 176 Kataoka, M., Rohani, L.P.S.,
(Forschungszentrum J€ ulich GmbH) Yamamoto, K., Wada, M., Kawabata, H.,
EP 456107. Kita, K., Yanase, H., and Shimizu, S.
162 Bradshaw, C.W., Hummel, W., and (1997) Appl. Microbiol. Biotechnol., 48,
Wong, C.-H. (1992) J. Org. Chem., 57, 699–703.
1532–1536. 177 Kataoka, M., Rohani, L.P.S., Wada, M.,
163 Matsuyama, A., Yamamoto, H., and Kita, K., Yanase, H., Urabe, I., and
Kobayashi, Y. (2002) Org. Proc. Res. Dev., Shimizu, S. (1998) Biosci. Biotechnol.
6, 558–561. Biochem., 62, 167–169.
164 Br€autigam, S., Dennewald, D., 178 Kataoka, M., Yamamoto, K.,
Sch€ urmann, M., Lutje-Spelberg, J., Kawabata, H., Wada, M., Kita, K.,
Pitner, W.-R., and Weuster-Botz, D. Yanase, H., and Shimizu, S. (1999)
(2009) Enzyme Microb. Technol., 45, Appl. Microbiol. Biotechnol., 51, 486–490.
310–316. 179 Kizaki, N., Yasohara, Y., Hasegawa, J.,
165 Wong, C.-H., Drueckhammer, D.G., and Wada, M., Kataoka, M., and Shimizu, S.
Sweers, H.M. (1985) J. Am. Chem. Soc., (2001) Appl. Microbiol. Biotechnol., 55,
107, 4028–4031. 590–595.
166 Wong, C.-H. and Drueckhammer, D.G. 180 Weckbecker, A. and Hummel, W. (2004)
(1995) Bio/Technology, 3, 649–651. Methods in Biotechnology, Microbial
167 Bastos, F.M., dos Santos, A.G., Jones, J. Enzymes and Biotransformations, vol. 17
Jr., Oestreicher, E.G., Pinto, G.F., and (ed. J.L. Barredo), Humana Press, Totowa,
Paiva, L.M.C. (1999) Biotechnol. Technol., pp. 225–238.
13, 661–664. 181 Gr€oger, H., Rollmann, C.,
168 Zhu, D., Mukherjee, C., and Hua, L. Chamouleau, F., Sebastien, I., May, O.,
(2005) Tetrahedron: Asymmetry, 16, Wienand, W., and Drauz, K. (2007)
3275–3278. Adv. Synth. Catal., 349, 709–712.
169 Kaluzna, I.A., Rozzell, J.D., and 182 Berkessel, A., Rollmann, C.,
Kambourakis, S. (2005) Tetrahedron: Chamouleau, F., Labs, S., May, O., and
Asymmetry, 16, 3682–3689. Gr€oger, H. (2007) Adv. Synth. Catal., 349,
170 Ma, S.V., Gruber, J., Davis, C., 2697–2704.
Newman, L., Gray, D., Wang, A., 183 Ema, T., Yagasaki, H., Okita, N.,
Grate, J., Huisman, G.W., and Nishikawa, K., Korenga, T., and Sakai, T.
Sheldon, R.A. (2010) Green. Chem., 12, (2005) Tetrahedron:Asymmetry, 16,
81–86. 1075–1078.
171 Review: Kataoka, M., Kita, K., Wada, M., 184 Zhang, J., Witholt, B., and Li, Z. (2006)
Yasohara, Y., Hasegawa, J., and Chem. Commun., 398–400.
Shimizu, S. (2003) Appl. Microbiol. 185 Wong, C.-H. and Whitesides, G.M. (1981)
Biotechnol., 62, 437–445. J. Am. Chem. Soc., 103, 4890–4899.
172 Yamada, H., Shimizu, S., Kataoka, M., 186 Drueckhammer, D.J., Sadozai, S.K.,
Sakai, H., and Miyoshi, T. (1990) FEMS Wong, C.-H., and Roberts, S.M. (1987)
Microbiol. Lett., 70, 45–48. Enzyme Microb. Technol., 9, 564–570.
173 Kataoka, M., Sakai, H., Morikawa, T., 187 Hummel, W. (1990) Appl. Microbiol.
Katoh, M., Miyoshi, T., Shimizu, S., and Biotechnol., 34, 15–19.
Yamada, H. (1992) Biochim. Biophys. Acta, 188 Feske, B.D. and Stewart, J.D. (2005)
1122, 57–62. Tetrahedron: Asymmetry, 16, 3124–3127.
174 Kita, K., Matuszaki, K., Hashimoto, T., 189 Wittman, M., Carboni, J.M., Attar, R.,
Yanase, H., Kato, N., Chung, M.C.M., Balasubramanian, B., Balimane, P.,
Brassil, P., Beaulieu, F., Chang, C., 202 Tosa, M.O., Podea, P.V., Paizs, C., and
Clarke, W., Dell, J., Eummer, J., Irimie, F.D. (2008) Tetrahedron:
Frennesson, D., Gottardis, M., Greer, A., Asymmetry, 19, 2068–2071.
Hansel, S., Hurlburt, W., Jacobson, B., 203 Acetti, D., Brenna, E., Fuganti, C.,
Krishnananthan, S., Lee, F.Y., Li, A., Gatti, F.G., and Serra, S. (2010)
Lin, T.-A., Liu, P., Ouellet, C., Sang, W., Eur. J. Org. Chem., 142–151.
Saulnier, M.G., Stoffan, K., Sun, Y., 204 Chen, Y., Lin, H., Xu, X., Xia, S., and
Velaparthi, U., Wong, H., Yang, Z., Wang, L. (2008) Adv. Synth. Catal., 350,
Zimmermann, K., Zoeckler, M., and 426–430.
Vyas, D. (2005) J. Med. Chem., 48, 205 Bertau, M., B€ urli, M., Hungerb€uhler, E.,
5639–5643. and Wagner, P. (2001) Tetrahedron:
190 Hanson, R.L., Goldberg, S., Goswami, A., Asymmetry, 12, 2103–2107.
Tully, T.P., and Patel, R.N. (2005) Adv. 206 He, J., Mao, X., Sun, Z., Zheng, P., Ni, Y.,
Synth. Catal., 347, 1073–1080. and Xu, Y. (2007) Biotechnol. J., 2,
191 Zhang, J., Witholt, B., and Li, Z. (2006) 260–265.
Adv. Synth. Catal., 348, 429–433. 207 Nakamura, K. and Matsuda, T. (2006)
192 Vrtis, J.M., White, A.K., Metcalf, W.W., Curr. Org. Chem., 10, 1217–1246.
and van der Donk, W.A. (2001) 208 Salvi, N.A. and Chattopadhyay, S. (2001)
J. Am. Chem. Soc., 123, 2672–2673. Tetrahedron, 57, 2833–2839.
193 Vrtis, J.M., White, A.K., Metcalf, W.W., 209 Kumaraswamy, G. and Ramesh, S. (2003)
and van der Donk, W.A. (2002) Green Chem., 5, 306–308.
Angew. Chem., 114, 3391–3393; (2002) 210 Yadav, J.S., Reddy, B.V.S.,
Angew. Chem. Int. Ed., 41, 3257–3259. Sreelakshmi, C., and Rao, A.B. (2009)
194 Johannes, T.W., Woodyer, R.D., and Synthesis, 1881–1885.
Zhao, H. (2007) Biotechnol. Bioeng., 96, 211 Bertau, M. and B€ urli, M. (2000) Chimia,
18–26. 54, 503–507.
195 (a) Reviews about bakers yeast as 212 Bertau, M., B€ urli, M., and
biocatalyst in organic synthesis: Servi, S. Hungerb€ uhler, E. (1998) Chim. Oggi., 16,
(1990) Synthesis, 1–25;(b)Csuk, R. and 58–61.
Glaenzer, B.I. (1991) Chem. Rev., 91, 213 Patel, R., Banerjee, A., Howell, J.,
49–97. McNamee, C., Brozozowski, D.,
196 Acetti, D., Brenna, E., and Fuganti, C. Mirfakhrae, D., Nanduri, V., Thottathil, J.,
(2007) Tetrahedron: Asymmetry, 18, and Szarka, L. (1993) Tetrahedron:
488–492. Asymmetry, 4, 2069–2084.
197 Ahmad, K., Kould, S., Taneja, S.C., 214 Chung, J.Y.L., Ho, G.-J., Charttrain, M.,
Singh, A.P., Kapoor, M., Hassan, R.-u., Roberge, C., Zhao, D., Leazer, J.,
Verma, V., and Qazi, G.N. (2004) Farr, R., Robbins, M., Emerson, K.,
Tetrahedron: Asymmetry, 15, Mathre, D.J., McNamara, J.M.,
1685–1692. Hughes, D.L., Grabowski, E.J.J., and
198 Katoh, T., Mizumoto, S., Fudesaka, M., Reider, P.J. (1999) Tetrahedron Lett., 40,
Nakashima, Y., Kajimoto, T., and 6739–6743.
Node, M. (2006) Synlett, 2176–2182. 215 Charttrain, M., Roberge, C., Chung, J.,
199 Johanson, T., Carlquist, M., Olsson, C., McNamara, J., Zhao, D., Olewinski, R.,
Rudolf, A., Frejd, T., and Gorwa- Hunt, G., Salmon, P., Roush, D.,
Grauslund, M.F. (2008) Appl. Microbiol. Yamazaki, S., Wang, T., Grabowski, E.,
Biotechnol., 77, 1111–1118. Buckland, B., and Greasham, R.
200 Wallentin, C.J., Orentas, E., Butkus, E., (1999) Enzyme Microb. Technol., 25,
and W€arnmark, K. (2009) Synthesis, 489–496.
864–867. 216 Liese, A., Seelbach, K., and
201 Katzberg, M., Wechler, K., M€ uller, M., Wandrey, C. (eds) (2006) Industrial
D€unkelmann, P., Stohrer, J., Biotransformations, 2nd edn,
Hummel, W., and Bertau, M. (2009) Wiley-VCH Verlag GmbH, Weinheim,
Org. Biomol. Chem., 7, 304–314. pp. 191–194.
References j1109
217 Anderson, B.A., Hansen, M.M., 232 Gelo-Pujic, M., Le Guyader, F., and
Harkness, A.R., Henry, C.L., Vicenzi, J.T., Schlama, T. (2006) Tetrahedron:
and Zmijewski, M.J. (1995) J. Am. Chem. Asymmetry, 17, 2000–2005.
Soc., 117, 12358–12359. 233 Pollard, D., Truppo, M., Pollard, J.,
218 Vicenzi, J.T., Zmijewski, M.J., Chen, C.-Y., and Moore, J. (2006)
Reinhard, M.R., Landen, B.E., Tetrahedron: Asymmetry, 17, 554–559.
Muth, W.L., and Marler, P.G. (1997) 234 Such a strategy based on protection
Enzyme Microb. Technol., 20, 494–499. of the hydroxyl group in
219 Liese, A., Seelbach, K., and Wandrey, C. 4-hydroxyacetophenone, asymmetric
(eds) (2006) Industrial Biotransformations, reduction with a chiral
2nd edn, Wiley-VCH Verlag GmbH, chlorodialkylborane and in situ-
Weinheim, pp. 164–166. deprotection has been applied for the
220 Westerhausen, D., Herrmann, S., synthesis of (S)-1-(4-hydroxyphenyl)
Hummel, W., and Steckhan, E. (1992) ethan-1-ol: Everhart, E.T. and Craig, J.C.
Angew. Chem., 104, 1496–1498; (1992) (1991) J. Chem. Soc., Perkin Trans. 1,
Angew. Chem. Int. Ed. Engl., 31, 1701–1707.
1529–1531. 235 Overview: Kleemann, A., Engel, J.,
221 Review: Steckhan, E. (1994) Top. Curr. Kutscher, B., and Reichert, D. (2009)
Chem., 170, 83–111. Pharmaceutical Substances, 5th edn,
222 Yuan, R., Watanabe, S., Kuwabat, S., and Thieme, Stuttgart.
Yoneyama, H. (1997) J. Org. Chem., 62, 236 Kumaraswamy, G. and Ramesh, S. (2003)
2494–2499. Green Chem., 5, 306–308.
223 van den Wittenboer, A., Schmidt, T., 237 Yadav, J.S., Reddy, B.V.S.,
M€ uller, P., Ansorge-Schumacher, M., Sreelakshmi, C., and Rao, A.B. (2009)
and Greiner, L. (2009) Biotechnol. J., 4, Synthesis, 1881–1885.
44–50. 238 Neupert, A., Ress, T., Wittmann, J.,
224 Stuermer, R., Hauer, B., Drew, D., Hummel, W., and Gr€oger, H. (2010)
Breuer, M., and Schroeder, H. (2009) Z. Naturforsch. Teil B, 65, 337–340.
(BASF AG), Eur. Pat. EP1815002B1. 239 Breuer, M., Pletsch, A., Hauer, B., and
225 Gupta, A., Zimmer, A., and Bobkova, M. Siegel, W. (2010) (BASF SE), PCT Pat.
(2009) (IEP GmbH), Eur. Pat. Appl. WO 2010/031776 A2.
EP1745134B1. 240 Lavandera, I., Kern, A., Ferreira-Silva, B.,
226 Matsuda, T., Nakajima, Y., Harada, T., and Glieder, A., de Wildeman, S., and
Nakamura, K. (2002) Tetrahedron: Kroutil, W. (2008) J. Org. Chem., 73,
Asymmetry, 13, 971–974. 6003–6005.
227 Keinan, E., Hafeli, E.K., Seth, K.K., and 241 Yang, W., Xu, J.-H., Xie, Y., Xu, Y.,
Lamed, R. (1986) J. Am. Chem. Soc., 108, Zhao, G., and Lin, G.-Q. (2006)
228 Ramachandran, P.V., Teodorovic, A.V., 242 Truppo, M.D., Pollard, D., and Devine, P.
and Brown, H.C. (1993) Tetrahedron, 49, (2007) Org. Lett., 9, 335–338.
1725–1738. 243 Shafiee, A., Motamedi, H., and King, A.
229 Bucciarelli, M., Forni, A., Moretti, I., and (1998) Appl. Microbiol. Biotechnol., 49,
Torre, G. (1983) Synthesis, 897–899. 709–717.
230 Doderer, K., Gr€oger, H., and May, O. 244 Rozzell, D. and Liang, J. (2008) Specialty
(2010) (Evonik Degussa GmbH), Eur. Chem. Mag., 28, 36–38.
Patent EP 1945785 B1. 245 King, A.O., Corley, E.G., Anderson, R.K.,
231 Patel, R.N., Goswami, A., Chu, L., Larsen, R.D., Verhoeven, T.R.,
Donovan, M.J., Nanduri, V., Goldberg, S., Reider, P.J., Xiang, Y.B., Belley, M.,
Johnston, R., Siva, P.J., Nielsen, B., Fan, Leblanc, Y., Labelle, M., Prasit, P., and
J., He, W.X., Shi, Z., Wang, K.Y., Eiring, Zamboni, R.J. (1993) J. Org. Chem., 58,
R., Cazzulino, D., Singh, A., and Mueller, 3731–3735.
R. (2004) Tetrahedron: Asymmetry, 15, 246 Pollard, D.J., Telari, K., Lane, J.,
1247–1258. Humphrey, G., McWilliams, C.,
Nidositko, S., Salmon, P., and Moore, J. 265 Kalaitzakis, D., Rozzell, J.D., Smonou, I.,
(2006) Biotechnol. Bioeng., 93, 674–686. and Kambourakis, S. (2006) Adv. Synth.
247 Ankati, H., Yang, Y., Zhu, D., Biehl, E.R., Catal., 348, 1958–1969.
and Hua, L. (2008) J. Org. Chem., 73, 266 Kalaitzakis, D., Kambourakis, S.,
6433–6436. Rozzell, D.J., and Smonou, I. (2007)
248 Manam, R.R., Macherla, V.R., and Tetrahedron: Asymmetry, 18, 2418–2426.
Potts, B.C.M. (2007) Tetrahedron Lett., 48, 267 Wolberg, M., Villela Filho, M.,
2537–2540. Bode, S., Geilenkirchen, P.,
249 Yang, Y., Wang, J., and Kayser, M. (2007) Feldmann, R., Liese, A., Hummel, W.,
Tetrahedron: Asymmetry, 18, 2021–2029. and M€ uller, M. (2008) Bioprocess. Biosyst.
250 Hussain, W., Pollard, D.J., Truppo, M., Eng., 31, 183–191.
and Lye, G. (2008) J. Mol. Cat. B: Enzym., 268 Ji, A.G., Wolberg, M., Hummel, W.,
55, 19–29. Wandrey, C., and M€ uller, M. (2001)
251 Kosjek, B., Nti-Gyabaah, J., Telari, K., Chem. Commun., 57–58.
Dunne, L., and Moore, J.C. (2008) 269 L€udeke, S., Richter, M., and M€ uller, M.
Org. Proc. Res. Dev., 12, 584–588. (2009) Adv. Synth. Catal., 351, 253–259.
252 Kosjek, B., Tellers, D.M., Biba, M., 270 Patel, R.N., Banerjee, A., McNamee, C.G.,
Farr, R., and Moore, J.C. (2006) Brzozowski, D., Hanson, R.L., and
Tetrahedron: Asymmetry, 17, 2798–2803. Szarka, L.J. (1993) Enzyme Microb.
253 Truppo, M.D., Kim, J., Brower, M., Technol., 15, 1014–1021.
Madin, A., Sturr, M.G., and Moore, J.C. 271 Liese, A., Seelbach, K., and Wandrey, C.
(2006) J. Mol. Cat. B: Enzym., 38, 158–162. (eds) (2006) Industrial Biotransformations,
254 Mamoli, L. and Vercellone, A. (1937) 2nd edn, Wiley-VCH Verlag GmbH,
Z. Physiol. Chem., 245, 93–101. Weinheim, pp. 162–163.
255 Mamoli, L. and Vercellone, A. (1937) 272 Hammond, R.J., Poston, B.W.,
Ber. Dtsch. Chem. Ges. B, 70, 2079–2082. Ghiviriga, I., and Feske, B.D. (2007)
256 Mamoli, L. and Vercellone, A. (1937) Tetrahedron Lett., 48, 1217–1219.
Ber. Dtsch. Chem. Ges. B, 70, 470–471. 273 Zhu, D., Ankati, H., Mukherjee, C.,
257 Kardinahl, S., Rabelt, D., and Reschke, M. Yang, Y., Biehl, E.R., and Hua, L. (2007)
(2006) Chem. Ing. Techn., 78, 209–217. Org. Lett., 9, 2561–2563.
258 Camerino, B., Alberti, C.G., and 274 Review: Cheetham, P.S.J. (1993) Trends
Vercellone, A. (1953) Helv. Chim. Acta, 36, Biotechnol., 11, 478–488.
1945–1948. 275 Fauconnier, M.-L., Mpambara, A.,
259 Kritchevsky, T.H., Garmaise, D.L., and Delcarte, J., Jacques, P., Thonart, P.,
Gallagher, T.F. (1952) J. Am. Chem. Soc., and Marlier, M. (1999) Biotechnol. Lett.,
74, 483–486. 21, 629–633.
260 Review: Zhdanov, R.I. and Corey, E.J. 276 Etschmann, M.M.W., Sell, D., and
(2009) Steroids, 74, 723–724. Schrader, J. (2003) Biotechnol. Lett., 25,
261 Hirakawa, H., Kamiya, N., Yata, T., and 531–536.
Nagamune, T. (2003) Biochem. Eng. J., 16, 277 Chamouleau, F., Hagedorn, C., May, O.,
35–40. and Gr€oger, H. (2007) Flavour Fragr. J., 22,
262 Okochi, M., Nakagawa, I., Kobayashi, T., 169–172.
Hayashi, S., Furusaki, S., and Honda, H. 278 Giacomini, D., Galletti, P.,
(2007) J. Biotechnol., 128, 376–382. Quintavalla, A., Gucciardo, G., and
263 Ema, T., Okita, N., Ide, S., and Sakai, T. Paradisi, F. (2007) Chem. Commun.,
(2007) Org. Biomol. Chem., 5, 1175–1176. 4038–4040.
264 Kalaitzakis, D., Rozzell, J.D., 279 Friest, J.A., Maezato, Y., Broussy, S.,
Kambourakis, S., and Smonou, I. (2005) Blum, P., and Berkowitz, D.B. (2010)
Org. Lett., 7, 4799–4801. J. Am. Chem. Soc., 132, 5930–5931.
j 1111
27
Reduction of C¼C Double Bonds
Despina J. Bougioukou and Jon D. Stewart
27.1
Introduction
Asymmetric hydrogenation of functionalized alkenes by chiral organometallic

catalysts has been extremely successful in delivering chiral building blocks (for
recent reviews and references, see References [1–3]). Such catalysts can also yield
either product enantiomer, depending on the chirality of the ligand field. One
limitation, however, is that achieving high stereoselectivities almost always depends
on the presence of a polar group such as an amide, acid, or alcohol near the reacting
alkene. These groups are essential in orienting the substrate with respect to the
catalytic metal center. Attempts to apply asymmetric hydrogenation strategies to
olefins conjugated with less polar groups such as aldehydes, ketones, esters, or nitro
groups have been much less successful, although a handful of notable exceptions
have been reported [4–8]. Organocatalytic strategies have also demonstrated excellent
performance in some olefin reductions [9–11]. In addition to these two chemical
approaches, biocatalytic methods for alkene reduction have also been explored.
Employing this strategy side-steps the need to prepare complex ligands or use high
hydrogen pressures. Here, we have summarized known biocatalysts with demon-
strated abilities to reduce alkenes. We have also summarized some of the preparative
reactions carried out with these biocatalysts to show how they can contribute
to organic synthesis. Earlier work in this area has been reviewed by several
authors [12–15].
27.2
Alkene Reduction by Whole Microbial Cells
Until recently, intact (wild-type) whole cells were used most often for
alkene reductions, both because the reducing equivalents (usually supplied by
NAD(P)H) can be regenerated inexpensively by cellular metabolism and because
the actual enzyme(s) responsible for alkene reduction were unknown. One drawback
and Oliver May.
1112 j 27 Reduction of C¼C Double Bonds
to the whole-cell approach, in particular when using wild-type cells, is that additional
reactions such as carbonyl reduction can also occur. Over the past 15 years, the
number of isolated and characterized alkene reductases has grown significantly and
their use in recombinant form (either as isolated enzymes or recombinant whole
cells) is now generally favored over wild-type whole cells. Nonetheless, microbial cells
offer the simplest and readily available methodology for biocatalytic alkene
reductions.
27.2.1
Bakers Yeast
Fischer and Wiedemann reported the first example of a preparative scale bakers
yeast-mediated alkene reduction in 1935 (Scheme 27.1) [16]. In addition to the
desired alkene reduction, carbonyl reduction was also observed. Bakers yeast
reductions of cyclopentenones, cyclohexenones, and their derivatives have been
used to prepare chiral building blocks for the syntheses of prostaglandins [17],
carotenoids [18], and other terpenoid compounds [19]. Whole bakers yeast cells have
also been used to reduce aliphatic a,b-unsaturated aldehydes and ketones, leading to
chiral building blocks for natural phytol [20], a-tocopherol [18, 21], and insect
pheromones [22, 23].
O Bakers' yeast O OH
1 kg sugar
7 days
+ *
CH 3 CH3 CH 3
* *
30 g 8 0% 2 0%
Scheme 27.1
The bakers yeast-mediated reduction of unsaturated aldehydes is usually accom-

panied by carbonyl group reduction by alcohol dehydrogenases [24]. The reversibility
of this process allows allylic alcohols to be used in place of enals, which are often
highly reactive and unstable. The process involves initial oxidation to the correspond-
ing aldehyde [25–27] to provide an activating group for alkene reduction. Scheme 27.2
shows an example of this type of sequence [24].
O O H O H O
fast slow fast
OH O O OH
H3 C H 3C H3 C H 3C
1 2 3 4
Scheme 27.2
With the exception of a-halogenated acrylates (where the stereochemical out-

come depends on the starting alkene geometry) [28], bakers yeast cannot reduce
a,b-unsaturated carboxylic acids [29]. Unsaturated esters are also inert to bakers
27.2 Alkene Reduction by Whole Microbial Cells j 1113
yeast, with the exception of a,b-unsaturated lactones (e.g., see Scheme 27.3) [30].
Yeast-mediated reductions of nitroalkenes afforded optically active nitroalkanes
[31, 32] that are important synthetic intermediates. However, when a-substituents
other than hydrogens are present, the nitroalkanes obtained are racemic at this
position [33].
Bakers'
yeast
+
Ph O O Ph O O Ph O O
rac-5 ( R )-goniothalamin 6 (R)-5
Scheme 27.3
Fronza et al. investigated the bakers yeast-mediated reduction of enone 7

(Scheme 27.4) in their synthesis of raspberry ketone, a key component of the
flavor of raspberry fruit [34], by natural means [35, 36]. When the reduction was
carried out with NADPH or NADH in D2O, deuterium was incorporated only at
the a-carbon, as expected. By contrast, deuterated nicotinamide cofactors led to
deuterium labeling at the b-carbon, although the facial selectivities for the two
cofactors were opposite (A-type for NADPH and B-type for NADH). Taken
together, these data indicate enzyme-mediated enone reduction involves the
addition of hydride (or an equivalent species) to the b-carbon with solvent-derived
protonation at the a-carbon. These studies have recently been extended to include
tri- and tetrasubstituted cinnamaldehydes [37].
Yeast enzyme
concentrate O OH
(Sigma type II)
CH3 + CH3
in D2O
O D D
HO HO
CH3 8 (major) 9 (minor)
Yeast enzyme
HO concentrate D O D OH
(Sigma type II)
7
CH3 + CH3
[(4R)-D]-NADPD
or HO HO
[(4S)-D]-NADD
10 (major) 11 (minor)
Scheme 27.4
Synthetic studies of raspberry ketone also yielded some empirical rules for pre-
dicting the stereochemical outcomes of bakers yeast bioreductions with acyclic
unsaturated compounds (Figure 27.1). Servi proposed that double bond reduction
occurs with net trans-addition of H2 with the facial selectivity shown when a-substi-
tuted alkenes are employed (A-type reaction, Figure 27.1). The opposite stereochem-
istry was proposed for b-substituted alkenes (B-type reaction). Since whole yeast cells
j 27 Reduction of C¼C Double Bonds
1114
Figure 27.1 Servis rules for (a) A- and (b) B-type reduction in bakers yeast.
were employed in these studies it is not possible to determine whether a single enzyme
is capable of both A-and B-type stereochemical pathways for H2 addition orwhether the
observed outcomes result from multiple alkene reductases within the yeast cell that
accept the same substrates [38]. It is also possible that a single enzyme might catalyze
both net cis- and trans-addition of H2 (see, for example, Reference [39]).
Bakers yeast has also been used in the production of levodione, a key intermediate
for synthesizing carotenoids [18] and flavor compounds [40]. Methods based on
organometallic catalysts have yielded relatively poor chemo- and regioselectiv-
ities [41–43]. In contrast, bakers yeast affords pure, crystalline levodione in 85%
yield and excellent optical purity that can be further converted into optically active
actinol (Scheme 27.5) [44].
27.2.2
Other Microbial Species
Other fungi have also been used to reduce simple cyclic unsaturated compounds
[45–47] as well as more complex derivatives [48, 49]. In addition, cyanobacteria and
plant cells lines have shown notable, and in certain cases, complementary
stereoselectivities as compared to bakers yeast [50–53]; however, their applicability
among organic chemists is narrow simply because complex and time-consuming
cultivation techniques are required. Goretti et al. recently reported a large-scale
screening of yeast strains for alkene reductase activity, and for providing
27.2 Alkene Reduction by Whole Microbial Cells j 1115
O OH
CH 3 CH3 CH3 CH3
CH3 CH3
O
CH 3 OH O
H2 /metal CH3
CH3 14 15
O OH
O CH 3 CH3
CH3 CH3
O 13 CH3 CH3
CH3 CH3
CH 3 OH O
16 17
O
12
O O O
CH 3 CH3 CH3 CH3 CH3 CH3
Bakers'yeast Bakers' yeast
CH3 CH3 + CH3
O OH OH
(S)-levodione 13 (R)-actinol 18 (R)-actinol 18

80 - 85% yield
98% ee
Scheme 27.5
additional microorganisms that may overcome some of the above-mentioned

problems [54].
While bakers yeast is the most frequently used microorganism for biocatalytic
reductions, it is certainly not the most powerful one. The reductive capabilities of
several Clostridia species are remarkable [55, 56] and their ability to reduce
unsaturated carboxylic acids in a stereospecific and NAD(P)H-free manner
renders them a very attractive system. The Simon laboratory has tested nearly
every trisubstituted acrylate as a substrate for these microorganisms [57, 58]. In
addition, several innovative techniques for supplying cheap and by-product-free
reducing equivalents have been applied to this system, including electrochemistry
and direct reduction by H2 gas [59, 60]. The simplest scheme consists of a
cathode electrode, an electron mediator such as methyl viologen, the unsatu-
rated acid, and the biocatalyst in a form of intact cells, a cell extract, or even a
purified enoate reductase. Rapid inactivation by oxygen is the major drawback of
these otherwise very useful organisms and enzymes. In comparison to bakers
yeast, the Clostridia species are more broadly applicable, reducing unsaturated
carboxylic acids in addition to most of the enones and enals accepted by bakers
yeast. Unfortunately, neither microorganism tolerates two bulky substituents at
the alkene b-position.
1116
27.3
Alkene Reductions by Isolated Enzymes
27.3.1
Saccharomyces pastorianus Old Yellow Enzyme
Old yellow enzyme (OYE) was isolated in 1932 by Christian and Warburg from
brewers bottom yeast (Saccharomyces carlsbergensis; later re-named Saccharomyces
pastorianus) [61]. This was the first protein shown to have a non-protein component,
later identified as a non-covalently-bound flavin mononucleotide (FMN). The name
old yellow enzyme came from its color and the need to distinguish it from a second
yellow protein isolated a few years later [62]. OYE has served as a model flavoprotein
and its properties have been investigated extensively by the Massey group.
The ability of oxidized OYE to bind aromatic compounds – especially phenols –
gives rise to new, long wavelength absorbance bands (green form of OYE) due to
transfer of charge from the phenolate to flavin [63–65]. This property was exploited in
a highly useful affinity purification of OYE using N-(4-hydroxybenzoyl)aminohexyl
agarose [65]. OYE binds with high selectivity to the column in its oxidized form; upon
in situ flavin reduction by sodium dithionite the protein has greatly diminished
affinity for the phenol and is eluted from the column in nearly pure form. Other small
molecules and pyridine nucleotide derivatives are also inhibitors for OYE, for
example, acetate (KD ¼ 3 mM), azide (KD ¼ 280 mM), pentafluorophenol (KD ¼ 30 mM),
p-hydroxybenzaldehyde (KD ¼ 100 nM), and nicotinate (KD ¼ 240 mM) [66].
The three-dimensional structure of S. pastorianus OYE was reported in 1994 by
Karplus [67]. One highly informative complex contains p-hydroxybenzaldehyde,
a competitive inhibitor. This structure revealed that the side chains of His191 and
Asn194 form hydrogen bonds with the phenol oxygen, positioning the aromatic ring
above the FMN in a way that likely reflects the manner of substrate binding
(Figure 27.2) [67].
b-NADPH is the likely physiological reductant for the OYE, even though the
uncommon a-anomer is slightly more efficient [66]. Sodium dithionite can also be
used; in this case, however, flavin reduction occurs via semiquinone intermediates
that disproportionate slowly into equal quantities of fully reduced and oxidized OYE.
An electron mediator such as methyl viologen may accelerate the disproportionation
process. In contrast to the limited scope of acceptable reducing agents, the reduced
flavin of OYE can be oxidized by a wide variety of substances. Molecular oxygen is an
opportunistic flavin oxidant, yielding hydrogen peroxide and superoxide. Other
electron acceptors include methylene blue, quinones, ferricyanide, ferric ion, and
cytochrome c [66].
That an electron-deficient alkene (2-cyclohexenone) could re-oxidize OYEs FMN
was not reported until 1993 [69]. Two years later, Massey reported the results of an
extensive study on alkene reductions by OYE in which nearly 50 compounds were
tested spectrophotometrically for their ability to reoxidize OYE under anaerobic
conditions (Figure 27.3) [70]. Several enals and enones were good substrates for
NADPH-mediated reduction by OYE. In contrast, a,b-unsaturated acids, esters,
27.3 Alkene Reductions by Isolated Enzymes j 1117
Figure 27.2 Active site of Saccharomyces bonds between the phenolic oxygen and OYE
pastorianus OYE with bound side-chains are indicated by dashed lines.
p-hydroxybenzaldehyde (1OYB). FMN is shown Distances from N5 of FMN and the hydroxyl of
in yellow and the inhibitor in blue. Hydrogen Tyr196 are also shown. [68].
Catalytic efficiencies 21 - 120%

O O O O O O O
CH3 CH3
H H H CH3 H H H H
OH
CH3 Ph Ph
acrolein methyl acrolein 2-ethyl acrolein
(100%) (88%) (78%)
crotonaldehyde cinnam- α-methyl CH3 4
aldehyde cinnamaldehyde
(88%)
(74%) (65%) 4-hydroxynonenal
(118%)
O O O O O O O
CH3
OH CH3 CH3 CH3 CH3
CH3 CH3 CH3
CH3 CH3 CH3
methyl vinyl
2-cyclo- cyclohexane- O O O
ketone 3-penten-2-one
hexenone 1,2-dione 4-oxo-isophorone menadione duroquinone
(100%) (56%) (88%) (76%) (21%) (21%) (78%)
Catalytic efficiencies 1 - 20%

O O O O O O
H CH3 O
CH3 CH3 Ph CH3

OHC
3-methyl-2- trans-4-phenyl- 3-methyl-2-
butenal 3-buten-2-one cyclohexenone
(19%) (2.2%) 3-oxodecalin- 3-oxodecalin- coumarin
(18%)
4-ene 4-ene-10- (3.5%)
(1.5%) carboxyaldehyde
(3.5%)
Figure 27.3 Substrate specificity of Saccharomyces pastorianus OYE. Reaction rates were
measured under anaerobic conditions by following NADPH oxidation and referenced to acrolein
(100%). Substrates that afforded <1% of this rate are not shown.
1118
Tyr 196
His 191
N
O O
H N N
H HH
Asn 194 O O H
R R'
R' R
H
H
Figure 27.4 Proposed mechanism for OYE-mediated enone reductions. Hydride is donated
by reduced FMN to the b-carbon and the a-carbon is protonated by the side-chain of Tyr196.
Hydrogen bonds from the side-chains of Asn194 and His191 orient and activate the enone for
reduction.
nitriles, and amines were not accepted. Substitution at the a- or b-position generally
decreased reaction rates, and the effects were particularly acute for 2-cyclohexenones.
According to the authors, the problem was not substrate binding per se, since
comparable perturbation of the flavin spectrum was observed in each case, but in
the oxidation rate of NADPH. This effect was also noted by Swiderska and Stewart,
who examined a homologous series of b-substituted 2-cyclohexenones [71]. While
3-methyl- and 3-ethyl-2-cyclohexenone were reduced by OYE, the rates of reduction
for the 3-n-propyl- and 3-n-butyl analogs were too low to be practically useful.
Using data from the X-ray crystal structure, along with other experimental results,
a mechanism was proposed in which net anti-addition of H2 occurs by hydride
transfer from the flavin with concomitant protonation by the side-chain of Tyr196
(Figure 27.4). Hydrogen bonds contributed by His191 and Asn194 position and
activate the carbonyl oxygen. The outcomes of other reactions can be predicted by
overlaying the appropriate functional groups onto the cyclohexenone structure.
Massey prepared several site-directed mutants and the results provided further
support for the proposed mechanism [72–75].
Massey also showed that some nitroalkenes are substrates for OYE. Like a carbonyl
group, a nitro moiety strongly polarizes a carbon–carbon double bond, facilitating
hydride attack at the b-carbon and enhancing the acidity of the proton at the
a-position to yield the corresponding saturated compound, and this activity was
demonstrated for three model substrates (Figure 27.5) [76]. S. pastorianus OYE also
NO 2 NO 2 NO 2
S
nitrocyclohexene
β-nitrostyrene nitrovinylthiophene
Figure 27.5 Nitroalkene substrates for Saccharomyces pastorianus OYE.

catalyzes the reductive cleavage of organic nitrates (Scheme 27.6). For propylene
1,2-dinitrate (19) OYE displayed a very strong preference for the primary position
(20 : 21 ¼ 9 : 1).
O O
N NADPH N
O O O O OH
+
O O OH O O
CH3 N CH3 CH3 N
O O
19 20 21
O O
N NADPH N
O O O O OH
+
O O O O O O OH O O O O
N N N N N
O O O O O
22 23 24
Scheme 27.6
NADPH-independent alkene dismutation is also catalyzed by S. pastorianus OYE.

It was initially observed that 2-cyclohexenone was converted into a 1 : 1 mixture of
cyclohexanone and phenol (Scheme 27.7) [69]. In the first step, cyclohexenone acts as
a hydride donor to reduce the FMN, yielding a cyclohexadienone that rapidly
tautomerizes to phenol. The reduced form of the enzyme adds the elements of
H2 to a second molecule of cyclohexenone, using the same mechanism as that
employed in the nicotinamide-dependent reduction. An isotopically labeled substrate
was used to show that trans-elimination was observed for the first step, as would be
expected. Notably, under aerobic conditions, molecular oxygen may competitively
reoxidize OYE, which diminishes the yield of reduced product. The dismutation
reaction is also not accessible to all cyclohexenones. For example, while 4,4-dimethyl-
2-cyclohexenone is reduced by OYE in the presence of NADPH it was not reduced by
the enzyme in the absence of the cofactor. Aromatization of a cyclohexadienone
intermediate is blocked by the geminal substitution, and this is likely responsible for
its failure in dismutation. The dismutation reaction of S. pastorianus OYE and other
related enzymes has recently been examined for preparative purposes by Faber and
coworkers [77].
Massey took advantage of the first step of the dismutation reaction (alkane
desaturation) to develop a kinetic resolution procedure for racemic aldehydes and
ketones at non-enolizable positions (Scheme 27.8) [78]. The mismatch between the
reduction potential of FMN and the saturated ketone substrates makes this process
inefficient for the native enzyme. However, by incorporating a synthetic flavin analog
(8-cyano-FMN) the potentials were better matched and the reaction was correspond-
1120
O O OH
spontaneous
50% yield
E·FMN ox E·FMN red
O O
50% yield
Scheme 27.7
O O
O P O O P O
O O
HO HO
OH OH
OH OH
CH3 N N O NC N N O
NH NH
CH3 N CH3 N
O O
~ -207 mV)
FMN (E°' = 8-cyano-FMN (E°' ~
= -50 mV)
O OYE O O
[8-CN-FMN]
+
CH3 CH3 CH3

rac- 25 26 (R)-25
CH3 H OYE CH3 H CH3 H

[8-CN-FMN]
O O + O
rac- 27 28 (R)-27
Scheme 27.8
ingly more efficient. Molecular oxygen was used to regenerate the oxidized flavin to
continue the catalytic cycles. While this study provided a solution to an otherwise
intractable synthetic problem (racemization of at an unactivated, non-enolizable
position), the reaction is hard to apply in practice because of the difficulty in
separating the desired product from its unsaturated counterpart, a maximal 50%
yield, and the need to prepare a semi-synthetic form of the enzyme.
Masseys extensive work to uncover the catalytic properties of S. pastorianus OYE
was critically important in bringing alkene reductases to the attention of those
engaged in biocatalysis for synthetic purposes. This led to a renewed interest in
identifying additional alkene reductase enzymes by marrying the modern tools of
genomics with classical strategies based on selective culturing and whole-cell activity
screening. The most pressing need was for additional alkene reductases to add to the
synthetic toolbox, and efforts to identify and characterize suitable biocatalysts are
described below.
27.3.2
Fungal Old Yellow Enzyme Superfamily Members
Old yellow enzyme family members are widely distributed in fungi, bacteria, and
plants [12]. All homologs contain flavin mononucleotide (FMN) as a non-covalent
prosthetic group, require NAD(P)H as a cofactor, and can reduce functionalized
alkenes, at least to some extent. The cellular roles are unknown for most members,
although recent experimental evidence suggests defense against acrolein stress [79]
and protection of the actin cytoskeleton in the case of yeast OYEs [80, 81].
Old yellow enzyme family members identified in fungi, along with their substrate
specificities, are listed in Table 27.1 [69, 82–88]. The OYE2 gene was detected in the
bakers yeast genome after screening an S. cerevisiae DNA library using the S. pastorianus
OYE1 gene as a probe [69]. The S. cerevisiae OYE3 gene was discovered when a DOYE2
yeast strain was unexpectedly found to retain OYE activity [82]. The Candida macedo-
niensis old yellow enzyme homolog was initially designated keto-isophorone (KIP)
reductase, based on its ability to reduce 4-oxoisophorone stereoselectively [86].
Estrogen binding protein was initially purified from C. albicans, the most common
fungal pathogen for humans [83]. The amino acid sequence showed no similarity to
human estrogen receptor, as had been anticipated. Instead, the protein shared 46%
amino acid identity with the S. cerevisiae OYE2 gene, which prompted Madani et al. to
assay the purified protein for old yellow enzyme catalytic activity. These studies
revealed that EBT1 reduces common OYE substrates and possesses desaturase
activity. Interestingly, the C. albicans enzyme reduces 19-nor-testosterone in an
NADPH-dependent manner; by contrast, S. pastorianus OYE can only aromatize
the same compound [89].
In the methylotrophic yeast Hansenula polymorpha, overexpression of the HYE
gene cluster confers resistance towards high concentrations of allyl alcohol in
presence of alcohol oxidase (AO) [84]. This observation suggested that the AOs
product (acrolein) is subsequently reduced by HYEs, thereby diminishing its
otherwise high toxicity.
1122
Table 27.1 Fungal OYE homologs.
Organism Protein (gene) name(s) Known substrate(s) Reference
O O O
O CH 3
CH 3 CH3 CH 3 CH3
Saccharomyces cerevisiae Old yellow enzyme 2 (OYE2) CH 3 [69, 82]
O2
Old yellow enzyme 3 (OYE3)
CH3 CH 3
O O O
CH 3 OH
O O
O
CH 3 CH 3
CH 3
Candida albicans ¼ Estrogen binding protein (EBP1) [83]
O
CHO CHO
CH3 Ph
Hansenula polymorpha Hansenula yellow enzyme 1 (HYE1) O [84]

Hansenula yellow enzyme 2 (HYE2) H
Hansenula yellow enzyme 3 (HYE3)
O O O O O
Yarrowia lipolytica N-Ethylmaleimide reductase [85]
(gene unknown) HN N N N Ph N
CH 3 HO2 C
O O CH 3 O O O
O O O
CH 3 CH3
N Ph N
CH 3
O O O
O O
O CH3
O CH3 CH 3 NO 2
CH 3
Candida macedoniensis Old yellow enzyme (oye) CH3 [86]
(Kluyveromyces marxianus)
CH3 O
O
CHO N
Ph
CH3
O
O O O O
Kluyveromyces lactis Kluyveromyces yellow enzyme (KYE1) CH 3 CH 3 [87, 88]
CH 3
CH3
CH3
CH3 O
CH 3 O O O
CH3 H Ph H CH 3 H
O O O
CH 3 CH3 O CH3
O HN N
CH3 H
O O O
27.3 Alkene Reductions by Isolated Enzymes
j 1123
1124
The purification of N-ethylmaleimide reductase from Yarrowia lipolytica arose

from earlier studies of an Escherichia coli protein [90]. In 1979, Mizugaki and
coworkers reported the isolation of an enzyme fraction that reduced cis-2-
alkenoyl-coenzyme A to the corresponding saturated acyl-CoA derivatives in the
presence of NADPH [91]. While characterizing this protein, they were surprised to
find that N-ethylmaleimide (NEM, a common labeling reagent for cysteine thiol
groups) failed to inhibit NADPH consumption; instead, it accelerated the rate of
NADPH oxidation. All attempts to separate these two catalytic activities by chroma-
tography were unsuccessful, which is consistent with the notion that both reactions
were associated with a single enzyme. Based on these results, a protein with similar
catalytic activity was purified from Y. lipolytica [85]. Neither the amino acid sequence
nor the UV-Vis spectrum of the Y. lipolytica protein were reported; however, this
protein most likely belongs to OYE family. A BLAST search using OYE1 as a probe
reveals six putative OYEs in Y. lipolytica.
Biotechnological applications of fungal OYEs have been rather limited. The OYE2
and OYE3 genes from S. cerevisiae, along with the oye gene from C. macedoniensis, have
been clonedand overexpressedin E. coli.All three recombinant enzymesafforded (6R)-
levodione in excellent enantiomeric purity from 4-oxoisophorone [92]. Moreover,
when genes encoding the C. macedoniensis OYE and glucose dehydrogenases were
co-expressed in E. coli, 100-fold higher productivity was achieved, compared to the
reaction with unmodified bakers yeast cells [93]. When a second enzyme (a ketor-
eductase) was added to the reaction mixture after completion of the alkene reduction
step, the doubly chiral compound (R)-actinol was obtained in 94% e.e.
Mergler et al. reported a clever application of OYE that took advantage of its ability
to bind phenols when the flavin is in the oxidized form [87]. The OYE from
Kluyveromyces lactis acted as a biological filter to bind bisphenol A, an important
monomer used in plastics manufacturing that has also been reported to be an
endocrine disruptor. The substrate selectivity of this enzyme was later studied by
Bommarius and coworkers [88].
27.3.3
Bacterial Old Yellow Enzyme Superfamily Members
Table 27.2 summarizes bacterial OYE homologs [94–102]. In 1979, old yellow enzyme
activity was detected in the bacterium Gluconobacter suboxydans. Its significance was
not clear at that time since the enone reductase activity of OYEs had not yet been
recognized [94].
In 1994, Bruces painstaking efforts to elucidate the degradation pathway of
morphine alkaloids in Pseudomonas species led to the isolation of morphinone
reductase from Pseudomonas putida M10 [95, 103]. The purified enzyme reduced
the olefin bonds of both morphinone and codeinone. The products obtained –
hydromorphone and hydrocodone – are difficult to synthesize by chemical means
and possess useful analgesic and antitussive properties [104]. Morphinone reductase
also reduced 2-cyclohexenone in an NADH-dependent manner. Both progesterone
and cortisone bound to the enzyme but were unreactive.
Table 27.2 Bacterial OYE homologs.
Gluconobacter suboxydans Old yellow enzyme (unknown) O2 [94]
HO CH 3O
O
Pseudomonas putida M10 Morphinone reductase (morB) O O [95]
N CH 3 N CH3
O O
CH 3
O O2N NO2
ONO2
Pseudomonas putida II-B Nitroester reductase (xenA) O 2NO ONO 2
[96]
NO 2
O O
O O
O CH3
Pseudomonas putida Xenobiotic reductase (xenA) HN N [88]
CH 3 CH 3 H
O O
CH 3
O O2N NO2
ONO2
Pseudomonas fluorescens I-C Nitroester reductase (xenB) O 2NO ONO2
[96]
(Continued )
j 1125
NO 2
1126
ONO2
Agrobacterium radiobacter Glycerol trinitrate reductase [97]
O 2NO ONO2
(ner)
O O
H3 C H 3C OH
O O
Enterobacter cloacae PB2 Pentaerythritol tetranitrate OH [98]

H3 C H 3C
reductase (orn)
O O
CH 3 OH
O2N NO2 O 2N NO 2 O 2NO ONO2
ONO 2
O2NO ONO 2
O2NO ONO 2
NO 2 NO2
O
NO2
Escherichia coli JM109 N-Ethylmaleimide reductase CHO [100]
CH 3
(nemA)
O2 NO ONO2 CHO
ONO2 NO2
O 2NO ONO2
O2NO ONO2
O O O O
Yersinia bercovieri Yers enoate reductase (Yers-ER) CH 3 CH3 [88]
CH3
CH 3
CH 3
CH 3 O
O O CH 3 CH3 O
Ph H CH3 H CH3 H
O O O
CH 3
O HN N
O O O
O O O O O
Thermoanaerobacter Thermoanaerobacter OYE CH 3 CH 3 [99]
CH 3
pseudethanolicus E39 (ZP 00777979)
CH 3 CH3
O
CH 3 O O
CH 3 O
CH 3 CH 3
H H
Ph H
CH3
O
O O
CH 3 CH3 O CH3 CH3 CH3
HN Ph N NO2
CH3 H Ph (Continued )
O O
j 1127
1128
Bacillus subtilis YqjM (yqjM) O [101]
N
CH 3
O
Shewanella oneidensis SYE1 (sye1) N [102]

CH 3
O
S. oneidensis SYE3 (sye3) N [102]

CH 3
O
O
O
S. oneidensis SYE4 (sye4) N [102]

CH3
O
Old yellow enzyme homologs have been also isolated from other Pseudomonas
species and other bacteria such as Agrobacterium radiobacter and Enterobacter cloacae
for their ability to reduce organic nitrate esters. Whether the XenA enzyme cloned
and expressed from P. putida by Bommarius is identical to that investigated by
Blehert et al. is not clear. Apart from XenA, the enone reductase activity of most of
these enzymes has been examined only superficially. Their ability to degrade the
aromatic explosive 2,4,6-trinitrotoluene (TNT), a persistent pollutant at military
sites [105, 106] has been of much greater interest. Since the first old yellow enzyme
(pentaerythritol tetranitrate reductase) was expressed in transgenic tobacco plants
and its bioremediation capabilities were demonstrated [107], the exact mechanism of
this reaction is under vigorous investigation [108–114]. Recently, Scrutton and
coworkers have published an extensive study of the substrate- and stereoselectivity
of E. cloacae PB2 pentaerythritol tetranitrate reductase [115].
In 2003, the first gene encoding an OYE from a Gram-positive bacterium (yqjM,
from Bacillus subtilis) was cloned and the protein was overexpressed and isolated from
E. coli in a particularly high yield (161 mg l1 culture) [101]. In addition to some
limited spectrophotometric studies regarding YqjMs substrate specificity, it was
found that the presence of xenobiotics such as TNT or agents such as hydrogen
peroxide that contribute to oxidative stress led to rapid induction of the protein in B.
subtilis. These observations further support the notion that the physiological role of
OYEs is detoxification.
Scrutton and coworkers recently reported data for a thermostable alkene reductase
from Thermoanaerobacter pseudethanolicus E39 [99]. This relative of the B. subtilis
YqjM protein has a melting temperature >70 C and also has high solvent tolerance.
Its stereoselectivity toward various model alkenes was determined along with its 3D
structure. This appears to be a very promising enzyme for future synthetic
applications.
The first comparative study among OYEs in the same bacterium was reported for
Shewanella oneidensis [102]. Out of the four putative OYEs in this species, three were
successfully overexpressed in E. coli and purified in both glutathione-S-transferase
(GST)-tagged and in their native forms. Their relative activities towards four different
substrates were investigated, and those reduced efficiently are shown in Table 27.2.
These studies were carried out with the native, untagged proteins. The
corresponding GST-SYEs gave similar results, although the kcat/KM values were
lower in most cases.
27.3.4
Plant Old Yellow Enzyme Superfamily Members
In the mid-1980s, Vick and Zimmerman elucidated the octadecanoid biosynthetic

pathway in several plant tissues, which led to the purification of 12-oxophytodienoic
acid reductase (OPR) from maize kernels [116]. This enzyme catalyzes the fourth step
in the biosynthesis of jasmonic acid from linolenic acid (29) (Scheme 27.9), reducing
the endocyclic double bond of 32 at the expense of NADPH to yield saturated
cyclopentanone 33 [117].
1130
CO2H CO2H
[O]
CH3 CH3
OOH
linoleinic acid 29
(13S)- 30
CH3
CO2H NADPH NADP +
O
CH3
O 12-oxophytodienoic
acid reductase
HO2C
31
(9S,13S)- 32
CH3 CH3
O O
HO2C
HO2C
jasmonic acid 34
33
Scheme 27.9
Ten years later, Schaller and Weiler purified a second OPR homolog from Corydalis
sempervirens after examining the OPR activities in five different plants [118]. They
noted that since the physiological role of OPR is to reduce 32, which is formed by the
allene oxidase/cyclase enzyme, the enzyme preparation would be expected to operate
preferentially on this compound, rather than the trans-analog. Initial GC/MS results
were encouraging (a 6 : 1 preference for the cis- versus the trans-isomer) but not
conclusive. Surprisingly, additional experiments showed that this enzyme prepara-
tion, known today as OPRI, as well as another OPRI homolog from Arabidopsis
thaliana, actually reduced the opposite enantiomer of 32, even though 32 was the
natural product [118–120]. This paradox was resolved by the isolation of a second
isoenzyme (OPRII), which showed a slight preference for the naturally occurring
enantiomer (9S,13S)-32 [121, 122]. Interestingly, the OYE from S. cerevisiae also
showed preference for the same enantiomer of 32. These enzymes are summarized
in Table 27.3 [116, 118, 122–125].
Based on these and other results, it is clear that OPR enzymes should be divided
into two subgroups [123]. Subgroup I consists of AtOPR1, AtOPR2, LeOPR1 (from
tomato), and OsOPR1 (from rice). The biological function of these enzymes is
uncertain. LeOPR1 is the only OYE in this category that is known to accept
compounds other than 2-cyclohexenone and 32 or related compounds
(Table 27.3) [126]. Subgroup II consists of AtOPR3 and LeOPR3, whose physiological
activity is the reduction of (9S,13S)-32, although the highest catalytic activity was
observed for N-ethylmaleimide. Interestingly, LeOPR3 reduces maleic, but not
fumaric, acid. The X-ray crystal structures of LeOPR1 and LeOPR3 have recently
been reported, along with some studies of site-directed mutants [127].
Table 27.3 Plant OYE homologs.
Organism Protein (gene) name(s) Known substrate(s) Reference(s)
Zea mays (corn) 12-Oxo-phytodienoic CH 3 [116]

acid reductase (gene
O
unknown)
HO2 C
CH 3 CH3
O O O
Corydalis sempervirens OPRI (gene unknown) [118]
(pink corydalis)
HO 2 C HO 2C
CH3 CH3
O O
Arabidopsis thaliana 12-Oxo-phytodienoate [120]
(thale cress) reductase 1, AtOPR1
(OPR1)
HO2C HO2C
CH3
O
A. thaliana (thale cress) 12-Oxo-phytodienoate [123]
reductase 2, AtOPR2
(OPR2)
j 1131
HO2C
(Continued )
1132
CH3
O
A. thaliana (thale cress) 12-Oxo-phytodienoate O [122]
reductase 3, AtOPR3
(OPR3)
HO2C
CH3 CH3
O O
HO2C HO2C
CH3 CH3
O O
HO2C HO2C
CH3 CH3
O O
Oryza sativa L. (rice) 12-Oxo-phytodienoic [123]
acid reductase,
OsOPR1 (opda)
HO2C HO2C
CH3
O
O
Lycopersicon esculentum cv. 12-Oxo-phytodienoate N HO2C CO2H [124]
Castlemart II (tomato) reductase 1, LeOPR1 CH3
(OPR1) O
HO2C
O
Pisum sativum PsOPR1 (PsOPR1) [125]
(garden pea)
O
P. sativum (garden pea) PsOPR2 (OPDRA) [125]
O
P. sativum (garden pea) PsOPR3 (PsOPR3) [125]
(Continued )
j 1133
1134
O
O
O
Matsui et al. classified the six OPR-like enzymes from pea into four groups based
on their ability to reduce 2-cyclohexenone [125]. They observed no catalytic activity for
PsOPR5, little for PsOPR3, moderate activities for PsOPR1, PsOPR4, and PsOPR6,
and highest activity for PsOPR2. A more recent study has extended this analysis to
include many additional sequences [128]. These authors also identified crucial amino
acids using bioinformatics tools and structural analysis.
Covello and coworkers cloned a gene encoding artemisinic aldehyde D11(13)
reductase from Artemisia annua, which is involved in the biosynthesis of the
antimalarial compound artemisinin [129]. The sequence of this enzyme is related
to those of 12-oxophytodienoate reductases described above. While the A. annua
reductase showed greatest catalytic activity for artemisinic aldehyde (the presumed
physiological substrate), it was also able to reduce 2-cyclohexenone and ( þ )-carvone.
27.3.5
Enoate Reductases
Enoate reductases belong to a rare class of flavoenzymes containing both FMN and
flavin adenine dinucleotide (FAD) [130]. Four iron and four labile sulfur atoms are
also present in each enzyme subunit. The mechanism of this class of enzymes is not
as well studied as for the OYEs but EPR studies have shown that electrons derived
from NADH flow via FAD and the [4Fe-4S] cluster to the FMN cofactor [131]. Enoate
reductases are very large, multidomain proteins (about 940 kDa). Their FMN
domains are very similar to that of OYE, and in that respect enoate reductases are
considered distant relatives.
Enoate reductases have the unique ability to catalyze reductions of non-activated
2-enoates. This stands in contrast to catalysis by enoyl-CoA and 2,4-dienoyl-
CoA reductases [132, 133], which only accept the corresponding CoA thioesters
[132, 134, 135]. Table 27.4 summarizes the enoate reductases and their known
substrates [135–138].
In 1975, Simon and coworkers observed that the reduction of (E)-2-methyl-
butenoate by some Clostridia species could occur even without prior conversion into
the corresponding CoA ester [139]. A key observation was the different stereochemical
outcome of this reduction, leading to the (2R)-enantiomer, compared to what had
been previously reported for the action of butyryl-CoA reductase on the corresponding
CoA ester, which gives the (2S)-enantiomer [140]. The (2R)-selective enzyme would be
designated 2-enoate reductase. Interestingly, the first preparation of this enzyme
from Clostridium kluyveri lacked the FMN cofactor, although catalytic activity was
detected for the compounds listed in Table 27.4 [137]. A shorter enzyme purification
protocol developed in the same laboratory allowed the enoate reductase from
Clostridium tyrobutyricum to be isolated [141]. This has been the most-studied enzyme
in this family. The C. tyrobutyricum enoate reductase prepared in this way contains
0.6–0.7 equivalents of FMN per subunit, underscoring the lability of the FMN cofactor
in these flavoproteins. In addition, the rapid deactivation of the catalyst in the presence
of oxygen (1–2 min for the reduced form) renders its purification quite laborious since
strictly anaerobic conditions are required in all steps. While attempts to express the
1136
Table 27.4 Enoate reductases.
CO2H CO2H CO2H

CH3 CH3 Ph
Clostridium kluyveri 2-Enoate reductase (enr) [136]
CH3
CH3
CO2H CO2H
Ph Ph
Clostridium tyrobutyricum 2-Enoate reductase (enr) [137]
CH3 CH3
CO2H CO 2H CO2CH3
HO2C H3CO2C HO 2C
CH3 CH3 CH3
CH3 CH3 CH3 CH3

CO2H
CH3 CH3
CO2H
Clostridium sporogenes 2-Enoate reductase (gene unknown) CO2H [138]

Ph
CO2H
CH3
Clostridium thermoaceticum 2-Enoate reductase (enr) [135]
CH3
C. tyrobutyricum enoate reductase in E. coli have failed so far, an enoate reductase from
Clostridium thermoaceticum was successfully overexpressed in E. coli when the
engineered strain was grown under anaerobic conditions [131].
The physiological role of Clostridium sporogenes enoate reductase in 3-phenylpro-
pionate formation has been elucidated by Simon and coworkers (Scheme 27.10) [142].
As might be anticipated, the substrate specificity of the C. sporogenes enoate reductase
is narrow and limited to cinnamic acid (Table 27.4). By contrast, enoate reductases
from C kluyveri and C. tyrobutyricum accept a broad range of enoates (reviewed in
Reference [57]).
O CO2 H3N CO2
CO2 CO2 NADH NAD +

CO2 CO2
NH3 O
L-phenylalanine 35 36
H2O
NADH NAD +
CO 2 CO 2 CO 2
OH enoate reductase
37 38 39
Scheme 27.10
As would be expected, purified enoate reductases display the same patterns of

substrate specificities as those determined from whole cell-mediated reductions (no
bulky disubstitution at the b-position and different stereochemical outcomes
depending on the starting alkene geometry). Geraniate, however, presents a special
case. Reductions of (E)- and (Z)-geraniate afforded (R)- and (S)-citronellate,
respectively, in 95% e.e. when purified C. tyrobutyricum enoate reductase was
employed. When whole Clostridium butyricum cells were substituted for the reaction
with (Z)-geraniate, (S)-citronellate was obtained in lower optical purity (60–85% e.
e.), presumably from isomerization to the thermodynamically more stable (E)-
geraniate prior to alkene reduction [57]. Finally, notably, b-halogenated enoates
undergo elimination upon reduction by enoate reductase, yielding an
a,b-unsaturated acid that is further reduced by the enzyme. The net result is a
dehalogenated, saturated acid coupled with consumption of 2 equivalents of NADH
per molecule of product [58].
Enals are also substrates for enoate reductases. It is important to remove the
saturated aldehyde product from the reaction mixture quickly, however, or low optical
purities result. Product removal can be achieved by in situ enzymatic reduction to the
corresponding alcohol or continuous extraction with an organic solvent [143]. The
erosion of optical purity is due to a desaturation reaction catalyzed by enoate
1138 j 27 Reduction of C¼C Double Bonds
reductases when aldehydes and an electron acceptor such as oxygen are present. This
reaction does not occur with saturated carboxylates. Given the extreme oxygen
sensitivity of enoate reductases, it is surprising that the half-life of enoate reductase
is more than 20 h in the presence of saturated aldehydes and oxygen.
27.3.6
Medium-Chain Dehydrogenases
Alkenal/one oxidoreductases (AORs) constitute a distinct family within the medium-

chain dehydrogenase/reductase (MDR) superfamily, called the LTD family [144, 145].
These enzymes are key players in eicosanoid inactivation and can act either as an allyl
alcohol dehydrogenase (leukotriene B4 12-hydroxydehydrogenase) or as an enone
reductase (15-oxo-prostaglandin 13-reductase) on major endogenous lipids media-
tors (Scheme 27.11) [146].
CH3 CO2H
arachidonic acid 40
CH3
OH O CO2H
CO2H
HO
OH OH
HO CH3
O CH3 O
CO2H
lipoxin A 4 42 15-oxo-prostaglandin E 2 (PGE2) 43
leukotriene B4 (LTB4) 41
Scheme 27.11
AOR is found in several mammalian species and in various tissues. It was first
isolated by Yokomizo et al. from the cytosolic fraction of porcine kidney and
designated LTB4 12-hydroxydehydrogenase (LTB4DH; LTB4 ¼ leukotriene B4) for its
alcohol dehydrogenase activity [147]. It showed 3.5 times higher activity for 6-trans-
LTB4 compared to that for LTB4, whereas its activity towards 6-trans-12-epi-LTB4 was
four times lower. The cDNA for the human enzyme was also cloned and over-
expressed in E. coli from the same laboratory and showed 84.7% identity at the
nucleotide level with that from pig [148]. Table 27.5 summarizes the properties of
medium-chain dehydrogenases [149–155].
In an independent study, Tai and coworkers purified a 15-oxoprostaglandin
reductase (PGR) from pig lung based on its ability to reduce an activated alkene [156].
Table 27.5 Medium-chain dehydrogenases.
Organism Protein (gene) Known substrate(s) Reference

name(s)
Rattus norvegicus NADP-dependent [149]

O O O O
(rat) leukotriene B4 12-
CH3
hydroxydehydrogen- H CH3 H H CH3 H
ase, 15-oxoprosta-
O O
glandin 13-reductase
CH3
(LTB4DH) CH3 H H
O O
CH3 H CH3 H
O O O
CH3 CH3
H CH3 H H
OH OH
O O O O O
CH3
Ph H Ph H CH3 CH3 CH3
CH3
O O O
CH3 CH3 Ph CH3 Ph Ph
HO2C O
(Continued )
CH3 OH
O
j 1139
1140
Organism Protein (gene) Known substrate(s) Reference

name(s)
O O O O
CH3
Arabidopsis thaliana P1-f-crystallin (P1) CH3 H CH3 H [150, 151]
(thale cress)
O O
CH3
CH3 H H
OH
O
H2 NOC
CH3
H
CONH 2
OH
O O
HO 2C
CH3 H
Hordeum vulgare Alkenal dehydroge- [152]
(barley) nase (ALH) O O
CH 3 CH3
H H
O CH 3
Mentha piperita L. ( þ )-Pulegone reduc- CH 3 [153]

cv. Black Micham tase (AY300163)
CH 3
HO O
Fragaria ananassa Quinone reductase [154]
(strawberry) (QR) CH 3 O
CO2H CO2H
Burkholderia sp. 2-Haloacrylate [155]
reductase (caa43) Cl Br
j 1141
1142
Surprisingly, the amino acid sequence of this novel alkene reductase differed from
that of LTB4DH at only a single position. When specific activities were compared,
alkene reduction of 15-oxo PGE2 (PGE2 ¼ prostaglandin E2) was 300-fold higher than
alcohol oxidation of LTB4, although both reactions appear to be physiologically
relevant. Subsequent work uncovered a role for this enzyme in inactivating lipoxin
A4. An X-ray crystal structure of the guinea pig alkene reductase/alcohol dehydro-
genase complexed with both NADP þ and the v-chain of 15-oxo-PGE2 suggested a
mechanism for alkene reduction [146]; by contrast, the catalytic mechanism of
alcohol oxidation remains obscure.
The significant sequence similarity between rat AOR and quinone reductase from
E. coli [157] prompted Kensler and coworkers to investigate an additional role for this
enzyme in chemoprotection by degrading toxic by-products of lipid peroxidation
such as enones and enals [149, 158]. The catalytic activity of recombinant rat AOR
toward several enones and enals was investigated spectrophotometrically (Table 27.5).
They found that enones are better substrates for the enzyme than enals, especially
when they bear a long aliphatic chain. Enone or enal substitution at either the a- or
b-positions was not tolerated by the enzyme, nor could it reduce endocyclic double
bonds.
In plants, the P1-f-crystallin (P1ZCr) quinone oxidoreductase from Arabidopsis
thaliana [150], an oxidative stress-induced enzyme, and the alkenal dehydrogenase
from barley possess substrate specificities similar to that of rat AOR
(Table 27.5) [152]. Owing to this function, the alternative term NADPH: 2-
alkenal/one a,b-hydrogenase (ALH) has been proposed to describe of this family
of enzymes [151].
Additional plant enzymes in this family have also been isolated. Pulegone
reductase from peppermint [153] and the quinone oxidoreductase from strawber-
ry [154] reduce the exocyclic double bond of ( þ )-pulegone and 4-hydroxy-5-methyl-2-
methylene-3(2H)-furanone (HMMF), respectively (Table 27.5). Interestingly, pule-
gone reductase lacks strict facial stereoselectivity, affording a mixture of
()-menthone and ( þ )-isomenthone in a 55 : 45 ratio. The catalytic activity of these
enzymes on other substrates has not been reported to date.
Kurata et al. have isolated an inducible alkene reductase from the soil bacterium
Burkholderia sp. WS grown on 2-chloroacrylate [155]. The purified protein
(2-haloacrylate reductase) catalyzed the reduction of chloro- and bromo-acrylates to
the corresponding (S)-products. This enzyme was paired with glucose dehydroge-
nase for the preparative-scale reduction of 2-chloroacrylate, yielding 37.4 g l1(S)-2-
chloropropionate (>99% e.e.) after a 30 h reaction [159]. While the Burkholderia
reductase shares significant sequence similarity (38.2% identity) with E. coli
quinone oxidoreductase, the former showed no detectable activity toward quinones.
Matsushima et al. reported the cloning and overexpression of two enzymes from
tobacco (Nicotiana tabacum) involved in reducing pulegone [160]. Unfortunately,
their catalytic efficiencies were relatively low [161].
In contrast to many other oxidoreductases belonging to the MDR superfamily, all
the alkene reductases listed in Table 27.5 are metal independent, lacking bound
Znþ 2 [134].
27.4 Applications of Alkene Reductases j 1143
27.3.7
Short-Chain Dehydrogenases
As part of the Croteau laboratorys study of monoterpene metabolism in plants,

()-isopiperitone reductase (IspR) was isolated from peppermint (Table 27.6) [153].
A BLASTsequence analysis and identification of conserved motifs placed IspR in the
short-chain dehydrogenase/reductase superfamily. One year later, Doorn et al.
showed that human carbonyl reductase (a classical short chain dehydrogenase),
converted 4-oxo-non-2-enal into a mixture of materials that included the alkene
reduction product [162]. Previous work had uncovered this enzymes ability to reduce
carbonyl groups in many endogenous and xenobiotic compounds and quinones [163].
Interestingly, it was found that two conjugated carbonyl groups are required for the
enone/al activity of human carbonyl reductase. If either carbonyl is converted into the
corresponding alcohol, the product is no longer a substrate for carbonyl reductase.
Both IspR and carbonyl reductase possess classic characteristics of the SDR
superfamily, particularly the NAD(P)H binding motif G-X-X-X-G-X-G [164, 165].
On the other hand, in the catalytic motif Y-X-X-X-K, the Tyr residue has been replaced
by Glu in the case of IspR. A Tyr residue at this position has been shown to participate
in catalysis by making a critical hydrogen bond with the carbonyl group [166].
Kreis and Gr€oger recently reported the application of D4,5-steroid 5b-reductase to
various enones [167]. While progesterone had been considered the native substrate,
the enzyme was even more efficient with 2-cyclohexenone itself. Interestingly, the
same enzyme also accepted two a-substituted acrylate esters.
27.4
Applications of Alkene Reductases
27.4.1
a,b-Unsaturated Aldehydes and Ketones
Table 27.7 summarizes the stereochemical properties of several enzymes within the
old yellow enzyme superfamily [168–170]. In these studies, various methods were
used to supply the reducing equivalents, and both NAD þ /NADH and NADP þ /
NADPH were investigated. One important lesson from these results is that the
enantioselectivity of old yellow enzyme family members is very highly conserved for a
given substrate, and only a few cases show reversal of stereochemistry. These
data also show that these enzymes generally have broad substrate acceptance,
although b-substitutions have a very negative effect on reaction rate.
This effect was also apparent in an independent study by Swiderska and Stewart,
who used whole E. coli cells that overexpressed S. pastorianus OYE to reduce a
homologous series of 2-cyclohexenones (Table 27.8) [71]. Only the methyl substituted
compounds were reduced completely; substrates with bulkier alkyl substituents were
reduced more slowly or not at all. The stereochemical outcomes of all of these
biotransformations are predictable based on the model shown in Figure 27.4.
1144
Table 27.6 Short-chain dehydrogenases.
O
Homo sapiens Human carbonyl reduc- CH3 [162]
H
tase (CBR1) O
Mentha piperita ()-Isopiperitenone CH3 [153]

L. cv. Black Mitcham reductase (AY300162) CH3
O
CH 3
O H 3C
EtO2 C CH3 EtO2C
Arabidopsis thaliana D4,5-Steroid 5b-reductase H3 C OH [167]
(thale cress)
O
Table 27.7 Stereochemical investigations of old yellow enzymes (Nd ¼ not determined).
Substrate Zymomonas Z. mobilis Z. mobilis Z. mobilis Tomato Tomato Bacillus

mobilis OYE-1; OYE-2; OYE-3; NCR; OPR1; OPR3; subtilis YqjM;
% conv. % conv. % conv. % conv. % conv. % conv. % conv.
[% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.]
[169] [169] [169] [169] [168, 170] [168, 170] [168]
O
CH3 97 [racemic] 99 [16 (R)] 50 [34 (S)] 99 [48 (S)] 82 [63 (S)] 38 [64 (S)] >99 [92 (S)]
54 [>99 (S)] 36 [>99 (S)] 25 [>99 (S)] 43 [>99 (S)] Nd 5 1

CH3
O
CH3 94 [87 (R)] 95 [94 (R)] 97 [92 (R)] 97 [93 (R)] 93 [75 (R)] 95 [68 (R)] 95 [93 (R)]
96 [>99 (S)] 91 [>99 (S)] 43 [>99 (S)] 61 [>99 (S)] Nd 1 Nd

CH3
O
CH3
CH3 98 [98 (R)] >99 [97 (R)] >99 [43 (R)] >99 [95 (R)] >95 [91 (R)] >95 [99 (R)] 91 [99 (R)]
(Continued )
27.4 Applications of Alkene Reductases
CH3
O
j 1145
1146
Substrate Zymomonas Z. mobilis Z. mobilis Z. mobilis Tomato Tomato Bacillus

mobilis OYE-1; OYE-2; OYE-3; NCR; OPR1; OPR3; subtilis YqjM;
% conv. % conv. % conv. % conv. % conv. % conv. % conv.
[% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.] [% e.e.]
[169] [169] [169] [169] [168, 170] [168, 170] [168]
HO2C CH3
Nd Nd Nd Nd >99 [>99 (R)] Nd Nd
HO2C
O
CH3
HN >99 [75 (R)] >99 [92 (R)] >99 [89 (R)] >99 [99 (R)] >99 [99 (R)] >99 [99 (R)] >99 [99 (R)]
O
O
CH3
Ph N >99 [>98 (R)] >99 [>98 (R)] >99 [>98 (R)] >99 [>98 (R)] >99 [99 (R)] >99 [99 (R)] >99 [99 (R)]
O
O
CH3
H Nd Nd Nd Nd 96 [47 (R)] 70 [19 (S)] 78 [10 (R)]
CH3
CH3 CH3 O
89 [20 (S)] 97 [20 (R)] 97 [42 (R)] >99 [>95 (S)] 79 [>95 (S)] 96 [>95 (S)] 59 [>95 (S)]
CH3 H
CH3
NO2 >99 [90 (R)] >99 [81 (R)] >99 [80 (R)] >99 [98 (S)] >90 [95 (R)] 75 [93 (S)] 50 [85 (S)]
Ph
Table 27.8 Reductions of 2-cyclohexenones by Saccharomyces pastorianus OYE.
Substrate Conversion (%) Optical purity (% e.e.)
O
CH3 100 96 (R)
CH3 16 90 (R)
100 94 (S)
CH3
76 94 (S)
CH3
25 89 (S)
CH3
O
CH3
18 90 (S)
CH3
O
Not recorded —
CH3
Kosjek and coworkers screened a commercially-available library of alkene

reductases to identify suitable enzymes for cyclopentenone 44 and
a,b-unsaturated nitrile 46 (Scheme 27.12) [171]. In the case of 44, several
reductases carried out the reduction with very high conversion and with consis-
tent stereoselectivity; substituting the ethyl ester yielded similar results. The
reduction of nitrile 46 was part of a model study for a more complex pharma-
ceutical intermediate. A screening study was carried out with several alkene
reductases and various aryl ring-substituted analogs of 46. In all cases, enzymes
were identified with good conversion rates and high stereoselectivities. Two
reactions were carried out on preparative scales (250 mg) using phosphite
dehydrogenase to regenerate NADPH.
1148
O O
F F
ERED114
CO2CH3 NADPH NADP + CO2CH3
44 45
HPO 4 - HPO 3 -
ERED112
CN CN
CH3 NADPH NADP + CH3

46 47
HPO 4 - HPO 3 -
Scheme 27.12
Rosche and Hauer have investigated biocatalytic reductions of citral as part

of a chemoenzymatic route to valuable intermediates such as menthol
(Scheme 27.13) [172, 173]. Citral is approximately a 1: 1 mixture of geranial (48) and
neral (49). It is a low-cost, nearly ideal starting material. The geometric isomers 48
and 49 interconvert readily in aqueous solution, making it much more feasible to use
citral (rather than pure geranial or neral) for preparative purposes. A strain collection
that included both fungi and bacteria was screened for the ability to reduce citral in both
aqueous and two-phase aqueous–organic mixtures. Biphasic conditions were more
successful, and organisms that produced both (R)- and (S)-51 in >99% e.e. were
identified. This led to the construction of recombinant E. coli strains that overexpressed
one of four OYE superfamily members from Z. mobilis [173]. While pure geometric
isomers (48 and 49) led to higher enantioselectivities, (R)-51 was produced in 89% e.e.
from citral using an E. coli strain that overexpressed the OYE2 gene.
CH3 CH 3
CHO
CH 3
geranial 48
CH3 CH3 CH3 CH3
citral 50 CHO + CHO
CH3 CH3
CH3 CH3 (R)-citronellal 51 (S)-citronellal 51
CH3 [H] [H]

CHO
neral 49
CH3 CH3 CH3 CH3
CH3 OH + CH3 OH
(R)-citronellol 52 (S)-citronellol 52
Scheme 27.13
Faber and coworkers also investigated biocatalytic solutions to the problem of citral
reduction [174]. In many cases, alcohol dehydrogenases presented serious compe-
tition, at the citral stage, the citronellal stage, or both. Nevertheless, several strains
with very high stereoselectivities for the alkene reduction step were identified.
27.4.2
Acrylates and Acrylate Esters
Swiderska and Stewart used the ability of S. pastorianus OYE to reduce highly
activated acrylate esters in a chemoenzymatic route to b2-amino acids
(Scheme 27.14) [175]. The substrates for OYE reduction were assembled by a simple,
two-step route that afforded preferentially the (Z)-alkenes 55a–d. Because alkene
geometry is important for substrate binding orientation and stereoselectivity (vide
infra), it was important to carry out the OYE-mediated reduction rapidly to out-
compete spontaneous (Z)-/(E)-isomerization. The reduced products were obtained
in 89–94% e.e., and they could be converted into the free b2-amino acids by nitro
group reduction and ester hydrolysis. A deuterium labeling study uncovered the
regiochemistry of the reduction step. That protonation occurred on the nitro-bearing
carbon demonstrated that the enzyme perceived the substrate to be a nitroalkene with
a carboethoxy substituent, rather than a nitro-substituted acrylate ester. Unfortu-
nately, it was not possible to extend this route to b2-amino acids with larger
substituents since the reaction rates of the OYE-mediated conversion were too slow
to be of practical use.
O R CH3NO2 R MsCl, Et3N

Amberlyst A-21 O2N
CO2Et HO CO2Et
53 54 a R = Me
b R = Et
c R = n-Pr
d R = i-Pr
S. carlsbergensis OYE,
1) NADP +, cofactor
R regeneration system
O2N CO2H
H 2N
CO2Et 2) H2, Ra-Ni
3) HCl, ∆ R
(Z)-55a-d 56a-d
Scheme 27.14
27.4.3
Nitroalkenes
Scrutton and coworkers have extensively investigated nitroalkene reduction by an

OYE family member, E. cloacae PB2 pentaerythritol tetranitrate reductase [115, 176].
1150
Interestingly, under biphasic reaction conditions, alkene geometry was irrelevant,

and both (E)- and (Z)-alkenes were converted into the same stereoisomer. In general,
higher optical purities were obtained from the (Z)-isomers. Preparatively useful
reactions have also been carried out with crude extracts from Clostridium
sporogenes [177].
27.5
Accessing Both Product Enantiomers
One general difficulty associated with biocatalytic strategies is to access both product
stereoisomers. When chemical catalysts are employed, it is simply a matter of
inverting the ligand field. Since only L-amino acids are used in protein biosynthesis,
however, this strategy cannot be employed for enzymes. It is therefore essential to
identify pairs of enzymes – either natural or engineered – that provide access to both
product enantiomers [178].
27.5.1
Using Wild-Type Enzymes
During our studies of rat LTB4 dehydrogenase we made the unexpected

discovery that it reduced the enantiomers of perillaldehyde with opposite
stereochemical courses (Scheme 27.15) [39]. By contrast, the stereochemical
Old yellow enzyme

or
CHO LTB4 dehydrogenase CHO
CH3 CH3
NADPH NADP +
(R)-perillaldehyde 57 cis-58
Old yellow enzyme CHO
CH3
NADPH NADP +
CHO
trans- 59
CH3
LTB4 dehydrogenase CHO

(S)-perillaldehyde 58
CH3
NADPH NADP +
cis- 58
Scheme 27.15
27.5 Accessing Both Product Enantiomers j 1151
outcomes of S. pastorianus OYE-mediated reductions were independent of the
side-chain configuration, as would be expected based on the substrate binding
model described above. A deuterium labeling study showed that LTB4 dehydro-
genase catalyzed net trans-addition of H2 to (R)-57, but net cis-addition of H2 to
(S)-58. To the best of our knowledge, such mechanistic divergence has not been
reported previously. Efforts to identify the amino acid(s) that act as the general
acid(s) were unsuccessful, and it may be that the enol(ate) intermediate is
protonated by solvent or a buffer species. In addition, we were unable to identify
additional LTB4 dehydrogenase substrates that show this stereochemical
divergence.
Faber has uncovered an interesting example of stereochemical divergence between
closely related OYE homologs (Scheme 27.16) [168]. The reversed outcomes are
particularly surprising given that the two enzymes share 55% sequence identity and
70% similarity.
Tomato OPR1 CH3

NO 2 >99% conv., 96% ee
+ Ph
CH3 NADPH NADP
(R)- 61
NO 2
Ph
60 Tomato OPR3 CH3
NO 2 72% conv., 87% ee
+ Ph
NADPH NADP
(S)-61
Scheme 27.16
Both Faber and Rosche have pointed out the importance of alkene geometry in
controlling the stereochemical outcomes of alkene reductase-mediated conver-
sions. As noted above, the geometric isomers within the citral mixture are reduced
with different enantiopreferences, and this contributes to the challenge of using
citral as a feedstock [173]. Hall et al. have shown similar behavior for a fumarate/
maleate pair (Scheme 27.17) [169].
CO2Me CO2Me Z. mobilis OYE1 68% conv., >99% ee

Old yellow enzyme
Z. mobilis OYE2 87% conv., >99% ee
CH3 CO2Me CH3 CO2Me Z. mobilis OYE3 99% conv., 97% ee
NAD(P)H NAD(P) +
62 (R)-63
CO2Me CO2Me Z. mobilis OYE1 99% conv., >99% ee

Old yellow enzyme
Z. mobilis OYE2 99% conv., >99% ee
MeO2C CH3 CH3 CO2Me Z. mobilis OYE3 99% conv., >99% ee
NAD(P)H NAD(P) +
64 (S)-63
Scheme 27.17
1152
Figure 27.6 Location of Trp 116 in the Saccharomyces pastorianus OYE active site. The side-chain of
Tyr196 (general acid) and the bound FMN are shown in stick form. A reasonable location for bound
3-ethyl-2-cyclohexenone (sticks) is also depicted. This figure was rendered in PyMOL [68].
27.5.2
Using Mutant Enzymes
During efforts to improve the ability of S. pastorianus OYE to accept alkenes with
larger b-substituents, Trp 116 was targeted for cassette mutagenesis. This residue
appears to form one wall of the active site, and the side-chain comes very close to the
predicted locations of b-substituents on 2-cyclohexenone substrates (Figure 27.6).
A library that potentially contained all possible replacements at position 116 was
created and screened in a Saccharomyces cerevisiae overexpression system [179].
Clones that retained significant levels of catalytic activity against 3-methyl-2-cyclo-
hexenone were examined further. Unfortunately, only a modest rate increase was
found for 2-cyclohexenones with larger b-substituents.
Surprisingly, when the catalytically active mutants were screened against addi-
tional substrates, very different behavior was noted in some cases. For example, the
wild-type and the W116F and W116I OYE mutants all reduced (R)-carvone 65 in the
manner predicted by the model in Figure 27.4. The results with (S)-carvone (67),
however, were very different (Scheme 27.18). The wild-type and W116F mutant OYE
provided the expected cis-product 68. Notably, the stereochemistry at the reacting
olefin is maintained. On the other hand, the W116I mutant OYE gave reversed
selectivity, yielding trans-69. Deuterium labeling showed that catalysis by the W116I
mutant proceeded by net trans-addition of H2, as does the wild-type enzyme. What is
different is that the isoleucine substitution makes it energetically more favorable for
(S)-carvone (67) to bind with opposite facial selectivity. A few other substrates showed
similarly divergent behavior and computational docking studies reproduced the
experimental results reasonably well. This result, along with the recent observations
by Reetz [180], provides hope that protein engineering can provide access to both
References j 1153
Conversion
Enzyme (%) % de
O O
CH3 Old yellow enzyme CH3 WT >98% 97%
W116F >98% 97%
CH3 CH3
NADPH NADP + W116I 77% >98%
(R)-carvone 65 trans-(1R,4R)- 66
O
wt, W116F
Old yellow enzyme CH3 WT 48% 93%
+ CH3 W116F 40% 77%
O NADPH NADP
CH3
cis-(1R,4S)- 68
CH3
O
W116I
Old yellow enzyme CH3
(S)-carvone 67 W116I >98% 88%
CH3
NADPH NADP +
trans-(1S,4S)- 69
Scheme 27.18
product enantiomers for a wide range of substrates. This will be a major task in this
research area over the next few years.
References
1 Minaard, A.J., Feringa, B.L., Lefort, L., 8472–8476 Richter, N., Gr€oger, H., and
and De Vries, J.G. (2007) Asymmetric Hummel, W. (2011) Appl. Microbiol.
hydrogenation using monodentate Biotechnol., 89, 79–89 Kraußer, M.,
phosphoramidite ligands. Acc. Chem. Winkler, T., Richter, N., Dommer, S.,
Res., 40, 1267–1277. Fingerhut, A., Hummel, W., and Gr€oger,
2 Zhang, W.C., Chi, Y.X., and Zhang, X.M. H. (2011) ChemCatChem, 3, 293–296.
(2007) Developing chiral ligands for 4 Ohta, T., Miyake, T., Seido, N.,
asymmetric hydrogenation. Acc. Chem. Kumobayashi, H., and Takaya, H. (1995)
Res., 40, 1278–1290; Roseblade, S.J. and Asymmetric hydrogenation of olefins
Pfaltz, A. (2007) Iridium-catalyzed with aprotic oxygen functionalities
asymmetric hydrogenation of olefins. catalyzed by BINAP Ru(II) complexes.
Acc. Chem. Res., 40, 1402–1411. J. Org. Chem., 60, 357–363.
3 For recent literature in this area see: 5 Dobbs, D.A., Vanhessche, K.P.M., Brazi,
Stueckler, C., Winkler, C.K., E., Rautenstrauch, V., Lenoir, J.Y., Genêt,
Bonnekessel, M., and Faber, K. (2010) J.P., Wiles, J., and Bergens, S.H. (2000)
Adv. Synth. Catal., 352, 2663–2666 Industrial synthesis of ( þ )-cis-methyl
Winkler, C.K., Stueckler, C., Mueller, dihydrojasmonate by enantioselective
N.J., Pressnitz, D., and Faber, K. (2010) catalytic hydrogenation; identification of
Eur. J. Org. Chem., 6354–6358 Stueckler, the precatalyst. Angew. Chem. Int. Ed., 39,
C., Mueller, N.J., Winkler, C.K., Glueck, 1992.
S.M., Gruber, K., Steinkellner, G., and 6 Lipshutz, B.H. and Servesko, J.M. (2003)
Faber, K. (2010) Dalton Trans., 39, CuH-catalyzed asymmetric conjugate
1154
reductions of acyclic enones. Angew. 19 Gramatica, P., Manitto, P., and Poli, L.
Chem. Int. Ed., 42, 4789–4792. (1985) Chiral synthetic intermediates via
7 Lipshutz, B.H., Servesko, J.M., Petersen, asymmetric hydrogenation of a-methyl-
T.B., and Papa, P.P. (2004) Asymmetric a,b-unsaturated aldehydes by Bakers
1,4-reductions of hindered b-substituted yeast. J. Org. Chem., 50, 4625–4628.
cycloalkenones using catalytic 20 Gramatica, P., Manitto, P., Monti, D., and
SEGPHOS-ligated CuH. Org. Lett., 6, Speranza, G. (1987) Microbial-mediated
1273–1275. syntheses of EPC. 4. Stereoselective total
8 Church, T.L. and Andersson, P.G. (2008) synthesis of natural phytol via double-
Iridium catalysts for the asymmetric bond reductions by Bakers yeast.
hydrogenation of olefins with Tetrahedron, 43, 4481–4486.
nontraditional functional substituents. 21 Fuganti, C. and Grasselli, P. (1977)
Coord. Chem. Rev., 252, 513–531. Transformations of non-conventional
9 Tuttle, J.B., Ouellet, S.G., and MacMillan, substrates by fermenting Bakers yeast –
D.W.C. (2006) Organocatalytic transfer production of optically active methyl-
hydrogenation of cyclic enones. J. Am. diols from aldehydes. Chem. Ind., 983.
Chem. Soc., 128, 12662–12663. 22 Fuganti, C., Grasselli, P., and Servi, S.
10 Martin, N.J.A. and List, B. (2006) Highly (1983) Synthesis of ()-frontalin from the
enantioselective transfer hydrogenation (2S,3R)-diol prepared from alpha-
of a,b-unsaturated ketones. J. Am. Chem. methylcinnamaldehyde and fermenting
Soc., 128, 13368–13369. Bakers-yeast. J. Chem. Soc., Perkin Trans.
11 Martin, N.J.A., Ozores, L., and List, B. 1 241–244.
(2007) Organocatalytic asymmetric 23 Gramatica, P., Giardina, G., Speranza,
transfer hydrogenation of nitroolefins. G., and Manitto, P. (1985) Bakers yeast
J. Am. Chem. Soc., 129, 8976–8977. hydrogenation of carbonyl activated
12 Williams, R.E. and Bruce, N.C. (2002) double-bonds – enantioselective
New uses for an old enzyme – the old synthesis of the (S)-form of the
yellow enzyme family of flavoenzymes. dihydroterpenediol secreted by Danaus
Microbiology, 148, 1607–1614. chrysippus and of a pheromone of
13 Stuermer, R., Hauer, B., Hall, M., and Callosobruchus chinensis L. Chem. Lett.,
Faber, K. (2007) Asymmetric bioreduction 1395–1398.
of activated C¼C bonds using enoate 24 DArrigo, P., Fuganti, C., Pedrocchi-
reductases from the old yellow enzyme Fantoni, G., and Servi, S. (1998)
family. Curr. Opin. Chem. Biol., 11, 1–11. Extractive biocatalysis: a powerful tool in
14 Matzinger, P.K. and Leuenberger, selectivity control in yeast
H.G.W. (1997) Stereoselective Synthesis, biotransformations. Tetrahedron, 54,
vol. 7 (eds J. Houben and T. Weyl), 15017–15026.
G. Thieme, Stuttgart. 25 Ferraboschi, P., Grisenti, P., Casati, R.,
15 Servi, S. (1999) Stereoselective Biocatalysis, Fiecchi, A., and Santaniello, E. (1987)
(ed R.N. Patel), Marcel Dekker, New York. Biohydrogenation of unsaturated
16 Fischer, F.G. and Wiedemann, O. (1935) compounds by Saccharomyces cerevisiae.
Über die Hydrierung Ungesättigter Part 1. Stereochemical aspects of the
Ketone durch Gärende Hefe. reaction and preparation of useful
Biochemische Hydrierungen II. Liebigs bifunctional chiral synthons. J. Chem.
Ann. Chem., 520, 52–70. Soc., Perkin Trans. 1, 1743–1748.
17 Jui, J. (1980) Microbial reactions in 26 Gramatica, P., Manitto, P., Monti, D., and
prostaglandin chemistry. Adv. Biochem. Speranza, G. (1988) Microbial mediated
Eng., 17, 37–62. syntheses of EPC. 5. Regioselective and
18 Leuenberger, H.G.W., Boguth, W., stereoselective hydrogenation of methyl-
Barner, R., Schmid, M., and Zell, R. substituted pentadien-1-ols by Bakers
(1979) Totalsynthese von natürlichem yeast. Tetrahedron, 44, 1299–1304.
a-Tocopherol. Helv. Chim. Acta, 62, 27 Ferraboschi, P., Casati, S., and
455–463. Santaniello, E. (1994) Bakers yeast-
References j 1155
mediated hydrogenation of 2-substituted (1996) Stereochemical aspects of flavour
allyl alcohols – a biocatalytic route to a biogeneration through Bakers yeast
new highly enantioselective synthesis of mediated reduction of carbonyl-activated
(R)-2-methyl alkanols. Tetrahedron: double bonds. Pure Appl. Chem., 68,
Asymmetry, 5 19–20. 2065–2071.
28 Utaka, M., Konishi, S., Okubo, T., Tsuboi, 37 Fronza, G., Fuganti, C., and Serra, S.
S., and Takeda, A. (1987) A facile (2009) Stereochemical course of Bakers
synthesis of optically pure L- yeast mediated reduction of the tri- and
armentomycin and its D-isomer. Highly tetrasubstituted double bonds of
enantioselective reduction of the C–C substituted cinnamaldehydes. Eur. J. Org.
double bond of methyl (E)- and (Z)-2,4,4- Chem., 6160–6171.
trichloro-2-butenoate by using Bakers 38 Koul, S., Crout, D.H.G., Errington, W.,
yeast. Tetrahedron Lett., 28, 1447–1450. and Tax, J. (1995) Biotransformation of
29 Utaka, M., Konishi, S., Mizuoka, A., a,b-unsaturated carbonyl compounds:
Ohkubo, T., Sakai, T., Tsuboi, S., and sulfides, sulfoxides, sulfones, nitriles and
Takeda, A. (1989) Asymmetric reduction esters by yeast species: carbonyl group
of the prochiral carbon-carbon double and carbon-carbon double bond
bond of methyl 2-chloro-2-alkenoates by reduction. J. Chem. Soc., Perkin Trans. 1,
use of fermenting Bakers yeast. J. Org. 2969–2988.
Chem., 54, 4989–4992. 39 Bougioukou, D.J. and Stewart, J.D. (2008)
30 Fuganti, C., Pedrocchi-Fantoni, G., Sarra, Opposite stereochemical courses for
A., and Servi, S. (1994) Stereochemistry enzyme-mediated alkene reductions of
of yeast mediated reduction of an enantiomeric substrate pair. J. Am.
a,b-unsaturated d-lactones in the Chem. Soc., 130, 7655–7658.
goniothalamin series. Tetrahedron: 40 Buque-Taboada, E.M., Straathof, A.J.J.,
Asymmetry, 5, 1135–1138. Heijnen, J.J., and van der Wielen, L.A.M.
31 Ohta, H., Ozaki, K., and Tsuchihashi, G. (2004) In situ product removal using a
(1987) Asymmetric hydrogenation of 2- crystallization loop in asymmetric
Aryl-1-nitropropenes by fermenting reduction of 4-oxoisophorone by
Bakers-yeast. Chem. Lett., 191–192. Saccharomyces cerevisiae. Biotechnol.
32 Kawai, Y., Inaba, Y., and Tokitoh, N. Bioeng., 86, 795–800.
(2001) Asymmetric reduction of 41 Pillai, U.R. and Sahle-Demessie, E.
nitroalkenes with Bakers yeast. (2003) Hydrogenation of 4-
Tetrahedron: Asymmetry, 12, 309–318. oxoisophorone over a Pd/Al2O3 catalyst
33 McAnda, A.F., Roberts, K.D., Smallridge, under supercritical CO2 medium. Ind.
A.J., Ten, A., and Trewhella, M.A. (1998) Eng. Chem. Res., 42, 6688–6696.
Mechanism of the yeast mediated 42 Brunner, H. and Fisch, K. (1993)
reduction of nitrostyrenes in light Asymmetric catalysis. 82.
petroleum. J. Chem. Soc., Perkin Trans. 1, Enantioselective hydration of 4-
501–504. oxoisophorone. J. Organomet. Chem.,
34 Serra, S., Fuganti, C., and Brenna, E. 456, 71–75.
(2005) Biocatalytic preparation of natural 43 von Arx, M., Mallat, T., and Baiker, A.
flavours and fragrances. Trends (1999) Unprecedented selectivity
Biotechnol., 23, 193–198. behaviour in the hydrogenation of an
35 Fronza, G., Fuganti, C., Mendozza, M., alpha,beta-unsaturated ketone:
Rallo, R.S., Ottolina, G., and Joulain, D. hydrogenation of ketoisophorone
(1996) Stereochemistry of the double over alumina-aupported Pt and Pd.
bond saturation in the formation of J. Mol. Catal. A. Chem., 148,
Bakers yeast of 4-(4-hydroxyphenyl)-2- 275–283.
butanone (raspberry ketone). 44 Cheetham, P.S.J. (2004) Bioprocesses for
Tetrahedron, 52, 4041–4052. the manufacture of ingredients for foods
36 Fronza, G., Fuganti, C., Mendozza, M., and cosmetics. Adv. Biochem. Eng.
Rigoni, R., Servi, S., and Zucchi, G. Biotechnol., 86, 83–158.
1156
45 Kergomard, A., Renard, M.F., and reduction of (4S)-( þ )-carvone by yeast

Veschambre, H. (1982) Microbiological enoate reductases. Enz. Microb. Technol.,
reduction of alpha,beta-unsaturated 45, 463–468.
ketones by Beauveria-sulfurescens. J. Org. 55 Simon, H., Rambeck, B., Hashimoto, H.,
Chem., 47, 792–798. Guenther, H., Nohynek, G., and
46 Fauve, A., Renard, M.F., and Neumann, H. (1974) Stereospecific
Veschambre, H. (1987) Inducibility of an hydrogenations with hydrogen gas and
enone reductase system in the fungus microorganisms as catalysts. Angew.
Beauveria sulfurescens: application in Chem. Int. Ed., 13, 608–609.
enantioselective organic synthesis. J. Org. 56 Bader, J., Guenther, H., Rambeck, B., and
Chem., 52, 4893–4897. Simon, H. (1978) Properties of 2
47 Matsumoto, K., Kawabata, Y., Takahashi, clostridia strains acting as catalysts for
J., Fujita, Y., and Hatanaka, M. (1998) preparative stereospecific hydrogenation
Asymmetric reduction of of 2-enoic acids and 2-Alken-1-ols with
a,b-unsaturated cyclic ketones by a yeast. hydrogen gas. Hoppe-Seylers Z. Physiol.
Chem. Lett., 283–284. Chem., 359, 19–27.
48 Sato, Y., Oda, T., and Saito, H. (1977) 57 Simon, H., Bader, J., G€ unther, H.,
Microbial transformation of Neumann, S., and Thanos, J. (1985)
dehydrogriseofulvin and griseofulvin: 2H Chiral compounds synthesized by
N. M. R. and mass spectrometric studies biocatalytic reductions. Angew. Chem. Int.
of stereochemical courses of microbial Ed. Engl., 24, 539–553.
hydrogenation and hydroxylation. J. 58 Simon, H. (1992) Properties and
Chem. Soc., Chem. Commun., 415–417. mechanistic aspects of newly found redox
49 Arnone, A., Cardillo, R., Nasini, G., and enzymes from anaerobes suitable for
de Pava, O.V. (1990) Enantioselective bioconversions on preparatory scale. Pure
reduction of racemic abscisic acid by Appl. Chem., 64, 1181–1186.
Aspergillus niger cultures. J. Chem. Soc., 59 Thanos, I.C.G. and Simon, H. (1987)
Perkin Trans. 1, 3061–3063. Electro-enzymatic viologen-mediated
50 Hirata, T., Takarada, A., Matsushima, A., stereospecific reduction of 2-enoates with
Kondo, Y., and Hamada, H. (2004) free and immobilized enoate reductase
Asymmetric hydrogenation of N- on cellulose filters or modified carbon
substituted maleimides by cultured plant electrodes. J. Biotechnol., 6, 13–29.
cells. Tetrahedron: Asymmetry, 15, 15–16. 60 Thanos, I., Bader, J., Guenther, H.,
51 Shimoda, K., Kubota, N., Hamada, H., Neumann, S., Krauss, F., and Simon, H.
Yamane, S.-y., and Hirata, T. (2004) (1987) Electroenzymatic and
Asymmetric transformation of enones electromicrobial reduction: preparation
with Synechococcus sp. PCC 7942. Bull. of chiral compounds. Methods Enzymol.,
Chem. Soc. Jpn., 77, 2269–2272. 136, 302–317.
52 Hamada, H., Miyamoto, Y., Ishihara, K., 61 Warburg, O. and Christian, W. (1932) Old
Nakajima, N., and Hamada, H. (2002) yellow enzyme. Naturwissenschaften, 20,
Stereoselective reduction of ketone and 688.
enone using plant cell cultures. Plant 62 Karplus, P.A., Fox, K.M., and Massey, V.
Biotechnol., 19, 203–205. (1995) Flavoprotein structure and
53 Shimoda, K., Kubota, N., Hamada, H., mechanism. 8. Structure-function
Kaji, M., and Hirata, T. (2004) relations for old yellow enzyme. FASEB
Asymmetric reduction of enones with J., 9, 1518–1526.
Synechococcus sp. PCC 7942. Tetrahedron: 63 Matthews, R.G., Massey, V., and Sweeley,
Asymmetry, 15, 1677–1679. C.C. (1975) Identification of para-
54 Goretti, M., Ponzoni, C., Caselli, E., hydroxybenzaldehyde as ligand in green
Marchigiani, E., Cramarossa, M.R., form of old yellow enzyme. J. Biol. Chem.,
Turchetti, B., Buzzini, P., and Forti, L. 250, 9294–9298.
(2009) Biotransformation of electron- 64 Abramovitz, A.S. and Massey, V. (1976)
poor alkenes by yeasts: asymmetric Interaction of phenols with old yellow
References j 1157
enzyme. Physical evidence for charge- The role of glutamine 114 in old yellow
transfer complexes. J. Biol. Chem., 251, enzyme. J. Biol. Chem., 277, 2138–2145.
5327–5336. 76 Meah, Y. and Massey, V. (2000) Old yellow
65 Abramovitz, A.S. and Massey, V. (1976) enzyme: stepwise reduction of nitro-
Purification of intact old yellow enzyme olefins and catalysis of aci-nitro
using an affinity matrix for the sole tautomerization. Proc. Natl. Acad. Sci.
chromatographic step. J. Biol. Chem., 251, USA, 97, 10733–10738.
5321–5326. 77 Stueckler, C., Reiter, T.C., Baudendistel,
66 Massey, V. and Schopfer, L.M. (1986) N., and Faber, K. (2010) Nicotinamide-
Reactivity of old yellow enzyme with independent asymmetric bioreduction of
alpha-NADPH and other pyridine C¼C-bonds via disproportionation of
nucleotide derivatives. J. Biol. Chem., 261, enones catalyzed by enoate reductases.
1215–1222. Tetrahedron, 66, 663–667.
67 Fox, K.M. and Karplus, P.A. (1994) Old 78 Murthy, Y.V.S.N., Meah, Y., and Massey,

yellow enzyme at 2 A resolution: overall V. (1999) Conversion of a flavoprotein
structure, ligand binding, and reductase to a desaturase by
comparison with related flavoproteins. manipulation of the flavin redox
Structure, 2, 1089–1105. potential. J. Am. Chem. Soc., 121,
68 DeLano, W.L. (2002) The PyMOL 5344–5345.
Molecular Graphics System, DeLano 79 Trotter, E.W., Collinson, E.J., Dawes, I.W.,
Scientific, San Carlos, CA; http:// and Grant, C.M. (2006) Old yellow
www.pymol.org. enzymes protect against acrolein toxicity
69 Stott, K., Saito, K., Thiele, D.J., and in the yeast Saccharomyces cerevisiae. Appl.
Massey, V. (1993) Old yellow enzyme. The Environ. Microbiol., 72,
discovery of multiple isozymes and a 4885–4892.
family of related proteins. J. Biol. Chem., 80 Haarer, B.K. and Amberg, D.C. (2004)
268, 6097–6106. Old yellow enzyme protects the actin
70 Vaz, A.D.N., Chakraborty, S., and Massey, cytoskeleton from oxidative stress. Mol.
V. (1995) Old yellow enzyme: Biol. Cell, 15, 4522–4531.
aromatization of cyclic enones and the 81 Odat, O., Matt, S., Khalil, H., Kampranis,
mechanism of a novel dismutation S.C., Pfau, R., Tsichlis, P.N., and Maris,
reaction. Biochemistry, 34, 4246–4256. A.M. (2007) Old yellow enzymes, highly
71 Swiderska, M.A. and Stewart, J.D. (2006) homologous FMN oxidoreductases with
Stereoselective enone reductions by modulating roles in oxidative stress and
Saccharomyces carlsbergensis old yellow programmed cell death in yeast. J. Biol.
enzyme. J. Mol. Catal. B: Enzym., 42, Chem., 282, 36010–36023.
52–54. 82 Niino, Y.S., Chakraborty, S., Brown, B.J.,
72 Brown, B.J., Deng, Z., Karplus, P.A., and and Massey, V. (1995) A new old yellow
Massey, V. (1998) On the active site of old enzyme of Saccharomyces cerevisiae.
yellow enzyme. Role of histidine 191 and J. Biol. Chem., 270, 1983–1991.
asparagine 194. J. Biol. Chem., 273, 83 Madani, N.D., Malloy, P.J., Rodriguez-
32753–32762. Pombo, P., Krishnan, A.V., and Feldman,
73 Kohli, R.M. and Massey, V. (1998) The D. (1994) Candida albicans estrogen-
oxidative half-reaction of old yellow binding protein gene encodes an
enzyme – the role of tyrosine 196. J. Biol. oxidoreductase that is inhibited by
Chem., 273, 32763–32770. estradiol. Proc. Natl. Acad. Sci. USA, 91,
74 Xu, D., Kohli, R.M., and Massey, V. (1999) 922–926.
The role of threonine 37 in flavin 84 Komduur, J.A., Leao, A.N., Monastyrka,
reactivity of the old yellow enzyme. Y., Veenhuis, M., and Kiel, J.A.K.W.
Proc. Natl. Acad. Sci. USA, 96, (2002) Old yellow enzyme confers
3556–3561. resistance of Hansenula polymorpha
75 Brown, B.J., Hyun, J.-W., Duvvuri, S., towards allyl alcohol. Curr. Genet., 41,
Karplus, P.A., and Massey, V. (2002) 401–406.
1158
85 Mizugaki, M., Nakazawa, M., Yamamoto, reduction. Appl. Environ. Microbiol., 69,
H., and Yamanaka, H. (1989) Purification 933–937.
and characterization of N- 93 Kataoka, M., Kotaka, A., Thiwthong, R.,
ethylmaleimide reducing enzyme from Wada, M., Nakamori, S., and Shimiza, S.
Candida lipolytica. J. Biochem., 105, (2004) Cloning and overexpression of the
782–784. old yellow enzyme of Candida
86 Kataoka, M., Kotaka, A., Hasegawa, A., macedoniensis, and its application to the
Wada, M., Yoshizumi, a., Nakamori, S., production of a chiral compound.
and Shimizu, S. (2002) Old yellow J. Biotechnol., 114, 1–9.
enzyme from Candida macedoniensis 94 Adachi, O., Matsushita, K., Shinagawa,
catalyzes the stereospecific reduction E., and Ameyama, M. (1979) Occurrence
of the C¼C bond of ketoisophorone. of old yellow enzyme in Gluconobacter
Biosci. Biotechnol. Biochem., 66, suboxydans, and the cyclic regeneration of
2651–2657. NADP þ . J. Biochem., 86, 699–709.
87 Mergler, M., Wolf, K., and Zimmermann, 95 French, C.E. and Bruce, N.C. (1994)
M. (2004) Development of a bisphenol a- Purification and characterization of
adsorbing yeast by surface display of the morphinone reductase from
kluyveromyces yellow enzyme on Pichia Pseudomonas putida M10. Biochem. J.,
pastoris. Appl. Micribiol. Biotechnol., 63, 301, 97–103.
418–421. 96 Blehert, D.S., Fox, B.G., and Chambliss,
88 Chaparro-Riggers, J.F., Rogers, T.A., G.H. (1999) Cloning and sequence
Vazquez-Figueroa, E., Polizzi, K.M., and analysis of two Pseudomonas flavoprotein
Bommarius, A.S. (2007) Comparison of xenobiotic reductases. J. Bacteriol., 181,
three enoate reductases and their 6254–6263.
potential use for biotransformations. Adv. 97 Marshall, S.J., Krause, D., Blencowe,
Synth. Catal., 349, 1521–1531. D.K., and White, G.F. (2004)
89 Buckman, J. and Miller, S.M. (1998) Characterization of glycerol trinitrate
Binding and reactivity of Candida albicans reductase (NerA) and the catalytic role of
estrogen binding protein with steroid and active-site residues. J. Bacteriol., 186,
other substrates. Biochemistry, 37, 1802–1810.
14326–14336. 98 French, C.E., Nicklin, S., and Bruce, N.C.
90 Miura, K., Tomioka, Y., Suzuki, H., (1996) Sequence and properties of
Yonezawa, M., Hishinuma, T., and pentaerythritol tetranitrate reductase
Mizugaki, M. (1997) Molecular cloning of from Enterobacter cloacae PB2.
the nema gene encoding N- J. Bacteriol., 178, 6623–6627.
ethylmaleimide reductase from 99 Adalbj€ornsson, B.V., Toogood, H.,
Escherichia coli. Biol. Pharm. Bull., 20, Fryszkowska, A., Pudney, C.R., Jowitt,
110–112. T.A., Leys, D., and Scrutton, N.S. (2009)
91 Mizugaki, M., Unuma, T., and Yamanaka, Biocatalysis with thermostable enzymes:
H. (1979) Studies on the metabolism of structure and properties of a
unsaturated fatty acids. 2. Separation and thermophilic ene-reductase related to
general properties of reduced old yellow enzyme. ChemBioChem, 11,
nicotinamide adenine dinucleotide 197–207.
phosphate dependent cis-2-enoyl- 100 Williams, R.E., Rathbone, D.A., Scrutton,
coenzyme a reductase from Escherichia N.S., and Bruce, N.C. (2004)
coli K-12. Chem. Pharm. Bull., 27, Biotransformation of explosives by the
2334–2337. old yellow enzyme family of
92 Wada, M., Yoshizumi, A., Noda, Y., flavoproteins. Appl. Environ. Microbiol.,
Kataoka, M., Shimizu, S., Takagi, H., and 70, 3566–3574.
Nakamori, S. (2003) Production of a 101 Fitzpatrick, T.B., Amrhein, N., and
doubly chiral compound, (4R,6R)-4- Macheroux, P. (2003) Characterization of
hydroxy-2,2,6-trimethylcyclohexanone, YqjM, an old yellow enzyme homolog
by two-step enzymatic asymmetric from Bacillus subtilis involved in oxidative
References j 1159
stress response. J. Biol. Chem., 278, tetranitrate reductase from Enterobacter
19891–19897. cloacae PB2. Acta Crystallogr., Sect. D, 54,
102 Brige, A., van den Hemel, D., Carpentier, 675–677.
W., De Smet, L., and Van Beeumen, J.J. 111 Barna, T., Messiha, H.L., Petosa, C.,
(2006) Comparative characterization and Bruce, N.C., Scrutton, N.S., and Moody,
expression analysis of the four old yellow P.C.F. (2002) Crystal structure of bacterial
enzyme homologues from Shewanella morphinone reductase and properties of
oneidensis indicate differences in the C191A mutant enzyme. J. Biol. Chem.,
physiological function. Biochem. J., 394, 277, 30976–30983.
335–344. 112 Blehert, D.S., Knoke, K.L., Fox, B.G., and
103 French, C.E. and Bruce, N.C. (1995) Chambliss, G.H. (1997) Regioselectivity
Bacterial morphinone reductase is related of nitroglycerin denitration by
to old yellow enzyme. Biochem. J., 301, flavoprotein nitroester reductases
671–678. purified from two Pseudomonas species.
104 Boonstra, B., Rathbone, D.A., and Bruce, J. Bacteriol., 179, 6912–6920.
N.C. (2001) Engineering novel 113 Pak, J.W., Knoke, K.L., Noguera, D.R.,
biocatalytic routes for production of Fox, B.G., and Chambliss, G.H. (2000)
semisynthetic opiate drugs. Biomol. Eng., Transformation of 2,4,6-trinitrotoluene
18, 41–47. by purified xenobiotic reductase b from
105 Ramos, J.L., Gonzalez-Perez, M.M., Pseudomonas fluorescens I-C. Appl.
Caballero, A., and van Dillewijn, P. (2005) Environ. Microbiol., 66, 4742–4750.
Bioremediation of polynitrated aromatic 114 Khan, H., Harris, R.J., Barna, t., Craig,
compounds: plants and microbes put up a D.H., Bruce, N.C., Munro, A.W., Moody,
fight. Curr. Opin. Biotechnol., 16, 275–281. P.C.E., and Scrutton, N.S. (2002) Kinetic
106 Stenuit, B., Eyers, L., El Fantroussi, S., and structural basis of reactivity of
and Agathos, S.N. (2005) Promising pentaerythritol tetranitrate reductase
strategies for the mineralisation of 2,4,6- with NADPH, 2-cyclohexenone,
trinitrotoluene. Rev. Environ. Sci. nitroesters and nitroaromatic explosives.
Biotechnol., 4, 39–60. J. Biol. Chem., 277, 21906–21912.
107 French, C.E., Rosser, S.J., Davies, G.J., 115 Fryszkowska, A., Toogood, H., Sakuma,
Nicklin, S., and Bruce, N.C. (1999) M., Gardiner, J.M., Stephens, G.M., and
Biodegradation of explosives by Scrutton, N.S. (2009) Asymmetric
transgenic plants expressing reduction of activated alkenes by
pentaerythritol tetranitrate reductase. pentaerythritol tetranitrate reductase:
Nat. Biotechnol., 17, 491–494. specificity and control of stereochemical
108 French, C.E., Nicklin, S., and Bruce, N.C. outcome by reaction optimisation. Adv.
(1998) Aerobic degradation of 2,4,6- Synth. Catal., 351, 2976–2990.
trinitrotoluene by Enterobacter cloacae 116 Vick, B.A. and Zimmerman, D.C. (1986)
PB2 and by pentaerythritol tetranitrate Characterization of 12-oxo-phytodienoic
reductase. Appl. Environ. Microbiol., 64, acid reductase in corn – the jasmonic acid
2864–2868. pathway. Plant Physiol., 80, 202–205.
109 Messiha, H.L., Bruce, N.C., Sattelle, B.M., 117 Schaller, F. (2001) Enzymes of the
Sutcliffe, M.J., Munro, A.W., and biosynthesis of octadecanoid-derived
Scrutton, N.S. (2005) Role of active site signalling molecules. J. Exp. Bot., 52,
residues and solvent in proton transfer 11–23.
and the modulation of flavin reduction 118 Schaller, F. and Weiler, E.W. (1997)
potential in bacterial morphinone Enzymes of octadecanoid biosynthesis in
reductase. J. Biol. Chem., 280, plants – 12-oxo-phytodienoate 10,11-
27103–27110. reductase. Eur. J. Biochem., 245, 294–299.
110 Moody, P.C.E., Shikotra, N., French, C.E., 119 Schaller, F., Hennig, P., and Weiler, E.W.
Bruce, N.C., and Scrutton, N.S. (1998) (1998) 12-Oxophytodienoate-10,11-
Crystallization and ppreliminary reductase: occurrence of two isoenzymes
diffraction studies of pentaerythritol of different specificity against
1160
stereoisomers of 12-oxophytodienoic Structural basis of substrate specificity of

acid. Plant Physiol., 118, 1345–1351. plant 12-oxophytodienoate reductases. J.
120 Schaller, F. and Weiler, E.W. (1997) Mol. Biol., 392, 1266–1277.
Molecular cloning and characterization of 128 Li, W., Liu, B., Yu, L., Feng, D., Wang, H.,
12-oxophytodienoate reductase, an and Wang, J. (2009) Phylogenetic
enzyme of the octadecanoid signaling analysis, structural evolution and
pathway from Arabidopsis thaliana. J. Biol. functional divergence of the 12-oxo-
Chem., 272, 28066–28072. phytodienoate acid reductase gene family
121 Costa, C.L., Arruda, P., and Benedetti, in plants. BMC Evol. Biol., 9, 90.
C.E. (2000) An Arabidopsis gene induced 129 Zhang, Y., Teoh, K.H., Reed, D.W., Maes,
by wounding functionally homologous to L., Goossens, A., Olson, D.J.H., Ross,
flavoprotein oxidoreductases. Plant Mol. A.R.S., and Covello, P.S. (2008) The
Biol., 44, 61–71. molecular cloning of artemisinic
122 Schaller, F., Biesgen, C., Mussig, C., aldehyde D11(13) reductase and its role in
Altmann, T., and Weiler, E.W. (2000) 12- glandular trichome-dependent
Oxophytodienoate reductase 3 (OPR3) is biosynthesis of artmisinin in
the isoenzyme involved in jasmonate Artemisia annua. J. Biol. Chem., 283,
biosynthesis. Planta, 210, 979–984. 21501–21508.
123 Sobajima, H., Takeda, M., Sugimori, M., 130 Steinbacher, S., Stumpf, M., Weinkauf, S.,
Kobashi, N., Kiribuchi, K., Cho, E.-M., Rohdich, F., Bacher, A., and Simon, H.
Akimoto, C., Yamaguchi, T., Minami, E., (2002) Enoate Reductase Superfamily in
Shibuya, N., Schaller, F., Weiler, E.W., Flavins and Flavoproteins (eds S. Chapman,
Yoshihara, T., Nishda, H., Nojiri, H., R. Perham, and N.S. Scrutton), Rudolf
Omori, T., Nishiyama, M., and Yamane, Weber, Berlin, pp. 941–949.
H. (2003) Cloning and characterization of 131 Caldeira, F.J., Feicht, R., White, H.,
a jasmonic acid-responsive gene Teixeira, M., Moura, J.J.G., Simon, H.,
encoding 12-oxophytodienoic acid and Moura, I. (1996) EPR and Mossbauer
reductase in suspension-cultured rice spectroscopic studies on enoate
cells. Planta, 216, 692–698. reductase. J. Biol. Chem., 271,
124 Strassner, J., Furholz, A., Macheroux, P., 18743–18748.
Amrhein, N., and Schaller, A. (1997) A 132 He, X.-Y., Yang, S.-Y., and Schulz, H.
homolog of old yellow enzyme in tomato (1997) Cloning and expression of the
– spectral properties and substrate fadH gene and characterization of the
specificity of the recombinant protein. J. gene product 2,4-dienoyl coenzyme a
Biol. Chem., 274, 35067–35073. reductase from Escherichia coli. Eur.
125 Matsui, H., Nakamura, G., Ishiga, Y., J. Biochem., 248, 516–520.
Toshima, H., Inagaki, Y., Toyoda, K., 133 Fillgrove, K.L. and Anderson, V.E. (2001)
Shiraishi, T., and Ichinose, Y. (2004) The mechanism of dienoyl-CoA
Structure and expression of 12- reduction by 2,4-dienoyl-CoA reductase is
oxophytodienoate reductase (subgroup I) stepwise: observation of a dienolate
genes in pea, and characterization of the intermediate. Biochemistry, 40,
oxidoreductase activities of their 12412–12421.
recombinant products. Mol. Genet. 134 Liang, X., Thorpe, C., and Schulz, H.
Genom., 271, 1–10. (2000) 2,4-Dienoyl-CoA reductase from
126 Straßner, J., F€urholz, A., Macheroux, P., Escherichia coli is a novel iron-sulfur
Amrhein, N., and Schaller, A. (1999) A favoprotein that functions in fatty acid
homolog of old yellow enzyme in tomato. beta-oxidation. Arch. Biochem. Biophys.,
Spectral properties and substrate 380, 373–379.
specificity of the recombinant protein. J. 135 Rohdich, F., Wiese, A., Feicht, R., Simon,
Biol. Chem., 274, 35067–35073. H., and Bacher, A. (2001) Enoate
127 Breithaupt, C., Kurzbauer, R., Schaller, F., reductases of Clostridia – cloning,
Stintzi, A., Schaller, A., Huber, R., sequencing, and expression. J. Biol.
Macheroux, P., and Clausen, T. (2009) Chem., 276, 5779–5787.
References j 1161
136 Bader, J. and Simon, H. (1980) The 145 Riveros-Rosas, H., Julian-Sanchez, A.,
activities of hydrogenase and enoate Villalobos, R., Pardo Juan, P., and Pina, E.
reductase in 2 clostridium species, their (2003) Diversity, taxonomy and evolution
interrelationship and dependence on of medium-chain dehydrogenase/
growth conditions. Arch. Microbiol., 127, reductase superfamily. Eur. J. Biochem.,
279–287. 270, 3309–3334.
137 Tischer, W., Bader, J., and Simon, H. 146 Hori, T., Yokomizo, T., Ago, H., Sugahara,
(1979) Purification and some properties M., Ueno, G., Yamamoto, M., Kumasaka,
of a hitherto-unknown enzyme reducing T., Shimizu, T., and Miyano, M. (2004)
the carbon-carbon double bond of alpha, Structural basis of leukotriene B4 12-
beta-unsaturated carboxylate anions. Eur. hydroxydehydrogenase/15-oxo-
J. Biochem., 97, 103–112. prostaglandin 13-reductase catalytic
138 Dickert, S., Pierik, A.J., Linder, D., and mechanism and a possible Src homology
Buckel, W. (2000) The involvement of 3 domain binding loop. J. Biol. Chem.,
coenzyme a esters in the dehydration of 279, 22615–22623.
(R)-phenyllactate to (E)-cinnamate by 147 Yokomizo, T., Izumi, T., Takahashi, T.,
Clostridium sporogenes. Eur. J. Biochem., Kasama, T., Kobayashi, Y., Sato, F.,
267, 3874–3884. Taketani, Y., and Shimizu, T. (1993)
139 Hashimoto, H., Guenther, H., and Enzymatic inactivation of leukotriene B4
Simon, H. (1975) Specificity and by a novel enzyme found in the porcine
stereospecificity of conversion of kidney – purification and properties of
different 2,3-unsaturated acids by leukotriene B4 12-
Clostridium kluyeri. Hoppe-Seylers Z. hydroxydehydrogenase. J. Biol. Chem.,
Physiol. Chem., 356, 1195–1201. 268, 18128–18135.
140 La Roche, H.J., Kellner, M., Guenther, H., 148 Yokomizo, T., Ogawa, Y., Uozumi, N.,
and Simon, H. (1971) Preparative Kime, K., Izumi, T., and Shimizu, T.
biochemical synthesis of compounds (1996) cDNA cloning, expression, and
stereospecifically labelled with hydrogen. mutagenesis study of leukotriene B4 12-
2. Stereochemistry of butyryl-coa hydroxydehydrogenase. J. Biol. Chem.,
dehydrogenase in Clostridium kluyveri. 271, 2844–2850.
Hoppe-Seylers Z. Physiol. Chem., 352, 149 Dick, R.A., Kwak, M.-K., Sutter, T.R., and
399–402. Kensler, T.W. (2001) Antioxidative
141 Kuno, S., Bacher, A., and Simon, H. function and substrate specificity of
(1985) Structure of enoate reductase from NAD(P)H-dependent alkenal/one
a Clostridium tyrobutyricum. Biol. Chem. oxidoreductase. J. Biol. Chem., 276,
Hoppe-Seyler, 366, 463–472. 40803–40810.
142 Buehler, M. and Simon, H. (1982) On the 150 Mano, J., Babiychuk, E., Belles-Boix, E.,
kinetics and mechanism of enoate Hiratake, J., Kimura, A., Inze, D.,
reductase. Hoppe-Seylers Z. Physiol. Kushnir, S., and Asada, K. (2000)
Chem., 363, 609–625. A novel NADPH: diamide oxidoreductase
143 Thanos, I., Deffner, A., and Simon, H. activity in Arabidopsis thaliana P1
(1985) Further reactions catalyzed by f-crystallin. Eur. J. Biochem., 267,
enoate reductase - reductions of 3661–3671.
2-enals, dehydrogenation of saturated 151 Mano, J., Torii, Y., Hayahi, S.-i.,
aldehydes and their racemization. Takomoto, K., Matsui, K., Nakamura, K.,
Biol. Chem. Hoppe-Seyler, 369, Inze, D., Babiychuk, E., Kushnir, S., and
451–460. Asada, K. (2002) The NADPH: quinone
144 Nordling, E., Jornvall, H., and Persson, B. oxidoreductase P1-f-crystallin in
(2002) Medium-chain dehydrogenases/ Arabidopsis catalyzes the
reductases (MDR) – family a,b-hydrogenation of 2-alkenals:
characterizations including genome detoxification of the lipid peroxide-
comparisons and active site modelling. derived reactive aldehydes. Plant Cell
Eur. J. Biochem., 269, 4267–4276. Physiol., 43, 1445–1455.
1162
152 Hambraeus, G. and Nyberg, N. (2005) 160 Matsushima, A., Sato, Y., Otsuka, M.,
Enzymatic hydrogenation of trans-2- Watanabe, T., Yamamoto, H., and Hirata,
nonenal in barley. J. Agric. Food Chem., 53, T. (2008) An enone reductase from
8714–8721. Nicotiana tabacum: cDNA cloning,
153 Ringer, K.L., McConkey, M.E., Davis, expression in Escherichia coli, and
E.M., Rushing, G.W., and Croteau, R. reduction of enones with the
(2003) Monoterpene double bond recombinant proteins. Bioorg. Chem., 36,
reductases of the ()-menthol 23–28.
biosynthetic pathway: isolation and 161 Hirata, T., Matsushima, A., Sato, Y.,
characterization of cDNAs encoding Iwasaki, T., Nomura, H., Watanabe, T.,
()-isopiperitenone reductase and Toyoda, S., and Izumi, S. (2009)
( þ )-pulegone reductase of peppermint. Stereospecific hydrogenation of the C¼C
Arch. Biochem. Biophys., 418, 80–92. double bond of enones by Escherichia coli
154 Raab, T., Lopez-Raez, J.A., Klein, D., overexpressing an enone reductase of
Caballero, J.L., Moyano, E., Schwab, W., Nicotiana tabacum. J. Mol. Catal. B:
and Munoz-Blanco, J. (2006) FaQR, Enzym., 59, 158–162.
required for the biosynthesis of the 162 Doorn, J.A., Maser, E., Blum, A., Claffrey,
strawberry flavor compound 4-hydroxy- D.J., and Petersen, D.R. (2004) Human
2,5-dimethyl-3(2H)-furanone, encodes an carbonyl reductase catalyzes reduction of
enone oxidoreductase. Plant Cell, 18, 4-oxonon-2- enal. Biochemistry, 43,
1023–1037. 13106–13114.
155 Kurata, A., Kurihara, T., Kamachi, H., and 163 Forrest, G.L. and Gonzalez, B. (2000)
Esaki, N. (2005) 2-Haloacrylated Carbonyl reductase. Chem.-Biol. Interact.,
reductase, a novel enzyme of the medium 129, 21–40.
chain dehydrogenase/reductase 164 J€ornvall, H., Persson, B., Krook, M.,
superfamily that catalyzes the reduction Atrian, S., Gonzalez-Duarte, R., Jeffery, J.,
of a carbon-carbon double bond of and Ghosh, D. (1995) Short-chain
unsaturated organohalogen compounds. dehydrogenases/reductases (SDR).
J. Biol. Chem., 280, 20286–20291. Biochemistry, 34, 6003–6013.
156 Ensor, C.M., Zhang, H., and Tai, H.-H. 165 Tanaka, N., Nonaka, T., Nakamura, K.T.,
(1998) Purification, cDNA cloning and and Hara, A. (2001) SDR: structure,
expression of 15-oxoprostaglandin 13- mechanism of action, and substrate
reductase from pig lung. Biochem. J., 330, recognition. Curr. Org. Chem., 5, 89–111.
103–108. 166 Ghosh, D., Sawicki, M., Pletnev, V.,
157 Thorn, J.M., Barton, J.D., Dixon, N.E., Erman, M., Ohno, S., Nakajin, S., and
Ollis, D.L., and Edwards, K.J. (1995) Duax, W.L. (2001) Porcine carbonyl
Crystal structure of Escherichia coli reductase – structural basis for a
QOR quinone oxidoreductase complexed functional monomer in short chain
with NADPH. J. Mol. Biol., 249, dehydrogenases/reductases. J. Biol.
785–799. Chem., 276, 18457–18463.
158 Dick, R.A. and Kensler, T.W. (2004) The 167 Burda, E., Kraußer, M., Fischer, G.,
catalytic and kinetic mechanisms of Hummel, W., M€ uller-Uri, F., Kreis, W.,
NADPH-dependent alkenal/one and Gr€oger, H. (2009) Recombinant D4,5-
oxidoreductase. J. Biol. Chem., 279, steroid 5b-reductases as biocatalysts for
17269–17277. the reduction of activated C¼C-double
159 Kurata, A., Fujita, M., Mowafy, A.M., bonds in monocyclic and acyclic
Kamachi, H., Kurihara, T., and Esaki, N. molecules. Adv. Synth. Catal., 351,
(2008) Production of (S)-2- 2787–2790.
chloropropionate by asymmetric 168 Hall, M., Stueckler, C., Ehammer, H.,
reduction of 2-chloroacrylate with 2- Pointner, E., Oberdorfer, G., Gruber, K.,
haloacrylate reductase coupled with Hauer, B., Stuermer, R., Kroutil, W.,
glucose dehydrogenase. J. Biosci. Bioeng., Macheroux, P., and Faber, K. (2008)
105, 429–431. Asymmetric bioreduction of C¼C bonds
References j 1163
using enoate reductases OPR1, OPR3 175 Swiderska, M.A. and Stewart, J.D.
and YqjM: enzyme-based stereocontrol. (2006) Asymmetric bioreductions of
Adv. Synth. Catal., 350, 411–418. b-nitroacrylates as a route to
169 Hall, M., Stueckler, C., Hauer, B., b2-amino acids. Org. Lett., 8,
Stuermer, R., Friedrich, T., Breuer, M., 6131–6133.
Kroutil, W., and Faber, K. (2008) 176 Toogood, H.S., Fryszkowska, A., Hare, V.,
Asymmetric bioreduction of activated Fisher, K., Roujeinikova, A., Leys, D.,
C¼C bonds using Zymomonas mobilis Gardiner, J.M., Stephens, G.M., and
NCR enoate reductase and old yellow Scrutton, N.S. (2008) Structure-based
enzymes OYE 1–3 from yeasts. Eur. J. insight into the asymmetric bioreduction
Org. Chem., 1511–1516. of the C¼C double bond of
170 Hall, M., Stueckler, C., Kroutil, W., a,b-unsaturated nitroalkenes by
Macheroux, P., and Faber, K. (2007) pentaerythritol tetranitrate reductase.
Asymmetric bioreduction of activated Adv. Synth. Catal., 350, 2789–2803.
alkenes using cloned 12-oxophytodienoate 177 Fryszkowska, A., Fisher, K., Gardiner,
reductase isoenzymes OPR-1 and OPR-3 J.M., and Stephens, G.M. (2008) Highly
from Lycopersicon esculentum (tomato): a enantioselective reduction of
striking change of stereoselectivity. Angew. b,b-disubstituted aromatic nitroalkenes
Chem. Int. Ed., 46, 3934–3937. catalyzed by Clostridium sporogenes. J. Org.
171 Kosjek, B., Fleitz, F.J., Dormer, P.G., Chem., 73, 4295–4298.
Kuethe, J.T., and Devine, P.N. (2008) 178 Mugford, P.F., Wagner, U.G., Jiang, Y.,
Asymmetric bioreduction of Faber, K., and Kazlauskas, R.J., (2008)
a,b-unsaturated nitriles and ketones. Enantiocomplementary enzymes:
Tetrahedron: Asymmetry, 19, 1403–1406. classification, molecular basis for their
172 M€ uller, A., Hauer, B., and Rosche, B. enantiopreference, and prospects for
(2006) Enzymatic reduction of the mirror-image biotransformations. Angew.
a,b-unsaturated carbon bond in citral. J. Chem. Int. Ed., 47, 8782–8793.
Mol. Catal. B: Enzym., 38, 126–130. 179 Padhi, S.K., Bougioukou, D.J., and
173 M€ uller, A., Hauer, B., and Rosche, B. Stewart, J.D., (2009) Site saturation
(2007) Asymmetric alkene reduction by mutagenesis of tryptophan 116 of
yeast old yellow enzymes and by a novel Saccharomyces pastorianus old yellow
Zymomonas mobilis reductase. Biotechnol. enzyme uncovers stereocomplementary
Bioeng., 98, 22–29. variants. J. Am. Chem. Soc., 131,
174 Hall, M., Hauer, B., Stuermer, R., Kroutil, 3271–3280.
W., and Faber, K. (2006) Asymmetric 180 Bougioukou, D.J., Kille, S., Taglieber, A.,
whole-cell bioreduction of an and Reetz, M.T., (2009) Directed
a,b-unsaturated aldehyde (citral): evolution of an enantioselective enoate-
competing prim-alcohol dehydrogenase reductase: testing the utility of iterative
and CC lyase activities. Tetrahedron: saturation mutagenesis. Adv. Synth.
Asymmetry, 17, 3058–3062. Catal., 351, 3287–3305.
j1165
28
Reductive Amination of Keto Acids
Werner Hummel and Harald Gr€oger
28.1
Introduction
Enzyme-catalyzed enantioselective reductive amination by NAD(P)H-dependent

amino acid dehydrogenases is a highly efficient way to synthesize enantiomerically
pure L- and D-amino acids from the corresponding a-keto acids. Scheme 28.1 shows
the general reaction scheme of this biotransformation. Major advantages of this
synthetic route are the straightforward approach without the need for protecting
groups in the substrate and the direct use of ammonia as nitrogen source. Notably,
such biotransformations have already found several applications on a technical scale
and certainly belong to one of the key technologies for producing (in particular non-
natural) amino acids, which are required in particular in the chemical and pharma-
ceutical industry.
Conceptually this biocatalytic reductive amination process consists of two enzy-
matic steps, which are illustrated in Scheme 28.2. The desired reductive amination
process is catalyzed by an amino acid dehydrogenase in combination with ammonia
as nitrogen source and the cofactor NAD(P)H as reducing agent. In this amino acid
dehydrogenase-catalyzed reductive amination of the a-keto acid, the reduced form of
the cofactor, NAD(P)H, is oxidized to NAD(P) þ . Owing to the high price and
molecular mass of the cofactor (thus making a stoichiometric use of this reducing
agent economically and ecologically unattractive) for preparative applications it is
necessary to couple this amino acid forming reaction with a second reaction for the
continuous in situ regeneration of the nicotinamide-type cofactor. By means of the
seconddehydrogenase,NAD(P) þ isre-transformedintoitsreducedform,NAD(P)H.
This in situ recycling enables the use of catalytic amounts of the cofactor. The
cosubstrate required in stoichiometric amount is typically a cheap reducing agent such
as, for example, formate or D-glucose, thus making the reductive amination process
economically attractive. When using formate, a formate dehydrogenase is required to
oxidize this cosubstrate to carbon dioxide, whereas when using glucose a glucose
dehydrogenase is applied to form D-gluconolactone, which after ring-open-
ing and neutralization ends up as sodium D-gluconate. Such transformations for
and Oliver May.
j 28 Reductive Amination of Keto Acids
1166
enzymatic
reductive
O NH2 NH2
amination
R CO2H R CO2H or R CO2H
α-keto acid L-amino acid D-aminoacid
Scheme 28.1 Synthesis of L- and D-amino acids via enzymatic reductive amination.
NH2 NH2
co-substrate or
NADP R CO2H R CO2H
L-amino acid D-amino acid
dehydrogenase amino acid

for cofactor dehydrogenase,
regeneration ammonia
O
oxidized
NAD(P)H R CO2H
co-substrate
α-keto acid
Scheme 28.2 Concept of enzymatic reductive amination under in situ cofactor regeneration.
cofactor regeneration represent irreversible steps and, thus, shift the equilibrium in the
direction of the products (see also Chapter 26) [1].
For the reductive amination step a broad range of amino acid dehydrogenases are
known, which are partly complementary to each other. A common feature of (most of)
these enzymes is the suitability for catalyzing both the reductive amination (when
NAD(P)H is provided) and the reverse reaction, namely, the oxidation of the amino
acid (when NAD(P) þ is provided; Scheme 28.3Þ. The direction of the reaction
(reductive versus oxidative mode) can be controlled by means of the corresponding
cofactor-regeneration method. Although this reaction is reversible, equilibrium with
a Keq in the range of 1014–1018 favors the amination direction [2].
enantioselective
biocatalytic
O NH2
reductive amination
R COOH + NAD(P)H + NH4+ R COOH + NAD(P)+
Scheme 28.3 General reaction scheme of a-amino acid dehydrogenases.
Currently 16 amino acid dehydrogenases are listed in the Enzyme Nomenclature

catalog (Table 28.1). Some members of this group show very limited substrate
specificity. For example, glutamate dehydrogenase only accepts a-ketoglutarate,
whereas leucine or valine dehydrogenases catalyze the synthesis of a broad range
Table 28.1 NAD(P)-dependent amino acid dehydrogenases listed by the Nomenclature Committee of the International Union of Biochemistry and Molecular Biology
(NC-IUBMB) (updated 2011).
Enzyme EC number Organisms Main amino acid Reference

substrate
Alanine dehydrogenase 1.4.1.1 Bacillus sp. L-Alanine

Glutamate dehydrogenase (NAD(P) þ -dependent 1.4.1.2–1.4.1.4 Ubiquitous L-Glutamate
Serine 2-dehydrogenase 1.4.1.7 Petroselinum crispum L-Serine [52]
Valine dehydrogenase 1.4.1.8 Streptomyces cinnamonensis ¼ L-Valine, aliphatic aa [115]
Leucine dehydrogenase 1.4.1.9 Bacillus sp., Geobacillus stear- L-Leucine and other ali- [20, 21, 23–27]
othermophilus, Thermoactino- phatic aa
myces intermedius, Corynebacteri-
um pseudodiphtheriticum, Clos-
tridium thermoaceticum
Glycine dehydrogenase 1.4.1.10 Mycobacterium tuberculosis Glycine [53, 54]
L-erythro-3,5-Diaminohexanoate dehydrogenase 1.4.1.11 Brevibacterium sp. Clostridium sp. L-erythro-3,5- [55, 56]
Diaminohexanoate
2,4-Diaminopentanoate dehydrogenase 1.4.1.12 Clostridium sticklandii 2,4-Diaminopentanoate [57, 58]
Glutamate synthase 1.4.1.13 (NADPH)
1.4.1.14 (NADH)
Lysine dehydrogenase 1.4.1.15
Diaminopimelate dehydrogenase 1.4.1.16
N-Methylalanine dehydrogenase 1.4.1.17
Lysine 6-dehydrogenase 1.4.1.18
Tryptophan dehydrogenase 1.4.1.19
Phenylalanine dehydrogenase 1.4.1.20 Brevibacterium sp. [33]
Rhodococcus sp. [34]
Rhodococcus maris [35]
(Continued )
28.1 Introduction
j1167
1168
Enzyme EC number Organisms Main amino acid Reference

substrate
Sporosarcina ureae [37]

Bacillus sphaericus [38]
Bacillus badius [39]
Nocardia sp. [40]
Thermoactinomyces intermedius [41]
Aspartate dehydrogenase 1.4.1.21
of aliphatic amino acids. A range of aromatic a-keto acids, on the other hand, are
preferred substrates of phenylalanine dehydrogenase. From a synthetic perspective,
in particular leucine dehydrogenases as well as phenylalanine dehydrogenases
turned out to have a broad substrate spectrum and have been widely used as
biocatalysts in many efficient reductive amination processes.
Regarding the biochemical relevance of amino acid dehydrogenases, the reactions
of these enzymes represent an important link between the chemistry of organic and
inorganic nitrogen compounds. The degradation of amino acids by amino acid
dehydrogenases starts with the removal of the amino group by oxidative deamination
followed by metabolism of the carbon skeleton. Contrary to other enzymatic routes,
for example, via transaminases or deaminases, this degradation is coupled with the
production of reduced nicotinamide coenzymes, which subsequently can be used to
generate ATP and to support various energy-consuming processes.
For synthetic applications, amino acid dehydrogenases are used in the asymmetric
reductive amination of a-keto acids leading to L-amino acids, practically in a highly
enantioselective manner. Suitable enzymes are known for a broad range of substrates
and are found ubiquitously in nature. In contrast, no naturally occurring NAD(P) þ -
dependent D-enantioselective enzymes with broad substrate range are known so far.
A range of D-amino acid dehydrogenases, which require FAD or artificial redox dyes
like methylene blue or 2,6-dichloroindophenol as electron acceptors, are found in
some bacteria such as the FAD-dependent D-arginine dehydrogenase from Pseudo-
monas aeruginosa [3], D-amino acid dehydrogenase from Helicobacter pylori [4], and the
membrane-bound dye-dependent enzymes from P. aeruginosa [5] or Escherichia
coli. [6]. All these enzymes catalyze the oxidative deamination of D-amino acids only
and no reductive amination of a-keto acids turned out to be possible. More recently,
however, the first example of a NAD(P)H-dependent D-amino acid dehydrogenase
with a broad substrate range has been developed by Codexis researchers by means of
rational and random mutagenesis [7].
For industrial purposes, amino acid dehydrogenases are well suited because of
their high catalytic activity, excellent enantioselectivities, and good stability under
operational conditions. Therefore, some industrial applications of amino acid
dehydrogenases for the large-scale manufacture of enantiomerically pure unnatural
amino acids have been already demonstrated.
Some amino acid dehydrogenases such as glutamate, leucine, or phenylalanine
dehydrogenase are well established meanwhile as catalysts for preparative applica-
tions. Besides these enzymes some other L-amino acid dehydrogenases bearing a
synthetic potential are characterized biochemically. All these enzymes show a strict
limitation concerning the amine compound to be coupled: Amino acid dehydro-
genases only accept ammonia as the nitrogen source, but there are some other
enzymes that react with amines instead of ammonia in a similar NAD(P) þ -depen-
dent reductive manner. The enzyme N-methyl-L-amino acid dehydrogenase isolated
from Pseudomonas putida uses alkylamines instead of ammonia. In addition, a group
of enzymes, opine dehydrogenases, catalyzes reactions in a similar fashion as amino
acid dehydrogenases but uses an amino acid instead of ammonia for the NAD(P)H-
dependent reductive condensation with an a-keto acid, for example, pyruvate. This
1170
kind of reaction produces N-carboxyalkyl-amino acids. Scheme 28.4 summarizes the

processes with these three groups of reductive aminating enzymes.
O L-amino acid NH 2
dehydrogenase
R1 COOH + NH3 R1 COOH
CH3
O N-methyl-L-amino acid HN
dehydrogenase
R1 COOH + CH 3-NH 2 R1 COOH
COOH
opine
O NH 2 HN R2
dehydrogenase
R1 COOH + R2 COOH R1 COOH
Scheme 28.4 Summary of different kinds of enzymes catalyzing reductive amination of a-keto
acids using different kinds of nitrogen compounds.
The L-amino acid dehydrogenases have been studied extensively and their bio-
chemical data and examples of their applications have been comprehensively
reviewed. The biochemistry and enzymology of these enzymes are summarized in
detail by Brunhuber and Blanchard [8], while preparative applications of amino acid
dehydrogenases, especially regarding their use for the synthesis of pharmaceutical
intermediates, are reviewed in several publications by Hanson and Patel [9–13] and
recently by Zhu and Hua [14]. This chapter gives an overview of the biochemical
properties of amino acid dehydrogenases, focusing on enzymes suited for synthetic
purposes and summarizing (particularly recent) examples of their preparative
applications for the synthesis of enantiomerically pure non-natural amino acids.
28.2
Biochemical Properties of Enzymes Catalyzing Reductive Amination Reactions
28.2.1
L-Amino Acid Dehydrogenases
In more detail the reductive amination of a-keto acids can be considered as a two-step
mechanism, which is exemplified in Scheme 28.5 for the reductive amination of
a-ketoglutarate (1). First, ammonia reacts with the keto group of 1 to form an imino
intermediate of type 2 (step A); second, the enantioselective step of reducing the
28.2 Biochemical Properties of Enzymes Catalyzing Reductive Amination Reactions j1171
O NH2
GluDH
HOOC COOH + NAD(P)H + NH4+ HOOC COOH + NAD(P)+
α-ketoglutarate (1) L-glutamate (L-3)
+ NH3 NAD(P)+
step A NH Step B
HOOC COOH NAD(P)H

2
Scheme 28.5 Two-step mechanism of reductive amination; exemplified for a glutamate

dehydrogenase (GluDH)-catalyzed reaction.
imine 2 with NAD(P)H as a coenzyme to the amine functionality of L-glutamate (L-3,

step B). The upper line in Scheme 28.5 shows the overall reductive amination
reaction of 1. As mentioned in Section 28.1, in general this reductive amination
process is reversible, and when providing the oxidized form of the cofactor NAD(P) þ
by means of a suitable in situ cofactor-regeneration (e.g., in the presence of an NAD(P)
H-oxidase) the amino acid dehydrogenase catalyzes the transformation of the amino
acid into the corresponding a-keto acid, according to the reverse direction to the one
shown in Scheme 28.5 (Scheme 28.5 shows only the direction of the reductive
amination).
The complete catalytic mechanism with participation of the amino acids forming
the active site was thoroughly studied at first in 1992 by the group of Rice for
glutamate dehydrogenase from Clostridium symbiosum [15], and was confirmed later
on for phenylalanine dehydrogenase from Rhodococcus sp. by the group of
Blanchard [16].
The synthetic applications of amino acid dehydrogenases are determined by their
distinct substrate specificity. So far most preparative applications have been carried
out with either leucine or phenylalanine dehydrogenase. Both enzymes accept a
broad range of a-keto acid substrates, but cover different and complementary
structures of the side chain-substituent. A broad range of natural and non-natural
aliphatic and aromatic a-amino acids are accessible. The use of a leucine dehydro-
genase is highly useful for the synthesis of a broad range of aliphatic amino acids with
linear and highly branched side chains, whereas phenylalanine dehydrogenase is the
catalyst of choice for the preparation of amino acids with aromatic and bulky side
chains. Thus, the biochemical data of these two enzymes are described in the next
subsections in detail, followed by the characterization of other amino acid
dehydrogenases.
28.2.1.1 Leucine Dehydrogenase

Leucine dehydrogenases are known from several Bacillus and Clostridium strains. For
Bacillus strains leucine dehydrogenase is an important key enzyme during the
germination of spores [17, 18]. The oxidative deamination of aliphatic amino acids
like leucine during germination supplies the cells in one step only with NADH,
which on the other hand can be regarded as a precursor of ATP, the main energy
1172
intermediate in living organisms. Additionally, branched-chain a-keto acids as

precursors of branched-chain fatty acids are available by oxidative deamination of
leucine, isoleucine, and valine.
The first leucine dehydrogenases intended for synthetic applications were isolated
by Ohshima and Soda from Bacillus sphaericus and B. stearothermophilus (renamed as
Geobacillus stearothermophilus) [19–22] and by Hummel and Sch€ utte from Bacillus
sphaericus and B. cereus [23, 24]. Meanwhile, leucine dehydrogenases have been
purified and characterized from strains of other genera such as Clostridium thermo-
aceticum [25], Corynebacterium pseudodiphtheriticum [26], and Thermoactinomyces
intermedius [27, 28]. All these enzymes are available in recombinant form and can
be produced by overexpression, for example, the recombinant enzymes from B.
cereus [29, 30] and B. stearothermophilus overexpressed in E. coli [19, 31].
All of the studied leucine dehydrogenases are NAD þ -dependent. Although
leucine dehydrogenases from different sources show different sequences and
quaternary structures, the substrate specificities are very similar. Leucine is the
preferred amino acid substrate but other straight- and branched-chain aliphatic
amino acids (and the corresponding a-keto acids thereof for the reductive mode) are
accepted with good activities. An extensive screening of enzymatic activity toward the
reductive aminating capability of four leucine dehydrogenases from different Bacillus
strains was carried out with natural and non-natural aliphatic a-keto acids by Krix
et al. (Table 28.2) [32].
These data show that leucine dehydrogenases from different Bacillus strains
display a similar substrate scope and activity pattern. Highest activities were found
for a-keto acids of the natural occurring amino acids leucine, isoleucine, and valine,
whereas structural deviations thereof result in lower (but – at least in part –
preparatively still useful) activities. Comparison of the unbranched a-keto acids
reveals the highest activity for C5 (2-oxopentanoid acid) and significantly lower values
for shorter (C4) or longer (C6) side chains. Branched side chains with one methyl
group at C3 seem to result in higher activities.
28.2.1.2 Phenylalanine Dehydrogenase

Phenylalanine dehydrogenase acting on a-keto acids with an aromatic side chain can
be regarded as a counterpart to leucine dehydrogenase (converting aliphatic a-keto
acids). Phenylalanine dehydrogenases were first described by Hummel et al., being
isolated from Brevibacterium sp. [33], followed by the isolation of a more active and
stable enzyme from Rhodococcus sp. M4 [34]. Both organisms were obtained by
enrichment from soil samples with phenylalanine as the sole carbon and nitrogen
source. Further enzymes have been isolated and characterized from Sporosarcina
ureae [35, 36], Bacillus sphaericus [37], B. badius [38], Rhodococcus maris [39, 40], and
Thermoactinomyces intermedius [41].
All phenylalanine dehydrogenases are NAD þ -specific, and very selective for
phenylalanine. Their (somewhat limited) substrate spectra for aromatic or aliphatic
a-keto acids are summarized in Table 28.3 for seven enzymes of different sources. As
well as phenylpyruvic acid, the corresponding a-keto acids of tyrosine and methi-
onine are further substrates with good activity. From a comparison of the nucleotide
Table 28.2 Relative activity values for four different leucine dehydrogenases from Bacillus strains. Activities with the different aliphatic keto acids are given
relative to 2-oxo-4-methylpentanoic acid; the absolute specific activities of the four enzyme samples with 2-oxo-4-methylpentanoic acid are mentioned in the
table headings. Activity measurements were carried out photometrically at pH 8.5.
a-Keto acid Source of leucine dehydrogenase
Bacillus stearothermophilus Bacillus cereus Bacillus sphaericus Bacillus sp. DSM 730
(204 U mg1) (19 U mg1) (5.6 U mg1)a) (1.4 U mg1)a)
COOH
O ¼ 100 ¼ 100 ¼ 100 ¼ 100

2-Oxo-4-methyl-
pentanoic acid
COOH
O 48.0 74.0 66.0 94.0

2-Oxobutyric acid
COOH
O 113.0 152.0 205.0 380.0

2-Oxo-3-methyl-
butyric acid
COOH
28.2 Biochemical Properties of Enzymes Catalyzing Reductive Amination Reactions
O 31.0 74.0 51.0 12.0

2-Oxo-3,3-dimethyl- (Continued )
butyric acid
j1173
1174
a-Keto acid Source of leucine dehydrogenase
Bacillus stearothermophilus Bacillus cereus Bacillus sphaericus Bacillus sp. DSM 730
(204 U mg1) (19 U mg1) (5.6 U mg1)a) (1.4 U mg1)a)
COOH
O 63.0 81.0 102.0 108.0

2-Oxopentanoic acid
COOH
O 110.0 114.0 88.0 270.0

2-Oxo-3-methyl-
pentanoic acid
COOH
O 2.0 11.0 5.0 2.0

2-Oxo-3,3-dimethyl-
pentanoic acid
COOH
O 7.0 14.0 11.0 12.0

2-Oxo-4,4-dimethyl-
pentanoic acid
COOH
O 15.0 63.0 75.0 19.0

2-Oxohexanoic acid
COOH
O 22.0 19.0 N.d. N.d.

2-Oxo-4-methyl-
hexanoic acid
COOH
O
1.0 11.0 N.d. N.d.
2-Oxo-4-ethyl-
hexanoic acid
COOH
O 0.5 1.2 0.2 N.d.

2-Oxo-4,4-dimethyl-
hexanoic acid
COOH
O 0.8 0.3 N.d. N.d.

2-Oxo-5,5-dimethyl-
hexanoic acid
COOH
O
0.8 0.1 0.3 N.d.
2-Oxo-3-cyclohexyl-
propanoic acid
a) N.d. ¼ not determined.
28.2 Biochemical Properties of Enzymes Catalyzing Reductive Amination Reactions
j1175
1176
Table 28.3 Relative reaction rate for reductive amination of a-keto acids with phenylalanine dehydrogenases from different sources.
Keto acid Rhodococcus Sporosarcina Bacillus Bacillus badius Rhodococcus Nocardia 239 Thermoactinomyces
sp. M4 ureae sphaericus maris intermedius
Phenylpyruvate 100 100 100 100 100 100 100

p-OH-Phenyl-pyruvate 5 24 136 53 91 28 —
Indolepyruvate 3 1 0 — 5 54 —
a-Oxo-4-methyl-mercaptobutyrate 33 27 11 16 9 — 14
a-Oxoisovalerate — 2 6 13 — — 6
a-Oxoisocaproate — 13 8 — — 240 —
a-Oxovalerate — 9 6 12 — — —
a-Oxocaproate — 32 0 31 9 — —
Reference [34] [37] [37] [38] [39] [40] [41]
and amino acid sequences of the enzymes from T. intermedius, B. sphaericus and S.
ureae, Takada et al. suggested that the enzymes are composed of two domains,
wherein the N-terminal part is responsible for binding of the amino acid and the C-
terminal domain binds the coenzyme [42].
28.2.1.3 Glutamate Dehydrogenase

Glutamate dehydrogenase plays a central role in nearly every organism during the
assimilation of ammonia by catalyzing the interconversion of a-ketoglutarate (1) and
þ þ
L-glutamate (L-3) using NAD or NADP and the corresponding reduced cofactors,
respectively (Scheme 28.6).
O NH 2
GluDH
+ NAD(P)H + NH4 + + NAD(P) +
HOOC COOH HOOC COOH
α-ketoglutarate (1) L-glutamate (L-3)
Scheme 28.6 Reaction scheme for reductive amination of a-ketoglutarate (1) with a glutamate
dehydrogenase (GluDH).
Glutamate dehydrogenases from higher eukaryotes accept both coenzymes

(EC 1.4.1.3), whereas the enzymes from lower eukaryotes and bacteria use either
NAD þ (1.4.1.2) or NADP þ (1.4.1.4). Usually, NADP þ -dependent enzymes are used
for anabolic processes, that is, the formation of glutamate, and NAD þ -specific
enzymes are involved in the degradation of glutamate. Bacteria usually express only a
single enzyme, which can be either NAD þ - or NADP þ -dependent. Higher fungi
express NAD þ - and NADP þ -specific glutamate dehydrogenases, whereas lower
fungi display only NAD þ -dependent ones. Owing to the limited substrate spectrum
of glutamate dehydrogenases, research interest so far has focused more on bio-
chemical properties than on preparative applications and, currently, structural
determinants for the thermal stability of glutamate dehydrogenases analyzing
enzymes from thermophilic organisms [43–47] and requirements for their substrate
specificity are under investigation [48–50].
Glutamate dehydrogenases generally are highly specific for L-glutamate or a-keto-
glutarate, respectively. The synthetic approach to L-glutamate, however, is based on
fermentation rather than on a biotransformation starting from a-ketoglutarate [51].
In organic synthesis only two preparative applications with glutamate dehydro-
genases have been described so far, namely, the reductive amination of a-keto-6-
hydroxyhexanoic acid to L-6-hydroxy-norleucine and the use for cofactor regeneration
in the enzymatic synthesis of non-proteinogenic amino acids in combination with a
D-amino acid aminotransferase, glutamate racemase, and formate dehydrogenase as
further biocatalysts (Section 28.3.4).
28.2.1.4 Further Amino Acid Dehydrogenases

A range of amino acid dehydrogenases listed in Table 28.1 is only poorly character-
ized. Serine dehydrogenase, a NAD þ -dependent enzyme isolated from Petroselinum
crispum, catalyzes the reductive amination of hydroxypyruvate [52]. Glycine
1178 j 28 Reductive Amination of Keto Acids
dehydrogenase is obtained from Mycobacterium tuberculosis; it catalyzes the NAD þ -
dependent oxidative deamination of glycine to glyoxylate [53, 54]. 3,5-Diaminohex-
anoate dehydrogenase, an enzyme found in Brevibacterium and Clostridium, is
involved in the metabolism of lysine. It catalyzes the NAD þ -dependent oxidative
deamination of L-erythro-3,5-diaminohexanoate to 5-amino-3-ketohexanoate
(Scheme 28.7, Reaction 1) [55, 56]. This is the only amino acid dehydrogenase
known to react with a b-amino acid. 2,4-Diaminopentanoate dehydrogenase,
obtained from Clostridium sticklandii, takes part of the oxidation of ornithine, which
is converted into 2-amino-4-ketopentanoate via 2,4-diaminopentanoic acid [57, 58].
Notably, this enzyme is one of the few amino acid dehydrogenases that requires an
a-amino acid as the substrate but reacts with the v-amino group (Scheme 28.7,
Reaction 2a). More slowly, the enzyme also acts on L-2,5-diaminohexanoate (L-8),
forming 2-amino-5-ketohexanoate (9, Scheme 28.7, Reaction 2b).
3,5-diaminohexanoate
NH 2 NH 2 O NH 2
dehydrogenase
HOOC + NAD + HOOC + NADH + NH4 + (1)
CH 3 CH 3
L-er ythr o-3,5-diaminohexanoate (S)-5-amino-3-ketohexanoate
(L-4) 5
NH 2 NH 2 2,4-diaminopentanoate NH 2 O
dehydrogenase
HOOC CH3 + NAD + HOOC CH 3 + NADH + NH4 + (2a)
L-2,4-diaminopentanoate L-2-amino-4-ketopentanoate
(L-6) 7
2,4-diaminopentanoate
NH2 dehydrogenase NH 2
CH 3 + NAD + CH 3 + NADH + NH4 + (2b)
HOOC HOOC
NH2 O
L-2,5-diaminohexanoate L-2-amino-5-ketohexanoate
(L-8) 9
Scheme 28.7 Oxidative deamination of diaminopentanoate to give L-2-amino-4-

diamino acids using diaminoalkanoate ketopentanoate catalyzed by 2,4-
dehydrogenases: (1) conversion of L-erythro-3,5- diaminopentanoate dehydrogenase from C.
diaminohexanoate into 5-amino-3- sticklandii and its side reaction (2b) converting L-
ketohexanoate catalyzed by 3,5- 2,5-diaminohexanoate into L-2-amino-5-
diaminohexanoate dehydrogenase from ketohexanoate.
Clostridium strains; (2a) main reaction of L-2,4-
Two different types of amino acid dehydrogenases catalyze the oxidative deam-
ination of L-lysine (Scheme 28.8). L-Lysine 6-dehydrogenase (EC 1.4.1.18) reacts with
the e-amino group of L-lysine (L-10) to form L-2-aminoadipate-6-semialdehyde (L-11)
(Scheme 28.8, pathway A). This compound in turn cyclizes non-enzymatically to L-
D1-piperideine-6-carboxylate (L-12), which is an intermediate for the synthesis of
optically active pipecolic acid. L-Lysine 2-dehydrogenase (EC 1.4.1.15), on the other
hand, catalyzes the oxidative deamination of the a-amino group (Scheme 28.8,
pathway B). To date, there has been only one report of the latter enzyme isolated from
human liver [59]. Lysine 6-dehydrogenase was first isolated from Agrobacterium
NH2
pathway A pathway B
H2N COOH
L-Lysine (L-10)
NAD+ NAD+
L-Lysine 6-dehydrogenase L-Lysine 2-dehydrogenase
NADH NADH
+ NH4+ + NH4+
H NH2 O
O COOH H2N COOH

2-aminoadipate-6-semialdehyde 2-keto-6-amino caproic acid
L-11) 13
- H2O - H2O
N COOH N COOH
∆1-piperideine-6-carboxylate ∆1-piperideine-2-carboxylate
(L-12) 14
Scheme 28.8 Reaction scheme for oxidative deaminations of L-lysine (L-10) with lysine
dehydrogenases and subsequent cyclization reactions.
tumefaciens. Meanwhile, a highly stable enzyme was isolated and characterized from
the thermophile Geobacillus stearothermophilus. The specific activity of the purified
recombinant enzyme was found to be 7.81 U mg1 [60]. This enzyme could be used
for the production of L-pipecolic acid and 2-aminoadipic acid from L-lysine.
Diaminopimelate dehydrogenase (DAPDH) is a most unusual enzyme, oxidizing
the D-amino group of meso-2,6-diaminopimelate (meso-15) and forming L-2-amino-6-
ketopimelate (L-16), which undergoes spontaneous dehydration to L-D1-piperideine-
2,6-dicarboxylate (L-17,Scheme 28.9) [61].
diaminopimelate
dehydrogenase HOOC COOH - H2O
HOOC COOH
(spontaneous) HOOC N COOH
NH2 NH2 NADP+ O NH2
NADPH + NH4+
meso-2,6-diaminopimelic acid L-2-amino-6-oxopimelic acid L-∆1-piperideine-
(meso-15) (L-16) 2,6-dicarboxylate
(L-17)
Scheme 28.9 Reaction pathway of meso-a,e-diaminopimelate D-dehydrogenase.
This enzyme has been found in several lysine overproducing bacteria where it is
involved in the biosynthesis of lysine. It was isolated and biochemically characterized
from Bacillus sphaericus [62], Brevibacterium sp. [63], or Corynebacterium glutamicum [64].
DAPDH is specific for NADP þ and active almost exclusively towards meso-2,6-
1180
diaminopimelate; neither L,L- nor D,D-isomers nor other alkyl and aryl D-amino
acids were converted [65].
28.2.2
D-Amino Acid Dehydrogenases
BioCatalytics researchers succeeded in creating a D-enantioselective amino acid

dehydrogenase with a broad substrate range by combining rational and random
mutagenesis for DAPDH from C. glutamicum [7]. This mutant enzyme accepts D-
amino acids with an aliphatic or aromatic side chain or cyclic D-amino acids such as D-
cyclohexylalanine. Activity for the reductive amination was found to be in the range of
0.11 (D-phenylalanine) to 7.8 U mg1 (D-2-aminooctanoate), and afforded D-amino
acids with a high enantiomeric excess of >99% e.e. in most cases.
28.2.3
N-Methyl-L-amino Acid Dehydrogenase
N-Methyl-a-amino acids are components of several biologically active natural pep-

tides, especially depsipeptides such as cyclosporine, vancomycin, enniatin, and
others. N-Methylated peptides are synthesized by specific N-methyltransferases
using S-adenosylmethionine as the methyl group donor. The methyl group of free
N-methyl amino acids, on the other hand, is derived from methylamine, which is
inserted by reaction with an a-keto acid.
Enzymes catalyzing such reactions belong to the family of N-methyl-L-amino acid
dehydrogenases. The first enzyme of this group was purified and characterized in
1975 by the group of Wagner [66]. They isolated an N-methylalanine dehydrogenase
from a Pseudomonas strain that catalyzes the NADPH-dependent reductive amina-
tion of pyruvate with methylamine. The biochemical characterization reveals that the
enzyme is highly specific for methylamine. Notably, ammonia, ethylamine, and
dimethylamine could not be reacted with the a-keto acid. Among several a-keto acids
tested, only oxaloacetate (64% activity related to the one of pyruvate) and a-ketobu-
tyrate (14% activity) can be used as substrates. The specific activity of this enzyme
with pyruvate and methylamine was 0.37 U mg1 for the purified protein, which
exhibits a high Km (75 mM) for methylamine.
Later on, the group of Esaki successfully screened bacterial strains for their ability
to form N-methylphenylalanine from phenylpyruvate and methylamine. The gene
coding for a N-methyl-L-amino acid dehydrogenase from a Pseudomonas putida strain
was cloned and sequenced followed by purification and characterization of the
enzyme [67]. The best a-keto acid substrate was pyruvate, but a-oxohexanoate
(52% activity related to the one of pyruvate), phenylpyruvate (30% activity), a-oxo-
butyrate (30% activity), fluoropyruvate (27% activity), a-oxovalerate (16% activity),
a-oxoisocaproate (9.1% activity), a-oxooctanoate (7.6% activity), and hydroypyruvate
(5.8% activity) were accepted, too. The enzyme depends on NADPH as a cofactor, and
the specific activity with pyruvate and NADPH was 42 U mg1. Concerning the
amine moiety, methylamine could be substituted by ethylamine with 4.4% activity,
28.3 Synthetic Applications of Enzymes Catalyzing Reductive Amination j1181
whereas no activity was found with ammonia. The enzyme was used for the synthesis
of N-methyl-L-phenylalanine and cyclic amino acids (Section 28.3.6).
28.2.4
Opine Dehydrogenases
Opine dehydrogenases catalyze the NAD(P)H-dependent reductive amination of

a-keto acids (e.g., pyruvate, 19) with an amino acid, furnishing the N-carboxylalkyl-
amino acids, the so-called called opines. Scheme 28.10 shows a representative example.
opine H
NH2 O CH3 dehydrogenase N CH3
+ + NADH + H+ + NAD +
COOH COOH CO2 H CO 2H
L-18 19 20
Scheme 28.10 Opine dehydrogenase-catalyzed reductive condensation of (S)-phenylalanine

(L-18) and pyruvate (19) to form N-[1-(R)-(carboxyl)ethyl]-(S)-phenylalanine (L-20).
The first opine was isolated from the muscle tissue of Octopus octopodia. The first
opine dehydrogenase, namely, the D-octopine dehydrogenase, catalyzing the forma-
tion of D-octopine by the reductive condensation of pyruvate and L-arginine, was
described in 1959 using muscle tissues from marine invertebrates [68] followed by its
purification from muscles from the scallop Pecten maximus (pilgrims scallop) in
1969 [69]. Meanwhile, several other opine dehydrogenases have been found in tissues
of marine and limnic invertebrates and also in some soil bacteria like Agrobacterium
tumefaciens [70] or Arthrobacter species [71, 72]. They all catalyze the reductive
coupling of pyruvate with one of the following amino acids as amine-donor (the
names of the associated amino acid dehydrogenases are given in parentheses):
glycine (D-strombine dehydrogenase), L-alanine (meso-alanopine dehydrogenase),
b-alanine (b-alanopine dehydrogenase), and taurine (tauropine dehydrogenase).
Regarding the potential of opine dehydrogenases for biotechnological applications
the resulting secondary amine dicarboxylic acids are useful chiral intermediates of
angiotensin converting enzyme (ACE) inhibitors, which are formed with high
enantioselectivity by opine dehydrogenases. A first application was published by
the group of Asano using the enzyme from Arthrobacter sp. (Section 28.3.7).
28.3
Synthetic Applications of Enzymes Catalyzing Reductive Amination
28.3.1
Introduction and General Remarks
Preparative applications of amino acid dehydrogenases are of particular interest for

the synthesis of non-natural amino acids. A key feature of all of these processes is the
1182
use of a (mostly enzymatic) cofactor-regeneration method (as illustrated in

Scheme 28.2), thus allowing the use of the costly cofactor in catalytic amounts only.
For natural amino acids, fermentative processes very often based on Corynebac-
terium glutamicum are well established and typically turned out to be the method of
choice for such molecules [51]. Strain improvement by mutation or metabolic
engineering led meanwhile to production strains that can produce high yields of
proteinogenic amino acids by consumption of cheap carbon and nitrogen sources.
Owing to the costs for enzyme production, coenzyme regeneration, and synthesis of
a-keto acid substrates, enzyme-catalyzed reductive amination processes have been
mainly applied for non-natural amino acids.
Most amino acid dehydrogenases prefer NADH, not NADPH, as the cofactor. For
preparative applications on laboratory and technical scales, the use of NAD þ as the
coenzyme is significantly preferred over NADP þ due to the increased stability, more
favorable price, and broader availability of cofactor-regeneration methods. Owing to
the high costs of the nicotinamide cofactors their stoichiometric use is not acceptable
from an economical point of view. For these reasons preparative (and consequently all
industrial) applications require an efficient in situ regeneration of the consumed
cofactors.
In recent years numerous concepts for the regeneration of nicotinamide cofactors
have been described, namely, enzymatic, chemical, photochemical, or electrochem-
ical methodologies [1]. Enzymatic ways of regenerating nicotinamide cofactors are
the best investigated and most convenient strategies. They show high selectivity for
the formation of the active form of the cofactor. Important parameters must be
considered for an appropriate regeneration system: the method should be inexpen-
sive, the regeneration system must be stable over a long period of time, and there
should be no cross reactions between reagents and products of the desired reaction
with the compounds involved in the cofactor regeneration. Furthermore, formation
of the product should be thermodynamically as well as kinetically preferred.
In principle, there are two main methods to regenerate nicotinamide cofactors
enzymatically. The first approach requires a second enzyme as well as a second
substrate for regeneration. The second method, substrate-coupled regeneration,
utilizes the same enzyme for the formation of the desired product as well as for
cofactor recycling. For amino acid dehydrogenases, only the first method using a
second enzyme can be applied because no suitable cheap second amino acid, which
could be used as the cosubstrate (in a similar fashion to isopropanol in the case of
alcohol dehydrogenases), is available so far.
Most of the described preparative applications of amino acid dehydrogenases
make use of either formate/formate dehydrogenase or glucose/glucose dehydroge-
nase as cofactor regeneration systems. Formate dehydrogenase has the advantage
that the substrate formate is very cheap and the by-product is carbon dioxide, which is
volatile and can easily be removed from the reaction mixture [73–75]. However, the
specific activity of formate dehydrogenase in the range 5–7 U mg1 is quite low.
Glucose dehydrogenase (GDH), isolated, for example, from Bacillus subtilis [76] or B.
megaterium, shows significantly higher activity (about 350 U mg1). It is a useful
catalyst for the regeneration of NADH as well as NADPH oxidizing b-D-glucose,
which also is a cheap substrate, to D-glucono-1,5-lactone. The formed lactone is
quickly converted into the corresponding D-gluconic acid as a by-product, which
makes this reaction irreversible.
The enzymes in reductive amination and in situ cofactor-regeneration have been
used as isolated enzymes in most synthetic examples. However, recombinant whole-
cell catalysts turned out to be a highly attractive alternative to the use of isolated
enzymes. Although isolated enzymes can be applied in high concentrations and
activity, their recovery and purification is a cost factor. Therefore, whole cells
overexpressing simultaneously the enzyme for the desired enantioselective reaction
as well as the cofactor-regenerating enzyme (so-called designer cells) are a very
promising catalyst for technical applications. Especially in the case of formate
dehydrogenase, which is very often used as NADH-regenerating enzyme, the use
of whole cells offers the possibility of stabilizing this enzyme. Furthermore, recom-
binant whole-cells can be produced in an economically attractive fashion by means of
high cell density fermentation, and are the most straightforward, simple, and
cheapest form for enzymes including cofactors, since downstream-processing steps
to obtain isolated enzymes, such as cell disruption and enzyme purification,
are avoided.
28.3.2
Leucine Dehydrogenase Catalyzed Reductive Amination
28.3.2.1 L-tert-Leucine
Enantiomerically pure L-tert-leucine (L-22) is an intermediate in the synthesis of
several pharmaceuticals, for example, protease inhibitors against several diseases
such as different tumors, rheumatic arthritis, and AIDS [73, 77]. In addition, L-tert-
leucine turned out to be a versatile starting material for the synthesis of ligands [78] in
chiral metal complexes, which serve as chemocatalysts in asymmetric syntheses.
Furthermore, derivatives of this non-natural amino acid L-22 have been widely used
as organocatalysts in asymmetric synthesis [79].
Owing to increasing interest in this building block, several chemical or enzymatic
preparation methods have been developed [73]. Among enzyme-catalyzed processes
asymmetric syntheses are favored compared to resolution of racemic mixtures
because a complete transformation of the starting material into the desired product
L-tert-leucine can be reached. An attractive prochiral starting material for an asym-
metric transformation is trimethylpyruvate (21). One asymmetric method is based on
a transamination process, and makes use of the side reactivity of a branched-chain
aminotransferase [80]. The a-keto acid trimethylpyruvate (2-oxo-3,3-dimethylbutyric
acid, 21) is aminated by L-glutamate, which is regenerated either by another
aminotransferase (aspartate aminotransferase) or by reductive amination of a-keto-
glutarate with glutamate dehydrogenase and the NADH-regenerating system for-
mate/formate dehydrogenase [80–83].
An alternative synthesis of L-tert-leucine (L-22) starting from trimethylpyruvate
(21), which turned out to be highly attractive, is based on an enzymatic reductive
amination process using a leucine dehydrogenase. For coenzyme regeneration the
1184
formate dehydrogenase system was found to be highly suitable. Scheme 28.11 shows
the reaction concept of this biotransformation. A detailed study by Kula et al. revealed
that several leucine dehydrogenases from Bacillus strains are suitable for the
reductive amination of trimethylpyruvate [32]. For example, a leucine dehydrogenase
from Bacillus stearothermophilus shows an acceptable specific activity for trimethyl-
pyruvate of 63 U mg1 (31% related to the specific activity for 2-oxo-4-methylpenta-
noic acid as the substrate for the synthesis of L-leucine; Table 28.2). This enzyme has
been used successfully for the reductive amination of trimethylpyruvate (21) on a
preparative scale at a a-keto acid concentration of 0.5–1 M and a pH of 8.5 at room
temperature. The cofactor regeneration has been carried out by a formate dehydro-
genase from Candida boidinii and with a surplus of ammonium formate. The
reductive amination of trimethylpyruvate (21) has been reported to proceed with
complete conversion, and after purification by cation exchange chromatography L-
tert-leucine (L-22) was obtained in 85% yield and with an excellent enantiomeric
excess of >99% e.e. [32]. Notably, after operating a batch process for the synthesis of L-
tert-leucine for three days the enzymes were recovered with high remaining activity.
In particular, the leucine dehydrogenase from Bacillus stearothermophilus showed an
excellent operational stability with >99% recovered activity, whereas the recovered
activity of the formate dehydrogenase from Candida boidinii was 70%.
NH 2
Me
CO2H
Me
- + Me
HCO2 NAD
(S)-ter t-leucine
(L-22)
>99% ee
leucine
formate dehydrogenase,
dehydrogenase ammonia
O
Me
CO 2 NADH CO2H
Me
Me
21
Scheme 28.11 Synthesis of optically pure L-tert-leucine (L-22) by reductive amination with a leucine
dehydrogenase and formate/formate dehydrogenase for the regeneration of NADH.
Using a repetitive batch process with ultrafiltration, a space–time yield of 638 g l1
1
d was reached. When increasing the reactant concentrations up to 1 M, formation
of an undesired precipitate has been observed, which consists of the two side-
products (racemic) a-N-pivaloyl-tert-leucinamide and its cyclic imidazole deriva-
tive [84, 85]. These undesired side-products result from a non-enzymatic reaction
of ammonium with two equivalents of the a-keto acid 21.
Detailed reaction engineering with respect to the development of a L-tert-leucine
synthesis running in a continuous fashion has been reported by Kragl et al.,
who applied an enzyme membrane reactor to realize such processes [86].
High space–time yields of up to 366 g l1 d1 have been achieved when using
a single continuously operated enzyme membrane reactor, and in addition a high
turnover number (TON) of 4230 for the cofactor NAD þ has been achieved.
Furthermore, Kragl et al. carried out a detailed modeling and simulation study for
the continuous production of L-tert-leucine (L-22) via enzymatic reductive
amination [87].
A technology suitable for retaining not only the enzymes but also the cofactor
components NAD þ and NADH efficiently in the synthesis of L-tert-leucine (L-22) has
been developed by the Kragl group by means of a nanofiltration membrane in an
enzyme membrane reactor [88]. Retention rates were high for such cofactors
(0.86–0.98), whereas retention rates for substrates and product as lower molecular
weight compounds were below 0.35. The reductive amination process is stable over a
period of ten days, and an excellent TON of 7900 was achieved.
Notably, the enzymatic reductive amination technology for the production of
enantiomerically pure L-tert-leucine (L-22) based on a leucine dehydrogenase and
formate dehydrogenase as isolated enzymes has been applied on the ton scale at
Degussa AG (now Evonik Degussa GmbH) [32, 73, 89]. This process is the first
example of an enzymatic redox process on a technical scale, which proceeds under in
situ cofactor regeneration (see also Chapter 29, Industrial Applications).
Despite the high synthetic efficiency of the enzymatic reductive amination process
with isolated enzymes, challenges remained since isolation of enzymes as well as
addition of an external (albeit catalytic) amount of the expensive cofactor NADH
represent cost factors. To overcome these limitations the direct use of a recombinant
whole cell catalyst, containing both enzymes in overexpressed form, and the use of
only the intracellular amount of cofactor have been desirable. The Esaki group
mentioned the suitability of a recombinant whole-cell-catalyst for the synthesis of L-
tert-leucine without, however, reporting experimental data [90]. More recently,
Degussa researchers jointly with the Altenbuchner and Hummel groups developed
a reductive amination process for L-tert-leucine (L-22) based on the use of a recom-
binant whole-cell biocatalyst with a leucine dehydrogenase from Bacillus cereus and a
formate dehydrogenase from C. boidinii [91]. With respect to the construction of an
efficient recombinant whole-cell catalyst the E. coli strain BW3110 was chosen as host
organism. A particular challenge was the successful co-expression of both genes
(leudh and fdh), encoding for the LeuDH (leucine dehydrogenase) and FDH (formate
dehydrogenase) and leading to comparable activities for both enzymes despite the
large differences in the specific activities of both enzymes by a factor of >50 (>400 U
mg1 for the LeuDH, 6 U mg1 for the FDH). Owing to the very high activity of
LeuDH and low activity of FDH, the gene encoding for the FDH was located on a high
copy plasmid and the one for the LeuDH on a medium plasmid. This resulted in
formation of a more FDH protein than LeuDH and, thereby, improved the enzyme
activity ratio to about 7 (compared with a ratio of >50 when considering the specific
activities of both enzymes). With this recombinant whole-cell catalyst in hand,
1186
process development has been carried out. To operate at a high overall substrate
concentration and overcome substrate inhibition effects, a batch process has been
developed in which the substrate concentration is kept below 500 mM by continuous
addition of a trimethylpyruvate solution until a final substrate concentration of 1 M,
corresponding to a substrate loading of 130 g l1, was reached (Scheme 28.12). Under
these reaction conditions the reductive amination proceeded highly efficiently,
leading to L-tert-leucine (L-22) with >95% conversion after a reaction time of 24 h.
Notably, this reaction runs without the need of external cofactor. Subsequent
separation of the biomass via centrifugation, ultrafiltration of the resulting product
mixture, and purification via ion-exchange chromatography gave the desired L-tert-
leucine (L-22) in 84% yield and with an excellent enantiomeric excess of >99% e.e.
(Scheme 28.12).
E. coli-
whole-cell catalyst
containing
leucine dehydrogenase,
O formate dehydrogenase, NH2
Me NAD+ Me
CO2H CO2H
Me ammonium formate, Me
Me Me
water
21 (S)-tert-leucine (L-22)
(substrate input: 130 g/l) >95% conversion
84% yield
>99% ee
Scheme 28.12 Synthesis of L-tert-leucine (L-22) using a recombinant whole-cell catalyst.
28.3.2.2 L-Neopentylglycine (and Further Aliphatic Amino Acids)

Several other non-proteinogenic a-amino acids have also been synthesized by
applying both enzymes leucine dehydrogenase and formate dehydrogenase
(Scheme 28.13). L-Neopentyl-glycine (L-28) was produced on a 30 kg-scale (see also
Chapter 29) by Degussa AG (now: Evonik Degussa GmbH). This process also serves
as a proof that the leucine dehydrogenase based asymmetric synthesis is synthetically
useful even when the enzyme shows a decreased activity for a a-keto acid (here: 2-oxo-
4,4-dimethyl-pentanoic acid; Table 28.2). When a leucine dehydrogenase from B.
stearothermophilus was used, the L-amino acid L-28 was isolated with a yield of 74% and
an excellent enantioselectivity of >99.9% e.e. (Scheme 28.13) [32]. The same method
was applied to produce several other L-amino acids on a gram-scale
(Scheme 28.13) [32]. Measurements of enzyme activity after application for the
synthesis of amino acids show that the leucine dehydrogenase from B. stearother-
mophilus is quite stable and can be recovered with nearly the complete activity
whereas formate dehydrogenase deactivates to about 30–50% of the original value
within three days of operation [32].
Several whole-cell biocatalysts useful for the reductive amination of a-keto acids
were developed by the group of Esaki and Soda [90]. Aliphatic L-amino acids were
leucine dehydrogenase
O NH2
from B. stearothermophilus
R COOH formate dehydrogenase, R COOH
ammonium formate, NAD+,
aq. buffer (pH 8.5)
3 d, rt
Selected examples
NH2 NH2 NH2 NH2
COOH COOH COOH COOH
L-2-aminobutyric acid L-2-amino-3,3-dimethyl- L-2-amino-5,5-dimethyl- L-2-amino-4-ethyl-

(L-23) pentanoic acid (L-24) hexanoic acid (L-25) hexanoic acid (L-26)
86% yield 83% yield 60% yield 80% yield
>99.9% ee >99.9% ee >99.9% ee >99.9% ee
NH2 NH2 NH2
COOH COOH COOH
L-cyclohexyl alanine L-neopentyl glycine L-ter t-leucine

(L-27) (L-28) (L-22)
63% yield 74% yield (30 kg scale) 85% yield
>99.9% ee >99.9% ee >99.9% ee
Scheme 28.13 Non-proteinogenic L-a-amino acids synthesized by reductive amination with a

leucine dehydrogenase and a formate dehydrogenase for regeneration of NADH.
produced with E. coli cells overexpressing a thermostable leucine dehydrogenase

from Thermoactinomyces intermedius [27] and FDH from Mycobacterium vaccae [92].
The L-enantiomers of the amino acids leucine (L-29), norleucine (L-30), valine (L-31),
a-aminobutyric acid (L-23), and methionine (L-33) were produced in high yields
(88–97%) and with 100% e.e. within 12 h reaction time. By converting 0.4 M
of a-ketoisocaproate (2-oxo-4-methylpentanoic acid; ketoleucine) crystals of L-leucine
(L-29) appeared after reaching an amino acid concentration of about 0.15 M.
For the proteinogenic L-amino acid L-alanine (L-34) a similar catalyst was developed
with an alanine dehydrogenase from Bacillus sp. [93] instead of a leucine dehydro-
genase, and after conversion of 0.6 M of pyruvate a yield of 75% was obtained. The
e.e.-value, however, decreased with prolonged incubation time, reaching 80% e.e.
after 10 h, which is probably due to the occurrence of an alanine racemase in the
E. coli host cells.
More recently, another whole-cell biocatalyst with recombinant leucine dehydro-
genase from Bacillus cereus and formate dehydrogenase from C. boidinii was
developed for the synthesis of L-neopentylglycine (L-28) [94]. Starting from the a-keto
acid 2-oxo-4,4-dimethylpentanoic acid as substrate at an overall substrate input of
88 g l1, enzymatic reductive amination under in situ cofactor regeneration with
formate and a formate dehydrogenase furnished the desired enantiomerically pure
L-amino acid L-28 with a conversion of >95% and in an enantiomeric excess of >99%
1188 j 28 Reductive Amination of Keto Acids
LeuDH
FDH
O NAD(H) NH 2
R COOH R COOH
Whole-cell catalysts:
A: LeuDH (T . inter medius) B: AlaDH (Bacillus sp.) C: LeuDH (B. cer eus)
+ FDH (M. vaccae) + FDH (M. v accae) + FDH (C. boidinii)
Selected examples
NH 2 NH 2 NH2 NH 2 NH 2
COOH COOH COOH COOH COOH
catalyst A catalyst A catalyst A catalyst A catalyst A

L-leucine (L-29) L-norleucine (L-30) L-valine (L-31) L-norvaline (L-32) L-2-aminobutyric
97% yield 95% yield 95% yield 95% yield acid (L-23 )
>99% ee >99% ee >99% ee >99% ee 88% yield
>99% ee
NH2 NH 2 NH 2 NH 2
S COOH COOH COOH COOH
catalyst A catalyst B catalyst C catalyst C

L-methionine (L-33) L-alanine (L-34) L-t er t-leucine (L- 22) L-neopentyl glycine (L-28)
88% yield 95% yield 85% yield 74% yield
>99% ee 80 % ee >99.9% ee >99.9% ee
(30 Kg scale)
Scheme 28.14 Synthesis of L-amino acids type B genes of alanine dehydrogenase from
from a-keto acids by reductive amination using Bacillus sp. and formate dehydrogenase from
whole-cell catalysts based on recombinant E. Mycobacterium vaccae [90]; type C gene of
coli cells. Whole cells type A expressing the leucine dehydrogenase from Bacillus cereus and
genes of leucine dehydrogenase from formate dehydrogenase from Candida
Thermoactinomyces intermedius and formate boidinii [91, 94].
dehydrogenase from Mycobacterium vaccae [90];
Scheme 28.14 summarizes preparative examples using whole-cell catalysts with a

leucine dehydrogenase as an amino acid dehydrogenase component.
28.3.2.3 L-b-Hydroxyvaline
L-b-Hydroxyvaline is a building block needed for the synthesis of the monobactam
tigemonam. The corresponding keto acid (a-keto-b-hydroxyisovalerate) is con-
verted by leucine dehydrogenase from Bacillus sphaericus ATCC 4525 with an
apparent vmax of 41% related to a-ketoisovalerate, the best substrate of leucine
dehydrogenase. Bristol-Myers Squibb developed a biocatalytic method converting
a-keto-b-hydroxyisovalerate (35) into enantiomerically pure L-b-hydroxyvaline
(L-36,Scheme 28.15) [95]. This reductive amination proceeds with a conversion of
71% and an excellent enantioselectivity of >99% e.e.
O NH 2
leucine dehydrogenase
HO COOH HO COOH
+ NH 3
35 L-β-hydroxyvaline
(substrate (L-36 )
NADH NAD +
concentration: 71% conversion
0.5 M) >99% ee
gluconolactone glucose
glucose dehydrogenase
(GDH)
Scheme 28.15 Reductive amination of a-keto-b-hydroxyisovalerate (35) catalyzed by leucine

dehydrogenase from Bacillus sphaericus, forming L-b-hydroxyvaline (L-36).
28.3.2.4 Isotopically Labeled L-Amino Acids

The synthesis of isotopically labeled amino acids has gained broad interest for studies
in medicinal and bioorganic chemistry. Enzymatic reductive amination has been
shown to be a versatile tool for a variety of such syntheses, and examples thereof are
summarized in the following.
A diastereoselective enzymatic reductive amination has been used in a che-
moenzymatic multistep synthesis of 13C-labeled L-leucine, (2S,4R)-38 (L-38;
Scheme 28.16) [96]. When starting from the corresponding ester 37, a two-step
one-pot process based on hydrolytic cleavage of the ester group of 37 with a lipase
under formation of the corresponding a-keto acid and subsequent enzymatic
reductive amination furnished the desired product (2S,4R)-5-[13C]leucine, (2S,4R)-
38 (L-38), in 79% yield. For the reductive amination a leucine dehydrogenase from
Bacillus sp. in combination with a formate dehydrogenase from Candida boidinii
was used.
lipase CRL,
H3 C CO2 Et formate dehydrogenase H 3C CO 2H
13 13
CH 3 O HCO 2NH4 CH 3 NH2
37 L-38
79% yield
Scheme 28.16 Enzymatic reduction for the synthesis of 13C-labeled L-leucine (L-38).
Starting from the corresponding ester 39 the same synthetic strategy has been
applied for the synthesis of L-leucine that is stereoselectively labeled with deuterium
at one of the diastereotopic methyl groups (L-40; Scheme 28.17) [96]. The desired
product (2S,4R)-[5,5,5-2H3]leucine [(2S,4R)-40,L-40] has been obtained in 85% yield
after lipase-catalyzed ester hydrolysis and leucine dehydrogenase-catalyzed reductive
amination under in situ cofactor regeneration with a formate dehydrogenase.
1190
lipase CRL,
H3C CO2Et formate dehydrogenase H3C CO2H
CD3 O HCO2NH4 CD3 NH2

39 L-40
85% yield
Scheme 28.17 Enzymatic reduction for the synthesis of deuterium-labeled L-leucine (L-40).
The synthesis of 15N-labeled L-amino acids by means of an enzymatic reductive

amination has also been accomplished, as exemplified by the Willis group for the
synthesis of b-hydroxy a-amino acids such as L-threonine [97]. The synthesis starts
from ethyl (S)-lactate, which was protected and then converted into the a-keto acid
methyl ester 41. Subsequent treatment with a lipase in aqueous medium (for ester
cleavage), reductive amination using a leucine dehydrogenase in combination with a
formate dehydrogenase and [15N]ammonium formate as an ammonia source,
and cleavage of the MOM-protecting group with HCl gave the desired L-[15N]
threonine (L-42) in 93% yield (Scheme 28.18, Reaction 1). In analogous fashion the
diastereomeric allo-[15N]-threonine (L-44) was obtained in a yield of 61% when
starting from MOM-protected methyl (R)-lactate 43 (Scheme 28.18, Reaction 2).
lipase CCL,
OMOM formate dehydrogenase OH
CO2Me CO2H (1)
H3C NADH, HCO215NH4, H3C
O 15NH
then 2M HCl 2
41 L-42
93% yield
lipase CCL,
phenylalanine dehydrogenase,
OR formate dehydrogenase OR
CO2Me CO2H (2)
H3C NADH, HCO215NH4, H3C
15
O then 2M HCl NH2
43 L-44
R=MOM R=H: 61% yield
Scheme 28.18 Enzymatic reduction for the synthesis of 15N-labeled L-amino acids.
28.3.2.5 L-Amino Acids with two Stereogenic Centers

Enzymatic reductive amination of a-keto acids has been extended to a kinetic
resolution process that affords a-amino acids of type 46, bearing two stereogenic
centers at the a- and b-position, respectively, when starting from racemic a-keto acids
rac-45. In the presence of a leucine dehydrogenase from Bacillus sp. this has been
demonstrated by the Willis group for the enantio- and diastereoselective synthesis of
a range of 3-substituted a-aminobutanoic acids, including allo-threonine [98]. Start-
ing from 3-methoxymethoxy 2-oxobutanoic acid such an enzymatic reductive amina-
tion gave a mixture of allo-threonine and L-threonine in a diastereomeric ratio of 4 : 1
and in 47% yield after incubation with a leucine dehydrogenase (Scheme 28.19;
experimental details, for example, with respect to cofactor regeneration have not been
published [98]).
H3C
O O incubation with a OH OH
CO2H leucine dehydrogenase
CO2H CO2H
H3C H3C + H3C
O then 2M HCl,
NH2 NH2
47% yield
rac-45 d.r.(anti/syn)=4:1 anti-46 syn-46
(2S,3S)-46 (2S,3R)-46
Scheme 28.19 Resolution via enzymatic reductive amination with a leucine dehydrogenase.
Furthermore, branched-chain racemic a-keto acids, which are not heteroatom-

substituted at the 3-position, served as substrates [98]. For example, the racemic
methyl ester of 5-methoxy-3-methyl-2-oxopentanoic acid (rac-47) was converted into
the corresponding amino acids anti-48 and syn-48 in 29% yield as a 1 : 1 diastereo-
meric mixture (Scheme 28.20). The initial step consists of a lipase-catalyzed ester
hydrolysis, followed by a subsequent enzymatic reductive amination process with a
leucine dehydrogenase. The in situ cofactor regeneration was carried out using
formate as a cosubstrate by means of a formate dehydrogenase.
H3C 1. lipase from Candida rugosa,

H3C H3C
O O 2. leucine dehydrogenase, O O O O
CO2Me formate dehydrogenase CO2H CO2H
H3C H3C + H3C
O NADH, HCO 2 NH4, NH2 NH2
rac-47 29% overall yield (2 steps) anti-48 syn-48
d.r.(anti/syn)=1:1 (2S,3S)-48 (2S,3R)-48
Scheme 28.20 Transformation of racemic ester rac-47 into amino acids of type 48 via combined
enzymatic steps.
28.3.3
Phenylalanine Dehydrogenase Catalyzed Reductive Amination
28.3.3.1 Synthesis of (S)-2-Amino-4-Phenylbutanoic Acid (L-Homophenylalanine)

(S)-2-Amino-4-phenylbutanoate is a building block of angiotensin converting
enzyme (ACE) and renin inhibitors. This phenylalanine homolog compound is
accessible by reductive amination of the corresponding a-keto acid using phenyl-
alanine dehydrogenase from Rhodococcus sp. as a biocatalyst [99]. Kinetic studies
1192
revealed severe substrate inhibition by 2-keto-4-phenylbutanoate, while phenylpyr-

uvate did not show such inhibition. The calculated substrate inhibition constant (Ki)
for 2-oxo-4-phenylbutanoate is 230 mM, which means that conversion at the maxi-
mum reaction rate was possible if the concentration of the keto acid was kept at
200 mM. Enantiomerically pure L-homophenylalanine (L-50) could be obtained using
reductive amination coupled with a formate dehydrogenase-catalyzed oxidation of
formate to carbon dioxide for regeneration of NADH, leading to the desired pure
product L-50 in 48% yield (Scheme 28.21).
O NH 2
phenylalanine dehydrogenase
COOH COOH
+ NH3
49 L-homophenylalanine (L-50)
NADH NAD + 63% conversion
48% yield
optically pure
CO 2 HCOOH
f ormate dehydrogenase
(FDH)
Scheme 28.21 Synthesis of enantiomerically pure homophenylalanine L-50 via enzymatic reductive
amination catalyzed by phenylalanine dehydrogenase from a Rhodococcus sp.
28.3.3.2 Synthesis of Allysine Ethylene Acetal ((S)-2-Amino-5-1,3-Dioxolan-2-

ylpentanoic Acid)
Allysine ethylene acetal is a chiral building block required for the synthesis of
omapatrilat, a vasopeptidase inhibitor. Bristol-Myers Squibb published several
methods of how to create the chiral center by reductive amination of the correspond-
ing a-keto acid [100–102]. Phenylalanine dehydrogenase from Thermoactinomyces
intermedius was used as the catalyst for the reductive amination, and formate
dehydrogenase to regenerate NADH. Heat-dried cells of E. coli harboring recombi-
nant phenylalanine dehydrogenase and heat-dried cells of Candida boidinii as the
source of formate dehydrogenase were used in several batch processes. A total of
197 kg of the L-amino acid L-52 was produced in three 1600-l batches using 5 mass%
of the a-keto acid (see also Chapter 29); L-52 was obtained in an average yield of 91.1%
and in excellent enantiomeric excess of >98% e.e. (Scheme 28.22).
28.3.3.3 Synthesis of the N-Terminal Amino Acid Portion of Nikkomycins

Nikkomycins are non-toxic nucleoside antifungals with an N-terminal amino acid
part. The enantiomerically pure L-amino acid L-55, bearing two stereogenic centers,
can be obtained by two consecutive enzyme-catalyzed steps starting from achiral
commodity chemicals [103]. As an example, pyridinecarboxaldehyde (53) and pyru-
vate (19) are converted by a KDPG aldolase (EC 4.1.2.14) into the a-keto acid 54, which
O O phenylalanine dehydrogenase O NH2
O COOH O COOH
+ NH3
51 (S)-allysine ethylene acetal
80 kg scale (L-52)
NADH NAD+ 92% conversion
>98% ee
CO2 HCOOH
formate dehydrogenase
(FDH)
Scheme 28.22 Phenylalanine dehydrogenase catalyzedsynthesis of(S)-allysine ethylene acetal(L-52).
is subsequently reductively aminated by means of a phenylalanine dehydrogenase

from Bacillus sphaericus to afford the desired L-amino acid L-55 in an overall yield of
75.9% (Scheme 28.23).
NH4+ H2O
O O KDPG OH O OH NH2
N OH aldolase N PheDH N
H + COOH COOH
O
53 19 (S)-54 NADH NAD+ (S,S)-55 (L-55)
99.7% ee 75.9% overallyield
FDH opticallypure
CO2 HCOOH
Scheme 28.23 Enzyme-catalyzed synthesis of the N-terminal amino acid fragment of nikkomycin
KX and KZ by a two-step process: initial formation of a-keto acid intermediate (S)-54 by an aldolase
followed by a phenylalanine dehydrogenase (PheDH)-catalyzed reductive amination.
A broad range of derivatives can be synthesized by this route because KDPG

aldolase accepts 2-, 3-, and 4-pyridinecarboxaldehydes and further related electro-
philes as well as fluoropyruvate, oxobutyrate, and hydroxypyruvate as nucleophiles.
28.3.3.4 Synthesis of (S)-3-Hydroxyadamantylglycine

Saxagliptin is an inhibitor of dipeptidyl peptidase IV currently being explored for
treatment of type 2 diabetes [104, 105]. A chiral intermediate for its synthesis is (S)-3-
hydroxyadamantylglycine (L-57), which can be obtained by reductive amination of the
a-keto acid 56. Bristol-Myers Squibb developed a biocatalytic process that was scaled
up to multi-kg scale (see also Chapter 29) using a phenylalanine dehydrogenase from
Thermoactinomyces intermedius and a formate dehydrogenase to regenerate NADH
(Scheme 28.24) [106]. Phenylalanine dehydrogenase was used as a C-terminal
modified enzyme bearing two changed amino acids and an additional 12 amino
acids at the C-terminus. This changed gene was introduced into a plasmid and
transformed into Pichia pastoris as the host. Formate dehydrogenase for NADH
1194j 28 Reductive Amination of Keto Acids
phenylalanine dehydrogenase
CO2 H CO 2H
+ NH 3
HO O HO NH2
56 (S)-3-hydroxyadamantylglycine
40 kg scale (L-57)
NADH NAD+
99% conversion
>99% ee
CO2 HCOOH
f ormate dehydrogenase
(FDH)
Scheme 28.24 Enzymatic synthesis of (S)-3-hydroxyadamantylglycine (L-57) by coupled use of a

phenylalanine dehydrogenase and a formate dehydrogenase.
regeneration was produced by P. pastoris cells during growth on methanol.

The reductive amination leading to (S)-3-hydroxyadamantylglycine (L-57) was carried
out at a pilot plant scale using 114.7 kg of frozen recombinant P. pastoris cells and
40 kg of the a-keto acid 56 in a 800-l vessel. After a reaction time of 49 h, a conversion
of 99% was observed.
28.3.4
Glutamate Dehydrogenase Catalyzed Reductive Amination
28.3.4.1 Synthesis of L-6-Hydroxynorleucine with Glutamate Dehydrogenase

L-6-Hydroxy-norleucine (L-59) is a chiral intermediate required for the synthesis of
Omapatrilat, an antihypertensive drug candidate that acts by inhibiting ACE and
neutral endopeptidase [107]. Bristol-Myers Squibb has published different methods
to obtain this amino acid by reductive amination of the corresponding a-keto acid 58.
One route makes use of NAD þ -dependent beef liver glutamate dehydrogenase,
which is coupled with glucose and a glucose dehydrogenase for regeneration of
NADH (Scheme 28.25) [108]. This group also developed successfully a route for
reductive amination of this a-keto acid 58 with a phenylalanine dehydrogenase.
O NH2
glutamate dehydrogenase
HO COOH HO COOH
+ NH3
2-keto-6-hydroxyhexanoic acid L-6-hydroxynorleucine (L-59)
(58) 91% conversion
(substrate concentration: NADH NAD + 80% yield
0.49 M) >99% ee
(GDH)
Scheme 28.25 Glutamate dehydrogenase catalyzed synthesis of L-6-hydroxynorleucine (L-59).

28.3.4.2 In Situ Synthesis of L-Glutamate as a Cosubstrate for Transamination
Processes
The use of an enzymatic reductive amination for the in situ-formation of L-glutamate
(L-3), which then serves directly as a cosubstrate and amino donor in a subsequent
enzymatic transamination process for the synthesis of D-amino acids, has been
successfully developed by the Soda group [109, 110]. D-Amino acids are useful
starting materials for the synthesis of antibiotics, bioactive peptides, and other
physiologically active compounds. For a long time D-amino acids have not been
synthetically accessible by reductive amination of their corresponding a-keto acid
due to the lack of suitable D-specific NAD(P) þ -dependent amino acid dehydro-
genases. In contrast, D-transaminase-catalyzed synthesis of D-amino acids has been
known, and the groups of Soda [109] and sung [110] developed an elegant method to
synthesize various D-amino acids from a-keto acids by coupling four enzymes,
including a D-transaminase and an L-glutamate dehydrogenase (Scheme 28.26). This
process is based on the strict enantioselective D-amino acid aminotransferase, which
requires D-glutamate (D-3) as the amino donor and glutamate dehydrogenase to
regenerate D-glutamate by a coupled process of reductive amination of the by-product
a-ketoglutarate (1) with glutamate dehydrogenase and an enzyme-catalyzed racemi-
zation by glutamate racemase. Starting with 100 mM of the a-keto acid, the D-
enantiomers of valine, alanine, a-aminobutyrate, aspartate, leucine, and methionine
were synthesized in yields >80%. Serine, histidine, phenylalanine, and tyrosine,
which are based on a-keto acids that are poor substrates for the aminotransferase,
were produced in yields of 50%, 36%, 28%, and 13%, respectively.
NH2 O
glutamate
HOOC COOH R COOH
racemase
D-3
NH2
D-amino acid
HOOC COOH aminotransferase
L-3
O
L-glutamate NH 2
dehydrogenase, HOOC COOH
NAD +, FDH, 1 R COOH
formate, NH 3 D-amino acid
Scheme 28.26 Synthesis of D-amino acids by the coupled use of four enzymes – with NADH as the
coenzyme for L-glutamate dehydrogenase and cofactor-regeneration by formate dehydrogenase
(FDH) and formate.
28.3.5
D-Amino Acid Dehydrogenase-Catalyzed Reductive Amination
For a long time, enzymatic reductive amination has been limited to the enantiose-
lective formation of L-enantiomers of amino acids, whereas synthetically applicable
1196
D-amino acid dehydrogenases based on the use of NAD(P)H as a cofactor had not
been available. However, in 2006 researchers from BioCatalytics reported the
creation of a D-amino acid dehydrogenase by means of rational and random
mutagenesis [7]. The mutagenesis was carried out via directed evolution of a
meso-2,6-D-diaminopimelic acid dehydrogenase as a starting enzyme. The most
active mutant, BC621, turned out to be suitable for the reductive amination of
various a-keto acids in a highly enantioselective fashion. Most D-amino acids were
obtained with an excellent enantiomeric excess of >99% e.e. The range of D-amino
acids obtained with >99% e.e. includes, for example, D-2-aminobutyrate, D-norvaline,
D-norleucine, D-leucine, D-cyclopentylglycine, D-cyclohexylalanine, D-methionine, D-
phenylalanine, and D-tyrosine. By means of this created D-amino acid dehydrogenase
a reductive amination process on a gram scale has also been successfully performed.
In combination with glucose and a glucose dehydrogenase for in situ cofactor
regeneration, cyclohexylpyruvate (60) was transformed at a substrate input of 40 g
l1 into the desired D-amino acid D-cyclohexylalanine (D-27) with a conversion of
>95% and excellent enantioselectivity (Scheme 28.27). The authors also comment
that this D-amino acid dehydrogenase could contribute to an inexpensive production
of enantiomerically pure D-amino acids [7].
D-amino acid dehydrogenase

O (mutant BC621) NH2
CO2H + NH3 CO2H

60 D-27
(substrate >95% conversion
input: 40 g/l) NADPH NADP+ ca. 70% yield
>99% ee
Scheme 28.27 Synthesis of D-cyclohexylalanine (D-27) using a D-amino acid dehydrogenase.
28.3.6
N-Methyl-amino Acid Dehydrogenase
A recombinant N-methyl-amino acid dehydrogenase from Pseudomonas putida

converts several aliphatic a-keto acids such as pyruvic or a-ketoisocaproate and
phenylpyruvic acid with a surplus of methylamine into N-methyl-L-amino acids. The
group of Esaki developed an E. coli-based recombinant whole-cell catalyst expressing
this amino acid dehydrogenase together with a glucose dehydrogenase from Bacillus
subtilis and used this catalyst to synthesize N-methyl-L-phenylalanine (L-62) at a 100
ml-scale from phenylpyruvic acid (61) [111]. After a reaction time of 7 h, 16 g l1 of L-
62 was produced, corresponding to a yield of >98% (Scheme 28.28). Furthermore,
the enantioselectivity of this reductive amination process is excellent, as indicated by
the >99% e.e. for the desired product L-62.
N-methyl-L-amino CH 3
O HN
acid dehydrogenase
+ CH 3NH 2
COOH COOH
61 N-methyl-L-phenylalanine
(L-62)
NADPH NADP+ 98% yield
>99% ee
(GDH)
Scheme 28.28 Synthesis of N-methyl-L-amino acid (L-62) by reductive amination using an E. coli
whole-cell catalyst co-expressing an N-methyl-L-amino acid dehydrogenase from Pseudomonas
putida and a glucose dehydrogenase (GDH) from Bacillus subtilis.
This group also reported the enantioselective reduction of cyclic imino acids by
means of this enzyme [112]. Imine compounds of type 65, such as, for example,
3,4,5,6-tetrahydropyridine-2-carboxylic acid or pyrroline-2-carboxylic acid, were
reduced to the L-form of the corresponding cyclic amino acids (Scheme 28.29).
Cyclic imino acids of type 65 are formed spontaneously from a-keto-v-amino acids
64, which are synthesized enzymatically in situ from a,v-diamino acids using amino
acid oxidases. For example, L-lysine was oxidized by an L-amino acid oxidase from
snake venom or a specific lysine oxidase from Trichoderma viridae [113]. Furthermore,
D-lysine in enantiomerically pure form or D-lysine in a racemic mixture are also well
suited as starting materials because the a-keto-v-amino acid a-keto lysine can be also
prepared by oxidation with a D-amino oxidase. Summing up, enantiomerically pure
COOH
X
NH2 LAO or LO
N-methyl-L-amino
NH2 COOH acid dehydrogenase
X X X
L-63a-f CO2H CO2H
O N NADP+, N
COOH DAO NH2 GDH, glucose H
X
64a-f 65a-f L-66a-f
NH2
NH2 for example:
L-pipecolicacid,(S)- 66a
D-63a-f 98% yield
X 100% ee
a (CH2)2
b CH2
c CHOHCH2
d (CH2)3
e CH2S
f (CH2)2S
Scheme 28.29 Synthesis of cyclic L-amino amino acid oxidases (LAO ¼ L-amino acid
acids of type L-66 by reduction of imino acids 65 oxidase; LO ¼ lysine oxidase; DAO ¼ D-amino
with N-methyl-L-amino acid dehydrogenase; acid oxidase) followed by spontaneous
imino acids 65 are formed in situ by oxidation of cyclization of the a-keto acid 64.
a,v-diamino acids L-63 or D-63 catalyzed by
1198
(a)
opine H
R NH 2 O H dehydrogenase R N H
+
COOH COOH cofactor, CO2H CO2 H
cofactor recycling
amino acid glyoxylate
Selected examples
H H H
N H N H N H
CO 2H CO2 H CO2 H CO2H CO 2H CO2 H

amino acid = amino acid = amino acid =
isoleucine leucine valine
>99% yield >99% yield >99% yield
(b)
opine H
R NH2 O CH3 dehydrogenase R N CH 3
+
COOH COOH cofactor, CO2 H CO 2H
cofactor recycling
amino acid pyruvate
Selected examples
H H H
S N CH3 N CH3 N CH 3
CO2H CO2 H CO2H CO2 H CO 2H CO2 H

methionine isoleucine leucine
96% yield >99% yield >99% yield
>99.9% de >99.9% de >99.9% de
>99.9% ee >99.9% ee >99.9% ee
H H
H N CH3 N CH3
N CH3 HO
CO2 H CO 2H CO 2H CO2 H
CO 2H CO2 H
valine phenylalanine serine
>99% yield 95% yield 97% yield
>99.9% de >99.9% de >99.9% de
>99.9% ee >99.9% ee >99.9% ee
(c)
opine H
R NH 2 O CH 2CH3 dehydrogenase R N CH 2 CH 3
+
COOH COOH cofactor, CO 2H CO2 H
cofactor recycling
amino acid α -keto butyric acid
Selected examples
H H
N CH 2 CH 3 N CH 2CH3
CO 2H CO2 H CO2 H CO 2H
amino acid = amino acid =
isoleucine valine
97% yield >99% yield
>99.9% de >99.9% de
>99.9% ee >99.9% ee
Scheme 28.30 Synthesis of opine-type secondary amine dicarboxylic acids from L-amino acid and
glyoxylate (a), pyruvate (b), or a-keto butyric acid (c) catalyzed by opine dehydrogenase from
Arthrobacter sp.
28.4 Summary j1199
cyclic L-amino acids L-66 can be produced from a,v-diamino acids D-63 or L-63 by
coupling an oxidative step, catalyzed by an amino acid oxidase, and a reductive step,
using the NADPH-dependent N-methyl-L-amino acid dehydrogenase.
28.3.7
Opine Dehydrogenase
Opine dehydrogenases have also been used in synthetic biotransformations, cata-

lyzing the NAD(P)H-dependent reductive condensation of a a-keto acid with an
amino acid to afford N-carboxylalkyl-amino acids (opines). Scheme 28.30 gives an
overview about such types of processes. The resulting amino acid derivatives are
useful intermediates of ACE-inhibitors such as enalapril or lysinopril.
In a first approach, the Asano group studied the preparative potential of the
recombinant NADH-dependent opine dehydrogenase from Arthrobacter sp. as a
biocatalyst [114]. This enzyme accepts a broad range of (S)-2-amino acids as amino
donors, for example, norvaline, 2-aminobutyrate, norleucine, b-chloro-alanine,
methionine, isoleucine, valine, phenylalanine, o-phospho-serine, or phenylglycine.
Vmax values between 100 and 600 U mg1 were calculated from kinetic studies for
these substrates when using pyruvate as the a-keto acid. Scheme 28.30b shows
selected synthetic examples of such biotransformations based on the use of pyruvate
as a-keto acid component.
Notably, the b-amino acid (R,S)-3-aminobutyrate is accepted, too, with a low Vmax
value of 6.5 U mg1; the amino alcohol (S)-phenylalaninol is also oxidized, with a
calculated Vmax of 18 U mg1. Other amino alcohols corresponding to aliphatic
amino acids were not suitable substrates.
Besides pyruvate, the homolog compounds glyoxylate or a-keto butyric acid can
also be used as alternative a-keto acid components (Scheme 28.30a and c). No activity,
however, was found with amino acid esters or amides as well as with amines such as
methylamine, ethylamine, 2-phenylethylamine, hydrazine, or hydroxylamine.
28.4
Summary
The enzyme-catalyzed reductive amination of a-keto acids is a highly efficient

method for the synthesis of a-amino acids with excellent enantiomeric excesses
typically exceeding 99% e.e. Meanwhile, a broad range of dehydrogenases are
available, forming a-amino acids, N-methyl a-amino acids, or opines (N-carbox-
ylalkyl a-amino acids). All known enzymes require NADH or NADPH, which can
be regenerated in situ using a second enzyme, for example, formate dehydroge-
nase or glucose dehydrogenase in combination with a suitable cosubstrate. Such
types of biotransformation reactions catalyzed by isolated enzymes or recombinant
whole-cell catalysts (so-called designer cells) have already found several applica-
tions on a technical scale and represent a current industrial key technology for the
production of, in particular, non-natural a-amino acids, which are required as
1200
chiral building blocks for the synthesis of drugs by the fine chemicals and
pharmaceutical industry.
References
1 Weckbecker, A., Gr€

oger, H., and 16 Brunhuber, N.M.W., Thoden, J.B.,
Hummel, W. (2010) Adv. Biochem. Eng./ Blanchard, J.S., and Vanhooke, J.L.
Biotechnol., 120, 195–242. (2000) Biochemistry, 39, 9174–9187.
2 Ohshima, T. and Soda, K. (2000) in 17 Obermeier, N. and Poralla, K. (1976)
Stereoselective Biocatalysis (ed. R. N. Patel), Arch. Microbiol., 109, 59–63.
Marcel Dekker, New York, 877–902. 18 Obermeier, N. and Poralla, K. (1979)
3 Fu, G.X., Yuan, H.L., Li, C.R., Lu, C.D., FEMS Microbiol. Lett., 5, 81–83.
Gadda, G., and Weber, I.T. (2010) 19 Nagata, S., Tanizawa, K., Esaki, N.,
Biochemistry, 49, 8535–8545. Sakamoto, Y., Ohshima, T., Tanaka, H.,
4 Tanigawa, M., Shinohara, T., Saito, M., and Soda, K. (1988) Biochemistry, 27,
Nishimura, K., Hasegawa, Y., 9056–9062.
Wakabayashi, S., Ishizuka, M., and 20 Ohshima, T., Misono, H., and
Nagata, Y. (2010) Amino Acids, 38, Soda, K. (1978) J. Biol. Chem., 253,
247–255. 5719–5725.
5 Marshall, V.P. and Sokatch, J.R. (1968) 21 Ohshima, T., Nagata, S., and
J. Bacteriol., 95, 1419–1424. Soda, K. (1985) Arch. Microbiol., 141,
6 Olsiewski, P.J., Kaczorowski, G.J., and 407–411.
Walsh, C. (1980) J. Biol. Chem., 255, 22 Ohshima, T. and Soda, K. (1989) Trends
4487–4494. Biotechnol., 7, 210–215.
7 Vedha-Peters, K., Gunawardana, M., 23 Hummel, W., Sch€ utte, H., and Kula,
Rozzell, J.D., and Novick, S.J. (2006) M.-R. (1981) Eur. J. Appl. Microbiol.
J. Am. Chem. Soc., 128, 10923–10929. Biotechnol., 12, 22–27.
8 Brunhuber, N.M.W. and Blanchard, J.S. 24 Sch€ utte, H., Hummel, W., Tsai, H., and
(1994) Crit. Rev. Biochem. Mol. Biol., 29, Kula, M.-R. (1985) Appl. Microbiol.
415–467. Biotechnol., 22, 306–317.
9 Patel, R.N., Hanson, R.L., Banerjee, A., 25 Shimoi, H., Nagata, S., Esaki, N., Tanaka,
and Szarka, L.J. (1997) J. Am. Oil Chem. H., and Soda, K. (1987) Agric. Biol. Chem.,
Soc., 74, 1345–1360. 51, 3375–3381.
10 Patel, R.N. (2001) Adv. Synth. Catal., 343, 26 Misono, H., Sugihara, K., Kuwamoto, Y.,
527–546. Nagata, S., and Nagasaki, S. (1990) Agric.
11 Patel, R.N. (2000) Adv. Appl. Microbiol., Biol. Chem., 54, 1491–1498.
47, 33–78. 27 Ohshima, T., Nishida, N.,
12 Patel, R.N. (1997) Adv. Appl. Microbiol., Bakthavatsalam, S., Kataoka, K.,
43, 91–140. Takada, H., Yoshimura, T., Soda, K., and
13 Patel, R., Hanson, R., Goswami, A., Esaki, N. (1994) Eur. J. Biochem., 222,
Nanduri, V., Banerjee, A., Donovan, M.J., 305–312.
Goldberg, S., Johnston, R., 28 Muranova, T.A., Ruzheinikov, S.N.,
Brzozowski, D., Tully, T., Howell, J., Sedelnikova, S.E., Baker, P.J., Pasquo, A.,
Cazzulino, D., and Ko, R. (2003) J. Ind. Galkin, A., Esaki, N., Ohshima, T., Soda,
Microbiol. Biotechnol., 30, 252–259. K., and Rice, D.W. (2002) Acta Crystallogr.,
14 Zhu, D. and Hua, L. (2009) Biotechnol. J., Sect. D, 58, 1059–1062.
4, 1420–1431. 29 Ansorge, M.B. and Kula, M.R. (2000)
15 Baker, P.J., Britton, K.L., Engel, P.C., Biotechnol. Bioeng., 68, 557–562.
Farrants, G.W., Lilley, K.S., Rice, D.W., 30 Ansorge, M.B. and Kula, M.R. (2000)
and Stillman, T.J. (1992) Proteins: Struct., Appl. Microbiol. Biotechnol., 53,
Funct., Genet., 12, 75–86. 668–673.
References j1201
31 Oka, M., Yang, Y.S., Nagata, S., Esaki, N., 47 Baker, P.J. (2004) Biochem. Soc. Trans., 32,
Tanaka, H., and Soda, K. (1989) 264–268.
Biotechnol. Appl. Biochem., 11, 48 Baker, P.J., Waugh, M.L., Wang, X.G.,
307–311. Stillman, T.J., Turnbull, A.P., Engel, P.C.,
32 Krix, G., Bommarius, A.S., Drauz, K., and Rice, D.W. (1997) Biochemistry, 36,
Kottenhahn, M., Schwarm, M., and 16109–16115.
Kula, M.R. (1997) J. Biotechnol., 53, 49 Stillman, T.J., Migueis, A.M.B.,
29–39. Wang, X.G., Baker, P.J., Britton, K.L.,
33 Hummel, W., Weiss, N., and Kula, M.-R. Engel, P.C., and Rice, D.W. (1999)
(1984) Arch. Microbiol., 137, 47–52. J. Mol. Biol., 285, 875–885.
34 Hummel, W., Sch€ utte, H., Schmidt, E., 50 Wang, X.G., Britton, K.L., Stillman, T.J.,
Wandrey, C., and Kula, M.-R. (1987) Rice, D.W., and Engel, P.C. (2001)
Appl. Microbiol. Biotechnol., 26, Eur. J. Biochem., 268, 5791–5799.
409–416. 51 Drauz, K., Grayson, I., Kleemann, A.,
35 Asano, Y. and Nakazawa, A. (1985) Agric. Krimmer, H.-P., Leuchtenberger, W., and
Biol. Chem., 49, 3631–3632. Weckbecker, C. (2007) Ullmanns
36 Asano, Y. and Nakazawa, A. (1987) Agric. Biotechnology and Biochemical
Biol. Chem., 51, 2035–2036. Engineering, vol. 1, Wiley-VCH Verlag
37 Asano, Y., Nakazawa, A., and GmbH, Weinheim, p. 253.
Endo, K. (1987) J. Biol. Chem., 262, 52 Kretovich, W.L. and Stepanovich, K.M.
10346–10354. (1966) Izv. Akad. Nauk SSSR Ser. Biol., 2,
38 Asano, Y., Nakazawa, A., Endo, K., 295–301.
Hibino, Y., Ohmori, M., Numao, N., and 53 Goldman, D.S. and Wagner, M.J.
Kondo, K. (1987) Eur. J. Biochem., 168, (1962) Biochim. Biophys. Acta, 65,
153–159. 297–306.
39 Misono, H., Yonezawa, J., Nagata, S., and 54 Wayne, L.G. and Lin, K.Y. (1982) Infect.
Nagasaki, S. (1989) J. Bacteriol., 171, Immunol., 37, 1042–1049.
30–36. 55 Baker, J.J., Jeng, I., and Barker, H.A.
40 De Boer, L., Van Rijssel, M., Euverink, (1972) J. Biol. Chem., 247, 7724–7734.
G.J., and Dijkhuizen, L. (1989) Arch. 56 Hong, S.C.L. and Barker, H.A. (1973)
Microbiol., 153, 12–18. J. Biol. Chem., 248, 41–49.
41 Ohshima, T., Takada, H., Yoshimura, T., 57 Somack, R. and Costilow, R.N. (1973)
Esaki, N., and Soda, K. (1991) J. Bacteriol., J. Biol. Chem., 247, 385–388.
173, 3943–3948. 58 Tsuda, Y. and Friedmann, H.C. (1970)
42 Takada, H., Yoshimura, T., Ohshima, T., J. Biol. Chem., 245, 5914–5926.
Esaki, N., and Soda, K. (1991) J. Biochem., 59 Buergi, W., Richterich, R., and
109, 371–376. Colombo, J.P. (1966) Nature, 211,
43 Sedelnikova, S.E., Yip, K.S.P., Stillman, 854–855.
T.J., Ma, K., Adams, M.W.W., Robb, F.T., 60 Heydari, M., Ohshima, T.,
and Rice, D.W. (1996) Acta Crystallogr., Nunoura-Kominato, N., and Sakuraba,
Sect D, 52, 1185–1187. H. (2004) Appl. Environ. Microbiol., 70,
44 Lebbink, J.H.G., Knapp, S., van der Oost, 937–942.
J., Rice, D., Ladenstein, R., and de Vos, 61 Misono, H., Togawa, H., Yamamoto, T.,
W.M. (1998) J. Mol. Biol., 280, 287–296. and Soda, K. (1979) J. Bacteriol., 137,
45 Vetriani, C., Maeder, D.L., Tolliday, N.J., 22–27.
Klump, H.H., Yip, K.S.P., Rice, D.W., and 62 Misono, H. and Soda, K. (1980) J. Biol.
Robb, F.T. (1998) New Dev. Mar. Chem., 255, 599–605.
Biotechnol., 221–225. 63 Misono, H., Ogasawara, M., and
46 Britton, K.L., Yip, K.S.P., Sedelnikova, Nagasaki, S. (1986) Agric. Biol. Chem., 50,
S.E., Stillman, T.J., Adams, M.W.W., Ma, 1329–1330.
K., Maeder, D.L., Robb, F.T., Tolliday, N., 64 Misono, H., Ogasawara, M., and
Vetriani, C., Rice, D.W., and Baker, P.J. Nagasaki, S. (1986) Agric. Biol. Chem., 50,
(1999) J. Mol. Biol., 293, 1121–1132. 2729–2734.
1202
65 Misono, H., Togawa, H., Yamamoto, T., 83 Tewari, Y.B., Goldberg, R.N., and
and Soda, K. (1976) Biochem. Biophys. Res. Rozzell, J.D. (2000) J. Chem. Thermodyn.,
Commun., 72, 89–93. 32, 1381–1398.
66 Lin, M.C.M. and Wagner, C. (1975) 84 Bommarius, A.S., Drauz, K.,
J. Biol. Chem., 250, 3746–3751. Hummel, W., Kula, M.-R., and
67 Mihara, H., Muramatsu, H., Wandrey, C. (1994) Biocatalysis, 10,
Kakutani, R., Yasuda, M., Ueda, M., 37–47.
Kurihara, T., and Esaki, N. (2005) FEBS J., 85 Bommarius, A., Schwarm, M., and
272, 1117–1123. Drauz, K. (1998) J. Mol. Catal. B: Enzym.,
68 Thoai, N.V. and Robin, Y. (1959) Bull. Soc. 5, 1–11.
Chim. Biol., 41, 735–742. 86 Kragl, U., Kruse, W., Hummel, W., and
69 Thoai, N.V., Huc, C., Pho, D.B., and Wandrey, C. (1996) Biotechnol. Bioeng.,
Olomucki, A. (1969) Biochim. Biophys. 52, 309–319.
Acta, 191, 446–453. 87 Kragl, U., Vasic-Racki, D., and
70 De Greve, H., Dhaese, P., Seurinck, J., Wandrey, C. (1996) Bioprocess Eng., 14,
Lemmers, M., Van Montagu, M., and 291–297.
Schell, J. (1982) J. Mol. Appl. Gen., 1, 88 Seelbach, K. and Kragl, U. (1997) Enzyme
499–512. Microb. Technol., 20, 389–392.
71 Asano, Y., Yamaguchi, K., and 89 Liese, A., Seelbach, K., and Wandrey, C.
Kondo, K. (1989) J. Bacteriol., 171, (2006) Industrial Biotransformations,
4466–4471. Wiley-VCH Verlag GmbH, Weinheim.
72 Dairi, T. and Asano, Y. (1995) Appl. 90 Galkin, A., Kulakova, L., Yoshimura, T.,
Environ. Microbiol., 61, 3169–3171. Soda, K., and Esaki, N. (1997) Appl.
73 Bommarius, A., Schwarm, M., Stingl, K., Environ. Microbiol., 63, 4651–4656.
Kottenhahn, M., Huthmacher, K., and 91 Menzel, A., Werner, H., Altenbuchner, J.,
Drauz, K. (1995) Tetrahedron: Asymmetry, and Groger, H. (2004) Eng. Life Sci., 4,
6, 2851–2888. 573–576.
74 Hummel, W. and Kula, M.-R. (1989) Eur. 92 Galkin, A., Kulakova, L., Tishkov, V.,
J. Biochem., 184, 1–13. Esaki, N., and Soda, K. (1995) Appl.
75 Shaked, Z. and Whitesides, G.M. Microbiol. Biotechnol., 44, 479–483.
(1980) J. Am. Chem. Soc., 102, 93 Kuroda, S., Tanizawa, K., Sakamoto, Y.,
7104–7105. Tanaka, H., and Soda, K. (1990)
76 Weckbecker, A. and Hummel, W. (2005) Biochemistry, 29, 1009–1015.
Microb. Enzym. Biotransform., 17, 94 Gr€oger, H., May, O., Werner, H.,
225–237. Menzel, A., and Altenbuchner, J. (2006)
77 Lehr, P., Billich, A., Charpiot, B., Org. Process Res. Dev., 10, 666–669.
Ettmayer, P., Scholz, D., Rosenwirth, B., 95 Hanson, R.L., Singh, J., Kissick, T.P.,
and Gstach, H. (1996) J. Med. Chem., 39, Patel, R.N., Szarka, L.J., and
2060–2067. Mueller, R.H. (1990) Bioorg. Chem.,
78 Zhang, Y.C. and Sammakia, T. (2004) 18, 116–130.
Org. Lett., 6, 3139–3141. 96 Fletcher, M.D., Harding, J.R.,
79 Berkessel, A. and Gr€oger, H. (2005) Hughes, R.A., Kelly, N.M., Schmalz, H.,
Asymmetric Organocatalysis, Wiley-VCH Sutherland, A., and Willis, C.L.
Verlag GmbH, Weinheim. (2000) J. Chem. Soc., Perkin Trans. 1,
80 Hong, E.Y., Cha, M., Yun, H., and Kim, 43–52.
B.-G. (2010) J. Mol. Catal. B: Enzym., 66, 97 Sutherland, A. and Willis, C.L. (1997)
228–233. Tetrahedron Lett., 38, 1837–1840.
81 Li, T., Kootstra, A.B., and Fotheringham, 98 Sutherland, A. and Willis, C.L.
I.G. (2002) Org. Process Res. Dev., 6, (1999) Bioorg. Med. Chem. Lett., 19,
533–538. 1837–1840.
82 Taylor, P.P., Pantaleone, D.P., Senkpeil, 99 Bradshaw, C.W., Wong, C.-H.,
R.F., and Fotheringham, I.G. (1998) Hummel, W., and Kula, M.-R. (1991)
Trends Biotechnol., 16, 412–418. Bioorg. Chem., 19, 29–39.
References j1203
100 Hanson, R.L., Howell, J.M., LaPorte, T.L., 107 Robl, J.A., Sun, C., Stevenson, J.,
Donovan, M.J., Cazzulino, D.L., Ryono, D.E., Simpkins, L.M.,
Zannella, V., Montana, M.A., Cimarusti, M.A., Dejneka, T.,
Nanduri, V.B., Schwarz, S.R., Eiring, R.F., Slusarchyk, W.A., Chao, S., Stratton, L.,
Durand, S.C., Wasylyk, J.M., Parker, W.L., Misra, R.N., Bednarz, M.S., Asaad, M.M.,
Liu, M.S., Okuniewicz, F.J., Chen, B.C., Cheung, H.S., Aboa-Offei, B.E.,
Harris, J.C., Natalie, K.J., Ramig, K., Smith, P.L., Mathers, P.D., Fox, M.,
Swaminathan, S., Rosso, V.W., Pack, S.K., Schaeffer, T.R., Seymour, A.A., and
Lotz, B.T., Bernot, P.J., Rusowicz, A., Trippodo, N.C. (1997) J. Med. Chem., 40,
Lust, D.A., Tse, K.S., Venit, J.J., 1570–1577.
Szarka, L.J., and Patel, R.N. (2000) 108 Hanson, R.L., Schwinden, M.D.,
Enzyme Microb. Technol., 26, 348–358. Banerjee, A., Brzozowski, D.B.,
101 Stengelin, M. and Patel, R.N. (2000) Chen, B.C., Patel, B.P., McNamee, C.G.,
Biocatal. Biotransform., 18, 373–400. Kodersha, G.A., Kronenthal, D.R.,
102 Patel, R.N. (2001) Biomol. Eng., 17, Patel, R.N., and Szarka, L.J. (1999) Bioorg.
167–182. Med. Chem., 7, 2247–2252.
103 Henderson, D.P., Shelton, M.C., 109 Nakajima, N., Tanizawa, K., Tanaka, H.,
Cotterill, I.C., and Toone, E.J. (1997) and Soda, K. (1988) J. Biotechnol., 8,
J. Org. Chem., 62, 7910–7911. 243–248.
104 Augeri, D.J., Robl, J.A., Betebenner, D.A., 110 Bae, H.-S., Hong, S.-P., Lee, S.-G.,
Magnin, D.R., Khanna, A., Robertson, Kwak, M.-S., Esaki, N., and Sung, M.-H.
J.G., Wang, A., Simpkins, L.M., Taunk, P., (2002) J. Mol. Catal. B: Enzym., 17,
Huang, Q., Han, S.-P., Abboa-Offei, B., 223–233.
Cap, M., Xin, L., Tao, L., Tozzo, E., 111 Muramatsu, H., Mihara, H., Kakutani, R.,
Welzel, G.E., Egan, D.M., Yasuda, M., Ueda, M., Kurihara, T., and
Marcinkeviciene, J., Chang, S.Y., Esaki, N. (2004) Tetrahedron: Asymmetry,
Biller, S.A., Kirby, M.S., Parker, R.A., and 15, 2841–2843.
Hamann, L.G. (2005) Med. Chem., 48, 112 Yasuda, M., Ueda, M., Muramatsu, H.,
5025–5037. Mihara, H., and Esaki, N. (2006)
105 Gwaltney, S.L. II and Stafford, J.A. (2005) Tetrahedron: Asymmetry, 17, 1775–1779.
Annu. Rep. Med. Chem., 40, 149–165. 113 Kusakabe, H., Kodama, K., Kuninaka, A.,
106 Hanson, R.L., Goldberg, S.L., Yoshino, H., Misono, H., and Soda, K.
Brzozowski, D.B., Tully, T.P., (1980) J. Biol. Chem., 255, 976–981.
Cazzulino, D., Parker, W.L., 114 Kato, Y., Yamada, H., and Asano, Y. (1996)
Lyngberg, O.K., Vu, T.C., Wong, M.K., J. Mol. Catal. B: Enzym., 1, 151–160.
and Patel, R.N. (2007) Adv. Synth. Catal., 115 Priestley, N.D. and Robinson, J.A. (1989)
349, 1369–1378. Biochem. J., 261, 853–861.
j1205
29
Industrial Application of Oxidoreductase catalyzed
Reduction of Ketones and Aldehydes
Katharina G€otz, Lutz Hilterhaus, and Andreas Liese
29.1
Introduction
This chapter addresses biotransformations in the field of the reduction of ketones

and aldehydes applied in industry. Numerous processes benefit from high regio-,
stereo-, or chemoselectivity, as well as high product purity by performing enzymatic
transformations, often resulting in simplified downstream processing. The enzy-
matic reduction of a carbonyl function catalyzed by oxidoreductases yields a broad
range of chiral alcohols or chiral amines, which play an important role as building
block in organic synthesis. A good overview of industrial relevant processes can be
found in the literature [1–3] and examples corresponding to carbonyl reduction are
summarized in Table 29.1. We focus in this chapter on examples that underline the
principles of cofactor regeneration systems and strategies to overcome limitations.
29.2
Reduction Processes Using Whole Cells
Whole-cell biotransformations for biocatalytic ketone reduction often benefit from

high enzyme stability, as the biocatalyst is located in its natural environment, efficient
in situ cofactor generation within the cell, and internal cofactor regeneration [2].
In the industrial production of the carbonic anhydrase inhibitor 6 the main
challenge was to establish a quantitative and enantioselective process for the
reduction of the keto-intermediate 3, which undergoes racemization in aqueous
media above pH 5 (Scheme 29.1).
AstraZeneca developed a sustainable process in which a natural plastics (1) is
depolymerized and further converted into methylketosulfone (6S)-3. To prevent
accumulation of the undesired by-product (6R)-3 the reduction is carried out using
whole cells of Neurospora crassa at pH 4. Thereby, the alcohol dehydrogenase (ADH)
only converts the (6S)-enantiomer (6S)-3 into the corresponding alcohol (4S,6S)-5,
with an enantiomeric excess >98%. The equilibrium of both enantiomers of 3 is
constantly shifted towards the (6S)-enantiomer, leading to an increase in product
and Oliver May.
1206
Table 29.1 Exemplary industrial biotransformations for carbonyl reduction. Oxidoreductases are applied as whole cells or as isolated enzymes
for example, in combination with formate dehydrogenase (FDH) or glucose dehydrogenase (GDH) for cofactor regeneration. A wide range of chiral alcohols
and amines is thereby accessible [1].
Substrate Product Company Biocatalyst
Zeneca Life Science Whole cells; Neurospora crassa

Molecules
5-6-Dihydro-6-methyl-4H-thieno 5,6-Dihydro-4-hydroxy-6-methyl-4H-thieno
[2,3b]thiopyran-4-one-7,7-dioxide [2,3b]thiopyran-7,7-dioxide
Forschungszentrum Isolated enzyme; Rhodococcus

J€
ulich erythropolis; þ FDH
1-Phenyl-2-propanone (S)-1-Phenylpropan-2-ol
Forschungszentrum Isolated enzyme; Rhodococcus

J€
ulich erythropolis þ FDH
(S)-4-Phenylbutan-2-one (S)-4-Phenylbutan-2-ol
Bristol-Myers Squibb Cell extract; Acinetobacter cal-

j 29 Industrial Application of Oxidoreductase catalyzed Reduction of Ketones and Aldehydes
coaceticus; þ GDH
6-Benzyloxy-3,5-dioxo-hexanoic (3R,5S)-6-Benzyloxy-3,5-dihydroxy-hexanoic
acid ethyl ester acid ethyl ester
Eli Lilly Whole cells; Zygosaccharomyces
rouxii
3,4-Methylenedioxyacetophenone 4-(3,4-Methylenedioxyphenyl)-
2-propanol
Ciba-Geigy Isolated enzyme; Staphylococcus

epidermidis; þ FDH
2-Oxo-4-phenylbutyric acid (OPBA)
(R)-2-Hydroxy-4-phenylbutyric acid
(2-HPBA)
Bristol-Myers Squibb Whole cells; Geotrichum candi-

dum SC5469
4-Chloro-3-oxo-butanoic acid methyl ester (S)-4-Chloro-3-hydroxy-butanoic acid methyl
ester
Merck Research Whole cells; Candida sorbophila

Laboratories
2-(4-Nitrophenyl)-N-(2-oxo-2-pyridin-3-ylethyl) (R)-N-(2-Hydroxy-2-pyridin-3-ylethyl)-2-(4-
acetamide nitrophenyl)acetamide
Evonik Industries Isolated enzyme; Bacillus

sphaericus; þ FDH
Trimethylpyruvic acid L-tert-Leucine
29.2 Reduction Processes Using Whole Cells
(Continued )
j1207
1208
Substrate Product Company Biocatalyst
Evonik Industries Isolated enzyme;

Bacillus sphaericus; þ FDH
4,4-Dimethyl-2-oxopentanoic acid L-Neopentylglycine
Evonik Industries Isolated enzyme;

Bacillus sphaericus; þ FDH
3,3-Dimethyl-2-oxopentanoic acid L-3,3-Dimethylpropane glycine

sphaericus; þ FDH
3-Ethyl-3-methyl-2-oxopentanoic acid L-3-Ethyl-3-methylpropane glycine

sphaericus; þ FDH
5,5-Dimethyl-2-oxohexanoic acid L-5,5-Dimethylbutyl glycine
29.2 Reduction Processes Using Whole Cells j1209
Scheme 29.1 Synthesis of TrusoptTM (6). The enzymatic reduction step using whole cells of
Neurospora crassa is highlighted.
yield compared to chemical synthesis. The excellent selectivity of the biotransfor-

mation is also reflected by the chemical purity of >99%, which facilitates the
downstream processing.
This biotransformation is carried out in a fed batch mode on a multi-ton scale. The
crystallized hydroxysulfone 5 can be further converted into compound 6, which is
traded as TrusoptÔ, an ophthalmic pharmaceutical topically applied for the treat-
ment of glaucoma [1, 4–8].
A further interesting biotransformation catalyzed by whole cells is carried out by
Eli Lilly (Scheme 29.2). Here, the (S)-configured (3,4,-methylenedioxyphenyl)-2-
Scheme 29.2 Whole-cell biotransformation for enantioselective reduction of 3,4-

methylenedioxyacetophenone (7) using whole cells of Zygosaccharomyces rouxii.
1210
Figure 29.1 Flow scheme for the production of (S)-(3,4-methylenedioxyphenyl)-2-propanol (8)

using XAD-7 resin for substrate/product adsorption and whole cells of Zygosaccharomyces rouxii.
propanol (8) is produced in 96% yield, >99.9% enantiomeric excess, and with a
chemical purity of 95%.
As both the substrate and product are toxic, when applied in higher concentra-
tions, the productivity is limited. To overcome this limitation a reaction system was
developed whereby the substrate is loaded on XAD-7 resin. The adsorbed substrate
is then continuously released to the reaction medium and converted by whole cells
of Zygosaccharomyces rouxii. As the product also has a high affinity for the
adsorbent complete conversion can be achieved without damaging the cells. In
addition the downstream processing becomes simple. After filtration, the resin is
retained and the product is released by washing with acetone. Figure 29.1 shows
this process.
This reactor concept coped with the substrate limitations to reach a space–time
yield of 75 g l1 day1. The alcohol 8 is the key intermediate for the synthesis of
Talampanel (10, Scheme 29.3), which is being investigated for the treatment of
amyotrophic lateral sclerosis [9–14].
The main demands on biotransformation processes are high substrate concen-
trations, high yield, good enantioselectivity, and short reaction times. In the previous
example, substrate limitation was overcome by adsorption. Processes also exist in
which organic solvents are added to the reaction media to increase the solubility of
substrates and products (see Scheme 29.9 below). A valuable concept for the
production of a broad range of (R)- and (S)-alcohols at high substrate concentrations
was developed by Degussa AG (nowadays Evonik Industries) (Scheme 29.4). Thereby,
so-called tailor made or designer cells were designed carrying an (R)-specific
ADH (alcohol dehydrogenase) from Lactobacillus kefir or an (S)-specific ADH from
Rhodococcus erythropolis in combination with glucose dehydrogenase from either
Bacillus subtilis or Thermoplasma acidophilum.
29.3 Reduction Processes Using Isolated Enzymes j1211
Scheme 29.3 Further conversion of (S)-(3,4-methylenedioxyphenyl)-2-propanol (8) towards

Talampanel (LY300164) (10).
Scheme 29.4 Designer cells for the production of either (R)- or (S)-alcohols 12.
The whole cells are suspended in water or buffer and ketone 11 is added in
concentrations up to 1 M or even higher. Thereby, the solubility limit is exceeded and
a second phase or an emulsion is formed. The system does not suffer from mass transfer
limitations, suggesting that the present organic phase contributes to cell membrane
permeabilization. Figure 29.2 underlines the broad applicability of the system.
Typically, ketone 11 is converted within 1–3 days to the corresponding alcohol 12,
whereby high conversions >90% and at least 90% e.e. are reached, making the
process economic and the downstream processing simple. After filtration of the cells
and acidification of the media, the alcohol is extracted in >95% purity and >90%
yield [15–18].
29.3
Reduction Processes Using Isolated Enzymes
29.3.1
Approaches for In Situ Cofactor Regeneration
The application of isolated oxidoreductases in redox reactions is only economically

attractive when expensive cofactors are recycled. The next two subsections discuss the
1212
Figure 29.2 Product spectrum of chiral alcohols produced with designer cells.
two major approaches for in situ cofactor regeneration, namely, substrate-coupled

and enzyme-coupled systems. In recent decades, these methods have become
essential on both laboratory and industrial scales, especially for the synthesis of
enantiomerically pure intermediates in life science and for pharmaceutical use.
29.3.1.1 Substrate-Coupled Cofactor Regeneration

In the substrate-coupled approach a cheap cosubstrate is applied to regenerate the
cofactor and to shift the thermodynamic equilibrium towards the desired product.
Wacker Chemie successfully established a well-thought-out process for the enzy-
matic production of chiral b-hydroxyesters. As an example, Scheme 29.5 shows the
biotransformation of the prochiral ketoester 13 via alcohol dehydrogenase (ADH)
from Lactobacillus brevis.
The reaction is carried out in batch mode with a capacity of 35 t a1, yielding 96% of
the enantiomerically pure hydroxyester 14 (e.e. >99.8%) with a space–time yield of
92 g l1 day1. Efficient cofactor regeneration is guaranteed via oxidation of isopro-
panol. Owing to continuous removal of the coproduct acetone (16) in a so-called
stripping-process the equilibrium is shifted towards complete conversion.
Scheme 29.5 Substrate-coupled approach for the asymmetric reduction of ethyl 3-oxobutanoate
(13) performed by Wacker Chemie using an ADH from Lactobacillus brevis [1].
Figure 29.3 Flow scheme of the biocatalytic process for the reduction of a ketoester with an internal
stripping process [1].
The reactor set-up highlighted in Figure 29.3 is a powerful tool for enantioselective
reduction of prochiral ketones in good to excellent yields. By means of extraction, the
product is isolated from the reaction mixture and the aqueous phase is recycled,
resulting in a dramatic reduction of contaminated aqueous waste. (R)-Ethyl 3-
hydroxybutyrate (14) is finally distilled and sold as an important building block in
organic synthesis, for example, for the production of b-lactams (Scheme 29.6) and
other pharmaceuticals, agrochemicals, and fragrances [19–22].
Scheme 29.6 Example of the further conversion of hydroxy-esters: methyl ester 17 is transformed
into versatile intermediate 19 in the production of b-lactam antibiotics.
1214
29.3.1.2 Enzyme-Coupled Cofactor Regeneration

Cofactor regeneration systems via enzyme-coupled reactions are also well estab-
lished in industry. The key benefits of formate dehydrogenases (FDHs, mostly
NAD þ specific) and glucose dehydrogenase (GDH, prefers NADP þ ) for cofactor
recycling are, especially, their proven applicability at technical and industrial scales,
low price, high accessibility, activity, and stability.
Ciba Spezialit€atenchemie AG utilizes two enzymes, namely, (R)-lactate-NAD
oxidoreductase from Staphylococcus epidermidis (LDH, lactate dehydrogenase) and
formate dehydrogenase from Candida boidinii (FDH), for the continuous production
of (2R)-hydroxy-4-phenylbutyric acid (21) in a stirred tank reactor, reaching a
space–time yield of 410 g l1 day1 and excellent enantiomeric excess (99.9% e.e.)
(Scheme 29.7) [23, 24].
Scheme 29.7 Continuous production of (R)-2-hydroxy-4-phenylbutyric acid (21) by Ciba

Spezialit€atenchemie AG via an enzyme-coupled cofactor regeneration process [1].
Owing to efficient FDH-catalyzed oxidation of formate (total turnover number:

900) the NAD þ concentration can be kept low. Additives like mercaptoethanol and
EDTA increase enzyme stability, making this process more economic. Here again,
the reactor set-up is kept simple. An ultrafiltration membrane retains the biocatalysts
and the product is easily extracted from the permeate (Figure 29.4).
Figure 29.4 Flow scheme of a continuous reduction process in a stirred tank reactor with an
ultrafiltration step for enzyme retention [1].
The recrystallized alcohol 21 serves as a chiral building block for the synthesis of
different ACE-inhibitors (ACE ¼ angiotensin converting enzymes), which are impor-
tant drugs in treating hypertension or congestive heart failures. For example, the
hydroxyester 24 can be further converted into the ACE-inhibitor Trandolapril (25,
Scheme 29.8).
Scheme 29.8 An example of the use of hydroxyester 24 is as a building block for the synthesis of
Trandolapril (25).
Beside the requirements for efficient cofactor regeneration, high volumetric pro-
ductivity is aspired to so as to render a process profitable. A common limitation in
biotransformation is the low solubility of hydrophobic substrates in an aqueous phase.
Degussa AG focused on this problem and came up with a two-phase system combined
with an alcohol dehydrogenase-based coupled enzymatic reaction system. Thereby, an
alcohol dehydrogenase (ADH) from Rhodococcus erythropolis and a mutant of the
formate dehydrogenase (FDH) from Candida boidinii were used and the concept of
the two-phase reaction system was successfully proven with different substrates on the
laboratory scale. Scheme 29.9 shows the biocatalytic reduction of phenoxyacetone (26).
Since many enzymes are sensitive towards organic solvents, the main challenge
was the development of an aqueous–organic medium in which both enzymes are
stable for a long time period. A water–n-hexane biphasic system (4 : 1) proved to be a
suitable mixture that was enzyme compatible and permitted an increase in substrate
concentration from about 20 up to 200 mM. Exemplarily, the conversion (>95%) of
ketone 26 to an important chiral building block 27 (>99% e.e.) for pharmaceutically
active molecules is shown in Scheme 29.9. The reaction takes place in aqueous
solution while the organic phase guarantees a high volumetric productivity as high
substrate and product concentrations are reached within this phase [25].
Codexis Inc. holds many patents on enzymatic processes for the production of 4-
substituted 3-hydroxybutyric acid derivatives and is still working on further optimi-
zation. The company designed a two-step, three enzyme process in which ethyl 4-
chloro-3-ketoburyrate (29) is converted into the key intermediate 31 for the produc-
tion of LipitorÒ (33) a cholesterol lowering drug (Scheme 29.10).
1216
Scheme 29.9 Two-phase system for the asymmetric reduction of phenoxyacetone (26) via an
enzyme-coupled approach.
Scheme 29.10 Multistep process for the synthesis of LipitorÒ (33).

In the first step ketone 29 is transformed by a mutant of an alcohol dehydrogenase
(ADH) into the desired (R)-alcohol 30 with 95% yield, in excellent chemical (>98%)
and chiral purity (>99.9%) within 8 h (Scheme 29.11). Cofactor regeneration is
performed via oxidation of glucose by a mutant glucose dehydrogenase (GDH).
Enzyme mutants were designed by means of DNA shuffling technology and high-
throughput screening. Selected mutants fulfill the demands of high activity, stability,
and enantioselectivity under desired reaction conditions.
Scheme 29.11 Process for the enzymatic reduction of ketoester 29 in an enzyme-coupled system [1].
In this example, the enzyme is stable at substrate concentration up to 160 g l1 and
the enzyme loading could be drastically reduced, which facilitates the downstream
process. The isolated alcohol 30 is then further converted by the newly designed
halohydrin dehalogenase (HHDH) (Scheme 29.12).
Scheme 29.12 Further enzymatic conversion of hydroxyester 30.
The precursor 31 is thereby produced in a highly economic and environmentally

attractive process [26–29].
Another industrial process that combines an alcohol dehydrogenase (ADH) and a
glucose dehydrogenase (GDH) is the enzymatic production of 6-benzyloxy-(3R,5S)-
dihydroxy-hexanoic acid ethyl ester (38) performed by Bristol-Myers Squibb
(Scheme 29.13).
In batch mode enantiomerically pure (99% e.e.) dihydroxy compound 38 is
produced (92% yield). Even if a substrate concentration of 36 mM is low, the process
is superior to chemical multistep synthesis. The hydrophobic product is easily
extracted from the aqueous phase and sold as an important building block for the
synthesis of anti-cholesterol drugs (e.g., 41,Scheme 29.14) [30–32].
1218
Scheme 29.13 Production of dihydroxy compound 38 with enzyme-coupled cofactor regeneration [1].
Scheme 29.14 Further conversion of 6-benzyloxy-(3R,5S)-dihydroxyhexanoic acid ethyl ester (38)

into the anti-cholesterol drug 41.
29.4
Reductive Amination in Industry
The market for natural and non-natural enantiomerically pure L-amino acids is vast,
as they are applied in various fields, for example, food industry, organic synthesis, and
pharmaceutical and cosmetic industries. One of the first industrial platforms for the
production of L-tert-leucine (43) on a tons scale was based on two enzymes, namely,
leucine dehydrogenase (LeuDH) from Bacillus sphaericus and formate dehydroge-
nase (FDH) from Candida boidinii (Scheme 29.15).
This reaction was carried out in a repetitive batch mode with a space–time yield
of 638 g l1 day1 and amino acid 43 was isolated in 74% yield. The same system
29.4 Reductive Amination in Industry j1219
Scheme 29.15 Enzyme-coupled system for the production of L-tert-leucine (43) [1].
could also be used for the production of non-natural amino acids, for example,
L-neopentylglycine. In ongoing interdisciplinary research a so-called second-gen-
eration process was developed and patented that is based on an efficient and less
expensive whole-cell biotransformation.
In this new approach recombinant whole cells from Escherichia coli are used that
host two plasmids, one carrying the gene encoding for leucine dehydrogenase from
Bacillus cereus and the other carrying the gene encoding for a formate dehydrogenase
mutant from Candida boidinii. Thereby, the expression of both enzymes was
successfully investigated. They are co-expressed in comparable activity. In contrast
to the application of isolated enzymes, no external addition of cofactor is needed and
cost-intensive down-stream processing for enzyme isolation is no longer required.
For the production of L-neopentylglycine (46) the biocatalyst converts the keto-acid 45
(88 g l1) in the presence of ammonium formate (3 equivalents) within 24 h
(Scheme 29.16). The cells are then separated and the product is isolated by means
of ion-exchange chromatography, affording enantiomerically pure (>99% e.e.) L-
amino acid 46 (83% yield).
Scheme 29.16 Transformation of keto-acid 45 into L-neopentylglycine (46) by whole-cell

biocatalysts.
Designer cells are a powerful tool for biotransformations. In ongoing work

designer cells were upgraded by an additional plasmid carrying a gene encoding
for a transaminase (TA). Thereby, the system can be used for the production of a
broad range of amines (Scheme 29.17) [33–37].
This reaction system was established in a fed-batch mode and a two-phase system
is patented by Evonik Industries. The well-understood reductive amination thereby
plays a key role, as the amino-function of the amino acid is continuously transferred
via transaminase, yielding new amines or amino acids.
1220
Scheme 29.17 Catalytic cycle for the production of amines/amino acids in a whole-cell/three-
enzyme system. Involved enzymes: formate dehydrogenase (FDH), for example, leucine
dehydrogenase (LeuDH), and a transaminase (TA).
29.5
Summary
Enzymatic reductions catalyzed by oxidoreductases have become a powerful tool in

industry. Even though some of the above-mentioned processes were established
years ago, research in this field is still ongoing, as can be seen by the number of new
patents and publications. The challenge of cost reduction due to the stoichiometric
demand of cofactor is overcome by the different regeneration systems discussed in
this chapter. Downstream processing is often simple as high conversions are reached
and the desired product is obtained in more than 95% chemical purity. Moreover,
strategies to overcome limitations like low substrate solubility in aqueous phase or
methods to increase biocatalysts half-life should have given an insight into how
processes can be optimized to render them economical. In future, the high demand
for environmental friendly and energy-saving processes will attract more scientists
and industries to engage in green technology.
References j1221
References
1 Liese, A., Seelbach, K., and Wandrey, C. polymeric adsorbent resins. Enzyme
(eds) (2006) Industrial Biotransformations, Microb. Technol., 20, 494–499.
2nd edn, Wiley-VCH Verlag GmbH, 10 Zmijewski, M.J., Vicenzi, J., Landen, B.E.,
Weinheim. Muth, W., Marler, P., and Anderson, B.
2 Goldberg, K., Schroer, K., L€ utz, S., and (1997) Enantioselective reduction of 3,4-
Liese, A. (2007) Biocatalytic ketone methylene-dioxyphenyl acetone using
reduction – a powerful tool for the Candida famata and Zygosaccharomyces
production of chiral alcohols – part II: rouxii. Appl. Microbiol. Biotechnol., 47,
whole-cell reductions. Appl. Microbiol. 162–166.
Biotechnol., 76, 249–255. 11 Anderson, B.A., Hansen, M.M.,
3 Goldberg, K., Schroer, K., L€ utz, S., and Harkness, A.R., Henry, C.L., Vicenzi, J.T.,
Liese, A. (2007) Biocatalytic ketone and Zmijewski, M.J. (1995) Application of
reduction – a powerful tool for the a practical biocatalytic reduction to an
production of chiral alcohols – part II: enantioselective synthesis of the 5H-2,3-
processes with isolated enzymes. Appl. benzodiazepine LY300164. J. Am. Chem.
Microbiol. Biotechnol., 76, 237–248. Soc., 117, 12358–12359.
4 Federsel, H.-J. (2005) Asymmetry on large 12 Zaks, A. and Dodds, D.R. (1997)
scale: the roadmap to stereoselective Application of biocatalysis and
processes. Nat. Rev. Drug Discovery., 4, biotransformations to the synthesis of
685–697. pharmaceuticals. Drug Discovery Today, 2,
5 Blacker, A.J. and Holt, R.A. (1997) CT 513–530.
Development of a multistage chemical 13 Iwamoto, F.M., Kreisl, T.N., Kim, L., Duic,
and biological process for an optically J.P., Butman, J.A., Albert, P.S., and Fine,
active intermediate for an anti-glaucoma H.A. (2010) Phase 2 trial of talampanel, a
drug, in Chirality in Industry II (eds A.N. glutamate receptor inhibitor, for adults
Collins, G.N. Sheldrake, and J. Crosby), with recurrent malignant gliomas. Cancer,
John Wiley & Sons, Inc., New York, 116, 1776–1782.
pp. 261–264. 14 Rascuzzi, R.M., Shefner, J., Chappell,
6 Holt, R.A. (1996) Microbial asymmetric A.S., Bjerke, J.S., Tamura, R., Chaudhry,
reduction in the synthesis of a drug V., Clawson, L., Haas, L., and Rothstein,
intermediate. Chim. Oggi., 9, 17–20. J.D. (2009) A phase II trial of talampanel in
7 Holt, R.A. and Rigby, S.R. (1996) Process subjects with amyotrophic lateral
for microbial reduction producing 4(S)- sclerosis. Amyotroph. Lateral Sclerosis, 11,
hydroxy-6(S)-methyl-thienopyran 266–271.
derivatives, Zeneca Limited, 15 Osswald, S., Doderer, K., Gr€oger, H., and
US 5580764. Wienand, W. (2007) Alcohol
8 Blacklock, T.J., Sohar, P., Butcher, J.W., dehydrogenases whole-cell catalysts – a
Lamanec, T., and Grabowski, E.J.J. (1993) broad technology platform for life science
An enantioselective synthesis of the applications. Chim. Oggi., 25, 16–18.
topically active carbonic anhydrase 16 Gr€oger, H., Chamouleau, F., Orologas, N.,
inhibitor MK-0507: 5,6-dihydro-(S)-4- Rollmann, C., Drauz, K., Hummel, W.,
(ethylamino)-(S)-6-methyl-4H-thieno Weckbecker, A., and May, O. (2006)
[2,3b]thiopyran-2-sulfonamide 7,7 dioxide Enantioselective reduction of ketones with
hydrochloride. J. Org. Chem., 58, designer cells at high substrate
1672–1679. concentrations: highly efficient access to
9 Vicenzi, J.T., Zmijewski, M.J., Reinhard, functionalized optically active alcohols.
M.R., Landen, B.E., Muth, W.L., and Angew. Chem. Int. Ed., 45, 5677–5681.
Marler, P.G. (1997) Large-scale 17 Gr€oger, H., Rollmann, C., Chamouleau,
stereoselective enzymatic ketone F., Sebastien, I., May, O., Wienand, W.,
reduction with in situ product removal via and Drauz, K. (2006) Enantioselective
1222
reduction of 4-fluoroacetophenone 26 Davis, C., Grate, J., Gray, D., Gruber, J.,
at high substrate concentration Huisman, G., Ma, S., Newman, L.,
using a tailor-made recombinant whole- Sheldon, R., and Wang, L. (2004)
cell catalyst. Adv. Synth. Catal., 349, Enzymatic process for the production of 4-
709–712. substituted 3-hydroxybutyric acid
18 Berkessel, A., Rollmann, C., Chamouleau, derivatives, WO 015132.
F., Labs, S., May, O., and Gr€oger, H. (2007) 27 Davis, C., Jenne, S., Drebber, A.,
Practical two-step synthesis of an Huisman, G., and Newman, L. (2005)
enantiopure aliphatic terminal (S)- Improved ketoreductase polypeptides and
epoxide based on reduction of relate polynucleotides, WO 017135.
haloalkanones with designer cells. Adv. 28 Davis, C., Grate, J., Gray, D., Gruber, J.,
Synth. Catal., 349, 2697–2704. Huisman, G., Ma, S., Newman, L.,
19 Daussmann, T., Hennemann, H.-G., Sheldon, R., and Wang, L. (2005)
Rosen, T.C., and D€ unkelmann, P. (2006) Enzymatic processes for the production of
Enzymatische technologien zur synthese 4-substituted 3-hydroxybutyric acid
chiraler alkohol-derivate. Chem. Ing. Tech., derivatives and vicinal cyano, hydroxyl
78, 249–255. substituted carboxylic acid esters, WO
20 Daussmann, T., Rosen, T.C., and 018579.
D€unkelmann, P. (2006) Oxidoreductases 29 Ma, S., Gruber, J., Davis, C., Newman, L.,
and hydroxynitrilase lyases: Gray, D., Wang, J., Grate, J., Huisman, G.,
complementary enzymatic technologies and Sheldon, R. (2010) A green-by-design
for chiral alcohols. Eng. Life Sci., 6, biocatalytic process for atorvastatin
125–129. intermediate. Green Chem., 12, 81–86.
21 Hummel, W. and Riebel, B. (2003) 30 Patel, R.N. (2001) Enzymatic synthesis of
Alkohol-Dehydrogenase und deren chiral intermediates for drug
verwendung zur enzymatischen development. Adv. Synth. Catal., 343,
herstellung chiraler 527–546.
hydroxyverbindungen, EP 0796914 A2. 31 Patel, R.N., McNamee, C.G., Banerjee, A.,
22 Cainelli, G., Galletti, P., and Giacomini, D. and Szarka, L.J. (1993) Stereoselective
(1998) A practical synthesis of a key microbial or enzymatic reduction of 3,5-
intermediate for the production of b- dioxo esters to 3-hydroxy-5-oxo, 3-oxo-5-
lactam antibiotics. Tetrahedron Lett., 39, hydroxy and 3,5-dihydroxy esters, EP
7779–7782. 0569998A2.
23 Schmidt, E., Ghisalba, O., Gygax, D., and 32 Patel, R.N., Banerjee, A., McNamee, C.G.,
Sedelmeier, G. (1992) Optimization of a Brzozowski, D., Hanson, R.L., and Szarka,
process for the production of (R)-2- L.J. (1993) Enantioselective microbial
hydroxy-4-phenylbutyric acid – an reduction of 3,5-dioxo-6-(benzyloxy)
intermediate for inhibitors of angiotensin hexanoic acid, ethyl ester. Enzyme Microb.
converting enzyme. J. Biotechnol., 24, Technol., 15, 1014–1021.
315–327. 33 Sit, S.Y., Parker, R.A., Motoc, I., Han, W.,
24 Schmidt, E., Blaser, H.U., Fauquex, P.F., and Balasubramanian, N. (1990)
Sedelmeier, G., and Spindler, F. (1993) Synthesis, biological profile and
Comparison of chemical and biochemical quantitative structure-activity relationship
reduction methods for the synthesis of of a series of novel 3-hydroxy-3-
(R)-2-hydroxy-4-phenylbutyric acid. methylglutaryl coenzyme A reductase
ChemInform, 24. inhibitors. J. Med. Chem., 33, 2982–2999.
25 Gr€oger, H., Hummel, W., Buchholz, S., 34 de Wildeman, S.M.A., Sonke, T.,
Drauz, K., van Nguyen, T., Rollman, C., Schoemaker, H.E., and May, O. (2007)
H€usken, H., and Abokitse, K. (2003) Biocatalytic reductions: from lab curiosity
Practical asymmetric enzymatic reduction to first choice. Acc. Chem. Res., 40,
through discovery of a dehydrogenase- 1260–1266.
compatible biphasic reaction media. Org. 35 Gr€oger, H., May, O., Werner, H., Menzel,
Lett., 5, 173–176. A., and Altenbuchner, J. (2006) A second-
References j1223
generation process for the synthesis of L- enantiomerically enriched amines, EP
neopentylglycine: asymmetric reductive 2183377.
amination using a recombinant whole cell 37 Kula, M.R. and Wandrey, C. (1987)
catalyst. Org. Process Res. Dev., 10, Continuous enzymatic transformation in
666–669. an enzyme membrane reactor with
36 Doderer, K., Wienand, W., Gr€ oger, H., and simultaneous NADH regeneration.
Rollmann, C. (2009) Process for preparing Methods Enzymol., 136, 9–21.
j1225
Part VII
Oxidations
and Oliver May.
j1227
30
Oxyfunctionalization of CH Bonds
Vlada B. Urlacher and Marco Girhard
30.1
Introduction
Hydroxylation, the conversion of a carbon–hydrogen into a carbon–hydroxyl bond, is

a key reaction of the oxidative metabolism of many organic compounds. It is one of
the most widespread enzyme activities, occurring in all forms of life from bacteria to
humans [1], and is considered as potentially the most useful of all biotrans-
formations [2]. Generally, the selective oxyfunctionalization of inert hydrocarbons
to useful chemicals of a higher oxidation state such as alcohols, carbonyl compounds,
or epoxides represents one of the most challenging topics in industrial (fine)
chemical synthesis [3]. The traditional methods of chemical oxidation, however,
often require the use of expensive, waste-generating, and hazardous reactants, as well
as employing catalysts with low conversion and/or selectivity, yielding product
mixtures that complicate product isolation [4]. Thus, there is a considerable demand
to replace these old technologies with cleaner alternative routes.
Oxygenases that perform such oxidations have attracted the attention of bioche-
mists, biotechnologists, and chemists, not only in academia but also in industry [5].
These biocatalysts are notable for many reasons: (i) they use dioxygen (O2) or
hydrogen peroxide (H2O2) as primary oxidant operate at ambient conditions and
thus present ideal systems for green organic synthesis; and (ii) they often exhibit
exquisite substrate specificities as well as regio- and/or stereoselectivities. Therefore
applications using those biocatalysts often provide compounds that are difficult to
produce by traditional chemical synthetic processes. Significant progress in the
understanding of oxygenase mechanisms has been made and extensive studies have
revealed some of the key principles that underlie their efficacy as biocatalysts. Based
on this knowledge, there are also numerous studies on engineering oxygenases with
altered substrate specificities, increased activity, and enhanced process stability.
In the following sections of this chapter we review CH bond oxyfunctionaliza-
tion. After a short introduction of the respective biocatalysts and the common
mechanism of oxygen activation, a brief summary for each type of oxygenases will
highlight their structure, catalytic mechanism, and some of the reactions catalyzed.
We will focus on substrates whose oxyfunctionalyzed derivatives are high-value
and Oliver May.
j 30 Oxyfunctionalization of CH Bonds
1228
compounds that represent interesting targets for biotechnological fine chemical

production.
30.2
Activation of Molecular Dioxygen
The first step in catalysis by oxygenases utilizing molecular dioxygen involves the
reductive cleavage of the O¼O bond. This reaction is exothermic and therefore, in
principle, energetically favorable. However, despite the strong thermodynamic
driving force, the kinetic reactivity of dioxygen with organic molecules at ambient
temperatures is intrinsically low, which is due to its triplet ground state (a diradical
with two unpaired electrons), whereas all stable organic molecules are singlets (all of
their electrons are paired). Direct reactions between triplet and singlet molecules to
yield a singlet product are spin-forbidden processes because chemical combination
reaction rates are much faster than spin inversion rates.
To overcome this high kinetic barrier oxygenases use either transition metal ions or
flavin cofactors. Flavin-dependent monooxygenases catalyze in most cases epoxida-
tions or Baeyer–Villiger oxidations and therefore will not be considered in this
chapter. In the case of metallo enzymes, iron and copper ions are often the metal ions
of choice for biological oxidation systems because of their abundance in the geo-
sphere, inherent electronic properties, and accessible redox potentials [6]. These ions
are incorporated in the active center of metallo oxygenases, for example, in heme
groups, non-heme diiron, and non-heme iron centers, or copper active sites.
Depending on the number of oxygen atoms that are introduced into the organic
substrate, monooxygenases and dioxygenases can be distinguished. Monooxy-
genases incorporate one oxygen atom from O2 into the substrate, while the second
one is reduced to water. Cytochrome P450s are probably the most famous
representatives of the group of monooxygenases [7]. Their active center contains
heme b, and therefore P450s follow an oxygen activation mechanism referred to as
heme paradigm [8]. Furthermore, there are several groups of non-heme diiron and
non-heme iron containing monooxygenases [9], and also some representatives of
metallo monooxygenases that contain copper ions in the active center, like particulate
methane monooxygenases or dopamine b-monooxygenase.
Another group of oxygenases that catalyze CH bond oxyfunctionalization is
represented by dioxygenases that incorporate both oxygen atoms from O2 into the
substrate. They can either be of the non-heme iron type (e.g., Rieske cis-diol
dioxygenases) [10] or have one or more heme iron units (e.g., indoleamine-2,3-
dioxygenase and tryptophan-2,3-dioxygenase) [11, 12].
The mechanisms by which oxyfunctionalizations are achieved differ strongly
depending on the type of metallo oxygenase. Despite those differences, however,
most mechanisms have in common that dioxygen activation involves the formation
of an initial dioxygen-adduct (superoxo-complex), followed by conversion into a
metal-peroxide (peroxo-complex) and subsequent O¼O bond cleavage to yield a high-
30.3 Heme Metallo Monooxygenases j1229
valent oxidant (oxo-complex). The oxo-complex is the so-called oxygen gun that
attacks and oxidizes the substrate (Scheme 30.1) [6].
Scheme 30.1 Parallels in metallo oxygenase monooxygenase; HPO, heme peroxidase;

mechanisms. All mechanisms involve the sMMO, soluble methane monooxygenase;
formation of an initial dioxygen-adduct RDO, Rieskecis-diol dioxygenase; TDM,
(superoxo), conversion into a metal-peroxide tetrahydropterin-dependent monooxygenase;
(peroxo), and subsequent O¼O bond cleavage KGDO, a-keto acid-dependent dioxygenase.
to yield the high-valent oxidant (oxo). (P), Adapted by permission from Macmillan
porphyrin; e, electron; P450, cytochrome P450 Publishers Ltd.: [6].
30.3
Heme Metallo Monooxygenases
30.3.1
Cytochrome P450 Monooxygenases
The most extensively studied enzymes with the ability to oxyfuntionalyze CH bonds
are cytochrome P450 monooxygenases (P450 or CYP). P450s belong to an ever-
growing superfamily of heme b containing monooxygenases found in all domains of
life [13]. They play a central role in drug metabolism and are involved in the
biosynthesis of important natural compounds. The numbers of P450 sequences
are constantly increasing. Currently there are more than 18 000 P450 sequences
available in several online databases [14]. Examples are CYPED1) [15, 16], the
1) http://www.cyped.uni-stuttgart.de (accessed16 August 2010).

1230
Fungal Cytochrome P450 Database2) listing more than 6800 fungal P450
sequences [17], and The cytochrome P450 homepage3) of D. Nelson, which
provides a classification for more than 12 000 P450 sequences [18].
Extensive studies have revealed the key chemical principles that underlie the
efficacy of P450s as biocatalysts for aerobic oxidations and several comprehensive
reviews and books on P450s have been written in the last decade [7, 19–26];
Reference [27] provides a survey of selected relevant publications on P450s published
between 2004 and 2006.
P450s were recognized and defined as a distinct class of heme-containing proteins
about 50 years ago [28, 29]. The name P450 is due to their unusual property to form
reduced (ferrous) iron/carbon monoxide complexes in which the heme absorption
Soret band shifts from 420 to 450 nm [30, 31]. Essential for this spectral charac-
teristic is the axial coordination of the iron by a cysteine thiolate that is common to all
P450s [32, 33]. This phylogenetically conserved cysteinate is the proximal ligand to the
iron, with the distal ligand generally assumed to be a weakly bound water molecule [34].
Despite relatively low sequence identity across the gene superfamily, crystal structures
of P450s show the same structural organization, with several structurally conserved
regions that are predominantly found in the core of the protein around the heme,
which reflects the common mechanism of electron- and proton-transfer and dioxygen
activation. Substrate recognition and binding is mainly arranged through six substrate
recognition sites (SRS1–SRS6) [35]. Mutations in these regions have a high impact on
substrate specificity. Furthermore, the substrate binding region is very flexible and
often susceptible to structural reorganization upon substrate binding, which accounts
for the broad substrate spectra of many P450s [36].
The catalytic cycle of P450s is by now well studied and was revised by Sligar and
colleagues in 2005 (Scheme 30.2) [37]. Briefly, substrate binding in the active site
induces the dissociation of a water molecule that is bound as sixth coordinating
ligand to the iron (1), thereby inducing a shift of the heme iron spin state from low-
spin to high-spin along with a positive shift in the reduction potential on the order of
130–140 mV [38]. The increased potential allows the delivery of the first electron,
which reduces the heme iron from the ferric (FeIII) (2) to the ferrous (FeII) form (3).
After the first electron transfer, the FeII binds dioxygen, resulting in a ferrous
superoxo-complex (4). The consecutive delivery of the second electron converts this
species into a ferric peroxo anion (5a). This species is then protonated to a ferric
hydroperoxy-complex (5b), which is known as compound 0. Next, protonation of the
ferric hydroperoxy-complex results in the so-called compound I – a high-valent ferryl-
oxo-complex (6). This process is accompanied by the release of a water molecule
through heterolytic scission of the dioxygen bond in the preceding intermediate (5b).
The exact mechanistic details of oxygen insertion into the CH bond are still a
subject of discussion, although it is widely accepted that compound I (6) is the
oxygenating species that transfers the activated oxygen atom to the substrate. The
most popular hypothesis is the so-called rebound mechanism, in which oxygen
2) http://p450.riceblast.snu.ac.kr (accessed 5 September 2010).

3) http://drnelson.uthsc.edu/CytochromeP450.html (accessed 5 September 2010).
H H RH RH
O RH e-
FeIII FeIII Fe II
S S S
Cys Cys Cys
1 2 3
H2 O O 2-
O2
ROH
autoxidation
shunt (-)
R H H2O2 H2O2 O
O 2 e- RH
O
2 H+
Fe III Fe III
peroxide
S S
shunt
Cys Cys
7 4
oxidase
shunt
H+ H+
e-
(-) (2-)
OH O
RH H 2O H+ RH RH
O O O
H+
FeIV FeIII Fe III
S S S
Cys Cys Cys
6 5b 5a
Scheme 30.2 Catalytic cycle of cytochrome P450 monooxygenases Adapted from Reference [22].
Copyright Wiley-VCH Verlag GmbH. Reproduced with permission.
insertion occurs through abstraction of one hydrogen atom from the substrate to give
a radical intermediate (8) followed by oxygen rebound to form COH (9) as shown in
Scheme 30.3 [39, 40]. The results from numerous studies of kinetics, stereoselectiv-
ity, and isotope effects for the hydroxylation reactions catalyzed by P450s conform to
this proposed mechanism [19]. Alternative hypotheses suggest that other oxy inter-
mediates, such as peroxo-iron, hydroperoxo-iron, or H2O2 coordinated iron, may also
be involved in the reaction cycle [41–43]. Compound 0 (5b), for example, is associated
with the epoxidation of C¼C double bonds [44, 45].
The rebound mechanism is not unique to P450s, but is commonly utilized for
oxygen insertion by several oxygenases, including soluble methane monooxygenases
(described in Section 30.4.1), or peptidylglycine a-hydroxylating monooxygenase and
dopamine b-monooxygenase (Section 30.4.3).
Since the oxyfunctionalization process by P450s requires the consecutive delivery of
two electrons to the heme iron, these enzymes utilize reducing equivalents (electrons
1232
P450
H H H
R R R
R' R' R'
H H O
O H-abstraction O rebound H
FeIV FeIV Fe III
S S S
Cys Cys Cys
6 8 9
sMMO
H H H
R R R
R' R' R'
H H OH
O H-abstraction O rebound
FeIV FeIV FeIII FeIV Fe III FeIII
O O O
Q R T
Scheme 30.3 Parallels in the rebound mechanisms of cytochrome P450 monooxygenases (P450)
and soluble methane monooxygenases (sMMOs).
in the form of hydride ions) ultimately derived from the pyridine cofactors NAD(P)H
and transferred to the P450 via special redox proteins [46, 47]. Depending on redox
partners, traditionally, two main classes of P450s were defined [22]. Class I P450s are
found in mitochondria and bacteria and use a small redox [2Fe-2S] iron-sulfur protein
(ferredoxin) and a FAD-containing ferredoxin reductase for transfer of electrons from
NAD(P)H to the P450 component. Microsomal P450s belong to class II P450 redox
systems that exploit a FAD- and FMN-containing cytochrome P450 reductase (CPR) –
and sometimes cytochrome b5 – for transfer of electrons from NADPH. In recent years
numerous genome sequencing projects have revealed many other types of electron
transfer proteins, which belong neither to class I nor to class II [23].
Under certain conditions P450s can also enter one of three so-called uncoupling
pathways (Scheme 30.2). The autoxidation shunt occurs if the second electron is not
delivered to reduce the ferrous superoxy-complex (4), which can decay to form
superoxide. Inappropriate positioning of the substrate in the active site is often the
molecular reason for the two other uncoupling cycles. The ferric hydroperoxy-
complex (5b) can collapse and release hydrogen peroxide (peroxide shunt), while
decay of compound I (6) is accompanied by the release of water (oxidase shunt).
For industrial applications it is particularly important to note that the uncoupling
pathways in all cases consume reducing equivalents from NAD(P)H without
product formation.
It is also notable that the peroxide shunt in some cases can also be utilized by P450s
to incorporate oxygen from H2O2 or other organic peroxides (e.g., cumene hydro-
peroxide or tert-butyl hydroperoxide [48]) as side activity [49–51]. Furthermore, there
is also the class of natural P450 peroxygenases that employ the peroxide shunt for
catalysis exclusively and therefore do not require exogenous redox proteins (see
Section 30.6.2 for further information) [52].
P450s oxidize a vast range of substrates and can catalyze more than 20 different
reaction types [12]. These reactions include oxidation of non-activated sp3 hybridized
carbon atoms, aromatic hydroxylation, epoxidation, CC bond cleavage, heteroatom
oxygenation, heteroatom release (dealkylation), oxidative ester cleavage, oxidative
phenol- and ring-coupling, isomerization via (abortive) oxidation, and oxidative
dehalogenation, as well as other complex reactions like dimer formation via Diel-
s–Alder reactions of products or Baeyer–Villiger-type oxidations [23, 53–55]. We will
focus on CH bond oxyfunctionalizations by P450s in detail in Section 30.6.
30.3.2
Heme Peroxidases
Other heme iron containing biocatalysts capable of oxyfunctionalizations are heme-

peroxidases that utilize one oxygen atom from H2O2 instead of molecular dioxygen
and produce water as coproduct. They therefore neither relay on electron delivery via
redox proteins nor on cofactor regeneration [56, 57]. Heme peroxidases are ubiq-
uitous in nature and belong to the subgroup of haloperoxidases. In their natural
function, which is the oxidative halogenation of organic substrates, heme peroxidases
generally perform one-electron rather than two-electron transfers [58]. However, a
few examples for the reactions where peroxidases under certain conditions can react
as peroxygenases and catalyze the oxyfunctionalization of CH bonds have been
reported in the literature [59–61]. The main obstacles in this respect are the
preference of one-electron versus two-electron transfers by most peroxidases
and the sterically restricted active site, which limits the access of the substrate to
the heme iron [62].
Three peroxidases are of particular interest for oxyfunctionalizations of CH
bonds: chloroperoxidase (CPO) from the fungus Caldariomyces fumago, horseradish
peroxidase (HRP) from Armoracia rusticana, and a haloperoxidase from the fungus
Agrocybe aegerita (AaP). CPO is capable of in vitro oxyfunctionalizations of various
useful compounds, if halide ions are absent. These reactions include allylic [63],
benzylic [64], and propargylic [65, 66] CH hydroxylation, as well as the regioselective
oxidation of indoles (Scheme 30.4) [67, 68]. HRP was shown to catalyze the
hydroxylation of benzene (10) to phenol (11) (Scheme 30.4) [69], as well as other
aromatic compounds, for example, L-tyrosine to L-3,4-dihydroxyphenylalanine,
L-()-phenylephrine to L-epinephrine (adrenaline), or phenol to catechol [56]. Recent-
ly, a second heme-thiolate haloperoxidase in the fungus Agrocybe aegerita (AaP) was
discovered, whose spectral properties bear great resemblance to P450s. AaP pos-
sesses different oxidative activities, including hydroxylation of aromatic rings.
Toluene oxidation by AaP, however, is unselective yielding among other compounds
benzyl alcohol (37%), methyl-p-benzoquinone (23%), benzaldehyde (12%), benzoic
acid (4%), and o- (4%) and p-cresol (2%). In contrast, naphthalene (12) hydroxylation
proceeds regioselectively to afford 1-naphthol (13) and traces (<2%) of 2-naphthol
1234
H H
N N
CPO
R + H2O2 R O + H2O
yield: 87- 97%
HO H
CPO
+ H2O2 + H2O
yield: 26%
ee: 91%
OH
CPO
+ H2O2 + H2O
yield: 20%
ee: 97%
OH
HRP
+ H2O2 + H2O
In benzene containing
1% phosphate buffer
10 yield: 29.1 nmol 11
OH
AaP
+ H2O2 + H2O
yield: 64%
12 13
Scheme 30.4 Regio- and stereoselective CH oxyfunctionalizations catalyzed by chloroperoxidase

(CPO), horseradish peroxidase (HRP), and Agrocybe aegerita peroxidase (AaP).
(Scheme 30.4) [70]. In catalyzing these oxygen-transfer reactions, CPO, HRP, and
AaP exhibit reactivities more typical of P450s than of classical peroxidases.
Consequently, the crystal structure of CPO shows that it shares structural
features with both P450s and heme peroxidases [71, 72]. As in P450s the proximal
heme ligand is a cysteine residue – in contrast with other heme peroxidases, in
which this coordination site is occupied by a histidine. On the other hand, the distal
heme pocket – which constitutes the hydrogen peroxide binding site – is occupied
by polar amino acids, whereas in P450s this pocket is lined primarily with nonpolar,
hydrophobic groups. Oxygenation in CPO proceeds via the abstraction of a proton
from an incoming hydrogen peroxide molecule yielding the peroxo anion coordi-
nated to FeIII. This step is followed by protonation of the terminal oxygen and
cleavage of the O¼O bond resulting in the FeIV-oxo porphyrin cation radical and a
molecule of water. In contrast to other heme peroxidases, which generally have
restricted access to the FeIV-oxo center, in CPO a small opening above the heme is
present, which is more similar to P450s. In the absence of halide ions relatively
small organic substrates can thus access the distal heme pocket and be oxyfunc-
tionalyzed by the FeIV-oxo species [73].
30.4 Non-heme Metallo Monooxygenases j1235
30.4
Non-heme Metallo Monooxygenases
30.4.1
Non-heme Diiron Monooxygenases
The best characterized non-heme diiron containing monooxygenases belong to the

family of bacterial multicomponent monooxygenases (BMM). BMM are soluble,
non-heme diiron containing protein complexes consisting of a Rieske-reductase, a
coupling/regulatory protein, and an oxygenase [74]. Four groups of BMM have been
identified including alkene monooxygenases, arene monooxygenases, and phenol
oxygenases. The fourth and best characterized group is soluble methane monoox-
ygenases (sMMO), which catalyze the oxidation of methane to methanol in metha-
notrophic bacteria. For activity, sMMO require NADH oxidoreductases containing
non-covalently bound FAD and a [2Fe-2S] cluster, as well as a small coupling/
regulatory protein. The oxygenase itself is hexameric (a2b2c2) and incorporates the
binuclear iron active sites. Crystal structures of sMMO in several states are avail-
able [75, 76]. Studies on sMMO using spectral and magnetic techniques coupled with
stopped-flow, rapid freeze quench methodologies, quantum mechanics/molecular
mechanics simulations, and density functional theoretical approaches [77–79] led to
the proposal of a mechanistic pathway for sMMO. This pathway consists of two major
processes: (i) dioxygen activation to give the oxo-intermediate built of a [FeIV.FeIV]
cluster (compound Q; MMOQ) followed by (ii) substrate oxidation (Scheme 30.5) [80].
Transient kinetic studies indicated that compound Q reacts directly with the substrate
to yield a species with hydroxylated product in the active site (compound T; MMOT),
probably via a radical mechanism [81, 82], in analogy to the rebound mechanism
proposed for cytochrome P450 monooxygenases (Scheme 30.3; see Section 30.3.1).
O O
O2
O O
FeII FeII FeII FeII FeIII FeII
MMOred O2° MMOred MMOsuperoxo
H2O
2 H+
O CH4
O O
FeIII FeIII FeIV FeIV FeIII FeIII
O O
CH3OH
MMOT MMOQ MMOperoxo
Scheme 30.5 Catalytic cycle of sMMO.

1236
Methane monooxygenases can be found either in soluble form (sMMO) or in

membrane-bound form (pMMO). pMMO is expressed when copper ions are
available in the culture medium. Its active site is surmised to contain a trimeric
copper cluster [9, 83] and therefore will be described later (Section 30.4.3).
sMMO can oxyfuntionalyze CH bonds of many abundant and potentially toxic
hydrocarbons [82–85], which has led industrial and environmental chemists to
explore applications for their catalytic capability (Scheme 30.6) [86].
Another class of non-heme diiron monooxygenases is represented by membrane-
spanning systems such as alkane monooxygenases (AlkB) and xylene monooxygen-
ase (XylM). Whereas the soluble diiron enzymes like sMMO are relatively well
studied, AlkB and XylM enzymes are less well characterized. Since no crystal
structure has been resolved so far, their active site structures are unknown. For
AlkB, spectroscopic analyses point to eight or nine histidine residues coordinating
the two iron ions and a carboxylate residue bridging the two metals [87, 88]. It is
proposed that both AlkB and XylM utilize a prototypical rebound mechanism similar
to that described for P450s (Section 30.3.1) [89–91].
30.4.2
Tetrahydropterin-dependent Monooxygenases
Tetrahydropterin-dependent monooxygenases (TDM) catalyze hydroxylations of aro-

matic amino acids: phenylalanine (14) to tyrosine (15), tyrosine to 3,4-dihydroxyphe-
nylalanine (L-dopamine; 16), and tryptophan to 5-hydroxytryptophan. They constitute a
fairly small oxygenase family. TDM contain FeII bound to two histidines and a
glutamate in the active site. They utilize tetrahydropterin (17) as cosubstrate that is
converted into 4a-hydroxypterin (18) during the catalytic cycle (Scheme 30.7) [92, 93].
TDM are found in human liver and the central nervous system, but also in a few
bacteria like Pseudomonas species or Chromobacterium violaceum [94, 95]. The exact
mechanism of substrate oxidation has not been resolved yet. It has been shown,
however, that both oxygen atoms in the hydroxylated amino acid and the pterin
product originate from molecular dioxygen. Further, ferric iron-(hydro)peroxo [FeIII-
O-OH] and high-valent iron-oxo [FeIV¼O] intermediates are discussed as promising
reactive iron derivatives [96].
30.4.3
Other Metallo Monooxygenases
Peptidylglycine a-amidating monooxygenase (PAM) and dopamine b-monooxygen-

ase (DbM) belong to a group of copper ion containing enzymes, which are capable of
reductive cleavage of O2 to form hydroxylated products and water [97]. Both enzymes
are of great importance in the central nervous system of high eukaryotes [98]. The
peptidylglycine a-hydroxylating monooxygenase (PHM) component of PAM catalyzes
the transformation of C-terminal glycine-extended peptides (19) to their a-hydroxyl-
ated products (20). DbM catalyzes the hydroxylation of dopamine (21) to norepineph-
rine (22) (Scheme 30.8). Both enzymes exist in soluble and membrane-bound forms
OH
HO
+ O2 + + H2O
NADH + H+ NAD+
HO O
+ O2 + + H2O
NADH + H+ NAD+
+ O2 + + H2O
OH
OH
NADH + H+ NAD+
+ O2 OH + + + H 2O
HO
HO
NADH + H+ NAD+
Scheme 30.6 Examples for CH bond oxyfunctionalizations catalyzed by sMMO.

30.4 Non-heme Metallo Monooxygenases
j1237
1238
OH
N N NH2
NH2 OH
14 NH
N
H
O2 17 OH O
17
18
OH N N NH2
OH
NH2
HO NH
N
15 H
OH
OH O
O2 17 18
18
HO
OH
NH2
HO
16
Scheme 30.7 Hydroxylation of phenylalanine (14) to tyrosine (15) by phenylalanine hydroxylase

and tyrosine to 3,4-dihydroxyphenylalanine (16) by tyrosine hydroxylase. The second oxygen atom in
both steps is transferred to tetrahydropterin (17), yielding 4a-hydroxypterin (18).
and play different physiological roles. However, the chemical mechanisms of these
forms are considered to be the same [98]. X-Ray analysis of PHM [99] and EPR
investigations of DbM [100, 101] revealed that these enzymes possess similar active
sites consisting of a mononuclear copper ion supported by two histidine imidazoles
and one methionine sulfur. Two electrons required for O2 activation are supplied
stepwise from an external reductant, such as ascorbate, through another mononuclear
copper site ligated by three histidine imidazoles [102]. The resulting CuI sites are
returned to the CuII state in the presence of substrate and O2 via a formal ping-pong
mechanism in which reductant and substrates interact with different forms of the
enzyme that are separated by irreversible chemical processes [103]. As key reactive
intermediate in PHM and DbM, a CuII-hydroperoxo- [104] or CuII-superoxo spe-
30.4 Non-heme Metallo Monooxygenases j1239
OH
Peptide OH PHM Peptide OH

N + O2 N + H 2O
H H
O 2e- + 2H+ O
19 20
OH
HO NH2 HO NH2
DßM
+ O2 + H2O
HO 2e- + 2H+ HO
21 22
Scheme 30.8 Reactions catalyzed by peptidylglycine a-hydroxylating monooxygenase (PHM) and

dopamine b-monooxygenase (DbM).
cies [105] have been suggested for the CH bond activation via a rebound mechanism
(like that in cytochrome P450 monooxygenases; see Section 30.3.1).
Another copper ion containing enzyme is the particulate methane monooxygenase
(pMMO). Most aspects of pMMO biochemistry remained elusive for a long time.
More recent work, however, has revealed the composition of pMMO systems [106,
107]. They were shown to consist of a hydroxylase that consists of three subunits and
an additional component, which is thought to be a methanol dehydrogenase – the
subsequent enzyme in the methane oxidation pathway [108]. The natural electron
donors and electron transfer pathways have not been identified to date.
For all monooxygenases described above, the reaction involves the consumption of a
pair of reducing equivalents – typically a reduced pyridine nucleotide cofactor or
ascorbate – and the generation of a reactive oxygen-containing species. By contrast to
these enzymes, enzymes from the molybdenum hydroxylase family use water as the
ultimate source of the oxygen atom incorporated into the CH bond of the substrate
and generate – rather than consume – reducing equivalents, typically in the form of
NADH, during substrate hydroxylation [109]. Among the members of the molybde-
num hydroxylase family are nicotinate dehydrogenase (NDH) from Eubacterium
barkeri [110, 111], the xanthine oxidoreductases (XDH) from Clostridium purinolyti-
cum [112], Clostridium acidiurici [113] and Eubacterium barkeri [114], and the purine
hydroxylase (PH) from Clostridium purinolyticum [112]. NDH catalyzing the hydrox-
ylation of nicotinate (23) to 6-hydroxynicotinate (24) is important in nicotinate and
nicotinamide metabolism. PH acts on various purines (25) and aldehydes and hydro-
xylates xanthine to uric acid. This enzyme plays an important role in the catabolism of
purines in some species, including humans. XDH catalyzes the oxidation of hypo-
xanthine (26) to xanthine (27) and further to uric acid (28) (Scheme 30.9).
Molybdenum hydroxylases can consist of a variable number of subunits and
contain different additional cofactors, which transfer electrons from a molybdopterin
cofactor to an external electron acceptor. Commonly, they possess a pair of [2Fe–2S]
clusters in two separate domains at the N-terminus and FAD in a third domain [109].
The molybdenum active center is located at the interface of two other elongated
1240
COOH COOH
NDH
N HO N
23 24
O
N N
N HN
N N O N N
H H H
25 27 O
H
N N
PH N XDH XDH HN
O
O N N N
H O N H
H H
26 28
Scheme 30.9 Hydroxylation of nicotinate (23) by nicotinate dehydrogenase (NDH), and oxidation
of purine (25) by purine hydroxylase (PH) and xanthine oxidoreductase (XDH).
domains. Several studies describe in detail mechanisms of molybdenum hydro-

xylases [115–117], which has been characterized best for XDH [118]. Basically, the
reactivity of the molybdenum center in the active site controls the chemical course of
the reaction, which involves base-catalyzed proton abstraction from a Mo-OH group
allowing a nucleophilic attack on the substrate and hydride transfer from the substrate
to a Mo¼S group in the active site. During the course of this reaction, the molybdenum
redox cycle forms MoVI to MoIV, with reoxidation of the MoIV species to generate the
active MoV intermediate. Kinetic investigations provided evidence for a two-electron
reduction mechanism rather than two subsequent one-electron reductions [119].
30.5
Dioxygenases
Dioxygenases catalyzing CH bond hydroxylations can contain either one or more
heme iron units or non-heme iron units. Independent from their type, however, all
dioxygenases have in common that they incorporate both oxygen atoms from
dioxygen into the substrate. Non-heme dioxygenases are further grouped into
intradiol and extradiol dioxygenases. Intradiol dioxygenases utilize a mononuclear
FeIII cofactor bound to a 2-tyrosine-2-histidine motif, whereas extradiol dioxygenases
utilize mononuclear FeII bound to a 2-histidine-1-carboxylate motif [120]. Most CH
bond oxyfunctionalization reactions are catalyzed by Rieske cis-diol dioxygenases that
belong to the non-heme iron extradiol type and are involved in oxidative cleavage of
catechol substrates as part of bacterial aromatic degradation pathways [121]. Other
groups of dioxygenases capable of CH bond oxyfunctionalizations are represented
by FeII/a-keto acid-dependent dioxygenases and lipoxygenases.
30.5 Dioxygenases j1241
30.5.1
Rieske cis-diol Dioxygenases
Rieske cis-diol dioxygenases (RDO) come along with different numbers and types of
protein components and subunits [10]. Typical RDO systems form a soluble electron
transport chain with either two or three separate protein components to harness the
reductive power of NAD(P)H and activate molecular oxygen. This electron transport
chain involves a reductase component for NAD(P)H oxidation and (in three-com-
ponent systems) a ferredoxin, either with a plant type [2Fe–2S] cluster, or with a
Rieske-type [2Fe–2S] cluster that stores and supplies an additional electron for the
reaction. The last component is the RDO enzyme consisting of an a- and sometimes
also a b-subunit. The a-subunit harbors the Rieske [2Fe–2S] cluster domain that
accepts electrons from the reductase or ferredoxin and passes them on to the catalytic
domain [122].
The mechanism of RDO for dioxygen activation and substrate oxidation is still
elusive. It is thought to proceed via a two-electron reduction leading to the hypothesis
that the reactive form of oxygen is a peroxo or hydroperoxo species bound to FeIII. For
the high-valent oxo-complexof RDO,aformal [FeV¼O]species is proposed,where both
atoms of oxygen in the metal coordination sphere form an FeV-oxo-hydroxo moiety
(Scheme 13.1), owing to the fact that both atoms of dioxygen must be retained for
incorporation into the substrate [123, 124]. However, alternative mechanistic hypoth-
eses exist, which suggest that after generation of the formal [FeV¼O] species either an
FeII-(hydro)peroxo or an FeIV-oxo-hydroxo complex is formed, which would presum-
ably be more readily stabilized in a biological system (Scheme 30.10) [125].
RDO fall into four families, in general correlating with the native substrates
oxidized by their members: the toluene/biphenyl family, the naphthalene (29) family,
the benzoate (30) family, and the phthalate family (Scheme 30.11) [126]. The number
of identified initial reaction products of RDO in the bacterial oxidation of aromatic
hydrocarbons is constantly increasing; more than 300 different arene cis-diols have
been described [127, 128].
Interestingly, toluene dioxygenase (TDO) from P. putida UV4 can also carry out
single hydroxylations on sp3 carbon atoms. The enzyme can convert indane and a
series of 2-substituted indane (31) substrates to yield enantiopure cis-indane-1-ols (32)
and cis,trans-indane-1-3-diols (33) (Scheme 30.12) [129]. The high stereoselectivity and
broad substrate specificity shown in this conversion makes this process a valuable one
for the production of enantiopure starting materials for chemical syntheses.
30.5.2
Iron(II)/a-Keto Acid-dependent Dioxygenases
Iron(II)/a-keto acid-dependent dioxygenases (KGDO) transfer one oxygen atom of

O2 to the actual substrate, while a-ketoglutarate (a-KG) acts as cosubstrate for the
second oxygen transfer. The active site of KGDO contains ferrous iron (FeII) bound by
three ligands found in a His-X-Asp/Glu-Xn-His motif. KGDOs catalyze a broad
diversity of reactions that result in protein side-chain modifications, repair of
1242
R R
O=O bond
cleavage O O
HO HO
FeV FeIV
R R R
H+
O O OH
HO R R OH FeIII OH FeIII
FeIII
radical
reaction
O O
OH III OH
Fe FeIV
Scheme 30.10 Possible radical-based mechanisms for cis-dihydroxylation by RDO. Reprinted from Reference [120],
with permission from Elsevier.
30.5 Dioxygenases j1243
H OH
OH
NDO
+ O2 H
+ +
NADH + H NAD
29
HO
HO O
C O C
BDO OH
+ O2
OH
NAD(P)H + H+ NAD(P)+
H
30
Scheme 30.11 RDO-catalyzed oxyfunctionalizations. NDO, naphthalene 1,2-dioxygenase; BDO,

benzoate 1,2-dioxygenase.
OH OH
TDO S TDO S
R R R R
O2 O2 S
31 32 33 OH
Scheme 30.12 Enantioselective hydroxylation of 2-substituted indanes (31) by toluene

dioxygenase (TDO).
alkylated DNA or RNA, biosynthesis of antibiotics and plant products, and biodeg-
radation of various compounds [130]. An FeIV-oxo species is predicted to be the key
intermediate for hydroxylations [131]. a-KG (34) chelates FeII, while it is oxidatively
decarboxylated to form CO2, succinate (35), and the activated FeIV-oxo species [130].
Some enzymes in the diverse group of KGDO catalyze monohydroxylation reac-
tions, for example, the specific hydroxylation of proline [132], lysine [133], or isoleu-
cine [134, 135]. Hydroxylated forms of proline (36) play an important role in collagen
synthesis and as precursor in the synthesis of antibiotics in various Streptomyces
strains. The proline 4-hydroxylase from Streptomyces griseoviridus, for example, pro-
duces trans-4-hydroxyproline (37) that is subsequently incorporated into the antibiotic
entamycin [136]. Similarly, proline 3-hydroxylase found in two Streptomyces strains
produces cis-3-hydroxyproline (38), a precursor of the peptide antibiotic telomycin
(Scheme 30.13) [137]. Lysyl hydroxylase is a homodimer catalyzing the formation of
4-hydroxylysine in collagen synthesis [133]. Two novel KGDOs from Bacillus thur-
ingiensis strain 2e2 AKU 0251 and the closely related strain B. thuringiensis ATCC 35646
were found that catalyze the hydroxylation of L-isoleucine (39) to 4-hydroxyisoleucine.
Both enzymes shows high stereoselectivity producing only (2S,3R,4S)-4-hydroxyiso-
leucine (40) out of eight possible diastereomers (Scheme 30.13) [134, 135]. 4-Hydro-
xyisoleucine enhances glucose-induced insulin secretion through a direct effect on
1244
COOH
HN
DO
COOH
G
pK
HN OH
3-
+ O2 38
4-
COOH
pK
36 COOH
G
HN
DO
C O COOH
CH2 CH2
+ CO2 OH
CH2 CH2
37
COOH COOH
34 35
O O
(S) + O2 (S) (S)

HO (S) HO (R)
4-iKGDO
NH2 NH2 OH
39 40
Scheme 30.13 Three examples for regio- and and the hydroxylation of isoleucine (39) by
stereoselective hydroxylations of amino acids isoleucine-4-hydroxylase (4-iKGDO). The
catalyzed by FeII/a-keto acid-dependent second oxygen atom is transferred to
dioxygenases (KGDOs): hydroxylation of a-ketoglutarate (34), yielding succinate (35)
proline (36) by proline 3-hydroxylase (3- and CO2.
pKGDO) or proline 4-hydroxylase (4-pKGDO),
pancreatic b cells in rats and humans and therefore is expected to be a novel orally active
drug for insulin-independent diabetes [138].
30.5.3
Lipoxygenases
Lipoxygenases (LOX) are non-heme iron containing fatty acid dioxygenases found
widely in higher eukaryotes like plants, fungi, and animals, but not in prokaryotes
and lower eukaryotes. They are notable because they are cofactor independent and
incorporate both oxygen atoms from O2 into the (1Z,4Z)-pentadienyl system of
polyunsaturated fatty acids to generate optically active hydroperoxy derivatives that
can either be an intermediate or an end product in the metabolic pathway of the
respective organism [139, 140].
LOX consist of a single polypeptide chain that is folded into two domains – an
a-helical catalytic domain, and a N-terminal b-barrel domain that is involved in
membrane binding. The catalytic site of LOX is of the non-heme iron type. The metal
ion is liganded to conserved histidines and the carboxyl group of a conserved
isoleucine. The conserved b-barrel domain shares significant homology to a similar
domain in mammalian lipases, giving a potential clue to mechanisms involved in
30.6 Oxyfunctionalization of CH Bonds for Production of Fine Chemicals j1245
substrate acquisition [141]. However, the understanding of the mechanisms of
substrate acquisition, substrate entry into the catalytic domain, and achievement
of positional and stereo-control is far from clear and still clouded by many
uncertainties [142].
As mentioned, LOX utilize unsaturated fatty acids with (1Z,4Z)-pentadiene – such
as arachidonic, linoleic (41), or linolenic acid – and synthesize an array of chiral
hydroperoxy derivatives from those substrates (Scheme 30.14). An individual
enzyme, however, inserts molecular oxygen on a single position on the carbon chain
and in a single stereo-configuration regulated by the orientation and depth of
substrate entry into the active site [143]. Some plant LOX-pathway products with
potential biotechnological applications are involved in the resistance to environmen-
tal stress and the defense against microbe and herbivore attack, whereas volatile
products, like jasmonic acid and short-chain aldehydes, have a function in plant–
plant communication [144].
OOH
O
C
OH
O2 13R-LOX
O
C
41
OH
O2
9S-LOX
O
C
OOH OH
Scheme 30.14 Two examples of regio- and stereoselective oxidations of linoleic acid (41) catalyzed
by LOX.
30.6
Oxyfunctionalization of CH Bonds for Production of Fine Chemicals
30.6.1
General Aspects
In this chapter we review CH bond oxyfunctionalizations of different types of

substrates. Of course, neither can all aspects of this rapidly developing field be
1246
considered nor can all recent publications be cited. One should keep in mind that in
2010 on P450s alone more than 280 review articles, as well as 2100 original papers
or monographs in books, were published according to a literature search in the ISI
Web of Science database. Therefore, we will focus on basic aspects and selected
substrates that represent interesting targets for biotechnological fine chemical
production.
Generally, technical applications of oxygenases face certain challenges: The most
important property that limits industrial applications of cytochrome P450 mono-
oxygenases, methane monooxygenases, and Rieske cis-diol dioxygenases is the fact
that they require the costly cofactor NAD(P)H and additional electron-transfer
proteins. Furthermore, some of these enzymes are membrane-bound proteins
and their efficient recombinant expression is still a challenge. Another important
factor for efficient biocatalysis is the coupling efficiency – the yield of product based
on NAD(P)H consumed. Besides reducing the efficiency of cofactor usage, uncou-
pling between NAD(P)H oxidation and product formation results in reactive oxygen
species (e.g., superoxide anions and hydrogen peroxide) that cause oxidative damage
of the oxygenase. Owing to these hurdles, industrial applications of oxygenases have so
far been restricted to whole-cell systems. In such instances, however, physiological
effects like limited substrate uptake, toxicity of substrate or product, product degra-
dation, and elaborate downstream processing must also be taken into account [145].
Many attempts – especially for P450s – have been made to overcome the hurdle of
cofactor dependency including direct chemical or electrochemical reduction of the
oxygenase, the use of inexpensive chemicals to directly replace NAD(P)H, and the
development of (enzymatic) cofactor recycling systems for in vitro or in vivo applica-
tions [23, 146–149].
30.6.2
Oxidation of Fatty Acids
Oxygenated fatty acids (and derivatives thereof) have multiple functions in diverse
forms of life and therefore are interesting target molecules for biotechnological
processing. In mammals oxygenated forms of arachidonic acid generated by LOX,
and their derivatives are involved in blood pressure regulation, and they are used as
signaling molecules in stress response to infection, allergy, and exposure to food,
drug, and environmental harmful substances [150]. The allylic hydroxylated deriva-
tives of linoleic acid and conjugated linoleic acids have been demonstrated to exhibit
in vitro cytotoxicity against a panel of human cancer cell lines [151]. Investigation of
natural sources and functions of hydroxylated fatty acids provides important infor-
mation for (i) detection of particularly useful oxygenating enzymes and (ii) suggest-
ing possible biotechnological applications in a technical context.
Numerous P450s – predominantly members of the CYP2, CYP4, CYP52, CYP505,
CYP102, and CYP152 family – use fatty acids and their derivatives as substrates. They
can be subdivided into terminal and subterminal fatty acid hydroxylases. The
mammalian CYP2 enzymes, bacterial CYP102A and CYP152, and fungal CYP505
enzymes belong to subterminal fatty acid hydroxylases. The group of terminal fatty
acid hydroxylases is much smaller and is represented by the mammalian CYP4- and
yeast CYP52 families.
Human CYP4 (e.g., CYP4F2, CYP4F3A, CYP4F3B) and CYP2 monooxygenases
(e.g., CYP2C19, CYP2E1) catalyze the hydroxylation of polyunsaturated fatty acids,
for example, eicosatrienoic acid, arachidonic acid, eicosapentaenoic acid, or doc-
osahexaenoic acid [152]. However, low expression levels and activity of these enzymes
do not correspond to the requirements of industrial processes.
The most intensely studied enzymes concerning fatty acid oxidation are the P450s
of the CYP102A family. CYP102A monooxygenases are self-sufficient flavocyto-
chromes that consist of a heme domain fused to a FMN- and FAD-containing
reductase domain. Therefore, their experimental setup in organic synthesis is a lot
easier compared to other P450s, which require one or two additional electron-
transport proteins for activity [153]. Fatty acid oxidation has been reported for
CYP102A1 (P450 BM3) from Bacillus megaterium, two P450s from Bacillus subtilis
(CYP102A2 and CYP102A3), one from Bacillus cereus ATCC 14579 (CYP102A5), and
one from Bacillus licheniformis ATCC 14580 (CYP102A7) [154, 155]. These enzymes
have commonly shown very high oxygenation rates with fatty acids of around
4600 nmol substrate (nmol P450)1 min1. Isolated reaction rates of even >15 000
min1 have been seen for oxidation of arachidonic acid catalyzed by P450 BM3 [156].
In contrast, typical rates of fatty acid oxidations by mammalian monooxygenase lie
around 1 min1 [157]. Many reviews on P450 BM3 have been written in the last
decade, where interested readers will find more details [156, 158–160].
Many studies aiming to engineer CYP102A mutants with altered regio- and
stereoselectivities and/or altered substrate specificities have also been undertaken:
Wild-type P450BM3 oxidizes saturated fatty acids at subterminal positions, producing
a mixture of v1, v2, and v3 hydroxylated products. Replacing F87 has profound
effects on regioselectivity of P450BM3 [158]. By combination with other mutations
located in the substrate binding pocket, triple mutants were constructed that oxidize
lauric acid at d-, c, and b-positions. Both d- and c-hydroxy lauric acid are valuable
synthons for the production of lactones, which are important commercial flavors with
a typical peachy odor [161].
Highly branched fatty acids and their derivatives are promising chiral precursors
for the synthesis of macrolide antibiotics. The key step in the utilization of these
compounds is their regioselective hydroxylation, which cannot be achieved in a
classical chemical approach. CYP102A2 and CYP102A3 do not show activity against
these substrates. However, P450 BM3 and its A74G/F87V/L188Q triple mutant
hydroxylate a variety of these compounds with high regioselectivity [162].
Another class of P450s capable of fatty acid hydroxylation is represented by the
H2O2-utilizing peroxygenases of the CYP152 family that employ the peroxide shunt
for catalysis exclusively and therefore do not require exogenous protein partners.
Three enzymes with a potential for biocatalytic applications are CYP152B1 (SPa)
from Sphingomonas paucimobilis [163], CYP152A1 (P450Bsb) from B. subtilis [164], and
CYP152A2 (P450CLA) from Clostridium acetobutylicum [165]. The products of this
reaction (a- and b-hydroxy fatty acids) can be utilized as precursors for synthesis of
antibiotic compounds like surfactin [166, 167].
1248
Candida strains have been widely used for the production of a,v-dicarboxylic acids
starting with long-chain fatty acids (>C12). a,v-Dicarboxylic acids are important
intermediates for the synthesis of polyesters, polyamides, or adhesives [168]. Mutated
Candida strains engineered for higher productivity were reported that produce up to
300 g l1 of dicarboxylic acids [169]. Among them the process with Candida tropicalis
was commercialized [170]. Microsomal CYP52 enzymes responsible for the first step
of this process – the terminal hydroxylation of fatty acids – obtain electrons from
NADPH via a cytochrome P450 reductase. Some CYP52 enzymes have been isolated,
expressed in recombinant hosts, and characterized [171].
Apart from yeasts several processes utilizing bacteria like Pseudomonas, Bacilli,
Rhodococci, or recombinant Escherichia coli (transformed with P450 BM3 and a
suitable fatty acid uptake system) have been studied to some extent [172–175], but
do not yet allow to produce oxyfunctionalized fatty acids at an economic price for
commercial exploitation.
Bioreactors using isolated enzymes are limited to lipoxygenases up to now, since
they are not cofactor dependent. Owing to the cofactor dependency of P450s their
application as isolated enzymes in bioreactors has so far been an issue of interest to
academia only (Section 30.6.1). In conclusion biocatalytic hydroxylation of fatty acids
is still in its infancy and large-scale applications in the near future seem to be possible
in very few cases only.
30.6.3
Oxidation of Alkanes
Depending on their chain length, different enzyme systems are involved in the first
oxidation step of alkanes: (i) methane to butane (C1–C4) are oxidized by methane
monooxygenase-like enzymes, (ii) pentane to hexadecane (C5–C16) can be oxidized
either by integral membrane non-heme iron monooxygenases or by P450s, and (iii)
longer alkanes of >C17 are accepted by a broad range of enzymes including P450s,
flavin-containing oxygenases, dioxygenases, and others [88].
Although chemical catalysts have been developed that can convert methane into
methanol in good yield (even using dioxygen as oxidant), most of them still suffer
serious drawbacks such as the requirement of high pressures and temperatures [176].
In contrast, nature provides us with efficient biocatalysts, MMO (described in
Section 30.4.1), which operate in neutral aqueous solution at moderate temperatures
and atmospheric pressure [177]. However, though having broad substrate spectra,
sMMO display quite low regio- and stereoselectivity, limiting their biotechnological
applications for synthetic purposes. Furthermore, their complex multidomain orga-
nization restricts their use to whole cell biocatalysis, for instance with Methylosinus
trichosporium OB3b [80]. By inhibiting methanol dehydrogenase in M. trichosporium
with sodium chloride, methanol production from methane was enhanced and a
methanol concentration of 7 mM was attained in a 36 h batch reaction [178].
Several microorganisms have been isolated for their ability to use gaseous n-
alkanes from ethane to butane as sole carbon source. Some of these bacteria are also
known to degrade various environmental pollutants (trichloroethylene, chloroform,
methyl ethers). Thus, from a biotechnological perspective, the enzymes participat-
ing in these oxidation pathways promise to be versatile biocatalysts. A novel soluble
butane monooxygenase from the C2–C9 alkane-utilizing bacterium Thauera buta-
nivorans has been characterized [179–181]. The enzyme has high sequence similarity
to the sMMO from M. trichosporium OB3b and exhibits a similar substrate range,
including gaseous and liquid C2–C5 alkanes, aromatics, alkenes, and halogenated
xenobiotics. Another example is the Gordonia sp. strain TY-5, which can grow in
propane as sole carbon source and produces 2-propanol [182]. The complete operon
encoding the putative diiron containing multicomponent monooxygenase, a
NADH-dependent reductase, and a regulatory protein was cloned. The hydroxylase
domain of this monooxygenase shows homology to the known sMMO and to
the butane monooxygenase from T. butanivorans, but accepts only propane as
substrate [182].
The enzymes introducing oxygen in alkane substrates with chain length C5–C16
belong to two distinctive groups: integral membrane non-heme iron monooxy-
genases (briefly described in Section 30.4.1) and cytochrome P450 monooxygenases
(Section 30.3.1). The first non-heme iron protein AlkB was discovered in P. putida
GPo1 (former P. oleovorans), which can grow on alkanes ranging from hexane to
dodecane and catalyze their terminal hydroxylation [183]. This alkane hydroxylase
system has been studied in detail. It consists of (i) a rubredoxin reductase (AlkT)
[184, 185], which transfers electrons from NADH to rubredoxin [186], (ii) rubredoxin
(AlkG), which is an iron-sulfur electron transfer protein, and (iii) AlkB [187], a
monooxygenase that catalyzes the oxidation of alkanes. The three alk genes are
clustered on the OCTplasmid in P. putida GPo1 [188]. AlkB is an integral-membrane
non-heme diiron monooxygenase [189], which requires phospholipids for catalytic
activity [190]. It has a very broad substrate spectrum and oxidizes C3–C12
alkanes [191]. It was also reported to oxidize N-benzylpyrrolidine for the preparation
of optically active N-benzyl-3-hydroxypyrrolidine [192]. Later, several similar systems
were identified in a wide range of bacteria from a-, b-, and c-proteobacteria and
Actinomycetales [193]. The genetic organization of the genes responsible for alkane
hydroxylation in, for example, Acinetobacter sp strain ADP1 is completely different
since they are not clustered or localized on a plasmid [194]. Interestingly, most of the
identified enzymes prefer alkane substrates longer than C10.
P450s belonging to the CYP153 family are found to catalyze terminal hydroxylation
of C5–C12 alkanes. These monooxygenase systems consist of a cytoplasmatic P450
enzyme, a [2Fe–2S] ferredoxin, and a ferredoxin reductase; however, corresponding
electron transfer partners could not be identified for all CYP153. The first member of
this family was isolated from Acinetobacter sp. EB104 and was shown to hydroxylate
hexadecane [195]. Later, biotransformation of octane using recombinant E. coli
expressing P450balk from Alcanivorax borkumensis SK2 was reported
[196, 197]. CYP153A from Acinetobacter OC4 was successfully co-expressed in E.
coli with its natural redox partners. In vivo oxidation of octane with the recombinant
cells produced 2250 mg l1 1-octanol and 722 mg l1 a,v-octandiol within 24 h [198].
A screening revealed 35 strains that possess CYP153 homologs, several of which
could be functionally expressed in P. putida [199]. Remarkably, some enzymes from
1250
the CYP153 family demonstrate a broad substrate spectrum and accept inter alia
limonene and four-, five-, and six-ring alicyclic compounds [200–203].
Some microorganisms are able to grow on alkanes longer than C16, for example,
Rhodococcus isolates that were shown to grow on pure alkanes up to C32 [204],
Pseudomonas fluorescens on C18–C28 alkanes [205], and Geobacillus thermodenitrificans
NG80-2 on C15–C36 alkanes [206]. In G. thermodenitrificans the key enzyme respon-
sible for the first oxidation is the long-chain alkane monooxygenase LadA that
converts alkanes of C15–C36 into the corresponding primary alcohols, but does
not accept C6–C14 alkanes [207]. Remarkably, LadA utilizes FAD for O2-activation and
is – in contrast to all other previously described oxygenases – an extracellular
enzyme [208].
Microsomal P450s from the CYP52-family – mostly found in various yeast strains –
with activity towards fatty acids (Section 30.6.2) are able to hydroxylate long-chain
alkanes as well [193]: Ten CYP52-genes were cloned from Candida tropicalis ATCC
20336, which oxidizes n-alkanes at the a- and v-positions to yield fatty acids and
dicarboxylic acids [171].
Remarkably, no efficient natural ethane oxidizing system has been identified yet.
Therefore, engineering of P450s to alkane hydroxylases is a topic of ongoing interest,
for example utilizing P450cam from P. putida (CYP101A1). The physiological
substrate of P450cam is ( þ )-camphor, which is hydroxylated regio- and enantiose-
lective to 6-hydroxy-champhor [209]. A step-by-step adaptation of the enzyme to
smaller n-alkanes beginning with hexane [210], then to butane and propane [211], and
finally to ethane [212] was undertaken. The best mutant with eight substitutions
oxidized propane at a rate of 500 min1 with 86% coupling, which was comparable
with that of the wild-type enzyme towards ( þ )-camphor, the natural substrate of
P450cam [212].
The activity and selectivity of P450 BM3 was altered by Arnold and coworkers from
hydroxylation of dodecane (C12) first to octane (C8) and hexane (C6) and further on to
gaseous propane (C3) [213] and ethane (C2) [214–216]. Some mutants were found
with moderate stereoselectivity, which lead either to the (R)- or the (S)-2-octanol
enantiomer products of alkane hydroxylation [217]. CYP102A3, which oxidizes n-
octane at subterminal positions with low activity, has also been engineered for
terminal hydroxylation of this substrate. The best mutant S189Q/A330V produced
48% 1-octanol [218].
30.6.4
Oxidation of Terpenes and Terpenoids
Oxidation of low value terpenes to higher value derivatives has been recognized for
some time as an attractive opportunity for synthetic chemistry [219, 220]. Terpenes
have the general formula (C5H8)n and are biosynthesized from isoprene units in the
form of isopentenyl pyrophosphate. The parent monoterpene hydrocarbons are often
readily available and their oxygenation gives derivatives that are sought-after fra-
grances and flavorings, pharmaceuticals, or building blocks for chemical synthesis.
The regiospecific introduction of carbonyl or hydroxyl groups in terpenes by chemical
means has proved difficult due to similar electronic properties of the primary and
secondary allylic positions. Moreover, the allylic hydroxylation often competes with
epoxidation of the corresponding C¼C double bond. As a consequence, classical
chemical oxidation procedures often lead to mixtures of different products. For
obvious reasons biocatalytic conversion of terpenes was considered as early as the
1960s [221]. Many terpene hydrocarbons are abundant in nature, for example,
limonene and pinene [222]. Owing to their chemical instability and poor sensory
impact, they are not qualified as flavorings, but represent ideal starting materials for
biocatalytic oxyfunctionalizations [223].
30.6.4.1 Monocyclic Monoterpenes: Limonene

The most notable oxygenated derivatives of (S)-()-limonene (L-limonene 42) and
(R)-( þ )-limonene (D-limonene 43) are perillyl alcohol (44), isopiperitenol (45),
carveol (46), carvone (47), and a-terpineol (48) (Scheme 30.15). Oxyfunctionalization
reactions on terpenes are usually attributed to P450s.
Two closely related enzymes, limonene-3-hydroxylase (L3H, CYP71D13) and
limonene-6-hydroxylase (L6H, CYP71D18), were identified and isolated from
Mentha species [224]. In the case of L3H, both enantiomers of limonene are
oxygenated with strict regio- and stereochemistry, leading to trans-isopiperite-
nol [225]. In the case of the L6H, L-limonene (42) is converted into ()-trans-carveol
as the only product. With D-limonene (43) as substrate, multiple products are
generated, with ( þ )-cis-carveol predominating [226].
A microbial screening approach with limonene revealed a vast variety of micro-
organisms with hydroxylating activity. Remarkably, trans-isopiperitenol – which is
OH
42 43
OH
44 OH O 48
HO
45 46 47
Scheme 30.15 Main oxygenated derivatives of L-limonene (42) and D-limonene (43).
1252
produced by plant P450s – has never been reported to be formed by bacterial strains,
but was the sole biotransformation product of the black yeast Hormonema sp. UOFS
Y-0067 [227]. Another fully regiospecific hydroxylation was found to be carried out by
the basidiomycete Pleurotus sapidus, which converts D-limonene (43) into cis-carveol,
trans-carveol, and carvone [228]. The latter biotransformation can also be catalyzed by
bacterial strains, for example, by Rhodococcus opacus PWD4, which converts
D-limonene (43) into trans-carveol by the action of regioselective toluene and/or
naphthalene dioxygenases [229] making it a promising candidate for industrial
applications [230]. Other microbial oxidations of limonene catalyzed by fungi or
bacteria include hydroxylations at the 1,2-position to limonene-1,2-diol [231, 232], at
the 8-position to a-terpineol [233–235], or at the 7-position yielding ()-perillyl
alcohol [236]. Perillyl alcohol is an anticancer drug and its extraction from plants is
expensive and insufficient. In an attempt to establish an efficient alternative, its
biocatalytic production was performed with the Mycobacterium sp. P450 alkane
hydroxylase from the CYP153 family, recombinantly expressed in P. putida cells [237].
The whole-cell process was performed in a two-phase system, resulting in 6.8 g l1
()-perillyl alcohol in the organic phase.
Several attempts have been undertaken to improve activity and specificity of
P450cam and P450 BM3 towards limonene. The F87W/Y96F/V247L mutant of
P450cam showed high regioselectivity (90%) for isopiperitenol formation [238].
The P450 BM3 wild-type converts D-limonene (43) into four different products:
racemic mixtures of limonene-1,2-epoxide (30%), limonene-8,9-epoxide (7%), iso-
piperitenol (54%), and carveol (9%). The two mutants F87A/A328F and F87V/A328F
epoxidize the C8 ¼ C9 double bond almost exclusively resulting in 94% and 97%
limonene-8,9-epoxide but no hydroxylated products [239].
30.6.4.2 Dicyclic Monoterpenes: Pinene

Other substrates of interest are the bicyclic monoterpenes ()-a-pinene (49) and
( þ )-a-pinene (50) – waste products from pulp processing. The primary products of
P450-catalyzed oxidations of a-pinene are verbenol (51), which is an active phero-
mone against various beetle species, and verbenone (52) (the product of the further
oxidation of verbenol), a sought-after compound of rosemary oil (Figure 30.1). A
challenge of selective enzymatic pinene oxidation is related to the simultaneous
autoxidation of this compound. Reports on the stability of pinene oxide in buffered
aqueous solutions at room temperature describe its rapid decomposition along with
the formation of cis-carveol and sobrerol [240]. Verbenone (52) and its precursor
OH O
49 50 51 52
Figure 30.1 Structures of ()-a-pinene (49), ( þ )-a-pinene (50), and two oxygenated derivatives
of ()-a-pinene: verbenol (51) and verbenone (52).
verbenol (51) have frequently been described as the main products of biotransforma-
tions of a-pinene with bacteria, yeasts, and fungi [223, 241]. Pleurotus sapidus was
found to selectively oxidize a-pinene to (E)-verbenol and verbenone at a ratio of 1 : 1.
A regiospecific a-pinene dioxygenase and a stereoselective (Z)-verbenol dehydro-
genase were proposed to be responsible for the microbial formation of verbenone.
The oxidation occurred via formation of (Z)- and (E)-verbenyl hydroperoxides [242].
Remarkably, no P450 wild-type enzymes have been identified so far that are
involved in pinene metabolism in plants and microorganisms, but several mutants of
P450cam and P450 BM3 have been constructed, which accept a-pinene. For example,
the F87W/Y96F/L244A mutant of P450cam gave 86% ( þ )-cis-verbenol and 5%
( þ )-verbenone and the Y96F/L244A/V247L mutant produced 55% ( þ )-cis-verbenol
and 32% ( þ )-verbenone [243]. Wild-type P450 BM3 shows no activity towards
()-a-pinene, but the triple mutant A74G/F87G/L188Q produces up to 77%
()-cis-verbenol [244].
30.6.4.3 Sesquiterpenoides: Valencene

Regioselective allylic hydroxylation of ( þ )-valencene (53) yields cis- (54) and/or trans-
nootkatol (55), which can further be oxidized to ( þ )-nootkatone (56) (Scheme 30.16).
( þ )-Valencene (53) is found in citrus oils and can be inexpensively extracted from
oranges. ( þ )-Nootkatone (56) is a high added-value commercial flavoring with a low
odor threshold of 1 mg l1 in water and with a broad application in the food,
cosmetics, and pharmaceutical industry. The products of the primary ( þ )-valencene
oxidation, cis- and trans-nootkatol, are useful flavoring compounds and might be used
in combination with ( þ )-nootkatone [245]. Traditionally, ( þ )-nootkatone is extracted
from grapefruits and its price and availability is dependent on the annual harvest,
which is restricted to a narrow producing area and very sensitive to weather conditions.
HO HO O
53 54 55 56
Scheme 30.16 Allylic oxidation of ( þ )-valencene (53).
Several biotechnological processes to achieve this oxidation have been designed.

Recent attempts include the manufacturing of ( þ )-nootkatone with green algae like
Chlorella or Euglena [246], or with fungi such as Aspergillus niger, Fusarium cul-
morum [247], Mucor sp. [248, 249], or the ascomycete Chaetomium globosum [250].
Cell-free enzymatic reactions for the conversion of ( þ )-valencene exploiting
enzymes from Cichorium intybus L. roots [251], lignin peroxidase [252], and fungal
laccase [253] have also been reported.
Recently, an efficient allylic oxidation of ( þ )-valencene to ( þ )-nootkatone with the
lyophilized mycelium of the basidiomycete Pleurotus sapidus has been reported; cis-
and trans-nootkatol represented the only side products. After 24 h biotransformation
1254
up to 320 mg l1 ( þ )-nootkatone was produced. The responsible enzyme was isolated
and characterized. The amino acid sequence demonstrated 50% similarity to a
putative lipoxygenases from Aspergillus fumigatus and Laccaria bicolor, as well as
26% similarity to the sequence of lipoxygenase-1 from soy bean [254]. Analysis of
the intermediates has revealed that the allylic oxidation to nootkatol and nootkatone
proceeds via the respective hydroperoxides [255]. The same mechanism was described
for oxidation of a-pinene by fungal dioxygenase (Section 30.6.4.2) [242]. Both
examples demonstrate that fungal oxyfunctionalization reactions of some common
terpene substrates might be catalyzed by dioxygenases rather than by P450s.
A new P450 from B. subtilis 168 – CYP109B1 – has been described, which
catalyzes regioselective allylic hydroxylation of ( þ )-valencene to nootkatol and
further to ( þ )-nootkatone. Recombinant E. coli cells expressing CYP109B1,
together with putidaredoxin and putidaredoxin reductase from P. putida, were
used for biotransformation. In a biphasic system with addition of nonpolar organic
solvents up to 120 mg l1 product was achieved [256].
Other reports describe the hydroxylation of ( þ )-valencene by mutants of P450cam
and P450 BM3 [257]. Conversion of ( þ )-valencene (9%) was achieved with the F87V/
Y96F/L244A triple mutant of P450cam. Trans-Nootkatol (38%) and ( þ )-nootkatone
(47%) represented the major transformation products [257]. Most P450 BM3
mutants had low chemo- and regioselectivity and produced up to six byproducts
besides nootkatol and ( þ )-nootkatone [257]. Later, a minimal P450 BM3 mutant
library of only 24 variants was constructed by combining five hydrophobic amino
acids (alanine, valine, phenylalanine, leucine, and isoleucine) at positions 87 and 328.
The best variant F87A/A328I revealed 94% preference for oxidation at the secondary
allylic position and produced 26% ( þ )-nootkatone [239].
30.6.4.4 Sesquiterpenoid Analogs: Ionone

Ionones and hydroxy-ionones are essential intermediates in the synthesis of several
carotenoids [258, 259] and can be used in the industrial synthesis of, for example-
zeaxanthin and cantaxanthin. Ionone derivatives constitute substantial aroma com-
ponents of floral scents [260, 261]. Furthermore, hydroxy-b-ionone is a versatile
compound in the synthesis of the phytohormone abscisic acid. All these useful
features make ionones and their hydroxy metabolites attractive for the fragrance and
flavor industry.
Aspergillus niger was found to convert b-ionone into two major products –
4-hydroxy-b-ionone and 2-hydroxy-b-ionone [262]. This bioconversion was carried
out using immobilized cells. Thus, this fungus could be used repeatedly for microbial
conversion of b-ionone in the presence of isooctane for more than 480 h [262].
Aspergillus awamori has also been shown to achieve hydroxylation of b-ionone in the
same manner [263]. The process has been explored to obtain a mixture of derivatives
that is utilized as an essential oil for tobacco flavoring.
Several Streptomyces strains were shown to hydroxylate a-ionone and b-ionone
regio- and stereoselectively [264]. Hydroxylation of racemic a-ionone (57) [(6R)-
()/(6S)-( þ )] resulted in the exclusive formation of the two enantiomers (3S,6S)-
hydroxy-a-ionone (58) and (3R,6R)-hydroxy-a-ionone (59) out of four possible
O O O
1
2 6 CYP105B1 (S) (R)
3 5
+
(S) (R)
4
HO HO
57 58 59
O O
1
2 6 CYP105B2
3 5
4 (S)
60 OH 61
Scheme 30.17 Regio- and stereoselective hydroxylation of a- (57) and b-ionone (60) by CYP105B
enzymes.
(Scheme 30.17). Later CYP105A1 and CYP105B1 from S. griseolus, as well as

CYP105D1 from S. griseus, were recombinantly expressed in E. coli together with
their own ferredoxins [265]. All three enzymes preferred the secondary allylic
position for hydroxylation. The best yield was obtained using CYP105B1 with
20% conversion of a-ionone after 48 h and 57% 3-hydroxy-a-ionone. b-Ionone
(60) biotransformation reached 40% conversion with the same recombinant system
with almost complete regioselectivity and moderate enantioselectivity, giving (4S)-
hydroxy-b-ionone (61) with 35% e.e. (Scheme 30.17).
In the area of protein engineering P450 BM3 was optimized for the selective
oxidation of ionones. The wild-type enzyme displayed only very low activity towards
b-ionone [266]. The triple mutant A74E/F87V/P386S, which was 300 times more
active than the wild-type produced exclusively 4-hydroxy-b-ionone with moderate
enantioselectivity for the (R)-enantiomer (up to 39% e.e.) [266].
Two new P450s from the CYP109 family – namely CYP109B1 from B. subtilis [267]
and CYP109D1 from Sorangium cellulosum So ce56 [268] – have been described that
are capable of a- and b-ionone oxidation. Both enzymes show 100% regioselectivity
for oxidation at allylic C-atoms, yielding exclusively 3-hydroxy-a-ionone and
4-hydroxy-b-ionone. However, the optical purity of the products was not determined
in this case.
30.6.5
Oxidation of Steroids
Steroid transformation by human P450s has been studied in detail [269]. For
example, the final steps in the synthesis of the major human glucocorticoid cortisol
and the most important mineralocorticoid aldosterone [270] are catalyzed by two
mitochondrial P450 isozymes, CYP11B1 and CYP11B2 [271]. Cortisol is synthesized
from 11-deoxycortisol through a hydroxylation reaction at position 11b catalyzed by
CYP11B1, whereas aldosterone is synthesized from 11-deoxycorticosterone through
1256
a series of reactions catalyzed by CYP11B2. The reduction equivalents are provided

via a [2Fe–2S]-containing adrenodoxin (Adx) and a NADPH-dependent FAD-
containing adrenodoxin reductase (AdR) [272, 273]. However, despite the high
potential of both enzymes they have so far not been employed in biotechnology due
to low expression yields and instability of the proteins. Successful co-expression of
AdR, Adx, and CYP11B1 in an auxotrophic strain of the fission yeast Schizosac-
charomyces pombe has been reported. Furthermore, through the introduction of a
single mutation in CYP11B1 the 11b-hydroxylase activity of the system could be
increased 3.4-fold leading to an average production of 1 mM hydrocortisone over a
period of 72 h [274].
Regio- and enantioselective hydroxylation of steroid hormones provide pharma-
ceuticals [275, 276] that have a broad therapeutic application in contraceptives, anti-
inflammatory, immunosuppressive, anabolic, and diuretic drugs. The complex
structure of steroid molecules requires complicated, multistep schemes for the
chemical synthesis of respective steroid compounds [275]. The preparation of
intermediates with protection groups and their subsequent removal is often neces-
sary making the chemical synthesis expensive and time consuming. Furthermore,
the basic ring structure of some steroid derivatives is sensitive to cleavage by a wide
variety of chemicals. Currently, chemical and biochemical processes are often
combined in the production of steroids. An adequate alternative to human P450s
in this case is represented by microbial systems, particularly by P450s from fungi.
During the last 30 years, microbial transformation using whole-cells has been used to
introduce different functional groups on the steroidal skeleton. These reactions
usually occur with high regio- and stereoselectivity. Application of whole-cell systems
allows to avoid the costs of enzyme isolation, purification, stabilization, and to solve
the problem of the NAD(P)H cofactor regeneration. Among very well-established
large-scale commercial applications is the 11b-hydroxylation of 11-deoxycortisol (62)
to hydrocortisone (63) using P450 of Curvularia sp. [277] applied by the Schering AG
(in 2006 acquired by Merck KgaA, Germany) at an industrial scale of approximately
100 tons per year (Scheme 30.18) [145]. Another example is the 11a-hydroxylation of
progesterone (64) by Rhizopus arrhizus to 11a-hydroxyprogesterone (65) developed in
the 1950s by Pharmacia & Upjohn (later acquired by Pfizer Inc, USA)
(Scheme 30.18) [278, 279]. Further examples cover 15a- and 16a-hydroxylated
products [280]. All processes are one-step biotransformations, which cannot be
achieved by chemical routes. The mentioned compounds are mainly used for the
production of adrenal cortex hormones and their analogues [275, 276, 280].
Although many bioconversions are known, there are still ongoing efforts to
increase their efficiency and to find new useful microorganisms. The application
of testosterone – the main male hormone – as substrate for various microorganisms
leads to many hydroxylation reactions, including 2b by Whetzelinia sclerotiorum [281],
6b by Fusarium culmorum [282], 7a by Botrytis cinerea [283], 7b by Phycomyces
blakesleeanus [284], 11a by Beauveria bassiana [285], 14a by Absidia coerulea [286] or
Thamnostylum piriforme [287], and 15b by Aspergillus fumigatus [288]. Another
common reaction was oxidation of 17b-OH to the corresponding ketone, which is
usually accompanied by hydroxylation [286, 287, 289]. Reports of biotransformation
OH OH
OH HO OH
H H
Curvularia sp.
H H H H
O O
62 63
O O
H HO H
H H
Rhizopus arrhizus
H H H H
O O
64 65
Scheme 30.18 11b-Hydroxylation of 11-deoxycortisol (62), and 11a-hydroxylation of progesterone

(64).
of testosterone using Fusarium oxysporum and F. culmorum confirmed the presence

of 15a-hydroxylase and 6b-hydroxylase in these species [282, 290]. Several reports
describe biotransformations using Rhizopus species to prepare 11a- and
6b-hydroxysteroids [291].
While steroid transformations by fungal P450s is well established, only a few
studies led to identification of bacterial P450 genes involved in steroid hydroxyla-
tions. However, bacterial P450s are more appropriate for biotechnological imple-
mentation, since they are soluble enzymes interacting with soluble electron-transfer
proteins. Moreover, they demonstrate higher activity and stability than their eukary-
otic counterparts and can easily be expressed in heterologous hosts. The search for
bacterial equivalents for steroid hydroxylating eukaryotic P450s has therefore
attracted attention of many research groups and pharmaceutical companies.
Bacillus stearothermophilus has been found to produce 6a-hydroxytestosterone and
6b-hydroxytestosterone, though as minor products, as well as androst-4-en-3,
17-dione [289]. Another example of a bacterial P450 capable of steroid biotransfor-
mation is CYP106A2 from B. megaterium ATCC 13368. The monooxygenase can
hydroxylate steroids such as progesterone, 11-deoxycortisol, and testosterone. It
hydroxylates specifically 3-oxo-D4-steroids, whereas 3b-hydroxy-D5-steroids are not
converted. CYP106A2 hydroxylates mainly at the 15b position [292]. Other hydrox-
ylation positions described are 6b, 9a, and 11a, when using progesterone as
substrate [293, 294]. CYP106A2 was expressed in E. coli [295] and its activity in vitro
was supported by bovine adrenodoxin and adrenodoxin reductase as electron
1258
transfer partners [296]. Furthermore, using methods of directed evolution 11-

deoxycortisol hydroxylation was improved by a factor of more than four and
progesterone conversion was improved about 1.4-fold [297].
In 2006, the construction and application of a novel P450-library, based on about
250 bacterial cytochrome P450 genes (about 70% from actinomycetes), co-expressed
with putidaredoxin and putidaredoxin reductase in E. coli was reported [298].
Screening of the P450-library with testosterone identified 24 bacterial P450s, which
stereoselectively monohydroxylate testosterone at the 2a-, 2b-, 6b-, 7b-, 11b-, 12b-,
15b-, 16a-, and 17-positions. Most of these hydroxylations are common for both
prokaryotic and human P450s. Thus the identified bacterial candidates can be further
applied for production of drug metabolites on a preparative scale.
30.7
Summary and Outlook
Many interesting oxygenase-mediated oxidations have been described in the liter-

ature so far. However, examples for process implementation and scale-up to pilot or
industrial scales are comparatively rare due to the complexity of these enzymes. The
invention of new methods of protein engineering in the past decade has led to the
construction of an abundance of mutants (mainly for P450s) with new tailored
properties. Recent major achievements include significant increases in productiv-
ities, yields, rates of catalytic turnover, and efficient multistep reactions in whole-cell
biocatalysts, coming one step closer to technical applications of oxygenases.
The number of identified novel oxygenases – associated with unknown pathways
of secondary metabolism – is constantly increasing through the sequencing of
genomes and microbial screenings. Mechanistic characterization and utilization of
these enzymes in drug development, fine chemical synthesis, bioremediation,
biosensors, and plant improvement presents new exciting perspectives for future
applications.
Acknowledgments
We wish to thank Matthias Gunne (Universit€at D€

usseldorf) for critical reading of the
manuscript.
References
1 Holland, H.L. and Weber, H.K. Biotransformations in Preparative

(2000) Curr. Opin. Biotechnol., 11, Organic Chemistry: The Use of
547–553. Isolated Enzymes and Whole-Cell
2 Davies, H.G., Green, R.H., Kelly, D.R., Systems in Synthesis, Academic Press,
and Roberts, S.M. (eds) (1989) London.
References j1259
3 Punniyamurthy, T., Velusamy, S., and (ed. R.H. Crabtree), Wiley-VCH Verlag
Iqbal, J. (2005) Chem. Rev., 105, GmbH, Weinheim, pp. 1–25.
2329–2363. 23 Urlacher, V.B. and Schmid, R.D.
4 Lucke, B., Narayana, K.V., Martin, A., and (2006) Curr. Opin. Chem. Biol., 10,
Jahnisch, K. (2004) Adv. Synth. Catal., 156–161.
346, 1407–1424. 24 Urlacher, V.B., Lutz-Wahl, S., and
5 Buhler, B. and Schmid, A. (2004) Schmid, R.D. (2004) Appl. Microbiol.
J. Biotechnol., 113, 183–210. Biotechnol., 64, 317–325.
6 Que, L. Jr. and Tolman, W.B. (2008) 25 Werck-Reichhart, D. and Feyereisen, R.
Nature, 455, 333–340. (2000) Genome Biol., 1, REVIEWS3003.
7 Bernhardt, R. (2006) J. Biotechnol., 124, 1–3003.9.
128–145. 26 Wong, L.L. (1998) Curr. Opin. Chem. Biol.,
8 Que, L. Jr. (2004) J. Biol. Inorg. Chem., 9, 2, 263–268.
643. 27 Ullrich, R. and Hofrichter, M. (2007)
9 Hakemian, A.S. and Rosenzweig, A.C. Cell Mol. Life Sci., 64, 271–293.
(2007) Annu. Rev. Biochem., 28 Garfinkel, D. (1958) Arch. Biochem.
76, 223–241. Biophys., 77, 493–509.
10 Kovaleva, E.G. and Lipscomb, J.D. (2008) 29 Klingenberg, M. (1958) Arch. Biochem.
Nat. Chem. Biol., 4, 186–193. Biophys., 75, 376–386.
11 King, N.J. and Thomas, S.R. (2007) 30 Omura, T. and Sato, R. (1964)
Int. J. Biochem. Cell Biol., 39, J. Biol. Chem., 239, 2370–2378.
2167–2172. 31 Raag, R. and Poulos, T.L. (1989)
12 Sono, M., Roach, M.P., Coulter, E.D., and Biochemistry, 28, 7586–7592.
Dawson, J.H. (1996) Chem. Rev., 96, 32 Bayer, E., Hill, H.O.A., R€oder, A., and
2841–2888. Williams, R.J.P. (1969) J. Chem. Soc.,
13 Nelson, D.R. (2006) Methods Mol. Biol., Chem. Commun., 109.
320, 1–10. 33 Hill, H.A.O., Roder, A., and
14 Nelson, D.R. (2011) Biochim. Biophys. Williams, R.J. (1969) Biochem. J., 115,
Acta, 1814, 14–18. 59P–60P.
15 Fischer, M., Knoll, M., Sirim, D., Wagner, 34 Poulos, T.L., Finzel, B.C., and
F., Funke, S., and Pleiss, J. (2007) Howard, A.J. (1986) Biochemistry, 25,
Bioinformatics, 23, 2015–2017. 5314–5322.
16 Sirim, D., Wagner, F., Lisitsa, A., and 35 Graham, S.E. and Peterson, J.A. (1999)
Pleiss, J. (2009) BMC Biochem., Arch. Biochem. Biophys., 369, 24–29.
10, 27. 36 Li, H. and Poulos, T.L. (2004) Curr. Top.
17 Park, J., Lee, S., Choi, J., Ahn, K., Park, B., Med. Chem., 4, 1789–1802.
Kang, S., and Lee, Y.H. (2008) BMC 37 Denisov, I.G., Makris, T.M.,
Genomics, 9, 402. Sligar, S.G., and Schlichting, I.
18 Nelson, D.R. (2009) Hum. Genomics, 4, (2005) Chem. Rev., 105,
59–65. 2253–2277.
19 Hamdane, D., Zhang, H., and 38 Sligar, S.G. (1976) Biochemistry, 15,
Hollenberg, P. (2008) Photosynth. Res., 98, 5399–5406.
657–666. 39 Filatov, M., Harris, N., and Shaik, S.
20 Kelly, S.L., Lamb, D.C., and Kelly, D.E. (1999) Angew. Chem. Int. Ed., 38,
(2006) Biochem. Soc. Trans., 34, 3510–3512.
1159–1160. 40 Groves, J.T. and Mcclusky, G.A. (1976)
21 Ortiz de Montellano, P.R. (ed.) J. Am. Chem. Soc., 98, 859–861.
Cytochrome P450: Structure, 41 Davydov, R., Makris, T.M.,
Mechanism and Biochemistry (2005) 3rd Kofman, V., Werst, D.E., Sligar, S.G., and
edn, Kluwer Academic/Plenum Press, Hoffman, B.M. (2001) J. Am. Chem. Soc.,
New York. 123, 1403–1415.
22 Urlacher, V.B. (2009) Green Catalysis, 42 Koppenol, W.H. (2007) J. Am. Chem. Soc.,
Vol. 3: Biocatalysis, 1st edn 129, 9686–9690.
1260
43 Vaz, A.D., McGinnity, D.F., and Coon, 62 Hersleth, H.P., Ryde, U., Rydberg, P.,
M.J. (1998) Proc. Natl. Acad. Sci. USA, 95, Gorbitz, C.H., and Andersson, K.K.
3555–3560. (2006) J. Inorg. Biochem., 100,
44 Jin, S., Bryson, T.A., and Dawson, J.H. 460–476.
(2004) J. Biol. Inorg. Chem., 9, 63 Zaks, A. and Dodds, D.R. (1995)
644–653. J. Am. Chem. Soc., 117, 10419–10424.
45 Sligar, S.G., Makris, T.M., and 64 Miller, V.P., Tschirret-Guth, R.A., and
Denisov, I.G. (2005) Biochem. Biophys. Ortiz de Montellano, P.R. (1995) Arch.
Res. Commun., 338, 346–354. Biochem. Biophys., 319, 333–340.
46 Bernhardt, R. (2004) Chem. Biol., 11, 65 Hu, S.H. and Hager, L.P. (1998) Biochem.
287–288. Biophys. Res. Commun., 253, 544–546.
47 Hannemann, F., Bichet, A., Ewen, K.M., 66 Hu, S.H. and Hager, L.P. (1999) J. Am.
and Bernhardt, R. (2007) Biochim. Chem. Soc., 121, 872–873.
Biophys. Acta, 1770, 330–344. 67 van de Velde, F., Bakker, M.,
48 Rabe, K.S., Spengler, M., Erkelenz, M., van Rantwijk, F., and Sheldon, R.A.
Muller, J., Gandubert, V.J., Hayen, H., (2001) Biotechnol. Bioeng., 72, 523–529.
and Niemeyer, C.M. (2009) 68 van Deurzen, M.P.J., van Rantwijk, F.,
ChemBioChem, 10, 751–757. and Sheldon, R.A. (1996) J. Mol. Catal. B
49 Cirino, P.C. and Arnold, F.H. (2003) Enzym., 2, 33–42.
Angew. Chem. Int. Ed., 42, 3299–3301. 69 Akasaka, R., Mashino, T., and Hirobe, M.
50 Kotze, A.C. (1999) Int. J. Parasitol., 29, (1995) Bioorg. Med. Chem. Lett., 5,
389–396. 1861–1864.
51 Rabe, K.S., Kiko, K., and Niemeyer, C.M. 70 Ullrich, R. and Hofrichter, M. (2005)
(2008) ChemBioChem, 9, 420–425. FEBS Lett., 579, 6247–6250.
52 Lee, D.S., Yamada, A., Sugimoto, H., 71 Kuehnel, K., Blankenfeldt, W.,
Matsunaga, I., Ogura, H., Terner, J., and Schlichting, I. (2006)
Ichihara, K., Adachi, S., Park, S.Y., and J. Biol. Chem., 281, 23990–23998.
Shiro, Y. (2003) J. Biol. Chem., 278, 72 Sundaramoorthy, M., Terner, J., and
9761–9767. Poulos, T.L. (1995) Structure,
53 Cryle, M.J., Stok, J.E., and De Voss, J.J. 3, 1367–1377.
(2003) Aust. J. Chem., 56, 749–762. 73 Leak, D.J., Sheldon, R.A.,
54 Guengerich, F.P. (2001) Curr. Drug. Woodley, J.M., and Adlercreutz, P.
Metab., 2, 93–115. (2009) Biocatal. Biotransform.,
55 Isin, E.M. and Guengerich, F.P. 27, 1–26.
(2007) Biochim. Biophys. Acta, 1770, 74 Lippard, S.J. (2005) Philos. Trans. A: Math
314–329. Phys. Eng. Sci., 363, 861–877.
56 Adam, W., Lazarus, M., 75 Rosenzweig, A.C., Frederick, C.A.,
Saha-Moller, C.R., Weichold, O., Hoch, Lippard, S.J., and Nordlund, P. (1993)
U., Haring, D., and Schreier, P. (1999) Nature, 366, 537–543.
Adv. Biochem. Eng. Biotechnol., 63, 76 Rosenzweig, A.C., Nordlund, P.,
73–108. Takahara, P.M., Frederick, C.A., and
57 Colonna, S., Gaggero, N., Richelmi, C., Lippard, S.J. (1995) Chem. Biol.,
and Pasta, P. (1999) Trends Biotechnol., 17, 2, 409–418.
163–168. 77 Jin, Y. and Lipscomb, J.D. (1999)
58 Murphy, C.D. (2003) J. Appl. Microbiol., Biochemistry, 38, 6178–6186.
94, 539–548. 78 Lee, S.K., Nesheim, J.C., and
59 Hofrichter, M. and Ullrich, R. Lipscomb, J.D. (1993) J. Biol. Chem.,
(2006) Appl. Microbiol. Biotechnol., 71, 268, 21569–21577.
276–288. 79 Rinaldo, D., Philipp, D.M., Lippard, S.J.,
60 Sugano, Y. (2009) Cell Mol. Life Sci., 66, and Friesner, R.A. (2007) J. Am. Chem.
1387–1403. Soc., 129, 3135–3147.
61 van Rantwijk, F. and Sheldon, R.A. (2000) 80 Kopp, D.A. and Lippard, S.J. (2002) Curr.
Curr. Opin. Biotechnol., 11, 554–564. Opin. Chem. Biol., 6, 568–576.
References j1261
81 Basch, H., Mogi, K., Musaev, D.G., and Coordination Chemistry II
Morokuma, K. (1999) J. Am. Chem. Soc., (eds J.A. McCleverty, and T.J. Meyer),
121, 7249–7256. Elsevier, Amsterdam, pp. 395–436.
82 Jin, Y. and Lipscomb, J.D. (2000) Biochim. 103 Evans, J.P., Ahn, K., and Klinman, J.P.
Biophys. Acta, 1543, 47–59. (2003) J. Biol. Chem., 278,
83 Chan, S.I. and Yu, S.S. (2008) Acc. Chem. 49691–49698.
Res., 41, 969–979. 104 Chen, P., Bell, J., Eipper, B.A., and
84 Colby, J., Stirling, D.I., and Dalton, H. Solomon, E.I. (2004) Biochemistry, 43,
(1977) Biochem. J., 165, 395–402. 5735–5747.
85 Green, J. and Dalton, H. (1989) 105 Francisco, W.A., Wille, G., Smith, A.J.,
J. Biol. Chem., 264, 17698–17703. Merkler, D.J., and Klinman, J.P.
86 Lipscomb, J.D. (1994) Annu. Rev. (2004) J. Am. Chem. Soc., 126,
Microbiol., 48, 371–399. 13168–13169.
87 Shanklin, J. and Whittle, E. (2003) FEBS 106 Kitmitto, A., Myronova, N., Basu, P., and
Lett., 545, 188–192. Dalton, H. (2005) Biochemistry, 44,
88 van Beilen, J.B. and Funhoff, E.G. (2005) 10954–10965.
Curr. Opin. Biotechnol., 16, 308–314. 107 Lieberman, R.L. and Rosenzweig, A.C.
89 Bertrand, E., Sakai, R., Rozhkova- (2005) Nature, 434, 177–182.
Novosad, E., Moe, L., Fox, B.G., Groves, 108 Myronova, N., Kitmitto, A., Collins, R.F.,
J.T., and Austin, R.N. (2005) J. Inorg. Miyaji, A., and Dalton, H. (2006)
Biochem., 99, 1998–2006. Biochemistry, 45, 11905–11914.
90 Groves, J.T. (2003) Proc. Natl. Acad. Sci. 109 Hille, R. (2005) Arch. Biochem. Biophys.,
USA, 100, 3569–3574. 433, 107–116.
91 Austin, R.N., Buzzi, K., Kim, E., 110 Dilworth, G.L. (1982) Arch. Biochem.
Zylstra, G.J., and Groves, J.T. (2003) Biophys., 219, 30–38.
J. Biol. Inorg. Chem., 8, 733–740. 111 Holcenberg, J.S. and Stadtman, E.R.
92 Fitzpatrick, P.F. (1999) Annu. Rev. (1969) J. Biol. Chem., 244,
Biochem., 68, 355–381. 1194–1203.
93 Pavon, J.A. and Fitzpatrick, P.F. (2009) 112 Self, W.T. and Stadtman, T.C. (2000) Proc.
J. Am. Chem. Soc., 131, 4582–4583. Natl. Acad. Sci. USA, 97, 7208–7213.
94 Carr, R.T., Balasubramanian, S., 113 Wagner, R., Cammack, R., and
Hawkins, P.C., and Benkovic, S.J. (1995) Andreesen, J.R. (1984) Biochim. Biophys.
Biochemistry, 34, 7525–7532. Acta, 791, 63–74.
95 Guroff, G. and Rhoads, C.A. (1969) 114 Schrader, T., Rienhofer, A., and
J. Biol. Chem., 244, 142–146. Andreesen, J.R. (1999) Eur. J. Biochem.,
96 Fitzpatrick, P.F. (2003) Biochemistry, 264, 862–871.
42, 14083–14091. 115 Huber, R., Hof, P., Duarte, R.O.,
97 Itoh, S. (2006) Curr. Opin. Chem. Biol., Moura, J.J., Moura, I., Liu, M.Y., LeGall, J.,
10, 115–122. Hille, R., Archer, M., and Romao, M.J.
98 Klinman, J.P. (2006) J. Biol. Chem., (1996) Proc. Natl. Acad. Sci. USA,
281, 3013–3016. 93, 8846–8851.
99 Prigge, S.T., Kolhekar, A.S., Eipper, B.A., 116 Romao, M.J., Archer, M., Moura, I.,
Mains, R.E., and Amzel, L.M. (1999) Nat. Moura, J.J., LeGall, J., Engh, R.,
Struct. Biol., 6, 976–983. Schneider, M., Hof, P., and Huber, R.
100 Blumberg, W.E., Goldstein, M., Lauber, (1995) Science, 270, 1170–1176.
E., and Peisach, J. (1965) Biochim. 117 Truglio, J.J., Theis, K., Leimkuhler, S.,
Biophys. Acta, 99, 187–190. Rappa, R., Rajagopalan, K.V.,
101 Ljones, T., Flatmark, T., Skotland, T., and Kisker, C. (2002) Structure, 10,
Petersson, L., Backstrom, D., and 115–125.
Ehrenberg, A. (1978) FEBS Lett., 118 Okamoto, K., Matsumoto, K., Hille, R.,
92, 81–84. Eger, B.T., Pai, E.F., and Nishino, T. (2004)
102 Halcrow, M.A. (2004) Monocopper Proc. Natl. Acad. Sci. USA, 101,
oxygenases, in Comprehensive 7931–7936.
1262
119 Stockert, A.L., Shinde, S.S., Anderson, 136 Lawrence, C.C., Sobey, W.J., Field, R.A.,
R.F., and Hille, R. (2002) J. Am. Chem. Baldwin, J.E., and Schofield, C.J. (1996)
Soc., 124, 14554–14555. Biochem. J., 313, 185–191.
120 Bugg, T.D. and Ramaswamy, S. (2008) 137 Mori, H., Shibasaki, T., Uozaki, Y.,
Curr. Opin. Chem. Biol., 12, 134–140. Ochiai, K., and Ozaki, A. (1996) Appl.
121 Gibson, D.T., Koch, J.R., Schuld, C.L., and Environ. Microbiol., 62, 1903–1907.
Kallio, R.E. (1968) Biochemistry, 7, 138 Broca, C., Manteghetti, M., Gross, R.,
3795–3802. Baissac, Y., Jacob, M., Petit, P., Sauvaire,
122 Ferraro, D.J., Gakhar, L., and Y., and Ribes, G. (2000) Eur. J. Pharmacol.,
Ramaswamy, S. (2005) Biochem. Biophys. 390, 339–345.
Res. Commun., 338, 175–190. 139 Brash, A.R. (1999) J. Biol. Chem., 274,
123 Ohta, T., Chakrabarty, S., Lipscomb, J.D., 23679–23682.
and Solomon, E.I. (2008) J. Am. Chem. 140 Feussner, I. and Wasternack, C.
Soc., 130, 1601–1610. (2002) Annu. Rev. Plant. Biol., 53,
124 Wolfe, M.D., Parales, J.V., Gibson, D.T., 275–297.
and Lipscomb, J.D. (2001) J. Biol. Chem., 141 Gillmor, S.A., Villasenor, A.,
276, 1945–1953. Fletterick, R., Sigal, E., and Browner, M.F.
125 Tarasev, M. and Ballou, D.P. (2005) (1997) Nat. Struct. Biol., 4, 1003–1009.
Biochemistry, 44, 6197–6207. 142 Schneider, C., Pratt, D.A., Porter, N.A.,
126 Gibson, D.T. and Parales, R.E. (2000) and Brash, A.R. (2007) Chem. Biol., 14,
Curr. Opin. Biotechnol., 11, 236–243. 473–488.
127 Boyd, D.R. and Sheldrake, G.N. (1998) 143 Coffa, G., Schneider, C., and Brash, A.R.
Nat. Prod. Rep., 15, 309–324. (2005) Biochem. Biophys. Res. Commun.,
128 Hudlicky, T., Gonzalez, D., and 338, 87–92.
Gibson, D.T. (1999) Aldrichim. Acta, 144 Liavonchanka, A. and Feussner, I. (2006)
32, 35–62. J. Plant Physiol., 163, 348–357.
129 Bowers, N.I., Boyd, D.R., Sharma, N.D., 145 van Beilen, J.B., Duetz, W.A., Schmid, A.,
Goodrich, P.A., Groocock, M.R., Blacker, and Witholt, B. (2003) Trends Biotechnol.,
A.J., Goode, P., and Dalton, H. (1999) 21, 170–177.
J. Chem. Soc., Perkin Trans. 1, 146 Faber, K. (ed.) (2004) Biotransformations in
1453–1461. Organic Chemistry, Springer-Verlag,
130 Hausinger, R.P. (2004) Crit. Rev. Biochem. Berlin.
Mol. Biol., 39, 21–68. 147 Gilardi, G. and Fantuzzi, A. (2001) Trends
131 Price, J.C., Barr, E.W., Tirupati, B., Biotechnol., 19, 468–476.
Bollinger, J.M. Jr., and Krebs, C. (2003) 148 Hlavica, P. (2009) Biotechnol. Adv., 27,
Biochemistry, 42, 7497–7508. 103–121.
132 Myllyharju, J. (2008) Ann. Med., 149 Hollmann, F., Hofstetter, K., and
40, 402–417. Schmid, A. (2006) Trends Biotechnol., 24,
133 Passoja, K., Rautavuoma, K., 163–171.
Ala-Kokko, L., Kosonen, T., and Kivirikko, 150 Nicolaou, K.C., Ramphal, J.Y.,
K.I. (1998) Proc. Natl. Acad. Sci. USA, Petasis, N.A., and Serhan, C.N. (1991)
95, 10482–10486. Angew. Chem. Int. Ed. Engl., 30,
134 Kodera, T., Smirnov, S.V., 1100–1116.
Samsonova, N.N., Kozlov, Y.I., 151 Goni, G., Zollner, A., Lisurek, M.,
Koyama, R., Hibi, M., Ogawa, J., Velazquez-Campoy, A., Pinto, S.,
Yokozeki, K., and Shimizu, S. (2009) Gomez-Moreno, C., Hannemann, F.,
Biochem. Biophys. Res. Commun., 390, Bernhardt, R., and Medina, M.
506–510. (2009) Biochim. Biophys. Acta, 1794,
135 Ogawa, J., Kodera, T., Smirnov, S.V., 1635–1642.
Hibi, M., Samsonova, N.N., Koyama, R., 152 Fer, M., Corcos, L., Dreano, Y.,
Yamanaka, H., Mano, J., Kawashima, T., Plee-Gautier, E., Salaun, J.P., Berthou, F.,
Yokozeki, K., and Shimizu, S. (2011) Appl. and Amet, Y. (2008) J. Lipid Res., 49,
References j1263
153 Urlacher, V. and Schmid, R.D. 168 Wilson, C.R., Craft, D.L., Eirich, L.D.,
(2002) Curr. Opin. Biotechnol., 13, Eshoo, M., Madduri, K.M., Cornett, C.A.,
557–564. Brenner, A.A., Tang, M., Loper, J.C., and
154 Dietrich, M., Eiben, S., Asta, C., Do, T.A., Gleeson, M. (2000) WO patent,
Pleiss, J., and Urlacher, V.B. (2008) 2000020566.
Appl. Microbiol. Biotechnol., 169 Green, K.D., Turner, M.K., and
79, 931–940. Woodley, J.M. (2000) Enzyme Microb.
155 Hilker, B.L., Fukushige, H., Hou, C., and Technol., 27, 205–211.
Hildebrand, D. (2008) Prog Lipid Res, 47, 170 Weiss, A. (2007) Selective microbail
1–14. oxidations in industry: Oxidations of
156 Warman, A.J., Roitel, O., Neeli, R., alkanes, fatty acids, heterocyclic
Girvan, H.M., Seward, H.E., compounds, aromatic compounds and
Murray, S.A., McLean, K.J., Joyce, M.G., glycerol using native or recombinant
Toogood, H., Holt, R.A., microorganisms, in Modern Oxidation:
Leys, D., Scrutton, N.S., and Enzymes, Reactions and Applications
Munro, A.W. (2005) Biochem. Soc. Trans., (eds R.D. Schmid and V.B. Urlacher),
33, 747–753. Wiley-VCH Verlag GmbH, Weinheim,
157 Guengerich, F.P. (2004) Drug Metab. Rev, pp. 193–210.
36, 159–197. 171 Craft, D.L., Madduri, K.M., Eshoo, M.,
158 Li, H. and Poulos, T.L. (1999) Biochim. and Wilson, C.R. (2003) Appl. Environ.
Biophys. Acta, 1441, 141–149. Microbiol., 69, 5983–5991.
159 Munro, A.W., Girvan, H.M., and 172 Kuo, T.M., Lanser, A.C., Nakamura, L.K.,
McLean, K.J. (2007) Biochim. Biophys. and Hou, C.T. (2000) Curr. Microbiol., 40,
Acta, 1770, 345–359. 105–109.
160 Munro, A.W., Leys, D.G., McLean, K.J., 173 Kuo, T.M., Nakamura, L.K., and
Marshall, K.R., Ost, T.W., Daff, S., Lanser, A.C. (2002) Curr. Microbiol., 45,
Miles, C.S., Chapman, S.K., Lysek, D.A., 265–271.
Moser, C.C., Page, C.C., and 174 Schneider, S., Wubbolts, M.G.,
Dutton, P.L. (2002) Trends Biochem Sci, Oesterhelt, G., Sanglard, D., and
27, 250–257. Witholt, B. (1999) Biotechnol. Bioeng., 64,
161 Dietrich, M., Do, T.A., Schmid, R.D., 333–341.
Pleiss, J., and Urlacher, V.B. (2009) 175 Schneider, S., Wubbolts, M.G.,
J. Biotechnol., 139, 115–117. Sanglard, D., and Witholt, B. (1998) Appl.
162 Budde, M., Morr, M., Schmid, R.D., and Environ. Microbiol., 64, 3784–3790.
Urlacher, V.B. (2006) ChemBioChem, 7, 176 Labinger, J.A. and Bercaw, J.E. (2002)
789–794. Nature, 417, 507–514.
163 Matsunaga, I., Yokotani, N., Gotoh, O., 177 Murrell, J.C., Gilbert, B., and
Kusunose, E., Yamada, M., and McDonald, I.R. (2000) Arch. Microbiol.,
Ichihara, K. (1997) J. Biol. Chem., 272, 173, 325–332.
23592–23596. 178 Lee, S.G., Goo, J.H., Kim, H.G., Oh, J.I.,
164 Matsunaga, I., Ueda, A., Fujiwara, N., Kim, Y.M., and Kim, S.W. (2004)
Sumimoto, T., and Ichihara, K. (1999) Biotechnol. Lett., 26, 947–950.
Lipids, 34, 841–846. 179 Cooley, R.B., Dubbels, B.L.,
165 Girhard, M., Schuster, S., Dietrich, M., Sayavedra-Soto, L.A., Bottomley, P.J., and
Durre, P., and Urlacher, V.B. (2007) Arp, D.J. (2009) Microbiology, 155,
Biochem. Biophys. Res. Commun., 362, 2086–2096.
114–119. 180 Dubbels, B.L., Sayavedra-Soto, L.A., and
166 Kaya, K., Ramesha, C.S., and Arp, D.J. (2007) Microbiology, 153,
Thompson, G.A. Jr. (1984) J. Biol. Chem., 1808–1816.
259, 3548–3553. 181 Dubbels, B.L., Sayavedra-Soto, L.A.,
167 Koch, A.K., Kappeli, O., Fiechter, A., and Bottomley, P.J., and Arp, D.J. (2009)
Reiser, J. (1991) J. Bacteriol., 173, Int. J. Syst. Evol. Microbiol.,
4212–4219. 59, 1576–1578.
1264
182 Kotani, T., Yamamoto, T., Yurimoto, H., 198 Fujii, T., Narikawa, T., Sumisa, F.,
Sakai, Y., and Kato, N. (2003) J. Bacteriol., Arisawa, A., Takeda, K., and Kato, J.
185, 7120–7128. (2006) Biosci. Biotechnol. Biochem., 70,
183 Witholt, B., de Smet, M.J., Kingma, J., 1379–1385.
van Beilen, J.B., Kok, M., Lageveen, R.G., 199 van Beilen, J.B., Funhoff, E.G., van Loon,
and Eggink, G. (1990) Trends Biotechnol., A., Just, A., Kaysser, L., Bouza, M.,
8, 46–52. Holtackers, R., Rothlisberger, M., Li, Z.,
184 Eggink, G., Engel, H., Vriend, G., and Witholt, B. (2006) Appl. Environ.
Terpstra, P., and Witholt, B. (1990) Microbiol., 72, 59–65.
J. Mol. Biol., 212, 135–142. 200 Chang, D., Feiten, H.J., Engesser, K.H.,
185 Eggink, G., van Lelyveld, P.H., van Beilen, J.B., Witholt, B., and Li, Z.
Arnberg, A., Arfman, N., Witteveen, C., (2002) Org. Lett., 4, 1859–1862.
and Witholt, B. (1987) J. Biol. Chem., 262, 201 Chang, D.L., Feiten, H.J., Witholt, B., and
6400–6406. Li, Z. (2002) Tetrahedron-Asymmetry, 13,
186 Peterson, J.A., Basu, D., and Coon, M.J. 2141–2147.
(1966) J. Biol. Chem., 241, 5162–5164. 202 Chang, D.L., Feiten, H.J., Witholt, B., and
187 Kok, M., Oldenhuis, R., van der Linden, Li, Z. (2004) Tetrahedron-Asymmetry, 15,
M.P., Raatjes, P., Kingma, J., van 571–572.
Lelyveld, P.H., and Witholt, B. (1989) 203 Li, Z. and Chang, D.L. (2004) Curr. Org.
J. Biol. Chem., 264, 5435–5441. Chem., 8, 1647–1658.
188 van Beilen, J.B., Panke, S., Lucchini, S., 204 van Beilen, J.B., Smits, T.H., Whyte, L.G.,
Franchini, A.G., Rothlisberger, M., and Schorcht, S., Rothlisberger, M.,
Witholt, B. (2001) Microbiology, 147, Plaggemeier, T., Engesser, K.H., and
1621–1630. Witholt, B. (2002) Environ. Microbiol., 4,
189 Benson, S., Oppici, M., Shapiro, J., and 676–682.
Fennewald, M. (1979) J. Bacteriol., 140, 205 Smits, T.H., Balada, S.B., Witholt, B., and
754–762. van Beilen, J.B. (2002) J. Bacteriol., 184,
190 Ruettinger, R.T., Olson, S.T., Boyer, R.F., 1733–1742.
and Coon, M.J. (1974) Biochem. Biophys. 206 Wang, L., Tang, Y., Wang, S., Liu, R.L.,
Res. Commun., 57, 1011–1017. Liu, M.Z., Zhang, Y., Liang, F.L., and
191 Johnson, E.L. and Hyman, M.R. (2006) Feng, L. (2006) Extremophiles, 10,
Appl. Environ. Microbiol., 72, 950–952. 347–356.
192 Li, Z., Feiten, H.J., van Beilen, J.B., Duetz, 207 Feng, L., Wang, W., Cheng, J., Ren, Y.,
W., and Witholt, B. (1999) Tetrahedron Zhao, G., Gao, C., Tang, Y., Liu, X., Han,
Asymmetry, 10, 1323–1333. W., Peng, X., Liu, R., and Wang, L. (2007)
193 van Beilen, J.B., Li, Z., Duetz, W.A., Proc. Natl. Acad. Sci. USA, 104,
Smits, T.H.M., and Witholt, B. (2003) Oil 5602–5607.
Gas Sci. Technol. - Rev. de lIFP, 58, 208 Li, L., Liu, X., Yang, W., Xu, F.,
427–440. Wang, W., Feng, L., Bartlam, M., Wang,
194 Ratajczak, A., Geissdorfer, W., and L., and Rao, Z. (2008) J. Mol. Biol., 376,
Hillen, W. (1998) Appl. Environ. 453–465.
Microbiol., 64, 1175–1179. 209 Gunsalus, I.C. and Sligar, S.G. (1976)
195 Maier, T., Forster, H.H., Asperger, O., and Biochimie, 58, 143–147.
Hahn, U. (2001) Biochem. Biophys. Res. 210 Stevenson, J.-A., Westlake, A.C.G.,
Commun., 286, 652–658. Whittock, C., and Wong, L.-L. (1996) J.
196 Nodate, M., Kubota, M., and Misawa, N. Am. Chem. Soc., 118, 12846–12847.
(2006) Appl. Microbiol. Biotechnol., 71, 211 Bell, S.G., Stevenson, J.A., Boyd, H.D.,
455–462. Campbell, S., Riddle, A.D., Orton, E.L.,
197 Kubota, M., Nodate, M., and Wong, L.L. (2002) Chem. Commun.,
Yasumoto-Hirose, M., Uchiyama, T., 490–491.
Kagami, O., Shizuri, Y., and Misawa, N. 212 Xu, F., Bell, S.G., Lednik, J., Insley, A.,
(2005) Biosci. Biotechnol. Biochem., 69, Rao, Z., and Wong, L.L. (2005) Angew.
2421–2430. Chem. Int. Ed., 44, 4029–4032.
References j1265
213 Fasan, R., Chen, M.M., Crook, N.C., and 231 Mukherjee, B.B., Kraidman, G., and
Arnold, F.H. (2007) Angew. Chem. Int. Hill, I.D. (1973) Appl. Microbiol., 25,
Ed., 46, 8414–8418. 447–453.
214 Farinas, E.T., Schwaneberg, U., Glieder, 232 Noma, Y., Yamasaki, S., and
A., and Arnold, F.H. (2001) Adv. Synth. Asakawa, Y. (1992) Phytochemistry, 31,
Catal., 343, 601–606. 2725–2727.
215 Glieder, A., Farinas, E.T., and Arnold, 233 Cadwallader, K.R., Braddock, R.J., Parish,
F.H. (2002) Nat. Biotechnol., 20, M.E., and Higgins, D.P. (1989) J. Food
1135–1139. Sci., 54, 1241–1245.
216 Meinhold, P., Peters, M.W., 234 Cheong, T.K. and Oriel, P.J. (2000) Appl.
Chen, M.M.Y., Takahashi, K., and Biochem. Biotechnol., 84–86, 903–915.
Arnold, F.H. (2005) ChemBioChem., 6, 235 Pescheck, M., Mirata, M.A., Brauer, B.,
1765–1768. Krings, U., Berger, R.G., and Schrader, J.
217 Peters, M.W., Meinhold, P., Glieder, A., (2009) J. Ind. Microbiol. Biotechnol., 36,
and Arnold, F.H. (2003) J. Am. Chem. 827–836.
Soc., 125, 13442–13450. 236 Duetz, W.A., Jourdat, C., and Witholt, B.
218 Lentz, O., Feenstra, A., Habicher, T., (2001) EP 1236802.
Hauer, B., Schmid, R.D., and 237 van Beilen, J.B., Holtackers, R., Luscher,
Urlacher, V.B. (2006) ChemBioChem., 7, D., Bauer, U., Witholt, B., and Duetz,
345–350. W.A. (2005) Appl. Environ. Microbiol., 71,
219 Dordick, J.S., Khmelnitsky, Y.L., and 1737–1744.
Sergeeva, M.V. (1998) Curr. Opin. 238 Bell, S.G., Chen, X., Xu, F., Rao, Z., and
Microbiol., 1, 311–318. Wong, L.L. (2003) Biochem. Soc. Trans., 31,
220 Maimone, T.J. and Baran, P.S. (2007) Nat. 558–562.
Chem. Biol., 3, 396–407. 239 Seifert, A., Vomund, S., Grohmann, K.,
221 Dhavalikar, R.S. and Bhattacharyya, P.K. Kriening, S., Urlacher, V.B., Laschat, S.,
(1966) Indian J. Biochem., 3, 144–157. and Pleiss, J. (2009) ChemBioChem., 10,
222 Duetz, W.A., Bouwmeester, H., van 853–861.
Beilen, J.B., and Witholt, B. (2003) Appl. 240 Colocousi, A., Saqib, K.M., and Leak, D.J.
Microbiol. Biotechnol., 61, 269–277. (1996) Appl. Microbiol. Biotechnol., 45,
223 Schrader, J. (2007) Microbial flavor 822–830.
production, in Flavours and Fragrances: 241 Bicas, J.L., Fontanille, P., Pastore, G.M.,
Chemistry, Bioprocessing and Sustainability and Larroche, C. (2008) J. Appl. Microbiol.,
(ed. R.G. Berger), Springer, Berlin, pp. 105, 1991–2001.
507–574. 242 Krings, U., Lehnert, N., Fraatz, M.A.,
224 Lupien, S., Karp, F., Wildung, M., and Hardebusch, B., Zorn, H., and Berger,
Croteau, R. (1999) Arch. Biochem. R.G. (2009) J. Agric. Food Chem., 57,
Biophys., 368, 181–192. 9944–9950.
225 Wust, M., Little, D.B., Schalk, M., and 243 Bell, S.G., Chen, X., Sowden, R.J., Xu, F.,
Croteau, R. (2001) Arch. Biochem. Williams, J.N., Wong, L.L., and Rao, Z.
Biophys., 387, 125–136. (2003) J. Am. Chem. Soc., 125, 705–714.
226 Wust, M. and Croteau, R.B. (2002) 244 Branco, R.J., Seifert, A., Budde, M.,
Biochemistry, 41, 1820–1827. Urlacher, V.B., Ramos, M.J., and Pleiss, J.
227 van Dyk, M.S., van Rensburg, E., and (2008) Proteins, 73, 597–607.
Moleleki, N. (1998) Biotechnol. Lett., 20, 245 Muller, B., Dean, C., Schmidt, C., and
431–436. Kuhn, J.-C. (1998) US patent, 5847226.
228 Onken, J. and Berger, R.G. (1999) J. 246 Hashimoto, T., Asakawa, Y., Noma, Y.,
Biotechnol., 69, 163–168. Murakaki, C., Tanaka, M., Kanisawa, T.,
229 Duetz, W.A., Fjallman, A.H., Ren, S., and Emura, M. (2003) Japanese patent,
Jourdat, C., and Witholt, B. (2001) Appl. 2003070492.
Environ. Microbiol., 67, 2829–2832. 247 Furusawa, M., Hashimoto, T., Noma, Y.,
230 Duetz, W.A., Jourdat, C., and Witholt, B. and Asakawa, Y. (2005) Chem. Pharm.
(2000) EP 1205556. Bull. (Tokyo), 53, 1513–1514.
1266
248 Furusawa, M., Hashimoto, T., Noma, Y., and Schmid, R.D. (1998) Appl. Environ.
and Asakawa, Y. (2005) Chem. Pharm. Microbiol., 64, 3878–3881.
Bull. (Tokyo), 53, 1423–1429. 265 Celik, A., Flitsch, S.L., and Turner, N.J.
249 Hashimoto, T., Asakawa, Y., Noma, Y., (2005) Org. Biomol. Chem., 3, 2930–2934.
Murakaki, C., Furusawa, M., Kanisawa, 266 Urlacher, V.B., Makhsumkhanov, A., and
T., Emura, M., and Mitsuhashi, K. Schmid, R.D. (2006) Appl. Microbiol.
(2003) Japanese patent, 2003250591. Biotechnol., 70, 53–59.
250 Kaspera, R., Krings, U., Nanzad, T., and 267 Girhard, M., Klaus, T., Khatri, Y.,
Berger, R.G. (2005) Appl. Microbiol. Bernhardt, R., and Urlacher, V.B.
Biotechnol., 67, 477–483. (2010) Appl. Microbiol. Biotechnol., 87,
251 de Kraker, J.W., Schurink, M., 595–607.
Franssen, M.C.R., Konig, W.A., de Groot, 268 Khatri, Y., Girhard, M., Romankiewicz,
A., and Bouwmeester, H.J. (2003) A., Ringle, M., Hannemann, F.,
Tetrahedron, 59, 409–418. Urlacher, V.B., Hutter, M.C., and
252 Willershausen, H. and Graf, H. Bernhardt, R. (2010) Appl. Microbiol.
(1996) CLB Chem. Labor Biotechnik, 47, Biotechnol., 88, 485–495.
24–28. 269 Miller, W.L. (2008) Endocr. Dev., 13, 1–18.
253 Huang, R., Christenson, P., and 270 Miller, W.L. and Tyrell, J.B. (1995)
Labuda, I.M. (2001) European patent, The adrenal cortex, in Endocrinology and
1083233. Metabolism (eds P. Felig, J. Baxter, and L.
254 Fraatz, M.A., Riemer, S.J.L., St€ober, R., Frohman), McGraw-Hill Press, New
Kaspera, R., Nimtz, M., Berger, R.G., and York, pp. 555–711.
Zorn, H. (2009) J. Mol. Catal. B Enzym., 271 White, P.C., Curnow, K.M., and
61, 202–207. Pascoe, L. (1994) Endocr. Rev., 15,
255 Kr€ugener, S., Krings, U., Zorn, H., and 421–438.
Berger, R.G. (2010) Bioresour. Technol., 272 Bernhardt, R. (1996) Rev. Physiol.
101, 457–462. Biochem. Pharmacol., 127, 137–221.
256 Girhard, M., Machida, K., Itoh, M., 273 Hakki, T. and Bernhardt, R. (2006)
Schmid, R.D., Arisawa, A., and Pharmacol. Ther., 111, 27–52.
Urlacher, V.B. (2009) Microb. Cell Fact., 274 Hakki, T., Zearo, S., Dragan, C.A.,
8, 36. Bureik, M., and Bernhardt, R. (2008) J.
257 Sowden, R.J., Yasmin, S., Rees, N.H., Biotechnol., 133, 351–359.
Bell, S.G., and Wong, L.L. (2005) Org. 275 Fernandes, P., Cruz, A., Angelova, B.,
Biomol. Chem., 3, 57–64. Pinheiro, H.M., and Cabral, J.M.S. (2003)
258 Eschenmoser, W., Uevelhart, P., and Enzyme Microb. Technol., 32, 688–705.
Eugster, C.H. (1981) Helv. Chim. Acta, 64, 276 Mahato, S.B. and Garai, S. (1997) Steroids,
2681. 62, 332–345.
259 Oritani, T. and Yamashita, K. (1984) 277 Petzoldt, K., Annen, K., Laurent, H., and
Kagaku Seibutsu, 22, 21. Wiechert, R. (1982) US patent,
260 Eugster, C.H. and Maerki-Fischer, E. 4353985.
(1991) Angew. Chem. Int. Ed. Engl., 30, 278 Hogg, J.A. (1992) Steroids, 57, 593–616.
654–672. 279 Peterson, D.H., Murray, H.C., Eppstein,
261 Sefton, M.A., Skouroumounis, G.K., S.H., Reineke, L.M., Weintraub, A.,
Massy-Westropp, R.A., and Williams, P.J. Meister, P.D., and Leigh, H.M. (1952)
(1989) Aust. J. Chem., 42, 2071–2084. J. Am. Chem. Soc., 74, 5933–5936.
262 Sode, K., Karube, I., Araki, R., and 280 Holland, H.L. (1999) Steroids, 64,
Mikami, Y. (1989) Biotechnol. Bioeng., 33, 178–186.
1191–1195. 281 Lamm, A.S., Chen, A.R., Reynolds, W.F.,
263 Kakeya, H., Sugai, T., and and Reese, P.B. (2007) Steroids, 72,
Ohta, H. (1991) Agric. Biol. Chem., 55, 713–722.
1873–1876. 282 Kolek, T. and Swizdor, A. (1998)
264 Lutz-Wahl, S., Fischer, P., Schmidt- J. Steroid. Biochem. Mol. Biol.,
Dannert, C., Wohlleben, W., Hauer, B., 67, 63–69.
References j1267
283 Farooq, A. and Tahara, S. (2000) J. Nat. 292 Berg, A., Carlstrom, K., Gustafsson, J.A.,
Prod., 63, 489–491. and Ingelman-Sundberg, M. (1975)
284 Smith, K.E., Latif, S., and Kirk, D.N. Biochem. Biophys. Res. Commun., 66,
(1989) J. Steroid. Biochem., 32, 445–451. 1414–1423.
285 Huszcza, E., Dmochowska-Gladysz, J., 293 Berg, A., Gustafsson, J.A., and
and Bartmanska, A. (2005) Z. Ingelman-Sundberg, M. (1976) J. Biol.
Naturforsch., Teil C., 60, 103–108. Chem., 251, 2831–2838.
286 Brzezowska, E., Dmochowska-Gladysz, 294 Lisurek, M., Kang, M.J., Hartmann, R.W.,
J., and Kolek, T. (1996) J. Steroid. Biochem. and Bernhardt, R. (2004) Biochem.
Mol. Biol., 57, 357–362. Biophys. Res. Commun.,
287 Hu, S., Genain, G., and Azerad, R. (1995) 319, 677–682.
Steroids, 60, 337–352. 295 Virus, C., Lisurek, M., Simgen, B.,
288 Mahato, S.B. and Mukherjee, A. (1984) Hannemann, F., and Bernhardt, R. (2006)
J. Steroid. Biochem., 21, 341–342. Biochem. Soc. Trans., 34, 1215–1218.
289 Al-Awadi, S., Afzal, M., and Oommen, S. 296 Berg, A., Ingelman-Sundberg, M., and
(2003) Appl. Microb. Biotechnol., 62, Gustafsson, J.A. (1979) J. Biol. Chem.,
48–52. 254, 5264–5271.
290 Wilson, M.R., Gallimore, W.A., and 297 Virus, C. and Bernhardt, R. (2008) Lipids,
Reese, P.B. (1999) Steroids, 64, 43, 1133–1141.
834–843. 298 Agematu, H., Matsumoto, N., Fujii, Y.,
291 Holland, H.L., Lakshmaiah, G., and Kabumoto, H., Doi, S., Machida, K.,
Ruddock, P.L. (1998) Steroids, 63, Ishikawa, J., and Arisawa, A. (2006) Biosci.
484–495. Biotechnol. Biochem., 70, 307–311.
j1269
31
Oxyfunctionalization of C–C Multiple Bonds
Bruno B€uhler, Katja B€
uhler, and Frank Hollmann
31.1
Introduction
The oxyfunctionalization of variably substituted C¼C double bonds is of special

interest for organic synthesis, as it typically leads to the formation of one or two
new chiral centers within prochiral substrates. However, the enantiomeric purity of
the product formed depends on the specificity of the catalyst, which is one of the
strengths of enzymatic catalysis. Furthermore, double bond oxyfunctionalization
allows the formation of highly reactive synthons with epoxides as the most prominent
example. Epoxidation is thus the most intensively investigated type of double bond
oxyfunctionalization, also in biocatalysis. However, biocatalytic dihydroxylations and
double bond cleavage have also been reported. Rare examples of triple bond
oxyfunctionalization exist, which, however, are associated with enzyme inactivation.
Without attempting to be exhaustive, we will focus in this chapter on enzymatic
activities allowing the direct oxyfunctionalization of olefins, using recombinant
or wild-type microbial cells or isolated enzymes. First, enzymes and respective
mechanisms enabling C–C multiple bond oxyfunctionalization will be reviewed,
followed by an overview of different functionalization schemes, of which olefin
epoxidation is the most prominent.
31.2
Enzymes Capable of C–C Multiple Bond Oxyfunctionalization
The oxidative introduction of oxygen atoms into C–C multiple bonds typically is
catalyzed by diverse classes of oxygenases and by peroxidases [1–5]. Even hydrolases,
for example, lipases, have been reported to mediate epoxidations via perhydrolysis of
carboxylic acids and esters with the resulting peroxyacids acting as the oxidizing
species to form alkene oxides [6].
Among the oxidoreductases, oxygenases are the most prominent enzyme class to
catalyze multiple bond oxyfunctionalizations. Except for double bond cleavage, C–C
multiple bond oxyfunctionalization by oxygenases depends on an electron donor,
and Oliver May.
j 31 Oxyfunctionalization of C–C Multiple Bonds
1270
which typically is NAD(P)H or an oxoacid cosubstrate. The high abundance and

versatility of oxygenases, their ability to specifically introduce oxygen from O2, and
the absence of enzyme-destabilizing peroxides as reactants have led to considerable
advances in process implementation and to first examples of industrial processes
[3, 7–13]. Most published biocatalytic processes for preparative multiple-bond
oxyfunctionalizations include oxygenase catalysis using whole microbial cells as
catalysts.
Catalytic mechanisms of oxygenases are as diverse as the active sites involved.
Different prosthetic groups are capable of reductive oxygen activation and C–C
multiple bond oxyfunctionalization, including flavins (typically FAD) [14, 15], heme
(cytochrome P450 monooxygenases) [16–18], and mono- or binuclear non-heme
iron [19–25].
31.2.1
Binuclear Non-heme Iron Oxygenases
Binuclear non-heme iron enzymes such as methane monooxygenase consist of two

to four protein components involving a soluble or membrane-bound hydroxylase
containing the diiron cluster in the active site, a reductase channeling electrons from
NADH either via a ferredoxin or rubredoxin component or directly to the hydroxylase,
and, in the case of the soluble hydroxylases, a small effector protein. The latter
modulates the conformation of the hydroxylase, with several roles in catalysis,
including control of substrate access as well as enhancement of the electron transfer
and O2 activation [26–29]. Typically, four glutamate and two histidine residues
coordinate the diiron center in soluble hydroxylases, whereas the diiron cluster of
membrane-bound hydroxylases is complexed by histidines only [30, 31]. The catalytic
mechanism was most thoroughly investigated for the prototype of this enzyme class,
the soluble methane monooxygenase (sMMO) [32, 33]. The catalytic cycle involves the
reaction of O2 with the reduced diferrous form to give a diferric peroxy species
(compound P in Figure 31.1) [34, 35], which spontaneously converts into a bis m-oxo
Fe(IV)2 cluster, the so-called diamond core (compound Q in Figure 31.1) [34, 36,
37]. It is, however, not clear if the latter conversion occurs via homolytic or heterolytic
O–O cleavage [33]. Both the peroxy and the oxo forms have been proposed to be the
oxygenating species, depending on the type of reaction to be catalyzed [38, 39]. Both
have also been proposed to be involved in the epoxidation of double bonds, depending
on the electronegativity of atoms/groups connected to the double bond [40, 41].
Whereas electron-rich substrates were proposed to follow a two-electron mechanism
with the peroxy form as the oxygenating species, a one-electron mechanism involving
compound Q was proposed for less electron-rich substrates (Figure 31.1).
31.2.2
Mononuclear Non-heme Iron Oxygenases
Mononuclear non-heme iron enzymes contain either a high-spin ferrous site that is
involved in O2 activation or a high-spin ferric site that activates substrates [21, 22, 24,
31.2 Enzymes Capable of C–C Multiple Bond Oxyfunctionalization j1271
NADH + H+ NAD+ + H2O
MMOR
MMOHox H MMOHred
O O
(III) (III)
Fe Fe H 2O
O O2
O O (II) (II)
R4 R3 Fe Fe
H
R2(O) R1
H2O
H2O
H H
O O
(III) (III) (III) (III)
T Fe Fe Fe Fe R4 R3
(III) O (III)
O O Fe Fe
R4 R3 O
R2O R1
R2(O) R1 +
R2 O R1
P
O
(IV) (III) O
Fe Fe (III) (III)
Fe Fe
O
O
R C
R2 R1
(IV)
O (IV)
Fe Fe
R4 R3 O
Q
R2 R1
Figure 31.1 Catalytic cycle of soluble MMO. oxygenating species, respectively. MMOR,
Compounds P, Q, R, and T are intermediates of reductase component of MMO; MMOH,
the catalytic cycle following the nomenclature hydroxylase component of MMO. Figure based
given in the literature. Two pathways for more on studies performed by Lipscomb and
and less electron-rich substrates are indicated, coworkers [42, 43] and Lippard and
with compounds P and Q as the proposed coworkers [27, 33, 41].
25]. The ferric site is coordinated by a variable histidine-rich ligand environment and is
known to catalyze intradiol aromatic ring cleavage and lipoxygenations, of which the
latter can be considered a C¼C double bond oxyfunctionalization. The ferrous site
typically is coordinated by two histidines and one aspartate or glutamate, a recurring
motif referred to as the 2-His-1-carboxylate facial triad (Figure 31.2). The remaining
three ligand sites are readily available for the binding of substrate, cofactor, and/or O2
during catalysis, making this ferrous site very versatile, catalyzing a wide variety of
reactions [44, 45]. Most oxygenases featuring the 2-His-1-carboxylate facial triad can be
classified into four main groups: extradiol cleaving catechol dioxygenases, 2-oxo acid
dependent enzymes, pterin-dependent hydroxylases, and Rieske oxygenases (Fig-
ure 31.2) [19, 46], of which the latter three have been reported to catalyze multiple bond
oxyfunctionalizations (not considering aromatic ring cleavage).
1272 j 31 Oxyfunctionalization of C–C Multiple Bonds
O O
R R OH His
Asp Asp (II)
HN H H S (III)
His
HN H H S (IV) His
Fe
Fe Fe R O His
O His His Glu
H O O
O N
N Me B: H
H H2O O Me
Me H
H COOH Me O O
COOH O
oxidase His
(II)
Fe
R O His
O2 Glu
O2
O extradiol dioxygenase
O H2O
H2O His H2N
(II) N
Fe
O N O His His
His H2O His O (II)
O (III) Fe
Fe O O2 O2 O
O O Glu/Asp HN NH
Asp Glu
N R R
H O2
COOH
O H2N
His His
CO2 N
O (IV)
N O Fe
O
O OH Glu
His
Fe
(III) HN NH
O O OH
O His O OH R
Asp Fe
(III)
His R
(V)
His Fe
N Asp His pterin dependent
H Asp
COOH hydroxylase
2-oxo acid
dependent dioxygenase Rieske type dioxygenase
Figure 31.2 Various modes of oxygen including oxidases, are shown together
activation by members of the 2-His/Glu facial with a model substrate (isopenicillin N,
triad family. In the central structure, the triad is substituted catechol, substituted benzene,
shown in bold and the three ligand sites shown naphthalene, proline) and, if applicable,
to be occupied by H2O can be vacant, occupied a cofactor (pterin) or cosubstrate
by OH, or occupied by a weak protein ligand in (2-oxoglutarate). Oxygen atoms derived from
different enzymes from the family. In a first molecular oxygen are shown in bold. B:
reaction, a substrate or cofactor is bound and, refers to a basic amino acid of the
concomitantly, solvent water is released to open enzyme, promoting OO bond cleavage by
an oxygen binding site at the iron. Active sites of providing a proton. Figure adapted from
the different enzymes, beside oxygenases, also Kovaleva et al. [46].
The 2-oxo acid dependent iron enzymes constitute a versatile family and initiate
catalysis by oxidative decarboxylation of a 2-oxo acid cosubstrate (e.g., 2-oxoglutarate),
which mediates the formation of a highly oxidizing Fe(IV)¼O intermediate
(Figure 31.2) [45, 47, 48]. In most cases, this intermediate activates C–H bonds in
substrates by abstraction of the H-atom followed by oxygen rebound leading to substrate
hydroxylation. However, different reactivities such as double bond epoxidation, triple
bond oxyfunctionalization, and oxidative ligand transfer involved in halogenations
have been reported [49–52]. With respect to oxyfunctionalizations, 2-oxoglutarate-
dependent hydroxylases can be classified as intermolecular dioxygenases and depend
directly on the central carbon metabolism providing 2-oxoglutarate.
In pterin-dependent hydroxylases, the pterin cofactor was proposed to fulfill a dual
role, the delivery of two electrons and the activation of O2 together with ferrous iron
resulting in the formation of hydroxylated pterin and a Fe(IV)¼O oxygenating
intermediate allowing double bond epoxidation (Figure 31.2) [53–55]. Regeneration
of tetrahydrobiopterin involves dehydration and NAD(P)H-dependent reduction of
the hydroxylated pterin cofactor.
Rieske dioxygenases, which primarily catalyze aromatic dihydroxylations but also
show alternative reactivities such as epoxidation catalysis, form a diverse group of
two- or three-component enzymes with a reductase component that obtains electrons
from NAD(P)H, often a ferredoxin component that shuttles the electrons, and an
oxygenase component where O2 activation and substrate oxidation occur [56, 57]. As a
special property of Rieske oxygenases, the mononuclear iron receives the electrons
needed for catalysis from the Rieske cluster of the neighboring oxygenase subunit.
Both an iron-(hydro)peroxide and a HO-Fe(V)¼O intermediate have been proposed
as oxygenating species [19, 24, 58].
As an exception, the ferrous site of recently characterized carotenoid oxygenases
is not coordinated by the 2-His-1-carboxylate facial triad but by four histidines [59, 60].
These enzymes catalyze the regiospecific double bond cleavage in carotenoids
to yield two carbonyl products (aldehydes or ketones depending on the substituents
of the vinyl group cleaved) [61, 62]. Two mechanisms have been proposed. For
the carotenase AtCCD1 from Arabidopsis thaliana, 18 O-incorporation studies pro-
vided convincing evidence that carotenoid cleavage follows a dioxygenase mecha-
nism via a dioxetane or related 1,2-dioxo-cyclic intermediate (Figure 31.3) [63].
Although this can be assumed to be the typical mechanism for carotenoid cleavage,
some enzymes such as the mammalian b-carotene-15,150 -monooxygenase have been
proposed to catalyze carotenoid cleavage via epoxide and diol intermediates [64, 65].
Mechanistic studies using a quantum chemical method indicated a lower activation
barrier for the dioxygenase mechanism while not ruling out the monooxygenases
mechanism [66], so that the mechanism followed may depend on the enzyme. Like
ring cleavage dioxygenases and lipoxygenases, carotenoid oxygenases do not depend
on an additional electron donor since all the electrons necessary to reduce molecular
oxygen are derived from the substrate.
Lipoxygenases, which catalyze the hydroperoxidation of polyunsaturated fatty acids
via O2 incorporation, involve ferric iron with a histidine-rich ligand environment.
Fe(III) has been proposed to activate the substrate by hydrogen abstraction, yielding
Fe(II) and a carbon-centered radical intermediate (Figure 31.4) [67]. This is followed
by O2 addition to the activated radical species, producing a peroxy radical interme-
diate, which can reoxidize Fe(II) to Fe(III), forming the hydroperoxy product [25].
Remarkably, the resting as-isolated Fe(II)-containing enzyme is activated by the fatty
acid hydroperoxide product [68]. Lipoxygenases are primarily involved in the syn-
thesis of secondary metabolites and the metabolism of endo- and xenobiotics [4] and
do not require an additional electron source. The two-electron oxidation of the
substrate is coupled to the one-electron reduction of two oxygen atoms from
molecular oxygen.
1274
R2 R = H, CH3
R1
carotenoid
O2 O2
R R
(II) (IV)
Fe R2 R2 Fe O
R1 R1
O O
O
dioxetane intermediate epoxide intermediate
H2O
R
(II) R
Fe R2 (IV)
O R2 R1 Fe O
R1 O HO
HO
diol intermediate
H2O
R (II)
Fe
O R2
R1 O
Figure 31.3 Proposed dioxygenase (left) and originating from molecular oxygen are
monooxygenase (right) mechanisms for the shown in bold. Oxygen atoms originating either
enzymatic cleavage of carotenoids catalyzed by from water or molecular oxygen are shown in
carotenoid oxygenases. Oxygen atoms grey.
31.2.3
Heme-Containing Monooxygenases
Heme monooxygenases harbor a protoporphyrin IX tetrapyrrole system containing

one catalytic iron nucleus; the resulting prosthetic group is designated a heme group.
Such heme monooxygenases usually belong to the class of cytochrome P450s and
catalyze a large range of reactions including C–C multiple bond oxyfunctionaliza-
tions (typically epoxidations). Mammalian cytochrome P450 monooxygenases
have been thoroughly studied in the context of drug metabolism and hormone
biosynthesis [69–71], whereas microbial P450 monooxygenases typically are inves-
tigated with respect to the degradation and biotransformation of xenobiotics and
hydrocarbons [72, 73]. Thus, although P450s so far have hardly been used for
biotechnological processes, these enzymes are highly interesting candidates for
(NHis) N His + R2 R1 (NHis) N His +

OAsn NHis OAsn N His
(III) (II)
Fe Fe +
NHis OH NHis OH2 R2 R1
O O O O
Ile O2
Ile
OO
H OO
R2 R1
R2 R1
Figure 31.4 Mechanism of lipoxygenase-catalyzed allylic peroxidation. Adapted from

Abu-Omar et al. [25].
biocatalytic applications [9, 11, 74]. Most cytochrome P450 systems reported to date
are multicomponent enzymes with additional proteins for the transport of reducing
equivalents from NAD(P)H to the terminal cytochrome P450 component. Typical
electron transfer chains of class I P450 systems (originating from bacteria or
mammalian adrenal mitochondria; soluble; NADH as typical electron donor) consist
of a NADH:ferredoxin oxidoreductase and a ferredoxin, whereas class II P450
systems (mammalian hepatic drug-metabolizing isoforms; membrane bound;
NADPH as typical electron donor) include a flavin containing P450 reductase [75].
An exception is cytochrome P450 BM-3 of Bacillus megaterium, which consists of a
single soluble polypeptide with a P450 domain and an electron transport domain of
the microsomal type (class II) [76]. The heme group of P450 monooxygenases is
directly involved in the oxidation process by activating O2. The catalytic cycle
(Figure 31.5) by which cytochrome P450-mediated oxyfunctionalization occurs is
still intensively studied [16–18, 77–82]. According to present understanding, sub-
strate binding is followed by a first electron transfer from NAD(P)H via the electron
transfer chain to the heme iron and reversible O2 binding to give a superoxide-iron
complex. A second reduction leads to peroxo-iron, which is protonated to hydro-
peroxo-iron. A second protonation results either in water abstraction and the
formation of an iron-oxo species, the prototype oxidant in P450 catalysis, or in
iron-complexed hydrogen peroxide, depending on whether the distal or the proximal
oxygen atom is protonated. Thereby, cytochrome P450-catalyzed oxyfunctionaliza-
tion was proposed to involve multiple mechanisms and oxidants, including hydro-
peroxo-iron, iron-complexed hydrogen peroxide, and two different spin states of the
iron-oxo species. Multiple mechanisms also have been proposed for C¼C double
bond epoxidation. However, recent studies give evidence that the iron-oxo species
(both spin states) are involved in P450-catalyzed epoxidations [17, 82]. Thereby, the
high-spin iron-oxo species has been proposed to give rise to long-lived radical
1276
O
FeIII
C
O
FeIII H2O FeIII e–
2e–
2H+
O H2O2
O2– FeII
FeV
H+
H2O2
H2O O2
– –
OH 2H+ O
O
O
H+ FeIII
O 2– FeIII
O
e–
FeIII
H+
Figure 31.5 Generalized CYP450 reaction cycle adapted for C¼C double bond epoxidation,
showing the three possible uncoupling reactions and the peroxide shunt pathway, which also
applies for heme-dependent peroxidases.
intermediates that can participate in side product formation and in scrambling of cis/
trans-stereochemistry, thus explaining the observed aldehyde formation and enzyme
inactivation by irreversible formation of a porphyrin N-alkylation product. At the
(hydro)peroxo stage, hydrogen peroxide can dissociate by one of the three possible
uncoupling mechanisms (Figure 31.5). This dissociation is reversible, and P450
enzymes can be shunted with hydrogen peroxide to give an active oxidant (see also
peroxidases below). The efficiency of this hydrogen peroxide shunt has been
improved by directed evolution [83, 84].
31.2.4
Flavin-Dependent Oxygenases
Flavin-dependent oxygenases functionalize a wide variety of substrates [15, 85] with

Bayer–Villiger oxidations [86, 87], aromatic hydroxylations [88], and epoxidations [89–
93] being the most prominent reactions catalyzed. In contrast to flavin-dependent
oxidases, the flavin is not covalently bound in flavoprotein monooxygenases.
Depending on the type of enzyme, flavin is tightly (non-covalently) bound and
reduced in the monooxygenase moiety itself or acts as an electron shuttle that is
reduced by a separate reductase component and then bound and stabilized by the
monooxygenase component. In either case, the electrons are derived from NAD(P)H.
R R R
O2
H3C N N O H3C N N O H3C N N O
+H+
NH NH -H+ NH
H3C N H3C N H3C N
H HO HO
O O O
reduced flavin peroxyflavin O hydroperoxyflavin OH
X, H+
X
nucleophilic electrophilic
oxygenation oxygenation
NAD(P)H
XO XO
R R
H3C N N O H3C N N O
NH NH
H3C N H3C N
HO
O NAD(P)+, H2O H O
oxidized flavin hydroxyflavin
Figure 31.6 General mechanism of oxygenation reactions catalyzed by flavin-dependent

monooxygenases, for which NAD(P)H delivers the necessary electrons. Figure adapted from
van Berkel et al. [15].
Peroxyflavin or hydroperoxyflavin are the activated oxygen species responsible for the
nucleophilic or electrophilic attack of substrates, respectively (Figure 31.6), and are
formed via the reaction of reduced flavin (in most cases FADH2) with O2 in the
following way [14, 15]: Upon a one-electron transfer from reduced flavin to O2, a
complex of superoxide and the flavin radical is formed. Upon spin inversion, for most
flavoprotein monooxygenases, a covalent adduct between the C4a of the flavin and
dioxygen is formed, yielding the above-mentioned reactive C4a-(hydro)peroxyflavin
species. Such a peroxyflavin is unstable and typically decays to form hydrogen
peroxide and oxidized flavin. However, flavoprotein monooxygenases are able to
stabilize this species in such a way that it can oxygenate a substrate. As a result of
nucleophilic or electrophilic attack on the substrate, a single oxygen atom is
incorporated into the substrate, while the other oxygen atom is reduced to water.
The specific type of oxygenation and selectivity depends on the shape and chemical
nature of the active site of the respective enzyme.
31.2.5
Peroxidases
Peroxidases also catalyze various multiple bond oxyfunctionalizations and, in con-

trast to most oxygenases, do not require any electron donor [94–97]. In such two-
electron oxidations, a peroxide serves as oxygen donor and electron acceptor and
one molecule of water (or alcohol in the case of organic peroxide driven reactions)
is produced as a coproduct. In catalyzing these oxygen transfer reactions, heme-
dependent peroxidases exhibit a reactivity that is more typical for cytochrome P450
monooxygenases than that of classical peroxidases, which typically catalyze oxidative
dehydrogenation reactions via one-electron processes. Heme-dependent peroxidases
are structurally and functionally related to P450 monooxygenases, which also catalyze
1278
substrate oxyfunctionalizations with hydrogen peroxide as oxygen source in the so-

called peroxide shunt pathway (Figure 31.5). The same oxidants and oxygen transfer
mechanisms have been proposed for these two enzyme groups. However, oxygen
transfer from oxidized peroxidase to the substrate is only possible when the heme
iron is accessible for the substrates. In contrast to P450 enzymes, the fifth (proximal)
ligand of the iron atom is a histidine, except in chloroperoxidase from Caldariomyces
fumago (CPO), which belongs to the group of haloperoxidases [98, 99]. With respect to
oxygen transfer reactions, this heme-thiolate peroxidase with iron ligated to cysteine,
as in P450 monooxygenases, is the most versatile enzyme of the known peroxidases.
An advantage of peroxidases is that they need no additional electron acceptor or donor
and thus no regeneration of cofactors such as NAD(P/H). However, a major
shortcoming is the low operational stability of peroxidases, generally resulting
from peroxide induced deactivation [100, 101]. An example is the facile oxidative
deterioration of the porphyrin ring in heme-dependent peroxidases such as CPO,
necessitating the maintenance of low hydrogen peroxide concentrations and/or the
in situ generation of hydrogen peroxide from O2 with a chemical reductant or an
oxidase [97, 101]. Typically, peroxidases are applied as isolated enzymes.
31.3
Epoxidation of C¼C Double Bonds
As discussed above, oxygenases and peroxidases provide highly activated oxygen

intermediates that can oxidize a wide range of functional groups. One of the most
studied among these has been the epoxidation of olefins [1–3, 5, 97, 102–104]. This
epoxidation is particularly interesting when applied to prochiral double bonds.
Spectacular success has been obtained in the field of asymmetric chemical epoxi-
dation, notably using Sharpless epoxidation catalysts for allyl alcohols and Jacobsen
catalysts for aryl olefins, which has made epoxides key intermediates in the synthesis
of chiral compounds [105–107]. However, these chemical catalysts often have a
limited substrate range and specificity and can be problematic from an environ-
mental perspective, as they are heavy metal-based. Biocatalysts can provide access to
complementary structural motifs and serve as an environmentally benign alternative
for selective epoxidation.
31.3.1
Aliphatic Olefins
Most reports on the biocatalytic epoxidation of aliphatic olefins involve heme-

dependent monooxygenases or non-heme diiron enzymes, with alkane monooxy-
genases, alkene monooxygenases, and methane monooxygenases (MMOs) as the
most prominent examples. As these enzymes typically are complex multicomponent
systems (2–6 components), which are cofactor-dependent (typically NAD(P)H)
and often membrane-associated, approaches for their in vitro application are scarce
(see below) and even expression in heterologous hosts proved to be difficult [2]. Thus,
31.3 Epoxidation of C¼C Double Bonds j1279
most studies focused on wild-type strains (i.e., native expression system in the native
host) or expression of cloned genes in homologous hosts. Nevertheless, these
enzymatic activities have been investigated and exploited for more than 30 years.
One of the earliest observations, made more than 50 years ago, implicating the
formation of epoxides during microbial olefin metabolism was the report by Bruyn
in 1954 on Candida lipolytica grown on 1-hexadecene producing 1-hexadecanediol
(about 5% of the hydrocarbon consumed was accounted for as the diol) [108]. 18 O from
molecular oxygen was shown to be incorporated into this diol and the 1,2-epoxide was
identified as one of the by-products of this metabolism [109, 110]. Several further
reports confirmed that enzymatic systems are able to achieve epoxidations. For
instance, van der Linden showed in 1963 that Pseudomonas aeruginosa grown on n-
heptane and resuspended in a buffer solution produced the epoxide from 1-
octene [111]. This led the authors to conclude that this epoxide was formed by enzymes
already present in the alkane-grown cells and that epoxidation might be catalyzed by the
same hydroxylases that would normally oxidize alkanes. A similar conclusion was
reached by Maynert and coworkers [112], who demonstrated that epoxides are
obligatory intermediates in the metabolism of simple olefins in rat liver microsomes.
However, the real breakthrough in the study of enzymatic epoxidations is due to
Abbot and coworkers [113] and to May and coworkers [114], who established
unequivocally that epoxides are formed from terminal olefins by the bacterial strain
Pseudomonas putida GPo1, formerly known as Pseudomonas oleovorans
(Scheme 31.1) [115]. They showed that 1-octene is epoxidized to 1,2-epoxyoctane
of (R)-configuration (70% e.e.) or hydroxylated to 7-octen-1-ol. The 1,7-diene is
exclusively epoxidized, affording (R)-( þ )-7,8-epoxy-1-octene (84% e.e.), which can
P. putida GPo1 O
+
HO
O2 H2O
P. putida GPo1 O P. putida GPo1 O
O
O2 H2O O2 H2O
P. putida GPo1 O
O2 H2O
P. putida GPo1
O
O2 H2O
O
O P. putida GPo1 O
O2 H2O
Scheme 31.1 Epoxidation of various alkenes by the alkane monooxygenase of P. putida GPo1.
1280
be further processed to the corresponding diepoxide [116]. This diepoxide was shown
to be essentially of (R,R) configuration, which indicated that the configuration of the
monoepoxide formed at one end of the molecule profoundly affects the stereochem-
ical course of the reaction. The enzymatic system responsible for these epoxidation
reactions was later shown to be a three-component alkane monooxygenase that
typically catalyzes the terminal hydroxylation of medium-chain alkanes and fatty
acids [117–127]. The three protein components include an NADH-rubredoxin
reductase, a rubredoxin, and a membrane-bound non-heme diiron oxygenase. The
monooxygenase from P. putida GPo1 can epoxidize terminal alkenes containing six
to twelve carbon atoms. This type of enzyme was shown to be widespread among
alkane-degrading microorganisms [115, 128–130]. Hydroxylation was shown to
predominate for the short substrates propylene and 1-butene, whereas epoxidation
is preferred for long substrates. For the medium length substrates, like for
instance 1-octene, both reactions occur. Thus, this substrate is epoxidized to 1,2-
epoxyoctane or hydroxylated to 7-octen-1-ol, while, for 1-decene, epoxidation largely
predominates. Interestingly, the epoxidation reaction exhibits a specificity far dif-
ferent from that expected from chemical reactivity, as terminal olefins are epoxidized
exclusively even in the presence of more highly substituted (electron-rich) double
bonds. Thus, cyclic and internal olefins were not epoxidized. In the reaction with
dienes, 1,5-hexadiene to 1,11-dodecadiene were epoxidized, while dienes with a
smaller number of carbon atoms were hydroxylated to the corresponding unsatu-
rated alcohols [131]. The reactivity was shown to be maximal for octadiene and
decreases rapidly as the carbon chain is shortened, but only slightly as the chain is
lengthened. The advantage of such alkane monooxygenase systems (and MMOs) as
compared to alkene monooxygenases is that often wild-type strains can be used as
pathways for the further degradation of epoxides are missing. Poor enantioselectivity
can be a disadvantage. Microbial strains containing both alkane and alkene specific
monooxygenases, however, have also been described.
To overcome inhibition by the toxic product, an organic solvent phase functioning
as product sink and substrate reservoir was used for the production of epoxyalkanes
from alkenes via whole-cell catalysis based on wild-type P. putida GPo1 [132–135].
This can be taken as a very early, if not the first, example of the application of the two-
liquid phase concept in whole-cell biocatalysis. Thereby, pure alkenes [134, 135] or
alkenes dissolved in cyclohexane [132, 136] were added as a second phase. In the
former case, repeated exchange of the aqueous phase with a fresh culture (cell
renewal procedure) allowed enhancement of the final product concentration [135]. In
this study, the highest overall productivity (0:08 g l1 1
tot h , corresponding to
1 1
0:16 g laq h ) was reached with an organic phase volume fraction of 50 vol.%,
resulting in a product concentration in the organic phase of 45 g l1 org . This concen-
tration could be increased up to 150 g l1 org by reducing the organic phase volume
fraction to 10 vol.%. In a later study on the epoxidation of 1,7-octadiene in two-liquid
phase systems, mass transfer was shown to determine the volumetric productivity,
depending on the ratio of the organic and aqueous phases, the degree of agitation,
and the cell concentration [136]. Furthermore, biocatalyst stability also depended on
the conditions in the two-liquid phase system, especially on the stirrer speed in a
stirred tank reactor. To avoid direct contact with the potentially toxic organic phase
and the formation of stable emulsions, thus simplifying product isolation,
separation of aqueous and organic phase by a membrane was applied for the
epoxidation of 1,7-octadiene to 1,2-epoxy-7,8-octene by P. putida GPo1 growing
continuously in a membrane bioreactor [137]. The organic phase consisted of a
mixture of the biotransformation substrate 1,7-octadiene and the growth substrate
heptane, which both partitioned over the membrane into the aqueous growth
medium. To avoid mass transfer limitation, the two constituents of the organic
phase were additionally fed directly into the aqueous medium. A dense silicone
rubber membrane was used to contact the two phases and no phase breakthrough of
either liquid, a major challenge associated with this technology, was observed.
A productivity of 0:23 g l1 aq h
1
was achieved after optimization of the aqueous
phase dilution rate. In another approach, the same epoxidation was performed in a
continuous closed-gas-loop reactor, in which the aqueous culture was contacted with
a gas stream saturated with heptane and 1,7-octadiene [138]. The gas stream also
stripped the product out of the reaction medium and was circulated through a
saturator/absorber module. This led to a long-term productivity of 0.11 g l1 h1. Co-
metabolism of heptane and 1,7-octadiene by the same enzyme was proposed to be
one limiting factor, which can be circumvented by recombinant oxygenase gene
expression in, for example, Escherichia coli. The latter approach was successfully
followed for alkane monooxygenase catalyzed hydroxylation reactions [139–143].
Pseudomonas putida GPo1 has also been used, among some other microorganisms,
for the stereospecific epoxidation of some aryl allyl ethers into ( þ )-aryl glycidyl
ethers (Figure 31.7). These intermediates were chemically converted into (S)-()-3-
substituted-1-alkylamino-2-propanols, which are the physiologically active compo-
nents of the b-adrenergic receptor blocking drugs. This method has been used to
synthesize (S)-()-Metoprolol and (S)-()-Atenolol with enantiomeric purities of
95.4–98.8% and 97%, respectively [144]. These applications are of great industrial
interest, since it has been shown that (S)-()-Metoprolol is 270–380 times more active
than its (R)-enantiomer [145]. In another study, phenyl glycidyl ether was produced
from allyl phenyl ether by the use of resting cells of Mycobacterium M156 [146].
Hexadecane was chosen as organic carrier solvent, and specific epoxidation rates in the
range 5–6 U g1 were reached for 1.5–2 h, whereupon activity was lost. This translates
to an aqueous phase based productivity of about 0:12 g l1aq h
1
for the applied small-
scale system with an aqueous phase volume fraction of 50 vol.%.
Various microorganisms have been screened for epoxidation activity [147].
Positive hits typically belonged to the genera Rhodococcus, Mycobacterium, Methylo-
sinus (and other methanotrophic bacteria [148]), Nocardia, Xanthobacter, and Pseu-
domonas. In Table 31.1 and below, several examples are described without being
exhaustive. For instance, it was shown that Corynebacterium equi (IFO 3730) grown on
n-octane can oxidize 1-hexadecene to give the corresponding optically pure (R)-
( þ )-epoxide (41% yield based on consumed substrate) [149, 150]. This strain also
assimilated other terminal olefins and produced the corresponding epoxides
from substrates that have a carbon chain longer than fourteen, although in very
low yields (<1%).
1282
O
Metoprolol
Organism Optical purity
R
(%)
O2
P. putida GPo1 Rhodococcus equi
95.4
H2O NCIB 12035
O
O Pseudomonas putida
98
NCIB 9571
R Pseudomonas putida
98.4
GPo1 NCIB 12035
Chemical
Pseudomonas aeruginosa
OH 98.8
NCIB 12035
O NHiPr
NHiPr
Figure 31.7 Synthetic route to Metoprolol via stereospecific epoxidation followed by chemical
epoxide cleavage [144]. The table gives optical product purities achieved with different strains.
R designates a methoxyethyl substituent.
In a commercial epoxidation process, the Japan Energy Corporation (formerly

Nippon Mining) used wild-type Rhodococcus rhodochrous B-276, formerly known as
Nocardia corralina B-276 [158], for the conversion of various alkenes [157, 159] (see
also Chapter 27). The alkene monooxygenase of this strain has a very large substrate
spectrum, is, in contrast to alkane monooxygenases such as the P. putida enzyme,
specific for double bonds, and also catalyzes subterminal epoxidations. Furthermore,
this strain has been shown to be able to epoxidize branched chain terminal olefins in
an asymmetric manner leading to (R)-epoxides showing optical purities of 76–90%
depending on the chain length [157]. These epoxides were used as chirons for further
synthesis of prostaglandin v-chains. C6–C12 alk-1-enes and various aromatic olefins
were diluted in a non-toxic carrier solvent (e.g., hexadecane or other alkanes) and
converted by glucose-grown resting cells [160], whereas C13–C18 alk-1-enes were
directly added to a culture growing on these alkenes [161]. Enantiomeric excesses
between 66% and 94% were reported for aliphatic products (see also Table 31.1).
With respect to the total volume, a product titer of 30 g l1 and a productivity of
0.25 g l1 h1 were achieved for (R)-1,2-epoxyoctane production on a 4 m3 scale (see
also Chapter 27). For (R)-1,2-epoxytetradecane production in a 5-l bioreactor, the
respective numbers were 80 g l1 and 0.56 g l1 h1. Glucose-grown Rhodococcus
rhodochrous B-276 has also been used in the epoxidation of smaller olefins (C3–C5),
which were added through the gas phase. Epoxide stripping via intense aeration
resulted in productivities of up to 17 mmol l1 h1 (for propylene oxide) and was
followed by extensive solvent absorption and distillation [157]. Thus, the use of the
Table 31.1 Epoxidation of different olefins by various bacterial strains containing methane, alkane,
and/or alkene monooxygenases.
Olefin Microorganism Epoxides Reference
(R) (%) (S) (%)
Methanotrophsa) 53–56 44–47 [147]

b)
Mycobacterium sp. E3 91 9 [147]
Mycobacterium sp. L1b) 99 1 [147]
Pseudomonas sp. P9yc) 45 55 [147]
Nocardia sp. IP1d) 91 9 [147]
Rhodococcus rhodochrous B-276c) 87 13 [147]
Xanthobacter sp. Py2e) 97 3 [151]
Nocardioides sp. JS614b) 99 1 [152]
Micrococcus sp. M90C 98 2 [153]
Cl
Mycobacterium sp. E3b) 1 99 [147]
Mycobacterium sp. Py8e) 1 99 [147]
R. rhodochrous B-276c) 11 89 [147]
Pseudomonas sp. P9yc) 45 55 [147]
Mycobacterium sp. Py8e) 93 7 [147]
OH
Br
Burkholderia cepacia G4f) 100 0 [155]
Pseudomonas mendocina KR1f) 54 46 [155]
P450 BM-3 variant RH-47 80 20 [156]
P450 BM-3 variant SH-44 13 87 [156]
(Continued )
1284
Olefin Microorganism Epoxides Reference
(R) (%) (S) (%)
P. putida GPo1g) 85 15 [135]

5 R. rhodochrous B-276 95 5 [157]
P450 BM-3 variant RH-47 92 8 [156]
P450 BM-3 variant SH-44 78 22 [156]
P. putida GPo1g) (epoxy) 92 8 [114]
4 P. putida GPo1g) (diepoxy) 83 17 [116]
P. putida GPo1g) 80 20 [135]

7 R. rhodochrous B-276
93 7 [157]
R. rhodochrous B-276 93 7 [157]
11
13 Corynebacterium equi IFO 3730
100 0 [150]
15
a) Grown on methane.
b) Grown on ethene.
c) Grown on propane.
d) Grown on isoprene.
e) Grown on propene.
f) Grown on toluene.
g) Grown on octane.
two-liquid phase concept as well as gas stripping enabled Japan Energy to commer-
cially produce a large variety of epoxides (see also Chapter 27).
With respect to the epoxidation of short-chain alkenes, an extensive study has been
conducted by de Bont and coworkers to prepare epoxides from gaseous olefins.
Different Mycobacterium species were shown to excrete ethylene oxide when grown on
ethylene, involving a monooxygenase type of reaction [162–164]. Experiments were
performed in a gas–solid reactor to prevent accumulation of the toxic ethylene oxide in
the immediate vicinity of the biocatalyst [164]. Using chiral gas chromatography,
eleven strains of alkene-utilizing bacteria were screened with respect to the stereo-
specific epoxidation of propene, 1-butene, and 3-chloro-1-propene. When grown on
alkenes, all strains produced 1,2-epoxypropane and 1,2-epoxybutane mainly in the (R)-
form (with 62– 94% e.e.; see also Table 31.1) [151]. Interestingly, when grown on
ethane, Mycobacterium E20 (one of the tested strains) showed low (S)-specificities (6%
and 18% e.e. for 1-propene and 1-butene epoxidation, respectively), pointing to the
presence of multiple differently regulated alkane- and alkene-specific monooxy-
genases with different enantiospecificities. Four strains were tested for the conversion
of 3-chloro-1-propene and mainly produced the (S)-epoxide (60–98% e.e.). Stereo-
selective epoxidation of 4-bromo-1-butene and of 3-buten-1-ol was similarly studied
using three strains (Table 31.1). Again, epoxides were predominantly obtained in the
(R)-form, with optical purities of 41–87% [154]. In general, the enantiomeric purities
of products depended on both the strain used and the substrate studied. Epoxides were
found to inactivate the whole-cell catalyst at the enzyme level (for alkene- and alkane
monooxygenases as well as MMO), thereby influencing the setup of a biotechnological
procedure for epoxide production [165]. Optimization was achieved by studying the
influence of various organic solvents on the retention of immobilized cell activity [166].
High activity retention was favored by a low polarity in combination with a high
molecular weight. Modeling the effects of mass transfer on the kinetics of propene
epoxidation was also achieved by the same authors [167, 168], and they showed that
product inhibition can be reduced by absorbing the epoxide in the gas phase in cold di-
n-octyl phthalate [169]. Furthermore, immobilization of the whole-cell biocatalysts has
also been reported to increase catalyst stability [170].
In addition to the Mycobacterium species, several other strains have been reported
to achieve epoxidation of olefins. Three distinct types of methane-grown methylo-
trophic bacteria (Methylosinus trichosporium, Methylobacterium capsulatus, and Methy-
lobacterium organophilum) were shown by Hou and coworkers [171] to be able to
oxidize terminal C2–C4 n-alkenes to their corresponding 1,2-epoxides. Results from
inhibition studies indicated, as in the case of the previously discussed alkane
monooxygenase of P. putida GPo1, that the same monooxygenase enzyme (in this
case particulate MMO) was responsible for the hydroxylation, in this case of methane,
and the epoxidation of alkenes. Further work by the same group showed that whole
cells of Methylosinus sp. CRL 31, immobilized by adsorption on glass beads, were able
to convert propylene into propylene oxide for several hours. The production period
could be prolonged by periodic addition of methanol as a source for reduction
equivalents and thus NADH regeneration. These authors also observed that
attempts to immobilize the cells by covalent binding or entrapment in polyacryl-
amide gel led to complete loss of propylene epoxidation activity [172]. Further studies
revealed that the epoxides were nearly racemic [173] as also observed for other
methanotrophs [147]. The major problem of these biotransformations was again
found to be product (epoxide) inhibition. Oxidation of propylene to propylene oxide
by Methylococcus capsulatus (Bath) was studied to optimize the biotransformation for
possible industrial production. However, the high rates obtained (up to 500 U g1 CDW ,
1 U g1
CDW ¼ 1 mmol product formed per min per gram cell dry weight) could only be
sustained for 3–4 min before loss of biocatalytic activity occurred [174]. The deac-
tivation was shown to depend on the propylene oxide concentration and the achieved
maximum activities. The latter could be limited by limiting cosubstrate (methanol)
availability. Limiting the maximal activity to 150 U g1 CDW and the concomitant
removal of propylene oxide allowed recovery of catalyst activity [175]. Reactivation
was shown to depend on the presence of growth substrates and on protein synthesis.
The use of the natural capability of microbial cells to self-immobilize as a biofilm
marks another interesting approach for immobilization, stabilization, and, thus,
continuous biocatalysis [176]. This concept was used for propene epoxidation by
applying a mixed culture of methanotrophs in a fluidized bed reactor, in which the
1286
cells formed biofilms on diatomite particles [177]. Again an appropriate co-feed, in

this case of methane, was necessary for cell growth and NADH regeneration. As
MMO was responsible for both methane catabolism and propene epoxidation,
optimization of the gaseous substrate feed in terms of methane, propene, O2, and
N2 content was crucial. The initial epoxypropane productivity on a 375 ml scale was
rather low at about 110–150 mmol day1, but the bioreactor operated continuously for
53 days without obvious loss of epoxypropane productivity.
Some Xanthobacter species were shown to be able to accumulate 1,2-epoxyethane
from ethene or, when grown on propene, to accumulate 2,3-epoxybutane from cis- or
trans-2-butene but with apparently low yields [178]. Similarly, Rhodococcus rhodo-
chrous, a propane-oxidizing strain, was shown to produce 1,2-epoxyalkanes from
short-chain terminal alkenes. Interestingly, its oxygenase enzyme appeared to be
capable of tolerating high levels of product without inhibition [179].
Mahmoudian and Michael [153, 180] isolated 18 bacterial strains, including
Aerococcus, Alcaligenes, Micrococcus, and Staphylococcus spp., able to produce optically
enriched epoxides from propene, 1-butene, and trans-2-butene with excellent e.e.s
for the (R)-enantiomers ranging from 90 to 96%, from 90 to 98%, and 64–88%,
respectively. However, in the case of trans-(2R,3R)-epoxybutane, it was shown that the
enantiomeric enrichment is in fact due to a second step enantioselective hydrolysis of
the epoxide, which is first produced in racemic form.
Recently, ethene-grown Nocardioides sp. strain JS614 cells were shown to be among the
most active alkene-oxidizers and chiral epoxyalkane accumulators in comparison to
previously studied ethene-grown bacteria (Table 31.1) [152]. High bioconversion rates and
yields were achieved combined with high enantiospecificities. Even though viability
plating and flow cytometry experiments indicated that JS614 suffers from product toxicity
during propene and 1-butene biotransformation, it was observed to be somewhat less
susceptible to the toxic effects of 1,2-epoxybutane than Mycobacterium strain E3.
A pharmaceutically interesting reaction is the stereospecific epoxidation of cis-
propenylphosphonate. Eighteen species of Penicillium, one of Oidium, and one of
Paecilomyces were found to effect this reaction, which affords directly ()-fosfomycin,
a broad-spectrum antibiotic. Using the strain Penicillium spinulosum MB 2843 at
optimum culture conditions, a 90% efficiency (based on olefin charged, 0.5 g l1) was
obtained after 6 days, leading to a product claimed to be optically pure [181]. Another
interesting example for the application of a fungal species was reported by Furstoss
and coworkers [182], who described the epoxidation of an (S)-sulcatol derivative by
Aspergillus niger, which affords the corresponding (2S,5S) enantiomer in 98% d.e. and
50% yield (Scheme 31.2). This product can be further transformed into the natural
enantiomer of the pheromone pityol.
Various toluene monooxygenases, also belonging to the non-heme diiron contain-
ing monooxygenase family, have also been reported to catalyze alkene epoxida-
tions [155]. Five toluene monooxygenase-containing bacterial strains have been
found to degrade linear alkenes and chloroalkenes three to eight carbons long at
specific rates between 20 and 660 U g1 CDW . The organism achieving the highest rate,
Pseudomonas mendocina KR1, obviously contained multiple enzyme systems capable
of alkene oxidation. However, recombinant E. coli expressing only the toluene-4-
OR OR
A. niger EtOH/NaOH O
O OH
O2 H2O
(2S) (2S,5S)-Pityol
(2S,5S)
R= CONHPh ee 100%; de 98%
Scheme 31.2 Epoxidation of a (S)-sulcatol derivative by the fungus Aspergillus niger; synthesis of the
pheromone pityol [182].
monooxygenase genes from Pseudomonas mendocina KR1 was able to epoxidize

butene, butadiene, pentene, and hexene. The enantiomeric ratios achieved with the
different organisms for butadiene and pentene epoxidation varied widely from very
high (100% (R): 0% (S)) to nearly racemic (54% (R): 46% (S)) depending on the
organism–substrate pair.
With respect to P450 monooxygenases, Bacillus megaterium was one of the first
organisms reported to catalyze cytochrome P450-dependent epoxidation reactions,
namely, the epoxidation of unsaturated fatty acids such as palmitoleic acid [183].
As in the case of alkane monooxygenases containing a non-heme diiron cluster,
epoxidation and hydroxylation are catalyzed by the same enzyme system. The enzyme
responsible for such epoxidations, the fatty acid hydroxylase cytochrome P450 BM-
3 [184], was the target of many directed evolution studies and variants have been
generated that catalyze hydroxylation and epoxidation of a wide range of non-native
substrates [185]. Variant 139-3 was shown to epoxidize propylene, cyclohexene, and
styrene (Scheme 31.3) [186]. Further engineering of BM-3 variants by saturation
mutagenesis of active site residues allowed not only for increased total turnover
numbers (TTNs) but also enhanced enantiospecificities [156]. Two variants with
opposite enantiospecificities have been generated, which convert terminal C5–C8
alkenes with 55–83% e.e. and 85–95% epoxidation selectivity (see also Table 31.1).
Peroxidases catalyze mainly two different types of C¼C double bond functiona-
lizations, namely, halohydrin formation by addition of hypohalous acid (see also
Chapter 37) and direct epoxidation by incorporation of one oxygen atom from a
peroxide [5, 94, 95]. The former reaction is not a classical oxyfunctionalization and
thus not the topic of this chapter, but can also be used to produce epoxides (typically
racemic) if coupled to a chemical or enzymatic cyclization reaction, for example,
catalyzed by a halohydrin epoxidase (Chapter 9). These C¼C double bond functio-
nalization reactions are catalyzed by haloperoxidases [98, 99, 187], of which chlor-
operoxidase from Caldariomyces fumago is the most commonly used enzyme [95, 96,
188, 189]. Studies on the direct epoxidation of alkenes showed that the best substrates
were aliphatic cis-disubstituted alkenes with a chain length up to C9 and the double
bond close to the chain terminus [188, 190]. In these cases, e.e. values as high as 95%
could be achieved with moderate to high yields. Lower yields and e.e. values were, in
general, obtained from differently substituted olefins such as trans-alkenes but
1288
O
P450 BM-3
NADPH NADP+
O2, H+ H2O
OH
P450 BM-3 +
O
NADPH NADP+
O2, H+ H2O (85%) (15%)
O
P450 BM-3
NADPH NADP+
O2, H+ H2O
P450 BM-3
2-5 2-5
O
NADPH NADP+
O2, H+ H2O
Scheme 31.3 Epoxidation of C¼C double bonds by cytochrome P450 BM-3 variants [156, 186].
also with cis-alkenes with the double bond far from the chain terminus (cis-3-alkenes).
In both cases, significant allylic hydroxylation was observed. Terminal alkenes gave
rise to heme alkylation and subsequent enzyme deactivation [191]. However, 2-
methyl-1-alkenes appeared to lead to interesting enantiomeric purities of epoxides in
certain cases (Table 31.2) [192, 193].
As mentioned in Section 31.2.5, the NAD(P/H)-independency is an advantage of
peroxidases. However, their low operational stability, generally resulting from
peroxide-induced deactivation [100, 101, 194] and heme alkylation, is a major
drawback. Different approaches have been followed to handle this instability prob-
lem [97], including the use of various reactor types and controlled hydrogen peroxide
addition, for example, resulting in a 20-fold increase of total turnover number and
space–time yield for the oxidation of indole to oxindole [195, 196]. Other approaches
are based on the in situ formation of the oxidant (see also Table 31.6 below), for
example, by co-immobilization of glucose oxidase to achieve efficient cis-2-heptene
epoxidation [197], directed evolution [198], solvent engineering [101], the use of
surfactants [199], and the use of ternary water–organic solvent systems [198, 200]. The
latter approach with a ternary system composed of a-pinene, t-butanol, and aqueous
buffer also was used for cis-2-heptene epoxidation.
Table 31.2 Epoxidation of different olefins by chloroperoxidase.
Olefin E.e. (%) Absolute Epoxide Reference

configuration yield (%)
96 (2R,3S) 78 [190]
92 (2R,3S) 82 [190]
95 (R) 23 [192]
O 94 (R) 34 [192]
89 (R) 22 [192]
62 (S) 61 [193]
Br
88 (R) 93 [193]
Br
95 (R) 89 [193]
Br
87 (R) 33 [193]
Br
50 (R) 42 [193]
Br
Hydrolases have also been reported to mediate epoxidations; however, not by direct
oxygen transfer catalysis but by catalyzing the perhydrolysis of carboxylic acids and
esters with the resulting peroxyacids acting as the oxidizing species to form alkene
oxides (Scheme 31.4) [6]. As in the case of peroxidases, hydrogen peroxide or organic
peroxides serve as the oxidants, in this case for ester or acid perhydrolysis. The actual
1290
R1 R3 R1 O R3
R2 R4 R2 R4
O O
R5 O OH R5 OH
Lipase
H2O H 2O 2
Scheme 31.4 Illustration of the chemoenzymatic epoxidation of an alkene, involving the

biocatalytic perhydrolysis of a carboxylic acid followed by a peroxy acid epoxidation.
epoxidation step is purely chemical, resulting in racemic products. Examples include

the epoxidation of terminal alkenes with a chain length of 8 to 16 carbon atoms as well
as some cycloalkenes [201–203], for example, cyclohexene [204] and 1-methylcyclo-
hexene [205]. As the reaction is not directly catalyzed by enzymes, it does not depend
on their substrate spectrum. This also enabled the epoxidation of polymers by this
strategy [206]. The addition of a carboxy-group-containing cosubstrate could be
avoided in the lipase-mediated epoxidation of unsaturated fatty acids or their esters.
After perhydrolysis of the carboxy group, a self -epoxidation could be obtained [203,
207–209]. Recently, the epoxidation of C¼N bonds [210] and even the direct lipase-
catalyzed epoxidation of a,b-unsaturated aldehydes have also been reported [211].
The latter was facilitated by exchanging the active-site serine of Candida antarctica
lipase B with alanine, which is a clear proof that this reaction does not follow the
indirect epoxidation mechanism but is the result of a novel catalytic promiscuity with
a histidine playing a key role in the direct epoxidation mechanism. In general, the
same peroxide-caused instability problems as in the case of peroxidases have to be
overcome for productive lipase-mediated epoxidation [6].
31.3.2
Vinylaromatic Substrates
The conversion of vinylaromatic compounds mostly covers epoxidation reactions.

The preferred catalysts for this transformation are flavin-dependent monooxy-
genases (see also Section 31.2.4) as well as heme-dependent monooxygenases
(Section 31.2.3) and peroxidases (Section 31.2.5), which will be discussed in this
section.
In addition to their well-described activity as Baeyer–Villiger oxidation catalysts,
flavin-dependent monooxygenases have been reported to be capable of epoxidizing
activated, vinylaromatic C¼C double bonds [15, 212]. These so-called styrene
monooxygenases have been described from various sources such as Pseudomonas [91,
213–215], Rhodococcus [216], or metagenome screenings [217].
Table 31.3 Substrate scope of styrene monooxygenase (StyAB).
Substrate Product E.e (%) Yield (%) Reference

(mass (g))
99.5 76.3 (12.6) [229]
99.9 46.5 (8.7) [229]
96.7 74.8 (11.3) [229]
99.8 87.2 (15.4) [229]
99.4 87.3 (18.3) [229]
99.4 72 (analytical) [230]
98.1 35 (analytical) [230]
05.6 27 (analytical) [230]
98.5 53.0 (10.2) [229]
98.0 47.9 (11.0) [229]
Styrene monooxygenase (StyAB) from Pseudomonas sp. strain VLB120 was the first
enzyme of this class, which was examined in detail for biocatalysis over more than a
decade. The two-component enzyme, for which homologues have been described for
different Pseudomonas strains [92, 218–220], consists of a FADH2-dependent oxyge-
nase (StyA), which transforms a broad range of styrenes into the corresponding (S)-
epoxides (Table 31.3) via reductive activation of molecular oxygen, and a reductase
1292
component (StyB) catalyzing the transfer hydrogenation between reduced nicotin-

amide cofactors and oxidized FAD [91, 221]. The highest productivities (9:1 g l1 aq h
1
1
on average over 8 h) and concentrations of toxic (S)-styrene oxide (72 g lorg ) have been
achieved with growing recombinant E. coli applied in an aqueous–organic two-liquid
phase system (phase ratio: 1 : 1) after extensive biocatalyst and process engineer-
ing [89, 213, 221, 222]. Scalability and economic as well as ecological feasibility have
also been evaluated [90, 222]. In this process, the product (S)-styrene oxide was
shown to cause acetic acid formation, membrane permeabilization, and cell lysis and
thus to become inhibiting (toxic) [89]. As a response to styrene oxide accumulation,
the specific activity of the recombinant cells as well as specific CO2 evolution, O2
uptake, and glucose uptake rates decreased. Since CO2 evolution during aerobic
glucose catabolism is directly proportional to the generation of reduction equivalents
in the form of NAD(P)H and FADH2 [223], the simultaneous decrease of glucose
uptake rates, specific CO2 evolution rates, and specific StyAB activities suggested that
decreasing cofactor regeneration rates may cause reduced biotransformation per-
formance. Limitations arising from mass transfer and intrinsic biocatalyst activity
could be excluded. As possible reasons for the reduced cofactor regeneration rates as
a consequence of product toxicity, metabolite and cofactor loss over permeabilized
membranes, enzyme deactivation, and regulatory phenomena have been dis-
cussed [89]. In addition, the 50% lower specific activity of growing cells as compared
to resting cells might be explained by NADH shortage in growing cells [224]. Such
shortage was also discussed for resting cells, which showed a correlation of
biocatalyst activity and glucose uptake rate when different central carbon metabolism
mutants of E. coli were tested [225]. An interesting approach to overcome toxicity and
cofactor regeneration problems involves the application of solvent tolerant Pseudo-
monas strains, which showed a high capacity for cofactor regeneration [226, 227]. For
example, the solvent tolerant styrene isomerase knock out strain Pseudomonas sp.
strain VLB120DC did not become limited by product toxicity or cofactor limitation,
but rather showed genetic variability. The same strain was used in its natural
immobilized form, that is, a biofilm, opening up a novel concept for productive
and stable biocatalysis. This concept has already been introduced for the continuous
production of epoxypropane (Section 31.3.1). Apart from application in continuous
bioprocessing, biofilms are generally much more robust against toxic substances, for
example, antibiotics and organic solvents, than their planktonic counterparts. By
utilizing Pseudomonas sp. strain VLB120DC grown in the biofilm mode it was
possible to achieve a constant productivity of 70 g styrene oxide l1
aq d
1
over a couple
of weeks [228].
In addition to the promising whole-cell approaches, also the cell-free application
of StyAB was evaluated. Using catalytic amounts of NAD þ together with the
formate dehydrogenase regeneration system, Hofstetter et al. performed gram-scale
epoxidation of various styrenes at excellent enantioselectivities and rates [231].
The rather complicated electron-transport chain delivering reduction equivalents
can be drastically simplified by using either a transition metal complex transferring
electrons from formate to the oxidized flavin cofactor [230] or, even more so, by direct
Table 31.4 Comparison of the catalytic performance of the different StyA regeneration approaches.
StyA Mediator E.e. (%) Reference
turnover total turnover TTN

frequency number (TTN)
(min1)
(a) 7.1 2870 90 (NAD) 99.7 [231]

(b) 12.1 701 12 (Rh) 99.4 [230]
(c) 8.4 42 0.1 (FAD) 99.8 [232]
2.3 1630 700 99.5 [233]
cathodic flavin reduction [232] (Table 31.4). Challenges related to the latter approach
include the undesired direct cathodic reduction of molecular oxygen, which results in
the formation of reactive oxygen species that impair biocatalyst stability. Using an
optimized electrochemical cell, Ruinatscha et al. could significantly reduce this
undesired side-reaction and thereby improve the productivity and robustness of the
electroenzymatic process [233]. Further significant improvements are expected in the
near future.
1294
The heme-containing monooxygenase P450cam from Pseudomonas putida also has

received some attention as styrene epoxidation catalyst. Especially, the Y96F mutant
showed significantly improved activity and selectivity for styrene epoxidation com-
pared to the wild-type enzyme [234]. To avoid the addition of the costly nicotinamide
cofactor, Vilker and coworkers chose an electrochemical strategy wherein the natural
electron donor for P450cam, putidaredoxin, was regenerated by direct cathodic
reduction. Molecular oxygen required for the monooxygenase reaction was generated
by anodic oxidation of water. This also proved to be necessary to suppress the
undesired reoxidation of putidaredoxin (Figure 31.8). Quite promising turnover
frequencies for the monooxygenase component of 10 min1 were observed. How-
ever, putidaredoxin had to be applied in huge molar surpluses, probably to com-
pensate for the oxidative uncoupling reaction.
Instead of performing reductive activation of molecular oxygen, heme-dependent
enzymes are also capable of using hydrogen peroxide to regenerate the catalytically
active oxyferryl heme-species (see also Section 31.2.5). This is particularly true for
heme-dependent peroxidases (vide infra); but also P450 monooxygenases such as
P450 BM-3 from Bacillus subtilis [186, 235] or P450s from thermophilic organ-
isms [236, 237] have been evaluated as styrene epoxidation catalysts utilizing the
peroxide shunt pathway. The major challenge encountered here is the poor stability of
the heme group in the presence of hydrogen peroxide. Thus, oxidative degradation
severely limits enzyme stability and thereby the applicability of this approach. This
may partially be overcome by means of protein engineering [84, 186].
Another class of heme-containing proteins, which has received some attention as
styrene epoxidation catalysts, is natural O2-transporting proteins such as hemoglobin
and myoglobin. Especially, the group around Rusling has been very active in
designing enzyme-modified cathodes for bioelectrocatalytic applications such as
styrene epoxidation [238–240]. Direct reduction of the electrode-absorbed enzymes
followed by activation of molecular oxygen thereby competes with direct regeneration
via the hydrogen peroxide shunt pathway by hydrogen peroxide generated at the
electrode. With this approach, only analytical transformations have been reported so
far, but exciting future developments may be expected.
Figure 31.8 Cathodic regeneration of reduced putidaredoxin (PDx) to promote P450cam-catalyzed

styrene epoxidation.
Chloroperoxidase (CPO) from the fungus Caldariomyces fumago was also reported
as selective epoxidation catalyst for the epoxidation of styrenes (Table 31.5).
Despite its high potential, CPO is not yet established in preparative organic
chemistry as a chiral catalyst as the effective recombinant expression of this highly
glycosylated protein is difficult [243–245].
The major limitation of CPO en route to practical application is its poor stability in
the presence of micromolar concentrations of H2O2, which results in the irreversible
inactivation of the prosthetic heme-group [100]. Therefore, the total turnover number
of CPO is usually 2–3 orders of magnitude too low to be economically attractive. To
overcome this limitation, various approaches such as the use of less reactive
hydroperoxides, the addition of antioxidants, or reduction of the biotransformation
rate are quite common [246].
Alternatively, a range of enzymatic, electrochemical, photochemical, and transi-
tion metal-based approaches have been evaluated for the in situ generation of H2O2
from dissolved O2 (Table 31.6).
31.3.3
Terpenes
The specific oxyfunctionalization of terpenes and steroids, which are secondary

metabolites with a high biological activity, is of significant interest for the pharma-
ceutical as well as for the flavor and fragrance industries. Interestingly, it appears that
terminal olefins (and only these) are epoxidized almost exclusively by bacteria. On the
other hand, more substituted double bonds are often preferentially oxidized by
higher organisms like fungi. The product is often the corresponding vicinal diol
arising from further metabolism (hydrolysis) of the primarily formed epoxides (see
Section 31.4 below). Numerous publications describe microbial and enzymatic
transformations of various terpenes [253–256].
One of the first examples of such a transformation was described by Marumo and
coworkers (Scheme 31.5) [257]. Their investigations, aimed at preparing optically
active insect juvenile hormone, showed that the monoterpene methylgeranate was
metabolized by the fungus Colletotrichum nicotianae, leading to 19.6% of (S)-
()-methyl-6,7-epoxygeranate and to 15.6% of (R)-( þ )-methyl-6,7-dihydroxygera-
nate after 9 h of incubation. Longer incubation times (24 h) produced only the
optically pure glycol with an isolated yield as high as 85%, showing that the first
epoxidation step had to be stereospecific. Unfortunately, this analytical study was not
pursued on a preparative scale, and no accurate results concerning the stereochem-
ical and kinetic aspects of these interesting biotransformations have been described.
A similar microbial oxidation of the isoprene double bond has been studied by
Veschambre and coworkers starting from linalool [258]. Thus, Streptomyces albus, a
strain that synthesized nigericine, transforms each enantiomer of linalool, as well as
the racemic compound, into a mixture (10–20% yield) of two diastereoisomeric
linalool oxides. In this case, the epoxide formed primarily is trapped by an intra-
molecular cyclization. Based on the reported product ratios, it can be estimated that
the e.e. of the formed epoxide was about 35%. Further work using several other
1296
Table 31.5 Epoxidation of various styrenes using chloroperoxidase (CPO) as biocatalyst.
Substrate Product Conversion E.e. TTN Reference

(%) value (%) (CPO)a)
O 67 96 N.d. [190]
O 85 97 N.d. [190]
O 40 49 N.d. [188]
About 30% aldehyde

and acid
No epoxide formed 0 — N.d. [188]
90 25 N.d. [188]
OH
OH
About 10% aldehyde

and acid
O 89 49 900 [192]
O 55 89 440 [192]
1 81 26 [192]
O
O 90 30 1000 [241]
Br
95 N.d. N.d. [242]
OH
Substrate Product Conversion E.e. TTN Reference

(%) value (%) (CPO)a)
Br 90 N.d. N.d. [242]

OH
Br 92 N.d. N.d. [242]

OH
microorganisms showed that Beauveria sulfurescens gave similar yields (15–20%

analytical) of an equimolar mixture of linalool oxides [259]. More specific linalool
epoxidation was reported by Abraham and coworkers [260] who showed that
()-linalool is processed by Diplodia gossypina exclusively to a mixture of (3S,6R)-
linalool oxide and the corresponding trans-tetrahydrofuran and trans-tetrahydro-
pyran. These were proposed to be formed by intramolecular cyclization of the
intermediate 3(S)-epoxide. Comparable results were obtained from linalool using
the strains Streptomyces cinnamonensis [261] and Aspergillus niger [262]. Similar
transformations were observed starting from 2-methyl-2-heptene-6-one [263] and
from sulcatol as shown in Scheme 31.2 [182]. With the aim of producing isomers of
the fragrance compounds lilac aldehyde and lilac alcohol, 19 fungal strains have been
Table 31.6 Comparison of various in situ H2O2 generation methods to promote peroxidase (CPO)-
catalyzed oxyfunctionalizations.
Regeneration catalysts Source of reducing TTN TTN (regeneration Reference

equivalents (CPO) catalyst)a)
GluOx Glucose 125 000 N.d. [96]

Cathode Cathode 58 900 N.d. [247–249]
Pd H2 6.500 13 [250]
[Cp Rh(bpy)(H2O)]2 þ /flavin Formate 80 (cytC) 8 [251]
Flavin/hn Formate/EDTA 22 400 1250 [252]

1298
19.6%
O O
C. nicotianae O
9h O
O OH
+ 15.6%
methyl geranate O
OH
S. albus
O
HO O HO OH
linalool (all steroisomers)
O
D. glossypina OH
A. niger HO
HO O HO
(-)-linalool (3S,6R)-linalool oxide

O
Scheme 31.5 Examples of microbial epoxidations of olefinic monoterpenes.
screened for linalool oxyfunctionalization products [264]. Indeed, A. niger and Botrytis
cinerea strains were found to produce low amounts of these fragrance compounds via
a postulated 8-hydroxylinalool intermediate. However, furanoid and pyranoid linal-
ool oxides remained the main products. With respect to their synthesis, Corynespora
cassiicola DSM was found to be the most efficient and selective biocatalyst with a
linalool oxide productivity of 120 mg l1 day1 and a conversion close to 100%.
Interestingly, human CYP2D6 was also found to catalyze pyranoid and furanoid
linalool oxide formation from ()-linalool [265].
The formation of stable epoxides (no further cyclization or rapid hydrolysis) has
been reported for the cyclic monoterpenes a-pinene and limonene. Xanthobacter sp.
C20 grown on cyclohexane was shown to epoxidize (R)-limonene at the ring (8–9
position) [266]. Apparently, the monooxygenase catalyzing cyclohexane hydroxylation
(a cytochrome P450 enzyme initiating cyclohexane catabolism) does not oxyfunctio-
nalize the cyclohexene ring of (R)-limonene, but catalyzes the epoxidation of the
propenyl-double bond instead, giving rise to (4R,8R)-limonene-8,9-oxide as the only
reaction product. (S)-Limonene was converted into a (78 : 22) mixture of (4S,8R)- and
(4R,8S)-limonene-8,9-oxide (Scheme 31.6). The maximal product concentration
reached (0.8 g l1) was limited by product inhibition.
()-a-Pinene oxyfunctionalization by means of recombinant E. coli containing a
quintuple mutant of P450 BM-3 led to a mixture of three products, a-pinene oxide,
trans-verbenol, and myrtenol, with a-pinene oxide being the predominant one
(Scheme 31.7) [267]. Efficient NADPH regeneration was guaranteed by co-expression
Xanthobacter sp.
(R)-Limonene (4R,8R)-Limonene-8,9-oxide
Xanthobacter sp.
+
O O
(S)-Limonene (4S,8R)-Limonene-8,9-oxide (4S,8S)-Limonene-8,9-oxide
Scheme 31.6 Epoxidation of the two limonene enantiomers by Xanthobacter sp. C20.
of genes for glucose dehydrogenase from Bacillus megaterium and a glucose uptake
facilitator from Zymomonas mobilis. Thereby, initial specific biocatalyst activities of 5.3
and 8:6 U g1 CDW have been achieved for a-pinene oxide and total product formation,
respectively. This approach allowed the formation of 20 mg g1 CDW of a-pinene oxide
within 1.5 h, after which biocatalyst activity was lost, most probably due to enzyme
deactivation. Similarly, in enzyme assays with P450cam from Pseudomonas putida
and mutants thereof, ( þ )-a-pinene and (S)-limonene were monooxygenated to a
OH
O
P450 BM-3 + + +
P450cam
OH O
α-pinene α-pinene oxide verbenol myrtenol verbenone
O
HO
P450cam
+ +
OH
(S)-limonene (-)-cis-limonene-1,2-oxide (-)-trans-isopiperitenol (-)-trans-carveol
Scheme 31.7 Monooxygenation of the monoterpenes a-pinene and (S)-limonene by P450 BM-3
and P450cam, respectively.
1300
mixture of products including oxides (Scheme 31.7) [268]. In this case, however,
( þ )-cis-verbenol and ()-trans-isopiperitenol were the respective major products with
( þ )-a-pinene oxide and ()-cis-limonene-1,2-oxide being respective by-products.
Further by-products were ( þ )-myrthenol and ( þ )-verbenone for ( þ )-a-pinene
oxyfunctionalization and ()-trans-carveol for (S)-limonene monooxygenation. Reac-
tion selectivities depended on the rationally introduced active site mutations, which
allowed a significant increase of catalytic rates and coupling efficiencies.
Limonene epoxidation was also achieved using plant cell cultures, for example, of
Picea abies, which led to a product mixture with limonene-1,2-oxide as a major
product [269]. Furthermore, plant cell cultures of Caragana chamlagu (Leguminosae)
were shown to convert b-ionone into 5,6-epoxy-b-ionone as the sole product with 87%
yield and a 60 : 40 ratio of the 5a- and 5b-epoxy-enantiomers [270]. The same cells
converted ()-a-ionone into ()-4,5a-epoxy-a-ionone (20% yield), ()-5a-hydroxy-
a-ionone (16% yield), and ()-3-oxo-a-ionone (50% yield) (Scheme 31.8).
Human P450 monooxygenases also have been shown to catalyze epoxidations of
monoterpenes. D3-Carene was converted into trans-D3-carene oxide by CYP1A2 [271].
The same product was formed via the lipase-mediated epoxidation (Scheme 31.4)
of D3-carene with 99% selectivity for the trans-product [272]. The lipase approach has
also been applied for a-pinene epoxidation [273, 274]. In both cases Candida antartica
lipase (CAL) B was used with octanoic acid as perhydrolysis substrate and toluene as a
solvent, giving good results.
Various cyclic sesquiterpenes have also been studied to explore the possibility of
achieving their microbiological oxyfunctionalization. Very often, these gave rise to
O O
O
C. chamlagu
β-ionone 5,6-epoxy-β-ionone
O O O
+
C. chamlagu
HO
O
(±)-3α-ionone (±)-4,5α-epoxy-α-ionone (±)-3α-hydroxy-α-ionone
+
O
(±)-3-oxo-α-ionone
Scheme 31.8 Oxyfunctionalization of ionones by plant cell cultures of Caragana chamlagu [270].
epoxidation products when at least one double bond was present in the substrate.
Germacrone, which is thought to be the precursor of various bicarbocyclic sesqui-
terpenoids, was transformed by the fungus Cunninghamella blakesleena. This led
primarily to regio- and stereoselective epoxidation of one of the intracyclic double
bonds of this prochiral triene, thus affording two epoxides (Scheme 31.9). The third
product isolated from this experiment was due to subsequent epoxidation of the
remaining intracyclic double bond. Interestingly, the exocyclic olefinic bond conju-
gated to the carbonyl function appeared to be resistant to oxidation [275].
O O O O
Cunninghamella O O
blakesleena
O O
O
Cunninghamella
blakesleena
germacrone
Scheme 31.9 Epoxidation of germacrone stereoisomers by the fungus Cunninghamella

blakesleena.
Valencene, another olefinic sesquiterpene, has been studied in the same context
using microorganisms isolated from soil. It was observed that these biotransforma-
tions led in reasonable yields to a mixture of three main metabolites, including an
epoxide and nootkatone, an interesting flavoring compound [276]. A mixture of
valencene oxidation products was also obtained when mutants of P450cam and P450
BM-3 were tested [277]. Whereas P450cam and its mutants converted ( þ )-valencene
into ( þ )-trans-nootkatol, ( þ )-nootkatone, and ( þ )-trans-nootkaton-9-ol as the main
products, P450 BM-3 and its mutants additionally formed the two epoxides cis-
( þ )-valencene-1,10-oxide and ( þ )-nootkatone-(13S),14-oxide in substantial
amounts (Scheme 31.10). As in the case of limonene oxidation, the rationally
introduced active site mutations heavily influenced reaction specificities.
The active site of the same enzyme also was engineered to accept amorpha-4,11-
diene and epoxidize it specifically to artemisinin-(11S),12-epoxide [278]. The latter
was produced at levels up to 250 mg l1 with the P450 BM-3 mutant genes expressed
in recombinant E. coli, which previously have been engineered to produce high levels
of amorpha-4,11-diene via an engineered terpenoid biosynthesis pathway. This
represents an important step towards a novel semi-biosynthetic route for the
production of the antimalaria drug artemisinin.
In general, it can be stated that the biocatalytic oxidation of terpenoids using whole
cells of bacteria and fungi but also of plants often involves the epoxidation of C¼C
double bonds, which then, however, often is followed or accompanied by further
reaction steps (e.g., oxidation or hydrolysis) finally leading to a product mixture.
1302
O
HO
P450 BM-3
+
mutants
(+)-valencene cis-(+)-valencene-1,10-oxide (+)-nootkatol

trans and cis
O O
+ O
+
S
(+)-nootkatone (+)-nootkatone-13S,14-oxide
OH
O
(+)-trans-nootkaton-9-ol
H H
P450 BM-3
mutants
H H O
S
amorpha-4,11-diene artemisinin-11S,12-epoxide
Scheme 31.10 Oxidation products formed by P450 BM-3 mutants from the sesquiterpenes
( þ )-valencene [277] and amorpha-4,11-diene [278].
These side reactions can be due to both the enzyme responsible for epoxidation
(low specificity) and the presence of other enzymes active on the substrate or product.
Epoxide hydrolysis leading to the formation of diols is very typical and can also occur
spontaneously in aqueous media depending on medium pH (see below).
31.4
Dihydroxylation of C¼C Double Bonds
Mononuclear non-heme iron oxygenases are well known to catalyze dihydroxylations

(Section 31.2.2). Typical substrates, which undergo biocatalytic dioxygenation, are
aromatic compounds with dihydrodiols being the typical products. The oxidation of
aromatic compounds is the topic of Chapter 34 and thus is not discussed here. Direct
biocatalytic dioxygenation of non-aromatic C¼C double bonds is rarely reported.
31.4 Dihydroxylation of C¼C Double Bonds j1303
Typically, the formation of diols from C¼C double bonds runs via the intermediary
formation of epoxides followed by spontaneous or enzyme-catalyzed hydrolysis. For
example, many terpene dihydroxylations are reported to follow this reaction
scheme [254–256]. Direct dioxygenations are primarily reported for conjugated
alkenes, as discussed in the next section.
31.4.1
Aliphatic Olefins and Conjugated Alkenes
The dihydroxylation of aliphatic olefins is actually a tandem reaction and proceeds via
an enzymatically formed epoxide, which subsequently is hydrolyzed, yielding the
respective diol. One example is the conversion of 1-methylcyclohexene into 1-methyl-
1,2-dihydroxycyclohexane catalyzed by the above-discussed chloroperoxidase from
the fungus Caldariomyces fumago (CPO). The reported reaction was carried out in a
ternary water–organic solvent system consisting of a-pinene and t-butanol using
isolated CPO. The conversion was rather unspecific, leading to a mixture of four
compounds, of which 1-methyl-1,2-dihydroxycyclohexane was identified as the main
product [200]. Epoxide hydrolysis is discussed in detail in Chapter 9, and will thus not
be further reviewed here.
Non-conjugated monoalkenes appear to be very poor substrates for bacterial
dioxygenase-catalyzed dihydroxylations, with only a few literature reports avail-
able [279, 280], whereas the asymmetric 1,2-dihydroxylation of several acyclic but
conjugated alkene substrates has been described. In these examples, the alkene
double bond was conjugated to an aryl group. One of the best known enzymes to
catalyze the direct dihydroxylation of aromatic as well as conjugated aliphatic
substrates is naphthalene dioxygenase (NDO), an enzyme commonly found in
Pseudomonas species. As its main physiological function, this enzyme dihydroxylates
naphthalene to 1,2-dihydroxynaphthalene, which is then converted into salicy-
late [281]. This enzyme belongs to the Riske-type dioxygenases, which have been
discussed in detail in Section 31.2.2. Apart from NDO [282], toluene dioxygenase
(TDO) is also well known for its ability to directly dihydroxylate various compounds
bearing conjugated alkene groups. Its substrate spectrum has been thoroughly
investigated and compared to NDO by Boyd and coworkers [283], using whole cells
of P. putida UV4 (a mutant strain containing TDO but no diol dehydrogenase activity)
and P. putida NCIMB8859 (a wild-type strain containing NDO and a diol dehydro-
genase). Besides the dihydroxylation of aromatic compounds, several types of meta-
substituted styrenes have been converted by NDO and TDO (Scheme 31.11).
While application of TDO yielded a mixture of (1S,2R)-arene-cis-dihydrodiols and
the corresponding alkene-1,2-diols, NDO exclusively formed the alkene diol
products with a conversion yield between 60% (for styrene to 1-phenyl-ethandiol)
and 12% (for meta-methylstyrene to the corresponding alkene diol).
In addition, the conversion of benzo-fused cyclic alkenes by NDO and TDO has
been studied, yielding the corresponding 1,2-diols of opposite absolute configuration
(Scheme 31.12) [282, 284–287]. Surprisingly, when methyl-substituted benzo-fused
cyclic alkenes were tested as substrates, P. putida UV4 only converted 2-methylin-
1304
R3 R3 R3
HO
OH
R2 R1 R2 R1 R2 R1
OH
TDO TDO or NDO
O2 O2
R OH R R
2a-g 1a-g 3a-g
TDO or NDO
O2
CH2OH
COOH
Scheme 31.11 Toluene and naphthalene R1 ¼ Me, (R, R2, R3) ¼ H; 1f R2 ¼ Me, (R, R1,
dioxygenase (TDO and NDO, respectively) R3) ¼ H; 1g R3 ¼ Me, (R–R2) ¼ H. At the
catalyzed dihydroxylation of variably substituted bottom, side products formed from 1e
styrenes. 1a R–R3 ¼ H; 1b R ¼ F, R1–3 ¼ H; 1c and 1g are shown. Scheme adapted from
R ¼ Cl, R1–3 ¼ H; 1d R ¼ Me, R1–3 ¼ H; 1e Boyd et al. [283].
OH OH
OH OH
(CH2) n (CH2)n
NDO TDO
NDO (CH2) n TDO
(CH2) n-1 (CH2) n-1
HO HO
Scheme 31.12 Conversion of benzo-fused cyclic alkenes (n ¼ 1–3) by TDO and NDO into the
corresponding cis-diols and mono-alcohols. Adapted from Boyd et al. [283].
dene into the corresponding cis-diol at a yield of 26%, with hydroxylation at the
benzylic carbon giving 2-methylindenol and 2-methylindanone as by-products [283].
Finally, the biotransformation of cyclic dienes and trienes has been reported using
TDO and NDO as biocatalysts (Scheme 31.13) [283, 288]. Cyclopentadiene was
converted into the corresponding cis-diol with a relatively low enantiomeric excess for
the (1R,2S)-enantiomer (20% and 40% e.e. for TDO and NDO, respectively), while all
other tested monocyclic dienes yielded the corresponding enantiopure 1,2-diols.
OH
TDO / NDO
X O2 X OH
Scheme 31.13 Conversion of cyclic dienes and trienes into the corresponding diols using whole
cells containing NDO or TDO. X ¼ CH2, (CH2)2, (CH2)3, (CH2)4, CH¼CHCH2, C¼CMe2, C¼CEt2,
C¼C(CH2)4, or C¼C(CH2)5. Adapted from Boyd et al. [283].
Trienes are also converted by both enzymes, but, as observed for the styrene
derivatives, the conversion by NDO is much more specific than that by TDO. For
example, NDO-catalyzed conversion of cycloheptatriene afforded the enantiopure
(1R,2S)-diene, while TDO gave an inseparable mixture of chiral diol (29% yield) and
achiral diol (15% yield), with the dihydroxylation of the middle double bond
(conjugated on both sides) in the latter case.
31.4.2
Terpenes
As described above, diols are major products of microbial terpene oxidation, which
typically result from sequential epoxidation and spontaneous or enzymatic hydrolysis
depending on conditions and catalyst (Section 31.3.3). Thus, the actual C¼C double
bond oxidation again is the epoxidation. Many microbial terpene dihydroxylations
have been reported [255, 256], of which some prominent examples are given here.
As in the case of terpene epoxidation, limonene was also studied as a model
substrate for microbial terpene dihydroxylation. A broad screen of 800 microorgan-
isms was performed using both (S)- and (R)-limonene as a starting substrate and
some other terpenes that were tested with the best suited strains (Scheme 31.14) [289].
The most interesting results were observed with Diplodia gossypina (ATCC 10936),
which afforded 380 mg (1R,2R,4S)-limonene-1,2-diol from 1 g of (S)-limonene. (R)-
Limonene was shown to afford (1S,2S,4R)-limonene-1,2-diol. Because of the rele-
vance of such products in flavor chemistry, the preparative-scale transformation
(100 l bioreactor filled with 70 l medium) of this enantiomer by the fungus Diplodia
gossypina has been undertaken: 1300 g were transformed to yield, within 96 h, 900 g of
the (1S,2S) diol showing high optical purity [290]. Similarly, Corynespora cassiicola
(DSM 62474) was described to yield 1.1 g of (1R,2R)-a-terpinene-1,2-diol from 1.8 g
of a-terpinene [289]. Furthermore, these two strains have been used to convert
c-terpinene, terpinolene, and (R)-a-phellandrene into the corresponding diols.
1306
OH
OH
Diploida
30%
gossypina
(S)-limonene (1R,2R,4S)-limonene-1,2-diol
OH OH
OH OH
Diploida
55% +
gossypina 0.6%
(R)-limonene (1S,2S,4R)-limonene-1,2-diol (1R,2R,4R)-limonene-1,2-diol
OH OH OH
OH O OH
Corynespora
49% + 2% + 1%
cassiicola
(1R)-α-terpinene-1-ol-2-one
α-terpinene
(1R,2R)-α-terpinene-1,2-diol (1R,2S)-α-terpinene-1,2-diol
Scheme 31.14 Stereospecific dihydroxylation of limonene enantiomers and a-terpinene.
Interestingly, substrates are converted rapidly with only low amounts of side
products. It is also notable that the obtained diols are almost exclusively trans
configurated (optical purities have not been indicated). This suggested that these
trans-diols were formed via intermediate epoxides being cleaved enzymatically.
Surprisingly, both these microorganisms were shown not to attack 3,3,5,5-tetra-
methyllimonene [291]. However, geranylacetone, nerylacetone, trans-nerolidol, cis-
nerolidol, farnesol, and 2,5-dimethyl-1,3-hexadiene were transformed by various
strains into the corresponding glycols in yields of up to 70% [292, 293] and interesting
optical purities of up to 98%. Rhodococcus erythropolis DCL14 was also shown to be
able to metabolize (R)-limonene, initiated by a double bond epoxidation, forming
(1S,2S,4R)-limonene-1,2-diol, (1S,4R)-1-hydroxy-2-oxolimonene, and (3R)-3-isopro-
penyl-6-oxoheptanoate. The opposite enantiomers (1R,2R,4S)-limonene-1,2-diol,
(1R,4S)-limonene-1-ol-2-one, and (3S)-3-isopropenyl-6-oxoheptanoate accumulated
when (S)-limonene was used as substrate, showing that the enzymes from this
pathway are not stereoselective [294].
The bio-oxygenation of geraniol derivatives is another interesting example of
terpene dihydroxylation (Scheme 31.15). The N-phenylcarbamate of geraniol was
transformed by the fungus Aspergillus niger into the corresponding 6,7-diol of (6S)
absolute configuration (95% e.e.), yielding 49% isolated product [295]. Thereby, 1 g
substrate treated for 36 h in 1 l of culture broth afforded 550 mg pure diol. Unique
OR A. niger OR A. niger OR
OH pH 2 pH 7 OH
50% 50%
OH OH
R = CONHPh
(6S) 6,7-diol (6R) 6,7-diol
ee =95% geraniol N-phenylcarbamate ee =95%
OH pH 2 pH 6 OH
60% 43%
OH OH
R = coumarin
(-)-marmine (+)-marmine
7-geranyloxycoumarin
OH pH 2 pH 6 OH
85% 60%
OH OH
R = CONHPh
(3R,6S) 6,7-diol (3R,6R) 6,7-diol
ee =90% (3R)-citronellyl N-phenylcarbamate ee =92%
Scheme 31.15 Stereoselective dihydroxylation of geranyl derivatives depending on pH.
stereochemical control could be simply achieved by modulating the pH of the

medium. If the pH was increased from 2 to 6–7, the diol of opposite (R) absolute
configuration was isolated in similar yields and with an e.e. again as high as 95%. A.
niger, being active across the pH 2–7 range, primarily formed the (6S)-epoxide, which
was hydrolyzed either following the classical acid-catalysis mechanism (in acidic
medium) to afford the (6S)-diol or, at pH 6, enzymatically to the (6R)-diol by hydrolytic
attack on the less substituted oxirane carbon atom [296]. These mechanisms were
confirmed by 95% 18 O incorporation on C6 and 5% on C7 at pH 2 and vice versa at pH
7. These results also prove the monooxygenation reaction, which leads to epoxide
formation followed by different hydrolysis, depending on the pH of the medium. The
resulting diols can be used as chirons for the synthesis of various natural or non-
natural products. For instance, they can be cyclized to the optically pure linalool
oxides [297] or the corresponding tetrahydropyranols [298].
1308
Bio-oxygenation of similar compounds, that is, 7-geranyloxycoumarin, citronellyl

N-phenylcarbamate (Scheme 31.15), and sulcatol N-phenylcarbamate, was also
studied [182, 299, 300]. The reaction was shown to be operative in all these cases,
leading, for instance, to both enantiomers of marmin (a member of the umbellifer-
one family). Similar results were obtained with both commercially available citro-
nellol enantiomers, leading to the corresponding diols showing e.e.s as high as 90
and 92%.
Dihydroxylation of larger terpenes such as sesquiterpenes has also been reported.
The oxidation of ( þ )-valencene and ( þ )-nootkatone was carried out using the
fungi A. niger, Fusarium culmorum, and Botryosphaeria dothidea [301]. ( þ )-Nootka-
tone was oxidized by A. niger and B. dothidea into stereoisomeric mixtures of
( þ )-nootkatone-11,12-diols with 12-hydroxy-11,12-dihydronootkatone and 7a-
hydroxy-nootkatone as respective by-products, while F. culmorum oxidized the same
substrate primarily to ( þ )-nootkatone-11(R),12-diol and 9-b-hydroxy-nootkatone
(Table 31.7). The biotransformation of ( þ )-valencene by A. niger also afforded both
( þ )-nootkatone-11,12-diol enantiomers with a slight preference for the 11(R)-diol
and was accompanied by the formation of several by-products (Table 31.7). These
product formation patterns were assigned to multiple oxidations including 11,12-
epoxidation followed by epoxide hydrolysis.
To conclude, fungal strains have been shown to be of special interest not only for
terpene dihydroxylation but also for terpene oxidation in general. However, microbial
and P450-catalyzed oxidation of terpenes typically involves the formation of a product
mixture, which necessitates careful work up to separate the products.
31.5
Oxidative Cleavage of Double Bonds
Different enzymatic strategies have been reported for C¼C double bond cleavage,
including mononuclear non-heme iron oxygenase catalysis via a monooxygenase- or
dioxygenase mechanism [66] (Figure 31.3), peroxidase catalysis [302], co-oxidation by
an enzymatically generated free radical species (e.g., lipoxygenase catalyzed lino-
leoylperoxyl radical formation) [303–305], and a novel radical-alkene cleaving mech-
anism exhibited by a metal-free active site [306].
Except for the last approach, the strategies primarily aimed at the cleavage of
tetraterpenoic carotenoids to yield commercially highly interesting bioactives and
flavor compounds such as b-ionone [307]. As an example, Scheme 31.16 shows the
cleavage of b,b-carotene by a carotenoid cleavage dioxygenase from Arabidopsis
thaliana at the 9,10- and 90 ,100 -double bonds giving b-ionone and apo-10,100 -car-
otendial as the oxygenated products [60].
Such dioxygenases from plants or cyanobacteria have been reported to cleave
various carotenoids, including, beside b,b-carotene, f-carotene, lycopene, torulene,
a-carotene, zeaxanthin, lutein, violaxanthin, and neoxanthin (Figure 31.9) [60, 62,
308–311]. Apocarotenoids such as b-apo-100 -carotenal are also accepted as substrates.
Depending on the enzyme and the substrate, different double bonds are cleaved,
31.5 Oxidative Cleavage of Double Bonds j1309
Table 31.7 Isolated yields (%) of different products resulting from ( þ )-nootkatone and
( þ )-valencene oxidation by different fungal species [301].
Product formed O
(+)-valencene
(+)-nootkatone
Asper- Fusarium Botryo- A. niger

gillus culmorum sphaeria
niger dothidea
OH 51.5 47.2 54.2 13.5

OH (S : R (S : R (S : R (S : R
¼ 3 : 2) ¼ 1 : 20) ¼ 3 : 2) ¼ 1 : 3)
(+)-nootkatone-11,12-diol
O 10.6 1
OH
12-hydroxy-11,12-dihydro-
nootkatone
OH 14.9
9b-hydroxy-nootkatone
O 20.9
OH
7a-hydroxy-nootkatone
1.1
OH (Continued )
OH
1310
Product formed O
(+)-valencene
(+)-nootkatone
Asper- Fusarium Botryo- A. niger

gillus culmorum sphaeria
niger dothidea
1.5
O OH
R OH
HO OH
OH
0.7
O OH
H OH
10′
9′
9
10
2O2
carotenoid cleavage dioxygenase
+ O
O O
2
β-ionone apo-10,10′-carotendial
Scheme 31.16 b,b-Carotene 9,10(90 ,100 )-cleavage by a carotenoid cleavage dioxygenase, for
example, from Arabidopsis thaliana.
31.5 Oxidative Cleavage of Double Bonds j1311
ζ-carotene
lycopene
torulene
α-carotene
OH
zeaxynthin
HO
OH
lutein
HO
OH
O
O
violaxanthin
HO OH
O
neoxanthin
OH
HO
β-apo-10′-carotenal
Figure 31.9 Typical substrates accepted by carotenoid cleavage oxygenases.
leading to the corresponding aldehydes and/or ketones. The most prominent

cleavage sites are the 15,150 -, 11,12-, 110 ,120 -, 9,10-, 90 ,100 -, 7,8-, 70 ,80 -, 5,6-, and
50 ,60 -double bonds, leading to various products, of which the most important
norisoprenoids, typically potent flavor compounds, are shown in Figure 31.10.
In vitro dioxygenase catalysis has been performed in aqueous micellar reaction
systems containing water-miscible solvents. Enzyme solubilization by means of
surfactants has been reported to play a crucial role [312].
1312
O O O
geranylacetone pseudoionone 6-methyl-5-heptene-2-one
O O O
O
retinal (vitamin A) β-ionone α-ionone β-cycloisocitral
O
O
O OH
HO O HO
HO
xanthoxin 3-hydroxy-β-ionone grasshopper ketone
Figure 31.10 Typical products arising from carotenoid cleavage by oxygenases.
The same carotenoids have been reported to be cleaved by secreted fungal

peroxidases [313, 314]. Cleavage of b,b-carotene to b-ionone and b-apo-100 -carotenal
was proposed to be initiated by abstraction of a hydrogen atom from the allylic methyl
group, resulting in a resonance-stabilized carbon radical [302]. Hydroperoxides
would then be formed intermediately through reaction with oxygen, and subsequent
Hock cleavage would yield two carbonyl compounds. Such peroxidase-mediated
catalysis led to further oxidation products such as epoxides and dihydroactinidiolide
(Figure 31.11). Such further oxidation also was observed with the co-oxidation
approach, in which radical species generated from a second substrate by, for example,
lipoxygenase or xanthine oxidase catalysis, oxidize carotenoids, finally leading to their
cleavage [303–305].
Another very interesting C¼C double bond cleaving enzyme activity was reported
by Kroutil and coworkers, who found that whole cells or cell-free extracts of the wood-
degrading fungus Trametes hirsuta FCC 047 cleave C¼C double bonds adjacent to an
aromatic ring by use of molecular oxygen as the sole oxidant to yield the correspond-
ing carbonyl compounds [316–318]. The as-yet unidentified enzyme was shown to
O O
O
β,β-carotene
β-ionone β-ionone-5,6-epoxide
O
O
O
O OH
2-hydroxy-2,6,6-trimethyl-
dihydroactinidiolide β-cyclocitral
cylclohexanone
Figure 31.11 Products formed upon b,b-carotene cleavage catalyzed by peroxidases isolated from
the secretome of Marasmius scorodonius [314, 315].
31.6 Triple Bond Oxyfunctionalization j1313
Table 31.8 Phenylalkenes cleaved by Trametes hirsuta FCC 047 in buffer (pH 6) at 2 bar oxygen
pressure. Adapted from Lara et al. [316]..
R2 R2
R3 cell-free extract R3
T. hirsuta FCC 047 O
O
+
buffer, O2 (2 bar)
R4 R4
R1 R1
R1 R2 R3 R4 Conversion Chemoselectivitya)
(%) (%)
p-OMe H CH3 H >99 >99

p-NH2 H H H 3 >99
p-OMe H H H 55 90
H H CH3 CH3 13 83
H H H H 22 >99
p-CH3 H H H 33 94
p-tBu H H H 23 92
m-NO2 H H H 5 67
m-CH3 H H H 20 92
m-Cl H H H 49 80
o-Cl H H H 77 85
o-CH3 H H H 20 91
H CH3 H H 27 >99
H H C3H7 H 27 64
H H CN H 1 >99
a) The chemoselectivity is defined as the ratio of formed aldehyde/ketone to all compounds

formed.
follow neither the classical dioxygenase mechanism nor a monooxygenase mecha-

nism [306]. Instead, an alternative radical-alkene cleaving mechanism was proposed
for the obviously metal-free active site, whereby two oxygen atoms of two different
oxygen molecules are incorporated. A broad range of phenylalkenes was shown to be
converted, typically with a high chemoselectivity (Table 31.8). Indene, 1,2-dihydro-
naphthalene, isosafrole, and 2-(prop-1-enyl)thiophene also were cleaved at the double
bond adjacent to the aromatic ring. For efficient preparative synthesis, high oxygen
pressures were required.
31.6
Triple Bond Oxyfunctionalization
Few reports are available on the oxyfunctionalization of triple bonds, including

catalysis by haloperoxidases [319, 320], cytochrome P450 monooxygenases [321, 322],
and a 2-oxoglutarate dependent mono-nuclear non-heme iron oxygenase [52, 323].
1314
Functionalization of the C:C triple bond by peroxidases occurs indirectly as the

actual oxidant is the enzymatically formed hypohalous acid, which oxidizes alkynes to
mono- and dihalogenated ketones (Scheme 31.17) [319, 320]. This reaction was
shown to proceed with chloroperoxidase or lactoperoxidase in the presence of
hydrogen peroxide and chloride or bromide ions.
O O X
R C CH + X + H2O2 R C C X + R C C X + H2O
H2 H
R = CH3, CH2CH3, C6H5 X = Cl, Br
Scheme 31.17 Haloperoxidase-mediated alkyne oxidation, leading to the formation of

halogenated ketones, with hypohalous acid as the oxidant. Depending on the substrate and the
halide, products can be halogenated at different positions.
P450 monooxygenases have been reported to directly catalyze C:C triple bond
oxyfunctionalization by two different mechanisms, of which one leads to mecha-
nism-based enzyme deactivation/inhibition via heme alkylation (Scheme 31.18) [321,
322]. Such mechanism-based enzyme inhibition during alkyne oxidation is ubiqui-
tous among P450 monooxygenases [324]. Cytochrome P450-catalyzed phenylacety-
lene oxidation resulted in a partitioning between phenylacetic acid formation and
heme alkylation in a ratio of 26: 1, whereby phenylacetylene is proposed to bind to the
active site in two orientations.
O
Heme
(inactive enzyme)
P450, O2
O
+ H2O
O
O
Scheme 31.18 Cytochrome P450-catalyzed oxidation of phenylacetylenes to phenylacetic acids

and parallel heme alkylation leading to enzyme inactivation.
The same type of alkyne oxyfunctionalization and enzyme inactivation has been
reported for 2-oxoglutarate dependent thymine hydroxylase [52]. With 5-ethynyluracil
as the substrate 5-(carboxyethyl)uracil was formed as a free product; also here, the
formation of a ketene intermediate was proposed, which then is trapped by water to
give the carboxy-product. In the case of the thymine hydroxylase, the relation between
free product formation and enzyme alkylation was as low as 3: 1 and adduct formation
was proposed to involve two oxidation steps with a phenylalanine residue in the active
site being modified [323].
However, no preparative application involving a biocatalytic triple-bond oxyfunc-
tionalization has been reported so far.
31.7
Summary and Outlook
This chapter has illustrated the very broad range of C–C multiple bond oxyfunctio-
nalizations that can be achieved using enzymes, of which most belong to the
oxidoreductase class with the exception of hydrolase-mediated epoxidation. Cata-
lyzed reactions include epoxidations, dihydroxylations, double bond cleavage, and
triple bond oxyfunctionalization. Since, depending on the enzyme(s) applied, many
substrates can lead to different oxidized products, the range of compounds that can
be generated is clearly enormous. Typically, these enzymatic reactions are highly
specific. However, especially with terpenoids as substrates, multiple products can be
formed, where side reactions can arise either from a low enzyme specificity or, if
whole cells are applied as catalyst, from unspecific intracellular enzyme activities.
Whereas biocatalytic oxyfunctionalizations provide a large tool box for organic
synthesis and also are of high interest for technical applications, the major hurdles
for industrial implementation derive from the complexity of these reactions. The
most practical way to use, for example, oxygenase-based biocatalysts for large volume
processes still is in whole-cell systems because of cofactor requirements and
problems with enzyme stability. The latter often arise from the enzyme-related
formation of reactive oxygen species (e.g., hydrogen peroxide, superoxide anion
radical, hydroxyl radical), for which living cells have developed efficient degradation
systems (e.g., catalase, superoxide dismutase). However, some oxygenases, such as
styrene monooxygenase and P450cam, can be isolated in sufficient quantities and
reconstituted for cell-free preparative scale biotransformations. This might be
particularly useful at a smaller scale for substrates that cannot penetrate cell walls
and are toxic to or unstable in the organism.
Substrates and products often are poorly water soluble and affect biocatalyst
stability due to toxic or deactivating effects. Several new technologies in genetics,
molecular biology, systems biology, and biochemical process engineering have
emerged and are applied to overcome such limitations on different levels: the
enzyme level, the cell level, and the reaction/process level. All these levels have to
be considered for the successful development of a technically and industrially
feasible process.
In conclusion, the application of biocatalysts for C–C multiple bond oxyfunctio-
nalizations is rapidly expanding in terms of practicality, substrate range, and
selectivity. A vast diversity of oxidoreductase genes is made available by genomics,
metagenomics, and mutagenesis. To guarantee adequate expression levels and
supply of reducing equivalents for O2 activation, suitable microbial hosts can be
engineered following the principles of metabolic engineering, systems biotechnol-
ogy, and synthetic biology. Combining these new technologies with innovative
approaches for biochemical process engineering, downstream processing, and
1316
process management, one can foresee a future where designer bio-oxidation cata-
lysts, tailored for a specific substrate, selectivity of reaction, high productivity, and
high stability, can be generated and implemented within short time spans.
References
1 Leak, D., Aikens, P., and Seyed- 18 Munro, A.W., Girvan, H.M., and
Mahmoudian, M. (1992) Trends McLean, K.J. (2007) Nat. Prod. Rep., 24,
Biotechnol., 10, 256–261. 585–609.
2 Leak, D.J., Sheldon, R.A., Woodley, J.M., 19 Bruijnincx, P.C.A., van Koten, G., and
and Adlercreutz, P. (2009) Biocatal. Gebbink, R.J.M.K. (2008) Chem. Soc. Rev.,
Biotransform., 27, 1–26. 37, 2716–2744.
3 B€uhler, B. and Schmid, A. (2004) 20 Ryle, M.J. and Hausinger, R.P. (2002)
J. Biotechnol., 113, 183–210. Curr. Opin. Chem. Biol., 6, 193–201.
4 Kulkarni, A.P. (2001) Cell. Mol. Life Sci., 21 Bugg, T.D. and Ramaswamy, S. (2008)
58, 1805–1825. Curr. Opin. Chem. Biol., 12, 134–140.
5 Archelas, A. and Furstoss, R. (1999) Top. 22 Solomon, E.I., Brunold, T.C., Davis, M.I.,
Curr. Chem., 200, 159–191. Kemsley, J.N., Lee, S.K., Lehnert, N.,
6 Carboni-Oerlemans, C., de Maria, P.D., Neese, F., Skulan, A.J., Yang, Y.S.,
Tuin, B., Bargeman, G., van der Meer, A., and Zhou, J. (2000) Chem. Rev., 100,
and van Gemert, R. (2006) J. Biotechnol., 235–350.
126, 140–151. 23 Wallar, B.J. and Lipscomb, J.D. (1996)
7 Schmid, A., Dordick, J.S., Hauer, B., Chem. Rev., 96, 2625–2658.
Kiener, A., Wubbolts, M., and Witholt, B. 24 Costas, M., Mehn, M.P., Jensen, M.P.,
(2001) Nature, 409, 258–268. and Que, L. (2004) Chem. Rev., 104,
8 Straathof, A.J., Panke, S., and Schmid, A. 939–986.
(2002) Curr. Opin. Biotechnol., 13, 25 Abu-Omar, M.M., Loaiza, A., and
548–556. Hontzeas, N. (2005) Chem. Rev., 105,
9 Julsing, M.K., Cornelissen, S., B€uhler, B., 2227–2252.
and Schmid, A. (2008) Curr. Opin. Chem. 26 Brazeau, B.J. and Lipscomb, J.D. (2003)
Biol., 12, 177–186. Biochemistry, 42, 5618–5631.
10 Urlacher, V.B. and Schmid, R.D. (2006) 27 Sazinsky, M.H. and Lippard, S.J. (2006)
Curr. Opin. Chem. Biol., 10, 156–161. Acc. Chem. Res., 39, 558–566.
11 Bernhardt, R. (2006) J. Biotechnol., 124, 28 Bailey, L.J., Mccoy, J.G., Phillips, G.N.,
128–145. and Fox, B.G. (2008) Proc. Natl. Acad. Sci.
12 Urlacher, V.B. and Schmid, R.D. (2007) USA, 105, 19194–19198.
Modern Biooxidation: Enzymes, Reactions 29 Schwartz, J.K., Wei, P.P., Mitchell, K.H.,
and Applications, Wiley-VCH Verlag Fox, B.G., and Solomon, E.I. (2008)
GmbH, Weinheim. J. Am. Chem. Soc., 130, 7098–7109.
13 Nolan, L.C. and OConnor, K.E. (2008) 30 Shanklin, J. and Whittle, E. (2003) FEBS
Biotechnol. Lett., 30, 1879–1891. Lett., 545, 188–192.
14 Massey, V. (1994) J. Biol. Chem., 269, 31 van Beilen, J.B. and Funhoff, E.G. (2007)
22459–22462. Appl. Microbiol. Biotechnol., 74, 13–21.
15 van Berkel, W.J.H., Kamerbeek, N.M., 32 Rinaldo, D., Philipp, D.M., Lippard, S.J.,
and Fraaije, M.W. (2006) J. Biotechnol., and Friesner, R.A. (2007) J. Am. Chem.
124, 670–689. Soc., 129, 3135–3147.
16 Coon, M.J. (2005) Annu. Rev. Pharmacol., 33 Tinberg, C.E. and Lippard, S.J. (2009)
45, 1–25. Biochemistry, 48, 12145–12158.
17 Meunier, B., de Visser, S.P., and 34 Shu, L., Nesheim, J.C., Kauffmann, K.,
Shaik, S. (2004) Chem. Rev., 104, M€ unck, E., Lipscomb, J.D., and Que, L. Jr.
3947–3980. (1997) Science, 275, 515–518.
References j1317
35 Lee, S.K. and Lipscomb, J.D. (1999) 56 Ferraro, D.J., Gakhar, L., and
Biochemistry, 38, 4423–4432. Ramaswamy, S. (2005) Biochem. Biophys.
36 Baik, M.H., Newcomb, M., Res. Commun., 338, 175–190.
Friesner, R.A., and Lippard, S.J. (2003) 57 Kweon, O., Kim, S.J., Baek, S.,
Chem. Rev., 103, 2385–2419. Chae, J.C., Adjei, M.D., Baek, D.H.,
37 Kopp, D.A. and Lippard, S.J. (2002) Kim, Y.C., and Cerniglia, C.E. (2008)
Curr. Opin. Chem. Biol., 6, 568–576. BMC Biochem., 9, 11.
38 Lee, S.-K., Nesheim, J.C., and 58 Chakrabarty, S., Austin, R.N., Deng, D.Y.,
Lipscomb, J.D. (1993) J. Biol. Chem., 268, Groves, J.T., and Lipscomb, J.D. (2007)
21569–21577. J. Am. Chem. Soc., 129, 3514.
39 Valentine, A.M., Stahl, S.S., and 59 Kloer, D.P., Ruch, S., Al-Babili, S.,
Lippard, S.J. (1999) J. Am. Chem. Soc., Beyer, P., and Schulz, G.E. (2005)
121, 3876–3887. Science, 308, 267–269.
40 Beauvais, L.G. and Lippard, S.J. 60 Schwartz, S.H., Qin, X.Q., and
(2005) J. Am. Chem. Soc., 127, Zeevaart, J.A.D. (2001) J. Biol. Chem., 276,
7370–7378. 25208–25211.
41 Murray, L.J. and Lippard, S.J. (2007) 61 Kloer, D.P. and Schulz, G.E. (2006)
Acc. Chem. Res., 40, 466–474. Cell. Mol. Life Sci., 63, 2291–2303.
42 Wallar, B.J. and Lipscomb, J.D. (2001) 62 Schwartz, S.H., Tan, B.C., Gage, D.A.,
Biochemistry, 40, 2220–2233. Zeevaart, J.A.D., and McCarty, D.R.
43 Brazeau, B.J., Austin, R.N., Tarr, C., (1997) Science, 276, 1872–1874.
Groves, J.T., and Lipscomb, J.D. 63 Schmidt, H., Kurtzer, R., Eisenreich, W.,
(2001) J. Am. Chem. Soc., 123, and Schwab, W. (2006) J. Biol. Chem., 281,
11831–11837. 9845–9851.
44 Koehntop, K.D., Emerson, J.P., and 64 Leuenberger, M.G., Engeloch-Jarret, C.,
Que, L. (2005) J. Biol. Inorg. Chem., 10, and Woggon, W.-D. (2001) Angew. Chem.
87–93. Int. Ed., 40, 2614–2617.
45 Purpero, V. and Moran, G.R. (2007) 65 Poliakov, E., Gentleman, S.,
J. Biol. Inorg. Chem., 12, 587–601. Chander, P., Cunningham, F.X.,
46 Kovaleva, E.G. and Lipscomb, J.D. (2008) Grigorenko, B.L., Nemuhin, A.V.,
Nat. Chem. Biol., 4, 186–193. and Redmond, T.M. (2009) BMC
47 Prescott, A.G. and Lloyd, M.D. (2000) Biochem., 10, 31.
Nat. Prod. Rep., 17, 367–383. 66 Borowski, T., Blomberg, M.R.A., and
48 Hausinger, R.P. (2004) Crit. Rev. Biochem. Siegbahn, P.E.M. (2008) Chem. Eur. J., 14,
Mol., 39, 21–68. 2264–2276.
49 Krebs, C., Fujimori, D.G., Walsh, C.T., 67 Glickman, M.H. and Klinman, J.P. (1996)
and Bollinger, J.M. (2007) Acc. Chem. Biochemistry, 35, 12882–12892.
Res., 40, 484–492. 68 Ruddat, V.C., Whitman, S., Holman,
50 Hashimoto, T. and Yamada, Y. (1987) T.R., and Bernasconi, C.F. (2003)
Eur. J. Biochem., 164, 277–285. Biochemistry, 42, 4172–4178.
51 Thornburg, L.D., Lai, M.T., 69 Guengerich, F.P. (1987) Mammalian
Wishnok, J.S., and Stubbe, J. (1993) Cytochromes P450, Vol. 1 and 2, CRC
Biochemistry, 32, 14023–14033. Press, Boca Raton, FL.
52 Thornburg, L.D. and Stubbe, J. (1993) 70 Guengerich, F.P. (2003) Arch. Biochem.
Biochemistry, 32, 14034–14042. Biophys., 409, 59–71.
53 Miller, R.J. and Benkovic, S.J. (1988) 71 Yun, C.H., Yim, S.K., Kim, D.H., and
Biochemistry, 27, 3658–3663. Ahn, T. (2006) Curr. Drug Metab., 7,
54 Eser, B.E., Barr, E.W., Frantorn, P.A., 411–429.
Saleh, L., Bollinger, J.M., Krebs, C., and 72 Urlacher, V.B., Lutz-Wahl, S., and
Fitzpatrick, P.F. (2007) J. Am. Chem. Soc., Schmid, R.D. (2004) Appl. Microbiol.
129, 11334. Biotechnol., 64, 317–325.
55 Pavon, J.A. and Fitzpatrick, P.F. (2006) 73 van Beilen, J.B. and Funhoff, E.G. (2005)
Biochemistry, 45, 11030–11037. Curr. Opin. Biotechnol., 16, 308–314.
1318
74 Urlacher, V.B. and Eiben, S. (2006) Trends 91 Otto, K., Hofstetter, K., R€
othlisberger, M.,
Biotechnol., 24, 324–330. Witholt, B., and Schmid, A. (2004)
75 Sono, M., Roach, M.P., Coulter, E.D., and J. Bacteriol., 186, 5292–5302.
Dawson, J.H. (1996) Chem. Rev., 96, 92 Kantz, A., Chin, F., Nallamothu, N.,
2841–2887. Nguyen, T., and Gassner, G.T. (2005)
76 Munro, A.W., Leys, D.G., McLean, K.J., Arch. Biochem. Biophys., 442, 102–116.
Marshall, K.R., Ost, T.W.B., Daff, S., 93 Laden, B.P., Tang, Y.Z., and Porter, T.D.
Miles, C.S., Chapman, S.K., (2000) Arch. Biochem. Biophys., 374,
Lysek, D.A., Moser, C.C., Page, C.C., 381–388.
and Dutton, P.L. (2002) Trends Biochem. 94 Adam, W., Lazarus, M.,
Sci., 27, 250–257. Saha-M€oller, C.R., Weichold, O.,
77 Schlichting, I., Berendzen, J., Chu, K., Hoch, U., H€aring, D., and Schreier, P.
Stock, A.M., Maves, S.A., Benson, D.E., (1999) Adv. Biochem. Eng. Biotechnol., 63,
Sweet, R.M., Ringe, D., Petsko, G.A., and 73–108.
Sligar, S.G. (2000) Science, 287, 95 van Deurzen, M.P.J., van Rantwijk, F.,
1615–1622. and Sheldon, R.A. (1997) Tetrahedron, 53,
78 Newcomb, M., Hollenberg, P.F., and 13183–13220.
Coon, M.J. (2003) Arch. Biochem. 96 van Rantwijk, F. and Sheldon, R.A.
Biophys., 409, 72–79. (2000) Curr. Opin. Biotechnol., 11,
79 Newcomb, M., Halgrimson, J.A., 554–564.
Horner, J.H., Wasinger, E.C., Chen, L.X., 97 van de Velde, F., van Rantwijk, F., and
and Sligar, S.G. (2008) Proc. Natl. Acad. Sheldon, R.A. (2001) Trends Biotechnol.,
Sci. USA, 105, 8179–8184. 19, 73–80.
80 Green, M.T. (2009) Curr. Opin. Chem. 98 Dembitsky, V.M. (2003) Tetrahedron, 59,
Biol., 13, 84–88. 4701–4720.
81 Sheng, X., Zhang, H.M., 99 Hofrichter, M. and Ullrich, R. (2006)
Hollenberg, P.F., and Newcomb, M. Appl. Microbiol. Biotechnol., 71, 276–288.
(2009) Biochemistry, 48, 1620–1627. 100 Valderrama, B., Ayala, M., and
82 Shaik, S., Hirao, H., and Kumar, D. Vazquez-Duhalt, R. (2002)
(2007) Nat. Prod. Rep., 24, 533–552. Chem. Biol., 9, 555–565.
83 Joo, H., Lin, Z., and Arnold, F.H. (1999) 101 Park, J.B. and Clark, D.S. (2006)
Nature, 399, 670–673. Biotechnol. Bioeng., 93, 1190–1195.
84 Cirino, P.C. and Arnold, F.H. (2003) 102 Park, J.B. (2007) J. Microbiol. Biotechnol.,
Angew. Chem. Int. Ed. Engl., 42, 17, 379–392.
3299–3301. 103 Archelas, A. and Furstoss, R. (1997)
85 Massey, V. (2000) Biochem. Soc. Trans., 28, Annu. Rev. Microbiol., 51, 491–525.
283–296. 104 May, S.W. (1979) Enzyme Microb. Technol.,
86 Mihovilovic, M.D. (2006) Curr. Org. 1, 15–22.
Chem., 10, 1265–1287. 105 Kolb, H.C., Finn, M.G., and
87 Kamerbeek, N.M., Janssen, D.B., van Sharpless, K.B. (2001) Angew. Chem.
Berkel, W.J.H., and Fraaije, M.W. (2003) Int. Ed., 40, 2004–2021.
Adv. Synth. Catal., 345, 667–678. 106 Katsuki, T. and Sharpless, K.B. (1980)
88 Moonen, M.J.H., Fraaije, M.W., J. Am. Chem. Soc., 102, 5974–5976.
Rietjens, I.M.C.M., Laane, C., and 107 Sharpless, K.B. (2002) Angew. Chem.
van Berkel, W.J.H. (2002) Adv. Synth. Int. Ed., 41, 2024–2032.
Catal., 344, 1023–1035. 108 Bruyn, J.K. (1954) Ned. Akad. Wet.
89 Park, J.B., B€
uhler, B., Habicher, T., (Amsterdam), 54, 41.
Hauer, B., Panke, S., Witholt, B., and 109 Ishikura, T. and Foster, J.W. (1961)
Schmid, A. (2006) Biotechnol. Bioeng., 95, Nature, 192, 892–893.
501–512. 110 Klug, M.J. and Markovetz, A.J. (1968)
90 Panke, S., Held, M., Wubbolts, M.G., J. Bacteriol., 96, 1115–1123.
Witholt, B., and Schmid, A. (2002) 111 van der Linden, A.C. (1963) Biochim.
Biotechnol. Bioeng., 80, 33–41. Biophys. Acta, 77, 157–159.
References j1319
112 Maynert, E.W., Foreman, R.L., and 129 Rozhkova-Novosad, E.A., Chae, J.C.,
Watabe, T. (1970) J. Biol. Chem., 245, Zylstra, G.J., Bertrand, E.M.,
5234–5238. Alexander-Ozinskas, M., Deng, D.Y.,
113 Abbott, B.J. and Hou, C.T. (1973) Appl. Moe, L.A., van Beilen, J.B., Danahy, M.,
Microbiol., 26, 86–91. Groves, J.T., and Austin, R.N. (2007)
114 May, S.W. and Abbott, B.J. (1973) J. Biol. Chem. Biol., 14, 165–172.
Chem., 248, 1725–1730. 130 Wang, L., Wang, W., Lai, Q., and
115 van Beilen, J.B., Panke, S., Lucchini, S., Shao, Z. (2010) Environ. Microbiol., 12 (5),
Franchini, A.G., R€othlisberger, M., and 1230–1242.
Witholt, B. (2001) Microbiology, 147, 131 May, S.W., Schwartz, R.D., Abbott, B.J.,
1621–1630. and Zaborsky, O.R. (1975) Biochim.
116 May, S., Steltenkamp, M.S., Biophys. Acta, 403, 245–255.
Schwartz, R.D., and McCoy, C.J. (1976) 132 Schwartz, R.D. and McCoy, C.J. (1977)
J. Am. Chem. Soc., 98, 7856–7858. Appl. Environ. Microbiol., 34, 47–49.
117 Peterson, J.A., Basu, D., and Coon, M.J. 133 de Smet, M.-J., Witholt, B., and
(1966) J. Biol. Chem., 241, 5162–5164. Wynberg, H. (1981) J. Org. Chem., 46,
118 Peterson, J.A., Kusunose, M., Kusunose, 3128–3131.
E., and Coon, M.J. (1967) J. Biol. Chem., 134 de Smet, M.-J., Wynberg, H., and
242, 4334–4340. Witholt, B. (1981) Appl. Environ.
119 McKenna, E. and Coon, M.J. (1970) Microbiol., 42, 811–816.
J. Biol. Chem., 245, 3882–3889. 135 de Smet, M.-J., Kingma, J., Wynberg, H.,
120 Ruettinger, R.T., Griffith, G.R., and and Witholt, B. (1983) Enzyme Microb.
Coon, M.J. (1977) Arch. Biochem. Biophys., Technol., 5, 352–360.
183, 528–537. 136 Harbron, S., Smith, B.W., and
121 Ruettinger, R.T., Olson, S.T., Boyer, R.F., Lilly, M.D. (1986) Enzyme Microb.
and Coon, M.J. (1974) Biochem. Biophys. Technol., 8, 85–88.
Res. Commun., 57, 1011–1017. 137 Doig, S.D., Boam, A.T., Livingston, A.G.,
122 van Beilen, J.B., Kingma, J., and and Stuckey, D.C. (1999) Biotechnol.
Witholt, B. (1994) Enzyme Microb. Bioeng., 63, 601–611.
Technol., 16, 904–911. 138 Steinig, G.H., Livingston, A.G., and
123 van Beilen, J.B., Penninga, D., and Stuckey, D.C. (2000) Biotechnol. Bioeng.,
Witholt, B. (1992) J. Biol. Chem., 267, 70, 553–563.
9194–9201. 139 Wubbolts, M.G., Favre-Bulle, O., and
124 Witholt, B., de Smet, M.-J., Kingma, J., Witholt, B. (1996) Biotechnol. Bioeng., 52,
van Beilen, J.B., Lageveen, R.G., and 301–308.
Eggink, G. (1990) Trends Biotechnol., 8, 140 Bosetti, A., van Beilen, J.B.,
46–52. Preusting, H., Lageveen, R.G., and
125 Austin, R.N., Chang, H.K., Zylstra, G.J., Witholt, B. (1992) Enzyme Microb.
and Groves, J.T. (2000) J. Am. Chem. Soc., Technol., 14, 702–708.
122, 11747–11748. 141 Favre-Bulle, O., Schouten, T.,
126 Austin, R.N., Luddy, K., Erickson, K., Kingma, J., and Witholt, B. (1991)
Pender-Cudlip, M., Bertrand, E., Bio/Technol., 9, 367–371.
Deng, D., Buzdygon, R.S., 142 Favre-Bulle, O., Weenink, E., Vos, T.,
van Beilen, J.B., and Groves, J.T. (2008) Preusting, H., and Witholt, B. (1993)
Angew. Chem. Int. Ed., 47, 5232–5234. Biotechnol. Bioeng., 41, 263–272.
127 Bertrand, E., Sakai, R., 143 Favre-Bulle, O. and Witholt, B. (1992)
Rozhkova-Novosad, E., Moe, L., Fox, B.G., Enzyme Microb. Technol., 14, 931–937.
Groves, J.T., and Austin, R.N. (2005) 144 Johnstone, S.I., Philips, G.T.,
J. Inorg. Biochem., 99, 1998–2006. Robertson, B.W., Watts, P.D.,
128 van Beilen, J.B., Smits, T.H., Whyte, L.G., Bertola, M.A., Koger, H.S., and Marx, A.F.
Schorcht, S., R€othlisberger, M., Plaggemeier, (1987) Stereoselective synthesis of S-
T., Engesser, K.H., and Witholt, B. (2002) (–)-b-blockers via microbially produced
Environ. Microbiol., 4, 676–682. epoxide intermediates in Biocatalysis in
1320
Organic Media (eds C. Laane, J. Tramper, 163 de Bont, J.A.M. and Harder, W. (1978)
and M.D. Lilly) Elsevier, Amsterdam, pp. FEMS Microbiol. Lett., 3, 89–93.
387–392. 164 de Bont, J.A.M., van Ginkel, C.G.,
145 Toda,F.N., Hayashi, S., Hatano, Y., Tramper, J., and Luyben, K.C.A.M.
Okunishi, H., and Miyazaki, M. (1978) (1983) Enzyme Microb. Technol.,
J. Pharmacol. Exp. Ther., 207, 311–319. 5, 55–59.
146 Prichanont, S., Leak, D.J., and Stuckey, 165 Habets-Cr€ utzen, A.Q.H. and de Bont,
D.C. (1998) Enzyme Microb. Technol., 22, J.A.M. (1985) Appl. Microbiol. Biotechnol.,
471–479. 22, 428–433.
147 Weijers, C.A.G.M., van Ginkel, C.G., and 166 Brink, L.E.S. and Tramper, J. (1985)
de Bont, J.A.M. (1988) Enzyme Microb. Biotechnol. Bioeng., 27, 1258–1269.
Technol., 10, 214–218. 167 Brink, L.E.S. and Tramper, J. (1986)
148 Hanson, R.S. and Hanson, T.E. (1996) Enzyme Microb. Technol., 8, 334–340.
Microbiol. Rev., 60, 439–471. 168 Brink, L.E.S. and Tramper, J. (1986)
149 Ohta, H. and Tetsukawa, H. (1978) Enzyme Microb. Technol., 8, 281–288.
J. Chem. Soc., Chem. Commun., 849–850. 169 Brink, L.E.S. and Tramper, J. (1987)
150 Ohta, H. and Tetsukawa, H. (1979) Agric. Enzyme Microb. Technol., 9, 612–618.
Biol. Chem., 43, 2099–2104. 170 Kovalenko, G. (1992) Biotechnol. Bioeng.,
151 Habets-Cr€ utzen, A.Q.H., Carlier, S.J.N., 39, 522–528.
de Bont, J.A.M., Wistuba, D., Schurig, V., 171 Hou, C.T., Patel, R., Laskin, A.I., and
Hartmans, S., and Tramper, J. (1985) Barnabe, N. (1979) Appl. Environ.
Enzyme Microb. Technol., 7, 17–21. Microbiol., 38, 127–134.
152 Owens, C.R., Karceski, J.K., and 172 Hou, C.T. (1984) Appl. Microbiol.
Mattes, T.E. (2009) Appl. Microbiol. Biotechnol., 19, 1–4.
Biotechnol., 84, 685–692. 173 Subramanian, V. (1986) J. Ind. Microbiol.,
153 Mahmoudian, M. and Michael, A. (1992) 1, 119–127.
J. Ind. Microbiol., 11, 29–35. 174 Stanley, S.H. and Dalton, H. (1992)
154 Archelas, A., Hartmans, S., and Tramper, Biocatalysis, 6, 163–175.
J. (1988) Biocatalysis, 1, 283–292. 175 Richards, A.O., Stanley, S.H., Suzuki, M.,
155 McClay, K., Fox, B.G., and Steffan, R.J. and Dalton, H. (1994) Biocatalysis, 8,
(2000) Appl. Environ. Microbiol., 66, 253–267.
1877–1882. 176 Gross, R., Hauer, B., Otto, K., and
156 Kubo, T., Peters, M.W., Meinhold, P., and Schmid, A. (2007) Biotechnol. Bioeng., 98,
Arnold, F.H. (2006) Chem. Eur. J., 12, 1123–1134.
1216–1220. 177 Xin, J.Y., Cui, J.R., Chen, J.B., Li, S.B., Xia,
157 Furuhashi, K. (1992) Biological routes to C.G., and Zhu, L.M. (2003) Process
optically active epoxides in Chirality in Biochem., 38, 1739–1746.
Industry (eds A.N. Collins, G.N. 178 van Ginkel, C.G., Welten, H.G.J.,
Sheldrake, and J. Crosby), John Wiley & and de Bont, J.A.M. (1986) Appl.
Sons, Ltd., pp. 167–186. Microbiol. Biotechnol., 24, 334–337.
158 Furuhashi, K., Taoka, A., Uchida, S., 179 Woods, N.R. and Murrell, J.C. (1990)
Karube, I., and Suzuki, A. (1981) Eur. J. Biotechnol. Lett., 12, 409–414.
Appl. Microbiol. Biotechnol., 12, 39–45. 180 Mahmoudian, M. and Michael, A.
159 Furuhashi, K. (1986) Chem. Econ. Eng. (1992) Appl. Microbiol. Biotechnol., 37,
Rev., 18, 21–26. 23–27.
160 Furuhashi, K., Shintani, M., and 181 White, R.F., Birnbaum, J., Meyer, R.T.,
Takagi, M. (1986) Appl. Microbiol. Tenbroek., J., Chemerda, J.M., and
Biotechnol., 23, 218–223. Demain, A.L. (1971) Appl. Microbiol., 22,
161 Furuhashi, K. (1984) Appl. Microbiol. 55–60.
Biotechnol., 20, 6–9. 182 Archelas, A. and Furstoss, R. (1992)
162 de Bont, J.A.M., Attwood, M.M., Tetrahedron Lett., 33, 5241–5242.
Primrose, S.B., and Harder, W. (1979) 183 Ruettinger, R.T. and Fulco, A.J. (1981)
FEMS Microbiol. Lett., 6, 183–188. J. Biol. Chem., 256, 5728–5734.
References j1321
184 Ruettinger, R.T., Wen, L.P., and 202 Bj€
orkling, F., Godtfredsen, S.E., and
Fulco, A.J. (1989) J. Biol. Chem., 264, Kirk, O. (1990) J. Chem. Soc., Chem.
10987–10995. Commun., 1301–1303.
185 Lewis, J.C. and Arnold, F.H. (2009) 203 Klaas, M.R.G. and Warwel, S. (1997)
Chimia, 63, 309–312. J. Mol. Catal. A, Chem., 117, 311–319.
186 Farinas, E.T., Alcalde, M., and Arnold, F. 204 Moreira, M.A., Bitencourt, T.B., and
(2004) Tetrahedron, 60, 525–528. Nascimento, M.D. (2005) Synth.
187 Geigert, J., Lee, T.D., Dalietos, D.J., Commun., 35, 2107–2114.
Hirano, D.S., and Neidleman, S.L. (1986) 205 Wiles, C., Hammond, M.J., and Watts, P.
Biochem. Biophys. Res. Commun., 136, (2009) Beilstein J. Org. Chem., 5 (27), 1–12.
778–782. 206 Jarvie, A.W.P., Overton, N., and St
188 Zaks, A. and Dodds, D.R. (1995) Pourcain, C.B. (1999) J. Chem. Soc., Perkin
J. Am. Chem. Soc., 117, 10419–10424. Trans. 1, 2171–2176.
189 Conesa, A., Punt, P.J., and van den 207 Warwel, S. and Klaas, M.R.G. (1995)
Hondel, C.A. (2002) J. Biotechnol., 93, J. Mol. Catal. B: Enzym., 1, 29–35.
143–158. 208 Klaas, M.R.G. and Warwel, S. (1996)
190 Allain, E.J., Hager, L.P., Deng, L., and J. Am. Oil Chem. Soc., 73, 1453–1457.
Jacobsen, E.N. (1993) J. Am. Chem. Soc., 209 Hilker, I., Bothe, D., Pruss, J., and
115, 4415–4416. Warnecke, H.J. (2001) Chem. Eng. Sci.,
191 Dexter, A.F. and Hager, L.P. (1995) 56, 427–432.
J. Am. Chem. Soc., 117, 817–818. 210 Bitencourt, T.B. and Nascimento, M.D.
192 Dexter, A.F., Lakner, F.J., Campbell, R.A., (2009) Green Chem., 11, 209–214.
and Hager, L.P. (1995) J. Am. Chem. Soc., 211 Svedendahl, M., Carlqvist, P.,
117, 6412–6413. Branneby, C., Allner, O., Frise, A.,
193 Lakner, F.J., Cain, K.P., and Hult, K., Berglund, P., and
Hager, L.P. (1997) J. Am. Chem. Soc., 119, Brinck, T. (2008) ChemBioChem, 9,
443–444. 2443–2451.
194 Grey, C.E., Hedstrom, M., and 212 Torres Pazmino, D.E., Winkler, M.,
Adlercreutz, P. (2007) ChemBioChem, 8, Glieder, A., and Fraaije, M.W. (2010)
1055–1062. J. Biotechnol., 146, 9–24.
195 Seelbach, K., van Deurzen, M.P.J., 213 Panke, S., Wubbolts, M.G., Schmid, A.,
van Rantwijk, F., Sheldon, R.A., and and Witholt, B. (2000) Biotechnol. Bioeng.,
Kragl, U. (1997) Biotechnol. Bioeng., 55, 69, 91–100.
283–288. 214 Sello, G., Orsini, F., Bernasconi, S., and
196 van Deurzen, M.P.J., Seelbach, K., Di Gennaro, P. (2006) Tetrahedron:
van Rantwijk, F., Kragl, U., and Asymmetry, 17, 372–376.
Sheldon, R.A. (1997) Biocatal. 215 Bernasconi, S., Orsini, F., Sello, G., and
Biotransform., 15, 1–16. Di Gennaro, P. (2004) Tetrahedron:
197 van de Velde, F., Lourenco, N.D., Asymmetry, 15, 1603–1606.
Bakker, M., van Rantwijk, F., and 216 Tischler, D., Eulberg, D., Lakner, S.,
Sheldon, R.A. (2000) Biotechnol. Kaschabek, S.R., van Berkel, W.J.H., and
Bioeng., 69, 286–291. Schlomann, M. (2009) J. Bacteriol., 191,
198 Rai, G.P., Sakai, S., Florez, A.M., 4996–5009.
Mogollon, L., and Hager, L.P. (2001) 217 van Hellemond, E.W., Janssen, D.B., and
Adv. Synth. Catal., 343, 638–645. Fraaije, M.W. (2007) Appl. Environ.
199 Park, J.B. and Clark, D.S. (2006) Microbiol., 73, 5832–5839.
Biotechnol. Bioeng., 94, 189–192. 218 Park, M.S., Bae, J.W., Han, J.H.,
200 Tzialla, A.A., Kalogeris, E., Gournis, D., Lee, E.Y., Lee, S.G., and Park, S.
Sanakis, Y., and Stamatis, H. (2008) (2006) J. Microbiol. Biotechnol., 16,
J. Mol. Catal. B: Enzym., 51, 24–35. 1032–1040.
201 Bj€orkling, F., Frykman, H., 219 OConnor, K.E., Dobson, A.D., and
Godtfredsen, S.E., and Kirk, O. (1992) Hartmans, S. (1997) Appl. Environ.
Tetrahedron, 48, 4587–4592. Microbiol., 63, 4287–4291.
1322
220 Hartmans, S., van der Werf, M.J., and Loehr, T.M., and de Montellano, P.R.O.
de Bont, J.A. (1990) Appl. Environ. (2000) J. Biol. Chem., 275, 14112–14123.
Microbiol., 56, 1347–1351. 238 Vaze, A., Parizo, M., and Rusling, J.F.
221 Panke, S., Witholt, B., Schmid, A., and (2004) Langmuir, 20, 10943–10948.
Wubbolts, M.G. (1998) Appl. Environ. 239 Munge, B., Estavillo, C., Schenkman, J.B.,
Microbiol., 64, 2032–2043. and Rusling, J.F. (2003) ChemBioChem, 4,
222 Kuhn, D., Kholiq, M.A., Heinzle, E., 82–89.
B€uhler, B., and Schmid, A. (2010) 240 Lvov, Y.M., Lu, Z.Q., Schenkman, J.B.,
Green Chem. 12, 815–827. Zu, X.L., and Rusling, J.F. (1998)
223 Hoffmann, F. and Rinas, U. (2001) J. Am. Chem. Soc., 120, 4073–4080.
Biotechnol. Bioeng., 76, 333–340. 241 Manoj, K.M., Lakner, F.J., and Hager, L.P.
224 B€uhler, B., Park, J.-B., Blank, L.M., and (2000) J. Mol. Catal. B: Enzym., 9,
Schmid, A. (2008) Appl. Environ. 107–111.
Microbiol., 74, 1436–1446. 242 Aoun, S. and Baboulene, M. (1998)
225 Blank, L.M., Ebert, B.E., B€ uhler, B., and J. Mol. Catal. B: Enzym., 4, 101–109.
Schmid, A. (2008) Biotechnol. Bioeng., 243 Kaup, J.A., Ehrich, K., Pescheck, M.,
100, 1050–1065. and Schrader, J. (2008) Biotechnol. Bioeng.,
226 Blank, L.M., Ionidis, G., Ebert, B.E., 99, 491–498.
B€uhler, B., and Schmid, A. (2008) 244 Yazbik, V. and Ansorge-Schumacher, M.
FEBS J., 275, 5173–5190. (2010) Process Biochem., 45, 279–283.
227 Park, J.B., B€uhler, B., Panke, S., 245 Conesa, A., van de Velde, F., van
Witholt, B., and Schmid, A. (2007) Rantwijk, F., Sheldon, R.A.,
Biotechnol. Bioeng., 98, 1219–1229. van den Hondel, C., and Punt, P.J. (2001)
228 Gross, R., Lang, K., B€ uhler, K., and J. Biol. Chem., 276, 17635–17640.
Schmid, A. (2010) Biotechnol. Bioeng., 246 Grey, C.E., Rundback, F., and
105, 705–717. Adlercreutz, P. (2008) J. Biotechnol., 135,
229 Schmid, A., Hofstetter, K., Feiten, H.J., 196–201.
Hollmann, F., and Witholt, B. (2001) Adv. 247 Lee, K. and Moon, S.H. (2003) J.
Synth. Catal., 343, 732–737. Biotechnol., 102, 261–268.
230 Hollmann, F., Lin, P.C., Witholt, B., and 248 Kohlmann, C. and Lutz, S. (2006) Eng. Life
Schmid, A. (2003) J. Am. Chem. Soc., 125, Sci., 6, 170–174.
8209–8217. 249 La Rotta, C.E., DElia, E., and Bon, E.P.S.
231 Hofstetter, K., Lutz, J., Lang, I., (2007) Electron. J. Biotechnol., 10, 24–36.
Witholt, B., and Schmid, A. (2004) 250 Karmee, S.K., Roosen, C., Kohlmann, C.,
Angew. Chem. Int. Ed., 43, 2163–2166. Lutz, S., Greiner, L., and Leitner, W.
232 Hollmann, F., Hofstetter, K., (2009) Green Chem., 11, 1052–1055.
Habicher, T., Hauer, B., and 251 Hollmann, F. and Schmid, A. (2009)
Schmid, A. (2005) J. Am. Chem. Soc., 127, J. Inorg. Biochem., 103, 313–315.
6540–6541. 252 Perez, D.I., Grau, M.M., Arends,
233 Ruinatscha, R., Dusny, C., Buehler, K., I.W.C.E., and Hollmann, F. (2009)
and Schmid, A. (2009) Adv. Synth. Catal., Chem. Commun., 44, 6848–6850.
351, 2505–2515. 253 Patel, R.N. (2000) Stereoselective
234 Mayhew, M.P., Reipa, V., Holden, M.J., Biocatalysis, Marcel Dekker, New York.
and Vilker, V.L. (2000) Biotechnol. Prog., 254 Mukherje, B.B., Kraidman, G., and
16, 610–616. Hill, I.D. (1973) Appl. Microbiol., 25,
235 Shoji, O., Fujishiro, T., Nakajima, H., 447–453.
Kim, M., Nagano, S., Shiro, Y., and 255 Borges, K.B., Borges, W.D., Duran-
Watanabe, Y. (2007) Angew. Chem. Int. Patron, R., Pupo, M.T., Bonato, P.S., and
Ed., 46, 3656–3659. Collado, I.G. (2009) Tetrahedron:
236 Rabe, K.S., Kiko, K., and Niemeyer, C.M. Asymmetry, 20, 385–397.
(2008) ChemBioChem, 9, 420–425. 256 Bicas, J.L., Dionisio, A.P., and
237 Koo, L.S., Tschirret-Guth, R.A., Pastore, G.M. (2009) Chem. Rev., 109,
Straub, W.E., Moenne-Loccoz, P., 4518–4531.
References j1323
257 Imai, K., Marumo, S., and Ohtaki, T. 276 Dhavlikar, R.S. and Albroscheit, G. (1973)
(1976) Tetrahedron Lett., 17, 1211–1214. Dragoco Rep., 12, 251–258.
258 David, L. and Veschambre, H. (1984) 277 Sowden, R.J., Yasmin, S., Rees, N.H.,
Tetrahedron Lett., 25, 543–546. Bell, S.G., and Wong, L.L. (2005)
259 David, L. and Veschambre, H. (1985) Org. Biomol. Chem., 3, 57–64.
Agric. Biol. Chem., 49, 1487–1489. 278 Dietrich, J.A., Yoshikuni, Y., Fisher, K.J.,
260 Abraham, W.R., Stumpf, B., and Woolard, F.X., Ockey, D., McPhee, D.J.,
Arfmann, H.A. (1990) J. Essent. Oil Res., 2, Renninger, N.S., Chang, M.C.Y.,
251–257. Baker, D., and Keasling, J.D. (2009)
261 Holmes, D.S., Ashworth, D.M., and ACS Chem. Biol., 4, 261–267.
Robinson, J.A. (1990) Helv. Chim. Acta, 279 Ziffer, H., Kabuto, K., Gibson, D.T.,
73, 260–271. Kobal, V.M., and Jerina, D.M. (1977)
262 Demyttenaere, J.C.R. and Willemen, H.M. Tetrahedron, 33, 2491–2496.
(1998) Phytochemistry, 47, 1029–1036. 280 Geary, P.J., Pryce, R.J., Roberts, S.M.,
263 Schwab, E., Bernreuther, A., Ryback, G., and Winders, J.A. (1990)
Puapoomchareon, P., Mori, K., and J. Chem. Soc., Chem. Commun., 204–205.
Schreier, P. (1991) Tetrahedron: 281 Davies, J.I. and Evans, W.C. (1964)
Asymmetry, 2, 471–479. Biochem. J., 91, 251–261.
264 Mirata, M.A., Wust, M., Mosandl, A., and 282 Resnick, S.M., Lee, K., and Gibson, D.T.
Schrader, J. (2008) J. Agric. Food Chem., (1996) J. Ind. Microbiol. Biotechnol., 17,
56, 3287–3296. 438–457.
265 Meesters, R.J.W., Duisken, M., and 283 Boyd, D.R., Sharma, N.D., Bowers, N.I.,
Hollender, J. (2007) Xenobiotica, 37, Brannigan, I.N., Groocock, M.R., Malone,
604–617. J.E., McConville, G., and Allen, C.C.R.
266 van der Werf, M.J., Keijzer, P.M., and van (2005) Adv. Synth. Catal., 347, 1081–1089.
der Schaft, P.H. (2000) J. Biotechnol., 84, 284 Wackett, L.P., Kwart, L.D., and Gibson,
133–143. D.T. (1988) Biochemistry, 27, 1360–1367.
267 Schewe, H., Kaup, B.A., and Schrader, J. 285 Boyd, D.R., McMordie, R.A.S.,
(2008) Appl. Microbiol. Biotechnol., 78, Sharma, N.D., Dalton, H., Williams, P.,
55–65. and Jenkins, R.O. (1989) J. Chem. Soc.,
268 Bell, S.G., Sowden, R.J., and Wong, L.L. Chem. Commun., 339–340.
(2001) Chem. Commun., 635–636. 286 Boyd, D.R., Dorrity, M.R.J., Malone, J.F.,
269 Lindmark-Henriksson, M., Isaksson, D., McMordie, R.A.S., Sharma, N.D.,
Vanek, T., Valterova, I., Hogberg, H.E., Dalton, H., and Williams, P. (1990)
and Sjodin, K. (2004) J. Biotechnol., 107, J. Chem. Soc., Perkin Trans. 1, 489–494.
173–184. 287 Boyd, D.R., Sharma, N.D., Kerley, N.A.,
270 Sakamaki, H., Itoh, K., Chai, W., McMordie, R.A.S., Sheldrake, G.N.,
Hayashida, Y., Kitanaka, S., and Horiuchi, Williams, P., and Dalton, H. (1996)
C.A. (2004) J. Mol. Catal. B: Enzym., 27, J. Chem. Soc., Perkin Trans. 1, 67–74.
177–181. 288 Bowers, N.I., Boyd, D.R., Sharma, N.D.,
271 Duisken, M., Benz, D., Peiffer, T.H., Kennedy, M.A., Sheldrake, G.N., and
Blomeke, B., and Hollender, J. (2005) Dalton, H. (1998) Tetrahedron: Asymmetry,
Curr. Drug Metab., 6, 593–601. 9, 1831–1834.
272 Moreira, M.A. and Nascimento, M.G. 289 Abraham, W.R., Stumpf, B., and Kieslich,
(2007) Catal. Commun., 8, 2043–2047. K. (1986) Appl. Microbiol. Biotechnol., 24,
273 Skouridou, V., Stamatis, H., and 24–30.
Kolisis, F.N. (2003) Biocatal. 290 Abraham, W.R., Hoffmann, J.M.R.,
Biotransform., 21, 285–290. Kieslich, K., Reng, G., and Stumpf, B.
274 Skouridou, V., Stamatis, H., and Kolisis, F.N. (1985) Microbial transformations of some
(2003) J. Mol. Catal. B: Enzym., 21, 67–69. monoterpenoids and sesquiterpenoids in
275 Hikino, H., Konno, C., Nagashim., T., Enzymes in Organic Synthesis (ed. S.C.R.
Kohama, T., and Takemoto, T. (1971) Porter), Pitman Press, London,
Tetrahedron Lett., 12, 337–340. pp. 146–157.
1324
291 Abraham, W.R., Stumpf, B., Kieslich, K., 308 Ilg, A., Beyer, P., and Al-Babili, S. (2009)
Reif, S., and Hoffmann, H.M.R. (1986) FEBS J., 276, 736–747.
Appl. Microbiol. Biotechnol., 24, 31–34. 309 Vogel, J.T., Tan, B.C., McCarty, D.R., and
292 Abraham, W.R., Arfmann, H.A., Klee, H.J. (2008) J. Biol. Chem., 283,
Stumpf, B., Washausen, P., and 11364–11373.
Kieslich, K. (1988) Microbial 310 Marasco, E.K., Vay, K., and
transformations of some terpenoids and Schmidt-Dannert, C. (2006) J. Biol.
natural compounds in Bioflavor Chem., 281, 31583–31593.
(ed. P. Schreier), Walter de Gruyter, 311 Rodriguez-Bustamante, E. and
New York, pp. 399–414. Sanchez, S. (2007) Crit. Rev. Microbiol., 33,
293 Arfmann, H.A., Abraham, W.R., and 211–230.
Kieslich, K. (1988) Biocatalysis, 2, 59–67. 312 Schilling, M., Haetzelt, F., Schwab, W.,
294 van der Werf, M.J., Swarts, H.J., and and Schrader, J. (2008) Biotechnol. Lett.,
de Bont, J.A.M. (1999) Appl. Environ. 30, 701–706.
Microbiol., 65, 2092–2102. 313 Zorn, H., Langhoff, S., Scheibner, M.,
295 Fourneron, J.D., Archelas, A., and Furstoss, and Berger, R.G. (2003) Appl. Microbiol.
R. (1989) J. Org. Chem., 54, 4686–4689. Biotechnol., 62, 331–336.
296 Zhang, X.M., Archelas, A., and 314 Zelena, K., Hardebusch, B., Hulsdau, B.,
Furstoss, R. (1991) J. Org. Chem., 56, Berger, R.G., and Zorn, H. (2009) J. Agric.
3814–3817. Food Chem., 57, 9951–9955.
297 Meou, A., Bouanah, N., Archelas, A., 315 Scheibner, M., Hulsdau, B., Zelena, K.,
Zhang, X.M., Guglielmetti, R., and Nimtz, M., de Boer, L., Berger, R.G., and
Furstoss, R. (1990) Synthesis, 752–753. Zorn, H. (2008) Appl. Microbiol.
298 Meou, A., Bouanah, N., Archelas, A., Biotechnol., 77, 1241–1250.
Zhang, X.M., Guglielmetti, R., and 316 Lara, M., Mutti, F.G., Glueck, S.M., and
Furstoss, R. (1991) Synthesis, 681–682. Kroutil, W. (2008) Eur. J. Org. Chem.,
299 Zhang, X.M., Archelas, A., Meou, A., and 3668–3672.
Furstoss, R. (1991) Tetrahedron: 317 Mang, H., Gross, J., Lara, M.,
Asymmetry, 2, 247–250. Goessler, C., Schoemaker, H.E.,
300 Zhang, X.M., Archelas, A., and Guebitz, G.M., and Kroutil, W. (2006)
Furstoss, R. (1992) Tetrahedron: Angew. Chem. Int. Ed., 45, 5201–5203.
Asymmetry, 3, 1373–1376. 318 Mang, H., Gross, J., Lara, M.,
301 Furusawa, M., Hashimoto, T., Noma, Y., Goessler, C., Schoemaker, H.E.,
and Asakawa, Y. (2005) Chem. Pharm. Guebitz, G.M., and Kroutil, W. (2007)
Bull., 53, 1423–1429. Tetrahedron, 63, 3350–3354.
302 Zorn, H., Langhoff, S., Scheibner, M., 319 Geigert, J., Neidleman, S.L., and
Nimtz, M., and Berger, R.G. (2003) Biol. Dalietos, D.J. (1983) J. Biol. Chem., 258,
Chem., 384, 1049–1056. 2273–2277.
303 Wu, L., Thompson, D.K., Li, G., 320 Geigert, J., Neidleman, S.L., Dalietos,
Hurt, R.A., Tiedje, J.M., and Zhou, J. D.J., and Dewitt, S.K. (1983) Appl.
(2001) Appl. Environ. Microbiol., 67, Environ. Microbiol., 45, 1575–1581.
5780–5790. 321 Komives, E.A. and Ortiz de Montellano,
304 Wache, Y., Bosser-DeRatuld, A., P.R. (1987) J. Biol. Chem., 262,
Lhuguenot, J.C., and Belin, J.M. (2003) 9793–9802.
J. Agric. Food Chem., 51, 1984–1987. 322 Ortiz de Montellano, P.R. and Komives,
305 Bosser, A. and Belin, J.M. (1994) E.A. (1985) J. Biol. Chem., 260,
Biotechnol. Prog., 10, 129–133. 3330–3336.
306 Lara, M., Mutti, F.G., Glueck, S.M., and 323 Lai, M., Wu, W., and Stubbe, J. (1995)
Kroutil, W. (2009) J. Am. Chem. Soc., 131, J. Am. Chem. Soc., 117, 5023–5030.
5368–5369. 324 Zhu, N.J., Lightsey, D., Liu, J.W.,
307 Auldridge, M.E., McCarty, D.R., and Foroozesh, M., Morgan, K.M.,
Klee, H.J. (2006) Curr. Opin. Plant Biol., 9, Stevens, E.D., and Stevens, C.L.K.
315–321. (2010) J. Chem. Crystallogr., 40, 343–352.
j1325
32
Oxidation of Alcohols, Aldehydes, and Acids
Frank Hollmann, Katja B€uhler, and Bruno B€
uhler
32.1
Introduction
There are various reasons for an organic chemist to consider biocatalysis to perform a
given oxidation. Enzymes are biobased catalysts also exhibiting superb biocompat-
ibility and biodegradability. Toxicologically, enzymes are clearly superior to transition
metals and many organocatalysts. Enzymatic reactions are usually performed under
milder reaction conditions compared to established chemical methodologies.
Furthermore, stoichiometric oxidants such as molecular oxygen, hydrogen peroxide,
and small organic compounds (e.g., acetone) are environmentally more acceptable
than classical chemical oxidants such as hypochlorite or osmium tetroxide. However,
one must be aware of the fact that biocatalysis is not per se more eco-friendly than well-
designed chemical methodologies. This claim, frequently raised especially by
biotechnologists, unfortunately too often lacks quantitative justifications. Only a
full life-cycle assessment (LCA, ISO 14000) comparing all steps of competing
chemical and biocatalytic methodologies can result in a realistic evaluation.
One clear advantage of bio- over chemical catalysts lies in their often higher
selectivity. Examples presented in this chapter consist of highly regio-, chemo-, and
enantioselective oxidative transformations. For many of these reactions, a chemical
counterpart with comparable selectivity is not known. Thus, biocatalysis can help to
drastically simplify synthesis strategies and circumvent tedious protection/deprotec-
tion steps.
32.2
Oxidation of Alcohols
The first part of this chapter discusses enzymatic methods for the oxidation of
alcohols. The relevant enzyme classes and some of the most important examples are
presented. An overview of mechanistic details and practical issues such as cofactor
regeneration is given. The second part focuses on practical application of biocatalytic
and Oliver May.
j 32 Oxidation of Alcohols, Aldehydes, and Acids
1326
Figure 32.1 Examples for alcohol oxidations of corresponding acids; (c) oxidative lactonization
preparative interest: (a) oxidation of primary of diols; (d) enantiopure alcohols via redox
alcohols selectively to the corresponding deracemization/stereoinversion;
aldehydes; (b) through oxidation to (e) regioselective oxidation in polyols.
alcohol oxidation in organic synthesis. Especially, examples of regioselective oxida-

tion of polyols and enantiospecific oxidation to generate optically pure alcohols via
(dynamic) kinetic resolution, aldehydes, and acids will be highlighted (Figure 32.1).
So-called oxidoreductases (EC 1.x.x.x) are the catalysts of choice for the oxidation of
alcohols. Within this diverse class of enzymes, mostly alcohol dehydrogenases
(ADHs, EC 1.1.1.x) and alcohol oxidases (AlcOxs, EC 1.1.3.x) are used for this task.
Compared to this, peroxidases (EC 1.11.1.x) are used to a lesser extent. These enzyme
classes are discussed below.
32.2.1
Alcohol Dehydrogenases (ADH) as Catalyst for the Oxidation of Alcohols
The most popular biocatalysts for the oxidation of alcohols are the so-called alcohol
dehydrogenases (EC 1.1.1.x, ADHs). They catalyze the reversible hydride abstraction
from the substrate and simultaneous transfer to the oxidized nicotinamide cofactors
(NAD(P) þ , Scheme 32.1).
ADHs (sometimes also called ketoreductases) can be classified based on bio-
chemical characteristics into short-chain and metal-free, medium-chain and Zn2 þ -
containing, and long chain and Fe2 þ -activated ADHs [1]. From a preparative point
of view, however, more useful is the classification according to stereochemical
properties. Thus, four types can be distinguished according to the stereopreference
towards the alcohol-hydride abstracted (Prelog, anti-Prelog) [2] and whether the
hydride is transferred to the pro-R or pro-S side of the nicotinamide ring [1].
32.2 Oxidation of Alcohols j1327
Scheme 32.1 Alcohol dehydrogenase (ADH)-catalyzed, reversible oxidation of alcohols. Reducing

equivalents liberated from the alcohol substrate are transferred as hydrides to the oxidized
nicotinamide cofactor NAD (non-phosphorylated) or NADP (phosphorylated).
Figure 32.2 shows a simplified oxidation mechanism of NAD(P) þ -dependent

ADH. The oxidized nicotinamide cofactor and the alcohol substrate bind sequentially
to the ADHs active site, which brings both in close proximity and arranges optimal
spatial orientation for the hydride transfer [3]. Coordination of the alcohol substrate to
Figure 32.2 Simplified ADH-oxidation mechanism [3].

1328
the metal ion further activates it for the hydride transfer to the oxidized nicotinamide
cofactor.
The standard redox potentials (vs. NHE) for the redox couples NAD(P)H/NAD
(P) þ , ethanol/acetaldehyde, and isopropanol/acetone are 320, 199, and 286
mV, respectively [4]. Thus, NAD(P) þ -coupled oxidation of primary and secondary
alcohols is thermodynamically uphill, resulting in equilibria lying far on the side of
the reduced alcohol substrates. This necessitates efficient methods to shift the
unfavorable equilibria towards the desired products (Le Châtaliers principle, vide
infra). Another important issue of ADH-catalysis is that the catalysts are very
often prone to pronounced product inhibition. Ways around this are discussed later
in this chapter.
32.2.1.1 Commonly Used ADHs

ADHs are endogenously found in all kingdoms of life. Thus virtually all organisms
can serve as source for ADHs. As of 2009, more than 300 different ADHs are known
http://www.brenda-enzymes.org. However, the most popular, commercially avail-
able ADHs originate from horse liver (HLADH) and microorganisms, such as
Thermoanaerobium brockii (TBADH), bakers yeast (YADH), Candida, and Lactoba-
cillus [1, 5, 6]. Depending on the desired selectivity Figure 32.3), a suitable ADH can
be chosen from the great variety of reported enzymes.
32.2.1.2 Horse Liver Alcohol Dehydrogenase (HLADH)

Horse liver alcohol dehydrogenase (HLADH) is certainly one of the most popular and
best-characterized ADHs [1]. The NAD-dependent enzyme exists as a dimer com-
posed of almost identical subunits designated E (for ethanol active) and S (for steroid
Figure 32.3 Substrate spectra of various LKADH ¼ Lactobacillus kefir ADH;

ADHs. YADH ¼ yeast ADH; HLADH ¼ ADH SSADH ¼ Sulfolobus solfataricus ADH;
from horse liver; TBADH ¼ Thermoanaerobium READH ¼ Rhodococcus erythropolis ADH
brockii ADH; HSADH ¼ hydroxysteroid ADH; Adapted from Faber et al. [5, 7].
active), differing in 6 out of 374 amino acids [1, 8–12]. HLADH exhibits a broad
substrate tolerance towards primary and secondary alcohols combined with an
almost invariable (S)-stereoselectivity, making it a highly predictable and therefore
valuable tool for stereoselective oxidations. HLADH also exhibits appreciable stereo-
selectivity towards chiral centers other than the alcohol group being converted. A
series of racemic a-amino- and a-hydroxy alcohols were oxidized to the correspond-
ing aldehydes with good yield and enantioselectivity [13]. Low enantioselectivity is
observed with bulky substituents at the b-position (Table 32.1). Examples of the many
uses of HLADH as catalyst for desymmetrization of meso-diols are given in
Table 32.1 Selection of HLADH substrates.
Substrate Product E.e. Reference

(product) (%)
OH OH
>97 [13, 23]
R OH R O
R: HOCH2-, HalCH2-,
H2NCH2-, H2C ¼ CH-, Et
OH OH
OH O >10 [13, 23]
O OH O O
>10 [13, 23]
O O
OH OH
>10 [13, 23]
OH O
NH2 NH2
OH O 96 [13, 23]
HO HO
OH O
R R
18–30 [24, 25]
X X
R: Me, Et; X: CH2, O, S
OH OH
OH O
86 [26]
Fe Fe
1330
Section 32.2.6. The stereoselectivity of HLADH was rationalized as early as 1982

using a modified cubic-space section model [14]. Today, various crystal structures
allow for a more detailed understanding of the factors governing HLADH stereo-
selectivity [15–17]. In contrast to homologous ADHs such as the ADH from yeast,
HLADH is quite stable, especially towards molecular oxygen. In addition, application
in water-saturated organic media [18–20] or under near-dry conditions in gas-phase
reactions [21, 22] has been reported.
32.2.1.3 Yeast Alcohol Dehydrogenase (YADH)

Even though the amino acid sequences differ significantly from each other, YADH
shows a high degree of structural similarity to HLADH [1, 27]. Nevertheless, YADH
exhibits a far lower stability against thermal and oxidative stress and in the presence
of organic solvents [28, 29]. Its stability may be enhanced by immobilization, for
example, to magnetic nanoparticles [30], by liposomal encapsulation [31], or by
incorporation into alginates [32]. Regarding the substrate spectrum, YADH is
somewhat more limited, covering primary short-chain alcohols and small 2-alkanols
(Figure 32.3). However, because of its low cost, YADH has been studied thoroughly as
regeneration enzyme for reduced nicotinamide cofactors [1, 4, 33]. However, both
ethanol as cosubstrate and corresponding acetaldehyde negatively influence YADH
stability even at low concentrations. This may be circumvented by using ethanol/
acetaldehyde only in catalytic amounts and in situ regeneration of ethanol from
acetaldehyde using inorganic hydrides such as NaBH4 [34, 35].
32.2.1.4 ADHs from Thermophilic Organisms

Redox enzymes attractive for organic synthesis are increasingly derived from
thermophilic microorganisms [36]. Interest stems from their thermostability and
activity at high temperatures. Furthermore, these enzymes have also been found to be
resistant toward common protein denaturants and organic solvents.
For example the ADH from Thermoanaerobacter brockii (TBADH) is mostly
recognized because of its superb thermostability (good stability even at 85 C)
[37, 38] and resistance to a range of water-miscible organic solvents [28]; activity
was observed in the presence of up to 87% (v/v) methanol, ethanol, or acetonitrile [39].
The substrate scope of the NADP-dependent TBADH [37, 38, 40] is somewhat
complementary to HLADH and YADH (Figure 32.3), which are both unable to
efficiently convert linear secondary alcohols [1].
In addition to TBADH, a range of other ADHs originating from (hyper)thermo-
philic host organisms have been evaluated. A NAD-dependent ADH from Sulfolobus
solfataricus (SsADH) was described [41] that exhibits very good thermal stability and
activity towards aliphatic alcohols and benzyl alcohols. However, the enantioselec-
tivity (e.g., towards 2-alkanols) is rather modest [42]. Another interesting NADP-
dependent secondary ADH (SADH) was obtained from Thermus ethanolicus [43]. This
enzyme was thoroughly investigated by Phillips and coworkers, elucidating the
influence of various parameters (such a enthalpic and entropic effects) on the
enantioselectivity [44–53]. For example, inversion of SADH enantioselectivity from
(S) to (R) was observed at 27 and 75 C for 2-butanol and 2-pentanol, respectively [44].
Recently, a (S)-selective, NAD-dependent ADH from Thermus sp. ATN1 (TADH) was
reported [54–56]. The enzyme was produced recombinantly in Escherichia coli from
which it can be purified to near-homogeneity by a simple one-step heat treatment
procedure [56].
32.2.1.5 ADH from Rhodococcus ruber (ADH-A)

Recently, a secondary ADH from Rhodococcus ruber DSM 44541 (ADH-A) has
attracted considerable interest as an exceptionally solvent-stable ADH [57]. Ace-
tone/isopropanol concentrations of up to 50% (v/v) are tolerated, allowing for
substrate-coupled redox processes (Scheme 32.4 below). Another advantage of this
high solvent-tolerance is that the cosubstrate/coproduct at the same time serves as
solubilizer for less hydrophilic substrates, thereby enabling high space–time yields.
The thermal stability is comparably good [58, 59]. ADH-A converts a broad range of
(S)-2-alkanols with E-values up to 100; benzyl alcohols are also converted efficiently
and highly stereoselectively whereas cyclic alcohols are somewhat sluggish sub-
strates (Table 32.2).
Further, interesting applications of ADH-A have been reported such as the
oxidation of rhododendrol [natural (S)-4-(p-hydroxyphenyl)butan-2-ol] to raspberry
ketone, which was scaled up to up to 500 g l1, reaching 83% isolated yield [61].
The kinetic resolution of the non-natural racemate proceeded with very high
enantioselectivity.
32.2.1.6 Glycerol Dehydrogenases (GDHs)

Glycerol dehydrogenases (GDHs) have been isolated from various organisms such as
Schizosaccharomyces pombe [62] and Cellulomonas sp. [63, 64]. Compared to HLADH,
GDH exhibits a somewhat complementary selectivity. With glycerol as substrate
HLADH selectively oxidizes the primary alcohol functionality yielding (S)-glyceral-
dehyde, whereas GDH is specific for secondary alcohols, producing only dihydroxy-
acetone, a valuable intermediate, for example, for cosmetic application as self-
tanning agent.
In aqueous media GDH-catalyzed oxidation reactions suffer from pronounced
product inhibition as, for example, shown for the oxidation of phenyl-1,2-ethane-
diol [65]. This problem could be overcome by in situ extraction of the product into an
organic phase (Scheme 32.2) [66].
Scheme 32.2 Kinetic resolution of rac-phenylethandiol using GDH.

Kinetic resolutions using lyophilized Rhodococcus ruber expressing ADH-A [60].
1332
Table 32.2

OH OH O
A DH -A
+
in R. ruber
R R R
NAD+ NADH
OH O
A DH -A
in R. ruber
20% v/v
Substrate Activity (mmol g1 h1)a) E
OH
1540 1.3
OH
560 33
OH
1120–1400 >100
3-5
OH
420–840 34–43
6-7
OH
5 2.7
OH
133–573 31–100
R
OH
Low 8.3
OH
Low 4.5
32.2 Oxidation of Alcohols

a) Determined as activity per g cell wet weight.
j1333
1334
Table 32.3 Oxidation of 1,2-diols catalyzed by GDH [62].
Substrate Product Activity (U mg1)
OH O
2.9
HO OH HO OH
OH O 0.87
HO HO
OH O
4.64
OH OH
OH O
2.61
OH OH
OH O
1.74
OH OH
OH O
OH OH 1.45
OH OH
OH O
0.87
HO O HO O
Further applications of GDH as catalyst for the kinetic resolution of some 1,2-diols
were reported as early as 1985 (Table 32.3) [62, 67]. GDH contains catalytically
relevant and autoxidizable thiol functionalities and, thus, anaerobic conditions are
mandatory.
32.2.1.7 Other ADHs

The number of novel ADHs described is constantly growing (http://www.brenda-
enyzmes.org/) and a description of all of them is clearly beyond the scope of this
chapter. Some prominent examples, however, are worth mentioning here. An ADH
from Lactobacillus kefir (LkADH) with a broad substrate range has attracted some
interest (however, mainly as reduction catalyst) [68–70]. Lactate dehydrogenase
(LDH, EC 1.1.1.27) converts a-hydroxy acids into the corresponding a -keto acids [71].
Owing to its rather limited substrate/product scope it is mostly used for cofactor
regeneration [72]. Some specific examples using LDH for the production of enan-
tiopure L-lactate are given later in this chapter. Carbohydrate dehydrogenases could
be valuable catalysts for the regioselective oxidation of polyols such as sugars. A wide
range of various DHs is available, for example, glucose DH (EC 1.1.1.47), mannitol
DH (EC 1.1.1.67), and fructose DH (EC 1.1.1.124). However, apart from some
examples discussed in this chapter, their importance for organic synthesis is rather
limited. Glucose dehydrogenase was used for the preparation of D-gluconic acid
using electrochemical NAD þ regeneration [73, 74]. Most applications of this
oxidation reaction, however, concentrate on NADH regeneration [75–78]. Likewise,
glucose-6-phosphate DH (G6PDH, EC 1.1.1.49) is a common NADPH regeneration
system [4, 79]. Hydroxysteroid dehydrogenases (HSDHs) represent another class of
synthetically useful catalysts. For virtually any given OH functionality in the steroid
backbone a specific HSDH is available for selective oxidation. In addition, many
HSDHs are also highly enantioselective, thereby discriminating between a- and
b-configured OH groups [80–86]. For example, cholic acid can be specifically oxidized
at 3-, 7-, or 12-position using the specific HSDH (Scheme 32.3).
Scheme 32.3 Regioselective oxidations of cholic acid using specific HSDHs.
By combination of stereo-complementary HSDHs Riva and coworkers performed

the stereoinversion at the 3-position of various bile acids [86] and of cholic acid into
12-ketoursodeoxycholic acid [81].
32.2.1.8 NAD(P) þ Regeneration Systems

Thermodynamically, the NAD(P) þ -dependent oxidation of alcohols is an unfavor-
able process. As a result the equilibrium usually lies far on the side of alcohols and
oxidized nicotinamide cofactor. Therefore, and due to the still high cost of the
nicotinamide cofactor, in situ regeneration of its oxidized form has to be applied. A
plethora of different regeneration approaches have been reported, which are dis-
cussed extensively in the recent review literature [4, 5, 87–92]. The different
regeneration methodologies can be classified based on the terminal electron acceptor
(cosubstrate) used (Table 32.4). Next to the ease of use and the price and availability of
the cosubstrates and regeneration enzymes, the thermodynamic driving force they
exert on the overall reaction and environmental factors like the amount and sort of
waste generated should also be taken into account. A common way to assess the
efficiency of a regeneration system is based on the total turnover number (TTN)
1336
Table 32.4 Common NAD(P) þ regeneration approaches.
Cosubstrate Coproduct Catalyst Waste DE 0 0 (mV)b)

(g mol1)a)
Acetone Isopropanol ADH 58 34

Pyruvate Lactate Lactate-DH 90 135
a-Ketoglutarate Glutamate Glutamate-DH 146 199
O2 H2O2 NADH oxidase 34 1136
O2 H2O NADH oxidase 18 1550
Electrochemical Anode/mediator — Variable
a) Calculated as mass of coproduct per mol of NAD(P) þ regenerated.

b) Relative to the NAD(P)H/NAD(P) þ redox couple (DE00 ¼ 320 mV versus NHE).
achieved for the nicotinamide cofactor. As a rule of thumb, a TTN of more than 1000
is considered to be sufficient for economic feasibility [93].
In the following, the most common NAD(P) þ regeneration approaches are briefly
discussed.
The substrate-coupled approach exploits the reversibility of ADH-catalyzed redox
reactions by utilizing the production enzyme also as regeneration enzyme
(Scheme 32.4). This approach can be considered as a biocatalytic variant of the
Oppenauer oxidation. It is the most widely used approach.
Scheme 32.4 Substrate-coupled regeneration approach. In this example acetone is used as

cosubstrate.
At first sight, this approach is most advantageous as only one enzyme is needed for
a functional production system. Furthermore, the nicotinamide cofactor does not
have to leave the ADHs active site to be regenerated and can stay immobilized (and
stabilized) within the enzyme. However, the high chemical similarity of substrate and
cosubstrate results in a very low thermodynamic driving force. As a result, usually
high molar surpluses are necessary to drive the desired reaction to completion. Often,
this leads to heavy stability losses of the enzyme. However, Kroutil, Faber, and
coworkers reported on an exceptionally chemotolerant ADH from Rhodococcus ruber
DSM 44541 that endures acetone concentrations of up to 50% (v/v) [58, 59, 94]. Thus,
the cosubstrate and coproduct also served as solubilizer for the rather hydrophobic
substrates of interest. The same group also recently reported a very elegant solution to
the driving force problem [95]. By cosubstrate and -product engineering they
eliminated the necessity for high molar surpluses. Suitably a-substituted (with H-
bond acceptors) cosubstrates such as acetoacetate or chloroacetone form thermo-
dynamically stable coproducts, thereby making the regeneration reaction practically
irreversible.
The enzyme-coupled approach separates the desired production reaction from the
regeneration reaction by coupling two enzymatic reactions via their opposing
demand for NAD(P) þ and NAD(P)H, respectively (Scheme 32.5). Additional ther-
modynamic driving force can be added to the system by suitable choice of the
cosubstrate/coproduct redox couple (Table 32.4). Prominent examples are glutamate
dehydrogenase (GluDH) and lactate dehydrogenase (LDH), which were used as early
as 1985 [13] to promote ADH-catalyzed oxidation reactions. On the downside,
however, significant amounts of (non-volatile) cosubstrates and coproducts (wastes)
are involved that can complicate product isolation. In addition, the atom-efficiency of
this approach is rather poor [96].
Scheme 32.5 Enzyme-coupled regeneration of NAD(P) þ using glutamate dehydrogenase

(GluDH), lactate dehydrogenase (LDH), NAD(P)H oxidase (NOx), or the laccase-mediator system
(LMS) as regeneration systems.
In recent years, several water- or hydrogen peroxide-forming NAD(P)H oxidases

(NOx) have been reported. [97–105] The latter systems combine attractive thermo-
dynamic driving force with the generation of only water (or hydrogen peroxide, which
can be easily dismutated to water and molecular oxygen) as by-product. Thus, exciting
new developments for this approach may be expected in the future. Recently, an
1338
aerobic NAD(P) þ -regeneration system based on the so-called laccase-mediator

system (LMS) was also reported [106], coupling the chemical NAD(P)H oxidation
catalyst 2,20 -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid (ABTS) to laccase-cata-
lyzed regeneration of its catalytically active, oxidized form. Thus, an efficient, aerobic
and water-forming regeneration system for both NAD þ and NADP þ was
established.
The so-called intrasequential regeneration represents a special case of the enzyme-
coupled approach. Both NAD(P)H and the carbonyl product of the ADH-catalyzed
oxidation are used for a subsequent oxidation. Scheme 32.6 shows a prominent
example [107]. Another elegant example of a substrate-coupled, intrasequential
regeneration was reported by Tanaka and coworkers [108].
Scheme 32.6 Intrasequential cofactor regeneration [107].
32.2.1.9 Miscellaneous (Non-enzymatic Approaches)

Despite their great success in preparative application, enzymatic NAD(P) þ regen-
eration systems bear a couple of intrinsic disadvantages. First, a given regeneration
enzyme is usually highly specific for either the phosphorylated (NADP) or the non-
phosphorylated (NAD) cofactor, thus lacking general applicability. Second, optimal
operational windows regarding temperature, pH, ionic strength, and so on of both
production and regeneration enzyme may differ significantly, leading sometimes to
poor compromises. In addition, the different robustness of both enzymes can
represent a significant challenge. Thus, a range of non-enzymatic regeneration
approaches have also been proposed in recent years.
Probably the earliest example for a non-enzymatic NAD(P) þ regeneration system
consists of simple flavins functioning as hydride acceptors from the reduced
nicotinamide cofactors [109]. The resulting reduced flavins are quickly reoxidized
in the presence of molecular oxygen, thus yielding an essentially irreversible NAD
(P) þ regeneration system. Its efficiency, however, is limited due to the rather
sluggish hydride transfer kinetics from NAD(P)H to the oxidized flavin. The catalytic
efficiency of the flavin catalyst expressed as turnover frequency (TF) lies in the range
of a few catalytic cycles per day [110–113]. This can be accelerated by several orders of
magnitude, for example, by co-catalysis using [Cp Rh(bpy)(H2O)]2 þ [114], which has
been employed in a double-kinetic resolution of 3-methyl cyclohexanol [56].
Electrochemical regeneration of oxidized nicotinamide cofactors also enjoys some
popularity [90–92, 115, 116]. Especially, the reagentless nature of the regeneration
reaction bears promise as an economically feasible and environmentally benign
alternative to the established regeneration approaches. The simplest approach
consists of direct anodic oxidation of NAD(P)H [74, 117, 118]. However, quite large
overpotentials in the range of 1 V are required to obtain significant turnover at bare
electrodes. Unfortunately, the number of enzymes, substrates, and products that can
withstand such strong oxidizing conditions is limited. Modified electrodes exhibiting
lower overpotentials are often reported [91, 119].
The high overpotentials needed for NAD(P)H oxidation can be considerably
lowered by the use of redox mediators. In particular, ortho- and para-quinones and
their derivates [120–130] undergoing two-electron transfer processes were found to
be ideal for NAD(P) þ regeneration. Among these, 1,10-phenanthroline-5,6-
diones [125, 126, 131] are probably the most potent mediators.
A very elegant example of mediated electrochemical regeneration of NAD þ was
reported by Willner and coworkers [132]. NAD was covalently attached to a gold
electrode via PQQ linkers that also served as electrical contact between the nicotin-
amide cofactor and the anode. The surface-exposed NAD moieties could bind LDH,
which after chemical crosslinking formed a stable layer of active LDH around the
electrode. This electrode–LDH assembly was used for the amperometric quantifi-
cation of lactate (Scheme 32.7).
Scheme 32.7 Direct electrochemical wiring of LDH to an anode [132].

1340
Figure 32.4 Typical examples of one- and two-electron mediators used in indirect electrochemical
NAD(P) þ regeneration.
Though originally designed for biosensor applications, this approach might also be
very useful for bioelectrochemical synthesis. Furthermore, quinoid mediators can be
generated on the surface of carbon electrodes by oxidative pretreatment [133]. Besides
these hydride acceptors, single-electron-transfer mediators (e.g., transition metal
complexes [127, 134], viologens, [135] heteropolyanions [136], conducting poly-
mers [137], or ABTS [66, 106, 138]) can also oxidize NAD(P)H.
Figure 32.4 shows examples of one- and two-electron acceptors. Many of the
quinone-based mediators react in their reduced states with molecular oxygen. This
aerobic regeneration has the advantage that no additional electrochemical equipment
is necessary to perform NAD(P) þ regeneration. On the other hand, reactive oxygen
species are generated, which might inactivate enzymes and which therefore need to
be removed from the reaction mixture.
Finally, photochemical approaches for the regeneration of NAD(P) þ are worth
mentioning [139]. Mechanistically, they are related to their electrochemical counter-
parts. The so-called photosensitizer mediates electron transfer from NAD(P)H to the
terminal electron acceptor. Photochemical methods are based either on the light-
induced excitation of a mediator, enabling it to oxidize NAD(P)H (reductive quench-
ing mechanism), or on the light-induced excitation of the already reduced mediator,
thus facilitating its reoxidation (oxidative quenching mechanism).
For reductive quenching, examples of suitable photosensitizers are tin porphyr-
ins [140] and methylene blue [141, 142]. Ruthenium(II) tris(bipyridine) complexes in
combination with viologens are used for oxidative quenching. After the oxidation of
NAD(P)H, the reduced Ru complex is excited by light. The resulting powerful
reducing agent transforms methyl viologen into the radical cation. The electrons
from NAD(P)H are usually transferred to molecular oxygen, protons, or the
anode [140, 143].
As well as soluble photosensitizers, semiconductors have been reported for NAD þ
regeneration [20, 144]. The advantage of these photochemical systems is that some of
them utilize visible light, pointing towards the possibility of using sunlight to drive
organic reactions. Disadvantageous, however, are the still low performances (TTN
and TF of the photosensitizers and coenzymes) and the fact that photoexcitation
results in the formation of strong oxidizing agents and the formation of free reactive
radicals. Therefore, photochemical regeneration has not become one of the standard
procedures, yet.
32.2.2
NAD(P)-Independent Dehydrogenases
In addition to the aforementioned nicotinamide-dependent ADH, in recent years

non-NAD(P)-dependent dehydrogenases have also enjoyed increasing interest. This
smaller group of dehydrogenases does not rely on NAD(P) þ as electron acceptor but
rather transfer the reducing equivalent liberated from the substrates to quinoid
electron acceptors. Prominent examples are the quinohemoprotein dehydrogenases
(QH-ADHs) [145]. QH-ADHs contain pyrroloquinoline quinone (PQQ) or related
prosthetic groups (Figure 32.5) [146–150].
Although the substrate specificity fingerprints show appreciable overlap with those
of NAD(P)-dependent alcohol dehydrogenases, their natural electron acceptors,
ubiquinone [151–154], cytochrome c, or blue copper proteins (such as azurin) [155,
156], are towards the oxidative end of the biological redox scale. Consequently,
Figure 32.5 Simplified quinoid-based oxidation mechanism of QH-ADH and some representative
prosthetic groups present in QH-ADHs.
1342
dehydrogenation reactions catalyzed by quinoprotein dehydrogenases are strictly

irreversible [145].
Type I QH-ADHs are soluble, monomeric proteins containing one molecule of
PQQ, Ca2 þ , and a single c-type heme. Type II QH-ADHs are membrane-associated
hetero oligomers. The large subunit exhibits the ADH activity and contains one PQQ,
one Ca2 þ , and one heme [157]. The medium-sized subunit contains three heme
moieties and is involved in electron transport; the small subunit is speculated to
mediate proper association of the two larger ones. The PQQ-catalyzed oxidation
mechanism is largely under debate – simple hydride abstraction, covalent intermedi-
ates between the alcohol, and the PQQ moiety or ene-reactions are discussed [145].
The so-called flavo-dehydrogenases make up yet another class of NAD(P) þ -
independent alcohol dehydrogenases [158]. Structurally, these enzymes are related
to the flavin-dependent oxidases (vide infra). In both cases, a reduced flavin is the
primary product of the enzymatic oxidation reaction. The oxidases directly utilize
molecular oxygen as terminal electron acceptor. Dehydrogenases in contrast transfer
the reducing equivalents to an intermediate electron acceptor and thereby make them
available for microbial endoxidation. Recently, van Berkel and coworkers demon-
strated the subtle differences between flavo-dehydrogenases and -oxidases by trans-
forming a dehydrogenase into an oxidase by just one amino acid exchange (alanine to
glycine) in the active site [159]. One example is cellobiose dehydrogenase (CDH, EC
1.1.99.18) an extracellular flavo-heme DH produced by wood-degrading fungi [160].
It oxidizes soluble cellodextrins, mannodextrins, and lactose efficiently to their
corresponding lactones using a wide spectrum of electron acceptors including
cytochrome c, quinones, phenoxy radicals, ABTS, Fe3 þ , Cu2 þ , and I3. Molecular
oxygen is a poor electron acceptor. L-Galactono-1,4-lactone dehydrogenase (GALDH)
is another example for this family [158]. In plants it completes the vitamin C synthetic
pathway by oxidizing L-galactono-1,4-lactone to ascorbic acid endoxidation [161].
Other interesting examples such as p-cresol methylhydroxylase (PCMH, EC
1.17.99.1) are discussed later in this chapter.
The genus Gluconobacter deserves special attention due to its importance for
biocatalytic oxidations. Its unique capacity to incompletely oxidize polyol substrates
has led to numerous whole-cell production processes for the synthesis of compounds
such as vitamin C (regioselective oxidation of D-sorbitol to L-sorbose as key step in the
so-called Reichenstein process) [162], (keto)gluconic acids, dihydroxyacetone [163],
vinegar [100], the synthesis of L-ribulose [164], and the synthesis of 6-deoxy-6-
butylamino sorbose [165]. Many other applications are discussed throughout
this chapter.
So far, this versatility is unsurpassed among the microorganisms known. Two
major groups of oxidoreductases in Gluconobacter (and acetic acid bacteria in general)
confer this versatility. The first group consists of membrane-associated dehydro-
genases with very strict substrate specificity, referred to as the Bertrand–Hudson
rule [100]. Table 32.5 shows the great variety of these DHs. These DHs are
cytochrome-related, thus channeling the reducing equivalents liberated from the
substrate via the membrane-associated electron transport chain to endoxidation.
However, artificial electron acceptors may also take this role.
Table 32.5 Membrane-bound dehydrogenases purified from Gluconobacter sp. [100].
Dehydrogenase Prosthetic group Substrates Reference
Alcohol-DH FAD, CytC Primary, straight chain [166]

alcohols
Aldehyde-DH PQQ, CytC Primary, straight chain [167]
aldehydes
D-Glucose-DH PQQ D-Glucose, maltose [168]
D-Gluconate-DH FAD (covalent) D-Gluconate [169]
2-Keto-D-gluconate-DH FAD (covalent), CytC 2-Keto-D-gluconate [170]
D-Sorbitol-DH (1) FAD (covalent), CytC D-Sorbitol, D-mannitol [171]
D-Sorbitol-DH (2) PQQ, CytC D-Sorbitol, D-ribitol [172]
D-Sorbitol-DH (3) PQQ D-Sorbitol, D-mannitol, glycerol [173]
L-Sorbose-DH NADP Sorbose [174]
Glycerol-DH PQQ D-Sorbitol, D-mannitol, glycerol [175]
D-Arabitol-DH PQQ D-Arabitol [176]
myo-Inositol-DH PQQ myo-Inositol [177]
Major polyol-DH PQQ D-Sorbitol, D-mannitol, glycerol, [178]
ribitol, 5-keto-D-gluconate
The second group contains classical, NAD(P)-dependent polyol dehydro-

genases found in the cytosol. Interestingly, some of these DHs exhibit different
optimal pH values for different substrates (Table 32.6) [100]. For in vivo regener-
ation of the oxidized nicotinamide cofactor, a H2O2-producing Old Yellow Enzyme
is suspected.
Table 32.6 NAD(P)-dependent sugar alcohol dehydrogenases.
Dehydrogenasea) Cofactor Substrates Reference
MDH-1 NADP D-Mannitol [179]

D-Fructose
MDH-2 NAD D-Mannitol [179]
D-Fructose
D-Arabitol
SDH-1 NAD D-Sorbitol [179]
D-Fructose
Xylitol
SDH-2 NADP D-Sorbitol [179]
L-Sorbose
RDH Ribitol [176]
Xylitol
L-Arabitol
D,L-Ribulose
a) MDH: mannitol-DH; SDH: sorbitol DH; RDH: ribitol DH.

1344
32.2.2.1 Regeneration of NAD(P)-Independent Dehydrogenases

In contrast to the above-discussed ADHs, NAD(P)-independent dehydrogenases
transfer the reducing equivalents liberated during the oxidation process to small (in)
organic cofactors as primary electron acceptors. The latter transfer the excess
electrons to the endoxidation. Under cell-free conditions, this electron-transport
chain can be shortcut to more or less directly transfer the reducing equivalents to
artificial electron acceptors. This has been performed extensively with p-cresol
methylhydroxylase (PCMH, EC 1.17.99.1). PCMH is a flavo-heme dehydrogenase
acting by a dehydrogenative mechanism on p-alkyl substituted phenols. The
actual dehydrogenation reaction is catalyzed by the flavin prosthetic group that then
transfers the electrons via an enzyme-bound heme to the natural electron acceptor
azurin [180].
Using the anode as terminal electron acceptor and ferrocene as mediator, Steckhan
and coworkers [181, 182] as well as Hill and coworkers [180] could quantitatively
transform p-cresol into the corresponding aldehyde (Scheme 32.8).
A continuous production system using the so-called electrochemical enzyme
membrane reactor was setup, reaching superb TTNs for both the enzyme and the
ferrocene mediator.
Scheme 32.8 Indirect electrochemical regeneration of PCMH using a water-soluble polymeric

ferrocene mediator [181, 182].
Though electrochemical regeneration is quite promising from an atom-economy

point-of-view, its broad applicability is hampered by the need for specialized
electrochemical equipment. A simpler, fully enzymatic regeneration system utilizing
the so-called laccase-mediator-system (LMS) was developed by Haltrich and cow-
orkers. Its practical usefulness was demonstrated by the example of cellobiose
dehydrogenase (CDH, EC 1.1.99.18) (Scheme 32.9) [183–187].
Quite a wide variety of mediators have proven to be efficient in this system, ranging
from single-electron mediators such as 2,20 -azino-bis(3-ethylbenzothiazoline-6-sul-
fonic acid) (ABTS), ferrocene, and so on to two-electron mediators such as benzo-
quinones and redox dyes such as Meldolas blue. In principle, the catalytically active,
oxidized forms of these mediators can also be regenerated by anodic oxidation. The
particular advantage of the LMS, however, is its apparent simplicity as well as the
Scheme 32.9 Use of the laccase-mediator system (LMS) for oxidative regeneration of cellobiose
dehydrogenase (CDH) [183–187].
circumvention of diffusion limitations that frequently hamper its electrochemical

counterparts.
32.2.3
Alcohol Oxidases
Oxidases are found in all kingdoms of life [188, 189]. Their natural role is not always
clear, especially from an energetic point-of-view it does not seem to be advantageous
to waste reducing equivalents liberated in the oxidation of alcohols such as
carbohydrates by circumventing endoxidation and direct transfer to O2, producing
hazardous hydrogen peroxide [188].
Their natural role seems to be the production of hydrogen peroxide, which then is
used by likewise excreted peroxidases to generate aromatic radicals for lignin
depolymerization [190].
Oxidases utilize molecular oxygen as terminal electron acceptor. This can be
considered as a direct aerobic regeneration of the prosthetic group. At first glance,
this seems a simpler oxidation procedure compared to the coenzyme dependant
dehydrogenases or monooxygenases. However, with few exceptions, such as some
NADH oxidases [103, 104] or laccases [191], which reduce molecular oxygen directly
to water in an overall four-electron transfer step, O2 reduction generally leads to
hydrogen peroxide (transfer of two electrons) or to the superoxide radical anion
(transfer of one electron) as primary reduction products. The occurrence of reactive
oxygen species (ROS) generally leads to oxidative inactivation of the enzymes, thereby
diminishing the efficiency of the whole production system. As a consequence,
significant research efforts have been devoted to avoid or at least diminish the
occurrence of ROS. Some of these approaches are summarized in the following.
32.2.3.1 Methods to Diminish/Avoid Hydrogen Peroxide

Auto-regeneration of oxidases with concomitant catalase-catalyzed disproportion-
ation of hydrogen peroxide is a simple and effective regeneration method
(Scheme 32.10); it is quite commonly used with oxidase reactions.
Hydrogen peroxide is highly reactive and irreversibly inhibits enzyme activity (also
catalase) even in low concentrations [192, 193]. Hydrogen peroxide can be avoided if
excess electrons are transferred to an electron acceptor other than molecular oxygen.
1346
Scheme 32.10 Alcohol oxidase (AlcOx)-catalyzed oxidations with concomitant catalase-catalyzed

disproportionation of H2O2.
This is realized by indirect electrochemical regeneration [181, 194, 195] and aerobic
regeneration utilizing the laccase mediator system (LMS) [183, 196–198] will be
shortly discussed here (Scheme 32.11).
Scheme 32.11 Hydrogen peroxide-free regeneration of alcohol oxidases (AlcOxs) either via the
laccase mediator system (LMS) or via indirect electrochemical regeneration.
Direct electron transfer between enzymes and electrodes is usually very slow,
because the enzymatic active sites are often deeply buried within the protein shell and
therefore inaccessible for the electrode (the tunneling probability of electrons is a
function of distance). To accelerate the electron transfer, low molecular weight redox
active substances (mediators) are used to shuttle the electrons between the enzyme
and the electrode. Interestingly, there is a significant overlap of the mediators used for
electrochemical regeneration and for the LMS.
Widely used mediators are, for example, ferrocenes [181, 195, 199–201], but also
bipyridine/phenanthroline, terpyridine, or hexacyano complexes are employed [202].
In addition, quinoid salts like TTF/TCNQ (tetrathiafulvalene/tetracyanoquinodi-
methane) [203] as well as benzoquinones [194] and redox dyes like phenazine and
phenothiazine derivatives (MPMS, thionin, azure A, and azure C) [204, 205] proved to
be useful redox agents for the indirect electron transfer. Even incorporation of
oxidases into conducting polymers made of polypyrrole or polythiophene derivatives
proved to function for electrochemical regeneration [206]. Likewise, co-immobili-
zation of GOx (glucose oxidase) with gold nanoparticles to electrodes proved to be a
very efficient approach to wire oxidases to electrodes [207]. Notably, most research
in the field of electrochemical oxidase regeneration concentrates on analytical
applications, inspired by the search for electrochemical biosensors.
However, it was demonstrated that indirect electrochemical methods are suitable
for prolonging oxidase operational stability [194]. For example, glucose oxidase (GOx,
EC 1.1.3.4) was immobilized on a carbon felt anode and regenerated with the
benzoquinone/hydroquinone redox couple (Scheme 32.12).
Scheme 32.12 Indirect electrochemical regeneration of glucose oxidase (GOx) using the
benzoquinone/hydroquinone redox couple.
Compared to aerobic regeneration the operational stability of GOx could be

increased at least 50 times. Productivities as high as 100 g l1 h1 have been reached.
The preparative usefulness of the LMS for H2O2-free regeneration of oxidases was
impressively demonstrated with pyranose oxidase (P2O, EC 1.1.3.10) [183, 197, 198].
Interestingly, P2O shows higher affinity for some mediators than for O2 (1,4-
benzoquinone KM value 120 mM compared to 650 mM for O2) with otherwise
comparable activities, yielding a six-times higher kcat/KM value. Preparative scale
biotransformations could be performed with twofold volumetric productivities. The
TTN were 1.1 106 and 800 for P2O and 1,4-benzoquinone, respectively, with a
residual enzyme activity of 85%.
32.2.3.2 Common Oxidases

Aliphatic alcohol oxidases (EC 1.1.3.13) are produced by methylotropic yeast such as
Candida, Hansenula, and Pichia (http://www.brenda-enyzmes.org/). Most popular
(and commercially available) oxidases are from Pichia pastoris and Candida boidinii.
Mostly primary alcohols are accepted and oxidized selectively to the aldehyde
stage [208–211]. Some interesting applications of the Pichia oxidase for the selective
oxidation of ethylene glycol into glycolaldehyde as well as for the in situ generation of
reactive aldehydes are discussed in Section 32.2.6 [212, 213]. In addition, aromatic
alcohol oxidases (EC 1.1.3.7) are known [214]. Overall, however, synthetic applica-
tions of the aforementioned aliphatic and aromatic alcohol oxidases are rather scarce.
Glucose oxidase (GluOx, EC 1.1.3.4) is probably the most prominent oxidase
known so far [215]. The dimeric flavoenzyme catalyzes the oxidation of b-D-glucose to
D-glucono-d-lactone, which spontaneously hydrolyzes to gluconic acid. Wild-type
1348
GluOx is highly specific for b-D-glucose. Other substrates such as D-maltose, D-xylose,
or L-sorbose are converted at only a fraction of the rate observed for glucose [216, 217].
Thus, on the one hand, the preparative usefulness of GluOx is rather limited. On the
other hand, this high specificity makes GluOx an ideal biosensor for glucose in
complex mixtures such as blood [218], fermentation broths [219], and beverages [220].
The analytical signal is either based on the stoichiometrically formed hydrogen
peroxide (typically in combination with a peroxidase/redox dye combination) or,
more elegantly, by contacting GluOx to an anode to generate an amperometric signal)
(Scheme 32.13).
Scheme 32.13 GluOx-based glucose sensors: (A) aerobic regeneration of oxidized GluOx with
concomitant spectrophotometric H2O2-quantification, (B) anaerobic, mediator-based regeneration
and amperometric quantification.
For the electrochemical biosensors, suitable mediators establishing the electrical

contact between GluOxs active site (reduced FAD) and the anode quinones [194],
conducting polymers [206], crosslinked GluOx-nano-gold particles [207], or transi-
tion metal complexes [200, 201] have been used. With chiral ferrocenes, even a certain
degree of enantiodiscrimination was observed [221].
In addition, in food and beverages GluOx is often used to remove either glucose
(responsible for Maillard-based color loss of the products) or O2 to increase the
product storage stability [215].
Furthermore, the controllable H2O2 formation by GluOx is frequently exploited
for the in situ dosage of H2O2 to promote peroxidase-catalyzed reactions
(Section 32.2.4) [222–224].
Galactose oxidase (GalOx, EC 1.1.3.9) is a single copper metalloenzyme that
catalyzes the oxidation of primary alcohols to corresponding aldehydes with strict
regioselectivity [225–227]. In its natural substrate galactose it specifically oxidizes the
6-OH group. Additionally, GalOx also accepts a range of unnatural substrates
(Table 32.7) [228, 229]. Glucose is not converted by wild-type GalOx [230].
Table 32.7 Selection of GalOx-substrates.
Substrate Product Reference
O OH O OH
HO O
HO OH HO OH
OH OH
D-Galactose meso-galacto-Hexodialdose
O O
HN HN
O O O O
O N O P O P O O O N O P O P O O CO 2H [231]
OH
OH OH OH OH
HO OH HO OH
HO OH HO OH
OH OH
UDP-[14C]galactose UDP-[14C]galacturonic acid
OH OH
HO HO [228]
OH O
OH OH
D-Threitol D-Threose
(Continued )
j1349
1350
Substrate Product Reference
OH OH OH OH
HO OH HO O [228]
OH OH
Xylitol L-Xylose
OH OH OH OH
OH O [228]
HO HO
OH OH OH OH
L-Glucitol L-Glucose
OH OH OH OH
HO HO [228]
OH O
OH OH OH OH
L-Galactitol L-Galactose
OH OH
[229]
HO OH O OH
L-(-)-Glyceraldehyde
OH OH OH
[225, 229]
HO Cl HO Cl O Cl
(S)-Halodiol þ (R)-aldehyde
O O O O
HO O
HO OH HO OH
OH OH
OH OH
HO OH HO OH
O O O O
HO O O OH HO O O O
OH OH
HO OH HO OH
O O
HO OH HO OH
OH OH
j1351
1352
The high selectivity of GalOx (and sugar oxidases in general) can be exploited for the
highly selective functionalization of complex polyols without the need for protection
chemistry [225, 232, 233]. Regeneration of the enzyme can be performed either
aerobically or indirect electrochemically using, for example, ferrocene mediators [195].
Pyanose-2-oxidase (P2O, EC 1.1.3.10) oxidizes sugars abundant in lignocellulose,
such as D-glucose, D-galactose, and D-xylose, to the corresponding 2-keto sugars
(osones). The essential structural requirement for P2O substrates is an equatorially
oriented 2-OH group in the carbohydrate-pyranoid form. Table 32.8 shows a selection
of P2O substrates [196, 234, 235].
Even though P2O exhibits a very high intrinsic stability (storage stability) it is very
labile towards H2O2, thereby limiting its operational stability [197, 236], which cannot
be fully circumvented even in the presence of a huge (1000-fold) molar excess of
catalase over P2O. This challenge can be (partially) overcome by the so-called laccase-
mediator system (LMS, Scheme 32.9). Using 1,4-benzoquinone as artificial electron
acceptor, which is reduced to the corresponding 1,4-catechol, H2O2-formation is
circumvented. The catechol is reoxidized by means of an O2-dependent, water-
forming laccase. However, in case of P2O there is significant competition from O2
and 1,4-benzoquinone for reoxidizing P2O-bound FADH2 [197]. As a consequence,
H2O2 cannot be avoided entirely, which still necessitates the co-administration of
catalase to dismutate trace amounts of H2O2 [198]. Protein engineering may be used
to increase the affinity of P2O towards artificial electron acceptors [237].
Application of P2O in the so-called Cetus process will be discussed in
Section 32.2.7. Furthermore, the use of P2O as entry reaction for the conversion
of unprotected sugars facilitates the development of convenient reaction routes for
the synthesis of rare sugars, sugar-derived synthons, and fine-chemicals [190, 238].
Cellobiose dehydrogenase (CDH) has been thoroughly investigated by Haltrich
and coworkers for the conversion of lactose into lactobionic acid. In particular, non-
aerobic regeneration of CDH proved to be efficient in avoiding H2O2 production and,
thereby, in stabilizing the enzyme. Laccase-mediator based [183–187] and electro-
chemical approaches [239] have been reported (Scheme 32.14).
Scheme 32.14 Artificial regeneration systems to promote CDH-catalyzed lactobionic acid

formation using either the laccase-mediator system (LMS, upper) or anodic reoxidation.
Table 32.8 A selection of P2O substrates [235].
Substrate Product Yield (%) Activity (%)
Oxidation of monosaccharides
O OH O OH
HO HO
100 100
HO OH HO O
OH OH
D-Glucose D-Glucosone
O OH O OH
HO HO
94 40
HO OH HO O
OH OH
D-Allose D-ribo-Hexos-2-ulose
O OH O OH
HO HO
70 8
HO OH HO O
OH OH
D-Galactose D-lyxo-Hexos-2-ulose
O OH O OH
5 Very low
HO OH HO O
OH OH
D-Ribose D-erythro-Pentos-2-ulose
O OH O OH
HO HO
100 96
HO OH HO O
3d-D-Glucose 3d-D-erythro-Hexos-2-ulose
O OH O OH
100 92
HO OH HO O
OH OH
6d-D-Glucose 6d-D-ribo-Hexos-2-ulose
O O
HO HO
100 75
HO OH HO O
OH OH
1,5-Anhydro-D-glucitol 1,5-Anhydro-D-fructose
1354
Recently, Fraaije and coworkers reported on a novel carbohydrate (alditol) oxidase

from Streptomyces coelicolor (AldO) (Table 32.9) [240, 241]. AldO is a 45-kDa soluble
monomeric flavoprotein containing a covalently bound FAD cofactor. Fusion to the
maltose-binding protein allowed efficient expression of the recombinant protein in E.
coli at levels up to 350 mg l1 culture [242]. AldO selectively converts primary OH
groups of a broad range of aliphatic and aromatic 1,2-diols in a kinetic resolution
preferring the (R)-configuration at the secondary OH group [241, 243]. vic-Diols are
over-oxidized to the acid stage, thereby giving access to enantiomerically pure
a-hydroxy acids. In contrast, non-vic-(di)ols are selectively oxidized to the
aldehyde stage.
Glycolate oxidase (GlyOx, EC 1.1.3.15) is a peroxisomal enzyme found in the leaves
of many green plants (e.g., spinach) or in mammalian liver. The partially purified
enzyme from spinach has been applied for the stereoselective (S)-oxidation of
a-hydroxy carboxylic acids [244, 245]. In addition to a-hydroxy carboxylic acids such
as lactic acid [246], GlyOx also accepts various homologues of lactic acid [244, 245].
Furthermore, 1,2-diols such as ethylene glycol are converted [247–249].
Cholesterol oxidase (ChOx, EC 1.1.3.6) is a FAD-dependent enzyme isolated from
various microbial sources [250]. In most cases ChOx catalyzes the selective oxidation
of cholesterol (cholest-5-en-3b-ol) to form cholest-4-en-3-one (Scheme 32.15) [251,
252]. ChOX from Rhodococcus erythropolis was applied for the kinetic resolution of
racemic mono- and bicyclic allyl alcohols [253]. Although the substrates tested were
much smaller than the native substrate, reasonable enantioselectivities (E) in the
range of 7–20 were found for the (S)-alcohols.
Scheme 32.15 ChOx-catalyzed oxidation of cholesterin [251, 252].
Other applications regarding selective conversion of steroids have been

reported [252, 254–256].
Analytical applications (amperometric biosensors) have been reported utilizing
organic redox dyes such as phenazine and phenothiazine or gold nanoparticles [204,
257]. In addition to these most common oxidases, the portfolio of novel interesting
biocatalysts is growing steadily [158].
32.2.4
Peroxidases
Peroxidases (EC 1.11.1.7) are ubiquitously in plants, microorganisms, and animals.

They depend on H2O2 to regenerate their oxidized, catalytically active prosthetic
Table 32.9 Substrate range of AldO (alditol oxidase) [241, 243].
Substrate Structure KM kcat kcat/KM

(mM) (s1) (s1 M1)
OH
Glycerol 350 1.6 4.6
HO OH
OH
L-Threitol HO 25 6.3 250
OH
OH
OH
Xylitol HO OH 0.32 13 41 000
OH OH
OH OH
D-Sorbitol
OH 1.4 17 12 000
HO
OH OH
OH OH
D-Mannitol OH 36 9.2 260
HO
OH OH
OH
L-Arabinose 430 1.7 4
O OH
OH OH
OH OH
D-Galactose
O — — 0.3
HO
OH OH
OH
1,2-Propanediol — — 0.3
OH
OH
1,2-Butanediol 150 0.29 1.9
OH
OH
1,2-Pentanediol 52 0.85 16
OH
OH
1,2-Hexanediol 97 2 21
OH
OH
1,3,5-Pentanetriol — — 7.8
HO OH
(Continued )
1356
Substrate Structure KM kcat kcat/KM

(mM) (s1) (s1 M1)
OH
1,2,4-Butanetriol OH 170 4.4 26
HO
OH
3-Butene-1,2-diol 250 0.34 1.4
OH
OH
4-Pentene-1,2-diol 42 0.35 18.3
OH
3-Butenol HO 480 0.1 0.2
1,4-Butanediol OH — — 0.5
HO
NH2
2-Amino-1-pentanol 35 0.017 0.6
OH
(R)-1-Phenyl-1, 101 0.74 7.3

OH
2-ethanediol
OH
(S)-1-Phenyl-1, OH 86 0.008 0.1

2-ethanediol
OH
group, which is protoporphyrin IX (heme peroxidases), selenium (glutathione

peroxidase) [258], vanadium (haloperoxidase) [259–264], or manganese (manganese
peroxidase) [265, 266].
Heme peroxidases are interesting for preparative application as they catalyze P450-
monooxygenase-like reactions – especially if the proximal iron ligand stemming
from the protein backbone is a cysteine (like in most P450 monooxygenases) [267,
268]. In contrast to monooxygenases, peroxidases do not need reduced nicotinamide
cofactors for catalysis. Instead, simple H2O2 or organic peroxides serve as oxidants to
regenerate the active oxyferryl species (Figure 32.6). Chloroperoxidase from Caldar-
iomyces fumago (CPO) in this respect is the most prominent peroxidase used for
catalysis (vide infra).
Figure 32.6 Simplified oxidation mechanism of CPO. Instead of hydrogen peroxides, organic
hydroperoxides can also be used. Inset: structure of protoporphyrin IX.
At a first glance, utilization of cheap hydrogen peroxide as electron acceptor seems

appealing. However, from their H2O2 dependency also stems one of the major
limitations for peroxidases en route to preparative usefulness: oxidative degradation
of the heme group is frequently observed [269]. For example, CPO exhibits a half-life
of less than 1 h even at a H2O2 concentration of only 30 mM [270–272]. Similarly, the
peroxidases from horseradish (HRP) and soybean (SBP) are also subject to peroxide-
related irreversible inactivation [273, 274]. Various reaction engineering approaches
such as immobilization [275–278], encapsulation in polymersomes [279], use of
antioxidants and/or radical scavengers [272, 280, 281], ionic liquids [282–284], or the
use of organic hydroperoxides such as tert-butyl hydroperoxide (tBHP) have been
reported [285, 286]. In addition, protein engineering to some extent yielded stabilized
CPO variants [287, 288].
However, the most efficient way to improve CPO stability and turnover is to
maintain the in situ H2O2 concentration low. For example, by sensor-controlled H2O2
addition the TTN(CPO) could be increased more than 20-fold to about 860.000 [270].
External H2O2 addition still has the disadvantage that locally high concentrations
occur at the entry points, resulting in CPO inactivation at these hot spots. This can be
circumvented via in situ generation of hydrogen peroxide. As a result of intensive
research in this area during recent years, a broad variety of such in situ H2O2
generation approaches have been reported. Among them are electrochemical [289,
290], photochemical [291], transition metal-based [292], and enzymatic variants
(Table 32.10). Currently, the latter are preferred and mostly alcohol oxidases are
used for this purpose [224, 293, 294].
CPO is mostly recognized as catalyst for selective epoxidation and heteroatom
oxyfunctionalization reactions. There are, however, a couple of reports demonstrat-
ing the usefulness of CPO for selective oxidation of alcohols. Table 32.11 gives a
representative selection. In general terms, CPO exhibits high activity towards
primary, activated (benzylic, allylic, propargylic) alcohols whereas non-activated
alcohols are rather poor substrates. Oxidations of secondary alcohols have not been
1358
Table 32.10 Selection of reported in situ H2O2 generation methods coupled to CPO.
O
red O2 +H 2O
reductant
R R'
catalyst CPO
OH
ox
reductant H2O2
R R'
Reductant Catalyst TTN(CPO) max Remark/reference
Cathode — 58.900 Specialized equipment [289, 290, 295]

Glucose GOx 250.000 So far, the best system;
significant by-products [224]
H2 Pd 6.500 Specialized equipment necessary[292]
EDTA or formate Riboflavin 22.000 Not optimized yet [291]
reported even though these might be interesting substrates for CPO-catalyzed kinetic
resolutions (e.g., of racemic 1-aryl-ethanols) [296].
32.2.5
Laccases
Laccases (EC 1.10.3.2) belong to the so-called blue-copper oxidases predominantly

found in fungi but also in plants and insects [191, 304]. Their physiological roles vary
depending on the host organism, from sclerotization in insects to the formation of
UV-resistant spores in some bacteria. Rot-white fungi excrete laccases to facilitate the
delignification processes. Laccases catalyze hydrogen abstraction reactions from
phenolic and related substrates, resulting in corresponding phenoxy radicals. Lac-
cases contain four copper ions classified in one T1 copper ion and a T2/T3 cluster. It
has been shown that the T1 site is the primary redox center accepting electrons from
the electron donors. Thus, the fully oxidized laccase is transformed via four
successive, fast single-electron transfer (SET) steps into the fully reduced laccase.
Molecular oxygen interacts with the fully reduced (T2/T3) cluster via a fast two-
electron-transfer process. The resulting peroxide is tightly bound so that H2O2
release prior the second two-electron-transfer is efficiently prevented. As a result, the
fully oxidized form of laccases consisting of a m3-oxo-bridged trinuclear (T2/T3) site is
formed. This structure is thermodynamically relatively stable and provides the
driving force for the overall process as well as an efficient electron transfer bridge
to re-generate the fully reduced state (Figure 32.7).
Along with their natural phenolic substrates, laccases also accept a range of non-
phenolic substrates such as ABTS, syringaldazine, b-diketones, and TEMPO (2,2,6,6-
tetramethyl-1-piperidinyloxyl) and derivates [191, 306]. Unfortunately, no direct
Table 32.11 Examples for CPO-catalyzed oxidation of activated primary alcohols.
Substrate Product Yield (%) Remarks/reference
OH O 81 [297]
OH O 95 E/Z Isomerization could be

controlled by the water
content [297]
OH O 99 [297]
OH O 97 [297]
O O
OH 70 Poor enantioselectivity at
O
room temperature, at 5 C up
to 99% e.e. [297–299]
O O
50 Poor enantioselectivity at
OH O
to 99% e.e. [297–299]
O O
OH O 46 Poor enantioselectivity at
to 99% e.e. [297, 298]
OH O 5–56 [300]
R R
OH O 4 [301]
OH O 2 [301]
OH O 29 [301]
OH O 26 [299, 301]
OH O 6 [301]
OH O 36 [301]
OH O 43 [301]
OH O 100 [302]
(Continued )
1360
Substrate Product Yield (%) Remarks/reference
OH O
32 [301]
OH O 27 [301]
N N
O O
34 [301]
OH O
40 [291]
O
Product mixture [303]
HO O
oxidations of alcohols have been reported so far. Such systems, however, would be
highly interesting from a preparative point-of-view as simple molecular oxygen would
be available as stoichiometric oxidant (such as in case of oxidases) while circumvent-
ing the formation of hazardous hydrogen peroxide.
Figure 32.7 Simplified oxidation mechanism of laccases [305, 306].

However, some laccase substrates can oxidize activated alcohols. The resulting so-
called laccase-mediator systems (LMSs) have attracted some preparative interest in
recent years (Scheme 32.16). Here, instead of direct enzymatic oxidation of the
alcohol, the laccase regenerates an oxidized mediator, which then oxidizes the
alcohol. Many of these mediators are established oxidants widely used in organic
synthesis [307]; thus the LMS can be considered as a green alternative to established
routes that often use problematic hypochlorite. However, the efficiency of the
currently known LMS systems is not high enough to enable preparative implemen-
tation – in particular the still low total turnover numbers of the mediators (seldom
exceeding a few dozen) represents a major limitation [96, 308].
Scheme 32.16 Schematic representation of the laccase-mediator-system (LMS) for the

chemoenzymatic oxidation of alcohols.
A broad range of mediators have been reported such as 2,20 -azino-bis(3-ethyl-

benzthiazoline-6-sulfonic acid) (ABTS), hydroxybenzotriazole (HBT), violuric acid
(VA), 3-hydroxyanthranilic acid (HAA), N-hydroxyphthalimide (HPI), and 2,2,6,6-
tetramethyl-1-piperidinyloxyl (TEMPO) [191]. The ABTS-based oxidation mecha-
nism seems to involve a sequence of single-electron transfer (SET) reactions with
intermediate deprotonation of the radical cation (Scheme 32.17).
Scheme 32.17 LMS-catalyzed oxidation of alcohols using ABTS as redox mediator. Upper part:
laccase-catalyzed regeneration of ABTS þ [306].
1362
Most N-OH mediators react via initial hydrogen atom abstraction (HAT) followed
by a SET step (Scheme 32.18) [309–313].
Scheme 32.18 LMS-catalyzed oxidation of alcohols using N-OH mediators.
TEMPO and its derivates, however, follow an ionic (oxamonium) oxidation

pathway (Scheme 32.19) [307, 311, 313–315]. An interesting further improvement
of this concept was reported by Vidziunaite and coworkers [316]. By co-catalysis of 10-
(3-propylsulfonate)phenoxazine (PSPX) the laccase-catalyzed regeneration rate of
oxidized TEMPO could be accelerated by more than two orders of magnitude, which
might lead to further developments making LMS feasible for the oxidation of
alcohols.
Scheme 32.19 LMS with TEMPO as catalytic oxidant [317].

Table 32.12 gives some representative examples of applications of the LMS for the
selective oxidation of the most activated alcohol functionality in the substrate
molecule.
32.2.6
Aldehydes/Acids from Primary Alcohols
This section discusses some basic principles and practical applications of enzymes to
the oxidation of primary alcohols. Oxidation of primary alcohols can either proceed
through to the acid stage or stop at the intermediate aldehyde level. Strategies to
selectively achieve either outcome will be introduced. In addition, stereochemical
aspects will be discussed.
32.2.6.1 Stopping the Oxidation at the Aldehyde Stage

The major challenge encountered if selective oxidation to the aldehyde stage is desired
is the high reactivity of the product. Aldehydes are highly reactive electrophiles,
reacting with various nucleophiles, which also stem from enzymes. The resulting
Schiff bases can impair enzyme activity in various ways. One is to sterically block access
to the active site, another is that surface polarity can be changed leaving insoluble or
structurally impaired biocatalysts. Furthermore, over-oxidation of aldehydes to the
acid stage is thermodynamically favored. This is a challenge particularly when
applying whole-cell biocatalysis where oxidation of aldehydes adds to the organisms
energy balance by regenerating one equivalent of NADH, but also when using isolated
enzymes such as ADHs or AlcOxs enzymatic over-oxidation may be observed.
One elegant solution to this problem is to make use of in situ product removal from
the reaction phase. Usually, a second liquid and less polar organic phase is applied.
Thus, the aldehyde preferentially partitions into the less polar phase and thereby is
extracted from the reaction mixture (Figure 32.8). Additionally, this strategy is
beneficial for downstream processing, which is greatly facilitated and overall higher
substrate payloads are possible than under one-phase conditions. In addition, toxic
and inhibitory effects of substrates and products can be alleviated by the second
phase, keeping the in situ aldehyde concentration in the biocatalyst phase low.
This concept was elegantly used by B€ uhler et al. for the selective oxidation of
pseudocumene to 3,4-dimethyl benzaldehyde using recombinant E. coli [324–327].
By using dioctyl phthalate as second organic phase, both the toxic effects of the
reagents and undesired overoxidation to the acid could be efficiently circumvented.
Overall, product titers of up to 0.22 M at 70% isolated yield were achieved.
Another application of the two liquid phase concept to control oxidation selectivity
simply by choice of reaction conditions was demonstrated by Molinari and coworkers
and others [328–332]. Using acetic acid bacteria as biocatalysts a range of primary
alcohols was converted selectively into either the aldehyde or the acid if and aqueous/
isooctane or a pure aqueous reaction mixture was used (Table 32.13).
A selective route to glycolaldehyde was reported by Isobe and coworkers [212, 213].
Using either AlcOx from Pichia pastoris or glycerol oxidase from Aspergillus japonicus
ethylene glycol was converted highly selectively (up to 97% yield and selectivity) into
1364
Table 32.12 Selection of LMS-catalyzed oxidation of alcohols.
Substrate Product Remarks/referencea)
OH OH
HO OMe O OMe
Violuric acid/LTv [318]

OMe OMe
OMe OMe
OMe OMe
OH O
O O
HO HO
HO O HO O
OH OH
MeO MeO TEMPO/LTb
NHAc NHAc Up to 86% conversion at 20 mol.% of
MeO MeO
TEMPO, the immobilized enzyme could be
recycled several times [319]
O O
MeS MeS
OH O
ABTS/LPc (R ¼ alkyl, alkoxy, halo,
R R nitro) [320, 321]
TEMPO/LTv [317]
R R
OH O TEMPO/LTv [321]
MeO MeO
OH O TEMPO/LTv [321]
yield: 94%
OH O
TEMPO/LTv [321]
yield: 44%
O
O TEMPO, N-hydroxyphthalimide, hydroxy-
O
R (CH2)n benzotriazole/LTv [322]
R (CH2)n
OH HOOC
O O
HO HO TEMPO/LTb
OR OR
HO OH various sugars (also polymeric) were selec-
HO OH tively oxidized at the primary alcohol, oxi-
dation went through to the acid, isolated
yields up to 50% [323]
a) LTv: laccase from Trametes villosa; LTb: laccase from Trametes pubescens; LPc: laccase from Pycnoporus coccineus; TEMPO: 2,2,6,6-tetramethyl-1-piperidinyloxyl; ABTS: 2,20 -
azino-bis(3-ethylbenzthiazoline-6-sulfonic acid.
j1365
1366
Figure 32.8 Prevention of over-oxidation by in situ removal of the intermediate aldehyde to an

organic phase. The organic phase also serves as substrate reservoir.
glycolaldehyde (Scheme 32.20). Overoxidation to glyoxal could be kept at less

than 1%.
Scheme 32.20 Selective oxidation of ethylene glycol to glycolaldehyde.
In addition, the selective conversion of glycolic acid into glyoxylic acid was
investigated using various oxidases [247–249, 333]. To prevent undesired overoxida-
tion, in situ removal of glyoxylic acid as imine can be used to optimize product yields.
The latter reaction was also applied to establish a chemoenzymatic route to N-
(phosphonomethyl)glycine (Scheme 32.21) [248].
Scheme 32.21 Chemoenzymatic synthesis of N-(phosphonomethyl)glycine [248].
Another nice example was reported by Siebum and coworkers [334]: AlcOx from
Pichia pastoris was used to in situ generate the reactive aldehyde substrate for the
Table 32.13 Control over the selectivity of a biocatalytic oxidation of primary alcohol by choice of the reaction conditions [328].
H2O Acetobactersp.
R OH
R O or
Gluconobacterasaii O
Acetobactersp.
R OH or
Gluconobacterasaii
R O
H2O/isooctane
Substratea) Yield of acid in pure H2O (%) Yield of aldehyde in H2O–isooctane (%)
HO >97 93
>97 90
HO
HO >97 91
16 29
HO
<5 <5
OH
>97 85
OH (Continued )
j1367
1368
H2O Acetobactersp.
R OH
R O or
Gluconobacterasaii O
Acetobactersp.
R OH or
Gluconobacterasaii
R O
H2O/isooctane
Substratea) Yield of acid in pure H2O (%) Yield of aldehyde in H2O–isooctane (%)
OH >97 96
S
20 24
OH
a) [Substrate]0 ¼ 2.5 g l1.

2-deoxyribo-5-phosphate aldolase (DERA)-catalyzed synthesis of b-hydroxyketones
(Scheme 32.22).
Scheme 32.22 Two-step, one-pot reaction, combining an oxidase and an aldolase.
Similarly, Wong and coworkers integrated the GalOx-catalyzed oxidation of

glycerol to L-glyceraldehyde into a four-enzyme one-pot cascade reaction to produce
fructose from simple precursors (Scheme 32.23) [335].
Scheme 32.23 One-pot synthesis of fructose by combining GalOx, rhamnulose-1-phosphate

aldolase (RhaD), and acid phosphatase (AP) [335].
32.2.6.2 Through Oxidations

Various examples of biocatalytic through oxidations of primary alcohols to the
corresponding acids have been reported over the years. In this field, whole-cell
biotransformations (mostly using acetic acid bacteria) clearly dominate over the use
of isolated enzymes. This can be attributed to mainly two factors. First, most
biocatalysts (such as ADHs, AlcOxs, etc.) are highly specific for one redox biotrans-
formation such as the oxidation of alcohols to aldehydes. Further oxidation may be
considered as a sort of catalytic promiscuity, generally occurring at significantly
reduced rates. As a result, two biocatalysts are generally required to achieve the full
oxidation of primary alcohols to the corresponding acid. These are already present in
whole cells whereas a two-enzyme system allowing through oxidation is more
complex. Second, as for many redox biotransformations, the issue of cofactor
regeneration is less severe compared to the use of isolated enzymes. Table 32.14
gives a representative selection of whole-cell transformations.
1370
Table 32.14 Representative selection of whole cell transformations for the oxidation of primary alcohols to carboxylic acids.
Catalyst Substrate Product Yield Remarks/reference
OH CO2H
Acetobacter Up to 23 g l1 Immobilized cells were advan-
tageous because of their
increased stability [336]
O O
Kinetic resolution, E ¼ 17
OH CO2H (S) [337–339]
OH OH
95% [340]
OH CO2H
OH
CO2H 1 g l1 Kinetic resolution, E > 200 (S),
strongly dependent on substrate
concentration; immobilized,
reusable cells [341]
HO Up to 80 mM Desymmetrization with >95%
HO OH CO2H
e.e. for the product. A pro-
nounced product inhibition was
overcome by extraction of the
acid into isooctane–trioctylpho-
sphine oxide (80 : 20) [342]
Up to 25 g l1 >97% e.e., fed batch [343]
OH
CO 2 H
60 g l1 [344]
CO 2H
OH
CO 2H
OH
45 g l1 [344]
CO 2H
OH
Gluconobacter OH CO2H Up to 76 g l1 (Na salt) [345]
oxydans
O O
OH O
O CO2H (R) [337]
O O
OH CO2H (S) [346]
10 g l1

OH Kinetic resolution with excellent
HO HO CO2H enantioselectivity below a sub-
strate concentration
<20 g l1 [347]
O O
Comamonas 5 g l1 Kinetic resolution, E ¼ 49 [337]
OH O
testosteroni O CO2H
(Continued )
j1371
1372
Catalyst Substrate Product Yield Remarks/reference
O OH O
Nocardia CO2H Up to 9 g l1 [348]
corallina
HO2C R R
OH
Rhodococcus CO2H 60% Up to 85% e.e. of the
sp. products [349]
R' R'
Cucurbita CO2H Low Very slow reaction [350]
OH
maxima 0-2 0-2
Many biocatalytic oxidations proceed stereoselectively via a kinetic resolution of the
substrate. Thus, only one of the substrate enantiomers is converted into the acid while
the other remains intact. Despite the fundamental disadvantage of kinetic resolutions
that only a maximal yield of 50% is possible this represents a convenient method to
obtain enantiopure alcohols. Whole-cell oxidations sometimes are hampered by low
overall enantioselectivity, mostly because the substrates are accepted by various
enzymes with complementary enantioselectivity [351]. For example, it is frequently
observed that the enantioselectivity of acetic acid bacteria-catalyzed oxidations
strongly varies with the substrate concentration [341, 347], which can be rationalized
by different affinities of the two (or more) enantiocomplementary enzymes towards
the substrate. Thus, maintenance of a low in situ substrate concentration by
continuous addition of the substrate or the use of cyclodextrins is a viable option [341].
Further, if both enzymes differ in their pH requirements, control over the reaction pH
value can help to increase the overall enantioselectivity [347].
Using isolated biocatalysts of course circumvents the aforementioned limitations,
thereby in principle allowing high substrate concentrations, albeit at the expense of
the necessity for a cofactor regeneration system. As early as 1985, Wong and
coworkers reported on a cell-free biocatalytic system for the kinetic resolution of
1,2-diols and -amino alcohols to produce optically pure a-hydroxy acids and a-amino
acids (Table 32.15) [13, 23]. Unfortunately, this very promising approach was not
followed further. More recently, Ohta and coworkers reported on an in vitro oxidation
system combining purified alcohol- [352] and aldehyde- [99] dehydrogenases from
Brevibacterium sp. [97] for the through-oxidation of various primary alcohols to the
corresponding acids. For in situ regeneration of the oxidized nicotinamide cofactor
the H2O-forming NADH oxidase from Lactobacillus brevis was used [98]. In addition,
hexanoic acid production using YADH coupled to NADH oxidase-cofactor regener-
ation has been reported recently [353].
A very interesting variant of the through oxidation of alcohols consists of the
HLADH-catalyzed oxidation of 1,4- and 1,5-diols. The primary aldehyde product
spontaneously cyclizes forming a hemiacetal that is further oxidized by HLADH,
giving access to enantiopure lactones. Very successfully, meso-diols have been
desymmetrized (Table 32.16).
32.2.7
Regioselective Oxidation in Polyols
Next to stereoselectivity, regioselectivity of enzymes is also one of the key-advantages

of biocatalysis over conventional counterparts. High regioselectivity is especially
appreciated when it comes to oxidation of carbohydrates. The poly-functionality of the
starting material makes extensive protection/deprotection steps inevitable to achieve
regioselective oxidation. Polyol-dehydrogenases and -oxidases on the other hand
exhibit an intrinsic high to exclusive regioselectivity (Section 32.2.1), which is
exploited in a range of preparative and industrial applications.
The so-called Cetus-process consists of a chemoenzymatic route for fructose
from glucose. In the first step, glucose is converted into the corresponding
1374
Table 32.15 Enantiopure a-hydroxy- and a-amino acids through HLADH-catalyzed kinetic resolution [13, 23].a).
X X X
HLADH AldDH
OH O
R R R CO 2H
NADH NAD + NADH NAD +
HO 2C CO 2H HO 2C CO 2H
GluDH GluDH
NH 2 O
+NH 3
Substrate Rateb) E.e. (%)c) Substrate Rateb) E.e. (%)c)
OH OH
1 >97 33 >97
HO OH OH
OH OH
6 >97 OH 26 0
F OH
OH NH2
5.5 >97 1 96
Cl OH HO OH
OH NH2
5.5 >97 16 84
Br OH OH
OH NH2
29.5 >97 OH 11 0
OH
OH O
3 >97 O 7 0
H2 N OH
OH
OH OH
30 >97 50 0
OH OH
a) HLADH: ADH from horse liver, AldDH: aldehyde dehydrogenase from yeast, GluDH: glutamate DH.
b) Relative rates.
c) Of the acid product.

j1375
1376
Table 32.16 Selection of HLADH-catalyzed lactone-formation from 1,4-diols.

OH O
OH O
HLADH OH O HLADH O
OH
NAD + NADH NAD+ NADH
Product Yield/e.e. (%) Regeneration Product Yield/e.e. (%) Regeneration system/reference
system/reference
O
O
64/>97 FMN/O2, [112] 79/>99.5 FMN/O2, GluDH/ketoadi-
O pate [113, 354]
O
(2R, 3S)
O O
86/>97 FMN/O2, [112] n ¼ 0 68/>99.5; n ¼ 1 FMN/O2, [113]
O O 90/>99.5; n ¼ 2 72/
(CH2)n
(2R, 3S) >99.5; n ¼ 3 79/
>99.5
R O
74/>97 FMN/O2, [112] R ¼ Me 70/90; R ¼ FMN/O2, [355]
O
O iProp 75/25; R ¼ Ph
O (2S, 3R) 60/21
O O
73/>97 FMN/O2, [112] High/99 FMN/O2, [356]
O
O
O (2S, 3R)
1374
Table 32.15 Enantiopure a-hydroxy- and a-amino acids through HLADH-catalyzed kinetic resolution [13, 23].a).
X X X
HLADH AldDH
OH O
R R R CO 2H
NADH NAD + NADH NAD +
HO 2C CO 2H HO 2C CO 2H
GluDH GluDH
NH 2 O
+NH 3
Substrate Rateb) E.e. (%)c) Substrate Rateb) E.e. (%)c)
OH OH
1 >97 33 >97
HO OH OH
OH OH
6 >97 OH 26 0
F OH
OH NH2
5.5 >97 1 96
Cl OH HO OH
OH NH2
5.5 >97 16 84
Br OH OH
OH NH2
29.5 >97 OH 11 0
OH
OH O
3 >97 O 7 0
H2 N OH
OH
OH OH
30 >97 50 0
OH OH
a) HLADH: ADH from horse liver, AldDH: aldehyde dehydrogenase from yeast, GluDH: glutamate DH.
b) Relative rates.
c) Of the acid product.

j1375
1376
Table 32.16 Selection of HLADH-catalyzed lactone-formation from 1,4-diols.

OH O
OH O
HLADH OH O HLADH O
OH
NAD + NADH NAD+ NADH
Product Yield/e.e. (%) Regeneration Product Yield/e.e. (%) Regeneration system/reference
system/reference
O
O
64/>97 FMN/O2, [112] 79/>99.5 FMN/O2, GluDH/ketoadi-
O pate [113, 354]
O
(2R, 3S)
O O
86/>97 FMN/O2, [112] n ¼ 0 68/>99.5; n ¼ 1 FMN/O2, [113]
O O 90/>99.5; n ¼ 2 72/
(CH2)n
(2R, 3S) >99.5; n ¼ 3 79/
>99.5
R O
74/>97 FMN/O2, [112] R ¼ Me 70/90; R ¼ FMN/O2, [355]
O
O iProp 75/25; R ¼ Ph
O (2S, 3R) 60/21
O O
73/>97 FMN/O2, [112] High/99 FMN/O2, [356]
O
O
O (2S, 3R)
O
O
64/>97 FMN/O2, [112] 81/>99.5 FMN/O2, [357]
O
O
(2S, 3R)
O
O
87/>97 FMN/O2, [112] 76/>99.5 FMN/O2, [357]
O
O
(2S, 3R)
O H3CO O
93.5/>99.5 FMN/O2; ABTS/ 42/>99.5 FMN/O2, [357]
anode, [113, 138]
O O
j1377
1378
2-keto-glucose by immobilized P2O and subsequently hydrogenated into practically

pure D-fructose (Scheme 32.24) [358, 359].
Scheme 32.24 Chemoenzymatic transformation of glucose into fructose in the so-called Cetus
process.
In the Reichenstein process for the production of ascorbic acid from glucose,
Gluconobacter oxydans catalyzes the key transformation (Scheme 32.25) [162].
Scheme 32.25 Gluconobacter oxydans-catalyzed selective oxidation of sorbitol as the key-step of the
Reichenstein process [162].
Scheme 32.26 summarizes other applications of Gluconobacter oxydans, consisting

of selective oxidation of, for example, glycerol [163], ribitol [164], and N-
butylglucamine [165].
Scheme 32.26 Preparative-scale applications of Gluconobacter oxydans for the regioselective

oxidation of ribitol (a), glycerol (b), and N-butylglucamine (c).
32.2.8
Kinetic Resolutions/Desymmetrizations
Oxidation of chiral secondary alcohols goes hand in hand with the elimination of a
chiral center. Using an enantioselective oxidation catalyst (enzyme) allows accumu-
lation of only one enantiomer in a kinetic resolution. The intrinsic disadvantage of
kinetic resolutions, however, is the maximal yield of only 50% (provided the catalyst
exhibits perfect enantioselectivity). Thus, from the point-of-view of enantioselective
synthesis, enantiospecific reduction of the prochiral ketones is usually preferred
(100% yield) [5]. Nevertheless, a range of oxidative kinetic resolutions has been
reported, which are summarized in Table 32.17.
The intrinsic disadvantage of kinetic resolutions can be easily overcome by using
symmetrical, prochiral alcohols, which upon oxidation are transformed into a chiral
product. Using this meso-trick on the one hand enables full conversion of the
starting material. On the other hand, the number of meso-compounds is naturally
limited. Nevertheless, quite efficient desymmetrization of syn-cyclohexanediol using
GDH [67] and of 2,3-butanediol using whole cells of Gluconobacter asaii [340] has
been reported.
Desymmetrization of meso-2,5-hexanediol was reported using ADH-A from Rho-
dococcus ruber, yielding valuable (R)-5-hydroxy-2-hexanone at 88% conversion within
2 h (>99% e.e.) (Scheme 32.27). Here, E. coli recombinantly expressing ADH-A was
used as biocatalyst [377].
Scheme 32.27 Desymmetrization of meso-2,5-hexanediol using ADH-A from Rhodococcus ruber.
32.2.9
Racemizations
Racemizations are generally considered as an unwanted side reaction rather than a

synthetically useful transformation. As a consequence, the controlled racemization
1380
Table 32.17 Overview of some representative biocatalytic kinetic resolutions.
Product Catalyst(s) E.e. (%) Yield (%) Scale (g l1) Remark/

reference
OH
Rhodococcus rubber 97.2 49 100 [57]
OH
Rhodococcus rubber 97.8 49.4 100 [57]
OH
OH
OH
OH
OH
ADH from Lactobacillus brevis Analytical [103]
OH
Acinetobacter sp. 96–99 6–40 Analytical [360]

R
R= o-, m-, p-Cl, Br
OH
Allium schoenoprasum >98 46 Analytical [361]
OH
Raphanus sativus 89 51 [361]
OH
OH Glycerol dehydrogenase (GDH) 95 40 Analytical [66]
O
Bacillus stearothermophilus 100 48 0.5 [362]
HO
S
Pseudomonas paucimobilis 100 40 0.5 [362]
HO
(Continued )
j1381
1382

reference
N P. paucimobilis 100 43 0.5 [362]

HO
N O P. paucimobilis 90 40 0.5 [362]

HO
N
P. paucimobilis 95 47 0.5 [362]
O
HO
N
Bacillus stearothermophilus 95 40 0.5 [362]
S
HO
N
P. paucimobilis 100 45 0.5 [362]
S
HO
CF3
OH P. paucimobilis 100 44 0.5 [362]

OH
Nocardia corallina B-276 Up to 86 Analytical [363]
Cl
OH
quinohemoprotein dehydrogenase E ¼ 13 Analytical [145, 364]
(CH2)n (n ¼ 1)–
800 (n ¼ 5)
OH
Rhodococcus rubber 33 49.5 100 [57]
OH
Rhodococcus rubber 0 36.9 100 [57]
OH
Rhodococcus rubber >99 53 30 E > 100 [365]
OH
Rhodococcus rubber 64 37 30 E > 100 [365]
OH
From Candida parapsilosis 95 48.5 25 [366]
OH rec. in E. coli
(Continued )
j1383
1384

reference
OH
CO2Me Rhodococcus erythropolis >98 48.5 0.5 [367]
O
R. erythropolis [368–370]
HO
OH Horse liver alcohol E ¼ 61 Analytical [108]

Si dehydrogenase (HLADH)
OH
GlOx >98 49–50 Analytical [245]
CO2H
1 ,2,4,7
OH
GlOx 99 50 Analytical [245]
CO2H
OH
GlOx 86 47 Analytical [245]
O CO 2 H
of organic compounds has been scarcely studied [372, 378]. They are, however, useful
in combination with a kinetic resolution (KR) reaction, thereby turning it into a
dynamic kinetic resolution (DKR). In this way, the intrinsic disadvantage of KRs of
allowing maximally 50% of the desired product can be elegantly overcome.
Scheme 32.28 shows the principle of a DKR is schematically [72, 371–376]. Secondary
alcohols are stereochemically stable (in contrast to, for example, easily enolizable
a-substituted carbonyl compounds) making mild in situ racemization rather chal-
lenging next to some transition metal-based approaches [379]; the groups around
Faber and Kroutil have developed various approaches [371–376]. The underlying
principle consists of the use of two, enantiocomplementary ADHs exhibiting the
same cofactor requirement. Thus, both enantiomers are in equilibrium; since
racemization is a thermodynamically favorable process (entropy driven) this process
occurs spontaneously. Alternatively, one (bio-)catalyst exhibiting poor or no enantio-
preference can be used [376].
Scheme 32.28 Principle of a dynamic kinetic resolution involving redox racemization of the
starting material.
In recent years, several biocatalytic racemizations of secondary alcohols have been

reported, which are summarized in Table 32.18.
32.2.10
Deracemizations
The basic requirement for redox-deracemization obviously is that at least one of the
steps, either oxidation or reduction, proceeds enantioselectively. Ideally, both steps
occur with high and complementary selectivity. Thus, only one redox cycle is
necessary to achieve complete deracemization (stereoinversion of the undesired
enantiomer). However, also if one step is unspecific, gradual enrichment of one
enantiomer occurs during each cycle. As a consequence, more redox cycles have to be
1386
Table 32.18 Examples of biocatalytic racemizations.
OH O OH
Cat-1 Cat-2
R R' R R' R R'
Starting material Catalyst E.e. (%) Reference/

(after x h) remarks
OH
Rhodococcus erythropolis rac (1) [371]
Both enantiomers
OH
R. erythropolis rac (1) [371]
Both enantiomers
OH
Raphanus sativus 43 (1) [371]
O
a)
rac (24) [371]
OH
O
b)
rac (24–48) [371]
OH
O
Bacillus megaterium, rac (72) [371]
OH Helminthosphorium sp
OH
Lactobacillus paracasei rac (24) [372, 373]
CO2 H
OH
L. paracasei rac (24) [372, 373]
CO2 H
OH
CO2 H
OH O OH
Cat-1 Cat-2
R R' R R' R R'

(after x h) remarks
OH
CO2 H L. paracasei rac (24) [372, 373]
OH
CO2 H
Both enantiomers
OH
OH
Cl
O
L. paracasei 6 (48) [374, 375]
OH
O
L. paracasei 30 (72) [374, 375]
OH
O
L. paracasei 4 (72) [374, 375]
OH
O
L. paracasei 10 (72) [374, 375]
OH
O
L. paracasei 11 (72) [374, 375]
OH
(Continued )
1388
OH O OH
Cat-1 Cat-2
R R' R R' R R'

(after x h) remarks
O
L. paracasei 20 (72) [374, 375]
OH
O
OH L. paracasei 6 (72) [374, 375]
OH
ADH-A & LK-ADH rac (24) [376]
OH
OH
OH
OH ADH-A & LK-ADH rac (24) [376]
OH
OH
OH
ADH-A & LK-ADH 19 [376]
a) Streptomyces caeruleus, Agrobacterium tumefaciens, Helminthosphorium sp., Rhizopus oryzae,

Syncephalastrum racemosum, Candida parapsilosis, and Bacillus megaterium.
b) Alcaligenes faecalis, Bacillus megaterium, Geotrichum candidum, Streptomyces caeruleus, and
Agrobacterium tumefaciens.
performed to achieve sufficiently high enantiomeric excess of the product. In contrast
to the above-discussed racemizations, deracemizations are thermodynamically
slightly uphill. Since both racemic starting material and enantiopure (or at least
enriched) products are secondary alcohols, the enthalpic term (DH) of the Gibbs-free-
energy change during deracemization reaction is zero. The entropic term (DS),
however, calculates as DS ¼ R ln 2, being about 0.4 kcal mol1 at 20 C [409, 410]. As a
consequence a thermodynamic driving force has to be added to the system to drive
any deracemization reaction to completion.
In the following some established strategies are briefly discussed: The use of two,
enantiocomplementary ADHs has gained considerable interest in recent times
(Scheme 32.29) [72, 81, 381–383, 411]. ADH-1 catalyzes the kinetic resolution of
the racemate to produce 0.5 equiv. of the desired enantiomer and 0.5 equiv. of ketone,
which is then transformed by enantiospecific reduction catalyzed by ADH-2. One
drawback of this methodology is that both ADHs need to differ in their cofactor
requirements to prevent undesired cross-activities. Likewise the corresponding
cofactor regeneration systems need to be selective for either the phosphorylated or
the non-phosphorylated nicotinamide cofactor. Cross-activities lead to futile cycles
and eventually heat generation by unnecessary consumption of the respective
cosubstrates.
Scheme 32.29 Bi-enzymatic deracemization utilizing enantiocomplementary ADHs with at the

same time differing cofactor requirements.
A somewhat simplified methodology is shown in Scheme 32.30. Here, a cofactor-

independent oxidase is used to catalyze the kinetic resolution step. Enantiospecific
reduction is achieved again by use of an ADH [244]. Compared to the system outlined
in Scheme 32.29, the cofactor-independent and irreversible oxidation step represents
a significant simplification.
An interesting exception to the above-mentioned requirement for enantiocom-
plementary catalysts was reported for Cunninghamella echinulata: here (S)-selective
oxidation was followed by (S)-selective reduction, albeit at lower enantioselectivity;
thus, though both steps exhibit the same enantiopreference, a deracemization is
achieved due to a mistake in the reduction step [385].
In addition to the aforementioned fully enzymatic deracemization systems, a range
of chemo- (Scheme 32.31) [244, 399, 408] and electroenzymatic (Scheme 32.32) [117]
1390
Scheme 32.30 Deracemization by combining oxidase-catalyzed kinetic resolution and ADH-

catalyzed stereospecific reduction.
Scheme 32.31 Chemoenzymatic deracemization of lactic acid by combining enantioselective

oxidation of only one enantiomer and NaBH4-mediated recycling of the resulting pyruvate into
racemic lactate.
Scheme 32.32 Electroenzymatic deracemization of lactic acid.

Figure 32.9 Deracemization of sec-alcohols or amines via a cyclic oxidation–reduction sequence;

adopted from Gruber et al. [409]. Dark and light gray: the relative concentration of both enantiomers,
line: e.e.-value of the reaction mixture.
approaches have also been reported. In both cases enantioselective enzymatic

oxidation is accompanied by an unselective reduction regenerating a fraction of
the previous racemate. Thus, during every cycle the enantiomeric excess of the
alcohol is increased. This was nicely visualized by Kroutil and coworkers (Fig-
ure 32.9) [409] by breaking down the continuously parallel working redox steps into
consecutive cycles. Under ideal conditions (perfectly enantioselective catalysts and
full theoretical conversion after each cycle) already after ten cycles an enantiomeric
purity of greater than 99.8% is attained. Obviously, under non-ideal conditions
(lower enantioselectivity) more cycles (¼ longer reaction times) are necessary to
achieve satisfactory optical purities. Table 32.19 shows a selection of biocatalytic
deracemization reactions.
32.2.11
Stereoinversions
Mechanistically, stereoinversions are closely related to the above-discussed derace-

mizations. From a chemical point-of-view they represent the biocatalytic counterpart
of the well-known Mitsunobu reaction [412]. However, compared to deracemizations
little attention has been paid to stereoinversions.
Geotrichium candidum [413], Cyanidioschyzon merolae [414], and Candida albi-
cans [415] were evaluated for the stereoinversion of some substituted arylethanols.
1392
Recently, Riva and coworkers reported on a oxidation/reduction sequence involving

stereoinversion at C7 of cholic acid (Scheme 32.33) [81].
Scheme 32.33 Redox sequence for the selective transformation of cholic acid into 12-
ketoursodeoxycholic acid [81].
Very selective stereoinversion of disaccharides was reported for the combination of

P2O and aldose reductase (Table 32.20) [196].
32.3
Oxidation of Aldehydes
32.3.1
Overview and Most Important Enzyme Classes/Applications
The oxidation of the aldehyde group yields a broad range of carboxylic acids, which
are important for the synthesis of polymers, as solvents, as synthons for drug
development, and as food additives. There are several chemical routes well known
for the oxidation of aldehydes, but examples of the biological conversion are scarce.
To date, few reports on synthetic enzymatic oxidations of aldehydes have been
Table 32.19 A selection of biocatalytic deracemization reactions.
Product Catalyst E.e. (%) Yield (%) Conc. Time Reference
Aliphatic alcohols
OH
Sphingomonas paucimobilis 99 90 1 g l1 4–6 d [380]
R R'
R: aryl, benzyl R0 : alkyl, vinyl, cyclohexyl
OH
(CH2)n S. paucimobilis 99 Up to 99 60–80 mM <22 h [381, 382]

n=5,7,9
Both enantiomers
OH
Rhodococcus spp. 93 Up to 94 50 mM 24 h [383]
OH
Rhodococcus spp. 99 99 60–80 mM 16 h [381, 382]
Both enantiomers
OH
Rhodococcus spp. [381, 382]

32.3 Oxidation of Aldehydes
HO (Continued )
j1393
1394
Both enantiomers
OH
Rhodococcus spp. 89 99 60–80 mM 16 h [381, 382]
OH
Both enantiomers
OH
Rhodococcus spp. 99 99 60–80 mM 4h [381, 382]
Both enantiomers
OH
Bacillus stearothermophilus/ 90 85 ? ? [384]
Yarrowia lipolytica
OH
Rhodococcus spp. 20 99 60–80 mM 16 [381, 382]
O
OH Rhodococcus spp. 30 99 60–80 mM 16 [381, 382]
Both enantiomers
OH
Candida rugosa 97 92 0.1 g l1 10
Benzylic alcohols
OH
Rhodococcus spp. Up to 60% 99 60–80 mM 16 [381, 382]
Both enantiomers
OH
H
N Cunninghamella echinulata 92 57 0.1 g l1 6d [385]
O
OH
F Aspergillus terreus CCT 99 86 0.4 g l1 1–17 d [386]
Both enantiomers
OH
A. terreus CCT 99 86 0.4 g l1 1–17 d [386]

X
Both enantiomers X: Cl, Br, NO2

OH
Arracacia xanthorrhiza (arraca- Up to 98 Up to 99 1.25 g l1 3–6 d [361]

cha), Polymnia sonchifolia
R
(yacon), Zingiber officinale
(ginger)
R: H, Br, CH3, NO2, SeCH3
OH
Candida rugosa 99 Up to 99 1.25 g l1 9–17 d [387]

N
O
Rhizopus oryzae 97 (R), 85 (S) Up to 76 1 g l1 15–21 d [388]
OH
(Continued )
j1395
1396
(R) pH 4–5 (S) pH 7.5-8

OH
OH Candida parapsilosis 98 92 15 g l1 48 h [389, 390]
OH
(1) LkADH/NADPH-oxidase/ Analytical [391]
catalase, (2) ReADH/FDH
OH
(1) Re ADH/NADH-oxidase/ [391]
catalase, (2) Lk ADH/NADPH

regeneration system
Propargylic alcohols
OH
Norcardia spp. 100 (R), 98 (S) Up to 83 5 g l1 1–5 d [392]
Vic-diols
OH
O OH Candida boidinii, Pichia Up to 100 Up to 88 0.2 g l1 2–7 d [393]
methanolica ¼
OH
OH Candida parapsilosis 98 92 8 g l1 48 h [390]
OH
C. parapsilosis Up to 100 99 Up to 10 g l1 1–3 d [394]
HO R
a-Hydroxy acids/esters
OH
CO2Et Candida parapsilosis 99 90 3 g l1 1h [395]
OH
Rhodococcus erythropolis 94 70 18 g l1 4d [396]
O
O
OH
O C. parapsilosis ? ? Analytical ? [397]
O
OH
CO2 Et C. parapsilosis 98 Up to 99 3 g l1 1.5 h [398]
R
R: p-Cl, p-CH3, o-Cl, m-NO2
OH
OH Lactate-oxidase/NaBH4 ? ? 5 mM 1.5 h [399]
O
OH
OH LDH/cathode 97.5 97 20 mM 32 h [117]
O
(Continued )
j1397
1398
OH
OH
R (1) GlyOx/catalase (2) LDH/FDH ? ? ? ? [244, 245]
O
R=Et, Pr, Bu
OH
OH (1) Pseudomonas polycolor IFO 99 60 300 mM 24 h [400, 401]
O 3918 (2) Micrococcus freudenrei-
chii FERM-P 13221
b-Hydroxy acids/esters
OH
CO2 Et
Candida parapsilosis 99 Up to 68 0.25 g l1 6h [402]
R
R: o-CH3, p-CH3, p-Cl, p-NO2
OH O
C. parapsilosis 96 99 60–80 mM 6h Both
O enantio-
mers
OH [381, 382]
CO2 Et
C. parapsilosis 99 62–75 0.25 g l1 6h [403]
R
R: p-OCH3, p-NO2, p-CH3
OH
Sphingomonas paucimobilis 96 84 1 g l1 6f [380]
OH
S. paucimobilis 99 90 1 g l1 5d [380]
OH
OH
Cl S. paucimobilis 99 79 1 g l1 5d [380]
OH
OH
OH

F (Continued )
j1399
1400
OH
O2N
OH
H 3CO
OH

N
OH
O S. paucimobilis 99 81 1 g l1 5d [380]
OH
S S. paucimobilis 99 78 1 g l1 5d [380]
OH
S S. paucimobilis 97 85 1 g l1 5d [380]
N
OH
OH
OH
OH Candida parapsilosis 98 92 Analytical 2d [390]
OH O
O C. parapsilosis 99 57 Analytical 6h [402–404]
OH O
OH O
OH O

Cl (Continued )
j1401
1402
OH O
O2 N
OH O
OH
C. parapsilosis 90–99 69–75 Analytical 3h [405]
R
OH
C. parapsilosis 76 71 Analytical 3h [405]
OH
C. parapsilosis 10 78 Analytical 3h [405]
OH
CO2 R
O
R
OH
CO2 R
S
R
OH
CO 2R C. parapsilosis 87–99 55–80 Analytical 3h [407]
OH
Geotrichum candidum/NaBH4 99 96 Analytical 24 h [408]
OH
G. candidum/NaBH4 95 87 Analytical 24 h [408]
OH
OH
(Continued )
j1403
1404
OH
OH
O G. candidum/NaBH4 99 66 Analytical 24 h [408]
OH
S G. candidum/NaBH4 99 73 Analytical 24 h [408]
OH
OH
OH
GlOx/LDH 99 100 Analytical 66 h [408]
CO 2H
Table 32.20 Stereoinversions catalyzed by P2O [196].
O OH O OH
O OH RO
RO RO
P2O AldR
HO OH HO O HO OH
OH OH
OH
O2 H 2O 2 NAD + NADH
H 2O + 1/ 2 O2 catalase CO2 FDH HCO 2H
Substrate Product Conversion (%)
OH OH
HO O O OH HO O O OH
O O
100
HO HO
OH OH OH OH
HO HO HO HO
Allolactose Allolactulose
OH OH
HO O O HO O O OH
O OH O
100
HO HO
OH OH OH
HO HO OH HO
HO
Meliobiose Meliobiulose
(Continued )
j1405
1406
O OH O OH
O OH RO
RO RO
P2O AldR
HO OH HO O HO OH
OH OH
OH
O2 H 2O 2 NAD + NADH
H 2O + 1/ 2 O2 catalase CO2 FDH HCO 2H
Substrate Product Conversion (%)
OH OH
HO O O HO O O
O OH O OH
100
HO HO
OH OH OH
HO HO OH HO
HO
Gentiobiose Gentiobiulose
OH OH
HO O O HO O O
O OH O OH
100
HO HO
OH OH OH
HO HO OH HO
HO
Isomaltose Palatinose
32.3 Oxidation of Aldehydes j1407
published. Preparative applications reported include bioconversions of natural
products such as retinal (Scheme 32.34a) [416] and various aliphatic and unsaturated
aldehydes (Scheme 32.34b) and aromatic aldehydes (Scheme 32.34c). Another
reported reaction type is the production of olefins from aldehydes by the oxidative
removal of formic acid from the substrate (Scheme 32.34d) [417–419]. The most
important enzyme classes, which will be discussed in this context, are alcohol
dehydrogenases, aldehyde dehydrogenases, monooxygenases, and oxidases. This
section will close with some examples from whole-cell catalysis.
Scheme 32.34 Selected enzymatic oxidations (c) oxidation of (hetero)aryl aldehydes;

of aldehydes: (a) oxidation of complex natural (d) oxidative cleavage of the aldehyde C–Ca
compounds, such as retinal; (b) oxidation of bond to yield terminal alkenes.
aliphatic and a,b-unsaturated aldehydes;
32.3.2
Alcohol Dehydrogenases
Alcohol dehydrogenases (ADHs) are a very diverse group of enzymes, which are
classified according to their prosthetic groups and structure into three groups. Group I
contains zinc-dependent long-chain ADHs (about 350 residues), group II are the so-
called short-chain zinc-independent ADHs (about 250 residues), and group III are
iron-activated ADHs of a subunit size of about 385 residues [6]. They are mainly
known to catalyze the reversible oxidation of primary and secondary alcohols to the
corresponding aldehydes and ketones, respectively, transferring the hydride equiva-
lent liberated to the oxidized nicotinamide cofactor NAD(P) þ [420]. These enzymes
have also attracted interest due to their ability to oxidize aldehydes, although this
reactionis considered to be rathera sidereactionand not theirmain function [421,422].
1408
The first ADH described to catalyze the oxidation of aldehydes was an alcohol
dehydrogenase isolated from horse liver (HLADH), which was shown to oxidize
formaldehyde [423, 424] and other aldehyde substrates like octanal, butanal, and so
on [425]. The oxidation activity of HLADH depends on the concentration of the
cofactor NAD þ . At stoichiometric concentrations of NAD þ and enzyme, HLADH
catalyzes the dismutation of aldehydes to equimolar quantities of the corresponding
acid and alcohol without net synthesis of NADH [424], while at higher cofactor
concentrations the aldehyde is oxidized with the concurrent reduction of NAD þ to
NADH [426]. Henehan and Oppenheimer proposed the reaction scheme shown in
Scheme 32.35 for this sequential reaction for TBADH.
NADH NADH
NAD+ R–CHO NAD+
K1 K2
R–CH 2OH R–COOH
K-1
NADH R–CH(OH)2
NAD+
Scheme 32.35 Reaction scheme of the sequential oxidation of alcohol to the corresponding
carboxylic acid catalyzed by TBADH as proposed by Henehan and Oppenheimer [425].
Essential for the reaction is the catalyst concentration. At low enzyme concentra-
tion and/or high aldehyde concentrations rapid dismutation of the aldehyde sub-
strate into the corresponding acid and alcohol is observed, until approx. 95%
conversion. Subsequently, an increase in the reduced cofactor NADH production
is detected, yielding a dynamic equilibrium between alcohol and aldehyde and
reduced and oxidized cofactor [425].
Therefore, by tuning the concentration of the different reactants involved the
system is capable of a total conversion of the alcohol or aldehyde substrate to the
carboxylic acid.
Various other ADHs from different species have also been described, which are
capable of the sequential oxidation of alcohol to the corresponding carboxylic acids.
These include ADHs from Drosophila melanogaster [427], TADH [54] and
TBADH [428], both isolated from extremophile bacteria, HLADH, YADH from
yeast [429], and HpCAD, an alcohol dehydrogenase isolated from Heliobacter
pylori [430].
32.3.3
Aldehyde Dehydrogenases
Aldehyde dehydrogenases belong to the superfamily of NAD(P) þ dependent

enzymes found in prokaryotes as well as in eukaryotes, catalyzing the irreversible
oxidation of various aldehydes to their corresponding carboxylic acids via the forma-
tion of a thiohemiacetal covalent intermediate with the aldehyde substrate [431]. This
enzyme class received a lot of attention in medical related research, as aldehydes
represent a group of rather toxic molecules for metabolism and a malfunction of
Figure 32.10 Reaction mechanism of the aldehyde oxidation reaction catalyzed by FDH as
described by Tsybovsky and coworkers [433].
enzymes responsible for aldehyde degradation leads to severe illnesses. On the other
hand, because of this essential role regarding detoxification these enzymes make a
good target for antibiotics in pathogenic microorganisms [432].
The reaction mechanism for aldehyde oxidation catalyzed by aldehyde dehydro-
genases has been investigated in the recent years for a couple of members of
the ALDH-family. As an example we discuss here the mechanism described for
10-formyltetrahydrofolate dehydrogenase from rat (Figure 32.10), which has been
described by Tsybovsky and coworkers and is exemplarily for many aldehyde
dehydrogenases [433].
Aldehyde oxidation is initiated by binding of the oxidized cofactor followed by the
binding of the aldehyde substrate. A key feature in the reaction mechanism seems to
be the orientation of the NAD coenzyme within the Rossmann fold, which is rather
unusual for NAD-binding proteins [434]. In this binding mode the pyrophosphate is
only weakly bound to the enzyme and it is assumed that this allows for the transition
between the hydride transfer and the hydrolysis positions of the nicotinamide ring
during the reaction [433, 435]. The two key amino acids involved in catalysis are Glu673
and Cys707. Owing to their orientation in the apoenzyme, Glu673 abstracts a proton
from Cys707, thus producing a more nucleophilic thiolate. While this proton is lost to
the solvent over a continuous network of water molecules and main chain carbonyl
groups [433], the thiolate attacks the carbonyl carbon of the aldehyde to form a
thiohemiacetal intermediate. Hydride-transfer to the oxidized nicotinamide cofactor
1410
allows for the re-formation of the carbonyl group, yielding an intermediate thioester.
The reduced cofactor now leaves the active center and is replaced by a water molecule,
which is bound to Glu673. This amino acid activates the water, allowing for a
nucleophilic attack at the carbonyl carbon, which finally leads to product release [433].
Examples of aldehyde dehydrogenases for biocatalytic applications are scarce. An
aldehyde dehydrogenase (EC 1.2.1.5) from yeast was applied to oxidize (Z,Z)-nona-
2,4-dienal, yielding the corresponding v-hydroxy carboxylic acid [436]. Recycling of
the necessary cofactor NAD þ was achieved in situ by addition of an alcohol
dehydrogenase, reducing (Z,Z)-nona-2,4-dienal to the corresponding alcohol. Since
both reactions are stoichiometrically linked via NAD, this corresponds to an overall
disproportionation of the aldehyde (Scheme 32.36) as described above.
Scheme 32.36 Enzymatic disproportionation of (Z,Z)-nona-2,4-dienal to the corresponding

alcohol and acid catalyzed by an alcohol and an aldehyde dehydrogenase, respectively [436].
This concept was extended to industrially relevant metabolites of linoleic acid,

which is an important building block for polymers and detergents. No isomerization
of the double bonds and yields of up to 90% were reported [436].
Higher vertebrates possess retinal specific aldehyde dehydrogenases, which often
occur in various isoforms. The main substrate of these enzymes is retinaldehyde,
which is converted into the corresponding carboxylic acid. Retinoic acid is one of the
most important metabolites of retinol that is involved in cell differentiation. NAD þ -
dependent aldehyde dehydrogenase isoenzyme I and II from bovine kidney was
characterized with respect to its activity towards retinal and other aldehydes
(Table 32.21a) [416].
Also very interesting, because of a broad substrate spectrum, is the benzaldehyde
dehydrogenase II from Acinetobacter calcoaceticus. The best substrate for this enzyme
was benzaldehyde, but it also converted other derivatives of benzaldehyde as well as
aliphatic aldehydes (Table 32.21b) [437, 438].
32.3.4
Monooxygenases
Monooxygenases belong to the very versatile enzyme class of oxygenases, which are
found in almost all kinds of living cells and several different catalytic centers and
Table 32.21 Substrates and kinetic constants for some AlDHs.
(a) Bovine kidney aldehyde dehydrogenase isoenzyme II [416]

Substrate Vmaxa) Km (mM)
(U mg1)
O
3.8 9.1
O
29 1
H
O
33 1.5
H
O
75 30
H
O
64 33.9
H
O
116 8.2
H
(b) Benzaldehyde dehydrogenase II from Acinetobacter calcoaceticus [437, 438]

Substrate Vmaxb) Km (mM) kcat (s1) kcat/Km
1
(U mg ) (s1 mM1)
H 78 0.57 72 126
H 70 2.8 64 23.1
F
O
F 44 3.9 41 10.3
H
H 53 16 49 3.08
H 3CO
O
H 3CO 36 6.5 33 5.07
H
(Continued )
1412
H 22 18 20 1.10
HO
O
HO 32 63 29 0.46
H
H 45 187 41 0.22
OH
O
H 29 15 26 1.82
H 72 20 66 3.30
N
H 75 26 68 2.62
S O
57 5.1 52 10.2
H
O O
50 59 46 0.79
H
H 19 152 17 0.11
O
29 460 26 0.06
H
O
31 72 29 0.39
H
a) U ¼ nmol min1 mg1.

b) U ¼ mmol min1 mg1.
mechanisms are known. They catalyze the introduction of one oxygen atom into the
substrate, under concurrent production of one molecule of water as a coproduct.
These enzymes are important for detoxification, for example, in mammalian liver
cells, for the biosynthesis of secondary metabolites, and in prokaryotic hydrocarbon
degradation. Most published biocatalytic processes for preparative oxyfunctionaliza-
tions are based on oxygenases as biocatalysts, although these enzymes are interesting
for their ability to oxyfunctionalize hydrocarbons, rather than for their ability to
oxidize aldehydes [439, 440]. Flavin-dependent monooxygenases represent a highly
interesting group among the oxygenases, for selective oxyfunctionalization of non-
activated hydrocarbons.
Luciferase (EC 1.14.14.3) was one of the first monooxygenases for which the
oxidation of aliphatic aldehydes was described [441]. During the reaction light of
about 490 nm (blue-green) is emitted as a by-product. Luciferase has been described
from several different species, mainly of marine origin, like Photobacterium
fisheri [442], Latia neritoides [443], Oplophorus gracilirostris [444], and many others.
Bacterial luciferase is a heterodimeric enzyme composed of a catalytically active
a-subunit and a non-catalytic b-subunit, which plays an important role in stabiliza-
tion of the catalytic complex. The reduced cofactor FMN is non-covalently bound to
the enzyme and activates molecular oxygen, forming a 4a,5-dihydro-4a-hydroperoxy-
flavin. It is assumed that upon substrate binding a tetrahedral intermediate is
formed, which decomposes to the corresponding carboxylic acid, oxidized FMN,
and light [445]. Excited flavin species have been discussed as emitters (Fig-
ure 32.11) [441, 446].
Figure 32.11 Mechanism proposed for light emission during the luciferase reaction. Adapted from
Reference [441].
1414
Another highly interesting group of enzymes are P450-monooxygenases. Many P450

systems are multicomponent enzymes with small protein cofactors such as putidar-
edoxin performing the electron mediation between NAD(P)H and the active site of the
enzyme [447, 448]. As discussed above for the flavin dependent monooxygenases the
cofactor NADPH can be omitted from the catalytic cycle by direct electrochemical
reduction of putidaredoxin. This was shown for P450cam (EC 1.14.15.1) oxidizing
camphor or styrene to 5-exo-hydroxycamphor and styrene oxide, respectively
[449–451]. Other approaches utilize Co sepulchrate as reducing agent, which can be
regenerated either chemically (via Zn) [452] or electrochemically [453, 454].
The oxidation of an aldehyde to the corresponding carboxylic acid with P450
systems has been reported for various substrates (Table 32.22). In some cases
oxidative decarboxylation is observed, yielding formic acid and an olefin that is one
carbon atom shorter than the substrate (Table 32.23). In the overall reaction with
NADPH present, an oxygen-derived heme-iron bound hydroperoxide is believed to
react with the electrophilic aldehyde carbonyl group to form an enzyme-bound
peroxyhemiacetal-like intermediate, the rearrangement of which yields the olefin and
formic acid by a concerted or sequential b-scission mechanism.
32.3.5
Oxidases
Oxidases are best known for their role as terminal electron acceptors in energy
storage pathways. Oxidases typically contain flavins, iron, or copper as catalytic
centers and do not require any additional coenzymes, as they couple the one-, two-, or
four-electron oxidation of substrates to the two- or four-electron reduction of
molecular oxygen yielding hydrogen peroxide or water. Aldehydes are rather untyp-
ical substrates for oxidases. An exception is xanthine oxidase (EC 1.17.3.2), a two-
component, molybdenum iron-sulfur flavoprotein hydroxylase. It oxidizes a wide
variety of purines, pyrimidines, pterins, and aldehydes. Interestingly, xanthin oxidase
and xanthine dehydrogenase represent alternate forms of the same gene product.
While the natural electron acceptor for xanthin oxidase is molecular oxygen, the
dehydrogenase form accepts NAD þ [461]. The oxidation of aldehydes catalyzed by
xanthin oxidase was described as early as 1938 and was accomplished under
anaerobic conditions using dyes like methylene blue, phenazine methosulfate, or
ferricyanide as artificial electron acceptors [462]. Dastoli and coworkers transferred
this reaction to nonpolar solvents with solid enzyme, which also worked at acceptable
rates [463]. Table 32.24 gives kinetic data for some substrates achieved under aerobic
conditions using the natural electron acceptor molecular oxygen [464].
32.3.6
Aldehyde Oxidations with Intact Microbial Cells
In principle all reactions discussed in this chapter are possible by applying the
respective catalysts either in isolated form or as recombinant whole-cell biocatalyst.
Regarding enzymes depending on cofactors like NAD(P) þ for activity, whole cell
Table 32.22 Oxidation of aldehydes to the corresponding carboxylic acids catalyzed by P450
monooxygenases.
O O
R H
P450 R OH
O2 , NAD(P)H H 2O, NAD(P) +
Substrate Reference
Aliphatic aldehydes [455, 456]

O
R H
H [457]
H [458]
H 3C
O H
[458]
H O
OH
[459]
N Cl
N CHO
[460]
N
N
N N
1416
Table 32.23 Oxidations and subsequent decarboxylations of aldehydes catalyzed by P450

monooxygenases.
O
P450 R + HCO2 H
R H
O2, NAD(P)H H 2 O, NAD(P)+
Substrate Reference
O
[417–419]
H
CO 2H
O H
applications are often preferred over isolated enzymes, as the cofactor recycling issue
is solved by the host cell metabolism. Other advantages are enhanced enzyme
stability and no tedious enzyme purification work. On the down side there are
unspecific side reactions, degradation of the product, and/or mass transfer problems
of the substrate over the cell membrane. Especially, carboxylic acids are often directly
processed further to the central carbon metabolism of the host.
Often, aldehyde oxidation is part of a whole degradation pathway, starting from
some unactivated hydrocarbon, as described for the synthesis of quinoxaline-2-
carboxylic acid utilizing wild-type Pseudomonas putida ATCC3315 as biocatalyst
(Scheme 32.37).
Scheme 32.37 Synthesis of quinoxaline-2-carboxylic acid using wild-type Pseudomonas putida

ATCC33015 as whole-cell biocatalyst. Adapted from Reference [465].
Table 32.24 Kinetic constants for xanthine oxidase [464].
Substrate KM (mM) Vmax (s1)
O
161.5 22.2
H H
O
130 100
H
O
430 23.3
H
O
142 2.4
H
O
H 0.36 3.4
N
H 0.046 2.7
N
O
H 1.7 4.2
H 1.03 7.7
OH
O
HO
H 0.068 15.7
OH
This is an example of a process that has been established at Pfizer Inc. [465].
Detailed information about the process is scarce. Pseudomonas putida is induced by
benzyl alcohol. Induction and substrate feed have to be carefully controlled, as the
catalyst starts to be inhibited at inducer concentrations above 1 g l1 and substrate
1418
concentrations exceeding 1.5 g l1 [466]. The final yield was 86% on a 14-l scale
(10.7 g l1 after 99 h). The responsible enzymes are reported to be a monooxygenase,
a benzyl alcohol dehydrogenase, and a benzaldehyde dehydrogenase. Further
degradation of the acid was not reported.
The efficient transformation of aromatic aldehydes originating from lignin into the
corresponding acids was reported for Burkholderia cepacia. Vanillin, para-hydroxy-
benzaldehyde, and syringaldehyde were converted into the corresponding acids with
high yields of 94%, 92%, and 72%, respectively (Scheme 32.38) [467]. The produced
acid is not further metabolized as long as the aldehyde substrate is still available for
the cells. This conversion was carried out with resting cells in distilled water,
containing only the aldehyde and the biocatalyst. The respective enzymes have not
been identified. However, for an industrial application this would be absolutely
essential, as Burkholderia cepacia is a pathogenic microorganism causing pneumonia
and therefore not suited as an industrial biocatalyst.
Scheme 32.38 Oxidation of aldehydes by Burkholderia cepacia [467].
32.4
Oxidation of Carboxylic Acids
32.4.1
Introduction
At first glance, synthetically relevant oxidations of carboxylic acids, except for

oxidations at positions other than the carboxylate group, can hardly be found in
literature. Some preparative applications in whole cell catalysis, however, were
reported and will be discussed in the following (Scheme 32.39a–c). In vitro, the
high thermodynamic driving-force of the oxidation of formate and pyruvate (Eo
(formate/CO2) ¼ 0.42 V [468]; Eo (pyruvate/(acetate, CO2) ¼ 0.70 V [469]) are
used for the regeneration of coenzymes such as NAD(P)H or indirectly of ATP
(Scheme 32.39d and e).
32.4.2
Pyruvate Oxidase (EC 1.2.3.3)
Pyruvate oxidase is a flavoprotein, forming homotetramers with the cofactors FAD,

TPP (thiamine pyrophosphate), and Mg2 þ . These enzymes are mainly found in lactic
32.4 Oxidation of Carboxylic Acids j1419
Scheme 32.39 Synthetic and preparative applications of oxidations of acids: (a) multistep
oxidations of benzoic acid initiated by dihydroxylation; (b) oxidative decarboxylation; (c) cis-
dihydroxylation; (d) and (e) energy coupling for the regeneration of enzymatic coenzymes.
acid bacteria like Streptococcus pneumoniae [470], Lactobacillus plantarum [471, 472], or
Streptococcus sanguis [473]. Under anaerobic conditions, they depend on homolactic
fermentation for energy generation, in which glucose is metabolized to pyruvate and
finally to the fermentation product lactate. Under aerobic conditions these organisms
would have a problem, as they lack the necessary cytochromes to feed electrons into
the respiratory chain. This is solved by the concerted action of two enzymes, namely
pyruvate oxidase and acetate kinase. Pyruvate is decarboxylated to acetate by the
oxidase, yielding CO2, H2O2, and the energy-rich metabolite acetyl phosphate. Acetyl
phosphate is an important substrate for the enzyme acetate kinase (EC 2.7.2.1), which
catalyzes the transphosphorylation of various nucleotide diphosphates like ADP,
GDP, TDP, IDP, or UDP to the activated triphosphates [474, 475], which
will concurrently generate ATP from ADP by the action of the acetate kinase
[476, 477] (Scheme 32.40). This reaction can be applied in vitro to regenerate ATP
in ATP-dependent enzymatic reactions.
Scheme 32.40 Decarboxylative phosphorylation of pyruvate by pyruvate oxidase (PyOx) as driving

force for the acetate kinase (AK)-catalyzed regeneration of ATP.
1420
Pyruvate oxidase catalyzed regeneration of ATP was successfully coupled to in vitro

protein biosynthesis [478]. Cell-free protein synthesis has become an interesting
option to understand and analyze genetic information [479, 480]. Since gene
translation involves many steps consuming ATP, a sufficient supply of this reduction
equivalent is essential for this approach. Under aerobic conditions, no additional
regeneration system for the pyruvate oxidase is necessary; catalase, however, has to be
added to destroy harmful hydrogen peroxide.
Other pyruvate oxidases (e.g., from E. coli) do not couple the decarboxylation of
pyruvate to the generation of an energy-rich metabolite, but simply feed two reducing
equivalents from the cytosol to the respiratory chain [481, 482]. Thus these variants
are not suited for in vitro ATP regeneration.
32.4.3
Formate Dehydrogenase (EC 1.2.1.2)
Probably the most prominent oxidation of a carboxylic acid is catalyzed by the enzyme
formate dehydrogenase (FDH). FDH was isolated from various bacteria, yeasts, and
plants, where its physiological role is the regeneration of NADH [483]. While most
enzymes require metals or prosthetic groups for a successful hydride transfer, FDH
is an example of an enzyme that has neither metal ions nor any other prosthetic group
bound to its active center. These enzymes make good models to study the mechanism
of hydride ion transfer in the active center because the reaction is devoid of proton
transfer steps.
FDH catalyzes the oxidation of formate to CO2, cleaving a single carbon–hydrogen
bond in the substrate and concurrently forming a new one in the nicotinamide-
coenzyme, which also accepts the electrons (Scheme 32.41). Owing to the favorable
thermodynamic equilibrium of the reaction and the volatility of the reaction product,
the enzyme is commonly applied for in situ regeneration of NADH during asym-
metric synthesis of chiral compounds [88, 89, 484, 485]. Electrostatic effects are
responsible for the correct hydrogen transfer during the FDH catalyzed reaction.
Multiple hydrogen bonds within a positively charged amino acid cluster direct the
orientation of the anionic substrate formate, while a negatively charged amino acid
cluster and hydrophobic side chains correctly orient the positively charged nicotin-
amide NAD þ , so that it is placed with one side facing a hydrophobic wall and the
Scheme 32.41 Reductive animation of 2-keto acids utilizing leucine dehydrogenase from Bacillus
sphaericus coupled to cofactor regeneration catalyzed by formate dehydrogenase (FDH) [465].
32.4 Oxidation of Carboxylic Acids j1421
other side facing the hydrophilic substrate binding site. During the reaction various
stabilizing and destabilizing interactions occur in the active center, one of the most
important of which is the improvement of the electrophilic properties of the
nicotinamide moiety (C4N) of the coenzyme. The NAD þ carboxamide group
(C4N) is polarized via interaction with negatively charged ligands and perturbation
of its ground state due to the twist of the carboxamide with respect to the pyridine
plane. During catalysis, a hydride anion from the substrate formate attacks the
electrophilic C4N of the positively charged nicotinamide moiety of NAD þ . Conse-
quently, two neutral species, CO2 and NADH, are produced and the nicotinamide
moiety becomes uncharged [483].
FDH from Candida boidinii is very selective for NAD þ and is widely used as
regeneration enzyme. At Evonik it is applied in a leucine dehydrogenase catalyzed
reductive amination of 2-keto acids yielding various amino acids (e.g., tert-leu-
cine) [465, 486, 487].
In the past decade or so several FDH variants have been designed, which have
either a changed substrate specificity for NAD(P) þ [487] or enhanced stability
regarding temperature or reaction conditions [488].
32.4.4
Oxidations with Intact Microbial Cells
32.4.4.1 Production of Benzaldehyde from Benzoyl Formate or Mandelic Acid

Benzaldehyde can be produced from benzoyl formate via an oxidative decarboxyl-
ation using whole cells of Pseudomonas putida ATCC 12633 as biocatalyst. Alterna-
tively, but less effectively, mandelic acid can be used as starting material
(Scheme 32.42) [489, 490].
Scheme 32.42 Degradation of D-mandelic acid with a crude cell-free extract from Pseudomonas
aeruginosa. ManRac: mandelate racemase, ManDH: mandelate dehydrogenase, BFD:
benzoylformate dehydrogenase, and BDH: benzaldehyde dehydrogenase.
Owing to the partial inactivation of the benzaldehyde dehydrogenase isoenzymes

and activation of the benzoyl formate decarboxylase, a pH of 5.4 was found to be
optimal for the accumulation of the product benzaldehyde. While fed batch culti-
vation prevented substrate inhibition, in situ product removal is necessary to prevent
product inhibition. Activated charcoal served as solid-phase adsorption device [489].
Thus, benzaldehyde and thiophene-2-carboxaldehyde were obtained from benzoyl
1422
formate from thiophene-2-glyoxylic acid in final concentrations of up to 4.8 g l1 and

molar yields exceeding 85%.
32.4.4.2 Microbial Production of cis,cis-Muconic Acid from Benzoic Acid

cis,cis-Muconic acid is used as raw material for the synthesis of resins and polymers
(precursor of adipic acid). Furthermore, it is widely used as building-block in the
synthesis of pharmaceuticals and agrochemicals. Significant effort was put into its
synthesis via the oxidation of benzoic acid, utilizing a multistep reaction catalyzed by
whole microbial cells from various genera. Toluene, benzoic acid, or catechol have
been used as starting material [491].
Mizuno and coworkers reported on a mutant Arthrobacter strain (lacking cis,cis-
muconate derivatization activity), which was exploited for the synthesis of cis,cis-
muconic acid from benzoic acid (Scheme 32.43) [492].
Scheme 32.43 Sequential oxidation of benzoate to cis,cis-muconic acid catalyzed by Arthrobacter

sp. BDO: benzoate dioxygenase, DH: dehydrogenase, and CDO: catechol-1,2-dioxygenase.
The reaction cascade is initiated by a dioxygenation of the benzylic ring followed by

decarboxylation yielding catechol, which is transformed into the product via dioxy-
genase catalyzed ring cleavage. Benzoic acid was fed continuously to the fermen-
tation medium. The space–time yield of the process including downstream proces-
sing amounts to 70 g l1 d1 [492, 493].
32.4.4.3 Biotransformation of Substituted Benzoates into the Corresponding cis-Diols

Enantiopure 1,2-cis-dihydroxycyclohexa-3,5-diene carboxylic acids have considerable
synthetic potential as building blocks in chiral synthesis. Such cis-diols can be
produced from benzoic acid derivatives by the action of toluate-1,2-dioxygenase of
Pseudomonas putida mt-2 [494] or homologous enzymes of a different origin
(Scheme 32.44).
Scheme 32.44 Dioxygenation of benzoate to the corresponding cis-1,2-diols.
Growing cells or recombinant Pseudomonas oleovorans GPo12 containing toluate-

1,2-dioxygenase efficiently transformed a whole range of meta- and para-substituted
References j1423
benzoates into the corresponding cis-diols, which were not further degraded by the
Pseudomonas host. In the ortho position only hydrogen and fluorine were accepted as
substituents. Toluate-1,2-dioxygenase activity was induced by o-toluate or the sub-
strates themselves.
Similar reactions were reported for the broad-substrate-specific benzoate dioxy-
genase of Rhodococcus sp. strain 19070 [495]. Recombinant E. coli containing this
enzyme transformed benzoate and anthranilate into 2-hydro-1,2-dihydroxybenzoate
and catechol, respectively.
References
1 Hummel, W. (1997) New alcohol 16 Eklund, H. and Ramaswamy, S. (2008)

dehydrogenases for the synthesis of Cell Mol. Life Sci., 65, 3907.
chiral compounds, in Adv. Biochem. Eng. 17 Ramaswamy, S., Eklund, H., and
Biotechnol., vol. 58 (ed. T. Scheper), Plapp, B.V. (1994) Biochemistry, 33, 5230.
Springer, Berlin, p. 145. 18 Andersson, M., Holmberg, H., and
2 Prelog, V. (1964) Pure Appl. Chem., Adlercreutz, P. (1998) Biocatal.
9, 119. Biotransform., 16, 259.
3 Silverman, R.B. (2000) The Organic 19 Lortie, R., Villaume, I., Legoy, M.D., and
Chemistry of Enzyme-Catalyzed Reactions, Thomas, D. (1989) Biotechnol. Bioeng., 33,
Academic Press, San Diego. 229.
4 Chenault, H. and Whitesides, G. (1987) 20 Kawamoto, T., Aoki, A., Sonomoto, K.,
Appl. Biochem. Biotechnol., 14, 147. and Tanaka, A. (1989) J. Ferment. Bioeng.,
5 Kroutil, W., Mang, H., Edegger, K., and 67, 361.
Faber, K. (2004) http://www.brenda- 21 Pulvin, S., Parvaresh, F., Thomas, D., and
enzymes.org, Adv. Synth. Catal., 346, 125. Legoy, M.-D. (1988) Ann. N. Y. Acad. Sci.,
6 Reid, M.F. and Fewson, C.A. (1994) Crit. 542, 434.
Rev. Microbiol., 20, 13. 22 Lamare, S., Legoy, M.D., and Graber, M.
7 Faber, K. (2004) Biotransformations in (2004) Green Chem., 6, 445.
Organic Chemistry, Springer, Berlin. 23 Matos, J.R., Smith, M.B., and Wong, C.H.
8 Dworschack, R.T. and Plapp, B.V. (1977) (1985) Bioorg. Chem., 13, 121.
Biochemistry, 16, 111. 24 Jones, J.B. and Takemura, T. (1984) Can.
9 Ryzewski, C.N. and Pietruszko, R. (1977) J. Chem., 62, 77.
Arch. Biochem. Biophys., 183, 73. 25 Jones, J.B. and Schwartz, H.M. (1981)
10 Pietruszko, R. (1982) Methods Enzymol., Can. J. Chem., 59, 1574.
89, 428. 26 Yamazaki, Y. and Hosono, K. (1988)
11 Eklund, H., Nordstr€om, B., Tetrahedron Lett., 29, 5769.
Zeppezauer, E., S€oderlund, G., 27 Ganzhorn, A.J., Green, D.W., Hershey,
Ohlsson, I., Boiwe, T., S€oderberg, B.-O., A.D., Gould, R.M., and Plapp, B.V. (1987)

Tapia, O., Br€anden, C.-I., and Akeson, A. J. Biol. Chem., 262, 3754.
(1976) J. Mol. Biol., 102, 27. 28 Miroliaei, M. and Nemat-Gorgani, M.
12 Cedergren-Zeppezauer, E.S., (2002) Int. J. Biochem. Cell Biol., 34, 169.
Andersson, I., Ottonello, S., and 29 Men, L. and Wang, Y.S. (2007)
Bignetti, E. (2002) Biochemistry, 24, 4000. J. Proteome Res., 6, 216.
13 Wong, C.H. and Matos, J.R. (1985) 30 Li, G.Y., Huang, K.L., Jiang, Y.R.,
J. Org. Chem., 50, 1992. Yang, D.L., and Ding, P. (2008) Int. J. Biol.
14 Jones, J.B. and Jakovac, I.J. (1982) Macromol., 42, 405.
Can. J. Chem., 60, 19. 31 Yoshimoto, M., Sato, M., Yoshimoto, N.,
15 Luo, J. and Bruice, T.C. (2001) and Nakao, K. (2008) Biotechnol. Prog.,
J. Am. Chem. Soc., 123, 11952. 24, 576.
1424
32 Xu, S.W., Lu, Y., Li, J., Zhang, Y.F., and 52 Ziegelmann-Fjeld, K.I., Musa, M.M.,
Jiang, Z.Y. (2007) J. Biomater. Sci.-Polym. Phillips, R.S., Zeikus, J.G., and
E., 18, 71. Vieille, C. (2007) Protein Eng., Design
33 Dodds, D.R. and Jones, J.B. (1982) Select., 20, 47.
J. Chem. Soc., Chem. Commun., 1080. 53 Hollmann, F. and Otten, L.G. (2009) in
34 Eikeren, P., Brose, D.J., Muchmore, D.C., Wiley Encyclopedia of Chemical Biology
and West, J.B. (1990) Ann. N. Y. Acad. Sci., (ed. T.P. Begley), John Wiley & Sons, Inc.,
613, 796. Hoboken, doi:10.1002/9780470048672.
35 Eikeren, P., Muchmore, D.C., wecb150.
Brose, D.J., and Colton, R.H. (1992) Ann. 54 H€ollrigl, V., Hollmann, F., Kleeb, A.,
N. Y. Acad. Sci., 672, 539. Buehler, K., and Schmid, A. (2008) Appl.
36 Radianingtyas, H. and Wright, P.C. Microbiol. Biotechnol., 81, 263.
(2003) FEMS Microbiol. Rev., 27, 593. 55 H€ollrigl, V., Otto, K., and Schmid, A.
37 Keinan, E., Sinha, S.C., and (2007) Adv. Synth. Catal., 349, 1337.
Singh, S.P. (1991) Tetrahedron, 56 Hollmann, F., Kleeb, A., Otto, K., and
47, 4631. Schmid, A. (2005) Tetrahedron:
38 Keinan, E., Sinha, S.C., and Asymmetry, 16, 3512.
Sinhabagchi, A. (1991) J. Chem. Soc., 57 Stampfer, W., Kosjek, B., Moitzi, C.,
Perkin Trans. 1, 3333. Kroutil, W., and Faber, K. (2002) Angew
39 Olofsson, L., Nicholls, I.A., and Chem. Int. Ed., 41, 1014.
Wikman, S. (2005) Org. Biomol. Chem., 58 Stampfer, W., Kosjek, B., Kroutil, W., and
3, 750. Faber, K. (2003) Biotechnol. Bioeng., 81,
40 Keinan, E., Hafeli, E.K., Seth, K.K., and 865.
Lamed, R. (1986) J. Am. Chem. Soc., 108, 59 Kosjek, B., Stampfer, W., Pogorevc, M.,
162. Goessler, W., Faber, K., and Kroutil, W.
41 Raia, C.A., Giordano, A., Rossi, M., (2004) Biotechnol. Bioeng., 86, 55.
Adams, M.W.W., and Kelly, R.M. (2001) 60 Stampfer, W., Kosjek, B., Faber, K., and
Methods in Enzymology, vol. 331, Kroutil, W. (2003) J. Org. Chem.,
Academic Press, p. 176. 68, 402.
42 Pennacchio, A., Esposito, L., Zagari, A., 61 Kosjek, B., Stampfer, W., Deursen, R.V.,
Rossi, M., and Raia, C.A. (2009) Faber, K., and Kroutil, W. (2003)
Extremophiles, 13, 751. Tetrahedron, 59, 9517.
43 Bryant, F.O., Wiegel, J., and 62 Marshall, J.H., May, J.W., and
Ljungdahl, L.G. (1988) Appl. Environ. Sloan, J. (1985) J. Gen. Microbiol., 131,
Microbiol., 54, 460. 1581.
44 Pham, V.T., Phillips, R.S., and Ljungdahl, 63 Nishise, H., Maehashi, S., Yamada, H.,
L.G. (1989) J. Am. Chem. Soc., 111, 1935. and Tani, Y. (1987) Agric Biol. Chem.,
45 Pham, V.T. and Phillips, R.S. (1990) 51, 3347.
J. Am. Chem. Soc., 112, 3629. 64 Nishise, H., Nagao, A., Tani, Y., and
46 Phillips, R.S. (1992) Enzyme Microb. Yamada, H. (1984) Agric Biol. Chem., 48,
Technol., 14, 417. 1603.
47 Phillips, R.S. (1996) Trends Biotechnol., 65 Liese, A., Karutz, M., Kamphuis, J.,
14, 13. Wandrey, C., and Kragl, U. (1996)
48 Tripp, A.E., Burdette, D.S., Zeikus, J.G., Biotechnol. Bioeng., 51, 544.
and Phillips, R.S. (1998) J. Am. Chem. 66 Degenring, D., Schroder, I., Wandrey, C.,
Soc., 120, 5137. Liese, A., and Greiner, L. (2004) Org.
49 Phillips, R.S. (2002) J. Mol. Catal., B: Process Res. Dev., 8, 213.
Enzym., 19–20, 103. 67 Lee, L.G. and Whitesides, G.M. (1986)
50 Secundo, F. and Phillips, R.S. (1996) J. Org. Chem., 51, 25.
Enzyme Microb. Technol., 19, 487. 68 Temi~ no, D.M.-R.D., Hartmeier, W.,
51 Musa, M.M., Ziegelmann-Fjeld, K.I., and Ansorge-Schumacher, M.B.
Vieile, C., Zeikus, J.G., and Phillips, R.S. (2005) Enzyme Microb. Technol.,
(2007) J. Org. Chem., 72, 30. 36, 3.
References j1425
69 Bradshaw, C.W., Hummel, W., and 85 Riva, S., Bovara, R., Pasta, P., and
Wong, C.H. (1992) J. Org. Chem., 57, Carrea, G. (1986) J. Org. Chem.,
1532. 51, 2902.
70 Tarhan, L. and Hummel, W. (1995) 86 Riva, S., Bovara, R., Zetta, L., Pasta, P.,
Process Biochem., 30, 49. Ottolina, G., and Carrea, G. (1988)
71 Kallwass, H.K.W., Hogan, J.K., J. Org. Chem., 53, 88.
Macfarlane, E.L.A., Martichonok, V., 87 Liu, W. and Wang, P. (2007) Biotechnol.
Parris, W., Kay, C.M., Gold, M., and Adv., 25, 369.
Jones, J.B. (1992) J. Am. Chem. Soc., 114, 88 Wichmann, R. and Vasic-Racki, D.
10704. (2005) Technology Transfer in
72 Koszelewski, D., Clay, D., Rozzell, D., and Biotechnology: from Lab to Industry to
Kroutil, W. (2009) Eur. J. Org. Chem., Production, vol. 92, Springer-Verlag,
2289. Berlin, p. 225.
73 Manj on, A., Obon, J.M., Casanova, P., 89 van der Donk, W.A. and Zhao, H. (2003)
Fernandez, V.M., and Ilborra, J.L. (2002) Curr. Opin. Biotechnol., 14, 421.
Biotechnol. Lett., 24, 1227. 90 Hollmann, F., Hofstetter, K., and
74 Fassouane, A., Laval, J.-M., Moiroux, J., Schmid, A. (2006) Trends Biotechnol.,
and Bourdillon, C. (1990) Biocatal. 24, 163.
Biotransform., 35, 935. 91 Hollmann, F. and Schmid, A. (2004)
75 Zhang, W., OConnor, K., Wang, D.I.C., Biocatal. Biotransform., 22, 63.
and Li, Z. (2009) Appl. Environ. Microbiol., 92 Kohlmann, C., M€arkle, W., and
75, 687. L€utz, S. (2008) J. Mol. Catal., B: Enzym.,
76 Schewe, H., Kaup, B.-A., and 51, 57.
Schrader, J. (2008) Appl. Microbiol. 93 Wandrey, C. (2004) Chem. Rec., 4, 254.
Biotechnol., 78, 55. 94 Stampfer, W., Kosjek, B., Faber, K., and
77 Chaparro-Riggers, J.F., Rogers, T.A., Kroutil, W. (2003) Tetrahedron:
Vazquez-Figueroa, E., Polizzi, K.M., and Asymmetry, 14, 275.
Bommarius, A.S. (2007) Adv. Synth. 95 Lavandera, I., Kern, A., Resch, V.,
Catal., 349, 1521. Ferreira-Silva, B., Glieder, A.,
78 Wada, M., Yoshizumi, A., Noda, Y., Fabian, W.M.F., de Wildeman, S., and
Kataoka, M., Shimizu, S., Takagi, H., and Kroutil, W. (2008) Org. Lett.,
Nakamori, S. (2003) Appl. Environ. 10, 2155.
Microbiol., 69, 933. 96 Sheldon, R.A. (2008) Chem. Commun.,
79 Stueckler, C., Hall, M., Ehammer, H., 3352.
Pointner, E., Kroutil, W., Macheroux, P., 97 Hirano, J.-I., Miyamoto, K., and
and Faber, K. (2007) Org. Lett., 9, 5409. Ohta, H. (2008) Tetrahedron Lett.,
80 Anelli, P.L., Brocchetta, M., Morosini, P., 49, 1217.
Palano, D., Carrea, G., Falcone, L., 98 Geueke, B., Riebel, B., and Hummel, W.
Pasta, P., and Sartore, D. (2002) Biocatal. (2003) Enzyme Microb. Technol.,
Biotransform., 20, 29. 32, 205.
81 Monti, D., Ferrandi, E.E., Zanellato, I., 99 Hirano, J.-I., Miyamoto, K., and
Hua, L., Polentini, F., Carrea, G., and Ohta, H. (2007) Appl. Microbiol.
Riva, S. (2009) Adv. Synth. Catal., 351, Biotechnol., 76, 357.
1303. 100 De Muynck, C., Pereira, C.S.S.,
82 Fossati, E., Polentini, F., Carrea, G., and Naessens, M., Parmentier, S.,
Riva, S. (2006) Biotechnol. Bioeng., 93, Soetaert, W., and Vandamme, E.J. (2007)
1216. Crit. Rev. Biotechnol., 27, 147.
83 Kristan, K., Stojan, J., Adamski, J., and 101 Findrik, Z., Simunovic, I., and
Rizner, T.L. (2007) J. Biotechnol., Vasic-Racki, -D
- . (2008) Biochem. Eng. J.,
129, 123. 39, 319.
84 Pedrini, P., Andreotti, E., Guerrini, A., 102 Jiang, R. and Bommarius, A.S.
Dean, M., Fantin, G., and Giovannini, (2004) Tetrahedron: Asymmetry,
P.P. (2006) Steroids, 71, 189. 15, 2939.
1426
103 Riebel, B.R., Gibbs, P.R., Wellborn, W.B., 122 Bailey, S.I. and Ritchie, I.M. (1985)
and Bommarius, A.S. (2003) Adv. Synth. Electrochim. Acta, 30, 3.
Catal., 345, 707. 123 Zare, H.R. and Golabi, S.M. (1999)
104 Riebel, B.R., Gibbs, P.R., Wellborn, W.B., J. Electroanal. Chem., 464, 14.
and Bommarius, A.S. (2002) Adv. Synth. 124 Zare, H.R. and Golabi, S.M. (2000)
Catal., 344, 1156. J. Solid State Electron., 4, 87.
105 €
Odman, P., Wellborn, W.B., and 125 Hilt, G. and Steckhan, E. (1993)
Bommarius, A.S. (2004) Tetrahedron: J. Chem. Soc., Chem. Commun., 1706.
Asymmetry, 15, 2933. 126 Hilt, G., Lewall, B., Montero, G.,
106 Aksu, S., Arends, I.W.C.E., and Utley, J.H.P., and Steckhan, E. (1997)
Hollmann, F. (2009) Adv. Synth. Catal., Liebigs Annalen/Recueil, 2289.
351, 1211. 127 Komoschinski, J. and Steckhan, E.
107 Willetts, A.J., Knowles, C.J., Levitt, M.S., (1988) Tetrahedron Lett., 29, 3299.
Roberts, S.M., Sandey, H., and 128 Retna Raj, C. and Ohsaka, T. (2001)
Shipston, N.F. (1991) J. Chem. Soc., Perkin Electrochem. Commun., 3, 633.
Trans. 1, 1608. 129 Raj, C.R. and Chakraborty, S. (2006)
108 Tsuji, Y., Fukui, T., Kawamoto, T., and Biosens. Bioelectron., 22, 700.
Tanaka, A. (1994) Appl. Microbiol. 130 Fotouhi, L., Raei, F., Heravi, M.M., and
Biotechnol., 41, 219. Nematollahi, D. (2010) J. Electroanal.
109 Jones, J.B. and Taylor, K.E. (1973) Chem., 639, 15.
J. Chem. Soc., Chem. Commun., 205. 131 Hilt, G., Jarbawi, T., Heineman, W.R., and
110 Irwin, A.J. and Jones, J.B. (1977) Steckhan, E. (1997) Chem. – Eur. J.,
J. Am. Chem. Soc., 99, 556. 3, 79.
111 Irwin, A.J. and Jones, J.B. (1977) 132 Bardea, A., Katz, E., Buckmann, A.F., and
J. Am. Chem. Soc., 99, 1625. Willner, I. (1997) J. Am. Chem. Soc.,
112 Lok, K.P., Jakovac, I.J., and 119, 9114.
Jones, J.B. (1985) J. Am. Chem. Soc., 133 Nowall, W.B. and Kuhr, W.G. (2002) Anal.
107, 2521. Chem., 67, 3583.
113 Jakovac, I.J., Goodbrand, H.B., 134 Ju, H. and Leech, D. (1997) Anal. Chimi.
Lok, K.P., and Jones, J.B. (1982) J. Am. Acta, 345, 51.
Chem. Soc., 104, 4659. 135 Malinauskas, A., Ruzgas, T., and
114 Hollmann, F., Witholt, B., and Gorton, L. (2000) J. Colloid. Interf. Sci.,
Schmid, A. (2002) J. Mol. Catal., B: 224, 325.
Enzym., 19–20, 167. 136 Sadakane, M. and Steckhan, E. (1998)
115 Steckhan, E., Arns, T., Heineman, W.R., Chem. Rev., 98, 219.
Hilt, G., Hoormann, D., J€orissen, J., 137 Atta, N.F., Galal, A., Karag€ozler, A.E.,
Kr€oner, L., Lewall, B., and Zimmer, H., Rubinson, J.F., and
P€utter, H. (2001) Chemosphere, Mark, H.B. (1990) J. Chem. Soc., Chem.
43, 63. Commun., 1347.
116 Radoi, A. and Compagnone, D. (2009) 138 Schr€oder, I., Steckhan, E., and Liese, A.
Bioelectrochemistry, 76, 126. (2003) J. Electroanal. Chem., 541, 109.
117 Biade, A.E., Bourdillon, C., Laval, J.M., 139 Willner, I. and Mandler, D. (1989) Enzyme
Mairesse, G., and Moiroux, J. (1992) Microb. Technol., 11, 467.
J. Am. Chem. Soc., 114, 893. 140 Handman, J., Harriman, A., and
118 Anne, A., Bourdillon, C., Daninos, S., and Porter, G. (1984) Nature, 307, 534.
Moiroux, J. (1999) Biotechnol. Bioeng., 141 Julliard, M. and Lepetit, J. (1982)
64, 101. Photochem. Photobiol., 36, 283.
119 Kim, Y.H. and Yoo, Y.J. (2009) 142 Julliard, M., Le Petit, J., and Ritz, P. (1986)
Enzyme Microb. Technol., 44, 129. Biotechnol. Bioeng., 28, 1774.
120 Degrand, C. and Miller, L.L. (2002) 143 Ruppert, R. and Steckhan, E. (1989)
J. Am. Chem. Soc., 102, 5728. J. Chem. Soc., Perkin Trans. 2, 811.
121 Murthy, A.S.N. and Sharma, J. (1998) 144 Freeman, R. and Willner, I. (2009) Nano
Talanta, 45, 951. Lett., 9, 322.
References j1427
145 Jongejan, A., Machado, S.S., and Theodoulou, F.L., and Foyer, C.H. (2003)
Jongejan, J.A. (2000) J. Mol. Catal., B: Plant Physiol., 133, 443.
Enzym., 8, 121. 162 Hancock, R.D. and Viola, R. (2002) Trends
146 Duine, J.A. (1991) Eur. J. Biochem., Biotechnol., 20, 299.
200, 271. 163 Bauer, R., Katsikis, N., Varga, S., and
147 Wang, S.X., Mure, M., Hekmat, D. (2005) Bioprocess Biosystems
Medzihradszky, K.F., Burlingame, A.L., Eng., 28, 37.
Brown, D.E., Dooley, D.M., Smith, A.J., 164 De Muynck, C., Pereira, C., Soetaert, W.,
Kagan, H.M., and Klinman, J.P. (1996) and Vandamme, E. (2006) J. Biotechnol.,
Science, 273, 1078. 125, 408.
148 Janes, S., Mu, D., Wemmer, D., Smith, A., 165 Landis, B.H., McLaughlin, J.K.,
Kaur, S., Maltby, D., Burlingame, A., and Heeren, R., Grabner, R.W., and
Klinman, J. (1990) Science, 248, 981. Wang, P.T. (2002) Org. Process Res. Dev.,
149 McIntire, W., Wemmer, D., 6, 547.
Chistoserdov, A., and Lidstrom, M. (1991) 166 Adachi, O., Tayama, K., Shinagawa, E.,
Science, 252, 817. Matsushita, K., and Ameyama, M. (1978)
150 Klinman, J.P., Dooley, D.M., Duine, J.A., Agric Biol. Chem., 42, 2045.
Knowles, P.F., Mondovi, B., and 167 Adachi, O., Tayama, K., Shinagawa, E.,
Villafranca, J.J. (1991) FEBS Lett., Matsuhita, K., and Ameyama, M. (1980)
282, 1. Agric Biol. Chem., 44, 503.
151 Cozier, G.E., Giles, I.G., and Anthony, C. 168 Ameyama, M., Shinagawa, E.,
(1995) Biochem. J., 308, 375. Matsushita, K., and Adachi, O. (1981)
152 Kondo, K. and Horinouchi, S. (1997) Agric Biol. Chem., 45, 851.
Appl. Environ. Microbiol., 63, 1131. 169 Shinagawa, E., Matsushita, K.,
153 Kondo, K., Beppu, T., and Horinouchi, S. Adachi, O., and Ameyama, M. (1984)
(1995) J. Bacteriol., 177, 5048. Agric Biol. Chem., 48, 1517.
154 Matsushita, K., Yakushi, T., 170 Shinagawa, E., Matsushita, K.,
Toyama, H., Shinagawa, E., and Adachi, O., and Ameyama, M. (1981)
Adachi, O. (1996) J. Biol. Chem., 271, Agric Biol. Chem., 45, 1079.
4850. 171 Shinagawa, E., Matsushita, K.,
155 de Jong, G.A.H., Caldeira, J., Sun, J., Adachi, O., and Ameyama, M. (1982)
Jongejan, J.A., de Vries, S., Loehr, T.M., Agric Biol. Chem., 46, 135.
Moura, I., Moura, J.J.G., and Duine, J.A. 172 Choi, E.-S., Lee, E.-H., and Rhee, S.-K.
(2002) Biochemistry, 34, 9451. (1995) FEMS Microbiol. Lett.,
156 Govardus, A.H.J., Arie, G., Joke, S., 125, 45.
Jaap, A.J., Simon, V., and Johannis, A.D. 173 Sugisawa, T. and Hoshino, T. (1999)
(1995) Eur. J. Biochem., 230, 899. Annual Meeting of the Japan-Society-for-
157 Matsushita, K., Toyama, H., Adachi, O., Bioscience-Biotechnology-and-
and Tempest, A.H.R.A.D.W. (1994) Agrochemistry, Japan Society for
Advances in Microbial Physiology, vol. 36, Bioscience, Biotechnology, and
Academic Press, p. 247. Agrochemistry, Fukuoka, p. 57.
158 Leferink, N.G.H., Heuts, D.P.H.M., 174 Sugisawa, T., Hoshino, T., Nomura, S.,
Fraaije, M.W., and van Berkel, W.J.H. and Fujiwara, A. (1991) Agric Biol. Chem.,
(2008) Arch. Biochem. Biophys., 474, 292. 55, 363.
159 Leferink, N.G.H., Fraaije, M.W., 175 Ameyama, M., Shinagawa, E.,
Joosten, H.J., Schaap, P.J., Mattevi, A., Matsushita, K., and Adachi, O. (1985)
and van Berkel, W.J.H. (2009) J. Biol. Agric Biol. Chem., 49, 1001.
Chem., 284, 4392. 176 Adachi, O., Fujii, Y., Ano, Y.,
160 Henriksson, G., Johansson, G., and Moonmangmee, D., Toyama, H.,
Pettersson, G. (2000) J. Biotechnol., Shinagawa, E., Theeragool, G.,
78, 93. Lotong, N., and Matsushita, K.
161 Millar, A.H., Mittova, V., Kiddle, G., (2001) Biosci. Biotechnol. Biochem.,
Heazlewood, J.L., Bartoli, C.G., 65, 115.
1428
177 Holscher, T., Weinert-Sepalage, D., and andez-Lafuente, R., Rodrıguez, V.,
192 Fern
Gorisch, H. (2007) Microbiol.-Sgm, Mateo, C., Fernandez-Lorente, G.,
153, 499. Arminsen, P., Sabuquillo, P., and
178 Matsushita, K., Fujii, Y., Ano, Y., Guisan, J.M. (1999) J. Mol. Catal., B:
Toyama, H., Shinjoh, M., Tomiyama, N., Enzym., 7, 173.
Miyazaki, T., Sugisawa, T., Hoshino, T., 193 Schussel, L.J. and Atwater, J.E.
and Adachi, O. (2003) Appl. Environ. (1996) Enzyme Microb. Technol.,
Microbiol., 69, 1959. 18, 229.
179 Adachi, O., Ano, Y., Moonmangmee, D., 194 Bourdillon, C., Lortie, R., and
Shinagawa, E., Toyama, H., Laval, J.M. (1988) Biotechnol. Bioeng.,
Theeragool, G., Lotong, N., and 31, 553.
Matsushita, K. (1999) Biosci. Biotechnol. 195 Petersen, A. and Steckhan, E. (1999)
Biochem., 63, 2137. Bioorg. Med. Chem., 7, 2203.
180 Hill, H.A.O., Oliver, B.N., Page, D.J., and 196 Leitner, C., Mayr, P., Riva, S.,
Hopper, D.J. (1985) J. Chem. Soc., Chem. Volc, J., Kulbe, K.D., Nidetzky, B., and
Commun., 1469. Haltrich, D. (2001) J. Mol. Catal., B:
181 Frede, M. and Steckhan, E. (1991) Enzym., 11, 407.
Tetrahedron Lett., 32, 5063. 197 Van Hecke, W., Salaheddin, C.,
182 Brielbeck, B., Frede, M., and Ludwig, R., Dewulf, J., Haltrich, D., and
Steckhan, E. (1993) European Symposium Langenhove, H.V. (2009) Bioresour.
on Biocatalysis, Harwood Academic Technol., 100, 5566.
Publisher GmbH, Graz, Austria, p. 49. 198 Sukyai, P., Rezic, T., Lorenz, C.,
183 Baminger, U., Ludwig, R., Galhaup, C., Mueangtoom, K., Lorenz, W.,
Leitner, C., Kulbe, K.D., and Haltrich, D. Haltrich, D., and Ludwig, R. (2008)
(2001) J. Mol. Catal., B: Enzym., J. Biotechnol., 135, 281.
11, 541. 199 Bartlett, P.N., Whitaker, R.G.,
184 Ludwig, R., Ozga, M., Zamocky, M., Green, M.J., and Frew, J. (1987)
Peterbauer, C., Kulbe, K.D., and J. Chem. Soc., Chem. Commun., 1603.
Haltrich, D. (2004) Biocatal. 200 Degani, Y. and Heller, A. (1988)
Biotransform., 22, 97. J. Am. Chem. Soc., 110, 2615.
185 Maischberger, T., Nguyen, T.-H., 201 Heller, A. (1990) Acc. Chem. Res.,
Sukyai, P., Kittl, R., Riva, S., Ludwig, R., 23, 128.
and Haltrich, D. (2008) Carbohydr. Res., 202 Johnson, J.M., Halsall, H.B., and
343, 2140. Heineman, W.R. (1985) Biochemistry,
186 Van Hecke, W., Bhagwat, A., Ludwig, R., 24, 1579.
Dewulf, J., Haltrich, D., and 203 Zhao, S. and Lennox, R.B. (2002) Anal.
Van Langenhove, H. (2009) Biotechnol. Chem., 63, 1174.
Bioeng., 102, 1475. 204 Nakaminami, T., Kuwabata, S., and
187 Van Hecke, W., Ludwig, R., Dewulf, J., Yoneyama, H. (1997) Anal. Chem., 69,
Auly, M., Messiaen, T., Haltrich, D., and 2367.
Van Langenhove, H. (2009) Biotechnol. 205 Shumakovich, G., Shleev, S.,
Bioeng., 102, 122. Morozova, O., Gonchar, M., and
188 van Hellemond, E.W., Leferink, N.G.H., Yaropolov, A. (2007) Appl. Biochem.
Heuts, D.P.H.M., Fraaije, M.W., Micro., 43, 15.
van Berkel, W.J.H., Laskin, S.S.A.I., and 206 Koopal, C.G.J., de Ruiter, B., and
Geoffrey, M.G. (2006) Advances in Nolte, R.J.M. (1991) J. Chem. Soc., Chem.
Applied Microbiology, vol. 60, Academic Commun., 1691.
Press, p. 17. 207 Yehezkeli, O., Yan, Y.M.,
189 Paulina, O., Marten, V., and Ida, J.K. Baravik, I., Tel-Vered, R., and
(2005) FEMS Yeast Res., 5, 975. Willner, I. (2009) Chem.-Eur. J.,
190 Giffhorn, F. (2000) Appl. Microbiol. 15, 2674.
Biotechnol., 54, 727. 208 Duff, S.J.B. and Murray, W.D. (1989)
191 Riva, S. (2006) Trends Biotechnol., 24, 219. Biotechnol. Bioeng., 34, 153.
References j1429
209 Borzeix, F., Monot, F., and 226 Whittaker, J.W. (2005) Arch. Biochem.
Vandecasteele, J.-P. (1995) Enzyme Biophys., 433, 227.
Microb. Technol., 17, 615. 227 Whittaker, J.W. (2003) Chem. Rev., 103,
210 Dienys, G., Jarmalavicius, S., 2347.
Budriene, S., Citavicius, D., and 228 Drueckhammer, D.G., Hennen, W.J.,
Sereikaite, J. (2003) J. Mol. Catal., B: Pederson, R.L., Barbas, C.F.,
Enzym., 21, 47. Gautheron, C.M., Krach, T., and
211 Karra-Chaabouni, M., Pulvin, S., Wong, C.H. (1991) Synthesis
Meziani, A., Thomas, D., Touraud, D., (Stuttgart), 499.
and Kunz, W. (2003) Biotechnol. Bioeng., 229 Klibanov, A.M., Alberti, B.N., and
81, 27. Marletta, M.A. (1982) Biochem. Biophys.
212 Ukeda, H., Ishii, T., Sawamura, M., and Res. Commun., 108, 804.
Isobe, K. (1998) Biosci. Biotechnol. 230 Sun, L., Bulter, T., Alcalde, M.,
Biochem., 62, 1589. Petrounia, I.P., and Arnold, F.H. (2002)
213 Isobe, K. and Nishise, H. (1995) ChemBioChem, 3, 781.
J. Mol. Catal., B: Enzym., 1, 37. 231 Basu, S.S., Dotson, G.D., and
214 Ferreira, P., Medina, M., Guillen, F., Raetz, C.R.H. (2000) Anal. Biochem.,
Martinez, M.J., Van Berkel, W.J.H., and 280, 173.
Martinez, A.T. (2005) Biochem. J., 389, 232 Schoevaart, R. and Kieboom, T. (2004)
731. Top. Catal., 27, 3.
215 Bankar, S.B., Bule, M.V., Singhal, R.S., 233 van Wijk, A., Siebum, A., Schoevaart, R.,
and Ananthanarayan, L. (2009) and Kieboom, T. (2006) Carbohydr. Res.,
Biotechnol. Adv., 27, 489. 341, 2921.
216 Kelley, R.L. and Reddy, C.A. (1988) 234 Prongjit, M., Sucharitakul, J.,
Methods Enzymol., 161, 307. Wongnate, T., Haltrich, D., and
217 Kelley, R.L. and Reddy, C.A. (1986) Chaiyen, P. (2009) Biochemistry,
J. Bacteriol., 166, 269. 48, 4170.
218 Santoni, T., Santianni, D., Manzoni, A., 235 Freimund, S., Huwig, A., Giffhorn, F.,
Zanardi, S., and Mascini, M. (1997) and K€opper, S. (1998) Chem. Eur. J., 4,
Talanta, 44, 1573. 2442.
219 Vodopivec, M., Berovic, M., Jancar, J., 236 Leitner, C., Neuhauser, W., Volc, J.,
Podgornik, A., and Strancar, A. (2000) Kulbe, K.D., Nidetzky, B., and Haltrich,
Anal. Chim. Acta, 407, 105. D. (1998) Biocatal. Biotransform.,
220 Chudobova, I., Vrbova, E., Kodıcek, M., 16, 365.
Janovcova, J., and Kas, J. (1996) Anal. 237 Spadiut, O., Pisanelli, I.,
Chim. Acta, 319, 103. Maischberger, T., Peterbauer, C.,
221 Ryabov, A.D., Firsova, Y.N., Goral, V.N., Gorton, L., Chaiyen, P., and Haltrich, D.
Ryabova, E.S., Shevelkova, A.N., (2009) J. Biotechnol., 139, 250.
Troitskaya, L.L., Demeschik, T.V., and 238 Giffhorn, F., K€opper, S., Huwig, A., and
Sokolov, V.I. (1998) Chem. Eur. J., Freimund, S. (2000) Enzyme. Microb.
4, 806. Technol., 27, 734.
222 Guillaume, D., Irene, C.R., 239 Dhariwal, A., Mavrov, V., and
Jeroen, J.L.M.C., and Roeland, J.M.N. Schroeder, I. (2006) J. Mol. Catal., B:
(2009) Chem. Eur. J., 15, 12600. Enzym., 42, 64.
223 Okrasa, K., Guibe-Jampel, E., and 240 Heuts, D.P.H.M., van Hellemond, E.W.,
Therisod, M. (2003) Tetrahedron: Janssen, D.B., and Fraaije, M.W. (2007)
Asymmetry, 14, 2487. J. Biol. Chem., 282, 20283.
224 van de Velde, F., Lourenço, N.D., 241 Forneris, F., Heuts, D., Delvecchio, M.,
Bakker, M., van Rantwijk, F., and Rovida, S., Fraaije, M.W., and Mattevi, A.
Sheldon, R.A. (2000) Biotechnol. Bioeng., (2008) Biochemistry, 47, 978.
69, 286. 242 van Bloois, E., Winter, R.T., Janssen, D.B.,
225 Parikka, K. and Tenkanen, M. (2009) and Fraaije, M.W. (2009) Appl. Microbiol.
Carbohydr. Res., 344, 14. Biotechnol., 83, 679.
1430
243 van Hellemond, E.W., Vermote, L., Vanderplas, H.C., Meijer, E.M., and
Koolen, W., Sonke, T., Zandvoort, E., Schoemaker, H.E. (1987) Biotechnol.
Heuts, D.P., Janssen, D.B., and Bioeng., 30, 607.
Fraaije, M.W. (2009) Adv. Synth. Catal., 260 Winter, J.M. and Moore, B.S. (2009)
351, 1523. J. Biol. Chem., 284, 18577.
244 Adam, W., Lazarus, M., 261 Renirie, R., Pierlot, C., Aubry, J.-M.,
Saha-M€oller, C.R., and Schreier, P. (1998) Hartog, A.F., Schoemaker, H.E.,
Tetrahedron: Asymmetry, 9, 351. Alsters, P.L., and Wever, R. (2003) Adv.
245 Adam, W., Lazarus, M., Boss, B., Synth. Catal., 345, 849.
Saha-Moller, C.R., Humpf, H.-U., and 262 Renirie, R., Pierlot, C., Wever, R., and
Schreier, P. (1997) J. Org. Chem., Aubry, J.-M. (2009) J. Mol. Catal., B:
62, 7841. Enzym., 56, 259.
246 Eisenberg, A., Seip, J.E., Gavagan, J.E., 263 Coughlin, P., Roberts, S., Rush, C., and
Payne, M.S., Anton, D.L., and Willetts, A. (1993) Biotechnol. Lett., 15,
DiCosimo, R. (1997) J. Mol. Catal., B: 907.
Enzym., 2, 223. 264 ten Brink, H.B., Schoemaker, H.E., and
247 Seip, J.E., Fager, S.K., Gavagan, J.E., Wever, R. (2001) Eur. J. Biochem., 268,
Gosser, L.W., Anton, D.L., and 132.
DiCosimo, R. (1993) J. Org. Chem., 58, 265 Kuwahara, M., Glenn, J.K.,
2253. Morgan, M.A., and Gold, M.H. (1984)
248 Gavagan, J.E., Fager, S.K., Seip, J.E., FEBS Lett., 169, 247.
Clark, D.S., Payne, M.S., Anton, D.L., and 266 Gold, M.H., Youngs, H.L., and
DiCosimo, R. (1997) J. Org. Chem., 62, Gelpke, M.D.S. (2000) Metal Ions in
5419. Biological Systems, vol. 37,
249 Seip, J.E., Fager, S.K., Gavagan, J.E., Marcel Dekker, New York, p. 559.
Anton, D.L., and Di Cosimo, R. (1994) 267 Manoj, K.M. and Hager, L.P. (2008)
Bioorg. Med. Chem., 2, 371. Biochemistry, 47, 2997.
250 MacLachlan, J., Wotherspoon, A.T.L., 268 Hofrichter, M. and Ullrich, R.
Ansell, R.O., and Brooks, C.J.W. (2006) Appl. Microbiol. Biotechnol.,
(2000) J. Steroid Biochem. Mol. Biol., 72, 71, 276.
169. 269 Valderrama, B., Ayala, M., and
251 Doukyu, N. (2009) Appl. Microbiol. Vazquez-Duhalt, R. (2002) Chem. Biol.,
Biotechnol., 83, 825. 9, 555.
252 Pollegioni, L., Piubelli, L., and Molla, G. 270 Van Deurzen, M.P.J., Seelbach, K.,
(2009) FEBS J., 276, 6857. van Rantwijk, F., Kragl, U., and
253 Dieth, S., Tritsch, D., and Biellmann, J.F. Sheldon, R.A. (1997) Biocatal.
(1995) Tetrahedron Lett., 36, 2243. Biotransform., 15, 1.
254 Khmelnitsky, Y., Hilhorst, L.R., and 271 Grey, C.E., Hedstr€om, M., and
Verger, C. (1988) Eur. J. Biochem., 176, Adlercreutz, P. (2007) ChemBioChem, 8,
265. 1055.
255 Labaree, D., Hoyte, R.M., and 272 Park, J.-B. and Clark, D.S. (2006)
Hochberg, R.B. (1997) Steroids, Biotechnol. Bioeng., 93, 1190.
62, 482. 273 Mahmoudi, A., Nazari, K.,
256 Alexander, D.L. and Fisher, J.F. (1995) Khosraneh, M., Mohajerani, B.,
Steroids, 60, 290. Kelay, V., and Moosavi-Movahedi, A.A.
257 Nakaminami, T., Ito, S., Kuwabata, S., and (2008) Enzyme Microb. Technol.,
Yoneyama, H. (1999) Anal. Chem., 71, 43, 329.
1068. 274 Durand, A., Lalot, T., Brigodiot, M., and
258 Mugesh, G., du Mont, W.W., and Marechal, E. (2000) Polymer,
Sies, H. (2001) Chem. Rev., 41, 8183.
101, 2125. 275 Perez, D.I., van Rantwijk, F., and
259 Deboer, E., Plat, H., Tromp, M.G.M., Sheldon, R.A. (2009) Adv. Synth. Catal.,
Wever, R., Franssen, M.C.R., 351, 2133.
References j1431
276 Jung, D., Paradiso, M., Wallacher, D., 294 Pezzotti, F. and Therisod, M. (2007)
Brandt, A., and Hartmann, M. (2009) Tetrahedron: Asymmetry, 18, 701.
ChemSusChem, 2, 161. 295 La Rotta, C.E., DElia, E., and
277 Wang, W., Xu, Y., Wang, D.I.C., and Bon, E.P.S. (2007) Electron. J. Biotechnol.,
Li, Z. (2009) J. Am. Chem. Soc., 131, 10, 24.
12892. 296 Baciocchi, E., Fabbrini, M.,
278 Roberge, C., Amos, D., Pollard, D., and Lanzalunga, O., Manduchi, L., and
Devine, P. (2009) J. Mol. Catal., B: Pochetti, G. (2001) Eur. J. Biochem., 268,
Enzym., 56, 41. 665.
279 de Hoog, H.M., Nallani, M., 297 Kiljunen, E. and Kanerva, L.T. (2000)
Cornelissen, J., Rowan, A.E., J. Mol. Catal., B: Enzym., 9, 163.
Nolte, R.J.M., and Arends, I. (2009) Org. 298 Kiljunen, E. and Kanerva, L.T. (1999)
Biomol. Chem., 7, 4604. Tetrahedron: Asymmetry, 10, 3529.
280 Grey, C.E., Rundb€ack, F., and 299 Lindborg, J., Tanskanen, A., and
Adlercreutz, P. (2008) J. Biotechnol., 135, Kanerva, L.T. (2009) Biocatal.
196. Biotransform., 27, 204.
281 Seelbach, K., van Deurzen, M.P.J., 300 Hu, S. and Hager, L.P. (1998) Biochem.
van Rantwijk, F., Sheldon, R.A., and Biophys. Res. Commun., 253, 544.
Kragl, U. (1997) Biotechnol. Bioeng., 301 Geigert, J., Dalietos, D.J.,
55, 283. Neidleman, S.L., Lee, T.D., and
282 Kohlmann, C., Greiner, L., Leitner, W., Wadsworth, J. (1983) Biochem. Biophys.
Wandrey, C., and L€ utz, S. (2009) Chem. Res. Commun., 114, 1104.
Eur. J., 15, 11692. 302 Samra, B.K., Andersson, M., and
283 Sanfilippo, C., DAntona, N., and Adlercreutz, P. (1999) Biocatal.
Nicolosi, G. (2004) Biotechnol. Lett., 26, Biotransform., 17, 381.
1815. 303 van Deurzen, M.P.J., van Rantwijk, F.,
284 Lichtenecker, R.J. and Schmid, W. (2009) and Sheldon, R.A. (1997) J. Carbohydr.
Monatsh. Chem., 140, 509. Res., 16, 299.
285 Lakner, F.J., Cain, K.P., and Hager, L.P. 304 Kunamneni, A., Camarero, S.,
(1997) J. Am. Chem. Soc., 119, 443. Garcia-Burgos, C., Plou, F., Ballesteros,
286 van Deurzen, M.P.J., van Rantwijk, F., A., and Alcalde, M. (2008) Microb. Cell
and Sheldon, R.A. (1997) Tetrahedron, 53, Fact., 7, 32.
13183. 305 Morozova, O., Shumakovich, G.,
287 van de Velde, F., Bakker, M., Gorbacheva, M., Shleev, S., and
van Rantwijk, F., Rai, G.P., Hager, L.P., Yaropolov, A. (2007) Biochemistry
and Sheldon, R.A. (2001) J. Mol. Catal., B: (Moscow), 72, 1136.
Enzym., 11, 765. 306 Witayakran, S. and Ragauskas, A.J.
288 Rai, G.P., Sakai, S., Florez, A.M., (2009) Adv. Synth. Catal.,
Mogollon, L., and Hager, L.P. (2001) Adv. 351, 1187.
Synth. Catal., 343, 638. 307 Ciriminna, R. and Pagliaro, M. (2009)
289 Lee, K. and Moon, S.-H. (2003) Org. Process Res. Dev., 14, 245.
J. Biotechnol., 102, 261. 308 Arends, I.W.C.E., Li, Y.-X., and
290 Kohlmann, C. and L€ utz, S. (2006) Eng. Life Sheldon, R.A. (2006) Biocatal.
Sci., 6, 170. Biotransform., 24, 443.
291 Perez, D.I., Mifsud Grau, M., 309 Astolfi, P., Brandi, P., Galli, C.,
Arends, I.W.C.E., and Hollmann, F. Gentili, P., Gerini, M.F., Greci, L., and
(2009) Chem. Commun., 6848. Lanzalunga, O. (2005) New J. Chem.,
292 Karmee, S.K., Roosen, C., Kohlmann, C., 29, 1308.
L€utz, S., Greiner, L., and Leitner, W. 310 Cantarella, G., Galli, C., and Gentili, P.
(2009) Green Chem., 11, 1052. (2004) New J. Chem., 28, 366.
293 Pezzotti, F., Okrasa, K., and 311 Galli, C. and Gentili, P. (2003) 9th
Therisod, M. (2005) Tetrahedron: Symposium on Organic Reactivity, John
Asymmetry, 16, 2681. Wiley & Sons Ltd., Oslo, p. 973.
1432
312 Baiocco, P., Barreca, A.M., Fabbrini, M., 330 Molinari, F., Aragozzini, F., Cabral,
Galli, C., and Gentili, P. (2003) Org. J.M.S., and Prazeres, D.M.F. (1997)
Biomol. Chem., 1, 191. Enzyme Microb. Technol., 20, 604.
313 dAcunzo, F., Baiocco, P., Fabbrini, M., 331 Molinari, F., Gandolfi, R., Aragozzini, F.,
Galli, C., and Gentili, P. (2002) Leon, R., and Prazeres, D.M.F. (1999)
Eur. J. Org. Chem., 4195. Enzyme Microb. Technol., 25, 729.
314 Arends, I.W.C.E., Li, Y.-X., Ausan, R., and 332 Çelik, D., Bayraktar, E., and
Sheldon, R.A. (2006) Tetrahedron, 62, Mehmetoglu, U. € (2004) Biochem. Eng. J.,
6659. 17, 5.
315 Bragd, P.L., van Bekkum, H., and 333 Isobe, K. and Nishise, H. (1999) J.
Besemer, A.C. (2004) Top. Catal., Biotechnol., 75, 265.
27, 49. 334 Siebum, A., van Wijk, A., Schoevaart, R.,
316 Kulys, J. and Vidziunaite, R. (2005) and Kieboom, T. (2006) J. Mol. Catal., B:
J. Mol. Catal., B: Enzym., 37, 79. Enzym., 41, 141.
317 Matijosyte, I., Arends, I.W.C.E., 335 Franke, D., Machajewski, T., Hsu, C.-C.,
de Vries, S., and Sheldon, R.A. (2010) and Wong, C.-H. (2003) J. Org. Chem., 68,
J. Mol. Catal., B: Enzym., 62, 142. 6828.
318 Barreca, A.M., Fabbrini, M., Galli, C., 336 Gandolfi, R., Cavenago, K.,
Gentili, P., and Ljunggren, S. (2003) Gualandris, R., Sinisterra Gago, J.V., and
J. Mol. Catal., B: Enzym., 26, 105. Molinari, F. (2004) Process Biochem.,
319 Baratto, L., Candido, A., Marzorati, M., 39, 749.
Sagui, F., Riva, S., and Danieli, B. (2006) 337 Geerlof, A., Stoorvogel, J., Jongejan, J.A.,
J. Mol. Catal., B: Enzym., 39, 3. Leenen, E.J.T.M., Dooren, T.J.G.M.,
320 Potthast, A., Rosenau, T., Chen, C.L., and Tweel, W.J.J., and Duine, J.A. (1994) Appl.
Gratzl, J.S. (1996) J. Mol. Catal., B: Microbiol. Biotechnol., 42, 8.
Enzym., 108, 5. 338 Geerlof, A., Jongejan, J.A.,
321 Fabbrini, M., Galli, C., Gentili, P., and van Dooren, T.J.G.M., Raemakers-
Macchitella, D. (2001) Tetrahedron Lett., Franken, P.C., van den Tweel, W.J.J., and
42, 7551. Duine, J.A. (1994) Enzyme Microb.
322 dAcunzo, F., Baiocco, P., and Galli, C. Technol., 16, 1059.
(2003) New J. Chem., 27, 329. 339 Wandel, U., Machado, S.S., Jongejan,
323 Marzorati, M., Danieli, B., J.A., and Duine, J.A. (2001) Enzyme
Haltrich, D., and Riva, S. (2005) Green Microb. Technol., 28, 233.
Chem., 7, 310. 340 Romano, A., Gandolfi, R., Nitti, P.,
324 B€uhler, B., Schmid, A., Hauer, B., and Rollini, M., and Molinari, F. (2002)
Witholt, B. (2000) J. Biol. Chem., 275, J. Mol. Catal., B: Enzym., 17, 235.
10085. 341 Gandolfi, R., Borrometi, A., Romano, A.,
325 B€uhler, B., Witholt, B., Hauer, B., and Sinisterra Gago, J.V., and Molinari, F.
Schmid, A. (2002) Appl. Environ. (2002) Tetrahedron: Asymmetry,
Microbiol., 68, 560. 13, 2345.
326 B€uhler, B., Bollhalder, I., 342 Leon, R., Prazeres, D.M.F.,
Hauer, B., Witholt, B., and Molinari, F., and Cabral, J.M.S.
Schmid, A. (2003) Biotechnol. Bioeng., (2002) Biocatal. Biotransform.,
81, 683. 20, 201.
327 B€uhler, B., Bollhalder, I., Hauer, B., 343 Molinari, F., Gandolfi, R., Villa, R.,
Witholt, B., and Schmid, A. (2003) Urban, E., and Kiener, A. (2003)
Biotechnol. Bioeng., 82, 833. Tetrahedron: Asymmetry, 14, 2041.
328 Gandolfi, R., Ferrara, N., and 344 Molinari, F., Villa, R.,
Molinari, F. (2001) Tetrahedron Lett., Aragozzini, F., Cabella, P., Barbeni, M.,
42, 513. and Squarcia, F. (1997) J. Chem. Technol.
329 Villa, R., Romano, A., Gandolfi, R., Biotechnol., 70, 294.
Sinisterra Gago, J.V., and Molinari, F. 345 Svitel, J. and Sturdık, E. (1995) Enzyme
(2002) Tetrahedron Lett., 43, 6059. Microb. Technol., 17, 546.
References j1433
346 Geerlof, A., Vantol, J.B.A., Jongejan, J.A., Colombi, N. (1998) Tetrahedron:
and Duine, J.A. (1994) Biosci. Biotechnol. Asymmetry, 9, 2317.
Biochem., 58, 1028. 363 Ramirez, M.A., Perez, H.I., Manjarrez,
347 Su, W., Chang, Z., Gao, K., and N., Solis, A., Luna, H., and Cassani, J.
Wei, D. (2004) Tetrahedron: Asymmetry, (2008) Electron. J. Biotechnol., 11, 7.
15, 1275. 364 Stigter, E.C.A., van der Lugt, J.P., and
348 Perez, H.I., Manjarrez, N., Solis, A., Somers, W.A.C. (1997) J. Mol. Catal., B:
Luna, H., Ramirez, M.A., and Cassani, J. Enzym., 2, 291.
(2009) African J. Biotechnol., 8, 2279. 365 Stampfer, W., Kosjek, B., Faber, K., and
349 Miyamoto, K., Hirokawa, S., and Kroutil, W. (2003) Tetrahedron:
Ohta, H. (2007) J. Mol. Catal., B: Enzym., Asymmetry, 14, 275.
46, 14. 366 Matsuyama, A., Yamamoto, H.,
350 Nagaki, M., Imaruoka, H., Kawakami, J., Kawada, N., and Kobayashi, Y. (2001) J.
Saga, K., Kitahara, H., Sagami, H., Mol. Catal., B: Enzym., 11, 513.
Oba, R., Ohya, N., and Koyama, T. 367 Nikodinovic, J., Dinges, J.M.,
(2007) J. Mol. Catal., B: Enzym., Bergmeier, S.C., McMills, M.C.,
47, 33. Wright, D.L., and Priestley, N.D. (2006)
351 Geerlof, A., Stoorvogel, J., Jongejan, J.A., Org. Lett., 8, 443.
Leenen, E.J.T.M., Dooren, T.J.G.M., 368 de Carvalho, C.C.C.R. and da Fonseca,
Tweel, W.J.J., and Duine, J.A. (1994) M.M.R. (2002) J. Mol. Catal., B: Enzym.,
Appl. Microbiol. Biotechnol., 42, 8. 19–20, 389.
352 Hirano, J.-I., Miyamoto, K., and Ohta, H. 369 de Carvalho, C.C.C.R. and da Fonseca,
(2005) J. Biosci. Bioeng., 100, 318. M.M.R. (2002) J. Mol. Catal., B: Enzym.,
353 Presecki, A.V. and Vasic-Racki, -D- . (2009) 19–20, 377.
Process Biochem., 44, 54. 370 Tecel~ao, C.S.R., van Keulen, F., and da
354 Matos, J.R. and Wong, C.H. (1986) J. Org. Fonseca, M.M.R. (2001) J. Mol. Catal., B:
Chem., 51, 2388. Enzym., 11, 719.
355 Irwin, A.J. and Jones, J.B. (1977) J. Am. 371 Nestl, B., Voss, C., Bodlenner, A.,
Chem. Soc., 99, 556. Ellmer-Schaumberger, U., Kroutil, W.,
356 Patel, R.N., Liu, M., Banerjee, A., and Faber, K. (2007) Appl. Microbiol.
Thottathil, J.K., Kloss, J., and Biotechnol., 76, 1001.
Szarka, L.J. (1992) Enzyme Microb. 372 Glueck, S.M., Larissegger-Schnell, B.,
Technol., 14, 778. Csar, K., Kroutil, W., and Faber, K. (2005)
357 Bridges, A.J., Raman, P.S., Ng, G.S.Y., Chem. Commun., 1904.
and Jones, J.B. (1984) J. Am. Chem. Soc., 373 Glueck, S.M., Pirker, M., Nestl, B.M.,
106, 1461. Ueberbacher, B.T.,
358 Liu, T.E., Wolf, B., Geigert, J., Larissegger-Schnell, B., Csar, K.,
Neidleman, S.L., Chin, J.D., and Hauer, B., Stuermer, R., Kroutil, W., and
Hirano, D.S. (1983) Carbohydr. Res., 113, Faber, K. (2005) J. Org. Chem.,
151. 70, 4028.
359 Geigert, J., Neidleman, S.L., and 374 Nestl, B.M., Glueck, S.M., Hall, M.,
Hirano, D.S. (1983) Carbohydr. Res., Kroutil, W., Stuermer, R., Hauer, B., and
113, 159. Faber, K. (2006) Eur. J. Org. Chem.,
360 Kagohara, E., Pellizari, V.H., 4573.
Comasseto, J.V., Andrade, L.H., and 375 Nestl, B.M., Kroutil, W., and Faber, K.
Porto, A.L.M. (2008) Food Technol. (2006) Adv. Synth. Catal., 348, 873.
Biotechnol., 46, 381. 376 Gruber, C.C., Nestl, B.M., Gross, J.,
361 Andrade, L.H., Utsunomiya, R.S., Hildebrandt, P., Bornscheuer, U.T.,
Omori, A.T., Porto, A.L.M., and Faber, K., and Kroutil, W. (2007) Chem.
Comasseto, J.V. (2006) J. Mol. Catal., B: Eur. J., 13, 8271.
Enzym., 38, 84. 377 Edegger, K., Mang, H., Faber, K., Gross, J.,
362 Fogagnolo, M., Giovannini, P.P., and Kroutil, W. (2006) J. Mol. Catal., B:
Guerrini, A., Medici, A., Pedrini, P., and Enzym., 251, 66.
1434
378 Strauss, U.T., Felfer, U., and Faber, K. 397 Shimizu, S., Hattori, S., Hata, H., and
(1999) Tetrahedron: Asymmetry, Yamada, H. (1987) Enzyme Microb.
10, 107. Technol., 9, 411.
379 Martin-Matute, B. and Backvall, J.E. 398 Baskar, B., Pandian, N.G., Priya, K., and
(2007) Curr. Opin. Chem. Biol., 11, 226. Chadha, A. (2005) Tetrahedron,
380 Allan, G.R. and Carnell, A.J. (2001) J. Org. 61, 12296.
Chem., 66, 6495. 399 Oikawa, T., Mukoyama, S., and
381 Voss, C.V., Gruber, C.C., and Kroutil, W. Soda, K. (2001) Biotechnol. Bioeng.,
(2008) Angew Chem. Int. Ed., 47, 741. 73, 80.
382 Voss, C.V., Gruber, C.C., Faber, K., 400 Takahashi, E., Nakamichi, K., and
Knaus, T., Macheroux, P., and Kroutil, W. Furui, M. (1995) J. Ferment Bioeng.,
(2008) J. Am. Chem. Soc., 130, 13969. 80, 247.
383 Voss, C.V., Gruber, C.C., and Kroutil, W. 401 Tsuchiya, S., Miyamoto, K., and Ohta, H.
(2007) Tetrahedron: Asymmetry, 18, 276. (1992) Biotechnol. Lett., 14, 1137.
384 Fantin, G., Fogagnolo, M., 402 Padhi, S.K. and Chadha, A. (2005)
Giovannini, P.P., Medici, A., and Tetrahedron: Asymmetry, 16, 2790.
Pedrini, P. (1995) Tetrahedron: Asymmetry, 403 Padhi, S.K., Pandian, N.G., and
6, 3047. Chadha, A. (2004) J. Mol. Catal., B:
385 Cardus, G.J., Carnell, A.J., Enzym., 29, 25.
Trauthwein, H., and Riermeir, T. (2004) 404 Padhi, S.K., Titu, D., Pandian, N.G., and
Tetrahedron: Asymmetry, 15, 239. Chadha, A. (2006) Tetrahedron,
386 Comasseto, J.V., Andrade, L.H., 62, 5133.

Omori, A.T., Assis, L.F., and Porto, 405 Titu, D. and Chadha, A. (2008)
A.L.M. (2004) J. Mol. Catal., B: Enzym., Tetrahedron: Asymmetry, 19, 1698.
29, 55. 406 Titu, D. and Chadha, A. (2008) J. Mol.
387 Takemoto, M. and Achiwa, K. (1995) Catal., B: Enzym., 52–53, 168.
Tetrahedron: Asymmetry, 6, 2925. 407 Thangavel, V. and Chadha, A. (2007)
388 Demir, A.S., Hamamci, H., Tetrahedron, 63, 4126.
Sesenoglu, O., Neslihanoglu, R., 408 Tanaka, T., Iwai, N., Matsuda, T., and
Asikoglu, B., and Capanoglu, D. (2002) Kitazume, T. (2009) J. Mol. Catal., B:
Tetrahedron Lett., 43, 6447. Enzym., 57, 317.
389 Nie, Y., Xu, Y., Lv, T.F., and Xiao, R. (2009) 409 Gruber, C.C., Lavandera, I., Faber, K., and
J. Chem. Technol. Biotechnol., Kroutil, W. (2006) Adv. Synth. Catal., 348,
84, 468. 1789.
390 Nie, Y., Xu, Y., and Mu, X.Q. (2004) Org. 410 Eliel, E.L., Wilen, S.H., and
Process Res. Dev., 8, 246. Doyle, M.P. (2001) Basic Organic
391 Hummel, W. and Riebel, B. (1996) Ann. Stereochemistry, Wiley Interscience,
N. Y. Acad. Sci., 799, 713. New York.
392 Xie, S.X., Ogawa, J., and Shimizu, S. 411 Hummel, W. and Riebel, B. (1996) Chiral
(1999) Appl. Microbiol. Biotechnol., 52, alcohols by enantioselective enzymatic
327. oxidation, in Enzyme Engineering Xiii,
393 Goswami, A., Mirfakhrae, K.D., and vol. 799 (eds J.S. Dordick and A.J.
Patel, R.N. (1999) Tetrahedron: Asymmetry, Russell), New York Academy of Sciences,
10, 4239. New York, p. 713.
394 Hasegawa, J., Ogura, M., 412 Swamy, K.C.K., Kumar, N.N.B.,
Tsuda, S., Maemoto, S.-I., Kutsuki, H., Balaraman, E., and Kumar, K. (2009)
and Ohashi, T., (1990) Agric Biol. Chem., Chem. Rev., 109, 2551.
54, 1819. 413 Nakamura, K., Fujii, M., and Ida, Y.
395 Chadha, A. and Baskar, B. (2002) (2001) Tetrahedron: Asymmetry, 12, 3147.
Tetrahedron: Asymmetry, 13, 1461. 414 Utsukihara, T., Misumi, O., Nakajima, K.,
396 Shimizu, S., Hattori, S., Hata, H., and Koshimura, M., Kuniyoshi, M.,
Yamada, H. (1987) Appl. Environ. Kuroiwa, T., and Horiuchi, C.A. (2008)
Microbiol., 53, 519. J. Mol. Catal., B: Enzym., 51, 19.
References j1435
415 Mantovani, S.M., Angolini, C.F.F., and Perozich, J., Lindahl, R., Hempel, J., and
Marsaioli, A.J. (2009) Tetrahedron: Wang, B.C. (1997) Nat. Struct. Biol., 4,
Asymmetry, 20, 2635. 317.
416 Bhat, P.V., Poissant, L., and Wang, X.L. 435 Moore, S.A., Baker, H.M., Blythe, T.J.,
(1996) Biochem. Cell Biol., 74, 695. Kitson, K.E., Kitson, T.M., and Baker,
417 Roberts, E.S., Vaz, A.D.N., and Coon, M.J. E.N. (1998) Struct. Fold. Des., 6, 1541.
(1991) Proc. Natl. Acad. Sci. USA, 88, 436 Nunez, A., Foglia, T.A., and Piazza, G.J.
8963. (1999) Biotechnol. Appl. Biochem., 29, 207.
418 Vaz, A.D.N., Kessell, K.J., and Coon, M.J. 437 MacKintosh, R.W. and Fewson, C.A.
(1994) Biochemistry, 33, 13651. (1988) Biochem. J., 255, 653.
419 Vaz, A.D.N., Roberts, E.S., and Coon, M.J. 438 MacKintosh, R.W. and Fewson, C.A.
(1991) J. Am. Chem. Soc., 113, 5886. (1988) Biochem. J., 250, 743.
420 Eklund, H., Plapp, B.V., Samama, J.P., 439 Buhler, B. and Schmid, A. (2004)
and Branden, C.I. (1982) J. Biol. Chem., J. Biotechnol., 113, 183.
257, 14349. 440 Leak, D.J., Sheldon, R.A., Woodley, J.M.,
421 Olson, L.P., Luo, J., Almarsson, O., and and Adlercreutz, P. (2009) Biocatal.
Bruice, T.C. (1996) Biochemistry, Biotransform., 27, 1.
35, 9782. 441 Walsh, C.T. and Chen, Y.C.J. (1988)
422 Velonia, K. and Smonou, I. (2000) Angew Chem. Int. Ed. Engl., 27, 333.
J. Chem. Soc., Perkin Trans. 1, 2283. 442 Hastings, J.W., Riley, W.H., and
423 Abeles, R.H. and Lee, H.A. (1960) Massa, J. (1965) J. Biol. Chem.,
J. Biol. Chem., 235, 1499. 240, 1473.
424 Dalziel, K. and Dickinson, F.M. (1965) 443 Shimomur, O., Kohama, Y., and
Nature, 206, 255. Johnson, F.H. (1972) Proc. Natl. Acad. Sci.
425 Henehan, G.T.M. and Oppenheimer, N.J. USA, 69, 2086.
(1993) Biochemistry, 32, 735. 444 Inouye, S. and Sasaki, S. (2007) Protein
426 Hinson, J.A. and Neal, R.A. (1972) Expression Purif., 56, 261.
J. Biol. Chem., 247, 7106. 445 Campbell, Z.T., Weichsel, A.,
427 Heinstra, P.W.H., Eisses, K.T., Montfort, W.R., and Baldwin, T.O. (2009)
Schoonen, W.G.E.J., Aben, W., Biochemistry, 48, 6085.
de Winter, A.J., van der Horst, D.J., 446 Baldwin, T.O. and Ziegler, M.M. (1992)
van Marrewijk, W.J.A., Beenakkers, Bioluminescence and
A.M.T., Scharloo, W., and Th€orig, G.E.W. chemiluminescence, in Chemistry
(1983) Genetica, 60, 129. and Biochemistry of Flavoenzymes
428 Lamed, R.J. and Zeikus, J.G. (1981) (ed. F. Muller), CRC Press, Boca Raton,
Biochem. J., 195, 183. Fl, p. 467.
429 Trivic, S., Leskova, V., and Winston, G.W. 447 Bernhardt, P. V. (2006) Aust. J. Chem., 59,
(1999) Biotechnol. Lett., 21, 231. 133.
430 Mee, B., Kelleher, D., Frias, J., 448 Julsing, M.K., Cornelissen, S.,
Malone, R., Tipton, K.F., Henehan, Buhler, B., and Schmid, A. (2008) Curr.
G.T.M., and Windle, H.J. (2005) FEBS J., Opin. Chem. Biol., 12, 177.
272, 1255. 449 Mayhew, M.P., Reipa, V., Holden, M.J.,
431 Ahvazi, B., Coulombe, R., Delarge, M., and Vilker, V.L. (2000) Biotechnol. Prog.,
Vedadi, M., Zhang, L., Meighen, E., and 16, 610.
Vrielink, A. (2000) Biochem. J., 349, 853. 450 Vilker, V., Reipa, V., Mayhew, M., and
432 Gonzalez-Segura, L., Rudino-Pinera, E., Holden, M. (1999) J. Am. Oil. Chem. Soc.,
Munoz-Clares, R.A., and Horjales, E. 76, 1283.
(2009) J. Mol. Biol., 385, 542. 451 Reipa, V., Mayhew, M.P., and
433 Tsybovsky, Y., Donato, H., Krupenko, N.I., Vilker, V.L. (1997) Proc. Natl. Acad. Sci.
Davies, C., and Krupenko, S.A. (2007) USA, 94, 13554.
Biochemistry, 46, 2917. 452 Schwaneberg, U., Appel, D.,
434 Liu, Z.J., Sun, Y.J., Rose, J., Chung, Y.J., Schmitt, J., and Schmid, R.D. (2000)
Hsiao, C.D., Chang, W.R., Kuo, I., J. Biotechnol., 84, 249.
1436
453 Estabrook, R.W., Faulkner, K.M., 470 Taniai, H., Iida, K., Seki, M., Saito, M.,
Shet, M., and Fisher, C.W. (1996) Shiota, S., Nakayama, H., and Yoshida, S.
Methods in Enzymology, vol. 272, (2008) J. Bacteriol., 190, 3572.
Academic Press, San Diego, p. 44. 471 Sedewitz, B., Schleifer, K.H., and Gotz, F.
454 Faulkner, K.M., Shet, M.S., (1984) J. Bacteriol., 160, 462.
Fisher, C.W., and Estabrook, R.W. 472 Sedewitz, B., Schleifer, K.H., and Gotz, F.
(1995) Proc. Natl. Acad. Sci. USA, 92, (1984) J. Bacteriol., 160, 273.
7705. 473 Carlsson, J. and Kujala, U. (1985) FEMS
455 Terelius, Y., Norstenhoog, C., Microbiol. Lett., 25, 53.
Cronholm, T., and Ingelmansundberg, 474 Vigenschow, H., Schwarm, H.M., and
M. (1991) Biochem. Biophys. Res. Knobloch, K. (1986) Biol. Chem. Hoppe
Commun., 179, 689. Seyler, 367, 951.
456 Watanabe, K., Matsunaga, T., 475 Skarstedt, M.T. and Silverstein, E. (1976)
Narimatsu, S., Yamamoto, I., and J. Biol. Chem., 251, 6775.
Yoshimura, H. (1992) Biochem. Biophys. 476 Spellerberg, B., Cundell, D.R.,
Res. Commun., 188, 114. Sandros, J., Pearce, B.J.,
457 Tomita, S., Tsujita, M., Matsuo, Y., IdanpaanHeikkila, I., Rosenow, C., and
Yubisui, T., and Ichikawa, Y. (1993) Int. J. Masure, H.R. (1996) Mol. Microbiol., 19,
Biochem., 25, 1775. 803.
458 Watanabe, K., Matsunaga, T., 477 Carlsson, J., Kujala, U., and
Yamamoto, I., and Yashimura, H. (1995) Edlund, M.B.K. (1985) Infect. Immun., 49,
Drug Metab. Dispos., 23, 261. 674.
459 Watanabe, K., Narimatsu, S., 478 Kim, D.M. and Swartz, J.R. (1999)
Matsunaga, T., Yamamoto, I., and Biotechnol. Bioeng., 66, 180.
Yoshimura, H. (1993) Biochem. 479 Kim, H.C. and Kim, D.M. (2009)
Pharmacol., 46, 405. J. Biosci. Bioeng., 108, 1.
460 Stearns, R.A., Chakravarty, P.K., 480 Schwarz, D., Klammt, C., Koglin, A.,
Chen, R., and Chiu, S.H.L. (1995) Drug Lohr, F., Schneider, B., Dotsch, V., and
Metab. Dispos., 23, 207. Bernhard, F. (2007) Methods,
461 Hille, R. and Nishino, T. (1995) FASEB J., 41, 355.
9, 995. 481 Hager, L.P. and Lipmann, F. (1961) Proc.
462 Booth, V.H. (1938) Biochem. J., 32, 494. Natl. Acad. Sci. USA, 47, 1768.
463 Dastoli, F.R. and Price, S. (1967) Arch 482 Abdel-Hamid, A.M., Attwood, M.M., and
Biochem. Biophys., 118, 163. Guest, J.R. (2001) Microbiology-Sgm, 147,
464 Morpeth, F.F. (1983) Biochim. Biophys. 1483.
Acta, 744, 328. 483 Popov, V.O. and Lamzin, V.S. (1994)
465 Liese, A., Seelbach, K., and Wandrey, C. Biochem. J., 301, 625.
(2006) Industrial Biotransformations, 484 Wichmann, R., Wandrey, C.,
Wiley-VCH Verlag GmbH, Weinheim. Buckmann, A.F., and Kula, M.R. (1981)
466 Wong, J.W., Watson, H.A., Bouressa, J.F., Biotechnol. Bioeng., 23, 2789.
Burns, M.P., Cawley, J.J., Doro, A.E., 485 Zhao, H. and van der Donk, W.A. (2003)
Guzek, D.B., Hintz, M.A., Curr. Opin. Biotechnol., 14, 583.
McCormick, E.L., Scully, D.A., 486 Bommarius, A.S., Schwarm, M., and
Siderewicz, J.M., Taylor, W.J., Truesdell, Drauz, K. (1998) J. Mol. Catal., B: Enzym.,
S.J., and Wax, R.G. (2002) Org. Process Res. 5, 1.
Dev., 6, 477. 487 Seelbach, K., Riebel, B.,
467 Tanaka, M. and Hirokane, Y. (2000) Hummel, W., Kula, M.-R., Tishkov, V.I.,
J. Biosci. Bioeng., 90, 341. Egorov, A.M., Wandrey, C., and
468 Woods, D.D. (1936) Biochem. J., Kragl, U. (1996) Tetrahedron Lett.,
30, 515. 37, 1377.
469 Krebs, H.A., Kornberg, H.L., and 488 Slusarczyk, H., Felber, S., Kula, M.R., and
Burton, K. (1957) Erg. Physiol. Biol. Ch., Pohl, M. (2000) Eur. J. Biochem., 267,
49, 212. 1280.
References j1437
489 Simmonds, J. and Robinson, G.K. Appl. Microbiol. Biotechnol.,
(1998) Appl. Microbiol. Biotechnol., 28, 20.
50, 353. 493 Yoshikawa, N., Ohta, K., Mizuno, S., and
490 Simmonds, J. and Robinson, G.K. (1997) Ohkishi, H. (1993) Bioprocess Technol.,
Enzyme Microb. Technol., 21, 367. 16, 131.
491 Yoshikawa, N., Mizuno, S., Ohta, K., and 494 Wubbolts, M.G. and Timmis, K.N.
Suzuki, M. (1990) J. Biotechnol., (1990) Appl. Environ. Microbiol.,
14, 203. 56, 569.
492 Mizuno, S., Yoshikawa, N., Seki, M., 495 Haddad, S., Eby, D.M., and Neidle, E.L.
Mikawa, T., and Imada, Y. (1988) (2001) Appl. Environ. Microbiol., 67, 2507.
j1439
33
Baeyer–Villiger Oxidations
Marko D. Mihovilovic
33.1
Introduction
The oxidative transformation of a ketone into either a lactone or ester functionality

was first discovered at the end of the nineteenth century by Adolf von Baeyer and
Victor Villiger [1]. The chemical reaction became particularly valuable for synthetic
applications, as the rearrangement process turned out to be highly predictable:
(i) oxygen incorporation commences with strict retention of configuration at the
migrating carbon center; (ii) the effects governing the regiochemistry of this reaction
are based on electron density at the migrating atom as well as a certain stereoelec-
tronic arrangement and have been well understood [2, 3]. Within current organic
chemistry, peroxyacids are largely applied as reagents for this oxygenation and also
asymmetric variants have been developed [4].
First indications of enzymatic Baeyer–Villiger oxygenations were reported in 1948
within the biocatalytic degradation of steroids [5]; subsequent studies proved this
biotransformation to be abundant in fungi and prokaryotic organisms [6, 7].
However, Baeyer–Villiger-type conversions are also encountered in the biosynthesis
of several natural products such as aflatoxins in fungi [8], shellfish toxins [9], and
iridoids in plants [10]. The enzymatic reaction was found to depend on oxygen as well
as on a nicotinamide reduction equivalent [11]. In the late 1960s Baeyer–Villiger
monooxygenases (BVMOs; EC 1.14.13.x) were first isolated and characterized as
responsible catalytic entities [12, 13].
In these biocatalysts flavin cofactors serve as natures analog to peracids used in
the chemical transformation. To activate the flavin prosthetic group for the reaction
with molecular oxygen, nicotinamide (NADH or NADPH) is required as electron
donor. Notably, most reported BVMOs are soluble proteins, in contrast to other
types of monooxygenases, which tend to be membrane associated. Based on the
flavin cofactor, a classification of BVMOs into two types was proposed early on [14]:
Type 1 BVMOs consist of a single polypeptide chain, contain FAD as cofactor, and
are NADPH dependent; the two cofactors are bound at separate dinucleotide
binding domains, as indicated by the presence of two Rossmann sequence motifs
and Oliver May.
j 33 Baeyer–Villiger Oxidations
1440
(GxGxxG). Type 2 BVMOs contain FMN and are dependent on NADH; they are
composed of two different subunits and display some relationship with flavin-
dependent luciferases.
Type 1 BVMOs are sequence related and also belong to subclass B flavoprotein
monooxygenases based on structural similarity [15]. A fingerprint sequence motif
(FxGxxxHxxxW) at the linker of the FAD- and the NADPH-binding domains is highly
conserved within this BVMO enzyme family and can serve as recognition probe, in
particular for annotation purposes [16]. This approach allowed the discovery and
characterization of several new BVMOs and was successfully complemented by
the methodology of polymerase chain reaction in combination with highly degen-
erate primers [17]. Initially, studies using these enzymes were conducted using native
organisms as whole-cell systems or protein isolates thereof. On a historic note, the
first recombinant expression systems utilized for preparative-scale biotransforma-
tions were based on Saccharomyces cerevisiae (designer yeasts) [18]. Nowadays, a
large collection of enzymes is available, preferably from Escherichia coli as recom-
binant host for fermentations or protein production. Members of type 1 BVMOs
represent the majority of biocatalysts studied within recent years and several previous
reviews and book chapters provide additional summaries [19, 20].
Notably, the distribution of BVMOs in nature is quite particular, as they are
abundantly encountered in prokaryotes and fungi but they seem to be absent in
higher eukaryotes and archaea. Many of these enzymes play a major role in catabolic
pathways enabling organisms to grow on alternative hydrocarbon sources for cell
energy as evolutionary advantage [21–23]. Historically, wastewater facilities of
chemical industries used to be a particularly rich source of organisms expressing
BVMOs. These species encounter an abundant supply of alkane precursors for
oxygenation reactions and elective growth techniques were applied to isolate such
expression strains. The technique of mRNA differential display was successfully
applied to such environmental isolates to identify metabolic genes for previously
unknown microbial BVMOs [24]. Additionally, BVMOs are involved in the produc-
tion of secondary metabolites [8, 25, 26]. Consequently, several of the thoroughly
investigated biocatalysts are remarkably promiscuous and display a broad substrate
tolerance for structurally diverse compounds, which makes them appealing catalysts
for the application in synthesis.
33.2
Mechanism and Enzyme Structure
The mechanistic cycle for BVMOs has been investigated using a BVMO from
Acinetobacter (Figure 33.1) [27, 28]. The biotransformation is initiated by reduction
of FAD mediated by the nicotinamide cofactor (usually NADPH). This generates an
activated FAD-intermediate that can undergo direct reaction with molecular oxygen
to give an FAD-4a-peroxyanion; this species represents the enzymatic equivalent of
an organic peracid. The peroxyanion is also in equilibrium with the corresponding
hydroperoxide; this protonation becomes a critical reaction channel in the absence of
33.2 Mechanism and Enzyme Structure j1441
R
N N O
NH
N
H2O H2O
FAD O
NADPH
R R R
N N O N N O N N O
NH NH NH
N N N
hydroxy-FAD H OH reduced FAD H hydroxy-FAD H OH
O O O
O O
O O2 S
R R'
R R'
product S
R R'
R R R
N N O N N O N N O
NH NH NH
N N N
H O H O H O
O O O
O O peroxy anion O hydroperoxide HO
O
R R' R R'
Criegee adduct substrate electrophilic reaction
nucleophilic reaction
Figure 33.1 Mechanism and catalytic cycle of flavin-containing BVMOs.
a target for the nucleophilic attack. The presence of these two species is suggested to
account for the ambivalent reactivity of BVMOs in both electrophilic and nucleophilic
oxidation processes. The nucleophilic peroxyanion facilitates Baeyer–Villiger reac-
tions, while the electrophilic hydroperoxide is believed to be responsible for hetero-
atom oxidations. Kinetic studies could confirm that formation of the oxygenated FAD
species is independent of substrate binding [29].
Nucleophilic attack of the anion at a carbonyl group of the substrate leads to
formation of the tetrahedral Criegee intermediate; the analogous chemical entity is
formed during the conventional reaction mechanism of the classical Baeyer–Villiger
oxidation [30].
At this stage, migratory preference is determined. Two major effects play a role in
determining the regioselectivity of the rearrangement: Usually, the more nucleo-
philic carbon migrates (i.e. the more substituted center). However, certain stereo-
electronic and conformational arrangements can override the effects of electron
density. Migration can only occur for the CC bond adopting the antiperiplanar
conformation vis-a-vis the leaving group. Consequently, the spatial and electrostatic
composition of the enzymes active-site also affects migratory preference. The
biotransformation product (ester or lactone) is released from the active site and the
catalytic cycle is closed by elimination of water to regenerate the flavin cofactor.
1442
Figure 33.2 Crystallographic structure determination of PAMOThermo as a ribbon diagram: FAD-

binding domain, NADPH-binding domain, FAD in stick representation, and Arg337 and His173 in
sphere representation.
Significant progress has been made in understanding this enzyme family in

general based on protein structure determination. After several failed attempts, a first
three-dimensional structure of a BVMO was reported for an enzyme from the
moderately thermophilic bacterium Thermobifida fusca [31]. Phenylacetone mono-
oxygenase (PAMOThermo) converts preferably linear ketones bearing an aromatic ring
with, however, a quite limited substrate promiscuity. The enzyme displayed sufficient
stability in purified form to allow crystallographic structure determination [32].
Recently, a second structure for a BVMO from Rhodococcus strain HI-31 with a
significantly larger substrate profile was reported [33] and suggested that conclusions
derived from PAMOThermo may represent quite general features for the whole type 1
BVMO family.
BVMOs exhibit a two-domain architecture containing an FAD- as well as an
NADPH-binding domain (Figure 33.2). This organization concept is typical for
flavoproteins; however, certain helical regions appear to be typical for BVMOs. The
active site is located in a cleft at the domain interface. Both domains are connected via
a linker carrying a highly conserved sequence motif and this fingerprint sequence
(FxGxxxHxxxW) is highly indicative for type 1 BVMOs [16]. Amino acid Arg337 (in
PAMOThermo) at the Re side of FAD is strictly conserved throughout the whole
enzyme group and seems to play a critical role in catalysis; it is the current hypothesis
that the highly polar residue is involved in stabilizing peroxygenated intermediates
within the active site.
All structures available so far have been reported with the FAD in the resting state;
no information on the actual situation after oxygenation of the cofactor is available.
33.3 Cofactor Recycling and Preparative Operations j1443
However, based on the enzyme mechanism and the arrangement of domains in the
resting state, substantial conformational changes can be expected during the catalytic
cycle. A certain flexibility of the domains seems mandatory to allow the different
processes to be conducted within the whole catalytic cycle. This is corroborated by the
fact that two conformations for Arg337 were found in the protein crystal of
PAMOThermo. Additionally, the nicotinamide cofactor contributes to the structural
integrity of the active site, as the stereoselectivity of biotransformations significantly
decreases when using alternative recycling systems in the absence of NADPH [34]
(see Section 33.3). A critical role in these conformational modifications seems to be
attributed to the highly conserved linker region between the two domains, which is
envisioned to act as hinge. Here, His173 (PAMOThermo) – as part of the fingerprint
region – adopts a fully solvent exposed position and, again, is highly conserved in
BVMOs.
Still, a better understanding of protein dynamics throughout the complete oxygen
transfer process is required to provide a highly predictive model for substrate
acceptance, regio-, and stereoselectivity of BVMOs.
33.3
Cofactor Recycling and Preparative Operations
Widespread application of BVMOs in chiral synthesis and production processes has

been hampered due to several reasons [35]: With the exception of PAMOThermo the
enzyme group displays limited operational stability towards temperature and solvent
additives. In some cases, low substrate and product inhibition levels have been
observed. An additional obstacle certainly is represented by the cofactor dependence,
as this requires the utilization of auxiliary agents for the recycling of NAD(P)H.
Application of some of the standard alcohol/ketone auxiliary substrates as recycling
agents is limited due to competing reaction within the Baeyer–Villiger process,
making the overall process more complicated.
A straightforward approach to utilize BVMOs on a preparative scale is to conduct
the biotransformation with purified enzymes. Availability of isolated BVMOs has
been significantly improved by utilizing E. coli based overexpression systems in
recent years. Protein isolation has been scaled-up to 300-l batches for the BVMO from
Acinetobacter, providing an enzyme titer of >8 105 U [36]. BVMO purification was
simplified by employing affinity chromatography incorporating a His6-tag at the
C-terminus [37].
The traditional approach to regenerate the reduction equivalents required for the
initial FAD activation step prior to reaction with molecular oxygen is based on
coupling the process with a second biotransformation reconstituting NADPH at the
expense of a sacrificial cosubstrate (Figure 33.3). Standard procedures include the
application of formate dehydrogenase (FDH) [38], glucose-6-phosphate dehydroge-
nase (G6PDH) [39], or the dehydrogenase from Thermoanaerobium brockii [40]. This
latter system was refined by using PEG-NADPH in a membrane reactor [41]. Cost
calculations of such systems utilizing cheap auxiliary alcohols such as 2-propanol
1444
O2
auxiliary NADP+ FAD-H2 R R'

substrate
dehydrogenase BVMO
additional O
NADPH FAD
product
O
R R'
H2O
Figure 33.3 General concept of a two-enzyme cofactor recycling system. In selected cases, the
redox product of the auxiliary substrate can serve as precursor for the Baeyer–Villiger
biotransformation.
suggest a certain economic potential in industrial-scale fermentations [42] and a

proof of concept at laboratory scale gave promising results [43]. In certain cases, this
system can be developed into a closed-loop if the product of the dehydrogenase-
mediated oxygenation can serve as substrate for the BVMO [44].
An option to overcome the limited stability of BVMOs and facilitate the down-
stream processing is the immobilization of the biocatalysts. However, few examples
have been reported for this enzyme group. The BVMO from Acinetobacter was
successfully co-immobilized with alcohol dehydrogenases on EupergitÒ (oxirane
acrylic beads) [45] with only a 20% loss of activity compared to free enzyme; this
system could be re-used for more than two weeks of repeated batch biotransforma-
tions. The same BVMO was also immobilized with a polyethylenimine-porous
agarose polymeric composite and enforced stability tests were conducted by
c-radiation; this protocol retained close to 90% of the initial activity and the catalyst
could be submitted to 16 production cycles [46].
Some radically different methods for cofactor regeneration have been investigated
on BVMOs more recently. A highly interesting variant of the two-enzyme recycling
concept has been exemplified by designing self-sufficient BVMOs via protein fusion.
Fusing of proteins has become a widely applied strategy in enzyme purification
protocols (e.g., GST- or His-tag) as well as in subcellular visualization of target
proteins (e.g., GFP-tag [47]). However, the concept has been largely neglected in the
context of synthetic applications for the combination of otherwise separate enzyme
groups (in contrast to the well-known assembly lines of polyketide syntheses).
Upon design of a self-sufficient BVMO/CRE (cofactor regeneration enzyme) phos-
phite dehydrogenase (PTDH) [48] was selected based on favorable thermodynamics
of the enzymatic oxidation of phosphite to phosphate as a nearly irreversible
process [49]. This auxiliary substrate offers the additional advantage of being
orthogonal to both substrate and product with respect to possible side-reactions.
Within a proof-of-concept study covalently connected fusion enzyme biocatalysts
Figure 33.4 Concept of self-sufficient fusion protein biocatalysts by covalent linkage of BVMOs
with PTDH to generate a bifunctional CRE-system.
between several BVMOs and PTDH were created (Figure 33.4) [50]. Remarkably, the
bifunctional catalytic entity displayed similar kinetic parameters to the separate
enzymes and both substrate specificity as well as stereoselectivity of the fusion
biocatalysts was hardly affected at all. While the first generation of BVMO/CRE
systems suffered from low stability of the PTDH domain, a second generation of
biocatalysts was introduced with improved thermal stability for the cofactor regen-
eration part [51]. While purified enzyme as well as whole-cell application gave
satisfactory results, fusion biocatalysts were preferably utilized as crude cell extract,
as this enzyme preparation usually contains sufficient quantities of NADPH from the
living host organism. This suggests a very intimate localization of the nicotinamide
cofactor close to the fusion enzyme shuttling between the two active sites.
Organometallic complexes have been identified as suitable reducing agents for
flavins and nicotinamides in particular within alternative cofactor regeneration
strategies [52]. [Cp Rh(bpy)(H2O)]2þ is a preferred choice in such bio-metal-organic
reaction cascades, as this complex can be easily reduced by formate (Figure 33.5). In
an attempt to adapt this strategy to BVMOs, a first approach was investigated by
completely substituting NADPH regeneration by the Rh-system. While conversion
O2
substrate
CO2 [Cp*Rh(bpy)H]+ NADP+ FAD-H2
BVMO
HCOOH [Cp*Rh(bpy)(H2O)]2+ NADPH FAD

product
H2 O
Figure 33.5 General concept of the Rh-mediated FAD regeneration; trace amounts of
nicotinamide cofactor are required to allow the correct composition of the active site in the BVMO.
1446
O2
O
solution enzyme
R
EDTA FADox FAD-H2
hν PAMO
mutant
O O
EDTA FADred FAD O R
decomposition R
products +
H2O
Figure 33.6 Concept of light-driven cofactor compound, a reaction cascade is initiated

recycling: the biotransformation is including redox-reaction between the dissolved
supplemented by the sacrificial cosubstrate flavin analog and the enzyme bound FAD
EDTA and a flavin derivative; upon light- cofactor.
mediated reduction of the soluble flavin
was observed, the biocatalytic system displayed essentially loss of stereoselectivity. It

was demonstrated that the presence of trace amounts of nicotinamide cofactor is
required to reconstitute the stereospecificity of the enzyme [53]. This suggests a
pivotal role of NADPH in the stabilization of the active site, which is further
supported by the fact that NADPH stays bound to BVMOs essentially throughout
the complete catalytic cycle.
Another alternative to replace the auxiliary substrate for the cofactor recycling was
proposed by implementing a photochemical strategy. Certain organometallic com-
plexes as well as rather simple compounds such as flavin derivatives have been
identified as suitable materials to collect solar energy within a photoreduction process
in the presence of sacrificial electron donors like EDTA (ethylenediaminetetraacetic
acid). Consequently, the conventional cofactor recycling process via a second
biocatalyst can be replaced by simple supplementation of the biotransformation
mixture with a flavin analog (FAD, FMN, riboflavin) and EDTA as sacrificial agent
(Figure 33.6) [54]. As expected from the above experience using Rh-based cofactor
substitutes, trace amounts of nicotinamide cofactor are required to constitute the
catalytically active enzyme complex. The performance of this light-driven reaction is
comparable to electrochemically coupled systems (as demonstrated for another
biotransformation); however, it is at least one order of magnitude below the
conventional coupled-enzyme approach with respect to TOF and TON [55].
For some BVMOs moderate solvent tolerance was reported. Hydroxyacetophe-
none monooxygenaseHAPMO from Pseudomonas is still active in biphasic mixtures
containing up to 50% hexane, while a mutant of PAMOThermo can tolerate up to 30%
of methanol as additive [56]. In particular with the latter biocatalyst, a beneficial effect
of methanol on the stereospecificity of the enzyme was observed on several
substrates [57].
The majority of recent applications took advantage of (recombinant) whole-cell
expression systems. This approach is based on intact internal cofactor regeneration
within a living cell as long as it is metabolically active. In general, this mode of
application is easier to implement in synthetic organic laboratories and is particularly
useful in rapid screening campaigns [58]. However, whole-cell mediated biotrans-
formations may have certain drawbacks: BVMOs in general tend to have rather low
limits for substrate and product inhibition; within cell-mediated conversions the
toxicity of chemicals vis-a-vis the host organism also has to be considered. In addition,
substrates have to be able to penetrate cell wall and membrane barriers to reach the
active biocatalyst; products have to be secreted into the broth. It is remarkable that
successful biotransformations of more than 200 ketones predominantly using whole-
cell conversions have proven that these obstacles can be overcome in many routine
laboratory applications.
Various recombinant E. coli hosts have been successfully applied also to fermen-
tation up-scaling, facilitating enzyme expression by strong promoters and under
highly controlled conditions (arabinose, IPTG) [59, 60]. This organism offers the
advantages of easy cultivation, rapid growth, tight control, and high yield of protein
production. Both glucose and glycerol have been used as feedstocks for efficient
cofactor regeneration [61, 62]. However, productivity under standard fermentation
conditions is usually moderate with average space–time yields in the range of g l1
h1 and average substrate concentrations of around 1 g l1 [63]. Fermentations under
non-growing conditions usually gives better results [61], which can be attributed in
part to the fact that membrane transport of substrates is facilitated when the total
concentration of organic compounds is increased. Such a protocol applying non-
growing conditions was also demonstrated to give improved results with compounds
previously reported as poor or non-substrates for BVMOs [64].
Oxygenation certainly is a critical aspect for BVMO-mediated biotransforma-
tions [65]: As the oxygen requirements of the Baeyer–Villiger reaction competes
with the needs of cell metabolism in particular at high cell density, biocatalyst
concentration has to be adjusted based on the oxygenation ability of a given reactor.
Excellent results have been obtained by optimizing the performance of aeration
spargers; this technology has progressed to pilot-plant size (50 l) [66].
Frequently encountered substrate inhibition problems could be overcome by
continuous feeding, keeping the limiting concentration below the critical value;
this strategy was successfully applied on 55-l and 200-l pilot-plant scale, yielding
multi-100 g product quantities [67].
Biphasic fermentation protocols were implemented to control product inhibition,
as well. A particularly successful example for liquid–liquid systems was reported for
the production of lauryl lactone using a BVMO from Pseudomonas applying hex-
adecane. By extending a conventional batch process using a two-phase semi-con-
tinuous reactor (TPSCR) substrate concentrations of 840 mM could be achieved with
largely improved downstream processing [68]. In a related strategy, a lipophilic
polymer was used as secondary phase, acting as reservoir for both substrate and
product, facilitating a two-in-one resin-based in situ substrate feeding product
removal (SFPR) method [69]. According to this concept, careful adjustment of the
substrate/resin ratio will generate a certain concentration of the precursor in
the fermentation broth. In addition, the still hydrophobic product will be trapped
1448
by the resin after biotransformation, which offers significantly improved product

isolation upon continuous extraction of the easily recovered resin containing the
biotransformation product. This approach has been implemented with several
chemically diverse substrates in product concentrations of >30 mM [60]. In com-
bination with optimized aeration techniques, this method was also applied at pilot-
plant size [70].
33.4
Synthetic Applications
33.4.1
Enzyme Platform
Many of the particular features of these enzymes were discovered with the cyclo-
hexanone monooxygenase from Acinetobacter NCIMB 9871 (CHMOAcineto), repre-
senting a prototype biocatalyst for this whole flavoprotein group and in particular for
type 1 BVMOs [71]. However, the identity of the natural substrates is often elusive and
could only be assigned unambiguously in very few cases.
Within the biosynthetic pathway towards mithramycin in Streptomyces sp., the
tricyclic structure core of the metabolite is generated via a corresponding tetracyclic
precursor. Oxidative cleavage of the additional ring system was demonstrated to be
facilitated by the monooxygenase MtmOIV, which seems to belong to a significantly
different enzyme class than most BVMOs (Scheme 33.1) [72]. This was also
documented by comparing the structure of MtmOIV to classical type-1 BVMOs [73].
MeO
OMe
H H OH
R1O OH R1O
MtmOIV
O O
OH OH O O O OH OH O O
OR2 OR2
premithramycin B R1 = disaccharide, R2 = trisaccharide
Scheme 33.1 Baeyer–Villiger oxidation of premithramycin B by MtmOIV.
The gene cluster in streptomycetes related to the biosynthesis of pentalenolactone

and pentalenic acid, respectively, also contains an open reading frame with the typical
BVMO fingerprint motif. However, upon cloning and characterization of the
corresponding protein PtlE, the alternative regiochemistry was observed for the
oxygenation process, leading to neopentalenolactone D instead (Scheme 33.2) [74].
However, for most (type-1) BVMOs a primary substrate type can be identified
based on substrate specificity and efficiency of compound conversion. Historically, a
single primary substrate was often used in elective culture methods to isolate BVMO
33.4 Synthetic Applications j1449
PPO
FPP
COOH COOH
[O]
O O
O O
O
COOH pentalenolactone D pentalenolactone
O COOH
PtlE
O
neopentalenolactone D
O
Scheme 33.2 Baeyer–Villiger oxidation within pentalenolactone biosynthesis.
producing wild-type strains and served as identification feature and designator for the
biocatalyst. It is important to note that these assignments (e.g., cyclohexanone
monooxygenase; see Table 33.1) can be misleading, as detailed substrate profiling
studies may have revealed compounds serving as better precursors for the given
enzyme. For clarity and continuation of previous assignments, the historical des-
ignations are still used in this chapter.
During recent years, the number of BVMOs available in recombinant form
increased significantly and also the structural diversity of potential substrates was
dramatically enlarged. From a historical point of view, certain wild-type organisms
were utilized as source of crude enzyme preparations as well as whole-cell systems to
conduct biotransformations [75–79]. In several cases, the genetic origin of these
BVMOs was not established unambiguously in later work, so a comparison based on
sequence homology and biocatalytic performance is not possible.
The general feature of BVMOs representing highly promiscuous enzymes was
also found for the large part of newly discovered enzymes. Nevertheless, certain
trends could be identified among sub-clusters of biocatalysts and the currently
available dataset already allows general predictions on the substrate acceptance
profile of enzymes based on their phylogenetic relationship (Figure 33.7 and
Table 33.1).
Based on phylogenetic analysis and the substantial amount of data from
substrate screenings certain preferences could be identified for typical structural
motifs required for successful conversion, regioselectivity of oxygen insertion, as
well as for enantiopreference. Consequently, three major substrate types can be
identified so far as dividing criterion for BVMOs: cyclic ketones (as most
abundantly studied compound class; typical representatives: CHMOAcineto,
CPMOComa), aryl containing ketones (typical representatives: HAPMOPflu,
1450
Table 33.1 Recombinant BVMOs (selection) and typical substrate types.
BVMO/origin Discovery Cloning Substrate type Representative

substrate profile
Baeyer–Villiger monooxygenase (BVMOMtb5) Mycobacterium tuberculosis H37Rv 2006 [80] 2006 [80] Cyclic ketones [80, 81]
Baeyer–Villiger monooxygenase (BVMOKT2440) Pseudomonas putida KT2440 2007 [82] 2007 [82] Linear ketones [82]
Baeyer–Villiger monooxygenase (BVMOPvero) Pseudomonas veronii MEK700 2008 [83] 2008 [83] Linear ketones (short) [83]
Cyclododecanone monooxygenase (CDMORhodo) Rhodococcus SC1 2001 [84] 2001 [84] Cyclic ketones [85]
Cyclohexanone monooxygenase (CHMOAcineto) Acinetobacter NCIMB 9871 1976 [86] 1988 [87] Cyclic ketones [71]
Cyclohexanone monooxygenase (CHMOArthro) Arthrobacter BP2 2000 [89] 2003 [88] Cyclic ketones [85, 90]
Cyclohexanone monooxygenase (CHMOBrachy) Brachymonas petroleovorans 2003 [91] 2003 [91] Cyclic ketones [85, 90]
Cyclohexanone monooxygenase (CHMOBrevi1&2) Brevibacterium HCU 2000 [92] 2000 [92] Cyclic ketones [85, 93]
Cyclohexanone monooxygenase (CHMORhodo1&2) Rhodococcus Phi1 & Phi2 2003 [88] 2003 [88] Cyclic ketones [85, 90]
Cyclohexanone monooxygenase (CHMORhodo-HI31) Rhodococcus HI-31 2009 [33] 2009 [33] Cyclic ketones [33]
Cyclohexanone monooxygenase (CHMOXantho) Xanthobacter sp. ZL5 2003 [94] 2003 [94] Cyclic ketones [95, 96]
Cyclopentadecanone monooxygenase (CPDMOPseudo) Pseudomonas HI-70 2006 [97] 2006 [97] Cyclic ketones [97]
Cyclopentanone monooxygenase (CPMOComa) Comamonas NCIMB 9872 1976 [98] 2002 [99] Cyclic ketones [90, 99]
Hydroxyacetophenone monooxygenase (HAPMOPflu) Pseudomonas fluorescence ACB 2001 [100] 2009 [100] Aryl ketones [101, 102]
Hydroxyacetophenone monooxygenase (HAPMOPput) Pseudomonas putida JD1 2009 [103] 2009 [103] Aryl ketones [103]
Linear ketone monooxygenase (BVMOPflu) Pseudomonas fluorescens DSM 50106 2006 [104] 2006 [104] Linear ketones [104]
Monooxygenases MO1-23 (BVMO library) Rhodococcus jostii RHA1 2009 [105] 2009 [105] Linear & cyclic ketones [105]
Phenylacetone monooxygenase (PAMOThermo) Thermobifida fusca 2004 [106] 2005 [107] Aryl ketones [108]
Steroid monooxygenase (STMORhodo) Rhodococcus rhodochrous IFO 3338 1999 [109] 1999 [109] Steroid side chain [109]
Figure 33.7 Phylogenetic relationships within BVMOs. Protein sequences of cloned biocatalyst
with confirmed BVMO activity were aligned; an unrooted phylogenetic tree was calculated using
ClustalW2 and represented graphically using Dendroscope.
PAMOThermo), and linear ketones (typical representative: BVMOPflu). Usually,

BVMOs display a large promiscuity within their specific substrate class, tolerating
the presence of lipophilic as well as polar functional groups (including heterocyclic
systems). The identification of certain sub-clusters of BVMOs (CHMO- and
CPMO-type ketone converting biocatalysts) with overlapping substrate specificity
was particularly relevant [90, 110]. Based on biocatalyst performance and phylo-
genetic relationship CHMOAcineto represents a prototype-enzyme for its sequence-
related neighborhood in a similar fashion as CPMOComa. Representatives of both
groups form enantiocomplementary lactone products and such divergent behavior
of catalysts was also observed in regiodivergent bio-oxygenations [111]. A recent
study on the first comprehensive expression library of all BVMOs present in the
genome of Rhodococcus jostii confirmed such classification approaches [105]. As
can be deduced in part from the enzyme designation, several BVMOs also differ in
the size of substrates that can be accommodated within the active site of the
biocatalyst. Cyclopentadecanone monooxygenase (CPDMOPseudo), for example,
1452
can even transform steroidal structures [112] (with, however, rather limited
efficiency), while BVMOPvero was reported to convert short-chain linear
ketones [83]. Nonetheless, we are only just starting to elucidate the general
principles of substrate acceptance among the various BVMOs. It remains a long
way from our present knowledge to a comprehensive and predictive understand-
ing, in particular of phylogenetics and biocatalyst performance.
33.4.2
Chemoselectivity
BVMOs can entertain two different oxygenation pathways, as already outlined in the
discussion of the enzyme mechanism (Figure 33.1). Within the main reaction cycle
the peroxy-flavin moiety will attack the substrate in a nucleophilic fashion following
the traditional mechanism for the Baeyer–Villiger reaction. However, the formation
of the corresponding flavin-hydroperoxide opens up an alternative electrophilic
pathway. Hence, the flavin moiety as key catalytic entity in BVMOs displays a certain
flexibility in reactivity [113, 114]. It is generally accepted that this electrophilic
reaction is responsible for oxygenations at heteroatom centers. BVMOs can conduct
various oxidation reactions at sulfur leading also to valuable chiral building
blocks [115–118]; more recent studies have also taken advantage of recombinant
BVMOs to provide access to enantiocomplementary sulfoxides [119]. Similar oxyge-
nations have been observed for nitrogen [120], boron [121], and selenium [122].
When a BVMO is challenged with a substrate bearing both a carbonyl center
and an oxidizable heteroatom, usually oxygenation of the ketone is favored. This
was demonstrated in a principal investigation of heterocyclic substrates contain-
ing sulfur or non-protected nitrogen (1); such compounds were transformed into
the corresponding lactone products (2) in moderate to good chemical yields
(Scheme 33.3) [123, 124].
O O
CHMOAcineto O
X = S, O, N-PG
R X R R X R
R = H, alkyl
1 2
Scheme 33.3 Preferred Baeyer–Villiger reaction to heteroatom oxygenation by BVMOs.
The functional tolerance of BVMOs towards groups labile to oxidation under

conventional chemical conditions was nicely confirmed in this survey: heterocyclic
systems bearing olefinic side chains (Scheme 33.3; R ¼ CH ¼ CH2) were cleanly
converted into the corresponding lactones without compromising the integrity of the
alkene group [125]. However, in a single case an electron-deficient Michael-type
acceptor alkene has been reported to undergo epoxidation in the case of alkenyl-
phosphonate precursors [126]; this reaction can also be interpreted as special case of
the nucleophilic oxygenation pathway. In a recent contribution is was observed, for
the first time, that even a non-activated C¼C double bond system can react to give an
epoxide and this biotransformation was clearly assigned to the activity of a CHMO-
type enzyme (Scheme 33.4) [96]. Notably, epoxide formation (5) by CHMOXantho is
only observed in the case of a heterocyclic substrate (3, X ¼ O), while the correspond-
ing carbocyclic analog (3, X ¼ CH2) was converted selectively (and in high optical
purity) into the expected lactone 4. This suggests a key role of the additional oxygen
atom in the substrate as coordinating center for the oxygenating species, and also
provides some indication for the exclusive diastereoselectivity of oxygen incorpo-
ration at the alkene. When using CPMOComa (the prototype enzyme for the CPMO-
type sub-cluster of BVMOs) for the biotransformation of the heterocyclic substrate 3
(X ¼ O) no unexpected olefin oxidation occurred but rather clean conversion into
lactone 4 was observed.
CHMOXantho
O CHMOXantho O O
X = CH2
X=O 99% e.e. O
O X CPMOComa X
X=O
95% e.e.
O 5 3 4
Scheme 33.4 Functional group selectivity of BVMOs.
33.4.3
Desymmetrizations
Desymmetrization reactions of prochiral substrates represent the most facile and

potent applications of enzymatic Baeyer–Villiger oxidations in synthesis, as they
allow for a theoretical yield of 100% of optically pure products bearing one or more
stereogenic centers established in a de novo fashion [127]. A large diversity of ketones
has been investigated, incorporating structural motifs of varying complexity. The
identification of BVMOs enabling access to enantiocomplementary lactones was a
major advance in the field [90], as broader application in bioactive compound
synthesis became feasible.
Prochiral cyclobutanones and cyclohexanones with various numbers of sub-
stituents were investigated (Table 33.2). BVMO-mediated formation of butyrolac-
tones by cycloketone converting enzymes was demonstrated as highly facile
transformation using recombinant expression hosts in a whole-cell format [128].
This facile reaction is most likely related to the liberation of ring-strain upon
rearrangement and ring expansion. Compounds containing oxygenated aryl
groups turned out to be particularly good substrates, opening up novel synthetic
routes to certain bioactive compounds and natural products. Similarly, 4-substi-
tuted cyclohexanones were successfully transformed into enantiocomplementary
lactones. Even sterically demanding precursors containing quaternary carbon
1454
Table 33.2 Monocyclic substrates for BVMO-mediated lactone formation (representative

examples).
Ketone Substitutiona) Enzyme E.e. (%)b) Yield (%) References
R ¼ Ph CHMOBrevi1 98 () 73 [128]

O CPMOComa 37 (þ) 66 [128]
R ¼ 4-Cl-Ph CHMOBrevi1 87 () 47 [128]
R CHMORhodo2 95 (þ) 63 [128]
R ¼ 3,4-(OCH2O)-Bn CHMORhodo2 98 () 52 [128]
CHMOBrevi1 75 (þ) 61 [128]
R ¼ 3,4,5-(MeO)3-Bn CHMOArthro 94 () 72 [128]
CHMOBrevi1 79 (þ) 72 [128]
R ¼ Me CHMOAcineto >98 () 83 [131]

O CPMOComa 46 (þ) 68 [132]
R''' R''' R¼I CHMOAcineto 97 () 60 [133]
CPMOComa 82 (þ) 65 [134]
R'' R''
R R'
R ¼ COOEt CHMOXantho 98 () >90 c)
[96]
CPMOComa 64 (þ) 83 [134]
R ¼ tert-Bu CHMOXantho 99 () 82 [96]
R ¼ Ph CHMOXantho 98 () 88 [96]
R ¼ Ph, R0 ¼ Me CHMOXantho 95 () 60 [96]
00
R ¼ Me CHMOBrevi1 97 () 61 [93]
CHMOBrevi2 99 (þ) 56 [93]
R ¼ R0 ¼ ¼CH2, CHMOBrevi1 99 (þ) 70 [93]
R00 ¼ Me CPMOComa 99 () 63 [90]
R000 ¼ Me CPDMOPseudo 99 (N.r.) 74c) [97]
a) R, R0 , R00 , R000 ¼ H unless specified otherwise.

b) Sign of specific rotation in N.r.: not reported.
c) Conversion data from screening experiments.
centers as 4-position and/or bulky substituents at positions 2/6 or 4 are well

accepted, in particular by CHMOXantho [96] and CPDMOPseudo [97]. A characteristic
rearrangement process of the expected lactone is encountered when the substrate
ketone also bears a hydroxyl function (Scheme 33.5). The initially formed seven-
membered ring system, which cannot be captured under biotransformation
conditions, collapses to the thermodynamically more stable five- and six-ring
lactone, respectively [129]. The rearrangement is independent of the diastereo-
meric relationship of the hydroxyl group relative to other substituents [130] and is
also observed for compounds containing a tert-alcohol structural motif [90].
O O
R
BVMO O
O OH
R R O
R R R'
HO R' R
HO R' R, R' = H, Me
O O
BVMO
O
OH
6 OH 7
Scheme 33.5 Rearrangement of hydroxyl-containing ketones subsequent to Baeyer–Villiger

bio-oxidation.
Both lactone enantiomers are accessible in high optical purity bearing func-
tional groups of diverse electronic properties based on a significant number of
substrates, especially in the cyclohexanone series. The sub-classification of
cycloketone converting BVMOs into a CHMO- and a CPMO-cluster based on
phylogenetic relationship was also reflected in the stereodivergent behavior of
these biocatalysts in cases of overlapping substrate acceptance profiles; represen-
tatives of each group provided access to antipodal lactones. While CHMOBrevi1 was
initially considered as a member of the CHMO-cluster (this was also in line with a
phylogenetic tree based on the limited number of sequences available at the time
of that study [90]), oxygenations of cyclobutanones in particular revealed some
special features of this enzyme. In the current phylogenetic analysis this enzyme
adopts a clearly separated position from both CHMO- and CPMO-groups, which is
also reflected by the enantiocomplementary oxygenation of cyclobutanones rel-
ative to other BVMOs.
Early studies on CHMOAcineto already indicated that the active site of BVMOs can
also accommodate sterically demanding bicyclic ketones [135]. Fused cycloketones
are converted by several enzymes accepting also functional decoration at the
carbocyclic core [136]. In contrast to poorly accepted a,a0 -disubstituted cyclohex-
anones, bicyclic precursors with the carbonyl function at the connecting bridge are
also readily converted by BVMOs; liberation of ring-strain upon oxygenation may be a
relevant contribution to facilitating this transformation of such otherwise sterically
constrained systems. Within the desymmetrization of such compounds, up to six
stereogenic centers can be established in a single biotransformation (certainly, the
relative configuration of the scaffolds has to be predetermined). Even when com-
bining the structural features of fused and bridged polycyclic systems such com-
pounds are converted by certain BVMOs, with, however, moderate efficiency and
limited stereoselectivity (Table 33.3).
1456
Table 33.3 Polycyclic substrates for BVMO-mediated lactone formation (representative examples).
Ketone Substitution Enzyme E.e. Yield Reference

(%)a) (%)
H R ¼ exo >CHCl CHMOAcineto >99 () 78 [90]

CHMOBrevi2 60 (þ) 59 [90]
R O
R ¼ CH ¼ CH CHMOXantho 88 () >90b) [90]
H CPMOComa >99 (þ) 76 [90]
R ¼ -C3H6- CHMORhodo2 99 () 63 [137]

O
CHMOBrevi2 92 (þ) 67 [137]
R ¼ -C4H8- CHMORhodo1 99 () 58 [137]
CHMOBrevi2 94 (þ) 78 [137]
R
R R ¼ -CH2OCH2- CHMOAcineto 92 () 53 [137]
CPMOComa 71 (þ) 49 [137]
R ¼ 1,3-connected cyclopentyl CHMOAcineto 98 () 51 [137]
X
X ¼ CH2 CHMOXantho >99 () >50b) [96]
O X¼O CPMOComa 95 (þ) 53 [89]
Endo CHMOXantho 94 (þ) <50b) [96]

O CPMOComa 36 () 73 [58]
Endo CHMOXantho 43 (þ) >90 b)
[96]
a) Sign of specific rotation in parentheses.

b) Conversion data from screening experiments.
33.4.4
Kinetic Resolutions
Classical kinetic resolution of racemic ketones has been conducted both utilizing
purified BVMOs as well as wild-type and recombinant whole-cell systems. A
particular focus was put on the oxygenation of cycloketone precursors as constrained
scaffolds for subsequent chemical elaboration. Under conventional biotransforma-
tion conditions, the conversion leads to a maximum 50% yield of chiral lactone
product while the antipodal substrate remains unchanged in optically enriched form
(Scheme 33.6). The regioselectivity for the enzymatic process is usually governed by
electronic effects and correlates with the chemical reaction. Consequently, preferred
migration of the more nucleophilic center is observed and oxygenation commences
at the CC bond between the carbonyl center and the higher substituted a-carbon.
This conventional outcome of the rearrangement process is usually referred to
as normal lactone product. Deviations from this expected behavior of BVMOs
are discussed below in Section 33.4.5 on regioselectivity; in such abnormal or
O O O
(R)
R BVMO O R
(S) +
( )n ( )n
( )n R
= CH2CH2, CH=CH
Scheme 33.6 Conventional kinetic resolution of racemic cycloketone substrates to normal chiral
lactones and optically enriched ketones (absolute configuration shown for CHMOAcineto
transformations).
non-conventional cases, stereoelectronic effects may override the nucleophilicity in

the carbon migration.
The generally accepted parameters to quantify selectivity of a particular BVMO in
kinetic resolution processes is the enantiomeric ratio, E [138], which corresponds to
the ratio of the relative second-order rate constants of the individual substrate
enantiomers and (theoretically) remains constant throughout the transformation.
To allow for a process to deliver yields close to the theoretical limit (50%), E is required
to be >30.
While a distinct sub-classification into groups of BVMOs enabling access to
enantiocomplementary lactones was observed within desymmetrization reactions,
most biocatalysts studied within kinetic resolutions, so far, display a pronounced
preference for the generation of (S)-lactones (Note: in certain cases (R/S) assign-
ment may be affected by chain branching or functionalization leading to a change in
priority numbering; however, the sense of chirality remains the same). BVMOs
selectively recognize the preexisting chiral center and only convert (S)-ketones, while
the opposite enantiomers remain unchanged (Table 33.4). Generally, kinetic reso-
lution proceeds with better selectivity when a-substituents contain at least three
carbon centers. Functional groups within the side-chain are well tolerated, including
substituents labile to oxidation (e.g., allyl).
While studies have clearly focused on cyclopentanone and cyclohexanone pre-
cursors, two particularly distinct and interesting biotransformations were reported:
5-hexyl-cyclopent-2-enone is the only substrate investigated with an unsaturated
cyclic core [139]; here, migration also commences towards the more substituted
center (and is not affected by the electron-rich olefin function). As indicated by the
designation of the enzyme, CPDMOPseudo can convert larger carbocycles, with,
however, only very moderate efficiency and selectivity [97].
Although the enzymatic Baeyer–Villiger oxygenation of aryl [146] and linear
ketones [147] was reported early on in the description of this biotransformation
(primarily using wild-type organisms), only recently was the systematic investigation
of recombinantly available enzymes initiated (Scheme 33.7, Table 33.5).
As indicated by their designations, PAMO and HAPMO-type enzymes are
particularly suited biocatalysts for aryl containing linear ketones. Within a detailed
study of linear a-substituted a-aryl-ketones and -aldehydes, good to excellent enan-
1458
Table 33.4 Kinetic resolution of racemic a-substituted cycloketonesa) (representative examples).
Ketone R Enzyme E.e. (%)b) Yield (%) E Reference
Me CHMOAcineto 32 (N.r.), 44 (N.r.) 36, 34 3.6 [140]

O
Et CHMOAcineto 39 (N.r.), 46 (N.r.) 44, 37 3.7 [140]
R
n-Pr CHMOAcineto 67 (), 72 () 51, 21 30 [140]
n-Hex CHMOAcineto >98 (), >98 () 32, 42 >200 [140]

n-Undec CHMOAcineto >98 (), >98 () 39, 37 >200 [140]
CH2OBn CHMOAcineto 97 (), 43 (N.r.) 43, 48 N.r. [141]
O
R n-Hex CPMOComa 75 (þ), 38 () 52, 43 10 [139]
O
Me CHMOAcineto 61 (), 35 (þ) 35, 52 6 [142]
R
CDMORhodo N.r. N.r. >200 [85]
CPDMORhodo N.r. N.r. >200 [97]
Et CHMOAcineto 95 (), >98 () 40, 35 >200 [143]
Allyl CHMOAcineto >98 (), >98 () 30, 29 >200 [143]
n-Non CHMOAcineto 85 (), 42 () 26, 32 20 [142]
Ph CHMOAcineto >98 (þ), 86 () 40, 48 >100 [142]
Bn CHMOAcineto >96 (), 78 () 22, 28 >100 [142]
CH2COOEt CHMOAcineto >99 (), 64 () 39, 60 N.r. [144]
CH2CH2OAc CPMOComa 42 (þ), 68 () 59, 37 5 [145]
O
R
Me CPDMORhodo 59 (N.r.) 10c) 6 [97]
a) Data for lactones in normal font, data for ketones in italics; N.r.: not reported.
b) Sign of specific rotation in parentheses.
c) Conversion data from screening experiments.
33.4 Synthetic Applications
j1459
1460
R'' R'' O R''

(a) R' BVMO R'
(S) O R' + (R)
O O
(b) O OPG BVMO OPG O OPG

R O +
R ( )n (S) ( )n R (R) ( )n
O
Scheme 33.7 Conventional kinetic resolution of racemic linear ketone substrates to normal
chiral esters and optically enriched ketones.
Table 33.5 Kinetic resolution of racemic linear cycloketonesa) (representative examples).
Ketone Substitution Enzyme Conversion E.e. E Reference

(%) (%)b)
R0 ¼ H, PAMOThermo 36 88 (S), 25 [149]

R''
R ¼ Me 50 (R)
R'
R0 ¼ R00 ¼ Me PAMOThermo 27 98 (S), 188 [149]
O 36 (R)
CHMOAcineto 41 8 (S), <3 [148]
3 (R)
CPMOComa 93 6 (S), <3 [148]
90 (R)
R0 ¼ R00 ¼ Et PAMOThermo 51 95 (S), 179 [149]
98 (R)
O OPG R ¼ Me, BVMOPflu 46 >99 (S), >200 [151]

PG ¼ H, n ¼ 3 85 (R)
R ( )n
CHMOAcineto 49 96 (S), 156 [151]
91 (R)
R ¼ Me, CPMOComa 50 99 (R), >200 [151]
PG ¼ CHO, n ¼ 3 99 (S)
R ¼ Me, PG ¼ H, BVMOPflu 41 96 (S), 100 [151]
n¼5 68 (R)
CHMOAcineto 50 93 (S), 90 [151]
92 (R)
R ¼ Me, PG ¼ Ac, CPMOComa 25 >99 (R), >200 [151]
n¼5 45 (S)
R ¼ Me, PG ¼ H, BVMOPflu 45 90 (S), 41 [104]
n¼7 74.5 (R)
a) Data for esters in normal font, data for ketones in italics.

b) Absolute configuration in parentheses.
tioselectivities were observed (Scheme 33.7a) [148, 149]. It was found that the
efficiency of the resolution largely depended on pH as well as reaction temperature
and could be facilitated by addition of organic co-solvents (especially MeOH when
using PAMO) [150]. Interestingly, cycloketone converting enzymes from the CHMO-
and CPMO-cluster were also capable of converting such precursors, with, however,
only low selectivity.
The discovery of BVMOPflu expanded the substrate profile towards linear pre-
cursors, significantly, as this biocatalyst can convert long-chain linear ketones [104]
without the necessity of an aryl substituent (Scheme 33.7b). The overall reaction can
be hampered to a certain extent by acyl migration reactions when using hydro-
xyketones (PG ¼ H), complicating the obtained product mixture. Still, it represents
an interesting novel utilization of the carbonyl group: Upon conversion of the ketone
into an ester, the carbonyl function serves as a masked hydroxyl group in a retro-
synthetic context.
In context of the above finding that previously characterized cycloketone convert-
ing enzymes were found to be capable of oxidizing straight chain carbonyls, it was
even more remarkable that representatives of the CHMO- and CPMO-type groups
could effectively give highly specific kinetic resolutions of linear hydroxyke-
tones [151]. Protection of the alcohol function even enabled formation of enantio-
complementary esters.
To overcome the limitation in yield for conventional kinetic resolutions, the
intrinsic acidity of a-protons at a carbonyl center can be exploited, which is also
reflected by the capacity of such a configuration to tautomerize into an enol form.
This allows in theory to set up an in situ equilibrium for the racemization of the
stereogenic center in the substrate as prerequisite for a dynamic kinetic resolution
(DKR). As long as the epimerization of the substrate does not affect the stereogenic
center of the product (the a-acidity of a ketone is usually significantly higher than that
of the corresponding lactone or ester) such an approach enables theoretically the
synthesis of enantiomerically pure compounds in 100% yield from a racemate. The
critical aspect in this concept is the ability of a substrate to racemize under conditions
compatible with the stability range of the biocatalyst. Enol formation usually requires
basic or acidic conditions. However, as the racemization rate may be slow relative to
the bio-oxidation rate, the choice of highly enantioselective biotransformations is
mandatory.
So far, this concept has been successfully implemented with substituted
cyclopentanone precursor 8 under whole-cell conditions using CHMOAcineto
(Scheme 33.8). Fermentations were conducted under basic conditions (pH 9) in
a straightforward fashion, enabling isolation of 85% chemical yield of the
corresponding lactone 9 in 96%ee, clearly indicating the operation of a
DKR [152]. Improved compound titers were achieved by employing the SFPR
concept due to milder cultivation conditions [153]. In this case, an anion-
exchanging resin was utilized to establish racemization conditions. Remarkably,
best results were obtained using a weakly basic resin compared to high-pH
materials.
1462
O O O
racemization CHMOAcineto
(S) (R)
O
OBn OBn (R)
OBn
(S)-8 (R)-8
(R)-9
75-85%, 96% e.e.
Scheme 33.8 Dynamic kinetic resolution using BVMOs; racemization conditions employed: basic
pH, SFPR using basic resins, and high concentration imidazole buffers.
Alternatively, DKRs were also carried out in the presence of a high concentration of
phosphate and imidazole buffers, as racemization was found to be one order of
magnitude faster in such buffers than in conventional fermentation broth. Com-
parable yields and optical purities were obtained for biotransformations at pH 7.2
without large amounts of solid phase.
33.4.5
Regioselectivity
Rearrangement of the Criegee intermediate as covalent connection between sub-

strate and oxygenated FAD is governed by two principal effects: (i) the pure electronic
effect reflects the electron density at the migrating center and determines the
nucleophilicity of the rearranging carbon; (ii) the stereoelectronic effect defines certain
arrangements of the bond system and lone electron pairs within the Criegee
intermediate that allow for a release of electrons and rearrangements of the bonding
system. It has been demonstrated that two prerequisites must be satisfied for
successful alkyl migration and carboxylic acid ejection: (i) the migrating CC bond
has to be in an antiperiplanar position relative to the peroxy-bond; (ii) effective
electron release from the hydroxyl oxygen towards the migration origin attached is
essential for the alkyl shift and requires a lone pair in an anti position at the oxygen
atom (Figure 33.8) [154]. Using isotopically labeled substrates it was also confirmed
for the enzymatic reaction that the fragmentation of the tetrahedral intermediate
proceeds with retention of configuration at the migrating center, analogous to the
chemical oxidation [155]. Hence, BVMOs can catalyze regio- and enantioselective
oxidations by permitting only one CC bond to adopt an antiperiplanar configuration
R O
antiperiplanar
O
O
O
migrating group non-migrating group
H
anti
Figure 33.8 Stereoelectronic requirements within the Criegee intermediate for successful
migration.
with the OO bond in the Criegee-type intermediate. Since the FAD is (relatively)
tightly bound to the enzyme, interaction of the substrate with amino acids in the
active site influences this alignment and the migrating group selection in the case of
two possible rearrangement options. In such systems, the compound expected for the
rearrangement process dominated by the electronic effect is often referred to as the
normal or conventional product, while overriding by the stereoelectronic effect
gives rise to the non-conventional isomer often referred to as the abnormal product.
In most cases, strict migratory preference of the more nucleophilic center is
observed, hence reflecting the dominance of the electronic effect. However, regio-
divergent oxygenations were observed for conversions of various 1-indanones: while
HAPMOPflu usually gave the expected lactone, a mutant of PAMOThermo (M446G)
selectively afforded the abnormal isomer (Scheme 33.9) [156]. The complementary
behavior of the two BVMOs tolerates both electron-donating and -withdrawing
substituents at the aromatic core and is particularly pronounced when HAPMO-
mediated transformations were conducted in the presence of 5% hexane, while
PAMO-mediated conversions were facilitated by adding 5% MeOH.
O O
R R R O O
O PAMO-mutant HAPMOPflu
abnormal lactone normal lactone
Scheme 33.9 Regiocomplementary oxygenation of 1-indanones by BVMOs leading to isomeric

lactones.
The chemical Baeyer–Villiger oxidation of substituted benzaldehydes represents a

particular scenario; the presence of electron-donating substituents promotes forma-
tion of phenol esters (and subsequently phenols), whereas neutral or electron-
withdrawing groups invert the regioselectivity, leading to benzoic acids [157].
Remarkably, (poly)fluorinated benzaldehydes were oxygenated by HAPMOPflu to
phenol esters (phenols were detected as hydrolysis products) predominantly as regio-
isomers to the corresponding chemical reaction (Scheme 33.10) [158]. This particular
R = electron donating
O O
R R chemical R
chemical O O
OH H
H
HAPMOPflu
R = electron withdrawing
R = electron withdrawing
Scheme 33.10 Regiocomplementary chemical and biological oxygenation of benzaldehydes

depending on the electronic nature of the aryl substituent.
1464
behavior of BVMOs can be regarded as the biological equivalent to the Dakin

reaction.
Formation of the conventional products is usually observed within the series of
a-substituted cycloketones (Table 33.4). Both chemical and enzymatic oxidation lead
to the expected proximal (normal) lactones independent of the nature of the
functional decoration in the side chain, as the difference in electron density between
carbons bearing the additional chain and without substitution is significant. So far,
only a single deviation from this behavior has been reported. In the case of cyano-
substituted cyclohexanone 10 CHMOAcineto displayed a distinct regiodivergent
oxygen insertion for the two enantiomeric ketones: racemic substrate was converted
selectively into the (R)-proximal lactone 11 and to the (S)-distal product 12 in high
optical purities (Scheme 33.11) [159]. This particular switch in migratory preference
was only observed for a methylene tether; a-cyano-cyclohexanone was not converted
by the enzyme as enolizable ketone and the ethyl-tethered analog displayed con-
ventional kinetic resolution to the proximal lactone, exclusively.
O O O
CHMOAcineto O (R) O (S)
CN +
CN CN
rac-10 (R)-11 (S)-12

32%, >99%ee 36%, >99%ee
Scheme 33.11 Stereodivergent oxygenation of a cyano-substituted cyclohexanone to

regioisomeric lactones in high optical purity.
As a consequence of the conversion of both substrate ketones into regioisomers,

different bonds have to occupy the antiperiplanar position required for successful
migration after the formation of the corresponding Criegee intermediate. Hence, two
concomitant binding modes have to be allowed for both enantiomers. A similar
regiodivergent behavior by BVMOs was discovered recently within the bio-oxygen-
ation of terpenones. In addition here, distal and proximal oxygen insertion was
observed for an a-substituted cyclohexanone system. Three enzymes were identified
to convert optically pure precursors 13 selectively into the two diastereomers (normal
and abnormal lactones) in both antipodal forms. This is the first reported access to all
four possible isomers in complementary fashion, which display remarkable prop-
erties as aroma compounds. The study impressively demonstrates how the exquisite
spatial arrangements within the individual active sites impose particular stereoelec-
tronic arrangements for different rearrangement preferences (Scheme 33.12) [160].
The regiodivergent bio-oxygenation of fused cyclobutanones was discovered early-
on in studies of BVMOs [161]. In addition, the racemic precursor is transformed by
the biocatalyst into two types of regioisomeric lactones upon diverse rearrangement
of the individual substrate ketones. Within this substrate class, a mechanistic
rationale was developed to explain the unusual accommodation of the two substrate
enantiomers in a different conformational state [162] and was actually confirmed by
O
O
O
O
CHMOAcineto
CHMOBrevi1
(+)-14
(-)-14
34% 77%
(-)-normal lactone O O (+)-normal lactone
CPMOComa
CHMOAcineto
(+)-13 (-)-13 O
O (+)-dihydrocarvone (-)-dihydrocarvone 18%
70% O
O
(+)-15
(-)-abnormal lactone (+)-abnormal lactone
(-)-15
Scheme 33.12 Regiodivergent BVMO-mediated oxygenation of dihydrocarvone.
successful oxygenation of a super-substrate combining both structural features of

the precursor antipodes [163]. Similar results were also found for various camphane
derivatives [164].
Based on the phylogenetic relationship of CHMO- and CPMO-type enzymes,
again a general trend in biocatalyst performance could be identified for the two
clusters (Scheme 33.13, Table 33.6): representatives of the CHMO-type clade display
a clear regiodivergent bio-oxygenation and racemic precursors are converted into
isomeric normal and abnormal lactones in approx. 1 : 1 ratios [regioisomeric
excess (r.e.) approx. 0] and in high optical purities. In contrast, CPMO-type bioca-
talysts preferentially yield normal lactones (high r.e.), in, however, racemic
form [111]; this enables chemoselective entry to chiral lactones starting from optically
pure precursors. Recent substrate profiling of various BVMOs originating from M.
O
H H
O
O + O
CHMO-type H H
normal lactone abnormal lactone
H O
O CPMO-type
O
H
normal lactone rac O
BVMOMtb5 H H O
O +
H H
abnormal lactone chiral ketone
Scheme 33.13 Divergent oxygenations of fused cyclobutanones by different BVMOs.

1466
Table 33.6 Regiodivergent bio-oxidation of fused bicyclobutanones (representative examples).
Ketone Enzyme S Yield (%)a) R.e. (%)b) E.e.P (%)c) E.e.s (%)d) Reference
CHMOAcineto 86 2 >95, >95 N.r. [161]

H O
CPMOComa 61 94 0, >99 N.r. [90]
BVMOMtub5 95e) 72 56, 86 >99 [81]
H
HAPMOPflu 56 36 75, 32 N.r. [102]
CHMOBrevi1 77 4 89, >99 N.r. [111]

H O
CPMOComa 89 74 14, >99 N.r. [111]
BVMOMtub5 68 e)
96 14, >99 >99 [81]
H
HAPMOPflu 79 16 0, 95 N.r. [102]
CHMOBrevi1 64 4 99, >99 N.r. [111]

H O
CPMOComa 87 82 6, >99 N.r. [111]
BVMOMtub5 50e) 100 N.a., >99 >99 [81]
O
H
HAPMOPflu 79 80 52, >99 N.r. [102]
a) Combined isolated yield of normal plus abnormal lactones.

b) Regioisomeric excess (r.e.) (normal font: favoring normal lactone, italics: favoring abnormal
lactone).
c) E.e.p ¼ e.e. of products (normal font – normal lactone, italics – abnormal lactone); N.a.: not
applicable.
d) E.e.s ¼ e.e. of substrate (where applicable); N.r.: not reported.
e) Conversion data from screening experiments.
tuberculosis revealed BVMOMtb5 as selective biocatalyst for the preparation of

abnormal lactones for certain substrates. This enzyme conducts a kinetic resolu-
tion of racemic bicyclobutanones to give optically pure ketone and lactone, hence
complementing the performance of the above-described BVMOs [80, 81]. As a
borderline case, HAPMOPflu displays predominant formation of abnormal lac-
tones starting from precursors containing a cyclopentane ring motif, while substrates
bearing a cyclohexane structural motif are predominantly converted into normal
lactones, in, however, only moderate optical purities in both cases [102].
In the case of b-substituted cycloketones the electronic properties of the potentially
migrating carbon centers are highly similar; hence, formation of product mixtures
composed of proximal and distal lactones is usually observed in chemical Baeyer–
Villiger oxidations (Scheme 33.14). The difference in nucleophilicity of the two
a-carbons is too insignificant to lead to selective migration; consequently, any
external geometry imposed on the Criegee intermediate will largely dominate
migratory preference via stereoelectronic interaction. BVMOs were again found to
effectively control formation of either proximal or distal products depending on the
configuration of the starting material [165]. Both CHMO- and CPMO-type enzymes
can clearly differentiate between antipodal forms of b-substituted cyclopentanones
and cyclohexanones to give clean regiodivergent oxidation. CPMOComa displayed a
O O
O O
+
( )n R ( )n R
O proximal products
( )n R O O
O O
n = 0,1
+
R ( )n R ( )n
distal products
Scheme 33.14 Bio-oxidation of b-substituted cycloketones to regioisomeric products.
high tolerance towards longer side chains and shows stereo-complementary behavior
to CHMOAcineto in some cases (Table 33.7). In addition, regiodivergent biotransfor-
mations for CPDMOPseudo [97] and CHMORhodo-HI31 [33] were reported, with,
however, incomplete characterization and assignment of structure.
During the elucidation of the bio-oxidative degradation of camphor [166] it became
apparent that several BVMOs are involved, operating in part on substrates containing
Table 33.7 Bio-oxidation of chiral b-substituted cycloketones (representative examples).a)
Ketone R E.e. (%)b) Enzyme Yield (%) E.e. (%)d) Reference

(ratio)c)
Me 100 (R) CHMOAcineto 77 (>99 : 1) 100 (R) [165]

O
CPMOComa 75 (>99 : 1) 100 (R) [165]
100 (S) CHMOAcineto 60 (1 : 99) 100 (S) [165]
R Et 81 (R) CPMOComa 87 (>99 : 1) 80 (R) [165]
n-Bu 83 (S) CPMOComa 85 (>99 : 1) 83 (S) [165]
Me >99 (R) CHMOAcineto 88 (10 : 90) 99 (R), 99 (R) [165]

O
CPMOComa 62 (100 : 0) >99 (R) [165]
Et 86 (R) CPMOComa 69 (>99 : 1) 86 (R) [165]
R n-Bu 90 (S) CPMOComa 84 (>99 : 1) 92 (S) [165]

(CH2)2Ph 88 (S) CHMOAcineto 80 (>99 : 1) 99 (S) [165]
a) Data for proximal lactones in normal font, data for distal lactones in italics.
b) The e.e. of the ketone after chemical reduction (absolute configuration in parentheses).
c) Combined isolated yield and ratio of proximal and distal lactones.
d) E.e. of lactones (absolute configuration in parentheses).
1468 j 33 Baeyer–Villiger Oxidations
more than one carbonyl center. Consequently, certain monooxygenases are clearly
capable of chemoselectively attacking a particular ketone function in the presence of
multiple additional C¼O groups [167]. However, systematic studies of such poly-
ketone substrates are lacking to a large extent. Regioselective oxygenation of the
sterically less congested carbonyl center was already reported early on in investigating
the biocatalytic performance of CHMOAcineto. This transformation proceeds via a
desymmetrization process in high regio- and stereoselectivity to the corresponding
keto-lactone 17 (Scheme 33.15a) [168]. Kinetic resolutions on multi-ketone substrates
have also been reported. The enzymatic oxidation of racemic Wieland–Miescher
ketone 18 with CHMOAcineto was fully selective for the non-conjugated carbonyl
center at position 1 (Scheme 33.15b). The resolution process gave access to the
expected normal lactone 19 with (S)-configuration in high optical purity [169].
Derivatives of this precursor were also converted by the BVMO in kinetic resolution
processes with regioselective oxygenation at position 1.
In a study screening for BVMO-mediated oxygenations of various steroids, a
particularly interesting observation was made in the androstane series. CPDMOPseudo
O O
(a) CHMOAcineto O
O 16 O 17
25%, >98%e.e.
O O O
(b) O
CHMOAcineto
1 (S) 2 + (R)
6 7
O O O
19 18
(rac)-18
35%, 99% e.e. 43%, 80% e.e.
O O
(c) O O
CPDMOPseudo 17a
+
O O
20 4 21 4 22
O O O
11%
30%
Scheme 33.15 Selective BVMO-mediated bio-oxidation of poly-ketone substrates.
was identified as a biocatalyst capable of converting such structures, with, however,

quite modest conversions and requiring high dilution conditions. With androstan-
3,17-dione (20), the steroidal A-ring was attacked preferentially and in high regios-
electivity to give 4-oxa-lactone 21 in 30% yield. Subsequent oxygenation of the D-ring
provided access to the corresponding bis-lactone 22 in minor quantities as follow-up
product to the initial Baeyer–Villiger process [112]. These findings underline the
chemo- and regioselectivity of BVMOs when challenged with multi-ketone precursors
and provide a perspective on the potential of such transformations for synthetic
applications; this particular feature of BVMOs has not been sufficiently exploited, yet.
33.4.6
Application in Bioactive Compound and Natural Product Synthesis
The first exploitation of BVMOs in the synthesis of natural products was outlined
by Taschner. Based on the observation that hydroxyl containing precursors (23)
provide access to ring-contracted lactones (25) upon biotransformation using
CHMOAcineto [168], structurally complex building blocks are available via desym-
metrization reactions. The advanced precursor 25 was elaborated for synthetic
approaches towards tirandamycin [170] and calyculin [171] (Scheme 33.16).
Recent applications in target oriented synthesis started to take advantage of the
presently available BVMO platform to access antipodal lactones (CHMO- and CPMO-
O O
CHMOAcineto O
O OH
O
R R H
25
OH OH
23
24
O O OH
O
O N OMe
H
O N OH NMe2
O
HO
O
O P OH calyculin
O O
OH O
CN
tirandamycin
N MeO
H O
HO
HO
Scheme 33.16 Desymmetrization of a substituted cyclohexanone towards an advanced precursor

of tirandamycin and calyculin.
type enzymes). Prochiral cyclobutanones (27) can be prepared in a straightforward

fashion, applying a [2 þ 2] cycloaddition strategy, and these compounds are highly
reactive substrates for enzymatic Baeyer–Villiger oxygenations due to their ring-
strain (Scheme 33.17). Desymmetrizations by enantiocomplementary BVMOs
enable access to butyrolactones 28 as appealing intermediates for various natural
products and bioactive structures [128]. This includes in particular structurally
diverse lignans of pharmacological relevance [172], drugs like baclofen [173], or
modified amino acids [174]. In most cases, access to the natural as well as the
1470
MeO
(+)-enterolactone O O
O
O O
NH2 O
(+)-hinokinin
Cl MeO
(R)-baclofen COOH
O
O
NH
MeO
H
MeO
HOOC R
(S) & (R)-ß-proline O
MeO
R MeO
O 28a
R Cl2C=C=O
microbial (-)-butyrolactone MeO
OH
BV-Ox
Zn/AcOH (+)-schizandrin
26 27 O H MeO
R 28b
O
O
O R
O
(+)-butyrolactone
MeO O
O MeO O
O
O O MeO
O O (-)-steganacine (R=OAc)
O (-)-steganol (R=OH)
(-)-deoxyisopodophyllotoxin
MeO OMe
OMe
MeO OMe
OMe (-)-trans-burseran
Scheme 33.17 Enantiocomplementary access to chiral butyrolactones via BVMO-mediated

desymmetrization as key intermediates for diverse target structures.
antipodal form of the various target structures is possible with the current set of
BVMOs.
Desymmetrization of bridged bicycloketones (29) also allowed for the preparation
of antipodal lactones suitable for subsequent conversion into natural products within
the indole alkaloid group. Formal total syntheses were outlined that aimed at allo-
yohimbane by exploiting ()-lactone 30, which is accessible via bio-oxidation with
CHMOBrachy in acceptable optical purity, as well as towards antirhine via the
antipodal (þ)-metabolite, which is obtained from CPMOComa-mediated biotransfor-
mations in excellent stereospecificity (Scheme 33.18a) [175].
The combination of biocatalysis with additional sustainable strategies enables
novel types of functional interconversion and opens up novel routes towards
difficult to access structures such as bicyclo[4.2.0]octanes. The metal assisted
H
O H
(a) CHMOBrachy N
85% e.e. O N
H H
H H (-)-30
(-)-allo-yohimbine H
O
29 H
H O H
N
CPMOComa N
O H H
99% e.e.
H
(+)-30 (-)-antirhine H
OH
H
(b) CHMOBrevi1 O
96% e.e.
O
H H
(-)-33
i) hν/Cu2+
OPG O
ii) chem.
oxidation HO
H H H
31 32 CPMOComa O
86% e.e.
O
H H
(+)-33 MeOOC 34
Scheme 33.18 Synthetic exploitation of enantiocomplementary lactones obtained via BVMO-

mediated desymmetrization reactions.
[2 þ 2] photocycloaddition of terminal olefins was significantly improved by rigid-

ifying the system upon incorporation of a bridge (31), hence moving the two alkenes
in immediate proximity for cyclization. Decorating the bridge with a keto func-
tionality allowed, in particular, for subsequent cleavage using a Baeyer–Villiger
process. Microbial oxygenation towards antipodal lactones (33) could be conducted
by applying CHMO- and CPMO-type enzymes [176] (Scheme 33.18b). The com-
bined photochemical/biocatalytic route provided full control over up to six stereo-
genic centers within various structural analogs [137] and the target bicyclo[4.2.0]
octane system 34 was obtained after chemical lactone hydrolysis.
Heteroatom-containing ketones are interesting substrates due to the paramount
role of heterocycles in medicinal chemistry. Bridged hetero-ketone systems are
particularly attractive compounds due to the structural complexity of the scaffolds,
allowing for exquisite stereocontrol in subsequent functional interconversions.
Based on the chemoselectivity of BVMOs desymmetrization of such precursors
opens up novel synthetic entries towards various target compounds. This approach
has been exemplified using oxygen-containing unsaturated bicycloketone 3, which
is available via a facile [4 þ 3] cycloaddition protocol utilizing sonochemistry [177].
CPMOComa-mediated oxygenation provided access to unsaturated bicyclo-lactone 4
in high optical purity. This advanced building block was synthetically elaborated
towards showdomycin as a prototype structure for the class of C-nucleoside
1472
antibiotics concomitantly establishing the absolute configuration of the novel bio-

oxidation product. Alternative synthetic exploitation of the residual functionalities
(alkene, lactone) opened up access to tetrahydrofuran natural products like kumau-
syne as well as goniofufurone analogs (Scheme 33.19) [178].
O NH
HO
O
HO OH
(+)-showdomycin
O O Br
O O
CPMOComa
O 1S
O 6S
95% e.e.
OAc
3 4 (+)-trans-kumausyne
Ph
H
O
HO
O O
HO H
goniofufurone
analogs
Scheme 33.19 Synthetic exploitation of heterocyclic lactones obtained via BVMO-mediated

desymmetrization reactions.
Kinetic resolutions have also been utilized for the production of pharmacologically
relevant compounds. Baeyer–Villiger oxygenation of functionalized racemic 35
provided chiral lactone 36, which was subsequently converted into (R)-(þ)-lipoic
acid as bioactive compound for the treatment of hepatitis, pancreatitis, and induced
carcinomas (Scheme 33.20a) [179]. Synthetic access to both enantiomers of a
pheromone from the oriental hornet Vespa orientalis was established by kinetic
resolution of racemic 37 to give (S)-lactone 38 as the natural product; the antipodal
lactone was obtained via chemical oxidation of the optically enriched ketone
(Scheme 33.20b) [180].
Regiodivergent Baeyer–Villiger oxygenations were utilized to obtain access to
several critical intermediates in bioactive compound synthesis. Within the carbo-
cyclic series, both regioisomeric products were exploited (Scheme 33.21): The
normal bio-oxygenation metabolite 40 represents an advanced entry point into
the synthesis of prostaglandins and structural analogs [181]. The abnormal
lactone 41 was utilized in the total synthesis of a series of brown algae phero-
mones [182] as well as in the preparation of the potent cytostatic sarkomycin [183].
In a similar fashion, regioisomeric lactones from norbornane-type precursors were
also utilized in the synthesis of pharmacological products such as carbocyclic
nucleosides [184].
(a)
OAc O
O COOH
CPMOComa O
S S
35 OAc
36
lipoic acid
(b)
O
O
CHMOAcineto
C11H23 O Vespa orientalis
(S) pheromone
C H
37 38 11 23
25%, 74%e.e.
Scheme 33.20 BVMO-mediated kinetic resolutions towards lipoic acid and a pheromone.
HO
H
O R
O
R'
O H 40
BVMO HO prostaglandins
OH
+
O
H
39
O viridene (R = CH3)
multifidene (R = =CH2)
H 41 R
COOH
sarkomycin
O
Scheme 33.21 Regiodivergent Baeyer–Villiger bio-oxygenations in natural product synthesis.
Oxygen-containing fused cyclobutanone 42 served as starting material for

the synthesis of the natural product clerodin. The normal lactone 43 can be
further elaborated towards the heterocyclic part of this insect antifeedant
(Scheme 33.22a) [185]. While oxygen-containing heterocycles have been demon-
strated as facile substrates for BVMOs, nitrogen analogs are more troublesome due to
their significantly increased polarity. Hence, it is noteworthy that the Cbz-protected
ketone 45 was successfully converted by CHMOAcineto in a regiodivergent biotrans-
formation [186]. The normal biotransformation product 46 represents a modified
Geisman–Waiss lactone, which is a critical intermediate for the synthesis of (þ)-
retronecine as prototype compound for necine-type pyrrolizidine alkaloids
(Scheme 33.22b).
1474
H O
O O H H
(a) O
O
O H 43 H
O BVMO
+ H
O
H
42 O
clerodin
O O OAc
OAc
H 44
OH
H HO H
O
(b) O
O N N
CHMOAcineto H 46
Cbz retronecine
N
+
O
45 H
Cbz
O
N
H 47
Cbz
Scheme 33.22 Regiodivergent Baeyer–Villiger bio-oxygenations of heterocycles in natural product

synthesis.
33.5
Enzyme Engineering
BVMOs were adapted by nature to facilitate specific transformations in various

microorganisms. Owing to their substrate promiscuity these enzymes have become
interesting catalysts for application in organic synthesis and asymmetric chemistry.
With the substantial expansion of the BVMO platform available to date a large range
of structurally diverse substrates can be converted into highly valuable chiral building
blocks. However, certain limitations are often met when searching for the best
biocatalyst for a given reaction. One approach to overcome this limitation is the
continuous search for novel BVMOs within the increasing number of genomes
becoming available constantly. In this context, the sub-classification of BVMOs based
on sequence comparison and prototype substrate profiles is a valuable tool to narrow
down the number of enzyme candidates for in-depth characterization.
A complementary approach is based on the modification of enzymatic properties
by mutagenesis methodology – a well-accepted and often-used approach in the field
of biocatalysis. By this approach, substrate specificities can be optimized and
additional key parameters of biocatalysts (e.g., thermostability, solvent tolerance,
salinity, pH, etc.) can be improved. Alteration of the biocatalyst performance can be
33.5 Enzyme Engineering j1475
knowledge-based, taking advantage of the available structure models and/or sub-
strate profiles of well-characterized BVMOs, as well as randomized. The latter
strategy has been employed successfully in biocatalysis in general and Baeyer–
Villiger bio-oxygenations in particular using the iterative protocol of directed evo-
lution [187, 188]. In such a high-diversity approach it is mandatory to develop an
efficient screening methodology to identify improved biocatalysts [189]. Both
approaches have their advantages and limitations: The randomized strategy turned
out to be particularly successful in fine-tuning already existing substrate acceptance,
for example, improving stereospecificity. This is in part a result of the lower
mutagenesis frequency close to the active site rather than at more remote loca-
tions [190]. Such distant mutations display typically a quite limited effect on the
substrate profile of an enzyme. Hence, modifications targeting the active site directly
can be conducted using site-specific mutations. Such a targeted mutagenesis of first
shell residues has been shown to result in dramatic changes in substrate specificity
and/or enantioselectivity [191]. However, detailed knowledge of the actual compo-
sition of the substrate binding site is required. To circumvent this significant
limitation, several protocols have been proposed leading to a site-restricted but
concomitantly more randomized modification of amino acids to enable rapid and
un-biased identification of the most beneficial alterations. In this context, the
complete active site saturation test (CAST) is a particularly useful tool and has also
been applied to adapt the design of BVMOs [192].
The first modification of a BVMO took advantage of an error-prone mutagenesis
approach aimed at improving the stereospecificity by directed evolution. In a case-
study, 4-hydroxycyclohexanone was used to improve the stereoselectivity of
CHMOAcineto, as the wild-type enzyme gives an almost racemic lactone (Table 33.8).
Already during the first round of mutagenesis a specific hot-spot was identified at
position 432 [193]: When replacing the original phenylalanine by a more lipophilic
amino acid (e.g., leucine) the optical purity of the ()-lactone could be improved;
changing to a polar group at this site (e.g., serine) led to an inversion in stereo-
preference. Saturation mutagenesis at this position provided access to mutants that
also display formation of antipodal lactones for other cyclohexanone precursors. In
addition, mutants from this library also showed expanded or improved substrate
acceptance as well as enantiocomplementary lactone formation on structurally
diverse ketones [194].
Biocatalyst re-design using the CASTing approach was exemplified with
CPMOComa [195]. Careful analysis of an enzyme model based on the crystal
structure of PAMO suggested several boxes for site restricted amino acid
exchange. Within this study the advantage of this approach was clearly demon-
strated, as the methodology led to interesting results with a comparably low
number of mutation candidates. Again, improvements in stereospecificity were
achieved and antipodal lactones could be accessed in certain cases (Table 33.8).
While PAMO possesses a comparably narrow substrate profile it is a remarkably
robust enzyme. It is therefore logical to use the PAMO scaffold to maintain the high
stability of the biocatalyst while extending the substrate scope by modifications close
to the active site. With the crystal structure of PAMO available, structure-inspired
1476
Table 33.8 Mutation studies on BVMOs with pronounced effects on biocatalyst performance
(representative examples).
Enzyme Lactone Amino acid Affected Reference

modificationsa) parameter
CHMOAcineto Wild type 10% e.e. (R)b) [193]

F432A, K500R 34% e.e. (R) [193]
O OH
O F432I 49% e.e. (R) [193]
L143F, E292G, 90% e.e. (R) [193]
L435Q, T464A
K78E, F432S 78% e.e. (S) [193]
F432S 79% e.e. (S) [193]
F432P 72% e.e. (R) [193]
O
O R ¼ OMe Wild type 78% e.e. (S) [193]
L143F, E292G, 25% e.e. (R) [193]
L435Q, T464A
R F432S 99% e.e. (S) [193]
R ¼ Cl F432P >95% e.e. (S) [193]
R ¼ Br F432P >95% e.e. (S) [193]
R¼I F432P >95% e.e. (S) [193]
CPMOComa O R ¼ OEt Wild type 35% e.e. (R) [64]

O F156G, L157F 94% e.e. (R) [64]
R ¼ OAll Wild type 52% e.e. (S) [64]
F156G, L157F 85% e.e. (R) [64]
R
PAMOThermo Indole M446G Indole ! indigo [196]
R ¼ Ph Wild type <10%, E ¼ 1.2 (S) [197]

O
DS441-A442 91%, E ¼ 100 (R) [197]
O
R R ¼ Bn Wild-type No conversion [197]
DS441-A442 40%, E > 200 (R) [197]
R ¼ aryl CASTing 441–444 E ¼ 43–>200 [198]
R ¼ aryl/alkyl P440N E ¼ 26–>200 [199]
a) Position and substitution of amino acids in one-letter coding.

b) Absolute configuration in parentheses.
modifications were considered when comparing the spatial composition of the active
site with other BVMOs. Within a study focusing on some site-directed mutations,
position M446G was identified as extending the catalytic repertoire of PAMO to
convert indole into indigo (Table 33.8) [196].
A first remarkable success along this line was achieved, when comparing the active
site model of CHMOAcineto with the structure of PAMO. This revealed a characteristic
bulge close to the position of FAD within PAMO (S441-S444), which is not present in
CHMO-type enzymes and was considered to limit access of sterically demanding
substrates to the active site of PAMO. Several deletion mutants were prepared to
increase space in this region of the biocatalyst. Remarkably, the substrate profile of
these PAMO-mutants could be expanded to aryl containing a-substituted cyclohex-
anones, which are not (well) accepted by wild-type PAMO [197]. This study was later
re-visited by comparing the loop from position 441–444 with a larger set of BVMOs
(Table 33.8). Choosing a greatly reduced amino acid alphabet, another round of
mutations was conducted, which generated a set of PAMO-variants capable of
conducting kinetic resolutions of 2-aryl-cyclohexanones in excellent stereoselectivity
(E > 150) [198].
To further optimize the stable PAMO scaffold towards larger substrate promis-
cuity, two alternative design plans were developed [199]. Docking studies of phenyl-
acetone based on the crystallographic structure of PAMO suggested 17 individual
regions for CASTing, which were investigated but only generated redundant hits
within the already investigated locations. However, significant improvements were
made when considering second sphere residues that are not in apparent direct
contact with the binding pocket close to FAD. In silico studies of this area suggested
P437 and P440 as interesting sites. These amino acids are rather conserved among
BVMOs. However, replacement by less rigid amino acids to possibly increase
structural flexibility during the catalytic cycle served as a working hypothesis.
CASTing of positions 437 and 440 using NNK degeneracy encoding all 20 proteino-
genic amino acids produced two focused libraries. While modifications at P437 led to
inactive biocatalysts, P440 could be replaced by a large number of alternative amino
acids. Several of these mutants were capable of conducting kinetic resolutions with
excellent selectivity (E > 100) and mutants accepted both 2-aryl as well as 2-alkyl
cyclohexanones (Table 33.8). In addition, the regiodivergent oxygenation of fused
cyclobutanones gave approximately equal amounts of normal and abnormal
lactones in high optical purity. Taken together, these results clearly indicate that the
performance of PAMO was successfully shifted towards CHMO-type enzymes, while
maintaining its thermostability.
33.6
Summary and Outlook
Since the early days of Baeyer–Villiger bio-oxidations using a handful of crude

biocatalysts originating from wild-type organisms with difficult to control enzyme
production, the application of BVMOs in synthetic chemistry has matured to a
platform technology. During recent years, a highly complementary collection of
BVMOs has become available for facile and mild oxygenation reactions on a large
number of structurally highly diverse precursors. Access to both enantiomers of a
lactone in optically pure form can be achieved and good synthetic yields using
purified enzymes or recombinant whole-cell systems are usually obtained. Kinetic
resolution and regiodivergent bio-oxidation of racemic ketones enable access to
1478
lactones difficult or impossible to obtain by conventional chemical methods in

similar efforts. Although versatile, BVMOs are also highly chemoselective and
tolerate the presence of functional groups incompatible with conditions of chemical
oxidations.
Access to novel BVMOs has become straightforward based on fingerprint
sequences and phylogenetic relationship to allow for the identification of new
enzymes within the constantly growing number of available genomes. In addition
to exploiting natural diversity, several approaches have successfully provided engi-
neered BVMOs with improved properties, especially with respect to selectivity and
stability. Together with progress in structure determination of BVMOs, a refined
picture of the dynamics within the catalytic process, and the identification of general
substrate patterns, we start to understand the architecture of this enzyme group. This
will facilitate future re-design of BVMOs to suit particular purposes in synthesis and
vice versa.
Scale-up of BVMO-mediated biotransformations has been demonstrated to
pilot-plant capacities, in part utilizing special techniques such as two-phase and
solid-phase fermentation. As this biotransformation commences under very
mild reaction conditions compared to chemical oxidation processes, the enzy-
matic Baeyer–Villiger reaction also offers the prospect of safe operation under
industrial conditions. Nevertheless, this is still a developing field and several
operational parameters (whole-cell vs. isolated enzyme, enzyme stability,
improved efficiency, inhibition, cofactor dependence) have to be addressed and
improved.
BVMO-mediated transformations enable access to a multitude of versatile build-
ing blocks for asymmetric synthesis. The number of case studies in the literature
utilizing these enzymes in natural product and bioactive compound synthesis is
constantly increasing. BVMOs offer efficient and facile access to complex scaffolds,
incorporating several stereogenic centers in a de novo fashion within a single bio-
oxygenation process. Still, BVMOs have not received general acceptance as
reagents among the community of synthetic chemists and further improvements
in the simplicity of application are required. The design of self-sufficient two-in-one
fusion biocatalysts combining BVMO activity and cofactor regeneration in a single
polypeptide chain may facilitate future application of these catalytic entities to
ultimately establish enzyme-mediated Baeyer–Villiger oxidation as an attractive
process in asymmetric organic synthesis from laboratory-scale to industrial
production.
References
1 Baeyer, A. and Villiger, V. (1899) Chem. 4 Mihovilovic, M.D., Rudroff, F., and
Ber., 32, 3625–3633. Gr€otzl, B. (2004) Curr. Org. Chem., 8,
2 Krow, G.R. (1993) Org. React., 43, 1057–1069.
251–798. 5 Turfitt,G.E.(1948)Biochem.J.,42,376–383.
3 Renz, M. and Meunier, B. (1999) Eur. J. 6 Fried, J., Thoma, R.W., and Klingsberg, A.
Org. Chem., 737–750. (1953) J. Am. Chem. Soc., 75, 5764–5765.
References j1479
7 Peterson, D.H., Eppstein, S.H., Meister, Biocatalytic Scope of Baeyer-Villiger
P.D., Murray, H.C., Leigh, H.M., Monooxygenases in Modern Biooxidation.
Weintraub, A., and Reinecke, L.M. (1953) Enzymes, Reactions and Applications (eds
J. Am. Chem. Soc., 75, 5768–5769. R.D. Schmid and V.B. Urlacher), Wiley-
8 Townsend, C.A., Christensen, S.B., and VCH Verlag GmbH, Weinheim, pp.
Davis, S.G. (1982) J. Am. Chem. Soc., 104, 77–97.
6154–6155. 21 Donoghue, N.A. and Trudgill, P.W.
9 Wright, J.L.C., Hu, T., McLachlan, J.L., (1975) Eur. J. Biochem., 60, 1–7.
Needham, J., and Walter, J.A. (1996) 22 Iwaki, H., Hasegawa, Y., Teraoka, M.,
J. Am. Chem. Soc., 118, 8757–8758. Tokuyama, T., Bergeron, H., and Lau,
10 Damtoft, S., Franzyk, H., and Jensen, P.C.K. (1999) Appl. Environ. Microbiol.,
S.R. (1995) Phytochemistry, 40, 773–784. 65, 5158–5162.
11 Prairie, R.L. and Talalay, P. (1963) 23 Cheng, Q., Thomas, S.M., Kostichka, K.,
Biochemistry, 2, 203–208. Valentine, J.R., and Nagarajan, V. (2000)
12 Forney, F.W. and Markovetz, A.J. (1969) J. Bacteriol., 182, 4744–4751.
Biochem. Biophys. Res. Commun., 37, 24 (a) Brzostowicz, P., Walters, D.M.,
31–38. Thomas, S.M., Nagarajan, V., and
13 Conrad, H.E., DuBus, R., Namtvedt, M.J., Rouviere, P.E. (2003) Appl. Environ.
and Gunsalus, I.C. (1965) J. Biol. Chem., Microbiol., 69, 334–342; (b) Brzostowicz,
240, 495–503. P.C., Gibson, K.L., Thomas, S.M.,
14 Willetts, A. (1997) Trends Biotechnol., 15, Blasko, M.S., and Rouviere, P.E. (2000)
55–62. J. Bacteriol., 182, 4241–4248.
15 van Berkel, W.J.H., Kamerbeek, N.M., 25 Wright, J.L.C., Hu, T., MacLachlan, J.L.,
and Fraaije, M.W. (2006) J. Biotechnol., Needham, J., and Walter, J.A. (1996)
124, 670–689. J. Am. Chem. Soc., 118, 8758–8759.
16 Fraaije, M.W., Kamerbeek, N.M., 26 Cheng, Q., Thomas, S.M., and
van Berkel, W.J.H., and Janssen, D.B. Rouviere, P. (2002) Appl. Microbiol.
(2002) FEBS Lett., 518, 43–47. Biotechnol., 58, 238–242.
17 Van Beilen, J.B., Mourlane, F., 27 Ryerson, C.C., Ballou, D.P., and
Seeger, M.A., Kovac, J., Li, Z., Smits, Walsh, C. (1982) Biochemistry, 21,
T.H.M., Fritsche, U., and Witholt, B. 2644–2655.
(2003) Environ. Microbiol., 5, 174–182. 28 Sheng, D., Ballou, D.P., and Massey, V.
18 Stewart, J.D., Reed, K.W., and (2001) Biochemistry, 40, 11156–11167.
Kayser, M.M. (1996) J. Chem. Soc., Perkin 29 Torres-Pazmino, D.E., Baas, B.-J.,
Trans. 1, 755–757. Janssen, D.B., and Fraaije, M.W. (2008)
19 (a) Kayser, M.M. (2009) Tetrahedron, 65, Biochemistry, 47, 4082–4093.
947–974; (b) Mihovilovic, M.D. (2006) 30 Criegee, R. (1948) Liebigs Ann. Chem.,
Curr. Opin. Chem., 10, 1265–1287; (c) 560, 127–141.
Kamerbeek, N.M., Janssen, D.B., van 31 Malito, E., Alfieri, A., Fraaije, M.W., and
Berkel, J.H., and Fraaije, M.W. (2003) Mattevi, A. (2004) Proc. Nat. Acad. Sci.
Adv. Synth. Catal., 345, 667–678; (d) USA, 101, 13157–13162.
Mihovilovic, M.D., M€ uller, B. and 32 Fraaije, M.W., Kamerbeek, N.M.,
Stanetty, P. (2002) Eur. J. Org. Chem., Heidekamp, A.J., Fortin, R., and
3711–3730. Janssen, D.B. (2004) J. Biol. Chem., 279,
20 (a) Alphand, V.A., Fraaije, M.W., 3354–3360.
Mihovilovic, M.D., and Ottolina, G. 33 Mirza, I.A.M., Yachnin, B.J., Wang, S.,
(2009) Second Generation Baeyer-Villiger Grosse, S., Bergeron, H., Imura, A.,
Biocatalysts in Modern Biocatalysis: Iwaki, H., Hasegawa, Y., Lau, P.C.K., and
Stereoselective and Environmentally Berghuis, A.M. (2009) J. Am. Chem. Soc.,
Friendly Reactions (eds W.D. Fessner and 131, 8848–8855.
T. Anthonsen), Wiley-VCH Verlag 34 de Gonzalo, G., Ottolina, G.,
GmbH, Weinheim, pp. 339–368; (b) Fraaije, M.W., and Carrea, G. (2005)
Fraaije, M.W. and Janssen, D.B. (2007) Chem. Commun., 3724–3726.
1480
35 Alphand, V., Carrea, G., Wohlgemuth, R., 48 (a) Metcalf, W.W. and Wolfe, R.S. (1998)
Furstoss, R., and Woodley, J.M. (2003) J. Bacteriol., 180, 5547–5558;
Trends Biotechnol., 21, 318–323. (b) Garcia Costas, A.M., White, A.K.,
36 Doig, S.D., OSullivan, L.M., Patel, S., and Metcalf, W.W. (2001) J. Biol. Chem.,
Ward, J.M., and Woodley, J.M. (2001) 276, 17429–17436; (c) Vrtis, J.M.,
Enzyme Microb. Technol., 28, 265–274. White, A.K., Metcalf, W.W., and
37 Cheesman, M.J., Kneller, B., Kelly, E.J., van der Donk, W.A. (2002) Angew. Chem.,
Thompson, S.J., Yeung, C.K., Eaton, D.L., 114, 3391–3393; (2002) Angew. Chem. Int.
and Rettie, A.E. (2001) Protein Expression Ed., 41, 3257–3259.
Purif., 21, 81–86. 49 Woodyer, R.D., van der Donk, W.A., and
38 Schwarz-Linek, U., Kr€odel, A., Zhao, H. (2003) Biochemistry, 42,
Ludwig, F.-A., Schulze, A., Rissom, S., 11604–11614.
Kragl, U., Tishkov, V.I., and Vogel, M. 50 Torres Pazmino, D.E., Snajdrova, R.,
(2001) Synthesis, 947–951. Baas, B.-J., Ghobrial, M., Mihovilovic,
39 de Gonzalo, G., Ottolina, G., M.D., and Fraaije, M.W. (2008) Angew.
Zambianchi, F., Fraaije, M.W., and Chem. Int. Ed., 47, 2275–2278.
Carrea, G. (2006) J. Mol. Catal. B.: 51 Pazmino, D.E.T., Riebel, A., de Lange, J.,
Enzym., 39, 91–97. Rudroff, F., Mihovilovic, M.D., and
40 (a) Willetts, A.J., Knowles, C.J., Fraaije, M.W. (2009) ChemBioChem, 10,
Levitt, M.S., Roberts, S.M., Sandey, H., 2595–2598.
and Shipston, N.F. (1991) J. Chem. Soc., 52 Hollmann, F., Hofstetter, K., and
Perkin Trans. 1, 1608–1610; (b) Grogan, Schmid, A. (2006) Trends Biotechnol., 24,
G., Roberts, S. and Willetts, A.J. (1992) 163–171.
Biotechnol. Lett., 14, 1125–1130. 53 de Gonzalo, G., Ottolina, G., Carrea, G.,
41 Secundo, F., Carrea, G., Riva, S., and Fraaije, M.W. (2005) Chem.
Battistel, E., and Bianchi, D. (1993) Commun., 3724–3726.
Biotechnol. Lett., 15, 865–870. 54 Hollmann, F., Taglieber, A., Schulz, F.,
42 Hogan, M.C. and Woodley, J.M. (2000) and Reetz, M.T. (2007) Angew. Chem. Int.
Chem. Eng. Sci., 55, 2001–2008. Ed., 46, 2903–2906.
43 Zambianchi, F., Pasta, P., Carrea, G., 55 Taglieber, A., Schulz, F., Hollmann, F.,
Colonna, S., Gaggero, N., and Rusek, M., and Reetz, M.T. (2008)
Woodley, J.M. (2002) Biotechnol. Bioeng., ChemBioChem, 9, 565–572.
78, 489–496. 56 Rioz-Martinez, A., de Gonzalo, G.,
44 (a) Willetts, A.J., Knowles, C.J., Torres-Pazmino, D.E., Fraaije, M.W., and
Levitt, M.S., Roberts, S.M., Sandey, H., Gotor, V. (2009) Eur. J. Org. Chem.,
and Shipston, N.F. (1991) J. Chem. Soc., 2526–2532.
Perkin Trans. 1, 1608–1610; 57 Rodrigez, C., de Gonzalo, G.,
(b) Grogan, G., Roberts, S., and Willetts, Torres-Pazmino, D.E., Fraaije, M.W., and
A.J. (1992) Biotechnol. Lett., 14, Gotor, V. (2008) Tetrahedron: Asymmetry,
1125–1130. 19, 197–203.
45 Zambianchi, F., Pasta, P., Carrea, G., 58 Mihovilovic, M.D., Snajdrova, R.,
Colonna, S., Gaggero, N., and Winninger, A., and Rudroff, F. (2005)
Woodley, J.M. (2002) Biotechnol. Bioeng., Synlett, 2751–2754.
78, 489–496. 59 Doig, S.D., Simpson, H., Alphand, V.,
46 Atia, K.S. (2005) Radiat. Phys. Chem., 73, Furstoss, R., and Woodley, J.M. (2003)
91–99. Enzyme Microbiol. Technol., 32, 347–355.
47 (a) Narita, J., Okano, K., Tateno, T., 60 Rudroff, F., Alphand, V., Furstoss, R., and
Tanino, T., Sewaki, T., Sung, M.-H., Mihovilovic, M.D. (2006) Org. Process Res.
Fukuda, H., and Kondo, A. (2006) Develop., 10, 599–604.
Appl. Microbiol. Biotechnol., 70, 564–572; 61 (a) Walton, A.Z. and Stewart, J.D. (2002)
(b) Liu, D., Schmid, R.D., and Rusnak, M. Biotechnol. Prog., 18, 262–268; (b) Walton,
(2006) Appl. Microbiol. Biotechnol., 72, A.Z. and Stewart, J.D. (2004) Biotechnol.
1024–1032. Prog., 20, 403–411.
References j1481
62 Simpson, H., Alphand, V., and 76 (a) Lebreton, J., Alphand, V., and
Furstoss, R. (2001) J. Mol. Catal. B: Enz., Furstoss, R. (1996) Tetrahedron Lett., 37,
16, 101–108. 1011–1014; (b) Lebreton, J., Alphand, V.,
63 Law, H.E.M., Baldwin, C.V.F., and Furstoss, R. (1997) Tetrahedron, 53,
Chen, B.H., and Woodley, J.M. (2006) 145–160.
Chem. Eng. Sci., 61, 6646–6652. 77 Quazzani-Chahdi, J., Buisson, D., and
64 Clouthier, C.M. and Kayser, M.M. (2007) Azerad, R. (1987) Tetrahedron Lett., 28,
J. Mol. Catal. B.: Enzym., 46, 32–36. 1109–1112.
65 (a) Hilker, I., Baldwin, C., Alphand, V., 78 (a) Conrad, H.E., DuBus, R.,
Furstoss, R., Woodley, J., and Namtvedt, M.J., and Gunsalus, I.C.
Wohlgemuth, R. (2006) Biotechnol. (1965) J. Biol. Chem., 240, 495–503; (b)
Bioeng., 93, 1138–1144; (b) Baldwin, Taylor, D.G., and Trudgill, P.W. (1986)
C.V.F. and Woodley, J.M. (2006) J. Bacteriol., 165, 489–497.
Biotechnol. Bioeng., 95, 362–369. 79 Ougham, H.J., Taylor, D.G., and
66 Hilker, I., Wohlgemuth, R., Alphand, V., Trudgill, P.W. (1983) J. Bacteriol., 153,
and Furstoss, R. (2005) Biotechnol. 140–152.
Bioeng., 92, 702–710. 80 Bonsor, D., Butz, S.F., Solomons, J.,
67 (a) Doig, S.D., Avenell, P.J., Bird, P.A., Grant, S., Fairlamb, I.J.S., Fogg, M.J., and
Gallati, P., Lander, K.S., Lye, G.J., Grogan, G. (2006) Org. Biomol. Chem., 4,
Wohlgemuth, R., and Woodley, J.M. 1252–1260.
(2002) Biotechnol. Prog., 18, 1039–1046; 81 Snajdrova, R., Grogan, G., and
(b) Baldwin, C.V.F., Wohlgemuth, R., and Mihovilovic, M.D. (2006) Bioorg. Med.
Woodley, J.M. (2008) Org. Proc. Res. Chem. Lett., 16, 4813–4817.
Develop., 12, 660–665. 82 Rehdorf, J., Kirschner, A., and
68 Yang, J., Wang, S., Lorrain, M.-J., Rho, D., Bornscheuer, U.T. (2007) Biotechnol. Lett.,
Abokitse, K., and Lau, P.C.K. (2009) Appl. 29, 1393–1398.
Microbiol. Biotechnol., 84, 867–876. 83 V€olker, A., Kirschner, A.,
69 (a) Hilker, I., Gutierrez, M.C., Furstoss, Bornscheuer, U.T., and Altenbuchner, J.
R., Ward, J., Wohlgemuth, R., and (2008) Appl. Microbiol. Biotechnol., 77,
Alphand, V. (2008) Nat. Protoc., 3, 1251–1260.
546–554; (b) Hilker, I., Alphand, V., 84 Kostichka, K., Thomas, S.M., Gibson,
Wohlgemuth, R., and Furstoss, R. (2004) K.J., Nagarajan, V., and Cheng, Q. (2001)
Adv. Synth. Catal., 346, 203–214; (c) J. Bacteriol., 183, 6478–6486.
Hilker, I., Gutierrez, M.C., Alphand, V., 85 Kyte, B.G., Rouviere, P., Cheng, Q., and
Wohlgemuth, R., and Furstoss, R. (2004) Stewart, J.D. (2004) J. Org. Chem., 69,
Org. Lett., 6, 1955–1958. 12–17.
70 Wohlgemuth, R. (2006) Eng. Life Sci., 6, 86 Donoghue, N.A., Norris, D.B., and
577–583. Trudgill, P.W. (1976) Eur. J. Biochem., 63,
71 Stewart, J.D. (1998) Curr. Org. Chem., 2, 175–192.
195–216. 87 Chen, Y.-C.J., Peoples, O.P., and
72 Gibson, M., Nur-e-alam, M., Lipata, F., Walsh, C.T. (1988) J. Bacteriol., 170,
Oliveira, M.A., and Rohr, J. (2005) J. Am. 781–789.
Chem. Soc., 127, 17594–17595. 88 Brzostowicz, P., Walters, D.M.,
73 Beam, M.P., Bosserman, M.A., Thomas, S.M., Nagarajan, V., and
Noinaj, N., Whenkel, M., and Rohr, J. Rouviere, P.E. (2003) Appl. Environ.
(2009) Biochemistry, 48, Microbiol., 69, 334–342.
4476–4487. 89 Cheng, Q., Thomas, S.M., Kostichka, K.,
74 Jiang, J., Tetzlaff, C.N., Takamatsu, S., Valentine, J.R., and Nagarajan, V. (2000)
Iwatsuki, M., Komatsu, M., Ikeda, H., J. Bacteriol., 182, 4744–4751.
and Cane, D.E. (2009) Biochemistry, 48, 90 Mihovilovic, M.D., Rudroff, F., Gr€otzl, B.,
6431–6440. Kapitan, P., Snajdrova, R., Rydz, J., and
75 Itagaki, E. (1986) J. Biochem., 99, Mach, R. (2005) Angew. Chem. Int. Ed., 44,
815–824. 3609–3613.
1482
91 Bramucci, M.G., Brzostowicz, P.C., Altenbucher, J., and Bornscheuer, U.T.

Kostichka, K.N., Nagarajan, V., Rouviere, (2007) Appl. Microbiol. Biotechnol., 73,
P.E., and Thomas, S.M. (2003) PCT Int. 1065–1072.
Appl. WO 2003020890.(2003) Chem. 105 Szolkowy, C., Eltis, L.D., Bruce, N.C., and
Abstr., 138, 233997. Grogan, G. (2009) ChemBioChem, 10,
92 Brzostowicz, P.C., Gibson, K.L., 1208–1217.
Thomas, S.M., Blasko, M.S., and 106 Malito, E., Alfieri, A., Fraaije, M.W., and
Rouviere, P.E. (2000) J. Bacteriol., 182, Mattevi, A. (2004) Proc. Natl. Acad. Sci.
4241–4248. USA, 101, 13157–13162.
93 Mihovilovic, M.D., Rudroff, F., 107 Fraaije, M.W., Wu, J., Neuts, D.P.H.M.,
M€ uller, B., and Stanetty, P. (2003) Bioorg. van Hellemond, E.W., Spelberg, J.H.L.,
Med. Chem. Lett., 13, 1479–1482. and Janssen, D.B. (2005) Appl. Microbiol.
94 van Beilen, J.B., Mourlane, F., Seeger, Biotechnol., 66, 393–400.
M.A., Kovac, J., Li, Z., Smits, T.H.M., 108 Fraaije, M.W., Kamerbeek, N.M.,
Fritsche, U., and Witholt, B. (2003) Heidekamp, A.J., Fortin, R., and
Environ. Microbiol., 5, 174–182. Janssen, D.B. (2004) J. Biol. Chem., 279,
95 Rial, D.V., Cernuchova, P., 3354–3360.
van Beilen, J.B., and Mihovilovic, M.D. 109 Morii, S., Sawamoto, S., Yamauchi, Y.,
(2008) J. Mol. Catal. B: Enzym., 50, 61–68. Miyamoto, M., Iwami, M., and Itagaki, E.
96 Rial, D.V., Bianchi, D.A., Kapitanova, P., (1999) J. Biochem., 126, 624–631.
Lengar, A., van Beilen, J.B., and 110 Mihovilovic, M.D., Rudroff, F., M€ uller, B.,
Mihovilovic, M.D. (2008) Eur. J. Org. and Stanetty, P. (2003) Bioorg. Med. Chem.
Chem., 1203–1213. Lett., 13, 1479–1482.
97 Iwaki, H., Wang, S., Grosse, S., 111 (a) Mihovilovic, M.D., Kapitan, P. and
Bergeron, H., Nagahashi, A., Kapitanova, P. (2008) ChemSusChem, 1,
Lertvorachon, J., Yang, J., Konishi, Y., 143–148; (b) Mihovilovic, M.D. and
Hasegawa, Y., and Lau, P.C.K. (2006) Kapitan, P. (2004) Tetrahedron Lett., 45,
Appl. Environ. Microbiol., 72, 2751–2754.
2707–2720. 112 Beneventi, E., Ottolina, G., Carrea, G.,
98 Griffin, M. and Trudgill, P.W. (1976) Eur. Panzeri, W., Fronza, G., and Lau, P.C.K.
J. Biochem., 63, 199–209. (2009) J. Mol. Catal. B: Enzym., 58,
99 Iwaki, H., Hasegawa, Y., Lau, P.C.K., 164–168.
Wang, S., and Kayser, M.M. (2002) Appl. 113 Bruice, T.C. (1980) Acc. Chem. Res., 13,
Environ. Microbiol., 68, 5681–5684. 256–262.
100 Kamerbeek, N.M., Moonen, M.J.H., 114 Ghisla, S. and Massey, V. (1989) Eur. J.
van der Ven, J.G.M., van Berkel, W.J.H., Biochem., 181, 1–17.
Fraaije, M.W., and Janssen, D.B. (2001) 115 Colonna, S., Gaggero, N., Richelmi, C.,
Eur. J. Biochem., 268, and Pasta, P. (2000) NATO Sci. Ser. 1, 33,
2547–2557. 133–160.
101 Kamerbeek, N.M., Olsthorn, A.J.J., 116 (a) Light, D.R., Waxman, D.J. and
Fraaije, M.W., and Janssen, D.B. Walsh, C. (1982) Biochemistry, 21,
(2003) Appl. Environ. Microbiol., 69, 2490–2498; (b) Secundo, F., Carrea, G.,
419–426. Dallavalle, S., and Franzosi, G. (1993)
102 Mihovilovic, M.D., Kapitan, P., Rydz, J., Tetrahedron: Asymmetry, 4, 1981–1982; (c)
Rudroff, F., Ogink, F.H., and Fraaije, Beecher, J., Richardson, P., and
M.W. (2005) J. Mol. Catal. B: Enzym., 32, Willetts, A. (1994) Biotechnol. Lett., 16,
135–140. 909–912; (d) Pasta, P., Carrea, G.,
103 Rehdorf, J., Zimmer, C.L., and Holland, H.L., and Dallavalle, S. (1995)
Bornscheuer, U.T. (2009) Appl. Environ. Tetrahedron: Asymmetry, 6, 933–936; (e)
Microbiol., 75, 3106–3114. Kelly, D.R., Knowles, J.C., Mahdi, J.G.,
104 (a) Kirschner, A. and Bornscheuer, U.T. Taylor, I.N., and Wright, M.A. (1996)
(2006) Angew. Chem. Int. Ed., 45, Tetrahedron: Asymmetry, 7, 365–368; (f)
7004–7006; (b) Kirschner, A., Colonna, S., Gaggero, N., Pasta, P., and
References j1483
Ottolina, G. (1996) Chem. Commun., 130 Mihovilovic, M.D., Rudroff, F., Gr€
otzl, B.,
2303–2307; (g) Colonna, S., Gaggero, N., and Stanetty, P. (2005) Eur. J. Org. Chem.,
Carrea, G., and Pasta, P. (1997) Chem. 1479–1482.
Commun., 439–440; (h) Holland, H.L., 131 Stewart, J.D., Reed, K.W., Martinez, C.A.,
Gu, J.-X., Kerridge, A., and Willetts, A. Zhu, J., Chen, G., and Kayser, M.M.
(1999) Biocatal. Biotransform., 17 (1998) J. Am. Chem. Soc., 120,
305–317. 3541–3548.
117 (a) Colonna, S., Gaggero, N., 132 Iwaki, H., Hasegawa, Y., Wang, S.,
Bertinotti, A., Carrea, G., Pasta, P., and Kayser, M.M., and Lau, P.C. (2002) Appl.
Bernardi, A. (1995) J. Chem. Soc., Chem. Environ. Microbiol., 68, 5671–5684.
Commun., 1123–1124; (b) Alphand, V., 133 Mihovilovic, M.D., Chen, G., Wang, S.,
Gaggero, N., Colonna, S., and Furstoss, R. Kyte, B., Rochon, F., Kayser, M.M., and
(1996) Tetrahedron Lett., 37, 6117–6120; Stewart, J.D. (2001) J. Org. Chem., 66,
(c) Alphand, V., Gaggero, N., Colonna, S., 733–738.
Pasta, P., and Furstoss, R. (1997) 134 Wang, S., Kayser, M.M., Iwaki, H., and
Tetrahedron, 53, 9695–9706. Lau, P.C.K. (2003) J. Mol. Catal. B:
118 Colonna, S., Gaggero, N., Carrea, G., and Enzym., 22, 211–218.
Pasta, P. (1998) Chem. Commun., 135 Taschner, M.J. and Peddada, L. (1992)
415–416. J. Chem. Soc., Chem. Commun.,
119 de Gonzalo, G., Ottolina, G., Zambianchi, 1384–1385.
F., Fraaije, M.W., and Carrea, G. (2006) 136 Mihovilovic, M.D., M€ uller, B.,
J. Mol. Catal. B: Enzym., 39, 91–97. Schulze, A., Stanetty, P., and
120 Ottolina, G., Bianchi, S., Belloni, B., Kayser, M.M. (2003) Eur. J. Org. Chem.,
Carrea, G., and Danieli, B. (1999) 2243–2249.
Tetrahedron Lett., 40, 8483–8486. 137 Snajdrova, R., Braun, I., Bach, T.,
121 Latham, J.A. Jr. and Walsh, C. (1986) Mereiter, K., and Mihovilovic, M.D.
J. Chem. Soc., Chem. Commun., 527–528. (2007) J. Org. Chem., 72, 9597–9603.
122 Latham, J.A. Jr., Branchaud, B.P. Jr., 138 (a) Chen, C.-S., Fujimoto, Y.,
Chen, Y.-C.J., and Walsh, C. (1986) Girdaukas, G., and Sih, C.J. (1982) J. Am.
J. Chem. Soc., Chem. Commun., 528–530. Chem. Soc., 104, 7294–7299;
123 Latham, J.A. and Walsh, C. (1987) J. Am. (b) Martin, V.S., Woodard, S.S.,
Chem. Soc., 109, 3421–3427. Katsuki, T., Yamada, Y., Ikeda, M., and
124 Mihovilovic, M.D., M€ uller, B., Sharpless, K.B. (1981) J. Am. Chem. Soc.,
Kayser, M.M., Stewart, J.D., Fr€ohlich, J., 103, 6237–6240; (c) Gredig, G. and
Stanetty, P., and Spreitzer, H. (2001) Fajans, K. (1908) Chem. Ber., 41, 752–763.
J. Mol. Catal. B: Enzym., 11, 349–353. 139 Bes, M.T., Villa, R., Roberts, S.M.,
125 Mihovilovic, M.D., Gr€otzl, B., Wan, P.W.H., and Willetts, A.J. (1996)
Kandioller, W., Muskotal, A., J. Mol. Catal. B: Enzym., 1, 127–134.
Snajdrova, R., Rudroff, F., and 140 Kayser, M.M., Chen, G., and
Spreitzer, H. (2008) Chem. Biodivers., 5, Stewart, J. (1998) J. Org. Chem., 63,
490–498. 7103–7106.
126 Colonna, S., Gaggero, N., Carrea, G., 141 Berezina, N., Alphand, V., and
Ottolina, G., Pasta, P., and Furstoss, R. (2002) Tetrahedron:
Zambianchi, F. (2002) Tetrahedron Lett., Asymmetry, 13, 1953–1955.
43, 1797–1799. 142 Alphand, V., Furstoss, R.,
127 Garcia-Urdiales, E., Alfonso, I., and Pedragosa-Moreau, S., Roberts, S.M., and
Gotor, V. (2005) Chem. Rev., 105, 313–354. Willetts, A.J. (1996) J. Chem. Soc., Perkin
128 Rudroff, F., Rydz, J., Ogink, F., Fink, M., Trans. 1, 1867–1872.
and Mihovilovic, M.D. (2007) Adv. Synth. 143 Stewart, J.D., Kieth, W.R., Zhu, J., Chen,
Catal., 349, 1436–1444. G., and Kayser, M.M. (1996) J. Org. Chem.,
129 Taschner, M.J., Black, D.J., and Chen, Q.- 61, 7652–7653.
Z. (1993) Tetrahedron: Asymmetry, 4, 144 Schwarz-Linek, U., Kr€odel, A.,
1387–1390. Ludwig, F.-A., Schulze, A., Rissom, S.,
1484
Kragl, U., Tishkov, V.I., and Vogel, M. 158 Moonen, M.J.H., Westphal, A.H.,
(2001) Synthesis, 947–951. Rietjens, I.M.C.M., and van Berkel,
145 Adger, B., Bes, M.T., Grogan, G., W.J.H. (2005) Adv. Synth. Catal., 347,
McCague, R., Podragosa-Moreau, S., 1027–1034.
Roberts, S.M., Villa, R., Wan, P.W.H., and 159 Berezina, H., Kozma, E., Furstoss, R., and
Willetts, A.J. (1997) Bioorg. Med. Chem., 5, Alphand, V. (2007) Adv. Synth. Catal., 349,
253–261. 2049–2053.
146 (a) Cripps, R.E. (1975) Biochem. J., 152, 160 Cernuchova, P. and Mihovilovic, M.D.
233–241.(b) Cripps, R.E., Trudgill, P.W., (2007) Org. Biomol. Chem., 5,
and Whateley, J.G. (1978) Eur. J. Biochem, 1715–1719.
86 175–186; (c) Higson, F.K. and Focht, 161 Alphand, V. and Furstoss, R. (1992) J. Org.
D.D. (1990) Appl. Environ. Microbiol, 56, Chem., 57, 1306–1309.
3678–3685. 162 Kelly, D.R., Knowles, C.J., Mahdi, J.G.,
147 (a) Forney, F.W., Markovetz, A.J., and Taylor, I.N., and Wright, M.A. (1995) J.
Kallio, R.E. (1967) J. Bacteriol., 93, Chem. Soc., Chem. Commun., 729–730.
649–655; (b) Britton, L.N., Brand, J.M., 163 Kelly, D.R., Knowles, C.J., Mahdi, J.G.,
and Markovetz, A.J. (1974) Biochim. Wright, M.A., Taylor, I.N., Hibbs, D.E.,
Biophys. Acta, 369, 45–49; (c) Hartmans, Hursthouse, M.B., Mishal, A.K.,
S. and DeBont, J.A.M. (1986) FEMS Roberts, S.M., Wan, P.W.H., Grogan, G.,
Microbiol. Lett., 36, 155–158. and Willets, A.J. (1995) J. Chem. Soc.,
148 Geitner, K., Kirschner, A., Rehdorf, J., Perkin Trans. 1, 2057–2066.
Schmidt, M., Mihovilovic, M.D., and 164 (a) Sandy, H. and Willetts, A.J. (1989)
Bornscheuer, U.T. (2007) Tetrahedron: Biotechnol. Lett., 2, 615–620; (b) Bes, M.T.,
Asymmetry, 18, 892–895. Villa, R., Roberts, S.M., Wan, P.W.H., and
149 Rodriguez, C., de Gonzalo, G., Willetts, A.J. (1996) J. Mol. Catal. B:
Fraaije, M.W., and Gotor, V. (2007) Enzym., 1, 127–134.
Tetrahedron: Asymmetry, 18, 1338–1344. 165 Wang, S., Kayser, M.M., and Jurkauskas,
150 Rodriguez, C., de Gonzalo, G., Torres V. (2003) J. Org. Chem., 68, 6222–6228.
Pazmino, D.E., Fraaije, M.W., and Gotor, 166 Bradshaw, W.H., Conrad, H.E., Corey,
V. (2008) Tetrahedron: Asymmetry, 19, E.J., Gunsalus, I.C., and Lednicer, D.
197–203. (1959) J. Am. Chem. Soc., 81, 5507.
151 Rehdorf, J., Lengar, A., Bornscheuer, 167 (a) Ougham, H.J., Taylor, D.G. and
U.T., and Mihovilovic, M.D. (2009) Trudgill, P.W. (1983) J. Bacteriol., 153,
Bioorg. Med. Chem. Lett., 19, 3739–3743. 140–152; (b) Taylor, D.G. and
152 Berezina, N., Alphand, V., and Trudgill, P.W. (1986) J. Bacteriol., 165,
Furstoss, R. (2002) Tetrahedron: 489–497; (c) Jones, K.H., Smith, R.T., and
Asymmetry, 13, 1953–1955. Trudgill, P.W. (1993) J. Gen. Microbiol.,
153 Gutierrez, M.C., Furstoss, R., and 139, 797–805.
Alphand, V. (2005) Adv. Synth. Catal., 347, 168 Taschner, M.J. and Black, D.J. (1988) J.
1051–1059. Am. Chem. Soc., 110, 6892–6893.
154 Noyori, R., Sato, T., and Kobayashi, H. 169 Ottolina, G., de Gonzalo, G., Carrea, G.,
(1983)Bull.Chem. Soc. Jpn.,56, 2661–2679. and Danieli, B. (2005) Adv. Synth. Catal.,
155 (a) Schwab, J.M. (1981) J. Am. Chem. Soc., 347, 1035–1040.
103, 1876–1878; (b) Schwab, J.M., Li, W.- 170 Taschner, M.J. and Aminbhavi, A.S.
B. and Thomas, L.P. (1983) J. Am. Chem. (1989) Tetrahedron Lett., 30, 1029–1032.
Soc., 105, 4800–4808. 171 (a) Yokokawa, F., Hamada, Y., and Shioiri,
156 Rioz-Martinez, A., de Gonzalo, G., T. (1996) Chem. Commun., 871–872; (b)
Torres Pazmino, D.E., Fraaije, M.W., and Duchamp, D.J., Branfman, A.R., Button,
Gotor, V. (2009) Eur. J. Org. Chem., A.C., and Rinehart, K.L. (1973) J. Am.
2526–2532. Chem. Soc., 95, 4077–4078.
157 Ludwig, T., Ermert, J., and 172 Bode, J.W., Doyle, M.P., Protopopova,
Coenen, H.H. (2002) Nucl. Med. Biol., 29, M.N., and Zhou, Q.-L. (1996) J. Org.
255–262. Chem., 61, 9146–9155.
References j1485
173 Mazzini, C., Lebreton, J., Alphand, V., and 185 Petit, F. and Furstoss, R. (1995) Synthesis,
Furstoss, R. (1997) Tetrahedron Lett., 38, 1517–1520.
1195–1196. 186 (a) Luna, A., Gutierrez, M.C., Furstoss, R.,
174 Mazzini, C., Lebreton, J., Alphand, V., and and Alphand, V. (2005) Tetrahedron:
Furstoss, R. (1997) J. Org. Chem., 62, Asymmetry, 16, 2521; (b) Petit, F. and
5215–5218. Furstoss, R. (1993) Tetrahedron:
175 Mihovilovic, M.D., M€ uller, B., Asymmetry, 4, 1341.
Kayser, M.M., and Stanetty, P. (2002) 187 Arnold, F.H. (2001) Nature, 409, 253–257.
Synlett, 700–702. 188 Reetz, M.T. (2009) J. Org. Chem., 74,
176 Braun, I., Rudroff, F., Mihovilovic, M.D., 5767–5778.
and Bach, T. (2006) Angew. Chem. Int. Ed., 189 Reetz, M.T., Becker, M.H., Klein, H.-W.,
45, 5541–5543. and St€ockigt, D. (1999) Angew. Chem.,
177 Mihovilovic, M.D., Gr€otzl, B., Kandioller, 111, 1872–1875.
W., Snajdrova, R., Muskotal, A., Bianchi, 190 Morley, K.L. and Kazlauskas, R.J. (2005)
D.A., and Stanetty, P. (2006) Adv. Synth. Trends Biotechnol., 23, 231–237.
Catal., 348, 463–470. 191 van den Heuvel, R.H.H., Fraaije, M.W.,
178 Mihovilovic, M.D., Bianchi, D.A., and Ferrer, M., Mattevi, A., and van Berkel,
Rudroff, F. (2006) Chem. Commun., W.J.H. (2000) Proc. Natl. Acad. Sci. USA,
3214–3216. 97, 9455–9460.
179 Adger, B.M., McCague, R., and 192 Reetz, M.T., Bocola, M., Carballeira, J.D.,
Roberts, S.M. (1996) Int. Pat. Zha, D., and Vogel, A. (2005) Angew.
Appl. WO 96/38437;Chem. Abstr. 126 Chem. Int. Ed., 44, 4192–4196.
(1997) 89208. 193 Reetz, M.T., Brunner, B., Schneider, T.,
180 Alphand, V., Archelas, A., and Furstoss, Schulz, F., Clouthier, C.M., and Kayser,
R. (1990) J. Org. Chem., 55, 347–350. M.M. (2004) Angew. Chem., 116,
181 Alphand, V., Archelas, A., and 4167–4170.
Furstoss, R. (1989) Tetrahedron Lett., 30, 194 Mihovilovic, M.D., Rudroff, F.,
3663–3664. Winninger, A., Schneider, T., Schulz, F.,
182 (a) Lebreton, J., Alphand, V., and and Reetz, M.T. (2006) Org. Lett., 8,
Furstoss, R. (1996) Tetrahedron Lett., 37, 1221–1224.
1011–1014; (b) Lebreton, J., Alphand, V., 195 Clouthier, C.M., Kayser, M.M., and Reetz,
and Furstoss, R. (1997) Tetrahedron, 53, M.T. (2006) J. Org. Chem., 71, 8431–8437.
145–160. 196 Torrez Pazmino, D.E., Snajdrova, R., Rial,
183 Andrau, L., Lebreton, J., Viazzo, P., D.V., Mihovilovic, M.D., and Fraaije,
Alphand, V., and Furstoss, R. (1997) M.W. (2007) Adv. Synth. Catal., 349,
Tetrahedron Lett., 38, 825–826. 1361–1368.
184 (a) Gagnon, R., Grogan, G., Levitt, M.S., 197 Bocola, M., Schulz, F., Leca, F., Vogel, A.,
Roberts, S.M., Wan, P.W.H., and Willetts, Fraaije, M.W., and Reetz, M.T. (2005)
A.J. (1994) J. Chem. Soc., Perkin Trans. 1, Angew. Chem., 347, 979–986.
2537–2543; (b) Gagnon, R., Grogan, G., 198 Reetz, M.T. and Wu, S. (2008) Chem.
Roberts, S.M., Villa, R., and Willetts, A.J. Commun., 5499–5501.
(1995) J. Chem. Soc., Perkin Trans. 1, 199 Reetz, M.T. and Wu, S. (2009) J. Am.
1505–1511. Chem. Soc., 131, 15424–15432.
j1487
34
Aromatic Oxidations
David J. Leak, Ying Yin, Jun-Jie Zhang, and Ning-Yi Zhou
34.1
Enzymology of Aromatic Hydrocarbon Oxidation
34.1.1
Metabolism of Aromatic Compounds
A bewildering number of different enzymes are able to oxidize aromatic compounds

(arenes), either as part of their natural function or fortuitously (co-metabolism). The
ability to change the specificity of biocatalysts through protein engineering intro-
duces an additional factor into the pool of possible variants to choose from. Therefore,
to be useful for applied biocatalysis it is important to establish not only the
possibilities but also the boundaries/limitations of different enzymes, to demystify
the field for new entrants.
The microbial degradation of aromatics has been extensively studied and well-
defined biodegradation pathways for most commonly encountered aromatic com-
pounds have been elucidated [1]. The wide phylogenetic diversity of microbes capable
of degradation of aromatic compounds provides multiple enzyme/gene variants for
each catabolic step and variations in initial catabolic pathways. For example, five
different biochemical pathways have been characterized for aerobic toluene degra-
dation [2]. Information on the degradation pathways, associated enzymes, and
genetic mechanism is available in databases such as the Kyoto Encyclopedia of
Genes and Genomes [3], MetaCyc [4], and University of Minnesota Biocatalysis/
Biodegradation Database (UM-BBD) [5].
Although there is clear evidence for arene oxidation under anaerobic conditions [6],
activities are low and we are a long way from being able to exploit this in biocatalysis.
Therefore, this chapter will focus on the aerobic reactions that involve molecular oxygen
as the oxidizing species and in which the reaction products are under enzymatic control.
This excludes laccases, which react with phenolics to generate phenoxy radicals, and
peroxidases, which initially form radical cations, both of which subsequently react in
various ways that are not under strict enzymatic control. With the exception of reactions
with free radicals, oxygen requires activation to take part in enzymatic reactions. This
and Oliver May.
j 34 Aromatic Oxidations
1488
has several consequences. First, to reduce oxygen to a sufficiently reactive state requires
an electron donor, which in many instances is NAD(P)H but also includes metabolic
intermediates such as a-keto glutarate. Second, oxygen has several reactive oxidation
states (superoxide, peroxide, hydroxyl radical). Each has a defined reactivity range that
may be modulated by interaction with enzyme cofactors such as flavin or metal ions.
Nature generally exploits this variation in reactivity to ensure that the reactive oxygen
species produced is just sufficient for the type of reaction required. This is a useful
guideline in selecting a potential biocatalyst. For instance, methane monooxygenase
needs to produce an active oxygen species able to break a single CH bond in methane
(435 kJ mol1). It is unsurprising to find that it catalyzes a range of aromatic hydro-
carbon hydroxylations. Although it may find application in certain simple aromatic
oxidations, the biotransformation of a complex molecule with both aromatic and non-
aromatic components will probably lead to mixed products from hydroxylation (and
epoxidation) at several sites. Choosing a less reactive enzyme such as toluene mono-
oxygenase should ensure that no products from oxidation of unactivated CH bonds
are produced (although it is capable of epoxidation and benzylic hydroxylation [7]).
Third, understanding the reactive oxygen species necessary for substrate oxidation
opens up the possibility of providing this directly, as an alternative to oxygen. Much of
the complexity of oxygenases arises from the need for controlled reduction of oxygen
and integration with the catalytic cycle. There are examples of oxygenase reactions
where the production and use of a reactive oxygen species is physically separated,
typically by the production of peroxide and subsequent use by peroxidases [8]. One of
the goals of current oxygenase research is to produce enzymes that can work directly
and efficiently with independently generated active oxygen species as this would allow
for much simpler enzymes and the use of cell-free oxygenase systems [9].
Aerobic biodegradation of structurally diverse aromatic compounds involves their
conversion into a few key intermediates [typically catechol, protocatechuate, gentisic
acid, hydroxyquinol, hydroquinone, or their derivatives (Scheme 34.1)], which are then
further degraded through a restricted repertoire of pathways to central cellular
metabolism. For the biodegradation of higher polycyclic aromatic compounds, the
rings are metabolized sequentially, yielding various substituted di- and mono-aromatic
compounds. The associated enzymes have been broadly grouped into peripheral
(or upper pathway) and ring-cleavage (or lower pathway) enzymes. Although enzymes
from different bacteria display similar function, there is significant diversity in the
peripheral pathways for specific aromatic compounds. However, the convergence on
catechols produces less diversity in the lower pathway, with ring fission involving either
intradiol- or extradiol-dioxygenases, the biochemical features of which are strongly
conserved. The ring-cleavage intermediates are then subjected to subsequent central
pathways leading to the formation of Krebs cycle intermediates (Scheme 34.2).
Biology supplies two options for initial hydroxylation of the aromatic ring. For the
substrate to be degraded as a carbon source, it needs to be dihydroxylated either
before or after modification of any substituents. Some prokaryotes do this by
sequential monohydroxylation steps, with groups of enzymes that are specialized
either for the first step or the second step (phenolic) in the process. However, many
prokaryotes do both steps at the same time, via dihydroxylation producing
34.1 Enzymology of Aromatic Hydrocarbon Oxidation j1489
CH3 OH
OH OH
OH
OH
COOH
CH3 Cl NO2
NO2
OH R=OH
COOH OH
CH3 OH
OH Cl
COOH COOH
R=OH,NO2 OH NO2
CH2COOH
COOH NO2
OH
COOH COOH
HO
OH OH OH OH
OH OH OH
OH COOH
HOOC HO HO
HO
Protocatechuate Gentisate Catechol Hydroxyquinol Hydroquinone
Ring Cleavage Pathway
Scheme 34.1 Biodegradation of mono-aromatic compounds through diverse peripheral pathways

leading to the formation of central pathway intermediates.
cis-dihydrodiols as intermediates. Although they are subsequently re-aromatized in

the course of metabolism, cis-dihydrodiols can be readily isolated and have been used
as the starting point for many synthetic exercises (Scheme 34.3). Further metabolism
of cis-dihydrodiols in prokaryotes proceeds via dehydrogenation to catechols and one
of two classes of oxidative ring cleavage reaction (intra-diol and extradiol) opening up
a further range of useful synthetic intermediates. Intradiol dioxygenase cleaves the
aromatic ring between two hydroxyl groups (also referred to as the ortho pathway)
whereas extradiol-dioxygenase cleaves adjacent to a hydroxyl group (the meta pathway).
Some eukaryotes can also degrade aromatic compounds using cytochrome P450
monooxygenases, producing phenols [10]; there is, though, limited evidence for
subsequent catechol formation. However, arenes can be fortuitously metabolized by
a wide range of cytochrome P450s (from prokaryotes and eukaryotes), producing
either phenols or trans-dihydrodiols (Scheme 34.3), the latter involving initial arene
epoxidation followed by hydrolysis with an epoxide hydrolase [11].
In the ensuing sections we review the enzymes of aromatic hydrocarbon metab-
olism (including side chain oxidation, particularly where it is affected by proximity to
1490 j 34 Aromatic Oxidations
C12DO O
COOH COOH COOH COOH Acetyl CoA
O O
OH
COOH O O COOH +
Succinic acid
Muconic aicd Muconolactone Ketoadipate enol-lactone 3-Ketoadipic acid
OH OH O OH O Pyruvic acid
Catechol
COOH COOH
CH3 CH CH2 C COOH +
C23DO Acetaldehyde
CHO CH3
2-Hydroxymuconate 2-Oxopenta-4-enoic acid 4-Hydroxy-2-oxovaleric acid
semialdehyde
HOOC
HOOC O Acetyl CoA
COOH COOH COOH COOH
O O +
COOH O O COOH
Succinic acid
PC34DO 2-Carboxymuconic 2-Carboxymuconolactone Ketoadipate enol-lactone 3-Ketoadipic acid
acid
HOOC HOOC CHO CHO COOH Pyruvic acid

OH PC23DO COOH COOH COOH
+
OH OH OH Acetaldehyde
OH
PC45DO 5-Carboxy-2-hydroxymuconic 2-Hydroxymuconic 2-Hydroxymuconic
Protocatechuate semialdehyde
semialdehyde acid
HOOC HOOC COOH HOOC

OH HOOC Oxaloacetic acid
COOH O OH +
CHO O COOH O COOH Pyruvic acid
COOH COOH
O
4-Carboxy-2-hydroxymuconic 2-Pyrone-4,6-dicarboxylic 4-Oxalomesoconic 4-Carboxy-4-hydroxy-2-oxoadipic
semialdehyde acid acid acid
HO O O
Acetyl CoA
OH COOH COOH
COOH COOH +
OH Succinic acid
Hydroxyquinol Maleylacetate 3-Ketoadipic acid
HO O O
HO Acetyl CoA
COOH COOH
CHO
COOH COOH COOH +
OH Succinic acid
Hydroquinone 4-Hydroxymuconic Maleylacetate 3-Ketoadipic acid
semialdehyde
NH2 NH2 NH2 O O
OH Acetaldehyde
COOH COOH COOH COOH
CHO COOH COOH +
2-Aminophenol 2-Aminomuconate 2-Aminomuconate 2-Oxo-3-hexene-1,6-dioate 2-Oxo-4-pentenoate Pyruvic acid
semialdehyde
COOH COOH
OH
O
Fumarylpyruvate Fumarate + Pyruvic acid
COOH
HO HO Maleate + Pyruvic acid
Gentisate Maleylpyruvate
D-Malate
Scheme 34.2 Products from major ring-cleavage dioxygenase-catalyzed reactions.
the aromatic ring), with a focus on complexity and reaction mechanism, followed by
an overview of applications in biocatalysis.
34.1.2
Dioxygenases
Aromatic ring hydroxylating dioxygenases (also known as Rieske non-heme iron

dioxygenases) add both atoms of O2 to the aromatic ring of the substrate. They are
multicomponent enzyme systems (EC 1.14.12.-) that catalyze dihydroxylation of their
H
OH
R Ring cleavage
OH
H Diol DH RCDO
DO
OH OH O
DIMO PH Tyr
R R R
R
O
Tyr OH
DIMO
NIH shift
P450
H
H OH
EH
R O R
H OH
H
Scheme 34.3 Routes of initial metabolism of seen with P450); P450 ¼ cytochrome P450. Step
aromatic compounds. Step 1: 2: Diol DH ¼ diol dehydrogenase; PH ¼ phenol
DO ¼ dioxygenase; DIMO ¼ diiron hydroxylase (diiron or flavin); Tyr ¼ tyrosinase;
monooxygenase (dotted line represents the EH ¼ epoxide hydrolase. Step 3: RCDO ¼ ring
possible formation of an epoxide intermediate cleavage dioxygenase; Tyr ¼ tyrosinase.
and rearrangement involving an NIH shift, as
substrates, and are distinct from ring cleavage dioxygenases (EC 1.13.11.-), which act
on the downstream catechol intermediates in many of the same catabolic pathways.
Over 100 aromatic hydroxylating dioxygenases have been identified based on
biological activity or nucleotide sequence identity. Many of these are quite promis-
cuous, catalyzing the oxidation of a wide range of compounds in addition to their
native substrates [12]. At the same time, however, many of these enzymes are highly
enantioselective, producing chiral cis-dihydrodiols or other chiral products in high
enantiomeric purity [12]. These properties have made aromatic ring hydroxylating
dioxygenases attractive as biocatalysts.
The initial reaction catalyzed by aromatic ring hydroxylating dioxygenases is cis-
dihydroxylation of the carbon–carbon double bond either of adjacent unsubstituted
carbon atoms (Scheme 34.4a) or at a substituted carbon and an adjacent unsub-
stituted carbon of substituted arenes with polar substituents, such as benzoate,
resulting in the formation of chiral cis-dihydroxylated cyclohexadiene carboxylic acids
(Scheme 34.4b). Dioxygenase-catalyzed dechlorination has also been demonstrated
for chlorinated benzoates, benzenes, and biphenyls (Scheme 34.4c and d). Dioxy-
genation at a chlorine-substituted carbon results in the subsequent spontaneous
elimination of chloride. Similar reactions have been shown with nitroaromatic,
aminoaromatic, and sulfoaromatic substrates (Scheme 34.4e–h), resulting in the
release of nitrite, ammonia, or sulfite [13–15] and the formation of dihydroxylated
(catecholic) products for further metabolism [16–21].
Historically, dioxygenases were classified using the Batie system, which was based
on the electron-transfer components present in the ten Rieske non-heme iron
1492j 34 Aromatic Oxidations
OH
Naphthalene H
dioxygenase OH
(a) H cis-dihydroxylation
O2
Naphthalene Naphthalene cis 1,2-dihydrodiol
COOH HOOC OH
Benzoate
dioxygenase OH
(b) H cis-dihydroxylation
O2
Benzoate Benzoate cis 1,2-dihydrodiol
Cl Cl OH OH
Chlorobenzene –
dioxygenase OH Cl OH
H cis-dihydroxylation
(c) and dehalogenation
O2
Chlorobenzene Chlorobenzene cis 1,2-dihydrodiol Catechol
COOH HOOC OH OH
Chlorobenzoate CO2+HCl
Cl dioxygenase
OH OH
(d) Cl cis-dihydroxylation,
dehalogenation and
O2 decarboxylation
2-chlorobenzoate Chlorobenzoate cis 1,2-dihydrodiol Catechol
NO2 O2N OH OH
Nitrobenzene
dioxygenase OH NO2- OH
(e) and nitrite elimination
O2
Nitrobenzene Nitrobenzene cis 1,2-dihydrodiol Catechol
NH2 H2N OH OH
Aniline
dioxygenase OH NH3 OH
(f) and deamination
O2
Aniline Aniline cis 1,2-dihydrodiol Catechol
COOH HOOC OH OH
Anthranilate CO2+NH3
NH2 dioxygenase
OH OH cis-dihydroxylation,
(g) NH2 deamination and
decarboxylation
O2
Anthranilate Anthranilate cis 1,2-dihydrodiol Catechol
COOH COOH
p-Sulfobenzoate COOH
dioxygenase HSO3-
(h) cis-dihydroxylation,
O2 H and desulfonation
OH
SO3H OH
SO3H
OH OH
p-Sulfobenzoate p-Sulfobenzoate cis 1,2-dihydrodiol Protocatechuate
Scheme 34.4 Dioxygenase regioselectivity and formation of (a) and (b) stable or (c)–(h) unstable
intermediates in the first step of metabolism of substituted aromatic compounds.
dioxygenase systems known at that time [22]. Two-component (reductase and

oxygenase; Class I) and three-component (reductase, ferredoxin, and oxygenase;
Class II, III) enzyme systems were represented, and the classes were further
subdivided based on the type of flavin cofactor (FAD or FMN) in the reductase, the
presence or absence of an iron-sulfur center in the reductase, the number of proteins
in the oxygenase, and, if a ferredoxin was involved, the type of iron-sulfur center
(plant or Rieske) in the ferredoxin. This classification system worked well with a
small number of known enzymes, but as more enzyme systems with diverse
properties were identified not all of the new enzymes fitted into this classification
system. A modified classification system based on amino acid sequence alignments
of available oxygenase a subunits was proposed that identified four dioxygenase
families (naphthalene, toluene/biphenyl, benzoate, and phthalate) (Table 34.1) [23].
This classification system (Werlen system) was based on the catalytic activity of the
enzymes since the a subunit of the oxygenase plays a major role in determining
substrate specificity. Nakatsu et al. subsequently built upon this classification system,
adding the an dioxygenases (those in the Batie system Class IA) to the Werlen system
and demonstrating that these an oxygenases formed a separate lineage [24]. In
general, the phylogenetic clustering of oxygenases into families correlates with the
native substrates oxidized by the members.
The toluene/biphenyl family includes enzymes for the degradation of toluene,
benzene, isopropylbenzene, chlorobenzenes, and biphenyl from both Gram-nega-
tive and Gram-positive organisms. Dioxygenases that initiate the pathway for the
degradation of benzene in Pseudomonas putida ML2 [26] and of benzene, toluene, and
ethylbenzene degradation in Pseudomonas putida F1 [27] are members of this family.
The naphthalene family consists of enzymes for the degradation of naphthalene
and phenanthrene. A branch from this group links it to a small group of enzymes for
nitrobenzene and nitrotoluene degradation. Based on sequence alignments of
naphthalene dioxygenase a subunit genes, the genes from Gram-negative bacteria
were found to fall into three distinct groups that have been designated nah-like, nag-
like, and phn-like. The nah-like group contains enzymes from various Pseudomonas
species and includes the well-studied naphthalene dioxygenase from Pseudomonas
sp. NCIB 9816-4 and P. putida G7 [28]. The nag-like group is represented by
dioxygenases for naphthalene and nitroarene compounds (nitrobenzene and nitro-
toluenes) from members of the b-proteobacteria. Comamonas sp. strain JS765,
Acidovorax sp. strain JS42, and Burkholderia sp. strain DNT contain nitroarene
dioxygenases that attack the nitro-substituted carbons of nitrobenzene, 2-nitroto-
luene, or 2,4-dinitrotoluene, respectively, resulting in the removal of the nitro group
as nitrite to form catechol [29–32]. These enzyme systems are very closely related to
naphthalene dioxygenase from Ralstonia sp. strain U2 (nagA; 94% nucleotide
sequence identity), and are believed to have evolved from an ancestral naphthalene
dioxygenase [33, 34]. The phn-like group is represented by the naphthalene/phen-
anthrene dioxygenase from Burkholderia sp. RP007 [35]. A series of naphthalene/
phenanthrene and pyrene dioxygenases has been identified in Gram-positive organ-
isms, including Rhodococcus [36, 37], Nocardioides [38], and Mycobacterium [39]. The
sequences of the a subunit genes from Gram-positive bacteria cluster together, and it
is of note that they are quite distantly related to those from Gram-negative bacteria.
The benzoate family is a diverse group consisting of enzymes that oxidize aromatic
acids (benzoate, toluate, anthranilate, 2-chlorobenzoate, trichlorophenoxyacetate,
and isopropylbenzoate). The benzoate family is thus far represented by two-
1494
Table 34.1 Classification scheme of Werlen system [25].
Rieske non-heme Example substrates Example members Structures

oxygenase family
Reductase Ferredoxin Oxygenase
Naphthalene Naphthalene, indole, nitroarenes, Naphthalene dioxygenase, nitroben-

NDO-F9816-4 NDO-O9816-4, NBDO-OJS765,

phenanthrene zene dioxygenase NDO-O12038
Toluene/biphenyl Toluene, cumene, biphenyl, PCBs, Benzene dioxygenase, toluene diox- BPDO-RKKS102 BPDO-FLB400 BPDO-ORHA1, BPDO-OB1,
benzene ygenase, biphenyl dioxygenase, CDO-OIP01
cumene dioxygenase
Benzoate Benzoate, toluate Benzoate dioxygenase BZDO-RADP1
Phthalate Phthalate Phthalate dioxygenase PDO-RPHK
component enzymes (reductase and dioxygenase) where the dioxygenase is
composed of both a and b subunits. Enzymes that have been characterized
include benzoate dioxygenases from Acinetobacter sp. ADP1 [40], P. putida
mt-2 [41], and Rhodococcus sp. strain 19 070 [42], and toluate dioxygenase from
P. putida pWWO [43].
The phthalate family is a large and diverse group of two-component enzymes,
each consisting of an an oxygenase component (lacking b subunits) and a reductase
component. This family contains enzymes that oxidize aromatic acids (vanillate,
phthalate, 3-chlorobenzoate, phenoxybenzoate, and p-toluenesulfonate). The phthal-
ate family represents the most diverse group in terms of amino acid sequence and
primary substrate, and members of this family have been identified in several different
Gram-positive and Gram-negative bacteria, including Arthrobacter keyseri 12B, Terra-
bacter sp. strain DBF63, and P. putida [44–46]. Like phthalate dioxygenase, isophthalate
dioxygenase is a two-component enzyme with a homomultimeric oxygenase.
Chlorobenzoate 3,4-dioxygenase from Alcaligenes sp. BR60 is a related enzyme that
catalyzes the cis-dihydroxylation and dechlorination of 3-chlorobenzoate [14].
Several dioxygenase systems have been purified and studied in detail to date. Here,
naphthalene dioxygenase (NDO) is used as a reference system and as a basis for
comparison to other related enzyme systems (Figure 34.1). All three protein
components of NDO have been purified from Pseudomonas sp. NCIB 9816-4
[47, 48]. The reductase is a 35 kDa monomer that contains one molecule of FAD
and a plant-type iron-sulfur center, which can accept electrons from either NADH or
NADPH [28, 48]. The ferredoxin is a Rieske [2Fe-2S] center-containing monomer of
approximately 11.4 kDa [28, 48]. The catalytic oxygenase component is an a3b3
hexamer consisting of large (a) and small (b) subunits [49]. Each a subunit contains
two redox centers, a Rieske [2Fe-2S] center, and mononuclear Fe2 þ at the active site.
Individually purified a and b subunits of the oxygenase were reconstituted [50, 51],
demonstrating that both subunits are essential for the reaction, which is consistent
with results obtained with biphenyl dioxygenase (BPDO) [52] and toluene dioxygen-
ase (TDO) [53]. In the reaction cycle (Figure 34.1), two electrons are transferred
sequentially from NAD(P)H to the iron at the active site of the oxygenase via the
reductase, the ferredoxin and the Rieske center of the oxygenase, with the Rieske
center of the oxygenase oxidized and the iron at the active site reduced. The reduced
oxygenase thus catalyzes the addition of both atoms of O2 to the aromatic ring
forming >99% ( þ ) 1R, 2S naphthalene cis-dihydrodiol.
2H+
2H+
NADH+H+ Reduced Reduced Reduced
Reductase Ferredoxin Naphthalene
Oxygenase
O2
NAD+ H OH
Oxidized Oxidized Oxidized OH
Prosthetic group FAD [2Fe-2S] [2Fe-2S]
[2Fe-2S] Fe2+ H
Naphthalene cis-1,2-dihydrodiol
Figure 34.1 Electron-transfer process and catalytic cycle of naphthalene dioxygenase.

1496
Among the aromatic ring hydroxylating dioxygenases that have been identified to
date, two types of oxygenase structures are known: those with both a and b subunits,
such as NDO, and those consisting of only a subunits, such as phthalate dioxygenase.
Studies of hybrid dioxygenases, in which the individual a and b subunits from
different enzymes were substituted, demonstrated that the a subunits of NDO and
the closely related enzymes 2-nitrotoluene dioxygenase and 2,4-dinitrotoluene control
substrate specificity [54, 55]. Similar results were reported with BPDO hybrids, TDO-
tetrachlorobenzene dioxygenase hybrids, and benzene (BDO)-BPDO hybrids [56–60],
which demonstrated that b subunit residues are not near the active site [49, 61, 62]. In
contrast, other studies suggested that b subunit may play a role in determining
substrate specificity in TDO, toluate dioxygenase, and other BPDOs [63–66]. Therefore,
it seems as though the b subunit has a structural function in most dioxygenases, but in
some cases the b subunit may be capable of modulating substrate specificity [12].
34.1.3
Monooxygenases (Di-iron)
Two types of monooxygenase are known to be capable of arene hydroxylation: the

ubiquitous cytochrome P450 and diiron monooxygenases. Extensive reviews have
been written about the properties [67, 68] and biocatalytic potential [11] of cytochrome
P450 (http://en.wikipedia.org/wiki/Cytochrome_P450), but partly because most
P450s with useful natural arene oxidation capabilities are integral membrane proteins
and because catalytic rates are relatively sluggish, P450-catalyzed conversions have not
featured heavily in applied biocatalysis with simple aromatic compounds (unlike the
situation with steroid bioconversions). However, the possibility of heterologous
expression [69, 70] and protein engineering of specificity indicates that this situation
may change in the future. Diiron monooxygenases (DIMOs) are multicomponent
enzyme complexes that utilize dioxygen to catalyze the initial hydroxylation step in
pathways for the oxidation of their respective aromatic compounds, and require NAD
(P)H as an electron donor. These enzyme systems all contain three or four components:
a dimeric hydroxylase protein composed of two or three subunits in an (abc)2 or (ab)2
structure, an NADH oxidoreductase with an N-terminal chloroplast-type ferredoxin
domain and a C-terminal reductase domain with FAD- and NAD(P)-ribose binding
regions, a small effector or coupling with no prosthetic groups, and, in some cases, a
Rieske-type ferredoxin protein [71–73]. Diiron monooxygenases have been the subject
of extensive research in recent years, largely because of interest in the nature of the
diiron center at the active site of these enzymes and its function in catalysis, as well as
the structure, arrangement, and biochemistry of electron transport chain and regu-
latory subunits that serve to deliver electrons to the oxygenase protein [74].
The diiron monooxygenases can be divided into two groups based on their
structure:
1) Three-component monooxygenases: structural analyses of two of these enzymes,
the Dmp (phenol hydroxylase) from Pseudomonas sp. strain CF600 [71, 75, 76], and
the toluene 2-monooxygenase, Tom, from Burkholderia cepacia G4 [72] have
shown that these enzymes are composed of an (abc)2 hydroxylase protein with a
diiron center, an FAD/[2Fe-2S] reductase, and an effector protein. The Dmp, P2
effector protein was shown to have similar secondary structural elements to the
homologous effector protein of the methane monooxygenase [77]. Both Dmp and
Tom catalyze the oxidation of phenol and certain methyl-substituted phenols as
the initial step in the catabolism of these compounds [75, 78]. Tom, however, is
also able to incorporate an oxygen atom into hydrocarbons with an unactivated
benzene nucleus, and sequentially oxidizes toluene to o-cresol and 3-methylca-
techol in the initial steps of the toluene 2-monooxgenase pathway [72]. The
toluene/benzene 2-monooxygenases of B. cepacia JS150 [79] and the phenol
hydroxylases of Comamonas testosterone R5 and Ralstonia eutropha E2 [80] have
similar substrate ranges to Tom and are able to hydroxylate benzene, toluene, and
other inactivated aromatic hydrocarbons (Scheme 34.5).
OH CH3
Toluene
R Phenol
Tom Tbu Tou/Aam
TDO Tmo
CH3
Dmp/Tom OH
CH3 CH3 CH3 p-, o-, m- Cresol
OH
OH
OH o-Cresol
OH OH
R
Toluene OH
Tom cis-dihydrodiol
p-Cresol m-Cresol
Catechol
CH3 CH2OH
OH
XylMA
OH
3-Methyl catechol Benzyl alcohol
Scheme 34.5 Specificities of different diiron enzyme consisting of XylM, a membrane-bound

monooxygenases in bacteria. TDO has also catalytic component with ferrous iron at the
been incorporated to illustrate the diversity of active site, and XylA, a NADH ferredoxin
toluene degradation pathways (see 34.2.2.4 for reductase that has a plant-type
discussion on the enantioselectivity of this [2Fe-2S] cluster and contains FAD [85, 86].
reaction). Methyl group oxidation involves Dmp, Tom, Tmo, Tbu, Tou, and Aam are
xylene monooxygenase, a two-component explained in the text.
2) Four-component monooxygenase: Toluene 4-monooxygenase (Tmo) from

Pseudomonas mendocina KR1 was the first four-component monooxygenase in
the diiron monooxygenase family [73], and consists of an (abc)2 hexameric
1498
hydroxylase with a diiron hydroxo-bridged center, NADH reductase, and an effector

protein homologous to those of methane monooxygenases [73, 81, 82]. Distinct
from three-component monooxygenases, Tmo also contains a separate Rieske-type
ferredoxin component similar to the ferredoxins of the three-component aromatic
ring dioxygenases [73, 83]. This small, soluble, iron-sulfur protein contains the
conserved iron ligands characteristic of other Rieske ferredoxins, required for the
transfer of electrons from reductase to hydroxylases [83]. The effector protein of
Tmo apparently influences catalysis through conformational changes in the
hydroxylase, as it does in other monooxygenases [82], and has also been identified
to play a role in regiospecificity and the efficiency of electron flow coupled with
hydroxylation [84]. Although the effector protein from Tmo has a similar secondary
structure topology to those of methane monooxygenase and Dmp, all three were
found to differ significantly in three-dimensional structures [82].
Tmo catalyzes the regiospecific hydroxylation of toluene to p-cresol, the initial step
in toluene degradation in P. mendocina KR1 [81]. Subsequent studies showed it to
have broad substrate specificity, including chlorobenzene, ethylbenzene, and so on
but excluding phenolic compounds [87]. Several other aromatic monooxygenases
related to Tmo have since been identified in other pseudomonads, including the
toluene 3-monooxygenase (Tbu) from Ralstonia pickettii PKO1 [88] and the toluene/2-
xylene monooxygenase (Tou) from P. stutzeri OX1 [89]. Tbu also hydroxylates
unactivated aromatic compounds but not phenols, and catalyzes the regiospecific
oxidation of toluene at C3 [90, 91]. Tou differs from Tbu and Tmo in two respects: it
hydroxylates both unactivated aromatic hydrocarbons and phenolic compounds, and
it can also hydroxylate toluene with relaxed regiospecificity, producing a mixture of o-,
m-, and p-cresols [89, 92]. Notably, the Aam (alkene monooxygenase) from Xantho-
bacter sp. strain Py2 exhibits a higher degree of similarity to the isofunctional proteins
of Tmo than to the AMO (alkene monooxygenase) from Rhodococcus corallinus B-276,
which has been shown to oxidize benzene to phenol then to catechol, and toluene to a
mixture of o-, m-, and p-cresols [93].
34.1.4
Monooxygenases (Flavoprotein)
The flavoproteins classified as monooxygenase enzymes catalyze the incorporation of

one atom of atmospheric dioxygen into the substrate with the simultaneous reduction of
the other oxygen atom to water [94]. The catalytic mechanism of flavoprotein mono-
oxygenase enzymes involves activation of oxygen to form a C4a-(hydro)peroxyflavin
intermediate [95], which is responsible for the oxygenation of the substrate. When
flavins are typically prosthetic groups that are covalently or tightly bound to the enzyme
they form a permanent part of the enzyme structure. In contrast, cofactors that bind with
a lower affinity are referred to as coenzymes, and can be regarded as substrates as they
often shuttle between different enzymes (e.g., NADH or coenzyme A).
The flavoprotein monooxygenases can be classified into two subclasses based on
their amino acid sequence similarity and the available structure [96].
Single component flavoprotein monooxygenases combine flavin reduction and
monooxygenation in one polypeptide chain. They only use FAD as prosthetic group
and NAD(P)H as electron donor, except that salicylate 1-monooxygenase requires
NADH [97]. Monooxygenases of this subclass have only one dinucleotide binding
domain and release NADPþ directly after reduction of FAD [98]. Well-studied
examples of enzymes belonging to this subclass are p-hydroxybenzoate hydroxylases
(PHBHs) from P. fluorescens, which catalyze the hydroxylation of 4-hydroxybenzoate
to 3,4-dihydroxybenzoate [99]. Another two examples: p-nitrophenol monooxygenase
(PnpA) from Pseudomonas sp. strain WBC-3 [100], which catalyzes the oxygenation of
p-nitrophenol to benzoquinone, and o-nitrophenol monooxygenase (OnpA) from
Alcaligenes strain NyZ215 [101], which converts o-nitrophenol into catechol, were also
characterized as NAD(P)H and FAD dependent.
Two distinct subclasses of two-component flavoenzymes have been identified. The
first of these, exemplified by 4-hydroxyphenylacetate-3-monooxygenase from
P. putida, is composed of a separate oxygenase and regulatory protein required for
the efficient coupling of pyridine nucleotide oxidation and substrate oxygenation
activities [102, 103]. Another member of this subclass, 2,4,6-trichlorophenol 4-
monooxygenase (TcpA) from Cupriavidus necator JMP134, uses both 2,6-dichloro-
phenol and 2,4,6-TCP (TCP ¼ trichlorophenol) as its substrates [104, 105]. When 2,6-
dichlorophenol was used, TcpA converted it into 2,6-dichloroquinol without detect-
able 6-chlorohydroxyquinol. When 2,4,6-TCP was the substrate, TcpA converted it
mainly into 6-chlorohydroxyquinol with minor accumulation of 2,6-dichloroquinol.
Hydroxylation of the 4-position of 2,6-dichlorophenol was catalyzed by TcpA through
a simple monooxygenase reaction without the formation of 2,6-dichloroquinone as
an intermediate. The formation of 2,6-dichloroquinone from 2,4,6-TCP oxidation
must be necessary for the subsequent conversion into 6-chlorohydroxyquinol.
Furthermore, chlorophenol 4-monooxygenase (TftD) from Burkholderia cepacia
AC1100 can convert 2,4,5-TCP into 2,5-dichloro-p-quinol and then into 5-chlorohy-
droxyquinol but converts 2,4,6-TCP only into 2,6-dichloro-p-quinol as the final
product [106].
More commonly encountered two-component flavoprotein monooxygenases
are composed of distinct flavin reductase and monooxygenase activity segregated
on separate peptide units. The flavin reductase provides the reduced flavin for the
monooxygenase enzyme. These enzymes can be further divided according to
their dependence for FMN or FAD but there are no obvious functional or
structural distinctions between the FMN-dependent and FAD-dependent enzyme
systems. Examples in this subclass are styrene monooxygenase [107], phenol
hydroxylases (PheA1A2) from Geobacillus thermoglucosidasius A7 [108], 4-hydro-
xyphenylacetate 3-hydroxylase (HpaBC) from Escherichia coli W ATCC
11105 [109], and p-hydroxyphenylacetate (HPA) hydroxylase (HPAH) from
Acinetobacter baumannii [110]. Several p-nitrophenol monooxygenases identified
in Gram-positive bacteria belong to this family, including NpdA1A2 in Arthro-
bacter sp. strain JS443 [111], NpcAB in Rhodococcus opacus SAO101 [112],
NpdA1A2 in Arthrobacter sp. strain NyZ415 (Liu PP, 2010), and NphA1A2 in
Rhodococcus sp. strain PN1 [113].
1500
34.1.5
Ring Cleavage Dioxygenases
The aerobic catabolism of aromatic compounds in bacteria usually proceeds via one
of five intermediates or its derivatives: catechol, protocatechuate, gentisate, hydro-
quinone, and hydroxyquinol. The ring cleavage of catecholic compounds is per-
formed by enzymes from one of two distinct classes: intradiol and extradiol
dioxygenases [114]. Intradiol dioxygenases utilize non-heme Fe(III) to cleave the
aromatic nucleus ortho to (between) the hydroxyl substituents and have an absolute
requirement for substrates with vicinal diols. In contrast, extradiol dioxygenases
utilize non-heme Fe(II) to cleave the aromatic nucleus meta (adjacent) to the
hydroxyl substituents (Tables 34.2 and 34.3). Although the distinctions between
intradiol and extradiol dioxygenases may appear to be minor, they are in fact a
manifestation of enzymes that have completely different structures and utilize
different catalytic mechanisms [115, 116]. Extradiol dioxygenases cleave a wider
variety of substrates, and occur in a wider variety of pathways, including some
biosynthetic pathways and pathways that degrade non-aromatic compounds. Thus,
extradiol dioxygenases appear to be more versatile than their intradiol counterparts.
Protocatechuate and hydroxyquinol, which are essentially substituted catechols,
can also act as the substrates of intradiol dioxygenases. By contrast, not all aromatic
compounds that are substrates of extradiol-type cleavage possess vicinal hydroxyl
groups. Non-catecholic compounds that are subject to extradiol-type cleavage
include the other intermediates gentisate, hydroquinone, and 2-aminophenol. In
comparison to the substrates of typical extradiol dioxygenases, these compounds
are either dihydroxylated in the para position and/or possess a carboxylate or an
amino group in place of the second hydroxyl group. Another difference between
intradiol and extradiol enzymes is that the former generally cleave catechols
possessing mildly electron-withdrawing substituents in vivo [117]. By contrast,
extradiol enzymes cleave catechols possessing electron-donating substituents
in vivo [117]. However, the physiological relevance of the cleavage type is unclear,
partly because in vitro studies indicate that the two classes of enzymes cleave similar
ranges of substrates. Intradiol enzymes are apparently unable to transform sub-
strates possessing strongly electron-withdrawing substituents in vitro [118, 119],
whereas extradiol enzymes can cleave compounds such as nitrocatechol at a low
rate [120].
34.1.5.1 Intradiol Dioxygenase
Catechol 1,2-Dioxygenase (C12DO) Catechol 1,2-dioxygenase (C12DO) catalyzes the

typical intradiol ring cleavage of catechol to cis,cis-muconic with the consumption of
1 mol of molecular oxygen [121]. C12DO from P. putida mt-2 was found to consist
of two identical subunits with the active site structure of [aa-Fe3þ ] [146], whereas
P. arvilla C-1 produces isozymes consisting of aa, ab, bb dimers ranging from 60 to
64 kDa with similar properties [123]. One atom of ferric iron is bound between a pair
of subunits, producing a single active site per dimer [147].
Table 34.2 The ortho- and meta-ring-cleavage dioxygenases.
Dioxygenases Substrate Products Subunit composition Cofactor Source material Reference
Intradiol dioxygenase
Catechol 1,2-dioxygenase Catechol cis,cis-Muconic acid aa,ab,bb Fe3þ Pseudomonas arvilla C-l [121–123]
Protocatechuate 3,4- Protocatechuate 3-Carboxy-cis, cis-muconic (ab)n (n ¼ 3–12) Fe3þ Pseudomonas aeruginosa [124–129]
dioxygenase acid
Hydroxyquinol 1,2- Hydroxyquinol Maleylacetate Homodimer Fe3þ Nocardioides simplex 3E [130, 131]
dioxygenase
Extradiol dioxygenase
Catechol 2,3-dioxygenase Catechol 2-Hydroxymuconate a4 Fe2þ Pseudomonas arvilla [132–135]
semialdehyde
Protocatechuate 4,5- Protocatechuate 2-Hydroxy-4-carboxymuco- a2b2 Fe2þ Pseudomonas testosteroni [136–138]
dioxygenase nate semialdehyde
Protocatechuate 2,3- Protocatechuate 2-Hydroxy-5-carboxymuco- a4 Fe2þ /Mn2þ Paenibacillus sp. strain JJ-1b [139, 140]
dioxygenase nate semialdehyde
2-aminophenol 2-Aminophenol 2-Aminomuconate a2b2 Fe2þ Pseudomonas pseudoalcali- [141, 142]
1,6-dioxygenase semialdehyde genes JS45
Hydroquinone 1,2- (Chloro)hydroquinone 4-Hydroxymuconic a2b2 Fe2þ Pseudomonas fluorescens ACB [143]
dioxygenase semialdehyde
Gentisate 1,2-dioxygenase Gentisate Maleylpyruvate a4 Fe2þ Pseudomonas acidovorans [144, 145]
34.1 Enzymology of Aromatic Hydrocarbon Oxidation
j1501
Table 34.3 Enzymology of ortho- and meta-ring-cleavage dioxygenases.
Catechol 1,2- Protocatechu- Hydroxyqui- Catechol 2,3- Protocatechuate 4,5- Protocatech- 2-Aminophe- Hydroqui- Gentisate 1,2-
dioxygenase ate 3,4- nol 1,2- dioxygenase dioxygenase uate 2,3- nol 1,6- none1 1,2- dioxygenase
dioxygenase dioxygenase dioxygenase dioxygenase dioxygenase
EC number 1.13.11.1 1.13.11.13 1.13.11.37 1.13.11.2 1.13.11.8 — — — 1.13.11.4

Source material Pseudomonas Pseudomonas Nocardioides Pseudomonas Pseudomonas testosteroni Paenibacillus Pseudomonas Pseudomonas Pseudomonas
arvilla C-l aeruginosa simplex 3E sp. OC1 ATCC49249 sp. strain pseudoalcali- fluorescens acidovorans
(LigAB) JJ-1b (PraA) genes JS45 ACB (HapCD)
(AmnAB)
Substrate speci- Catechol (100) Protocatech- Hydroxyqui- Catechol (100) Protocatechuate Protocatech- 2-Aminophe- Hydroqui- Gentisate
ficity (relative uate (100) nol (100) uate (100) nol (100) none (100) (100)
activity in
parentheses)
4-Methylcate- 3-Hydroxyca- 5-Chlorohy- 4-Methylcate- 3,4-Dihydroxyphenyl 2,3,4-Trihy- 6-Amino-m- Methylhydro- 3-Methylgenti-
chol (90) techol (2.4) droxyquinol chol (100) droxybenzo- cresol (25) quinone (120) sate (43)
(2.4) ate (30)
3-Methylcate- Catechol (0.4) 6-Chlorohy- 3-Methylcate- 5-Methoxygallic acid Protocatech- 2-Amino-m- Methoxyhy- 4-Methylgenti-
chol (8) droxyquinol chol (62) uate methyl cresol (1) droquinone sate (6)
(5.0) ester (10) (50)
4-Chlorocate- 3-Methylcate- Catechol 4-Chlorocate- 5-Methylprotocatechuate 3,4-Dihy- 2-Amino-4- Chlorohydro- 3-Ethylgenti-
chol (3.6) chol (0.4) (0.6) chol (51) droxy-5- chlorophenol quinone (70) sate (31)
methyl-ben- (15)
zoate (10)
3-Methyoxyca- 4-Methylcate- 3-Hydroxyca- Gallic acid 4-Clorocate- Catechol (13) Bromohydro- 3-Isopropyl-
techol (0.8) chol (0.2) techol (33) chol (2) quinone (30) gentisate (17)
3-Hydroxyca- 3,4-Dihydrox- 3-Hydroxyca- Sulfonylcatechol Catechol 2,3-Difluoro- 3-Bromogenti-
techol (0.6) yphenylacetic techol (21) (1.5) hydroquinone sate (26)
acid (0.2) (80)
Protocate- 3,4-Dihydrox- Protocatechu- Protocatech- 2,5-Difluoro- 3-Fluorogenti-
chualdehyde yphenylman- ate (0.15) uate ethyles- hydroquinone sate (33)
(0.02) delic acid ter (3) (75)
(0.1)
3,4,5-Trihy- 3,5-Difluoro- 4-Fluorogenti-
droxybenzo- hydroquinone sate (13)
ate (1.2) (90)
Catechol 1,2-dioxygenases fall into two types [148]. Type I dioxygenases (P. arvilla)
are relatively specific enzymes primarily using catechol and methylcatechol as
substrates [146]. Type II enzymes (Rhodococcus opacus 1CP) are relatively nonspecific,
being able to convert chlorinated catechols more rapidly than catechol and accom-
modate a wide range of methyl- or methoxy-substituted catechols [149]. By either type
of catechol 1,2-dioxygenase the physiological substrate, catechol, and the 4-
substituted catechol derivatives were cleaved exclusively in the intradiol manner.
However, when 3-methylcatechol was used as a substrate both 2-methylmuconic acid
from intradiol cleavage and 2-hydroxy-6-oxo-2,4-heptadienoic acid produced by extra-
diol cleavage between the methyl carbon and a hydroxyl carbon were formed [150]. In
contrast, C12DO from Brevibacterium fuscum [150] and Arthrobacter [151] catalyzes
only the intradiol cleavage even with 3-substituted catechol as substrate.
Protocatechuate 3,4-Dioxygenase (PC34DO) Protocatechuate 3,4-dioxygenase

(PC34DO) catalyzes the intradiol ring cleavage of protocatechuate to form b-
carboxy-muconic acid [124]. The enzyme obtained from p-hydroxybenzoate-induced
cells of P. putida had a molecular weight of approximately 700 kDa and consisted of twelve
identical subunits [124, 125, 152]. Each subunit was reported to dissociate further into two
non-identical polypeptides [ab-Fe3 þ ] with a non-heme ferric ion bound at the interface of
the subunits creating a single active site [153]. All of the well-characterized P34DOs seem
to be similarly made up of equimolar amounts of two non-identical subunits organized as
ab protomers in complexes of three to twelve subunits [154]. The sequences and tertiary
structures of the two polypeptides are very similar, suggesting that the ancestral enzyme
was originally a homodimer with two active sites [129, 152].
Hydroxyquinol 1,2-Dioxygenase (HQ12DO) Hydroxyquinol, a central intermediate

in the degradation of polychloro- and nitroaromatic pollutants, is subject to intradiol
cleavage to form 3-hydroxy-cis,cis-muconate, which occurs in solution in the keto
form, that is, as maleylacetate [155]. HQ12DO (hydroxyquinol 1,2-dioxygenase) has
been characterized from Gram-negative bacteria (Burkholderia cepacia AC1100),
Gram-positive bacteria (Nocardioides simplex 3E), and also from fungi (Trichosporum
cutaneum). HQ12DO from N. simplex 3E is a homodimer of about 65 kDa with a
quaternary structure (a FeIII)2 [156] with markedly high selectivity in hydroxyquinol
ring cleavage. DNA sequencing showed that HQ12DOs are most closely related to
catechol and chlorocatechol dioxygenases [112, 157–159]. Nevertheless, HQ12DOs
appear to have a distinct substrate specificity and do not convert catechol and only
slowly convert substituted catechols [158, 160–162]. Solvent exposure in the upper
part of the active site together with differences in several active site residues, revealed
in the crystal structure of HQ12DO from N. simplex 3E, could explain the preferential
cleavage of hydroxyquinol versus catechols [131].
34.1.5.2 Extradiol Dioxygenase
Catechol 2,3-Dioxygenase (C23DO) Catechol 2,3-dioxygenase (C23DO), first iden-

tified in Pseudomonas sp. OC1, cleaves catechol in the extradiol proximal manner to
1504
yield a-hydroxymuconate semialdehyde [163]. The enzyme is easily inactivated by

oxidizing agents, such as air or H2O2 [135]. However, a low concentration of organic
solvents such as acetone and ethanol can protect the enzyme from inactivation.
Hence, buffer solutions containing 10% acetone were used in all procedures to
achieve crystallization of the enzyme [132, 133]. Unlike catechol 1,2-dioxygenase,
catechol 2,3-dioxygenase is colorless, and requires ferrous iron as its cofactor [135].
Steady state kinetic analyses and studies with inhibitors are consistent with an
ordered bi-uni mechanism in which the substrate first combines with the enzyme
and then reacts with oxygen to form a ternary complex, quite similar to that of
protocatechuate 3,4-dioxygenase [126, 164].

(PC45DO), catalyzing the extradiol ring cleavage of protocatechuate to yield
2-hydroxy-4-carboxymuconate semialdehyde, has been purified from several strains
grown on 4-hydroxybenzoate. The enzyme from Rhizobium leguminosarum [165] was
shown to be a homodimer [165], whereas PC45DO from P. testosterone and S.
paucimobilis SYK-6 were tetramers, loose dimers of two tightly bound a-b hetero-
dimers with four ferrous ion atoms per molecule [136, 137, 166]. The a subunit
interacts extensively with one face of the b subunit of the same protomer and with the
b subunit of the second protomer. The latter a–b contacts stabilize the a2/b2 dimer,
which lacks a–a or b–b contacts. In PC45DO from S. paucimobilis SYK-6, the
subunits appear to be unrelated, and the b subunit, which contains the active site,
is similar to the protomers of the homo-oligomeric enzymes [142, 167].

(PC23DO) has been described thus far only among strains of Bacillus, for example,
Paenibacillus sp. (formerly Bacillus macerans) JJ1b [139, 168]. PC23DO catalyzes
proximal extradiol ring cleavage of protocatechuate to form 2-hydroxy-5-carbox-
ymuconate semialdehyde [140]. The quaternary structure of PC23DO is similar to
that of most other extradiol catecholic dioxygenases in having a single type of
subunit, in contrast to the distal enzyme PC45DO protocatechuate despite the fact
that it utilizes the same substrate [137]. The substrate range of PC23DO is similar
to that of C23DO and chlorocatechol dioxygenase [169], although the latter do not
readily accommodate bulky side chains in the position equivalent to C1 of
protocatechuate. Thus, the active site pocket of PC23DO seems to be sufficiently
open to bind substrate analogs with significant alterations in structure from
protocatechuate. Moreover, the ring cleavage substrate does not appear to be as
sensitive to inductive effects of substituents as proposed for other dioxy-
genases [170]. Therefore, in PC23DO the binding orientation of the substrate
relative to the reactive oxygen species in the active site is primarily controlled by the
enzyme structure rather than the functional groups of the substrates them-
selves [140]. This specificity might be achieved by chelation of the vicinal hydroxyl
groups of the substrate to the iron as the primary binding determinant. Such
chelate structures have been observed for other extradiol dioxygenase–substrate
complexes [144, 145, 171].
2-Aminophenol 1,6-Dioxygenase (AP16DO) 2-Aminophenol is subject to ring fission
by 2-aminophenol 1,6-dioxygenase (AP16DO) to form 2-aminomuconate semialde-
hyde in the nitrobenzene-degrading strain P. pseudoalcaligenes JS45 [172]. This
enzyme has an a2b2 subunit structure containing 2 moles of ferrous iron per mole
of protein [141]. Four iron chelating histidines characteristic of meta-ring-cleavage
dioxygenases are conserved in the b subunit [173, 174]. Aromatic compounds not
possessing either two vicinal hydroxyl groups or an amino group adjacent to a
hydroxyl group were not substrates and, with the exception of guaiacol, had no
inhibitory effect on 2-aminophenol oxidation [141]. AP16DO cleaves neither proto-
catechuate nor gentisate, indicating that the active site must be considerably different
from that of PC45DO and GDO. Moreover, catechol is a suicide substrate of
AP16DO [142], possibly resulting from oxidation of the active-site iron from the
ferrous state to the ferric state as proposed for the suicide inactivation of C23DO from
P. putida mt2 by 4-methylcatechol [175]. The minor effect of 2-aminophenol on the
activity of C23DO suggests that the amino group of 2-aminophenol prohibits access
to the active site [142].
Hydroquinone 1,2-Dioxygenase (HDO) Hydroquinone 1,2-dioxygenase (HDO)

catalyzes the ring fission of a wide range of hydroquinones to the corresponding
4-hydroxymuconic semialdehydes. HDO involved in the catabolism of 4-hydroxya-
cetophenone in P. fluorescens ACB was shown to be an a2b2 heterotetramer [143], and
differed from a proposed aminohydroquinone dioxygenase from Cupriavidus necator
JMP134, which was identified as a homodimer. These two enzymes show little
similarity and may belong to different families of extradiol dioxygenases, although
they share similar substrate specificity. Substrate profiling of HDO from P. fluorescens
ACB showed that both para-hydroxyl groups of hydroquinone are crucial for enzyme
activity. Hydroquinones with an electron-donating methyl or methoxy group were
readily converted, while those containing electron-withdrawing substituents were
converted at lower rates. Moreover, the number and position of fluorine substituents
determined both the reaction rate and the regioselectivity of dioxygenation. However,
introduction of a hydroxyl (hydroxyquinol) or carboxyl (gentisate) group at the ortho-
position both resulted in enzyme inhibition. This suggests that the activity of HDO is
determined by both electronic and steric constraints and also that substrates may
become differently oriented in the active site [143]. However, HDO was reversibly
inhibited by a large number of phenolic compounds. 4-Hydroxybenzylic compounds
and 4-hydroxycinnamates were strong inhibitors, which may result from the
4-hydroxyl group serving as the iron ligand and the fact that the para-substituent
of the phenol is an important determinant for discriminating between weak and tight
binding [143].
Gentisate 1,2-Dioxygenase (GDO) Gentisate 1,2-dioxygenase (GDO) catalyzes the

ring cleavage of gentisate with the formation of maleylpyruvate in the catabolism of
aromatic compounds such as 3- and 4-hydroxybenzoates [168, 176] and salicy-
late [177]. GDO from P. acidovorans and P. testosterone have an (aFe)4 quaternary
structure. However, homotrimeric and dimeric GDOs have also been identified
1506
[178, 179]. As proposed for other extradiol class of catecholic dioxygenases [171], the
Fe2þ appears to play a central role. It is likely that substrate is bound to gentisate 1,2-
dioxygenase with the carboxylate close to the iron, since the site of cleavage for
gentisate is adjacent to the carboxylate [144, 145]. The ability of the enzyme to bind
gentisate analogs with very bulky substituents on the side of the ring away from the
cleavage site is consistent with the proposed substrate binding orientation. It is
evident from the inductive effects of fluorine-substituted gentisates that ring
substituent effects also occur. However, both electron-donating and electron-with-
drawing substituents appear to decrease the rate. Moreover, the rate is decreased by
substituents located both ortho (or para) and meta to any potential reactive ring
position. Thus, the inductive effects are complex and may differentially affect more
than one step of the reaction [144, 145]. Other features of the gentisate substrate
might also play crucial roles in these electronic effects. In particular, the 5-hydroxyl
group may be important in resonance stabilization of an intermediate in the reaction,
as has been proposed for the vicinal hydroxyl groups of the substrate of both intra- and
extradiol catecholic dioxygenases [180].
34.2
Biotransformations of Aromatic Compounds
34.2.1
Whole Cell versus Cell-Free Reactions and Strategic Approaches
For provision of precursors for small-scale synthesis, the simplicity of incubation of

an aromatic substrate with a biocatalyst in air is attractive. However, when
considering the development of an industrial process the direct biological use of
oxygen causes problems with productivity [181]. As highlighted in the introduction,
the triplet ground state of the oxygen molecule is relatively un-reactive, which
means that for biological reaction it requires initial activation by partial reduction.
For reaction with unactivated aromatic substrates this typically requires NAD(P)H
and much of the structural complexity of metal containing oxygenases comes from
the requirement to convert the two electron delivery from NAD(P)H oxidation into
discrete one-electron reduction steps. Furthermore, it is evident from comparison
of kcat values (Table 34.4) that the complexity and control of this electron-transport
pathway makes such enzymes relatively sluggish compared to oxygenases that do
not use an external reductant, and positively pedestrian compared to hydrolytic
enzymes. So, although it would be possible to provide an NAD(P)H recycling
system for extended use of cell-free enzyme systems, the low kcat combined with
relatively poor stability of these complex enzymes means that total turnover
numbers (TTNs) are generally small and any advantage of working with cell-free
systems is usually outweighed. Furthermore, some oxygenases are membrane
bound, with the membrane helping to concentrate substrate. Thus, for most
monooxygenase and dioxygenase reactions on unactivated aromatic substrates,
whole cell biocatalysts are preferred.
34.2 Biotransformations of Aromatic Compounds j1507
While the use of whole cell oxygenase biocatalysts solves the problem of NAD(P)H
recycling, it introduces the problem of product metabolism. Thus, dihydrodiols
formed by the action of dioxygenases on aromatic substrates do not accumulate in
their wild-type host strains as they are rapidly converted into catechols and on into
intermediary metabolites. To obtain product accumulation requires one of the
following: mutant strains that are unable to catalyze the next step in metabolism,
fortuitous oxidation where the substrate is not the natural substrate for the oxygenase
and thus the enzymes for product metabolism may not be present, or heterologous
expression of the oxygenase in a host lacking enzymes capable of further metabolism.
With many simple hydrolytic biocatalysts the latter approach has proved to be very
successful. Heterologous expression can often achieve higher enzyme yields and
specific productivities than growth of the natural host and enzymes can be tailored to
improve the ease of subsequent recovery. However, heterologous expression of
oxygenases has met with limited success, at least with respect to production of
improved biocatalysts (Table 34.4). Several factors have been suggested as contrib-
uting to the relatively low specific activity of recombinant oxygenases, such as poor
assembly of the multicomponent enzymes (and possibly higher turnover of non-
assembled components), different membrane composition affecting substrate
uptake, and the absence of host components necessary for stabilization or reactiva-
tion. An alternative approach to heterologous expression (usually in E. coli) is self
cloning or homologous overexpression in the natural or closely related host strain.
Even if problems of heterologous overexpression (or high level homologous
expression) can be resolved there are additional issues that need to be considered
before high volumetric productivity biotransformations can be achieved. In partic-
ular, the relatively high Km of oxygenases for molecular oxygen [181] combined with
Table 34.4 Comparison of catalytic turnover of cofactor-requiring dioxygenases and

monooxygenases and cofactor-independent dioxygenases, and maximum specific activities of
oxygenases expressed in cells of wild-type (P) and recombinant strains (R). Data taken from
Reference [181].
Enzyme Substrate kcat (s1) Activity

(U [gm dry wt]1)
Dioxygenase
Toluene dioxygenase Toluene 9.4 750 (P)
235 (R)
Naphthalene dioxygenase Naphthalene 1.8
Monooxygenase
CYP2F2 Methylnaphthalene 1.1
Toluene-4-monooxygenase Toluene 2
Styrene monooxygenase Styrene 1.6 200 (P)
180 (R)
No cofactor
Catechol 2,3 oxygenase Catechol 187
Protocatechuate 3,4 dioxygenase Protocatechuate 758
1508
competition from cellular respiration means that the oxygen supply will soon become
limiting in high activity or high cell density processes. The rate of supply of NAD(P)H
could become limiting in high intensity processes, while product toxicity is a
perennial problem that usually requires some form of in situ product recovery. A
preliminary cost analysis [181] indicated that if all of these issues were successfully
addressed then biotransformations yielding product at a cost of 1$ kg1 were feasible.
However, given that most of these problems can be traced back to the use of molecular
oxygen as a substrate, the engineering of simpler catalysts capable of using peroxide
in place of molecular oxygen is an attractive goal.
34.2.2
Dihydroxylations
Arene dihydroxylation, in which both atoms of dioxygen are inserted into the substrate
simultaneously by a dioxygenase, forms a cis-dihydrodiol. With a substituted arene
there is the potential to achieve both regioselectivity and stereoselectivity, yielding
enantiomerically pure dienediols as precursors for further reaction. Early synthetic
studies used a mutant strain 39/D of Pseudomonas putida that expresses a toluene
dioxygenase but lacks the diol dehydrogenase [182], which Hudlicky exploited for an
improved, four-step synthesis of PGE2a [183]. In the early 1980s ICI started producing
small amounts of the meso benzene cis-diol, initially for the production of polyphe-
nylene [184]. This allowed Ley et al. [185] to devise a synthetic scheme for pinitol.
Subsequently, the range of enzymes and specificities has been extended to include
benzene dioxygenase (BDO), naphthalene dioxygenase (NDO), biphenyl dioxygenase
(BPDO), benzoic acid dioxygenase (BZDO), chlorobenzene dioxygenase (CBDO),
nitrobenzene dioxygenase (NBDO), and variants from different organisms (see
above). Most of these are available in recombinant form in E coli [23, 27, 30–32,
186–188], providing a platform for targeted and random mutagenesis leading to
variants with altered substrate specificities, for example [189]. Additionally, a scheme
has been devised to detect dioxygenase genes in environmental isolates, using PCR
amplification of the specificity determining portion of the gene encoding the a-sub-
unit, and expressing it as a hybrid in a backbone consisting of the bphA gene cluster
from Burkholderia LB400 cloned into an E coli expression system [190].
With an increasing range of choices available it is important to have some
guidelines for selection of the most suitable biocatalyst. The range of factors that
needs to be considered is discussed in the rest of this section.
34.2.2.1 Substrate Specificity

As a general rule it can be assumed that enzymes that are naturally active against
monocyclic substrates (e.g., BDO, TDO) are poor catalysts for, or do not recognize,
polycyclic substrates (with the exception of biphenyl) and vice versa. Although TDO
will dihydroxylate naphthalene the rate of reaction is poor in comparison to rates with
monocyclic arenes [191]. Clearly, given that discrimination based on size is primarily
determined by fit into the active site, mutant and hybrid enzymes may bridge that
gap. However, whereas enzymes such as BDO and TDO typically hydroxylate the 2,3-
positions of monosubstituted benzenes, enzymes that naturally attack carboxylates
(BZDO) [192, 193] and nitrobenzenes (NBDO) [32, 62] specifically require these
functional groups which direct the formation of 1,2-(ipso)-diols. With carboxylates,
these can be isolated as stable intermediates, but with nitrobenzenes the presumed
ipso-intermediate is unstable, converting spontaneously into the catechol with
elimination of nitrite (Scheme 34.4).
34.2.2.2 Reaction Selectivity

As discussed in the introduction, because oxygenases operate through initial
activation of oxygen, it is not surprising that where readily oxidizable groups can
enter the active site, whether attached to an arene or not, they can also be hydroxylated
by dioxygenases. Thus, dioxygenases have been shown to catalyze sulfoxidation
[194, 195], alkene dihydroxylation [196–198], benzylic hydroxylation [197–200], and
even desaturation reactions (e.g., ethylbenzene to styrene [201–203]). Thus, with
complex substrates multiple products may be obtained and there is a possibility that
initial products will be substrates for further reaction. While some of this may not be
avoidable (or may be desirable) studies by Boyds group produced some general rules,
most of which are interpretable in terms of a balance between reactivity and substrate
binding. Thus, NDO and BPDO catalyze a wide range of reactions on substituted
monocyclic arenes, presumably because the substrate is free to rotate in the active
site, exposing the most reactive (electron rich) sites and the enzyme cannot impose
selectivity. TDO has a binding pocket for a small alkyl group, which tends to protect
the benzylic (and equivalent akylbenzyl sulfide and styrene) positions from attack.
However, factors that affect the normal binding, such as ring disubstitution and
longer alkyl chains (e.g., alkyl aryl sulfides), reduce the rate of dihydroxylation
(presumably as a result of poor/slow binding), thus increasing the extent of
alternative reaction products. Similarly, as previously highlighted, dioxygenases with
polar group binding pockets tend to direct the reaction in favor of dihydroxylation. At
the present time this picture is far from complete and enzymes with better R group
binding pockets may be available in nature, or could possibly be engineered.
In a typical enzyme reaction, binding of the product is disfavored, so further
reaction of ring dihydrodiols is unlikely. However, with unnatural products such as
sulfoxides a second round of binding and reaction is possibly yielding the ring
dihydroxylated sulfoxide [204, 205].
34.2.2.3 Regioselectivity
As indicated above, benzoate and nitrobenzene dioxygenases specifically direct
dihydroxylation to the 1,2-position, and this also seems to be the case with aniline
dioxygenase [206]. This is efficient in terms of microbial degradation, and subsequent
enzymatic or spontaneous re-aromatization eliminates the R group to yield the
catechol. However, only the benzoate derivatives are sufficiently stable to be isolated
as synthetic precursors. Regardless of the presence of other substituents the directing
effect of the carboxyl and nitro groups predominates with these enzymes, although
upon NBDO oxidation of 1-nitronaphthalene the predominant product was cis-(1,2)-
dihydroxy-1,2-dihydro-8-nitronaphthalene [32]. BDO, TDO, and CBDO all direct
1510
dihydroxylation to the 2,3-position of those monosubstituted benzenes that are

accepted as substrates, and although nitrobenzene and benzoic acid do not appear
to be substrates for TDO, 4-nitrotoluene is dihydroxylated between the two substi-
tuents (but this is not the only product). Interestingly, NDO attacks 2-naphthoic acid
at the 1,2-position, whereas with all other monosubstituted naphthalenes, including
2-nitronaphthalene, the substituent directs dihydroxylation to the other ring.
Although the natural product of NDO is 1,2-dihydroxydihydronaphthalene, with
biphenyl as substrate a small amount of the 3,4-dihydroxylated product was
obtained [207, 208]. Site-directed mutagenesis of NDO has subsequently produced
a version in which 3,4-dihydroxylation was the major outcome [208], demonstrating
that there is no inherent chemical reason for the 2,3-specificity of dioxygenases. A
recent analysis of an o-xylene dioxygenase from Rhodococcus sp. showed that
with toluene and ethylbenzene a mixture of 2,3- and 3,4-substitutions was
obtained [209, 210]. In time, other variants with 3,4-specificity may be produced,
although achieving 100% conversion from 2,3- to 3,4-specificity is a major challenge.
With the 2,3-specific enzymes such as BDO, TDO, and CBDO the rule for attack on
disubstituted compounds, when it occurs, is that dihydroxylation is sterically directed
by the bulkiest group. However, this is rarely absolute so mixtures of dihydrodiols are
frequently obtained and, as indicated above, the poor fit of these substrates into the
active site means that rates of production are considerably lower than for the
monosubstituted compounds and attack at other-non-aromatic positions can occur.
In particular, Boyd et al. observed a meta effect with TDO oxidation of substituted
ethyl and propyl benzene [211, 212]. With ortho- and para-substitution cis-dihydro-
diols were obtained according to the rules above. However, with meta-substituted
alkylbenzenes, benzylic alcohols were obtained in addition to cis-dihydrodiols, and
these did not appear to be further ring hydroxylated.
34.2.2.4 Stereoselectivity
Formation of cis-dihydrodiols from a monosubstituted benzene produces two inter-
dependent chiral centers of opposite configuration (1R,2S or 1S,2R). Boyd and
Bugg [191] point out that technically these are diastereomers because the diene
backbone can adopt one of two enantiomeric helical conformations (M and P
conformers) that can interconvert via bond rotation. However, the ratios of the
different conformers obtained depend on the R groups and will vary during any
subsequent synthetic steps; consequently, the most important consideration from
the perspective of synthetic chemistry is the configuration of the stereogenic centers.
Given that that active site of an enzyme is fixed within its 3D structure, stereo-
selectivity results from substrate binding in a single orientation in the active site. The
presence of an R group can influence this by either binding in a pocket in the active
site or being excluded from the active site. Thus with all substituents except F
(the smallest R group), including monosubstituted, 1,2-disubstituted, and 1,3-
disubstituted, the dihydrodiols produced by TDO show >98% e.e. (1S,2R) with
respect to the sterically dominant R group in the 3-position. As would be expected
from the discussion above, with 1,4-disubstituted benzenes the e.e. of the resulting
dihydrodiols depends on the relative size of the substituents.
34.2.2.5 Effect of Ring Heteroatoms
Although several dihydroxylations of heteroarene substrates have been recorded
[201, 202, 213], there is no evidence that the products have been used as precursors
for synthesis. This is partly because some of the products are unstable, undergoing
spontaneous rearrangements [213] and also because multiple products tend to be
obtained. Monocyclic furans, pyrroles, and thiophenes are probably all attacked by
TDO, but only the latter yield stable diols, and 3- but not 2-subtituted thiophenes are
dihydroxylated [199, 200], the former at the 4,5-position (Scheme 34.6). However,
the cis-diol spontaneously isomerizes to a trans-diol and S-oxidation of thiophenes
can lead to dimerization. Similarly, dihydroxylation products from pyridines are
probably unstable, undergoing spontaneous dehydration. Bicyclic benzothio-
phenes, benzofurans, and indoles are dihydroxylated in both the heterocyclic and
carbocyclic rings [201, 202] with the former being relatively unstable; indole cis-
dihydrodiol spontaneously dehydrates, forming indoxyl that autoxidizes to indi-
go [213], a reaction that has been investigated as a commercial route to indigo
production and also exploited as a useful test of oxygenase activity [214]. The
carbocyclic dihydroxylation products are more stable and can be isolated, while the
ratio of heterocyclic to carbocyclic products can be altered by substitution with
sterically bulky groups.
34.2.2.6 Using cis-Dihydrodiols in Synthesis

There are numerous examples of syntheses exploiting the cis-dihydrodiol configu-
ration and many have been cited in recent reviews. In this chapter we limit our
coverage to a few with historical significance and selected examples that elegantly
exploit additional features of the cis-dihydrodiol platform. Following the early
examples from the Ley group [185], the defined regio- and stereochemistry of cis-
dihydrodiols has lent itself to several studies in cyclitol, and carbasugar production.
Hudlickys early work [183] demonstrated that acetonide protection of the cis-
dihydrodiols opened up the reactivity of the cyclohexadiene ring, either for ring
opening in terpene and prostanoid synthesis or further reaction in the synthesis of
complex alkaloids [215]. A few years later [216] it was demonstrated that by bulking-up
the cis-diol [as the bis(tert-butyldimethylsilyl)ether] it was possible to drive the
stereochemistry of cyclopropanation at the adjacent double bond to generate a key
chiral intermediate in pyrethroid synthesis. The F39/D mutant of P. putida and a
recombinant toluene dioxygenase expressed in E. coli were shown to oxidize indene
to (–)-cis-(1S,2R)-indandiol, which can be directly converted into cis-(1S)-amino-(2R)-
indanol [217, 218], a key intermediate in the chemical synthesis of Mercks HIV-1
protease inhibitor Indinivir Sulfate [219]. Subsequently, directed evolution of toluene
dioxygenase was used to create a variant that produced less of the indene by-products,
1-indenol and 1-indenone, while maintaining the highest (–)-cis-(1S,2R)-indandiol
enantiomeric purity [220]. Although early work exploited Gibsons F39/D mutant
of P. putida to produce the natural toluene cis-dihydrodiol [183] it was soon
recognized that additional options were available from the halogenated cis-
dihydrodiols (Figure 34.2). Most syntheses since then have involved alkyl and
halo-benzene precursors [215].
1512
R R
R
OH
TDO
+ Bis-sulfoxide
dimers
S OH S
S
O
TDO
R R Bis-sulfoxide
S dimers
S
O
OH
H OH H
TDO + HO
OH H
H
S S S
H OH
TDO +
OH H
H O
O O HO
H
OH
OH
H OH
TDO
OH
N H N
H N H
H
Indoxyl
O
H
N
N
H
O
Indigo
Scheme 34.6 Oxidation of electron-rich heteroaromatic rings by toluene dioxygenase.

General transformations
Suzuki-Stille-Heck
coupling
Br
H
OH
Cycloadditions Claisen
rearrangements
OH
H
Epoxidation
dihydroxylation
aziridination
cyclopropanation
Specific transformations
Oxidative
cleavage
Br
H
OH
Electrophilic
Nucleophilic tethers
tether OH
O H
Figure 34.2 Reaction options for arene cis-dihydrodiols. Adapted from Reference [215].
Although chloro-, bromo-, and iodobenzene yield highly enantiomerically pure

(1S,2R)-3-halo cis-dihydrodiols, until recently it has been difficult to prepare the
unnatural enantiomers and regioisomers. For the former, the observation that
fluorobenzene and disubstituted benzenes can give rise to mixtures of enantiomers
has been exploited together with biological and chemical resolution methods [221].
For biological resolution, the specificity of the cis-dihydrodiol dehydrogenase, the
second enzyme in the pathway, can be exploited to remove the natural enantiomer
(Scheme 34.7a) [222] and in recombinant strains expressing both dioxygenase and
dehydrogenase this can yield the unnatural enantiomer (albeit non-stoichiometri-
cally and mixed with the catechol). Diol dehydrogenases with different enantiopre-
ference are available that could be used to produce designer combinations of
dioxygenase with low stereospecificity [188], to yield both enantiomers in separate
single-pot biotransformations. A chemical approach to the same end is to produce
mixed diols from para substituted iodobenzene that can be separated and hydro-
genolyzed [223]. Strategies have also been devised to produce the equivalent trans-
diols (Scheme 34.7c) from the corresponding cis-diol [224, 225].
As yet, it has not been possible to engineer a dioxygenase that is specific for
3,4-dihydroxylation, but it is possible to produce the equivalent of both 3,4- and
1,2-dihydroxylation by conversion of the natural diol product via a three-step
synthesis (Scheme 34.7c) and conversion of benzene dioxide intermediates [224,
225]. Alternatively, application of the biotransformation and hydrogenolysis route to
ortho-substituted iodobenzene can generate both enantiomers of these unnatural
regioisomers (Scheme 34.7b) [223].
1514
(a)
R R
H
OH OH
2 +
R R
R H H OH OH
OH OH H
1
+
R
3 H R
OH OH OH
H H OH
+
OH OH
H
(b)
H
OH
5
OH
H
R3
R1,R2=H
I I H H
R1 OH OH
R1 R1,R3=H
4
R2 OH R2 OH
R2 H
H
R3 R3
R2,R3=H
H
R1 OH
(c) OH
H
R
R R R
H H
6
6
H O
5 Steps O H
3 Steps
OH H H
HO OH
O H O H H
H H R H OH
H
OH
R R OH
HO O H R OH
H R
OH H O H
6 6 OH
H O 5 Steps 3 Steps H
O
H
H
Scheme 34.7 Strategies to obtain unnatural using syn and anti benzene dioxides. [1] Low
regio- and stereoisomers of dihydrodiols, selectivity dioxygenase (e.g., chlorobenzene
starting with natural cis-dihydrodiols: dioxygenase) or R ¼ F; [2], for example, benzene
(a) enzymatic resolution of mixed isomers using diol dehydrogenase; [3], for example,
diol dehydrogenases; (b) exploiting the naphthalene diol dehydrogenase; [4] toluene
dominance of a bulky and removable iodine dioxygenase; [5] H2, Pd/C; [6] Pd(OAc)2, CO,
substituent in the regioselectivity of TDO; K2CO3, THF, H2O.
(c) interconversion of regio- and stereoisomers
More recently, there has been a trend towards utilization of more complex
substrates for biotransformation and, with the recognition that altering the specificity
of dioxygenases is feasible, this is likely to continue. meta-Dibromobenzene was used
as the precursor for the alkaloid narciclasine [226] and although benzoate is not a
good substrate for TDO, benzoate esters are dihydroxylated and have been used as
precursors for pseudo-sugar synthesis and Tamiflu [227]. Naphthalene has been
converted into the 1,2-diol as a precursor for ( þ )-gonodiol [228] and the tricyclic
heteroarene dictamine was dihydroxylated on the carbocycle to yield a synthetic
precursor for a range of furoquinoline alkaloids [197, 198]. Boyds group has also
demonstrated the generation of tetra-hydroxylated products via the dihydroxylation
of acetonides produced from initial cis-dihydrodiol formation [211, 212].
34.2.2.7 Catechols
The combination of dioxygenase activity and diol dehydrogenase activity in the same
recombinant strain, or mutation in the ring opening enzymes, will yield catechols with
the same 2,3-substitution as seen with dihydrodiol formation [229]. A 3,4-substitution
pattern is potentially more valuable, being found in L-DOPA, adrenaline, and nor-
adrenaline. This can be generated by combining two monooxygenase steps, with the
first exhibiting specificity for the 4-(para)-position, for example, toluene 4-monoox-
ygenase. Although several aromatic monooxygenases (see below) can catalyze the
sequential two-step oxidation of arenes to catechols, the second step is generally slower
than initial monohydroxylation, meaning that long incubation times are required for
high catechol yields. Additionally, two sequential monooxygenase steps create high
oxygen and NAD(P)H demands, which are undesirable for an intensified process.
Nolan and OConnor [230] combined the activities of T4MO with tyrosinase, which
catalyzes the cofactor-independent oxidation of phenols to catechols. Although this
reduced NAD(P)H requirement, the tyrosinase catalyzed further oxidation of catechols
to o-quinones was difficult to avoid. In practice, despite the increased cofactor
requirement, a 4-monooxygenase followed by a specific phenol hydroxylase [108]
catalyzed oxidation is probably a better strategy to achieve 3,4-catechols, although
evidence is emerging for a class of tyrosinase with a high tyrosine/L-DOPA oxidation
ratio [231]. Although there are no good examples to date, there is no fundamental
reason why the specificity of diol dehydrogenases could not be modified by protein
engineering to convert 3,4-dihydrodiols into their corresponding catechols.
Although direct biocatalytic production of catechols is clearly feasible, their toxicity
to the producing organisms requires separation in situ, to maintain high specific
productivity. This can be achieved in a two-phase (aqueous–organic) system, par-
ticularly if the two phases are separated by a hydrophobic membrane [232, 233].
34.2.3
Monohydroxylations
Aromatic monohydroxylation may be carried out using a wide range of

monooxygenases, some of which are specific for arene oxidation while others
1516
fortuitously oxidize aromatic compounds. Prokaryotic monooxygenases such as

toluene 4-monooxygenase [73], which are specifically involved in arene mono-
hydroxylation, are typically binuclear non-heme iron enzymes with similarities to
methane monooxygenase. However, the active oxygen species generated for
arene oxidation is not as reactive as that produced by methane monooxygen-
ase [234] and is unable to oxidize unactivated aliphatic C–H bonds. It follows that,
where the required regioselectivity exists, aromatic monohydroxylation is best
catalyzed by arene-specific monooxygenases as this reduces the chance of non-
specific hydroxylation of aliphatic side chains. However, as with dioxygenases,
where substrate binding allows, arene monooxygenases will also attack more
easily oxidizable substituents, such as sulfoxide, methoxy and benzyl posi-
tions [235]. Thus, the actual choice of enzyme for any particular reaction will
need to be a combination of both appropriate reaction specificity and sterically
controlled regioselectivity. In eukaryotes and some prokaryotes aromatic hydrox-
ylation is carried out by a cytochrome P450. Although cytochrome P450s are
unable to oxidize methane, they are commonly associated with unactivated
alkane oxidation [236]. Given that all variants of P450 appear to produce the
same type of active oxygen species, believed to be an FeIV ¼ O species, unlike
the non-heme iron proteins non-specific oxidation will be a persistent problem as
the only selectivity available is determined by substrate orientation within the
active site. However, a much wider range of diversity in natural substrates is
available in cytochrome P450s, meaning that with more bulky substrates they are
likely to be the catalysts of choice [11]. Cytochrome P450s also have a greater
potential for use in vitro, particularly using the peroxide shunt [237], and are
simpler single component enzymes [238].
A further level of reaction specificity is evident in enzymes that specifically
hydroxylate phenols. Although both non-heme iron and cytochrome P450 mono-
oxygenases can hydroxylate phenols they also hydroxylate unactivated arenes. How-
ever, two groups of enzymes specifically require the presence of the initial hydroxyl
group to convert phenols into catechols, which are natural intermediates in their
associated metabolic pathways. The first group are the flavoprotein monooxygenases
such as 4-hydroxybenzoate hydroxylases from P. fluorescens, which converts
4-hydroxybenzoate into 3,4-dihydroxybenzoate [99], phenol hydroxylases from G.
thermoglucosidasius A7 [108], and p-hydroxyphenylacetate 3-hydroxylases [109, 110].
The second group is the copper-containing tyrosinases that catalyze cofactor-
independent o-hydroxylation of phenols to catechols, but tend to further oxidize these
through to quinones [239], although more recent work suggests that separate
tyrosinase and catechol oxidase activities might exist in some organisms [231].
Electron distribution influenced by the presence of the initial hydroxyl group would
favor the second hydroxylation at both ortho and para positions; consequently, it is
clear that these natural phenol hydroxylases are sterically or mechanistically (in the
case of tyrosinases) constrained to give ortho-hydroxylation. Where non-specific
enzymes such as P450 hydroxylate phenols, steric interactions can give a prepon-
derance of para-hydroxylation, although typically some ortho-hydroxylation is also
observed [237].
34.2.4
Side Chain Oxidation
It has already been stated that electron-rich substituents such as sulfides attached to
aromatic rings may be oxidized selectivity or concomitantly with aromatic ring
hydroxylation by both monooxygenases and dioxygenases. In many cases this is
simply a reflection of the greater ease of hydroxylation of these moieties, possibly
involving an oxidation state earlier in the natural catalytic cycle, together with an
aromatic vehicle for enzyme recognition and binding. However, the benzylic
position in alkyl arenes and some other substituents attached to an aromatic ring
can be electronically activated, giving them a unique reactivity. This leads to two
useful types of monooxygenase, those that exploit this reactivity to give reaction
specificity and those that exploit steric effects to mask the most reactive site from
attack.
Xylene monooxygenase from P. putida mt-2 [240] and cymene monooxygenase
from P. putida F1 [241] are interesting case studies. Both clearly have methyl group
binding sites that sterically constrain the substrate to enable hydroxylation to be
directed to the benzylic position. Indeed, xylene monooxygenase has been exploited
by Lonza for the oxidation of 2,5-dimethylpyrazine to 5-methylpyrazine-2-carboxylic
acid with whole cells of P. putida mt-2 [242]. It was, therefore, unsurprising to find
that these enzymes both selectively epoxidize styrene. As non-heme iron enzymes
they can produce a form of activated oxygen that is capable of unactivated CH bond
oxidation, which is more difficult than double bond epoxidation. The FAD enzyme
styrene monooxygenase, on the other hand, produces a less reactive peroxy-flavin
intermediate that is capable of styrene oxidation but incapable of benzyl hydroxyl-
ation [107, 214]. Intriguingly, styrene monooxygenase is a poor epoxidizer of isolated
alkenes and nature typically exploits a de-activated (to discriminate between alkene
and alkane) form of the non-heme iron enzyme for general alkene epoxidation.
Therefore, although many enzymes are capable of styrene epoxidation, the combi-
nation of greater simplicity of the styrene monooxygenase and the greater reaction
selectivity would make this the enzyme of choice, particularly where protein
engineering is envisaged. All of the enzymes considered preferentially produce the
(S)-enantiomer, often with very high enantiomeric excess. Notably, as a rider to this,
however, styrene epoxidation is frequently the predominant reaction of more reactive
monooxygenases, suggesting reaction selectivity consistent with utilization of the
peroxy-iron intermediate in the catalytic cycle.
34.2.5
Products from Ring-Cleavage Reactions
As yet, there are few examples of the application of ring cleavage enzymes in synthetic
biocatalysis. However, given the probability that biocatalysis and metabolic pathway
engineering will play an increasingly important role as green chemistry comes to
the fore this situation is likely to change. It is probably also true that the ring cleavage
dioxygenases are less familiar as biocatalytic tools.
1518
A generic application of extradiol dioxygenases is in the formation of picolinic

acids, which can subsequently be converted into pyridines by exploiting the spon-
taneous re-aromatization of hydroxymuconic semialdehydes (Scheme 34.8a). Asano
et al. [243] used C23O to convert catechol, 3-methyl, 4-methyl, and 4-chlorocatechol
into their respective picolinic acids. He and Spain [244] subsequently demonstrated
that AP16DO could produce picolinic acid directly from 2-aminophenol, without the
addition of ammonia. There is also an elegant example of the use of this approach to
synthesize a more complex picolinic acid from diphenylacetylene (DPA). This
involved screening several toluene-degrading bacteria for the ability to produce a
yellow color, typical of meta-ring cleavage products from DPA. The identity of the
expected intermediate was confirmed and it was converted into 6-phenylacetylene
picolinic acid, which has potential as an improved (cf. DPA) crosslinker or dienophile
for thermosetting polymers [245].
(a)
NH3
OH OH
C23O
COOH
OH N COOH
CHO
2 Hydroxymuconic acid semi-aldehyde Picolinic acid
(b)
OH
C12O COOH Pt/C, H2
COOH
HOOC
COOH
OH
cis,cis muconic acid
Scheme 34.8 Useful ring-opening biotransformations of catechol: (a) via catechol 2,3-dioxygenase
to picolinates; (b) via catechol 1,2-dioxygenase to adipic acid.
cis,cis-Muconic acid, the product of ortho-cleavage of catechol (by C12O) can be

catalytically reduced to adipic acid, the precursor of nylon-6,6 (Scheme 34.8b). Whole-
cell processes for production of cis,cis-muconic acid from toluene, benzoic acid, and
catechol have been described [246, 247], together with direct conversion of catechol by
immobilized C12O [248]. In anticipation of the switch from petrochemicals to
renewables, Frost et al. [249] also used C12O in a metabolic engineering strategy
to produce cis,cis-muconic acid from glucose, via the shikimate pathway.
Although not the most carbon efficient route, van der Werf et al. found that the maleate
hydratase (malease, EC 4.2.1.31) from P. pseudoalcaligenes NCIMB 9867 converted
maleate into D-malate, a useful chiral synthon [250], with an enantiomeric excess of
more than 99.97% e.e. [251]. This organism uses GDO to cleave gentisate, producing
maleylpyruvate, which was subsequently hydrolyzed to pyruvate and maleate.

Enzyme Catalysis in Organic Synthesis, Third Edition (2012) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Enzyme Catalysis in Organic Synthesis, Third Edition (2012) PDF

Uploaded by

Copyright:

Available Formats

Edited by

Karlheinz Drauz, Harald Gröger,

Enzyme Catalysis in Organic

Whittall, J., Sutton, P.

Gotor, V., Alfonso, I., García-Urdiales, E. (eds.)

Crabtree, R. H. (ed.) Enzymes in Industry

Fessner, W.-D., Anthonsen, T. (eds.)

Enzyme Catalysis in Organic Synthesis

With a Foreword by Herbert Waldmann

Third, Completely Revised and Enlarged Edition

Prof. Dr. Harald Gröger British Library Cataloguing-in-Publication Data

All rights reserved (including those of translation

Cover Design Formgeber, Eppelheim

Printed in the Federal Republic of Germany

Print ISBN: 978-3-527-32547-4

Part I Principles of Enzyme Catalysis 1

1 Introduction – Principles and Historical Landmarks of Enzyme

1.4 Chemical Processes with Isolated Enzymes: The Impact

4 Rational Design of Enzymes 89

5 Directed Evolution of Enzymes 119

5.2 Short History of Directed Evolution 119

6 Production and Isolation of Enzymes 191

6.5 Extraction of Enzymes 197

7 Reaction and Process Engineering 217

7.4 Process Operation 231

Part II Hydrolysis and Formation of C–O Bonds 249

8 Hydrolysis and Formation of Carboxylic Acid Esters 251

8.2.2.6 Resolution of Carboxylates with a Hydroxy-(or oxo-)

8.7.2.2 Desymmetrization of 2,2-Disubstituted 1,3-Diols 342

9 Hydrolysis and Formation of Epoxides 363

9.3.7 Towards 100% Yield 403

10 Hydrolysis and Formation of Glycosidic Bonds 417

11 Addition of Water to C¼C Bonds and its Elimination 467

11.3.1.7 2-Hydroxy-4-Dienoate Hydratases 478

12 Industrial Application and Processes Forming C–O Bonds 503

Part III Hydrolysis and Formation of C–N Bonds 531

13 Hydrolysis of Nitriles to Amides 533

14 Hydrolysis of Nitriles to Carboxylic Acids 545

15 Hydrolysis of Amides 561

16 Hydrolysis and Formation of Hydantoins 651

16.5 L-N-Carbamoylase 659

17 Hydrolysis and Synthesis of Peptides 675

17.3.6.2 Cationic Substrate Mimetics 720

18 C–N Lyases Catalyzing Addition of Ammonia, Amines,

18.5.10 L-Threonine Dehydratase/Deaminase 767

19 Application of Transaminases 779

20 Industrial Applications and Processes Using Enzymes

Part IV Formation and Cleavage of C–C Bonds 855

21 Aldol Reactions 857

22 Acyloin and Benzoin Condensations 919

23 Cleavage and Formation of Cyanohydrins 947

23.8 Resolution of Racemates 973

24 Industrial Application and Processes Using Carbon–Carbon Lyases 991

Part V Hydrolysis and Formation of P–O Bonds 1001

25 Hydrolysis and formation of P–O Bonds 1003

25.5 Phosphorylases 1021

Part VI Reductions 1035

26 Reduction of Ketones and Aldehydes to Alcohols 1037

26.3.4 Enzyme-Coupled Cofactor-Regeneration Using a Glucose

27 Reduction of C¼C Double Bonds 1111

27.5 Accessing Both Product Enantiomers 1150

28 Reductive Amination of Keto Acids 1165

29 Industrial Application of Oxidoreductase catalyzed Reduction