Novel enzyme technology for food applications

Modifying lipids for use in food (ISBN 9781855739710) Any oil or fat should have the optimum physical, chemical, and nutritional properties dictated by its end use. Modification of natural fats and oils is therefore important to improve the quality of lipids for use in foods. When lipids are modified, though, compromises have to be made as the physical, chemical and nutritional properties of lipids are not always mutually compatible and this provides an important challenge for food technologists. Edited by an eminent specialist, this collection shows how these challenges have been met in the past, how they are being met today, and how they may be met in the future. Starch in food – Structure, function and applications (ISBN 9781855737310) Starch is both a major component of plant foods and an important ingredient for the food industry. Starch in food reviews starch structure and functionality and the growing range of starch ingredients used to improve the nutritional and sensory quality of food. Part I illustrates how plant starch can be analysed and modified, with chapters on plant starch synthesis, starch bioengineering and starchacting enzymes. Part II examines the sources of starch, from wheat and potato to rice, corn and tropical sources. The third part of the book looks at starch as an ingredient and how it is used in the food industry. There are chapters on modified starches and the stability of frozen foods, starch–lipid interactions and starchbased microencapsulation. Part IV covers starch as a functional food, including the impact of starch on physical and mental performance, detecting nutritional starch fractions and analysing starch digestion. Proteins in food processing (ISBN 9781855737235) Proteins are essential dietary components and have a significant effect on food quality. Edited by a leading expert in the field and with a distinguished international team of contributors, Proteins in food processing reviews how proteins may be used to enhance the nutritional, textural and other qualities of food products. After two introductory chapters, the book first discusses sources of proteins, examining the caseins, whey, muscle and soy proteins and proteins from oilproducing plants, cereals and seaweed. Part II illustrates the analysis and modification of proteins, with chapters on testing protein functionality, modelling protein behaviour, extracting and purifying proteins and reducing their allergenicity. A final group of chapters is devoted to the functional value of proteins and how they are used as additives in foods.

Starch in food – Structure, function and applications

(ISBN 978-1-85573-731-0)

Starch is both a major component of plant foods and an important ingredient for the food

industry Starch in food reviews starch structure and functionality and the growing range of

starch ingredients used to improve the nutritional and sensory quality of food Part I illustrates how plant starch can be analysed and modified, with chapters on plant starch synthesis, starch bioengineering and starch-acting enzymes Part II examines the sources of starch, from wheat and potato to rice, corn and tropical sources The third part of the book looks at starch as an ingredient and how it is used in the food industry There are chapters

on modified starches and the stability of frozen foods, starch–lipid interactions and starch-based microencapsulation Part IV covers starch as a functional food, including the impact of starch on physical and mental performance, detecting nutritional starch fractions and analysing starch digestion.

Proteins in food processing

(ISBN 978-1-85573-723-5)

Proteins are essential dietary components and have a significant effect on food quality Edited by a leading expert in the field and with a distinguished international team of

contributors, Proteins in food processing reviews how proteins may be used to enhance the

nutritional, textural and other qualities of food products After two introductory chapters, the book first discusses sources of proteins, examining the caseins, whey, muscle and soy proteins and proteins from oil-producing plants, cereals and seaweed Part II illustrates the analysis and modification of proteins, with chapters on testing protein functionality, modelling protein behaviour, extracting and purifying proteins and reducing their allergenicity A final group of chapters is devoted to the functional value of proteins and how they are used as additives in foods.

Details of these books and a complete list of Woodhead’s titles can be obtained by:

• visiting our web site at www.woodheadpublishing.com

• contacting Customer Services (e-mail: sales@woodhead-publishing.com; fax: +44 (0)

1223 893694; tel.: +44 (0) 1223 891358 ext 130; address: Woodhead Publishing Ltd, Abington Hall, Abington, Cambridge CB21 6AH, England)

Novel enzyme technology for food applications

Edited by Robert Rastall

CRC Press Boca Raton Boston New York Washington, DC

Cambridge England

Published by Woodhead Publishing Limited, Abington Hall, Abington,

Cambridge CB21 6AH, England

www.woodheadpublishing.com

Published in North America by CRC Press LLC, 6000 Broken Sound Parkway, NW, Suite 300, Boca Raton, FL 33487, USA

First published 2007, Woodhead Publishing Limited and CRC Press LLC

© 2007, Woodhead Publishing Limited Chapters 12 and 14 were prepared by

US government employees; those chapters are therefore in the public domain

and cannot be copyrighted.

The authors have asserted their moral rights.

Every effort has been made to trace and acknowledge ownership of copyright The publishers will be glad to hear from any copyright holders whom it has not been possible

to contact.

This book contains information obtained from authentic and highly regarded sources Reprinted material is quoted with permission, and sources are indicated Reasonable efforts have been made to publish reliable data and information, but the authors and the publishers cannot assume responsibility for the validity of all materials Neither the authors nor the publishers, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book.

Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or

by any information storage or retrieval system, without permission in writing from Woodhead Publishing Limited.

The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying Trademark notice: Product or corporate names may be trademarks or registered trade- marks, and are used only for identification and explanation, without intent to infringe British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.

Library of Congress Cataloging in Publication Data

A catalog record for this book is available from the Library of Congress.

Woodhead Publishing ISBN 978-1-84569-132-5 (book)

Woodhead Publishing ISBN 978-1-84569-371-8 (e-book)

CRC Press ISBN 978-1-4200-4396-9

CRC Press order number WP4396

The publishers’ policy is to use permanent paper from mills that operate a

sustainable forestry policy, and which has been manufactured from pulp

which is processed using acid-free and elementary chlorine-free practices.

Furthermore, the publishers ensure that the text paper and cover board used

have met acceptable environmental accreditation standards

Typeset by Ann Buchan (Typesetters), Middlesex, England

Printed by TJ International Limited, Padstow, Cornwall, England

Contributor contact details xi

Preface xv

Part I Principles of industrial enzyme technology 1 Discovering new industrial enzymes for food applications 3

Thomas Schäfer, Novozymes A/S, Denmark 1.1 Introduction 3

1.2 Where to screen for new enzymes 4

1.3 How to screen for new enzymes 8

1.4 Summary: which option to choose? 13

1.5 References 13

2 Improving enzyme performance in food applications 16

Ronnie Machielsen, Sjoerd Dijkhuizen and John van der Oost, Wageningen University, The Netherlands; Thijs Kaper and Loren Looger, Carnegie Institution of Washington, USA 2.1 Introduction 16

2.2 Laboratory evolution 17

2.3 Examples of improving enzyme stability and functionality by laboratory evolution 24

2.4 Rational and computational protein engineering 28

2.5 Examples of improving enzyme stability and ability by rational protein engineering 30

2.6 Examples of combined laboratory evolution and computational design 34

vi Contents

2.7 Summary and future trends 35

2.8 Sources of further information and advice 35

2.9 References 36

3 Industrial enzyme production for food applications 43

Carsten Hjort, Novozymes A/S, Denmark 3.1 Introduction 43

3.2 Traditional sources and processes for industrial enzyme production 44

3.3 Design of expression systems for industrial enzyme production 46

3.4 Development of an enzyme production process 54

3.5 Future trends 56

3.6 Sources of further information and advice 56

3.7 References 57

4 Immobilized enzyme technology for food applications 60

Marie K Walsh, Utah State University, USA 4.1 Introduction 60

4.2 Immobilized enzyme technology for modification of acylglycerols 62

4.3 Immobilized enzyme technology for modification of carbohydrates 68

4.4 Immobilized enzyme technology protein modification 73

4.5 Immobilized enzyme technology for production of flavor compounds 75

4.6 Future trends 77

4.7 References 78

5 Consumer attitudes towards novel enzyme technologies in food processing 85

Helle Søndergaard, Klaus G Grunert and Joachim Scholderer, MAPP, University of Aarhus, Denmark 5.1 Introduction 85

5.2 Theoretical approaches to how consumers form attitudes to new food production technologies 86

5.3 Studies of consumer attitudes to enzyme technologies 88

5.4 Implications of consumer attitudes to enzyme technologies 94

5.5 Future trends 95

5.6 Sources of further information and advice 95

5.7 Acknowledgements 96

5.8 References 96

Part II Novel enzyme technology for food applications

6 Using crosslinking enzymes to improve textural and

other properties of food 101

Johanna Buchert, Emilia Selinheimo, Kristiina Kruus, Maija-Liisa Mattinen, Raija Lantto and Karin Autio, VTT, Finland 6.1 Introduction 101

6.2 Types of crosslinking enzymes 103

6.3 Application of crosslinking enzymes in baking and pasta manufacture 109

6.4 Application of crosslinking enzymes in meat and fish processing 114

6.5 Application of crosslinking enzymes in dairy applications 118

6.6 Other applications of crosslinking enzymes in food manufacture 122

6.7 Analysing the chemistry of crosslinks formed by enzymes 122

6.8 Effect of biopolymer crosslinking on nutritional properties of food 124

6.9 Conclusions 126

6.10 References 126

7 Enzymatically modified whey protein and other protein-based fat replacers 140

Jacek Leman, University of Warmia and Mazury in Olsztyn, Poland 7.1 Introduction 140

7.2 Enhancing the fat mimicking properties of proteins 142

7.3 Applications in low-fat foods 149

7.4 Future trends 152

7.5 References 153

8 Enzymatic production of bioactive peptides from milk and whey proteins 160

Paola A Ortiz-Chao and Paula Jauregi, University of Reading, UK 8.1 Introduction 160

8.2 Angiotensin I-converting enzyme inhibitory peptides 161

8.3 Other bioactive peptides and their health benefits 165

8.4 Production of bioactive peptides from milk and whey proteins 170

8.5 Future trends 177

8.6 Sources of further information and advice 177

8.7 References 177

viii Contents

9 Production of flavours, flavour enhancers and other protein-based

speciality products 183

Stuart West, Biocatalysts Ltd, UK 9.1 Introduction 183

9.2 Production and usage of monosodium glutamate (MSG) 186

9.3 Chondroitin sulphate 188

9.4 Production of aspartame 190

9.5 Enzymes for vanilla extraction 191

9.6 Enzyme modified cheese as a flavour ingredient 193

9.7 Enzymes used in savoury flavouring 198

9.8 Enzymes used in yeast extract manufacture 199

9.9 Future trends 200

9.10 Sources of further information and advice 202

9.11 References 203

10 Applications of cold-adapted proteases in the food industry 205

A Guðmundsdóttir and J Bjarnason, University of Iceland, Iceland 10.1 Introduction 205

10.2 Use of proteolytic enzymes in food processing 208

10.3 Application of cold-adapted serine proteases in food processing 209

10.4 Modifying marine proteases for industrial use 211

10.5 Future trends 212

10.6 References 212

11 Health-functional carbohydrates: properties and enzymatic manufacture 215

Simon Hughes and Robert A Rastall, University of Reading, UK 11.1 Introduction 215

11.2 Dietary fibre 215

11.3 Prebiotics 217

11.4 Inulin 219

11.5 Transgalacto-oligosaccharides 222

11.6 Gluco-oligosaccharides 223

11.7 Alternansucrase–maltose acceptor oligosaccharides 224

11.8 Resistant starch 226

11.9 Arabinoxylan 228

11.10 Oligosaccharides from non-starch polysaccharides 230

11.11 Pectins 232

11.12 Oligodextran 234

11.13 Conclusion 237

11.14 References 237

12 Flavorings and other value-added products from sucrose 243

Gregory L Côté, United States Department of Agriculture, USA 12.1 Introduction 243

12.2 Di- and oligosaccharides from sucrose 244

12.3 Polysaccharides from sucrose 257

12.4 Other products 260

12.5 Future trends 261

12.6 Sources of further information and advice 262

12.7 References 262

13 Production of structured lipids with functional health benefits 270

Xuebing Xu, Janni B Kristensen and Hong Zhang, BioCentrum-DTU, Technical University of Denmark, Denmark 13.1 Introduction 270

13.2 Production of diglyceride oils 271

13.3 Production of healthy oils containing medium chain fatty acids 278 13.4 Future trends 282

13.5 Acknowledgements 282

13.6 References 282

14 Lipase-catalyzed harvesting and/or enrichment of industrially and nutritionally important fatty acids 285

George J Piazza and Thomas A Foglia, US Department of Agricul-ture, USA; and Xuebing Xu, BioCentrum-DTU, Technical University of Denmark, Denmark 14.1 Introduction 285

14.2 Lipase selectivity 286

14.3 Fatty acid harvesting 294

14.4 Structured triacylglycerols 295

14.5 Single reaction step process for the production of STAG 301

14.6 Multiple reaction step processes for the production of STAG 307

14.7 Nutritional and other uses of structured lipids 307

14.8 Summary and future trends 308

14.9 References 309

Index 315

R Machielsen* and J van der Oost

Hesselink van Suchtelenweg 4

email: cahj@novozymes.com

Chapter 4

M K WalshUtah State University

8700 Old Main HillNFS 318

Logan

UT, 84322-870USA

email: mkwalsh@cc.usu.edu

Chapter 5

H Søndergaard*, K G Grunert and

J ScholdererMAPP

Aarhus School of BusinessHalslegaardsvej 10DK-8210 Aarhus VDenmark

email: hals@asb.dk

xii Contributor contact details

Chapter 6

J Buchert*, E Selinheimo, K Kruus,

M L Mattinen, R Lantto and K

Faculty of Food Sciences

University of Warmia and Mazury in

P A Ortiz-Chao and P Jauregi*

School of Food Biosciences

Chapter 10

A Guðmundsdóttir*

Science InstituteUniversity of IcelandLæknagardi

Vatnsmýrarvegi 16

101 ReykjavíkIcelandemail: ag@raunvis.hi.is

J B BjarnasonDunhaga 3

107 ReykjavíkIceland

Chapter 11

S Hughes and R A Rastall*School of Food Biosciences

PO Box 226WhiteknightsReadingRG6 6APUKemail: r.a.rastall@reading.ac.uk

Chapter 12

G CôtéNCAUR/ARS/USDA

1815 N University StPeoria

IL 61604USAemail: greg.cote@ars.usda.gov

Agricultural Research Service

Eastern Regional Research Center

600 East Mermaid LaneWyndmoor

PA 19038USAemail: george.piazza@ars.usda.gov

X XuBioCentrum-DTUTechnical University of DenmarkBuilding 227

DK-2800 Kgs LyngbyDenmark

email: xx@biocentrum.dtu.dk

Enzymes have been used in the food industry for many years They have largelybeen used as processing aids and they have many attributes that make them fit forthis purpose They are generally non-toxic and speed up chemical reactions withgreat specificity at low temperatures and pressures and at near-neutral pH A largeindustry exists to serve this need across the world.

One of the limitations of enzyme application in the food industry is the lack ofavailability of enzymes with the required properties at an acceptable price Whilstdesired enzyme activities are frequently known somewhere in the biologicalworld, they are often unsuitable for commercial application In recent years,however, there has been increasing sophistication in our ability to isolate novelenzymes from biological sources and an expansion of the range of sources ofenzymes to include, for example, extremophiles Such organisms frequently haveenzymes with higher pH and temperature optima and can extend the range ofprocesses in which enzymes can be used We now have the ability to rationallyengineer or artificially evolve desired catalytic properties into enzyme molecules.These new technologies will ultimately remove many of the limitations cur-rently restricting the application of enzymes in the food industry and will open upmany more possibilities Technological aspects are dealt with in Part I – Principles

of industrial enzyme technology Chapters 1 and 2 deal with the discovery of novelenzymes for food applications and the improvement of enzymes for food applica-tions Chapters 3 and 4 then examine the production of industrial enzymes and theirimmobilisation in the context of food applications Part I is concluded by Chapter

5 on consumer attitudes to novel enzyme technologies

Concurrent with these technological developments has been the advance in ourknowledge of the role of specific food components in health and disease This hasled to a significant increase in interest in functional food ingredients – compoundsthat are specifically added (or whose levels are deliberately increased) in foods to

xvi Preface

provide a specific health attribute beyond nutrition Examples include prebioticoligosaccharides to improve gut health, bioactive peptides to help reduce bloodpressure, and nutritionally enhanced fats Governments around the world are alsotaking heed of modern nutritional knowledge and are increasingly looking to thefood industry to manufacture foods with a healthier profile These nutritionaldevelopments are starting to provide a new range of application areas for novelenzymes and enzyme technologies and it is these applications that are discussed inPart II – Novel enzyme technology for food applications Chapters 6, 7 and 8 dealwith enzymatic modification of proteins to achieve cross-linking, to generate fatreplacers and to manufacture bioactive peptides respectively Protein modificationalso features in Chapter 9 on production of flavours and flavour enhancers and inChapter 10 on the application of novel cold-adapted proteases The focus thenmoves to carbohydrates, in Chapter 11 on health-functional carbohydrates andChapter 12 on value-added products from sucrose Finally, the manufacture oflipids with health and other functional attributes is discussed in Chapters 13 and 14.This volume aims to give the reader an overview of recent developments inenzyme technology in the food industry rather than an exhaustive account oftraditional applications The aim is to increase awareness of and stimulate interest

in developing novel enzyme technologies to meet the new and changing needs ofthe food industry

Professor Robert Rastall University of Reading

Part I

Principles of industrial enzyme technology

Discovering new industrial enzymes for food applications

Thomas Schäfer, Novozymes A/S, Denmark

Enzymes have been exploited by humans for thousands of years Traditional foodsand beverages like cheese, yoghurt and kefir, bread, beer, vinegar, wine and otherfermented drinks, as well as paper and textiles, were produced with the help ofenzymes which were present in starting materials as early as 6000 BC in China,Sumer and Egypt The epoch of classical biotechnology was marked by thelandmark discoveries of microbes by Leeuwenhook, of fermentations as biologicalprocesses by Pasteur, of enzymes as proteins by Buchner and of the first enzymecrystal structures by Sumner

The modern era of industrial enzymology began in 1913 when Otto Röhm wasgranted a patent for the use of a crude protease mixture isolated from pancreases inlaundry detergents In the following years an increasing number of enzymes werefound in microorganisms and these microbes were cultured in large scalefermentations to produce enzymes However, the number of enzymes that could beproduced in this fashion was limited, because not all microbes are amenable tolarge scale fermentation The pioneering work of Avery and MacLeod, Hersheyand Chase, Watson and Crick, Cohen and Boyer and many others who introducedthe era of recombinant biotechnology revolutionized industrial enzyme screeningand production

With the advent of genetic engineering, genes encoding interesting enzymescould be transferred to and expressed in selected host microbes for production on

an industrial scale Today, gene technology plays a major role in both the discovery

4 Novel enzyme technology for food applications

of novel enzymes and the optimization of existing proteins, and is basis for theproduction of the majority of industrial enzymes Food applications of enzymesrepresent a wide and highly diverse field including baking, dairy, juice, vegetableprocessing and meat The enzymes are used to obtain a number of benefits, likemore efficient processes, leading to reduced use of raw materials, improved orconsistent quality, replacement of chemical food additives and avoidance ofpotential harmful by-products in the food

1.1.1 Technologies for discovery of industrial enzymes

Nature holds a wonderful diversity of organisms and the corresponding wealth ofenzymes and has often been the starting point for the identification of novelenzymes For a variety of applications even Nature’s assortment faces somelimitations or it is too time consuming and difficult to look into Nature’s diversity.This imposes a challenge for scientists to optimize existing natural enzymes and togenerate additional ‘artificial’ diversity to tailor-make enzymes for a given appli-cation Natural diversity approaches and optimization strategies are complementaryroutes and both are equally important in developing a high-quality diversity of

enzymes (Nedwin et al., 2005).

Today, discovery of enzymes for the food industry is not only amultidisciplinary effort involving a wide array of different screening technolo-gies, but is also based on close interaction between food scientists whounderstand or model the application and biotechnologists who can deliver en-zymes for initial trials Each screening project is new and challenging.Accordingly, each project needs to be uniquely designed to solve the specifiedapplication problems in a certain industrial application and for each project theexpert team needs to have members with exactly the competencies needed tofind a solution It is obvious that major enzyme companies have to master avariety of technologies which, often in combination with each other, lead to thesolution For all approaches it is important to stress that it is not the broadestpossible diversity, but rather the highest possible quality of diversity which willlead to the ultimate goal, namely a novel product that addresses exactly thespecific demands of the industrial application In this respect selection/deselection via perfectly designed assays is of utmost importance, indicating thesignificance of linking process understanding to biochemistry

One of the main questions which has to be answered in the very beginning of each

discovery initiative is ‘where to look for diversity?’ (Bull et al., 2000; Fig 1.1).

There are various potential sources, as input to screening programs is basicallydivided into (a) natural enzyme diversity and (b) artificial diversity, which arecomprehensively reviewed by Schäfer and Borchert (2004) and Aehle (2004).Here the basic principles will be summarized

Fig 1.1 Overview of the main approaches to diversity input in screening programs.

1.2.1 Nature’s diversity: an unlimited source of enzymes

The challenge is that Nature’s diversity is virtually infinite and that living organisms have inhabited virtually all ecological niches on planet Earth during 3.5billion years of evolution The number of described bacterial and fungal species ishuge, new isolations are added daily so that the actual number can only beextrapolated roughly From bioinformatics analysis of the genomes it can beassumed that a bacterial genome on average contains about 4000 enzyme codinggenes, while for fungi the number of enzyme encoding genes can be up to 20,000

micro-(Hirose et al., 2000; Dunn-Coleman and Prade, 1998) The art of screening

obviously consists of having the right tools to find the ‘needle’ in this ‘haystack’

of biodiversity; no scientist will start looking into the totally available biodiversity,but will look into groups of carefully selected microorganisms Considering thesenumbers and using best practice, it is obvious that all screening efforts face alimitation in that we are only scratching the surface Microorganisms, namelybacteria, fungi and archaea, which are normally stored in culture collections of thegroups performing the screening or in public collections, where the strains areaccessible for everyone who is interested, often comprise the biological startingmaterial

On top of the cultivated diversity, complex gene libraries compiled from naturalmaterial without prior cultivation (Handelsman, 2005) can be generated and used

to discover industrial enzymes (Short, 1997) and other natural compounds (Brady

et al., 2001) Today, it is generally accepted that only minor numbers among the

whole of the microbial diversity have been cultured or might even be amenable to

growth in the laboratory (Torsvik et al., 2002) thereby leaving not only a huge set

of questions concerning our understanding of the role of microbes in their habitats,but also an enormous potential for yet undiscovered physiological and biochemical

6 Novel enzyme technology for food applications

traits including enzyme genes in the so-called metagenome (Lorenz and Schleper,

2002; Rondon et al., 1999) It is estimated that 1 g of soil contains more than 4000

different bacterial genomes, that is about 16 million enzyme encoding genes Byisolating the genetic material, be it DNA or RNA, directly from the soil and cloningthis into suitable host complexes, ‘environmental libraries’ can be constructed.These gene libraries need to be screened as described below using either sequence-based techniques or activity assays including some novel constraints caused by thecomplexity of the library

1.2.2 Bioinformatics and genomics

Input to screening efforts can also come from genes or genomes described byresearchers worldwide over time The updated status of established genomes andthose underway can be obtained by visiting the homepage of the TIGR institute(http://www.tigr.org/) or the homepage of the DOE Joint Genome Institute (http://genome.jgi.org/) The use of existing gene information can potentially shortcutthe flow to a new product candidate, although in most instances a gene is described

by a sequence only, that is, no biochemical information is available By usingsophisticated software tools within the new discipline of bioinformatics thosegenes can be aligned to existing ones, grouped into enzyme families in order topredict ideally their putative biochemistry, that is, enzyme activity (Henrissat andBork, 1996) This is also where one of the major pitfalls lies, namely that theoriginal description of the enzyme can turn out to be incorrect An additionalcomplexity is the fact that roughly 30% of all gene sequences from genomes arenew, that is they do not resemble any biochemical description of the correspondingprotein Interesting hits found in this way can subsequently be analysed in moredetail but this requires cloning and expression of the gene (see below) followed bypurification and characterization of the corresponding enzyme, which is a tediousand resource-intensive effort when many genes are of potential interest Accord-ingly, this comprises one of the major bottlenecks in genomics as the protein canonly be characterized very late in the process and the chance of failure is high.Searching of gene databases, both generated in-house and external ones, is a dailycomplement to the work of a screening scientist In addition to screening theexternal world of sequence data for novel enzymes, the discovery scientist mustalso determine whether any enzymes found are novel and whether their use isprotected by patents

Whole genome sequencing combined with bioinformatics, array studies andproteomics are novel key technologies for the targeted improvement of production

strains This has illustratively been described for lysine production in bacterium glutamicum (Ohnishi et al., 2002) Whole genome sequencing which

Coryne-completely maps all genes is, however, not ideal for discovery of selected genes,for example those encoding for enzymes and especially for those enzymes thatmatch defined application criteria Assuming an average genome size for a

bacterium of about 4 Mb, for yeast of 13 Mb (Zagulski, et al., 1998) and for

filamentous fungi in the order of 30–40 Mb (Dunn-Coleman and Prade, 1998;

Radford and Parish, 1997), the costs of sequencing programs of total genomes areunreasonably high for discovery purposes, especially considering the wide diver-sity of microbes that are interesting for enzyme screening From the 4100 open

reading frames (ORFs) of the Bacillus subtilis genome, only a fraction may be

relevant for industrial enzymes For many industrial applications, extracellular

enzymes are of major importance and it is estimated that B subtilis produces 150–

180 secreted proteins (Hirose et al., 2000), while the number of secreted enzymes

for filamentous fungi might be in the order of 200–400 corresponding to theirlarger genome sizes This indicates that only 2–5% of the ORFs in a completegenome are of primary interest for enzyme discovery Accordingly, whole genomesequencing can hardly be justified for enzyme discovery purposes As a conse-quence alternative approaches have been developed to mine selectively microbialgenomes for secreted enzymes Those will be described in more detail in Section1.3.4

1.2.3 Protein optimization of enzymes

In cases where enzymes found from natural sources cannot provide the ance needed for a given application, protein optimization offers an attractiveoption In many applications the enzymes are very much stressed by, for example,high temperatures, extreme pH values or the presence of metal ions, which areknown to induce unfolding of the protein Several strategies can be followed inorder to optimize the properties of enzymes found in Nature A simplified way oflooking at protein optimization technologies is to divide the field into rationalprotein engineering and random molecular evolution (Fig 1.1) This is discussedbriefly below and in more detail in Chapter 3

perform-Rational protein engineering is based on the knowledge of a given enzymestructure and the corresponding biochemistry, for example the substrate specificity,the temperature tolerance, inhibition by metal ions, and so on, the combinationthereof comprising the structure–function relationship This is the parameter thatwill be modified as changes in structure will lead to changes in functionality Thechallenge is to introduce changes that lead to improved functionality rather thaninferior variants of an enzyme A key is the ability to create protein variants withdesigned and deliberate amino acid alterations at any desired position that providesthe capability for precise probing of structure–function relationships in proteins(Bott, 2005)

The gene is mutated selectively at specified sites and the corresponding enzyme

is expressed and subsequently tested to verify the hypothesis behind the mutation.Positive mutations are collected and analysed in more detail, for example whichamino acids in the enzymes were changed, at the position where the mutations arelocated These mutations are eventually combined to find the ultimate combination

of positive mutation events As easy as this might sound, this approach represents

a considerable challenge for researchers and in many instances the experimentshave failed Several years of hands-on experience of biochemistry, bioinformaticsand structure–function analysis are a prerequisite for success Importantly, it must

8 Novel enzyme technology for food applications

be acknowledged that we still only have limited understanding of protein functionand that only a small part of the huge natural enzyme diversity has been analysed

on a structural level, resulting in severe limitations for rational protein engineering

We are still far from predictability, that is, knowing which amino acid change willresult in which consequence Years of trial and error have, however, increased ourknowledge, especially of selected enzyme classes which are of major importancefor industrial enzymes Accordingly, engineered variants of a number of hydro-lytic enzymes such as proteases, amylases, lipases and cellulases are commerciallyavailable today

In contrast, no prior knowledge of structure–function relationships is needed tocarry out molecular evolution experiments Here the basic principle is to carry outrandom introduction of mutations, thereby generating DNA libraries consisting of

up to millions of variant genes The DNA variation is expressed into protein

diversity in a variant library, for example in Escherichia coli or Saccharomyces cerevisiae, the library is subjected to a screening procedure using a functional

assay (see below) and the best performing mutants are collected There is a tightconnection between the selected variant protein and its encoding gene, whichmakes it easy to sequence the enzyme coding gene and detect the mutation, in thiscase after the modified phenotype was detected

Random mutation leads to millions of mutants of a given gene and smartscreening systems are needed to identify the best performing variants Roboticsequipment for colony picking, colony transfer into screen-able formats, oftenmicrotitre plates, and addition of assay components are a prerequisite for this

approach (Eijsink et al., 2005).

After the question ‘where to screen?’ has been answered the next question is ‘how

to screen?’ (Bull et al., 2000) There are several possibilities for this and variations

of these themes (Fig 1.2) The following paragraphs will describe some of themost prominent screening approaches

1.3.1 Functional biochemical assays

The most preferred screening route for novel enzymes is via functional screeningassays, where the biochemical activity can be detected Ideally, this will also showhow well the positive hits meet application requirements to be tested; a detectedamylase can, for example, also be tested for activity at elevated temperature, high

or low pH or under other hostile conditions by using the same assay principle.Biochemical assays allow the screening of living microorganisms, of gene librariesconstructed from cultivated microbes and from the environment (metagenomes),

of artificial evolution libraries, as well as of rationally designed protein variants for

a wanted enzyme activity Many of the assays can be implemented on agar plates,where growing colonies can be tested for activity, or on a smaller scale using

Fig 1.2 Overview of the main technologies for screening for industrial enzymes.

microtitre plates The latter is often a challenge but is a prerequisite for throughput screening which is needed especially for screening metagenome andartificial evolution libraries Accordingly, a lot of effort is invested in developingscreening assays and novel screening technologies including high-throughputtechnologies that give enhanced freedom in assay design Accordingly a variety of

high-publications and patent applications cover this field (Joo et al., 1999; Ruijssenaars and Hartmans, 2001; Meeuwsen et al., 2000; Preisig and Byng, 2001; Kongsbak

et al., 1999; Short and Keller, 2001; Schellenberger, 1997).

To include application-relevant parameters is very difficult, as the biophysicalmatrix in the target application is often very complex and it can be difficult toidentify relevant screening criteria, for example for enzymes in bread making.Potential enzyme candidates can be identified (for example an amylase), but it isdifficult to qualify these in small scale assays which mimic the application.Accordingly, full scale baking trials are the only way of evaluating the enzymecandidates, which is both a time and resource requiring approach To overcomethis, hypotheses concerning the most relevant application parameters are made andassays are developed which are as close to real conditions as possible Ideally,functional screening procedures help to select the best performers in a givenlibrary The best candidates might be chosen as product candidates directly, oralternatively they can form the starting point for a repeated round of the directedevolution cycle if a performance gap still exists In the second round, enzymesfrom natural diversity screens or protein engineered variants might be included ifthey show beneficial characteristics for a given application, again pointing at thefact that screening approaches are complementary and the combination of several

of them often leads to success

Expression

Proteomics

Protein production

Protein analytics Transcriptomics Gene regulation/expression

Secretomics Gene trapping technologies

Genomics/

Bioinformatics

Genome mapping and analysis

Biochemical screening assays

Sequence based screening

Screening output or ‘how to screen’

10 Novel enzyme technology for food applications

1.3.2 Primary and secondary screening

The above illustrates one of the main approaches to screening, that is, sequentialprimary and secondary screening This is the general pattern for all new screeningprogrammes First, screening criteria are defined which include the substrate to beconverted and conditions such as temperature, pH, presence or absence of ions.These are used to design functional screening assays that are able to (a) detect thedesired enzyme activity in the primary round and (b) select the best performingenzymes in the secondary round

The primary screening is often very broad and all microbes, clones or variantswhich are positive in the chosen assay are selected Accordingly, the primaryscreening assay has low selectivity in order to capture a wide variety of positivehits In the next round these hits are subjected to a secondary screening where ahighly selective enzyme assay is applied that allows ranking and selection of thebest-performing enzyme candidates It must be noted that screening is seldom asimple linear process but usually involves learning loops running in several cycles

1.3.3 Expression cloning for further characterization and application testing

Ideally, the enzymes that are tested in industrial applications are monocomponent,that is, free of other potentially misleading enzyme activities To deliver these pureenzymes, cloning of genes and expression of proteins is an important step In mostcases a technique called expression cloning is used which is an effective means ofisolating a gene from a gene library based on its encoded activity (Dalbøge andHeldt-Hansen, 1994) In brief, the genetic material, either DNA or mRNA, isisolated and purified, cut into small pieces (in case of DNA) to separate all genesphysically and randomly from each other, set into suitable vector systems, andtransferred into a screening host to form a gene library Host strains ideally do not

produce the targeted enzyme activity and often E coli or S cerevisiae are used for

this purpose In this step a 10 000–50 000 colony gene library is tested with thebiochemical assay, and single colonies expressing the cloned enzyme can beisolated Ideally the selected clones only produce the enzyme of interest Subse-quently the corresponding gene can be isolated, sequenced and used for furtheranalysis and optimization

Monocomponent enzymes produced by recombinant DNA technology arepreferred in small-scale applications to correlate measured effects clearly to agiven protein rather than a complex mixture of enzymes This is important whencomparing the protein with existing enzymes (Can the enzyme be patented? Is itbetter than the benchmark?), for further improvement by protein optimization andalso to obtain an initial idea of whether the enzyme can be produced undereconomically promising conditions (Chapter 3) Only at the point of working with

a pure enzyme can the hypothesis underlying the assay be tested, that is, do theselected candidates fail or pass the real application test? Both failure and successcan be used to optimize the assay and thereby generate even more and betterdiversity Accordingly the quality and nature of the screening assay has a central

role during the whole selection/deselection programme as the quality of the assaydetermines the quality of the resulting candidate The key to successful screening

of industrial enzymes is not to detect a large diversity of proteins, but rather thosefew that can perfectly match the application conditions Again, the importance ofcooperation between food scientists and the biochemists performing the screeningcannot be ranked highly enough

1.3.4 Molecular screening approaches for identification of genes and proteins

Sequence-based approaches, also described as molecular screening, complementfunctional screens They are based on similarities between enzyme-encoding gene

sequences (Dalbøge and Lange, 1998, Precigou et al., 2001) Sequence

information from a set of related enzyme genes is used to identify evolutionarilyconserved regions and to design polymerase chain reaction (PCR) primers toamplify genes from other organisms Using this method, a number of genes whichare homologous to the initial gene sequences can quickly be identified Thelimitation of the method is that enzyme variants rather than totally novel enzymesare detected The advantages, on the other hand, are that this method is notdependent on growing the strains in the laboratory or on the active expression of

a protein, which makes this an interesting option for screening metagenomiclibraries

Secretome studies, transcriptomics and proteomics are gaining more

impor-tance as screening tools (Nedwin et al., 2005) A fast and efficient approach for

trapping of genes which encode secreted enzymes from a genome (hencesecretomics) is transposon-assisted signal trapping (TAST) of gene libraries

(Duffner et al., 2001) A genomic or cDNA library is treated with a transposon

carrying a reporter gene which codes for a secreted protein with its own secretionsignal sequence removed A signalless beta-lactamase has been used as reporterwhich can, upon insertion in a gene with an active secretion signal, be transportedout of the cell as a fusion protein This results in ampicillin-resistant phenotypesthat can be selected on agar plates Genes encoding secreted proteins aresubsequently sequenced and identified by comparison to databases usingbioinformatic tools In contrast to traditional screening of gene libraries withfunctional assays for selected enzyme activities, the entire genome, as represented

in the library, is trapped for known and novel enzymes simultaneously on a genelevel The disadvantage of the method is that genes are identified but are notdirectly available for testing which means that all relevant ORFs have to beexpressed individually

Alternative approaches have been developed to mine microbial genomesselectively for secreted enzymes One alternative is to clone and sequence ex-pressed sequenced tags (ESTs) randomly This has been used to identify novel

enzymes in Fusarium (Berka et al., 2003) The principle is based on large-scale

isolation and (partial) sequencing of randomly selected, anonymous cDNA cloneswhich express the enzyme linked to the gene derived from the cDNA By

12 Novel enzyme technology for food applications

sequencing the gene and comparing it with other genes in the databases, previouslyunknown genes have been identified in a variety of organisms

When gene sequences are analysed from cells grown on different nutrientsources, it is possible to discover and catalogue novel enzymes that are producedspecifically on those nutrients For this kind of comparative analysis signaltrapping, EST analysis, transcriptomics using DNA microarrays or proteomicsmight be used

DNA microarray technology is a popular tool for studying gene expression and

gene regulation by monitoring mRNA formation on a genomic scale (DeRisi et al., 1997; Berka et al., 2003) This technology can also be applied in enzyme screening

as it is possible to detect genes induced under specific physiological conditions

(Yaver et al., 2003; Diehn et al., 2000; Bashkirova and Berka, 2003) In principle

the genome of the microorganism to be investigated needs to be known, all ORFs,that is enzyme-encoding genes, need to be amplified by PCR and plotted on acarrier material, for example a glass slide

When a strain is grown on different specific substrates, genes involved inutilization of that specific substrate are induced and RNA is collected The mRNAcan be isolated and labelled with fluorescent dyes of different colours, correspond-ing to, for example, the different growth substrates This differential, mediumdependent expression allows detection of genes that are important for degradation

of the particular substrate by comparing hybridization of differently labelledmRNAs to the DNA on the glass slide Using this technology, global gene

expression profiling of Ceriporiopsis subvermispora was performed to discover

novel peroxidases enzymes whose expression is induced during growth on

thermomechanical pulp (Yaver et al., 2003) Thus, shotgun genomic DNA

microarrays appear to be a viable approach for identifying novel enzymes involved

in the degradation of complex substrates

A method combining suppression subtractive hybridization (SSH) and DNAmicroarray techniques was used to identify biomass-induced genes in the cellulo-

lytic fungus Trichoderma reesei (Bashkiro and Berka, 2003) The degradation of

cellulosic biomass is the result of the concerted effort of many fungal enzymes,though only a few enzymes have been identified and characterized The cDNAlibraries generated by SSH allowed for the selection of differentially expressedmRNAs, as well as enrichment of rare mRNAs and equalization of the cDNA pool.DNA sequence analysis and bioinformatics were used to assemble the clones intoapproximately 90 previously unrecognized genes/proteins Thus, the combination

of SSH and cDNA microarray technologies has proved to be a useful tool fordiscovering new differentially expressed enzyme genes involved in biomassutilization

All these different genetic approaches have drastically increased the amount ofdata available in both industrial and public databases Additionally, hundreds ofgenome projects have led to an explosion of data, which again underlines the needfor new tools within the discipline of bioinformatics to compare, sort and finallyselect the most relevant data

1.4 Summary: which option to choose?

The discussion above has outlined the complexity, not only of the individualmethods which have to be mastered, but also of the interplay between thesetechniques As each screening program is new and possesses unique challenges, avariety of questions have to be answered: Which technology or which combination

of technologies will most probably result in a new product candidate? Which routegives highest chance of success? Which route is fastest to pursue? And which willdeliver the optimal result in terms of quality and patentability? Accordingly, noclear, straightforward answer can be given, but often accumulated knowledge iscrucial in deciding which route to follow

Technological improvements have contributed to shorten the time from idea

to product significantly In the mid-1990s, it took approximately five years fromthe creation of a gene bank to selling the product in the market place In 2000,this was reduced to approximately 26 months Today, in many cases, it ispossible to go from enzyme identification to shipment of large (tonnes) quanti-ties of safe and approved product in the technical field in approximately 12 to 24months Within the food and animal feed areas the approval process extends thelaunch time in the order of an additional one to two years Once enzymediscovery is made, enzyme scale-up progression begins This includes strainevaluation, selection, fermentation development and optimization These tech-nologies are described in Chapter 3

Aehle W (2004), Enzymes in Industry, Wiley-VCH, Weinheim.

Bashkirova E and Berka R M (2003) ‘Towards the discovery of new enzymes involved

in biomass degradation: combination of SSH and microarray technologies to identify

Trichoderma reesei biomass-induced genes’, 25th Symposium on Biotechnology for Fuels and Chemicals, Breckenridge, Colorado, USA, May 4–7.

Berka R M., Nelson B A., Zaretsky E J., Yoder W T and Rey M W (2003) ‘Genomics

of Fusarium venenatum: An alternative fungal host for making enzymes’, Applied

Mycology and Biotechnology, 4, 5682–5687.

Bott R (2005) ‘Analyzing the three-dimensional structures of variant enzymes’ In

Svendsen A., Enzyme Functionality, Marcel Dekker, New York, p 35

Brady S F., Chao C J., Handelsman J and Clardy J (2001) ‘Cloning and heterologous

expression of a natural product biosynthetic gene cluster from eDNA’, Organic Letters,

3(13), 1981–1984.

Bull A T., Ward A C and Goodfellow M (2000) ‘Search and discovery strategies for

biotechnology: the paradigm shift’, Microbiology and Molecular Biology Reviews, 64(3),

573–606,CP3.

Dalbøge H and Heldt-Hansen H P (1994) ‘A novel method for efficient expression cloning

of fungal enzyme genes’, Molecular and General Genetics, 243(3), 253–260.

Dalbøge H and Lange L (1998) ‘Using molecular techniques to identify new microbial

biocatalysts’, Trends in Biotechnology, 16(6), 265–272.

DeRisi J L., Iyer V R and Brown P O (1997) ‘Exploring the metabolic and genetic control

of gene expression on a genomic scale’, Science, 278(5338), 680–686.

Diehn M., Eisen M B., Botstein D and Brown P O (2000) ‘Large-scale identification of

14 Novel enzyme technology for food applications

secreted and membrane-associated gene products using DNA microarrays’, Nature

Genetics, 25(1), 58–62.

Duffner, F., Wilting R and Schnorr K (2001), Signal Sequence Trapping, Novozymes

A/S, Patent: WO 20017–1315-A.

Dunn-Coleman N and Prade R (1998) ‘Toward a global filamentous fungus genome

sequencing effort’, Nature Biotechnology, 16(1), 5.

Eijsink V G H., Gåseidnes S., Borchert T V and van den Burg B (2005), “Directed

evolution of enzyme stability’, Biomolecular Engineering, 22(1–3), 21–30,

Handelsman J (2005) ‘Metagenomics or megagenomics?’, Nature Reviews Microbiology,

3(6), 457–458.

Henrissat B and Bork P (1996) ‘On the classification of modular proteins’, Protein

Engineering, 9(9), 725–726.

Hirose I., Sano K., Shioda I., Kumano M., Nakamura K and Yamane K (2000) ‘Proteome

analysis of Bacillus subtilis extracellular proteins: a two-dimensional protein

electro-phoretic study’, Microbiology-UK, 146, 65–75.

Joo H., Arisawa A., Lin Z L and Arnold F H (1999) ‘A high-throughput digital imaging

screen for the discovery and directed evolution of oxygenases’, Chemistry and Biology,

6(10), 699–706.

Kongsbak L., Jørgensen K S., Jørgensen C T., Husum T L., Ernst S and Møller S (1999), A Fluorescence Polarisation Screening Method, Novozymes A/S, Patent WO 9945–143A.

Lorenz P and Schleper C (2002) ‘Metagenome – a challenging source of enzyme

discovery’, Journal of Molecular Catalysis B: Enzymatic, 19, 13–19,

Meeuwsen P J A., Vincken J P., Beldman G and Voragen A G J (2000) ‘A universal

assay for screening expression libraries for carbohydrases’, Journal of Bioscience and

Bioengineering, 89(1), 107–109.

Nedwin G E., Schäfer T and Falholt P (2005) ‘Enzyme discovery – Screening, cloning,

evolving’, Chemical Engineering Progress, 101(10), 48–55.

Ohnishi J., Mitsuhashi S., Hayashi M., Ando S., Yokoi H., Ochiai K and Ikeda M (2002).

‘A novel methodology employing Corynebacterium glutamicum genome information to

generate a new L-lysine-producing mutant’, Applied Microbiology and Biotechnology,

58(2), 217–223.

Precigou S., Goulas P and Duran R (2001) ‘Rapid and specific identification of nitrile

hydratase (NHase)-encoding genes in soil samples by polymerase chain reaction’, FEMS

Microbiology Letters, 204(1), 155–161.

Preisig C and Byng G (2001) ‘Applications of mass spectrometry in screening for new

biocatalysts’, Journal of Molecular Catalysis B: Enzymatic, 11(4–6), 733–741.

Radford A and Parish J H (1997) ‘The genome and genes of Neurospora crassa’, Fungal

Genetics and Biology, 21(3), 258–266.

Rondon M R., Goodman R M and Handelsman J (1999) ‘The Earth’s bounty: Assessing

and accessing soil microbial diversity’, Trends in Biotechnology, 17(10), 403–409.

Ruijssenaars H J and Hartmans S (2001) ‘Plate screening methods for the detection of

polysaccharase-producing microorganisms’, Applied Microbiology and Biotechnology,

55(2), 143–149.

Schellenberger V (1997), Compartmentalization Method for Screening Microorganisms,

Genencore International, Patent WO 9737–036A1.

Schäfer T and Borchert T V (2004) ‘Bioprospecting for industrial enzymes: importance

of integrated technology platforms for successful biocatalyst development’, in Bull A T., Microbial Diversity and Bioprospecting, ASM Press, Washington, 375–390.

Short J M (1997) ‘Recombinant approaches for accessing biodiversity’, Nature

Biotech-nology, 15(13), 1322–1323.

Short J M and Keller M (2001), High-Throughput Screening for Novel Enzymes, Diversa

Corporation, Patent US 6,174, 673 B1.

Torsvik V., Ovreas L and Thingstad T F (2002) ‘Prokaryotic diversity – Magnitude,

dynamics, and controlling factors’, Science, 296(5570), 1064–1066.

Yaver D S., Weber B and Murrell J (2003) ‘Global expression profiling of the lignin degrading fungus Ceriporiopsis subvermispora for the discovery of novel enzymes’,

Applied Mycology and Biotechnology, 3, 261–269.

Zagulski M., Herbert C J and Rytka J (1998) ‘Sequencing and functional analysis of the

yeast genome’, Acta Biochimica Polonica, 45(3), 627–643.

Biocatalysis is gradually taking over from chemical catalysis in many industrialapplications Enzymes are environmentally friendly, biodegradable, efficient, andlow cost in terms of resource requirements; as such they provide benefits comparedwith traditional chemical approaches in various industrial processes In manyinstances, however, natural enzymes do not perform optimally in a particularunnatural process and, as such, can be unsuitable for large-scale industrial applica-

tions (Schoemaker et al., 2003) Reflecting their participation in complex metabolic

networks inside living cells, enzymes are often inhibited by their own substrates orproducts, either of which may severely limit the productivity of an industrialbiocatalytic process During natural evolution, enzymes are optimized and oftenhighly specialized for certain biological functions within the context of a livingorganism In contrast, biotechnology needs enzymes that have (i) a high activityover longer periods of time (a feature that might clash with the need for rapidprotein turnover inside a cell), (ii) a high stability under harsh physical (hightemperature) and chemical (non-aqueous solvents) conditions and (iii) a highspecificity and selectivity that does allow the enzyme to generate efficientlyspecific products that are not necessarily present in nature

There are three major and principally different routes to obtain enzyme variantswith improved features: (i) isolating enzyme variants from organisms living inappropriate environments, (ii) rationale-based mutagenesis and (iii) laboratory

evolution The first option assumes that the desired enzyme is around and has beengenerated by natural evolution To allow rationale-based engineering, a high-resolution three-dimensional model and insight into the structure–function relations

of the biocatalyst of interest are required Laboratory evolution offers a way tooptimize enzymes randomly in the absence of structural or mechanistic informa-tion (Bornscheuer and Pohl, 2001)

In this chapter, the available techniques for engineering enzymes are described.The engineering approaches are subdivided into directed (rational protein engineer-ing) and random (laboratory evolution) techniques In addition, different selectionand (high-throughput) screening methods are described, a crucial developmentthat allows screening of large mutant libraries

Laboratory evolution has emerged as a powerful tool for improving biocatalysts aswell as for broadening our understanding of the underlying principles of substratespecificity, stereoselectivity and the responsible catalytic mechanism In contrast

to rational protein design (discussed below), laboratory evolution does not requireknowledge of the three-dimensional structure of a given enzyme or about therelationship between structure, sequence and mechanism

Laboratory evolution experiments implement a simple, iterative Darwinianoptimization algorithm Molecular diversity is typically created by random muta-genesis [for example error-prone polymerase chain reaction (PCR)] and/orrecombination of a target gene or a family of related target genes Improvedvariants are identified in a screen (or selection) that accurately reflects theproperties of interest The gene(s) encoding those improved enzymes are, ifnecessary, used as parents for the next round of evolution

The basis of laboratory evolution, also referred to as directed evolution, waslaid by Pim Stemmer in 1994 He proposed a recombination system for genes(‘gene shuffling’, or ‘molecular breeding’) by using random fragmentation of two

or more genes and their subsequent reassembly (Stemmer, 1994 a,b) Comparedwith the error-prone PCR method in which few point mutations are introduced,gene shuffling results in blocks of mutations Since then, numerous variations onthis theme have been developed, each specific to certain types of proteins or

desired outcomes (Yuan et al., 2005) Currently, laboratory evolution principles

have been used to improve a variety of enzyme properties: enantioselectivity

(Reetz et al., 2004), catalytic efficiency or rate (Van der Veen et al., 2004, 2006), enzyme stability (Kim et al., 2003; Eijsink et al., 2005), pH activity profile (Bessler

et al., 2003), enzyme functionality in organic solvents (Castro and Knubovets, 2003), product inhibition (Rothman et al., 2004) and substrate specificity (Zhang

et al., 1997; Yano et al., 1998; Oue et al., 1999) Further developments in the field

of laboratory evolution take the procedures to a different level Enzymatic

path-ways (Masip et al., 2004; Umeno et al., 2005) and genomes (Patnaik et al., 2002; Zhang et al., 2002; Dai and Copley, 2004) are now subjected to various shuffling

18 Novel enzyme technology for food applications

Fig 2.1 Key steps in a typical laboratory evolution experiment.

protocols, with various outcomes Unlike natural recombination, the geneticmaterial (gene, operon, genome) of more than two parents may be shuffled in asingle laboratory evolution experiment

2.2.1 Techniques used in laboratory evolution

The major steps in a typical laboratory evolution experiment are outlined inFig 2.1 The genetic diversity for evolution is created by mutagenesis and/orrecombination of one or more parent sequences These altered genes are clonedback into a plasmid for expression in a suitable host organism Clones expressingimproved enzymes are identified in a high-throughput screen (see Section 2.2.2) or

in some cases by selection, and the gene(s) encoding those improved enzymes areisolated and if necessary recycled to the next round of laboratory evolution.The most widely used approaches for generating diversity are random point

mutagenesis and in vitro recombination The commonly used technique to create

random point mutants is error-prone PCR The fidelity of the DNA polymerase isdecreased (i) by adding divalent cations (for example Mn2+) to substitute for theoptimal Mg2+ in the PCR mixture or (ii) by employing high error rate DNApolymerases (for example Taq polymerase lacking proofreading activity, Mutazymewhich is designed to make mistakes) that will incorporate mismatching bases at a

Fig 2.2 Basic DNA shuffling scheme (A) The starting pool of homologous genes can be either a library of random mutants of a parental gene (e.g by error-prone PCR) or a family

of related genes The pool of genes is fragmented with DNaseI (B) Pool of random DNA fragments (C) The fragments are reassembled into full length genes by repeated cycles of PCR without added primers During this step the fragments prime each other in the homologous regions, resulting in recombination when fragments derived from one parental gene prime another one, causing a template switch (crossovers shown by the black line) Reassembly of the random fragments into full length genes results in frequent template switching and recombination (D) A selected pool of improved recombinants provides the

starting point for another round of mutation and/or recombination.

controllable rate during gene amplification A low error rate (2–3 base tions or ~1 amino acid changes) accumulates mostly adaptive mutations, whereas

substitu-higher error rates merely generate neutral and deleterious mutations (Arnold et al., 2001) DNA shuffling methods enable the in vitro recombination of DNA

sequences that are more or less related, either orthologous genes or error-pronePCR variants (Fig 2.2) DNA shuffling requires a degree of identity >55–65%.Techniques such as incremental truncation for the creation of hybrid enzymes(ITCHY) are variant methods that allow for recombination of sequences with alower degree of homology In the last decade a multitude of methods have beendeveloped to enable shuffling of genes with lower homology, to improve mutantlibraries by negative selection for wild-type sequences or to obtain a less biased

library (Kurtzman et al., 2001) Most new methods have been invented to solve the

drawbacks of the original protocol or to circumvent patent limitations Althoughthese alternative procedures aim to solve one problem, they usually appear tocreate another one For example, it is difficult to start with highly different parentgenes and still reach a high recombination frequency Methods aiming at higher

Fragmentation

Reassembly of fragments (recombination)

20 Novel enzyme technology for food applications

recombination frequencies start with more homologous parents, thereby ing fewer mutations On the other hand, methods have been developed to recombinegenes without any homology These processes result mainly in only one crossover,although some methods do generate multiple crossovers at fixed places

introduc-Another way of obtaining a high recombination frequency is by starting fromsynthetic oligonucleotides The advantage of this method is that any mutant can beconstructed and thereby the largest possible sequence space can be explored Anadditional advantage is the possibility of using codons other than the original ones

in order to obtain more homology Furthermore, the preferred codon usage for theexpression host can be applied in the synthetic oligonucleotides The largestdisadvantages of synthetic methods are the high costs and the large size of thelibrary, which quickly exceeds the most elaborate screening and selection methods(Ostermeier, 2003) Other conflicting parameters seem to be speed and bias.Recombination methods that aim to generate unbiased libraries all consist ofnumerous steps in order to achieve this, resulting in more elaborate procedures,while quick protocols normally result in more wild-type backgrounds and a biasedlibrary

Most of the methods mentioned can result in good mutant libraries Therefore,the choice for one or another strategy is usually led by the size of the protein, thegoal of the research, the existence of homologous proteins, the selection and/orscreening capacity and practical issues like the equipment and expertise in theresearch group Table 2.1 summarizes some of the methods that have beensuccessfully utilized for laboratory evolution of a variety of proteins This is not acomplete list, as new techniques and strategies for laboratory evolution areconstantly arising

2.2.2 Selection and screening

The success of a laboratory evolution experiment depends greatly on the methodthat is used to select the best mutant enzyme Since most laboratory evolutionexperiments generate a huge mutant library, it is very important to develop anefficient method of screening this library for the desired property Both selectionand screening strategies have been developed for all kinds of enzyme functions

(Boersma et al., 2007) The big challenge in these strategies is generally to make

the improved function quantifiable, that is to differentiate signal from noice.Enzymatic assays have to be sufficiently sensitive and specific to identify positivemutants (Zhao and Arnold, 1997)

Selection is based on the fact that mutants with the desired enzyme functionhave an advantage over wild-type enzymes; variants are selected because, undercertain conditions, they enable the host to grow Selection in the laboratory mimicsthe natural survival-of-the-fittest strategy and is the most efficient method to findthe best mutant, since only mutants of interest will appear Unfortunately, this

approach is not possible for all enzymatic activities For in vivo selection this

means that only enzyme activities with a growth or survival advantage can be used.Only a few industrially interesting enzymes are essential for the bacterial cell

themselves, so most selection methods are based on enzymatic activities that lead

to the generation of a product that is essential for growth of the expression host.The coupling of the desired enzymatic reaction to survival in the selection stepoften requires the development of complex, non-trivial and intelligent assays

(Taylor et al., 2001) Sometimes, this means that the substrates in these selection

systems are not the desired substrates, but analogues thereof This may result in theselection of undesired mutants with activity towards the analogue and not towardsthe wanted substrate It is, therefore, very important to choose the selectionsubstrate carefully, since the first law of laboratory evolution is: ‘you get what youselect/screen for’ (Schmidt-Dannert and Arnold, 1999)

In vitro selection is usually based on binding the enzyme to the desired substrate

or a transition state analogue, although strategies in which catalytic properties are

used for selection are also described (Boersma et al., 2007) These methods are

mostly based on a physical linkage between phenotype and genotype The firstestablished and most used technique is phage display, which has been successfullyused to find improved enzymes In this system, the enzyme of interest is fused to

a coat protein of a filamentous phage and as such displayed on the outside of thephage, where in principle it is able to retain enzymatic activity Since the geneencoding the displayed protein is present in the phage particle, the gene of themutant enzyme with the desired property is linked to its phenotype (Fernandez-

Gacio et al., 2003) When the displayed proteins may be toxic to filamentous phage

assembly or incompatible with the bacterial secretion pathway, lytic phages can beused that allow displayed sequences to minimize negative selection

Other in vitro selection methods with a physical phenotype–genotype linkage

are cell-surface display, ribosome display, plasmid display and mRNA-protein

fusion (Lin and Cornish, 2002; Becker et al., 2004) Recently, a different approach was described to maintain a linkage between genotype and phenotype In vitro

compartmentalization is a method in which compartments are formed as aqueousdroplets in water-in-oil emulsions which contain only one gene and a completetranscription/translation machinery (Tawfik and Griffiths, 1998) These dropletsmimic a bacterial cell by keeping the gene and its product together The dropletscontaining an enzyme with the desired activity can be selected by fluorescenceactivated cell sorting (FACS) or, when the gene is physically bound to thesubstrate, by breaking the droplets and fishing out the desired product (Griffiths

and Tawfik, 2000) The advantages of in vitro over in vivo selection are the larger

sample size of a mutant library that can be handled and as such the larger amount

of possible enzyme variants to be tested Drawbacks are that making the rightwater-in-oil emulsions with only one gene per droplet is tricky, and that the

efficiency of the in vitro transcription and translation can be a bottleneck.

Another way of finding the desired mutant enzyme is by screening In screeningmethods all mutants have to be tested for the desired enzymatic reaction, eventhose that might not be active or accurately folded The advantage is, however, thatalmost every enzymatic reaction can be tested, since the activity does not have to

be dependent on growth rate or the formation of essential products This can bedone in a qualitative way by relatively simple visual screens such as the formation

Error-prone PCR Introduces random point mutations by imposing imperfect, and thus Leung et al., 1989

gene to be mutagenized to a mutator E coli strain These strains lack DNA Camps et al., 2003

repair mechanisms or contain a modified polymerase with lower fidelity.

Saturation mutagenesis With this approach, it is possible to create a library of mutants containing Miyazaki and Arnold, 1999

all possible mutations at one or more pre-determined target positions in a Zheng et al., 2004

Wong et al., 2005

DNA shuffling DNA is randomly digested and allowed to recombine to form novel Stemmer, 1994 a,b

Synthetic shuffling Using degenerate oligonucleotides, every amino acid from a set of Ness et al., 2002

parents is allowed to recombine independently of every other amino acid Ostermeier, 2003 Physical starting genes are unnecessary, and additional design criteria

such as optimal codon usage can also be incorporated.

Staggered extension process (StEP) StEP consists of priming the template sequence(s) followed by repeated Zhao et al., 1998

catalyzed extension In each cycle the growing fragments anneal to Zhao, 2004 different templates based on sequence complementarity and extend

further This is repeated until full-length sequences form Due to template switching, most of the polynucleotides contain sequence information from different parental sequences.

Random chimeragenesis on transient DNA shuffling method for generating highly recombined genes The Coco et al., 2001

templates (RACHITT) approach relies on the ordering, trimming and joining of randomly cleaved Coco, 2003

parental DNA fragments annealed to a transient polynucleotide scaffold.

Improving enzyme performance in food applications

chimeragenesis (SISDC) the recombination of sequences that are not related at all.

Structure based combinatorial protein A semi-rational protein engineering approach that uses information from O’Maille et al., 2002

engineering (SCOPE) protein structure coupled with established DNA manipulation techniques O’Maille et al., 2004

to design and create multiple crossover libraries from non-homologous genes.

Incremental truncation for creation of Incremental truncation, a method for creating a library of every one base Ostermeier et al., 1999

hybrid enzymes (ITCHY) truncation of dsDNA, creates diversity by changing the length of a gene. Horswill et al., 2004

The combination of two incremental truncation libraries creates diversity

by fusing two gene fragments Performing ITCHY between two different genes generates libraries of fusion proteins in a DNA-homology independent fashion.

Creating multiple-crossover DNA This technique is based on the ITCHY technique but includes an extra Lutz et al., 2001

identity (SCRATCHY)

Sequence homology-independent This procedure maintains alignment between two parental genes, and Sieber et al., 2001

protein recombination (SHIPREC) produces cross-overs mainly at structurally related sites along the sequences. Udit et al., 2003

Combinatorial libraries enhanced by A combination of PCR reassembly and in vivo recombination in yeast Abecassis et al., 2000

recombination in yeast (CLERY) produces highly shuffled libraries.

Degenerate oligonucleotide gene Procedure for gene shuffling using degenerate primers that allows control Gibbs et al., 2001

shuffling (DOGS) of the relative levels of recombination between the genes that are shuffled

and reduces the regeneration of unshuffled parental genes.

Degenerate homoduplex gene family This approach seeks to randomly swap polymorphisms between a collection Coco et al., 2002

recombination (DHR) of polymorphic genes This technique is different from shuffle techniques

in that random segments of genes are not recombined, but homologous segments containing point mutations are.