Starch in food: structure, function and applications [1 ed.] 9780849316524, 0-8493-1652-9, 0-8493-2555-2, 1-85573-706-X, 1-85573-723-X, 1855737310

Citation preview

Starch in food

ß 2004, Woodhead Publishing Limited

Related titles from Woodhead's food science, technology and nutrition list: Proteins in food processing (ISBN 1 85573 723 X) Proteins are essential dietary components and also have a significant effect on food quality. This collection reviews the wide range of protein sources available, ways of modifying them, and their use in food processing to enhance the nutritional, textural and other qualities of food products. Yeasts in food (ISBN 1 85573 706 X) Yeasts play a crucial role in the sensory quality of a wide range of foods. This book provides a comprehensive review of the methods for their detection, identification and analysis as well as the role of yeasts in several food products including dairy products, meat, fruit, bread and beverages. Handbook of minerals as nutritional supplements (ISBN 0 8493 1652 9) This handbook provides a comprehensive analysis of the current status of research on various mineral supplements. Details of these books and a complete list of Woodhead's food science, technology and nutrition titles can be obtained by: · visiting our web site at www.woodhead-publishing.com · contacting Customer Services (email: [email protected]; fax: +44 (0) 1223 893694; tel.: +44 (0) 1223 891358 ext. 30; address: Woodhead Publishing Limited, Abington Hall, Abington, Cambridge CB1 6AH, UK) Selected food science and technology titles are also available in electronic form. Visit our web site (www.woodhead-publishing.com) to find out more. If you would like to receive information on forthcoming titles in this area, please send your address details to: Francis Dodds (address, tel. and fax as above; e-mail: [email protected]). Please confirm which subject areas you are interested in.

ß 2004, Woodhead Publishing Limited

Starch in food Structure, function and applications Edited by Ann-Charlotte Eliasson

ß 2004, Woodhead Publishing Limited

Published by Woodhead Publishing Limited Abington Hall, Abington Cambridge CB1 6AH England www.woodhead-publishing.com Published in North America by CRC Press LLC 2000 Corporate Blvd, NW Boca Raton FL 33431 USA First published 2004, Woodhead Publishing Limited and CRC Press LLC ß 2004, Woodhead Publishing Limited The authors have asserted their moral rights. 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 the publishers. The consent of Woodhead Publishing Limited and CRC Press LLC 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 or CRC Press LLC for such copying. Trademark notice: product or corporate names may be trademarks or registered trademarks, 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 Limited ISBN 1 85573 731 0 (book); 1 85573 909 7 (e-book) CRC Press ISBN 0-8493-2555-2 CRC Press order number: WP2555 The publisher's policy is to use permanent paper from mills that operate a sustainable forestry policy, and which have been manufactured from pulp which is processed using acid-free and elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper and cover board used have met acceptable environmental accreditation standards. Project managed by Macfarlane Production Services, Markyate, Hertfordshire (e-mail: [email protected]) Typeset by MHL Typesetting Limited, Coventry, Warwickshire Printed by TJ International Limited, Padstow, Cornwall, England

ß 2004, Woodhead Publishing Limited

Contents

Contributor contact details Part I

Analysing and modifying starch

1

Plant starch synthesis J. Preiss, Michigan State University, USA 1.1 Introduction: localization and function of starch in plants 1.2 Starch synthesis: enzyme reactions in plants and algae and glycogen synthesis in cyanobacteria 1.3 Properties of plant glucan synthesizing enzymes: ADP-glucose pyrophosphorylase 1.4 Properties of plant glucan synthesizing enzymes: starch synthase 1.5 Properties of plant glucan synthesizing enzymes: branching enzymes 1.6 Initiation of starch synthesis using a glucosyl-protein 1.7 Locating starch synthesis in plants: the plastid 1.8 In vivo synthesis of amylopectin 1.9 Regulating starch synthesis in plants 1.10 References

2

Analysing starch structure E. Bertoft, AÊbo Akademi University, Finland 2.1 Introduction: characterising structures of starch components 2.2 Fractionation of starch 2.3 Analysis of amylose

ß 2004, Woodhead Publishing Limited

2.4 2.5 2.6 2.7 2.8 2.9

Analysis of amylopectin structure Analysis of intermediate materials Analysis of chemically modified starches Future trends Sources of further information and advice References

3

Starch bioengineering A. Blennow, The Royal Agricultural and Veterinary University, Denmark Introduction: the importance of starch 3.1 3.2 Technologies for genetic modification and starch profiling 3.3 Improving starch yield and structure 3.4 Physical and chemical properties of modified starches 3.5 Functionality and uses of modified starches in food processing 3.6 Ensuring successful modification of starch 3.7 Future trends 3.8 References

4

Starch-acting enzymes D. P. Butler, Marc J. E C. van der Maarel and P. A. M. Steeneken, TNO Nutrition and Food Research Institute, The Netherlands Introduction: the importance of enzymes 4.1 4.2 Using enzymes to modify starch 4.3 Developing starch-modifying enzymes for food processing applications 4.4 Future trends 4.5 References

5

Understanding starch structure and functionality A. M. Donald, University of Cambridge, UK 5.1 Introduction: overview of packing at different lengthscales 5.2 The effect of amylopectin chain architecture on packing 5.3 Improving packing within starch granules 5.4 The gelatinisation process 5.5 Food processing: implications of starch granule structure 5.6 Conclusions and future trends 5.7 Sources of further information and advice 5.8 References

6

Measuring starch in food M Peris-Tortajada, Polytechnic University of Valencia, Spain 6.1 Introduction 6.2 Sample preparation

ß 2004, Woodhead Publishing Limited

6.3 6.4 6.5 6.6 6.7 Part II

Methods of analysing starch in food Determining starch in food: recent technological developments Future trends Sources of further information and advice References Sources of starch

7

The functionality of wheat starch H. Cornell, RMIT University, Australia 7.1 Introduction: manufacture of wheat starch for the food industry 7.2 Granular and molecular structure of wheat starch 7.3 Functionality of wheat starch: granules, films and pastes 7.4 Rheological properties of starch pastes and gels 7.5 Improving and chemically modifying wheat starch for use in the food industry 7.6 Wheat starch syrups 7.7 Analysing starch-based products 7.8 Future trends 7.9 Sources of further information and advice 7.10 References

8

Developments in potato starches W. Bergthaller, Federal Centre for Nutrition and Food, Germany 8.1 Introduction 8.2 Components and rheological properties of potato starch 8.3 Techniques for producing potato starch 8.4 Improving the functionality of potato starch for use in the food industry 8.5 Future trends 8.6 References

9

The functionality of rice starch J. Bao and C. J. Bergman, Texas A&M University, USA 9.1 Introduction 9.2 Rice flour and starch as food ingredient 9.3 Constituents of rice starch 9.4 Structure and functionality of rice starch 9.5 Gelatinization and the structure of rice starch 9.6 Retrogradation and other properties of rice starch 9.7 Improving rice starch functionality for food processing applications 9.8 Future trends

ß 2004, Woodhead Publishing Limited

9.9 9.10

Sources of further information and advice References

10

New corn starches P. J. White and A. Tziotis, Iowa State University, USA 10.1 Introduction: the use of corn starch in food processing 10.2 Improving the functionality of corn starch for food processing applications: natural corn endosperm mutants 10.3 Chemically modifying corn starches for use in the food industry 10.4 Genetically modifying corn starches for use in the food industry 10.5 Future trends 10.6 Sources of further information and advice 10.7 References

11

Tropical sources of starch S. N. Moorthy, Central Tuber Crops Research Institute, India 11.1 Introduction: tropical sources of starch 11.2 Characteristics and properties of cassava starch 11.3 Characteristics and properties of sweet potato starch 11.4 Characteristics and properties of yam and aroid starches 11.5 Characteristics and properties of other minor root starches 11.6 Modifying `tropical' starches for use in the food industry 11.7 Future trends 11.8 References

Part III

Applications

12

Starch as an ingredient: manufacture and applications P. Taggart, National Starch and Chemical, UK 12.1 Introduction 12.2 Manufacture 12.3 Structure 12.4 Modifications 12.5 Technical data 12.6 Uses and applications 12.7 Regulatory status: European label declarations 12.8 Acknowledgements 12.9 Bibliography

13

Utilizing starches in product development T. Luallen, Cargill Inc., USA 13.1 Introduction

ß 2004, Woodhead Publishing Limited

13.2 13.3 13.4 13.5 13.6 13.7

Components of starch Food applications for natural and modified starches Methods of starch selection Factors affecting starch in food products Using the functional properties of starch to enhance food products References

14

Modified starches and the stability of frozen foods H. D. Goff, University of Guelph, Canada 14.1 Introduction 14.2 The structure and stability of frozen foods 14.3 The role of modified starch in stabilizing frozen foods 14.4 Future trends 14.5 Sources of further information and advice 14.6 References

15

Starch-lipid interactions and their relevance in food products A-C. Eliasson and M. Wahlgren, Lund University, Sweden 15.1 Introduction 15.2 The structure and properties of the starch-lipid complex 15.3 Analysis of starch: lipids and emulsifiers 15.4 The effect of lipids on starch behaviour 15.5 Enzymatic degradation of amylose-lipid complexes 15.6 Future trends 15.7 References

16

Starch-based microencapsulation P. Forssell, VTT Biotechnology, Finland 16.1 Introduction: using microencapsulation in food processing 16.2 Using starch in microencapsulation: starch hydrolysates, derivatives, polymers and granules 16.3 Starch-based shell matrices for food ingredients 16.4 Future trends 16.5 References

Part IV 17

Starch and health

Development of a range of industrialised cereal-based foodstuffs high in slowly digestible starch V. Lang, Danone Vitapole, France 17.1 Introduction 17.2 Characteristics and properties of starch and starchy foods 17.3 Low G I diets and their associated health benefits

ß 2004, Woodhead Publishing Limited

17.4

17.5 17.6 17.7 17.8

Case study: low glycaemic index, high slowly digestible starch plain biscuits, the EDPÕ (`Long-lasting energy') range developed by Danone, Vitapole Future trends Sources of further information and advice Acknowledgements References

18

Starch: physical and mental performance F. Brouns, Cerestar Vilvoorde R & D Centre, Belgium and University of Maastricht, Netherlands and L. Dye, University of Leeds, England 18.1 Introduction 18.2 Physical performance: energy requirements, delivery and availability 18.3 Mental performance: the effects of glucose 18.4 Mental performance: the effects of CHO and glucose during the day 18.5 Future trends 18.6 References

19

Detecting nutritional starch fractions K. Englyst and H. Englyst, Englyst Carbohydrates, UK 19.1 Introduction 19.2 Methods of determining RAG, SAG and RS fractions 19.3 Quality control and troubleshooting 19.4 Carbohydrate bioavailability data for selected foods 19.5 Conclusion and future trends 19.6 Acknowledgement 19.7 References

20

Resistant starch M. Champ, INRA-UFDNH/CRNH, France 20.1 Introduction 20.2 Effects of resistant starch on the digestive system 20.3 Improving the functional effects of resistant starch 20.4 Future trends 20.5 Sources of further information and advice 20.6 References

21

Analysing starch digestion R. E. Wachters-Hagedoorn, M. G. Priebe and R. J. Vonk, University Hospital Groningen, The Netherlands 21.1 Introduction 21.2 Starch and the prevention of hypo- and hyperglycemia

ß 2004, Woodhead Publishing Limited

21.3 21.4 21.5 21.6 21.7 21.8

The determinants of the rate of absorption of starch-derived glucose Techniques for monitoring starch digestion Current applications of slowly available starch and the prevention of hyper- and hypoglycemia Future trends Sources of further information and advice References

ß 2004, Woodhead Publishing Limited

Contributor contact details

Chapter 1 Professor J. Preiss Department of Biochemistry Michigan State University East Lansing MI 48824 USA Tel: 517 353 3137 Fax: 517 353 9334 E-mail: [email protected]

Chapter 2 Dr E. Bertoft Department of Biochemistry and Pharmacy Ê bo Akademi University A PO Box 66 FIN 20521 Turku Finland Tel: 358 2 215 4272

ß 2004, Woodhead Publishing Limited

Fax: 358 2 215 4745 E-mail: [email protected]

Chapter 3 Dr A. Blennow The Royal Agricultural and Veterinary University Denmark E-mail: [email protected]

Chapter 4 Dr D. P. Butler, Dr M. J. E. C. van der Maarel and Dr P. A. M. Steeneken TNO Nutrition and Food Research Institute Groningen The Netherlands E-mail: [email protected] [email protected]

Chapter 5

Chapter 8

Professor A. Donald Department of Physics Cavendish Laboratory University of Cambridge

Dr W. Bergthaller Federal Research Centre for Nutrition and Food Location Detmoed and Muenster Institute for Cereal, Potato and Starch Technology PO Box 1354 Detmold 32756 Germany

Madingley Road Cambridge CB3 0HE Tel: +44 1223 337382 Fax: +44 1223 337000 E-mail: [email protected]

Tel: + 49 5231 741320 Fax: + 49 5231 741300 E-mail: [email protected]; [email protected]

Chapter 6 Professor M. Peris-Tortajada Department of Chemistry Polytechnic University of Valencia 46071 Valencia Spain E-mail: [email protected]

Chapter 7 Professor H. Cornell Department of Applied Chemistry RMIT University City Campus GPO Box 2476V Melbourne Victoria 3001 Australia Tel: 0061 3-9925 2117 Fax: 0061 3-9039 1321 E-mail: [email protected]

ß 2004, Woodhead Publishing Limited

Chapter 9 Dr C. Bergman University of Nevada Las Vegas Nevada 89154 USA Dr J. Bao Zhejiang University Huajiachi Hangzhou 310029 China E-mail: [email protected] [email protected]

Chapter 10

Chapter 13

Professor P. J. White and Ms A. Tziotis Department of Food Science and Human Nutrition and Centre for Crops Utilization Research 2312 Food Sciences Building Iowa State University Ames Iowa USA

Dr T. Luallen Cargill Inc Specialty Food and Pharma Solutions BU PO Box 1467 Cedar Rapids IA 52406 USA

Tel: 515 294 9688 Fax: 515 294 8181 E-mail: [email protected]

Tel: 319 399 6187 Fax: 319 399 6123 E-mail: [email protected]

Chapter 14 Chapter 11 Dr S. N. Moorthy Central Tuber Crops Research Institute Sreekariyam Thiruvananthapuram 695 017 Kerala India Tel: 0471 598551 Fax: 0091 471 590063 E-mail: [email protected]

Professor H. D. Goff Department of Food Science University of Guelph Guelph Ontario N1G 2W1 Canada Tel: 519 824 4120 Fax: 519 824 6631 E-mail: [email protected]

Chapter 15 Chapter 12 Dr P. Taggart National Starch and Chemical Ltd Prestbury Court Green Court Business Park 333 Styal Road Manchester UK Tel: +44(0)161 435 3200 Fax: +44(0)161 435 3221 E-mail: [email protected]

ß 2004, Woodhead Publishing Limited

Professor A-C. Eliasson and M. Wahlgren Food Technology Division Lund University Box 124 22100 Lund Sweden Tel: +46 222 9674 Fax: +46 222 9517 E-mail: [email protected]

Chapter 16

Chapter 19

Dr P Forssell VTT Biotechnology Tietotie 2 PO Box 1500 02044 VTT Finland

Dr K. Englyst and Dr H. Englyst Englyst Carbohydrates 2 Venture Road Chilworth Science Park Southampton SO16 7NP UK

E-mail: [email protected]

Chapter 17 Dr V. Lang Danone Vitapole R&D Center of the Groupe Danone Nutrition Research Department Route DeÂpartementale 128 91767 Palaiseau Cedex - France Tel: +33 1 69 35 72 34 Fax: + 33 1 69 35 76 89 E-mail: [email protected]

Tel: +44 (0) 23 80 769650 Fax: +44 (0) 23 80 769654 Email: [email protected]

Chapter 20 Dr M. Champ INRA-UFDNH/CRNH Rue de la GeÂraudieÁre BP 71627 44316 Nantes Cedex 03 France Email: [email protected]

Chapter 18 Professor F. Brouns and Dr L. Dye Cerestar Vilvoorde R&D Centre Havenstraat 84 B-1800 Vilvoorde Belgium Tel: 00 32 2 2570736 Fax: 00 32 2 2570740 Email: [email protected]

Chapter 21 Professor R. J. Vonk Laboratory of Nutrition and Metabolism Laboratory Centre CMC V, Y2147 University Hospital Groningen Hanzeplein 1 P.O. Box 30.001 9700 RB Groningen The Netherlands Tel: +31-50-3632675 Fax: +31-50-3611746 Email: [email protected]

ß 2004, Woodhead Publishing Limited

Part I Analysing and modifying starch

ß 2004, Woodhead Publishing Limited

1 Plant starch synthesis J. Preiss, Michigan State University, USA

1.1

Introduction: localization and function of starch in plants

This chapter reviews starch synthesis in higher plants and algae. Since the reactions leading to glycogen synthesis in the cyanobacteria are similar to those observed in the higher plants there will be some referral to studies in those organisms particularly in regulation of cyanobacterial 1,4-glucan synthesis. The enzymology and biochemistry of the various enzymes in the plant, algal and cyanobacterial systems will also be described. In view of the existing information available on the properties of the starch biosynthetic enzymes and the effects of certain mutants on starch structure a pathway of starch synthesis is described which postulates specific functions for the starch synthases and branching enzymes. Finally regulation of starch synthesis at the enzymatic level will be discussed and in relation to this regulation, recent results indicating how starch content has been increased in certain plants will be descibed. A previous chapter1 in the second edition of Starch Chemistry and Technology which reviewed starch biosynthesis discussed the various maize endosperm mutants or mutant combinations, 26 of them, that showed an effect on the quantity or the nature of the starch formed. This information remains of interest and the reader is referred to that review. However, for many of those mutants the biochemical basis for mutation effects on starch quantity or quality was unknown. This review will deal with only the mutants where the biochemical process affected by the mutation has been elucidated to at least some extent. There are some recent reviews on starch biosynthesis2±11 that discuss many of the areas presented in this chapter.

1.1.1 Leaf starch Starch is deposited in granules in almost all green plants and in various types of plant tissues and organs, e.g., leaves, roots, shoots, fruits, grains, and stems.

ß 2004, Woodhead Publishing Limited

Illumination of the leaf in bright light causes the formation of starch granules in the chloroplast organelle and was demonstrated in the nineteenth century.12 Disappearance of the starch occurs either by exposure of the leaf to low light or by extended exposure in the dark (24±48 hours). This is readily observed by iodine staining of the tissue13 or by light or electron microscopy.14 Starch accumulates due to carbon fixation during photosynthesis and the starch formed in the light is degraded in the dark to products that are in most cases utilized for sucrose synthesis. Mutants of Arabidopsis thaliana unable to synthesize starch, grow at the same rate as the wild type in a continuous light regime because they are able to synthesize sucrose,15 but their growth rate is drastically reduced if grown in a day-night regime. The reason for this is that the accumulated starch is required for sucrose synthesis at night; the sucrose is transported from the leaf to the sink tissues. Biosynthesis and degradation of starch in the leaf is therefore a dynamic process having diurnal fluctuations in its stored levels. Starch also plays an important role in the operation of stomatal guard cells, where it is degraded during the day. In the late afternoon or evening while the stomata are open, the starch is resynthesized. Leaf starch is lower in amylose content than what is observed in storage tissues.16 The amylose structure is also of a smaller molecular size.

1.1.2 Starch in storage tissues In storage organs, fruit or seed, during the development and maturation of the tissue, synthesis of starch occurs. At the time of sprouting or germination of the seed or tuber, or ripening of the fruit, starch degradation in these tissues then occurs and the derived metabolites are used as a source both for carbon and energy. The degradative and biosynthetic processes in the storage tissues may therefore be temporally separated. However, there is some possibility that during each phase of starch metabolism some turnover of the starch molecule occurs. The main site of starch synthesis and accumulation in the cereals is the endosperm, with starch granules that are located within the amyloplasts. Starch content in potato tuber, maize endosperm, and in roots of yam, cassava and sweet potato ranges between 65 and 90% of the total dry matter. Patterns of starch accumulation during development of the tissue are specific to the species and are related to the unique pattern of differentiation of the organ. Starch granules in storage tissues can vary in shape, size and composition. The shape and size of the granules depends on the source, but in each tissue there is a range of sizes and shapes. The diameter of the starch granule changes during the development of the reserve tissue. There are also some fine features, characteristic of each species, e.g., the `growth rings', spaced 4±7 m apart, and the fibrillar organization seen in potato starch, which allows one to identify the botanical source of the starch by microscopic examination. Two polymers are distinguished in the starch granule. Amylose, which is essentially linear, and amylopectin, highly branched. Amylose is mainly found as linear chains of about 840 to 22,000 units of -D-glucopyranosyl residues linked by -(1->4) bonds (molecular weight around 136,000 to 3.5  106). The number

ß 2004, Woodhead Publishing Limited

of anhydroglucose units, however, varies quite widely with plant species and stage of development. Some of the amylose molecules are branched to a small extent (-1->6-D glucopyranose; one per 170 to 500 glucosyl units). Amylopectin, in contrast, which usually comprises about 70% of the starch granule, is more highly branched with about 4 to 5% of the glucosidic linkages being -1->6. Amylopectin molecules are large flattened disks consisting of -(1,4)-glucan chains joined by frequent -(1,6)-branch points. Many models of amylopectin structure have been proposed but from these the most satisfactory models, i.e., those that best fit the experimental data available, are those proposed by Robin et al.17, Manners and Matheson18 and by Hizukuri.19 These are known as cluster models. The chemical and physical aspects of the starch granule and its components amylose and amylopectin have been discussed in some recent excellent reviews by Morrison and Karkalis20 and Hizukuri.21

1.2 Starch synthesis: enzyme reactions in plants and algae and glycogen synthesis in cyanobacteria 1.2.1 Enzyme reactions of starch synthesis The sugar nucleotide utilized for synthesis of the -1,4 glucosidic linkages in amylose and amylopectin is ADP-glucose and not UDP-glucose. ADP-glucose synthesis is catalyzed by ADP-glucose (synthetase) pyrophosphorylase (reaction 1, E.C. 2.7.7.27; ATP:-D-glucose-1-phosphate adenylyltransferase). ATP ‡ -glucose-1-P () ADP-glucose ‡ PPi

1:1

ADP-glucose ‡ -1,4 glucan ˆ)-1,4-glucosyl--1,4 glucan ‡ ADP

1:2

Elongated -1,4-oligosaccharide chain ˆ)-1,4--1,6 branched-glucan (pro-amylopectin) (phytoglycogen)

1:3

Reaction 2 is catalyzed by starch synthase (E.C. 2.4.1.21; ADP-glucose;1,4-D-glucan 4--glucosyltransferase). A similar reaction is noted for glycogen synthesis in cyanobacteria and other bacteria (see references 22 and 23 but the reaction is referred to as glycogen synthase (also E.C. 2.4.1.21). Reaction 3 is catalyzed by branching enzyme (E.C. 2.4.1.18; 1,4--D-glucan 6--(1,4-glucano)-transferase). The branch chains in amylopectin are longer (about 20 to 24 glucose units long) and there is less branching in amylopectin (~5% of the glucosidic linkages are -1,6 as seen in glycogen (10±13 glucose units long and 10% of linkages are -1,6). Thus, the starch branching enzymes may have different properties with respect to size of chain transferred, or placement of branch point, than enzyme that branches glycogen. Alternatively, the interaction of the starch branching enzymes with the starch synthases may be different from the interaction of the bacterial branching enzymes with their respective glycogen synthases. The chain elongating properties of the starch synthases could be different from those observed for the bacterial glycogen synthases and may account for some of the differences observed in the amylopectin structure. The

ß 2004, Woodhead Publishing Limited

differences in the catalytic properties of the starch synthases and branching enzymes isolated from different plant sources may also account for the differences observed in the various plant starch structures. Isozymic forms of plant starch synthases (cited in references, 3, 4, 24±28) and branching enzymes (cited in references 3, 4, 24, and in recent literature, 29±34) have been reported. They seem to play different roles in the synthesis of the two polymers of starch, amylose and amylopectin and are products from different genes. In many different plants35±40 as well as in Chlamydomonas reinhardtii 41 a granule-bound starch synthase involved in the catalysis of reaction (5) has been shown to be involved in the synthesis of amylose. Mutants of many different plants defective in this enzyme, are known as `waxy' mutants and give rise to starch granules having only amylopectin. Another enzyme, a debranching enzyme, most probably is involved in synthesis of the starch granule and its polysaccharide components amylose and amylopectin.42±45 Soluble -glucan formed by reactions 1±3, is debranched to form the amylopectin present in the starch granule9, 46 and possibly provide a primer in the starch granule for amylose synthesis by the granule-bound starch synthase. The data strongly suggesting the role of a debranching enzyme in synthesis of amylopectin and the starch granule is discussed in a later section. Reaction 2 was first described by Leloir et al.47 with UDP-glucose as the glycosyl donor, but it was later shown that ADP-glucose was more efficient in terms of maximal velocity and km value.48 Leaf starch synthases and the soluble starch synthases of reserve tissues are specific for ADP-glucose. In contrast, the starch synthases bound to the starch granule in reserve tissues do have some low activity with UDP-glucose as compared to activity seen with ADP-glucose.

1.3 Properties of plant glucan synthesizing enzymes: ADPglucose pyrophosphorylase 1.3.1 Structure-function relationships The ADP-glucose pyrophosphorylases (ADPGlc Ppase) of higher plants and green algae, as well as the cyanobacteria, are proteins under allosteric control and an important enzyme site for regulation of starch synthesis. The enzymes are highly activated by 3-phosphoglycerate and inhibited by inorganic phosphate. The regulation of starch and cyanobacterial glycogen synthesis via regulation of the photosynthetic ADPGlc Ppase will be discussed in a later section. The structural properties of the ADPGlc Ppase will be described in this section. The kinetic and regulatory properties of the ADPGlc Ppases from the leaf extracts of spinach, barley, butter lettuce, kidney bean, maize, peanut, rice, sorghum, sugar beet, tobacco, and tomato have been studied in detail and are similar.49±52 The spinach leaf ADPGlc Ppase has been purified by either preparative disc gel electrophoresis,53 by hydrophobic chromatography51 and the use of FPLC.54 The enzyme has a molecular mass of 206,000 and is composed of two different subunits, of molecular masses of 51 and 54 kDa.54±57 These subunits are also

ß 2004, Woodhead Publishing Limited

distinguished with respect to amino acid composition, amino-terminal sequences, peptide patterns of the tryptic digests on high-performance liquid chromatography (HPLC), and antigenic properties. The two subunits are therefore quite distinct and are the products of two different genes. The enzyme may be considered as having an 2 2 structure. Bacterial ADPGlc Ppases including the cyanobacterial enzymes, in contrast, are homotetrameric, i.e., composed of only one subunit, with a molecular mass of 50 to 55 kDa depending on the species.58 Other plant ADPGlc Ppases have been shown to be composed of two dissimilar subunits. The maize endosperm ADPGlc Ppase, which has a molecular mass of 230 kDa, is composed of subunits of 55 and 60 kDa, corresponding to the spinach leaf 51 and 54 kDa.59 The maize endosperm mutants shrunken 2 (sh 2) and brittle 2 (bt2) are ADPGlc Ppase activity deficient (reviewed references 3 and 4). In immunoblotting experiments using antibodies raised against the native spinach leaf enzyme and the individual subunits, it was found that the mutant bt2 endosperm lacks the 55 kDa subunit and the mutant sh 2 endosperm lacks the 60 kDa subunit. These results59 strongly suggested that the maize endosperm ADPGlc Ppase is composed of two immunologically distinctive subunits and that the sh 2 and bt2 mutations cause reduction in ADPGlc Ppase activity through the lack of one of the subunits; the sh 2 gene would be the structural gene for the 60 kilodalton protein while the bt2 gene would be the structural gene for the 55 kDa protein. Consistent with this hypothesis is the isolation of an ADPGlc Ppase cDNA clone from a maize endosperm library60 which hybridized with the small subunit cDNA clone from rice.61 This maize ADPGlc Ppase cDNA clone was found to hybridize to a transcript present in maize endosperm but absent in bt2 endosperm. Thus, the bt2 mutant appears to be the structural gene of the 55 kDa subunit of the ADPGlc Ppase. Hylton and Smith62 proposed the existence of not two, but four polypeptides of MW around 50 kDa for the ADPglucose PPase of pea embryo, and a molecular mass for the holoenzyme of about 110 kDa. The relationship of the four subunits to the constitution of the native enzyme was not explained. However, the available information of most systems indicates that both the seed and leaf ADPglucose pyrophosphorylase are heterotetramers composed of two different subunits, and that, on the basis of immunoreactivity and sequence data,63 there is close homology between the subunits in the leaf enzyme and with the subunits of reserve tissue enzyme. Another point brought out by comparison of the amino acid sequences of the the two different plant subunits with each other and with the bacterial ADPGlc Ppase subunit amino acid sequences is that the plant subunits may have evolved from the bacterial subunit.63

1.3.2 Chemical modification of ligand binding sites of substrates and effectors Substrate sites Because the plant native ADPGlc Ppases are tetrameric and composed of two different subunits it was of interest to understand why two subunits are required

ß 2004, Woodhead Publishing Limited

for optimal catalytic activity in contrast to the bacterial ADPGlc Ppase. The enzyme must contain ligand-binding sites for the activator, 3PGA, and inhibitor, Pi, as well as catalytic sites for the two substrates, ATP and glucose-1-P, and it is possible that these sites may be located on different subunits. Chemical modification was used to obtain information on the catalytic mechanism and on the catalytic site of the ADPGlc Ppase. Chemical modification studies on the ADPGlc Ppase have involved the use of the following affinity labels: pyridoxal5-phosphate (PLP), an analogue of the activator 3-PGA as well as an analog of the substrate, glucose-1-P; 8-azido-ATP and 8-azido-ADPGlc, photoaffinity analogs of the substrates ATP and ADPGlc, respectively64, 3 and phenylglyoxal, for the identification of arginine residues that may be involved in substrate or effector binding. These types of studies have provided information on the catalytic and regulatory sites of the spinach ADPGlc Ppase and on the role of the large and small subunits. In studies with the E. coli ADPGlc Ppase, Lys residue 195 has been identified as the binding site for the phosphate of glucose-1-P65 and tyrosine residue 114 has been identified as involved in the binding of the adenosine portion of the substrate, ATP.66 The overall amino acid sequence identity of the E. coli enzyme when aligned with the plant and cyanobacterial ADPGlc Ppases ranges from 30 to 40%.63, 67, 68, 69 However, there is greater sequence identity when the E. coli ATP and glucose-1-P binding sites (Tables 1.1 and 1.2) are compared with the corresponding sequences of the plant and cyanobacterial enzymes suggesting that those sequences are still important in the plant enzyme, probably having the same function. Indeed, in a recent preliminary experiment with the potato tuber ADPGlc Ppase expressed in E. coli,70 site-directed mutagenesis on the lysine residue K198 of the 50kDa subunit (equivalent to the E. coli ADPGlc Ppase K195 to a glutamate residue, increased the S0.5 value (concentration required for 50% of maximal activity) for glucose-1-P from 57 M to about 31 mM without any perceptible change on the km or Ka for the other substrates, Mg+2, ATP or for activator, 3PGA (Y. Fu and J. Preiss, unpublished results, Table 1.3). The apparent affinity of glucose-1-P was lowered over 500-fold. Even a conservative mutation such as arginine replacing lysine at residue 198, caused a 135-fold decrease in the glucose-1-P apparent affinity. These results indicate an involvement of Lys residue 198 of the plant ADPGlc Ppase in the binding of glucose-1-P. In the case of the putative ATP binding site instead of tyrosine there is a phenylalanine residue in the corresponding sequences of the plant and cyanobacterial enzymes (Table 1.2). It would be of interest to determine in future site-directed mutagenesis and chemical modification studies whether the WFQGTADAV region of the plant enzyme is indeed a portion of the ATP binding region or whether the conservative change of two amino acids in the sequence has affected the function of that portion of the protein. The amino acids completely conserved are the tryptophan, glycine, threonine, alanine and aspartate residues and possibly mutation of those residues would indicate the relevancy of this region as an ATP binding site.

ß 2004, Woodhead Publishing Limited

Table 1.1 Conservation in plant ADPGlc Ppases of the Glucose-1-Phosphate (65) sites present in E. coli ADPGlc PPase. References to these sequences for the plant ADPGlc Ppases are in Smith-White and Preiss.63 The sequences for the Anabaena enzyme is in Charng et al.,67 for the Synechocystis enzyme in Kakefuda et al.68 and for the wheat endosperm small subunit, in Ainsworth et al.69 The number 195 corresponds to Lys195 of the E. coli enzyme and | signifies the same amino acid as found in the E. coli enzyme Source PROKARYOTES E. coli S. typhimurium Anabaena Synechocystis

Glucose-1-P Site 195 IIEFVEKP-AN ||D|||||-|| V|D|S|||KGE |TD|S|||QGE

PLANT SMALL SUBUNIT A. thaliana leaf Barley endosperm Maize endosperm 54kDa Potato tuber 50kDa Rice seed Spinach leaf 51kDa Wheat endosperm

||||A|||KGE ||||A|||KGE ||||A|||KGE ||||A|||QGE |V||A|||KGE ||||A|||KGE ||||A|||KGE

PLANT LARGE SUBUNIT A. thaliana leaf Barley endosperm Maize endosperm 60 kDa Potato tuber 51kDa Spinach leaf 54kDa Wheat endosperm (large subunit)

L|S|S|||KGD V|Q|S|||KGD VLQ|F|||KGA VVQ|A|||KGF VLS|S|||KGD VVQ|S|Q|KGD

Activator sites The binding site for pyridoxal phosphate in the spinach leaf ADPGlc Ppase small subunit was isolated revealing a lysine residue close to the C-terminus which may be important for 3PGA activation.71 When PLP is covalently bound, the plant ADPGlc Ppase is much less dependent on 3PGA for activation. The reductive phosphopyridoxylation is also prevented by the allosteric effectors, 3PGA and Pi. These observations showing that the modified enzyme no longer requires an activator for high activity and that the covalent modification is prevented by the presence of the allosteric effectors, strongly indicate that the activator analog, PLP, is binding at the activator site. Ball and Preiss72 showed also that three lysine residues of the spinach leaf large subunit are also involved or close to the binding site of pyridoxal-P and, presumably, of the activator, 3PGA. The chemical modification of these Lys residues by pyridoxal-P was prevented by the presence of 3PGA during the reductive phospho-pyridoxylation process and in the case of the Lys residue of site 1 of the small subunit and site 2 of the large subunit (Table 1.4), Pi also prevented them from being modified by reductive pyridoxylation.

ß 2004, Woodhead Publishing Limited

Table 1.2 Conservation in plant ADPGlc Ppases of the ATP binding sites present in E. coli ADPGlc Ppase.22 References to these sequences for the plant ADPGlc Ppases are in Smith-White and Preiss.63 The sequences for the Anabaena enzyme is in Charng et al.,67 for the Synechocystis enzyme in Kakefuda et al.68 and for the wheat endosperm small subunit, in Ainsworth et al.69 The number 114 corresponds to tyrosine of the E. coli enzyme and | signifies the same amino acid as found in the E. coli enzyme PROKARYOTES E. coli S. typhimurium Anabaena Synechocystis

ATP Site 114 WYRGTADAV ||||||||| |FQ|||||| |FQ||||||

PLANT SMALL SUBUNIT A. thaliana leaf Barley endosperm Maize endosperm 54kDa Potato tuber 50kDa Rice seed Spinach leaf 51kDa Wheat endosperm

|FQ|||||| |FQ|||||| |FQ|||||| |FQ|||||| |FQ|||||| |FQ|||||| |FQ||||||

PLANT LARGE SUBUNIT A. thaliana leaf Barley endosperm Maize endosperm 60 kDa Potato tuber 51kDa Spinach leaf 54kDa | Wheat endosperm (large subunit)

|FQ|||||| |FR|||||| |FQ||||SI |FQ|||||| |FQ|||||| |FR|||||W

Similar results were otained via reductive phosphopyridoxylation of the Anabaena ADPGlc Ppase.73 The modification was also prevented by 3PGA and Pi. Lys419 was the modified residue and the adjacent sequences about that residue are very similar to those observed for site 1 sequences of the higher plants (Table 1.4). Site-directed mutagenesis of Lys 419 to either Arg, Ala, Gln, or Glu produced mutant enzymes having 25- to 150-fold lower apparent affinities, A0.5 (concentration of activator needed for 50% of maximal activation), than that of wild-type enzyme (Table 1.5). No other kinetic constants such as affinity (km) for substrates and the inhibitor, Pi, were affected. The heat stability or the catalytic efficiency of the enzyme were also not affected. These mutant enzymes, however, were still activated to a great extent at higher concentrations of 3PGA suggesting that an additional site was involved in the binding of the activator. The Lys419Arg mutant was chemically modified with the activator analog, PLP. Modification of Lys382 in the Arg mutant was observed and caused a dramatic alteration in the allosteric properties of the enzyme which could be prevented by the presence of 3PGA or Pi during the chemical modification process. Lys382 was thus identified as another site involved in the binding of the activator and as seen in Table 1.4, the sequence about Lys382 in the Anabaena enzyme is very similar to that seen for the higher plants site 2 which is situated on the large subunit. Thus the site directed

ß 2004, Woodhead Publishing Limited

Table 1.3 The effect of site-directed mutagenesis of Lys residue 198 of the small subunit of potato tuber ADPGlc Ppase on the Km of glucose-1-P. S refers to the 50 kDa (small) subunit and L refers to the 51 kDa (large) subunit of the potato tuber enzyme. The letters, K, R, A and E, are of the one alphabet code corresponding to the amino acids, lysine, arginine, alanine and glutamate Enzyme

Km (mM)

Wild-type (normal) SK198R L-Wt SK198A L-Wt SK198E L-Wt

0.057 7.7 22.0 31.1

mutagenesis along with the reductive phosphopyridoxylation experiments strongly indicate that in higher plants as well as in the cyanobacteria, lysine residues near the carboxyl terminii of the ADPGlc Ppase subunits are part of the binding domain of the allosteric activator. Site 1 of the Anabaena enzyme corresponds to the lysyl residue near the C terminus, Lys440, that is phosphopyridoxylated in the spinach leaf small subunit,71 and corresponds to Lys468 in the rice seed small subunit and to Lys441 in the potato tuber ADPGlc Ppase small subunit. Lys404 of the potato tuber large subunit corresponds to Site 2 of the Anabaena enzyme, Lys382. Table 1.4 Plant and cyanobacterial ADPglucose pyrophosphorylase activator binding sites. The sequences are listed in one letter code and were taken from Smith White and Preiss,63 and from references indicated in the text. The Lys residues, covalently modified by pyridoxal-P are in outline. The potato tuber enzyme Lys residue was identified via site-directed mutagenesis. The numbers 419 and 382 correspond to the Lys residues in the Anabaena ADPGlc Ppase subunit. Site 1 is present in the small subunit of the plant ADPGlc Ppase while Site 2 is present in the large subunit Activator Site 1 419 KNAVITDGTII SGIVVVLK NGIVVVIKNVTIADGTVI

Activator Site 2 382 QRRAIIDK KNAR IRRAIIDKNAR

Higher plants

Activator Site 1, (small subunit)

Activator Site 2, (large subunit)

Arabidopsis thaliana Barley endosperm Maize endosperm Potato tuber Rice seed Spinach leaf Wheat leaf Wheat seed

SGIVTVIKDALIPTGVI SGIVTVIKDALLPSGTVI GGIVTVIKDALLPSGTVI KDALIPSGIII SGIVTVIK SGIVTVIKDALLLAEQLY KDALIPSGTVI SGIVTVIK

IQECIIDKNAR ISNCIIDMNAR IRNCIIDMNAR IRKCIIDKNAK INNCIIDMNAR IKDAIIDK KNAR IKRAIIDKNAR IQNCIIDKNAR

Cyanobacteria Anabaena Synechocystis

SGIVTVIKDALLPSGTVI

ß 2004, Woodhead Publishing Limited

Table 1.5 Effect of mutagenesis of the allosteric activator binding site Lys residues of cyanobacterial and plant ADPGlc Ppases. The amino acid residues are indicated by a oneletter code. The Anabaena ADPGlc Ppase data was obtained from Charng et al.73 and from Sheng et al.74 The potato tuber is from Preiss et al.75 ADPGlc Ppase Anabaena

3PGA A0.5 M

Wild-type K419R K419A K419Q K419E K382R K382A K382Q K382E K419R, K382R

40 1,200 2,200 5,500 6,000 530 1,900 7,800 5,900* 5,200

ADPGlc Ppase Potato tuber

3PGA A0.5 M

Wild-type 50kDa, K441R 50kDa, K441A 50kDa, K441E

100 180 3,000 8,300

*3PGA is not an activator of this mutant. It is an inhibitor and the value represents the concentration that gives 50% inhibition.

Table 1.4 also shows that the amino acid sequence of the spinach leaf small subunit peptide containing the modified lysyl residue of Site 1 is highly conserved in the barley,76 potato tuber, maize,77 rice seed and wheat endosperm small subunits69 and the Anabaena67, 78 and Synechocystis68 ADPGlc Ppase subunits. The amino acid sequence of Site 2 of the spinach leaf large subunit is highly conserved in the barley endosperm,79 maize80 potato tuber, wheat seed and wheat leaf 81 large subunit ADPGlc Ppases. Phenylglyoxal, a modifier of arginine residues causes inactivation of the enzyme that can be prevented by 3PGA or by Pi. This is evidence for one or more arginine residues being present at the allosteric sites of the spinach leaf enzyme.82 Both subunits were labelled when [14C]phenylglyoxal was used. Where the Arg residues are located in the sequence is presently unknown but there is a possibility it may be close to the Lys residue at activator Site 2. cDNA clones encoding the putative mature forms of the large and small subunits of the potato tuber ADPGlc Ppase have been expressed together, using compatible vectors, in an E. coli mutant deficient in ADPGlc Ppase activity.70, 83 The ADPGlc Ppase activity expressed was high and the enzyme displayed catalytic and allosteric kinetic properties very similar to the ADPGlc Ppase purified from potato tuber.83 The enzyme activity was also neutralized by antibody prepared against potato tuber and not by antibody prepared against the E. coli ADPGlc Ppase.83 This expression system is a very useful tool to perform site directed mutagenesis to further characterize the allosteric function of the lysyl residues identified via chemical modification with pyridoxal-P of the spinach enzyme. Indeed in preliminary results, shown in Table 1.5, site directed mutagenesis of Lys441 of the potato ADPGlc Ppase small subunit to Glu and Ala results in mutant enzymes lower in their affinity, 30- to 83-fold, respectively, for 3PGA.75 The

ß 2004, Woodhead Publishing Limited

conservative mutation to arginine resulted in only a twofold increase in A0.5, thus indicating that the positive charge of the cationic amino acid was important for the binding of the activator.

1.3.3 Possible functions of the small and large subunits of the higher plant ADPGlc Ppase The ability to express cDNA clones representing the potato tuber small and large subunits together in E. coli70 to obtain a highly active enzyme enables one to express the subunits separately to determine their specific functions. It was found that the potato tuber small subunit when expressed alone had high catalytic activity provided that the 3PGA concentration was increased to 20 mM. The 3PGA saturating concentration for the expressed transgenic or (normal) potato tuber heterotetrameric enzyme is 3 mM. It was found that the Ka of the transgenic enzyme in ADPGlc synthesis is 0.10±0.16 mM while for the small subunit alone it is 2.4 mM. Thus, the small subunit by itself has about 15- to 24fold lower affinity for the activator. The small subunit expressed alone is more sensitive to Pi inhibition than the transgenic heterotetrameric enzyme with an 8fold lower Ki. The kinetics of 3PGA activation and the Pi inhibition were the main kinetic differences between the homotetrameric small subunit and the recombinant heterotetrameric ADPGlc Ppase. These results are consistent with those obtained for the Arabidopsis thaliana mutant ADPGlc Ppase lacking the large subunit84 where the enzyme had lower affinity for the activator and higher sensitivity towards Pi inhibition than the heterotetrameric normal enzyme.85 The potato tuber large subunit expressed by itself has negligible activity. The dominant function of the small subunit therefore is catalysis while the major function of the large subunit is to affect the sensitivity of the small subunit to allosteric activation and inhibition. N-terminal sequence analysis of the purified catalytic (50 kDa) subunit expressed in E. coli. and of the purified expressed heterotetrameric enzyme having both subunits (50 and 51 kDa) are shown in Table 1.6. The expected initial methionine residue has been cleaved for the most part from both subunits. On the basis that there are equal amounts of tyrosine and valine at residues two for the subunits and roughly equal amounts of valine and aspartate at residues 4, it may be concluded that the expressed heteromeric enzyme has an 2 2 structure.

1.3.4 Comparison of ADPGlc Ppase sequences A high degree of amino acid sequence identity is observed when comparing the sequences of corresponding ADPGlc Ppase subunits from different species, a result that could be expected from the spinach leaf catalytic (lower molecular weight) subunit antibody effectively reacting with the equivalent subunits of the enzymes from maize endosperm,59, 86 rice seed,61, 87 A. thaliana88 and potato tuber.89 The antibody for the lower MW spinach leaf subunit reacts weakly with the regulatory (higher molecular mass) subunit of the other species ADPGlc Ppases. Not much homology, therefore, was expected between sequences of the

ß 2004, Woodhead Publishing Limited

Table 1.6 Amino terminal sequence analysis of the potato tuber ADPGlc Ppase expressed in E. coli either as the homotetrameric 50 kDa subunit or as the heterotetrameric enzyme having both the 50 and 51 kDa subunits. Only those PTH amino acid residues found in substantial number (> 4 pmol) are shown. Other amino acids in the HPLC analysis were of lower amounts. Because the PTH-Ser derivative is less stable it is found in smaller amounts than the other PTH derivatives. The analysis indicates that all of the initial methionine is cleaved from the 50 kDa subunit while about 14% of the 51 kDa subunit retains the initial methionine residue 50 kDa subunit Amino acid Ala Val Ser Asp

1 184 3.7 3.7 6.9

2 8.8 194 3.7 6.9

Residue number pmol 3 4 4.8 6.1 6 3.6 40 4.4 2.1 105

5 10.4 6.2 22 22

Concluded sequence of 50 kDa subunit: Ala-Val-Ser-Asp-Ser 50 and 51 kDa subunits together Amino acid Tyr Val Ser Asp Ile Ala Met

1 3.5 0 7.0 2.7 1.9 244 23.7

2 113 141 6.3 9 10.8 14 3.3

3 0.4 10.8 66.5 7.5 19.5 0 3.0

4 8.5 99 8.5 122 27 15 3.8

5 10 24 42 24 116 17 4

Concluded sequence for the 51 kDa subunit: Ala-Tyr-Ser-Val-Ile

small and large subunits. The degree of identity between the large and small subunits (obtained by Edman degradation or deduced from nucleotide sequences of cDNAs or genomic DNA) is around 40 to 60%.63 Sequence analyses indicate a greater identity between the 54 kDa subunit of the spinach leaf enzyme, the subunit coded by the Sh-2 gene from maize, and the subunit encoded by the cDNA insert, we7, from wheat endosperm81 suggesting that the latter two correspond to the large molecular weight subunit of the ADPGlc Ppase. Because of the relatively low but certain homology between the two subunits of the ADPGlc Ppase it is speculated that they have arisen originally from the same gene. The bacterial ADPGlc Ppase has been shown to be a homotetramer composed of only one subunit.58 The cyanobacterial ADPGlc Ppase has 3PGA as an allosteric activator and Pi as an inhibitor, similar to the enzyme from higher plants,90 and unlike the bacterial enzymes. Both bacterial22, 58 and cyanobacterial91 ADPglucose pyrophosphorylases are homotetrameric, unlike the higher plant enzymes, indicating that regulation by 3PGA and Pi is not related to the heterotetrameric nature of the higher plant enzyme. It is quite possible that during evolution there was duplication of the ADPGlc Ppase gene and then divergence of the genes produced two different genes coding for the two peptides, both required for optimal activity of the native higher plant enzyme.

ß 2004, Woodhead Publishing Limited

As previously indicated above, one can tentatively assign catalytic function to the small subunit of the ADPGlc Ppase and this is consistent with the identity and similarity in sequence between the small subunits isolated from different plants and tissues. In the case of the large subunit in which amino acid sequences have lower similarity to what is observed for the small subunits, it could be postulated that the different large subunits promote or affect different regulatory properties to the heterotetrameric ADPGlc Ppases of different species and in different tissues. As the needs and amounts of starch required for each type of tissue and plant are different, a different sensitivity in regulation of the ADPGlc Ppase may be required and that difference in regulation may be affected by the different large subunits. Because sequences of the large subunits reflect their occurrences in different plant tissues, e.g., leaf, stem, guard cells, tuber, endosperm, root,63 it is possible these sequence differences evoke different allosteric properties for the enzyme from each tissue.

1.3.5 Prediction of the two-dimensional structure of the ADPGlc Ppase subunits Unfortunately, no satisfactory crystals of the ADPGlc Ppase from any source are available for determination of the three-dimensional structure of the enzyme. However, the secondary structures can be analyzed on the basis of their amino acid sequences and by using partial proteolysis to see if the peptide fragments obtained are consistent with the analysis. Hydrophobic cluster analysis (HCA)92 was applied to ADPGlc Ppases from different sources representing enzymes having different allosteric effectors, and from different tissues, such as E. coli, Anabaena, potaoto tuber small subunit and the large subunits from Chlamydomonas, maize embryo, maize endosperm, potato tuber and Arabidopsis thaliana. The HCA shows that the ADPGlc are identical in the position of many clusters, and in others the differences are small (Fig. 1.1). Even though the similarities in amino acid sequences are lower between the bacterial and plant enzymes, all the clusters present in the E.coli enzyme are also present in both subunits of the potato tuber ADPGlc Ppase. There are some insertions and deletions between the sequences but the general pattern of the clusters is not altered due to the insertions not having any buried amino acids. The important point of this analysis is that the small insertions or differences seen among the ADPGlc Ppases are not part of the `core' of the protein and this indicates that the ADPGlc Ppases from different sources share a common folding pattern, despite having different quatenary structure, the plant heterotetramers and the bacterial homotetramers, and different activator specificities (E. coli, fructose-1,6-bis-P and Anabaena and higher plants, 3PGA). If ADPGlc Ppases from different sources have similar three-dimensional structures, the structure of one should help predict the secondary structure of another. The sequences of the above seven ADPGlc Ppase subunits were also analyzed using the PHD program.93 Fig. 1.2 shows the predicted general structure that fits all of these proteins. The ADPGlc Ppase is an alpha/beta protein with some parts being mainly beta as the C-terminal region and what is noted in Fig. 1.2, as Domain 3. The model is consistent with results obtained through controlled

ß 2004, Woodhead Publishing Limited

Fig. 1.1

Hydrophobic cluster analysis plots of the E. coli and Anabaena ADPGlc Ppases. b stands for `buried' amino acids, usually hydrophobic. The asterisk represents glycine residues and black squares represent proline residues.

Fig. 1.2 Profile neural network (PHD) prediction of the secondary structure of the ADPGlc Ppase. Domain 1 contains the fructose-1,6-bis-P activator site, KRAKPAV of E. coli in a loop. Domain 2 has the putative ATP binding site, Tyr114, also in a loop region between a -strand and an -helix starting at GTAD. The glucose-1-P binding site is also seen in a loop series of predicted -strands. The secondary structure between Domains 1 and 2 cannot be determined and is drawn as a dotted line.

proteolysis of the Anabaena and E. coli enzymes with trypsin (Charng, Y. Y. and Preiss, J., unpublished results). Exposed loops would be more sensitive to proteolysis and the protease studies, do show the protease cleaving at sites predicted as loops in the model. Proteolysis is seen in the alpha helix predicted near the C-terminal of the Anabaena enzyme and may be regarded as contrary to the prediction. However this may be regarded as an insertion of about 20 amino acids, absent in the E. coli enzyme and it it is not predicted as being buried in Anabaena in the HCA study. Thus, this helix may be regarded as not part of the core but part of a loop and sensitive to partial proteolysis. It is also important to point out that most of the amino acids shown to have roles in the binding of the substrates, ATP and glucose-1-P, the binding of the activators, are situated in loops or are very close to loops. Also a common super secondary structure, the glycine loop motif, seen in nucleotide binding proteins,94 is also present in the model in what is labelled in Fig. 1.2 as Domain 1. This domain would bind the phosphates of the ATP, and with the Domains 2 beta and alpha helices comprise a Rossman fold where the purine portion of the ATP is bound. It is likely that Domains 1, 2 and 3 form a catalytic domain, having a typical alpha/ beta structure where the substrates bind in the loop parts exposed to the aqueous media. The secondary structure predicted for the ADPGlc Ppase in Domains 1 and 2 is identical to the accepted structure of the oncogenic protein H-Ras21 which is used as one of the folding models for binding of GTP.95

ß 2004, Woodhead Publishing Limited

In Domain 2, the loops on the N- side of the beta sheets (the C-end of the alpha helices) have no amino acids conserved in all the sequences of the known ADPGlc Ppases. This is in agreement with the idea that the ATP binding site is located on the other side of the alpha/beta structure. For topological reasons these loops would not be accesible to the substrate and as a consequence, evolutionary pressure to conserve the amino acids in these loops would be lower than in the loops located at the C-end of the beta sheets.

1.4 Properties of plant glucan synthesizing enzymes: starch synthase 1.4.1 Enzyme properties Starch synthase activity is measured as the transfer of [14C]glucose from ADPglucose into a primer such as rabbit glycogen or amylopectin.96, 97 There are many unknowns regarding starch synthases that remain to be answered. The gaps in our understanding of starch biosynthesis arise partly from the difficulties inherent to the starch granule itself, which is insoluble in water, has a very complicated structure and is still the object of much research and speculation.20 In vivo, starch synthesis occurs by deposition on the granule surface by the concerted action of starch synthases and branching enzyme. Starch synthase activity is associated with the starch granules or in the supernatant of crude extracts. Thus we can have granule-bound forms and soluble forms. The elucidation of the roles of the multiple forms of starch synthase and branching enzyme in the biosynthesis of starch, and the determination of its fine structure will require purification and characterization of each isoform. In maize endosperm there are at least four starch synthases, two soluble98 and at least two granule-bound.97 The number of isoforms may vary with the plant species and the developmental stage, but those that have been studied more carefully seem to have a similar number of isoforms. Indeed as in the case of pea embryo an isozyme of starch synthase, starch synthase II, can exist as a soluble and starch-granule bound.99 The question remains whether the activities soluble and granule bound are both functional. Indeed, Mu Forster et al.,100 have reported that in maize endosperm, more than 85% of the starch synthase I protein may be associated with the starch granule. This was determined by using antibody prepared against the starch synthase. However no evidence was presented to indicate that the starch synthase I was active in the particulate stage.The cDNA clones that encode the two isozymes of granule bound starch synthase of pea embryo are optimally expressed at different times during development;40 while isozyme II is expressed in every organ, isozyme I is not expressed in roots, stipules or flowers.40 Purification of the starch synthase and branching enzymes in large amounts and to a large specific activity has proven to be difficult, and partly for this reason it has not been possible so far to find out how the enzymes interact to produce the two carbohydrates, amylose and amylopectin, that form the starch granule. At

ß 2004, Woodhead Publishing Limited

present, knowledge of the elongating properties of the starch synthase isozymes are unknown and whether there are differences between the isozymes is also unknown. Also whether the isozymes have a preference in elongating A,B or even C chains is not known. Purification of the isoforms to high specific activity and lack of interfering activities will facilitate the characterization of the isoforms with respect to primer specificity and interaction with isoforms of branching enzyme, supplying information about their role in vivo.

1.4.2 Identification of the waxy locus as the structural gene for the granule bound starch synthase Genetic studies implicate one granule-bound starch synthase (GBSS) isoform in the synthesis of amylose. In waxy (wx) mutants there is virtually no amylose, GBSS activity is deficient35, 36, 101 and the wx protein is missing. The final product of the wx locus is a protein of molecular weight 58 kDa associated with the starch granule. Shure et al.36 prepared cDNA clones homologous to wx mRNA and, in subsequent experiments102 restriction endonuclease fragments containing part of the wx locus were cloned from strains carrying the different wx alleles to further characterize the controlling insertion elements activator (ac) and dissociation (ds). Excision of the ds element from certain wx alleles produces two new alleles encoding for wx proteins with altered starch synthase activities.103 The DNA sequence of the wx locus of Zea mays was determined by analysis of both a genomic and an almost full length cDNA clone103 and the wx locus from barley has been cloned and its DNA sequenced.104 Amino-acid sequences are also available for rice,105 potato,38 cassava,106 wheat107 and pea isozymes.40 Table 1.7 compares three regions of the deduced amino acid sequences from barley, cassava, maize, potato, rice, isozyme I of pea embryo and wheat wx clones with the amino acid sequence for the E. coli glycogen synthase,108 and the rice soluble starch synthase.109 Region 1 starts with the first 27 amino acids of the N terminal of the E.coli glycogen synthase. Thirteen are identical to the amino acid sequences deduced for the plant wx proteins. Of particular significance is the sequence starting at residue Lys 15 of the bacterial enzyme ...KTGGL... . The lysine in the bacterial glycogen synthase has been implicated in the binding of the substrate, ADPglucose110 on the basis of the chemical modification of that site by the substrate analogue ADP-pyridoxal. The similarity of sequences between the bacterial glycogen synthase, the soluble starch synthase and the wx protein provides further evidence that the wx gene is indeed the structural gene for the granule bound starch synthase. There are two other regions of high conservation of the various GBSSs with the E.coli glycogen synthase. In region II, only one or two amino acids of the thirteen amino acids are different from the E. coli sequence and in region III, all the GBSSs are completely identical with respect to the amino acid sequence while the bacterial enzyme differs in only two of nine amino acids, an Arg for a Ser and an Ala for a Val.

ß 2004, Woodhead Publishing Limited

Table 1.7 Conserved regions of amino acid sequences of the E. coli glycogen synthase, rice soluble starch synthase and the various granule-bound starch synthases. The numbers preceding the sequence indicate the residue number from the N-terminus in the sequence. The underlined sequence in KTGGL, has been shown for the E. coli glycogen synthase to be involved in binding of the sugar nucleotide substrate.110 References to the other sequences may be obtained from Preiss and Sivak.9

1.4.3 The wx protein is a starch synthase The genetic evidence points to the wx locus as the structural gene for a starch synthase bound to the starch granule. However, direct biochemical evidence was lacking, mainly because of the difficulties involved in studying the proteins associated with starch. Starch was solubilized using amylases, and the starch proteins liberated into the supernatant were fractionated by chromatography on

ß 2004, Woodhead Publishing Limited

DEAE97 The GBSS I was clearly associated with the wx protein (recognized by its mobility on SDS polyacrylamide gels and its reaction with antibodies raised against the pure wx protein) throughout purification. The molecular mass of the GBSSI, determined by gel filtration or by sucrose density gradients, was about 59 kDa.111 Because of the failure to demonstrate that the wx protein from pea endosperm had starch synthase activity, Smith112 suggested that the waxy protein of pea is not the major granule-bound starch synthase. However, when starch was extracted from developing embryos of pea a starch synthase activity was solubilized and was determined to be the waxy protein on the basis of the following results.111 The MW of the pea starch synthase is about 59 kDa, as determined by ultracentrifugation in sucrose density gradients. The pea granulebound starch synthase preparation displayed a relatively high specific activity and when this enzymatic fraction was subjected to SDS polyacrylamide gel electrophoresis it migrated the same as the waxy protein and gave a strong immunoblot with antibody prepared against the waxy protein either from pea embryo or maize.111 Thus the immunological data indicated that the activity assayed by Sivak et al.111 was due to the granule bound starch synthase (waxy protein) and not due to the truncated soluble starch synthase of 60 kDa as suggested by Edwards et al.99 When the gene coding for the mature waxy protein from maize kernel was expressed in E. coli, the recombinant protein had a MW similar to the maize protein as determined by SDS PAGE, reacted with antibody raised against the plant protein and had starch synthase activity (M. N. Sivak, H. P. Guan and J. Preiss, unpublished data). Thus, the biochemical re-examination of starch synthase present in starch granules from two species, maize and pea, strengthens the genetic evidence supporting the role of the wx protein as a granule-bound starch synthase with a major role in the determination of amylose content of starch. It has been shown by many experiments involving anti-sense RNA in potato39, 113 and in rice,114 that disappearance of amylose correlates very well with the loss of wx gene expression. It is possible that the interior of the granule is devoid of branching enzyme or, if branching enzyme is in the granule itself, it is not appreciably active. The presence of an active chain-elongating enzyme, i.e., starch synthase, without an active branching enzyme present (in the presence or absence of some debranching activity), may then explain why amylose formation occurs. However, this situation may be more complicated since more than one isozyme of the GBSS has been found for a number of plants. It is quite possible that the GBSS may also be involved in the initial formation of amylopectin near the exterior portion of the granule along with the soluble starch synthases. In Chlamydomonas reinhardtii a wx mutant deficient in GBSS was isolated.41 In this mutant not only was the isolated starch deficient in amylose but also one of the amylopectin fractions, amylopectin II, was significantly lower. Amylopectin II has longer chains than the amylopectin I fraction as judged from the increase in max of the glucan-I2 complex. When

ß 2004, Woodhead Publishing Limited

GBSS is active, it would not be rate limiting and thus amylose and amylopectin are seen as normal components of the starch granule. When there is a loss of the major GBSS activity, then the rate of formation of the amylose and initial amylopectin structures may be limiting in the wx mutant and only the higher branched amylopectin I fraction would be present. A normal amount of starch is made even if the GBSS activity is deficient and this may suggest that there is sufficient activity of the minor granule-bound starch synthases present or that the soluble starch synthases activity may substitute to form sufficient unbranched chains allowing normal amylopectin synthesis to proceed at the same rate as in the wild type. However these limiting amounts of GBSS isozyme activity are not sufficient to produce substantial amounts of amylose. Alternatively, if the GBSS activity is deficient and is involved in amylopectin synthesis then it is possible that the soluble starch synthase activities may substitute for the normal GBSS function for only amylopectin synthesis.The involvement of GBSS in amylopectin and amylose synthesis synthesis is discussed in a later section.

1.4.4 Characterization of the soluble starch synthases A variety of plant systems have shown the presence of multiple forms of soluble starch synthases (SSS). Studies with barley, maize, pea, rice, sorghum and teosinte seeds, and wheat endosperms, spinach leaf and potato tuber extracts, have indicated the presence of at least two major forms of SSS (reviewed in references 3,4, 9, 24) designated as types I and II. In most cases, SSS I usually elutes from an anion exchange column at lower salt concentrations than SSS II. Although starch synthase I (SSSI) has been partly purified from maize kernels,115, 116 starch synthase II (SSSII), a more unstable isoform, has been more difficult to purify. The properties observed for the isoforms of maize endosperm tissue reflect the properties of the corresponding enzyme forms in other plant materials and the properties of the starch synthase isozymes have been reviewed.9 The apparent affinity for ADPglucose, measured by the km, is similar for the two forms. The maximal velocity of the type I enzyme is greater with rabbit liver glycogen than with amylopectin and the type II enzyme is less active with glycogen than with amylopectin. Citrate stimulation of the primed reaction is greater for type I than for type II. Both forms can use the oligosaccharides maltose and maltotriose as primers when present at high concentrations. Starch synthase I seemed to have more activity than SSSII with these acceptors. The lower activity for SSSI with amylopectin as a primer as compared to glycogen, suggests that SSSI may prefer the short exterior chains (A-chains) that are more prevalent in glycogen than in amylopectin. The reverse may be true for SSSII where it may prefer the longer chains (B-chains) seen in amylopectin. Differences were also noted in the apparent affinities with respect to primer. For example, the km of the type I enzyme for amylopectin is nine times lower than that of the type II enzyme. It is worth noting that the type I enzyme is active

ß 2004, Woodhead Publishing Limited

without added primer in the presence of 0.5 M citrate while the type II enzyme is inactive in these conditions. Citrate decreases the km of amylopectin for both types of enzymes; 160 fold for the type I enzyme and about 16 fold for the type II starch synthase with 0.5 M citrate. The starch synthase isozymes in maize endosperm have different molecular masses. The GBSS isozyme I has a molecular mass of 60 kDa, that of GBSS II 95 kDa, the SSS I a molecular mass of 72 kDa, and SSS II, 95 kDa (reviewed in 9). Mu et al.,117 have reported the molecular mass of maize endosperm SSSI as 76 kDa which is similar to the value reported previously for SSSI. These molecular mass values for the starch synthases are all higher than that of the E. coli glycogen synthase with a molecular weight of 52 kDa.108 It appears that the maize endosperm SSS I and II are immunologically distinct.97 Antibody prepared against maize endosperm SSSI showed very little reaction with SSSII in neutralization tests. In summary, the maize SSSI and II seem to be distinct forms distinguished on the basis of their physical, kinetic and immunological properties and are probably products of two different genes. Because of their different kinetic properties and different specificities with respect to primer activities they may have different functions in the formation of the starch granule. In rice, three isoforms of soluble starch synthase were separated by anion exchange chromatography which, in immunoblot, reacted with antibodies raised to the rice waxy protein.109 After affinity chromatography of the active fractions, amino-terminal sequences were obtained for the protein bands of 55±57 kDa (separated by SDS PAGE) that cross-reacted weakly with serum raised against the rice waxy protein. It is worth noting that this experimental approach does not exclude the possibility that other soluble starch isoforms were present which did not cross-react with the antiserum, and the authors indicate that other results suggest that another soluble starch synthase isoform, with a mw of 66 kDa, is also present in seed extracts. Other forms of starch synthase may be present in plants. Recently, Marshall et al.118 have reported the presence of a starch synthase, 140 kDa, in potato tubers which may account for 80% of the total soluble starch synthase activity. A cDNA representing the protein gene was isolated. Expression of an anti-sense mRNA caused a reduction of about 80% of the soluble starch synthase activity in the tuber extracts. Of interest was that the severe reduction in activity had no effect on starch content or on the amylose/amylopectin ratio of the starch. However, there was a change in the morphology of the starch granules suggesting an alteration in the starch structure. The specific change in structure causing the morphology change remains to be determined.

1.4.5 Cloning of the soluble starch synthases Baba et al.109 isolated cDNA clones coding for the putative soluble starch synthase from maize from an inmature rice seed library in gt 11 using as probes synthetic oligonucleotides designed on the basis of the amino-terminal amino

ß 2004, Woodhead Publishing Limited

acid sequences available. The insert of about 2.5 kb, was sequenced and shown to code for a 1878-nucleotide open reading frame. Comparison with the corresponding amino-terminal sequences led the authors to conclude that the protein is initially synthesized as a precursor, carrying a long transit peptide at the amino acid terminus and that the same gene would be expressed both in seeds and in leaves.

1.4.6 Soluble starch synthase II mutants of Chlamydomonas reinhardtii In order to understand the various functions of the different starch synthases Ball and his associates in Lille, France set out to isolate various mutants of Chlamydomonas deficient in starch synthase activities. They have been successful in isolating a soluble starch synthase II deficient mutant119 and double mutants deficient both in GBSS and in SSS II.120 These studies have provided significant information on the function of both these enzymes and their involvement in amylopectin biosynthesis. The SSSII mutant contained only 20± 40% of the starch seen in the wild-type organism and the percent amylose of the total starch increased from 25% to 55%. This mutant also contained a modified amylopectin which had an increased amount of very short chains (2 to 7 DP) and a concomitant decrease of intermediate size chains (8 to 60 DP). This suggested that the SSSII was involved in the synthesis or maintenance of the intermediate size chains (B-chains) in amylopectin. The higher amylose content could be explained because if an unbranched amylose-like intermediate was a precursor for amylopectin synthesis, in the SSSII mutant it could not be effectively utilized. It is quite possible that this amylose fraction may be more highly branched than the usual amylose. The absorption spectra of its I2 complex has lower maximal wavelength than the wild-type amylose fraction, suggesting that more branching has occurred. The mutant amylose fraction may therefore have a greater amount of branched amylose intermediates on the route to amylopectin biosynthesis. The double mutants defective in SSSII and a GBSS,120 had an even lower starch content, 2% to 16% of the wild-type and the amount of starch present was inversely correlated with the severity of the GBSS defect of the double mutant. The authors suggest that the GBSS is required to form the basic structure of the amylopectin and these effects of the GBSS absence are exacerbated due to the diminished SSSII activity. Of interest is that the SSSI may, in addition to a small amount of starch, synthesize a small water soluble polysaccharide. Analysis of both fractions suggests that they may be intermediate in structure between amylopectin and glycogen with respect to the extent of branching. These studies of the Chlamydomonas mutants by Ball and his colleagues are important in that they provide good evidence for involvement of the GBSS in amylopectin as well as in amylose synthesis and suggest that an important function for SSSII would be in its involvment in synthesis of the intermediate size (B) branches in amylopectin.

ß 2004, Woodhead Publishing Limited

1.5 Properties of plant glucan synthesizing enzymes: branching enzymes 1.5.1 Purification and characterization of the branching enzyme isozymes Assay procedures Branching enzyme has been primarily assayed in three different ways. An iodine assay is based on the decrease in absorbance of the glucan-iodine complex resulting from the branching of amylose or amylopectin by the enzyme. During incubation of the assay mixture containing amylose or amylopectin, aliquots are taken at different times and iodine reagent is added.32, 121 For amylose, the decrease of absorbance is measured at 660 nm and, for amylopectin, at 530 nm. A unit of activity is defined as decrease in absorbance of 1.0 per min. at 30 ëC at the defined wavelength. The phosphorylase-stimulation assay121-123 is based on the stimulation of the 'unprimed' (without added glucan) phosphorylase activity of the phosphorylase a from rabbit muscle as the branching enzyme present in the assay mixture increases the number of non-reducing ends available to the phosphorylase for elongation. One unit is defined as 1 mol transferred from glucose-1-P per min. at 30 ëC. In contrast to the two assays described above, the branch-linkage assay124 is an assay that measures the actual number of branch chains formed by branching enzyme catalysis. The enzyme is incubated with the substrate, a NaBH4-reduced amylose, and the reaction is then stopped by boiling. The product is incubated with purified Pseudomonas isoamylase for debranching. Finally, the reducing power of the oligosaccharide chains transferred by the enzyme is measured by a highly sensitive reducing sugar assay like the Park-Johnson method. Reduction of amylose with borohydride gives about 2% of the reducing power of the nonreduced amylose, resulting in lower blanks. The branching-linkage assay is the most quantitative assay for branching enzyme but amylolytic activity would interfere with this assay. The phosphorylase stimulation assay is the most sensitive particularly if labelled glucose-1-P is used and the I2 assay although not very sensitive, does allow assaying branching enzyme specificity with various -1,4-dextrins and providing information on the possible role of the different branching enzyme isoforms. It is best to employ all three assays when studying the properties of the branching enzymes, but, above all, before studying the branching enzymes properties, they must be purified to the extent that all degradative enzymes are eliminated if reliable information is being sought. Enzyme characterization of the isozymes Maize endosperm has three branching enzyme isoforms.31, 121, 125 Reports on other tissues are consistent with the presence of more than one isoform, such as castor bean.126 BE I, IIa and IIb from maize kernels31, 121, 125 have been purified to the extent that they no longer contained amylolytic activity.31, 125 Molecular weights were 82,000 for isoform I and 80,000 for isoforms IIa and IIb.121, 122

ß 2004, Woodhead Publishing Limited

Table 1.8 summarizes the properties of the various maize endosperm BE isozymes from the studies of Takeda et al.124 and Guan and Preiss.31 Of the three isoforms, BEI had the highest activity (using the iodine assay) in branching amylose (Table 1.8) and its rate of branching amylopectin was about 3% of that with amylose. In contrast, the BEIIa and IIb isoforms branched amylopectin at twice the rate they branched amylose, and catalyzed branching of amylopectin at 2.5 to 6 times the rate observed for BEI. Takeda et al.124 have analyzed the branched products made in vitro, from amylose by each BE isoform. This was done by debranching the products of each reaction using isoamylase, followed by gel filtration. BEIIa and BEIIb are very similar in their affinity for amylose and the size of chain transferred. When presented with amyloses of different average chain length, the three BEs have higher activity with the longer chain amylose, but while BEI could still catalyze the branching of an amylose of average c.l. of 197 with 89% of the activity shown with the c.l. of 405, the activity of BEII dropped sharply with chain length. The study of the reaction products showed that in vitro, the action of BEIIa and BEIIb results in the transfer of shorter chains than those transfered by BEI. These results suggest that BEI catalyzes the transfer of longer branched chains and that BEIIa and IIb catalyze transfer of shorter chains. Thus, it is quite possible that BEI may produce slightly branched polysaccharides which serve as substrates for enzyme complexes of BE II isoforms and starch synthases to synthesize amylopectin; BE II isoforms may play a major role in forming the short chains present in amylopectin. Also BEI may be more involved in producing the more interior (B-chains) chains of the amylopectin while BE IIa and BEIIb would be involved in forming the exterior (A) chains.

Table 1.8 Properties of the maize endosperm branching isozymes. The units are for phosphorylase stimulation and branching linkage assays, mol/min. and for the iodine stain assay, a decrease of one absorbance unit per min. Branching enzymes

BE I

BE IIa

BE IIb

Phosphorylase stimulation (a)

1196

795

994

Branching linkage assay (b)

2.6

0.32

0.14

Iodine stain assay (c) Amylose (c1) Amylopectin (c2)

800 24

29.5 59

39 63

Ratio of activity a/b a/c1 a/c2 c2/c1

460 1.5 49.8 0.03

2484 27 13.5 2

7100 25 15.8 1.6

ß 2004, Woodhead Publishing Limited

In potato tubers, Vos-Scheperkeuter et al.127 purified a single form of branching activity of 79 kDa molecular mass. Antibodies were prepared to the native potato enzyme and they were found to react strongly with maize BEI and very weakly with maize BEIIb. In neutralization tests the antiserum inhibited the activities of both the potato tuber BE and of maize BEI. It was concluded that the potato branching enzyme shows a high degree of similarity to the maize BEI and to a lesser extent with the other maize BE. However, whether potato tubers have two isoforms of branching enzyme such as BEI and BEII has not been determined. Borovsky et al.128 isolated from potato tubers a BE of molecular mass of 85 kDa. This is close to the mass of 79 kDa found by Vos-Scheperkeuter et al.127 Recently, it has been claimed that branching enzymes of molecular mass of 97 and 103 kDa can be isolated129,130 and these results suggested that the previous lower molecular mass values of 79 and 85 kDa are the results of proteolysis during purification of the 103 kDa BE . Khoshnoodi et al.131 showed that limited proteolysis of the 103 kDa enzyme either with trypsin or chymotrypsin produced an enzyme, still fully active, of a molecular size of 80 kDa. Indeed, up to four cDNA clones have been isolated for BE; one for 91 to 99 kDa.131±133 All of these allelic clones have sequences similar to the BEI type. It is still not resolved whether or whether the 97 and 80 kDa proteins could be the products of different allelic forms of the BE gene or different BE genes. Also the sbeIc allele codes for a mature enzyme of 830 amino acids and a molecular weight of 95,180. The sbeIc BE protein product, expressed in E. coli, migrates as a 103 kDa protein.131 It is of interest to note that BE isolated from other plants or bacteria or mammals have molecular masses ranging from 75 to about 85 kDa. These molecular masses have been consistent with the molecular weights obtained from deduced amino acid sequences obtained from isolated genes or cDNA clones. Mizuno et al.29 has reported four forms of branching enzyme from immature rice seeds that were separated by chromatography on DEAE-cellulose chromatography. It seems that two of the forms, BE1 and BE2 (composed of BE2a and BE2b) were the major forms while BE3 and BE4 were minor forms comprising less than 10% of the total branching enzyme activity. The MW of the branching enzymes were, BE1, 82 kDa, BE2a, 85 kDa, BE2b, 82 kDa, BE3, 87 kDa, BE4a, 93 kDa and BE 4b 83 kDa. However, BE1 and 2a and b seem to be immunologically similar in their reaction to maize endosperm BEI antibody. Moreover, the rice seed BE1, BE2a and BE2b had very similar N-terminal amino acid sequences. All three BEs had two N-terminal sequences, TMVXVVEEVDHLPIT and VXVVEEVDHLPITDL. The latter sequence is very similar to the first but lacking just the first two N-terminal amino acids. Thus, although these activities came out in separate fractions from the DEAEcellulose column they seem to be the same protein on the basis of immunology and N terminal sequences, although BE2a, is 3 kDa larger. Antibody raised against BE3 reacted strongly against BE3 but not towards BE1 and 2a, 2b. Thus,

ß 2004, Woodhead Publishing Limited

rice endosperm as noted for maize endosperm had essentially two different isoforms of BE. Because of the many isoforms existing for the rice seed branching enzymes, Yamanouchi and Nakamura134 studied and compared the BEs from rice endosperm, leaf blade, leaf sheath, culm and root. The BE activity could be resolved into two fractions, BE1 and BE2, and both fractions were found in all tissues studied in different ratios of activity. On native gel electrophoresis rice endosperm BE2 could be resolved into two fractions, BE2a and BE2b. Of interest was that electrophoresis of the other tissue BE2 forms, only BE2b was found. BE2a was detected only in the endosperm tissue. It appears that in rice there could be tissue specific isoforms of BE. Three forms of branching enzyme from developing hexaploid wheat (Triticum aestivum) endosperm have been partially purified and characterized.135 Two forms are immunologically related to maize branching enzyme I and one form with maize BE II. The N-terminal sequences are consistent with these relationships. The wheat BE IB gene is located on chromosome 7B while the wheat BE IAD peptides genes are located on chromosomes 7A and D. The BE classes in wheat are differentially expressed during endosperm development in that BE II is constitutively expressed throughout the whole cycle while BE IB and BE IAD are expressed in late endosperm development.

1.5.2 Genetic studies on branching enzyme deficient mutants There are some maize endosperm mutants which appear to increase the percentage amylose of the starch granule. The normal starch granule contains about 25% of the polysaccharide as amylose with the rest as amylopectin. In contrast, amylose extender mutants may have as much as 55±70% of the polysaccharide as amylose and may have an amylopectin fraction with fewer branch points and with the branch chains longer in length compared to those of normal amylopectin. Results with the recessive maize endosperm mutant, amylose extender, ae, originally suggested that ae is the structural gene for either branching enzyme IIa or IIb122, 136, 137 as activity of BEI was not affected by the mutation. In gene dosage experiments, Hedman and Boyer138 reported a near-linear relationship between increased dosage of the dominant ae allele and BE IIb activity. Since the separation of form IIa from IIb was not very clear, it is possible that the ae locus was also affecting the level of IIa. Singh and Preiss125 concluded that although some homology exists between the three starch branching enzymes, there are major differences in the structure of branching enzyme I when compared to IIa and IIb as shown by its different reactivity with some monoclonal antibodies, and differences in amino acid composition and in proteolytic digest maps. It was also concluded125 that branching enzymes IIa and IIb are very similar and perhaps the product of the same gene. However, recent studies by Fisher et al.,139, 140 in analyzing 16

ß 2004, Woodhead Publishing Limited

isogenic lines having independent alleles of the maize amylose extender (ae) locus, suggest that BEIIa and BEIIb are encoded by separate genes and the BEIIb enzyme is encoded by the ae gene. They isolated a cDNA clone labelled Sbe 2b, which had a cDNA predicted amino acid sequence at residues 58 to 65 exactly as the N-terminal sequence of the maize BE IIb that they had purified.139 Moreover, they did not detect in ae endosperm extracts any mRNA with the Sbe 2b cDNA clone. Some BE activity was observed in the ae extracts which chromatographed similarly with BE IIa. The finding that the enzyme defect in the ae mutant is BEIIb is consistent with the finding that, in vitro, BEII is involved in transfer of small chains. Besides having an increase in amylose, the altered amylopectin structure in the ae mutant consists of fewer and longer chains and a lesser number of total chains. In other words, there are very few short chains. The wrinkled pea has a reduced starch level; about 66±75% of that seen in the round seed and whereas the amylose content is about 33% in the round form, it is 60±70% in the wrinkled pea seed. Edwards et al.141 measured the activities of several enzymes involved in starch metabolism in wrinkled pea at four different developmental stages. In this variety it was found that branching enzyme activity was, at its highest, only 14% of that seen for the round seed. The other starch biosynthetic enzymes and phosphorylase had similar activities in the wrinkled and round seeds. These results were confirmed by Smith142 who also showed that the r (rugosus) lesion (as found in the wrinkled pea of genotype rr ) was associated with the absence of one isoform of branching enzyme. Edwards et al.141 proposed that the reduction in starch content observed in the mutant seeds is caused indirectly by the reduction in BE activity through an effect on the starch synthase. The authors suggested that, in the absence of branching enzyme activity, the starch synthase forms an -1->4-glucosyl elongated chain which is a poor glucosyl acceptor (primer) for the starch synthase substrate, ADPGlc, therefore decreasing the rate of -1->4-glucan synthesis. Indeed, in a study of rabbit muscle glycogen synthase143 it was found that continual elongation of the outer chains of glycogen caused it to become an ineffective primer, thus decreasing the apparent activity of the glycogen synthase. The observation that ADPglucose in the wrinkled pea accumulates to higher concentrations than in round or normal pea, was considered evidence that activity of the starch synthase was restricted in vivo. Under optimal in vitro conditions, in which a suitable primer like amylopectin or glycogen is added, starch synthase activity in the wrinkled pea was equivalent to that found in the wild type. Amylose extender mutants have been found in rice and studied.144 The alteration of the starch structure is very similar to that reported for the maize endosperm ae mutants. The defect is BE3 isozyme and BE3 of rice is more similar in amino acid sequence to maize BEII than to BEI.33, 144 Thus rice BE3 may catalyze the transfer of small chains rather than long chains.

ß 2004, Woodhead Publishing Limited

1.5.4 Isolation of cDNA clones encoding the branching enzyme isozyme genes The r locus of pea seed has been cloned by using an antibody towards one of the pea branching enzyme isoforms and screening a cDNA library,145 It appears that the branching enzyme gene in the wrinkled pea contains an 800 bp insertion causing it to express an inactive branching enzyme. The authors indicated that the sequence of the 2.7 kb clone showed over a 50% homology to the glycogen branching enzyme of E. coli 146 and proposed that the cDNA that they had cloned corresponded to the starch branching enzyme gene of pea seed. The glg B gene sequence has been determined for a cyanobacterium147 and its deduced amino acid sequence has extensive similarity to the amino acid sequence (62% identical amino acids) in the middle area of the E. coli protein. It appears, therefore, that branching enzymes in nature have extensive homology irrespective of the degree of branching of their products, which is higher (about 10% -1,6 linkages) in glycogen, the storage polysaccharide in enteric bacteria and in cyanobacteria, than in the amylopectin (about 5% -1,6 linkages) present in higher plants. cDNA clones of genes representing different isoforms of branching enzyme from various plants have been isolated from potato tuber,131±133 from maize kernel,31, 32, 139, 140 from cassava148 and from rice seeds (branching enzyme I; references, 29, 149, branching enzyme 3, reference 144). The cDNA clones of the maize BE I and BE II have been over-expressed in Escherichia coli and purified.32, 33 The transgenic enzymes had the same properties as seen with the natural maize endosperm BEs with respect to specific activity and specifity towards amylose and amylopectin.31

1.5.5 Reserve tissue branching enzyme is localized in the plastid The data available on the localization of branching enzyme within the plastid have been obtained with potato150 using antibodies raised against potato BE and immunogold electron microscopy. The enzyme (which would be the equivalent of the BEI isoform of maize, as discussed above) was found in the amyloplast, concentrated at the interface between stroma and starch granule, rather than throughout the stroma, as is the case with the ADPglucose pyrophosphorylase.151 This would explain how amylose synthesis is possible when the enzyme responsible for its formation, i.e., the wx protein, one of the granulebound starch synthases, is capable of elongating both linear and branched glucans. The spatial separation of branching enzyme and granule-bound starch synthase, even if only partial, would allow the formation of amylose without it being subsequently branched by the branching enzyme. However, even if spatial separation did not exist, starch crystallization would have the same effect, i.e., prevent further branching. M. Morell and J. Preiss (unpublished experiments) found about 5% of the total branching enzyme activity associated with the starch granule after amylase digestion. Whether this branching enzyme was similar to the soluble branching enzymes was not determined, but M. Sivak,9 using SDS

ß 2004, Woodhead Publishing Limited

polyacrylamide electrophoresis, found among the proteins present in maize and pea starch of maize starch, a polypeptide of about 80 kDa that reacted with antibodies raised against maize BEII. It is worth noting, however, that small amounts of BE are expected to sediment with the starch granule because of its affinity for the polysaccharide. These results have been confirmed by MuForster et al.100

1.5.6 Branching enzyme belongs to the -amylase family The relationship in amino acid sequences between that of branching enzyme (BE) and amylolytic enzymes such as -amylase, pullulanase, glucosyltransferase and cyclodextrin glucanotransferase was first reported by Romeo et al.152 particularly at those sequences believed to be contacts between the substrate and the hydrolytic enzymes. Baba et al.,153 reported that there was a marked conservation in the amino acid sequence of the four catalytic regions of amylolytic enzymes in maize endosperm BE I. As shown in Table 1.9, four regions that putatively constitute the catalytic regions of the amylolytic enzymes are conserved in the starch branching isoenzymes of maize endosperm, rice seed, potato tuber and the glycogen branching enzyme of E. coli. A very good and extensive analysis of this high conservation in the -amylase family has been reported by Svensson and her group154, 155 with respect to sequence homology and also in the prediction the ( =†8-barrel structural domains with a highly symmetrical fold of eight inner, parallel -strands, surrounded by eight helices, in the various groups of enzymes in the family. The ( =†8-barrel structural domain was determined from the crystal stucture of some -amylases and cyclodextrin glucanotransferases. The conservation of the putative catalytic sites of the -amylase family in the glycogen and starch branching enzymes may be anticipated as the BE catalyzes two consecutive reactions; cleavage of an -1,4-glucosidic linkage in an 1,4-D-glucan chain yielding a non-reducing end oligosaccharide chain which is transferred to a C 6 hydroxyl group of the same chain, or to another 1,4--Dglucan, with formation of an -1,6-glucosidic linkage.

1.5.7 Amino acid residues that are functional in branching enzyme catalysis As indicated in Table 1.9, four regions which constitute the catalytic regions of the amylolytic enzymes are conserved in the starch branching isoenzymes of maize endosperm, rice seed, potato tuber and the glycogen branching enzymes of E. coli.154, 155 It would be of interest to know whether the seven highly conserved amino acid residues of the -amylase family listed in bold letters in Table 1.8, aspartate are also functional in branching enzyme catalysis. Further experiments such as chemical modification and analysis of the threedimensional structure of the BE would be needed to determine the precise functions and nature of its catalytic residues and mechanism.

ß 2004, Woodhead Publishing Limited

Table 1.9 The primary structures of branching enzymes showing the four most conserved regions of the -amylase family. The sequences have been derived from references cited in the text. Only two examples of enzymes from the amylase family are shown for comparison. Over 40 enzymes ranging from amylases, glucosidases, various -1,6-debranching enzymes as well as four examples of branching enzymes are compared by Svensson (154) The invariant amino acid residues believed to be involved in catalysis are in bold letters

Potato tuber BE Maize endosperm BE 1 Maize endosperm BE 11 Rice seed BE 1 Rice seed BE 3 E. coli glycogen BE B. subtilis -amylase B.sphaericus cyclodextrinase

Region 1

Region 2

Region 3

Region 4

355 277 315 271 337 335 100 238

424 347 382 341 404 400 171 323

453 402 437 396 459 453 204 350

545 470 501 461 524 517 261 414

DVVHSH DVVHSH DVVHSH DVVHSH DVVHSH DWVPGH DAVINH DAVFNH

GFRFDGITS GFRFDGVTS GFRFDGVTS GFRFDGVTS GFRFDGVTS ALRVDAVAS GFRFDAAKH GWRLDVANE

VTMAEEST TVVAEDVS VTIGEDVS TIVAEDVS ITIGEDVS VTMAEEST FQYGEILQ IIVGEVWH

CVTYAESHD CIAYAESHD CVTYAESHD CVTYAESHD CVTYAESHD NVFLPLNHD LVTWVESHD SFNLLGSHD

Indeed, the seven highly conserved amino acid residues of the -amylase family appear to be also functional in branching enzyme catalysis. Preliminary experiments,156 where amino acid replacements have been done by site-directed mutagenesis do suggest that the conserved Asp residues of regions 2 and 4 and the Glu residue of region 3 (Table 1.9, in bold letters) are important for BE II catalysis. Their exact functions however, are unknown and further experiments such as chemical modification and analysis of the three-dimensional structure of the BE would be needed to determine the precise functions and nature of its catalytic residues and mechanism. Arginine residues are also important as suggested by chemical modification with phenylglyoxal157 as well as histidine residues as suggested by chemical modification studies with diethyl pyrocarbonate (K. Funane and J. Preiss, unpublished experiments). It would also be of interest to determine the regions of the C-terminus and N-terminus which are dissimilar in sequence and in size in the various branching isoenzymes. It may be these areas that are important with respect to BE preference with respect to substrate (amylose-like or amylopectinlike) as well as in size of chain transferred or to the extent of branching.

1.6

Initiation of starch synthesis using a glucosyl-protein

Initiation of starch synthesis via a glucosyl-protein as seen for glycogen synthesis is a viable hypothesis. Tandecarz and Cardini158 described a system which comprises at least two enzymatic reactions in which proplastid membranes from potato tuber glucosylate a membrane protein at a serine or threonine residue using UDPGlucose to form a glucoprotein. This product, a 38 kDa protein, in turn, is used as an acceptor for a long chain of glucoses sequentially added in a -1,4 bond using either ADPglucose or UDPglucose as donors. This system has been further characterized and one of the enzymes has been purified159, 160 and the potato enzyme catalyzes its own glycosylation161. The reaction requires Mn2+ and thus the reaction is similar to the selfglycosylation carried out by glycogenin. However, although the enzymatic formation of the putative glucosyl-protein has been demonstrated in maize endosperm162 not much information is available yet on the fate of the putative glucan protein.

1.7

Locating starch synthesis in plants: the plastid

The site of starch synthesis in leaves and other photosynthetic tissues is the chloroplast and literature on confinement of the starch biosynthetic enzymes to leaf chloroplasts is reviewed in Preiss4 and Okita.163 The starch formed during the day is degraded at night and the carbon is utilized to synthesize sucrose. All the starch biosynthetic enzymes are present solely in the chloroplast. The seed imports carbon and energy from the source tissues in the form of sucrose. The

ß 2004, Woodhead Publishing Limited

site of starch synthesis in the seed is the amyloplast, a non-photosynthetic organelle. Amyloplasts resemble chloroplasts in that they are enclosed by an envelope comprising two membranes164 and in that they develop from proplastids. Sucrose cannot be taken up by the amyloplast because the inner envelope is practically impermeable to sucrose and thus other metabolites derived from sucrose are taken up by the amyloplast. Although the events that lead to the flow of carbon into starch have been fairly well established for photosynthetic tissues, the situation is far less clear for storage tissues. This is because the amyloplast, the organelle in which starch is stored in sink tissues, is even more fragile than the chloroplast. The metabolism of the amyloplast, which is dependent on the cytosol for carbon and energy, is bound to differ, in many ways, from that of the chloroplast, which generates ATP and fixes CO2. Information regarding the amyloplasts has been obtained in a number of ways, e.g., localization of the starch biosynthetic enzymes using immunocytochemical studies, measurement of enzyme activity in isolated amyloplasts, and measurement of uptake of labelled metabolites by isolated plastids. To study the metabolism of a plastid it is essential to isolate active plastids that are intact, free of cytosolic contamination and of other organelles and in good yield. If the isolated plastids are good enough, they can provide reliable information on the enzymes present in them, what metabolites they can take up, at what rate, whether transport of a particular metabolite is passive or active. Keeling et al.165 supplied developing wheat endosperm with glucose or fructose labelled in [1-13C]- or in [6-13C]- and then examined the extent of redistribution of 13C between carbons 1 and 6 in the starch glucosyl moieties. The redistribution that would have been expected if carbon flow into starch had been by the C3 pathway via triosephosphate isomerase was lower (12±20%). The authors suggested that hexose monophosphates (rather than triose phosphates) were more likely to be the main source of energy and carbon for the amyloplast. The major carbon transport system for the wheat grain amyloplast does not seem to involve triose-P and most likely, involves hexose-P as the transport metabolite. Entwistle et al.166 found that wheat endosperm lacks significant amyloplastic fructose-1,6-bisphosphatase, an enzyme that would be required if a triose-P/Pi transport system (like the one in the chloroplast envelope) were required for starch synthesis. In search of a transport system capable of supplying carbon for starch synthesis in the wheat endosperm, Tyson and ap Rees167 incubated intact amyloplasts with different 14C-labelled compounds, i.e., glucose, glucose-1-P, glucose-6-P, fructose-6-P, fructose-1,6-bisP, dihydroxyacetone-P and glycerol-P. Only glucose-1-P was incorporated into starch and this incorporation was dependent on the integrity of the amyloplast. These results are consistent with the results of Keeling et al.165 Direct import of six carbon compounds has been reported for amyloplasts of potato, fava beans,168 maize endosperm, and other tissues.169 Hill and Smith170 reported that

ß 2004, Woodhead Publishing Limited

glucose 6-phosphate was the preferred metabolite for starch synthesis by pea embryo amyloplasts and that ATP was also required. In pea roots171, 172 the Pi translocator is active with dihydroxy-acetone-P, 3PGA, gluose-6-P and P-enolpyruvate. There appears, therefore, to be a great diversity between the various translocators that exchange Pi with phosphorylated compounds (for review see references 169 and 173) and it seems that the major transport system for most reserve non-photosynthetic plant systems may be at the hexose-P level and not at the triose-P level. Studies for the past five years are consistent with this view. A glucose -6-P translocator was found in intact cauliflower-bud plastids174±176 and in maize endosperm.175 A glucose-1-P translocator for storage tissue amyloplasts of potato suspension cultured cells has also been reported.177 In all these cases the uptake of the hexose-P into the amyloplast was much higher than that observed for dihydroxy-acetone-P. Also if six-week-old cut spinach leaves were incubated in a 50 mM glucose solution for over four days, a glucose-6-P transporter was induced in the chloroplasts178. Fruit chloroplasts assimilate very little of the CO2 that is fixed in the leaves but appear to import carbohydrates. The intact chloroplasts of green pepper fruits179 and tomato fruit chloroplasts and chromoplasts180 also have systems that translocate hexosephosphates. In the tomato systems the solubilzed envelope proteins were reconstituted into liposomes and it was found that the leaf chloroplasts translocated with Pi, dihydroxy-acetone-P and 3-PGA and had low activity with glucose-6- or glucose-1-P. However, the fruit chloroplast and chromoplast envelope proteins in additon to good translocation with the triose phosphates had good translocation activity with P-enol-pyruvate, glucose-6-P and glucose-1-P. The properties of the glucose-6-P translocater of the cauliflower-bud and maize endosperm plastids175, 176 and of the green pepper fruit chloroplast179 are interesting. The translocator identified in the cauliflower-bud plastid is 31.6 kDa.175 Translocation of glucose-6-P is measured by its incorporation of its label into the plastid starch. This incorporation was stimulated 6- to 40-fold by the presence of ATP and 3-PGA. The authors176 interpret the effect due to the need for ATP and 3PGA for starch synthesis. They postulate that these compounds in the cytosol act as feed-forward signals for starch synthesis and are also translocated into the plastid and utilized for synthesis of ADP-glucose. 3PGA is the allosteric activator of ADPGlc Ppase, the enzyme synthesizing ADP-glucose. The glycolytic scheme in the amyloplast may then take on a more important function than the one it has in the chloroplast in that it would contribute to the production of amyloplastic ATP. Thus, the concentration of 3PGA could be a indicator of the ATP supply and of the availability of carbon in the amyloplast. If this were true, the regulatory effect of 3PGA on the ADPGlc Ppase from non photosynthetic tissues (i.e. stimulation and/or reversal of its inhibition by Pi) would have a physiological role similar to the one it has for the leaf enzyme. Two recent reports have indicated, however, that a significant portion of the ADPGlc Ppase activity may be present in the cytosol and not solely in the

ß 2004, Woodhead Publishing Limited

amyloplast. Before these reports it was generally thought that the ADPGlc Ppase was exclusively in the amyloplast as indicated in studies with soybean cell culture,181 wheat endosperm,182 pea embryo142 and oilseed rape embryos.183 In barley endosperm isolation of amyloplasts was achieved with intactness ranging from 41 to 89%.184 The amount of total endosperm activity of plastidial enzymes, recovered in the amyloplast, starch synthase and alkaline pyrophosphatase, ranged from 13 to 17% while the percentage of total ADPGlc Ppase activity residing in these amyloplasts was 2.5%. On this basis, the amount of ADPGlc Ppase activity residing in the amyloplast was estimated to be 15% with the rest present in the cytosol. On the basis of antibody studies two different isoforms of the ADPGlc Ppase were detected, one mainly in the cytosol and the other mainly plastid. The authors indicate, however, that there is an excess of ADPGlc Ppase activity in the amyloplast to accommodate the starch synthetic rate and the function of the cytosolic ADPGlc Ppase to them at present is obscure. Moreover, the kinetic propeties of the plastid ADPGlc Ppase have not been characterized. In another report,185 preparations enriched in maize endosperm plastids contained 24 to 47% of the total activity of the plastid marker enzymes, starch synthase and alkaline pyrophosphatase, while only containing 3% of the total ADPGlc Ppase activity. On this basis they estimated that more than 95% of the ADPGlc Ppase activity was non-plastid. Using antibodies prepared against the bt2 subunit of the maize endosperm ADPGlc Ppase it was demonstrated that most of the bt2 protein was confined to the supernatant but also occurred in the plastid. In bt2 mutant kernels, the cytosolic bt2 protein was not detected but there was a plastidial form of ADPGlc Ppase observed. These data are somewhat different from those obtained by Miller and Chourey186 and by J. L. Prioul (private communication) who detected the bt2 protein by immunogold labelling in the amyloplast. Miller and Chourey186 however, could not eliminate the possibility that there was also a cytosolic ADPGlc Ppase. There may be many different, possible routes for ADPGlc synthesis and these are shown in Fig. 1.3. The carbon may be translocated into the plastid via a glucose-6-P or a glucose-1 P translocator and then converted to ADPGlc via plastidial phosphoglucomutase and ADPGlc Ppase catalysis.175 The results of Denyer et al.,185 indicate that ADPGlc synthesis in the amyloplast can be catalyzed via a plastidial ADPGlc Ppase, either containing the bt2 protein as one of its subunits, or by another ADPGlc Ppase isozyme. Since the major activity of ADPGlc synthesis resides in the cytosol according to Denyer et al.185 and if ADPGlc is synthesized in the cytosol then it must be translocated into the plastid. An ADPGlc transporter then would be required. Thus far, no protein having those properties has been identified. ADPGlc uptake by the Acer pseudoplatanus amyloplasts for starch synthesis has been reported.187 However, as shown by Borchert et al.172 and by Batz et al.,188 the ADPGlc transport may not be physiologically relevant. In vitro, ADPGlc may be translocated via the ATP/ADP translocator since both ADP and ATP at concentrations lower than their physiological concentrations effectively inhibit ADPGlc uptake in the pea root and cauliflower-bud amyloplasts. The hypothetical ADPGlc transporter if present, remains to be characterized.

ß 2004, Woodhead Publishing Limited

Fig. 1.3 Proposed transport of carbohydrates into maize amyloplasts. The figure shows metabolism of sucrose to glucose-6-P, glucose-1-P and ADPGlc in the cytosol and their transport into the amyloplast. The enzyme reactions shown are, 1, sucrose synthase, 2, fructokinase 3, UDPGlc PPase 4, cytosolic P-hexoseisomerase 5, cytosolic Pglucomutase 6, cytosolic ADPGlc PPase 7, plastidial P-glucomutase 8, plastidial ADPGlc PPase and 9, starch synthase.

There is a possibility that the bt1 gene product maybe the ADPGlc transporter. The bt1 gene encodes a plastidial membrane associated protein189, 190 whose deduced amino acid sequence shows similarity to known adenine nucleotide transporters.191 The bt1 mutant is starch deficient and shows a high level of ADPGlc concentration in the endosperm compared to the normal endosperm.192 Thus, the bt1 protein remains to be studied in order to determine its exact function. Thus, much remains to be done with respect to the characterisitics of the plastidial and cytosolic barley and maize endosperm ADPGlc Ppases with respect to their functions in starch synthesis and with respect to what is the more important route for hexose incorporation into starch.

1.8

In vivo synthesis of amylopectin

The reactions of starch biosynthesis, both in the leaves and in storage tissues, have been described above. The regulation of the first enzyme in the pathway, the ADPglucose pyrophosphorylase, is described below and as will be indicated, it has been possible to increase the starch content of potato tuber and tomato fruit, the first time that an increase in the accumulation of a useful natural product has been achieved by genetic transformation.193, 194 However, it is still

ß 2004, Woodhead Publishing Limited

not possible to provide a precise description of how synthesis of the starch granule starts, how amylopectin and amylose are made, or why starch granules from different species differ in their size, number per cell and composition. The cluster model of amylopectin structure as postulated by Hizukuri19, 21 is the currently accepted structure. A feature of the model is the clustering of the -1,6 linkage branch points in certain regions of the amylopectin and the occurrence of B-chains of varying sizes B1, about 19 glucose units long, B2 with 41, B3 with 69, and B4, about 104±115 glucose units long. The number of B3 and B4 chains are few as compared to the number of B2 and B3 chains. The B1 chains extend into only one area of clusters while the B2, B3 and B4 chains extend into two, three and four cluster areas of -1,6 branch linkage areas, respectively, and these areas are usually separated by 39 to 44 glucose units.19 What roles would the starch synthases (SS) and branching enzyme isozymes play in the formation of the crystalline starch granule and amylopectin structures? How is amylose formed? Why are starch granules from different species different in size and in the number per cell? These differences most probably are related to the SS specificities in chain elongation and in the size transfer of glucose chain units by BE and where the -1,6 bond is formed after transfer. As indicated above, in the discussion of BEs,131, 24 it was shown that in vitro, BEI transfers long chains (DP 40 to >100) while BEII transfers only shorter size chains (DP 6 to 14). Amylose is the preferred substrate for BEI while amylopectin is the preferred substrate for BEII. Thus, BEI may be more involved in synthesis of the interior B-chains while BEII is involved in the synthesis of exterior A- and B1-chains. However, it should be pointed out that most probably in the in vivo situation, amylose is not the actual substrate for either branching enzyme I or II and that synthesis of the branched amylopectin occurs via continual elongation of A- and B-chains by the starch synthases and then branching by the BE isozymes. Nevertheless the mode of actions seen in vitro, appear to occur to some extent in vivo as the isozymes were expressed in E. coli and maize BE I did transfer longer chains than BEII and BEII transferred shorter chains.195 On the basis of the above in vitro data obtained with the BE isozymes of maize and on mutant data obtained with the starch synthases, a pathway leading to amylopectin and amylose can be postulated. As mentioned before, Ball and associates have isolated various mutants of Chlamydomonas deficient in starch synthase activities; a granule bound starch synthase (GBSS) deficient mutant, a soluble starch synthase II (SSSII) deficient mutant119 and a double mutant deficient both in GBSS and in SSS II.120 As indicated in the discussion of the starch synthases, the SSSII mutant had only 20±40% of the wild-type starch content and the amylose fraction of the starch increased from 25 to 55%. This mutant also had a modified amylopectin with an increased amount of short chains of 2 to 7 DP, and a decrease of intermediate size chains of 8 to 60 DP. This suggests that the SSSII is involved in the synthesis or maintenance of the intermediate size chains (mainly Bchains) in amylopectin. The higher amylose content could be explained because of the failure of the SSSII mutant to make extended chains.

ß 2004, Woodhead Publishing Limited

Studies prior to those on the GBSS of Chlamydomonas showed that a deficiency of GBSS activity in plants such as maize, rice, barley and sorghum endosperms and in potato tuber resulted in a loss of the amylose fraction in the starch granule with little effect on the amylopectin fraction. Thus, GBSS was considered to play a major role in amylose synthesis. Moreover, the role of GBSS in amylose synthesis in potato plants was also demonstrated by transforming them to produce antisense RNA from a gene construct containing GBSS cDNA in reverse orientation. Total suppression of GBSS activity gives rise to tubers containing amylose-free starch.39 Once again, the amylopectin fraction did not appear to be affected. The double mutants defective in SSSII and a GBSS in C. reinhardtii,120 however, had a starch content of only 2±16% of the wild type. The severity of the GBSS defect of the double mutant dictated the amount of starch present in the double mutant with an almost null mutant having very little starch. The authors suggest that GBSS is very important for synthesis of not only amylose, but also the internal structure of the amylopectin and the effect of GBSS deficiency is worsened by the diminished SSSII activity. These studies, using the Chlamydomonas mutant, provide evidence for the involvement of the GBSS not only in amylose but also in amylopectin synthesis, and suggest that a function for SSSII is in synthesis of the intermediate size B branch chains in amylopectin. Thus, a possible route for amylopectin and amylose can be proposed and is shown in Fig. 1.4. A reaction with the potential of being the initating reaction for synthesis has been observed in potato tuber160 and maize endosperm162 and a transfer of glucose from UDPGlc to serine or threonine residues has been observed. The resulting glucosylated 38 kDa protein can serve as a primer for the synthesis of starch via the starch synthase reactions. Whether there is an acceptor protein that could be glucosylated by ADPGlc has not been demonstrated; the acceptor protein and reaction has not been characterized as well as the other starch biosynthetic enzymes to show its presence in other plants besides potato tubers and maize endosperm. After formation of the unbranched maltodextrin-protein primer of undetermined size, high rates of polysaccharide formation may occur at the surface of the developing starch granule, where GBSS, SSSII and branching enzyme I interact with the glucosylated protein primer to form a branched glucan having both long and intermediate size chains. The postulation of phase 2 in Fig. 1.4 is also based on the studies of the polysaccharide structures observed in the Chlamydomonas SSSII and GBSS mutants as well as the ae mutants of rice144 and maize136 which are defective in BEII. BEII deficient mutants have altered oligosaccharides with fewer branches and longer size branched chains. In phase 3, SSSI and BE II are responsible for the synthesis of the A and exterior B-chains to complete the first cluster region in the glucan. Continued synthesis in phase 4 is essentially a repeat of phases 2 and 3 to synthesize a highly branched -glucan termed pro-amylopectin or phytoglycogen. This highly -branched glucan is water soluble and non-

ß 2004, Woodhead Publishing Limited

Fig. 1.4 Proposed synthesis of amylose and amylopectin. Phase 1. Initiation of -glucan synthesis via synthesis of a glucoprotein glucosyl acceptor for starch synthases. Phase 2. Formation of the internal cluster structure of the ultimate amylopectin product by GBSS, SSSII and BEI. Phase 3. Formation of the external cluster stucture (exterior A- and Bchains) by SSSI and BEII. Phase 4. Continual repeats of phase 2 and 3 reactions to form completion of a highly branched pro-amylopectin (phytoglycogen). Phase 5. Debranching of pro-amylopectin to form amylopectin which can now crystallize and the amylose primer, `pro-amylose'. Phase 6. Formation of amylose by elongation by GBSS.

crystalline. In phase 5, a debranching enzyme debranches the pre-amylopectin to form amylopectin which can now crystallize. In phase 6, the chains liberated by the debranching action of the pro-amylopectin (phytoglycogen) are used as primers by GBSS to form amylose. Amylose synthesis may occur only inside the starch granule and only GBSS would be involved because it may be the only starch synthase present at the site of amylose synthesis. Inside the granule, branching enzyme activity is quite restricted and, therefore, the amylose would be only slightly branched. Possibly the slight branching observed in the amylose fraction had occurred previously before the debranching phase, 5. Debranching of the pro-amylopectin may have liberated primer for GBSS that had some branch chains. The reason for postulating a water soluble pro-amylopectin is based on the existence of the sugary 1-mutation of maize endosperm which contains reduced amounts of amylopectin and starch granules, The mutant accumulates about

ß 2004, Woodhead Publishing Limited

35% of its dry weight as phytoglycogen, a highly branched water soluble polysaccharide.42 The sugary 1 mutation was shown to be deficient in debranching enzyme activity. The evidence that the sugary 1mutation affects the structural gene for a debranching enzyme is supported by the isolation of a cDNA of the su 1 gene. Its deduced amino acid sequence is most similar to a bacterial isoamylase.43 It remains to be shown whether the su 1 gene product debranching enzyme activity is actually an isoamylase, a pullulanase or an R enzyme.21 Moreover, the specificity of the reaction needs to be studied with respect to the factors that determine which -1,6 linkages are cleaved and which remain resistant to debranching action. It is quite possible that the crowding of the -1,6 linkages in the cluster region causes some steric difficulties for the debranching of the linkages in the cluster region, but at present this is only a hypothesis. The involvement of a debranching enzyme for formation of amylopectin in Chlamydomonas reinhardtii has also been demonstarted.45 A mutant, sta 7, leads to a complete loss of starch (amylose + amylopectin) content with formation of a water soluble -1,4-glucan reminiscent of a phytoglycogen structure. The phytoglycogen content is about 8% of the starch content seen in the normal algal strain. The mutant seems to be deficient in an 88 kDa protein and debranching activity.45 The authors also postulate a phase (phase 5 in Fig. 1.3) in starch synthesis requiring debranching activity as well as in a review.46 The Chlamydomonas reinhardtii mutant, sta 7, is interesting in that it suggests that both amylose and amylopectin synthesis require debranching activity and that the debranched fragments of proamylopectin are the primers for amylose synthesis. These reactions do not have to occur in perfect sequence and the phases may have some overlap, e.g., phases 2, 3, 4 may overlap, and possibly even 5 and 6. However, the present evidence, such as intermediate products formed by starch mutants of Chlamydomonas and of higher plants, support the sequence of reactions shown in Fig. 1.4 for amylopectin and amylose biosynthesis. Further experiments are certainly required to test the proposed scheme in Fig. 1.4.

1.9

Regulating starch synthesis in plants

In bacteria and in plants the only physiological function for ADP-glucose is as a donor of glucose for -1,4-glucosyl linkages. It would be advantageous to conserve the ATP utilized for synthesis of the sugar nucleotide for the procaryote and plant cell to regulate glucan synthesis at the level of ADPglucose formation. Over 50 ADPGlc Ppases have been studied with respect to regulatory properties. In almost all cases, glycolytic intermediates activate ADPGlc synthesis while AMP, ADP and/or Pi are inhibitors. Glycolytic intermediates in the cell can be considered as indicators of carbon excess and therefore, under conditions of limited growth with excess carbon in the media, accumulation of glycolytic intermediates would be signals for the activation of

ß 2004, Woodhead Publishing Limited

ADPGlc synthesis. Thus, the enzyme seems to be modulated by the availability of ATP in the cell and the presence of glycolytic intermediates. A group of procaryotes which are capable of oxygenic photosynthesis are the cyanobacteria having ADPGlc Ppases activated by 3-phosphoglycerate (3PGA), the primary CO2 fixation product of photosynthesis.78, 90, 91 3PGA is also the primary activator of the green algae196,197 and higher plant ADPGlc Ppases.3, 4, 9, 198 Thus the initial product of photosynthesis serves as an allosteric activator for the synthesis of a reserve product, glycogen or starch. The specificity of the activation is the same whether the enzyme is from a plant incorporating CO2 via the C3 pathway or the C4 pathway. The ADP-glucose pyrophosphorylases of Chlorella pyrenoi dosa, Chlorella vulgaris, Scenedesmus obliquus, and Chlamydomonas reinhardtii and of the cyanobacteria are also activated by 3-phosphoglycerate to a lesser extent than by other glycolytic intermediates and inhibited by Pi . 3-Phosphoglycerate (3PGA) also increases the apparent affinity of all the substrates for the spinach leaf enzyme from 2- to 13-fold. ADP-glucose synthesis catalyzed by the spinach leaf enzyme is inhibited 50% by 22 M Pi in the absence of activator at pH 7.5.49 In the presence of 1 mM 3-phosphoglycerate, 50% inhibition required 1.3 mM phosphate. Thus, the activator decreased sensitivity to Pi inhibition about 450fold. However, Pi at 0.5 mM increased the concentration of 3-phosphoglycerate needed for activation. The peptide domains involved in the inhibitor binding sites remain to be identified.

1.9.1 Effect of activators on inhibition and substrate kinetics Similar interactions can be seen in Table 1.10 with the potato tuber ADPGlc Ppase which also show that large effects can be seen on the rate of ADP-glucose synthesis with relatively small changes in the 3PGA, the activator, and Pi concentrations, particularly at low concentrations of 3PGA where the activation is minimal in the presence of Pi. At 1.2 mM Pi and 0.2 mM 3PGA, ADPGlc synthesis is inhibited over 95%. However, if the Pi concentration decreases 33% to 0.8 mM and the 3PGA concentration increases 50% to 0.3mM there is an 8.5fold increase in the rate of ADPGlc synthesis. Conversely, at 0.4mM 3PGA and 0.8 mM Pi, the rate of ADPGlc synthesis is 7.5 nmol per 10 min. This is reduced to 2.2 nmol per 10 min. (70% decrease) if only the 3PGA concentration decreases 50% to 0.2mM. If the Pi concentration increases as well to 1.2mM (a 50% increase) then the synthetic rate is reduced to 0.65 nmol which is a reduction of ADPGlc synthesis of 91%. The reason for the small changes in the effector concentrations giving such large effects in the synthetic rate is due to the sigmoidal nature of the curves particularly at the low concentrations of 3PGA. Evidence will be presented below that strongly indicates that the ratio of activator/inhibitor modulates the activity of ADP-glucose pyrophosphorylase in vivo in algae and in plants, thereby regulating the synthesis of starch in these systems.

ß 2004, Woodhead Publishing Limited

Table 1.10 The interaction of phosphate and 3-P-glycerate concentrations on potato tuber ADPGlc PPase activity [Pi], mM

0.4

0.8

[3PGA], mM

ADP-glucose formed, nmol/10 min.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 1.0

0.16 2.1 7.2 9.9 11.3 11.8 12.8 14.2 16.9

0.18 0.49 2.2 5.5 7.5 8.5 10.3 11.0 13.0

1.2

0.08 0.16 0.65 2.7 3.8 6.6 7.8 9.0 10.6

1.6

0.02 0.15 0.25 0.64 2.2 3.8 5.2 5.8 8.3

1.9.2 In vivo evidence in support of the ADP-glucose pyrophosphorylase activator-inhibitor interaction regulating starch synthesis The leaf ADP-glucose pyrophosphorylase is highly sensitive to 3phosphoglycerate, the primary CO2 fixation product of photosynthesis, and to inorganic phosphate. Thus, it has been suggested that these compounds play a significant role in vivo in regulating starch biosynthesis in algae, higher plants and in the regulation of glycogen synthesis in cyanobacteria.3±8, 198, 199 For example, inorganic phosphate levels in leaves may decrease during photosynthesis because of photophosphorylation, and glycolytic intermediates increase in the chloroplast in the light. This situation would thus contribute to conditions necessary for optimal starch or glycogen synthesis via the increased rate of formation of ADP-glucose. In the light, the levels of ATP and reduced pyridine nucleotides are also increased, leading to the formation of sugar phosphates from 3-phosphoglycerate. In the dark, there is an increase in phosphate concentration with concomitant decreases in the levels of 3phosphoglycerate, ATP, and reduced pyridine nucleotides. This would lead to inhibition of ADP-glucose synthesis and therefore of starch or glycogen synthesis. Data correlating altered accumulation of starch or altered rates of starch synthesis due to changes of either cellular Pi or 3PGA levels have been discussed in previous reviews3, 4, 24 and suggest that in vivo, 3PGA and Pi levels affect starch synthetic rates via modulation of ADPGlc Ppase activity. Recent evidence has also been obtained strongly suggesting that the regulatory effects seen for the plant and algal ADPGlc Ppase are important in vivo, for starch synthesis. Kacser-Burns control analysis methods200, 201 vary enzyme activity, either by using mutants deficient in that enzyme or by varying the physiological conditions and correlating the effect of these changes on the rate of a metabolic process (e.g., starch synthesis). If the enzyme activity is rate limiting or

ß 2004, Woodhead Publishing Limited

important in controlling the metabolic process, then a large effect on that process should be seen. Conversely, if there is no or little effect, then the enzyme level of activity is not considered to be rate-limiting for the metabolic process being measured. The effect is measured as a flux coefficient ratio. If the ratio correlates well with the variation of enzyme activity with the change in rate of the process measured, a correlation ratio close to one should be observed. Analysis on photosynthate partitioning was done on starch synthesis in Arabidopsis thaliana.202, 203 The results showed that leaf ADPGlc Ppase is a major site of regulation for starch synthesis203 and that regulation of the enzyme by 3PGA is an important determinant of the rate of starch synthesis in vivo.202 Arabidopsis thaliana mutant strains containing only 7% of the normal activity of ADPGlc Ppase and a hybrid strain between the mutant and normal strain having 50%, had 90% and 39% reduction in the starch synthetic rate respectively, at a high level of light, as compared to the wild type.203 Thus, there is fairly good correlation between the activity of the ADPGlc Ppase and the rate of starch synthesis. The flux control coefficient was determined to be 0.64. A possible underestimation of the Kacser-Burns coefficient control Despite the fairly high value seen for the ADPGlc Ppase flux control coefficient it is quite possible that it may be underestimated due to the allosteric properties of the enzyme. In flux control analysis the maximal enzyme activity is measured. In the case of an allosteric enzyme the potential maximal enzyme activity may not be as critical as the allosteric effector concentrations that establish the enzyme activity. Therefore, with an allosteric enzyme a valid flux control coefficient just based on potential maximum activity may not be observed. With ADPGlc Ppases, activation by 3PGA can be anywhere from 10to 100-fold. Moreover, inhibition by the allosteric inhibitor, Pi and variations in the [3PGA]/[Pi] ratio could cause greater fluctuations in the potential maximal activity. Thus, a flux coefficient control value analysis based only on the enzyme's maximal activities of the Arabidopsis thaliana mutants and normal ADPGlc Ppases may underestimate the regulatory potential of the ADPGlc Ppase reaction. That is why the claim that the pea embryo ADPGlc Ppase having a low control coefficient ratio (0.1) may be invalid204 since only the maximal activity was considered and not the actual activity in the embryo determined by the 3PGA and Pi concentrations. In the experiments utilizing a Clarkia xantiana mutant, deficient in leaf cytosolic phosphoglucoseisomerase (having only 18% of the activity seen in the wild type) lower sucrose synthetic rates and the increased starch synthesis rates were observed.202 The chloroplast 3PGA concentration increased about twofold, suggesting that the increase of starch synthetic rate measured in the mutant deficient in cytosolic phosphoglucoseisomerase was due to activation of the ADPGlc Ppase by the increased 3PGA concentration and the 3PGA/Pi ratio. Important evidence indicating that the in vitro activation of the ADPGlc Ppase is truly functional in vivo comes from isolation of a class of mutants

ß 2004, Woodhead Publishing Limited

where the mutation directly affects the allosteric properties of the ADPGlc Ppase. Such mutants were found easily for the bacteria E. coli and Salmonella typhimurium.22, 23, 58 Similar mutants have been found for Chlamydomonas rheinhardtii and for maize endosperm. A significant finding was made by Ball et al.205 who isolated a starch deficient mutant of C. reinhardtii in which the defect was shown to be in the ADPGlc Ppase, which could not be effectively activated by 3PGA. The inhibition by Pi was similar to the wild type (Iglesias, Koornhysue Ball and Preiss, unpublished results). The starch deficiency was observed in the mutant whether the organism was grown photautotrophically with CO2 or in the dark with acetate as the carbon source. Thus, the allosteric mechanism seems to be operative for photosynthetic or non-photosynthetic starch biosynthesis. Another putative ADPGlc Ppase allosteric mutant from maize endosperm, which has 15% more dry weight (in addition to starch) than the normal endosperm, has been isolated and described by Giroux et al.206 The mutant allosteric ADPGlc Ppase was less sensitive to Pi inhibition than the normal enzyme. The Chlamydomonas starch deficient mutant and higher dry weight maize endosperm mutant studies strongly suggest that the in vitro regulatory effects observed with the photosynthetic and non-photosynthetic plant ADPGlc Ppases are highly functional in vivo and that ADPGlc synthesis is rate-limiting for starch synthesis. Thus, data continue to accumulate showing the importance of the plant ADPGlc Ppase in regulating starch synthesis and that the allosteric effectors, 3PGA and Pi, are important in vivo, in photosynthetic as well as in non-photosynthetic starch synthesis. Are starch synthases or branching enzyme rate-limiting in starch synthesis? Can starch synthase and branching enzymes be rate-limiting under certain physiological conditions? The wrinkled pea has a reduced starch content; about 66±75% of that seen in the round seed. The amylose content is about 33% in the round normal form it is 60±70% in the wrinkled pea seed. As indicated before, Smith142 who also showed that the r (rugosus) lesion (as found in the wrinkled pea of genotype rr) was associated with the absence of one isoform of branching enzyme concentrations than in round or normal pea. Smith et al.207 showed that in mutant rr leaves, in high light intensity there was a 40% decrease in the rate of starch synthesis. Control coefficient analysis207 showed essentially no effect on the rate of starch synthesis in low light intensity while in high light intensity the flux control coefficient a small value (meaning very little control) of 0.13 was obtained. An 86% reduction of branching enzyme activity had a small effect on regulation of starch synthesis. It should be noted that the BE control coefficient was only 20% of the value found for the ADPglucose pyrophosphorylase which was 0.64.203 At temperatures above 30 ëC both maize208 and wheat endosperm209, 212 had a reduction of starch deposition compared to lower temperatures. In wheat the starch biosynthetic enzyme affected was soluble starch synthase.209±212 Using flux control coefficient analysis, Keeling et al.211 showed a control coefficient

ß 2004, Woodhead Publishing Limited

of 1.15 between the rate of starch synthesis and the level of starch synthase activity in the wheat endosperm extracts. In vitro, the endosperm starch synthase activity was sensitive to heat treatment in the range of 30±40 ëC if the treatment was for longer than 15 min. A similar study with maize endosperm showed a reduction of starch synthetic rate and a decrease in starch synthase activity in the heat-stressed maize endosperm.208 However, in the heat-stressed maize, the endosperm ADPGlc Ppase activity was also reduced and even to a greater extent than the soluble starch synthase.208 Thus, in wheat and maize there may be a relationship, under some environmental conditions, between reduction of starch synthase activity and decreased starch synthesis. However, as the maize data suggests,208 other factors besides starch synthase activity, not yet studied, may be the primary reason for the reduction of starch synthesis in the heat-stressed plants. In the case of maize endosperm another enzyme involved in starch synthesis, ADPGlc Ppase, is also affected in the heat-stressed plant. Cheikh and Jones213 showed that in maize soluble invertase activity was inhibited by short- and long-term heat stress. It is also quite possible that other critical steps leading to starch biosynthesis are also affected in both plants, such as carbon flow (sucrose transport?) from source to sink tissues. Those processes were not studied in the heat-stressed plants. Thus, it is premature on the basis of the above experiments to designate starch synthase as a major control point and other data should be obtained. Flux control coefficients for an enzyme can only be determined for any process if only that enzyme's activity is affected. In the case of the heat-stressed plants it has been shown that other activities beside starch synthase activity are affected. Could starch synthetic rate can be increased by over-expressing soluble starch synthase activity in the amyloplast? As shown below starch accumulation can be increased by expressing a bacterial ADPGlc Ppase allosteric mutant in plants.193 Increasing starch content by transforming plants with an allosteric mutant ADPGlc Ppase gene A preponderance of evidence strongly suggests that a rate-limiting and regulatory enzyme of starch biosynthesis in algae or higher plants is ADPGlc Ppase. Also, control analysis experiments have shown that ADPGlc Ppase is rate-limiting and important for regulation of leaf starch synthesis.203 Reduced ADPGlc Ppase activity in mutants also lead to a reduction in the rate of starch synthesis in Arabidopsis leaves84, 88 and in potato tubers.214 It was of interest to see if starch content in a plant could be augmented by increased expression of activity of one of the enzymes involved in starch biosynthesis. Overexpression of a plant ADPGlc Ppase activity however, would require expression of two distinct genes to reconstitute its ADPGlc Ppase activity. It is also possible the plant would compensate for the overexpression by altering the ratio of the effector metabolites, 3PGA and Pi, to a value that starch synthesis would not be elevated. A different strategy was therefore chosen. An E. coli ADPGlc Ppase, glg C gene of allosteric mutant 618, labelled as glg C16,215 which encodes for an enzyme independent of the presence of an

ß 2004, Woodhead Publishing Limited

activator for activity, was used for the transformation. Expression of this bacterial mutant gene would have two advantages. First, only one gene needs be expressed for ADPGlc Ppase activity and, second, the mutant enzyme would be less sensitive to inhibition by its allosteric inhibitor, 5'AMP, insensitive to the plant enzyme's inhibitor, Pi and independent of activator for good activity.215 Since starch synthesis occurs in the plastid, a nucleotide sequence encoding transit peptide of the Arabidopsis ribulose 1,5 bisphosphate carboxylase chloroplast transit peptide was fused to the translation initiation site of the glg C16 gene (Fig. 1.5). The chimeric gene was then cloned behind either a tuberspecific patatin promoter or a cauliflower mosaic virus (CaMV) enhanced 35S promoter, or, in the case of tomato plants, the Arabidopsis plant promoter from the rbcS gene (193, Fig. 1.5). The chimeric gene containing promoter was placed in a cloning vector with a 35S-neomycin phosphotransferase gene as a selectable marker193 and used for transformation of various plants193 Starch levels were increased over the controls lacking the glg C gene product by about 1.7±8.7-fold in tobacco calli where the glg C gene product activity was detected.193 When the CaMV-chimeric gene was electroporated into tobacco protoplasts extracts of the transformed protoplasts gave rise to ADPGlc synthesis resistant to Pi inhibition and activated by fructose 1,6 bis -P. This is indicative of the E. coli ADPGlc Ppase activity. Almost all plant ADPGlc Ppases are highly sensitive to inhibition by Pi and ADPGlc synthesis in the control protoplast extract was totally inhibited by Pi as expected. Comparison of transgenic tobacco with control calli, show a very large increase in the number of starch granules.193 Similarly, tomato shoots excised from the transformed calli containing the transit peptide-glg C16 gene stained black with I2 while the controls were essentially negative.193 Similar results have been obtained for transformed Russet-Burbank potato tubers where the chimeric gene with transit peptide under control of a tuberspecific patatin promoter, increased starch in the tuber 25±60 % over controls not containing the bacterial enzyme (193, Table 1.11). Bacterial ADPGlc Ppase glg C16 gene expressed in the tuber lacking the transit peptide gene portion, gave no increase in starch content (Table 1.11). ADPGlc Ppase activity was expressed but presumably was not present in the amyloplast and was not able to supply ADPGlc to the amyloplast localized starch synthases. When wild-type E. coli glg C gene was used for transformation, no increase in starch was observed.193 Thus, the allosteric properties of the ADPGlc Ppase in the normal

Fig. 1.5 Construction of the promoter-plastid transit peptide glg C16 ADPGlc Ppase gene vector. The transit peptide is from a modified Arabidopsis thaliana chloroplast transit peptide which contained 23 amino acids of the mature N-terminal of the Rubisco small subunit and an extra protease cleavage site (193). The Nos terminator is the nopaline synthase 3' poly A signal.

ß 2004, Woodhead Publishing Limited

Table 1.11 Starch content in potato tubers transformed with the glg C16 and glg C genes as shown by reference (193) Average starch content, % wet wt. A. control; untransformed chlorplast transit peptide-glg C16 glg C16, no transit peptide B. control; untransformed chloroplast transit peptide-glg C

12.3Ô1.15 16.0Ô2.00 12.4Ô0.24 13.2Ô0.12 13.1Ô0.07

situation are important for regulation and that alteration of the allosteric properties was needed to increase starch levels. A relationship between the expression levels of the ADPGlc Ppase of glg C16 as measured by Western blotting of the potato extracts and the increase in starch content was observed, particularly in tubers at a lower range of starch content.193 With lower levels of expressed ADPGlc Ppase increases of 21±63 % increase in starch were seen. Intermediate levels of the expressed ADPGlc Ppase gave 33± 118% increases in starch and the high expressed levels of the transit peptide-glg C16 gave increases of 33±167%. Thus, increasing starch synthesis is possible by transforming the tuber with an ADPGlc Ppase having allosteric properties minimized to permit higher rates of ADPGlc synthesis under physiological conditions. These results also strongly suggest that the ADPGlc Ppase is a rate-limiting enzyme for starch synthesis even in non-photosynthetic plant tissues. Further studies are currently being carried out to determine the relationship between increased enzyme activity (due to the transformation of the bacterial gene) and rate of starch synthesis by feeding labelled glucose or sucrose to potato slices. The possibility of other genes involved in starch metabolism being indirectly affected by the glg C16 transformation is also being investigated and thus far little change has been seen with starch synthase activity (H. Davies, unpublished results). In independent studies216, 217 similar studies were done with a different variety of potato and there was a very positive correlation with expressed glg C16 ADPGlc Ppase that increased increased activity 2±4-fold in potato tuber and increased flux of glucose incorporation in potato tubers. A 20±50% increase in starch synthase activity was seen. The authors concluded that increased activity of ADPGlc Ppase could indeed increase the flow of carbon into starch about 2±7 fold.217 However, no increase in starch was noted in the transgenic tubers and an increase in amylase activity was noted in the transgenic potatoes.217 It was concluded that the expected increase in starch content in the transgenic potatoes was offset by the increase in amylase activity. These results are in contrast to the results of Stark et al.193 Many different potato varieties have been transformed with the glg C16 ADPGlc Ppase and the increase in starch has been noted (unpublished experiments, Monsanto Co.). Moreover, no increase in starch degradative activity has been observed (H.

ß 2004, Woodhead Publishing Limited

Davies, unpublished experiments) in the glg C16 ADPGlc Ppase transgenic potatoes. Also recent success has been achieved in producing high starch in rice and wheat endosperm, and tomato fruit (193; unpublished experiments) by transforming the plants with the glg C16 ADPGlc Ppase. It is conceivable therefore, that similar methods can be used to change in addition to quantity, starch quality via expression of the isoforms of starch synthase and branching enzymes in plants. These `new starches' may have greater usefulness in food and industrial processes. The production of modified `speciality' starches via molecular biology techniques is promising and certainly there is an increased demand for starch in many non-food and specialized uses.

1.10 References 1.

2. 3. 4. 5. 6. 7. 8.

9.

10. 11. 12. 13.

14. 15. 16. 17.

and GARWOOD, D. L. (1984) in Starch Chemistry and Technology, 2nd edn, (Whistler, R. L., BeMiller, J. N. and Paschall, E. F., eds) Academic Press, Inc. San Diego, pp. 25±86. MARTIN, C. and SMITH A. M. (1995) Plant Cell 7, 971±985. PREISS, J. (1988) in The Biochemistry of Plants: Vol. l4, Carbohydrates, Structure and Function (Preiss, J., ed.) Academic Press, New York, pp. 184±249. PREISS, J. (1991) in Oxford Survey of Plant Molecular and Cellular Biology Vol. 7 (Miflin, B. J., ed.) Oxford University Press, Oxford, UK, pp. 59±114. PREISS, J. (1992) in Frontiers in Carbohydrate Research (Chandrasekaran, R., ed.) Elsevier Science Publishers, Ltd, Essex, UK, pp. 208±230. PREISS, J. (1996) in Biotechnology Annual Review Vol. 2. (El-Gewely, R., ed.) Elsevier Science, Amsterdam, The Netherlands, pp. 259±279. PREISS, J. (1996) Carbohydrates in Europe 15, 11±14. PREISS, J. (1997) in Engineering improved carbon and nitrogen resource use efficiency in higher plants (C. Foyer and P. Quick, eds) Taylor and Francis, Publishers, London and Washington, DC In Press, pp. 81±104. PREISS, J. and SIVAK, M. N. (1996) in Photoassimilate distribution in plants and crops: Source-sink relationships (Zamski, E. and Schaffer, A. A., eds) Marcel Dekker Inc. New York, pp. 63±96. SIVAK, M. N. and PREISS, J. (1995) in Seed development and germination (Kigel, J. and Galili, G., eds) Marcel Dekker, New York, pp. 139±168. SMITH, A. M., DENYER, K. and MARTIN, C. (1995) Plant Physiol. 107, 673±677. SACHS, J. (1887) in Lectures of the physiology of plants (translated by H. M. Ward) Clarendon Press, Oxford, pp. 304±325. EDWARDS, G. and WALKER, D. A. (1983) in C3,C4: Mechanisms of cellular and Environmental regulation, photosynthesis, University of California Press, Berkeley, Calif, pp. 204±207. BADENHUIZEN, I. P. (1969) The biogenesis of starch granules in higher plants, Appleton-Century Crofts, New York. CASPAR, C., HUBER, S. C. and SOMERVILLE, C. (1986) Plant Physiol. 79, 1±7. MATHESON, N. K. (1996) Carbohyd. Res. 282, 247±262. ROBIN, J. P., MERCIER, C., CHARBONNIERE, R. and GUIBOT, A. (1974) Cereal Chemistry 51, 389±406. SHANNON, J. C.

ß 2004, Woodhead Publishing Limited

18. 19. 20. 21. 22. 23.

24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

and MATHESON, N. K. (1981) Carbohyd. Res. 90, 99±110. (1986) `Polymodal distribution of the chain lengths of amylopectins, and its significance', Carbohydr. Res. 147, 342±347. MORRISON, W. R. and KARKALAS, J. (1990) in Methods in Plant Biochemistry (Dey, P. M., ed.) Academic Press, Ltd. London, pp. 323±352. HIZUKURI, S. (1995) in Carbohydrates in Food (Eliasson, A-C., ed.) Marcel Dekker, Inc. New York, pp. 347±429. PREISS, J. and ROMEO, T. (l989) Advances in Bacterial Physiology, Vol. 30, 184± 238. PREISS, J. (1996) in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, 2nd edn, (F. C. Neidhardt, editor in Chief) Amer. Soc. Microbiol., Washington, D. C. Vol. 1, pp. 1015±1024. PREISS, J. and LEVI, C. (1980) in The Biochemistry of Plants: Vol. 3, Carbohydrates, Structure and Function (Preiss, J., ed.) Academic Press, New York, pp. 371±423. DENYER, K. and SMITH, A. M. (1992) Planta 186, 609±617. DENYER, K. SIDEBOTTOM, HYLTON, C. M. and SMITH, A. M. (1993) Plant J. 4, 191 198. DENYER, K. HYLTON, C. M., JENNER, AND SMITH A. M. (1995) Planta 196, 256±265. HYLTON, C. M., DENYER, K., KEELING, P. L., CHANG, M.-T. and SMITH, A. (1996) Planta, 198, 230±237. MIZUNO, K., KIMURA, K., ARAI, Y., KAWASAKI, T., SHIMADA, H. and BABA, T. (1992) J. Biochem. 112, 643±651. BHATTACHARYYA, M., MARTIN, C. and SMITH, A. M. (1993) Plant Molec. Biol. 22, 525±531. GUAN, H. P. and PREISS, J. (1993) Plant Physiol. 102, 1269±1273. GUAN, H. P., BABA, T. and PREISS, J. (1994) Plant Physiol. 104, 1449±1453. GUAN, H. P., BABA, T. and PREISS, J. (1994) Cell. Molec. Biol. 40, 981±988. MANNERS, D. J.

HIZUKURI, S.

BURTON, R. A., BEWLEY, J. D., SMITH, A. M., BHATTACHARYYA, M., TATGE, H., RING, S.,

and MARTIN, C. (1995) Plant J. 7, 3±15. and CHANG, M. T. (1978) Plant Physiol. 62, 383 386. SHURE,M., WESSLER, S. and FEDEROFF, N. (1983) Cell 35, 225±233. SANO, Y. (1984) Theor. Appl. Genetics 68, 467±473. VAN DER LEIJ, F. R., VISSER, R. F. G., PONSTEIN, A. S., JACOBSEN, E. and FEENSTRA, W. J. (1991) Mol. Gen. Genetics 228, 240±248. 39. VISSER, R. F. G., SOMHORST, I., KUIPERS, G. J., RUYS, N. J., FEENSTRA, W. J. and JACOBSEN, E. (1991) Molec. Gen. Genetics 225, 289±296. 40. DRY, I., SMITH, A., EDWARDS, A. BHATTACHARYA, M., DUNN, P. and MARTIN, C. (1992) Plant J. 2, 193±202. 41. DELRUE, B., FONTAINE, T., ROUTIER, F., DECQ, A., WIERUSZESKI, J.-M., VAN DEN KOORNHUYSE, N., MADDELEIN, M.-L., FOURNET, B. and BALL, S. (1992) J. Bacteriol. 174, 3612±3620. 42. PAN, D. and NELSON, O. E. (1984) Plant Physiol. 74, 324±328. 43. JAMES, M. G., ROBERTSON, D. S. and MEYERS, A. M. (1995) Plant Cell 7, 417±429. 44. NAKAMURA, Y., UMEMOTO, T., TAKAHATA, Y. KOMAE, K., AMANO, E. and SATOH, H. (1996) Physiol. Plant. 97, 491±498. 45. MOUILLE, G., MADDELEIN, M.-L., LIBESSART, N., TALAGA, P. DECQ, A., DELRUE, B. and BALL, S. (1996) Plant Cell 8, 1353±1366. 46. BALL, S., GUAN, H.-P., JAMES, M., MYERS, A., KEELING, P., MOUILLE, G., BULEÂON, A., COLONNA, P. and PREISS, J. (1996) Cell 86, 349±352. 47. LELOIR, L. F., DEFEKETE, M. A. R. and CARDINI, C. E. (1961) J. Biol. Chem. 236, 636± 641. BULL, V., HAMILTON, D. O.

35. 36. 37. 38.

NELSON O. E., CHOUREY, P. S.

ß 2004, Woodhead Publishing Limited

48. 49. 50. 51. 52. 53. 54. 55.

56.

and LELOIR, L. F. (1961) Biochem. Biophys. Res. Commun. 6, 85, 88. and PREISS, J. (1966) J. Biol. Chem. 241, 4491±4505. PREISS, J., GHOSH, H. P. and WITTKOP, J. (1967) in Biochemistry of Chloroplasts (Goodwin, T. W., ed.), Vol. 2, pp. 131±153. Academic Press, New York. COPELAND L. and PREISS, J. (1981) Plant Physiol. 68, 996±1002. SANWAL, G., GREENBERG, E., HARDIE, J., CAMERON, E. and PREISS, J. (1968) Plant Physiol. 43, 417±427. RIBEREAU-GAYON, G. and PREISS, J. (1971) Methods Enzymol. 23, 618±629. MORELL, M. K., BLOOM, M., KNOWLES, V. and PREISS, J. (1987) Plant Physiol. 85, 185± 187. MORELL, M. BLOOM, M. LARSEN, R., OKITA, T. W. and PREISS, J. (1987) in Plant gene systems and their biology. (Key, J. L. and McIntosh, L. eds), Alan R. Liss, Inc. New York, pp. 227±242. RECONDO, E. GHOSH, H. P.

PREISS, J., BLOOM, M., MORELL, M. KNOWLES, V., PLAXTON, W. C., OKITA, T. W. LARSEN,

and PUTNAM-EVANS, C. (1987) in Tailoring genes for crop improvement. An agricultural perspective. (Bruening, G., Harada, J., Kosuge, T. and Hollaender, eds), Plenum Press, New York, pp. 133±152. PREISS, J., MORELL, M., BLOOM, B., KNOWLES, V. L. and LIN, T. P. (1987) in Progress in photosynthesis research, Vol. III (Biggins, J. ed.) Nijhoff, The Hague, pp. 693± 700. PREISS, J. (1984) Annual Review of Microbiology 38, 419±458. PREISS, J., DANNER, S., SUMMERS, P. S., MORELL, M., BARTON, C. R., YANG, L. and NIEDER, M. (1990) Plant Physiol. 92, 881±885. BARTON, C., YANG, L., GALVIN, M., SENGUPTA-GOPALAN, C. and BORELLI, T. (1986) in Regulation of Carbon and Nitrogen Reduction and Utilization in Maize. (Shannon, J. C., Knievel, D. P. and Boyer, C. D. eds), American Society of Plant Physiologists, Rockville MD, pp. 363±365. ANDERSON, J. M., HNILO, J., LARSON, R., OKITA, T. W., MORELL, M. and PREISS, J. (1989) J. Biol. Chem. 264, 12238±12242. HYLTON, C. and SMITH, A. M. (1992) Plant Physiol. 99, 1626±1634. SMITH-WHITE, B. and PREISS, J. (1992) J. Molec. Evolution 34, 449±464. PREISS, J., CRESS, D., HUTNY, J., MORELL, M., BLOOM, M., OKITA, T. and ANDERSON, J. (1989) in Biocatalysis in agricultural biotechnology. (Whitaker, J. R. and Sonnet, P. E., eds) Amer. Chem. Soc. Washington, D. C, pp. 84±92. HILL, M. A., KAUFMANN, K. OTERO, J. and PREISS, J. (1991) J. Biol. Chem. 266, 12455± 12460. LEE, Y. M. and PREISS, J. (1986) J. Biol. Chem. 261, l058±l064. CHARNG, Y.-Y., KAKEFUDA, G., IGLESISAS, A. A., BUIKEMA, W. J. and PREISS, J. (1992) Plant Mol. Biol. 20, 37±47. KAKEFUDA, G., CHARNG, Y.-Y., IGLESIAS, A. A., MCINTOSH, L. and PREISS, J. (1992) Plant Physiology 99, 359±361. AINSWORTH, C., TRAVIS, M. and CLARK, J. (1993) Plant Molec. Biol. 23, 23±33. BALLICORA, M. A., LAUGHLIN, M. J., FU, Y., OKITA, T. W., BARRY, G. F. and PREISS, J. (1995) Plant Physiol. 109, 245±251. MORELL, M., BLOOM, M. and PREISS, J. (1988) J. Biol. Chemi. 263, 633±637. BALL, K. L. and PREISS, J. (1994) J. Biol. Chem. 269, 24706±24711. CHARNG, Y.-Y., IGLESIAS, A. A. and PREISS, J. (1994) J. Biol. Chem. 269, 24107± 24113. SHENG, J., CHARNG, Y.-Y. and PREISS, J. (1996) Biochem. 35, 3115±3121. R., HARMON, A. C.

57.

58. 59. 60.

61. 62. 63. 64.

65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

ß 2004, Woodhead Publishing Limited

75.

76. 77. 78. 79. 80. 81. 82. 83.

and BALLICORA, M. A. (1996) in Proc. Xth International Photosynthesis Congress (P. Mathis, ed.) Kluwer Academic Publishers. Vol. 5 pp. 47±52. THORNBJéRNSEN, T., VILLAND, P., DENYER, K., OLSEN, O.-A. and SMITH, A. (1996) Plant J. 10, 243±250. BAE, J. M., GIROUX, J. and HANNAH, L. C. (1990) Maydica 35, 317±322. CHARNG, Y.-Y., SHENG, J. and PREISS, J. (1995) Arch. Biochem. Biophys. 318, 476 480. VILLAND, P., OLSEN, O.-A. KILIAN, A. and KLECKOWSKI, L. A. (1992) Plant Physiol. 100, 1617±1618. BHAVE, M. R., LAWRENCE, S., BARTON, C. and HANNAH, L. C. (1990) Plant Cell, 2, 581± 588. OLIVE, M. R., ELLIS, R. J. and SCHUCH, W. W. (1989) Plant Molec. Biol. 12, 525, 538. BALL, K. L. and PREISS, J. (1992) J. Prot. Chem. 11, 231±238.

PREISS, J., SHENG, J. FU, Y.

IGLESIAS, A. A., BARRY, G. F., MEYER, C., BLOKSBERG, L., NAKATA, P. A., GREENE, T., LAUGHLIN, M. J., OKITA, T. W., KISHORE, G. M.

and PREISS, J. (1993) J. Biol. Chem. 268,

1081±1086. 84.

and PREISS, J. (1988). Plant Physiol. 88, 1175± 1181. 85. LI, L. and PREISS, J. (1992) Carbohyd. Res. 227, 227±239. 86. PLAXTON, W. C. and PREISS, J. (1987) Plant Physiol. 83, 105±112. 87. KRISHNAN, H. B., REEVES, C. D. and OKITA, T. W. (1986) Plant Physiol. 81, 642±645. 88. LIN, T. P., CASPAR, T., SOMERVILLE, C. and PREISS, J. (1988). Plant Physiol. 86, 1131± 1135. 89. OKITA, T. W., NAKATA, P. A., ANDERSON, J. M., SOWOKINOS, J., MORELL, M. and PREISS, J. (1990) Plant Physiol. 93, 785±790. 90. LEVI, C. and PREISS, J. (1976) Plant Physiol. 58, 753±756. 91. IGLESIAS, A. A., KAKEFUDA, G. and PREISS, J. (1991) Plant Physiol. 97, 1187±1195. 92. LEMESLE-VARLOOT, L., HENRISSAT, B., GABORIAUD, C., BISSERY, V., MORGAT, A. and MORNON, J. P. (1990) Biochimie 72, 555±574. 93. ROST, B. and SANDER,R. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 7558±7562. 94. ROSSMAN, M. G., MORAS, D. and OLSEN, K. W. (1974) Nature 250, 194±199. 95. TONG, L. A., DE VOS, A. M., MILBURN, M. V. and KIM, S. H. (1991) J. Mol. Biol. 217, 503± 516. 96. MACDONALD, F. D. and PREISS, J. (1983) Plant Physiology 73, 175±178. 97. MACDONALD, F. D. and PREISS, J. (1985) Plant Physiology 78, 849±852. 98. OZBUN, J. L., HAWKER, J. S. and PREISS, J. (1971) Biochem. Biophys. Res. Comun. 43, 631±636. 99. EDWARDS A., MARSHALL, J., DENYER, K., SIDEBOTTOM, C., VISSER,R. G., MARTIN, C. and SMITH, A. M. (1996) Plant Physiol. 112, 89±97. 100. MU-FORSTER, C., HUANG, R., POWERS, J. R., HARRIMAN, KNIGHT, M., SINGLETARY, G. W., KEELING, P. and WASSERMAN, B. P. (1996) Plant Physiol. 111, 821±829. 101. NELSON, O. E. and RINES, H. W. (1962) Biochem. Biophys. Res. Commun. 9, 297±300. 102. FEDOROFF, N., WESSLER, S. and SHURE, M. (1983) Cell 35, 235±242. 103. KLOÈSGEN, R. B., GIERL, A., SCHWARTZ-SOMMER, Z. and SAEDLER, H. (1986) Molec. Gen. Genet. 203, 237±244. 104. RHODE, W., BECKER, D. and SALAMINI, F. (1988) Nucleic Acids Res. 16, 7185±7186. 105. WANG, Z., WU, Z., XING, Y., ZHENG, F., GUO, X., ZANG, W. and HONG, M. (1990) Nucleic Acids Res. 18, 5898. LIN, T. P., CASPAR, T., SOMERVILLE, C.

ß 2004, Woodhead Publishing Limited

106.

SALEHUZZAMAN, S. N. I. M., JACOBSEN, E.

and VISSER, R. G. F. (1993) Plant Molec. Biol.

23, 947±962. 107. 108. 109.

AINSWORTH, C., CLARK, J. KUMAR, A., LARSEN, C. E.

and BALSDON, J. (1993) Plant Molec. Biol. 22, 67±82. and PREISS, J. (1986). J. Biol. Chem. 261, 16256±16259.

BABA, T., NISHIHARA, M., MIZUNO, K., KAWASAKI, T., SHIMADA, H. KOBAYASHI, E., OHNISHI, S., TANAKA, K.

and

ARAI, Y.

(1993) Plant Physiol. 103, 565±573. and FUKUI, T. (1990) J. Biol. Chem.

110.

FURAKAWA, K., TAGAYA, M. INOUYE, M., PREISS J.

111. 112. 113. 114.

and PREISS, J. (1993) Plant Physiol. 103, 1355±1359. (1990) Planta 182, 599±604. KUIPERS, A. G. J., JACOBSEN, E. and VISSER, R. G. F. (1994) Plant Cell 6, 43±52. SHIMADA, H., TADA, Y., KAWASAKI, T. and FUJIMURA, T. (1993) Theor. Appl. Genet. 86, 665±672. BOYER, C. and PREISS, J. (1979) Plant Physiol. 64, 1039±1042. POLLOCK, C. and PREISS, J. (1980) Arch. Biochem. Biophys. 204, 578±588. MU, C., HARN, C., KO, Y-T., SINGLETARY, G. W., KEELING, P. L. and WASSERMAN, B. P. (1994) Plant J. 6, 151±159. MARSHALL, J., SIDEBOTTOM, C., DEBET, M., MARTIN, C., SMITH, A. M. and EDWARDS, A. (1996) Plant Cell 8, 1121±1135.

265, 2086±2090.

115. 116. 117. 118. 119.

SIVAK, M. N., WAGNER, M. SMITH, A

FONTAINE, T., D'HULST, C., MADDELEIN, M.-L., ROUTIER, F., PEPIN, T. M., DECQ, A., WIERUSZESKI, J.-M. DELRUE, B., VAN DEN KOORNHUYSE, N., BOSSU, J.-P., FOURNET, B.

and 120.

BALL, S.

(1993) J. Biol. Chem. 268, 16223±16230.

MADDELEIN, M.-L., BELLANGER, F., DELRUE, B., LIBESSART, N., D'HULST, C., VAN DEN

and BALL, S. (1994) J. Biol Chem. 269, 25150±25157. 121. BOYER, C. D. and PREISS, J. (1978) Carbohydrate Res. 61, 321±334. 122. BOYER, C. D. and PREISS, J. (1978) Biochem. Biophys. Res. Comm. 80, 169±175. 123. HAWKER, J. S., OZBUN, J. L., OZAKI, H., GREENBERG, E. and PREISS, J. (1974) Arch. Biochem. Biophys. 160, 530±551. 124. TAKEDA, Y., GUAN, H. P. and PREISS, J. (1993) Carbohydrate Res. 240, 253±263. 125. SINGH, B. K. and PREISS, J. (1985) Plant Physiol. 78, 849±852. 126. GOLDNER, W. and BEEVERS, H. (1989) Phytochem. 28, 1809±1812. 127. VOS-SCHEPERKEUTER, G. H., DE WIT, J. G., PONSTEIN, A. S., FEENSTRA, W. J. and WITHOLT, B. (1989) Plant Physiol. 90, 75±84. 128. BOROVSKY, D., SMITH, E. E. and WHELAN, W. J. (1975). Eur. J. Biochem. 59, 615±625. 129. BLENNOW, A. and JOHANSSON, G. (1991) Phytochem. 30, 437±444. 130. KHOSHNOODI, J., EK, B., RASK, L. and LARSSON, H. (1993) FEBS 332, 132±138. 131. KHOSHNOODI, J., BLENNOW, A., EK, B., RASK, L. and LARSSON, H. (1996) Eur. J. Biochem. 242, 132±138. 132. KOûMANN, J., VISSER, R. G. F., MUÈLLER-ROÈBER, WILLMITZER, L. and SONNEWALD, U. (1991) Molec. Gen. Genet. 230, 39±44. 133. POULSEN, P. and KREIBERG, J. D. (1993) Plant Physiol. 102, 1053±1054. 134. YAMANOUCHI, H. and NAKAMURA, Y. (1992) Plant Cell Physiol. 33, 985±981. 135. MORELL, M. K., BLENNOW, A., KOSAR-HASHEMI, B. and SAMUEL, M. S. (1997) Plant Physiol. 113, 201±208. 136. BOYER, C. D. and PREISS, J. (1981) Plant Physiol. 67, 1141±1145. 137. PREISS, J. and BOYER, C. D. (1980) in Mechanisms of saccharide polymerization and depolymerization (Marshall, J. J., ed.). Academic Press, New York, pp. 161±174. 138. HEDMAN, K. D. and BOYER, C. D. (1982) Biochem. Genet. 20, 483±492. KOORNHUYSE, N., FONTAINE, T., WIERUSZESKI, J.-M. DECQ, A.

ß 2004, Woodhead Publishing Limited

139. 140.

FISHER, D. K., BOYER, C. D.

and HANNAH, L. C. (1993) Plant Physiol. 102, 1045 1046. and GUILTINAN, M. J. (1996) Plant

141. 142. 143. 144.

EDWARDS, J., GREEN, J. H.

145.

BHATTACHARYYA, M. K., SMITH, A. M., NOEL-ELLIS, T. H., HEDLEY, C.

FISHER, D. K., GAO, M., KIM, K.-N., BOYER, C. D.

Physiol. 110, 611±619. and AP REES, T. (1988) Phytochem. 27, 1615±1620. (1988) Planta 175, 270±279. and SMITH, E. E. (1978) Carbohydr. Res. 61, 395±406.

SMITH, A. M. CARTER, J.

MIZUNO, K., KAWASAKI, T., SHIMADA, H., SATOH, H., KOBAYASHI, E., OKAMURA, S., ARAI, Y.

and

BABA, T.

(1993) J. Biol. Chem. 268, 19084±19091. and

MARTIN, C.

(1990) Cell 60, 115±122. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173.

and PREISS, J. (1986) J. Biol. Chem 261, 8738, 8743. and VENEMA, G. (1990) Gene 89, 77, 84. SALEHUZZAMAN, S. N. I. M., JACOBSEN, E. and VISSER, R. G. F. (1992) Plant Molec. Biol. 20, 809±819. NAKAMURA, Y. and YAMANOUCHI, H. (1992) Plant Physiol. 99, 1265±1266. KRAM,, A. M., OOSTERGETEL, G. T. and VAN BRUGGEN, E. F. J. (1993) Plant Physiol. 101, 237±243. KIM, W. T., FRANCHESCHI, V. R., OKITA, T. W., ROBINSON, N. L., MORELL, M. and PREISS, J. (1989) Plant Physiology 91, 217±220. ROMEO, T., KUMAR, A. and PREISS, J. (1988) Gene 70, 363±376. BABA, T., KIMURA, K., MIZUNO, K., ETOH, H., ISHIDA, Y., SHIDA, O. and ARAI, Y. (1991) Biochem. Biophys. Res. Commun. 181, 87±94. SVENSSON, B. (1994) Plant Molec. Biol. 25, 141±157. JESPERSON, H. M., MACGREGOR, E. A., HENRISSAT, B., SIERKS, M. R. and SVENSSON, B. (1993) J. Protein Chem. 12, 791±805. KURIKI, T., GUAN, H., SIVAK, M. and PREISS, J. (1996) J. Prot. Chem. 15, 305±313. CAO, H. and PREISS, J. (1996) J. Prot. Chem. 15, 291±304. TANDECARZ, J. S. and CARDINI, C. E. (1978) Biochim. Biophys. Acta 543, 423±429. MORENO, M., CARDINI, C. E. and TANDECARZ, J. S. (1986) Eur. J. Biochem. 157, 539± 545. MORENO, M., CARDINI, C. E. and TANDECARZ, J. S. (1987) Eur. J. Biochem. 162, 609± 614. ARDILLA, F. J. and TANDECARZ, J. S. (1992) Plant Physiol. 99, 1342±1347. ROTHSCHILD, A. and TANDECARZ, J. S. (1994) Plant Sci. 97, 141±148. OKITA, T. W. (1992) Plant Physiol. 100, 560±564. BADENHUIZEN, I. P. (1969) The biogenesis of starch granules in higher plants, Appleton-Century Crofts, New York. KEELING, P. L., WOOD, J. R., TYSON, R. H. and BRIDGES, I. G. (1988) Plant Physiol. 87, 311±319. ENTWISTLE, G., TYSON, R. H. and AP REES, T. (1988) Phytochem. 25, 2033±2039. TYSON, R. H. and AP REES, T. (1988) Planta 175, 33±38. VIOLA, R., DAVIES, H. V. and CHUDECK, A. R. (1991) Planta 183, 202±208. È GGE, U.-I. and BORCHERT, S. (1991) Plant Physiol. 95, 341±343. HELDT, H. W., FLU HILL, L. M. and SMITH, A. M. (1991) Planta. 185, 91±96. BORCHERT, S., GROSSE, H. and HELDT, H. W. (1989) Febs Lett. 253, 183±186. BORCHERT, S., HARBORTH, J., SCHUÈNEMANN, D., HOFERICHTER, P. and HELDT, H. W. (1993) Plant Physiol. 101, 303±312. È GGE, U. I. and HELDT, H. W. (1991) Ann. Rev. Plant Physiol. Plant Mol. Biol. 42, FLU 129±144. BAECKER, P. A., GREENBERG, E.

KIEL, J. A. K. W., BOELS, J. M., BELDMAN, G.

ß 2004, Woodhead Publishing Limited

174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194.

(1993) Biochem. J. 294, 15±17. and SCHEIBE, R. (1993) Biochem. J. 296, 395±401. NEUHAUS, H. E., HENRICHS, G. and SCHEIBE, R. (1993) Plant Physiol. 101, 573 578. KOSEGARTEN, H. and MENGEL, K. (1994) Physiol. Plant. 91, 111±120. QUICK, W. P., SCHEIBE, R. and NEUHAUS, H. E. (1995) Plant Physiol. 109, 113 121. BATZ, O., SCHEIBE, R. and NEUHAUS, H. E. (1995) Planta 196, 50±57. È NEMANN, D. and BORCHERT, S. (1994) Bot. Acta 107, 461±467. SCHU MACDONALD, F. D. and AP REES, T. (1983) Biochim. Biophys. Acta 755, 81±89. ENTWISTLE, G. and AP REES, T. (1988) Biochem. J. 255, 391±396. KANG, F. and RAWSTHORNE, S. (1994) Plant J. 6, 795±805. THORNBJéRNSEN, T., VILLAND, P., DENYER, K., OLSEN, O.-A. and SMITH, A. (1996) Plant J. 10, 243±250. DENYER, K., DUNLAP, F., THORNBJéRNSEN, T., KEELING, P. and SMITH, A. M. (1996) Plant Physiol. 112, 779±785. MILLER, M. E. and CHOUREY, P. (1995) Planta 197, 522±527. POZUETA-ROMERO, J. FREHNER, M., VIALE, A. M. and AKAZAWA, T. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 5769±5773. BATZ, O., MAAû, U., HEINRICHS, G., SCHEIBE, R. and NEUHAUS, H. E. (1994) Biochim. Biophys. Acta 1200, 148±154. CAO, H., SULLIVAN, T. D., BOYER, C. D. and SHANNON, J. C. (1995) Physiol. Plant. 95, 176±186. SULLIVAN, T. and KANEKO, Y. (1995) Planta 196, 477±484. SULLIVAN, T., STRELOW, L. I., ILLINGWORTH, C. A., PHILLIPS, C. A. and NELSON, O. E. (1991) Plant cell 3, 1337±1348. SHANNON, J. C., PIEN, F.-M. and LUI, K.-C. (1996) Plant Physiol. 110, 835±843. STARK, D. M., TIMMERMAN, K. P., BARRY, G. F., PREISS, J. and KISHORE, G. M. (1992) Science, 258, 287±292.

BATZ, O., SCHEIBE

and

NEUHAUS, H. E.

NEUHAUS, H. E., BATZ, O., THOM, E.

PREISS, J., STARK, D. BARRY, G. F., GUAN, H. P., LIBAL-WEKSLER, Y., SIVAK, M. N., OKITA,

and KISHORE, G. M. (1994) in Proceedings of Symposium, Improvement of Cereal Quality by Genetic Engineering (Henry, R. J. and Ronalds, J. A., eds) Plenum Press, New York and London, pp. 115±127. GUAN, H., KURIKI, T., SIVAK, M. and PREISS, J. (1995) Proc. Nat. Acad. Sci. U. S. A. 92, 964±967. SANWAL, G. G. and PREISS, J. (1967) Arch. Biochem. Biophys. 119, 454±469 . IGLESIAS, A. A., CHARNG, Y.-Y., BALL, S. and PREISS, J. (1994) Plant Physiol. 104, 1287±1294. PREISS, J. (1982) Ann. Rev. Plant Physiol. 54, 431-454. PREISS, J. (1978) in, Advances in Enzymology and Related Areas of Molecular Biology, Vol. 46, (A. Meister, ed.), John Wiley and Sons, Inc, pp. 317±381. KACSER, H. (1987) in (Davies, D. D., ed.) The Biochemistry of Plants, Vol. 11, Academic Press, Inc. New York, pp. 39±67. KACSER, H. and BURNS, J. A. (1979) Biochem. Soc. Trans. 7, 1149±1160. NEUHAUS, H. E., KRUCKEBERG, A. L., FEIL, R. and STITT, M. (1989) Planta 178, 110±122. NEUHAUS, H. E. and STITT, M. (1990) Planta 182, 445±454. DENYER, K., FOSTER, J. and SMITH, A. M. (1995) Planta 197, 57±62. BALL, S., MARIANNE, T., DIRICK, L., FRESNOY, M., DELRUE, B. and DECQ, A. (1991) Planta 185, 17±26. GIROUX, M. J. SHAW, J., BARRY, G., COBB, B. G., GREENE, T., OKITA, T. and HANNAH, L. C. (1996) Proc. Nat. Acad. Sci., USA 93, 5824±5829.

T. W.

195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206.

ß 2004, Woodhead Publishing Limited

207. 208.

SMITH, A. M., NEUHAUS, H. E.

209. 210. 211. 212.

HAWKER, J. S.

and

SINGLETARY, G. W., BANISADR, R.

STITT, M. (1990) Planta 181, 310±315. and KEELING, P. L. (1994) Austral. J. Plant Physiol.

21, 829±841.

213. 214. 215. 216. 217.

and JENNER, C. F. (1993) Austral. J. Plant Physiol. 20, 197±209. (1994) Austral. J. Plant Physiol. 21, 791±806. KEELING, P. L., BACON P. J. and HOLT, D. C. (1993) Planta 191, 342±348. KEELING, P. L., BANISADR, R., BARONE, L., WASSERMAN, B. P. and SINGLETARY, G. W. (1994) Austral. J. Plant Physiol. 21, 807±827. CHEIKH, N. and JONES, R. J. (1995) Physiol. Plant. 95, 59±66. È BER, B. T., SONNEWALD, U. and WILLMITZER, L. (1992) EMBO J. 11, È LLER-RO MU 1229±1238. LEUNG, P., LEE, Y. M., GREENBERG, E., ESCH, K., BOYLAN, S. and PREISS, J. (1986) J. Bacteriol. l67, 82±88. SWEETLOVE, L. J., BURRELL, M. M. and AP REES, T. (1996) Biochem. J. 320, 493, 498. SWEETLOVE, L. J., BURRELL, M. M. and AP REES, T. (1996) Biochem. J. 320, 487, 492.

JENNER, C. F.

ß 2004, Woodhead Publishing Limited

2 Analysing starch structure Ê bo Akademi University, Finland E. Bertoft, A

2.1 Introduction: characterising structures of starch components Most naturally occurring starch granules, regardless of the plant source or tissue, contain two principal types of polysaccharides, namely amylose and amylopectin. Both are polymers of solely -D-glucose connected by (1!4)-linkages into shorter or longer chains. Amylopectin, the major component of most starches, consists of a large number of shorter chains that are bound together at their reducing end side by a (1!6)-linkage, which makes this very large polysaccharide extensively branched.1 Amylose consists only of either a single or a few long chains, thus making the molecule linear or slightly branched.2 The amylose content of most starches is 20±30%.3±5 However, certain mutant plants, commonly called waxy because of the waxy appearance of the seed endosperm, have a much lower content, or even lack the amylose component completely.4±8 Other types possess an increased amylose content.5,7±10 The shape of the granules of high-amylose starches are often more or less deformed,11 and they may contain an additional polysaccharide component referred to as intermediate material because of its structure that in many aspects apparently is intermediate to the two main components.12 Though the chemical composition of the starch components is very simple, the analysis of their structure is not. For any polymer the sequence of the chemical units, by which it is built up, is of prime importance. Proteins are readily characterised by the sequence of their 20 different types of amino acids, and nucleic acids are routinely analysed for their nucleotide sequences. Many polysaccharides are composed of at least two carbohydrate species but in the starch components and other polyglucans it is practically impossible to organise the single type of carbohydrate into a meaningful sequence. Special parameters

ß 2004, Woodhead Publishing Limited

Fig. 2.1 Definition of types of chains and chain segments in the branched structure of amylopectin. Circles denotes glucosyl residues, horizontal lines (1!4) and vertical lines (1!6) linkages.

are therefore in use by which the starch components are characterised. These parameters are especially useful for analysis of the more complex amylopectin component. In amylopectin, short unit chains are clustered together,13±16 and the units of clusters are interconnected by longer chains.17 The definitions of the different types of chains, and segments of chains, included in this structure are outlined in Fig. 2.1. In the classical nomenclature by Peat et al.,18 A-chains are defined as unsubstituted, whereas B-chains are substituted by other chains. The

ß 2004, Woodhead Publishing Limited

macromolecule also contains a single C-chain that carries the sole reducing end group. This chain is not, however, distinguished from the B-chains in most experiments. The B-chains are further subdivided into Ba-chains, defined19 as those substituted with at least one A-chain (and possibly other A- and/or Bchains), and Bb-chains that are only substituted with one or more B-chains. The B-chains are also divided according to their positions in the cluster structure model proposed by Hizukuri.17 Thus, B1±chains are short chains, being components of the units of clusters, whereas B2, B3, etc., are long chains that span over two, three, or more clusters, thereby interconnecting them. This latter division of B-chains is not used for amylose, because clusters are not found in that component. The chains are also characterised as long and short chains (see Fig. 2.1), but there is no exact definition of their lengths. Note also that the definition can be very different in amylose compared to amylopectin. The chains are further divided into characteristic segments. An external chain is the part of a chain that extends from the outermost branch point to the non-reducing end.1 Thus, all Achains are external, whereas a part of the B-chains is external. The rest of a Bchain was called the total internal chain and includes all the glucosyl residues involved in branch points.20 An alternative definition is core chain, from which the outermost branched residue is excluded.21 Finally, internal chains are defined as the segments of the B-chains between the branches, excluding the branch point residues.1 For practical reasons, the segment at the reducing end side of the molecule is also considered an internal chain. The parameters above are all conveniently obtained by specific hydrolysis of the starch components with enzymes. There are three major classes of enzymes useful for analytical purposes.22 Probably most used are the debranching enzymes that specifically hydrolyse the -D-(1!6)-linkages, thereby releasing the units of chains from the macromolecule. The second group are the exoacting enzymes that attack the -D-(1!4)-linkages near the non-reducing ends. Because the enzymes cannot bypass a branch point, most of the external chains are removed, leaving a limit dextrin in which all branches are preserved. Glucoamylases are, however, an exception and can, under certain circumstances, also attack the branch and thereby eventually degrade the starch components completely into glucose. The third group are endo-acting enzymes, for which internal chain segments are the principal substrate. However, external chains are also attacked. The principles for the structural analysis of the starch components, focused on enzymic methods and amylopectin as the major subject, are highlighted in this chapter.

2.2

Fractionation of starch

When a structural analysis is to be performed, the starch sample generally has to be fractionated into its components amylose, amylopectin and, in some cases,

ß 2004, Woodhead Publishing Limited

intermediate material. It is necessary to remove the fatty material from the starch granules by Soxhlet extraction in 85% methanol23 and then completely dissolve the sample before the fractionation. Granules are dissolved in 90% dimethylsulphoxide (DMSO) by stirring at room temperature or on a boiling water bath;24, 25 6±10 M urea solution,26 or a mixture of 90% DMSO and 10% 6 M urea has also been used.27 As an alternative, 0.5 M KOH or NaOH act as an optimal solvent for starch.28 In many applications the starch granules are dispersed in 1±2 M hydroxide solution and then diluted.29±31 When extremely high pH is used, heat should be used very carefully, or avoided, to prevent any alkaline degradation of the carbohydrates. Molecular aggregates of amylopectin tend to remain in the solution, however.30±32 To obtain a more completely destructured sample, the starch is precipitated in ethanol and then redissolved.24, 25, 30 The classical method for starch fractionation is that of Schoch,33 and modifications thereof,34,35 in which an insoluble complex between amylose and n-butanol is allowed to settle. Amylopectin remains in the supernatant. Lansky et al.36 used a commercial mixture of amyl alcohols (Pentasol) to precipitate amylose. In a slightly modified procedure a mixture of aqueous n-butanol and 3-methyl-n-butanol was used.37,38 A very high molecular weight amylopectin component was obtained from some maize starches as a middle, loose layer on the amylose fraction in a similar butanol mixture.39 Several re-precipitations increase the purity of the fractions.38, 40 Banks and Greenwood41 described a procedure suitable for fractionation of cereal starches, in which the amylose was precipitated as a thymol complex. Another approach for fractionating starch was described by Matheson and Welsh.42 In this method the amylopectin component, rather than the amylose, is precipitated as a complex with the lectin concanavalin A. The lectin (a protein) is then destroyed by hydrolysis with a protease.43 Though the method gives very pure amylopectin preparations, they will also contain branched intermediate material if such exist in the original starch sample.42,44 The binding of concanavalin A to amylopectin is dependent on the concentration of the lectin and the structure of the polysaccharide.45 Gel-permeation chromatography (GPC) has also been used for small-scale fractionation.46,47 Amylopectin is eluted at the void volume of the column and is easy to collect,46 whereas amylose is partly included into the gel particles. In some cases intermediate materials were reported to elute with volumes intermediate to that of amylopectin and amylose.24,44 A common media is Sepharose CL 2B eluted with an alkaline solution.48±50

2.3

Analysis of amylose

Though amylose is the minor component in most granules, it has a large influence on the properties of starch. The traditional definition of amylose is a molecule composed of a long, linear chain of (1!4)-linked -D-glucosyl

ß 2004, Woodhead Publishing Limited

residues. Most starch preparations also contain, however, slightly branched amylose molecules. There is no specified smaller size for the definition of amylose. In some cases even comparatively short linear dextrins, obtained by hydrolysis of starch, were called amylose. This is unfortunate, because such fragments can also be produced by hydrolysis of amylopectin, and may therefore confuse the reader.

2.3.1 Amylose content of starch Amylose was probably the first biopolymer for which a helical structure was proposed.51 The well-known deep blue complex formed with iodine was later proved to involve amylose in a helical conformation,52 details of which have been investigated.53±55 The colour and intensity of the complex depends on the chain length (CL) of the amylose.56 At CL >80 the wavelength maximum (max) of the absorption of light is >610 nm. The max shifts to lower wavelengths and the colour shifts to red for shorter chains.56 Thus, the short chains of amylopectin possess a max at 530±575 nm.57±60 The blue value61 (BV) is defined as the absorbance at 680 nm (though sometimes measured at 640 nm) of 1 mg starch in 100 ml of a mixture containing 2 mg I2 and 20 mg KI. For a `true' BV the absorbance should be multiplied with the factor 4 because old colorimeters used 4-cm cyvettes rather than the 1-cm cyvettes used in modern instruments. BV for amylose38±40,57,58,60,62 is 1.01±1.63, whereas that for amylopectin39,57,58,60,62 is 0.08±0.38. Though the BV is easy to measure, it should be considered mainly as a qualitative test for amylose. The most frequently used quantitative test for amylose is to measure the iodine affinity (IA) potentiometrically.63 An automated amperometric titration method was also described.59 In most cases approximately 20g of iodine is bound per 100g amylose, in contrast to only 0.5±1.1g iodine/100g amylopectin.38,39,57,59,60 The amylose content is calculated from IA of a defatted starch sample, and assuming the IA value of amylose to be 20, as:59 % Apparent amylose ˆ IAstarch =20  100

‰2:1Š

It should be noted that the amylopectin component in some starches, notably high-amylose starches, possesses unusually long average chain length,25,64±67 which increases IA for the starch and leads to overestimation of the amylose content. The amylopectin of indica varieties of rice contain unusually high levels of long chains,59 which also interfere. It is, however, possible to correct the obtained apparent amylose values to true values if the IA of the purified amylopectin is taken into account:59 % Amylose ˆ …IAstarch ÿ IAsamylopectin †=…IAamylose ÿ IAamylopectin †  100 ‰2:2Š A considerable part of the amylose in many starches, especially cereal starches, is complexed with lipids, mainly lysophospholipids.3,5,68,69 These lipid-amylose complexes (LAM) have no affinity for iodine. A colorimetric determination of LAM and the total (true) amylose content was described by Morrison and

ß 2004, Woodhead Publishing Limited

Laignelet.27 In this, starch granules are dissolved in hot 90% DMSO-10% 6 M urea, treated with a mixture of I2 and KI, and the absorbance at 635 nm is measured, from which the fraction of apparent amylose (free amylose, FAM) is obtained. A part of the dissolved sample is defatted in ethanol before the measurement to obtain the total amylose. In both cases the content is calculated from the general equation: % Amylose ˆ …28:414  absorbance† ÿ 6:218

‰2:3Š

LAM is obtained as the difference between the total amylose and FAM.27 A rapid method for the estimation of amylose in maize starches, including high-amylose starch, was developed by Knutson and Grove.70 The starch granules are gelatinised without heat in 3 M CaCl2 and then further sonicated at 60±70ëC in an iodine-DMSO mixture. After dilution in water, the absorbance is read at 600 nm and the amylose content is determined from a standard curve. A micro scale method was described by Mohammadkhani et al.,71 in which only 2±3 cereal seeds are needed for the amylose determination. Campbell et al.72 suggested that near-infra-red transmittance spectroscopy could partly replace the iodine staining methods and enable rapid screening of apparent amylose contents in breeding programmes. Amylose contents of starches were also estimated directly by GPC24,31,73 or by high-performance size-exclusion chromatography (HPSEC).67,74,75 There is a risk, however, that solubilisation problems related to the amylopectin component affect the result.30 Though no baseline separation of amylose and amylopectin normally is obtained, GPC on a column packed with TSK HW75 S was found to be the most accurate among several different amylose determination methods.44 The fraction of amylose in GPC is identified by measuring the BV76 of the collected fractions,24,50,73,77 or on-line by postcolumn iodine injection in the case of HPSEC.78±80 An alternative method is to enzymatically debranch the starch sample prior to GPC3,24,68,77,81,82 or HPSEC.67, 83 The long amylosic chains are eluted from the column before the short chains of the amylopectin.84 Also in this case, however, a baseline separation between the two fractions is difficult to achieve.

2.3.2 Structural analysis of amylose The size of amylose molecules is more frequently given as the degree of polymerisation (DP) than as molecular weight, and is obtained by light scattering,85 from the limiting viscosity number,86 or by reducing end analysis with the Park-Johnson reagent87 as modified by Hizukuri et al.37,88 The numberaverage DP (DP n ) from different plants varies between 0.51± 6:34  103.37,38,40,57,58,85,89±91 Weight-average values are higher (Table 2.1).38,60,85 The size-distribution was analysed by HPSEC on Toyosoda TSKGEL PW columns, coupled to simultaneous refractive index (RI) and low-angle laser light scattering (LALLS) detection, for a range of samples.85,92 Broad distributions were obtained for amylose from potato (DPw 0.84±21.8103) and

ß 2004, Woodhead Publishing Limited

Table 2.1

Properties of amyloses

Source

Rice (japonica)38, 97 Rice (indica)38, 97 Maize90 Wheat58 Barley60 Water chestnut111 a b

Branched fraction as ±limit dextrin

Whole fraction DPw

DPn

DPw

DPn

CLn

NCa

MFb

2820±2950 2290±2720 2270±2500 2360±5450 5580 4210

1100±1110 920±1040 930±990 830±1570 1570 800

3080±3500 2480±3460 2700±3000 1670±5880 5930 4150

890±1030 790±890 790±850 700±1430 1440 1200

105±115 90±105 140±160 50±71 120 108

8.5±9.0 5.7±9.7 5.3±5.4 12.9±20.7 12.0 11.1

0.31±0.43 0.32±0.49 0.44±0.48 0.26±0.44 0.35 0.11

Average number of chains per branched amylose molecule. Molar fraction of branched amylose.

tapioca (0.58±22.4103), whereas amylose from kuzu possessed a more narrow distribution (0.48±12.3103). A considerable portion of the amylose fraction, from 10±50% depending on the sample, contains slightly branched macromolecules.37,93,94 These are generally larger than the linear amyloses2,95 and possess between 5±20 chains per molecule (Table 2.1). The C-chain of branched amylose from maize was tritium labelled at the reducing end by treatment with sodium 3H-borohydride and CL ranged from 200±700.96 Takeda and Hizukuri with co-workers2,94,97 studied branched amyloses by preparing the -amylolysis limit dextrins ( -LDs) of the amylose fraction. All linear amylose, together with the external chains of the branched amylose, is hydrolysed. (For a more detailed description of the use of enzymes, see Section 2.4.) The molar fraction of branched amylose (MFbranched) is calculated58, 94 from: MFbranched ˆ NCwhole amylose sample =…NC ÿLD ÿ 1†

‰2:4Š

NC is the average number of chains per molecule. Takeda et al.98 subfractionated maize amylose in aqueous n-butanol. A minor, branched component with high molecular weight remained in the supernatant. It was shown that the component contained a fraction of very short chains that were suggested to form small `immature' clusters.

2.4

Analysis of amylopectin structure

The molecular weight of amylopectin is considerably larger than that of amylose and no suitable media are available for a size-distribution analysis by GPC or SEC. Because of the tendency to form molecular aggregates,31,32,99,100 as well as the risk of fragmentation,101 it is also difficult to obtain accurate average estimations of the true size of this gigantic macromolecule. Mw-values range from 2±700106, depending on the source of plant, method of determination, and solvent for the amylopectin sample.6,32,99,101±106 Mn-values are much lower.107,108 When the reducing power is measured with the modified Park-Johnson method,37,88 DPn values are 4.8±15.0103 in different samples,39,58,90,109±111 which is only slightly higher than those of amyloses and corresponds to Mn 0.8±2.5106. By fluorescent labelling of the reducing end followed by gel-permeation chromatography, the amylopectin component from several plants was divided into three molecular species with DPn values from 0.7±26.5103.112 Thus the Mw/Mn- ratio is of the magnitude 101±102. It should be kept in mind, therefore, that an amylopectin preparation includes a very broad range of molecular sizes and, eventually, the fine structure of different size-classes is not identical. The use of different enzymes to investigate the amylopectin structure is outlined in Fig. 2.2. Different amylases are active in a concentration of DMSO of 2.5 M or less,113 which should be taken into account if the sample is dissolved in DMSO prior to the enzymic degradation. Note that several of the methods described in this section are also applicable to amylose and intermediate materials.

ß 2004, Woodhead Publishing Limited

Fig. 2.2

Principles of the use of enzymes in structural studies of amylopectin and other branched starch components.

2.4.1 Unit chain length and distribution The most common, though not necessarily the most informative, structural analysis is to measure the length and the size-distribution of the unit chains in amylopectin. The average chain length (CL) is obtained from CL ˆ Gtot =NC

ß 2004, Woodhead Publishing Limited

‰3:5Š

in which Gtot is the total number of glucosyl residues (or total carbohydrate content) and NC is the number of chains in the sample. NC equals the number of non-reducing ends and is obtained by a modified37,40 rapid Smith degradation method.114 The glycerol liberated from the terminal non-reducing ends is then measured enzymatically.37 Alternatively, the sample is first debranched enzymatically, as described below, and then the reducing end of each chain is measured with the Nelson reagent.115 With the modified37,88, 116 Park-Johnson reagent87 approximately ten times higher sensitivity is obtained. A very sensitive alternative is the 2,2'-bincinchoninate reagent,117,118 which also was adopted to micro-scale applications.119 CL can also be measured from the signals obtained by 1H- and 13C-NMR spectroscopy.120 CL equals approximately the ratio of -(1!4) to -(1!6) linkages, because amylopectin is a very large molecule. The H-1 proton at the non-reducing terminal was assigned a chemical shift separate from the other H-1 protons, which increases the precision.62 With NMR there is no need to debranch the sample prior to the analysis, and the agreement with the enzymic methods is good.62,120 Considerably more information than is given by the CL value is obtained from the unit chain distribution, often called a unit chain profile analysis. Prior to the analysis, the sample is debranched with isoamylase or pullulanase1 (Fig. 2.2). The enzymes have different action patterns against branched oligosaccharides and against intact amylopectin.121 Isoamylase attacks the branches by an endo-like pattern, whereas pullulanase was characterised to perform exo-attack.122 The Km-value for isoamylase with amylopectin as the substrate is lower than for pullulanase,123 and it is easier to obtain a more complete debranching with the former.122, 124 Whether all the branches have been hydrolysed is conveniently controlled by the addition of -amylase after the debranching experiment. This enzyme degrades the sample completely into maltose (and small amounts of maltotriose) if no branches remain.124 The unit chain profile has been analysed by GPC35,39,50,125±132 or HPSEC3,19,47,128,133±135 for a broad range of amylopectins. All samples, regardless of the source, possess typically a major group of short chains and a minor group of long chains, the division between the groups being approximately at DP 30±40. Most samples have, however, a polymodal distribution, which is best obtained by HPSEC.17,25,58±60,66,109±111,136 The long chains, thought to span between the clusters of short chains, are therefore subdivided into B2 (DP approx. 35±60), B3 (60±80), etc.17 Extremely long chains with DP in the order of 103 were also described for several samples.19,58±60,109±111 Their connection to the clusters is not known, but they were suggested to be Bchains with widely spaced side-chains109 or possibly represent the C-chain of the amylopectin macromolecules.110 The peak of short chains possesses in some samples a shoulder and was therefore also sub-divided.17,58,60,134 The shortest chains were suggested to represent A-chains, whereas the other sub-group is short B-chains (B1-chains).17 The unit chain profile was also transformed into gaussian distributions, by which two groups of short A-chains were described in

ß 2004, Woodhead Publishing Limited

Table 2.2

Chain lengths and molar ratios of some amylopectins

Source

CL

ECL

ICL

S:La

A:Bb

Waxy maize21 Waxy maize163 Waxy maize20,168 Maize21 Maize163 Waxy rice21 Waxy rice167 Waxy rice172,174 Rice167 Pea43 Pea283 Potato132 Potato173 Amylose-free potato279

18 23 22 22 26 20 18 17 19 26 n.a. 28 21 24

12 15 14 14 17 12 12 11 13 16 n.a. 17 14 15

6 8 6 7 8 6 5 5 5 9 n.a. 9 6 8

n.a.c 8.1 11.5 n.a. 9.0 n.a. n.a. 10.5 n.a. 11.5 7.3 4.4 6.6 n.a.

1.3 1.1 1.0 1.2 1.2 1.2 1.3 0.8 1.2 n.a. 1.4 n.a. 1.5 1.2

a b c

Molar ratio of short:long chains. Molar ratio of A:B chains obtained from debranching of ,  or ; -limit dextrins. Not analysed.

addition to several groups of B-chains.136 Table 2.2 shows typical CL values of some amylopectins. The distribution of the C-chain of amylopectin was described recently by Hanashiro et al.134 Samples were labelled at the reducing end with the fluorescent dye 2±aminopyridine prior to the debranching and then fractionated by SEC. The distribution of the C-chains was detected from the fluorescence emission and showed that C-chains were distributed over the DP range 10±130, with a peak around DP 40.112,134 The true nature of the extra long chains found in some samples110 remains therefore unclear. High-performance anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD)137±139 has largely increased the resolution of individual chains up to the order of DP 50±70.4,31,60,140±146 Unfortunately, the amperometric detection is not directly proportional to the carbohydrate content,137,147 but it is possible to adjust this by calibration of the response with fractions of known CL and carbohydrate content.46 The problem was also avoided by coupling a short column with immobilised glucoamylase after the main column. The glucoamylase hydrolyses the chains completely into glucose before entering the PAD detector thus giving a proportional response regardless of the chain length.128 An additional benefit is that the sensitivity of PAD for the longer chains increases. Two unit chain profiles, presented as bar graphs, are shown in Fig. 2.3. Waxy maize is a typical A-crystalline starch sample135 with a comparatively high molar ratio of short:long chains (S:L ratio, Table 2.2). A typical B-crystalline starch is represented by the amylopectin from an amylose-free potato, which has a lower S:L ratio. Note that the shortest chain in both samples is DP 6, which is a

ß 2004, Woodhead Publishing Limited

68

Starch in food

Fig. 2.3 Bar graphs redrawn from HPAEC-PAD chromatograms showing the unit chain distribution of two amylopectin samples and their limit dextrins. Bars for DP >60 are only approximate. Groups of different chains in amylopectin and limit dextrins are based on the nomenclature of Hizukuri17 and Bertoft and Koch,174 respectively.

general feature of all amylopectins described so far (though trace amounts of shorter chains may exist). The distribution pattern of the shortest chains with DP 6±8 is characteristic for many naturally occurring A- and B-crystalline types, respectively.137 High-amylose maize amylopectin samples of mutant plants, which have B-crystalline granules,135 possess a pattern similar to normal maize amylopectin. 4 C-crystalline starches, being mixtures of A- and Bpolymorphs,148,149 possess intermediate patterns.137,141 The unit chain profile was also analysed by capillary electrophoresis,150 or by using a DNA sequencer.151 In this method, the chains of the debranched sample are labelled at their reducing ends with the negatively charged fluorophore 8-amino1,3,6-pyrenetrisulphonic acid, which is registered by a fluorescence detector. The resolution is at least as high as in HPAEC, and the result represents directly the molar distribution of the unit chains, rather than the weight distribution.150, 152

2.4.2 External chain length and internal chain distribution Though the unit chain length distribution provides knowledge about the overall composition of chains, a more detailed understanding of the fine structure of amylopectin needs information of the distribution, or positions, of the chains within the macromolecule.153 If exo-acting enzymes remove the external chains, limit dextrins containing the internal parts together with all original branches remain and are used for the study of the internal structure. In addition, information about different types of chains is better achieved.

ß 2004, Woodhead Publishing Limited

Analysing starch structure

69

Fig. 2.4 Action pattern of sweet potato -amylase and rabbit muscle phosphorylase a on a branched substrate. The residual external chain stubs of the -limit dextrin, adapted from Summer and French,154 and the - and ; -limit dextrins, adapted from Bertoft,168 are shown as filled circles.

The enzyme -amylase hydrolyses every second -D-glucosidic (1!4)linkage from the non-reducing ends, thereby producing maltose.1 The enzyme cannot bypass the branches, thus leaving a resistant -limit dextrin ( -LD) with remaining short external chain-stubs of characteristic lengths154,155 (Fig. 2.4). By performing the hydrolysis inside a dialysis membrane, the maltose is directly removed from the -LD.156,157 Alternatively, the dialysis is made after the reaction18,21,89,155,158±160 or the -LD is precipitated in methanol160±162 or ethanol.163 Maltose was also removed by GPC on a short de-salting column.164 The -limit value is defined as the relative amount of maltose formed in the reaction. Typically, amylopectins possess -limit values in the order of 50± 65%.6, 21, 45, 58±60, 110, 165±167 The -limit value is used to calculate the average external chain length (ECL) from1 ECL ˆ CL  …% -limit=100† ‡ 2

‰2:6Š

in which it is assumed that the sample possesses an equal number of A- and Bchains, as well as of chains with odd and even numbers of residues therefore the average length of the external stubs in -LD is two glucosyl residues. Phosphorylase a, an enzyme isolated from rabbit muscle, produces glucose 1phosphate from the non-reducing ends1 and leaves at average 3.5 residues next to a branch point168 (Fig. 2.4). The cleavage of the glucosidic linkage proceeds by phosphorolysis, rather than hydrolysis, and the phosphate is provided as a phosphate buffer. The reaction is readily reversible,169 however, and in order to obtain a true -LD, the concentration ratio of phosphate:glucose 1-phosphate should exceed 10 at the end of the reaction.168 If the reaction is made inside a dialysis bag,155 the concentration of glucose 1-phosphate is continuously kept at

ß 2004, Woodhead Publishing Limited

a low level. Because the residual external chain stubs are longer, the -limit value is typically somewhat lower than the -limit value45, 168±171 and the ECL is obtained from:168 ECL ˆ CL…%-limit=100† ‡ 3:5

‰2:7Š

If the -LD is further treated with -amylase,124,156,168,172,173 each external chain stub will be attacked once and the molar amount of maltose formed equals the number of unit chains.168 The average length of the external stubs of the resulting limit dextrin (; -LD) is 1.5 and the structure168 equals the smallest possible -LD (Fig. 2.4). Because the number of internal chains equals the number of branches, which are one less than the number of chains (NC), the internal chain length (ICL) is obtained from:168 ICL ˆ ‰…CL ÿ ECL†  NCŠ=…NC ÿ 1† ÿ 1

‰2:8Š

When the NC is large or unknown, the following approximation is used:1 ICL ˆ CL ÿ ECL ÿ 1

‰2:9Š

Typical ECL and ICL for some amylopectins are shown in Table 2.2. Pullulanase attacks short maltosyl chain stubs more efficiently than isoamylase19 and is used to completely debranch - or ; -LDs. The chain distribution can be analysed chromatographically and the maltose obtained from the ; -LD equals the number of A-chains, whereas all longer chains represent the B-chains173, 174 (see Fig. 2.3). From a -LD, half of the A-chains are obtained as maltose and half as maltotriose.89,126,131,144,150,156,159,163,175±180 A small amount of the shortest B-chains are also found as maltotriose, however, and might interfere slightly on the estimation of A-chains. The ratios of A:Bchains of some amylopectins estimated from limit dextrins are shown in Table 2.2. Note that when using limit dextrins, the estimation of A-chains is not dependent on their original length. It was shown that the ratio of A:B-chains obtained from limit dextrins tends to be lower than that estimated from debranched native amylopectin, when assuming that the shortest chains represent A-chains.180 The distribution of the B-chains of ; -LDs represents their total internal lengths plus the single external residue (Fig. 2.3). The sub-division of long Bchains is similar to the unit chain profile of the native samples, but the short B1chains are also found as sub-groups. These were designated B1a and B1b,130 and the latter group was also produced as a result of -amylase attack at longer internal chain segments between clusters.20, 174 B1a-chains were further sub-divided into very short B1a(s)-chains and somewhat longer B1a(l)-chains.174 These are thought to be the main components of the individual clusters and their external chain segments build up the crystalline lamellae inside the starch granules together with the A-chains.174 The B1a(s)-group was suggested to represent the internal chains of building blocks inside clusters181 (see Section 2.4.3). Isoamylase attacks branched maltosyl chain-stubs very slowly compared to longer branches, and this was used by Hizukuri and Maehara19 to perform a

ß 2004, Woodhead Publishing Limited

partial debranching of -LD from wheat amylopectin (see Fig. 2.2). By careful choice of the conditions for the reaction, half of the A-chains (being the maltosyl stubs) remained bound to the B-chains. Zhu and Bertoft173 performed a partial debranching of the ; -LD of potato amylopectin, for which all the A-chains remained bound to the B-chains. Chains carrying only other B-chains (Bbchains) were completely debranched and hydrolysed into maltose by the addition of -amylase. Remaining branched chains were the Ba-chains. Wheat amylopectin possessed a ratio of Ba:Bb chains of 1.5:1,19 and the corresponding ratio for potato amylopectin was estimated to 2.2:1.173 The ratio of A:Ba-chains was 2.1:1 in both samples. Hizukuri182 defined the span length (SL) as the average length between A-chains along a Ba-chain. SL increased with the length of the Ba-chains in wheat amylopectin from 2.7±10.9.182 The number of Achains increased from 1±2 per short Ba-chain (DP of the branched chain approx. 9) to 3±4 per long Ba-chain (DP 46).19

2.4.3 Analysis of units of clusters The unit chains of amylopectin are packed into clusters with densely grouped branches.13±16, 183 An exact definition of a unit of cluster does not exist, however, and comparatively few attempts to study the clusters in more detail have been published. Thurn and Burchard103 presented a model in which one cluster contains, on average, only 4.22 chains, whereas Hizukuri17 suggested that some 22±25 chains are involved. Gallant et al.184 concluded that a cluster is built up by 18±34 chains in the form of double helices. Endo-acting enzymes have been used to cleave the internal chain segments found between the units of clusters, thereby releasing them (see Fig. 2.2). Because different enzymes have different action patterns and affinity to the chain segments between branches, the results from different laboratories are somewhat contradictory, reflecting the need for a more exact definition of the cluster. Finch and Sebesta185 treated -LD of amylopectin with a maltotetraoseforming amylase from P. stutzeri. The resistant limit dextrins obtained were suggested to represent the unit clusters. The Mn of the dextrins from potato amylopectin was 23,000 (corresponding to DP approximately 140), whereas wheat amylopectin possessed clusters only one-third this size. In both samples the clusters were uniform in size.185 Bender et al.161 studied the action of cyclodextrin glycosyltransferase (CGT). In addition to producing cyclodextrins by attack at the external chains, CGT also attacks longer internal chains.186 The intermediate, non-cyclic dextrins produced from waxy maize and potato amylopectin were precipitated in methanol into three major sizes suggested to represent clusters.161 The DP of the -LD of these clusters ranged from 40±140 and was similar in both samples. The clusters were, however, suggested to be more tightly packed in waxy maize. Bertoft and co-workers used the -amylase from B. amyloliquefaciens (earlier called liquefying amylase of B. subtilis) to isolate clusters (Fig. 2.2). This enzyme is similar to CGT in its action pattern. It contains nine subsites

ß 2004, Woodhead Publishing Limited

Table 2.3

Sizes of clusters in various amylopectinsa Cluster sizeb

Sample

Waxy rice172 Waxy barley284 Waxy maize160 Wxdu maize164 Aewx maize164 Potato173 Amylose-free potato181 Smooth pea283

XXL

XL

L

M

S

200

150 130 117 94

90 85 88 84 47 70 48

47 65 68 38 32 50 40 55

34 27 25 17 33 31 32

a The average degree of polymerisation (DP) or peak-DP of ,  or ; -limit dextrins in GPC analyses. b Extra large, large, medium, and small clusters found within a sample.

unevenly distributed around the catalytically active site,187, 188 so that attack at the external chains preferentially results in the formation of linear (instead of cyclic) maltohexaose.160,187 Simultaneously, preferential attack at long internal chains between clusters occurs.168 The reaction is initially fast when all subsites are filled with D-glucosyl residues, but slows down considerably when this criterion is not fulfilled.189 At this stage, units of clusters dominate the reaction mixture and can be size-fractionated in methanol.190 The estimated DP of clusters in the form of limit dextrins from a range of amylopectin sources is shown in Table 2.3. All samples analysed so far contained clusters of at least two or more sizes. Because the result is based on the rate of the reaction, and the dextrins under study are intermediate products from the hydrolysis, the smaller dextrins obtained in a mixture represent either very small, true clusters or subcluster units formed from larger clusters. It should be noted that other amylases, with different number and distribution of subsites around the catalytic site, could give a different result, but have not been tested so far. The structure of the clusters of some samples were analysed in more detail. LD of waxy maize amylopectin clusters were debranched and the chain profile analysed.20 The number of chains in small clusters was 5±8, whereas larger units contained 11±19 chains. The ratio of A:B chains remained unchanged (1:1) during the -amylolysis and was similar in all size-fractions, thus suggesting that the mode of enzyme attack resulted in an even production of each type of chains. A similar result was obtained from waxy rice amylopectin and for this sample it was also shown that the ratios of A:Ba and Ba:Bb chains remained practically constant.174 However, the long B-chains were largely reduced in number due to their cleavage, and B-chains of intermediate sizes corresponding to B1b-chains were preferentially formed.20, 174 This was in agreement with the suggestion that the long chains are involved in the interconnection of the clusters.17 For waxy rice ; -LDs, the shortest possible B-chain, maltotriose, also increased in number.174 Though a part of these chains possibly derived from

ß 2004, Woodhead Publishing Limited

A-chains as a consequence of -amylase attack at external segments,143 a major part was thought to represent new B-chains from the attack at internal chains and suggested a preferential mode of the interconnection of clusters in that sample.174 For potato amylopectin, the ratios of A:B and A:Ba chains increased slightly in smaller dextrins, whereas Ba:Bb decreased. Such changes were also suggested to reflect certain preferential modes of cluster interconnection.173 Fractions containing ; -LD of intermediate -dextrins have also been treated a second time with -amylase.172, 173 For waxy rice, two types of hydrolysis patterns were found among the fractions,172 suggesting that they represented different types of structures. Such differences were further suggested to represent structural domains172 though it remained unclear whether this appeared at an intra- or inter-molecular level or possibly inter-granular level. Because the waxy rice lacked amylose completely, a possible domain structure could be found between amylopectin in the semi-crystalline and amorphous growth rings of the granules.172 The clusters of one of the potential domains (DomA) were of at least three sizes with DP from 100±200, whereas DomB apparently had more uniform cluster size around DP 90. For potato amylopectin, one out of four fractions was hydrolysed by a clearly faster rate and possibly, therefore, also represented a different structure.173 Whereas a diluted -amylase preparation was used for the isolation of clusters, a 100 times more concentrated solution was used to produce -LD from the clusters of waxy rice amylopectin172 (Fig. 2.2). These -LDs contain only very short internal chains that are resistant to further attack and were suggested to represent building blocks from which the units of clusters are constructed. A cluster contained 3±8 branched building blocks depending on its size,172 and the building blocks were composed of 2±7 chains, mainly B1a(s).174 Interestingly, this number is of the same order as suggested by Thurn and Burchard for the unit cluster.103 The size-distribution of the building blocks analysed by GPC was of two different types corresponding to the tentative domains.172 Dextrins of DomA type possessed building blocks of the range DP 5±40, with a peak at DP 35±40 on a weight basis. DomB had a more uniform distribution with building blocks mainly of intermediate and small sizes (DP 5±30). Building blocks of similar types as in waxy rice was found in a mutant maize starch (waxy:dull) that also is of the Acrystalline polymorph.164 A predominately B-crystalline maize mutant starch sample (amylose-extender:waxy) possessed a different distribution with smaller building blocks.164 The sizes of the clusters in the latter sample were also small with average number of blocks per cluster from 1.2±2.3. Thus, it is possible that the architecture of the clusters is connected to the type of crystallinity of the starch granules. The composition of building blocks181 and other hydrolysis mixtures of dextrins of similar sizes were also analysed by HPAEC147, 191±193 and NMR.191

2.4.4 Starch phosphate esters Some starches, in particular from potato and other tubers or roots,31, 141, 194 possess phosphate esters substituted mostly at C-6 and C-3 positions of the D-

ß 2004, Woodhead Publishing Limited

glucosyl units of the amylopectin component, whereas the amylose is essentially free from phosphate.33 A dikinase enzyme was found in potato and suggested to be responsible for the phosphorylation.195,196 Though the total phosphorus content is low, in potato starch corresponding to approximately one phosphorylated residue in 200±500, 31,194 it greatly differs between cultivars197,198 and is of importance for the functionality of starch. Surprisingly, amylose-free potato starch was reported to contain similar levels199 or less phosphorus,141 whereas a very high amylose containing potato starch contained higher levels than normal starch.10 The reason might be the higher affinity for long chains of the branched substrate possessed by the phosphorylating enzyme.196 Cereal starches contain also phosphorus, but this is only to a minor degree covalently bound to amylopectin.19,58,59,90,109 Instead, most of this phosphorus is found as lysophospholipids that are complexed, not covalently linked, to the amylose fraction of the granules.200,201 The esterified phosphorus content and positions on the D-glucosyl residues have been analysed by 31P-NMR.194,197,198,202±204 Total starch bound phosphorus was also analysed chemically.197,205,206 After controlled acid hydrolysis of all glycosidic linkages in hot 0.7 M HCl for four hours,207 the amount of glucose 6±phosphate residues was analysed enzymatically by the glucose 6±phosphate dehydrogenase catalysed reduction of NADP+.197,207,208 Glucose 6±phosphate and glucose 3±phosphate were also separated by HPAECPAD.209 However, the quantitative estimation of the more labile component glucose 3±phosphate is difficult because of partial degradation. Generally, about 70% of the phosphorus is found at the C-6 position and 30% at C-3.198, 202, 207 The location of phosphate esters on the amylopectin macromolecule was analysed using the enzyme glucoamylase of Rhizopus delemar207,210 or Aspergillus niger.173,203,210,211 This is an exo-acting enzyme that produces glucose by hydrolysis of the -D-(1!4)-linkages. If the entire external part of the B-chain is removed together with all -D-(1!4)-linkages of the A-chain, the residual single glucosyl stub is also removed by attack at the -D-(1!6)linkage, and the amylopectin is completely degraded.22 If, however, the requirements for attack at the (1!6)-linkage are not fulfilled, a -limit dextrin is produced. This is the case if a substituent, like a phosphate group, is found in either the A- or the B-chain. Such groups act as barriers and cannot be by-passed by the enzyme210,211 (Fig. 2.5). Only a minor part of the macromolecule remains resistant to glucoamylase attack; reported -limit values are 81±97%.173,211 Phosphorylated amylopectin has also been debranched enzymatically and the phosphorylated chains were isolated by anion-exchange chromatography on DEAE-Sephadex A-50.210,212 By further treatment with beta-amylase, Takeda and Hizukuri213 showed that the phosphate is largely associated with the long Bchains of potato amylopectin, and the degree of phosphorylation seems to increase with the length of the chains.194 Moreover, no phosphate is found closer than 9 D-glucosyl residues from the non-reducing ends or in the vicinity of the branches.213 Blennow et al.202 performed partial acid hydrolysis of the starch granules, after which mainly the crystalline lamellae remains, and showed that

ß 2004, Woodhead Publishing Limited

Fig. 2.5 Action pattern of fungal glucoamylase on phosphorylated branched and linear substrates. ", phosphate ester at C-6; #, phosphate ester at C-3. The structures of limit dextrins are adapted from Takeda and Hizukuri210 and the Handbook of amylases and related enzymes.22

most of the phosphate is associated with the amorphous parts. In another report, however, only 35.5% of the phosphorus was located in the amorphous regions of a high phosphate containing potato variety.214 C-6 phosphate was assumed to disturb the crystallisation of amylopectin215 and, accordingly, the major part of the C-6 substituted phosphate was found in the amorphous parts, whereas the phosphate at C-3 position was equally distributed between the amorphous and crystalline areas.202 In addition, Blennow et al.194,216 analysed a series of phosphorylated starches and concluded that the degree of phosphorylation is strongly correlated to the distribution of the non-phosphorylated chains in the amylopectin component. Jane and Shen217 used a chemical gelatinisation technique, in which the granules are gradually solubilised in 4 M CaCl2.218 They showed that the hilum of potato starch granules is more phosphorylated than the peripheral parts. This correlated with a longer length of the long B-chains of the amylopectin component towards the interior of the granules. They also showed that the small starch granules, which possess lower amylose content, contain a higher degree of phosphorylation than large ones.217 In addition, Bay-Smith et al.197 found that individual potato tubers are more phosphorylated at the interior parts. Thus, the phosphate esters are unevenly distributed on both the molecular and the granular levels, as well as on the tuber level.

2.5

Analysis of intermediate materials

Some starches, notably the high-amylose types, contain intermediate materials (IM). Beside mutant maize varieties7,12,24,44,50,64,129,219±221 and wrinkled

ß 2004, Woodhead Publishing Limited

peas,9,42,47,222,223 IM was described in normal maize,25 oats,41,224±226 wheat,41,227 rye, barley,41 and potato226 and in high-amylose varieties of barley9 and potato.10 The nature of this material is somewhat confusing, and might be unique for each type of starch.9 The method of fractionation of the starch could also influence the result, especially when IM is isolated together with either the amylose or the amylopectin fraction. In such cases the IM might even remain unknown for the investigator and influence the structural analysis of the other components. In several cases, the measured amylose content of high-amylose starches by iodine binding is largely overestimated due to IM with a high iodine affinity or amylopectin of longer CL than normal, 44,73,220,228 or both.12,24,64,225,228 Because IM is structurally related to either amylose or amylopectin, it is analysed by the same methods. Banks and Greenwood41 described a method adopted for cereal starches, in which an amylose-thymol complex was separated from the amylopectin component. The complex was reprecipitated in butanol into amylose and a soluble fraction containing anomalous amylose and/or amylopectin, the latter with unusually long CL. Adkins and Greenwood,34 working with maize starches, precipitated amylose in 1-butanol and from the supernatant (the amylopectin fraction) they precipitated a glucan in complex with iodine. The amount of the complex increased with increasing amylose content of the maize and was characterised as short-chain amylosic material. Wang et al.228 also found IM included in the supernatant fraction together with amylopectin. IM was fractionated by GPC on Sepharose CL 2B and found to be a branched component smaller than the amylopectin. It possessed similar types of chains as in amylopectin, but in different proportions depending on the type of mutation. Klucinec and Thompson25 precipitated amylose together with an intermediate fraction from the amylopectin in an aqueous mixture of 6% 1-butanol and 6% isoamyl alcohol. The amylose was then reprecipitated in 1-butanol, whereas IM remained in the supernatant. They found that IM in normal maize had rather similar types of chains as amylopectin, but in high-amylose starches the composition changed, and IM possessed a large group of long chains. Simultaneously, the apparently true amylopectin component contained increased amounts of the long chains.25 More, and/or longer, long chains is a typical feature associated with amylopectin in maize starches containing the amyloseextender mutation.50, 64, 73, 129, 219, 228 Klucinec and Thompson25 concluded that IM, like the amylopectin, was branched, though its structural features resulted in altered physical behaviour, as shown by its precipitation in the 1-butanolisoamyl alcohol mixture. In a double (dull:waxy) and triple mutant (amyloseextender:dull:waxy) maize starch an IM component was found in high concentration (40 and 80%, respectively).229 The material, which contained only slightly altered chain distribution, was apparently more resistant to amylase attack than the normal amylopectin and the behaviour of the -dextrins in methanol was different. Bertoft et al.229 suggested that this IM possessed a more regularly branched structure that prevented the action of the -amylase and caused the altered behaviour of the molecule.

ß 2004, Woodhead Publishing Limited

Colonna and Mercier222 isolated a very low-molecular weight component from wrinkled pea starch (also a high-amylose type starch). This IM was branched and possessed similar types of chains as the amylopectin component, but the iodine affinity was high and the S:L chain ratio low. Bertoft et al.223 sizefractionated the IM and found that the proportion of the long chains slightly increased with decreasing molecular weight. In fact, the molecular weight of the small IM resembled that of the clusters of the amylopectin component and was suggested to be composed of small, cluster-like structures interconnected by the long chains, thus increasing the proportion of these chains.223 The IM of wrinkled pea has also been described as a mixture of very short linear amylose chains and branched, either normal9 or long-chained amylopectin.73 Biliaderis89 also reported long-chained materials and Matheson43 found that both ECL and ICL were larger than in the normal amylopectin of smooth pea starch. The different opinions regarding the nature of IM probably reflect differences among varieties of wrinkled peas.

2.6

Analysis of chemically modified starches

Chemically substituted starches are of great importance in many industrial applications. The chemical modification is mostly made directly on the granular starch, which is in a convenient form to be collected and handled after the reaction. Therefore, the granular structure influences greatly the substitution pattern of the starch components.230 The chemical reagents penetrate with different efficiencies through the surface or via channels in the granules into the interior.231,232 Generally, the amorphous parts of the granules are substituted more easily than the crystalline areas,233 though granules from different plants can be affected in various ways depending on their architecture, type of crystallinity, amylose content, and the presence of minor components. Amylose tends to become more strongly substituted than amylopectin,234±237 which is preferentially substituted at the branched regions.159,234,237 Common substituents introduced to starch are hydroxypropyl, hydroxyethyl, methyl, and acetyl groups.238 Oxidised starches contain carboxyl and carbonyl groups238±241 and become largely depolymerised as a result of the chemical reactions.29, 239±242 It is of prime importance to determine the amount of substituents because it affects the functional properties of the starch. The degree of substitution (DS) is the average number of hydroxyl groups on the D-glucosyl units that have been substituted.243 DS range from 0±3. If DS is three all possible hydroxyl groups are substituted on each residue, and if DS is less than one, the average number of substituents per D-glucosyl residue is less than one. The molar substitution (MS) is the average number of moles of the substituent per Dglucosyl residue.243 Commonly, MS equals DS, but if more than one substituent can be attached at the same position on the residues, MS can exceed three. DS is often determined chemically, the method being dependent on the type of substituent. NMR is an alternative method and was used for oxidised,244±246

ß 2004, Woodhead Publishing Limited

acetylated,247,248 hydroxypropylated,248±250 and cationised samples.251 Besides DS, NMR offers the possibility to analyse the substitution pattern on the Dglucosyl residues. This was also possible with gas-liquid chromatography mass spectrometry (GLC-MS) on methylated,234 hydroxypropylated250 and cationic starches.252,253 C-2 is the preferential site to become substituted by several types of chemical groups. In addition to the substitution pattern on the D-glucosyl residues, the positions of the substituted residues within the polymers, and their distribution between the starch components, is of interest. For this purpose enzymic methods have been used. The introduced substituents act as barriers to hydrolytic attack159,254,255 in a similar way as the branches stop the exo-acting enzymes and restrict the action of endo-amylases. Substitution at C-2 was reported to inhibit -amylase attack more than C-6 substitutions.230,256 New types of limit dextrins are therefore obtained and give information about the substitution pattern. Steeneken and Woortman230 treated methylated potato starch consecutively with the -amylase from B. amyloliquefaciens and glucoamylase from A. niger. When derivatised in a granular form, larger amounts of glucose and highly substituted high molecular weight dextrins were obtained than from the non-granular counterpart with similar DS. They concluded that the granular form resulted in a heterogeneous distribution of the methyl groups within the starch polymers, whereas non-granular starch was homogeneously substituted. Similar results were obtained with cationised253 and hydroxypropylated starch.250 Kavitha and BeMiller237 investigated the amylose fraction of hydroxypropylated potato starch with the same combination of enzymes and found heterogeneous substitution within this component. Burght et al.257 used a combination of -amylase and -amylase to investigate the distribution pattern in methylated amylopectin close to the branch points. -Amylase can be used for the estimation of substituents along the external chains. Zhu and Bertoft258 used the enzyme on oxidised potato starches (DS 0.04±0.05) that were analysed by GPC before and after the hydrolysis. The limit value for hypochlorite oxidised starch was lower (44%) than for native starch (60%) and all polymers (that in oxidised starch are of low molecular weight) were obtained as -LDs. Hydrogen peroxide oxidised starch, however, had only slightly decreased -limit value and 24% of the sample was completely degraded.258 Thus, the method used for the oxidation influenced the distribution pattern of oxidised carbons. Kavitha and BeMiller237 obtained a similar decrease of the -limit value to 47% in hydroxypropylated potato starch and concluded that some substituents were found near the non-reducing ends. Richardson et al.250 reported an ECL of 10.2 in the -LD of hydroxypropylated amylose-free potato starch. The -limit value of a cross-linked hydroxypropylated manioc (tapioca) starch159 was 34%, and Hood and Mercier159 reported that the -limit value of the same sample decreased from 93% to 49%. Debranching enzymes are expected to be blocked in their action if a substituent is found at, or in the vicinity of, a branched D-glucosyl residue. Hood and Mercier159 obtained incomplete debranching of the hydroxypropylated

ß 2004, Woodhead Publishing Limited

manioc starch with pullulanase. Similar results were reported when hydroxypropylated237 and oxidised259 potato starches were debranched with isoamylase. Oxidised starches were more completely attacked by a successive treatment with isoamylase and pullulanase than with isoamylase only.258 After successive -amylolysis of the debranched samples, 20±26% of the apparent chains remained resistant to complete attack. Their CL values were, in fact, higher than for the sample before -amylolysis, suggesting that most of the oxidised carbons were found at the non-reducing ends.258 Similar results were found in cationised starches.260, 261 Oxidised potato starches, being anionic in character, were fractionated by ion-exchange chromatography on DEAE-Sepharose.245 A major part of the samples was bound to the ion-exchange matrix and had larger DP than unbound dextrins. Cationic waxy maize starch was debranched prior to ion-exchange chromatography on CM-Sepharose.251 The bound dextrins were eluted in a series of fractions with a sodium acetate gradient and then analysed both enzymatically and with proton-NMR. It was shown that the unbound fraction consisted of unsubstituted linear chains. The series of bound dextrins possessed increased DP, CL and degree of branching.251 Though the DS of the different fractions was fairly similar, the average number of substituents per dextrin molecule increased from 1.2 to 7.9. The distribution of substituents between amorphous and crystalline areas inside the starch granules was analysed by comparing the level of substitution before and after a mild acid hydrolysis. Burgt et al.233 showed that MS of the amorphous domains of methylated potato starch granules was about twice as high as in crystalline parts. This was thus in agreement with the heterogeneous substitution pattern suggested from -amylase and glucoamylase treatments. In oxidised potato starches the relative level of carboxyl groups decreased after acid hydrolysis to 64% of the original level, which also showed that the major site of modification was in the amorphous parts.260 In a cationised waxy maize starch, the level of substituents decreased to 40%.261 This was in contrast to a similar cationised potato sample, in which the level corresponded to 90% of the original before the acid treatment and indicated an almost even distribution inside the granules.260 It was suggested260 that the cationising reagent could penetrate deep into the crystals through the water-filled central cavities found in the B-type potato starch.262 In waxy maize starch granules, which are A-type, the cavity is filled with a double helix of the amylopectin external chains, making the crystals more compact.263,264

2.7

Future trends

During the last decades, research on starch structure has made substantial progress, especially regarding the architecture of starch granules. Part of this progress is the result of new microscopic techniques.184 Another part is due to progress in solid state NMR and crystallographic techniques, which made it

ß 2004, Woodhead Publishing Limited

possible to analyse the conformation of amylopectin inside the granules.265,266 Even sections of individual starch granules were examined.149 In this light, it would have been expected that the structures of the starch components were established earlier, especially since the knowledge that starch is composed of two major macromolecules is more than 60 years old.267 Though there is a broad consensus regarding the cluster model for the amylopectin, knowledge concerning the details of this structure is still limited. These details have to be resolved in order finally to understand the structure of the starch granules and the functionality of starch. Genetic modification of plants by biotechnology has made it possible to manipulate the biosynthesis of starch.268 Tailor-made starches for different industrial applications are predicted to become increasingly important in the reasonably near future. The development of these starches, and the understanding of the interaction of the enzymic activities involved in their synthesis, demands a range of suitable methods for the analysis of the structure of the altered macromolecules. Indeed, the alterations can only be understood if the structure of the unaltered, native starch also is fully investigated. Though the methods described in this chapter provide tools for a thorough structural analysis, they cannot give the complete answers to all details. For this improved methodology is needed. Existing methods were developed for basic research purposes; many are laborious and thereby time consuming. Thus, there is also an obvious need of improved technology suitable for routine analyses. Fortunately, progress in both methodology and technology is continuously made. For the investigation of the size and dimensions of starch components, flow field-flow fractionation coupled to multi-angle light scattering and refractive index detectors (FIFFF-MALS-RI) was tested.269 The method is based on the diffusion coefficient of the macromolecules and offers the advantage of obtaining the size distribution of virtually any substance without upper size limits,270 and without forcing them through a stationary network phase with the accompanying risk of partial degradation or low yields.101 The latter especially is a common problem in HPSEC of high molecular weight substances. Density-gradient ultracentrifugation was used to fractionate wheat,271 maize and potato starch.226 NMR is an established important tool for structural analyses of both native120,272±275 and modified starches.203,244,247± 250 Additional new methods are, however, also tested. Gas-liquid chromatography coupled to mass spectrometry (GLC-MS) was used for analysis of modified starches.234,250,252,253 While this method includes a derivatisation of the carbohydrates prior to analysis, liquid chromatography (LC) does not. HPAEC-PAD-MS was recently used for enzymatically degraded hydroxypropylated starch.276 Matrix-assisted laser desorption ionisation timeof-flight mass spectrometry (MALDI-TOF MS) was tested for the analysis of the unit chain profile of amylopectin.277 On-line microdialysis sampling could offer routine application of enzymatically treated samples to LC systems.278, 279 Thus, the technology will be improved in the near future if the demand, and market, is strong enough. The underlying methodology in the new techniques will,

ß 2004, Woodhead Publishing Limited

however, to a large extent remain dependent on the enzymic methods described in this chapter.

2.8

Sources of further information and advice

Some general principles for starch structure analysis were outlined above. It should be noted, however, that almost every starch sample possesses unique properties and therefore may require certain modifications of the general methods. This is probably one of the reasons for the great diversity of method modifications found between different laboratories. Often it starts with the choice of method for starch granule solubilisation. The reader is therefore strongly advised to consult the original sources referred to in the text before planning any experiments in detail. Regarding the use of enzymes in structural studies, it is of the greatest importance that the enzyme preparation is free from any disturbing activity that could influence the final results. Generally, only enzyme preparations intended for analytical purposes should be considered, whereas preparations used for industrial scale production of starch derived products may contain considerable levels of minor activities. In a few situations the disturbing influence of a sideactivity can be reduced, or eventually obliterated, by changing the analytical method for measuring the products. New books concerning the topic of this chapter have not been published recently. Starch, chemistry and technology, edited by Whistler, BeMiller and Paschall,280 was published in 1984. Another much referenced book from 1986 is Wurzburg's Modified starches: properties and uses.238 The sources, isolation, properties and applications of enzymes are described in Handbook of amylases and related enzymes22 and in Enzyme chemistry and molecular biology of amylases and related enzymes.281 Both volumes were edited by the Amylase Research Society of Japan. For further information about starch structure and analysis, the reviews by Manners,1 Gallant et al.,184 BuleÂon et al.,282 and Thompson153 are recommended. Several international journals frequently publish new findings in starch structural research, among others the reader might consider Carbohydrate Polymers, Carbohydrate Research, Starch/StaÈrke, International Journal of Biological Macromolecules, Journal of Chromatography A, Cereal Chemistry, and Journal of Cereal Science.

2.9

References

1.

MANNERS D J, `Recent developments in our understanding of amylopectin structure', Carbohydr Polym, 1989 11 87±112. 2. TAKEDA Y, MARUTA N AND HIZUKURI S, `Structures of amylose subfractions with different molecular sizes', Carbohydr Res, 1992 226 279±285. 3. FREDRIKSSON H, SILVERIO J, ANDERSSON R, ELIASSON A-C and AÊMAN P, `The influence

ß 2004, Woodhead Publishing Limited

4.

5. 6. 7. 8.

9. 10.

11.

12.

13.

14. 15. 16. 17. 18. 19. 20.

21.

22. 23.

of amylose and amylopectin characteristics on gelatinization and retrogradation properties of different starches', Carbohydr Polym, 1998 35 119±134. JANE J, CHEN Y Y, LEE L F, MCPHERSON A E, WONG K S, RADOSAVLJEVIC M and KASEMSUWAN T, `Effects of amylopectin branch chain length and amylose content on gelatinization and pasting properties of starch', Cereal Chem, 1999 76 629±637. MORRISON W R, MILLIGAN T P and AZUDIN M N, `Relationship between the amylose and lipid contents of starches from diploid cereals', J Cereal Sci, 1984 2 257±271. BANKS W, GREENWOOD C T and WALKER J T, `Studies on the starches of barley genotypes: the waxy starch', StaÈrke, 1970 22 149±152. GEÂRARD C, COLONNA P, BULEÂON A and PLANCHOT V, `Order in maize mutant starches revealed by mild acid hydrolysis', Carbohydr Polym, 2002 48 131±141. VASANTHAN T and BHATTY R S, `Physicochemical properties of small- and largegranule starches of waxy, regular, and high-amylose barleys', Cereal Chem, 1996 73 199±207. BANKS W, GREENWOOD C T and MUIR D D, `Studies on starches of high amylose content. Part 17. A review of current concepts', StaÈrke, 1974 26 289±328. SCHWALL G P, SAFFORD R, WESTCOTT R J, JEFFCOAT R, TAYAL A, SHI Y-C, GIDLEY M J

and JOBLING S A, `Production of very-high-amylose potato starch by inhibition of SBE A and B', Nature Biotech, 2000 18 551±554. JANE J-L, KASEMSUWAN T, LEAS S, ZOBEL H and ROBYT J F, `Anthology of starch granule morphology by scanning electron microscopy', Starch/StaÈrke, 1994 46 121±129. SHI Y-C, CAPITANI T, TRZASKO P and JEFFCOAT R, `Molecular structure of a lowamylopectin starch and other high-amylose maize starches', J Cereal Sci, 1998 27 289±299. ROBIN J P, MERCIER C, CHARBONNIEÁRE R and GUILBOT A, `Lintnerized starches. Gel filtration and enzymatic studies of insoluble residues from prolonged acid treatment of potato starch', Cereal Chem, 1974 51 389±406. MANNERS D J and MATHESON N K, `The fine structure of amylopectin', Carbohydr Res, 1981 90 99±110. FRENCH D, `Fine structure of starch and its relationship to the organization of starch granules', J Jpn Soc Starch Sci, 1972 19 8±25. NIKUNI Z, `Studies on starch granules', Sta È rke, 1978 30 105±111. HIZUKURI S, `Polymodal distribution of the chain lengths of amylopectins, and its significance', Carbohydr Res, 1986 147 342±347. PEAT S, WHELAN W J and THOMAS G J, `Evidence of multiple branching in waxy maize starch', J Chem Soc, Chem Commun, 1952 4546±4548. HIZUKURI S and MAEHARA Y, `Fine structure of wheat amylopectin: the mode of A to B chain binding', Carbohydr Res, 1990 206 145±159. BERTOFT E, `Investigation of the fine structure of alpha-dextrins derived from amylopectin and their relation to the structure of waxy-maize starch', Carbohydr Res, 1991 212 229±244. YUN S-H and MATHESON N K, `Structures of the amylopectins of waxy, normal, amylose-extender, and wx:ae genotypes and of the phytoglycogen of maize', Carbohydr Res, 1993 243 307±321. THE AMYLASE RESEARCH SOCIETY OF JAPAN, Handbook of Amylases and Related Enzymes, Oxford, Pergamon Press plc, 1988. SCHOCH T J, `Non-carbohydrate substances in cereal starches', J Am Chem Soc, 1942 64 2954±2956.

ß 2004, Woodhead Publishing Limited

24.

25.

26.

27.

28.

29.

30.

31.

32.

33. 34.

35. 36. 37. 38. 39. 40. 41. 42.

and JANE J, `Characterization of starch structures of 17 maize endosperm mutant genotypes with Oh43 inbred line background', Cereal Chem, 1993 70 171±179. KLUCINEC J D and THOMPSON D B, `Fractionation of high-amylose maize starches by differential alcohol precipitation and chromatography of the fractions', Cereal Chem, 1998 75 887±896. PATIL N B, SOMVANSHI B S, GUPTE S P and KALE N R, `Isolation and characterisation of high-molecular-weight amylose by dispersion of starch granules with urea', Makromol Chem, 1974 175 1979±1994. MORRISON W R and LAIGNELET B, `An improved colorimetric procedure for determining apparent and total amylose in cereal and other starches', J Cereal Sci, 1983 1 9±20. LEHTONEN P, `Effect of method of solution on observed molecular weight distribution of barley starch and starch derivatives in size exclusion chromatography', Chromatographia, 1988 26 157±159. È S H, `A gel chromatographic study of molecular weight BRUUN H and HENRIKSNA distribution in native wheat starch and its oxidation products', Starch/StaÈrke, 1977 29 122±126. CHEN Y, FRINGANT C and RINAUDO M, `Molecular characterization of starch by SEC: dependence of the performances on the amylopectin component', Carbohydr Polym, 1997 33 73±78. BLENNOW A, BAY-SMIDT A M and BAUER R, `Amylopectin aggregation as a function of starch phosphate content studied by size exclusion chromatography and on-line refractive index and light scattering', Int J Biol Macromol, 2001 28 409±420. MILLARD M M, DINTZIS F R, WILLETT J L and KLAVONS J A, `Light-scattering molecular weights and intrinsic viscosities of processed waxy maize starches in 90% dimethyl sulfoxide and H2O', Cereal Chem, 1997 74 687±691. SCHOCH T J, `Fractionation of starch by selective precipitation with butanol', J Am Chem Soc, 1942 64 2957±2961. ADKINS G K and GREENWOOD C T, `Studies on starches of high amylose-content. Part X. An improved method for the fractionation of maize and amylomaize starches by complex formation from aqueous dispersion after pretreatment with methyl sulphoxide', Carbohydr Res, 1969 11 217±224. JANE J-L and CHEN J-F, `Effect of amylose molecular size and amylopectin branch chain length on paste properties of starch', Cereal Chem, 1992 69 60±65. LANSKY S, KOOI M and SCHOCH T J, `Properties of the fractions and linear subfractions from various starches', J Am Chem Soc, 1949 71 4066±4075. HIZUKURI S, TAKEDA Y, YASUDA M and SUZUKI A, `Multi-branched nature of amylose and the action of de-branching enzymes', Carbohydr Res, 1981 94 205±213. TAKEDA Y, HIZUKURI S and JULIANO B O, `Purification and structure of amylose from rice starch', Carbohydr Res, 1986 148 299±308. TAKEDA Y and PREISS J, `Structures of B90 (sugary) and W64A (normal) maize starches', Carbohydr Res, 1993 240 265±275. TAKEDA Y, SHIRASAKA K and HIZUKURI S, `Examination of the purity and structure of amylose by gel-permeation chromatography', Carbohydr Res, 1984 132 83±92. BANKS W and GREENWOOD C T, `The fractionation of laboratory-isolated cereal starches using dimethyl sulphoxide', StaÈrke, 1967 19 394±398. MATHESON N K and WELSH L A, `Estimation and fractionation of the essentially unbranched (amylose) and branched (amylopectin) components of starches with WANG Y-J, WHITE P, POLLAK L

ß 2004, Woodhead Publishing Limited

43.

44. 45. 46.

47.

48.

49. 50.

51. 52.

53.

54. 55.

56. 57.

58. 59. 60. 61.

concanavalin A', Carbohydr Res, 1988 180 301±313. `A comparison of the structures of the fractions of normal and high-amylose pea-seed starches prepared by precipitation with concanavalin A', Carbohydr Res, 1990 199 195±205. GEÂRARD C, BARRON C, COLONNA P and PLANCHOT V, `Amylose determination in genetically modified starches', Carbohydr Polym, 2001 44 19±27. COLONNA P, BITON V and MERCIER C, `Interactions of concanavalin A with -Dglucans', Carbohydr Res, 1985 137 151±166. Ê MAN P, `Quantitative analysis of amylopectin unit KOCH K, ANDERSSON R and A chains by means of high-performance anion-exchange chromatography with pulsed amperometric detection', J Chromatogr A, 1998 800 199±206. LLOYD J R, HEDLEY C L, BULL V J and RING S G, `Determination of the effect of r and rb mutations on the structure of amylose and amylopectin in pea (Pisum sativum L.)', Carbohydr Polym, 1996 29 45±49. ROLLINGS J E, OKOS M R, TSAO G T and BOSE A, `System development for size exclusion chromatography of starch hydrolysates', J Appl Polym Sci, 1982 27 2281±2293. COLONNA P, BULEÂON A and LEMARIEÂ F, `Action of Bacillus subtilis -amylase on native wheat starch', Biotech Bioeng, 1988 31 895±904. BOYER C D and LIU K-C, `The interaction of endosperm genotype and genetic background. Part I. Differences in chromatographic profiles of starches from nonmutant and mutant endosperms', Starch/StaÈrke, 1985 37 73±79. HANES C S, `The action of amylases in relation to the structure of starch and its metabolism in the plant. Part IV-VII', New Phytologist, 1937 36 189±239. RUNDLE R E and FRENCH D, `The configuration of starch in the starch-iodine complex. III. X-ray diffraction studies of the starch-iodine complex', J Am Chem Soc, 1943 65 1707±1710. Â N I, LAETTIG S, WELFE H, SHELDRICK G M and SAENGER W, `XNIMZ O, GEûLER K, USO ray structure of the cyclomaltohexaicosaose triiodide inclusion complex provides a model for amylose-iodine at atomic resolution', Carbohydr Res, 2003 338 977± 986. ELMGREN H, `Helical conformation of amylose in aqueous solution. I. Viscosity measurements', Biopolymers, 1984 23 2525±2541. 13 JANE J-L, ROBYT J F and HUANG D-H, ` C-N.M.R. study of the conformation of helical complexes of amylodextrin and of amylose in solution', Carbohydr Res, 1985 140 21±35. BAILEY J M and WHELAN W J, `Physical properties of starch. I. Relationship between iodine stain and chain length', J Biol Chem, 1961 236 969±973. È RINEN P and HIZUKURI S, `Structural SCHULMAN A H, TOMOOKA S, SUZUKI A, MYLLA analysis of starch from normal and shx (shrunken endosperm) barley (Hordeum vulgare L.)', Carbohydr Res, 1995 275 361±369. SHIBANUMA K, TAKEDA Y, HIZUKURI S and SHIBATA S, `Molecular structures of some wheat starches', Carbohydr Polym, 1994 25 111±116. TAKEDA Y, HIZUKURI S and JULIANO B O, `Structures of rice amylopectins with low and high affinities for iodine', Carbohydr Res, 1987 168 79±88. TAKEDA Y, TAKEDA C, MIZUKAMI H and HANASHIRO I, `Structures of large, medium and small starch granules of barley grain', Carbohydr Polym, 1999 38 109±114. GILBERT G A AND SPRAGG S P, `Iodimetric determination of amylose. Iodine sorption: ``Blue Value'' ', Methods Carbohydr Chem, 1964 4 168±169. MATHESON N K,

ß 2004, Woodhead Publishing Limited

62.

63. 64. 65.

66. 67.

68.

69. 70. 71.

72.

73.

74.

75. 76. 77. 78. 79.

80. 81.

and GORTON L, `Determination of the degree of branching in normal and amylopectin type potato starch with 1H-NMR spectroscopy. Improved resolution and two-dimensional spectroscopy', Starch/ StaÈrke, 1996 48 352±357. SCHOCH T J, `Iodimetric determination of amylose', Methods Carbohydr Chem, 1964 4 157±160. BABA T and ARAI Y, `Structural characterization of amylopectin and intermediate material in amylomaize starch granules', Agric Biol Chem, 1984 48 1763±1775. MONTGOMERY E M, SEXSON K R, DIMLER R J and SENTI F R, `Physical properties and chemical structure of high-amylose corn starch fractions', StaÈrke, 1964 16 345± 351. TAKEDA C, TAKEDA Y and HIZUKURI S, `Structure of the amylopectin fraction of amylomaize', Carbohydr Res, 1993 246 273±281. BRADBURY A G W and BELLO A B, `Determination of molecular size distribution of starch and debranched starch by a single procedure using high-performance sizeexclusion chromatography', Cereal Chem, 1993 70 543±547. Ê MAN P, ANDERSSON L, FREDRIKSSON H, OSCARSSON-BERGH M, ANDERSSON R and A `Characterisation of starch from inner and peripheral parts of normal and waxy barley kernels', J Cereal Sci, 1999 30 165±171. MORRISON W R, LAW R V and SNAPE C E, `Evidence for inclusion complexes of lipids with V-amylose in maize, rice and oat starches', J Cereal Sci, 1993 18 107±109. KNUTSON C A and GROVE M J, `Rapid method for estimation of amylose in maize starches', Cereal Chem, 1994 71 469±471. MOHAMMADKHANI A, STODDARD F L and MARSHALL D R, `Survey of amylose content in Secale cereale, Triticum monococcum, T. turgidum and T. tauschii', J Cereal Sci, 1998 28 273±280. CAMPBELL M R, YEAGER H, ABDUBEK N, POLLAK L M and GLOVER D V, `Comparison of methods for amylose screening among amylose-extender (ae) maize starches from exotic backgrounds', Cereal Chem, 2002 79 317±321. BOYER C D, DAMEWOOD P A and MATTERS G L, `Effect of gene dosage at high amylose loci on properties of the amylopectin fractions of the starches', StaÈrke, 1980 32 217±222. KOBAYASHI S, SCHWARTZ S J and LINEBACK D R, `Rapid analysis of starch, amylose and amylopectin by high-performance size-exclusion chromatography', J Chromatogr, 1985 319 205±214. MUA J-P and JACKSON D S, `Gelatinization and solubility properties of commercial oat starch', Starch/StaÈrke, 1995 47 2±7. KRISMAN C R, `A method for the colorimetric estimation of glycogen with iodine', Anal Biochem, 1962 4 17±23. 2+ MANELIUS R and BERTOFT E, `The effect of Ca -ions on the -amylolysis of granular starches from oats and waxy-maize', J Cereal Sci, 1996 24 139±150. SUORTTI T and PESSA E, `Gel permeation chromatrographic determination of starches using alkaline eluents', J Chromatogr A, 1991 536 251±254. PESSA E, SUORTTI T, AUTIO K and POUTANEN K, `Molecular weight characterization and gelling properties of acid-modified maize starches', Starch/StaÈrke, 1992 44 64±69. AUTIO K, POUTANEN K, SUORTTI T and PESSA E, `Heat-induced structural changes in acid-modified barley starch dispersions', Food Structure, 1992 11 315±322. SARGEANT J G, `Determination of amylose:amylopectin ratios of starches', Sta È rke, NILSSON G S, BERGQUIST K-E, NILSSON U

ß 2004, Woodhead Publishing Limited

1982 34 89±92. 82.

83. 84. 85.

86.

87. 88. 89.

90. 91. 92.

93. 94.

95. 96. 97. 98. 99. 100. 101.

102.

and SUNDBERG B, `Amylose content and chain profile of amylopectin from normal, high amylose and waxy barleys', Starch/StaÈrke, 1994 46 325±328. WANG L Z and WHITE P J, `Structure and physicochemical properties of starches from oats with different lipid contents', Cereal Chem, 1994 71 443±450. KARKALAS J AND TESTER R F, `Continuous enzymic determination of -glucans in eluates from gel-chromatographic columns', J Cereal Sci, 1992 15 175±180. HIZUKURI S and TAKAGI T, `Estimation of the distribution of molecular weight for amylose by the low-angle laser-light-scattering technique combined with highperformance gel chromatography', Carbohydr Res, 1984 134 1±10. COWIE J M G and GREENWOOD C T, `Physicochemical studies on starches. Part VIII. Further observations on the fractionation of potato starch', J Chem Soc, Chem Commun, 1957 4640±4644. PARK J T and JOHNSON M J, `A submicrodetermination of glucose', J Biol Chem, 1949 181 149±151. HIZUKURI S, SHIRASAKA K and JULIANO B O, `Phosphorus and amylose branching in rice starch granules', Starch/StaÈrke, 1983 35 348±350. BILIADERIS C G, GRANT D R and VOSE J R, `Structural characterization of legume starches. I. Studies on amylose, amylopectin, and beta-limit dextrins', Cereal Chem, 1981 58 496±502. TAKEDA Y, SHITAOZONO T and HIZUKURI S, `Molecular structure of corn starch', Starch/StaÈrke, 1988 40 51±54. LII C-Y and CHANG Y-H, `Study of starch in Taiwan', Food Rev Int, 1991 7 185±203. TAKAGI T and HIZUKURI S, `Molecular weight and related properties of lily amylose determined by monitoring of elution from TSK-GEL PW high performance gel chromatography columns by the low-angle laser light scattering technique and precision differential refractometry', J Biochem, 1984 95 1459±1467. MANNERS D J and BATHGATE G N, `-1,4±Glucans. Part XX. The molecular structure of the starches from oats and malted oats', J Inst Brew, London, 1969 75 169±175. TAKEDA Y, HIZUKURI S, TAKEDA C and SUZUKI A, `Structures of branched molecules of amyloses of various origins, and molecular fractions of branched and unbranched molecules', Carbohydr Res, 1987 165 139±145. TAKEDA Y, TOMOOKA S and HIZUKURI S, `Structures of branched and linear molecules of rice amylose', Carbohydr Res, 1993 246 267±272. TAKEDA Y, MARUTA N and HIZUKURI S, `Examination of the structure of amylose by tritium labelling of the reducing terminal', Carbohydr Res, 1992 227 113±120. TAKEDA Y, HIZUKURI S and JULIANO B O, `Structures and amounts of branched molecules in rice amyloses', Carbohydr Res, 1989 186 163±166. TAKEDA Y, SHITAOZONO T and HIZUKURI S, `Structures of sub-fractions of corn amylose', Carbohydr Res, 1990 199 207±214. CALLAGHAN P T and LELIEVRE J, `The size and shape of amylopectin: A study using pulsed-field gradient nuclear magnetic resonance', Biopolymers, 1985 24 441±460. ERLANDER S R and TOBIN R, `Effect of phosphate buffer on dispersed starch. II. Extrapolated molecular weight of amylopectin', Starch/StaÈrke, 1967 19 94±99. YOKOYAMA W, RENNER-NANTZ J J and SHOEMAKER C F, `Starch molecular mass and size by size-exclusion chromatography in DMSO-LiBr coupled with multiple angle laser light scattering', Cereal Chem, 1998 75 530±535. ABERLE T, BURCHARD W, VORWERG W and RADOSTA S, `Conformational contributions SALOMONSSON A-C

ß 2004, Woodhead Publishing Limited

103. 104. 105. 106.

107. 108.

109.

110. 111.

112.

113. 114.

115. 116. 117. 118. 119. 120. 121. 122.

of amylose and amylopectin to the structural properties of starches from various sources', Starch/StaÈrke, 1994 46 329±335. THURN A and BURCHARD W, `Heterogeneity in branching of amylopectin', Carbohydr Polym, 1985 5 441±460. LELIEVRE J, LEWIS J A and MARSDEN K, `The size and shape of amylopectin: a study using analytical ultracentrifugation', Carbohydr Res, 1986 153 195±203. FRANCO C M L, WONG K-S, YOO S-H and JANE J-L, `Structural and functional characteristics of selected soft wheat starches', Cereal Chem, 2002 79 243±248. WONG K-S, KUBO A, JANE J-L, HARADA K, SATOH H and NAKAMURA Y, `Structure and properties of amylopectin and phytoglycogen in the endosperm of sugary-1 mutants of rice', J Cereal Sci, 2002 37 139±149. STACY C J and FOSTER J F, `Molecular weight heterogeneity in starch amylopectins', J Polym Sci, Part A, 1957 25 39±50. POTTER A L, SILVEIRA V, MCCREADY R M and OWENS H S, `Fractionation of starches from smooth and wrinkled seeded peas. Molecular weights, end-group assays and iodine affinities of the fractions', J Am Chem Soc, 1953 75 1335±1338. TAKEDA Y, MARUTA N, HIZUKURI S and JULIANO B O, `Structures of indica rice starches (IR48 and IR64) having intermediate affinities for iodine', Carbohydr Res, 1989 187 287±294. HIZUKURI S, TAKEDA Y, MARUTA N and JULIANO B O, `Molecular structure of rice starch', Carbohydr Res, 1989 189 227±235. HIZUKURI S, TAKEDA Y, SHITAOZONO T, ABE J, OHTAKARA A, TAKEDA C and SUZUKI A, `Structure and properties of water chestnut (Trapa natans L. var. bispinosa Makino) starch', StaÈrke, 1988 40 165±171. TAKEDA Y, SHIBAHARA S and HANASHIRO I, `Examination of the structure of amylopectin molecules by fluorescent labeling', Carbohydr Res, 2003 338 471± 475. LINEBACK D R and SAYEED A L, `Effect of dimethyl sulfoxide on amylase activity', Carbohydr Res, 1971 17 453±456. HIZUKURI S and OSAKI S, `A rapid Smith-degradation for the determination of nonreducing, terminal residues of (1!4)--D-glucans', Carbohydr Res, 1978 63 261± 264. NELSON N, `A photometric adaptation of the Somogyi method for the determination of glucose', J Biol Chem, 1944 153 375±380. TAKEDA Y, GUAN H-P and PREISS J, `Branching of amylose by the branching isoenzymes of maize endosperm', Carbohydr Res, 1993 240 253±263. WAFFENSCHMIDT S and JAENICKE L, `Assay of reducing sugars in the nanomole range with 2,20 -bicinchoninate', Anal Biochem, 1987 165 337±340. MCFEETERS R F, `A manual method for reducing sugar determinations with 2,20 bicinchoninate reagent', Anal Biochem, 1980 103 302±306. FOX J D and ROBYT J F, `Miniaturization of three carbohydrate analyses using a microsample plate reader', Anal Biochem, 1991 195 93±96. GIDLEY M J, `Quantification of the structural features of starch polysaccharides by n.m.r. spectroscopy', Carbohydr Res, 1985 139 85±93. KAINUMA K, KOBAYASHI S and HARADA T, `Action of Pseudomonas isoamylase on various branched oligo- and polysaccharides', Carbohydr Res, 1978 61 345±357. HARADA T, MISAKI A, AKAI H, YOKOBAYASHI K and SUGIMOTO K, `Characterization of Pseudomonas isoamylase by its actions on amylopectin and glycogen: Comparison with Aerobacter pullulanase', Biochim Biophys Acta, 1972 268 497±505.

ß 2004, Woodhead Publishing Limited

123.

124. 125. 126.

127.

128. 129.

130. 131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

and HARADA T, `Comparison of the kinetic parameters of Pseudomonas isoamylase and Aerobacter pullulanase', Biochim Biophys Acta, 1973 293 197±202. YOKOBAYASHI K, MISAKI A and HARADA T, `Purification and properties of Pseudomonas isoamylase', Biochim Biophys Acta, 1970 212 458±469. È RSCH P and HOOD L F, `Structure of starch from mango seed', Sta WU È rke, 1981 33 217±221. AKAI H, YOKOBAYASHI K, MISAKI A and HARADA T, `Structural analysis of amylopectin using Pseudomonas isoamylase', Biochim Biophys Acta, 1971 252 427±431. MACGREGOR A W and MORGAN J E, `Structure of amylopectins isolated from large and small starch granules of normal and waxy barley', Cereal Chem, 1984 61 222± 228. WONG K S and JANE J, `Quantitative analysis of debranched amylopectin with a postcolumn enzyme reactor', J Liq Chromatogr, 1997 20 397±310. INOUCHI N, GLOVER D V, TAKAYA T and FUWA H, `Development changes in fine structure of starches of several endosperm mutants of maize', StaÈrke, 1983 35 371± 376. BERTOFT E, `Chains of intermediate lengths in waxy-maize amylopectin', Carbohydr Res, 1991 212 245±251. REDDY K R, ALI S Z and BHATTACHARYA K R, `The fine structure of rice-starch amylopectin and its relation to the texture of cooked rice', Carbohydr Polym, 1993 22 267±275. YUSUPH M, TESTER R F, ANSELL R and SNAPE C E, `Composition and properties of starches extracted from tubers of different potato varieties grown under the same environmental conditions', Food Chem, 2003 82 283±289. KLUCINEC J D and THOMPSON D B, `Amylopectin nature and amylose-to-amylopectin ratio as influences on the behavior of gels of dispersed starch', Cereal Chem, 2002 79 24±35. HANASHIRO I, TAGAWA M, SHIBAHARA S, IWATA K and TAKEDA Y, `Examination of molar-based distribution of A, B and C chains of amylopectin by fluorescent labeling with 2±aminopyridine', Carbohyd Res, 2002 337 1211±1215. HIZUKURI S, `Relationship between the distribution of the chain length of amylopectin and the crystalline structure of starch granules', Carbohydr Res, 1985 141 295±306. ONG M H, JUMEL K, TOKARCZUK P F, BLANSHARD J M V and HARDING S E, `Simultaneous determinations of the molecular weight distributions of amyloses and the fine structures of amylopectins of native starches', Carbohydr Res, 1994 260 99±117. KOIZUMI K, FUKUDA M and HIZUKURI S, `Estimation of the distributions of chain length of amylopectins by high-performance liquid chromatography with pulsed amperometric detection', J Chromatogr, 1991 585 233±238. WONG K S and JANE J, `Effects of pushing agents on the separation and detection of debranched amylopectin by high-performance anion-exchange chromatography with pulsed amperometric detection', J Liquid Chromatogr, 1995 18 63±80. LU T-J, LACOURSE W R and JANE J-L, `Evaluating sodium salts as pushing agents on high-performance anion-exchange chromatography with pulsed amperometric detection for maltodextrin analysis', Starch/StaÈrke, 1997 49 505±511. NODA T, TAKAHATA Y, SATO T, SUDA I, MORISHITA T, ISHIGURO K and YAMAKAWA O, YOKOBAYASHI K, AKAI H, SUGIMOTO T, HIRAO M, SUGIMOTO K

ß 2004, Woodhead Publishing Limited

Analysing starch structure

141. 142.

143. 144.

145.

146.

147.

148.

149.

150.

151.

152.

153. 154. 155. 156. 157.

158.

89

`Relationships between chain length distribution of amylopectin and gelatinization properties within the same botanical origin for sweet potato and buckwheat', Carbohydr Polym, 1998 37 153±158. MCPHERSON A E and JANE J, `Comparison of waxy potato with other root and tuber starches', Carbohydr Polym, 1999 40 57±70. Ê MAN P, `The effect of SILVERIO J, FREDRIKSSON H, ANDERSSON R, ELIASSON A-C and A temperature cycling on the amylopectin retrogradation of starches with different amylopectin unit-chain length distribution', Carbohydr Polym, 2000 42 175±184. Ê RD T, ANDERSSON R and A Ê HMAN P, `Gradual enzymatic modification of barley FRIGA and potato amylopectin', Carbohydr Polym, 2002 47 169±179. SHI Y-C and SEIB P A, `Fine structure of maize starches from four wx-containing genotypes of the W64A inbred line in relation to gelatinization and retrogradation', Carbohydr Polym, 1995 26 141±147. HANASHIRO I, ABE J-I and HIZUKURI S, `A periodic distribution of chain length of amylopectin as revealed by high-performance anion-exchange chromatography', Carbohydr Res, 1996 283 151±159. YASUI T, MATSUKI J, SASAKI T and YAMAMORI M, `Amylose and lipid contents, amylopectin structure, and gelatinisation properties of waxy wheat (Triticum aestivum) starch', J Cereal Sci, 1996 24 131±137. AMMERAAL R N, DELGADO G A, TENBARGE F L and FRIEDMAN R B, `High-performance anion-exchange chromatography with pulsed amperometric detection of linear and branched glucose oligosaccharides', Carbohydr Res, 1991 215 179±192. CAIRNS P, BOGRACHEVA T Y, RING S G, HEDLEY C L and MORRIS V J, `Determination of the polymorphic composition of smooth pea starch', Carbohydr Polym, 1997 32 275±282. BULEÂON A, GEÂRARD C, RIEKEL C, VUONG R and CHANZY H, `Details of the crystalline ultrastructure of C-starch granules revealed by synchrotron microfocus mapping', Macromolecules, 1998 31 6605±6610. MORELL M K, SAMUEL M S and O'SHEA M G, `Analysis of starch structure using fluorophore-assisted carbohydrate electrophoresis', Electrophoresis, 1998 19 2603±2611. O'SHEA M G and MORELL M K, `High resolution slab gel electrophoresis of 8±amino1,3,6±pyrenetrisulfonic acid (APTS) tagged oligosaccharides using a DNA sequencer', Electrophoresis, 1996 17 681±688. NAKAMURA Y, SAKURAI A, INABA Y, KIMURA K, IWASAWA N and NAGAMINE T, `The fine structure of amylopectin in endosperm from Asian cultivated rice can be largely classified into two classes', Starch/StaÈrke, 2002 54 117±131. THOMPSON D B, `On the non-random nature of amylopectin branching', Carbohydr Polym, 2000 43 223±239. SUMMER R and FRENCH D, `Action of -amylase on branched oligosaccharides', J Biol Chem, 1956 222 469±477. LEE E Y C, `The action of sweet potato -amylase on glycogen and amylopectin: formation of a novel limit dextrin', Arch Biochem Biophys, 1971 146 488±492. LII C-Y AND LINEBACK D R, `Characterization and comparison of cereal starches', Cereal Chem, 1977 54 138±149. QI X, TESTER R F, SNAPE C E and ANSELL R, `Molecular basis of the gelatinisation and swelling characteristics of waxy rice starches grown in the same location during the same season', J Cereal Sci, 2003 37 363±376. MANNERS D J and MARSHALL J J, `Studies on carbohydrate metabolising enzymes.

ß 2004, Woodhead Publishing Limited

159. 160. 161.

162.

163. 164.

165. 166. 167. 168. 169. 170. 171.

172. 173. 174. 175. 176.

177.

178.

Part XXVIII. Preparation and some properties of -amylase isoenzymes from malted rye', StaÈrke, 1972 24 3±8. HOOD L F and MERCIER C, `Molecular structure of unmodified and chemically modified manioc starches', Carbohydr Res, 1978 61 53±66. BERTOFT E, `Investigation of the fine structure of amylopectin using alpha- and beta-amylase', Carbohydr Res, 1989 189 195±207. È KE A, `Can cyclodextrin glycosyltransferase BENDER H, SIEBERT R and STADLER-SZO be useful for the investigation of the fine structure of amylopectins? Characterisation of highly branched clusters isolated from digests with potato and maize starches', Carbohydr Res, 1982 110 245±259. LUNDQVIST H, NILSSON G S, ELIASSON A-C and GORTON L, `Changing the amylopectin-sodium dodecyl sulphate interaction by modifying the exterior chain length', Starch/StaÈrke, 2002 54 100±107. KLUCINEC J D and THOMPSON D B, `Structure of amylopectins from ae-containing maize starches', Cereal Chem, 2002 79 19±23. GEÂRARD C, PLANCHOT V, COLONNA P and BERTOFT E, `Relationship between branching density and crystalline structure of A- and B-type maize mutant starches', Carbohydr Res, 2000 326 130±144. BANKS W and GREENWOOD C T, `Molecular properties of the starch components and their relation to the structure of the granule', Ann N Y Acad Sci, 1973 210 17±33. BILIADERIS C G, GRANT D R and VOSE J R, `Molecular weight distributions of legume starches by gel chromatography', Cereal Chem, 1979 56 475±480. ENEVOLDSEN B S and JULIANO B O, `Ratio of A chains to B chains in rice amylopectins', Cereal Chem, 1988 65 424±427. BERTOFT E, `Partial characterization of amylopectin alpha-dextrins', Carbohydr Res, 1989 189 181±193. HESTRIN S, `Action pattern of crystalline muscle phosphorylase', J Biol Chem, 1949 179 943±955. LIDDLE A M and MANNERS D J, `-1:4±Glucosans. Part VIII. Multiple-branching in glycogen and amylopectin', J Chem Soc, Chem Commun, 1957 4708±4711. WALKER G J and WHELAN W J, `The mechanism of carbohydrase action. 8. Structures of the muscle-phosphorylase limit dextrins of glycogen and amylopectin', Biochem J, 1960 76 264±268. BERTOFT E, ZHU Q, ANDTFOLK H and JUNGNER M, `Structural heterogeneity in waxyrice starch', Carbohydr Polym, 1999 38 349±359. ZHU Q and BERTOFT E, `Composition and structural analysis of alpha-dextrins from potato amylopectin', Carbohydr Res, 1996 288 155±174. BERTOFT E and KOCH K, `Composition of chains in waxy-rice starch and its structural units', Carbohydr Polym, 2000 41 121±132. ATWELL W A, HOSENEY R C AND LINEBACK D R, `Debranching of wheat amylopectin', Cereal Chem, 1980 57 12±16. ASAOKA M, OKUNO K and FUWA H, `Effect of environmental temperature at the milky stage on amylose content and fine structure of amylopectin of waxy and nonwaxy endosperm starches of rice (Oryza sativa L.)', Agric Biol Chem, 1985 49 373±379. INOUCHI N, GLOVER D V and FUWA H, `Chain length distribution of amylopectins of several single mutants and the normal counterpart, and sugary-1 phytoglycogen in maize (Zea mays L.)', StaÈrke, 1987 39 259±266. MERCIER C, `The fine structure of corn starches of various amylose-percentage:

ß 2004, Woodhead Publishing Limited

179. 180.

181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191.

192.

193.

194.

195.

196.

197.

waxy, normal and amylomaize', StaÈrke, 1973 25 78±83. ROBIN J P, `Study of -limit dextrins from various native starches. Interpretation in term of amylopectin structure', Sci Aliments, 1981 1 551±567. YUAN R C, THOMPSON D B and BOYER C D, `Fine structure of amylopectin in relation to gelatinization and retrogradation behavior of maize starches from three wxcontaining genotypes in two inbred lines', Cereal Chem, 1993 70 81±89. BERTOFT E, `The structure of clusters from potato amylopectin', Zywnosc/Food, 2002 9 112±123. HIZUKURI S, `Starch: analytical aspects', in Eliasson A-C, Carbohydrates in Food, New York, Marcel Dekker, 1996, 347±429. IMBERTY A and PEÂREZ S, `Conformational analysis and molecular modelling of the branching point of amylopectin', Int J Biol Macromol, 1989 11 177±185. GALLANT D J, BOUCHET B and BALDWIN P M, `Microscopy of starch: evidence of a new level of granule organization', Carbohydr Polym, 1997 32 177±191. FINCH P and SEBESTA D W, `The amylase of Pseudomonas stutzeri as a probe of the structure of amylopectin', Carbohydr Res, 1992 227 c1±c4. BENDER H, `Enzymology of the cyclodextrins', in I. Int. Symp. on Cyclodextrins, Budapest, Dordrecht, Reidel, 1982, 77±87. ROBYT J and FRENCH D, `Action pattern and specificity of an amylase from Bacillus subtilis', Arch Biochem Biophys, 1963 100 451±467. THOMA J A, BROTHERS C and SPRADLIN J, `Subsite mapping of enzymes. Studies on Bacillus subtilis amylase', Biochemistry, 1970 9 1768±1775. BERTOFT E, `Hydrolysis of amylopectin by the alpha-amylase of B. subtilis', Carbohydr Res, 1986 149 379±387. BERTOFT E and SPOOF L, `Fractional precipitation of amylopectin alpha-dextrins using methanol', Carbohydr Res, 1989 189 169±180. JODELET A, RIGBY N M and COLQUHOUN I J, `Separation and NMR structural characterisation of singly branched -dextrins which differ in the location of the branch point', Carbohydr Res, 1998 312 139±151. GELDERS G G, BIJNENS L, LOOSVELD A-M, VIDTS A and DELCOUR J A, `Fractionation of starch hydrolysates into dextrins with narrow molecular mass distribution and their detection by high-performance anion-exchange chromatography with pulsed amperometric detection', J Chromatogr A, 2003 992 75±83. WHITE D R, HUDSON P and ADAMSON J T, `Dextrin characterization by highperformance anion-exchange chromatographypulsed amperometric detection and size-exclusion chromatography-multi-angle light scattering-refractive index detection', J Chromatogr A, 2003 997 79±85. BLENNOW A, BAY-SMIDT A M, WISCHMANN B, OLSEN C E and LINDBERG-MéLLER B, `The degree of starch phosphorylation is related to the chain length distribution of the neutral and the phosphorylated chains of amylopectin', Carbohydr Res, 1998 307 45±54. LORBERTH R, RITTE G, WILLMITZER L and KOSSMANN J, `Inhibition of a starchgranule-bound protein leads to modified starch and repression of cold sweetening', Nature Biotech, 1998 16 473±477. RITTE G, LLOYD J R, ECKERMANN N, ROTTMANN A, KOSSMANN J and STEUP M, `The starch-related R1 protein is an -glucan, water dikinase', Proc Nat Acad Sci, 2002 99 7166±7171. BAY-SMIDT A M, WISCHMANN B, OLSEN C E and NIELSEN T H, `Starch bound phosphate in potato as studied by a simple method for determination of organic phosphate and

ß 2004, Woodhead Publishing Limited

31

198. 199.

200. 201. 202.

203. 204.

205. 206. 207.

208. 209.

210.

211. 212.

213. 214. 215.

216.

P-NMR', Starch/StaÈrke, 1994 46 167±172. and TELLIER C, `Determination of the phosphorylation of starch from native potato varieties by 31P NMR', Starch/StaÈrke, 1991 43 25±27. VISSER R G F, SUURS L C J M, BRUINENBERG P M, BLEEKER I and JACOBSEN E, `Comparison between amylose-free and amylose containing potato starches', Starch/StaÈrke, 1997 49 438±443. MORRISON W R, `Starch lipids and how they relate to starch granule structure and functionality', Cereal Foods World, 1995 40 437±446. SHAMEKH S, FORSSELL P, SUORTTI T, AUTIO K and POUTANEN K, `Fragmentation of oat and barley starch granules during heating', J Cereal Sci, 1999 30 173±182. BLENNOW A, BAY-SMIDT A M, OLSEN C E and MéLLER B L, `The distribution of covalently bound phosphate in the starch granule in relation to starch crystallinity', Int J Biol Macromol, 2000 27 211±218. LIM S and SEIB P A, `Location of phosphate esters in a wheat starch phosphate by 31 P-nuclear magnetic resonance spectroscopy', Cereal Chem, 1993 70 145±152. KASEMSUWAN T and JANE J, `Location of amylose in normal starch granules. II. Locations of phosphodiester cross-linking revealed by phosphorus-31 nuclear magnetic resonance', Cereal Chem, 1994 71 282±287. LANZETTA P A, ALVAREZ L J, REINACH P S and CANDIA O A, `An improved assay for nanomole amounts of inorganic phosphate', Anal Biochem, 1979 100 95±97. BERGTHALLER W, `Untersuchungen u È ber die Herstellung von PhosphorsaÈureestern der StaÈrke', Starch/StaÈrke, 1971 23 73±79. HIZUKURI S, TABATA S and NIKUNI Z, `Studies on starch phosphate. Part 1. Estimation of glucose-6-phosphate residues in starch and the presence of other bound phosphate(s)', StaÈrke, 1970 22 338±343. MICHAL G, `D-Glucose 6-Phosphate and D-Fructose 6-Phosphate', Methods Enzyme Anal, 1984 6 191±198. BLENNOW A, BAY-SMIDT A M, OLSEN C E and LINDBERG-MéLLER B, `Analysis of starch-bound glucose 3-phosphate and glucose 6-phosphate using controlled acid treatment combined with high-performance anion-exchange chromatography', J Chromatogr A, 1998 829 385±391. TAKEDA Y and HIZUKURI S, `Actions of Aspergillus oryzae alpha-amylase, potato phosphorylase, and rabbit muscle phosphorylase a and b on phosphorylated (1!4)--D-glucan', Carbohydr Res, 1986 153 295±307. ABE J-I, TAKEDA Y and HIZUKURI S, `Action of glucoamylase from Aspergillus niger on phosphorylated substrate', Biochim Biophys Acta, 1982 703 26±33. TAKEDA Y and HIZUKURI S, `Re-examination of the action of sweet-potato betaamylase on phosphorylated (1!4)--D-glucan', Carbohydr Res, 1981 89 174± 178. TAKEDA Y and HIZUKURI S, `Location of phosphate groups in potato amylopectin', Carbohydr Res, 1982 102 321±327. WISCHMANN B and ADLER-NISSEN J, `Melting properties and lintnerisation of potato starch with different degrees of phosphorylation', Starch/StaÈrke, 2002 54 499±507. MUHRBECK P, SVENSSON E and ELIASSON A-C , `Effect of the degree of phosphorylation on the crystallinity of native potato starch', Starch/StaÈrke, 1991 43 466±468. BLENNOW A, ENGELSEN S B, MUNCK L and MéLLER B L, `Starch molecular structure and phosphorylation investigated by a combined chromatigraphic and chemometric approach', Carbohydr Polym, 2000 41 163±174. MUHRBECK P

ß 2004, Woodhead Publishing Limited

217. 218. 219.

220.

221.

222. 223. 224. 225. 226. 227. 228.

229.

230. 231. 232. 233.

and SHEN J J, `Internal structure of the potato starch granule revealed by chemical gelatinization', Carbohydr Res, 1993 247 279±290. JANE J-L, `Mechanism of starch gelatinization in neutral salt solutions', Starch/ StaÈrke, 1993 45 161±166. IKAWA Y, GLOVER D V, SUGIMOTO Y and FUWA H, `Amylose percentage and distribution of unit chain-length of maize starches having a specific genetic background', Carbohydr Res, 1978 61 211±216. PERERA C, LU Z, SELL J and JANE J, `Comparison of physicochemical properties and structures of sugary-2 cornstarch with normal and waxy cultivars', Cereal Chem, 2001 78 249±256. YUN S-H and MATHESON N K, `Structural changes during development in the amylose and amylopectin fractions (separated by precipitation with concanavalin A) of starches from maize genotypes', Carbohydr Res, 1992 227 85±101. COLONNA P and MERCIER C, `Macromolecular structure of wrinkled- and smoothpea starch components', Carbohydr Res, 1984 126 233±247. BERTOFT E, QIN Z and MANELIUS R, `Studies on the structure of pea starches. Part 4: Intermediate material of wrinkled pea starch', Starch/StaÈrke, 1993 45 420±425. PATON D, `Oat starch: Some recent developments', Starch/Sta È rke, 1979 31 184± 187. WANG L Z and WHITE P J, `Structure and properties of amylose, amylopectin, and intermediate material of oat starches', Cereal Chem, 1994 71 263±268. YOON J-W and LIM S-T, `Molecular fractionation of starch by density-gradient ultracentrifugation', Carbohydr Res, 2003 338 611±617. 13 DAIS P and PERLIN A S, `High-field, C-n.m.r. spectroscopy of -D-glucans, amylopectin, and glycogen', Carbohydr Res, 1982 100 103±116. WANG Y-J, WHITE P, POLLAK L and JANE J, `Amylopectin and intermediate materials in starches from mutant genotypes of the Oh43 inbred line', Cereal Chem, 1993 70 521±525. Ê VALL A-K, `Observations on the BERTOFT E, BOYER C, MANELIUS R and A amylolysis pattern of some waxy maize starches from the inbred line Ia453', Cereal Chem, 2000 77 657±664. STEENEKEN P A M and WOORTMAN A J J, `Substitution patterns in methylated starch as studied by enzymic degradation', Carbohydr Res, 1994 258 207±221. HUBER K C and BEMILLER J N, `Location of sites of reaction within starch granules', Cereal Chem, 2001 78 173±180. GRAY J A and BEMILLER J N, `Accessibility of starch granules to fatty acyl amides', Cereal Chem, 2001 78 236±242. JANE J-L

BURGT Y E M V D, BERGSMA J, BLEEKER I P, MIJLAND P J H C, HOOF A V D K-V,

and VLIEGENTHART J F G, `Distribution of methyl substituents over crystalline and amorphous domains in methylated starches', Carbohydr Res, 1999 320 100±107.

KAMERLING J P

234.

BURGT Y E M V D, BERGSMA J, BLEEKER I P, MIJLAND P J H C, HOOF A V D K-V,

and VLIEGENTHART J F G, `Distribution of methyl substituents over branched and linear regions in methylated starches', Carbohydr Res, 1998 312 201±208.

KAMERLING J P

235.

BURGT Y E M V D, BERGSMA J, BLEEKER I P, MIJLAND P J H C, HOOF A V D K-V,

and VLIEGENTHART J F G, `Distribution of methyl substituents in amylose and amylopectin from methylated potato starches', Carbohydr Res, 2000 325 183±191.

KAMERLING J P

ß 2004, Woodhead Publishing Limited

94 236. 237. 238. 239. 240. 241.

242.

243. 244.

245. 246.

247. 248.

249.

250.

251.

252.

253. 254.

255.

Starch in food and THEANDER O, `The cationic distribution in a cationised potato starch', Starch/StaÈrke, 1991 43 81±82. KAVITHA R and BEMILLER J N, `Characterization of hydroxypropylated potato starch', Carbohydr Polym, 1998 37 115±121. WURZBURG O B, Modified starches: properties and uses, Boca Raton, CRC Press, Inc., 1986. PAROVUORI P, HAMUNEN A, FORSSELL P, AUTIO K and POUTANEN K, `Oxidation of potato starch by hydrogen peroxide', Starch/StaÈrke, 1995 47 19±23. FORSSELL P, HAMUNEN A, AUTIO K, SUORTTI T and POUTANEN K, `Hypochlorite oxidation of barley and potato starch', Starch/StaÈrke, 1995 47 371±377. AUTIO K, SUORTTI T, HAMUNEN A and POUTANEN K, `Microstructural and physicochemical properties of oxidized potato starch for paper coating', Starch/ StaÈrke, 1992 44 393±398. È S H and BRUUN H, `Molecular weight distribution in starch solutions HENRIKSNA when hydrolised with alpha-amylase and when oxidised with sodium hypochlorite', Starch/StaÈrke, 1978 30 233±237. WURZBURG O B, `Converted starches', in Wurzburg O B, Modified starches: properties and uses, Boca Raton, CRC Press, Inc., 1986, 17±40. 1 SALOMONSSON A-C B, ANDERSSON R E, TORNEPORT L J and THEANDER O, `A H- and 13 C-n.m.r. study of bromine-oxidised potato starch', Carbohydr Res, 1991 217 221±225. È HOLM R, NURMI K and BERTOFT E, `Structural characterization of oxidized ZHU Q, SJO potato starch', Carbohydr Res, 1998 309 213±218. È MMA È LAHTI E, BUCHERT J and NURMI K, `Structure of TELEMAN A, KRUUS K, A dicarboxyl malto-oligomers isolated from hypochlorite-oxidised potato starch studied by 1H and 13C NMR spectroscopy', Carbohydr Res, 1999 315 286±292. FALK H, STANEK M and WUTKA R, `Structural aspects of partially acetylated degraded amylopectins ± a 13C NMR study', Starch/StaÈrke, 1997 49 488±491. GRAAF R A D, LAMMERS G, JANSSEN L P B M and BEENACKERS A A C M, `Quantitative analysis of chemically modified starches by 1H-NMR spectroscopy', Starch/ StaÈrke, 1995 47 469±475. XU A and SEIB P A, `Determination of the level and position of substitution in hydroxypropylated starch by high-resolution 1H-NMR spectroscopy of alpha-limit dextrins', J Cereal Sci, 1997 25 17±26. RICHARDSON S, NILSSON G S, BERGQUIST K-E, GORTON L and MISCHNICK P, `Characterisation of the substituent distribution in hydroxypropylated potato amylopectin starch', Carbohydr Res, 2000 328 365±373. MANELIUS R, MAAHEIMO H, NURMI K and BERTOFT E, `Characterisation of fractions obtained by isoamylolysis and ion-exchange chromatography of cationic waxy maize starch', Starch/StaÈrke, 2002 54 58±65. WILKE O and MISCHNICK P, `Analysis of cationic starches: determination of the substitution pattern of O-(2-hydroxy-3-trimethylammonium)propyl ethers', Carbohydr Res, 1995 275 309±318. WILKE O and MISCHNICK P, `Determination of the substitution pattern of cationic starch ethers', Starch/StaÈrke, 1997 49 453±458. FISCHER S K and PILLER F, `New knowledge on degradation of starch by hypochlorite. Part 6. Enzymatic degradation of oxidized starch', Starch/StaÈrke, 1977 29 262±265. FISCHER S K and PILLER F, `New knowledge on degradation of starch by SALOMONSSON B A-C, FRANSSON G M B

ß 2004, Woodhead Publishing Limited

256.

257.

258. 259. 260.

261. 262. 263. 264. 265.

266.

267. 268. 269.

270.

271.

272.

273. 274.

hypochlorite. Part 7. Study on the effect of hypochlorite oxidation on amylose by means of degradation by -amylase and determination of the iodine binding capacity', Starch/StaÈrke, 1978 30 4±7. YOSHIDA M, YAMASHITA T, MATSUO J and KISHIKAWA T, `Enzymic degradation of hydroxyethyl starch. Part I. Influence of the distribution of hydroxyethyl groups on the enzymic degradation of hydroxyethyl starch', Starch/StaÈrke, 1973 25 373±376. BURGT Y E M V D, BERGSMA J, BLEEKER I P, MIJLAND P J H C, KAMERLING J P and VLIEGENTHART J F G, `Substituent distribution in highly branched dextrins from methylated starches', Carbohydr Res, 2000 327 423±429. ZHU Q and BERTOFT E, `Enzymic analysis of the structure of oxidized potato starches', Int J Biol Macromol, 1997 21 131±135. TORNEPORT L J, SALOMONSSON B A-C and THEANDER O, `Chemical characterization of bromine oxidized potato starch', Starch/StaÈrke, 1990 42 413±417. MANELIUS R, BULEÂON A, NURMI K and BERTOFT E, `The substitution pattern in cationised and oxidised potato starch granules', Carbohydr Res, 2000 329 621± 633. MANELIUS R, NURMI K and BERTOFT E, `Enzymatic and acidic hydrolysis of cationized waxy maize starch granules', Cereal Chem, 2000 77 345±353. IMBERTY A and PEÂREZ S, `A revisit to the three-dimensional structure of B-type starch', Biopolymers, 1988 27 1205±1221. IMBERTY A, CHANZY H, PEÂREZ S, BULEÂON A and TRAN V, `New three-dimensional structure for A-type starch', Macromolecules, 1987 20 2634±2636. IMBERTY A, CHANZY H, PEÂREZ S, BULEÂON A and TRAN V, `The double-helical nature of the crystalline part of A-starch', J Mol Biol, 1988 201 365±378. COOKE D and GIDLEY M J, `Loss of crystalline and molecular order during starch gelatinisation: origin of the enthalpic transition', Carbohydr Res, 1992 227 103± 112. WAIGH T A, DONALD A M, HEIDELBACH F, RIEKEL C and GIDLEY M J, `Analysis of the native structure of starch granules with small angle X-ray microfocusing scattering', Biopolymers, 1999 49 91±105. MEYER K H, BRENTANO W and BERNFELD P, `Recherches sur l'amidon II. Sur la nonhomogeÂneÂite de l'amidon', Helv Chim Acta, 1940 23 845±853. KOSSMANN J and LLOYD J, `Understanding and influencing starch biochemistry', Crit Rev Plant Sci, 2000 19 171±226. BRUIJNSVOORT M V, WAHLUND K-G, NILSSON G and KOK W T, `Retention behaviour of amylopectins in asymmetrical flow field-flow fractionation studied by multi-angle light scattering detection', J Chromatogr A, 2001 925 171±182. WAHLUND K-G and ZATTONI A, `Size separation of supermicrometer particles in asymmetrical flow field-flow fractionation. Flow conditions for rapid elution', Anal Chem, 2002 74 5621±5628. MAJZOOBI M, ROWE A J, CONNOCK M, HILL S E and HARDING S E, `Partial fractionation of wheat starch amylose and amylopectin using zonal ultracentrifugation', Carbohydr Polym, 2003 52 269±274. FALK H, MICURA R, STANEK M and WUTKA R, `Structural aspects of native and acid or enzyme degraded amylopectins ± a 13C NMR study', Starch/StaÈrke, 1996 48 344± 346. MCINTYRE D D, HO C and VOGEL H J, `One-dimensional nuclear magnetic resonance studies of starch and starch products', Starch/StaÈrke, 1990 42 260±267. TAMAKI S, HISAMATSU M, TERANISHI K and YAMADA T, `Structural change of potato

ß 2004, Woodhead Publishing Limited

96

275. 276.

277. 278.

279.

280. 281. 282. 283. 284.

Starch in food starch granules by ball-mill treatment', Starch/StaÈrke, 1997 49 431±438. TAMAKI S, HISAMATSU M, TERANISHI K, ADACHI T and YAMADA T, `Structural change of maize starch granules by ball-mill treatment', Starch/StaÈrke, 1998 50 342±348. RICHARDSON S, COHEN A and GORTON L, `High-performance anion-exchange chromatography±electrospray mass spectrometry for investigation of the substituent distribution in hydroxypropylated potato amylopectin starch', J Chromatrogr A, 2001 917 111±121. KOCH K, `Methods for studying starch characteristics' (thesis), Uppsala, Swedish University of Agricultural Sciences, 1999. Ê NG J, RICHARDSON S, NILSSON G S, GORTON L, LAURELL T and MARKOTORTO N, BA VARGA G , `Optimal membrane choice for microdialysis sampling of oligosaccharides', J Chromatogr A, 1998 806 265±278. NILSSON G S, RICHARDSON S, HUBER A, TORTO N, LAURELL T and GORTON L, `Microdialysis clean-up and sampling in enzyme-based methods for the characterisation of starch', Carbohydr Polym, 2001 46 59±68. WHISTLER R L, BEMILLER J N and PASCHALL E F, Starch: chemistry and technology, Orlando, Academic Press, Inc., 1984. THE AMYLASE RESEARCH SOCIETY OF JAPAN, Enzyme chemistry and molecular biology of amylases and related enzymes, Boca Raton, CRC Press, 1995. BULEÂON A, COLONNA P, PLANCHOT V and BALL S, `Starch granules: structure and biosynthesis', Int J Biol Macromol, 1998 23 85±112. BERTOFT E, QIN Z and MANELIUS R, `Studies on the structure of pea starches. Part 3: Amylopectin of smooth pea starch', Starch/StaÈrke, 1993 45 377±382. Ê VALL A-K, `Structural analysis on the amylopectin of waxy-barley BERTOFT E and A large starch granules', J Inst Brew (London), 1992 98 433±437.

ß 2004, Woodhead Publishing Limited

3 Starch bioengineering A. Blennow, The Royal Agricultural and Veterinary University, Denmark

3.1

Introduction: the importance of starch

Starch is the principal energy reserve in plants and is one of the most abundant carbohydrates in the biosphere. Storage starch, which is deposited in storage organs of our main starch crops such as corn, potato, wheat, cassava, etc., serves as the most important energy source for human consumption. Moreover, being a biodegradable polymer with well-defined chemical properties, it has a huge potential as a versatile renewable resource for various material applications. A vast range of native starches with highly different functionalities are already on the market (Swinkels, 1985). However, in order to meet steadily increasing demands for secured and improved specific starch functionalities produced by environmentally friendly starch production systems, yet at low costs, new strategies are required. Biotechnology, specifically targeted to improve starch yield, structure and functionality directly in planta can help to realise these requirements. The advantage is evident for our main starch crops, however, by applying agricultural biotechnology starch can also be functionalised in `local' starch crop genotypes in order to accomplish added value for the farmers as well as increased quality for consumers in the developing parts of the world. The global abundance, economic and nutritional importance, the simple chemical structure and relatively well characterised biochemistry of starch deposition in our chief starch crops makes starch the ideal target for in planta functionalisation using a biotechnological approach. Nevertheless, further efforts are needed to advance towards predictable functionalisation of starch biopolymers directly in the plant. This chapter will highlight promising data in the fields of molecular biology, biochemistry, starch chemistry and modelling

ß 2004, Woodhead Publishing Limited

and suggest strategies for predictable modification of starch in plants stressing the importance of highly cross-disciplinary approaches.

3.1.1 The structure and assembly of starch granules Starch is composed exclusively of glucose residues linked by only two types of bonds: -1,4 and -1,6 glucosidic linkages. The newly synthesised starch molecules are deposited in quite organised arrays to form liquid-crystalline granules, up to 100 mm in size (Fig. 3.1). The major constituents of starch granules are amylopectin and amylose. Amylopectin (~75%) is a semicrystalline highly branched polysaccharide with an -1,4 backbone and 4±5% -1,6 branch points enabling structuring of the starch granule while amylose (~25%) is amorphous in the native starch granule and is composed of essentially linear chains of -1,4 linked glucose units. As a result of the simple chemistry of starch its biosynthesis requires relatively few enzymatic steps and it is an ideal model system for biotechnological modification of plant biopolymers. However, the simple chemistry of starch does not reflect the extreme complexity of the final starch molecules. The structure and self-assembly processes of the resulting starch biopolymers in the plant are not at all understood at the level that is required to predict the starch configurational effects of a given genetic alteration. The revolution in plant biotechnology and carbohydrate analysis that has taken place during the last decades has given starch biotechnology a promising

Fig. 3.1 Schematic representation of the molecular organisation of the starch granule. Concentric growth rings with approx. 1 m width are composed of radially orientated arrays of liquid crystalline aligned double helical -glucan chains (Waigh et al., 2000). A 9 nm lamella consists of a thin -1,6, branched base and a crystalline matrix, approx 4±7 nm wide, of -1,4 linked glucose residues. Some of the amylopectin chains are phosphorylated. Glucose units are represented by dark grey spheres. Amylose (light grey spheres) is interspersed in the amorphous regions.

ß 2004, Woodhead Publishing Limited

future and paved the way for agricultural biotechnology applied to our main starch producing crops. Systems biology approaches, i.e., taking into account the whole crop plant functionality utilizing and interpreting the information in enormous gene sequence data banks, are widely adopted to exploit new and high value functionalities of crops. Even though starches isolated from different genotypes have widely different functional properties, there is a continuously rising demand for new and improved specific functional properties of starch. It should be noted that any significant improvement, if only minor, of a starch crop will result in massive societal impact as a result of the huge volumes of starch consumed globally (Burrell, 2003). If these starch functionalities can be produced directly in planta, a strategy also known as `molecular farming', not only new added-value starch biopolymer products will be obtained but this approach will also contribute to shortening the production chain and, in many cases, reducing the need for energy-demanding, expensive and polluting industrial processes such as chemical phosphorylation, carboxylation, hydroxypropylation, acetylation, etc. (Light, 1990). Hence, plants can be used as efficient bioreactors for the production of functionalised starch biopolymers directly deposited in the seeds of cereals, tubers of, e.g., potato and roots using, e.g., cassava. It should however be borne in mind that deposition of starch, like all other biopolymers, is optimised for plant functionality and not for the consumers' desires. Hence, structural alterations that impede the storage and remobilisation functionalities will inevitably result in decreased crop production. However, for storage starch deposition, which is not an absolute requirement for plant survival, quite a wide range of structural alterations are accepted.

3.2

Technologies for genetic modification and starch profiling

Starch is synthesised by a quite limited number of enzymes coded for by genes that can be suppressed or alternatively over expressed directly in the plant. The levels of these can be manipulated in the plant, in planta, by mutation or by transgene technology to generate entirely novel starch biopolymers. Hence, alterations at the gene or genome level are manifested at the phenome level generating a novel phenotype, i.e., an altered starch structure. The analysis of the generated biopolymers is a quite complex and demanding task. For starch analysis, a number of analytical technologies have been developed both to permit high throughput screening and produce much more reliable and detailed data of starch chemical and physical properties.

3.2.1 Genetic technologies The main approaches used to obtain a genetic alteration of a plant are chemical or radiational mutation which brings about direct and specific alteration, suppression or deletion of the expression of a gene or alternatively, gene transfer

ß 2004, Woodhead Publishing Limited

which can be used to express or suppress desired genes in a crop in a more predictive way. Mutants of a variety of starch crops including maize, wheat and barley can be generated and screened for altered starch structure without the prior knowledge of the gene(s) sequences affected, a strategy called forward genetics. A number of mutants with altered starch structure are now available for our most important starch crops including maize (Nelson, 1985), pea (Wang et al., 1998), rice, barley and wheat (James et al., 2003). The maize sugary (high sugar content and deposition of highly branched phytoglycogen instead of starch) and waxy mutants (devoid of amylose) are examples of genotypes that were used commercially long before the actual genes affected were identified and cloned (Nelson, 1985). However, the identification of the primary metabolic step affected is not always trivial. This is the case for the sugary mutation which is located in a starch debranching enzyme gene but whose suppression results in highly pleiotropic effects, i.e., a complex series of side effects taking place that are indirectly linked to the target mutation (Dinges et al., 2003). However, in some cases the exact genetic lesions have been identified, e.g., for the rugosous (r) mutation in pea (Bhattacharyya et al., 1993), the amylose extender phenotypes in maize (Stinard et al., 1993), and rice (Nishi et al., 2001) the expression of starch branching enzymes are suppressed. In contrast to the untargeted radiational or chemical approaches for generation of mutants, predictable generation of modified starch biopolymers in planta requires a general approach termed reverse genetics. By using the information of the known DNA gene sequence of a mutated locus (gene position) the function of the gene with respect to, e.g., starch structure, is assessed. The fact that genome programmes have not only provided full DNA sequence data for the model plant Arabidopsis but also partial sequences for main starch crops genomes provides the basis for successful identification, cloning and modification of relevant genes in starch biosynthesis. The first genome sequence to be available will probably be rice for which a draft sequence in now available (Goff et al., 2002). Forward genetics approaches have facilitated the isolation of mutants with specific and predictable alterations in key metabolic reactions (Maes et al., 1999). The strategies include the T-DNA (Agrobacterium mediated gene transfer) and transposon (`jumping genes') insertion systems and the specific suppression of a target gene expression using homology-dependent gene silencing approaches such as antisense RNA and RNA interference (RNAi) technologies. During the last two decades, transgenic approaches have been employed to modify starch in plants for which mutants are not easily generated. Potato and cassava can be readily and efficiently transformed with desired genetic constructs using the Agrobacterium gene transfer system. Hence, these crops have become important modes for transgenic modification. For altering starch functionality of the potato tuber, the antisense RNA and overexpression approaches have proven to be highly successful (Kossmann and Lloyd, 2000). An example where the transposon technology was successfully employed is the maize Mutator insertional mutant showing an amylose-extender phenotype, which was mapped to the Sbe2b gene (Stinard et al., 1993).

ß 2004, Woodhead Publishing Limited

The first generation of plants engineered for directed production of novel biopolymers is already under way and involves modification of the natural polymers in the plant adapted to fulfil specific food functionalities. Some functionalities obtained using the potato system is shown in Fig. 3.3 on page 103. However, detailed structural and functional analysis of most of these starches remains to be completed.

3.2.2 Analytical technologies for starch structural profiling State-of-the-art analytical technologies that provide detailed molecular mass or chain length structural data for starch samples but cannot be used for screening of very large numbers of plants or starch samples involve size exclusion chromatography (SEC) providing detailed profiling information about starch molecular mass and amylose content, the Dionex (BioLC) chromatography system (Blennow et al., 2000) and the FACE system (O'Shea and Morell, 1997) giving very detailed chain length structural data. Rheological and melting profiles can be generated using rapid visco analysis (RVA, Walker et al., 1988) used for quick gelatinisation profiling and differential scanning calorimetry (DSC) for generation of melting and recrystallisation profiles (LelieÁvre and Liu et al., 1994). At the granular size level down to nm dimensions, light- electron- and confocal laser light microscopy (Blennow et al., 2003) and atomic force microscopy (AFM, Ohtani et al., 2000) provide topographical and sometimes internal structural information. Spectroscopy, including nuclear magnetic resonance (NMR), Xray and infra-red (IR) spectroscopic techniques can provide very detailed information on starch chemistry, conformation and phase characteristics. NMR discloses chemical bonds in the starch molecules where high field 13C MAS NMR (Tang and Hills, 2003) gives information related to conformations, e.g., double helices, crystallinity, and water binding and low field NMR supply detailed information on hydration characteristics (Thygesen et al., 2003). X-ray crystallography crystallinity type (A-, B-, and C-type, Imberty et al., 1988; Imberty and PereÂz, 1988) and extent of crystallinity (wide angle X-ray spectroscopy, WAXS) whereas repeated structures at the nano scale, e.g., the 9 nm lamellae is analysed by small angle X-ray spectroscopy (SAXS, Jenkins et al., 1993). IR and related spectroscopic techniques, e.g., near infra-red (NIR), raman, fourier transform IR spectroscopy (FTIR, Sevenou et al., 2002), and NIR (Thygesen et al., 2001) are promising with respect to in situ screening of larger number of mutant and transgenic collections for altered biopolymer alterations. These data can in many cases be related and correlated to more detailed structural and functional data such as crystallinity and viscosity. Outcome from some of those technologies applied to starch are shown in Fig. 3.3 below.

ß 2004, Woodhead Publishing Limited

Fig. 3.2 Key target reactions of starch biosynthesis for in planta modification of starch yield and structure. The actual sequence of the reactions is not exactly known. ADPglucose pyrophosphorylase (AGPase), soluble starch synthase (SS), granule bound starch synthase (GBSS), starch branching enzyme (SBE), glucan water dikinase (GWD) and starch debranching enzyme (DBE).

3.3

Improving starch yield and structure

By engineering metabolic reactions in close connection to starch biosynthesis or altering the flow of carbohydrate in the plant (Fig. 3.2), the yield and structure of starch can be affected. The main structural elements that can be readily analysed are the amylose concentration, the amylopectin and amylose chain length distribution, the amylopectin and amylose molecular mass, phosphate substitution and starch granule size distribution. All these molecular alterations have profound effects on the physical properties of the starch granules and their gels (Fig. 3.3 and Table 3.1). Most of the enzyme reactions responsible for synthesis of these constituents have been characterised to some extent and their effects demonstrated in vivo. However, the existence of pleiotropic effects is rather a rule than an exception and complicates the identification of the site of the mutation. These effects could be a result of alterations of enzyme activities other than the target gene, the cooperative nature of starch biosynthesis, etc. In principle, chain length can be altered either by engineering the elongation reaction or by engineering the branching reaction and phosphate substitution can be engineered both by altering the phosphorylating enzyme and by engineering the chain length profile. The results may be very different and the pleiotropic effects make it hard to interpret the data and generate predictive models. In this section, the strategies for modifying starch yield and structure will be described in the light of these effects. The enzyme reactions with closest relation

ß 2004, Woodhead Publishing Limited

Fig. 3.3 Examples of the transgenic approach to produce functionalised starch in planta. Suppression of specified key enzymes of starch biosynthesis has been performed using antisense technology using the potato tuber as a model and the resulting structural, pasting and gel texture alterations are examined.

to starch biosynthesis are outlined in Fig. 3.2. These involve reactions for sugar activation (ADPglucose pyrophosphorylase, AGPase), chain elongation (soluble starch synthase, SS, granule bound starch synthase, GBSS), chain branching (starch branching enzyme, SBE, phosphorylation, glucan water dikinase, GWD), and debranching (starch debranching enzyme, DBE). The primary effects taking place as a result of their expression or suppression, respectively, in the plant are outlined in Table 3.1.

3.3.1 Engineering starch yield The yield of starch in a specific crop can be changed with the only aim of increasing starch yield but not affecting the starch functionality. However, an unwanted and quite usual side effect of in planta functionalisation of starch is the `yield penalty', i.e., a significant decrease in starch content or crop production is linked to the altered phenotype (Visser and Jacobsen, 1993). For both situations, strategies to increase starch deposition in the starch storing organs have to be developed. Current knowledge of starch metabolism suggests that the main approaches to obtain increased starch content in a starch producing plant is to increase flow of carbon into starch facilitated by increasing the production of ADPglucose in the cells or decreasing the rates of starch degradation by suppressing amylolytic activities. Yield penalties for in planta starch modifications occur in cases where the starch structure is grossly altered so as to prevent accurate deposition and formation of the starch granules as demonstrated in pea (Smith et al., 1990; Wang et al., 1998) or when the flow of metabolism towards starch synthesis is directly affected as demonstrated in

ß 2004, Woodhead Publishing Limited

Table 3.1 Some structural and physical properties of transgenic and mutant starches. Roman type (or -) denotes decreased concentrations and effects and bold type (or +) denotes increased concentrations and effects. The resulting main physical effects indicated should be regarded as guidelines. Opposite effects may occur, e.g., for extreme concentrations and in a mixed system Enzymes

ÿ ADPG concentration1

ÿ GBSS

ÿ SSS

Main structures affected

Amylose

Amylose

Amylopectin Amylopectin branching branching Amylose Phosphate

Starch Phosphate branching Amylose

Amylopectin Amylopectin branching branching Phosphate Phosphate

Granule crystallinity2 Gelatinisation temperature Water solubility Gel clarity Gel viscosity (G'')3 Gel stickiness Gel retrogradation Gel freeze-thaw stability Gel forming capacity (G')3 Gel syneresis

+ + + + + -

+ + + + + -

+ + + + +

+ + + + +

+ + + + +

+? + + + +

+ + + + +

+ + + + +

Starch crops

Potato Pea

Potato Cassava Maize Pea Rice

Potato Pea

Potato Maize Pea Rice

maize

Potato

Potato

Potato

1 2 3

ÿ SBE

ÿ DBE

ÿ GWD

+ GS

+ GBE

ADPG content decreased by modulating the AGPase, ATP transporter, Sucrose synthase, Phosphoglucomutase, malic enzyme as described in the text. In cases where a transformation between A- and B-type crtstalline polymorphs occurs, e.g., for pea (Bogrecheva et al., 1999) crystallinity can instead decrease. G' = elastic modulus G'' = viscous modulus.

potato (Tjaden et al., 1998). However in other cases, the yields are only moderately affected as shown for, e.g., potato (Tjaden et al., 1998; Lorberth et al., 1998; Viksù et al., 2001), sweet potato (Noda et al., 2002) and pea (Wang et al., 1998). For example, two of the most used maize mutants producing high amylose and high amylopectin (waxy) starch gives 75% and 91% of normal yield, respectively (Nelson, 1985). The initial and first committed step of starch biosynthesis catalysed by AGPase is the obvious target for attempts to increase the starch yield in the crop plant. Even though upstream metabolic reactions, e.g., sucrose synthase and phosphoglucomutase (Wang et al., 1998), are vital for starch deposition (see also below), several lines of evidence suggest that AGPase is the prime rate-limiting step of starch biosynthesis (Stark et al., 1992). Hence, increasing the activity of AGPase or increasing the immediate substrates for this enzyme, i.e., G1P and ATP or side products, i.e., pyrophosphate (PPi) would increase the rate of production of ADPglucose which in principle would result in increased starch yield. The importance of AGPase for starch yield was verified by transgenic potatoes expressing the E. coli AGPase which accumulated up to 60% more starch (Stark et al., 1992). By expressing the E.coli pyrophosphatase in potato, pyrophosphate, which is one of the products of the AGPase catalysed reaction, was removed from the site of starch biosynthesis and the formation of ADPglucose was stimulated resulting in 20±30% increased starch accumulation in the young tubers (Geigenberger et al., 1998). However, mature tubers instead showed a decrease in starch content (Sonnewald 1992) addressing the importance of using the appropriate model system for investigation of yields. Moreover, if the total yield of starch storing organs (e.g. potato tubers) is significantly decreased, no increase in starch production is achieved. A promising result was achieved by antisensing the NAD malic enzyme in potato tubers. Using this strategy, the glycolytic metabolism was affected and as a side effect, starch content in the tubers increased with no effects on tuber yield (Jenner et al., 2001). In an attempt to increase the concentration of ATP in the starch storing organ, the transport of ATP into the amyloplast was increased by expression of the plastidic ADP/ATP transporter in the potato tuber resulting in increased starch concentration (Tjaden et al., 1998). However, tuber production was slightly decreased. Very recently, a significant increase in the starch content and tuber yield was achieved (Regierer et al., 2002) which is a major step towards high yield production of a variety of different starches by in planta biotechnology. In this study the adenylate pool in potato tubers was significantly increased by antisense suppression of the adenylate kinase. A general problem is that the structure of the starch is often altered as an effect of increased flow of carbon into starch, which hampers the possibility to exclusively increase the starch content in a starch crop. Especially amylose tends to increase with increased starch deposition as shown for potato (Tjaden et al., 1998; Viksù-Nielsen, et al., 2001), pea (Clarke et al., 1999) and Arabidopsis leaf starch (Critchley et al., 2001). This effect is probably caused by different

ß 2004, Woodhead Publishing Limited

affinities of ADPglc, for different starch synthases (Clarke et al., 1999), i.e., GBSS that synthesises amylose becomes more active. However, no significant alterations of the starch structure were detected by instead decreasing starch content by suppressing the activities of phosphoglucomutase (PGM) sucrose synthase (susy) or AGPase in pea (Wang et al., 1998). By inhibition of starch degradation, the deposition of starch can be increased. The leaf starch excess phenotype in (sex) Arabidopsis is an interesting example where starch degradation is partly suppressed either by mutation of the GWD gene (Zeeman and ap Rees, 1999; Yu et al., 2001) or in a putative amylase genes (Zeeman et al., 1998a), debranching enzyme (Zeeman et al., 1998b) or in disproportionary enzyme gene (Critchley et al., 2001). Antisense suppression of a -amylase in potato gave a starch excess phenotype in the leaves (Scheidig et al., 2002). However when the GWD enzyme was suppressed in the potato system, starch excess was seen only in the leaves, not in the tubers (Lorberth et al., 1998) stressing that it is necessary to specify and investigate the appropriate organ, i.e., tuber or seed and that transient starch accumulation is very different from storage starch deposition. Attempts to increase starch content by affecting enzymes directly involved in starch polymerization and modification naturally leads to more dramatic effects. Most often, starch yield is severely decreased. However, antisensing of the starch-phosphorylating enzyme GWD in potato results in starch and amylose excess in the leaves but no significant starch excess in the tubers. In the tubers, the amylopectin molecular mass is reduced and, as expected, phosphate level in the starch is severely decreased (Viksù-Nielsen et al., 2001). These, and similar effects will be discussed in more detail next.

3.3.2 Engineering chain elongation The elongation step of starch biosynthesis is catalysed by isoforms of soluble starch synthase (SSS) and granule bound starch synthase (GBSS) all of which differ in substrate specificity and product structure. The engineering of amylose and amylopectin structure and concentration can be performed by altering one or several starch synthase activities (Fulton et al., 2002). The presence of several activities complicates the engineering of starch chain elongation since there is a partial redundancy, i.e., the suppressed activity of one SS isoform can be partly compensated for by the activity of another isoform. Moreover, the activities of the isoforms are dependent on each other complicating the interpretation of the effect on starch structure of a single SS isoform alteration. Engineering amylose content The biosynthesis of amylose is controlled by the plant at numerous metabolic steps (Denyer et al., 2001). However, the isoform enzyme responsible for amylose biosynthesis GBSS, was identified early in maize (Nelson and Rines, 1962), and suppression of this enzyme specifically results in a complete removal of amylose. Low amylose or `waxy' type starches were found very early in

ß 2004, Woodhead Publishing Limited

mutants of maize (Weatherwax, 1922) but have now been isolated from a vast number of mutant and transgenic starch crops including potato (Flipse et al., 1996; Visser et al., 1997a) barley (Ono and Suzuki, 1957; Hylton et al., 1996) rice (Murata et al., 1965), wheat (Nakamura et al., 1995), pea (Denyer et al., 1996) sorghum (Hsieh et al. 1996), and sweet potato (Noda et al., 2002). GBSS is entirely bound within the starch granule matrix and works processively on each amylose molecule (Denyer et al., 1999). Engineering amylose content in the starch granule by suppression of GBSS does not result in any alterations of starch granule morphology or topography. The reasons might be that amylopectin structure is not significantly altered by the presence of amylose in normal granules as the amylose molecules are synthesised and deposited in existent small cavities in the semicrystalline scaffold of the starch granules and do not normally interact significantly with amylopectin (Tatge et al., 1999; Denyer et al., 2001). Finally is should be noted that when starch yield is increased, amylose content is also often somewhat increased as described in Section 3.3.1. Engineering amylopectin chain length Amylopectin chain elongation can be engineered by the suppression or expression of soluble starch synthase activities. Antisense suppression of the soluble starch synthase activity in potato changes the chain length profile of the amylopectin. In pea, the rug5 mutant devoid of a soluble starch synthase, also contains increased amylose content (Bogracheva et al., 1999). As expected, suppression of elongating activities in potato leads to an increase in shorter chains in the amylopectin but unexpectedly also synthesis of a pool of very long chains when the SSSIII isoform is suppressed. The resulting effect of suppressing SSSIII is severely misshapen and fissured starch granules (Craig et al., 1998; Edwards, et al., 1999; Lloyd et al., 1999). Interestingly, the simultaneous suppression of SSS and GBSS causes a reversion to the original morphology of the starch granules indicating that the presence of amylose can have detrimental effects in granules with erroneous amylopectin structure (Fulton et al., 2002). The presence of very long amylopectin chains in starches obtained from antisense starch synthase III lines was suggested to be responsible for severe granule fissuring. It is plausible that granule morphological disturbances may be caused by `forbidden' interactions between non-structural amorphous amylose and structural semi-crystalline amylopectin segments. Using the barley system, removal of the SSIIa isoform resulted in smaller amylopectin chains and, surprisingly, higher amylose content (Morell et al., 2003). The latter effect, being unusual, again addresses the problem with unpredictable side effects. Another example is the starch produced by potato tubers expressing a bacterial glycogen synthase (GS, Shewmaker et al., 1994). Since the elongation activity is increased in these tubers, the starch would be expected to be less branched and contain longer chains. However, these potatoes yield a more branched starch indicating the specificity of GS determines the structure yielding a more glycogen-like product.

ß 2004, Woodhead Publishing Limited

3.3.3 Engineering chain branch point formation -1,6 chain branch point formation in starch is in principle reduced by the suppression of the endogenous starch branching enzyme (SBE) activities in the plant. The resulting amylopectin molecules will contain less -1,6 branch points and as a consequence, longer -1,4 backbone chains in the amylopectin but also increased amount of amylose. The opposite effects are obtained by the overexpression of branching enzymes from other organisms. The two main different SBE isoforms present in plants, SBE I (SBEB) and SBEII (SBEA), catalyse chain transfer resulting in isoform specific distribution of the amylopectin side chains (Guan and Keeling 1998; Andersson et al., 2002). In these investigations, the SBE I isoform has been shown to transfer longer chains than the SBE II form in vivo as well as in vitro. As for all starch synthesizing enzymes, isoforms of SBE work in a highly cooperative way and as for the SSs, the activities are partly redundant. Suppressing the SBE level results in increased amylose concentrations, increased lengths of the amylopectin unit chains as demonstrated for potato (Jobling et al., 1999; Schwall et al., 2000) pea (Lloyd et al., 1996; Craig et al., 1998; Bogracheva et al., 1999) and maize (Boyer et al., 1976; Bhattacharyya et al., 1993). The resulting starch granules are severely altered with respect to morphology and topography showing irregular, elongated, multi-lobed or fissured shapes (Boyer et al., 1976; Blennow et al., 2003). It is likely that these effects likely stem from the altered (long chain) amylopectin and amylose structures and increased amylose contents found in these starches. It is unlikely that pure amylose can form an ordered starch granules since granule integrity is strictly dependent on stabilising forces of the branch points. In extreme cases, the identification of pure amylopectin and amylose components is impossible (Shi et al., 1998). A possible explanation is that these two molecules converge in branch structure and molecular mass in the mutants resulting in `forbidden' amylose-amylopectin interactions in the granule distorting the organised amylopectin framework. The combined suppression of GBSS and SBE in potato tuber results in an amylose free starch with properties that are somewhat different from those of the pure amylose free starch (Flipse et al., 1996). Increased branching of the amylopectin can be achieved by overexpression of the E. coli glycogen branching enzyme in potato tubers (Kortsee et al., 1996). An interesting pleiotropic effect of moderation of branch points in starch is the inverse correlation to the on starch phosphate content (Section 3.3.5).

3.3.4 Engineering chain debranching and disproportionation By affecting enzymes involved in the degradation of starch, novel starch structures can be generated in planta. Phytoglycogen is a water-soluble -glucan present in the sugary maize mutant whose structure resembles that of animal glycogen (Morris and Morris, 1939). As a consequence of its highly branched structure this -glucan does not have the capability to form semi-crystalline

ß 2004, Woodhead Publishing Limited

granules. These genotypes also clearly demonstrate the importance of the amylopectin molecule to be optimised in order to generate semicrystalline and ordered starch granules. The accumulation of phytoglycogen in plants is dependent on the suppression of a debranching step catalysed by isoforms of debranching enzyme (DBE) supposed to be in close balance with the branching activity (Nelson and Pan, 1995) or as an integrated step in starch biosynthesis as formalised recently (Ball et al., 1996; Mouille et al., 1996). Two distinct isoforms of DBE exist, namely the isoamylase and the pullullanase isoforms. Both isoforms have been indicated to be involved in the biosynthesis of starch as shown for rice (Nakamura et al., 1997; Kubo et al., 1999) and maize (Dinges et al., 2003). Hence, the production of highly branched starch biopolymers, phytoglycogen, can be achieved by suppressing the DBE activity (Nakamura, 1996). However, many pleiotropic effects have been identified in DBE mutants hampering a straightforward interpretation of the enzyme activity±starch structure relationship since many other enzyme activities of starch metabolism are changed in these mutants (Dinges et al., 2003). Another starch degrading enzyme that is indicated to play a role during starch biosynthesis is the disproportionary enzyme (DPE), an enzyme that cleaves 1,4 linkages and subsequently condenses the liberated chain onto another chain forming a new -1,4 bond thereby disproportinating the chain length distribution of the target starch molecules (Takaha et al., 1993). Even though no changes were observed on the antisense suppression of the DPE in potato tubers (Takaha et al., 1998), effects have been found when analysing leaf starch (Critchley et al., 2001) hence, possibly DPE has a limited function for the deposition of storage starch.

3.3.5 Engineering chain phosphate substitution Phosphorylation of starch occurs simultaneously with starch biosynthesis (Nielsen et al., 1994) but the function of starch phosphorylation is not clear (Blennow et al., 2002). The starch phosphorylating enzyme glucan water dikinase (GWD) catalyses the formation of starch phosphate monoesters at the C-3 and C-6 positions of the glucose units in the amylopectin using ATP as the phosphate donor (Ritte et al., 2002). The degree of starch phosphorylation is rather low, i.e., only one out of approx 300 glucose units is phosphorylated in the highly phosphorylated potato starch (Blennow et al., 1998). GWD has a substrate recognition requirement for -1,6 bonds (Mikkelsen et al., 2003) and preferably phosphorylates the long chains in the amylopectin molecule (Blennow et al., 1998; Mikkelsen et al., 2003). Mainly three strategies can be used to engineer the bound phosphate concentration in starch. The most obvious is the direct change of expression of the starch phosphorylating enzyme GWD. However, changing the degree of branching of the starch by altering the expression of SBE also generates changes in the degree of phosphorylation. Finally, altering the expression of SS also affects starch phosphate content. While quite dramatic changes in the phosphate content are

ß 2004, Woodhead Publishing Limited

obtained by altering GWD and SBE levels, only minor alterations are accomplished by antisense suppression of starch synthases. The effects upon suppression of the GWD in potato (Lorberth et al., 1998; Ritte et al., 2002) or the SEX1 (starch excess) homologue in Arabidopsis (Yu et al., 2001) results in starch that is almost totally depleted in phosphate. However, in the Arabidopsis leaf, a starch excess phenotype is observed and in Arabidopsis and potato, the amylose content is increased and in potato the amylopectin molecular weight is decreased (Viksù-Nielsen et al., 2001). It remains to be investigated if overexpression of GWD will result in increased phosphorylation. As mentioned in Section 3.3.3, the suppression of SBE activity results in increased amylose concentrations and increased lengths of the amylopectin unit chains. An interesting additional effect is the dramatic ~3-fold increase in the level of starch bound phosphate (Schwall et al., 2000). The reason for the latter effect is likely that the GWD activity is highly stimulated by long (DP30±100) amylopectin chains (Mikkelsen et al., 2003) which are enriched in these plants. It should, however, be noted that GWD still has a requirement for the -1,6 branch points in the amylopectin. Hence, amylose is not likely to be phosphorylated in these plants despite of its long chains. Suppression of GBSS also increases the phosphate concentration in the total starch. However, this is only an effect of the removal of amylose (approx 20± 30%) which is not phosphorylated. Suppression of the SSS also affects the phosphate content but in a more unpredictable way. When the total SSS activity is repressed, the starch phosphate content is somewhat reduced (Lloyd et al., 1999) whereas suppression of only the SS III isoform increases the phosphate content by 70% (Abel et al., 1996).

3.3.6 Engineering starch molecular mass As a result of the rather complicated technology for investigating starch molecular mass (see, e.g., Blennow et al., 2001) few thorough investigations have been performed on the effects of starch molecular mass upon genetic engineering of a starch crop. No strategies have been put forward to enable a specific increase or reduction in the starch molecular mass. It was, however, notified that the apparent mass of the amylopectin was severely reduced by antisensing the GWD in potato tuber (Viksù-Nielsen et al., 2001). However, this effect was accompanied by decreased starch phosphate content (the prime effect of reduced GWD) and increased amylose content. Future possibilities for engineering the molecular mass of starch could involve the disclosure of possible glucan primers whose concentration would be likely to govern starch molecular mass and the effects of expression of endo amylases in plants which would in principle generate lower molecular mass starch molecules.

ß 2004, Woodhead Publishing Limited

3.4

Physical and chemical properties of modified starches

Biotechnological modification of starch molecular structures directly affects fundamental physical properties of the starch. The results on starch granule morphology, topography, crystallinity and assembly have been briefly described in Section 3.3. The chief molecular alterations made by in planta starch biotechnology are amylose content, amylopectin chain structure and phosphate content. All these structural elements exert profound effects on the physical properties of the starch. These involve properties of the intact starch granule as well as the solutions, gels and recrystallised material obtained for starch granules treated in aqueous and other environments. In this section relevant examples of functionalities of engineered and mutant starches will be given. The amylose content is perhaps the constituent that exerts the most influence on starch physical properties. As described in Section 3.3, biotechnological over-production of amylose results in a concomitant increase in the amylopectin chain length and (in tuberous starches) increased phosphate content, both of which also affect the starch functionality. On the contrary, suppression of amylose can be performed quite specifically. Suppression of the amylose content can be achieved without significant alterations of the amylopectin structure as demonstrated for potato starch (Visser et al., 1997a,b). Hence, increasing the amylose content in maize affects chain length of the amylopectin molecule and as a consequence starch crystallinity and crystalline polymorph, i.e., a shift from A-type to B-type polymorph of the native starch granules (Cheetham and Tao, 1998). The most evident changes that are linked to the amylose content are the swelling power and retrogradation behaviour. Swelling power or hydration capacity is controlled by the amylose content which can be quite specifically adjusted by controlling the levels of GBSS (Ortega-Ojeda and Eliasson, 2001; Visser et al., 1997b). The retrogradation properties are mainly governed by the amylose content but also by other structural variables including the amylopectin chain length and phosphate content (Jacobson et al., 1997, Thygesen et al., 2003). Increasing the chain length of the amylopectin enhances the re-crystallisation properties of the starch (Moates et al., 1997; Jobling et al., 1999; Jane and Chen, 1992; Fredriksson et al., 1998; Yuan and Thompson, 1998). Hence, retrogradation can in principle be engineered and controlled by specific modification of these parameters. It should be noted that while the high amylose content results in irreversible recrystallisation of the starch, amylopectin crystallisation is reversible. The length of the amylose molecules is important so that a certain length is required for crystallisation and retrogradation (Miles et al., 1985). However, amylose chains above approx. 1000 glucose units tend to form a gel instead of crystallising (Gidley and Bulpin, 1989; Clark et al., 1989). However, there is no clear strategy on how to engineer the chain length and molecular size of amylose. Chemical substitution of starch is routinely performed in order to stabilise the starch in aqueous media. However, tuberous starches are naturally phosphorylated which brings about functionalities similar to those of chemically modified

ß 2004, Woodhead Publishing Limited

starches. In potato starch the degree of phosphorylation is influenced by the elongation of the side chains which stimulates the phosphorylation of the starch (Schwall et al., 2000). The covalently bound phosphate content not only increases gel viscosity but also clarity as shown for the potato system (Veselovsky, 1940; Wiesenborn et al., 1994). Other expected effects are increasing solubility and suppression of retrogradation. Hence it is expected that starch engineered for increased phosphate content by suppressing SBE will display these functionalities. However, the increased chain length of the amylopectin and the increased amylose content supposedly overrule these expected effects since the starch behaves as a typical high amylose starch. However, specific suppression of the starch phosphorylating activity, GWD, yields the expected results, i.e., stronger gels and decreased gelatinisation viscosity (Lorberth, 1998; Viksù-Nielsen, 2001). It should, however, be noted that in these plants the molecular mass of the amylopectin is decreased and the amylose content somewhat increased in these starches (Viksù-Nielsen et al., 2001). Short chain starches can be generated by reduction of SSII as shown for pea in the rug5 mutant (Craig et al., 1998) and antisense of the SSII and SSIII in potato (Edwards et al., 1999; Lloyd et al., 1999). The shorter amylopectin chains deposited in the barley SSIIa mutant results in a reduced gelatinisation temperature of the starch (Morell et al., 2003). Recently, the simultaneous antisense inhibition of GBSS, SSII and SSIII in potato abolishes also the amylose synthesis (Jobling et al., 2002) yielding a starch that after gelatinisation forms solutions that are virtually totally stable. With respect to these functionalities, novel starches with specific functionalities, e.g., short chains are continuously being discovered, e.g., the specific functionality of floridean starch (Yu, et al., 2002). Structure±functionality relationships generated for these natural starches can be used to help to qualitatively foresee functionalities of a biotechnologically modified starch. The examples given above clearly demonstrated that starch engineering is a multivariate problem and hence, at least initially has to be dealt with using appropriate strategies to allow for the identification of the structural factors that are of most importance for starch functionality. These approaches will be more thoroughly discussed in Section 3.6.

3.5 Functionality and uses of modified starches in food processing Native starch has many applications in a number of food products. Bioengineering starch directly in the plant can generate native starches with combined functionalities that are not easily or cheaply obtained by post-harvest modifications. These functionalities can be controlled by moderating the key enzymatic reactions of starch biosynthesis as described above. However, while mutant starches, primarily the waxy and high amylose maize starches, are extensively used to impart specific functionalities in various food applications

ß 2004, Woodhead Publishing Limited

(Light, 1990; Deis, 1998), the documentation in literature for transgenic food systems is yet limited. However, in a unique study, a set of transgenic potato starches with altered amylopectin chain length structure and phosphate substitution was used in a food model system (Wischmann et al., (2002). For starch applications, properties such as the melting-, solubilisation-, recrystallisation- (retrogradation-), swelling and visco-elastic properties are of great importance. Parameters with importance for food quality such as adhesiveness, chewiness, gumminess and organoleptic characteristics could be linked to the specific altered structural and rheological properties. A complication with food structure-functionality models is that they are highly dependent on the process model system used, e.g., water content, temperatures, and other components of the model. For example, salt and pH affect the viscosity of phosphorylated starches but not very much that of unsubstituted starches. The addition of only small amounts of salt dramatically decreases gelatinisation viscosity (Muhrbeck and Eliasson, 1987). For the pure starch preparations additional minor components such as lipids integrated in the amylose fraction during biosynthesis or starch processing in cereal starches (Swinkels, 1985, Lin and Czuchajowska, 1998) and proteins present mainly at the granule surface (Baldwin, 2001) are known to affect starch functionality. Since lipids are complexed mainly to the amylose (Lim and Seib, 1993) engineering amylose concentration in starch, especially in cereal starch, also affects the lipid content which can have profound effects on starch gelatinisation behaviour by decreasing the gelatinisation viscosity and slowing down retrogradation (Lin and Czuchajowska, 1998). Protein attached to the processed starch can give unwanted odour and colour and restrict enzyme attack and hydration (Baldwin, 2001). Functionalities of the pure starch components that are expected to become important for starch-based food systems include freeze-thaw stability, water solubility, resistance towards enzyme attack, extreme viscosity and rigid and clear gels. Traditionally, starches that give freeze-thaw stable pastes are generated by post-harvest chemical cross-linking (Yangsheng and Seib, 1990) and phosphorylation (Liu et al., 1999) of the native starch. For native, unmodified starches, freeze-thaw stability requires high amylopectin concentrations as shown for rice starch (Varavinit et al., 2002), barley (Bhatty, 1999) and a very short chained starch recently isolated from red algae and functionally analysed (Yu et al., 2002). Hence, strategies for obtaining freezethaw stable starches rely on reducing the amylose content and-or decreasing the chain length of the amylopectin by completely reducing the GBSS activity suppressing the SSS activities (Jobling et al., 2002). Phytoglycogen, which is entirely soluble and can be produced by mutant or transgenic plants devoid of DBE, would be an ideal freeze thaw stable starch analogue. However, thus far, no strategies have been proven to efficiently produce large amounts of readily extractable phytoglycogen in the plant. High amylose starches generated in planta, e.g., by suppression of the SBE can be made highly resistant towards enzymatic degradation (`resistant starch,

ß 2004, Woodhead Publishing Limited

see Chapters 19±21 this volume) as shown for, e.g., high amylose barley starch (Szczodrak and Pomeranz, 1991). The increased gelling and retrogradation behaviour of high amylose starches makes it ideal to achieve gelled, short and chewy bite of gelled products. Hence, high amylose starch may have specific functionalities to partly substitute for gelatine (Richardson et al., 2000) but is also useful for generating crispy surfaces of fried foods, e.g., chips (Light, 1990). However, high amylose starches also require high process temperatures. This is an incentive for producing low amylose starches where functionality is not of prime importance. High amylopectin starch pastes yield weak, stringy and cohesive functionalities and give increased freeze-thaw stability to the product. High amylopectin starches have been investigated for the production of food foam formulations (Shogren et al., 2002). Finally, the high phosphate content of potato starch gives very clear and viscous gels (Hoover, 2001; Viksù-Nielsen et al., 2001) which are important characteristics in many food systems. Biotechnology can contribute to fine tune the native functionalities or to achieve more extreme or unique combinations of functionalities not found in existing starches. It is a future challenge to produce starch functionalities with combined tasty and healthy properties. One example could be the generation of a starch based -glucan which behaves as a soluble fibre. An extremely phosphate substituted starch could show impaired enzyme degradability combined with a high solubility provided by the phosphate groups. The strategy for achieving this goal involves the overexpression in the starch crop of the starch phosphorylating enzyme GWD (Blennow et al., 2002). Another example is the generation of a starch with combined gel clarity, high gelling capacity and freeze-thaw stability, a combination that is inherently impossible to achieve for existing natural unmodified starches.

3.6

Ensuring successful modification of starch

Many fundamental problems can be identified that impede a simple straightforward and predictable strategy for biotechnological modification of starch. Many of the problems originate from the fact that starch biosynthesis, starch granule self-assembly and the resulting starch functionalities are highly cooperative processes at the various levels. Many of these effects have been exemplified previously and are also outlined in Fig. 3.3 and Table 3.1. To date, the following circumstances can be identified that hamper a clear-cut interpretation of a given single genetic alteration. 1.

Alteration of starch biosynthesis, being a primary metabolism, inevitably will result in profound changes in the crop's life-cycle. Generally, metabolic pathways cannot be regarded as separate units but rather metabolic grids or networks. The impact of one enzymatic step inevitably results in a cascade of linked metabolic side effects and altered expression of other enzymes. These are referred to as pleiotropic effects.

ß 2004, Woodhead Publishing Limited

2.

3.

4.

5. 6.

7.

8.

9.

The product of one enzyme can be the substrate for the same enzyme which may or may not act possessively (on the same substrate molecule) on the polymer substrate. Moreover, both substrates and products are highly polydisperse. The enzymes of starch synthesis act interdependently, i.e., the activity of an enzyme in starch biosynthesis is highly dependent on the -glucan product produced by the activity of the previous enzyme in the pathway. Starch metabolism is a good example of such a metabolic network in which the enzyme-catalysed reactions are intimately related. The mechanism of the self-assembly of the starch granule is not understood. The enzymes of starch biosynthesis act on polymeric substrates for which polymeric products are generated, which firstly are not monodisperse and secondly go from soluble to partly crystalline states during biosynthesis. Hence, it is a physically dynamic system where one has to consider the distribution of substrates and products in time and space. The biosynthetic mechanisms are not known for all relevant structural parameters, e.g., molecular mass of the starch molecules. The activities of the biosynthetic enzymes vary over time. The enzyme activities may oscillate diurnally, e.g., GBSS, (Tenorio et al., 2003), or change with development, e.g., SBE (Morell et al., 1997) the starch-storing organ. Most likely, not all the enzymes that affect starch deposition have been identified. The most recent enzyme in starch biosynthesis, the starch phosphorylating enzyme (GWD), was reported first in 1998 (Lorberth et al., 1998) and new isoforms of well known enzymes are continuously being discovered, e.g., the GWD homologues in Arabidopsis (Yu et al., 2001). Environmental conditions during growth affect starch functionality by altered metabolism and starch synthesis. Some structural elements such as amylose content and phosphate content may be highly dependent upon growth temperature (Tester et al., 1999; Nikuni et al., 1969; Tester, 1997; Tester and Karkalas, 2001; Haase and Plate, 1996; Burrell, 2003). Using barley and potato model plant systems the growth temperature, without detectable effects on the chemical structure of the starch, changed the crystallinity and gelatinisation behaviour of the starch (Tester, 1997; Tester and Karkalas, 2001) an effect referred to as `in vivo annealing'. Other examples are the physicochemical properties demonstrated to be altered during growth of the potato (Madsen and Christiansen, 1996; Liu et al., 2003), the phosphate content in the starch shown to be dependent on the extent of phosphate fertiliser given (Jacobsen et al., 1998) and altered, e.g., malting and feed quality of barley demonstrated to be dependent on the barley growth conditions (Molina-Cano et al., 1997). Starch structure-function models and food models are not yet developed to the level where functionality can be predicted from structure only. This will involve a very detailed understanding of the relationships between chemical structure and physical chemistry of different starch systems (Parker and

ß 2004, Woodhead Publishing Limited

Ring, 2001) and the effects of minor constituents of starch preparations such as proteins, lipids and salts. For food models, the functionality of the starch in mixed systems must be established and generalised. Clearly, the complexity of the starch system at the genetic, biosynthetic, chemical and physical levels requires multidisciplinary efforts. In spite of insufficient data, schematic models attempting to describe the highly multivariate processes of starch biosynthesis have been put forward for, e.g., amylopectin biosynthesis (Mouille et al., 1996; Ball et al., 1996; Nakamura, 2003) and amylose biosynthesis (Ball et al., 2000; Denyer et al., 2001). Even though these models cannot be used as predictive models for the generation of specific starch structures, they are very useful working models for inspiring the development of new biotechnological strategies for starch modifications. Prior to the generation of more detailed models, much more inductive and non-biased, yet standardised, strategies need to be developed in order to identify only the entities that are most important for alterations of starch deposition and functionality for a given model system. This strategy involves the analysis of mutants and transgenic plants, many of which already have been generated, using efficient data mining tools including multivariate analysis of the starch structural phenome and for prediction of starch structure±functionality relationships. It also involves a quite extensive data reduction, and only the indispensable enzyme reactions and starch structural elements required to build starch granules with high yield and desired functionalities shall be identified and their contributions quantified for given starch crop models and for functionality models that are most easily generalised. The necessity of balancing the key enzymes AGPase, SS, SBE, DBE and GWD and resulting change in starch structural motifs has already been addressed. The requirements for proper organisation of the amylopectin into a structured constant 9±nm concentric repeated lamellar body seems to be inherently confined by the very simple basic structures of starch, i.e., the glucose units linked by only -1,4 and -1,6, glucosidic linkages and the generated chain length pools including the essential absence of chains below DP 6, and the three principal chain length profiles disclosed by principle component analysis (Blennow et al., 2000). Finally, amylose and starch phosphate seem unnecessary to build a starch granule but are of immense importance to achieve specific functionalities and should consequently be included in starch structure±functionality models. In order to provide a tool to describe or quantify the effects of a given genetic alteration of the major part of the metabolome for different isogenic mutants or transgenic plants, multivariate analysis was recently introduced (Roessner et al., 2001). With this strategy, phenotyping based on the plant metabolome, i.e., the quantified pool of metabolites in the plant, is accomplished and the plant lines grouped with respect to their metabolite levels. Biopolymers are not normally included in the metabolome analysis. However, by including starch in the metabolome, combined with data on starch biosynthesis, final structures and self-assembly mechanisms phenotyping of starch crops is performed using the

ß 2004, Woodhead Publishing Limited

same approach. Likewise, using the same strategy, functionality modelling can be applied to generate models for prediction of functionality from structure. These models would be important contributions, or even requirements, for registration of genetically modified crops for consumption in the future since it provides a quantification of the plant's phenotype. Multivariate analysis or chemometrics has recently been employed to study and predict structural and functional properties of complex starch systems (Fredriksson et al., 1998; Blennow et al., 2000; Thygesen et al., 2001, 2003) including food models (Wischmann et al., 2002). Hence, the starch synthesising machinery must be analysed and evaluated at all levels of its organisation in order to provide an understanding of, e.g., the pleiotropic effects by providing a cross-talk between competences within different functional genomics disciplines. However, and most important the generated models will permit the prediction and development of biotechnological strategies from the structure±functionality data to create novel functionalities in between, or extending, from the model data set.

3.7

Future trends

After a period of rather exaggerated optimism for applied gene technology, a more realistic awareness of the power of this technology is progressively evolving. The time frame of producing accepted crops has been highly underestimated and the complexity of starch metabolism has not been sufficiently understood. Through modern genetics the principal mechanisms of starch metabolism and metabolism in general are assessed. Direct use of transgenic technology to produce novel starch functionalities as well as providing know-how generated from gene technological achievements to assist classical starch crop breeding, are currently of equal importance to generate the desired starch functionalities in the relevant crops. The balance between using classically bred crops, mutant crops and transgenic crops. A possible shift of this balance towards using transgenic crops is indicated by the transgenic generation of high and low amylose potato starch functionalities which are substantially equivalent to mutant starches, e.g., waxy and high amylose maize mutant starches being generally accepted for consumption. The use of genetic engineering to generate improved crops is continuously subjected to thorough and intense debate (Uzogara, 2000; Falk et al., 2002; Sharma et al., 2002). It is, however, interesting to note that the products of modern genetics are actually much less complex and more directed than the partly unpredictable domestication processes that took place thousands of years ago starting with teosinte, the ancestor of modern maize (Falk et al., 2002), emmer that was domesticated to yield modern wheat (Peng et al., 2003), and ancient Solanum tubers, that have been selected and bred to obtain our modern potato varieties (Van den Berg et al., 1996). This would suggest that the application of modern genetics for the production of improved starch crops has a

ß 2004, Woodhead Publishing Limited

future and could be accepted along with the increased public awareness of the risks and benefits for this new technology. However, the driving force will certainly be the increasing demands for healthier foods and, at the global scale, starch crops that can be grown and give high yields under harsh growth conditions, e.g., in arid areas (Burrell, 2003). Increased public awareness and often relevant criticism of applied modern genetics are opening up debates on food and environmental safety which hopefully will not only be restricted to genetics but also result in a general discussion and a re-evaluation on how to secure a future production of high quality functional foods both for the industrial and for the developing parts of the world. Most importantly, the consumers must be provided with safe products with significantly improved functionality at reasonable cost. Modern genetics could not only fulfil these demands but the debate on genetically modified plants can also help to set a new higher standard for food safety regulations in general. Ongoing research programmmes continuously accumulate the knowledge necessary to allow for the predictable design of tailored starch biopolymers with well-defined functional properties in planta by using first principles. The first set of transgenic plants is already on the market. Those involve crops that are resistant towards diseases and herbicides and have resulted in a substantial reduction in the use of herbicides and insecticides. The second wave of plants will be modified with respect to output traits (Willmitzer, 1999) and will include crops with improved starch functionality and yield as described in this chapter. The current strategies mentioned for engineering specific starch motifs will be refined further. For example, amylolytic activities or priming activities should be explored as a means to engineer starch molecular mass, enzyme engineering enabling GWD to catalyse phosphoryl transfer to linear chains to generate polymers forming stronger and clearer gel networks, more efficient SBE generating more soluble starches, etc. Starch functionality modelling will help to disclose similar functionality for different polymers, e.g., starch vs. gelatin systems. Together with advanced starch processing technology these new biopolymers can be integrated into the processing of improved foods. Having established general starch structure-functionality models, there are in principle possibilities for screening mutant and transgenic collections for novel starch functionality. High throughput screening technology for starch structure is a requirement to efficiently identify interesting phenotypes in randomly generated mutant and `wild' collections. While the genome is now readily databased and analysed by bioinformatics, the phenome is still a rather undefined unit. However, recent results (Munck et al., 2001) indicate the possibility to `quantify the phenome' in barley using spectroscopy combined with extraction of data by multivariate data analysis, yielding spectroscopic indicators that allow deduction of the starch structural effect of a target genetic alteration. Specifically, in situ spectroscopic techniques for high-throughput analysis of carbohydrate structure, methods for determining starch functional properties, and data-mining tools are being developed (e.g. Thygesen et al., 2001) and can be applied to mutants and registered varieties that are established

ß 2004, Woodhead Publishing Limited

for different applications. The analytical methods from genes to starch, when combined with data-mining software, will enable the targeting of key genes and the identification of their genetic positions associated with desired starch functionalities. However, this approach will also efficiently disclose, at the phenome level, structure, functionality and interaction of pools of other carbohydrate biopolymers and storage protein. Of specific interest for functional foods are the intricate balances between starch content/structure, -glucan deposition, and storage protein composition, which is directly related to the different performances of barley in food and feed applications. By aiming at isolating and breeding for new functionalities, it can be more efficient to digitise and screen the phenome rather than the genome.

3.8

References

and KOSSMANN J (1996), `Cloning and functional analysis of a cDNA encoding a novel 139 kDa starch synthase from potato (Solanum tuberosum L.)', Plant J 10, 981±991. Ê MAN P (2002), ANDERSSON L ANDERSSON R ANDERSSON R E RYDBERG U LARSSON H and A `Characterisation of the in vitro products of potato starch branching enzymes I and II', Carbohydr Polym 50, 249±257. BALDWIN P M (2001), `Starch granule-associated proteins and polypeptides: a review', Starch/StaÈrke 53, 475±503. BALL S GUAN H P JAMES M MYERS A KEELING P MOUILLE G BULEÂON A COLONNA P and PREISS J (1996), `From glycogen to amylopectin: A model for the biogenesis of the plant starch granule' , Cell 86, 349±352. BALL S DE WAL M and VISSER R G F (2000), `Progress in understanding the biosynthesis of amylose', Trends Plant Sci 3, 462467. BHATTACHARYYA M MARTIN C and SMITH A (1993), `The Importance Of Starch Biosynthesis In the Wrinkled Seed Shape Character', Plant Mol Biol 22, 525±531. BHATTY R S (1999), `The potential of hull-less barley', Cereal Chem 76, 5, 589±599. BLENNOW A BAY-SMIDT A M WISCHMANN B OLSEN C E and MéLLER B L (1998), `The degree of starch phosphorylation is related to the chain lenght distribution of the neutral and the phosphorylated chains of amylopectin', Carbohydr Res 307, 45±54. BLENNOW A ENGELSEN S B MUNCK L and MéLLER B L (2000), `Starch molecular structure and phosphorylation investigated by a combined chromatographic and chemometric approach', Carbohydr Polym 41, 163±174. BLENNOW A BAY-SMIDT A M and BAUER R (2001), `Amylopectin aggregation as a function of starch phosphate content studied by size exclusion chromatography and on-line refractive index and light scattering', Int J Biol Macromol 28, 409±420. BLENNOW A ENGELSEN S B NIELSEN T H and BAUNSGAARD L (2002), `Starch phosphorylation Ð a new front line in starch research', Trends Plant Sci 7, 445±450. BLENNOW A HANSEN M SCHULZ A JéRGENSEN K DONALD A M and SANDERSON J (2003), `The molecular deposition of transgenically modified starch in the starch granule as imaged by functional microscopy', J Struct Biol (in press). BOGRACHEVA T Y CAIRNS P NOEL T R HULLEMAN S WANG T L MORRIS W J RING S G and HEDLEY C L (1999), `The effect of mutant genes at the r, rb, rug3, rug4, and lam loci ABEL G J W SPRINGER F WILLMITZER L

ß 2004, Woodhead Publishing Limited

on the granular structure and physico-chemical structure of pea seed starch', Carbohydr Res 39, 303±314. BOYER C D DANIELS R R and SHANNON J C (1976), `Abnormal starch granule formation in Zea mays L. endosperms possessing the amylose extender mutant', Crop Sci 16, 298±301. BURRELL M M (2003), `Starch: the need for improved quality or quantity ± an overview' J Exp Bot, 54 (382), 451±456. CHEETHAM N W H and TAO L (1998), `Variation in crystalline type with amylose content in maize starch granules: an X-ray powder diffraction study', Carbohydr Polym 36, 277±284. CLARK A H GIDLEY M J RICHARDSON R K and ROSS-MURPHY S B (1989), `Rheological studies of aqueous amylose gels: The effect of chain length and concentration on gel modulus', Macromolecules 22, 346±351. CLARKE B R DENYER K JENNER C F and SMITH A M (1999), `The relationship between the rate of starch synthesis, the adenosine 5'-diphosphoglucose concentration and the amylose content of starch in developing pea embryos', Planta 209, 324±329. CRAIG J LLOYD J R TOMLINSON K BARBER L EDWARDS A WANG T L MARTIN C HEDLEY C L and SMITH A M (1998), `Mutations in the gene encoding starch synthase II profoundly alter amylopectin structure in pea embryos', Plant Cell 10, 413±426. CRITCHLEY J H ZEEMAN S TAKAHA T SMITH A M and SMITH S M (2001), `The role of disproportionary enzyme in starch breakdown is revealed by a knock-out mutation in Arabidopsis', Plant J 26, 89±100. DEIS R C (1998). `The new starches', Food Prod Design Feb, 34±58. DENYER K BARBER L M BURTON R HEDLEY C L HYLTON C M JOHNSON S JONES D A MARSHALL J SMITH A M TATGE H TOMLINSON K and WANG T L (1996), `The isolation and characterization of novel low-amylose mutants of pisum sativum L', Plant Cell & Environment 18, 1019±1026. DENYER K WAITE D MOTAWIA M S MéLLER B L and SMITH A M (1999), `Granule-bound starch synthase I in isolated starch granules elongates malto-oligosaccharides processively', Biochem J 340, 183±191. DENYER K JOHNSON P ZEEMAN S and SMITH A M (2001), `The control of amylose synthesis', J Plant Physiol 158, 479±487. DINGES J R COLLEONI C JAMES M G and MYERS A M (2003), `Mutational analysis of the pullulanase-type debranching enzyme of maize indicates multiple functions in starch metabolism', Plant Cell 15, 666±680. È SSNER U MARTIN C and EDWARDS A FULTON D C HYLTON C M JOBLING S A GIDLEY M J RO SMITH A M (1999), `A combined reduction in activity of starch synthases II and III of potato has novel effects on the starch of tubers', Plant J, 17, 251±261. FALK M C CHASSY B M HARLANDER M H HOBAN T J MCLOUGHLIN M N and AKHLAGI A R (2002), `Food biotechnology. Benefits and concerns', J Nutr 132, 1384±1390. FLIPSE E, SUURS L, KEETELS C J A M, KOSSMANN J, JACOBSEN E and VISSER R G F (1996), `Introduction of sense and antisense cDNA for branching enzyme in the amylosefree potato mutant leads to physico-chemical changes in the starch', Planta 198, 340±347. Ê MAN P (1998), `The influence FREDRIKSSON H, SILVERIO J, ANDERSSON R, ELIASSON A C and A of amylose and amylopectin characteristics on gelatinization and retrogradation properties of different starches', Carbohydr Polym 35, 119±134. FULTON D C EDWARDS A PILLING E ROBINSON H L FAHY B F SEALE R KATO L DONALD A M GEIGENBERGER P MARTIN C

and

ß 2004, Woodhead Publishing Limited

SMITH A M

(2002), `Role of granule-bound starch

synthase in determination of amylopectin structure and starch structure morphology in potato', J Biol Chem 277, 10834±10841. GEIGENBERGER P HAJIREZEAI M GEIGER M DEITING U SONNEWALD U and STITT M (1998), `Overexpression of pyrophosphatase leads to increased sucrose degradation and starch synthesis, increased activities of enzymes for sucrose-starch interconversions, and increased levels of nucleotides in growing potato tubers', Planta 205, 428±437. GIDLEY M J and BULPIN P V (1989), `Aggregation of amylose in aqueous systems: The effect of chain length on phase behaviour and aggregation', Macromolecules 22, 341±346. GOFF S A RICKE D LAN T H PRESTING G WANG R DUNN M GLAZEBROOK J SESSIONS A OELLER P VARMA H HADLEY D HUTCHINSON D MARTIN C KATAGIRI F LANGE B M MOUGHAMER T XIA Y BUDWORTH P ZHONG JP MIGUEL T PASZKOWSKI U ZHANG S P COLBERT M SUN W L CHEN L L COOPER B PARK S WOOD TC MAO L QUAIL P WING R DEAN R YU Y S ZHARKIKH A SHEN R SAHASRABUDHE S THOMAS A CANNINGS R GUTIN A PRUSS D REID J TAVTIGIAN S MITCHELL J ELDREDGE G SCHOLL T MILLER R M BHATNAGAR S ADEY N RUBANO T

and BRIGGS S (2002), `A draft sequence of the rice genome (Oryza sativa L. ssp. Japonica)', Science 296, 92±100. GUAN H and KEELING P (2000),`Starch biosynthesis: Understanding the functions and interactions of multiple isozymes of starch synthase and branching enzyme' Trends in Glycosci Glycotechnol 10, 307±319. HAASE N U and PLATE J (1996), `Properties of potato starch in relation to varieties and environmental factors' Starch/StaÈrke 48, 167±171. HOOVER R (2001), `Composition, molecular structure, and physicochemical properties of tuber and root starches: a review', Carbohydr Polym 45, 253±267. HSIEH J S LIN CL HUANG Y C and HSING Y C (1996), `An EMS-induced sorghum waxy mutant that still express the 60 kD waxy protein', Plant Physiol 111, 741. HYLTON C M DENYER K KEELING P L CHANG M T and SMITH A M (1996), `The effect of waxy mutations on the granule-bound starch synthases of barley and maize endosperms', Planta 198, 230±237. IMBERTY A and PEÂREZ S (1988) `A revisit to the three-dimensional structure of B-type starch', Biopolymers 27, 1205±1221. IMBERTY A CHANZY H PEÂREZ S BULEÂON A and TRAN V (1988), `The double-helical nature of the crystalline part of A-starch', J Mol Biol 201, 365±378. JACOBSEN H B MADSEN M CHRISTIANSEN J and NIELSEN T H (1998), `The degree of starch phosphorylation as influenced by phosphate deprivation of potato (Solanum tuberosum L.) plants', Potato Res 41, 109±116. JACOBSON M R, OBANNI M and BEMILLER J N (1997), `Retrogradation of starches from different botanical sources', Cereal Chem 74, 511±518. JAMES M G DENYER K and MYERS A M (2003), `Starch synthesis in the cereal endosperm', Curr Opin Plant Biol 6, 215±222. JANE J-L and CHEN J F (1992), `Effect of amylose molecular size and amylopectin branch chain length on paste properties of starch', Cereal Chem 69, 60±65. JENKINS, P J CAMERON, R E and DONALD A M (1993), `A universal feature in the structure of starch granules from different botanical scources' Starch/StaÈrke 45, 417±420. JENNER H L WINNING B M MILLAR A H TOMLISON K L LEAVER C J and HILL S A (2001), `NAD malic enzyme and the control of carbohydrate metabolism in potato tubers', Plant Physiol 126, 1139±1149. TUSNEEM N ROBINSON R FELDHAUS J MACALMA T OLIPHANT A

ß 2004, Woodhead Publishing Limited

and (1999), `A minor form of starch branching enzyme in potato (Solanum tuberosum L.) tubers has a major effect on starch structure: cloning and characterisation of multiple forms of SBE A', Plant J 18, 163±171. JOBLING S A WESTCOTT R J TAYAL A JEFFCOAT R and SCHWALL G P (2002), `Production of a freeze-thaw stable potato starch by antisense inhibition of three starch synthase genes', Nature Biotechnol 20, 295±299. KORTSEE A J VERMEESCH A M DE VRIES B J JACOBSEN E and VISSER R G (1996), `Expression of Escherichia coli branching enzyme in tubers of amylose-free transgenic potato leads to an increased branching degree of the amylopectin', Plant J 10, 83±90. KOSSMANN J and LLOYD J (2000) `Understanding and influencing starch biochemistry', Crit Rev Plant Sci 19,171±226. KUBO A FUJITA N HARADA K MATSUDA T SATOH H and NAKAMURA Y (1999), `The starchdebranching enzymes isoamylase and pullulanase are both involved in amylopectin biosynthesis in rice endosperm', Plant Physiol 121, 399±409. LELIEÁVRE J and LIU H (1994), `A review of thermal analysis studies of starch gelatinization', Thermochim Acta 246, 309±315. LIGHT J M (1990), `Modified food starches: why, what, where and how', Cereal Foods World 35, 1081±1092. LIM S and SEIB P A (1993), `Location of phosphate esters in a wheat starch phosphate by 31P-nuclear magnetic resonance spectroscopy', Cereal Chem, 70, 145±152. LIN P-Y and CZUCHAJOWSKA Z (1998), `Role of phosphoros in viscosity, gelatinization, and retrogradation of starch', Cereal Chem 75, 705±709. LIU H J RAMSDEN L and CORKE H (1999), `Physical properties and enzymatic digestibility of phosphorylated ae, wx, and normal maize starch prepared at different pH levels' Cereal Chem 76, 938±943. LIU Q WEBER E CURRIE V and YADA R (2003), `Physicochemical properties of starches during potato growth', Carbohydr.Polym. 51, 213±221. LLOYD R J HEDLEY C L BULL V J and RING S G (1996), `Determination of the effect of r and rb mutations of the structure of amylose and amylopectin in pea (Pisum sativum L.)', Carbohydr Polym 29, 45±49. È TZE V and KOSSMANN J (1999), `Simulaneous antisense inhibition of LLOYD J R LANDSCHU two starch-synthase isoforms in potato tubers leads to accumulation of grossly modified amylopectin', Biochem J 338, 515±521. LORBERTH R RITTE G WILLMITZER L and KOSSMANN J (1998), Inhibition of a starch-granulebound protein leads to modified starch and repression of cold sweetening', Nature Biotechnol. 16, 473±477. MADSEN M and CHRISTIANSEN D H (1996), `Changes in viscosity properties of potato starch during growth', Starch 48, 245±249. MAES T DE KEUKELEIRE P and GERATS T (1999), `Plant tagnology', Trends Plant Sci 4, 90±96. MIKKELSEN R BAUNSGAARD L MéLLER, B L and BLENNOW A (2003), `Functional characterization of -glucan, water dikinase, the starch phosphorylating enzyme' Biochem. J. in press. MILES M J MORRIS V J ORFORD P D and RING S G (1985), `The roles of amylose and amylopectin in the gelation and retrogradation of starch', Carbohydrate Research 135, 271±281. MOATES G K NOEL T H PARKER R and RING S G (1997), `The effect of chain length and solvent interactions on the dissolution of the B-type crystalline polymorph of amylose in water', Carbohydr Res 298 327±333. JOBLING S A SCHWALL G P WESTCOTT R J SIDEBOTTOM C DEBET M GIDLEY M J JEFFCOAT B SAFFORD R

ß 2004, Woodhead Publishing Limited

and BRUFAU J (1997), `Genetic and environmental variation in malting and feed quality of barley', J Cereal Sci 25, 37±47. MORELL M K BLENNOW A KOSAR-HASHEMI B and SAMUEL M S (1997), `Differential expression and properties of starch branching enzyme isoforms in developing wheat endosperm', Plant Physiol 113, 201±208. MOLINA-CANO J-L FRANSESCH M PEREZ-VENDRELL A M RAMO T VOLTAS J

MORELL M K KOSAR-HASHEMI B CMIEL M SAMUEL M S CHANDLER P RAHMAN S BULEÂON A

and LI Z (2003), `Barley sex6 mutants lack starch synthase IIa activity and contain a starch with novel properties', Plant J, 34, 185. MORRIS D L and MORRIS C T (1939), `Glycogen in the seeds of Zea Mays (variety Bantam)', J. Biol. Chem. 130, 535±544. MOUILLE G MADDELEIN M L LIBBESART N TALAGA P DECQ A DELRUE B and BALL S (1996), `Preamylopectin processing: A mandatory step for starch biosynthesis in plants' The Plant Cell 8, 1353±1366. MUHRBECK P and ELIASSON A C (1987), `Influence of pH and ionic strength on the viscoelastic properties of starch gels ± A comparison of potato and cassava starches', Carbohydr Polym 7, 291±300. BATEY I L

MUNCK L PRAM NIELSEN J MéLLER B JACOBSEN S SéNDERGAARD I ENGELSEN S B NéRGAARD L

and BRO R (2001), `Exploring the phenotypic expression of a regulatory proteomealtering gene by spectroscopy and chemometrics', Anal Chem Acta 446, 171±186. MURATA T SUGIYAMA T and AKAZAWA T (1965), `Enzymic mechanism of starch synthesis in glutitious rice grains', Biochem Biophys Res Commun 18, 371±376. NAKAMURA A (2003), `Towards a better understanding of the metabolic system for amylopectin biosynthesis in plants: Rice endosperm as a model tissue.' Plant Cell Physiol 43, 718±725. NAKAMURA T YAMAMORI M HIRANO H HIDAKA S and NAGAMINE T (1995), `Production of waxy (amylose-free) wheats', Mol Gen Genet 248, 253. NAKAMURA Y (1996), `Some properties of starch debranching enzymes and their possible role in amylopectin biosynthesis', Plant Sci 121, 1±18. NAKAMURA Y KUBO A SHIMAMUNE T HARADA K and SATOH H (1997), `Correlation between activities of starch debranching enzyme and -polyglucan structure in endosperms of sugary-1 mutants of rice', Plant J 12, 143±153. NELSON O (1985), `Genetic control of starch synthesis in maize endosperms ± a review. In: New approaches to research on cereal carbohydrates', ed. R. D. Hill and L. Munck, Amsterdam: Elsevier, 19±28. NELSON O and PAN D (1995), `Starch synthesis in maize endosperms', Ann Rev Plant Physiol Plant Mol Biol 46, 475±496. NELSON O E and RINES H E (1962), `Enzymic deficiency in waxy mutant of maize', Biochem Biophys Res Commun 9, 297±300. NIELSEN TH WISCHMANN B ENEVOLDSEN K and MOLLER B L (1994), `Starch phosphorylation in potato tubers proceeds concurrently with de novo biosynthesis of starch', Plant Physiol 105, 111±117. NIKUNI Z HIZUKURI S KAMAGI K HASEGAVA H MORIWAKI T FUKUI T DOI K NARA S and MAEDA I (1969), `The effect of temperature during maturation period on the physicochemical properties of potato and rice starches', Mem Sci Indstr Res Osaka Univ 26, 1±27. NISHI A NAKAMURA Y TANAKA N and SATOH H (2001), `Biochemical and genetic analysis of the effects of amylose-extender mutation in rice endosperm', Plant Physiol 127, 459±472 Oct. 2001.

ß 2004, Woodhead Publishing Limited

and SUDA I (2002), `Physicochemical properties of amylose-free starch from transgenic sweet potato', Carbohydr, Polym 49, 253±260. OHTANI T YOSHINO T USHIKI T HAGIWARA T and MAEKAWA T (2000), `Structure of rice starch granules in nanometre scale revealed by atomic force microscopy', J Electron Microsc 49, 487±489. ONO T and SUZUKI H (1957), `Endosperm characters in hybrids between barley varieties with starchy and waxy endosperm', Siehen Ziho 8, 11±19. ORTEGA-OJEDA F E and ELIASSON A-C (2001), `Gelatinisation and retrogradation behaviour of some starch mixtures', Starch/StaÈrke 53, 520±529. O'SHEA M G and MORELL M K (1997), `High resolution slab gel electrophoresis of 8±amino1,3,6-pyrenetrisulfonic acid (APTS) tagged oligosaccharides using a DNA sequencer', Electrophoresis 17, 681±688. PARKER R and RING S G (2001), `Aspects of the physical chemistry of starch', J Cereal Sci 34, 1±17. PENG J H RONIN Y FAHIMA T RODER M S LI Y C NEVO E and KOROL A (2003), `Domestication quantitative trait loci in Triticum dicoccoides, the progenitor of wheat', Proc Nat Acad. Sci. Amer 100, 2489±2494.

NODA T KIMURA T OTANI M IDETA O SHIMADA T SAITO A

REGIERER B FERNIE AR SPRINGER F PEREZ-MELIS A LEISSE A KOEHL K WILLMITZER L

and KOSSMANN J (2002), `Starch content and yield increase as a result of altering adenylate pools in transgenic plants', Nature Biotechnol 20, 1256±1260. RICHARDSON P H JEFFCOAT R and SHI Y-C (2000), `High-amylose starches: from biosynthesis to their use as food ingredients', MRS Bulletin Dec., 20±24. RITTE G LLOYD J R ECKERMANN N ROTTMANN A KOSSMANN J and STEUP M (2002), `The starch-related R1 protein is an alph-glucan, water dikinase', PNAS 99, 7166± 7171. ROESSNER U LUEDEMANN A BRUST D FIEHN O LINKE T WILLMITZER L and FERNIE A R (2001), `Metabolic profiling allows comprehensive phenotyping of genetically or environmentally modified plant systems', Plant Cell 13, 11±29. SCHEIDIG A FROHLICH A SCHULZE S LLOYD J and KOSSMANN J (2002), `Down regulation of a chloroplast-targeted bet-amylase leads to a starch-excess phenotype in leaves', Plant J 30, 581±591. SCHWALL G P SAFFORD R WESTCOTT R J JEFFCOAT B TAYAL A SHI Y-C GIDLEY M J and JOBLING S A (2000), `Production of very-high-amylose potato starch by inhibition of SBE A and B', Nat Biotechnol 18, 551±554. SEVENOU O HILL S E, FARHAT I A and MITCHELL J R (2002), `Organisation of the external region of the starch granule as determined by infrared spectroscopy', Int J Biol Macromol 31, 79±85. SHARMA H C CROUCH J H SHARMA K K SEETHARAMA N and HASH C T (2002), `Applications of biotechnology for crop improvements: prospects and constraints', Plant Sci 163, 381±395. SHEWMAKER C K BOYER C D WIESENBORN D P THOMPSON D B BOERSIG M R OAKES J V and STALKER D M (1994), `Expression of Escherichia coli glycogen synthase in the tubers of transgenic potatoes (Solanum tuberosum) results in a highly branched starch', Plant Physiol 104, 1159±1166. SHI Y-C CAPITANI T TRZASKO P and JEFFCOAT R (1998), `Molecular structure of a lowamylopectin starch and other high-amylose maize starches', J Cereal Sci, 27, 289± 299. GEIGENBERGER P

ß 2004, Woodhead Publishing Limited

and TIEFENBACHER K F (2002), `Baked starch foams: starch modifications and additives improve process parameters, structure and properties', Industr Crop Prod 16, 69±79. SMITH A M NEUHAUS H E and STITT M (1990), `The impact of decreased activity of starchbranching enzyme on photosynthetic starch synthesis in leaves of wrinkled-seeded peas', Planta 181, 310±313. SONNEWALD U (1992), `Expression of E. coli inorganic pyrophosphatase in transgenic plants alters photo assimilate partitioning', Plant J 2, 571±581. STARK, D M TIMMERMAN K P BARRY G F PREISS J and KISHORE G M (1992), `Regulation of the amount of starch in plant tissues by ADP glucose pyrophosphorylase' Science 258, 287±292. STINARD P S ROBERTSON D S and SCHNABLE P S (1993), `Genetic isolation, cloning and analysis of Mutator-induced dominant antimorph of the maize amylose-extender1 locus', Plant Cell 5, 1555±1566. SWINKELS J J M (1985), `Composition and properties of commercial native starches', Starch/StaÈrke 37, 1±5. SZCZODRAK J and POMERANZ Y (1991), `Starch and enzyme-resistant starch from highamylose barley', Cereal Chem 68, 589±596. TAKAHA T YANESE M OKADA S and SMITH S M (1993), `Disproportionating enzyme (4-glucanotransferase; EC 2.4.1.25) of potato', J Biol Chem, 268, 1391±1396. TAKAHA T CRITCHLEY J OKADA S and SMITH S M (1998), `Normal starch content and composition in tubers of antisense potato plants lacking D-enzyme', (4glucanotransferase)' Planta 205, 445±451. TANG H and HILLS B P (2003), `Use of 13C MAS NMR to study domain structure and dynamics of polysaccharides in the native starch granule', Biomacromol in press. TATGE H MARCHALL J MARTIN J C EDWARDS A and SMITH A M (1999), `Evidence that amylose synthesis occurs within the matrix of the starch granule in potato tubers' Plant Cell Env 22, 543±550. Â (2003), `Oscillation of mRNA level and TENORIO G, OREA A, ROMERO J M and MEÂRIDA A activity of granule-bound starch synthase I in Arabidopsis leaves during the day/ night cycle', Plant Mol Biol 51, 949±958. TESTER R (1997), `Influence of growth conditions on barley starch properties', Int J Biol Macromol 21, 37±45. TESTER R and KARKALAS J (2001), `The effects of environmental conditions on the structural features and physico-chemical propertes of starches', Starch/StaÈrke 53, 513±519. TESTER R F DEBON S J J DAVIES H V and GIDLEY M J (1999), `Effect of temperature on the synthesis, composition and physical properties of potato starch', J Sci Food Agric 79, 2054±2051. THYGESEN L G, ENGELSEN S B, MADSEN M H and SORENSEN O B (2001), `NIR spectroscopy and partial least squares regression for the determination of phosphate content and viscosity behaviour of potato starch', J Near Infrared Spectrosc. 9, 133±139. THYGESEN L G BLENNOW A and ENGELSEN S B (2003), `The effects of amylose and starch phosphate on starch gel retrogradation studied by low-field 1H NMR relaxometry', Starch/StaÈrke 55, 241±249). È HLMANN T KAMPFENKEL K HENRICHS G and NEUHAUS H E (1998), `Altered TJADEN J MO plastidic ATP/ADP-transporter activity influences potato (Solanum tuberosum L.) tuber morphology, yield and composition of tuber starch', Plant J 16, 531±540. UZOGARA S G (2000), `The impact of genetic modification of human foods in the 21st SHOGREN R L LAWTON J W

ß 2004, Woodhead Publishing Limited

century', Biotechnol Adv 18, 179±206. and KARDOLUS J P (1996), `The wild ancestors of cultivated potato: The brevicaule-complex', Acta Bot Neerland 45, 2, 157±171. VARAVINIT S, SHOBSNGOB S, VARANYANOND W, CHINACHOTI P and NAIVIKUL O (2002), `Freezing and thawing conditions affect the gel stability of different varieties of rice flour', Starch/StaÈrke 54, 1, 31±36. VESELOVSKY I A (1940), `Biochemical and anatomical properties of starch of different varieties of potatoes and their importance for industrial purposes', Amer Pot J 17, 330±339. VIKSé-NIELSEN A BLENNOW A JéRGENSEN K KRISTENSEN K H JENSEN A and MéLLER B L (2001), `Structural, physicochemical, and pasting properties of starches from potato plants with repressed r1-Gene', Biomacromol 2, 836±843. VISSER R G F and JACOBSEN E (1993), `Towards modifying plants for altered starch content and composition. TIBTECH 11, 63±68. VISSER R G F SUURS L C J M BRUINENBERG P M BLEEKER I and JACOBSEN E (1997a), `Comparison between amylose-free and amylose containing potato starches', Starch/StaÈrke 49, 438±443. VISSER R G F SUURS L C J M STEENEKEN P A M and JACOBSEN E (1997b), `Some physicochemical properties of amylose-free potato starch', Starch/StaÈrke 49, 443± 448. WAIGH T A GIDLEY M J KOMANSHEK B U and DONALD A M (2000), `The phase transformations in starch during gelatinisation: a liquid crystalline approach' Carbohydr Res 328, 165±176. WALKER C E, ROSS A S, WRIGLEY C W and MCMASTER G J (1988), `Accelerated starch-paste characterization with the rapid visco-analyzer', Cereal Foods World 33, 491±493. WANG T L BOGRACHEVA T Y and HEDLEY C L (1998), `Starch: as simple as A, B, C?', J Exp Bot 49, 481±502. WEATHERWAX P (1922), `A rare carbohydrate in waxy maize', Genetics 7, 568±572. WIESENBORN D P, ORR P H, CASPER H H and TACKE B K (1994), `Potato starch paste behaviour as related to some physical/chemical properties', J Food Sci 59, 644±648. WILLMITZER L (1999), `Biotechnology: output traits ± the second generation of plant biotechnology products is gaining momentum', Curr Opin Biotechnol 10, 1±162. WISCHMANN B BLENNOW A MADSEN F and BANDSHOLM O (2002), `Functional properties of genetically modified potato starches', Ann. Transact. Nordic Rheol. Soc 10. YANGSHENG W U and SEIB P A (1990), `Acetylated and hydroxypropylated distarch phospates from waxy barley: Paste properties and freeze-thaw stability', Cereal Chem 67, 202±208. YU S BLENNOW A BOJK, M MADSEN F OLSEN C-E and ENGELSEN S B (2002), `Physicochemical characterisation of floridean starch of red algae', Starch/StaÈrke 54, 66± 74.

VAN DEN BERG R G GROENDIJKWILDERS N

È USLER R E HILLE D FLU È GGE H-I ZEEMAN S C SMITH A M KOSSMANN J YU T-S KOFLER H HA LLOYD J RITTE G STEUP M LUE W CHEN J and WEBER A (2001), `The Arabidopsis sex1 mutant is defective in the R1 protein, a general regulator of starch degradation in plants, and not in the chloroplast hexose transporter', Plant Cell 13, 1907±1918. YUAN R C and THOMPSON D B (1998), `Rheological and thermal properties of aged starch pastees from three waxy maize genotypes', Cereal Chem. 75, 117±123. ZEEMAN S C and AP REES T (1999), `Changes in carbohydrate metabolism and assimilate export in starch-excess mutants of Arabidopsis', Plant Cell and Env 22, 1445± 1453.

ß 2004, Woodhead Publishing Limited

and AP REES T (1998a), `A starch accumulating mutant of Arabidopsis thaliana deficient in a chloroplastic starch-hydrolysing enzyme', Plant J 15, 357±365. ZEEMAN S C UMEMOTO T LUE W-L AU-YEUNG P MARTIN C SMITH A M and CHEN J (1998b), `A mutant of Arabidopsis lacking a chloroplastic isoamylase accumulates both starch and phytoglycogen', Plant Cell 10, 1699±1611. ZEEMAN S C, NORTHROP F, SMITH A M

ß 2004, Woodhead Publishing Limited

4 Starch-acting enzymes D. P. Butler, Marc J. E. C. van der Maarel and P. A. M. Steeneken, TNO Nutrition and Food Research Institute, The Netherlands

4.1

Introduction: the importance of enzymes

Starch is one of the most abundant carbohydrates found in nature, and is a major source of nutrition for both man and animals. Its importance as a food stuff may be judged by the fact that it accounts for over 30% of the `average' diet on a dry weight basis and more than 25% on an available energy basis (Galliard, 1989). The utilisation of starch and starch derivatives in many industrial processes, particularly food processing, has led to the development of numerous methods of starch processing and breakdown. The preparation of modified starch was originally performed by the use of chemical additives. In recent years, enzymatic methods have largely replaced the use of chemicals. This change is partly due to enzymes being safer and healthier for both the environment and consumers of starch-containing products. Enzymes are also advantageous as they perform more specific hydrolysis reactions, give higher yields and also create fewer by-products and consequently require less purification (Kennedy et al., 1985). The industrial environment in which enzymes are required to act on starch is in most cases different from the native environment where most of the enzymes that are commercially used today were originally isolated. Consequently, either the reaction conditions have to be altered to match those required by the enzyme or alternatively the enzyme itself has to be modified to function as part of the industrial process being performed. This has added further impetus to the isolation and study of new and improved starch-acting enzymes. The function of starch-acting enzymes is the digestion of starch-containing food, which is achieved by the -amylases and -glucosidases in the gastrointestinal tract. Enzymes involved in starch biosynthesis have been discussed in Chapter 1. In the present chapter we will deal with enzymes which have

ß 2004, Woodhead Publishing Limited

potential use in post-harvest modification of starch on an industrial scale. Categorised by their activity, two types of starch-acting enzymes can be distinguished. -Glucanohydrolases catalyse the splitting of -1,4- and/or 1,6-glucosidic linkages, which results in the overall depolymerisation of starch. -Glucanotransferases promote the transfer of monomers or oligomers from a glucosyl donor to a glucosyl acceptor. Donor and acceptor may be located at the same molecule or at different molecules. The number average degree of polymerisation may not change by a transfer reaction. The transfer reaction may be accompanied by a change in linkage type, e.g., in -1,4-glucosyl -1,1glucanotransferase which catalyses the formation of trehalose from starch, or it may use a non-glucose polymer or oligomer as an acceptor. Up to now, only the -glucanohydrolases have been applied on a commercial scale. Henrissat classified all enzymes which have activity on carbohydrates into a number of families according to amino acid and DNA homology. The enzymes which act on starch and its derivatives are classified into several of these families and may be viewed on the CAZY website (://afmb.cnrs-mrs.fr/ CAZY/) (Coutinho and Henrissat, 1999). Currently, this website includes 91 distinct families, of which families 4, 13, 31, 57, 70 and 77 include enzymes which show activity on starch or its derivatives. Although every family includes enzymes which show sequence similarity and contain a number of conserved amino acid residues which are specific to that family, each family also includes a number of different reaction specificities. This is also seen with the families which contain enzymes that act on starch. Family 13 includes enzymes with amylase (EC 3.2.1.1), pullulanase (EC 3.2.1.41), isoamylase (EC 3.2.1.68), glucan branching (EC 2.4.1.18) and cyclodextrin glycosyltransferase (EC 2.4.1.19) activities. Family 57 posseses -amylase (EC 3.2.1.1), -galactosidase (EC 3.2.1.22) and 4--glucanotransferase (EC 2.4.1.-) activities, while family 77 includes amylomaltase and 4--glucoantransferase (EC 2.4.1.-) activities. Family 4 contains many activities including maltose-6-phosphate glucosidase (EC 3.2.1.122), -glucosidase (EC 3.2.1.20), -galactosidase (EC 3.2.1.22), 6phospho- -glucosidase (EC 3.2.1.86) and -glucuronidase (EC 3.2.1.139). Family 31 includes the following activities -glucosidase (EC 3.2.1.20); glucoamylase (EC 3.2.1.3); sucrase-isomaltase (EC 3.2.1.48) (EC 3.2.1.10); -xylosidase (EC 3.2.1.-); -glucan lyase (EC 4.2.2.13); isomaltosyltransferase (EC 2.4.1.-) (Fig. 4.1). In this chapter, we briefly address the industrially important starch hydrolysing enzymes and their use in achieving functional changes in starch and starch-based foodstuffs. We will also examine the potential of glucanotransferases in the design of modified starches with special functionality. The second part of this chapter is devoted to molecular biological strategies and techniques employed to obtain enzymes with novel starch-modifying activities. In particular the potential implications of the availability of increasing numbers of genome sequences will be assessed. A number of methods which aid in the successful prediction of the function of proteins will also be described. In addition, recent advances in both high throughput screening and protein

ß 2004, Woodhead Publishing Limited

Fig. 4.1 The enzymatic hydrolysis of starch and its derivatives. For reasons of clarity not all activities and products are included.

engineering will be addressed with specific emphasis on their potential to create enzymes with novel activities on starch. Examples of both gene shuffling and directed evolution will be presented, demonstrating that both these technologies have the scope to create new enzymatic entities with possible applications in the field of starch modification.

4.2

Using enzymes to modify starch

4.2.1 Starch hydrolysing enzymes A very significant amount of the world production of starch, which amounts to at least 45106 tonnes/year (Gordon, 1999) is converted into starch hydrolysis products (SHP). At least two textbooks on the production of SHP have appeared (Kearsley and Dziedzic, 1995; Schenck and Hebeda, 1992). Originally, the production of SHP almost always involved acid hydrolysis. In a typical operation, a 30±40% starch slurry is acidified to pH 1.5±2, converted at 140± 160 ëC for a specified time depending on the desired degree of hydrolysis, cooled, and neutralised. The degree of hydrolysis is expressed as dextrose equivalent (DE), which is defined as the amount of reducing sugar with respect to the total amount of starch carbohydrate on a weight basis. For a maltooligosaccharide with degree of polymerisation p, DE ˆ 18000/ [180 + 162 (p1)]. Conversion of starch to a rather low DE value is referred to as liquefaction, whereas saccharification signifies the further hydrolysis to a high DE. A feature of acid conversion is that glucosidic bonds are cleaved randomly, resulting in a fixed maltooligomer distribution at a given DE. An advantage is that high starch concentrations can be used. On the other hand, colour formation occurs due to glucose dehydration reactions, which become more severe at increasing DE. As a consequence of this discoloration and retrogradation, acid conversion is only suitable for the production of syrups of DE 30±55. The major grade of DE 42 is still predominantly produced by this method while higher and lower DE SHPs are never made by acid conversion alone. -Amylases (1,4--D-glucan glucanohydrolases) catalyse the cleavage of arbitrary -1,4-glucosidic linkages without affecting -1,6-linkages. An excellent overview on the -amylases and their applications has appeared recently (Van der Maarel et al., 2002). The action pattern of -amylase is not random but depends on its origin, more specifically on the spatial conformation of its active site (Robyt, 1985). This implies that SHP oligosaccharide composition at a given DE may vary with the enzyme used, which allows for much greater flexibility in SHP production. Nevertheless, it took a long time before amylases superseded acid conversion in the liquefaction process. The main reason for this was that the first available amylases were not active at the temperature conditions necessary for the efficient gelatinisation and liquefaction of 35% starch suspensions and were not able to prevent recrystallisation of partially degraded starch. The introduction of the highly thermostable amylases

ß 2004, Woodhead Publishing Limited

from Bacillus licheniformis (Madsen et al., 1973) and Bacillus stearothermophilus (Reeve, 1992; Tamuri et al., 1979) allowed the application of continuous cooking equipment for the liquefaction of concentrated starch suspensions without viscosity problems and the risk of recrystallisation. This was the breakthrough necessary for the development of totally enzyme-based processes. These enzymes have their pH optimum of 6.0 and require 20±100 ppm Ca2+. The most used saccharification enzyme is amyloglucosidase or glucoamylase (AMG, glucan 1,4--glucosidase), usually from Aspergillus niger. This enzyme cleaves successive glucosidic linkages, both -1,4 and -1,6, starting from the non-reducing end of the malto-oligosaccharide molecules to a DE of 95 and higher with no Ca2+ requirement. Typical process conditions are pH 4±4.5 and 60 ëC. A recent development in saccharification is the use of Bacillus megaterium amylase, which combines transferase with endo-amylase activity. As well as cleaving -1,4 bonds, this enzyme catalyses the transfer of small oligosaccharides to glucose, thereby simultaneously providing an increase in non-reducing end groups which act as a substrate for AMG. It also results in a temporary decrease of glucose concentration to diminish product inhibition and reversion reactions. As a result, the final DE is increased by several percent (David et al., 1987; Hebeda et al., 1988; Reeve, 1992). Other developments include the introduction of -amylases and AMG's with pH optima in the range 5±5.5, making pH adjustments during processing nearly unnecessary (Shetty and Allen, 1988; Mitchinson et al., 1996). -Amylases with no or limited Ca2+ requirement are also now available (Norman et al., 1997; Peckous, 1999). In addition, AMGs with higher thermostability (Chen et al., 1994) and less tendency for reversion reactions (Fang et al., 1998a,b; Liu et al., 1998), and the use of pullulanases and isoamylases, which assist in the cleavage of branching points, are interesting advances. The latter two enzymes may also be used without -amylase for the production of amylodextrin, a mixture of short and long linear starch molecules which are obtained by the debranching of starch. This subject will be discussed further below. Almost all of these improvements have been attained by making use of the molecular biological techniques to be discussed in Section 4.3. High-glucose syrups are predominantly used for the production of highfructose syrups (HFS), especially in the USA. The breakthrough was marked by the development of immobilised glucose-isomerases at the end of the 1960s (MacAllister, 1980; Takasaki et al., 1969; Takasaki and Tanabe, 1971; White, 1992). The most utilised enzymes originate from Bacillus coagulans and several Streptomyces species. All enzymes require Co2+ or Mg2+ as a co-factor. Because Ca 2+ inhibits enzyme activity, the glucose feed streams have to be demineralised. In addition the DE of the substrate syrup should to be as high as possible. The basic isomerisation product contains 42% fructose. This in turn is enriched by an anion exchange chromatographic separation which results in a 90% fructose syrup, while glucose is returned to the feed stream. By blending, a 55% fructose syrup is obtained as the major commercial product (HFS55).

ß 2004, Woodhead Publishing Limited

High-maltose syrups are produced from the liquefact by the use of maltogenic -amylases, e.g., from Aspergillus oryzae which operates at pH 5 and 55 ëC (Reeve, 1992). At the lower end of the dextrose equivalent range (DE 1±20), the maltodextrins are produced by controlled -amylolysis at 70±90 ëC using a bacterial amylase. A heating step at 100 ëC is applied to gelatinise the starch completely and to inactivate the enzyme. On cooling, the hydrolysate can be subjected to further amylolysis to attain the desired DE (Alexander, 1992; Armbruster, 1974; Kennedy et al., 1995; Morehouse et al., 1972). Alternative processes, wherein maltodextrins are produced by liquefaction prior to controlled hydrolysis are also described (Alexander, 1992).

4.2.2 Overview of properties and applications of SHPs Glucose syrups include SHPs with a medium range of between 20 and 80 DE and cover a wide range of compositions, properties, and applications. One of the most important applications of SHPs, sucrose-based confectionery, will now be discussed. In hard candies, substitution of part of the sucrose by glucose syrups diminishes the tendency of sucrose to crystallise (graining) during production and storage. At the same time this allows a higher total sugars content in the product. The mechanism of graining has been explained by Cakebread (1972), who describes this phenomenon as an autocatalytic process. Because of the relatively low water activity (0.53±0.65 in hard candies), water is taken up. As a consequence, viscosity is lowered and sucrose molecules are mobilised which are then able to crystallise. As the sucrose crystals are anhydrous, water bound to the sucrose molecules is liberated and becomes available for further mobilisation of sucrose. This also explains why graining is associated with increased stickiness and why graining starts at the surface, as it is there that moisture sorption takes place (Kit, 1993). Remedies against graining include lowering the sucrose concentration and increasing the aw and viscosity. The first one is achieved by replacing sucrose by so-called `doctor sugars'; glucose syrups or invert sugar. These increase the saturation concentration of the sugar, probably by lowering the concentration of each individual sugar, and thereby its nucleation tendency. Nevertheless, even with high glucose/invert sugar additions, supersaturation remains (Keysers, 1982). Moreover, aw is lowered by the addition of monosaccharides. An increase of aw can be achieved by partial replacement of sucrose by starch-derived syrups with lower DE. High molar mass sugars offer a number of benefits, which include a higher aw, reduced stickiness, a higher viscosity, a higher Tg, and a reduced tendency to crystallise (CPC, 1986). This is offset by more difficult handling and reduced sweetness (effects on taste). In fondants, glucose syrups are used to obtain a fine grained crystalline texture. Sweetness, water activity (aw), and product texture are strongly related to sugar composition. Sweetness increases with DE. High DE glucose sugars have a low aw and hence tend to attract moisture during storage. The reverse is

ß 2004, Woodhead Publishing Limited

true for low DE sugars. Product stability in a general sense is governed by the glass transition temperature Tg, which is inversely related to DE (Roos, 1995). Below the Tg, large-scale mobility is suppressed. This prevents molecular reorganisation resulting in crystallisation and it delays water migration. This formulates low DE SHPs which add to product stability (Levine and Slade, 1984), However, it has to be borne in mind that products below Tg are essentially brittle. Tg is strongly affected by plasticisers, e.g., product water and glycerol. Low molar mass sugars in the product also act as plasticisers, though less effective than water. On the other hand, stability to microbial deterioration is enhanced by a low aw. The art of confectionery making is to reconcile the often conflicting demands regarding taste, texture, and shelf-life (moisture control and crystallisation inhibition) by a proper formulation of the ingredient mix. At the high end of the DE range we find the syrups with a high glucose content. These are predominantly used for the production of HFSs. The major application of HFS55 is as a replacer of 50% of the sucrose in carbonated drinks, especially in the USA (White, 1992). This is probably the largest single-outlet application of starch-based products. The main application of maltodextrins is as a bulking agent and for flavour encapsulation, where it has replaced gum arabic to a large extent (Alexander, 1992). Dried flavours are generally prepared by spray-drying emulsions containing the flavour and maltodextrin. By modifying maltodextrin by hydrophobic groups, e.g., octenyl succinate, or combining it with a hydrophobic starch, an emulsifier may be omitted (Alexander, 1992). The interesting, recently developed application in fat replacement relies on the observation that low DE maltodextrins at moderate concentration (>10%) form thermoreversible gels after heating and cold storage, i.e., these gels melt at temperatures below 100 ëC (Richter et al., 1975). The gels have a particulate structure and provide a spreadable rather than a rubber-like texture, which provides a fat-like mouthfeel (Kaper and Gruppen, 1987). Potato starch is the favoured source of gelling maltodextrins, because the lower degree of branching results in a larger proportion of linear segments, which promote gelation. Amylodextrins are prepared by the debranching of starch with isoamylase or pullulanase and consist of a mixture of amylose and rather short linear maltooligosaccharide chains with a very broad molar mass distribution. The short chain amylodextrins can be prepared by selecting an amylose-free (waxy) starch as the starting material. It is known that aqueous starch solutions are inherently unstable at ambient temperature, which is caused by starch recrystallisation, also termed retrogradation (Parker and Ring, 2001). In conditions of excess water and moderate temperatures, double helical crystallites of the B-polymorph are formed. Being composed of linear chains, amylose has a much greater propensity to crystallise than amylopectin. In contrast to amylopectin, amylose crystals are thermo-irreversible, i.e., they melt only at temperatures higher than 100 ëC in the presence of water. The retrogradation rate of linear starch chains is dependent on the degree of polymerisation p and is highest at p 50±100 (PfannemuÈller et al., 1971; Gidley and Bulpin, 1989). In contrast to amylose,

ß 2004, Woodhead Publishing Limited

where retrogradation results in gel formation at concentrations exceeding 1% (Miles et al., 1984), amylodextrin chains are generally too short for gelation and semi-crystalline particles are obtained. Two applications of amylodextrin will be discussed. Because of its crystalline nature, amylodextrin escapes digestion in the gut to a rather large extent, i.e., it contains a large amount of resistant starch. The main nutritional property of resistant starch is that it is available for fermentation in the colon. A second application is as an excipient in pharmaceutical tablets. Excipients are pharmaceutically inactive. Their function as a drug carrier is to maintain tablet integrity and stability, while also assisting drug release. Tablets containing amylodextrin show breaking strengths superior to those composed of microcrystalline cellulose, although performance is strongly dependent on the recovery of the amylodextrin crystals. Drying from the glassy state, i.e., when large-scale mobility is inhibited by freeze-drying or by exchange with nonaqueous solvents, gives the best results (Bergsma and Te Wierik, 1996). Low porosity tablets containing amylodextrin may exhibit controlled release, i.e., a release linear with time instead of the expected diffusional release with the square root of time (Besemer and Lerk, 1995; Arends-Scholte et al., 1996). This controlled release property depends on the linear migration of a solvent front into a glassy tablet matrix, where release can only start as soon as a tablet layer has been softened by the solvent (Te Wierik et al., 1997). Because this process is counteracted by polymer swelling, the best results are obtained with short chain crystalline starches which do not swell. Resistance to amylolytic degradation in the gastro-intestinal tract is warranted by the aforementioned resistant-starch like nature of amylodextrin (Guraya et al., 2001a,b). -Amylase splits maltose starting from the non-reducing ends of starch molecules. However it cannot circumvent -1,6-linkages. Therefore the net result is an overall shortening of amylopectin side-chains. Limited -amylolysis has been claimed to reduce the tendency of starch to retrograde (Wursch and Gumy, 1994). All known -amylases originate from plant sources and are not thermostable, which hampers their industrial use. Splitting of maltose from nonreducing chain ends is also a major activity of maltogenic -amylases. However these enzymes show a limited endo-activity. A thermostable enzyme has been isolated from Bacillus stearothermophilus, which is active during baking and is currently used as a commercial anti-staling enzyme in bread (Diderichsen and Christiansen, 1988; Christophersen et al., 1998). Its major effect is an in situ reduction of the exterior chain length of amylopectin.

4.2.3 Starch granule degrading enzymes Because of its semi-crystalline nature, an ungelatinised starch granule is digested much more slowly by starch hydrolysing enzymes than a gelatinised granule or a starch solution. During hydrolysis a range of different morphologies may be observed, which range from pitting of small holes, shell formation, and surface erosion. The mode of enzymatic degradation is dependent both on the

ß 2004, Woodhead Publishing Limited

starch and the type of enzyme employed (Planchot et al., 1995). Enzymatic attack on large wheat starch granules starts at the equatorial groove (Gallant et al., 1973). Potato starch is digested much more slowly than the cereal starches. Moreover, potato starch is digested by surface erosion rather than by pitting. Because of these widely different morphologies and the lack of proof of enrichment in crystallinity in extensively degraded granules it cannot be stated unequivocally that the amorphous domains of the starch granules are preferentially degraded (BuleÂon et al., 1998). The differences in digestibility and digested granule morphology may be connected to crystalline type: B in potato and A in cereal starch. Crystalline lamellar structures in potato starch are organised in larger more coherent domains when compared to cereal starches (Oostergetel and Van Bruggen, 1993), which may in part explain the different mode of granule degradation. A screening programme for micro-organisms capable of growth on ungelatinised potato starch granules yielded two different mesophilic microorganisms exhibiting amylolytic activity. They were able to degrade potato starch granules as fast as maize or wheat starch (Wijbenga et al., 1991; Wijbenga and Binnema, 1997). One of the strains was initially assigned as a Bacillus firmus/lentus, but was later identified as a Paenibacillus granivorans. The amylase complex from this strain consists of at least five different amylases, two of which were active on intact starch granules. Whilst maltopentaose was the main product when soluble starch was used as a substrate, glucose was formed almost exclusively from starch granules. In the Microbacterium aureum preparation only two amylases could be detected. The mode of degradation of the enzymes from both micro-organisms was completely different. Whereas the amylases from Paenibacillus granivorans act by surface erosion, those from Microbacterium aureum form numerous small pores (Fig. 4.2). Porous starch granules have potential as adsorbents and carriers, e.g. of flavours. As a result of hydrolysis with Microbacterium amylase, the specific surface area of potato starch granules increased by a factor ten at 40% degradation, whereas the capacity of vanillin adsorption was enhanced by a factor five. (Wijbenga and Binnema, 1997). Production of porous granules from cereal starches by

Fig. 4.2

Potato starch granules before (left) and after degradation by the amylases from Paenibacillus granivorans (middle) and Microbaterium aureum (right).

ß 2004, Woodhead Publishing Limited

enzymatic degradation has also been described. Particles with size 1±2 m could be produced from these by ball milling (Niemann and Whistler, 1992).

4.2.4 Starch transferases -1,4-Glucanotransferases catalyse the cleavage of -1,4 glucosidic bonds and the transfer of the newly formed reducing end group (donor) to a non-reducing saccharide unit (acceptor) with the formation of a new -1,4 glucosidic bond. Donor and acceptor may originate from the same type of molecule or from different types of molecules. Several action patterns may be distinguished. Firstly, a single low molar mass malto-oligosaccharide molecule may act both as donor and acceptor. In this case, cyclic saccharides are formed. Familiar examples are the cyclodextrins (CDs) with 6, 7, or 8 glucose units in the ring, which are formed by cyclodextrin glucosyltransferase. The net result of this type of activity is a decrease in the average molar mass. CDs are prepared by incubating a starch liquefact of DE < 10 with CGT-ase from Bacillus macerans (Riisgaard, 1990) or from thermostable micro-organisms like Thermoanaerobacter species (Starnes, 1990). By addition of a complexing solvent, a crystalline inclusion complex is formed that can be separated from the mother liquor and refined by redissolution in water and recrystallisation. In a nonsolvent process, the linear malto-oligosaccharides are degraded by amylolysis and cyclodextrin is recovered by partial evaporation and crystallisation. Like amylose, CDs have the ability to form crystalline complexes with a number of inclusion compounds. CDs have been widely explored as drug and flavour carriers (Szejtli, 1998). A novel development of randomly methylated -CD is its use as an aid in emulsion polymerisation, e.g. in the surfactant-free production of polystyrene and polymethylmethacrylate latex particles in aqueous media (Storsberg et al., 2003). The avoidance of surfactants leads to lattices with a narrow particle size distribution. This application exploits the complexing ability of some monomers with CD, making them water-soluble. A different situation arises when the donor is transferred either to a different acceptor molecule or to a different side chain on the same molecule, e.g., in amylopectin. An interesting example is the so-called disproportioning enzyme or D-enzyme, also named amylomaltase because it catalyses the transfer of maltose from starch to glucose. A well known source of this enzyme is the potato, but a thermostable amylomaltase has also been isolated from Thermus thermophilus (Euverink and Binnema, 1998; Van der Maarel et al., 2000). Limited action of this enzyme on gelatinised starch leads to the progressive disappearance of amylose and the formation of amylopectin with a broader chain-length distribution (Fig. 4.3). During this process the granular remnants dissolve completely and a low viscous molecular solution is obtained with a shift in iodine absorption maximum to lower wavelengths. Interestingly, these solutions form turbid rubber-like gels with a relatively high modulus on cooling at concentrations as low as 3%, which melt again on heating to c. 70 ëC (Fig. 4.4). Amylomaltase-modified starch is a potential gelatin replacer in applications

ß 2004, Woodhead Publishing Limited

Fig. 4.3 Chain-length distributions of debranched native potato starch (left) and amylomaltase-modified potato starch, as measured with high-performance anion exchange chromatography with pulsed amperometric detection. The figures correspond to the degree of polymerisation of the original side-chains.

where gel transparency is not an issue. At higher conversions with this enzyme, cyclisation reactions on both linear and branched molecules occur, leading to relatively high molar mass starch products with a limited tendency to retrograde (Takaha et al., 1996, 1997). A third situation arises when a different linkage type is formed during the transfer reaction. The action of starch branching enzyme (SBE) involves breaking of an -1,4-linkage and the concomitant formation of an -1,6-linkage.

Fig. 4.4 Response of gel modulus (G0 ) of 5% native potato starch and 5% amylomaltase-modified potato starch (AMAZ) to temperature cycling, as measured by small deformation rheology.

ß 2004, Woodhead Publishing Limited

Thermostable branching enzymes have been isolated from a limited number of sources, e.g., Bacillus stearothermophilus (Kiel et al., 1991; Takata et al., 1994). The main effect is an overall reduction of amylopectin chain length, thereby reducing the propensity of starch to retrograde and increasing the solubility of starch in water. The fact that this action also results in the elimination of amylose and is not accompanied by the formation of maltose as a by-product makes SBE more useful than -amylase. Involvement of a newly formed reducing end group into an -1,6 linkage on the same molecule will also lead to the formation of cyclic molecules (Takata et al., 1996, 1997). An example of glucanotransferases that depend on a non-starch molecule as a glucosyl donor are the phosphorylases which utilise glucose-1-phosphate (G-1P) as a donor. If a short chain malto-oligosaccharide Gn with DP4 is used as an acceptor, a linear glucan chain is formed the length of which correlates positively to the G-1-P/Gn ratio in the initial reaction mixture. A peculiar feature of this reaction is that the molar mass distribution of these glucan chains is extremely narrow, especially when high molar masses are aimed at. The transfer mechanism closely resembles that of an addition polymerisation in synthetic polymer chemistry (PfannemuÈller, 1975). A major application of phosphorylase is to produce synthetic amyloses with a precisely defined molar mass, which find application in fundamental studies. With the use of these synthetic preparations it has been proven that amylose exhibits a rather sharp peak in retrogradation rate at DP 50±100 (PfannemuÈller et al., 1971; Gidley and Bulpin, 1989). When the phosphorylase catalysed reaction is applied for the production of shorter chain amylose, the molar mass distribution is much broader (Niemann et al., 1990). Large-scale application of phosphorylase has been precluded by the high cost of G-1-P. Starch itself can also serve as an acceptor for the phosphorylase reaction. In that case starch molecules with extended side chains are formed, which tend to retrograde much more rapidly than amylose (PfannemuÈller et al., 1976).

4.2.5 Starch lyases Starch lyases (EC 4.2.2.13) catalyse the conversion of one molecule of linear 1,4-glucans containing n+2 monomers to n+1 molecules of 1,5- Anhydro-Dfructose(n+1) and one molecule of D-glucose. In 1988 the first -1,4-glucan lyase activity was identified in fungal species and later in 1993 -1,4-glucan lyase activity was purified from red seaweed (Yu and Pedersen, 1993). Starch lyases act by degrading the non-reducing ends of the starch molecule in a similar manner to that of exo-acting starch hydrolases such as amyloglucosidases and amylases. The lyase will continue to act on the reducing end of the starch molecule until it reaches either the first glucose unit at the reducing end in the case of amylose or a branching point in amlyopectin. (Yu et al., 1995). Sequence analysis of these enzymes revealed that they share no sequence similarity with other known polysaccharide lyases (EC 4.2.2.1-4.2.2.12). Significantly, however, they do show homology to the family 31 -glucosidases particularly

ß 2004, Woodhead Publishing Limited

in the segments of the enzymes which are thought to be involved in substrate binding and catalysis (Yu et al., 1999). It is therefore speculated that the mechanism of action of the starch lyases is similar to that of the -glucosidases, and that they differ in that the reaction site of the lyase is more compact resulting in absence of water molecules in the site. This absence of the water molecules then in turn leads to the lytic cleavage of the glucosidic bond as opposed to the hydrolytic cleavage preformed by the hydrolases (Yu et al., 1999).

4.2.6 Trehalose and non-reducing oligosaccharide synthesis from starch Trehalose (-D-glucopyranosyl -D-glucopyranoside) is a very unreactive sugar and its high stability is the result of the very low energy of the glycoside oxygen bond joining its two hexose rings. (Schiraldi et al., 2002). Recently a number of groups have reported the production of trehalose from starch or maltooligosaccharides. The process may be performed by the action of -1,4-D-Glucan phosphorylase on maltose producing glucose-1-phosphate followed by the subsequent action of trehalose phosphorylase on D-glucose and glucose-1phosphate producing trehalose and orthophosphate (Wada et al., 1995). This method is not suitable for industrial-scale processes as the purification methods used for the enzymes from their hosts (yeast) is complicated and does not give a high yield. Furthermore the reversibility of the reactions means that the yield of trehalose is also relatively low (Schiraldi, et al., 2002). Since then, a novel nonreducing saccharide forming enzyme (malto-oligosyl trehalose synthase) has been detected in cultures of micro-organisms as Rhizobium sp. M-11 and Arthrobacter sp. Q36 (Maruta et al., 1994b) and the thermoacidophilic archaebacteria Sulfolobus acidocaldarius and Sulfolobus solfataricus by Italian, Japanese, and Dutch groups (Lama et al., 1991; Nakada et al., 1996a; Kato et al., 1996a; Binnema et al., 1996). This enzyme acts by transferring the reducing end of a malto-oligosaccharide to C1OH of glucose, thus forming a glucosyl trehalose (Kato et al., 1996b). One advantage of non-reducing malto-oligosaccharides over their reducing counterparts is that the former are much less prone to Maillard reactions with amino acids and peptides on heating and hence may find use in applications where discoloration is unwanted. In the cultures of the aforementioned micro-organisms the maltooligosyl trehalose synthase activity is accompanied by an amylase (malto-oligosyl trehalose trehalohydrolyse) activity, which catalyses the cleavage of the -1,4 linkage next to the terminal trehalose unit in the malto-oligosyl trehaloses (Maruta et al., 1994a; Nakada et al., 1996b; Kato et al., 1996c). By the concerted action of both enzymes, trehalose is formed from soluble starch or maltodextrins (Maruta, 1994a,b; Schiraldi et al., 2002). Another route towards trehalose is by the action of trehalose synthase on maltose. This enzyme was isolated from Pimelobacter sp. R48 and the more thermostable Thermus aquaticus ATCC 33923 (Nishimoto et al., 1995, 1996). Chen et al. (2000) have also reported the cloning and expression of a maltoologosyl trehalose synthase from Sulfolobus shibatae. This enzyme is able to produce trehalose from partial starch hydrolysates. De Pascale et al. (2002)

ß 2004, Woodhead Publishing Limited

created a fusion protein which was formed by the fusion of the trehalosyl dextrin forming enzymes and the trehalose forming enzyme from Sulfolobus solfataricus MT4. This enzyme can produce trehalose from maltooligosaccharides at 75 ëC. Potential applications of trehalose include the use in living tissue preservation, as moisturiser in cosmetics, and as food processing aid in drying and concentration operations (Schiraldi et al., 2002).

4.3 Developing starch-modifying enzymes for food processing applications 4.3.1 Enzyme development: general requirements and strategies The starch industry already has a number of -amylase and -glucosidase enzyme preparations which are available for commercial use and these perform their required tasks quite competently (Van der Maarel et al., 2002). Two types of enzyme may be of use and interest to the industry, the first are those enzymes which are able to withstand the extremes of temperature and pH to which they will be subjected in the industrial environment. The second enzyme category which is of interest are those which perform reactions which are not within the range of the currently available commercial enzymes. These may include, for example, starch-branching enzymes which are able to produce highly branched starch molecules. Enzymes such as starch lysases may also have potential for development into commercially viable entities (Yu et al., 1999). The identification of novel starch-acting enzymes within a reasonable time frame is best achieved by designing an efficient and sensitive screening strategy and using it to screen the greatest possible input of diverse genes and organisms (Lorenz et al., 2002). The approach which is to be adopted when screening for amylase-like enzymes comes down essentially to a choice between two alternatives. Firstly there is the more traditional activity-based approach by which an enzyme is identified because it shows activity on a selected substrate (usually starch or one of its derivatives) (Fig. 4.5, grey tinted boxes on left-hand side). This method usually starts with a bacterial or fungal culture which shows activity on starch. Often these bacteria and fungi are isolates from samples taken from extreme environments such as hot springs, hydrothermal vents or soda lakes as microbes from these environments must have enzymes which function in these conditions (Sunna et al., 1997; Niehaus et al., 1999; Veille and Zeikus, 2001). Examples of activities which have been obtained using this method over the last few years include a thermostable pullulanase from Fervidobacterium pennavorans (Bertoldo et al. 1999) and an -amylase from isolated from Pyrococcus woesei (Antranikian, et al., 1990).

4.3.2 Genome mining: the exploitation of complete genome sequences The second approach which has also been successful, is the use of molecular biology techniques to either screen and isolate new genes of interest or

ß 2004, Woodhead Publishing Limited

Fig. 4.5

The approaches adopted to identify and purify enzymes which have an activity on starch.

alternatively genetically engineering can be employed to change the activity of a known enzyme (Fig. 4.5, clear box on right-hand side). The isolation or selection of genes from the genome of an organism can be performed using a number of techniques. Genome libraries are one such method. A library entails a large number of clones each containing small segments of genomic DNA which are to be screened for genes encoding amylase or starch-modifying activity. This screening is performed by growing the clones in a media which contains the substrate and accessing the activity of the clones on the substrate. A large number of substrates which may be used for the identification of enzymes with different activities on starch have also been developed. Modern high-throughput screening (HTS) methods are now greatly enhancing the efficiency of screening protocols which previously required more time and manpower to perform successfully. Both total DNA libraries and cDNA (copy DNA) libraries have been successfully screened for starch-acting genes in this way. Excellent examples of starch-acting enzymes which were identified by employing this strategy are the pullulanase, maltodextrin phosphorylase and -glucosidase enzymes isolated by screening a genome library of the hyperthermophilic bacteria Thermotoga maritima MSB8 (Bibel et al., 1997). In this case a genome library was screened on a number of different substrates each selected to identify clones which entailed a particular activity. HTS methods have the advantage that knowledge of the DNA or amino acid sequence of the enzyme is not required to isolate a suitable biocatalyst. Although for genome libraries in particular the presence of an active enzyme is very dependent on the expression system and bacterial host used as the enzyme will only be detected if the gene is correctly transcribed, translated and then processed into an active protein. The exploitation of the genetic material of 99% of the microbial community which is not picked up by standard cultivation methods is also now receiving increasing attention. Until recently this was a largely unused reservoir of novel and potentially valuable enzymes including some which act on starch (Lorenz et al., 2002). One approach used to obtain these enzymes involves the use of the conserved domains of the nucleotide or amino acid sequences of known amylase families such as family 13 (see earlier in this chapter) to design degenerate primers for use in PCR. These primers are then used with template DNA samples obtained from an extreme environment. The PCR products obtained can then be utilised to obtain genes which encode novel enzymes which act under extreme conditions. An indication of the potential of this strategy is given by looking at the list of genes picked up in Icelandic hot springs by Fridjonsson et al. (2003). An array of different family 13 and family 57 (see earlier in chapter) genes were isolated many of which display novel homologies and activities. A potential drawback of this method is the inability to reliably determine where the DNA originates. Consequently the necessity to express it in a heterologous host may result in problems particularly if the enzyme is to be used in the food industry. An interesting

ß 2004, Woodhead Publishing Limited

variation on these techniques was recently adopted by Tsutsumi et al. (1999) who used degenerate primers to successfully clone novel isoamylase genes from a species of Sulfolobus and of Rhodothermus marinus. The sequencing of whole genome sequences of many microorganisms, particularly those of thermophiles and fungi has opened the way for the mining of these genomes for starch acting enzymes. This, the bioinformatics approach, is inherently more conservative when compared to activity screening methods as the searches performed (irrespective of the sophistication of the programs employed to scan the genomes) rely on similarity to amino acid or DNA segments from already well characterised starch-acting enzymes. Nevertheless, due to the great diversity which is available in nature it is still possible to pick up surprisingly novel enzymes using these techniques (Lorenz et al., 2002). The families which include enzymes which act on starch include glycosyl hydrolase families 5, 13, 31, 57 and 77. The conserved regions of the protein can be used as search parameters. One of these families, family 13, includes the majority of enzymes used to act on starch, and has four well-defined amino acid motifs which are conserved in all family members. Additionally, all members of the family have three or more protein domains which are known to perform specific function. (van der Maarel et al., 2002) These conserved characteristics allow searches to be performed looking for proteins with similar amino acid patterns or similar domains. Homology identification and prediction can be performed using either primary methods such as Blast searches for homology between DNA and amino acid sequences to the more complex weighted methods such as pFAM (protein families) analysis and the use of Hidden Markov Models (HMMs). (Altschul et al., 1997, Gough et al., 2001).

4.3.4 Methods for identification of genes which encode starch-acting enzymes HMMs originally emerged in the domain of speech recognition where they were the result of attempts to model speech generation statistically. They have also been employed in the analysis of protein sequences as they offer a more systematic approach to modelling the consensus sequence of a protein family or domain. A model can be created which represents the `consensus profile' of amino acids for a set of protein sequences belonging to the same family. The `Superfamily' website (http:/supfam.mrclmb.cam.ac.u/ SUPERFAMILY/ assignment.html) (Gough et al., 2001) includes a HMM library and also a genome assignment server which assigns functions to the open reading frames on a large number of genomes based on homology to the models created. Of particular interest for the identification and selection of starch acting enzymes are two HMMs for domains found in enzymes with activities on starch. The first of these superfamily models is the (Trans)Glycosidase model, which entails the TIM beta/alpha-barrel fold. The second model is the -amylase C-terminal

ß 2004, Woodhead Publishing Limited

domain model, which is a model of a domain found at the C-terminus side of the catalytic region of amylase family enzymes. Microarrays, which are now available for a number of organisms and currently being created for many more, are also an interesting approach to the challenge of identifying novel starch-acting enzymes. The recently published work by Chhabra et al. (2003) on the caybohydrate acting enzymes Thermotoga maritima OT3 indicates that along with genes which are already known to encode starch-acting enzymes other genes (some of which are of unknown function) are induced by growth on a starch background. Such information may again be used as a guide to identifying new genes and is particularly useful if the gene/proteins do not share any homology with other know starch-acting genes. The future may also see the screening of the complete `Proteome' of a given organism by the coupled cloning of all predicted genes or open reading frames of a given organism's genome followed by their overexpression and activity testing using a high throughput screening approach. Protein engineering of -amylase-like enzymes -Amylases are employed in several industrial process such as starch liquefaction, laundering, dye removal and feed pre-processing (GuzmanMaldonado and Paredes-Lopez, 1995). Many of these processes require both pH values and temperatures which are quite different from those at which these enzymes perform optimally (Nielson et al., 2001). Amylases which have been engineered usually fall into two broad categories, firstly there are enzymes which have been engineered to increase their stability in a particular environment, usually high temperatures or low pH. The work which has been performed in the Bacillus licheniformis alphaamylase (BLA)(Termamyl) illustrates the improvements which can be attained from the study and engineering of these enzymes. The B. licheniformis amylase was compared with the amylase from B. amyloliquefaciens, with which is shares over 80% homology. This resulted in the identification of two regions which were found to be responsibly for the thermostability of BLA. Deletion or substitution of these regions, Gln178 (region I) and the 255th and 270th residues (region 2) resulted in the loss of thermostability (Suzuki et al., 1989). Conrad et al. (1995) performed a similar study in which a number of hybrids were created which entailed segments of both the B. licheniformis and B. amyloliquefaciens enzymes. This allowed the identification of four regions which contain amino acids important for thermostability. These regions are found between positions 34±76, 112±142, 174±179 and 263±276 respectively of the hybrid enzymes and illustrate the importance of the N-terminal portion of the enzymes in thermostability. Interestingly, two of the regions overlap with regions I and II as determined by Suzuki et al. (1989). Declerck et al. (1995) performed site-saturated mutagenesis, a technique whereby the targeted amino acid residue is substituted with all 20 other amino acid residues, on the BLA alpha-amylase. This revealed that substitutions at two critical positions for thermostability, namely Histidine 133 and Alanine 209, the

ß 2004, Woodhead Publishing Limited

half-life of the enzyme at 80 ëC could be increased by a factor of three. The greatest increases in stability were found when these residues were substituted with the amino acids Isoleusine and Valine respectively. This work also revealed that double mutants where both residues were simultaneously changed had a cumulative effect on the thermostability. Computer modelling was also employed to generate 3-D structures for both the wild-type and mutated versions of the enzymes. From these it was ascertained that Alanine 209 lies in the conserved / barrel domains of the alpha-amylase. Furthermore, the modelling suggests that the substitution of this amino acid residue with valine would result in increased intra- and interhelical compactness and hydrophobic interactions. Structure based mutagenesis was also used to determine the function of the amino acids previously proposed as important for the thermostability of the Bacillus licheniformis alpha-amylase (Declerck et al., 2000). Residues found to be involved in salt-bridges, calcium binding or potential deamination processes were selected and replaced with other amino acids using site-directed mutagenesis. This approach resulted in the identification of three asparagine residues (Asn172, Asn188 and Asn190) which, if replaced, could significantly increase the thermostability when compared to the wild-type enzyme. The greatest stabilisation effect found resulted from the substitution of phenylaline in place of asparagine at position 190, which gave a sixfold increase in the enzymes half-life at 80 ëC. Interesting work by Ma et al. (2000) on the barley -amylase showed that the -amylase found in germinated barley had a notably higher thermostability and substrate binding affinity for starch compared to that of mature barley grains. It was later elucidated that elimination of four C-terminal glycine rich repeats significantly increased the enzymes thermostability. The stability of the BLA enzyme at low pH has also been the subject of a number of studies in recent years. Random mutagenesis was employed by Shaw et al. (1999) to increase the stability of BLA enzyme at pH 5.0 and 83 ëC up to 23 times. Furthermore, Nielson et al. (2001) successfully altered the pH-profile of the Bacillus licheniformis -amylase by the mutation of selected neutral amino acid residues which are positioned close to the enzymes active site. This result indicated that point mutations which change the dynamics of the active site can also change the pH-activity profile. Gene shuffling is a technique whereby a number of genes which have desirable traits are fragmented using molecular biological techniques and `shuffled' together to create a library of hybrid gene constructs which are then screened to identify genes which encode enzymes with enhanced activities. Usually these enzymes display varying combinations of the traits encoded by the `parent' genes selected for `shuffling' (Chang et al., 1999). This method was successfully employed by Bozonnet et al. (2003) who created a number of chimeras from two barley -amylase genes, namely AMY1 and AMY2. Although these genes share over 80% homology with each other they also have quite different enzymatic properties. In particular they show major variations in calcium affinity, sensitivity to the inhibitor BASI (barley -amylase/subtilisin

ß 2004, Woodhead Publishing Limited

inhibitor) and in isoelectric point. In addition AMY2 is expressed in very low amounts in the host strains Saccharomyces cerevisiae and Pichia partoris. The gene shuffling method followed by screening on starch plates led to the isolation of chimeras with improved activity on starch. Directed evolution of the Thermus maltogenic amylase (ThMA) was recently performed by Young-Wan et al. (2003). Four rounds of DNA shuffling were performed followed by recombination of the mutants producing a mutant designated ThMA-DM. This mutant was found to have a total of seven amino acid substitutions and its optimal reaction temperature was found to be 75 ëC which is 15 ëC higher than the wild-type ThMA. HTS may now also be coupled with these genetic engineering techniques to allow the efficient screening of large numbers of variant clones to identify and isolate enzymes with relevant activities. A potentially powerful technique for the selection and development of industrial enzymes was recently used by Verhaert et al. (2002) to select and develop -amylases with improved low pH starch-binding. The technique used was phage surface display, this entails the display of the active amylases on the surface of phage fd. A selection system was then developed based on the ability of the -amylase displaying phages to bind to cross-linked starch. This system was then used to select for those phage which displayed amylases which bound the strongest to the cross-linked starch at low pH. In this way the random mutagenesis of the -amylase gene was used in conjunction with phage display as a selection technique to develop an enzyme which had a higher starch-binding ability at low pH. The second category of engineered amylases which are created are those which either act on a different substrate to the wild-type enzyme or produce a different product. The increasing availability of 3-D structural information has greatly aided the process of determining the structure-function relationships of these enzymes and also made it easier to determine which amino acids are important for reaction specificity of the enzyme. In such methods the substitution of a number of amino acid residues can significantly alter the substrates acted on and the products produced by an enzyme. The conversion of a cyclodextrin glycosyltransferase from Thermoanerobacterium thermosulfurigenes EMI with a transglycosylase activity into a starch hydrolase by the mutation of the phenylalanine amino acid residue at position 184 in the enzymes substrate binding site has recently been demonstrated. (Leemhuis et al., 2002). Work by Kuriki et al. (1996) showed that the reaction specificity of the neopullulanase from Bacillus stearthermophilus TRS40 could be altered to give up to a 45% increase in transglycosylation activity by substituting a number of individual amino acids. Although an enzyme may have an interesting activity under certain conditions it does not necessarily follow that it will be industrially viable. This can be achieved only if the enzyme can firstly be produced at high levels in either a bacterial or fungal host. Additionally, the enzyme must be adaptable to large-scale fermentation and purification. Genetic manipulation has also been employed to over come some of these problems. The work of Fukuda and

ß 2004, Woodhead Publishing Limited

Svensson (2003) shows this. The barley amylases AMY1 and AMY2 have 80% amino acid identity. Despite this, AMY1 is expressed at 0.5 g/ml while AMY2 is expressed at 0.05 g/ml in Saccharomyces cerevisiae strain S150-2B. To overcome this, ten non-conserved amino acid residues in AMY2 were replaced with the corresponding residues in AMY1. These mutants were obtained by degenerate oligonucleotide gene shuffling (DOGS) (Gibbs et al., 2001). This resulted in the creation of a mutant which had a three-fold increase in the expression level compared to the wild-type enzyme. (Fukuda and Svensson, 2003). Industrial conditions are another hurdle which most enzymes have to overcome if they are to be of practical use. Typically an -amylase needs to be able to function effectively in a solution of 30% dry solids if it is to be use in the degradation of starch (Schafer et al., 2000).

4.4

Future trends

Although there are many amylase-like enzymes which are already being used successfully in the starch industry there are also a number of areas where there is much scope for improvement in the enzymes that are available. There is also potential for new enzymatic activities which catalyse reactions that are currently not able to be performed using enzymatic methods. The desire for amylases which are able to work optimally at extremely high temperatures, low pH and also in the absence of calcium, is still not satisfied and it is probable that in the coming years further improvement in the biocatalysts performing these reactions will be achieved. The methods which will be used to attain these improvements will include many of the molecular biological techniques that have been used to effect improvements over the last few years. Better analysis and annotation of the increasing number of complete genomes that are now sequenced should lead to a number of useful enzymes for the industry. The new focus on proteomics and metabolomics, where complete metabolic pathways involved in the synthesis or degradation of molecules or even polymers are elucidated, may also result in the identification of new enzymatic activities. The growing ability of protein engineering to alter the activities of enzymes and do so in a predictable and logical manner will be complemented by the increasing knowledge which is now being compiled about the structure of amylases and how their structure relates to function. The alterations in the structure of the enzymes when they interact with both their substrate and cofactors is also shedding light on the mechanism employed by these enzymes to perform reactions. It is probable that the structure of products made by engineered amylases can be tailored by the logical manipulation of the amino acids involved in both the active site and substrate binding site of the enzymes. Such products may have a specified molecular structure or degree of polymerisation. The possibility of performing such modifications in a cheap way may open the way for many new application for starch and its derivatives.

ß 2004, Woodhead Publishing Limited

4.5

References (1992) `Maltodextrins: production, properties, and applications' in Schenck F.W. and Hebeda, R.E., Starch hydrolysis products. Worldwide technology, production and applications, NewYork, VCH Publishers, pp. 233±75.

ALEXANDER, R.J.

ALTSCHUL, S.F., MADDEN, T.L., SCHAFFER, A.A., ZHANG, J., ZHANG, Z., MILLER, W., LIPMAN, D.

(1997) `Gapped BLAST and PSI-BLAST: a new generation of protein database search programs', Nucleic Acid Res., 25(17), 3389±3402. ANTRANIKIAN, G., KOCH, R., SPREINAT, A. (1990) Novel hyperthermostable alpha-amylase. Patent application WO 90/11352. ARENDS-SCHOLTE, A.W., BERGSMA, J., EISSENS, A.C., GOTLIEB, K.F., LERK, C.F., SWINKELS, J.J.M.,

(1996) Starch products as tabletting excipient, method for preparing same, and method for making tablets. Patent application WO 96/09815. ARMBRUSTER, F.C. (1974) `Process for producing non-waxy hydrolysates', Patent application US 3 853 706. BERGSMA, J., TE WIERIK, G.H.P. (1997) `Debranched starch products as excipients in pharmaceutical tablets', in Van Doren, H.A. and Van Swaaij, A.C., Starch 96 ± the book, Den Haag, Zestec bv/Carbohydrate Research Foundation, pp. 105±12. BERTOLDO, C. DUFFNER, F., JORGENSEN, P.L., ANTRINIKAN, G. (1999) `Pullulanase type I from Fervidobacterium pennavorans Ven5: cloning, sequencing, and expression of the gene and biochemical characterization of the recombinant enzyme'. Appl. Environ. Micobiol., 65, 2084±2091. BESEMER, A., LERK, C.F. (1995) `Composition for controlled release of an active substance and method for the preparation of such a composition', Patent application EP 0648 116. BIBEL, M., BRETTL, C., GOSSLAR, U., KRIEGSHAUSER, G., LIEBL, W. (1997) `Isolation and analysis of genes for amylolytic enzymes of the hyperthermophilic bacterium Thermotoga maritima', FEMS Microbiology Letters, 158, 9±15. BINNEMA, D.J., WIJBENGA, D.J., KLIP, G.R., VAN HOOGMOED, C.G. (1996) `Thermostable nonreducing glucan-forming enzyme and conversion of starch to non-reducing glucans useful for preparation of foodstuffs, cosmetics, pharmaceuticals, and detergents', Patent application EP 727 485. BOZONNET S., KIM, T-J., SVENSSON, B. (2003) `Directed evolution of barley -amylase genes'. Abstracts of the 5th Carbohydrate Bioengineering Meeting, University of Groningen, Groningen, The Netherlands, p. 9. BULEÂON, A., COLONNA, P., PLANCHOT, V., BALL, S. (1998) `Starch granules: structure and biosynthesis', Int. J. Biol. Macromol., 23, 85±112. CAKEBREAD, S.H. (1972) `Physical properties of confectionery ingredients. Viscosity and high boilings stability (continued)', Confectionery Production, 38 (1), 16±19. CHANG, C.C. CHEN, T.T. COX, B.W. DAWES G.N., STEMMER, W.P., PUNNONEN J., PATTERN P.A. (1999) `Evolution of a cytokine using DNA family shuffling', Nat. Biotechnol., 17, 793±7. CHEN, H.M., BAKIR, U., REILLY, P.J., FORD, C. (1994) `Increased thermostability of Asn-182 ! Ala mutant Aspergillus awamori glucoamylase', Biotech. Bioeng., 43, 101±5. CHEN, W., LIU, L., SUN, P., JIN, C. (2000) `Cloning and expression of the gene encoding maltoologosyl trehalose synthase from Sulfolobus shibatae in E. coli', Wei Sheng Wu Xue Bao, 40, 57-61 TE WIERIK, G.H.P.

CHHABRA, S.R., SHOCKLEY, K.R., CONNERS, S.B., SCOTT, K.L., WOLFINGER, R.D., KELLY R.M.

(2003) `Carbohydrate-induced differential gene expression patterns in the hyperthermophilic bacterium Thermotoga maritima', J. Biol. Chem., 278 (9), 7540±52.

ß 2004, Woodhead Publishing Limited

(1998) `Enzymatic characterization of Novamyl, a thermostable -amylase', Starch/ StaÈrke, 50, 39±45. CONRAD B., HOANG V., POLLEY A., HOFEMEISTER J. (1995) `Hybrid Bacillus amyloliquefaciens X Bacillus licheniformis alpha-amylases. Construction, properties and sequence determination', Eur. J. Biochem., 230, 481±90. COUTINHO, P.M., HENRISSAT, B. (1999) `Carbohydrate-active enzymes: an integrated database approach', in Gilbert, H.J., Davies, G., Henrissat, B. and Svensson, B. (eds) Recent Advances in Carbohydrate Engineering, Cambridge, The Royal Society of Chemistry. CPC (1986) `Introduction to the role of glucose syrups in sugar confectionery', Confectionery Production, 52, 359±60. È PER, H. (1987) `Catalytic properties of Bacillus megaterium È NTHER, H., RO DAVID, M.H., GU amylase', Starch/StaÈrke, 39, 436±40. DE PASCALE, D., DI LERNIA, I., SASSO, M.P., FURIA, A., DE ROSSA, M., ROSSI, M. (2002) `A novel thermophilic fusion enzymes for trehalose production', Extremophiles, 6, 463±68. DECLERCK, N., JOYET, P., TROSSET, J.Y., GERNIER, J., GAILLARDIN, C. (1995) `Hyperthermostable mutants of Bacillus licheniformis alpha-amylase:multiple amino acid replacements and molecular modelling', Protein Eng., 8, 1029±37. DECLERCK, N., MACHUIS, M., WIEGAND, G., HUBER, R.,GAILLARDIN, C. (2000) `Probing structural determinants specifying high thermostability in Bacillus licheniformis alpha-amylse', J. Mol. Biol., 301, 1041±57. DIDERICHSEN, B., CHRISTIANSEN, L. (1988) `Cloning of a maltogenic alpha-amylase from Bacillus stearothermophilus', FEMS Microbiol. Lett., 56, 53±60. EUVERINK, G.J.W., BINNEMA, D.J. (1998) `Use of modified starch as an agent for forming a thermoreversible gel', Patent application WO 98/15347. FANG, T.Y., COUTINHO, P.M., REILLY, P.J., FORD, C. (1998) `Mutations to alter Aspergillus awamori glucoamylase selectivity. I', Protein Eng., 11, 119±26. FANG, T.Y., HONZATKO, R.B., REILLY, P.J., FORD, C. (1998) `Mutations to alter Aspergillus awamori glucoamylase selectivity. II', Protein Eng., 11, 127±33. CHRISTOPHERSEN, C., OTZEN, D.E., NORMAN, B.E., CHRISTENSEN, S., SCHaÈFER, T.

FRIDJONSSON, O., HREGGVIDSSON, G.O., SKIRNISDOTTIR, S., HAUKSDOTTIR, S., KRISTJANSSON, K.J. (2003) `Retrieval of amylolytic enzyme through GENEMINING of environmental DNA', Proceedings of the 5th Carbohydrate Bioengineering Meeting, Groningen, The Netherlands, p. 12. FUKUDA, K, SVENSSON, B. (2003) `Degenerate oligonucleotide gene shuffling of the Nterminal region in barley -amylase to improve secretion in S. cerevisiae and in cereal activity', Abstracts of the 5th Carbohydrate Bioengineering Meeting, University of Groningen, Groningen, The Netherlands, p. 71. GALLANT, D., DERRIEN, A., AUMAITRE, A., GUILBOT, A. (1973) `DeÂgradation in vitro de l'amidon par le suc pancreÂatique. Etude par microscopie eÂlectronique aÁ transmission et aÁ balayage', Starch/StaÈrke, 25, 56±64. GALLIARD, T. (1989) Starch availability and utilization. Chapter 1, Starch: Properties and potential, Chichester, Wiley, pp. 1±15. GIBBS, M.D., NEVALAINEN, K.M., BERGQUIST, P.L. (2001) `Degenerate oligonucleotide gene shuffling (DOGS): a method enhancing the frequency of recombination with family shuffling', Gene, 271 (1), 13±30. GIDLEY, M.J., BULPIN, P.V. (1989) `Aggregation of amylose in aqueous systems: the effect of chain length on phase behavior and aggregation kinetics', Macromolecules, 22, 341±6.

ß 2004, Woodhead Publishing Limited

(1999) `Starches from differing sources supply, demand, price formation', Starch/StaÈrke, 51, 193±6. GOUGH, J., KARPLUS, K., HUGHEY, R., CHOTHIA, C. (2001) `Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure', J. Mol. Biol., 313 (4), 903±19. GURAYA, H.S., JAMES, C., CHAMPAGNE, E.T. (2001a) `Effect of cooling and freezing on the digestibility of debranched rice starch and physical properties of the resulting material', Starch/StaÈrke, 53, 64±74. GURAYA, H.S., JAMES, C., CHAMPAGNE, E.T. (2001b), `Effect of enzyme concentration and storage temperature on the formation of slowly digestible starch from cooled debranched rice starch', Starch/StaÈrke, 53, 131±9. GUZMAN-MALDONADO, H., PAREDES-LOPES, O. (1995) `Amylolytic enzymes and products derived from starch: review', Crit. Rev. Food Sci. Nutr., 35 (5), 373±403. HEBEDA, R.E., STYRLUND, C.R., TEAGUE, W.M. (1988), `Benefits of Bacillus megaterium amylase in dextrose production', Starch/StaÈrke, 40, 33±36. KAPER, F., GRUPPEN, H. (1987) `Replace oil and fat with potato-based ingredient', Food Technol., 41, 112±13. KATO, M., MIURA, Y., KETTOKU, M., SHINDO, K., IWAMATSU, A., KOBAYASHI, K. (1996a), `Purification and characterization of new trehalose-producing enzymes isolated from hyperthermophilic archaeaum, Sulfolobus solfataricus KM1', Biosci. Biotechnol. Biochem., 60, 546±50. KATO, M., MIURA, Y., KETTOKU, M., SHINDO, K., IWAMATSU, A., KOBAYASHI, K. (1996b), `Reaction mechanism of a new glycosyltrehalose-producing enzyme isolated from the hyperthermophilic archaeum, Sulfolobus solfataricus KM1', Biosci. Biotechnol. Biochem., 60, 921±4. KATO, M., MIURA, Y., KETTOKU, M., KOMEDA, T., IWAMATSU, A., KOBAYASHI, K. (1996c), `Reaction mechanism of a new glycosyltrehalose-hydrolyzing enzyme isolated from the hyperthermophilic archaeum, Sulfolobus solfataricus KM1', Biosci. Biotechnol. Biochem., 60, 925±8. KEARSLEY, M.W., DZIEDZIC, S.Z. (1995) Handbook of starch hydrolysis products and their derivatives, Glasgow, Blackie Academic & Professional. KENNEDY, J.F., KNILL, C.J., TAYLOR, D.W. (1995) `Maltodextrins', in Kearsley, M.W. and Dziedzic, S.Z. (1995), Handbook of starch hydrolysis products and their derivatives, Glasgow, Blackie Academic & Professional, pp. 65±82. KEYSERS, H. (1982) `Unterschiede in der KristallisationsbestaÈndigkeit zwischen Invertzucker und High-Fructose-Glukosesirup sowie deren Weichhaltewirkung', Zucker u. SuÈûwarenwirtschaft, 35 (5), 147±8. KIEL, J.A.K.W., BOELS, J.M., BELDMAN, G., VENEMA, G. (1991) `Molecular cloning and nucleotide sequence of glycogen branching enzyme (glg B) from Bacillus stearothermophilus and expression in Escherichia coli and Bacillus subtilis', Mol. Gen. Genet., 230, 136±44. KIT, J.S. (1993) `Hard candy graining, causes and prevention', The Manufacturing Confectioner, 73 (11), 47±8. GORDON, I.

KURIKI, T., KANEKO, H., YANASE, M., TAKATA, H., SHIMADA, J., HANDA, S., TAKADIA, T., UMEYAMA, H., OKADA, S. (1996) `Controlling substate preference and transglycosylstion activity of neopullulanase by manipulating steric constraint and hydrophobicity in active center', J. Biol. Chem., 271 (29), 17321±9. LAMA, L., BARBARA, N., TRINONE, A., MORZILLO, P., CALANDRELLI, V., GAMBACORTA, A. (1991) `Thermostable amylolytic activity from Sulfolobus solfataricus', Biotechnol.

ß 2004, Woodhead Publishing Limited

Forum Eur., 8, 201±3. (2002) `Mutations converting cyclodextrin glycosyltransferase from a transglycosylase into a starch hydrolase', FEBS Letters, 514 (2±3), 189±92. LEVINE, H., SLADE, L. (1984), `A polymer physico-chemical approach to the study of commercial starch hydrolysis products', Carbohydr. Polym., 6, 213±44. LIU, H.L., COUTINHO, P.M., REILLY, P.J., FORD, C. (1998) `Mutations to alter Aspergillus awamori glucoamylase selectivity. III', Protein Eng., 11 389±98. LORENZ, P., LIEBETON, K., NIEHAUS, F, ECK, J. (2002) `Screening for novel enzymes for biocataltic processes: accessing the metagenome as a resource of novel function sequence space', Curr. Opin. in Biotechnolo, 13, 572±7. MA, Y.F., EGLINTON, J.K., EVANS, D.E., LOGUE, S.J., LANGRIDGE, P. (2000) `Removal of the four C-terminal glycine-rich repeats enhances the thermostability and substrate binding of barley beta-amylase', Biochemistry, 39 (44), 13350±5. MACALLISTER, R.V. (1980) in Pitcher Jr, W.H., Immobilized enzymes for food processing, Boca Raton, CRC Press, pp. 81±111. MADSEN, G., NORMAN, B.E., SLOTT, S. (1973) `A new, heat stable bacterial amylase and its use in high temperature liquefaction', Starch/StaÈrke, 25, 304±8. MARUTA, K., KUBOTA, M., SUGIMOTO, T., MIYAKE, T. (1994a) `Microbial trehalose-releasing enzyme, method for its production and purification, method for trehalose manufacture, and its uses in pharmaceuticals, foods, and cosmetics', Patent application EP 0628 630. MARUTA, K., SUGIMOTO, T., KUBOTA, M., MIYAKE, T. (1994b) `Non-reducing saccharideforming enzyme, and its purification from microorganisms, its uses', Patent application EP 0606 753. MILES, M.J., MORRIS, V.J., RING, S.G. (1984) `Gelation of amylose', Carbohydr. Res., 135, 257±69. MITCHINSON, C., REQUADT, C., ROPP, T., SOLHEIM, L.P., RINGER, C., DAY, A. (1996), `Bacillus amylase mutant recombinant production, improved low pH starch liquefaction, thermal stability and activity, and use as a laundry detergent or dishwashing agent', WO 96 39528. MOREHOUSE, A.L., MALZAHN, R.C., DAY, J.T. (1972) `Hydrolysis of starch', Patent application US 3 663 369. LEEMHUIS, H., DIJKSTRA. B.W., DIJKHUIZEN, L.

NAKADA, T., IKEGAMI, S., CHAEN, H., KUBOTA, M., FUKUDA, S., SUGIMOTO, T., KURIMOTO, M., TSUJISAKA, Y. (1996a) `Purification and characterization of thermostable maltooligosyl trehalose synthase from the thermoacidophilic archaebacterium Sulfolobus acidocaldaricus', Biosci. Biotechnol. Biochem., 60, 263±6. NAKADA, T., IKEGAMI, S., CHAEN, H., KUBOTA, M., FUKUDA, S., SUGIMOTO, T., KURIMOTO, M.,

(1996b) `Purification and characterization of thermostable maltooligosyl trehalose trehalohydrolase from the thermoacidophilic archaebacterium Sulfolobus acidocaldaricus', Biosci. Biotechnol. Biochem., 60, 267±70. NIEHAUS, F., BERTOLDO, C., KAHLER, M., ANTRANIKIAN, G. (1999) `Extremophiles as a scource of novel enzymes for industrial applications', Appl. Microbiol. Biotechnol., 51, 711±29 NIELSON, J.E., BORCHERT, T.V., VRIEND, G. (2001) `The determinants of -amylase pHactivity profiles', Proteins Eng., 14 (7), 505±12. NIEMANN, C., WHISTLER, R.L. (1992) `Effect of acid hydrolysis and ball milling on porous corn starch', Starch/StaÈrke, 44, 409±14. È LLER, B., SAENGER, W. (1990) `Phosphorolytische NIEMANN, C., NUCK, R., PFANNEMU TSUJISAKA, Y.

ß 2004, Woodhead Publishing Limited

Synthese von niedermolekularen Amylosen mit modifizierten Endgruppen', Carbohydr. Res., 197, 187±96. NISHIMOTO, T., NAKANO, M., IKEGAMI, S., CHAEN, H., FUKUDA, S., SUGIMOTO, T., KURIMOTO, M., TSUJISAKA, Y. (1995) `Existence of a novel enzyme converting maltose into trehalose', Biosci. Biotechnol. Biochem., 59, 2189±90. NISHIMOTO, T., NAKADA, T., CHAEN, H., FUKUDA, S., SUGIMOTO, T., KURIMOTO, M., TSUJISAKA, Y.

(1996) `Purification and characterization of a thermostable trehalose synthase from Thermus aquaticus', Biosci. Biotechnol. Biochem., 60, 835±9. NORMAN, B.E., PEDERSEN, S., BISGAARD-FRANDSEN, H., BORCHERT, T.V. (1997) `Innovations in enzymatic starch liquefaction', Lecture 48th Starch Convention, Detmold. OOSTERGETEL, G.T., VAN BRUGGEN, E.F.J. (1993) `The crystalline domains in potato starch granules are arranged in a helical fashion', Carbohydr. Polym., 21, 7±12. PARKER, R., RING, S.G. (2001) `Aspects of the physical chemistry of starch', J. Cereal Sci., 34, 1±17. PECKOUS, L.W. (1999) `A novel alpha-amylase for liquefaction of whole grain mash', Lecture 50th Starch Convention, Detmold. PFANNEMUÈLLER, B. (1975) `Living Polymerization und enzymatische Polysaccharidsynthese', Naturwissenschaften, 62, 231±3. È FER, H., SCHULZ, R.C. (1971) `Conformation of amylose in PFANNEMUÈLLER, B., MAYERHO aqueous solution: optical rotatory dispersion and circular dichoism of amyloseiodine complexes and dependence on chain length of retrogradation of amylose', Biopolymers, 10, 243±61. PFANNEMUÈLLER, B., BURCHARD, W., FRANKEN, I. (1976) `Untersuchungen an enzymatisch verzweigten Polysacchariden. I. Sternpolymere aus Glykogen und Amylopekin mit phosphorolytisch verlaÈngerten Zweigen', StaÈrke, 28, 1±9. PLANCHOT, V., COLONNA, P., GALLANT, D.J., BOUCHET, B. (1995) `Extensive degradation of native starch by alpha-amylase from Aspergillus fumigatus', J. Cereal Sci., 21, 163±71. REEVE, A. (1992) `Starch hydrolysis: processes and equipment', in Schenck, F.W. and Hebeda, R.E., Starch hydrolysis products. Worldwide technology, production and applications, NewYork, VCH Publishers, pp. 79±120. RICHTER, M., SCHIERBAUM, F., AUGUSTAT, S. (1975) `Starch hydrolysis products for use in calorie-reduced foods. Use of new types of gel-forming starch hydrolysates in the food industry', ErnaÈhrungsforschung, 20, 168±71. RIISGAARD, S. (1990) `The enzyme industry and modern biotechnology', in Christiansen, C., Munck, L., Villadsen, J. (eds), Proceedings of the fifth European congress on biotechnology, Copenhagen, Munksgaard, pp. 31±40. ROBYT, J.F. (1985) `Enzymes in the hydrolysis and synthesis of starch', in Whistler, R.L., BeMiller, J.N. and Paschall, E.F. (eds), Starch: chemistry and technology, Orlando, Academic Press, pp. 87±123. ROOS, Y.H. (1995) Phase transition in foods, Helsinki, Academic Press. SCHAFER, T., DUFFNER, F., BORCHERT, T.V. (2000) `Extremophilic enzymes in industry: screening protein engineering and application', Proceedings of the 3rd International Congress on Extremophiles, Hamburg, Germany, pp. 306±7. SCHENCK, F.W., HEBEDA, R.E. (1992), Starch hydrolysis products. Worldwide technology, production and applications, New York, VCH Publishers. SCHIRALDI, C., DI LERNIA, I., DE ROSSA, M. (2002) `Trehalose production: exploiting novel approaches', Trends in Biotechnology, 20 (10), 420±5. SHAW, A., BOTT, R., DAY, A.G. (1999) `Protein engineering of alpha-amylase for low pH

ß 2004, Woodhead Publishing Limited

performance', Curr. Opin. Biotechnol., 10 (4), 349±52. (1988) `An acid stable thermostable alpha amylase for starch liquefaction', Cereal Foods World, 33, 929±34. STARNES, R.L. (1990) `Industrial potential of cyclodextrin glycosyl transferases', Cereal Foods World, 35, 1094±9. STORSBERG, J., VAN AERT, H., VAN ROOST, C., RITTER, H. (2003) `Cyclodextrins in polymer synthesis: a simple and surfactant free way to polymer particles having narrow particle size distribution', Macromolecules, 36, 50±3. SUNNA, A., MORACCI, M., ROSSI, M., ANTRANIKIAN, G. (1997) `Glycosyl hydrolases from hyperthermophiles', Extremophiles, 1, 2±13. SUZUKI Y., ITO, N., YUUKI T., YAMAGATA H., UDAKA S. (1989) `Amino acid residues stabilizing a Bacillus alpha-amylase against irreversible thermoinactivation', J. Biol. Chem., 264 (32), 18933±8. SZEJTLI, J. (1998) Cyclodextrin technology, Dordrecht, Kluwer Academic Publisher. TAKAHA, T., YANASE, M., TAKATA, H., OKADA, S., SMITH, S.M. (1996) `Potato D-enzyme catalyzes the cyclization of amylose to produce cycloamylose, a novel cyclic glucan', J. Biol. Chem., 271, 2902±8. TAKAHA, T., YANASE, M., TAKATA, H., OKADA, S., SMITH, S.M. (1997) `Cyclic glucans produced by the intramolecular transglycosylation activity of potato D-enzyme on amylopectin', Biochem. Biophys. Res. Commun., 247, 493±7. TAKASAKI, Y., TANABE, O. (1971) `Enzymatic method for converting glucose in glucose syrups to fructose', Patent application US 3 616 221. TAKASAKI, Y., KOSUGI, Y., KANABAYASHI, A. (1969) Fermentation advances, New York, Academic Press. TAKATA, H., TAKAHA, T., KURIKI, T., OKADA, S., TAKAGI, M., IMANAKA, T. (1994) `Properties and active center of the thermostable branching enzyme from Bacillus stearothermophilus', Appl. Environ. Microbiol., 60, 3096±104. TAKATA, H., TAKAHA, T., OKADA, S., TAKAGI, M., IMANAKA, T. (1996) `Cyclization reaction catalyzed by branching enzyme', J. Bacteriol., 178, 1600±6. TAKATA, H., TAKAHA, T., OKADA, S., HIZUKURI, S., TAKAGI, M., IMANAKA, T. (1997) `Structure of the cyclic glucan product from amylopectin by Bacillus stearothermophilus branching enzyme', Carbohydr. Res., 295, 91±101. TAMURI, M., KANNO, M., ISHII, Y. (1979) `Heat and acid-stable -amylase enzymes and processes for producing the same', Patent application US 4 284 722. TE WIERIK, G.H.P., EISSENS, A.C., BERGSMA, J., ARENDS-SCHOLTE, A.W., LERK, C.F. (1997) `A new generation of starch products as excipients in pharmaceutical tablets. 2. High surface area retrograded pregelatinized starch products in sustained-release tablets', J. Controlled Release, 45, 25±33. TSUTSUMI, N., BISGARD-FRABNTZEN, H., SVENDSEN A. (1999) Starch conversion process using thermostable isoamylases from Sulfolobus. Patent application WO 99/01545. VAN DER MAAREL, M.J.E.C., EUVERINK, G.J.W., BINNEMA, D.J., BOS, H.TH.P., BERGSMA, J. (2000) `Amylomaltase from the hyperthermophilic bacterium Thermus thermophilus: enzyme characteristics and applications in the starch industry', Med. Fac. Landbouwuniv. Gent, 65, 231±4. SHETTY, J.K., ALLEN, W.G.

VAN DER MAAREL, M.J.E.C., VAN DER VEEN, B., UITDEHAAG, J.C.M., LEEMHUIS, H., DIJKHUIZEN, L. (2002) `Properties and applications of starch-converting enzymes of the amylase family', J. Biotechnol., 94, 137±55. VEILLE, C., ZEIKUS G.J. (2001) `Hyperthermophilic enzymes: sources, use, and molecular mechanisms for thermostability', Microbiol. Mol. Biol. Rev., 65, 1±43.

ß 2004, Woodhead Publishing Limited

(2002), `Phage display selects for amylases with improved low pH starch binding', 96 (1), 103±18. WADA, T., KUBOTA, S., OOGUCHI, M. (1995) `Enzymic manufacture of trehalose', Jpn Kokai Tokkyo Koho JP 95203983. WHITE, J.S. (1992) `Fructose syrup: production, properties, and applications', in Schenck, F.W. and Hebeda, R.E., Starch hydrolysis products. Worldwide technology, production and applications, NewYork, VCH Publishers, pp. 177±99. WIJBENGA, D.J., BELDMAN, G., VEEN, A., BINNEMA, D.J. (1991) `Production of native-starchdegrading enzymes by a Bacillus firmus/lentus strain', Appl. Microbiol. Biotechnol., 35, 180±4. WIJBENGA, D.J., BINNEMA, D.J. (1997) `Enzymatic modification of starch granules', in Van Doren, H.A. and Van Swaaij, A.C. (eds), Starch 96 ± the book, pp. 97±104. WURSCH, P., GUMY, D. (1994) `Inhibition of amylopectin retrogradation by partial betaamylolysis', Carbohydr. Res., 256, 129±37. VERHAERT, R.M., BEEKWILDER, J., OLSTHOORN, R., VAN DUIN J., QUAX, W.J.

YOUNG-WAN, K., JI-HYE, C., JUNG-WAN, K., CHEONSEOK, P., JUNG-WOO, K., BYOUNG-HA, O., TAEWHA, M., KWAN-HWA, P. (2003) `Directed evolution of thermus maltogenic amylase towards enhanced thermal resistance', Proceedings of the 5th Carbohydrate Bioengineering Meeting, Groningen, The Netherlands, p. 4. YU, S., PEDERSEN, M. (1993) `Alpha-1,4-glucan lyase, a new class of starch/glycogen degrading enzyme. II. Subcellular localisation and partial amino-acid sequence', Planta, 191 (1), 137±42. YU, S., AHMAD, T., KENNE, L., PEDERSEN, M. (1995) `Alpha-1,4-glucan lyase, a new class of starch/glycogen degrading enzyme. III Substrate specificity, mode of action, and cleavage mechanism', Biochim Biophys Acta, 1244 (1), 1±9. YU, S., BOJSEN, K., SVENSSON, B., MARCUSSEN, J. (1999) `Alpha-1,4-glucan lyases producing 1,5-anhydro-D-fructose from starch and glycogen have sequence similiarity to alpha-glucosidases', Biochim Biophys Acta, 1433 (1±2) 1±15.

ß 2004, Woodhead Publishing Limited

5 Understanding starch structure and functionality A. M. Donald, University of Cambridge, UK

5.1 Introduction: overview of packing at different lengthscales The basic structure of starch granules has been recognised for decades, but the implications of a granule's detailed internal packing, and its interplay both with the biochemical pathways of production and the consequences for functionality and end-use, are still being unravelled. In this chapter these interrelationships will be explored, with the aim of providing a physical framework in which to understand recent developments in applications discussed elsewhere in this book. Successful utilisation of starch can no longer be viewed as simply the preserve of starch chemists, but a whole range of different disciplines must be brought to bear on the topic. The first section covers a historical overview of our understanding of the structure of the granule, and how modern methodologies are improving this understanding. Different frameworks for describing the structures at lengthscales from angstroms (i.e interatomic) to microns (the whole granule) are considered. Building on this, the implications of the high degree of branching of the amylopectin molecule are considered, and how this can be accommodated within the regular structure of the granule in the next section. Having explored the basic nature of the chain packing within the granule, Section 5.3 turns to consider ways in which this can be altered. In particular the effects of plasticisation and thermal treatment are considered, commonly used approaches in manufacturing. For many applications, the native granule is not used. Instead the process known as gelatinisation is initiated, in which heat and moisture are applied, leading to disruption and ultimately destruction of the granule. How these processes occur is considered in Section 5.4 in the light of

ß 2004, Woodhead Publishing Limited

the structural insight provided earlier. The final section then considers other types of processing including milling and extrusion. Early reviews of starch granule structure can be found in (French, 1984, Blanshard, 1987). Every variety of starch has its own characteristic shape and average size. Wheat even exhibits two different populations, so-called A and B type. The experienced eye can distinguish starches simply by close examination under either light or electron microscopy. Optical microscopy reveals the overall shape, which ranges from simply spherical, to ellipsoidal and lenticular (Banks and Muir, 1980, Jane et al., 1994). Some granules, such as rice, can also exhibit quite sharp edges (Rani and Bhattacharya, 1995), whereas some of the genetically modified versions show highly complex morphologies (Wang et al., 1997). Examination of the granule under crossed polarisers shows a characteristic Maltese cross, which rotates as the polarisers are rotated. Analysis of this birefringent structure shows, subject to some assumptions about the correlation between molecular axis and principal optic axis, that the underlying packing of the molecules is radial (Banks et al., 1972). This conclusion is confirmed in a more direct way by microfocus X-ray diffraction (Waigh et al., 1997). Optical microscopy has, of course, rather limited resolution, and much effort has also been directed at both scanning and transmission electron microscopy (SEM and TEM respectively) and, most recently atomic force microscopy, to provide more detailed information on internal packing. SEM can either be used simply to examine the gross morphology or, by the use of an etch on fractured samples, to study variations in internal packing (see e.g. Gallant and Guilbot, 1973, Banks and Muir, 1980, Baldwin et al., 1994). TEM provides yet higher resolution images, but has the additional limitation that thin sections must be cut in order to permit penetration by the electron beam. In addition, in order to obtain contrast either etching or staining is, in general, required (Gallant and Massy, 1969, Gallant et al., 1997, Atkin et al., 1998). Any of the microscopies can be used to reveal the existence of a series of concentric rings, which have become known as growth rings (Fig. 5.1). Optically these can be demonstrated by the use of stains (Badenhuizen, 1939, 1959), or by cytochemical markers (Atkin et al., 1999). For either SEM or TEM, etches can be used to highlight these rings, demonstrating that these correspond to alternating regions more and less resistant to the etch (which can be an acid or alkali, or an enzyme such as -amylase) (Buttrose, 1960, 1962, 1963, FreyWyssling and Buttrose, 1961, Gallant and Massy, 1969, Gallant and Guilbot, 1973, Atkin et al., 1998, Fulton et al., 2002, Zeeman et al., 2002). The origin of the rings is not entirely clear. Early work (Buttrose, 1962) indicated that they arose from diurnal fluctuations, but exactly what is fluctuating is less clear (light, temperature or substrate supply for instance). This explanation of diurnal fluctuations giving rise to growth rings is supported by the observations of wheat and barley granules lacking growth rings after growth in constant environmental conditions (Buttrose, 1962). However, if potato tubers are grown under constant conditions, the appearance of the resulting growth rings is unchanged (Buttrose,

ß 2004, Woodhead Publishing Limited

Fig. 5.1 Potato starch (cv Desiree) fractured at low temperature and etched with -amylase before viewing uncoated in the environmental scanning electron microscope (photograph courtesy of Jane Crawshaw).

1962), which suggests that for potatoes at least a circadian rhythm may be playing a role.The growth rings are spaced at distances of 100 nm and upwards. The spacing is variable and can depend on position within the granule. Their dimensions are hard to determine accurately since the width of the readily etched regions will be affected by the time the etch is applied. If one now turns to much smaller distances within the granules, then it is well known that crystallinity is present. Many different techniques can probe this. Most straightforwardly, wide-angle X-ray diffraction confirms the existence of crystals. Three different types of crystal structures have been identified, classified by Katz and Itallie (1930). These are known as A-type, characteristic of cereal starches, B-type found in tuber starches, and C-type found in legumes. Figure 5.2 shows the crystal structure for A- and B-type starches (Imberty et al., 1991); it is now known that C-type starches actually consist of a mixture of Aand B-type, with the centre of legume starch granules exhibiting the B-form, and the outermost regions the A-type (Gidley, 1987, Bogracheva et al., 1998, Buleon et al., 1998). A-type starches have a monoclinic unit cell (a ˆ 2.124 nm, b ˆ 1.172 nm, c ˆ 1.069 nm and ˆ 123.5ë) with space group B2 and with eight water molecules per unit cell; in the B-type crystal structure there is a hexagonal unit cell (a ˆ b ˆ 1.85 nm, c ˆ 1.04 nm) with 36 water molecules per cell. In both cases the basic building blocks are double helices of glucose residues. The crystal structures described are actually derived from studies on amylose (Imberty et al., 1991), but in fact within the starch granule it is not the amylose which crystallises but the amylopectin, although both of these polysaccharides are believed to crystallise in the same way. Nevertheless, based upon the fact that the waxy mutant of maize, which contains no amylose, shows essentially the same crystal structure, it has long been recognised that it is the amylopectin

ß 2004, Woodhead Publishing Limited

Fig. 5.2 The crystal structure for A-type and B-type starches of amylose (after Imberty et al., 1991). The small black circles represent water molecules within the structure; glucose molecules are represented by open circles.

which crystallises within the granule. Amylopectin is a highly branched molecule, and its side chain branches intertwine to form the double helices which are the basis of the crystals. Different species of starch not only exhibit different proportions of amylose to amylopectin (although for most wild-type species amylopectin usually makes up around 70% of the polysaccharide content), but the precise architecture of the amylopectin molecule is also species dependent. This is discussed in more detail in Chapter 2. For the purposes of this chapter, the simple schematic shown in Fig. 5.3 will suffice. Figure 5.3 identifies the different types of sub-chains within the amylopectin molecule, termed A, B and C. The A chains are the shortest (Degree of polymerisation, DP, 6±15) and are linked to the amylopectin molecule by a single  1±6 linkage. Being linked by a single  1±6 bond identifies the A chains as `outer' chains. B chains are identified as those that support A chains and other B chains. B chains are then further classified into B1, B2, B3 and B4 depending on their respective length and the number of `clusters' they span (see below) (Kainuma and French, 1972, Robin et al., 1975b). B1 chains span one cluster, B2 span two clusters, and so on. By spanning clusters these longer chains provide structural integrity to the whole molecule, and also the long range correlations within the granule. B1 chains have typical DPs of 15±25, B2 typically have a DP of 40±50 with B3 and B4 chains being longer. There is only one C chain per amylopectin molecule and it is identified as having the only non-reducing end. Within this structure the branch-points are located in the low molecular order region, and the linear chains within the high molecular order region. These linear chains can then form double helices and pack into a crystal structure, A- or B-type according to the type of starch as indicated above. The actual distribution of the different populations of side chain type varies between species, and modern methods now enable accurate determinations of

ß 2004, Woodhead Publishing Limited

Fig. 5.3 Schematic representation of the sub-chains within an amylopectin molecule. Where the branch points sit are regions of low order, and neighbouring B chains can form double helices to make up the crystal structure.

these to be made (see Chapter 2) building on the original work of Hizukuri et al., (1983), Hizukuri (1985), (1986), Hizukuri and Maehara (1990). Broadly speaking A-type starches have shorter side chain branches than B-type, although exactly why this should be so is not yet clear. There are two other aspects of the hierarchical packing of the chains in the starch granule which need to be considered. TEM has been used in an unconventional way by Oostergetel and van Bruggen (1993) using tomographic reconstruction of images of negatively stained potato starch granule fragments coupled with electron diffraction recorded at ÿ170ëC. Using this approach it was concluded that the crystalline regions were formed from a continuous network of left-handed helices; in other words that there is a super-helical structure present. These studies have not been repeated on other starches, but studies on whole, unstained potato starch granules using microfocus small angle X-ray scattering were shown to be consistent with this type of packing (Waigh et al., 1999), although not uniquely so. Secondly, TEM has also been used to suggest the existence of `blocklets'. The concept of blocklets within starch granules has been around for more than 50 years (Hanson and Katz, 1934, Badenhuizen, 1936), the original idea being

ß 2004, Woodhead Publishing Limited

that crystalline regions or blocklets were distributed in an amorphous matrix. Much more recently this idea has been revived, based on evidence from a range of microscopies. High resolution SEM images taken on granules etched by amylase, reveal spherical entities which are less etched than the surroundings, and these are found in a range of different starches (Gallant et al., 1992). The use of TEM to study stained sections supports the existence of this structure, with the stain being predominantly excluded from the more crystalline blocklets than from the surrounding amorphous material (Gallant et al., 1997). However, what is puzzling, is that the blocklets are apparently found in both the semicrystalline and amorphous growth rings; if the blocklets are actually crystalline entitities it is not clear why they should persist in the amorphous growth ring. Most recently AFM has also confirmed the presence of a blocklet structure in maize granules (Ridout et al., 2002), although this paper highlighted the dangers of sample mounting for the introduction of artefacts which may have been present in earlier papers claiming AFM support for blocklet structures (Baker et al., 2001).

5.2

The effect of amylopectin chain architecture on packing

All of the above descriptions of the internal structure of the starch granule start from the belief that it is the amylopectin molecule, in all its complexity, which is involved in the crystalline component. However, as plant biochemists unravel the enzymatic pathways to the synthesis of the starch polysaccharides, it is important to be able to relate alterations in the amylopectin architecture to packing within the granule, and ultimately to end use and functionality. Starting from the cluster model (see Fig. 5.3), one can immediately see that alterations in the distribution of A, B and C chain lengths are likely to have a profound effect on the ability of the chains to pack efficiently. One way of trying to rationalise this ± even without biochemical methods being able to give very precise insight into the exact distribution of side chain branching ± is to borrow the concept of side chain liquid crystallinity from the synthetic polymer literature. In conventional synthetic side chain liquid crystalline polymers, rigid moieties (the `mesogen') are attached to a flexible backbone via so-called `flexible spacers', as in Fig. 5.4. The mesogen is typically a para-linked group of aromatic rings attached to the backbone through an alkyl group (for a full discussion of liquid crystalline polymers see (Donald and Windle, 1992). If the flexible spacer is not sufficiently flexible, or is too short so that entropic effects of the backbone dominate over any tendency for the side chains to organise, then the mesogens cannot effectively line up and a nematic phase forms (see Fig. 5.4a). If on the other hand, there is sufficient decoupling between the mesogens and the backbone, then the tendency for the former to line up in register can dominate and a smectic phase forms (see Fig. 5.4b). The smectic phase is a lamellar, or planar, structure and there are, in fact, a variety of different variants with different degrees of symmetry.

ß 2004, Woodhead Publishing Limited

Fig. 5.4 Schematic representation of the structure of a side chain liquid crystalline polymer. (a) shows the less ordered nematic phase, whereas (b) shows a smectic phase in which the mesogens line up into layers.

The relevance of this framework to starch can be seen when it is recognised that the double helices formed from the side chains of the amylopectin molecule are effectively stiff and rigid, and can therefore be equated with the mesogens in Fig. 5.4. Clearly there is some short segment of the sidechains which must permit neighbouring side chains to come together and intertwine to form the double helical structure, and these segments comprise the flexible spacers. The biochemical characterisation of the side chains does not permit a ready separation of these two different segments of each side chain, so that it is difficult to know exactly what the distribution of flexible spacer lengths is, or what proportion of the chain actually sits in a double helix, but some inferences about these parameters can be made (Jane et al., 1997, O'Sullivan and Perez, 1999). The recognition that a liquid crystalline analogy might be fruitful for starch was first put forward in (Waigh et al., 1996), and since then the idea has been further developed (Waigh et al., 1998, 2000a,b). Small angle scattering (using either X-rays or neutrons) reveals that there is a regular repeat in a starch granule that occurs at around 9 nm (Sterling, 1962, Blanshard et al., 1984), and this spacing has been found to be remarkably constant across a large number of starches (Jenkins et al., 1993). It is also known that this spacing is not present if the starch is viewed when dry (Waigh et al., 1996). These observations can be rationalised by considering that when the starch granule is hydrated a regular, smectic structure develops in which layers of double helices are well-aligned and correlations exist both laterally (which gives rise to the crystalline structure familiar from wide-angle X-ray scattering) and between the layers (which gives rise to the 9 nm spacing). On the other hand when the granule is dried out the flexible spacer becomes less flexible, pulling the double helices out of register and leading to the development of a nematic phase in which the long range correlations between the layers ± and hence the 9 nm spacing present in the smectic phase ± is lost, although correlations persist within the double helices retaining the presence of appropriate peaks in the wide-angle X-ray scattering pattern. Further thought will be given to this in Section 5.3.

ß 2004, Woodhead Publishing Limited

However, what is the impact of chain architecture within this model? Taking first the broad classification into A and B type starches, it is known that on average the B chain lengths in B type starches such as potato are longer than in A type starches, such as the cereals, based on debranching studies (Hizukuri, 1986), and more specifically that the double helical length is longer for B type starches than for A, but the flexible spacer length is shorter. As a consequence one might expect potato starch, for instance, to have more stable crystals than wheat or maize, because the width of the lamellar crystal is greater. On the other hand, the fact that the flexible spacer units are shorter is likely to make it harder for the smectic order to develop without compromising the backbone conformation ± the mesogen and the backbone are more strongly coupled. Thus the layers may be more distorted in potato than maize. Small-angle X-ray (SAXS) scattering provides a very good tool for studying these issues. Figure 5.5 shows a typical SAXS pattern, in this case for waxy maize in water so that there are no complicating effects from amylose. Using a simple paracrystalline model, based on semicrystalline stacks of alternating amorphous and crystalline lamellae embedded in an amorphous matrix (the amorphous growth ring) (Wenig and Bramer, 1978, Cameron and Donald, 1992) a three phase model to fit the scattering of starch has been developed and found to work well for different starches in different situations (see (Donald et al., 1997) for a review). Figure 5.6 outlines the model: the mesogens sit in the semicrystalline lamellae, the spacers and branch points of the amylopectin in the spacer lamellae, and the amorphous growth ring (amorphous background) must contain any amylose present since it is assumed this does not crystallise within the granule itself.

Fig. 5.5

Small-angle X-ray scattering pattern for waxy maize in water (45% dispersion w/w).

ß 2004, Woodhead Publishing Limited

Fig. 5.6

Schematic representation of the structures present within the starch granule at different lengthscales (after Jenkins et al., 1993).

This schematic representation of the packing within the granule must not be taken too literally. SAXS can only provide average information and, except in the case of microfocus experiments of which very few have been done, the averaging is done across not only a single granule but over many granules. Local variations are necessarily lost in this averaging process. Thus, although the growth rings are drawn simplistically as being continuous around the granule, this is not likely to be accurate ± but SAXS can provide no insight into what discontinuities might be present. Various authors, when discussing the blocklet structure, have implied that blocklets are inconsistent with the model of Fig. 5.6, because blocklets require discontinuities, but in fact there need be no inconsistency. The dimensions quoted for the blocklets, typically 50±250 nm are sufficiently large that their impact on the breadth of the SAXS peak is likely to be small, and SAXS by itself cannot prove or disprove their existence. Returning now to the nature of the flexible spacer, as indicated above this is thought to be shorter for B type starches than for A. The paracrystalline model described in Cameron and Donald (1992) assumes flat layers, which can be accurate only if the double helices are sufficiently decoupled from the backbone. Otherwise, either the double helices cannot line up themselves into layers or the layers must be distorted. Distorted, curved layers are familiar in other branches of biology in the form of cell membranes. Their distortions and fluctuations have been well-modelled (see e.g., (Safran, 1994)). Building on the understanding developed there, it is possible to refine the lamellar model for starch to take the possibility of layer fluctuations into account. What is found in this case (Daniels and Donald, 2003) is that B type starches have a much more distorted lamellar structure than the A type cereal starches, in line with the expectations expressed above regarding the shorter length of the flexible spacer. A key problem in analysing the packing of the amylopectin side chains is that so little is actually known about the spatial distribution of branches. Whereas debranching followed by various chromatographic techniques can provide insight into the distribution of side chain branch lengths, as originally discussed

ß 2004, Woodhead Publishing Limited

by Hizukuri (1986), this only provides average information across many granules. To understand the details of the local packing within each cluster, much more detailed, local information is required ± which as yet is unavailable. Thompson has discussed this problem in the context of how random the branching of the amylopectin molecule is (Thompson, 2000), concluding that the branches are actually not random at all, and that this will relate both to the enzymatic pathways by which the molecule is formed, and also to the packing and consequent properties of the starch. O'Sullivan and Perez (1999) have examined the clustering of branches using computer simulation, and the nature of the clusters in A and B type starches was compared by GeÂrard et al. (2000) using mutant maizes. From this still rather limited number of studies, it is clear that understanding the detail of packing within the granule requires a greater level of detail about the architecture of the amylopectin molecule than is currently available. Although much is known, there is still a long way to go before we have a complete understanding of the interplay between synthesis, chain architecture, packing and consequent properties.

5.3

Improving packing within starch granules

There are various routes to improving the packing within the granule, with corresponding consequences for the ease with which the granule breaks down during subsequent processing. They can broadly be subdivided into classes; one in which dry heat is applied, frequently known as annealing, and one in which a solvent is used to facilitate rearrangements, a process often known as plasticisation when organic solvents are used, or heat-moisture treatments when water is the solvent.

5.3.1 Annealing Annealing refers to heating at low moisture contents; Jacobs and Delcour (1998) have distinguished the case of high moisture (i.e. greater than ~35% moisture) treatments from true annealing. In general, annealing will lead to both an increase in gelatinisation temperature, and a sharpening and increase in magnitude of the gelatinisation endotherm. These thermal properties indicate that annealing is associated with better local packing of the amylopectin side chain branches, facilitating an improvement in local order. Annealing must be carried out over an appropriate temperature range if it is to be effective. Clearly the temperature applied must be above the glass transition temperature Tg, and equally clearly it must lie below the gelatinisation temperature Tgel. Various measurements can be used to explore what effect such annealing treatments have. Wide-angle X-ray scattering (WAXS) indicates that neither crystal type nor percentage crystallinity is altered by annealing (Gough and Pybus, 1971, Stute, 1992). Likewise 13C-CP/MAS-NMR (carbon-13 cross polarisation/magic angle spinning nuclear magnetic resonance) has shown that

ß 2004, Woodhead Publishing Limited

the number of double helices is unchanged by annealing, so that no new double helices can be being formed in the process (Tester et al., 1998). However, smallangle X-ray scattering (SAXS) and DSC experiments reveal alterations have occurred post-annealing, of which DSC has been most extensively applied. SAXS experiments show that the peak, visible in Fig. 5.5 and present in all hydrated wild-type starches, does not shift in position upon annealing. This indicates that the overall lamellar repeat is unchanged in spacing. However, the peak does increase in intensity, indicating an increase in contrast between the amorphous and crystalline lamellae (Jacobs et al., 1998b), consistent with the idea of an increase in perfection of packing within the crystals. This is confirmed by DSC which shows conclusively that changes are occurring during the annealing process, as well as studies on granule stability post-annealing. Although there are differences between species (Tester et al., 2000), a generic interpretation only is presented here, since the basic principles appear consistent across different types of starch. If DSC is used to look at the gelatinisation (see Section 5.4) properties of starches after annealing, it is found that the onset of gelatinisation is delayed by annealing, and the breadth of the gelatinisation endotherm is reduced. However the enthalpy associated with gelatinisation may or may not always increase (Yost and Hoseney, 1986, Stute, 1992, Hoover and Vasanthan, 1994, Jacobs et al., 1995), and the magnitudes of the changes observed depends on how close to the gelatinisation range the annealing is carried out (Tester et al., 2000). The changes observed have been related, as indicated above, to an increase in perfection of packing. Tester and Debon (2000) conclude that annealing leads to enhanced lateral registration of the amylopectin double helices in the crystalline lamellae, possibly accompanied by a slight increase in the average length of double helices, and a higher rigidity of the amorphous regions. Both these alterations would make granule disruption during annealing harder to achieve, as will be discussed below. These ideas of increasing perfection are very similar to ideas about annealing for conventional synthetic crystalline polymers such as polyethylene, for which it is well known that annealing reduces the concentration of defects and, in this case, pushes up the melting point of the crystals (Bassett, 1981).

5.3.2 Plasticisation As indicated at the start of this section, improvements in order can also be achieved in the presence of solvents including (but not limited to) water. The case of starch in glycerol can be used as a useful model system to explore the changes which occur during plasticisation in a general way, and these can also be related to the mechanisms of annealing referred to above. If a SAXS experiment is carried out on starch in glycerol at room temperature, the familiar peak seen in Fig. 5.5 is absent. This is equivalent to the observation that dry starch granules do not show this peak (see Section 5.2 and Fig. 5.4; in dry starch the ordering is nematic, but smectic in the hydrated case). However, whereas in the case of water simply allowing hydration to occur permits the lateral register

ß 2004, Woodhead Publishing Limited

Fig. 5.7

Evolution of the SAXS peak for potato starch (40% w/w starch) mixed with ethylene glycol upon holding at room temperature.

of the double helices to develop, this is not the case for glycerol. No peak is seen when starch is held in glycerol unless the sample is heated to elevated temperatures (above ~80 ëC in the case of waxy maize). Similar effects are seen for a variety of other solvents including sugar solutions and polyols (Perry and Donald, 2000). In all cases, upon initial immersion of the starch granules in the solvent, the 9 nm peak is absent from the SAXS pattern, but either after waiting for prolonged times at room temperature, or by raising the temperature, the peak is seen to develop. As an example, Fig. 5.7 shows, for the specific example of potato starch in ethylene glycol, the development of the peak when the sample is held at room temperature over long periods. Simultaneously, a significant increase in the crystalline peaks is seen in the WAXS patterns. The interpretation of these SAXS findings is that a sufficient amount of the plasticising solvent must enter the granule to facilitate the transition from the disordered, nematic organisation in the dry granule, to the more perfectly ordered smectic phase. The plasticiser acts on the flexible spacers in the amorphous lamellae to permit the double helices to slot into register. As this happens both the interhelix repeat (revealed by the 100 spacing in the WAXS pattern of potato) and the long-range 9 nm repeat develop. What determines the conditions which add up to `a sufficient amount' will depend on the specific solvent and species of starch. Larger solvent molecules, high viscosity solvents, and more densely packed starches will all hinder ingress of the solvent and thus require either longer times or higher temperatures. In other words kinetics matter as well as equilibrium, in determining the state of order in the granule. If the solvent ingress is changing the local packing, one might expect the transitions to show up in DSC also. This is indeed so. Using glycerol, to which

ß 2004, Woodhead Publishing Limited

Fig. 5.8 DSC traces, vertically offset, for 25% waxy maize solutions in aqueous glycerol solutions of the concentrations marked (given in % glycerol) heated at a rate of 10oC/min. The arcs mark the loci of the self-assembly exotherm, and the gelatinisation endotherm as indicated (after Perry and Donald, 2000).

small amounts of water have been added (glycerol and water are miscible over the entire composition range), it is possible to see in Fig. 5.8 an exotherm occurring for solutions of greater than 80% glycerol for samples of waxy maize. The temperature at which this occurs (marked by a curve on the figure indicated as the self-assembly exotherm) depends on the glycerol concentration, moving upwards in temperature as the glycerol concentration increases. This exotherm (which corresponds to heat being given out) occurs significantly below the temperature at which the more familiar gelatinisation endotherm occurs (also shown in Fig. 5.8, and discussed later). The presence of the exotherm confirms that some transition is occurring which leads to an altered state of organisation and hence the release of energy, consistent with the description given above. In fact, for a variety of different starches and different plasticising solutions, there is good correlation between the temperature at which DSC exhibits this exotherm and the appearance of the SAXS peak (Perry and Donald, 2000). Therefore both the DSC and the SAXS support a picture in which at room temperature the flexible spacers are insufficiently plasticised in many organic solutions to give rise to smectic ordering, but as the temperature is raised solvent ingress and accompanying plasticisation and enhanced mobility permit the transition shown in Fig. 5.3, from right to left. Glycerol is a substantially larger molecule than water, and it is not surprising that it enters the granule with more difficulty hence requiring higher temperatures than water. For low glycerol concentrations in water, the exotherm is not seen in DSC, and the SAXS patterns always exhibits the 9 nm peak, since enough solvent can enter the granule in this case.

ß 2004, Woodhead Publishing Limited

Pulling together the ideas from annealing and from plasticising solvents, in both cases it is clear that the organisation within the native granule can be improved by imparting sufficient mobility to the flexible spacers to encourage the double helical mesogens to line up in good register to give both lateral correlations and correlations between the different lamellar layers. Both heat and solvent ingress can contribute to the generation of mobility.

5.4

The gelatinisation process

When heated in the presence of water, starch undergoes an irreversible orderdisorder transition termed gelatinisation. Various changes can be observed: the granules swell, absorb water, lose crystallinity, and leach amylose. Many techniques have been employed to study these events in an attempt to understand the precise structural changes underlying gelatinisation. Techniques used include DSC, X-ray scattering, light scattering, optical microscopy, thermomechanical analysis (TMA) and NMR. Based on these studies various models have been proposed but, as yet, a universally accepted explanation of the gelatinisation process is still lacking. A useful early review describing some of these studies is given in Blanshard (1987). As we have seen above, the starch granule is known to be semicrystalline, exhibiting birefringence when viewed between crossed polars. As the starch granules gelatinise and their structure is disrupted, this birefringence is lost. Many studies have attempted to characterise the point at which all birefringence is lost for a sample studied under an optical microscope. This point is termed the birefringence end point temperature (BEPT). Whereas the loss of birefringence occurs over quite a large temperature range for the whole sample (a BEPT range of 56±64 ëC has been measured for 12% starch (w/w) suspensions of wheat and potato starches heated at 1.5 ëC/min-1 (Hill and Dronzek, 1973)), individual granules are observed to lose birefringence over a much smaller range, generally less than 1 ëC (French, 1984). During the gelatinisation process, substantial swelling occurs, together with loss of crystallinity, as demonstrated by WAXS studies. Crystallinity loss has been quantitatively correlated with thermal events (as measured by DSC) in a detailed study by Liu et al. (1991). A similar investigation was made by Cooke and Gidley (1992) with the additional use of NMR to measure double helix content. They observed that crystalline and molecular order are lost concurrently during gelatinisation. More recently gelatinisation has been studied via the use of synchrotron radiation, which permits real time X-ray studies to be made during the heating process (as opposed to post mortem as in the case of lab-based X-ray sources or NMR measurements). In addition, it is possible in this case to carry out simultaneous SAXS/WAXS/DSC experiments (Jenkins et al., 1994, Jenkins and Donald, 1998). Early results for wheat starch indicated that electron density changes take place first within the amorphous growth ring, suggesting that initial

ß 2004, Woodhead Publishing Limited

water absorption occurs within this region (Cameron and Donald, 1993). This is consistent with the fact that no change in either the 9 nm spacing or the position of the crystal peaks is observed during the early stages of swelling of the whole granule. The initial swelling observed optically occurs without the semicrystalline lamellae expanding radially (LeszczynÄski, 1987). In addition, small angle neutron scattering (SANS), which can be used to examine the distribution of water within the granule, also indicates that the amorphous growth ring is the site of the first increase in water content (Jenkins and Donald, 1998). (Proton NMR has also been used to try to explore changes in the distribution of water during gelatinisation, but with somewhat limited success (Tang et al., 2001).) Pulling together results from many studies it appears that, as originally postulated by Donovan (1979), gelatinisation is primarily a swelling driven process. Water uptake by the amorphous growth ring is accompanied by swelling within this region. Although the semicrystalline lamellar region does not expand radially, the amylopectin molecules at the edges of the lamellar stack will be connected through bonds with the amorphous regions, thereby coupling the semicrystalline regions to the amorphous regions. As a consequence, swelling within the amorphous regions imposes a stress upon the amylopectin crystallites, and ultimately this stress will cause the amylopectin double helices within the crystallites to dissociate, hence leading to the breakdown of granule integrity. This scenario assumes that there is an excess of water present. It has been known for many years that in conditions of limiting water the DSC traces during gelatinisation are altered: the single, so-called G endotherm, now splits and a second M1 endotherm appears. In conditions of limiting water there is insufficient water present for all the crystallites to be disrupted in the manner described above. Indeed what constitutes `limiting water' conditions is usually defined by whether or not a second endotherm is seen (Donovan, 1979). Limiting water can occur either due to a high starch:water ratio, or due to the presence of solutes (including sugars) in aqueous solution. It is generally concluded that in limiting water, melting of some or all of the crystallites must occur as opposed to the swelling-driven process described for excess water (Donovan, 1979). These ideas can also be reformulated within the liquid crystalline model for starch presented earlier. Within this model one can envisage two different processes which are both requisite for loss of all order within the granule, but which need not occur simultaneously. The first is that the double helices, aligned originally in the smectic phase, pull back out of register into the nematic phase under swelling-driven stresses. Secondly, the double helices themselves must unwind, undergoing a helix-coil transition (Waigh et al., 2000a). In general, such a helix-coil transition is expected to be a sharp, cooperative process (Zimm and Bragg, 1959). Within this framework it is clear why there may be only one transition (the G endotherm) during gelatinisation, when both processes occur essentially simultaneously, or two, in which the higher temperature M1 transition corresponds to the unwinding of the double helices, delayed due to the

ß 2004, Woodhead Publishing Limited

lack of mobility in insufficient water/solvent. Further details of this model can be found in Waigh et al. (2000a).

5.4.1 The sequence of plasticisation and gelatinisation and the role of solutes As we have seen for water, plasticisation by water occurs so fast at room temperature that essentially one always sees the smectic organisation. Likewise, during gelatinisation in excess water, the process by which the double helices are disrupted (both in terms of register and the unwinding of the helices themselves) occurs sufficiently rapidly that in DSC experiments only a single endotherm is seen, as observed originally by Donovan (1979). However, for limiting water, water containing solutes such as sugars, and other solvents this situation does not apply. In this case it is necessary to map out the whole set of events which lead to disruption, and several thermal processes can be seen in the DSC experiments. In addition, if prior annealing has taken place, this will also affect the response of the granule to heat (Jacobs et al., 1998b, Tester and Debon, 2000). The sequence of events which must occur are shown in Fig. 5.9. First, the dry unplasticised granule must undergo sufficient plasticisation for the nematic organisation to transform into the smectic form, with its corresponding register. As discussed in the previous section this will require sufficient mobility imparted by a combination of heat and solvent ingress. Thereafter the nematic phase may reform (not shown) as the double helices are pulled out of register, followed by their complete disruption (Perry and Donald, 2002), at which point gelatinisation will be regarded as being complete. So what happens to the effectiveness of water as solutes are added? It is well known that elevation of the gelatinisation temperature occurs as either the molecular weight of the solute or the concentration increases (see e.g. Blanshard, 1987). To explain this, Evans and Haisman (1982) hypothesised that the gelatinisation temperature could be related to water activity and the

Fig. 5.9 Sequential steps required for complete granule disruption to occur. First the double helices must gain sufficient mobility to form the regular lamellar structure. As further heating occurs, the double helices lose registry again and ultimately unwind.

ß 2004, Woodhead Publishing Limited

volume fraction of water in the system. This approach excluded the possibility of gelatinisation in the absence of any water. It has also been suggested (Slade and Levine, 1989) that sugars have an antiplasticising effect when compared with water alone. This approach emphasises the importance of the solute on the glass transition of starch, and considers how (in)effective an aqueous solution will therefore be on the requisite plasticisation before gelatinisation can occur. Building on this framework van Soest et al. (1994) considered how sugars might stabilise the structure of water, with hydrogen bonding between water and solute playing a key role. Alternatively specific starch-solute interactions could occur, thereby stabilising the amorphous regions of the granule and raising the temperature for gelatinisation (Spies and Hoseney, 1982, Ghiasi et al., 1983). Finally one could imagine that the presence of solutes reduces the amount of water available to starch, thereby prompting behaviour reminiscent of limiting water conditions (Labuza, 1975). A key test of what is the mechanism underlying the elevation of Tgel is therefore examination of DSC traces. Figure 5.10 shows the occurrence of a single gelatinisation endotherm for all concentrations of water: solute without the appearance of the additional M1 peak. The peak temperature rises by an amount depending on the sugar and its concentration (further quantitative data can be found in Perry and Donald, (2002)), but there is no sign of a change in peak shape or breadth or of a second higher temperature transition appearing. Even if the concentration of sugar is pushed into the regime in which the starch:water ratio reaches a level at which the sample is effectively in `limiting water' conditions the gelatinisation behavour is not qualitatively altered. Thus gelatinisation in sugar solutions is not equivalent to moving into limiting water conditions. The behaviour of a whole range of different types of solutions can instead be understood by extending the plasticisation model described above, to

Fig. 5.10 Peak temperature of waxy maize starch gelatinisation as revealed by DSC traces for a range of aqueous solutions (after Perry and Donald, 2002).

ß 2004, Woodhead Publishing Limited

consider how the nature of the solution (type of solute, molecular size, and concentration) will affect the initial ingress of the solution into the granule, impede or facilitate the organisation of the lamellae and finally their subsequent breakdown. As shown in Fig. 5.8 (showing glycerol solutions of increasing concentration with waxy maize), although there is not a constant temperature difference between the (lower temperature) self-assembly exotherm and the gelatinisation endotherm, nevertheless they both follow the same upward trend. For both processes a critical degree of plasticisation/mobility is required for initiation. The details of the exact temperature at which this critical degree is reached depends on the effectiveness of the system for plasticisation, which in turn is related the size of solute molecule, possibility of specific interactions, as well as amount present. What matters is the properties of the solution as a whole, and it is wrong to consider the water and the solute as separate entities, with only the water able to cause the plasticisation.

5.4.2 The role of amylose Much of what has been written above, both relating to the overall packing within the granule, and the subsequent gelatinisation behaviour, has been discussed in terms of the amylopectin alone. This may be fine for waxy maize, but obviously is an incomplete description for non-waxy varieties, with typically around 30% amylose present in wild-type starches. Much work on studying the location of amylose within the granule has approached the problem through consideration of mutants whose ability to synthesise either amylose or amylopectin has been severely impaired. However, these studies have the limitation that amylose and amylopectin are not strictly distinct molecules in a certain sense, although it is convenient to classify them as such. Altering the starch enzymatic pathway can produce intermediate molecules which could either be regarded as highly branched amylose or amylopectin with fewer, but longer, side chains, and the whole granule morphology may be significantly altered (Smith, 2001). Nevertheless, broadly speaking, one can conclude that a significant proportion of the amylose must sit in the amorphous growth rings , but whether there is any cocrystallisation, as envisaged by Blanshard in his early model of the cluster packing (Blanshard, 1987) remains unclear. However, some clues about the location and impact of amylose can be found from indirect evidence. Using iodine as a selective stain for amylose, it is possible to examine its distribution within the granule, particularly when its overall level has been decreased by genetic manipulation (Kuipers et al., 1994, Tatge et al., 1999). The presence of `blue rings', which (when the amylose content is significantly reduced) are predominantly confined to the centre of the granule, confirm that the amylose is predominantly located in the (presumed) amorphous growth ring. However there is little difference between Tgel for wildtype and waxy maize species, so this differentiation in location does not appear to be significant for gelatinisation. However, a recent observation has indicated that amylose can interfere with lamellar organisation if the temperature is raised

ß 2004, Woodhead Publishing Limited

into a regime where it may be mobilised (Waigh et al., 2000a). In these experiments both maize and waxy maize were taken up into the gelatinisation range, and then recooled before complete gelatinisation had occurred. In this case, SAXS showed that although the lamellar organisation had disappeared at the elevated temperature, below a certain temperature it was possible to reform the 9 nm repeat simply by dropping the temperature. The critical temperature, beyond which the lamellae were irreversibly lost, was lower for regular maize than for waxy, suggesting that the amylose could move in such a way it was able to block partially the perfect reconstruction of the lamellae. Part of the problem of understanding the role of amylose (aside from the basic question as to why Nature has evolved to produce two such similar polysaccharides within the granule) is that methods to distinguish chemically very similar molecules are currently severely limited. A better understanding of the contribution of amylose to the overall properties of the granule will only be forthcoming when novel approaches enable greater discrimination between amylose and amylopectin in the native granule.

5.5

Food processing: implications of starch granule structure

Starch granules are rarely consumed as intact granules, nor are they frequently used industrially in that form. Processing can take many forms: milling, enzymatic degradation, cooking, extrusion and chemical modification. Chemical modification is itself an enormous subject and will be covered elsewhere later in this book. Most of what follows in this section will deal with physical changes, but enzyme treatments will also briefly be discussed.

5.5.1 Milling Milling, a necessary process for the large scale extraction of cereals such as wheat starch, can lead to substantial damage of the granules. This is both externally visible, as in cracking, but may also include internal alterations. For instance it is known that damage leads to a greater degree of water absorption (Mok and Dick, 1991a), and a consequent increase in the amount of swelling which occurs in water (Mok and Dick, 1991b). Severe milling also reduces the crystallinity measured by WAXS (Lelievre, 1974, Morrison et al., 1994) and the double helix content as measured by 13C-CP/MAS-NMR (Morrison et al., 1994). The explanation for these changes is thought to reside in the generation of shear and compressive structures during the milling process (frequently laboratory tests are carried out using ball milling, which is particularly severe). As these stresses are transmitted to the amorphous regions of the granule, scission within the amylopectin molecule may occur, thereby removing restraints upon swelling and hence facilitating subsequent gelatinisation (Morrison et al., 1994). This is consistent with the substantial decrease in chain length reported by Yamada et al. (1997) for potato starch.

ß 2004, Woodhead Publishing Limited

Considering the gelatinisation of mill damaged starches, it is clear that both the onset temperature and also the enthalpy decrease (Anjum and Walker, 1991, Morrison et al., 1994, Yamada et al., 1997, Tester, 1997). In other words, the changes induced by gelatinisation ± drop in molecular weight, increase in swelling at room temperature, and loss of constraints ± will all render the granule disruption which accompanies gelatinisation easier. And the accompanying reduction in crystallinity will lead to a decrease in the enthalpy associated with gelatinisation. These are general trends. In fact there will obviously be quantitative variations between species and cultivars. For instance it is known that hard wheats (so classified according mainly to their protein content) are more susceptible to mill damage than soft (Mok and Dick, 1991b). Thus for a given duration of milling the alterations in the behaviour of hard wheat cultivars will be more extreme than for soft cultivars. Furthermore waxy maize is known to be more susceptible than wild-type maize (Han et al., 2002).

5.5.2 Amylolysis and hydrolysis Amylolysis is the breakdown of the granule by the action of the enzyme amylase. This enzyme is naturally occurring and is found in both the grain and human saliva. Its action on granules is therefore extremely important. Damaged starch is more susceptible to the enzyme, and therefore this is an additional factor which must be taken into account in controlling milling conditions. As with the other factors discussed in the previous section, it is believed that cracking and increased water uptake lead to the increase in rate of amylolysis upon damage. The products of -amylase attack are glucose, maltose and limit dextrins (in contrast -amylases produce maltose and -limit dextrins, but their action is more restricted (Colonna et al., 1992)). In the human gut, it is difficult to follow the degradation of granules readily, but an important issue surrounds those starches which are known to be more `resistant' to digestion, and whose presence may confer health benefits (Englyst et al., 1992). Thus in vivo as well as in vitro, mill damage of granules leading to rapid amylolysis may have significant impact on functionality and properties of starch. The precise pattern of degradation can be followed by various microscopies, and has been found to vary between species. As we have seen in Section 5.1.2, the enzyme preferentially attacks the amorphous growth rings, as can be seen in the SEM of Fig. 5.1. This is because the amorphous regions will be attacked more readily than the semicrystalline growth rings. However, such images are obtained from sectioned granules; if intact granules are exposed to the enzyme, then the first alteration in structure is seen to be pitting at the surface. In the case of maize these small pits become holes, after which endocorrosion (i.e. from the inside out) appears to occur. In this case the internal preferential loss of material from the amorphous growth rings becomes clear. However potato starch behaves differently, with substantial cracking occurring at the surface first (exocorrosion) only followed later by endocorrosion (Mathias et al., 1997).

ß 2004, Woodhead Publishing Limited

The different behaviours of cereal (A-type) and potato (B-type and certain mutants of maize) have formed the focus of a body of work from the group of Colonna, which has also covered parallel studies of acid hydrolysis, although this work has tended to concentrate on `model' granules, rather than native ones. The conclusion reached in each of the studies was that A-type starches were always more susceptible to degradation than B-type, with the suggestion that the resistance to amylolysis was related to the distribution of B-type crystallites rather than just their number (Gerard et al., 2001, Planchot et al., 1997b). Various possibilities were suggested for what was perceived as an intrinsic greater resistance of the B-type crystals, including three-dimensional size, their size along the helical length, and a greater crystal perfection (Planchot et al., 1997a). Hydrolysis more usually refers to the action of acid upon starches, and can be sub-divided into the action of sulphuric acid which produces so-called NaÈgeli amylodextrins (NaÈgeli, 1874) ± and hydrochloric acid (usually at temperatures slightly above room temperature) which leads to lintnerised starches (Lintner, 1886). These two routes for hydrolysis have formed the focus of much work over the years, both because they provide a convenient way of exploring the packing of the granule, and also because the end-products have commercial uses. In general a two stage hydrolysis route is seen to apply, postulated by Robin et al., 1974, 1975a) and subsequently supported by others (Muhr et al., 1984); the first stage is associated with hydrolysis of the amorphous regions, whereas the subsequent slower second stage relates to degradation of more crystalline parts of the sample. One consequence of the loss of amorphous regions initially is an apparent increase in crystallinity, but a concomitant decrease in the SAXS peak (Muhr et al., 1984). This observation can be rationalised by recognising that as the amorphous growth rings are lost, the apparent magnitude of the crystallinity increases. However, the loss of connectivity imposed by the degradation of the amorphous growth rings, also means that there is a reduction in the electron density differences between the different regions of the granule, and this has the effect of reducing the SAXS peak height (Jenkins and Donald, 1997). In addition, there will be loss of constraints as the amylopectin chains themselves are degraded, which may also lead to loss of long range correlations, as well as an increasing ability of the granule to swell as it takes up water. Annealing, as described in Section 5.3.1, has the effect of improving local packing. For potato starch it has been found that annealing has the effect of reducing the rate of hydrolysis, but not for wheat or pea starch. This can be explained by considering that these starches are readily hydrolysed anyhow, and any improvements in local packing are still insufficient to reduce the rate of hydrolysis, whereas for B-type potato (which is more resistant to acid attack anyhow) annealing can produce a significant effect (Jacobs et al., 1998a). These various treatments can therefore be seen to make significant alterations to granule structure, and therefore corresponding changes in the ease with which other subsequent processes can occur. For instance, gelatinisation temperature ranges will be modified by hydrolysis (or amylolysis), although the nature of the

ß 2004, Woodhead Publishing Limited

changes in this appear to be species-specific, as well as obviously dependent on extent of hydrolysis (Muhr et al., 1984, Shi and Seib, 1992). The structural changes occurring during these different processing treatments, must therefore be taken into account when considering functionality.

5.5.3 Extrusion and plasticisation Extrusion is a common and very important means of processing starch, as much outside the food industry as within. It is not possible to cover the substantial literature here, much of which does not relate specifically to the role of the granule itself. Only a very brief account of the extrusion process is given here, relating directly to the breakdown of the granule under different conditions of temperature, plasticisation and mechanical input. Extrusion combines a high temperature with a large amount of mechanical energy input. Different extruder configurations lead to different forms of mechanical disruption, as can be seen from comparing twin and single screw extruders (see e.g. Pearson (1985)). Melting/plasticisation occurs in a very short transition section, after which the material is compacted under a high shear rate. After starch extrusion, all granular features of the starch should be lost, or inhomogeneities will remain in the extrudate. Exactly how the granule breaks down is still not entirely clear. In general either water (for the food industry) or some other plasticiser (such as glycerol in the manufacture of thermoplastic starch (Tomka, 1991)) is added to facilitate breakdown at temperatures below that at which degradation of the macromolecules will occur. Recent work has suggested the following series of events during destructuring of the granule during extrusion of pea and waxy maize starch (Barron et al., 2001). First the granule is disrupted and fragmented, due to the action of shear, and there is loss of organisation starting in the amorphous regions. As shearing proceeds, breakup between the crystalline lamellae was found to occur; in other words the amorphous lamellae act as sites of weakness for fracture. This type of fracture also implies the amylopectin molecules themselves are broken, facilitating crystallite breakdown. That extrusion leads to molecular weight degradation has been well documented (Colonna et al., 1984, Diosady et al., 1985), with greater degradation occurring at high temperatures, high screw speeds and lower water contents all conditions favouring a high viscosity melt. Ultimately a fully amorphous, homogeneous melt results. By comparing the two starches (Barron et al., 2001) it was concluded that the presence of amylose might be leading to a higher melt viscosity with correspondingly more heat dissipation in the molten phase, favouring a greater loss of crystallinity in the pea starch compared with waxy maize.

5.6

Conclusions and future trends

Starch has been the subject of research for over 100 years because of its huge importance to mankind. Our understanding has improved steadily, with many

ß 2004, Woodhead Publishing Limited

new experimental techniques being brought to bear on our understanding of the internal arrangements of the molecules within the starch granule. Much progress has been made, but substantial difficulties remain. Firstly, although we understand broadly the different levels of organisation within the granule, there are still uncertainties ± for instance can Oostergetel's model of superhelices (Oostergetel and Bruggen, 1993) be reconciled with the idea of blocklets (Gallant et al., 1997)? Secondly, the detailed distribution of amylose is not known, and in the case of mutants it is hard even to know exactly what the molecular make-up of the granule is, when molecules intermediate between amylose and amylopectin may form. And, perhaps most importantly, we have no quantitative predictive capability when it comes to unraveling precise transition temperatures (e.g. for gelatinisation) or lengthscales (how far apart are amorphous growth rings?) in the wide range of different species. As biochemists interrogate the enzymatic pathways more fully, correlations between different types of structures with molecular architecture may provide further clues, but we are certainly still a long way off a complete understanding of the granule structure at all levels, and how this structure impacts on functionality and subsequent end-use.

5.7

Sources of further information and advice

There are some useful early reviews to be found in French (1984), Blanshard (1987), Zobel (1988). Readers who want to understand the interplay between biosynthesis and starch granule structure should refer to Smith (2001). The books based around the International Conferences on Starch Structure and Functionality held in Cambridge UK, in 1996 and 2000 are also useful references (Frazier et al., 1997, Barsby et al., 2001), including some of the papers referred to in this chapter. Useful websites include those attached to the following Universities: Iowa (http://www.public.iastate.edu/~pkeeling/Welcome.htm), Purdue (http://www.whistlercenter.purdue.edu/Research _Focus.htm) Cambridge (http://www.poco.phy.cam.ac.uk/research/starch/starch1.htm) Southern Illinois (for an archaelogical context http://www.siu.edu/~ebl/ amylose.htm) and attached to various Institutes (International Starch Institute, Denmark http://www.starch.dk, Royal Veterinary and Agricultural University, Denmark http://www.plbio.kvl.dk/plbio).

5.8

References

ANJUM, F. M. and WALKER, C. E., (1991), `Review on the significance of starch and protein to

wheat kernel hardness', Journal of the Science of Food and Agriculture, 56, 1±13. and ROBARDS, A. W., (1998), `An experimentally-based predictive model for the separation of amylopectin subunits during starch gelatinization', Carbohydrate Polymers, 36, 173±192. ATKIN, N. J., CHENG, S. L., ABEYSEKERA, R. M. and ROBARDS, A. W., (1999), `Localisation of

ATKIN, N. J., ABEYSEKERA, R. M., CHENG, S. L.

ß 2004, Woodhead Publishing Limited

amylose and amylopectin in starch granules using enzyme-gold labelling', Starch/ StaÈrke, 51, 163±172. BADENHUIZEN, N. P., (1936), `Abhandlung zur physikalischen Chemie der StaÈrke und der Brotbereitung. XXV Weiter Beobachtungen uÈber die BloÈckchenstruktur der StaÈrkerkoÈrner', Z fuÈr Physik Chemie, A175, 383±95. BADENHUIZEN, N. P., (1939), `Growth and corrosion of the starch granule in connection with our present knowledge of the microscopical and chemical organisation.' Zeitschrift f. Bot., 33, 140±468. BADENHUIZEN, N. P., (1959), In Handbuch der Protoplasmaforschung (eds, Heiltrun, L. V. and Weber, F.) Springer-Verlag, pp. 1±73. BAKER, A. A., MILES, M. J. and HELBERT, W., (2001), `Internal structure of the starch granule revealed by AFM', Carb. Res., 330, 249±256. BALDWIN, P. M., ADLER, J., DAVIES, M. C. and MELIA, C. D., (1994), `Holes in starch granules: confocal SEM and light microscopy studies of starch granule structure', Starch, 46, 341±346. BANKS, W. and MUIR, D., (1980), In The Biochemistry of Plants, Vol 3: Carbohydrates: structure and function, Vol. 321±69 (ed. Preiss, J.) Academic Press, New York. BANKS, W., GEDDES, R., GREENWOOD, C. T. and JONES, I. G., (1972), `Physicochemical studies on starches', Starch, 24, 245±280. BARRON, C., BOUCHET, B., DELLA VALLE, G., GALLANT, D. and PLANCHOT, V., (2001), `Microscopical study of the destructuring of waxy maize and smooth pea starches by shear and heat at low hydration', J Cer Sci, 33, 289±300. BARSBY, T. L., DONALD, A. M. and FRAZIER, P. J. (eds.) (2001) Starch Advances in Structure and Function, Cambridge RSC. BASSETT, D., (1981), Principles of Polymer Morphology, Cambridge, CUP. BLANSHARD, J. M. V. (1987) In Starch: Properties and Potential (ed. Galliard, T.) Wiley, pp. 17±78. BLANSHARD, J. M. V., BATES, D. R., MUHR, A. H., WORCESTER, D. L. and HIGGINS, J. S., (1984), `Small angle neutron scattering studies of starch granule structure', Carbohydrate Polymers, 4, 427±442. BOGRACHEVA, T., WANG, T., MORRIS V.J., RING, S. G. and HEDLEY, C. L., (1998), `The granular structure of C-type pea starch and its role in gelatinisation.' Biopolymers, 45, 323± 32. BULEON, A., GERARD, C., RIEKEL, C., VUONG, R. and CHANZY, H., (1998), `Details of the crystalline ultrastructure of C-starch granules revealed by synchrotron microfocus mapping', Macromols, 31, 6605±10. BUTTROSE, M. S., (1960), `Submicroscopic development and structure of starch granules in cereal endosperms', Journal of Ultrastructure Research, 4, 231±257. BUTTROSE, M. S., (1962), `Influence of environment on the shell structure of starch granules', J.Cell Biol., 14, 159±167. BUTTROSE, M. S., (1963), `Electron microscopy of acid-degraded starch granules', Sta È rke, 15, 85±92. CAMERON, R. E. and DONALD, A. M., (1992), `A SAXS study of the annealing and gelatinisation of starch', Polymer, 33, 2628±35. CAMERON, R. E. and DONALD, A. M., (1993), `A SAXS study of the absorption of water into the starch granule', Carb Res, 244, 225±236. COLONNA, P., DOUBLIER, J. L., MELCION, J. P., DE MONREDON, F. and C., M., (1984), `Extrusion cooking and drum drying of wheat starch. I. Physical and macromolecular modifications', Cereal Chem., 61, 538±543.

ß 2004, Woodhead Publishing Limited

and BULEON, A., (1992), `Limiting factors of starch hydrolysis', European Journal of Clinical Nutrition, 46, S17±S32. COOKE, D. and GIDLEY, M. J., (1992), `Loss of crystalline and molecular order during starch gelatinisation: origin of the enthalpic transition', Carbohydrate Research, 227, 103±112. DANIELS, D. R. and DONALD, A. M., (2003), `An improved model for analysing the SAXS of starch granules', Biopolymers, 69, 165±75. DIOSADY, L. L., PATON, D., ROSEN, N., RUBIN, L. J. and ATHANASSOULIAS, C., (1985), `Degradation of wheat starch in a single-screw extruder: mechano-kinetic breakdown of cooked starch', J. Food Sci., 50, 1697±1699. DONALD, A. M. and WINDLE, A. H., (1992), Liquid Crystalline Polymers, CUP. DONALD, A. M., WAIGH, T. A., JENKINS, P. J., GIDLEY, M. J., DEBET, M. and SMITH, A. (1997) In Starch Structure and Function (eds, Frazier, P. J., Donald, A. M. and Richmond, P.) Cambridge RSC. DONOVAN, J., (1979), `Phase transitions of the starch water system', Biopolymers, 18, 263±75. ENGLYST, H. N., KINGMAN, S. M. and CUMMINGS, J. H., (1992), `Classification and measurement of nutritionally important starch fractions', European Journal of Clinical Nutrition, 46, 533±550. EVANS, I. and HAISMAN, D., (1982), `The effect of solutes on the gelatinisation temperature range of potato starch', StaÈrke, 34, 224±231. FRAZIER, P. J., RICHMOND, P. and DONALD, A. M. (eds.) (1997) Starch Structure and Functionality, Cambridge RSC. FRENCH, D. (1984) In Starch: Chemistry and Technology (eds, Whistler, R. L., BeMiller, J. N. and Paschall, E. F.) Academic Press, Inc., London, pp. 183±247. FREY-WYSSLING, A. and BUTTROSE, M. C., (1961), `Makromolekulare Feinstlamellen in den KoÈrnern der KartoffelstaÈrke', Makromolekulare Chem, 44, 173±8. COLONNA, P., LELOUP, V.

FULTON, D. C., EDWARDS, A., PILLING, E., ROBINSON, H. L., FAHY, B., SEALE, R., KATO, L.,

and SMITH, A. M., (2002), `Determination of Starch Granule Morphology in Potato', J Biol Chem, 277, 10834±41. GALLANT, D. and GUILBOT, A., (1973), `De veloppement des connaissances dur l'ultrastructure du grain d'amidon', Starch, 25, 335±42. GALLANT, P. D. and MASSY, A. G., (1969), `Etude de l'ultrastructure du grain d'amidon aÁ l'aide de nouvelles meÂthodes de preÂparation en microscopic eÂlectronique', Di StaÈrke, 156±163. GALLANT, D. J., BOUCHET, B., BULEON, A. and PEREZ, S., (1992), `Physical characteristics of starch granules and susceptibility to enzymatic degradation', European Journal of Clinical Nutrition, 46, S3±S16. GALLANT, D. J., BOUCHET, B. and BALDWIN, P. M., (1997), `Microscopy of starch: evidence of a new level of granule organization', Carb Poly, 32, 177±191. GEÂRARD, C., PLANCHOT, V., COLONNA, P. and BERTOFT, E., (2000), `Relationship between branching density and crystalline structure of A and B type maize mutant starches', Carb Res, 326, 130±44. GEÂRARD, C., COLONNA, P., BULEON, A. and PLANCHOT, V., (2001), `Amylolysis of maize mutant starches', J Sci Food Agri, 81, 1281±7. GHIASI, K., HOSENEY, R. and VARIANNOMARSTON, E., (1983), `Effects of flour components and dough ingredients on starch gelatinisation', Cer Chem, 60, 58±61. GIDLEY, M., (1987), `Factors affecting the crystalline type (A-C) of native starches and model compounds: A rationalisation of observed effects in terms of polymorphic structures.' Carb Res, 161, 301±4. DONALD, A. M., GEIGENBERGER, P.

ß 2004, Woodhead Publishing Limited

and PYBUS, J. N., (1971), `Effect on the gelatinisation temperature of wheat starch granules of prolonged treatment with H2O at 50 ëC', Starch, 23, 210. HAN, X., CAMPANELLA, O., MIX, N. and HAMAKER, B., (2002), `Consequences of starch damage on rheological properties of maize starch pastes', Cer Chem, 79, 897±901. È ber Versuche die gewachsene Struktur des HANSON, E. and KATZ, J., (1934), `U StaÈrkerkorns mikroskopisch sichtbar zu machen, besonders an lintnerisierter StaÈrke.' Z fuÈr Physik Chemie, A168, 339±52. HILL, R. D. and DRONZEK, B. L., (1973), `Scanning electron microscopy studies of wheat, potato and corn during gelatinisation.' StaÈrke, 25, 367±372. HIZUKURI, S., (1985), `Relationship between the distribution of the chain length of amylopectin and the crystalline structure of starch granules', Carbohydrate Research, 141, 295±306. HIZUKURI, S., (1986), `Polymodal distribution of the chain length of amylopectins and its significance', Carb. Res., 147, 342±347. HIZUKURI, S. and MAEHARA, Y., (1990), `Fine structure of wheat amylopectin: the mode of A to B chain binding', Carb. Res., 206, 145±159. HIZUKURI, S., KANEKO, T. and TAKEDA, Y., (1983), `Measurement of the chain length of amylopectin and its relevance to the origin of crystalline polymorphism of starch granules', Biochima and Biophysica Acta, 760, 188±191. HOOVER, R. and VASANTHAN, T., (1994), `Effect of heat-moisture treatment on the structure and physicochemical properties of cereal, legume, and tuber starches', Carbo. Res., 252, 33±53. IMBERTY, A., BULEON, A., TRAN, V. and PEREZ, S., (1991), `Recent advances in knowledge of starch structure', Starch, 43, 375±84. JACOBS, H. and DELCOUR, J., (1998), `Hydrothermal modifications of granular starch, with retention of the granular structure: a review', J Agric Food Chem, 46, 2895±2905. JACOBS, H., EERLINGEN, R., CLAUWAERT, W. and DELCOUR, J., (1995), `Influence of annealing on the pasting properties of starches from varying botanical sources.' Cer Chem, 72, 480±7. JACOBS, H., EERLINGEN, R. C., SPAEPEN, H., GROBET, P. J. and DELCOUR, J. A., (1998a), `Impact of annealing on the susceptibility of wheat, potato and pea starches to hydrolysis with pancreatin', Carbohydrate Research, 305, 193±207. JACOBS, H., MISCHENKO, N., KOCH, M. H. J., EERLINGEN, R. C., DELCOUR, J. A. and REYNAERS, H., (1998b), `Evaluation of the impact of annealing on gelatinisation at intermediate water content of wheat and potato starches: A differential scanning calorimetry and small angle X-ray scattering study', Carbohydrate Research, 306, 1±10. JANE, J.-L., KASEMUWAN, T., LEAS, S., ZOBEL, H. and ROBYT, J., (1994), `Anthology of starch granule morphology by scanning electron microscopy', Starch, 46, 121±9. JANE, J.-L., WONG, K.-S. and MCPHERSON, A., (1997), `Branch structure difference in starches of A- and B-type X-ray patterns revealed by their NaÈgeli dextrins', Carb Res, 300, 219±27. JENKINS, P. J. and DONALD, A. M., (1997), `The effect of hydrolysis on native starch granule structure', StaÈrke, 49, 462±7. JENKINS, P. J. and DONALD, A. M., (1998), `Gelatinisation of starch: a combined SAXS/ WAXS/SANS study', Carb Res, 308, 133±147. JENKINS, P. J., CAMERON, R. E. and DONALD, A. M., (1993), `A universal feature in the structure of starch granules from different botanical sources', StaÈrke, 45, 417±20. JENKINS, P. J., CAMERON, R. E., DONALD, A. M., BRAS, W., DERBYSHIRE, G. E., MANT, G. R. and RYAN, A. J., (1994), `In situ simultaneous small and wide angle Xray scattering: a GOUGH, B. M.

ß 2004, Woodhead Publishing Limited

new technique to study starch gelatinisation', J Poly Sci Phys Ed., 32, 1579±83. and FRENCH, D., (1972), `NaÈgeli amylodextrin and its relationship to starch granule structure II. Role of water in crystallization of B-starch.' Biopolymers, 11, 2241±2250. KATZ, J. R. and VAN ITALLIE, T. B., (1930), `Abhandlungen zur physikalischen Chemie der StaÈ rke und der Brotbereitung. V. Alle StaÈ rkearten haben das gleiche Retrogradationsspektrum', Z. Physikal Chem., 150, 90±99. KUIPERS, A. G. J., JACOBSEN, E. and VISSER, R. G. F., (1994), `Formation and deposition of amylose in the potato tuber starch granule are affected by the reduction of granulebound starch synthase gene expression.' The Plant Cell, 6, 43±52. LABUZA, T. (1975), in Water Relations in Foods (ed., Duckworth, R.) Academic Press, New York. LELIEVRE, J., (1974), `Starch Damage', Sta È rke, 26, 85±88. Ä SKI, W., (1987), `New methods for determination of starch gelatinisation LESZCZYN temperatures', StaÈrke, 39, 375±378. LINTNER, C. J. (1886), J. Prakt. Chem., 34, 378. LIU, H., LELIEVRE, J. and AYONG-CHEE, W., (1991), `A study of starch gelatinisation using differential scanning calorimetry, X-ray and birefringence measurements', Carbohydrate Research, 210, 79±87. MATHIAS, P., BAILEY, K., MCEVOY, J., CUFFE, M., SAVAGE, A. and ALLEN, A. (1997), in Starch Structure and Functionality (eds, Frazier, P. J., Donald, A. M. and Richmond, P.) RSC, Cambridge. MOK, C. and DICK, J. W., (1991a), `Moisture absorption of damaged wheat starch', Cereal Chemistry, 68, 405±409. MOK, C. and DICK, J. W., (1991b), `Response of starch of different wheat classes to ball milling', Cereal Chemistry, 68, 409±412. MORRISON, W. R., TESTER, R. F. and GIDLEY, M. J., (1994), `Properties of damaged starch granules II. crystallinity, molecular order and gelatinization of ball-milled starches.' J Cer Sci, 19, 209±17. MUHR, A. H., BLANSHARD, J. M. V. and BATES, D. R., (1984), `The effect of lintnerisation on wheat and potato starch granules', Carb Poly, 4, 399±425. È GELI, W. (1874), Liebigs Annalen der Chemie, 173, 218. NA OOSTERGETEL, G. T. and VAN BRUGGEN, E. F. J., (1993), `The crystalline domains in potato starch granules are arranged in a helical fashion.' Carb Poly, 21, 7±12. O'SULLIVAN, A. C. and PEREZ, S., (1999), `The Relationship between Internal Chain Length of Amylopectin and Crystallinity in Starch', Biopolymers, 50, 381±390. PEARSON, J., (1985), Mechanics of polymer processing, London, Elsevier. PERRY, P. A. and DONALD, A. M., (2000), `The role of plasticisation in starch granule assembly', Biomacromols, 1, 424±32. PERRY, P. A. and DONALD, A. M., (2002), `The Effect of Sugars on the Gelatinisation of Starch', Carb Res, 49, 155±65. PLANCHOT, V., COLONNA, P. and BULEON, A., (1997a), `Enzymatic hydrolysis of a-glucan crystallites', Carb Res, 298, 319±26. PLANCHOT, V., COLONNA, P., BULEON, A. and GALLANT, D. J. (1997b), in Starch Structure and Functionality (eds, Frazier, P. J., Donald, A. M. and Richmond, P.) RSC, Cambridge. RANI, M. R. S. and BHATTACHARYA, K. R., (1995), `Microscopy of rice starch granules during cooking', Starch, 47, 334±337. RIDOUT, M., GUNNING, A., PARKER, M., WILSON, R. and MORRIS, V.J., (2002), `Using AFM to KAINUMA, K.

ß 2004, Woodhead Publishing Limited

image the internal structure of starch granules', Carb Poly, 50, 123±132. and GUILBOT, A., (1974), `Lintnerized starches. Gel filtration and enzymatic studies of insoluble residues from prolonged acid treatment of potato starch', Cereal Chemistry, 51, 389±405. ROBIN, J. P., MERCIER, C., DUPRAT, F., CHARBONNIeÁRE, R. and GUILBOT, A., (1975a), `Lintnerized starches. Chromatographic and enzymatic studies of insoluble residues from acid hydrolysis of various cereal starches, particularly waxy maize starch', StaÈrke, 27, 36±45. ROBIN, J. P., MERCIER, C., DUPRAT, F., CHARBONNIERE, R. and GUILBOT, A., (1975b), `Amidons Lintnerises', Die StaÈrke, 27, 36±45. SAFRAN, S., (1994), Statistical thermodynamics of surfaces, interfaces and membranes, Addison Wesley. SHI, Y.-C. and SEIB, P. A., (1992), `The structure of four waxy starches related to gelatinisation and retrogradation.' Carbohydrate Research, 227, 131±145. SLADE, L. and LEVINE, H., (1989), in Frontiers in Carbohydrate Research -1: Food Applications (eds, Milane, R. P., Bemiller, J. N. and Chandrasekaran, R.) Elsevier Applied Science, London. SMITH, A. M., (2001), `The Biosynthesis of Starch Granules', Biomacromolecules, 2, 335± 41. SOEST, J. J. G. VAN, WIT, D. D., TOURNOIS, H. and VLIEGENTHART, J. F. G., (1994), `The influence of glycerol on structural changes in waxy maize starch as studied by Fourier transform infra-red spectroscopy', Polymer, 35, 4722±4727. SPIES, R. and HOSENEY, R., (1982), `Effect of sugars on starch gelatinisation', Cer Chem, 59, 128±31. STERLING, C. J., (1962), `A low angle spacing in starch', J.Poly. Sci., 56, S10±12. STUTE, R., (1992), `Hydrothermal modification of starches ± the difference between annealing and heat-moisture treatment.' Starch, 44, 205±14. TANG, H.-R., BRUN, A. and HILLS, B., (2001), `A proton NMR relaxation study of the gelatinisation and acid hydrolysis of native potato starch', Carb Poly, 46, 7±18. TATGE, H., MARSHALL, J., MARTIN, C., EDWARDS, E. and SMITH, A., (1999), `Evidence that amylose synthesis occurs within the matrix of the starch granule in potato tubers', Plant Cell and Env, 22, 543±50. TESTER, R., (1997), `Properties of damaged starch granules: composition and swelling properties of maize, rice, pea and potato starch fractions in water at various temperatures', Food Hydrocolloids, 11, 291±301. TESTER, R. F. and DEBON, S. J., (2000), `Annealing of starch ± a review', International Journal of Biological Macromolecules, 27, 1±12. TESTER, R. F., DEBON, S. and KARKALAS, J., (1998), `Annealing of wheat starch', J Cer Sci, 28, 259±72. TESTER, R. F., DEBON, S. and SOMMERVILLE, M., (2000), `Annealing of maize starch', Carb Poly, 42, 287±99. THOMPSON, D. B., (2000), `On the non-random nature of amylopectin branching', Carbohydrate Polymers, 43, 223±239. TOMKA, I., (1991), in Water Relationships in Starch (eds, Levine, H. and Slade, L.) Plenum, New York, pp. 627±637. WAIGH, T. A., JENKINS, P. J. and DONALD, A. M., (1996), `Quantification of water in carbohydrate lamellae using SANS', Far Disc, 103, 325±337. WAIGH, T. A., HOPKINSON, I. M., DONALD, A. M., BUTLER, M. F., HEIDELBACH, F. and RIEKEL, C., (1997), `Analysis of the native structure of starch granules with X-ray microfocus ROBIN, J. P., MERCIER, C., CHARBONNIERE, R.

ß 2004, Woodhead Publishing Limited

diffraction', Macromols, 30, 3813±20. and DONALD, A. M., (1998), `Side chain liquid crystalline properties of starch', Macromols, 31. WAIGH, T. A., DONALD, A. M., HEIDELBACH, F., RIEKEL, C. and GIDLEY, M. J., (1999), `Analysis of the native structure of starch granules with SAXS microfocus diffraction', Biopolymers, 49, 91±105. WAIGH, T. A., GIDLEY, M. J., KOMANSHEK, B. U. and DONALD, A. M., (2000a), `The phase transformations in starch during gelatinisation: a liquid crystalline approach', Carb Res, 328, 165±76. WAIGH, T. A., KATO, K. L., DONALD, A. M., GIDLEY, M. J. and RIEKEL, C., (2000b), `Side-Chain Liquid-Crystalline Model for Starch', StaÈrke, 52, 252±60. WAIGH, T. A., RIEKEL, C., GIDLEY, M.

WANG, T. L., BARBER, L., CRAIG, J., DENYER, K., HARRISON, C., LLOYD, J. R., MACLEOD, M., SMITH, A. and HEDLEY, C. L., (1997), in Starch Structure and Functionality (eds, Frazier, P. J., Richmond, P. and Donald, A. M.) RSC, Cambridge, pp. 188±95. WENIG, W. and BRAMER, R., (1978), `Die Bedeutung der mittleren probendichte fu È r die quantitative interpretation der rontgenkleinwinkelstreuung teilkristalliner polymerer', Colloid and Poly. Sci., 256, 125. YAMADA, T., TAMAKI, S., HISAMATSU, M. and TERANISHI, K., (1997), in Starch Structure and Functionality (eds, Frazier, P., Richmond, P. and Donald, A.) RSC, Cambridge. YOST, D. A. and HOSENEY, R. C. (1986), `Annealing and Tg of starch', Starch, 38, 289. ZEEMAN, S. C., TIESSEN, A., PILLING, E., KATO, K. L., DONALD, A. M. and SMITH, A. M., (2002), `Starch synthesis in arabidopsis. Granule synthesis, composition and structure', Plant Phys, 129, 516±29. ZIMM, B. and BRAGG, J., (1959), J Chem Phys, 31, 526. ZOBEL, H. F. (1988), `Molecules to granules: a comprehensive starch review', Starch, 40, 44.

ß 2004, Woodhead Publishing Limited

6 Measuring starch in food M. Peris-Tortajada, Polytechnic University of Valencia, Spain

6.1

Introduction

6.1.1 Why analyze for starch content? Starch is the most important, abundant, digestible food polysaccharide and is therefore a major source of energy in our diets. Common food starches are derived from seed (wheat, maize, rice, barley) and root (potato, cassava/tapioca) sources. On the other hand, they are modified to improve desired functional characteristics and are added in relatively small amounts to foods as food additives; for instance, starch is used as thickening and gelling agent in sweetened fruit preparations, its analysis being of interest regarding regulatory aspects, and a technological necessity because it may modify the rheology of products into which sweetened fruit preparations are incorporated. With all these features in mind, there is clearly a need for both qualitative and quantitative determination of starch; not only the amount of starch present in the food sample, but also the types of starch (modified starch, resistant starch, etc.) are key parameters in food quality control. It is undoubtedly a very important carbohydrate in many foodstuffs, be it as a natural component or as an additive. In this last case, it is often necessary to determine the proportions of amylose and amylopectine (with different physicochemical properties) in starchy foods, since they may vary according to the specific purposes for which these types of starch are produced.

6.1.2 Regulations Relevant institutions and governing bodies have established rules and recommendations with respect to the methodologies for starch analysis in

ß 2004, Woodhead Publishing Limited

foodstuffs adopted by them. Details may be found in the references and in the specialized literature; we will mention only the more important ones. The Association of Official Analytical Chemists (AOAC) has adopted several methods for official action in the determination of starch in most foodstuffs in which it occurs. The reader is then directed to its last compendium of official methods (Horwitz, 2000) to update the official, recommended procedures. Depending on the type of food to be analyzed, a determined method is proposed, some of them with certain modifications with respect to the original procedure. On the other hand, the American Association of Cereal Chemists (AACC) also approved standard methods for the determination of starch in cereals and derivatives. They have been recently revised and updated (AACC, 2000), and are to be applied to nearly all products containing starch. Similar action has also been taken by the International Association for Cereal Science and Technology.

6.2

Sample preparation

The starch content of most foods cannot be determined directly because the starch is contained within a structurally and chemically complex food matrix. In particular, starch is often present in a semi-crystalline form, be it granular or retrograded starch (Tsuge et al., 1990), that is inaccessible to the chemical reagents used to determine its concentration. It is therefore necessary to isolate starch from the other components present in the food matrix prior to carrying out a starch analysis. As in most quantitative determinations, it is clear that the sample treatments may affect the analytical results, thus it should be considered a key step in the chemical analysis of starch, especially given the complexity of the matrix in foodstuffs. A general scheme for sample pretreatment in the determination of starch in foods in the presence of sugars and dietary fiber is shown in Fig. 6.1. Obviously, some steps may vary slightly according to the type of food analyzed; moreover, additional, less important actions should also be taken in some cases. But the sample pretreatment remains basically the same in nearly all instances. In principle it comprises three basic stages: removal of the soluble sugars, disintegration of the sample and solubilization of the starch, and conversion of the starch into glucose (for its subsequent quantification). First of all, care should be taken to grind the sample in a Wiley mill if it is not in powder form. In practice, nearly all products except canned foods are milled to a fine powder. Canned foods are homogenized in a high-speed blender. High-fat foodstuffs (e.g. crackers, cookies) are defatted with n-hexane or petroleum ether. The extract is discarded and the residue dried overnight, followed by a further extraction with 80% aqueous methanol or ethanol. After centrifugation, soluble sugars remain in the solution, whereas the residue obtained (containing the starch) is gelatinized in water at 121 ëC and then incubated with pH 4.8 acetate buffer and an enzyme (usually amyloglucosidase or a mixture of amyloglucosidase and -amylase) solution. In this way, the starch is converted into glucose,

ß 2004, Woodhead Publishing Limited

Fig. 6.1

General schematic flowchart for sample pretreatment in the determination of starch in foodstuffs.

which ± after centrifugation ± remains in the supernatant. Once dried, it is ready for analysis by any of the analytical techniques mentioned in Section 6.3. For sample disintegration and starch solubilization other protocols have been described, including autoclaving, and treatment with dispersants such as dimethyl sulfoxide (DMSO), alone or in mixtures with water. The former is, however, usually insufficient to obtain a good accessibility of the starch for the enzymatic hydrolysis and results therefore in too low measured starch contents.

ß 2004, Woodhead Publishing Limited

In certain cases, such as the Ewers' polarimetric method (see pages 191±2), there are two alternatives to the enzymatic conversion of the starch into glucose: (i) dissolution of the starch with hydrochloric acid (ii) dissolution of the starch with a boiling calcium chloride solution. In both instances, dissolved proteinanceous substances are precipitated with Carrez solution (I and II, 30% zinc acetate and 15% potassium ferrocyanide, respectively) and filtered off, prior to the analysis of glucose in the filtrate. In general, acid hydrolysis (Faithfull, 1990) (Rose et al., 1991), which is easier to carry out, is utilized in the case of foods with a low content of non-starch polysaccharides, whereas enzymatic hydrolysis is suitable for most foods, hence it is the most recommended and widely used method for sample pretreatment in the analysis of starch in foodstuffs. The conversion of starch into glucose has recently been discussed (Brunt et al., 1998). With HCl complete conversion can be achieved, but its lack of selectivity may be a serious drawback. Starch contents measured are often unexpectedly high due to glucose which is set free from non-starch polysaccharides, occurring in the sample, that are partially degraded by the hydrochloric acid. On the other hand, the use of a mixture of enzymes has proved to be unsuitable for the hydrolysis of starch, mainly because of the large differences in the pH and temperature optima of the applied enzymes. The best results are generally obtained with pure amyloglucosidase, starch conversion being about 99% after four hours incubation. Finally, dialysis is sometimes used to remove interfering agents from the sample solution prior to the starch hydrolysis. For example, starch is added as a thickening and gelling agent in sweetened fruit preparations and its determination is usually difficult because of the high sucrose concentrations in these products, since a significant quantity of glucose is released from sucrose during starch hydrolysis and this phenomenon hinders the determination of the glucose issued from starch. Cellulose acetate molecular porous membranes are then employed to dialyze the sample solution, thus eliminating the interfering sucrose molecules (Chatel et al., 1996).

6.3

Methods of analysing starch in food

Various physical, chemical, and biochemical methods (Bernetti et al., 1990) have been reported in the last decades for the analysis of total starch (see Table 6.1). An excellent review of them, though not updated, is the chapter by Morrison and Karkalas (Dey, 1990), who also discuss the problem of solubilization and lipid removal. More recently, Brunt and co-workers (Brunt et al., 1998) have tested and discussed different principles and protocols for each of the three most important steps involved with the starch determination, from the disintegration of the food sample and solubilization of the starch to the quantification of the glucose obtained. This work has been later abridged through an interlaboratory, collaborative study (Brunt, 2000).

ß 2004, Woodhead Publishing Limited

Table 6.1

Major methods of starch analysis

Type

Method

Classical

Iodine-iodide (qualitative) Gravimetry Redox titration

Modern

Near-infra-red reflectance (qualitative) Polarimetry High-performance liquid chromatography Gas-liquid chromatography Near-infra-red spectroscopy Enzymes Capillary electrophoresis UV-visible spectroscopy

Table 6.2 Examples of recommended methods for starch analysis according to the type of food analyzed Food

Method

Meat Cereals Cacao products Confectionery Fruits Baking powder Roasted coffee Peanut butter Food dressings Prepared mustard

Redox titration of starch hydrolyzate Colorimetry of starch hydrolyzate Redox titration of starch hydrolyzate Redox titration of starch hydrolyzate Iodine-iodide (qualitative) Redox titration of starch hydrolyzate Redox titration of starch hydrolyzate Gravimetry Redox titration of starch hydrolyzate Redox titration of starch hydrolyzate

In this section we present an overview of all these methods with special emphasis on the more recent ones and/or those commonly utilized in food analysis laboratories. A distinction is made between classical, traditional methods and modern (more recent), instrumental methods. The latter are, of course, of greater importance and will therefore receive more attention in this chapter. The choice of a given method will often depend on the matrix and, in this sense, Table 6.2 shows selected examples of recommended methods for starch analysis according to the type of food analyzed.

6.3.1 Classical methods Qualitative analysis The food sample is milled (if necessary) and well homogenized. A small amount is added to a suitable volume of boiling water and kept for five minutes. After

ß 2004, Woodhead Publishing Limited

that, the resulting mixture is cooled and an aliquot of the lower part (the upper layer is mainly fat) is mixed with some drops of iodine/potassium iodide reagent. A dark blue color appears in the presence of starch (Burriel et al., 2000). Details of the reaction involved in this procedure are given on pages 192± 4. Furthermore, this qualitative method of starch analysis is widely described in most organic chemistry books. Quantitative analysis Iodine can be added to the starch solution (resulting from the starch being extracted) to form an insoluble starch-iodine complex that can be determined gravimetrically by collecting, drying and weighing the precipitate formed or titrimetrically by determining the amount of iodine required to precipitate the starch (redox titration with sodium thiosulfate, also known as the Luff-Schoorl titration). Both are well known and have been widely used for many years. The reader is directed to the bibliography of Skoog et al., 1992 if further information is required. Another alternative consists of titrating the glucose obtained by hydrolysis of the starch. In this sense, two procedures can be followed: (i) the aforementioned Luff-Schoorl method; (ii) the modified Lane-Eynon's method (Horwitz, 2000), according to which a determined volume of boiling Soxhlet reagent (Fehling A + Fehling B + potassium ferrocyanide) is titrated with the sample solution, using methylene blue as indicator. Nevertheless, none of these methods is specific for glucose. The eventual presence of fructose or other reducing sugars would interfere, and it requires well experienced and skilled technicians. If there are no other components present in the solution that would interfere with the analysis, then the starch concentration could be determined using physical methods, e.g., density, refractive index, etc.

6.3.2 Modern methods Numerous books have been written on the fundamentals, apparatus, and applications of instrumental techniques in quantitative chemical analysis. Therefore, it is not the author's intention to describe them. Rather, the reader can refer to the bibliographic sources mentioned in the references if background information is required. In this case, two of them are especially recommended (Pare and BeÂlanger, 1997) (Meyers, 2000). In this section, the applications of the major instrumental methods for the analysis of starch in foods will be discussed. Qualitative analysis Although not as important as quantitative analysis, characterization of starch from different origins and its properties has also received attention in the last two decades. It can provide valuable information on possible adulterations, frauds, and/or unexpected changes in the starch characteristics. Nevertheless, it must be remarked that the presence of starch in a determined type of food is not necessarily an indication of its addition as adulterant. For instance, starch usually occurs in small amounts in apples and occasionally in other fruits, and

ß 2004, Woodhead Publishing Limited

unless it is found in a fruit product in considerable quantity its presence may be due to these natural sources. Qualitative analysis includes the use of near-infra-red reflectance analysis (NIRA) for evaluating the amount of cross linkage in chemically modified starch (Wetzel, 1983) and other techniques to investigate physicochemical properties, namely: particle size (low-angle laser light-scattering), granule shapes (light microscopy), and starch gelatinization (differential scanning calorimetry) (Farhat et al., 1999). For this last study, starch was isolated from yam tubers by obtaining a slurry with a sodium hydroxide solution. After filtering, the resulting suspension is centrifuged and the starch granule layer is finally cleaned and dried, prior to the examination by the above-mentioned techniques. In the case of sweet fruit preparations, starches are identified by optical microscopy and acetyl and hydroxypropyl modifications are characterized by Fourier transform infra-red spectroscopy (FT-IR) (Forrest, 1992; Chatel et al., 1997). It is a simple technique without long sample preparation, which permits the identification of acetylated and hydroxypropylated starches from the same spectrum scanning. Polarimetry The Ewers' polarimetric method (Ewers, 1908) or rather its modified version has widely been used for the determination of starch in cereal products (Mitchell, 1990). The procedure involves several stages. Starch standards are prepared by dissolving the appropriate amounts of starch in either diluted hydrochloric acid or trifluoroacetic acid. The mixture is then stirred in a bath of boiling water for precisely 15 minutes and thereafter cooled to room temperature. A certain volume of 4% sodium phosphotungstate is added and, finally, the solution obtained is shaken and filtered, discarding the first ml of filtrate. The same steps are to be followed to prepare the sample; care should be taken to grind the sample in a Wiley mill if it is not in powder form. Once both standards and samples are ready, they are measured by means of a saccharimeter. On the other hand, thermal gravimetric analysis is conducted on each standard and sample to determine the temperature at which the sample loses moisture. Then the percent moisture in each of the standard samples is calculated according to the following expression: % moisture ˆ …A ÿ B†=A  100 where A ˆ weight (grams) of undried sample and B ˆ weight (grams) of dried sample. The weight of starch, on a dry weight basis, is also calculated in each of the standard samples using the formula: Dry weight …in grams† of starch ˆ A ÿ AC where A ˆ weight (grams) of undried sample and C ˆ % moisture determined for each of the samples. A standard curve of saccharimeter readings versus the dry weight of starch in each standard sample is plotted. At the same time, the percent moisture in each

ß 2004, Woodhead Publishing Limited

of the unknowns is calculated as shown in the case of the standards. The dry weight `d' (grams) of starch in the unknown samples is then determined by reading off the standard curve and, finally, the percent starch on a dry weight basis is calculated for each of the unknown samples as follows: % starch ˆ d=…a ÿ ac† where a ˆ weight (grams) of undried sample, c ˆ % moisture in each unknown and d ˆ weight (grams) read off the graph. The saccharimeter reading may be plotted versus the percent starch, so that a factor is calculated from a slope of the line connecting the points, which can be used to multiply the saccharimeter reading to obtain percent starch directly. An alternative (in this method) to the acid hydrolysis of starch consists of dispersing it in hot calcium chloride solution; then the solution obtained is clarified and the starch content is calculated from the measurement of optical rotation as mentioned above. Nevertheless, and for ease of use, the calcium chloride procedure has advantages over its acid hydrolysis counterpart. In any case, and taking into account the fact that Dr Ewers first proposed the method in 1908, it is a compliment to him that nearly one hundred years later his method is still used in a manner only little changed from its conception. Visible spectroscopy (colorimetry) Early methods for estimating total starch used a basic `iodine-blue color' to compare different starches. The reaction between starch and iodine has been known for over a century, Iÿ 3 being considered as the iodine species responsible for the characteristic blue color formed consequently. Nevertheless, the reaction has become better understood in more recent times by studying the structure of the iodine complex using resonance Raman and iodine-129 MoÈssbauer spectroscopy; the results obtained showed that the major chromophore was the pentaiodide anion, Iÿ 5 (Teitelbaum et al, 1980). In practice, those methods reflected the amylose content, rather than total starch content, since only amylose forms helical inclusion complexes with iodine which display a strong blue color. By contrast, since amylopectin is highly branched, it binds only a small amount of iodine, thus producing a much paler purple-red color. Therefore, the estimation of total starch by the absorbance of its iodine complex is not wholly reliable because results are dependent on the amylose/amylopectin ratio which varies widely according to the type of food. There are many spectrophotometric (colorimetric) methods for determining the concentration of the amylose/iodine complex, but none of them give accurate results for amylose concentration directly because the amylose/amylopectin iodine complexes both absorb in the 620±680 nm region and therefore measurement of the iodine-blue color to estimate amylose inevitably includes amylopectin. The amount of amylose may be calculated by calibration with a standard amylose/amylopectin mixture, but any variation in the sample from the standard ratio may result in errors. There is another method that determines only the percentage of amylose in a starch (Morrison and Laignelet, 1983). To

ß 2004, Woodhead Publishing Limited

determine the amount of amylopectin, total starch must be measured by an alternative method and the amylose subtracted from it. Several years later, a two-wavelength spectrophotometric method was reported for determining amylose/amylopectin ratios in potato leaf and tuber tissue, although this approach does not quantify amylose or amylopectin directly, and the use of absorbances at only two wavelengths does not give as accurate results as does the procedure proposed by Jarvis and Walker (1993). They developed a spectrophotometric method for the simultaneous measurement of total starch, amylose, and amylopectin, in lipid-free food samples using a diode-array spectrophotometer. It involves, firstly, measurement of the spectrum of the amylose/amylopectin-iodine complexes at six wavelengths, namely: 1.

2. 3. 4. 5.

6.

504 nm: wavelength of greatest difference between the amylose and amylopectin spectra and where the absorbance due to amylopectin was greater than that due to amylose. 548 nm: amylopectin peak. 580 nm: peak for 20% amylose and 80% amylopectin. 630 nm: amylose peak. 700 nm: wavelength of greatest difference between the amylose and amylopectin spectra and where the absorbance due to amylose was greater than that due to amylopectin. 800 nm: wavelength of greatest absorbance due to amylose and where the absorbance due to amylopectin approaches zero.

Secondly, it involves calculation, after correction for the calculated iodine blank value at each of these wavelengths, of amylose and amylopectin concentrations utilizing multi-component analysis and a spreadsheet computer program. Lipids can affect the final absorbance of the iodine-blue complex and, therefore, starches from sources other than lipid-free foodstuffs (e.g. cereals) may require prior treatment to remove lipids (Morrison and Laignelet, 1983). The Morrison and Laignelet method includes the use of 85% methanol-water (v/v) or watersaturated butanol to extract these materials from the starch before carrying out the analysis. More recently, an improved, rapid method has been proposed for the determination of amylose in starch-containing foodstuffs (McGrance et al., 1998). It requires measurement at only one wavelength (600 nm) and avoids the use of harsh dispersants for the starch, dimethyl sulfoxide being utilized for this purpose, since this compound has been shown to be superior to strong alkalis in this regard. A linear relationship is obtained between absorbance and amylose concentration for mixtures of amylose and amylopectin standards, and this forms the basis of the determination. The method is rapid, simple, accurate and does not require the use of multi-component analysis of spectra, since a wavelength is chosen that suits the particular starch being analyzed. In the case of meat and meat products, simple sugars are extracted with 80% hot ethanol, and the starch residue is solubilized with diluted perchloric acid. The resulting solution is heated with a reagent composed of sulfuric acid and

ß 2004, Woodhead Publishing Limited

anthrone (9,10-dihydro-9-oxoanthracene) (Clegg, 1956). The greenish reaction product is monitored spectrophotometrically at 630 nm. The calibration curve is run with glucose standards and, after calculating the percentage of glucose (G) in the analyzed sample, the following formula is applied: % starch ˆ 1.06 G. Enzymatic methods In these procedures, specific enzymes are used not only to break down the starch to glucose, but also to catalyze the reaction of the latter with the subsequent spectrophotometric measurement. A rapid and quantitative method has been developed for the determination of total starch in a wide range of materials, including high-amylose maize starches and food materials containing resistant starch (McCleary et al., 1994, 1997). The method allows the analysis of 20 samples in three hours. For a range of samples, the total starch values obtained with this method are significantly higher than those provided by current standard methods. Two assay formats are proposed. According to one of them, the sample is treated with the chaotropic agent dimethyl sulfoxide to gelatinize the starch, which is then solubilized and partially depolymerized by controlled incubation at ~100 ëC with a defined level of thermostable -amylase in 3-(N-morpholino)propane sulfonic acid buffer (pH 7.0) at room temperature. This allows near-complete solubilization of most starches. The remaining starch is then solubilized and the starch fragments are converted to maltose and maltotriose by the combined action of highly purified pullulanase and -amylase. After volume adjustment and filtration (if necessary), the malto-oligosaccharides are hydrolyzed by high-purity amyloglucosidase to glucose, which is measured spectrophotometrically at 510 nm through the reaction with glucose oxidase-peroxidase/4-amino-antipyrine/p-hydroxybenzoic acid reagent (GOPOD) in phosphate buffer (pH 7.4). This assay format gives quantitative starch determination in all native starch samples, including highamylose maize starches. In the second alternative proposed, which is applicable to most starches and cereals flours, the dimethyl sulfoxide pretreatment step is omitted. Samples containing glucose and/or maltosaccharides are pre-washed with aqueous ethanol before carrying out the analysis. The Holm's method (Holm et al., , 1986) has widely been utilized because of its high accuracy and precision in the analysis of starch contents, especially those of cereal products. This procedure consists of four steps: (i) gelatinization and incubation with thermamyl, (ii) removal of protein, (iii) incubation with amyloglucosidase, (iv) spectrophotometric measurement of the glucose released by using glucose-oxidase/peroxidase. An important modification (Morales et al., 1997) was the step of dispersion and solubilization of retrograded amylose in 2M potassium hydroxide, after removal of protein. This change in the original method ensures a high recovery for amylose, as reported when wheat and potato starches as well as potato amylose and amylopectin standards were analyzed. The treatment with 2M KOH increases the solubilization of retrograded amylose, which makes this modification highly recommendable for the analysis of total starch in food, especially in those cases (cereals, legumes, vegetables) with a high retrograded amylose content (Tsuge et al., 1990).

ß 2004, Woodhead Publishing Limited

As regards resistant and digestible starch, their simultaneous determination in foods and food products has also been dealt with (Morales et al., 1997), this approach being mainly based on a previous work (GonÄi et al., 1996) that studied the analysis of resistant starch in detail. The procedure for the determination of the digestible starch fraction is then optimized. It consists of three steps: the removal of digestible starch, the hydrolysis of products, and the quantification of digestible starch as released glucose by using a glucose-oxidase/peroxidase kit for determination of glucose. Stomach and intestine physiological conditions (pH, transit time) are simulated approximately. The method was evaluated by the analysis of the starch content in wheat flour, spaghetti, lentils, beans, peas and potatoes. Some variables, such as employment of sodium azide, incubation time of -amylase and amyloglucosidase amount, which might affect the levels of digestible and resistant starch, were also evaluated and discussed. Test kits are also commercially available for the enzymatic determination of starch in foodstuffs such as bread and bakery goods, cereal products, dietetic food, flour, meat products (sausages, hamburger, meat balls), potatoes, and the determination of partially hydrolyzed starch (glucose-syrup, starch-sugar) in beverages and jam. The chemical process involved when the test kit is applied is as follows (Beutler, 1978): Starch + (n-1) H2O ! n D-glucose (in the presence of amyloglucosidase) D-glucose + ATP ! glucose-6-phosphate + ADP (by means of hexokinase) Glucose-6-phosphate + NADP+ ! D-gluconate-6-phosphate + NADPH + H+ (in the presence of glucose-6-phosphate-dehydrogenase) The absorbance of the final reaction product (NADH) is then measured spectrophotometrically at 340 nm. High Performance Liquid Chromatography (HPLC) The liquid chromatography characterization of starch has always posed a difficult analytical challenge. HPLC techniques with either metal-loaded cation exchange or amino-bonded silica columns coupled with refractive index (RI) detection have been widely used for the determination of sugars and small oligosaccharides, but in many cases they have proved inadequate for separating the complex mixtures of oligo- and polysaccharides resulting from the hydrolysis of starch. A major limitation is the incompatibility of RI with gradient elution, but equally important is the requirement for a column to resolve carbohydrate polymers of DP > 20 differing only by a single sugar unit. Chromatographic methods such as gel permeation chromatography (GPC) and high-performance size-exclusion chromatography (HPSEC) in conjunction with RI or light scattering detection are very useful for determining gross characteristics such as amylose/amylopectin ratios and average molecular weight (Kobayashi et al., 1985), but they are not capable of resolving individual polysaccharides and therefore provide only partial information on chain length distribution. Hydrophilic interaction chromatography coupled with pulsed amperometric detection on amino-bonded silica columns have been used to separate gluco-oligosaccharides derived from hydrolyzates of

ß 2004, Woodhead Publishing Limited

starch, up to a DP of ~30. This technique requires gradient elution with acetonitrile-water in the range 67:33 to 50:50 over 70 min and also post-column addition of sodium hydroxide to the eluent prior to the detector. Good resolution was achieved for DP 30 in the acetonitrile-water eluent is a limiting factor (Henshall, 1996). There was clearly a need for a better detection method for carbohydrates (and starch in particular) that is more specific and sensitive than RI and is compatible with gradient elution. This led to the development and use of high-performance anion-exchange chromatography (HPAE-PAD) (also referred to as high-pH anionexchange chromatography or HPAEC-PAD), which is proving to be a powerful tool for the analysis of starch and complex carbohydrates. The technique is direct (no derivatization required), highly sensitive and specific, and compatible with gradient elution procedures. On the other hand, resolution of polysaccharides up to DP >60 can also be achieved routinely in ~1 hour. Nowadays this technique is currently being applied in research and quality control laboratories in the characterization and determination of starch in many food products (Hauffe, 1997). Several interesting approaches have been reported in the 1990s concerning the determination of starch in foodstuffs by liquid chromatography. A reversedphase HPLC method using a 5 m Spherisorb NH2 column, a refractive index detector, and 85:15 v/v acetonitrile/water as mobile phase was developed and applied to green beans and other starchy vegetables (LoÂpez-HernaÂndez et al., 1994). Starch is extracted from the samples and hydrolyzed with amyloglucosidase solution. After the elution process, the starch derived glucose concentration, C, is calculated by comparison with the chromatogram peak of a glucose standard run prior to the sample: C ˆ Cs A=As where Cs is the concentration of the glucose standard, A is the area of glucose peak from the sample, and As is the area of glucose peak from the standard. Starch concentration as a weight percentage of the original sample, is determined by Starch (%) ˆ 0.90  100 C (mg/mL)/W (mg) ˆ 90 C/W where W is the mass of sample and the factor 0.90 corrected for incorporation of a molecule of water in each glucose unit during hydrolysis (Horwitz, 2000). Another interesting proposal is a separation scheme for the determination of starch in processed food (plain cereals, sugar-coated cereals, canned fruits, canned vegetables, crackers, cookies, and so on). It is based on AOAC Method 985.29, some modifications having been added (Casterline et al., 1999). Samples are milled to a fine powder and, if necessary, defatted with petroleum ether. pH 6.0 phosphate buffer is added and samples are stored at 4 ëC overnight to ensure hydration of the matrix. After centrifugation, the two fractions obtained (soluble and insoluble materials) are separated and the following procedure is applied to each one of them: they are treated with heat-stable amylase, and amyloglucosidase to hydrolyze starch. Acetonitrile is then added to precipitate substances such as soluble proteins and fibers, which create backflow and thus interfere with the chromatographic resolution of peaks; residues are

ß 2004, Woodhead Publishing Limited

removed by centrifugation and filtration through a nylon filter, and the filtrate is passed through a solid-phase extraction cartridge to remove HPLC-interfering substances, the eluate being analyzed for starch (soluble or insoluble, depending on the fraction analyzed) by liquid chromatography. Starch content is calculated from the increase in the amount of glucose. Gas-liquid chromatography A general scheme was developed (Li, 1996) to determine starch in half-gram freeze-dried samples of selected high-consumption foods such as bread, rice, spaghetti, potatoes, and beans. Samples are extracted for free sugars with 80% methanol, and the residues are incubated with a solution containing amyloglucosidase in acetate buffer. Starch hydrolyzates are then centrifuged and aliquots are removed for glucose determination by gas-liquid chromatography (GLC), after the corresponding derivatization to trimethylsilylated oximes or ethers. Samples are thereafter ready for GLC analysis or may be kept in a refrigerator for several weeks. As for the chromatographic determination, a 25±m capillary column containing cross-linked methyl silicone is used. The usual operating conditions are as follows: injection volume, 1 or 2 L; oven temperature programmed from 160 ëC (with 1 min hold) to 270 ëC at 6 ëC minÿ1; injection port temperature, 250 ëC; detector temperature, 310 ëC; carrier gas (helium) constant flow pressure, 20 psi; split ratio, 70:1. Starch is then calculated according to the following formula: Starch content ˆ glucose…g=100 g†  0:9 GLC has also been used for the determination of resistant starch (the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuals) in cooked dried beans. As an alternative to the abovementioned procedure, the pretreatment before enzyme hydrolysis consists of autoclaving the samples at 121 ëC in the presence of dimethyl sulfoxide (Li and Zhao, 1997). This treatment leads to extensive if not complete solubilization of the starches in the analyzed samples. Near-infra-red spectroscopy Near-infra-red spectroscopy analysis is an instrumental method for rapidly and reproducibly measuring the chemical compositions of samples with little or no sample preparations (Schwedt, 1997). Since the near-IR spectroscopic analysis offers four principal advantages: speed, simplicity of sample preparation, multiplicity of analyses from a single spectrum, and intrinsic nonconsumption of the sample, it has been widely used to measure constituents of many agricultural commodities and food products as well as on-line analysis at food processing sites. Especially, it makes the analyses of starch in foodstuffs an easy, rapid, and nondestructive routine analysis. This is very important, for instance, when the viscosity of gravy in gravy-containing food products is to be evaluated (viscosity is a major factor affecting the final quality of the products). The viscosity of gravy depends on the type and content of the starch used in preparation, the

ß 2004, Woodhead Publishing Limited

former being directly proportional to the amount of starch and to its chemical structure (the more branched the starch prepared for the gravy is, the higher viscosity the gravy has). Therefore, direct determination of the starch content in gravy is important for further monitoring the quality of canned products. In this sense, near-IR spectroscopy has proved to be very useful when compared to labor-intensive, time-consuming, and destructive conventional wet chemistry methods (Zeng et al., 1996). The procedure is based on the use of a near-IR spectrum composition analyzer equipped with a built-in computer that integrates with the instrument and can independently carry out the analysis according to the calibration constants. Modified food starch, pure corn starch, and beef flavoring may act as gravy models. This method can be used at a varied range of temperatures, and permits the on-line analysis of starch content during gravycontaining food processing. Capillary electrophoresis Capillary electrophoresis (CE) has been playing an increasingly important role in liquid-phase chemical analysis. Several features including speed of analysis, high resolution, and efficiency account for the present acceleration in the acceptance of this technique. High electric fields are used in CE to force ionic solutes to migrate through a buffer- or gel-filled capillary. The species, injected in minute amounts, are separated along the tube on the basis of charge, size, or both and are subsequently detected near the capillary end (GuÈnzler and Williams, 2001). The various forms of this technique have been successfully demonstrated in the analysis of carbohydrates, starch being no exception. In the case of starch, the absence of charge does not offer any problem by resorting to the use of iodine complexation to impart charge and permit detection of starch components in capillary electrophoresis (Brewster and Fishman, 1995). Amylopectin and amylose in potato starch are resolved in less than 10 min using iodine-containing buffers (over the pH range 4.0±7.5) in unmodified silica capillaries. The primary basis for separation is shown to be iodine binding affinity, which can be manipulated through control of temperature and iodine concentration. Iodine concentrations between 0.1 and 0.5 mg mlÿ1 have yielded the best results. At lower iodine concentrations amylopectin is not resolved, whereas Joule heating due to the high conductivity of KI solutions prevents separations in 50 m capillaries at higher concentrations. Furthermore, in contrast to other approaches for CE of carbohydrates, lengthy (4±24 h) derivatization steps are avoided, and separations can be carried out using commonly available detectors.

6.4 Determining starch in food: recent technological developments The use of biosensors in food analysis and, in particular, in the determination of starch, is based on enzymatic methods and avoids sample pretreatment, apart

ß 2004, Woodhead Publishing Limited

from saving time and money in many cases. State-of-the-art biosensors utilize membranes which have the enzyme (usually oxidases) covalently immobilized. These membranes (10 mm diameter) are set in the measuring device by means of an o-ring. The sample is led into a chamber where the oxygen dissolved is transformed into hydrogen peroxide, which is then electrochemically oxidized (using a platinum electrode) into oxygen. The electrical current produced between the platinum anode and the silver cathode is amperometrically measured and is proportional to the analyte concentration. Calibration is carried out with the corresponding standards that are commercially available (Bucsis, 1998). In the case of starch analysis, after autoclaving the sample it is incubated with amyloglucosidase and the resulting glucose is analyzed using a suitable biosensor as described above. If free glucose occurs in the sample, a previous determination of glucose should take place, and thereafter the starch content is calculated. The whole procedure has been automated to a greater or lesser extent by some specialized firms. Biosensors are straightforward with important advantages, namely: (i) No sample pretreatment is normally required, which even allows for suspensions to be directly measured. This is due to the fact that amperometry, instead of light absorption, is the measurement technique (ii) high sensitivity and specificity, typical of enzymatic analysis (iii) high flexibility and analysis frequency (ca. 40 samples per hour). Flow Injection Analysis (FIA) is a technique which has gained increasing popularity in laboratories all around the world (Trojanowicz, 2000). Many of the time-consuming batch procedures can be replaced by fully automated flow injection systems which can dramatically decrease the time and cost for analysis. In this sense, the use of immobilized enzyme reactors combines the selectivity of enzyme reactions with the speed and simplicity of FIA (Karkalas, 1991). Starch hydrolyzing enzymes are commonly used to convert this polysaccharide into glucose, which may then be determined by different flow injection procedures, such as the one proposed by Peris-Tortajada and co-workers, whose scheme is shown in Fig. 6.2. In this case, the sample (glucose) solution is injected into a CuII-neocuproine (2,9-dimethyl-1,10-phenanthroline) stream and later merged with a basic (NaOH) stream, the reaction product being monitored at 460 nm. Moreover, since the reduction of glucose requires drastic conditions for sufficiently fast development, the reactor in which the chemical reaction takes place is immersed in a water thermostatic bath at a suitably high temperature (70 ëC). Starch has also been determined using a FIA system composed of an immobilized glucoamylase reactor followed by pulsed amperometric detection of the glucose produced. Amyloglucosidase, immobilized onto porous silica and packed into a short stainless steel column, is capable of nearly quantitative (98%) conversion of the starch to glucose. The sensitivity of pulsed amperometric detection for soluble starch is increased 26-fold by first passing the starch through the immobilized glucoamylase reactor. The proposed method turned out to be simple, rapid, and sensitive for starch (Larew et al., 1988). Additionally, the half-life of immobilized glucoamylase was estimated to be

ß 2004, Woodhead Publishing Limited

Fig. 6.2 Flow injection configuration for the analysis of hydrolysis products (sugars) of starch. (P = peristaltic pump; I.V. = injection valve; T.B. = water thermostatic bath; R = reactor; D = UV-V spectrophotometric detector; W = waste).

over 500 days, which means that the reactor can be used during reasonably long periods of time. Another interesting FIA approach also based on the use of immobilized enzymes has been developed for the determination of starch from different origins (wheat, corn, rice). However, in this case the glucose obtained by hydrolyzing starch reacts with NAD+ (nicotinamide adenine dinucleotide) conveniently immobilized, and the resulting NADH is monitored spectrophotometrically (Emneus et al., 1986, 1993). The authors of this approach immobilize a heat stable -amylase on controlled pore glass and amyloglucosidase on a ceramic silica support. These enzyme supports are packed into two separate immobilized enzyme reactors which together with a third reactor containing co-immobilized glucose dehydrogenase/mutarotase are incorporated into a flow injection system. -Amylase and amyloglucosidase convert (hydrolyze) the starch into glucose which, by entering the glucose dehydrogenase/mutarotase reactor, is oxidized by glucose dehydrogenase to D-gluconate in the presence of NAD+ which at the same time is reduced to NADH (measured at 340 nm); this product is then proportional to the amount of glucose produced from the hydrolyzed starch samples. In this case, mutarotase plays an important role, since glucose dehydrogenase is only active on the anomeric form of glucose, and therefore the produced -D-glucose also formed in the enzymatic hydrolysis of starch (and through spontaneous mutarotation) needs to be converted to the -form by mutarotase. An alternative to the UV spectrophotometric measurement is the electrochemical oxidation of the hydrogen peroxide formed after the complete oxidation of the produced glucose in a third reactor containing coimmobilized mutarotase and glucose oxidase (instead of glucose dehydrogenase) (Emneus and Gorton, 1990). A fully automated method for the determination of starch in foodstuff has been developed (Velasco-Arjona and Luque de Castro, 1996). The method is based on the standard procedure involving leaching of sugars from the sample,

ß 2004, Woodhead Publishing Limited

hydrolysis of the analyte, and colorimetric determination of the hydrolysis products (sugars) by the neocuproine method (Peris-Tortajada et al., 1992). A focused microwave digester is used as peripheral for the robotic station in order to expedite analyses by accelerating the hydrolysis step. The station is connected to a flow injection manifold, which minimizes the dilution and derivatization times. In this way, duplicate analyses of three solid samples take 5 h rather than 15 h required when the whole analytical sequence is performed by the robotic station and 5 h per sample taken by the manual (conventionally applied) method.

6.5

Future trends

Some work is currently in progress concerning sonic spray ionization (SSI), a liquid chromatography-mass spectrometry interfacing technique in which a liquid flow is sprayed from the tip of a capillary under atmospheric pressure with a gas flow coaxial to the capillary. In contrast to the electrospray or atmospheric pressure chemical ionization techniques, SSI forms charged droplets without heating the capillary or applying an electric field to the tip of the sprayer. Therefore, it is ideally suited for the analysis of thermally unstable compounds such as carbohydrates. Furthermore, ionization without additional high voltages increases the flexibility in the choice of mobile phases. SSI is proving to be very useful in the determination of small saccharides, oligo- and polysaccharides (among them starch), a very high sensitivity being achieved (Volmer, 2000). Additional work will probably be conducted in the future taking into account the promising results obtained so far in this field. The use of biosensors is expected to consolidate its leading position in starch analysis in foods. Enzyme electrodes should allow the continuous assay of this important component, not only for the purpose of classical at-line analysis but also with a view to on-line and in-line monitoring of food production processes. The latter is already a well-established analytical procedure and the growing availability of commercial biosensors will open a promising path. Furthermore, the increasing stability of many enzymes that can be achieved by crosslinking and immobilization techniques will improve the determination of starch and derivatives. Finally, Expert Systems (ES) are currently being applied to a greater or lesser extent in various fields of analytical chemistry, including food analysis (Peris, 2002). The typical scheme consists of a centralized system, implemented around a personal computer, which not only has the necessary sensors and actuators to connect itself to the process to be controlled but also runs the knowledge-based system. Especially remarkable is the outstanding role ESs play in the monitoring and control of food analysis processes. Presumably, the determination of starch will soon take advantage of this technology, which can be of great utility in terms of a higher automation of the analysis process. For instance, ESs can easily control a flow injection system for the analysis of the glucose resulting from the hydrolysis of starch, as shown in the on-line determination of reducing

ß 2004, Woodhead Publishing Limited

sugars in the course of an alcoholic fermentation process using a rule-based system (Peris et al., 1997). Moreover, in the case of starch-containing foods, taking representative samples, handling them in such a way that no significant changes occur, and preparing the samples for starch analysis are of the utmost importance, since insufficient attention to these factors may very easily lead to distorted data and erroneous conclusions. Sample preparation is often the most critical and time-consuming step in the entire analytical process. For instance, when determining the concentration of starch in peanut butter, the main difficulty is not the measurement procedure, but achieving an aqueous solution containing all (or a known fraction) of the starch in the sample. Therefore, a good sample preparation method leads to the removal of (potential) interferences and to a concentration of the compounds of interest. Thus, there is nothing strange in the fact that several attempts have recently been made to develop expert systems in this area, especially relevant in food analysis. Nevertheless, we believe that such efforts must be strongly encouraged, since the cornerstone of expert system applications in food analysis is still the analysis itself, preanalysis steps being somewhat left aside. On the other hand, the use of computer networks offers the possibility of interchanging information and even permits the remote access from computers located very far from the places where the analysis is being performed. These and other facts have led to the development of the so-called Distributed Expert Systems (DES) (Bonastre et al., 2000). DESs yield significant advances over the already existing expert systems in the field of Analytical Chemistry, namely, the distribution itself, with the aforementioned features, the remote access to the data through the Internet, response times that are low enough to permit the implementation of real-time systems, and ease of programming and depuration of its functioning for analytical chemists without a solid background in computer science (Bonastre et al., 2001). In any case, ESs/DESs offer the possibility of revolutionizing the practice of food analysis in general, and starch in particular. There are, however, two major problems which must still be overcome, namely, the time required to build the knowledge base and the amount of computation needed to solve a given problem, since the most useful applications tend to be the most complex. Anyway, the possible benefits of these systems in food analysis are so great that these problems will probably be solved, making expert systems more cost effective. The future of ES applications in food industries will probably bring exciting new approaches. The time will come when the use of expert systems will be an integral part of the practice of food analysis in quality control laboratories. Expert systems will not replace food chemists, they will rather be very useful assistants which handle the details and allow the chemist to concentrate on the more challenging problems. In the current decade, we will see that the increasing computational power of individual instruments and their expanding capability to communicate with each other point to a unified, networked laboratory. In this laboratory of the near future, expert systems will be in charge

ß 2004, Woodhead Publishing Limited

of making decisions on the development, implementation, and testing of toplevel analysis processes as well as of managing automated systems in the food quality control laboratory. Additionally, remote process monitoring will be guaranteed by local area networks and/or the Internet, real-time data of different analyses being easily accessible from several locations. In this sense, we predict that the use of centralized knowledge-based systems ± though still not consolidated ± will probably be replaced by distributed expert systems, due to their above-mentioned significant advantages. However, although ESs unarguably help to fill the gap when a human expert is not available, and under pressure conditions may even outperform the operator, in many cases the optimal solution will still be provided by a combination of artificial and natural intelligence. In any event, some time will undoubtedly elapse until expert systems have the typical `common sense' of human beings. For now, and unlike us, they are not able to react properly to many unpredictable situations. Time will tell whether they can achieve this goal.

6.6

Sources of further information and advice

A detailed monograph The Analytical Chemistry of Carbohydrates (Scherz and Bonn, 1998) presents the whole field of qualitative and quantitative methods for the analysis of nearly all kinds of carbohydrates, including polysaccharides such as starch. Although it mainly focuses on the determination procedures of monoand oligosaccharides, its importance lies in the fact that many methods for the analysis of starch involve the prior hydrolysis to glucose and determination of this monosaccharide. This book covers theoretical aspects as well as practical applications, with special emphasis on instrumental techniques such as chromatography, photometry, and electrophoresis. Another extensive ± though not so updated ± compilation, Determination of Food Carbohydrates (Southgate, 1991) also deals with major methods and techniques utilized for the analysis of sugars and starch in food. It can be considered as an excellent review in this field and the reader is directed to this bibliographic source, especially if he/she is a primer in the determination of food carbohydrates. Methods in Agricultural Chemical Analysis: A Practical Handbook (Faithfull, 2003) is a recent handbook that reports ± among other things ± the determination of starch in starch-containing foods, with a special mention of sampling procedures and sample pretreatment (acid digestion, acid hydrolysis, extraction procedures). Methods are described in a step-by-step approach and contain many practical details. Calculations are also carefully explained. The book is designed as a laboratory sourcebook with useful Internet addresses. Considered for some food chemists as the `bible' of food analysis, Handbook of Food Analysis (Nollet, 1996; second edition due in 2004) devotes an entire chapter to the analysis of carbohydrates and starch. Special attention is paid to the latter given its important presence in many foodstuffs. It covers all the stages

ß 2004, Woodhead Publishing Limited

in its analytical determination from sampling to the analysis itself, be it by classical methods or by modern instrumental techniques. Practical applications are also given along with the corresponding bibliographic references. When the reader has this book in his/her hands, Starch (Whistler, 2004) will already be available in its third edition. Written by one of the leading authorities in the field of polysaccharides, Roy Whistler, it deals with everything related to starch, including methods of analysis. It provides the analytical chemists with an excellent view of the complex world of starch and will help them with a great amount of information on the chemistry of this compound. Undoubtedly, a valuable tool in the chemical analysis of starch, most notably in foods. For over 40 years the journal Starch/StaÈrke (StaÈrke is the German word for starch) has focused on the most important carbohydrates ± from a renewable resources point of view ± such as cellulose, starch, and sugars produced by photosynthesis. Comprehensive and topical, it publishes original articles dealing with fundamental and applied studies. Particular attention is given to recent studies on new analytical methods for starch, modified starches, starch derivatives, and starch saccharification products. Book reviews, an extensive documentation service, patent reviews and previews of symposia complete the package. This journal is published monthly in English by John Wiley & Sons, Inc. Last but not least, since its inception in 1965 the journal Carbohydrate Research has gained a reputation for its high standard and wide scope. Articles published include all aspects of starch analytical chemistry and biochemistry. Normal length research papers, perspectives, notes, rapid communications and book reviews, together with notices of relevant meetings can be found in this journal, which is published in English by Elsevier Science.

6.7

References

(2000), Approved methods of the AACC 10th edn, The Association, St. Paul MN. BERNETTI R, KOCHAN D, TROST V, and YOUNG S (1990), `Modern methods of analysis of food starches', Cereal Foods World, 35, 1100±1105. BEUTLER H O (1978), `Enzymatische Bestimmung von StaÈrke in Lebensmitteln mit Hilfe der Hexokinase-Methode', Starch/StaÈrke, 30, 309±312. BONASTRE A, ORS R, and PERIS M (2000), `Monitoring of a wort fermentation process by means of a distributed expert system', Chemom Intell Lab Syst, 50, 235±242. BONASTRE A, ORS R, and PERIS M (2001), `Distributed expert systems as a new tool in analytical chemistry', Trends in Anal Chem, 20, 263±271. BREWSTER J D and FISHMAN M L (1995), `Capillary electrophoresis of plant starches as the iodine complexes', J Chrom A, 693, 382±387. BRUNT K (2000), `Collaborative study concerning the enzymatic determination of starch in food, feed, and raw materials of the starch industry', Starch/StaÈrke, 52 (2±3), 73±75. BRUNT K, SANDERS P, and ROZEMA T (1998), `The enzymatic determination of starch in food, feed, and raw materials of the starch industry', Starch/StaÈrke, 50 (10), 413±419. AMERICAN ASSOCIATION OF CEREAL CHEMISTS

ß 2004, Woodhead Publishing Limited

(1998), `Biosensoren in der Lebensmittelanalytik', LaborPraxis, 22 (5), 44±47. and HERNAÂNDEZ J (2000), QuõÂmica AnalõÂtica Cualitativa, Thomson Publishing, Madrid. CASTERLINE J L, OLES C J, and KU Y (1999), `Measurement of sugars and starches in foods by a modification of the AOAC total dietary fiber method', J AOAC Int, 82 (3), 759±765. CHATEL S, VOIRIN A, LUCIANI A, and ARTAUD J (1996), `Starch identification and determination in sweetened fruit preparations', J Agric Food Chem, 44, 425±430. CHATEL S, VOIRIN A, and ARTAUD J (1997), `Starch identification and determination in sweetened fruit preparations. 2. Optimization of dialysis and gelatinization steps, infrared identification of starch chemical modifications', J Agric Food Chem, 45, 425±430. CLEGG K M (1956), `The Application of the Anthrone Reagent to the Estimation of Starch in Cereals' J Sci Food Agric, 7, 40±44. DEY P M (ED) (1990), Methods in Plant Biochemistry 2, Carbohydrates, Academic Press, London. EMNEUS J, APPELQVIST R, MARKO-VARGA G, GORTON L, and JOHANSSON G (1986), `Determination of starch in a flow injection system using immobilized enzymes and a modified electrode', Anal Chim Acta, 180, 3±8. EMNEUS J, and GORTON L (1990), `An analytical flow system for starch determination based on consecutive enzyme steps and amperometric detection at a chemically modified electrode', Anal Chem, 62 (8), 263±268. EMNEUS J, NILSON G, and GORTON L (1993), `A flow injection system for the determination of starch in starch from different origins with immobilized -amylase and amyloglucosidase reactors', Starch/StaÈrke, 45 (8), 264±270. È ber die Bestimmung des StaÈrkegehaltes auf polarimetrischem Wege', EWERS E (1908), `U Z offent Chemie, 14, 150±157. FAITHFULL N (1990), `Acid hydrolysis prior to automatic analysis for starch', J Sci Food Agric, 50, 419±421. FAITHFULL N T (2003), Methods in Agricultural Chemical Analysis: A Practical Handbook, CABI Publishing, Cambridge MA. FARHAT, I A, OGUNTONA T, and NEALE R J (1999), `Characterisation of starches from West African yams', J Sci Food Agric, 79, 2105±2112. FORREST B (1992), `Identification and quantification of hydroxypropylation of starch by FTIR', Starch/StaÈrke, 5, 179±183. Ä I I, GARCIÂA-DIZ, L, MAN Ä AS E, and SAURA-CALIXTO, F (1996), `Analysis of resistant GON starch: a method for foods and food products', Food Chem, 56 (4), 445±449. È NZLER H, and WILLIAMS A (2001), Handbook of analytical techniques, Wiley-VCH, GU Weinheim. HAUFFE D (1997), `Analytik von StaÈrke und komplexen Novel Food Kohlenhydraten durch HPLC', GIT Fachz Lab, 41 (5), 460±467. HENSHALL A (1996), `Analysis of starch and other complex carbohydrates by liquid chromatography', Cereal Foods World, 41 (5), 419±424. È RCK, I, DREWS A, and ASP N G (1986), `A rapid method for the analysis of HOLM J, BJO starch', Starch/StaÈrke, 38, 224±226. HORWITZ W (ed) (2000), Official Methods of Analysis of AOAC International, AOAC, Arlington VA. JARVIS C E and WALKER J R L (1993), `Simultaneous, rapid, spectrophotometric determination of total starch, amylose and amylopectin', J Sci Food Agric, 63, 53±57. BUCSIS L

BURRIEL F, LUCENA F, ARRIBAS S,

ß 2004, Woodhead Publishing Limited

(1991), `Automated enzymatic determination of starch by flow injection analysis', J Cereal Sci, 14, 279±286. KOBAYASHI S, SCHWARTZ S J, and LINEBACK D R (1985), `Rapid analysis of starch, amylose, and amylopectin by high performance size exclusion chromatography', J Chromatogr, 319, 205±212. LAREW L A, MEAD D A, and JOHNSON D C (1988), `Flow-injection determination of starch and total carbohydrate with an immobilized glucoamylase reactor and pulsed amperometric detection', Anal Chim Acta, 204, 43±51. LI B W (1996), `Determination of sugars, starches, and total dietary fiber in selected highconsumption foods', J AOAC Int, 79 (3), 718±723. LI B W and ZHAO Z (1997), `Determination of starches and dietary fiber polysaccharides in cooked dried beans: Comparison of different temperatures and dimethyl sulfoxide treatments', J Agric Food Chem, 45, 2598±2601. Â PEZ-HERNAÂNDEZ J, GONZA Â LEZ-CASTRO M J, VA Â ZQUEZ-BLANCO M E, VA Â ZQUEZ-ODEÂRIZ M L, LO and SIMAL-LOZANO J (1994), `HPLC determination of sugars and starch in green beans', J Food Sci, 59 (5), 1048±1049. MCCLEARY B V, SOLAH V, and GIBSON T S (1994), `Quantitative measurement of total starch in cereal flours and products', J Cereal Sci, 20 (1), 51±58. MCCLEARY B V, GIBSON T S, and MUGFORD D C (1997), `Measurement of total starch in cereal products by amyloglucosidase--amylase method: Collaborative study', J AOAC Int, 80 (3), 571±579. MCGRANCE S J, CORNELL H J, and RIX C J (1998), `A simple and rapid colorimetric method for the determination of amylose in starch products', Starch/StaÈrke, 50 (4), 158± 163. MEYERS R A (ED) (2000), Encyclopedia of Analytical Chemistry, Wiley, New York. MITCHELL G A (1990), `Methods of starch analysis', Starch/Sta È rke, 42 (4), 131±134. Â LEZ M C (1997), `Simultaneous determination of MORALES M D, ESCARPA A, and GONZA resistant and digestible starch in foods and food products', Starch/StaÈrke, 49 (11), 448±453. MORRISON W R, and LAIGNELET B (1983), `An improved colorimetric procedure for determining apparent and total amylose in cereal and other starches', J Cereal Sci, 1, 9±20. NOLLET L M (ed) (1996), Handbook of Food Analysis 1st edn (Chapter 15), Marcel Dekker, New York. PAREÂ J R, and BEÂLANGER J M (eds) (1997), Instrumental methods in food analysis, Elsevier, Amsterdam. PERIS M (2002), `Present and future of expert systems in food analysis', Anal Chim Acta, 454, 1±11. PERIS M, ORS R, BONASTRE A, GIL P, and SERRANO J (1997), `Advanced application of rule nets to the automation of chemical analysis systems', Anal Chim Acta, 354, 249±253. PERIS-TORTAJADA M, PUCHADES R, and MAQUIEIRA A (1992), `Determination of reducing sugars through the neocuproine method by flow injection analysis', Food Chem, 43, 65±69. ROSE R, ROSE C, OMI S, FORRY K, DURALL D, and BIG W (1991), `Starch determination by perchloric acid vs enzymes: evaluating the accuracy of six colorimetric methods', J Agric Food Chem, 39, 2±11. SCHERZ H and BONN G (1998), The Analytical Chemistry of Carbohydrates, Thieme, Stuttgart. SCHWEDT G (1997), The essential guide to Analytical Chemistry, Wiley, New York. KARKALAS J

ß 2004, Woodhead Publishing Limited

Measuring starch in food SKOOG D A, WEST D M,

207

and HOLLER F J (1992), Fundamentals of Analytical Chemistry 6th edn, Saunders College Publishing, Orlando FL. SOUTHGATE T (1991), Determination of Food Carbohydrates 2nd edn, Elsevier, London. TEITELBAUM R C, RUBY S L, and MARKS T J (1980), `A resonance Raman/iodine Mo È ssbauer investigation of the starch-iodine structures. Aqueous solution and iodine vapor preparations', J Am Chem Soc, 102, 3322±3328. TROJANOWICZ M (2000), Flow Injection Analysis. Instrumentation and Applications, World Scientific Publishing, Singapore. TSUGE H, HISHIDA M, IWASAKI H, WATANABE S, and GOSHIMA G (1990), `Enzymatic evaluation for the degree of starch retrogradation in foods and foodstuffs', Starch/ StaÈrke, 6, 213±216. VELASCO-ARJONA A and LUQUE DE CASTRO M D (1996), `A robotic-flow injection approach to the fully automated determination of starch in food', Anal Chim Acta, 333, 205± 213. VOLMER D A (2000), `Analysing thermally unstable compounds by LC-SSI-MS', LC-GC Europe, November 2000, 838±846. WETZEL D L (1983), `Near-infrared reflectance analysis. Sleeper among spectroscopic techniques', Anal Chem, 55 (12), 1165 A±1176 A. WHISTLER R L (2004), Starch 3rd edn, Academic Press, San Diego CA. ZENG W, ZHANG H Z, and LEE T C (1996), `Direct determination of the starch content in gravy by near-infrared spectroscopy', J Agric Food Chem, 44, 1460±1463.

ß 2004, Woodhead Publishing Limited

Part II Sources of starch

ß 2004, Woodhead Publishing Limited

7 The functionality of wheat starch H. Cornell, RMIT University, Australia

7.1 Introduction: manufacture of wheat starch for the food industry Starch is the most abundant carbohydrate in cereal grains, being found in the endosperm. Wheat starch, like other starches, plays a major role in the food industry, where its thickening and gel-forming properties are paramount. Starches from various plants have a major role as an energy source, providing 70±80% of the calories consumed by humans worldwide, this being greatly in excess of all other food hydrocolloids combined (Be Miller and Whistler, 1996). The calorific value of starch is 1550 kJ/100 g (370 kcal/100 g). Starch, being a mixture of polysaccharides, differs from sugars in having less crystallinity, lower solubility in water and a bland taste, unlike the sweet taste of sugars. Starch is composed of the polymers amylose and amylopectin, both based on anhydroglucose units and its chains can thus be represented as (C6H10O5)n, where n is variable depending on the distribution of the different polymers present. Starch is the predominant food reserve material in plants as glycogen is in animals. Apart from wheat starch, discussed in this chapter, starch is available from corn (including high amylose and waxy corn) and rice, as well as from tubers and roots such as potato and cassava (tapioca). The manufacture of wheat starch, its properties and functionality in the food industry are the key elements of this chapter, while a very important aspect is its versatility brought about by chemical modifications. The manufacture of starch from wheat flour necessitates the separation of the starch in granular form from the gluten complex (a mixture of many proteins) and many other materials which are present in smaller amounts. The process is carried out in aqueous environment so that the starch granules and gluten complex remain behind, while water-soluble proteins, carbohydrates, minerals,

ß 2004, Woodhead Publishing Limited

vitamins, etc., are carried through with the washing procedures. Basically, the process involves making a dough or batter from the flour, separating the gluten complex and refining the starch suspension by centrifugal and water-washing techniques. The process is usually called wet-milling of wheat flour.

7.1.1 The Martin Process The Martin Process is the earlier of the two main processes of the above type. A dough is made from a white flour and about 85% by weight of water. Usually, a small amount of sodium chloride is also added to improve the separation of the gluten. After doughing, the gluten is recovered as curds after the addition and mixing with water. This can be achieved using a screw conveyor and a series of shaker sieves on which washing of the gluten can be carried out. The starch `milk' from these sieves can be processed continuously by centrifugal methods which include concentrators and washers such as hydrocyclones (Svarovsky, 1984) before being recovered as a starch `cake' on basket-type centrifuges (known generally as `hydros'). Finally, the starch is dried to produce a white powder of about 12% moisture. Some of this water is bound chemically forming a spherocrystal (Zobel, 1988). The yield of prime grade starch is normally about 60% of the weight of flour and about 15% of lower grade starch, generally not suitable for food use. Because of centrifugal techniques being central to the processing, the prime grade starch has a higher percentage of large granule starch and, as a consequence, lower protein and lipid content. The low grade starch is associated with higher levels of pentosans, compounds which are mainly arabinoxylans originating from the cell walls of the lignified bran layer (Lineback and Rasper, 1988). For further details, refer to Cornell and Hoveling (1998).

7.1.2 The batter process In this process, a batter, rather than a dough, is prepared as a first step. The flour is mixed with about 1.25 times its weight of water, so that the consistency is lower than in the case of a dough and, furthermore, higher temperatures are used (30±35 ëC) in order to hasten hydration of the gluten and reduce mixing times and power consumption. Otherwise, the process is similar to the Martin Process. Other processes in which starch splitter decanters, such as those from AlfaLaval, Tumba, Sweden are employed, together with decanter centrifuges, allow reductions in the use of water (Cornell and Hoveling, 1998). In wet-milling, it is important to remove not only gluten, but soluble proteins and proteins associated with the starch. Eliasson and Tjerneld (1990) studied the adsorption of wheat proteins on wheat starch under different conditions and showed that pH and ionic strength were governing factors. The protein adsorbed increased as the pH was increased from 3.1 to 7.6. Seguchi (1986) showed that there was a protein-rich layer on wheat starch granules using the microscope and protein-specific dye binding techniques.

ß 2004, Woodhead Publishing Limited

Furthermore, it was also shown that lipids were extracted by sodium dodecyl sulfate containing 2-mercapto ethanol, but not peptides (Seguchi, 1995). Some of the proteins on the surface of wheat starch granules are referred to as `friabilins' and have molecular weights of about 15 000 (Greenblatt et al., 1995).

7.1.3 Production of high-quality wheat starch The wet-milling of wheat flour produces starch of good quality but there is some variation in quality depending upon how much of the lower grade starch (`tailing starch') is incorporated into the final product. Two separate streams of quality can usually be arranged by concentration of the effluents from separation of the prime starch. These two streams can be differentiated on the basis of their particle size distribution (see Sect. 7.2) and their content of nitrogenous material, chiefly proteins and complex lipids. Starch of the highest quality will have nitrogen contents of about 0.035% (determined by Kjeldahl or Dumas method) and this is ideal for food use and for the manufacture of starch syrups (see Sect. 7.5). Much of this nitrogen is due to lysophosphatidyl choline, the major lipid in wheat starch (Morrison, 1978) but in the lower grades of starch, protein and pentosan contamination is significant and its pasting properties may not be of high enough standard to warrant its use in foodstuffs. Cornell (1970) showed that both proteins and lipids could be extracted by treatment of wheat starch with NaOH at pH 11.7 at 40 ëC for 96 hours. This resulted in a product of only 0.014% N, well below what is normally regarded as being suitable for individuals with coeliac disease, where gluten is the causative agent (Anderson et al., 1952). The extracted materials could include friabilins, lipoproteins, gluten proteins and complex lipids. Friabilins seem to control wheat endosperm softness and are partly removed from starch during wet-milling of wheat flour. Further extraction is possible after removal of the polar lipids. The reason for the improvement of soft wheat flours for cakemaking by chlorination may be due to its effects on these surface proteins.

7.2

Granular and molecular structure of wheat starch

7.2.1 Granular structure Starch is seen as discrete granules within the endosperm of the wheat kernel, being embedded in a protein matrix. Wheat endosperm contains mainly two types of granules a larger type, mostly about 20±35 micrometers (m) across (Astarch), being lenticular in shape, and a smaller spherical type, ranging from 2± 8 m in diameter (B-starch). Starches from different botanical sources have different size distributions, this being useful for identification. Scanning electron microscopy of the larger granules shows concentric shells, exposed after treatment with enzymes (French, 1984). The layers are due to the fact that starch grows by apposition ± the layers depending upon the amount of carbohydrate synthesised in each growth period.

ß 2004, Woodhead Publishing Limited

Starch is usually referred to as a semicrystalline material and a `fringed micelle' model has been proposed which is comprised of microcrystalline regions bound to amorphous regions of flexible chain segments (Slade and Levine, 1987). That it has a significant degree of crystallinity is supported by polarised light microscopy and by X-ray diffraction (de Man, 1990) and the birefringence, which shows in the form of a Maltese cross under polarised light, suggests a significant degree of molecular order (Hoseney, 1994). However, the arrangement of the two major components of the starch amylose and amylopectin within the starch granules is not fully understood because the amylose is intermingled through the amylopectin structure (Lineback and Rasper, 1988). It is, however, generally accepted that the amylopectin is present in a tiered cluster structure (Robin et al., 1974), the tiers being perpendicular to the growth rings. Particle size distribution Several methods have been used to determine the particle size distribution of wheat starch. Early methods relied on the use of a sedimentation balance in which the amount of starch settling on a balance pan over a given period of time was weighed and the appropriate calculations made. In the case of wheat starch, the larger granules settle more rapidly than the smaller ones and the resultant graph is useful for assessing the quality of the starch, this being directly related to the percentage of large granules (Cornell and Hoveling, 1998). Laser beam diffraction methods are now widely used and have the advantages of being rapid and able to use small amounts of sample. Low angle laser light scattering (LALLS) utilises a gas laser which passes through a stirred suspension of the test material and requires measurement of intensity of the focused rays by a photosensitive detector. Figure 7.1 shows a typical profile for commercial wheat starch. The distribution curve shows a major peak at about

Fig. 7.1 Particle size distribution of commercial wheat starch (Bunge Bioproducts Ltd) in water, as determined by the use of the Malvern Mastersizer X. The line shows the cumulative results. (Reprinted from Cornell HJ et al., Starch/Staerke, Vol. 46. pp. 203±7, 1994, with kind permission of Wiley VCH, Weinheim, Germany).

ß 2004, Woodhead Publishing Limited

22 m in aqueous suspension. Under the same conditions, the peak for potato starch was about 40 m. Particle size distribution can also be determined by the well-established gas absorption methods and the newer acoustic spectroscopic methods (Cornell and Hoveling, 1998).

7.2.2 Molecular structures of the major wheat starch components Wheat starch, like the other grain and tuber starches, consists chiefly of amylopectin with lower amounts of amylose, averaging about 25% (range 20± 30%). Exceptions are the starches from mutant varieties of grains, which can either contain high amounts of amylose, or starches which contain virtually all amylopectin, called `waxy starches' (Whistler and Daniel, 1984). Amylose is a molecule formed by -1,4 linkages between glucose units and is thus a glucan or, more generally, a glycan. The structure of amylose is shown by Haworth representation in Fig. 7.2. It is essentially a linear polymer, but the chains can form helical structures as shown in Fig. 7.3, where the anhydro-glucose units are depicted in the chair form and six such units are present in one turn of the helix. Amylose can form inclusion complexes with iodine in aqueous potassium iodine, due to the pentaiodide ion (I5-), a fact made use of in the determination of amylose in starch (McGrane et al., 1998). In contrast to amylose, amylopectin is a highly branched molecule, caused by -1,6 glycosidic bonds, in addition to having -1,4 glycosodic bonds. Fig. 7.2 also shows a Haworth representation of amylopectin, in which can be seen the

Fig. 7.2

Haworth representations of amylose and amylopectin.

ß 2004, Woodhead Publishing Limited

Fig. 7.3 Helical conformation of amylose, a left-handed helix containing six anhydroglucose units per turn. Reprinted from Cornell HJ and Hoveling AW WheatChemistry and Utilization p 139, 1998, with kind permission of CRC Press, Boca Raton, FL, USA.

potential for branching. Three types of chains are present A-chains, consisting of -1,4-linked glucose units, B-chains, consisting of -1,4- and -1,6-linked glucose units and C-chains, those with both types of linkages together with a reducing group. The ratio of the A-chains to B-chains is usually about 1:1 to 1.5:1 (Whistler and Daniel, 1984). These chains have a degree of helical structure (Hoseney, 1994). Amylose and amylopectin from different starches will differ in their molecular structures and various chromatographic techniques have been employed to determine these differences and make use of them in selecting the best starches for the food in mind. Takeda et al. (1984) examined amylose from different starches using gel-permeation chromatography on Toyopearl

ß 2004, Woodhead Publishing Limited

HW-75F (Toyo Soda Co Ltd, Tokyo, Japan) columns and were able to show differences in molecular size and numbers of chains per molecule. Molecular weight determinations It is important to compare the molecular weight of amylose with that of amylopectin. However, this is not an easy task as the values depend not only on the botanical source of the starch from which they were derived, but also on the method used. One of the most commonly used techniques is based on viscosity studies in dilute solution, making use of the Mark-Houwink-Sakurada equation (Kitamura et al., 1989) which relates limiting viscosity number to molecular weight. ‰Š ˆ KM

7:1

where ‰Š ˆ limiting viscosity number, K ˆ a constant, M ˆ molecular weight (weight average), and  ˆ an exponent which depends upon the shape of the polymer in the particular solvent used. The molecular weight of the sample is able to be calculated by solving simultaneous equations for K and  using two standards of known molecular weight. Experiments are commonly carried out using an Ubberlohde capillary flow viscometer (a modification of the Ostwald viscometer). Values of  for amylose in solution in dimethyl sulfoxide (DMSO) are usually in the range 0.7± 0.8 indicating it is present as some type of coiled molecule. The molecular weight of amylose (weight average) determined in DMSO varies from about 150 000 to 400 000 (900±2500 anhydroglucose units) but this depends upon the botanical source of the starch. Wheat amylose, prepared by 1butanol precipitation in presence of potassium thiocyanate (Cornell et al., 1999) appears to be similar in size to corn amylose (molecular weight approximately 250 000 in DMSO). Wheat amylopectin, prepared by the same method, appears to have a weight average molecular weight of about 10 million, compared to 12 million for corn amylopectin and 14 million for waxy corn amylopectin (Cornell et al., 2002). These estimates were obtained using a mixture of DMSO and 2 mol/LKSCN in a ratio of 1:3 (v/v) which is required to limit intermolecular hydrogen bonding between ±OH groups on amylopectin and between amylopectin and water molecules. Physicochemical and other properties Apart from molecular weight, the stability of aqueous solutions, the gel properties and ability to form films are important physicochemical properties. Differences in the colours of the respective iodine complexes and the digestibility with the enzyme -amylase are also evident (Zobel, 1988) reflecting the contrasting structures of amylose and amylopectin. The differences in functionality between amylose and amylopectin are also the consequence of the differences in molecular structures and the properties of solutions, pastes and gels are important considerations for the food industry.

ß 2004, Woodhead Publishing Limited

Table 7.1

Summary of the physicochemical properties of amylose and amylopectin

Property

Amylose

Amylopectin

Molecular structure Molecular weight Dilute aqueous solutions* Gels Films Iodine colour Digestibility by -amylase

Essentially linear 150,000±400,000 Unstable Firm, irreversible Coherent Blue 100%

Highly branched 10±15 million Stable Soft, reversible Not readily formed Reddish-purple Approx. 60%

*Amylose is, however, soluble in 0.5 mol/L KOH, DMSO and formamide.

Some of the physicochemical properties of each component are summarised in Table 7.1. Separation of major components The separation of amylose from amylopectin without degradation of either polymer has been a difficult undertaking. The basis of several methods has been the selective precipitation of amylose by formation of the inclusion complex with 1-butanol (Schoch, 1942). The complex precipitates from the hot sol on cooling and after centrifugation, amylose is obtained by heating the complex under vacuum. Although the purity of the amylose obtained is governed to some extent by the conditions of precipitation, a high degree of association between amylose and amylopectin prevents clean separations. However, it has been shown that if precipitation of the complex is carried out in the presence of KSCN, good quality amylose can be obtained from wheat starch (Cornell et al., 1999). Amylopectin remains in solution and can be recovered by precipitation with excess ethanol. At the present time, high amylose and high amylopectin (waxy) starches are manufactured from mutant varieties of grain and serve the same purpose.

7.3

Functionality of wheat starch: granules, films and pastes

7.3.1 Granules Before discussing the major uses of starch, which are connected with its pasting properties, it is necessary to point out that granular starch is used in several ways. Wheat starch is used in the confectionery industry as a moulding powder for the various shapes of sweets and can be re-used many times. In the pharmaceutical industry it is used as an excipient, a type of binding agent for the active medicament. Starch, by virtue of its amylose content, is able to form inclusion complexes with many food ingredients such as essential oils, flavourings and fatty acids, e.g., omega-3 fatty acids. It thus acts as an encapsulating agent, increasing the shelf life of the product. The ability of these small molecules to form inclusion

ß 2004, Woodhead Publishing Limited

complexes depends on their size and shape. Scatchard plots can be used to give information about the various binding parameters of the system.

7.3.2 Films Films of amylose, being biodegradable, are of commercial value as edible coatings. Amylose films form light crisp batters on food meant for frying and prevent large take-up of oil, thus giving health benefits. Amylose forms clear films which adhere well to most food formulations and high amylose starch is also useful in this respect. Amylopectin, by contrast, does not form useful films.

7.3.3 Pastes The physicochemical properties of amylose and amylopectin are quite diverse and it is understandable that these two major components of starch contribute in different ways to the pasting properties of starch. This is its most important functional property. When starch is heated with water the insoluble granules are disrupted by the energy supplied, resulting in a loss of the molecular organisation responsible for the crystallinity and the penetration of water. Swelling of the granules results in an increase in viscosity and a complete loss of crystallinity, as adjudged by the loss of birefringence that occurs when the heating is continued. The temperature at which birefringence is first lost is called the gelatinisation temperature. For wheat, the temperature range over which this occurs is 53±64 ëC, a little lower than for most other starches (Be Miller and Whistler 1996). The increase in viscosity continues to a maximum when most granules burst and form a network of interlocking chains. Upon cooling, the amylose chains `zipper' together and, aided by the amylopectin, form a gel, provided the concentration is high enough. This process is referred to as `gelation'. The starch gel can be considered a composite, in which swollen starch granules reinforce an intimate amylose gel matrix. Changes brought about by heating are detectable by microscopy and differential scanning calorimetry (DSC). Light microscopy and DSC were used by Ghiasi et al. (1982) to study the changes in wheat starch suspensions. They found that at a water:starch ratio of 2:1, birefringence was lost over 57±64 ëC, but at a lower water:starch ratio of 0.5:1, the temperature range extended to 87 ëC. Thermograms obtained in their DSC studies are shown in Fig. 7.4. The pasting properties of a particular sample of starch are governed by the concentration, rate of heating and the presence of other food components. In general, the variety of the starch and the climatic and soil conditions in which the plant is grown also govern these properties. There can be some enzymatic activity in the grain before harvesting due to periods of rain, causing a reduction in viscosity of the paste. Similarly, if the wet-milling of the flour is not carried out efficiently so that minimal bacterial action is allowed, the pastes will have lower viscosity than normal. A simple test for paste viscosity is the `falling

ß 2004, Woodhead Publishing Limited

Fig. 7.4 Differential scanning calorimeter thermograms of heated wheat starch suspensions with water-to-starch ratios of 2.0 (a), 1.0 (b), 0.75 (c), 0.50 (d), 0.44 (e) and 0.35 (f). Sensitivity was 1 mcal and chart speed 10 mm/min. (Reprinted from Ghiasi K et al., Cereal Chem, Vol. 59, pp. 258±62, 1982, with kind permission of American Association of Cereal Chemists, St Paul, MN, USA).

Fig. 7.5 Cooking curves for wheat, potato and corn (maize) starches (8% w/w) in the Brabender Amylograph. (Reprinted from Cornell HJ and Hoveling AW WheatChemistry and Utilization p. 149, 1998, with kind permission of CRC Press, Boca Raton, FL, USA).

ß 2004, Woodhead Publishing Limited

number', obtained by taking the time for a metallic sphere to fall through a starch paste prepared under standard conditions. More advanced testing for pasting properties requires the use of an instrument such as the Brabender Amylograph (Viscograph) which heats the stirred starch suspension at a selected rate, holds the paste near boiling and then cools it to a selected temperature, all with constant stirring and measurement of viscosity. Fig. 7.5 shows typical cooking curves for starches from wheat, potato and corn. The parameters provided by the Amylograph are: · · · · · ·

Pasting temperature Peak viscosity (in Brabender Units) Holding strength (reading at the end of the 95ëC heating period) Breakdown (peak viscosity ± holding strength) Final viscosity (reading at the end of the cooling period) Setback (final viscosity ± holding strength)

Another instrument, designed to test the pasting characteristics is the Rapid Visco Analyser (Newport Scientific, Sydney, Australia) which uses smaller amounts of sample.

7.4

Rheological properties of starch pastes and gels

The rheological properties of starch pastes have been studied extensively. The rheology of biopolymers is now an important area of research in the food industry because it has implications in the formulation of foods and starch is one of the most important of these biopolymers. Many such materials exhibit a time-dependent behaviour unlike a Hookean material, which when placed under a constant stress, deforms immediately to a constant strain and instantly and completely recovers when the stress is removed (like a spring). They are also unlike Newtonian materials (e.g. mineral oils) in that they deform immediately under stress, but do not recover immediately when that stress is removed. Elastic shear moduli are features of starch pastes and gels, the values being governed by the molecular weight distribution of the polymers present (Nielsen, 1977). Shear thinning is the type of behavour encountered with all starch pastes, but when these pastes are cooled, a sharp increase in the elastic modulus, G, occurs (see equation 7.4) due mainly to the gel-forming properties of amylose. The combination of viscous and elastic behaviour is termed `viscoelasticity' (Barnes et al., 1989). Studies of this type require sophisticated apparatus, preferably that equipped with parallel plates. One such instrument is the Rheometrics Fluids Spectrometer RFS II (Rheometrics, New Jersey, USA). It has titanium parallel plates of radius 25.0 mm, and a gap size of about 1.1 mm may be used.

ß 2004, Woodhead Publishing Limited

7.4.1 Mathematical relationships and models A useful method for describing linear viscoelastic behaviour, i.e., where strain is directly proportional to the stress, is by the use of mechanical models. These models consist of springs (ideal elastic or Hookean deformation) and dashpots (ideal viscous or Newtonian flow) which can be combined in different configurations such as in series, in parallel or combinations thereof (Blocksma, 1972). With starch pastes, viscosity decreases with increased shear rate (`shear thinning') and is measured at different shear rates using Newton's Law:  ˆ = 0 where 0 ˆ
=
t

7:2

and  ˆ viscosity  ˆ shear stress

0 ˆ shear rate The SI units for viscosity are Pascal seconds (Pa.s). A useful parameter which is a measure of the material's ability to resist deformation is the modulus, G, the proportionality constant in the relationship, which applies for shear  ˆ G

7:3

With dynamic mechanical measurements, an oscillatory strain is applied to the sample, allowing the viscous and elastic properties of the sample to be measured simultaneously and gives information on network strength and relaxation times, very important in bread dough rheology. In these situations,  ˆ G

7:4

where G is the complex modulus, which includes a viscous component, G00 , and an elastic component G0 . The SI units for moduli are Pascals. Parallel plate rheometers utilise the angle, , (in radians) such that

0 ˆ K 0

7:5


=
 ˆ K
=
t

7:6

or

where  is the actuator angular velocity (in radians/second), K ˆ R/H (the strain constant), R is the plate radius (mm), and H is the gap width (mm).

7.4.2 Physicochemical aspects of pastes and gels The thixotropic behaviour of starch pastes is probably due to the progressive orientation of molecules in the direction of flow and the rupture of H-bonds formed in the amylose-amylopectin-water structure. When the shear force is removed, the viscosity finally recovers to the value before the application of this

ß 2004, Woodhead Publishing Limited

force. Doublier et al. (1987) found that wheat and corn starch pastes displayed less thixotropy than oat starch pastes of the same concentration but were weaker in structure. Wong and Lelievre (1982) studied the rheology of wheat starch pastes under steady shear conditions and found that the volume occupied by the gel particles and the size distribution of the particles were important factors governing apparent viscosity and yield stress. Most starches form elastic gels when cooked above 6% (w/w) concentration in water and there is an approximately linear relationship between the rigidity modulus and concentration. The gelation process for starch has been described as occurring by way of two possible mechanisms: 1. 2.

A thermodynamic process of phase separation followed by development of crystallinity. Non-covalent interactions, e.g., hydrogen bonding that initiate gel formation followed by aggregation (Case et al., 1998).

Miles et al. (1984, 1985) found that the development of G0 in amylose gels was concentration dependent, attaining a saturation value which was also concentration dependent. The development of opacity and gel structure seemed to be related to phase separation. Eliasson and Bohlin (1982) studied the rheological and thermal properties of concentrated wheat starch pastes from 30±80% (w/w) heated to different temperatures. At 30% (w/w) water content a true gel was not formed as the DSC thermogram did not show a peak. At 50% and 60% (w/w) water content, a maximum value of the relaxation modulus was obtained when the temperature interval for gelatinisation was passed and the DSC thermograms showed a gelatinisation and a second endotherm which was interpreted as due to a process which makes the granules softer. At 80% (w/w) water content, only the gelatinisation endotherm was observed.

7.4.3 Properties of mixtures of amylose and amylopectin Doublier and Llamas (1993) examined the viscoelastic properties of aqueous mixtures of amylose and amylopectin and found that these properties were governed by a phase separation process. Below 15% amylose, the continuous phase contained predominantly amylopectin and beyond 22% amylose, the amylose was concentrated sufficiently in the continuous phase to form a strong gel. The phases are the result of differences in affinity for the water and reflect some incompatibility of amylose and amylopectin in this solvent.

7.4.4 Properties of amylose compared with amylopectin In order to better understand the functionality of wheat starch, it is important to look at the functionality of its two major components, which are both available commercially. Their functional properties differ considerably. Dilute solutions

ß 2004, Woodhead Publishing Limited

of amylose (1%) in DMSO or 0.5 mol/L aqueous KOH are fairly stable but in water it requires autoclaving (130±150 ëC) before it forms solutions (Case et al., 1998). Even so, the latter solutions are unstable because the amylose molecules associate and precipitate. At higher concentrations (2%), the ability to form gels is governed by the solvent used to disperse the amylose. In DMSO and DMSO-water mixtures containing mainly DMSO, amylose does not form gels, but in water and waterrich mixtures, gels form on standing (Cornell et al., 2002). Amylopectin does not form gels in water or in DMSO, so it is clear that the gel-forming ability of starch is due to the amylose component. This is because of amylose forming extensive hydrogen bonding (H-bonding) networks with water molecules, as well as intramolecular H-bonding. Amylopectin pastes are softer due to fewer intermolecular H-bonds and at 2% concentration (w/w) in water, free-flowing solutions are obtained. This is because of its highly branched, dendritic-type molecular structure. Amylopectin is more readily soluble in water than amylose but gelation occurs only at concentrations greater than 10% (w/w) where amylopectin chains become heavily entangled, especially after storage at 1 ëC (Ring et al., 1987). Turbidity reached a limiting value after about five days with the elastic modulus (G0 ) only showing a small increase. Thus, under these conditions, some tendency to crystallisation is suggested, but it is largely reversible, unlike with amylose. Studies by Bello-Perez and Paredes-Lopez (1994) showed that amylopectin contributed more significantly to the viscous modulus (G00 ) in starch and concluded that amylose must contribute more significantly to the elastic modulus. Parovuori et al. (1996) examined the effect of amylopectins of different molecular weight and found that the elastic modulus was increased by amounts of high molecular weight (native) amylopectin less than 50% (w/w) of the total amylose plus amylopectin, but decreased as the percentage of amylopectin exceeded 50%. Low molecular weight amylopectins did not act in this way to reinforce the gel. Several workers have used solutions of amylose in DMSO/water to study gelation (Hayashi et al., 1983; Takeyama et al., 1993; Cheetham and Tao, 1997, 1998; McGrane, 2001) showing the importance of water content in gelation and measuring turbidity and other parameters related to gelation. McGrane (2001) found that at approximately 30% (w/w) water content and above, the apparent viscosities of 5% (w/w) and higher corn amylose concentration increased significantly and the solution exhibited pronounced shear thinning behaviour, suggesting a large increase in intermolecular reactions at these higher proportions of water. Contrary to these observations, water content had little effect on the apparent viscosity of corn amylopectin. McGrane (2001) proposed that in DMSO, amylose and the outer linear chains of amylopectin have a rigid helical conformation which is maintained by intramolecular hydrogen bonding that is stabilised by DMSO molecules. As water is added, the intramolecular hydrogen bonds are replaced with intermolecular bonds to water, causing the amylose molecules to extend until they adopt an extended interrupted helical conformation. This is supported by

ß 2004, Woodhead Publishing Limited

studies of formation of the amylose-iodine inclusion complex which only forms in water/DMSO mixtures greater than 53% water. Rheological studies indicated that the minimum water content required for the onset of gelation was about 15% (w/w) water. It should be understood that whilst it is useful to look at the pasting properties of amylose and amylopectin separately, the properties of the starches are not directly related to each component. Jane and Chen (1992) have found synergistic effects on paste viscosities, particularly when long chain amylopectins were present.

7.5 Improving and chemically modifying wheat starch for use in the food industry 7.5.1 Improving the functionality of wheat starch Before looking at the ways in which we may improve the functionality of wheat starch by chemical modification, it is worthwhile contemplating the necessary requirements for the particular foods and trying to determine if they can be met by the native starch, corresponding amylose-rich, amylopectin-rich starches or mixtures. Wheat starch cannot be considered the ideal starch for food applications, particularly when the food requires a smooth, stable paste or gel of high clarity. However admixture of wheat starch with small amounts of amylopectin improves the smoothness of the paste or gel and its stability (Cornell, 1963). Gels of amylose become very firm on standing due to the formation of crystallites, whereas amylopectin gels are softer, more stable and less crystalline because of extensive branching. In each case the processes are reversible, although amylose gels require autoclaving to achieve this reversal. The stability of wheat starch products, particularly the gels, is improved by addition of amylopectin because the latter prevents association of amylose molecules which follows the release of H-bonded water molecules, as shown in Fig. 7.6. This process is termed `retrogradation', which means the return from a solvated, dispersed state to an insoluble, aggregated state. The change is accompanied by an increase in cloudiness as well as formation of free water, the latter being termed `syneresis'. Amylopectin undergoes retrogradation more slowly and to a much lesser extent than amylose, because of its highly branched structure. Unlike the linear amylose, amylopectin molecules are not able to align and Hbond so readily; this only happens over limited regions (Thomas and Atwell,

Fig. 7.6

Syneresis of starch gel, exemplified by release of water from hydrogen bonded amylose gel.

ß 2004, Woodhead Publishing Limited

1999). The contribution of amylopectin to the retrogradation of wheat starch is more noticeable at higher concentrations, such as in the staling of bread. Effect of molecular differences in starch components The properties required for certain food applications will be governed by the molecular structures of the amylose and amylopectin, not only for pasting properties, but other properties, such as those required for Japanese noodle manufacture. Shibanuma et al. (1994) studied the molecular structures of amyloses and amylopectins from different varieties of wheat. They found that wheats containing starches with higher amounts of branched amylose gave better noodle quality. Pregelatinised starch A native starch can be gelatinised prior to use, making it possible to be used in products that do not require cooking. This type of starch, called `pregelatinised starch' is very useful for instant desserts and baby foods. Instant desserts are made with pregelatinised starch, fine sugar, tetrasodium pyrophosphate (coagulant) and calcium acetate (accelerator) plus colouring and flavouring. When mixed with cold milk (140 g/l of milk) and left to stand for 15 minutes a firm textured custard is formed. Pregelatinised starch also has the advantage of reducing the cooking time of gravies, soups and sauces. These products require only the addition of hot water with good stirring before serving. Pregelatinised starch is manufactured by feeding an aqueous starch slurry onto steam-heated rollers. The gelled starch is removed by a scraper, crushed and sieved.

7.5.2 Chemical modification of starch Chemical modification of wheat starch is aimed at improving its functional properties for use in modern food formulations. The properties generally chosen for improvement are: · · · ·

paste stability at high temperatures, particularly at low pH values paste clarity, particularly at low temperatures texture; `mouth feel' and smoothness of consistency, reduced syneresis freeze-thaw stability.

These properties are best improved by chemical modification as all native starches are lacking, to different extents, in most of these respects. Most chemical modifications depend upon the reactions of ±OH groups of the amylose and amylopectin. Amylose contains three ±OH groups per anhydroglucose unit, but there are less than this in amylopectin because of -1,60 bonds. Hence the degree of substitution (DS) of these ±OH groups is an important criterion. Chemical modification is of great benefit to wheat starch because wheat starch gels tend to be `rubbery' in texture and are somewhat opaque. They also have poor freezethaw stability.

ß 2004, Woodhead Publishing Limited

Thin boiling starch This is an example of where the chemical modification depends upon the hydrolysis of -1,40 and -1,60 glycosidic bonds, and not upon reactions of the ±OH groups. Thin boiling starch is normally manufactured by treating a starch slurry (approximately 40% solids) below its gelatinisation temperature, with hydrochloric acid or sulfuric acid (0.1±0.2 mol/l) at 30±45 ëC. The time required depends upon the end use of the product and viscosity tests are carried out during processing in order to determine when the batch is within the specifications. At this point, which may take several hours, the mixture is neutralised with soda ash and the product is separated, washed and dried as for normal starch processing. Thin boiling starches have much lower pasting viscosity than unmodified starch, a slower rate of gel formation on cooling and higher clarity gels. They are widely used in the manufacture of soft eating confectionery such as fruit gums and children's sweets. Dextrins Dextrins are also products of acid degradation of starch and find use as encapsulating agents and in emulsions. Encapsulating agents improve flavour retention in foods and after the cooking process (kettle or jet cooking at 130 ëC) the products are then spray dried. In emulsions, dextrins improve stability and mouthfeel of the products. Starch ethers Starch ethers are typified by hydroxyethyl starch, prepared by the action of ethylene oxide, a substance with an oxirane ring which makes it very reactive with ±OH groups of the starch as shown in Fig. 7.7. Products with DS 0.05±0.10 have been manufactured by reacting a starch suspension with ethylene oxide (a gas, BP 10.7 ëC) at 50 ëC. This DS results in a product which has a lower gelatinisation temperature and, even more importantly, a large improvement in paste clarity. Products of this type may also be made using ethylene

Fig. 7.7

Reaction of starch with ethylene oxide (top) and ethylene chlorohydrin (bottom) to produce hydroxyethylated starch.

ß 2004, Woodhead Publishing Limited

chlorohydrin (Fig. 7.7). Being a liquid, it is easier to use but care is required in use because of its toxicity. Hydroxyethyl starches are extremely useful for fruit pie fillings because high clarity leads to greater eye appeal of such products and retrogradation of the product made from wheat starch is much less severe than that of the parent starch (Cornell and Hoveling, 1998). Hydroxypropyl starches are likewise made by the use of propylene oxide (a liquid, BP 35 ëC) and the products have similar uses to those of the hydroxyethyl starches, salad dressings being such an additional use. Kim et al. (1993) have studied the rheological properties of hydroxypropyl potato starch (molar substitution 0.125) cooked under different conditions and have suggested that the changes in parameters such as the complex modulus (G*) are due to structural changes, signalled by coarse aggregation. Another starch ether is carboxymethyl starch, requiring the reaction of the starch with chloroacetic acid. As with ethylene chlorohydrin, the reactive chlorine atom is involved. Carboxymethyl starch is used as a stabiliser in foods such as ice cream. Starch esters The most commonly used starch esters are the phosphates. These products are of two main types: 1. 2.

Monostarch phosphates. Distarch phosphates.

They are manufactured using various phosphorus compounds, e.g., phosphorus oxychloride, sodium orthophosphates, sodium tripolyphosphate (STPP) or sodium trimetaphosphate (STMP) (Kerr and Cleveland, 1962). The two reactions with phosphorus oxychloride are shown in Fig. 7.8. With lower ratios of the oxychloride to starch, distarch phosphate is formed as seen in the same figure. The HCI formed can be neutralised, as formed, by the constant addition of NaOH or Na2CO3 solution. The POCl3 also reacts to some extent with water. A typical process using STMP involves reaction of a wheat starch suspension (45% solids) at pH10 with STMP (2±3% on solids) for one hour followed by heating the starch cake at 130 ëC for 2±3 hours. This procedure brings about further reaction, and unreacted salts can be washed out of the cake before drying. The reactions involved are shown in Fig. 7.9. At lower pH values, monostarch phosphates are the main products. Likewise, STPP (5% on solids) forms mainly distarch phosphates, whilst at pH7, starch pyrophosphate is the major product formed (Lim and Seib, 1993). The phosphorus content of the modified starch may be determined on a sample which has been well washed with water (to remove unreacted phosphate) using a colorimetric method based on the formation of a complex with molybdate and metavanadate (Whattham and Cornell, 1991). Starch phosphates are usually prepared with DS 0.1±0.4 and are characterised by having higher paste viscosities, especially if a high degree of cross linking is achieved. Hence wheat starch after modification resembles potato starch

ß 2004, Woodhead Publishing Limited

Fig. 7.8

Reaction of starch with phosphorus oxychloride showing the formation of a starch phosphate and a distarch phosphate.

because of increased viscosity and higher hot-paste clarity, the latter probably being due to effects of repulsion between negatively charged phosphate groups. Starch phosphates also display good resistance to breakdown in the presence of acids as well as good freeze-thaw stability. Hence they find widespread use in acidic products such as tomato soup and fruit pie fillings. Distarch phosphates have similar uses, and are even more stable to heat and agitation, particularly at low pH values. They develop higher viscosity at high temperatures, which is an

Fig. 7.9 Reaction of starch with sodium trimetaphosphate (STMP) showing the formation of a monostarch phosphate and distarch phosphate.

ß 2004, Woodhead Publishing Limited

added advantage in keeping large food particles in suspension at these temperatures. Starch acetates Starch acetates are readily prepared by reacting the starch with acetic anhydride in the presence of sodium acetate as catalyst. Again, products may be prepared by reaction of a starch slurry at 25 ëC and pH9±11 (Fig. 7.10). Starch acetates have higher paste clarity than the parent starches and are readily made to about 0.5 DS. However, high DS can lead to lower gel strength. Because of high paste clarity, products of DS 0.1±0.2 are very acceptable for fruit pie fillings. Takahashi et al. (1989) prepared mixed acetate-phosphate starches by the use of acetic anhydride and phosphorus oxychloride. Corn starch and wheat starch products were thus prepared, and although paste viscosities of the modified wheat starch were of lower viscosity and the gels of lower elastic modulus than those from corn starch, the freeze-thaw stability was markedly improved at DS 0.1 acetylation and 0.1 molar phosphate substitution. Starch adipates and succinates Starch adipate is another modified starch which has higher viscosity due to cross-linking of the starch. Reaction can be brought about by use of the reactive diacyl chloride and again, mixed products such as the acetate-adipate can be very useful as the problems encountered with acetates in regard to diminished paste viscosity can be offset by the increase due to cross links formed because of the difunctional adipyl dichloride (see Fig. 7.10). It has uses in condensed soups because of improved storage characteristics. Other applications are based on high paste clarity, a major feature of all these ester and ether modifications. Succinic anhydride offers the advantageof high reactivity with starch, like acetic anhydride. However, being bifunctional, it can cause cross linking and

Fig. 7.10 Reaction of starch with acetic anhydride to form starch acetate (top) and structure of a mixed acetate-adipate starch.

ß 2004, Woodhead Publishing Limited

Fig. 7.11 Reaction of starch with succinic anhydride to form a distarch succinate (top) ± the normal product ± and some monostarch succinate, an acidic by-product.

therefore effect an increase in paste viscosity. Starch succinate is easier to prepare than starch phosphate, but does not quite match the stability to hot acidic environment of the latter. Typical structures of starch succinates are shown in Fig. 7.11. Betancur-Ancona et al. (2002) have studied the properties of jack bean (Canavalia ensiformis) starch after succinylation to varying DS. They found that succinylation to DS 0.026 increased viscosity and solubility, resulting in pastes and gels with better thickening capacity, clarity and stability at low temperatures and pH.

7.6

Wheat starch syrups

Although these products are not as widely used as corn syrups, wheat syrups serve the same purposes and represent a large percentage of the output of wheat starch manufacturers, particularly in Australia. Starch syrups are sweet edible products, widely used in the food industry, particularly in the confectionery, jam and beverage industries. They are manufactured from the starch by a process of hydrolysis, carried out by acids, enzymes or a combination of both. Besides syrups, solid glucose (dextrose) can be obtained by crystallisation from completely hydrolysed starch liquors after the reaction enzymes ; heat …C6 H10 O5 †n ‡ nH2 O ÿÿÿÿÿÿÿÿÿÿÿÿÿ!nC6 H12 O6 (refer to Section 7.6.2). Starch syrups are products not hydrolysed to the same extent and the syrups are normally distributed in the range of 70±85% solids content. Their calorific value (1715 kJ/100 g, 410 kcal/100 g) (based on solids content) is a little higher than that of starch. There are several types of syrups on the market but the three main classes are those made by:

ß 2004, Woodhead Publishing Limited

1. 2. 3.

acidic hydrolysis (regular syrups) acidic hydrolysis followed by enzymic hydrolysis (dual conversion syrups) enzymic hydrolysis alone.

7.6.1 Syrups from acidic hydrolysis Hydrolysis of starch at temperatures of 130±170 ëC using hydrochloric acid has been carried out for decades. The hydrolysates are partially neutralised with soda ash, treated with activated carbon, filtered and evaporated to the desired solids content. The degree of hydrolysis (degree of conversion) is measured by the dextrose equivalent (DE) which is the percentage of reducing sugars calculated as glucose (dextrose). For batch conversion, the contents of the acid resistant converter are cooled and neutralised when the required DE is achieved. This can be between 34 DE, the product used for toffee and hard confectionery and 55 DE, the product for soft confectionery. The major product used is about 42 DE, having applications, not only in confectionery, but in jams, preserved fruit, etc. Carbohydrate composition of syrups of this type are shown in Table 7.2. Modern practice is to carry out the conversion in continuous converters such as the type manufactured by Karl Kroyer (Aarhus, Denmark). Acidified starch slurry (40% solids content) is pumped through a series of superheated steam pipes, the desired degree of hydrolysis being achieved by control of temperature, acidity (based on conductivity) and retention time (Cornell and Clack, 1965). The quality of the wheat starch used is critical to the quality of the resultant syrup and it is important to control the conditions for optimal removal of lipids and peptides (Cornell and Hoveling 1998). 7.6.2 Syrups from enzymic hydrolysis Enzymes provide a means of tailoring the properties of the syrup to meet specialised needs. One example of this is high maltose syrup, usually prepared Table 7.2 Composition of acid-converted corn starch hydrolyzates as determinedby chromatographic analyses Dextrose equivalent 10 20 30 40 50 60 65

Mono

Di

Tri

Tetra

2.3 5.5 10.4 16.9 25.8 36.2 42.5

2.8 5.9 9.3 13.2 16.6 19.5 20.9

2.9 5.8 8.6 11.2 12.9 13.2 12.7

3.0 5.8 8.2 9.7 10.0 8.7 7.5

Penta

Hexa

Hepta

Higher

3.0 5.5 7.2 8.3 7.9 6.3 5.1

2.2 4.3 6.0 6.7 5.9 4.4 3.6

2.1 3.9 5.2 5.7 5.0 3.2 2.2

81.7 63.3 45.1 28.3 15.9 8.5 5.5

Based on recent chromatographic analyses of 91 samples from all US manufacturers. The data are considered accurate to 1.0 percent.

ß 2004, Woodhead Publishing Limited

by initial acid hydrolysis, followed by treatment with -amylase (EC 3.2.1.2), an exoglucosidase that hydrolyses -1,4- glycosidic bonds from the non-reducing ends of chains to form maltose. Thus by taking the acid hydrolysis only to 15 DE, when the glucose and maltose contents are each only about 4%, treating with -amylase (24 hours, 55 ëC, pH5) to about 42 DE a syrup with a maltose content of 44% of solids can be obtained. The reaction can be terminated by deactivation of the enzyme at 90 ëC and pH3. This syrup has a much higher content of maltose than that of regular starch syrup of 42 DE (14%) and results in a product which gives high boiling confectionery (hard sweets) less stickiness and less `graining' (tendency to cloudiness due to crystallisation), i.e. superior keeping qualities. This type of confectionery is normally based on mixtures of syrup and cane sugar (sucrose) often about 35:65 respectively. Syrups of the high maltose type can also be made using bacterial -amylase instead of acid in the first stage. High DE syrups (63±65 DE) are also produced by dual conversion except that acidic hydrolysis is allowed to proceed to about 30 DE, followed by reaction with -amylase (6 hours, 55 ëC, pH 5) when the DE is about 48, and finally reaction with amylo --1,4--1,6- glucosidase (EC 3.2.1±), an exoglucosidase that hydrolyses -1,4- and -1,6-glycosidic bonds from the non-reducing ends of amylose and amylopectin producing glucose (sometimes referred to as glucoamylase or -amylase). To achieve 63±65 DE, the reaction is continued for about 24 hours and then terminated by deactiviting the enzymes as before. This high DE syrup is ideally suited for soft-eating confectionery, which when made with regular syrup tend to dry out and become tough on storage. Marshmallows, fudge and soft jellies are examples of confectionery use, while in the baking industry buns, cakes and sponges also have enhanced shelf life because of greater moisture retention. Being less sweet than cane sugar, it is of advantage also in the manufacture of jams and canned fruit, where replacement of the part of the sugar results in enhancement of flavour. The extent of replacement in many applications will be governed by economic considerations. The carbohydrate composition of a typical syrup is shown in Table 7.2. High sweetness syrups contain high amounts of fructose, which is much sweeter than glucose. This is achieved by the use of glucose isomerase (EC 5.3.1.18) an enzyme which brings about the isomerisation of glucose to fructose. The process is often carried out in columns of immobilised enzyme where the enzyme is linked to a carrier and is not released into the medium. Several types of syrups are available, containing 55±90% fructose (Whistler and Daniel, 1984). They are used chiefly as sugar substitutes, fructose being sweeter than sucrose. Hence they have applications in soft drinks and cordials. Glucose has, in the past, been manufactured by acidic hydrolysis but the high DE syrups required suffered from bitter flavour due to the presence of undesirable side products (reversion products). This process was improved by the application of amyloglucosidase after partial acidic hydrolysis but nowadays, glucose is manufactured using a thermostable bacterial amylase for the first stage, converting the starch to dextrins, followed by a saccharification stage to

ß 2004, Woodhead Publishing Limited

98 DE using amyloglucosidase. When the activated carbon-treated syrup is evaporated to 75% solids content and seeded with glucose crystals, the bulk of the glucose in the batch crystallises and the mixture is then centrifuged to give glucose which is milled and sieved. The final product is close to 100 DE, while the supernatant material from centrifugation (`Hydral') is about 85 DE. Syrups for fermentation are manufactured using a similar enzyme treatment, but do not require seeding and centrifugation. These syrups are required to have a high percentage of yeast-fermentable sugars, on a solids basis, without any undesirable flavour and need to be of low cost in order to replace cane sugar in beverages such as beer, ale and stout. With syrups for yeast manufacture, the off-flavour is not so important and products of about 90 DE are acceptable, provided that the cost per tonne of fermentable sugar solids is competitive with cane sugar. All these examples show that chemical modification of starch has again been useful in providing a different class of product and great variety within that class. Syrups, being a major outlet for starch producers, provide a more versatile range of products than can be obtained from sugar and is a viable alternative to sugar in many countries. A combination of these products with sugar still continues to be widely used in the food industry, as the addition of glucose syrups, with their range of simple and oligosaccharides, prevents crystallisation of the cane sugar. In the ice-cream industry, corn syrups and dextrose are used to control sweetness and improve body and texture. Regular DE syrups depress the freezing point somewhat less than sucrose (on equal solids basis) but glucose (dextrose) lowers the freezing point more than sucrose. Hence a good deal of control of physical properties is provided by the starch syrups. Glucose syrups are an excellent example of chemical modification of starch which leads to a wide range of useful products acting as sweeteners, stabilisers and fermentable syrups designed not only to replace or partly replace sucrose, but to provide benefits to the formulation.

7.7

Analysing starch-based products

7.7.1 Starches Starch, is required to be of high quality for food applications. Quite apart from the analytical procedures used for elucidation of structural characteristics, which have been discussed in earlier chapters, quality control is an essential and practical aspect which requires close attention by the starch manufacturer and the user alike. For most applications, the starch manufacturer will ensure that a product with standard specifications for protein, pH, moisture content and ash (mineral) content will be supplied as these do not vary greatly. However, paste characteristics, microbiological standards and foreign matter (such as specks) will vary from time to time and the user, as well as the supplier should check all batches and correlate them with deliveries. Obviously, food manufacturers rely

ß 2004, Woodhead Publishing Limited

on the bland taste of starch and organoleptic tests should not indicate any foreign odour or taste. Taste panels of experienced personnel are required to test starch pastes made under standardised conditions. Modified starches will require further tests designed to ensure that the chemical modification has brought about the desired effects. These will include paste clarity and viscosity characteristics, freeze-thaw stability, high temperature stability and stability to acid. Most of these specifications will be set by the user, to suit particular products, who will test each batch delivered. Viscosity characteristics will be observed in an instrument like the Brabender Amylograph and samples after cooking can be diluted with hot water for measurement of paste clarity in a UV-visible spectrophotometer (Wiesenborn et al., 1994). Tests can be performed on 7% w/v starch gels kept at 4 ëC (if necessary at low pH), using syneresis as a measure of the degree of retrogradation (Tjahjadi and Breene, 1984). These tests can be modified to test freeze-thaw stability ÿ20 ëC/+20 ëC cycles as required. High temperature stability should be carried out in an autoclave, with and without the presence of acid.

7.7.2 Starch syrups Starch syrups also require different types of tests. Tanker deliveries of syrups can be readily tested for solids content using refractive index (Corn Industries, 1956). Determination of DE may involve a hot titration with mixed Fehling's solutions (an alkaline cupric tartrate solution). Quantitative reduction of Cu (II) to Cu (I) occurs at boiling point and the end point may be detected with methylene blue (a `redox' indicator). By comparison of the titres of the diluted syrup and a standard glucose solution, the DE of the syrup can be determined. For the confectionery industry, some form of practical cooking test with cane sugar is required. This is often referred to as a barley sugar or candy test and requires heating of the mixture to about 175 ëC. After pouring the hot mixture into a stainless steel dish, the colour of the candy, foam height reached around the side of the dish and the time for the foam to disperse can all be used to indicate the quality of the syrup. For acid-hydrolysed syrups, the salt content of the syrup can be determined volumetrically by the traditional Mohr titration or an ion-specific electrode. Sodium metabisulfite is often used as a preservative in starch syrups and the level of SO2 can be determined iodometrically. Analysis of the different sugars and oligosaccharides present in syrups can be carried out by HPLC with estimation of the total carbohydrate present in each fraction by the resorcinol-sulfuric acid reaction. This type of analysis is important for comparison of syrups made by enzymic processes against those made by acidic hydrolysis. Trace elements such as Cu, Fe and Ca may need to be determined and this can be done using atomic absorption spectroscopy or colorimetric techniques.

ß 2004, Woodhead Publishing Limited

7.8

Future trends

Starch is becoming more important as the size of the international food market expands. Where starch per se is not so important as flour, for economic reasons, the quality of the starch in that flour requires careful examination. Toyokawa et al. (1989) showed the importance of starch swelling to Japanese noodle texture and Batey and Mueller (1991) chromatographed starch digests from different wheat cultivars and examined the oligosaccharides produced in regard to noodle-making quality. Further experiments along these lines will be fundamental for the growing Asian food market. Wheat starch has many features which can be improved by chemical modification and a large effort must be put into making modified wheat starch with improved functionality, e.g., improved gel texture, paste clarity and stability to conditions imposed by modern food formulations. Unless this is done, low protein wheat flours, such as those produced by air classification, will replace starch produced by wet-milling of flour. It is likely that many chemical modifications will be less successful with these low protein flours. More use should be made of amylose and amylopectin, the preferred forms being from genetically modified grain. Amylose has excellent qualities as a film former. As an edible, non-toxic, biodegradable polymer it is unsurpassed. Amylopectin has the ability to soften the texture of wheat starch gels and lessen retrogradation. At present, a very significant proportion of wheat starch is modified enzymically to produce a range of syrups for the food industry and the same approach is already being used with low protein air classified flour. The wet-milling of wheat flour for starch is supported by the demand for wheat gluten. So long as there is a strong demand for gluten to reinforce weaker flours and to produce high protein breads, the concomitant production of wheat starch will remain a viable industry, especially while the wheat syrup industry allows better quality specialised products for the food industry.

7.9

Sources of further information and advice

There is a considerable amount of research being carried out on starch because of its applications in so many industries. Food applications are numerous amongst these references, many of which are in conventional sources such as technical books and journals and proceedings of conferences. Connections to the World Wide Web (www) can be made through `Uncover' (University of Colorado), `Chemical Abstracts', `Current Contents' and other data bases and home pages of various institutions are available through Netscape, Internet Explorer, Google and others. 7.9.1 Journals on starch research Starch research is included in a number of journals on carbohydrates such as Carbohydrate Research, Methods in Carbohydrate Chemistry, and

ß 2004, Woodhead Publishing Limited

Carbohydrate Polymers. The most important of the journals specialising in starch is Starch/StaÈrke, published by Wiley-VCH.

7.9.2 Journals on cereal foods Many articles on starches for food applications appear in journals concerned with food chemistry, particularly those specialising in cereals such as Cereal Chemistry, Cereal Foods World, and Journal of Cereal Science. Other journals with food applications include Food Technology, Journal of Food Science, Journal of the Science of Food and Agriculture and the Journal of Agriculture and Food Chemistry. Journals on nutrition such as the American Journal of Clinical Nutrition, Nutrition Research and others of this type include articles on starch from time to time.

7.9.3 Books on wheat, food and starch The following books include the chemistry and applications of wheat starch: Food Chemistry 2nd edn, ed. Fennema O, 1985, New York, Marcel Dekker Inc. Handbook of Cereal Science and Technology 2nd edn, eds Lorenz KJ and Kulp IG, 2000, New York, Marcel Dekker Inc. Wheat Chemistry and Technology Vols 1 and 2, 1988, ed. Pomeranz, Y St Paul Amer. Assn Cereal Chemists Wheat Chemistry and Utilization, 1998, Cornell HJ and Hoveling AW, Baton Rouge, CRC Press Starch: Chemistry and Technology, 2nd edn, 1984, eds Whistler RL, Be Miller JN and Paschall EF, New York, Academic Press Inc.

7.9.4 Proceedings from conferences and institutional reports Starch is included in the programmes of international cereal conferences and reports from university and other centres of excellence in agricultural chemistry. Among these are: Kansas State University, USA, Department of Agriculture Ohio State University, USA, Department of Agriculture CSIRO, Australia, Division of Food Science, Wheat Research Unit USA Departments of Agriculture, Peoria and Albany Conference Proceedings of the American Association of Cereal Chemists.

7.9.5 Starch on the www The following topics presented in this chapter are accessible on the www: Wheat starch, production, properties, foods, rheology, chemical modification, starch syrups, analysis

ß 2004, Woodhead Publishing Limited

Courses and specialist groups Journals: electronic and non-electronic Meetings: includes abstracts of conferences

7.10

References

and SMELLIE JM (1952) `Coeliac disease:gastrointestinal studies and the effect of dietary wheat flour' Lancet i 836±842. BARNES HA, HUTTON JF and WALTERS K (1989) An Introduction to Rheology, New York, Elsevier Science Publishing Company Inc., 1±54. BATEY IL and MUELLER BM (1991) `Variations in the structure of Australian wheat starch and the relationship to starch quality', Cereals International, eds Martin DJ and Wrigley CW, Melbourne, Royal Australian Chemical Institute, 147±149. BE MILLER JN and WHISTLER RL (1996) `Carbohydrates' in Fennema OR (ed.), Food Chemistry 3rd edn, New York, Marcel Dekker Inc., 191±204. BELLO-PEREZ LA and PAREDES-LOPEZ O (1994) `Starch and amylopectin:rheological behaviour of gels', Starch/StaÈrke 46 411±413. BETANCUR-ANCONA D, GARCIA-CERVERA E, CANIZARES-HERNANDEZ E and CHEL-GUERRERO L (2002) `Chemical modification of Jack bean starch by succinylation' Starch/StaÈrke 54 540±546. BLOCKSMA AH (1972) `Rheology of wheat flour doughs', J Texture Studies 3 3±17. CASE SE, CAPITANI T, WHALEY JK, SHI YC, TRZASKO P, JEFFCOAT R and GOLDFARB HB (1998) `Physical properties and gelation behaviour of a low-amylopectin maize starch and other high-amylose maize starches', J Cereal Sci 27 301±314. CHEETHAM NWH and TAO L (1997) `Amylose conformational transitions in binary DMSO/ water mixtures', Starch/StaÈrke 49 407±415. CHEETHAM NWH and TAO L (1998) `Amylose conformational transitions in binary DMSO/ water mixtures', Carbohydrate Polymers 35 287±295. CORN INDUSTRIES RESEARCH FOUNDATION (1956) Corn syrups and sugars, Washington DC. CORNELL HJ (1963) `The effect of amylopectin on the properties of starch gels', Starch/ StaÈrke 15 43±50. CORNELL HJ (1970) `Some of the minor constituents of wheat starch and its hydrolysatesPart 1', Starch/StaÈrke 22 55±64. CORNELL HJ and CLACK ID (1965) `The application of conductivity measurement to the study of starch hydrolysis kinetics', Starch/StaÈrke 17 7±13. CORNELL HJ and HOVELING AW (1998) Wheat-chemistry and utilisation Boca Raton, CRC Press. CORNELL HJ, MCGRANE SJ and RIX CJ (1999) `A novel and rapid method for the partial fractionation of starch using 1±butanol in the presence of thiocyanate', Starch/ StaÈrke 51 335±341. CORNELL HJ, RIX CJ and MCGRANE SJ (2002) `Visometric properties of solutions of amylose and amylopectin in aqueous potassium thiocyanate' Starch/StaÈrke 54 517±526. DE MAN JM (1990) Principles of Food Chemistry, 2nd edn, New York, Van Nostrand Reinhold, 163±168. DOUBLIER JL and LLAMAS G (1993) `A rheological description of amylose-amylopectin ANDERSON CM, FRAZER AC, FRENCH JM, GERRARD JW, SAMMONS HG

ß 2004, Woodhead Publishing Limited

mixtures', in Dickinson E and Walstra P (eds), Food Colloids and Polymers: Stability and Mechanical Properties, Cambridge, The Royal Society of Chemistry, 138±146. DOUBLIER JL, PATON D and LLAMAS G (1987) `A rheological investigation of oat starch pastes', Cereal Chem, 64(1), 21±26. ELIASSON AC and BOHLIN L (1982) `Rheological properties of concentrated wheat starch gels', Starch/StaÈrke, 34(8), 267±271. ELIASSON AC and TJERNELD E (1990) `Adsorption of wheat proteins on wheat starch granules', Cereal Chem 67 366±372. FRENCH D (1984) `Organization of starch granules' in Starch:Chemistry and Technology 2nd edn, Whistler RL, BeMiller JN and Paschall EF eds, New York, Academic Press 183±247. GHIASI K, HOSENEY RC and VARRIANO-MARSTON E (1982) `Gelatinization of wheat starch III. Comparison by differential scanning calorimetry and light microscopy', Cereal Chem 59 258±262. GREENBLATT GA, BETTGE AD and MORRIS CF (1995) `Relationship between endosperm texture and the occurrence of friabilin and bound polar lipids in wheat starch', Cereal Chem 72 172±176. HAYASHI A, KINOSHITA K, MIYAKE Y and CHO C-H (1983) `Gelation and its related changes in amylose solution', Agricultural and Biological Chemistry 47, 1699±1704. HOSENEY C (1994) Principles of Cereal Science and Technology, 2nd edn, St Paul, American Association of Cereal Chemists, 29±64. JANE JL and CHEN JF (1992) `Effect of amylose molecular size and amylopectin branch chain length on paste properties of starch', Cereal Chem 69 60±65. KERR RW and CLEVELAND FC, JR (1962) `Thickening agent and method for making the same', US Patent 2 961 440. KIM HR, MUHRBECK P and ELIASSON AC (1993) `Changes in rheological properties of hydroxypropyl potato starch pastes during freeze-thaw treatments III. Effect of cooking conditions and concentration of the starch paste', J. Sci Food Agric 61 109±116. KITAMURA S, KOBAYASHI K, TANAHASHI H, OZAKI T and KUGA T (1989) `On the MarkHouwink-Sakurada equation for amylose in aqueous solvents', Denpur Kagaku 36 257±264. LIM S and SEIB PA (1993) `Preparation and pasting properties of wheat and corn starch phosphates', Cereal Chem 70 137±144. LINEBACK DR and RASPER VF (1988) `Wheat carbohydrates' in Wheat: Chemistry and Technology, Pomeranz Y ed. St Paul, Am Assn Cereal Chem 377±72. MCGRANE SJ (2001) `Rheological properties of starch polysaccharides' PhD Thesis, RMIT University, Melbourne. MCGRANE SJ, CORNELL HJ and RIX CJ (1998) `A simple and rapid colorimetric method for the determination of amylose in starch products', Starch/StaÈrke 50 158±163. MILES MJ, MORRIS VJ and RING SG (1984) `Some recent observations on the retrogradation of amylose', Carbohydrate Polymers 4 73±77. MILES MJ, MORRIS VJ and RING SG (1985) `Gelation of amylose', Carbohydrate Research 135 257±269. MORRISON WR (1978) `Wheat-lipid composition', Cereal Chem 55 548±558. NIELSEN LE (1977) Polymer Rheology, New York, Marcel Dekker Inc., 1±10. PAROVUORI P, MANELIUS R, SUORTTI T, BERTOFT E and AUTIO K (1996) `Effect of amylopectin molecular weight on the rheological and structural properties of

ß 2004, Woodhead Publishing Limited

amylose-amylopectin mixtures', in Phillips GO, Williams PA and Wedlock DJ (eds). Gums and Stabilisers for the Food Industry 8, Oxford, IRL Press at Oxford University Press, 217±223. RING SG, COLONNA P, L'ANSON KJ, KALICHEVSKY MT, MILES MJ, MORRIS VJ and ORFORD PD (1987) `The gelation and crystallisation of amylopectin', Carbohydrate Research 162 277±293. ROBIN JP, MERCIER C, CHARBONNIERE R and GUILBOT A (1974) `Lintnerized starches. Gel filtration and enzymic studies of insoluble residues from prolonged acid treatment of potato starch', Cereal Chem 51 389±406. SCHOCH TJ (1942) `Fractionation of starch by selective precipitation with butanol', J Am Chem Soc 64 2957±2961. SEGUCHI M (1986) `Dye binding to the surface of wheat starch granules', Cereal Chem 63 518±520. SEGUCHI M (1995) `Surface staining of wheat starch granules with Remazol brilliant blue R dye and their extraction, with aqueous sodium dodecyl sulfate and mercaptoethanol', Cereal Chem 72 602±608. SHIBANUMA K, TAKEDA Y and HIZUKURI S (1994) `Molecular structures of some wheat starches', Carbohydrate Polymers 25 111±116. SLADE L and LEVINE H (1987) `Influences of the glassy and rubbery states on the thermal, mechanical and structural properties of doughs and baked products', in Pearson AM, Dutson TR and Bailey A (eds). Advances in Meat Research, Volume 4, Collagen as a Food. New York, The AVI Publishing Company Inc., 251±266. SVAROVSKY L (1984) Hydrocyclones, Lancaster, Technomic Publishing Co., PA. TAKAHASHI S, MANINGAT CC and SEIB PA (1989) `Acetylated and hydroxypropylated wheat starch: Paste and gel properties compared with modified maize and tapioca starches', Cereal Chem 66 499±506. TAKEDA Y, SHIRASAKA K and HIZUKURI S (1984) `Examination of the purity and structure of amylose by gel-permeation chromatography', Carbohyd Res 132 83±92. TAKEYAMA H, KOBAYASHI M, YAJIMA H, ENDO R, KOHYAMA K and NISHINARI K (1993) `Gelation Process of Amylose-DMSO-Water System', Macromolecular Chemistry 76 83±88. THOMAS DJ and ATWELL WA (1999) Starches St Paul, USA, American Association of Cereal Chemists. TJAHJADI C and BREENE WM (1984) `Isolation and characterization of Adzuki bean starch', J Food Sci 49 558±562. TOYOKAWA H, RUBENTHALER GL, POWERS JR and SCHANUS EG (1989) `Japanese noodle quality II Starch components', Cereal Chem 66 387±391. WHATTAM J and CORNELL HJ (1991) `Distribution and composition of the lipids in starch fractions from wheat flour', Starch/StaÈrke 43 152±156. WHISTLER RL and DANIEL JR (1984) `Molecular structure of starch', in Starch: Chemistry and Technology, 2nd edn, eds Whistler RL, Be Miller JN and Paschall EF, New York, Academic Press Inc. WIESENBORN DP, ORR PH, CASPER HH and TACKE BK (1994) `Potato starch paste behaviour as related to some physical/chemical properties', J Food Sci 59 644±668. WONG RBK and LELIEVRE J (1982) `Rheological characteristics of wheat starch pastes measured under steady shear conditions', J Appl Polymer Sci 27 1433±1440. ZOBEL HF (1988) `Molecules to granules.A comprehensive starch review', Starch/Sta È rke 40 44±50.

ß 2004, Woodhead Publishing Limited

8 Developments in potato starches W. Bergthaller, Federal Centre for Nutrition and Food, Germany

8.1

Introduction

In 1984 Mitch described in a concise way the functional properties of potato starch, which allow us to understand the preference of this type of starch for diverse fields of application including specific food preparations. The most important characteristics mentioned were the formation of high consistency on pasting and excellent flexible films, the binding power and lower gelatinisation temperature (Mitch, 1984). Pastes of potato starch are also very sensitive towards heating and agitation and react with a significant decrease in viscosity. The unique reaction of paste viscosity towards the electrolytes present in aqueous media is remarkable, in particular the effects of monovalent cations like sodium and potassium as well as divalent cations like calcium and magnesium (Nutting, 1952; RuÈggeberg, 1953). While sodium ions increase paste viscosity to maximum levels, calcium ions produce the contrary by `ionic cross-linking' (Bergthaller and Fehn, 1977). The reaction is connected with the presence of covalently bound phosphate ester groups generally described as integral parts of the amylopectin fraction of potato starch (RuÈggeberg, 1953; Radomski and Smith, 1963; Hizukuri et al., 1970; Bay-Smidt et al., 1994) and their ability to replace their ionic load. Depending on electrolyte type, concentration and condition the differently charged phosphate groups easily undergo ion exchange (Wegner, 1957; Pa•asinski, 1963). When comparing the property profile of potato starch with other industrially produced starch types (maize starch, wheat starch, tapioca starch) not only the viscosity behaviour of its paste is outstanding, but also other characteristics like granule size/granule size distribution, the concentration of protein, lipids and minerals and, as mentioned in the foregoing, especially the

ß 2004, Woodhead Publishing Limited

phosphorous concentration is unique. Further important aspects deal with the high transparency of formed starch pastes and gels and the low tendency towards formation of new structures and retrogradation (Alexander, 1995). The indicated property profile depends greatly on the genetic basis of the potatoes from which the starch has been extracted (Putz and Tegge, 1976; Putz et al., 1978; Cotrell et al., 1995, Haase and Plate, 1996; Brunt and Zinsmeester, 1997). Even when the genotype determines the basic properties of potato starch, growing conditions like fertilisation and environmental conditions also contribute to the formation of many aspects of the property profile (de Willigen, 1951a/b; GoÈrlitz, 1963; Putz and Tegge, 1976; Putz et al., 1978; Cotrell et al., 1995, Haase and Plate, 1996). After all, also tuber size and harvesting (tuber maturity) (Haase, 1993; Brunt and Zinsmeester, 1997) and selected technologies in processing, in particular the separate recovery of a small-size starch fraction in 3-phase nozzle separator and hydrocyclone washing may affect the quality spectrum of recovered starch. In contrast to its potential, economic constraints and the fact that potato starch is a low-price commodity provokes only a limited interest in recovering and utilising the full range of quality aspects.

8.2

Components and rheological properties of potato starch

8.2.1 Impurities of potato starch Available literature (Mitch, 1984; Gaillard and Bowler, 1987; Alexander, 1995; Brunt and Zinsmeester, 1997; Vasanthan et al., 1999) describes native potato starch as very pure in comparison to competing cereal starches (Table 8.1). In Germany, limits concerning the main quality criteria of national regulations for commercial potato starch (Bergthaller et al., 1999) document that, in general, concentration of minor components in industrially produced starch must be smaller and the quality even better than the reported levels may indicate. The high level of purity, in particular of lipids, grants potato starch a neutral taste and prevents it from developing an off flavour under conditions of long-term storage. With respect to reported levels of proteins (0.05 to 0.2%) and lipids (traces to 0.2%) the level of these impurities originates in the development of starch granules within tissue cells of potato tubers during cultivation and storage where Table 8.1

Composition of potato starch-moisture content and minor components

Reference Mitch (1984) Gaillard and Bowler (1987) Alexander (1995) Haase and Plate (1996) Vasanthan et al. (1999)

Moisture (%)

Protein

17±18 12±17

traces 0.05±0.20

0.35 0.3±0.4

16±18 ± 13±17

0.06 ± 0.07±0.14

0.04 0.08 ± 0.04±0.12 0.18±0. 32 0.05±0.08

ß 2004, Woodhead Publishing Limited

Minerals

Phosphorus

Lipids

(% dry weight basis) ± 0.04±0.13

± 0.0±0.20 0.05 ± 0.08±0.17

the granules remain loosely distributed in contact with the cells potato juice (Bergthaller et al., 1999). When removed from destroyed potato cells and washed with softened or highly purified water in a counter-current stream during starch isolation, nitrogen-containing organic and inorganic compounds are separated to a maximum amount. For mineral components the same situation is evident disregarding the phosphorus bound to amylopectin in the form of phosphate ester groups and the metal cations bound to them. All surplus minerals are washed out during processing (Nutting, 1952; RuÈggeberg, 1953; Winkler, 1958; Wegner, 1959; Bergthaller et al., 1999).

8.2.2 The importance of phosphate ester groups The relation of phosphate ester group concentration and potato starch granule size is evident and has been documented occasionally since the early 1950s (GoÈrlitz, 1963; Putz et al., 1978). On the other hand, the connection of starch granule particle size and phosphate concentration with concepts of gelatinisation, paste formation and viscosity is even older and a result of manifold investigations during the last century. Based on contemporary knowledge Putz and co-workers (Putz and Tegge, 1976; Putz et al., 1978; Putz and Tegge, 1980) as well as Haase and co-workers (Haase and Putz, 1991; Haase and Plate, 1996) concluded that granule particle size is especially relevant for the rheological behaviour of potato starch and defined amylose content, phosphorous content, paste peak viscosity and a specific large granule fraction are prominent quality characteristics to be used in the selection of potato lines and varieties to be utilised in conventional breeding to suit future starch production.

8.2.3 Rheology of aqueous dispersions Concerning functionality in food systems, potato starch is well known for providing high viscosity and forming pastes and gels of long consistency. This behaviour of pastes and gels is combined with high clarity. This property profile is not only advantageous for technical purposes but also for applications in food production, where clarity is especially desired (Mitch, 1984). However, inherent to potato starch are also shear and heat sensitivity which are serious disadvantages. They are critical under conditions of modern food processing, for example in sterilising, cooling and freezing and following storage of the products. The use of native potato starch as a thickening agent or binder is represented by specific rheological characteristics that can be explained by the molecular dimensions of amylose and amylopectin and especially the structure of the amylopectin fraction (Blanshard, 1987) and the contribution of phosphate ester groups, their charge and the type and quantity of cations (K+, Na+, Ca2+ and Mg2+) associated with potato starch (de Willigen, 1951b; RuÈggeberg, 1953; Wegner, 1957; Winkler, 1958; Wegner; 1959; Pa•asinksi, 1963; Bergthaller et al., 1999). The repulsion of negatively-charged phosphate ester groups are in part

ß 2004, Woodhead Publishing Limited

also regarded as responsible for the rapid hydration, swelling and the typically high viscosity of potato starch pastes (Nutting, 1952; RuÈggeberg, 1953). When heated in aqueous suspension, potato starch loses irreversibly its granular structure as soon as a critical temperature range, the so-called gelatinisation temperature, is reached. Under such conditions granules swell tangentially at first and disintegrate under formation of a colloidal dispersion. For potato starch a temperature range reaching from 56 to 66 ëC has been reported (Tegge, 1984). The disintegration is connected with a sharp increase in viscosity, which is unusually pronounced with potato starch as a result of the additional hydration of the characteristic phosphate ester groups. Furthermore, cations connected with potato starch bind water in their typical way and withdraw water needed for hydration of starch molecules thus contributing to further viscosity increase. The indicated effects of cations become visible, for example, in viscosity measurements (Fig. 8.1) where the resistance of a pasting starch suspension developed along a time-temperature programme is registered under rotation as torque by a measuring unit, for example, in a Brabender Viscograph or a Rapid Visco Analyzer. After the initial rise of viscosity following gelatinisation a peak maximum (1st maximum) is reached during or after heating to 95 ëC. This is followed by a sharp and for potato starch very pronounced breakdown to a so-called trough (minimum) as result of the shear forces applied by the mixing elements at high temperature. The recorded decrease of viscosity (breakdown) characterises the sensitivity of potato starch towards shearing and heat treatment. The following cooling phase allows potato starch pastes to settle again and gels form a so-

Fig. 8.1 Cation (Na+, Ca2+) effects on viscosity development of potato starches after completed exchange of mineral load in comparison to native potato starch (commercial product).

ß 2004, Woodhead Publishing Limited

called set back (2nd maximum) (Tegge, 1984). In contrast to the well known firm and stiff gels of cereal starches (effect of pudding formation), potato starch gels are characterised in general by a stringy and slimy texture. However, differences can be pronounced and eventually rated in a more appropriate way especially when sodium or potassium starches, according to Winkler (1958) are used (de Willigen, 1951a; Tegge et al., 1976). In characterising the rheological quality of native potato starch one has to consider steadily the differently orientated contribution of cations being present and affecting the network of the amylopectin molecules in pasting (Bergthaller et al., 1999). Despite the demonstrated partly problematic property profile, native potato starch is still an important thickening agent in soups and sauces, different types of directly and indirectly expanded, coated and baked snacks, in fish and meat products, starch and flour-based noodles and many baked goods.

8.2.4 Resistance against retrogradation Another property generally disadvantageous to cereal starch dispersions and gels is their inclination towards structural changes indicated as retrogradation or set back. Dispersed starch systems tend to change their state of order and at the same time to release bound water. The visible effect is called syneresis and can be understood as an indication of intensive changes in gel structures. With respect to food preparations such changes are highly undesirable, especially in products that require freeze-thaw stability. Another advantage of potato starch is that dispersions and gels hardly ever demonstrate effects of retrogradation (Mitch, 1984). Even after repeated application of freeze-thaw cycles the formation of newly ordered structures as an indication for retrogradation effects could hardly be observed in DSC studies. (Gaida et al., 1987) Nevertheless, in concentrated systems retrogradation effects during long-term storage cannot be excluded.

8.3

Techniques for producing potato starch

Due to the tissue structure of potatoes and the fact that starch is deposited loosely in the fruit water of potato cells the process of starch production is relatively simple. Nevertheless, in modern potato starch technology there are divergent processes with different requirements concerning water consumption and different techniques to recover high molecular proteins dissolved in fruit water, in particular. Starch isolation begins, in principle, with disintegration of the storage parenchyma of the tuber followed either by immediate removal of the concentrated potato juice and subsequent washing out of starch or in other concepts by extracting the cell content with different amounts of water (Bergthaller et al., 1999; van der Ham, 2002). Disintegration is limited by the fact that the greater part of cell wall tissue residues should remain at a size separable by wet sieving mill starch streams. From the resulting diluted

ß 2004, Woodhead Publishing Limited

suspensions the starch gets rid of impurities, like fibre and protein residues as well as minerals, by a combination of concentration and washing steps. Countercurrent washing with soft fresh water in a multicyclone station or nozzle separators followed by conventional concentration and dewatering systems produces a starch cake, which is dried in the moving bed of a heated air stream to a moisture content of less than 20% in a flash dryer (Bergthaller et al., 1999; van der Ham, 2002). To guarantee a high level of purity and the optimum functionality of potato starch a well functioning process water and waste water regime are vital. A recent concept realised in a German potato starch factory covers its full wash water supply used in counter-current starch refining by RO water recovered from de-proteinized potato juice of delivered potatoes (Wilhelm et al., 2003).

8.4 Improving the functionality of potato starch for use in the food industry 8.4.1 General aspects Application areas of potato starches cover a wide range of uses. Almost every segment within the food industry where thickening agents or binders are generally used can be supplied with native or modified potato starch or other speciality products based on it. The most important areas of application are represented by the following food segments: soups and sauces, meat, fish, potato, and milk products, spreads, pasta and noodles, canned fruits and vegetables, ready-to-eat menus, bakery goods and confectionery, snacks. Besides thickeners and binders starch and starch derived products can be added to varied examples of the said segments also as fat mimetics or fillers. For many of the envisaged areas of application, however, potato starch must be adapted by selected modification reactions to fulfil the requirements of processors to display its property profile fully even under production conditions of modern industrial food preparation, e.g., acid pH, freezing, heat treatment, shearing and their combinations. It is evident that the indicated conditions are often disadvantageous to the maintenance of desired functionalities typical of native starch use. With respect to the type of modification one can differentiate chemical, physical and enzymatic treatments that are used separately or in sequence in producing starch derivatives to get them adjusted in an optimum way to the requirements of specific applications. For food processing the most promising modifications result from various chemical reactions including acid-thinning or oxidising, formation of esters with phosphates, phosphorous oxychloride and anhydrides of short chain organic acids and formation of ethers with propylene oxide. To adapt and, in particular, stabilise the main functional properties, esters and ethers are often cross-linked, for instance, by multifunctional groups (Wurzburg, 1986b; Blanshard, 1987). Amongst physical modifications of starch, including chemically modified starch types disintegration of the granular products by gelatinisation and drying

ß 2004, Woodhead Publishing Limited

on a heated drum (Fritze, 1971) or destructurisation by extrusion cooking (Colonna et al., 1989) are in general the preferred techniques. In this way instant starches are achieved that can be dispersed without heating (`cold soluble starches') or that swell at room temperature. A further recent method of physical pre-treatment can be found in spray cooking (Boersen, 2000). Further advantages can be seen in heat treatments where granular structure remains unchanged although property profiles gain a new characteristic. The respective procedures are known as heat-moisture treatment and annealing (Stute, 1992). Finally, even enzymatic treatments opened a segment of starch modification by limited degradation of gelatinised starch by -amylase resulting in multiple-use maltodextrins (Alexander, 1992; Kennedy et al., 1995).

8.4.2 Chemically modified potato starches Although the range of potential modification reactions is manifold, only a small segment is acceptable and in use, in particular for food applications (Table 8.2). It is not only the reaction pattern and type of the introduced group used that restricts food application of modified starches, but also the level of substitution that affects digestibility. The higher the degree of substitution (DS) is, the more Table 8.2

Modified potato starches and their preferred uses

Modification product

E-no.

Preferences in application

Oxidised starch

E1404

Monostarch phosphate Distarch phosphate

E1410 E1412

Acetylated distarch phosphate

E1414

Acetylated starch

E1420

Acetylated distarch adipate

E 1422

Hydroxypropyl starch Hydroxypropyl distarch phosphate

E1440 E1442

Starch octenylsuccinate

E1450

Oxidised starch acetate

E1451

Soups and sauces, canned fruits, ready-toeat menus, sugar confectionery, coated and extruded snacks Emulsion stabiliser, pudding, Soups and sauces, coated and extruded snacks, mashed potatoes, canned fruits, ready-to-eat menus Soups and sauces, canned fish, cooked and formed meat products, ready-to-eat desserts, canned fruits, ready-to-eat menus Confectionery, diverse snacks, cooked meat products, mashed potatoes, flour-base noodles, baked goods, bakery fillings, ready-to-eat menus Soups and sauces, canned fish, cooked meat products, ready-to-eat desserts Soups and sauces, starch noodles Soups and sauces, canned fish, cooked meat products, bakery fillings, ready-to-eat desserts, canned fruits, marmalade, readyto-eat menus Emulsion stabiliser, dressing, encapsulation Sugar confectionery

ß 2004, Woodhead Publishing Limited

digestibility is reduced. Therefore, modified starches provided for food applications are prepared only at a very low to low level of substitution. Oxidised starch The purpose of oxidising potato starch is to decrease the characteristic high viscosity and result in a thin boiling starch that allows the development of a paste of high clarity at the same time. In general, oxidation is induced in aqueous starch suspensions under slightly alkaline conditions and careful control of temperature by addition of sodium hypochlorite (Wurzburg, 1986a). It is orientated towards the desired level of fluidity when pasted in hot water at higher concentration. The high gel-forming capacity combined with increased stability can be utilised in specific fields of confectionery, for instance, as a gelatine replacement for jellies and wine gums or in the production of convenience foods where body formation and mouthfeel should be regulated. When potato starch is only slightly oxidised a short and smooth texture can be induced. The main application is also connected with convenience foods. Its use is preferred in instant puddings, toppings, sauces and soup powders. Snack products and coatings for roasted nuts are other examples. Starch esters With respect to application, the most important group of esters produced with potato starch comprises monophosphate esters and acetylated starch. To further expand the interesting potential of acetyl esters of potato starch, acetylation is combined with different methods of cross-linking (preferably phosphate diesters and distarch adipates) (Wurzburg, 1986b). Starch phosphate monoesters The esterification of starch with inorganic phosphate salts affects dramatically the colloidal properties of the starch. Monosodium orthophosphate and sodium tripolyphosphate can be used for modification. Phosphate groups are introduced by a `dry heat' reaction (140 to 160 ëC with orthophosphates and 100 to 120 ëC with tripolyphosphate). Previously, starch was impregnated with phosphate salts in an aqueous suspension. After pH adjustment (4 to 6.5 with orthophosphate solutions or 5 to 8.5 with tripolyphosphate) and mixing the starch is filtered, dried and finally heated. Monosodium hydrogen phosphate and disodium hydrogen phosphate can result in DS up to 0.2, while sodium tripolyphosphate produces starch monoesters with 0.02 DS. Starch phosphate monoesters yield higher viscosity and form more clear and very stable dispersions, even under freezing conditions including freeze-thaw cycles. Increasing phosphorylation reduces gelatinisation temperatures. Passing DS 0.07 makes the monoesters swell in cold water. Water softness then plays a significant role in determining the level of swelling. The ionic properties of monoesters also suit them for emulsifying agents (Solarek, 1986).

ß 2004, Woodhead Publishing Limited

Acetylated starch For food uses low-DS acetylated starches ranging from 0.01 to 0.2 are of interest. The main advantage of this DS range is the preservation of the granular structure during derivatisation, which allows purification by washing with water and recovery by centrifugation or filtration. In the presence of dilute alkali (depending on temperature ÿ25 to 38 ëC ± pH 8.5 to 7), preferred sodium hydroxide or sodium carbonate, acetic anhydride reacts with granular (or gelatinised starch) with high efficiency (> 70%). Alternatively also vinyl acetate can be used for acetylation in aqueous alkaline (pH 7.5 to 12.5) suspension. Acetylated potato starch offers colloidal dispersibility combined with high clarity on cooking, film formation, binding, a capability of acting as thickener stabilising agent and acts also as texturising agent. Gelatinisation temperature is markedly lowered (by approximately 10 ëC) by the introduction of acetyl groups depending on DS. Although hot and cold paste viscosity is decreased, consistency and viscosity remain high. Susceptibility to acidic pH, freezing and heat and shearing is medium to low (Jarowenko, 1986). The application of starch acetate is widespread, for instance, in extruded and coated snacks, frozen fish and cooked meat products, flour based noodles, bakery products and in various frozen or cold stored ready-to-eat menus. Because of the remaining deficiency in heat and shear stability starch actetates are frequently reinforced by cross-linking (Wurzburg, 1986b). When oxidation of starch is combined with acetylation a modified potato starch especially suited for sugar confectionery (candy) is created. Concerning functionality, high clarity combined with medium pH and heat stability and low viscosity are desired properties. Starch octenyl succinate To provide starch the ability of an emulsion stabiliser for oil in aqueous systems a substituted cyclic dicarboxylic acid anhydrides can be used in a standard esterification reaction. A starch suspension is mixed with the reagent under alkaline conditions. For food purposes octenylsuccinic acid anhydride is allowed for production of starch octenyl succinate with a DS up to 0.02 (Trubiano, 1986). This type of starch ester, however, belongs to food additives of limited admission only, which means that a specific list clearly regulates application. Beverage emulsions, salad dressings and encapsulation are named as preferred fields. Cross-linking Although cross-linking can be used as a single reaction in modifying starch it is generally applied to stabilise starch molecules as thickeners, binding or texturising agents against the drastic effects of modern food production. As such, high shear in the mixing of food components or in an acidic environment can be indicated. In utilising phosphorous oxychloride or sodium trimetaphosphate a covalent phosphate diester bond can be introduced into starch that links its molecules and stabilises the high susceptibility of potato starch towards

ß 2004, Woodhead Publishing Limited

shearing, pH and freezing effects. In phosphate diester formation even very low substitution (DS approximately 10ÿ5) may have dramatic effects on viscosity, gelling and textural properties of the educts (Solarek, 1986). On the other hand, and as mentioned previously, cross-links reduce the dispersibility (DS > 10ÿ3) and can also be critical in starches determined for food uses. Another cross-linking agent of commercial significance and suitable for food purposes is adipic acetic mixed anhydride, which forms distarch adipates. These are made by esterifying granular starch in aqueous suspension under mildly alkaline conditions. In distarch adipates, the cross-links are combined with starch hydroxyls through organic ester linkages that are resistant to acidic conditions existing frequently in food. Under mildly alkaline conditions they hydrolyse readily alike phosphate linkages which are indicated to tolerate alkaline conditions a little more (Wurzburg, 1986b). Starch ethers Within starch ethers the outstanding and unique modified starch suited for food uses is hydroxypropyl starch. The alkaline-catalysed reaction of starch with propylene oxide in an aqueous slurry system was important for new developments in food industry. Full integrity of starch granules must be guaranteed to wash the product completely free of by-products after reaction. The modification for food uses (DS in general 0.1, approximately) improves coldstorage stability, clarity and textural properties of starch pastes made therefrom. Since the reaction is base catalysed 0.5 to 1% of sodium hydroxide relative to dry weight of starch is added to the starch slurry of 18 to 21 ëBeÂ. Salts (sodium chloride or sodium sulfate, 5 to 10%) are added, too, to protect granular integrity. In general, temperatures of 45 to 52 ëC together for about 24 h are used for reaction in the case of potato starch. With the introduction of hydroxypropyl groups internal structures of starch are weakened and gelatinisation temperature is decreased, in principle, until starch becomes cold-water soluble (Tuschhoff, 1986). In the development of conventional and new convenience-type food products hydroxypropyl starches have a wide spectrum of applications. According to Tuschhoff (1986) it improves shelf life, freeze/thaw stability, cold storage stability, cold water swelling and reconstituting properties to a formulated product. In case of potato starch-based hydroxypropyl starches, application concentrates on soups and meat-based sauces and starch noodles. Although the improvement of properties is impressive, in general, stability of viscosity is still critical, but can be improved significantly by cross-linking, which is done in the form of distarch phosphates, in general. The level of crosslinking depends greatly on cooking conditions in application. Paste texture turns also towards shorter characteristics. Tailoring of this type of starch needs careful control in the introduction of hydroxyl groups and phosphate cross-linking (Wurzburg, 1986b). Hydroxypropyl distarch phosphate has an impressive variability in application that starts as cooking starch, and in some cases also as instant

ß 2004, Woodhead Publishing Limited

starch, with soups and sauces, sterilised fish products, cooked meat products, fillings for confectionary, ready-to-eat desserts, canned fruits and jams and ends with different ready-to-eat menus especially with microwave-suited, cooled and frozen food preparations.

8.4.3 Physically modified potato starches Physical modification comprises treatments of controlled heat and water/ moisture applications like annealing and heat/moisture treatment as well as full disintegration of granules by precooking/pregelatinisation, extrusion or spraycooking. In annealing and heat/moisture treatment the integrity of starch granules (particle size, Maltese cross) remains untouched while swelling capacity is reduced. Along with this effect gelatinisation behaviour and viscosity undergoes nearly equivalent changes that are visible in Brabender curves by increased gelatinisation temperature and reduced peak viscosity. By modified annealing procedures potato starch may also gain gelling properties comparable to cereal starches (Stoof et al., 1998; Schmiedl et al., 1998) or results in resistant starch. In semi-dry conditions of heat/moisture treatments a much broader range of modification conditions can be applied with B and C type starches and allows finally the formation of a higher variability in product properties. Although general functionality of both treatments is comparable, for example, as increased viscosity, the structural changes in starch granules induced by heat/moisture treatment differ significantly, for instance with B and C type starches (Stute, 1992). Precooking/pregelatinisation was the preferred method that turns granular starch from a cold water (room temperature) indispersable into a fully dispersable form that is commercialised as an instant starch product. The most common and elaborate technique is conventional drum drying where a starch slurry is gelatinsed between or on the surface(s) of heated drum(s) and is distributed and dried on the surface until moisture contents of _DP/ 24 ratio is less than 20%, S-type: the ratio is higher than 24%.

Table 9.4 Amylose content, swelling power and solubility of starches from different rice cultivars (Sodhi and Singh, 2003) Cultivar

Amylose content (%) Swelling power (g/g) Solubility (%)

PR-106 PR-114 IR-8 PR-103 PR-113

16.1 bc 16.1 c 15.6 7.8 18.9

28.8 28.6 30.1 33.2 26.1

bc b c d a

0.319 0.360 0.307 0.287 0.346

b d b a c

Values with similar letter in column do not differ significantly (P100) and short chain amylopectin chains (average DP 17) were negatively and positively correlated with paste breakdown, respectively. However, Vandeputte et al. (2003b) using a more diverse set of rice starches including five waxy and ten non-waxy starches, found no significant correlation between amylopectin chain length distribution

ß 2004, Woodhead Publishing Limited

Fig. 9.7

RVA profiles of three waxy rices and three non-waxy rices, JY293 = Jiayu 293 (AAC: 25%), ZF504 = Zhefu 504 (12%), XS11 = Xiushui 11 (18%).

and PV, BD, SB and CPV. As with many apparently confusing situations in cereal chemistry, a study is needed that evaluates the association between amylopectin and amylose characteristics (i.e., content and structure) and starch pasting properties across a set of samples that are representative of the world's germplasm. A fractionation and reconstruction study could be telling as well. The pasting viscosity of rice flour, compared to its starch counterpart is reportedly different due to the influence of proteins and lipids which are removed during the production of starch (Singh et al. 2000; Fitzgerald et al. 2003). Experiments that analyzed pasting viscosities before and after protein and/or lipid flour removal indicate that protein has a major influence on flour pasting viscosity (Martin and Fitzgerald 2002; Fitzgerald et al. 2003). Changes in rice starch granule rigidity also are reported to change the pasting viscosity of starch compared to flour (Lai 2001). Wang et al. (2002) compared the pasting viscosity profiles of flours and starches from whole, broken and yellowed rice grain (Fig. 9.9). Compared to whole rice flour, broken rice flour had a similar pasting temperature but lower HPV and CPV, whereas yellowed rice flour exhibited a higher pasting temperature, less breakdown, and significantly higher HPV and CPV. However, the differences in the pasting properties between the three starches were not as distinct as compared to their flour counterparts. The pasting viscosity of rice flour or starch is affected by environmental and post-harvest treatments. Cultural management practices that increase rice grain

ß 2004, Woodhead Publishing Limited

Fig. 9.8 RVA pasting profiles of rice flour from the varieties `Jodon' and `Dixiebelle' which are similar in amylose content (25% AAC).

protein have been associated with decreased peak viscosity (Martin and Fitzgerald 2002). Martin and Fitzgerald (2002) concluded that the decreases seen in peak viscosity during rice storage result from an increased number of disulfide bonds. Teo et al. (2000) reported that ageing markedly affected the pasting behavior of rice flour, but had no effect on isolated starch. These authors suggested that modification of the protein component, rather than starch, was primarily responsible for the rheological changes associated with aging of rice flour (Teo et al. 2000; Zhou et al. 2003). Changes in protein content and disulfide bonds thus appear to be, at least in-part, controlling the differences seen in rice pasting viscosity during storage.

9.5.2 Gelatinization: rheological properties Rheological properties during rice starch gelatinization have been characterized by many researchers using different rheometers (Lii et al. 1995, 1996b; Sodhi and Singh 2003). Lii et al. (1996b) reviewed their results and concluded that the starch granular properties are the major factors responsible for starch rheological behavior, followed by the degree of amylose leaching during the gelatinization process, especially in high concentration systems. The change in viscoelastic properties of rice starch suspensions during gelatinization can be placed into three or four transition stages: starch suspension into sol, sol transition to gel, network destruction and network strengthening. During the early heating of starch granules in water, the increase of storage modulus G0 and tan  is relatively small, which indicates that amylose molecules are dissolved and the suspension has been transformed into a `sol' (Sodhi and Singh 2003). Then the G0 and G00 increase to a maximum in which the

ß 2004, Woodhead Publishing Limited

Fig. 9.9

Brabender amylographs of (A) 10% flours (dry basis); (B) 6% starches (dry basis) of whole, broken, and yellowed rice (Wang et al. 2002).

temperature coincides with the onset temperature, which is attributed to the close-packed network of swollen starch granules. Tan  decreases simultaneously, which indicates that a three-dimensional gel network has been constructed from amylose, reinforced by strong interactions among the swollen starch particles (Lii et al. 1995, 1996b; Sodhi and Singh 2003). The G0 increases during the gelatinization of rice starch are reported to be mainly governed by granule characteristics, which include swollen granule rigidity and the interaction between the close-packed granules (Lii et al. 1996b). In the third stage, continued heating beyond TG0 , the G0 decreases and tan  increases, indicating that the gel structure has been destroyed during prolonged heating (Lii et al. 1995, 1996b; Sodhi and Singh 2003). The destruction is likely due to the `melting' of the crystalline regions remaining in the swollen starch granule or results from the disentanglement of the amylopectin molecules in the swollen particles, which softens the particles. Another reason the network collapses may be due to the loss of interaction between particles and the network.

ß 2004, Woodhead Publishing Limited

The fourth stage, G0 and G00 are reported to increase and tan  increase even higher after an inflection point (Fig. 9.10). These authors attributed the G0 increase to the leached low molecular weight amylopectin, which interacted with the amylose matrix to strengthen the continuous phase (network). However, a larger tan  indicated that the dispersed phase (particles) became softer due to continuing dissolution of amylopectin.

9.6

Retrogradation and other properties of rice starch

Retrogradation describes the process in which a heated starch paste cools to below the melting temperature of starch crystallites, and the amylase and amylopectin reassociate and unite with the swollen starch grains in an ordered structure that results in viscosity increase, gel firming, and textural staling of predominantly starch-containing systems. The retrogradation properties can be measured by DSC (Qi et al. 2003; Vandeputte et al. 2003c), rheological properties, starch gel hardness (Vandeputte et al. 2003c), and NMR (Yao et al. 2002; Qi et al. 2003c). Methods to study retrogradation of starch have been reviewed by Karim et al. (2000). This phenomenon is generally regarded as a crystallization or recrystallization (i.e. formation and subsequent aggregation of double helices) process of amylose and amylopectin. The rapid initial rate of retrogradation relates to the loss of networked amylose, the development of amylose aggregates, and binding of granule remnants into assemblies by amylose and amylose aggregates. Thus, amylose is responsible for short-term (less than one day) changes during retrogradation (Zhou et al. 2002). Amylopectin forms shorter double helices which can be attributed to restrictions imposed by the branching structure of the amylopectin molecules and the chain lengths of the branches. Because the amount of amylopectin in most starches is greater than amylose, most of the crystallites formed during starch retrogradation are related to the association of amylopectin chains. Thus, amylopectin retrogradation proceeds slowly over several weeks of storage and contributes to the long term reheological and structural changes of starch systems (Lii et al. 1998; Zhou et al. 2002) The retrogradation kinetics of starch have received wide attention though the underlying mechanism of retrogradation has not been concluded. The Avrami equation is generally accepted as being able to simulate retrogradation kinetics. Lai et al. (2000) reported the retrogradation kinetics of pure amylopectin from 13 rice cultivars. Generally, the amylopectin systems showed two stages of retrogradation behavior during early ( 7 days) and late (> 7days) storage. Correlation analysis suggested that the kinetics of early stage retrogradation were more correlated than the late stage retrogradation with the number-average molecular weight and chain lengths of the amylopectin molecules. The proportion of short, long and extra long chain fractions appeared to have greater effects on the enthalpy

ß 2004, Woodhead Publishing Limited

Fig. 9.10 Tenperature sweep data for gelatinization of 25% TCW70 (a), waxy rice and TCS10 (b), normal rice with amylose content of 17.1% rice starches suspension, and comparison of G' between TCW70 and TCS10 (c). Symbols: G' (-s-), G'' (-
-) and tan (-*-).

ß 2004, Woodhead Publishing Limited

changes and late stage kinetics than the other structural factors (Lai et al. 2000). Tako and Hizukuri (2000) proposed some mechanisms for rice starch retrogradation which are based on the formation of hydrogen bonding at various molecular levels. It is assumed that intramolecular hydrogen bonding may take place between OH-6 and the adjacent hemiacetal oxygen atom of the D-glucosyl residues. Intermolecular hydrogen bonding may take place between OH-2 of the amylopectin and an adjacent O-6 of the amylose. Another intermolecular hydrogen bond may form between OH-2 of a D-glucose residue of the former molecule and O-6 of a D-glucose residue of a short side chain (A and B1) of the latter molecule. After saturation of intermolecular hydrogen bonding between amylose and amylopectin molecules, an intermolecular association may also take between amylopectin molecules through hydrogen bonding. The mechanism of retrogradation is complicated because retrogradation rate may vary from one cultivar to another due to differences in the proportion and interaction of amylopectin and amylose, chain length distribution, and molecular size of branched molecules (Hizukuri 1986; Eliasson and Gudmundsson 1996). Amylose content, GT, amylopectin structure in relation to rice starch retrogradation has been studied. Studies of the amylopectin staling of non-waxy starch gels with similar amylose content and GT have also been made. Storage modulus (G0 ) of 25% starch gels measured by small-strain oscillatory shear showed greater increase during staling for 17.5 hr at 8 oC in intermediate- to high-GT starches than in low-GT starches. Yao et al. (2002) indicated that amylose content after defatting and the peak value of amylopectin short-chain length has a significant positive association, while the amount of amylose-lipid complex has a negative relationship with the amylopectin crystallization rate constant. Qi et al. (2003) found that retrogradation of the solubilized native high- and low-GT amylopectin molecules confirmed that the high-GT amylopectin molecules form higher dissociation temperatures and enthalpy crystalline domains. This confirms that the higher proportion of short and especially relatively longer chains promotes crystalline formation. Vandeputte et al. (2003c) indicated that both amylopectin retrogradation and gel textural characteristics were related to absolute, free and lipid-complexed amylose contents and amylopectin chain length distributions. Other factors such as storage temperature and moisture content also have significant effects on the retrogradation properties of rice flour (Fan and Marks 1999).

9.6.1 Clarity Starch clarity is very important for many food applications. This starch property is thought to be dependent on the associative bonds between starch molecules in granules. The clarity of gelatinized starch suspensions varies among different rice cultivars and may be attributed to amylose content and granular size. Sodhi and Singh (2003) reported that the clarity of starch suspensions progressively decreased with the increase in storage duration up to three or four days, further

ß 2004, Woodhead Publishing Limited

storage caused less of a decrease in clarity. The decrease in clarity has been attributed to leached amylose and amylopectin chains that lead to the development of functional zones that scatter light (Perera and Hoover 1999). Differences in clarity may also be due to phosphate monoester derivatives and phospholipid content (Jane et al. 1996). The phosphate-monoester derivatives have been reported to increase paste clarity whereas addition of phospholipids made starch paste opaque (Kasemsuwan and Jane 1996).

9.6.2 Freeze-thaw stability There have been many reports from the food industry that waxy rice starch has substantial freeze-thaw stability, but this property has not been explored very much in the scientific literature. It has been reported that when non-waxy rice starch gel has undergone a freeze-thaw treatment, syneresis occurs due to increased molecular association between starch chains, at reduced temperature, thus excluding water from the gel structure. Waxy rice starch gel was reported to be more resistant to syneresis after a freeze-thaw cycle due to the formation of fewer inter-molecular associations.

9.6.3 Digestibility Starch that is resistant to digestion is thought to offer humans some protection from the development of various chronic diseases (Topping and Clifton 2001). There have been several reports that have linked amylose content to rice starch digestibility (Goddard et al. 1984, Noda et al. 2003). However, it has been found that similar hydrolysis of raw starch by amylase occurred with waxy and nonwaxy rice cultivars. Perez et al. (1991) reported that rice cultivars with similar amylose content varied in starch digestibility, the difference being associated with other properties, such as GT, cooking time, amylograph consistency, and volume expansion upon cooking. Therefore, amylose content alone is not a predictor of starch digestion rate (Perez et al. 1991). Zhang et al. (1996) showed that japonica waxy rice flour with different X-ray diffraction patterns differed in starch digestion, the percent digestion of group II flour was higher than group I flour, either with glucoamylase or with alpha-amylase. The authors suggested that the starch in group II had a looser structural arrangement, at the molecular level, than did those in group I. This allowed the enzymes to penetrate rapidly into the starch granules in group II. Perhaps, then, amylopectin structure may play a role in conferring the rate of rice starch digestion. Processing techniques are also reported to impact the rate of rice starch digestion. Parboiling reportedly decreases rice starch rate of digestion (Tetens et al. 1997). Rashmi and Urooj (2003) found that the steaming of rice created more resistant starch than boiling or pressure cooking. Storage under refrigeration also has been reported to slow the rate rice starch digestion (Frei et al. 2003).

ß 2004, Woodhead Publishing Limited

9.7 Improving rice starch functionality for food processing applications Although native starch has been widely used in food processing, the physical properties of these starches and the colloidal sols produced from them are limited in terms of commercial applications with which they are suited. For example, native rice starches are limited in their food applications due to their instability under various temperature, shear, and pH conditions. Because starch in its native form is relatively inert, it must be modified to achieve more desirable functional properties. Such treatments may be chemical, physical or genetic. The many different methods of chemically modify starch are the most widely used means of changing native starch functionality. 9.7.1 Chemical modification of rice starch Chemical modification of starch generally occurs via the introduction of functional groups that change starch properties and extend their applications in food systems. The aim of many chemical modifications is to modify cooking characteristics, decrease retrogradation, decrease gelling tendencies of pastes, increase freeze-thaw stability, decrease paste and/or gel syneresis, improve film formation, improve adhesion, or improve emulsion stability (see Chapter 12). As with native rice starch, the amount of modified rice starch produced annually is much less than that of corn and potato. Research that evaluated the effects of chemical modification of rice starch is discussed below. Acid thinned rice starch The modification of rice starch through acid hydrolysis allows the starch to be used at a higher solids concentration for quick gelling and provides gum or jelly with shorter texture and flexible properties. Wang and Wang (2001b) reported the structure and physicochemical properties of acid thinned rice starch. Acid reportedly attacked the amorphous regions within the starch granule and both amylose and amylopectin were hydrolyzed simultaneously. Acid modification decreased the longer chain fraction and increased the shorter chain fraction of rice starch. The authors suggested that the short-term development of gel structure by acid thinned starches was dependent on amylose content, whereas the long-term gel strength was due to the long amylopectin chains. Acetylated rice starch Low degree of substitution (DS) acetylated starch can be prepared by reacting starch with acetic anhydride in aqueous medium in the presence of dilute sodium hydroxide, while high DS acetylated starches can be prepared using pyridine. Jae et al. (1993) indicated that acetylation of rice starch increased solubility, swelling power, viscosity and the firmness, adhesiveness and cohesiveness of gels. This chemical process also decreased GT. Gonzalez and Perez (2002) reported that the acetylation of rice starch caused an increase in water absorption, swelling power and solubility, and lower initial pasting temperature,

ß 2004, Woodhead Publishing Limited

and setback. Rice starches with different amylose content can differ in their response to acetylation. For example, Liu et al. (2000) reported that acetylation increased the swelling power and freeze-thaw stability of non-waxy starch, but decreased these properties in waxy rice. Octenyl succinic anhydride modified starch Rice starch can be modified using 2-octenyl succinic anhydride. Octenyl succinic anhydride (OSA) modified rice starch reportedly increased paste viscosity, swelling volume, and reduced GT (Bao et al. 2003; Shih and Daigle 2003). The pasting viscosities of OSA modified starch were influenced by pH and salt content, the lower the pH or the higher the sodium chloride content the lower the pasting viscosity and increased pasting temperature (Shih and Daigle 2003). Higher DS has also been reported by Bao et al. (2003) to result in firmer OSA starch gels. Hydroxypropylated rice starch Low DS hydroxypropylated starch can be prepared by etherification of starch with propylene oxide, in water, under alkali conditions. The properties of the low-DS hydroxypropylated starches are similar to those of low-DS acetylated starches. Yeh and Yeh (1993) reported that hydroxypropylation of starch deceased its GT and shear stability, and increased its solubility in dimethyl sulfoxide and freeze-thaw stability. Yook et al. (1993) found that hydroxypropylation of rice starch caused a large decrease in GT and reduced the retrogradation rate of cooked rice starch. In rice starch treated by both hydroxypropylation and cross-linking, retrogradation was reduced more than for hydroxypropylated rice starch (Yook et al. 1993). Seow and Thevamalar (1993) found a progressive shift of biphasic gelatinization endotherms to lower temperatures, as well as a broadening and shortening of the endotherm with increasing molar substitution. The gelatinization enthalpy of the hydroxypropylated rice starch was also reduced. The authors suggested that the change in thermal properties resulted from an increase in internal plasticization and destabilization of amorphous regions. Phosphorylated rice starch Phosphorylated starch is produced through esterification of the hydroxyl groups with a phosphorylating agent such as sodium tripolyphosphate (STPP), sodium trimetaphosphate (STMP) or phosphorus oxychloride (POCl3). Reaction of two starch molecules with STMP or POCl3 produces cross linked starch (di-starch phosphate ester, see below). Rice starch phosphates can also be prepared by extrusion of rice starch with phosphate salts (Singh et al. 1999). Increased barrel temperature (120 to 180 oC) resulted in greater phosphorus incorporation into the starch. Pastes produced from the starch phosphate esters were clear, of high viscosity and consistency, had acceptable freeze-thaw stability and resistance to retrogradation. Sitohy et al. (2000) found that phosphorylation of rice starch at low DS resulted in greater solubility, swelling power, and paste viscosity and

ß 2004, Woodhead Publishing Limited

clarity. It was concluded that a low degree of phosphorylation can improve clarity and hence quality of starch paste (Sitohy et al. 2000). Sitohy and Ramadan (2001a) reported that starch granule size impacted the effects of phosphorylation on solubility, swelling and paste clarity. Phosphorylated rice starch has also been reported to experience a reduced degree of hydrolysis after acid or alpha-amylase hydrolysis (Sitohy and Ramadan 2001b). Cross-linked rice starch Rice starch can be cross linked with STPP, STMP, POCl3, and epichlorhydrin. Yeh and Yeh (1993) reported that rice starch cross-linked with POCl3 had greater
H and shear stability, and reduced solubility in dimethyl sulfoxide and freeze-thaw stability. Yook et al. (1993) reported a lower
H in cross-linked rice starch, compared to native starch, due to a reduced portion of starch able to be gelatinized. By way of explanation of these apparently conflicting conclusions, Liu et al. (2000) suggested that the effects of cross-linking differ in waxy versus non-waxy starch. Specifically, cross-linking increased shear stability and decreased swelling power and solubility of non-waxy and waxy starch but it increased the pasting temperature and
H of waxy starch and resulted in the opposite effect in non-waxy starch. These authors also studied the freeze-thaw stability of cross-linked starch and found a decrease in retrogradation of non-waxy starch, but an increase in waxy starch. Similar results were obtained for non-waxy starch by Chatakanonda et al. (2000), who also reported that they observed no effect of cross-linking on the glass transition temperature of starch gels. The melting endotherms of non-waxy rice starch shifted to a higher temperature with an increasing degree of cross-linking, while there was no dramatic change in enthalpy.

9.7.2 Physical modification of rice starch Physical modification of starch involves the simultaneous action of several conditions such as temperature, pressure, shear and moisture on starch. The starch granules can either be conserved or completely disorganized. Hydrothermal treatment Annealing and heat-moisture treatments are hydrothermal treatments that modify the physicochemical properties of starch, without destroying its granular structure. These processes involve incubation of starch at excess or intermediate water content (heat-moisture) or at low moisture levels (annealing), at a temperature above the glass transition temperature but below the GT (Jacobs and Delcour 1998). Heat-moisture treatments of rice starch have been reported to increase GT and decrease solubility (Lu et al. 1996, Kim and Shin 1990) and transition enthalpy (Lu et al. 1996). Kim and Shin (1990) reported that the swelling power of heat-moisture treated rice starch decreased. Heat-moisture treatment of rice starch was reported by Lu et al. (1994) to have decreased the swelling power and pasting viscosity of non-waxy rice starch, but caused the

ß 2004, Woodhead Publishing Limited

opposite effect in waxy rice starch. Anderson et al. (2002) reported that all viscosity parameters decreased after heat-moisture treatment in a conventional oven for both non-waxy and waxy rice starches. These authors also used a DSC to perform a heat-moisture treatment and found a reduced rate of amylase digestibility. Jacobs et al. (1995) reported that annealing rice starch increased its GT, enthalpy, peak viscosity, and cool paste viscosity, but narrowed the GT range. This treatment also decreased the amount of soluble amylose. Extrusion Extrusion cooking, because of its low cost and continuous processing capability, is a popular means of modifying the functional characteristics of cereal grains. This high temperature and short time processing technique is capable of thoroughly gelatinizing the starch even at low moisture levels. Extrusion does not change the total starch content or degrade the starch moiety to low molecular weight sugars; however, it does cause some macromolecular degradation. After extrusion, it has been reported that the viscosity profiles for non-waxy rice and its parboiled counterpart, initially increased in viscosity, then decreased (Bryant et al. 2001). Extruded waxy flour had a low initial viscosity, which greatly increased upon cooling. The flours extruded at 100 oC digested at a slower rate compared to those extruded at 125 or 150 oC. Sonication Ultrasonication has many applications in the food industry. Chung et al. (2002) studied the effects of ultrasonic waves on the functional properties of rice starch. They found that the apparent and inherent viscosities of rice starch decreased after sonication. The starch paste became more transparent and the viscosity measured at 70 oC showed a marked decrease after sonication. The authors indicated that the functional property changes resulted from the disruption of swollen granules rather than the breakage of glucosidic linkages. Sonication is thus able to create a modified rice starch that has good clarity and low viscosity without decreased chain length. Gamma irradiation Gamma irradiation is used to protect foods from insect infestation and microbial contamination during storage. The effects of gamma irradiation on rice starch properties has been studied and found to cause dramatically reduced pasting viscosity, while the amylose content and GT was only slightly reduced (Bao et al. 2001, Bao and Corke 2002).

9.7.3 Genetic modification of rice starch Across the world traditional breeding methods produce rice cultivars with different combinations of amylose and amylopectin content and structure. Also, natural rice mutants have been found and other mutants created using mutation breeding techniques. Several of these have starch properties very

ß 2004, Woodhead Publishing Limited

different from native genotypes. For example, the following rice mutants have unique starch properties, amylose extender, sugary and dull (Satoh and Omura 1981, Mizuno et al. 1993, Nakamura et al. 1996). The tools for genetic engineering offer the means to create rice varieties with novel starch functional properties suited for use in new or improved food products. Currently the bioengineering of genes related to rice starch properties have been performed primarily to better understand aspects of starch synthesis. Some examples are as follows. Terada et al. (2000) transformed rice plants with a 2.3 kb Wx cDNA, several transformants with various degrees of amylose reduction were created, as well as were two lines with no amylose. Perhaps it is also possible to create rice with very high amylose content and as a consequence affect human blood glucose levels at a very slow rate. A retrotransposon inserted into the gene encoding starch synthase I resulted in amylopectin with more chains of 6 and 7 DP and less 8±12 DP chains (Nakamura 2002). As time goes on and our tools develop, all of the genes involved in starch synthesis will likely be targets for genetic engineering.

9.8

Future trends

Rice starch offers the food industry functional advantages and financial disadvantages. Rice is a source of starch that is hypoallergenic for many, bland in taste, white in color and as a gel is smooth in texture. In contrast, for example, corn starch is yellowish white, has a `protein' flavor and forms a firm gel. Thus, in terms of end-use quality characteristics rice offers the food processing industry some advantages compared to other sources of starch. However, using rice starch in food to perform a specific function comes at a greater cost than if another starch source were used. Rice is a more costly crop to produce compared to other sources of starch. It also has no well developed markets for its co-products. As a consequence of future research it is possible that the cost of rice production will decrease compared to other sources of starch, or the price of other starches will increase, but these things are not likely. What is more probable is that the cost of producing rice starch will decrease with the development of new processing technology. It is also possible that the co-products of rice, namely, flour protein, bran, rice bran oil and rice bran's health beneficial components will bring rice producers and millers added value in the future. Also, as the demand for novel products and product quality increases we may see an increased demand for rice starch due to its unique properties inspite of its cost. What will happen in the future regarding the demand for rice starch is not known. However, what is sure is that many, many factors are involved in controlling the future production and use of rice starch. Rice germplasm contains a great deal of genetic diversity in terms of enduse quality properties. This variation has not been fully explored and thus it has not been capitalized on for product development purposes. The market

ß 2004, Woodhead Publishing Limited

today has access to rice starch from waxy cultivars and low GT and intermediate GT amylose types. As a consequence much of the food industry thinks that these are the only types of rice that exist. An increased awareness of the various end-use quality types that currently exist today as cultivars and the potential for creating new cultivars with novel properties could result in an increased utilization of rice starch. Also, the potential application of starches from the amylose extender and sugary mutants has not been examined. Nor has the potential use of genetic engineering to develop rice with novel starch properties even been fully discussed. Clearly, the waxy gene which is responsible for amylose synthesis will receive a great deal of future research attention, as will the genes involved in amylopectin synthesis such as those coding for starch branching enzymes, starch synthases, debranching enzymes and disproportionating enzyme. Alone or in combination, the results of such experiments will have the potential to create rice starch properties never before known and likely to become in great demand. The food processing industry frequently clamors for starch functionality stability across growing environments and post-harvest handling conditions. Achieving such a thing would require rice geneticists, breeders and chemists to work together to find, or create, starch synthesis genes, or their regulators, that are not sensitive to the environment. With such cultivars available, food processors would substantially decrease their need to chemically evaluate incoming rice flour and starch, and would no longer need to constantly modify the parameters of their processing equipment.

9.9

Sources of further information and advice

9.9.1 Organizations Asia Rice Foundation http://www.asiarice.org FLAR ± The Latin American Fund for Irrigated Rice http://www.flar.org/ ingles.htm IRRI ± International Rice Research Institute http://www.irri.org WARDA ± The Africa Rice Center http://www.warda.cgiar.org

9.9.2 Industries A & B Ingredients http://www.abingredients.com California Natural Products http://www.californianatural.com Japan Rice Market http://www.japan-rice.com Oryza http://www.oryza.com Remy http://www.remy-industries.be Rice Growers' Association of Australia http://www.rga.org.au Sage V http://www.sagevfoods.com Zumbro (division of Primera Foods) http://www.primerafoods.com

ß 2004, Woodhead Publishing Limited

9.9.3 Literature Champagne E (ed) 2004. Third Edition. Rice Chemistry and Technology. American Association of Cereal Chemists St. Paul, USA. pp. 656. (ISBN 1891127349) http://www.aaccnet.org Owen S (2003) The Rice Book. Frances Lincoln, LTD. London England pp. 404. (ISBN 0711222606) http://www.franceslincoln.com Alford J and Duguid N (1998) Seductions of Rice. Artisan Publishers, Muskogee OK, USA. pp. 480 (ISBN 1579651135) http://www.artisanpublishers.com International Rice Research Notes (IRRN). Available for free on the internet in PDF format. http://www.irri.org/publications/irrn/index.asp

9.10 References (2002) Digestibility and pasting properties of rice starch heat-moisture treated at the melting temperature (Tm). Starch 54: 401±409. BAO JS, CORKE H (2002) The pasting properties of gamma-irradiated rice starches as affected by pH. J Agri Food Chem., 50: 336±341. BAO JS, HE P, XIA YW, CHEN Y, ZHU LH (1999) Starch RVA profile parameters of rice are mainly controlled by Wx gene. Chinese Science Bulletin, 44: 2047±2051. BAO JS, SHU QY, XIA YW, BERGMAN C, MCCLUNG A. (2001) Effects of gamma irradiation on aspects of milled rice (Oryza sativa) end-use quality. J Food Quality, 24: 327±336. BAO JS, XING J, PHILLIPS DL, CORKE H (2003) Physical properties of octenyl succinic anhydride modified rice, wheat and potato starches. J Agric Food Chem 51: 2283± 2287. BHATTACHARYA M, ZEE SY, CORKE H (1999) Physiochemical properties related to quality of rice noodles. Cereal Chemistry 76: 861±867. BRYANT RJ, KADAN RS, CHAMPAGNE ET, VINYARD BT, BOYKIN D (2001) Functional and digestive characteristics of extruded rice flour. Cereal Chem 78: 131±137. CHAMPAGNE ET (1996) Rice starch composition and characteristics. Cereal Foods World 41: 833±838. CHATAKANONDA P, VARAVINIT S, CHINACHOTI P (2000) Relationship of gelatinization and recrystllization of cross-linked rice to glass transition temperature. Cereal Chemistry 77: 315±319. CHEN MH, BERGMAN CJ (2004) Analysis of rice starch structure using HPLC-GPC with MALLS-RI. Carbohydrate Polymers (submitted). CHEN MH, BERGMAN CJ, FJELLSTROM RG. (2003) Genetic variation at the waxy locus associated with amylose content in international rice germplasm. Plant and Animal Genome Conference, San Diego, CA USA, January 11±15. CHIANG PY, YEH AI (2002) Effect of soaking on wet-milling of rice. Journal of Cereal Science 35: 85±94. CHIOU H, MARTIN M, FITZGERALD M (2002) Effect of purification methods on rice starch structure. Starch, 54: 415±420. CHUNG KM, MOON TW, KIM H, CHUN JK (2002) Physicochemical properties of sonicated mung bean, potato, and rice starches. Cereal Chem 79: 631±633. COOKE D, GIDLEY MJ (1992) Loss of crystalline and molecular order during starch ANDERSON AK, GURAYA HS, JAMES C, SALVAGGIO L

ß 2004, Woodhead Publishing Limited

gelatinization. Origin of the enthalpic transition. Carbohydrate Research 227: 103±112. CORNELL HJ, MCGRANE SJ, RIX, CJ (1999) A novel and rapid method for the partial fractionation of starch using 1-butanol in the presence of thiocyanate. Starch 51: 335±341. DEIS, RC (1997) Functional ingredients from rice. Food Product Design. January 1±8. ELIASSON AC, GUDMUNDSSON M (1996) Starch: physicochemical and functional aspects. Pages 431±503 in: Carbohydrates in Foods. AC Eliasson, ed. Marcel Dekker: New York. FAN J, MARKS BP (1999) Effects of rough rice storage conditions on gelatinization and retrogradation properties of rice flours. Cereal Chem 76: 894±897. FAN J, MARKS BP, DANIELS MJ, SIEBENMORGEN TJ (1999) Effects of postharvest operations on the gelatinization and retrogradation properties of long-grain rice. Trans ASAE 42: 727±731. FISHMAN ML, RODRIGUEZ L, CHAU HK (1996) Molar masses and sizes of starches by highperformance size exclusion chromatography with on-line multiangle laser light scattering detection. Journal of Agriculture and Food Chemistry, 44: 3182±3188. FITZGERALD MA, MARTIN M, WARD RM, PARK WD, SHEAD HJ (2003) Viscosity of rice flour: a rheological and biological study. J Agric Food Chem, 51: 2295±2299. GODDARD M, YOUNG G, MARCUS R (1984) The effect of amylase content on insulin and glucose responses to ingested rice. Am J Clin Nutr 39: 388±392. GONZALEZ Z, PEREZ E (2002) Effect of acetylation on some properties of rice starch. Starch. 54: 148±154. HAN XZ, HAMAKER BR (2001) Amylopectin fine structure and rice starch paste breakdown. Journal of Cereal Science 34: 279±284. HIZUKURI S (1986) Polymodal distribution of the chain lengths of amylopectin and its significance. Carbohydrate Research 147: 342±347. HIZUKURI S (1993) Towards an understanding of the fine structures of starch molecules. Denpun Kagaku 40: 133±147. HOGAN JT (1967) The manufacture of rice starch. Starch Chem Technol 2: 65±78. JACOBS H, DELCOUR JA (1998) Hydrothermal modification of granular starch, with retention of the granular structure: a review. J Agric Food Chem 46: (2895±2905). JACOBS H, EERLINGEN RC, CLAUWAERT W, DELCOUR JA (1995) Influence of annealing on the pasting properties of starches from varying botanical sources. Cereal Chemistry 72: 480±487. JAE HJ, JUNG SB, MAN JO (1993) Physicochemical properties of acetylated rice starches. Korean Journal of Food Science and Technology 25: 123±129. JANE J, KASEMSUWAN T, CHEN LF (1996) Phosphorus in rice and other starches. Cereal Food World, 41: 827±832. JANE J, CHEN YY, LEE LF, MCPHERSON A, WONG KS, RADOSAVLJEVIC M, KASEMSUWAN T

(1999) Effect of amylopectin branch chain length and amylose content on the gelatinization and pasting properties of starch. Cereal Chemistry 76: 629±637. JULIANO BO (1998) Varietal impact on rice quality. Cereal Food World 43: 207±222. JULIANO BO, VILLAREAL RM (1987) Varietal differences in physicochemical properties of waxy rice starch. Starch 39: 298±301. KARIM AA, NORZIAH MH, SEOW CC (2000) Methods for the study of starch retrogradation. Food Chemistry 71: 9±36. KASEMSUWAN T, JANE L (1996) A quantitative method for the survey of starch phosphate derivatives and starch phospholipids by 31P nuclear magnetic resonance

ß 2004, Woodhead Publishing Limited

spectroscopy. Cereal Chemistry 73: 702±707. (2003) Analysis of food composition data on rice from a plant genetic resources perspective. Food Chem 80: 589±596. KIM SK, SHIN MS (1990) Modification of physicochemical properties of rice starch by heatmoisture treatment. Journal of the Korean Agricultural Chemical Society. 33: 1±7. LAI HM (2001) Effects of hydrothermal treatment on the physiochemical properties of pregelatinized rice flour. Food Chem 72: 455±463. LAI VMF, LU S, LII CY (2000) Molecular characteristics influencing retrogradation kinetics of rice amylopectin. Cereal Chem. 77: 272±278. LII CY, SHAO YY, TSENG KH (1995) Gelation mechanism and rheological properties of rice starch. Cereal Chem 72: 393±400. LII CY, TSAI ML, TSENG KH (1996a) Effect of amylose content on the rheological property of rice starch. Cereal Chem 73: 415±420. LII CY, LAI MF, TSAI ML (1996b) Studies on starch gelatinization and retrogradation with dynamic rheometry- the influence of starch granular structure and composition. Zywnosc Technologia Jakosc. (Poland) 2: 27±53. LII CY, LAI VMF, LU S, TSAI ML (1998) Correlation between the physicl property, eating quality and the molecular structure of rice-starchy systems. Zywnosc Technologia Jakosc (Poland) 4: 72±86. LIM ST, KASEMSUWAN T, JANE J (1994) Characterization of phosphorus in starch by 31Pnuclear magnetic resonance spectroscopy. Cereal Chem 71: 488. LITTLE RR, HILDER GB, DAWSON EH (1958) Differential effect of dilute alkali on 25 varieties of milled white rice. Cereal Chem 35: 115±126. LIU HJ, RAMSDEN L, CORKE H (2000) Physical properties of cross-linked and acetylated normal and waxy rice starch. Starch. 51: 249±252. LU S, FANG WT, LII CY (1994) Studies of the effects of different hydrothermal treatments on the physicochemical properties of nonwaxy and waxy rices. Journal of the Chinese Agricultural Society. 32: 372±383. LU S, CHEN CY, LII CY (1996) Gel Chromatography fractionation and thermal characterization of rice starch affected by hydrothermal treatment. Cereal Chem 73: 5±11. LU S, CHEN LN, LII CY (1997) Correlation between the fine structure, physicochemical properties, and retrogradation of amylopectins from Taiwan rice varieties. Cereal Chem 74: 34±39. LUMDUBWONG N, SEIB PA (2000) Rice starch isolation by alkaline protease digestion of wet-milled rice flour. J. Cereal Sci 31: 63±74. MARTIN M, FITZGERALD MA (2002) Proteins in rice grains influence cooking properties. Journal of Cereal Science 36: 285±294. KENNEDY G, BURLINGAME B

MIZUNO K, KAWASAKI T, SHIMADA H, SATOH H, KOBAYASHI E, OKUMURA S, ARAI Y, BABA T

(1993) Alterationof the structure of starch components by the lack of an isoform of starch branching enzyme in rice seeds. J Biol Chem 268: 19084±19091. MORRISON WR, AZUDIN MN (1987) Variation in the amylose and lipid contents and some physical properties of rice starches. J Cereal Sci 5: 35±39. MORRISON WR, MILLIGAN TP, AZUDIN MN (1984) A relationship between the amylose and lipid contents of starches from diploid cereals. J. Cereal Sci 2: 257±260. MURUGESAN G, HIZUKURI S, FUKUDA M, JULIANO BO (1992) Structure and properties of waxy-rice (IR29) starch during development of the grain. Carbohydrate Res. 223: 235±242. NAKAMURA Y (2002) Towards a better understanding of the metabolic system for

ß 2004, Woodhead Publishing Limited

amylopectin biosynthesis in plants: Rice endosperm as a model tissue. Plant Cell Physiology 43: 718±725. NAKAMURA Y, UMEMOTO T, TAKAHATA Y, KOMAE K, AMANO E, SATOH H (1996) Changes in structure of starch and enzyme activities affected by sugary mutations in developing rice endosperm. Possible role of starch debranching enzyme (Renzyme) in amylopectin biosynthesis. Physiol Plant 97: 491±498. NAKAMURA Y, SAKURAI A, INABA Y, KIMURA K, IWASAWA N, NAGAMINE T (2002) The fine structure of amylopectin in endosperm from Asian cultivated rice can be largely classified into two classes. Starch 54: 117±131. NODA T, TAKAHATA Y, SATO T, IKOMA H, MOCHIDA H (1996) Physicochemical properties of starches from purple and orange fleshed sweet potato roots at two levels of fertilizer. Starch, 48: 395±399. NODA T, NISHIBA Y, SATO T, SUDA I (2003) Properties of starches from several low-amylose rice cultivars. Cereal Chem 80: 193±197. NORMAND FL, MARSHALL WE (1989) Differential scanning calorimetry of whole grain milled rice and milled rice flour. Cereal Chem 66: 317±320. ONG MH, BLANSHARD JMV (1995) Texture determinants in cooked, parboiled rice. I: rice starch amylose and the fine structure of amylopectin. Journal of Cereal Science 21: 251±260. PERERA C, HOOVER R (1999) Influence of hydroxypropylation on retrogradation properties of native, defatted and heat moisture treated potato starches. Food Chemistry 64: 361±375. PEREZ CM, THOMPSON LU, JULIANO BO (1991) Rice varieties with similar amylose content differ in starch digestibility and glycemic response in humans. Am J Clin Nutr 54: 871±877. QI X, TESTER RF, SNAPE CE, ANSELL R (2003) Molecular basis of the gelatinization and swelling characteristics of waxy rice starches grown in the same location during the same season. J Cereal Sci 37: 363±376. RAMESH M, MITCHELL JR, JUMEL K, HARDING SE (1999) Amylose content of rice starch. Starch, 51: 311±313. RASHMI S, UROOJ A (2003) Effect of processing on nutritionally important starch fractions in rice varieties. Int J Food Sci Nutr 54: 27±36. REDDY KR, ALI SZ, BHATTACHARYA KR (1993) The fine structure of rice starch amylopectin and its relation to the texture of cooked rice. Carbohydrate Polymer 22: 267±275. SATOH H, OMURA T (1981) New endosperm mutations induced by chemical mutagens in rice, Oryza sativa L. Japn J Breed 31: 316±326. SEOW CC, THEVAMALAR K (1993) Internal plasticization of granular rice starch by hydroxypropylation: effects of phase transitions associated with gelatinization. Starch 45: 85±88. SHENG DY (1995) Rice-based ingredients in cereals and snacks. Cereal Food World 40: 538±540. SHIH FF, DAIGLE KW (2003) Gelatinization and pasting properties of rice starch modified with 2-octen-1-ylsuccinic anhydride. Nahrung/Food 47: 64±67. SIMPSON JE, ADAIR CR, KOHLER G.O, DAWSON EH, DEOBALD GJ, KESTER EB, HOGAN JT, BATCHER OM AND HALICK JV (1965) Quality evaluation studies of foreign and domestic rices. Tech. Bull. 1331. Agric. Res. Serv. U.S. Dept. Agric., Washington, DC p. 186. SINGH N, KAUR K, SINGH B, SEKHON KS (1999) Effects of phosphate salts on extrusion behaviour of rice. Food Chem 64: 481±488.

ß 2004, Woodhead Publishing Limited

(2000) Thermal and physicochemical properties of rice grain, flour and starch. J. Agri Food Chem 48: 2639±2647. SITOHY MZ, RAMADAN MF (2001a) Granular properties of different starch phosphate monoesters. Starch 53: 27±34. SITOHY MZ, RAMADAN MF (2001b) Degradability of different phosphorylated starches and thermoplastic films prepared from corn starch phosphomonoesters. Starch 53: 317±322. SITOHY MZ, EL-SAADANY SS, LABIB SM, RAMADAN MF (2000) Physicochemical properties of different types of starch phosphate monoesters. Starch 52: 101±105. SLADE L, LEVINE H (1988) Non-equilibrium melting of native granular starch. Part I. Temperature location of the glass transition associated with gelatinization of Atype cereal starches. Carbohydrate Polymers 8: 183±208. SODHI NS, SINGH N (2003) Morphological, thermal and rheological properties of starches separated from rice cultivars grown in India. Food Chem 80: 99±108. SUZUKI Y, SANO Y, ISHIKAWA T, SASAKI T, MATSUKURA U, HIRANO H (2003) Starch characteristics of the rice mutant du2-2 Taichung 65 highly affected by environmental temperatures during seed development. Cereal Chem 80: 184±187. TABATA S, NAGATA K, HIZUKURI S (1975) Studies on starch phosphates. Part 3. On the esterified phosphates in some cereal starches. Starch 27: 333. TAKEDA Y, HIZUKURI, S, JULIANO, BO (1986) Purification and structure of amylase from rice starch. Carbohydrate Research 148: 299±308. TAKO M, HIZUKURI S (2000) Retrogradation mechanism of rice starch. Cereal Chem 77: 473±477. TANG SX, KHUSH GS, JULIANO BO (1989) Variation and correlation of four cooking and eating quality indices of rice. Philipp J. Crop Sci. 14: 45±49. TEO CH, KARIM AA, CHEAH PB, NORZIAH MH, SEOW CC (2000) On the roles of protein and starch in the ageing of non-waxy rice flour. Food Chemistry, 69: 229±236. TERADA R, NAKAJIMA M, ISSHIKI M, OKAGAKI RJ, WESSLER SR, SHIMAMOTA K (2000) Antisense Waxy genes with highly active promoters effectively suppress Waxy gene expression in transgenic rice. Plant Cell Physiol. 41: 881±888. TESTER RF (1997) Starch: the polysaccharide fractions. In Frazier PJ, Richmond P and Donald AM (eds) Starch: structure and functionality, pp. 163±171, Royal Society of Chemistry. TESTER RF, MORRISON WR (1990a) Swelling and gelatinization of cereal starches. I. Effects of amylopectin, amylose and lipids. Cereal Chem 67: 551±557. TESTER RF, MORRISON WR (1990b) Swelling and gelatinizationof cereal starches. II. Waxy rice starches. Cereal Chem 67: 558±563. TETENS I, BISWAS SK, GLITO LV, KABIR KA, THILSTED SH, CHOUDHURY NH (1997) Physicochemical characteristics as indicators of starch availability from milled rice. Journal of Cereal Science 26: 355±361. TOPPING DL, CLIFTON PM (2001) Short-chain fatty acids and human colonic function: roles of resistant starch and nonstarch polysaccharides. Physiol Rev 81: 1031±1064. UMEMOTO T, NAKAMURA Y, SATOH H, TERASHIMA K (1999) Differences in amylopectin structure between two rice varieties in relation to the effects of temperature during grain-filling. Starch 51: 58±62. VANDEPUTTE GE, VERMEYLEN R, GEEROMS J, DELCOUR JA (2003a) Rice starches, I. Structural aspects provide insight into crystallinity characteristics and gelatinisation behaviour of granular starch. J Cereal Sci 38: 43±52. SINGH V, OKADOME H, TOYOSHIMA H, ISOBE S, OHTSUBO K

ß 2004, Woodhead Publishing Limited

(2003b) Rice starches, II. Structural aspects provide insight in swelling and pasting properties. J Cereal Sci 38: 53±59. VANDEPUTTE GE, VERMEYLEN R, GEEROMS J, DELCOUR JA (2003c) Rice starches, III. Structural aspects provide insight in amylopectin retrogradation properties and gel texture. J Cereal Sci 38: 61±68. VILLAREAL CP, JULIANO BO, HIZUKURI S (1993) Varietal differences in amylopectin staling of cooked waxy milled rices. Cereal Chem 70: 753±758. WANG LF, WANG YJ (2001a) Comparison of protease digestion at neutral pH with alkaline steeping method for rice starch isolation. Cereal Chem 78: 690±692. WANG LF, WANG YJ (2001b) Structures and physicochemical properties of acid-thinned corn, potato and rice starches. Starch 53: 570±576. WANG YJ, WANG LF (2002) Structures of four waxy rice starches in relation to thermal, pasting and textural properties. Cereal Chem 79: 252±256. WANG YJ, WANG LF, SHEPHARD D, WANG FD, PATINDOL J (2002) Properties and structures of flours and starches from whole, broken, and yellowed rice kernels in a model study. Cereal Chem 79: 383±386. WONG KS, JANE J (1997) Quantitative analysis of debranched amylopectin by HPAECPAD with a post-column enzyme reactor. J. Liq Chrom. 20: 297±310. WONG KS, KUBO A, JANE JL, HARADA K, SATOH H, NAKAMURA Y (2003) Structures and properties of amylopectin and phytoglycogen in the endosperm of sugary-1 mutant of rice. J Cereal Sci 37: 139±149. YANO M, OKUNO K, KAWAKAMI J, SATOH H, OMURA T (1985) High amylose mutants of rice, Oryza sativa L. Theor Appl Genet 69: 253±257. YAO Y, ZHANG JM, DING XL (2002) Structure-retrogradation relationship of rice starch in purified starches and cooked rice grains: a statistical investigation. J Agri Food Chem 50: 7420±7425. YEH AI, YEH SL (1993) Some characteristics of hydroxypropylated and cross-linked rice starch. Cereal Chem. 70: 596±601. YOO SH, JANE JL (2002) Molecular weights and gyration radii of amylopectins determined by high-performance size-exclusion chromatography equipped with multi-angle laser-light scattering and refractive index detectors. Carbohydrate Polymer 49: 307±414. YOOK C, PEK UH, PARK KH (1993) Gelatinization and retrogradation characteristics of hydroxypropylated and cross-linked rices. J Food Sci 58: 405±407. ZHANG Q, ABE T, TAKAHASHI T, SASAHARA T (1996) Variations in in vitro starch digestion of glutinous rice flour. J Agri Food Chem. 44: 2672±2674. ZHOU ZK, ROBARDS K, HELLIWELL S, BLANCHARD C (2002) Composition and functional properties of rice. International Journal of Food Science and Technology. 37: 849± 868. ZHOU ZK, ROBARDS K, HELLIWELL S, BLANCHARD C (2003) Effect of rice storage on pasting properties of rice flour. Food Research International 36: 625±634. VANDEPUTTE GE, DERYCKE V, GEEROMS J, DELCOUR JA

ß 2004, Woodhead Publishing Limited

10 New corn starches P. J. White and A. Tziotis, Iowa State University, USA

10.1 Introduction: the use of corn starch in food processing Corn provides a high-quality starch used widely in the food industry in many applications requiring particular viscosities and textures. Starch is a carbohydrate consisting of two distinct molecules; amylopectin, with anhydroglucose units linked to create a highly branched molecule; and amylose, a primarily linear glucose molecule. Starch functionality depends greatly on the molecular weight, size, and structure of its components, amylose and amylopectin, which, differing greatly in molecular-weight distribution and molecular structures, display different pasting, retrogradation, viscoelastic, and rheological properties (Morrison and Tester, 1991, Bahnassey and Breene, 1994, White et al., 1989). Normal corn starch is made up of about 25% amylose and 75% amylopectin. Amylopectin is the major component of most starches, and its fine structure plays a critical role in the characteristics of starch. Amylopectin (AP), the highly branched component of starch, consists of chains of -D-glucopyranosyl residues linked together mainly by 1,4 linkages, with 5 to 6% of 1,6 bonds forming branch points. Within amylopectin, the crystalline component is composed of parallel arrays of linear chains packed tightly in double helices (Thompson, 2000). Amylose (AM) has been defined as a linear molecule of 1,4 linked -D-glucopyranosyl units, but it is now well established that the linear molecules are slightly branched by -1,6-linkages. Amylose is the smaller of the two fractions (105±106 Da; degree of polymerization (DP) 500±5000) and possesses very few branches, 9±20 per molecule, with chain-lengths (CL) of between 4 and 100 glucose units and greater. Differences among corn starches in granule swelling (onset of viscosity), peak temperature, peak viscosity, shear

ß 2004, Woodhead Publishing Limited

thinning during pasting, and gel firmness during storage, have been mostly attributed to differences in amylopectin structure (Ring and Stainby, 1985, Doublier et al., 1987a, Bahnassey and Breene, 1994), whereas differences in setback and final viscosity during pasting have been attributed to amylose structure (Ott and Hester, 1965, Leloup et al., 1991, Vasanthan and Hoover, 1992).

10.1.1 Intermediate materials in corn starch Differences in granule swelling (onset of viscosity), peak temperature of gelatinization, peak viscosity, shear thinning, and firmness of gel during storage among starches have been mostly attributed to AP. Meanwhile, differences in setback and final viscosity during pasting have been attributed to AM. Lansky et al. (1949) proposed that a third component in normal corn starch exists, called intermediate material (IM), with properties different from those of AM and AP. This component could also play an important role in determining the functional properties of starch. The presence of a great number of short branch chains in this component could contribute to lower granular crystallinity, gelatinization temperature, enthalpy change, viscosity, and degree of retrogradation; and greater degree of digestibility by enzymatic hydrolysis, On the other hand, molecules with longer branch CL and a lesser degree of branching would contribute to greater crystallinity, gelatinization temperature enthalpy change, degree of retrogradation, viscosity and gel firmness (Campbell et al., 1995, Perera et al., 2001). Based on indirect evidence from iodine affinities, Lansky et al. (1949) suggested that 5 to 7 % of normal corn starch consists of material intermediate between the strictly linear and highly branched fractions. Subsequently, several types of branched polysaccharides have been recovered by various modifications of the previously described fractionation procedures. Erlander et al. (1965) recovered a low-molecular-weight component from the supernatant following AM precipitation with thymol and removal of AP by centrifugation. The polysaccharide remaining in the supernatant had a -amylolysis limit and degree of branching similar to that of AP. Perlin (1958) obtained an intermediate component following removal of AP by centrifugation and precipitation of AM with amyl alcohol. The polysaccharide remaining in the supernatant was more highly branched than AP, based on reduced -amylolysis limits, and was of lower molecular weight. With regard to starch structure in other plant species, a related highly branched polysaccharide with viscosity similar to AP was recovered from the supernatant following recomplexing of the AM fraction of starch from potato tuber, rubber seed, barley kernels, and oat kernels. A polysaccharide with a lower degree of branching than AP, but with greater average CLs and higher -amylolysis limits, was recovered from rye and wheat starches (Banks and Greenwood, 1967) and also from normal corn starch (Whistler and Doane, 1961). Another polysaccharide reported in small amounts in starch of corn (Adkins and Greenwood, 1969) is short-chain-length AM. In

ß 2004, Woodhead Publishing Limited

normal corn starch, this linear polysaccharide has an average CL of 58 (Adkins and Greenwood, 1969). Banks and Greenwood (1975) suggested that the type and amount of IM in corn starch depended primarily on the AM percentage of the starch, although it could significantly vary among different starches. Variations in the amounts and structures of AP, AM and IM can result in starch granules with very different physicochemical and functional properties that can have a major impact on the utilization of these starches in food products (Kobayashi et al., 1986, Yuan et al., 1993). Baba and Arai (1984) found a longer average chainlength from AP and IM of amylomaize than from the AP of waxy corn. The IM from 50% amylomaize had an average degree of polymerization of 250± 300 glucose units per molecule, with four or five branches having a CL of about 50 glucose units. Wang et al. (1993) reported that about 15% of the starch in dull (du1) mutant endosperms was composed of IM, distinguished from the other components by the properties of its starch-iodine complex. Starch from du1 mutants had the highest degree of branching among a wide variety of normal and mutant kernels analyzed (Inouchi et al., 1987, Wang et al., 1993). Also, the degree of branching of AP and IM decreased when the amylose extender (ae) gene was present. The ae gene has a greater impact on the degree of branching of the IM than of the AP. The greater iodine affinity of ae IM than of AP indicated that the IM had a longer CL than did AP. On the other hand, the presence of more highly branched molecules was indicated in du1 starch. In addition, the degree of branching for IM was less than that of AP of the same starch (Wang et al., 1993). The gel-permeation chromatograms of corn starch from the dominant amylose extender (ae) mutant showed greater proportions of AM and IM than from normal corn starch (Kasemsuwan et al., 1995). Both AP and IM had a similar molecular structure, except the IM had more, shorter chains than did AP, and AP had more chains of DP 16 to 30 than did the IM.

10.1.2 Role of the environment on corn starch characteristics Corn starch and its structural features can be affected by environmental factors during the development of plants, even though the impact of the environment has not been reported to be as severe as that associated with other plant species and varieties. White et al. (1991) found differences in the differential scanning calorimetry (DSC) profiles of genotypes grown in temperate (warmer growing temperatures) versus tropical locations (cooler growing temperatures). The planting location affected the peak temperature of gelatinization and the gelatinization enthalpy change, which increased with later planting dates. Narrower gelatinization ranges were obtained from corn grown in a tropical environment than from corn grown in a temperate environment. Starch biosynthesis is subject to changes with environmental temperature, leading to the formation of different starch structures that can cause different functional properties. In a study of corn from two different backgrounds, each

ß 2004, Woodhead Publishing Limited

grown at two different temperatures, Lu et al. (1996) showed that kernel dry weight and density of corn decreased as the environmental temperature increased, in agreement with findings by White et al. (1991), who also showed that later planting dates were related to decreased kernel weight for corn from different lines. Lu et al. (1996) showed that starch in corn developed at a higher environmental temperature had a higher gelatinization temperature and wider temperature range than did that developed at a lower temperature, but the gelatinization enthalpy change was not affected by the developmental temperatures. A higher environmental temperature resulted in AM of smaller molecular size, and decreased iodine affinity, as well as apparent and true AM contents. The content of longer branch chains in AP increased as the developmental temperature increased, which correlated with the increase of gelatinization onset temperature. It was concluded that environmental temperature affects the enzymes that play a major role in the starch biosynthesis, such as branching enzyme I, which preferentially transfers long chains and has an optimum reactivity temperature higher than that of branching enzymes IIa and IIb, which transfer short chains and the different isoforms of soluble starch synthases.

10.2 Improving the functionality of corn starch for food processing applications: natural corn endosperm mutants Structurally and functionally, AP is the more important of the two starch fractions. Corn, and other grains, are capable of generating starch granules devoid of AM, as is observed in some mutant starches. Plants can be bred that produce starches with AM to AP contents outside the `normal' range (~25% AM and ~75% AP); for example, corn can be grown with an AM content as high as 85% (amylomaize) or as low as zero (waxy corn). Such families of starches are useful for studying the structural and functional roles of AM and AP within the granule, and also are useful in studying the synthesis and development of granules themselves. These starches generally contain AP to AM ratios that are much different than that of normal corn starch, and likely have molecules with altered structures, thus leading to an array of properties and specific applications of potential interest in food manufacturing. The normal and the corresponding endosperm mutant corn types have been produced through traditional plant breeding techniques, thus offering a natural alternative to corn developed via biotechnology techniques (genetically modified (GM)) for modifying and extending the range of functionality of a specific starch type. Native mutant starches could be marketed as `all natural' and non-GM. Many natural mutanttype corn types in which starch is affected were studied by Wang et al. (1993). A list of some of these single mutants is shown Table 10.1, along with their starch granule size distribution.

ß 2004, Woodhead Publishing Limited

Table 10.1 Size distribution of starch granules of maize genotypes from OH43 inbred corn measured by scanning electron micrographs (Wang et al., 1993) Genotypeã

Range (m)

Average ÔSDb (m)

Normal ae bt1 bt2 du1 h sh2 su1 wx

6±17 4±11 4±9 6±19 4±11 8±22 2±9 2±10 6±14

11.6Ô4.5 7.0Ô2.1 6.1Ô1.3 10.8Ô3.4 7.8Ô2.1 13.8Ô3.6 6.3Ô1.9 5.4Ô2.6 10.3Ô2.6

ãae = Amylose extender; bt = brittle; du = dull; h = horny; sh = shrunken; su = sugary; and wx = waxy. b Average ÔSD of 30 starch granules, 15 each from two micrographs.

10.2.1 The waxy mutation The waxy (wx) mutant is unique to all other known mutants relative to its lack of accumulation of AM (Shannon and Garwood, 1984). Waxy corn, containing essentially 100% amylopectin, is the most important raw material for modified starches (Yuan et al., 1993). This mutant lacks granule bound starch synthase (GBSS), which inhibits its production of AM (Doehlert and Kuo, 1994). The introduction of the wx gene into any corn type, or the combination of any other mutant with the wx gene results in a starch granule devoid of AM (Shannon and Garwood, 1984; Wang et al., 1993). When the corn kernel is cut with a knife, its lack of AM makes the cut surface appear shiny and waxy: thus, the name, waxy, was introduced. In general, wx and normal corn development, such as starch and dry weight production, and the accompanying starch granule morphology are similar. Both types of granules have A-type x-ray diffraction patterns.

10.2.2 The amylose extender (ae) mutation The ae mutation results in a loss of starch branching enzyme IIb activity (Boyer and Preiss, 1978). Amylomaize, containing the ae mutant gene, has an apparent AM content of up to 85%, and is associated with the presence of abnormal AP (Mercier, 1973, Ikawa et al., 1978, Boyer et al., 1980, Ikawa et al., 1981). The estimation of AM content is difficult because of the presence of branched components with long external chains, which can lead to an overestimation of the AM content. On the other hand, the presence of short chain-length AM can be responsible for an underestimation. Wolf et al. (1955) proposed that the unusual properties of ae AP could be attributed to a polymer of greater linearity than normal AP and of an intermediate structure between those of AM and AP. Boyer et al. (1980) observed that starch from ae wx double-mutant corn consists of an unusual AP similar to that of the intermediate fraction of ae starches. Wolf

ß 2004, Woodhead Publishing Limited

et al. (1955) showed that the AP fraction in a 50% amylomaize starch had longer inner and outer chains than those of normal AP, where the inner and the outer chains were divided at the point of -1,6-glucoside bonds. Montgomery et al. (1964) supported the proposal of Wolf et al. (1955), who suggested that the AP in amylomaize starch was less highly branched than was normal AP. More recently, however, it was proposed that the abnormal AP was a result of the presence of contaminating short-chain AM. The discrepancy likely originates because the same techniques for the dispersion and fractionation of starch are seldom used among researchers (Banks and Greenwood, 1975). Although the fine structure of abnormal AP from amylomaize has not been completely clarified, there are reports showing the presence of abnormal AP in amylomaize starch (Mercier, 1973, Ikawa et al., 1978, Boyer et al., 1980, Ikawa et al., 1981). In addition to the high apparent AM content, it has been shown (Takeda et al., 1993, Klucinec and Thompson, 1998) that ae starches contain AP with a higher proportion of longer chains (DP>30) than present in the AP of common corn starch. In addition to the ae gene, the combinations of a dull-1 (du1) or sugary-1 (su1) gene with other mutant genes (except the wx gene) also have produced starches with increased AM contents (Ikawa et al., 1981, Yeh et al., 1981, Inouchi et al., 1983, 1987, Boyer and Liu, 1985, Wang et al., 1993). Wang et al. (1993) reported increased AM as well as increased intermediate material content of these starches.

10.2.3 The du mutation The mutation of du1, when homozygous in otherwise non-mutant backgrounds, results in mature kernels with a tarnished, glassy, and somewhat dull appearance, which is referred to as to the `dull phenotype'. Total carbohydrate content in mature du1 mutant kernels is slightly lower than in normal corn (Creech, 1965, Creech and McArdle, 1966). The apparent AM content in starch from du1 mutants is slightly to greatly higher than normal, depending on the genetic background (Shannon and Garwood, 1984). Starch granules from du1 mutants seem to have normal structural and physical properties, although some abnormally shaped granules are found in the mutant endosperm (Shannon and Garwood, 1984). The du1 mutation causes reduced activity in the endosperm of two seemingly unrelated starch biosynthetic enzymes, starch synthase II (SSII) and starch branching enzyme IIa (SBEIIa) (Boyer and Preiss, 1981, Gao et al., 1998). The relative AM content of starch in du1 mutant kernels is significantly greater than it is in wild-type kernels. Approximately 15% of the starch in the du1 mutant endosperm is thought to consist of AP chains that are abnormally highly branched. Mutant du1 alleles, when combined with other mutations affecting starch synthesis, result in a broad range of alterations more severe than those in the single mutants. Double mutants containing du1, together with either a wx, ae, su1, or su2 mutation, have greater amounts of soluble sugars and lower total starch content than do any of the single mutants (Nelson and Pan, 1995).

ß 2004, Woodhead Publishing Limited

10.2.4 The su2 mutation The recessive su2 allele in corn identified by Eyster (1934) was found to reside on chromosome six. Perera et al. (2001) reported that the su2 starch granules consisted of lobes that resembled starch mutants deficient in soluble SSs, resulting in starch with a greater content of AM and a lower gelatinization temperature than that of normal corn starch (Pfahler et al., 1957, Kramer et al., 1958, White et al., 1994, Li and Corke, 1999, Perera et al., 2001). The su2 starches also have been shown to retrograde less during storage than do normal starches (Inouchi et al., 1991, White et al., 1994, Campbell et al., 1994). Li and Corke (1999) reported that the swelling power of su2 starch was significantly lower than that of normal corn starch. Starch isolated from genotypes containing su2 allele, in combination with du and su1, contained about 77% AM (Dvonch et al., 1951, Dunn et al., 1953). Also, su2 starch has an improved nutritional quality as a result of its high susceptibility to -amylase digestion; thus, its use has been suggested in improving animal feed value (Sandstedt et al., 1962). The su2 mutant also has been shown to enhance grain quality, because this allele in combination with opaque-2, resulted in kernel density nearly equal to that of ordinary dent corn (Glover et al., 1975). The su2 mutant has not been associated with any genetic lesion. Inouchi et al. (1984) and Boyer and Liu (1985) found that the su2 gene is associated with increased AM content to different extents. White et al. (1994), in describing the properties of su2 starch, found an AM content of 35%, smaller starch granules than normal starch, and suitable pasting properties for application in starch-thickened acidic foodstuffs. Inouchi et al. (1984) described starch of su2 as having an A-type X-ray diffraction pattern. The diffraction peaks, however, were broad and weak, reflecting a lower degree of crystallinity than found in normal starch. It was suggested that starch of su2 may differ from normal starch because of differences in the bonding of starch molecules or anomalous linkages within the molecules. Little difference was found in the fine structures of AP from su2 and normal starch on the basis of unit chain-length distributions (Inouchi et al., 1987). On the other hand, Takeda and Preiss (1993) determined that, compared with that of normal starch, the AP of su2 starch was composed of larger sized long B-chains, which were poorly branched, leading to an increased iodine affinity value. Several novel starches resulting from su2 alone and in combination with other alleles, such as du, ae, and wx, produced starches with properties resembling those of some modified starches (Friedman et al., 1988a,b, Wang et al., 1993). In addition, the use of starches from genotypes possessing the su2 allele, alone or in combination with other mutant genes, have been patented because of the favorable physical properties (White et al., 1994).

10.2.5 The su1 mutation Creech (1965) reported that sugary-1 (su1) corn kernels are wrinkled and have reduced amounts of dry material. The concentration of sugars is higher and the

ß 2004, Woodhead Publishing Limited

starch content is much lower than in normal corn. Extraction of su1 endosperm by different methods resulted in different distributions of starch and phytoglycogen (Boyer et al., 1981). The su1 mutants of corn have been known for decades to accumulate, in addition to starch, a novel form of water-soluble polysaccharide, termed phytoglycogen. Summer and Somers (1944) reported that the principal polysaccharide storage product in su1 endosperms was not starch, but a highly branched, water-soluble polysaccharide of high-molecular weight, which they called phytoglycogen. Phytoglycogen has twice the frequency of branch linkages as AP, a shorter average CL (average DP is approximately 10 versus an average of 20±25 for AP), and a significantly different chain-length distribution (Yun and Matheson, 1993). Thus phytoglycogen is multiply branched and lacks the packed crystalline helices of AP. These structural alterations cause the molecule to be water soluble, whereas AP in endosperm cells is insoluble. In 1958, Erlander proposed that glycogen could be considered as a precursor of starch biosynthesis, further suggesting that AP would be produced by debranching this precursor and AM would be subsequently generated through debranching AP. Pan and Nelson (1984) first reported that the su1 phenotype is caused by the loss of the activity of one of the three isoforms of R enzyme (RE) (a pullulanase-type starch-debranching enzyme) and possesses low activities of the other two RE isoforms, suggesting that the debranching enzyme also is involved in starch biosynthesis and that the debranching enzyme participates in the organization of regularly spaced clusters within AP. Doehlert et al. (1993) reported activities of total amylase and -amylase, as well as of RE, that were lower in su1 kernels of corn than in normal cultivars. This fact contrasts with the well-accepted idea that the starch debranching enzymes (pullulanase and isoamylase) are only involved in starch degradation in conjunction with other hydrolytic activities. James et al. (1995) showed, by cloning the su1 gene, that the su1 gene of corn encodes an isoamylase-like enzyme. Recently, it was reported that the su1 gene product possesses isoamylase activity, and that su1 mutants are deficient in both isoamylase and pullulanase (Rahman et al., 1998, Beatty et al., 1999). In work by Inouchi et al. (1983), phytoglycogen had a constant distribution of CL during kernel development, and AM content of su1 starches was somewhat higher than that of normal starches during kernel development. Inouchi et al. (1987) reported no long B chains in su1 phytoglycogen. Indeed, B chains of the phytoglycogen had comparatively uniform CL, which were shorter than the CL of the other corn starches. The su1 corn contained particulate granules, made up of phytoglycogen and AM (Matheson, 1975). Yeh et al. (1981) reported widely different AM percentages in su1 starch in different backgrounds ranging from 0±65 %, likely because of different environmental conditions during kernel development, kernel age, and methods of starch isolation and AM measurement.

ß 2004, Woodhead Publishing Limited

10.2.6 Comparisons of starch fractionation methods and functionality among mutant corn types in the same genetic background In a recent study involving most of the endosperm corn mutants just discussed (Tziotis, 2001), the fine structures of starch fractions obtained from a wild-type (normal) corn starch and amylose-extender25, dull39, sugary2, and sugary1 corn mutants in the same genetic background (ExSeed68) were isolated, evaluated, and compared by using three different fractionation procedures based on gelpermeation chromatography or alcohol precipitation methods. Starch fractions obtained from each of the three methods were enzymatically debranched and analyzed by using a high-performance anion-exchange chromatograph with a post-column amyloglucosidase reactor and a pulsed amperometric detector. No apparent differences in the molecular weight distributions of AP or of AM among the different starches were observed. The separations were performed by fractionation on a GPC column, by precipitation with 1-butanol, and by preferential precipitation with 1-butanol and isoamyl alcohol. The proportions of branch-CL of the starch components obtained by the various fractionation methods were very similar among methods for each of the starch types analyzed, such as the predominance of long branch-chains in ae25 corn and that of the short branch-chains in su2 corn. Overall, the effect of the mutations was more important in the differences observed among the starch types than was the method of analysis used. The thermal and functional properties of these starches also were evaluated and related to the structural features of the starches examined. X-ray diffraction patterns of the starches were obtained. The onset temperature of gelatinization values of starches from all mutant lines ranged from 52.0 ëC for su2 to 62.9 ëC for du39, temperatures that were all lower than that of the normal starch (64.5 ëC) as measured by using DSC (Table 10.2). The change of enthalpy of gelatinization of starch from su2 (7.7 J/g) was less than that of the wild type starch (14.1 J/g). The viscosity of the su2 starch over the cooking process was Table 10.2 Differential scanning calorimetry thermal properties of starch from ExSeed68 (wild-type) corn and mutants, dull39, amylose-extender25, sugary2, and sugary1, in the ExSeed background (Tziotis, 2001) Starch typea

Peak onset temperature (ëC)

Gelatinization range (ëC)

Change in enthalpy (J/g)

Retrogradation (%)

Wild-type du39 ae25 su2 su1

64.5a 62.9b 62.5b 52.0d 60.4c

8.4e 9.8d 16.8a 13.1b 11.3c

14.1b 16.1a 15.7ab 7.7c 8.5c

58.5 33.3 57.5 33.1 38.1

a

b

Values reported are means of four replicates. Numbers followed by the same lower-case letter within each column are not significantly different at P < 0.05. b Value reported for ae25 starch is the sum of enthalpy change obtained from this peak and another peak not reported in the table.

ß 2004, Woodhead Publishing Limited

relatively stable as measured by Rapid Visco Analyser (RVA), showing only a small breakdown of the peak viscosity during cooking, and suggesting high stability of the starch granules against mechanical shear. The su1 mutant starch formed the strongest gel among all starch-gel samples during measurements of both fresh and stored gels evaluated by using a Texture Analyzer.

10.2.7 Influence of genetic background on functional properties Variations in the functions of mutant starches related to structural differences were noted when the mutant was placed in corn with different genetic backgrounds. For example, Li and Corke (1999) developed five different corn inbred lines and evaluated the thermal, pasting, and gel textural properties of each isogenic line consisting of normal corn along with its du and su2 mutants. Differences were reported among the same mutant but in a different background, in the AM percentage (varying between 29.5 and 43.2% for su2, and 29.0 and 37.6% for the du mutants); swelling power; solubility; digestibility; onset, peak, and conclusion temperatures of gelatinization; change of enthalpy of gelatinization; pasting properties; and gel firmness and adhesiveness. These values compare with the work of Tziotis (2001) just mentioned (Table 10.2). The gelatinization onset temperature (To) values of starches from mutant lines all placed in the same experimental background line of ExSeed68 were all lower than that of the normal starch, which agrees with relative values reported by others for these mutants in the same (Perera et al., 2001) and other genetic backgrounds (Pfahler et al., 1957, Kramer et al., 1958, Brown et al., 1971, Ninomya et al., 1989, Campbell et al., 1994, Ng et al., 1997, Li and Corke, 1999). The change in enthalpy of gelatinization, gelatinization range values, and retrogradation percentage values reported by Tziotis (2001) also reflected differences among the mutants that were similar to results from these other studies. Both absolute and relative thermal property values of corn starches can vary, however, based on their genetic backgrounds, as indicated by the following reports on To values. One should bear in mind that the values, although all obtained by DSC, did not necessarily result from the same heating parameters and starch moisture contents. The du mutant starch in the IA5125 inbred line had a To of 61.0 ëC, the ae mutant starch had a To of 68.8 ëC, and normal starch a To of 64.2 ëC (Sanders et al., 1990). Similarly the To for the ae starch in the W64A background (70.6 ëC) was greater than that of normal starch (63.9 ëC) (Krueger et al., 1987). The su2 mutant in an Oh43 background had a variable To, which decreased with increased dosage of su2, reaching 58.3 ëC (normal 67.3 ëC) for the complete su2 mutant background (Campbell et al., 1994). Alternatively, Inouchi et al. (1991) reported that normal corn starch in the Oh43 background had a To of 61 ëC, with the mutants as follows: ae (65 ëC), du (64 ëC), and su2 (45 ëC). Thus, all mutant starch values in that study, except for su2, were greater than that of the normal starch. Wang et al. (1992) also reported DSC values for corn mutants in the Oh43 inbred line with slightly greater

ß 2004, Woodhead Publishing Limited

relative and absolute To values than those of Inouchi et al. (1991). The To of the ae, du, and su1 mutant starches were 68.7, 67.2, and 64.6 ëC respectively, with the normal counterpart having a To of 67.2 ëC. Li and Corke (1999) studied corn starch from five different genetic backgrounds (A632, Oh43, Hz85, Hz101, and Hz47) consisting of the du and su2 mutants and their normal counterparts. In all cases in that study, the properties of du mutant starch were very similar to those of the normal starch, as noted by Wang et al. (1992), and the su2 mutants had To values approximately 10 C lower than those of the normal. The su1 mutant corn starch in the P39±5XP51±B background had a To of 62 C, with the normal counterpart at 65 C (Ninomya et al., 1989).

10.3 Chemically modifying corn starches for use in the food industry The use of starch in many industrial applications depends on its granular structure; cold water solubility; colloidal dispersability during heating; film forming, binding, adhesion, thickening and stabilizing abilities; and textural contribution (Jarowenko, 1986). In addition to the altered structures of starches from the naturally occurring corn endosperm mutants, chemical modifications serve as a powerful tool to expand the spectrum of corn starch properties and uses. Chemical modifications usually involve the addition of moieties on the linear chains of the glucose units of the starch molecules, thus changing the molecular size and viscosity characteristics of the starch. The response of starch to chemical treatments depends on its origin, type, and history. Chemical modification of starch, by placing substituent groups along its polymeric backbone, decreases gelatinization temperature, and increases translucency, viscosity, freeze-thaw stability, solubility, swelling power, hardness, cohesiveness and adhesiveness of the starch gel (Betancur-Ancona et al., 1997).

10.3.1 Succinylation Succinylation of starch, by the addition of succinic anhydride (dihydro-2,5furandione) to its molecules, may be desirable in the food industry because the modification improves properties, such as decreasing gelatinization temperature; increasing freeze-thaw stability, thickening power, viscosity stability, clarity, ability to swell in cold water, and stability in acid and salt; and reducing the tendency to retrograde (Tessler and Wurzburg, 1983, Trubiano, 1987, Bhandari and Singhal, 2002), although this chemical process is used more in industrial applications than in food applications. The beginning material for starch succinylation is generally waxy starch, because the combination creates very useful starches. The peak viscosity increases slightly with increased treatment level (with succinic anhydride), whereas the final viscosity decreases, and the low-temperature stability improves. Cooked, unmodified corn starch forms a gel

ß 2004, Woodhead Publishing Limited

on cooling, whereas starch succinates give smooth, stable, high-viscosity colloidal suspensions. Extrusion also has been employed to prepare starch succinates with succinic anhydride (Tomasik et al., 1995). Bhandari and Singhal (2002) compared the swelling power of corn starch succinates to that of the native corn. They reported that the swelling power of native corn starch increased from 2.1 to 11 over the 45 to 95 ëC temperature range, whereas the swelling power of corn starch succinate with degree of substitution (DS) was 2.14 at 45 ëC and 17 at 95 ëC for corn starch succinates of DS 0.20. They considered that the increase in swelling power with the DS could be due to easy hydration, an indication of the increasing number of hydrophilic groups incorporated in the starch. Regarding the Brabender characteristics of corn starch succinates, the peak and cold paste viscosities increased with an increase in DS from 0.05 to 0.20. When studying the freeze-thaw stability of the same starches, native corn starch showed 66% syneresis after ten cycles, whereas with an increase in DS, resistance to syneresis improved. Also, percent syneresis decreased with an increase in DS. For example, a corn starch succinate with 0.20 DS demonstrated excellent freeze-thaw stability. In addition, the corresponding gelatinization temperature dropped with increased treatment level. Corn starch treated with succinic anhydride produces starch pastes that are relatively clear, stable to viscosity changes, but gummy in texture, which is not always desirable in food applications. To avoid the gummy texture, succinic anhydride can be reacted with lightly cross-linked granular starch so that the resulting texture of the cook is short, smooth, and shiny, with retention of the stability and much of the original clarity. Cross-linking also improves the resistance of the derivative to breakdown at high temperatures. Succinate derivatives are used in the food industry as binders and thickening agents in soups, snacks, canned, and refrigerated food products. Their properties also are desirable in pharmaceuticals, where they are used as tablet disintegrants. Crosslinked high-AM starch (Hylon VII) was introduced a few years ago as an excipient (ContramidTM) for controlled drug release (Ispas-Szabo et al., 2000). This starch swells in water to form an elastic gel. Its ability to regulate the swelling controls drug release in aqueous media as a function of cross-linking density. So this hydrogel is particularly suitable as a pharmaceutical excipient. These properties are strongly dependent on the degree of the cross-linking of the starch. Best release properties and highest mechanical hardness were obtained from cross-linked high-AM starch matrices with low-to-moderate crystallinity, where the V- and the B-type structures coexist with amorphous regions (Ravenelle et al., 2002).

10.3.2 Acetylation In a study, Wilkins et al. (2003a) evaluated the variability of starch acetylation caused by hybrid influence. Six wx corn hybrids grown in 1998 and five wx corn hybrids grown in 1999 were wet-milled in the laboratory and modified, with the reaction efficiencies monitored. Reaction efficiencies were highly variable (47

ß 2004, Woodhead Publishing Limited

to 73%), and were significantly lower for 1998 hybrids (50.0%) than for the same hybrids grown in 1999 (62.7%). Regarding the pasting properties of the modified starches analyzed by using an RVA, acetylated starch from 1999 had increased peak, trough and final viscosities and increased reaction efficiency as compared with acetylated starch from 1998. Differences in setback were observed among 1998 hybrids for acetylated samples. Differences in trough and final viscosity were observed among 1999 hybrids for acetylated and native (unmodified) samples, whereas differences in breakdown among 1999 hybrids were observed for native samples. Also, Wilkins et al. (2003b) evaluated the differences in the pasting properties and reaction efficiencies of acetylated dent corn starch from ten dent corn hybrids grown during 1998 and nine dent corn hybrids grown during 1999, all wet-milled in the laboratory. Acetyl content (reaction efficiency) was measured by a spectrophotometric method and ranged from 35 to 56%. Reaction efficiencies for starches from 1998 hybrids were lower than for starches from 1999 hybrids. Differences in peak viscosity, trough viscosity, final viscosity, setback, and pasting temperature occurred among 1998 hybrids, whereas differences only in trough viscosity, final viscosity, and breakdown were found among 1999 hybrids. Wang and Wang (2002) studied the effects of the catalyst used in acetylation, including sodium hydroxide (NaOH), potassium hydroxide (KOH), and calcium hydroxide (Ca(OH)2), on the chemical and physicochemical properties of acetylated waxy corn starch. The Ca(OH)2-catalyzed acetylated starch had a slightly higher pasting temperature and a lower -amylolysis limit than did the acetylated starch prepared under NaOH or KOH catalysis, but there were no differences related to their thermal properties. The isoamylase-debranched acetylated starches catalyzed by Ca(OH)2-and their -limit dextrins showed an elution profile, analyzed by high-performance size-exclusion chromatography, that was different from those of the other two acetylated starches with a greater proportion of saccharides eluted at a longer retention time. However, the differences in pasting temperature, -amylolysis limit, and carbohydrate profile among the acetylated starches diminished when ethylenediaminetetraacetic acid (EDTA) was added. The results suggested that calcium might induce intermolecular cross-linking through chelation with oxygen of the anhydroglucose units and that this type of cross-linking was promoted in acetylation catalyzed by Ca(OH)2.

10.3.3 Hydroxypropylation The effects of modification sequence on chemical structures and physicochemical properties of hydroxypropylated and cross-linked waxy corn starch were recently reported (Wang and Wang, 2000), where the chemical structures of dual-modified starches and their -limit dextrins were characterized with high-performance liquid-chromatography. The hydroxypropylatedcross-linked starch had higher Brabender viscosity than did the cross-linkedhydroxypropylated starch at both pH 7 and 3; but both starches has similar

ß 2004, Woodhead Publishing Limited

gelling properties. The hydroxypropylated-cross-linked starch had significantly higher onset and peak gelatinization temperatures, gelatinization enthalpy, and lower retrogradation. Also, the latter starch exhibited a significantly higher amylolysis limit and higher content of low molecular weight saccharides in its isoamylase-debranched starch, suggesting a structure more accessible to enzymatic digestion than that of the cross-linked-hydroxypropylated starch. Further structural analyses revealed different distribution patterns of modifying groups between the two modified starches. The results indicated that the modification sequence was responsible for changing the susceptibility to enzymes, the locations of substitution, and the physicochemical properties of the hydroxypropylated and cross-linked wx corn starches. Shi and Be Miller (2002) recently evaluated the aqueous leaching of the hydroxypropylated common corn starches at different times and temperatures. Results indicated that the greater the modification the easier it was for AM to leach out and that the preference for leaching of derivatized AM decreased as molar substitution of the whole starch increased. McPherson and Jane (2000) studied the effects of extrusion variables on granular morphology and molecular weight of the starch components of native and cross-linked (0.0 to 0.028% POCl3) hydroxypropylated (8%) corn starches. Extrusion of starches caused substantial morphological changes in granular structure. Starches extruded at 60 ëC showed distorted and fragmented granules, whereas the extruded starches at 100 ëC showed no granular structure and were completely amorphous. Extrusion conditions affected the molecular weights of the extruded starches. Increased starch moisture content reduced AP degradation during extrusion. Also, cross-linking prevented AP degradation. However, the magnitude of AP degradation increased at higher levels of cross-linking as shear increased. Increased temperature of extrusion decreased AP molecular weight of extruded native and hydroxypropylated corn starches, whereas the opposite effect was observed in cross-linked hydroxypropylated starches, likely related to the glass transition temperature. Bae and Lim (1998) performed the hydroxylation of normal (25% AM) and high-amylose (70% AM) corn starches to 0.1 degree of molar substitution with propylene oxide in an alkaline-ethanol medium (70% ethanol). Stearic acid, glycerol monostearate, or lecithin (3%, based on starch) was added to each mixture to examine the effects on the physical properties of the extrudate. Highamylose, alone and with all additives, showed lower die swelling in extrusion than did normal corn starch, whereas hydroxypropylated normal corn starch and hydroxypropylated high-amylose starch showed higher die swelling than the corresponding unmodified starches. Hydroxypropylation increased the water absorption for both starches (high-amylose extrudates from 22 to 35% and normal corn starch extrudates from 68 to 97% at 25 ëC). Differential scanning calorimetry showed that during extrusion, the lipid additives formed a helical complex with AM in normal and high-amylose starch, but that the complex was weak with their hydroxypropylated derivatives. The extruded strands of highamylose starch, alone and with additives, exhibited higher tensile and bending

ß 2004, Woodhead Publishing Limited

strengths (37.1±58.4 and 2.16±5.07 MPa, respectively), than did the normal corn starch strands (12.4±59.3 and 1.06±4.10 MPa, respectively) at the same moisture content (7.5±8.5%). Both tensile strength and percent of elongation of the starch strands were reduced by the presence of a lipid additive. Hydroxypropylation increased elongation and flexibility of the extrudates. Also, hydroxypropylated high-amylose starch exhibited the greatest mechanical strength and flexibility among all starches.

10.3.4 Phosphorylation Liu et al. (1999) reported the preparation of phosphorylated starches with sodium tripolyphosphate at pH 6, 8, and 10 from wx (3.3% AM), normal (22.4% AM), and two high-amylose (ae, 47 and 66% AM) corn starches. The resulting starches had a decreased gelatinization peak temperature and an increased pasting peak viscosity except for wx, which showed a slight increase in gelatinization temperature. There was a substantial effect of phosphorylation pH on paste viscosity. Also, more cross-linking was found in ae starches with phosphorylation at pH 10. It was indicated that sodium ions decreased the paste viscosity of all the phosphorylated starches, whereas only slightly affected the paste viscosity of native starches. Phosphorylation increased swelling power of some of the starches, with greatest swelling power at phosphorylation pH 8 and least at pH 10. In the case of wx starch, maximum swelling power was obtained after preparation at pH 8 and minimum swelling power at pH 6. With phosphorylation, the clarity and freeze-thaw stability of all the starches was greatly increased compared to the native starches. Digestibility of phosphorylated ae starches was increased, but phosphorylation had little effect on wx and normal starches. All phosphorylated derivatives showed an increase in the adhesiveness, springiness, and cohesiveness; additionally, the hardness of 47% ae and wx starches increased, and that of normal starches decreased. Enthalpy of gelatinization decreased after phosphorylation with the ae starches exhibiting the greatest decrease. Lim and Seib (1993a) produced phosphorylated corn starches with 5% sodium tripolyphosphate (STPP) and/or 2% sodium trimetaphosphate (STMP). All phosphorylation reactions were done using 5% sodium sulfate and adjusting the initial reaction pH by adding aqueous sodium hydroxide or hydrochloric acid to the pre-reaction slurries. The degree of phosphorylation decreased 40 to 50% with STPP as reaction pH increased from 6 to 11, whereas it increased by 100% with STMP. Corn starch phosphate with the most desirable pasting properties was obtained at an initial pH of 11 with STPP and contained 0.16% P (including 0.02% from lipid). When corn starch was treated with a mixture of 5% STPP and 2% STMP, the best product when pasted at 95 ëC was obtained at the initial reaction pH of 9.5. Paste clarity of the phosphorylated starches indicated that cross-linking accelerated rapidly above pH 8 with STMP, but above pH 10 with STPP.

ß 2004, Woodhead Publishing Limited

10.3.5 Cross-linking Cross-linking is the most widely used technology for the improvement of the properties of native starches, which are usually sensitive to shear, high temperature, and acid treatment when cooked in water. Cross-linked wx cereal starches generally show a `short' spoonable texture, higher paste stability, and resistance to cooking shear, temperature, and low pH as compared to native starches (Whistler and Be Miller, 1997). On the other hand, cross-linking reduces paste clarity and stability to cold storage. In that case, the undesirable properties can be improved by further modification treatments, such as esterification or etherification. Cross-linked starch is essential in the manufacture of foods to thicken, stabilize, and provide texture. The creation of cross-linked starch originated from the need for starch granules, which resist disintegration on cooking with water. To avoid a thick, pasty mass, a process was designed to chemically treat native starch with acid chlorides including phosphorous oxychloride (POCl3) in water (Felton and Schopmeyer, 1943). Other researchers followed suit with novel chemical approaches to cross-linking starch using other reagents such as epichlorohydrin (EPI) or sodium trimetaphosphate (STMP) (Konigsburg, 1950, Hofreiter et al., 1960, Lloyd, 1970). Waxy starches (lacking AM) are often used as the base for cross-linked starches because AM retrogrades on cooling and forms an irreversible gel (Katzback, 1972). The mechanism of the reaction is as follows: cross-linking agents bind neighboring anhydroglucose units in the amorphous regions of the wx corn AP. The cross-links prevent the granules from fully swelling and ultimately disintegrating. The covalent cross-link network also makes the granules less susceptible to pH extremes and high shear processes common to food manufacturing. The extent of the effects of cross-linking on swelling and viscosity depends both on the treatment conditions of raw starch and on how the starch is prepared in the final application. Major factors in the cross-linking reaction include chemical composition of reagent, reagent concentration, pH, reaction time and temperature (Lim and Seib, 1993b). The degree of crosslinking for food starches is very low, making the extent of reaction and yield of cross-linked starch difficult to measure chemically, consequently requiring the measurement of physical properties. Maximum extent of cross-linking reaction for EPI with corn starch, assuming that the percentage of reacted EPI that results in cross-links is constant, was reported at 90% (42 hr at 25 C) by Hammerstrand et al. (1960). No such quantitative values are reported for POCl3 or STMP. Cross-linked starches are produced when unswollen, native granules are mixed in an aqueous system with reagents capable of reacting with at least two of the hydroxyl groups of neighboring molecules (Wurzburg and Szymanski, 1970). The type of reagent used and cross-linking conditions determine the ratio of mono and di-type bonds (esters with phosphorous based agents and glycerols with epichlorohydrin) caused by the cross-linking reaction mechanism and available starch hydroxyls (Koch et al., 1982). Starch thickening properties can be controlled by changing the degree of cross-linking and manipulating the extent of swelling. A relationship between

ß 2004, Woodhead Publishing Limited

rheological properties and swelling capacity of starch granules has been demonstrated (Evan and Haisman, 1979, Bagley and Christianson, 1982). The flow behavior and textural properties of cross-linked starch are very complex due to the effects of starch concentration, heating rate, heating temperature, and amount of shear, as well as competition with other dissolved solutes and polymers (Doublier et al., 1987b, Steeneken, 1989, Gluck-Hirsch and Kokini, 1997). Relative effects that the different cross-linking agents have on physical properties have been studied (Evans and Haisman, 1979 Eliasson, 1986, Steeneken, 1989, Evans and Lips, 1992); however, it is uncertain how crosslinking is achieved at the macromolecular level of the 3-D granule structure. Jane et al. (1992) reported that when granular starch was cross-linked, a greater amount of AP than AM was found cross-linked. When corn starch was treated with a cross-linking reagent (0.07% epichlorohydrin, at pH 10.5 for 24 hr), 91% of its AP and 45% of its AM became insoluble. Cross-linking of pregelatinized and dispersed starch caused less difference in the proportion of soluble AM and AP than did the cross-linking of native granular starch. After the starch had been cross-linked in the granular form, there was no increase in the size of AM as a result of cross-linking between two or more AM molecules. However, susceptibility of the AM to sequential hydrolysis by isoamylase and -amylase decreased. The relative blue values of AP peaks indicated that AM was cross-linked to AP, which was confirmed when the AP isolated from crosslinked starches was debranched with isoamylase.

10.4 Genetically modifying corn starches for use in the food industry It is important to search both inside and outside the Corn Belt in the United States for corn with desirable properties, because of the potentially beneficial properties these new corn starches can contribute to produce high-quality food products, and because of the introduction of new genes to diversify the gene pool. Pollak and White (1997) examined the range of variability of starch functional traits in Corn Belt inbred lines and exotic inbred lines from Argentina, Uruguay, and South Africa. Reciprocal hybrids of some of the lines within each set were compared with their parents. Functional traits were examined by using DSC on starch extracted from single kernels of genotypes (Table 10.3). The Corn Belt lines had a wider range of values for most traits than did the set of exotic lines. For both sets of lines, the maximum value for peak height index (a measure of the enthalpy change divided by the range) was as high as that previously reported for the wx endosperm mutant (data not shown). The exotic lines showed a wider range of values for percentage retrogradation. Hybrid values were not consistently higher, lower, midpoint, or similar with respect to the values of their parents. This finding was true regardless of germplasm type or functional trait. Reciprocal cross values showed trends suggesting reciprocal differences, although there was no trend suggesting greater

ß 2004, Woodhead Publishing Limited

Table 10.3 Means, maximum and minimum values and standard errors for starch thermal traits of corn belt and exotic inbred lines and their crosses (Pollak and White, 1997) Peak onset temperature (ëC)

Gelatinization range (ëC)

Change in enthalpy (J/g)

Retrogradation (%)

66.4 71.2 61.2 1.16

12.0 17.6 8.2 1.13

12.13 13.39 10.88 0.25

50.7 53.0 46.0 0.76

Corn belt crosses Mean 68.9 Maximum 70.1 Minimum 66.4 Standard error 0.26

8.0 9.3 6.0 0.19

12.60 13.80 10.88 0.21

48.1 53.0 42.0 0.62

Exotic lines Mean Maximum Minimum Standard error

65.7 68.1 62.9 0.65

10.3 13.5 7.6 0.73

11.18 12.55 8.37 0.42

52.9 63.0 42.0 2.35

Exotic crosses Mean Maximum Minimum Standard error

68.2 70.6 65.9 0.35

8.5 10.9 6.9 0.28

12.55 14.64 11.30 0.21

51.3 57.0 45.0 0.70

Corn belt lines Mean Maximum Minimum Standard error

effect of the female parent. These traits seemed to be controlled by many modifying effects in addition to major effects. Results indicated that sufficient variability exists even within the Corn Belt germplasm to conduct breeding and inheritance studies effectively and that there should be potential for breeding for functional traits. Campbell et al. (1995) examined the genetic modifiers of corn starch thermal properties by using DSC. The su2 kernels from segregating ears were identified based on textural appearance of starches following crosses between an exotic corn accession with the inbred OH43, which was homozygous for the su2 allele (OH43 su2). The two exotic corn accessions used were the PI213768 and PI451692. Germs were retained from su2 kernels and used to produce an F2 population of su2 plants containing 50% exotic germplasm. With few exceptions, F2 ears from the populations were homozygous for the su2 allele. Significant (P less than or equal to 0.05) differences were observed between the exotic populations and OH43 su2 for gelatinization onset temperature, range, enthalpy, and retrogradation. The number of DSC values with significant withinpopulation variations was greater among F2 ears within the exotic populations

ß 2004, Woodhead Publishing Limited

Table 10.4 Differential scanning calorimetry values for starches from exotic and inbred corn lines crossed with the corn mutant, sugary-2 (Campbell et al., 1995) Su2 F2 population

MeanÔSD

(Exotic X OH42su2)

Peak onset temperature (ëC)

Gelatinization range (ëC)

Change in enthalpy of gelatinization (J/g)

Retrogradation (%)

P1213768su2 P145169su2 OH43su2

52.8Ô1.6 53.5Ô1.3 54.6Ô0.8

13.4Ô1.7 12.2Ô1.3 10.6Ô0.7

6.3Ô0.67 5.4 0.88 5.4Ô0.21

34.5Ô4.8 32.9Ô5.1 29.0Ô4.1

than among ears within the inbred line OH43 su2. Standard deviations for DSC values were consistently greater for exotic su2 populations than for those of OH43 su2 (Table 10.4). Also, starch from the population PI213768 su2 had mean values by DSC that were significantly different from those of from starch of OH43 su. It was concluded that examining the texture of starches from single kernels can be useful in identifying and developing populations homozygous for the su2 allele. In addition, the increased variability for DSC values within populations containing 50% exotic germplasm suggests that genetic modifiers might be used to alter thermal properties and, possibly, functional properties of su2 starch. Other variations in corn starch structure and function may arise from the introduction of ancient relatives of corn into modern corn lines. Teosinte, a variable wild grass, has some resemblance to maize in that it has tassels (male flowers) at the end of stalks, and broad, flat, pointed leaves. Reeves (1950) determined that crossing teosinte with corn resulted in greater resistance of the corn to heat and drought damage. In a recent study, the structural and chemical properties were compared of starches isolated from modern corn (maize), Chalco teosinte, and a Chalco teonsinte-maize cross (Keppel, 2001). The Chalco teosinte was developed for growth in the US Corn Belt, with adaptation for photo sensitivity of the region. Teosinte contained only 25% starch compared with 70% starch in the corn. The starch content of the cross was similar to that of the corn. Teosinte starch had a gelatinization onset temperature of 61 ëC, a gelatinization enthalpy change of 11.4 J/g, and a retrogradation of 56.2%, compared with values for the corn of 66.1 ëC, 15.2 J/g and 41%, respectively. The cross had intermediate values of 64.0 ëC, 13.9 J/g and 49.7%, respectively, suggesting some modification of corn starch with the introduction of the teosinte genes. Starch granules from the teosinte were similar in size and shape to corn starch granules, but had the most irregularly shaped, broken or hollow granules, as noted by scanning electron microscopy. All granule types were similar in x-ray diffraction patterns (A-type) and amylose contents, but teosinte starch had the greatest proportion of short chains (18.5%) and shortest average chain-lengths

ß 2004, Woodhead Publishing Limited

(23.5 DP), with corn starch having values of 16.6% and 25.1 DP, respectively. Once again, the cross had values that were intermediate between those of corn and teosinte starches. Pasting properties by Rapid Visco Analysis showed that teosinte starch had the lowest peak viscosity, final viscosity, and setback values, but the highest pasting temperature among the starch types, again suggesting promise for introducing variability into Corn Belt lines. Trypsacum, another ancient maize-like plant, may provide variance in corn starch structure and function when crossed with modern corn lines to create unusual corn properties. Duvick et al. (2003) have a patent pending on the use of such crosses to create corn with altered starch, oil, and protein quality and quantity.

10.5

Future trends

As noted in this chapter, there are many available choices for selecting a specific corn starch with properties desirable for the food industry, including normal corn from different genetic backgrounds, naturally occurring endosperm mutants, as well as chemically modified corn starches. Another potential source of corn starches with altered properties is genetically modified (GM) crops, but the future of GM materials is uncertain, pending government and consumer acceptance in the field and market. GM corn can enhance the successful growing of corn in the field with the absence of pesticides and other chemicals, thus leading to a more profitable product for the growers. But the benefits and risks of these crops has not yet been fully evaluated to the satisfaction of some countries.

10.6

Sources of further information and advice

Valuable information regarding the latest issues, as accessed June 2003, can be found on the web at: American Association of Cereal Chemists http://www.aaccnet.org Iowa Corn Promotion Board: http://www.iowacorn.org National Corn Growers Association: http://www.ncga.com

10.7

References and GREENWOOD C T (1969), `Studies on starches of high amylose-content. Part X. An improved method for the fractionation of maize and amylomaize starches by complex formation from aqueous dispersion after pretreatment with methyl sulfoxide', Carbohydr. Res., 11, 217±224.

ADKINS V L,

ß 2004, Woodhead Publishing Limited

and ARAI Y (1984) `Structural characterization of amylopectin and intermediate material in amylomaize starch granules'. Agric. Biol. Chem., 48, 1763±1775. BAE S O, and LIM S T (1998), `Physical properties of extruded strands of hydroxypropylated normal and high-amylose corn starch', Cereal Chem., 75, 449±454. BAGLEY E B, and CHRISTIANSON D D (1982), `Swelling capacity of starch and its relationship to suspension viscosity ± effect of cooking time, temperature, and concentration', J. Texture Stud., 13, 114±126. BAHNASSEY Y A, and BREENE W M (1994), `Rapid viscoanalyzer (RVA) pasting profiles of wheat, corn, waxy corn, tapioca, and amaranth starches (A. hypochondriacus and A. cruentus) in the presence of konjac flour, gellan, guar, xanthan, and locust bean gums', Starch, 48, 134±141. BANKS W, and GREENWOOD C T (1967), `The fractionation of laboratory-isolated cereal starches using dimethyl sulfoxide', Starch, 19, 394±398. BANKS W, and GREENWOOD C T (1975), `Fractionation of the starch granule, and the fine structures of its components', in Banks W and Greenwood C T, Starch and Its Components, Edinburgh, Edinburgh University Press, 4±66. BEATTY M K, RAHMAN A, CAO H, WOODMAN W, LEE M, MYERS A M, and JAMES MG (1999), `Purification and molecular genetic characterization of ZPU1, a pullulanase-type starch debranching enzyme from maize', Plant Physiol., 119, 254±266. BETANCUR-ANCONA D, CHEL-GUERRERO L, and CANIZARES-HERNANDEZ E (1997), `Acetylation and characterization of Canavalia ensiformis starch', J. Agric. Food Chem., 45, 378±382. BHANDARI P N, and SINGHAL R S (2002), `Effect of succinylation on the corn and amaranth starch pastes', Carbohydr. Polym., 48, 233±240. BOYER C D, and LIU K C (1985), `The interaction of endosperm genotype and genetic background. Part I. Differences in chromatographic profiles of starches from nonmutant and mutant endosperms', Starch, 37, 73±79. BOYER C D, and PREISS J, (1978), `Multiple forms of (1,4)--D-glucan-6±glucosyl transferase from developing Zea mays L. kernels', Carbohydr. Res., 61, 321±334. BOYER C D, and PREISS J (1981), `Evidence for independent genetic control of the multiple forms of maize endosperm branching enzymes and starch synthases', Plant Physiol., 67, 1141±1145. BOYER C D, DAMEWOOD P A, and MATTERS G L (1980), `Effect of gene dosage at high amylose loci on the properties of the amylopectin fractions of the starches', Starch, 32, 217±222. BOYER C D, DAMEWOOD P A, and SIMPSON E K J (1981), `The possible relationship of starch and phytoglycogen in sweet corn. I. Characterization of particulate and soluble polysaccharides', Starch, 33, 124±130. BROWN R P, CREECH R G, and JOHNSON L J (1971), `Genetic control of starch granule morphology and physical structure in developing maize endosperms', Crop Sci., 11, 297±302. CAMPBELL M R, WHITE P J, and POLLAK L M (1994), `Dosage effect at the sugary-2 locus on maize starch structure and function', Cereal Chem., 71, 464±468. CAMPBELL M R, WHITE P J, and POLLAK L M (1995), `Properties of sugary-2 maize starch: influence of exotic background', Cereal Chem., 72, 389±392. CREECH R G (1965), `Genetic control of carbohydrate synthesis in maize endosperm', Genetics, 52, 1174±1186. CREECH R G, and MCARDLE F J (1966), `Gene interaction for quantitative changes in carbohydrates in maize kernels', Crop Sci., 6, 192±194. BABA T,

ß 2004, Woodhead Publishing Limited

and KUO T M (1994), `Gene expression in developing kernels of some endosperm mutants of maize', Plant Cell Physiol., 35, 411±418. DOEHLERT D C, KUO T M, JUVIK J A, BEERS E P, and DUKE S H (1993), `Characteristics of carbohydrate metabolism in sweet corn (sugary-1) endosperms', J. Am. Soc. Hort. Sci., 118, 661±666. DOUBLIER J L, LLAMAS G, and LEMEUR M (1987a), `A rheological investigation of cereal starch pastes and gels. Effect of pasting procedures', Carbohyd. Polym., 7, 251± 175. DOUBLIER J L, PATON D, and LLAMAS G (1987b), `A rheological investigation of oat starch pastes', Cereal Chem., 64, 21±26. DUNN G M, KRAMER H H, and WHISTLER R L (1953), `Gene dosage effects on corn endosperm carbohydrates', Agron. J., 45, 101±104. DUVICK, S A, POLLAK L M, and WHITE PJ (2003), `Altered fatty acid, protein, oil and starch corn lines and method for producing same', US Patent pending 09/285,368. DVONCH W, KRAMER H H, and WHISTLER R L (1951), `Polysaccharides of high amylose corn', Cereal Chem., 28, 270±279. ELIASSON A-C (1986), `Viscoelastic behaviour during the gelatinization of starch', J Texture Stud., 17, 253±265. ERLANDER S R (1958), `A proposed mechanism for the synthesis of starch from glycogen', Enzymologia, 19, 273±283. ERLANDER S R, MCGUIRE J P, and DIMLER R J (1965), `An anomalous low-molecular-weight branched component in dent corn starch', Cereal Chem., 42, 174±186. EVANS I D, and HAISMAN D R (1979), `Rheology of gelatinized starch suspensions', J Texture Stud., 10, 347±370. EVANS I D, and LIPS A (1992), `Viscoelasticity of gelatinized starch suspensions', J Texture Stud., 23, 69±86. EYSTER W H (1934), `Genetics of Zea mays', Bibliogr. Genet., 11, 187±392. FELTON G E, and SCHOPMEYER H H (1943), `Thick bodied starch and method of making', US Patent 2,328,537. FRIEDMAN R, GOTTNIED D J, FARON E J, PUSTEK F J, and KATZ F R (1988a), `Food stuffs containing starch of a dull sugary-2 genotype', US patent 4,792,458. FRIEDMAN R, GOTTNIED D J, FARON E J, PUSTEK F J, and KATZ F R (1988b), `Food stuffs containing starch of an amylose sugary-2 genotype', US patent 4,798,735. GAO M, WANAT J, STINARD P S, JAMES M G, and MYERS A M (1998), `Characterization of dull1, a maize gene coding for a novel starch synthase', Plant Cell, 10, 399±412. GLOVER D V, CRANE P L, MISRA P S, and MERTZ E T (1975), `Genetics of endosperm mutants in maize as related to protein quality and quantity', in Mertz E T, High Quality Protein Maize, Stroudsburg, PA, Dowden, Hutchinson and Ross, Inc, 228±240. GLUCK-HIRSCH J, and KOKINI J L (1997), `Determination of the molecular weight between cross-links of waxy maize starches using the theory of rubber elasticity', J. Rheol., 41, 129±139. HAMMERSTRAND G E, HOFREITER B T, and MEHLTREITER C L (1960), `Determination of the extent of reaction between epichlorohydrin and starch', Cereal Chem., 37, 519± 524. HOFREITER B T, MEHLTREITER C L, BENNIE J, and HAMMERSTRAND G E (1960), `Starch products of stable viscosity', US patent 2,929,811. IKAWA Y, GLOVER D V, SUGIMOTO Y, and FUWA H (1978), `Amylose percentage and distribution of unit chain-length of maize starches having a specific background', Carbohydr. Res., 61, 211±216. DOEHLERT D C,

ß 2004, Woodhead Publishing Limited

and FUWA H (1981), `Some structural characteristics of starches of maize having a specific genetic background', Starch, 33, 9±13. INOUCHI N, GLOVER D V, TAKAYA T, and FUWA H (1983), `Development changes in fine structure of starches of several endosperm mutants of maize', Starch, 35, 371±376. INOUCHI N, GLOVER D V, SUGIMOTO Y, and FUWA H (1984), `Developmental changes in starch properties of several endosperm mutants of maize', Starch, 36, 8±12. INOUCHI N, GLOVER D V, and FUWA H (1987), `Chain length distribution of amylopectins of several single mutants and the normal counterpart, and sugary-1 phytoglycogen in maize (Zea mays L.)', Starch, 39, 259±266. INOUCHI N, GLOVER D V, SUGIMOTO Y, and FUWA H (1991), `DSC characteristics of retrograded starches of single-, double-, and triple-mutants and their normal counterpart in the inbred Oh43 maize (Zea mays L.) background', Starch, 43, 473± 47. ISPAS-SZABO P, RAVENELLE F, HASSAN I, PREDA M, and MATEESCU M L, (2000), `Structureproperties relationship in cross-linked high-amylose starch for use in controlled drug release', Carbohydr. Res., 323, 163±175. JAMES M G, ROBERTSON D S, and MYERS A M (1995), `Characterization of the maize gene sugary-1, a determinant of starch composition in kernels', The Plant Cell, 7, 417± 429. JANE J, XU A, RADOSAVLJEVIC M, and SEIB P A, (1992), `Location of Amylose in Normal Starch Granules. I. Susceptibility of Amylose and Amylopectin to Cross-Linking Reagents', Cereal Chem., 69, 404± 409. JAROWENKO W (1986), `Acetylated starch and miscellaneous organic esters', in Wurzburg O B, Modified Starches: Properties and Use, Boca Raton, FL, CRC Press, 54±77. KASEMSUMAN T, JANE J L, SCHNABLE P, STINARD P, and ROBERTSON D, (1995), `Characterization of the dominant mutant amylose-extender (Ae1±5180) maize starch', Cereal Chem., 72, 457±464. KATZBACK W (1972), `Phosphate cross-bonded waxy corn starches solve many food application problems', Food Technol., 4, 32±36. KEPPEL C (2001), A study of glucose storage polymers: teosinte starch, starch crystallinity, and cyanobacterial glycogen, MS. Thesis. Ames, IA, Iowa State University. KLUCINEC J D, and THOMPSON D B (1998), `Fractionation of high-amylose maize starches by differential alcohol precipitation and chromatography of the fractions', Cereal Chem., 75, 887±896. KOBAYASHI S, SCHWARTZ S J, and LINEBACK D R (1986), `Comparison of the structures of amylopectin from different wheat varieties', Cereal Chem., 63, 71±74. KOCH V H, BOMMER H D, and KOPPERS J (1982), `Analytical investigations on phosphate cross-linked starches', Starch, 34, 16±21. KONIGSBURG M (1950), `Ungelatinized starch ethers from polyfunctional etherifying agents', US patent 2,500,950. KRAMER H H, PFAHLER P L, and WHISTLER R L (1958), `Gene interaction in maize affecting endosperm properties', Agron. J., 50, 207±210. KRUEGER B R, WALKER C E, KNUTSON C A, and INGLETT G E (1987), `Differential scanning calorimetry of raw and annealed starch isolated from normal and mutant maize genotypes', Cereal Chem., 64, 187±190. LANSKY S, KOOI M, and SCHOCH T J (1949), `Properties of various fractions of different Starches', J. Am. Chem. Soc., 71, 4066±4075. LELOUP V M, COLONNA P, and BULEON A (1991), `Influence of amylose-amylopectin ratio IKAWA Y, GLOVER D V, SUGIMOTO Y,

ß 2004, Woodhead Publishing Limited

gel properties', J. Cereal Sci., 13, 1±13. and CORKE H (1999), `Physicochemical properties of maize starches expressing dull and sugary-2 mutants in different genetic backgrounds', J. Agric. Food Chem., 47, 4939±4943. LIM S and SEIB P A (1993a), `Preparation and pasting properties of wheat and corn starch phosphates', Cereal Chem., 70, 137±144. LIM H and SEIB P A (1993b), `Location of phosphate esters in a wheat starch phosphate by p-nuclear magnetic resonance spectroscopy', Cereal Chem., 70, 144±152. LIU H, RAMSDEN L, and CORKE H (1999), `Physical properties and enzymatic digestibility of phosphorylated ae, wx, and normal maize starch prepared at different pH levels', Cereal Chem., 76, 938±943. LLOYD N E (1970), `Starch esters', US patent 3,539,551. LU T-J, JANE J-L, KEELING P L, and SINGLETARY G W (1996), `Maize starch fine structures affected by ear developmental temperature' Carbohydr. Res., 282, 157±170. MATHESON N K (1975), `The alpha (1-4)(1-6) glucans from sweet and normal corns', Phytochemistry, 14, 2017±2021. MCPHERSON A E and JANE J (2000), `Extrusion of cross-linked hydroxypropylated corn starches II. Morphological and molecular characterization', Cereal Chem., 77, 326±332. MERCIER C (1973), `The fine structure of corn starches of various amylose-percentage: waxy, normal, and amylomaize', Starch, 25, 78±82. MONTGOMERY E M, SEXTON K R, DIMLER R J, and SENTI F R (1964), `Physical properties and chemical structure of high-amylose corn starch fractions', Starch, 16, 314±318. MORRISON W R and TESTER R F (1991), `Chemical and physical factors that affect cereal starches', in Martin D J and Wrigley C W, Cereals International, Melbourne, Royal Australian Chem. Inst., 134±138. NELSON O and PAN D (1995), `Starch synthesis in maize endosperms', Annu. Rev. Plant Mol. Biol., 46, 474±496. NG K Y, DUVICK S A, and WHITE P J (1997), `Thermal properties of starch from selected maize (Zea mays L.) mutants during development', Cereal Chem., 74, 288±292. NINOMYA Y, OKUNO K, GLOVER D V, and FUWA H (1989), `Some properties of starches of sugary-1; brittle-1 maize (Zea mays L.)', Starch, 41, 164±167. OTT M and HESTER E E (1965), `Gel formation as related to concentration of amylose and degree of starch swelling', Cereal Chem., 42, 476±484. PAN D and NELSON O E (1984), `A debranching enzyme deficiency in endosperms of the sugary-1 mutants of maize', Plant Physiol., 74, 324±328. PERERA C, LU Z, SELL J, and JANE J L (2001), `Comparison of physicochemical properties and structures of sugary-2 cornstarch with normal and waxy cultivars', Cereal Chem., 78, 249±256. PERLIN A S (1958), `Radiochemical evidence for heterogeneity in wheat starch', Can. J. Chem., 36, 810±813. PFAHLER P L, KRAMER H H, and WHISTLER R L (1957), `Effect of genes on birefringence endpoint temperature of starch grains in maize', Science, 125, 441±442. POLLAK L M and WHITE P J (1997), `Thermal starch properties in corn belt and exotic corn inbred lines and their crosses', Cereal Chem., 74, 412±416. RAHMAN A, WONG K S, JANE J L, MYERS A M, and JAMES M G (1998), `Characterization of SU1 isoamylase, a determinant of storage starch structure in maize', Plant Physiol., 117, 424±435. RAVENELLE F, MARCHESSAULT R H, LEGARE A, and BUSCHMANN M D, (2002), `Mechanical LI J

ß 2004, Woodhead Publishing Limited

properties and structure of swollen crosslinked high amylase starch tablets', Carbohydr. Polym., 47, 259±266. REEVES R.G (1950), `The use of teosinte in the improvement of corn inbreds', J. Agron., 42, 248±251. RING S G and STAINBY G J (1985), `A simple method for determining the shear modulus of food dispersions and gels', J. Sci. Food Agric., 36, 607±613. SANDERS E B, THOMPSON D B, and BOYER C D (1990), `Thermal behavior during gelatinization and amylopectin fine structure for selected maize genotypes as expressed in four inbred lines', Cereal Chem., 67, 594±602. SANDSTEDT R M, STRAHAN D, UEDA S, and ABBOT R C (1962), `The digestibility of highamylose corn starches compared to that of other starches: The apparent effect of the ae gene on the susceptibility to amylase action', Cereal Chem., 39, 123±131. SHANNON J C and GARWOOD D L (1984), `Genetics and physiology of starch development', in Whistler R L, Be Miller J N, and Pachall E F, Starch: Chemistry and Technology, Orlando, Academic Press, 26±86. SHI X and BE MILLER J N (2002), `Aqueous leaching of derivatized amylose from hydroxypropylated common corn starch granules, Starch, 54, 16±19. STEENEKEN P A M (1989), `Rheological properties of aqueous suspensions of swollen starch granules', Carbohydr. Polym., 11, 23±42. SUMMER J B and SOMERS G F (1944), `The water soluble polysaccharides of sweet corn', Arch. Biochem., 4, 7±9. TAKEDA C, TAKEDA Y, and HIZUKURI S (1993), `Structure of the amylopectin fraction of amylomaize', Carbohydr. Res., 246, 273±281. TAKEDA Y and PREISS J (1993), `Structures of B90 (sugary) and W64A (normal) maize starches', Carbohydr. Res., 240, 264±275. TESSLER M M and WURZBURG O B (1983), `Starch sulfomaleate half ester and their use toprepare starch disulfosuccinate half ester', US patent US 4379 919 (Cited from Chem. Abstr., 1983, 98, 217532 h). THOMPSON D B (2000), `On the non-random nature of amylopectin branching', Carbohydr. Polym., 43, 223±239. TOMASIK P, WANG Y-J, and JANE J L (1995), `Facile route to anionic starches. succinylation, maleination and phthalation of corn starch on extrusion', Starch, 47, 96±99. TRUBIANO P C (1987), `Succinate and substituted succinate derivatives of starch', in Wurzburg O B, Modified starch: Properties and uses, Boca Raton, FL, CRC Press, 131±148. TZIOTIS A (2001), Characterization of starch fractions from maize endosperm mutants, MS. Thesis. Ames, IA, Iowa State University. VASANTHAN T and HOOVER R (1992). `Effect of defatting on starch, starch structure and physicochemical properties', Food Chem., 45, 337±347. WANG Y J and WANG L (2000), `Effects of modification sequence on structures and properties of hydroxyproylated and crosslinked waxy maize starch', Starch, 52, 406±412. WANG Y J and WANG L (2002), `Characterization of acetylated waxy maize starches catalyzed by different alkalis', Starch, 54, 24±30. WANG Y J, WHITE P J, and POLLAK L (1992), `Thermal and gelling properties of maize mutants from the Oh43 inbred line', Cereal Chem., 69, 328±334. WANG Y J, WHITE P J, POLLAK L, and JANE J L (1993), `Amylopectin and intermediate materials in starches from mutant genotypes of the Oh43 inbred line', Cereal Chem., 70, 521±525.

ß 2004, Woodhead Publishing Limited

and BE MILLER J N (1997), `Starch', in Whistler R L and Be Miller J N, Carbohydrate Chemistry for Food Scientists, St Paul MN, Eagan Press, 117±151. Whistler R L and Daniel J R (1984), `Molecular structure of starch', in Whistler R L, Be Miller J N, and Paschall E F, Starch: Chemistry and Technology, Orlando, Academic Press, 153±182. WHISTLER R L and DOANE W M (1961), `Characterization of intermediate fractions of highamylose corn starches', Cereal Chem., 38, 251±255. WHITE P J, ABBAS I R, and JOHNSON L A (1989), `Freeze-thaw stability and refrigeratedstorage retrogradation of starches', Starch, 41, 176±180. WHITE P J, POLLAK L M, and BURKHART S (1991), `Thermal properties of starches from corn grown in temperate and tropical environments'. (Abstr.) Cereal Foods World 36:524. WHITE P J, POLLAK L M, and JOHNSON L A (1994), `Starch thickened acidic foodstuffs and method of preparation', US patent 5,356,655. WILKINS M R, WANG P, XU L, NIU Y, TUMBLESON M E, and RAUSCH K D (2003a), `Variability in starch acetylation efficiency from commercial waxy corn hybrids', Cereal Chem., 80, 68±71. WILKINS M R, WANG P, XU L, NIU Y, TUMBLESON M E, and RAUSCH K D (2003b), `Variability of reaction efficiencies and pasting properties of acetylated dent corn starch from various commercial hybrids', Cereal Chem., 80, 72±75. WOLF I A, HOFREITER B T, WATSON P R, DEATHEARAGE W L, and MACMASTERS J E (1955), `The structure of a new starch of high amylose content', J. Amer. Chem. Soc., 77, 1654±1659. WURZBURG O B and SZYMANSKI C D (1970), `Modified starches in the food industry', J. Agric Food Chem., 18, 997±1001. YEH J Y, GARWOOD D L, and SHANNON J C (1981), `Characterization of starch from maize endosperm mutants', Starch, 33, 222±230. YUAN R C, THOMPSON D B, and BOYER C D (1993), `Fine structure of amylopectin in relation to gelatinization and retrogradation behavior of maize starches from three wxcontaining genotypes in two inbred lines', Cereal Chem., 70, 81±89. YUN S H and MATHESON N K (1993), `Structures of the amylopectins of waxy, normal, amylose-extender, and wx:ae genotypes and of the phytoglycogen of maize', Carbohydr. Res., 243, 307±321. WHISTLER R L

ß 2004, Woodhead Publishing Limited

11 Tropical sources of starch S. N. Moorthy, Central Tuber Crops Research Institute, India

11.1 Introduction: tropical sources of starch The tropical belt, which covers around 40% of the total land area encompassing five continents and many countries, harbours a number of starch bearing crops that include cereals, tree, fruit and vegetable crops and most important the root crops.1±8 However, commercial use of these for starch extraction has been limited to a few of these crops. Among the non-cereal sources, the most important ones are sago starch from sago palm, potato, cassava and sweet potato starches from the corresponding tubers. Minor quantities of starch are extracted from other crops such as screw pine fruits, and the tuber crops like colocasia, amorphophallus, yams, arrowroot, Canna and Curcuma sp. but they have no commercial importance. Among these different starches, only cassava and sweet potato starches have been studied in detail and this chapter tries to bring out the available information on the tropical starches and also possible avenues of utilisation based on their physicochemical and functional characteristics. The starches and their properties are dealt with under different sections

11.1.1 Tree crops Among the starch bearing tree crops, the most important ones are sago palm, mango, plantain, jackfruit, breadfruit and screw pine. The starch is found either in the stem, fruit or seed. Sago palm (Metroxylon sagu) is a nonbranching palm cultivated in South East Asia. It grows to 9±12 metres high and flowers after 10± 15 years and then dies. At the time of flowering, the palms are felled, the stems are sliced the pith is rasped, sieved and the starch granules are allowed to settle. The starch is collected, dried and used to make sago pearls by granulation. Each

ß 2004, Woodhead Publishing Limited

palm can give up to 250 kg starch. Originally, all the sago pearls were obtained exclusively from sago starch, but now the sago pearls are obtained from cassava starch which is more abundantly available. Mango (Mangifera indica) is widely grown in the tropics for its delicious fruits. The trees grow up to 40 metres in height and one tree can yield 8000±30,000 fruits per year. The seeds which form around 20±40% of the fruit, contain up to 30% starch. The starch can be extracted by crushing the seeds, filtering and allowing to settle. Very little work has been carried out on this starch. Jackfruit (Artocarpus integra) is a large tropical fruit tree with a dense crown and reaches up to 50 metres in height. The large fruits contain a fleshy portion covering oval shaped seeds. The seeds form around 5% of the fruit and contain nearly 30% starch. Often the seeds are roasted and eaten. The starch properties have not been reported. Breadfruit (Artocarpus altilis) is a tall tree 30 metres high and branching. It yields about 600±700 medium size round fruits per year. They are consumed as a vegetable and have a very delicious taste. The fruits contain around 20% starch. Data on the starch properties is limited. Screw pine (Pandanus leram) is a tall slender palm, 12±18 metres in height found in various tropical countries and particularly grown widely in the Andaman and Nicobar islands of the Indian Ocean where it is very common. Its fruit (10±18 kg) is fleshy and contains nearly 20% starch. The local population consume the tubers after cooking and also extract starch by crushing the fruits, sieving and drying. The starch has not been studied.

11.1.2 Cereals Maize, rice, sorghum and a number of other minor millets like ragi, bajra, etc., are cultivated in the tropics and form important crops in many areas. They serve as a cheap source of food. Most of them contain a good quantity of starch, the most important being maize. Maize (Zea Mays) and rice (Oryza sativa) are annual herbs grown in large areas of the tropical and subtropical belt and form a staple food. The yield of the crops varies widely between 5±29 t/ha and the starch content is as high as 70% on a dry weight basis. The extraction of starch from maize has been well standardised and is carried out extensively and used in various foods and industries. Extraction of starch from rice is less common. The starches have been characterised thoroughly and dealt with in other chapters. Sorghum (Sorghum bicolor) and other millets are grown widely in the tropics and considered as poor man's cereal. The yield of grain varies from 1±3 t/ha and the starch content is quite high in the grains (75%), but very little work has been carried out on its starch properties. Tef (Eragrostis tef ) is an annual branching grass and serves as a major cereal staple in Ethiopia and its flour is used for making various products. The starch content in the grains is similar to other cereals (60%) and can be easily extracted using water.

ß 2004, Woodhead Publishing Limited

11.1.3 Herbs/shrubs There are a number of herbs/shrubs which produce starch in their grains, fruits or seeds. But these have not been exploited as a starch source. Among them a few important ones are outlined below. The Amaranthus (Amaranthus sp.) family includes more than 60 species that grow in many tropical regions. Although most of them produce edible leaves, a few species are cultivated for their grains. The chief ones are A. hypochondriacus, A. caudatus, and A. edulis. They are perennial herbs widely differing in height and appearance. The grain yield is normally around 3±4 t/ha. The grain contains 40±50% starch mainly waxy starch. The small size of the granules and low amylose content in the starch make them very valuable as a type of starch with special food applications. Plantain/banana (Musa sp.) is grown extensively in most parts of the tropics for its fleshy fruits and there are a number of species having widely varying taste and texture. It is a straight annual yielding 10±20 t/ha of fruits. The unripe fruits contain nearly 35% starch which has rarely been extracted. Okenia (Okenia hypogeae) is a perennial herbaceous crop distributed in central and south America. The seed yield is up to 2 t/ha and the seeds contain nearly 35% starch. Quinoa (Chenopodium quinoa) is the ancient Andes crop widely grown in S. America, produces grain which serves as an important food source. These are perennial herbs yielding 1±2 t/ha of grains which contain approximately 40±50% starch.

11.1.4 Pulses A large number of pulses are grown for food uses in most tropical countries and give a yield of 2±5 t/ha. The pulses also contain starch to a good (20±30%) extent. However, they are never obtained pure due to the presence of lipids and proteins accompanying the starch. The extraction of starches from pulses is rarely reported except for some isolated reports on the study of their starches. A few of these are given below. · Chick pea (Cicer arietinum) is an annual branching leguminous herb producing seed bearing pods. The seeds are used as food in most parts of the tropic and contain 40±50% starch. · Cow pea (Vigna sinensis) is a leguminous crop having long trailing vines producing seeds of black, white or red colour used widely in various food preparations. The starch content is over 45%. · Horse gram. (Dolichos biflorus) is a branching suberect annual bearing flat pods containing 5±6 seeds which contain around 45% starch. · Lima bean (Phaseolus lunatus) is a perennial/annual legume producing thin long pods which bear seeds (2±3 t/ha) containing nearly 55±60% starch. · Velvet beans (Mucuna pruriens) is a trailing vine yielding seeds which vary in colour and shape. The seed yield is 1 t/ha and has a starch content of 51.5%.

ß 2004, Woodhead Publishing Limited

11.1.5 Tropical root crops Tropical root and tuber crops are important food crops serving either as subsidiary or subsistence food in different parts of the tropical belt. They are rich sources of starch9±12 besides many vitamins, minerals, etc. Although there has been some decline in their use as food, their industrial application, especially that of cassava, is making rapid advances. Cassava and to a small extent, sweet potato (Ipomoea batatas Lam.) are used for starch extraction in countries like India, Brazil, Thailand, Indonesia, Philippines and China. Studies at different laboratories have brought to light the wide diversity in the starch characteristics of tuber crops and the possibility of using these native starches instead of chemically modified starches.12,13 Cassava (Manihot esculenta Crantz) is a sturdy perennial crop grown in many parts of Asia, Africa and South America. The yield of the crop is normally around 20 t/ha and the starch yield forms nearly one quarter of the total yield. The starch has been studied in detail and finds use in a large range of industries. Sweet potato (Ipomoea batatas Lam.) is a herbaceous perennial vine and is grown extensively in the tropics and also in some parts of the USA for its tubers. The tubers have different sizes, shapes and colour. The yield varies anywhere from 8±30 t/ha and it has been possible to have three crops in a year thus giving a very high annual starch yield. The starch content in the fresh tubers varies from 12±30%. Taro (Colocasia esculenta) is a small herbaceous plant with large leaves found in most parts of the tropics and is very important in the pacific regions. The crop is harvested at 8±10 month stage and produces a number of cormels around a corm. The yield is 5±10 t/ha and starch content in the tubers is 12±20%. Tannia (Xanthosoma sagittifolium) is a large herbaceous plant grown widely for its cormels which are much larger than taro cormels. The yield is around 10± 25 t/ha and starch content in the tubers nearly 20%. Elephant Foot Yam (Amorphophallus paeoniifolius) is grown extensively for its huge corms. It is a big perennial herb harvested after one year and the corms can weigh over 15 kg. The yield is over 20 t/ha and starch content in the tubers is approximately 20%. Yams comprise a large genus with over 600 species from which a few are more commonly cultivated. Most of them are trailers. The tubers are harvested at 8±12 months after planting and the tubers especially those of D. alata, are very large. Some of the species produce aerial tubers also. The yield of the crop and starch yield also vary considerably among the different species and among these D. rotundata tubers have the highest starch content. In addition, there are a number of other minor tuber and root crops which contain starch, but their utilisation is limited. African yam bean (Sphenostylis stenocarpa) belongs to Leguminosae. It is a vigorous herbaceous climbing vine reaching 1.5±2 m in height producing pods as well as small spindle shaped tubers about 5±8 cm long similar to sweet potato. The crop is found mainly in Africa yielding up to 4 t/ha. The tubers are rich in starch (25%) but there are no studies on this starch.

ß 2004, Woodhead Publishing Limited

Arracacha, Peruvian carrot, (Arracacia xanthorrhiza) is a stout semicaulescent herb, resembling celery grown in south America and parts of Africa mainly at high altitudes. The edible secondary tubers are usually 6±10 in number with a yield of 3±18 t/ha. The tubers which contain nearly 20% starch, are used as a source of edible starch. Chinese water chestnut (Eleocharis dulcis) is a variable annual stout aquatic plant producing corms found in the Asian region. The yield is 20±40 t/ha and starch content is around 7%. In China the starch is extracted by rasping and settling. East Indian arrowroot (Tacca leontopetaloides) is a perennial herb with a tuberous rhizome producing large/small sized tubers, found in tropical areas. The tubers are similar to potato are harvested after 8±10 months and can weigh up to 1 kg. They are used in Tahiti to make `poi' a traditional food. The tubers are rich in starch (20±30%) and it is extracted and used widely. Giant taro (Alocasia macrorrhiza) is a tall succulent herbaceous plant up to 4.5 metres in height producing big corms up to 18 kg. It is grown in Asia and South America and harvested 10±12 months after planting. The tubers contain 17±25% starch. Coleus (Plectranthes rotundifolius) is a small herbaceous annual 15±30 cm high found in Africa and Asia. At maturity (6 months) they yield round to oval tubers that are very prized for their delicate flavour. The yield is 7±15 t/ha and the starch content is nearly 15%. Kudzu (Pueraria lobata), known as arrowroot vine, is a small perennial twining herb or shrub with elongated tuberous roots often weighing up to 40 kg and yield of 5±7 t/ha. Roots are starchy, 30±60 cm long and used as a source of edible starch in place of arrowroot starch or gelatine in many foods. The starch content is over 20% and is extracted on a small scale in Japan but the starch has not been studied. Lotus root (Nelumbo nucifera) is a perennial aquatic herb, rooting in mud found in South and South-east Asia, Africa and Australia. The white globulous rhizomes which are harvested at 6±9 months measure 60±120 cm in length. The root yield is 5 t/ha and starch content is nearly 18%. In China, a fine starch is isolated from the rhizomes. Oca (Oxalis tuberosa) is a small compact annual tuberous herb 20±30 cm high. Oca is a ancient food plant of the Andes and found in many parts of South America. The rhizomatous tubers are harvested at eight months maturity and are similar to potato (5±8 cm in length). The tuber yield is 4±5 t/ha while the starch content is 12%. Queensland arrowroot (Canna sp) is perennial herbaceous monocotyledon found in many parts of Asia, Africa and South America. The shape of the rhizomes varies from cylindrical to tapering and 5±9 cm in size. The tuber yield is 15±40 t/ha and the starch content varies from 24±30%. Shoti (Curcuma zedoaria) known as Indian arrowroot, is a robust perennial with a fleshy branching rhizomes cultivated in Asia. The starchy finger shaped rhizomes are greyish in colour, grow to 15 cm in length and have a musky odour.

ß 2004, Woodhead Publishing Limited

The tuber yield is 8±12 t/ha and starch content is 12±15%. The starch is extracted by rasping the tubers, sieving and settling and serves as a source of easily digested starch. Swamp taro (Cyrtosperma chamiossonis) is a giant herbaceous perennial 3±4 m in height with huge leaves. It is cultivated in many parts of Asia, Africa and the Pacific islands. The yield of corms is 7±10 t/ha and the starch content 28± 30%. Winged bean (Psophocarpus tetragonolobus) is a leguminous, climbing perennial found in Asia and Africa. Tubers are obtained 5±8 months after planting with a yield of 2±6 t/ha. Root tubers are 5±12 cm in length and contain nearly 20% starch.

11.1.6 Other minor sources In addition there are a few other sources of starch which have not been exploited. They include Bamboo (Guadua flabellata) which is a perennial crop grown largely as a source of cellulose. However the culm of the crop has nearly 20% starch which can be easily extracted and can serve as a source of starch. Black pepper (Piper nigrum), termed king of spices, is grown in Asia and Africa and is a branching climbing perennial shrub. The annual yield of dry pepper is 100±300 kg/ha and the seeds contain over 40% starch but it has not found any application. Buffalo Gourd (Cucurbita foetidissima) produces roots which can be very large ± even up to 75 kg. The roots contain 15% starch and can be used a source of starch.

11.2

Characteristics and properties of cassava starch

Among the vast sources of starch outlined above, only cassava and maize starches have been commercially exploited for some time and continue to be major sources of starch. The extraction of maize starch is slightly complicated because of the need for steeping the dried cobs for facilitating starch extraction and use of whitening agents. On the other hand, cassava starch is easily extractable since the tubers contain a very low quantity of proteins, fats, etc. Hence the extraction process is simple and the starch obtained is pure white in colour if the process is carried out properly. Since the lipid content in the starch is very little (4.5 b. 25 ft. hot c. vacuum f. _______________

ß 2004, Woodhead Publishing Limited

Table 13.1 Continued 12.

Will the process involve one or more of the following? a. hot filling e. blast or quick freezing b. ambient filling f. re-heating (reconstitution) c. refrigeration g. steam tables d. slow freezing h. _________________

13.

If fat is used, what type? a. liquid b. solid c. shortening

d. lard e. __________

14.

Are salts utilized? a. what type b. percent c. blends d. _____________

15.

Are other hydrocolloids used? a. what type b. blends c. ______________

16.

Is the final product a dry mix?

17.

If yes to 16, what does the process involve? a. blending c. extruding b. agglomerating d. _____________

18.

Is moisture content of the added starch critical?

19.

If yes to 18, what is the anticipated packaging? a. paper d. heat sealed b. wax coated e. ______________ c. poly-lined

20.

If not a dry mix, what is the anticipated packaging? a. glass e. paper b. can f. plastic c. pouch g. tote d. drums h. ______________

21. Is the added starch to be used in more than one product? If yes, consider similar events for the other products and identify the critical functions contributed by the starch. 22.

How important is ingredient economics? a. very c. quality dependent b. not very d. _______________

ß 2004, Woodhead Publishing Limited

Table 13.2

Consider the impact of processing conditions on viscosity stability

Equipment

Time

Degree of shear

Kettle Swept surface Plate heat exchanger Jet cooker low back pressure high back pressure PDPPositive Displacement Centrifugal pump Piping Emulsification Colloid mill Homogenization

long intermediate short

low moderate ÿ! high high

short short short short variable

low high low high variable

short very short

high very high

Pump

considered. Both of these are essential for providing a fixed set of conditions that can be monitored and repeated. Knowing the rates for both conditions will also help with identifying the necessary equipment for scaling the system for commercialization.

13.5.2 Shear Earlier, shear was described for its degrading effect upon starch. To know about equipment and its shear potential is very helpful to a food technologist. Knowing when and where shear can and cannot be introduced is critical also. Many of the modified starches designed for food processing can withstand some shear. A few are specifically engineered to withstand very high shear systems. In all cases though, starch is much more shear stable at ambient or cool temperatures as compared to elevated temperatures. Studies have shown that below 57.2 ëC starch solutions can be processed with little shear damage. Starch in solution prior to gelatinization is very shear stable.

13.5.3 Packaging One mistake common to many products prepared with starch is that of packaging. Too many times the same product is packaged in a variety of containers. This in most cases is not a reliable practice. Many starch stabilized food products cannot be packaged over a large volume range of filled quantities. An example would be hot fudge topping. Utilizing a single correctly modified starch to achieve viscosity, texture and stability is possible for product packaged in containers from one pint (half-liter) to five gallons (20 liters). However, to take this same product, an attempt to package it in 55-gallon drums (110 liter) will most likely be unsuccessful. Experience has shown that to facilitate this requires the use of two modified starches or a starch/gum blend. In both cases

ß 2004, Woodhead Publishing Limited

starch levels had to be adjusted and some characteristic differences were noted in the finished topping.

13.5.4 Storage All too often the storage of product is overlooked when initiating product development. Just as important as the container, are the storage conditions. In addition to the conditions, the time of storage is critical. Some foods can be stored much longer than others under the same conditions. Changes usually occur in texture during storage. Some products thin, others thicken. Others could become lumpy, grainy or eventually separate. It is essential to regulate the warehouse, cooler or freezer conditions. The food product should also be pretested in as many storage conditions as possible; one never knows what the consumer or transportation department will do.

13.5.5 Water Water may be the most common ingredient utilized with starch. It is also one of the most underestimated ingredients to affect starch negatively. Water as a food ingredient has been emphasized as essential to make starch, native or modified multi-functional in food products. This being true, we must consider also the differences in water quality around the globe. Water obtained from a well has certain measurable properties and constituents. The same can be said for water taken from a city water system. Unfortunately, these two waters are significantly different and could dramatically alter the functional properties of added starch. Consider the treatment facility for any city and the number of cities around the world. We now have several approved treatments for the purification of water and not many cities use exactly the same process, let alone similar levels of reagents within a process. Regulatory guidelines leave a broad range, providing room for variability as to the effect on starches. Experience has proven that if a water supply smells like a swimming pool it probably has enough chlorine present to oxidize starch. Also, well water systems can offer a vast variety in mineral content. Some of these can lead to starch complexes that inhibit the functionality of the added starch. Water softened with salts may be great to bathe or wash dishes, but have the potential to negatively affect the added starch to your food system. Because of these differences, many development labs formulate with distilled or de-ionized water. Unfortunately, as soon as a product is taken to the plant for production, a complete reformulation is necessary. So, what is thought to be the simplest ingredient for foods can still have a very complex variety of components.

13.5.6 Sweeteners Some facts have already been established about sweeteners. They offer a lot more to the food product than just sweetness. Sweeteners for this review will

ß 2004, Woodhead Publishing Limited

include maltodextrins, sweetener solids greater than 20 DE, high fructose syrup, syrups classified as medium and high conversion, as well as sucrose. The high intensity sweeteners contribute sweetness in foods, but are predominately carried by maltodextrins. Therefore, I will forgo discussion about their affect, as it relates primarily to that of maltodextrins. When utilizing maltodextrins with starch, the food technologist is attempting first, to increase solids or secondly, control sweetness or thirdly, generate a desired color. In some instances a maltodextrin may be added to assist with stickiness. This is best done with very low (< 5 DE) maltodextrin products. These same maltodextrins have proven to provide fat-like properties. Because of the blandness and solubility of maltodextrins, they are very compatible with the incorporation of instant starches. They conveniently offer a method for premixing and not affecting the functionality of the added starch. In fact, the low DE maltodextrins are somewhat similar to modified starches processed with enzyme. Sweetener solids possessing a DE greater than 20 are not classified as maltodextrins. In fact, those derived from maize starches are called corn syrup solids. Similar products are commercially available today, derived from other starch sources, i.e., rice, tapioca, potato, etc. They now represent a small but reasonable share of the dry solids and liquid sweetener market. A primary functional characteristic is that of adding sweetness to food products differing from that of maltodextrins. However, one similar characteristic is that the higher DE sweetener solids are also excellent humectants. Water activity can be significantly reduced through the utilization of low DE sweetener solids.

13.5.7 Salts Most technologists consider this ingredient to be predominately sodium chloride. For many food systems this may be correct. However, salts can complex with a number of compounds, thus creating unique structures, resulting in unique eating qualities. An example of such a complex is the preparation of instant puddings. Many food scientists as well as consumers consider the starch matrix as the building block for producing the set and eating quality for instant puddings. However, the starch is functioning as the control mechanism for the water phase, will enhance the mouth-feel and provide stability during storage. The structure and eating quality is generated from added salts forming complexes with the protein, typically the dairy portion of the mixture. This protein could come from added liquid or dry milk solids. Other sources of proteins can influence the texture and eating quality, such as proteins from soy, cereal grains or extracts. Quantity and blend ratio of the added salts, influence the rate of set, strength of the gel and the mouth-feel of the finished pudding. Calcium and potassium salts, either phosphate or chloride derivatives significantly add to or detract from the quality of the puddings. Calcium usually increases gel strength and shortens set time, whereas potassium has the reverse effect. The chloride derivatives generally produce a distinct off flavor

ß 2004, Woodhead Publishing Limited

and this is considered unacceptable for commercial products. However, the effect is worth noting.

13.5.8 Other food ingredients (spices, fruits, flavors, etc.) Many food ingredients can and usually do contribute to soluble solids of the food matrix. This then adds ingredients that compete for water. Ingredients competing for the water typically retard the hydration of the added starch. However, many ingredients are added for flavor, particulate characteristics or enhancement of other functional ingredients. Many times these unique additives contain alpha-amylase. Alpha-amylase is an enzyme utilized to degrade starches for the production of sweeteners. It is also the type of enzyme utilized by the body to break down carbohydrates. Starch is a carbohydrate and therefore will be digested by this enzyme. Peppers, some smoke flavors, fruits such as pineapple and fresh blueberries are known to contain significant quantities of alpha-amylase. Dairy products such as cheese (blue cheese) and low heat processed milk solids can be a source. In the early 1990s the flour industry experienced a significant setback with flour. For a two-year period the wheat crop experienced a very wet season and a second growth occurred within the crop prior to harvest. Not only were protein related enzymes formed, but so was alpha-amylase. Many bakery products experienced breakdown of the starch phase of the flour as well as the digestion of added starches. In many products the effect was not noticed until the product reached the consumer (refrigerated and frozen dough). There are test procedures available, as well as test kits to assist with analysis. Many contract laboratories assay for enzyme presence. A simple non-quantitative test can be utilized using triplicate containers of a control starch paste and inoculating with suspect materials. It requires only 24 hours at room temperature to study the effect, or at elevated temperatures (60 ëC) about 4±6 hours yields an indication of enzyme activity. As with all analytical testing, it is recommended to use a control sample for comparison.

13.5.9 Proteins and other starches Ingredients that contain proteins typically contain carbohydrates, therefore always consider the source of the protein and what you anticipate its contribution to be to the food product. Proteins, as mentioned earlier, complex very easily with added salts. These complexes can create gray to dark particulates. These particulates can create undesirable textures, flavor, mouthfeel and appearance. In foods that will be fried, baked or subjected to long-term processing at high temperatures, can potentially produce dark spots on the surface (Maillard reaction). Protein can also contribute to texture. In breads it may be a desirable characteristic, however the elastic effect of gluten-like dough in some foods is not acceptable. Agglomerated proteins generate small to large particulates that can create uneven surface texture as well as a gray color.

ß 2004, Woodhead Publishing Limited

Starch added from other sources or as another ingredient not intended to yield functional properties is often overlooked. One source that can contribute a significant percentage of starch, and has already been mentioned, is that found in powdered sugar. The 3% starch allowed in powdered sugar (North America) could significantly alter the cook, flavor, clarity, texture and shelf-life of a food product. When a technologist has identified all sources of starch or other hydrophilic compounds prior to formulating the final product, a relative amount of correctly modified starch can be added allowing for the affect from the other ingredients. In many cases the technologist may need to seek advice from suppliers as to ingredient composition and potential functionality when mixed with other ingredients. Some ingredients utilized can be pre-blended compounds from many sources. A supplier may not quantify the amounts, and in some instances, not disclose the detailed listing of ingredients. An example of such a mixture could be a blend of common and waxy corn starches. Suppliers of blended maize systems can identify those mixtures as containing `corn starch'. The supplier would be correct in doing so.

13.6 Using the functional properties of starch to enhance food products As noted earlier in section 13.4 most examples were generic in description. This section will attempt to identify the type of starch, modifications and possible levels of use for a few specific food processes.

13.6.1 Thermal processing (canned-jarred; retorted-sterilized-hot filled) Thermal processing can involve severe heat treatment. Typically a food technologist would recommend a modified starch possessing the treatment of mono-substitution in conjunction with moderate to high levels of cross-linking. These modifications will provide stability, texture and consistency. Processes such as retorting or heat-sterilization require such starches. The correct starch will depend upon other functional attributes, such as flavor, length of shelf life and appearance. Usually a modified common or waxy corn is a good choice to initiate studies. In the high acid or low pH foods, heat treatment is usually minimal. A cross-linked only starch could be the best choice. However, monosubstitution could be necessary, dependent upon desired product characteristics and storage, i.e., refrigeration and/or freezing (use ranges from 1.5±6.0%). Starches that possess unique characteristics can offer process advantages in thermally processed foods. Native or specifically modified starches offer thick to thin viscosity profiles, aid with processing and retaining particulate integrity. The thick to thin concept is limited in starch choice and utilization. This concept is generally for food preparations that either require specific particulate addition to containers prior to processing or a very critical viscosity profile post

ß 2004, Woodhead Publishing Limited

processing. Native potato starch is a natural thick to thin starch and has been used for such. The final texture can be undesirable for some food formulations. The starch industry suppliers offer modified starches, some mono-substituted only, while others include the addition of specific acids to create a hydrolysis reaction during sterilization. A smoother consistency and reduced viscosity results. These starches have also found unique functionality in microwave foods today (use can range from 0.5±4.5%) Other degrees of modification yield starches that can provide thin to thick properties and have shown a reduction in sterilization time for some food systems. This could increase productivity and improve product quality. Depending upon the food product, quality issues related to color, texture and mouth-feel could result. These starches are usually very highly cross-linked and generally possess some degree of mono-substitution, dependent upon final viscosity and textural parameters. Waxy maize is one starch used widely for this process. Foods not designed to be clear, are referred to as opaque (cloudy). Starches have been engineered to provide such a characteristic. Foods such as gravies, sauces or beverage mixes are excellent applications for such modified starches. These starches are usually stable during high temperature processing. Modified common maize and/or wheat starches are typically used. Starches designed for opacity are not suggested for use as water binders or phase stabilizers. Author's comment Considering the variety of base starches that a starch processor has to utilize for the conversion into the various levels of modified products, the food scientist now has an extensive spectrum of stabilizers and thickeners to choose from. Thus, utilize your ingredient supplier. They have the expertise to suggest a good starting point for the initial formulation.

13.6.2 Frozen The freezing of food is a growing method for the preservation of quality food products. It can offer more than a year of stability for many food systems. Industrial studies have shown the freezing of foods typically maintains high quality, and in some cases rivaling that of fresh. This is true for a given period of time for several food systems. Foods sold as frozen may require thermal processing prior to storage. The thermal processing generally is to enhance flavor, consistency, texture and appearance. The primary reason it is done is to stabilize or manage the water phase of the food matrix. Some foods can be prepared without heat and can usually utilize instant starches, with or without the incorporation of other hydrocolloids (gums). But, in many of these food products, the frozen type is reheated prior to consumption thus requiring some type of stabilizer to ensure that water is not lost during heating and which in many cases generates an undesirable texture.

ß 2004, Woodhead Publishing Limited

As was written earlier, a food scientist should be aware of the type of freezing system used, as well as how long the product is going to be subjected to continuous freezing. The use of rapid or blast freezing reduces ice crystal size. However it is handled it can still be a thermal shock to the starch granule, food particulates or other temperature sensitive components. The use of spiral freezers that are significantly slower is more common and generally creates larger ice crystals. Within the freezing process the slower the water freezes, the greater the chance for large ice crystal build-up. This can be detrimental to added starch. Ice crystals can cause the fragmentation of granules and/or food particulate cell walls. Dependent upon the storage conditions, ice crystals can continue to grow, thus increasing storage or product degradation and shortening the overall shelf-life of the food product. One positive attribute for freezing, regardless of method, is that freezing retards bacterial and enzyme activity. When extended shelf-life is desired (greater than a year), industrial studies have shown that starch blended with small amounts of a gum (hydrocolloid) maintain a higher product quality than starch alone. For foods that are to be frozen (ÿ20 to ÿ40 ëC), it is suggested that a minimum of 1.5% of a properly modified starch and approximately 0.025% of gum (xanthan) be used (formula weight). Actual amount required will vary based on product characteristics and properties desired. Any starch can be used with proper modifications. The desired source is up to the food scientist and product parameters. It should be noted that neither cross-linking nor mono-substitution, as an individual modification will yield a stable frozen product. The modifications alone do not support extended freezer or freeze-thaw functionalities for the purpose of maintaining fresh characteristics. To achieve these functional properties, it requires the proper levels for both cross-linking and mono-substitution. For all food stuffs, the starch of choice must be determined on several parameters. However, personal experience has shown that tapioca, wheat and waxy maize starches are extremely stable in frozen foods, especially those anticipated for extended freezer-life. Freeze-thaw can be accomplished with almost any properly modified starch. Waxy maize starch can yield exceptional stability for freeze-thaw conditions when modified utilizing the proper degrees of crosslinking and mono-substitution. The ultimate starch selection must include all facets of preparation and handling.

13.6.3 Instant products (soups, sauces and gravies) Most instant foods are sold as dry mixes. These mixes thus require the consumer to prepare by reconstitution. Usually this involves water or other basic liquids commonly sold, i.e., juices, milk, broth, etc. The reconstitution could require the use of heat. The commercial process to manufacture these dry mixes usually involves the blending of several ingredients prior to packaging. Starches used in these foods are typically of low moisture content. Instant starches are usually less than 5% moisture and if cook-up starches are incorporated, they would most likely possess a moisture of less than 10% (8±15% based on mix weight).

ß 2004, Woodhead Publishing Limited

Instant foods can and are produced commercially via spray drying preblended mixtures, or utilization of freeze drying, drum drying or extrusion. Starches used in foods of this type generally contribute to the product functionality and characteristics. Viscosity, texture, stability, appearance and eating quality can be either controlled or affected by the incorporation of starch. Instant foods processed via spray drying, drum drying or extrusion may not support the use of added starches. The process may adequately gelatinize any natural starches within the mix. Instant mixes formulated with instant modified starch usually require the addition of ingredients for dispersion. Such products may already be included in the product formula, but if not, ingredients such as sucrose, dextrose, low DE sweetener solids, maltodextrins and/or flours can be used as the dispersing aids. These will eliminate or significantly reduce the potential of lumping or `fish eyeing', as it might be referred to. Some manufacturers have utilized the technique of agglomeration, as a method to reduce lumping during hydration. Another method may be to request not only a modified instant starch, but one of coarser particle size. The larger particle reduces the surface area available for hydration and thus results in slower water uptake. It can produce a somewhat initial grainy texture as compared to the other accepted procedures. Also the reverse is possible by utilizing a finer particle size, thus increasing surface area and increasing hydration rate.

13.6.4 Functional differences based on process of production Instant starches available commercially today offer a vast range of viscosity and functional parameters for the food scientist. It is very difficult to attempt to prepare similar functional properties with instant starches processed via different production systems. Therefore, if you are attempting to secure two starches having similar functional characteristics, it is suggested you acquire them from similar production processes. As mentioned earlier regarding drum drying and extrusion, it is similar for spray drying, freeze drying, or other physical modifications being utilized today. Genetically derived instant starches have not been developed to date, therefore we must wait for property comparison.

13.6.5 Snack foods Both native and modified starches are commonly utilized in snack products. Their use has been for expansion, adhesion, texture and color generation. Puffed (extruded-baked) Common maize, tapioca, potato, wheat and other grains are commonly used. Modified starches, however, usually require very specific modifications and are manufactured to possess specific functional properties. These properties are as much related to the starch origin as the degree and type of modification

ß 2004, Woodhead Publishing Limited

incorporated. Native, mono-substituted and/or hydrolyzed starches generate unique expanded products. Dextrinized and/or thermally processed starches offer adhesion, color and textural properties. Depending on the degree of modification or the origin, a food scientist can prototype a product for extrusion, baking or frying. Fried Native or unmodified starches are not predominately used in these products; typically, these snacks consist of modified starches and flavorings. The shaping or forming of the dough is dependent upon the finished snack product. Products are shaped or formed prior to frying. Sometimes snacks are frozen prior to the frying process. They can also be frozen after par frying. Par frying is the process used to only partially cook the food item. In many of the fried snack products, blended ingredients are used. Starches, both modified and native, as well as flours, dextrins and low DE sweeteners are common ingredients. Level of usage is dictated by texture, color, flavor, storage and frying conditions. Typically, for simple coatings, (batters) 20±80% starch is used in the dry mix. More complex coatings can involve several starches at significantly differing levels (1±10%). All ingredients are chosen based on product quality and functional attributes to be contributed. Baked, micro-waved, impinged air processed The above terms can be also referred to as non-fried. Within the snack industry, this is one of the fastest growing market areas. For starch use, this too has been a significant growth area. Starches have been able to create the unique products required for this growth. Those products are call `halfproducts'. A half-product represents a formulation that for commercial use requires the consumer or commercial operation to complete the texture or functional properties incorporated into the product. An example could be an extruded particulate that requires baking or microwaving to complete the expansion, viscosity development or crisping of the ultimate product for consumption. Dependent upon texture and other functional properties, starch choice could range from common corn to a highly modified starch (origin optional based on function).

13.6.6 Dressings, sauces, gravies and other condiments The product categories of dressings, sauces, gravies, etc., usually consist of very high water content products. This excess water usually demands the use of starches that are classified as hydrophilic; starches that will hydrate several times their weight in water. This capability allows the food scientist to provide viscosity and texture to the finished food product. Starches of choice also have to have the functionality of high acid or low pH stability. This is especially true for the product line of dressings. These products are shelf stable and utilize low water activity as well as high acidity to maintain

ß 2004, Woodhead Publishing Limited

a safe product of good quality. Dressings involve high shear processing and with many products steam injected cooking. Both of these processing systems can cause starch damage. Starches of choice for the above processing conditions will usually be mono-substituted and highly cross-linked starches thus providing the water binding capacity and the processing stability. The above is true for the spoonable type dressings, those that require cooking to prepare the starch matrix, and then utilize homogenization to incorporate the lipid portion. Starch usage is typically 3±6% of the total dressing. For the family of dressings known as pourable, many producers do not use the cooked starch system. Pourable dressings are typically stabilized with instant starches and other hydrocolloids (gums-xanthan). These starches do not require cooking to hydrate, thus offering the dressing processor the capability to limit processing conditions that could damage starch granules. With many pourable dressings, homogenization is very limited and with some brands not required. Starch usage is from 0.5±2.0% in the total dressing. For the preparation of sauces and gravies, a food scientist must consider these to be neutral food products and will require sterilization, refrigeration or freezing when producing a hydrated product. Therefore, knowledge of the heating process is very important and should be considered necessary prior to selecting the correct starch. Many gravies and sauces are prepared utilizing a dry mix procedure initially, followed by the incorporation of water and then heated prior to serving. In some instances, such as food service preparations, it is necessary to use modified starches that will provide stability for long periods at elevated temperatures. Water control is very important in these foods, as well as the ability to reconstitute after storage via refrigeration and/or freezing. Storage could range from 8±72 hours, dependent upon the activities of the food service facility. Again, the starch of choice will depend upon the textural, eating properties and processing characteristics outlined for the particular food system. Always refer back to the starch origin and consider the basic properties to enhance the correct choice. Usage could range from 1.5±4.0%, wet basis, depending upon the overall functional characteristics.

13.6.7 Bakery products In this section we will discuss bakery systems typical for baked food products recognized by the consumer today. Those previously mentioned as `half-products' were unique to that discussion and will not be included in this section. Products will be considered totally processed utilizing a bakery operation. In most bakery products the primary starch source is from wheat flour. However, to develop and retain unique functional properties in today's baked foods, the food scientist will most likely incorporate modified starches and/or gums to enhance consumer appeal or improve shelf-life. The balance of moisture in differing systems has become a very difficult formulation task for the food scientist. Water activity and the knowledge of ingredient interaction has become a challenge for today's food

ß 2004, Woodhead Publishing Limited

scientist. This becomes more significant as the product undergoes mixing, handling, baking, storage and possible reconstitution. Now not only do the ingredients have to interact correctly, but packaging must also complement the final product to retain quality through handling and storage. Those baked products that are formulated as dough could require special handling when incorporating added starches or stabilizers. These dough products could be refrigerated or frozen. Some may consider them as half-products, however, for this writing I am referring to them as a finished mix in a transition phase. Moisture hydration by all hydrophilic ingredients must be considered, primarily due to the dough mixing, sheeting, handling, etc. It is not just the total hydration capacity of the added starch, but the rate of hydration by all ingredients that could be significant. Another important factor could be the rate at which the starch loses moisture and what physical properties are generated after processing. Wheat starch should be of significant consideration, both as a native source and/or modified product. It derives from origin grain and therefore offers unique synergy with flour. Tapioca (cassava) and waxy maize starches have also performed very well in many baked foods. Dependent upon how the baked item is to be handled, cook-up starches can be beneficially utilized; however, if boil-out or blow-out becomes a problem either higher soluble solids are necessary or the utilization of cook-up and/or instant starches. Gums (another hydrocolloid) may also contribute to the control of viscosity, thus reducing the chance of boil-out, however, texture could become an issue if levels of use are too great. Similar concerns can be said about excess amounts of starch within the same products. As with any food system, always consider the other ingredients when selecting a starch. The lipid and sweetener matrix can become very critical as it is related to total functionality produced by the added starch. As discussed earlier in the section on `ingredients', the type of lipid, when added and what it is added with, can significantly contribute to the hydration and functionality of the starch. The same can be said regarding the sweetener system. It is important to balance the sweetener and flour for certain baked goods. The choice of sweetener is critical. The effect that sucrose has on starch and flour is significantly different from that of sweeteners derived from other sources. Will this cause a problem with your starch of choice? It could. It may be necessary to evaluate more than one sweetener for lipid and starch chosen for the baked system. Sucrose used as the primary sweetener in baked/frozen or frozen/baked product could cause crystal growth and shorten the desired shelf-life of the finished product. For frozen baked items such as fruit systems, sucrose should be kept at about 10% of the sweetener system and the total solids utilized for the fillings should be between 45±60%. The added starch can range from 3±4.5% dependent upon the desired texture.

13.6.8 Pet products Do not be blinded by the word `pet' when deciding product choices. Pet foods undergo as many regulatory issues and processing parameters as do food

ß 2004, Woodhead Publishing Limited

products for human consumption. There are a few ingredients utilized within pet foods that we do not use in human food products. The care of formulation and product safety, as well as the standards are very high for pet foods. Formulation, processing and consumer satisfaction is taken very seriously. Starch utilization is very similar to foods formulated and processed for humans. The canned foods are sterilized with similar standards, therefore requiring starches with similar food approval and functional characteristics. Those pet foods extruded or formed for snacks also require starch and ingredient mixtures requiring similar modification and processing characteristics as those for human consumption. In some instances, the different eating characteristics between humans and animals are considered and the standards for functionality may be altered to enhance texture or flavor for pets. Hardness and bite are important issues to the pet food manufacturer. What a human may consider hard, a pet would not consider hard enough. This then requires animal eating, tasting and quality panels for product analysis. These are very similar to human taste panel studies. The same holds true for diet contribution, digestion, health benefit, fecal discharge and many other physical and physiological properties. Starch as utilized in pet foods will be dictated by the basic functional properties of the native starch and the acceptance of it by the consumer (the pet).

13.6.9 Meat products Starch has gained approval for general use in many meat or analog products around the world. Several meat products commercially produced at present have very strict standards of identity. For the improvement of storage and processing, starch and related products have been approved for use in both standardized and non-standard meat products. The use of starch (native or modified) in meat products is regulated for the United States by the US Department of Agriculture's Food Safety and Inspection Service. Specifications are available in Part 318.7 of the Code of Federal Regulations (Title 9). Prior to the approval of starch and related products, water management in many meat products was controlled through the use of high levels of salt. For meats such as frankfurters, bologna, luncheon loafs, etc., (emulsified or chopped), starch contributes a method to reduce and control purge (free water/ brine). Flour and other grain-based ingredients have been utilized over the years, none contributed to the water absorption and control functionality compared to modified starches. In seafood derivatives such as surimi (seafood analogs), starch contributes more than water management. Texture, process improvement and most significantly, economic advantages have been achieved with the addition of starches. It has been identified that not only can modified starches be used, but also significant levels of native (common/unmodified) starches too. Unmodified tapioca, potato and wheat are commonly used in conjunction with one or more modified starches to generate the functional properties desired in a meat analog. Modified waxy maize, tapioca and potato are typical starches of choice for meat

ß 2004, Woodhead Publishing Limited

analog systems. Again, the final product being processed and the desired texture determines the types of starches utilized. A meat analog of this type is a complex formulation and requires functionality knowledge of starches prior to their selection for use. It could take several evaluations to determine the ideal blend of starches and/or related products to produce a commercial meat analog.

13.6.10 Cereals, pasta, bars and related products Cereals today are predominately whole grains, with or without added starches, sweeteners and other nutritional additives. Many cereals depend on the expansion capability of the flour(s) to develop and maintain the texture and eating quality for the finished product being produced. Referring back to the sections pertaining to `extrusion' and `instant starches', we can relate to the potential contribution of starches to cereals. Cereal manufacturers have utilized extrusion as one method for preparing ready-to-eat cereals (RTE). Partial to total gelatinization of the starchy portion of the formulation is developed during the extrusion processing of cereal systems during production. Therefore, the effect of processing can be as significant as the functional properties generated during and post processing by added starch from the flour or specific grain used for the cereal type. Cereal producers today utilize added starch (native wheat or maize) to enhance the texture, processing and functional properties of the finished cereal. In addition to the above processing, many cereal products commercially available today are coated with individual ingredients or complex blends to provide additional functional attributes. Examples of these additions would be the use of sucrose coatings, dextrin coatings, or the inclusion of native starches, lipids or proteins to enhance the bowl life of a cereal. Basically, these are added to enhance the reduction of the absorption of liquids by the cereal. The property of extended bowl life has been a key research issue for many cereal producers over the years, a property that offers longer crisp texture to a cereal when mixed with a liquid, usually milk. The coating of cereal pieces has been an accepted method for years, however the desire for a cereal product requiring no post-extrusion coating, while providing equal stability still exists. Improvement has been achieved, but none to date match what can be accomplished with a coating. Typical starch usage in cereal products varies based on the functional attribute of the starch for the cereal being produced. Amounts of 3±20% could be utilized if the starch was an inclusion ingredient. If used in the coating application it could range from 10±30%. The coating process has now led to another line of cereal-like products. It also created an extension to the snack industry with cereal-like products, the granola or whole grain coated snacks. We have bars, clusters and bits for our enjoyment. All are just a step from the cereal product itself. Wheat, oat, barley and rice flours are popular starters for cereals. Wheat, oat, rice, tapioca and potato starches are commonly used in many cereals as well as the cereal-like snacks. Usage in bars, clusters, etc., has been from 2±15%. The level is dependent upon the functional characteristic desired.

ß 2004, Woodhead Publishing Limited

Pasta in most of the world has standards for processing and product names. However, some manufacturers have become creative and some standards have been softened to allow the use of non-standard ingredients to help enhance stability and processing both during production and post consumer/retail use. Foods that have been developed through thermal processing (canning) required the food scientist to introduce non-standard ingredients into the pasta portion of the food system, thus creating a fanciful name item. Mono-substituted and crosslinked starches within specific modification levels can provide the unique functional properties desired. Waxy maize, common maize, and tapioca are very typically used in these products at 0.5±2.0% for the pasta and then again at 2±5% in the sauce portion, if one is present. For the production of pasta products starches can provide greater stability and improve reconstitution at the consumer level. Added starch from common and waxy maize will reduce friability and improve water uptake during consumer processing. Levels of 1±3% typically are added as a flour replacement and this should contribute to these properties. Low DE maltodextrins can also be utilized when formulating pasta type products. They too, will provide improvements as stated above, but the food scientist should consider color development, as maltodextrins are a type of sweetener solid.

13.6.11 Confections (candy) When considering confections (candy) as the products category, the food scientist should review the different confections being marketed today. Confections are generally listed in one of three categories: soft candies, gummy (chewable) and hard (non-chewable). Soft candies are those products similar to chocolate-based or flavor-coated products (i.e. circus peanuts) that may or may not be chewable. Gummy (chewable) type candies are gum drops, jelly-beans, jube-jube, etc. The hard (non-chewable) category would be those confections that are related to products such as cough drops, lozenges, etc. Starches utilized in these confections can differ dramatically. Soft candies usually require starches that generate a tender texture and typically are not gelling. Another type could require a starch possessing little heat stability and would totally degrade if exposed to an enzyme system. An example would be the starch used for the production of a confection called a `chocolate covered cherry' product. Others confection products may be for the preparation of caramel or fruit centers. In the gummy and hard candies starches are similar, but differ with the degree of modification and are blended as necessary to accomplish the desired texture. Gummy candies are produced using hydrolyzed starches. In some instances lightly modified or unmodified high amylose starch may be utilized. For the production of hard candies this type of starch would be used because of the set generated by the greater percentage of amylose. In all confections the food scientist will have to decide on a sweetener system. It could contain sucrose or more likely will be composed of differing types of liquid sweeteners derived from a carbohydrate modification. It is probable that the sweetener system will be derived from a hydrolyzed grain,

ß 2004, Woodhead Publishing Limited

based syrup, generally supplied as a liquid. Some sweeteners are available derived from other starch bases, but are not commercially available and do not offer quite the economics, as yet. Solids for most confections are in excess of 80%, thus requiring elevated or greater than atmospheric conditions to process. To develop the textures desired for the gummy or hard candies starch is typically added at a 6±13% level. The added starch could be either a single unit or more likely a blend of starches to achieve the desired results. Flavors and colors are usually not processed at the extreme conditions as they could be compromised in functionality. Because of the levels required to generate desired color and flavor profiles and their cost, usage is typically minimal in these formulations. In the confections marketed as low or no-sugar considerable efforts are made to limit variance from the above confections. These low or reduced calorie confections provide the food scientist with another challenge for balancing the excess water with other carbohydrates and hydrocolloids, while hopefully duplicating the original full sugar product. In many cases, reformulation is easier in confections with the no-sugar claim, than formulations attempting to make a low or no-fat claim. The area of fat reduction is discussed in the section on starch-based products used for total or partial fat replacement later in this chapter.

13.6.12 Dairy and related products For the purpose of this chapter I have classified `dairy foods' as those foods containing all or a significant percentage (>50%) of dairy derived ingredients or those that meet given sets of criteria related to a standard of identity for a specific product category. `Dairy related products' are those products usually sold in the dairy section of commercial and/or retail outlets, however, they have been formulated to simulate a standardized dairy product but do not meet the identity standards associated with that product category. Many of these dairy related products will contain a portion (>0 70 hours, equivalent to an energy expenditure of > 70 000 kcal, if fat were the only substrate (Newsholme and Leech, 1983b). However, this would be possible only if fat could deliver an adequate amount of energy and if pain in the muscles and joints did not occur as limiting factors. At endurance competition speeds, however, CHO availability will be one of the performance limiting factors, because fat alone would be inappropriate for resynthesizing ATP at a high rate (Newsholme and Leech, 1983a, 1983b). CHOs have `an economic advantage' in that the amount of oxygen required for energy production from CHO is about 10% lower than that of fat (McGilvery, 1973). Sufficient carbohydrate availability is also required to make it possible for fatty acids to be oxidized in the citric acid cycle by providing the necessary

ß 2004, Woodhead Publishing Limited

intermediates to keep the citric acid cycle running. Thus, the basic energy requirements for muscle work will always be covered by the mobilization and subsequent metabolism of substrates from the CHO and fat pools in muscle, liver and adipose tissue (Bergstrom and Hultman, 1966, 1967a, b; Bjorntorp, 1991; Hultman, 1974; Newsholme and Leech, 1983,a; Sherman and Lamb, 1988). The use of any of these pools will never be exclusive but one or the other may be dominant, depending on availability! If an athlete were to use fat as the prime energy source, their performance capacity would fall to about 50% of maximum capacity, a level that has been observed in glycogen depletion studies. Indeed, several lines of evidence suggest that intense and lasting muscle work cannot be performed without appropriate availability of CHO. When specific muscles or muscle fibers become glycogen depleted and become impaired in their ability to perform repeated high-intensity contractions (Bergstrom and Hultman, 1967a, b; Costill, 1988; Hultmann, 1974; Maughan, 1990). Research has shown that glycogen depletion, either by exercise or by a combination of exercise and low CHO intake, leads to a reduction in work capacity of about 50% of the normal maximal working capacity (Costill, 1988; Maughan, 1990; Newsholme and Leech, 1983a). Conversely, when the CHO stores in muscle and liver are increased by diet manipulation, athletes are able to perform for longer at high exercise intensity. These examples show that the availability of glucose and the size of the glucose/glycogen stores are important and limiting factors for endurance performance. Manipulation of the size of the glycogen store may therefore impact on performance capacity (Brouns, 1997), as will CHO supply during exercise (Jeukendrup et al., 1999, for review see Brouns, 1997, 2002). As such, supply of CHOs that are an efficient fuel for muscles to optimize performance is not only recommended for long-distance runners and

Fig. 18.2 Muscle glycogen is stored as `starch' granules (C) within the muscle fibers (A) between the mitochondria (B). The amount of glycogen stored has classically been measured by biochemical analysis of a muscle biopsy sample. Nuclear magnetic resonance is a more modern technique which is non-invasive and makes it possible to study relative changes in glycogen content.

ß 2004, Woodhead Publishing Limited

Fig. 18.3 Metabolism in a glycogen depleted state. Because of a lack of glycogen in the liver, blood glucose falls; lactic acid production decreases and fat metabolism increases to compensate for energy deficits. The consequence is that the performance capacity will drop to approximately 50% of the maximally achievable level. Reproduced from Wagenmakers et al. (1991), with permission from the American Physiological Society.

cyclists but has also been shown to improve performance in downhill skiing, ice hockey, tennis, football and other repeated sprint-like sports (Brouns et al., 2002b; Balsom et al., 1999).

18.2.3 Carbohydrate reserves: a limited stock In the body glucose is stored as glycogen, long chains of glucose units, in the liver and in the muscles. The amount of glycogen stored in the liver amounts to approximately 100 g. This quantity may change periodically depending on the amount of glycogen that is broken down for the supply of blood glucose in periods of fasting and the amount of glucose that is supplied to the liver after food intake. Accordingly, liver glycogen reserves increase after meals but diminish in between,

ß 2004, Woodhead Publishing Limited

Fig. 18.4 Glycogen depletion rate depends on exercise intensity. Trained subjects use less glycogen and more fat. At maximal intensities these differences are less pronounced.

especially during the night, when the liver constantly delivers glucose into the bloodstream to maintain a normal blood glucose level (Hultman, 1967, 1974, 1981; Newsholme and Start, 1973; Richter et al., 1986). The amount of glycogen that is stored in total muscle in the body amounts to approximately 300 g in sedentary people and may be increased to > 500 g in trained individuals by a combination of exercise and the consumption of a glucose-rich diet (Bergstrom and Hultmann, 1967a, b; Hultman, 1974). The total intramuscularly stored glucose, in energetic equivalent, may thus range from 1200 to 2000 kcal. During physical exercise a number of metabolic and hormonal stimuli will lead to an increased uptake of blood glucose by the working muscles to serve as a fuel for muscular contractions. To avoid blood glucose levels falling below the normal physiological value, the liver will be stimulated to supply more glucose to the bloodstream. This supply is mainly derived from the liver glycogen pool and to a small degree from the process of gluconeogenesis (Ahlborg et al., 1974; Hultman, 1974; Newsholme and Start, 1973; Richter et al., 1986). Thus, an appropriate amount of stored glucose in the liver is a key factor for maintenance of a normal blood glucose level during prolonged exercise. When the liver glycogen store is depleted and exercise is executed without concomitant glucose intake, blood glucose may fall to hypoglycemic levels (Felig and Wahren, 1975; Felig, 1982). Glucose uptake by the muscles from blood will drop to marginal levels and the working muscles will then totally depend on the local glucose supply from remaining muscle glycogen. Signs of fatigue may then occur leading to early exhaustion.

ß 2004, Woodhead Publishing Limited

18.2.4 Habitual starchy diet: stocking up the glycogen reserves After exercise the endogenous glucose pools should be replenished. The time available for total recovery, i.e., the time elapsing between finishing exercise and the next sport activity, will determine any need for interventions to help maximize glucose storage. Glycogen synthesis has been shown to be most rapid during the first few hours after exercise. Thereafter, the synthesis rate declines gradually (Coyle, 1991a, b; Hultman, 1974) Glycogen synthesis itself is possible only when glucose is supplied and is possible both at rest and during exercise (Brouns and Beckers, 1993; Kuipers et al., 1987). The glycogen synthesis rate depends largely on the glycogen synthase activity and the quantitative glucose supply (Costill, 1988; Coyle, 1991a, b). The latter depends largely on the types of CHO and content of the foods ingested. Glucose favors muscle glycogen recovery, while fructose is primarily taken up by the liver for liver glycogen resynthesis (Bergstrom and Hultman, 1967a, b; Hultman, 1981). When the next activity takes place after one or two days, the athlete can recover properly by ingesting normal meals with a relatively high CHO content (55±65% of total daily energy intake) These meals optimally contain plant starch sources such as whole grains, cereals, pulses, fruits, vegetables, etc. A relatively slow digestion and absorption rate is favorable in this condition. Under these conditions, 400±600 g of glucose/day should be sufficient to recover glycogen stores for meeting energy requirements of up to 4000 kcal/day (Costill, 1988). However, when daily energy expenditures are very high, such as during multi-day competitions (e.g. Tour de France cycling), glucose requirements may reach > 12 g/kg body weight/day. Under these conditions, glucose intake by normal meals composed of low glycemic index CHO sources may result in too much gastrointestinal bulk (due to the dietary fiber ingested along with the CHO) causing gastrointestinal distress. Therefore, athletes who ingest only normal meals in such circumstances will not be able to ingest enough food, which will result in negative energy balance and insufficient glucose intake. The next exercise day will commence in a state of incomplete recovery. Sports practice and also controlled experiments have shown that the high needs for energy and glucose during days with an energy expenditure exceeding 4500 kcal/day can only be covered appropriately by the ingestion of high glycemic index CHO foods/drinks (Brouns et al., 1989; Lamb and Snyder, 1991; Saris et al., 1989). Processed, cooked and mashed potato, rice, noodles or cornstarch belong to this category. Additional CHO drinks can be taken during the first few hours of recovery, and will help to promote muscle glucose storage (Coyle, 1991a; Kiens et al., 1990; Dreher et al., 1984; Lamb and Snyder, 1991).

18.2.5 Carbohydrate ingestion before, during and after exercise The rate of utilization of glucose from stored glycogen in the body can be reduced by supplying oral CHO (Jeukendrup et al., 1999). For example, when starch-containing food is ingested, digested and absorbed, glucose will enter the circulation and reduce the need to break down liver glycogen for the maintenance of an appropriate blood glucose level. Additionally, glucose

ß 2004, Woodhead Publishing Limited

supply to and glucose uptake by the muscles will be elevated. Indeed, a large body of scientific evidence shows that oral glucose intake reduces liver glucose output by about similar rates as the supply to blood (Ahlborg et al., 1974; Jeukendrup et al., 1999). The increase blood glucose after intake will stimulate insulin release and with it glucose uptake by the muscles as well as subsequent CHO oxidation (DeÂcombaz et al., 1985; Guezennec et al., 1989; Hawley et al., 1992; Mosora et al., 1981; Pallikarikis et al., 1986). Theoretically this will reduce the rate of muscle glycogen and protein degradation for energy production and delay the onset of fatigue/improve performance (Wagenmakers et al., 1991; Yaspelkis and Ivy, 1991). For exercise lasting longer than 45 minutes it is recommended that at least 20 g, but optimally up to 60 g, be consumed, with sufficient fluid, during every following hour of exercise (Costill, 1988). Such amounts have also been shown not to delay gastric emptying to a physiologically important degree and to stimulate water absorption in the intestine (Brouns, 1998, for review see Brouns, 2002a; Rehrer, 1991). This aspect is of particular importance in endurance events in hot conditions, where both CHO and fluid availability may be performance-limiting factors (Yaspelkis and Ivy, 1991). The CHO sources used should be rapidly digestible and absorbable. Most efficient are (soluble) CHO sources, which can be ingested with fluid. The gastric emptying rate should be relatively fast and the physical form of the CHO should allow rapid digestion/enzymatic hydrolysis. This is not the case with all CHO sources (Leijssen et al., 1995). Also food matrix may have an effect (Jarvis et al., 1992; Keizer et al., 1987). For example, the dietary fiber in which some CHO sources are `packed' may form a physical barrier to digestive enzymes (Crapo, 1985) and may also reduce gastric emptying rate. Normal daily meals should primarily contain foods that are rich in slowly digestible CHO and dietary fiber resulting in a low glycemic response. Examples of such foods are whole grain products and cereals. However, foods taken shortly before and during exercise should be low in dietary fiber and have a high glycemic index, in order to allow for a rapid gastric emptying and digestion/absorption (Brouns et al., 1991a,b; Coyle, 1991a). The reason for this apparent paradox is that dietary fiber may reduce gastric emptying and intestinal transit while reducing the degree in which enzymes can reach the starch for hydrolysis. Dietary fiber also increases gastrointestinal bulk due to water uptake and swelling. Fiber enhances transit in the gut and may be subject to bacterial fermentation causing gas production. Softening of the intestinal contents by fiber and the related improved intestinal transit are desirable in sedentary individuals but may pose a problem during intensive endurance exercise lasting more than four hours. These factors may explain why athletes who ingest slowly digestible whole grain foods, both before and during exercise, experience more gastrointestinal problems than athletes who ingest low fiber products (Brouns et al., 1991a, b; Browns and Beckers, 1993; Rehrer et al., 1992). When dietary fiber is excluded from the CHO source, the starch will be fully accessible to enzymatic digestion. Accordingly, starch and glucose polymers but also sucrose have been shown to be very effective in increasing blood glucose levels resulting in high oxidation rates during exercise (Coyle et al., 1984,

ß 2004, Woodhead Publishing Limited

Fig. 18.5 (a) Oxidation of CHO taken orally during exercise. (b) Oxidation of oral CHO as a percentage of the given dose (grams). M ˆ maltodextrin with a mean chain length of 20 glucose units, S ˆ sucrose. Oral carbohydrate is oxidized during exercise, and thus contributes to energy production. Increasing CHO intake to about 100 g/hour increases its oxidation. Higher CHO intakes have no effect, most probably because of the induction of a delayed gastric emptying rate resulting in a drop in the percentage of CHO oxidized. Interestingly, glucose polymers (maltodextrins) are as well oxidized as sucrose. From data of Wagenmakers et al. (1993).

1991a; Guezennec et al., 1989; Hawley et al., 1992; Wagenmakers et al., 1993). Whether such CHOs are ingested as solid foods or as liquids does not seem to play an important role (Mason et al., 1993). Effects on blood insulin levels during exercise also do not appear to be different for these CHOs (Coyle et al., 1991a, b). These types of CHO have the additional benefit of being easily dissolved in fluids, which is an important feature in that both the requirement for CHO and fluid are determined by the exercise intensity and duration. With respect to palatability and

ß 2004, Woodhead Publishing Limited

gastrointestinal comfort, starch-hydrolysate and starch may have the benefit of being less sweet than the mono- and disaccharides. They also have less effect on fluid osmolality and have been shown to maximize quantitative glucose absorption, which is of advantage at concentrations of > 80 g/l (Rehrer et al., 1991).

18.2.6 Effect of glycemic response on metabolism and performance Some early studies showed that an intake of 50±75 g of rapidly absorbable CHO prior to exercise induces a rapid rise in blood glucose and insulin, followed by a rebound hypoglycemia caused by the induced high insulin levels. The latter was observed to impair performance during subsequent exercise, most probably by reduced glucose availability and at the same time low fat oxidation rates due to the inhibiting action of insulin on lipolysis. The suggestion was made that this would be the case for all rapidly available glucose sources. Along similar lines, hypoglycemia may also affect brain metabolism and mental performance (see following sections). Accordingly, ingesting low GI carbohydrates in similar conditions will induce fewer changes in insulin and blood glucose response and subsequent counter-regulatory effects and may thus improve performance (Thomas et al., 1991) However, all studies that showed such an effect were performed after an overnight fast, after which CHO was ingested in the resting state, 45±60 minutes prior to exercise and no CHO ingestion was allowed during exercise. These conditions are not comparable to those of the endurance athlete, who will always eat a pre-game meal and continue to ingest some CHO during exercise. In studies conducted under real competition, food ingestion conditions do not show any rebound hypoglycemia (Brouns et al., 1989, 1991a). Also, CHO intake during exercise will counteract pre-exercise diet effects (Burke et al., 1998). To date, a large number of studies have shown that pre-game starch and glucose intake can be beneficial in delaying fatigue (for review see Coyle et al., 1991a, b). For example, Chryssanthopoulos et al. (1994) compared the effect of ingesting a CHO drink during a 30 km treadmill trial to ingestion of a CHO rich meal four hours prior to the run. Performance times were identical and there were no differences in self-selected running speeds. In both cases, blood glucose was maintained above 4.5 mmol/l.

18.2.7 Carbohydrate availability affects performance The source and the molecular structure of the carbohydrate will determine the rate of digestion, absorption and subsequent fate in metabolism. For example whole grain products that are rich in dietary fiber and have intact grain granules will influence gastric emptying and intestinal transit resulting in a reduced rate of digestion and absorption of glucose into the circulation. This would be a disadvantage in any condition where rapid glucose availability is desired as, for example, during intensive physical endurance work. On the contrary, in resting conditions where generally high blood glucose excursions are less desirable, the characteristic of low glycemic response may be desirable in terms of preventing

ß 2004, Woodhead Publishing Limited

large fluctuations in blood glucose and insulin as well as reducing the risks of protein glycation. Cooked starchy meals that have been cooled down may contain significant amounts of resistant starch, for example, pasta and potato salad. This fraction of indigestible starch will be subject to bacterial degradation in the colon and may lead to gas production. The latter is highly undesirable during an endurance sports event such as marathon running since it may induce intestinal distress that will hinder performance. However, a relatively high content of resistant starch may be desirable for a healthy colonic environment (Brouns et al., 2002a). CHO rich meals that also contain high amounts of dietary fiber will exert similar effects and will also cause an increased degree of filling of the large bowel. The latter may induce an urge to defecate, a difficult situation when in competition. Some carbohydrate sources may lead to gastrointestinal distress because the matrix that they belong to contains substances that are less well tolerated in a number of individuals. Examples are fruit juices and milk. Studies have shown that different simple sugars are emptied from the stomach and are absorbed at different rates in the gut, while osmolarity does not play an important role at low CHO concentrations as usually consumed during exercise (Brouns, 1998; Saris et al., 1993). The latter will strongly affect the glycaemic and insulinaemic responses and uptake in tissues, thus the availability for energy metabolism. Examples are glucose, which is rapidly absorbed and oxidized in energy metabolism compared to fructose and galactose, which are slowly absorbed and oxidized (Leijssen et al., 1995; Massicotte et al., 1989; Hawley et al., 1992). Such low absorption rates have also been associated with gastrointestinal distress, most probably related to osmotic effects in the gut lumen (Murray et al., 1989; Brouns and Beckers, 1993). Some researchers have specifically addressed the study of the rates of glucose oxidation from various starch types. This is possible by using starch that is naturally enriched with Carbon 13 (cornstarch) or by adding a C13 glucose tracer to starch. Both methods have their limitations. C13 starch studies are difficult to perform in US populations because their high consumption of cornstarch containing products causes high levels of endogenous C13 enrichment of glycogen and also fat depots. In a European population this is less of a problem (Wagenmakers et al., 1993). Alternatively, a C14 glucose label can be added (Hawley et al., 1991) but is not allowed in most countries for reasons of radiation exposure. Adding a C13 or a C14 label to a starch, in order to quantify the rate of oxidation of the label and assume that this equals the rate of starch oxidation is a practice that has been used. While doing this it has also been assumed that the rate of digestion of the starch is non-limiting and that glucose derived from starch digestion behaves similarly to the glucose label added (Hawley et al., 1991, 1992). However, Saris et al. (1993) were able to show that this leads to an overestimation of the oxidation rate. Accordingly, appropriate measures of starch oxidation can only be obtained using intrinsically labeled starch. This was done by Guezennec et al. (1993) who compared the rates of oxidation of bread, potatoes, rice, spaghetti and glucose,

ß 2004, Woodhead Publishing Limited

given as isocaloric meals. Their study showed that the glucose and insulin responses were as follows in descending order: glucose > potatoes > bread > rice > spaghetti. The oxidation rate of spaghetti was lower than that of glucose. The oxidation rate of amylopectin equaled that of glucose while a progressive decline in oxidation rate was observed from gelatinized amylase and crude amylopectin to crude amylose. These data confirm earlier published data from van Amelsvoort and Westrate (1992) From the data described above it appears that the type of carbohydrate will determine its suitability for use during exercise. Effects may be similar to those required for supporting mental performance (see later).

18.2.8 Macronutrient interactions influence metabolic fate of carbohydrates The metabolism of carbohydrate may also be influenced by the presence of other macronutrients. For example, recent studies have shown that a mixture of CHO and protein or selected amino acids leads to a significantly higher insulin response compared to the ingestion of CHO alone (Zawadski et al., 1992; van Loon et al., 2000). This may be of advantage in circumstances where a high insulin level may be of benefit to the rate of uptake of glucose in tissues. Accordingly, it was observed that the resynthesis rate of muscle glycogen after exhausting exercise proceeds more rapidly when such a combination is ingested (van Loon et al., 2000). This finding may be of benefit to athletes who have to train intensively or to compete more than once a day.

18.2.9 Carbohydrate-protein ratio and serotoninergic function CHOs may not necessarily exert effects purely by altering availability of glucose but might alter serotoninergic function with consequences for mental and physical performance. The theoretical background linking CHO and serotonin implies that the dietary effects are mediated by changes in plasma tryptophan:large neutral amino acid (TRP:LNAA) ratio. A good deal of work has been prompted by the idea that foods varying in the proportions of protein and CHO influence performance via changes in brain serotoninergic function. The mechanism by which a high CHO meal increases brain tryptophan and enhances serotoninergic neurotransmission has been well documented (Fernstrom and Wurtman, 1971; Young, 1991). However, it should be kept in mind that as little as 4% protein added to high carbohydrate can abolish the tryptophan effect (Teff et al., 1989). Fat may also influence brain tryptophan uptake (Huether et al., 1998). Another issue is the extent to which the plasma tryptophan/LNAA ratio must increase in order to elevate brain serotonin. There is disagreement in the literature. Some authors maintain that a 50% elevation of TRP:LNAA ratio must occur in humans to produce a meaningful rise in brain serotonin (Ashley et al., 1985). However, 20±40% increases in the plasma ratio of TRP:LNAA have been demonstrated to produce neuroendocrine alterations indicative of altered brain serotonin (Anderson et al., 1990; Kaye et al., 1998;

ß 2004, Woodhead Publishing Limited

Lyons and Truswell, 1988). Furthermore, animal studies suggested that an elevation of the order of 35% produced an increase in brain serotonin (Fernstrom and Wurtman, 1971). There is also some debate about whether the level of free or albumin-bound tryptophan is the most important factor in influencing brain serotonin levels. Chaouloff (1993) presents evidence suggesting that both forms can be transported across the blood-brain barrier. Whether CHO effects on mental performance are dependent on the promotion of glucose availability or the provision of tryptophan and concomitant increased serotonin will be discussed in relation to the form of CHO ingested, the particular domains of cognition, the time course of effects and individual susceptibility to cognitive and nutritional challenges. Newsholme (1990) has suggested that serotonin may be also involved in the development of physical fatigue. Based on data from animal research, showing that low blood levels of branched chain amino acids (BCAAs) may facilitate the entry of tryptophan into the brain, Newsholme hypothesized that a decrease in serum levels of BCAAs, as often observed during the later stage of endurance exercise, may be a contributing factor to fatigue. Thus, theoretically, reducing the rate at which BCAA can be utilized for energy production, by supplying rapidly available CHO as fuel, or supplying BCAA as supplement along with CHO during endurance exercise may help to delay the onset of fatigue by decreasing the rate of tryptophan uptake into the brain. However, no good data are available to support this hypothesis (for review see Brouns, 2002).

18.2.10 Future research directions for athletic performance Based on historical developments, focus is at present on fully digestible but slowly absorbed carbohydrates. Examples of such carbohydrates are the disaccharides trehalose and isomaltulose as well as high amylose starches. The reason for this focus is the assumption that slowly released CHOs will trigger less insulin response in favor of a higher degree of fatty acid oxidation. The latter may be of benefit to athletes, in order to spare muscle glycogen and delay fatigue, as well as to leisure sports and fitness sports, in order to reduce fat storage and maintain body weight in a desired range. The food industry is also working on designing novel types of slowly digestible carbohydrates. It is clear that research will need to be done on the rate of gastric emptying, subsequent digestion, absorption and oxidation as well as its impact on fluid absorption and intestinal tolerance before recommendations to athletes can be made regarding suitability for performance support.

18.3

Mental performance: the effects of glucose

18.3.1 Effects of glucose on cognition Blood glucose is normally 4±5.5 mmol/l (Ewing et al., 1998) and is controlled by a series of hormonal mechanisms. When it falls below 2.2 mmol/l,

ß 2004, Woodhead Publishing Limited

hypoglycaemia occurs (Amiel et al., 1991). Hypoglycemia occurs when the uptake of glucose by the tissues including the brain, exceeds the rate of supply from the liver and/or meals. The decline in glucose, occurring in both mild and severe hypoglycaemia, impairs cognitive performance. The threshold for impairment of cognitive tasks has been found to be within a range of 2.2 to 2.8 mmol/l (Widom and Simonson, 1990). However, some individuals maintained normal performance at below 2.2 mmol/l, and others showed disruption of function at 4 mmol/l. (Bellisle et al., 1998) suggesting that glycaemic control and glucose tolerance may be more critical to cognitive function than absolute blood glucose levels (Manning et al., 1990). There is substantial evidence however, that higher baseline levels of blood glucose can facilitate cognitive performance (Benton and Owens, 1993; Benton et al., 1994). Cognitive disruption at higher glucose levels occurs in non-diabetics whose symptoms of hypoglycaemia are relieved by food intake (Snorgaard et al., 1991). Subjective effects of neuroglycopenia (low blood glucose at the neuronal level) include negative mood (e.g., increased tension and anger) and attenuated arousal (e.g. drowsiness) but not all individuals will be aware of their hypoglycaemic status.

18.3.2 Hypoglycaemia and mental performance Hypoglycaemia is a common side effect of insulin treatment in individuals with insulin dependent (Type 1) diabetes mellitus (IDDM). It produces somatic symptoms such as trembling, sweating, lightheadedness, mental lethargy, headache and irritability. Effects on mental performance have been demonstrated following induced hypoglycemia in patients with Type I diabetes (IDDM) (Gold et al., 1995). Visual and auditory information processing is commonly affected. Decreased cognitive capacity on standard tests of performance and lowered IQ have been proposed to be enduring effects in follow-up studies of IDDM patients (Deary et al., 1993). It has been suggested that these effects result from repeated episodes of hypoglycaemia (Langan et al., 1991). However, Austin and Deary (1999) found no association between repeated episodes of hypoglycaemia and cognitive decrement or baseline cognitive ability in a reanalysis of long-term data. Type II diabetes (non-insulin dependent diabetes mellitus ± NIDDM) has a later onset, is often associated with obesity, and also has effects on cognitive function which have been proposed to lead to a higher incidence of dementia (for reviews see Strachan et al., 1997; Stewart and Liolitsa, 1999). However, the cognitive domains affected appear to be different from IDDM. In IDDM complex tasks which require speed and sustained attention (e.g. Digit Symbol Substitution Test ± DSST) are more reliably disrupted by hypoglycaemia than more simple psychomotor performance tests (e.g. Simple Reaction Time ± SRT). Cerebral blood flow alterations have been observed during hypoglycaemia in Type I diabetics (McLeod et al., 1996), in particular an increase in the frontal lobes. In Type II diabetes verbal memory and concept formation are the most commonly

ß 2004, Woodhead Publishing Limited

affected cognitive functions (Strachan et al., 1997; Elias et al., 1997) and frontal lobe and psychomotor function is less consistently affected (Biessels, 1999; Vanhanen et al., 1999). It is unclear why there should be a difference in the cognitive capacities impaired during hypoglycaemia in these two types of diabetes. It is possible that delivery of glucose to the cortex is prioritized so that supply to the frontal lobes is protected. An alternative possibility is that as age of onset and length of time of undetected diabetes differs greatly between NIDDM and IDDM, the incidence of repeated (and untreated) episodes of hypoglycaemia and potential cerebral disruption of glucose availability also differs (Elias et al., 1997).

18.3.3 Induced hypoglycaemia Effects on mental performance have been demonstrated following induced hypoglycaemia in normal subjects (Gold et al., 1995; McCrimmon et al., 1996). Although verbal memory and concept formation in particular, are the most commonly affected cognitive functions in Type II diabetes, these functions are not impaired in experimentally induced hypoglycaemia. A number of variables influence response variability to hypoglycaemia (Deary, 1997). These include gender (men are more affected than women); previous hypoglycaemic episodes and hypoglycaemic unawareness (heightens effect); IDDM (diabetics show stronger effects): IQ (subjects with higher IQ show greater impairment and this may interact with task type), though the evidence for the effect of intelligence is rather equivocal and could represent regression to the mean. In summary, lack of glucose availability produces definite impairment of cognitive performance in diabetics of both types and in normal subjects (Deary, 1998, 1999). It follows that enhancing glucose availability via the ingestion of carbohydrate should ameliorate impairment or indeed enhance performance. In addition, glucose tolerance and glucose regulation should also be important for the protection of cognitive function. On a wider scale this is important because of the significant increase in the incidence of impaired glucose tolerance (IGT) and type 2 diabetes. Recent figures from the International Diabetes Federation suggest that 8.2% of the global adult population may have IGT, a condition that precedes the development of diabetes in 70% of cases of NIDDM. These figures are expected to rise ± 9% of the adult population will have IGT by 2025 and 6.3% will have diabetes with two-thirds of all cases in the developed world. The implications for cognitive function and age-related memory impairment are significant.

18.3.4 Insulin The pathogenesis of diabetes mellitus (IDDM and NIDDM) and a number of other neurological diseases which have cognitive deficits as a central feature, e.g. Alzheimer's disease may be related to insulin level or insulin insensitivity (Park, 2001). Reduced insulin concentration and insulin receptor densities have

ß 2004, Woodhead Publishing Limited

been observed in such vulnerable populations, compared to healthy controls (Zhao and Alkon, 2001). Additionally, Craft et al. (1996) identified no memory improvement when the natural insulin rise in response to a glucose load, which has previously been associated with memory enhancement, was prevented. Such effects maybe a result of increased insulin mediated chaperoning of glucose into brain neurons maintaining a stable and sufficient energy supply to active brain regions (Zhao and Alkon, 2001). Insulin has also been suggested to account for cognitive improvements following the ingestion of glucose.

18.3.5 Glucose administration and mental performance Glucose administration has generally been shown to have beneficial effects in young adults and the elderly (Gonder-Frederick et al., 1987; Hall et al., 1989; Lapp, 1981). A series of studies in elderly subjects have shown enhanced effects on working memory (short-term) but not digit span, tests of attention or motor function (Manning et al., 1990). Similar effects in the same cognitive domains have been demonstrated in young adults (Benton and Sargent, 1992). In particular, glucose administration seems to facilitate memory performance and speed of reaction time. Reaction time tasks appear to be sensitive to changes in blood glucose. However, glucose administration leaves performance on digit span, tests of attention and motor function, unaltered (Manning et al., 1990). The performance-enhancing effects of glucose have been suggested to be due simply to the provision of readily utilizable energy. For example, in one study, the performance of subjects who had consumed breakfast, or those who had not, were compared on three memory tasks (Benton and Parker, 1998). The `no breakfast' condition was associated with poorer recall, but this effect was reversed by the administration of a glucose drink at breakfast time, in two of the three memory tasks. However, other studies have demonstrated that readily available energy from a high CHO drink enhanced the number of words correctly recalled, compared to an equicaloric low CHO drink (Lluch et al., 2000). Moreover, change in blood glucose was a significant covariate in the high CHO condition, but not in the low CHO condition for tests of arousal (CFFT) and memory performance. Poorer performance may be due to a lack of available energy rather than energy per se and consequently lower blood glucose.

18.3.6 Cognitive demand Taken together research demonstrates fairly conclusively that glucose can enhance memory and attentional processes provided tests of sufficient difficulty are used to produce high cognitive demand in normal adults who are less vulnerable to cognitive impairment than undernourished or elderly samples. Cognitive demand is therefore a key feature of studies which demonstrate facilitatory effects of glucose on mental performance. Test difficulty influences the apparent effects of glucose on cognition. More difficult tests produce a high cognitive demand. Cognitive demand increases glucose uptake (Scholey et al., 2001). Scholey et al. (2001)

ß 2004, Woodhead Publishing Limited

administered 25 g glucose and tested memory performance after either serial sevens (a highly demanding task), or a control task. Both tasks were followed by word retrieval (verbal fluency). Glucose consumption improved performance on serial sevens and verbal fluency. Peripheral measures of blood glucose decreased under the high cognitive load. This suggests that intense cognitive processing leads to increased neural glucose utilization. The process by which glucose exerts this effect is not clear. Frontal lobe functioning is enhanced by glucose, suggesting that effects may be due to an increased supply of glucose to the brain areas involved in memory tasks (Jonides et al., 1997; Kennedy and Scholey, 2000). For example, evidence from Positron Emission Tomography (PET) suggests that increased neural activity (e.g. learning a complex visuo-spatial motor task or verbal working memory) is associated with an increased use of glucose by the brain (Biessels, 1999). After learning a task, use of glucose in extraneous brain areas decreased (Haier et al., 1992). Cognitive demand therefore, seems to be associated with increased glucose metabolism in localized areas of the brain (Benton et al., 1996; Kennedy and Scholey, 2000) and consistent with this cognitively demanding situations can deplete the brain of glucose. Alternatively or additionally, glucose or a metabolite may activate release of acetylcholine during learning to produce improved memory performance (Korol and Gold, 1998). Indeed, systemic injections of glucose facilitate the hippocampal release of acetylcholine which may account in part for effects on memory (Ragozzino et al., 1996).

18.3.7 Glucose tolerance Glucose tolerance (i.e., the ability to uptake glucose from the bloodstream to tissues and to the brain) is posited as important in mediating nutritional effects on cognitive function. There are substantial individual differences in glucose tolerance (Jonides et al., 1997), as reflected in susceptibility to impairment of cognitive performance. Subjects with higher levels of blood glucose and good glucose tolerance respond most efficiently to the demand of cognitive tasks (Elias et al., 1997). Glucose tolerance can be improved by exercise and frequent, small meals, but effects on cognitive performance remain unknown. Recent studies suggest that cognitive functions may be impaired before gluco-regulation is disrupted to the extent evident in type II diabetes (Messier et al., 2003). Some studies suggest that stable performance is related to balanced glucose metabolism and state of metabolic activation (Fischer et al., 2001). This may be mediated via an action of glucose on the cholinergic system (Owens and Benton, 1994). A series of studies have demonstrated that good glucose tolerance and the ability to respond efficiently to a glucose load or a cognitively demanding memory task are important features predicting optimal mental performance (Benton, 2001). The effect of 50 g glucose on memory for trigrams depended on the initial blood glucose levels of the subject (Martin and Benton, 1999). Those with higher levels of blood glucose showed better performance. However, after the glucose drink, falling levels of blood glucose were associated with better memory

ß 2004, Woodhead Publishing Limited

performance. The decline in blood glucose during task performance is proposed to indicate that increased neural uptake of glucose results in better memory. It may be that absolute levels of blood glucose are unimportant in relation to cognitive function, except when hypoglycaemia occurs and cognitive performance is affected by glucose tolerance. Donohoe and Benton (1997) reported that recovery from nadir (how quickly the subject recovers from a fall in glucose) showed a stronger relationship to cognitive performance than the hypoglycaemic index or area under the curve (AUC). Whether levels of blood glucose were rising or falling was also an important predictor of performance in Benton and Owens (1993) study. Rising levels were associated with better recall than falling levels. In addition, higher levels of blood glucose correlated with faster information processing, better word recall and improvement on the Stroop test. Thus glucose tolerance may account for changes in cognitive function following administration of glucose.

18.4 Mental performance: the effects of CHO and glucose during the day 18.4.1 Effects of CHO on cognition The vast majority of studies which have examined the effects of nutrients on mental performance have used interventions in liquid form, notably glucose or high CHO drinks. It is relatively straightforward to produce an identical placebo for a drink manipulation. This is far more difficult when real foods are used or solid macronutrient manipulations are to be made. Thus CHO and glucose have been reported to both enhance and impair cognitive function (Korol and Gold, 1998; Bellisle et al., 1998). A simple experimental strategy is to compare noenergy versus some-energy manipulations.

18.4.2 Macronutrient manipulations involving a no-energy versus someenergy manipulation Two studies using high CHO (high energy) snacks have examined effects on four cognitive tasks: digit span, arithmetic reasoning, reading and attention (Kanarek and Swinney, 1990). In comparison with a diet soft-drink (containing aspartame) the high CHO snack improved digit span performance (greater recall) and enhanced performance in the attention task (continuous performance) in the late afternoon. When calorie-containing yoghurts were compared with soft drinks, subjects solved more arithmetic problems in a shorter time after the yoghurt. The consumption of a prior lunch had no effect on this performance.

18.4.3 Direct comparisons of pure carbohydrate Comparing the effects of pure macronutrients is difficult because few studies, especially in humans, have administered each macronutrient in the same

ß 2004, Woodhead Publishing Limited

experimental situation. Thus studies generally differ in terms of the energy consumed, the type of macronutrient, e.g., glucose or starch, MUFA or PUFA; and the measure of cognitive performance employed. There have been only two studies to date, which have administered isoenergetic loads of each macronutrient in liquid or spoonable cream form and subjected participants to a battery of cognitive tests. Cunliffe et al. (1997) gave 400 kcal of mixed CHO, fat and protein, or of pure CHO (maltodextrin) or pure fat as breakfast and found slower reaction times and greater subjective feelings of fatigue after the pure carbohydrate drink. Fat ingestion did not affect reaction time but decreased flicker fusion frequency indicating increased central fatigue or decreased visual information processing capacity. In a well-controlled study Fischer et al. (2001) used test meals (400 kcal) of either pure CHO (glucose, maltodextrin and rice starch), protein or fat (a combination of saturated and unsaturated fatty acids). Postprandial cognitive performance was best after fat ingestion. In contrast, CHO and protein ingestion affected different aspects of cognitive performance. CHO ingestion resulted in relatively better short-term memory and accuracy of tasks at the expense of efficiency (fewer errors on RT) whereas protein ingestion resulted in better attention and efficiency of tasks. The effects on peripheral versus central tasks reflected a different mechanism of action of CHO and protein such that CHO resulted in relatively lower peripheral performance and better central performance while the opposite pattern was observed following protein ingestion. Fat had no effect on reaction time. The authors attributed this to the steady plasma concentrations of glucose, insulin and glucagon, produced by fat consumption. Moreover, they observed that the ingestion of pure CHO induced slower reaction time, whilst pure protein produced faster reaction times on demanding complex (but not simple) reaction time tasks. This effect was only apparent two hours after ingestion, reflecting differences in the absorption and metabolism of the two macronutrients. The first hour following ingestion saw the reverse pattern. CHOs may not necessarily exert their effect purely by altering availability of glucose but might alter serotoninergic function with consequences for performance (Altman and Normile, 1988; Markus, 1999).

18.4.4 Tryptophan, 5-HT and mental performance If the effects of CHO are mediated via changes in 5-HT function, then administering the amino acid tryptophan should produce similar effects to eating CHO. Lieberman et al. (1983) compared tryptophan and tyrosine on four performance tasks: simple auditory reaction time, two-choice reaction time, grooved pegboard and Thurstone tapping task. The tryptophan treatment was different from tyrosine only on simple auditory reaction time but neither amino acid treatments differed from its own placebo. However, tryptophan did increase subjective ratings of fatigue-inertia and decreased feelings of vigour-activity on rating scales. Reaction time and visual information processing (measured by Critical Flicker Fusion Threshold) were significantly increased by a

ß 2004, Woodhead Publishing Limited

decaffeinated Coca-Cola drink with added tryptophan when compared with a decaffeinated Coca-Cola drink alone (Cunliffe et al., 1998). These amino acid manipulations suggest impairment of performance follows protein consumption. The increased subjective fatigue observed is similar to that found after CHO consumption (Reid and Hammersley, 1999). In addition, providing a supplementation of branched-chain amino acids (BCAAs) has been shown to prevent impairment of performance of shape rotation and identification tasks and to improve performance of the Stroop task after exercise (Hassman et al., 1994). Since exercise increases the plasma concentration ratio of free tryptophan to other LNAAs leading to an elevation of 5-HT in the brain, BCAAs may therefore minimize central and mental fatigue during and after sustained exercise.

18.4.5 Carbohydrate and protein manipulations A number of psychological changes are predicted from diets varying in CHO and protein content but the evidence is quite weak. However, several studies have found that high CHO meals tend to produce greater drowsiness, sleepiness and calmness (for review see Spring et al., 1987; Reid and Hammersley, 1999). In general, studies comparing CHO and protein have found slower reaction time and impaired attention following high CHO meals. High-protein meals produce susceptibility to distraction and slower memory scanning. These results suggest that the two macronutrients are operating on different cognitive processes and may be task specific. For example, Smith et al. (1988) examined performance on categoric search and focused attention tasks following either a high starch, high sugar or high protein lunch. Slower reaction times to peripheral stimuli on a visual search task were observed after both CHO meals than after the high protein meal, indicating little difference between the simple and complex CHO on this aspect of cognitive performance. High protein meals resulted in greater distraction from stimuli in the focused attention task. Cadets undergoing a demanding military combat training course were given either five daily doses of a protein-rich drink containing 2 g tyrosine, or a carbohydrate rich drink each with 255 kcal (Deijen et al., 1999). Assessments were made both immediately prior to the combat course and on the sixth day of the course. The group supplied with the tyrosine-rich drink performed better on a memory and a tracking task than the group supplied with the carbohydrate-rich drink. Tyrosine may also sustain working memory in situations requiring multitasking (Thomas et al., 1999). Stress may potentiate the effects of food on mental performance by modulating brain serotonin function. Chronic stress may produce increased serotonin activity leading to a functional shortage of tryptophan and brain serotonin concentrations. Markus et al. (1999) demonstrated that performance of stress-prone subjects on a Sternberg memory-scanning task improved after a carbohydrate-rich, protein-poor (CR/PP) diet compared to a protein-rich, carbohydrate-poor (PR/CP). This effect was observed after exposure to a

ß 2004, Woodhead Publishing Limited

controllable stressor but not following exposure to an uncontrollable stressor. An earlier study by the same author did not find an effect on cognitive performance but this was attributed to a low level of cognitive demand and the lack of a controllable stressor condition (Markus et al., 1998).

18.4.6 Carbohydrate and fat manipulations Although much more research has compared CHO and fat, few decisive conclusions can be drawn. This is due to a series of methodological inconsistencies between studies rendering comparison difficult. In comparisons of low/high CHO and low/high fat lunches, no effects were seen on psychomotor tasks (Kelly et al., 1994) but with only six subjects, the power was obviously low. A further study showed that optimal performance was seen with medium fat/medium CHO lunch. Reaction times were worse following the high fat/low CHO lunch and the low fat/ high CHO lunch (Lloyd et al., 1994). A similar non-significant trend was found following a high fat/low CHO breakfast (Lloyd et al., 1996). Higher than usual proportions of either fat or CHO caused subjects to be more drowsy, uncertain and muddled. Although no clear differences in performance were observed the authors suggest that cognitive efficiency may also be impaired. In another study, a comparison of a high fat/low CHO and a low fat/high CHO meal eaten either as a brunch or a lunch showed that subjective lassitude increased following ingestion of all meals. Little change was found in performance (Wells and Read, 1996). It was alleged that fat in the morning caused a greater depression of alertness and mood. Smith et al. (1994a) showed that a high fat lunch did not affect logical reasoning or vigilance but did produce slower but more accurate performance on selective attention tasks. In a wellcontrolled, repeated measures design study, reaction time was slower after a high fat lunch compared with a low fat lunch (Lluch et al., 2000). These studies demonstrate the lack of clarity of effect of macronutrient manipulations on various measures of cognitive performance. Some of these inconsistencies are due to confounding variables which apply to many studies in this area and which are discussed later in this chapter.

18.4.7 Performance and the GI of foods The glycaemic index (GI) represents the glucose raising potential of CHO foods and is defined as the incremental area under the glucose response curve (AUC) following the ingestion of a 50 g CHO reference food, most commonly glucose, or bread (Wolever et al., 2003). The glycaemic response to different starchy foods, classified by the glycaemic index can vary over a three to four fold range (Jenkins et al., 1981) probably because of differences in the rates at which foods are digested and absorbed from the gastrointestinal tract (Jenkins et al., 1982). The GI of foods is usually determined by tests conducted in the morning after an overnight fast (Wolever et al., 1991). Tables of the GI of hundreds of food items are now available (Foster-Powell et al., 2002). Carbohydrates can also be

ß 2004, Woodhead Publishing Limited

categorized as providing slowly available glucose (SAG) and rapidly available glucose (RAG) (Englyst et al., 1999). The RAG fraction is rapidly absorbed and largely determines the GI of a food. However, in research and real world situations, a number of other factors may influence the actual glycaemic response to particular foods. Firstly, blood samples are required to measure blood glucose in response to a nutritional challenge. Capillary blood samples have been associated with greater reliability and higher glycaemic responses to food ingestion than those obtained via venous plasma (Wolever and Bolognesi, 1996). Secondly, time of day may affect the absolute and relative glycaemic effect of foods. For instance, high and low GI cereal based foods given at breakfast and at lunch produced different glycaemic response curves. Glycaemic response was greater at breakfast than at lunch and most closely approximated the calculated GI. Differences between two test meals were smaller. The low and high GI test meals barely differed at lunchtime (21% versus 50%) and responses were more variable at lunchtime (Wolever and Bolegnesi, 1996). It has also been suggested that the blood glucose concentration in response to foods (the assumed GI) reflects a balance of both the entry and removal of glucose into and from the blood rather than simply the rate of digestion and entry of glucose into the systemic circulation (Schenk et al., 2003). Effects on glycaemic response and cognitive performance produced via the ingestion of pure macronutrients are likely to be less subtle than macronutrients ingested in the form of meals of mixed composition. The CHO, fat and protein content of a meal affects rate of gastric emptying and modifies the glycaemic index and insulin response. Ingestion of mixed composition meals means that macronutrient interactions may occur. For instance, a reduced glucose response has been demonstrated following the combined ingestion of protein and glucose, compared with glucose consumption alone. The effects of such interactions on metabolism and the implications of this for mental performance have not been considered by any previous studies. In order to achieve ecological validity the three macronutrients need to be examined concurrently in experimental manipulations.

18.4.8 Effects of GI on cognitive performance Strong theoretical evidence exists which supports the ameliorating role of blood glucose on cognitive performance (for review see Benton, 2001; Benton and Nabb, 2003). Interest in the potential impact of foods varying with respect to GI is growing and although there have been only a handful of studies to date, findings are conflicting. In a recent breakfast study Kaplan et al. (2000) compared the ingestion of 50 g glucose, mashed potato and barley. Barley consumption produced the strongest and most rapid memory enhancing effects compared with reduced and slower amelioration following consumption of the two high glycaemic index (HGI) meals (glucose and mashed potato). The barley meal induced a blood glucose rise that did not exceed the normal range (4.0± 5.5 mmol/l). These results imply that a lower or more stable level of blood glucose may be optimal for cognitive function.

ß 2004, Woodhead Publishing Limited

Benton et al. (2003) examined the effects of slow and rapid acting glucose given as breakfast in a between subjects design. Blood glucose response to the RAG was higher than to the SAG breakfast in the first 120 minutes post consumption but memory performance differed only at 150 and 210 minutes post consumption by which time blood glucose did not differ. The authors concluded that a high SAG breakfast was associated with better memory throughout the morning. To date, no studies examining the effect of dietary manipulations at lunchtime on cognitive performance have controlled the GI of the meal. However, Fischer et al. (2001) administered CHO and PRO in a pure, cream-like form. Overall improvement of accuracy and attention was observed following the ingestion of the PRO meal which was concomitant with minimal metabolic activation and subsequent stability of blood glucose. A similar relationship between lunchtime fat intake and cognitive performance has also been found (Smith et al., 1994b). The CHO meal in the Fischer et al. study was also associated with better performance in central, versus peripheral tasks. In a study examining the effects of high and low GI meals on cognitive performance at lunchtime, Handley et al. (2003) compared CHO based low and high GI meals. Consumption of the LGI meal was related to improved accuracy across various cognitive domains and was accompanied by a stable blood glucose profile. This lends support for the role of a stable supply of neural fuel to the brain in performance amelioration (Fischer et al., 2002). Handley et al. (2003) also demonstrated that calculated GI does not necessarily predict the blood glucose response to mixed composition meals. Interactions can occur on a number of levels and include the physiology of individual, their glucose tolerance and insulin sensitivity as well as macronutrient interactions which may be compounded by processing and cooking methods. It is therefore, impossible at present to predict the effect of a combined macronutrient meal on cognitive performance.

18.4.9 Effect of palatability on performance The palatability of a food is a result of the integration of orosensory and postingestive stimuli. Like hunger, palatability is a hypothetical construct, that is, an explanatory concept that cannot itself be directly observed but is inferred from operationally defined and measurable events (Blundell and Rogers, 1991). Whilst the relationship between palatability and food intake has been examined in many studies, the relationship between palatability and cognitive performance has not. There have been no studies to date that have directly measured the effects of palatability per se on cognitive performance. It is, however logical that palatability could influence cognitive performance via the mediating effects of mood (e.g. palatable food-induced increase in endorphins). That is, if an increase in palatability causes an increase in positive mood, then performance could also be enhanced. Despite the lack of evidence on this topic, some experiments intended to examine the effects of foods (e.g. macronutrient content, energy

ß 2004, Woodhead Publishing Limited

value, energy density) on cognitive performance have inadvertently allowed taste or palatability characteristics to vary. Recent research has demonstrated that circulating glucose is higher after a palatable meal than after a meal composed of the same constituents presented in a non-palatable form. (Sawaya et al., 2001) Although cognitive performance was not measured in this study, it raises the possibility that cognitive performance may be influenced by the mediating effect of palatability on blood glucose rather than as supposed above via effects on mood. Handley et al. (2003) found that a highly palatable high GI CHO based meal resulted in faster reaction times and augmented blood glucose response than a high GI protein based meal although mood did not differ in relation to the two meals.

18.4.10 Time of day affects metabolic responses to food Circadian patterns in mood may interact with food intake and physiological responses to meals such as glucose and cortisol secretion vary across the day (Biston et al., 1996). For instance, alertness rises to peak in midmorning, then falls throughout the day (DeCastro, 1987). Cortisol's natural rhythm declines over the course of the day. These rhythms could explain the different effects of meals on mood and performance over the day. Thus, the effects of alertnessenhancing foods are most likely to be detected if they are ingested at a time when alertness is naturally declining, and cognitive capacity is challenged by an appropriate cognitive test. For instance, the beneficial effects of afternoon snacks do not occur in late morning (Kanarek and Swinney, 1990). Clearly, then, the selection of tasks for a test battery should be made with these considerations in mind. Furthermore, physiological measures, e.g., blood glucose, insulin, fatty acids are important biomarkers of the metabolic state of the individual following food consumption.

18.4.11 Effect of CHO at breakfast On waking, people have effectively fasted for 6±8 hours and cognitive loads which demand extra energy are likely to be identified. One experimental strategy which takes advantage of this naturally occurring event is to examine the effect of breakfast consumption on mental performance. Poorer performance may be due to a lack of available energy and consequently lower blood glucose. Most breakfast studies to date have not considered the macronutrient composition of the meal but have quantified breakfasts in terms of high, low or no energy. A number of studies, beginning with the Iowa breakfast studies (Tuttle et al., 1949, 1950, 1952, 1954), have demonstrated that missing breakfast can have detrimental effects on performance of reaction time tasks, spatial memory and immediate word recall (Benton and Sargent, 1992; Smith et al., 1994a). While some aspects of memory seem to be susceptible to the effects of missing breakfast, other aspects of performance are not affected, e.g., addition, sentence verification (Smith et al., 1994a; Dickie and Bender, 1982; Smith and

ß 2004, Woodhead Publishing Limited

Kendrick, 1992). Benton and Parker (1998) compared subjects who had and had not eaten breakfast on three memory tasks. The no-breakfast condition was associated with poorer recall but this effect was reversed by administration of a glucose drink at breakfast time for two of the three memory tasks. These studies examined well-nourished populations and the negative effects of breakfast omission may be due to the disruption of a regular pattern of consumption rather than lack of available energy. Effects are greatest in low income or undernourished children whose scholastic performance increases following breakfast (Pollitt et al., 1998; Pollitt and Matthews, 1998).

18.4.12 Effects of CHO at lunchtime No studies examining the effects of CHO at lunchtime have described the Glycaemic Index (GI) of the dietary manipulation. Manipulations of the proportions of CHO and fat in lunchtime meals have produced diverse outcomes; some have shown CHO to produce a better effect on performance than fat, whilst others have shown the converse (Lloyd et al., 1994; Lluch et al., 2000; Smith and Miles, 1986; Spring et al., 1983). Higher than usual proportions of fat or CHO may induce drowsiness or uncertainty, or potentially affect cognitive efficiency and motor performance (Reid and Hammersley, 1999). Two studies (Spring et al., 1983; Lieberman et al., 1986) compared the performance effects of high CHO and high protein meals. One found no effect of either meal on auditory reaction time, irrespective of whether the meal was given at breakfast or lunch (Spring et al., 1983). The nature of the CHO, and the timing of ingestion, may be important in explaining the different effects in each study.

18.4.13 Post-lunch dip Performance on sustained attention tasks is impaired in the early afternoon compared with late morning irrespective of food consumption (Smith and Miles, 1986). A review of six studies examining circadian rhythms in mental performance (Monk and Folkard, 1985) reported evidence of a deficit in performance in the early afternoon, with lapses of attention most commonly reported. It is difficult to untangle the effects of consumption of lunch from the effects of the underlying circadian rhythm or `post-lunch dip'. Some tasks seem to be impaired, irrespective of whether lunch was consumed and this has been interpreted to reflect a post-lunch dip. Equally though, it could reflect the effect of negative energy balance or a lack of available glucose. Increasing meal size at lunchtime produces more momentary lapses of attention). This post-lunch dip effect is abolished by caffeine (Smith et al., 1994a).

18.4.14 Effects of CHO in the evening Few studies have examined the effects of evening meals. One study showed that logical memory was improved by a three-course meal of between 1200 and

ß 2004, Woodhead Publishing Limited

1500 kcals, compared to no meal. No differences in sustained attention, word recall or word recognition were noted (Smith et al., 1994b). However, the meal was a free choice selection from a college refectory and one can only assume that it had a mixed macronutrient composition and similar amounts were consumed by each participant.

18.4.15 Methodological considerations concerning mental performance A number of methodological issues which limit the comparability of studies of CHO on mental performance have been discussed (Dye et al., 2000; Lieberman, 2003). These include the failure to recognize the importance of differences in total energy supplied by different nutritional manipulations and the lack or appropriate placebo or control treatments. In a number of studies, non-identical control treatments have been employed or it has been assumed that a non-nutritive control is appropriate even though this could exert an effect by affecting GI response. Cognitive demand reportedly increases the uptake of glucose. The order of tests in a battery therefore warrants careful consideration and piloting with respect to the proposed intervention. A lack of effect may occur because the most demanding task is placed in a particular order in the battery (e.g. first or at the peak of a nutrient effect), when resources are most plentiful. Thus, the timing of tests after consumption and the relative order of tests within a battery must be considered in conjunction with data about the GI of the food, details of its likely course of digestion, and metabolism. In addition, measurements of blood glucose immediately after cognitive testing may reflect arousal or stress produced by the cognitive tests (perhaps exacerbated by evaluation or apprehension of the participants), rather than an exclusive response to a nutrient intervention.

18.5 Future trends Varying the glycaemic index (GI) offers a considerable opportunity for functional food development, with potentially positive effects on the extent and timing of post lunch inefficiency. Maintaining stable blood glucose appears to be beneficial. This can inform advice about breakfast consumption, snacking, and the development of functional foods, which modulate blood glucose. The effect of CHO throughout the day is therefore an important consideration. Impaired glucose tolerance, sedentary lifestyles and increasing incidence of diabetes all have implications for cognitive function. The prevalence of agerelated memory decline is likely to increase and this feature of mental performance is most susceptible to manipulations involving glucose and CHO. Hence there is tremendous scope for the development of functional foods to prevent cognitive impairment. However, a great deal of research is needed to understand what aspects of mental performance are influenced by which aspects of the nutritional supply. In relation to CHO, the impact of the GI of foods on cognitive function is poorly understood and further research is required.

ß 2004, Woodhead Publishing Limited

18.6

References

et al., Substrate turnover during prolonged exercise in man. J Clin Lab Invest 1974; 53: 1080±1090. ALTMAN HJ, NORMILE HJ, What is the nature of the role of the serotonergic nervous system in learning and memory: prospects for development of an effective treatment strategy for senile dementia. Neurobiology of Ageing 1988; 9: 627±638. AMELSVOORT JMM VAN, WESTRATE JA VAN, Amylose-amylopectin ration in a meal affects post prandial variables in male volunteers. Am. J. Clin. Nutr. 55; 712±718, 1992. AMIEL SA, POTTINGER RC, ARCHIBALD HR, et al., Effect of antecedent glucose control on cerebral function during hypoglycaemia. Diabetic Care 1991; 14: 109±118. ANDERSON IM, PARRY-BILLINGS M, NEWSHOLME EA, FAIRBURN CG, COWEN PJ, Dieting reduces plasma tryptophan and alters brain 5-HT function in women. Psychological Medicine 1990; 20: 785±791. ASHLEY DVM, LIARDON R, LEATHWOOD PD, Breakfast meal composition influences plasma tryptophan to large neutral amino acid ratios of healthy lean men. Journal of Neural Transmission 1985; 63: 271±283. AUSTIN EJ, DEARY IJ, Effects of repeated hypoglycemia on cognitive function: a psychometrically validated reanalysis of the Diabetes Control and Complications Trial data. Diabetes Care 1999; 22: 1273±1277. BALSOM PD, WOOD K, OLSSON P, EKBLOM B, Carbohydrate intake and multiple sprint sports: with special reference to football (soccer). Int J.Sports Med. (1999) 20; 48±52. BELLISLE F, BLUNDELL JE, DYE L, et al., Functional food science and behaviour and psychological functions. British Journal of Nutrition 1998; 80: S173±193. BENTON D, The impact of the supply of glucose to the brain on mood and memory. Nutrition Reviews. 2001, 59, (1) S20±S21 Part 2. BENTON D, Carbohydrate ingestion, blood glucose and mood. Biobehavioural Reviews. 2002, 26(3): 293±308. BENTON D, NABB S, Carbohydrate, memory, and mood. Nutrition Reviews. 2003, 61(5 pt 2): 561±67. BENTON D, OWENS D, Blood glucose and human memory. Psychopharmacology 1993; 113: 83±88. BENTON D, PARKER PY, Breakfast blood glucose and cognition. American Journal of Clinical Nutrition 1998; 67: 772S±778S. BENTON D, SARGENT J, Breakfast blood glucose and memory. Biological Psychology 1992; 33: 207±210. BENTON D, OWENS DS, PARKER PY, Blood glucose influences memory and attention in young adults. Neuropsychologia 1994; 32, 595±607. BENTON D, PARKER PY, DONOHOE R, The supply of glucose to the brain and cognitive functioning. Journal of Biosocial Science 1996; 28: 463±479. BENTON D, RUFFIN M-P, LASSEL T; NABB S, MESSAOUDI M, VINOY S, DESOR D, LANG V, The delivery rate of dietary carbohydrates affects cognitive performance in both rats and humans. Psychopharmacology. 2003, 166 (1), 86±90. È M J, HULTMAN E, The effect of exercise on muscle glycogen and electrolytes in BERGSTRO normals. Scand J Clin Lab Invest 1966; 18: 16±20. È M J, HULTMAN E, A study of the glycogen metabolism during exercise in man. BERGSTRO Scand J Clin Lab Invest 1967a; 19: 218±228. È M J, HULTMAN E, Synthesis of muscle glycogen in man after glucose and BERGSTRO fructose infusion. Acta Med Scand 1967b; 182: 93±107. AHLBORG G, FELIG P, HAGENFELDT L

ß 2004, Woodhead Publishing Limited

BIESSELS GJ,

Cerebral complications of diabetes: clinical findings and pathogenetic mechanisms. Netherlands Journal of Medicine 1999; 54: 35±45. BISTON P, VAN CAUTER E, OFEK G, LINKOWSKI P, POLONSKY KS, DEGAUTE JP, Diurnal variations in cardiovascular function and glucose regulation in normotensive humans. Hypertension. 1996; 28 (5): 863±71. BJORNTORP P, Importance of fat as a support nutrient for energy: metabolism of athletes. J Sports Sci 1991; 9: Special Issue, pp. 71±76. BLUNDELL JE, ROGERS PJ, Hunger, hedonics, and the control of satiation and satiety. In Chemical Senses. Appetite and Nutrition, vol. 4, 1991; pp. 127±148 (M Friedman, MG Tordoff and M Kare, eds). New York: Marcel Dekker, Inc. BROUNS F, Etiology of gastrointestinal disturbances during endurance events. Scand J Med Sci Sports 1: 66±77, 1991a. BROUNS F, Gastrointestinal Symptoms in Athelets: Physiological and Nutritional Aspects. In; Advances in Nutrition and Top Sport, Brouns F, Saris WHM, Newsholme EA (eds) Med Sport Sci 1991b; 32: 166±199, Karger, Basel. BROUNS F, The effect of athletic training and dietary factors on the modulation of muscle glycogen. Proceedings of The Clinical Pharmacology of Sport and Exercise, T Reilly and M Orme (eds) Sitges, Spain, Elsevier Science Publishers, pp. 181±193, 1997. BROUNS F, Gastric emptying as a regulatory factor in fluid uptake. Int J Sports Med 19: S125±S128, 1998. BROUNS F, Sport Nutrition Essentials, Wiley Press, London 2002. BROUNS F, BECKERS E, Is the gut an athletic organ? Sports Med 15 (4): 242±257,1993. BROUNS F, SARIS WHM, TEN HOOR F, Dietary problems in the case of strenuous exercise: part 1±a literature review, Part 2 ± the athletes diet: nutrient dense enough? J Sportsmed Phys Fitness 1986; 26: 306±319. BROUNS F, SARIS WHM, REHRER NJ, Abdominal complaints and gastrointestinal function during long-lasting exercise. Int J Sports Med 1987; 8: 175±189. BROUNS F, REHRER NJ, SARIS WHM, BECKERS E, MENHEERE P, TEN HOOR F, Effect of carbohydrate intake during warming-up on the regulation of blood glucose during exercise. Int J Sports Med 1989; 10: (suppl 1) S68±S75. BROUNS F, REHRER NJ, BECKERS E, SARIS WHM, MENHEERE P, TEN HOOR F, Reaktive HypoglykaÈmie. Deutsche Zeitschr Sportmed 1991; 42 (5): 188±200. BROUNS F, SENDEN J, BECKERS EJ, SARIS WHM Osmolarity does not affect the gastric emptying rate of oral rehydration solutions. J. Parenteral and Enteral Nutrition 19 (5): 387±391, 1995. BROUNS F, ARRIGONI E, KETTLITZ B, Resistant starch and the butyrate revolution. Trends in Food Science and Technology, 13; 251±261, 2002a. BROUNS F, VAN NIEUWENHOVEN M, JEUKENDRUP A, VAN MARKENLICHTENBELT W, Functional foods and food supplements for atheletes: from myths to benefit claims substantiation through the study of selected biomarkers. Brit J. Nurt. 88, Suppll 2, S177±S186, 2002b. BURKE LM, CLAASSES A, HAWLEY JA, NOAKES TD, Carbohydrate intake during prolonged cycling minimizes effect of glycemic index of preexercise meal. J. Appl. Physiol (1998) 85 (6) 2220±2226. CHAOULOFF F, Physiopharmacological interactions between stress hormones and central serotonergic systems, Brain Research Reviews 1993; 18: 1±32. CHRYSSANTHOPOULOS CC, WILLIAMS W, WILSON et al. Comparison between carbohydrate feedings before and during exercise on running performance during a 30km

ß 2004, Woodhead Publishing Limited

treadmill time trial. Int. J. Sport Nutr. 1994, 4: 374±386. Carbohydrates for exercise: dietary demands for optimal performance. Int J Sports Med 1988; 9: 1±18. COYLE EF, Carbohydrate Feedings: Effects on Metabolism, Performance and Recovery. In: Advances in Nutrition and Top Sport, Brouns F, Saris WHM, Newsholme EA (eds), Med Sport Sci 1991a; 32: 1±14, Karger, Basel. COYLE EF, Timing and method of increased carbohydrate intake to cope with heavy training, competition and recovery. J Sports Sci 1991b; 9: Special Issue, pp. 29±52. COYLE EF, COGGAN AR, Effectiveness of carbokydrate feeding in delaying fatigue during prolonged excercise. Sports med 1984; 1: 446±458.

COSTILL DL,

CRAFT S, NEWCOMER J, KANNE S, DAGOGO-JACK S, CRYER P, SHELINE Y, LUBY J, DAGOGO-JACK A, ALDERSON A,

(1996) Memory improvement following induced hyperinsulinemia in Alzheimer's Disease. Neurobiology of Aging, 17, 123 130. CRAPO PA, Simple versus complex carbohydrate use in the diabetic diet. Ann Rev Nutr 1985; 5: 95±114. CUNLIFFE A, OBEID OA, POWELL-TUCK J, Post-prandial changes in measures of fatigue: effect of a mixed or a pure carbohydrate or pure fat meal. European Journal of Clinical Nutrition 1997; 51: 831±838. CUNLIFFE A, OBEID OA, POWELL-TUCK J, A placebo controlled investigation of the effects of tryptophan or placebo on subjective and objective measures of fatigue. European Journal of Clinical Nutrition 1998; 52: 425±430. DE CASTRO JM, Circadian rhythms of the spontaneous meal pattern, macronutrient intake, and mood of humans. Physiology and Behaviour 1987; 40: 437±46. DEARY IJ, The impact of acute hypoglycaemia on cognitive function and mood in humans. Proceedings of the British Psychological Society 1997; 5: 24 (Abstract). DEARY IJ, The effects of diabetes on cognitive function. Diabetes Annual 1998; 11: 97±118. DEARY IJ, Symptoms of hypoglycaemia and effect on mental performance and emotions. In Frier BM, Fisher BM, eds, Hypoglycaemia in Clinical Diabetes. Chichester: Wiley 1999: 29±54. DEARY IJ, CRAWFORD JR, HEPBURN D et al. Severe hypoglycemia and intelligence in adult patients with insulin-treated diabetes. Diabetes 1993; 42: 341±344. DEÂCOMBAZ J, SARTORI D, ARNAUD MJ et al. Oxidation and metabolic effects of fructose and glucose ingested before exercise. Int J Sports Med 1985; 6: 282±286. DEIJEN JB, WIENTJES CJE, VULLINGHS HFM, CLOIN PA, LANGEFELD JJ, Tyrosine improves cognitive performance and reduces blood pressure in cadets after one week of a combat training course. Brain Research Bulletin 1999; 48: 203±209. DICKIE NH, BENDER AE, Breakfast and performance in schoolchildren. British Journal of Nutrition 1982; 48: 483±96. DONOHOE RT, BENTON D, Glucose tolerance and psychological functioning. Proceedings of the British Psychological Society 1997; 5: 124. DREHER ML, DREHER CJ, BERRY JW, Starch digestibility of foods: a nutritional perspective. CRC Critical Rev, 1984; 20: 47±71. DYE L, LLUCH A, BLUNDELL JE, Macronutrients and mental performance. Nutrition 2000; 16: 1021±34. ELIAS PK, ELIAS MF, DAGOSTINO RB et al. NIDDM and blood pressure as risk factors for poor cognitive performance ± The Framingham Study. Diabetes Care 1997; 20: 1388±1395. ENGLYST KN, ENGLYST HN, HUDSON GJ, COLE TJ, CUMMINGS JH, Rapidly available glucose in foods: an in vitro measurement that reflects the glycemic response. American

ß 2004, Woodhead Publishing Limited

Journal of Clinical Nutrition. 1999; 69 (3): 448±54. EWING FME, DEARY IJ, MCCRIMMON RJ, STRACHAN MWJ, FRIER BM,

Effect of acute hypoglycemia on visual information processing in adults with type 1 diabetes mellitus. Physiol Behav 1998; 64: 653±660. FELIG P, Hypoglycemia during prolonged excercise in normal man. New Engl J Med 1982; 306: 895±910. FELIG P, WAHREN J, Fuel homeostasis in excercise. New Engl J Med 1975; 293: 1078±1084. FERNSTROM JD, WURTMAN RJ, Brain serotonin content: increase following ingestion of carbohydrate diet. Science 1971; 174: 1023±1025. FISCHER K, COLOMBANI PC, LANGHANS W, WENK C, Cognitive performance and its relation to postprandial metabolic changes after different macronutrient ingestion in the morning. British Journal of Nutrition 2001; 85, 393±405. FISCHER K, COLOMBANI PC, LANGHANS W, WENK C, Carbohydrate to protein ratio in food and cognitive performance in the morning. Physiology & Behavior. 2002: 75 (3): 411±23. FOSTER-POWELL K, HOLT SH, BRAND-MILLER JC, International table of glycemic index and glycemic load values: 2002. American Journal of Clinical Nutrition. 2002; 76 (1): 5±56. GOLD AE, DEARY IJ, MACLEOD KM, FRIER BM, Hypoglycaemia-induced cognitive dysfunction in patients with IDDM: effect of hypoglycaemia unawareness. Physiology and Behavior 1995; 58: 501±511. GONDER-FREDERICK L, HALL JL, VOGT J et al. Memory enhancement in elderly humans: effects of glucose ingestion. Physiology and Behavior 1987; 41: 503±504. GUEZENNEC CY, SABATIN P, DUFOREZ F, Oxidation of corn starch, glucose and fructose ingested before exercise. Med Sci Sports Exerc 1989; 21: 45±50. GUEZENNEC C.Y., SABATIN P., DUFOREZ F, KOZIET J, ANTOINE JM, The role of type and structure of complex carbohydrates response to physical exercise. Int. J. Sports Med. Vol. 14, no. 4, 224±231, 1993. HAIER RJ, SIEGEL B, TANG C, ABEL L, BUCHSBAUM MS, Intelligence and changes in regional cerebral metabolic rate following learning. Intelligence 1992; 16: 415-426. HALL JL, GONDER-FREDERICK LA, CHEWNING WW, SILVEIRA J, GOLD PE, Glucose enhancement of performance on memory tests in young and aged humans. Neuropsychologia 1989; 27: 1129±1138. HANDLEY RC, KING N, DYE L, The effects of lunches varying in macronutrient composition and glycaemic index index on cognitive performance. Proceedings of the British Psychological Society, 2003. HASSMAN P, BLOMSTRAND E, EKBLOM B, NEWSHOLME EA, Branched-chain amino acid supplementation during 30±km competitive run: mood and cognitive performance. Nutrition 1994; 10: 405±410. HAWLEY JA, DENNIS SC, LAIDLER BJ, BOSCH A, NOAKES TD, BROUNS F, High rates of exogenous carbohydrate oxidation from starch ingested during prolonged exercise. J. Appl. Physiol 71 (5): 1801±1806, 1991. HAWLEY JA, DENNIS SC, NOAKES TD, Oxidation of carbohydrate ingested during prolonged endurance exercise. Sports Medicine 14: 27±42, 1992. È THER E, Nutritional effects on central seronergic activity: a HUETHER G, SCMIDT S, RU hypothesis on the unconscious self-manipulation of mood by food intake and dietary selection. Nutritional Neuroscience 1998; 1: 3±7. HULTMAN E, Dietary manipulations as an aid to preparation for competition. In: Proceedings of the World conference on Sports medicine, Melbourne 1974: pp. 239±265.

ß 2004, Woodhead Publishing Limited

HULTMAN E,

Liver glycogen in man: effect of different diets.and muscular exercise. In: Pernow B, Saltin B (eds) Muscle metabolism during exercise. New York, Plenum, 1981, pp. 143±152. HULTMAN E, Studies on metabolism of glycogen and active phosphate in man with special reference to exercise and diet. Scan J Clin Lab Invest 1967; 19: suppl 94. JARVIS JK, PEARSALL D, OLINER CM, SCHOELLER DA The effect of food matrix on carbohydrate utilization during moderate exercise. Med. Sci. Sports. Exerc., vol. 24, no. 3, 320±326, 1992. JENKINS DJ, WOLEVER TM, TAYLOR RH, BARKER H, FIELDEN H, BALDWIN JM, BOWLING AC, NEWMAN HC, JENKINS AL, GOFF DV,

Glycemic index of foods: a physiological basis for carbohydrate exchange. American Journal of Clinical Nutrition. 1981; 34 (3): 362±6.

JENKINS DJ, THORNE MJ, CAMELON K, JENKINS A, RAO AV, TAYLOR RH, THOMPSON LU, KALMUSKY J, REICHERT R, FRANCIS T,

Effect of processing on digestibility and the blood glucose response: a study of lentils. American Journal of Clinical Nutrition. 1982; 36 (6): 1093±101. JEUKENDRUP AE, SARIS WHM, BROUNS F, KESTER ADM, A new validated endurance performance test. Med Sci Sports Exerc. 1996; 28(2): 266±270. JEUKENDRUP AE, WAGENMAKERS AJM, STEGEN JH et al. Carbohydrate ingestion can completely suppress endogenous glucose production during exercise. Am. J. Physiol. (1999) 276 (endocrinol metab 39) E672±E683. JONIDES J, SCHUMACHER EH, SMITH EE, LAUBER EK, AWH E, SATOSHI M, KOEPPE RA Verbal working memory load affects regional brain activation as measured by PET. J Cognit Neurosci, 1997; 9, 462±475. KANAREK RB, SWINNEY D, Effects of food snacks on cognitive performance in male college students. Appetite 1990; 14, 15. KAPLAN RJ, GREENWOOD CE, WINOCUR G, WOLEVER TM, Cognitive performance is associated with glucose regulation in healthy elderly persons and can be enhanced with glucose and dietary carbohydrates. American Journal of Clinical Nutrition. 2000; 72 (3): 825±36. KAYE WH, GWIRTSMAN HE, BREWERTON TD, GEORGE DT, WURTMAN RJ, Bingeing behavior and plasma amino acids: a possible involvement of brain serotonin in bulimia nervosa. Psychiatric Research 1998; 23: 31±43. KEIZER HA, KUIPERS H, VAN KRANENBURG G, GEURTEN P, Influence of liquid and solid meals on muscle glycogen resynthesis, plasma fuel hormone response, and maximal physical working capacity. Int J. Sports Med 1987; 8: 99±104. KELLY TH, FOLTIN RW, ROLLS BJ, FISCHMAN MW, Effect of meal macronutrient and energy content on human performance. Appetite 1994; 23: 97±111. KENNEDY DO, SCHOLEY AB, Glucose administration, heart rate and cognitive performance: effects of increasing mental effort. Psychopharmacology 2000; 149: 63±71. KIENS B, RABEN AB, VALEUR AK, RICHTER E, Benefit of dietary simple carbohydrates on the early post exercise glycogen repletion in male athletes. Med Sci Sports Exerc 1990; 22: 588 (abstract). KOROL DL, GOLD PE, Glucose, memory, and ageing. American Journal of Clinical Nutrition 1998; 67: 764S±771S. KUIPERS H, KEIZER HA, BROUNS F, SARIS WHM, Carbohydrate feeding and glycogen synthesis during exercise in man. PfluÈgers Arch. 1987; 410: 652±656. LAMB DR, SNYDER AC, Muscle glycogen loading with a liquid carbohydrate supplement. Int J Sports Nutr 1991; 1: 52±60.

ß 2004, Woodhead Publishing Limited

LANGAN SJ, DEARY IJ, HEPBURN DA, FRIER BM,

Cumulative cognitive impairment following recurrent severe hypoglycaemia in adult patients with insulin-treated diabetes mellitus. Diabetologia 1991; 34, 337±344. LAPP JE, Effects of glycemic alterations and noun imagery on the learning of paired associates. Journal of Learning Disabilities 1981; 14: 35±38. LEIJSSEN DPC, SARIS WHM, JEUKENDRUP AE, WAGENMAKERS AJM, Oxidation of Exogenous [13C]Galactose and [13C]Glucose during Exercise. In J Appl Physiol. 1995; 79, 720±725. LIEBERMAN HR, Nutrition, brain function and cognitive performance. Appetite 2003; 40 (3), 245±254. LIEBERMAN HR, CORKIN S, SPRING BJ, GOWDEN JH, WURTMAN RJ, Mood, performance and pain sensitivity: Changes induced by food constituents. Journal of Psychiatric Research 1983; 17: 135±145. LIEBERMAN HR, SPRING B, GARFIELD GS, The behavioural effects of food constituents: strategies used in studies of amino acids, protein, carbohydrates and caffeine. Nutr. Rev. 1986; 44: S61±69. LLOYD HM, GREEN MW, ROGERS PJ, Mood and cognitive performance effects of isocaloric lunches differing in fat and carbohydrate content. Physiology and Behavior 1994; 56: 51±57. LLOYD HM, ROGERS PJ, HEDDERLEY DI, WALKER AF, Acute effects on mood and cognitive performance of breakfasts differing in fat and carbohydrate content. Appetite 1996; 27: 151±164. LLUCH A, DYE L, BLUNDELL JE, Effects of nutrient manipulations on motivation to eat and performance in high fat and low fat phenotypes. Appetite 2000; 35; 202. LOON LJC VAN, SARIS WHM, KRUIJSHOOP M, WAGENMAKERS AJM, Maximizing post-exercise glycogen synthesis: carbohydrate supplementation and application of aminoacid or protein hydrolysate mixtures. Am J. Clin. Nutr 72, 106±111, 2000. LYONS PM, TRUSWELL AS, Serotonin precursor influenced by type of carbohydrate meal in healthy adults. American Journal of Clinical Nutrition 1988; 47: 433±439. MANNING CA, HALL JL, GOLD PE, Glucose effects on memory and other neuropsychological tests in elderly humans. Psychological Science 1990; 1: 307±311. MARKUS CR, Relations between stress, food and mood: a role for brain serotonin. PhD thesis University of Utrecht, 1999. MARKUS CR, PANHUYSEN G, TUITEN A et al. Does carbohydrate-rich, protein-poor food prevent a deterioration of mood and cognitive performance of stress-prone subjects when subjected to a stressful task. Appetite 1998; 31: 49±65. MARKUS CR, PANHUYSEN G, JONKMAN LM, BACHMAN M, Carbohydrate intake improves cognitive performance of stress-prone individuals under controllable laboratory stress. British Journal of Nutrition 1999; 82: 457±467. MARTIN PY, BENTON D, The influence of a glucose drink on a demanding working memory task. Physiology and behaviour; 1999; 67, 69±74. MASON WL, MCCONNEL G, HARGREAVES M, Carbohydrate ingestion during exercise: liquid vs solid feedings. Med. Sci. Sports. Exerc. Vol. 25, no. 8, 966±969, 1993. MASSICOTTE D, PERONNET F, BRISSON G, Oxidation of a glucose polymer during exercise: comparison with glucose and fructose. J. Appl. Physiol 66 (1): 179±183, 1989. MAUGHAN RJ, Effects of diet composition on the performance of high intensity exercise. In: H Monod (ed.), Nutrition et sport, Masson, Paris. 1990: pp. 201±211. MCCRIMMON RJ, DEARY IJ, HUNTLY BJP, MACLEOD KJ, FRIER BM, Visual information processing during controlled hypoglycaemia in humans. Brain 1996; 119: 1277±1287.

ß 2004, Woodhead Publishing Limited

MCGILVERY RW,

The use of fuels for muscular work. In: Howald H, Poortmans JR (eds) Metabolic adaptation to prolonged physical exercise. Proc. of the second int. symp on biochemistry. Basel, BirkhaÈuser Verlag, 1973, pp. 12±20. MCLEOD KM, GOLD AE, EBMEIER KP et al. The effects of acute hypoglycaemia on relative cerebral blood flow distribution in patients with type 1 (insulin-dependent) diabetes and impaired hypoglycaemia awareness. Metabolism 1996; 45: 974±980. MESSIER C, TSIAKAS M, GAGNON M, DESROCHERS A, AWAD N, Effects of age and glucoregulation on cognitive performance. Neurobiology of Aging, 2003; 24, 985±1003. MONK TH, FOLKARD S, Shiftwork and performance. In: Folkard S, Monk TH, (eds) Hours of Work: Temporal Factors in Work-scheduling (Chichester: Wiley) 1985. MOSORA F, LACROIX M, LUYCKX AS et al., Glucose oxidation in relation to the size of the oral glucose loading dose. Metabolism 1981; 30: 1143±1149. MURRAY R, PAUL GL, SEIFERT JG, EDDY D, HALABY GA, The effects of glucose, fructose and sucrose ingestion during exercise. Med Sci. sports Exerc. Vol. 21, no. 3, 275±282, 1989. NEWSHOLME E, Effects of exercise on aspects of carbohydrate, fat, and amino acid metabolism. In: Bouchard C, Shephard R, Stephens T, et al. (eds) Exercise, Fitness and Health, Human Kinetics, Champaign, Il., 1990. NEWSHOLME EA, LEECH AR Metabolism in exercise. In: Biochemistry for the medical sciences. John Wiley & Sons, Chichester 1983a, pp. 357±381. NEWSHOLME EA, LEECH AR Integration of carbohydrate and lipid metabolism. In: Biochemistry for the medical sciences. John Wiley & Sons, Chichester 1983b, pp. 336±356. NEWSHOLME EA, START C (eds), Regulation of carbohydrate metabolism in liver. In: Regulation in Metabolism. John Wiley & Sons, Chichester 1973, pp. 247±323. OWENS DS, BENTON D, The impact of raising blood glucose on reaction times. Neuropsychobiology, 1994; 30, 106±113. PALLIKARAKIS N, JANDRAIN B, PIRNAY F, et al. Remarkable metabolic availability of oral glucose during long duration exercise in humans. J Appl Physiol 1986; 60: 1035± 1042. PARK CR, Cognitive effects of insulin in the central nervous system. Neuroscience & Biobehavioral Reviews. 2001; 25 (4), 311±323. POLLITT E, MATHEWS R, Breakfast and cognition: an integrative summary. American Journal of Clinical Nutrition 1998; 67: 804S±813S. POLLITT E, CUETO S, JACOBY ER, Fasting and cognition in well- and undernourished schoolchildren: a review of three experimental studies. American Journal of Clinical Nutrition 1998; 67: 779S±784S. RAGOZZINO ME, UNICK KE, GOLD PE, Hippocampal acetylcholine release during memory testing in rats: augmentation by glucose. Proc. Natl. Acad. Sci. USA, 1996; 93, 4693±4698. REHRER NJ, Aspects of Dehydration and Rehydration during Exercise. In: Advances in Nutrition and Top Sport, Brouns F, Saris WHM, Newsholme EA (ed.) Med Sport Sci 1991; 32: 128±146, Karger, Basel. REHRER NJ, VAN KEMENADE MC, MEESTER TA, BROUNS F, SARIS WHM, Gastroinstestinal complaints in relation to dietary intakes in triathletes. Int J Sport Nutr 1992; 2: 48±59. REID M, HAMMERSLEY R, The effects of carbohydrates on arousal. Nutrition Research Reviews 1999; 12: 3±23. RICHTER EA, SONNE B, PLOUGH T et al., Regulation of carbohydrate metabolism in exercise. In: Saltin B (ed.): Biochemistry of exercise 6. Champaign Il, Human Kinetic Publ, 1986: pp. 151±166.

ß 2004, Woodhead Publishing Limited

ROETHLISBERGER FJ,

The elusive phenomena: An autobiographical account of my work in the field of organized behavior at the Harvard Business School. Cambridge, MA: Division of Research, Graduate School of Business Administration. Harvard University Press, 1977. SARIS WHM, ANTOINE JM, BROUNS F, et al. PASSCLAIM-Physical performance and fitness Eur J Nutr 2003, Suppl 1, 150±195. SARIS WHM, VAN ERP BAART MA, BROUNS F et al. Study on food intake and energy expenditure during extreme sustained exercise: the Tour de France. Int J Sports Med 1989; 10, S26±S31 (supplement). SARIS WHM, GOODPASTER BH, JEUKENDRUP AE, BROUNS F, HALLIDAY D, WAGENMAKERS AJM, Exogenous carbohydrate oxidation from different carbohydrate sources during exercise J Appl Physiol. 1993; 75: 2168±2172. È RCK I, BLAAK E, BORNET F, BROUNS F, FRAYN KN, Functional food SARIS WHM, ASP NGL, BJO science and substrate metabolism. Br J Nutr 1998; 80: S47±S75. SAWAYA AL, FUSS PJ, DALLAL GE, TSAY R, MCCROY MA, YOUNG V, ROBERST SB, Meal palatability, substrate oxidation and blood glucose in young and older men. Physiology and Behavior, 2001; 72, 5±12. SCHENK S, DAVIDSON CJ, ZDERIC TW, BYERLEY LO, COYLE EF, Different glycemic indexes of breakfast cereals are not due to glucose entry into blood but to glucose removal by tissue. American Journal of Clinical Nutrition 2003; 78 (4): 742±8. SCHOLEY AB, HARPER S, KENNEDY DO, Cognitive demand and blood glucose. Physiology and Behavior, 2001; 73, 585±592. SHERMAN WM, LAMB DR, Nutrition and Prolonged Exercise. In: Lamb DR, Murray R (eds) Perspectives in Exercise Science and Sports Medicine, Vol. 1, Prolonged Exercise. Benchmark Press, Inc., Idianapolis, Indiana 1988, pp. 213±280. SMITH AP, KENDRICK AM, Meals and performance. In: Smith AP, Jones DM (eds) Handbook of human performance, Vol. 2, Health and performance (London: Academic Press) 1992; pp. 2±23. SMITH AP, MILES C , Effects of lunch on selective and sustained attention. Neuropsychobiology 1986; 16: 117±20. SMITH AP, LEEKHAM S, RALPH A, MCNEILL G, The influence of meal composition on postlunch performance efficiency and mood. Appetite 1988; 10: 195±203. SMITH AP, KENDRICK A, MABEN A, SALMON J, Effects of fat content, weight, and acceptability of the meal on postlunch changes in mood, performance, and cardiovascular function. Physiol Behav 1994a; 55: 417±422. SMITH AP, MABEN A, BROCKMAN P, Effects of evening meals and caffeine on cognitive performance, mood and cardiovascular functioning. Appetite 1994b; 22: 7±65. SNORGAARD O, LASSEN LH, ROSENFALCK AM, BINDER C, Glycaemic thresholds for hypoglycaemic symptoms impairment of cognitive function and release of counter regulatory hormones in subjects with functional hypoglycaemia. Journal of Internal Medicine 1991; 229: 343±350. SPRING B, MALLER O, WURTMAN J, DIGMAN L, GOZOLINI L, Effects of protein and carbohydrate meals on mood and performance: interactions with sex and age. J. Psychiatric Research. 1983; 17: 155±167. SPRING B, CHIODO J, BOWEN DJ , Carbohydrates, tryptophan, and behaviour: a methodological review. Psychological Bulletin 1987; 102: 234±256. STEWART R, LIOLITSA D, Type 2 diabetes mellitus, cognitive impairment and dementia. Diabetic Medicine. 1999; 16 (2): 93±112. STRACHAN MW, DEARY IJ, EWING FM, FRIER BM, Is type II diabetes associated with an increased risk of cognitive dysfunction? A critical review of published studies.

ß 2004, Woodhead Publishing Limited

Diabetes Care 1997; 20: 438±445. The effects of protein or carbohydrate breakfasts on subsequent plasma amino acid levels, satiety and nutrient selection in normal males. Pharmacology, Biochemistry and Behavior 1989; 34: 829±837. THOMAS DE, BROTHERHOOD JR, BRAND JC, Carbohydrate feeding before exercise: Effect of glycemic index. Int J. Sports Med. Vol. 12, no. 2, 180±186, 1991. THOMAS JR, LOCKWOOD PA, SINGH A, DEUSTER PA, Tyrosine improves working memory in a multitasking environment. Pharmacology, Biochemistry and Behavior 1999, 64: 495±500. TUTTLE WW, WILSON M, DAUM K, Effects of altered breakfast habits on physiologic response. Journal of Applied Physiology 1949; 1: 545. TUTTLE WW, DAUM K, MYERS L, MARTIN C, Effects of omitting breakfast on the physiologic response of men. Journal of the American Dietetic Association 1950; 26: 332±335. TUTTLE WW, DAUM K, IMIG CJ, RANDALL S, SCHUMACHER MT, Effect of omitting breakfast on the physiologic response of the aged. Journal of the American Dietetic Association 1952; 28: 117±120. TUTTLE WW, DAUM K, LARSEN R, SALZANO J, ROLOFF L, Effect on schoolboys of omitting breakfast: physiologic responses, attitudes and scholastic attainments. Journal of the American Dietetic Association 1954; 30: 674±677. VANHANEN M, KUUSISTO J, KOIVISTO K et al. Type-2 diabetes and cognitive function in a non-demented population. Acta Neurologica Scandinavia 1999; 100: 97±101. WAGENMAKERS AJM, BECKERS EJ, BROUNS F, KUIPERS H, SOETERS VD, VUSSE G, SARIS WHM, Carbohydrate supplementation, glycogen depletion and amino acid metabolism during exercise. Am J Physiol 260 (endocrinol metab 23): E883±E890, 1991. WAGENMAKERS AJM, BROUNS F, SARIS WHM, HALLIDAY D, Oxidation rates of orally ingested carbohydrates during prolonged exercise in men J Appl Physiol 75 (6): 2774±2780, 1993. WELLS AS, READ NW, Influences of fat, energy and time of day on mood and performance. Physiology Behavior 1996; 59: 1069±1076. WIDOM B, SIMONSON DC, Glycemic control and neurophysiologic function during hypoglycemia in patients with insulin dependent diabetes mellitus. Annals of Internal Medicine 1990; 112: 904±912. WOLEVER TM, BOLOGNESI C, Time of Day influences relative glycaemic effect of foods. Nutrition Research, 1996, 16 (3), 381±384. WOLEVER TM, JENKINS DJ, JENKINS AL, JOSSE RG, The glycemic index: methodology and clinical implications. American Journal of Clinical Nutrition. 1991; 54 (5): 846±54. TEFF KL, YOUNG SN, BLUNDELL JE,

WOLEVER TM, VORSTER HH, BJORCK I, BRAND-MILLER J, BRIGHENTI F, MANN JI, RAMDATH DD, GRANFELDT Y, HOLT S, PERRY TL, VENTER C, XIAOMEI WU, Determination of the glycaemic index of foods: interlaboratory study. European Journal of Clinical Nutrition. 2003; 57 (3): 475±82. YASPELKIS BB, IVY JL, Effect of carbohydrate supplements and water on exercising metabolism in the heat. J. Appl. Physiol. 71: 680±687, 1991. YOUNG SN, Some effects of dietary components (amino acids, carbohydrate, folic acid) on brain serotonin synthesis, mood and behaviour. Canadian Journal of Physiology and Pharmacology 1991, 69, 893±903. ZAWADSKI KM, YASPELKIS BBD, IVY JL, Carbohydrate-protein complex interactions increase the rate of muscle glycogen storage after exercise. J. Appl. Physiol 1992, 72; 1854±1859. ZHAO W, ALKON DL, Role of insulin and insulin receptor in learning and memory. Molecular and Cellular Endocrinology; 2001; 177, issue no. 1±2, 125±134.

ß 2004, Woodhead Publishing Limited

19 Detecting nutritional starch fractions K. Englyst and H. Englyst, Englyst Carbohydrates Ltd., UK

19.1 Introduction 19.1.1 Dietary carbohydrate classification Dietary carbohydrates have a range of physiological fates, determined by their chemical identity and their exposure to food processing. The 1998 WHO/FAO consultation on the role of dietary carbohydrates in human nutrition recommended that dietary carbohydrates should be classified primarily on the basis of chemical characteristics, and then with consideration of physiological properties (FAO/WHO, 1998). In conjunction with human studies, we have developed analytical procedures for the measurement of non-starch polysaccharides (NSP) as a marker of plant cell walls (dietary fibre) (Englyst et al., 1982, 1994, 2000a), resistant short-chain carbohydrates (RSCC; also referred to as non-digestible oligosaccharides (Quigley et al., 1999), resistant starch (RS) as a measure of the starch that escapes digestion in the human small intestine (Englyst and Cummings, 1985, 1986, 1987; Englyst and Kingman, 1990; Englyst et al., 1992; Englyst and Hudson, 1996), rapidly available glucose (RAG) and slowly available glucose (SAG) as glucose rapidly or slowly available from sugar and starch (Englyst et al., 1996, 1999, 2000b, 2003). These measurements have been incorporated into the proposed classification and measurement scheme shown in Table 19.1, which is based primarily on the measurement of chemically identified components and incorporates knowledge of carbohydrate bioavailability and likely physiological properties. The RSCC, NSP, RS and most sugar alcohols escape digestion and absorption in the small intestine, and enter the large intestine, where they can be substrates for fermentation (Cummings and Englyst, 1987). The chemical and

ß 2004, Woodhead Publishing Limited

Table 19.1

Classification of dietary carbohydrates Main components

Free sugars Mono- and disaccharides

Sugar alcohols Mono- and disaccharides Short chain carbohydrates Maltodextrins Resistant short-chain carbohydrates (Nondigestible oligosaccharides)

Polysaccharides Starch Rapidly digestible starch (RDS) Slowly digestible starch (SDS) Resistant starch (RS)

Non-starch polysaccharides (NSP) Plant cell-wall NSP

Other NSP

Comments

(Soluble in 80% ethanol; 2 sugar units) Glucose, fructose, sucrose Glucose, maltose and sucrose rapidly digested. maltose, lactose Fructose and lactose may, in part, escape digestion and absorption in the small intestine. Physiological response depends on identity. Free glucose + glucose from sucrose = FSG. (Soluble in 80% ethanol; 2 sugar units) Sorbitol, inositol, mannitol Poorly absorbed in the small galactitol, maltitol intestine. May reach the large intestine. (Soluble in 80% ethanol; > 2 sugar units) -Glucans Partly hydrolysed starch. Normally included in the measurement of starch. Escape digestion in the small Fructo-oligosaccharides intestine and are fermented to galacto-oligosaccharides, pyrodextrins, polydextrose different extents. Some may stimulate growth of bifidobacteria. Physiological effect largely unknown. (Insoluble in 80% ethanol) -Glucans The most abundant dietary carbohydrates. Rapidly released glucose Rapidly digested in the small intestine. RDS + rapidly released FSG = RAG. Slowly released glucose Slowly digested in the small intestine. SDS + slowly released FSG = SAG. RS1 (physically The three types of RS escape inaccessible) digestion in the small intestine RS2 (resistant granules) and are fermented to different RS3 (retrograded starch) extents. Physiological effect largely unknown. Many different types of Escape digestion in the small polysaccharides intestine and are fermented to different extents. Main constituents: arabinose Encapsulation and slow xylose, mannose, galactose absorption of nutrients. Good glucose, uronic acids marker for naturally high-fibre diets for which health benefits have been shown. Many types of constituents Food additives. Minor components of the human diet. The amounts added to foods are known and regulated.

FSG, free-sugar glucose; RAG, rapidly available glucose; SAG, slowly available glucose.

ß 2004, Woodhead Publishing Limited

physical properties of this diverse group of carbohydrates determine their very different effects on the physiology of the large intestine. The focus of this chapter is the rate of digestion and absorption of dietary carbohydrates in the small intestine with its implications for the GI concept, which ranks foods by the extent they raise blood glucose levels (Jenkins et al., 1981). Low GI diets have been shown to improve metabolic profiles in diabetes (Jenkins et al., 2002) and have been associated with a reduced risk of developing type 2 diabetes and heart disease (Salmeron et al., 1997; Liu et al., 2000). The GI and the RAG and SAG measures are different but complementary approaches to describing carbohydrate quality.

19.1.2 Carbohydrate bioavailability The digestion of starch in the human small intestine is an ongoing process that, for many sources of starch, does not go to completion. The rate and extent of starch digestion is determined by numerous factors including amylose:amylopectin ratio, structure of the food matrix and the extent of starch gelatinisation and retrogradation, which is largely controlled by food processing. The RAG, SAG and starch fraction measures (see Table 19.1) have been developed to describe the likely rate and extent of glucose release from foods in the small intestine. These fractions describe physicochemical characteristics of the foods, which in turn determine the bioavailability and likely physiological fate of the carbohydrate when it is consumed. The method described in this chapter has been developed from a series of in vivo and in vitro studies. Due to the variation in both the rate and extent of starch digestion that is seen both within and between individuals in human studies, in vitro measurements are based on the average of measurements observed in vivo. Several studies have used human ileostomy subjects as a model to investigate the digestive physiology of the small intestine. Carbohydrate analysis of the ileostomy effluent allows determination of the amount of starch and starch digestion products escaping digestion in the small intestine (Englyst and Cummings, 1985, 1986, 1987). The amount of starch recovered in the ileostomy effluent varied between individuals by  20% around the mean. The in vitro methodology has been tuned to yield values for RS that match the mean proportion of starch recovered in ileostomy studies for both single foods and mixed meals (Englyst et al., 1996; Silvester et al., 1995). RS is subdivided into three categories that reflect the reasons why starch escapes digestion and absorption in the human small intestine (physically inaccessible starch (RS1), resistant starch granules (RS2) and retrograded starch (RS3)). The physiological relevance of the division between RAG and SAG has been demonstrated in a series of studies investigating their relation to glycaemic response and glycaemic index values. A study of 39 starchy foods showed that RAG values were correlated strongly to published GI values (Englyst et al., 1996). This relation was further confirmed in a study investigating the determinants of GI and insulinaemic index values of cereal products (Englyst

ß 2004, Woodhead Publishing Limited

Fig. 19.1 Relationship between (A) RAG and (B) SAG and mean glycemic response (measured as the incremental area under the curve; iAUC) for 8 subjects, each consuming 8 meals (four foods with varied proportions of RAG and SAG, each fed as 25 g and 50 g glycemic glucose portion sizes). Modified from Englyst et al., 1999.

et al., 2003). The carbohydrate model including the RAG and SAG fractions described 68% of the variance in GI between products compared to 33% described by the model with only starch and sugar. Another study investigated four foods with differing proportions of RAG and SAG consumed as either 25 g or 50 g carbohydrate portions. This demonstrated that it was RAG, by virtue of its rapid digestion and absorption in the small intestine that was the main determinant of the postprandial rise in blood glucose concentrations (Englyst et al., 1999), and furthermore that SAG exerted its lowering effect on glycaemic response by replacing RAG in the test meal (Fig. 19.1).

19.2

Methods of determining RAG, SAG and RS fractions

19.2.1 Methodological considerations This section describes the in vitro methodology that has been developed to quantify RAG, SAG and the starch fractions. Briefly, the main method quantifies the various carbohydrate fractions through enzymatic hydrolysis and measurement of the glucose released. Samples are analysed `as eaten' and are treated with protease to disrupt any starch-protein interaction. This is followed by incubation with amylolytic enzymes under conditions controlled for temperature, pH, viscosity and rate of mechanical mixing. Subsamples are taken at 20 min and 120 min as measures of the rate and extent of starch digestion. Any remaining starch is dispersed before enzymatic hydrolysis to give a value for total glucose (TG). A value for RS is calculated as the difference between glucose released at 120 min hydrolysis and the TG value. The purpose of the methodology is to measure the food mediated variations in the rate and extent of carbohydrate digestion. The divisions between the RAG, SAG and RS fractions are not based on chemical differences alone. The challenge is to divide the total glucose fraction from starch and sugars into

ß 2004, Woodhead Publishing Limited

subgroups, which reflect the physicochemical characteristics of interest. When dealing with such physicochemical divisions within the total glucose fraction it is essential that the values obtained for any given food can be reproduced from one batch to the next. In this respect the method differs considerably from other analytical methods where the emphasis is to achieve complete recovery of the compound of interest. The physicochemical properties, described by the RAG, SAG and starch fractions, can only be compared between different foods on a like-with-like basis when such a high level of reproducibility is guaranteed. Achieving such a degree of reproducibility requires robust analytical methodology in which the hydrolysis conditions that can influence the rate and extent of glucose release from the food sample are strictly controlled. Three reference samples, selected due to their specific starch digestion profiles (see Section 19.3.1), have been used in the optimisation of the hydrolysis conditions of this procedure. It is not possible to prepare mixtures of hydrolytic enzymes that are identical between every batch, therefore the hydrolytic potential of the preparations is designed to be in excess so as not to represent a rate-limiting step which could mask food mediated variations. The shape of the hydrolysis curve means that it is the early stages of the incubation, where the RAG:SAG division lies (20 min), that display the steepest hydrolysis rates and therefore batch to batch variability in hydrolysis conditions will result in greater errors in RAG and SAG values, than in RS values. The 20 minute division is not intended to correspond directly to the amount of glucose absorbed in vivo by 20 min. Glucose release and absorption in the small intestine does not occur as discrete increments, but is rather a continuous process starting when the food enters the small intestine and ending when all the glucose has either been absorbed or entered the large intestine. Similarly, glucose release during in vitro amylolytic hydrolysis for the determination of RAG and SAG is a continuous process, even though measurements are taken at only two time points. Therefore, the RAG and SAG terms should be viewed as measures of the rate at which glucose is `likely' to become available for absorption in the small intestine. For foods with a well-defined particulate structure, the rate and extent of starch digestion are critically dependent on the way the food sample is divided and the method chosen should reflect the average disruption of food structure achieved by chewing. Comparison of the starch digestibility in food samples that were chewed before analysis and those minced, as described in this method, showed that, while both gave similar mean values, the values for the minced samples had a smaller standard deviation (Englyst et al., 1992). The addition of glass balls to the incubation tubes provides further mechanical disruption of particulate matter during hydrolysis with pancreatic amylase. This allowed the determination of in vitro RS values for polished rice commensurate with the starch recovered in ileostomy studies with similar products (Englyst et al., 1992). To prevent excessive grinding of starch granules by the glass balls, the viscosity of the sample mixture was increased by the addition of guar gum, which keeps the sample material in suspension. The extent

ß 2004, Woodhead Publishing Limited

of mechanical disruption by the glass balls is controlled by the speed of the shaking waterbath, which is calibrated using a reference sample (see Section 19.3.1). Three approaches are used to disperse the starch fraction that resists hydrolysis during the incubation with pancreatic -amylase. Physical disruption of particulate material (RS1) by vigorous vortex-mixing at several stages is aided by the presence of the glass balls. A 30 min boiling step is used to gelatinise starch granules (RS2). Retrograded starch (RS3) is dispersed by 30 min treatment with 2 M KOH, at 0 ëC to avoid the destruction of sugars that can occur at higher temperatures. These treatments in combination have been shown to totally disperse starch in all but the hardest whole grains, which may require milling to obtain a reliable value for total starch (Englyst et al., 1992). During the early development of the method, released glucose was measured by an enzyme linked colourimetric assay. This meant that all the pipetting steps in the procedure necessarily needed to be extremely accurate as no internal standardisation was available to compensate for variations in volume. The nature of the sample, which is viscous due to the guar gum and can contain heterogeneous food particles, can reduce the reproducibility of pipetting. The measurement of sugars by chromatography has the advantage that an internal standard may be added early on in the procedure and allows the simultaneous analysis of other sugar moieties such as fructose. The HPLC version of the procedure is described here.

19.2.2 Reagents Distilled water, or water of equivalent purity should be used throughout the method. All sugars should be dried to constant weight under reduced pressure with phosphorus pentoxide before use. Acetic acid, 1 mol lÿ1 plus 4 ml of 1 mol lÿ1 calcium chloride per litre Amyloglucosidase, EC 3.2.1.3 (Englyst Carbohydrate Services Ltd, Cat. no. 61002) Amyloglucosidase solution, amyloglucosidase diluted 1:4 (v/v) with water Benzoic acid, saturated. Prepare a saturated solution of benzoic acid at room temperature Enzyme mixture. For 18 samples/standards, into each of six centrifuge tubes weigh 3.0 g of pancreatin and suspend in 20 ml of water using a vortexmixer. Add a magnetic stirring bar and mix for 10 min. Centrifuge at 1500 g for 10 min, then remove 15 ml of the cloudy supernatant from each tube and combine (90 ml total). Add 8 ml of amyloglucosidase and 6 ml of invertase. Mix well. Enzyme mixture I should be prepared immediately before use Guar gum powder (Sigma, Cat no. G-4129) Hydrochloric acid, 0.05 mol lÿ1 Internal standard. Dissolve 40 g of arabinose in water, add 200 ml of saturated benzoic acid and make to 1 l with water

ß 2004, Woodhead Publishing Limited

Invertase, EC 3.2.1.26 (Merck, Cat. no. 39020) Methanol, absolute, electrochemical grade Pancreatin (Sigma, 8 USP, Cat. no. P-7545) Pepsin powder, EC 3.4.23.1 (Sigma, Cat. no. P-7000) Pepsin/guar gum solution. For 18 samples/standards, add 1 g of pepsin powder to 200 ml of 0.05 mol lÿ1 hydrochloric acid and mix with a magnetic stirring bar. Just before use, add 1 g of guar gum and mix well. The pepsin/guar gum solution should be prepared immediately before use Potassium hydroxide, 7 mol lÿ1 Reference sample 1 (white wheat flour: Homepride, UK) Reference sample 2 (raw potato starch: Kartoffelmel Centralen, Denmark) Reference sample 3 (corn flakes: Kellogg's, UK) Stock sugar mixture. Dissolve 50 g of glucose and 25 g of fructose in water, add 200 ml of saturated benzoic acid and make to 1 l with water Sodium acetate, 0.5 mol lÿ1. Weigh 68 g of sodium acetate trihydrate, and make to 1 l with water Sodium hydroxide, 50 %, w/v.

19.2.3 Apparatus Balance, accurate to 0.1 mg Centrifuge, capable of exerting 1500 g Centrifuge tubes. Polypropylene (Copolymer) centrifuge tubes of length 11.5 cm and diameter 3 cm (50 ml capacity), with screw caps Glass balls (marbles) of approximately 1.5 cm diameter, sufficient to provide 5 balls per sample HPLC system and carbohydrate separation columns Magnetic stirrer Mincer, hand-driven, with a plate of 0.9 cm diameter holes Test tubes, glass or plastic, of 15 ml and 30 ml capacity, preferably with lids, capable of withstanding low-speed centrifugation Vortex-mixer Waterbath, capable of maintaining 100 ëC Waterbaths, capable of maintaining temperatures in the range 35±70 ëC. This or another similar bath must have a linear shaking capacity of not less than 160 strokes/min and a stroke length of approximately 35 mm. The shaking bath must be fitted with the means to hold all the centrifuge tubes exactly horizontally under the water, with the long axis of each tube exactly parallel with the direction of movement. Each bath should be of sufficient capacity that there is no significant change in temperature when a rack containing all the tubes is placed into it.

ß 2004, Woodhead Publishing Limited

Table 19.2 Dry matter (%)

Examples

Weight (g)

75±100 75±100 55±75 35±55 15±35

Starch Flours, breakfast cereals Bread, cakes Beans, pasta, rice Canned foods, sauces

0.6 0.8 1.0±2.0 1.5±3.0 3.0±4.0

19.2.4 Sample preparation Starch digestibility is greatly influenced by food processing and samples must be analysed as eaten. Foods normally eaten dry are analysed dry, and foods normally eaten hot and wet are cooked and maintained at 70±80 ëC immediately before analysis. Foods with a recognisable structure that would normally require chewing (e.g. pasta, rice, maize) are passed through the mincer. Determine the moisture content of samples as weight loss after overnight incubation at 104 ëC. Sample weights should be chosen to contain 500±600 mg of starch and sugars, which can be estimated from food tables. Examples of suitable sample weights are given in Table 19.2.

19.2.5 Main procedure Sample weight and internal standard Weigh an appropriate amount of sample, to the nearest 0.1 mg, into 50-ml tubes. If the sample is hot, cap the tube and place into a waterbath at 70±80 ëC until ready to analyse. Samples should be analysed at least in duplicate. Add 5 ml of internal standard. Incubation with pepsin Add 10 mL of the freshly prepared pepsin/guar gum solution into each tube, vortex-mix, and place the tubes into a waterbath at 37 ëC for 30 min. Remove the samples from the waterbath and add 5 glass balls and 5 ml of 0.5 mol lÿ1 sodium acetate to each tube. Shake the tubes gently to disperse the contents, and replace them in the waterbath at 37 ëC to equilibrate. Enzymatic hydrolysis of starch Remove one sample tube from the 37 ëC waterbath and add 5 ml of enzyme mixture. Immediately cap the tube and mix the contents gently by inversion before securing the tube horizontally in the 37 ëC shaking waterbath. Start the shaking action of the waterbath; this is time zero for the incubation and the shaking action is not interrupted until all the G120 portions have been collected (see below). Repeat the addition of enzyme mixture for the rest of the sample tubes, at 1 min intervals to aid timing of the procedure, and place them into the shaking waterbath. Remove each tube from the bath at exactly 20 min after

ß 2004, Woodhead Publishing Limited

addition of the enzyme mixture, take 0.2 ml into 4 ml of methanol and vortexmix to stop the hydrolysis; this is the G20 portion to be measured in step 2.7. Return the sample tube to the shaking waterbath immediately after the sample has been taken. After a further 100 min (a total of 120 min incubation), take 0.2 ml from the sample tube into 4 ml of methanol and vortex-mix; this is the G120 portion to be measured in step 2.7. Dispersion of resistant starch Having removed all the tubes from the shaking waterbath, vortex-mix vigorously to break up any large particles, then place the rack of tubes into a boiling waterbath for 30 min. Remove the tubes from the bath, vortex-mix and place the tubes in ice-water until thoroughly chilled. Add 10 ml of 7 mol Lÿ1 KOH, cap the tube, and mix the contents by inversion. Immerse the tubes horizontally in a shaking waterbath containing ice-water for 30 min. Hydrolysis of starch to glucose Remove the tubes singly from the ice-water and take 0.2 ml of the contents into a tube containing 1 ml of 1 mol lÿ1 acetic acid. To these tubes add 40 l of amyloglucosidase solution. Mix and place the tubes into a 70 ëC waterbath for 30 min followed by 10 min in a boiling waterbath. Cool the tubes to room temperature before adding 8 ml of methanol; this is the total glucose (TG) portion to be measured in step 2.7.

19.2.6 Measurement of free-sugar glucose Weigh an appropriate amount of sample, to the nearest 0.1 mg, into 50-mL tubes and add 5 glass balls to each. Add 5 ml of internal standard to each sample tube. Add 20 ml of water, cap, and vortex-mix vigorously in order to begin breaking up large particles. Place the rack of tubes into a boiling waterbath. After 30 min remove the rack of tubes from the boiling waterbath and vortex-mix vigorously to break up any remaining particles of sample. Cool to 37 ëC, then add 0.2 ml of invertase to each tube. Cap the tubes and immerse horizontally in the shaking waterbath at 37 ëC for 30 min. Vortex-mix vigorously before removing 0.2 ml of each sample or standard solution into a test-tube containing 4 ml of methanol; this is the free-sugar glucose (FSG) portion to be measured in step 2.7.

19.2.7 HPLC analysis of sugars Prepare two sugar standards in 50-mL tubes. Standard 1:1 ml of stock sugar mixture, 19 ml of water, 5 ml of internal standard. Standard 2: 10 ml of stock sugar mixture, 10 ml of water, 5 ml of internal standard. Mix well, and take 0.2 ml from each into a tube containing 4 ml of methanol. Before HPLC analysis, centrifuge all the fractions for 5 min at 500 g. The amount to be taken for analysis varies according to sugar content: typically 70 l for the sugar standards and the G20 and G120 portions; 200 l for the TG

ß 2004, Woodhead Publishing Limited

portions; and 70±120 l for the FSG portions. Place the samples into HPLC vials, add 1 ml of deionized water and vortex-mix. A Dionex AS3500 autoinjector may be used for injection of 20 l of the diluted fractions. Sugar separation may be achieved with a Dionex Carbopac PA100 column using a Dionex GP40 gradient pump and a Dionex Carbopac PA10 guard column. A Dionex Aminotrap may be used to prevent pancreatic peptides from reaching the analytical column. The eluents are high-purity water and 200 mol lÿ1 NaOH (16 ml of 50% w/v NaOH solution per litre, with a 20 mol lÿl NaOH solution used for separation. Monosaccharide detection may be achieved with a Dionex ED40 electrochemical detector with the following pulse potentials, E, and durations, t: E1 0.05 V, t1 400 ms; E2 0.75 V, t2 200 ms; E3 0.15 V, t3 400 ms; a response time of 1 s, and the output on the detector set at 300 nA.

19.2.8 Calculations Values for RAG, SAG, RDS, SDS, RS and total starch (TS) are calculated from the measured FSG, G20, G120 and TG values (see below). Values for the starch fractions are expressed as polysaccharides, using a factor of 0.9 to convert the measured glucose value to polysaccharide. Calculations RAG ˆ G20 ˆ FSG + RDS SAG ˆ G120 G20 RDS ˆ 0.9 (G20 FSG) SDS ˆ 0.9 (G120 G20) TS ˆ 0.9 (TG FSG) RS ˆ 0.9 (TG G120)

(1) (2) (3) (4) (5) (6)

19.2.9 Measurement of RS1, RS2 and RS3 There are three forms of RS: RS1 is physically inaccessible starch present in foods having a dense or rigid structure, e.g., boiled rice, pasta, whole-grain bread, maize and legumes; RS2 is resistant starch granules present in raw foods, e.g., bananas, and raw cereal flours and grains, in foods cooked with very little water, e.g., some dry-baked biscuits, and in legumes; RS3 is retrograded starch present in foods that have been cooked and then cooled, e.g., bread, breakfast cereals and cold potatoes. (Note: for a few high-amylose products, e.g., Hylon VII, the boiling step in the procedure will not gelatinise all of the starch. Any starch that remains ungelatinised after this step will therefore be included as RS3.) A value for total RS is obtained by the main procedure. Values for the individual RS1, RS2 and RS3 fractions may be obtained by the parallel analysis of three samples of the test food, which are prepared and analysed in different ways. Sample A is prepared in the form in which the food would normally be eaten, and is then passed through a hand-operated mincer. The analytical sample is

ß 2004, Woodhead Publishing Limited

weighed into 50-ml tubes and treated as described for the main procedure. Only the G120(A) value is required for calculation of the RS1 fraction but it is recommended that G20(A) and TG(A) are measured during the procedure and the values used for calculation of RAG, SAG and total RS. Sample B is homogenised for moist foods or milled for dry foods to disrupt the food matrix before analysis but ball milling should not be used, as this may damage starch granules. The analytical sample is weighed into 50-ml tubes and treated as described for the main procedure. Only the G120(B) value is required for calculation of the RS2 fraction. Sample C is homogenised or milled as described for sample B. The analytical sample is weighed into 50-ml tubes and treated as described for the main procedure up to and including the addition of sodium acetate. At this stage, the tubes are placed into boiling water for 30 min to gelatinise any native starch granules that would normally be gelatinised during cooking. The tubes are cooled to 37 ëC and the remainder of the main procedure followed to obtain the G120(C) value that is required for calculation of the RS2 and RS3 fractions. The analysis is continued to obtain a TG(C) value, which is required for calculation of the RS3 fraction. The step in the boiling waterbath in the main procedure after the G120(C) subsample has been withdrawn may be omitted, instead continuing with the dispersion of retrograded starch by treatment with KOH. Calculations RS1 ˆ 0.9 (G120(B) G120(A)) RS2 ˆ 0.9 (G120(C) G120(B)) RS3 ˆ 0.9 (TG(C) G120(C))

(7) (8) (9)

19.3 Quality control and troubleshooting 19.3.1 Quality control For reproducible results to be obtained within and between laboratories it is essential that stringent quality control is applied. The procedures described here are based on the measurement of the rate and extent of starch digestion in vitro and require standardised enzyme preparations and incubation conditions. It is suggested that reference sample 1 and reference sample 2 are included in every batch of samples analysed. The reference samples have been selected for their specific starch digestion profiles. Reference sample 1: starch in raw wheat flour is present as type A starch granules which are partially disrupted by the milling process. The starch is hydrolysed slowly but almost completely during 120 min of amylolytic hydrolysis. In wheat flour the crystalline structure of the starch granules determines the rate at which starch is hydrolysed into soluble dextrins. Reference 1, with its high SDS content, is used to check the efficiency of starch hydrolysis during the amylolytic incubation. Values for G120 that are too low

ß 2004, Woodhead Publishing Limited

Table 19.3 Target values (g/100g) for the reference samples.

Reference sample 1 Reference sample 2 Reference sample 3

G20

G120

TG

35(Ô1) 4(Ô0.5) 79Ô1)

77(Ô1) 26(Ô1) 81(Ô1)

78(tl) 89(Ô1) 85(Ô1)

(see target values above) may indicate that the activity of the amylolytic enzymes is insufficient. Reference sample 2: starch in raw potato starch is present as largely intact type B starch granules. Only 30% of the starch is hydrolysed during the 120 min amylolytic hydrolysis, in a slow linear fashion. Glucose release from potato starch is limited by the rate at which the surface of the starch granule becomes available to the amylolytic enzymes, which is determined by the extent of mechanical disruption, i.e., by glass balls and the effect of the speed of the shaking water-bath. Reference sample 2 is used to establish the optimum stroke speed of the shaking waterbath during the amylolytic incubation. If the G120 value for the potato starch is too high (see target values below), decrease the stroke speed and vice versa. Reference sample 3: starch in corn flakes has been fully gelatinised, some of which has then retrograded. The majority of the starch is rapidly digested during the first 20 min of amylolytic hydrolysis. The retrograded starch is not hydrolysed during the amylolytic hydrolysis, and requires dispersion by potassium hydroxide before it can be hydrolysed. Reference sample 3, with its high RDS content, is used to check the efficiency of starch and maltose hydrolysis during the first stage of amylolytic incubation. Values for G20 that are too low (see target values above) may indicate that the amyloglucosidase activity is insufficient. Target values (g/ 100 g) for the reference samples are given in Table 19.3. The method yields CV values of 3.7% intra-assay and 6.6% inter-assay for the measurement of RAG, 2.2% and 3.9% for G120 (RAG + SAG) and 1.6% and 2.5% for TG for the three reference samples.

19.3.2 Troubleshooting Hydrolysis steps Accurate timing during subsampling for G20 and G120 measurements is critical. Pitted glass balls may increase the mechanical disruption of the samples, and should be replaced. Failure to ensure that tubes are held exactly horizontal in the shaking waterbath and exactly parallel with the direction of movement may result in differences in mechanical disruption of samples. Variation in reference sample values between batches Use the reference samples to check enzyme activity levels. Consistently low results may indicate incorrect pH during hydrolysis; check the buffer system.

ß 2004, Woodhead Publishing Limited

Variation in G120 values between batches for reference sample 2 may be due to changes in the shaking speed or the temperature of the waterbath. Variation in replicate analysis The sample heterogeneity of many foods analysed `as eaten' can make it difficult to take fully representative subsamples, leading to greater variation in results than for homogenised or milled samples. It is recommended that an increased number of replicates are analysed for such problem samples. Spurious peaks on the chromatogram Peaks that significantly interfere with the integration of sugar peaks may be due to amino acids/peptides in the sample that have eluted onto the separator column from the guard column intended for their retention. Clean the guard column or alter the timing of the column-switching mechanism.

19.4 Carbohydrate bioavailability data for selected foods Table 19.4 shows the carbohydrate digestibility profiles for a number of starchy foods. The fructose, RAG and SAG values are expressed in g/100 g of food as eaten, and as a percentage of glycemic carbohydrates. Many breakfast cereals and bakery products contain fully gelatinised starch and/or are without a dense food matrix, which makes the carbohydrates easily dispersed and hydrolysed, and consequently have high RAG contents. Similarly the sugar component of these foods is usually rapidly digested, so although the higher fructose contents reduce the RAG, the beneficial SAG component will be lower too. Those products which contain a dense food matrix, such as whole kernel cereal grains and pasta, are dispersed only slowly and will have a high SAG content. Similarly, some types of biscuits produced under low moisture conditions will have a high SAG content due to the slow digestion of the ungelatinised starch present in them. Information on the SAG composition of products is useful in identifying those low GI products with a high content of slowly digested and absorbed carbohydrates.

19.5 Conclusion and future trends This chapter has described the development of the RAG, SAG and RS fractions and the associated methodology. The discussion describes the basis of this approach to carbohydrate bioavailability (Section 19.5.1), explores the relevance of this approach in relation to the glycemic index concept (Section 19.5.2) and finally describes their contribution to the field of carbohydrate quality (Section 19.5.3).

ß 2004, Woodhead Publishing Limited

Table 19.4 Rapidly available glucose (RAG) and slowly available glucose (SAG) in some plant foods Glycemic carbohydrates g/l00g food as eaten

Glycemic carbohydrates composition

Fructosey

RAG

SAG

3.8 17.1 5.0 19.6

81.7 38.6 85.7 68.1

2.5 4.7 1.5 1.7

4.3 28.3 5.4 21.9

92.8 63.8 92.9 76.2

2.8 7.8 1.6 1.9

Whole-grain cereals Long-grain rice Parboiled rice Bulgar wheat Pearl barley

0.0 0.0 0.2 0.2

21.0 19.4 11.5 8.5

8.3 16.3 7.5 9.3

0.0 0.0 1.2 0.9

71.8 54.4 59.7 47.1

28.2 45.6 39.0 52.0

Bakery products White bread, wheat Wholemeal bread, Rye bread Sponge cake Blueberry muffin

0.3 0.4 1.7 16.4 12.2

43.5 35.5 27.5 41.5 33.6

3.3 1.5 5.1 1.0 0.1

0.6 1.1 4.9 27.9 26.6

92.3 95.0 80.2 70.4 73.2

7.1 3.9 15.0 1.7 0.2

Pasta Wholemeal spaghetti Macaroni Egg noodles

0.3 0.3 0.2

18.1 23.3 21.1

9.4 10.5 11.7

0.9 0.7 0.6

65.2 68.5 64.0

33.8 30.8 35.4

Biscuits and crackers Cream crackers Water biscuit Shortbread Digestive

0.4 0.9 8.2 4.8

67.8 72.7 31.1 44.2

3.5 5.1 21.9 11.4

0.6 1.1 13.5 8.0

94.6 92.4 50.8 73.1

4.8 6.5 35.7 18.9

Breakfast cereals Corn flakes Luxury muesli Rice Krispies Chocolate Rice Krispies

Fructose (%) RAG (%)

SAG (%)

y

Includes fructose from sucrose.

19.5.1 Basis for carbohydrate bioavailability approach Starch has been incorrectly perceived as being slowly but completely digested in the small intestine. This has proved to be an over-simplified view with the reality being that, depending on biological origin and food processing, starch exhibits varied gastrointestinal fates. This ranges from rapid digestion and absorption comparable to that of sugars, to resistance to the amylolytic enzymes of the small intestine. There was a need to characterise these variations in carbohydrate digestibility in order to allow their nutritional value to be assessed.

ß 2004, Woodhead Publishing Limited

The in vitro bioavailability measures RAG, SAG, RS have been developed to meet this demand, providing information on the rate and extent of glucose release from starch and free sugar glucose (including that derived from sucrose; see Table 19.1) during controlled in vitro enzymatic hydrolysis. In addition, values for fructose, the other principal glycemic carbohydrate from plants, can be determined by the HPLC method described. Lactose is not of plant origin and therefore its glucose portion is not included in the RAG, SAG, RS measures, rather separate values for lactose should be included alongside the other carbohydrate fractions where appropriate. The profile of RAG, SAG, RS values are determined by the physicochemical characteristics of the food (i.e. structural properties of the food matrix, presence of intact plant cell walls, type and integrity of starch granules). It is these physicochemical characteristics of foods that determine the gastrointestinal fate or bioavailability of the carbohydrate component. Inevitably, the rate and extent of carbohydrate digestion of a food or meal will be subject to the effects of biological variation associated with physiological differences between individuals and within the same individual in different circumstances. This variance does not diminish the value of the RAG, SAG, RS values as measures of carbohydrate bioavailability, as what is required and provided is a consistent means of characterising these specific properties of foods, allowing a carbohydrate quality comparison between different products.

19.5.2 Relation between the in vitro carbohydrate bioavailabilty measures and glycemic index The GI is an in vivo measure that takes into account the complex interaction of food factors that determine the glycemic response. In addition to the rate of carbohydrate digestion, the GI can be influenced by food mediated effects on both gastrointestinal events and postabsorptive metabolism. Gastric emptying is affected by food particle size (Thomsen et al., 1994), and fat content (Gannon et al., 1993), as well as by viscous fibre, which also limits enzymatic hydrolysis in the small intestine by restricting access to the food bolus (Jenkins et al., 1978). Postabsorptive factors which can influence GI include the identity of the sugar moieties, which are metabolised differently (Lee and Wolever, 1998) and the insulinotropic effect of protein which can increase the clearance rate of circulating glucose (van Loon et al., 2000). This emphasises the fact that GI values do not represent a direct measure of carbohydrate absorption from the small intestine. Rather, the GI values are determined by the combined effect of all the properties of a food that influence the rate of influx and removal of glucose from the circulation. The hypothesis is that as the RAG fraction is rapidly released and absorbed it will be a major determinant of the glycemic response. In contrast, the SAG fraction is released and absorbed slowly and is not expected to contribute to the glycemic response above that which can already be accounted for by the RAG fraction. The physiological relevance of the RAG and SAG fractions has been

ß 2004, Woodhead Publishing Limited

demonstrated in the series of studies described in the introduction (Englyst and Hudson, 1996; Englyst et al., 1999, 2003). The RS fraction should not elicit a glycemic response as it is not absorbed in the small intestine. This approach to characterising carbohydrate quality focuses solely on the in vitro rate of carbohydrate release from foods and the demonstrated relation with glycemic response. It does not describe the glycemic impact of other dietary components that exert their effect via gastric emptying or incretin hormones. This is very different from the GI measurement, which incorporates all the food mediated factors, such as protein and fat, that can influence glycemic response. This ability to encompass the overall glycemic effect of all food factors is the power of the GI, but the negative aspect of this is that it is not possible to identify which food factors are responsible for the GI of any specific food. The utility of the GI in mixed meal situations has been an area of contention. One concern that has been raised is that the overall fat and protein contents of mixed meals will influence the glycemic response to the meal in a way that is not adequately reflected by the combination of individual food GI values, especially when the meal includes added fat or protein components. In addition, the glycemic response to a meal is influenced by other factors such as the amount of carbohydrate, meal volume, previous meals and the recent physical activity of the subject. This may reduce the value of the GI in the prediction of individual glycemic responses, but does not diminish the real value of the GI, which is the promotion of low GI diets which have been shown to improve glycemic control and other metabolic parameters over longer periods (Jenkins et al., 2002). The in vitro bioavailability values of RAG and SAG for individual food items can be added to provide a definitive description of the carbohydrate component of the meal or diet. It is this unambiguous description of the carbohydrate release characteristics that provides the power of the RAG and SAG measures.

19.5.3 Contribution to carbohydrate quality concept There is an urgent need to readdress the public perception of carbohydrates, especially with the recent popularity of low-carbohydrate diets. In part, this situation has arisen with the widespread acceptance that the high glycemic and insulinemic responses associated with many modern starch and sugar containing foods is not conducive to optimal health. Instead of the misinterpretation of this as support for a general reduction in carbohydrate intake, the challenge is to encourage consumption of slow release carbohydrates for which health benefits have been demonstrated. What is required for this message to succeed is a greater understanding of the concept of carbohydrate quality and how this can assist in promoting increased intakes of appropriately processed plant based food products. The GI and the in vitro bioavailability measures have the same overall objective of improving health by providing greater awareness of the concept of carbohydrate quality. The two measures differ in their approach to characterising dietary carbohydrates, and the values they provide are not interchangeable, although they are complementary.

ß 2004, Woodhead Publishing Limited

The in vitro RAG, SAG and RS measures are valuable in distinguishing between different types of low GI diets. The majority of evidence in support of health benefits of low GI diets is associated with foods containing slowly released carbohydrate. Less clear are the health implications of other types of low GI diets with high contents of fructose, lactose, fat or protein. The in vitro bioavailability measures described here provide the means with which to characterise different types of low GI diets and investigate whether they are equally beneficial. This may be especially applicable to epidemiological assessments where a single diet GI value is obtained, which is difficult to interpret without additional information. Similarly, the in vitro bioavailability characterisation of carbohydrates should help distinguish the relative contributions of the quantity and quality of carbohydrates to dietary glycemic load scores. The combination of the GI together with the, RAG and SAG measures therefore provides a powerful tool, which could be utilised in epidemiological studies relating dietary carbohydrates and health, and in studies of postprandial metabolism to provide insight into the mechanisms by which carbohydrates influence longer-term metabolic risk factors. For the purposes of food selection, the GI does not provide information on either the carbohydrate content or other nutritional attributes of the food. The in vitro bioavailability approach provides information on carbohydrate quality, whilst remaining in a format (g/100 g) that can easily be compared with other nutritional attributes of a food. For example the SAG fraction, either in g/100 g or as a percentage of carbohydrate, directly identify foods and diets rich in slow release carbohydrate. The classification scheme (Table 19.1) allows the provision of values for food tables and databases, and the direct calculation of total carbohydrate as the sum of its identified components, as recommended by FAO/WHO.

19.6 Acknowledgement This work was in part supported by the UK Food Standards Agency.

19.7 References (1987) `Fermentation in the Human Large Intestine and the Available Substrates', Am. J. Clin. Nutr., 45, 1243±1255. ENGLYST H.N., CUMMINGS J.H. (1985) `Digestion of the Polysaccharides of Some Cereal Foods in the Human Small Intestine', Am. J. Clin. Nutr., 42, 778±787. ENGLYST H.N., CUMMINGS J.H. (1986) `Digestion of the Carbohydrates of Banana (Musa paradisiaca sapientum) in the Human Small Intestine', Am. J. Clin. Nutr.,44, 42± 50. ENGLYST H.N., CUMMINGS J.H. (1987) `Digestion of the Polysaccharides of Potato in the Small Intestine of Man', Am. J. Clin. Nutr., 45, 423±431. CUMMINGS J.H., ENGLYST H.N.

ß 2004, Woodhead Publishing Limited

(1996) `The Classification and Measurement of Dietary Carbohydrates', Food Chem., 57, 15±21. ENGLYST H.N., KINGMAN S.M. (1990) `Dietary Fiber and Resistant Starch. A Nutritional Classification of Plant Polysaccharides', in Dietary Fiber, eds D. Kritchevsky, C. Bonfield, J.W. Anderson, Plenum Press, New York, 49±65. ENGLYST H.N., WIGGINS H.S., CUMMINGS J.H. (1982) `Determination of the Non-starch Polysaccharides in Plant Foods by GasLiquid Chromatography of Constituent Sugars as Alditol Acetates', Analyst, 107, 307±318. ENGLYST H.N., KINGMAN S.M., CUMMINGS J.H. (1992) `Classification and measurement of nutritionally important starch fractions', Eur. J. Clin. Nutr., 46, S33±S50. ENGLYST H.N., QUIGLEY M.E., HUDSON G.J. (1994) `Determination of Dietary Fiber as Nonstarch Polysaccharides with GasLiquid Chromatographic, High-performance Liquid Chromatographic or Spectrophotometric Measurement of Constituent Sugars', Analyst, 119, 1497±1509. ENGLYST H.N., VEENSTRA J., HUDSON G.J. (1996) `Measurement of rapidly available glucose (RAG) in plant foods: a potential in vitro predictor of the glycaemic response', Brit. J. Nutr. 75, 327±337. ENGLYST K.N., ENGLYST H.N., HUDSON G.J., COLE T.J., CUMMINGS J.H. (1999) `Rapidly Available Glucose in Foods. An in vitro measurement that Reflects the Glycemic Response', AM. J. CLIN. NUTR., 69, 448±454. ENGLYST H.N., QUIGLEY M.E., HUDSON G.J. (2000a) `Dietary Fiber Analysis as Non-starch Polysaccharides' in: Meyers RA, (ed.) Encyclopaedia of Analytical Chemistry, John Wiley & Sons, Chichester 3912±3929. ENGLYST K.N., HUDSON G.J., ENGLYST H.N. (2000b) `Starch analysis in food', in: Meyers R.A., (ed.) Encyclopaedia of Analytical Chemistry, John Wiley & Sons, Chichester 4246±4262. ENGLYST K.N., VINOY, S., ENGLYST H.N., LANG V. (2003) `Glycaemic index of cereal products explained by their content of rapidly and slowly available glucose', Brit. J. Nutr., 89, 329±339. FAO/WHO (1998) `Carbohydrates in Human Nutrition. Report of a Joint FAO/WHO Expert Consultation, Rome, 1418 April 1997'. FAO Food and Nutrition Paper No. 66, FAO, Rome. GANNON M.C., NUTTALL F.Q., WESTPAL S.A., SEAQUIST E.R. (1993) `The effect of fat and carbohydrate on plasma glucose, insulin C-peptide and triglycerides in normal male subjects', Journal of American College of Nutrition 12, 36±41. ENGLYST H.N., HUDSON G.J.

JENKINS D.J.A., WOLEVER T.M.S., LEEDS A.R., GASSULE M.A., DILAWARI J.B., GOFF D.V., METZ G.L., ALBERTI K.G.M.M. (1978) `Dietary fibres, fibre analogues and glucose tolerance: importance of viscosity', British Medical Journal 1, 1392±1394. JENKINS D.J.A., WOLEVER T.M.S., TAYLOR R.H., BARKER H., FIELEN H., BALDWIN J.M., BOWLING A.C., NEWMAN H.C., JENKINS A.L., GOFF D.V. (1981) `Glycemic Index of Foods: A Physiological Basis for Carbohydrate Exchange', Am. J. Clin. Nutr., 34, 362±366. JENKINS D.J.A., KENDALL C.W.C., AUGUSTIN L.S.A., FRANCESCHI S., HAMIDI M., MARCHIE A.,

LEE

JENKINS A.L., AXELSEN M. (2002) `Glycemic index: overview of implications in health and disease', Am. J. Clin. Nutr., 76, 266S±273S. B.M., WOLEVER T.M.S. (1998) `Effect of glucose, sucrose and fructose on plasma glucose and insulin responses in normal humans: comparison with white bread', European Journal of Clinical Nutrition 52, 924±928.

LIU S., WILLETT W.C., STAMPFER M.J., HU F.B., FRANZ M., SAMPSON L., HENNEKENS C., MANSON J.E.

(2000) `A prospective study of dietary glycemic load, carbohydrate intake, and

ß 2004, Woodhead Publishing Limited

Detecting nutritional starch fractions

559

risk of coronary heart disease in US women', Am. J. Clin. Nutr., 71, 1455±1461. (1999) `Determination of Resistant Short-chain Carbohydrates (Non-digestible Oligosaccharides) using GasLiquid Chromatography', Food Chem., 65(3), 381±390. SALMERON J., MANSON J., STAMPFER M., COLDITZ G., WING A., WILLETT W. (1997) `Dietary fiber, glycemic load, and risk of non insulin dependent diabetes mellitus in women', JAMA 277 472±477. SILVESTER K.R., ENGLYST H.N., CUMMINGS J.H. (1995) `Ileal recovery of starch from whole diets containing resistant starch measured in vitro and fermentation of ileal effluent', Am. J. Clin. Nutr. 62, 403±411.

QUIGLEY M.E., HUDSON G.J., ENGLYST H.N.

THOMSEN C., RASMUSSEN O.W., CHRIASTIANSEN C., ANDREASEN F., POULSEN P.L., HERMANSEN

VAN

K. (1994) `The glycaemic index of spaghetti and gastric emptying in non-insulindependent diabetic patients', European Journal of Clinical Nutrition 48, 776±780. LOON L.J.C., SARIS W.H.M., VERHAGEN H., WAGEMAKERS A.J.M. (2000) `Plasma insulin responses of different amino acid or protein mixtures with carbohydrate', Am. J. Clin. Nutr. 72, 96±105.

ß 2004, Woodhead Publishing Limited

20 Resistant starch M. Champ, INRA/CRNH, France

20.1

Introduction

Resistant starch (RS) has been defined as the sum of starch and products of starch degradation not absorbed in the small intestine of healthy individuals (Asp, 1992). RS is now classified in most definitions as dietary fibre (Champ et al., 2003a). Indeed, being mostly a polysaccharide, it is fermented in the large intestine as are most non-starch polysaccharides. Different types of RS have been identified (Table 20.1); the classification is based on origin and physical characteristics of the starches usually considered `as eaten' as most technological treatments, including reheating, can affect RS content in foods. RS type 1 (RS1) is found in starchy foods which are not fractionated and refined and mostly in pulses and some cereals (Table 20.2). Pulses are characterised by thick cell walls which might not be disrupted during the preparation of the meal and in the stomach. Cereal grains when not finely ground can also appear at the end of the small intestine. RS2 are native resistant starch granules, mostly Btype (X-ray pattern) starches. Until recently, such starches were only ingested as unripe banana (Table 20.2), most of our starchy foods being thermally treated. However, the interest for resistant starches has made available a number of high amylose starches (often corn starches, such as EurylonÕ, Novelose240Õ, HylonÕVII, Hi-maizeTM) which are also B-type starches. This second type of RS2 is characterised by a high resistance to gelatinisation as temperatures above 120 ëC are needed to gelatinise these high-amylose starches. Native potato starch is also a source of RS2 but is easily gelatinised in the presence of water at around 60 ëC. RS3 are retrograded starches. Retrogradation occurs when starches are gelatinised then cooled and/or kept at room temperature, 4 ëC (fridge) or ÿ20 ëC (freezer) from a few hours to

ß 2004, Woodhead Publishing Limited

Table 20.1

Classification of resistant starches

Type

Main characteristic

Examples of RS or RS sources

RS 1 RS 2 RS 3

Physically inaccessible starch Resistant starch granules Retrograded starch

RS 4

Chemically modified starch

Grain legumes (beans, lentils) Unripe banana, raw potatoes Cooked and cooled potatoes, stale bread, Novelose 330Õ Cross-bonded starches (with high degree of reticulation; >> to common chemically modified starches)

several days or months. Retrogradation is a reorganisation of linear chains of starch after de-crystallisation due to gelatinisation. It needs a minimum of water and first affects amylose then linear fractions of amylopectin. Retrogradation may affect most starchy foods which are cooked and not immediately dried or eaten. The optimal temperature for retrogradation seems to be around 4 ëC. As a consequence, a classical example of RS3, first mentioned by Englyst and Cummings (1987), is the potatoes cooked and cooled (Table 20.2). Several cycles of reheating and cooling further increase the amount of RS3. RS3 are now commercially available to be introduced in foods such as breads, biscuits or dairy products. Most of these commercial RS3 are derived from high-amylose corn starches (NoveloseÕ330) but a new generation of RS3 (CPActistar, Neo-amylose) which Table 20.2

Ileal digestibilities of starches

Source of starch

Ileal digestibility Value (nb) Method(s)

References

White bread

87.0, 97.5 (2)

IM

Wholemeal bread Oat flakes Corn flakes Freshly cooked potatoes Cooked and cooled potatoes Lentils Beans Banana (ripe) Banana (unripe)

89 (1) 97.8 (1) 95.0 (1) 96.7 (1)

IM IM IM IM

Wolever et al. (1986); Englyst and Cummings Wolever et al. (1986) Englyst and Cummings Englyst and Cummings Englyst and Cummings

HACS, retrograded CPActistar

49.1 (1) 40.7 (1)

88.1 (1) IM 79 (1) IM 73.5 (1) IH 44.3 (1) IM 16.3, 16.0 (2) IM, IH IH IM

(1985) (1985) (1985) (1985)

Englyst and Cummings (1985) Wolever et al. (1986) Noah et al. (1998) Englyst and Cummings (1986) Englyst and Cummings (1986), Faisant et al. (1995) Champ et al. (1998) Langkilde et al.*

IM: Ileostomy model, IH: intubation of healthy subjects, HACS: High amylose corn starch, nb: number of studies. * Langkilde A.M., Andesson H., Brouns F. Personal communication.

ß 2004, Woodhead Publishing Limited

has been introduced does not have some of the disadvantages, in some of its uses, of the most `classical' commercial RS3. Indeed CPActistar, for instance, is obtained by retrogradation of tapioca maltodextrins, themselves derived from tapioca starch by enzymatic amylolysis. As a consequence, it is much more soluble than `classical RS3'; it does not have a `lumpy' texture and can be introduced in drinks. Neo-amylose is described as a linear starch (alpha 1-4 amylose) which is also classified as an RS3. Finally, a new category of RS, RS4 has been introduced lately. RS4 are chemically modified starches with a number of modifications (starch esters, ethers and cross-bonded) much higher than in the usual chemically modified starches authorized in the EC and used in many processed foods such as soups or baby foods. Until now, this type of RS4 has not been authorized in Europe though it is in Japan. The first papers on RS were published in the mid-eighties (Englyst and Cummings, 1985) though starches with high levels of amylose were studied earlier for their functional properties, on cholesterol and bile acid metabolism, for example (Sacquet et al., 1983). Since that period, much work has been done both in private industries (mostly to develop sources of starch high in RS) and in public research institutes and Universities to explain the resistance of starch and mainly to evaluate the potential of RS in health. A concerted European action (EURESTA FLAIR CONCERTED ACTION N11 (COST 911)) has greatly stimulated research on RS in Europe and has certainly contributed to its use in the food industry.

20.2

Effects of resistant starch on the digestive system

Being a neutral polymer and mostly fermented in the colon, the main properties of RS are linked to its fermentation rate and profile. Some studies, however, demonstrated some effects on the absorption of nutrients.

20.2.1 Fate of RS in the digestive tract By definition, RS is the fraction of starch not digested in the small intestine. Thus RS is only digested (and fermented) in the colon. Part of the RS can even escape colonic fermentation and be excreted in the faeces. None of the starch of the starchy foods and ingredients is constituted of 100% RS (Table 20.2). As a consequence, the starch fraction of RS-containing foods as well as its environment is usually drastically modified during digestion in the upper part of the gastrointestinal tract. It is the reason why in vitro quantification of RS, evaluation of its fermentability or physico-chemical properties have to take into account these digestive events. This pre-digestion in the mouth, the stomach and the small intestine involves enzymatic hydrolysis but also mechanical disruption (chewing then gastric mixing) and acidic treatment (chlorhydric acid secretion in the stomach). RS can be native whole starch granules (for instance in unripe banana) or fractions of starch granules which have mostly been digested by -amylases (as in

ß 2004, Woodhead Publishing Limited

Table 20.3 Pattern of short chain fatty acids production from various substrates (from Kritchevsky, 1995) Percentages of total SCFA Substrates Resistant starch Starch Oat bran Wheat bran Cellulose Guar gum Ispaghula Pectin

Acetate 41 50 57 57 61 59 56 75

Propionate 21 22 21 15 20 26 26 14

Butyrate 38 29 23 19 19 11 10 9

most RS3 sources). In both cases, the starch granules are entrapped in a matrix which can mostly be constituted of proteins, lipids, sugars and/or fibres (cell walls or pure added fibres). Proteins, lipids and sugars are in most cases extensively digested before the colon whereas fibres can undergo only minor modifications. These modifications are mostly disruptions of the macrostructure of the fibre when intense chewing has to take place due to a compact structure of the food and large particles or minor chemical changes such as demethylation of pectins. In the colon, the huge number of bacteria colonising this part of the intestine is able to digest then to metabolise most organic polymers due to an impressive number of species and enzyme activities. RS are thus first digested by bacterial amylases (i.e. -amylases, glucoamylase, isomaltase) then glucose is metabolised into organic acids (mainly short chain fatty acids (SCFA), lactic acid) and gases (CO2, H2, CH4). One of the interesting characteristics of RS is its pattern of fermentation and mainly the profile of SCFA which is produced as well as its relatively slow rate of fermentation. Indeed butyrate production, one of the three main SCFA produced during colonic fermentation, is higher than for most dietary fibres (Table 20.3). This SCFA is known as the main nutrient of the colonocyte and a lack of butyrate would increase the risk for some colonic diseases such as colon cancer. The rate of fermentation might be important to predict the site of production of the SCFAs and to explain the relatively low risk of flatus after ingestion of significant amounts of RS. Indeed, it is probable that gases are absorbed from the colon at a sufficient rate to avoid gas accumulation in the lumen which is responsible for the symptoms of flatulence.

20.2.2 Effects on stool weight and bowel transit time As mentioned above, RS is mostly fermented in the large intestine. It then contributes to bacterial proliferation and can be expected to increase bacterial mass. This is indeed observed in some of the studies that have been performed on healthy subjects (Phillips et al., 1995; Cummings et al., 1996; Heijnen et al.,

ß 2004, Woodhead Publishing Limited

1998; Hylla et al., 1998) but a minimal intake of 17 g (more in some of the studies) would be needed to expect a significant effect on stool weight. According to Cummings et al. (1996), it is likely that substrate interaction occurs in the colon and that RS may well be fermented in preference to NSP. As a consequence, a small part of the laxative properties of RS may be ascribed to additional NSP in faeces holding water. No significant effects have been described after RS consumption on bowel transit time (Flourie et al., 1986; Shetty and Kurpach, 1986; Tomlin and Read, 1990; Van Munster et al., 1994).

20.2.3 Effects on the absorption of nutrients If it is obvious that the substitution of part of the digestible starch in a food by RS will decrease the appearance of glucose in the blood, an effect of RS fraction on the absorption of glucose produced by the digestion of the digestible starch fraction has not been demonstrated (Ranganathan et al., 1994). No evidence of an influence of RS on amino-acids and vitamins has been described in the literature. On the contrary, it is suggested that some RS (RS2 but not RS3 according to Schulz et al. (1993)) would be able to improve calcium and magnesium absorption by an enhancing of mineral solubility in the caecum and/ or large intestine due to acidification of the content induced by RS fermentation (Younes et al., 1993).

20.2.4 Interest of RS in prevention and treatment of colonic pathologies Due to its relatively high production of butyrate during its colonic fermentation, RS has been used in a number of trials to evaluate its potential in the prevention or treatment of colonic pathologies such as colon cancer and inflammatory diseases. For obvious reasons, most of these studies have first been conducted on tumour cell lines (in vitro) and animal models of cancer or inflammation. As far as cancer is concerned, the main animal model which is used is rats with chemically induced cancer (dimethylhydrazine (DMH) or azoxymethane (AOM)). With this model, the positive effect of RS3 (retrograded high amylose corn starch) has been observed (fewer aberrant crypt foci) in rats fed a diet containing 5.6% RS in the food (Perrin et al., 2001). However the same group found no beneficial effect (same number of tumours in the control group) of the same RS in Min mice which are animals with a mutation on the APC gene on chromosome 5q which is also found in patients with familial adenomatous polyposis (FAP) (Pierre et al., 1997). The difference between both models could be explained by an effect of RS only in early phases of the development of a cancer whereas it may even be deleterious when the procedure of cancerisation has started. It is difficult to compare the other studies which have different animal models, doses of RS, duration and RS type (Table 20.4). The effects of RS are negative in one of the studies (with Apc mice), positive in two of them (both on AOM treated rats) and finally show no effect on three of them (Min mice, AOM and DMH treated rats).

ß 2004, Woodhead Publishing Limited

Table 20.4

RS and colon cancer ± some data obtained with animal models

Animal model

RS nature

Results

Reference

RS2: more intestinal tumours RS2+aspirin: no effect (compared to chow or western-style diets) RS3 had no effect on the number of tumours (whereas FOS reduced it) RS2 did not enhance carcinogeninduced apoptosis (whereas wheat bran did) RS3 (and FOS) reduced the number of aberrant crypt foci RS2 reduced the number of aberrant crypt foci (lower daily caloric intake)

Williamson et al. (1999)

g/100g diet (duration)

Apc1638N mice

RS2: Mixture 1:1 of raw potato starch 12.9 (1±5 and Hylon VII (Ôaspirin) months)

C57BL/6J-Min/+ mice (Min mice) Azoxymethane-treated rats

RS3: retrograded high amylose corn starch RS2: Hi-maize

5.6 (42 days) 4 weeks

Azoxymethane-treated rats Azoxymethane-treated rats

RS3: retrograded high amylose corn starch RS2: potato starch

1,2-dimethylhydrazinetreated rats

RS2: potato starch

5.6 (44 days) Approx. 53 (67% potato starch) (16 weeks) 3 and 10

Pierre et al. (1997) Le Leu et al. (2002) Perrin et al. (2001) Thorup et al. (1995)

RS2: no effect on total cancer volume Sakamoto et al. (1996) (whereas cellulose did)

Table 20.5 RS and colon cancer ± some data obtained on healthy subjects RS

Results

nature

g/day (duration)

`Chinese' diet `Australian' diet

2 g RS/MJ 0.8g RS/MJ (3 weeks)

RS2 : Hylon VII

RS2: Hylon VII RS3: retrograded Hylon VII

Fecal pH: CA Fecal ammonia: C>A Fecal phenols: C>A 55.2 (high RS) Fecal weight: HRS>LRS 7.7 (low RS) Fecal conc. of SCFA: HRS=LRS (4 weeks) Bacterial -glucosidase activity: HRS