Advances in Food and Nutrition Research Volume 45

ADVANCES IN

FOOD AND NUTRITION RESEARCH VOLUME 45

ADVISORY BOARD BRUCE CHASSY University of Illinois, USA

PATRICK FOX University College Cork, Republic of Ireland

DENNIS GORDON North Dakota State University, USA

ROBERT HUTKINS University of Nebraska, USA

RONALD JACKSON Quebec, Canada

DARYL B. LUND University of Wisconsin, USA

CONNIE WEAVER Purdue University, USA

LOUISE WICKER University of Georgia, USA

HOWARD ZHANG Ohio State University, USA

SERIES EDITORS GEORGE F. STEWART EMIL M. MRAK C. O. CHICHESTER BERNARD S. SCHWEIGERT JOHN E. KINSELLA STEVE L. TAYLOR

(1948–1982) (1948–1987) (1959–1988) (1984–1988) (1989–1993) (1995–1993)

ADVANCES IN

FOOD AND NUTRITION RESEARCH VOLUME 45

Edited by

STEVE L. TAYLOR Department of Food Science and Technology University of Nebraska Lincoln, Nebraska USA

Amsterdam Boston London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo

This book is printed on acid-free paper. Copyright © 2003 by ACADEMIC PRESS All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Explicit permission from Academic Press is not required to reproduce a maximum of two figures or tables from an Academic Press chapter in another scientific or research publication provided that the material has not been credited to another source and that full credit to the Academic Press chapter is given.

Academic Press An Imprint of Elsevier Science 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com Academic Press An Elsevier Science Imprint 84 Theobald’s Road, London WC1X 8RR, UK http://www.academicpress.com

ISBN 0-12-016445-0

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

Typeset by RDC, Malaysia Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall 03 04 05 06 07 08 MP 9 8 7 6 5 4 3 2 1

CONTENTS

CONTRIBUTORS TO VOLUME 44 . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Inositol Phosphates in Foods Brian Q. Phillippy I. II. III. IV. V. VI. VII. VIII. IX.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemistry of Inositol Phosphates . . . . . . . . . . . . . . . . . . Metabolism of Inositol Phosphates . . . . . . . . . . . . . . . . . Inositol Phosphates in Seeds . . . . . . . . . . . . . . . . . . . . . . Inositol Phosphates in Fruits and Vegetables . . . . . . . . . . Inositol Phosphates in Animals . . . . . . . . . . . . . . . . . . . . Nutritional Importance of Inositol Phosphates . . . . . . . . . Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . Future Research Needs . . . . . . . . . . . . . . . . . . . . . . . . . . Disclaimer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1 2 7 12 22 23 32 41 42 43 43 43

Pyrrolizidine Alkaloids in Foods Roger A. Coulombe, Jr I. II. III. IV.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plant Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Structures of Pyrrolizidine Alkaloids . . . . . . . . Pyrrolizidine Alkaloids in Foods and Herbal Medicines . .

61 62 65 66

vi

CONTENTS

V. VI. VII. VIII.

Toxicity of Pyrrolizidine Alkaloids . . . . . . . . . . . . . . . . . Metabolism of Pyrrolizidine Alkaloids . . . . . . . . . . . . . . Mechanism of Toxic Action . . . . . . . . . . . . . . . . . . . . . . Control of Pyrrolizidine Alkaloids and Future Prospects . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

78 81 83 91 93 93

Ultrasonic Sensors for the Food Industry John N. Coupland and Raffaella Saggin I. II. III. IV. V.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measurement Methods . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

102 103 121 129 157 158 159

Ozone and Its Current and Future Application in the Food Industry Jin-Gab Kim, Ahmed E. Yousef and Mohammed A. Khadre I. II. III. IV. V. VI. VII. VIII. IX. X.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ozone Chemistry and Physics: An Overview . . . . . . . . . Medium for Ozone Treatment . . . . . . . . . . . . . . . . . . . . . Reactor and Equipment Considerations . . . . . . . . . . . . . . Application of Ozone in Food Processing . . . . . . . . . . . . Selected Food Applications . . . . . . . . . . . . . . . . . . . . . . . Combination Treatments . . . . . . . . . . . . . . . . . . . . . . . . . Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regulatory Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Limitations, Toxicity and Safety . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

168 170 178 182 186 193 200 204 206 206 208 208

CONTENTS

vii

The High Molecular Weight Subunits of Wheat Glutenin and Their Role in Determining Wheat Processing Properties Peter R. Shewry, Nigel G. Halford, Arthur S. Tatham, Yves Popineau, Domenico Lafiandra and Peter S. Belton I. II. III. IV. V. VI. VII. VIII. IX.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The HMW Subunits of Glutenin . . . . . . . . . . . . . . . . . . . The Sequences and Structures of HMW Subunits . . . . . . Experimental Evidence for the Role of HMW Subunits in Dough Mixing and Gluten Viscoelasticity . . . . . . . . . . Manipulating HMW Subunit Composition . . . . . . . . . . . Experimental Evidence for Differential Effects of Individual HMW Subunits on Mixing and Rheological Properties . . The Molecular Basis for Correlations between HMW Subunits and Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical Basis for Properties of the HMW Subunits and their Role in Gluten Viscoelasticity and Dough Mixing . . Conclusions and Future Prospects . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

220 224 228 242 249 256 265 273 289 290 291 303

This Page Intentionally Left Blank

CONTRIBUTORS TO VOLUME 45

Numbers in parentheses indicate the page on which the authors’ contributions begin.

Peter S. Belton, School of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ, UK (219) Roger A. Coulombe Jr., Graduate Program in Toxicology and Department of Veterinary Sciences, Utah State University, 4620 Old Main Hall, Logan, UT 84332-4620, USA (61) John N. Coupland, Department of Food Science, Pennsylvania State University, 111 Borland Laboratory, University Park, PA 16802-2504, USA (101) Nigel G. Halford, IACR-Long Ashton Research Station, Department of Agricultural Sciences, University of Bristol, Long Ashton, Bristol BS41 9AF, UK (219) Mohammed A. Khadre, Department of Food Science and Technology, The Ohio State University, 2015 Fyffe Road, Parker Hall, Columbus, OH 43210, USA (167) Jin-Gab Kim, Department of Food Science and Technology, The Ohio State University, 2015 Fyffe Road, Parker Hall, Columbus, OH 43210, USA (167) Domenico Lafiandra, Dipartimento di Agrobiologia ed Agrochimica, Universitè degli Studi della Tuscia, Via San Camillo de Lellis, Viterbo 01100, Italy (219)

x

CONTRIBUTORS

Brian Q. Phillippy, United States Department of Agriculture, Agricultural Research Service, Southern Regional Research Center, 1100 Robert E. Lee Boulevard, New Orleans, LA 70124, USA (1) Yves Popineau, Institut National de la Recherche Agronomique, Centre de Recherches de Nantes, Laboratoire de Biochimie et de Technologie des Protéines, B.P. 71627, Rue de la Géraudière, Nantes 44316, Cedex 03, France (219) Raffaella Saggin, Department of Food Science, Pennsylvania State University, 111 Borland Laboratory, University Park, PA 16802-2504, USA (101) Peter R. Shewry, IACR-Long Ashton Research Station, Department of Agricultural Sciences, University of Bristol, Long Ashton, Bristol BS41 9AF, UK (219) Arthur S. Tatham, IACR-Long Ashton Research Station, Department of Agricultural Sciences, University of Bristol, Long Ashton, Bristol BS41 9AF, UK (219) Ahmed E. Yousef, Department of Food Science and Technology, The Ohio State University, 2015 Fyffe Road, Parker Hall, Columbus, OH 43210, USA (167)

INOSITOL PHOSPHATES IN FOODS BRIAN Q. PHILLIPPY United States Department of Agriculture Agricultural Research Service Southern Regional Research Center New Orleans, LA 70124 USA

I. Introduction II. Chemistry of Inositol Phosphates A. Nomenclature B. Analysis III. Metabolism of Inositol Phosphates A. Synthesis B. Degradation IV. Inositol Phosphates in Seeds A. Whole Raw Seeds B. Processed Foods V. Inositol Phosphates in Fruits and Vegetables VI. Inositol Phosphates in Animals A. Absorption and Tissue Content B. Biological Functions VII. Nutritional Importance of Inositol Phosphates A. Bioavailability of Minerals B. Prevention of Health Disorders VIII. Summary and Conclusions IX. Future Research Needs Disclaimer Acknowledgement References

I. INTRODUCTION

Since the middle of the twentieth century, myo-inositol hexakisphosphate, which is commonly known as phytate, has been recognized as an antinutrient for its ability to bind to, precipitate and decrease the bioavailability of diand trivalent cationic minerals. Phytate is present in all seeds, usually at ADVANCES IN FOOD AND NUTRITION RESEARCH VOL 45 ISBN: 0-12-016445-0

Copyright © 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

2

B. Q. PHILLIPPY

levels between approximately 0.5 and 2% of their dry weight. In diets containing a large proportion of calories derived from grains and/or legumes, an imbalance of phytate and minerals can lead to nutritional deficiencies. Trace minerals such as zinc and iron bind to phytate most tightly and are affected to a greater degree than calcium. A comprehensive review of this subject was published by Reddy et al. (1989) and updated by Zhou and Erdman (1995) and Weaver and Kannan (2002). Until the 1980s, phytate in foods was almost always quantified using methods that were not specific for inositol hexakisphosphate. During that decade, high-performance liquid chromatography (HPLC) procedures for inositol phosphate analysis opened the way for significant advances in the collection of more accurate data. Recognition that inositol phosphates containing different numbers of phosphate groups were present in appreciable amounts in many foods also created a dilemma as to how the older data should be interpreted. This problem was magnified by the discovery that the different inositol phosphates had different effects on the bioavailability of minerals. An additional layer of complexity arose at that same time with the ability of some HPLC methods to separate some of the inositol phosphate isomers. Although isomers containing the same number of phosphate groups have not been shown to differ much in their effects on mineral availability, some chemical properties and biological functions have been linked to specific structures. Appreciation of the fundamental importance of inositol phosphates in basic cell physiology coupled with experimental data involving humans and other animals has led to a re-evaluation of the roles of inositol phosphates in food. In this review the current knowledge about these compounds is compiled and organized in an attempt to provide a context for its interpretation and to create a framework from which scientists can formulate ideas for future research.

II. CHEMISTRY OF INOSITOL PHOSPHATES A. NOMENCLATURE

There have been several changes in the rules and conventions for naming inositol phosphates since the first myo-inositol monophosphate isomer was identified in soybeans by Ballou and Pizer (1959). This resulted in some confusion in the literature of the following years, when the names assigned to enantiomers became switched and subsequently simplified. The current guidelines were issued in 1989 (NC-IUB, 1989).

INOSITOL PHOSPHATES IN FOODS

3 P

OH OH

HO HO

2

P

P P

2

OH

P

HO

P

I

II

Thorough discussions of the naming and numbering of the structures of inositol and inositol phosphates are provided by Posternak (1965) and Cosgrove and Irving (1980). Myo-inositol is one of nine isomers of inositol, which is the common name for cyclohexanehexols. It contains five equatorial and one axial hydroxyl groups numbered from one to six, with the axial hydroxyl designated as position number 2 (I). The molecule is symmetrical on either side of an axis formed by positions 2 and 5. Thus positions 1 and 3 are equivalent, as are positions 4 and 6. The first numbering convention was to depict the hydroxyls as a fraction, arranged with the most numerous hydroxyls above the inositol ring and listed in the numerator and those below the ring in the denominator. Cis hydroxyl groups were assigned the lowest possible numbers. Accordingly, phytate was referred to as myo-inositol 1,2,3,5/4,6-hexakisphosphate (II). Originally inositol phosphate isomers were numbered according to the conventions of carbohydrate chemistry. Optically active isomers were named based on the configuration of D- and L-glyceraldehyde (Lardy, 1954). Phosphates were given the lowest possible numbers and assigned the prefix D when numbered clockwise with the hydroxyl or phosphate at position 1 projecting upward, and L when numbered counterclockwise. Symmetrical isomers, also known as meso compounds, such as inositol 1,3-bisphosphate, had no D or L. A major reversal of these rules occurred in 1967, when the International Union of Pure and Applied Chemistry and the International Union of Biochemistry decided to change the nomenclature system for cyclitols (IUPAC-IUB, 1968). Following these rules, the positions were still numbered in the direction that would have the lowest possible number for the first phosphate group, but designated as D if the numbers proceeded in a clockwise direction when the hydroxyl or phosphate at position 1 projected downward, or L if the numbers rotated counterclockwise when the hydroxyl or phosphate at position 1 projected downward. The result of this change was that the D and L assignments of all of the inositol

4

B. Q. PHILLIPPY

HO

HO

1

5 6

P

HO

P

P

5

2

4

1 6

3

HO

P

P

3 2

4

P

P

P

III

IV

phosphate names in the literature prior to the implementation of these changes became switched. It was also decided at that time to use the Greek prefixes bis, tris, tetrakis, pentakis and hexakis with inositol phosphates to show that the phosphates are singly attached to the inositol carbons. In contrast, pyrophosphate linkages are represented with the Latin prefixes di, tri, tetra, etc., as in nucleotides and inositol polyphosphate pyrophosphates. An example of these rules is shown below for the enantiomeric pair of Dand L-myo-inositol 1,2,3,4-tetrakisphosphate (III, IV). Pairs of enantiomers are referred to as D and/or L when the proportion of isomers is unequal or unknown, and DL when the isomers are in equal proportion, i.e. a racemic mixture. Another revision followed the discovery in 1983 that D-myo-inositol 1,4,5-trisphosphate was a second messenger in signaling events that released calcium from intracellular stores to activate various biochemical reactions. In just a few years a complex pathway of inositol phosphate metabolism was uncovered, and hundreds of scientific articles were published. The abbreviation that became widely used at this time was Ins, preceded by a D or L, as needed, and followed by the phosphate positions enclosed in parentheses, and finally a capital P with a subscript to denote the number of phosphates when more than one. In addition, the prefix myo was omitted. Since most of the isomers in these studies had the D configuration, in 1988 the Nomenclature Committee of IUB again decided to modify the rules and name all isomers according to the D orientation and omit the D or L (NCIUB, 1989). Using the new rules, structures III and IV would be inositol 1,2,3,4-tetrakisphosphate and inositol 1,2,3,6-tetrakisphosphate, or, in abbreviated form, Ins(1,2,3,4)P4 and Ins(1,2,3,6)P4, respectively. Inositol phospholipids, or phosphoinositides, are abbreviated similarly, using PtdIns to represent phosphatidylinositol, e.g. PtdIns(4,5)P2. B. ANALYSIS

The specificity and accuracy of the analytical methods for phytate and other myo-inositol phosphates has evolved to the point where the means

INOSITOL PHOSPHATES IN FOODS

5

available to collect information may be greater than the needs of most food and nutrition scientists. The diversity of methods currently in use poses some intricate questions about exactly what data are most desirable and how they should be presented. There is no doubt that the older methods for phytate analysis are not very specific, but some are still used, and, for many food products, provide the only data available in the literature. Prior to the development of HPLC methods that separate inositol phosphates from one another, most of these compounds usually were measured together to give data intended to represent phytate. These procedures included numerous variations of the ferric chloride precipitation method and the ion exchange method that was approved by the Association of Official Analytical Chemists in 1988. Components of foods such as oxalic acid (McKenzie-Parnell and Guthrie, 1986), gallic acid (Bos et al., 1991), chlorogenic acid (Bos et al., 1991) and polyphosphate compounds that do not contain inositol, such as nucleotides (Phillippy et al., 1988), could also give elevated phytate values. Nevertheless, data obtained from seeds in their native state should be reasonably accurate, since any interfering compounds are likely to be present in small amounts compared to the large quantities of phytate. However, values from foods in which the phytate may have been partially degraded by enzymatic or thermal processes must be viewed with some caution. The low phytate values reported in fruits and vegetables are especially questionable, since those data are the most likely to be significantly inflated due to the presence of nucleotides, oxalate, etc. HPLC methods specific for phytate first appeared in the early 1980s and were soon expanded to quantify different inositol phosphates present in foods. At about the same time, the burgeoning research in signal transduction revolving around inositol 1,4,5-trisphosphate led to extremely sensitive methods for measuring inositol phosphates in animal cells. The methods developed in these related disciplines have been reviewed (Irvine, 1990; Xu et al., 1992; Skoglund and Sandberg, 2002). The currently used HPLC methods for the separation of inositol phosphates fall into two basic categories: ion pair and ion exchange. The ion pair procedure developed by Sandberg and Adherinne (1986) has been used the most often by food scientists because it separates inositol tris-, tetrakis-, pentakis- and hexakisphosphates based only on the number of phosphate groups. This simplifies their quantification and provides all the information that is usually wanted. The disadvantages are that isomers are not separated and that other polyphosphates such as nucleotides can interfere (Morris and Hill, 1996). However, nucleotides such as ATP do not appear to be present in sufficient quantities in mature ungerminated

6

B. Q. PHILLIPPY

seeds to significantly elevate the data, so this is probably a minor limitation. Some modifications have been suggested to improve the original method (Lehrfeld, 1994). Recently, an inability to use the ion pair method to measure the inositol phosphates in infant cereals was attributed to the combined high mineral and low phytate contents of these foods (Brooks and Lampi, 2001). The other type of HPLC procedure separates some of the inositol phosphate isomers by ion exchange. The method developed by Phillippy and Bland (1988) separates phytate and some isomers of inositol tris-, tetrakis- and pentakisphosphate. An improved method can now also separate inositol bisphosphates and a few more of the other isomers (Skoglund et al., 1997a, 1998; Carlsson et al., 2001). The ion exchange procedures provide more extensive data than those using ion pairing and are most useful in identifying the specific isomers present in food products or enzymatic reactions. In many cases, however, routine quantification of the numerous individual isomers present in foods may be neither practical nor justified unless the additional information is specifically desired. Whereas the above ion exchange methods employing acidic eluants give better separation of the most highly phosphorylated inositol phosphates, high pH eluants have been found useful in separating those with the lowest numbers of phosphates (Skoglund et al., 1997b, 1998). Ion exchange HPLC can also be used to separate inositol bis- to hexakisphosphates based solely on the number of phosphate groups (Rounds and Nielsen, 1993), but one should beware of the possible interferences from nucleotides such as ADP, which is present in mature soybean seeds in almost a ten-fold excess over ATP (Phillippy et al., 1994). Recently, ion exchange chromatography with high pH eluants and conductivity detection has been used to analyze phytate, other inositol phosphates and inorganic polyphosphates in foods (Sekiguchi et al., 2000; Talamond et al., 2000). Additional selective methods to analyze inositol phosphates include capillary chromatography and nuclear magnetic resonance (NMR). Several capillary electrophoresis methods have been developed to separate phytic acid and other inositol phosphates, but they do not appear to have been adopted by the scientific community, perhaps because they need further refinement (Skoglund and Sandberg, 2002). NMR methods can be used to simultaneously determine phytic acid and other inositol phosphates in a mixture (Johnson et al., 1995), and 31P NMR has been used to quantify some of the inositol phosphates in complete and digested feeds (Kemme et al., 1999). The newest evaporative light scattering detectors have the potential to significantly lower the HPLC detection limits for the most highly phosphorylated inositol phosphates in food extracts, provided that the background from interfering compounds is not too high.

INOSITOL PHOSPHATES IN FOODS

7

III. METABOLISM OF INOSITOL PHOSPHATES A. SYNTHESIS

The pathway for the synthesis of phytate has not yet been defined with complete certainty. One reason for this is that there are numerous branch points between myo-inositol and phytate, and a single direct pathway may simply be inadequate to portray this complex web of interconnected reactions. Another reason may be that there is more than one possible route, resulting in redundancy to help ensure the production of sufficient phytate to meet the needs of a particular type of cell. In nature different pathways for phytate synthesis appear to have been favored during the evolution of the diverse species of microorganisms and higher life forms. It is believed that all cells of all organisms probably contain some phytate. Studies of the simplest life forms have provided insights into the possible pathways for phytate synthesis in higher plants and animals. A complete pathway has been reported for the slime mold Dictyostelium (Stephens and Irvine, 1990); the intermediates were identified as Ins(3)P, Ins(3,6)P2, Ins(3,4,6)P3, Ins(1,3,4,6)P4 and Ins(1,3,4,5,6)P5. Each of these isomers was detected in Dictyostelium cells or homogenates incubated with [3H]inositol, and each of them was converted to InsP6 by Dictyostelium homogenates in separate experiments. Ins(1,2,4,5,6)P5 and Ins(1,2,3,4,6)P5 were also detected in the cells and homogenates and could be phosphorylated to InsP6. However, these pentakisphosphates were probably not the main precursor of InsP6, because they contained lower specific radioactivities than Ins(1,3,4,5,6)P5, and because they, and not the latter, were observed as dephosphorylation products of InsP6. Dictyostelium contains inositol 3-kinase activity, but Ins(3)P may also be derived in part from glucose 6-phosphate or phosphatidylinositol 3-phosphate (Stephens et al., 1990). Another pathway for InsP6 synthesis in Dictyostelium has been observed, starting with Ins(1,4,5)P3, which is the inositol phosphate second messenger involved in signaling via calcium release. This pathway was located in cell nuclei, and the intermediates were identified as Ins(1,3,4,5)P4 and Ins(1,3,4,5,6)P5 (Van der Kaay et al., 1995). Somewhat similarly, a soluble fraction from the yeast Schizosaccharomyces pombe converts Ins(1,4,5)P3 into Ins(1,3,4,5,6)P5 and InsP6 primarily through Ins(1,4,5,6)P4, but also partially through Ins(1,3,4,5)P4 (Ongusaha et al., 1998). In animal cells the first pathway identified for InsP6 synthesis begins with the cleavage of Ins(1,4,5)P3 from PtdIns(4,5)P2, followed by the sequential formation of Ins(1,3,4,5)P4, Ins(1,3,4)P3, Ins(1,3,4,6)P4 and Ins(1,3,4,5,6)P5 (Shears, 1989). The latter is by far the predominant InsP5 isomer in animal cells, which also contain the 2-kinase that phosphorylates

8

B. Q. PHILLIPPY

it to form InsP6 (Ji et al., 1989; Stephens et al., 1991). Evidence from avian erythrocytes suggests that Ins(1,3,4,5,6)P5 could also be produced from Ins(1)P via the stepwise formation of Ins(1,6)P2, Ins(1,4,6)P3, Ins(1,3,4,6)P4, Ins(3,4,6)P3 and Ins(3,4,5,6)P4 (Stephens and Downes, 1990). In addition, a 3-kinase that phosphorylates Ins(1,4,5,6)P4 as well as Ins(1,2,4,5,6)P5 has been detected in rat liver (Craxton et al., 1994). Recently, an inositol polyphosphate multikinase that can convert Ins(4,5)P2 to Ins(1,3,4,5,6)P5 via Ins(1,4,5)P3 and Ins(1,3,4,5)P4 was cloned from a rat cDNA library (Saiardi et al., 2001). The enzymes mentioned above and others form a complex network leading to InsP6 that may provide redundancy to ensure its synthesis and/or regulatory control over the cellular concentrations of its metabolites. Nevertheless, the fundamental route of InsP6 synthesis in animals is currently unresolved (Irvine and Schell, 2001). The inositol polyphosphate pathway in plants also appears to have alternate routes to make InsP6. As in animals, a likely precursor is Ins(1,3,4,5,6)P5. Kinases that can phosphorylate Ins(1,3,4)P3 to Ins(1,3,4,5)P4 and Ins(1,3,4,5,6)P5 have been identified in Arabidopsis and soybean seeds (Wilson and Majerus, 1997; Phillippy, 1998a). Some evidence has been obtained for the sequential phosphorylation of Ins(3)P, Ins(3,4)P2, Ins(3,4,6)P3, Ins(3,4,5,6)P4, and Ins(1,3,4,5,6)P5 in Spirodela polyrhiza and Commelina communis (Brearly and Hanke, 1996, 2000). Kinases that phosphorylate several of the inositol pentakisphosphates already containing a phosphate at position 2 have also been observed in mung bean and soybean seeds (Stephens et al., 1991; Phillippy et al., 1994), but those isomers may arise from the degradation of InsP6 rather than its synthesis. In addition, an Ins(1,4,5)P3 6-kinase has been identified in pea roots (Chattaway et al., 1992). Transcripts of the genes for L-Ins(1)P, which is the same as D-Ins(3)P, were observed in the embryo and aleurone layer of developing rice seeds shortly before the appearance of phytatecontaining particles called globoids (Yoshida et al., 1999). Furthermore, two maize mutants deficient in InsP6 synthesis produced reduced amounts of InsP6, and one had increased levels of other inositol phosphates, though the affected genes have yet to be identified unequivocally (Raboy et al., 2000). One of two types of barley mutants produced less than one fourth as much InsP6 as the parent line and accumulated 15% of the inositolbound phosphorus in D and/or L-Ins(1,3,4,5)P4, hypothetically due to a mutated Ins(1,3,4,5)P4 6-kinase gene (Hatzack et al., 2000, 2001). B. DEGRADATION

Animals, plants and microorganisms all make enzymes that break down phytic acid. Phytases are phosphatases with the ability to use InsP6 as a

INOSITOL PHOSPHATES IN FOODS

9

substrate, whereas phytases and other phosphatases can hydrolyze the various InsP5, InsP4, InsP3, InsP2 and InsP isomers. Monogastric animals including humans lack sufficient phytase in their guts to adequately break down the InsP6 in diets high in whole grains or legumes. Therefore phytases from plants and microorganisms are sometimes utilized to help degrade the inositol phosphates before and/or after foods are eaten. Inositol phosphates in foods can also be degraded by high temperatures and pressures during thermal processes such as frying and canning. The reduction of phytate content during food processing has been thoroughly reviewed by Reddy et al. (1989) and Sathe and Venkatachalam (2002). InsP6 and other inositol phosphates in food may be hydrolyzed by phytases as the food is prepared or while it passes through the gastrointestinal tract. The endogenous phytases of seeds can remove much of the InsP6 in grains and legumes if they are soaked in aqueous solutions for a number of hours prior to cooking (Larsson and Sandberg, 1992; Gustafsson and Sandberg, 1995; Fredlund et al., 1997; Bergman et al., 1999). The absorption of water initiates the germination of seeds and activates any phytases already present. New enzymes are also synthesized from freshly transcribed RNA. Some ungerminated seeds such as rye contain significant amounts of phytase (Greiner et al., 1998), while others such as maize have very little (Laboure et al., 1993). Fermentation can destroy much of the InsP6 in foods due to the action of microbial and plant phytases (Sutardi and Buckle, 1985; Gustafsson and Sandberg, 1995; Türk et al., 1996). After food is eaten, any phytases present in the food can break down phytate in the stomach. The amount of phytase activity in different types of cells and tissues does not appear to be correlated with their InsP6 content. Ungerminated rye grain contained phytase activity of 3.2 µmol min–1 g–1 at 35°C, and this activity remained fairly constant during 10 days of germination (Greiner et al., 1998). Although ungerminated spelt and barley have negligible phytase activity, after 2 and 4 days, respectively, of germination 1.1 and 1.35 µmol min–1 g–1 of activity at 35°C had accumulated (Koneitzny et al., 1994; Greiner et al., 2000b). Maize roots contained 0.50 µmol min–1 g–1 of phytase activity at 40°C (Hübel and Beck, 1996). In rat small intestine the duodenum, jejunum and ileum contained respectively 6.0, 1.3 and 1.0 µmol min–1 g–1 of activity at 60°C (Rao and Ramakrishnan, 1985), and in rat intestinal mucosal tissue the phytase activity was 0.36 µmol min–1 g–1 at 37°C (Yang et al., 1991a). Vegetables contain phytase activities at levels up to 0.15 µmol min–1 g–1 at 37°C, which is present in green onions (Phillippy, 1998b). The most widely utilized microbial phytases are secreted, although the Escherichia coli enzyme is strictly an intracellular protein.

10

B. Q. PHILLIPPY

Numerous phytases and other inositol polyphosphate phosphatases have been purified, and some of them have been cloned for experimental and industrial production. The specific activities of some of the purified phytases along with their pH optima are shown in Table I. The phytases with the highest activities in vitro are from E. coli and Peniophora lycii. Though few of the animal phytases have been purified and studied, their potential contributions to the breakdown of phytate in foods may justify a more thorough exploration of ways to increase their impact. TABLE I ACTIVITIES OF PURIFIED PHYTASES

Source

Specific activity µmol min–1 mg–1 (°C) pH optimum

Reference

Rice bran Maize seedlings Spelt seedlings Maize roots

50 2.3 262 64

(40) (55) (35) (40)

4.4 4.8 6.0 5.0

Tomato roots Rye seeds Scallion leaves Wheat bran Oat seedlings

285 517 500 260 307

(37) (35) (37) (37) (35)

4.3 6.0 5.5 6.0 5.0

Barley seedlings Faba bean seedlings Rat intestinal mucosa Rat liver Schwanniomyces castelli Aspergillus niger

117 636 5.7 0.01 441

(35) (35) (37) (37) (70)

5.0 5.0 7.5 7.4 4.4

126

(58)

5.5

110

(37)

6.5

Ullah and Gibson (1987) Berka et al. (1998)

987 142 26 29 42

(37) (37) (37) (37) (37)

4.5 6.5 6.4 6.5 5.5

Lassen et al. (1998) Wyss et al. (1999) Wyss et al. (1999) Wyss et al. (1999) Wyss et al. (1999)

811 8016 205 20 88

(37) (35) (35) (37) (37)

4.6 4.5 5.0 7.0 7.5

Wyss et al. (1999) Greiner et al. (1993) Greiner et al. (1997) Kim et al. (1998) Kerovuo et al. (1998)

Thermophilus lanuginosus Peniophora lycii Aspergillus terreus Aspergillus fumigatus Emericella nidulans Myceliophthora thermophila E. coli M15 E. coli K12 Klebsiella terrigena Bacillus sp. DS11 Bacillus subtilis

Hayakawa et al. (1989) Laboure et al. (1993) Konietzny et al (1994) Hübel and Beck, (1996) Li et al. (1997) Greiner et al. (1998) Phillippy (1998b) Nakano et al. (1999) Greiner and Alminger (1999) Greiner et al. (2000b) Greiner et al. (2001b) Yang et al. (1991a) Nogimori et al. (1991) Segueilha et al. (1992)

INOSITOL PHOSPHATES IN FOODS

11

Phytases are often categorized by their specificity in removing the first phosphate group from InsP6 to form InsP5. To some extent, the initial site of attack appears to be related to the pH optimum of the enzyme. At pH 2.0, which is the lower of its two pH optima, the phytase from the fungus Aspergillus niger produces mainly Ins(1,2,4,5,6)P5 (Irving and Cosgrove, 1972). At its higher optimum of pH 5.5 this enzyme still prefers to hydrolyze the 3-phosphate, but forms an increased proportion of Ins(1,2,3,4,5)P5. The phytase from the fungus P. lycii appears to form the same products, with increased hydrolysis of the 4- or 6-phosphate at pH 5.5 compared to pH 3.5, but at both pH values Ins(1,2,3,4,5)P5 or Ins(1,2,3,5,6)P5 is the major product (Lassen et al., 1998). NMR was used to identify the products from P. lycii phytase but, like HPLC, this technique cannot differentiate between the chemically equivalent 4 and 6 or the 1 and 3 positions. The phytase from the yeast Saccharomyces cerevisiae appears to form only Ins(1,2,4,5,6)P5 at pH 4.5 (Greiner et al., 2001a). E. coli phytase has an optimum of pH 4.5 and degrades InsP6 via Ins(1,2,3,4,5)P5 (Greiner et al., 2000a). The most studied plant phytase, from wheat bran, has maximal activity at pH 5.2 and produces predominantly Ins(1,2,3,5,6)P5 (Tomlinson and Ballou, 1962). Paramecium phytase has a pH 7.0 activity optimum (Freund et al., 1992) and initially forms Ins(1,2,3,4,5)P5 (Van der Kaay and Van Haastert, 1995). Little is known about the plant alkaline phytases such as the lily pollen enzyme, which has an optimum rate at pH 8 and yields Ins(1,2,3,4,6)P5 (Barrientos et al., 1994). The relatively nonspecific multiple inositol polyphosphate phosphatase (MIPP) from rat liver forms similar amounts of (Ins(1,2,4,5,6)P5 and/or Ins(2,3,4,5,6)P5), (Ins(1,2,3,4,5)P5 and/or Ins(1,2,3,5,6)P5), Ins(1,2,3,4,6)P5 and Ins(1,3,4,5,6)P5 from InsP6 at pH 7.4 (Nogimori et al., 1991). However, in the presence of 200 µM Al3+, the dominant product from MIPP is Ins(1,2,3,4,6)P5 (Ali et al., 1995). Following the removal of the first phosphate from InsP6, some phytases and other phosphatases proceed to remove additional phosphates adjacent to a hydroxyl group. Thus the wheat phytase hydrolyzes Ins(1,2,3,5,6)P5 to Ins(1,2,5,6)P4 and Ins(1,2,3,6)P4, followed by the formation of Ins(1,2,6)P3, Ins(1,2,3)P3 and Ins(1,5,6)P3, and so on, until only Ins(1)P, Ins(2)P and possibly myo-inositol remain (Tomlinson and Ballou, 1962; Phillippy, 1989; Nakano et al., 2000). Phytases from other cereal seeds follow a similar sequence, producing Ins(1,2,3,5,6)P5, Ins (1,2,5,6)P4, Ins(1,2,6)P3, Ins(1,2)P2 and Ins(2)P (Greiner and Alminger, 2001). A. niger phytase leaves only Ins(2)P, whereas a combination of the phytase and the pH 2.5 optimum acid phosphatase from A. niger removes all six phosphates (Wyss et al., 1999). In contrast to wheat phytase, the

12

B. Q. PHILLIPPY

predominant isomer formed by Dictyostelium phytase from Ins(1,2,3,6)P4 is Ins(2,3,6)P3 (Adelt et al., 2001). The calcium-dependent phytase from Bacillus subtilis appears to cleave alternate rather than adjacent phosphates, resulting in Ins(2,4,6)P3 and Ins(1,3,5)P3 as the end products (Kerovuo et al., 2000). Inositol phosphates that are intermediates in the synthesis of InsP6 can be degraded by various phosphatases with characteristic specificies. For example, Dictyostelium and the rat liver MIPP can produce Ins(1,4,5)P3 from Ins(1,3,4,5,6)P5 using either Ins(1,3,4,5)P4 or Ins(1,4,5,6)P4 as an intermediate (Van Dijken et al., 1995). Additional inositol phosphate phosphatases have been reviewed by Shears (1989). Thermal degradation of phytate is accelerated by low pH and high pressure. Upon hydrolysis of pure InsP6 in solution by autoclaving for 1 h at 121°C, the reduction in InsP6 content at pH 4.0, 7.0 and 10.7 was 81, 64 and 43%, respectively (Phillippy et al., 1987). The inositol phosphate breakdown products from InsP6 autoclaved at pH 4.0, resembled those formed by A. niger phytase, with Ins(1,2,4,5,6)P5 and/or Ins(2,3,4,5,6)P5 being the predominant InsP5 isomer, followed by Ins(1,2,3,4,5)P5 and/or Ins(1,2,3,5,6)P5 (Phillippy and Bland, 1988). At pH 10.8 the hydrolysis showed much less specificity, which resulted in a more even distribution of isomers. IV. INOSITOL PHOSPHATES IN SEEDS A. WHOLE RAW SEEDS

The predominant inositol phosphate in whole grains, legumes and nuts is InsP6, which may account for approximately 0.4–6% of their dry weight (Reddy, 2002). Compared to the InsP6 content, the other inositol phosphates in raw and dried seeds are present in relatively minor amounts. For this reason, data obtained using some of the nonspecific methods for phytate analysis that measure all polyphosphate compounds, although not 100% accurate, provide estimates of the InsP6 in these materials. InsP6 content is partly determined by the phosphate level of the soil, and InsP6 accumulates mainly during the final stages of soybean seed development (Raboy and Dickinson, 1987). Levels of substances such as nucleotides that interfere in its analysis would not be expected to be influenced by variations in growing conditions to as great an extent as InsP6. Mechanical processing of dry seeds such as grinding and milling do not result in significant enzymatic or thermal degradation of InsP6. However, since InsP6 is often concentrated in specific areas of seeds (Reddy et al., 1989), separation of different anatomical parts may lower or raise the percentage of InsP6 in the products as compared to the whole seeds. This is especially true for grains such as rice and wheat, where the InsP6 is found mostly in

INOSITOL PHOSPHATES IN FOODS

13

the bran and germ. Processing strategies to extract InsP6 from foods derived from seeds have been extensively reviewed (Reddy et al., 1989; Sathe and Venkatachalam, 2002). The inositol phosphate contents of some of the most widely utilized seeds used as food are listed in Table II. Literature values have been uniformly converted to g/100 g as shown in parentheses because these units are usually used to portray the nutritional composition of foods and they are easy to compare with percentage data, which has traditionally been used to report phytate content. Values of InsP6, InsP5, InsP4 and InsP3 were converted from µmol/g to g/100 g by multiplying by 0.066, 0.058, 0.050 and 0.042, respectively. As mentioned above, InsP6 levels can fluctuate considerably due to environmental and processing effects, and additional variation can result from genetic differences. Therefore the data in Table II represent random examples that may be more or less typical of each kind of seed as it exists throughout the world. InsP6 comprises more than 75% of the inositol phosphates in most of these seeds, and the InsP6 values fall within or near the ranges of the nonspecific “phytate” values reported elsewhere in the literature. InsP6 and InsP5 account for more than 95% of the total inositol phosphates in most raw grains and legumes. InsP3 is undetectable in most of these seeds and InsP4 is usually less than 5% of the total. Therefore, it may not always be necessary to perform routinely quantification of the InsP4 and InsP3 fractions of raw food materials derived from seeds. Some raw seeds and related products have not been analyzed for the different groups of inositol phosphates, but data have been obtained using HPLC and NMR methods that specifically measured only InsP6. The InsP6 contents of white rice determined by HPLC and polished rice determined by NMR were 0.23 and 0.42%, respectively (O’Neill et al., 1980; Graf and Dintzis, 1982). These values are comparable to the range of 0.14–0.34% obtained by nonspecific methods (Reddy et al., 1982). Although InsP6 accounted for only 55% of the total inositol phosphates in wild rice (Table II), in a nonquantitative study of farm rice the only inositol phosphate detected other than InsP6 was the monophosphate (Asada et al., 1969). Soybean meal was found to contain 1.44% InsP6 by HPLC (Bos et al., 1991), which is similar to the 1.24% InsP6 in raw soybeans, also determined by HPLC (Talamond et al., 2000). Other values for InsP6 obtained by NMR include 0.84–0.97% for 100% extraction wheat flour, 0.13–0.18% for white wheat flour and 0.75% for Chinese millet (O’Neill et al., 1980). Pearl millet and peanuts analyzed by HPLC contained 0.74 and 0.68% InsP6, respectively (Talamond et al., 2000), and sunflower seeds were determined by HPLC to contain 4.48% InsP6 (Cilliers and van Niekert, 1986).

14

TABLE II INOSITOL PHOSPHATES IN RAW SEEDS

InsP3 µmol/g (g/100 g)

Wheatabc Cornac Wild riced Oatabc Ryeab Barleyabcef Sorghumac Triticalea Sesamea Luping Pinto beansadh Great Northern beansdh Navy beansdh Baby Lima beansh Lima beansd Roman beansh Red kidney beansh Red chili beansh Red (Guerniquesa) beansg Brown beansi Black beans fhj Black (Tolosana) beansg Faba beansg Soybeans f Chickpeas (garbanzo beans) fgh

– – 0.67 – – 0.05–0.36 – – – – – – – – – – – – 0.21 – – 0.21 0.43 – 0.62

µmol/g

InsP4 (g/100 g)

0.52 – (0.03) 1.40 0.58 0.79 (0.00–0.01) 0.12–0.84 – 0.48 1.30 – 0.17–2.96 0.19–0.48 0.14–0.96 0.23 0.77 0.02 0.16 0.02 (0.01) – – 0.10–0.16 (0.01) – (0.02) 0.26–0.94 0.28 (0.03) 0.00–0.22

(0.03) (0.07) (0.03) (0.04) (0.01–0.04) (0.02) (0.06) (0.01–0.15) (0.01–0.02) (0.01–0.05) (0.01) (0.04) (0.00) (0.01) (0.00) (0.00–0.01) (0.01–0.05) (0.01) (0.00–0.01)

µmol/g

InsP5 (g/100 g)

0.55–3.64 0.47–0.95 3.14 0.25–3.16 0.68–4.24 0.44–3.16 0.47–1.55 3.12 17.81 0.00–1.03 0.89–8.48 1.04–2.19 1.24–1.80 2.13 1.50 1.95 1.84 2.18 0.33 1.01 1.05–1.87 0.34 1.29–3.66 1.41 0.53–1.76

(0.03–0.21) (0.03–0.05) (0.18) (0.01–0.18) (0.04–0.25) (0.03–0.18) (0.03–0.09) (0.18) (1.03) (0.00–0.06) (0.05–0.49) (0.06–0.13) (0.07–0.10) (0.12) (0.09) (0.11) (0.11) (0.13) (0.02) (0.06) (0.06–0.11) (0.02) (0.07–0.21) (0.08) (0.03–0.10)

InsP6 µmol/g

(g/100 g)

10.27–16.36 14.18–15.91 6.36 9.80–17.25 9.31–15.30 2.94–17.87 7.12–13.98 15.15 81.21 4.76–10.86 8.94–14.1 12.7–17.0 12.4–16.5 9.96 12.78 10.6 13.5 11.9 9.52 12.6 5.82–15.91 8.55 5.23–9.24 5.79 3.91–6.00

(0.68–1.08) (0.94–1.05) (0.42) (0.65–1.14) (0.61–1.01) (0.19–1.18) (0.47–0.92) (1.00) (5.36) (0.31–0.72) (0.59–0.93) (0.84–1.12) (0.82–1.09) (0.66) (0.84) (0.70) (0.89) (0.78) (0.63) (0.83) (0.38–1.05) (0.56) (0.34–0.61) (0.38) (0.26–0.40)

B. Q. PHILLIPPY

Seed

TABLE II (continued) INOSITOL PHOSPHATES IN RAW SEEDS

InsP3 µmol/g (g/100 g)

µmol/g

InsP4 (g/100 g)

Blackeye peas (cowpeas) h Pigeon peash Green split peasah Yellow split peas h Lentils ghk

0.01 (0.00) – – – 0.32–0.38 (0.01–0.02)

0.26 0.04 0.17–0.58 0.12 0.21–0.50

µmol/g

(0.01) 2.52 (0.00) 2.41 (0.01–0.03) 1.36–2.47 (0.01) 1.49 (0.01–0.02) 0.83–1.39

aData

gData

bData

hData

derived from Lehrfeld (1989). derived from Larsson and Sandberg (1992). cData derived from Kasim and Edwards (1998). dData derived from Lehrfeld (1994). eData derived from Bergman et al. (1999). fData derived from Trugo et al. (1999).

InsP5 (g/100 g) (0.15) (0.14) (0.08–0.14) (0.09) (0.05–0.08)

InsP6 µmol/g

(g/100 g)

12.6 7.96 6.06–6.48 8.82 3.70–10.77

(0.83) (0.52) (0.40–0.43) (0.58) (0.24–0.71)

derived from Burbano et al. (1995). derived from Morris and Hill (1996). iData derived from Gustafsson and Sandberg (1995). jData derived from Greiner and Konietzny (1998). kData derived from Kozlowska et al. (1996).

INOSITOL PHOSPHATES IN FOODS

Seed

15

16

B. Q. PHILLIPPY B. PROCESSED FOODS

Processing methods that degrade InsP6 result in the formation of inositol phosphates containing fewer than six phosphate groups. As the degradation proceeds, the inositol phosphate concentrations individually peak and then fall at rates determined by the processing conditions and the composition of the food. InsP6 may remain the predominant inositol phosphate even after most of it has been hydrolyzed if the other inositol phosphates accumulate to only low levels due to a slow InsP6 hydrolysis rate. Intermediates will accumulate when their rate of formation is faster than their rate of degradation. One factor that can limit their rate of formation is that the solubilities of their salts with multivalent minerals are inversely proportional to the number of phosphate groups. The solubilities of the inositol phosphates help to determine their aqueous concentrations and the reaction rates of phytases and other phosphatases during enzymatic hydrolysis. In addition, the rate of hydrolysis depends on the amount of enzyme, and higher concentrations of intermediates transiently accumulate when more enzyme is present. InsP6 in foods may be destroyed enzymatically, nonenzymatically or by a combination of these methods. Canning at high temperatures and pressures quickly inactivates phytases and results in nonenzymatic InsP6 degradation. Similarly, the toasting of breakfast cereals at very high temperatures results in nonenzymatic destruction of InsP6. Fermentation of breads results in the enzymatic breakdown of much of the InsP6 due to yeast and cereal phytases (Harland and Harland, 1980; Larsson and Sandberg, 1991; Türk et al., 1996, 2000). Soaking and hydrothermal processing of legumes and grains prior to cooking also cause much enzymatic loss of InsP6 (Gustaffson and Sandberg, 1995; Fredlund et al., 1997). Quick soaking for 2 h and overnight soaking for 18 h gave similar levels of InsP3, InsP4, InsP5 and InsP6 in various cooked beans (Morris and Hill, 1996). The inositol phosphates in various processed foods are shown in Table III. The data are on a dry weight basis except for some of the flours and breakfast cereals, which are low moisture foods. Compared to the levels in whole raw seeds (Table II), the level of InsP6 was higher in wheat bran, rice bran and wheat germ but lower in corn bran, wheat flour, corn flour and sorghum flour. InsP6 was the predominant inositol phosphate in breakfast cereals, cooked beans, peas and lentils, and breads, except for two of the sour breads. There was much greater variation in the InsP6 values for the different breakfast cereals (0.6–8.82 µmol/g) and breads (0.17–8.26 µmol/g) compared to the cooked legumes (4.93–10.1 µmol/g). The levels of InsP6 in the cooked legumes were greater than 50% of their levels in the raw seeds.

TABLE III INOSITOL PHOSPHATES IN PROCESSED FOODS

Wheat branabcd Corn brana Rice brande Wheat germa Wheat flour, 85% extraction f Wheat flour, 55% extraction f Corn flour, 95% extractiong Corn flour, 65% extractiong Corn floure Oat whole meal flourc Rye whole meal flourc Whole rye meal (1.8–2.0% ash)h Rye meal (0.5–0.6% ash)h Sorghum flourg Quinoa flouri Soy-based infant formula j Bran breakfast cerealsk Bran flake cereal j Wheat breakfast cerealsk Corn breakfast cerealsk Corn flakes cerealsh Rice breakfast cerealsk Oat breakfast cerealsk Mixed grain breakfast cerealsk Müsli cerealsh White wheat breadgj

– – – – – – – – 1.0 – – – – – – – 0.4 1.41 0.7 – – 0.1 0.2 0.1 – –

(0.04)

(0.02) (0.06) (0.03) (0.00) (0.01) (0.00)

µmol/g

InsP4 (g/100 g)

0.8–2.48 – 1.30 1.58 – – – – 1.9 – 0.06 – – 0.1 0.18 1.7 2.63 1.7 – – 0.2 1.2 0.5 1.5 0.17

(0.04–0.12) (0.06) (0.08)

(0.09) (0.00) – (0.01) (0.01) (0.08) (0.13) (0.08) (0.01) (0.06) (0.02) (0.07) (0.01)

µmol/g

InsP5 (g/100 g)

1.0–11.93 – 9.82–13.36 9.48 0.07 0.00–0.02 1.33 0.88 2.5 0.22 0.50 1.4 0.8 1.03 0.1–0.2 0.95 4.9 6.42 3.1 0.1 0.52 0.6 3.3 1.6 2.6 0.25–0.34

(0.06–0.69) (0.57–0.77) (0.55) (0.00) (0.00) (0.08) (0.05) (0.14) (0.01) (0.03) (0.08) (0.05) (0.06) (0.01) (0.05) (0.28) (0.37) (0.18) (0.01) (0.03) (0.03) (0.19) (0.09) (0.15) (0.01–0.02)

InsP6 µmol/g

(g/100 g)

46.9–84.4 0.8–1.1 99.2–103.5 30.1 1.24 0.14–0.17 10.1 2.14 1.5 10.8 10.3 15.0 7.2 5.92 8.6–11.4 3.64 12.0 8.82 4.9 0.6 1.04 0.9 7.6 3.3 8.3 0.17–0.93

(3.09–5.57) (0.05–0.07) (6.55–6.83) (1.99) (0.08) (0.01) (0.67) (0.14) (0.10) (0.71) (0.68) (0.99) (0.47) (0.39) (0.57–0.75) (0.24) (0.79) (0.58) (0.32) (0.04) (0.07) (0.06) (0.50) (0.22) (0.55) (0.01–0.06)

17

InsP3 µmol/g (g/100 g)

INOSITOL PHOSPHATES IN FOODS

Food

18

TABLE III (continued) INOSITOL PHOSPHATES IN PROCESSED FOODS

InsP3 µmol/g (g/100 g)

µmol/g

French breadh Wheat and oat breadh Wheat and rye breadh Sour rye and wheat breadh Sour wheat and potato breadh Sour buckwheat breadh Crispbreadh Whole wheat pastah Regular pastash Four cereals pastah Wheatmeal porridge flakes, rawh Ricemeal porridge flakes, rawh Oatmeal porridge flakes, rawh Ryemeal porridge flakes, rawh Barleymeal porridge flakes, rawh Pinto beans, cookedl Great Northern beans, cookedl Navy beans, cookedl Baby Lima beans, cookedl Roman beans, cookedl Red kidney beans, cookedl Red kidney beans, canned j Red chili beans, cookedl Black beans, cookedl Chickpeas, cookedl Blackeye peas, cookedl Pigeon peas, cookedl

– 3.13 – 3.02 1.43 – – – – – – – – – – 0.20 0.23 0.13 0.25 0.08 0.19 0.45 0.07 0.18 0.10 0.22 0.22

– 1.99 – – 1.59 – 1.0 – 1.1 – – – 1.3 – – 0.91 1.05 0.68 1.08 0.73 1.02 0.91 0.81 0.98 0.56 0.89 0.96

(0.13) (0.13) (0.06)

(0.01) (0.01) (0.00) (0.01) (0.00) (0.01) (0.02) (0.00) (0.01) (0.00) (0.01) (0.01)

InsP4 (g/100 g) (0.10) (0.08) (0.05) (0.05)

(0.06) (0.04) (0.05) (0.03) (0.05) (0.04) (0.05) (0.04) (0.04) (0.05) (0.03) (0.04) (0.05)

µmol/g – 2.56 – 0.60 – – 1.8 1.5 0.8 – 2.5 0.4 4.0 2.3 0.7 3.33 3.60 3.07 3.07 3.25 2.81 3.19 3.37 3.62 2.04 3.38 2.77

InsP5 (g/100 g) (0.15) (0.03) (0.10) (0.09) (0.05) (0.14) (0.02) (0.23) (0.13) (0.04) (0.19) (0.21) (0.18) (0.18) (0.19) (0.16) (0.18) (0.19) (0.21) (0.13) (0.20) (0.16)

InsP6 µmol/g

(g/100 g)

0.75 8.26 2.62 0.60 0.48 0.61 5.9 13.5 6.5 5.4 13.9 2.7 14.1 14.0 4.2 8.14 9.24 8.80 7.08 9.17 9.12 7.51 10.1 9.96 5.18 9.67 5.97

(0.05) (0.54) (0.17) (0.04) (0.03) (0.04) (0.39) (0.89) (0.43) (0.36) (0.92) (0.18) (0.93) (0.92) (0.28) (0.54) (0.61) (0.58) (0.47) (0.60) (0.60) (0.50) (0.67) (0.66) (0.34) (0.64) (0.39)

B. Q. PHILLIPPY

Food

TABLE III (continued) INOSITOL PHOSPHATES IN PROCESSED FOODS

InsP3 µmol/g (g/100 g)

µmol/g

Green split peas, cookedl Yellow split peas, cookedl Lentils, cookedl Bean sprouts, canned j Tofu j Corn doughg Cassava doughg Bankug Ekuegbemig Fufug Garig Fanti kenkeyg Ga kenkeyg Wheat-based meal, cookedm Rice-based meal, cookedm Maize-based meal, cookedm Pearl millet-based meal, cookedm Finger millet-based meal, cookedm Sorghum-based meal, cookedm

0.07 0.05 0.44 0.20 – – – – – – – – – 2.79 1.62 1.86 2.29 2.17 2.21

0.45 0.52 0.97 0.22 – – – – – – – – – 2.98 1.98 1.42 2.74 1.68 2.30

(0.00) (0.00) (0.02) (0.01)

(0.12) (0.07) (0.08) (0.10) (0.09) (0.09)

InsP4 (g/100 g) (0.02) (0.03) (0.05) (0.01)

(0.15) (0.10) (0.07) (0.14) (0.08) (0.11)

µmol/g 1.73 1.53 3.62 0.60 1.21 3.31 0.34 2.34 1.03 0.60 0.17 2.26 2.05 4.53 2.53 2.84 4.40 3.50 3.66

aData

hData

bData

iData

(0.10) (0.09) (0.21) (0.03) (0.07) (0.19) (0.02) (0.14) (0.06) (0.03) (0.01) (0.13) (0.12) (0.26) (0.15) (0.16) (0.25) (0.20) (0.21)

InsP6 µmol/g

(g/100 g)

4.93 7.35 7.09 1.57 22.0 2.24 0.21 2.12 1.52 0.98 0.59 1.15 1.18 12.3 5.32 7.73 11.0 10.9 9.82

(0.32) (0.48) (0.47) (0.10) (1.45) (0.15) (0.01) (0.14) (0.10) (0.06) (0.04) (0.08) (0.08) (0.81) (0.35) (0.51) (0.72) (0.72) (0.65)

derived from Plaami and Kumpulainen (1995). derived from Valencia et al. (1999). jData derived from Phillippy et al. (1988). kData derived from Morris and Hill (1995). lData derived from Morris and Hill (1996). mData derived from Agte et al. (1999).

19

derived from Lehrfeld (1989). derived from Sandberg and Ahderinne (1986). cData derived from Sandberg and Svanberg (1991). dData derived from Kasim and Edwards (1998). eData derived from Lehrfeld (1994). fData derived from Brune et al. (1992). gData derived from Ferguson et al. (1993).

InsP5 (g/100 g)

INOSITOL PHOSPHATES IN FOODS

Food

20

B. Q. PHILLIPPY

Brune et al. (1992) studied the breakdown of InsP6 during breadmaking. Wheat flour of 85% extraction contained 0.829 and 0.42 g/kg of InsP6 and InsP5, respectively. Bread made from this flour had 0.049, 0.008, 0.011 and 0.040 g/kg InsP6, InsP5, InsP4 and InsP3, respectively, based on the dry weight of the flour used. In contrast, wheat flour of 55% extraction contained 0.089–0.106 and 0.005–0.006 g/kg InsP6 and InsP5, respectively. The levels of InsP6, InsP5, InsP4 and InsP3 in bread made from this flour were 0.003–0.013, 0.001–0.005, 0.001–0.002 and 0.005–0.020 g/kg, respectively. The loss of InsP6 during breadmaking was 94% for the 85% extraction flour. The effect of pH on the hydrolysis of InsP6 in wheat breads containing varying amounts of rye bran, oat flour or oat bran was investigated by Larsson and Sandberg (1991). The acidities of the doughs were raised by the addition of sour dough or lactic acid to pH 3.9–5.5, which resulted in breads with pH 4.2–5.7. The sum of InsP6 and InsP5 was reduced by 79–97% in the breads containing sour dough, whereas the total of InsP6 plus InsP5 decreased 44–97% in breads with lactic acid. Optimum hydrolysis occurred when the pH was between 4.3 and 5.1 in both the dough and the bread. Less InsP6 plus InsP5 was destroyed in breads containing oat bran than in the breads made with rye bran or oat flour owing to lower phytase activities in the oat bran. In breads made from equal parts of whole and white wheat flours, the addition of lactic acid and A. niger phytase completely eliminated InsP6 and InsP5 (Türk and Sandberg, 1992). Results in the same study also showed that unfermented milk inhibited the destruction of InsP6 and InsP5 to a greater extent than could be accounted for by its calcium content. Türk et al. (1996) investigated the isomeric structures of the inositol phosphates present in whole wheat bread. Ion chromatography revealed that D- and/or L-Ins(1,2,3,4,5)P5 was the predominant InsP5, followed by D- and/or LIns(1,2,4,5,6)P5, Ins(1,2,3,4,6)P5 and Ins(1,3,4,5,6)P5. Similar amounts of what appeared to be D- and/or L-Ins(1,2,3,4)P4 and D- and/or L-Ins(1,2,5,6)P4 were also present, most likely due to the action of the wheat and yeast phytases. Kozlowska et al. (1996) ground raw lentils containing 10.77, 1.07, 0.50 and 0.37 µmol/g InsP6, InsP5, InsP4 and InsP3, respectively, into flour which was then fermented for 4 days. Suspensions of 79 g lentil flour per liter of sterile tap water fermented at 28 or 42°C lost 73 and 78%, respectively, of their InsP6. In contrast, suspensions of 221 g flour per liter fermented similarly had InsP6 decreases of only 63 and 68% at 28 and 42°C, respectively. Interestingly, InsP3 appeared to increase to fairly high levels, 1.51 and 1.91 µmol/g, respectively, in the latter two suspensions after 4 days. However, additional work is needed to determine the

INOSITOL PHOSPHATES IN FOODS

21

composition of these fractions, since nucleotides such as ATP will elute along with InsP3 in the ion pair HPLC analysis (Morris and Hill, 1996). Optimized procedures have been developed to degrade the inositol phosphates in pea flour (Fredrikson et al., 2001b). At pH 7.5 and 45°C, 66% of the sum of InsP6, InsP5, InsP4 and InsP3 was eliminated by endogenous phytase activity within 10 h. The predominant InsP5 and InsP4 breakdown products were Ins(1,2,3,4,6)P5 and D- and/or LIns(1,2,3,4)P4, which is a pattern similar to that of the alkaline phytase of lily pollen (Barrientos et al., 1994). A modified process employing an exogenous microbial phytase for just 1 h was used to prepare a dephytinized pea protein isolate flour intended for test production of infant formulas (Fredrikson et al., 2001a). To increase the reduction of InsP6 during the soaking of black beans containing 15.91, 1.72, 0.16 and 0 µmol/g of InsP6, InsP5, InsP4 and InsP3, respectively, the effects of pH, temperature and exogenous enzymes were evaluated (Greiner and Konietzny, 1999). When the pH of the soaking buffer was adjusted to pH values from 4.5 to 8.0, the greatest degradation of InsP6 occurred at pH 6.0 during a 15 h incubation at 50°C and resulted in 7.64, 1.40, 3.85 and 4.75 µmol/g of InsP6, InsP5, InsP4 and InsP3, respectively. As mentioned above for lentils, large apparent InsP3 values could conceivably be due to nucleotides such as ATP. Whereas soaking overnight in water at room temperature followed by cooking resulted in only an 8% decrease in the sum of InsP6 and InsP5, soaking at pH 6.0 and 60°C for 15 h led to a 54% reduction after cooking. Addition of E. coli or rye phytase to the soaking buffer, 50 mM sodium acetate pH 5.5, during the last 2 h of a 15 h soak, had no effect at 25°C, but reduced the sum of InsP6 and InsP5 by 29 and 34%, respectively, at 50°C. Cooking led to decreases in beans soaked with E. coli or rye phytases at both temperatures such that the sums of InsP6 and InsP5 decreased by 46 and 39%, respectively, in beans soaked at 25°C and by 82 and 70%, respectively, in beans soaked at 50°C. Germination of black beans at 25°C in the dark required 14 days to achieve a 47% reduction in the sum of InsP6 and InsP5. Grains can also be soaked during their initial treatment in procedures such as steeping, malting or hydrothermal processing. Soaking leads to the breakdown of InsP6 by endogenous phytases in wheat, rye, oat, barley and corn (Sandberg and Svanberg, 1991; Larsson and Sandberg, 1992; Hull and Montgomery, 1995; Fredlund et al., 1997). Bergman et al. (1999) optimized the hydrothermal processing of barley to degrade InsP6 and increase the level of myo-inositol. Variables were the temperature (T1) of the first 1 h wet and 5 h dry steeps, the temperature (T2) of the second 1 h wet and 15 h dry steeps, and the concentration and volume of lactic

22

B. Q. PHILLIPPY

acid in the wet steeps. InsP6 was lowered by 96% to a final concentration of 0.5 µmol per g dry weight using 48 and 50°C for T1 and T2, respectively, and 3.2 volumes of 0.8% lactic acid. Myo-inositol was increased from 0.56 to 2.68 µmol per g dry weight with 48°C for T1, during which the dry steep was prolonged to 21 h, 1.5 volumes of 0.8% lactic acid and without the second steeps. The best combination, a 95% decrease in InsP6 and a myo-inositol value of 2.23 µmol per g dry weight, was achieved using 48 and 50°C for T1 and T2, respectively, and 1.5 volumes of 0.8% lactic acid. Under the latter conditions the final contents of InsP6, InsP5, InsP4, InsP3, InsP2 and InsP were 0.66, 0.15, 0.79, 1.73, 1.1 and 0.9 µmol per g dry weight, respectively. Germination is responsible for the InsP6 breakdown in malting processes for legumes and grains. Trugo et al. (1999) studied the effect of malting on the inositol phosphates in soybean, black bean, chickpea and barley seeds. After 2 days of malting the sum of InsP5 and InsP6 was reduced by 25% in black beans, which had the greatest losses. Higher levels of InsP3 and InsP4 were generated during malting, but InsP6 was the predominant inositol phosphate present in all of the malted products. Additional reports on seed germination showed that little of the inositol phosphates in lentils was lost after 3 days but more than 70% was hydrolyzed by day 6 (Ayet et al., 1997), and 22–38% of the inositol phosphates in two species of lupin was destroyed in 4 days (de la Cuadra et al., 1994). Slight increases in InsP3 and InsP4 were observed during germination of the lentils and one of the species of lupin. Valencia et al. (1999) compared soaking and fermentation of raw and germinated quinoa flours followed by cooking. Cooking alone lowered the InsP6 content of raw flour from 8.6–11.4 µmol/g to 6.9–9.5 µmol/g. Soaking in a suspension with three parts of water for 12–14 h at room temperature prior to cooking lowered the InsP6 to 2.0–4.3 µmol/g, whereas fermentation with Lactobacillus plantarum prior to cooking gave 1.0–2.0 µmol/g InsP6. The most extensive reduction of InsP6 was in germinated flour that was fermented prior to cooking, which gave 0.2–0.3 µmol/g InsP6, 0.0 µmol/g InsP5 and 0.0–0.1 µmol/g each of InsP4 and InsP3. Soaking of maize flour or pounded maize was recently found to reduce the InsP5 and InsP6 content by more than half and was preferable to fermentation because the conditions were easier to control (Hotz and Gibson, 2001). V. INOSITOL PHOSPHATES IN FRUITS AND VEGETABLES

Extremely little is known about the inositol phosphates in fruits and vegetables. Data obtained using nonspecific methods for phytate analysis

INOSITOL PHOSPHATES IN FOODS

23

indicate that the amounts of InsP6 in these foods are considerably lower than in seeds. A notable exception is avocado fruit, which contains approximately 0.5% InsP6 on a dry weight basis (Phillippy and Wyatt, 2001). Avocado fruit, like seeds, contains a high level of fat that needs to be protected from oxidation. Therefore the fact that avocado fruit contains large amounts of InsP6 bolsters the hypothesis that in vivo InsP6 serves as an antioxidant to prevent iron-catalyzed free radical formation (Graf et al., 1987). Some vegetables that grow underground in the form of bulbs, roots and tubers may have higher amounts of inositol phosphates than green vegetables. Onions, parsnips and carrots have been reported to contain 0.38, 0.24 and 0.04% InsP6, respectively, according to dry weight, whereas turnips, beet roots, celery and cabbage all had 0.02% or less InsP6 (Harland and Morris, 1995). In all of these vegetables InsP6 was the predominant inositol phosphate, and smaller amounts of InsP5, InsP4 and InsP3 were also measured. Potatoes analyzed by NMR were reported to contain 0.09% InsP6 (O’Neill et al., 1980), but HPLC results with several varieties of store-bought potatoes have revealed InsP6 concentrations on average of about 0.3% on a dry weight basis (B. Q. Phillippy, unpublished data). Post-harvest changes in the inositol phosphates of these foods have never been studied, and it is not known whether the above data are typical for these foods or how much variation might be expected. Plant cells also contain Ins(1,4,5)P3, which is involved in calcium signaling and growth (Stevenson et al., 2000). Ins(1,4,5)P3 is produced upon the hydrolysis of PtdIns(4,5)P2 by phospholipase C following stimulation. Ins(1,4,5)P3 then binds to receptors in the vacuole to release calcium stores, which evoke a biological response (Munnik et al., 1998). The levels of Ins(1,4,5)P3 are low compared to those of InsP6 and some of the other inositol phosphates. Red beets contain 8–11 pmol Ins(1,4,5)P3/g of fresh weight (Beno-Moualem et al., 1995), and the stems of maize plants have 139–276 pmol Ins(1,4,5)P3/g of fresh weight (Perera et al., 1999). In vegetative plant tissues Ins(1,4,5)P3 appears to be a minor component of the InsP3 fraction, which includes Ins(1,2,3)P3, Ins(3,4,6)P3, D- and/or LIns(1,5,6)P3, D- and/or L-Ins(2,4,5)P3 and D- and/or L-Ins(1,2,5)P3 (Brearley and Hanke, 2000). Other inositol phosphates identified in Spirodela polyrhiza L. (duckweed) turions, which are similar to tubers, include Ins(3,4,5,6)P4, Ins(1,3,4,5,6)P5 and D- and/or L-Ins(1,2,4,5,6)P5 (Brearley and Hanke, 1996). VI. INOSITOL PHOSPHATES IN ANIMALS A. ABSORPTION AND TISSUE CONTENT

The fate of inositol phosphates eaten by animals is particularly complex. Many factors interact as the inositol phosphates make their way through

24

B. Q. PHILLIPPY

the digestive tract to determine whether they will be excreted, degraded or absorbed. The composition of the diet may be most important, since food can contribute phytases and other phosphatases, minerals that affect the solubilities of the inositol phosphates, ionic components that can bind to the inositol phosphates, and components that may interact more indirectly. Nondietary factors that may also play a role in the destiny of the inositol phosphates include the genetic disposition of the animal as well as its physiological and nutritional status. Significant enzymatic degradation of inositol phosphates can occur in the stomach of monogastric animals including humans if the right type and amount of phosphatases are present in the diet (Sandberg and Andersson, 1988; Sandberg et al., 1996; Mullaney et al., 2000). Additional phytases and phosphatases will be active in the intestines and may come from cells in the intestinal mucosa (Yang et al., 1991b; Maenz and Classen, 1998) or from microorganisms (Moore and Veum, 1983; Lopez et al., 1998). Adaptation to InsP6 increased the phytase activity in the duodenum and ileum of rats, whereas wheat bran increased phytase activity only in the ileum (Lopez et al., 2000b). Resistant starch fed to rats adapted to wheat bran doubled the disappearance of InsP6 in the feces, most likely by promoting the growth of digestive microflora that express phytase activity (Lopez et al., 2000a). Inositol liberated from inositol phosphates during digestion is actively absorbed in the small intestine and is transported through the blood for absorption by other tissues (Holub, 1982). Inositol phosphates may have some limited ability to be absorbed, but the situation is far from clear. Sakamoto et al. (1993) found that [3H]InsP6 in drinking water appeared in gastric mucosal cells as a mixture of inositol and inositol mono- through hexakisphosphates, but only inositol and the monophosphate were detected in the blood plasma. Similarly, Vucenik and Shamsuddin (1994) showed that [3H]InsP6 incubated with HT-29 (human colon adenocarcinoma) cells appeared as a mixture of inositol and inositol mono- through hexakisphosphates in the cytosol. In contrast, YAC-1 (mouse lymphoma) and K562 (human erythroleukemia) cells exposed to [3H]InsP6 contained some of the less phosphorylated inositols but no detectable InsP6. These results show that the cells contain phytase that can degrade InsP6, but they do not show which compound(s) were absorbed, since the phytase and other phosphatases could have acted outside of the cells. Grases et al. (2001a) recently demonstrated that the addition of dodecasodium phytate at a level of 1% by weight to diets devoid of InsP6 and other forms of inositol for 12 weeks resulted in dramatic increases by more than ten-fold in the level of InsP6 in the brain, liver, kidney, bone, urine and plasma of rats. Because inositol was not used as a control, it

INOSITOL PHOSPHATES IN FOODS

25

could not be estimated how much of the tissue InsP6 may have originated from synthesis via inositol liberated from InsP6 by phytases in the gut. However, the results do show that tissue levels of InsP6 respond to dietary manipulation in rats. The addition of InsP6 to the rat diet resulted in a smaller increase in InsP5 compared to InsP6 in the tissues and fluids (Grases et al., 2001b), indicating that InsP6 absorbed intact may have served as a precursor for the InsP5. In a study with human volunteers, Grases et al. (2001c) determined that the plasma levels of InsP6 were normally 260±30 µg L–1 but fell to 70±10 µg L–1 after 15 days on an diet containing no cereal products, legumes, nuts, potatoes or coffee. Maximum concentration of InsP6 in plasma was reached 4 h after ingestion of 1400 mg dodecasodium phytate, and urinary InsP6 levels were directly related to plasma levels. It was noted that the overall percentage of absorption was low and that maximum urinary excretion, and thus maximum plasma levels, could be obtained by consumption of a diet containing normal amounts of InsP6. Inositol levels in the diets were not measured nor was inositol used as a control to compare the effects of dietary inositol and dietary InsP6 on InsP6 levels in plasma and urine. Adaptor protein 2 (AP-2) is an InsP6 receptor associated with the plasma membrane (Voglmaier et al., 1992). By binding to AP-2, cytosolic InsP6 prevents the binding of clathrin to AP-2 to form coated pits that take part in endocytosis (Gaidarov et al., 1996). However, the assembled coat structures containing clathrin and AP-2 have a greater affinity for dioctanoylphosphatidylinositol 3,4,5-trisphosphate than for InsP6, suggesting that endogenous phosphoinositides occupy the AP-2 binding sites in the plasma membrane (Gaidarov et al., 1996). It is not known whether InsP6 can be absorbed by endocytosis via these or other receptors. Good evidence for the intact absorption of inositol phosphates was reported by Ozaki et al. (2000). Using polyamines including aminoglycosides such as neomycin, synthetic spherical dendrimeric polyamines with 12 or 32 primary amines and polybasic nuclear histone proteins as carriers, phosphoinositides or inositol phosphates were translocated into a variety of cells. Although the efficiency of cellular uptake was best for inositides with lipophilic moieties, an undetermined portion of 77 µM Ins(1,4,5)P3 preincubated with 50 µM type III-S histone from calf thymus was taken up by NIH3T3 mouse fibroblasts and induced rapid cytosolic calcium mobilization. Thus polybasic compounds can neutralize the charge of inositol phosphates in a manner analogous to the intracellular delivery of oligonucleotides. Foods obtained from plants and animals contain a variety of compounds containing amines that could be tested for their ability to carry inositol phosphates into cells.

26

B. Q. PHILLIPPY

The large number of InsP2, InsP3 and InsP4 isomers identified in animal cells has made quantification of their individual concentrations a tedious prospect. Some effort has been made to determine the amounts of those isomers which are most abundant. Ins(1,4,5)P3 is present in animal cells at resting concentrations of about 0.1–2 µM with several-fold increases observed upon stimulation (Shears, 1989). Unstimulated levels of Ins(1,3,4)P3 in some cells are 1–4 µM (Shears, 1989), whereas Ins(1,2,3)P3 levels of 0.6–13.1 µM have been reported in various cell types (Barker et al., 1995). The predominant inositol phosphates in most types of animal cells appear to be InsP6 and Ins(1,3,4,5,6)P5. In heart, kidney, spleen, liver and blood but not muscle of Buffalo rats InsP6 and Ins(1,3,4,5,6)P5 were typically present at levels of at least 5–15 nmol/g wet weight (Szwergold et al., 1987). InsP6 levels alone were determined to be 1.04–2.80, 3.20–4.10, 30.0–42.0 and 1.07–1.21 µg/g in kidney, liver, brain and bone, respectively, of rats, and 0.14–0.31 and 0.43–2.52 µg/ml in plasma and urine, respectively, of humans (March et al., 2001). In various types of human blood cells InsP6 ranged from 10 to 105 µM, whereas Ins(1,3,4,5,6)P5 concentrations were 3–55 µM (Pittet et al., 1989; Bunce et al., 1993; Guse et al., 1993). Other isomers of InsP5, InsP4 and InsP3 were present in these cells at lower levels, several of them between 1 and 10 µM. InsP2 isomers were somewhat more abundant than InsP4 or InsP3, and the total monophosphates amounted to 27–93 µM (Bunce et al., 1993). The inositol polyphosphate pyrophosphates consist of myo-inositol esters where five or six of the hydroxyl groups are substituted with different combinations of monophosphate and diphosphate groups yielding various configurations of InsP6, InsP7 and InsP8. Radiolabeling experiments in AR4-2J pancreatoma cells indicates that their concentrations are probably low compared to concentrations of many of the other inositol polyphosphates (Shears et al., 1995). Quick-frozen muscles from frogs and rats contained about 2–3 nmol InsP6/g wet tissue (Table IV). Assuming that the muscles contained approximately 75% H2O, the InsP6 levels on a dry weight basis were about 10 nmol g–1, which is 1000-fold less than the InsP6 concentration found in raw seeds (Table II). After InsP6, the most abundant inositol phosphate in muscle appears to be D- and/or L-Ins(1,4)P2. The inositol phosphate concentrations in animal products other than muscle used as food have been largely ignored owing to their low levels in comparison to the levels found in seeds. In fresh turkey blood the predominant inositol phosphate was Ins(1,3,4,5,6)P5 at 1.1 mM, followed by 27 µM Ins(1,4,5,6)P4, 6.6 µM InsP6, 5.3 µM Ins(1,3,4,6)P4, 1.8 µM Ins(1,5,6)P3, 1.2 µM Ins(1,3,4,5)P4, 1.1 µM Ins(2,4,5)P3, 1.0 µM Ins(1,4,5)P3 and 0.4 µM Ins(1,3,4)P3 (Radenberg et al., 1989). In contrast, refrigerated

INOSITOL PHOSPHATES IN FOODS

27

TABLE IV MASSES OF SOME INOSITOL PHOSPHATES IN SKELETAL MUSCLESa

InsP6 Ins(1,3,4,5,6)P5 D- and/or L-Ins(1,2,4,5,6)P5 D- and/or L-Ins(1,4,5,6)P4 D- and/or L-Ins(1,2,5,6)P4 D- and/or L-Ins(1,3,4,5)P4 D- and/or L-Ins(1,4,5)P3 D- and/or L-Ins(1,3,4)P3 Ins(1,3,5)P3 and/or Ins(2,4,6)P3 D- and/or L-Ins(1,4)P2 Ins(1,3)P2 a

Frog (nmol/g wet weight)

Rat (nmol/g wet weight)

2.44–2.91 0.52–0.65 0.14–0.24 0.12–0.17 0.03–0.05 0.12–0.19 1.21–1.46 0.07–0.18 0.05–0.13 2.03–2.69 0.28–0.87

1.85–3.03 0.48–0.65 <0.02–0.15 0.24–0.50 <0.03–0.06 0.17–0.59 0.69–1.48 0.13 <0.03 3.29–4.07 1.33–1.85

Data compiled from Mayr and Thieleczek (1991).

calf brains contained mostly InsP6 followed by Ins(1,3,4,5,6)P5 (Phillippy and Bland, 1988). Interestingly, in rats injected with [3H]inositol, more labeled InsP6 was present in the hippocampus after 24 h than in the other regions of the brain (Vallejo et al., 1987). The intracellular level of InsP6 in the rat hippocampus was estimated to be 13 µM, and similar concentrations were detected in the cerebellum, cortex and striatum (Yang et al., 2001). B. BIOLOGICAL FUNCTIONS

All of our knowledge about the biological functions of inositol phosphates has come about within the last twenty years (reviewed in Irvine and Schell, 2001). In 1983 Streb and coworkers discovered that Ins(1,4,5)P3, which is formed by the action of phospholipase C on PtdIns(4,5)P2, releases calcium from a nonmitochondrial source in pancreatic acinar cells (Streb et al., 1983). Subsequently other inositol phosphates were identified and their metabolic relationships were elucidated. In addition to calcium mobilization, a variety of other signaling functions have been associated with certain inositol phosphates (Shears, 1998). In particular, InsP6 seems to be involved in numerous cellular processes as a result of its proclivity to bind to cationic minerals and proteins. Binding of ligands such as hormones, neurotransmitters and growth factors to their receptors in the plasma membrane causes the hydrolysis of PtdIns(4,5)P2 by phospholipase C to yield diacylglycerol and Ins(1,4,5)P3, which translocates through the cytoplasm as a second messenger (Berridge, 1993). Ins(1,4,5)P3 receptors are calcium channels found in the

28

B. Q. PHILLIPPY

membranes of cellular organelles including the endoplasmic reticulum (Taylor et al., 1999), sarcoplasmic reticulum (Tasker et al., 2000), nuclear membrane (Humbert et al., 1996) and plasma membrane (Tanimura et al., 2000). Upon binding, Ins(1,4,5)P3 activates its receptor, which results in the opening of the calcium channel and the release of stored calcium or uptake of extracellular calcium. Many cellular processes including growth, fertilization, secretion, contraction and sensation have been linked to Ins(1,4,5)P3 signaling (Berridge, 1993). In addition, Ins(1,4,5)P3 is involved in the regulation of cellular proliferation and apoptosis through this pathway (Patel et al., 1999; Jayaraman and Marks, 2000). Different inositol phosphates can bind to and activate Ins(1,4,5)P3 receptors in Xenopus oocytes or Chinese hamster ovary cells. Some metabolites of Ins(1,4,5)P3 such as Ins(1,3,4,5)P4 and Ins(1,3,4,6)P4 display activity, although they are less potent than the former (Delisle et al., 1994; Burford et al., 1997). Highly active isomers that are not naturally abundant include Ins(4,5)P2, Ins(2,4,5)P3, DL-Ins(1,4,6)P3 and DL-Ins(1,2,4,5)P4. Interestingly, some isomers formed by plant phytases, such as Ins(1,2,3)P3, may have a low calcium-releasing activity. Other isomers that released calcium from Xenopus oocytes included DL-Ins(1,2,6)P3, DL-Ins(1,5,6)P3, DL-Ins(1,2,3,6)P4, DL-Ins(1,2,5,6)P4 and DL-Ins(1,2,3,5,6)P5, although the relative activities of the enantiomers in these pairs were not determined (Delisle et al., 1994). In addition to membrane-bound receptors, inositol phosphates also interact with soluble proteins. Ins(1,4)P2 binds to and activates 6-phosphofructokinase (Mayr, 1989). Ins(1,4)P2, Ins(1,4,5)P3 and Ins(1,3,4,5)P4 bind to and inhibit fructose 1,6-bisphosphate aldolase A (Koppitz et al., 1986). Aldolase-bound Ins(1,4,5)P3 may also act as a pre-existing pool of the second messenger Ins(1,4,5)P3 that is discharged by fructose 1,6-bisphosphate during glycolysis in muscle (Thieleczek et al., 1989). Ins(1,2,6)P3 is known to be produced from Ins(1,2,5,6)P4 by phytases from plants and microbes (Phillippy, 1989; Türk et al., 2000). D- and/or LIns(1,2,6)P3 has been identified as a minor inositol trisphosphate in some animal cells (McConnell et al., 1992). Also known as the pharmaceutical drug α-trinositol, Ins(1,2,6)P3 has analgesic and anti-inflammatory properties and is an antagonist of neuropeptide Y (Bell and McDermott, 1998). Although its mode of action is unknown, Ins(1,2,6)P3 may inhibit signal transduction by binding to proteins such as receptors for Ins(1,3,4,5)P4 (Bell and McDermott, 1998). Ins(1,2,3)P3 contains the functional 1,2,3-trisphosphate grouping that binds iron such that it cannot catalyze the formation of hydroxyl free radicals (Hawkins et al., 1993; Spiers et al., 1995, 1996). While acting as intracellular antioxidants, Ins(1,2,3)P3 and other inositol phosphates containing this

INOSITOL PHOSPHATES IN FOODS

29

grouping might safely transport iron between sites within the cell (Barker et al., 1995). Chelation of iron to the 1,2,3-trisphosphate grouping may also reduce the likelihood for lipid peroxidation catalyzed by iron bound to other anionic compounds (Phillippy and Graf, 1997). In vitro Ins(1,2,3)P3 was nearly as effective as InsP6 at preventing iron-catalyzed hydroxyl radical formation (Spiers et al., 1995, 1996), whereas InsP6 was significantly better than Ins(1,2,3)P3 at inhibiting iron-catalyzed lipid peroxidation (Phillippy and Graf, 1997). Ins(1,3,4,5)P4 appears to act synergistically with Ins(1,4,5)P3 in the mobilization of calcium from intracellular stores. However, results from different studies have been variable and its mechanism is unclear (Smith et al., 2000). Low concentrations of Ins(1,3,4,5)P4 may facilitate calcium influx by inhibiting Ins(1,4,5)P3 5-phosphatase, whereas higher concentrations may inhibit calcium signaling by binding to Ins(1,4,5)P3 receptors (Hermosura et al., 2000). Ins(1,3,4,5)P4 and other inositol phosphates with structural similarities to PtdIns(3,4,5)P3 also compete with the latter for binding to proteins containing pleckstrin homology (PH) domains. This helps to regulate the recruitment of signaling molecules containing PH domains such as Gap1, protein kinase B (also known as Akt) and phospholipase C to cellular membranes (Kavran et al., 1998). Ins(3,4,5,6)P4 inhibits chloride secretion by epithelial cells following prolonged stimulation of Ins(1,4,5)P3 production (Vajanaphanich et al., 1994). Receptor-mediated inositol phosphate turnover increases the conversion of Ins(1,3,4,5,6)P5 to Ins(3,4,5,6)P4, which inactivates chloride channels in the plasma membrane (Xie et al., 1996). In human pancreatoma epithelial cells, Ins(3,4,5,6)P4 specifically attenuates the longer term activation of calcium-dependent chloride channels by type II calmodulindependent protein kinase following the acute phase of calcium mobilization (Ho et al., 2001). Deficits in chloride channel activity regulated by Ins(3,4,5,6)P4 may be involved in the kidney and lung pathology resulting from diabetes and cystic fibrosis, respectively (Ismailov et al., 1996). However, effective Ins(3,4,5,6)P4 concentrations may be essential for the chloride ion regulation of metabolic functions such as the salt and fluid secretion of intestinal epithelial cells (Vajanaphanich et al., 1994). Ins(1,3,4,5,6)P5 has not been assigned any specific functions other than being a key intermediate in the formation of Ins(3,4,5,6)P4 and InsP6, and more recently as a substrate for the tumor suppressor protein PTEN (phosphatase and tensin homolog deleted on chromosome ten), which is a protein phosphatase, a PtdIns(3,4,5)P3 3-phosphatase and an Ins(1,3,4,5,6)P5 3-phosphatase (Caffrey et al., 2001). However, its structural similarity to certain InsP4 isomers and InsP6 results in some sharing of functionality. For example, both Ins(1,4,5,6)P4 and Ins(1,3,4,5,6)P5 interact strongly

30

B. Q. PHILLIPPY

with the PH domain of protein kinase B (Razzini et al., 2000). Similarly, both Ins(1,3,4,5,6)P5 and InsP6 bind L- and P-selectins (Cecconi et al., 1994). In birds and some reptiles Ins(1,3,4,5,6)P5 binds to hemoglobin and regulates its affinity for oxygen (Gersonde and Ganguly, 1986). InsP6 has the ability to serve the same function, but neither inositol phosphate is naturally present in mammalian erythrocytes in sufficient amounts. Therefore, InsP6 has been incorporated into human red blood cells in order to increase the delivery of oxygen to tissues (Boucher et al., 1996). Numerous roles for InsP6 in cells have been suggested to date (Shears, 2001), and it is very likely that more remain to be discovered. With its high negative charge density, much of InsP6 is probably bound to cellular membranes through bridges of metal cations such as Mg2+ and Fe3+ (Poyner et al., 1993). By chelating iron in a manner that prevents it from catalyzing the formation of hydroxyl free radicals, InsP6 may be a critical antioxidant (Graf and Eaton, 1990; Hawkins et al., 1993). Some of the proteins InsP6 strongly binds to include L- and P-selectins (Cecconi et al., 1994), AP-2 (Timerman et al., 1992; Voglemaier et al., 1992), AP-3 (Norris et al., 1995; Ye et al., 1995), coatomer (Fleisher et al., 1994), synaptotagmin (Fukuda et al., 1994; Llinas et al., 1994), myelin proteolipid protein (Yamaguchi et al., 1996) and guanylate cyclase (Suzuki et al., 2001). A protein kinase stimulated by InsP6 phosphorylates pacsin/syndapin I and thereby increases its association with dynamin I at nerve terminals (Hilton et al., 2001). InsP6 also binds to PH domains, although with less affinity than some of the other inositol phosphates (Kavran et al., 1998). In the hippocampus, cerebellum, cortex and striatum regions of rat brain, InsP6 levels were elevated upon activation and lowered by inhibition of neuronal activity (Yang et al., 2001). Specific functions of InsP6 in insulin-secreting cells are derived through the inhibition of phosphatases and the activation of protein kinase C. By inhibiting serine-threonine phosphatases, InsP6 may enhance phosphorylation of voltage-gated L-type calcium channels resulting in calcium influx over the plasma membrane (Larsson et al., 1997). This in turn leads to an increase in cytoplasmic free calcium and insulin release. Activation of protein kinase C by InsP6 may also lead to insulin secretion by promoting the recruitment and transport of granules to the site of exocytosis or by altering the conformation of proteins responsible for vesicle fusion (Efanov et al., 1997). Enhancement of calcium influx by InsP6 has also been observed in other cells and organelles such as cerebellar neurons and liver mitochondria (Nicoletti et al., 1989; Copani et al., 1991). Recently, InsP6 was observed to increase L-type calcium channel activity in hippocampal neurons by increasing the activity of adenylyl cyclase, which raised cyclic AMP levels, which in turn enhanced the activity of protein

INOSITOL PHOSPHATES IN FOODS

31

kinase A (PKA) (Yang et al., 2001). Thus L-type calcium channel activity in hippocampal neurons may be enhanced by InsP6 through increased phosphorylation at PKA phosphorylation sites of the channel, in addition to the inhibition of serine/threonine protein phosphatases. InsP6 is also required for the export of mRNA from the nucleus to the cytoplasm, where it can be translated into protein (York et al., 1999; Feng et al., 2001). It is possible that InsP6 functions by binding proteins associated with the nuclear pore complex or the shuttling heterogeneous nuclear ribonucleoprotein complexes. The three enzymes needed for synthesis of the InsP6 required for mRNA export in Saccharomyces cerevisiae were phospholipase C, Ins(1,4,5)P3 6-kinase/Ins(1,4,5,6)P4 3-kinase and Ins(1,3,4,5,6)P5 2-kinase (York et al., 1999; Saiardi et al., 2000a). The intermediate Ins(1,4,5,6)P4 was also shown in S. cerevisiae to be needed to regulate the transcription of six genes involved in the synthesis or breakdown of argenine (Odom et al., 2000). The requirement of InsP6 for efficient DNA repair of double-stranded breaks is an additional nuclear function that was recently discovered (Hanakahi et al., 2000). InsP6 is part of the nonhomologous end-joining apparatus that consists of the XRCC4/DNA ligase IV complex and DNAdependent protein kinase (DNA-PK), which is comprised of a catalytic subunit and the DNA end-binding protein Ku. InsP6 binds to DNA-PK, but the mechanism by which InsP6 promotes end-joining is unknown. Maximum activity was obtained with 1 µM InsP6, although Ins(1,3,4,5,6)P5 and Ins(1,3,4,5)P4 were also somewhat effective ligands. Ins(1,3,4,5,6)P5 and InsP6 can have phosphates enzymatically added onto one or two of their carbon-bound phosphates to form diphosphorylated inositol phosphates such as InsP7 and InsP8. These inositol pyrophosphates are found in animals, plants and microbes and thus far include [PP]2InsP4, PP-InsP5, [PP]2-InsP3 and PP-InsP4 (Yang et al., 1999; Saiardi et al., 2000b). InsP6, PP-InsP5, and [PP]2-InsP4 are synthesized by Ins(1,3,4,5,6)P5 2-kinase, InsP6 kinase and PP-InsP5 kinase, respectively, which can also catalyze their reverse reactions to produce ATP (Phillippy et al., 1994; Voglmaier et al., 1996; Huang et al., 1998). Thus InsP6 and the inositol pyrophosphates may function partly as cellular energy stores. InsP7 and/or InsP8 are involved in the homologous recombination mode of repair of double-stranded DNA breaks (Luo et al., 2002). The family of diphosphoinositol polyphosphates may regulate assembly of vesicles used for endocytosis and the trans-Golgi transport of proteins (Saiardi et al., 2000b), and InsP7 may also have a role in vesicle exocytosis (Luo et al., 2001). In addition, the fact that InsP6 kinase stimulates the uptake of inorganic phosphate may mean that InsP7 or InsP8 also help to regulate that process (Schell et al., 1999). An especially intriguing new finding is

32

B. Q. PHILLIPPY

that the post-transcriptional induction of an InsP6 kinase is largely responsible for the growth inhibition and apoptosis of human ovarian carcinoma cells following treatment with interferon-β (Morrison et al., 2001).

VII. NUTRITIONAL IMPORTANCE OF INOSITOL PHOSPHATES A. BIOAVAILABILITY OF MINERALS

The history of the inositol phosphates from a nutritional perspective has followed an uneven path. As described in the review by Reddy et al. (1989), the propensity of InsP6 to impair the absorption of minerals such as calcium and iron was recognized early in the twentieth century. By chelating and precipitating multivalent cationic minerals, InsP6 was found to lower the bioavailability of the macrominerals calcium and magnesium as well as trace minerals such as iron and zinc. Since the precipitated mineral complexes contained InsP6, the inositol and phosphate moieties of this compound would also be unavailable for absorption, although this has not been viewed as a significant concern for humans. A change in the paradigm started in the 1980s when some positive nutritional attributes of InsP6 were discovered. Presently, nutritionists are attempting to clarify the roles of inositol phosphates in the diet and to determine the most appropriate levels for consumption. There is general agreement that diets containing high levels of InsP6 can reduce the bioavailability of polyvalent cationic minerals (Zhou and Erdman, 1995; Rickard and Thompson, 1997; Harland and Narula, 1999). What is not clear is how much InsP6 is too much. As data on this subject have grown more extensive, it has become apparent that a simple rule cannot encompass all situations. Instead, special considerations must be made for different minerals, different groups of people and diets of different composition. Although phytate can probably chelate all polyvalent cationic minerals, some minerals have received more attention than others in rough correlation to their perceived nutritional importance. Zinc and iron have been of primary concern because they readily form poorly available insoluble complexes with InsP6 and are critical for growth and development (Zhou and Erdman, 1995). Calcium has also received considerable attention, but it is now recognized that InsP6 inhibits calcium absorption less than oxalic acid does (Heaney et al., 1991; Frossard et al., 2000). Interactions of InsP6 with magnesium, copper, selenium, manganese, cobalt, nickel, cadmium, lead, aluminum and mercury have also been studied to varying degrees

INOSITOL PHOSPHATES IN FOODS

33

and cannot be ignored when considering the nutritional importance of inositol phosphates. In vivo studies with Japanese quail, rats and humans have shown that moderate amounts of InsP6 and InsP5 decrease the absorption of zinc, whereas InsP4 and InsP3 have no effect (Tao et al., 1986; Lönnerdal et al., 1989; Sandström and Sandberg, 1992). In Caco-2 cells, the inhibition of the uptake and transport of zinc was directly proportional to the number of phosphate groups on InsP3, InsP4, InsP5 and InsP6 (Han et al., 1994). In vitro studies showing that InsP6 and InsP5 bind more zinc than InsP4 and InsP3 at an inositol phosphate to zinc molar ratio of 1 : 14 indicated that chelation and precipitation are most likely the mechanisms responsible for inhibiting zinc absorption (Persson et al., 1998). Simpson and Wise (1990) also found that zinc was more soluble in the presence of InsP3 and InsP4 than InsP5 and InsP6 at molar ratios of 1 : 1, but at higher ratios of inositol phosphate to zinc the highest percentage of soluble zinc was observed in the presence of InsP6. However, raising the molar ratio of calcium to zinc above 15 : 1 resulted in lower zinc solubilities in the presence of any of those inositol phosphates. In recent studies in rats and humans, respectively, whole wheat flour and oat bran, which have high levels of InsP6, decreased the fractional absorption of zinc in comparison with lowfiber diets but enhanced the total zinc absorption owing to the high zinc contents of those ingredients (Levrat-Verny et al., 1999; Sandström et al., 2000). In rats fermentation of fibers in the large intestine increased the absorption of various minerals by increasing their solubility via a lower pH and microbial production of phytase (Lopez et al., 1998, 2000a). Thus the high zinc content of whole grains may compensate for the negative effect of InsP6 on zinc absorption in some diets. Additional work is needed to determine whether these findings can be reproduced with other diets such as those containing legumes and the effects of different levels of calcium. In a study of iron absorption from bread in humans, the sum of InsP3, InsP4, InsP5 and InsP6 was related to the inhibition of absorption (Brune et al., 1992). In experiments performed at molar ratios of inositol phosphate to iron of 0.04 : 1 to 3.6 : 1, the solubility of iron following in vitro digestion was decreased by InsP6 and InsP5 but slightly increased by InsP4 and InsP3 (Sandberg et al., 1989). The reason for the apparent different responses in the above studies to InsP4 and InsP3 in humans and in vitro was recently discovered. While InsP3 and InsP4 added individually to wheat rolls had no effect on iron absorption in humans, they did increase the negative effect of small amounts of InsP5 and InsP6 when all four were added together (Sandberg et al., 1999). Thus InsP3 and InsP4 potentiated the negative effect of InsP5 and InsP6, presumably by binding some of the soluble iron and

34

B. Q. PHILLIPPY

cross-linking it to insoluble complexes formed by iron and InsP5 and InsP6, thereby increasing the portion of iron that was insoluble. In Caco2 cells 10 : 1 molar ratios of inositol phosphate to iron gave small differences in uptake of iron between InsP3, InsP4, InsP5 and InsP6, but the transport of iron across the cells was dramatically reduced in proportion to the number of phosphates (Han et al., 1994). Additional Caco-2 experiments with a 2 : 1 molar ratio of inositol phosphate to iron showed InsP6 and InsP5 inhibited iron uptake after 1 h, but various isomers of InsP4 and InsP3 only inhibited uptake after 4 h (Skoglund et al., 1999). Similar to zinc, whole wheat flour enhanced the total iron absorption in rats, but unlike with zinc, the per cent iron absorption was also increased relative to white wheat flour (Levrat-Verny et al., 1999). In an earlier rat study, Saha et al. (1994) found that the fractional absorption of iron and other minerals from whole wheat flour decreased with increasing phytate content, but was still high, 66% for iron, at the highest level of phytate. It would be helpful to do a similar study in humans. A consequence of the addition of dodecasodium phytate at a level of 1% to the diet of rats was a decrease in the iron concentration in their brains (Grases et al., 2001d). Degradation of InsP6 by phytase has been used to increase the bioavailability of iron from soybased infant formula and wheat bread (Davidsson et al., 1994; Sandberg et al., 1996). Several strategies have been used to lower the amount of InsP6 in seeds. Maize, barley, rice and soybeans with mutations in genes involved in InsP6 synthesis have been developed with levels of InsP6 reduced by about half or more (Larson and Raboy, 1999; Hatzack et al., 2000; Larson et al., 2000; Wilcox et al., 2000). When the low-phytic acid maize was used to make tortillas, iron absorption by men was 8.2% compared to 5.5% from tortillas made from wild-type maize (Mendoza et al., 1998). When porridge was prepared from the two types of maize, which had been fortified with additional iron, no effect on the iron absorption in women was observed (Mendoza et al., 2001). Because 92% of the activity of transgenic Aspergillus fumigatus phytase expressed in rice was lost after boiling the seeds for 20 min in water, phytase expression targeted to the endosperm to prevent the storage of phytic acid in the edible part of the seeds is currently under investigation (Lucca et al., 2001). Prospective work with an E. coli phytase introduced into Arabidopsis has produced transgenic seeds with reduced levels of phytate and increased free phosphate (Coello et al., 2001). Low-phytic acid grains and legumes have much anticipated potential for those who must satisfy their nutritional needs with a limited food intake and cannot supplement or fortify their diets with extrinsic minerals. A recent study unexpectedly found that variations in InsP5 plus InsP6 from 19.6 to 29.2 µmol/g in 24 genotypes of beans (Phaseolus

INOSITOL PHOSPHATES IN FOODS

35

vulgaris L.) with 52–157 µmol/g of endogenous iron had no effect on iron bioavailability in rats (Welch et al., 2000). However, the iron bioavailability from the beans was high, 53–76%, possibly due to unknown promoter substances and the rat intestinal phytase. Two studies have compared the effects of various inositol phosphates on the availability of calcium. The absorption of calcium in fasted rats was 83% from a solution containing InsP6, whereas InsP5, InsP4 and InsP3 resulted in absorption of more than 98% of the dose at an inositol phosphate to calcium molar ratio of 4 : 1 (Lönnerdal et al., 1989). In another rat study using purified inositol phosphate isomers, InsP6 decreased calcium absorption, while Ins(1,2,3,5,6)P5 and Ins(1,2,5,6)P4 had no effect (Shen et al., 1998). However, Ins(1,2,3,6)P4 significantly increased calcium absorption at an inositol phosphate to calcium molar ratio of 1 : 25. It is possible that a portion of the Ins(1,2,3,6)P4 entered cells and bound to receptors in the plasma membrane, thereby opening calcium channels, or this isomer may have formed a soluble calcium complex that was more readily absorbed. The bioavailability of calcium from grain and legume products such as whole wheat flour and soy flour is generally high (Mason et al., 1993; Saha et al., 1994), and components besides InsP6 in wheat bran and beans may be more inhibitory to its absorption (Weaver et al., 1993, 1996). In recent studies calcium transport across Caco-2 cell monolayers was reduced 16% by 2 mM InsP6 (Kennefick and Cashman, 2000), and the fractional calcium absorption was decreased 16% by a diet containing 7.5 mmol/kg InsP6 in rats (Harrington et al., 2001). Although the effects of different inositol phosphates on the bioavailability of polycationic minerals other than zinc, iron and calcium have not been investigated, it can be assumed that InsP6 and InsP5 will generally have the greatest negative impact. Interactions with other components of the diet, especially calcium and other minerals, also play an important role in determining availability. In vivo studies have continued to investigate concerns for the adverse effects of InsP6 on the absorption of selenium and magnesium (Saha et al., 1994; Pallauf et al., 1998; Rimbach and Pallauf, 1999). In 10 mM metal chloride and 62.5 µM InsP6 solutions at pH 6.0 and 7.0 more than half the Mg2+ precipitated within 2 h while all the Ca2+ remained in solution, but below pH 5.5 and 6.0, 10 mM Ca2+ and Mg2+, respectively, were 100% soluble at all concentrations of InsP6 from 62.5 µM to 20 mM (Nolan et al., 1987). In solutions of 1 mM InsP6 and 1–3 mM copper(II) or zinc(II) ions at pH 5.9, copper(II) ions were more soluble than zinc(II) ions (Champagne and Hinojosa, 1987), which explains why InsP6 can enhance the bioavailability of copper in the rat (Lee et al., 1988). However, soluble zinc to copper molar ratios increased as the total metal ion to InsP6 molar ratios increased from 2 : 1 to 12 : 1 (Champagne and Hinojosa, 1987), which

36

B. Q. PHILLIPPY

may help to explain why InsP6 has also been found to inhibit copper absorption in rats (Lopez et al., 1998). In pH 7.0 solutions containing 10 mM InsP6 and 1 mM metal ions, copper(II) remained soluble while zinc phytate slowly precipitated (Champagne and Fisher, 1990). Between pH 5 and pH 6, InsP6, InsP5, InsP4 and InsP3 can bind more copper than zinc, and the amount of metal bound is proportional to the number of phosphate groups (Persson et al., 1998). Additional mineral nutrients including manganese, cobalt and nickel are known to bind to InsP6 (Vohra et al., 1965), and others such as chromium, molybdenum and vanadium are likely to do so as well. Aluminum and the heavy metals also form complexes with inositol phosphates. Aluminum binds to the second messenger Ins(1,4,5)P3 more strongly than to ATP, but the biological significance of this is not clear (Mernissi-Arifi et al., 1995). Aluminum and lead formed insoluble complexes with InsP6 at metal to InsP6 molar ratios of 5 : 1 to 3 : 1, whereas mercury was completely soluble at all InsP6 concentrations (Bullock et al., 1995). At pH 5.0 in solutions containing 10 mM InsP6 and 10 mM metal ions the solubilities of aluminum and lead were 69% and 4%, respectively. Mercury competes with calcium in binding to InsP6 (Bullock et al., 1995) and forms complexes with Ins(1,2,6)P3 (Lapp and Speiss, 1991). Thus one must consider the possibility that inositol phosphates might enhance the absorption of mercury, as has been shown for InsP6 and copper. Cadmium binds to InsP6, InsP5, InsP4 and InsP3 less strongly than copper and zinc do at low pH values in the vicinity of pH 4 (Persson et al., 1998). In a recent study in rats, Aspergillus niger phytase addition to a maize-soybean diet to increase zinc absorption did not alter cadmium concentrations in the liver and kidney, but nonsignificant increases in femur lead were observed (Rimbach et al., 1998).

B. PREVENTION OF HEALTH DISORDERS

The backbone of most inositol phosphates in cells is myo-inositol. The nutritional importance of myo-inositol has long been recognized for its roles in the utilization of fat, as a growth promoter, and its ability to improve nerve conductance in diabetics (Holub, 1982, 1986). These functions may be partly or mostly derived from the use of inositol as a precursor of phosphatidylinositols and inositol phosphates. An extensive review of the metabolism of myo-inositol in plants was published recently (Loewus and Murthy, 2000). Inositol phosphates from seeds are a significant food source of myo-inositol, as are the phospholipids and free inositol from many plant- and animal-based foods (Berdanier, 1992). The

INOSITOL PHOSPHATES IN FOODS

37

total myo-inositol contents of the majority of fruits, vegetables, grains and nuts analyzed after digestion with 6 N HCl for 40 h at 120°C were between 0.2 and 2.0 mg/g (Clements and Darnell, 1980). Myo-inositol has been evaluated for its ability to improve the mental health of patients with various psychiatric disorders (Kofman et al., 1998; Seedat and Stein, 1999; Kofman et al., 2000; Einat and Belmaker, 2001; Nemets et al., 2001). In addition to myo-inositol, smaller amounts of epi- and scyllo-inositol are present in human brains (McLaurin et al., 2000). Phosphatidyl-scyllo-inositol appears to be synthesized more rapidly than phosphatidyl-myo-inositol in barley seeds (Carstensen et al., 1999), but little is known about the metabolism or function of scyllo-inositol in animals. D-Chiro-inositol, which may be of benefit to diabetics (Steadman et al., 2000), and myo-inositol levels in urine of older men and women, appear to be related to insulin secretion (Campbell et al., 2001). Myo-inositol and InsP6 have synergistic or additive effects in inhibiting the development of cancer (Shamsuddin, 1999). In mice, dietary myoinositol has been shown to be effective in preventing cancer of the colon (Shamsuddin et al., 1989), lung (Estensen and Wattenberg, 1993; Hecht et al., 1999; Wattenberg et al., 2000; Hecht et al., 2001), forestomach (Estensen and Wattenberg, 1993) and liver (Nishino et al., 1999). The anticancer action of InsP6 is extensively documented as reviewed by Shamsuddin (1995, 1999) and Jenab and Thompson (2002). In rats, mice or humans InsP6 has antitumor effects in cells or tissues of the blood, colon, liver, lung, mammary gland, prostate and skin. Although exogenous InsP6 and wheat bran containing a similar amount of InsP6 had similar effects on biomarkers of colon cancer risk in rats (Jenab and Thompson, 1998, 2000), the former was more effective in reducing the number of mammary tumors (Vucenik et al., 1997). While most studies on the anticancer effects of InsP6 have yielded positive results, a few contradictory reports have raised concerns regarding cancers of the urinary tract and rhabdomyosarcomas, which are muscle tumors occurring mainly in young people. Rats given drinking water containing 1.25% or 2.5% InsP6 ad lib for 2 years passed blood in the urine from hemorrhage of necrotic renal papillae and developed renal papillomas (Hiasa et al., 1992). In rats given a combination of three cancer initiators and fed diets with or without 2% InsP6 for 32 weeks, InsP6 increased the incidence of urinary bladder papillomas (Hirose et al., 1999). However, in rats given a single different initiator and treated similarly with InsP6 or its salts, InsP6 alone had no effect while the dodecasodium salt of InsP6 significantly increased the incidence of urinary bladder hyperplasias and papillomas (Hirose et al., 1999). It was concluded that the effect of InsP6 itself was equivocal, but the dodecasodium salt of InsP6 promoted carcino-

38

B. Q. PHILLIPPY

genesis. Alkaline salts are known to promote urinary bladder carcinogenesis by raising the urinary pH (Lina et al., 1994), which causes the formation of cytotoxic calcium phosphate precipitates (Cohen et al., 2000). In serum-free medium, micromolar levels of InsP6 stimulated the growth of two human rhabdomyosarcoma cell lines but inhibited the growth of a third rhabdomyosarcoma and two human colon carcinoma cell lines (Germain and Houghton, 1997). When rhabdomyosarcoma cells susceptible to growth inhibition by InsP6 were xenografted into nude mice, tumor size after 2 and 5 weeks was 25-fold and 49-fold smaller, respectively, in mice treated with InsP6 than in untreated controls (Vucenik et al., 1998). Additional studies to further clarify the potential risks associated with the consumption of InsP6 in regards to urinary tract cancers and rhabdomyosarcomas would be helpful. There are a variety of mechanisms by which inositol and InsP6 may inhibit the development of cancer. There is evidence that myo-inositol suppresses the phosphatidylinositol 3-kinase pathway and thereby protects against the inhibition by carcinogens of cell differentiation (Jyonouchi et al., 1999). The effect of myo-inositol within the cell is likely to be mediated through one or more of its phosphorylated metabolites. Signal transduction through Ins(1,4,5)P3 is also elevated in human carcinomas, and inhibitors of phosphatidylinositol 4-kinase and phosphatidylinositol 4-phosphate 5-kinase induce differentiation and apoptosis of cancer cells (Weber et al., 1999). Razzini et al. (2000) propose that Ins(1,4,5,6)P4 and Ins(1,3,4,5,6)P5, but not InsP6, inhibit the growth of various types of cancer cells by binding to pleckstrin homology (PH) domains of Akt (protein kinase B), which is activated by the lipid products of phosphatidylinositol 3-kinase. It has been suggested that InsP6 inhibits cell transformation into a cancerous phenotype by directly inhibiting phosphatidylinositol 3-kinase (Huang et al., 1997; Dong et al., 1999) or by inhibiting the phosphorylation of extracellular signal-related protein kinases (Erks), c-Jun NH2-terminal kinases (JNKs) or an inhibitor of nuclear factor κβ (Iκβ) (Chen et al., 2001). Proteins that become upregulated following InsP6 treatment include p53 in HT-29 colon carcinoma cells (Saied and Shamsuddin, 1998) and hepatic glutathione S-transferase in mice (Singh et al., 1997). Ornithine decarboxylase, which is essential for the promotion of some tumors, is downregulated by InsP6 in mouse keratinocytes (Nickel and Belury, 1999). Furthermore, InsP6 inhibits endocytosis by impairing the binding of erbB1 to AP2 in prostate cancer cells (Zi et al., 2000). Although it seems possible that some InsP6 may enter cells intact, convincing proof for this is lacking. An alternative mechanism through which extracellular InsP6 conceivably could combat cancer is through mineral deprivation; intracellular zinc deficiency leads to cell death by apoptosis (Truong-Tran et al., 2000). Iron chelators such

INOSITOL PHOSPHATES IN FOODS

39

as O-Trensox and desferrioxamine induce apoptosis in human hepatoblastoma HepG2 and hepatocarcinoma HBG cells, although addition of iron or zinc during treatment restores both proliferation and inhibition of apoptosis (Rakba et al., 2000). Another relevant observation is that InsP6 enhances natural killer cell activity (Baten et al., 1989). InsP6 is also a substrate for the InsP6 kinase that transduces the signal from interferon-β for the growth inhibition and apoptosis of ovarian carcinoma cells (Morrison et al., 2001). The antioxidant attributes of inositol phosphates may contribute to their anticancer activity as well as the prevention and amelioration of other conditions associated with excessive oxidation or inflammation. InsP6 chelates iron within its 1,2,3-trisphosphate grouping, thus preventing ironcatalyzed hydroxyl free radical formation (Hawkins et al., 1993). Ins(1,2,3)P3, D/L-Ins(1,2,3,4)P4, Ins(1,2,3,5)P4, D/L-Ins(1,2,3,4,5)P5 and Ins(1,2,3,4,6)P5 also possess this property (Hawkins et al., 1993; Spiers et al., 1995, 1996; Phillippy and Graf, 1997). InsP3, InsP4 and InsP5 fractions derived from InsP6 by hydrolysis with microbial phytase prevent the iron-catalyzed decomposition of lipid peroxides, which liberates peroxyl and/or alkoxyl radicals, whereas InsP2 has no effect (Miyamoto et al., 2000). Studies have shown that dietary InsP6 reduces lipid peroxide formation in the liver of lactating mice and the colon of pigs with high iron intake (Singh et al., 1997; Porres et al., 1999). In rats subjected to oxidative stress, dietary InsP6 and Ins(1,2,3,6)P4 decreased the production of lipid peroxides in the small intestine and colon, whereas only InsP6 gave an antioxidative response in the lung (Burgess and Gao, 2002). However, InsP6 had no effect on lipid peroxides, protein oxidation, α-tocopherol or reduced glutathione in the liver of growing rats (Rimbach and Pallauf, 1998). In HL-60 human leukemia cells and calf thymus DNA exposed to H2O2, InsP6 reduced the formation of 8-oxo-7,8-dihydro-2′-deoxyguanosine, a biomarker of oxidative DNA damage; and cleavage at the 5′-guanine of GG and GGG sequences in c-Ha-ras-1 and p53 gene DNA fragments in the presence of H2O2 and copper was decreased by InsP6 (Midorikawa et al., 2001). In vitro InsP6 inhibited the oxidation of 6-hydroxydopamine by iron or manganese, accelerated catalysis by vanadium and had no effect in the presence of copper (Bandy et al., 2001). Formation of the p-quinone oxidation product proceeded most rapidly when the reduction potential of various metal-ligand complexes fell between the reduction potentials of 6-hydroxydopamine and molecular oxygen. Copper was the most effective catalyst, and the rate of 6-hydroxydopamine oxidation by copper phytate was increased three-fold in the presence of 100 mM Na2SO4. The effectiveness of InsP6 as an antioxidant food preservative has been demonstrated repeatedly (Empson et al., 1991; Lee and Hendricks, 1995;

40

B. Q. PHILLIPPY

Hix et al., 1997; Lee et al., 1998; Cornforth, 2002). Unlike iron, copper appears to bind preferentially to the 5-phosphate of InsP6 (Champagne et al., 1990), but the ability of copper to produce hydroxyl free radicals is unhindered by InsP6 (Madurawe et al., 1997). In contrast to InsP6, the common food additive ethylenediaminetetraacetic acid (EDTA) stimulates iron-catalyzed but inhibits copper-catalyzed hydroxyl radical formation (Madurawe et al., 1997). Ins(1,2,6)P3, which is the drug α-trinositol and is produced by many plant and microbial phytases, exhibits analgesic and anti-inflammatory properties. It has been speculated that Ins(1,2,6)P3 may function in part by binding to one or more of the proteins in the phosphatidylinositol signaling pathway (Bell and McDermott, 1998), but its effectiveness has only been reported for parenteral systemic administration and topical treatment of burned skin (Tarnow et al., 1998). It might be revealing to investigate the antioxidant and anti-inflammatory potencies of Ins(1,2,3)P3 and Ins(1,2,6)P3 in some model feeding studies. InsP6 lowers blood glucose, cholesterol and triglycerides (Rickard and Thompson, 1997; Jariwalla, 1999). This would be especially helpful for people susceptible to diabetes or heart disease. The mechanism for these effects is not completely clear but may be related to the inhibition of digestive enzymes by InsP6 and/or a reduction in the plasma ratio of zinc to copper. Myo-inositol and InsP6 were equally effective in preventing the elevation in liver lipids following sucrose feeding in rats (Katayama, 1999). It was proposed that both compounds worked by depressing hepatic lipogenesis rather than by inhibiting intestinal enzymes. Nonetheless, the old saying that a meal that keeps a person from getting hungry for a long time ‘sticks to the ribs’ may refer to foods high in InsP6, which probably slows digestion by adhering to enzymes and other proteins in the lumen of the stomach and on the surface of the gastric mucosa. Indeed, α-amylase and lipase activities were significantly inhibited in vitro by 2–4 mM inositol phosphates containing one to six phosphates, and the decrease in digestibility and degree of inositol phosphorylation were highly correlated (Knuckles and Betschart, 1987; Knuckles, 1988). Various other health benefits from InsP6 consumption have been postulated. Some of those who eat diets high in red meat may accumulate excess iron, which can promote the development of infections, cancer and other degenerative diseases, and foods rich in InsP6 may help prevent the overaccumulation of absorbed iron (Weinberg, 1999). This is a somewhat complex issue, especially in light of the recent observation that foods containing InsP6 may lead to increased iron absorption simply because they increase the amount of iron consumed (Levrat-Verny et al., 1999). In some instances the iron from foods naturally high in InsP6 may be less

INOSITOL PHOSPHATES IN FOODS

41

available on a per cent basis than the iron in meat and thus less likely to be absorbed in overabundance. There has been an accumulation of evidence that dietary InsP6 helps to prevent the formation of kidney stones (Zhou and Erdman, 1995). It has been suggested that InsP6 excreted in the urine is responsible for preventing stone formation by inhibiting crystallization of calcium salts (Grases and Costa-Bauzá, 1999; Grases et al., 2000). However, 12 h following intragastric administration of [3H]InsP6 to rats the radioactivity in the urine appeared to be associated with Ins and InsP but not with any of the inositol polyphosphates (Sakamoto et al., 1993). Thus any intact absorption and subsequent excretion of InsP6 must be very low. InsP6 has also been linked to upregulation of neutrophil functions (Eggleton, 1999) and the inhibition of platelet aggregation (Vucenik et al., 1999).

VIII. SUMMARY AND CONCLUSIONS

Advances in the analytical methods for inositol phosphates have enabled an increase in our knowledge of their nutritional roles in recent years. HPLC methods provide the separations needed to identify and quantify individual inositol phosphates in foods. The metabolic pathways for the synthesis and degradation of InsP6 are multibranched and dependent upon the particular mix of enzymes and substrates present in the same cellular compartment. InsP6 is the most abundant inositol phosphate in the raw seeds of most grains and legumes and generally is present at concentrations between 0.4 and 1.2% of the dry weight. InsP6 and InsP5 account for more than 95% of the total inositol phosphates in most raw grains and legumes and predominate in processed foods, which sometimes also contain substantial levels of InsP4 and InsP3. Avocado fruit and some vegetables contain appreciable amounts of InsP6, but very little data is available in this area. Inositol phosphates appear to be mostly hydrolyzed to inositol prior to absorption in the guts of animals. Numerous inositol phosphate isomers in animal cells display an increasingly diverse range of biological functions. The fractional absorption of dietary minerals such as zinc and iron is decreased by InsP6 and InsP5, and these effects might be potentiated by InsP4 and InsP3. Myo-inositol and InsP6 may help to prevent various health disorders by a number of possible mechanisms. More research is needed before the optimum levels of InsP6 in the diets of people differing in age, sex and health concerns can be estimated.

42

B. Q. PHILLIPPY IX. FUTURE RESEARCH NEEDS

The most immediate research need is to determine the appropriate levels of InsP6 in human diets. Some InsP6 may be desirable for its potential ability to prevent or delay various health disorders, but too much can result in mineral deficiencies. It may be helpful to establish a tolerance zone or range, which lies between the optimal and toxic doses, as has been suggested for essential trace elements such as selenium and chromium (Katz, 1996). This could be accomplished by feeding studies in which the effects of InsP6 on indices of potential health benefits and on the bioavailabilities of minerals are monitored simultaneously. Since the tolerance zone likely depends on factors such as age, sex, health status and diet composition, experiments performed on different population groups would be necessary in order to be able to estimate the appropriate level of InsP6 in someone’s diet. It will be important to identify trade-offs between the positive and negative nutritional aspects of InsP6 if overlapping effects are observed and to attempt to resolve any potential conflicts. Then we will have a better idea of how much InsP6 it would be prudent to remove from our foods by breeding or during processing. More accurate data on the inositol phosphates in foods are needed. Since a number of inositol phosphates are bioactive and may be absorbed intact by cells under certain conditions, more data on their concentrations in foods should be obtained. The inositol phosphates and their natural variation in fruits and vegetables need to be analyzed, since comprehensive and accurate data for these foods is lacking. It might be a good idea to replace the nonspecific Association of Official Analytical Chemists’ method for phytate analysis with one of the HPLC methods if there is sufficient interest among researchers. Biological studies are needed to clear up some lingering questions about the fate of inositol phosphates and their associated minerals within the gut. The possibility of adaptation to InsP6 in the human diet needs to be investigated more thoroughly. Recent reports showing greater overall absorption of zinc and iron from diets containing high-phytate foods should be followed up with more definitive studies. The potential for absorption of inositol phosphates complexed with food components should be evaluated. It is not currently known to what extent any of the bioactive inositol phosphates consumed in foods can be absorbed before they are enzymatically degraded or whether there may be significant population group differences in inositol phosphate absorption. Potential health benefits of myo-inositol and inositol phosphates in the diet have been identified. More research is needed to define their mechanisms of action and whether the effects are mediated predominately

INOSITOL PHOSPHATES IN FOODS

43

within the lumen of the gut or following absorption, on the surface of cells or intracellularly.

DISCLAIMER

Mention of names of companies or commercial products is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the United States Department of Agriculture over others not mentioned.

ACKNOWLEDGEMENT

The author thanks Dr John M. Bland for the computer-generated chemical structures. REFERENCES Adelt, S., Plettenburg, O., Dallmann, G., Ritter, F. P., Shears, S. B., Altenbach, H.-J. and Vogel, G. 2001. Regiospecific phosphohydrolases from Dictyostelium as tools for the chemoenzymatic synthesis of the enantiomers D-myo-inositol 1,2,4-trisphosphate and D-myo-inositol 2,3,6-trisphosphate: non-physiological, potential analogues of biologically active D-myo-inositol 1,3,4-trisphosphate. Bioorg. Med. Chem. Lett. 11, 2705–2708. Agte, V. V., Tarwadi, K. V. and Chiplonkar, S. A. 1999. Phytate degradation during traditional cooking: significance of the phytic acid profile in cereal-based vegetarian meals. J. Food Compos. Anal. 12, 161–167. Ali, N., Craxton, A., Sumner, M. and Shears, S. B. 1995. Effects of aluminum on the hepatic inositol polyphosphate phosphatase. Biochem. J. 305, 557–561. Asada, K., Tanaka, K. and Kasai, Z. 1969. Formation of phytic acid in cereal grains. Ann. N. Y. Acad. Sci. 165, 801–814. Ayet, G., Burbano, C., Cuadrado, C., Pedrosa, M. M., Robredo, L. M., Muzquiz, M., de la Cuadra, C., Castaño, A. and Osagie, A. 1997. Effect of germination, under different environmental conditions, on saponins, phytic acid and tannins in lentils (Lens culinaris). J. Sci. Food Agric. 74, 273–279. Ballou, C. E. and Pizer, L. I. 1959. Synthesis of an optically active myo-inositol 1-phosphate. J. Am. Chem. Soc. 81, 4745. Bandy, B., Walter, P. B., Moon, J. and Davison, A. J. 2001. Reaction of oxygen with 6-hydroxydopamine catalyzed by Cu, Fe, Mn, and V complexes: identification of a thermodynamic window for effective metal catalysis. Arch. Biochem. Biophys. 389, 22–30. Barker, C. J., French, P. J., Moore, A. J., Nilsson, T., Berggren, P. O., Bunce, C. M., Kirk, C. J. and Michell, R. H. 1995. Inositol 1,2,3-trisphosphate and inositol 1,2- and/or 2,3-bisphosphate are normal constituents of mammalian cells. Biochem. J. 306, 557–564. Barrientos, L., Scott, J. J. and Murthy, P. P. N. 1994. Specificity of hydrolysis of phytic acid by alkaline phytase from lily pollen. Plant Physiol. 106, 1489–1495.

44

B. Q. PHILLIPPY

Baten, A., Ullah, A., Tomazic, V. J. and Shamsuddin, A. M. 1989. Inositol-phosphate-induced enhancement of natural killer cell activity correlates with tumor suppression. Carcinogenesis 10, 1595–1598. Bell, D. and McDermott, B. J. 1998. D-myo inositol 1,2,6, triphosphate (alpha-trinositol, pp56): selective antagonist at neuropeptide Y (NPY) Y-receptors or selective inhibitor of phosphatidylinositol cell signaling? Gen. Pharmacol. 31, 689–696. Beno-Moualem, M. D., Naveh, L. and Jacoby, B. 1995. Responses of red beet tissue to hypertonic salt-shock: Inositol 1,4,5-trisphosphate, ATPase activation and protein phosphorylation. Plant Physiol. Biochem. 33, 311–318. Berdanier, C. D. 1992. Is inositol an essential nutrient? Nutr. Today 27, 22–26. Bergman, E. L., Fredlund, K., Reinikainen, P. and Sandberg, A. S. 1999. Hydrothermal processing of barley (cv. Blenheim): optimisation of phytate degradation and increase of free myo-inositol. J. Cereal Sci. 29, 261–272. Berka, R. M., Rey, M. W., Brown, K. M., Byun, T. and Klotz, A. V. 1998. Molecular characterization and expression of a phytase gene from the thermophilic fungus Thermomyces lanuginosus. Appl. Environ. Microbiol. 64, 4423–4427. Berridge, M. J. 1993. Inositol trisphosphate and calcium signalling. Nature 361, 315–325. Bos, K. D., Verbeek, C., van Eden, C. H. P., Slump, P. and Wolters, M. G. E. 1991. Improved determination of phytate by ion-exchange chromatography. J. Agric. Food Chem. 39, 1770–1772. Boucher, L., Chassaigne, M. and Ropars, C. 1996. Internalization and distribution of inositol hexakisphosphate in red blood cells. Biotechnol. Appl. Biochem. 24, 73–78. Brearley, C. A. and Hanke, D. E. 1996. Metabolic evidence for the order of addition of individual phosphate esters to the myo-inositol moiety of inositol hexakisphosphate in the duckweed Spirodela polyrhiza L. Biochem. J. 314, 227–233. Brearley, C. A. and Hanke, D. E. 2000. Metabolic relations of inositol 3,4,5,6-tetrakisphosphate revealed by cell permeabilization. Identification of inositol 3,4,5,6-tetrakisphosphate 1-kinase and inositol 3,4,5,6-tetrakisphosphate phosphatase activities in mesophyll cells. Plant Physiol. 122, 1209–1216. Brooks, S. P. J. and Lampi, B. J. 2001. Problems associated with measuring phytate in infant cereals. J. Agric. Food Chem. 49, 564–569. Brune, M., Rossander, H. L., Hallberg, L., Gleerup, A. and Sandberg, A. -S. 1992. Iron absorption from bread in humans: Inhibiting effects of cereal fiber, phytate and inositol phosphates with different number of phosphate groups. J. Nutr. 122, 442–449. Bullock, J. I., Duffin, P. A., Nolan, K. B. and Smith, T. K. 1995. Effect of phytate on the in-vitro solubility of Al3+, Ca2+, Hg2+ and Pb2+ as a function of pH at 37°C. J. Sci. Food Agric. 67, 507–509. Bunce, C. M., French, P. J., Allen, P., Mountford, J. C., Moor, B., Greaves, M. F., Michell, R. H. and Brown, G. 1993. Comparison of the levels of inositol metabolites in transformed haemopoietic cells and their normal counterparts. Biochem. J. 289, 667–673. Burbano, C., Muzquiz, M., Osagie, A., Ayet, G. and Cuadrado, C. 1995. Determination of phytate and lower inositol phosphates in Spanish legumes by HPLC methodology. Food Chem. 52, 321–325. Burford, N. T., Nahorski, S. R., Chung, S. K., Chang, Y. T. and Wilcox, R. A. 1997. Binding and activity of the nine possible regioisomers of myo-inositol tetrakisphosphate at the inositol 1,4,5-trisphosphate receptor. Cell Calcium 21, 301–310. Burgess, J. R., and Gao, F. 2002. The antioxidant effects of inositol phosphates. In “Food Phytates” (Reddy, N. R. and Sathe, S. K., eds.), pp. 189–197. CRC Press, Boca Raton, FL. Caffrey, J. J., Darden, T., Wenk, M. R. and Shears, S. B. 2001. Expanding coincident signaling by PTEN through its inositol 1,3,4,5,6-pentakisphosphate 3-phosphatase activity. FEBS Lett. 499, 6–10.

INOSITOL PHOSPHATES IN FOODS

45

Campbell, W. W., Ostlund, R. E., Joseph, L. J., Farrell, P. A. and Evans, W. J. 2001. Relationships of plasma C-peptide and gender to the urinary excretion of inositols in older people. Horm. Metab. Res. 33, 44–51. Carlsson, N. -G., Bergman, E. -L., Skoglund, E., Hasselblad, K. and Sandberg, A. -S. 2001. Rapid analysis of inositol phosphates. J. Agric. Food Chem. 49, 1695–1701. Carstensen, S., Pliska Matyshak, G., Bhuvarahamurthy, N., Robbins, K. M. and Murthy, P. P. N. 1999. Biosynthesis and localization of phosphatidyl-scyllo-inositol in barley aleurone cells. Lipids 34, 67–73. Cecconi, O., Nelson, R. M., Roberts, W. G., Hanasaki, K., Mannori, G., Schultz, C., Ulich, T. R., Aruffo, A. and Bevilacqua, M. P. 1994. Inositol polyanions: Noncarbohydrate inhibitors of L- and P-selectin that block inflammation. J. Biol. Chem. 269, 15060–15066. Champagne, E. T. and Fisher, M. S. 1990. Binding differences of Zn(II) and Cu(II) ions with phytate. J. Inorg. Biochem. 38, 217–223. Champagne, E. T. and Hinojosa, O. 1987. Independent and mutual interactions of copper(II) and zinc(II) ions with phytic acid. J. Inorg. Biochem. 30, 15–33. Champagne, E. T., Fisher, M. S. and Hinojosa, O. 1990. NMR and ESR studies of interactions among divalent cations, phytic acid, and N-acetyl-amino acids. J. Inorg. Biochem. 38, 199–215. Chattaway, J. A., Drobak, B. K., Watkins, P. A. C., Dawson, A. P., Letcher, A. J., Stephens, L. R. and Irvine, R. F. 1992. An inositol 1,4,5-trisphosphate-6-kinase activity in pea roots. Planta 187, 542–545. Chen, N., Ma, W.-Y. and Dong, Z. 2001. Inositol hexaphosphate inhibits ultraviolet B-induced signal transduction. Mol. Carcinog. 31, 139–144. Cilliers, J. J. L. and van Niekert, P. J. 1986. LC determination of phytic acid in food by postcolumn colorimetric detection. J. Agric. Food Chem. 34, 680–683. Clements, R. S., Jr. and Darnell, B. 1980. Myo-inositol content of common foods: development of a high-myo-inositol diet. Am. J. Clin. Nutr. 33, 1954–1967. Coello, P., Maughan, J. P., Mendoza, A., Philip. R., Bollinger, D. W., Veum, T. L., Vodkin, L. O. and Polacco, J. C. 2001. Generation of low phytic acid Arabidopsis seeds expressing an E. coli phytase during embryo development. Seed Sci. Res. 11, 285–291. Cohen, S. M., Arnold, L. L., Cano, M., Ito, M., Garland, E. M. and Shaw, R. A. 2000. Calcium phosphate-containing precipitate and the carcinogenicity of sodium salts in rats. Carcinogenesis 21, 783–792. Copani, A., Raciti, G., Bruno, V., Sortino, M. A., Nicoletti, F., Canonico, P. L. and Cambria, A. 1991. Inositolhexakisphosphate stimulates calcium-45 influx in purified mitochondria from rat liver. Ital. J. Biochem. 40, 289–294. Cornforth, D. P. 2002. Potential use of phytate as an antioxidant in cooked meats. In “Food Phytates” (Reddy, N. R. and Sathe, S. K., eds), pp. 199–209. CRC Press, Boca Raton, FL. Cosgrove, D. J. and Irving, G. C. J. 1980. “Inositol Phosphates: Their Chemistry, Biochemistry, and Physiology. Studies in Organic Chemistry; 4.” Elsevier Scientific Pub. Co., New York. Craxton, A., Erneux, C. and Shears, S. B. 1994. Inositol 1,4,5,6-tetrakisphosphate is phosphorylated in rat liver by a 3-kinase that is distinct from inositol 1,4,5-trisphosphate 3-kinase. J. Biol. Chem. 269, 4337–4342. Davidsson, L., Galan, P., Kastenmayer, P., Cherouvrier, F., Juillerat, M. A., Hercberg, S. and Hurrell, R. F. 1994. Iron bioavailability studies in infants: the influence of phytic acid and ascorbic acid in infant formulas based on soy isolate. Pediatric Res. 36, 816–822. de la Cuadra, C., Muzquiz, M., Burbano, C., Ayet, G., Calvo, R., Osagie, A. and Cuadrado, C. 1994. Alkaloid, α-galactoside and phytic acid changes in germinating lupin seeds. J. Sci. Food Agric. 66, 357–364. Delisle, S., Radenberg, T., Wintermantel, M. R., Tietz, C., Parys, J. B., Pittet, D., Welsh, M. J. and Mayr, G. W. 1994. Second messenger specificity of the inositol trisphosphate

46

B. Q. PHILLIPPY

receptor: Reappraisal based on novel inositol phosphates. Am. J. Physiol. 266, C429–C436. Dong, Z., Huang, C. and Ma, W. Y. 1999. PI-3 kinase in signal transduction, cell transformation, and as a target for chemoprevention of cancer. Anticancer Res. 19, 3743–3747. Efanov, A. M., Zaitsev, S. V. and Berggren, P. O. 1997. Inositol hexakisphosphate stimulates non-Ca2+ mediated and primes Ca2+ mediated exocytosis of insulin by activation of protein kinase C. Proc. Natl Acad. Sci. USA 94, 4435–4439. Eggleton, P. 1999. Effect of IP6 on human neutrophil cytokine production and cell morphology. Anticancer Res. 19, 3711–3715. Einat, H. and Belmaker, R. H. 2001. The effects of inositol treatment in animal models of psychiatric disorders. J. Affective Disord. 62, 113–121. Empson, K. L., Labuza, T. P. and Graf, E. 1991. Phytic acid as a food antioxidant. J. Food Sci. 56, 560–563. Estensen, R. D. and Wattenberg, L. W. 1993. Studies of chemopreventive effects of myo-inositol on benzo(a)pyrene-induced neoplasia of the lung and forestomach of female A/J mice. Carcinogenesis 14, 1975–1977. Feng, Y., Wente, S. R. and Majerus, P. W. 2001. Overexpression of the inositol phosphatase SopB in human 293 cells stimulates cellular chloride influx and inhibits nuclear mRNA export. Proc. Natl Acad. Sci. USA 98, 875–879. Ferguson, E. L., Gibson, R. S., Opare, O. C., Osei, O. F., Stephen, A. M., Lehrfeld, J. and Thompson, L. U. 1993. The zinc, calcium, copper, manganese, nonstarch polysaccharide and phytate content of seventy-eight locally grown and prepared African foods. J. Food Compos. Anal. 6, 87–99. Fleischer, B., Xie, J., Mayrleitner, M., Shears, S. B., Palmer, D. J. and Fleischer, S. 1994. Golgi coatomer binds, and forms K+-selective channels gated by, inositol polyphosphates. J. Biol. Chem. 269, 17826–17832. Fredlund, K., Asp, N. G., Larsson, M., Marklinder, I. and Sandberg, A. -S. 1997. Phytate reduction in whole grains of wheat, rye, barley and oats after hydrothermal treatment. J. Cereal Sci. 25, 83–91. Fredrikson, M., Alminger, M. L., Carlsson, N. -G. and Sandberg, A. -S. 2001a. Phytate content and phytate degradation by endogenous phytase in pea (Pisum sativum). J. Sci. Food Agric. 81, 1139–1144. Fredrikson, M., Biot, P., Alminger, M. L., Carlsson, N. -G. and Sandberg, A. -S. 2001b. Production process for high-quality pea-protein isolate with low content of oligosaccharides and phytate. J. Agric. Food Chem. 49, 1208–1212. Freund, W. D., Mayr, G. W., Tietz, C. and Schultz, J. E. 1992. Metabolism of inositol phosphates in the protozoan Paramecium: Characterization of a novel inositol-hexakisphosphatedephosphorylating enzyme. Eur. J. Biochem. 207, 359–367. Frossard, E., Bucher, M., Maechler, F., Mozafar, A. and Hurrell, R. 2000. Potential for increasing the content and bioavailability of Fe, Zn and Ca in plants for human nutrition. J. Sci. Food Agric. 80, 861–879. Fukuda, M., Aruga, J., Niinobe, M., Aimoto, S. and Mikoshiba, K. 1994. Inositol-1,3,4,5tetrakisphosphate binding to C2B domain of IP4BP/synaptotagmin II. J. Biol. Chem. 269, 29206–29211. Gaidarov, I., Chen, Q., Falck, J. R., Reddy, K. K. and Keen, J. H. 1996. A functional phosphatidylinositol 3,4,5-trisphosphate/phosphoinositide binding domain in the clathrin adaptor AP-2 alpha subunit: Implications for the endocytic pathway. J. Biol. Chem. 271, 20922–20929. Germain, G. S. and Houghton, P. J. 1997. Phytic acid stimulates the growth of a human rhabdomyosarcoma. In Vitro Cell. Dev. Biol. Animal 33, 581–583.

INOSITOL PHOSPHATES IN FOODS

47

Gersonde, K. and Ganguly, T. 1986. Inositol phosphates as modulators of oxygen affinity in hemoglobin. In “Phytic Acid: Chemistry & Applications” (E. Graf, ed.), pp. 195–248. Pilatus Press, Minneapolis, MN. Graf, E. and Dintzis, F. R. 1982. Determination of phytic acid in foods by high-performance liquid chromatography. J. Agric. Food Chem. 30, 1094–1097. Graf, E. and Eaton, J. W. 1990. Antioxidant functions of phytic acid. Free Radic. Biol. Med. 8, 61–69. Graf, E., Empson, K. L. and Eaton, J. W. 1987. Phytic acid. A natural antioxidant. J. Biol. Chem. 262, 11647–11650. Grases, F. and Costa-Bauzá, A. 1999. Phytate (IP6) is a powerful agent for preventing calcifications in biological fluids: Usefulness in renal lithiasis treatment. Anticancer Res. 19, 3717–3722. Grases, F., March, J. G., Prieto, R. M., Simonet, B. M., Costa, B. A., Garcia, R. A. and Conte, A. 2000. Urinary phytate in calcium oxalate stone formers and healthy people: Dietary effects on phytate excretion. Scand. J. Urol. Nephrol. 34, 162–164. Grases, F., Simonet, B. M., Prieto, R. M. and March, J. G. 2001a. Phytate levels in diverse rat tissues: influence of dietary phytate. Brit. J. Nutr. 86, 225–231. Grases, F., Simonet, B. M., Prieto, R. M. and March, J. G. 2001b. Variation of InsP4, InsP5 and InsP6 levels in tissues and biological fluids depending on dietary phytate. J. Nutr. Biochem. 12, 595–601. Grases, F., Simonet, B. M., Vucenik, I., Prieto, R. M., Costa-Bauzá, A., March, J. G. and Shamsuddin, A. M. 2001c. Absorption and excretion of orally administered inositol hexaphosphate (IP6 or phytate) in humans. BioFactors 15, 53–61. Grases, F., Simonet, B. M., Prieto, R. M. and March J. G. 2001d. Dietary phytate and mineral bioavailability. J. Trace Elem. Med. Biol. 15, 221–228. Greiner, R. and Alminger, M. L. 1999. Purification and characterization of a phytate degrading enzyme from germinated oat (Avena sativa). J. Sci. Food Agric. 79, 1453–1460. Greiner, R. and Alminger, M. L. 2001. Stereospecificity of myo-inositol hexakisphosphate dephosphorylation by phytate-degrading enzymes of cereals. J. Food Biochem. 25, 229–248. Greiner, R. and Konietzny, U. 1998. Endogenous phytate-degrading enzymes are responsible for phytate reduction while preparing beans (Phaseolus vulgaris). J. Food Process. Preserv. 22, 321–331. Greiner, R. and Konietzny, U. 1999. Improving enzymatic reduction of myo-inositol phosphates with inhibitory effects on mineral absorption in black beans (Phaseolus vulgaris var. Preto). J. Food Process. Preserv. 23, 249–261. Greiner, R., Konietzny, U. and Jany, K. D. 1993. Purification and characterization of two phytases from Escherichia coli. Arch. Biochem. Biophys. 303, 107–113. Greiner, R., Haller, E., Konietzny, U. and Jany, K. D. 1997. Purification and characterization of a phytase from Klebsiella terrigena. Arch. Biochem. Biophys. 341, 201–206. Greiner, R., Konietzny, U. and Jany, K. D. 1998. Purification and properties of a phytase from rye. J. Food Biochem. 22, 143–161. Greiner, R., Carlsson, N. and Alminger, M. L. 2000a. Stereospecificity of myo-inositol hexakisphosphate dephosphorylation by a phytate-degrading enzyme of Escherichia coli. J. Biotechnol. 84, 53–62. Greiner, R., Jany, K. D. and Larsson, A. M. 2000b. Identification and purification of myoinositol hexakisphosphate phosphohydrolases (phytases) from barley (Hordeum vulgare). J. Cereal Sci. 31, 127–139. Greiner, R., Alminger, M. L. and Carlsson, N. -G. 2001a. Stereospecificity of myo-inositol hexakisphosphate dephosphorylation by a phytate-degrading enzyme of baker’s yeast. J. Agric. Food Chem. 49, 2228–2233.

48

B. Q. PHILLIPPY

Greiner, R., Muzquiz, M., Burbano, C., Cuadrado, C., Pedrosa, M. M. and Goyoaga, C. 2001b. Purification and characterization of a phytate-degrading enzyme from germinated faba beans (Vicia faba var. Alameda). J. Agric. Food Chem. 49, 2234–2240. Guse, A. H., Greiner, E., Emmrich, F. and Brand, K. 1993. Mass changes of inositol-1,3,4,5,6pentakisphosphate and inositol hexakisphosphate during cell cycle progression in rat thymocytes. J. Biol. Chem. 268, 7129–7133. Gustafsson, E. L. and Sandberg, A. -S. 1995. Phytate reduction in brown beans (Phaseolus vulgaris L.). J. Food Sci. 60, 149–152. Han, O., Failla, M. L., Hill, A. D., Morris, E. R. and Smith, J. C., Jr. 1994. Inositol phosphates inhibit uptake and transport of iron and zinc by a human intestinal cell line. J. Nutr. 124, 580–587. Hanakahi, L. A., Bartlet-Jones, M., Chappell, C., Pappin, D. and West, S. C. 2000. Binding of inositol phosphate to DNA-PK and stimulation of double-strand break repair. Cell 102, 721–729. Harland, B. F. and Harland, J. 1980. Fermentative reduction of phytate in rye, white, and whole wheat breads. Cereal Chem. 57, 226–229. Harland, B. F. and Morris, E. R. 1995. Phytate: A good or a bad food component? Nutr. Res. 15, 733–754. Harland, B. F. and Narula, G. 1999. Food phytate and its hydrolysis products. Nutr. Res. 19, 947–961. Harrington, M. E., Flynn, A. and Cashman, K. D. 2001. Effects of dietary fibre extracts on calcium absorption in the rat. Food Chem. 73, 263–269. Hatzack, F., Johansen, K. S. and Rasmussen, S. K. 2000. Nutritionally relevant parameters in low-phytate barley (Hordeum vulgare L.) grain mutants. J. Agric. Food Chem. 48, 6074–6080. Hatzack, F., Hübel, F., Zhang, W., Hansen, P. E. and Rasmussen, S. K. 2001. Inositol phosphates from barley low-phytate grain mutants analyzed by metal-dye detection HPLC and NMR. Biochem. J. 354, 473–480. Hawkins, P. T., Poyner, D. R., Jackson, T. R., Letcher, A. J., Lander, D. A. and Irvine, R. F. 1993. Inhibition of iron-catalysed hydroxyl radical formation by inositol polyphosphates: A possible physiological function for myo-inositol hexakisphosphate. Biochem. J. 294, 929–934. Hayakawa, T., Toma, Y. and Igaue, I. 1989. Purification and characterization of acid phosphatases with or without phytase activity from rice bran. Agric. Biol. Chem. 53, 1475–1483. Heaney, R. P., Weaver, C. M. and Fitzsimmons, M. C. 1991. Soybean phytate content: Effect on calcium absorption. Am. J. Clin. Nutr. 53, 745–747. Hecht, S. S., Kenney, P. M. J., Wang, M., Trushin, N., Agarwal, S., Rao, A. V. and Upadhyaya, P. 1999. Evaluation of butylated hydroxyanisole, myo-inositol, curcumin, esculetin, resveratrol and lycopene as inhibitors of benzo(a)pyrene plus 4-(methylnitrosamino)-1-(3pyridyl)-1-butanone-induced lung tumorigenesis in A/J mice. Cancer Lett. 137, 123–130. Hecht, S. S., Kenney, P. M. J., Wang, M. Y. and Upadhyaya, P. 2001. Dose-response study of myo-inositol as an inhibitor of lung tumorigenesis induced in A/J mice by benzo(a)pyrene and 4-methylnitrosamino)-1-(3-pyridyl)-1-butanone. Cancer Lett. 167, 1–6. Hermosura, M. C., Takeuchi, H., Fleig, A., Riley, A. M., Potter, B. V. L., Hirata, M. and Penner, R. 2000. InsP4 facilitates store-operated calcium influx by inhibition of InsP3 5-phosphatase. Nature 408, 735–740. Hiasa, Y., Kitahori, Y., Morimoto, J., Konishi, N., Nakaoka, S. and Nishioka, H. 1992. Carcinogenicity study in rats of phytic acid ‘Daiichi’, a natural food additive. Food Chem. Toxicol. 30, 117–125. Hilton, J. M., Plomann, M., Ritter, B., Modregger, J., Freeman, H. N., Falck, J. R., Krishna, U. M. and Tobin, A. B. 2001. Phosphorylation of a synaptic vesicle-associated protein by an inositol hexakisphosphate-regulated protein kinase. J. Biol. Chem. 276, 16341–16347.

INOSITOL PHOSPHATES IN FOODS

49

Hirose, M., Fukushima, S., Imaida, K., Ito, N. and Shirai, T. 1999. Modifying effects of phytic acid and gamma-oryzanol on the promotion stage of rat carcinogenesis. Anticancer Res. 19, 3665–3670. Hix, D. K., Klopfenstein, C. F. and Walker, C. E. 1997. Physical and chemical attributes and consumer acceptance of sugar-snap cookies containing naturally occurring antioxidants. Cereal Chem. 74, 281–283. Ho, M. W. Y., Kaetzel, M. A., Armstrong, D. L. and Shears, S. B. 2001. Regulation of a human chloride channel. A paradigm for integrating input from calcium, type II calmodulindependent protein kinase, and inositol 3,4,5,6-tetrakisphosphate. J. Biol. Chem. 276, 18673–18680. Holub, B. J. 1982. The nutritional significance, metabolism, and function of myo-inositol and phosphatidylinositol in health and disease. Adv. Nutr. Res. 4, 107–141. Holub, B. J. 1986. Metabolism and function of myo-inositol and inositol phospholipids. Ann. Rev. Nutr. 6, 563–597. Hotz, C. and Gibson, R. S. 2001. Assessment of home-based processing methods to reduce the phytate content and phytate/zinc molar ratio of white maize (Zea mays). J. Agric. Food Chem. 49, 692–698. Huang, C., Ma, W. Y., Hecht, S. S. and Dong, Z. 1997. Inositol hexaphosphate inhibits cell transformation and activator protein 1 activation by targeting phosphatidylinositol-3′ kinase. Cancer Res. 57, 2873–2878. Huang, C. F., Voglmaier, S. M., Bembenek, M. E., Saiardi, A. and Snyder, S. H. 1998. Identification and purification of diphosphoinositol pentakisphosphate kinase, which synthesizes the inositol pyrophosphate bis(diphospho)inositol tetrakisphosphate. Biochemistry 37, 14998–15004. Hübel, F. and Beck, E. 1996. Maize root phytase. Purification, characterization, and localization of enzyme activity and its putative substrate. Plant Physiol. 112, 1429–1436. Hull, S. R., and Montgomery, R. 1995. Myo-inositol phosphates in corn steep water. J. Agric. Food Chem. 43, 1516–1523. Humbert, J. P., Matter, N., Artault, J. C., Koppler, P. and Malviya, A. N. 1996. Inositol 1,4,5trisphosphate receptor is located to the inner nuclear membrane vindicating regulation of nuclear calcium signaling by inositol 1,4,5-trisphosphate: Discrete distribution of inositol phosphate receptors to inner and outer nuclear membranes. J. Biol. Chem. 271, 478–485. Irvine, R. F. 1990. “Methods in Inositide Research.” Raven Press, New York. Irvine, R. F. and Schell, M. J. 2001. Back in the water: The return of the inositol phosphates. Nature Rev. Mol. Cell Biol. 2, 327–338. Irving, G. C. J. and Cosgrove, D. J. 1972. Inositol phosphate phosphatases of microbiological origin: the inositol pentaphosphate products of Aspergillus ficuum phytases. J. Bacteriol. 112, 434–438. IUPAC-IUB 1968. Tentative rules for cyclitol nomenclature. J. Biol. Chem. 243, 5809–5819. Ismailov, I. I., Fuller, C. M., Berdiev, B. K., Shlyonsky, V. G., Benos, D. J. and Barrett, K. E. 1996. A biologic function for an ‘orphan’ messenger: D-myo-inositol 3,4,5,6-tetrakisphosphate selectively blocks epithelial calcium-activated chloride channels. Proc. Natl Acad. Sci. USA 93, 10505–10509. Jariwalla, R. J. 1999. Inositol hexaphosphate (IP6) as an anti-neoplastic and lipid-lowering agent. Anticancer Res. 19, 3699–3702. Jayaraman, T. and Marks, A. R. 2000. Calcineurin is downstream of the inositol 1,4,5trisphosphate receptor in the apoptotic and cell growth pathways. J. Biol. Chem. 275, 6417–6420. Jenab, M. and Thompson, L. U. 1998. The influence of phytic acid in wheat bran on early biomarkers of colon carcinogenesis. Carcinogenesis 19, 1087–1092.

50

B. Q. PHILLIPPY

Jenab, M. and Thompson, L. U. 2000. Phytic acid in wheat bran affects colon morphology, cell differentiation and apoptosis. Carcinogenesis 21, 1547–1552. Jenab, M. and Thompson, L. U. 2002. Role of phytic acid in cancer and other diseases. In “Food Phytates” (Reddy, N. R. and Sathe, S. K., eds), pp. 225–248. CRC Press, Boca Raton, FL. Ji, H., Sandberg, K., Baukal, A. J. and Catt, K. J. 1989. Metabolism of inositol pentakisphosphate to inositol hexakisphosphate in Xenopus laevis oocytes. J. Biol. Chem. 264, 20185–20188. Johnson, K., Barrientos, L. G., Le, L. and Murthy, P. P. N. 1995. Application of two-dimensional total correlation spectroscopy for structure determination of individual inositol phosphates in a mixture. Anal. Biochem. 231, 421–431. Jyonouchi, H., Sun, S., Iijima, K., Wang, M. and Hecht, S. S. 1999. Effects of anti-7,8dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene on human small airway epithelial cells and the protective effects of myo-inositol. Carcinogenesis 20, 139–145. Kasim, A. B. and Edwards, H. M., Jr 1998. The analysis for inositol phosphate forms in feed ingredients. J. Sci. Food Agric. 76, 1–9. Katayama, T. 1999. Hypolipidemic action of phytic acid (IP6): Prevention of fatty liver. Anticancer Res. 19, 3695–3698. Katz, S. A. 1996. Are essential micronutrients toxic? Food Test. Anal. 2, 19–23. Kavran, J. M., Klein, D. E., Lee, A., Falasca, M., Isakoff, S. J., Skolnik, E. Y. and Lemmon, M. A. 1998. Specificity and promiscuity in phosphoinositide binding by pleckstrin homology domains. J.Biol. Chem. 273, 30497–30508. Kemme, P. A., Lommen, A., De Jonge, L. H., Van der Klis, J. D., Jongbloed, A. W., Mroz, Z. and Beynen, A. C. 1999. Quantification of inositol phosphates using 31P nuclear magnetic resonance spectroscopy in animal nutrition. J. Agric. Food Chem. 47, 5116–5121. Kennefick, S. and Cashman, K. D. 2000. Inhibitory effect of wheat fiber extract on calcium absorption in Caco-2 cells: Evidence for a role of associated phytate rather than fiber per se. Eur. J. Nutr. 39, 12–17. Kerovuo, J., Lauraeus, M., Nurminen, P., Kalkkinen, N. and Apajalahti, J. 1998. Isolation, characterization molecular gene cloning, and sequencing of a novel phytase from Bacillus subtilis. Appl. Environ. Microbial. 64, 2079–2085. Kerovuo, J., Rouvinen, J. and Hatzack, F. 2000. Analysis of myo-inositol hexakisphosphate hydrolysis by Bacillus phytase: Indication of a novel reaction mechanism. Biochem. J. 352, 623–628. Kim, Y. O., Kim, H. K., Bae, K. S., Yu, J. H. and Oh, T. K. 1998. Purification and properties of a thermostable phytase from Bacillus sp. DS11. Enzyme Microb. Technol. 22, 2–7. Knuckles, B. E. 1988. Effect of phytate and other myo-inositol phosphate esters on lipase activity. J. Food Sci. 53, 250–252. Knuckles, B. E. and Betschart, A. A. 1987. Effect of phytate and other myo-inositol phosphate esters on α-amylase digestion of starch. J. Food Sci. 52, 719–721. Kofman, O., Agam, G., Shapiro, J. and Spencer, A. 1998. Chronic dietary inositol enhances locomotor activity and brain inositol levels in rats. Psychopharmacology 139, 239–242. Kofman, O., Einat, H., Cohen, H., Tenne, H. and Shoshana, C. 2000. The anxiolytic effect of chronic inositol depends on the baseline level of anxiety. J. Neural Transm. 107, 241–253. Konietzny, U., Greiner, R. and Jany, K. D. 1994. Purification and characterization of a phytase from spelt. J. Food Biochem. 18, 165–183. Koppitz, B., Vogel, F. and Mayr, G. W. 1986. Mammalian aldolases are isomer-selective highaffinity inositol polyphosphate binders. Eur. J. Biochem. 161, 421–433. Kozlowska, H., Honke, J., Sadowska, J., Frias, J. and Vidal Valverde, C. 1996. Natural fermentation of lentils: influence of time, concentration and temperature on the kinetics of hydrolysis of inositol phosphates. J. Sci. Food Agric. 71, 367–375. Laboure, A. M., Gagnon, J. and Lescure, A. M. 1993. Purification and characterization of a

INOSITOL PHOSPHATES IN FOODS

51

phytase (myo-inositol-hexakisphosphate phosphohydrolase) accumulated in maize (Zea mays) seedlings during germination. Biochem. J. 295, 413–419. Lapp, C. and Speiss, B. 1991. Complexation studies on inositol phosphates III. Cadmium(II), lead(II), and mercury(II) complexes of D-myo-inositol 1,2,6-trisphosphate. J. Inorg. Biochem. 42, 257–266. Lardy, H. A. 1954. Inositols. I. Nomenclature and formula. A. Discussion of terminology. In “The Vitamins. Chemistry, Physiology, Pathology” (Sebrell, W. H. Jr. and Harris, R. S., eds), Vol. II, pp. 323–329. Academic Press, New York. Larson, S. R. and Raboy, V. 1999. Linkage mapping of maize and barley myo-inositol 1-phosphate synthase DNA sequences: Correspondence with a low phytic acid mutation. Theor. Appl. Genet. 99, 27–36. Larson, S. R., Rutger, J. N., Young, K. A. and Raboy, V. 2000. Isolation and genetic mapping of a non-lethal rice (Oryza sativa L.) low phytic acid 1 mutation. Crop Sci. 40, 1397–1405. Larsson, M. and Sandberg, A. S. 1991. Phytate reduction in bread containing oat flour, oat bran or rye bran. J. Cereal Sci. 14, 141–149. Larsson, M. and Sandberg, A. S. 1992. Phytate reduction in oats during malting. J. Food Sci. 57, 994–997. Larsson, M., Minekus, M. and Havenaar, R. 1997. Estimation of the bioavailability of iron and phosphorus in cereals using a dynamic in vitro gastrointestinal model. J. Sci. Food Agric. 74, 99–106. Lassen, S. F., Bech, L., Fuglsang, C. C., Breinholt, J., Ohmann, A. and ∅stergaard, P. R. 1998. Peniophora phytase. International Patent # WO 98/28408. Lee, B. J. and Hendricks, D. G. 1995. Phytic acid protective effect against beef round muscle lipid peroxidation. J. Food Sci. 60, 241–244. Lee, B. J., Hendricks, D. G. and Cornforth, D. P. 1998. Antioxidant effects of carnosine and phytic acid in a model beef system. J. Food Sci. 63, 394–402. Lee, D. Y., Schroeder, J., III and Gordon, D. T. 1988. Enhancement of Cu bioavailability in the rat by phytic acid. J. Nutr. 118, 712–717. Lehrfeld, J. 1989. High-performance liquid chromatography analysis of phytic acid on a pHstable, macroporous polymer column. Cereal Chem. 66, 510–515. Lehrfeld, J. 1994. HPLC separation and quantification of phytic acid and some inositol phosphates in foods: problems and solutions. J. Agric. Food Chem. 42, 2726–2731. Levrat-Verny, M. A., Coudray, C., Bellanger, J., Lopez, H. W., Demigne, C., Rayssiguier, Y. and Remesy, C. 1999. Wholewheat flour ensures higher mineral absorption and bioavailability than white wheat flour in rats. Br. J. Nutr. 82, 17–21. Li, M., Osaki, M., Honma, M. and Tadano, T. 1997. Purification and characterization of phytase induced in tomato roots under phosphorus-deficient conditions. Soil Sci. Plant Nutr. 43, 179–190. Lina, B. A. R., Hollanders, V. M. H. and Kuijpers, M. H. M. 1994. The role of alkalizing and neutral potassium salts in urinary bladder carcinogenesis in rats. Carcinogenesis 15, 523–527. Llinas, R., Sugimori, M., Lang, E. J., Morita, M., Fukuda, M., Niinobe, M. and Mikoshiba, K. 1994. The inositol high-polyphosphate series blocks synaptic transmission by preventing vesicular fusion: A squid giant synapse study. Proc. Natl Acad. Sci. USA 91, 12990–12993. Loewus, F. A. and Murthy, P. P. N. 2000. myo-Inositol metabolism in plants. Plant Sci. 150, 1–19. Lönnerdal, B., Sandberg, A. S., Sandstrom, B. and Kunz, C. 1989. Inhibitory effects of phytic acid and other inositol phosphates on zinc and calcium absorption in suckling rats. J. Nutr. 119, 211–214. Lopez, H. W., Coudray, C., Bellanger, J., Younes, H., Demigne, C. and Remesy, C. 1998. Intestinal fermentation lessens the inhibitory effects of phytic acid on mineral utilization in rats. J. Nutr. 128, 1192–1198.

52

B. Q. PHILLIPPY

Lopez, H. W., Coudray, C., Bellanger, J., Levrat-Verny, M. A., Demigne, C., Rayssiguier, Y. and Remesy, C. 2000a. Resistant starch improves mineral assimilation in rats adapted to a wheat bran diet. Nutr. Res. 20, 141–155. Lopez, H. W., Vallery, F., Levrat-Verny, M. A., Coudray, C., Demigne, C. and Remesy, C. 2000b. Dietary phytic acid and wheat bran enhance mucosal phytase activity in rat small intestine. J. Nutr. 130, 2020–2025. Lucca, P., Hurrell, R. and Potrykus, I. 2001. Approaches to improving the bioavailability and level of iron in rice seeds. J. Sci. Food Agric. 81, 828–834. Luo, H. R., Saiardi, A., Nagata, E., Ye, K. Q., Yu, H. B., Jung, T. S., Luo, X. J., Jain, S., Sawa, A. and Snyder, S. H. 2001. GRAB: A physiologic guanine nucleotide exchange factor for Rab3A, which interacts with inositol hexakisphosphate kinase. Neuron, 31, 439–451. Luo, H. R., Saiardi, A., Yu, H., Nagata, E., Ye, K. and Synder, S. H. 2002. Inositol pyrophosphates are required for DNA hyperrecombination in protein kinase C1 mutant yeast. Biochemistry 41, 2509–2515. Madurawe, R. D., Lin, Z., Dryden, P. K. and Lumpkin, J. A. 1997. Stability of lactate dehydrogenase in metal-catalyzed oxidation solutions containing chelated metals. Biotechnol. Prog. 13, 179–184. Maenz, D. D. and Classen, H. L. 1998. Phytase activity in the small intestinal brush border membrane of the chicken. Poultry Sci. 77, 557–563. March, J. G., Simonet, B. M. and Grases, F. 2001. Determination of phytic acid by gas chromatography–mass spectrometry: Application to biological samples. J. Chromatogr. B 757, 247–255. Mason, A. C., Weaver, C. M., Kimmel, S. and Brown, R. K. 1993. Effect of soybean phytate content on calcium bioavailability in mature and immature rats. J. Agric. Food Chem. 41, 246–249. Mayr, G. W. 1989. Inositol 1,4-bisphosphate is an allosteric activator of muscle-type 6-phosphofructo-1-kinase. Biochem. J. 259, 463–470. Mayr, G. W. and Thieleczek, R. 1991. Masses of inositol phosphates in resting and tetanically stimulated vertebrate skeletal muscles. Biochem. J. 280, 631–640. McConnell, F. M., Shears, S. B., Lane, P. J. L., Scheibel, M. S. and Clark, E. A. 1992. Relationships between the degree of cross-linking of surface immunoglobulin and the associated inositol1,4,5-trisphosphate and calcium signals in human B cells. Biochem. J. 284, 447–455. McKenzie-Parnell, J. M. and Guthrie, B. E. 1986. The phytate and mineral content of some cereals, cereal products, legumes, legume products, snack bars, and nuts available in New Zealand. Biol. Trace Elem. Res. 10, 107–121. McLaurin, J., Golomb, R., Jurewicz, A., Antel, J. P. and Fraser, P. E. 2000. Inositol stereoisomers stabilize an oligomeric aggregate of Alzheimer amyloid β peptide and inhibit A β-induced toxicity. J. Biol. Chem. 275, 18495–18502. Mendoza, C., Viteri F. E., Lonnerdal, B., Young K. A., Raboy, V. and Brown K. H. 1998. Effect of genetically modified, low-phytic acid maize on absorption of iron from tortillas. Am. J. Clin. Nutr. 68, 1123–1127. Mendoza, C., Viteri, F. E., Lonnerdal, B., Raboy, V., Young, K. A. and Brown, K. H. 2001. Absorption of iron from unmodified maize and genetically altered low-phytate maize fortified with ferrous sulfate or sodium iron EDTA. Am. J. Clin. Nutr. 73, 80–85. Mernissi-Arifi, Bieth, H., Schlewer, G. and Spiess, B. 1995. Complexation studies on inositolphosphates, VI. A13+ complexes of DL-MYO-inositol 1,4,5-triphosphate and D-MYO-inositol 1,2,6-triphosphate. J. Inorg. Biochem. 57, 127–133. Midorikawa, K., Murata, M., Oikawa, S., Hiraku, Y. and Kawanishi, S. 2001. Protective effect of phytic acid on oxidative DNA damage with reference to cancer prevention. Biochem. Biophys. Res. Commun. 288, 552–557.

INOSITOL PHOSPHATES IN FOODS

53

Miyamoto, S., Kuwata, G., Imai, M., Nagao, A. and Terao, J. 2000 Protective effect of phytic acid hydrolysis products on iron-induced lipid peroxidation of liposomal membranes. Lipids 35, 1411–1413. Moore, R. J. and Veum, T. L. 1983. Adaptive increase in phytate digestibility by phosphorusdeprived rats and the relationship of intestinal phytase (EC 3.1.3.8) and alkaline phosphatase (EC 3.1.3.1) to phytate utilization. Br. J. Nutr. 49, 145–152. Morris, E. R. and Hill, A. D. 1995. Inositol phosphate, calcium, magnesium, and zinc contents of selected breakfast cereals. J. Food Compos. Anal. 8, 3–11. Morris, E. R. and Hill, A. D. 1996. Inositol phosphate content of selected dry beans, peas, and lentils, raw and cooked. J. Food Compos. Anal. 9, 2–12. Morrison, B. H., Bauer, J. A., Kalvakolanu, D. V. and Lindner, D. J. 2001. Inositol hexakisphosphate kinase 2 mediates growth suppressive and apoptotic effects of interferon-β in ovarian carcinoma cells. J. Biol. Chem. 276, 24965–24970. Mullaney, E. J., Daly, C. B. and Ullah, A. H. J. 2000. Advances in phytase research. Adv. Appl. Microbiol. 47, 157–199 Munnik, T., Irvine, R. F. and Musgrave, A. 1998. Phospholipid signalling in plants. Biochim. Biophys. Acta 1389, 222–272. Nakano, T., Joh, T., Tokumoto, E. and Hayakawa, T. 1999. Purification and characterization of phytase from bran of Triticum aestivum L. cv. Nourin #61. Food Sci. Technol. Res. 5, 18–23. Nakano, T., Joh, T., Narita, K. and Hayakawa, T. 2000. The pathway of dephosphorylation of myo-inositol hexakisphosphate by phytases from wheat bran of Triticum aestivum L. cv. Nourin #61. Biosci. Biotechnol. Biochem. 64, 995–1003. NC-IUB 1989. Numbering of atoms in myo-inositol. Recommendations 1988. Biochem. J. 258, 1–2. Nemets, B., Fux, M., Levine, J. and Belmaker, R.H. 2001. Combination of antidepressant drugs: The case of inositol. Human Psychopharmacol. 16, 37–43. Nickel, K. P. and Belury, M. A. 1999. Inositol hexaphosphate reduces 12-O-tetradecanoylphorbol13-acetate-induced ornithine decarboxylase independent of protein kinase C isoform expression in keratinocytes. Cancer Lett. 140, 105–111. Nicoletti, F., Bruno, V., Fiore, L., Cavallaro, S. and Canonico, P. L. 1989. Inositol hexakisphosphate (phytic acid) enhances Ca2+ influx and D-[3H]aspartate release in cultured cerebellar neurons. J. Neurochem. 53, 1026–1030. Nishino, H., Murakoshi, M., Masuda, M., Tokuda, H., Satomi, Y., Onozuka, M., Yamaguchi, S., Bu, P., Tsuruta, A., Nosaka, K., Baba, M. and Takasuka, N. 1999. Suppression of lung and liver carcinogenesis in mice by oral administration of myo-inositol. Anticancer Res. 19, 3663–3664. Nogimori, K., Hughes, P. J., Glennon, M. C., Hodgson, M. E., Putney, J. W. and Shears, S. B. 1991. Purification of an inositol (1,3,4,5)-tetrakisphosphate 3-phosphatase activity from rat liver and the evaluation of its substrate specificity. J. Biol. Chem. 266, 16499–16506. Nolan, K. B., Duffin, P.A. and McWeeny, D. J. 1987. Effects of phytate on mineral bioavailability. In vitro studies on Mg2+, Ca2+, Fe3+, Cu2+, and Zn2+ (also Cd2+) solubilities in the presence of phytate. J. Sci. Food Agric. 40, 79–85. Norris, F. A., Ungewickell, E. and Majerus, P. W. 1995. Inositol hexakisphosphate binds to clathrin assembly protein 3 (AP-3/AP180) and inhibits clathrin cage assembly in vitro. J. Biol. Chem. 270, 214–217. Odom, A. R., Stahlberg, A., Wente, S. R. and York, J. D. 2000. A role for nuclear inositol 1,4,5trisphosphate kinase in transcriptional control. Science 287, 2026–2029. O’Neill, I. K., Sargent, M. and Trimble, M. L. 1980. Determination of phytate in foods by phosphorus-31 Fourier transform nuclear magnetic resonance spectrometry. Anal. Chem. 52, 1288–1291.

54

B. Q. PHILLIPPY

Ongusaha, P. P., Hughes, P. J., Davey, J. and Michell, R. H. 1998. Inositol hexakisphosphate in Schizosaccharomyces pombe: Synthesis from Ins(1,4,5)P3 and osmotic regulation. Biochem. J. 335, 671–679. Ozaki, S., DeWald, D. B., Shope, J. C., Chen, J. and Prestwich, G. D. 2000. Intracellular delivery of phosphoinositides and inositol phosphates using polyamine carriers. Proc. Natl Acad. Sci. USA. 97, 11286–11291. Pallauf, J., Pietsch, M. and Rimbach, G. 1998. Dietary phytate reduces magnesium bioavailability in growing rats. Nutr. Res. 18, 1029–1037. Patel, S., Joseph, S. K. and Thomas, A. P. 1999. Molecular properties of inositol 1,4,5-trisphosphate receptors. Cell Calcium. 25, 247–264. Perera, I. Y., Heilmann, I. and Boss, W. F. 1999. Transient and sustained inceases in inositol 1,4,5-trisphosphate precede the differential growth response in gravistimulated maize pulvini. Proc. Natl. Acad. Sci. USA 96, 5838–5843. Persson, H., Türk, M., Nyman, M. and Sandberg, A. S. 1998. Binding of Cu2+, Zn2+, and Cd2+ to inositol tri-, tetra-, penta-, hexaphosphates. J. Agric. Food Chem. 46, 3194–3200. Phillippy, B. Q. 1989. Identification by two-dimensional NMR of myo-inositol tris- and tetrakis(phosphates) formed from phytic acid by wheat phytase. J. Agric. Food Chem. 37, 1261–1265. Phillippy, B. Q. 1998a. Identification of inositol 1,3,4-trisphosphate 5-kinase and inositol 1,3,4,5-tetrakisphosphate 6-kinase in immature soybean seeds. Plant Physiol. 116, 291–297. Phillippy, B. Q. 1998b. Purification and catalytic properties of a phytase from scallion (Allium fistulosum L.) leaves. J. Agric. Food Chem. 46, 3491–3496. Phillippy, B. Q. and Bland, J. M. 1988. Gradient ion chromatography of inositol phosphates. Anal. Biochem. 175, 162–166. Phillippy, B. Q. and Graf, E. 1997. Antioxidant functions of inositol 1,2,3-trisphosphate and inositol 1,2,3,6-tetrakisphosphate. Free Radical Biol. Med. 22, 939–946. Phillippy, B. Q. and Wyatt, C. J. 2001. Degradation of phytate in foods by phytases in fruit and vegetable extracts. J. Food Sci. 66, 535–539. Phillippy, B. Q., White, K. D., Johnston, M. R., Tao, S. -H. and Fox, M. R. S. 1987. Preparation of inositol phosphates from sodium phytate by enzymatic and nonenzymatic hydrolysis. Anal. Biochem. 162, 115–121. Phillippy, B. Q., Johnston, M. R., Tao, S. -H. and Fox, M. R. S. 1988. Inositol phosphates in processed foods. J. Food Sci. 53, 496–499. Phillippy, B. Q., Ullah, A. H. J. and Ehrlich, K. C. 1994. Purification and some properties of inositol 1,3,4,5,6-pentakisphosphate 2-kinase from immature soybean seeds. J. Biol. Chem. 269, 28393–28399. Pittet, D., Schlegel, W., Lew, D. P., Monod, A. and Mayr, G. W. 1989. Mass changes in inositol tetrakis- and pentakisphosphate isomers induced by chemotactic peptide stimulation in HL-60 cells. J. Biol. Chem. 264, 18489–18493. Plaami, S. and Kumpulainen, J. 1995. Inositol phosphate content of some cereal-based foods. J. Food Compos. Anal. 8, 324–335. Porres, J. M., Stahl, C. H., Cheng, W. H., Fu, Y., Roneker, K. R., Pond, W. G. and Lei, X. G. (1999). Dietary intrinsic phytate protects colon from lipid peroxidation in pigs with a moderately high dietary iron intake. Proc. Soc. Exp. Biol. Med. 221, 80–86. Posternak, T. 1965. “ The Cyclitols.” Holden-Day, San Francisco, CA. Poyner, D. R., Cooke, F., Hanley, M. R., Reynolds, D. J. M. and Hawkins, P. T. 1993. Characterization of metal ion-induced [3H]inositol hexakisphosphate binding to rat cerebellar membranes. J.Biol. Chem. 268, 1032–1038. Raboy, V. and Dickinson, D. B. 1987. The timing and rate of phytic acid accumulation in developing soybean seeds. Plant Physiol. 85, 841–844.

INOSITOL PHOSPHATES IN FOODS

55

Raboy, V., Gerbasi, P. F., Young, K. A., Stoneberg, S. D., Pickett, S. G., Bauman, A. T., Murthy, P. P. N., Sheridan, W. F. and Ertl, D. S. 2000. Origin and seed phenotype of maize low phytic acid 1-1 and low phytic acid 2-1. Plant Physiol. 124, 355–368. Radenberg, T., Scholz, P., Bergmann, G. and Mayr, G. W. 1989. The quantitative spectrum of inositol phosphate metabolites in avian erythrocytes, analysed by proton n.m.r. and h.p.l.c. with direct isomer detection. Biochem. J. 264, 323–333. Rakba, N., Loyer, P., Gilot, D., Delcros, J. G., Glaise, D., Baret, P., Pierre, J. L., Brissot, P. and Lescoat, G. 2000. Antiproliferative and apoptotic effects of O-Trensox, a new synthetic iron chelator, on differentiated human hepatoma cell lines. Carcinogenesis 21, 943–951. Rao, R. K. and Ramakrishnan, C. V. 1985. Studies on inositolphosphatase in rat small intestine. Enzyme 33, 205–215. Razzini, G., Berrie, C. P., Vignati, S., Broggini, M., Mascetta, G., Brancaccio, A. and Falasca, M. 2000. Novel functional PI 3-kinase antagonists inhibit cell growth and tumorigenicity in human cancer cell lines. FASEB J. 14, 1179–1187. Reddy, N. R. 2002. Occurrence, distribution, content, and dietary intake of phytate. In “Food Phytates” (Reddy, N. R. and Sathe, S. K. eds), pp. 25–51. CRC Press, Boca Raton, FL. Reddy, N. R., Sathe, S. K. and Salunkhe, D. K. 1982. Phytates in legumes and cereals. Nutritional significance of phytic acid. Adv. Food Res. 28, 1–92. Reddy, N. R., Pierson, M. D., Sathe, S. K. and Salunkhe, D. K. 1989. “Phytates in Cereals and Legumes.” CRC Press, Boca Raton, FL. Rickard, S. E. and Thompson, L. U. 1997. Interactions and biological effects of phytic acid. In “Antinutrients and Phytochemicals in Food”, ACS Symposium Series 662, pp. 294–312. American Chemical Society, Washington, DC. Rimbach, G. and Pallauf, J. 1998. Phytic acid inhibits free radical formation in vitro but does not affect liver oxidant or antioxidant status in growing rats. J. Nutr. 128, 1950–1955. Rimbach, G. and Pallauf, J. 1999. Effect of dietary phytate on magnesium bioavailability and liver oxidant status in growing rats. Food Chem. Toxicol. 37, 37–45. Rimbach, G., Walter, A., Most, E. and Pallauf, J. 1998. Effect of microbial phytase on zinc bioavailability and cadmium and lead accumulation in growing rats. Food Chem. Toxicol. 36, 7–12. Rounds, M. A. and Nielsen, S. S. 1993. Anion-exchange high-performance liquid chromatography with post-column detection for the analysis of phytic acid and other inositol phosphates. J. Chromatogr. A 653, 148–152. Saha, P. R., Weaver, C. M. and Mason, A. C. 1994. Mineral bioavailability in rats from intrinsically labeled whole wheat flour of various phytate levels. J. Agric. Food Chem. 42, 2531–2535. Saiardi, A., Caffrey, J. J., Snyder, S. H. and Shears, S. B. 2000a. Inositol polyphosphate multikinase (ArgRIII) determines nuclear mRNA export in Saccharomyces cerevisiae. FEBS Lett. 468, 28–32. Saiardi, A., Caffrey, J. J., Snyder, S. H. and Shears, S. B. 2000b. The inositol hexakisphosphate kinase family. Catalytic flexibility and function in yeast vacuole biogenesis. J. Biol. Chem. 275, 24686–24692. Saiardi, A., Nagata, E., Luo, H. R., Sawa, A., Luo, X., Snowman, A. M. and Snyder, S. H. 2001. Mammalian inositol polyphosphate multikinase synthesizes inositol 1,4,5-trisphosphate and an inositol pyrophosphate. Proc. Natl. Acad. Sci. USA 98, 2306–2311. Saied, I. T. and Shamsuddin, A. M. 1998. Up-regulation of the tumor suppressor gene p53 and WAF1 gene expression by IP6 in HT-29 human colon carcinoma cell line. Anticancer Res. 18, 1479–1484. Sakamoto, K., Vucenik, I. and Shamsuddin, A. M. 1993. Tritiated phytic acid (inositol hexaphosphate) is absorbed and distributed to various tissues in rats. J. Nutr. 123, 713–720.

56

B. Q. PHILLIPPY

Sandberg, A.-S. and Ahderinne, R. 1986. HPLC method for determination of inositol tri-, tetra, penta-, and hexaphosphates in foods and intestinal contents. J. Food Sci. 51, 547–550. Sandberg, A. -S. and Andersson, H. 1988. Effect of dietary phytase on the digestion of phytate in the stomach and small intestine of humans. J. Nutr. 118, 469–473. Sandberg, A. -S. and Svanberg, U. 1991. Phytate hydrolysis by phytase in cereals; effects on in vitro estimation of iron availability. J. Food Sci. 56, 1330–1333. Sandberg, A. -S., Carlsson, N. -G. and Svanberg, U. 1989. Effects of inositol tri-, tetra-, penta-, and hexaphosphates on in vitro estimation of iron availability. J. Food Sci. 54, 159–161. Sandberg, A. -S., Hulthén, L. R. and Türk, M. 1996. Dietary Aspergillus niger phytase increases iron absorption in humans. J. Nutr. 126, 476–480. Sandberg, A.-S., Brune, M., Carlsson, N. -G., Hallberg, L., Skoglund, E. and RossanderHulthén, L. 1999. Inositol phosphates with different numbers of phosphate groups influence iron absorption in humans. Am. J. Clin. Nutr. 70, 240–246. Sandström, B. and Sandberg, A. S. 1992. Inhibitory effects of isolated inositol phosphates on zinc absorption in humans. J. Trace Elem. Electrolytes Health Dis. 6, 99–103. Sandström, B., Bugel, S., McGaw, B. A., Price, J. and Reid, M. D. 2000. A high oat-bran intake does not impair zinc absorption in humans when added to a low-fiber animal protein-based diet. J. Nutr. 130, 594–599. Sathe, S. K. and Venkatachalam, M. 2002. Influence of processing technologies on phytate and its removal. In “Food Phytates” (Reddy, N. R. and Sathe, S. K., eds), pp. 157–188. CRC Press, Boca Raton, FL. Schell, M. J., Letcher, A. J., Brearley, C. A., Biber, J., Murer, H. and Irvine, R. F. 1999. PiUS (Pi uptake stimulator) is an inositol hexakisphosphate kinase. FEBS Lett. 461, 169–172. Seedat, S. and Stein, D. J. 1999. Inositol augmentation of serotonin reuptake inhibitors in treatment-refractory obsessive–compulsive disorder: An open trial. Int. Clin. Psychopharmacol. 14, 353–356. Segueilha, L., Lambrechts, C., Boze, H., Moulin, G. and Galzy, P. 1992. Purification and properties of the phytase from Schwanniomyces castellii. J. Ferment. Bioeng. 74, 7–11. Sekiguchi, Y., Matsunaga, A., Yamamoto, A. and Inoue, Y. 2000. Analysis of condensed phosphates in food products by ion chromatography with an on-line hydroxide eluent generator. J. Chromatogr. A 881, 639–644. Shamsuddin, A. M. 1995. Inositol phosphates have novel anticancer function. J. Nutr. 125, 725S–732S. Shamsuddin, A. M. 1999. Metabolism and cellular functions of IP6: A review. Anticancer Res. 19, 3733–3736. Shamsuddin, A. M., Ullah, A. and Chakravarthy, A. K. 1989. Inositol and inositol hexaphosphate suppress cell proliferation and tumor formation in CD-1 mice. Carcinogenesis 10, 1461–1463. Shears, S. B. 1989. Metabolism of the inositol phosphates produced upon receptor activation. Biochem. J. 260, 313–324. Shears, S. B. 1998. The versatility of inositol phosphates as cellular signals. Biochim. Biophys. Acta 1436, 49–67. Shears, S. B. 2001. Assessing the omnipotence of inositol hexakisphosphate. Cell. Signal. 13, 151–158. Shears, S. B., Ali, N., Craxton, A. and Bembenek, M. E. 1995. Synthesis and metabolism of bisdiphosphoinositol tetrakisphosphate in vitro and in vivo. J. Biol. Chem. 270, 10489–10497. Shen, X., Weaver, C. M., Kempa-Steczko, A., Martin, B. R., Phillippy, B. Q. and Heaney, R. P. 1998. An inositol phosphate as a calcium absorption enhancer in rats. J. Nutr. Biochem. 9, 298–301. Simpson, C. J. and Wise, A. 1990. Binding of zinc and calcium to inositol phosphates (phytate) in vitro. Br. J. Nutr. 64, 225–232.

INOSITOL PHOSPHATES IN FOODS

57

Singh, A., Singh, S. P. and Bamezai, R. 1997. Modulatory influence of arecoline on the phytic acid-altered hepatic biotransformation system enzymes, sulfhydryl content and lipid peroxidation in a murine system. Cancer Lett. 117, 1–6. Skoglund, E. and Sandberg, A. -S. 2002. Methods for analysis of phytate. In “Food Phytates.” (Reddy, N. R. and Sathe, S. K., eds), pp. 127–137. CRC Press, Boca Raton, FL. Skoglund, E., Carlsson, N. -G. and Sandberg, A. -S. 1997a. Determination of isomers of inositol mono- to hexaphosphates in selected foods and intestinal contents using high-performance ion chromatography. J. Agric. Food Chem. 45, 431–436. Skoglund, E., Carlsson, N. -G. and Sandberg, A. -S. 1997b. Analysis of inositol mono- and diphosphate isomers using high-performance ion chromatography and pulsed amperometric detection. J. Agric. Food Chem. 45, 4668–4673. Skoglund, E., Carlsson, N. -G. and Sandberg, A. -S. 1998. High-performance chromatographic separation of inositol phosphate isomers on strong anion exchange columns. J. Agric. Food Chem. 46, 1877–1882. Skoglund, E., Lönnerdal, B. and Sandberg, A. -S. 1999. Inositol phosphates influence iron uptake in Caco-2 cells. J. Agric. Food Chem. 47, 1109–1113. Smith, P. M., Harmer, A. R., Letcher, A. J. and Irvine, R. F. 2000. The effect of inositol 1,3,4,5tetrakisphosphate on inositol trisphosphate-induced Ca2+ mobilization in freshly isolated and cultured mouse lacrimal acinar cells. Biochem. J. 347, 77–82. Spiers, I. D., Freeman, S., Poyner, D. R. and Schwalbe, C. H. 1995. The first synthesis and iron binding studies of the natural product, myo-inositol 1,2,3-trisphosphate. Tetrahedron Lett. 36, 2125–2128. Spiers, I. D., Barker, C. J., Chung, S. K., Chang, Y. T., Freeman, S., Gardiner, J. M., Hirst, P. H., Lambert, P. A., Michell, R. H., Poyner, D. R., Schwalbe, C. H., Smith, A. W. and Solomons, K. R. H. 1996. Synthesis and iron binding studies of myo-inositol 1,2,3trisphosphate and (±)-myo-inositol 1,2-bisphosphate, and iron binding studies of all myoinositol tetrakisphosphates. Carbohydr. Res. 282, 81–99. Steadman, K. J., Burgoon, M. S., Schuster, R. L., Lewis, B. A., Edwardson, S. E. and Obendorf, R. L. 2000. Fagopyritols, D-chiro-inositol, and other soluble carbohydrates in buckwheat seed milling fractions. J. Agric. Food Chem. 48, 2843–2847. Stephens, L. R. and Downes, C. P. 1990. Product-precursor relationships amongst inositol polyphosphates. Incorporation of [32P]Pi into myo-inositol 1,3,4,6-tetrakisphosphate, myoinositol 1,3,4,5-tetrakisphosphate, myo-inositol 3,4,5,6-tetrakisphosphate and myo-inositol 1,3,4,5,6-pentakisphosphate in intact avian erythrocytes. Biochem. J. 265, 435–52. Stephens, L. R. and Irvine, R. F. 1990. Stepwise phosphorylation of myo-inositol leading to myo-inositol hexakisphosphate in Dictyostelium. Nature 346, 580–583. Stephens, L. R., Kay, R. R. and Irvine, R. F. 1990. A myo-inositol D-3 hydroxykinase activity in Dictyostelium. Biochem. J. 272, 201–210. Stephens, L. R., Hawkins, P. T., Stanley, A. F., Moore, T., Poyner, D. R., Morris, P. J., Hanley, M. R., Kay, R. R. and Irvine, R. F. 1991. Myo-inositol pentakisphosphates. Structure, biological occurrence and phosphorylation to myo-inositol hexakisphosphate. Biochem. J. 275, 485–499. Stevenson, J. M., Perera, I. Y., Heilmann, I., Persson, S. and Boss, W. F. 2000. Inositol signaling and plant growth. Trends Plant Sci. 5, 252–258. Streb, H., Irvine, R. F., Berridge, M. J. and Schulz, I. 1983. Release of Ca2+ from a nonmitochondrial intracellular store in pancreatic acinar cells by inositol-1,4,5-trisphosphate. Nature 306, 67–69. Sutardi and Buckle, K. A. 1985. Reduction in phytic acid levels in soybeans during tempeh production, storage and frying. J. Food Sci. 50, 260–261, 263. Suzuki, T., Suematsu, M. and Makino, R. 2001. Organic phosphates as a new class of soluble guanylate cyclase inhibitors. FEBS Lett. 507, 49–53. Szwergold, B. S., Graham, R. A. and Brown, T. R. 1987. Observation of inositol pentakis- and

58

B. Q. PHILLIPPY

hexakis-phosphates in mammalian tissues by 31P NMR. Biochem. Biophys. Res. Commun. 149, 874–881. Talamond, P., Doulbeau, S., Rochette, I., Guyot, J. P. and Treche, S. 2000. Anion-exchange high-performance liquid chromatography with conductivity detection for the analysis of phytic acid in food. J. Chromatogr. A 871, 7–12. Tanimura, A., Tojyo, Y. and Turner, R. J. 2000. Evidence that type I, II, and III inositol 1,4,5trisphosphate receptors can occur as integral plasma membrane proteins. J. Biol. Chem. 275, 27488–27493. Tao, S. -H., Fox, M. R. S., Phillippy, B. Q., Fry, B. E., Jr., Johnson, M. L. and Johnston, M. R. 1986. Effects of inositol phosphates on mineral utilization. Fed. Proc. 45, 819. Tarnow, P., Jonsson, A., Lindblom, L., Gustafsson, T. and Cassuto, J. 1998. Topical D-myoinositol-1,2,6-trisphosphate and hexylbetaine treatment reduces albumin extravasation in experimental rat skin burn injury. Burns 24, 460–463. Tasker, P. N., Taylor, C. W. and Nixon, G. F. 2000. Expression and distribution of InsP3 receptor subtypes in proliferating vascular smooth muscle cells. Biochem. Biophys. Res. Commun. 273, 907–912. Taylor, C. W., Genazzani, A. A. and Morris, S. A. 1999. Expression of inositol trisphosphate receptors. Cell Calcium 26, 237–251. Thieleczek, R., Mayr, G. W. and Brandt, N. R. 1989. Inositol polyphosphate-mediated repartitioning of aldolase in skeletal muscle triads and myofibrils. J. Biol. Chem. 264, 7349–7356. Timerman, A. P., Mayrleitner, M. M., Lukas, T. J., Chadwick, C. C., Saito, A., Watterson, D. M., Schindler, H. and Fleischer, S. 1992. Inositol polyphosphate receptor and clathrin assembly protein AP-2 are related proteins that form potassium-selective ion channels in planar lipid bilayers. Proc. Natl. Acad. Sci. USA. 89, 8976–8980. Tomlinson, R. V. and Ballou, C. E. 1962. Myoinositol polyphosphate intermediates in the dephosphorylation of phytic acid by phytase. Biochemistry 1, 166–171. Trugo, L. C., Muzquiz, M., Pedrosa, M.M., Ayet, G., Burbano, C., Cuadrado, C. and Cavieres, E. 1999. Influence of malting on selected components of soya bean, black bean, chickpea and barley. Food Chem. 65, 85–90. Truong-Tran, A. Q., Ho, L. H., Chai, F. and Zalewski, P. D. 2000. Cellular zinc fluxes and the regulation of apoptosis/gene-directed cell death. J. Nutr. 130, 1459S–1466S. Türk, M. and Sandberg, A. -S. 1992. Phytate degradation during breadmaking: Effect of phytase addition. J. Cereal Sci. 15, 281–294. Türk, M., Carlsson, N. -G. and Sandberg, A. -S. 1996. Reduction in the levels of phytate during wholemeal bread making; effect of yeast and wheat phytases. J. Cereal Sci. 23, 257–264. Türk, M., Sandberg, A. -S., Carlsson, N. -G. and Andlid, T. 2000. Inositol hexaphosphate hydrolysis by baker’s yeast. Capacity, kinetics, and degradation products. J.Agric. Food Chem. 48, 100–104. Ullah, A. H. J. and Gibson, D. M. 1987. Extracellular phytase (E.C. 3.1.3.8) from Aspergillus ficuum NRRL 3135: Purification and characterization. Prep. Biochem. 17, 63–91. Vajanaphanich, M., Schultz, C., Rudolf, M. T., Wasserman, M., Enyedi, P., Craxton, A., Shears, S. B., Tsien, R. Y., Barrett, K. E. and Traynor, K. A. 1994. Long-term uncoupling of chloride secretion from intracellular calcium levels by Ins(3,4,5,6)P4. Nature 371, 711–714. Valencia, S., Svanberg, U., Sandberg, A. -S. and Ruales, J. 1999. Processing and quinoa (Chenopodium quinoa, Willd): Effects on in vitro iron availability and phytate hydrolysis. Int. J. Food Sci. Nutr. 50, 203–211. Vallejo, M., Jackson, T., Lightman, S. and Hanley, M. R. 1987. Occurrence and extracellular actions of inositol pentakis- and hexakisphosphate in mammalian brain. Nature 330, 656–658.

INOSITOL PHOSPHATES IN FOODS

59

Van der Kaay, J. and Van Haastert. P. J. M. 1995. Stereospecificity of inositol hexakisphosphate dephosphorylation by Paramecium phytase. Biochem. J. 312, 907–910. Van der Kaay, J., Wesseling, J. and Van Haastert, P. J. M. 1995. Nucleus-associated phosphorylation of Ins(1,4,5)P3 to InsP6 in Dictyostelium. Biochem. J. 312, 911–917. Van Dijken, P., de Haas, J. R., Craxton, A., Erneux, C., Shears, S. B. and Van Haastert, P. J. M. 1995. A novel, phospholipase C-independent pathway of inositol 1,4,5-trisphosphate formation in Dictyostelium and rat liver. J. Biol. Chem. 270, 29724–29731. Voglmaier, S. M., Keen, J. H., Murphy, J. E., Ferris, C. D., Prestwich, G. D., Snyder, S. H., and Theibert, A. B. 1992. Inositol hexakisphosphate receptor identified as the clathrin assembly protein AP-2. Biochem. Biophys. Res. Commun. 187, 158–163. Voglmaier, S. M., Bembenek, M. E., Kaplin, A. I., Dorman, G., Olszewski, J. D., Prestwich, G. D. and Snyder, S. H. 1996. Purified inositol hexakisphosphate kinase is an ATP synthase: Diphosphoinositol pentakisphosphate as a high-energy phosphate donor. Proc. Natl Acad. Sci. USA 93, 4305–4310. Vohra, P., Gray, G. A. and Kratzer, F. H. 1965. Phytic acid-metal complexes. Proc. Soc. Exp. Biol. Med. 120, 447–449. Vucenik, I. and Shamsuddin, A. M. 1994. [3H]Inositol hexaphosphate (phytic acid) is rapidly absorbed and metabolized by murine and human malignant cells in vitro. J. Nutr. 124, 861–868. Vucenik, I., Yang, G. Y. and Shamsuddin, A. M. 1997. Comparison of pure inositol hexaphosphate and high-bran diet in the prevention of DMBA-induced rat mammary carcinogenesis. Nutr. Cancer 28, 7–13. Vucenik, I., Kalebic, T., Tantivejkul, K. and Shamsuddin, A. M. 1998. Novel anticancer function of inositol hexaphosphate: Inhibition of human rhabdomyosarcoma in vitro and in vivo. Anticancer Res. 18, 1377–1384. Vucenik, I., Podczasy, J. J. and Shamsuddin, A. M. 1999. Antiplatelet activity of inositol hexaphosphate (IP6). Anticancer Res. 19, 3689–3693. Wattenberg, L. W., Wiedmann, T. S., Estensen, R. D., Zimmerman, C. L., Galbraith, A. R., Steele, V. E. and Kelloff, G. J. 2000. Chemoprevention of pulmonary carcinogenesis by brief exposures to aerosolized budesonide or beclomethasone dipropionate and by the combination of aerosolized budesonide and dietary myo-inositol. Carcinogenesis 21, 179–182. Weaver, C. M. and Kannan, S. 2002. Phytate and mineral bioavailability. In “Food Phytates” (Reddy, N. R. and Sathe, S. K., eds), pp. 211–223. CRC Press, Boca Raton, FL. Weaver, C. M., Heaney, R. P., Proulx, W. R., Hinders, S. M. and Packard, P. T. 1993. Absorbability of calcium from common beans. J. Food Sci. 58, 1401–1403. Weaver, C. M., Heaney, R. P., Teegarden, D., and Hinders, S. M. 1996. Wheat bran abolishes the inverse relationship between calcium load size and absorption fraction in women. J. Nutr. 126, 303–307. Weber, G., Shen, F., Yang, H., Prajda, N. and Li, W. 1999. Regulation of signal transduction activity in normal and cancer cells. Anticancer Res. 19, 3703–3709. Weinberg, E. D. 1999. Iron loading and disease surveillance. Emerg. Infect. Dis. 5, 346–352. Welch, R. M., House, W. A., Beebe, S. and Cheng, Z. 2000. Genetic selection for enhanced bioavailable levels of iron in bean (Phaseolus vulgaris L.) seeds. J. Agric. Food Chem. 48, 3576–3580. Wilcox, J. R., Premachandra, G. S., Young, K. A. and Raboy, V. 2000. Isolation of high seed inorganic P, low-phytate soybean mutants. Crop Sci. 40, 1601–1605. Wilson, M. P. and Majerus, P. W. 1997. Characterization of a cDNA encoding Arabidopsis thaliana inositol 1,3,4-trisphosphate 5/6-kinase. Biochem. Biophys. Res. Commun. 232, 678–681. Wyss, M., Brugger, R., Kronenberger, A., Rémy, R., Kimbel, R., Oesterhelt, G., Lehmann, M. and van Loon, A. P. G. M. 1999. Biochemical characterization of fungal phytases (myo-

60

B. Q. PHILLIPPY

inositol hexakisphosphate phosphohydrolases): Catalytic properties. Appl. Environ. Microbiol. 65, 367–373. Xie, W., Kaetzel, M. A., Bruzik, K. S., Dedman, J. R., Shears, S. B. and Nelson, D. J. 1996. Inositol 3,4,5,6-tetrakisphosphate inhibits the calmodulin-dependent protein kinase IIactivated chloride conductance in T84 colonic epithelial cells. J. Biol. Chem. 271, 14092–14097. Xu, P., Price, J. and Aggett, P. J. 1992. Recent advances in methodology for analysis of phytate and inositol phosphates in foods. Prog. Food Nutr. Sci. 16, 245–262. Yamaguchi, Y., Ikenaka, K., Niinobe, M., Yamada, H. and Mikoshiba, K. 1996. Myelin proteolipid protein (PLP), but not DM-20, is an inositol hexakisphosphate-binding protein. J. Biol. Chem. 271, 27838–27846. Yang, W. J., Matsuda, Y., Sano, S., Masutani, H. and Nakagawa, H. 1991a. Purification and characterization of phytase from rat intestinal mucosa. Biochim. Biophys. Acta 1075, 75–82. Yang, W. J., Matusda, Y., Inomata, M. and Nakagawa, H. 1991b. Developmental and dietary induction of the 90K subunit of rat intestinal phytase. Biochim. Biophys. Acta 1075, 83–87. Yang, X., Safrany, S. T. and Shears, S. B. 1999. Site-directed mutagenesis of diphosphoinositol polyphosphate phosphohydrolase, a dual specificity NUDT enzyme that attacks diadenosine polyphosphates and diphosphoinositol polyphosphates. J. Biol. Chem. 274, 35434–35440. Yang, S. -N., Yu, J., Mayr, G. W., Hofmann, F., Larsson, O. and Berggren, P. -O. 2001. Inositol hexakisphosphate increases L-type Ca2+ channel activity by stimulation of adenylyl cyclase. FASEB J. 15, 1753–1763. Ye, W., Ali, N., Bembenek, M. E., Shears, S. B. and Lafer, E. M. 1995. Inhibition of clathrin assembly by high affinity binding of specific inositol polyphosphates to the synapsespecific clathrin assembly protein AP-3. J. Biol. Chem. 270, 1564–1568. York, J. D., Odom, A. R., Murphy, R., Ives, E. B. and Wente, S. R. 1999. A phospholipase Cdependent inositol polyphosphate kinase pathway required for efficient messenger RNA export. Science 285, 96–100. Yoshida, K. T., Wada, T., Koyama, H., Mizobuchi-Fukuoka, R. and Naito, S. 1999. Temporal and spatial patterns of accumulation of the transcript of myo-inositol-1-phosphate synthase and phytin-containing particles during seed development in rice. Plant Physiol. 119, 65–72. Zhou, J. R. and Erdman, J. W., Jr. 1995. Phytic acid in health and disease. Crit. Rev. Food Sci. Nutr. 35, 495–508. Zi, X., Singh, R. P. and Agarwal, R. 2000. Impairment of erbB1 receptor and fluid-phase endocytosis and associated mitogenic signaling by inositol hexaphosphate in human prostate carcinoma DU145 cells. Carcinogenesis 21, 2225–2235.

PYRROLIZIDINE ALKALOIDS IN FOODS ROGER A. COULOMBE, JR Graduate Program in Toxicology and Department of Veterinary Sciences Utah State University Logan, UT 84322-4620 USA

I. II. III. IV. V. VI. VII. VIII.

Introduction Plant Sources Chemical Structures of Pyrrolizidine Alkaloids Pyrrolizidine Alkaloids in Foods and Herbal Medicines A. Pyrrolizidine Alkaloids in Foods B. Pyrrolizidine Alkaloids in Traditional Remedies and Medicines Toxicity of Pyrrolizidine Alkaloids Metabolism of Pyrrolizidine Alkaloids Mechanism of Toxic Action Control of Pyrrolizidine Alkaloids and Future Prospects Acknowledgements References

I. INTRODUCTION

In addition to the many well-known major nutrients (protein, fat, carbohydrate and fiber) and minor nutrients (vitamins, minerals and nonessential compounds), foods contain thousands of naturally present toxic plant compounds. Some are carcinogenic in animals, and thus may be potentially carcinogenic in people. Many of these compounds are commonly termed “nature’s pesticides” because they are often toxic to predators, such as insects and animals, thereby conferring a competitive advantage to the plant that produces them. Although these chemicals are in every meal we eat, they have received little attention compared to that given to the relatively minor residues of synthetic chemicals such as polychlorinated biphenyls (PCBs) and pesticides. Our food contains greater than 10 000-fold more natural toxins than the synthetic kind, and ADVANCES IN FOOD AND NUTRITION RESEARCH VOL 45 ISBN: 0-12-016445-0

Copyright © 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

62

R. A. COULOMBE

in terms of metabolic reactions, our bodies are not able to distinguish between the two. Despite the popular notion equating “natural” and “healthy,” it is clear that natural toxins pose a far greater health risk than that posed by synthetic chemicals in our foods. One important and well-known class of naturally occurring chemicals in foods is the pyrrolizidine alkaloids. Pyrrolizidine alkaloids are a large group of compounds; more than 350 pyrrolizidine alkaloids have been isolated from over 6000 plant species. The majority of pyrrolizidine alkaloids are toxic. Many of these have been shown to cause cancer in animals, and are therefore potentially carcinogenic in people. People become intoxicated by pyrrolizidine alkaloids in a variety of ways. Epidemics of pyrrolizidine alkaloid poisoning have typically occurred when large numbers of people eat foods made from small grains contaminated with seeds from pyrrolizidine alkaloid-containing plants. Other less spectacular poisoning incidents occur from the use of certain dietary supplements and traditional “remedies,” direct consumption of pyrrolizidine alkaloid-containing plants, or through residues in food products, such as eggs, meat, milk and honey. The real extent of human poisoning by pyrrolizidine alkaloids has probably gone woefully underreported. Pyrrolizidine alkaloid poisonings in people have been referred to as an “iceberg disease” because reported cases of pyrrolizidine alkaloid poisoning are probably only a small percentage of the true incidence (Huxtable et al., 1977). This is due to several factors, not the least of which is the difficulty inherent in determining the etiology of a chronic disease that often has a long latency period. In addition, the toxic effects have characteristics similar to those of other diseases, such as chronic alcoholism; pyrrolizidine alkaloids are seldom analyzed in dietary supplements and traditional remedies; and there has been a lack of medical inquiry stemming from the misconception that natural compounds are healthful rather than potentially harmful. This is not intended to be an exhaustive review of all aspects of pyrrolizidine alkaloids. For that, I recommend Robin Mattock’s book Chemistry and Toxicology of Pyrrolizidine Alkaloids (Mattocks, 1986). Although it was published more than 15 years ago, it remains the definitive source for in-depth descriptions of nearly all aspects of pyrrolizidine alkaloids. II. PLANT SOURCES

Pyrrolizidine alkaloids are produced by thousands of species of flowering plants in several higher plant families, but the genera responsible for most of the outbreaks of human poisonings are in Fabaceae (aka Leguminosae),

PYRROLIZIDINE ALKALOIDS IN FOODS

63

TABLE I SOME PLANTS CONTAINING (OR SUSPECTED OF CONTAINING) PAs WHICH HAVE BEEN USED AS EITHER HERBAL MEDICINES OR FOODS

Plant Apocynaceae Holarrhena antidysenterica Asteraceae (Compositae) Adenostyles alliariae Ageratum conyzoides Cacalia decomposita (matarique) Cacalia hastate, hupensis C. yatabei Chromolaena odoranta Eupatorium cannabinum, chinense, fortunei, japonicum Farfugium japonicum Liatris punctata

Medicine (M) or food (F)

Country or region

Reference

M

Sri Lanka

Arseculeratne et al. (1981)

M M M

Italy China United States

Sperl et al. (1995) Roeder (2000) Sullivan (1981)

M F M M

China Japan China China

Roeder (2000) Hikichi et al. (1978) Roeder (2000) Roeder (2000)

M M

Furuya et al. (1971) Mead et al. (1992)

Ligularia dentata Packera candidissma

F M

Petasites japonicus Senecio abyssinicus

F, M M

Japan Southwestern US Japan Mexico, Southwestern US Japan Nigeria

S. aureus S. bupleuroides

M M

United States Africa

S. burchelli S. coronatus S. discolor

F, M M M

South Africa South Africa Jamaica

S. doronicum S. inaequidens S. jacobaea (ragwort)

M F M

Germany South Africa Europe

S. longilobus (S. douglassi)

M

United States

S. monoensis S. nemorensis ssp. fuchsii S. pierotii S. retrorsus (S. latifolius) S. vulgaris (common groundsel) Trichodesma africana

M M F M F M

United States Germany Japan South Africa Japan Asia

Asada and Furuya (1984) Bah et al. (1994) Hirono et al. (1973) Williams and Schoental (1970) Wade (1977) Watt and Breyer-Brandwijk (1962) Rose (1972) Rose (1972) Asprey and Thornton (1955) Roder et al. (1980) Rose (1972) Schoental and Pullinger (1972); Wade (1977) Stillman et al. (1977); Huxtable (1979) Huxtable (1980) Habs et al. (1982) Asada and Furuya (1982) Rose (1972) Hikichi and Furuya (1976) Omar et al. (1983)

64

R. A. COULOMBE TABLE I (continued)

SOME PLANTS CONTAINING (OR SUSPECTED OF CONTAINING) PAs WHICH HAVE BEEN USED AS EITHER HERBAL MEDICINES OR FOODS

Plant Tussilago farfara (coltsfoot) Boraginaceae Anchusa officinalis Arnebia euchroma Cordia myxa Cynoglossum amabile, lanceolatum, zeylanicum C. geometricum C. officinale Hackelia floribunda (western false forget-me-not) Heliotropium eichwaldii

Medicine (M) or food (F)

Country or region

Reference

M M

Japan; China? Norway

Culvenor et al. (1976) Borka and Onshuus (1979)

M M M M

Europe China China China

Broch-due et al. (1980) Roeder (2000) Roeder (2000) Roeder (2000)

M M M

East Africa Iran United States

Schoental and Coady (1968) Coady (1973) Hagglund et al. (1985)

M

India

F, M M

Datta et al. (1978a, b); Gandhi et al. (1966) India; Greece IARC (1976) India, Africa, Schoental and Coady South American, (1968); Hoque et al (1976) China Roeder (2000) Arabia Macksad et al. (1970); Coady (1973) Tanzania Schoental and Coady (1968) Japan (and Hirono et al. (1978) elsewhere) General Hills (1976) Europe Roeder et al. (1992)

M, F M

Sri Lanka; India Arseculeratne et al. (1981) China Roeder (2000)

F M

East Africa Jamaica

C. incana C. juncea

M M, F

East Africa India

C. laburnifolia

M

Tanzania

C. mucronata

F M

Asia Tanzania

H. europaeum H. indicum

M M

H. ramossissimum (ramram)

M

H. supinum Symphytum officinale (comfrey) S. x uplandicum S. x uplandicum

M F, M

Fabaceae (Leguminosae) Cassia auriculata Crotalaria albina, assamica, mucronata, sessiliflora, tetragona C. brevidens C. fulva

Coady (1973) Barnes et al. (1964); McLean (1970, 1974) Schoental and Coady (1968) Chopra (1933); Watt and Breyer-Brandwijk (1962) Schoental (1968) Coady (1973) Coady (1973)

PYRROLIZIDINE ALKALOIDS IN FOODS

65

TABLE I (continued) SOME PLANTS CONTAINING (OR SUSPECTED OF CONTAINING) PAs WHICH HAVE BEEN USED AS EITHER HERBAL MEDICINES OR FOODS

Plant

Medicine (M) or food (F)

Country or region

Reference

C. recta

M, F

Tanzania

C. retusa

M, F

Africa; India

C. verrucosa

M

Sri Lanka

Schoental and Coady (1968); Coady (1973) IARC (1976); Watt and Breyer-Brandwijk (1962) Arseculeratne et al. (1981)

Orchidaceae Liparis nervosa

M

China

Roeder (2000)

Scrophulariaceae Castilleja integra

M

Mead et al. (1992)

C. rhexifolia

M

Pedicularis sp.

M

Southwestern US Southwestern US Southwestern US

Stermitz and Suess (1978) Schneider and Stermitz (1990)

Updated from Mattocks (1986).

Asteraceae (Compositae) and Boraginaceae (Table I). Pyrrolizidine alkaloidproducing plants are common worldwide, and are considered noxious weeds in many geographical areas. So abundant are pyrrolizidine alkaloids that they are estimated to be in 3% of the world’s flowering plants (Winter and Segall, 1989). While alkaloids are found in many parts of the plant, they are typically associated with seeds or fruit, but other parts may also contain pyrrolizidine alkaloids. In fact, the highest recorded pyrrolizidine alkaloid content of any plant, 18% dry weight, was found in the leaves of Senecio riddellii (Molyneux and Johnson, 1984). III. CHEMICAL STRUCTURES OF PYRROLIZIDINE ALKALOIDS

The majority of pyrrolizidine alkaloids, and all of the ones discussed in this review, contain an eight-membered necine base, which can either be saturated or contain a double bond in the 1,2-position (Figure 1). The presence of this double bond is an important determinant in the hepatoxicity of pyrrolizidine alkaloids in that only those with 1,2-unsaturation are hepatotoxic. Other variants of the necine base are also found, such as the

66

R. A. COULOMBE

FIG. 1. Chemical structures of the various component features of pyrrolizidine alkaloids and their N-oxides.

discontinuous ring structure, otonecine, in petasitenine and senkirkine (Figure 2). The necine moiety is often esterifed to constituents called necic acids, which vary significantly in structure. Necic acids may be absent (as in retronecine; Figure 2) or they may be present as open esters (heliosupine) or as macrocyclic esters (senecionine, monocrotaline, petasitenine and senkirkine). Other features of necic acids include an epoxide (petasitenine), saturation (monocrotaline) or unsaturated carbon nuclei (senecionine). Pyrrolizidine alkaloids with a cyclic diester moiety are generally more potently toxic, and are more likely to form more cellular DNA cross-links than the other forms (Kim et al., 1993, 1995). Usually coexisting in the plant with pyrrolizidine alkaloids, often in greater quantities, are N-oxides, such as indicine N-oxide (Figure 2), which often may represent the majority of the total pyrrolizidine alkaloid content. N-Oxides usually become reduced to the basic alkaloids during the process of extraction. IV. PYRROLIZIDINE ALKALOIDS IN FOODS AND HERBAL MEDICINES

Since pyrrolizidine alkaloid-containing plants are distributed worldwide, poisoning incidences are, predictably, seen in many geographical locations.

PYRROLIZIDINE ALKALOIDS IN FOODS

67

FIG. 2. Chemical structures of several representative pyrrolizidine alkaloids. Some pyrrolizidine alkaloids have necic acids esterified to the necine base. Necic acids may be absent (retronecine), or may be present as open esters (heliosupine) or as macrocyclic esters (senecionine, monocrotaline, petasitenine and senkirkine). Other possible features of necic acids may include an epoxide (petasitenine), saturation (monocrotaline) or unsaturated carbon nuclei (senecionine). N-Oxides, represented here by indicine N-oxide, usually coexist with the reduced pyrrolizidine alkaloid in the plant.

68

R. A. COULOMBE

The majority of reports of outbreaks of pyrrolizidine alkaloid poisoning have largely been limited to third world countries. Generally, these have been outbreaks where hundreds, and sometimes thousands, have been poisoned from eating staple foods made from cereal crops contaminated with seeds from pyrrolizidine alkaloid-containing weeds. More recently, however, along with an increasing reliance on unconventional medicine and the use of herbal supplements and traditional medicines, there has been a sharp rise in the number of poisonings in industrialized countries. A survey of some of the plants reported to be responsible for incidences of human poisonings, whether from foods or medicines, is presented in Table II. A. PYRROLIZIDINE ALKALOIDS IN FOODS

Numerous sporadic outbreaks of pyrrolizidine alkaloid intoxications have been recorded (Table II). These are largely attributed to consumption of staple small grains contaminated with seeds from Senecio, Heliotropium,

TABLE II SOME LARGE-SCALE POISONING OUTBREAKS CAUSED BY PA-CONTAMINATED FOODS OR BY HERBAL MEDICINES

Plant

PA (if identified)

Country or region

Senecionine?

South Africa

Senecio ilicifolius; S. burchelli Senecio spp.

–

Heliotropium lasiocarpum Crotalaria fulva

Heliotrine; lasiocarpine Fulvine

Central Asia West Indies Central Asia Afghanistan

Crotananine; cronaburmine Heliotropium popovii Heliotrine

India

Adapted from Mattocks (1986).

80 12

Heliotropium lasiocarpum Heliotropium popovii Heliotrine Crotalaria nana

Number of cases

Tajikistan

28 Endemic 61 7800 67 3906

Reference Willmont and Robertson (1920) Selzer and Parker (1951) McLean (1970) Bras et al. (1954); Bras and Hill (1956) McLean (1970) Mohabbat et al. (1976); Tandon et al. (1978) Tandon et al. (1976), Siddiqi et al. (1978) Chauvin et al. (1993); Mayer and Luthy (1993)

PYRROLIZIDINE ALKALOIDS IN FOODS

69

Crotalaria and other genera of pyrrolizidine alkaloid-containing plants. As with many episodes of food intoxication caused by naturally occurring toxins from plants and fungi, they are often associated with consumption of a contaminated staple food supply that is locally grown. Pyrrolizidine alkaloids are one of the few classes of natural toxins to be regulated in foods or herbal supplements. In 1992, the German Federal Health Agency established a maximum allowable daily intake of 0.1 µg of pyrrolizidine alkaloids and N-oxides in herbal supplements. One µg per day is allowed if the intake is limited to 6 weeks per year (German Federal Health Bureau, 1992; Edgar and Smith, 2000). The United States has no such regulation. Currently, there is no limit on intake of pyrrolizidine alkaloidcontaining foods, herbal supplements or other products, or the amount of pyrrolizidine alkaloids that a herbal supplement can contain. Pyrrolizidine alkaloids enter the food chain in a variety of ways – through grain, milk, honey, eggs and herbal medicines. Because these items can at times be contaminated at concentrations that exceed the German regulations, a real risk is posed to human health by pyrrolizidine alkaloids in foods.

1. Pyrrolizidine alkaloids in staple foods The first report of human poisonings due to pyrrolizidine alkaloids recorded in scientific literature occurred in 1918 in the George District of Cape Province, South Africa (Willmont and Robertson, 1920). Dubbed “Senecio disease,” it was apparently caused by contamination of cereal grains with seeds of Senecio ilicifolius and S. burchelli, which were made into bread. The problem appeared to be exacerbated by “old-fashioned” mills that did not efficiently winnow the grain, and the fact that bread was the staple diet of the largely poor populace. The outbreak was responsible for some 80 cases over a 10-year period, most of whom were children. The symptoms, which were often fatal, included nausea, vomiting, acute gastric pain, ascites and hepatic distension, and had an onset of 2 weeks to 2 years or longer. The investigators noted similarities in symptoms and post-mortem signs between this outbreak and Moltendo disease in South Africa, Winton’s disease in New Zealand, and Picton disease in Nova Scotia, all diseases in livestock known to be caused by Senecio. Seltzer and Parker (1951) published details of clinical and post-mortem findings of Senecio poisoning in Cape Town, South Africa, similarly attributed to contamination of wheat with seeds from Senecio sp. Of the 12 cases described, six were fatal. The most common and most constant early symptoms in all patients were abdominal pain and swelling, followed by rapidly developing ascites and hepatomegaly. Surviving victims recalled

70

R. A. COULOMBE

that the tainted bread had an abnormal taste, some describing it as “musty,” others as “bitter.” These investigators noted the similarity of these signs to those of Budd–Chiari syndrome, a hepatic disease characterized by obstruction of the trunk and large branches of the hepatic vein. An outbreak of veno-occlusive disease (VOD) that was probably caused by consumption of cereals contaminated with seeds of Crotalaria sp. containing pyrrolizidine alkaloids occurred in the Sarguja district of India in November–December 1975 (Tandon et al., 1976). Forty-two per cent of the 67 recorded cases died. A follow-up study reported post-mortem signs characteristic of pyrrolizidine alkaloid poisoning, which included changes ranging from acute hemorrhagic centrilobular necrosis, progressive sclerosis to nonportal cirrhosis, and in the terminal phases, occlusion of the terminal veins (Tandon et al., 1977). A much larger outbreak of VOD occurred in northwestern Afghanistan. This was caused by consumption of bread made from wheat contaminated with seeds of Heliotropium, which contained heliotrine (Figure 3) (Mohabbat et al., 1976; Tandon et al., 1978). Approximately 7800 were affected. In a small number of cases, clinical improvement was observed after several months of hospitalization. Another extensive outbreak of pyrrolizidine alkaloid poisoning occurred in 1992 in southern Tajikistan, caused by consumption of bread contaminated with Heliotropium popvoii (Chauvin et al., 1993; Mayer and Luthy, 1993). The outbreak was exacerbated by political instability that led to a blockade of the Farkhar region, causing a 2-month delay of the wheat harvest and subsequent famine. Because of this delay, Heliotropium had time to go to seed, and then contaminate the wheat at harvest. In the area affected by the outbreak, a poisoning rate of 4% occurred, and a total of 3906 cases of VOD was recorded. The overall fatality rate was approximately 1.3%. Attack rates favored younger children and adolescents, while the case–fatality rate increased with age. The most dominant alkaloid in samples of contaminated wheat was heliotrine (Mayer and Luthy, 1993). 2. Pyrrolizidine alkaloids in milk Another potential source of pyrrolizidine alkaloids in the human food chain is from the milk of animals that have ingested pyrrolizidine alkaloidproducing plants. The pyrrolizidine alkaloid, jacoline, was detected in the milk of four lactating cows given Senecio jacobaea (tansy ragwort) via rumen canula at the dose rate of 10 g kg–1 day–1 for 2 weeks. The plants, which contained five pyrrolizidine alkaloids, had an average pyrrolizidine alkaloid content of 0.16% dry weight. The dosed cows exhibited marked changes in blood leukocyte count, sorbitol dehydrogenase values, and mild

PYRROLIZIDINE ALKALOIDS IN FOODS

71

FIG. 3. Selected pyrrolizidine alkaloids found in folk medicines and foods that have been involved in human poisoning incidents. Fulvine, from Crotalaria fulva, an ingredient in some “bush teas,” is responsible for endemic veno-occlusive disease (VOD) in the West Indies. Heliotrine, from Heliotropium popvoii and spp., is the etiological agent of large endemics of VOD in Afganistan and Tajikistan. Seneciphylline, a potent hepatotoxin in Senecio-based teas, is extensively used in southwestern United States and Northern Mexico, where it is known under various names. It has also been found in honey. Intermedine is one of the major hepatotoxic alkaloids found in the widely used traditional herb comfrey Symphytum officinale and Siberian comfrey S. uplandicum. Lycopsamine is another major alkaloid in comfrey. Echimidine was found in commercial Australian honey made from bees pollinating Echium plantagineurm. Australian researchers recently found lasiocarpine, among other alkaloids, in commercial eggs at concentrations ranging from 1.2 to 9.7 µg of total alkaloid per egg. Chickens were given a wheat-based feed grain contaminated with Heliotropium europaeum. Intermedine, lycopsamine and senecionine were found in Liatris punctata, also known as “cachana,” “grey feather” and “blazing star.” This herb is traditionally used by some Native Americans and Hispanics in the southwestern United States as a diuretic, for throat inflammation, laryngitis, cough and nosebleeds.

72

R. A. COULOMBE

liver pathology, consistent with hepatotoxicity. However, no such changes were noted in calves nursing from these cows. Jacoline was found in concentrations ranging from 9.4 to 16.7 µg per 100 mL of milk (Dickinson et al., 1976). Milk from goats fed tansy ragwort (1% body weight per day) caused either a negative or a mildly positive mutagenic response in the Salmonella/mammalian microsome (Ames) assay (White et al., 1984). Another study confirmed that transfer of pyrrolizidine alkaloids in milk can cause hepatic toxicity. Rats fed a diet of milk from goats fed tansy ragwort (7.5 ng of pyrrolizidine alkaloid per g dry weight) for 180 days with a calculated total pyrrolizidine alkaloid intake of 0.96 mg per rat had swollen hepatocytes of centrilobular distribution and biliary hyperplasia, indicating hepatic toxicity due to pyrrolizidine alkaloids (Goeger et al., 1982). Importantly, these hepatic signs were similar to those found in rats fed tansy directly at concentrations of 1%, 0.1%, 0.01% and 0.001% (corresponding to pyrrolizidine alkaloid intakes of 39.77, 5.04, 0.52, and 0.05 mg per rat). In this study, no changes were noted in two calves fed the pyrrolizidine alkaloid-containing goats’ milk. 3. Pyrrolizidine alkaloids in honey Honey is another food source found to contain pyrrolizidine alkaloids. Initial investigations found pyrrolizidine alkaloids at relatively low concentrations of 0.3–3.9 ppm in honey produced from bees foraging fields of S. jacobea (Deinzer et al., 1977). The type of alkaloids found in the honey – senecionine, seneciphylline, jacoline, jaconine, jacobine and jacozine – reflected those found in the local plant. In one study, Senecio pollen counts in honey samples were correlated with pyrrolizidine alkaloid concentrations. Honey that contained S. jacobea pollen (15–21 grains g–1) with total alkaloid concentrations of 0.011–0.056 mg kg–1 contained jacobine, jacozine, seneciophylline and senecionine. Honey samples containing two grains or less did not contain detectable alkaloids (Crews et al., 1997). These authors did not detect pyrrolizidine alkaloids in honey purchased in retail outlets. Australian researchers discovered that honey produced from stands of the purple flower Echium plantagineurm, known as “Patterson’s Curse” or “Salvation Jane,” contained a pyrrolizidine alkaloid content ranging between 0.27 and 0.95 ppm (Culvenor et al., 1981). The main alkaloid found in honey obtained from four suppliers in Victoria and New South Wales contained echimidine, with smaller amounts of 7-acetyllycopsamine, 7-acetylintermedine, echiumine, uplandicine, lycopsamine, intermedine (Figure 3) and a novel alkaloid. Given usual patterns of honey consumption and the relatively low concentrations of alkaloids found, there is probably little, if any, human health risk posed by pyrrolizidine alkaloids in honey.

PYRROLIZIDINE ALKALOIDS IN FOODS

73

4. Pyrrolizidine alkaloids in eggs While there have been many instances of poultry being poisoned by pyrrolizidine alkaloids in their feed, one report has indicated that pyrrolizidine alkaloids can enter the human food chain via eggs (Edgar and Smith, 2000). Three flocks of chickens from a small-scale egg producer were given wheat-based feed grain contaminated with Heliotropium europaeum, at a concentration estimated to be 0.6% by weight. Analysis of the wheat uncovered the presence of the pyrrolizidine alkaloids supinine, heleurine, heliotrine, europine and lasiocarpine. These pyrrolizidine alkaloids (as well as others, probably as a result of metabolism) were found at concentrations ranging from 1.2 to 9.7 µg per egg. 5. Pyrrolizidine alkaloids in meat Residues of pyrrolizidine alkaloids have not been found in meat from animals ingesting alkaloid-containing plants. There are several possible reasons for this. These compounds are metabolized into reactive pyrrolic intermediates with relatively short half-lives, which rapidly bind to cellular macromolecules, such as GSH, proteins and DNA. Thus, it is unlikely that sufficiently large amounts of residues of pyrrolizidine alkaloids per se would exist in meat products to be of any risk to the consumer. One possible exception to this would be if an animal is slaughtered only hours after grazing on contaminated pasture. In any event, it is not likely that pyrrolizidine alkaloids in meat pose a significant health threat. B. PYRROLIZIDINE ALKALOIDS IN TRADITIONAL REMEDIES AND MEDICINES

In North America, where crop weeds are controlled by herbicides and foods regularly inspected, the biggest risk of exposure to pyrrolizidine alkaloids comes primarily from so-called herbal “remedies,” herbal teas and folk medicines (Huxtable, 1989). In contrast to the sporadic outbreaks of poisonings in many third world countries, poisonings from consumption of pyrrolizidine alkaloid-contaminated botanical folk medicines account for a consistent, regular occurrence of cases. Taken as a whole, the majority of herbal drinks are harmless and in some cases may even be beneficial. However, some herbal teas that are widely sold in health food stores contain pyrrolizidine alkaloids that have been implicated in many cases of human poisonings. The serious health risk posed by pyrrolizidine alkaloid-containing plant-based supplements compelled the German Federal Health Bureau to

74

R. A. COULOMBE

establish regulations limiting intake. Unfortunately, in the United States, recent legislation has moved in the opposite direction, away from consumer protection. The ironically misnamed “Dietary Supplement Health and Education Act” of 1994, now exempts “natural” food products such as folk medicines and teas from safety regulations that normally apply to prescription drugs. In fact, this law placed dietary supplements into a new category, distinct from food or drugs, that is now exempt from the rules the US Food and Drug Administration used against questionable products. Protected by this legislation, manufacturers and dispensers of herbal supplements are not required to provide information on content, effectiveness (or lack thereof), or possible adverse effects. Furthermore, in the case of botanical supplements, there is usually no reliable information available about how and where the plant was collected, or what part of the plant went into the product. Such factors have a bearing on the content of potentially toxic compounds. Compounded by the commonly held belief that equates “natural” with health and well-being, the popularity of botanicals used for dietary supplements has dramatically increased in industrialized countries. The occurrence of liver disease resulting from ingestion of herbal teas containing pyrrolizidine alkaloids have been known for some time. There is a longstanding tradition of medicinal herb use in Jamaica and the West Indies (Asprey and Thornton, 1955). In the 1950s, Bras and co-workers reported on endemic VOD in Jamaica (Bras et al., 1954). Liver biopsies, taken mostly from children, showed nonportal liver pathology typical of pyrrolizidine alkaloid ingestion with the key features of inflammation and veno-occlusion with intact hepatic portal triads, termed “nonportal” cirrhosis. More than 50% of the patients affected in Jamaican outbreaks fully recovered after sodium and fluid restriction. In the West Indies, acute VOD, which largely affects children, is associated with consumption of folk medicines made from Crotalaria fulva, which contains the pyrrolizidine alkaloid fulvine (Figure 3) (McLean, 1970). These medicines, or “bush teas,” which are hot-water infusions prepared from various plants, are generally bitter tasting, and are traditionally given to children as a tonic for ailments as diverse as a cold, teething pain and gastric problems. A number of reports described clinically evident liver diseases resulting from either short- or long-term use of pyrrolizidine alkaloid-containing herbs. Probably the most popular and widely used pyrrolizidine alkaloidcontaining medicinal herb is comfrey (Symphytum). Since Greek and Roman antiquity, this herb has been part of the offical pharmacopeia of many cultures, and has been used as a “cure-all” for a variety of ailments and complaints. In fact, the species name officinale indicates its place in medieval herbs (Huxtable, 1989). Comfrey and Russian comfrey (S. uplandicum) are cultivated throughout Europe, North America and Australia.

PYRROLIZIDINE ALKALOIDS IN FOODS

75

The name comfrey may have been derived from the Latin confirmare, “to strengthen,” owing to its reputed ability to promote overall health and well-being, as well as heal fractured bones more quickly. Other uses for comfrey have included external use for wounds, and internal use for joint inflammation, gout, haematomas, gastritis, diarrhea, rheumatoid arthritis, bronchitis, backache and various allergies (Stickel and Seitz, 2000). Comfrey can also be found in a range of cosmetics and personal-care products, such as shampoo, skin creams, bath oils and various ointments. Another popular form is a comfrey-pepsin capsule marketed as a digestive aid. According to Huxtable (1989), one preparation claiming to contain comfrey leaves contained 40 mg kg–1 pyrrolizidines and 230 mg kg–1 pyrrolizidine N-oxides, while another claiming to contain comfrey root had a total content of 2900 mg kg–1, consisting of 400 mg kg–1 pyrrolizidines and 2500 mg kg–1 of the N-oxides. As people can take several of these capsules per day, these products represent a major source of exposure to pyrrolizidines. Comfrey contains a mixture of pyrrolizidine alkaloids, including intermedine, acetylintermedine, lycopsamine, acetyllycopsamine, symphytine, echimidine and symviridine, all of which are hepatotoxic (Huxtable, 1989). The overall alkaloid content of the plant varies from 0.003 to 0.2% for dry leaves and 0.2 to 0.4% for root (Roitman, 1981). These variations are mostly due to species, age of plant, location and season collected. Roitman found 8.5 mg of total alkaloids in a cup of comfrey root tea. Mattocks observed that because the leaf contains a relatively low concentration of alkaloids (which are largely N-oxides), and because the acute toxicity in rats of comfrey alkaloids is approximately half that of other pyrrolizidine alkaloids such as heliotrine (Culvenor et al., 1980), it is unlikely that people can be acutely poisoned by drinking comfrey tea (Mattocks, 1986). He predicted that a person weighing 60 kg would require 700 cups of comfrey tea (with 8 mg total alkaloids) to be so poisoned. However, poisoning by pyrrolizidine alkaloids is accumulative and generally chronic. Predictably, chronic comfrey use has been involved in a number of poisonings in people. Ridker et al. (1985) reported hepatic VOD and centrilobular necrosis in a 49-year-old woman who was a “heavy consumer” of herbs, vitamins, and natural food supplements and a regular drinker of commercial comfrey tea. Additionally, she took six comfrey-pepsin capsules per day. From these sources, and from the regularity of use, the authors calculated that she had consumed 15 µg kg–1 day–1 or a minimum of 85 mg of pyrrolizidine alkaloids in the 4-months previous to her admission to hospital. In another case, a 13-year-old boy was diagnosed with classic VOD who had for 2–3 years regularly consumed comfrey tea as a home treatment prescribed by a homeopath (Weston et al., 1987). The authors were con-

76

R. A. COULOMBE

fident that comfrey ingestion was the only plausible explanation for this condition. That regular use of comfrey products has serious health consequences was confirmed in a later clinical report of hepatic VOD diagnosed in a 47-year-old woman who was given a prescription of comfrey from a homeopathic doctor to cure abdominal pain, fatigue and allergies (Bach et al., 1989). The patient consumed as many as 10 cups of comfrey tea per day, as well as comfrey-pepsin capsules “by the handful.” A fatal case of hepatic VOD associated with regular comfrey use was reported in a 23-year-old man who ate 4–5 steamed comfrey leaves 1–2 weeks before the onset of symptoms (Yeong et al., 1990). Other possible causes of VOD were excluded by these investigators. In another report, McDermott and Ridker described classic VOD associated with heavy comfrey use (McDermott and Ridker, 1990). Other, less common, medicinal plants that contain pyrrolizidine alkaloids have been reported to cause poisonings. Daily consumption of the pyrrolizidine alkaloid-containing herbal medicines by a pregnant woman resulted in transmission of fatal VOD to her newborn infant (Roulet et al., 1988). The responsible pyrrolizidine alkaloid was identified as senecionine from the herb Tussilago farfara, which the mother purchased from a pharmacy. This herb, which is also known as coltsfoot, or the Old English designation coughwort, has been used since antiquity as a cough suppressant. Assuming a pyrrolizidine alkaloid content in the tea of 0.6 mg per kg dry weight, the authors calculated a cumulative transplacental exposure to the baby of approximately 0.125 mg total pyrrolizidine alkaloids per kg body weight. This was the first report of transplacental transmission of a pyrrolizidine alkaloid in people. It was later reported that the tea involved in this fatal case of poisoning also contained roots of Petasites officinalis (Spang, 1989). Veno-occlusive disease was diagnosed in an 18-month-old boy who had regularly consumed a herbal tea mixture since the 3rd month of life (Sperl et al., 1995). The tea, given to the boy to promote “healthy development,” was gathered by the boy’s mother in her own garden. It consisted of peppermint leaves and Adenostyles alliariae or “Alpendost,” which the mother misidentified as coltsfoot (also likely to be toxic!). The two plants can easily be confused, especially after the flowering period. Analysis of the tea mixture revealed high amounts of seneciphylline (Figure 3) and its N-oxide. The authors calculated that the child had consumed at least 60 µg per kg body weight per day of pyrrolizidine alkaloids over 15 months. Fortunately, the child recovered completely within 2 months following conservative treatment. Several wildflowers native to the southwestern United States and northern Mexico in common use as folk medicines have been shown to

PYRROLIZIDINE ALKALOIDS IN FOODS

77

contain pyrrolizidine alkaloids. There have been several poisoning incidents in which these plants are implicated. A 6-month-old girl regularly given home-brewed tea made from Senecio longilobus developed VOD which progressed over 2 months to extensive hepatic fibrosis (Stillman et al., 1977). Tea made from this plant had 3 mg g–1 of pyrrolizidine alkaloids, and 10.5 mg g–1 of N-oxides. By preparing a tea made according to the mother’s recipe, the authors calculated that during the 2 weeks before admission, the child received between 70 and 140 mg of pyrrolizidine alkaloids, clearly a toxic dose. Indigenous to the southwestern United States and northern Mexico, and known by Hispanics as “gordolobo yerba,” hot-water infusions of Senecio longilobus are used as a gargle and cough suppressant. Another medicinal plant native to the southwestern United States, Pedicularis, also known as “betony,” “lousewort,” “Indian Warrior,” and “Elephant Head,” among other appellations, contains pyrrolizidine as well as toxic quinolizidine alkaloids through transfer by root parasitism from alkaloid-containing host plants, such as Senecio triangularis (Schneider and Stermitz, 1990). The major alkaloid found in this plant was senecionine. This is in stark contrast with the commonly held view that this plant is safe for use by children. For example, this plant was described in the “Bible” of herbal remedies entitled Medicinal Plants of the Mountain West (Moore, 1979) in this way: Betony is an effective sedative for children [my emphasis]. It acts as a mild relaxant … quieting anxieties and tension … Large quantities may cause a befuddled lethargy and some interference of motor control … It wouldn’t hurt to test a particular collection before administering freely, since the potency of various species is variable. Castellija sp., or Indian paint brush, a native southwestern plant that has some traditional medicinal use, may contain pyrrolizidine alkaloids through root parasitism from another medicinal pyrrolizidine alkaloidcontaining plant, Liatris punctata (Mead et al., 1992). Also known as “cachana,” “grey feather” and “blazing star,” Liatris is used as a diuretic, to treat throat inflammation and laryngitis, and as a cough suppressant. Burning Liatris root is inhaled as a cure for nosebleeds and tonsillitis (Moore, 1979). Liatris punctata was found to contain senecionine, intermedine and lycopsamine, in addition to a novel open-diester pyrrolizidine alkaloid, punctanecine (Mead et al., 1992). Another medicinal Indian paint brush, C. rhexifolia, contains pyrrolizidine alkaloids (Stermitz and Suess, 1978). Senecionine and other hepatotoxic pyrrolizidine alkaloids and N-oxides (0.36% in aerial parts, 0.76% in roots) were found in Packeria candidissima, a Mexican medicinal herb commonly used to treat kidney ailments, ulcers,

78

R. A. COULOMBE

and for its antiseptic properties (Bah et al., 1994). Among Mexicans in rural and urban areas of the state of Chihuahua and by Hispanics in southwestern United States, Packeria is commonly known as “chucaca,” “lechuguilla de la sierra,” “té de milagros” and “hierba de milagro.” There are several other reports of fatal and non-fatal liver disease resulting from use of various pyrrolizidine alkaloid-containing herbal products (Lyford et al., 1976; McGee et al., 1976; Margalith et al., 1985; Jones and Taylor, 1989). V. TOXICITY OF PYRROLIZIDINE ALKALOIDS

Many pyrrolizidine alkaloids are carcinogenic in a number of animal models. (Schoental, 1968; Hirono, et al. 1973, 1978), although there are not sufficient data to conclude that they are carcinogenic in people. Much of the quantitative data on the toxicity of pyrrolizidine alkaloids are from shorter-term lethality studies. These data can give some indication about the relative toxicity of various pyrrolizidine alkaloids. There is a considerable difference in the toxicity among pyrrolizidine alkaloids (Table III). For example, the more potent alkaloids to which people are exposed, such as trichodesmine, senecionine and seneciphylline, have acute lethal toxicity (i.e. LD50) values of 25, 50 and 77 mg kg–1, respectively, while for heliotrine, lycopsamine and heliotrine N-oxide the LD50 values are 296, >1000 and c. 5000, respectively (Table III). Monocrotaline is in the middle range of these toxins. Similarly, there is a considerable range of susceptibility among animal species to pyrrolizidine alkaloids. Acute lethal toxicity of retrorsine, a TABLE III COMPARATIVE ACUTE TOXICITY OF SOME PYRROLIZIDINE ALKALOIDS

Alkaloid Trichodesmine Senecionine Seneciphylline Lasiocarpine Monocrotaline Heliotrine Lycopsamine Heliotrine N-oxide a

Male rats, intraperitoneal (i.p.) administration. Adapted from Mattocks (1986) and references therein.

LD50 (mg kg–1, i.p.)a 25 50 77 77 109 296 >1000 c. 5000

PYRROLIZIDINE ALKALOIDS IN FOODS

79

12-membered α,β-unsaturated pyrrolizidine alkaloid similar to senecionine, varies from 34 mg kg–1 in rats to 279 mg kg–1 in quail and over 800 mg kg–1 in guinea pigs (White et al., 1973). Like most other toxins, a major factor in species susceptibility to pyrrolizidine alkaloids is the ability of that animal to metabolize the parent compound into the active, electrophilic intermediate. Generally, susceptible animals form more of the reactive pyrrole than do resistant animals. Some evidence indicates that pyrrolizidine alkaloids may be degraded and hence detoxified by rumen microflora in resistant species. Postulated as a possible etiology of Indian childhood cirrhosis, which is characterized by excessive hepatic copper accumulation (Morris et al., 1994; Aston et al., 1996), copper and pyrrolizidine alkaloids have been shown to exhibit true hepatotoxic synergism in a number of animal models. In rats, retrorsine and copper given concurrently result in significantly greater mortality, liver pathology and hepatic copper accumulation compared to that seen when either of these agents was administered alone (Morris et al., 1994). Likewise, heliotrope hepatotoxicity in sheep was markedly enhanced by concurrent or subsequent administration of copper (Howell et al., 1991). Those authors noted that hepatic copper concentrations were higher in sheep given heliotrope + copper compared to those given the same amount of copper alone. Serum copper concentrations have been used as a noninvasive indicator for monocrotaline-induced cardiopulmonary toxicity (Molteni et al., 1988). Retrorsine passing to rat neonates via breast milk results in the accumulation of hepatic copper, an impairment of the rise in serum ceruloplasmin, and a decrease in hepatic metallothionein and serum albumin levels (Aston et al., 1996). The authors suggested that accumulation of liver copper and reduction of copper-binding proteins could result in an increase of free copper that might enter into pro-oxidant Fenton-type reactions, resulting in the generation of oxygen free radicals. The mechanism of the synergy between copper and pyrrolizidine alkaloids probably centers around the fact that both compete for the same pool of cellular glutathione (GSH). Copper uptake and incorporation into metallothionein is intimately associated with reduced GSH, and GSH is generally thought to act as a protectant against oxidant damage (Freedman et al., 1989). In fact, depletion of GSH potentiates metal toxicity in a variety of animals and cell systems (Freedman et al., 1989). Similarly, GSH depletion resulted in an increase of the release of active pyrroles from rat liver, which are then available for macromolecular alkylation and toxicity (Yan and Huxtable, 1995). As a group, pyrrolizidine alkaloids are hepatotoxic, but depending on the alkaloid, other extra-hepatic effects may result from ingestion. One important distinction of pyrrolizidine alkaloid poisoning is that it is

80

R. A. COULOMBE

progressive, once it has been initiated by exposure to even a single moderate dose. The most prevalent pyrrolizidine alkaloid-caused diseases in people are hepatic VOD, pulmonary hypertension and cor pulmonale, or congestive right heart failure (Huxtable, 1990). Hepatotoxicity commonly results from pyrrolizidine alkaloids found in Senecio species, such as senecionine, retrorsine, seneciphylline and riddelliine, while cardiopulmonary toxicity is more likely to result from Crotalaria alkaloids, such as monocrotaline. Trichodesmine, an alkaloid very similar to monocrotaline (Figure 2), is a neurotoxin (Yan et al., 1995). In people, the most common clinical sequela is VOD caused by an occlusion of the smaller branches of the hepatic vein due to endothelial proliferation and medial hypertrophy (Huxtable, 1989). This occlusion leads to centrilobular congestion and a pooling of blood. Clinically, VOD can be divided into an acute, a subacute and a chronic phase. The acute phase is characterized by a rapid onset of nausea, emesis, abdominal pain, distension, portal hypertension, reduced urinary output, hepatomegaly and ascites (Stillman et al., 1977; Sperl et al., 1995). In this phase, poisoning victims may either expire, recover completely (especially if pyrrolizidine alkaloid ingestion is discontinued), or progress to the subchronic and chronic phases (Stillman et al., 1977). The subchronic phase involves persistent hepatomegaly and recurrent ascites. The chronic phase consists of cirrhosis and liver failure, and may be delayed for months or years following exposure. There is no specific antidote for pyrrolizidine alkaloid poisoning, except supportive treatment. Aside from being occasionally seen as a result of chemotherapy and bone marrow transplantation, VOD is diagnostic for pyrrolizidine alkaloid poisoning. It has a poor long-term prognosis, and death may occur anywhere from 2 weeks to over 2 years following exposure. Children are more vulnerable than adults to VOD. A small number of pyrrolizidine alkaloids, of which monocrotaline is the best example, result in cardiopulmonary toxicity. The pathology and pulmonary toxicology of monocrotaline is well studied. As with other pyrrolizidine alkaloids, the toxicity depends upon hepatic metabolic activation, which, in this case, results in the formation of monocrotaline pyrrole, or dehydromonocrotaline. Unlike hepatotoxic pyrrolizidine alkaloids such as senecionine, monocrotaline exerts its primary toxicity in an organ distant to the location of metabolic activation. The lung appears not to activate monocrotaline to any appreciable extent. Monocrotaline had no effect when perfused through isolated rat lungs, but induced pulmonary toxicity when first perfused through isolated liver (Lafranconi and Huxtable, 1984). The selectivity of monocrotaline for the lung is dependent upon the binding of the monocrotaline pyrrole by red blood

PYRROLIZIDINE ALKALOIDS IN FOODS

81

cells, where it is stabilized during transport to the lung (Wilson et al., 1992). Toxicity to the lung proceeds through an inflammatory response, for which platelet activation plays a role. In animals, this results in pulmonary hypertension in which an abrupt elevation in pulmonary arterial pressure occurs, and acute cor pulmonale. This pathology, while not known to occur in people exposed to monocrotaline, closely follows acute respiratory distress syndrome (ARDS), a syndrome known to be caused by a variety of environmental agents. As mentioned earlier, diagnosis pyrrolizidine alkaloid poisoning is often difficult, especially when the disease takes a chronic course. There are several factors that confound a correct diagnosis. These include the varying interval between ingestion and development of symptoms, and that health care providers may not query the patient about herbal supplement use. Unfortunately, there is little general appreciation that natural “organic” supplements and tonics may indeed be harmful. Furthermore, pyrrolizidine alkaloid-related diseases can easily be confused with other pathologies. For example, chronic hepatic cirrhosis induced by pyrrolizidine alkaloids is clinically indistinguishable from that caused by alcohol, infection, or any number of other etiologies. Veno-occlusive disease has been confused with viral hepatitis (Datta et al., 1978a). Pyrrolizidine intoxication may also have a few symptoms in common with Reye’s syndrome, as was documented in the fatal poisoning case of a 2-year-old child (Fox et al., 1978). Despite the fact that pyrrolizidine alkaloids are known to be carcinogenic in animals, and are genotoxic in a variety of short-term in vitro and in vivo systems, there is insufficient evidence to conclude that long-term exposure to pyrrolizidine alkaloids causes cancer in people. VI. METABOLISM OF PYRROLIZIDINE ALKALOIDS

Like thousands of other “pro-toxicants” such as aflatoxin B1, benzo[a]pyrene, and dimethylnitrosamine, pyrrolizidine alkaloids are not toxic per se, but must first be metabolized by endogenous enzymes following ingestion of the plant material. Several classes of enzymes act on pyrrolizidine alkaloids. In mammals, the most important examples are the cytochromes P450 (CYP). Cytochromes P450 are a large group of hemoprotein enzymes that are present in greatest quantities in the liver, but are also found in the lung, kidney, brain and other organs. CYPs utilize reduced NADPH + H+ as a cofactor together with NADPH cytochrome P450 reductase to reduce active site iron to the ferrous form to catalyze insertion of oxygen into the substrate. Hundreds of CYP isoforms have been isolated, each having various substrate specificities.

82

R. A. COULOMBE

Metabolic conversion of pyrrolizidine alkaloids results in several possible products with varying degrees of toxicity. The three main phase I enzymatic reactions acting on pyrrolizidine alkaloids in the liver are: hydrolysis, resulting in the formation of a free necic acid and necic base; N-oxidation, to form the N-oxides, which are of modest toxicity (this pathway is therefore considered a detoxification); and dehydrogenation to form reactive and toxic pyrroles, also called dehydroalkaloids (Figure 4). As with many other proximal, reactive intermediates (such as the aflatoxin 8,9-epoxide, and benzo[a]pyrene diol-epoxide), pyrroles are electrophilic and inherently unstable, and react quickly with endogenous nucleophilic macromolecules. The current hypothesis for the mechanism for production of the pyrrolic intermediate involves an initial CYP-mediated hydroxylation of the pyrrolizidine alkaloid at C-8, producing a chemically unstable carbinolamine, which spontaneously loses the hydroxyl group, then a proton, giving the dehydropyrrolizidine (Mattocks, 1986; Rajski and Williams, 1998) (Figure 5). There are several fates for these reactive pyrroles, which have a half-life of a few seconds. In liver tissue from people and rodents, CYP 3A4 is an important isoform catalyzing dehydrogenation. This isoform is the most prevalent in human liver (approximately 60% of total), and is inducible by compounds such as barbiturates, dexamethazone and erythromycin (Guengerich, 1989). This enzyme has broad substrate specificity; indeed, approximately half of the drugs currently on the market are substrates for CYP 3A4 (Ueng et al., 1997). The prototype activity of CYP 3A4, which is the basis of diagnosing 3A4 activity of various tissues, is the oxidative conversion of the calcium channel blocker nifedipine (NF) to its main metabolite, dehydronifedipine. Good evidence of the critical role of CYP 3A4 in pyrrolizidine alkaloid activation was provided by the observation that two mechanism-based inhibitors of human liver CYP 3A4, gestodene and triacetyloleandomycin, inhibited the conversion of senecionine to dehydrosenecionine and senecionine N-oxide in human liver microsomes (Miranda et al., 1991). N-Oxidation of pyrrolizidine alkaloids appears to be catalyzed by CYP 2C11 and to a lesser extent by CYP 3A. In hepatic microsomal preparations from guinea pig, N-oxide formation is also catalyzed by flavincontaining monooxygenases (FMO) (Williams et al., 1989). The formation of the toxic pyrroles and detoxified N-oxides occurs by different pathways. This conclusion has been reached largely from the observation that plant N-oxides are not activated to pyrroles in microsomal preparations in vitro. The actual mechanism of N-oxide formation has not been elucidated.

PYRROLIZIDINE ALKALOIDS IN FOODS

83

An important route of detoxification is mediated by the universal detoxifying enzyme, glutathione S-transferase (GST), to produce various glutathione (GSH) conjugates that are water soluble and hence more easily excretable than the parent compound (Figure 4). Glutathione S-transferases are a group of cytosolic homo- and heterodimeric proteins that catalyze the detoxification of a large number of compounds which, like pyrrolizidine alkaloids, are metabolized to electrophilic intermediates (Boyer, 1989). Like the CYPs, the GSTs are a multienzyme family with varying substrate specificities. Which GST isoforms are involved in catalysis of pyrrolic detoxification has not yet been elucidated. Various GSH conjugates were frequently the major detoxified products of pyrrolizidine alkaloids in isolated, perfused rat liver (Yan et al., 1995). Activated pyrroles may polymerize, resulting in detoxified products. A variety of cellular nucleophiles are known to react with pyrrolic pyrrolizidine alkaloids and hence may be involved in detoxifying Sn1type reactions. The thiols GSH and cysteine (Robertson et al., 1977), and thiol resins (Glowaz et al., 1992) efficiently react with pyrroles. Dietary cysteine resulted in protection against pyrrolizidine alkaloid-induced hepatoxicity compared to control rats without cysteine (Miranda et al., 1982). Glutathione and cysteine, but not methionine, competed with λ-phage DNA for pyrrolic cross-linking, indicating that thiols with a free sulfhydryl group are sufficiently reactive to compete with DNA for reaction with pyrroles (Coulombe et al., 1999). VII. MECHANISM OF TOXIC ACTION

Enzymatic oxidation of pyrrolizidine alkaloids to the pyrrole results in the production of a bifunctional intermediate with reactive, electrophilic centers at C-7 and C-9 by conjugation of the pyrrole nitrogen lone pair. These electrophilic centers react with a variety of nucleophilic cellular macromolecules, possibly the most critical of which are various nucleoside bases in DNA, resulting in the formation of cross-links between two strands of DNA (DNA–DNA cross-links) or between one strand of DNA and some cellular protein (DNA–protein cross-links; DPCs) (Figure 5). It has been postulated that nucleophilic substitution at C-7 is favored over C-9 owing to greater stablilization of the secondary carbonium ion at C-7 (Huxtable and Cooper, 2000). In cultured cells, various pyrrolic pyrrolizidine alkaloids form an approximately equal proportion of these cross-links, and are inherently DNA repair-resistant (Kim et al., 1993, 1995). A number of reports strongly support the hypothesis that the formation of DNA cross-links is an important molecular event in the toxicity of

84

OH

GS H

OH HO

GSH/GST

H

OH HO

H

dehydroretronecine

[O]

GSH/GST HO

N Retronecine + HO H3C

CH 3 O

H3C hydrolysis [H +]

O

CH3 O

CH3 O O H

[O]

N O

CH 3 OH OH

CYP 3A4mediated oxidation

Senecionine

HO H3C O

O

CH3 O

CH3 O

N

Senecic acid [O]

CYP 3A, 2C11, FM O (N-oxidation)

Dehydrosenecionine

R. A. COULOMBE

react with tissue nucleophiles (see Figure 5)

N CYP

glutathione conjugate

N

CYP 3A, 2C11, FM O (N-oxidation)

[O]

O

O

CH3 O H N O

+ –

Senecionine N-oxide FIG. 4. Metabolic fate of a representative pyrrolizidine alkaloid senecionine showing the various phase I and phase II reactions. Alkaloids are metabolized by a variety of enzymes, the most important of which are the cytochromes P450 (CYP). Activated pyrrolic intermediates may react with glutathione (GSH) under catalysis of glutathione S-transferase (GST), resulting in a water-soluble conjugate that is easily excreted. Pyrrolic intermediates may also form cross-links with tissue nucleophiles, such as DNA and proteins, resulting in toxicity. Alternatively, the pyrrole may be detoxified by CYP or flavin monooxygenases (FMO) to N-oxides, or hydrolyzed to retronecine and senecic acid.

PYRROLIZIDINE ALKALOIDS IN FOODS

CH3 O

HO

85

86

R. A. COULOMBE

PYRROLIZIDINE ALKALOIDS IN FOODS

FIG. 5. Postulated sequential mechanism of cytochrome P450-(CYP)-mediated oxidative activation of senecionine, and formation of a cross-link. Generation of the pyrrolic intermediate results in activation to electrophilic centers at the C-7 and C-9 by conjugation of the pyrrole nitrogen lone pair. Once formed, the activated, bifunctional pyrrolic intermediate can form cross-links with cellular nucleophiles (Nu), such as DNA, and protein, such as actin. It has been postulated that nucleophilic substitution at C-7 is favored over C-9 owing to greater stablilization of the secondary carbonium ion at C-7. Adapted from Huxtable and Cooper (2000) and Rajski and Williams (1998).

87

88

R. A. COULOMBE

pyrrolizidine alkaloids. For example, the degree to which various pyrrolic pyrrolizidine alkaloids cause cytotoxic end points such as megalocyte formation and inhibition of colony formation correlates with the formation of DNA cross-links in cultured mammalian cells (Kim et al., 1993). Megalocytosis is seen in the liver of animals exposed to pyrrolizidine alkaloids. It is a condition where cells have an unusually large cellular and nuclear volume, due to the antimitotic effect of the pyrrolizidine alkaloids where biosynthesis of cellular precursors occur, but without formation of mitotic spindles and mitosis (Figure 6). Another report demonstrated that DNA cross-link formation by pyrrolizidine alkaloids interferes with DNA replication. Dehydrosenecionine (DHSN) and dehydromonocrotaline (DHMO) interrupt the polymerase chain reaction amplification of a segment of pBR322 (Kim et al., 1999). This implies that cross-linking by activated pyrrolizidine alkaloids is functionally significant in the cell. There is a significant difference in cross-link potency among pyrrolizidine alkaloids. In a series of structure–activity studies, it was found that structural features, most notably the presence of α,β-unsaturation and a macrocyclic diester, confer potent cross-link and megalocytic activity to pyrrolizidine alkaloids in cultured bovine kidney (MDBK) cells (Kim et al., 1993, 1999). In this system, macrocyclic pyrrolizidine alkaloids with α,β-unsaturation (seneciphylline, senecionine, riddelliine and retrorsine) were significantly more potent DNA cross-linkers and inducers of megalocytosis than was the α,β-saturated monocrotaline. The open diester pyrrolizidine alkaloids (latifoline and heliosupine) were less potent than monocrotaline, and the simple necine base retronecine was the least potent of any of these. Ironically, indicine N-oxide, which had demonstrated activity against acute leukemia in clinical trials, but was discontinued because of myelosupression and severe hepatotoxicity in human subjects (King et al., 1987), did not induce any measurable DNA cross-links (Kim et al., 1993). The rank order of relative DNA cross-link potencies is often reflected in the animal toxicity of these pyrrolizidine alkaloids. For example, the acute toxicities in rats of senecionine, seneciphylline and retrorsine are approximately equal, but significantly higher than that of monocrotaline (Huxtable and Cooper, 2000). Chemically reactive pyrrolic pyrrolizidine alkaloids share a common pyrrolic substructure with reductively activated bifunctional mitomycins, such as mitomycin C, which preferentially cross-link 5′-CG sequences within DNA (Woo et al., 1993). There is conflicting evidence as to whether activated pyrrolizidine alkaloids share similar cross-linking base or DNA sequence specificity. Several DNA bases have been shown to be involved in covalent interactions and/or cross-links by pyrrolic pyrrolizidine alkaloids. For example, dehydroretronecine reacts with purine and pyrimidine

PYRROLIZIDINE ALKALOIDS IN FOODS

89

FIG. 6. Phase-contrast photomicrographs of pyrrolizidine alkaloid-induced megalocytosis in Madin–Darby bovine kidney (MDBK) cells. Cells were cultured with either (a) vehicle (dimethylsulfoxide, DMSO) or (b) seneciphylline (500 µM) for 2 h, washed, supplemented with fresh medium, then cultured for 6 weeks (×360). Reproduced from Kim et al. (1993), by permission.

nucleosides such as the N2 of deoxyguanosine (dG) and N6 of deoxyadenosine (dA). The O2 sites of uridine and deoxythymidine or thymidine (dT) have also been identified as targets (Robertson, 1982; Wickramanayake et al., 1985). Dehydromonocrotaline and dehydroretrorsine preferentially cross-

90

R. A. COULOMBE

linked dG-to-dG at a 5′-CG sequence in synthetic duplex DNA (Weidner et al., 1990). Dehydromonocrotaline was shown to cross-link at the N-7 position of guanine in a 35 bp fragment of pBR322 with a preference for 5′-GG and 5′-GA sequences. However, because these authors examined only alkylation at guanyl residues, no light was shed on whether DHMO alkylated at other sites (Pereira et al., 1998). Our laboratory has recently shown that the pyrrolic pyrrolizidine alkaloids dehydrosenecionine, dehydroretrorsine and dehydromonocrotaline had little, if any, discernible sequence specificity when cross-linked to a series of synthetic 32P end-labeled oligonucleotides of varying DNA sequence (Coulombe and Rieben, 2002). These oligonucleotide targets were 14- or 24-base poly dA-T that had different central base sequences such as 5′-d(CG), 5′-d(GC), 5′-d(TA); 5′-d(CGA), 5′-d(GCA) or 5′-d(TAA). Similar model oligonucleotides have been used to determine possible sequence recognition of other bifunctional electrophiles, such as mitomycin C (Weidner et al., 1990). Less is known about the characteristics of the pyrrolizidine alkaloidinduced DNA–protein cross-link. Initial studies from our laboratory showed that a major protein released by DNAse I treatment of DNA–protein complexes purified from pyrrolizidine alkaloid-treated bovine kidney cell (MDBK) nuclei had a molecular weight of c. 43 kD, was acidic (pI ≈ 4.5) and had a two-dimensional electrophoretic pattern similar to that of cells treated with cisplatinum, a benchmark bifunctional cross-linker known to cross-link DNA with actin (Kim et al., 1995). A follow-up study using an anti-actin multiple antigen peptide (MAP) antibody confirmed that actin is the major protein involved in pyrrolizidine alkaloid-induced DNA–protein cross-links in human breast carcinoma (MCF-7) and in Madin–Darby bovine kidney (MDBK) cells (Coulombe et al., 1999). The cross-link pattern by two pyrrolic pyrrolizidine alkaloids dehydrosenecionine and dehydromonocrotaline was similar to that of cisplatinum as well as mitomycin C, a pyrrolic anti-cancer drug, although these benchmark compounds were more potent cross-linkers than the pyrrolizidine alkaloids (Figure 7). The involvement of actin in pyrrolizidine alkaloid-induced DPCs is a reasonable expectation due to the abundance of this protein in the nuclear matrix. Actin is also a target for cross-linkers such as cisplatinum and trivalent chromium (Miller et al., 1991), and mitomycin C (Coulombe et al., 1999). Inasmuch as the megalocytic and antimitotic effects of pyrrolizidine alkaloids coincide with cross-linking potency (Kim et al., 1993), it is plausible that the antimitotic action of pyrrolic pyrrolizidine alkaloids in vitro and in vivo may be explained, at least in part, by their ability to cross-link DNA with actin. In addition to actin, other protein

PYRROLIZIDINE ALKALOIDS IN FOODS

91

targets that were later identified in dehyrdomonocrotaline-treated cultured pulmonary artery endothelial cells included galectin-1 and proteindisulfide isomerase (Lame et al., 2000). The functional significance of a monocrotaline cross-link to these other proteins is not known. VIII. CONTROL OF PYRROLIZIDINE ALKALOIDS AND FUTURE PROSPECTS

Usual practices of treating cereal crops with herbicides to reduce weed infestation and subsequent quality assurance steps such as grain inspection can help prevent large-scale outbreaks of poisoning due to contamination of pyrrolizidine alkaloid-containing seeds or plants in staple food. Modern winnowing techniques readily separate toxic Heliotropium, Senecio and Crotalaria seeds from cereal grain, at least to acceptable tolerances. However, it is likely that drought, famine and political instability will again conspire to produce another large-scale outbreak of human poisonings. Periodic exposure to small amounts of pyrrolizidine alkaloids in foods such as milk, honey and eggs is also likely to occur. University-based outreach and information programs may be one mechanism by which producers can become better educated on ways to minimize potential pyrrolizidine alkaloid contamination in their products. For example, honey producers may benefit from learning if plants their bees are likely to pollenate contain pyrrolizidine alkaloids. They can then move their hives to other fields, thereby preventing or reducing contamination. The many poisoning cases resulting from consumption of “safe” and “natural” supplements emphasizes the urgent need for greater awareness of the potential adverse health effects of herbal products. Years of traditional use of a botanical supplement is no guarantee of safety. This caveat is especially true in the case of pyrrolizidine alkaloids and other chemicals, whose effects can be delayed months or even years. Such chronic toxicoses are notoriously difficult to diagnose. Public health campaigns against the use of Crotalaria fulva have been very effective in reducing VOD in Jamaica (Mattocks, 1986). In the United States, unfortunately, much of the popular literature on herbal supplements, as well as nearly all of the product “fact sheets” provided by health food stores (produced by the companies that produce the herb products), are misleading. In the case of pyrrolizidine alkaloid-containing plants like comfrey and coltsfoot, new legislation is needed to protect consumers. Clearly, existing laws, such as the Dietary Supplement Health and Education Act Congress, need to be significantly modified in the interest of consumer protection. In an article entitled “Herbal Roulette,” Consumer’s

92

R. A. COULOMBE

FIG. 7. Western immunoblots showing the presence of actin isolated from proteins released from purified DNA–protein cross-links in Madin–Darby bovine kidney (MDBK; left) and human breast carcinoma MCF-7 (right) nuclei treated with DMSO vehicle (control), and the pyrrolic pyrrolizidine alkaloids dehydrosenecionine (DHSN), or dehydromonocrotaline (DHMO). Anti-cancer cross-linking agents mitomycin C (MMC), or cisplatinum are included for reference. Released proteins that were originally cross-linked to 40 µg nuclear DNA were loaded onto SDS-PAGE (11% running, 4.5% stacking) gels, electorphoresed, transferred to membranes, then probed with a monoclonal anti-actin multiple antigen peptide (MAP) antibody that recognizes an epitope conserved in all actin isoforms. α-Skeletal actin was used as a standard (α-actin). Relative densitometric intensities of actin signals were normalized to that observed from cisplatinum-induced DNA–protein cross-links. Reproduced from Coulombe et al. (1999) by permission.

Union recommended that changes to the law should include: clearer disclaimers, in large type, stating that any claims of safety and efficacy are strictly the claims of the manufacturer, and have not been confirmed by the FDA; consistent manufacturing and content standards; and banning clearly dangerous supplements (Consumer’s Union, 1995). Another worthwhile step would be to adapt the approach of the German Commission E, which was established to evaluate the safety and efficacy of herbal supplements on the basis of clinical trials and comprehensive risk assessment analyses. The Commission has published more than 320 monographs on herbs

PYRROLIZIDINE ALKALOIDS IN FOODS

93

(Klepser and Klepser, 1999). In the absence of these steps, people in the United States will continue to receive misinformation about the usefulness and safety of their “health” foods. In any event, an industry and governmentsponsored campaign to increase awareness among herbalists, natural food proprietors, and Native American shamans on the deadly potential of their products will help. A few high-profile product liability cases may be the only impetus to motivate the industry to self-regulate. Medical providers and users of herbs should be aware that young people are especially susceptible to the toxic effects of pyrrolizidine alkaloidcontaining supplements. Medical providers should also query their patients on their possible use of home remedies and herbal supplements, especially when there are unexplained symptoms. An increased recognition of the relatively uncommon symptoms of pyrrolizidine alkaloid poisoning by the medical community will also increase the accuracy of incident reporting. ACKNOWLEDGEMENTS

I wish to acknowledge generous support from NIH-NIEHS (ESO4813), USDA-NRI (02-2780, 98-3754, 97-3081, AS-79), the Willard L. Eccles Foundation, the Dr W. C. Swanson Family Foundation, and from the Utah Agricultural Experiment Station, where this is designated as number 7452. This review is dedicated to Centennial Professor Frank R. Stermitz of Colorado State University – natural product chemist, teacher, mentor and friend. REFERENCES Arseculeratne, S. N., Gunatilaka, A. A. L., and Panabokke, R. G. 1981. Studies on medicinal plants of Sri Lanka: Occurrence of pyrrolizidine alkaloids and hepatotoxic properties in some traditional medicinal herbs. J. Ethnopharmacol. 4, 159–177. Asada, Y. and Furuya, T. 1982. Neosenkirkine and senkirkine from Sencio pierotti. Planta Med. 44, 182. Asada, Y. and Furuya, T. 1984. Studies on Constituents of Crude Drugs. XV. New pyrrolizidine alkaloids from Ligularia dentata. Chem. Pharm. Bull. 32, 475–482. Asprey, G. F. and Thornton, P. 1955. Medicinal plants of Jamaica. West Indian Med. J. 4, 145–168. Aston, N., Morris, P. and Tanner, S. 1996. Retrorsine in breast milk influences copper handling in suckling rat pups. J. Hepatol. 25, 748–755. Bach, N., Thung, S. N. and Schaffner, F. 1989. Comfrey herb tea-induced hepatic veno-occlusive disease. Am. J. Med. 87, 97–99. Bah, M., Bye, R. and Pereda-Miranda, R. 1994. Hepatotoxic pyrrolizidine alkaloids in the Mexican medicinal plant Packera candidissima (Asteraceae: Senecioneae). J. Ethnopharmacol. 43, 19–30. Barnes, J. M., Magee, P. N., and Schoental, R. 1964. Lesions in the lungs and livers of rats poisoned with the pyrrolizidine alkaloid fulvine and its N-oxide. J. Pathol. Bacteriol. 88, 521–531.

94

R. A. COULOMBE

Borka, L. and Onshuus, I. 1979. Senkirkine content in the leaves of Tussilago fafara L. Medd. Norsk Farm. Selsk 41, 165–168. Boyer, T. D. 1989. The glutathione s-transferases: an update. Heptology 9, 486–496. Bras, G. and Hill, K. R. 1956. Veno-occlusive disease of the liver-essential pathology. Lancet 2, 161–163. Bras, G., Jelliffe, D. B. and Stuart, K. L. 1954. Veno-occlusive disease of liver with nonportal type of cirrhosis, occurring in Jamaica. Am. Med. Assoc. Arch. Pathol. 57, 285–300. Broch-due, A. I. and Aasen, A. J. 1980. Alkaloids of Anchusa officinalis L. Identification of the pyrrolizidine alkaloid lycopsamine. Acta Chem. Scand. B34, 75–77. Chauvin, P., Dillon, J.-C., Moren, A., Talbak, S. and Barakaev, S. 1993. Heliotrope poisoning in Tadjikistan. Lancet 341, 1663. Chopra, R. N. 1933 “Indigenous Drugs of India”. The Art Press, Calcutta. Coady, A. 1973. “Evidence for the exposure of human populations in Ethiopia and elsewhere to liver toxins, including carcinogens, of plant orgin.” Dissertation submitted to University of Manchester. Consumer’s Union 1995. Herbal Roulette. Consumer Reports, pp. 698–705. Coulombe, R.A. and Rieben, R.K. 2002. Lack of apparent base sequence preference of activated pyrrolizidine alkaloid cross-links with DNA. In “Poisonous Plants and Related Toxins”. (T. Acamovic, C. S. Stewart and T. W. Pennycott, eds.) CAB International (in press). Coulombe, R.A., Drew, G. L. and Stermitz, F. R. 1999. Pyrrolizidine alkaloids crosslink DNA with actin. Toxicol. Appl. Pharmacol. 154, 198–202. Crews, C., Startin, J. R. and Clarke, P. A. 1997. Determination of pyrrolizidine alkaloids in honey from selected sites by solid phase extraction and HPLC-MS. Food Addit. Contam. 14, 419–428. Culvenor, C. C. J., Edgar, J.A., Smith, L.W. and Hirono, I. 1976. The occurrence of senkirkine in Tussilago farfara. Aust. J. Chem. 29, 229–230. Culvenor, C. C. J., Clarke, M., Edgar, J. A., Frahn, J. L., Jago, M. V., Peterson, J. E. and Smith, L. W. 1980. Structure and toxicity of the alkaloids of Russian comfrey (Symphytum X uplandicum Nyman), a medicinal herb and item of human diet. Experientia 36, 377–379. Culvenor, C. C. J., Edgar, J. A. and Smith, L. W. 1981. Pyrrolizidine alkaloids in honey from Echium plantagineum L. J. Agric. Food Chem. 29, 958–960. Datta, B. V., Khuroo, M. S., Mattocks, A. R., Aikat, B. K. and Chhuttani, P. N. 1978a. Herbal medicines and veno-occlusive disease in India. Postgrad. Med. J. 54, 511–515. Datta, D. V., Khuroo, M. S., Mattocks, A. R., Aikat, B. K. and Chhuttani, P. N. 1978b. Venoocclusive disease of liver due to heliotropium plant, used as medicinal herb. J. Assoc. Physicians India 26, 383–393. Deinzer, M. L., Thomson, P. A., Burgett, D. M. and Isaacson, D. L. 1977. Pyrrolizidine alkaloids: their occurrence in honey from tansy ragwort (Senecio jacobaea L.). Science 195, 497–499. Dickinson, J. O., Cooke, M. P., King, R. R. and Mohamed, P. A. 1976. Milk transfer of pyrrolizidine alkaloids in cattle. J. Am. Vet. Med. Assoc. 169, 1192–1196. Edgar, J. A., and Smith, L. W. 2000. Transfer of pyrrolizidine alkaloids into eggs: food safety implications. In “Natural and Selected Synthetic Toxins Biological Implications”, (A. Tu, and W. Gaffield, eds), pp. 118–128. American Chemical Society, Washington, DC. Fox, D. W., Hart, M. C., Bergeson, P. S., Jarrett, P. B., Stillman, A. E. and Huxtable, R. J. 1978. Pyrrolizidine (Senecio) intoxication mimicking Reye syndrome. J. Pediatrics 93, 980–982. Freedman, J. H., Ciriolo, M. R. and Peisach, J. 1989. The role of glutathione in copper metabolism and toxicity. J. Biol. Chem. 264, 5598–5605. Furuya, T. Murakami, K. and Hikichi, M. 1971. Senkirkine, a pyrrolizidine alkaloid from Farfugium japonicum. Phytochemistry 10, 3306–3307. Gandhi, R. N. Rajagopalan, T.R. and Seshadri, T.R. 1966. Chemical components of Heliotropium eichwaldi. Curr. Sci. 35, 121–122.

PYRROLIZIDINE ALKALOIDS IN FOODS

95

German Federal Health Bureau 1992. Bundesrepublik Deutschland/pyrrolizidine alkaloide, stufe ii abwehr von arzneimittelrisiken. Pharm. Ztg. 137, 2088–2089. Glowaz, S. L., Michinika, M. and Huxtable, R. J. 1992. Detection of a reactive pyrrole in the hepatic metabolism of the pyrrolizidine alkaloid, monocrotaline. Toxicol. Appl. Pharmocol. 115, 168–173. Goeger, D. E., Cheeke, P. R., Schmitz, J. A. and Buhler, D. R. 1982. Effect of feeding milk from goats fed tansy ragwort (Senecio jacobaea) to rats and calves. Am. J. Vet. Res. 43, 1631–1633. Guengerich, F. P. 1989. Characterization of human microsomal cytochrome P-450 enzymes. Ann. Rev. Pharmacol. Toxicol. 29, 241–264. Habs, H., Habs, M., Marquardt, H., Roder, E., Schmahl, D. and Wiedenfeld, H. 1982. Carcinogenic and mutagenic activity of an alkaloidal extract of Senecio nemorensis ssp. fuchsii. Arzneimittelforschung 32, 144–148. Hagglund, K. M., L’Empereur, K. M., Roby, M. R. and Stermitz, F. R. 1985. Latifoline and latifoline-n-oxide from Hackelia floribunda, the western false forget-me-not. J. Nat. Products 48, 638–639. Hikichi, M. and Furuya, T. 1976. Studies on constituents of crude drugs. VII. syneilesine and acetylsyneilesine from Syneilesis palmata. Chem. Pharm. Bull. 24, 3178–3184. Hikichi, M., Furuya, T. and Iitaka, Y. 1978. Yamataimine, A new pyrrolizidine alkaloid form Cacalia yatabei. Tetrahedron Lett 8, 767–770. Hills, L. D. 1976. “Comfrey, Past, Present and Future”. Faber & Faber, London. Hirono, I. Shimizu, M., Fushimi, K., Mori, H. and Kato, K. 1973. Carcinogenic activity of Petasites japonicus Maxim, a kind of coltsfoot. Gann 64, 527–528. Hirono, I., Mori, H. and Haga, M. 1978. Carcinogenic activity of Symphytum officinale. J. Natl Cancer Inst. 61, 865–869. Hoque, M. S., Ghani, A., and Rashid, H. 1976. Alkaloids of Heliotropium indicum Linn. grown in Bangladesh. Bangladesh Pharm. J. 5, 13–15. Howell, J. M., Deol, H. S., Dorling, P. R., and Thomas, J. B. 1991. Experimental copper and heliotrope intoxication in sheep: morphological changes. J. Comp. Pathol. 105, 49–74. Huxtable, R. J. 1979. New aspects of the toxicology and pharmacology of pyrrolizidine alkaloids. Gen. Pharmacol. 10, 159–167. Huxtable, R. J. (1980). Herbal teas and toxins: novel aspects of pyrrolizidine poisoning in the United States. Pers. Biol. Med. 24, 1–14. Huxtable, R. J. 1989. Human health implications of pyrrolizidine alkaloids and herbs containing them. In “Toxicant of Plant Origin” (P. R. Cheeke, ed.), pp. 41–86. CRC Press, Inc., Boca Raton, FL. Huxtable, R. J. 1990. Activation and pulmonary toxicity of pyrrolizidine alkaloids. Pharmac. Ther. 47, 371–389. Huxtable, R. J., and Cooper, R. A. 2000. Pyrrolizidine alkaloids: physicochemical correlates of metabolism and toxicity. In “Natural and Selected Synthetic Toxins Biological Implications” (A. Tu and W. Gaffield, eds), pp. 100–117. American Chemical Society, Washington, DC. Huxtable, R. J., Stillman, E. A., and Ciaramitaro, D. 1977. Characterization of alkaloids involved in human Senecio (pyrrolizidine) poisoning. Proc. West. Pharmacol. Soc. 20, 455–459. IARC (1976). IARC Monogr. Eval. Carcinog. Risk Chem. Man 10, 265–342. Jones, J. G., and Taylor, D. E. 1989. Hepatic veno-occlusive disease and herbal remedies. Ann. Rheum. Dis. 48, 791. Kim, H.-Y., Stermitz, F. R., Molyneux, R. J., Wilson, D. W., Taylor, D. and Coulombe, R. A., Jr 1993. Structural influences on pyrrolizidine alkaloid-induced cytopathology. Toxicol. Appl. Pharmacol. 122, 61–69. Kim, H.-Y., Stermitz, F. R. and Coulombe, R. A., Jr 1995. Pyrrolizidine alkaloid-induced

96

R. A. COULOMBE

DNA–protein cross-links. Carcinogenesis 16, 2691–2697. Kim, H.-Y., Stermitz, F. R., Li, J. K. and Coulombe, R. A., Jr 1999. Comparative DNA crosslinking by activated pyrrolizidine alkaloids. Food Chem. Toxicol 37, 619–625. King, S. A., Suffness, M., Leyland-Jones, B., Hoth, D. F. and O’Dwyer, P. J. 1987. Indicine N-oxide: Clinical use of a pyrrolizidine alkaloid. Cancer Treatment Rep. 71, 517–523. Klepser, T. B. and Klepser, M. E. 1999. Unsafe and potentially safe herbal therapies. Am. J. Health Syst. Pharmacol 56, 125–138. Lafranconi, W. M., and Huxtable, R. J. 1984. Hepatic metabolism and pulmonary toxicity of monocrotaline using isolated perfused liver and lung. Biochem. Pharmacol. 33, 2479–2484. Lamé, M. W., Jones, A. D., Wison, D. W., Dunston, S. K., and Segall, H. J. (2000). Protein targets of monocrotaline pyrrole in pulmonary artery endothelial cells. J. Biol. Chem. 275, 29091–29099. Lyford, C. L., Vergara, G. G., and Moeller, D. D. 1976. Hepatic veno-occlusive disease originating in Ecuador. Gastroenterology 70, 105–108. Macksad, A., Schoental, R. and Coady, A. 1970. The hepatotoxic action of a traditional Bedu plant remedy “RamRam”. J. Kwt. Med. Assoc. 4, 297–299. Margalith, D., Heraief, E., Schindler, A. M., Birchler, R., Mosimann, F., Aladjem, D., and Gonvers, J. J. 1985. Veno-occlusive disease of the liver due to the use of tea made from senecio plants. A report of two cases. J. Hepatol. 1, S280. Mattocks, A. R. 1986. “Chemistry and Toxicology of Pyrrolizidine Alkaloids”. Academic Press, Orlando, FL. Mayer, F., and Luthy, J. 1993. Heliotrope poisoning in Tadjikistan. Lancet 342, 246–247. McDermott, W. V., and Ridker, P. M. 1990. The Budd–Chiari syndrome and hepatic venoocclusive disease. Recognition and treatment. Arch. Surg. 125, 525–527. McGee, J. O., Patrick, R. S., Wood, C. B., and Blumgart, L. H. 1976. A case of veno-occlusive disease of the liver in Britain associated with herbal tea consumption. J Clin. Pathol. 29, 788–794. McLean, E. K. 1970. The toxic actions of pyrrolizidine (Senecio) alkaloids. Pharmacol. Rev. 22, 429–483. McLean, E. K. 1974. Senecio and other plants as liver poisons. Israel J. Med. Sci. 10, 436–440. Mead, E. W., Looker, M., Gardner, D. R., and Stermitz, F. R. 1992. Pyrrolizidine alkaloids of Liatris punctata and its root parasite, Castilleja Integra. Phytochemistry 31, 3255–3257. Miller, C. A., III, Cohen, M. D., and Costa, M. 1991. Complexing of actin and other nuclear proteins to DNA by cis-diamminedichloroplatinum(II) and chromium compounds. Carcinogenesis 12, 269–276. Miranda, C. L., Buhler, D. R., Ramsdell, H. S., Cheeke, P. R., and Schmitz, J. A. 1982. Modifications of chronic hepatotoxicity of pyrrolizidine (Senecio) alkaloids by butylated hydroxyanisole and cysteine. Toxicol. Lett. 10, 177–182. Miranda, C. L., Reed, R. L., Guengerich, F. P., and Buhler, D. R. 1991. Role of cytochrome P450IIIA4 in the metabolism of the pyrrolizidine alkaloid senecionine in human liver. Carcinogenesis 12, 515–519. Mohabbat, O., Srivastava, R. N., Younos, M. S., Sediq, G. G., Merzad, A. A., and Aram, G. N. 1976. An outbreak of hepatic veno-occlusive disease in north-western Afghanistan. Lancet 2, 269–271. Molteni, A., Ward, W. F., Ts’ao, C. H., and Fitzsimons, E. J. 1988. Serum copper concentration as an index of cardiopulmonary injury in monocrotaline-treated rats. Ann. Clin. Lab Sci. 18, 476–483. Molyneux, R. J., and Johnson, A. E. 1984. Extraordinary levels of production of pyrrolizidine alkaloids in Senecio riddellii. J. Nat. Products 47, 1030–1032.

PYRROLIZIDINE ALKALOIDS IN FOODS

97

Moore, M. 1979. Medicinal Plants of the Mountain West, p. 36. The Museum of New Mexico Press, Santa Fe, NM. Morris, P., O’Neill, D., and Tanner, S. 1994. Synergistic liver toxicity of copper and retrorsine in the rat. J. Hepatol. 21, 735–742. Omar, M., Defeo, J., and Youngken, H. W. Jr. 1983. Chemical and toxicity studies of Trichodesma africanum L. J. Nat. Products 46, 153–156. Pereira, T. N., Webb, R. I., Reilly, P. E. B., Seawright, A. A., and Prakash, A. S. 1998. Dehydromonocrotaline generates sequence-selective N-7 guanine alkylation and heat and alkali stable multiple fragment DNA crosslinks. Nucleic Acids Res. 26, 5441–5447. Rajski, S. R., and Williams, R. M. 1998. DNA Cross-linking agents as antitumor drugs. Chem. Rev. 98, 2723–2796. Ridker, P. M., Ohkuma, S., McDermott, W. V., Trey, C., and Huxtable, R. J. 1985. Hepatic venocclusive disease associated with the consumption of pyrrolizidine-containing dietary supplements. Gastroenterology 88, 1050–1054. Robertson, K. A. 1982. Alkylation of N2 in deoxyguanosine by dehydroretronecine, a carcinogenic metabolite of the pyrrolizidine alkaloid monocrotaline. Cancer Res. 42, 8–14. Robertson, K. A., Syemour, J. L., Hsia, M.-T., and Allen, J. R. 1977. Covalent interaction of dehydroretronecine, a carcinogenic metabolite of the pyrrolizidine alkaloid monocrotaline with cysteine and glutathione. Cancer Res. 37, 3141–3144. Roder, E., Wiedenfeld, H., and Frisse, M. 1980. Pyrrolizidinalkaloide aus Senecio doronicum. Phytochemistry 19, 1275–1277. Roeder, E. 2000. Medicinal plants in China containing pyrrolizidine alkaloids. Pharmazie 55, 711–726. Roeder, E., Bourauel, T., and Neuberger, V. 1992. Symviridine, a new pyrrolizidine alkaloid from Symphytum species. Phytochemistry 31, 4041–4042. Roitman, J. N. 1981. Comfrey and liver damage. Lancet 1, 944. Rose, E. F. 1972. Senecio species: Toxic plants used as food and medicine in the Transkei. S. Afr. Med. J. 46, 1039–1043. Roulet, M., Laurini, R., Rivier, L., and Calame, A. 1988. Hepatic veno-occlusive disease in newborn infant of a woman drinking herbal tea. J. Pediatrics 112, 433–436. Schneider, M. J., and Stermitz, F. R. 1990. Uptake of host plant alkaloids by root parasitic Pedicularis species. Phytochemistry 29, 1811–1814. Schoental, R. 1968. Toxicology and carcinogenic action of pyrrolizidine alkaloids. Cancer Res. 28, 2237–2246. Schoental, R., and Coady, A. 1968. The hepatotoxicity of some Ethiopian and east African plants, including some used in traditional medicines. East African Med. J. 45, 577–580. Schoental, R., and Pullinger, B. D. 1972. On the alleged oestrogenic and other medicinal properties of pyrrolizidine (senecio) alkaloids. East African Med. J. 49, 436–439. Selzer, G., and Parker, R. G. F. 1951. Senecio poisoning exhibiting as Chiari’s syndrome. Am. J. Pathol. 27, 885–907. Siddiqi, M. A., Suri, K. A., Suri, O. P., and Atal, C. K. 1978. Genus Crotalaria: Part XXXIV — Cronaburmine, a new pyrrolizidine alkaloid from Crotalaria nana Burm. Indian J. Chem. 16B, 1132–1133. Spang, R. 1989. Toxicity of tea containing pyrrolizidine alkaloids. J. Pediatrics 115, 1025. Sperl, W., Stuppner, H., Gassner, I., Judmaier, W., Dietze, O., and Vogel, W. 1995. Reversible hepatic veno-occlusive disease in an infant after consumption of pyrrolizidine-containing herbal tea. Eur. J. Pediatrics 154, 112–116. Stermitz, F. R., and Suess, T. R. 1978. Pyrrolizidine alkaloids in Castilleja rhexifolia (Schrophulariaceae). Phytochemistry 17, 2142.

98

R. A. COULOMBE

Stickel, F., and Seitz, H. K. 2000. The efficacy and safety of comfrey. Pub. Health Nutr. 3, 501–508. Stillman, A. E., Huxtable, R., Consroe, P., Kohnen, P., and Smith, S. 1977. Hepatic veno-occlusive disease due to pyrrolizidine (Senecio) poisoning in Arizona. Gastroenterology 73, 349–352. Sullivan, G. 1981. Detection of pyrrolizidine-type alkaloids in matarique (Cacalia decomposita). Vet. Hum. Toxicol. 23, 6–7. Tandon, B. N., Tandon, H. D., Tandon, R. K., Narndranathan, M., and Joshi, Y. K. 1976. An epidemic of veno-occlusive disease of liver in central India. Lancet 2, 271–272. Tandon, H. D., Tandon, B. N., Tandon, R., and Nayak, N. C. 1977. A pathological study of the liver in an epidemic outbreak of veno-occlusive disease. Indian J. Med. Res. 65, 679–684. Tandon, H. D., Tandon, B. N., and Mattocks, A. R. 1978. An epidemic of veno-occlusive disease of the liver in Afghanistan. Pathologic features. Am. J. Gastroenterol. 70, 607–613. Ueng, Y.-F., Kuwabara, T., Chun, Y.-J., and Guengerich, F. P. 1997. Cooperativity in oxidations catalyzed by cytochrome P450 3A4. Biochemistry 36, 370–381. Wade, A. (1977). “Martindale. The Extra Pharmacopoeia”. Pharmaceutical Press, London. Watt, J. M., and Breyer-Brandwijk, M. G. 1962. “The Medicinal and Poisonous Plants of Southern & Eastern Africa”. Livingstone, Edinburgh and London. Weidner, M. F., Sigurdsson, S.T., and Hopkins, P.B. 1990. Sequence preferences of DNA interstrand cross-linking agents: dG-to-dG cross-linking at 5′-CG by structurally simplified analogues of mitomycin C. Biochemistry 29, 9225–9233. Weston, C. F. M., Cooper, B. T., Davies, J. D., and Levine, D. F. 1987. Veno-occlusive disease of the liver secondary to ingestion of comfrey. Brit. Med. J. 295, 183. White, I. N., Mattocks, A. R., and Butler, W. H. 1973. The conversion of the pyrrolizidine alkaloid retrorsine to pyrrolic derivatives in vivo and in vitro and its acute toxicity to various animal species. Chem. Bil. Int. 6, 207–218. White, R. D., Krumperman, P. H., Cheeke, P. R., Deinzer, M. L., and Buhler, D. R. 1984. Mutagenic responses of tansy ragwort (Senecio jacobaea) plant, pyrrolizidine alkaloids and metabolites in goat milk with the Salmonella/mammalian-microsome mutagenicity test. J. Anim. Sci. 58, 1245–1254. Wickramanayake, P. P., Arbogast, B.L., Buhler, D.R., Dienzer, M.L. and Burlingame, A.L. 1985. Alkylation of nucleosides and nucleotides by dehydroretronecine; characterization of covalent adducts by liquid secondary ion mass spectrometry. J. Am. Chem. Soc. 107, 2485–2488. Williams, A. O., and Schoental, R. 1970. Hepatotoxicity of Senecio abyssinicus experimental and ultrastructural studies. Trop. Geog. Med. 22, 201–210. Williams, D. E., Reed, R. L., Kedzierski, B., Ziegler, D. M., and Buhler, D. R. 1989. The role of flavin-containing monooxygenase in the N-oxidation of the pyrrolizidine alkaloid senecionine. Drug Metab. Dispos. 17, 380–386. Willmont, F. C., and Robertson, G. W. 1920. Senecio disease or cirrhosis of the liver due to Senecio poisoning. Lancet 2, 848–849. Wilson, D. W., Segall, H. J., Pan, L. C., Lame, M. W., Estep, J. E., and Morin, D. 1992. Mechanisms and pathology of monocrotaline pulmonary toxicity. Crit. Revs. Toxicol. 22, 307–325. Winter, C. K., and Segall, H. J. 1989. Toxicants of plant origin. In “Metabolism of Pyrrolizidine Alkaloids” (P. R. Cheeke, ed.), pp. 24–40. CRC Press, Inc., Boca Raton, FL. Woo, J., Sigurdsson, S. T., and Hopkins, P. B. 1993. DNA interstand cross-linking reactions of pyrrole-derived bifunctional electrophiles: evidence for a common target site in DNA. J. Am. Chem. Soc. 115, 3407–3415. Yan, C. C., and Huxtable, R. J. 1995. Relationship between glutathione concentration and metabolism of the pyrrolizidine alkaloid, monocrotaline, in the isolated, perfused liver. Toxicol. Appl. Pharmacol. 130, 132–139.

PYRROLIZIDINE ALKALOIDS IN FOODS

99

Yan, C. C., Cooper, R. A., and Huxtable, R. J. 1995. The comparative metabolism of the four pyrrolizidine alkaloids, seneciphylline, retrorsine, monocortaline, and trichoddesmine in the isolated, perfused rat liver. Toxicol. Appl. Pharmacol. 133, 277–284. Yeong, M. L., Swinburn, B., Kennedy, M., and Nicholson, G. 1990. Hepatic veno-occlusive disease associated with comfrey ingestion. J. Gastroenterol. Hepatol. 5, 211–214.

This Page Intentionally Left Blank

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY JOHN N. COUPLAND AND RAFFAELLA SAGGIN Department of Food Science The Pennsylvania State University University Park, PA 16802–2504 USA

I. Introduction II. Theory A. Ultrasonic Waves B. Ultrasonic Propagation in Homogeneous Materials C. Ultrasonic Propagation in Inhomogeneous Media D. Waves at Interfaces E. Diffraction III. Measurement Methods A. Pulsed Modes B. Noncontact Measurements C. Resonance Methods D. Reflectance Methods E. Love Wave Sensors IV. Applications A. Level Sensors B. Fluid Flow C. emperature D. Composition E. Phase Transitions F. Rheological Properties G. Dispersed Systems and Food Microstructure H. Polymeric Systems I. Miscellaneous Applications J. Imaging V. Conclusions Acknowledgements References

ADVANCES IN FOOD AND NUTRITION RESEARCH VOL 45 ISBN: 0-12-016445-0

Copyright © 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

102

J. N. COUPLAND AND R. SAGGIN I. INTRODUCTION

Sensing and measurement of food properties are crucial to improving the quality and profitability of food manufacturing operations. Instrumental measurements can reduce the dependence on time-consuming chemical and sensory analysis. The ideal would be to make an assessment of the food quality as close to the processing step as possible. The measurement should provide some information about the food (e.g. temperature, composition, structure, concentration) useful in controlling the final product quality. The response time is crucial, so while laboratory tests on finished product are valuable, measurements, made at-line or on-line of the freshly made or in-process food are better. On-line sensors in the food industry must also be adequately cheap and robust to survive in the frequently hot and wet environment of a food processing plant. They should also be both nondestructive of the food products and amenable to hygienic design principles (ideally noninvasive) and provide output usefully to an operator or automated control system. Sensing modalities meeting some or all of these conditions have found specific applications in the food industry (e.g., microwaves for water content, thermocouples for temperature, nuclear magnetic resonance (NMR) for solid fat content, and infrared (IR) for composition). Ultrasound has also been used, but to a much lesser extent. The different physical principles sensed ultrasonically provide a unique set of advantages (and disadvantages) as a food sensor. Ultrasound has become invaluable as a nondestructive sensor for materials as diverse and valuable as a human fetus and an aircraft wing. The same principles that enable us to “see” these structures with sound have also allowed researchers to attempt to characterize foods ultrasonically. In many ways food is a more difficult material for nondestructive evaluation as low unit costs and high production rates make it uneconomic to have a trained technician make a piece-by-piece evaluation, and the sensor must be directly incorporated into the process. There have been some important practical successes (e.g. ultrasonic flowmeters, carcass evaluation, concentration sensing and particle size determination), but a far wider set of studies that have never progressed beyond a few academic publications. One crude distinction between the successes and failures is that the physics of sound–material interactions is frequently underexploited by food scientists; another is that the complexities of food manufacture are often underestimated by nondestructive evaluation specialists. One of the goals of this chapter is to clearly set out some of the underlying physical principles and then, in a review of some applications, stress how they can lead to useful measurement.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

103

We start by discussing the physics of sound and the interactions of sound with matter. We then describe the practicalities of ultrasonic measurement, and finally review some applications of ultrasonic measurements to foods, grouping them by the type of parameter measured rather than by food group. The field of ultrasonic evaluation of foods has been previously reviewed elsewhere, and the works of Povey (1997; Povey and Mason, 1998), McClements (1997), Javanaud (1988) and Kress-Rogers (1993) are all recommended. The bibliography provided by Povey (1998a) is particularly comprehensive.

II. THEORY A. ULTRASONIC WAVES

A wave is a collective phenomenon, implying coherent local variation of elements of the structure. In an acoustic wave, the variation is in the position of volume elements of the media through which the sound is passing, and so the acoustic properties of a material depend on both its density (the amount of material to be moved) and mechanical properties (how easy it is to move). The strains within the material caused by the passing sound wave are coupled with other parameters including local energy, pressure, density and temperature of the medium. The rate of the temperature fluctuations is usually more rapid than heat flux; therefore, most theories of sound rely on the adiabatic assumption (i.e. no heat flux) (Strutt, 1945; Wood, 1955). Ultrasound is qualitatively similar to audible sound except that the vibrations occur at frequencies (> 20 kHz) too high to be detected by the human ear. Indeed, a sound wave is part of the whole spectrum of material periodic motion (not including translational motion) that at very high frequencies is seen as thermal waves. Just as classical electromagnetic spectroscopy can probe different structures in the same material by interaction with light of different frequencies, ultrasonic spectra can in principle reveal detail at different levels of material structure. A further interesting parallel between electromagnetic and mechanical waves is that very high frequency sound is better thought of as a particle – a phonon – and the paradoxes of wave–particle duality are relevant; see Povey (1997) for further discussion. The material oscillations associated with an acoustic wave can occur both normal to and parallel with the direction of transmission. When ultrasonic waves are transmitted through an elastic, inert material, each element will experience forces due to the fluctuating pressure and will

104

J. N. COUPLAND AND R. SAGGIN

tend to move from its equilibrium position. For a sinusoidally varying pressure, the motion of the particle will also be sinusoidal and, depending on the balance of forces acting, will follow an elliptical path about its equilibrium position (Wood, 1955). The frequency of the motion is the same as the frequency of the sound and the amplitude of the motion is related to both the magnitude of the wave and the mechanical properties of the material. In most sensing applications, the acoustic power levels and hence the material movements are small and within the elastic limit of the material. It is therefore assumed that as the wave passes the material will return to its unperturbed state and will be physically and chemically unchanged (i.e. ultrasonic testing is nondestructive). The displacement (ξ) of a particle in a periodic wave from its equilibrium position (x) with time (t) is given by:

冢

ξ = a sin 2π vt –

x λ

冣

(1)

where ␯ is the frequency of the oscillation, and a and λ are the amplitude and wavelength of the sound. Eqn (1) can be formulated in terms of wave velocity (c = ␯λ) so the microscopic motion of volume elements of the material is described solely by the bulk properties of the wave. Although sound can propagate by the oscillation of volume elements in three dimensions, ultrasonic measurements are often conducted using waves to induce motion in simpler paths. When an oscillating stress is applied perpendicular to the surface of a material, vibrations take place as a series of compressions and expansions in the direction of sound propagation (particles oscillate back and forth) and longitudinal or compression waves are generated. When the oscillating force is applied parallel to the surface of a material, the vibrations take place perpendicular to the direction of wave propagation and shear or transverse waves are generated (Figure 1). Shear waves can be generated by a shear transducer or by mode conversion of longitudinal waves at an interface (see below). The different types of wave are sensitive to different material properties but, as longitudinal waves have been predominantly used in food systems, we will, unless noted, restrict our discussion to this mode. Longitudinal and shear waves are both bulk waves but it is also possible to generate waves trapped in a surface (which are often useful for detecting surface defects or composition). These waves are produced when longitudinal waves are directed at a critical angle (see section D below) so the refracted energy passes into the surface (Raichel, 2000). Two types of surface waves are Love waves, which are shear waves polarized in the plane of the surface, and Raleigh waves, which contain longitudinal and

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

105

FIG. 1 Diagram illustrating the deformation of matter in a longitudinal or shear pressure wave. The parallel lines represent volume elements of the material that are deformed from their equilibrium either parallel or normal to the direction of sound propagation (i.e. longitudinal or shear waves).

shear components (Blitz, 1963). Surface waves can travel long distances and are sometimes used to detect flaws in pipes. B. ULTRASONIC PROPAGATION IN HOMOGENEOUS MATERIALS

A material responds to the passing ultrasonic wave both elastically and viscously. These two phenomena can be expressed together as a complex wavenumber, k: k = ω/c+ iα

(2)

where ω is the angular frequency (= 2π v), α is the attenuation coefficient and i is √–1. Velocity is the distance traveled by the sound wave in unit time or the product of wavelength and frequency. Attenuation describes the logarithmic decrease in wave energy with distance (x), i.e.:

106

J. N. COUPLAND AND R. SAGGIN

A = A0 exp (–αx)

(3)

where A0 and A are the amplitudes at distance 0 and x. The attenuation coefficient (α) is measured in Nepers (Np) or decibels (dB) per meter (1 Np = 8.686 dB). The propagation of sound through a material depends on the physical properties of that material: k

2

冢ω冣

=

ρ E

(4)

where ρ is the density and E an appropriate elastic modulus. For materials where the attenuation is not large (i.e. α << ω /c), the imaginary component of the parameters described in Eqn (4) can be neglected and the expression can be rewritten as: E c2 = ρ

(5)

The elastic modulus used here is distinct from the elastic modulus used in ordinary static measurements. In static measurements, deformation is slow and the material is at thermal equilibrium prior to measurement, so it is the isothermal elastic modulus that is measured. However, when ultrasonic waves travel through a material, the deformations are extremely rapid and the local temperature rises during the compression phase of the cycle and decreases during rarefactions. Since the rate of pressure change is rapid, no heat exchange occurs between the hot and cold regions and ultrasonic propagation is dependent on the adiabatic elastic modulus. Frequently we shall be concerned with measurements of either the real (velocity) or imaginary (attenuation) part of the wavenumber, but Eqn (4) is a useful reminder that these are parts of a single phenomenon. Similarly the complex parts of the other measured parameters can give independent information on material properties. The ultrasonic properties of selected materials are listed in Table I. 1. Ultrasonic velocity Gases are able to support only longitudinal waves and in practice even these can be hard to measure (see below). For ideal gases the appropriate adiabatic elastic modulus is the bulk elastic modulus (K), equal to the product of the hydrostatic pressure P and γ (= Cp/Cv, where Cp and Cv are the specific heats at constant pressure and volume). Depending on the

TABLE I ULTRASONIC PROPERTIES OF SELECTED MATERIALS. UNLESS NOTED, MEASUREMENTS WERE CONDUCTED AT 20°C

Ultrasonic velocity (m s–1)

Non-food materials Aluminum Brass Cast iron Epoxy resin Gold Perspex Plexiglas Quartz glass Rubber (soft) Steel Teflon Air Hydrogen Oxygen

Reference

0.4 1 1

6320 4400 3500–5800 2400–2900 3240 2730 2670 5570 1480 5900 1350 330 1300 310

3130 2200 2200–3200 1100 1200 1430 1120 3520 – 3230 550 – – –

17 37 25–42 2.7–3.6 63 3.22 3.2 14.5 1.4 45 3 4.3 × 10–4 1.1 × 10–4 4.5 × 10–4

Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Birks and Green (1991) Blitz (1963) Blitz (1963) Blitz (1963)

2.25

1482.3 3840 1597

– 1200 –

1.48 4 1.7

DelGrosso and Mader (1972) Povey (1989) Saggin and Coupland (2001a)

2.25 2.25 2.25 2.25

1511 1527 1650 1530

– – – –

1.57 1.63 1.88

Saggin and Coupland (2001a) Saggin and Coupland (2001a) Saggin and Coupland (2001a) Zacharias and Parnell (1972)

107

Food materials Water Ice Sodium chloride solution (10 wt%) Sucrose solution (10 wt%) Glycerol (10 wt%) Ketchup Apple juice

Shear

Longitudinal

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

Material

Ultrasonic impedance (106 kg m2 s–1) Longitudinal

Frequency (MHz)

Ultrasonic velocity (m s–1) Material

Longitudinal

2.25

1854–1900

1.2

0.037 0.05 0.12 0.5 0.37 0.5 0.5 0.5 0.8

1465 1469 1459 1470 1471 2000–2070 114 274–283 160 700–850 440 114 1020–1740 900–1000 2050

2.5 2.1 1 3.5 1 1

1560 1500 1590 1572 1522 1623

Shear

Ultrasonic impedance (106 kg m2 s–1) Longitudinal 1.9–2.3

Zacharias and Parnell (1972)

1.35 1.35 1.35 1.35 1.35 2.0–2.1 0.09 0.16 0.7–0.85 0.44 141 – – –

McClements and Povey (1988) McClements and Povey (1988) McClements and Povey (1988) McClements and Povey (1988) McClements and Povey (1988) McClements (1997) Liljedahl and Abbott (1994) Mizrach et al. (1989) Hayes and Chingon (1982) Povey (1989) Nielsen and Martens (1997) Povey (1998b) Povey (1998b) Povey (1998b) Povey (1998b)

1.6 1.4 1.65–1.74 55 1.5 –

Povey (1998b) Javanaud (1988) Goss et al. (1978) Ghaedian et al. (1997) McClements (1997) Saggin and Coupland (2001c)

– – – – – – – – – – – – – – – – – –

Reference

–

J. N. COUPLAND AND R. SAGGIN

Syrup (light and dark corn syrup and maple syrup) Olive oil Corn oil Palm oil Soybean oil Sunflower oil Solid animal fat (31°C) Apple (golden delicious) Avocado Papaya Potato Carrot Unyeasted bread dough Milk chocolate (25°C) Aerated milk chocolate (25°C) De-aerated milk chocolate (25°C) Egg thin white Egg yolk Muscle (39°C) Cod fillet (30°C) Skim milk (28°C) Light cheddar cheese

Frequency (MHz)

108

TABLE I (continued) ULTRASONIC PROPERTIES OF SELECTED MATERIALS. UNLESS NOTED, MEASUREMENTS WERE CONDUCTED AT 20°C.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

109

gases present, K has a value between 1.3 and 1.7. Ultrasonic velocity in an ideal gas can also be calculated in terms of its molecular weight, M: c2 =

γ RT M

(6)

where R is the gas constant and T absolute temperature. Liquids can readily support longitudinal waves but can only marginally support shear waves. In liquids the difference between adiabatic and isothermal elastic modulus is negligible and the modulus in Eqn (5) is equivalent to the bulk compressional modulus, K. Eqn (5) is often rewritten in terms of the compressibility κ (= K –1), neglecting the imaginary (loss) part: c–2 = κρ

(7)

In ideal cases the speed of sound in a mixture is given by the volume fraction weighted sum of the components in Eqn (7). When the beam of sound is narrower than the material through which it passes, there is some shearing at the beam edges, and therefore the modulus is equal to K + (4/3)G. However, since in fluids K >> G, it is often possible to neglect the shear modulus and measurements of longitudinal ultrasonic velocity have little direct value as a measurement of shear rheology. The elements in a solid are more directly connected than in a fluid; therefore, solids are better capable of supporting both longitudinal and shear waves. When the sound wave passes through a narrow structure where the cross-sectional dimensions are small compared to the wavelength, the speed of longitudinal ultrasound is given by: Y c2 = ρ

(8)

where Y is the Young modulus. However, when the longitudinal waves pass through a material in bulk form, some shearing motion occurs at the beam edges and the material shear modulus also affects the wave propagation and, as for liquids, the modulus is given by K + (4/3)G. For shear waves, the speed of sound in solids is equal to (G/ρ)1/2 and is typically much slower than for longitudinal waves. 2. Ultrasonic attenuation Any factors that cause sound energy to be either converted to other energy forms (usually heat) or directed into another direction so they are not

110

J. N. COUPLAND AND R. SAGGIN

detected will contribute to the attenuation. In practice the measured attenuation can be taken as an arithmetic sum of several contributing factors:

αmeasured = αclassical + αresonance + αscattering + αother + …

(9)

where the subscripts refer to the various mechanisms. Measuring attenuation as a function of frequency can help to differentiate between the loss mechanisms and to probe the structure and dynamics of the sample. The heating and cooling due to the passing pressure wave is not completely efficient. Some energy is lost as heat (molecular friction), which contributes to measured attenuation. Classical attenuation accounts for only viscous and thermal losses:

αclassical 4π 2 4η (1 – γ )τ = ρc + Cp ␯2 3

冢

冣

(10)

where η is the viscosity and τ is the thermal conductivity. The first term of the equation describes the viscous losses due to intermolecular friction, while the second describes the thermal losses due to inefficient heat transfer. Except for the simplest fluids, other attenuation mechanisms will be much more significant than these, and Eqn (10) usually greatly underestimates the actual attenuation. Important other loss mechanisms include scattering and the disruption of certain chemical equilibria. When a system is in equilibrium and the conditions change, the position of the equilibrium will also change. A passing ultrasonic wave creates periodic zones of high (and low) pressure and temperature that will tend to cause such a shift in the position of the equilibrium if it can respond rapidly enough. The rate of response of the system will depend on the kinetics of the processes, e.g. for a system at equilibrium between two states interconvertable by first-order processes with rate constants k1 and k–1, the relaxation time, tr (i.e. the time to respond to a change) is (k1 + k–1) –1 (Dickinson and McClements, 1995). More complex expressions exist for more complex transitions. If the rate of change in temperature/pressure is high (high frequency) then the system has no time to respond and there is little additional energy lost, while at low frequency the rate of change in pressure/temperature is slow and the reaction can respond fast enough to remain in equilibrium throughout the passage of the wave. However, at intermediate frequencies the position of the equilibrium is constantly attempting to respond to the varying conditions. At the low and high frequency limits, the ultrasonic velocity is constant (and different) and the attenuation (due to this mechanism) is low. Over the intermediate region the velocity increases from the low to the high frequency limit and there

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

111

FIG. 2 Variation of velocity and attenuation across a resonance. The velocity increases from a low to a high frequency plateau while the attenuation per cycle shows a bell-shaped peak.

is a peak in the attenuation spectrum. This process is known as a relaxation. It is expressed mathematically as follows and is illustrated in Figure 2: c = c0 +

αr =

(c∞ – c0) ␯ 2 ␯ 2r + ␯ 2

(11)

Kr ␯ 2r ␯ 2 ␯ 2r + ␯ 2

(12)

In Eqns (11) and (12) c0 and c∞ are the ultrasonic velocities at zero and infinite frequency (the plateau values above and below the relaxation), ␯r is the frequency of the resonance, αr is the attenuation due to the resonance and Kr is a constant describing its magnitude (Dickinson and McClements, 1995). Measuring the resonance frequency by ultrasonic spectroscopy can allow calculation of the relaxation times of the process and is extremely useful in probing the kinetics of fast chemical reactions (tr ~ 1 ns to 1 ms) (Slutsky, 1981; Bryant and McClements, 1999b).

112

J. N. COUPLAND AND R. SAGGIN C. ULTRASONIC PROPAGATION IN INHOMOGENEOUS MEDIA

All foods are heterogeneous over some length scales. Very small heterogeneities are not distinguishable by a sound wave and very large structures are best thought of as bulk phases (see sections describing reflection and imaging). Intermediate sized structures where a characteristic length scale is of similar magnitude to the wavelength of the sound can scatter the acoustic wave, leading to increased attenuation and changes in ultrasonic velocity. For longitudinal sound propagating in water, the wavelength of 1 MHz sound is about 1.5 mm, so many very small structures found in food can act as scatterers. The extent of scattering depends on the size, shape, composition and number of inhomogeneities, and so ultrasonic measurements can be used to characterize fine structure. Mathematical descriptions of scattering interactions are often extremely complex and in many cases of only moderate complexity cannot be readily solved. One successful application of scattering theory is the ultrasonic characterization of emulsions. (Other, more complex, theories are available for nonspherical particles (Ahuja and Hendee, 1978).) Emulsion droplets are spherical and (usually) considerably smaller than the wavelength of sound. These two conditions make the problem much more approachable. The two stages to predicting the ultrasonic properties of a dispersion are: (1) calculate the scattering from an individual particle; and (2) calculate the interactions between the scattered and nonscattered waves from an ensemble of particles. Because the particle is small compared to the wavelength of the sound, it experiences an effectively uniformly changing pressure. The pressure gradient induces fluid flow and both the particle and the surrounding medium will tend to move. However, because they have a different inertia, the droplet and continuous phase will move slightly out of phase and the particles will be seen to oscillate in the acoustic wave (Figure 3). The moving particle itself acts as a dipolar source, scattering sound mainly forwards and backwards. As well as moving in the pressure gradient, the particle and surrounding media will also be compressed. If the particle has a different compressibility to the surrounding fluid it will be differentially compressed and seen to pulsate. A pulsating droplet acts as a monopolar sound source scattering sound waves homogeneously in all directions (Figure 3). In fact the actual amount of sound energy scattered by these mechanisms in the long wavelength regime is very small and the losses are dominated by the inefficiencies of the interactions. (Scattering becomes dominant as the particle radius approaches the ultrasonic wavelength.) The friction between the oscillating particle and the surrounding fluid converts sound

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

113

FIG. 3 Diagram illustrating viscoinertial and thermal scattering from an emulsion droplet in an ultrasonic wave. Viscoinertial scattering is due to the particle oscillating in the continuous phase as a result of the inertial mismatch and acting as a dipolar source. Thermal scattering is due to pulsations of the particle and scatters waves in all directions (i.e. monopolar).

energy to heat (viscoinertial losses) and the compression of the particles causes a differential heating effect. Thermal diffusion is inefficient and energy is lost as heat. Once the scattering from a single particle is understood, the next stage of the problem is to combine their effects to predict the ultrasonic properties of a reasonably concentrated ensemble of particles. The main approaches to this problem are the core–shell models and the use of multiple scattering theories. Both of these are well described in the literature (McClements, 1992; Hermar et al., 1997) and only the former will be described further here. Multiple scattering theory (Waterman and Truel, 1961) uses an arithmetic sum of terms to calculate the interactions between the primary acoustic wave and the waves scattered from other particles. Each of the terms can be calculated from a series of complex simultaneous equations (McClements and Povey, 1989) and complete solution of the problem can require the extremely laborious calculation of more than 20 of these. However, in the long wavelength limit for moderate concentrations of scatterers, adequate

114

J. N. COUPLAND AND R. SAGGIN

precision can be achieved by considering only the first- and second-order terms (McClements, 1996): K

2

冢k 冣 1

=1–

i4πN(A0 + 3A1) 48π 2 A0 A1 – k13 k16

(13)

where k1 is the complex wavenumber of the continuous phase, N is the number of droplets per unit volume (= 3φ/4π r 3, φ is the volume fraction and r the droplet radius). The velocity and attenuation predicted from the theory are contained in the real and imaginary part of K, the complex wavenumber of the emulsion (= ω /c + iα). The terms A0 and A1 are analytical solutions to the scattering equations describing the thermal and viscoinertial losses. They are functions of the particle radius, wavenumber (i.e. velocity and attenuation), the physical properties (density, thermal conductivity, viscosity, cubic expansion coefficient and specific heat at constant pressure) of the component phases and the frequency of the sound (McClements, 1996). Using published values for the physical constants it is possible to use Eqn (13) to predict velocity and attenuation as a function of frequency and the concentration and size of the particles present. If necessary, the terms in radius can be replaced by a size distribution

(a)

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

115

(b)

FIG. 4 Theoretical prediction of the ultrasonic (a) attenuation and (b) velocity spectra of a dispersion of corn oil droplets (φ = 1%, 10% and 30%) in distilled water. The calculations were performed using a simplification of multiple scattering theory as described in the text using literature data for the physical properties of the component phases. Note that the spectra are size dependent over a limited frequency range and concentration dependent over the whole range.

function – most commonly a log-normal function that only requires the introduction of polydispersity as an additional unknown. Theoretical predictions of the acoustic spectra for various concentrations of fine corn oil droplets in water are shown in Figure 4. (Note that in Figure 4 frequency and particle size are both contained in the x-axis and losses are expressed as attenuation per wavelength, αλ.) At high and low frequencies the ultrasonic properties of a dispersion are independent of the size of the particles, but over a critical range velocity increases sigmoidally and there is a peak in attenuation. For micrometer-sized droplets this occurs in the range of 104–108 MHz and size measurements are possible from ultrasonic spectra made in this range. (More complete theories greatly extend this range.) Particle sizing can then be achieved by measuring the ultrasonic properties of the emulsion over this region and iteratively finding the best value of radius (and perhaps also concentration and polydispersity) to fit the theory to the experimental data. Both velocity and attenuation can be

116

J. N. COUPLAND AND R. SAGGIN

used for particle sizing (Coupland and McClements, 2001b), although the latter is usually preferred as it is often less sensitive to temperature fluctuations (Chanamai et al., 1998a). These calculations can be somewhat laborious with current computing power and simpler solutions are available under more limited circumstances (Wang and Povey, 1999). As with any curve-fitting operation, the fewer adjustable parameters (i.e. size, volume fraction, polydispersity and physical constants), the more rapid and satisfactory the solution. Where possible the physical properties of the component phases should be measured independently for the system under investigation, but, failing that, literature values are available (Coupland and McClements, 1997). Babick et al. (2000) used numerical analyses to show the effects of uncertainty in material parameters on the results of a sizing operation. If possible, it is advantageous to determine the volume fraction of the emulsion from an independent chemical or density measurement, leaving only the hard-to-measure microstructural parameters unknown. The simplification of multiple scattering theory described here has provided a reasonable description of experimental results up to moderate particle concentrations (McClements, 1992), but the full theory is required for concentrated (>~10%) particles or outside the long wavelength limit. More complex theory based around multiple scattering or core–shell theories have been shown to work over a wide range of particle sizes and concentrations (Chanamai et al., 1999; Herrmann and McClements, 1999). Recently, a theory has been developed to describe the ultrasonic properties of flocculated emulsions (McClements et al., 1998). As we have seen, the compression of particles in a sonic field generates waves of heat that are inefficiently dispersed into the continuous phase. When the droplets are closely associated, for example in a floc, the thermal waves overlap and heat loss is reduced, leading to a decreased ultrasonic attenuation. This effect is treated theoretically by assuming the flocculated emulsion can be treated as a two-phase system, which consists of spherical “particles” (the flocs) dispersed in a continuous phase. The flocs are treated as an “effective medium” whose physical properties (e.g. density, viscosity) depend on the size, concentration and packing of the droplets within them. The ultrasonic properties of the flocculated emulsion can then be calculated by considering the interactions of sound with (a) a single particle within the effective medium of the floc and (b) the floc with the continuous phase. Sample calculations for flocculated emulsions are presented in Figure 5. In this figure the total oil–water ratio is held as a constant but the proportion of the (1 µm) droplets present in (10 µm) flocs is changed. This theory has been shown to give good agreement with experimental measurements of flocculated oil-in-water emulsions (Chanamai et al., 1998b). At high frequencies the flocs scatter sound themselves and cause

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

117

FIG. 5 Theoretical prediction of the attenuation of a fine (1 µm water emulsion with varying proportions of the droplets present in larger (10 µm diameter) flocs. The more flocculated droplets have a higher attenuation at high frequency and a lower attenuation at low frequency.

the additional attenuation observed. At low frequencies the thermal overlap effects discussed above are responsible for the reduced attenuation. D. WAVES AT INTERFACES

Inevitably to make an ultrasonic measurement the wave must pass across an interface where it will be partly reflected (Figure 6). The proportion of transmission and reflectance at an interface is governed by the acoustic impedances of the component phases. The specific acoustic impedance, Za, is defined as: Za =

P ωρ = κ = Ra + iXa ξ

(14)

where P is the acoustic pressure. (Note the shear impedance will be different to the longitudinal impedance.) As with all acoustic parameters characterizing both the elastic and inelastic parts of the wave, acoustic impedance

118

J. N. COUPLAND AND R. SAGGIN

is complex, expressed as the complex sum of Ra the resistive (real) and Xa the reactive (imaginary) part. In many cases, the attenuation is low and the reactive part of the impedance can be neglected, so it is possible to simplify Eqn (14) to Z = ρc, where Z is the characteristic impedance. The characteristic impedances of some materials are provided in Table I. The amount of energy reflected at a plane interface can be expressed in terms of the ratios of either the amplitude or the intensity of the reflected (subscript r) to the incident (subscript i) wave. (In fact it is possible to define the reflection in terms of almost any characteristic parameter, for example the local pressure maximum, particle velocity, particle acceleration.) As intensity (I) is proportional to the square of amplitude (a), this leads to two commonly used, and frequently confused, reflection coefficients (R): Ra = RI =

ar Z2 – Z1 = ai Z1 + Z2

Ir Z – Z1 = 2 Ii Z2 + Z1

冢

(15) 2

冣

(16)

When sound that is traveling from a medium of low acoustic impedance encounters a boundary with a second medium of high acoustic impedance (e.g. traveling from air to a solid), the ultrasonic waves are almost entirely reflected (R tends to unity). When the boundary is between media of similar impedance, then R tends to zero and the materials are said to be acoustically matched, or ideally coupled. By a similar analysis, the transmission coefficient (the ratio of the transmitted and incident waves) is: T=

4Z1 Z2 It = (Z1 + Z2)2 Ii

(17)

The situation is more complex when the wave encounters an oblique interface and part of the sound is reflected and part of it is refracted, as illustrated in Figure 6(a). The (ultrasonic) refractive index (n) is given by Huygens’ principle and Snell’s law as: n=

sin θ1 c1 = sin θ2 c2

(18)

where c1 and c2 are, respectively, the velocity of waves in medium 1 and 2, θ1 is both the angles of incidence and the angle of reflection, and θ 2 is

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

119

the angle of refraction. The reflection and the transmittance coefficients are given by: Z2 cos θ1 – Z1 cos θ2 2 cos θ1 – Z1 cos θ2

2

R=

冢Z

冣

(19)

T=

4Z1 Z2 cos θ1θ2 (Z2 cos θ1 + Z1 cos θ2)2

(20)

When θ1 and θ2 are both equal to 90°, i.e. normal incidence, cos θ = 1 and Eqns (19) and (20) reduce to Eqns (16) and (17). When longitudinal waves reflect with a certain critical angles at a boundary, they are subjected to a mode conversion and they can generate different types of wave (Rose, 1999). For example, when the velocity of

(a)

(b)

120

J. N. COUPLAND AND R. SAGGIN

(c)

FIG. 6 Interaction of a wave with an interface. (a) Oblique incidence leads to partial reflection and partial transmission (refraction). Because the incident wave is oblique to the interface, shear and longitudinal components will be generated (not shown). (b) When the angle of incidence is equal to the first critical angle (φcr1), the longitudinal transmitted component is guided along the interface. (c) When the angle of incidence is equal to the second critical angle (φcr2), the shear transmitted component is guided along the interface.

sound in medium 2 is greater than in medium 1, the refracted longitudinal waves are deviated away from the normal. The first critical angle (θcr1, Figure 6(b)) occurs when the angle of normal wave refraction θ2, is equal to 90° so that the transmitted longitudinal sound moves along the interface and the shear component continues into the second medium. The second critical angle (θcr2, Figure 6(c)) is when the shear component is refracted along the interface and the longitudinal component is completely reflected.

E. DIFFRACTION

Our discussion of sound propagation so far has treated the acoustic waves as a coherent beam, and it is possible to find analogous expressions for many of the equations used here in the field of laser optics. However, sound waves are not coherent and we must also account for diffraction. Diffraction is particularly important as beam spreading can mean the detector can receive only a portion of the sound energy produced and the apparent attenuation of the material is overestimated. An ultrasonic transducer can be considered as a piston source. A crystal vibrates under an applied voltage and the force wave propagates away from it as ultrasound. The area in front of the acoustic source is divided in two

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

121

FIG. 7 Diagram showing the near- and far-field regions in an ultrasonic wave emitted from a plane source. In the near-field region the wavefronts are parallel to the source but the pressure fluctuates wildly with distance. In the far-field the pressure fluctuates sinusoidally with distance and the wave appears to diverge from the center of the source.

regions (Figure 7): the near and far fields. In the near zone, the beam of sound is parallel to the generating source, while in the far zone it diverges apparently from a point at the center of the source. The value of dnz, i.e. the limit of the near zone, is calculated from the following equation: dnz =

D2 4λ

(21)

where D is the circular source diameter. Measurements in the far-field region must account for the loss in signal either mathematically (Khimunin, 1972) or through appropriate calibration. Beam spreading can also lead to ultrasonic reflections from the container walls that can make accurate measurement more difficult, particularly as certain angles of incidence lead to mode conversion. In the near-field region the pressure fluctuates wildly with distance and it is extremely difficult to make an accurate measurement (Povey, 1997); it is almost always preferable to make measurements in the far field. III. MEASUREMENT METHODS

In the theory section above we discussed the various ways that the ultrasonic and material properties of materials are linked. Later we will proceed to describe various applications of these relationships to food

122

J. N. COUPLAND AND R. SAGGIN

characterization, but we must first consider the practicalities of making ultrasonic measurements. In all cases we seek to define the principles and practicalities of the method as well as provide some suggestions for avoiding common errors. Several important elements of measurement system design are reviewed by McClements and Fairley (1991, 1992), Papadakis (1990a, 1990b), Crecraft (1983), and Sarvazyan (1982). In general, all the measurement systems share some common features: (a) Ultrasonic transducers. Ultrasonic transducers convert an electrical to an ultrasonic signal and vice versa. There are various approaches to transducer design but most depend on the electrically induced vibration of a crystal, i.e. the piezoelectric effect. Cutting the crystal at different angles to the atomic structure can yield transducers with different modes of vibration. When an electrically insulating crystal is compressed it will experience a deformation and electric charges will be created on the surface, i.e. positive charge on one surface and negative on the other, generating an electrostatic field within the crystal. On the other hand, when a crystal is placed in an electric field, it experiences a deformation, i.e. transducers work both as transmitter and receiver. The back of the crystal is usually protected by an acoustically insulating backing material to absorb the energy released in that direction and the front by an acoustically conducting wear plate. (b) Signal generator. A signal generator provides a controlled voltage to the transducer to generate the acoustic signal. The simplest measurements use a board-band electrical pulse to excite the transducer, i.e. an energy pulse containing a wide range of frequencies. This is analogous to striking a bell with a hammer; it is possible to make a loud noise but many frequencies of sound are generated simultaneously. More sophisticated signal generators can provide a burst of known energy composition and/or duration. Acoustic signal generators are frequently quite high voltage (~100 V) and provide a synchronization pulse to zero the time measurement apparatus. (c) Digitizer. The acoustic signal captured and converted to an electrical signal by the transducer must be stored and analyzed. This is commonly achieved using a digital storage oscilloscope or an equivalent computer card. Whatever mode of data capture is selected, it is important to ensure that the resolution is sufficient to capture the high frequency detail in the signal – usually at least three times faster than the highest frequency of interest. The digitizer must also be able to store sufficient data points to capture all the signals of interest and should have a wide dynamic range to cope with highly attenuating samples. Modern oscilloscopes provide many other useful features such as

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

123

Fourier domain editing and arithmetic functions. A single ultrasonic measurement is often rapid (~ 20 µs) so a useful amount of signal averaging can be conducted in apparently “real-time” within the oscilloscope. All of these items need to be connected with appropriate cabling. Unless very high frequency measurements are considered (> 10 MHz), conventional coaxial cabling is sufficient. The cabling should be kept as short as practically possible and should not be changed between calibration and when performing measurements. Defects in connections and cabling can cause errors in measurement often overlooked when troubleshooting a system. A. PULSED MODES

A real-world application of a pulsed sonic measurement is shouting in a canyon and waiting for the echo to return. Using the wait time (t) and the speed of sound in air (c ~ 330 m s–1), it is trivial to calculate the width of the canyon (d/2) from: d = ct

(22)

Ultrasonic measurement methods seek to apply this principle in a systemized and controlled manner (McClements and Fairley, 1991). The measurement can be conducted in either one- or two-transducer mode, as illustrated in Figure 8. In one-transducer mode the signal reflects from a material interface and is detected by the original (speaker) transducer; in two-transducer mode the second transducer acts as a receiver. In both

FIG. 8 Diagram illustrating pulsed mode measurements. Ultrasonic velocity is calculated from the time taken for the pulse to travel a known distance and attenuation is calculated from the energy loss. Measurements may either be made (a) through transmission mode, where a second transducer is required to detect the signal, or (b) in pulse-echo mode, where a single transducer is used to generate and detect the sound.

124

J. N. COUPLAND AND R. SAGGIN

cases the ultrasonic velocity is calculated using Eqn (22) as c = d/t. The attenuation of the material can also be measured from the loss in signal amplitude with distance (Eqn (3)), but it is important to account for the nonattenuation losses due to diffraction and imperfect reflection. Frequency dependence of the acoustic parameters can be determined from a Fourier transformation of the various echoes (McClements and Fairley, 1992) or by a series of a few cycles of the pure frequency of interest (tone burst). The time of flight of the pulse can be measured extremely precisely and so for precise velocity measurements a good measurement of pathlength is also required (see Eqn (23)). Conventional micrometers are not adequately precise so instead the sample cell is calibrated using a material of known ultrasonic velocity (i.e. water; DelGrosso and Mader, 1972). Calibration should be performed regularly and certainly whenever the temperature is changed or the transducers moved. For the most precise measurements it is good practice to check the calibration with another fluid of well-defined properties. B. NONCONTACT MEASUREMENTS

A significant limitation of conventional ultrasonic measurements is the huge impedance mismatch between air and the transducer delay line. Indeed, until recently it was considered impossible to propagate high frequency (~MHz) ultrasound through air. However, a new transducer technology has been developed to overcome the air–transducer impedance barrier (Bhardwaj, 1986). In particular, so-called “quarter-wave matching layers” have been devised which gradually, i.e. layer-by-layer, adapt the traveling waves to lower and lower acoustic impedances until they match that of air and thus provide transmission into air with reasonable efficiency. This procedure has been so successful that it is now possible to transmit sound through air at frequencies of several MHz. The key feature in the transducer construction is the use of new materials, specifically the use of polymers filled with microballoons whose number and size can be adjusted to tailor the acoustic impedance. Ideally the acoustic impedance of the matching layer needs to be equal to the square root of the product of the two impedances to be matched. The thickness of the matching layer needs to be equal to one-fourth of the wavelength of the wave propagating through the layer (Tittmann et al., 1998). Clearly noncontact measurements are applicable to many sensing applications where direct contact with the food surface is impractical. A typical experimental setup is shown in Figure 9. The ultrasonic signal is passed to one of two ultrasonic transducers arranged a fixed distance

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

125

FIG. 9 Diagram showing the arrangement of transducers used in a noncontact measurement. By using each transducer to measure the time for a pulse of sound to echo from the sample surface and to transmit to the second transducer, it is possible to simultaneously measure the ultrasonic velocity in the sample and its thickness.

apart, in good alignment with each other. The ultrasonic transducer converts the electrical signal to sound that propagates through the air and is detected by the second transducer at a time t1 later. If there is a sample between the transducers, part of the signal is reflected from the front face of the sample and the reflection detected by the first transducer after an elapsed time of t2. The operation is then repeated in reverse, so measuring a transmitted time signal t4 at the first transducer and, if there is a sample present, an echo at time t3 at the second transducer. Measurements of t1 and t4 are recorded in the absence of a sample and t1–4 in the presence of a sample. The thickness (d) and speed of sound (c) of the sample can be calculated as follows (the superscripts a and s refer to the air and sample, respectively): d2s = cat1a – cs =

ca (t s + t s ) 2 2 3

–2d2s (–2t 1s + t 3s + t 3s )

(23) (24)

(Note that in Eqns (23) and (24), t1 could be replaced by t4 without changing the result.) Attenuation can be calculated from similar measurements of signal power.

126

J. N. COUPLAND AND R. SAGGIN

Aside from the technological breakthroughs required in its development, noncontact ultrasonic measurements are a variation of the pulsed methods described above. A key advantage is that thickness is determined directly by the acoustic wave and not measured independently. Precise, contactmode measurements are restricted to fluids confined in defined geometries and solids can only be measured by (imprecisely) measuring the thickness with calipers. C. RESONANCE METHODS

Although pulsed modes are the most frequently used methods of ultrasonic food characterization, they have an important disadvantage of requiring a long pathlength to allow the signal to decay sufficiently for an attenuation measurement and for the arrival time to change enough to distinguish velocities. Consequently the sample volumes required are rather large and it can become difficult to maintain adequate temperature control for the highest precision measurements. An alternative approach is to exploit the resonance properties of a container holding the sample (Sarvazyan, 1982). A pair of aligned transducers is placed either directly in the fluid or around a container holding the fluid. One transducer generates a continuous pulse of pure varying frequency and the second acts as the detector. (A second, less common, type of resonator uses a fixed frequency and variable pathlength to generate the resonance peaks. Aside from this, the principles of measurement are similar.) The wave will echo backwards and forwards between the two transducers and when the wavepath is a whole number of half-wavelengths it will constructively interfere and appear as a peak in the frequency/amplitude plot, i.e. the condition for resonance is given by: n

λ =l 2

(25)

where n is the (integer) number of the resonance peak, λ is the wavelength, and l is the pathlength of the resonator. The wavelength is hard to measure independently, but can be calculated as c␯–1. Therefore the change in resonance frequency (dv) of a given peak can be related to a change in velocity: dc d␯ = c ␯

(26)

Alternatively, the velocity can be calculated from measurements with a calibration liquid:

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

127

c = 2l(␯n – ␯n–1)

(27)

where vn and vn–1 are the resonance frequencies of adjacent peaks. This approach is somewhat simplistic since it neglects the effects of diffraction and nonideal reflection, but these factors can also be dealt with by additional calibration, often using sodium chloride solutions. The losses of a filled resonance cell are often expressed as the Q-factor of the system (cell plus sample) calculated from the width of a resonance peak (Buckin and Smyth, 1999). The contribution of the cell to the Q-factor can be calculated from a suitable attenuation calibration (often magnesium or manganese sulfate solutions) to give the loss properties of the fluid. The Q-factor of the sample is a function of attenuation: Q sample = π/αλ

(28)

The most precise ultrasonic characterization of fluids is by resonance methods; approaching 0.3 × 10–4% for the best systems (Buckin and Smyth, 1999). Ultra-high precision requires optimization of all of the factors in the measurement system, importantly including temperature. Commercial water baths can easily control temperature to 0.01°C, but ultrasonic velocity precision of the type described above requires a two order of magnitude improvement over this. In practice, precision measurements are made using paired cells in differential mode – one cell measures temperature with a calibration fluid while the other makes high precision measurement of the fluid under consideration. D. REFLECTANCE METHODS

The magnitude of an echo returning from an interface is related to both the ultrasonic properties* of the two phases (Eqn (16)) and the magnitude of the original pulse. If one of these materials is fixed, then the magnitude of the returning echo is solely dependent on the acoustic properties of the second material and on the magnitude of the incident pulse. This can be exploited as a sensing modality, as changes in the magnitude of an echo returning from the interface between the container and its contents will depend only on changes in the contents. Alternatively, a delay line can be coupled to the transducer and the reflection from the interface between the delay line and the sample surface measured. In practice it is difficult to ensure that the performance of the transducer and signal generator is constant on a day-to-day basis, so it is better to *Specifically impedance; these techniques are also known as impedance methods.

128

J. N. COUPLAND AND R. SAGGIN

normalize the reflection amplitude to that of a calibration material such as water. Alternatively it is possible to use a dual delay line made of two materials linked in series so there are echoes from the interface within the delay line and the delay line–sample interface. The two echoes can be used to solve the two unknowns in the measurements system (i.e. magnitude of the generated signal and ultrasonic properties of the unknown sample). When necessary, reflectance (R) can be calculated as a complex parameter from the magnitude (M) and phase (θ) of the reflected echo from the sample (subscript s) and calibrant (subscript c) obtained from a Fourier transformation (McClements and Fairley, 1992): Rs = Rc

Ms exp [i(θs – θc)] Mc

(29)

Reflectance measurements are often suitable for on-line sensing, as they do not require a fixed and invariant pathlength for the sound within the material of interest, which may not be available in some process equipment. However, reflectance is clearly only sensitive to the surface of material close to the container and may not be representative of the bulk. This may be a particular problem for foods in containers prone to surface fouling. Reflectance is the only effective way to characterize the shear ultrasonic properties of fluids, as shear waves cannot penetrate sufficient distances for transmission measurements. Related to normal reflectance measurements are the so-called “guided wave” sensors where an acoustic wave is partly trapped inside a container or pipe wall and propagates a significant distance (several meters) along by multiple internal reflections. At each reflection a portion of the sound leaks into the surrounding material so the detected signal is sensitive to changes in composition. Guided wave sensors are particularly valuable in detecting fouling in pipes and hard to reach equipment (Rose, 1999).

E. LOVE WAVE SENSORS

Love waves are shear, horizontally polarized waves that propagate in a thin surface layer. There are some losses into the media surrounding the guiding layer and the wave properties are therefore sensitive to material properties. A major advantage of Love wave sensors is they can be manufactured very cheaply and compactly using printed circuit board technology onto silicon wafers of thickness in the order of 500–1000 µm. They can also be wirelessly interrogated and can operate in pulse echo, through transmission, or in resonance modes.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

129

Curtin and coworkers (1998) developed a micromachined wet cell for Love wave measurements particularly suitable for foods and fluids characterization. A micromachined channel shields all but the wave path from the influence of the liquid and a heater controls the local temperature. A similar design, incorporating paired Love wave sensors, was used by Varadan and Gardner (1999) for a variety of sensing applications. In these cases, one of the sensors is shielded from the fluid and is thus only sensitive to temperature (or mechanical stress), while the other is exposed and responds to both temperature and fluid composition. With two such sensors it is possible to measure two unknowns simultaneously. Arrays of Love wave sensors are particularly exciting food sensors as controlling the masking layer can make them sensitive to different components. In the previous example a single mask allowed simultaneous temperature and fluid property measurement. Using selective masks, similar to those currently employed for ion selective electrodes, it would be possible to develop a genuinely multicomponent ultrasonic sensor.

IV. APPLICATIONS

Even good quality measurements of ultrasonic velocity and attenuation have little value in themselves, but only as they relate to properties determining food quality and to parameters useful for process control. Under certain circumstances the relationships used can be based directly on the underlying physics outlined in the theory section, but in most cases they are empirical. Empiricism should, as always, be treated with caution. Any established link is valid only for the sample set considered and apparently small variations outside this set may render the results meaningless (for example, a correlation between ripeness and ultrasonic velocity for one variety of apple may not hold for another variety). With a mechanistic relationship, it is explicit what is being measured and so wherever possible the physical principles should be at least applied qualitatively to justify an apparent correlation. As we introduce some applications, we shall attempt to stress their theoretical basis wherever possible. A. LEVEL SENSORS

Perhaps the simplest ultrasonic measurement, one exploiting the gross differences between a fluid and a gas along with the useful ability of sound to make measurements through an opaque solid, is level sensing. Various modes of measurement are available:

130

J. N. COUPLAND AND R. SAGGIN

(a) Time-of-flight sensors. An ultrasonic transducer is placed at the base of the tank and the time taken for a pulse passing up through the liquid and reflecting from the surface is used as a measure of depth. This is precise if the surface is flat and smooth enough to clearly detect a returning echo. This approach has been applied to controlling the filling operation for canned beverages using either a transducer mounted below the fluid (Ridgway et al., 1999) or an air-coupled transducer mounted above (Griffin et al., 2001). (b) Presence/absence sensors. A series of pairs of ultrasonic transducers are aligned with each other at various heights in a column of liquid. The depth of the container is given by the highest pair that can transmit a signal (based on the principle that air cannot support ultrasound but the fluid can. The resolution of this method is limited by the spacing of the transducers. (c) Internal reflection sensors. A pulse of sound is passed obliquely into a bar of solid immersed in the liquid. The sound echoes repeatedly within the solid waveguide and is detected either by a second transducer or in reflectance mode. Each echo depends on the material outside the bar at that point, i.e. either liquid or gas. In general more sound energy will be lost into the liquid so total attenuation of the signal can be related to the amount of the bar immersed in the liquid. Placing the bar appropriately inside a tank can provide a measure of depth. B. FLUID FLOW

Another widely used group of methods determine the rate of flow of liquids from longitudinal ultrasonic velocity measurements. Again, there are three groups of methods (Lynnworth, 1989): (a) Transit-time mode (Figure (10a)). Two transducers are clamped outside a pipe at a known distance from each and angled to the direction of flow either in V-mode (the transducers are mounted on the same side of the pipe; illustrated in Figure (10a)), W-mode (the sound traverses the pipe four times) or in Z-mode (the transducers are mounted on opposite sides of the pipe). By alternately using the paired transducers as transmitters and receivers, two times of flight are measured, with and against the direction of flow. The speed of sound measured with the flow (downstream) is greater than that measured upstream as the translational motion of the fluid supporting the wave adds to (and subtracts from) the movement of the ultrasonic wave. (If the fluid were not moving there would be no difference in measured ultrasonic

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

131

FIG. 10 Diagrams of various types of ultrasonic flowmeter. (a) Transit time mode: Paired transducers acting alternately as sources and detectors measure the time of flight of an acoustic pulse with and against the flow and use the difference to calculate the speed of flow. (b) Doppler flowmeters: Sound reflected from scattering particles entrained in the fluid is used to calculate flow rate. (c) Vortex shedding: Sound is scattered from vortices downstream from a strut introduced into the flow and related to flow rate.

velocity.) The liquid velocity (V) in the pipe can be calculated from the difference in time of flight (∆t) of the sound traveling downstream and upstream through the following equation: V = BL ∆t

(30)

where B is a constant and L is the distance between the two transducers. The constant B is usually either fixed empirically or calculated as a function of the Reynolds number. Commercial transittime mode devices are available working in the temperature range –40 to 150°C, in pipe of diameter 1–500 cm, and typically perform best at high flow rates. (b) Doppler flowmeters (Figure (10b)) are based on the principle that the frequency of ultrasonic waves reflected from suspended particles or gas bubbles (discontinuities) within the moving medium is shifted in proportion to the velocity of the moving discontinuities. Current technologies require for a 1 MHz transducer approximately 25 ppm

132

J. N. COUPLAND AND R. SAGGIN

of entrained particles in the liquids, and these inhomogeneities should be at least 30 µm in diameter to provide an adequate reflected signal. To obtain a good signal and high precision during measurement, some conditions need to be satisfied; specifically the pipe must be full and flowing above a certain minimum velocity. Also the scattering particles must be moving at the same rate as the liquid. In many cases turbulence is also required. (c) Vortex shedding (Figure (10c)). A bluff body (the vortex shedder) is fixed to the wall of the pipe, forcing the liquid to flow turbulently around it, producing a series of downstream vortices known as the von Kármán vortex street. These vortices have a characteristic frequency that is proportional to the fluid velocity. Using an ultrasonic device either in transmission or in reflection mode, and knowing the geometry of the shedder and the Reynolds number, the velocity of a liquid can be measured with an accuracy of 1%. Vortex shedders are particularly vulnerable to fouling in fluid streams containing a large amount of suspended solids. C. TEMPERATURE

In most media the speed of sound is a function of temperature (see Figure 11 for examples) and so measurements of ultrasonic velocity can be potentially used as a thermometer. For example, at room temperature the speed of sound increases approximately 3 m s–1 per degree Celsius. A typical pulse echo instrument is capable of measuring velocity to within 0.1 m s–1, implying a precision in temperature of 0.3°C. (Clearly the more precise resonance methods would be at least an order of magnitude better.) However, when Richardson and Povey (1990) used a tone-burst technique to measure the speed of sound of a fluid flowing in a pipe, they were only able to make measurements within confidence limits of ±1°C. The velocity–temperature function used as a calibration for this technique depends on composition and structure so any changes in the food caused by heating would make the temperature measurements unreliable. Recently Sigfusson and coworkers (2001) measured the speed of sound in slabs of food undergoing unsteady state cooling. They calculated the theoretical speed of sound as a function of temperature in their food using composition-dependent equations developed by McClements and coworkers (Ghaedian et al., 1998; Chanamai and McClements, 1999; Sigfusson et al., 2001) and showed that the measured values were in good agreement with an integral of this function over the thickness of the slab calculated from the modeled thermal gradient. They further showed measured ultrasonic velocity is a product-specific linear function of core temperature and

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

133

FIG. 11 Speed of sound (~1–2 MHz) as a function of temperature for (䊊) distilled water, (䉭) 20% sucrose solution, (䊉) corn oil, and a fine corn oil-in-water emulsion (䊏) φ = 10%, (䊐) φ = 50%. Note that while the speed of sound in water and aqueous solutions tends to increase (to a maximum at ~70°C) with temperature, it decreases in oils. The ultrasonic velocity in mixed oil–water systems shows a response to temperature somewhere between the constituents depending on their volume fraction. At the temperature where the speed of sound is similar in oil and water (~20°C), it is insensitive to their ratio in a mixed system. Data were either measured by the authors or taken from various literature sources.

therefore a good sensor for product chilling. In a somewhat similar study, Haeggstrom and Luukkala (2000) measured various ultrasonic properties of fried beef burgers during cooling (74–45°C) and found the best prediction of cooling from the attenuation of 300 kHz sound. An alternative approach to ultrasonic thermometry is to place a rod in contact with the food. The rod expands with temperature and its length can be accurately measured by propagating an ultrasonic pulse along its length and measuring the reflection from the far (free) end. An advantage of this approach is that it does not need to be recalibrated for each food, but its important disadvantage is that it is invasive. Richardson and Povey (1990) evaluated several rod shapes and were able to achieve confidence limits to their measurements of 0.3°C using a tapered design. D. COMPOSITION

Some of the most simple and most widely used ultrasonic applications are composition sensors. The speed of sound in many solutions increases nonlinearly with increasing concentration; several examples are shown in

134

TABLE II REPORTED DETERMINATIONS OF COMPOSITION FROM ULTRASONIC MEASUREMENTS

Ultrasonic parameters

Several simple sugars

Velocity (by resonance 1 MHz) Velocity

Alcohol and solids in wine

Measurement conditions 20–45°C 30 and 65°C 12–20 vol% ethanol 2–14% wt% sucrose 20°C 13–18% yeast 10.8–11.3°Plato 5–50°C

Yeast slurry and wort (°Plato) in brews

Velocity (3.6 MHz)

Fruit and vegetable juices, oils, sauces, wines, syrups Sugar in solution and in juices Aqueous solutions of starch and gelatin

Velocity

Absorption (0.6–9 MHz)

Various monosaccharides

Velocity

20–80°C 0–1.2 M

Sucrose

Velocity and absorption (5–25 MHz) Velocity

20–40°C 0–5 mM

Monosaccharides, disaccharides, and methyl pyranosides

Velocity (0.1–10 MHz)

10–110°C 0–60 wt% sugar 30–60°C 0.1–0.6 wt%

5, 15, 25°C 0–50 mM

Comments

Reference

Study of hydration properties of sugars Good correlation (r2=0.99) with composition

Shiio (1957) Winder et al. (1970)

Calibration curve for wort and Feil and Zacharias, yeast and then measurements (1971) in brews Effect of temperature, water content Zacharias and Parnell and alcohol content on (1972) ultrasonic velocity Error ±0.25% Fedotkin et al. (1981) Excess absorption was noted in Reddy and starch solutions at various Suryanarayana concentrations and temperatures (1981) Adiabatic compressibility was Smith and Winder calculated from velocity and (1983) density. Berchiesi et al. (1987) Molar compressibility and volume were calculated from velocity and density

Kaulgud and Dhondge (1988)

J. N. COUPLAND AND R. SAGGIN

Food

TABLE II (continued) REPORTED DETERMINATIONS OF COMPOSITION FROM ULTRASONIC MEASUREMENTS

Ultrasonic parameters

Nucleic bases

Velocity (7.1–7.4 MHz)

Sugar content in fruit juices and beverages

Velocity (2.5 MHz)

Cod fillet

Velocity and attenuation (3.5 MHz) Velocity and attenuation (3.5 MHz)

Catfish, cod, flounder, mackerel and salmon Fish analogs made from dried cod powder, sunflower oil and water

Velocity and attenuation (3.5 MHz)

Atlantic mackerel

Velocity and attenuation (1–6 MHz)

Measurement conditions

Comments

15–35°C Apparent molar adiabatic and isothermal compressibility were 0.5–1.5 mg per g H2O calculated for seven nucleic basis 10–30°C Ultrasonic measurements predicted sugar content to within 0.2% (pure sugar) or 0.5% (mixed sugars) and was sensitive to sugar type 5–35°C Relation between moisture and speed of sound (r 2 > 0.8). No relation with attenuation 5–35°C Measurement at different temperatures for multicomponent (fat, moisture, ash) analysis 5–35°C Measurement at different temperatures for multicomponent (fat, moisture, ash) analysis 5–25°C Measurement at different temperatures for multicomponent (fat, moisture, ash) analysis. Attenuation was insensitive to food composition

Reference Buckin (1988) Contreras et al. (1992)

Ghaedian et al. (1997) Suvanich et al. (1998)

Ghaedian et al. (1998)

Sigfusson et al. (2001)

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

Food

135

136

TABLE II (continued) REPORTED DETERMINATIONS OF COMPOSITION FROM ULTRASONIC MEASUREMENTS

Ultrasonic parameters

Measurement conditions

Sausages

Velocity (1 MHz)

4–25°C

Glycerol, sucrose, sodium chloride, tomato ketchup solutions

Velocity and reflectance (2.25 MHz)

20°C

Comments

Reference

Explained variance 99.6% for fat, Benedito et al. (2001) 98.7% for moisture and 85.4% for protein. Both velocity and reflection Saggin and Coupland coefficient could be used as (2001b) concentration sensors with comparable precision.

J. N. COUPLAND AND R. SAGGIN

Food

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

137

(a)

(b)

FIG. 12 Ultrasonic (a) velocity and (b) normalized reflection coefficient of solutions of various food samples (䉲, sodium chloride; 䊏, ethanol; 䊉, glycerol; 䉮, sucrose; and 䊊, tomato ketchup). Measurements were made using a pulse-echo device (2.25 MHz) at 25°C. Reflectance was measured from the interface between Plexiglas and the sample and was normalized to the signal from water.

Figure 12(a) and Table II. Clearly any of these correlations could be used, with caution, as a calibration curve for a concentration. The first major problem is that ultrasonic velocity is a function of overall composition as well as other factors, importantly including microstructure and temperature, and so concentration measurements require the other factors to remain unchanged. It is usually impossible to measure multiple components simultaneously. Ultrasound can, however, be combined as

138

J. N. COUPLAND AND R. SAGGIN

part of a measurement suite, for example for the simultaneous determination of alcohol and sugars in beverages from combined velocity and density measurements. One way for making multicomponent measurements using ultrasound exploits the different temperature dependencies of the speed of sound in oil and water (Figure 11). At approximately 20°C, the speed of sound in an oil–water system is independent of the concentration of oil and so measurements at this temperature can be used to measure the composition of the aqueous phase (Chanamai et al., 1998a). Moving away from this temperature the dependence on oil/water ratio increases and can be measured. This principle has been used in a dairy products analyzer developed in the United States by Winder and coworkers (1961) and the BIOTEST instrument developed in the Soviet Union (cited in Buckin and Smyth, 1999). It has also been used for the determination of fat and nonfat solids in various fish (Ghaedian et al., 1997, 1998; Suvanich et al., 1998; Sigfusson et al., 2001) and meat products (Chanamai and McClements, 1999; Benedito et al., 2001) and for the determination of alcohol in wine (Winder et al., 1970). Other ultrasonic parameters can also be used to measure concentration, but it should be stressed these often covary with ultrasonic velocity so may not be compatible as multicomponent sensors. Recently we have shown (Saggin and Coupland, 2001b) that the proportion of a longitudinal ultrasonic pulse that is reflected from the surface of a variety of foods depends on concentration (Figure 12) and that the precision available using a simple pulsed methodology was comparable to that for velocitybased measurements (for example, the amount of water added to a ketchup sample could be determined to within 2% from velocity measurements and within 1.5% from reflectance). Reflectance measurements can be a better technique for on-line analysis when the geometry of the system is not conducive to velocity measurements, but only in cases where the surface is representative of the bulk. Ultrasonic attenuation tends to increase only slightly with concentration for many food solutions and so is rarely used as a concentration sensor. However in some cases, particularly for some ionic solutions and for dispersions, attenuation has a useful dependency on concentration. An especially important application of composition sensing is using ultrasonic measurements to predict the composition of animals and carcasses. One of the simplest versions of this is exploiting the correlation between backfat thickness (measurable ultrasonically) and bulk carcass composition (Fisher, 1997). Knowing the different speed of sound for lean and fat tissue, it is also possible to measure fat deposition in the carcass. However, ultrasonic velocity in muscle tissue is a function of both composition and

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

139

microstructure (texture), so the results may be somewhat convoluted (Abouelkaram et al., 2000). E. PHASE TRANSITIONS

The speed of sound is typically much greater in a solid than in a liquid, and thus ultrasonic sensors are particularly applicable to monitoring phase transitions in both lipid- and water-based systems. There are two main groups of methods – isothermal and temperature scanning – although these are based on the same principle. 1. Isothermal. Various authors (Miles et al., 1985; McClements and Povey, 1987) have shown that the speed of sound in liquid oil increases by approximately 3 m s–1 per 1 wt% increase solid fat present and exploited this to measure fat crystallization (Coupland, 2001). If ultrasonic velocity can be measured to within 0.1 m s–1, then this implies that a change of as little as 0.03% in solids content can be determined ultrasonically. Some theoretical justification for the observed data can be gained by extending Eqn (7) for a two-phase system by using volume fraction weighted averages of density and adiabatic compressibility: κ = (1 – φ)κ1+φκ2 and ρ = (1 – φ)ρ1+φρ2 (subscripts 1 and 2 refer to the dispersed and continuous phases, respectively). The compressibility of each phase can be calculated from independent measurements of velocity and density and the solid fat content (SFC) as the positive solution of the quadratic equation: – B – B2 – 4AC 2A

SFC =

(31)

where

ρ ρ A = c 21 1 – ρ1 + c 22 1 – ρ2 B = c 22 C = c 22

冢 ρ 冢ρ

2 1

冣 冢 ρ – 2冣 + c ρ 2

c21 1– 2 c

冢

2 1

1

冣

1 2

冣

and c is the speed of sound in the semi-solid fat. This model is based on several at least partly unrealistic assumptions (i.e. small density difference between phases and solids fat particles much smaller than ultrasonic

140

J. N. COUPLAND AND R. SAGGIN

wavelength), yet gave a good prediction up to approximately 20% solids (McClements and Povey, 1987). These workers also showed that this method of SFC determination was superior to NMR for low levels of solids (McClements and Povey, 1987). However, an important practical difficulty arises at higher solid fat content when the fat begins to scatter sound significantly, perhaps from voids within the structure (Cebula et al., 1990) or from air entrained in the viscous mixture, and it becomes extremely hard to make a measurement. Another problem with many fats (particularly cocoa butter) is that the contraction of solidification is sufficient to pull the sample away from the wall of the container and form an air gap large enough to prevent ultrasonic propagation. Despite these difficulties, Garbolino et al. (2000) used an ultrasonic velocity meter as an on-line sensor to demonstrate the effect of applied shear on the onset of lipid crystallization. Neither of these disadvantages arises when the oil is present as the dispersed phase of an emulsion, and some of the strongest applications of SFC determination have been made in dispersed systems. The relevant theory is largely similar and McClements (1988) showed that the ultrasonic velocity in a mixed alkane oil emulsion increases almost linearly with SFC. The technique has been used to track the crystallization kinetics of emulsified alkanes (Hindle et al., 2000), palm oil (Hodate et al., 1997; Kloek et al., 2000) and cocoa butter (Hindle et al., 2000). Most of the published work reported in this field suggests that the increase in ultrasonic velocity with solid fat is largely independent of solids composition (i.e., polymorphism), but Kloek and coworkers (2000) have suggested there may be some dependency on the polymorphic form of the fat. As polymorphic forms differ in density and crystal packing, it seems likely that there is at least some effect. The same principles apply equally well for aqueous phase transitions, but have been applied less widely. Miles and Cutting (1974) showed that the speed of sound is much greater in frozen meat and thus the solid/liquid ratio (and thus enthalpy) of partly frozen product could be calculated from an ultrasonic velocity measurement.

2. Temperature scanning measurements Changes in ultrasonic properties, usually velocity, measured as a function of temperature, can be used to identify phase transitions in lipid and aqueous systems. For example, the speed of sound in a lipid system is changed by temperature according to two factors: (1) the speed of sound in liquid oils decreases with temperature (Figure 11) and (2) the speed of sound in oils

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

141

FIG. 13 Ultrasonic velocity in a fine, 20 wt% n-hexadecane-in-water emulsion stabilized with 2 wt% polyoxyethylene sorbitan monolaurate (d32 ~ 0.6 µm) as a function of temperature showing the hysteresis between cooling and heating cycles.

increases with SFC. These factors overlay to produce a typical sigmoidal function, as shown in Figure 13. The solid fat content can be measured by extrapolating the lines from the solid and liquid regions over the transition range then calculating SFC as: SFC =

(1/c2) – (1/c 21) (1/c 2s) – (1/c 21)

(32)

where c is the measured ultrasonic velocity and cs (cl) the velocity in pure solid fat (liquid oil) extrapolated to measurement temperature (Miles et al., 1985). Eqn (32) has a similar theoretical basis as Eqn (31) and relies on similar, partially justified, assumptions. This approach is comparable to differential scanning calorimetry in that it assumes that the extrapolations drawn are from pure liquid or solid fat and suffers from similar limitations. The temperature scanning approach has also been applied to emulsified oils and has identified transitions in alkanes (McClements et al., 1993; Hindle et al., 2000) and triglycerides (Hindle et al., 2000; Kloek et al., 2000). Although the changes in attenuation associated with dispersed phase transitions tend to be more modest than those seen in velocity measurements, and thus less widely reported, McClements et al (1993) noted an anomalously high attenuation as emulsified alkanes melted. The magnitude of this peak decreased with frequency (0.5–3.5 MHz). These workers

142

J. N. COUPLAND AND R. SAGGIN

postulated that the ultrasonic energy was absorbed by perturbing the fatcrystal oilliquid equilibria and was thus related to the molecular dynamics of melting. Other workers have used temperature scans of other ultrasonic properties to measure phase transitions. Recently Garbolino and coworkers (2000) showed that it is possible to measure the onset of lipid crystallization by the changing reflection coefficient of longitudinal ultrasonic waves. Shore and others (1986) that showed the attenuation coefficient was approximately four times larger in frozen beef than in unfrozen samples, but no attempt was made to measure percentage frozen from the data. Although not classically a phase transition, Mulet et al. (1999) showed the temperature dependency of the ultrasonic velocity in cheese changes as the cheese melts and were able to identify three distinct zones that could be related to thermal events in a thermogram. F. RHEOLOGICAL PROPERTIES

There have been many efforts made to correlate the texture of foods, particularly fruits and vegetables, with ultrasonic properties; some of these investigations are summarized in Table III. The correlation coefficients reported are often quite low and probably fortuitous. The factors that allow sound to propagate through complex biological structures are poorly understood and there is no reason to assume they will correlate with the factors that determine bulk texture. This problem is compounded in most fruit and vegetables by the strong scattering of sound by intracellular air, which makes transmissions measurements at high frequencies extremely difficult. Lower frequency sound tends to be attenuated less and several of the publications listed in Table III use sonic resonance techniques, a related method based on measurements of the frequencies generated by the fruit when struck. The limitations of some of these approaches arise from the lack of mechanistic basis for a relationship for the relationships sought. Longitudinal waves in fluids are only weakly related to changes in shear modulus and as K >> G, longitudinal ultrasonic properties are effectively exclusively dependent on bulk modulus. Bulk modulus is rarely measured for foods as it is thought to have limited practical significance and it is unsurprising that attempts to correlate longitudinal ultrasonic parameters with shear viscous properties have met with only limited success. Another significant problem is that there is no reason to expect the ultrasonic properties of a food to be at all related to the large-deformation or failure properties of a food (e.g. fracture, chewability) and again the correlations seen are indirect, although frequently very useful. An example of this type of relation was

TABLE III REPORTED DETERMINATIONS OF FOOD TEXTURE FROM SONIC MEASUREMENTS.

Ultrasonic measurement

Texture property

Comments

Reference

Poisson ratio (v), density, Young modulus (E), shear modulus (G)

The velocity of propagation through the whole fruit is related to velocity through quarter sections

Garett and Furry (1972)

Yamamoto et al. (1980)

Apples

Bar velocity, Vb; apparent wave propagation, Vwt; dilatation velocity, Vd

Apples

(a) Hitting sound wave Mechanical firmness and and (b) power spectrum sensory data (20–1.25 kHz) Acoustic transmission Firmness, soluble solids, profile (by sonic and maturity resonance) (5–10 000 Hz) Amplitude resonance Puncture test (0–2 kHz)

Best correlation, r = 0.77

Watermelon

Amplitude resonance (0–2 kHz)

Puncture test

Best correlation, r = 0.73

Biscuits

Velocity, attenuation

Mechanical tests and sensory brittleness

Poor correlation with sensory Povey and Harden data (r = 0.13) and only weak (1981) with mechanical data (r > 0.5)

Apples (red delicious) Apples (golden delicious)

Single and multiple frequency Affeldt and Abbott (1989) models were developed to compare acoustic data with mechanical methods. Resonance frequencies showed Liljedahl and Abbott relationship to apple softening. (1994) Resonance method was more reproducible than puncture method Liljedahl and Abbott (1994)

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

Food

143

144

TABLE III (continued) REPORTED DETERMINATIONS OF FOOD TEXTURE FROM SONIC MEASUREMENTS.

Food

Ultrasonic measurement

Texture property

Comments

Reference Sarkar and Wolfe (1983)

Reflectance

Reflectance depends on skin roughness

Tomatoes (cracks)

Reflectance

Significantly more scattering in Sarkar and Wolfe reflectance from cracked (1983) tomatoes, correlated (r = 0.71) with crack width

Potato, avocado, carrot

Velocity (50 kHz), attenuation (50 kHz), reflectance (500 kHz)

Large velocity difference between fresh and ripe avocado

Mizrach et al. (1989)

Winter-grown melons

Velocity, attenuation (50 kHz)

Modulus of elasticity, and tangent modulus

Poor correlation of velocity with internal properties of melons.

Mizrach et al. (1991)

Avocado

Velocity, attenuation

Firmness

Distinct minimum in velocity Mizrach et al. (1996) during ripening possibly related to changes in internal structure

Avocado

Velocity, attenuation

Firmness

Correlation between firmness and attenuation related to cold-softening of avocado

Flitsanov et al. (2000)

J. N. COUPLAND AND R. SAGGIN

Oranges (skin texture)

TABLE III (continued) REPORTED DETERMINATIONS OF FOOD TEXTURE FROM SONIC MEASUREMENTS.

Texture property

Avocado

Velocity, attenuation

Dry weight during ripening Negative correlation between attenuation and dry weight

Mizrach et al. (1999)

Mango

Velocity, attenuation

Softness, acidity, and sugar content during ripening

Mizrach et al. (1997)

Cooked carrots

Velocity, attenuation (37 kHz)

Compressive Young modulus Major changes in all properties Nielsen and Martens (1997) and strain at failure in the first 25 min of cooking associated with changes in air and water content.

Mahon cheese

Velocity (1 MHz)

Cheddar cheese

Texture profile analysis

Comments

Statistical model used to relate changes in ultrasonic signal to ripening in a packing house

Reference

Ultrasonic velocity decreases Benedito et al. (2000c) with curing time probably due to drying. The authors predicted a model (92% of variance) of the dependence of velocity on temperature and deformability modulus. Best measurement temperatures 0–17°C

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

Ultrasonic measurement

Food

Benedito et al. (2000b)

145

146

J. N. COUPLAND AND R. SAGGIN

seen by Benedito et al. (2000a), who showed the speed of sound increased (approximately 1650–1720 m s–1) in Mahon cheese during aging up to 1 year along with instrumental measures of large-deformation elasticity and yield values. Although the correlations were good (0.92 with modulus and 0.88 with the force maximum) the cheeses were also shown to lose about 30% of their water content over the aging period. These authors recognize that the value of ultrasound as a cheese texture sensor in this case is most likely due to both the speed of sound and texture varying with water content (Benedito et al., 2000c) and showed similar results with Cheddar cheese (Benedito et al., 2000b). A similar correlation-type approach was taken by Povey and Harden (1981) to relate the crispness of cookies to ultrasonic velocity. An interesting way around both of these problems was taken by Faeth and Chem (1999), who used low frequency (0.1 MHz), noncontact ultrasonics to measure the position of the surface of baked goods as they were deformed by a jet of air. In this case ultrasound was simply being used as a strain gauge and the ultrasonic properties of the food were not measured. These workers were able to correlate the induced deformability of the bread with baking time (r = –0.85). A more direct set of texture measurements is based on the direct relationship between shear acoustic properties and fluid viscosity. Viscous liquids are only marginally able to support shear waves and very strongly attenuate them. For low viscosity liquids (~1 cP), the shear velocity and attenuation coefficient are expressed as follows (Blitz, 1963): c2 =

2ηω ρ

(33)

α=

ωρ 2η

(34)

Both speed of sound and attenuation are frequency dependent and, while speed of sound increases with increasing viscosity, the attenuation decreases. For example, for water according to Eqns (33) and (34), 1 MHz shear waves would have a velocity of 3.5 m s–1 and an attenuation of 15 000 dB mm–1, making it impossible to transmit waves over even a few micrometers. As the viscosity increases the characteristic time for the shear wave to collapse becomes less than the (reciprocal) frequency of the wave and propagation becomes more possible (Blitz, 1963). One solution to this measurement problem is to generate a pulse of shear waves in a solid and measure the amount of this pulse reflected from

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

147

an interface with water. Recently Kulmyrzaev and McClements (2000) and Saggin and Coupland (2001a) took this approach to measuring the viscosity of honey and oils, respectively. Both groups reported good correlations between ultrasonically and conventionally measured data and Kulmyrzaev and McClements noted the magnitudes of the viscosity measured by the two methods were significantly different – perhaps due to a relaxation over the wide span of frequencies separating the two methods. Buckin and coworkers (Buckin and Kudryashov, 2001; Kudryashov et al., 2001) used a more sensitive shear wave resonator to track the sol–gel transition in milk during yogurt manufacture. The formation of a gel was monitored in time using high-frequency shear wave transducers (5–15 MHz) and dynamic rheology at a frequency of 0.1 Hz. Both methods monitored the same kinetics of formation of a gel network, which was accompanied by a significant increase of the elastic and storage modulus. However, the absolute values of G ′ and G″ of the casein gel at frequencies of 5 and 15 MHz were approximately 1000 times higher than the values at the low frequency (0.1 Hz). Furthermore, the measured viscosity at high frequencies was closer to the viscosity of liquid milk. The interpretation given by these workers was that at high frequencies (time scale 10–7–10–8 s) the movements of the gel network are “frozen” and do not contribute to the viscous losses. Lee and coworkers (1992) were, somewhat surprisingly, able to propagate shear waves over reasonable distances (1.1 cm) in bread dough and cheese. The shear wave measurements were used to calculate G ′ and G″ and there was reasonable qualitative agreement between the acoustic and conventional measures of material texture; however, the quantitative relationship was weak. More recently Letang and coworkers (Letang et al., 2001) combined longitudinal velocity and shear wave resonance to study the effect of water content on complex shear and longitudinal moduli of bread dough and showed that the magnitude of the moduli increase with frequency. Love waves are also shear ultrasound, and Varadan and Gardner (1999) noted a concentration and viscosity-dependent change in attenuation of 0.5 dB wt%–1 on inserting a micromachined liquid cell into glycerol solutions. A Love wave sensor is a surface measurement and again if the material close to the sensor is not representative of the bulk will not be reliable. An interesting and novel approach to the ultrasonic characterization of fluid viscosity was recently taken by McCarthy and Choi (personal communication) based on the principle that the velocity profile across a fluid flowing in a pipe is dependent on both the rheological properties of the fluid and the pressure drop across the pipe. Pressure drop is easily measured by nonultrasonic methods and any method capable of measuring a flow profile could be used to calculate the rheological properties of interest.

148

J. N. COUPLAND AND R. SAGGIN

There are two observations that are used to characterize fully developed, steady laminar flow in tomographic-based viscometry. First, independent of the constitutive equation for the material, the conservation of linear momentum demands that the shear stress, σ, depends upon the radial position in the tube, r:

σ (r) = –

∆P r 2L

(35)

where ∆P/L is the pressure drop taken from the downstream direction of the flow over the tube length L. Second, the shear rate, γ˙, is given by:

γ˙(t) =

冷

dυ (r) dr

冷

(36)

and is locally computed using a global curve fit to the tomographic velocity data. Using both Eqns (35) and (36) at each radial position yields the shear stress as a function of shear rate. Repeating this evaluation at all radial positions in a single image of velocity profile, the fluid viscosity can be characterized over a wide range of shear rates. Theoretically the shear rate ranges from zero at the tube center to a maximum at the tube wall. Then the shear viscosity, η(γ˙), is given by:

η(γ˙) =

σ(γ˙) γ˙

(37)

A transducer is placed at an angle to the pipe wall and is used to transmit and receive pulsed ultrasound. The transducer used in this system operates at 5 MHz in a pulse mode. The time from transmission to reception yields the distance the pulse has traveled and the frequency shift of the wave yields the velocity of the particle reflecting the wave. The particles are assumed to move at the speed of the fluid and hence the measurement yields the velocity profile of the fluid. Ultrasonic Doppler velocimetry (UDV) measurements of the fluid velocity profile are in excellent agreement with simultaneous magnetic resonance imaging measurements of the velocity profile. Figure 14 demonstrates the UDV velocity image for tomato sauce and the extracted velocity profile. G. DISPERSED SYSTEMS AND FOOD MICROSTRUCTURE

The ultrasonic properties of dispersed systems depend on the frequency of the sound and the size, shape and number of scattering particles (Figure 4).

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

149

Velocity /cms

(b)

-1

(a)

5 4 3 2 1 0 0

1

2

3

4

5

6

Radial distance /cm FIG. 14 Ultrasonic Doppler velocimetry of the flow pattern of tomato concentrate suspension (4.2° brix) and the calculated velocity profile. Note the decrease in signal intensity across the velocity image due to signal attenuation. Signal attenuation does not adversely affect the estimation of rheological parameters as long as the velocity profile can be extracted. (Figure courtesy of Dr M. McCarthy and Y.C. Choi, University of California, Davis, CA.)

150

J. N. COUPLAND AND R. SAGGIN

In food emulsions the scattering theory has been developed to such an extent that good predictions can be made of the ultrasonic properties of a wide range of emulsion sizes and concentrations. These devices are particularly interesting as they remain the only practical technology for the rapid determination of the particle size of realistic concentrations of emulsions and are commercially available from a number of sources. Some examples of successful particle sizing include: emulsified vegetable oils (Coupland and McClements, 2001b), salad dressings (McClements et al., 1990; Chanamai et al., 2000), milk fat globules (Miles et al., 1990), and parenatal fat emulsions (Kippax et al., 1999). Similar methods have also been used to measure the size of casein micelles (Povey, 1997; Povey et al., 1999). A number of experimental studies have shown that droplet flocculation causes an alteration in the ultrasonic properties of emulsions (McClements, 1994; Hermar et al., 1997; Hibberd et al., 1997). This theory has been applied to the study of droplet aggregation in protein-stabilized emulsions in which flocculation was induced by decreasing the electrostatic repulsion between droplets (Demetriades and McClements, 1999) or by adding a nonabsorbing biopolymer to the continuous phase (Chanamai et al., 1998c; Chanamai and McClements, 2001). These studies have shown that ultrasound is sensitive to the spatial distribution of the droplets within an emulsion. The same ultrasonic spectroscopy technique has been used to study the disruption of flocs in a shear field (Chanamai et al., 1998c). As the emulsions are exposed to higher shear rates the flocs become disrupted and their attenuation spectra becomes closer to that of nonflocculated droplets. An interesting modification of the scattering theories introduced above allows the calculation of the emulsion ζ-potential. Briefly, the acoustic wave causes a charged particle to oscillate within its associated ionic environment. At certain frequencies the ions cannot realign rapidly enough to keep pace with the moving particle and the ionic friction generates a voltage and causes additional ultrasonic losses, i.e. the electroacoustic effect (Dukhin et al., 2000). Instrumentation can be based on this principle by either measuring the current generated from the particles in an applied ultrasonic wave or the ultrasound generated by the particles in an oscillating electrical field. The latter method is believed to be the most practical as the ultrasound generated by the particles is directly proportional to their electrophoretic mobility. Electroacoustic techniques have been applied to the determination of ζ-potential in concentrated suspensions including oil-in-water emulsions (Kong et al., 2001), casein micelles (Wade et al., 1996) and dairy emulsions (Wade and Beattie, 1997). Bubbles and concentrated foams are important structures common to many foods. In many cases the foam is an integral and valued feature (e.g.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

151

bread, ice cream, beer) but in other cases excessive foaming is a problem during processing (e.g. in bioreactors). Foams are transient and fragile and thus very difficult to characterize. Gravimetric analysis can usually provide a good estimate of volume fraction, but size analysis usually requires optical microscopy or other imaging techniques. Ultrasound is extremely sensitive to both the volume fraction and size of air bubbles and is a promising analytical approach. However, the extremely high attenuation means it is practically impossible to transmit high-frequency sound through an appropriate thickness of foam. Reflectance is a promising alternative and Kulmyrazaev et al. (2000) measured the reflectance spectra (1–7 MHz) from a series of model food foams and noted clear spectral differences between the foams (different sizes and concentrations), but only qualitative agreement with theoretical calculations. Foreign bodies, such as glass, steel, plastic, wood or stones, can be hard to detect in the finished product and may injure consumers. However, their significant acoustic mismatch with most other food materials makes them relatively easy to detect ultrasonically. Haeggstrom and Luukkala (2001) used ultrasonic signal analysis to detect small pieces of each of these materials at a depth of 20–75 mm in homogeneous materials (margarine, smooth cheese) but to lesser depths (50 mm) in coarser materials (marmalade). In some samples these workers were able to distinguish between the materials based on the ultrasonic echo pattern. H. POLYMERIC SYSTEMS

The adiabatic compressibility of a solution can be calculated from measurements of its ultrasonic velocity and density (Eqn (7)). The compressibility of a solute is usually taken as the molecular increment of solution compressibility measurements made at very low concentrations. The extremely dilute (~1 mg ml–1) solutions used in these measurements are extremely advantageous for high value or hard to obtain materials but also require extremely sensitive measurement apparatus. Precise density measurements can be made with a vibrating U-tube densitometer and ultrasonic velocity measurements with a precision resonator. Apenten and coworkers (2000) recently showed that the compressibility of proteins can be measured at relatively high concentrations (>10 mg ml–1) if thermal scattering effects are accounted for. High concentrations means that less-precise pulse echo methodologies may be used. Compressibility measurements are a useful way to probe polymer, particularly protein, hydration. The measured value is taken to be the sum of contributions from the hydration layer and the intrinsic compressibility of the polymer. Compressibility measurements have been used to reveal

152

J. N. COUPLAND AND R. SAGGIN

protein hydrophobicity, polar/nonpolar residues, and aspects of secondary structure (Gekko and Hasegawa, 1986) as well as changes associated with protein denaturation (Kamiyama and Gekko, 1997) and ligand binding (Gekko and Yamagami, 1998). Protein–solvent interactions are believed to be important determinants of functionality, and compressibility is unsurprisingly correlated to protease susceptibility, foaming capacity, and free energy of unfolding of a number of proteins (Gekko and Yamagami, 1991). Gelation and melting of polymer systems has proved a more difficult process to probe ultrasonically. This may be because longitudinal ultrasound is sensitive to the bulk modulus and gelation usually involves changes to the shear modulus. However, in some cases attenuation has proved sensitive to polymer aggregation phenomena, perhaps due to the scattering by aggregates. Gunasekaran and Ay (1994, 1996) used a through-transmission pulsed method to measure the ultrasonic properties (velocity and attenuation) of milk after renneting. They showed that while the changes in ultrasonic velocity were small, the attenuation decreased, perhaps due to the same thermal overlap effects thought to be responsible for the attenuation decrease when an emulsion flocculates. Similar studies on the isoelectric precipitation of other proteins have also revealed a change in attenuation as the proteins aggregated (Pavlovskaya et al., 1992; Bryant and McClements, 1999b). Buckin and Smyth (1999) were concerned with the tendency of calcium-fortified milk to aggregate and gel or precipitate on heating and used a high-precision resonator method to detect the onset of aggregation of their samples. Audebrand and coworkers (1995) demonstrated an increase in attenuation (greater at higher frequencies) as alginate gels. Shore and coworkers (Shore et al., 1986; Shore and Miles, 1988b, 1988c) saw a peak in some muscle preparations (but not in liver or kidney) at pH 5 corresponding to a maximum shortening and density and minimum waterholding capacity of the meat. By suspending the myofibrils in solutions of different density and viscosity they were able to show that viscous scattering was not a significant contributor (~20%) to the total measured attenuation losses (Shore and Miles, 1988a), but they were not able to identify the dominant mechanism. Bryant and McClements used scattering theory to interpret the attenuation spectra of aggregating whey protein in terms of the sizes of structures present (Bryant and McClements, 1999a, 1999b). Various investigators have seen an ultrasonic attenuation maxima at pH ~ 11 for solutions of amino acids, polypeptides (Appelgate et al., 1968; Saravazyan and Kharakoz, 1979) and proteins (Kanda et al., 1972; O’Brien and Dunn, 1972) owing to the excess absorption caused by the resonance of the protonation equilibrium at amine groups. Miles and others observed a similar attenuation peak at pH~11 in homogenates (Shore and Miles,

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

153

1988c) and suspensions (Shore and Miles, 1988b) of bovine skeletal muscle as well as other beef tissues (liver and kidney). By repeated fractionation of the protein components, they ascertained that this was largely due to a molecular relaxation process rather than to any scattering phenomena. Bryant and McClements (1999b) measured a similar pH-dependent peak in solutions of whey protein isolate and used ultrasonic spectroscopy to estimate the relaxation time of the process (~10–8 s) I. MISCELLANEOUS APPLICATIONS

Varadan and Gardner (1999) reported on the use of a Love wave-based smart tongue and nose. Love wave resonance (108.7 MHz) depended on the properties of the materials in contact with the sensor (air, water, orange juice, deicer). Using similar sensors and interpreting the data using principle component analysis, Varadan (personal communication, 2000) was able to distinguish fruit juices (prune, grapefruit, grape, apple, orange) by type. Vegetable oils deteriorate during frying, breaking down to form offflavor volatiles and polymerizing to produce a dark brown color and an increase in viscosity. Lacey and Payne (1994) showed that the ultrasonic properties of oil also change with aging, but the correlations are probably too weak to detect spoilage before it becomes obvious. Withers (1996) considered methods to evaluate pipe fouling and showed that ultrasound can quantify films of thickness between 0.5 and 6 mm (the lower limit being set by the wavelength of the ultrasound). A disadvantage of this method is that it is a single-point inspection and may not be representative of an entire piece of equipment. In non-food applications, Rose (1999) has pioneered the use of guided waves for fouling detection and has shown that a single transducer can inspect several meters of pipe (or kilometers of railroad track). J. IMAGING

By making several measurements at different positions throughout the food it is possible to construct an image either of the fundamental ultrasonic properties or, by using several of the methods set out above, code the image to composition, temperature, etc. As with all ultrasonic measurements, it is important to remember that the measurements are a function of the composition, microstructure and physicochemical properties of the sample, and converting a time-of-flight into another variable for imaging purposes is not trivial. Pinfield and coworkers (1994, 1996) provide a good example of this problem, demonstrating that the microstructure of

154

J. N. COUPLAND AND R. SAGGIN

a creaming emulsion as well as the local oil concentration can have an effect on the measured velocity profile. A single ultrasonic measurement is analogous to drawing a line (of finite width) through the body under consideration. The detected signal will contain information, in terms of echo times and amplitude, about the structures in its path. By taking many measurements at different positions on (or angles through) the body, it is possible to provide information about all the detectable internal structures. The process of reconstructing an image of these structures from a set of one-dimensional signals is known as tomography, and while the theory is well understood (Hoyle, 1998) and commonly exploited in medical (Shung, 1990) and materials imaging, it has not been adopted in its full form in foods. It is worth remembering that imaging is a way of generating two-dimensional data from a much richer data set and is therefore a data reduction operation. The full set of measurements is an A-scan, the image generated from selecting the amplitude at a set time is a B-scan, and that generated from the amplitude of a selected signal feature is a C-scan. It is important to bear in mind that both B-scans and C-scans are data reduction tools and there is often some valuable information lost in generating an image (Coupland and McClements, 2001a). Perhaps the most sophisticated imaging regularly used in foods is in the field of carcass grading. Images can be generated based on a number of ultrasonic principles and used to estimate the texture and composition of carcasses, as well as quality attributes such as marbling and defects (Griffin et al., 1999; Brethour, 2000). An interesting extension to conventional imaging is the use of elastography, where an image coded to elastic modulus is generated by measuring the movement of an ultrasonic signal reflected from the internal structures in a carcass on gentle bulk compression (Ophir et al., 1994). Ahvenainen et al. (1989) and Wirtanen et al. (1992) used ultrasonic imaging to measure the spoilage of milk in cartons (presumably through the formation of gas bubbles by the active microorganisms present). Acton et al. (1986) used a medical imaging system to generate some images of gels. Other workers have simplified the geometry of their system to allow only one-dimensional changes and recorded images of emulsion creaming (Howe et al., 1986; Gunning et al., 1989; Dickinson et al., 1994; Basaran et al., 1998b) and the diffusion of small molecules in gels (Basaran et al., 1998a) and emulsions (Basaran and McClements, 1999). Ozguler et al. (1998) measured reflected (backscattered) ultrasonic energy (17.4 MHz) and used the signal to detect leaks (air- or water-filled channels and food inclusions) in laminated foil or plastic retortable pouches. The defects that could be detected were in the order of tens of micrometers. Frazier et al. (2000) used a modification of this approach to increase the resolution of the structures imaged.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

155

An important limitation with acoustic imaging is that the sample under investigation must be placed in a tank of water to allow good acoustic propagation from the transducer to and through the food. The measurements are also quite slow and limited by the speed of the robot arm used to position the transducer. (Simultaneous measurements with arrays of transducers can greatly increase the speed of imaging.) These limitations are acceptable for certain high-value items, for example in the aerospace and materials science industries, but it is clearly unacceptable for the routine inspection of foods. However, the use of noncontact, air-coupled transducers offers a way to extend imaging to practical food characterization. Saggin and Coupland (2001c) have shown that noncontact ultrasonic velocity measurements can provide a good description of the size of various food items. For the various soft solids used in this study the ultrasonic measurement of thickness was superior to measurements taken with calipers. A particularly important aspect of the noncontact methodology to size measurement is that the ultrasonic properties of the material are measured simultaneously and do not need to be known a priori. A more sophisticated image based on noncontact ultrasonic velocimetry is reported in Figure 15. Noncontact imaging is potentially competitive with NMR imaging currently used in medical diagnostics and, just as NMR imaging has proved invaluable in various types of food characterization, we expect that it will also see wide applications in the food industry.

FIG. 15 Image of a coin generated from the sound reflected from the surface in aircoupled (noncontact) mode. The image was produced using an NCA-1000 (SecondWave Systems, Boalsburg) and an NCT230-R6 (3 MHz, 6.4 mm diameter and 6.4 mm focal length) transducer. (Image courtesy of Dr M. Bhardwaj.)

156

J. N. COUPLAND AND R. SAGGIN

Scanning acoustic microscopy (SAM) is a useful nondestructive evaluation (NDE) method often applied to the measurement of micro-cracks or other defects in ceramics, composite materials and electronic materials. An important advantage of ultrasonic microscopy over other imaging techniques is in the detection of defects inside opaque materials without requiring sectioning or other destructive operation. Conventionally SAM works in contact mode, detecting reflections of high frequency (tone bursts at 5–230 MHz) sound from the microstructures of interest. The reflected signal is converted to an image dependent on the thickness of the layers and their impedance mismatch. Because of the low impedance of air, delaminations, bubbles and other voids are most easily imaged by this method (Adams, 2000). Aside from the differences in transducer size and frequency, ultrasonic microscopy is similar to other imaging methods and suffers from similar limitations. A couplant material (often water), located between the acoustic lens and the specimen, is usually necessary to allow the propagation of ultrasound into the sample; therefore, the samples are generally immersed in a water tank and repositioned using a stepper motor. Ultrasonic microscopy has been used to detect sealworms (a parasite) in fish fillets (Hafsteinsson and Rivzi, 1984), but other workers have generated SAM images of biological systems, particularly for medical applications (Kent and Lee, 1997; Kinoshita et al., 1998; Karl and Bereiter-Hahn, 1999). The resolution of ultrasonic imaging is limited because of the long wavelength of low-frequency ultrasound and the large size of the transducers used. It is possible to modify both of these in a technique known as acoustic microscopy to achieve resolutions approaching an optical microscope. An estimate of the minimum resolvable distance (∆r) is as follows: ∆r = Fλ = F

cc v

(38)

where F is a constant related to the ultrasonic lens geometry and cc is the velocity of sound in the coupling medium. Therefore increasing the frequency of sound would decrease the resolution of the instrument (for example if 5 MHz sound could resolve points 0.1 mm apart, than 230 MHz sound could resolve points 2 µm apart). An intriguing alternative method to improve resolution is to use a coupling material with a lower ultrasonic velocity and hence shorter wavelength (for example, if a system could resolve points 0.1 mm apart through a water couplant, then it could distinguish points 0.02 mm apart in an air-coupled system). The aircoupled transducers discussed previously are therefore interesting candidates for high-resolution ultrasonic imaging (Bhardwaj et al., 2000a,

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

157

2000b), particularly as the alternatives (liquid nitrogen or helium) are particularly disruptive of delicate structures (Briggs, 1992). It is also important to remember that attenuation increases with frequency so the penetration depth available at high resolution may be limited. Some recent advances on this technique make use of shear wave lens and noncontact mode, as proposed by Miyasaka and Tittmann (2000). Shear waves offer some great potential, such as a better resolution than longitudinal waves, i.e. they have a shorter wavelength, less sensitivity to surface roughness and lower spherical aberration caused by reflection. V. CONCLUSIONS

Ultrasound becomes attractive as a sensor in the food industry when it has a unique capability to make a measurement or can otherwise outperform other technologies. Many of the strongest applications gain their advantage from the capacity of ultrasound to propagate through many optically opaque materials, particularly container and piping walls, to make measurements on-line. The significant disadvantages of ultrasound often center around the fact that propagation depends on a complex set of variables – compositional, structural and dynamic – within the sample and it becomes hard to deconvolute the parameter of interest. Practically, the most significant complicating factors tend to be fluctuations in temperature and/or the presence of air in the sample. The ultrasonic properties of food are strongly temperature dependent and even a 1°C deviation from the calibration conditions can obscure other important changes. The large acoustic mismatch between air and food means that even a tiny amount of entrained gas will dominate the signal or attenuate it to a such a degree as to make measurement impossible. Furthermore, the difficulty of transmitting ultrasound through air limits its application as an on-line sensor for discrete food items. Another disadvantage is that ultrasonic measurements are often poor multicomponent sensors. Acoustic spectra can be used to identify multiple effects simultaneously but the various relaxations responsible for the changing signal are broad, complex and highly overlapping, so can be difficult to exploit in practice. Velocity and attenuation respond to different properties of the system so are sometimes useful in combination. As a rule of thumb: velocity tends to be more sensitive to composition and temperature while attenuation is more sensitive to structure. Weighing the advantages and disadvantages, we can see why some ultrasonic applications have seen such a wide application. For example,

158

J. N. COUPLAND AND R. SAGGIN

the use of ultrasonic attenuation spectra for emulsion characterization allows measurement of a parameter not otherwise accessible (size of concentrated dispersions) under conditions where the other variables (importantly temperature) can be controlled. In cases when the conditions are less well satisfied, the acceptance of ultrasonics is also less, e.g. ultrasonic velocity can be easily used to measure the concentration of sugar syrup in a pipe, but on-line refractive index sensors offer similar functionality and compete for this market. In cases where ultrasound substantially fails to meet the conditions set out above, e.g. multicomponent compositional analysis based upon a temperature scan, we expect limited application. The successful technologies will be improved and refined through applications by food scientists, but it is also interesting to speculate how this set may be extended by some of the recent innovations discussed in this chapter. Noncontact transducers are particularly interesting as they allow practical application of ultrasonic methods in cases where an air gap between the food and sensor is unavoidable. The use of printed circuit boards to make Love wave sensors that can be wirelessly interrogated has the potential to completely change the economics of sensing. If a fully functional sensor can be made for a few cents, one could be incorporated into each package of food. The use of selective masks for the sensor could allow the detection of spoilage compounds or the metabolic products of microbial growth. Building a reader into domestic refrigerators or cooking equipment could “ask” each piece if it were fresh, safe, or adequately cooked. No single sensor will meet all the needs of the food industry but, through a critical understanding of both the strengths and weaknesses of ultrasonics, we expect the number of practical applications to grow.

ACKNOWLEDGEMENTS

We are grateful to Dr Mike McCarthy and Y. J. Choi (UC Davies), Dr Mahesh Bhardwaj (Ultran Laboratories/SecondWave Systems, Boalsburg, PA), and Julian McClements (University of Massachusetts) for contributing unpublished data, figures and valuable discussion to this paper, and to Melissa Goff for technical assistance in preparing the manuscript. This work was partly supported by a grant from the Center for Food Manufacturing (Penn State).

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

159

REFERENCES Abouelkaram, S., Suchorski, K., Buquet, B., Berge, P., Culioli, J., Delachartre, P. and Basset, O. 2000. Effects of muscle texture on ultrasonic measurements. Food Chem. 6(4), 447–455. Acton, J. C., Clay, D. L., Robinson, K. E., Dick, R. L. and Acton, W. C. 1986. Structural characteristics of protein gels as determined by an ultrasound imaging technique. J. Food Sci. 51(2), 524–525. Adams, C. 2000. Very high resolution reflection-mode acoustic microscopy. The American Microscopy and Analysis July, 11–12. Affeldt, H. and Abbott, J. 1989. Apple firmness and sensory quality using contact acoustic transmission. In “Proceedings of the 7th International Congress on Agricultural Engineering,” Dublin, 4–8 September 1989 (V. A Dodd and P. M. Grace, eds), vol. 3, pp. 2037–2045. Brookfield, Rotterdam. Ahuja, A. S. and Hendee, W. R. 1978. Effects of particle shape and orientation on propagation of sound in suspensions. J. Acoust. Soc. Am. 63, 1074–1080. Ahvenainen, R., Wirtanen, G. and Manninen, M. 1989. Ultrasound imaging a non-destructive method for monitoring the microbiological quality of aseptically packed milk products. Lebens. Wiss. Technol. 22, 382–386. Apenten, R. K. O., Buttner, B., Mignot, B., Pascal, D. and Povey, M. J. W. 2000. Food Hydrocolloids 14, 83–91. Appelgate, K., Slutsky, L. J. and Parker, R. C. 1968. Kinetics of proton-transfer reactions of amino acids and simple polypeptides. J. Am. Chem. Soc. 90(25), 6909–6913. Audebrand, M., Doublier, L. J., Durand, D. and Emery, J. R. 1995. Investigation of gelation phenomena of some polysaccharides by ultrasonic spectroscopy. Food Hydrocolloids 9, 195–203. Babick, F., Hinze, F. and Ripperger, S. 2000. Dependence of ulrasonic attenuation on the material properties. Colloids Surf. A172(1–3), 33–46. Basaran, T. K. and McClements, D. J. 1999. Nondestructive monitoring of sucrose diffusion in oil-in-water emulsions by ultrasonic velocity profiling. J. Colloids Interface Sci. 220(2), 429–435. Basaran, T. K., Coupland, J. N. and McClements, D. J. 1998a. Monitoring molecular diffusion of sucrose in xanthan solutions using ultrasonic velocity measurements. J. Food Sci. 64(1), 125–128. Basaran, T. K., Demetriades, K. and McClements, D. J. 1998b. Ultrasonic imaging of gravitational separation in emulsions. Colloids Surf. A 136(1–2), 169–181. Benedito, J., Carcel, J., Clemente, G. and Mulet, A. 2000a. Cheese maturity assessment using ultrasonics, J. Dairy Sci. 83(2), 248–254. Benedito, J., Carcel, J. A., Sanjuan, N. and Mulet, A. 2000b. Use of ultrasound to assess Cheddar cheese characteristics. Ultrasonics 38(1–8), 727–730. Benedito, J., Gonzalez, R., Rossello, C. and Mulet, A. 2000c. Instrumental and experrt assessment off Mahon cheese texture. J. Food Sci. 65(7), 1170–1174. Benedito, J., Carcel, J. A., Rossello, C. and Mulet, A. 2001. Composition assessment of raw meat mixtures using ultrasonics. Meat Sci. 57(4), 365–370. Berchiesi, G., Amico, A., Vitali, G., Amici, L. and Litargini, P. 1987. Ultrasonic Investigation in aqueous-solutions of sucrose. J. Mol. Liq. 33(2–3), 157–181. Bhardwaj, M. 1986. Importance and use of very high-frequency ultrasound in nondestructive characterization of materials. Am. Ceram. Soc. Bull. 65(11), 1461–1461. Bhardwaj, M., Neeson, I., Langron, M. and Vandervalk, L. 2000a. Contact-free ultrasound: the final frontier in non-destructive materials characterization. In “24th Conference: An International Conference on Engineering Ceramics and Structures.”

160

J. N. COUPLAND AND R. SAGGIN

Bhardwaj, M., Neeson, I. and Stead, G. 2000b. Introduction to contact-free ultrasonic characterization and analysis of consolidated materials. J. Nondestructive Testing Ultrasonics 5(6). Birks, A. S. and Green, R. E. 1991. “Ultrasonic Testing.” American Society for Nondestructive Testing, Columbus, OH. Blitz, J. 1963. “Fundamentals of Ultrasonics.” Butterworths, London. Brethour, J. R. 2000. Using serial ultrasound measures to generate models of marbling and backfat thickness changes in feedlot cattle. J. Anim. Sci. 78(8), 2055–2061. Briggs, G.A.D. 1992. Acoustic microscopy – A summary. Rep. Progr. Phys. 55(7), 851–909. Bryant, C. M. and McClements, D. J. 1999. Ultrasonic spectrometry study of the influence of temperature on whey protein aggregation. Food Hydrocolloids 13(6), 439–444. Bryant, C. M. and McClements, D. J. 1999. Ultrasonic spectroscopy study of relaxation and scattering in whey protein solutions. J. Sci. Food Agric. 79(12), 1754–1760. Buckin, V.A. 1988. Hydration of nucleic basis in dilute equeous solutions. Apparent molar adiabatic and isothermal compressibilities, apparent molar volumes and their temperature slopes at 25°C. Biophys. Chem. 29(3), 283–292. Buckin, V.A. and Kudryashov, E. 2001. Ultrasonic shear wave rheology of weak particle gels. Adv. Colloid Interf. Sci. 89, 401–422. Buckin, V.A. and Smyth, C. 1999. High-resolution ultrasonic resonator measurements for analysis of liquids. Semin. Food Anal. 4(2), 95–112. Cebula, D. J., McClements, D. J. and Povey, M. J. W. 1990. Small angle neutron scattering from voids in crystalline trilaurin. J. Am. Oil Chem. Soc. 67, 76–78. Chanamai, R. and McClements, D. J. 1999. Ultrasonic determination of chicken composition. J. Agric. Food Chem. 47(11), 4686–4692. Chanamai, R. and McClements, D. J. 2001. Depletion flocculation of beverage emulsions by gum arabic and modified starch. J. Food Sci. 66(3), 457–463. Chanamai, R., Coupland, J. N. and McClements, D. J. 1998a. Effect of temperature on the ultrasonic properties of oil-in-water emulsions. Colloids Surf. A 139, 241–290. Chanamai, R., Hermmann, N. and McClements, D. J. 1998b. The influence of flocculation on the ultrasonic properties of emulsions: experiment. J. Appl. Phys. D 31, 2956–2963. Chanamai, R., Hermmann, N. and McClements, D. J. 1998c. Ultrasonic spectroscopy study of flocculation and shear-induced floc disruption in oil-in-water emulsions. J. Colloid Interface Sci. 204(4), 268–276. Chanamai, R., Herrmann, N. and McClements, D. J. 1999. Influence of thermal overlap effects on the ultrasonic attenuation spectra of polydisperse oil-in-water emulsions. Langmuir 15(10), 3418–3423. Chanamai, R., Alba, F. and McClements, D. J. 2000. Ultrasonic spectroscopy study of salad dressings. J. Food Sci. 65(3), 507–513. Contreras, N. I., Fairley, P., McClements, D. J. and Povey, M. J. W. 1992. Analysis of the sugar content of fruit juices and drinks using ultrasonic velocity measurements. Int. J. Food Sci. Technol. 27, 515–529. Coupland, J. N. 2001. Ultrasonic characterization of lipid crystallization. In “Crystallization and Solidification of Properties of Lipids” (N. Widlak, R. Hartel and S. Narine, eds). American Oil Chemists Society Press, Champaign, IL. Coupland, J. N. and McClements, D. J. 1997. Physical properties of liquid edible oils. J. Am. Oil Chem. Soc. 74(12), 1559–1564. Coupland, J. N. and McClements, D. J. 2001a. Ultrasonics. In “Nondestructive Food Evaluation: Techniques to Analyze Properties and Quality” (S. Gunasekaran, ed.), pp. 233–242, Marcel Dekker, New York. Coupland, J. N. and McClements, D. J. 2001b. Droplet size determination in food emulsions: Comparison of ultrasound and light scattering methods. J. Food Eng. 50(2), 117–120.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

161

Crecraft, D. I. 1983. Ultrasonic instrumentation: principles, methods, and applications. J. Phys. E.: Sci. Instrum. 16, 181–189. Curtin, S. D., Jakoby, B., Berthold, A., Varadan, V. K. and Varadan, V. V. 1998. A micromachined wet cell for a Love-wave liquid sensor. In “Proceedings of the SPIE Conference on Smart Electronics and MEMS,” San Diego, 2–4 March 1998 (V. K. Varadan, P. J. McWhorter, R. A. Singer and M. J. Vellekoop, eds), vol. 3328, pp. 194–200. SPIE, Bellingham. DelGrosso, V. A. and Mader, C. W. 1972. Speed of sound in pure water. J. Acoust. Soc. Am. 52(5), 1442–1445. Demetriades, K. and McClements, D. J. 1999. Ultrasonic attenuation spectroscopy study of flocculation in protein stabilized emulsions. Colloids Surf. A 150(1–3), 45–54. Dickinson, E. and McClements, D. J. 1995. In “Advances in Food Colloids.” Blackie Academic & Professional, Glasgow. Dickinson, E., Ma, J. and Povey, M. J. W. 1994. Creaming of concentrated oil in water emulsions containing xanthan. Food Hydrocolloids 8, 481–497. Dukhin, A. S., Goetz, P. J., Wines, T. H. and Somasundaran, P. 2000. Acoustic and electroacoustic spectroscopy. Colloids Surf. A 173(1–3), 127–158. Faeth, L. and Chem, D. 1999. On-line, noncontact baking monitor for bread. Cereal Food World 44(3), 155–158. Fedotkin, I., Sukauskas, V., Chepurnoi, M., Shnaider, V. and Klimenko, M. 1981. Food Sci Technol. Abstr. 13, 459. Feil, M. and Zacharias, E. 1971. The determination of yeast slurry consistency and wort plato by sonic solution analysis. The Brewers Digest 29, 76–80. Fisher, A. V. 1997. A review of the technique of estimating the composition of livestock using the velocity of ultrasound. Comput. Electron. Agric. 17(2), 217–231. Flitsanov, U., Mizrach, A., Liberzon, A., Akerman, M. and Zauberman, G. 2000. Measurement of avocado softening at various temperatures using ulrasound. Postharvest Biol. Technol. 20(3), 279–286. Frazier, C. H., Tian, Q., Ozguler, A., Morris, S. A. and O’Brien, W. D. 2000. High contrast ultrasound images of defects in food package seals. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 47(3), 530–539. Garbolino, C., Ziegler, G. R. and Coupland, J. N. 2000. Ultrasonic determination of the effect of shear on lipid crystallization J. Am. Oil Chem. Soc. 77, 157–162. Garett, R. E. and Furry, R. B. 1972. Velocity of sonic pulses in apples. Trans. Am. Soc. Agric. Engng 15, 770–774. Gekko, K. and Hasegawa, Y. 1986. Comprressibility structure relationship of globular-proteins. Biochemistry 25(21), 6563–6571. Gekko, K. and Yamagami, K. 1991. Flexibility of food proteins as revealed by compressibility. J. Agric. Food Chem. 39, 57–62. Gekko, K. and Yamagami, K. 1998. Compressibility and volume changes of lysozyme due to inhibitor binding. Chem. Lett. 8, 839–840. Ghaedian, R., Decker, E. A. and McClements, D. J. 1997. Use of ultrasound to determine cod fillet composition. J. Food Sci. 62(3), 500–504. Ghaedian, R., Coupland, J. N., Decker, E. A. and McClements, D. J. 1998. Ultrasonic determination of fish composition. J. Food Eng. 35(3), 323–337. Goss, S., Johnston, R. and Dunn, F. 1978. Comprehensive compilation of empirical ultrasonic properties of mammalian tissues. J Acoust. Soc. Am. 64(2), 423–457. Griffin, D. B., Savell, J. W., Recio, H. A., Garrett, R. P. and Cross, H. R. 1999. Predicting carcass composition of beef cattle using ultrasound technology. J. Anim. Sci. 77(4), 889–892. Griffin, S. J., Hull, J. B. and Lai, E. 2001. Development of a novel ultrasound monitoring system for container filling operations. J. Mater. Process. Technol. 109(1–2), 72–77.

162

J. N. COUPLAND AND R. SAGGIN

Gunasekaran, S. and Ay, C. 1994. Evaluating milk coagulation with ultrasonics. Food Technol. 12, 74–78. Gunasekaran, S. and Ay, C. 1996. Milk coagulation cut-time determination using ultrasonics. J. Food Process. Eng. 19, 63–73. Gunning, P. A., Hibberd, D. J., Howe, A. M. and Robbins, M. M. 1989. Use of velocity of ulrasound to monitor gravitational separation in dispersions, J. Soc. Dairy Technol. 42(3), 70–77. Haeggstrom, E. and Luukkala, M. 2000. Ultrasonic monitoring of beef temperature during roasting. Lebens. Wiss. Technol. 33(7), 465–470. Haeggstrom, E. and Luukkala, M. 2001. Ultrasound detection and identification of foreign bodies in food products. Food Control 12(1), 37–45. Hafsteinsson, H. and Rivzi, S. S. H. 1984. Acoustic microscopy – principles and applications in the study of biomaterial microstructure. Scan. Electron Microsc. III, 1237–1247. Hayes, C. F. and Chingon, H. T. G. 1982. Acoustic properties of Papaya. J. Texture Stud. 13, 397–402. Hermar, Y., Hermann, N., Lemarechal, P., Hocquart, R. and Lequeux, F. 1997. Effective medium model for ultrasonic attenuation due to the thermo-elastic effect in concentrated emulsions. J. Physique II 7, 637–647. Herrmann, N. and McClements, D. J. 1999. Ultrasonic propagation in highly concentrated oilin-water emulsions. Langmuir 15(23), 7937–7939. Hibberd, D., Holmes, A., Garrood, M., Fillery-Travis, A., Robbins, M. and Challis, R. 1997. Ultrasonic monitoring of oil-in-water emulsions undergoing depletion flocculation. J. Colloid Interface Sci. 193(1), 77–87. Hindle, S., Povey, M. J. W. and Smith, K. 2000. Kinetics of crystallization in n-hexadecane and cocoa butter oil-in-water emulsions accounting for droplet collision-mediated nucleation. J. Colloid Interface Sci. 232(2), 370–380. Hodate, Y., Ueno, S., Yano, J., Katsuragi, T., Tezuka, Y., Tagawa, T., Yoshimoto, N. and Sato, K. 1997. Ultrasonic velocity measurement of crystallization rates of palm oil in oil water emulsions. Colloids Surf. A 128(1–3), 217–224. Howe, A. M., Kackie, A. R. and Robbins, M. 1986. Technique to measure emulsion creaming by the velocity of ultrasound. J. Dispersion Sci. Technol. 7, 231–243. Hoyle, B. S. 1998. In “Ultrasound in Food Processing” (M. J. W. Povey and T. J. Mason, eds), pp. 183–192. Blackie Academic & Professional, London. Javanaud, C. 1988. Applications of ultrasound to food systems. Ultrasonics 26(3), 117–123. Kamiyama, T. and Gekko, K. 1997. Compressibility and volume changes of lysozyme due to guanidine hydrochloride denaturation. Chem. Lett. 10, 1063–1064. Kanda, H., Ookubo, N., Nakajima, H., Suzuki, Y., Minato, M., Ihara, T. and Wada, Y. 1972. Ultrasonic absorbtion in aqueous solutions of lysozyme. Biopolymers 15, 783–795. Karl, I. and Bereiter-Hahn, J. 1999. Tension modulates cell surface motility: A scanning acoustic microscopy study. Cell Motil. Cytoskeleton 43(4), 349–359. Kaulgud, M. V. and Dhondge, S. S. 1988. Apparent molal volumes and apparent molal compressibilities of some carbohydrates in dilute aqueous-solutions at different temperatures. Ind. J. Chem. A 27(1), 6–11. Kent, S. D. and Lee, H. 1997. Scanning tomographic acoustic microscopy: Development and applications. Int. J. Imag. Syst. Technol. 8(3), 255–262. Khimunin, A. S. 1972. Numerical calculation of the diffraction corrections for the precise measurement of ultrasound absorption. Acustica 27, 173–181. Kinoshita, A., Senda, S., Mizushige, K., Masugata, H., Sakamoto, S., Kiyomoto, H. and Matsuo, H. 1998. Evaluation of acoustic properties of the live human smooth-muscle cell using scanning acoustic microscopy. Ultrasound Med. Biol. 24(9), 1397–1405. Kippax, P., Sherwood, J. D. and McClements, D. J. 1999. Ultrasonic spectroscopy study of

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

163

globule aggregation in parenteral fat emulsions containing calcium chloride. Langmuir 15(5), 1673–1678. Kloek, W., Walstra, P. and Van Vliet, T. 2000. Nucleation kinetics of emulsified triglyceride mixtures. J. Am. Oil Chem. Soc. 77(6), 643–652. Kong, L., Beattie, J. K. and Hunter, R. J. 2001. Electroacoustic study of concentrated oil-inwater emulsions. J. Colloid Interface Sci. 238(1), 70–79. Kress-Rogers, E. (ed.) 1993 “Instrumentation and Sensors for the Food Industry.” ButterworthHeinemann, Oxford. Kudryashov, E. D., Hunt, N. T., Arikainen, E. O. and Bickin, V. A. 2001. Monitoring of acidified milk gel formation by ultrasonic shear wave measurements. High-frequency viscoelastic moduli of milk and acidified milk gel. J. Dairy Sci. 84(2), 375–388. Kulmyrzaev, A. and McClements, D. J. 2000. High frequency dynamic shear rheology of honey. J. Food Eng. 45(4), 219–224. Kulmyrazaev, A., Cancelliere, C. and McClements, D. J. 2000. Characterization of aerated foods using ultrasonic reflectance spectroscopy. J. Food Eng. 46, 235–241. Lacey, R. E. and Payne, F. A. 1994. Ultrasonic velocity in used corn oil as a measure of oil quality. Trans. Am. Soc. Agric. Engng 37, 1583–1589. Lee, H. O., Luan, H. and Daut, D. G. 1992. Use of an ultrasonic technique to evaluate the rheological properties of cheese and dough. J. Food Eng. 16, 127–150. Letang, C., Piau, M., Verdier, C. and Lefebvre, L. 2001. Characterization of wheat-flour-water doughs: a new method using ultrasound. Ultrasonics 39(2), 133–141. Liljedahl, L. A. and Abbott, J. A. 1994. Changes in sonic resonance of delicious and golden delicious apples undergoing accelerated ripening. Trans. Am. Soc. Agric. Engng 37, 907–912. Lynnworth, L.C. 1989. “Ultrasonic Measurements for Process Control.” Academic Press, San Diego. McClements, D. J. 1988. The use of ultrasonics for characterizing fats and emulsions. PhD thesis, Procter Department of Food Science, Leeds University, Leeds. McClements, D. J. 1992. Comparison of multiple scattering theories with experimental measurements in emulsions. J. Acoust. Soc. Am. 91, 849–853. McClements, D. J. 1994. Ultrasonic determination of depletion flocculation in oil-in-water emulsions containing a non-ionic surfactant. Colloids Surf. A 90, 25–35. McClements, D. J. 1996. Principles of ultrasonic droplet size determination in emulsions. Langmuir 12(14), 3454–3461. McClements, D. J. 1997. Ultrasonic characterization of foods and drinks: Principles, methods, and applications. Crit. Rev. Food Sci. Nutr. 37, 1–46. McClements, D. J. and Fairley, P. 1991. Ultrasonic pulse echo reflectometer. Ultrasonics 29, 58–62. McClements, D. J. and Fairley, P. 1992. Frequency scanning ultrasonic pulse echo reflectometer. Ultrasonics 30(6), 403–405. McClements, D. J. and Povey, M. J. W. 1987. Solid fat content determination using ulrasonic velocity measurements. Int. J. Food Sci. Technol. 22, 491–499. McClements, D. J. and Povey, M. J. W. 1988. Ultrasonic velocity measurements in some liquid triglycerides and vegetable oils. J. Am. Oil Chem. Soc. 65(11), 1791–1796. McClements, D. J. and Povey, M. J. W. 1989. Scattering of ultrasound by emulsions. J. Phys. D: Appl. Phys 22, 38–47. McClements, D. J., Povey, M. J. W., Jury, M. and Betsanis, E. 1990. Ultrasonic characterization of a food emulsion. Ultrasonics 28, 266–272. McClements, D. J., Povey, M. J. W. and Dickinson, E. 1993. Absorption and velocity dispersion due to crystallization and melting of emulsion droplets. Ultrasonics 31(6), 433–437. McClements, D. J., Herrmann, N. and Hemar, Y. 1998. Influence of flocculation on the ultrasonic properties of emulsions: Theory. J. Phys. D 31, 2950–2955.

164

J. N. COUPLAND AND R. SAGGIN

Miles, C. A. and Cutting, G. L. 1974. Technical note: Changes in the velocity of ultrasound in meat during freezing. J. Food Technol. 9(1), 119–122. Miles, C. A., Fursey, G. A. J. and Jones, R. C. D. 1985. Ultrasonic estimation of solid/liquid ratios in fats, oils, and adipose tissue. J. Sci. Food Agric. 36(3), 218–228. Miles, C. A., Shore, D. and Langley, K. R. 1990. Attenuation of ultrasound in milks and creams. Ultrasonics 28(6), 394–400. Miyasaka, C. and Tittmann, B. R. 2000. Recent advances in acoustic microscopy for nondestructive evaluation. J. Pressure Vessel Technol. – Trans. Am. Soc. Mech. Engrs 122(3), 374–378. Mizrach, A., Galili, N. and Rosenhouse, G. 1989. Determination of fruit and vegetable properties by ultrasonic excitation. Trans. Am. soc. Agric. Engng 32, 2053–2058. Mizrach, A., Galili, N., Rosenhouse, G. and Teitel, D. C. 1991. Acoustical, mechanical, and quality parameters of winter-grown melon tissue. Trans. Am. Soc. Agric. Engng 34, 2135–2138. Mizrach, A., Galili, N., Gan-mor, S., Flitsanov, U. and Prigozin, I. 1996. Models of ultrasonic parameters to access avocado properties and shelf life. J Agric. Engng. Res. 65, 261–267. Mizrach, A., Flitsanov, U. and Fuchs, Y. 1997. An ultrasonic nondestructive method for measuring maturity of mango fruit. Trans. Am. Soc. Agric. Engng 40(4), 1107–1111. Mizrach, A., Flitsanov, U., El-Batsri, R. and Degani, C. 1999. Determination of avocado maturity by ultrasonic attenuation measurements. Sci Hort. 80(3–4), 173–180. Mulet, A., Benedito, J., Bon, J. and Rossello, C. 1999. Ultrasonic velocity in Cheddar cheese as affected by temperature. J. Food Sci. 64(6), 1038–1041. Nielsen, M. and Martens, H. J. 1997. Low frequency ultrasonics for texture measurements in cooked carrots (Daucus carota L.) J. Food Sci. 62(6), 1167–1170, 1175. O’Brien, W. D. and Dunn, F. 1972. Ultrasonic absorbtion mechanisms in aqueous solutions of bovine haemoglobin J. Phys. Chem. 76(4), 528–533. Ophir, J., Miller, R. K., Ponnekanti, H., Cespedes, I. and Whittaker, A. D. 1994. Elastography of beef muscle. Meat Sci. 36, 239–250. Ozguler, A., Morris, S. A. and O’Brien, W. D. 1998. Ultrasonic imaging of micro-leaks and seal contamination in flexible food packages by the pulse-echo technique. J. Food Sci. 63(4), 673–678. Papadakis, E. P. 1990a. The measurement of ultrasonic velocity. In “Ultrasonic Measurement Methods” (R.N. Thurston and A.D. Pierce, eds), vol. XIX, pp. 81–106. Academic Press, San Diego. Papadakis, E. P. 1990b. The measurement of ultrasonic attenuation. In “Ultrasonic Measurement Methods” (R.N. Thurston and A. D. Pierce, eds), vol. XIX, pp. 107–155. Academic Press, San Diego Pavlovskaya, G., McClements, D. J. and Povey, M. J. W. 1992. Preliminary study of the influence of dextran on the precipitation of leguminin from aqueous salt solutions. Int. J. Food Sci. Technol. 27, 629–635. Pinfield, V. J., Dickinson, E. and Povey, M. J. W. 1994. Modelling of concentration profiles and ultrasonic velocity profiles in a creaming emulsion: impotance of scattering effects. J. Colloid Interface Sci. 166, 363–374. Pinfield, V. J., Povey, M. J. W. and Dickinson, E. 1996. Interpretation of ultrasonic sound velocity in creaming profiles. Ultrasonics 34, 695–698. Povey, M. J. W. 1989. Ultrasonics in food engineering. Part 2: Applications. J. Food Eng. 9, 1–20. Povey, M. J. W. 1997. “Ultrasonic Techniques for Fluids Characterization.” Academic Press, San Diego. Povey, M. J. W. 1998a. Rapid determination of food material properties. In “Ultrasound in Food Processing” (M. J. W. Povey and T. J. Mason, eds), pp. 30–65. Blackie Academic and Professional, London. Povey, M. J. W. 1998b. Ultrasonics of food. Contemp. Phys. 39, 467–478.

ULTRASONIC SENSORS FOR THE FOOD INDUSTRY

165

Povey, M. J. W. and Harden, C. A. 1981. An application of the ultrasonic pulse echo technique to the measurement of crispness of biscuits. J. Food Technol. 16, 167–172. Povey, M. J. W. and Mason, T. J. (eds) 1998. “Ultrasound in Food Processing.” Blackie Academic and Professional, London. Povey, M. J. W., Golding, M., Higgs, D. and Wang, Y. T. 1999. Ultrasonic spectroscopy studies of casein in water. Int. Dairy J. 9(3–6), 299–303. Raichel, D. R. 2000. “The Science and Applications of Acoustics.” AIP Press, New York. Reddy, M. A. and Suryanarayana, C. V. 1981. Ultrasonic-absorption in aqueous-solutions of starch and gelatin. Acustica 48, 19–22. Richardson, P. and Povey, M. J. W. 1990. Ultrasonic temperature measurement and its potential for food processing systems. Food Control, 54–57. Ridgway, J. S., Henthorn, K. S. and Hull, J. B. 1999. Controlling of overfilling in food processing. J. Mater. Process. Technol. 93, 360–367. Rose, J. 1999. “Ultrasonic Waves in Solid Media.” Cambridge University Press, Cambridge. Saggin, R. and Coupland, J. N. 2001a. Oil viscosity measurement by ultrasonic reflectance. J. Am. Oil Chem. Soc. 78(5), 509–511. Saggin, R. and Coupland, J. N. 2001b. Concentration measurement by acoustic reflectance. J. Food Sci. 66(5), 681–685. Saggin, R. and Coupland, J. N. 2001c. Non-contact ultrasonic measurementts in food materials. Food Res. Int. 34(10), 865–870. Saravazyan, N. R. and Kharakoz, D. P. 1979. Ultrasonic investigation of the pH dependent solute-solvent interactions in aqueous solutions of amino-acids and proteins. J. Phys. Chem. 83(13), 1786–1789. Sarkar, N. and Wolfe, R. R. 1983. Potential of ultrasonic measurements in food quality evaluation. Trans. Am. Soc. Agric. Engng 26, 624–629. Sarvazyan, A. P. 1982. Development of methods of pecise ulrasonic measurements in small volumes of liquids. Ultrasonics 20, 151–154. Shiio, H. 1957. Ultrasonic interferometer measurements of the amount of bound water. Saccharides. J Am. Chem. Soc. 80, 70–73. Shore, D. and Miles, C. A. 1988a. Experimental estimation of the viscous component of ultrasound attenuation in suspensions of bovine skeletal-muscle myofibrils. Ultrasonics 26(1), 31–36. Shore, D. and Miles, C. A. 1988b. Attenuation of ultrasound in suspensions of bovine muscle myofibrils and myosin. Ultrasonics 26(3), 164–167. Shore, D. and Miles, C. A. 1988c. Attenuation of ultrasound in homogenates of bovine skeletalmuscle and other tissues. Ultrasonics 26(4), 218–223. Shore, D., Woods, M. O. and Miles, C. A. 1986. Attenuation of ultrasound in post-rigor bovine skeletal-muscle. Ultrasonics 24(2), 81–87. Shung, K. K. 1990. Basic principles of ultrasound tissue characterization. In “Noninvasive Techniques in Biology and Medicine” (S.E. Freeman, E., Fukushima and E.R. Greene, eds), pp. 205–217. San Francisco Press Inc., San Francisco, CA. Sigfusson, H., Decker, E. A. and McClements, D. J. 2001. Ulttrasonic characterization of Atlantic mackerel (Scomber scombrus) Food Res. Int. 34(1), 15–23. Slutsky, L. J. 1981. Ultrasonic chemical relaxation spectroscopy. Meth. Exp. Phys. 19, 179–235. Smith, D. E. and Winder, C. 1983. Effects of temperature, concentration, and solute structure on the acoustic properties of monosaccharide solutions. J. Food Sci. 48, 1822–1825. Strutt (Lord Rayleigh), J. W. 1945. “ The Theory of Sound.” Dover, New York. Suvanich, V., Ghaedian, R., Chanamai, R., Decker, E. A. and McClements, D. J. 1998. Prediction of proximate fish composition from ultrasonic properties: Catfish, cod, flounder, mackerel and salmon. J. Food Sci. 63(6), 966–968.

166

J. N. COUPLAND AND R. SAGGIN

Tittmann, B. R., Ettinger, K., Kalternbacher, M. and Bhardwadj, M. 1998. Air-coupled and noncontact ultrasonics. In “Proceedings of the ASNT Spring Conference and 7th Annual Research Symposium,” Anaheim 23–27 March 1998. American Society for Nondestructive Testing, Columbus, OH. Varadan, V. K. and Gardner, J. W. 1999. Smart tongue and nose. In “Proceedings of the SPIE Conference on Smart Electronics and MEMS,” Newport Beach, 1–3 March 1999 (V. K. Varadan, ed.), pp. 67–76, SPIE, Bellingham. Wade, T. and Beattie, J. K. 1997. Electroacoustic determination of size and zeta potential of fat globules in milk and cream emulsions. Colloids Surf. B 10(2), 73–85. Wade, T., Beattie, J. K., Rowlands, W. N. and Augustin, M. A. 1996. Electroacoustic determination of size and zeta potential of casein micelles in skim milk. J Dairy Res. 63, 387–404. Wang, Y. T. and Povey, M. J. W. 1999. A simple and rapid method for the determination of particle size in emulsions from ultrasound data. Colloids Surf. B 12(3–6), 417–427. Waterman, P. C. and Truel, R. 1961. Multiple scattering of waves. J. Math. Phys. 2, 512–537. Winder, W. C., Consigny, N. P. and Rodriguez-Lopez, B. 1961. An ultrasonic method for measurement of solids non-fat and milk fat in fluid mulk. II. Evaluation of the method. J. Dairy Sci. 44, 1165. Winder, W. C., Aulik, D. J. and Rice, A. C. 1970. An ultrasonic method for direct and simultaneous determination of alcohol and extract content of wines. Am. J. Enol. Vint. 21, 1–11. Wirtanen, G., Ahvenainen, R. and Mattila-Sandholm, T. 1992. Non-destructive detection of spoilage of aseptically-packed milk products: effect of frequency and imaging parameters on the sensitivity of ultrasound imaging. Lebens. Wiss. Technol. 25, 126–132. Withers, P. M. 1996. Ultrasonic, acoustic and optical techniques for the non-invasive detection of fouling in food processing equipment. Trends Food Sci. Technol. 7, 293–298. Wood, A. B. 1955. “A Textbook of Sound.” Bell, London. Yamamoto, H., Iwamoto, M. and Hagimuna, S. 1980. Acoustic impulse response method forr measuring natural frequency of intact fruits and preliminary applications to internal quality evaluation of apples and watermelons. J. Texture Stud. 11, 117–136. Zacharias, E. M. and Parnell, R. A. 1972. Measuring the solids content of foods by sound velocimetry. Food Technol. 26, 160–166.

OZONE AND ITS CURRENT AND FUTURE APPLICATION IN THE FOOD INDUSTRY JIN-GAB KIM, AHMED E. YOUSEF and MOHAMMED A. KHADRE Department of Food Science and Technology The Ohio State University Columbus, OH 43210 USA

I. Introduction II. Ozone Chemistry and Physics: An Overview A. Ozone Solubility B. Ozone Stability C. Ozone Reactivity III. Medium for Ozone Treatment A. Temperature B. Relative Humidity C. Residual Ozone D. Ozone Demand of the Medium IV. Reactor and Equipment Considerations A. Ozone Interaction with Processing Equipment V. Application of Ozone in Food Processing A. Targeted Microorganisms B. Inactivation Kinetics and Mechanisms C. Ozone as an Alternative Sanitizer in Food Processing D. Food Properties and Ozone Applicability E. Ozone Applications at Different Stages of Processing VI. Selected Food Applications A. Raw Poultry and Meats B. Fruits and Vegetables C. Fish Processing and Storage D. Dry Food and Food Ingredients E. Packaging Material and Food Contact Surfaces F. Pesticides on Agricultural Commodities VII. Combination Treatments A. Ozone and Hydrogen Peroxide B. Ozone and Chlorine C. Ozone and Other Gases D. Ozone and Heat

ADVANCES IN FOOD AND NUTRITION RESEARCH VOL 45 ISBN: 0-12-016445-0

Copyright © 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

168

J.-G. KIM ET AL.

E. Ozone and Ultraviolet Radiation F. Ozone and Pulsed Electric Field VIII. Analytical Methods IX. Regulatory Status X. Limitations, Toxicity and Safety Acknowledgement References

ABSTRACT

The food industry is interested considerably in using ozone to enhance the shelf-life and safety of food products and in exploring new applications of the sanitizer. This interest was recently accompanied by a US governmental approval of ozone for the safe use, in gaseous and aqueous phases, as an antimicrobial agent on food, including meat and poultry. Ozone has a strong microbicidal action against bacteria, fungi, parasites and viruses when these microorganisms are present in low ozone-demand media. Readily available organic constituents in food, however, compete with microorganisms for applied ozone and thus efficacy of the treatment is minimized. Ozone is suitable for washing and sanitizing solid food with intact and smooth surfaces (e.g., fruits and vegetables) and ozone-sanitized fresh produce has recently been introduced in the US market. Use of ozone to sanitize equipment, packaging materials, and processing environment is currently investigated. Efforts to decontaminate bean sprouts and remove biofilm with ozone have not been successful. The antimicrobial efficacy can be enhanced considerably when ozonation is combined with other chemical (e.g., hydrogen peroxide) or physical (e.g., ultraviolet radiation) treatments. Mechanical action is also needed as a means to dislodge microorganisms from the surface of food and expose them to the action of the sanitizer. The food industry also is interested in using ozone to decontaminate processing water and decrease its chemical and biological oxygen demand. This application improves the reusability of processing water and allows for environment-friendly processing operations. I. INTRODUCTION

Ozone has been applied industrially for many years, mostly in water treatments, because of its high oxidizing power and superior antimicrobial properties. Use of ozone in the food industry, however, has been limited mainly to shelf-life extension of commodities during storage. Recently,

OZONE APPLICATION IN THE FOOD INDUSTRY

169

there has been a renewed interest in ozone and its application in food processing. Application of ozone for decontamination of poultry chiller water seems promising (Sheldon and Chang, 1987; Waldroup et al., 1993; Diaz and Law, 1999) and washing fruits and vegetables with ozone is gradually gaining acceptance. Novel application of this powerful sanitizer will be addressed in this chapter. Current sanitization technologies are crucial to maintaining the quality and enhancing the safety of fresh agricultural commodities. These technologies, however, have many drawbacks and some treated products are potentially hazardous to consumers. Safety of produce, which is commonly treated with chlorine (or occasionally consumed unsanitized), is currently questionable because of frequent disease outbreaks associated with these products. Pathogens resistant to preservation factors, such as acid-tolerant Salmonella spp., Listeria monocytogenes and Escherichia coli O157:H7, have emerged as a serious threat to the fresh produce industry (Beuchat, 1995; Odumeru et al., 1997; NACMCF, 1999). Salmonellosis outbreaks have been associated with pre-cut watermelons and cantaloupes and fresh tomatoes contaminated with S. Javiana, S. Montevideo and S. Poona (Gayler et al., 1955; Ries et al., 1990; Wood et al., 1991; CDC, 1993; LA Times, 2001). Water used to wash the tomatoes was implicated as the source of contamination. Outbreaks linked to consumption of unpasteurized apple juice and cider (Besser et al., 1993; CDC, 1996b) also may be attributed to the failure in sanitizing apples properly. Oocysts of Cryptosporidium parvum, a zoonotic protozoan parasite, have been detected in unpasteurized cider which caused a disease outbreak (Millard et al., 1994). In 1996, a large epidemic in the United States and Canada involving another protozoan parasite, Cyclospora cayetanensis, was epidemiologically linked to raspberries that were imported from Guatemala (CDC, 1996a). Use of nonpotable water for spraying the plant with fungicide was suggested as the source of the pathogen. In addition to fresh produce, the meat industry can also benefit greatly from new developments in sanitization technology. Meat products have caused numerous foodborne disease outbreaks. Listeria monocytogenes has been associated with ready-to-eat meat products resulting in multistate outbreaks of listeriosis in 1998 and 2000 (CDC, 1998, 2000). Escherichia coli O157: H7 is traditionally linked to beef products (CDC, 1996c, 1997). This pathogen has caused several disease outbreaks due to consumption of undercooked meat, and the bacterium most likely entered the processing chain on contaminated carcasses. It is obvious that effective, reliable, economical and industry-relevant alternative sanitization methods are needed. At present, chemical disinfectants such as chlorine and hypochlorites are commonly used as

170

J.-G. KIM ET AL.

sanitizing agents in the food industry. Treatments with 50–100 ppm free chlorine solutions reduce initial contamination of vegetables (Carlin et al., 1995), but these levels of the sanitizer may lead to discoloration and production of off-flavors in fresh produce (Hurst and Schuler, 1992). Additionally, chlorination of food could lead to the formation of toxic and carcinogenic chlorinated compounds (Brungs, 1973; Page et al., 1976; Kirk and Mitchell, 1980). Ozone is an effective, chlorine alternative, sanitizer with superior antimicrobial properties (Kessel et al., 1943; Ito and Seeger, 1980; Korich et al., 1990). It is capable of inactivating bacteria, bacterial spores, molds, yeasts, protozoan cysts and viruses at relatively low concentration and in short exposure time when applied to pure cell suspensions (Giese and Christenser, 1954; Scott and Lesher, 1963; Kim, 1998). Ozone has been tested on nearly every type of food during storage and processing to improve the safety and to extend the shelf-life of these products. The ability of ozone to inactivate contaminant microflora on food is variable; in some instances, however, ozone decreased food microflora more than 5 log units (Yousef and Rodriguez-Romo, 2001). Ozone not only inactivates microbial contaminants, but is also potentially useful in decreasing the level of pesticides, such as azinphosmethyl, captan, formethanate-HCl and ethylenethiourea, on fresh produce (Ong et al., 1996; Hwang, 1999). The chemical oxygen demand (COD) and biological oxygen demand (BOD) of water used in washing and processing of foods can be decreased appreciably by ozonation (Sheldon and Brown, 1986). Thus, use of ozone minimizes the accumulation of inorganic waste in the environment (Horvath et al., 1985). Moreover, rapid decomposition to oxygen and lack of toxic residues make ozone a favorable environment-friendly sanitizer. Ozone is currently used in many countries and its use in food processing has been approved recently in the United States (Federal Register, 2001). Additionally, ozone-treated produce has just been introduced in the United States market. This chapter addresses current applications of ozone in the food industry, problems that were recently encountered in attempts to apply ozone in food processing, and some probable and challenging future applications. Some of the application problems originated from lack of basic knowledge on sanitization and others are ozone-specific. Recent ozone findings are presented with emphasis on improving the safety of fresh produce. II. OZONE CHEMISTRY AND PHYSICS: AN OVERVIEW

Ozone is formed in the stratosphere (15–35 km altitude) by the action of short ultraviolet (UV) solar radiation (< 240 nm) on molecular oxygen and

OZONE APPLICATION IN THE FOOD INDUSTRY

δ+

δ−

.. .. O

δ−

.O . −

δ

..

O

..

..

.. ..O..

.. O .. ..

O

.. O .. ..

δ

.. .. O

.. O ..

..

.O . −

δ+

..

.. O

..

δ+

171

.. .. O

δ+

FIG. 1. Resonance structures in ozone molecules (Trambarulo et al., 1953).

TABLE I OXIDATION–REDUCTION POTENTIAL (VOLTS) OF DIFFERENT CHEMICAL OXIDANTS

Agent Fluorine Ozone Hydrogen peroxide Potassium permanganate Hydrobromous acid Hypochlorous acid Chlorine Chlorine dioxide Oxygen Chromic acid Bromine Nitric acid Iodine

Molecular formula

Oxidation–reduction potential

F2 O3 H2O2 KMnO4 HOBr HOCl Cl2 ClO2 O2 H2CrO4 Br2 HNO3 I2

2.87 2.07 1.78 1.70 1.59 1.49 1.36 1.27 1.23 1.21 1.09 0.94 0.54

a small portion of ozone is transported to the troposphere (< 15 km altitude). About 10% of the atmospheric ozone is present in the troposphere but a very small concentration of ozone occurs naturally at the Earth’s surface (Wojtowicz, 1996). Large amounts of the gas can be synthesized by generators for industrial use. Ozone is a triatomic molecule (O3) that is considered to be an allotropic modification of oxygen. It has a relative molecular mass of 48 and its molecular structure is a resonance hybrid of the four canonical forms having delocalized bonding (Figure 1). Pure ozone is a pale blue gas and bluish liquid with pungent and characteristic odor. Ozone exists in the gaseous state at room and refrigeration temperature and it is partially soluble in water. Ozone has an oxidation–reduction potential of 2.07 V (Brady and Humiston, 1978), which makes it the strongest oxidant currently available for food applications (Table I). The density of ozone in the gaseous state is 2.14 g L–1 at 0°C and 101.3 kPa (Wojtowicz, 1996), which is greater than that of air (1.28 g L–1) under similar conditions.

172

J.-G. KIM ET AL. A. OZONE SOLUBILITY

Ozone is partially soluble in water and its solubility depends on several physical parameters. By Henry’ law, ozone solubility in liquid is directly proportional to the pressure that the gas exerts above the liquid (Bablon et al., 1991). The most important parameter affecting the solubility of ozone is probably water temperature. Based on Henry–Dalton constants, the solubility of ozone in water is higher at lower temperatures. Watson (1943) reported that the solubility ratio (ozone concentration in water phase/ ozone concentration in gas phase) was 0.26 at 20°C. Meddows-Taylor (1947) reported a solubility ratio of ~ 0.4 at 20°C. According to Horvath et al. (1985), the solubility ratio of ozone in water was 0.31–1.13, depending on the water temperature. Our data showed that solubility ratio varied with the source of water used to solubilize ozone. The solubility ratio was 0.16 for distilled water at ~ 22°C, after ozone was bubbled in the water at 29.4 mL min–1 for 18 min (Kim, 1998). When ozone was sparged for 18 min into tap (from two different sources) and deionized water under similar conditions, the solubility of ozone was 38, 56 and 95% of that for distilled water, respectively (Figure 2). Variations in published solubility data may be attributed to differences in the reactor design, gas flow rate and analytical method used to quantitate ozone. 1.0

Distilled

Absorbance at 258nm

0.8 Deionized 0.6 Tap I 0.4 0.2

Tap II

0.0 0

4

8

12

16

20

Minutes FIG. 2. Treatment of water from different sources with ozone gas (~ 2.5%, v/v) at a flow rate of 29.4 mL min–1 (Kim, 1998). Tap I, tap water in the research laboratory; Tap II, drinking fountain water.

OZONE APPLICATION IN THE FOOD INDUSTRY

173

High pH interferes with the solubility of molecular ozone. As the pH of ozone solutions increases, the rate of decomposition of molecular ozone into hydroxyl radical also increases (Adler and Hill, 1950; Hewes and Davison, 1973). Researchers attributed the rapid decomposition of ozone in aqueous solutions with high pH to the catalytic activity of the hydroxyl ion. Farooq et al. (1977) noted a greater survival rate of Mycobacterium fortuitum during ozone treatment when the pH of the treatment medium was increased. The authors attributed this increased survival to a smaller ozone residual concentration as the pH of water increased. Hydroxide ions are consumed in initiating the ozone decomposition process in water; therefore, ozonation of a solution can decrease its pH value. Sheldon and Chang (1987) found that the pH of poultry-processing water, containing a high organic load, decreased from 6.9 to 5.6 after 50 min ozonation. B. OZONE STABILITY

Ozone is more stable in the gaseous than in the aqueous phase (Stumm, 1958). Stability of dissolved ozone (measured as half-life) is affected by its concentration, pH of the aqueous medium, exposure to UV radiation, presence of radical scavengers (Tomiyasu et al., 1985; Kim, 1998), application of turbulence (Shechter, 1973), temperature (Sease, 1976), and presence of organic matter and metal ions (Horvath et al., 1985). Decomposition of ozone in water does not always follow the first-order rate law (Gurol and Singer, 1982; Tomiyasu et al., 1985; Yurteri and Gurol, 1988). The stability of ozone in water decreases when the pH of the medium increases. High pH is also believed to interfere with the solubility of molecular ozone (Roth and Sullivan, 1981; Ouederni et al., 1987). Kim (1998) found that ozone decomposes rapidly in phosphate buffer when the pH is greater than 8.0 (Figure 3). The researcher bubbled ozone in different types of water and phosphate buffer (pH 7.0) at 25°C, to attain 1.6–2.5 ppm, and monitored ozone decomposition spectrophotometrically. The rate of ozone decomposition was greater in tap water and buffer than it was in distilled, deionized and HPLC-grade water (Figure 4). Ozone stability in the water is greatly influenced by the presence of contaminants, particularly metal ions. According to Bablon et al. (1991), however, ionic strength resulting from mineralization of drinking water (< 1000 mg L–1 total dissolved solids) does not affect the solubility of ozone appreciably. Temperature plays an important role in the stability of ozone in solutions. The half-life of ozone in the gaseous state is approximately 12 h at room temperature and in pure, clean water (pH 7–8) it is 20–30 min (Graham, 1997). The stability of ozone in solutions also depends greatly on the amount of ozone-demand material in the water. Rosenthal (1974)

174

J.-G. KIM ET AL.

Ozone concentration(ppm)

14.0 12.0

IC=13.4

10.0

IC=11.0

8.0

IC=9.4

6.0 IC=6.5

4.0

IC=4.6

2.0

IC=2.6

0.0 5.0

6.0

7.0

8.0

9.0

pH FIG. 3. Ozone stability in phosphate buffer having different pH values (Kim, 1998). IC, calculated initial ozone concentration in the ozone–buffer mixture.

Residual ozone fraction (N/N0)

1.0

0.8

0.6

0.4

0.2

0.0 0.0

Distilled Deionized Tap water I Tap water II HPLC Buffer

2.0

4.0

6.0

8.0

Minutes FIG. 4. Ozone decomposition in water from different sources (Kim, 1998). N0, ozone concentration initially; N, residual ozone concentration.

OZONE APPLICATION IN THE FOOD INDUSTRY

175

1. Cycloaddition δ+ O δ−O O

C

δ− O δ+ O O C δ+

C

O

C δ−

O

O

C

C

O C

C

+ H2O2

Ozonide

2. Electrophilic reactions O

O3

O δ+

Oδ−

H

δ−

+

O O O

OH

3. Nucleophilic reactions

O

O3

δ−O δ+

O

δ+ O O δ−

O

+ O=O

O-

O C=O

COOH COOH

FIG. 5. Ozone in cycloaddition, electrophilic (Komissarov and Komissarova, 1973) and nucleophilic reaction (Leffler, 1949).

reported that the half-life of ozone was much longer in double-distilled water (> 85 min) at 20°C than in distilled or tap water (~ 20 min). Kim (1998) found that the half-life of ozone at 25°C in deionized and tap water was 12 and 6 min, respectively. The half-life of ozone in a solution containing 0.01 mol perchloric acid is ~ 18 h at 25°C and in the presence of 0.001 mol iron perchlorate is ~ 83 h. The half-life of free radicals is commonly measured in microseconds. C. OZONE REACTIVITY

1. Molecular ozone Ozone undergoes three types of reactions in organic solvent media (Bailey, 1978): (1) dipolar cyclo-addition with unsaturated carbon–carbon bonds; (2) electrophilic reaction with aromatic compounds, amines and sulfides having strong electronic density; and (3) nucleophilic reaction with carbons carrying electron-withdrawing groups (Figure 5). Therefore, the molecular ozone reactions are selective and limited to unsaturated aromatic and aliphatic compounds as well as to specific functional groups.

176

J.-G. KIM ET AL.

2. Products of decomposition – free radical species Ozone gas is very unstable and decomposes quickly in the air. Because of its instability, the gas is generally produced at the point of application, sparged in water and applied immediately in a closed system. The following discussion addresses ozone decomposition in aqueous solutions. At low concentrations, auto-decomposition of dissolved ozone is an apparent first-order reaction with respect to ozone (Masschelein, 1982; Finch et al., 1988; Peeters et al., 1989). Auto-decomposition of ozone is accompanied by the production of numerous free radical species such as hydroperoxyl (HO2˙), hydroxyl (˙OH) and superoxide (˙O2–) radicals (Adler and Hill, 1950; Hoigné and Bader, 1975). The high reactivity of ozone is attributed to the oxidizing power of these free radicals. According to Jans and Hoigné (1998), when a mole of aqueous ozone decomposes, ~ 0.5 mol of ˙OH is produced, regardless of whether the transformation is catalyzed by hydroxide anions (i.e. elevated pH), addition of H2O2, or exposure to UV

M Moxid

O3 OH– O2

In In+

Initiation HO2–

•O – 2

H2O2 ROO •

HO2•

•O – 3

R•

Poxid O2 Promotion

H+ O2 HO3•

P

•OH Hydroxyl radical

Ih

HO2• / •O2– are not formed

Inhibition

FIG. 6. Ozone decomposition reactions (adapted from Jans and Hoigné, 1998). M, Solute; In, initiator (e.g. OH–, HO2–, Fe2+, HCOO–, –SH, UV); P promotor (e.g. O3, –SH, R–CH2OH, Aryl); Ih, inhibitor (e.g. Alkyl, HCO3–, CO32–, t-BuOH, –SH).

OZONE APPLICATION IN THE FOOD INDUSTRY

177

irradiation. The hydroxyl radical is an important transient species and chain propagating radical. The reactions of hydroxyl radical with many substrates are very fast (Hoigné and Bader, 1975). Typical rate constants for reactions of hydroxyl radical with organic solutes are in the range 108 to 1010 M–1 s–1 (Farhataziz and Ross, 1977). Ozone decomposition occurs in a chain reaction process (Figure 6) including initiation, propagation, and termination steps (Weiss, 1935; Staehelin et al., 1984; Jans and Hoigné, 1998). The initiators are compounds capable of inducing the formation of the superoxide radical (˙O2–); these include hydroxyl ions (OH–), hydroperoxide ions (HO2–), and some cations and organic compounds (e.g. glyoxylic acid, formic acid and humic substances). Ultraviolet radiation at 253.7 nm also can initiate the free-radical generation process. Promotion reactions regenerate the superoxide radical from the hydroxyl radical. Common promoters include aryl groups, formic acid, glyoxylic acid, primary alcohols and humic acids. Phosphate species are important inorganic promoters. The superoxide anion also can promote the decomposition of ozone. The inhibition reactions lead to the consumption of OH radicals without regenerating ˙O2–. Some of the common inhibitors include bicarbonate and carbonate ions, alkyl groups, tertiary alcohols and humic substances (Hoigné and Bader, 1985). Antioxidants such as tocopherol and ascorbic acid from food also can scavenge the free radicals and thus block the chain reaction.

3. Reactions with inorganic compounds Minerals, metal ions, hydroxyl ions and halogens (e.g. chlorine) catalyze ozone decomposition and this increases ozone demand (Alder and Hill, 1950; Hewes and Davidson, 1973; Hoigné and Bader, 1976). Reaction of ozone with inorganic compounds found in water usually follows a firstorder kinetics, with regard to ozone and the oxidizable compound. Ozone oxidizes ferrous (Fe2+) into ferric (Fe3+) species, which precipitate in water as ferric hydroxide, Fe(OH)3, and is easily removed by filtration. Similarly, the manganous ion (Mn2+) is oxidized by ozone into the manganic (Mn4+) state, forming manganic oxide (MnO2), which also precipitates out and can be filtered. These reactions are important for the removal of contaminant metals from drinking water (Dore, 1989). The oxidation–reduction potential values for ozone, chloride, bromide and iodide are 2.07, 1.49, 1.33 and 0.99 V, respectively (Table I). Hence, ozone can oxidize chloride ions slowly, but it reacts moderately with bromide and rapidly with iodide ions, producing elemental bromine and iodine, respectively (Haag and Hoigné, 1983).

178

J.-G. KIM ET AL. III. MEDIUM FOR OZONE TREATMENT

When applied in food processing, ozone gas is used for food storage applications and the aqueous form is used in the surface decontamination of food equipment or packaging materials. Municipal water is commonly used for various washing purposes and dissolution of sanitizers; therefore, this water is the medium of choice for most aqueous ozone applications in food processing. The properties of the treatment medium considerably affect the efficacy of ozone treatment of food. Ozone demand, for example, resulting from dissolved substances in municipal water should be met before the desired sanitizing action occurs. The properties of the medium of significance to ozonation efficacy (i.e. temperature, relative humidity, residual ozone and ozone demand) will be addressed in this section. A. TEMPERATURE

Several researchers have tested the relationship between ozone efficacy and treatment temperature, but their results seem inconclusive. Kuprianoff (1953) found that ozone was more effective against microorganisms when applied at low (< 10°C) than at high temperatures. Herbold et al. (1989) also reported that the effectiveness of ozone against hepatitis A virus (HAV) and E. coli diminished when the temperature increased from 10°C to 20°C. Katzenelson et al. (1974), however, indicated that lowering the temperature from 5°C to 1°C had only a minor effect on the inactivation kinetics of microorganisms. According to Kinman (1975), there is no difference in the disinfection rate by ozone when applied at 0°C or 30°C. Achen and Yousef (2001) treated apples inoculated with E. coli O157:H7 in bubbling ozone water at 4, 22 and 45°C for 3 min. The researchers found that residual ozone concentration following the treatments were 36, 22 and 18 mg L–1, respectively, and no significant difference in ozone efficacy between the treatments at three different temperatures was found (P > 0.05). The disagreement among researchers may be due to the changes in ozone properties at different temperatures. A decrease in the temperature of an aqueous medium increases the solubility and stability of ozone. On the contrary, an increase in temperature enhances the reactivity of residual ozone. The relative contribution of these two factors (solubility/stability and reactivity) to ozone efficacy may vary with the experimental setup. B. RELATIVE HUMIDITY

High humidity is needed for inactivation of microorganisms by ozone gas. It is believed that hydration of dry microorganisms in humid atmospheres

OZONE APPLICATION IN THE FOOD INDUSTRY

179

makes them susceptible to ozone. The optimum relative humidity (RH) for microbial inactivation by gaseous ozone is 90–95%, and the gas loses its bactericidal effect at ⭐ 50% RH (Kuprianoff, 1953). Ozone, however, decomposes more rapidly at high than at low RH values. Elford and van den Ende (1942) used low ozone concentrations and long exposure time at variable relative humidity to disinfect airborne microorganisms. At < 45% RH, the germicidal power of ozone was negligible. Inactivation was substantial at high humidities even when ozone concentration was < 0.1 mg L–1. Ewell (1946) also demonstrated that microorganisms were killed more readily by ozone in an atmosphere having high rather than low RH. Hoffman (1971) indicated that not only were desiccated microorganisms more resistant than hydrated cells to sterilization by ozone, but once desiccated, some cells were difficult to rehydrate sufficiently to be susceptible to ozone sterilization. Ozone, therefore, is an effective sanitizer only against well-hydrated microbial cells. J. G. Kim and A. E. Yousef (Table II, unpublished data) found a similar reaction of ozone in a powdered, food-grade, anticaking agent containing natural contaminants. Application of 200 ppm gaseous ozone caused a minimal decrease in the microbial load of the anticaking agent with a water activity (aw) of ⭐ 0.84. When an anticaking agent that contained aw of 0.96 was treated with 150 ppm gaseous ozone, the microbial load decreased by more than 2 log units. Application of 300 ppm gaseous ozone decreased the microbial load of this agent to an undetectable level. When aw of a drier anticaking agent TABLE II INACTIVATION BY GASEOUS OZONE OF NATURAL MICROBIAL CONTAMINANTS ON SILICA-BASED ANTICAKING AGENTS WITH DIFFERENT WATER ACTIVITIES AND PH VALUES

Ozone dose (ppm)a

Count (CFU g–1)

8.03

0 200

2.7 × 102 1.7 × 102

0.95

8.23

0 150 300

8.2 × 103 3.0 × 101 < 10

2

0.96

3.16

0 150 300

5.0 × 103 1.5 × 101 < 10

3

0.10

7.00

0 200

3.4 × 103 2.3 × 103

Anticaking agent

aw

pH

1

0.84

1 (water added)

CFU, colony-forming unit. a µg gaseous ozone per g powder.

180

J.-G. KIM ET AL.

was increased from 0.84 to 0.95, ozone was as effective in decreasing the microbial load as it was in the product that naturally contained a high water activity. Microbial contaminants in the powder were mostly fungal and bacterial spores, which can survive in a suboptimal growth environment. C. RESIDUAL OZONE

The term “residual ozone” is used in this chapter to refer to the detectable concentration, in the treatment medium, of ozone after it has been applied to the targeted food product. The effectiveness of ozone against microorganisms depends on the amount applied, but more so on the residual ozone in the medium. The stability of ozone under application conditions and the presence of ozone-demanding material in the treatment medium greatly affect the level of residual ozone available for disinfection of the food product. Venosa (1972) pointed out that one of the most serious failures by various investigators has been their inability to distinguish between the concentration of applied ozone and residual ozone necessary for effective sanitization. Therefore, in addition to the applied dose, the availability and the decay of ozone during the course of the treatments should be reported, otherwise the actual effective dose used may be overestimated. The results of the following studies illustrate the importance of monitoring residual ozone. Izat et al. (1990) chilled eviscerated broiler carcasses in chlorinated water (20 ppm) or ozonated water at 1.7°C to 4.4ºC. The oxidation-reduction potential (ORP) of the control (chlorinated water) was 900 mV and this value decreased as carcasses were introduced into the water. The ozone generator produced 20 g h–1, resulting in an average ORP of 270 mV in the chiller water. After the first 20 carcasses were chilled, total microbial and presumptive coliform counts were significantly higher (unexpectedly) in the ozonated side of the chiller than in the control side. This suggests that either insufficient ozone was generated or the ozone was not remaining in solution for a sufficient time to affect bacterial numbers. In another study, shrimp meat was inoculated with Vibrio cholerae, V. parahaemolyticus, Flavobacterium aquatile, Pseudomonas aeruginosa, P. putida, P. fluorescens, E. coli, Salmonella typhimurium and Staphylococcus aureus. Inoculated meat was immersed in ozonated, 2% saline solution (2.9–4.8 mg L–1 ozone, 5ºC) and flushed with ozone at 150 mL min–1 continuously for 60 min (Chen et al., 1992). Ozone concentration decreased by more than 1.4 mg L–1 within 15 min but gradually increased thereafter. Reduction of bacterial count during the first 15 min of flushing was < 1 log unit except for E. coli, which had a 2-log unit decrease. A low ozone

OZONE APPLICATION IN THE FOOD INDUSTRY

181

concentration, therefore, is ineffective for disinfection of food products when extraneous organic matter is present. Bullock et al. (1997) used ozone to treat water in a recirculating rainbow trout (Oncorhynchus mykiss) culture to reduce the heterotrophic bacterial counts in system water and to prevent bacterial gill disease (BGD). Applied ozone was 25 or 36–39 g per kg of feed. Less than 90% reduction of F. branchiophilum in water or on gill tissue was achieved but BGD outbreaks were prevented. The limited decrease in bacterial count may be attributed to the short exposure time to ozone (35 s contact chamber) and rapid loss of oxidation caused by suspended organic matter in the medium. D. OZONE DEMAND OF THE MEDIUM

When compared with other treatment media, pure water has the least ozone demand. Impurities in water react with applied ozone and generate demand. Some of these impurities may initiate ozone decomposition (Hoigné and Bader, 1976); these include glyoxylic acid, formic acid and humic substance. Kim (1998) bubbled several types of water used in the laboratory with ozone gas (1.1 mM) at a flow rate ~ 30 mL min–1 (Figure 2). Ozone dissolved faster in deionized and distilled water than in tap water, resulting in a higher maximum ozone concentration in the former water. The solubility of ozone was > 2-fold greater in deionized and distilled water than in tap water. Yang and Chen (1979) reported that the bactericidal efficacy of ozone was lower in Ringer solution, and in solutions containing NaCl (5%) or egg albumin, compared with distilled water. The tested substances decomposed ozone and thus decreased the amounts of residual ozone available for reaction with microorganisms. According to Restaino et al. (1995), death rates of some microorganisms in ozonated water, which contained organic matter, were not significantly affected by 20 ppm soluble starch but were significantly reduced by the addition of 20 ppm bovine serum albumin (BSA). Residual ozone in water containing BSA was significantly lower than in deionized water and water with soluble starch. Achen (2000) varied BSA concentration in E. coli O157:H7 suspension and treated the mixture with ozone. The researcher observed no inactivation of the pathogen when > 0.1% BSA was added (Figure 7). Ogawa et al. (1990), however, found that addition of 0.5 g loamy soil per liter of ozone solution (1.5–3.8 mg L–1) did not affect the ability of ozone to inactivate spores of Mucor piriformis, Botrytis cinerea and Phytophthora parasitica. Antioxidants originating from food may generate ozone demand by scavenging radicals formed during ozone decomposition. Food additives such as acids, surfactants or sugars can stabilize or destabilize ozone, depending on their properties.

182

J.-G. KIM ET AL. 10 9 8

Log CFU/mL

7 6

Control

5

1%BSA

4

0.1%BSA 3

0.01%BSA

2 1 0 0

0.5

1

1.5

2

2.5

3

3.5

4

Ozone concentration (ppm)

FIG. 7. Counts (CFU mL–1) of Escherichia coli O157:H7 after exposure to different concentrations of ozone at 4°C for 20 s, in the presence of bovine serum albumin (Achen, 2000).

IV. REACTOR AND EQUIPMENT CONSIDERATIONS

For the treatment of food with aqueous ozone, the sanitizer must be transferred from the gas to the liquid phase. The design of treatment chambers and diffusion systems is important to maximize ozone transfer for the intended purpose and to make the process economically feasible. Dissolution/contacting units to provide aqueous ozone for application vary, depending on the specific functions of ozone at the points of application. A number of techniques are available for dissolution of ozone in the liquid. These include conventional fine bubble diffusion, turbine mixers, injectors, packed columns, spray chambers, deep U-tubes, porous plate diffuser contactors, and submerged static radial turbine contactors (Bellamy et al., 1991). The mass transfer of ozone occurs via diffusion through the gas–water interface. Favorable conditions for ozone mass transfer include high concentration of ozone in the carrier gas and high pressure (Gomella, 1972). According to Henry’s law, high pressure above the process liquid

OZONE APPLICATION IN THE FOOD INDUSTRY

183

increases ozone solubility. White (1986), however, emphasized the importance of contactor efficiency, since maintaining high partial pressures of ozone above the process liquid is sometimes not attainable. Decreasing the diameter of the ozone bubbles increases the total area of exchange and the contact time between water and gaseous ozone (Harris, 1972). MeddowsTaylor (1947) defined “contact value” as the total area of gas bubbles multiplied by the time required to rise a unit distance. The author found that bubbles with a diameter of 0.1 cm have ~ 32 times more contact value than those with a diameter of 1.0 cm. In our experimental work (e.g. Kim and Yousef, 2000b), a sparger with 10 µm pore size was used effectively to decrease the bubble diameter of ozone gas and to increase the transfer rate of gaseous ozone into the water phase. In addition, stirring was carried out during ozonation to ensure sufficient turbulence. Efficient ozone use in food processing can be achieved by adding a filtration apparatus prior to the contactor. Filtration may keep the level of nontarget demand substances to a minimum and improve ozone dissolution in the contactor (Hampson and Fiori, 1997). Prefiltration of poultry chiller water to decrease the organic load prior to ozone treatment is also recommended for optimum reduction of microbiological levels and efficient use of ozone (Sheldon, 1986; EPRI, 1999). When Sheldon (1986) filtered spent poultry processing water with diatomaceous earth (DE) prior to ozonation, the researcher obtained high-quality water. The filtration–ozonation combination decreased COD, total solids, fats/oil/grease (FOG), total aerobic microorganisms, coliforms and Salmonella spp. by 87%, 65%, 95%, 3 log units, 2.7 log units and 3 log units, respectively. A similar combined treatment was applied on whole-bird rinse water from commercial poultry processing plants (Sheldon and Chang, 1987). Measurable reductions in levels of COD (92%), absorbance at 280 nm (88%), total solids (59%), total volatile solids (82%), aerobic plate count (> 3 log units), coliforms (2 log unit), E. coli (2 log units), and Salmonella spp. (3 log units) were detected following the filtration–ozonation treatment. Filtration was more effective and rapid when DE rather than sand was used in removing contaminants from processing water.

A. OZONE INTERACTION WITH PROCESSING EQUIPMENT

Ozone not only reacts with contaminants (including microorganisms) in the treatment medium and food, it may also interact with the reactor and equipment materials. Efficacy of ozone treatment, therefore, may be influenced by the type of materials used to manufacture the equipment.

184

J.-G. KIM ET AL. TABLE III COMPATIBILITY OF DIFFERENT MATERIALS WITH OZONEa.

Material

Theoretical ratingb

304 stainless steel 316 stainless steel Aluminum Bronze Copper ABS plastic Acetal (Delrin) CPVC EPDM Hypalon Hytrel Kel-F LDPE Polycarbonate Polypropylene PTFE (Teflon) PVC PVDF (Kynar) Silicone Viton Natural rubber Neoprene Nylon Buna N (Nitrile)

Good Excellent Good Good Excellent Good Fair Excellent Excellent Excellent Fair Excellent Fair Excellent Good Excellent Good Excellent Excellent Excellent Severe effect Fair Severe effect Severe effect

aCole

Parmer, Vernon Hills, IL. concentration not specified.

bOzone

1. Reaction with metals Ozone oxidizes metals except gold, platinum and iridium to oxides of the metals in their highest oxidation states. Ozone converts oxides to peroxides, sulfides to sulfates, carbon to carbon dioxide, and NH3 to NH4NO3. In most oxidation reactions, ozone is reduced to molecular oxygen. Ozone at high concentrations corrodes equipment, but such high concentrations are found only inside the generator or in the ozone-towater contacting system. Materials used in food processing are usually compatible with ozone at low concentrations (Table III). Stainless steel (e.g. 316L) is corroded less by ozone than by chlorine (Greene et al., 1999; Singh and Singh, 1999). Glass-lined steel (Grosse and Streng, 1960) and steel with a phosphated inside surface (Waller and McTurk, 1965) are

OZONE APPLICATION IN THE FOOD INDUSTRY

185

resistant to ozone and suitable for maximum half-life storage of ozone. According to a recent report (Viera et al., 2000), ozone in aqueous solution at 0.1 and 0.2 ppm, did not affect containers made of stainless steel and titanium, whereas copper alloys were susceptible to corrosion. Metals commonly promote the decomposition of ozone and some of these reactions are catalytic. Good catalysts for ozone decomposition include iron, particularly if rusted, zinc, mercury, platinum and silver (Horvath et al., 1985). Large specific surfaces of absorbents such as activated carbon, molecular sieves, silica gel and activated alumina, strongly catalyze the decomposition of ozone (Mahieux, 1962). Greene et al. (1999) reported that pulsing ozone into water at ambient temperature for 20 min per day for 7 days caused greater weight loss of aluminum, carbon steel, copper, 304 stainless steel and 316 stainless steel samples than that observed in the untreated controls; however, weight loss for carbon steel only was significant (P < 0.05). Severe pitting was noted on ozone-treated copper samples when observed by scanning electron microscope. Black striations were observed on ozone-treated carbon steel surfaces. Brass and copper also should be avoided for concentrations > 1.0 ppm aqueous ozone (Hampson, 2000). In an ozone atmosphere of 25–40 mg m–3, metal surfaces must be protected by appropriate ozone-resistant painting; however, gas mixtures containing ozone ⭐ 5 mg m–3 cause minimal corrosion (Kuprianoff, 1953). The author recommended stainless steel and anodized aluminum for construction of ducts and pipes that carry ozone.

2. Reaction with rubber and plastics Common plastics used in food processing are generally resistant to ozone (Table II); these include polychlorotrifluoroethylene (EFTFE), polydichlorodifluoroethylene (PDFE), polytetrafluoroethylene (PTFE), polyvinylidenefluoride (PVDF), polyvinylchloride (PVC), and silicon tubing and gaskets (Grosse and Streng, 1960; Mahieux, 1962). Some plastic materials (e.g. PVC and polyethylene, PE) are useful in applications where low ozone concentrations are used. Rubber in seals, pipes and other components reacts actively with ozone, leading to a total disintegration into powder. Synthetic rubbers, however, are resistant to ozone (Horvath et al., 1985). Fluorinated hydrocarbon lubricants have a good resistance to ozone, but polymers of monochlorotrifluoroethylene are the most suitable as lubricants. Silicon grease is adequate for short-term use, but it is oxidized on extended exposure to ozone. The sealing materials on doors and windows of fruit storage rooms should be made of ozone-resistant synthetic materials (Kuprianoff, 1953).

186

J.-G. KIM ET AL. V. APPLICATION OF OZONE IN FOOD PROCESSING

Although ozone is highly effective against microorganisms in pure cell suspensions, it is unlikely to be used directly in food containing high ozone-demand materials. Organic constituents of such food compete with microorganisms for ozone, and thus high doses of this agent may be needed for effective elimination of microorganisms. These high levels of ozone may also alter sensory attributes, and adversely affect the acceptability of food. Current ozone applications in the food industry are mostly related to decontamination of processing water. Ozone, however, has been used with mixed success to inactivate microbial contaminants on meat products, eggs and dry food. Ozone is most suitable for treatment of solid food with intact surfaces such as fresh vegetables and fruits.

A. TARGETED MICROORGANISMS

Gaseous and aqueous ozone, at a low dose and with short contact time, is effective against numerous bacteria, molds, yeasts, parasites and viruses (Kim et al., 1999b; Rodgers et al., 1999). The efficacy of ozone as a sanitizer, however, depends on the target microorganism and treatment conditions. Microorganisms inherently vary in sensitivity to ozone. Ozone was tested recently against Gram-positive vegetative cells, bacterial spores, mold conidiospores and yeast ascospores that are commonly isolated from fruits and known to spoil fruit juices. Results show that bacterial spores are the most resistant and bacterial vegetative cells are the most sensitive to ozone (Kim and Yousef, 2000a). Spores of Bacillus spp. varied in susceptibility to ozone (Khadre and Yousef, 2001b). Among eight Bacillus spp. tested, spores of B. stearotherophilus were most resistant, while spores of B. cereus were most sensitive to ozone. The physiological status of the treated microorganism may affect its susceptibility to ozone. Stationary-phase cells are more resistant to ozone than are cells from the exponential phase. The resistance of microorganism to ozone is greater with natural microflora on food than with microorganisms frequently cultured in the laboratory (Kim et al., 1999a). E. T. Ryser (personal communication) compared the efficacy of sanitizers against natural and artificial contaminants on produce. In produce inoculation studies, ozone (3 ppm) treatments for 5 min decreased the microbial population ⭓ 5 log units. In contrast, naturally occurring background populations generally decreased ~ 3 log units after similar treatments. Dormant microbial cells from a dry environment are extremely resistant to gaseous ozone (J. G. Kim and A. E. Yousef, unpublished data). Washing microbial cells

OZONE APPLICATION IN THE FOOD INDUSTRY

187

decreases their resistance to ozone because of the removal of ozonedemand material from the cell surface (Schechter, 1973). Bacteria in biofilms have a polysaccharide architecture that protects them from antiseptics (e.g. hydrogen peroxide and chlorine-releasing preparations) and antibiotics (Dixon, 1998). Microorganisms attached to inert surfaces are less susceptible to the effect of chemical sanitizers than their free-living (planktonic) counterparts (Le Chevaillier et al., 1988). Lethal dose increases when cells have complex envelopes and capsules, specially when cells are in a clump/biofilm. A higher lethal dose is required for spores than for vegetative cells. Existing colonies and clumped cells on the surface of food are hard to destroy because the outer layer of the clump protects inner cells. Mechanical treatment may help break the biofilm structure and increase accessibility of ozone to inner cells (Khadre and Yousef, 2001a). B. INACTIVATION KINETICS AND MECHANISMS

Ozone kills microbial cells rapidly so that inactivation rates are difficult to measure. Difficulties in obtaining meaningful kinetic parameters have been addressed in a recent publication (Kim and Yousef, 2000b). Uniform procedures to measure inactivation kinetics and establish dose–response plots also were discussed. Suitable indicator microorganisms or spores should be used to measure ozone efficacy. Khadre and Yousef (2001b) suggested Bacillus stearothermophilus spore as an indicator of ozone sanitization. A clearer understanding of the mechanism of inactivation is needed to optimize the effectiveness of ozone and supporting technologies. Ozone activity is likely related to its molecular form (Hunt and Marinas, 1997) or intermediate reactive species such as free radicals and singlet oxygen (Kanofsky and Sima, 1991). It appears that ozone causes damage to the following cellular constituents: (1) unsaturated lipids in the microbial cell envelope; (2) the lipopolysaccharides layer of Gram-negative bacteria; (3) intracellular enzymes; and (4) microbial genetic materials. Earlier studies suggest that ozone reacts with microbial cell membranes (Giese and Christenser, 1954). Ozone is believed to cause the oxidation of lipids on the cell envelope of bacteria (Murray et al., 1965; Scott, 1975). Further oxidation leads to leakage of intracellular cell contents, damage of genetic material and death (Prat et al., 1968; Shechter, 1973). Ozone reacts with cell dehydrogenases (Ingram and Haines, 1949), DNA (Scott, 1975) and RNA (Kim et al., 1980). Khadre and Yousef (2001b) examined spores of B. subtilis with the electron microscope after these spores were treated with ozone. The authors observed damage to the outer spore coat layer, which may serve as a primary target of ozone.

188

J.-G. KIM ET AL.

Electron microscopic analysis revealed damage to cellular structures after ozone treatment (Kim, 1998; Dave, 1999; Khadre and Yousef, 2001b). Damage was more pronounced in Gram-negatives, P. fluorescens and E. coli O157:H7, than in Gram-positives, Leuconostoc mesenteroides and Listeria monocytogenes (Kim, 1998). When treated with ozone under similar conditions, Gram-positive bacteria seemed to lose only some mucoid material outside the cell wall, whereas Gram-negative cells tended to collapse and lose cellular components. Therefore, ozone at low concentrations damages the outer membrane of Gram-negative bacteria and thus causes dramatic changes in cell structure. Similar concentrations of ozone cause less visible damage to the cell wall of Gram-positive bacteria, but the agent causes intracellular damage and effectively inactivates these cells (Kim, 1998). Kim (1998) also tested the injury of microorganisms by ozone. The degree of injury varied depending on the microorganism, ozone concentration and exposure time. Maximum injury occurred at ozone concentrations that caused mild lethality. Combined treatment of ozone and pulsed electric field (PEF) resulted in a synergistic lethal effect against E. coli O157:H7 and L. monocytogenes (Unal et al., 2001). It was suggested that the synergy may result from cell injury during the ozone treatment and rapid inactivation of injured cells when they were subsequently treated with PEF. C. OZONE AS AN ALTERNATIVE SANITIZER IN FOOD PROCESSING

The food industry has traditionally used chlorine to limit microbial growth on processing equipment and in wash water. Fresh-cut produce is commonly treated with chlorine to extend product shelf-life and minimize the risk of foodborne pathogens. Despite the benefits of chlorine, chlorination may lead to the formation of toxic or carcinogenic chlorinated organic compounds in water (Brungs, 1973; Page et al., 1976; Kirk and Mitchell, 1980) and food, or on food contact surfaces (Wei et al., 1985). Although considerable chlorine-related research has been done to determine optimal parameters for reducing pathogens and extending product shelf-life, its effectiveness on fruits and vegetables remains variable and unpredictable (Nguyen-the and Carlin, 1994). Some studies suggest that chlorine, at high concentrations, causes only modest inactivation of pathogens on food. When Brussels sprouts were treated with 200 mg L–1 chlorine, the count of L. monocytogenes decreased by only 2 log units (Brackett, 1987). Beuchat and Brackett (1990) also reported that treating shredded lettuce with chlorine did not prevent the growth of L. monocytogenes after the lettuce was packaged in modified atmosphere. Chlorine dioxide treatments only inactivated 1.1 log units of L. monocytogenes population on lettuce and 0.4 log unit on cabbage (Zhang and Farber, 1996). To improve the safety

OZONE APPLICATION IN THE FOOD INDUSTRY

189

of sprouts, treatment of seeds with 20 000 ppm chlorine has been recommended (US-FDA, 1999). Given the drawbacks of chlorination, researchers are actively seeking alternatives to chlorine use in food processing. Chlorine dioxide offers several advantages over chlorine as a sanitizing agent and disinfectant (Dychdala, 1991). Based on availability of oxidative species in solution, the oxidation capacity of chlorine dioxide is 2.5 times greater than that of chlorine. Compared with chlorine, chlorine dioxide is also slower in dissociation/hydrolysis, stable over a broader pH range, less corrosive to metal equipment, and less likely to form chlorinated by-products and potential carcinogens. Chlorine dioxide has been used for many years as a water disinfectant because of its bactericidal qualities and its ability to react with humic substances without the formation of trihalomethanes. Hydrogen peroxide and peracetic acid are potential alternatives to chlorine as a food sanitizer. In a study of various sanitizers for beef brisket adipose tissue, Gorman et al. (1995) found that hydrogen peroxide is one of the most effective sanitizers. While Bundegaard-Nielsen and Nielsen (1996) reported that peracetic acid was not effective as a fungicide, Orth and Mrozek (1989) demonstrated the effectiveness of peracetic acid in reducing numbers of several foodborne bacteria. Ozone is emerging as a viable alternative to chlorine. The strong germicidal action and the high oxidation potential are some of ozone’s properties that make it an attractive alternative to the traditionally used chemical disinfectants. In addition, ozone decomposes rapidly to oxygen and it leaves no toxic residues; this makes ozone a favorable sanitizer for users concerned about the environment. Ozone, compared with chlorine, is a more powerful and efficient antimicrobial agent against spores, fecal and pathogenic microorganisms, and viruses in an environment containing a high proportion of organic matter (Gomella, 1972). Ozone is also effective for pesticide residue reduction (Ong et al., 1996), food preservation, shelflife extension, equipment sterilization and improvement of food plant effluents (Horvath et al., 1985; Hampson, 2000). The superiority of ozone over chlorine compounds was reported by many researchers (Kessel et al., 1943; Scarpino et al., 1972; Korich et al., 1990). Of particular importance to the fruit and vegetable industry is a report indicating that greater than 90% of C. parvum oocysts added to water were inactivated after 5 min of exposure to 1 ppm ozone (Korich et al., 1990). In contrast, approximately 90 min of exposure to 80 ppm chlorine were required to achieve similar results. Ozone is effective in surface decontamination of fresh produce. Ozone was recommended recently as an alternative to chlorine (Kim, 1998) and hydrogen peroxide (Khadre and Yousef, 2001b).

190

J.-G. KIM ET AL. D. FOOD PROPERTIES AND OZONE APPLICABILITY

1. Food composition Commodities with different chemical composition require different ozone dose for effective sanitization. Fresh meat, which contains high fat contents, for example, requires more ozone than do fruits and vegetables, which contain low fat and high carbohydrates. Fournaud and Lauret (1972) treated beef with gaseous ozone during refrigeration and thawing to reduce the surface microorganisms. Gaseous ozone concentrations as high as 500 ppm caused little microbial inactivation. The authors attributed treatment inefficacy to the reaction of ozone with fat and proteins in the meat rather than with the contaminating microorganisms. Kaess and Weidemann (1968b) found that the ozone consumption per unit area of fatty surface tissue was considerably smaller than that of muscle tissue. 2. Surface structure The nature of the surface of food contributes substantially to the efficacy of ozone treatment. Bacteria on poultry carcasses are located primarily on skin surfaces, within feather follicles and on exposed muscle surfaces. Microorganisms located within the feather follicles are generally protected from the bactericidal action of disinfectants, as shown by relatively small reduction in carcass microbial counts during washing and ozonation (Barnes and Impey, 1968). Bullock et al. (1997) used an ozone treatment in a recirculating rainbow trout (Oncorhynchus mykiss) culture system and prevented bacterial gill disease (BGD) outbreaks but found that the causative bacterium, F. branchiophilum, was still colonizing gill tissues. In recent studies, bubbling ozone in wash water for 3 min was effective in reducing microbial counts on the surface of apples by up to 3.7 log units (Klingman and Christy, 2000; Achen and Yousef, 2001). Ozone treatment, however, was less effective in decontaminating the calyx and stem areas of the apple. Spraying water on these areas prior to the ozonation helped dislodge the cells and reduced the counts by 1.5 log CFU g–1. When E. coli O157:H7 was permitted to attach to the apple surface, the efficacy of disinfection by ozone diminished. In conclusion, strongly attached surface microorganisms and those attached to areas that are not freely exposed to ozone cannot be eliminated by mere dipping in ozonated water. In addition, microorganisms embedded in product surfaces are more resistant to ozone than those suspended in water. Therefore, when ozone is applied in food processing, good contact between the sanitizer and the target microorganisms on the treated food

OZONE APPLICATION IN THE FOOD INDUSTRY

191

should be ensured. A variety of methods have been used to accomplish this goal including stirring, pumping, fluming, bubbling, sonication, abrasion and pressure washing. 3. Release of exudates Most fruits and vegetables have a hard protective layer of peel, skin or rind, and the outer surface is usually covered with a waxy material. These products commonly have a limited ozone demand. However, in minimal processing, fresh vegetables and fruits are usually trimmed, peeled (or cut if necessary) and washed, thus tissues are exposed and cellular fluids are released that increase ozone demand. When hot water or steam is used for blanching, vitamins, flavors, colors, carbohydrates and other water-soluble components are inevitably leached (Ihl et al., 1998); this released organic matter constitutes ozone-demand. Treating meat with ozone is a high ozone demand process, since the meat surface has crevices and it leaches ozonedemanding organic substances such as fats and proteins. 4. Injuries and wounds Gaseous ozone treatment was not effective in decreasing infection in inoculated wounds in apples (Schomer and McColloch, 1948). Ogawa et al. (1990) readily inactivated spores of B. cinerea on the surface of noninjured tomato fruit using 3.8 mg L–1 aqueous ozone for 10 min. However, spores placed on the surface of injured tomato fruit were not inactivated. Smock and Watson (1941) reported that when spores were protected by moist surfaces of apple flesh or by other organic protectants, ozone had no effect on their germination. When pear fruits were wound-inoculated with Penicillium expansum and then treated with 5.5 mg L–1 aqueous ozone for 5 min, levels of decay were similar to those of a control treated with water only (Spotts and Cervantes, 1992). This suggests that plant tissues and extracellular biochemicals at wound sites react with ozone and render it ineffective. E. OZONE APPLICATION AT DIFFERENT STAGES OF PROCESSING

Ozone may be applied directly on raw agricultural commodities at the preprocessing stage, during processing, or on the finished product while at storage. It is usually advantageous to apply ozone on the raw than the processed product. Whole grains, for example, require less ozone to disinfect than does the powder product (Naitoh et al., 1987). Raw solid food commonly has intact surface and most of the natural contaminants are limited

192

J.-G. KIM ET AL.

to the surface. Elimination of readily accessible surface contaminants is feasible using most sanitizers including ozone. Sanitizers, however, may differ in dealing with contaminants that are attached or embedded in food surfaces. Aqueous ozone can be used to decontaminate beef and beef brisket fat (Gorman et al., 1995), poultry meat (Dave, 1999), salmon (Goche and Cox, 1999), apples (McLoughlin, 2000; Achen and Yousef, 2001), strawberries (Lyons-Magnus, 1999), lettuce (Kim et al., 1999a), broccoli and cauliflower (broccoflower) (Hampson and Fiori, 1997) and other commodities. Microbial studies typically show a reduction of 2 log total count and a significant reduction of spoilage and potentially pathogenic species that are most commonly associated with fresh fruits and vegetables. When these raw foods carry high organic loads, effectiveness of ozone treatment is likely to diminish. Multiple-stage wash system may be needed in this case with a pre-wash and a rinse preceding the ozone treatment. Some researchers used ozone to treat ingredients before they are included into food formulation. Kim et al. (1993) treated various spices, used to prepare Kimchi, with ozone and improved the fermentation of the final product. J. G. Kim and A. E. Yousef (unpublished) also used ozone to decontaminate the ingredients of fruit juices such as high-fructose corn syrup. These researchers speculated that ozone treatment of ingredients, rather than the final juice, reduces ozone usage and minimizes damage to the sensory quality of the final product. Naitoh et al. (1989) reported that the treatment of wheat flour with ozone inhibited microbial growth in namamen products and increased their storage life. Application of ozone in the processing facility to minimize environmental contaminants has been studied (Naitoh, 1993; Hampson, 2000). Combinations of ozone with other oxidants such as hydrogen peroxide were used to sanitize packaging films (Gardner and Sharma, 1998), a confectionary plant (Naitoh, 1989), and hatchery equipment (Whistler and Sheldon, 1989). Ozone decreased surface flora by ~ 3 log units when tested in wineries for barrel cleaning, tank sanitation, and clean-in-place (CIP) operations (Hampson, 2000). Some wineries have embraced the use of ozone for multiple purposes, including barrel and tank cleaning and sanitation, CIP systems, and for general purpose sanitation (Duca, 1999; Hampson, 2000). Aqueous ozone, at 1.5 ppm, was tested in a food-processing facility. The treatment decreased the microbial load on a stainless steel kettle, table top, shroud, plastic shipping container, floor surface and drain (Hampson, 2000). The author reported reductions of up to 3 log total plate count on floor surfaces and shrouds, but less than 1 log unit in floor drains. The author recommended using ozone in a complementary sanitizing regime to maintain the overall cleanliness and sanitation of wineries and any other food-processing facility.

OZONE APPLICATION IN THE FOOD INDUSTRY

193

Gaseous ozone can be used to extend the shelf-life of food during storage. Ozone minimized growth of surface contaminants on meat (Greer and Jones, 1989), grapes (Sarig et al., 1996) and broccoli florets (Zhuang et al., 1996) when the gas was applied during storage of these commodities.

VI. SELECTED FOOD APPLICATIONS A. RAW POULTRY AND MEATS

Count and diversity of the microbial population dictates the shelf-life of raw poultry and meats. In addition to numerous spoilage microorganisms, these products occasionally carry pathogenic microorganisms such as Campylobacter spp., Salmonella spp., L. monocytogenes and pathogenic E. coli. The types and number of microorganisms present on raw poultry and meats depend on the microorganisms that colonize the gastrointestinal tract. Food contamination with these microorganisms can occur at multiple steps along the food chain including slaughtering, handling, storage and distribution (Zhao et al., 2001). Use of sanitizers on carcasses and cut meat is currently limited. Chlorine may be used in the poultry chiller tanks or as spray on carcasses, but alternative and effective sanitizers have been pursued. Several investigators tested decontamination of beef and beef brisket fat by ozone; results were variable (Gorman et al., 1995, 1997; Reagan et al., 1996). Other research groups found that gaseous ozone minimized or prevented growth of microorganisms on the meat surface (Kaess and Weidemann, 1968a; Greer and Jones, 1989; Mitsuda et al., 1990). Numerous investigators have demonstrated the microbicidal efficacy and safety of ozone for use in washing poultry carcasses (Barnes and Impey, 1968; Yang and Chen, 1979; Sheldon and Brown, 1986; Caracciolo, 1990; Izat et al., 1990; Jindal et al., 1995), reconditioning poultry chiller water (Sheldon, 1986; Sheldon and Chang, 1987; Waldroup et al., 1993; Diaz and Law, 1999), and sanitizing hatchery equipments (Whistler and Sheldon, 1989). Advanced oxidation processes, including ozone and adjuncts such as hydrogen peroxide and UV radiation, enhanced the efficacy of ozone as an antimicrobial control agent in poultry chiller water (Diaz and Law, 1999; EPRI, 1999). Prefiltration of the chiller water prior to ozone treatment is recommended for optimum reduction of microbiological levels and efficient use of ozone (Sheldon, 1986; EPRI, 1999). The presence of Salmonella enterica ser. Enteritidis in shell eggs has serious public health implications. Several thermal and chemical treatments have been developed to control or eliminate this pathogen in

194

J.-G. KIM ET AL.

eggs. Yousef and Rodriguez-Romo (2001) used ozone at low temperatures and under mild pressure for cold sanitization of shell-eggs. Shell-eggs were externally contaminated with S. Enteritidis so that shells contained ~ 106 CFU g–1. Eggs were treated with gaseous or liquid ozone for 1–20 min, at 4–25°C and 0–15 psi (0–100 kPa). Gaseous ozone treatment without pressure decreased the count of S. Enteritidis on shells by 2.2–2.7 log units. Treating contaminated eggs with gaseous ozone for 10 min at 22–25°C and 15 psi decreased Salmonella population by more than 5 log units. Such a treatment may be used industrially to produce “cold-sanitized” eggs. Cox et al. (1995) patented a “hyperpasteurization” process, which uses vacuum, heat and ozone, to eliminate Salmonella spp. from the contents of shell eggs. This method includes, heating shell eggs at higher than 54.4°C for longer than 15 min with subsequent application of ozone. According to this report, the combined treatment extended the shelf-life and reduced the microbial load of shell eggs. B. FRUITS AND VEGETABLES

There are many steps in the food production chain with multiple potential sources of contamination at each step. For example, dirty irrigation water, manure fertilizer, and improper worker hygiene have been cited as probable causes for pre-harvest contamination of fresh fruits and vegetables. Following harvest, improper handling and storage, use of contaminated wash water, processing equipment and transportation facilities as well as cross-contamination from other produce contribute to the microbial hazards associated with fresh fruits and vegetables. Of special concern is the quality of wash water and the potential hazard associated with cross contamination (Tauxe et al., 1997). Following cutting, shredding and slicing of fresh-cut fruits and vegetables, the loss of surface integrity can lead to penetration and rapid growth of microorganisms. In a study on processing conditions of chopped lettuce and coleslaw, shredders were identified as the major source of in-plant contamination (Garg et al., 1990). Other contributing factors include the expansion of production facilities and longer food marketing chains that allow the distribution of more heavily contaminated produce to wider populations (Tauxe et al., 1997). Refrigerated fresh-cut produce is susceptible to microbial spoilage (Nguyen-the and Carlin, 1994) and growth of pathogenic microorganisms (Beuchat, 1995). Ozone application for improving the safety and extending the shelf-life of fresh-cut produce seems feasible. Improving water reusability by the fruit and vegetable processing industry is an additional benefit of ozone application. Ozone-sanitized produce has been introduced recently in the United States market after several years of developing and testing (TVA,

OZONE APPLICATION IN THE FOOD INDUSTRY

195

2001). The following are examples of studies and successful attempts to sanitize fresh produce with ozone. 1. Apples Apples are subject to contamination with pathogenic microorganism on the farm. Use of contaminated apples to produce unpasteurized apple juice and cider resulted in E. coli O157:H7 foodborne infections (Besser et al., 1993; CDC, 1996b). The Food and Drug Administration (FDA) now regulates the production of cider with recommendations for the pasteurization of apple cider and other juice products or the use of alternative processing steps to reduce the counts of the pathogen in question by 5 log units (FDA, 1998). Use of effective sanitizers on whole apples prior to pressing is a feasible option. Chlorine and hydrogen peroxide with surfactants and isothiocyanate have been investigated as sanitizers (Beuchat et al., 1998; Lin et al., 2000). In a recent study, Achen and Yousef (2001) inoculated apples with E. coli O157:H7 prior to treatment with an aqueous solution containing 21–28 mg L–1 ozone. Decontamination treatments were more effective when ozone was bubbled during apple washing than by dipping apples in pre-ozonated water. Maximum decreases in surface counts of E. coli O157:H7 were 2.6–3.7 log units, compared to unwashed inoculated controls. However, counts of E. coli O157:H7 decreased by less than 1 log unit in the stem-calyx region by the ozone treatment. Rinsing the apples with an inorganic wetting agent (tetrasodium pyrophosphate) increased the efficacy of the ozonation process. The wetting agent may have enhanced the contact between ozone and bacterial cells that are attached to the hydrophobic surface of the apple, decreased cell attachment on the stem and calyx areas, and assisted in exposing entrapped cells to ozone. The authors speculated that in conventional apple washing environments, the efficacy of ozone against microbial contaminants may become limited because of the high organic loads in the washing tanks resulting from debris, soils, and fruit saps; these contaminants impose an ozone demand. 2. Lettuce Kim et al. (1999a) tested ozone against natural contaminants in fresh lettuce and results were compared with those obtained from chlorine treatment. Ozone (1.3 mM) or chlorine (1 mM) inactivated mesophilic and psychrotrophic bacteria by 1.4 and 1.8 log units in 3 min, respectively. Counts of these microorganisms on lettuce, from a different batch, decreased 3.9 and 4.6 log units, respectively, during 5 min of ozone treatment. In another

196

J.-G. KIM ET AL.

experiment, shredded lettuce was treated with gaseous ozone, or mixed with aqueous solution of ozone (1 : 20 w/w) with or without bubbles. Results show that bubbling gaseous ozone in water, combined with stirring, is the most effective ozonation method for shredded lettuce. 3. Alfalfa sprouts Alfalfa sprouts received great attention in recent years owing to the incidence of pathogens in this product and associated disease outbreaks. In 1998, the FDA issued a statement warning consumers of high-risk groups to avoid consumption of sprouts due to the potential health hazard associated with these products. J. M. Boff and A. E. Yousef (unpublished data) determined the effectiveness of ozone in reducing the natural flora on alfalfa sprouts. Treatment of alfalfa sprouts with ozone for 5 min decreased their microbial load by 1.2 log units. An additional 0.8 log unit decrease in population was observed when ozonation was accompanied with agitation. As an alternative ozonation process, alfalfa seeds were pretreated with ozone and the microbial count during seed germination and growth was monitored. Additionally, ozonated water was used to water the sprouts twice daily during the growth period. The initial microbial count on the seeds was ~ 105 CFU g–1 and ozone treatment decreased the count by ~ 2 log units. The difference in count between ozone-treated and nontreated seeds diminished during the growth period and the population in both treatments reached ~ 109 CFU g–1 after 4 days of seed incubation. Most of the growth of the natural flora occurred in the first 2 days after setting. Treatment with ozone water during the growth period only temporarily decreased the count, but the counts after 4 days of growth were identical in ozone-treated and nontreated sprouts. An end treatment consisting of bubbling ozone into the sprouts for 2.5 min decreased the microbial count from 5.0 × 109 to 2.0 × 107 CFU g–1. Alternatively, the sprouts were placed in ozonated water (30–32 ppm) and stirred for 20 min; the average count decreased to 4.8 × 107 CFU g–1. The latter method may be preferable to the former one because of better quality and texture of the resulting sprouts. In conclusion, ozone treatments, as tested in this study, are not sufficient to bring about a substantial reduction in the microbial population on sprouts. 4. Fruit juice ingredients Technological advances in citrus processing led to development of convenient, shelf-stable, ready-to-drink juices. Survival of heat-resistant bacteria and fungi during pasteurization of juice can be a serious concern

OZONE APPLICATION IN THE FOOD INDUSTRY

197

to citrus processors. Ozone is potentially useful in inactivating heat-resistant spores in juice components with low ozone demand. Therefore, the efficacy of ozone against spores of Alicyclobacillus acidocaldarius, Neosartorya fischeri and Zygosaccharomyces bailii, which are known to be problematic in juices, was tested in selected juice ingredients (Kim and Yousef, 2001). Fruit juice components, high fructose corn syrup (HFCS), orange juice concentrate (OJC), and pine apple juice concentrate (PJC), spiked with cells or spores (c. 107 mL–1), were treated with gaseous ozone. The sensitivity of spores to ozone varied depending on the type of spores and the juice component. The ozone dose required for inactivating 5 log units of A. acidocaldarius spores was 0.31, 0.28 and 0.41 mg ozone per mL spore suspension in HFCS, OJC and PJC, respectively. The ozone dose for inactivating 5 log units of N. fischeri spores was 0.12–0.51 mg mL–1, depending on the juice ingredient. The amount of ozone required to inactivate N. fischeri spores was four times greater for PJC than for HFCS. In order to achieve a decrease of 5 log units of Z. bailii spores, 0.04–0.24 mg mL–1 was needed, depending on the juice component. Inactivation of Z. bailii spores by ozone was faster in HFCS than in other juice components even though HFCS has higher solids than the other components. In conclusion, the ozone dose required to achieve a 5-log unit decrease of the targeted microorganisms varied from 0.04 to 0.5 mg mL–1 with A. acidocaldarius, and N. fischeri spores were three to four times more resistant to ozone than were Z. bailii spores. The authors do not recommend applying ozone directly to juice concentrates or reconstituted juice. Ozone, however, may be safely applied to the HFCS. Ozonation can also be combined with different inactivation technologies to minimize the changes in product quality while maximizing the inactivation of contaminants. C. FISH PROCESSING AND STORAGE

Ozone was tested for decontaminating shrimp (DeWitt et al., 1984), mussels (Abad et al., 1997), and various fish such as jack mackerel (Haraguchi et al., 1969), sockeye salmon (Lee and Kramer, 1984), Japanese flounder (Mimura et al., 1998) and rockfish (Kötters et al., 1997). The antimicrobial efficacy of ozone was equal to or better than that of chlorine in some applications studied (Arimoto et al., 1996; Goche and Cox, 1999). Ozone reduced disease incidence in hatcheries with less mortality and shorter growth cycles (Blogoslawski et al., 1993; Arimoto et al., 1996; Bullock et al., 1997). The absence of adverse sensory effects and harmful oxidation by-products confirms the desirability of ozone use in processing fish products for human consumption. Ozone treatment of shrimp meat extract, however, was ineffective against microorganisms in the product (Chen et

198

J.-G. KIM ET AL.

al., 1992). Ozone may have reacted with ozone-demand substances in the extract, instead of microbial cells. D. DRY FOOD AND FOOD INGREDIENTS

Gaseous ozone can eliminate Bacillus spp. and Micrococcus spp., which are dominant in cereal grains, peas, beans and spices, by up to 3 log units, depending on concentration, temperature and relative humidity (Naitoh et al., 1988). The surface area of the dry food is an important factor for the effectiveness of ozone treatment. Cereal flour and ground pepper, for example, require higher concentration of ozone and longer contact time than whole cereal and pepper to achieve the same degree of microbial inactivation (Naitoh et al., 1989; Zagon et al., 1992). In addition to antimicrobial effects, ozone destroys or greatly reduces aflatoxins from peanut (Dollear et al., 1968) and cottonseed meal (Rayner et al., 1971) and oxidizes odors produced during dehydration of onions and garlic (McGowan et al., 1979). Ozone treatment may cause lipid oxidation (Naitoh et al., 1988), decrease amino acid (e.g. thiamine) content (Naitoh et al., 1989) and essential oil content (Zhao and Cranston, 1995), and contribute negative effects on the sensory quality of some dry food. J. G. Kim and A. E. Yousef (Table II, unpublished data) used ozone to inactivate natural contaminants in a silica-based anticaking agent. High water activity with limited organic matter provided suitable conditions for growth of bacteria and fungi in this product. Ozone gas at 30–40 ppm was sufficient to sterilize an anticaking agent but failed to decontaminate other types of the substance. The authors hypothesized that water activity is an important factor for cell inactivation by ozone in the anticaking agent. E. PACKAGING MATERIAL AND FOOD CONTACT SURFACES

Microbial contaminants on food packaging materials are commonly small in number, but they may survive conventional decontamination processes and cause food spoilage. Currently, sterility of packaging materials is achieved by several methods including heat, hydrogen peroxide and UV radiation (Stefanovic and Dickerson 1986; Yokoyama 1990; Gardner and Sharma 1998). Sterilization by hydrogen peroxide is a tedious and variable method (Yokoyama 1990), and unacceptable levels of hydrogen peroxide residues may remain and interact with some of the polymer in the packaging material (Stefanovic and Dickerson 1986; Castle et al. 1995). Sanitization of equipment and food-contact surfaces is essential for safety and quality of processed food. When biofilms develop on wet food contact surfaces such as those made from stainless steel, the micro-

OZONE APPLICATION IN THE FOOD INDUSTRY

199

organisms inside the biofilms are usually protected from sanitizers (Frank and Koffi, 1990; Carpentier and Cerf, 1993; Dixon, 1998). Hence, alternative methods for decontamination of packaging materials and stainless steel are being sought. A multilaminated aseptic food packaging material and stainless steel were treated with ozone to inactivate natural contaminants, bacterial biofilms (P. fluorescens) and dried films of B. subtilis spores (Khadre and Yousef, 2001a). Sterility of the multilaminated packaging material, which contained a low natural, mostly mesophilic, contaminants was achieved when it was treated with 5.9 mg L–1 ozone in water for 1 min. Counts of bacteria in dried films decreased by 4.6–5.1 log units for the multilaminated packaging material and 5.5 log units for stainless steel when treated with 5 mg L–1 aqueous ozone for 1 min. Dried films of spores (108 per 6.3 cm2 surface) decreased below detection (< 10 spores per 6.3 cm2 surface) by application of 13 mg L–1 aqueous ozone for the multilaminated packaging material and 8 mg L–1 in the case of the stainless steel. Repeated exposure to ozone and agitation during the treatment were needed to decrease effectively the population in biofilms. Ozone inactivated P. fluorescens in biofilms more effectively on stainless steel than on the multilaminated packaging material. The relatively low efficiency of ozone against bacteria in the biofilm, compared to that in dried films, is probably due to the tenacious adherence of bacteria to the surface of packaging material when biofilms are formed. It is concluded that ozone is an effective sanitizer with potential applications in the decontamination of packaging materials and equipment food-contact surfaces. Removal of biofilms by ozone, however, requires additional mechanical action during the treatment. F. PESTICIDES ON AGRICULTURAL COMMODITIES

Use of pesticides on fruits and vegetables is crucial for insuring highquality products. Residues of these pesticides, however, raise the concern of consumers and effective means to remove these residues are sought. The use of chlorine, ozone, chlorine dioxide and peracetic acids as postharvest treatments has been effective in remediating several different pesticides on apples (Ong et al., 1996; Siler, 1998; Hwang, 1999). These same treatments also have the added benefit of reducing microbial populations on the surface of fruits and vegetables by up to 5 log units (Rodgers et al., 1999). Organic food processors and consumers of organic food also are concerned about the presence of chemical residues, including chlorine and chlorinated by-products. However, these residue levels are likely reduced by ozonation in solution.

200

J.-G. KIM ET AL. VII. COMBINATION TREATMENTS

Molecular ozone, or free radicals resulting from its degradation, interact with pollutants or microorganisms and cause their destruction. Direct reaction of ozone with organic compounds is generally relatively slow. Therefore, most ozone reactivity is associated with the radical-chain reaction rather than with the direct reaction with solutes (Hoigné and Bader, 1979). Typical rate constants for reactions of the hydroxyl radical with organic compounds are in the range 108 to 1010 M–1 s–1 (Farhataziz and Ross, 1977). Thus, microorganisms are inactivated faster in the presence of radicals formed during ozone decomposition than by ozone molecules themselves because of the higher reaction rates of the former. In order to improve ozone action, free radical generation and the combination of ozonation with other technologies have been studied. Advanced oxidation process techniques are designed to promote the formation of hydroxyl free radicals, resulting in increased microbicidal activity above that of ozone itself. Effective ozonation procedures are also being developed through improved delivery systems in order to overcome the physical barriers that diminish the efficacy of sanitization of products, and to maximize the biocidal action of ozone.

A. OZONE AND HYDROGEN PEROXIDE

The combination of ozone and hydrogen peroxide in aqueous solution generates hydroxyl free radicals (Figure 6). Combinations are obtained by adding the necessary amount of hydrogen peroxide to the water being treated and then passing the solution through an ozone-contacting apparatus (Graham, 2000). Hydrogen peroxide is a weak acid, which partially dissociates into hydroperoxide ion in aqueous solutions. H2O2 + H2O → HO2– + H3O+ O3 + HO2–

→ ˙OH + ˙O2– + O2

The hydrogen peroxide molecule reacts slowly with ozone, whereas the hydroperoxide anion is highly reactive (Taube and Bray, 1940). Consequently, the rate of ozone decomposition by hydrogen peroxide increases with increasing pH. The reaction rate of ozone decomposition with H2O2 is theoretically 100 000 times greater than that of ozone decomposition initiated by hydroxyl ions. Therefore, very low concentrations of H2O2 are kinetically effective in initiating O3 decomposition.

OZONE APPLICATION IN THE FOOD INDUSTRY

201

For optimum oxidative performance, a specific weight ratio of peroxide to ozone is required to destroy each pollutant. For example, the taste- and odor-causing compounds, geosmin and 2-methylisoborneol in potable water, require a peroxide to ozone ratio of about 1 : 0.3. Pesticides sometimes require weight ratios as high as 1 : 0.8. Therefore, it is advisable to determine experimentally the optimum peroxide to ozone ratio that is most suitable for a given application (Graham, 2000). Glaze et al. (1987) found that the rate of oxidation of organic load in water decreases when the weight ratio of peroxide to ozone is greater than one. The Electric Power Research Institute (EPRI, 1999) evaluated the efficacy of ozone and hydrogen peroxide combinations in poultry chiller operation. When a broiler was rinsed with chiller overflow ultrafiltrate containing 1–2 mg L–1 ozone and 0.5 mg L–1 hydrogen peroxide, the average decrease in total count was more than 2 log units. B. OZONE AND CHLORINE

Processors of fresh-cut produce maintain a minimum chlorine concentration in wash water to ensure the efficacy of the treatment. Therefore, chlorine residues are monitored as a critical control point in these processes. Processors who look for ozone as an alternative to chlorine treatments are concerned that the efficacy of sanitization cannot be similarly measured, since ozone treatments may not result in measurable residues. The potential interaction of ozone and chlorine should be considered when combining these two sanitizers in a processing line. Ozone oxidizes residual chlorine in water to form chlorate and perchlorate, which are weaker oxidants (Kolle, 1968; Siddiqui, 1996). However, Buydens and Fransolet (1971) did not notice any interaction of ozone and chlorine. Therefore, sequential application of ozone and chlorine is potentially a useful combination treatment in such facilities. C. OZONE AND OTHER GASES

Tahoe Food Technology, Inc. (1998) invented an apparatus that produces a continuous stream of mixed gas containing ozone, carbon dioxide and argon. Navel oranges were treated with this gas mixture for 2 h in a sealed chamber to control bean thrips, red scale and fuller rose beetle at 20°C, 30% RH and 9.5 psi (66 kPa). The concentrations of ozone, carbon dioxide and argon were 4.0, 10.0 and 1.0%, respectively. All adults, larvae and eggs (fuller rose beetle only) were killed in the process. The treated naval oranges were incubated for 28 days after treatment to ensure that all three life cycles had been destroyed.

202

J.-G. KIM ET AL.

Mitsuda et al. (1990) tested the decontamination of fresh cucumbers with a mixture of ozone and carbon dioxide gases in polyvinylchloride film bags. The concentration of ozone gas generated was 20–40 g m–3. The food in each bag was exposed for 5 min to ozone, carbon dioxide, and their mixture. The best results were obtained when mixing ozone and carbon dioxide at 3 : 1 and 2 : 1 (v/v). After 14 days storage, ~ 1 log unit decrease of bacteria was obtained with 2 : 1 and 1 : 1 mixtures. The survival of microorganisms during storage was lower for the mixed gases of ozone and carbon dioxide than it was for the individual gases. The authors reasoned that this synergistic action was a result of the quenching effect of carbon dioxide on the chain decomposition reaction of ozone, and by the bacteriostatic effect of carbon dioxide. D. OZONE AND HEAT

Ozone degrades rapidly with heat; therefore, simultaneous application of these two preservation factors is ineffective against food microflora. Sequential application of ozone and heat, however, may be beneficial. J. G. Kim and A. E. Yousef (unpublished data) applied sublethal concentrations of ozone to B. subtilis spores and then measured thermal inactivation of pre-treated spores. Results (Table IV) show that ozone treatment greatly sensitized bacterial spores to heat. The D-value was decreased considerably by the sublethal ozone pre-treatment. Khadre and Yousef (2001b) examined ozone-treated bacterial spores using the transmission electron microscope. The authors found that treatment of spores with aqueous ozone (10 mg L–1) for 1 min caused substantial damage to the outer coat layer. Therefore, weakening of the outer coats by sublethal ozone concentrations may have sensitized B. subtilis spores to heat. E. OZONE AND ULTRAVIOLET RADIATION

An advanced oxidation process, based on a combination of ozone and UV radiation, enhances the antimicrobial action of ozone (Diaz and Law, TABLE IV D-VALUES (MIN) OF BACILLUS SUBTILIS SPORES AFTER PRETREATMENT WITH 6 PPM OZONE SOLUTION FOR 1 MIN AND SUBSEQUENTY HEATING AT 85–95°C

Temperature Treatment

85°C

90°C

95°C

Control Ozone

294 26.3

74.6 9.3

27.0 4.0

OZONE APPLICATION IN THE FOOD INDUSTRY

203

1999; EPRI, 1999). Ozone has a maximum absorption for UV radiation at 253.7 nm. In a gaseous phase enriched with water vapor, photolysis of ozone involves the release of a molecule of oxygen and an atom of oxygen (McGrath and Norrish, 1960); the latter may react with water to produce hydroxyl radicals. Treatment of water by ozone and UV combination is achieved by placing a UV bulb in the ozone-contacting chamber (Graham, 2000). As water flows through the chamber, the UV bulb is turned on to initiate the decomposition of ozone molecules to hydroxyl radicals. Kruithof and Kamp (1999) applied advanced oxidation processes in drinking waters spiked with pesticides or with clostridia spores. Ozone/ peroxide and UV/peroxide combinations rapidly destroyed the pesticides tested; however, ozone treatment alone was not effective. Clostridia spores (2.2 × 104 spores mL–1) were resistant to ozone/peroxide and ozone/UV combination treatments, but, spore count decreased to below detection level with the UV/peroxide treatment. Diaz and Law (1999) evaluated UV-ozonation system for the treatment of unscreened overflow poultry chiller water samples and obtained a more than 60% decrease in total plate count, coliforms and E. coli, and more than 80% of the light transmission of fresh water. Synergy between ozone and UV radiation accounted for > 0.8 log unit decrease in total plate count. A synergistic bactericidal effect also was reported between ozone and UV radiation when poultry overflow chiller water, inoculated with nalidixic acid-resistant S. typhimurium, was treated with the combination for 4–8 min.

F. OZONE AND PULSED ELECTRIC FIELD

Unal et al. (2001) explored the potential enhancement of inactivation of Lactobacillus leichmannii, E. coli O157:H7 and L. monocytogenes by use of a pulsed electric field (PEF) when they are pretreated with ozone. The authors found a synergistic bactericidal effect. The E. coli count decreased by 3.6 log units and the L. monocytogenes count by 3.9 log units when these bacteria were treated sequentially with 0.75 mg ozone mL–1 and 15 kV cm–1. However, ozone at 0.75 mg L–1 inactivated E. coli and L. monocytogenes by 1.8 and 3.0 log units, and PEF at 15 kV cm–1 inactivated the microorganisms by 1.8 and 0.8 log units, respectively. The synergy was more apparent at mild than severe doses of ozone, and when the combination treatment was applied to Lb. leichmannii as opposed to E. coli or L. monocytogenes. Oshima et al. (1997) treated E. coli with combinations of PEF and ozone. They reported that simultaneous application of PEF and ozone synergistically inactivated E. coli. Their

204

J.-G. KIM ET AL.

data, however, show that ozone and PEF combinations had an additive rather than a synergistic effect. VIII. ANALYTICAL METHODS

Ozone can be produced by electric discharge, photochemical, chemical, thermal, chemonuclear, and electrolytic methods (Horvath et al., 1985). The corona discharge method is commonly used to produce large amounts of ozone but the UV-based methods generate smaller yield and concentrations of the gas. The corona discharge produces ozone when a highvoltage alternating current is applied across a discharge gap in the presence of oxygen or air (Kim et al., 1999b). There are physical, physicochemical, and chemical methods for the measurement of ozone. Most of the ozone analytical methods are modifications of chlorine residual methods, which are based on determining the total oxidation in solution. Physical methods measure direct absorption in the UV, visible or infrared region of the spectrum. Physicochemical methods are dependent upon reaction outputs such as heat or chemiluminescence. Chemical methods quantitate products released from the reaction between ozone and a chemical reagent such as potassium iodide. Determination of residual ozone in aqueous solutions is quite difficult because of the rapid decomposition of ozone, volatility from solution, and reactivity with many organic and inorganic chemicals. The iodometric method has been widely used (Gordon and Grunwell, 1983). This method, however, measures ozone and other oxidizing species present in an ozone reaction solution. In addition, several factors, including pH, buffer composition, buffer concentration, iodide ion concentration, sampling techniques and reaction time, affect the accuracy of the iodometric method. Hence, measurement of residual ozone by the iodometric method is not recommended. The indigo method currently is widely used for determining residual ozone in water and waste water (APHA, 1998); it is relatively sensitive, precise and fast, with a detection limit of 0.005 mg L–1 ozone (Bader and Hoigné, 1981). Indigo has a relatively high molar absorptivity (~ 20 000 mol–1 cm–1) and the dye absorbs light at 600–610 nm. Ozone reacts selectively with the carbon–carbon double bond of the sulfonated indigo molecule; therefore, ozone measurement by this method is not affected by the presence of hydrogen peroxide, organic peroxides, manganous ions and oxidized species in the aqueous medium. For gaseous ozone measurements, a UV spectrophotometric method, which is based on UV absorbance at 258 nm and a molar absorptivity of 2900 mol–1 cm–1, is most suitable.

OZONE APPLICATION IN THE FOOD INDUSTRY

205

Ozone concentration (mg/L)

16.00

12.00

O3 (Indigo) O3 (UV) O3 (Monitor)

8.00

4.00

0.00 0.00

4.00

8.00

12.00

16.00

Time (min) FIG. 8. Monitoring ozonation of water for 14 min using the indigo-based chemical method, a UV-spectrophotometric method, and a commercial ozone analyzer (Monitor).

In-line measurements of ozone are based on spectrophotometric, calorimetric or potentiometric methods. Several manufacturers produce instruments to monitor ozone concentration based on UV absorption. Calorimetric methods of ozone measurement depend upon decomposition of ozone in the presence of a catalyst. Residual ozone in water can also be measured by amperometric or potentiometric methods. The latter methods depend on the oxidation–reduction potential of ozone. In our laboratory, we compared ozone measurements using a commercial ozone analyzer (Ebara Jitsugyo Co., Ltd., Tokyo, Japan), the indigo-based chemical method (Hoigné and Bader, 1986) and a UV spectrophotometric continuous procedure (Kim and Yousef, 2000b). Results (Figure 8) show that ozone measurements by the commercial analyzer were consistently smaller, compared with those from the other methods (unpublished data). The length of the tube between the ozone reservoir and the analyzer seems to affect the discrepancy between the results. In addition, the length of the tube between the ozone reservoir and the UV spectrophotometer also cause variations in ozone concentration readings by this instrument. Since ozone is such a dynamic oxidizing agent, there should be a minimal time lag between collection of a sample and analysis. Consideration should also be given to the ozone demand of materials in contact with the sanitizer, e.g., the holding and handling apparatus.

206

J.-G. KIM ET AL. IX. REGULATORY STATUS

Ozone use is permitted in many European and Asian countries. Ozone has been safely and effectively used in water treatment for several decades in European countries (Bryant et al., 1992). Industrial use of ozone in the United States was limited to the removal of metal ions, color, taste and odors in water (O’Donovan, 1965). In 1975, the US-FDA permitted the use of gaseous ozone up to 0.1 ppm in meat-aging coolers and in 1982, ozone was approved as a Generally Recognized As Safe (GRAS) substance for treatment of bottled water (USDA, 1982). According to the 1982 ruling, other uses of ozone (e.g. in food) require a food additive petition. Subsequently, the United States Department of Agriculture (USDA) permitted the use of ozone in poultry chiller water (USDA, 1984). The Electric Power Research Institute and Agriculture and Food Technology Alliance submitted a direct food additive petition to FDA so that ozone could be used in food processing without limitations (Graham, 2000). In response to the petition, the FDA amended the food additive regulations to provide for the safe use of ozone in gaseous and aqueous phases as an antimicrobial agent on food, including meat and poultry (Federal Register, 2001). This approval should boost a broad use of ozone in food processing. X. LIMITATIONS, TOXICITY AND SAFETY

The reactivity of ozone and the potential deterioration in the quality of treated product may limit uses of this sanitizer in food processing. Sensory attributes may be altered, depending on the chemical composition of food, ozone dose and treatment conditions. Surface oxidation of food by ozone results in discoloration, undesirable odors and oxidative spoilage. Ozone decreased vitamins and amino acid contents and increased lipid oxidation and activity of some enzymes such as superoxide dismutase, ascorbate peroxidase, and glutathione reductase in lettuce leaves (Kang et al., 1999). The acute and chronic effects of excessive exposure to ozone were investigated (Stockinger, 1965). Small concentrations of ozone gas in air (0.3 ppm) may cause discomfort to susceptible people. Scott and Lesher (1963) reported that as little as 0.02–0.04 mg L–1 can be detected by humans and 0.1 mg L–1 is objectionable to all normal humans because of irritation to the nose, throat and eyes. The respiratory tract is the primary site of ozone toxicity, where the gas may cause pulmonary congestion. Symptoms resulting from exposure to ozone include headaches and dryness of the throat, nose and eyes (Stockinger, 1965; Mustafa et al., 1980). Some researchers demonstrated that repeated exposures have progressively

OZONE APPLICATION IN THE FOOD INDUSTRY

207

lesser effects, suggesting that tolerance may develop (Nadel, 1979). Thorp (1950) indicated that with an hour exposure symptomatic, irritant, toxic and irreversible lethal effects can be induced by ozone concentrations of 2, 4, 15 and 95 ppm, respectively. Menzel (1984) postulated that ozone may generate toxic substance when tissue proteins, unsaturated fatty acids or other components of food are oxidized. However, immersion of shrimp meat in saline containing 5 mg L–1 ozone, for 120 min, did not generate mutagens in the product (Chen et al., 1992). In a recent direct food additive petition, Graham (2000) argued that ozone can be used comparatively safely in industrial applications. According to the author, ozone is detected by human olfactory senses at concentrations as low as 0.01 ppm, and higher concentrations of the agent exert only temporary acute symptoms in humans. Since ozone has a high oxidation potential, disinfection can be accomplished with less concentration and shorter exposure time, compared to other oxidizing agents. Ozone is manufactured on-site at relatively low concentrations and pressures, therefore, an uncontrolled, widespread, and immediate release of large quantities of ozone is not possible. Graham (2000) also indicated that the relatively short half-life of ozone minimizes the persistence of the gas in the environment, and ozone breakdown product (diatomic oxygen) cause no harm. When ozone is generated and used in food applications, precautions and personal safety always should be observed. Dissolving ozone in water is commonly accompanied by excess undissolved gas that remains entrapped in the solution (Hampson, 2000). Excess ozone should be degassed or separated from the water stream prior to delivery to equipment or the processing environment. Ozone detection and destruction systems and respirators are needed for the safety of workers in food processing facilities. Good manufacturing practice (GMP) and hazard analysis and critical control point (HACCP) systems are needed to control high ozone demand materials in food processing; this helps optimize ozone use in the processing facility. Workplace monitoring for ozone off-gas should be performed, and records should be maintained to ensure compliance with regulation. Pryor and Rice (2000) discussed ozone exposure threshold limits. In the United States, current permissible exposure level-time weighted average (PEL-TWA) for ozone exposure in the work place environment is 0.1 ppm, as recommended by the American Conference of Governmental Industrial Hygienists (ACGIH, 1986) and adopted by the United States Occupational Safety and Health Administration (OSHA). Susceptible individuals can be exposed continually to this ozone concentration during a normal 8-h day/40-h working-week without adverse effects (CFR, 1997). The short-

208

J.-G. KIM ET AL.

term exposure limit is 0.3 ppm for an exposure less than 15 min and four times per day.

ACKNOWLEDGEMENTS

Funds for compiling this article were provided by the Center for Advanced Processing and Packaging Studies (Raleigh, NC; Columbus, OH; Davis, CA), and the US National Science Foundation. REFERENCES Abad, F. X., Yinto, R. M., Gajardo, R. and Bosch, A. 1997. Viruses in mussels: Public health implications and depuration. J. Food Prot. 60(6), 677–681. Achen, M. 2000. Efficacy of ozone in inactivating Escherichia coli O157:H7 in pure cell suspensions and on apples. M. Sc. Thesis, The Ohio State University, Columbus, OH. Achen, M. and Yousef, A. E. 2001. Efficacy of ozone against Escherichia coli O157: H7 on apples. J. Food Sci. 66, 1380–1384. ACGIH 1986. “Threshold limit values for chemical substances in the work environment.” American Conference of Governmental Industrial Hygienists, Cincinnati, OH. Adler, M. G. and Hill, G. R. 1950. The kinetics and mechanism of hydroxide ion catalyzed ozone decomposition in aqueous solution. J. Am. Chem. Soc. 72, 1884–1886. APHA 1998. Ozone (residual). In “Standard Methods for the Examination of Water and Wastewater” (A. E. Greenberg, A. D. Eaton and L. S. Clesceri, eds; M. A. H. Franson, managing ed.), Ch. 4, pp. 137–139. American Public Health Association, Washington DC. Arimoto, M., Sato, J., Maruyama, K., Mimura, G. and Furusawa, I. 1996. Effect of chemical and physical treatments on the inactivation of striped jack nervous necrosis virus (SJNNV). Aqua Culture 143(1), 15–22. Bablon, G., Bellamy, W. D., Bourbigot, M-M., Daniel, F. B., Dore, M., Erb, F., Gordon, G., Langlais, B., Laplanche, A., Legube, B., Martin, G., Masschelein, W.J., Pacey, G., Reckhow, D. A. and Ventresque, C. 1991. Fundamental aspects. In “Ozone in Water Treatment, Application and Engineering” (G. Langlais, D.A. Reckhow and D. R. Brink, eds), pp. 11–132. Lewis Publishers, Inc., Chelsea, MI. Bader, H., and Hoigné, J. 1981. Determination of ozone in water by the indigo method. Water Res. 15, 449–456. Bailey, P. S. 1978. “Ozonation in Organic Chemistry.” Academic Press, Inc., New York. Barnes, E. M., and Impey, C. 1968. Psychrophilic spoilage bacteria of poultry. J. Appl. Bacteriol. 31, 97–107. Bellamy, W. D., Langlais, B., Lykins, B. Jr., Rakness, K. L., Robson, C. M. and Schulhof, P. 1991. Economics of ozone systems: New installations and retrofits. In “Ozone in Water Treatment, Applications and Engineering” (B. Langlais, D. A. Reckhow and D. R. Brink, eds), pp. 491–542. Lewis Publishers, Inc., Chelsea, MI. Besser, R. E., Lett, S. M., Weber, J. T., Doyle, M. P., Barrett, T. J., Wells, J. G., and Griffin, P. M. 1993. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157: H7 in fresh pressed apple cider. J. Am. Med. Assoc. 269, 2217–2220. Beuchat, L. R. 1995. Pathogenic microorganisms associated with fresh produce. J. Food Prot. 59, 204–216.

OZONE APPLICATION IN THE FOOD INDUSTRY

209

Beuchat, L. R., and Brackett, R. E. 1990. Survival and growth of Listeria monocytogenes on lettuce as influenced by shredding, chlorine treatment, modified atmosphere packaging and temperature. J. Food Sci. 55, 755–758, 870. Beuchat, L. R., Nai, B. V., Adler, B. B., and Clavero, M. R. S. 1998. Efficacy of spray application of chlorinated water in killing pathogenic bacteria on raw apples, tomatoes, and lettuce. J. Food Prot. 61, 1305-1311. Blogoslawski, W. J., Perez, C., and Hitchens, P. 1993. Ozone treatment of sea water to control vibriosis in marine culture of penaeid shrimp, Penaeus vannameif. In Proceedings of the International Symposium on Ozone-oxidation Methods for Water and Wastewater Treatment, 26–28 April, pp. I.5.1–I.5.11. Int. Ozone Assoc., Wasser, Berlin. Brackett, R. E. 1987. Antimicrobial effect of chlorine on Listeria monocytogenes. J. Food Prot. 50, 999–1003. Brady, J. E., and Humiston, G. E. 1978. “General Chemistry Principles and Structure,” 2nd edn. John Wiley & Sons, New York. Brungs, W. A. 1973. Effects of residual chlorine on aquatic life. J. Water Poll. Control Fed. 45, 2180–2193. Bryant, E. A., Fulton, G. P., and Budd, G. C. 1992. “Disinfection Alternatives for Safe Drinking Water.” Van Nostrand Reinhold, New York. Bullock, G. L., Summerfelt, S.T., Noble, A. C., Weber, A. L., Durant, M. D., and Hankins, J. A. 1997. Ozonation of a recirculating rainbow trout culture system. I. Effects on bacterial gill disease and heterotrophic bacteria. Aquaculture 158, 43–55. Bundegaard-Nielsen, K., and Nielsen, P. V. 1996. Fungicidal effect of 15 disinfectants against 25 fungal contaminants commonly found in bread and cheese manufacturing. J. Food Prot. 59, 268–275. Buydens, R., and Fransolet, G. 1971. L’action de l’ozone sur le chlore, le dioxide de chlore et le chlorite contenus dans les eaux trainees. Tribune Cebedeau 326, 4–6. Caracciolo, L. 1990. Ozone treatment of turkey scap meat. Data submitted to US FDA in support of a Food Additive Petition #0A4207. Carlin, F., Nguyen-the, C., and Abreu da Silva, A. 1995. Factors affecting the growth of Listeria monocytogenes on minimally processed fresh endive. J. Appl. Bacteriol. 78, 636–646. Carpentier, B., and Cerf, O. 1993. Biofilms and their consequences, with particular reference to hygiene in the food industry. J. Appl. Bacteriol. 75, 499–511. Castle, L., Mercer, A. J., and Gilbert, J. 1995. Chemical migration from polypropylene and polyethylene aseptic food packaging as affected by hydrogen peroxide sterilization. J. Food Prot. 58, 170–174. CDC 1993. Multistate outbreak of Salmonella serotype Montevideo infections. EPI-AID 93-79. CDC 1996a. Update: Outbreaks of Cyclospora cayetanensis infection – United States and Canada, 1996. MMWR 45(28), 611–612. CDC 1996b. Outbreak of Escherichia coli O157:H7 infections associated with drinking unpasteurized commercial apple juice – British Columbia, California, Colorado, and Washington, October 1996. MMWR 45(44), 975. CDC 1996c. Outbreak of Escherichia coli O157:H7 infection – Georgia and Tennessee, June 1995. MMWR 45(12), 249–251. CDC 1997. Escherichia coli O157:H7 infections associated with eating a nationally distributed commercial brand of frozen ground beef patties and burgers – Colorado, 1997. MMWR 46(33), 777–778. CDC 1998. Multistate outbreak of listeriosis – United Sates, 1998. MMWR 47(50), 1085–1086. CDC 2000. Multistate outbreak of listeriosis – United States, 2000. MMWR 49(50), 1129–1130.

210

J.-G. KIM ET AL.

CFR 1997. “Air Contaminants,” Title 29, Vol. 6, Part 1910. Office Federal Register, Washington, DC. Chen, H. C., Huang, S. H., Moody, M. V., and Jiang, S. T. 1992. Bacteriocidal and mutagenic effects of ozone on shrimp (Penaeus monodon) meat. J. Food Sci. 57, 923–927. Cox, J. P., Cox, J. M., Duffy Cox, R. W. 1995. Hyperpasteurization of food. US Patent 5 431 939. Dave, S. 1999. Efficacy of ozone against Salmonella enteritidis in aqueous suspensions and on poultry meat. M.Sc. Thesis, The Ohio State University, Columbus, Ohio. DeWitt, B. J. III, McCoid, V., Holt, B. L., Ellis, D. K., Finne, G., and Nickelson, R. II. 1984. The potential use of ozonated ice for on-board storage of Gulf of Mexico shrimp. In Proceedings of the Ninth Annual Tropical and Subtropical fisheries Technological Conference of the Americas, 9–12 January, pp. 260–279. Brownsville, TX. Diaz, M. E., and Law, S. E. 1999. UV-enhanced ozonation for reduction of pathogenic microorganisms and turbidity in poultry-processing chiller water for recycling. Proceedings of the 14th Ozone World Congress, IOA, pp. 391–403. Dearborne, MI. Dixon, B. 1998. Biofilms: cultural diversity in action. ASM News 64, 484–485. Dollear, F. G., Mann, G. E., Codifier, L. P. Jr., Garder, H. K. Jr., Koltun, S. P., and Vix, H. L. E. 1968. Elimination of aflatoxins from peanut meal. J. Am. Oil. Chem. Soc. 45, 862–865. Dore, M. 1989. Ozonation of molecules constituting cellular matter. Proceedings of the 9th Ozone World Congress, International Ozone Association, New York. Duca, J. G. 1999. Latest practices of CIP operations of the brewery. Tech. Q. Master Brewers’ Assoc. Americas 36(4), 407–411. Dychdala, G.R. 1991. Chlorine and chlorine compounds. In “Disinfection, Sterilization and Preservation” (S.S. Block, ed.), 4th edn, Ch. 7, pp. 131–151. Lea & Febiger, Philadelphia, PA. Elford, W. J., and van den Ende, J. 1942. An investigation of the merits of ozone as an aerial disinfectant. J. Hygiene 42, 240–265. EPRI 1999. Membrane filtration and ozonation of poultry chiller overflow water: Study of membrane treatment to reduce water use and ozonation for sanitation at a poultry processing plant. Technical Report-114435, EPRI, Palo Alto, CA. Ewell, A. W. 1946. Recent ozone investigation. J. Appl. Phys. 17, 908–911. Farhataziz, A., and Ross, B. 1977. Selected specific rates of reactions of transients from water in aqueous solution, National Bureau of Standards, Washington, DC. NSRDS-NBS59. Farooq, S., Chian, E. S. K., and Engelbrecht, R. S. 1977. Basic concepts in disinfection with ozone. J. Water Poll. Control Fed. 49, 1818–1831. FDA 1998. Guide to minimize microbial food safety hazards for fresh fruits and vegetables. US Department of Health and Human Services. Federal Register 2001. Secondary direct food additives permitted in food for human consumption. Federal Register 66 (123), 33829–33830. Finch, G. R., Smith, D. W., and Stiles, M. E. 1988. Dose–response of Escherichia coli in ozone demand-free phosphate buffer. Water Res. 22, 1563–1570. Fournaud, J., and Lauret, R. 1972. Influence of ozone on the surface microbial flora of beef during refrigeration and thawing. Technologia Alimentari 6, 12. Frank, J. F., and Koffi, R. A. 1990. Surface adherent growth of Listeria monocytogenes is associated with increased resistance to surfactant sanitizers and heat. J. Food Prot. 53, 550–554. Gardner, D. W. M., and Sharma, G. 1998. The kinetics of Bacillus subtilis spore inactivation on filter paper by ultraviolet light and ultraviolet light in combination with hydrogen peroxide. J. Appl. Microbiol. 84, 633–641. Garg, N., Churey, J. J., and Splittstoesser, D. F. 1990. Effect of processing conditions on the

OZONE APPLICATION IN THE FOOD INDUSTRY

211

microflora of fresh-cut vegetables. J. Food Prot. 53, 701–703. Gayler, G. E., MacCready, R. A., Reardon, J. P., and McKernan, B. F. 1955. An outbreak of salmonellosis traced to watermelon. Publ. Health Rep. 70, 311–313. Giese, A. C., and Christenser, E. 1954. Effects of ozone on organisms. Physiol. Zool. 27, 101–115. Glaze, W. H., Kang, J. W., and Chapin, D. H. 1987. The chemistry of water treatment and processes involving ozone, hydrogen peroxide, and ultraviolet radiation. Ozone Sci. Engng. 9, 335–352. Goche, L., and Cox, B. 1999. Ozone treatment of fresh H and G Alaska salmon. Report to Alaska Science and Technology Foundation and Alaska Department of Environmental Conservation. Seattle, WA, Surefish, November 1999. Gomella, C. 1972. Ozone practice in France. J. Am.Water Works Assoc. 64, 39–46. Gordon, G., and Grunwell, J. 1983. Comparison of analytical methods for residual ozone. Proceedings of the 2nd National Symposium – Municipal Wastewater Disinfection. EPA Report, EPA 600/9-83-009. Gorman, B. M., Sofos, J. N., Morgan, J. B., Schmidt, G. R., and Smith, G. C. 1995. Evaluation of hand-trimming, various sanitizing agents, and hot water spray-washing as decontamination interventions for beef brisket adipose tissue. J. Food Prot. 58, 899–907. Gorman, B. M., Kochevar, S. L., Sofos, J. N., Morgan, J. B., Schmidt, G. R., and Smith, G. C. 1997. Changes on beef adipose tissue following decontamination with chemical solutions or water of 35°C or 74°C. J. Muscle Foods 8, 185–197. Graham, D. M. 1997. Use of ozone for food processing. Food Technol. 51, 72–75. Graham, D. M. 2000. Direct food additive petition (Ozone as an antimicrobial agent for the treatment, storage, and processing of foods in gas and aqueous phases. Submitted to FDA on 2 August, 2000; approved June 2001. Greene A. K., Smith, G. W., and Knight, C. S. 1999. Ozone in dairy chilling water systems: effect on metal materials. Int. J. Dairy Technol. 52(4), 126–128. Greer, G. G., and Jones, S. D. M. 1989. Effects of ozone on beef carcass shrinkage, muscle quality and bacterial spoilage. Can. Ins. Food Sci. Technol. J. 22, 156–160. Grosse, A. V., and Streng, A. G. 1960. Storage of explosive gases. Research Institute, Temple University. Patent 2928529 (1958/1960). Gurol, M. D., and Singer, P. C. 1982. Kinetics of ozone decomposition, a dynamic approach. Environ. Sci. Technol. 16, 377–383. Haag, W. R., and Hoigné, J. 1983. Ozonation of bromide-containing waters: kinetics of formation of hypobromous acid and bromate. Environ. Sci. Technol. 17, 261. Hampson, B. C. 2000. Use of ozone for winery and environmental sanitation. Practical Winery & Vineyard, Jan–Feb, 27–30. Hampson, B. C., and Fiori, S. 1997. Application in food processing operations. In Proceedings of the IOAPAG Annual Conference Lake Tahoe, NV, pp. 261–267, International Ozone Association, Pan American Group, Stamford, CT. Haraguchi, T., Simidu, U., and Aiso, K. 1969. Preserving effect of ozone to fish. Bull. J. Soc. Sci. Fish. 35, 915–919. Harris, W. C. 1972. Ozone disinfection. J. Water Works Assoc. 64, 182–183. Herbold, K., Flehmig, B., and Botzenhart, K. 1989. Comparison of ozone inactivation, in flowing water, of hepatitis A virus, poliovirus 1, and indicator organisms. Appl. Environ. Microbiol. 55, 2949–2953. Hewes, C. G., and Davison, R. R. 1973. Renovation of wastewater by ozonation. Am. Inst. Chem. Engng. Symp. Ser. 69(129), 71–80. Hoffman, R. K. 1971. Toxic gases. In “Inhibition and Destruction of the Microbial Cell” (W.B. Hugo, ed.), pp. 225–258. Academic Press, New York. Hoigné, J., and Bader, H. 1975. Ozonation of water: role of hydroxyl radicals as oxidizing

212

J.-G. KIM ET AL.

intermediates. Science 190, 782–784. Hoigné, J. and Bader, H. 1976. The role of hydroxyl radical reactions in ozonation processes in aqueous solutions. Water Res. 10, 377–386. Hoigné, J., and Bader, H. 1979. Ozonation of water: Selectivity and rate of oxidation of solutes. Ozone Sci. Engng. 1, 73–85. Hoigné, J., and Bader, H. 1985. Rate constants of reactions of ozone with organic and inorganic compounds in water. III. Inorganic compounds and radicals. Water Res. 19, 993–1004. Hoigné, J., and Bader, H. 1986. Determination of ozone and chlorine dioxide in water by the indigo method. In “Analytical Aspects of Ozone Treatment of Water and Wastewater” (R.G. Rice, L. J. Bollyky and W. J. Lacy, eds), Ch. 9. pp. 129–149. Lewis Publishers, Inc. Chelsea, MI. Horvath, M., Bilitzky, L., and Huttner, J. 1985. “Ozone.” Elsevier, New York. Hunt, N. K., and Marinas, B. J. 1997. Kinetics of Escherichia coli inactivation with ozone. Water Res. 31, 1355–1362. Hurst, W. C. and Schuler, G. A. 1992. Fresh produce processing – an industry perspective. J. Food Prot. 55, 824–827. Hwang, E. S. 1999. Post harvest treatment to reduce or remove ethylene bisdithiocarbamate (EBDC) fungicide residues from apples and apple products and elucidation of possible degradation by-products and pathways. Ph.D. Dissertation, Michigan State University. Ihl, M., Monslaves, M., and Bifani, V. 1998. Chlorophylase inactivation as a measure of blanching efficacy and color retention of artichokes (Cynara scolymus L.). Lebensm. Wiss. Technol. 31, 50–56. Ingram, M., and Haines, R. B. 1949. Inhibition of bacterial growth by pure ozone in the presence of nutrients. J. Hyg. (Cambr.) 47, 146–158. Ito, K. A., and Seeger, M. L. 1980. Effects of germicides on microorganisms in can cooling water. J. Food Prot. 43, 484–487. Izat, A. L., Adams, M., Colberg, M., and Reiber, M. 1990. Effects of ozonated chill water on microbiological quality and clarity of broiler processing water. Arkansas Farm Res. 39(2), 9. Jans, U., and Hoigné, J. 1998. Activated carbon and carbon black catalyzed transformation of aqueous ozone into OH-radicals. Ozone Sci. Engng. 20(1), 67–90. Jindal, V., Waldroup, A. L., Forsythe, R. H., and Miller, M. 1995. Ozone and improvement of quality and shelflife of poultry products. J. Appl. Poul. Res. 4, 239–248. Kaess, G., and Weidemann, J. F. 1968a. Ozone treatment of chilled beef. I. Effect of low concentrations of ozone on microbial spoilage and surface colour of beef. J. Food Technol. 3, 325–334. Kaess, G., and Weidemann, J. F. 1968b. Ozone treatment of chilled beef. II. Interaction between ozone and muscle. J. Food Technol. 3, 335–343. Kang, S. J., Oh, J. Y., and Chung, J. K. 1999. Changes of antioxidant enzyme activities in leaves of lettuce exposed to ozone. J. Korean Soc. Horti. Sci. 40(5), 541–544. Kanofsky, J. R., and Sima, P. 1991. Singlet oxygen production from the reactions of ozone with biological molecules. J. Biol. Chem. 266, 9039–9042. Katzenelson, E., Kletter, B., and Shuval, H. I. 1974. Inactivation kinetics of viruses and bacteria in water by use of ozone. J. Am. Water Works Assoc. 66, 725–729. Kessel, J. F., Allison, D. K., Moore, F. J., and Kaime, M. 1943. Comparison of chlorine and ozone as virucidal agents of poliomyelitis virus. Proc. Soc. Exp. Biol. Med. 53, 71–73. Khadre, M. A., and Yousef, A. E. 2001a. Decontamination of a multilaminated aseptic food packaging material and stainless steel by ozone. J. Food Safety 21, 1–13. Khadre, M. A., and Yousef, A. E. 2001b. Sporicidal action of ozone and hydrogen peroxide, a comparative study. Int. J. Food Microbiol. 71, 131–138. Kim, C. K., Gentile, D. M., and Sproul, O. J. 1980. Mechanism of ozone inactivation of

OZONE APPLICATION IN THE FOOD INDUSTRY

213

bacteriophage f2. Appl. Environ. Microbiol. 39, 210–218. Kim, J. G. 1998. Ozone as an antimicrobial agent in minimally processed foods. Ph.D. thesis, The Ohio State University, Columbus, OH. Kim, J. G., and Yousef, A. E. 2000a. Ozone treatment to inactivate bacterial and fungal spores. In “Book of Abstracts”, p. 187. Institute of Food Technology, Chicago, IL. Kim, J. G., and Yousef, A. E. 2000b. Inactivation kinetics of foodborne spoilage and pathogenic bacteria by ozone. J. Food Sci. 65, 521–528. Kim, J. G., and Yousef, A. E. 2001. Ozone treatment to inactivate bacterial and fungal spores in fruit juices and juice-based beverages. American Society for Agricultural Engineering Annual Meeting, Sacramento, CA. Kim, J. G., Yousef, A. E., and Chism, G. W. 1999a. Use of ozone to inactivate microorganisms on lettuce. J. Food Safety 19, 17–34. Kim, J. G., Yousef, A. E., and Dave, S. 1999b. Application of ozone for enhancing the microbiological safety and quality of foods: a review. J. Food Prot. 62, 1071–1087. Kim, M. J., Oh, Y. A., Kim, M. H., Kim, M. K., and Kim, S. D. 1993. Fermentation of Chinese cabbage kimchi inoculated with Lactobacillus acidophilus and containing ozone-treated ingredients. J. Korean Soc. Food Nutr. 22, 165–174. Kinman, R. N. 1975. Water and wastewater disinfection with ozone, a critical review. Crit. Rev. Envir. Contr. 5, 141–152. Kirk, J. R., and Mitchell, S. K. 1980. Risks and benefits associated with chlorine in the food industry. Water Chlorination: Environ. Impact Health Eff. 3, 283–303. Klingman, M. H., and Christy, A. D. 2000. Design of an ozone treatment system for disinfection of cider apples. Trans. Am. Soc. Agric. Engng 43, 1989–1996. Kolle, W. 1968. Problem der gemeinsamen anwendung verschiedener oxygationsmittel bei der wasseraufbereitung. Vom Wasser, pp. 367–381. Quoted in Y. Richard and L. Brener, Interferences between ozone and chlorine. In “Handbook of Ozone Technology and Applications” (R.G. Rice and A. Netzer, eds) 1982, Vol. 1, pp. 277–284. Butterworth Publishers, Ann Arbor, MI. Komissarov, V. D., and Komissarova, I. N. 1973. Formation of phenol during benzene ozonolysis. Izv. Akad. Nauk. USSR, Ser. Khim. 3, 677–679. Kötters, J., Prahst, A., Skura, B., Rosenthal, H., Black, E. A. and Rodigues-Lopez, J. 1997. Observations and experiments on extending shelflife of ‘rockfish’ (Sebastes spp.) products with ozone. J. Appl. Ichthyol. 13, 1–8. Korich, D. G., Mead, J. R., Madore, M. S., Sinclair, N. A., and Sterling, C. R. 1990. Effects of ozone, chlorine dioxide, chlorine, and monochloramine on Cryptosporidium parvum oocyst viability. Appl. Environ. Microbiol. 56, 1423–1428. Kruithof, J., and Kamp, P. C. 1999. Pesticide control and disinfection by UV/H2O2 treatment without bromated formation. Proceedings of the American Water Works Association Annual Meeting, pp. 138–152. American Water Works Association, Washington, DC. Kuprianoff, J. 1953. The use of ozone for the cold storage of fruit. Z. Kalentechnik. 10, 1–4, 6. LA Times 2001. Tainted cantaloupes kill 1, sicken 30. 16 May, 2001. Le Chevaillier, M. W., Cawthou, C. D. and Lee, R. G. 1988. Inactivation of biofilm bacteria. Appl. Environ. Microbiol. 34, 2492–2499. Lee, J. S. and Kramer, D. E. 1984. Effectiveness of ozone-treated wash water and ice on keeping quality and stability of Sockeye salmon. Report FITC 84/T-1 to Alaska Sea Grant College Program and Alaska Department of Environmental Conservation, July 1984. Leffler, J. E. 1949. Cleavages and rearrangements involving oxygen radicals and cations. Chem. Rev. 45, 385–417. Lin, C. M., Kim, J., Du, W. and Wei, C. 2000. Bactericidal activity of isothiocyanate against

214

J.-G. KIM ET AL.

pathogens on fresh produce. J. Food Prot. 63, 25–30. Lyons-Magnus 1999. Ozone use survey data. Ozone treatment of fresh strawberries, data submitted to EPRI Agriculture and Food Alliance, 28 September 1999, Fresno, CA. Mahieux, F. 1962. Genie chemique. Chemie et Industrie 87, 15. Masschelein, W. J. 1982. Contacting of ozone with water and contactor offgas treatment. In “Handbook of Ozone Technology and Applications” (R.G. Rice and A. Netzer, eds.). Ann Arbor Science Publishers, Ann Arbor, MI. McGowan, C. L., Bethea, R.M., and Tock, R. W. 1979. Feasibility of controlling onion and garlic dehydration odors with ozone. Trans. Am. Soc. Agri. Engng. 22, 899–905, 911. McGrath, W. D., and Norrish, R. G. W. 1960. Studies of the reactions of excited oxygen atoms and molecules produced in the flash photolysis of ozone. Proc. Roy. Soc. (Lond.), Ser. A 254, 317. McLoughlin, G. 2000. Apple processor switches from chlorine to ozone treatment. Water Technol. 23(2), 53–54. Meddows-Taylor, J. 1947. Some characteristics of ozone in relation to water treatment. J. Inst. Water Eng. 1, 187–201. Menzel, D. B. 1984. Ozone: an overview of its toxicity in man and animal. J. Toxicol. Environ. Health 13, 183–204. Millard, P. S., Gensheimer, K. F., Adiss, D. G., Sosin, D. M., Beckett, G. A., Houck-Jankoski, A., and Hudson, A. 1994. An outbreak of cryptosporidiosis from fresh pressed apple cider. JAMA 272, 1592–1596. Mimura, G., Nakamitu, T., Nagase, T., and Namba, K. 1998. Qualitative assay of residual oxidant in seawater and effect of several oxidants on Japanese flounder, Paralichthys olivaceous, eggs. Suisanzoshoku 46(4), 579–587. Mitsuda, H., Ominami, H., and Yamamoto, A. 1990. Synergistic effects of ozone and carbon dioxide gases for sterilizing food. Proc. Japan. Acad. 66, 68–72 (Ser. B). Murray, R. G., Pamela, S., and Elson, H. E. 1965. The location of the mucopeptide of sections of the cell wall of Escherichia coli and other Gram negative bacteria. Can. J. Microbiol. 11(3), 547–560. Mustafa, M. G., Faeder, E. J., and Lee, S. D. 1980. Biochemical basis of pulmonary response to ozone and nitrogen dioxide injury. In “Molecular Basis of Environmental Toxicity” (R.S. Bhatnagar, ed.), pp. 151–172. Ann Arbor Science, Ann Arbor, MI. NACMCF 1999. Microbiological safety evaluations and recommendations on fresh produce. National Advisory Committee on Microbiological Criteria for Foods. Food Control 10, 117–143. Nadel, J. A. 1979. Airway hyperirritability induced by ozone. Report to the California Air Resources Control Board, Contract A8-053-30. Naitoh, S. 1989. Studies on the utilization of ozone in food preservation: Effect of ozone treatment on airborne microorganisms in a confectionery plant. J. Antibact. Antifung. Agents 17(10), 483–489. Naitoh, S. 1993. Studies on the application of ozone in food: Effect of ozone treatment on aerial contaminants in a plastic film factory. J. Antibact. Antifung. Agents 21, 445–451. Naitoh, S., Okada, Y., and Sakai, T. 1987. Studies on utilization of ozone in food preservation: III. Microbicidal properties of ozone on cereal grains, cereal grain powders, peas, beans, and whole spices. J. Jap. Soc. Food Sci. Technol. 34, 788–793. Naitoh, S, Okada, Y., and Sakai, T. 1988. Studies on utilization of ozone in food preservation: V. Changes in microflora of ozone-treated cereals, grain, peas, beans, and spices during storage. J. Jap. Soc. Food Sci. Technol. 35, 69–77. Naitoh, S., Sawada, Y., and Yamaguchi, N. 1989. Studies on utilization of ozone in food

OZONE APPLICATION IN THE FOOD INDUSTRY

215

preservation: Effect of ozone treatment on storage of packaged namamen Japanese raw noodle. J. Antibact. Antifung. Agents 17, 517–526. Nguyen-the, C., and Carlin, F. 1994. The microbiology of minimally processed fresh fruits and vegetables. Crit. Rev. Food Sci. Nutrit. 34, 371–401. O’Donovan, D. C. 1965. Treatment with ozone. J. Am. Water Works Assoc. 57, 1167–1192. Odumeru, J. A., Mitchell, S. J., Alves, D. M., Lynch, J. A., Yee, A. J., Wang, S. L., Styliadis, S., and Farber, J. M. 1997. Assessment of the microbiological quality of ready-to-use vegetables for health-care food services. J. Food Prot. 60, 954–960. Ogawa, J. M., Feliciano, A. J., and Manji, B. T. 1990. Evaluation of ozone as a disinfectant in postharvest dump tank treatments for tomato. Phytopathology (Abstr.) 80, 1020. Ong, K. C., Cash, J. N., Zabik, M. J., Siddiq, M., and Jones, A. L. 1996. Chlorine and ozone washes for pesticide removal from apples and processed apple sauce. Food Chem. 55, 152–160. Orth, R., and Mrozek, H. 1989. Is the control of Listeria, Campylobacter and Yersinia a disinfection problem? Fleischwirtshaft. 69, 1575–1576. Oshima, T., Sato, K., Terauchi, H., and Sato, M. 1997. Physical and chemical modifications of high-voltage pulse sterilization. J. Electrostatics 42, 159–166. Ouederni, A., Mora, J. C., and Bes, R. S. 1987. Ozone absorption in water: Mass transfer and solubility. Ozone Sci. Eng. 9, 1–12. Page, T., Harris, R. H., and Epstein, S. S. 1976. Drinking water and cancer mortality in Louisiana. Science 193, 55–57. Peeters, J. E., Mazas, E. A., Masschelein, W. J., Martinez de Maturana, I. V., and Debacker, E. 1989. Effect of disinfection of drinking water with ozone or chlorine dioxide on survival of Cryptosporidium parvum oocysts. Appl. Environ. Microbiol. 55, 1519–1525. Prat, R., Nofre, C., and Cier, A. 1968. Effets de L’Hypochlorite sodium, de l’ozone et des radiations ionisantes sur les constituents pyrimidiques d’Escherichia coli. Ann. Inst. Pasteur 114, 595–607. Pryor, A., and Rice, R. G. 2000. Ozone toxicology and guidelines for safe use in food processing systems. Ozone News 28(4). Rayner, E. T., Dwarakanath, C. T., Mann, G. E., and Dollear, F. G. 1971. Aflatoxin reduction. US Patent 3592641. Reagan, J. O., Acuff, G. R., Buege, D. R., Buyck, M. J., Dickson, J. S., Kastner, C. L., Marsden, J. L., Mortan, J. B., Nickelson, R.II, Smith, G. C., and Sofos, J. N. 1996. Trimming and washing of beef carcasses as a method of improving the microbiological quality of meat. J. Food Prot. 59, 751–756. Restaino, L., Frampton, E. W., Hemphill, J. B., and Palnikar, P. 1995. Efficacy of ozonated water against various food-related microorganisms. Appl. Environ. Microbiol. 61, 3471–3475. Ries, A. A., Zaza, S., Langkop, C., Tauxe, R. V., and Blake, P. A. 1990. A multistate outbreak of Salmonella Chester linked to imported cantaloupe. p. 238, Abstract no. 915. 30th Interscience Conference on Antimicrobical Agents and Chemotherapy, American Society for Microbiology, Washington, DC. Rodgers, S. L., Cash, J. N., and Ryser, E. T. 1999. Inactivation of E. coli 0157:H7 and L. monocytogenes using ozone, chlorine dioxide, sodium hypochlorite and peracetic acid. IAMFES. Annual Meeting, Dearborn, MI. Rosenthal, H. 1974. Selected bibliography of ozone, its biological effects and technical applications. Fisheries research board of Canada technical report No. 456, Fisheries and Marine Service, Pacific Biological Station, Nanaimo, BC. Roth, J. A., and Sullivan, D. E. 1981. Solubility of ozone in water. Indus. Engng. Chem. Fund. 20, 137. Sarig, P., Zahavi, T., Zutkhi, Y., Yannai, S., Lisker, N., and Ben-Arie, R. 1996. Ozone for control

216

J.-G. KIM ET AL.

of post-harvest decay of table grapes caused by Rhizopus stolonifer. Physiol. Mol. Plant Pathol. 48, 403–415. Scarpino, P. V., Berg, G., Chang, S. L., Dahling, D. and Lucas, M. 1972. A comparative study of inactivation of viruses in water by chlorine. Water Res. 6, 959. Schomer, H.A., and McColloch, L.P. 1948. Exposure of apple to ozone atmospheres. US Department of Agriculture Circular 765. Scott, D. B. M. 1975. The effect of ozone on nucleic acids and their derivatives. In “Aquatic Applications of Ozone” (W.J. Blogoslawski, and R.G. Rice, eds.), pp. 226–240. International Ozone Institute, Syracuse, NY. Scott, D. B. M., and Lesher, E. C. 1963. Effect of ozone on survival and permeability of Escherichia coli. J. Bacteriol. 85, 567–576. Sease, W.S. 1976. Ozone mass transfer and contact system. In “Proceedings of the Second International Symposium on Ozone Technology” (R. G. Rice, P. Pichet and M. A. Vincentr, eds), pp. 1–14. International Ozone Association, Vienna, VA. Shechter, H. 1973. Spectrophotometric method for determination of ozone in aqueous solutions. Water Res. 7, 729–739. Sheldon, B. W. 1986. Effects of ozonation of poultry processing water on the characteristics of process water and potential for recycling. Industry Summary, Southwestern Poultry and Egg Association. Sheldon, B. W., and Brown, A. L. 1986. Efficacy of ozone as a disinfectant for poultry carcasses and chill water. J. Food Sci. 51, 305–309. Sheldon, B. W., and Chang, Y. H. 1987. The application of ozone and other physical processes for treating spent poultry chiller water. Proceedings of the Food Processing Waste Conference. Atlanta, GA. Siddiqui, M. S. 1996. Chlorine-ozone interactions: formation of chlorate. Water Res. 30, 2160–2170. Siler, M. D. 1998. The use of preharvest intervals, post harvest wash treatments and processing in the removal of pesticides from apple fruit. M.Sc. Thesis, Michigan State University. Singh, G., and Singh, A. K. 1999. Localized corrosion in chlorine dioxide bleach media of paper industry. Bull. Electrochem. 15(3–4), 127–130. Smock, R. M., and Watson, R. D. 1941. Ozone in apple storage. Refrig. Engng 92, 97–101. Spotts, R. A., and Cervantes, L. A. 1992. Effect of ozonated water on postharvest pathogens of pear in laboratory and packinghouse tests. Plant Dis. 76(3), 256–259. Staehelin, J., Buhler, R. E., and Hoigné, J. 1984. Ozone decomposition in water studied by pulse radiolysis. 2. OH and HO4 as chain intermediates. J. Phys. Chem. 88, 5999–6004. Stefanovic, S. and Dickerson, R. W. Jr 1986. Removal of hydrogen peroxide from flat packaging material used in aseptic packaging of food. Curr. Technol. Flexible Packag. 912, 24–36. Stockinger, H. E. 1965. Ozone toxicology. A review of research and industrial experience: 1954–1964. Arch. Environ. Health. 10, 719–31. Stumm, W. 1958. Ozone as a disinfectant for water and sewage. J. Boston Soc. Civ. Eng. 46, 68. Tahoe Food Technology, Inc. 1998. Dynamic Ox biological burden reduction. US Patent Application No. 09/217.581, Sparks, NY. Taube, H., and Bray, W. C. 1940. Chain reactions in aqueous solutions containing ozone, hydrogen peroxide and acid. J. Am. Chem. Soc. 62, 3357. Tauxe, R., Kruse, H., Hedberg, C., Potter, M., Madden, J., and Wachsmuth, K. 1997. Microbial hazards and emerging issues associated with produce. A preliminary report to the national advisory committee on microbiologic criteria for foods. J. Food Prot. 60, 1400–1408. Thorp, E. D. 1950. The toxicity of ozone: A report and bibliography. Ind. Med. Surg. 15, 49–57. Tomiyasu, H., Fukutomi, H., and Gordon, G. 1985. Kinetics and mechanisms of ozone

OZONE APPLICATION IN THE FOOD INDUSTRY

217

decomposition in basic aqueous solution. Inorg. Chem. 24, 2962–2966. Trambarulo, R., Ghosh, S. N., Burrus, C. A. Jr and Gordy, W. 1953. The molecular structure, dipole moment, and g factor of ozone from its microwave spectrum. J. Chem. Phys. 21, 851–855. TVA. 2001. TVA, partners announce new food processing technology using ozone, April 10, 2001. http://www.tva.gov/news/releases/04-10-ozone.htm. Unal, R., Kim, J. G., and Yousef, A. E. 2001. Inactivation of Escherichia coli O157:H7, Listeria monocytogenes, and Lactobacillus leichmannii by combinations of ozone and pulsed electric field. J. Food Protect. 64, 777–782. USDA 1982. 5 November. Code of Federal Regulations. Title 21 Part 184 – GRAS status of ozone. Vol. 47, No. 215. Office Federal Register, Washington, DC. USDA 1984. 13 March. Code of Federal Regulations. Title 9 Part 381 – Poultry Products; Chiller Water Reuse. Vol. 49, No. 50. Office Federal Register, Washington, DC. US-FDA 1999. National advisory committee on microbiological criteria for food. http:// vm.cfsan.fda.gov/~mow/sprouts2.html#summary. Venosa, A. D. 1972. Ozone as a water and wastewater disinfectant. In “Ozone in Water and Wastewater Treatment” (F. L. Evans, III, ed.), p. 95. Ann Arbor Science Publishers Inc., Ann Arbor, MI. Viera, M. R., Guiamet, P. S., de Mele, M. F. L. and Videla, H. A. 2000. The effect of dissolved ozone on the corrosion behavior of heat exchanger structural materials, biocidal efficacy on bacterial films. Corrosion Rev. 18(2–3), 205–270. Waldroup, A. L., Hierholzer, R. E., Forsythe, R. H., and Miller, M. J. 1993. Recycling of poultry chill water using ozone. J. Appl. Poul. Res. 2, 330–336. Waller, J. G., and McTurk, G. 1965. Storage of compressed gaseous ozone. J. Appl. Chem. 15, 363. Watson, G. 1943. Some factors influencing the toxicity of ozone to fungi in cold storage. Refrig. Engng. 46, 103. Wei, C-I., Cook, D. L., and Kirk, J. R. 1985. Use of chlorine compounds in the food industry. Food Technol. 39, 107–115. Weiss, J. 1935. Investigation on the radical HO2 in solution. Trans. Faraday Soc. 31, 668. Whistler, P. E., and Sheldon, B. W. 1989. Biocidal activity of ozone versus formaldehyde against poultry pathogens inoculated in a prototype setter. Poult. Sci. 68, 1068–1073. White, G. C. 1986. Ozone. In “The Handbook of Chlorination,” 2nd edn, Ch.13, pp. 893–961. Van Nostrand Reinhold Co., New York. Wojtowicz, J. A. 1996. Ozone. In “Encyclopedia of Chemical Technology” (R. E. Kirk, and D. E. Othmer, eds), 4th edn, Vol. 17, pp. 953–994. Wiley-Interscience, New York. Wood, R. C., Hedberg, C., and White, K. 1991. A multistate outbreak of Salmonella javiana associated with raw tomatoes. Abstract. p. 69, Epidemic Intelligence Service 40th Annual Conference, CDC, Atlanta, GA. Yang, P. P. W., and Chen, T. C. 1979. Stability of ozone and its germicidal properties on poultry meat microorganisms in liquid phase. J. Food Sci. 44, 501–504. Yokoyama, M. 1990. Aseptic Packaged Foods. In “Food Packaging” (T. Kadoya, ed.), Ch.12, pp. 213–228. Academic Press, New York. Yousef, A. E., and Rodriguez-Romo, L. 2001. Methods for decontaminating shell eggs. US Patent Application, Dockett Number: 22727–04099. Yurteri, C., and Gurol, M. D. 1988. Ozone consumption in natural water: Effects of background organic matter, pH and carbonate species. Ozone Sci. Engng. 10, 277–290. Zagon, J., Dehne, L. I., Wirz, J., Linke, B., and Boegl, K. W. 1992. Ozone treatment for removal of microorganisms from spices as an alternative to ethylene oxide fumigation or irradiation. Results of a practical study. Bundesgesundheitsblatt 35, 20–23. Zhang, S., and Farber, J. M. 1996. The effects of various disinfectants against Listeria

218

J.-G. KIM ET AL.

monocytogenes on fresh-cut vegetables. Food Microbiol. 13, 311–321. Zhao, C., Ge, B., De Villena, J., Sudler, R., Yeh, E., Zhao, S., White, D. G., Wagner, D., and Meng, J. 2001. Prevalence of Campylobacter spp., Escherichia coli, and Salmonella serovars in retail chicken, turkey, pork, and beef from the greater Washington, DC, area. Appl. Environ. Microbiol. 67, 5431–5436. Zhao, J., and Cranston, P. M. 1995. Microbial decontamination of black pepper by ozone and the effect of the treatment on volatile oil constituents of the spice. J. Sci. Food Agric. 68, 11–18. Zhuang, H., Hildebrand, D. F. and Barth, M. M. 1996. Short-term ozonated water treatment affects quality and physiology of broccoli during postharvest storage. Institute of Food Technologists Annual Meeting: book of abstracts. p. 99.

THE HIGH MOLECULAR WEIGHT SUBUNITS OF WHEAT GLUTENIN AND THEIR ROLE IN DETERMINING WHEAT PROCESSING PROPERTIES PETER R. SHEWRY, NIGEL G. HALFORD AND ARTHUR S. TATHAM IACR-Long Ashton Research Station Department of Agricultural Sciences, University of Bristol Long Ashton, Bristol BS41 9AF, UK

YVES POPINEAU Institut National de la Recherche Agronomique Centre de Recherches de Nantes Laboratoire de Biochimie et de Technologie des Protéines B.P. 71627, Rue de la Géraudière, Nantes 44316, Cedex 03, France

DOMENICO LAFIANDRA Dipartimento di Agrobiologia ed Agrochimica Universitè degli Studi della Tuscia Via San Camillo de Lellis, Viterbo 01100, Italy

PETER S. BELTON School of Chemical Sciences University of East Anglia, Norwich NR4 7TJ, UK

I. Introduction A. Overview of Wheat Gluten Proteins II. The HMW Subunits of Glutenin A. Genetics and Polymorphism of HMW Subunits B. Correlations between HMW Subunit Composition and Grain Processing Quality III. The Sequences and Structures of HMW Subunits A. Amino Acid Sequences B. Size and Shape of HMW Subunits C. Structures of HMW Subunit Terminal Domains D. Structure of HMW Subunit Repetitive Domains E. Cross-link Formation between HMW Subunits

ADVANCES IN FOOD AND NUTRITION RESEARCH VOL 45 ISBN: 0-12-016445-0

Copyright © 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

220

P. R. SHEWRY ET AL.

IV. Experimental Evidence for the Role of HMW Subunits in Dough Mixing and Gluten Viscoelasticity A. Gluten Fractionation and Reconstitution B. Incorporation of Protein Fractions into Dough C. Incorporation of Purified HMW Subunits into Dough D. Incorporation of HMW Subunit Peptides into Dough V. Manipulating HMW Subunit Composition A. Near Isogenic Lines B. Transgenic Lines VI. Experimental Evidence for Differential Effects of Individual HMW Subunits on Mixing and Rheological Properties A. Variation between Cultivars B. Near Isogenic Lines C. Transgenic Lines VII. The Molecular Basis for Correlations between HMW Subunits and Quality A. Protein Amount B. Protein Sequences C. Stability of Allelic Subunits D. Subunit Interactions VIII. Theoretical Basis for Properties of the HMW Subunits and Their Role in Gluten Viscoelasticity and Dough Mixing A. Why are HMW Subunits Insoluble? B. Interactions between HMW Subunits C. High Molecular Weight Subunits and Dough Viscoelasticity D. Explanation for the Role of the HMW Subunits in Dough Mixing E. Is the Loop and Train Hypothesis Consistent with Other Models? IX. Conclusions and Future Prospects Acknowledgements References

I. INTRODUCTION

The mature wheat grain contains about 9–15% protein, approximately half of which is storage proteins deposited in the starchy endosperm cells. These cells are separated from the outer layers (including aleurone) and embryo during milling to give white flour. When the white flour is mixed with water and kneaded to form dough the storage proteins form a continuous network, called gluten, which confers unique mechanical properties that underpin the utilization of wheat in many food systems. These properties are a combination of elasticity and viscosity, and the precise balance between these properties (dough strength) is important in determining the end use. In particular, relatively strong (i.e. highly elastic) doughs are required to make pan breads and weaker doughs to make flat breads, noodles, cakes and biscuits. Wheat gluten can be readily prepared by washing dough with water, giving a cohesive mass comprising about 80% protein, 10% starch, 5%

HMW SUBUNITS OF WHEAT GLUTENIN

221

lipid and 5% other components (minerals, fibre, etc.) on a dry weight basis. This has facilitated studies of the composition and biophysical properties of gluten but it should not be forgotten that these properties are also likely to be affected by interactions with other components in dough. Because of their importance in wheat processing, the structures and properties of the wheat gluten proteins have been studied for many years, starting with the first description of the isolation of gluten by Beccari in 1745 (Beccari, 1745). Subsequent studies showed that gluten proteins could be separated into two fractions, which were either soluble or insoluble in alcohol (Taddei, 1819), and this division, with some modifications, has remained in use to the present day, with the gluten proteins that are readily soluble in alcohol–water mixtures (e.g. 60–70% ethanol) being called gliadins and those that are insoluble being called glutenins. These fractions are also important as they have functional significance, with the glutenins being mainly responsible for gluten and dough viscosty and elasticity and the gliadins for plasticity and extensibility (as discussed below). However, we now know that the two fractions contain proteins that are structurally related, with the differences in solubility resulting from their presence as monomers that interact by noncovalent forces (gliadins), or as high molecular mass polymers stabilized by inter-chain disulphide bonds. When present as reduced subunits the glutenin proteins are also soluble in alcohol–water mixtures and can therefore be defined together with the gliadins as prolamins. The ratio of gliadin to glutenin proteins in dough and gluten is generally about 1 : 1, although this ratio may vary with genotype and growth conditions with resulting effects on dough strength (Doekes and Wennekes, 1982; Graybosch et al., 1995; Vereijken et al., 2000; Johansson et al., 2001). A. OVERVIEW OF WHEAT GLUTEN PROTEINS

The gliadins and glutenins are not single homogeneous proteins but complex mixtures that can be separated using various electrophoretic and other procedures. In fact, the true extent of the complexity has never been conclusively established as it is difficult to determine the correspondence of components separated by different procedures or whether apparently single components actually contain two or more closely related proteins. However, it is generally accepted that at least 50 different gluten proteins are present in hexaploid bread wheat, and that only partial separation can be achieved by any single procedure, even by two-dimensional systems such as that shown in Figure 1. A high level of complexity is supported by the analysis of gluten protein genes, although the existence of pseudogenes (i.e. nonexpressed genes) makes it difficult to establish precise gene copy numbers (see, for example, Anderson and Greene, 1997).

222

P. R. SHEWRY ET AL.

FIG. 1. The polymorphism of wheat gluten proteins (c.v. Chinese Spring) demonstrated by two-dimensional electrophoresis (isoelectric focusing followed by SDS-PAGE). The groups of HMW prolamins (HMW subunits), S-poor prolamins (ω-gliadins) and S-rich prolamins (α-gliadins, γ-gliadins and LMW subunits) are indicated. Taken from Shewry et al. (1987) with permission.

The gliadin proteins are traditionally separated into four groups, called (α, β- γ- and ω-gliadins, based on their mobilities when separated by electrophoresis at low pH (ω-gliadins being slowest) (Figure 2). We now know that α- and β-gliadins are closely related in sequence and structure and it is usual to refer to both as α-type gliadins, as opposed to the γ-type gliadins which are more distantly related. Furthermore, the α-type and γ-type gliadins are often classified together as sulphur-rich (S-rich) prolamins, while the ω-gliadins form a separate S-poor group. The glutenin polymers can be separated on the basis of their size, as discussed below. However, in order to separate their component subunits it is necessary to first reduce the inter-chain disulphide bonds. Once this is done the subunits can be separated by electrophoresis in the presence of the detergent sodium dodecylsulphate (SDS) into two broad groups of subunits called high molecular weight (HMW) and low molecular weight (LMW) subunits, as shown in Figure 2. The low molecular weight subunits can be further classified into three groups: the D-type subunits, which appear to be related to ω-gliadins; the C-type, which comprise components related to α-type and γ-type gliadins, and the B-type, which form a discrete group of S-rich prolamins. The

HMW SUBUNITS OF WHEAT GLUTENIN

223

FIG. 2. The groups of gliadins and glutenin subunits separated by lactate-PAGE and SDS-PAGE, respectively. Taken from Shewry et al. (1999), with permission.

HMW subunits are not closely related to any other gluten proteins and form a distinct group called the HMW prolamins. This classification is summarized in Figure 3. WHEAT GLUTEN PROTEINS (PROLAMINS)

S-POOR

S-RICH

gliadins

ω-gliadins

γ-type gliadins

α-type gliadins

glutenin subunits

D-type LMW Subunits

C-type LMW Subunits

C-type LMW Subunits

HMW

B-type LMW Subunits

FIG. 3. The classification and nomenclature of wheat gluten proteins.

HMW Subunits

224

P. R. SHEWRY ET AL.

Although the HMW subunits were only identified and defined as a group about 20 years ago, they have since become the most widely and intensively studied group of gluten proteins. This is because of work started in the late 1970s which demonstrated correlations between HMW subunit composition and grain quality. However, before discussing these correlations in detail it is first necessary to discuss the polymorphism and genetic control of the HMW subunits.

II. THE HMW SUBUNITS OF GLUTENIN A. GENETICS AND POLYMORPHISM OF HMW SUBUNITS

The use of sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) for the analysis of the subunits resulting from the reduction of glutenin polymers allowed the determination of their genetic control. Wheat genetic stocks, such as the nulli-tetrasomic and ditelocentric lines analysed by Bietz et al. (1975), gave the first clear evidence for the chromosomal location of glutenin subunit genes in the bread wheat cv. Chinese Spring. Subsequently, the introduction of the discontinous SDSPAGE system of Laemmli (1970) and two-dimensional electrophoretic separations have given more detailed information on the genetics of HMW glutenin subunits and the extent of their polymorphism in different bread wheat cultivars (Lawrence and Shepherd, 1980; Holt et al., 1981; Payne and Lawrence, 1983). In particular, it is now firmly established that genes controlling the synthesis of HMW subunits are located on the long arms of the homoeologous group 1 chromosomes of hexaploid bread wheat at loci designated Glu-A1, Glu-B1 and Glu-D1, with each Glu-1 locus containing two tightly linked genes encoding subunits of high and low Mr, which are termed x- and y-type, respectively. These gene loci have also been mapped but contrasting results have been obtained (Payne et al., 1981b; Singh and Shepherd, 1988). DNA sequencing of the genes corresponding to HMW glutenin subunits has revealed the structural characteristics of the corresponding subunits, including the presence of a large repetitive central domain that may provide a basis for major and rapid structural changes in the Glu-1 genes by duplication and/or deletion of large segments as a result of unequal crossing over (Shewry et al., 1989). These processes, together with the accumulation of small insertions, deletions or point mutations, have resulted in the existence of large numbers of allelic forms of subunits encoded by each locus, as detected by current electrophoretic techniques (Figure 4). Based on electrophoretic analyses of about 300 wheat varieties, Payne and

HMW SUBUNITS OF WHEAT GLUTENIN

FIG. 4. SDS-PAGE of HMW subunits from a range of genotypes of wheat showing allelic variation in the mobilities of proteins encoded by the Glu-A1, Glu-B1 and Glu-D1 loci. The numbering is according to Payne and Lawrence (1983), with subsequent modifications by other workers.

225

226

P. R. SHEWRY ET AL.

Lawrence (1983) identified three alleles at the Glu-A1 locus, eleven at the Glu-B1 locus and five at the Glu-D1 locus, and proposed a numbering system to designate the different alleles. Subsequent analyses of cultivars from different countries and of wheat collections stored in gene banks has resulted in a continued increase in the number of alleles detected at each of the three loci (Figure 4). The detection of genes at the Glu-D1 and Glu-A1 loci encoding rare subunits with unusual high molecular mass (see for example lanes 6, 13 and 14 of Figure 4) and their characterization by polymerase chain reaction (PCR) have provided further evidence for the role of the sequence encoding the repetitive central domain in the molecular evolution of glutenin subunit genes (D’Ovidio et al., 1994; Tahir et al., 1996). Despite the fact that bread wheat possesses six HMW subunit genes, the number of expressed subunits ranges from three to five because of gene silencing processes that have occurred during its evolutionary history. In particular, the y-type gene present at the Glu-A1 locus is always silent in cultivated wheat, while the x-type gene at the same locus and the y-type gene at the Glu-B1 locus are expressed only in some cultivars. In contrast, the y-type subunit encoded by the Glu-A1 locus is expressed in cultivated and wild diploid wheats (Triticum monococcum ssp. monococcum and ssp. boeoticum, T. urartu), in the wild tetraploid wheat T. turgidum ssp. dicoccoides (Waines and Payne 1987; Levy et al., 1988) and in cultivated and wild forms of tetraploid wheats with the genomic formula AAGG (Margiotta et al., 1998). Studies of variation in HMW glutenin subunits present in old cultivars or landraces have also resulted in the identification of unusual allelic variants or mutant types characterized by the absence of subunits that are normally present in current bread wheat cultivars. For example, the absence of both HMW glutenin subunits encoded by chromosome 1D was reported by Bietz et al. (1975) in seeds of the landrace Nap Hal, while lines lacking either x- or y-type subunits encoded by the Glu-D1 or GluB1 loci have also been identified (Payne et al., 1984; Lafiandra et al., 1988). This type of material is being used for the development of genetic stocks suitable for elucidating the composition–functionality relationships, as discussed in a later section. B. CORRELATIONS BETWEEN HMW SUBUNIT COMPOSITION AND GRAIN PROCESSING QUALITY

The first direct correlation between HMW subunit composition and grain processing quality was reported by Payne et al. (1979), who showed a correlation between the presence of HMW subunit 1Ax and quality

HMW SUBUNITS OF WHEAT GLUTENIN

227

(measured using the indirect SDS sedimentation test) in the progeny of a cross between wheat cultivars of good (Maris Wigeon) and poor (Maris Ranger) breadmaking quality. Further studies showed a similar correlation with the subunit pair 1Dx5 + 1Dy10 (Payne et al., 1981a). Although Payne and coworkers used an indirect test for breadmaking quality, an association of subunits 1Dx5 + 1Dy10 and 1Ax2* with loaf volume was subsequently reported by Moonen et al. (1982, 1983), while Burnouf and Bouriquet (1980) demonstrated that subunits 1Dx5 and 1Bx7 were present in cultivars with good breadmaking quality and high gluten strength. The association of specific subunits or pairs of subunits with good or poor breadmaking quality (i.e. high or low gluten strength) has since been confirmed by many studies carried out using a wide range of germplasm and by workers in many countries (as reviewed by Payne, 1987 and Shewry et al., 1989, 1992), with general agreement on the following: 1. Subunits encoded by all three genomes (A, B, D) may be associated with quality. 2. The subunit pair 1Dx5 + 1Dy10 encoded by chromosome 1D is associated with the highest quality, whereas the allelic pairs 1Dx2 + 1Dy12, 1Dx3 + 1Dy12 and 1Dx5 + 1Dy12 are all associated with poor quality. 3. The presence of an x-type subunit encoded by chromosome 1A (1Ax1 or 1Ax2*) is superior to the null (i.e. silent) allele. 4. The subunit pair 1Bx17 + 1By18 is generally superior to other alleles encoded by chromosome 1B. Furthermore, the combination of information from a wide range of studies has allowed “quality scores” to be assigned to individual subunits or subunit pairs (Payne et al., 1987b; Branlard et al., 1992), as shown in Table I. TABLE I QUALITY SCORES ASSIGNED TO INDIVIDUAL HMW SUBUNITS OR SUBUNIT PAIRS

Locus Score

Glu-A1

Glu-B1

Glu-D1

4 3 3 2 2 1 1

– 1* 2* – – null –

– 17 + 18 7+8 7+9 – 7 6+8

5 + 10 – – 2 + 12 3 + 12 4 + 12 –

Taken from Payne et al. (1987b) with permission.

228

P. R. SHEWRY ET AL.

Quantitative analyses have shown that the HMW subunits account for up to about 12% of the total grain proteins, corresponding to about 1–1.7% of the flour dry weight (Seilmeier et al., 1991; Halford et al., 1992; Nicolas, 1997). Nevertheless, they account for between about 45 and 70% of the variation in breadmaking performance within European wheats (Branlard and Dardevet, 1985, Payne et al., 1987b, 1988a). Consequently, the quality scores assigned to the HMW subunits can be exploited to select for breadmaking performance in breeding programmes. III. THE SEQUENCES AND STRUCTURES OF HMW SUBUNITS

Although early studies resulted in the direct determination of the N-terminal amino acid sequences of a number of HMW subunits purified from wheat grain (Field et al., 1982; Shewry et al., 1984), the determination of full sequences resulted only from the isolation and sequencing of the corresponding genes. As a result we now know the sequences of nine HMW subunit proteins from bread wheat, including forms encoded by all three genomes (see Table II), and of homeologues from related cultivated (T. timopheevi) and wild (T. tauschii, Aegilops cylindrica) species (Mackie et al., 1996; Wan et al., 2002). A. AMINO ACID SEQUENCES

The subunits which occur most widely in bread wheat comprise between 627 (1Dy10) and 827 (1Dx5) amino acid residues, with Mr ranging from 67 476 to 88 128 (Table II). Furthermore, their amino acid sequences can be divided into three distinct parts, or domains, with the central domains consisting of highly repeated blocks of amino acids and ranging in length from 481 to 696 residues (Figure 5). The availability of complete amino acid sequences of HMW subunit proteins from cultivated wheat and related wild species of grasses allows a detailed comparison of their repeat motifs to be made. Because there are no apparent differences between the sequences of proteins from modern bread wheat, ancient cultivated wheats (T. timopheevi) and wild related species (A. cylindrica, T. tauschii), data from these species are combined in Table III and in the following discussion. A major difference between x-type and y-type subunits is that the former contain tripeptide, hexapeptide and nonapeptide motifs (Table III), whereas the y-type subunits contain only hexapeptides and nonapeptides (Table III). The tripeptides in x-type subunits always occur in tandem with hexapeptides, forming essentially a nine-residue repeat motif (see Figure 5).

TABLE II SUMMARY OF THE TOTAL NUMBER OF RESIDUES (RES), NUMBER OF CYSTEINE RESIDUES (CYS) AND NUMBER OF TRI-, HEXA- AND NONAPEPTIDE REPEAT UNITS (TRI, HEXA, NONA, RESPECTIVELY) IN EIGHT HMW SUBUNITS OF GLUTENIN FROM BREAD WHEAT AND THEIR THREE STRUCTURAL DOMAINS

N-terminal domain

C-terminal domain

Repetitive domain

Whole protein

Cultivar

Mr

Res

Cys

Res

Cys

Tri

Hexa

Nonaa

Res

Cys

Res

Cys

1Ax1 1Ax2* 1Bx7 1Bx17 1By9 1Dx2 1Dx5 1Dy10 1Dy12

Hope Cheyenne Cheyenne L88-69 Cheyenne Yamhill Cheyenne Cheyenne Chinese Spring

87680 86309 82865 80750 73518 87000 88137 67495 68696

86 86 81 81 104 88 89 104 104

3 3 3 3 5 3 3 5 5

681 666 647 611 538 687 696 481 493

0 0 0 0 1 0 1 1 1

15 16 4 4 0 20 23 0 0

65 67 66 64 56 73 73 47 49

23 23 25 23 22 21 21 21 21

42 42 42 42 42 42 42 42 42

1 1 1 1 1 1 1 1 1

809 794 770 734 684 817 827 627 639

4 4 4 4 7 4 5 7 7

aIncludes

two long (11/12 residue) degenerate repeats. Based on sequences reported in Halford et al. (1987, 1992); Anderson and Greene (1989); Anderson et al. (1989); Thompson et al. (1985); Sugiyama et al. (1985); Reddy and Appels (1993).

HMW SUBUNITS OF WHEAT GLUTENIN

Subunit

229

230

SH SH

P. R. SHEWRY ET AL.

SH SH SH

SH (1Dx5 only)

SH (1By - and 1Dy-type only)

Deleted in x-type

N-terminal domain

Repetitive domain

81-104 residues 3 cysteine (x-type) ) 5 cysteine (y-type) )

481-696 residues Comprises hexapeptides, nonapeptides and tripeptides (x-type only) 0/1 cysteine

SH

C- terminal domain 42 residues 1 cysteine

FIG. 5. Schematic summary of the sequences of x-type and y-type HMW subunits.

Similarly, in both types of subunit the hexapeptides occur either in tandem arrays or interspersed with nonapeptides, the latter forming a 15-residue repeat (Figure 6). Unlike the hexapeptides, the tripeptides and nonapeptides never occur in tandem arrays. The hexapeptides have the same consensus motif in x-type and y-type subunits (Pro.Gly.Gln.Gly.Gln.Gln.), but subtle differences are apparent in the frequencies of substitutions at different positions. Notably, the replacement of Pro with Ser at position 1 is much more common in x-type subunits, as is the replacement of Gln with Pro at position 6. However, the latter only occurs in hexapeptides within a 15-residue (i.e. 6 + 9) motif, not in tandemly arranged hexapeptides. Most, but not all, of the 15 amino acid repeats in x-type subunits have Pro in this position. Greater differences are present between the nonapeptides in the x-type and y-type subunits. Replacement of Tyr with His at position 2 and of Thr with Ala at position 5 are both common in y-type subunits but rare or absent in x-type. These two substitutions also usually occur together, resulting in two consensus sequences for y-type nonapeptides: Gly.Tyr.Tyr.Pro.Thr.Ser.Leu.Gln.Gln. and Gly.His.Tyr.Pro.Ala.Ser.Leu.Gln.Gln. The consensus motif of the x-type subunits also differs in having Pro in place of Leu at position 7. The data in Table III also show clear differences in the frequency of amino acid substitutions at different position in the motifs. In particular, Gln is highly conserved wherever it occurs (positions 3, 5 and 6 of the nonapeptide and 2 and 3 of the tripeptide), with the exception of the substitution with Pro at position 6 of hexapeptides present in 15-residue repeats of x-type subunits, as discussed above. These conserved glutamine residues can, therefore, be regarded as forming a glutamine backbone to the repetitive domains.

HMW SUBUNITS OF WHEAT GLUTENIN

231

TABLE III FREQUENCY OF OCCURRENCE OF DIFFERENT AMINO ACID RESIDUES IN EACH POSITION OF: (A) THE 347 HEXAPEPTIDE, 103 NONAPEPTIDE AND 81 TRIPEPTIDE REPEAT MOTIFS OF x-TYPE HMW SUBUNITS 1Ax1, 1Bx7 AND 1Dx5 (T. AESTIVUM); 1Ax (T. TIMOPHEEVI); 1Dx (Ae. CYLINDRICA). (B) THE 339 HEXAPEPTIDE AND 123 NONAPEPTIDE REPEAT MOTIFS OF y-TYPE HMW SUBUNITS 1Ay (NOT EXPRESSED), 1By9 AND 1Dy10 (T. AESTIVUM); 1Ay (T. TIMOPHEEVI); 1Cy AND 1Dy (Ae. CYLINDRICA); 1Dy (T. TAUSCHII)

Hexapeptides (%) A 1 Pro 62 Ser 26 Leu 10 Ile 1 Other 2

2 3 Gly 84 Gln 99 Ala 7 Other 1 Glu 4 Arg 3 Thr 2

Tripeptides (%)

4 Gly 75 Trp 9 Leu 7 Glu 4 Arg 2 Ala 1 Other 1

5 Gln 94 Leu 3 Other 3

6 Gln 80 Pro 15 Ser 2 Arg 1 Leu 1 Other 2

1 2 Gly 89 Gln 99 Asp 5 Arg 1 Ala 2 Arg 2 His 1

3 Gln 99 Arg 1

Nonapeptides (%) 1 2 Gly 84 Tyr 98 Arg 6 His 2 Glu 3 Trp 3 Val 2 Ala 1 Lys 1

3 Tyr 97 Asp 2 Phe 1

4 5 6 Pro 90 Thr 96 Ser 100 Leu 8 Ile 4 Ser 2

7 Pro 70 Ser 13 Leu 11 Ala 2 Glu 2 --- 2 Arg 1

8 Gln 88 Leu 8 Trp 3 Arg 1

9 Gln 94 Leu 4 Glu 2

7 8 Leu 54 Gln 97 Pro 21 His 3 Gln 19 Val 4 Gly 2 Ser 2 Ala 1

9 Gln 90 His 7 Glu 2 Stop 1

Hexapeptides (%) B 1 Pro 65 Ser 12 Leu 10 Ile 7 Thr 4 Gln 1 Other 1

2 3 Gly 92 Gln 96 Glu 6 Lys 4 Lys 2 Other 1

4 Gly 76 Glu 7 Trp 7 Arg 4 Ala 2 Val 2 Lys 1 Other 2

5 Gln 94 His 2 Lys 2 Other 1

6 Gln 94 Glu 2 His 2 --- 2

Nonapeptides (%) 1 Gly 96 Trp 2 Arg 1 Tyr 1

2 Tyr 54 His 41 Gln 5

3 Tyr 85 Cys 4 Asp 2 His 2 Ile 2 Phe 2 Arg 1 Asn 1 Glu 1

4 Pro 91 Leu 5 Arg 2 Ser 1 Thr 1

5 Thr 60 Ala 37 Ser 2 Ile 1

6 Ser 97 Tyr 2 Phe 1

Percentages may not add up to 100 because of rounding. Residues present at less than 1% are either included as ‘Other’ if together they add up to 1% or are not shown.

232

P. R. SHEWRY ET AL. 1Dx5

1Dy10

RYYPSVTCPQQ VSYYPGQASPQR PGQGQQ PGQGQQGYYPTSPQQ PGQWQQ PEQGQPRYYPTSPQQ SGQLQQ PAQGQQ PGQGQQGQQ PGQGQPGYYPTSSQLQ PGQLQQ PAQGQQ GQQ PGQAQQGQQ PGQGQQ PGQGQQGQQ PGQGQQ PGQGQQGQQ LGQGQQGYYPTSLQQ SGQGQPGYYPTSLQQ LGQGQSGYYPTSPQQ PGQGQQ PGQLQQ PAQGQQ PGQGQQGQQ PGQGQQGQQ PGQGQQ PGQGQPGYYPTSPQQ SGQGQPGYYPTSSQQ PTQSQQ PGQGQQGQQ VGQGQQAQQ PGQGQQ PGQGQPGYYPTSPQQ SGQGQPGYYLTSPQQ SGQGQQ PGQLQQ SAQGQKGQQ PGQGQQ PGQGQQGQQ PGQGQQGQQ PGQGQPGYYPTSPQQ SGQGQQ PGQWQQ PGQGQPGYYPTSPLQ PGQGQPGYDPTSPQQ PGQGQQ PGQLQQ PAQGQQ GQQ LAQGQQGQQ PAQVQQ GQR PAQGQQ GQQ PGQGQQGQQ LGQGQQGQQ PGQGQQGQQ PAQGQQ GQQ PGQGQQGQQ PGQGQQGQQ PGQGQQ PGQGQPWYYPTSPQE SGQGQQ PGQWQQ PGQGQPGYYLTSPLQ LGQGQQGYYPTSLQQ PGQGQQ PGQWQQ SGQGQHWYYPTSPQL SGQGQR PGQWLQ PGQGQQGYYPTSPQQ PGQGQQ LGQWLQ PGQGQQGYYPTSLQQ TGQGQQ SGQGQQGYY

GYYPGVTSPRQ GSYYPGQASPQQ PGQGQQ PGKWQE PGQGQQWYYPTSLQQ PGQGQQ IGKGQQGYYPTSLQQ PGQGQQGYYPTSLQH TGQRQQ PVQGQQ PEQGQQ PGQWQQGYYPTSPQQ LGQGQQ PRQWQQ SGQGQQGHYPTSLQQ PGQGQQGHYLASQQQ PGQGQQGHYPASQQQ PGQGQQGHYPASQQQ PGQGQQGHYPASQQE PGQGQQGQIPASQQQ PGQGQQGHYPASLQQ PGQGQQGHYPTSLQQ LGQGQQIGQ PGQKQQ PGQGQQ TGQGQQ PEQEQQ PGQGQQGYYPTSLQQ PGQGQQ QGQGQQGYYPTSLQQ PGQGQQGHYPASLQQ PGQGQ PGQRQQ PGQGQH PEQGKQ PGQGQQGYYPTSPQQ PGQGQQ LGQGQQGYYPTSPQQ PGQGQQ PGQGQQGHCPTSPQQ SGQAQQ PGQGQQ IGQVQQ PGQGQQGYYPTSVQQ PGQGQQ SGQGQQ SGQGHQ PGQGQQ SGQEQQGYD

FIG. 6. Amino acid sequences of the repetitive domains of typical x-type (1Dx5) and ytype (1Dy10) HMW subunits arranged to show their repeat unit structures.

HMW SUBUNITS OF WHEAT GLUTENIN

233

The x-type subunits are also highly conserved at positions 2, 3, 5, 6 and 9 of the nonapeptide, with the y-type subunits being more variable at these positions. Similarly, Ser is always present at position 6 of the x-type nonapeptides and is highly conserved (97%) in the y-type motifs. However, comparison of the properties of the amino acids that occur at specific positions provides no evidence for conservative substitutions (i.e. replacement with an amino acid with similar properties). Instead, the most common replacements can be accounted for by single base mutations in the DNA codons, with substitutions requiring two mutations occurring more rarely. For example, 55% of the x-type hexapaptides have Pro (codon CCA) at position 1, with 12% containing Leu (CTA), 30% Ser (TCA) but only 3% Ile (ATA). Nevertheless it is possible that at least some of the differences in the degree of conservation within motifs relate to the role of individual residues in determining protein structure. A number of studies have been carried out to determine the structures of the HMW subunits and their domains. However, it must be borne in mind that most of these have been carried out on proteins or peptides in the solution state rather than as hydrated solid, which is the “natural” state in protein bodies in the developing grain and in gluten and dough. B. SIZE AND SHAPE OF HMW SUBUNITS

Data from a range of studies indicate that the HMW subunits have an extended rod-like structure, both in solution and the hydrated solid state (Table IV). Field et al. (1987) used intrinsic viscosity measurements to calculate the dimensions of subunit 1Bx20 purified from pasta wheat, showing overall dimensions ranging from 49 × 1.8 nm in 50% (v/v) aqueous propan-1-ol to 62 × 1.5 nm in trifluroethanol. Small angle X-ray scattering (SAXS) has shown similar lengths but different diameters, of 56.7 × 8.0 and 78.6 × 6.3 nm for undefined subunits in 50% (v/v) aqueous propan-1-ol and 0.1 M acetic acid, respectively (Matsushima et al., 1992), and of 69 × 6.4 nm for subunit 1Bx20 in 50% (v/v) aqueous propan-1-ol (Thomson et al., 1999). Scanning tunnelling microscopy (STM) of the hydrated solid protein showed aligned rods of diameter about 1.8 nm (Miles et al., 1991). It can therefore be suggested that the diameters obtained by SAXS represent side-to-side aggregates rather than individual proteins. Atomic force microscopy (AFM) of subunit 1Dx5 deposited onto mica or graphite substrate has indeed shown such filaments, with a diameter of about 20 nm (Figure 7). An Mr 58 000 peptide derived from the central repetitive domain of subunit 1Dx5 also formed filamentous structures with diameters about 20 nm, with connection points every 45–50 nm (Humphris et al., 2000).

234

P. R. SHEWRY ET AL. TABLE IV

COMPARISON OF THE MOLECULAR DIMENSIONS DETERMINED AND CALCULATED FOR HMW SUBUNIT MOLECULES

Dimensions (nm) Method

Subunit

State

Viscometry Viscometry

1Bx20 1Bx20

49 50

1.8 1.75

1 1

Viscometry SAXS SAXS SAXS STM Modelling Model 1 Model 2 Model 3

1Bx20

50% propan-l-ol 0.05 M acetic acid/ 0.01 M glycine Trifluoroethanol 0.01 M acetic acid 50% propan-l-ol 50% propan-l-ol hydrated solid

62 79 57 69 –

1.5 6.3 8.0 6.4 1.8

1 2 2 3 4

冧

53.5 37.7 54.5

1.7 2.0 2.1

5 5 5

1Bx20 1Bx20

冧

650a residues

modelled in solvent sheath

Length

Diameter Reference

aModelled

as repetitive domain comprising 650 residues of a 9 + 6 + 9 residue motif. References: 1, Field et al. (1987); 2, Matsushima et al., 1992; 3, Thomson et al. (1999); 4, Miles et al. (1991); 5, Parchment et al. (2001).

C. STRUCTURES OF HMW SUBUNIT TERMINAL DOMAINS

The N- and C-terminal domains of the HMW subunits have proved to be difficult to study in isolation from the central repetitive domain. Early structure predictions indicated that these domains were essentially globular in structure, being rich in α-helix (Tatham et al., 1984, 1985). More recently, van Dijk et al. (1998) have predicted that residues 1–33 of subunit 1Dx5 (and the corresponding regions of subunits 1Ax1, 1Ax2* and 1Bx7) form α-helices while residues 40–48, 52–59, 64–67 and 75–84 form β-sheet. This is supported by more detailed prediction and modelling studies reported by Köhler et al. (1997). They proposed that residues 5–32 of subunit 1Dx5 form a continuous α-helix, as shown in Figure 8. In contrast, several short sections of α-helix are predicted for subunit 1Bx7 (residues 6–13, 16–20, 24–26), with an inverse β-turn between residues 14 and 16 allowing the formation of an intra-chain disulphide bond between the cysteine residues at positions 10 and 17 (Figure 8) (see below for a discussion of disulphide bond formation). It has not so far proved possible to isolate a peptide corresponding to the N-terminal domain of an authentic HMW subunit, but van Dijk et al. (1998) have reported the expression in Escherichia coli and

HMW SUBUNITS OF WHEAT GLUTENIN FIG. 7. Atomic force microscopy of reduced and alkylated subunit 1Dx5 deposited from 0.05 M acetic acid on a highly orientated pyrolytic graphite (HOPG) substrate shows network formation. Taken from Humphris et al. (2000), with permission.

235

236

P. R. SHEWRY ET AL.

FIG. 8. Molecular models of the N-terminal domains of subunits 1Dx5 and 1Bx7, showing the backbone only in cartoon form. Sulphur atoms (10, 25 and 40 in 1Dx5; 10, 17 and 32 in 1Bx7) are shown in grey. Taken from Köhler et al. (1997), with permission.

characterization of a peptide corresponding to the N-terminal domain of subunit 1Dx5. The peptide was insoluble in water except in the presence of SDS and circular dichroism spectroscopy (in 0.1% SDS) indicated the presence of 26% α-helix and 33% β-sheet at pH 8.17 and 35% α-helix and 36% β-sheet at pH 3.59. The authors concluded that the insolubility of the N-terminal domain was responsible for the solubility properties of the whole subunits. However, the expressed peptide differed from the native N-terminal domain in the presence of a 21-residue signal sequence and it is possible that this hydrophobic sequence may have affected the structure and properties of the peptide, including the solubility. Tatham et al. (1984) also predicted that the short (42-residue) C-terminal domains of the HMW subunits were α-helical in structure. Bekkers et al. (1996) subsequently synthesized a peptide corresponding to the C-terminal domain of subunit 1Dx5 and showed that it was readily soluble in aqueous buffers over a wide pH range and adopted a random coil structure when dissolved in water. Nuclear magnetic resonance (NMR) spectroscopy in the “structure-inducing” solvent 40% (v/v) aqueous trifluoroethanol allowed a low-resolution structure to be determined which was “molten globule” like, with α-helical regions formed by residues 5–20 and 26–32. It is not possible to conclude whether this structure is related to that adopted in the hydrated solid state (i.e. as in gluten) and it is also possible that the

HMW SUBUNITS OF WHEAT GLUTENIN

237

structure of this domain, and of the N-terminal and central domains, is affected by interactions with other domains of the subunit or with other gluten proteins. D. STRUCTURE OF HMW SUBUNIT REPETITIVE DOMAINS

The repetitive domains form the largest part of the HMW subunits, ranging from 481 to 696 residues. Initial studies using secondary structure prediction (Tatham et al., 1984, 1985), circular dichroism (CD) and infrared (IR) spectroscopies of the whole protein (Tatham et al., 1985) and synthetic peptides based on the repeat motifs (Tatham et al., 1990) indicated β-turns as the dominant structural feature. β-Turns were predicted over residues QPGQ, QQGY, YPTS and SPQQ, and it was proposed that regular β-turn predictions led to the formation of a spiral structure, similar to the β-spiral described for a synthetic polypentapeptide based on the repeat motif present in the connective tissue protein elastin (Tatham et al., 1985). Van Dijk et al. (1996a) used 2D-NMR, CD and IR of synthetic cyclic peptides to confirm the presence of β-turns: a type II β-turn at QPGQ, type I β-turns at YPTS and SPQQ and a type I or II β-turn at QQGY. They also reported cis/trans proline isomerism, with 50% of proline residues in the cis conformation in YPTS, and the other proline residues being more than 90% trans-conformation. Conversion of the cis form in YPTS to the transconformation destabilized the type I turn, but increased the stability of the β-turns at SPQQ and QQGY. Van Dijk et al. (1996b) also studied the solution structures of HMW subunits 1Bx6 and 1Bx7 and of an Mr 16 802 heterologously expressed peptide from the central repetitive domain of subunit 1Dx5. Using CD and IR, they concluded that the structure was compatible with β-turns stabilized by hydrogen bonds both within and between turns. More recently, variable temperature CD studies of an Mr 58 000 peptide derived from the central repetitive domain of subunit 1Dx5 indicated the presence of β-turn structures in equilibrium with a poly-Lproline II-like structure (an extended hydrated structure), with supporting evidence for this coming from IR studies (Gilbert et al., 2000). Matsushima et al. (1990) attempted to model the nonapeptide repeat motif of an x-type subunit (GYYPTSPQQ), but concluded that as the repeat contained two proline residues, there were many degrees of freedom and hence many models could be constructed. They modelled structures with two or three β-turns per repeat and calculated diameters of 1.4–1.6 nm, but reported no detailed structures. Kasarda and coworkers have developed models based on hexapeptide and nonapeptide repeats and a hexanonapeptide (15-mer) repeat, fitted to the hexapeptide template (Kasarda, 1994; Kasarda et al., 1994). Using energy minimization and molecular

238

(A)

P. R. SHEWRY ET AL.

(B)

(C)

FIG. 9. Space filling models of backbone structures of hexapeptide + nonapeptide + hexapeptide repeat motifs (PGQGQQ GYYPTSPQQ PGQGQQ): (A) with type II β-turns at QPGQ, YPTS, SPQQ and QGQQ; (B) with type II β-turns at QGYY and QPGQ and type-I/III β-turns at SPQQ and YPTS; (C) with distance-based β-turns at QPGQ, YPTS and SPQQ, β-sheet at GQGQQGY. Taken from Parchment et al. (2001) with permission.

dynamics they reported that when type II β-turns were used a distorted spiral with a flat ribbon shape resulted, but that this was unstable and other turn types did not improve the stability. When inverse γ-turns (a threeresidue turn) were modelled a more stable spiral structure resulted, with a diameter of about 2.4 nm and a pitch of about 1 nm. The stability of the spiral arose from hydrogen bonding of the glutamine side chain amide groups to the backbone amide groups and to other glutamine side chains. However, the model did not include interactions with water, which might break up some or all of the hydrogen bonding that was proposed (Kasarda et al., 1994). Parchment et al. (2001) used structure prediction and molecular dynamics to generate three alternative spiral structures based largely on β-turns with diameters of about 2 nm. The β-turns were placed over residues identified by prediction and spectroscopic studies and spiral structures were generated in the presence of a water shell (Figure 9). The simplest model placed a type II β-turn at QPGQ and turns were forced at YPTS, SPQQ and QGQQ in the repeat motif, resulting in a spiral with a diameter of 1.7 nm and a pitch of 1.3 nm (Figure 9). In the second model type II β-turns were located at QPGQ and QGYY and type I/III β-turns at SPQQ and YPTS, resulting in a spiral structure with a diameter of about 2 nm and

HMW SUBUNITS OF WHEAT GLUTENIN

239

pitch of 1.6 nm (Figure 9). The third model did not use standard β-turn types, but the positions of β-turns were based on distance criteria. These were located at QPGQ, YPTS and SPQQ, with the sequence GQGQQGY forming a β-sheet structure. The resulting spiral was flattened, with a cross-section of about 1.7 × 2.5 nm (Figure 9). Calculated dimensions for the models are shown in Table IV. Arkin et al. (2000) used multicanonical simulation to model the structures of five common tetrapeptide sequences (QPGQ, QSGQ, YPTS, SPQQ and QPGY) in the central repetitive domains. They found that QOGQ and QPGY had the highest probabilities for β-turns, with lower probabilities for the others. They also considered the probabilities of γ-turn formation and found probabilities in all five tetrapeptides. They concluded that these may make a contribution to the overall structures of the central repetitive domains. Arkin et al. (2001) also modelled two hexapeptide repeats, PGQGQQ and SGQGQQ; these agreed with about 40% of the total occurrence of β-turns predicted by Tatham et al. (1985), but also concluded γ-turns may contribute to the proposed spiral structure. The studies discussed above were either carried out on the HMW subunits in dilute solution, or were modelled in vacuo or hydrated in a water shell. However, the environment of the HMW subunits in dough and gluten is as a hydrated solid protein mass. Belton et al. (1995) used Fourier transform infrared (FT-IR) and NMR spectroscopy to study the hydration behaviour of subunits and found that the proportions of β-turns and β-sheet varied in relation to the water content. Thus, the content of intermolecular β-sheet structure increased when the protein was hydrated from the dry solid. Gilbert et al. (2000) also expressed an Mr 58 000 peptide from the central repetitive domain of subunit 1Dx5 in E. coli and studied its behaviour during hydration using FT-IR. They reported that β-turn rich structures formed in the dry and hydrated solid states, with significant contents of intermolecular β-sheet structure. In comparison with intact subunits, the Mr 58 000 peptide showed a lower propensity to form β-sheet structure, indicating that the N- and C-terminal domains may play a role in assembling the molecules to allow β-sheet formation to occur in the central repetitive domains. NMR studies indicate that the central repetitive domain is flexible, with increasing hydration resulting in increased flexibility (Belton et al., 1995; Alberti et al., 2001). However, two populations of glutamine residues were identified: one in a mobile environment that was tentatively identified with β-turn conformations, and a second population that was more hindered, possibly by hydrogen bonding, in protein segments containing glycine residues and possibly adopting a β-sheet conformation (Alberti et al., 2001).

240

P. R. SHEWRY ET AL. E. CROSS-LINK FORMATION BETWEEN HMW SUBUNITS

The HMW subunits are only present in oligomers or polymers that are stabilized by inter-chain disulphide bonds. However, little is known about their organization within these polymers. Partial reduction of gluten with reducing agents followed by stabilization of the free cysteine sulphydryl groups with cystamine diHCl has been shown to result in the release of HMW subunit dimers (Lawrence and Payne, 1983; Werner et al., 1992). These include a preponderance of x–y dimers, and in particular dimers involving 1Dy with 1Ax, 1Bx or 1Dx subunits. These studies indicate that x–y HMW subunit dimers may act as “building blocks” in glutenin polymers. Evidence for head-to-tail bonds between HMW subunits was reported by Tao et al. (1992). They prepared a glutenin-enriched fraction from flour and digested this with the endoproteinase LysC. Fractionation of the digest led to the isolation of two peptides containing intact disulphide bonds, both of which consisted of the C-terminal part of subunit 1Bx17 linked to the N-terminal domain of subunit 1Dy10. However, the precise cysteine residue involved in subunit 1Dy10 could not be identified. The isolation of x–y dimers and the identification of head-to-tail disulphide bonds are both consistent with the suggestion by Graveland et al. (1985) that alternating x-type and y-type subunits form the “backbone” structure of the glutenin polymers. Further disulphide bonds involving HMW subunits have been reported by Köhler and coworkers (Köhler et al., 1991, 1993, 1994; Keck et al., 1995). These include one intra-chain bond within the N-terminal domain of subunit 1Bx7 (involving Cys 10 and Cys 17) (Figure 8) and one crosslink between a cysteine residue within the repetitive domain of subunit 1By9 (Cys 564) or 1Dy10 (Cys 507) and a low molecular weight subunit of glutenin. This could lead to the presence of LMW subunits as “branches” on the HMW subunit backbone, as proposed by Graveland et al. (1985).

x-type

LMW y-type y-type

FIG. 10. Schematic summary of disulphide bonds identified involving HMW subunits. Based on details in Tao et al. (1992), Köhler et al. (1991, 1993, 1994) and Keck et al. (1995).

LMW y-type x-type

LMW y-type y-type y-type 1Dx5 LMW

x-type

HMW?

N- terminal domain C-terminal domain

repetitive domain disulphide bonds

LMW subunits

241

FIG. 11. Hypothetical structure for wheat glutenin polymers, based on mapped disulphide bonds (Figure 10). Not all cysteine residues are included. The additional cysteine residue present in the repetitive domain of subunit 1Dx5 is proposed to link to another HMW subunit, based on the effect on gluten rheology of transformation with additional gene copies.

HMW SUBUNITS OF WHEAT GLUTENIN

LMW

x-type

242

P. R. SHEWRY ET AL.

In addition, two disulphide bonds were identified between the adjacent cysteine residues (Cys 44, Cys 45) in the N-terminal domains of two ytype subunits (1By9 and/or 1By10), linking the two subunits in parallel. Since hetero- or homodimers comprising two y-type subunits were not detected by Werner et al. (1992) or Lawrence and Payne (1983), it can be concluded that such cross-links between y-type subunits are either readily reduced or occur at low frequency. Based on these studies, our current knowledge of the disulphide bonds formed by HMW subunits is summarized in Figure 10. This information can be used to propose a wider structural model for wheat glutenin polymers, based on x–y subunit dimers with branches to LMW subunits via y-type subunits (Figure 11). A more detailed discussion of the disulphide bond structure of gluten proteins is provided by Shewry and Tatham (1997). Although disulphide bonds are classically considered to be the only type of covalent cross-links in wheat glutenin polymers, Tilley et al. (2001) have recently reported that a novel type of cross-link may occur. These are tyrosine cross-links formed specifically between the TyrTyr sequences present in the HMW subunit repeat motifs. Analysis of synthetic peptides (Tyr.Tyr.; Tyr.Tyr.Pro.Thr.Ser. and Gln.Gln.Gly.Tyr.Tyr.Pro.Thr.Ser.) demonstrated that the formation of such cross-links was enhanced by KBrO3 which acts as an “improver” in breadmaking. The authors suggest that tyrosine bonds formed during mixing and baking contribute to the structure of the gluten network. IV. EXPERIMENTAL EVIDENCE FOR THE ROLE OF HMW SUBUNITS IN DOUGH MIXING AND GLUTEN VISCOELASTICITY

Experimental evidence for the role of the HMW subunits in determining the viscoelastic properties of doughs comes from a range of studies in which gluten fractions, individual proteins and peptides have been studied in reconstituted systems. A. GLUTEN FRACTIONATION AND RECONSTITUTION

A valuable approach to elucidate the role played by the different protein components in gluten and dough functionality is to fractionate and reconstitute gluten or flour and to observe the effects of adding or omitting components on the technological or rheological properties. The earliest work used this approach to investigate the role of gliadinand glutenin-rich fractions. In this way Finney and coworkers (for a synthesis, see Finney, 1985) showed that glutenin was involved in the mixing requirement

HMW SUBUNITS OF WHEAT GLUTENIN

243

and that glutenins extracted from good or poor quality wheats had different properties. Further fractionation of glutenin into acid-soluble and acid-insoluble fractions, corresponding broadly to components with different polymer sizes and/or aggregation behaviour, showed a positive relationship between the acid-insoluble (or high molecular mass) glutenin aggregates and dough mixing strength (Orth and Bushuk, 1972; Huebner and Wall, 1976; Hamada et al., 1982). More recently, it has been shown that increasing the glutenin to gliadin ratio (while maintaining a constant total protein content) increased the mixing time and the overmixing tolerance of dough. Extensibility was reduced by the addition of glutenin whereas resistance to extension was increased (Uthayakumaran et al., 1999), corresponding to higher rupture viscosity and a lower rupture strain when dough was submitted to uniaxial elongation. In contrast, gliadin contributed to increased extensibility of dough by lowering the viscosity and enhancing the rupture strain (Uthayakumaran et al., 2000a). Rheological studies have also been carried out on isolated gliadin and glutenin fractions separated by the Osborne procedure. Gliadins were described as an elasto-viscous liquid whereas glutenins behaved like a viscoelastic solid (Khatkar et al., 1995). The addition of protein fractions to gluten and gluten reconstitution with varied ratios of gliadins to glutenins showed that the storage (related to elasticity) and the loss (related to viscosity) moduli (G′ and G″, respectively) of gluten were positively related to its content of glutenin and that increasing the glutenin content increased gluten elasticity (Khatkar et al., 1995; Janssen et al., 1996), However, the results of this approach could be invalidated if the extraction procedure modified protein functionality (MacRitchie, 1985; Skerritt et al., 1996) and it was considered that the extraction of gliadin with alcohol/ water in the classical Osborne procedure could irreversibly change the aggregation of the glutenin fraction (Hoseney et al., 1969). MacRitchie (1987) therefore proposed a new method for the sequential extraction of gluten proteins in dilute hydrochloric acid, a procedure which was considered to preserve their original functionality. The ten fractions obtained in this way contain mixtures of gliadin and glutenin in varying proportions, with the size distribution of the glutenin polymers also varying between the fractions (Lundh and MacRitchie, 1989). The effects of these fractions on the mixing properties and the storage modulus (G′) of the dough, when added to a control flour, were related to their contents of glutenin polymers and to their aggregation state (MacRitchie, 1987; Eliasson and Lundh, 1989). The same fractionation procedure was also applied to gluten extracted in the laboratory. The fractions were analysed by size-exclusion highperformance liquid chromatography (SE-HPLC) under dissociating conditions (Cornec et al., 1994), showing that the most readily extracted material

244 P. R. SHEWRY ET AL. FIG. 12. Viscoelastic properties of gluten subfractions differing in their content and size distribution of glutenin polymers. The content of glutenin and the proportion of the largest glutenin polymers increase from fractions F3 to F8. Dynamic rheological assays in shear were performed at 20°C on gluten fractions fully hydrated with water using a Carri-Med CSL 100 constant stress rheometer (cone-plate geometry: cone angle, 4°; diameter, 2 cm; amplitude of strain at all frequencies 3%). (A) Storage (G′) and loss (G″) moduli: closed symbols, G′; open symbols, G″; diamonds, F3; circles F4; squares, F6; triangles, F8. (B) Tangent of the loss angle. Symbols: diamonds, F3; circles F4; squares, F6; triangles, F8. Taken from Cornec et al. (1994), with permission.

HMW SUBUNITS OF WHEAT GLUTENIN

245

contained mostly gliadin monomers. With decreasing extractability, the fractions became enriched in glutenin polymers of medium and large size with the HMW subunits being concentrated in the largest, most insoluble polymers. The mechanical spectra of the fully hydrated fractions were also recorded in a dynamic assay in shear (Cornec et al., 1994). The viscoelastic properties of pure gliadin were too low to be measured with the Carrimed SL 100 stress rheometer but an increasing content of glutenin polymers was associated with increases in both the storage (G′) and loss (G″) moduli but with a decrease in the G″/G′ ratio (tangent δ), indicating that the elasticity of the material increased (Figure 12). For all the fractions except those containing the lowest amount of glutenin, G′ was higher than G″ over the whole range of frequencies (with a cross-over of G′ and G″ being observed for gliadin-rich fractions, as shown for F3 in Figure 12). The frequency range corresponding to the elastic plateau was also shifted towards higher values when the content of glutenin polymers increased. This behaviour is typical of a transient network structure and the connectivity of the network, measured by the height of the elastic plateau (GN0), increased when the extractability of the fractions decreased. Thus, GN0 was between about 3000 and 30 000 N m–2 for the glutenin-rich fractions and was strongly positively correlated with the content of large glutenin polymers (measured as the excluded peak in SE-HPLC), but not with that of medium-size glutenin polymers. Therefore, it was assumed that the density of transient cross-links in hydrated gluten was determined primarily by the proportion of the largest glutenin polymers. In this network, monomeric gliadins act as plasticizing elements (Cornec et al., 1994). Using a different fractionation procedure, Tsiami et al. (1997a, b) confirmed that the major variable affecting the rheological properties of gluten proteins, and particularly of glutenins, is the size of the concatenations. B. INCORPORATION OF PROTEIN FRACTIONS INTO DOUGH

In other studies, different types of isolated gliadins or glutenin subunits have been incorporated into gluten and dough to discriminate between the effects of individual proteins and to relate functionality to particular structural features. The addition of isolated gliadins confirmed that monomeric prolamins have a weakening effect on dough (Fido et al., 1997), by increasing the extensibility and decreasing resistance. However, the extent of the effect depended on the gliadin type, and conflicting results were reported by different authors (Fido et al., 1997; Uthayakamuran et al., 2001), with both molecular mass and hydrophobicity being reported to determine gliadin functionality. Such experiments are relatively easy to perform with gliadins because they are present in the flour as monomeric proteins. A

246

P. R. SHEWRY ET AL.

simple mixing after addition of the protein into a control flour is sufficient to observe the changes in dough properties induced by the added components. In the case of the glutenin subunits a more complicated procedure must be used, because the subunits must be incorporated by covalent disulphide bonds into glutenin polymers in order to express their functionality in the same manner as in the native gluten. Special procedures to reduce and reoxidize flour proteins were therefore developed and applied in solution (Schropp et al., 1995; Veraverbeke et al., 2000a, b) or in dough (Békés et al., 1994a; Uthayakamuran et al., 2000b). When reoxidation of HMW subunits was performed in solution, only partial polymerization was obtained, with the extent depending on protein concentration and type and on the concentration of oxidizing agent (Schropp et al., 1995; Veraverbeke et al., 2000a). About 20–40% of the initial subunits remained as monomers, with intramolecular instead of intermolecular bonds being formed. Some conflicting results were also reported concerning the respective oxidizing efficiencies of KBrO3, KIO3 and H2O2 (Schropp et al., 1995, Veraverbeke et al., 2000a, b). Although no differences in the degree or pattern of polymerization were observed when a mixture of subunits 1Bx5, 1Bx7, 1By9 and 1Dy10 was compared with a mixture of subunits 1Dx2, 1Bx6, 1By8 and 1Dy12 (Schropp et al., 1995), y-type subunits were found to be less highly polymerized than x-type subunits (Veraverbeke et al., 2000b) and subunit 1Dy10 was found to form the lowest proportion of polymers (Antes and Wieser, 2001b). Antes and Wieser (2001b) also noted that subunit 1Dx5, which contains an additional cysteine residue, yielded a higher proportion of polymers than subunit 1Bx7, whereas the opposite was noted by Veraverbeke et al. (2000a). It is also notable that LMW glutenin subunits showed a higher propensity to polymerize in vitro than HMW subunits and that the polymers formed by HMW and LMW subunits did not differ in their size distribution (Veraverbeke et al., 2000b). These results are not consistent with those obtained with native gluten, where the largest glutenin polymers contain a higher proportion of HMW subunits (Payne and Corfield, 1979; MacRitchie, 1989; Lundh and MacRitchie, 1989), nor with results obtained with lines depleted in HMW subunits, where the absence of HMW subunits increased the extractability and decreased the average size of the glutenin polymers (Popineau et al., 1994; Gupta et al., 1995; Lefebvre et al., 2000). These conflicting results indicate that it is difficult to synthesize glutenin polymers in vitro from isolated subunits and that it may be unwise to draw conclusion about glutenin functionality in native gluten from such experiments. Nevertheless, in vitro polymerized glutenins have been incorporated into gluten and dough to test their effect on the rheological properties. The addition of glutenin polymers composed solely of HMW subunits clearly

HMW SUBUNITS OF WHEAT GLUTENIN

247

increased the resistance of gluten and dough to extension, provided some free thiol groups were available to enable the added polymers to be linked to endogenous proteins (Schropp and Wieser, 1996; Antes and Wieser, 2001a). If no thiol groups were available, extensibility was decreased. Incorporation of polymers comprising LMW glutenin subunits also resulted in decreased extensibility, while the incorporation of mixed polymers decreased both the resistance and extensibility (Antes and Wieser, 2001a). In other experiments, isolated glutenin subunits were directly incorporated into the dough using a partial reduction/reoxidation procedure, initially developed by Békés et al. (1994b). Provided the conditions were carefully controlled, reduction by dithiothreitol (DTT) could be completely reversed by addition of KIO3, as judged by the mixing properties. However, the optimal conditions of reduction/reoxidation for mixing studies were recently shown to be unsuitable for extension and baking testing and specific procedures were therefore developed for each test (Uthayakumaran et al., 2000b). Extension tests were also shown to be more discriminating than mixing tests for determining the rheological properties of the treated doughs with the shape of the resistance vs. extensibiliy curve being especially sensitive to redox conditions. Even under the conditions determined as “optimal”, the overall shape of the curve after reduction and reoxidation was different from the original one, indicating that the structure of the protein network and the rheological behaviour of the dough were not restored to an identical condition, even if maximal resistance and extension were similar. Incorporation of HMW subunits always had a positive effect on strength, whereas the effects of incorporation of LMW subunits depended on the type of subunit added (Sissons et al., 1998). Increasing the HMW : LMW subunit ratio improved all parameters in the mixing test, with increased resistance and decreased extensibility of the dough. These changes were correlated with an increase in the amount of “unextractable polymeric proteins” (i.e. the largest glutenin polymers) (Uthayakumaran, 2000b). C. INCORPORATION OF PURIFIED HMW SUBUNITS INTO DOUGH

Different HMW glutenin subunits have been incorporated into control doughs using reduction/reoxidation procedures. Incorporation of subunit 1Bx20 from the durum wheat cultivar Bidi 17 increased dough strength (i.e. increased mixing time and peak resistance and, decreased the breakdown in Mixograph tests). The amount of the largest glutenin polymers was also increased and the 1Bx20 subunit was recovered mainly from this polymer fraction (Békés et al., 1994b). On the other hand, the simple addition of the subunit (with no reduction and reoxidation to form inter-

248

P. R. SHEWRY ET AL.

chain disulphide bonds) resulted in a decrease in dough strength. This demonstrated that increasing the proportion of HMW glutenin subunit in glutenin polymers was associated with increases in their size and/or aggregation state, and significantly enhanced the technological properties of the flour. Using the same method, the effects of partially purified 1Dx and 1Dy subunits expressed in E. coli were investigated. Incorporation of the subunits increased dough strength, but the effect of 1Dx subunits was greater than that of 1Dy subunits. When incorporated in pairs, the most efficient associations were 1Dx + 1Dy, and subunit 1Dx5 combined with 1Dy10 resulted in a greater enhancement of strength than subunit 1Dx2 with 1Dy10 (Békés and Gras, 1994). The subunit pair 1Dx5 + 1Dy10 was also superior to the 1Dx2 + 1Dy12 pair. These results were confirmed by mixing, extension and baking tests on dough in which single subunits or

A Mixing time(s)

600 500 400 300 200

B Resistance breakdown (%)

5+10 10

7

7+8 8

10

5+10

8

Banks 5

7

7+8

0

Hartog 5

100

40 30 20

10

5+10

8

Hartog 5

7+8

7

10

5

5+10

Banks

8

7+8

0

7

10

FIG. 13. Effect of incorporation of HMW glutenin subunits 1Bx7, 1By8, 1Dx5 and 1Dy10 on the mixing properties of two base flours, Banks and Hartog. The purified HMW glutenin subunits labeled on the abscissa were incorporated into the base flours using the reduction–oxidation procedure of Békés and Gras (1994). (A) Mixing time; (B) resistance of dough to breakdown. Taken from Uthayakumaran et al. (2000c), with permission.

HMW SUBUNITS OF WHEAT GLUTENIN

249

pairs of subunits, extracted from flours or expressed in E. coli, were incorporated (Uthayakumaran et al., 2000c) (Figure 13). These experiments showed that chromosome 1B-encoded subunits were less able to increase dough strength than chromosome 1D-encoded subunits. When single subunits were compared, subunit 1Dx5 had a greater effect than subunit 1Bx7 while subunit 1Dy10 had a greater effect than subunit 1By8. Synergy was also noted for the chromosome 1D-encoded subunits in that the subunit pairs 1Dx5 + 1Dy10 or 1Dx2 + 1Dy12 had greater positive effects than each single component incorporated at the same total concentration (Békés and Gras, 1994; Uthayakumaran et al., 2000c). This was not the case for the chromosome 1B-encoded subunits 1Bx7 and 1By8. The greatest effect of subunits 1Dx5 + 1Dy10 was also demonstrated using baking tests (Uthayakumaran et al., 2000c). The results obtained with the subunit pairs are consistent with the presence in gluten of dimers composed of x- and y- subunits (see above) and also provide experimental support for the quality scores of glutenin subunits based on correlations between the HMW composition of genotypes and their technological properties (Payne et al., 1987b; Branlard et al., 1992). D. INCORPORATION OF HMW SUBUNIT PEPTIDES INTO DOUGH

In order to determine whether the HMW subunit repetitive domain was alone able to affect the mixing properties of dough, Buonocore et al. (1998) expressed a series of repetitive peptides in E. coli and then incorporated them into dough using the 2 g Mixograph. The peptides had Mr of about 58 000 and consisted of residues 103 to 643 of subunit 1Dx5 (comprising most of the repetitive domain), with the addition of short linking sequences at the N- and C-termini. In addition, a series of mutants were constructed with 0, 1 and 2 cysteine residues as substitutions close to the N- and/or C-termini. When incorporated into dough, the peptide with two additional cysteine residues at the N- and C-termini (called 2 + 2) resulted in substantial increases in the mixing time and peak resistance (Figure 14), demonstrating that the nonrepetitive N- and C-terminal domains of the HMW subunits are not essential for the formation of the viscoelastic glutenin polymers. V. MANIPULATING HMW SUBUNIT COMPOSITION

Although it is possible to obtain information on the functional properties of individual HMW subunits using in vitro approaches, it is important to

250

A 89

Subunit 1Dx5 M r 88,000

NH 2 SH SH SH SH 103

SH

Mr 58,000 peptide

106 112 Cys Cys

COOH 643

634 Cys

640 Cys

B

Expression in E.coli

P. R. SHEWRY ET AL.

C

785 827

REPEATS

Incorporation into dough for Mixograph

a

b

c

d

e

FIG. 14. The effect of an Mr 58 000 repetitive peptide from subunit 1Dx5 on the mixing properties of dough. A subclone encoding residues 103 to 643 from subunit 1Dx5 mutated to encode cysteine residues at positions 106, 112, 634 and 640 (A) was expressed in E. coli, purified (B) and incorporated into dough using the 2g Mixograph (C). Tracks in part (B) are a, Mr marker proteins; b, total proteins from uninduced E. coli cells; c, total proteins from E. coli cells after induction to express the Mr 58 000 peptide (arrowed); d, the proteins extracted from induced cells of E. coli with 70% (v/v) ethanol and e, the purified Mr 58 000 peptide. Taken from Buonocore et al. (1998) with permission.

HMW SUBUNITS OF WHEAT GLUTENIN

251

also determine their properties when incorporated into glutenin polymers in vivo. In order to do this it is advantageous to compare the properties of individual subunits when expressed in the same genetic background. This can be achieved by classical crossing to produce near isogenic lines, or by transformation to insert additional genes. These approaches can be regarded as complementary, with the transgenic approach being used to generate variation beyond that which is available naturally, for example, to add additional copies of expressed genes or to express single genes or novel combinations of single genes which are usually inherited as tightly linked allelic pairs. A. NEAR ISOGENIC LINES

The demonstration by Payne and coworkers (see review by Payne, 1987) that specific HMW subunit alleles are associated with good and poor breadmaking performance has led to their selection in plant breeding programmes. Although work was initially carried out on defined crosses (Payne et al., 1981a), the correlations demonstrated have since been confirmed and extended using collections of genotypes and different genetic stocks such as biotypes, null lines and near isogenic lines. Of these approaches, the development of near isogenic lines (NIL) is of greatest value for studies of structure–functionality relationships as they contain different combinations of HMW subunits transferred into a common genetic background by repeated back-crossing. A number of such series of lines are now available, in addition to near isogenic pairs such as the recently reported transfer of a novel 1Bx subunit into the breadwheat cultivar Fiorello (Margiotta et al., 2000). 1. The Sicco NIL Payne et al. (1987a) have described the production of lines in the cultivar Sicco differing in their number (two to five) and composition of HMW subunits. These contained the following subunit combinations: Sicco 2 + 12 1Ax null 1D null 1Ax/1D null

Glu-A1 1 1 null 1 null

Glu-B1 7+9 7+9 7+9 7+9 7+9

Glu-D1 5 + 10 2 + 12 5 + 10 null null

Subsequently, Rogers et al. (1991) used the same Sicco background to

252

P. R. SHEWRY ET AL.

produce three additional NIL differing in the absence of x-type or y-type subunits. They had the following compositions: Glu-A1 1 1 1 1

Sicco 1Dx null 1Dy null 1Bx null

Glu-B1 7+9 7+9 7+9 null + 8

Glu-D1 5 + 10 null + 36 2 + null 5 + 10

2. The Gabo NIL Lawrence et al. (1988) combined null mutations at all three Glu-1 loci to develop NIL in the cultivar Gabo with subunit numbers ranging from zero to five, as listed below. Glu-A1 1 null 1 1 null null 1 null

Glu-B1 17 + 18 17 + 18 null 17 + 18 17 + 18 null null null

Glu-D1 5 + 10 5 + 10 5 + 10 null null 5 + 10 null null

3. The Galahad NIL Payne and Seekings (1996) described a series of NIL containing only single 1B subunits (either 1Bx6, 1Bx7 or 1By8) in the cultivar Galahad. These four lines have been used for detailed studies of the functional properties of HMW subunits as discussed below. In addition, two more recently produced series of NIL will undoubtedly also prove to be valuable in this respect. 4. The Pegaso NIL Margiotta et al. (2000) have repored the development of lines in the bread wheat cultivar Pegaso (Figure 15). They include lines expressing novel xtype and y-type subunits with higher and lower mobilities on SDS-PAGE than the more widely occurring alleles, which may reflect the presence of differences in the length of their repetitive domains. Furthermore, a GluA1 locus expressing genes for both x-type and y-type subunits was

HMW SUBUNITS OF WHEAT GLUTENIN

253

FIG. 15. SDS-PAGE of HMW subunits from near isogenic lines of the bread wheat cultivar Pegaso, showing the expression of novel 1Ax (21*), 1Ay (21*y), 1Dx (2.2, 2.2*) and 1Dy (12*, 121) subunits.

introduced from the wild tetraploid wheat T. dicoccoides, allowing a NIL expressing six subunits to be produced. 5. Single subunit lines Lafiandra et al. (2000) have reported the production of a series of lines expressing single x-type (1Ax1, 1Dx2, 1Bx7) or y-type (1By8, 1Dy10, 1Dy12) subunits in the background of Gabo, by combining spontaneous mutations identified in collections of old wheat varieties. Crossing of these lines will allow the properties of novel combinations of x-type and y-type subunits (e.g. 1Dx2 + 1Dy10) to be determined. B. TRANSGENIC LINES

The development of methods for the transformation of wheat lagged behind those for most other major crops (including rice and maize), with the first fertile transformed plants being reported barely a decade ago (Vasil et al., 1992). Nevertheless, wheat transformation has now been established in a number of laboratories world-wide, resulting in renewed interest in quality targets. The most widely used wheat transformation system is based on direct gene transfer, the DNA being coated onto the surface of microscopic gold

254

P. R. SHEWRY ET AL.

particles and literally shot into the cells using high-pressure helium gas. It is necessary to use a recipient tissue that can be regenerated into a whole plant and it is usual to use immature embryos for bread wheat and either immature embryos or immature inflorescence explants for pasta wheat (see Barcelo et al., 2001). The exogenous DNA appears to integrate in a random fashion into the genome of cells of the recipient tissue, but only some cells are transformed. This could result after regeneration in a chimaeric plant containing a mixture of transformed and non-transformed cells. Consequently it is usual to also use a “selectible marker” gene which will usually co-integrate with the gene of interest and allow the selective regeneration of transformed cells. The most widely used selectible marker genes confer resistance to toxic herbicides or antibiotics, which can therefore be used to kill non-transformed cells. An alternative gene delivery system, based on using the bacterium Agrobacterium tumefaciens as a vector, has been applied to wheat (Cheng et al., 1997) but is not widely used. The delayed development of wheat transformation systems meant that genes for HMW subunits and information on their relationship to grain processing quality were already available when reliable systems become available, with the result that the HMW subunits were the earliest target selected for the improvement of wheat by transformation. The first success was reported by Blechl and Anderson (1996), who constructed a chimaeric gene encoding a hybrid subunit comprising residues 1 to 124 of the mature 1Dy10 subunit fused to residues 130 to TABLE V PUBLISHED REPORTS OF THE EXPRESSION OF HMW SUBUNIT GENES IN TRANSGENIC WHEAT

Species

Line

Subunit

Reference

Bread wheat Bread wheat Bread wheat

Bobwhite Bobwhite L88-6 L88-31 Bobwhite Canon Cadenza Pro INTA Federal L88-6 L88-31 L35 Ofanto Three lines

1Dy10/1Dx5 hybrid 1Ax1 1Ax1 1Dx5 1Dx5+1Dy10 1Ax1 1Ax1 1Ax1 1Dx5 1Dx5 mutants 1Dx5 mutants 1Ax1 1Dx5 1Ax1 1Dx5

Blechl and Anderson (1996) Altpeter et al. (1996) Barro et al. (1997)

Bread wheat Bread wheat Bread wheat Bread wheat Pasta wheat Tritordeum

冧

Anderson and Blechl (2000) Pastori et al. (2000) Alvarez et al. (2000) He et al. (2000) He et al. (1999) Rooke et al. (1999a)

HMW SUBUNITS OF WHEAT GLUTENIN

255

848 of subunit 1Dx5. This essentially combined the N-terminal domain of subunit 1Dy10 with the repetitive and C-terminal domains of 1Dx5. The novel subunit was readily resolved from the endogenous subunits present in the recipient cultivar (Bobwhite) by SDS-PAGE and analysis of seeds showed accumulation at levels comparable to those of the native proteins. However, subsequent studies (Shimoni et al., 1997) showed that the novel subunit formed circular monomeric structures stabilized by head-to-tail disulphide bonds rather than becoming incorporated into glutenin polymers. Since then HMW subunit genes have been successfully expressed in a number of lines of bread and pasta wheat, including commercial cultivars as well as model lines (Table V). They include two mutant forms of subunit 1Dx5 in which the length of the repetitive domain has been increased or decreased to determine the effects on the gluten properties. Also, HMW subunit genes have been used to transform cultivars of durum wheat and lines of tritordeum, the latter being a novel cereal produced by

FIG. 16. SDS-PAGE of ten single seeds of the transgenic line B73-6-1 showing expression of the 1Dx5 transgene in the L88-6 background. The minor band indicated by an arrow is observed to segregate between the progeny. Taken from Rooke et al. (1999b), with permission.

256

P. R. SHEWRY ET AL.

combining the genomes of pasta wheat and the wild barley species Hordeum chilense (Martin et al., 1999). In all cases expression levels up to or exceeding those of the endogenous genes have been reported, with expression being stable over a number of generations. However, both Blechl et al. (1998) and Alvarez et al. (2000) reported that some lines exhibited silencing or reduced expression of endogenous subunits, presumably via a co-suppression mechanism (see Barcelo et al., 2001). Of particular interest are some lines in which the levels of expression of the transgenes greatly exceed those of the endogenous genes. An example of this is shown in Figure 16 (Rooke et al., 1999b). The line B736-1 was produced by transforming the line L88-6 (expressing subunits 1Ax1, 1Dx5, 1Bx17, 1By18, 1Dy10) with about 15 copies of the 1Dx5 transgene, resulting in a four-fold increase in the proportion of subunit 1Dx5 (from about 2.7 to 10.7% of the total seed proteins) and a 1.6-fold increase in total HMW subunits (from about 12.7 to 20.5% of the total). B73-6-1 and several other transgenic lines have now been grown in replicate field trials at two UK sites (Long Ashton Research Station near Bristol and Rothamsted near London) over four seasons (1998–2001) (Fido et al., 2000; Popineau et al., 2001). The effects of the transgenes on the functional properties are discussed in a later section. VI. EXPERIMENTAL EVIDENCE FOR DIFFERENTIAL EFFECTS OF INDIVIDUAL HMW SUBUNITS ON MIXING AND RHEOLOGICAL PROPERTIES

In vivo evidence for differential effects of individual HMW subunits comes from comparative analyses of the functional properties of lines with different compositions of HMW subunits. These may be collections of cultivars, which led to the concept of “quality scores” discussed above, or more defined near isogenic and transgenic lines. A. VARIATION BETWEEN CULTIVARS

The close association of subunits 1Dx5 + 1Dy10 with good breadmaking quality when compared with subunits 1Dx2 + 1Dy12 has been largely confirmed by analysis of a range of cultivars from around the world (Branlard and Dardevet, 1985; Cressey et al., 1987; Campbell et al., 1987; Payne et al., 1988a; Ng and Bushuk, 1988; Lukow et al., 1989; Mosleth and Uhlen, 1991; Dong et al., 1991; Gupta et al., 1991a; Dong et al., 1992; Manley et al., 1992). Kolster et al. (1991) have reported that the alleles present at the Glu-D1 locus also modify the effects of alleles at the Glu-

HMW SUBUNITS OF WHEAT GLUTENIN

257

A1 and Glu-B1 loci on breadmaking quality. Thus, in the absence of subunits 1Dx5 + 1Dy10 correlations may be observed between Glu-A1 and Glu-B1 alleles and dough or gluten properties, as demonstrated by analysis of a collection of cultivars from southern Japan which contained subunits 1Dx2 + 1Dy12 or subunits 1Dx2.2 + 1Dy12 but not subunits 1Dx5 + 1Dy10 (Nagamine et al., 2000). Furthermore, genetic variation in the LMW subunit composition must also be taken into account when predicting technological properties based on glutenin subunit composition. A study including 101 genotypes (48 from Australia and 53 from around the world) showed that allelic variation in both LMW and HMW glutenin subunit composition explained the rheological properties of dough as determined by extensimetry (Gupta et al., 1991a). In the world set of genotypes, most of the dough resistance was explained by the HMW subunits, with a marked difference between subunits 1Dx5 + 1Dy10 and 1Dx2 + 1Dy12. However, this was not observed in Australian lines, where the LMW subunit composition was more strongly correlated with resistance than the HMW subunit composition. This association was attributed to the presence of particular combinations of HMW and LMW subunit (Glu-1 and Glu-3) alleles in these genotypes, with high-quality Glu-1 alleles often being associated with low quality Glu-3 alleles. Analysis of the dough properties of biotypes with identical gliadin compositions but different HMW subunit alleles indicated that those with the Glu-D1 subunits 1Dx5 + 1Dy10 had higher dough resistances than those with the subunits 1Dx3 + 1Dy12 and that these lines also contained a greater proportion of unextractable glutenin polymers (Gupta and MacRitchie, 1994). The effects of alleles at the Glu-A1 and Glu-B1 loci was much smaller (Lawrence et al., 1987). The dough properties and HMW glutenin subunit composition of the F3 population of the cross Nuri 70 (1Ax1, 1Bx17 + 1By18, 1Dx5 + 1Dy10) × UL 72 (1Ax null, 1Bx7 + 1By8, 1Dx2 + 1Dy12) were determined by Lagudah et al. (1988). No significant correlations were observed between dough properties and gliadin composition, but the resistance of the dough was strongly dependent on the HMW subunits alleles, especially those encoded at the Glu-D1 locus. In particular, the progeny containing subunits 1Dx5 + 1Dy10 showed higher dough resistance and lower resistance breakdown than those with subunits 1Dx2 + 1Dy12. On the other hand, no differences were observed between the Glu-B1-encoded subunit pairs 1Bx7 + 1By8 and 1Bx17 + 1By18. However, no relationship was found between HMW subunit composition and dough extensibility. Similar studies with other parental lines confirmed these results. Allelic variation at the Glu-D1 locus (i.e. subunits 1Dx5 + 1Dy10 vs 1Dx2 + 1Dy12)

258

P. R. SHEWRY ET AL.

explained most of the variation in the SDS sedimentation test values of random lines from the Cheyenne (1Ax2*, 1Bx7 + 1By9, 1Dx5 + 1Dy10) × MG 27116 (1Ax null, 1Bx7 + 1By8, 1Dx2 + 1Dy12) cross, whereas the Glu-B1 alleles had a much smaller effect (Benedettelli et al., 1992). Analysis of recombinant inbred lines also showed that subunits 1Dx5 + 1Dy10 were associated with superior qualitative traits compared with subunits 1Dx2 + 1Dy12 (Rousset et al., 1992). The loci encoding the HMW and LMW glutenin subunits have therefore been ranked as follows according to their contribution to the dough quality (essentially resistance) (Gupta et al., 1994): Glu-D1 > GluB1 > Glu-B3 > Glu-A3 > Glu-D3 = Glu-A1. However, additive and epistatic effects between the loci were also noted, and the differences in quality could only be assessed by taking into account the alleles present at the Glu-1 loci (encoding HMW subunits) and the Glu-3 loci (encoding LMW subunits) (Rousset et al., 1992; Gupta et al., 1994). It was also noted that factors other than the quantity of the HMW subunits were responsible for the effects of alleles at the Glu-D1 and Glu-B1 loci on the development time and the resistance of the doughs, in particular the size distribution of the polymeric proteins (Singh et al., 1990a, b; Gupta and MacRitchie, 1994), which was largely determined by allelic variation in the HMW (and LMW) glutenin subunits. Thus, lines with subunits 1Dx5 + 1Dy10 contained a higher proportion of unextractable polymers and showed a longer dough development time than other lines. Similarly, the presence of high molecular mass glutenin polymers and of unextractable polymeric glutenin proteins had previously been shown to be strongly correlated with dough strength (Dachkevitch and Autran, 1989; Singh et al., 1990b; Gupta et al., 1992, 1993). This indicates that the individual HMW subunits do not have the same propensity to polymerize, which could be the basis for the differences in dough properties associated with allelic variation in their composition. This is substantiated by preliminary studies of the in vitro polymerization of isolated subunits (Candler et al., 1996). Subunits encoded by the Glu-1 locus oxidized more slowly than subunits encoded by the Glu-D1 locus, with the x-type subunits oxidizing faster than the y-type for each locus. A synergistic effect was observed when xand y-type subunits were mixed, accelerating the oxidation and leading to a higher proportion of large polymers. Glutenin macropolymer (GMP) has been defined as the wheat protein fraction unextractable in 1.5% (w/v) SDS (Graveland et al., 1980). It contains principally glutenin polymers and its composition is strongly affected by dough mixing and resting (Hamer and Lichtendonk, 1987). The highest contents of GMP were found in flours, but mixing resulted in a significant decrease in the GMP content of dough and dough resting in

HMW SUBUNITS OF WHEAT GLUTENIN

259

an increase. These changes in GMP content (i.e. in the extractability of glutenin polymers) were interpreted as arising from de-polymerization and re-polymerization of glutenin (Hamer and Lichtendonk, 1987). The contents of GMP in flours and doughs of 14 cultivars were also compared and related to the mechanical properties and breadmaking performance of doughs (Weegels et al., 1997a). The de-polymerization/re-polymerization during dough mixing and resting were observed in all cultivars. The GMP contents also influenced dough rheology, with the contents after a 45 min rest explaining about 90% of the dough resistance to extension. Globally, GMP contents were better related to quality parameters than the classical Osborne protein fractions (i.e. albumins, globulins, gliadins and glutenins). GMP is composed of about 70% LMW and 30% HMW glutenin subunits (Weegels et al., 1995). However, each type of subunit was not affected to the same extent by de-polymerization and re-polymerization (Weegels et al., 1997b). After mixing, GMP contained a lower proportion of HMW subunits than flour GMP and a selective de-polymerization of x-type HMW subunits was also observed during mixing (Skerritt et al., 1999). However, subunit 1Dx5 was found to be particularly resistant to de-polymerization (Aussenac et al., 2001). HMW subunits were also more prone to become extractable through de-polymerization, but resting of the dough restored their initial concentrations in GMP. In this respect, the HMW subunits can be considered to be more reactive than the LMW glutenin subunits; this could be due either to their structures and conformations or to the conformations of the glutenin polymers. Furthermore, y-type subunits were re-incorporated in GMP more rapidly than x-type subunits during dough resting. As a consequence, GMP became more enriched in y-type subunits after glutenin re-polymerization (Weegels et al., 1997b). These data on the dynamics of glutenin polymers during dough processing showed that they undergo complex rearrangements in their composition and properties in dough. B. NEAR ISOGENIC LINES

Direct experimental evidence for a functional relationship between dough or gluten properties and HMW subunits comes from analyses of the series of NIL. The first important result from analysis of the Sicco isogenics was to show that the absolute amount of HMW subunit protein had a significant effect on the dough properties, with the absence of one or three subunits having significant effects on the SDS sedimentation values and breadmaking performance. The absence of subunit 1Ax had a small but significant negative effect on the quality parameters, whereas the absence

260

P. R. SHEWRY ET AL.

of subunits 1Dx5 + 1Dy10 had a major negative effect, as observed previously with two sister lines of wheat exhibiting contrasting breadmaking properties (Payne et al., 1988b). The absence of the 1Ax and 1Dx + 1Dy subunits had an even more pronounced negative effect. Furthermore, substitution of subunits 1Dx5 + 1Dy10 by subunits 1Dx2 + 1Dy12 gave less elastic dough (Payne et al., 1987a). This demonstrated that a qualitative factor, probably related to the specific structures of the individual subunits, was also involved in the determination of the dough elasticity. Further studies of gluten extracted from these Sicco NIL showed that absence of HMW subunits encoded by Glu-A1 and Glu-D1 decreased the total content of HMW subunits by about 50%. This increased the extractability of the glutenin, with the amounts of large glutenin polymers, as determined by sequential extraction and SE-HPLC, being reduced. The rheological properties of the hydrated gluten were also affected, with considerable change in both the storage and loss moduli (Popineau et al., 1994b) (Figure 17). The height of the elastic plateau dropped to less than 1/30 the value for the standard line and the G″/G′

10000

1000

GN0 N/m2

100

10

1

Sicco 5+10

Sicco 2+12

Sicco double null

Gabo 5+10

Gabo 17+18

Gabo triple null

FIG. 17. The height of the elastic plateaux (GN0) of glutens extracted from near-isogenic lines of Sicco and Gabo wheats differing in their composition of HMW glutenin subunits encoded at the Glu-A1, Glu-B1 and Glu-D1 loci. Dynamic rheological assays in shear were performed at 20°C on glutens fully hydrated with water using a Carri-Med CSL 100 constant stress rheometer (cone-plate geometry: cone angle, 4°; diameter, 2 cm, amplitude of strain at all frequencies, 3%). Sicco 5+10 (1Ax1, 1Bx7 + 1By9, 1Dx5 + 1Dy10); Sicco 2 + 12 (1Ax1, 1Bx7 + 1By9, 1Dx2 + 1Dy12); Sicco double null: (1A null, 1Bx7+1By9, null); Gabo 5+10 (1A null, 1B null, 1Dx5 + 1Dy10); Gabo 17 + 18 (1A null, 1Bx17 + 1By18, 1D null); Gabo triple null (1A null, 1B null, 1D null).

HMW SUBUNITS OF WHEAT GLUTENIN

261

increased four-fold, representing a dramatic collapse of gluten elasticity. The substitution of the subunit pair 1Dx5 + 1Dy10, normally present in the Sicco genotype, by the 1Dx2 + 1Dy12 pair did not modify the total amounts of HMW subunits or the concentrations of individual components. Nevertheless, the gluten in the 1Dx2 + 1Dy12 line contained a lower proportion of large glutenin polymers and the height of the elastic plateau was decreased five-fold. The relationship between the contents of the largest glutenin polymers in gluten fractions and the height of the viscoelastic plateau, previously shown in fractionation experiments (Cornec et al., 1994), was also valid with the Sicco NIL, showing that the HMW subunit composition determines gluten viscoelasticity by modifying the polymer size distribution and the aggregative properties of glutenin (Popineau et al., 1994b). Analysis of the Gabo NIL (Lawrence et al., 1988) confirmed these data. The mixing time of the dough (determined by Mixograph analysis) was decreased when HMW subunits were absent, the shortest time being recorded for the triple null line lacking all of the HMW subunits. In this study, a quantitative effect of the HMW subunits was observed and no differences were found between the contributions of subunits 1Ax1, 1Dx5, 1Bx17, 1By18 and 1Dy10 to dough functionality. Complementary experiments on the same lines indicated, however, that the presence of subunits 1Dx5 + 1Dy10 resulted in a longer dough mixing time and higher maximum resistance than subunits 1Bx17 + 1By18 (Gupta et al., 1995). The viscoelastic properties of glutens extracted from four of these NIL (the control, 1A/1B null, 1A/1D null and triple null lines) were compared by dynamic assay in shear (Lefebvre et al., 2000). Similar behaviour to that of the Sicco NIL was observed (Figure 17). When HMW glutenin subunits were absent, the rheological behaviour of the gluten was drastically modified: the height of the viscoelatic plateau decreased to a value equal to only 1/250th of the plateau in the standard line, and the position of the plateau was shifted to lower frequencies. Glutenin polymers accounted for 35% of total proteins in this triple null line but they were composed only of LMW subunits and contained only 5% of unextractable glutenin. When only HMW subunits encoded by the Glu-B1 locus (1Bx17 + 1By18) were present, the height of the plateau was lowered to about one half the value in the standard line. Furthermore, gluten from the 1Dx5 + 1Dy10 NIL exhibited the same plateau value as the standard line (Lefebvre et al., 2000). The size distribution of the glutenin also depended on the number and the type of HMW subunits that were absent. Thus, the absence of all of the subunits resulted in a very large reduction in the proportion of unextractable glutenin polymers (i.e. those requiring sonication to be extracted) in the flour (Gupta et al., 1995),

262

P. R. SHEWRY ET AL.

and their quasi-absence from the gluten, whereas the presence of the subunits 1Dx5 + 1Dy10 resulted in a higher amount of unextractable polymers than in the line with subunits 1Bx17 + 1By18. This was in agreement with the rheological properties of the glutens (Hargreaves et al., 1996, Lefebvre et al., 2000). From these assays, it was concluded that HMW glutenin subunits are practically indispensable for the formation of large aggregative polymers and that these cannot be formed by LMW subunits alone. It can therefore be concluded that HMW subunits constitute the basis for gluten viscoelasticity, because polymers below a critical size limit cannot efficiently entangle and so cannot contribute to the gluten strength properties (Gupta and MacRitchie, 1994; Bangur et al., 1997). Furthermore, a minimum amount of large glutenin polymers is necessary in order to confer viscoelastic behaviour to gluten fractions (Cornec et al., 1994). In addition, the different alleles have different “viscoelastic potential ”, which is higher for subunits 1Dx5 + 1Dy10 than for subunits 1Dx2 + 1Dx12 or 1Bx17 + 1By18 and is related to their ability to form aggregates or polymers. These differences must arise from structural features of the individual subunits such as the presence of an additional cysteine residue in the sequence of subunit 1Dx5 or to the regularity of the conformation of the repetitive domain (Anderson et al., 1989; Flavell et al., 1989; Goldsborough et al., 1989; Shewry et al., 1992). This is discussed in more detail in a later section. In the studies reported above the effect of allelic variation was analysed for pairs of subunits encoded by the Glu-B1 and Glu-D1 loci. The contributions of individual x-type and y-type subunits was studied by Rogers et al. (1991) using the Sicco null lines. The two Glu-D1 subunits showed a much greater positive effect on gluten strength and dough quality than the y-type Glu-1B subunit, confirming the ranking of the loci reported previously (Gupta et al., 1994) (Table I). Although the 1Dx subunit appeared to be superior to the 1Dy subunit, this observation was not conclusive because allelic variation was superimposed on the deletion of subunits (Rogers et al., 1991). Further experiments on the Galahad NIL (Galahad 6, Galahad 7 and Galahad 8, containing only the 1B-encoded HMW subunits 6, 7 and 8, respectively) demonstrated that gluten containing only a single y-type subunit was less extensible and more sensitive to heat treatment (Payne and Seekings, 1996). This was attributed to the additional cysteine residues present in y-type subunits, which could allow more extensive cross-linking of glutenin polymers. C. TRANSGENIC LINES

The creation of transgenic lines of wheat differing in their HMW subunit

HMW SUBUNITS OF WHEAT GLUTENIN

263

composition makes it possible to determine the effects of individual subunits on the rheological and technological properties of wheat gluten and thus to interpret functionality in terms of protein structure. Transgenes encoding subunits 1Ax1 (containing two cysteines available for intermolecular disulphide bonds) and subunit 1Dx5 (with three cysteines available for intermolecular disulphide bonds) were therefore expressed in bread wheat (Barro et al., 1997, Rooke et al., 1999b), pasta (durum) wheat (He et al., 1999) and tritordeum (Rooke et al., 1999a). The introduction of one (1Ax1) or two (1Ax1 + 1Dx5) transgenic HMW glutenin subunits in a Gabo NIL line (L88-31) containing only the Glu-B1-encoded subunits 17 and 18 resulted in progressive increases in the mixing time of dough (Barro et al., 1997), demonstrating that the subunits encoded by the transgenes were incorporated into the gluten structure in the same way as normal glutenin subunits and contributed to dough strength. In contrast, expression of subunit 1Ax1 in the cultivar Bobwhite showed only small effects on mixing time and loaf volume (Vasil et al., 2001), although this could be related to the fact that this cultivar contains the 1BL/1RS chromosome translocation, which results in “sticky dough” and poor mixing and baking performance. Similarly, although Anderson and Blechl (2000) reported that the expression of subunits 1Dx5 and 1Dy10 (presumably in the cultivar Bobwhite) resulted in greatly increased mixing time (from about 4 to 17 minutes), the peak resistance was greatly reduced. The 1Ax1 and 1Dx5 transgenes were also expressed in durum wheat lines that lack the D genome associated with high gluten viscoelasticity and were also silent for the Glu-1A locus. The absence of subunits 1Ax1 and 1Dx5 from the donor durum wheat lines therefore made it easy to evaluate the technological effects of transgene expression. Both transgenic subunits increased the strength and stability of the dough, with the presence of subunit 1Dx5 resulting in an overstrong dough that could not be properly mixed under normal conditions. However, blending of the transgenic 1Dx5 flour with a weak flour resulted in fortification when added at medium doses (He et al., 1999). The same subunits were also expressed in tritordeum, a fertile amphiploid between wild barley and durum wheat, which is also generally unsuitable for breadmaking. These lines also lacked subunits 1Ax1 and 1Dx5. The expression of subunit 1Ax1 resulted only in a small increase in the dough strength, whereas the expression of the subunit 1Dx5 improved both the mixing time and the resistance breakdown (Rooke et al., 1999a), but this difference could have resulted partially from the different expression levels of the two subunits. More detailed studies have been carried out on the effects of the 1Ax1 and 1Dx5 transgenes on Gabo NIL expressing HMW subunits 1Ax1,

264

P. R. SHEWRY ET AL.

1Bx17 + 1By18, 1Dx5 + 1Dy10 (the high-quality line, L88-6) and only subunits 1Bx17 + 1By18 (the low-quality line, L88-31). Over-expression of subunit 1Dx5 in the transgenic line B73-6-1 doubled the proportion of the HMW subunits in the total proteins, reaching 20%, and increased the concentration of subunit 1Dx5 by four-fold (Rooke et al., 1999b), but the gliadin/glutenin ratio was only slightly altered. Dough from this transgenic line did not develop properly under the condition of hydration and mixing speed (88 rpm) used in the 2g Mixograph test, whereas the control line L88-6 behaved normally. A high-speed mixing was necessary to form a continuous and cohesive dough. Blending experiments with a weak flour also indicated that the B73-6-1 line was overstrong (Rooke et al., 1999b). Further experiments were carried out on the L88-6 and L88-31 lines and the transgenic lines B73-6-1 (1Dx5 in L88-6), B102-1-2 (1Ax1 in L8831) and B72-8-11b (1Dx5 in L88-31) grown in the field in 1998, including protein fractionation and determination of the size distribution and rheological properties of extracted gluten (Popineau et al., 2001). Subunits 1Ax1 and 1Dx5 accounted for about 50 and 70% of the HMW subunits in the transformed lines, respectively, compared with 0% (1Ax1 in L88-31) and 26% (1Dx5 in L88-6) in the control lines. Overexpression of subunits

100000

10000

GN0 N/m2

1000

100

10

1

L88-31

B102-1-2

B72-8-11b

L88-6

B73-6-1

FIG. 18. The height of the viscoelastic plateaux (GN0) of glutens extracted from control and transgenic lines of wheat. Dynamic rheological assays in shear were performed at 20°C on glutens fully hydrated with water using a Carri-Med CSL 100 constant stress rheometer (cone-plate geometry: cone angle, 4°; diameter, 2 cm; amplitude of strain at all frequencies, 3%). L88-31: control line (1A null, 1Bx17 + 1By18, 1D null); B102-1-2: transformed line expressing subunit 1Ax1 transgene in the L88-31 background. B72-8-11b: transformed line expressing subunit 1Dx5 transgene in the L88-31 background. L88-6: control line (1Ax1, 1Bx17 + 1By18, 1Dx5 + 1Dy10); B73-6-1: transformed line expressing the subunit 1Dx5 transgene in the L88-6 background.

HMW SUBUNITS OF WHEAT GLUTENIN

265

1Ax1 and 1Dx5 in the transgenic lines B102-1-2 and B72-8-11b doubled the proportion of the HMW subunits in the total proteins, when compared which their respective control lines. Clear differences were observed, however, between the effects of subunits 1Ax1 and 1Dx5 on gluten physicochemical properties, emphasizing that differences in subunit structure can influence their contribution to gluten structure and rheology. Sequential extraction and SE-HPLC showed that the expression of the subunit 1Ax1 transgene increased glutenin aggregation, but did not appear to result in increased cross-linking by disulphide bonds. Thus, only the average size of glutenin polymers may have been increased. Gluten viscoelasticity was only moderately altered by the expression of the subunit 1Ax1 transgene (with slightly higher storage and loss moduli) (Figure 18), which mainly increased the dough resistance to elongation during mixing (Figure 19). In contrast, overexpression of subunit 1Dx5 generated very large and insoluble aggregates, probably through covalent cross-linking of polymers. As a result, the glutenin was only completely extracted after reduction of disulphide bonds. The connectivity and viscoelastic moduli of the gluten network were also greatly increased (Figure 18). This effect can be attributed primarily to the presence of an additional cysteine residue available for intermolecular cross-linking in subunit 1Dx5 as compared with subunit 1Ax1 (see Shewry et al., 1992). The very high gluten strength also resulted in abnormal dough mixing behaviour, irrespective of whether the genetic background was L88-6 or L88-31 (Figure 19). It can be postulated that an excess of subunit 1Dx5 modified the glutenin (gluten) structure and hindered the formation of a homogeneous protein network. Furthermore, subunit 1Dx5 is always expressed as a pair with subunit 1Dy10 and there is evidence that dimers between these two subunits, and between other x-type and y-type subunits, are present as “building blocks” in the glutenin polymers (Lawrence and Payne, 1983; Gao et al., 1992; Tao et al., 1992; Werner et al., 1992; Shani et al., 1994). Over-expression of subunit 1Dx5 in the absence of additional subunit 1Dy10 (or another y-type subunit) could therefore have resulted in extensive restructuring of the glutenin polymers, with important consequences for gluten strength and for the mixing and baking properties of dough. VII. THE MOLECULAR BASIS FOR CORRELATIONS BETWEEN HMW SUBUNITS AND QUALITY

It is clear from the preceding sections that we now know a great deal about the amino acid sequences of individual HMW subunits, which has led to

266

P. R. SHEWRY ET AL. 80

A

B

70 60

1

50

Torque (%)

5 17+18 10

40 30 20 10

a

b

c

d

e

0 0

100

200

300

400

500

600

Time (s) 80

80

C

70

60

50

Torque (%)

Torque (%)

D

70

60

40 30

50 40 30

20

20

10

10

0

0 0

100

200

300

400

500

600

0

100

200

Time (s)

400

500

600

80

80

E

F

70

60

60

Torque (%)

Torque (%)

300

Time (s)

40

20

50 40 30 20 10

0 0

100

200

300

Time (s)

400

500

600

0 0

100

200

300

400

500

600

Time (s)

FIG. 19. Analysis of the mixing properties of transgenic wheats expressing additional HMW subunits using the 2g Mixograph. (A) SDS-PAGE of the HMW subunits from a, L88-31: control line (1A null, 1Bx17 + 1By18, 1D null); b, B72-8-11b: transformed line expressing 1Dx5 subunit transgene in the L88-31 background; c, B102-1-2: transformed line expressing 1Ax1 subunit transgene in the L88-31 background; d, L88-6: control line (1Ax1, 1Bx17 + 1By18, 1Dx5 + 1Dy10); e, B73-6-1: transformed line expressing 1Dx5 subunit transgene in the L88-6 background. (B–F) are mixographs of (B) L88-31; (C) B102-1-2; (D) B72-8-11B; (E) L88-6; (F) B73-6-1. The resistance is given as torque % and the mixing time in seconds. Taken from Popineau et al. (2001), with permission.

HMW SUBUNITS OF WHEAT GLUTENIN

267

the development of generalized models for their structures, interactions and role in gluten structure and properties. However, it has proved more difficult to identify the precise molecular basis for the differential effects of individual subunits on processing properties that have allowed “quality scores” to be assigned to them (Table I). We will now consider our current knowledge of the molecular basis for these effects. A. PROTEIN AMOUNT

There is good evidence from several studies that the presence of a 1Ax subunit (1Ax1 or 1Ax2*) is associated with an increase in the total amount of HMW subunit protein present in the grain, by about 1.5 to 2.0% of the total grain proteins when compared with the null allele (Seilmeier et al., 1991; Halford et al., 1992). It is probable that this quantitative difference accounts for the quality score of 3 assigned to subunits 1Ax1 and 1Ax2*, as broad correlations between the total amount of HMW subunit protein and good processing properties have been reported by other workers (Payne et al., 1988a; Lawrence et al., 1988; Gupta et al., 1991b; Békés et al., 1994b; Popineau et al., 2001). B. PROTEIN SEQUENCES

The availability of complete amino acid sequences for allelic pairs of 1Dx and 1Dy subunits associated with good (1Dx5 + 1Dy10) and poor (1Dx2 + 1Dy12) quality allows comparisons to be made to identify features which may relate to their different properties (Figure 20). Comparisons between the sequences of subunits 1Dy10 and 1Dy12 show that they have identical N- and C-terminal domains, but their repetitive domains differ by 12 single amino acid substitution and by the deletion of two hexapeptides and two adjacent residues in subunit 1Dy10 and of two adjacent residues in subunit 1Dy12. Although these differences are minor, they may nevertheless affect the structure and stability of the protein domain. Thus, Flavell et al. (1989) noted that the differences in sequence resulted in a higher proportion of consensus repeat motifs in subunit 1Dy10, which should result in a more regular pattern of β-turns, while Hickman (1995) predicted the presence of 130 β-turns present mainly in the repetitive domain of subunit 1Dy10 but only 125 β-turns in subunit 1Dy12 (Figure 20). This difference in regularity could affect the stability and intrinsic elasticity of the subunits and their ability to form “trains” in gluten. Similar minor differences between the amino acid sequences of the repetitive domains of subunits 1Dx2 and 1Dx5 are also observed with 13 single amino acid substitutions and the deletion of two hexapeptides and

268

P. R. SHEWRY ET AL.

HMW SUBUNITS OF WHEAT GLUTENIN

269

FIG. 20. Alignment of the amino acid sequences of subunits 1Dx2, 1Dx5, 1Dy10 and 1Dy12. The lines indicate β-turns predicted by the method of Chou and Fasman (1978), using a significance level of 1 × 10–4.

insertion of three nonapeptides and two hexapeptides in subunit 1Dx5. In addition, they differ in one single amino acid deletion in the N-terminal domain of subunit 1Dx2. These differences are predicted to have little effect on the β-turn structure, with 211 and 212 turns predicted to be present mainly in the repetitive domains of subunits 1Dx2 and 1Dx5, respectively (Figure 20).

270 P. R. SHEWRY ET AL. FIG. 21. Transverse urea gradient gel electrophoresis of a range of x-type (1Ax1, 1Ax2*, 1Dx2, 1Bx6, 1Bx14) and y-type (1By8, 1By15, 1Dy10, 1Dy12) HMW subunits. Taken from Lafiandra et al. (1999), with permission.

HMW SUBUNITS OF WHEAT GLUTENIN

271

However, one of the single amino acid substitutions could have an impact on the interactions of the subunits in gluten. This is the substitution of a cysteine residue for serine at position 97 (at the N-terminal end of the repetitive domain) of subunit 1Dx5. This could result in the formation of more highly cross-linked, and hence more elastic, polymer. Evidence for this comes from the expression of the 1Dx5 transgene in developing grain, as discussed above. C. STABILITY OF ALLELIC SUBUNITS

It is possible that the differences in the degree of conservation of the repeat motifs present in allelic subunits lead to differences in protein conformation and/or conformational stability that affect their functional properties. A simple way to determine conformational stability is by gel electrophoresis in a transverse gradient of 0–8 M urea, as described by Goldenberg and Creighton (1984). Data obtained with this technique can provide information on the thermodynamics and the kinetics of the unfolding process and demonstrate changes in stability following introduction of a given mutation in the protein (Goldenberg, 1992). In general, proteins that unfold in a single, cooperative, two-state transition exhibit an abrupt decrease in mobility, with a single inflection point at the midpoint of the transition. Studies on the unfolding behaviour of different allelic variants of HMW subunits have been carried out (Lafiandra et al., 1999), demonstrating different behaviour for x-type and y-type subunits (Figure 21). The x-type subunits generally show a broad unfolding pattern with no clear transition as the urea concentration increases, suggesting the existence of several conformational intermediates. In contrast, most y-type subunits show an abrupt decrease in mobility on unfolding, with a single inflection point at the midpoint of the transition indicative of a protein unfolding in a single, cooperative, two-state transition. The free energy values associated with the unfolding process were calculated for the ytype subunits, which showed a two-state transition. Different values were calculated for the allelic subunits 1Dy10 and 1Dy12, with the former showing greater stability. This is consistent with the differences in conservation of the repetitive domains of these two subunits, as discussed above, and may provide a partial explanation for association of the subunit pair 1Dx5 + 1Dy10 with good quality when compared with subunits 1Dx2 + 1Dy12. D. SUBUNIT INTERACTIONS

These studies indicate, therefore, that both the structure and stability of subunit 1Dy10 and the cross-linking behaviour of subunit 1Dx5 may

272

P. R. SHEWRY ET AL. 100

% IMMOBILE PROTEIN

80 60 40 20 0 0

1

2

3

4

5

PROTEIN TO WATER RATIO

FIG. 22. Relationship between the proportion of mobile protein and the water content of HMW subunits. Triangles: data from Belton et al. (1994) calculated assuming that all populations of relaxation times less than or equal to 125 µs contributed to the immobile signal. Open squares: data from Gil et al. (2001) using the average figures for the subunits 1Dx5 + 1Dy10.

contribute to the greater dough strength associated with subunits 1Dx5 + 1Dy10 compared with 1Dx2 + 1Dy12. However, it is also possible that specific interactions occur between the individual subunits of the allelic pairs (i.e. 1Dx5 with 1Dy10, 1Dx2 with 1Dy12), resulting in synergistic effects on quality. One of the well-documented effects of adding water to high molecular weight subunits is the change in mobility observed by NMR. These observations have contributed to the development of the loop and train hypothesis discussed below (see p. 281). It seems logical, therefore, that any interactions between different subunits that affect the rheology should be reflected in changes in mobility. Gil et al. (2001) have compared the 13C CPMAS (cross-polarization magic angle spinning) spectra of subunits 1Dx5, 1Dx2, 1Dy10 and 1Dy12 separately and in combination. The CPMAS spectra are only sensitive to the immobile parts of the proteins. Therefore, the experiment directly measures only those parts of the protein that are not moving rapidly and therefore correspond to regions in which the hydrogen-bonded interprotein interactions are present (these are referred to as chains in the following section). Similar measurements of the amount of immobile material can also be made using proton NMR relaxation time measurements. (Belton et al., 1994). Figure 22 shows relaxation time data obtained for a mixture of high molecular weight subunits, together with average values for subunits 1Dx5 and 1Dy10 calculated using CPMAS data. While the scatter in both sets of data is

HMW SUBUNITS OF WHEAT GLUTENIN

273

TABLE VI VARIATION IN THE AMOUNT OF IMMOBILE PROTEIN IN HMW SUBUNIT MIXTURES

Percentage of immobile protein Water : protein ratio Dry 0.6 1.8

1Dx5 + 1Dy10 Predicted value

1Dx5 + 1Dy10 measured value

1Dx2 + 1Dy12 predicted value

1Dx2 + 1Dy12 measured value

100 36 32

100 50 40

100 38 52

100 95 29

Data recalculated from Gil et al. (2001).

high, the same trends are evident. The scatter arises, in part, because of the intrinsic difficulty of measurement, but also because the two methods of measurement are not exactly the same. When different subunits are compared using the CPMAS method differences in behaviour are observed. Table VI compares the amounts of immobile protein measured for mixtures of subunits 1Dx5 + 1Dy10 and 1Dx2 + Dy12 with the predicted values arrived at by simply averaging the values of the two components in the mixture. It is difficult to estimate the errors in the data, but the variation between the predicted and measured values for subunits 1Dx5 + 1Dy10 is probably not significant. For the mixture of subunits 1Dx2 + 1Dy12 the differences are much larger and suggest that the sensitivity to water content has been greatly changed by interactions of the two subunits. In particular, at intermediate water contents the interactions between the two subunits appear to result in a higher population of trains (see p. 281). If this were the case it would seem to run counter to the observation that enhanced dough strength arises from interactions of the subunits 1Dx5 + 1Dy10 rather than subunits 1Dx2 + 1Dy12. This is an area where further work is needed to compare the molecular level information obtained from spectroscopy with the macroscopic rheological information. VIII. THEORETICAL BASIS FOR PROPERTIES OF THE HMW SUBUNITS AND THEIR ROLE IN GLUTEN VISCOELASTICITY AND DOUGH MIXING

Much of the above discussion has focused on providing a biochemical and genetic explanation for the role of the HMW subunits in determining breadmaking quality. However, recent work allows us to develop more sophisticated biophysical models to explain the fundamental properties of

274

P. R. SHEWRY ET AL.

HMW subunits and the molecular basis for their role in viscoelasticity and mixing. A. WHY ARE HMW SUBUNITS INSOLUBLE?

It is often stated that gluten proteins are insoluble in water or that particular prolamins are insoluble or soluble in specific solvents. Unfortunately, the terminology is often used with respect to some pragmatic measure such as the ability to dissolve the protein to a sufficient extent to carry out some biochemical analysis. However, in thermodynamic terms the solubility of a component in a liquid at equilibrium can be defined by the equilibrium condition: µS = µ0L + RT ln aL

(1)

where µ is the chemical potential of the component in the solid (S) or the liquid (L). The subscript 0 refers to the standard state of the component, while a is the activity of the component. Thus, for a system at equilibrium with a pure solid µS = µ0S, and RT ln aL = µ0S – µ0L

(2)

with the term on the right-hand side being a constant. In practice it is very difficult to achieve this condition for proteins since the standard state is a crystalline solid. In general, therefore, the equation must be modified to take account of the fact that the solid is not in its standard state. Assigning an activity to the noncrystalline protein may achieve this, if pseudoequilibrium between solid and liquid is assumed. Thus: µ0S + RT ln aS = µ0L + RT ln aL

(3)

This equation reflects the importance of the activities in both the phases. The term aS is a measure of the deviation of the solid phase from its standard state, which for proteins will be a function of sample history. The solubility as measured, therefore, will depend on the state of the solid sample and should be measured in a steady state for defined and reproducible solids and solvents. Unfortunately, pragmatic biochemical observations are often treated as if they were thermodynamic measurements. Nevertheless, it is obvious that addition of excess water to gluten or native high molecular weight subunits does not result in the formation of a clear solution as is the case with soluble proteins such as lysozyme or bovine serum albumin, and it is therefore legitimate to infer that the

HMW SUBUNITS OF WHEAT GLUTENIN

275

factors favouring the formation of a solution are insufficient in the case of gluten. In order for a solution to form there must be net decrease in the free energy of the protein/solvent system on mixing. Conceptually one can imagine that two steps have to take place: the first is the formation of a liquid from the solid phase, and the second is the mixing of the liquid solute and the solvent. The total free energy change (∆Gsol) is the free energy of the formation of the liquid solute (∆Gfus) plus the free energy of formation of the mixture (∆Gmix) ∆Gsol = ∆Gfus + ∆Gmix

(4)

In order for solution to occur there must be net decrease of the free energy, i.e. ∆Gsol must be negative, ∆Gfus will be positive and thus for solution to occur ∆Gmix must be negative and greater in magnitude than ∆Gfus. The magnitude of ∆Gfus will depend on the strength of the intermolecular interactions in the solid; if these are strong then solution will not be favoured. The importance of considering both terms seems to have been neglected by some authors. For example, Singh and MacRitchie (2001) only consider the Flory free energy of mixing a liquid polymer with a liquid solvent (Flory, 1953) and ignore the role of ∆Gfus. If this logic were applied to the solubility of inorganic salts one would expect compounds such as barium sulphate or lithium fluoride to be readily soluble, as the free energies of hydration of the component ions are quite favourable. However, in both of these cases the very high energy of formation of the crystalline state results in lack of solubility. Free energy contains both enthalpic and entropic contributions, and some authors have suggested that gluten proteins (including HMW subunits) are insoluble because the entropy of solution is unfavourable. Assumptions about the importance of entropic terms can arise because of the neglect of the ∆Gfus term (Singh and MacRitchie, 2001), or because of the assumption that hydrophobic interactions in gluten make it insoluble. If a substance is hydrophobic, its interactions with water cause an increase in the hydrogen bonding of the water. As this is a more ordered state, the entropy of the system decreases on solution. Because this is not compensated by any enthalpic interaction, solution is not favoured. This view of gluten protein solubility is also based on the assertion that gluten proteins are hydrophobic. However, consideration of the amino acid composition does not provide support for this. In fact, gluten proteins are very rich in hydrophilic amino acids. The hydrophilic/hydrophobic balance can also be calculated using published data for individual amino acids and typical amino acid composition of whole gluten and HMW subunits (see

276

TABLE VII FREE ENERGIES OF HYDRATION FOR AMINO ACIDS IN GLIADINS, GLUTENINS, WHOLE GLUTEN AND AN AVERAGE OF THE AMINO ACID CONTENTS OF 314 PROTEINS

Amino acid

Free Mol% energy of amino hydration acid in (kcal/mol) gliadins

–3.11 –1.69 –2.36 –3.15 0.23 –0.23 –0.06 –0.27 0.04 –0.1 0.07 0.07 –2.82 –0.28 –3.77 –2.18 –6.85 –0.88

2.8 2.4 6.1 34.5 16.2 3.1 3.3 3.3 4.8 1.3 4.4 6.9 1.8 4.3 0.6 1.9 2 0.4

Mol% amino acids in 1Dx5

3.7 3.5 7 28.9 11.9 7.5 4.4 2.6 4.8 1.4 3.7 6.5 2.5 3.6 2 2 3 1.3

0.48 2.9 5.68 37.97 13.8 20.07 3.02 0.61 1.69 0.24 0.48 4.36 5.56 0.29 0.73 0.48 1.21 1.09

Mol% Average Free Free Free Free Free amino amino energy energy energy energy energy acids in content of of of of of of 1Dy10 314 hydration hydration hydration hydration hydration proteins for an for for for 1Dx5 for (mol%) average gliadins glutenins (kcal/mol) 1Dy10 protein (kcal/mol) (kcal/mol) (kcal/mol) (kcal/mol) 0.64 3.83 6.7 35.57 11 18.02 3.67 1.12 2.55 0.48 0.64 3.83 5.42 0.32 1.12 2.07 2.07 0.96

9.8 6.1 7 9.9 5.2 8.4 8.6 2.9 6.6 1.7 4.5 7.4 3.4 3.6 6.6 2 4.9 1.3 Totals

–30.478 –8.708 –10.309 –4.056 –16.52 –14.396 –31.185 –108.675 1.196 3.726 –1.932 –0.713 –0.516 –0.198 –0.783 –0.891 0.264 0.192 –0.17 –0.13 0.315 0.308 0.518 0.483 –9.588 –5.076 –1.008 –1.204 –24.882 –2.262 –4.36 –4.142 –33.565 –13.7 –1.144 –0.352 –164.147 –159.794

–11.507 –5.915 –16.52 –91.035 2.737 –1.725 –0.264 –0.702 0.192 –0.14 0.259 0.455 –7.05 –1.008 –7.54 –4.36 –20.55 –1.144 –165.817

–1.4928 –1.9904 –4.901 –6.4727 –13.4048 –15.812 –119.606 –112.046 3.174 2.53 –4.6161 –4.1446 –0.1812 –0.2202 –0.1647 –0.3024 0.0676 0.102 –0.024 –0.048 0.0336 0.0448 0.3052 0.2681 –15.6792 –15.2844 –0.0812 –0.0896 –2.7521 –4.2224 –1.0464 –4.5126 –8.2885 –14.1795 –0.9592 –0.8448 –169.616 –177.224

The data for this table are taken from Janssen (1992) (amino acid contents) and Oobtake and Tatsuo (1993) (hydration free energies) 1 cal = 4.18 J.

P. R. SHEWRY ET AL.

asp thr ser gln pro gly ala cys val met ile leu tyr phe lys his arg trp

Mol% amino acid in glutenins

HMW SUBUNITS OF WHEAT GLUTENIN

277

Table VII). The result is consistent with the calculations of Belton (1995) and shows that the free energies of hydration for HMW subunits are very close to the average value calculated from a sample of 314 proteins of –164 kcal mol–1 (–686 kJ mol–1). From this point of view, gluten and high molecular weight subunits are typical hydrophilic proteins. This is entirely consistent with the swelling behaviour of gluten in water (Grant et al., 1999). Furthermore, comparison of the behaviour of the truly hydrophobic protein elastin with HMW subunits (Belton et al., 1994) shows that the latter absorb water on heating whilst elastin ejects it, clearly demonstrating the hydrophilic nature of HMW subunits. These results do not imply that hydrophobic interactions do not exist, but they do imply that the reason for lack of solubility cannot solely be hydrophobicity. Singh and MacRitchie (2001) have argued that the lack of charged groups may also contribute to the lack of solubility of gluten proteins. This is undoubtedly the case, but the authors fail in their analysis to take account of the importance of polar groups in contributing to polymer solubility. Polar groups such as OH groups can account for the solubility of many polysaccharides; gluten is rich in glutamine and thus polar NH2 groups, which would be expected to interact favourably with water. It is clear, therefore, that the insolubility of HMW subunit proteins needs some explanation. We consider that the answer lies mainly in the ∆Gfus term, not in the ∆Gmix term. The polypeptide polyglutamine is not soluble in water, but exists as an insoluble material with a β-sheet conformation. Similarly, Perutz (1996) has identified the formation of “polar zippers” formed between stretches of glutamine residues in pathological forms of human proteins. Recent modelling work (Belton et al., 2000) has suggested that the formation of intermolecular β-sheet involving glutamine residues present in HMW subunits may also occur. The hydrogen bonding pattern will also have the effect of placing the hydrophobic side chains within the interior of the protein. The role and importance of β-sheet formation will be discussed in more detail later, but for the present we note that the reason for the insolubility of the HMW subunits and other gluten proteins in water is probably the formation of strong inter- and intra-protein interactions rather than hydrophobic interactions with water. The question still remains as to why it is possible to dissolve HMW subunits in alcohol–water mixtures, and in this case the answer may lie in the balance of hydrophobic and hydrophilic forces. The addition of alcohol will favour the interaction of the hydrophobic groups with the solvent; in addition alcohols are polar and so can interact with the amide groups of the glutamine residues. This interaction will clearly be less favourable than the interaction with water but, combined with favourable

278

P. R. SHEWRY ET AL.

interactions with hydrophobic side chains, may be enough to tip the balance in favour of solubility. It is important to recognize that disulphide bonds interlink much of the protein present in doughs. This creates very large polymeric units (Lindsay and Skerrit, 1999) that appear to contain both HMW and LMW subunits, with the LMW subunits acting to both extend and terminate the chains of polymers. However, the extractability of glutenins from dough has been shown to be a strong function of mixing (Weegels et al., 1996) and it is assumed that extractability is related to polymer size. Clearly, if the polymer network is very extended, then the behaviour of the dough may be better understood as akin to a swollen polymer gel. However, it is important to note that the presence of the polymer does not alone account for the insolubility of the proteins in water since reduction of the disulphide bonds does not result in water solubility. B. INTERACTIONS BETWEEN HMW SUBUNITS

As discussed above, consideration of the factors affecting subunit solubility suggest that hydrogen bonding may play a role. Early NMR relaxation time studies of C hordein, the barley homologue of the wheat ω-gliadins, indicated that significant changes in the dynamics of the protein occurred as water was added, even though the protein did not dissolve (Belton and Gil, 1993; Belton et al., 1994). Similar results have since been observed with HMW subunits (Belton et al., 1995) and ωgliadins (Belton et al., 1998). The C hordeins and ω-gliadins both resemble HMW subunits in containing repeat motifs that are rich in proline and glutamine residues. The similar behaviour of these proteins may therefore be attributed to competition with water for hydrogen bonding with glutamine. Increasing the water content results in a decrease in glutamine–glutamine hydrogen bonds and an increase in glutamine–water hydrogen bonds, leading to an increase in the mobility of the polymer chains. A number of authors have used FT-IR spectroscopy to determine the structure of the proteins in gluten and related systems (Pezolet et al., 1992; Popineau et al., 1994b; Belton et al., 1995; Wellner et al., 1996; Grant et al., 1999; Gilbert et al., 2000). The broad conclusion from these studies is that gluten and the individual proteins contain β-sheet, β-turn and some α-helix, the latter mainly being associated with nonrepetitive sequences in the C- and N-termini of the HMW subunits (see above). Detailed studies of the behaviour of HMW subunits and ω-gliadins show that the proteins also undergo changes in conformation with water content. When dry, the proteins are largely disordered, but as water is added there are initially increases in the amounts of β-sheet and then increases in β-turns. More

HMW SUBUNITS OF WHEAT GLUTENIN

279

recent work (Feeney et al., 2002) has shown that similar behaviour is observed with model peptides based on the repeat motifs of the HMW subunits. The behaviour of the HMW subunits on hydration can be explained in terms of the formation and breaking of hydrogen bonds: at low water levels there are many intra- and inter-chain interactions. However, because there is very little molecular motion, the proteins remain in the disordered state that they adopt during sample preparation (typically freeze-drying). As more water is added, some molecular rearrangements to the lowest energy state will occur, which for glutamine-rich materials will be the β-sheet conformation. As yet more water is added, the water and the side chain amide residues will compete for hydrogen bond formation, decreasing the population of inter-chain β-sheet conformers and increasing the population of extended hydrated β-turn conformers. The transition from β-sheet to β-turn may be correlated with the changes in mobility observed by NMR; the β-sheet regions are of low mobility and the hydrated extended β-turns of high mobility. C. HIGH MOLECULAR WEIGHT SUBUNITS AND DOUGH VISCOELASTICITY

This interpretation of the spectroscopic results has led to a model for the interactions of HMW subunits in doughs that offers an explanation for dough viscoelasticity. However, before considering this it is important to recognize that a theory of dough rheology is not a theory of breadmaking quality. Although appropriate dough mechanical properties are necessary for breadmaking, they are not sufficient. Bubble stability, oven spring, and the stabilization of structure by starch gelation depend on other factors. Theories of dough rheology can, therefore, only be expected to explain those observations that relate to the mechanical properties of dough. Before considering theoretical models in more detail, it is important to clarify the thermodynamic basis of elasticity. It is often assumed that the elastic component of gluten viscoelasticity is purely entropic. This can arise because of a misunderstanding of the basic thermodynamics involved. For example, Ewart (1989) assumed that the second law of thermodynamics ensures that all spontaneous changes are entropy-driven. Although it is true that the entropy must increase in a spontaneous change for whole closed systems, this may occur because of enthalpic effects causing a rise in temperature. It is certainly not true for any individual component of that system, and if it were crystallization could not occur. The second law does not, therefore, imply that elasticity must be the result of entropic forces; the total free energy change is critical. In addition to these a priori reasons

280

P. R. SHEWRY ET AL.

for not assuming that elasticity is driven by entropy in gluten, recent studies by Tatham et al. (2001), in which the temperature coefficient of the elasticity of high molecular weight subunits was measured, have shown empirically that elasticity is not entropic. Some of the key observations on the mechanical behaviour of dough are as follows: 1. Dough is viscoelastic. 2. The mechanical behaviour of dough is strongly dependent on water content. 3. The mechanical behaviour of dough is strongly dependent on the nature and concentration of HMW subunits present. 4. Mixing in D2O rather than H2O results in stronger dough. 5. Esterification of glutamine residues reduces the coherence and the resistance to extension of doughs. 6. The presence of oxidants and reducing agents affects dough rheology. 7. The resistance to extension of doughs reaches a maximum on mixing. A reasonable starting assumption is that dough consists of a protein matrix in which are embedded starch granules. This matrix contains a crosslinked polymer network with some noncovalently bonded readily extractable proteins. The readily extractable proteins are unlikely to contribute to the elasticity of the dough since elasticity requires an interconnected network, albeit not necessarily a covalently bonded one. Elasticity therefore resides in the network. The network comprises both HMW and LMW subunits, which may interact with each other through covalent disulphide bonds and non-covalent interactions. Although the exact organization and role of the low molecular weight subunits in the polymer is not clear (Lindsay and Skerrit, 1999; Shewry et al., 2001), it is possible that they can act both as chain terminators and extenders. Our current knowledge of the disulphide bond structure of glutenin indicates that most disulphide bonds between subunits are end-to-end, although some branches also occur (see above). Located between these cross-links are the repetitive domains that interact with the repetitive domains of other proteins by hydrogen bonds. Interactions between HMW subunits will therefore tend to increase the cohesiveness of the dough, while interactions between HMW and LMW may increase or decrease cohesiveness, depending on the precise interactions. If the interaction results in the masking of the “interactive surface” of a HMW subunit, then the cohesiveness may be diminished. If, however, the external surface of the LMW subunit can hydrogen bond, then cohesiveness may be unaffected or increased. Interactions between HMW and LMW subunits could, therefore, affect dough rheology by both hydrogen bonding and disulphide interactions. This is illustrated in Figure 23.

HMW SUBUNITS OF WHEAT GLUTENIN

281

C D

F A

E G

H

B FIG. 23. Schematic representation of the interactions in a glutenin macropolymer. (A) Cand N-termini of HMW subunits containing cysteine residues; (B) HMW C- and N-termini blocked by chain-terminating LMW subunit; (C) train region of HMW subunit; (D) loop region of HMW subunits; (E) disulphide cross-linked C- and N-termini; (F) LMW subunit acting as chain extender; (G) LMW subunits acting as a hydrogen bonding cross-linking unit; (H) LMW subunit preventing the formation of inter-chain hydrogen bonds.

Irrespective of the exact role of LMW subunits, the starting point for the theory of dough rheology is that elasticity originates mainly from interaction between the HMW subunits, although these may be mediated by the LMW subunits. The hydrogen bonded interactions between repeat regions (which we term chains) of these would be expected to be mediated by water. Thus, in the dry state strong interactions would be expected. Addition of water would result in competition between water and amino acid residues to form hydrogen bonds, breaking some of the inter-chain interactions. However, the density of glutamine residues and the proximity of the chains would result in many inter-chain interactions, and it is unlikely that all would be broken simultaneously. Two types of interaction will therefore be present in equilibrium in the hydrated system: hydrated regions, which, as above, are identified with mobile, extended β-turn structures, and regions of subunit interactions identified with rigid β-sheet structures. These two regions have been termed “loops and trains” (Belton, 1995, 1999a), in analogy with the interactions of polymers with surfaces (Dickinson, 1992). In the latter the polymer has many interactions with the surface, making it almost impossible to wash it from the surface once it has been absorbed. Even if the individual interactions

282

P. R. SHEWRY ET AL. LOOPS

TRAINS

FIG. 24. Schematic depiction of loops and trains formed by a polymer at an interface.

are quite weak, the statistical probability of all the interactions being broken simultaneously is very low. Polymers at interfaces thus exist in train regions where there are a number of interactions between the polymer and surface and solvated loop region. This is illustrated in Figure 24. The “loop and train” hypothesis has been supported by recent work of Alberti et al. (2001), who showed, using proton magic angle spinning NMR, that hydrated HMW subunits contained two types of glutamine residues. It is probable that one of these is associated with loops and the other with trains. As mechanical extension is applied to the gluten network, the first effect will be to extend the loops, as this will require relatively little energy. Further extension will strain the loops and pull the β-sheet train regions apart, while even further extension may result in breakage of the inter-chain disulphide bonds. When the extending force is relaxed the tendency will be to restore the balance between loops and trains, since this is the equilibrium situation. If the time scale for extension is faster than the intrinsic relaxation rate of the polymers, then the application of mechanical force will result in the build up of stored elastic energy of the system. This is illustrated diagrammatically in Figure 25, in which the polymeric network is simplified and shown as a group of cross-linked linear polymers and the LMW subunits omitted. The model leads to an explanation of some of the points made above. It explains why dough is viscoelastic (point 1) as it provides a mechanism for both extension and elastic recovery. The effects of water (point 2) may be seen in the changes that hydration causes in the ratio of loops to trains. Thus, increasing the water content will result in an increase in the loop to train ratio, which will make extension easier. However, the loop to train ratio cannot increase without limit as it is constrained by the disulphide cross-links that will keep the proteins in close proximity, by the solubility of the proteins which is in turn limited by the interchain interactions, and by breadmaking practice that typically maintains the water content of

HMW SUBUNITS OF WHEAT GLUTENIN

283

A

B

C

FIG. 25. Schematic depiction of changes in the polymer network in dough during extension. A two-dimensional network is shown for clarity; in reality the network is threedimensional. (A) Network before extension; (B) the extended network; (C) the network after relaxation. Note that after relaxation some polymers have become detached from the network.

dough at about 65%. Assuming that the water is evenly distributed between all components, this would result in about six water molecules per amino acid residue. A similar calculation has been reported by Ewart (1989). This is a very high protein concentration but it is still not high enough for glass formation at room temperature. The system is, therefore, in a rubbery state (Noel et al. 1995; Kalichevsky et al., 1992) and is liable to plasticization by water by the mechanism proposed. It should be noted that the term rubbery is used in this context to denote the opposite of glassy and does not imply that dough behaves according to rubber elasticity theory. In the hypothesis outlined above the HMW subunits play a crucial role and are assumed to be responsible for most of the gluten viscoelasticity. This may seem surprising, as they only account for about 10% of the total grain protein. However, it is not unusual for the rheology of a food system to be governed by the behaviour of a minor component, and high

284

P. R. SHEWRY ET AL.

molecular mass polymers can radically change the rheology of systems even when concentrations are low. A good example of this phenomenon is agar, which at 1% (w/v) will transform liquid water to a solid selfsupporting gel by formation of a continuous network. Thus, the crucial role of the HMW subunits, which account for only about 20% of the total glutenin subunits, is not surprising. Any change in the concentration of the HMW subunits or any changes that might make a difference to the ratio of loops to trains will therefore directly affect the mechanical properties of the system (point 3). The role of the polymer interactions can also explain the effects of mixing in the presence of D2O and of the esterification of glutamine residues. Deuterium would form stronger hydrogen bonds than hydrogen in the β-sheet regions, which would result in stronger interactions and stronger doughs. Conversely, the esterification of glutamine residues would reduce the likelihood of β-sheet formation and thus weaken doughs (points 4 and 5). Oxidizing and reducing agents (point 6) would be expected to have effects on disulphide bonds, with oxidizing agents being assumed to increase the number of disulphide bonds and reducing agents to decrease them (Weegels et al., 1996). Thus, the effects of oxidizing and reducing agents may be through their effects on the connectivity of the network. However, both oxidizing and reducing agents also increase the extractability during mixing. It is probable that reducing agents increase the extractability by reducing the number of disulphide bonds. In contrast, increasing cross-links by oxidation would increase dough stiffness and work input. This could result in more bond breakage, resulting in a fragmented structure and thus increased protein extractability (Weegels et al., 1996). It should be noted that implicit in the interpretation of extractability measurements is the assumption that the controlling factor for extraction is the nature of the polymer network. However, as noted above, equilibrium considerations do not apply in such systems and the controlling factor may be the kinetics of the extraction process. This may depend on such factors as fissures and cracks in the dough improving solvent ingress and solute egress and the effective dough surface area. So far these factors do not appear to have been investigated. D. EXPLANATION FOR THE ROLE OF THE HMW SUBUNITS IN DOUGH MIXING

In order to understand the maximum resistance observed in doughs during mixing (point 7), it is necessary to explore the process in a more quantitative manner using a simple kinetic model. Figure 25(A) shows a

HMW SUBUNITS OF WHEAT GLUTENIN

285

diagrammatic representation of the network of polymers in dough. In this model the HMW subunit network is simplified to a network of polymers that are cross-linked at the ends and are elastic. The elasticity, as explained above, results from the tendency of the HMW subunit network to return to its equilibrium ratio of loops to trains after extension. After extension the rate of restoration of equilibrium is relatively slow compared with the extension rate in the mixer. Thus, if there are N0 polymers in the network, on extension due to mixing, the network is stretched and there will be NS stretched polymers formed. The effect of mechanical extension of the stretched network is to break NB polymers away from the network with NU polymers remaining unstretched. If the system is left for a suitably long time after mixing it relaxes to its original unstretched state, but NB polymers remain broken from it. Following the work of Gras et al. (2000), the best characterized mixing process is that of the Mixograph. They showed that the action of the Mixograph was purely extensional and that extension resulted in the storage of elastic energy in the dough. They reported the behaviour of three different wheat flours in the system. The flours showed the usual behaviour in reaching a maximum in resistance to extension, the height of which deceased with increased water content. Water content also affected the behaviour of the dough after resting, with the resistance to extension decreasing with water content and the degree of extension to break increasing. Longer mixing decreased the resistance to extension and decreased the degree of extension to break of doughs of constant water content, which were rested after different mixing times. The effects of extension on the network shown in Figure 25 may for simplicity be modelled by first-order kinetics (these are, of course, not realistic for a mixing experiment, but the kinetics of the process are complex and currently not known). First-order models are, therefore, used to illustrate the principles involved and demonstrate that realistic models based on a network theory, combined with the loop and train model, can reproduce some of the features observed during mixing. In the following it is assumed that each revolution of the Mixograph results in the creation of stretched polymers and that the rate of relaxation of the polymers from their stretched state is slow compared with the rate of mixing. Thus, mixing stores mechanical energy in the dough. Further, it is assumed that the formation of the network and the uniform hydration of the dough are fast compared with the mixing process. The latter assumption may not be true for all mixing processes and some of the observed increase in the resistance of the dough may be due to network build up. However, this would not explain the observation of maxima in dough resistance, merely a monotonic increase and levelling,

286

P. R. SHEWRY ET AL.

and for simplicity it is ignored here. Such an effect could be modelled, however, by simply adding a term for the rate of increase in the number of polymers in the network. After R revolutions the system can be described by the following equations: dNS/dR = k1NU – k2 NS

(5)

dNU /dR = – k1 NU

(6)

dNB/dR = k2 NS

(7)

and NU + NS + NB = N0

(8)

where k1 is the rate constant for the rate of creation of stretched polymers and k2 is the rate constant for the rate of breakage of the polymers. In order to relate these equations to resistance to extension, additional assumptions about the elasticity of the polymer networks need to be made. Resistance, ρ, is related to the number of polymers by the simple linear equation

ρ = EUNU + ESNS + EBNB

(9)

where E is the resistance to extension per polymer unit. The most important factor controlling the extensibility of the network will be the number of polymers it contains that are already extended, because as more polymers are extended the possibilities for further extension will decrease. Unextended polymers that are attached to the network will resist extension since they are coupled to it by trains and will have the natural resistance to extension of any polymer. Broken polymers will only contribute viscous drag to the system, thus ES > EU > EB. The variation of the numbers of the various types of polymers with the number of revolutions is shown in Figure 26. Clearly the number of stretched polymers and hence the resistance reaches a maximum. Figure 27 shows the effects of varying the rate constant for the formation of broken polymers. A large value of k2 results in a sharp maximum whereas a low value results in a flattening of the curve. It may be expected that stiffer doughs will result in the input of more mechanical energy for dough extension and hence to more breakages in the polymer network. The effects of water content may be understood in terms of the behaviour EU and EB: as water content increases it would be expected that the loop to train ratio would increase. This would result in easier extension

Relative number of polymers

HMW SUBUNITS OF WHEAT GLUTENIN 1.00

287

BROKEN POLYMERS

0.80

UNSTRETCHED POLYMERS

0.60

STRETCHED POLYMERS

0.40

0.20

0.00

0

50

100

150

200

250

300

350

400

450

Number of revolutions

Resistance to extension

FIG. 26. The effects of the number of mixer revolutions on the populations of unstretched, stretched and broken polymers in a dough.

0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00

1

51

101 151 201 251 301 351 401 451 Number of revolutions

FIG. 27. The effect of changing the rate constant for polymer breakage (k2) on the shape of the mixing curve. The flattened curve has a low value of k2 and the curve with a sharper maximum has a higher value of k2. Values of k1 are kept constant and resistance values are normalized to be on a similar scale.

of the unstretched polymers and would therefore decrease the resistance to extension of the dough. This would in turn result in a lower maximum resistance, since the total resistance is that of the stretched polymers plus that of the other polymers. It would also result in a lower breakage rate of the polymers since less energy would be required to extend the dough. The

288

P. R. SHEWRY ET AL.

changes in resistance to extension after resting can also be understood. Energy input into the doughs will result in an increase in the number of broken polymers. After resting, the polymer network will relax to its equilibrium configuration but the number of polymers in the network will have decreased. This will result in decreased resistance to extension, as the number of polymers in the network available to be stretched will be lowered. In addition, there will be a reduced degree of extension to break since the network integrity will be reduced by the loss of polymers from it. E. IS THE LOOP AND TRAIN HYPOTHESIS CONSISTENT WITH OTHER MODELS?

The combination of the loop and train model with a simple kinetic approach offers an explanation of the role of the HMW subunits in determining many of the phenomena observed in dough viscoelasticity and mixing. By recognizing the importance of network formation by covalent linkages and the role of hydrogen bonding in understanding of the elastic component of dough behaviour, a testable hypothesis is formulated that is consistent with both mechanical and spectroscopic behaviour. Some other theories of the mechanical behaviour of doughs have also been formulated. (For a thoughtful review of some earlier work see El-Dash, 1991.) Ewart (1989) has formulated a hypothesis that assumes that the “glutenin molecules” are arranged in chains of folded units; on stretching these units unfold and the elastic restoring force is said to be entropic. As pointed out above, this argument is based on a misunderstanding. The glutenin molecules are assumed to be oriented by the shearing process in dough and overlap between oriented molecules is suggested to augment noncovalent interactions between these molecules. This proposal is not dissimilar to the loop and train model in that it recognizes the role of noncovalent interactions and a role for the stretching of the molecules in elasticity. However, explicit mechanisms are lacking and, apart from the prediction of the effects of shear on orientation, little is offered in the way of testable predictions. Singh and MacRitchie (2001) consider the applications of polymer science to the properties of gluten. Some of the problems of their approach have been discussed above. The core of their argument is that resistance to extension of the dough arises through chain entanglements. However, these are not the generalized “entanglements” of classical polymer theory in which interactions between polymers are represented by a constraining tube (Doi and Edwards, 1986; Belton, 1999b), but specific points at which it appears that the polymers form tangles analogous to those observed on the macroscopic scale in string. Slippage through these entanglements is

HMW SUBUNITS OF WHEAT GLUTENIN

289

proposed to cause resistance to extension. The authors then consider the effects of molecular weight on slippage and the draw ratio of the material. The problem with this approach is that it is nonspecific; it does not explain why any polymer of suitable molecular weight does not behave like dough or why specific effects of D2O and esterification are observed. The effects of water are discussed in terms of the glass transition temperature of polymer–water systems and it is suggested that zein proteins of maize do not show dough-like properties at room temperature because of glass transition effects. In fact, the glass transition temperatures of wheat gluten are below 20°C for gluten when it has greater than 15% water content (Kalichevsky et al., 1992; Noel et al., 1995) and similar figures are reported for zein (Madeka and Kokini, 1996). Arguments about a role for glass transitions in dough rheology seem therefore to be untenable. El-Dash (1991) has proposed a mechanism of viscoelasticity that has some features in common with the loop and train model and with Ewart’s model. He emphasizes that the relative rarity of cysteine residues in glutenin ensures that disulphide interchange is unlikely to be a cause of network formation and assumes that the disulphide-bonded network preexists. Noncovalent interactions are considered to be responsible for holding coiled, but not necessarily ordered, molecules together in chains and sheets; these are interconnected by the relatively rare disulphide bridges. Mixing enhances noncovalent intermolecular interactions by bringing more molecules into contact with each other while continued mechanical stress causes unfolding of the coiled molecules. The recovery of the coiled state is the mechanism of elasticity. Interestingly, he points out that any polymeric system in which covalent cross-links exist together with noncovalent intermolecular interactions might be made to demonstrate similar viscoelasticity to gluten. All of the other models proposed therefore offer some insights, but none of them have been sufficiently developed to make testable predictions about the nature of dough viscoelasticity or to generate testable hypotheses. IX. CONCLUSIONS AND FUTURE PROSPECTS

The initial demonstration of correlations between HMW subunit composition and breadmaking quality of wheat stimulated a wide range of studies (biochemical, biophysical and genetic) that demonstrated that these proteins play a crucial role in determining the viscoelastic properties of hydrated gluten and the mixing properties of dough. Furthermore, the relationship between the allelic composition of the HMW subunits and dough mechanical properties has been shown to result

290

P. R. SHEWRY ET AL.

from both quantitative (i.e. subunit amount) and qualitative effects, the latter relating to differences in the structures and properties of individual subunits. In this respect the glutamine-rich repeated sequence domain may be particularly important in establishing inter-molecular interactions that contribute to both cohesiveness and elasticity. Although the primary and secondary structures of the HMW subunits have been studied in some detail, we still know very little about their higher-order structures. However, it is clear that the functional state of gluten consists of proteins assembled into polymers or aggregates, stabilized by both covalent disulphide bonds and noncovalent forces (principally hydrogen bonds). These protein assemblies are essential for the expression of the mechanical properties; isolated HMW subunits are not viscoelastic. Despite the paucity of our knowledge of the structures of these assemblies, it is possible to propose a plausible physicochemical mechanism for the role of the HMW subunits in gluten viscoelasticity, based on studies of the structures and interactions of isolated subunits. In addition to the genetic effects discussed above, it is clear that environmental factors also affect the structure and properties of the gluten polymers and hence grain processing quality. This results in extensive year-to-year and site-to-site variation, posing problems for the grain utilizing industries. Furthermore, there is also evidence for the existence of genotype × environment (G × E) interactions, meaning that cultivars may be differentially affected by environmental factors. The elucidation of the detailed structures and properties of the glutenin polymers is clearly an important goal for wheat scientists. This will not only require the analysis of glutenin polymers isolated from dough, but also the determination of their mechanisms of synthesis and its regulation, their interactions with other dough components (notably lipids, starch and arabinoxylans) and dynamic aspects of their changes in structure during grain maturation and dough mixing. This is clearly a formidable challenge, but one that can be faced using the combination of modern genetic, molecular, biochemical and biophysical approaches discussed above.

ACKNOWLEDGEMENTS

IACR receives grant-aided support from the Biotechnology and Biological Sciences Research Council of the United Kingdom. Much of the work carried out in the authors’ laboratories was supported by the European Union FAIR program CT96-1176 “Improving the Quality of EU Wheats for the Food Industry – EUROWHEAT ”.

HMW SUBUNITS OF WHEAT GLUTENIN

291

REFERENCES Alberti, E., Humpfer, E., Spraul, M., Gilbert, S. M., Tatham, A. S., Shewry, P. R. and Gil, A. M. 2001. A high resolution 1H magic angle spinning NMR study of a high Mr subunit of wheat glutenin. Biopolymers 58, 33–45 Altpeter, F., Vasil, V., Srivastava, V. and Vasil, I. K. 1996. Integration and expression of the highmolecular-weight glutenin subunit 1Ax1 gene into wheat. Nature Biotechnol. 14, 1155–1159. Alvarez, M. L., Guelman, S., Halford, N. G., Lustig, S., Reggiardo, M. I., Ryabushkina, N., Shewry, P., Stein, J. and Vallejos, R. H. 2000. Silencing of HMW glutenins in transgenic wheat expressing extra HMW subunits. Theoret. Appl. Genet. 100, 319–327. Anderson, O. D. and Blechl, A.E. 2000. Transgenic wheat – challenges and opportunities. In “Transgenic Cereals” (L. O’Brien and R. Henry, eds), pp. 1–27. American Association of Cereal Chemists, St Paul, MN. Anderson, O. D. and Greene, F. C. 1989. The characterisation and comparative analysis of highmolecular-weight glutenin genes from genomes A and B of hexaploid wheat. Theoret. Appl. Genet. 77, 689–700. Anderson, O. D. and Greene, F. C. 1997. The α-gliadin gene family. II. DNA and protein sequence variation, subfamily structure, and origins of pseudogenes. Theoret. Appl. Genet. 97, 59–65. Anderson, O. D., Greene, F.C., Yip, R. E., Halford, N.G., Shewry, P.R. and Malpica-Romero J.-M. 1989. Nucleotide sequences of the two high-molecular-weight glutenin genes from the D-genome of a hexaploid bread wheat, Triticum aestivum L. cv Cheyenne. Nucl. Acids Res. 17, 461–462. Antes, S. and Wieser H. 2001a. Effects of high and low molecular weight glutenin subunits on rheological dough properties and breadmaking quality of wheat. Cereal Chem. 78, 157–159. Antes, S. and Wieser H. 2001b. Reoxidation behavior of wheat and rye glutenin subunits. Cereal Chem. 78, 8–13. Arkin, H., Yasar, F., Celik, T., Celik, S. and Köksel, H. 2000. Multicanonical simulation of five tetrapeptide sequences in the central domain of HMW glutenin. Int. J. Mod. Phys. C. 11, 1596–1606. Arkin, H., Yasar, F., Celik, T., Celik, S. and Köksel, H. 2001. Molecular modeling of two hexapeptide repeat motifs of HMW glutenin subunits. Int. J. Mod. Phys. C. 12, 281–292. Aussenac, T., Carceller, J.L. and Kleiber, D. 2001. Changes in SDS solubility of glutenin polymers during dough mixing and resting. Cereal Chem. 78, 39–45. Bangur, R., Batey, I. L., McKenzie, E. and MacRitchie, F. 1997. Dependence of extensograph parameters on wheat protein compostion measured by SE–HPLC. J. Cereal Sci. 25, 237–241. Barcelo, P., Rasco-Gaunt, S., Thorpe, C. and Lazzeri, P. A. 2001. Transformation and gene expression. Adv. Bot. Res. 34, 59–126. Barro, F., Rooke, L., Békés, F., Gras, P., Tatham, A. S., Fido, R. J., Lazzeri, P., Shewry, P. R. and Barcelo, P. 1997. Transformation of wheat with HMW subunit genes results in improved functional properties. Nature Biotechnol. 15, 1295–1299. Beccari 1745. De Frumento. De Bononiensi Scientiarum et Artium Instituto atque Academia Commentarii, II, Part I, pp. 122–127. Békés, F. and Gras, P. W. 1994. Effect of individual glutenin subunits on mixing properties. In “Gluten 93, Association of Cereal Research, Detmold”, pp. 170–179. Békés, F., Gras, P. W. and Gupta, R. B. 1994a. Mixing properties as a measure of reversible reduction and oxidation of doughs. Cereal Chem. 71, 44–50.

292

P. R. SHEWRY ET AL.

Békés, F., Gras, P. W., Gupta R. B., Hickman, D. R. and Tatham, A. S. 1994b. Effects of a high Mr glutenin subunit (1Bx20) on the dough mixing properties of wheat flour. J. Cereal Sci. 19, 3–7 Bekkers, A. C., van Dijk, A., de Boef, E., van Swieten, E., Robillard, G. and Hamer, R. J. 1996. HMW glutenins: structure–function relationships step by step. In “Gluten 96 – Proceedings of the 6th International Wheat Gluten Workshop”, Sydney, September 1996, pp. 190–194. Belton, P. S. 1995. A hypothesis concerning the elasticity of high molecular weight subunits In “Wheat Kernel Proteins: Molecular and Functional Aspects”, Università Degli Studi Della Tuscia, Consiglio Nazionale Delle Ricerche, Viterbo, Italy, pp. 159–165. Belton, P. S. 1999a. On the elasticity of wheat gluten. J. Cereal Sci. 29, 103–107 Belton, P. S. 1999b. NMR of Food Biopolymers. In “Advances in Magnetic Resonance in Food Science”, pp. 115–125. Royal Society of Chemistry, Cambridge. Belton, P. S. and Gil, A. M. 1993. Proton nmr line shapes and transverse relaxation in a hydrated barley protein. J. Chem. Soc. Faraday Trans. 89, 4203–4206. Belton, P. S., Colquhoun, I .J., Field, J. M., Grant, A., Shewry, P. R. and Tatham, A. S. 1994. 1H and 2H NMR relaxation studies of a high Mr wheat glutenin and comparison with elastin. J. Cereal Sci. 19, 115–121 Belton, P. S., Colquhoun, I. J., Grant, A., Wellner, N., Field, J. M., Shewry, P. R. and Tatham, A. S. 1995. FT-IR and NMR studies on the hydration of a high Mr subunit of glutenin. Int. J. Biol. Macromol. 17, 74–80. Belton, P. S., Gil A. M., Grant A., Alberti E. and Tatham A. S. 1998. Proton and carbon NMR measurements of the effects of hydration on the wheat protein omega gliadin. Spectrochim. Acta A 54, 955–966. Belton, P. S., Wellner, N., Mills, E. N. C., Grant, A. and Jenkins, J. 2000. Do high molecular weight subunits form polar zippers? In “Wheat Gluten” (P. R. Shewry and A. S. Tatham, eds), pp. 363–367. Royal Society of Chemistry, Cambridge, UK. Benedettelli, S., Margiotta, B., Porceddu, E., Ciaffi, M. and Lafiandra, D. 1992. Effect of the lack of proteins controled by genes at the Gli-D1/Glu-D3 loci on the breadmaking quality of wheat. J. Cereal Sci. 16, 69–79. Bietz, J. A., Shepherd, K. A. and Wall, J. S. 1975. Single–kernel analysis of glutenin: use in wheat genetics and breeding. Cereal Chem. 52, 513–532. Blechl, A. E. and Anderson, O. D. 1996. Expression of a novel high-molecular-weight glutenin subunit gene in transgenic wheat. Nature Biotechnol. 14, 875–879. Blechl, A. E., Le, H. Q. and Anderson, O. D. 1998. Engineering changes in wheat flour by genetic transformation. J. Plant Physiol. 152, 703–707. Branlard, G and Dardevet, M. 1985. Diversity of grain protein and bread wheat quality. II. Correlation between high molecular weight subunits of glutenin and flour quality characteristics. J. Cereal Sci. 3, 345–354. Branlard, G., Pierre, J. and Rousset, M. 1992. Selection indices for quality evaluation in wheat breeding. Theoret. Appl. Genet. 84, 57–64. Buonocore, F., Bertini, L., Ronchi, C., Bekes, F., Caporale, C., Lafiandra, D., Gras, P., Tatham, A. S., Greenfield, J.A., Halford, N.G. and Shewry, P.R. 1998. Expression and functional analysis of Mr 58000 peptides derived from the repetitive domain of high molecular weight glutenin subunit 1Dx5. J. Cereal Sci. 27, 209–215. Burnouf, T. and Bouriquet, R. 1980. Glutenin subunits of genetically related European hexaploid wheat cultivars: their relation to breadmaking quality. Theoret. Appl. Genet. 58, 107–111. Campbell, W. P., Wrigley, C. W., Cressey, P. J. and Slack, C. R. 1987. Statistical correlations between quality attributes and grain–protein composition for 71 hexaploid wheats used as breeding parents. Cereal Chem. 64, 293–299.

HMW SUBUNITS OF WHEAT GLUTENIN

293

Candler, D., Szabo, C., Murray, D. and Bekes, F. 1996. In vitro polymerization of glutenin subunit using protein disulphide isomerase. In “Gluten 96 – Poceedings of the 6th International Wheat Gluten Workshop”, Sydney, September 1996, pp. 133–136. Cheng, M., Fry, J. E., Pang, S., Zhou, H., Hironaka, C. M., Duncan, D. R., Conner, T. W. and Wan, Y. 1997. Genetic transformation of wheat mediated by Agrobacterium tumefaciens. Plant Physiol. 115, 971–980. Chou, P. Y. and Fasman, G. D. 1978. Empirical predictions of proteins conformations. Ann. Rev. Biochem. 47, 251–276. Cornec, M., Popineau, Y. and Lefebvre, J. 1994. Characterisation of gluten subfractions by SE–HPLC and dynamic rheological analysis in shear. J. Cereal Sci. 19, 131–139. Cressey, P. J., Campbell, W. P., Wrigley, C. W. and Griffin, W. B. 1987. Statistical correlations between quality attributes and grain–protein composition for 60 advanced lines of crossbred wheat. Cereal Chem. 64, 299–301. D’Ovidio R., Porceddu, E. and Lafiandra, D. 1994. PCR analysis of genes encoding allelic variants of high-molecular-weight glutenin subunits at the Glu–D1 locus. Theoret. Appl. Genet. 88, 175–180. Dachkevitch, T. and Autran, J. C. 1989. Prediction of baking quality of bread wheats in breeding programs by size-exclusion high performance liquid chromatography. Cereal Chem. 66, 448–456. Dickinson, E. 1992. “An Introduction to Food Colloids.” Oxford University Press, Oxford. Doekes, G. J. and Wennekes, L. M. J. 1982. Effect of nitrogen-fertilization on quantity and composition of wheat-flour protein. Cereal Chem. 59, 276–278 Doi, M. and Edwards, S. 1986. “The Theory of Polymer Dynamics.” Oxford Science Publications, Oxford. Dong, H., Cox, T. S., Sears, R. G. and Lookhart, G. 1991. High molecular weight glutenin genes: effects on quality in wheat. Crop Sci. 31, 974–979. Dong, H., Sears, R. G, Cox, T. S., Hoseney, R. C., Lookhart, G. and Shogren M. D. 1992. Relationships between protein composition and mixograph and loaf characteristics in wheat. Cereal Chem. 69, 132–136. El-Dash, A. 1991. Molecular structure of gluten and viscoelastic properties of dough: a new concept. In “Proceedings of the First Brazilian Congress on Proteins” (B. Olivera, and V. Sgarbieri, eds), pp. 513–530. Editoria da Unicamp, Campinas, Brazil. Eliasson, A. C. and Lundh, G. 1989. Rheological and interfacial behavior of some wheat protein fractions. J. Texture Studies 20, 431–441. Ewart, J. A. D. 1989. Hypothesis for how linear gluten is held in dough. Food Chem. 32, 135–150. Feeney, K. A., Wellner, N., Gilbert, S., Halford, N. G., Tatham, A. S., Shewry, P. R. and Belton, P. S. 2002. Molecular structures and interactions of repetitive peptides based on wheat glutenin subunits depend on chain length. (Submitted.) Fido, R., Bekes, F., Gras, P. W. and Tatham, A. S. 1997. Effects of α-, β-, γ-, and ω-gliadins on the dough mixing properties of wheat flour. J. Cereal Sci. 26, 271–277. Fido, R. J., Darlington, H. F., Cannell, M. E., Jones, H., Tatham, A. S., Békés, F. and Shewry, P. R. 2000. Expression of HMW glutenin subunits in field grown transgenic wheat. In “Wheat Gluten” (P.R. Shewry and A.S. Tatham, eds), pp. 77–80. Royal Society of Chemistry, Cambridge. Field, J. J., Shewry, P. R., Miflin, B. J. and March, J. F. 1982. The purification and characterization of homologous high molecular weight storage proteins from grain of wheat, rye and barley. Theoret. Appl. Genet. 62, 329–336. Field, J. M., Tatham, A. S. and Shewry, P. R. 1987. The structure of a high Mr subunit of durumwheat (Triticum durum) gluten. Biochem. J. 247, 215–221. Finney, K. F. 1985. Experimental breadmaking studies, functional (breadmaking) properties, and related gluten protein fractions. Cereal Foods World 30, 794–801.

294

P. R. SHEWRY ET AL.

Flavell, R. B., Goldsbrough, A. P., Robert, L. S., Schnick, D. and Thompson, R. D. 1989. Genetic variation in wheat HMW glutenin subunits and the molecular basis of breadmaking quality. Bio/Technology 7, 1281–1285. Flory, P. J. 1953. “Principles of Polymer Chemistry,” pp. 507–509. Cornell University Press, Ithaca, NY. Gao, L., Ng, P. K. W. and Bushuk, W. 1992. Structure of glutenin based on farinograph and electrophoresis results. Cereal Chem. 69, 452–455. Gil, A. M., Alberti, E. and Santos, D. 2001. An insight into the structure of foods using 13C and 1H magic angle spinning (MAS) NMR: application to wheat dough. In “Magnetic Resonance in Food Science: A View to the Future” (G.A. Webb, P.S. Belton, A.M. Gil, and I. Delgadillo, eds), p. 43. Royal Society of Chemistry, Cambridge. Gilbert, S. M., Wellner, N., Belton, P. S., Greenfield, J. A., Siligardi, G., Shewry, P. R. and Tatham, A. S. 2000. Expression and characterisation of a highly repetitive peptide derived from a wheat seed storage protein. Biochim. Biophys. Acta 1479, 135–146. Goldenberg, D. P. 1992. Mutational analysis of protein folding and stability. In “Protein Folding” (T.E. Creighton, ed.), pp. 353–403. W. H. Freeman, New York. Goldenberg, D. P. and Creighton, T. E. 1984. Gel electrophoresis in studies of protein conformation and folding. Anal. Biochem. 138, 1–18. Goldsbrough, A. P., Bulleid, N. J., Freedman, R. B. and Flavell, R. B. 1989. Conformational differences between two wheat (Triticum aestivum) high-molecular-weight glutenin subunits are due to a short region containing six amino acid differences. Biochem. J. 263, 837–842. Grant, A., Belton, P. S., Colquhoun, I.J., Parker, M. L., Plijter, J., Shewry, P. R., Tatham, A. S. and Wellner, N. 1999. The effects of temperature on the sorption of water by wheat gluten determined using deuterium NMR. Cereal Chem. 76, 219–226. Gras, P. W., Carpenter, H. C. and Anderssen, R. S. 2000. Modelling the developmental rheology of wheat–flour dough using extension tests. J. Cereal Sci. 31, 1–13. Graveland, A., Bosveld, P., Lichtendonk, W. J. and Moonen, J. H. E. 1980. Superoxide involvement in the reduction of disulphide bonds of wheat gel proteins. Biochem. Biophys. Res. Commun. 93, 1189–1195. Graveland, A., Bosveld, P., Lichtendonk, W. J., Marseille, J. P., Moonen, J. H. E. and Scheepstra, A. 1985. A model for the molecular structure of the glutenins from wheat flour. J. Cereal Sci. 3, 1–16. Graybosch, R. A., Peterson, C. J., Baenziger, P. S. and Shelton D. R. 1995. Environmental modification of hard red winter wheat flour protein composition. J. Cereal Sci. 22, 45–51. Gupta, R. B. and MacRitchie, F. 1994. Allelic variation at glutenin subunit and gliadin loci, Glu1, Glu-3 and Gli-1 of common wheats. II. Biochemical basis of the allelic effects on dough properties. J. Cereal Sci. 19, 19–29. Gupta R. B., Békés, F. and Wrigley C. W. 1991a. Prediction of physical dough properties from glutenin subunit composition in bread wheats: correlation studies. Cereal Chem. 68, 328–333. Gupta, R. B., MacRitchie, F., Shepherd, K. W. and Ellison, F. 1991b. Relative contribution of LMW and HMW glutenin sununits to dough strength and dough stickiness of bread wheat. In “Gluten Proteins 1990” (W. Bushuk and R. Tkachuk, eds), pp. 71–80. American Association of Cereal Chemists, Saint–Paul, MN. Gupta R. B., Batey, I. L. and MacRitchie, F. 1992. Relationship between protein composition and functional properties of wheat flours. Cereal Chem. 69, 125–131. Gupta, R. B., Kahn, K. and MacRitchie, F. 1993. Biochemical basis of flour properties in bread wheats. I. Effects of variation in the quantitity and size distribution of polymeric proteins. J. Cereal Sci. 18, 23–41.

HMW SUBUNITS OF WHEAT GLUTENIN

295

Gupta, R. B., Paul, J. G., Cornish, G. B., Palmer G. A., Bekes F. and Rathjen A. J. 1994. Allelic variation at glutenin subunit and gliadin loci, Glu-1, Glu-3 and Gli-1, of common wheats. I. Its additive and interactions effects on dough properties. J. Cereal Sci. 19, 9–17. Gupta, R. B., Popineau, Y., Lefebvre, J., Cornec, M., Lawrence, J. G. and MacRitchie, F. 1995. Biochemical basis of flour properties in bread wheats. II. Changes in polymeric protein formation and dough/gluten properties associated with the loss of low Mr or high Mr glutenin subunits. J. Cereal Sci. 21, 103–116. Halford, N. G., Forde, J., Anderson, O. D., Greene, F. C. and Shewry, P. R. 1987. The nucleotide and deduced amino acid sequences of an HMW glutenin subunit gene from chromosome 1B of bread wheat (Triticum aestivum L.), and comparison with those of genes from chromosomes 1A and 1D. Theoret. Appl. Genet. 75, 117–126. Halford, N. G., Field, J. M., Blair, H., Urwin, P., Moore, K., Robert, L., Thompson, R., Flavell, R. B., Tatham, A. S. and Shewry, P. R. 1992. Analysis of HMW glutenin subunits encoded by chromosome 1A of bread wheat (Triticum aestivum L.) indicates quantitative effects on grain quality. Theoret. Appl. Genet. 83, 373–378. Hamada, A. S., McDonald, C. E. and Sibbitt, L. D. 1982. Relationship of protein fractions of spring wheat flour to baking quality. Cereal Chem. 59, 296–301. Hamer, R. J. and Lichtendonk, W. J. 1987. Structure–function studies on gluten proteins. Reassembly of glutenin protein after mixing. In “Proceedings of the 3rd International Workshop on Gluten Proteins” (R. Lasztity and F. Békés, eds), pp. 227–237. World Scientific Publishers, Singapore. Hargreaves, J., Popineau, Y., Cornec, M. and Lefebvre, J. 1996. Relation between aggregative, viscoelastic and molecular properties in gluten from genetic variant of bread wheat. Int. J. Biol. Macromol. 18, 69–75. He, G., Rooke, L., Steele, S., Békés, F., Gras, P., Tatham, A. S., Fido, R., Barcelo, P., Shewry, P. R. and Lazzeri, P. 1999. Transformation of pasta wheat (Triticum durum L.) with HMW glutenin subunit genes and modification of dough functionality. Mol. Breed. 5, 377–386. He, G. Y., D’Ovidio, R., Anderson, O. D., Fido, R., Tatham, A. S., Jones, H. D., Lazzeri, P. A. and Shewry, P. R. 2000. Modification of storage protein composition in transgenic bread wheat. In “Wheat Gluten” (P.R. Shewry and A.S. Tatham, eds), pp. 84–87. Royal Society of Chemistry, Cambridge. Hickman, D. R. 1995. Biochemical studies of the high molecular weight glutenin subunits of bread wheat. PhD Thesis, University of Bristol. Holt, L. M., Astin, R. and Payne, P. I. 1981. Structural and genetical studies on the highmolecular-weight subunits of wheat glutenin. Part II. Relative isoelectric points determined by two-dimensional fractionation in polyacrylamide gels. Theoret. Appl. Genet. 60, 237–243. Hoseney, R. C., Finney, K. F., Pomeranz, Y. and Shogren, M. D. 1969. Functional (breadmaking) and biochemical properties of wheat flour components. IV. Gluten protein fractionation by solubilizing in 70% ethanol and in dilute lactic acid. Cereal Chem. 46, 495–502. Huebner, F. R. and Wall, J. S. 1976. Fractionation and quantitative differences of glutenin from wheat varieties ranging in baking quality. Cereal Chem. 51, 228–240. Humphris, A. D. L., McMaster T. J., Miles, M. J., Gilbert, S. M., Shewry, P. R. and Tatham, A. S. 2000. An AFM study of the interactions of the HMW subunits of wheat glutenin. Cereal Chem. 77, 107–110. Janssen, A. M. 1992. “Obelisk and Katepwa wheat gluten, a study of factors determining bread making performance”. University of Groningen, Netherlands Janssen, A. M., van Vliet, T. and Vereijken, J. M. 1996. Rheological behaviour of wheat glutens at small and large deformations. Comparison of two glutens differing in bread making potential. J. Cereal Sci. 23, 19–31.

296

P. R. SHEWRY ET AL.

Johansson, E., Prieto–Linde, M. L. and Jönson, J. Ö. 2001. Effect of wheat cultivar and nitrogen application on storage protein composition and breadmaking quality. Cereal Chem. 78, 19–25. Kalichevsky, M. T., Jaroszkiewicz, E. M. and Blanshard, J. M. V. 1992. Glass transition of gluten1: gluten and gluten–sugar mixtures. Int. J. Biol. Macromol. 14, 257–266. Kasarda, D. D. 1988. Glutenin structure in relation to wheat quality. In “Wheat is Unique” (Y. Pomeranz, ed.), pp. 277–302. American Association of Cereal Chemists, St Paul, MN. Kasarda, D. D. 1994. Contrasting molecular models for HMW-GS. In “Wheat Kernel Proteins: Molecular and Functional Aspects,” Università Degli Studi Della Tuscia, Consiglio Nazionale Delle Recherche, Viterbo, Italy, pp. 63–68. Kasarda, D. D., King, G. and Kumonsinski, T.F. 1994. Comparison of β–spiral structures in wheat high molecular weight glutenin subunits and elastin, by molecular modelling. In “Proceedings of the Symposium on Molecular Modelling” (T.F. Kumonsinski and M. Liebman eds), pp. 209–220. American Chemical Society, Washington, DC. Keck, B., Köhler, P., and Wieser, H. 1995. Disulphide bonds in wheat gluten: cystine peptides derived from gluten proteins following peptic and thermolytic digestion. Z. Lebensm. Unters. Forsch. 200, 432–439. Khatkar, B. S., Bell, A. E. and Schofield, J. D. 1995. The dynamic rheological properties of glutens and gluten sub-fractions from wheat of good and poor bread making quality. J. Cereal Sci. 22, 29–44. Köhler, P., Belitz, H.-D. and Wieser, H. 1991. Disulphide bonds in wheat gluten: isolation of a cystine peptide from glutenin. Z. Lebensm. Unters. Forsch. 192, 234–239. Köhler, P., Belitz, H.-D. and Wieser, H. 1993. Disulphide bonds in wheat gluten: further cystine peptides from high molecular weight (HMW) and low molecular weight (LMW) subunits of glutenin and from γ-gliadins. Z. Lebensm. Unters. Forsch. 196, 239–247. Köhler, P., Keck, B., Müller, S. and Wieser, H. 1994. Disulphide bonds in wheat gluten. In “Wheat Kernel Proteins: Molecular and Functional Aspects”, Università Degli Studi Della Tusci, Consiglio Nazionale Delle Recherche, Viterbo, Italy, pp. 45–54. Köhler, P., Keck-Gassenmeier, B., Wieser, H. and Kasarda, D. D. 1997. Molecular modelling of the N-terminal regions of high molecular weight glutenin subunits 7 and 5 in relation to intramolecular disulphide bond formation. Cereal Chem. 74, 154–158. Kolster, P., van Euwijk, F. A. and van Gelder, W. M. J. 1991. Additive and epistatic effects of allelic variation at the high molecular weight glutenin subunit loci in determining the bread-making quality of breeding lines of wheat. Euphytica 55, 277–285. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophge T4. Nature 227, 680–685. Lafiandra D., Benedettelli S., Porceddu E., 1988. Null forms for storage proteins in bread wheat and durum. In “Proceedings of the 7th International Wheat Genetic Symposium”, Cambridge, UK (T.E. Miller and R.M.D. Koebner, eds), pp. 963–968. Lafiandra D., Turchetta, T., D’Ovidio, R., Anderson, O.D., Facchiano, A.M. and Colonna, G. 1999. Conformational polymorphism of high Mr glutenin subunits detected by transverse urea gradient gel electrophoresis. J. Cereal Sci. 30, 97–104. Lafiandra D., Masci, S., D’Ovidio, R. and Margiotta, B. 2000. The genetics of wheat gluten proteins: an overview. In “Wheat Gluten” (P.R. Shewry, and A.S. Tatham eds), pp. 3–10. Royal Society, Cambridge. Lagudah, E. S., MacRitchie, F. and Halloran, G. M. 1988. The influence of high-molecularweight subunits of glutenin from Triticum tauschii on flour quality of synthetic hexaploid wheat. J. Cereal Sci. 5, 129–138. Lawrence G. J. and Shepherd, K. W. 1980. Variation in glutenin protein subunits of wheat. Aust. J. Biol. Sci. 33, 221–233.

HMW SUBUNITS OF WHEAT GLUTENIN

297

Lawrence G. J. and Payne, P. I. 1983. Detection by gel electrophoresis of oligomers formed by the association of high-molecular-weight glutenin protein subunits of wheat endosperm. J. Exp. Bot. 34, 254–267. Lawrence, G. J., Moss, H. J., Shepherd, K. W. and Wrigley, C. W. 1987. Dough quality of biotypes of eleven Australian wheat cultivars that differ in high-molecular-weight glutenin subunit composition. J. Cereal Sci. 6, 99–101. Lawrence, G. J., MacRitchie, F. and Wrigley, C. W. 1988. Dough and baking quality of wheat lines deficient in glutenin subunits controlled by the Glu-A1, Glu-B1 and Glu-D1 loci. J. Cereal Sci., 7, 109–112. Lefebvre, J., Popineau, Y., Deshayes, G., Lavenant, L. 2000. Temperature–induced changes in the dynamic rheological behavior and size distribution of polymeric proteins for glutens from wheat near–isogenic lines differing in HMW glutenin subunit composition. Cereal Chem. 77, 193–201. Levy A. A., Galili G. and Feldman M., 1988. Polymorphism and genetic control of high molecular weight glutenin subunits in wild tetraploid wheats Triticum turgidum var. dicoccoides. Heredity 61, 63–72. Lindsay, M. P., and Skerrit, J. H. 1999. The glutenin macropolymer of wheat flour doughs: structure–function perspectives. Trends Food Sci. Technol. 10, 247–253. Lindsay, M. P., Tamas, L., Appels, R. and Skerrit, J. H. 2000. Direct evidence that the number and location of cysteine residues affect glutenin polymer structure. J. Cereal Sci. 31, 321–334. Lukow, O. M., Payne P. I. and Tkachuk, R. 1989. The HMW glutenin subunit composition of Canadian wheat cultivars and their association with bread-making quality. J. Sci. Food Agric. 46, 451–460. Lundh, G. and MacRitchie, F. 1989. SE–HPLC characterisation of gluten protein fractions varying in bread–making potentials. J. Cereal Sci. 10, 247–253. Mackie, M. A., Sharp, P. J. and Lagudah, E. S. 1996. The nucleotide and derived amino acid sequence of a HMW glutenin gene from Triticum tauschii and comparison with those from the D genome of bread wheat. J. Cereal Sci. 24, 73–78. MacRitchie, F. 1985. Studies of the methodology for fractionation and reconstitution of wheat flours. J. Cereal Sci. 3, 221–230. MacRitchie, F. 1987. Evaluation of contributions from wheat protein fractions to dough mixing and breadmaking. J. Cereal Sci. 6, 259–268. MacRitchie, F. 1989. Identifying quality related components of wheat flours. Cereal Foods World 34, 548–552. Madeka, H and Kokini, J. L. 1996. Effect of glass transition and cross-linking on rheological properties of zein: development of a preliminary state diagram. Cereal Chem. 73, 433–438 Manley, N, Randall, P. G. and McGill, A. E. J. 1992. The prediction of dough properties of South African wheat cultivars by SDS-PAGE analysis of HMW glutenin subunits. J. Cereal Sci. 15, 39–47. Margiotta, B., Urbano, M., Colaprico, G., Turchetta, T. and Lafiandra, D. 1998. Variation of high molecular weight glutenin subunits in tetraploid wheats of genomic formula AAGG. In “Proceedings of the 9th International Wheat Genetic Symposium” (A.E. Slinkard, ed.), Vol. 4, pp. 195–197. University Extension Press, University of Saskatchewan, Saskatoon, Canada. Margiotta, B., Pfluger, L., Roth, M. R., MacRitchie, F. and Lafiandra, D. 2000. Isogenic bread wheat lines differing in number and type of high Mr glutenin subunits. In “Wheat Gluten” (P.R. Shewry and A.S. Tatham eds), pp. 29–33. Royal Society of Chemistry, Cambridge. Martín, A., Alvarez, J. B., Martín, L. M., Barro, F. and Ballesteros, J. 1999. The development of tritordeum: A novel cereal for food processing. J. Cereal Sci. 30, 85–95. Matsushima, N., Creutz, C. E. and Kretsinger, R. H. 1990. Polyproline, β–turn helices. Novel

298

P. R. SHEWRY ET AL.

secondary structures proposed for the tandem repeats within rhodopsin, synaptophysin, synexin, gliadin, RNA polymerase II, and gluten. Proteins 7, 125–155. Matsushima, N., Danno, G., Sasaki, N. and Izumi, Y. 1992. Small angle X-ray scattering study by synchrotron orbital radiation reveals that high molecular weight subunit of glutenin is a very anisotropic molecule. Biochem. Biophys. Res. Commun. 186, 1057–1064. Miles, M. J., Carr, H. J., McMaster, T., Belton, P. S., Morris, V. J., Field, J. M., Shewry, P. R. and Tatham, A. S. 1991. Scanning tunnelling microscopy of a wheat gluten protein reveals details of a spiral supersecondary structure. Proc. Natl Acad. Sci. USA 88, 68–71. Moonen, J. H. E., Scheepstra, A. and Graveland, A. 1982. Use of SDS-sedimentation test and SDS-polyacrylamide gel electrophoresis for screening breeders’ samples of wheat for breadmaking quality. Euphytica 31, 677–690. Moonen, J. H. E., Scheepstra, A. and Graveland, A. 1983. The positive effects of the high molecular weight subunits 3 + 10 and 2* of glutenin on the breadmaking quality of wheat cultivars. Euphytica 32, 735–742. Mosleth, E. and Uhlen A. K. 1991. Association between the composition of gliadins and HMW glutenin subunits and the gluten quality in wheat (Triticum aestivum L.). In “Gluten Proteins 1990” (W. Bushuk and R. Tkachuk, eds), pp. 112–128. American Association of Cereal Chemists, Saint Paul, MN. Nagamine, T., Kai, Y., Takayama, T., Yanagisawa, T. and Taya, S. 2000. Allelic variation at the Glu-1 and Glu-3 loci in southern Japanese wheats, and its effects on gluten properties. J. Cereal Sci. 32, 129–135. Ng, P. K. W. and Bushuk W. 1988. Statistical relationships between high molecular weight subunits of glutenin and breadmaking quality of Canadian grown wheats. Cereal Chem. 65, 408–413. Nicolas, Y. (1997) Les prolamines de blé: extraction exhaustive et développement de dosages chromatographiques en phase inverse et de dosages immunochimiques à l’aide d’anticorps anti-peptides. PhD thesis, University of Nantes. Noel, T. R., Parker, R., Ring, S. G. and Tatham, A. S. 1995. The glass–transition behaviour of wheat gluten proteins. Int. J. Biol. Macromol. 17, 81–85. Oobtake, M. and Tatsuo, O. 1993. Hydration and heat stability effects on protein unfolding Progr. Biophys. Mol. Biol. 59, 237–284. Orth, R. A. and Bushuk W. 1972. A comparative sudy of the proteins of wheats of diverse baking quality. Cereal Chem. 49, 268–275. Parchment, O., Tatham, A. S., Shewry, P. R. and Osguthorpe, D. J. 2001. Molecular modelling of the central repetitive domains of the high molecular weight subunits of wheat. Cereal Chem. 78, 658–662. Pastori, G. M., Steele, S. H., Jones, H. D. and Shewry, P. R. 2000. Transformation of commercial wheat varieties with high molecular weight glutenin subunit genes. In “Wheat Gluten” (P.R. Shewry and A.S. Tatham, eds), pp. 88–92. Royal Society of Chemistry, Cambridge. Payne, P. I. 1987. Genetics of wheat storage proteins and the effect of allelic variation on breadmaking quality. Ann. Rev. Plant Physiol. 38, 141–153. Payne, P. I. and Corfield, K. G. 1979. Subunit composition of wheat glutenin proteins isolated by gel filtration in a dissociating medium. Planta 145, 83–88. Payne, P. I. and Lawrence G. J. 1983. Catalogue of alleles for the complex gene loci, Glu-A1, Glu-B1, Glu-D1 which code for high-molecular-weight subunits of glutenin in hexaploid wheat. Cereal Res. Commun. 11, 29–35. Payne, P. I. and Seekings J. A. 1996. Characterisation of Galahad-6, Galahad-7 and Galahad-8, isogenic lines that contain only one HMW glutenin subunit. In “Gluten 96 – Proceedings of the 6th International Wheat Gluten Workshop”, Sydney, September 1996, pp. 14–17. Royal Australian Chemical Institute, North Melbourne, Australia.

HMW SUBUNITS OF WHEAT GLUTENIN

299

Payne, P. I., Corfield, K. G. and Blackman, J. A. 1979. Identification of a high molecular weight subunit of glutenin whose presence correlates with breadmaking quality in wheats of related pedigree. Theoret. Appl. Genet. 55, 153–159. Payne, P. I., Corfield, K. G., Holt, L. M. and Blackman, J. A. 1981a. Correlations between the inheritance of certain high molecular weight subunits of glutenin and breadmaking quality in progenies of six crosses of bread wheat. J. Sci. Food Agric. 32, 51–60. Payne, P. I., Holt, L. M., Worland, A. J. and Law, C. N. 1981b. Structural and genetical studies on the high-molecular-weight subunits of wheat glutenin. Part III. Telocentric mapping of the subunit genes on the long arms of the homoeologous group 1 chromosomes. Theoret. Appl. Genet. 63, 129–138. Payne, P. I., Holt, L. M., Jackson, E. A. and Law, C. N. 1984. Wheat storage proteins: their genetics and their potential for manipulation by plant breeding. Phil. Trans. Roy. Soc. Lond. B 304, 359–371. Payne, P. I., Holt, L. M., Harinder, K., Macartney, D. P. and Lawrence, G. J. 1987a. The use of near–isogenic lines with different HMW glutenin subunits in studying breadmaking quality and glutenin structure. In “Proceedings of the 3rd International Workshop on Gluten Proteins” (R. Lasztity and F. Békès, eds), pp. 216–226. World Scientific Publishers, Singapore. Payne, P. I., Nightingale, M. A., Krattiger, A. F. and Holt, L. M. 1987b. The relationship between HMW glutenin subunit composition and the breadmaking quality of British grown wheat varieties. J. Sci. Food Agric. 40, 51–65. Payne, P. I., Holt, L. M., Krattiger, A. F. and Carrillo, J. M. 1988a. Relationship between seed quality characteristics and HMW glutenin subunit composition determined using wheats grown in Spain. J. Cereal Sci. 7, 229–235. Payne, P. I., Lookhart, G. L. and Forsyth, S. A. 1988b. The high molecular weight glutenin subunit composition of two closely related lines of wheat that have contrasting breadmaking qualities. J. Cereal Sci. 8, 285–288. Perutz, M. F. 1996. Glutamine repeats and inherited neurodegenerative diseases: molecular aspects. Curr. Opin. Struct. Biol. 6, 848–858. Pezolet, M., Bonenfant, S., Douseau, F. and Popineau, Y. 1992. Conformation of wheat gluten proteins. FEBS Lett. 3, 247–250. Popineau, Y., Cornec, M., Lefebvre, J. and Marchylo, B. 1994a. Influence of high Mr glutenin subunits on glutenin polymers and rheological properties of gluten and gluten subfractions of near-isogenic lines of wheat Sicco. J. Cereal Sci. 19, 231–241. Popineau, Y., Bonenfant, S, Cornec, M. and Pezolet, M., 1994b. A study by infrared spectroscopy of the conformations of gluten proteins differing in their gliadin composition. J. Cereal Sci. 20,15–22. Popineau, Y., Deshayes, G., Lefebvre, J., Fido, R., Tatham, A. S. and Shewry, P. R. 2001. Prolamin aggregation, gluten viscoelasticity, and mixing properties of transgenic wheat lines expressing 1Ax and 1Dx high molecular weight glutenin subunit transgenes. J. Agric. Food Chem. 49, 395–401. Reddy, P. and Appels, R. 1993. Analysis of a genomic DNA segment carrying the wheat highmolecular-weight (HMW) glutenin Bx17 subunit and its use as an RFLP marker. Theoret. Appl. Genet. 85, 616–624. Rogers, W. J., Payne, P. I., Seekings, J. A. and Sayers, E. J. 1991. Effect on breadmaking quality of x-type and y-type high molecular weight subunits of glutenin. J. Cereal Sci. 14, 209–221. Rooke, L., Barro, F., Steele, S., Békés, F., Gras, P., Martin, A., Tatham, A. S., Fido, R., Lazzeri, P., Shewry, P. R. and Barcelo, P., 1999a. Improved functional properties by transformation of tritordeum with HMW glutenin subunit genes. Theoret. Appl. Genet. 99, 851–858. Rooke, L., Békés, F., Fido, R., Barro, F., Gras, P., Tatham, A. S., Barcelo, P., Lazzeri, P. and

300

P. R. SHEWRY ET AL.

Shewry, P. R. 1999b. Overexpression of a gluten protein in transgenic wheat results in greatly increased dough strength. J. Cereal Sci. 30, 115–120. Rousset, M., Carrillo, J. M., Qualset, C. O. and Kasarda, D. D. 1992. Use of recombinant inbred lines of wheat for study of associations of high-molecular-weight glutenin subunit alleles to quantitative traits. 2. Milling and bread-making quality. Theoret. Appl. Genet. 83, 403–412. Schropp, P. and Wieser, H. 1996. Effects of high molecular weight subunits of glutenin on the rheological properties of wheat gluten. Cereal Chem. 73, 410–413. Schropp, P., Belitz, H. D., Seilmeier, W. and Wieser, H. 1995. Eoxidation of high molecular weight subunits of glutenin. Cereal Chem. 72, 406–410. Seilmeier W., Belitz, H.-D. and Wieser, H. 1991. Separation and quantitative determination of high-molecular-weight subunits of glutenin from different wheat varieties and genetic variants of the variety Sicco. Z. Lebensm. Unters. Forsch. 192, 124–129. Shani, N., Rosenberg, N., Kasarda, D. D. and Galili, G. 1994. Mechanism of assembly of wheat high molecular weight glutenins inferred from expression of wild type and mutant subunits in transgenic tobacco. J. Biol. Chem. 269, 8924–8930. Shewry, P. R. and Tatham, A. S. 1997. Disulphide bonds in wheat gluten proteins. J. Cereal Sci. 25, 207–227. Shewry, P. R., Field, J. M., Faulks, A. J., Parmar, S., Miflin, B. J., Dietler, M. D., Lew, E. J-L. and Kasarda, D. D. 1984. Purification and N-terminal amino acid sequence analysis of high molecular weight (HMW) gluten polypeptides of wheat. Biochim. Biophys. Acta 788, 23–34. Shewry, P. R., Parmar, S. and Pappin, D. J. C. 1987. Characterization and genetic control of the prolamins of Haynaldia villosa: relationship to cultivated species of the Triticeae (Rye, Wheat and Barley). Biochem. Genet. 25, 309–325. Shewry, P. R., Halford, N. G. and Tatham, A. S. 1989. The high molecular weight subunits of wheat, barley and rye: genetics, molecular biology, chemistry and role in wheat gluten structure and functionality. Oxford Surveys of Plant Molec. Cell Biol. 6, 173–219. Shewry, P. R., Halford, N. G. and Tatham, A.S. 1992. The high molecular weight subunits of wheat glutenin. J. Cereal Sci. 15, 105–120. Shewry, P. R., Tatham, A. S. and Halford, N. 1999. The prolamins of the triticeae. In “Seed Proteins” (P.R. Shewry and R. Casey, eds), pp. 35–78. Kluwer Academic Publishers, Dordrecht. Shewry, P. R., Popineau, Y., Lafiandra, D. and Belton, P. 2001. Wheat gluten subunits and dough elasticity. Trends Food Sci. Technol. 11, 433–441. Shimoni, Y., Blechl, A. E., Anderson, O. D. and Galili, G. 1997. A recombinant protein of two high molecular weight glutenins alters gluten polymer formation in transgenic wheat. J. Biol. Chem. 272, 15488–15495. Singh, N. K. and Shepherd, K. W. 1988. Linkage mapping of the genes controlling endosperm proteins in wheat. II. Genes on the long arms of group 1 chromosomes. Theoret. Appl. Genet. 75, 642–650. Singh, H. and MacRitchie, F. 2001. Application of polymer science to properties of gluten. J. Cereal Sci. 33, 231–243. Singh, N. K., Donovan, G. R., Batey, I. L. and MacRitchie, F. 1990a. Use of sonication and sizeexclusion high-performance liquid chromatography in the study of wheat flour proteins. I. Dissolution of total proteins in the absence of reducing agents. Cereal Chem. 67, 150–161. Singh, N. K., Donovan, G. R. and MacRitchie, F. 1990b. Use of sonication and size-exclusion high-performance liquid chromatography in the study of wheat flour proteins. II. Relative quantity of glutenin as a measure of breadmaking quality. Cereal Chem. 67, 161–170. Sissons, M. J., Békés, F. and Skerritt, J. H. 1998. Isolation and functionality testing of low molecular weight glutenin subunits. Cereal Chem. 75, 30–36. Skerritt, J. H., Békés, F. and Murray, D. 1996. Isolation treatments and effects of gliadin and glutenin fractions on dough mixing properties. Cereal Chem. 73, 644–649.

HMW SUBUNITS OF WHEAT GLUTENIN

301

Skerritt, J. H., Hac, L., Lindsay, M. P. and Békés, F. 1999. Depolymerization of the glutenin macropolymer during mixing. II. Differences in retention of specific glutenin subunits. Cereal Chem. 76, 402–409. Sugiyama, T., Rafalski, A., Peterson, D., and Sol, D. 1985. A wheat HMW glutenin subunit gene reveals a highly repeated structure. Nucl. Acids Res. 13, 8729–8737. Taddei, G. 1819. Ricerche sul glutine del frumento. Giornale di fisica, chimica, e storia naturale, Brugnatelli 2, 360–361. Tahir, M., Pavoni, A., Tucci, G. F., Turchetta, T. and Lafiandra, D. 1996. Detection and characterization of a glutenin subunit with unusual high Mr at the Glu-A1 locus in hexaploid wheat. Theoret. Appl. Genet. 92, 654–659. Tao, H. P., Adalsteins, A. E. and Kasarda, D. D. 1992. Intermolecular disulphide bonds link specific high-molecular-weight glutenin subunits in wheat endosperm. Biochim. Biophys. Acta 1159, 13–21. Tatham, A. S. and Shewry, P. R. 2000. Elastomeric proteins: biological roles, structures and mechanisms. Trends Biochem. Sci. 25, 567–571. Tatham, A. S., Shewry, P. R. and Miflin, B. J. 1984. Wheat gluten elasticity: a similar molecular basis to elastin? FEBS Lett. 177, 205–208. Tatham, A. S., Miflin, B. J., and Shewry, P. R. 1985. The (β-turn conformation in wheat gluten proteins: relationship to gluten elasticity. Cereal Chem. 62, 405–412. Tatham, A. S., Drake, A. F. and Shewry, P. R. 1990. Conformational studies of synthetic peptides corresponding to the repetitive region of the high molecular weight (HMW) glutenin subunits of wheat. J. Cereal Sci. 11, 189–200. Tatham, A. S., Hayes, L., Shewry, P. R. and Urry, D. W. 2001. Wheat seed proteins exhibit a complex mechanism of protein elasticity. Biochim. Biophys. Acta. 1548, 187–193. Thompson, R. D., Bartels, D. and Harberd, N. P. 1985. Nucleotide sequence of a gene from chromosome 1D of wheat encoding a HMW glutenin subunits. Nucl. Acids Res. 13, 6833–6846. Thomson, N. H., Miles, M. J., Popineau, Y., Harries, J., Shewry, P. R. and Tatham, A. S., 1999. Small angle X-ray scattering of wheat seed storage proteins: α-, β- and γ-gliadins and the high molecular weight (HMW) subunits of glutenin. Biochim. Biophys. Acta 1430, 359–366. Tilley, K. A., Benjamin, R. E., Bagorogoza, K. E., Okot–Kotber, B. M., Prakash, O. and Kwen, H. 2001. Tyrosine cross-links: molecular basis of gluten structure and function. J. Agric. Food Chem. 49, 2627–2632. Tsiami, A. A., Bot A., and Agterof, W. G. M. 1997a. Rheology of mixtures of glutenin subfractions. J. Cereal Sci. 26, 279–287. Tsiami, A. A., Bot, A., Agterof, W. G. M. and Groot, R. D. 1997b. Rheological properties of glutenin subfractions in relation to their molecular weight. J. Cereal Sci. 26, 15–27. Uthayakumaran, S., Gras, P. W., Stoddard, F. L and Békés, F. 1999. Effect of varying protein content and glutenin-to-gliadin ratio on the functional properties of wheat dough. Cereal Chem. 76, 389–394. Uthayakumaran, S., Newberry, M., Keentok, M., Stoddard, F. L and Békés, F. 2000a. Basic rheology of bread dough with modified protein content and glutenin-to-gliadin ratios. Cereal Chem. 77, 744–749. Uthayakumaran, S., Stoddard, F. L, Gras, P. W. and Békés, F. 2000b. Optimized methods for incorporating glutenin subunits into wheat dough for extension and baking studies. Cereal Chem. 77, 731–736. Uthayakumaran, S., Stoddard, F. L, Gras, P. W., and Békés, F. 2000c. Effects of incorporated glutenins on functional properties of wheat doughs. Cereal Chem. 77, 737–743. Uthayakumaran, S., Tömösközi, S., Tatham, A. S., Savage, A.W. J., Gianibelli, M. C., Stoddard, F. L and Békés, F. 2001. Effects of gliadin fractions on functional properties of wheat dough depending on molecular size and hydrophobicity. Cereal Chem. 78, 138–141.

302

P. R. SHEWRY ET AL.

van Dijk, A. A., van Wijk, L. L., van Vliet, E., Haris, P., van Swieten, E., Tesser, G. I. and Robillard, G. T. 1996a. Structure characeterization of the central repetitive domain of high molecular weight gluten proteins. 1. Model studies using cyclic and linear peptides. Protein Sci. 6, 637–648. van Dijk, A. A., DeBoef, E., Bekkers, A., van Wijk, L. L., van Swieten, E., Hamer, R. J. and Robillard, G. T. 1996b. Structure characterization of the central repetitive domain of high molecular weight gluten proteins. 2. Characterization in solution and the dry state. Protein Sci. 6, 649–656. van Dijk, A. A., van Swieten, E., Kruize, I. T. and Robillard, G. T. 1998. Physical characterisation of the N-terminal domain of high-molecular-weight gluten subunit Dx5 from wheat. J. Cereal Sci. 28, 115–126. Vasil, V., Castillo, A. M., Fromm, M. E. and Vasil, I. K. 1992. Herbicide resistant fertile transgenic wheat plants obtained by microprojectile bombardment of regenerable embryonic callus. Bio/Technology 10, 667–674. Vasil, I.K., Bean, S., Zhao, J., McCluskey, P., Lookhart, G., Zhao, H.-P., Altpeter, F. and Vasil, V. 2001. Evaluation of baking properties and gluten protein composition of field grown transgenic wheat lines expressing high molecular weight glutenin gene 1Ax1. J. Plant Physiol. 158, 521–528. Veraverbeke, W. S., Larroque, O. R., Békés, F. and Delcour, J. A. 2000a. In vitro polymerization of wheat glutenin subunits with inorganic oxidizing agents. I. Comparison of single–step and stepwise oxidations of high molecular weight subunits. Cereal Chem. 77, 582–588. Veraverbeke, W. S., Larroque, O. R., Békés, F. and Delcour, J. A. 2000b. In vitro polymerization of wheat glutenin subunits with inorganic oxidizing agents. II. Stepwise oxidation of low molecular weight glutenin subunits and a mixture of high and low molecular weight glutenin subunits. Cereal Chem. 77, 589–594. Vereijken, J. M., Klostermann, V. L. C., Beckers, F. H. R., Spekking, W. T. J. and Graveland, A. (2000) Intercultivar variation in the proportions of wheat protein fractions and relation to mixing behaviour. J. Cereal Sci. 32, 159–167. Waines, J. G. and Payne P. I. 1987. Electrophoretic analysis of the high-molecular-weight glutenin subunits of Triticum monococcum, T. urartu and the A genome of bread wheat (T. aestivum). Theoret. Appl. Genet 74, 71–76. Wan, Y. F., Wang, D., Shewry, P. R. and Halford, N. G. 2002. Isolation and characterization of five novel high molecular weight subunit genes from Triticum timopheevi and Aegilops cylindrica. Theoret. Appl. Genet. 104, 828–839. Weegels, P. L., Hamer, R. J. and Schofield J. D. 1995. RP–HPLC and capillary electrophoresis of subunits from glutenin isolated by SDS and Osborne fractionation. J. Cereal Sci. 22, 211–224. Weegels, P. L., Hamer, R. J. and Schofield, J. D. 1996. Functional properties of wheat gluten. J. Cereal Sci. 23, 1–18 Weegels, P. L., van de Pijpekam, A. M., Graveland, A., Hamer, R. J. and Schofield J. D. 1997a. Depolymerisation and re-polymerisation of wheat glutenin during dough processing. I. Relationship between glutenin macropolymer content and quality parameters. J. Cereal Sci. 23, 103–111. Weegels, P. L., Hamer, R. J. and Schofield J. D. 1997b. Depolymerisation and re-polymerisation of wheat glutenin during dough processing. II. Changes in composition. J. Cereal Sci. 25, 155–163. Wellner, N., Belton, P. S. and Tatham A. S. 1996. Fourier transform IR spectroscopic study of hydration induced structure changes in the solid state of omega gliadins. Biochem. J. 319, 741–747. Werner, W. E., Adalsteins, A. E. and Kasarda, D. D. 1992. Composition of high-molecularweight glutenin subunit dimers formed by partial reduction of residue glutenin. Cereal Chem. 69, 535–541.

INDEX

A Acetylintermedine, 72, 75 Acetyllycopsamine, 72, 75 Acoustic impedance, 117–118 Acoustic signal generator, 122 Actin–DNA cross-links, pyrrolizidine alkaloid antimitotic activity, 90–91, 92 Adenostyles alliariae (Alpendost), 76 Aegilops cylindrica, 228 Aflatoxins, 198 Agrobacterium tumefaciens transformation vector, 254 Alfalfa sprouts, ozone treatment, 196 Alginate gels, ultrasonic analysis, 152 Alicyclobacillus acidocaldarius, 197 Alkanes, ultrasonic phase transition monitoring, 141 Alpendost (Adenostyles alliariae), 76 α-gliadins, 222 α-trinositol (inositol(1,2,6)trisphosphate), 28, 40 Aluminium–inositol phosphate interactions, 32, 36 Animal products, inositol phosphate content, 26–27 Anticaking agent, gaseous ozone treatment, 179–180, 198 Anticancer activity, inositol phosphates, 37, 38–39 Antioxidants inositol phosphates, 28, 29 anticancer activity, 39

ozone interactions, 177 AP-2 (adaptor protein 2), 25, 30, 38 AP-3 (adaptor protein 3), 30 Apple juice, fruit washing water-related disease outbreaks, 169, 195 Apples, ozone treatment, 192, 195 of washing water, 190 Arabidopsis Escherichia coli phytase transgenic expression, 34 phytate (inositol hexakisphosphate) synthetic pathway, 8 Argon, ozone combined sanitization procedures, 201 Aspergillus fumigatus, phytase transgenic expression, 34 Aspergillus niger, 11, 12, 36 phytase, 20 Asteraceae (Compositae), pyrrolizidine alkaloids, 65 Avocado, inositol phosphate content, 23 B Bacillus cereus, 186 Bacillus spp., 186, 198 Bacillus stearothermophilus, 186, 187 Bacillus subtilis, 187, 199, 202 phytase, 12 Barley inositol phosphate degradation during processing, 21 malting, 22 phytases, 9

304

INDEX

reduced phytate gene mutations, 34 Beet root, inositol phosphate content, 23 β-gliadins, 222 Betony (Pedicularis), 77 BIOTEST, 138 Black beans, inositol phosphate degradation during processing, 21, 22 Blazing star (Liatris punctata), 77 Boraginaceae, pyrrolizidine alkaloids, 65 Botrytis cinerea, 181, 191 Bread, ultrasonic texture sensing, 146, 147 Breadmaking inositol phosphates degradation, 16, 20 see also Glutenin high molecular weight subunits Broccoli, ozone treatment, 192 Brussel sprouts, chlorination, 188 C Cabbage chlorination, 188 inositol phosphate content, 23 Cachana (Liatris punctata), 77 Cadmium–inositol phosphate interactions, 32 Calcium, inositol phosphate effects, 27, 29, 30–31 absorption impairment, 32, 35 inositol(1,4,5)PU3u signaling, 27–28 Campylobacter spp., 193 Canning filling operations, 130 phytate degradation, 16 Capillary chromatography, inositol phosphates, 6 Carbon dioxide, ozone combined sanitization procedures, 201–202 Cardiopulmonary toxicity, pyrrolizidine alkaloids, 80–81 Carrots, inositol phosphate content, 23 Casein, ultrasonic micelle particle sizing, 150 Castellija rhexifolia, 77 Castellija sp. (Indian paint brush), 77 Cauliflower, ozone treatment, 192 Celery, inositol phosphate content, 23 Cereal grains inositol phosphates, 5–6 degradation during processing, 16, 21–22 ozone treatment, 198

pyrrolizidine alkaloids contamination, 68, 69–70 Cheese ultrasonic texture sensing, 146, 147 ultrasonic velocity changes during melting, 142 Chickpeas, inositol phosphate degradation during malting, 22 Chinese millet, phytate content, 13 Chlorine dioxide sanitization, 188, 189 Chlorine sanitization, 169–170, 188–189 ozone combined procedures, 200–201 Chromium–inositol phosphate interactions, 36 Cider fruit washing water-related disease outbreaks, 169, 195 pasteurization recommendations, 195 Coatomer, 30 Cobalt–inositol phosphate interactions, 36 Coltsfoot (Tussilago farfara), 76, 91 Comfrey (Symphytum), 74–76, 91 Commelina communis, phytate (inositol hexakisphosphate) synthetic pathway, 8 Compositae (Asteraceae), pyrrolizidine alkaloids, 65 Compressibility measurements, polymer hydration analysis, 151–152 Cookies, ultrasonic texture sensing, 146 Copper hepatotoxic synergism with pyrrolizidine alkaloids, 79 phytate (inositol hexakisphosphate) interactions, 32, 35–36, 40 Corn bran/flour phytate content, 16 inositol phosphate degradation during processing, 21 Coughwort (Tussilago farfara), 76 Crotalaria, 80 contamination of staple foods, 69, 70 prevention, 91 Crotalaria fulva, 74, 91 Cryptosporidium parvum, 169, 189 Cucumbers, ozone/carbon dioxide mixed gas treatment, 202 Cyclospora cayetanensis, 169 CYP 2C11, 82 CYP 3A, 82

INDEX CYP 3A4, 82 Cytochromes P450, pyrrolizidine alkaloids metabolism, 81, 82 D Dairy products, ultrasonic composition sensors, 138 Dictyostelium phytase, 12 phytate (inositol hexakisphosphate) synthetic pathway, 7 Dietary Supplement Health and Education Act (1994), 74, 91 Digital storage oscilloscope, 122–123 DNA cross-links formation during pyrrolizidine alkaloid metabolism, 83, 88–90 DNA–protein cross-links, 90, 93 DNA repair, inositol phosphate requirement, 31 Doppler ultrasound flowmeters, 131–132 viscosity determination, 148, 149 E Echimidine, 71, 72, 75 Echium plantagineum (Patternson’s Curse; Salvation Jane), 72 Echiumine, 72 Eggs ozone inactivation of microbial contaminants, 186, 193–194 pyrrolizidine alkaloid concentrations, 73, 91 Elastography, carcass grading, 154 Elephant Head (Pedicularis), 77 Emulsions, ultrasonic characterization, 112, 114–116, 158 floculated emulsions, 116, 150 imaging, 154 particle sizing, 115–116, 150 practical applications, 150 Epi-inositol, 37 Escherichia coli, 178, 180, 183, 193 phytase, 9, 10, 11, 21 transgenic expression in Arabidopsis, 34 Escherichia coli O157:H7, 169, 178, 181, 188, 190, 195, 203 Europine, 73

305

F Fabaceae (Leguminosae), pyrrolizidine alkaloids, 62 Fermentation, phytate degradation, 9, 16, 20, 22 Fish ozone treatment, 197-198 ultrasonic composition sensors, 138 ultrasonic microscopy for parasite detection, 156 Flavin-containing monooxygenases, 82 Flavobacterium aquatile, 180 Flavobacterium branchiophilum, 181, 190 Fluid flow, ultrasonic sensing, 130–132 Doppler flowmeters, 131–132 transit-time mode, 130–131 vortex shedding, 132 Foams, ultrasonic analysis, 150–151 Foil pouches, ultrasonic leak detection, 154 Food processing equipment, ozone treatment, 192 Foreign bodies, ultrasonic detection, 151 Fourier transform infrared (FT-IR) spectroscopy, gluten proteins, 239, 278 Fruit chlorine treatment, 188 inositol phosphates, 22-23 ozone treatment, 190, 191, 192, 194–195 pesticide residues reduction, 199 texture (rheological properties), ultrasonic sensing, 142 washing water-related disease outbreaks, 169, 194 Fruit juices Love wave-based component analysis, 153 ozone treatment of ingredients, 196–197 Fulvine, 71, 74 G Gabo near isogenic line, 252 dough properties, 260, 261 transgenic HMW glutenin subunit studies, 263–264 Galahad near isogenic line, 252 γ-gliadins, 222 Garlic, ozone treatment, 198 Gelation, ultrasonic analysis, 152

306

INDEX

Gliadins, 221 dough viscoelasticity gluten fractionation/reconstitution experiments, 242, 243, 245 protein fraction incorporation studies, 245–246 electrophoretic separation, 222, 223 structural studies, 278 Glu-A1, 224, 226, 256–257, 258, 260, 263 near isogenic lines, 251, 252 Glu-A3, 258 Glu-B1, 224, 226, 257, 258, 262, 263 near isogenic lines, 251, 252 Glu-B3, 258 Glu-D1, 224, 226, 256, 257, 258, 260, 262 near isogenic lines, 251, 252 Glu-D3, 258 Glutathione metabolism, 79 Glutathione S-transferase, 83 Gluten insolubility, thermodynamic aspects, 274–275, 276 macropolymer, 258–259 dough performance relationship, 259 viscoelasticity, 279–280 Glutenin high molecular weight subunits, 222, 223 amino acid sequences, 228-233 atomic force microscopy, 233, 235 composition manipulation, 249, 251–256 cross-link formation disulphide bonds, 240-242, 278, 284 low molecular weight subunit branches, 240, 242, 278 mechanical behaviour of dough, 280 transgenic lines, 265 tyrosine bonds, 242 dough viscoelasticity influence, 242–249 dough mixing process, 284–288 gluten fractionation/reconstitution experiments, 242–245 in vitro polymerization, 246 ‘loop and train’ hypothesis, 281–283, 284, 285, 286, 288–289 macroploymer interactions, 280–281 protein fraction incorporation studies, 245–247 purified subunits incorporation into dough, 247–249 subunit peptides incorporation into dough, 249, 250

theoretical models, 279–284 genetics, 224–226 grain breadmaking quality relationship, 226–228, 259–265, 266 biophysical models, 273–289 gluten macropolymer composition, 259 molecular basis, 265, 267–273 quality scoring system, 227 ranking of subunit loci, 258 relation to dough rheology, 279 variation between cultivars, 256–259 hydration-related molecular rearrangements, 278–279, 284 insolubility, thermodynamic aspects, 274–278 mixing/rheological properties influence, 256-265 molecular basis of processing properties amino acid sequences, 267–269, 271 subunit interactions, 271–273 subunit stability, 270, 271 total amount of protein, 267 molecular dimensions, 233, 234 near isogenic lines, 251–253 dough quality differences, 259–262 Gabo, 252, 260, 261 Galahad, 252 Pegaso, 252–253 Sicco, 251–252, 259–261 single subunit, 253 polymorphisms, 224–226 SDS-polyacrylamide gel electrophoresis, 224, 225 structure repetitive domains, 237–239 terminal domains, 234, 236–237 transgenic lines, 253–256 dough quality differences, 262–265, 266 Glutenin low molecular weight subunits, 222, 223 B-type, 222 C-type, 222 cross-links, 240, 242, 278 mechanical behaviour of dough, 280, 281 D-type, 222 dough viscoelasticity, incorporation studies, 247 genetic variation, influence on technological properties, 257, 258

INDEX gluten macropolymer composition, 259 in vitro polymerization, 246 Glutenins, 221 dough viscoelasticity gluten fractionation/reconstitution experiments, 242–245 protein fraction incorporation studies, 245–246 electrophoretic subunit separation, 222, 223 subunits see Glutenin high molecular weight subunits; Glutenin low molecular weight subunits Grey feather (Liatris punctata), 77 Guided wave sensors, 128 H Heleurine, 73 Heliosupine, 66, 67 Heliotrine, 70, 71, 73, 78 Heliotrine N-oxide, 78 Heliotropium contamination of staple foods, 68, 70 prevention, 91 Heliotropium europaeum, 73 Heliotropium popvoii, 70 Hepatic veno-occlusive disease, pyrrolizidine alkaloid toxicity, 69–70, 72, 75–76, 77, 79–80, 91 herbal teas, 74 synergism with copper, 79 Hepatitis A, 178 Herbal medicines, pyrrolizidine alkaloids adverse health effects awareness, 91, 93 sources, 66, 68, 73-78 High-performance liquid chromatography (HPLC), inositol phosphates analysis, 2, 5–6, 13, 41 Honey pyrrolizidine alkaloids contamination, 72, 91 ultrasonic viscosity measurement, 147 Hordeum chilense, 256 Hydrogen peroxide sanitization, 189 ozone combined procedures, 200–201 Hypochlorite sanitization, 169–170 I Indian childhood cirrhosis, 79 Indian paint brush (Castellija), 77 Indian Warrior (Pedicularis), 77

307

Indicine N-oxide, 66, 67, 88 Inositol hexakisphosphate (InsPU6u; phytate), 1–2 analytical methods, 4–5 anticancer activity, 37, 38–39 antioxidant activity, 23, 39 food preservative effects, 39–40 blood glucose-lowering effects, 40 cholesterol/triglyceride-lowering effects, 40 degradation, 8–12 digestive processes, 9, 24 food processing/preparation, 9, 16 thermal, 12 extraction during seed processing, 13 functions in animal tissues, 30–31 calcium flux regulation, 30–31 DNA repair, 31 nuclear mRNA export, 31 high-performance liquid chromatography (HPLC), 5, 6 mineral absorption impairment, 2, 32–33 iron, 32, 33–34, 40–41 optimal dietary levels, 42 seed content processed foods, 16–22 reduction strategies, 34–35 whole raw seeds, 12–15 synthetic pathway, 7–8 Inositol kinases, 7, 8, 31-32 Inositol phosphate phosphatases, 11, 12 Inositol phosphates, 1-43 analytical methods, 4-6 animals, 23–32 absorption, 23–25 biological functions, 27–32 tissue content, 26–27 anticancer activity, 37, 38–39 antioxidant activity, 23, 28, 29, 30, 39 calcium mobilisation, 27–28, 29, 30–31 chemistry, 2–6 chloride channel regulation, 29 degradation, 8–12 breadmaking processes, 16, 20 digestive processes, 9, 24 malting, 22 fruit/vegetable content, 22–23 hemoglobin oxygen affinity regulation, 30 iron binding, 28–29, 30 metabolism, 7–12

308

INDEX

nomenclature, 2–3 nutritional importance, 32–41, 42 health disorders prevention, 36–41 mineral bioavailability influence, 32–36, 42 receptors, 27, 28 seed contents, 12–23 processed foods, 16–22 whole raw seeds, 12–15 signaling activities, 7, 27, 28, 29 synthesis, 7–8 animal cells, 7–8 microorganisms, 7 plants, 8 Inositol pyrophosphates, 31 Inositol(1,2,6)trisphosphate (alphatrinositol), 28, 40 Inositol(1,4,5)trisphosphate (Ins(1,4,5)PU3u), 5, 7 animal cells, 26 receptors (calcium channels), 27–28 signaling function, 27–28 plant cells, 23 Intermedine, 71, 72, 75, 77 Iron, inositol phosphates binding, 28-29 absorption impairment, 32, 33–34, 40–41 low-phytic acid grains, 34–35 anticancer activity, 38–39 overall dietary absorption impact, 40–41, 42 J Jack mackerel, ozone treatment, 197 Jacobine, 72 Jacoline, 70, 72 Jaconine, 72 Jacozine, 72 Japanese flounder, ozone treatment, 197 K Kimchi preparation, 192 L Lactobacillus leichmannii, 203 Lactobacillus plantarum, 22 Lasiocarpine, 71, 73 Lead–phytate (inositol hexakisphosphate) interactions, 32 Legumes, phytate degradation during processing, 16

Leguminosae (Fabaceae), pyrrolizidine alkaloids, 62 Lentils, inositol phosphates degradation, 20, 22 Lettuce chlorination, 188 ozone treatment, 192, 195–196, 206 Leuconostoc mesenteroides, 188 Level sensors, ultrasonic, 129–130 Liatris punctata (cachana; grey feather; blazing star), 77 Lily pollen phytase, 11 Listeria monocytogenes, 169, 188, 193, 203 Lousewort (Pedicularis), 77 Love wave sensors, 128-129, 147 arrays, 129 smart tongue and nose, 153 Love waves, 104 Low-phytic acid grains, 34–35 Lupin, inositol phosphates degradation, 22 Lycopsamine, 71, 72, 75, 77, 78 M Magnesium absorption, phytate (inositol hexakisphosphate) impairment, 32, 35 Maize phytases, 9 phytate (inositol hexakisphosphate), 9 degradation during processing, 22 reduced content gene mutations, 34 Malting, inositol phosphates degradation, 22 Manganese absorption, inositol phosphate effects, 36 Meat elastographic carcass grading, 154 foodborne disease outbreaks, 169 ozone treatment, 186, 190, 191, 192, 193–194 protein ultrasonic analysis, 152–153 pyrrolizidine alkaloid residues, 73 sanitizing agents, 189 ultrasonic composition sensors, 138 ultrasonic imaging of carcasses, 154 Megalocytosis, pyrrolizidine alkaloidinduced, 88, 89 Mercury–inositol phosphate interactions, 32, 36

INDEX Micrococcus spp., 198 Milk contamination with pyrrolizidine alkaloids, 70, 72, 91 ultrasonic analysis aggregates, 152 fat globule particle sizing, 150 imaging for spoilage measurement, 154 Mineral bioavailability, inositol phosphate influences, 32-36 Mixograph, 285 Molybdenum–inositol phosphate interactions, 36 Monocrotaline, 66, 67, 78, 79, 80, 88 Mucor piriformis, 181 Multiple inositol polyphosphate phosphatase (MIPP), 11, 12 Multiple scattering theory, 113–116 Mung bean, phytate (inositol hexakisphosphate) synthetic pathway, 8 Mussels, ozone treatment, 197 Mycobacterium fortuitum, 173 Myelin proteolipid protein, 30 Myo-inositol, 3, 36 anticancer activity, 37, 38 food sources, 36–37 plasma lipid effects, 40 N Necic acids, 66 Neosartorya fischeri, 197 Neper (Np), 106 Nickel–phytate (inositol hexakisphosphate) interactions, 32, 36 Nuclear magnetic resonance (NMR) spectroscopy high molecular weight glutenin subunits, 272–273, 279, 282 domain structures, 236, 237, 239 inositol phosphates, 6 whole raw seed studies, 13 O Oats inositol phosphate degradation during processing, 21 zinc absorption influence, 33

309

Oils, ultrasonic analysis particle sizing, 150 spoilage correlations, 153 viscosity, 147 Ω-gliadins, 222 structural studies, 278 Oncorhynchus mykiss (rainbow trout) culture, ozone water treatment, 181, 190 Onions inositol phosphates content, 23 ozone treatment, 198 Oranges, ozone/carbon dioxide/argon mixed gas treatment, 201 Ozone, 167–208 analytical methods, 204–205 atmospheric, 170–171 biological oxygen demand (BOD) of water, 170 chemical oxygen demand (COD) of water, 170 chemistry/physics, 170–177 cycloaddition reactions, 175 decomposition, 176–177 catalysis by inorganic compounds, 177 free radical species generation, 176 inhibition, 177 initiation, 177 promotion, 177 electrophilic reactions, 175 microbial inactivation, 170 biofilms, 187, 198–199 effect of oxygen demand of medium, 181, 182 food characteristics influencing efficiacy, 190, 191 food sanitizing applications, 189 kinetics, 187–188 mechanisms, 187–188 pulsed electric field combined treatment, 188, 203–204 relative humidity/microbial cell hydration, 179 residual ozone values, 180-181 spores, 186, 187 targeted microorganisms, 186–187 nucleophilic reactions, 175 oxidation–reduction potential, 171 pesticide residue reduction, 170, 189, 199

310

INDEX

physical properties, 171 processing equipment interactions, 183–185 materials compatibility, 184 metals, 184–185 plastics, 185 rubber, 185 production methods, 204 reactivity, 175–177 inorganic compounds oxidation, 177 molecular ozone, 175 solubility, 172–173 stability, 173–175 structure, 171 toxicity, 206–207 transfer from gas to liquid phase (contacting systems), 182–183 treatment chambers, 182 treatment medium, 178–181 oxygen demand, 170, 181, 182 relative humidity, 178–180 residual ozone, 180–181 temperature, 178 workplace exposure thresholds, 207–208 Ozone sanitization combination treatments, 200–204 carbon dioxide, 201–202 chlorine, 201 heat, 202 hydrogen peroxide, 200–201 pulsed electric field, 188, 203–204 ultraviolet radiation, 202–203 dry foods/ingredients, 198 equipment treatment, 192 filtration of treatment water, 183 fish, 197–198 food processing applications, 186–199 chlorine treatment alternative, 188–189 environmental contaminants reduction, 192 food characteristics influencing ozone-demand, 190–191 food surface structure, 190–191 limitations, 206 safety issues, 207 fruit, 190, 191, 192, 194–195 meat, 186, 189, 190, 191, 192, 193–194 metal contaminants removal from drinking water, 177

packaging materials, 192, 198–199 pesticide residue reduction, 170, 189, 199 poultry carcasses, 183, 190, 192, 193–194, 201 raw products/ingredients, 191–192 regulatory status, 206 shelf-life extension, 193 storage applications (ozone gas), 178 surface decontamination, 178 food contact surfaces, 198–199 vegetables, 190, 191, 194, 195–196 P Packaging materials, ozone treatment, 192, 198–199 Packeria candidissima, 77–78 Paramecium phytase, 11 Parenteral fat emulsions, ultrasonic particle sizing, 150 Parsnips, inositol phosphates content, 23 Patternson's Curse (Echium plantagineum), 72 Pea flour, inositol phosphate degradation during processing, 21 Peanuts aflatoxins, ozone treatment reduction, 198 phytate content, 13 Pedicularis (betony), 77 Pegaso near isogenic line, 252–253 Penicillium expansum, 191 Peniophora lycii phytase, 10, 11 Pepper, ozone treatment, 198 Peracetic acid food sanitization, 189 Pesticide residues, reduction by ozone treatment, 170, 189, 199 Petasitenine, 66, 67 Petasites officinalis, 76 Phaseolus vulgaris L., iron availability of low-phytic acid varieties, 34–35 Phonons, 103 Phosphatases, 24 Phospholipase C, 27 Phytases, 8-9 activation in seeds/grains before cooking, 9 animal digestive tract, 24 cell/tissue activity, 9–10 animal cell lines, 24

INDEX classification, 11 phytate degradation during food processing, 16, 20, 21 purified enzymes activity, 10 Phytate see Inositol hexakisphosphate Phytophthora parasitica, 181 Pipe fouling, ultrasonic detection, 153 Plant pyrrolizidine alkaloid sources, 62–65 Plastic pouches, ultrasonic leak detection, 154 Polymer hydration, ultrasonic analysis, 151–152 Potatoes, inositol phosphate content, 23 Poultry carcasses, ozone treatment, 190, 192, 193–194 hydrogen peroxide combined treatment, 201 prefiltration of treatment water, 183 residual ozone values, 180 Protein hydration, ultrasonic analysis, 151–152 Pseudomonas aeruginosa, 180 Pseudomonas fluorescens, 180, 188, 199 Pseudomonas putida, 180 PTEN, 29 Pulsed electric field/ozone combined sanitization procedures, 188, 203–204 Pulsed mode ultrasonic measurements, 123-124 Punctanecine, 77 Pyrrolizidine alkaloids, 61–93 carcinogenic potential, 62, 66, 78, 81 chemical structure, 65–66, 67, 71 consumer protection measures, 69, 73–74, 91–93 food contamination, 66, 68–73 cereal grains contaminated with weeds, 68, 69–70, 91 eggs, 73, 91 honey, 72, 91 meat, 73 milk, 70, 72, 91 preventive measures, 91 herbal medicine sources, 66, 68, 73–78 adverse health effects awareness, 91, 93 metabolism, 81–83, 84–85, 86–87 dehydrogenation, 82 detoxification, 83 N-oxidation, 82–83 plant sources, 62–65

311

toxicity, 62, 78–81 acute, 78 cardiopulmonary, 80–81 diagnostic problems, 81 DNA cross-links formation, 83, 88–90 DNA–protein cross-links, 90, 92 hepatic veno-occlusive disease, 69–70, 72, 75–76, 77, 79–80, 91 hepatotoxic synergism with copper, 79 human poisoning incidents, 62, 68, 73, 75–77 mechanism, 83, 88–91 megalocytosis, 88, 89 progressive nature, 79–80 species differences in susceptibility, 78–79 structural determinants, 65 transplacental transmission, 76 Q Quinoa floor, inositol phosphates degradation during processing, 22 R Rainbow trout (Oncorhynchus mykiss) culture, ozone water treatment, 181, 190 Raleigh waves, 104–105 Retronecine, 66, 67 Retrorsine, 78, 79, 80, 88 Rice Aspergillus fumigatus phytase transgenic expression, 34 phytate (inositol hexakisphosphate) content, 12, 13, 16 reduced phytate gene mutations, 34 Riddelliine, 80, 88 Rockfish, ozone treatment, 197 Russian comfrey (Symphytum uplandicum), 74–75 Rye phytase, 9, 21 phytate (inositol hexakisphosphate), 9 degradation during processing, 21 S Saccharomyces cerevisiae phytase, 11 Salad dressings, ultrasonic particle sizing, 150 Salmon, ozone treatment, 192

312

INDEX

Salmonella enterica ser. Enteritidis, 193, 194 Salmonella spp., 169, 183, 193, 194 Salmonella typhimurium, 180, 203 Salvation Jane (Echium plantagineum), 72 Sanitization methods ozone applications see Ozone sanitization requirements for food safety, 169–170 Scanning acoustic microscopy, 156 Schizosaccharomyces pombe, phytate (inositol hexakisphosphate) synthetic pathway, 7 Scyllo-inositol, 37 Seeds, inositol phosphates content, 12–15 Selectins, 30 Selenium absorption, phytate (inositol hexakisphosphate) effects, 32, 35 Senecio, 80 contamination of staple foods, 68, 69–70 prevention, 91 Senecio burchelli, 69 Senecio disease, 69 Senecio ilicifolius, 69 Senecio jacobaea (tansy ragwort), 70, 72 Senecio longilobus, 77 Senecio riddellii, 65 Senecio triangularis, 77 Senecionine, 66, 67, 72, 76, 77, 78, 80, 88 metabolism, 82, 84-85 Seneciphylline, 71, 72, 76, 78, 80, 88 Senkirkine, 66, 67 Shelf-life extension, ozone treatment, 193 Shrimp meat, ozone treatment, 197–198 residual ozone values, 180 Sicco near isogenic line, 251–252 dough properties, 259–261 Sockeye salmon, ozone treatment, 197 Solid fat content, ultrasonic sensing isothermal scanning, 139–140 temperature scanning, 141 Sorghum flour, phytate content, 16 Sour dough breads, phytate content, 20 Soybean inositol phosphates calcium absorption influence, 35 degradation during malting, 22 phytate (inositol hexakisphosphate), 12 reduced content gene mutations, 34 synthetic pathway, 8 Spirodela polyrhiza, 8, 23

Staphylococcus aureus, 180 Strawberries, ozone treatment, 192 Sunflower seeds, phytate content, 13 Supinine, 73 Symphytine, 75 Symphytum (comfrey), 74–76, 91 Symphytum uplandicum (Russian comfrey), 74–75 Symviridine, 75 Synaptotagmin, 30 T Tansy ragwort (Senecio jacobaea), 70, 72 Temperature, ultrasonic sensing, 132–133 Texture, ultrasonic sensing, 142–148 viscosity measurements, 147–148, 149 Thermal degradation, inositol phosphates, 16 Thermal sanitization, ozone sequential procedures, 202 Tomato sauce, ultrasonic Doppler velocimetry, 148, 149 Tomatoes ozone treatment, 191 washing water-related salmonellosis outbreaks, 169 Transgenic wheat breadmaking quality differences, 262–265 transformation system, 253–254 Trichodesmine, 67, 78, 80 Triglycerides, ultrasonic phase transition monitoring, 141 Triticum dicoccoides, 253 Triticum monococcum ssp. boeoticum, 226 Triticum monococcum ssp. monococcum, 226 Triticum timopheevi, 228 Triticum tsudchii, 228 Triticum turgidum ssp. dicoccoides, 226 Triticum urartu, 226 Tritordeum transformation, 255-256 transgenic line breadmaking qualities, 263 Turnip, inositol phosphates content, 23 Tussilago farfara (coltsfoot; coughwort), 76, 91 U Ultrasonic composition sensors, 133–139 application to animal carcasses, 138–139

INDEX Ultrasonic flow meters, 130–132 Doppler flowmeters, 131–132 transit-time mode, 130–131 vortex shedding, 132 Ultrasonic imaging, 153–157 carcass grading, 154 high-resolution, 156–157 limitations, 155 noncontact methodology, 155 scanning acoustic microscopy, 156 Ultrasonic measurement systems, 121–129 air–transducer impedance barrier, 124 digitizer, 122-123 guided wave sensors, 128 Love wave sensors, 128–129 noncontact measurements, 124–126 pulsed modes, 123–124 pathlength measurement, 124 reflectance methods, 127–128 resonance methods, 126–127 resonance cell losses (Q-factor), 127 signal generator, 122 ultrasonic transducers, 122 Ultrasonic phase transition monitoring, 139–142 isothermal scanning, 139–140 temperature scanning, 140–142 Ultrasonic sensors, 101–158 advantages/disadvantages, 157–158 applications, 129–157 composition, 133–139 dispersed systems, 148, 150–151 fluid flow, 130–132 imaging, 153–157 level sensors, 129–130 phase transitions, 139–142 polymeric systems, 151–153 rheological properties (texture), 142–148 temperature, 132–133 viscosity measurements, 147–148 nondestructive testing, 104 theoretical aspects, 103–121 Ultrasonic thermometry, 132–133 Ultrasonic transducers, 122 Ultrasonic wave propagation, 103–104 attenuation, 105, 109–111 coefficient, 105–106 overestimation, 120 relaxation, 111

313

diffraction, 120–121 homogeneous materials, 105–111 adiabatic elastic modulus, 106 properties of materials, 106, 107–108 inhomogeneous media, 112–117 emulsions, 112, 114–116 floculated emulsions, 116–117 multiple scattering theory, 113–116 particle sizing, 115–116 scattering interactions, 112–113 interfaces, 117–120 acoustic impedance, 117–118 reflection coefficient, 118, 119 refractive index, 118-119 transmission coefficient, 118, 119 longitudinal (compression) waves, 104, 105, 106, 109 critical angle interactions at boundaries, 119–120 near-field/far-field regions, 121 shear (transverse) waves, 104, 105, 109 surface waves, 104–105 Love, 104 Raleigh, 104–105 velocity, 105, 106, 109 adiabatic compressibility calculation, 151 gases, 106, 109 liquids, 109 solids, 109 wave characterisitcs, 103–105 Ultraviolet radiation, ozone combined sanitization procedures, 202–203 Uplandicine, 72 V Vanadium, inositol phosphate interactions, 36 Vegetables chlorine treatment, 188 in-plant contamination, 194 inositol phosphates, 22-23 ozone treatment, 190, 191, 192, 194, 195–196 pesticide residues reduction, 199 sanitization methods, 170 texture, ultrasonic sensing, 142 Vibrio cholerae, 180 Vibrio parahaemolyticus, 180

314

INDEX

Viscosity, ultrasonic measurement, 147–148, 149 von Kármán vortex street, 132 Vortex shedding, 132 W Wheat flour calcium bioavailability, 35 iron absorption influence, 33–34 ozone treatment, 192 zinc absorption influence, 33 gluten, 220-221 polymorphisms, 221, 222 proteins, 221–224 glutenin, 219–290 inositol phosphates, 12, 13, 16, 20 degradation during digestion, 24

degradation during processing, 21 phytase, 11 storage proteins, 220 transformation system, 253–254 Whey protein, ultrasonic analysis, 152, 153 Wine, ultrasonic alcohol determination, 138 Winery equipment, ozone treatment, 192 Y Yogurt, ultrasonic sol–gel transition monitoring, 147 Z Zinc absorption, inositol phosphates influence, 32, 33, 35, 36, 42 Zygosaccharomyces bailii, 197