Chemistry and Physics of Baking: Materials, Processes, and Products 0851869955

1,207 198 16MB

English Pages 276 [145] Year 1986

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Chemistry and Physics of Baking: Materials, Processes, and Products
 0851869955

Citation preview

1 Special Publication No. 56

Chemistry and Physics of Baking Materials, Processes, and Products

The Proceedings of an International Symposium Organized by the Food Chemistry Group of The Royal Society of Chemistry and the School of Agriculture of the University of Nottingham The School of Agriculture, Sutton Bonington, 10th-12th April 1985 Edited by

J. M. V. Blanshard School of Agriculture, University of Nottingham

P. J. Frazier Dalgety UK Ltd., Cambridge T. Galliard RHM Research Ltd., High Wycombe

The Royal Society of Chemistry _r,; Burlington House, London W1 V 0B~w

Preface

Copyright

@

I 986

The Royal Society of Chem,istry All Rights Reserved No part of this book may be reproduced or transmitted in any form or by any means-graphic, electronic, including photocopying, recording, taping or information storage, and retrieval systems--without written permission from The Royal Society of Chemistry British Library Cataloguing in Publication Data

Chemistry and physics of baking: materials, processes, and products: the proceedings of an international symposium organised by the Food Chemistry Group of the Royal Society of Chemistry and the School of Agriculture of the University of Nottingham, the · School of Agriculture, Sutton Bonington, 10th-12th April 1985.--(Special publication; no. 56) 1. Baked products I. Blanshard, J. M. V. IL Frazier, P. J. 111. Galliard, T. IV. Royal Society of Chemistry, Food Chemistry Group V. University of Nottingham. School of Agriculture VI. Series 664'.752 TX763 ISBN 0-85186-995-5

Typeset by Bath Typesetting Ltd., Bath and printed by J. W. Arrowsmith Ltd., Bristol Made in Great Britain

If the Baking Industry were a new development in the twentieth century, then undoubtedly it would have had a higher scientific profile than has been the case. In fact, the process of baking to our knowledge has been in existence for at least 6000 years and has been exploited variously by both simple and sophisticated societies. Nevertheless it is sensible to remind ourselves of the magnitude of the industry. Figures, worldwide, would be difficult to estimate but in the UK alone the total annual retail value of the bread, cakes, and biscuit market is estimated as £3000 million. It is a giant operation. Technologically one might argue that it has been a slumbering giant, and there is some truth in such a description. However, economic pressures and an increasing awareness of nutritional issues have posed sharp questions to those wishing to feed the nations of today wisely and at the same time remain commercially viable. Such a purpose requires a detailed knowledge of materials and processes which, from biochemical and physicochemical points of view, represent highly complex and indeed variable systems. Although there are several journals devoted to the publication of research and annual reviews of specific topics related to the baking industry, there has been a noticeable absence of a comprehensive, up-to-date examination of this field, particularly in the light of recent information derived from basic physical, chemical, and biochemical studies. In an attempt to remedy this situation The Food Chemistry Group of the Royal Society of Chemistry, jointly with the University of Nottingham Department of Applied Biochemistry and Food Science, organized an international symposium at Sutton Bonington on the theme of the title of this volume. Some two hundred and twenty participants gathered from the international cereal science community for what proved to be a most stimulating meeting. The objectives of the organizers were to bring together cereal scientists and technologists with chemists, biochemists, physicists, and engineering scientists having common interests in the raw materials, processes, and products of baking. In order to achieve the most effective exchange of ideas and experiences, it was recognized that an attempt should be made to assemble experts from around the world, since no one country, or even continent, could claim a monopoly of knowledge. We were indeed fortunate that all of the invited speakers willingly agreed to contribute the reviews presented in this volume. Thus, the proceedings cover a wide range of relevant subjects. The articles follow a logical sequence, early chapters dealing with the major components of baked products: polysaccharides, proteins, fats and emulsifiers, enzymes, yeast, and water-a most important functional ingredient, unfortunately neglected by many food scientists. A series of reviews then covers both physical and chemical interactions that are important in the baking process. With current trends favouring a reduction in the use of 'additives' in food, it is increasingly important to gain better understanding and control of physical and chemical processes such as mixing, rheological changes, component interactions, and oxidation-reduction reactions in the various stages of the baking process. The V

VI

Preface

recent and rapid growth in importance of 'high fibre' products and of extrusion cooking has relied largely on empirical practice; articles on both these areas review ways in which the underlying scientific aspects differ from those in conventional baked products. We trust that these published proceedings will not only provide a more permanent record for those who attended the symposium, but will also be of interest and benefit to a wider readership. We are extremely grateful to all those who contributed to the conference in so many ways, and to the following organisations who provided generous financial support: Associated British Foods Ltd., Baker Perkins Ltd., Dalgety (UK) Ltd., The Distillers Company Plc., Marks and Spencer Pie., Northern Foods Plc., Pedigree Petfoods, Rank Hovis McDougall Research Ltd., Rowntree Mackintosh Plc., Staufi;er Chemical Ltd., Unilever Research, United Biscuits (UK) Ltd.

Contents Basic Constituents of Baked Products

The Significance of the Structure and Function of the Starch Granule in Baked Products

By J. M. V. Blanshard 2

Flour Proteins: Structure and Functionality in Baked Products

3 The Significance of Water in the Baking Process J.M. V. Blanshard P. J. Frazier T. Galliard

14

By J. D. Schofield

30

By S. Ablett, G. E. Attenburrow, and P. J. Lillford 42

4 Non-starch Polysaccharides

By F. Meuser and P. Suckow 5 Functionality of Wheat Lipids in Relation to Gluten Gel Formation

62

By K. Larsson 6 The Role of Fats and Emulsifiers in Baked Products

75

By S. Tamstorf, T. Jonnson, and N. Krog 7 Protein Functionality in Bakery Products

89

By A. P. Davies 105

8 Enzymes in Baking

By Yu-Yen Linko and P. Linko 9 The Biotechnology of Baker's Yeast: Old or New Business?

117

By H. W. van Dam Fundamental Interactions: Consequences, Control

l O Physicochemical Processes in Mixing

132

By F. MacRitchie 11

Protein-Lipid and Protein-Carbohydrate Interactions in Flour-Water Mixtures

147

By W. Bushuk 12

155

Redox Systems in Dough

By W. Grosch vii

Contents

viii 13

Rheological Aspects of Structural Changes during Baking

170

1

179

The Significance of the Structure and Function of the Starch Granule in Baked Products

By A.H. Bloksma

14 Action 'of Oxidants and Other lmprovers By C. S. Fitchett and P. J. Frazier

15

Wholemeal Flour and Baked Products: Chemical Aspects of Functional Properties 0

,By J.M. V. Blanshard 199

DEPARTMENT OF APPLIED BI.OCHEMISTR Y & FOOD SCIENCE, UNIVERSITY OF NOTTINGHAM, SUTTON BONINGTON, LOUGHBOROUGH LE12

By T. Galliard

5RD, UK

16 Component Interaction during Heating and Storage of Baked Products

216 1 Introduction

By R. C. Hoseney Developments in Processes and Products

17

Extrusion Cooking versus Conventional Baking

227

By R. C. E. Guy

18

Optimization of Products and Processes

236

By J. R. Mitchell, H. Back, K. Gregson, S. Harding, and S. Mather

19 The Way Ahead: Wheat Breeding for Quality Improvement

In Western European and North American markets, wheat flour is the major cereal component of baked products. Furthermore, within wheat flour itself, starch is the principal element constituting ca. 70% of the material. It would be surprising, therefore, if the starch adopted a strictly neutral role either in terms of interactions with other components or in the macroscopic rheological properties ofthe,system. The situation within a baked .product in terms of potential interactions is illustrated in Figure I. This indicates that the starch granule may interact in

--- - - .... I

251

By P. R. Day, J. Bingham, P. I. Payne, and R. D. Thompson

proteins

'

I

hyc:l(~philic

Baking-The Way Ahead

/

262

,

By B. Spencer

Subject Index

\

/

li~Jsf / /

20

.,~,

endo- / / genous, /

1

1 1

/

/

exo-

\ \

\ .

genous /

,, /-....._______ .,,,-/ /

I

271

I /

\

sfar,:h

\

I

I

gr~nules /

/

/ water

sugar

Figure 1 Potential interactions between the five major components of baked products: starch, water, hydrophilic macromolecules, lipid, and sugar

I

Chemistry and Physics of Baking

2

physicochemical terms with one or more of the other four principal components: water, sugar, lipid, and hydrophilic macromolecules such as proteins and pentosans. The actual extent of these interactions will vary from system to system and, indeed, the aim of the innovative technologist may be described in terms of either containing or exploiting one or more specific interactions. That the starch is affected by the baking process, and variably so in different products, can be readily demonstrated by microscopy; Greenwood, for example, has described 1 the stages of progressive disordering 0f the granule in a number of systems varying from a maximal effect in wafers to a minimal change in Scottish shortbread (Figure 2).

swollen

Structure and Function of the Starch Granule

3

Table 1 Typical percentage composition of starch granules of different biological origin (based on dry starch content after Gui/bot and Mercier, 2 Blanshard, 3 and Morrison et al. 4 )

Amylase Lipids Protein Phosphorus Ash

Wheat

Maize (normal)*

Potato

26--31 0.48-1.12 0.2 -0.33 0.05 0.14---0.3

24--32 0.60---0.80 0.27---0.39 0.o!5 0.05---0.1

23 0.09 0.05 0.04 0.3

Hence amylopectin content = ca. 70-75%. * i.e. not 'waxy' or 'high amylase' cultivars.

gelatinized disrupted

Bread

dispersed

Iwaters

enzymieally degraded Figure 2 Stages ofgranular dispersion in baked goods ( Greenwood, ref 1)

If, therefore, we recognize that starch is both modified by the baking process and yet exercises a definite and sometimes a distinctive role in the final product, we may address ourselves to three questions:

(i) What is the composition and what are the structual elements in the starch granule in relation to the baking process and product? (ii) How do the other components in a dough/batter affect gelatinization of the starch granule? (iii) What are the important structures, where starch is involved in the final baked product? We shall examine these one by one.

highly branched, racemose structure.) We are therefore obliged to assume that there are more subtle differences. For example, though the composition of amylase and amylopectin may be the same, it is possible that the molecules have significantly different structures and properties or, more likely, that the architecture of the granule is different. In other words, similar proportions of the critical components, and of similar structure, are compounded into a macrostructure which is very different functionally. A variety of physical techniques have been used to explore the architecture and have provided valuable information. For example, wide-angle X-ray diffraction (WAXS) was first reported by Katz 5 •6 to distinguish two principal types of a polymer crystallinity within starches, so-called A-starches (found largely in cereals) and B-starches (largely of tuberous origin); a third type, C-starch, is widely believed to be a mixture of A- and B-types and is found especially in legume starches. Contrary to what might be expected, it is the branched amylopectin which is responsible for the crystallinity (Meyer,7 Montgomery and Senti 8 ) whereas the amylase is believed to be largely in the amorphous state. Molecular models have recently been proposed 9 • 10 for A- and B-starches, though in this instance amylase fibres provided the source of the necessary X-ray diffraction data and there is some doubt as to whether the suggested antiparallel packing of double helices is either biosynthetically feasible or indeed fits sufficiently well with the

2 The Composition and Distinctive Structure of the Wheat Starch Granule 2

An examination of the composition of the three most common starches in the Western world, namely wheat, maize, and potato, as shown in Table 1, suggests that there is little significant difference in terms of the amylase, amylopectin, lipid, protein, phosphorus, or ash content, though there is no doubt that these starches in certain situations may behave very differently. (The amylase is, of course, an essentially linear, polymeric molecule whereas the evidence of biochemical, physicochemical, and microscopical studies of amylopectin points to it being a 1

C. T. Greenwood, 'Starch' in 'Advances in Cereal Science and Technology', ed. Y. Pomeranz, American Association of Cereal Chemists Inc., St. Paul, MN, 1976.

3

4 5

6 7

8 9

10

A. Guilbot and C. Mercier, 'Starch', in 'The Polysaccharides', ed. G. 0. Aspinall, Vol. 3, Academic Press, 1985. J. M. V. Blanshard, 'Starch Granule Structure and Function: A Physicochemical Approach', in 'Starch: Properties and Potential', ed. T. Galliard, Critical Reports in Applied Chemistry, Society of Chemical Industry, London, 1986 (in press). W.R. Morrison, T. P. Milligan, and M. N. Azudin, J. Cereal Sci., 1984, 2,257. J. R. Katz, in 'A Comprehensive Survey of Starch Chemistry', ed. R. P. Walton, Vol. 1, p. 68, The Chemical Catalog Co., New York, 1928. J. R. Katz and T. B. van Hallie, Z. Phys. Chem., 1930, AlSO, 90. K. H. Meyer, Adv. Colloid Sci., 1942, l, 183. E. M. Montgomery and F. R. Senti, J. Po/ym. Sci., 1958, 18, I. H. C. Wu and A. Sarko, Carhohydr. Res., 1978, 61, 7. H. C. Wu and A. Sarko, Carhohydr. Res., 1978, 61, 27.

4

Chemistry and Physics of Baking

observed scattering patterns of starch granules whether they be A- or B-types (French, 11 Wild and Blanshard 12 ). WAXS measurements have also been used 13 to show that the percentage absolute crystallinity within the starch granule is about 25-39%. Various other techniques provide additional insights. For example, the well known optical birefringence of starch granules when viewed under polarized light (French 14), particularly when the evidence arising from the use of a A/4 wave plate, 15 strongly suggests that the polymer chains are predominantly radially arranged. This situation obtains (as in the case for the X-ray crystallinity) not only when amylopectin is the major component as in wheat, maize, and potato starches, but also when it is the ·~xclusive element as in waxy maize starch. 16 Electron microscopical information points to similar conclusions while the characteristic four-leaved, clover type scattering observed with small-angle light scattering (SALS) in the Hv mode supports the same viewpoint but, in addition, suggests that the granules may be considered as polymer spherulites. 17 The use of both small-angle X-ray scattering (SAXS) 18 and small-angle neutron scattering (SANS) 19 has given evidence of a regular periodicity of 9-10 nm, and the present interpretation of this is that the 7 nm periodicity proposed from the results of electron microscopy is somewhat low owing to preparative artefacts (particularly dehydration) and that the 9.0---J0.0 nm figure is much closer to the situation prevailing in the native hydrated granule. All starch granules examined, except those of the wrinkled pea, exhibit this periodicity. 19 An extremely useful extension of the SANS technique is that of contrast matching, which facilitates, under appropriate conditions, the masking out of one component so as to visualize a second component more clearly. This method has been used to obliterate the major polysaccharide element of the starch granule which is 'invisible' in a 52: 48 (v/v) D 2 O-H 2 O mixture, thereby permitting one to see residual lipid. By comparing native and defatted wheat starches, a peak has been observed at 15-16 nm. Such information is particularly interesting when put alongside deductions from line-broadening measurements of the 100 peak observed by WAXS (this peak is a single reflection at 5.6° 20 in potato starch) that the tangential crystallite size is ca. 14-15 nm.11· 20 We may therefore hypothesize that the lipid is arranged radially, perhaps acting as a hydrophobic sheathing to the amylopectin crystallites. A further point that needs to be considered is the location of the amylose molecules. In addition to the evidence of electron microscopists, 3 further information can be derived from infrared dichroism studies of starch granules. The basis of 11

12

13

14 15

16

17 18 19

20

D. French. 'Organization of Starch Granules' in 'Starch: Chemistry and Technology', ed. R. L. Whistler, J. N. BeMiller, and E. F. Paschall. 2nd Edn., Academic Press Inc., 1984. D. L. Wild and J.M. V. Blanshard, Carbohydr. Polyrn., 1986 (in press). H. F. Zobel and F. R. Senti, Prqgram of 45th Annual Meeting of American Association of Cereal Chemists, 1960, p. 40. D. French, Denpun Kagaku, i972. 19, 8. W. Banks and C. T. Greenwood, 'Starch and its Components', Edinburgh University Press, Edinburgh, 1975. M. Z. Bhuiyan and J.M. V. Blanshard, Stiirke, 1982, 34,262. J. Barch, A. Sarka, and R.H. Marchessault, Stiirke, 1969, 21,279. C. Sterling, J. Polym. Sci., 1962, 56, SI0. J.M. V. Blanshard, D.R. Bates, A.H. Muhr, D. L. Worcester, and J. S. Higgins, Carbohydr. Po/ym., 1984, 4, 427. S. Hizukuri and Z. Nikuni, Nature, 1957, 180,436.

Structure and Function of the Starch Granule

5

this approach is that, when starch granules are stained with iodine, amylose molecules present in the amorphous state may be expected to form the amylose--iodine complex. If the molecules are arranged radially such amylose-iodine complexes will display a dichroism in the infrared. This dichroism is observed in wheat starch but, surprisingly, the dichroism is weak or almost absent in the strongly birefringent potato starch. A plausible suggesfion is that the amylose molecules in the potato starch granules are not entirely in the amorphous state but partly co-crystanized with amylopectin chains in the polymer crystallite and, therefore, largely unavailable for complex formation with iodine. Some support has been given to this view from recent studies of the comparative efficiency of etherifying reactions on native and heat/moisture-treated starch granules; 21 it was concluded that heat/moisture treatments led to a significant reduction in absolute crystallinity of the granules but-more importantly to the present argument-an increased accessibility of amylose to etherification. Such results can be interpreted as the partial disruption of the crystallites with an accompanying release of the amylose molecules. Much of the information that has been referred to can conveniently be summarized in the hypothetical model· for starch granule crystallites shown in Figure 3.

------15nm

l

6nm

!+

4nm

i

KEY:

amylopect in

helix

hybrid amylose/amylopectin helix V-amylose helix free

lipid

free amylose Figure 3 Hypothetical model of crystallites in the starch granule showing typical distribution of components and dimensions ( Blanshard, ref 3) 21

P.A. M. Steeneken, Stiirke, 1984, 36, 13.

6

Chemistry and Physics of Baking

A number of interesting questions immediately come to mind. For example: (i) How significant is the X-ray crystalline form to the functionality of the starch granule? One obvious method of examining this issue is to replace wheat starch (A-type) by potato starch (B-type) in baked products and determine whether this can be done without a serious deterioration in quality. Such an experiment has been performed by Hoseney et al., and the results 22 show that potato starch is not satisfactory. However, we might conclude that any difference could be due to inherent structural differences in the granule quite distinct from the order within the crystallites. It is interesting therefore that Lorenz and Kulp have observed 23 that pot,ato starch subjected to a heat/moisture treatment (27% moisture for 16 h at 100 °C) and substituted for wheat starch gave a significant improvement in cake volume and total cake score in comparison with untreated potato starth. A somewhat similar effect has been noted for other tuberous starches exposed to comparable heat/moisture treatments. 24 It does seem, then, that the heat/moisture treatment introduces a structural change which ameliorates the functionality of potato starch in a cake system. (ii) Are there structural/compositional differences evident from a wider study of starches that can be related to functional differences? Hoseney et al. attempted 22 to substitute wheat starches in bread by alternative starches and found, generally, that rye and barley starches were satisfactory, whereas the starches of corn, sorghum, rice, and· potato were unsatisfactory. Ghiasi et al. showed 25 that the use of waxy barley starch as a substitute gave loaves which showed excessive shrinkage. Other investigations by workers at the Flour Milling and Baking Research Association established 1 that in a high-ratio yellow cake system, the use of maize starch as a substitute for wheat starch was satisfactory, but that waxy maize resulted in a collapse of the system post-baking; amylomaize, riot unexpectedly, also produced a product which collapsed since no gelatinization had occurred. On surveying this information there are definite pointers to the importance of amylose as a functional ingredient in baking if, on gelatinization, it is in such a form that it can be released into the intergranular space. 3 How do the Other Components in a Dough/Batter Affect Gelatinization of the Starch Granule?

At this stage it is impossible to give a comprehensive answer to this question but certain lines of work examining three-component systems of starch, sugar, and water have given encouraging leads. To gain a satisfactory understanding it is necessary to have a correct perception of the gelatinization process of a semi-crystalline material like the starch granule as well as to understand the physical chemistry and the potential and limitations of techniques that have been employed.

Structure and Function of the Starch Granule

An interpretation and viewpoint on starch gelatinization that has gained wide credence in the last decade is to treat it as a melting phase transition, promoted by the presence of solvent which, of course, is usually water. Such a system may be handled by equations devised tor synthetic polymer-solvent systems and based on equilibrium thermodynamics. The approach has proved valuable in systematizing the relation between gelatinization ten;iperature and solvent content. More recently, other workers have sounded 26 a valuable cautionary note by emphasizing the essentially non-equilibrium character of a semi-crystalline material like the starch granule and have regarded the solvent as essentially a plasticizing agent. The subject has been fully reviewed elsewhere. 3 Two approaches that have been made recently to study gelatinization have cast light upon the gelatinization process. Firstly, calorimetric techniques [with the twin methods of differential scanning calorimetry (DSC) and differential thermal analysis (DTA)] have been used by many groups of workers. The two methods record the temperatures and enthalpies of gelatinization. The usual technique involves using a uniform heating rate of 5 or 10 °C min - 1 . In this respect the work of Shiotsubo and Takahashi 27 is of particular interest as they know that gelatinization occurs only under non-kinetically controlled conditions at heating rates less than 0.5 °C min - 1 , substantially less than those used by most workers. In contrast, the second approach using the SALS technique, which was developed in these laboratories, provides both dynamic and equilibrium information. The gelatinization is effected by small (2 °C) temperature jumps, but a second temperature perturbation is not initiated until loss of birefringence from the first jump has ceased and an equilibrium situation has been attained. The technique can therefore be used to compare the kinetics of gelatinization of different starches and also to produce an 'equilibrium' (or at the very least a quasi-equilibrium) plot of birefringence versus temperature for a given system. 28 With these concepts in mind we can briefly outline a possible thermodynamic treatment of the system. In a thermodynamically ideal solution, the interaction energies between like and unlike molecules of suitable size are the same, which is evidenced by the enthalpy of mixing being zero. Few systems behave in this manner and deviations from ideality are handled in a variety of ways. Where small molecules are being considered, the activity coefficient is used as a 'fudge' factor or, alternatively, if one is more mathematically inclined, a virial expansion is employed. These approaches are feasible with large molecules too, but a device widely used in the synthetic polymer field is the polymer-solvent interaction coefficient x. This in many situations is, again, a 'fudge' factor even though it is given some physical form [x = zl1w/kT, where z is the co-ordination number, 11w = w12 - ½(w 11 + w22 ), i.e. the energy difference per polymer solvent contact, k the Boltzmann constant, and T absolute temperature]. This latter approach was developed particularly by P. J. Flory 29 and employed by many succeeding polymer physical chemists in understanding the solution properties of polymers, particularly in concentrated systems. A derivative of this approach was the 26

22 23

24 25

R. C. Hoseney, K. F. Finney, Y. Pomeranz, and M. Shogren, Cereal Chem., 1971, 48, 191. K. Lorenz and K. Kulp, Cereal Chem., 1981, 58, 49. K. Lorenz and K. Kulp, Stiirke, 1982, 34, 5076. K. Ghiasi, R. C. Hoseney, K. Zelezuak, and D. E. Rodgers, Cereal Chem., 1984, 61,281.

7

27 28

29

L. Slade and H. Levine, ACS NERM 14 Presentation: Abstract Water Soluble Polymer Symposium, 12/6/1984. T. Shiotsubo and K. Takahashi, Agric. Bio/.,Chem., 1984, 48, 9. J.M. V. Blanshard, 'Physicochemical Aspects of Starch Gelatinization', in 'Polysaccharides in Food', ed. J.M. V. Blanshard andJ. R. Mitchell, Butterworths, London, 1979. P. J. Flory, 'Principles of Polymer Chemistry', Cornell University Press, Ithaca, New York, 1953.

Chemistry and Physics of Baking

8

expression which relates the variation of the melting point (Tm) of a polymer under equilibrium conditions to the volume fraction of solvent (v 1 ). However, the treatment for such a two-component system 30 can be extended to a three-component system (e.g. starch, sugar, and water) and also performed by Lelievre. For a three-component system (see Figure 4) there would be three interaction coefficients (x 12 , Xu, and X23 ). The problem is to solve these interactions with multiple interaction coefficients. Lelievre 31 simplified the solution by assuming that x12 = x13 and that x23 = O; i.e. the interaction coefficients of water with the sugar and ,starch were equal and there was a negligible interaction between the sugar and the starch. Using a valu,e of x = 0.5 derived from a two-component study, 30 Lelievre found that there was good agreement between the experimental and predicted values of Tm when maltose was used, but results were somewhat less satisfactory for other sugars. However, using a simplex search routine, it is feasible to lpok for possible solutions of the equation and values of x12 , Xu, and

3-Component System:

Lelievre (1976) assumed: (i)

X23

= 0

(ii)

X13

=

X12

and hence the equation assumed the form:

Figure 4 Lattice thermodynamics treatment ( after Flory, ref 29) of a three component system (water, starch, and sugar) illustrating the three interaction coefficients and the resultant equations 30 31

J. Lelievre, J. Appl. Polym. Sci., 1973, 18,293. J. Lelievre, Polymer, 1976, 7, 854.

Structure and Function of the Starch Granule

9

x23 which give optimum fit. A number of sugars have been investigated in this way using both the DSC and SALS techniques to follow the gelatirrization process. 32 The results of the SALS and DSC techniques (see Table 2) are interesting on a number of grounds.

Table 2 Sample values of X12 , Xu, and X23 for water-starch-sugar systems for four different sugars calculated by a simplex fitting routine. The experimental data were determined in ( a) SALS and in (b) DSC ( Pitt et al. 32 ) Experimental 1echnique

Sugar

(a) SALS

fructose sucrose xylose glucose

0.488 0.506 0.516 0.511

0.373 0.210 0.380 0.402

1.017 0.920 0.888 1.069

(b) DSC

glucose

0.497

-0.590

0.541

X12

X13

X23

(i) The X12 values (i.e. the starch-water interaction coefficient) for all sugars by both techniques are very close to 0.5, which accords with the original value determined by Lelievre 30 for the starch-water system and one of his subsequent assumptions for the three-component system. However, the Xu (sugar-water) and x23 (sugar-starch) interactions certainly do not equal 0.5 and zero respectively, and furthermore vary with the sugar that is being examined. It is therefore not surprising that the different sugars exert a considerably different effect upon the texture of the final baked product. (ii) Further interesting features are the results for glucose when investigated by DSC. It is evident that in these circumstances, although x12 is effectively 0.5, x13 assumes a value of -0 ..59 and x23 falls to a value of 0.54. The difference in these values obtained when the DSC was operated at 5 °C min - 1 demonstrates the danger of using the Flory-type approach based on equilibrium thermodynamics under this mode of operation with a fundamentally kinetically controlled process and comparing the results obtained under more truly equilibrium conditions. However, although such an approach may be suspect from a thermodynamic viewpoint it does point the way to a semi-empirical quantifying of the interactions between the different components. The negative x13 value suggests that the interactions between the water and glucose are those characteristic of an excellent solvent. Clearly such an approach offers the possibility of examining changes over a wider range of heating rates and with the addition of other components, e.g. proteins and lipids.

32

S. R. Pitt, S. Wynne-Jones, and J. M. V. Blanshard, in preparation.

Chemistry and Physics of Baking

10

1200

)(

GI "O C:

1000

GI

E

....

800

GI ti)

C: 0 C.

. ti)

GI

600

"O GI

E E ::I

u,

400

200

20

40

60

80

100

g sugar /10Qg water Figure 5 Variation of the. summed response time index ( see ref 3) for water-starch-sugar systems where the sugar is varied in type and concentration (g sugar/100 g water)

(iii) It should also be noted (Figure 5) that the kinetics of the gelatinization process as monitored by the SALS technique vary with the sugar. For example, at a concentration of 100 g sugar/JOO g water there is a broad correlation with molecular weight in that the highest summed residence time index (SRT) value (slow gelatinization) is caused by a pentose (xylose) with the lowest molecular weight and lowest gelatinization temperature, whereas the lowest SRT value is given by the disaccharide sucrose with the highest molecular weight and gelatinization temperature. Values for the two hexoses, glucose and fructose, are intermediate. 4 What Are the Important Starch-based Structures in the Baked Product?

Structure and Function of the Starch Granule

systems, substantial amounts of added lipid may inhibit hydration and formation of the gluten network or be present as a distinct phase. Similarly, sugars may be present in different amounts limiting gluten formation, gelatinization, or even contributing to the formation of a sugar glass. It is these materials, in variable proportions with different products, that contribute to the solid matrix surrounding the air cells and spaces which are characteristic features of the structure and perceived texture of baked, cereal products. Microscopical investigations have given broad support to the above conclusions, in terms of both the formation or the absence of a gluten network, 34 - 37 as well as the significance of starch 38 and of flours other than wheat. 39 Conceptually, we can regard baked products as complex composites. Since no one model is appropriate to the whole range of products, we shall discuss one which is possibly helpful in illuminating the behaviour of those products which have a partially or well developed gluten network and a moisture content between 20-40% (i.e. encompassing the cake/bread range). For such systems we can consider the solid matrix to consist of a filled network, the network being composed of a variably developed gluten but other proteins also. Also present is amylase and some amylopectin, while the principal filler is starch present as native, partially or wholly gelatinized granules. There is clear evidence that, though they have lost their birefringence and swollen, nevertheless, within the product, they do maintain their integrity. 38 With this background we can seek to establish what are the distinctive roles of amylase and amylopectin in the structures that are developed during baking and storage. Some insight can be obtained by the use of waxy starch variants and such systems have been examined. Reference has already been made 1 to the substitution of wheat starch by waxy maize starch in a high-ratio cake system while Hoseney et al. have investigated 40 the functionality of waxy barley starch in a reconstituted bread dough. Both of these systems which lacked amylase led to a collapsed structure. Hoseney et al. believed that this showed that the amylase fraction, perhaps as a result of rapid retrogradation, was responsible for setting the crumb structure. Ghiasi et al. have extended this work and shown 25 that bread baked with a partial reduction of the amylase : amylopectin ratio [through dilution of a high-protein (17.2%) wheat with waxy barley starch to a protein content ofl2.5%] led to a product that was initially softer. These workers, in line with Hoseney's earlier suggestion, proposed that this softer structure was a consequence of the lower concentration of amylase which was not then available for retrogradation. Such a conclusion is derived, no doubt, from the fact that in dilute solution amylase retrogrades rapidly whereas an amylopectin solution is stable. Such a situation, however, is a far cry from that prevailing in a baked product. Laboratory evidence, for example, shows that a 50 : 50 amylase-water gel is stable and does not retrograde

From the most elementary considerations we can see that in a product prepared from whole flour the starch granules, swollen or not, will be surrounded by protein (which may or may not have aggregated into a gluten network) admixed with pentosan molecules. In addition, on gelatinization, there will be the release from the granule of linear amylase and also variable amounts of amylopectin. 33 In some

34

33

40

K. Ghiasi, R. C. Hoseney, and D.R. Lineback, Cereal Chern., 1979, 56,485.

11

35 36

37

38 39

B. Francis and C.H. Groves, J. R. Microsc. Soc., 1962, 81, 53. D. J. Stevens, Starke, 1976, 28, 5. L. Wasserman and H. H. Dorfner, Getreide Mehl Bro/, 1974, 28,324. R. E. Angold, 'Cereals and Bakery Products', in 'Food Microscopy', ed. J. G. Vaughan, Academic Press, London, 1979. D. B. Bechtel, Y. Pomeranz, and A. de Francisco, Cereal Chern., 1978, 55,392. Y. Pomeranz, D. Meyer, and W. Seibel, Cereal Chern., 1984, 61, 53. R. C. Hoseney, D.R. Lineback, and P.A. Seib, Baker's Digest, 1978, 52, (4), 11.

12

Chemistry and Physics of Baking

at ambient temperatures (as measured by X-ray diffraction and DSC) though there is a significant change in the rheological properties after formation of such a gel. Present studies of the effect and significance of flour chlorination have also provided informative insights into the role of amylose. A variety of techniques have been used in the past 41 to determine what, if any, effect chlorination has upon the structure and gelatinization behaviour of starch granules and for the most part their conclusions have not been particularly illuminating. Telloke has, however, recently shown 42 that, whereas when untreated wheat starch granules are gelatinized in 30--60% w/w sucrose solutions they exude relatively little amylose, chlorinated flours do release amylose, and the level of exudation is linearly related to the degree of chlorination; there is, not surprisingly, a corresponding increase in paste viscosity and cake crumb firmness which alleviate the problem of collapse. It is interesting to notice that the absence of chlorination and the presence of high sugar concentrations leads to a situation very similar to the substitution of wheat starch by waxy maize flours with accompanying collapse. The evidence we have accumulated thus fat would suggest that amylose in the intergranular spaces exercises an important structural and textural role in at least sotne baked products. It is further suggested that, as already indicated, the textural changes observed immediately after baking are not retrogradation (i.e. the consequence of crystallization phenomena) but the result of the product cooling and transforming through a continuum of rubbery states to possibly (depending on composition) the glassy state. Such a state, if attained; will remain remarkably stable, physicochemically, in the absence of the intrusion of water which, if this occurs, might reduce the glass transition temperature (Tg) to below ambient temperature and permit further significant textural changes. With such a model in mind, we may wish to speculate further about possible means of manipulating the properties of the system. For instance, if we use the data provided by van den Berg43 and appropriate polymer theory44 we can calculate the variation of the Tg of starch (perhaps more especially amylose) with water content. If we do so, then we can demonstrate that in an amylose gel with a 20% water content Tg is approximately ambient temperature and, hence, amylose would be stable as a polymer glass. In practice, however, short-chain amylodextrins may also be present with multiple chain ends, which tend to lower Tg. Similarly, the scanty evidence available suggests that proteins have lower values of Tg, 45 and such may be the case for pentosans also. Overall, therefore, although the amylose per se in these given conditions might be present as a glass, the combined effects of proteins, pentosans, and other polysaccharides may be to impose a rubbery behaviour upon the intergranular material. It is important, however, to appreciate that the transition from a highly elastic gel to a rubber and, thereafter, to a glass are frequently not sharp transitions (lnd are time-dependent processes. If we now turn our attention to the gelatinized granules containing the amylopectin, the evidence of DSC and X-ray diffraction clearly points to the amylopectin

Structure and Function of the Starch Granule

within the granules being responsible for the increase in crystallinity and for the retrogradation process. It is also of interest to note that sugars, 46 surfactants, e.g. glycerol monostearate (GMS), 47 and cyclohepta-amylose 48 inhibit (and delay) the granule swelling process, though their effect is less pronounced on the gelatinization temperature as monitored by loss of birefringence. Presumably the reduced swelling means that the 'filler' in the filled network is then less closely packed and yields, as observed, a softer structure. We can see, therefore, that staling may involve changes, due to amylopectin within the granules, that involve crystallization and which are largely discrete from and kinetically distinct from events in the intergranular space. Nevertheless the processes that occur in this latter region, though the molecules may be in the amorphous (rubbery or glassy) state, may exert a profound influence on the overall rheological behaviour of the system. Such a view helps to explain the effects of enzymes (e.g. bacterial a-amylase) which may be added during baking as an antistaling agent. It certainly results in a scission of polymers and polysaccharide networks yielding a. softer product. We may anticipate that the enzyme will more rapidly attack the intergranular polysaccharides rather than those in the partially swollen starch granules. Such a view may explain the observations of Zobel and Senti 49 and Dragsdorf and Varriano-Marston, 50 who found that the X-ray crystallinity in such materials is not diminished, i.e. the retrogradation continues in the (partially) swollen granules. However, it should be mentioned that the amylodextrins produced by amylolytic attack will have a lower Tg and therefore retrograde much more readily; such materials would contribute to the observed X-ray crystallinity, but be less important in increasing product firmness. Monoglycerides, proteins, and pentosans have all been examined and their effects recently reviewed. 51 Their role (either through specific interactions, e.g. GMS with amylose, or by moderating the physical properties of the system) is complex. Furthermore, their role may change with time, particularly as crystallization of the amylopectin leads to redistribution of water within the system. 52 In summary, therefore, we have a system in which the gelatinized starch granules, initially, are semi-spherical, elastic inclusions which progressively firm with the development of crystallinity as monitored by DSC and X-ray. They are, however, the 'filler' in a semi-continuous network of polysaccharides and proteins in the glassy or rubbery states among which amylose may exert a dominating influence. The actual physical state and rheological character of the network is not only affected by the components and their molecular weight and the presence of plasticizers (water or water-sugar systems), but is also susceptible to time-dependent changes, either through diffusion of water into or out of the network, or as a consequence of interactions between the macromolecules.

46

47

41 42 43 44

45

B. M. Gough, C. T. Greenwood, and M. E. Whitehouse, CRC Crit. Rev. Food Sci. Nutr., 1978, 10, 91. G, W. Telloke, Starke, 1985, 37, 17. C. van den Berg, Thesis, Agricultural University, Wageningen, The Netherlands, 1981. F. Bueche, 'Physical Properties of Polymers', 2nd Edn., Krieger, 1979. V. N. Morozov and S. G. Gevorkian, Biopolymers, 1985, 24, 1785.

13

48

49 50 51 52

M. M. Bean and W. T. Yamazaki, Cereal Chem., 1978, 55, 936. K. Larsson, Stiirke, 1980, 32, 125. H. 0. Kim and R. D. Hill, Cereal Chem., 1984, 61, 432. H.F. Zobel and F. R. Senti, Cereal Chem., 1959, 36,441. R. D. Dragsdorf and E. Varriano-Marston, Cereal Chem., 1980, 57. 310. K. Kulp and J. G. Ponte, CRC Crit. Rev. Food Sci. Nutr., 1981, 15, 1. S. Wynne-Jones and J.M. V. Blanshard, Carbohydr. Polym., 1986 (in press).

Flour Proteins: Structure and Functionality in Baked Products

2 Flour Proteins: Structure and Functionality in Baked Products By J. D. Schofield FLOUR MILLING AND BAKING RESEARCH ASSOCIATION,

15

Another aspect of flour protein functionality concerns the possible involvement of proteins associated with starch granules in the improvement ?f 'hi~h-ratio' ca_ke flours brought about by chlorine treatment. A group of protems with properties unlike the bulk of the proteins in flour is associated with starch granules. Recent work has produced indirect evidence that these might be the site of action of chlorine in producing the improver effect, and the proteins of the starch granules are being characterized. The nature of wheat proteins will be examined here with respect to these different aspects of functionality in baked products.

CHORLEYWOOD, HER TS. WD3 SSH, UK

2 Flour Proteins and Breadmaking Quality ' 1 Introduction

The proteins of wheat flour are crucial in determining a flour's breadmaking quality, both protein quantity and qualitative characteristics being important. 1' 2 The ability of wheat flour to be baked into bread is due to the physico-chemical properties of its gluten protein fraction and differences in breadmaking quality between wheat varieties have been shown to be due to differences in this same fraction.1·3 This discovery has prompted an intensive effort to understand the chemical basis for gluten's functionality and its variation from one wheat variety to another, which is largely controlled genetically. The conditions under which wheat is grown may also give rise to modification of the inherent protein quality characteristics of any particular wheat sample. Imbalance in nitrogen and sulphur supply to the developing grain has been of particular interest recently in this respect. 4 - 7 The biochemical basis for the effects of nitrogen~sulphur imbalance on flour protein functionality has received attention, both to define the problem and to identify means by which it might be ameliorated. In Western Europe, in particular, a significant recent development in flour and baking technology has been the use of commercial gluten to enhance the protein level of breadmaking flours. Flours produced in this way have supplanted, to a significant extent, those produced by the traditional procedure of milling flours from blends of home-grown, lower-protein wheats and higher-protein wheats imported from countries outside Europe, especially Canada and the USA. Commercial glutens vary in quality, however, probably owing to different conditions employed during drying, and research has been carried out to define the mechanisms involved_s- io 1 2 3

4 5

6 7 8

9

10

K. F. Finney, Cereal Chem., 1943, 20,381. K. F. Finney and M.A. Barmore, Cereal Chem., 1948, 25,291. M. R. Booth and M.A. Melvin, J. Sci. Food Agric., 1979, 30, 1057. H. J. Moss, P. J. Randall, and C. W. Wrigley, J. Cereal Sci., 1983, 1, 255. C. W. Wrigley, D. L. du Cros, D. G. Fullington, and D. D. Kasarda, J. Cereal Sci., 1984, 2, 15. M. F. Timms, R. C. Bottomley, J. R. S. Ellis, and J. D. Schofield, J. Sci. Food Agric., 1981, 32,684. M. Byers and J. Bolton, J. Sci. Food Agric., 1979, 30,251. M. R. Booth and M. F. Timms, Sixth International Cereal and Bread Congress, Winnipeg, Canada, 1978, paper A4--8. . J. D. Schofield and M. R. Booth, in 'Developments in Food Proteins-2', ed. B. J. F. Hudson, Applied Science Publishers Ltd., London and New York, 1983, p. l. J. D. Schofield, R. C. Bottomley, M. F. Timms, and M. R. Booth, J. Cereal Sci., 1983, 1,241.

14

Classification of Wheat Proteins.-lt is the gluten proteins that are recognized generally as being those of importa~ce in breadmakin~. They r_epresent ab~ut 80-90% of the total proteins of white flour and compnse essentially two maJor protein groups, gliadins and glutenins, which are present in ap~roxi~ately equal amounts. There is controversy over the nomenclature and class1ficat10n of wheat proteins. Gliadins are regarded usually as being those proteins that are not extracted by aqueous salt solutions but that are extractable by concentrated aqueous aliphatic alcohol solutions. Glutenins are extracted neither by sal~ sol_utions, nor, in the-main, by alcohol solutions; they are extractable, however, m dissociating solvents, such as dilute acid, chaotropic agents, soaps, and ionic detergents. However, fractionations based on extractability do not discriminate rigorously between gliadins and glutenins. 11 - i 3 . Gliadins and glutenins are synthesized only in the developmg endosperm of the wheat grain; they are deposited in protein storage bodies, and they ~ave _similar amino-acid compositions, being rich in proline, glutamate, and amide mtrogen and poor in charged residues, particularly the basic amino-acids. 14 The genes controlling the synthesis of gliadin and glutenin polypeptides are located on homoeologous group 1 and group 6 chromosomes and the genes for certain gliadin and glutenin polypeptides occur in very tightly linked groups. 15 Furthermore, glutenin polypeptides may be extracted in concentrated aqueous alcohol solutions in the presence of a reducing agent and dilute acid, particularly at elevated temperatures (up to 80 °C). 16 There is much to support the proposal, therefore, that both gliadin and glutenin proteins should be classified as prolamins.14 Important chemical and physical_ differences exist between the gliadin and glutenin proteins, however, that are relevant to their functioual cha~acteri~tics. The gliadin fraction comprises monomeric proteins; where present, d1s~lph1de bon~s are intra-chain. 17 The glutenin fraction, on the other hand, compnses polymenc

11 12

13 14

1s 16 17

J. A. Bietz and J. S. Wall, Cereal Chem., 1975, 52, 145. p I Payne and K. G. Corfield, Plan/a, 1979, 145, 83. E: Jackson, L. M. Holt, and P. I. Payne, Theor. Appl. Genet., 1983, 66, 29. P.R. Shewry, B. J. Miflin, and D. D. Kasarda, Phi/os. Trans. R. Soc. Lond., B, 1984, 304,297. p_ r. Payne, L. M. Holt, E. A. Jackson, and C. N. Law, Philos. Trans. R. Soc. Lond., B, 1984, 304,359. M. Byers, B. J. Miflin, and S. J. Smith, J. Sci. Food Agric., 1983, 34, 447. . J. S. Wall, in 'Recent Advances in the Biochemistry of Cereals', ed. D. L. La1dman and R. G. Jones, Phytochem. Soc. Eur. Symp. Ser. No. 16, Academic Press, London, 1979, p. 275.

A.

Chemistry and Physics of Baking

16

proteins, whose component subunits are almost certainly linked by mter-chain disulphide bonds since they cannot be dissociated in strongly dissociating solvents but are readily dissociated on addition ofreducing agents. 18 •19 The glutenin proteins are also extremely polydisperse with respect to molecular size. 12 • 17 • 19 As many as 50 different polypeptides may compose the gliadin fraction 20 and as many as 20 the glutenin fraction as determined by two-dimensional electrophoresis techniques. 21 • 13 The polypeptides present in each fraction are quite distinct, as shown by high-resolution 2,D electrophoresis. 13 The two fractions also have quite different physical properties when hydrated, the gliadin fraction behaving as a viscous liquid and the glutenin fraction as a cohesive elastic solid. 1 7 The Component of Gluten Responsible for Breadmaking Quality .-The rheological properties pf gl~ten (i.e. its combination of viscous, elastic, and cohesive properties) are generally considered to be responsible for conferring breadmaking quality on wheat flour, and both the gliadin and glutenin fractions contribute to the viscoelastic nature of gluten. Differences in the glutenin fraction appear to account in large part for differences in gluten quality between wheat varieties. 22 A model has been proposed for the structure of glutenin, called the 'linear glutenin hypothesis', that can account for many of the technologically important characteristics of dough. 23 •24 Glutenin molecules are considered as linear chains of polypeptide subunits called 'concatenations', the subunits being joined headto-tail by inter-chain disulphide bonds (Figure 1). An important feature of

~ ~ S

Stretching ~

T T~-~, a-helix region

13-turn

~ s

s

s

~

region

Figure 1 Schematic representation of a polypeptide subunit of glutenin within a linear concatenatio,n. The subunits are joined head-to-tail via disulphide (S-S) bonds to form polymers with molecular weights of up to several million. 23 •24 The subunits are considered to have a conformation that may be stretched when tension is applied to the polymers, but when the tension is released.the native conformation is regained through elastic recoil. 23 •24 The N- and C-terminal ends of some high molecular weight subunits, where interchain S-S bonds are located, are now thought to be a-helix-rich domains whereas the central domains are thought to be rich in repetitive P-turn structures. 37 The presence of repetitive P-turn structures may result in a P-spiral structure, which may confer elasticity 31

18

19

2

°

21 22

23

24

D. D. Kasarda, J. E. Bernardin, and C. C. Nimmo, Adv. Cereal Sci. Technol., 1977, 1, 158. R. C. Bottomley, H. F. Kearns, and J. D. Schofield, J. Sci. Food Agric., 1982, 33, 481. C. W. Wrigley and K. W. Shepherd, Ann. N. Y. Acad. Sci., 1973, 209, 154. P. I. Payne, C. N. Law, and E. E. Mudd, Theor. Appl. Genet., 1980, 58, 113. B. J. Millin, J. M. Field, and P. R. Shewry, in 'Seed Proteins', ed. J. Daussant, J. Masse, and J. Vaughan, Phytochem. Soc. Eur. Symp. Ser. No. 20, Academic Press, London, 1983, p. 255. J. A. D. Ewart, J. Sci. Food Agric., 1977, 28, 191. J. A. D. Ewart, J. Sci. Food Agric., 1979, 30,482.

Flour Proteins: Structure and Functionality in Baked Products

17

the model also is the nature of the individual subunits. The conformation of each subunit is envisaged as being such that when tension is applied to the ends of a subunit the conformation can be stretched out of its native state and, conversely, when the tension is removed the structure recoils to regain its original conformation, i.e. its lowest-energy state (Figure 1). Also considered in this hypothesis was how individual glutenin molecules would interact with each other, and with the gliadin proteins, and what effect such interactions would have on the technological and chemical properties of dough systems. In a freshly wetted dough the glutenin polymers are envisaged as being oriented randomly with respect to each other (Figure 2). Although there are numerous interactions between molecules at their points of contact through weak, non-co_valent bonds, the overall contribution of such secondary forces to the mechanical strength of dough is small because of the random orientation ofthe molecules.

I ->Jr

.1

p~ Figure 2 Schematic representation of how glutenin polymers may become aligned as a result of the shear stresses imposed during dough mixing. The numerous weak secondary bonds between the polymers in an unmixed dough contribute little to dough strength because of the random orientation of the polymers. Alignment of the polymers during dough mixing results in these secondary forces now acting co-operatively and dough strength is increased markedly 23 •24

On application of shear, as during dough mixing, the weak secondary bonds between polymers are broken easily and the polymers tend to become aligned in the plane of the shear stress (Figure 2). New secondary bonds form as polymers come into contact, but now these individually weak forces act co-operatively because of the alignment of the concatenations. The result is that dough strength is increased substantially: such a mechanism essentially accounts for dough development at a molecular level. Stress relaxation in a dough occurs through slippage of molecules past one another and through breaking and reformation of the disulphide bonds linking subunits together. Dough breakdown, on the other hand, occurs when the instantaneous tension in a concatenation reaches such a magnitude that molecular slippage and disulphide cleavage and reformation are too slow to relieve that tension. In this situation, covalent bonds holding the subunits together are literally pulled apart, scission most likely occurring at disulphide bonds in the central

Chemistry and Physics of Baking

18

region, of the concatenations where tension is greatest because of the probability that the largest numbers of effective secondary forces are distributed about the mid-point of the chain. Polymeric Structure of Glutenin Molecules.-A basic tenet of the model described above is that glutenin molecules are polymers comprising polypeptide subunits linked by inter-chain disulphide bonds. Glutenin extracted from flour or dough using strongly diss@ciating splvents has been shown by gel filtration chromatography to comprise extremely large structures with estimated molecular weights ofup to several million (Figure 3). 25 •12 • 19 Glutenin molecules are not homogeneous with respect to molecular/size, however, but rather they are extremely polydisperse with molecular weights ranging from about 10 5 upwards. Addition of reducing agents to glutenin extracts diminishes the molecular size drastically such that on gel filtration chromatography essentially all of the protein appears as components with molecular weights of about 10 5 or below. 19 These findings have been confirmed by viscosity measurements. Glutenin solutions have very high intrinsic viscosities that are diminished markedly on addition of reducing agents. 26 •24 Such evidence indicates that glutenin molecules are indeed polymers comprising subunits linked together by disulphide bonds.

E

0·1

r-r-N

~ C

ro

1: 0 V)

0 0·2

.0

Oat

Rye

-I!"'. t

Stearoyl

Table 2 Effects of emulsifiers and flour lipids on loaf volume and characteristics of bread baked from partially de-fatted flour ( data from ref 11)

t

~~-i·.','.,J,..,S.yl,( Polyglycerol esters of fatty acids

Figure 1 Chemical formulae and structural models of emulsifiers

185 273 227 232 155 159

6---7 6 6 7 7 7

261 270 255 262 259 260

6---7 6 5--6 6---7 5 6---7

175 285 210 296 157 255

"Diacetyltartaric acid esters. 'Polar lipids from wheat flour. 'Non-polar lipids from wheat flour. 'Total lipids from wheat flour. 7

8

Diacetyl tartaric acid esters of monoglycerides

6---7 6 5-6 6---7 6 5-6

9

G. Schuster and W. F. Adams in, 'Advances in Cereal Science & Technology', Vol. VI, ed. Y. Pomeranz, American Association of Cereal Chemists, St. Paul, MN., USA, 1984, p. 139. R. C. Hoseney, K. F. Finney, and Y. Pomeranz, Cereal Chem., 1970, 47, 135. F. MacRitchie, in 'Lipids in Cereal Technology', ed. P. J. Barnes, Academic Press, London, 1983, p. 165.

II

Chemistry and Physics of Baking

82

starch granules in a dough and, thereby, contribute to dough elasticity, allowing 10 gas cells to expand and resulting in an increased volume of baked goods. Added polar surfactants probably act in the same way as polar flour lipids. All hydrophilic emulsifiers like DATEM and SSL form lamellar liquid-crystalline phases in water at the temperature of dough mixing. Therefore, they may combine with native, polar flour lipids in membrane-like structures in the dough and thus enhance the polar flour lipids in their baking function, i.e. improve the viscoelastic properties and reduce the rate of gas diffusion in t.he dough. In a recent study by Pomeranz 11 the functions of DATEM and lecithin were compared to those of natural p~lar and non-pola; flour lipids as shown in Table 2. Non-polar flour lipids were found to be detrimental to the baking quality, when added back to defatted flour. The polar flour lipids, added lecithin, and particularly the DATEMs improved loaf volume and crumb characteristics. Fats added in a 5% concentration (based on flour weight) also improved volume and overall score, but not as much· as 0.3% DATEM. Liquid oil (5%) performed poorly when added alone, but in conbination with 0.3% DATEM or polar flour lipids gave the best volume and overall score. Studies of the expansion of yeast-raised doughs made with Danish flour during the baking process, using an electrical resistance oven, show the effect of various surfactants as dough strengtheners (Figure 2). This demonstrates the relative effect of 3% shortening (lard) and 0.3---0.5% emulsifiers (based on flour weight) on set time and on increases in dough height during baking. As a general trend, all added

50

e

.s

750c

40

Q)

/n"C--------•ssL,

"'.,, Q)

ti

0--------------- _ -□ DATEM, 0.3%

- - - - - - - - - - - - , , . Shortening, 3.0% 0.5%

30

·1.-1°c·····••""···--······-·······•·A GMS Hydrate, 0.5%

...E .c

83

The Role of Fats and Emulsifiers in Baked Products

lipids gave the same rate of increase in dough height, but the doughs stopped expanding at different set points. GMS-hydrate and SSL gave, respectively, initial set points at 50 °C and 55 °C, after which the rate of height increase changed until the final set point of 77-79 °C was reached. Shortening (3%) gave a set point at 60 °C and a height increase of 36 mm, whereas 0.3% DATEM gave a set point of 75 °C and· a height increase of 40 mm. The control dough, without added lipids, gave a much lower rate of height increase, resulting in a low final dough volume although the set point was as high as 77 °C. The differences in rate of height increase may be due to the effect of the added lipids ort the rheological properties of the dough. The fact that monoglycerides also contribute to the volume of baked goods, although they normally are referred to as crumb softeners rather than dough conditioners, shows that there is no sharp distinction between the dough strengthening effect and crumb softening effect of the commercial emulsifiers in use. Most emulsifiers have multifunctional effects in baking products, but with one function dominating over the others. However, the relative effect on volume of baked goods with different emulsifiers varies with the type of flour and the baking technique used. Table 3 Effects of emulsifiers on volume of Danish hard rolls Fat or emulsifiers added(% on flour basis)

Volume/cm 3 often rolls each o/60 g dough

Lard (2.0%) Lard(4.0%) CSL (0.2%) CSL (0.4%) DATEM (0.2%) DATEM(0.4%) Monoglycerides (0.2%) Monoglycerides (0.4%) Control, no addition

3025 3140 2500 3100 2950 3290 2760 2915 2125

In Europe, DA TEMs are used especially in the production of hard rolls, and Table 3 shows the effect of DATEMs and CSL, compared with that oflard, on the volume of rolls made with Danish flour.

en

·; :r

20

.c

en

:, 0 Q

~-----------o 77°c

10

2

6

10

14

18

Control (no lipids added)

22

Time (min.)

Figure 2 Effects offats and emulsifiers on dough volume and set time during baking of Danish wheat bread. Concentrations of added lipids are based on flour weight 1

° K. Larsson, in 'Lipids in Cereal Technology', ed. P. J. Barnes, Academic Press, London, 1983, p. 237.

11

Y. Pomeranz, Cereal Chem., 1984, 61, 136.

Anti-staling or Crumb Softening Effects of Emulsifiers.-The types of emulsifiers which are used primarily for their function as anti-staling agents or crumb softeners are distilled, saturated monoglycerides, but SSL, SMG, and DATEMs also provide some crumb softening effect in bakery products, although generally not to the same extent as distilled monoglycerides. Monoglycerides are manufactured from edible fats and oils of animal or vegetable origin by interesterification with glycerol. The resulting product is a mixture of mono-, di, and tri-glycerides, containing from 35% to about 60% of the total as monoglycerides. Concentrated monoglycerides are made by subjecting the monoglyceride blend to a molecular distillation process, whereby the monoglycerides are separated from the di- and tri-glycerides. The resulting products are usually referred to as distilled monoglycerides and contain a minimum of 90% as 1-monoglycerides

Chemistry and Physics of Baking

84

(usually 93-95%), 2-3% 2-monoglycerides, and 3-5% diglycerides, with traces of triglycerides, free fatty acids, and glycerol. Depending on the type of fat or oil used as a basis for the monoglyceride manufacture, the melting point and physical form may vary considerably. The high-melting, saturated monoglycerides are used predominantly in low-fat, or fat-free bakery products, as anti-staling agents in bread, or as aerating agents in cakes. The soft, plastic monoglycerides are used especially, in combination with fats, in high-fat products such as sweet goods. Distilled monoglycerides for use in bread and speciality breads as an anti-staling agent or crumb softener are ,traditionally used in the USA and in the UK in the form of so-called hydrates. A hydrate is a suspension of monoglyceride crystals (B-form), usually 20-25%, in water. The hydrate has a smooth, creamy consistency and is easy to disperse in a dough system, ensuring good distribution of the lnstron g/cm2

130

------------

120

- - • Control

,,,.• SDM Powder .,,,,✓

110

.,,

I

I I

100

,,,. /

/

/

I

I

I

90

I

80

70

/ It

I

GMS Hydrate

I

I

I

I

0

•✓

I

I

I I

0

:jl

/

48

monoglycerides. Blends of saturated and unsaturated monoglycerides in the form of a fine powder are also used as anti-staling agents and have the advantages of being dispersible in cold water. They can therefore be added directly to the dough in a concentrated form. The anti-staling and crumb softening effect of monoglycerides is primarily due to their interaction with the starch fraction of bread. The addition of monoglycerides to doughs results in a decrease in starch gelatinizatlon and precipitation of water-soluble amylose, inhibiting formation of amylose gels which would increase crumb firmness. Monoglycerides are also able to complex with amylose inside the starch granules. The retrogradation of the amylopectin fraction of starch is considered to be the major cause of the staling process and, therefore, any delay in retrogradation of amylopectin will decrease the rate of staling. The amylosemonoglyceride complex formed in doughs may possibly interfere with the retrogradation of amylopectin and consequently reduce the staling rate. The relative effects of various monoglycerides on the retardation of firmness of bread are shown in Figure 3, which demonstrates that saturated monoglycerides are very effective as anti-staling agents. Cakes and Cake Mixes.-The types of emulsifiers permitted for use in cakes, biscuits, etc. in the European countries are listed in Table 4. In the UK, cakes are not standardized food products and therefore all permitted emulsifiers can be used. Mono-diglycerides or distilled monoglycerides represent by far the most widely used group of emulsifiers in cakes and related products. They may be used alone in shortenings, to increase cake volume and shelf life, or used in combination with other emulsifiers to achieve special benefits. The use of monoglycerides in cake shortenings, which began in the 1930s, has led to a change in cake formulae from the old fashioned pound-to-pound cake, to today's high-ratio cake. The use of fluid shortenings began about 1970, initiated by a desire for bulk handling of shortenings. Fluid shortenings are mixtures of hard fat and surfactants in vegetable oil. The surfactants used in fluid shortenings are combinations of monoglycerides, lactylated monoglycerides (LMG), and propylene glycol esters (PGMS).

40 □

24

The Role of Fats and Emulsifiers in Baked Products

72

96

120

144

Hours

Storage Time Figure 3 Effects of emulsifiers on crumb firmness of German wheat bread. Concentrations

based on flour weight, control bread, no emulsifiers added. SDM: 0.5% partially unsaturated, distilled monoglycerides ( Amidan SDM)*; GMS: 0.5% saturated, distilled monoglycerides ( Dimodan PM)* as a 25% hydrate • Man~factured by Grindsted Products A/S, Denmark.

The lactylated monoglycerides are reaction products of saturated monoglycerides (often distilled) and lactic acid (88% pure). Direct esterification of glycerol, lactic acid, and a blend of palmitic and stearic acids is an alternative method. LME is not polymorphic like the corresponding monoglycerides, but stable in an a-like crystal form; the melting point is about 42 °C. Propylene glycol esters (PGME) are produced by esterifying propylene glycol with blends of palmitic and stearic acids, yielding a mixture of propylene glycol mono- and di-esters. The monoesters can be concentrated by a molecular distillation process, and a product containing at least 90% of propylene glycol monostearate (PGMS) is obtained. PGMS is also stable in an a-like crystal form (m.p. 43 °C) and, together with LMG and acetylated monoglycerides, is often referred to as an a-tending surfactant. The formulae of LMG and PGMS are shown in Figure 1. LMG and especially PGME are important surfactants in shortenings for cake mixes, in which they are used in an amount of6--10%, of the weight of shortening

Chemistry and Physics of Baking

86 0

""

0

""

0

0

"" ""

The Role of Fats and Emulsifiers in Baked Products

87

content. PGME is often used in combination with monoglycerides in fats for household cake mixes. A recent development is the use of liquid vegetable oil (salad oil etc.) in cake manufacturt. Liquid oils were originally found to be unsuitable for cakes, since addition of oils gave low volume, open grain, and poor structure. However, by reducing the total level of fat and by using the correct types and blends of hydrophilic and lipophilic emulsifiers, good quality cakes can be produced with liquid oils instead of shortenings. Distilled PGMS can be used in a concentration of 8-12% of the oil content in high ratio cakes, where the fat content may be reduced from 21 % shortening to 14% oil, based on flour weight. A considerable reduction in fat level is thus obtained, yielding a reduced-calorie product with equivalent quality in terms of volume, grain, and texture. Aqueous gels of hydrated emulsifiers, such as distilled monoglycerides, PGMS, and polysorbate 60, can be added directly to the cake batter with all other ingredients before mixing. Table 5 demonstrates the use of emulsifiers in yellow layer cakes made with plastic shortenings or liquid oil. Table 5 Use of emulsifiers in yellow layer cakes Shortening-oil-emulsifier level onfiour weight

0 M

0 M

0 M

0 M

36% shortening with 7% PGME 36% non-emulsified shortening + 5% emulsifier gel" 18% liquid soy bean oil + 5% emulsifier gel

Crumb

Batter density/ gJ-1

Cake volume/ cm 3

texture

664

1340

Good, slightly dense

628

1395

Good

524

1815

Good, very light

I!

"The gel contains 30% emulsifiers. 1. 0

N

0

N

Sponge cakes contain no added fats and, therefore, represent bakery products in which aqueous gels of hydrated emulsifiers are used as aerating agents to facilitate 'all in' mixing of products and a short mixing time. Emulsifiers also function as a replacement for part of the egg content in sponge cakes and thus give savings on a costly ingredient. At the same time, there is a reduction in calorie content of the finished product. The hydrated gel contains 20-30% of a blend of distilled monoglycerides and PGMS together with 1-2% of an anionic co-emulsifier. The blend of emulsifiers must be stable in the a-crystal form to obtain optimal functionality. Aqueous gels ofpolyglycerol esters of fatty acids are also used as aerating agents in sponge cakes or other low-fat cakes (Swiss rolls). Spray-dried emulsions of cake emulsifiers, such as distilled monoglycerides, PGMS, and LMG, containing non-fat milk powder as a carrier, are used in sponge cakes with the same benefit as aqueous gels. The advantages of using cake emulsifiers are not only as a processing aid to reduce labour and mixing time but also as a cost-reducing factor functioning as an egg replacer and a fat-sparing ingredient. Industrially produced cakes need a long shelf life in order to allow transport and storage on the way from manufacturer to the consumer. The emulsifiers available to the industry make it possible to produce such products.

I

I

I

I

I I

88

Chemistry and Physics of Baking

Legal Aspects.-In Europe, the EEC Council directive of July 1978 (C78/666/EEC) lists permitted emulsifiers for use in foods. The emulsifiers are designated by an E-number and may be authorized for use by all member states of the EEC. In the USA food emulsifiers are regulated by the Food and Drug Administration (FDA). All emulsifiers which are approved by the EEC or the US-FDA can be considered safe food ingredients when used within the limits of their acceptable daily intake (ADI) level, which has been established by the FAO/WHO Joint Expert Committee on Food Additives (JECFA) in 1974.

7 Protein Functionality in Bakery Products By A. P. Davies UNILEVER RESEARCH LABO RA TORY, COL WORTH HOUSE, SHARNBROOK, BEDFORD MK44 I LQ, UK

1 Introduction

The eating properties of a bakery product are influenced by the nature and interaction of ingredients and the processing regime used in its manufacture. What we recognize as properties or attributes of the product are a manifestation of its physical characteristics and these are determined, in part, by the chemical reactions/interactions which have occurred during the assembly and heat-setting processes. The wide range of bakery products which exist in the market place today is largely to the credit of the skills of the artisan bakers, but the application of scientific understanding to the area is now generating new products, ingredients, and processes. Irrespective of the identity of the final baked product, there are stages in the manufacture of baked products which are essentially common to all, viz. (i) (ii) (iii) (iv)

Hydration of flour and ingredients Development (by mixing) of a raw batter/dough structure Introduction of gas phase-by design or otherwise Heat-setting of batter/dough into a rigid crumb structure which maintains its integrity when the product is cooled

The kinetics of these events and the temperatures at which they occur will have a strong bearing on the physical characteristics of the product. Proteins play a dominant role in creating baked products which are desirable to the consumer, their multifunctionality enabling them to take part in a number of the stages occurring between raw materials and product. An appropriate product system, to illustrate the roles which the proteins may play, is the sponge cake.

2 Stages in the Preparation of a Sponge Cake The main ingredients in sponge cakes are eggs, sugar, and flour and a gas phase, which is introduced partly in the mixing process and partly from the release of carbon dioxide (CO 2 ) from baking powders. In recent years, when sponge cakes have been produced on a large scale and where shelf-life and ease of preparation are important factors, other ingredients-particularly fats and lipid surfactants-are added. In the traditional method of sponge cake preparation the air is beaten into the sugar-plus-egg and the flour is folded into the beaten foam. The use of lipid surfactants (aerating agents) removes the need for the two-stage 89

90

Chemistry and Physics of Baking

process and enables cakes to be made by the all-in procedure. This was an important development in sponge cake manufacture and influences the roles played by proteins. Before considering the role which the proteins-from wheat, egg, milk-may play let us first explore the stages which occur in sponge cake manufactm.::, namely batter formation, heating, drying, and, finally, cooling. The action of mixing and beating a cake allows hydration of the dry components and entrainment of air. The nature of the physical energy input and the presence of surfactants determines the phase volume of entrained air and the air cell size distribution and number. Initially, large air bubbles are incorporated by mixing and these are stabilizyd by surfactants and subsequently distended/stretched by the mixing action, until "break-down into an increased number of smaller bubbles occurs. Their immediate integrity is dependent upon there being available adequate surfactant for surface coverage and stability. Since the pressure (p) in smaller bubbles is larger than that in big bubbles (!).p = 2y/r) [where /).p = p(bubble) - p(atmosphere), y represents surface tension, and r is the bubble radius] there is a tendency for small bubbles to disproportionate so as to lower the surface energy of the system. Bubble instability will also result from film drainage but is considered to be less important than disproportionation in relation to cake batter instability. 1 The hydration of flour components-proteins, damaged starch, hemicellulose, etc.-is accompanied by a reorganization of the proteins and the formation of a high, non-Newtonian viscosity batter. 2 This has a marked effect on the rate of migration of bubbles and therefore exerts control on the stability of batters by retarding disproportionation and bubble rise. Most batters, however, permit slow migration of bubbles and, over a period of time, bubbles disproportionate and rise towards the surface with a consequent change in size distribution/number and lead to a loss of gas-phase volume. A third aspect which contributes to batter structure is the generation of carbon dioxide gas. Baking powders are usually incorporated in sponge cake recipes and will release CO 2 gas during the various stages of manufacture-during mixing, during resting, and during baking. The CO 2 generated dissolves in the water as carbonate and bicarbonate ions and is in equilibrium with free and dissolved undissociated CO 2 gas. Diffusion of CO 3 z- /HCO 3 - ions will occur throughout the aqueous phase, lowering the pH and saturating air bubbles with CO 2 • Consequently, bubble pressure will increase and bubbles expand, providing that the continuous phase allows; a portion of the CO 2 will diffuse into the atmosphere and will be lost from the batter. Hence the final phase voiume and bubble number/size distribution of gas in a batter will be a function of (i) the nature of the physical energy input and presence of surface active agents in the batter, (ii) the stability of the gas phase towards coalescence and flotation, and (iii) the volume of the CO 2 generated during the mixing/resting period. 3 Most UK sponge batters have a 'Batter Specific Volume' of2.0-3.0. 1 2

3

R. Brown and M. Galashan, personal communication. A G. Naylor and M. Hale, unpublished work. J. Robb in FM BRA Bull., 1984, No. 6.

Protein Functionality in Bakery Products

9i

As the temperature of a flour batter is raised, numerous chemical amd physical changes occur and these are manifested in the final structure of the baked product. Heat transfer occurs by conduction, with virtually no mass movement of the batter while bubble expansion progresses during baking from the outer hot layer through to the central region. 2 The fluidity of the batter increases until starch gelatinization and protein gel formation cause a substantial increase in batter rigidity. The solubility of CO 2 in the aqueous phase decreases with increasing temperature which enhances its diffusion into the gas phase cells (and additionally out into the atmosphere) causing a volume increase (Figure 1). Additional CO 2 will also be

0 I 0 L..

1.4

70 X

1.2

60

(lJ

cl

:: 1.0

""E ""---E

?

1i ::::J

!

50 'E z -"'

0.8

40

(lJ L..

::::J

"'"'

0.6

(lJ

30 a. L..

0

"'

L..

::::J 0

N

u 0.4

0

20

a.

CJ

>

L..

(lJ

0.2

0

10 40

50

60

70

80

90

100

0

Temperature (OC)

Figure 1 Temperature effect on CO 2 solubility and water vapour pressure

generated from the residual baking powder. The gas phase, which now comprises air, CO 2 , and water vapour, expands with increasing temperature owing, primarily, to the water vapour pressure. Expansion is dramatic after 60 °C and continues until starch gelatinization and protein gel formation increase the viscosity of the batter to a point at which further expansion is arrested (Figure 2). The structure eventually changes to a sponge from a viscous foam, and this change is essential if the cake is to remain in its fully expanded form. This event starts at the surfaces where bubbles burst, owing to the excess pressure within the bubbles compared with atmospheric pressure, and to the loss of wall flexibility due to drying. A slight volume decrease occurs in the final stages of baking when the bubbles of a viscous foam burst and the matrix structural elements buckle owing

o

3:

Chemistry and Physics of Baking

92

Protein Functionality in Bakery Products

93

The above discussion summarizes available information on the events taking place in sponge cake manufacture. Are proteins involved and, if so, in what capacity?

12

3 Composition and Functional Properties of Proteins Used in Sponge Cakes

The multifunctionality of proteins has resulted in extensive usage in foods where they may contribute to water absorption, water binding/thickening/gelf[tion, emulsification, elasticity, fat absorption, foaming, etc. 6 The type of functional property required of a protein ingredient depends upon the nature of the particular food in which the protein is to be incorporated. In sponge cakes, five types of functionality are readily defined where the contribution from protein may be desirable but where other ingredients may be fulfilling these requirements, viz.

9

E u

'.... £

(JI

.ii, £ Cb

-"'

6

Cl

u

Water Vapour

(i) (ii) (iii) (iv) (v)

60°C

3

CO2

5

10 Time/min

15

Figure 2 Cake expansion during baking

(A.G. Na:ylor and M. Hale, to be published)

to incomplete solidification. 2 In addition, severe collapse of a cake can occur after its removal from the oven, The excess pressure in the bubbles-due to water vapour pressure-is reduced when the temperature decreases and, if this cannot be equilibrated with the atmosphere (foam structure) and if the cake matrix has some flexibility, then cake collapse is the result. This will occur in the upper centre of the cake, which is the last part to reach the final temperature. This event can be prevented by either (i) allowing the cake to cool under reduced pressure 4 or (ii) dropping the hot cake 6-20 cm on to a solid surface. 5 The former method reduces the effect of the· atmospheric pressure on the cake allowing the structure time to solidify whereas the latter encourages rupture of the cell walls by creating a shock wave in the cake. Neither method, in our experience, provides a cake with eating qualities comparable with that of a cake which does not collapse naturally. 4 The permeability of non-collapsed cakes is greater than that of collapsed cakes.2

Several factors 7 - 9 affect the functional properties of a protein: (i) the intrinsic properties of the protein, which are related to its amino-acid composition and sequence, its secondary and tertiary structures, and whether it is native or denatured (effect of isolation/processing) and (ii) interactions and reactions with other components of the product, e.g. pH, ionic strength, lipids, proteins, enzymes, redox agents, water, etc. The composition and physical/chemical properties in molecular terms are readily determined, but the relationship of molecular structure to product properties is only partly achievable, since we do not have a sound understanding of the nature of the many interactions that can occur between proteins and other components of a food product. In this paper we shall examine the properties of a number of proteins and consider their functionality in a sponge cake. The proteins selected are (i) wheat proteins, since they are ubiquitous to baked products, (ii) egg proteins, which provide many of the desirable eating qualities of sponge cakes, and (iii) milk proteins which have been used as potential egg (white) replacers or which are added as 'carriers' for lipid surfactants or are included for ill-defined but likely beneficial purposes, e.g. flavour, appearance, etc. Wheat Proteins.-Proteins play a major functional role in flour, affecting water absorption, reduction/oxidation, rheology, gas retention, and product quality. The major proteins in flour are gliadin and glutenin which possess similar aminoacid compositions. Few of the carboxy-groups are free to ionize while the low 6

7 8

4 5

R. Yoell, personal communication. H. Ohtsubo, T. Kanbe, Y. Kaneko, and S. Nomura, Cereal Foods World, 1978, 23,361.

Air incorporation Air stabilization Fluidity of batter during gas expansion stage Structural setting--coagulation Transformation of foam to sponge structure

9

J.E. Kinsella, 'Food Proteins', ed. P. F. Fox and J. J. Condon, Applied Science Publishers, London, 1982. J.E. Kinsella, Crit. Rev. Food Sci. Nutr., 1976, 7,219. J.E. Kinsella, J. Am. Oil Chem. Soc., I 976, 56, 242. A. Pour-El, in 'Protein Functionality in Foods', ed. J. Cherry, Am. Chem. Soc. Symp. Ser., No. 147, 1981.

Chemistry and Physics of Baking

94

levels of lysine, histidine, and arginine, which are capable of acquiring positive charges in solution, result in the low solubility of the proteins. The relatively high levels of non-polar groups facilitate hydrophobic interactions, the unusually high glutamine content facilitates hydrogen bonding,- and the presence of proline discourages a-helix formation. Overall, wheat proteins have a high tendency towards aggregation. 10 Gliadins are a heterogeneous range of proteins with an average molecular weight of ca. 40 000 and composed of single polypeptide chains. lntramolecular disulphide bonds are present which confer stability on folded random coil structures. Glutenins are also a heterogeneous range of proteins with molecular weights of up to 10 7 . They consist''of polypeptide subunits ranging in molecular weight from ca. 40 000 to 150 000, which are capable of forming both intra- and interpolypeptide disulphide bonds.11 - 13 The existence of glutamine residues and the solubility of a proportion of the glutenin in urea, guanidine, or detergent (sodium dodecyl sulphate, SDS) indicates that both hydrogen bonding and hydrophobic interactions are also involved in the association of glutenins. Albumins and globulins 14 are more soluble than glutenins or gliadins, and have molecular weights ranging from 2000 to 16 000 for albumins and from 20 000 to 200 000 for globulins. ,06

';'

E

u

(I)

10 5

C

>-

:s (!)

,/·~·

ti)

w••

~ :::, ""CJ 0

E

b

,'.'

····················,,,

10 4 ••••

u

,,,1

~

C_____ .,. ,,/

Protein Functionality in Bakery Products

95

The two major protein classes present in wheat, viz, glutenins and gliadins, have elastic and extensible/cohesive properties respectively. Addition of wheat gliadin 15 to gluten (an association of glutenins and gliadins together with other minor components) results in an increase in extensibility while the resistance to extension decreases. As the temperature of gluten is raised the fluidity increases (elastic modulus G' decreases) until the starch begins to gelatinize when there is a dramatic increase in the elastic modulus. 16 In the absence of starch (or when present at low levels) an increase (Figure 3) in the elastic modulus occurs at ca. 80 °C. At this temperature additional protein-protein aggregation is occurring and a network is being formed; cysteine residues are implicated since the solubility of the proteins in, e.g., SDS is decrea~cd but restored in the presence of mercaptoethanol. In summary: Flour proteins are viscoelastic, thermally set at ca. 80 °C, and have poor aeration properties.* Egg White Proteins.-Liquid whole egg 1 7 consists of ca. 64% albumen, 36% yolk (dry weight basis), and contains 74% water. Egg albumin 18 - 21 is essentially an aqueous solution of various globular proteins within a network of ov9mucin fibres. Twelve protein types are present, the major components of which are ovalbumin, conalbumin, and ovomucoid. Ovalbumin (mo!. wt. 45 000) is the major protein constituent. It is classified as a phosphoglycoprotein and during storage changes to S-ovalbumin, 22 probably via thiol--disulphide interchange. In this form 1t is more heat stable, but otherwise indistinguishable in physical and chemical properties. Hence in product use, one could expect different functionality from stored egg and fresh egg. Conalbumin (mo!. wt. 80 000) is a glycoprotein which readily binds iron, aluminium, copper, and zinc 23 and is more sensitive to heat than ovalbumin, but contains no free thiol groups. Ovomucoid (mo!. wt. 28 000) is a non-heat-coagulable glycoprotein which has been shown to consist of three components differing in sialic acid content. 24 Lysozyme, in contrast to all the other egg albumin proteins, is a basic protein; its mo!. wt. is 17 000, it contains no free sulphydryl groups, and is classified as a globulin. Two other globulins are present (ovoglobulins E 2 and E 3 ) and, together with lysozyme, are rated the most effective foaming agents of the albumin proteins.

0

w 10 3

* Editor's footnote: A more detailed discussion of wheat protein functionality is given in Chapter 2 of this

20

30

40

50

60

70

80

90

100

Temperature (°C)

volume. 15 16

Figure 3 Influence of starch and temperature on gluten rheology ( starch content decreasing in the order a> b > c)

17 18

19 10 11 12

13

14

Y. Pomeranz, in 'Wheat Chemistry and Technology', Am. Assoc. Cereal. Chem., St. Paul, MN, 1971. J. Ewart, J. Sci. Food Agric., 1972, 23. 687. J. S. Wall, in 'Recent Advances in B10chemistry of Cereals', ed. D. Laidman and R. Wynn-Jones, Academic Press, N.Y., 1979. J. Ewart, J. Sci. Food Agric., 1968, 19,618. E. Porceddu, D. Lafiandr&, and G. T. Scarascia-Mugnozzo, in 'Seed Proteins', ed. Gottschalk and Mullen. Martinus Nijhoff, 1983.

20 21

22 23

24

H. D. Belitz et al., personal communication. A. P. Davies, R. K. Richardson, and S. B. Ross-Murphy, unpublished work. A. L. Romanoff and A. J. Romanoff, The Avian Egg', J. Wiley & Sons Inc., N.Y., 1949. J. D. Hom, in 'Applied Protein Chemistry', ed. R. A. Grant, Applied Science Publishers, London, 1980. W. J. Stadelman, 0. J. Cotteril, et aL, 'Egg Science and Technology', AVI Publishing Co. Inc., Westport, Connecticut, 1977. US Patent 3 268 346. J. C. Carter et al., 'Egg Quality', British Egg Marketing Board Symposium, 1968, No. 4, Oliver & Boyd Ltd., Edinburgh. T. L. Parkinson, J. Sci. Food Agric., 1966, 17, 101. M. B. Smith, Aust. J. Biol. Sci., 1964, 17,261. M. B. Rhodes, N. Bennett, and R. E. Feeney, J. Biol. Chem., 1960, 235, 1686.

Chemistry and Physics of Baking

96

The remaining proteins make up a low percentage of the total protein and have not yet been related significantly to the functionality of egg white. Proteins are amphipathic and, consequently, are surface active. In their native state the polypeptide chains are folded to minimize the free energy of conformation. 25 Adsorption at an interface can be sufficient to cause denaturation since an alternative way of folding the molecule to minimize the free energy becomes possible. 26 Thus, besides folding to the interior of the globule in the aqueous phase, apolar side chains can be located in the air phase and lead to the formation of adsorbed films containing,trains of amino-acids which are in contact with the surface and loops and tails of residues residing in either bulk phase: air or water (Figure 4). Foamability is assiste"d by a decreased interfacial tension (y) of the a1r-water interface as protein molecules are absorbed. Thus, flexible protein molecules, which can. rapidly reduce y, are considered to give good foamability, whereas tightly ordered globular molecules, which are relatively resistant to surfacedenaturation, give poor foamability. Aqueous solutions of egg white proteins reduce the surface tension of water from 71.8 dyne cm - l to values between 39 and 51.8 dyne cm - l. 27 Their ability to form/stabilize foams, however, differs considerably. In the case of aqueous foam the foam volume decreased in the order egg albumin = ovoglobulins > conalbumin > ovalbumin > lysozyme with the foa)Il stability decreasing in the

Protein Functionality in Bakery Products

97

order ovoglobulins > egg albumin> ovalbumin > conalbumin > lysozyme. 28 Addition of sucrose to the aqueous phase, however, increases the foam volumes and changes the stability, but without significantly altering the order. When proteins are used in combination their performance changes yet .again. Addition of lysozyme-a basic globular protein-to the other egg white proteins (with the exception of ovoglobulins) has a positive effect on both foam volume and foam stability when sucrose is present. In the absence of sugar the effect is zero or negative on both the volume and the stability of the foam. 28 An additional source of complexity in egg white aeration is the presence of metal ions. References have been made 29 - 31 to the benefits of using copper vessels for the preparation of egg white foams and a recent scientific-rather than culinary -study has confirmed the benefits of such equipment and established a plausible explanation. The foams took longer to prepare in copper vessels in comparison with glass bowls, but their resistance to breakdown (on standing) was much improved. Conalbumin binds Cu 2 + and Fe 3 + and, since both metal-protein complexes are more resistant than native conalbumin to various denaturing treatments, the copper-conalbumin complex might be expected to modify excessive surface denaturation and foam collapse. Absorption spectroscopy (Figure 5) provided evidence for a conalbumin-copper complex when egg whites were beaten in a copper bowl or in the presence of copper ions. 32 In summary, the aeration properties of egg white proteins will depend on composition and on the presence of other proteins, metal ions, and carbohydrates (sucrose).

"'u C

Cl

-E

b

~

..0

M > Mmax/3 owing to chains breaking preferentially near their centres. The net result is that the molecular mass distribution of the original sample is narro.wed and the highest molecular masses will always be lost. This is shown in Figure 9. The changes that have been observed for gluten proteins during mixing correspond closely to those predicted by the Bueche theory. Additionally extractable protein, formed by mixing, falls in the glutenin class. A decrease in extractability of protein during 'unmixing' found by Paredes-Lopez and Bushuk 33 suggests that bonds can re-form during relaxation. There is good evidence that it is ,o p; Bueche, J. Appl. Polym. Sci., 1960, 4, IOI. 31 T. L. Smith and N. W. Tschoegl, Rhea/. Acta, 1970, 9, 339. F. MacRitchie, J. Polym. Sci., 1975, Symposium No. 49, p. 85. 33 0. Paredes-Lopez and W. Bushuk, Cereal Chem., 1983, 60, 19.

32

10 5

10 4

where Eis the energy needed to break a chain bond, K is a constant of magnitude approximately equal to the bond vibration frequency, and 28 is approximately the length of the chain bond. F 0 is given by 11 F0 = (constant) ( \

145

10 6

molecular mass Figure 9 Schematic diagram illustrating the effect of shear degradation on the molecular mass

distribution of gluten protein. The hatched area represents material broken down by mixing and the dashed line shows the change in distribution as a result offormation of new lower molecular mass material

+

+

E

E 30 ~ 0

-g

+'

>"

>

Q./

-a

20

C Cl)

C

Vl

d

Cl)

Vl

d

Q)

,_ u

-

Q./

Q./

Cl) ,._ u

10

C

C

~

0

~ 0

0

40

80

40

80

- ¾ protein extracted- ¾ protein extracted Figure IO Effects on mixograph peak development time and loaf volume of additions of 1%

gluten protein fractions to a base flour. Successively extracted fractions are shown from left to right

146

Chemistry and Physics of Baking

the S-S bonds that are broken in mixing. 34 This is consistent with their lower strength compared with peptide bonds of the main chain. One of the consequences of chains breaking near their centres is that the position of S-S bonds in glutenin molecules could be important. Those occurring near the centres of molecules would be more susceptible to scission. Alternatively, glutenin molecules not having S-S bonds near their centres would be expected to confer mixing stability to doughs. This may be relevant to the observation reported by several workers that there are rheologically effective S-H and S-S groups in doughs and that the number of S-S bonds contr~buting to rheological properties is a relatively small fraction of the total number present. Mixing and Baking Quality.-Mixing behaviour can give information for predicting baking quality of flours. It is therefore useful to examine how these two properties are related. Finney and Shogren 35 have reported that flours with mixograph peak development times below a certain value gave inferior baking performance. Above this value, little benefit in terms of baking performance was observed as mixing time _increased. Figure 10 shows the effects of adding fractions of gluten protein on mixograph peak development time and loaf volume in an optimized baking test. The gluten fractions were prepared by differential solubility in dilute hydrochloric acid and additions of 1% by flour weight were made to a base flour of 11.2% protein. The results show that mixing requirements increase progressively with successively extracted fractions. However, the loaf volume pattern is different in that an intermediate glutenin fraction has the greatest improving effect. It has not, however, been confirmed whether the same pattern would be maintained if a higher mixing intensity were used in the baking test.

11 Protein-Lipid and Protein-Carbohydrate Interactions in Flour-Water Mixtures ByW. Bushuk DEPARTMENT OF PLANT SCIENCE, UNIVERSITY OF MANITOBA, WINNIPEG, CANADA R3T 2N2

1 Introduction

All of the major constituents of flour---carbohydrates, lipids, and proteins---contribute to its unique ability to form a dough which can be leavened and baked into a loaf of bread of large volume and appropriate texture. 1 - 3 Differences in breadmaking potential among flours milled from grain of different wheat varieties result mainly from intervarietal differences in the proteins 4 •5 and, to a lesser degree, from intervarietal differences in the lipids. 6 •7 Both quantitative and qualitative aspects of these constituents are involved in breadmaking quality differences. The scientific literature contains numerous publications on the structure-functionality relationship of individual constituents of flour or specific components of those constituents. Although it has been alluded to ii} many publications that interactions of different constituents contribute to dough properties, it is only recently that those interactions have been the subject of specific investigations. Complete understanding of the structure and behaviour of doughs and other flour-water mixtures (e.g. cake batters, slurries as in starch and gluten separation, etc.) will require detailed information on the formation and properties of inter-constituent complexes. This article reviews recent advances in research on interactions of proteins with flour lipids and with carbohydrates in the context of their potential contribution to technological properties of doughs. The emphasis will be on specific interactions; the functionality of complexes resulting from the interactions will receive only cursory comment. 2 Protein-Lipid Interactions

It is common knowledge that addition of water to flour induces gluten proteins to 'bind' certain flour lipids. 8 If the flour-water mixture is subsequently mixed (e.g. as in dough development), lipid binding is modified, depending on the nature of the 1 2 3

4 5 6

34 35

G. Danno and R. C. Hoseney, Cereal Chem., 1982, 59, 196. K. F. Finney and M. D. Shogren, Baker's Digest, 1972, 46 (2), 32, 38, 77.

7 8

W. Bushuk, Cereal Foods World, 1984, 29,508. B. D'Appolonia and D.R. Shelton, Cereal Foods World, 1984, 29, 508. Y. Pomeranz, Cereal Foods World, 1984, 29,508. K. F. Finney and M. Barmore, Cereal Chem., 1948, 25,291. R. A. Orth, R. J. Baker, and W. Bushuk, Can. J. Plant Sci., 1972, 52, 139. 0. K. Chung, Y. Pomeranz, and K. F. Finney, Cereal Chem., 1982, 59, 14. U. Zawistowska, F. Bekes, and W. Bushuk, Cereal Chem., 1984, 61, 527. 0. K. Chung and Y. Pomeranz, Baker's Digest, 1981, 55 (5), 38.

147

Chemistry and Physics of Baking

148

mixing action and on the atmosphere. 9 The first report on the specific binding of lipids by gluten proteins was published more than 40 years ago. 10 In 1947, Olcott and Mecham 11 proposed the name 'lipoglutenin' for a highly stable lipid-protein complex that they isolated from gluten. Most early research on lipid-protein interactions in flour-water mixtures focused on the lipid component; several models have been proposed for that complex. 12 Until recently, little attention has been paid to the lipid-protein complex from the point of view of the protein moiety. A possible reason for this apparent deficiency could be the fac.t that much of the research on flour protein has been carried out on totally or partially defatted flour. There is also a serious lack of consistency in the published data on the lipids associated with specific flour proteins. It is now known that the amount and the nature of the lipids associated with various proteins of gluten are extremely sensitive to the preparative procedure used. 13

Gliadin-Lipid Complexes.-Work in our laboratory on the gliadin fraction prepared from untreated flour showed that it contained 0.6% carbohydrate and 9.4% unidentified material, presumed to be lipid. 14 Subsequently, we confirmed that the unidentified constituent was indeed lipid and also that the carbohydrate, mostly galactose, was a component of that lipid and not a part of a simple carbohydrate. 15 The lipid associated with the gliadin fraction turned out to be

Interactions in Flour-Water Mixtures

149

mostly digalactosyldiglyceride. Gliadin fractions prepared without the use of ethanol also contained the glycolipid. 13 On the basis of this evidence we have concluded that the presence of lipid in the gliadin fraction was not an artifact of the use of ethanol in the method of preparation. When the gliadin fraction from untreated flour was fractionated by gel-filtration chromatography cin Sephadex G-200 (Figure 1), both the lipid and the galactose co-eluted with fraction I protein in the void volume of the column. An analogous fraction of gliadin preparations from totally defatted flour was considerably smaller and, as expected, contained no lipid or carbohydrate. The ratio of fraction I to the sum of fractions II and III (the two included fractions from Sephadex G-200) increased with increasing lipid content in the original gliadin preparation 15 (Figure 2). We have recently demonstrated that fraction I is an aggregate comprising mostly the high molecular weight components (65-80 kD) of gliadin and substantial amounts of low molecular weight protein. 13 A major part of the aggregation is reversible and is mediated by the glycolipid. 16 On the basis of this evidence, we have postulated that flour galactolipids are functionally active in doughs through their involvement in the aggregation of the high molecular weight gliadin components.

F

1.4

l=l

L2

G

1.2

0

(X)

~ 0

~ 0.8

40

0.6

E

+

c,,

30 W

"'

::i..

Q)

>

~

0.

0. 2

Q) ~

C:

ca

0 .Q

4



Q)

-

1

Q)

( OJ o-

.s."" -9

~

"O 0

~~

s~

.

~

~~

~

. c-e ""

.;

C

-;,

~

z

..... .,"' r.r..

Q

.. O·B

~

z

~ 0:

+ 2RSH protein thiol

11 H H II 0 0 biurea

z

0·4

'

,...;.-.'\ 0·4

CJ

0:

UJ

\

0·2

>

0

Scheme 1

~

9 10 11

12 13

C. C. Tsen, Cereal Chem., 1963, 40, 638. C. C. Tsen, Cereal Chem., 1964, 41, 22. E. A. Alesch, Food Eng .. 1968, 40, 115. A. J. Shukis, Bakers' Weekly, 1969, 216, 27. E. D. Weak, R. C. Hoseney, P.A. Seib, and M. Biag, Cereal Chem., 1977, 54, 794.

.: /

I

Its action in dough is to form a cohesive, dry dough that can tolerate high water absorption while resulting in a good texture and volume in the loaf. Unlike some other improvers, ADA is not a bleach, does not destroy vitamins in the flour, acts extremely quickly in converting SH groups into SS, thus strengthening the dough, and leaves a harmless and perfectly acceptable residue, biurea. Tsen 9 suggested ADA as an ideal replacement for iodate-a similar fast-acting improver that is less acceptable as a food additive (see below). Alesch 11 indicated the usefulness of ADA as a replacement for iodate if used in a mixture with bromate. Such a combination gave a 20% reduction in mixing time and a better loaf with respect to texture and volume. Shukis 12 advised the careful use of oxidants, particularly ADA, and outlined the characteristics of an 'over oxidized' loaf, namely grey, streaky crumb of poor volume resulting from a tight, inextensible dough. The rheology of doughs improved using ADA has been studied by Weak 13 using the Mixograph and by Frazier 6 using stress relaxation measurements. Both have recognized the fast action of ADA and particularly the rapid breakdown of dough with further mixing in the presence of ADA. This results from ADA progressively moving peak dough development to lower work levels as illustrated in Figure I. Visualization of the complex surface is aided by reference to selected cross-sections. In this example, 6 using soft English flour, development of the dough without oxidant proceeded to a maximum relaxation time of just over 1.2 logs at 150 kJ kg- 1 , followed by breakdown. As little as 0.4 mEq ADA kg- 1 was sufficient to move peak development from 150 to 60 kJ kg - 1 in addition to increasing peak relaxation time. At a fixed, commercial work level of 40 kJ kg- 1 , as used in the Chorleywood Bread Process, this peak shift makes the relative improving effect of ADA very great [in this case from 1.0 to over 1.4 logs at an optimum addition of

1·4

__ .,.,-....__,

0

100

200

300

400 WORK

0 INPUT

./\.

'.

·,

\

0

1·2

----■ .....____,._____

'./·""·'.

0

z

i:

m

\.

0:

0..

0

...,

. -~ ·,. ~.



0

protein disulphide

>
..

:D

0·8

1·8

)>

z

>


38

39 40 41 42

43 44 45

1·2

a.

~

0

1·0

0

R. M. Sandstedt and B. D. Hites, Cereal Chem., 1945, 22, 161. C. C. Tsen, Cereal Chem., 1965, 42, 86. N. I. Proskuryakov and T. L. Auerman, Biokhimiya, 1959, 24,297. D. R. Grant and V. K. Sood, Cereal Chem., 1980, 57, 46. D. W. Lillard, Jr., Diss. Abstr. Int., 1981, B41, 2421. D, W. Lillard, P.A. Seib, and R. C. Hoseney, Cereal Chem., 1982, 59,291. M. Elkassabany, R. C. Hoseney, and P.A. Seib, Cereal Chem., 1980, 57, 85. J. E. Carter and J. Pace, Cereal Chem., 1965, 42, 20 I. T. Kuninori and H. Matsumoto, Cereal Chem., 1964, 41, 39.

0

"'

"'

0 a:

Scheme 5

37

s:

m

0

reductase

0 z

z

(.)

via glutathione

m

r

100

200

300

400 WORK

0

INPUT

100

200

300

400

kJ kg-1

Figure 4 Combined effect ofL-ascorbic acid concentration and mechanical work input on the development of dough structure as determined by compressive stress relaxation measurement. Left: Three-dimensional plot representing the development surface by contour lines joining points of equal relaxation time. Right: Development profile cross-sections taken at three levels of LAA addition ( arrowed on contour plot) (from Ref 6)

46

P. Meredith, Chem. Ind. (London), 1966, 948.

47

M. Elkassabany and R. C. Hoseney, Cereal Chem., 1980, 57, 88. N. Chamberlain, N. J. H. Dodds, and G. A.H. Elton, Chem. Ind. ( London), 1966, 1456.

48

r Chemistry and Physics of Baking

190

Clearly the mechanism of LAA improvement of wheat dough is not yet fully understood, but its complex series of self-limiting stages is of considerable benefit to the baker since this most likely confers the high tolerance to over-treatment that is the outstanding property of LAA-based improvers. LAA is currently permitted up to 200 p.p.m. in the UK for use as a bread improver and is the only improver permitted in wholemeal bread. L-Cysteine Hydrochloride.-The use of the amino-acid L-cysteine as a reducing

agent that could accelerate \he development of wheat dough was first patented in 1962 by Henika and Rodgers. 49 So-called chemical dough development, or activated dough development (ADD), is an alternative to high-speed, mechanical dough development. L-Cysteine is able to break disulphide bonds between gluten proteins by thiol--disulphide interchange, causing dispersion of disulphide-bonded aggregates 5_0 (Scheme 6). SSR 1 SH

I Cl+

CHz

I

H3N-CH

I

+

RSSR

1

----

I Cl +

CHz

I

H3N-CH

I

+

RSH

COOH

COOH

L - cysteine hydrochloride

protein disulphide

L - cysteine disulphide bonded to protein

protein thiol

Scheme 6

Under mechanical development conditions, addition of L-cysteine on its own softens the dough, producing a development curve of lower relaxation time than the untreated flour, but with peak development at a similar work level. Small amounts ofL-cysteine may therefore be used to modify or tone-down the development profile for a mixed oxidant system in mechanical development. However, L-cysteine-based improvers are usually designed to replace the requirement for much of the mechanical work that would otherwise be used to encourage disulphide interchange. Substantial cost advantages are reported as lower-power mixing is used to produce bread doughs. Although mixing energy is only a very small proportion of total (oven) energy consumption, the need for special capital equipment for high-power mixing is avoided. L-Cysteine alone cannot cause dough development and ADD processes also include oxidizing improvers such as bromate and L-ascorbic acid to oxidize the broken sulphur bonds and form new, rheologically effective disulphides. Typical ingredients in the ADD process are 35 p.p.m. L-cysteine hydrochloride (more soluble •than L-cysteine); 50 p.p.m. L-ascorbic acid, and 50 p.p.m. bromate, but

T

AcOon of Oxkkmt, Md Uthe, Impmm

levels as high as 70 p.p.m. L-cysteine HCl may be used for strong flours of the North American types. Over-treatment with L-cysteine should be avoided as this produces a very sticky dough which is difficult to handle. Baking tests have shown that ADD can produce bread of equal quality to that from mechanically developed doughs and the ADD process is especially useful for the manufacture of fruited breads, when excessive mechanical mixing would damage the texture and appearance of the fruit. Pace arn;l Stewart 51 devised a two-stage process where the reducing agent, L-cysteine, was added prior to the oxidant, mixed, allowed to rest, and then mixed again with oxidant present. Such a process presumably overcomes any tendency for the L-cysteine to react directly with the oxidant component of the improver blend. Currently, up to 75 p.p.m. ofL-cysteine hydrochloride (0.5 mEq kg- 1 , see Table I) is allowed in UK bread flours, while 300 p.p.m. is permitted as a dough softener for biscuit baking as an alternative to sodium metabisulphite (see below). The cheaper ingredient, metabisulphite, has been used in the ADD process, 52 but cysteine was found to be generally more effective and is, of course, more acceptable as a food ingredient, being a naturally occurring amino-acid. The use of chemical dough development in conjunction with 'overstrong' North American flours has been reported by Finney 53 and by Kilborne and Tipples 54 who recognized that such slow-developing, strong flours could be more useful for breadmaking if developed using cysteine to carry out thiol--disulphide interchange, rather than rely on the considerable amounts of mechanical work that these particular wheat flours otherwise require. Peroxides.-Acetone peroxide (AP), described by Ferrari, 55 was introduced by Sterwin Chemicals as a flour bleach and improver in 1961. This oxidant, produced by careful oxidation of acetone by hydrogen peroxide, is generally used on a carrier of starch, although it is possible to use the liquid itself to treat wheat flour. The dry starch concentrate contains mainly monomeric acetone peroxide (ca. 90%) with the remainder as the acyclic dimer and some residual hydrogen peroxide. The product is known commercially as 'Ketox' and usually represents an activity equal to 10% of its weight as hydrogen peroxide. Acetone peroxide is a very fast-acting oxidant which possesses the advantage of good tolerance to over-treatment. This oxidant differs from improvers such as ADA in that it is active in the dry flour and completes its action within 24 h of treatment. 56 In keeping with this behaviour, AP has poorer storage properties than ADA. In general, acetone peroxide represents an alternative to ADA and, as such, is not permitted for use in the UK (Table I). Benzoyl peroxide is permitted, however, being a bleaching agent only, up to a level of 50 p.p.m. in all flours except wholemeal.

51 52 53

54

R. G. Henika and N. E. Rodgers, Cereal Chem., 1965, 42, 397. so C. C. Tsen, Baker"s Digest, 1973, 47 (5), 44.

49

191

55

56

J. Pace and B. A. Stewart, Milling, 1966, 146, 317. N. Chamberlain, T. H. Collins, G. A.H. Elton, and R. Lipton, F.M.B.R.A. Report, No. 4, 1967, p. I. K. F. Finney, C. C. Tsen, and M. D. Shogren,'Cerea1 Chem., 1971, 48,540. R. H. Kilborn and K. H. Tipples, Cereal Chem., 1973, 50, 70. C. G. Ferrari, K. Higashiuchi, and J. A. Podliska, Cereal Chem., 1963, 40, 89. J. A. Johnson, D. Miller, and lJ. C. Fryer, Baker's Digest, 1962, 36 (6), 50.

. Chemistry and Physics of Baking

192

The action of AP has been studied by Tsen 10 and found to involve the oxidation of thiol groups in dough (Scheme 7). Although AP will react with flour before a dough is mixed, if its rate of action is compared with ADA in the dough state, it is found to act more slowly than ADA or iodate.

CH 3

CH 3

I

I

H0-0-C-0-,0H + 4RSH -

C=O + 2RSSR + 3H 20

I

I

I

I

In recent years, other workers, notably Wade, 5 8 - 62 have investigated the quantitative aspects of dough softening by sulphite ion. Optimum treatment of wheat flour for biscuit baking can be achieved at ca. 200 p.p.m. (2 mEq kg- 1 ; see Table 1) leading to reduced water and energy requirements during mixing, increased extensibility of the dough, and better shape of the resulting product. In mechanical dough development, sulphite dramatically decreases relaxation times, especially at low work levels (Figure 5). However, in air, increasing work input tends to reverse the effect and leads to recovery of much of the dough structure. This does not happen in the absence of oxygen

CH 3

CH 3

acetone

protein thiol

acetone perox'ide

193

Action of Oxidants and Other Jmprovers

2·0

protein disulphide

1·8

Scheme 7 (/)

1-6

/~ -~-~- ~=

"'

0

4 Biscuit and Pastry Improvers w :;

Sodium Metabisulphite.-When wheat flour is to be used for baking biscuits or

pastry, it is often desirable to produce a less resilient, softer dough rather than a strongly elastic bread dough. This can be achieved by the addition of sulphite ions as sodium metabisulphite (SMS). Early work by Hlynka 57 characterized the effect of sodium hydrogen sulphite (bisulphite) on wheat gluten as via cleavage of disulphide crosslinks between protein molecules. He noted that acetaldehyde can reverse this softening effect by complexing with hydrogen sulphite ions, allowing disulphides to reform. The chemistry of these effects is now known to involve the hydrolysis ofmetabisulphite to bisulphite (hydrogen sulphite) which then forms an equilibrium mixture with sulphite ion. The sulphite ion cleaves disulphide to produce an S-sulpho derivative and a free thiol (see Scheme 8). Acetaldehyde will remove bisulphite irreversibly, leading to a reverse of sulphite formation and sulphitolysis.

(1)

2-

-

HzO + SzOs -

-

2HS03

metabisulphite

(2)

HS0

3

bisulphite

( 3)

so;-+ RSSR sulphite

bisulphite

H+ + sor

---

protein disulphide

sulphite

Rsso;

+

RS-

protein protein thiol s-sulphocysteine derivative Scheme 8

1·2

z 2

I-

< ~ :r/ '·~-:

;:::

60

57

,:/·~-~

1-4

62

P. Wade, Food Trade Rev., 1970, 40, 34. P. Wade, J. Sci. Food Agric., 1972, 23,333. P. Wade, Food Trade Rev., 1971, 41, 19. B. H. Thewlis and P. Wade, J. Sci. Food Agric., 1974, 25, 99. P. Wade, Baking Ind. J., 1970, 2, 24.

194

Chemistry and Physics of Baking

Action of Oxidants and Other Improvers

195

5 Treatment of Cake Flour Chlorine Gas.-The use of chlorine gas as a flour improver was patented by Beans in 1879. Originally, chlorine was intended as a bleaching agent for bread baking, to be used at ca. 200-400 p.p.m., but quantitative handling of the gas was barely possible until about 30 years later, when it became commonly used. Today chlorine is not permitted as an improver for bread flour, but is used extensively in the UK for the preparation of flour for use in the High-Ratio Cake Process. This process requires that the batter can hold a high ratio of sugar-toflour (often 130% sugar on'~ flour weight basis), only possible following treatment of the flour with chlorine. This treatment is carried out at the mill by the 'Nuchlor' process at levels betwee~900 and 2000 p.p.m. chlorine (Table 1) aiming for a flour pH of 4.8-5.2 (exact chlorine requirements depending on protein and moisture content of the flour). During this process only half the chlorine reacts with the flour solid's, the remainder reacting with moisture to form hypochlorite and chloride. Estimates of the distribution of chlorine between carbohydrate, protein, and lipid vary, but there is agreement that the majority of the chlorine absorbed by the flour reacts with the lipids and proteins, with 5% or less reacting with the carbohydrate fraction. 63 Nevertheless, the principal action of chlorine on wheat flour is believed to involve changes in the starch component of the flour. Chamberlain 63 noted in the early sixties that chlorine treatment led to increasc:d swelling of starch granules (up to 30% larger than untreated) and more easily dispersed gluten. Both these changes result in higher water absorption, which is desirable for the High-Ratio Process. The effect on wheat starch is similar to that of dry heat treatment (100--140 °C) which forms the basis of a patent by Lyons & Co. for producing cake flour as an alternative to chlorine treatment in those cases where chlorine is not permitted. 64 Many workers have studied the effects of chlorine on wheat flour. Tsen 65 found that improvement in cake baking quality was achieved at treatment levels up to 2000 p.p.m., leading to higher water absorption and protein dispersion. Above 2000 p.p.m. treatment, he detected a decrease in quality associated with extensive depolymerization of the starch molecules and enormous swelling of the starch granules. When water is limiting, the action of high levels of chlorine can depolymerize starch without the appearance of reducing sugars, owing to formation of a second ring structure 66 (1-6 anhydro ring: see Scheme 9). Further effects on other flour components have been suggested, notably the effect of chlorine on flour lipids. The reaction of chlorine with wheat lipids is reported by Daniels 67 and other workers have indicated that it may be these oxidation reactions that are responsible for the improved baking quality of chlorinated cake flours. Rees 68 proposed that the effect of chlorine may be in altering the structure of the lipid coat on starch granules, thereby allowing greater

CH 20H (1)

64 65 66

67

68

N. Chamberlain, B.B.l.R.A. Bull., 1962, 6, 160. J. Lyons & Co. Ltd., British Patent, No. I 110 711. C. C. Tsen, K. Kulp, and C. T. Daly, Cereal Chem., 1971, 48,247. R. L. Whistler and R. E. Pyler, Cereal Chem., 1968, 45, 183. N. W.R. Daniels, D. L. Frape, P. W. Russell Eggitt, and J.B. M. Coppock, J. Sci. Food Agric., 1963, 14,883. R. F. Rees, Baking Ind. J., 1971, 4, 27, 28, 30, 31, 35.

H2 0

ClOQ+HCl

( water present)

starch chain (water absent)

-

(2)

(3) c1 2

+

H

H

-c=cunsaturated lipid

-

E

H

H

1 Cl

I Cl

-c-cchlorinated lipid (dichlorostearic has been identified 67 )

Scheme 9

water absorption. Morrison, 69 however, found no effect on the starch lipids and only detected chlorination and hydrolysis of free, non-starch lipid components. Extensive studies of the role of starch-chlorine interaction by Johnson and Hoseney 7 o- 72 have indicated that oxidative changes in the starch molecules are responsible for the improving action of chlorine. They have also made frequent comparisons with the action of heat and storage on wheat flour and have concluded that storage for 8 months at 4 °C gives improvement similar to chlorination. These changes are accelerated by heating and by defatting flour. Rheological studies by various workers have identified distinct changes in the physical properties of cake batters following chlorination of the flour used. Amylograph first-stage peak viscosity is lowered significantly 73 on chlorination and the gelatinization behaviour of the starch is modified. At temperatures in the range 70--90 °C, starch from chlorine-treated flour exhibits a greater tendency to swell, 74 both faster and more extensively than untreated granules, resulting in a greater gel strength to support the cake crumb structure. 75 Very recently, these effects have been attributed to greater amylase exudation from chlorinated starch granules during baking. 76 Changes in structural properties of the starch granule were identified by Allen 77 as being limited to the outer surface where there was a noticeable roughening due to chlorine treatment. His experiments using differential scanning calorimetry 69

10 63

-

Cl 2 +

71

72

73 74

75 76 77

W. R. Morrison, J. Sci. Food Agric., 1978, 29, 365. A. C. Johnson and R. C. Hoseney, Cereal Chem., 1979, 56, 443. A. C. Johnson and R. C. Hoseney, Cereal Chem., 1980, 57, 92. A. C. Johnson, R. C. Hoseney, and K. Ghaisi, Cereal Chem., I 980, 57, 94. A. C. Johnson, Diss. Abstr. Int., 1978, B39, 147. R. C. E. Guy and H. R. Pithawala, J. Food Technol., 1981, 16, 153. P. J. Frazier, F. A. Brimblecombe, and N. W.R. Daniels, Chem. Ind. (London), 1974, 1008. G. W. Telloke, Starch/Stiirke, 1985, 37, 17. J.E. Allen, Diss. Abstr. Int., 1978, B38, 182.

Chemistry and Physics of Baking

196

indicated that less energy was required to gelatinize chlorine-treated starch. (Recent studies on the effects of flour chlorination on starch surface structure are described by Schofield in Chapter 2 of this volume.) The role of starch in improvement of cake flour by chlorine is now well established and further evidence for the involvement of lipids has been produced by Donelson 78 • 79 suggesting that starch involvement is responsible for crumb stickiness but that lipid reactions can cause improvement in cake volume. Active research is still continuing in this area to clarify these findings. The current Flour and Bread Regulations in the UK permit chlorine to be used to treat cake flours up to a level of 2500 p. p.m. 6 Wheat Variety Considerations

Although much of the basic chemistry of improver action is now understood, the dough-mixing behaviour of wheat flours is observed to vary considerably, depending on the individual varieties of wheat present in the flour grist. Consequently, predictions of flour quality and the requirement for improver can only be determined by a careful study of the dough development characteristics and improver response of those pure varieties. Gluten Quality Index.-The method of stress relaxation testing, as described earlier (Section 2) can be applied effectively to the study of the development characteristics of doughs mixed from pure wheat varieties. Figure 6 shows the 2·0

197

Action of Oxidants and Other Improvers

development profiles, in the absence of added improvers, for four UK wheat varieties together with a strong Canadian Western Red Spring wheat. The development characteristics are not merely dependent on protein quantity, but are influenced significantly by the protein 'quality', which is mainly determined genetically for the different varieties. The development profiles classify the five flours into their order of bread baking q,.uality, the strongest flour showing the highest relaxation time. A single numerical value, which effectively classifies each variety in terms of its development potential, can.be measured accurately from the relaxation time at the peak or plateau of each curve. Multiplied by ten (for convenience) the log relaxation time value then forms a quality scale ranging from I to 20, with most wheats falling in the range 9-18. Values of this Gluten Quality Index (GQI) for the flours in Figure 6 are thus: Mardler 9.5, Kador 11, Flanders 12, Avalon 14, CWRS 17-18. Response to Improvers.-Although the mixing characteristics of wheat flours may be usefully studied in the absence of added improver to establish the inherent gluten quality, it is very important,. when fully evaluating the commercial usefulness of particular varieties, to consider the response of the flour to the improvers that will be used in the bakery. This response effectively determines how much of the gluten quality potential (usually revealed at high work levels) can be realized effectively at the CBP work input level, or lower, by improver interaction changing the shape of the development profile. Figure 7 illustrates the response to ADA of two pure varieties which demonstrate similar gluten quality in the absence of improvers (GQI = 14). 2·0

1·8

AVALON

1·8 CWRS

,,,

1·6

,,,

""

0

1·6

.. \:;x/./:-· / : "'· /,~:-:

E"' AVALON 10·0'\,

1-4 UJ

~ FLANDERS 12·5'\,

f-

1·2 KADOR

11·1 'I,_

..

z 2 f-




0 DEPARTMENT

OF

GRAfN

0

~B~

By R. C. Hoseney

"' SCIENCE

AND

INDUSTRY,

KANSAS

STATE

UNIVERSITY, MANHA TT AN, KS 66506, USA

The scientific literature pertaining to doughs, batters, and the products made from them is extensive. This literature can be broken into sections covering (1) dough and batter formation, (2) dough and batter rheology, (3) changes occurring during fermentation, (4) changes occurring during heating, and (5) changes occurring during ageing or storage of the products. It is the latter two sections that I shall discuss here.

0 0

1 Interaction during Heating

There is much less literature pertaining to changes occurring during heating than for any of the other sections. This must be because of the difficulty of studying such events and not because of their lack of importance. Clearly, proofing doughs to a constant size does not assure equal bread volumes, nor does producing batters of equal specific gravities assure equal cake volumes, nor does using a standard cutter ensure the same size biscuit or cookie. Thus, in all baked systems, the differences in the size of the final product are determined by interactions that occur during heating. Therefore it would follow that, if we are to understand these systems, we must understand the interactions occurring during heating. Bread.-A major problem in studying these interactions is that temperature gradients develop when a proofed dough is placed into a hot oven. Therefore, temperature-controlled reactions occur over the entire course of baking. This makes them virtually impossible to follow or study. To solve this problem, we need a method to heat the dough uniformly so that all temperature-controlled reactions occur at once. Baker 1 described an electrical resistance oven for producing crustless bread. Our version of an electrical resistance oven 2 is shown in Figure I. In such an oven, the dough is used as a resistor and thus heats uniformly. By controlling the distance between the electrodes, the size of the dough mass (area in contact with electrode), and the applied voltage, we can control the heating rate. Figure 1 Resistance baking oven: A, electrical connection; B, stainless steel plates; C, carrier gas intake; D, carrier gas outflow *Contributio·· No. 85-379A, Kansas Agricultural Experiment Station. 1 2

J.C. Baker, Cereal Chem., 1939, 16, 513. R. C. Junge and R. C. Hoseney, Cereal Chem., 1981, 58, 408.

216

!

ChRm.,t,y ami Phy,k, af BakU,g

218

As the dough is heated, of course, it increases in size. Because the other dimensions are fixed, that increase in size occurs as an increase in height. Thus, we can monitor the increase in height (volume) as a function of time. By inserting a thermocouple into the dough, we can also follow the increase in temperature. This method produces curves of height or volume increase as a function of time or temperature (Figure 2).

e.....,

1

Campa~nt lnte,ac/Wn,J,umg H,atmg ~d Swrng,

I

219

B

A

5.0

D

(.)

4.0

IJ:

(!)

w

3.0

Cl

2.0

F

J:

w

Figure 3 Schematic drawing of the system used to quantify carbon dioxide evolved during baking: A, nitrogen gas; B, soda lime CO 2 trap; C, gas flow meter; D, electrical resistance oven; E, variable transformer; F, infrared CO 2 detector

Cl)

c--~-=½'7'?'7'f7'2"%C7~-,.-r="""""-:='.·N ••55/:•=--;=yy=:.'-.;':-•

!.

?7>?7

n

!: : ;I

, 242

Chemistry and Physics of Baking

620 600

580 560 540

540

'j'520

520

cj

~

3500

500

~480

480

~

u

Figure 3 Example of a second-order polynomial response surface for cake volume. The axes 'water solubles increasing' and 'tailings increasing' are the coded variables for X 1 and X3 which are defined in the text; X 2 was fixed at the median of its range (Reproduced with permission from Cereal Chem., 1960, 37, 241).

volume, top contour appearance was evaluated subjectively on a 1-10 scale. A central composite design was used giving 25 formulations (16 from the 24 factorial, 8 star points, and the centre point); again the results were fitted by a second-order polynomial. In a later paper Kissel 17 investigated a more commercial formulation containing egg albumen and milk powder a~ additional ingredient variables plus constant low levels of salt and vanilla. This gave seven ingredient variables and therefore six ratio variables. The central composite design thus had 77 treatments (2 6 + 12 + 1). Volume, top contour, and _total internal score were measured and the usual second-order polynomial model was used. It may sometimes be helpful to transform the model to a canonical form (e.g. Wilson and Donelson 18 ). That is to say the response is given by

Optimization of Products and Processes

243

optimized may be easily detected and omitted from the subsequent analysis. The dimensionality of the problem may thus be reduced with consequent reduction in the number of experiments needed. A more recent example of the use of second-order polynomial models applied to cakes is the work of Lee and Hoseney, 19 who optimized the fat-emulsifier system and the gum-egg-water system for a laboratory single-stage cake mix. The two systems were treated separately in individual three-variable experiments. The variables in the fat-emulsifier system were the levels of shortening, propylene glycol monostearate, and a mono-diglyceride emulsifier. Specific gravity, viscosity of mix, volume, top contour, shrinkage, and cake structure were measured. Splitting the design in this way reduces the amount of work and simplifies the interpretation, but implies that there is no interaction between the two systems; i.e. the optimum level for the emulsifier-shortening system is independent of the levels of the gum-egg-water system. An example of the application of the extreme vertices design of McClean and Anderson 13 to cakes is given by Johnson and Zabik. 20 They were interested in the effect of the levels of five egg proteins, in a blend, on the properties of angel food cakes. The procedure for selecting the treatments consisted of initially defining the admissible ranges for proteins in the blend; e.g. the minimum and maximum values for ovomucin were 0 and 2.50%, whereas the corresponding figures for ovalbumin were 30% and 89%. The vertices were constructed by considering all possible treatments with five of the proteins at the maximum and minimum of their range and adjusting the level of the sixth protein (ovalbumin) to make up the blend to 100%. In view of the large allowed range for ovalbumin all 32 formulations calculated in this way were feasible. The design was completed by including the centroids of the three-dimensional faces. This was achieved by averaging the.vertices. The authors used a technique described by Becker 21 to detect components with a linear blending behaviour. This resulted in a model containing terms of the form

as well as linear, second-order, and third-order terms. Bread. A simple fractional 33 factorial design has been used by Henselman et al. 22 to investigate breads fortified with added protein from different sources. The variables were the type of added protein (soya, milk, or fish) and the protein level (0, 4, and 8%). Bread volume, specific volume, structure, and flavour were determined. The results were fitted by a second-order polynomial. Paloheimo et al. 23 investigated the effect of oven variables on bread quality using a half replicate 2 5 - i factorial design with four replicates at the centre point. Variables were baking temperature, the height of the bread in the oven, humidity, air circulation velocity, and the presence or absence of a separate heatable hearth.

Y =flv;)

where the vectors V; correspond to the principal component vectors. One advantage of this approach is that variables with little influence on the property to be

19 20

21 22

17

18

L. T. Kissel, Cereal Chem., 1967, 44, 253. J. T. Wilson and D. H. Donelson, Cereal Chem., 1965, 42, 25.

23

C. C. Lee and R. C. Hoseney, Cereal Chem., 1982, 59, 392. T. M. Johnson and M. E. Zabik, J. Food Sci., 1981, 46, 1226. N. G. Becker, J. R. Statis. Soc., Ser. B, 1968, 30, 349. M. R. Henselman, S. M. Donatoni, and R. G. Henika, J. Food Sci., 1974, 39, 943. M. Paloheimo, Y. Maleki, and S. Kaijaluoto, in 'Thermal Processing and Quality of Foods', ed. P. Zeuthen et al., Elsevier Applied Science, London, 1984.

a

4 244

Chemistry and Physics of Baking

Measured parameters included theoretical energy consumption, volume and specific volume, and thickness and colour of the crust. The results were expressed in ~erms of a linear model only, as it was considered that there were insufficient experimental data to justify determining cross-product (interaction) terms. Biscuits. A good example of the transformation of variables can be found in the studies of Connor and Keagy on vitamin retention during the baking of vitaminfortified biscuits. 4 •24 Here the primary objective of the work was to study the retention of thiamin or folaciµ. The variables were biscuit thickness baking time temperature, rest time, and levels of sodium bicarbonate and ammonium bicarbonate. Rest time was defined as the time from the end of mixing to the beginning of baking. ,· It has been shown that the thermal destruction of the vitamins follows firstorder kinetics, i.e. lnC0 /C = Kt L'

'

'

where C 0 is the initial vitamin concentration and C the concentration at time t. In the above equation K = Aexp(-E./RT) Where E. is the activation energy. Substituting and taking logs gives lnln(C0 /C) = lnA-E./RT+ Int Thus the measured vitamin retention is expressed as In In (C0 /C) and time and temperature are expressed in terms of Int and 1/T respectively. The centre point of the design and an appropriate value for the + 1 level were decided by the investigator. It was then possible to generate the transformation equation. For example in the case of baking time the real value at the centre point was 9.7 min and the+ I level was 12.2 min. The transformation equation takes the general form X=B+Clnt

where Xis the coded level of the time variable which will be used in the design and subsequent analysis.Band Care constants. We thus have 0=B+Cln9.7 and 1 = B + Cln 12.2 Solving these two equations gives the appropriate transformation relationship: X = -9.84 + 4.33 ln t Extrusion. A large proportion of the recent extensive research on extrusion cooking utilizes response surface methods (RSM). Space precludes a detailed examination of specific examples. A very good discussion of the applicability of central co~p~:ite _rotatabl~ designs to extrusion cooking is given by Olkku and Hagqv1st. This paper mcludes examples of the use of RSM to model the extrusion processing of barley, wheat, and rye flours. Another excellent example of the use of RSM to optimize extrusion process variables is given in the case of soya extrusion by Frazier et al. 26 ~; M.A. Connor and K. P. M. Keagy, Cereal Chem., 1981, 58, 239. J. Olkku and A. Hagqvist, J. Food Eng., 1983, 2, 105. P. J. Frazier, A. Crawshaw, N. W. R. Daniels, and P. W. Russell-Eggit, Food Eng., 1983, 2, 79.

26

Optimization of Products and Processes

245

3 Evolutionary and Sequential Methods Evolutionary Operation (EVOP).-Evolutionary operation (generally termed EVOP for convenience) is as much a management philosophy as a statistical technique. It is described in detail in the book by Box and Draper. 27 An optimization exercise carried out in the laboratory or pilot plant will normally be only partly applicable to a full-scale process. The EVOP approach is to make small systematic changes in the full-scale process. The results are then analysed and decision variables are altered in the direction which it is predicted will most improve the process. The changes are sufficiently small to preclude the production of unsaleable product. The statistical designs used are very straightforward so production personnel are directly involved in the exercise, which is a continuous process rather than a project of limited duration. Most frequently 2 2 or 2 3 factorial designs are used with the current process being taken as the centre point. In view of the fact that the 'experiment' is run by process operators it is generally considered that a design involving more than two variables will be too demanding and result in a negative reaction from the personnel involved. The experimental cycle is repeated a number of times, the exact number depending upon the significance of the changes that are being detected. The approach can also be applied to the case where the responses are ranked rather than evaluated quantitatively. 28 •29 Simplex Methods.-A simplex is a figure in n dimensions with n + I vertices. Thus in two dimensions it is a triangle and in three dimensions a tetrahedron. In the basic simplex approach the responses are measured at the vertices and a new simplex is created by elimating the worst vertex by reflection through the centroid of the remaining vertices. A simple example of the early stages of this approach is shown in Figure 4. 30 Here the gel strength of a blend of the three polysaccharides guar, locust bean gum, and carrageenan is being measured. Although there are three ingredients the dimensionality of the problem is reduced to two by giving all formulations tested the same total cost. That is to say the relationship 4X1

+ 2X2 + X 3

= 27

is obeyed where X 1 , X 2 , and X3 are the concentrations of carrageenan, locust bean gum, and guar gum respectively in g 1- 1 . (At the time of writing these relative costs would not be realistic.) It is clear that a quantitative evaluation of the response at the vertices is not required; all that is necessary is that the worst response is identified. It has been suggested that for this reason the simplex approach would be useful in product development since it is far easier for a sensory panel to select one product from a series than quantitatively to evaluate or rank all of them. 31 However, an attempt in our laboratory to hit a target ice cream formulation starting

27 28 29

30

31

G. E. P. Box and N. R. Draper, 'Evolutionary Operation', Wiley, New York, 1969. A. Kramer, Food Technol., 1964, 19 (]), 37. A. Kramer and B. A. Twigg, 'Quality Control for the Food Industry', 3rd Edn. AVI, Connecticut, 1970, Chapter 16. J. Y. Y. Lee, 'Optimisation of a Gel Formulation by Simplex and Response Surface Methods', BSc. Thesis, University of Nottingham, Faculty of Agricultural Science, 1984. J. R. Mitchell, Food Manuf, 1983, 58 (8), 27.

Chemistry and Physics of Baking

246

s

4

FORBIDDEN REGION -VE GUAR CONCENTRATION

3

2+----.------,----,---r-----,-------,--~-, 3

4

5

6 LOCUST

7 BEAN GUM

8 g dm" 3

9 (X2l

10

Figure 4 The early stages of a simplex, run experimentally, to determine an optimum gel blend formulation. Although the blend contains the three polysaccharides, carrageenan, locust bean gum, and guar gum, the use of the constant-cost constraint discussed in the text reduces the dimensionality of the problem to 2. The numbers (i)-(iii) correspond to the progress of the optimization procedure and the arrowed lines show the successive reflections. The next move will trespass into a region of negative guar concentration, which is experimentally inaccessible, so a decision is now required. The numbers at the vertices are F.1.R.A. gel strengths obtained for a 30° paddle turn. The blends have been autoclaved at 121.l °C for 60 min prior to measurement (Adapted from J. Y. Y. Lee, BSc. Thesis, University of Nottingham, 1984)

from random levels of four ingredient variables using this approach was not successful. There have been extensive modifications of the basic simplex idea. These have aimed at speeding up the search and dealing with the problem, illustrated in Figure 4, of what to do when the next vertex requested by the algorithm trespasses into a forbidden region. -Examples include the modified simplex of Nelder and Mead, 32 where expansion and contraction steps are included if the reflected vertex is respectively the best or the worst of the new simplex. Trespassing into forbidden regions is handled by assigning a very unfavourable value of y to that region, when a contraction into the allowed area will automatically follow. in the super modified simplex of Routh et al. 33 a second-order polynomial is fitted to the responses at the



I: Optimization of Products and Processes

centroid, the worst, and the reflected vertices. If the maximum of this curve lies between the worst and reflected vertices, then this is taken at the new simplex vertex; if, however, the curve is concave upward then the vector joining the worst and reflected vertices is extended by an appropriate interval. If a boundary is transgressed then the point where the vector crosses the boundary is taken as the new vertex. Nakai 34 compared a number of different optimization techniques by evaluating the efficiency with which they found the maxima or minima of surfaces described by second-order polynomials containing three or five variables. Overall, a modified super simplex algorithm performed best. This algorithm followed the approach by Routh et al. 33 described above, with the addition of a subprogram of quadratic factorial regression analysis to calculate new vertices at various stages in the process. To improve the efficiency further Nakai et al. 35 added a graphical mapping procedure. It is cla~med that the mapping super simplex procedure has been applied to more than 20 food processing experiments and has markedly improved the efficiency of research. Most experiments on food analysis and proceising have beefl optimized within 25-35 iterative experiments depending on the number of factors. One of the few examples of the application of the simplex method to a baked product is reported by MacDonald and Bly. 36 The objective of the investigation was to determine optimum combination of four emulsifiers (monoglycerides, sorbitan monostearate, polysorbate 60, and glyceryl lactopalmitate) for cake mix shortenings. Initially a designed experiment was carried out in which the four emulsifiers could take one of three levels where each level was a percentage of the total weight of the shortening-emulsifier mixture. This latter total weight was kept constant. Cakes were evaluated for volume (measured by seed displacement), volume index (the sum of cake height measured at three points), symmetry index, grain, and texture. Total score was obtained as sum of grain, volume, and texture. A second-order polynomial [equation (l)] was fitted for each of the responses. Initially an attempt to determine the optimum ingredient combination region was made by constructing a series of two-way tables for each of the responses, i.e. two of the independent variables were kept constant and the response was tabulated for a range of combinations of the other two independent variables. Minimum standards for each response were decided on and areas of unacceptable results were eliminated in turn. Hence a region of acceptable formulations suitable for further examination was attained. (This is a manual method of solving the problem discussed in the Appendix.) Having determined a near optimum formulation in this way, a simplex design was then employed to refine the optimum further. Thus a decision was made as to the scale length for each of the four emulsifiers and the size of the simplex was defined. A five-point simplex was then constructed and five cakes were baked corresponding to the emulsifier levels at the vertices. The worst cake as judged by total score was eliminated and a new formulation predicted. Clearly at this stage each additional experimental run consisted of baking one cake alone since data for four of the simplex points were known. An optimum formulation was arrived at after eight experimental trials. 34

32 33

J. A. Nelder and R. Mead, Comput. J., 1965, 7,308. M. W. Routh, P.A. Swartz, and M. B. Denton, Anal. Chem., 1977, 49-, 1422.

247

35 36

S. Nakai, J. Food Sci., 1982, 47, 144. S. Nakai, K. Koidei, and K. Eugster, J. Food Sci., 1984, 49, I 143. I. A. MacDonald and D. A. Bly, Cereal Chem., 1966, 43, 571.

Chemistry and Physics of Baking

248

Acknowledgement The authors are grateful to Mr. P. K. Skeggs of R.H.M. Research for drawing our attention to some of the literature on constant-weight mixture designs.

Optimization of Products and Processes

It was decided that the product would be satisfactory if the following conditions held: air cell size ;, 3.0 3.0,;; air cell size consistency ,;; 4.5 firmness ;, 3.0 colour ;, 3.0

Appendix: An Example of the Determination of the Optimum Parameters from the Response Surface Model Using Standard.Numerical Algorithms

A number of standard numerieal pa~kages are now available for the minimization or maximization of a non-linear function involving several variables, with constraints (Saguy et a/. 14 ). One such routine which appears particularly well suited to the problem is the NAG (F'oRTRAN) routine E04UAF. 37 In this study this routine is used to optimize a dry mix formulation. The NAG ,routine E04UAF employs a method using sequential augmented Lagrangian multipliers 38 •39 which are successively iterated with the aid of a 'penalty parameter' until a minimum is found. The gradient of the augmented Lagrangian functions is estimated by finite differences during the iterations. Mather 40 used a conventional central composite design to model a commercial dry cake mix. In addition to volume (which was measured as an averaged height), air cell size, air cell size consistency, colour, and firmness were evaluated on a 1-5 scale. The results were expressed in terms of second-order polynomial models, non-significant terms being eliminated. The optimization· approach used was to maximize one response within constraints placed on the coded levels of the ingredient variables and the other responses. The question considered was: what is the formulation for the sponge mix that gives the greatest volume for the same total ingredient cost as the current formulation? The appearance and texture of the sponge must be satisfactory. We now seek to obtain the optimum formulation using the standard package E04UAF. The constraints -1 ,;; X, ,;; I (i

= 1-6)

(Al)

(where X 1 , X 2 , X 3 , X 4 , X 5 , and X 6 are the coded levels of flour, sugar, fat, dextrose, salt, and cornflour respectively) confine the six ingredient variables to a range where the second-order polynomial might be expected to give a reasonable representation of the true response. The centre point of the design (X; = 0), ~ 1_ 6 corresponds to the currently used formulation. Cost will be a linear function of the coded ingredient levels and the condition that the cost of the new optimum formulation is the same as the original formulation is given by 0.375X1

37 36

39

•0

+ 0.88X2 + 0.52X3 + 0.0277X4 + 0.0017X5 + 0.040X6 = 0

(A2)

NAG Fortran Library Manual, 1978. Numerical Algorithms Group, Banbury Road, Oxford, UK. P. E. Gill and W. Murray, 'Numerical Methods for Constrained Minimisation', Academic Press, New York, 1974. W. Murray, in 'Optimisation in Action', ed. L. C. W. Dixon, Academic Press, New York, 1976, Chapter 12. S. Mather, 'Optimisation of a Sponge Mix Formulation', BSc. Thesis, University of Nottingham, Faculty of Agricultural Science, 1985.

249

In terms of the response surface model these conditions can be expressed respectively by equations (A3)-(A6): 0.056

+ 0.27X2 + 0.42X6

-0.34X3 X 6

;,

0

3.0,;; 4.01 -0.32X,2 -0.23X/ -0.32X/ -0.41X3 X 5 0.22

+ 0.72X1 -0.39X2 ;,

0.17

+ 0.22X2 + 0.23X4

(A3)

+ 0.34X4 X 6

,;;

4.5

0

(A4) (A5)

-0.28X1 X 3

;,

0

(A6)

The volume response Yv (measured as an averaged height) is given by

= 1.057 + 0.098X1 + 0.053X3 + 0.113X1 X2 -0.052X2 X 5

Yv

-0.092X,2 -0.067X/ (A7)

The problem is thus defined as maximizing Yv subject to the conditions Al-A6. For the NAG routine E04UAF, equation (Al) for i = 1-6 is supplied as the fixed bounds, (A2) as an equality constraint, equations (A3), (AS), and (A6) as inequality constraints and (A4) as a range constraint. Equation (A7) has to be rearranged slightly since E04UAF minimizes rather than maximizes a function FC; this is simply done by changing positive coefficients for negative (and vice versa): FC = -1.057 - 0.098X1

-

0.053X3

+ 0.092X,2 + 0.067 X/-

0.113X1 X 2

+ 0.052X2 X 5

The user has to supply standard workspace parameters etc., a starting point (for this case a value for the X; of Owas chosen, corresponding to the original formulation), and also has to specify the accuracy in the variables X; to which the solution is required (for this example, to two decimal places). A satisfactory minimization was obtained after less than 2 seconds CPU time on the Cambridge University IBM 3081 Model B. The following values for the parameters were obtained: Yv (volume)

X 1 (flour) X 2 (sugar) x, (fat) X4 (dextrose) X 5 (salt) X 6 (cornflour)

1.112 0.044 = -0.588 0.893 = -0.145 1.000 1.000

where the value for Yv corresponds to an increase in volume (averaged height) of ca. 5.2% as compared with the original formulation.

I

I

-

'1',

250

Chemistry and Physics of Baking

Although we have used a numerical package (E04UAF) originally designed for mainframe computers, it is to be expected that such routines will be available on popular microcomputers in the near future: indeed many NAG routines are already available for the IBM Personal Computer. This availability will no doubt enhance the general application of such procedures to similar optimization problems in the food industry.

19 The Way Ahead: Wheat Breeding for Quality Improvement By P.R. Day, J. Bingham, P. I. Payne, and R. D. Thompson PLANT BREEDING INSTITUTE, MARIS LANE, TRUMPINGTON, CAMBRIDGE CB22LQ, UK

1 Introduction

At a time of cereal surplus in the UK, wheat producers face the challenge of increasing efficiency and reducing costs, and of providing a product finely tuned to the needs of the marketplace and the end user. Quality wheat is used principally by the breadmaking industry, which sets the standards that plant breeders strive to meet. Breadmaking currently uses 3.8 mt, biscuits 0.6 mt, and other 1.0 mt of grain per year. Millers and bakers require a stable, assured supply of inexpensive grain of high specific weight and good milling texture that is low in a-amylase. This grain must have a high content of endosperm proteins that give bread doughs the crucial visco-elastic properties that bakers depend on. At the same time, breeders have to attend to the producers' need for high grain yield, resistance to lodging, and ability to resist the depredation of pests, diseases, and climatic extremes. Until fairly recently, UK breadmaking grists were very dependent on imports of high quality North American hard red wheat. For example, in 1970 home-produced wheat made up less than 30% of the grists. Progress in breeding and growing better UK wheat varieties, 1 and improvements in breadmaking technology, have reduced imports so that in 1984 they made up only 20% of UK grists, and several bakeries are regularly producing white bread from 100% home-grown flour. North American wheat is now used principally to make wholemeal bread, although even the need for this may be reduced by gluten supplementation. In this paper, we review the major problems faced by breeders whose objective is to increase the quality of home-produced grain. We discuss work in progress to find solutions and speculate on the likely shape of the developments to come. 2 The Basis of Quality Milling Texture.-For breadmaking, millers require wheat with a hard endosperm which resists fracturing across the dehydrated protoplasm because of strong adhesion between protein and starch. When cleavage occurs mainly along the lines of the cell walls, large, regular shaped particles are formed which make for a freerunning flour. During milling, the starch granules tend to remain embedded in the protein matrix, and so become damaged, more so than in a soft milling wheat.

1

J. Bingham, J. A. Blackman, and R. A. Newman, 'Wheat. A guide to varieties from the Plant Breeding Institute,' National Seeds Development Organisation, 1985, 80 pp.

251

252

Chemistry and Physics of Baking

Wheat Breeding for Quality Improvement

253

As a result, they absorb more water when mixed into a dough. The higher moisture content of bread made from hard milling wheats lengthens shelf life by retarding staling. Biscuits are made from soft milling wheats because the lower moisture content of the dough reduces baking time, energy inputs, and the moisture gradient between the surface and the interior of the biscuit so that they are less likely to crack. A single gene Ha on chromosome 5D controls the difference between hard and soft milling wheats 2 and is thus a simple character for breeders to manage. The growing interest in the commercial extraction of gluten and starch from flour may favour the production of wheats with large starch grains from which the endosperm protein is more readily separable than is the case for traditional milling wheats. Large starch graiils are desirable since they lessen the problems of starch loss in the effluent of extraction plants. There is also interest in finding industrial uses for this starch co-product so that it may compete with traditional sources of starch.

fibrils of glutenin molecules. 5 The molecules consist of two groups of subunits, high molecular weight (HMW) and low molecular weight (LMW), which are connected together by disulphide bonds. 2 The HMW glutenin subunits are considered to be the key components 2 and their structure is currently being intensively studied. Each end of these molecules consists mainly of hydrophilic amino-acid residues, including cysteine, 6 which give rise to disulphide bonds, whereas the central region is strongly hydrophobic and consists of highly repetitive amino-acid sequences arranged in the form of ~-turns. 7 Well characterized elastic proteins, such as elastin, are also known to have ~-turns. The terminal location of cysteine residues in hydrophilic regions ensures that they occur at the surface of the protein molecules and enables disulphide bond formation with other subunits to form long, linear glutenin molecules. 8 The process of mixing and working causes the long fibrils to become aligned and enables secondary (non-covalent) forces to form between them and other proteins in the dough. A visco-elastic dough has the gas retention properties needed to produce bread with the desired crumb structure.

Endosperm Protein Content.-There is a strong tendency for protein content to be inversely related to grain yielding ability. Protein content is heavily influenced by environment and by cultural treatments. For example, high rates of nitrogen fertilizer application, timed so that nitrogen is available for grain filling, are required at high yield levels to achieve the 11 % protein that may be needed for a breadmaking premium. Even so, intensive selection for protein content has enabled breeders to make some compensation for its decline in higher yielding varieties. It is unlikely that further progress can be made along the same lines to produce UK wheats with protein ofup to 16% as found in the best Canadian grain. The average yield in the wheatlands of Canada is about 2 tonne ha - i and is limited principally by rainfall. The national average yield in the UK in 1984 was more than three times this at 7.6 tonne ha - 1 . However, the much larger area of wheat land in Canada ensures a high total production of high quality grain. While North American production is high, the total cereal production in the UK in 1984 of 26 mt was greater than that for Australia, a country which was exporting wheat to the UK in 1970.

Genetic Controls of Endosperm Proteins.-Genes responsible for the allelic variation found among these proteins have been identified at nine loci on the chromosomes belonging to groups 1 and 6. Wheat has 21 pairs of chromosomes made up of three genomes or sets with seven pairs each. These genomes are derived from three ancestors of cultivated wheat and are designated A, B, and D (Table 1). The seven pairs of chromosomes in each genome are numbered I to 7. The addition of a letter allows identification of each one of the 21 pairs. Thus chromosomes IA, IB, and ID belong to a group called a homoeologous group. An example of the functional homoeology within a group is shown by the allelic variation of a particular set of HMW glutenin subunits which are distinguished by their slow mobility

Table 1 The chromosomes of bread wheat A genome (Triticum urartu)

Possible donors Bgenome (unknown)

Dgenome (Aegilops squarrosa)

IA

1B

ID

2 3 4 5 6

2A 3A 4A 5A 6A

2B 3B 4B 5B 6B

2D 3D 4D 5D 6D

7

7A

7B

7D

Chromosome

Endosperm Protein Quality.-Although protein content is important, there is now

ample evidence that the kinds of protein present are also very important. Wheat endosperm contains a large number of storage protein components, probably in excess of 100. 3 There are two major groups, the gliadins, which consist of a complex mixture of polypeptides without a subunit structure, and the glutenins, which consist of large aggregates with molecular weights of up to 20 x I 06 •4 These proteins are easily extracted in the presence of a reducing agent and appropriate solvents, and are usually characterized by polyacrylamide gel electrophoresis (PAGE). Studies of the physico-chemical properties of these proteins show that the small gliadin molecules enable doughs to extend or stretch. The elasticity, or tendency of a dough to retract after stretching, results from the presence of long

The chromosomes that carry the storage protein genes are underlined.

5 2

3

4

C. N. Law, C. F. Young, J. W. S. Brown, J. W. Snape, and A. J. Worland, in 'Seed Improvement by Nuclear Techniques', International Atomic Agency, Vienna, 1978, pp. 483-502. P. I. Payne, L. M. Holt, E. A. Jackson, and C. N. Law, Philos. Trans. R. Soc. London B, 1984, 304,359. R.R. Huebner and J. S. Wall. Cereal Chem., 1976, 53,258.

6

7 8

J. S. Wall, in 'Recent Advances in the Biochemistry of Cereals', ed. D. L. Laidman and R. G. WynJones, Academic Press, London, 1979, pp.275-311. J.M. Field, P. R. Shewry, and B. J. Millin, Theor. Appl. Genet., 1982, 62, 329. A. S. Tatham, P.R. Shewry, and B. J. Millin, FEBS Lett., 1984, 177, 205. J. A. D. Ewart, J. Sci. Food Agric., 1977, 28, 191.

Chemistry and Physics of Baking

254

in SDS-PAGE (see bracket, in Figure 4). The addition of 2-mercaptoethanol causes dissociation of the glutenin complexes, and the addition of the detergent SDS (sodium dodecyl sulphate) gives them negative charges allowing separation of the HMW subunits by electrophoresis. The variation is controlled by genes at three loci, Glu-Al, Glu-Bl, and Glu-Dl, at least some of which was derived from the Glu-1 loci of the original diploid ancestors. The genes are co-dominant so that the products of each gene are present in the grain endosperm. If one is represented by a null allele its protein subunit bands are absent from the pattern observed after SDS-PAGE. ,,

Wheat Breeding for Qual(ty Improvement

'ii,

1 C

Cl

~

"'

::=: 10-12

Q Q f ------------'------------------------1

j

10-2 10-1

-1--7,...7

"'

18-8

-8

Gene locus Glu-Bl

-13-13-14 -20_15-19-15=1{;

-8

-9 Q

1A

-21

-6

10-12

~

I CJ:1===============l:::i1

Gene locus Glu-Al

..0

Gli-A1

Glu-B1

-2·

·1:

:::,

Tc,_,-----1,

1

18-8

:::,

Glu-A1

1-

255

12

f

. . . 2 . . .3 _ 4

18

g

-s

~

f

-2

9

h

l

Gene locus

k

GI

u-D1

18-8

7 77:J,__1--=JlGlu-D1

.

-12-12-12

1D

=r

-12

Figure 2 SDS-PAGE pattern variants of HMW subunits of glutenin coded at the gene loci Glu-Al, Glu-Bl, Glu-D1. The patterns were obtained from the analysis of some 300

Gli-A2

:::=====================:::!C.,..1-----T-----

-10-10

.QQfg~f

varieties of wheat. The arrows indicate the direction of electrophoresis. On the left hand side of each of the three allelic groups are the HMW subunits of the variety Chinese Spring. This, standard, subunit composition enables the relative positions of variants from each of the three groups to be superimposed. A variety will have any combination of one variant type from each of the three allelic groups. Subunit 21 is derived from rye and occurs in the few wheat varieties which contain the JR chromosome of rye substituted for chromosome I B

6A

:::=========================1-c,_1______

68

::::=========================1-c,..,_____7=_'..,i...0_2_

6D

Figure 1 Chromosomal location of the storage protein genes in wh.eat. The long arms of the

chromosomes are on the left hand side of the Figure. Locations are shown for genes controlling gliadin (Gli) and HMW glutenin subunits (Glu)

The chromosomal location of the nine loci is shown in Figure 1. Figure 2 shows the number of identified alleles for three of them. The pattern for the variety Chinese Spring (shown as a standard) is but one of some 200 different combinations, each of which could occur in a distinct wheat variety. Clearly this is far too many for breeders to test individually. Genetic studies in which the alleles ~t some of these loci are varied, while keeping the others constant, enable compansons to be made of the relative contributions of individual glutenin subunits to baking quality. This is illustrated in Figure 3 where glutenin subunits are ranked in order of their contribution to breadmaking.

The HMW subunit composition of breadmaking. wheats bred at the Plant Breeding Institute (PBI) and now in cultivation (Table 2), shows that each of these varieties only carries the most favourable alleles at one of the important loci on chromosomes IA, !B, or ID. There is thus the potential for improvement by selecting varieties which combine, for example, subunits I, 17 + 18, 5 + I 0. Current breeding work is directed to this end, by using SDS-PAGE to select the parents for crosses. This ensures that the progeny will include recombinants with the desired features of both parents. In subsequent generations a third of a single grain without the embryo can be subject to SDS-PAGE (see Figure 4) while the embryo-containing remainder is sown to produce a seedling. This allows precise single plant selection of the desired endosperm protein subunits to be carried out. Coupled with the reduced generation times from using the method of single-seed descent (SSD), 9 much more rapid advances in the production of wheats with improved quality should be possible. The SSD method accelerates progress to homozygosity in inbreeding species such as wheat. 9 Single seeds are sown in small 9

J. W. Snape and E. Simpson, in 'Efficiency in Plant Breeding', Pudoc, Wageningen, 1984, pp. 82-86.

Chemistry and Physics of Baking

256

18-8

-11r-~!-7 - 18

-6

-8 -9

....,.5

-10

>_£ > .9. ;;,

Q

-2-3_4

--i2 --i2 --12

f!>.9.=_Q>f.



Quality

Figure 3 Allelic variation in the HMW subunits ofglutenin and their relationship to breadmaking quality. A proportion of the subunits shown in Figure 2 were ranked for quality by analysing random lines from crosses by SDS-PAGE and by the SDS sedimentation test

Table 2 Composition of high molecular weight glutenin subunits of bread-making varieties bred at the Plant Breeding Institute Variety Bounty Avalon Brimstone Moulin Mercia*

IA (Glu-Al)

1 (good) I (good) null (poor) null (poor) null (poor)

* Currently in official trials.

Chromosome (locus) !B (Glu-Bl)

7 (poor) 6 + 8 (poor) 6 + 8 (poor) 17 + I 8 (good) 6 + 8 (poor)

257

amounts of soil (ca. 10 g) under very crowded conditions (spacing ca. 2-3 cm in each direction). The seedlings are vemalized and then flower, each usually producing one small ear from which a single seed is taken to sow a next generation. After three such generations, possible in a growth chamber in a 12 month period, plants may be so~n in the field to selec.t for agronomic characters. A disadvantage of the SDS-PAGE procedure in selecting progeny is that it is restricted by the number of samples that can be screened. A realistic upper limit is about 500 samples per worker per week. For this reason it is used as a secondary screen, progeny being selected first for agronomic characters and disease resistance in the field, and general protein quality by the SDS-sedimentation test in the laboratory. Furthermore, if only those progeny which are homozygous for the desired proteins are selected, they need never be analysed again by electrophoresis in later generations.

-1 -q

-8

i > f ' _Q

Wheat Breeding for Quality Improvement

ID (Glu-Dl) 2+ 2+ 2+ 2+ 5+

12 12 12 12 IO

(mediocre) (mediocre) (mediocre) (mediocre) (very good)

Novel Proteins from Diverse Germplasm.-To widen the range of storage-protein variants in commercial wheat varieties, genes coding for novel protein types are being transferred from primitive bread-wheat landraces and from diploid species related to wheat. Bread-wheat landraces from many countries have been screened by SDS-PAGE and several proteins with unique electrophoretic properties have been discovered: 2 Thus, several landraces from Japan contain a chromosome ID-encoded HMW glutenin subunit (called subunit 2.2) whose apparent size is much greater than those of commonly occurring subunits. One landrace, containing subunit 2.2, was crossed to the European spring wheat, Sicco, and the Fl progeny backcrossed to Sicco. Grains containing subunit 2.2 were germinated, grown to maturity, and backcrossed again to Sicco. After five such backcrosses, with selection at each generation for subunit 2.2, a line was developed by selfing and was genetically very similar to Sicco, except for containing genes for subunit 2.2 (Figure 4) . Sicco and several near-isogenic lines, including the one containing subunit 2.2, will be grown in field trials during the season of 1985-86 and subsequent work on the grain will determine the effect of the introduced proteins on breadmaking quality. Any near-isogenic lines whose quality exceeds that of Sicco will be crossed to a winter wheat of the wheat breeder's choice in preparation for its introduction into the bread-quality breeding programme. The wild relatives of wheat, such as the goat grasses (Aegilops species), contain the most extensive range of storage-protein variants. 10 Unfortunately, only the chromosomes from the most closely related Triticum spp. will readily recombine with those from bread wheat. However, genetic methods are available which enable recombination to be induced between diverse species and this has allowed the transfer of the G!l-UJ locus of Aegilops umbellulata into the genome of bread wheat. 11 Further protein gene transfers are in progress at the PBI. A disadvantage of our current approaches to transferring storage-protein genes is that the potential breadmaking quality of the proteins being transferred is not

1

° C. N. Law and P. I. Payne, J. Cereal Sci., 1983, l, 79.

11

C. N. Law, P. I. Payne, A. J. Worland, T. E. Miller, P.A. Harris, J. W. Snape, and S. M. Reader, in

'Cereal Qrain Protein Improvement', International Atomic Energy Agency, Vienna, 1984, pp. 279-300.

Chemistry and Physics of Baking

258

Wheat Breeding for Quality Improvement

259

known. As discussed below, when molecular biologists begin to relate protein sequence with the contribution of the protein to breadmaking quality, it is hoped that, in the future, they may be able to predict which of the many proteins available for transfer are most likely to improve breadmaking quality. 3 Molecular Biology of Endosperm Proteins

5

1

5

6

10

Figure 4 SDS-PAGE of Sicco near-isogenic lines: slots 1-5, lines having an identical protein composition to Sicco; slots 6-10, lines having subunits 2.2 and 12 from Danchi which replace the allelic subunits of Sicco, 5 and 10. The HMW subunits of g/utenin occur within the bracket

Cloning Glutenin Genes.~The physico-chemical basis of the difference in quality among glutenin subunits is still not fully understood. The number and position of cysteine residues in the polypeptides are important. However, the effect of other amino-acids and the role of the tertiary structure (three-dimensional shape) of the protein in contributing to dough properties is a subject of intensive research in several laboratories. Current research at PBI and Rothamsted Experimental Station is directed to obtaining a better understanding of the role of certain glutenin subunits through examination of their structure by isolating the wheat genes that direct their synthesis. This research has several objectives. The DNA sequence provides the aminoacid sequence of the protein; this could lead to a far more objective comparison of different subunits, once the relationship between protein structure and function is better understood. The introduction of genes for specific subunits by transformation is a more long-term goal for breeders; once transformation has been achieved in wheat, it will be possible to adjust the amino-acid sequence of individual subunits by altering the DNA sequences of isolated genes and, by re-introducing the genes into wheat, observe their effects. In Cambridge, Bartels and Thompson 12 collected developing wheat endosperm tissue by passing inflorescences, at the right stage of development, through the rubber rollers of a small 'wringer'. 13 They assumed that the storage protein genes would be actively transcribed and that their mRNA transcripts would be likely to form a sizeable fraction of the poly A+ RNA present in the tissue. Total poly A+ RNA isolated from several wheat lines was prepared and translated in a rabbit reticulocyte lysate system. When the in vitro protein products were run on SDS-PAGE, HMW glutenin bands were observed that corresponded exactly in relative mobilities with bands produced in vivo by the same lines. The next step was to produce, by reverse transcription, a bank of cDNA plasmid clones using the poly A+ RNA as a template. Most of these encoded gliadin-like proteins; however, a small group of clones was found to encode HMW glutenin subunits. One of these clones contained an insert of ca. 1500 base pairs. 14 The coding properties of the insert were demonstrated by hybrid-arrested translation experiments. When the cDNA clone was hybridized with endosperm poly A+ RNA, four bands corresponding to HMW glutenin subunits were absent after in vitro translation. However, when the DNA-RNA hybrid molecules were melted by heating before translation, the glutenin subunits were present. Thus, the arrest of translation as a result of highly specific DNA-RNA hybridization confirms that the DNA encodes the missing products.

12 13

14

D. Bartels and R. D. Thompson, Nuc/. Acids Res., 1983, 11, 2961. M. O'Dell and R. D. Thompson, J. Sci. Food Agric., 1982, 33,419. R. D. Thompson, D. Bartels, N. P. Harberd, and R. B. Flavell, Theor. Appl. Genet., 1983, 67, 87.

I I

260

Chemistry and Physics of Baking

The first 730 nucleotides of the sequence were determined, and one of the possible reading frames revealed a polypeptide fragment with an amino-acid composition similar to that determined for HMW glutenin by workers at Rothamsted. 15 Further studies showed that the clone hybridizes to DNA sequences found only on the long arms of chromosomes IA, IB, and ID. This is where the Glu-1 loci for HMW glutenins are known to be located (Figure 1). The number of copies of this DNA in the wheat genome was estimated at between 5 and 10. This cDNA clone has been used to isolate HMW subunit genes from a gene library and the complete primary sequence of one gene, coding for subunit 12, has been determined. 18 Molecular Markers for Quality.-Work now in progress will soon lead to direct comparisons of the DNA sequences of alleles correlated with 'good' and with 'poor' breadmaking quality. If it is possible to identify 'good' quality genes by their DNA'sequences, then it should be possible to detect their presence in alien diploid species. It is impossible to carry out normal quality tests with grains from these plants bij,.,ause their genetic backgrounds are so unlike those of normal cultivated bread wheats. However, alien diploid species that readily cross with cultivated wheat contain a very wide range of endosperm proteins not found in hexaploid wheats. If those likely to confer quality could be first identified, they could then be transferred to breeding lines by hybridization. At the same time, comparisons of different 'good' genes should reveal those protein structural features that are critically important. This knowledge should then provide a rationale for incorporating the most effective combinations of alleles in a variety. In the future, it will be possible to introduce directed changes in the DNA sequences to test the effect of altering the amino-acid sequ~nces in individual HMW subunits. 16 Wheat Breeding and Transformation.-In 1984 several reports of success in transforming monocotylodenous plants with Agrobacterium tumefaciens plasmid DNA appeared. This system is now well established for dicotylodenous hosts of this plant pathogenic bacterium. 17 The bacterium introduces a piece of DNA (TDNA) from a plasmid (Ti plasmid) that it carries into the host cell where it becomes integrated in the host DNA, either in chromosomes in the nuclei or in the DNA of chloroplasts. The TDNA can be manipulated (disarmed) so that, not only does it have no deleterious effect on the host cell, but it can carry foreign DNA spliced into it. The hitch-hiking DNA, provided that it has suitable controlling sequences, is expressed in the host cell. For example, if it is a gene for kanamycin resistance, it enables the plant cell to make the enzyme neomycin phosphotransferase, so that its growth is not inhibited by this antibiotic. Characters introduced by TDNA appear to be stably inherited and are expressed in seedlings raised from the transformed mother plant. Progress towards achieving transformation in wheat is slow. It has so far proved very difficult to regenerate wheat plants from naked protoplasts, although this 15 16 17

18

J. Forde, B. G. Forde, M. Kreis, R. Frey, P. R. Shewry, and B. J. Millin, FEBS Lett., 1983, 162, 360. R. B. Flavell, P. I. Payne, R. D. Thompson, and C. N. Law, Biotechnol. Genet. Eng. Rev., 1985, 2, 157. M. Bevan, Nucl. Acids Res., 1984, 12, 8711. R. D. Thompson, N. Harberd, and D. Bartels, Nucl. Acids Res., 1985, 13, 6833.

Wheat Breeding for Quality Improvement

261

may well prove to be technically possible one day, as it is now with an increasing number of other plant species. An alternative approach, of microinjecting transforming DNA into developing wheat ovules shortly after fertilization, may provide a solution to this problem. When transformation can be achieved 'in wheat, it may be possible to introduce the genes from alien diploid relatives by this means. As knowledge of the relationship between the amino-acid sequence of endosperm proteins, their tertiary structure, and their function in bread doughs advances, so we may eventually progress to the insertio.n of synthetic genes. However, the skills of plant breeders will still be needed to shape these new materials into the agronomically adapted varieties needed by the farmers, millers, and bakers of the future. Acknowledgements. The progress reported here has depended on help from many colleagues. Special thanks are due to J. A. Blackman, R. B. Flavell, and C. N. Law.

Q, Baking-The Way Ahead 2 The Way Ahead

20 Baking-The Way Ahead By B. Spencer FLOUR MILLING AND BAKING RESEARCH ASSOCIATION, CHORLEYWOOD, HER TS. WO 3 SSH, UK

1 Introduction·

Baking is one of the oldest and most ubiquitous of the traditional skills and one whose basic principles have changed little with time. The baking of unleavened bread has hardly altered from the dawn of civilization and the present day bulk fermentation systems would be readily recognized by a Roman or even an ancient Egyptian baker; only the use of mechanical and electrical aids would seem novel. Of course, there have been many technological changes over the millennia. 1 Alterations in agronomic methods, in wheat breeding, and in milling technology, particularly over the past century, have provided good wheat flour of consistent quality, cheaply and in abundance so that the baker can be provided with flours of all grades for specific baking requirements. Dough mixers have been produced in a myriad of forms. Heating the dough has progressed from the use of open fires and hot stones to ovens, and ovens have varied in size and shape, have been of static, rotatory, or travelling design, and have utilized different forms of energy for heating from dried camel dung to microwave energy generated by nuclear-produced electricity. Considerable ingenuity and imagination have been expended in creating a vast panoply of ethnic, national, regional, and local bread types of every conceivable shape, size, and texture, thickness and colour of crumb; biscuits and cakes too in wide ranges and often incorporating a variety of foreign materials from mammary secretions to the normally thrown-away fibrous portions of plants. Changes in the breadmaking process itself, rather than refinements of practice, have been few. The most fundamental change that occurred in some 6000-7000 years was the introduction by the Egyptians, and most probably by a happy accident, of yeast fermentation which develops the dough to a viscoelastic network which then entraps carbon dioxide bubbles produced by further yeast fermentation to make leavened bread. Some 2500 years later a variation of this change was introduced when mechanica!2 and chemica!3 methods replaced yeast for dough development while retaining yeast to raise the developed dough during proof and baking. High-pressure extrusion cooking, borrowed from plastic technology, which is connected with baked products in the use of some of the same raw materials and in producing products which compete with conventional baked products for market share, should be regarded as an alternative form of food preparation rather than a change in baking method. 1 2 3

263

T. H. Collins, D. G. Hodge, and D. Thacker, British Baker, 1985, 182, 68. C. 0. Swanson and E. B. Working, Cereal Chem., 1.926, 3, 65. R. G. Hennika and S. F. Zenner, Baker's Digest, 1960, 34 (3), 36.

262

The chances of further fundamental changes in the short term are slight and hardly predictable and in this paper it is the refinements of practice that are dealt with and particularly how they may be influenced in the future by external factors. The breadmaking process was developed in an empirical manner and it might be considered that there has been ample time and enough persons involved over the centuries for most possible empirical permutations and combinations of changes in recipe and in process parameters to have been tried already. Much that modern science might have offered if breadmaking had only been invented 50 years ago has already been anticipated. However, although in most cases the principles had been long discovered, many of the adjuncts used in baking today have been introduced during this period, as more efficient products became available through advances in different areas. For example, the use of oxidants, such as azodicarbonamide, which came from the plastics industry, and the use of chemically synthesized emulsifiers. Empiricism will always play a role in development but as we learn more about the composition and characteristics of the raw materials and how they interact during the successive stages of baking so it will be possible to substitute empiricism with rational and theoretical argument based on fundamental knowledge. The introduction of a blend of reducing and oxdizing agents for 'instant dough development' at low mixing speeds 3 is an example of such rationally inspired change. As the previous papers have indicated, much of the fundamental work on baking has been directed towards explaining the empiricisms of the past and in many cases a level of knowledge has been reached which at least allows correction of baking processes when they go wrong and for the optimization and regularization of these processes. While it is a truism to say that there can be no substitute for fundamental work in the 'way ahead', in an applied field such as baking this dogma of faith perhaps needs re-emphasis, not the least for the benefit of those concerned with funding its research. 3 Additives and Attitudes

The acquisition. of fundamental knowledge of baking will lead to the development of new methods of manufacture and better adjuncts for baking than are now available. However, the baker has eventually to sell baked goods to the public, and public perception of what it wants, or can be persuaded to want, will exert considerable influence on the applied aspects of the work of cereal scientists. In Western countries, now that hunger is no longer a problem and gastronomic demands are satiated by a plethora of products, public interest is fifillly focused on the quality, safety, and nutritional aspects of food. Moreover, governments seem anxious to advise, inform, and legislate on these matters. A growing public awareness of the extent to which additives are used in the food industry has raised doubts in the minds of some as to their safety for human consumption. Many hitherto unexplained illnesses are being attributed to additives, often on less than adequate grounds, but in the absence of controlled experiments to show otherwise it is sometimes difficult to allay these fears, irrational as sometimes they may be. The uncertainties associated with present toxicological testing

-----.----·Chemistry and Physics of Baking

264

Baking-The Way Ahead

methods and the difficulties of extending. full testing to the complete range of additives, particularly colours and flavours, are often raised to fuel the argument. The result is considerable public pressure, stronger at the moment in some countries than in others, for a restriction in the number and type of additives used in foods. The introduction oflabelling legislation, for example, has exposed a formidable list of additives on some food products, either as chemical names or as 'E' numbers, both terminologies being equally distasteful to the anti-additive lobby. Major retailers are concerned, 'and since in the UK five retailers command some 60% of food sales, increasing pressures from this sector to restrict the use of additives will be influential. Furtl}ermore, Government regulating agencies, reacting to public pressure, may become much more stringent in permitting the introduction of new additives on the plea of technological need. The trend, already clearly seen. in many Continental countries, is to less rather than more additives and, scientifically irrational as it may seem, a so-called 'natural' additive, such as ascorbic acid, is more acceptable to the public than a 'synthetic' chemical. Attention then will be centred on producing the present range of baked goods using fewer or no additives. Improvements in baking and solutions to baking problems will be directed towards the use of 'natural products' and to alterations of physical process conditions to achieve the desired results in an 'acceptable' way.

• COMA, 'Diet and Cardiovascular Disease', Committee on Medical Aspects of Food Policy, Report of the Panel on Diet in Relation to Cardiovascular Disease, Dept. of Health and Social Security, Report on Health and Social Subjects 28, HMSO, London, 1984. · 5 COMA, 'Nutritional Aspects of Bread and Flour', Committee on Medical Aspects of Food Policy, Report of the Panel on Bread, Flour and Other Cereal Products, Dept. of Health and Social Security, Report on Health and Social Subjects 23, HMSO, London, 1981.

265

Another recommendation that affects baked products is that to lower salt consumption. Government in the UK is taking the attitude that it is the diet, not the food, that can be bad and, for its part, is intoducing nutritional labelling so as to allow an informed public a choice in planning its diet along the recommended lines. Recommendations stemming from Government committees and from semi-official bodies such as the National Advisory Committee on Nutrition Education (NACNE) and the Health Education Council (HEC) are powerful stimuli and are taken up by women's magazines, Sunday paper supplements, and TV food programmes. The messages appear to be successful as witness the decline in butter and sugar consumption. In the bread field, the dietary fibre concept has been eagerly embraced by the public and the media and the effects on the type of bread consumed in most Western countries have been quite remarkable. In the UK brown and wholemeal bread consumption in the home has risen from about 8% to just over 20% at the expense of white bread during a period of overall decline in consumption. 6 Such changes give rise to baking problems of the type that Dr. Galliard has dealt with in Chapter 15. How far the move to higher-extraction breads will go in the UK is a debatable point but other previously 'white bread' countries, such as Holland, have already changed to more than 50% of high-extraction bread. Much of our knowledge of the baking process has been derived from studies of white flours and in pointing the way ahead investigations should increasingly be directed to include brown and wholemeal breads. Some lessons may have to be relearned.

4 Bread Consumption

Not only do bread, cakes, or biscuits have to be visually attractive, o'rganoleptically desirable, and 'safe', but they should also be nutritionally acceptable according to the dictates of the time. Bread consumption in the UK started to fall from late Victorian times. Since 1952 the National Food Survey has shown that consumption of bread in the home has fallen by about 2.3% compound per annum but with some signs in the past couple of years that the rate is slowing down a little. All Western countries have shown similar declines with shallower or steeper curves. The explanation of the decline is multifactorial but one of the reasons has been the poor nutritional image of bread as a starchy fattening food. Modern nutritional thought has now reversed its conception of the role of bread and other complex carbohydrates in the diet and dietary guidelines, starting with the McGovern Report in the USA and repeated in many other countries, have been issued by authorities to encourage bread consumption. In the UK the Committee on the Medical Aspects of Food Policy has made recommendations 4 to lower sugar and fat consumption and to change the polyunsaturated/saturated fat (P/S) ratio in favour of polyunsaturated fat. Greater consumption of bread is urged, particularly brown and wholemeal bread, to substitute for loss of energy due to decreased fat and sugar intake and to increase dietary fibre consumption. 5

----

5 Dietary Fibre, Including Resistant Starch

White bread, however, is still the preferred bread in the UK, USA, France, .and many other countries. The lack of the major minerals and vitamins.in white bread compared with higher-extraction breads is very marginal and of no consequence within the context of the overall diet, particularly after fortification and restoration as legally required in the UK, but the dietary fibre level is lower than that of the other breads. Even so, white bread remains the greatest single commodity contributor to fibre consumption in the average UK diet because of the amount eaten. Dietary fibre is not a single chemical entity and there is as yet no universally applicable definitive assay method. A number of rapid check methods are available such as various modifications of the neutral detergent fibre method of van Soest (e.g. ref. 7). The reference methods for differential fibre analysis, such as the Southgate method, 8 tend to overestimate total fibre owing to failure to remove starch completely. A modification by Englyst and Cummings 9 has largely solved this problem but has raised the question of resistant starch. 10 Resistant starch is starch that has become resistant to digestion by mammalian carbohydrases as a result of physical changes that occur on cooling after baking in 6

7 8

~ 1

MAFF, 'Household Food Consumption and Expenditure: 1983', Annual Report of the National Food Survey Committee, Ministry of Agriculture, Fisheries and Food, HMSO, London, 1985. AACC Committee on Dietary Fiber, Cereal Foods World, 1981, 26, 295. D. A. T. Southgate, 'Determination of Food Carbohydrates', Applied Science Publishers, London, 1976, p. 137. H. N. Englyst and J. H. Cummings, Analyst ( London), 1984, 109, 937. C. S. Berry, Bull. FMBRA, 1984, p. 236.

""d«h' = + k - / - , '

,o,:\;_,-'

,/

0,'11''

266

Chemistry and Physics of Baking

reasonably moist conditions. It appears to be a retrograded form of amylase and, unlike retrograded amylopectin responsible for staling and which is attacked by human a-amylase, it does not further increase with time after first being formed. The workers at the Dunn Nutritional Laboratory have shown that when fed to humans with colostomies the resistant starch can be recovered after passage through the small intestine and it can be similarly recovered from the faeces of rats which have been simultaneously treated with Nebacitin which reduces fermentation in the hind gut. 11 In normal rats, resistant starch increases faecal bulk 10 but, like other soluble dietaF'y fibres, it does not decrease transit time. Resistant starch is thought to increase faecal bulk by a proliferation effect on the bacterial flora of the large bowel, thereby producing a neutral or acid stool with the possibility that bile acid'. absorption and recirculation could be affected. The long-term effect of resistant starch on health and disease prevention remains to be established, as indeed it does for other classes within the dietary fibre complex, but it does fulfil the criteria generally accepted for definition of dietary fibre 12 and its determination should be included in any official analytical method for dietary fibre. While resistant starch contributes to the dietary fibre level of all breads it has most significance in the traditionally preferred white bread where it contributes a greater proportion of the total dietary fibre. Furthermore, there is the possibility that the resistant starch and total dietary fibre level of white bread can be increased either by changes in the baking method to produce more resistant starch in situ or by adding preformed resistant starch, prepared from wheat or high amylase maize, at the dough mixing stage. Meanwhile, the addition of other dietary fibres to white bread, such as bran and pea fibre, continues and bread is a good vehicle for such nutritional additives. The elucidation of the physiological efficacy of the different types of dietary fibre and their effects on health will dictate which may be most advantageously added in the future and the subsequent problems to maintain volume, crumb texture, and colour are offered for solution. 6 Fat and Sugar

Fat labelling in the UK is the subject of consultation between Government and industry with the Government pressing for total fat and the ratio polyunsaturated/(saturated + trans-fatty acids) (P/S) to be declared. Bread is very low in fat and sugar but biscuits and cakes have already been highlighted in the COMA Report 4 as being rich in both these components. Relatively small amounts of biscuits and cakes are eaten by the average person and their contribution to total fat and sugar consumption in the average UK diet is not high (e.g. biscuits account for 4.4% of total fat consumed) but the publicity already given will undoubtedly affect their sales. There is considerable scope for new products in these two areas with lower sugar and fat levels and higher P/S ratios. One approach would be to reduce the fat and sugar levels in traditional recipes while 11

12

I. Bjorck, M. Nyman, B. Pedersen, M. Siljestrom, N.-G. Asp, and B. 0. Eggum, J. Cereal Sci., 1986, 4, I. H. Trowell, D. A. T. Southgate, T. M. S. Wolever, A. E. Leeds, M.A. Gassull, and D. J. A. Jenkins, Lancet, 1976, i, 967.

Baking-The Way Ahead

267

retaining recognizable product characteristics. Use could be made of the fat-sparing properties of 'acceptable' emulsifiers and, provided the structural roles can be retainep, new formula fats with higher P/S ratios and sugar syrups and sweeteners with lower cariogenic and calorific properties could also play a role. Improved (undamental understanding of component interactions in biscuit and cake doughs could be of considerable help in effecting changes of this nature. The alternative of producing totally new ranges of low fat and sugar products might technically be simpler but will involve considerable marketing effort. A combination of both forms of attack may be necessary. 7 Wheat

Plant breeders need to be better informed of the milling and baking qualities required of wheat varieties for specific end uses. The present criteria on which promising varieties are judged, although effective in measuring final performance, are insufficiently discriminatory to indicate reasons for varietal shortcomings. A greater understanding of the role of protein, starch, lipids, and enzymes in the baking process, coupled with methods to measure their contribution to the performance of a specific variety, are needed if wheat breeding is to proceed along more rational lines. Some progress, albeit still empirical, has been made by relating certain glutenin bands seen on electrophoresis to good bread-baking performance. 13 Whereas millers in the USA, Canada, and a few other countries can look forward to consistent wheat of excellent quality, millers in the UK and many other countries with uncertain climates are handicapped with often inadequate and inconsistent quality wheats with which to provide flours for specific baking uses. In the UK there is, or should be, no problem in growing low-protein, soft-milling biscuit wheats except that, as more is known about the influence of flour components in baking of the different types of biscuits, so criteria for selection of varieties will sharpen. Most UK wheat is too low in protein for satisfactory breadmaking and in the first part of this century bread grists were composed mainly of highprotein Canadian wheat, filler wheats from the Argentine and Australia, plus a token 10% of UK-grown wheat. An unforeseen benefit of the introduction of the Chorleywood Bread Process (CBP) was the ability to make bread comparable to that made by bulk fermentation using a flour with I% less protein, i.e. an 11 % protein flour instead of 12% protein. This immediately allowed the use of more of the low-protein UK (or French) wheat and in the 1960s filler wheats disappeared from the grist. North American wheats were still relatively cheap and continued to provide more than 60% of the grist (Figure). When the UK joined the Common Market a levy on third country wheats imported into the UK was gradually introduced and, by the late 1970s, the levy could be termed swingeing. This economic consideration stimulated the industry to reduce North American imports to try to keep the price of flour and baked goods as low as possible consistent with acceptable quality and in 1984 some 70% of UK wheat was used in the average bread grist. This change has been made possible by a number of factors. The hugely increased UK harvests have offered a greater selection of wheat of breadmaking 13

P. I. Payne, L. M. Hall, R. D. Thompson, D. Bartels, N. P. Harberd, P. A. Harris, and C. N. Law, Proceedings of the 6th International Wheat Symposium, ed. S. Sakamoto, Kyoto, Japan, 1983, p. 827.

Chemistry and Physics of Baking

268 80

Baking-The Way Ahead

stage extend the period of loaf expansion before setting takes place. Further knowledge of the mechanism of fungal a-amylase action may lead to more efficient ways to effect the same action. Continued economic pressure, coupled with improvements in the quality of UK wheat and perhaps even further technological advances, will probably result in the virtual disappearance of Canadian wheat from grists in the UK except for specialist uses where the baker is prepared to pay the price differential.

70

UK 60

50 ~L. 01

8 Wheat a-Amylase

40

C a,, 01

~

Third country

30

C a,

~

a, Cl.

20

10

EEC 0 1973/74 75176

269

77178

79/80

81/82

83/84

85/86

Year Figure 1 Composition of the average bread grist in the UK ( data from Dr. N. Chamberlain, personal communication)

quality in sufficient quantity. The CBP has been used to its full potential particularly in plant bakeries where baking conditions can be carefully controlled. Since flours for master bakers, who need more tolerance in the flour, still tend to contain high proportions of Canadian wheat, a figure of 70% of home-grown wheat in the average grist means that many CBP flours are much in excess of this figure and many are already milled from 100% UK wheat. Added gluten is now used extensively instead of Canadian wheat to raise the protein levels of flours milled with high proportions of UK wheat. About twice the amount of added gluten protein is needed as Canadian-wheat native protein to effect the same improvement in baking. Although marked differences in performance can be shown using more sensitive laboratory systems, little difference between glutens can be seen when used as supplements at commercial levels. There remains much room for improvement to increase the efficiency of gluten addition both in the method of production of gluten and in its use in baking. A further technological advance to improve the performance of low-protein flours has been the use of high levels of fungal a-amylase which effectively raise loaf volume. The critical phase of the enzyme activity is in the temperature range 55-57 °c where dough setting, starch gelatinization, and rapid enzyme thermal denaturation are taking place. Softening effects caused by enzyme action at this

As greater reliance is placed on UK wheat for breadmaking the problems stemming from a wet harvest, with resultant incipient sprouting, could prove most troublesome. Even with the present range of UK wheat varieties, where considerable care has been taken in breeding for sprout resistance, such conditions would lead to high a-amylase levels in wheat. Because cereal a-amylase is relatively heat stable, flours milled from such wheat give loaves of poor colour and dark crust with excessive volume leading to collapse of side crusts and with sticky crumb giving slicing problems. On the last occasion a wet harvest occurred, in 1977, the usage of UK wheat in the average bread grist, which had been about 25%, fell in that year to less than 10% (Figure). With 70% of home-grown wheat now in the grist, the potential problem is proportionately much greater and although the industry will rely on importing wheat this will prove expensive, particularly if other EEC countries are similarly affected by bad weather and third country wheat has to be obtained. The long-term solution to the a-amylase problem is the breeding of varieties of wheat of low intrinsic a-amylase which are even less prone to sprouting than present varieties but will still germinate. The possibility oflowering the heat stability of the a-amylase and the use of sprayed-on growth regulators and herbicides to suppress sprouting 14 are other alternatives. Ameliorative steps to reduce the problems experienced in bakeries with higha-amylase flours, 15 such as restriction of dough water and use of acid calcium phosphate, can be taken but their scope is limited. 9 Microwave Baking

Microwave baking offers an alternative solution to the a-amylase problem. 16 Microwaves penetrate the whole of the dough after proof and denature the wheat a-amylase before the enzyme produces the sticky dextrins. The loaf baked by microwave heating alone is crustless and pale and conventional heat has to be used simultaneously to form a crust which gives flavour and attractiveness to the loaf. There are still technological difficulties to overcome, chiefly the production of a robust plastic bread pan which must be heat stable. The original investigation of microwave baking at Chorleywood was originally aimed not only at preventing trouble from wet harvests that occur every 5-10 14 15 16

S. Salmon, A. D. Evers, and T. Fearn, Cereal Res. Comm., 1981, 9, 167. N. Chamberlain, T. H. Collins, and E. E. McDermott, J. Food Technol., 1981, 16, 127. N. Chamberlain, Food Trade Rev., 1973, 43, 8.

270

Chemistry and Physics of Baking

years but also at the possibility of baking a loaf from low-protein UK flours. This latter problem has since been adequately covered by the use of gluten and fungal a-amylase and industry is perhaps more reluctant than it was in the past to invest heavily in microwave ovens which are untried at the commercial level. However, microwave baking might be useful for specialized types Gfbaking or in breadmaking if economic factors change, for example if the country runs out of natural gas, used for heating most ovens, and has to rely on electrical energy. 10 Other Areas

Numerous other problems in J:>aking remain tri be solved or improvements made. Prolongation of the shelf life of bread by preventing staling and mould growth would be advantageous. Yeast with a metabolism that was suppressed at room temperatures could be useful as also would a more robust yeast resistant to freezing and thawing. Improvements in the efficiency of ovens will no doubt follow from more detailed studies of heat and mass transfer during baking. Further control technology, based on a more detailed understanding of the baking process, would enable more consistent products to be made. Advances in machinery design will facilitate the fully automatic production of even the most elaborate flour confectionery. Consumer acceptability of baked products and of the manner of their production, together with economic considerations, will be major factors dictating the direction of advances in baking technology. Fundamental understanding of the baking process is, however, the basis of any such advance irrespective of the direction it may take.

Subject Index

Acetone peroxide, 181 Acoustic emission, 38 Activated dough development (ADD), 108, 190 Active dry yeast (ADY) keeping quality, I 19 leavening activity, 119 rehydration, 119, Additives, 263 Ageing of bread, 51 Ageing of flour, 106 Air incorporation in sponge cake, 93 occlusion of, 134 stabilization in sponge cake, 93 Albumins, 94 Am/low process, 187 a-Amylase, 107 bacterial, 13, 109 cereal, 269 fungal, l09, 268 inactivation, 107 poor baking quality, I 07 retardation of staling, 108,110 supplementation, l08 P-Amylase, 105, I JO Amylodextrins, 12 Amylograph measurements, 195 Amylolytic enzymes, controlled fermentation, 111 Amylose biosynthesis, 28 location in granule, 4

retrogradation, 11 Arabinogalactan, protein complex, 52 Arabinogalactan, structure, 45 Arabinogalactan peptide, structure, 46 Arabinogalactans, 43 Arabinoxylan, protein complex, 52 Arabinoxylan, structure; 45 Arabinoxylans, 43 Ascorbic acid, improver effect of, 159

filled network, I I role of water, 30 state of water, 33 texture, 36 production process, 120 production process control, 121 Baker's yeast, 117 raw materials for production, 120 Baking microwave, 269

the way ahead, 262 Baking performance, 19 Baking powder, 174 Balady bread, I 76 Batter fluidity, 93 Batter specific volume, 91 Benzoyl peroxide, I 9 I Biscuits, 2, 113, 192 Biscuits, vitamin fortified, 244 Biurea, 182 Bleaching agent, 191 Bleaching of carotenoid pigments, 167 Bound water, 32 Bread, I, 171 brown, 170, 199 consumption, 264 dough, 174 doughs, gas cell structure of, 135 flavour, 129 grists, 267 improvers, 80 Breadmaking quality, 19 Bromate, 180 Bromelain, 113 Brown bread, 170, 199 By-product utilization, 112 Cake, 2 Cake crumb, treated as cellular material, 40 Cake flour, 194 Cake flours chlorine treatment, 194 dry heat treatment, 194 Cake mix, optimization of emulsifier combination, 247 Cakes constant weight mixture study, 241 high ratio, 11, 76 madeira type, 76 modelling of formulation, 241 optimization of fat-emulsifier system, 243 optimization of gum-egg-water system, 243 sponge, 87 yellow layer, 87 Canadian Western Red Spring wheat (CWRS), 197 Carbon dioxide, diffusion in dough, 219

Ascorbic acid, stereoisomers, 159

L-ascorbic acid, I 81, 20 I D-ervthro-ascorbic acid, 160 D-lh~eo-ascorbic acid, 159 L-erythro-ascorbic acid, 159 ff L-threo-ascorbic acid, 159 ff Ascorbic acid oxidase, 114, 160 L-Ascorbic acid oxidase, 188 Aspergil/us awamori, 110 Aspergil/us niger, 110 Aspergillus oryzae, 105, I 12 Average maximum extension of molecule, 143 Azodicarbonamide (ADA), 180,201,263 Bacillus subtilis, 105, 113 Baked products complex composites, I I

271

. 7"·····

272 S-Carboxymethylglutathione, 159 Carotenoid pigments, 167 Casein (Us 1 and Us 2 ), 99 P-Casein, 99 K-Casein, 99 Catabolite repression, 125 Catalase, 114 Cellular materials, compressive behaviour of, 40 Chain breakdown, shear degradation mechanism, 144 Chemical dough development, 190 Chlorination alteration of amyloglucosidase attack, 29 modification of surface proteins of granule, 28 use in cake flours, 12, 28 Chlorine, 181, 194 Chlorine, effects on flour constituents, 194 Chlorine dioxide, I 8 I Chorleywood bread process (CBP), 76, 108, 182, 187,201,267,268 CM proteins, 150 Collapse of cake structure, 92, I 02 Committee on the Medical Aspects of Food Policy (COMA),264 Component interactions, 216 Compressed yeast, 118 Compressive stress relaxation, 180 Compudomixer, 179 Conalbumin, 95 Constant rate of work input, 179 Constant weight mixture designs, 240 Conventional baking development stage, 228 expansion and setting, 228 mixing stage, 228 product shaping, 228 Cookie spread effect of protein fraction, 223 factors controlling, 223 Cookies, 109 'Crabtree' effect, 125 Crackers, 109, 113 Crispness, 36 sensory, 38 sound, 38 Crumb stickiness, 196 Crumb structure, 170 Crust, 172 Crust checking, 183 Cubic phase (emulsifiers), 69 Cycloheptaamylose, 13 Cysteine, 164 L-Cysteine hydrochloride, i 08, 181, 190 Cysteinylglycine, 158 DATEM, effects on loaf volume, 213 Dehydroascorbic acid, 209 D-erythro-dehydroascorbic acid, 161 D-threo-dehydroascorbic acid, 162 L-erythro-dehydroascorbic acid, 162 L-threo-dehydroascorbic acid, 162 Die swell, 234 Dietary fibre, 265 Dietary guidelines, 264 Diferulic acid, 47

Chemistry and Physics of Baking Differential scanning calorimeter, 7, 9,170,223, 225 Digalactosyl diacyl glycerols, 69, 70 Digalactosyl diglyceride, 149 'Diox' process, 187 Disulphide bonds, 15 Disulphide content of wheat flours, 156 Domaker process, 187 Dough breakdown, 17, 143 development, 136 molecular description, 17 molecular processes in, 141 liquid phase of, 134 mixing, 17,136 resting, 143 rheology, materials science approach, 33 strength, 19, 23 strengtheners, 78 Doughs, mechanical development of, 151 Dried yeast, 118 Durum wheat, 23 Dynamic optimization problems, 237 Egg white, Raman spectra, 98 Egg white proteins, 95 effect of metal ions, 97 Egg yolk components, 98 Electrical resistance oven, 216 Emulsifiers anti-staling effect, 83 crumb softening effect, 83 DATEM, 78, 79, 80, 82, 83,213 distilled monoglycerides, 78, 83 effect on loaf volume of wholemeal bread, 212 ethoxylated monoglycerides (EMG), 78 lactylated monoglycerides (LMG), 85 legal aspects, 88 liquid crystalline phases in dough, 82 u-monoglyceride, 109 propylene glycol esters (PGME), 85 propylene glycol monostearate (PGMS), 85, 87 role in baked products, 75 stearoyl lactylates (CSL, SSL), 78 succinylated monoglycerides (SMG), 78 use in cake mixes, 85

use in cakes, 85, 136 use in shortenings, 80 Endogenous lipoxygenase, 165 Endosperm protein quality, 252 Endosperm proteins amino-acid sequence of HMW subunits, 260 manipulation: of amino-acid sequence, 259 molecular biology of, 259 Energy consumption in mixing, 190 Entanglement coupling, 142 Enzymes in baking, I 05 Essential fatty acids, 187 Esterases, 114 Ethanol, vaporization of, 222 N-Ethylmaleimide (NEMI), 156, 158 Evolutionary method, 237 Evolutionary operation (EVOP), 245 Expansion, 170 ff Extruded product, nature of, 234 Extrusion, 244

Subject Index Extrusion cooking, I 08, 227 action of shearing elements, 233 development in direct expansion process, 231 die swell, 234 direct process, 229 expansion of extrudate, 233 'half product' process, 229 high pressure, 262 mass temperature profiles, 232 mixing stage (direct expansion), 231 molecular degradation, 233 pressure profiles, 232 shape control of product, 233 Farinograph, 144 Farinograph absorption, 179 Fat consumption, 264 Fat labelling, 266 Fats effect on ambient stored wholemeal, 213 effect on frozen wholemeal, 213 effect on product volume, 75 effects in wholemeal baking, 213 role in baked products, 75 role in biscuits, 76 role in bread, 75 role in cake mixes, 76 role in cakes, 76 role in pastry, 77 role in yeast-raised products, 75 Fatty acids, sodium soaps, 21 Feb-batch process, 125 Ferulic acid, 47 oxidative gelation mechanism, 48 Firming correlation with retrogradation, 226 effect of sucrose recrystallization, 226 Firmness, 36 Flat bread, I 08 Flour defatted, 135 free lipids in, 221 shelf-life, 199 Flour doughs, state of water, 32 Flour lipids, 165 Flour proteins, 14 Flour-water mixtures, 147 Foam, stability of, 134 Foaming ease of, 134 factors controlling, 135 Fortified breiids, 243 Free fatty acid (FFA) levels, 205 Freezing, effect on yeast, 127 Fresh yeast, 118 Freshness of bread, 51 Frozen doughs, 111 Fruit cakes, 109 Fruited breads, 191 Galactolipids, 135, 149 Galactose, 148 P-Galactosidase, 111 Gas cell structure of bread doughs, 135 Gas cells, 170, 171, 173 Ge)atinization

273 effect of plasticizing agents, 7 kinetics, IO melting phase transition, 7 non-equilibrium character, 7 Genetic engineering applied to yeast, 130 Glass transition temperature, 12, 37 Glassy state, plasticization of, 38 Gliadin, 142 functional role, 24 y-gliadin bands, 23 patterns, 23 Gliadin--