Gums and Stabilisers for the Food Industry 10

Gums and Stabilisers for the Food Industry 10

Edited by

Peter A. Williams North East Wales Institute, Wmxham, UK

Glyn 0.Phillips Research Transfer Ltd, Cardifi UK

RSC

The proceedings of the tenth International Gums and Stabilisersfor the Food Industry: The Past, Resent and Future of Food Hydrocolloids conference held at The North East Wales Institute, Wrexham, Wales on 5 - 9 July 1999

Special Publication No. 25 1 ISBN 0-85404-820-0 A catalogue record for this book is available from the British Library Q The Royal Society of Chemistry 2000

All rights reserved. Apartfrom any fair dealing for the purpose of research or private study, or criticism or review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988, this publication may not be reproduced, stored or transmitted, in any form or by any means, without the prior permission in writing of The Royal Society of Chemistry, or in the case of reprographic reproduction only in accordance with the terms of the licences issued by the Copyright Licensing Agency in the UK, or in accordance with the terms of the licences issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries concerning reproduction outside the terms stated here should be sent to The Royal Society of Chemistry at the address printed on this page. Published by The Royal Society of Chemistry, Thomas Graham House, Science Park,Milton Road, Cambridge CB4 O W , UK For further information see our web site at www.rsc.org Printed by MF'G Books Ltd,Bodmin, Cornwall, UK

This volume is the 10th in the series, and from the point of view of readership, sales and reviews, the volumes have been universally well received. They have also been widely quoted in the scientific literature. Taken together, therefore, it is fair to say that they have fidfilled an important need in the field of food hydrocolloids.To support this claim let me quote briefly from some of the numerous positive reviews of Volume 9: Food and Nutrition Press....provides a snapshot of the current state of research in the gums and stabilisers area ...definitely worth adding to your boohhelf J. Sci.Food and m c . . . ...a testimony to the new developments made in thefield ... ...a valuable reference to researchers Food Australia... .. . .the series has proved to be of h t i n g value ...is fiequently cited ....and well thumbed copies can be seen on many shelves Nahrung/Food 43( 1999)H.2...... ..it is a very useful book and recommended for scientists as well asfor all persons connected with development of new products. This last comment is particularly welcome, for the Conferences on which the books are based, have sought to bring together scientists, producers and users of food hydrocolloidstogether. The Wrexham Conferencecelebrated its loth birthday in the July 1999 Conference, on which this book is based, which represents some 20 years of discussions on Gums and Stabilisers for the Food Industry. The fist Conference was held in July 1981 with the Organising Committee having worked together for over a year previously. I have never before used the term “Wrexham Conference”, but on this occasion perhaps you will permit me to use the expression that almost everyone else in the business uses about our Conference. Of course some participants have come and gone. But, by and large, we have had a very loyal membership, with a stream of friends coming to one Conference after another. There are three, apart from myself, who have been ever present from the very first Committee; George Barber ( Treasurer). Peter Williams (Secretary) and Bob Morley, our correspondingmember from the USA. Continuity has nevertheless been provided since the same companies have continued to keep faith with us and ensured that we have senior representatives on our Committee, and of course we have had new companies join us.

Our home over the 20 years has been The North East Wales Institute (NEW for the subject arose naturally out of our research into hydrocolloids. We started on the Cartrefle Campus where the notorious “hospitalitysuites’ were situated! We have now become more fiscally prudent and possibly more respectable ! But I hope that the enjoyment and the welcome are as friendly and as warm. We moved to the new Campus in Wrexham only for the 1997 Conference, due to the change in status of NEW into an exclusively a higher education institution. It was pleasing that the response to that Conference and 10th birthday Conference was even more complimentary than ever before.

iv

Gum and Stabilisersfor the Food Industry I0

In my Introduction to the Proceedings of the first conference, I wrote ‘‘Gums and Stabilisers have not been the most fashionable area to study in recent times” Indeed the purpose of the Conferences has been to focus attention on and inspire new work into this field . As already noted we have sought to encourage a partnership between industrial users, producers and academics working in the field of hydrocolloids. We took a giant step forward in this respect when the FOOD HYDROCOLLOIDS TRUST was formed, which enabled the Conferencesto become a recognised charitable activity, and so provide a vehicle to safeguard the fbture of the Conferences. The papers published in this volume are a testimony to the success of the venture. Between the posters, presentations as Plenary Lectures, Parallel Session talks and Posters there were close on a hundred contributions. The main presentations are published in this book. This does not include the new initiative this year of the “MASTER CLASS’, providing a teach-in for the younger scientists about the production , structure and uses of individual hydrocolloids, which was chaired by a constant friend over the years George Sanderson , who has supplied sense and music to our meetings. We have observed that companies were sending their young new scientists to the Conference to gain an introduction to the field, Thus a structured teach-in on the various hydrocolloids was a great introduction to the field for such individuals. This time I will refrain from previewing the content of this volume and simply reflect on the success of the Conferences without which the books would not have appeared. This book follows in the distinguished tradition of its predecessors and covers the new work in the subject authoritatively. I wish to thank all the contributors and my friends on the Organising Committee for their dedicated service and steadfast support. Glyn 0. Phillips Chairman, Organising Committee and Food Hydrocolloids Trust

Contents

Polysaccharide Characterisation Determination of the distribution of non-esterified galacturonic acid in pectin with endo-polygalacturonase P.J.H. Daas, G.J. W.M. van Alebeek, A.G.J. Voragenand H.A. Schols Multi-angle light scattering estimation of pectin molecular weight and the effect of homogenization L. Wicker, M.Corredig and W.L. Kerr

3

19

Extraction and characterisation of pectin from pomelo fruit peels M.H. Norziah, E.O. Fang and A. Abd Karim

27

Gum Arabic - quality and quantity assured K.A. Karamallah

37

Identification of Gum Arabic using PAGE and IEF S.Motlagh, P. Ravines, Q. Ma and F. Jaksch

53

Relevant structural features of the gum from Enterolobium cyclocarpum G. U o n de Pinto, M. Martinez, 0.Beltrcin, C. Clamens, F. Rincbn and L. Sanabria

59

Preliminary study of the gelling properties of the polysaccharideisolated from the fruit of the Cordia abyssinica M.A.N. Benhura and C. Katayi-Chidewe

69

What is the true amylose content of rice starch? M. Ramesh, J.R. Mitchell, K. Jumel and S.E. Harding

76

The dispersibilityof polysaccharidesin water and water-cadoxan mixtures Q. Wang, P.J. Wood, W. Cui and S.B. Ross-Murphy

82

Polysaccharide Gelation Biopolymer gelation - the structure-property relationship A.H. Clark

91

Vi

Gums and Stabilisersfor the Food Industry 10

Rheological and thermal properties near the sol-gel transition of gellan gum aqueous solutions E. Myoshi and K. Nishinari

111

Comparison of texture analyser and rheometer measurements on carrageenan and pectin gels S. Ndoni, B.U. Marr, H. Neilsen, I.-L. Vedersl and L. Borregaard

129

Syneresis of potassium K-carrageenan gels at different KCl and LBG concentrations D. Dunstan, R. Salvatore, M. Jonsson and M-L. Liao

137

Rheological studies of hydroxypropylated and cross-linked potato starch K. Morikawa and K. Nishinari

148

Gelling mechanisms of non-starch polysaccharides (NSP) from pre-processed wheat bran fraction W. Cui, P.J. Wood and Q. Wang

156

Mixed Biopolymer Systems Phase separation in mixed biopolymer systems L. Lundin, I.T. Norton, T.J. Foster, M.A.K. Williams, A.-M. Hermansson and E. Bergstrom

167

Effect of temperature on the rheological properties of starchkarrageenan mixtures C. Loisel, A. Tecante, P. Cantoni and J.-L. Doublier

181

Interactions between K-carrageenan and P-lactoglobulin in gelling and non-gelling aqueous systems N.E. Hotrum, J. Lucey and H. Singh

188

Casein micelles and their interaction with exo-polysaccharides; turbidity and viscosity C.G. de Kruifand R. Tuinier

196

A description of micellar caseink-carrageenan mixed systems by means of calorimetry and rheology S. Bourriot, C. Gamier and J.-L. Doublier

203

Effect of heat treatment on K-carrageenan gelation in milk A, Tziboula and D.S. Home Solvent structure and the influence of anions on the gelation of K-carrageenan and its synergistic interaction with locust bean gum D. Oakenfull, J. Naden and J. Paterson

21 1

22 1

Vii

Contents

Heterotypic interactions of deacetylated xanthan with a galactomannan of high galactose substitution during synergistic gelation F.M. Goycoolea, M. Milas and M. Rinaudo

229

High Solid Systems Hydrocolloids in low water and high sugar environments J.R. Mitchell

243

Effect of sugars on gelatinisation and rheological properties of sago starch F.B. Ahmad and P.A. Williams

255

The rheological properties and enzymatic digestibility of amylose and amylopectin gels in the presence of maltitol E. Vesterinen, P. Forssell, P. Myllarinen and K Autia

262

In vivo and in vitro annealing of starches

270

S.J.J. Debon and R.F. Tester Solvent structure and gelation of polysaccharides in concentrated solutions of simple sugars D. Oakenfull

277

Effect of sugar concentration on the properties of gellan gum gels G. Sworn and C. Johnson

285

Effect of sucrose on milk protein, LBG and their interactions C. Schorsch, M.G. Jones and I. T. Norton

292

Glass transitions in high sugadbiopolymermixtures - some recent developments S.Kasapis and I.M.A. Al-Marhoobi

303

Proteins and Emulsions Recent advances in protein interactions

317

N.K. Howell Surfactant-proteininteractions at air-water and oil-water interfaces observed by atomic force microscopy V.J. Morris, P.J. Wilde, A.R. Mackie and A.P. Gunning

328

Rheological investigations of transient gel in a depletion-flocculated polydisperse emulsion P. Manoj, A.J. Fillery-Travis, D.J. Hibberd, A.D. Watsonand M.M. Robins

337

Gunk and Stabilisers for the Food Industry 10

Viii

Effect of pH and NaCl on rheological and textural properties of lupin protein emulsions A. Raymundo, J. Empis and I. Sousa Effects of lipid on whey protein gelation S. Ikeda and E.A. Foegeding The influence of different calcium-sequestering salts on the hydration characteristics of rennet casein in a simple model system M.P. Ennis, A. Thornton and D.M. Mulvihill

350

366

313

Recent Developments, Future Trends Commercial requirements and interests: an update P.J. Lillford

387

Genetic engineering as a means to modify polysaccharides G. Tucker

397

Substitution of gelatin in low-fat spread: a rheological characterisation F. Madsen

41 1

Designing galactomannans for the food industry M. Brooks, K. Philp, G. Cooney and L. Horgan

42 1

Hydrolysed and deodorized guar gum including other guar specialty products: functional properties and applications F.M. Ward

429

Mechanical and moisture barrier properties of hydroxypropyl rice starch and hydroxypropyl rice starch-poly(acry1ic acid) graft copolymer film B.M.N. MOM Azemi, I.Z. Arifin, A. R. Mazidah and A. Abd Karim

439

Selectivity of various pectins to heavy metal ions M.T. Kartel, L.A. Kupchik and I.G. Levchenko

452

List of Participants

457

Subject Index

466

Acknowledgements This Tenth Conference owed its success to the diligence and invaluable assistance of the Organising Committee. MEMBERS OF THE ORGANISING COMMITTEE Mr G A Barber (Honorary Treasurer)

Mr P Cowburn

Danisco Ingredients Ltd

Dr T Foster

Unilever Research

Mr D Gregory

D G Associates

Dr I Hodgson (Vice Chairman)

Kelco Alginates

Prof D Howling

Kelloggs

Mr H Hughes (Secretariat)

The North East Wales Institute

Mr D Lloyd

Cerestar UK Ltd

Dr R G Morley

Delphi Consultant Services Inc. USA

Prof J R Mitchell

University of Nottingham

Prof E R Morris

University College Cork, Ireland

Dr V J Moms

Institute of Food Research

Dr P Murphy

National Starch & Chemical Co

Dr J C F Murray

Hercules Ltd

Dr Mei Ong

Nestle Ltd

Ms L Paterson

Du Pont (UK) Ltd

Prof G 0 Phillips (Chairman)

Research Transfer Ltd

Ms V Sharpe

Rhone Poulenc Rorer

Dr R White

Optokem Instruments Ltd

Prof. P A Williams (Secretariat)

The North East Wales Institute

DETERMINATION OF THE DISTRIBUTION OF NON-ESTERIFIED GALACTURONIC ACID IN PECTIN WITH ENDO-POLYGALACTURONASE

P. J. H. Daas, G.J. W.M. van Alebeek, A. G.J. Voragen, andH. A. Schols,

Wageningen University, Department of Food Technology and Nutritional Sciences, Food Chemistry Group, Bomenweg 2, 6703 HD Wageningen. E-mail: [email protected]

1 INTRODUCTION Pectin is defined as a mixture of heteropolysaccharides consisting predominantly of partially methyl-esterified galacturonic acid residues. Pectins are present in relative large proportions in plant tissue, where they are a part of the cell wall, playing a role in cell growth control and e.g. in the defence against invasions of micro-organisms’. Pectins also play an important role in the physical and sensorial properties of k s h fiuit and vegetables (ripeness, texture) and their processing characteristics (canned h i t s and vegetables, purees, juices, etc.)’. On an industrial scale, pectins extracted from agricultural plant materials (by-products) like apple pomace and lemodlime peel are used in food applications for their ability to form gels under specific conditions or to increase the viscosity of liquids. They are also widely applied as stabilisers in acid milk product^^^^. Increased interest for pectins from a nutritional, medical and pharmaceutical point of view originates fkom the (claimed) role of pectin as e.g. dietary fibre, cholesterol lowering compound and anti-tumour agent‘. 1.1 Chemical structure in general

The main building block of pectin is galacturonic acid (GalA)which can be methylesterified at the C-6 position. The DM @ e w e of Methyl esterification) is defined as the moles of methoxyl groups per 100 moles of GalA residues. Pectins are referred to as ’highmethoxyl pectins’ when the DM is 50% or higher, while ‘low- methoxyl pectins’ have a DM value below 30%’. Acetyl groups may occur as substituents at C-2 and/or C-3 of GalA’. Apart fkom GalA, pectins always contain (varying) amounts of other sugar moieties like rhamnose, arabinose and galactose. Native pectins are considered to consist of rather pure homogalactururonanparts (polymers of a-(1+4)-linked GalA residues) in addition to more complex ’subunits’. Examples of such complex segments of pectic molecules are rhamnogalacturonan I and 11 (RG-I, RG-II)5*6.These highly branched parts of pectin are usually referred to as ‘hairy’ or ramified regions. The homogalacturonan regions are often referred to as ‘smooth’ regions (Figure 1)2.The proportion of ‘smooth’ to ‘hairy’ region can vary greatly depending on the type of tissue or its development stage.

4

Gums and Siabilisers for the Food Indusiry I0

Figure 1 Schematic structure of pectin including the homogalacturonan (smooth region: SR) and the rhamnogalacturonan (hairy region: HR) backbone. To understand the many functionality's of pectin and pectin fractions -both in plants and in industrial applications- it is essential to know their chemical structure. Important characteristics are GalA content, neutral sugar (linkage) composition, amount and distribution of methyl esters (and acetyl groups), and molecular weight. In addition, the ratio between smooth and hairy regions, the length and chemical fine structure of the neutral side chains and the distribution of the various subunits over the backbone may influence the pectins properties as well. Many of these characteristics can be determined by conventional chemical methods', however pectin-degrading enzymes in a pure form in combination with modem spectroscopic techniques have opened new avenues in elucidation of their fine

1.2 Enzymes active on pectin Two families of enzymes active towards pectic polysaccharides can be distinguished: homogalacturonan- and rhamnogalacturonan- modifymgldegrading enzymes. Below, a short summary of the enzymes active on homogalacturonan is given. We have also dedicated substantial research towards the purification and characterisation of the rhamnogalacturonan enzymes. For a description of these enzymes the reader is referred to the review of Beldman et al.''. Homo-galacturonan degrading enzymes are classified according to their mode of attack''"2. Many reviews have been published discussing the mode of action and characteristics of these common pectic enzymes and they will therefore only be discussed briefly. Pectin methylesterase (PE) is able to de-esterify high-methoxyl pectin resulting in low-methoxyl pectins and released methanol. PE may act blockwise (plant PE) or in a more random fashion (fungal PE). Polygalacturonase (PG), pectin lyase (PL), and pectate lyase (PAL) all belong to the class of depolymerases. PG cleaves the linkage between two adjacent non-esterified GalA residues in a hydrolytic way, where PL and PAL split this linkage by a mechanism of R-elimination resulting in a 4,5-unsaturated galacturonosyl residue at the non-reducing end of the degradation product. PL needs high-methoxyl pectins to act on, while PAL splits next to non-esterified galacturonosyl moieties. Both endo- and exo-types are known of PG and PAL. The endo-types can split the pectin chain at random. Exo-PGs split off mono- or dimers from the reducing end of the chain, while exo-PALS split off unsaturated dimers from the non-reducing end''. The sites of attack of endo-PG, endo-PAL and PL are indicated in Figure 2. The availability of the different enzymes in a purified and!or cloned form has accelerated the structural characterisationof pectins and specific pectin fractions.

5

PolysaccharideCharacterisation

PE

endo-PG, endo-PAL

PL

Figure 2 Homogalacturonanf e n t ofpectin and points of attack ofpectic enzymes. 2 STRUCTURAL STUDY OF PECTIN WITH ENZYMES 2.1 Structural elucidation using enzymes

The group of Albersheim was the first to use enzymes in structure elucidation of pectic material. By treatment of suspension-cultured sycamore cell walls with endo-PG and endoglucanase they solubilised RG-I and RG-II”. The isolation and the elucidation of the composition of these structural units by chemical methods has been described in many publications that were conveniently reviewed by O’Neill et al.’. De Vries et al. also used enzymes in structure studies of pectins7. After extensive enzymatic degradation of the galacturonan backbone of pectin and fiactionation of the remaining material with sizeexclusion chromatography (SEC), polymeric fiagments were obtained. These nondegraded fragments contained virtually all of the neutral sugars residues found in the parental pectin. From this work, it was concluded that there is an intramolecular distribution in pectin in which the neutral sugars are concentrated in blocks of more highly substituted rhamnogalacturonan regions (“hairy”), separated by (“smooth”) regions containing almost exclusively D-galactosyluronic residues. The discovery of the rhamnogalacturonandegrading enzymes”, greatly advanced the elucidation of structure of the hairy regionsi4.The use of enzymes in establishing the distribution of methyl esters over the homo-galacturonanbackbone of pectin is the topic of the remainder of this paper and extensively discussed below.

2.2 Overview of the enzymatic study of the methyl ester distribution The methyl ester distribution of pectin has been the subject of study for many years.

This distribution is very complex due to the fact that the methyl ester distribution should be revealed on a intramolecular level (within one molecule) and on a intermolecular level (distribution of methyl esters over various pectin molecules within a mi~ture’’”~. Methods reported until 1982 have been reviewed by Taylor”. The majority of the studies described in this report used a ‘chemical’, an enzymatical, or a combined approach. Crude enzyme preparations were almost exclusively used. By choosing conditions that favored the enzyme reaction(s) desired, the problem of side activity was attempted to be avoided as much as possiblei7. The majority of the studies almost exclusively describe the intramolecular methyl ester distribution. In a chronological order, the various approaches used are discussed. In 1982 Tuerena et al. developed a method that allowed for the study of the intramolecular distribution of h e (non-esterified) carboxyl groups in pectin with enzyme^'^"^. The non-esterified carboxyl groups of pectin were glycolated with ethylene oxide and a mixture of pecticenzymes was used to degrade the polymer. The non-

6

Gums and Stabilisersfor the Food Industry I0

degraded glycolated oligomers were separated and the molecular weight distribution was determined by SEC. Computer studies were performed to determine the distributions expected for random methyl esterified samples of varying DM’8*19. The experimental results were in reasonable agreement with the calculated data”. However, because the SEC-technique employed was not able to fully separate the various oligomers produced, the methyl ester distributioncould not be resolved in great detail. The methyl ester distribution of carefully extracted apple pectin was meticulously studied by De Vries et al. with both highly purified pectin lyase and endo-pectate lyase”. Comparison of the SEC patterns of the ‘native’ apple pectin digests with those of enzyme digests of a ‘transesterified’ (an assumed random esterified) apple pectin of similar DM, revealed clear dissimilarities. The ‘native’ and transesterified pectins were also degraded to a different extent. Fractionation of the pectin digests by anion-exchange chromatography enabled the separation and quantification of the oligomers produced according to the number of free carboxyl groups2’. Here, only minor differences were observed, suggesting a similar intramolecular methyl ester distribution”. Ion-exchange fractionation of the undigested ‘native’ apple pectin did reveal the presence of an nonhomogeneous intermolecular methyl ester distribution. Similar results were found for carefully extracted lemon (albedo) pectinI6.Using pectin lyase, de Vries also investigated the methyl ester distribution of commercially extracted lemon, lime, orange peel, and apple pectin2’.Here, data suggesting the occurrence of quite large sequences of methyl esterified GalA in the commercially extracted pectins was obtained. A non-homogeneous intermolecular methyl ester distribution was reported for such pectins as wellz2. Unfortunately, the anion-exchange separation employed was only able to separate the oligomers liberated by their overall negative charge, greatly reducing the information obtained. Kiyohara et al. used purified endo-PG to characterize the methyl-ester distribution of complementary activating pectic polysaccharides extracted from the root of Angelica acutiloba2’. Separation of the extract by anion-exchange chromatography resulted in the isolation of four pure homogeneous fractions (a-d) which were subsequently extensively degraded by the enzyme. Of the four polymeric fiactions purified, one (d) produced much more mono- to octa-GalAs after enzyme action than the others (a-c). The pour degradabilityof the a and b fractions resulted from their much higher methyl ester content. Fraction c, however, had a DM nearly identical to that of the good degradable fraction d. Meticulous analysis and more detailed fractionation demonstrated that the homogalacturonan regions of the polysaccharide d contained much more homogeneous sequences of non-esterified GalA, as opposed to polysaccharide c24.The composition of the oligomeric mixtures was only poorly resolved by the SEC-method employed, again preventing the resolution of the methyl ester distributionin great detail. The methyl ester distribution pattern of commercially extracted lemon and apple pectins of high DM was also investigated by Kravtchenko et a1.2s.Highly purified endoPG was used for the study of the intramolecular distribution of non-esterified GalA. The intermolecular distribution of these pectins was independently determined by both anionexchange and size exclusion chromatographJ6.27. The latter revealed a non-homogeneous intermolecular methyl ester distribution, which mostly resembled that of a mixture of pectic polymers with blocks of non-methyl esterified GalA of various sizes. Highperformance anion-exchange chromatography (HPAEC) of the endo-PG degradation products of the commercial lemon and apple pectins enabled the separation and

Polysaccharide Characrerisation

I

identification of oligomers ranging from DP 1 to 17. Because of the high pH employed during HPAEC separation, the information regarding the methyl ester content of the oligomers produced was lad'. Computer simulation studies revealed that the amount of DP 1-3 oligomers produced after endo-PG degradation of the commercial pectins was much higher than the amount expected for random methyl esterified pectins of identical DM. The amounts of DP 4-17 oligomers found were lower than the amounts expected for random esterified pectins2’. These results again confirmed the non-random methyl-ester distribution of commercially extracted lemon and apple pectin. In general, it was concluded that the commercial lemon and apple pectins investigated contained at least part of the non-esterified carboxylic acids grouped in blocks: sequences of non-esterified GalA residues2’. Unfortunately, Kravtchenko calculated the amount of oligomers of each DP produced assuming a response of the pulsed amperometric detector (PAD) proportional to the amount of HCOH groups present. Particularly for large oligomers, this could result in an underestimation of the exact amount of oligomers detected (Van Alebeek and Schols, unpublished results). Though very exact data was obtained on the size of the oligomers produced, the unavoidable loss of methyl groups during HPAEC analysis at high pH, resulted in the concomitant loss of very important information. 3NEW METHOD FOR THE DETERMINATION OF THE METHYL ESTER DISTRIBUTION

3.1 Detection of methyl esterified galacturonides One important breakthrough of the last years was the development of HPAEC at high pH followed by PAD-detection2’ for the analysis of GalA olig~mers~~*’~*~’. This separation technique has been adapted to elution conditions at lower pH (5-7) to increase the number of oligomers ~eparated’~ or to improve the separation between pectin- and pectate lyase degradation product^'^. This procedure was adapted by us for the examination of the composition of methyl esterified oligomers”. Post-column alkali addition was used to enable PAD-detection. When this method was used to separate an endo-PG digest of a low methoxyl (DM = 30%) pectin at pH 5 , the complex elution pattern of Figure 3 (line a) was obtained. In addition to a series of non-esterified GalA oligomers @P 1,2, 3, and some 4) other compounds were also eluted in a regular fashion. The latter were identified as methyl-esterified oligomers after fiactionation by HPAEC (in absence of post-column detection and alkali addition) and analysis with Matrix Assisted Laser Desorption / Ionisation Time of Flight Mass Spectroscopy (Maldi-Tof MS)’4. The peaks corresponding to the compounds detected are listed in the resolved elution profile of Figure 3 (line b). All peaks but one could be assigned to specific galacturonides’. With the two techniques, methyl-esterified GalA oligomers can be easily detected even when present in complex mixtures. Maldi-TOF-MS can also be used for the identification of the oligomers produced by PL and Unfortunately, quantificationof these components is not yet possible with the MS-technique and forms a topic for future research. With HPAEC, the oligomers can be quantified after determination of the PAD-responsefactors’. This -for the first timeenabled the exact determination of specific structural features related to the methyl-ester distribution of pectin by using enzymes. Recently, techniques have become available that are able to differentiate the various isomers of partially methyl-esterified GalA

8

Gums and Stabilisers f o r the Food Industry 10

3'

4'

5'

72!y

1"1\ 10

15

20

25

30

35

40

40

&-

45

50

55

Tlnu irnln)

I

Figure 3 Separation of an endo-PG digest of a DM 30 pectin with HPAEC at p H 5. Elution profiles before (A) and after integration (B) are shown. The arabic number indicates the DP. The number in superscript denote the number of methyl esters. o l i g o m e ~ s ~(Van ~ , ~ ' Alebeek, unpublished results). In this way, the exact location of the methyl-ester(s) within a GalA oligomer can be determined. This is illustrated in figure 4 in which a (Maldi-TOF MS) post-source decay fragmentation spectrum of (chemically prepared) mono-esterified tri-GalA is shown. By labeling the reducing GalA with '*Owater, fragments originating from the reducing end of the molecule can be identified. From the degradation spectrum it can be concluded that the GalA located at the reducing end of the trimer is e~terified~~. Enzymatically formed oligomers have also been Intensity

40000.

30000.

R-G,1

2c@w

409.7

100

200

300

400

500

M/Z ratio

Figure 4 Post-source decay fragmentation pattern of mono-esterlfied tri-GalA (G31). The GalA at the reducing end of the oligomer is labelled with ' 8 0 . The origin of thefragment is indicated: R for reducing and NR for non-reducing end.

Polysaccharide Characterisation

9

investigated in this way". Both the study of the methyl ester distribution and the study of the composition of the active site of pectin degrading enzymes will greatly benefit from this detailed knowledge. 3.2 Non-esterified regions in pectin For the study of the methyl ester distribution, commercially extracted lemon pectins of varying DMs (17-74%) as well as PE de-esterified (DMs 56-71%), highly esterified (DM 85 and 93%), and random esterified (DMs 30-70%) pectins and polygalacturonic acid were used. Highly esterified pectins were prepared by methanolhlfuric acid treatment, whereas the random esterified pectins were prepared by sodium hydroxide dee~terification~. Endo-PG of Kluyveromyces fragilis was used to extensively degrade the pectic polymers. HF'AEC analysis at pH 5 was employed to identify the components formed. With the exception of the DM 93% pectin digest, non-estenfied mono-, di-, and tri-GalA were always formed. Non-esterified tetra-GalA was never observed in DM > 35% digests and only in small amounts in pectin digests of lower DM. No methyl-esterified oligomers other than the compounds shown in Figure 3 (line b) were found. Upon analysis of pectins that were sequentially de-esterified with tomato PE, we found that large amounts of non-esterified mono-, di-, and tri-GalA were released. The more methyl esters removed with PE, the more non-esterified residues liberated'. Random methyl esterified pectins liberated the lowest amounts of non-esterified GalA, even when the DM was relatively low. In general it could be concluded that, the absolute amount of non-esterified mono-, di-, and tri-GalA residues produced after extended endo-PG degradation of pectin indicates the occurrence of sequences of non-esterified GalA (socalled blocks) in pectin. The more of these blocks present and/or the larger the average size of these blocks, the more non-esterified residues formed. In Figure 5, the HPAEC elution profiles of endo-PG digests of three DM f67% pectins are shown that release different amounts of non-esterified GalA; indicating a clearly different methyl ester

0

10

za

30

4a

Figure 5 Three non-identical pH 5 HPAEC elution patterns ofpectin endo-PG digests of nearly identical DM (- 67%). The elution positions of non-esterified mono-,di-, and tri-GalA are shown. B, blockwise (PE) de-ester$ed; C, commercially extracted.

Gum and Stabilisersfor the Food Industry 10

Figure 6

Percentage of non-esterifed GalA residues liberated by endo-PG versus the DM ofpectin. Random esterified (M93,R70,R32 and PGA; -+-), alkaline de-esterifed (C67,CRS2, CR32; --0--), commercial extracted (C68- C23; --A+,highly esterified (C67,M8S, M93; -V-), and PE deesterified (C77,B71,B63,BS7; +) and C72 (C72,B67,B62,BS6; U) pectins are shown. CR, commercial random de-esterifed; M, additionally esterified; PGA, polygalacturonic acid; R, random esterified.

distribution. For the total amount of non-esterified GalA liberated, expressed as the percentage of the total number of non-esterified GalA residues present in pectin, the term “degree of blockiness” (DB) was introduced’. In Figure 6, the DBs found for pectins of various DM are shown. In this figure, the results of similar series of pectins are connected with lines to better illustrate the information obtained. Pectins having the same DM, but a different DB have a non-identical methyl-ester distribution. The pectin of high DB will have more and/or larger regions of non-esterified residues present. With random methylesterified pectins as a reference, the DB readily reveals to what extent the overall methylester distribution differs from a random one. The DB, however, does not discriminate between the methyl-ester distribution of two pectins (of similar DM) of which one contains a few large and the other more but smaller non-esterified blocks if both release the same amount of non-esterified GalA residues upon enzyme degradation. Information on the average size of the blocks can, however, be obtained from the chromatograms. 3.3 Sue of the non-esterifiedblocks

Even if both the DM and DB were nearly identical, some pectin digests gave dissimilar elution patterns. In Figure 7, the chromatograms of the endo-PG digest of two blockwise de-esterified pectins of DM 56.4 and 57.1% are shown. Both digests contain large amounts of non-esterified mono-, di-, and tri-GalA. Besides the DM, the total amount of GalA residues liberated is also nearly identical; resulting in DBs of 32.7 and 34.4%, respectively. However, from Figure 7, it is clear that the ratio at which the non-esterified mono-, di-, and tri-GalA molecules were produced differed greatly. Varying ratios of mono-, di-, and tri-GalA were also discerned in a lot of the other pectin digests. Inclusion of the small amounts of non-esterified tetra-GalA present in the DM < 35% digests had

Polysaccharide Characterisation

11

Figure 7 HPAECpH 5 elution patterns of 856 and BS 7pectin endo-PG digests.

0

10

20

30

40

50

60

70

80

90

100

Mono-GalA (%)

Figure 8 Ternary graph of the normalized percentages of non-esterijied mono-,di-, and tri-GalA observed in pectin endo-PG digests. Commercial (C69-C30; A), PE deestenjied (B71-BS6; a) and highly esterified A485 @)pectins are shown. The random esterified (R70-R32, PGA; -+) and commercial de-estenjied (C67, CR.52-CR32; --0--) pectin series are connected with lines (see texq.

Gums and Stabilisersfor the Food Industry 10

12

no effect and were therefore ignored3’. When the individual amounts of mono-, di-, and triGalA were expressed as the percentage of the total number of non-esterified molecules released clear differences were observed. These differences can best be viewed when plotted in a ternary graph (Figure 8). In this figure, the data for the random esterified samples (R70-R32 and polygalacturonic acid) are connected with lines to better illustrate the information revealed. Overall, a gradual increase in the relative occurrence of mono-, di-, and tri-GalA can be observed for random esterified pectins of decreasing DM. The higher the methyl ester content, the larger the relative percentage of mono- and di-GalA released by the enzyme. For random esterified pectins, it can be calculated that the average size and occurrence of non-esterified blocks decreases with increasing DM3’. As a result, endo-PG will predominantly have to degrade blocks close to the limit of its action in high DM, random esterified pectins (e.g. R70). For the latter this will result in the release of (relative) a lot of small (di- and mono-)GalA molecules. During the degradation of polygalacturonic acid, endo-PG is not hindered by methyl esters and produces mono-, di-, and tri-GalA in a ratio strictly determined by the mode of action of the enzyme. In Figure 8, it can be observed that this “preferred” ratio greatly differs from the ratio observed in most of the pectin digests. Only during the degradation of DM 30 pectins (C30, CR31, and R32) endo-PG is apparently not hindered by any of the substituents. From the above, it is obvious that the ratio in which non-esterified mono-, di-, and tri-GalA are released after endo-PG action indicates the average size of all non-esterified blocks in pectin. Pectins containing (a whole range 00 different sized blocks will produce the non-esterified components in ratios depending on the pre-dominance of one or more of these block sizes. This effect is illustrated in Figure 9, were the endo-PG degradation of three DM 50 pectins with different methyl ester distributions and different average sized

ar Figure 9 Schematic representation of endo-PG action on three DM 50pectins with different methyl ester distributions.Methyl- and non-esterifed GalA are represented by black and open circles. White and black arrows indicate the first and second linkage split, respectively. The non- and methyl-esterrfied GalA molecules released are indicated with solid and dashed underscores. Here, it is assumed that endo-PG needs 4 adjacent none-esterified GalA residues to act.

Polysaccharide Characterisation

13

blocks is shown. Here, it is assumed that the Kluyveromycesfiagilis enzyme employed at least requires a sequence on 4 non-esterified GalA residues to split the chain9.’*.Though all three pectins in Figure 9 release the exact identical amount of non-esterified GalA residues after enzyme action, the amounts of mono-, di-, and tri-GalA released clearly differ. The smaller the average size of the “endo-PG degradable” blocks, the smaller the GalA molecules produced. 3.4 Distribution of non-esterified blocks

Endo-PG action on pectins of similar DM, DB, and non-esterified mono-, di-, and tri-GalA ratio could still result in non-identical elution profiles. This is illustrated in the chromatograms of the R32 and C30 pectin digests of Figure 10. Here, for example, clear differences in the areas of the peaks corresponding to mono-esterified trimer and monoesterified tetramer can be observed. In the C30 digest higher amounts of these components are present. Because the methyl-esterified GalA oligomers were not available in sufficient quantities, no PAD-response factors could be determined. Therefore, no exact quantification of the oligomers was possible so peak areas were used instead. As a means to express the (relative) amount of all esterified components liberated, the area of all peaks corresponding to methyl-esterified oligomers were summarized and divided by the total area of the non-esterified GalA peaks3*.In Figure 11, the ratios obtained are plotted versus the DM of the corresponding pectins. For all digests, the methyl- and non-esterified peak area ratio was highest for random esterified pectins. The information indicated by the methyl- and non-esterified peak area ratio is again best explained by the results of the random esterified pectin series. As expected, the methyl- vs. non-esterified area ratio will be high for pectins releasing large amounts of methyl-esterified oligomers and small numbers of non-esterified GalAs. For random esterified pectins the occurrence of “endoPG degradable” blocks will increase with decreasing DM whereas, simultaneously, the chance of liberating a methyl esterified oligomer after endo-PG action gradually decreases. As a result, starting from a highly, random esterified pectin, the ratio of methyl- vs. non-

Figure 10 Non-identical pH 5 HPAEC chromatograms of C3O and R32 pectin endo-PG digests. Only the peak corresponding to non- and mono-methyl-esterifed GalA are indicated

Gums and Stabilisers for the Food Industry 10

14

20

18

-

16

Figure 11

0

?I 14 L.

z

t

p

12

f

10

E 2 c

*t 5

08

f

06

04

Ratio of the methyl- versus the non-esterzfted GalApeak areas of pectin endo-PG digests plotted versus the DM. Commercial (C69C30; A), PE de-esterified (B71B56; a) and highly esterijied M85 (0) pectins are shown. The random esterified (R 70-R32,PGA; and alkaline de-esterijied (C67, CR52-CR32; --0--)pectin series are connected with lines as explained in the text.

-+-)

02

00

esterified GalA production will increase up to a specific DM and then gradually decrease. From Figure 11, it can be concluded that this transition occurs at a DM of around 50%. With the HPAEC method employed, esterified oligomers up to DP 10 can be detected’. As a result of this and the composition of the active site of the enzyme, “endoPG degradable” blocks located at distances of more than a certain number of GalA residues (6 with the assumed active site of Figure 9) will not release any detectable methyl-esterified oligomers. Endo-PG digests of pectins with a lot of closely located blocks (so-called clustered blocks), in contrast, will contain large amounts of detectable esterified oligomers. Hence, a high ratio of the total methyl-esterified peak area versus the total non-esterified peak area is indicative for the occurrence of blocks located at distance closer than -6-10 GalA residues3*.These observations are also illustrated in Figure 9. The three DM 50 pectins shown all contain “endo-PG degradable” blocks and release different amounts of methyl-esterified oligomers after enzyme action. Pectin A, releases 2 methylesterified oligomers (a hexamer with 2 methyl esters and an octamer with 4 methyl esters), whereas Pectin B only produces a decamer with 5 methyl esters. Pectin C does not release any esterified oligomers after endo-PG action. The data in Figure 11 demonstrates that all of the commercial lemon pectins, with the exception of the DM 30% and 56 pectins, contain almost no clustered blocks. Random esterified pectins contain considerable amounts of such blocks.

Polysaccharide Characterisation

15

3.5 Combined results With the enzymatic approach developed, three characteristics indicative for the distribution of non-esterified GalA residues in pectin are obtained: i) the degree of blockiness9, ii) the normalized percentages of non-esterified mono-, di-, and tri-GalA liberated-”, and iii) the ratio of the total methyl-esterified peak versus the non-esterified peak area”. These three characteristics all indicate a typical aspect of the non-esterified GalA distribution. By combining these aspects -including the DM- a better and more complete overview of the GalA distribution in pectins can be obtained. Computational techniques commonly employed in the determination of the sequence similarity of DNA and proteins were used to discriminate the various type of distributions found and construct a distance tree. The approach used will be published in detail elsewherem.In Figure 12, the distance tree obtained for the non-esterified sequence similarity data is shown. In the tree, the more closely the GalA distribution resemble one another, the smaller the distance between the pectins. In general, three types of GalA distributions can be discerned: A) random, B) high esterified, and C) blockwise esterified. Group A is almost solely composed of chemically modified pectins. These pectins all have DMs of 30 to 70 and quite high methyl vs. non-methyl esterified peak area ratios (2 0.6), indicative for a random or close to random methyl ester distribution. Group B is solely composed of highly esterified pectins of DMs 2 74. A typical feature of these polymers is the presence of very small non-esterified blocks. Group C pectins have, from left to right, increasing amounts and/or sizes of non-esterified blocks. This group also includes polygalacturonic acid. Preliminairy results indicate that within these groups even some subgroups might be distinguished. Since the distribution of the methyl esters of pectin is very important for its the hctional properties, it is highly conceivable that the three general types of distributions found could be correlated to one or more of the typical applications of pe~tin~~’*~.

Figure 12 Non-rooted tree showing the similarity of the methyl ester distribution in pectin. 7ke letters in parenthesis denote the three groups discerned (see tart).

Gums and Stabilisersfor the Food Industry 10

16

4 CONCLUSIONS It can be concluded that as a result of the development of two techniques capable of detecting methyl-esterified GalA oligomers, great progress has been made in the study of the chemical fine structure of pectin. As a result of this, a vast amount of information has become available with respect to the precise distribution of methyl esters over the homogalacturonan backbone of pectin. In our opinion, the same techniques and some of the approaches developed can also be applied to study the physicochemical and catalytic properties of pectic enzymes resolving their mode of action in much more detail. The methyl ester distribution information obtained is essential for understanding the relation between chemical structure of pectins and their hctional properties. When more even information becomes available about the structure of pectin (e.g. inter- and intramolecular distribution of methyl esters / GalA blocks, ramified regions, and/or acetyl groups) specifically suited for specific applications, new strategies for the extraction and the modification of industrial pectins may lead to newly formulated pectin preparations. Also other (food) applications of pectins may be recognised. The obtained knowledge may also contribute to improve the processing conditions of fruit and vegetables and may lead to a better use of agricultural by-products.

Acknowledgement

The work of Piet J.H. Daas was financially supported by Hercules Incorporated. Brigitte Boxma of the University of Nijmegen (the Netherlands) is gratefully acknowledged for her skilfbl assistance in the construction of the pectin distance tree.

References

1. A. Bacic, P.J. Harris, and B.A. Stone. In: i%e Biochemistry of Plants 14, Carbohydrates. J. Preiss (Ed.), Academic Press, London, 1988,297-371. 2. A.G.J. Voragen, W. Pilnik, J-F. Thibault. M.A.V. Axelos, and C.M.G.C. Renard. In: Food polysaccharides and their applications. A.M. Stephen (Ed.), Marcel Dekker Inc., New York, 1995,287-339. 3. C. Rollin and J. De Vries. In: Food Gels, P. Harris (Ed.), Elsevier, Amsterdam, 1996, 401-434. 4. K.W. Waldron and R.R. Selvendran. In: Food and Cancer Prevention, Chemical and bioZogica1 aspects. K.W. Waldron, I.T.Johnson, and G.R. Fenwick (Eds), Royal SOC. Chem. 1993,307-326. 5. M. O'Neill, P. Albersheim, and A.G. Darvill. In: Methods in Plant Biochemistry 2, Carbohydrates. P.M. Dey (Ed.), Academic, London, 1990,415-441. 6. P. Albersheim, A.G. Darvill, M.A. OWeill, H.A. Schols, and A.G.J. Voragen. In: Progress in Biotechnology 14: Pectins and pectinases. J. Visser and A.G.J. Voragen (Eds), Elsevier, Amsterdam, 1996,47-55. 7. J.A. De Vries, F.M. Rombouts, A.G.J. Voragen, and W. Pilnik. Carbohydr. Polym. 2 (1982) 25-33.

Polysaccharide Characterisation

17

8. H.A. Schols and A.G.J. Voragen. In: Progress in Biotechnology 14: Pectins and pectinuses. J. Visser and A.G.J. Voragen (Eds), Elsevier, Amsterdam, 1996, 3-19. 9. P.J.H. Daas, K. Meyer-Hansen, H.A. Schols, G.A. De Ruiter, and A.G.J. Voragen. Curbohydr. Res., 318 (1999) 135-145. 10. G. Beldman, M. Mutter, M.J.F. Searle-van Leeuwen, L.A.M. van den Broek, H.A. Schols, and A.G.J. Voragen. In: Progress in Biotechnology 14: Pectins and pectinuses. J. Visser and A.G.J. Voragen (Eds), Elsevier, Amsterdam, 1996,23 1-245. 1 1 . W. Pilnik and F.M. Rombouts. In: Enzymes and Food Processing. G.G. Birch, N. Blakebrough, and K.J. Parker (Eds), Applied Science Publishers LTD, London, 1981, 105-128. 12. W. Pilnik and A.G.J. Voragen. In: Food Enzymology 1 . P.F. Fox (Ed), Elsevier, London, 1991,303- 336. 13. K. Keegstra, K.W. Talmadge, W.D. Bauer, and P. Albersheim. Plant Physiol. 51 (1973) 188-197. 14. H.A. Schols, J.M. Ros, P.J.H. Daas, E.J. Bakx, and A.G.J. Voragen. In: Gums and Stubilisersfor the Food Industry 9. P.A.Williams and G.O. Phillips (Eds), IRL Press, Oxford, 1997,3-15. 15. J.A. De Vnes, F.M. Rombouts, A.G.J. Voragen, and W. Pilnik. Curbohydr. Polym. 3 (1983) 245-258. 16. J.A. De Vries, F.M. Rombouts, A.G.J. Voragen, and W. Pilnik. Curbohydr. Polym. 4 (1983) 89-101. 17. A.J. Taylor Curbohydr. Polym. 2 (1982) 9-17. 18. C.E. Tuerena, A.J. Taylor, and J.R. Mitchell. Curbohydr.Polym. 2 (1981) 193-203. 19. C.E. Tuerena, A.J. Taylor, and J.R. Mitchell. J. Sci. FoodAgric. 35 (1984) 797-804. 20. S.B. Tjan, A.G.J. Voragen, and W. Pilnik. Curbohydr. Res. 34 (1973) 15-32. 21. J.A. De Vnes, M. Hansen, J. Snderberg, P.E. Glahn, and J.K. Pedersen. Curbohydr. Polym. 6 (1986) 165-176. 22. Glahn and C. Rolin. In: Gums and Stabilizersfor the Food Industry 8. G.O. Philips, D.J. Wedlock, and P.A. Williams (Eds), DRL Press, Oxford, 1996, 393-402. 23. H. Kiyohara, J.-C. Cyong, and H. Yamada. Curbohydr. Res. 182 (1988) 259-275. 24. H. Kiyohara and H. Yamada. Curbohydr. Polym. 25 (1994) 117-122. 25. T.P. Kravtchenko, M. Penci, A.G.J. Voragen, and W. Pilnik. Curbohydr. Polym. 20 (1993) 195-205. 26. T.P. Kravtchenko, A.G.J. Voragen, and W. Pilnik. Curbohydr. Polym. 19 (1992) 115124. 27. T.P.Kravtchenko, A.G.J. Voragen and W. Pilnik. Curbohydr. Polym.18 (1992) 253263. 28. T.P. Kravtchenko. Studies on the structure of industrial high methoxyl pectins. Thesis Wageningen Agricultural University, 1992. 29. Y.C.Lee. J. Chromutogr.A ,720 (1996) 137-149. 30. N.A. Shanley, L.A.M. van den Broek, A.G.J. Voragen, and M.P. Coughlan. J. Biotechn. 28 (1993) 179-197. 31. N.A. Shanley, L.A.M. van den Broek, A.G.J. Voragen, and M.P. Coughlan. J. Biotechn. 28 (1993) 199-218. 32. A.T. Hotchkiss, Jr. and K.B. Hicks. Curbohydr.Res. 247 (1993) 1-7. 33. H.P. Lieker, K.Thielecke, K. Buchholz, and P.J. Reilly. Curbohydr. Res. 238 (1993) 307-3 1 1 .

18

Gums and Stabilisersfor the Food Industry 10

34. P.J.H. Daas, P.W. Arisz, H.A. Schols, G.A. de Ruiter, and A.G.J. Voragen. Anal. Biochem. 257, 195-202. 35. R. Komer, G. Limberg, J. Dalgaard Mikkelsen, and P. Roepstorff J. Mass.Spectrom. 33 (1998) 836-842. 36. G.J.W.M. van Alebeek, 0. Zabotina, G. Beldman, H.A. Schols, and A.G.J. Voragen. Carbohydr.Polym. submitted. 37. R. Komer, G. Limberg, T.M.I.E. Christensen, J. Dalgaard Mikkelsen, and P. Roepstorff. Anal. Chem. 71 (1999) 1421-1427. 38. P.J.H. Daas, A.G.J. Voragen, and H.A. Schols. Carbohydr.Res. submitted. 39. E.M.W. Chen and A.J. Mort. Carbohydr.Polym. 29 (1996) 129-136. 40. P.J.H. Daas, B. Boxma, A.M.C.P. Hopman, A.G.J. Voragen, and H.A. Schols. Biopolymers. submitted.

Multi-Angle Light Scattering Estimation of Pectin Molecular Weight and the Effect of Homogenization

LOUISE WICKER, MILENA CORREDIG, AND WILLIAM L. KERR Department of Food Science and Technology,Food Process Research and Development Laboratory, University of Georgia, Athens, GA 30602 USA

ABSTRACT A method has been presented to estimatethe molecular weight of commercial pectin based on separation on linear mix gel permeation chromatographyand multi-angle light scatteringdetection. Although pectin cannot be cleanly separated by GPC, analysis of individual slices of the chromatogramallows some estimation of the heterogeneity of commercial pectins. Although the pectic substances differed in molecular weight, the RMS values were similar. This technique was used to estimate a decrease in pectin molecular weight after high pressure homogenization. Pectin molecular weight was not changed by homogenizationat 17 MPa but was decreased by about 25% by homogenization at 124 MPa. The polydispersity (calculated as M,,./M,,) was not affected by homogenization.

1

INTRODUCTION

Pectin is a heteropolysaccharideof partially esterified u-l,4linked D-galacturonides, containing varying amounts of covalentl attached rhamnose and branches of L-arabinose, D-galactose, D-xylose, and L-rhamnose.7 Pectin functions as a gelling, thickening and stabilizing ingredient in foods. Functionality is related to the molecular weight, degree of esterification, distribution of esterification, presence of non-uronide components and other structural characteristics. In addition, the heterogeneity of pectin may result from the strong tendency to aggregate in solution depending on source, pre-treatment, method of extraction, and concentration. Likewise, elution solvent, temperature, pH, ionic strength, degree of esterification (% DE) and distribution of esterification (DDE) influence the propensity to aggregate. The propensity of pectin to aggregate has resulted in estimates of average molecular weight that are as varied as the methods that were employed. Some of the classical work using gel permeation chromatography(GPC), viscometric with single angle light scattering detection and component anal sis has shown wide variation in estimates of average pectin molecular weight.** ’ ’ ’ ’ Multi-angle light scattering detection was used to observe the removal of large pectate aggregates by successive filtration.’

‘

20

Gums and Stabilisersfor the Food Industry 10

Furthermore, it has been shown that the estimation of average molecular weight, % DE and neutral sugar coptent does not necessarily predict performance of pectins in foods.9310 Although pectinmethylesterase (PME) influences charge, charge distribution of pectin and its propensity to aggregate, our laboratory became interested in characterizing molecular weight of pectin because of the increasing evidence that aggregates of pectin effect PME activity. The complex of PME-pectin in the cell wall stabilizes the enzyme.''*'* Cations compete for the carboxylic groups of pectin and displace the enzyme from an inactive PME/pectin ~ o m p l e x . ' ~ ~Cation ' ~ * ' ~type, concentration and pH influence PME acti~ity.'~,''It has been recently shown that the detectable thermostable pectinmethylesterase(TS-PME) was significantly higher in some purified PME fractions than that measured in the heated crude extract.I8The authors suggested that heat solubilizes an inactive form of TS-PME from an insoluble inactive complex possibly composed by pectin." Finally, estimation of pectin molecular weight is necessary to produce and characterize "tailored" pectins. Commercial fractionation of pectins with specific properties and accurate prediction of functionality would be useful for food applications. For industrial application of pectin, it is necessary not only to characterize the structure/functionof the initial ingredient, but also to have a clear understanding of changes which may occur during processing. In a recent report, high shear and turbulence during microfluidizationdecreased the molecular weight of xanthan gum from about 25 to 4 million dalton~.'~ The high pressure, shear treated xanthan gum had lower viscosity and was less pseudoplastic than the untreated sample. Work carried out in our laboratory has demonstrated that high pressure throttling from 3 10 MPa to atmospheric pressure decreases the viscosity of pectin stabilized acidified milk dispersions by 3-4 times and improves the stability of the dispersions?' The objective of this study was to estimate the molecular weight of pectin using linear mix columns in-line with a refractive index and multi-angle light scattering detector and to determine if low pressure homogenization affects pectin molecular weight. 2

MATERIALS AND METHODS

Commercial citrus pectins which varied in methoxyl content and molecular weight were donated by Citrus Colloids (Hereford, UK), Hercules Inc. (Wilmington, DE) and TIC Gums (Belcamp, MD). Pectins were dispersed in water and dialyzed prior to freeze drying. Before analysis, pectins were dissolved based on solids content to obtain 2 to 6 mg/ml and filtered through 0.45 pm or 0.2 pm filters (Acrodisc, Gelman Science) prior to injection. For pectins which were pressure treated a 3.0 mg/ml solution was made in 50 mM sodium nitrate pH 5.8 and hydrated overnight at 4°C. An Avestin Emulsiflex C-5 (Avestin, Ottawa, Canada) was used to homogenize the samples at 17 MPa or 124 MPa for 5 passes through the homogenizer. Samples were immediately transferred in an ice bath. Samples were kept overnight at 4"C, and chromatography analyses was carried out the next day. The HPSEC-multi-angle light scattering (MALLS) was previously described by Corredig et al." The system consisted of a Waters P500 pump with an in-line de-gasser (Waters, Milford, MA), two in-line filters (0.22 pm and 0.1 pn)Millipore (Bedford, MA), a Spectrasystem autosampler with a 50 pl injection loop (AS 3000, Thermoseparation Products, San Jose, CA), and two linear mix ( 8 p n pore size) PL-aquagel-OH columns in series, enclosed in the autosampler oven at 35°C. In-line detectors included a multi-angle

21

Polysaccharide Characterisation

light scattering (LS) detector (DAWN DSP-F) equipped with an F2 flow cell and a He-Ne laser light source (L=633 nm) and a refractive index @ detector I) (Optilab DSP, Wyatt Technologies, Santa Barbara, CA) operating at 633 nm and 40°C. The LS detector was calibrated with 0.2 pm filtered HPLC grade toluene. The specific refractive index increments (dddc) were determined with the RI detector using a syringe pump (Raze1 Scientific, Stamford, CT). Pectins were dispersed in elution buffer (50 mM NaNO3, pH 5.8) and the dddc was estimated using the Wyatt software (dddc version 5.2). Elution data were analyzed using the AstraiEasi SEC software (version 4.1) and molecular weight as number average @In), weight average @Iw), z-average (Mz), root mean square (Rh4S) and polydispersity were determined from Debye plots. Data presented are representative of at least two injections per treatment.

3

RESULTS AND DISCUSSION

Molecular weight strongly influences pectin functionality as a thickening and gelling agent in foods. Some of the heterogeneity of reported pectin molecular weight is related to the propensity of pectin to aggregate. The data in Figure 1 shows that the GPC columns did not resolve pectin fractions.

0

5

10

15

20

25

Volume (ml)

Figure 1: Light scattering and refiactive index signals of high methoxylpectin (3.6 mdml),filtered using 0.45 p filters, eluted at I mUmin in 50 mM nitrate buffer. However, differences in the LS and RI signals distinctly show the polydispersity of the commercial pectin by the presence of a high molecular weight species at a relatively low concentration. In previous work it was shown that elution conditions such as solvent, flow, concentration, and filtration did not significantly influence the molecular weight for four commercial pectins estimated by LS?' In the RI signal, most of the pectins elute early in the chromatogramand show a distinctive tail at higher elution volumes. The higher LS signal at elution volumes near 11-12 min with a corresponding low RI signal reinforces the concept that commercialpectins are polydisperse with respect to molecular weight. The A S M A S 1 DAWN software allows the independent analysis of slices of the detector signal to further characterize the pectin. Figure 2 depicts the molecular

22

Gums and Stubilisers for the Food Industry 10

weight distributionplot for replicate injections and chromatography for one of the pectins. The slight differences in differential weight fraction with molecular weight are within experimental error.

1.

1.Orlo.

l.Orl0'

1.0x10.

Molecular Weight (g/mOl)

Figure 2: Molecular weight distribution of high methoxyl pectin (replicate samples are shown). Injected concentration 3.0 mg'ml(50 pl loop). Elution was performed with 50 mMsodium nitrate, at aflow rate of 0.5 ml/min. The estimated M,, Mw, and Mzvalues were 84,100,120,000,and 170,700, respectively (Table 1). The M a n ratio was 1.45,indicating the level of polydispersity in molecular weight for this pectin. Because of the angular dependence of the intensity of Table 1. Summary of Molecular Weight calculated by ASTMEASI of a High Methoxyl Pectin before and ajier Homogenization.

x Pal 84,100 120,000 170,700 1.45

x [Dal 91,800 126,700 193,700 1.39

X [Dal 63,600 91,400 168,200 1.44

Control 0

15,000 5,600 9,100 0.209 Pressure at 17 MPa c 7,800 1 1,000 28,700 0.05 Pressure at 124 MPa c 5,900 2,800 9,300 0.10

YOerror 17.8 4.6 5.3 14.4

YOerror 8.5 3 .O 5.5 3.6 % error 9.2 3.0 5.5 6.6

Polysaccharide Characterisation

23

scatteringof light, detection of scatter at different angles can be used to estimate the radius of gyration or root mean square.22Figure 3 illustrates the RI signal of four commercial pectins with elution volume. Figure 3 also depicts the change in RMS with elution volume. The RMS remained near 50-55 nm for the four pectins except at high elution volumes, where the signal to noise ratio was too great for an accurate estimate of RMS.

10.0

12.0

14.0

16.0

18.0

Volume (ml)

Figure 3: Root mean square radius of 4 commercialpectins as afinction of the elution volume. Concentrationofpectin was 3 mg/ml,filtered using 0.45 ,umfilters, eluted at 0.5 ml/min in 50 mMsodium nitrate, pH 5.8. It has been demonstratedthat high pressure homogenization (microfluidization) and shear disrupt aggregates of xanthan gum.I9Changes in molecular weight resulted in changes in the thickening properties of the polysaccharide. In this study, more moderate homogenizationwith a valve type homogenizer also resulted in a decrease of molecular weight. The molecular weight as a function of the elution volume (as RI signal) for pectin homogenized at different pressures is depicted in Figure 4. Very littie effect of homogenizationat 17 MPa was observed compared to the non homogenized control. However, a clear increase in elution volume of pectins was observed after homogenization at 124 MPa. A linear relationship between log Mwand elution volume was observed except at the extremes of the separation. Homogenizationhad no effect on the linear relationship,but altered the relative elution of 124 MPa treated sample. The effect of homogenization is more clearly depicted with a differentialmolecular weight plot (Figure 5). Homogenizationat 17 MPa had no measurable effect on the weight average molecular weight (Mw) distribution of the control and 17 MPa treated pectins, which were not significantly different at 120,000 and 126,700, respectively (Table 1). However, homogenizationat 124 MPa decreased the M, to 91,400. Figure 6 illustrates the RMS and the RI trace (elution volume) of control pectin and pectin homogenized at 124 MPa. The elution profile is shifted towards higher volumes after homogenizationindicating a decrease in molecular weight. The M,JMnratio varied between 1.3-1.7 for all replications and all treatments and was not significantly affected by homogenization. Estimation of RMS was not accurate due to the variation in signal to noise ratio of the scattering detector between the replicate samples.

Gums and Stabilisersfor the Food Industry 10

24

8.0

8.0

(0.0

12.0

14.0

18.0

Volume (ml)

Figure 4: Molecular weight and elution volume of high methoxyl pectin at 3.0 mg/ml in 50 mM sodium nitrate, as a finction of homogenizationpressure. (Elution volume for molecular weight is indicated on top ofjgure).

1.0~10'

I.0XlV

Molecular Weight (g/mol)

Figure 5: Differential weight distribution ofpressure treatedpectins and control pectin. Pectin concentration was 3 mglml, eluted with 50 mMsodium nitrate, pH 5.8.

Polysaccharide Characterisation

25

12.0

lw.o

14.0

10.0

t

v)

10.0

2 d

f I

1.o 8.0

10.0

.

*

>

>

!

,

12.0

14.0

10.0

Volume (ml)

Figure 6: Root Mean Square radius and elution volume (XI signal) of control and pectin treated at 124 MPa. Pectin concentration was 3 mgiml, eluted with 50 mMsodium nitrate, pH 5.8.

4

ACKNOWLEDGEMENTS

This work was supported in part by the BARD project IS-2793-96.The contributions of pectins by Citrus Colloids and Hercules are gratefully acknowledged. 5

REFERENCES

1 A.G.J. Voragen, W. Pilnik, J.-F. Thibault, M.A.V. Axelos, and M.G.C. Renard, ‘Food Polysaccharidesand Their Applications’, Marcel Dekker, Inc. New York, 1995,287. 2 M.L. Fishman, P.E. Pfeffer, R.A. Barford, and L.W. Doner. J. Agric. and Food Chem., 1984,32,372. 3 M.L. Fishman, D.T. Gillespie, and S.M. Sondey. Curb. Res., 1991,215,91. 4 M.L. Fishman, P. Cooke, B. Levaj, D.T. Gillespie, S.M. Sondey, and R. Scorza. Arch. Biochem. and Biophy., 1992,294,253. 5 M.L. Fishman, P. Cooke, A. Hotchkiss, and W. Damert. Carb. Res., 1993,248,303. 6 P.D. Hoagland, M.L. Fishman, G. Konja, and E. Clauss, J Agric. and Food Chem., 1993,41,1274-1281. 7 P.D. Hoagland, G. Konja, E. Clauss, and M.L. Fishman, J. Food Sci., 1997,62,69. 8 A. Malovikova, M.Rinaudo, and M. Milas, Carb. Polym., 1993,22,87. 9 T.P. Kravtchenko, A.G.J. Voragen, and W. Pilnik, Carb. Polym., i 992,19, 1 15. 10 T.P. Kravtchenko, G. Berth, A.G.J. Voragen, and W. Pilnik, Carb. Polym., 1992,18, 253. 1 1 L.Wicker, E. Echeverria, and M. Vassallo. J. Food Sci., 1988,53,1171.

26

G u m and Srabilisers for rhe Food Industry 10

12 H. M. Macdonald, R. Evans, and W.J. Spencer. J. Sci FoodAgric., 1993,62,163. 13 L.R. MacDonell, E.F. Jansen, and H. Lineweaver. Arch. Biochem., 1945,6,389. 14 J. Nari, G. Noat, and J. Richard. Biochem. J., 1991,279,343. 15 D. Chamay, J. Nari, and G. Noat. Eur. J. Biochem., 205,711. 16 V.A. Leiting, and L. Wicker. J. FoodSci., 1997,62, 1. 17 D. Sun,and L. Wicker. J. Agric. and Food Chem. 1999,47,1471. 18 M. Corredig, W.L. Kerr, and L. Wicker. J. Agric Food Chem., 1999, submitted. 19 N.Lagoueyte and P. Paquin, Food Hydrocolloiak, 1998,12,365. 20 L. Wicker, R.T. Toledo, J. Moorman and R. Pereyra. Milchwiss., 1999, Accepted. 21 M. Corredig, W.L. Kerr, and L. Wicker, Food Hydrocolloiak, 1999, Accepted. 20 P.J.Wyatt, Analytica Chimica Acta, 1993,272, 1.

EXTRACTION AND CHARACTERISATION OF PECTIN FROM POMELO FRUIT PEELS

M. H. NOniah, E. 0.Fang and A. Abd Karim Food Division, School of Industrial Technology Universiti Sains Malaysia 11800 Penang, Malaysia

ABSTRACT Pectin was extracted from pomelo h i t (Ciaus g r d s ) by three different extraction methods. Pomelo, a tropical h i t which belongs to the citrus family has a very thick skin which can be used as a source of pectin. It is currently widely cultivated in the country but little is known of the properties of pomelo pectin which could be extracted. This work describes the extraction of pectin 6om pomelo fruit peels using ethanol, aluminium salt and acetone precipitation extraction methods. The pectin obtained &om these methods were compared in terms of yield, color and gelling characteristics. Acetone extraction using sodium hexametaphosphate was the best method which yielded 20.8% pectin (dry weight basis) fiom the raw material. The isolated pectin contained 12.7% moisture, 74.9% anhydrouronic acid (AUA), 5.3% methoxyl content, 40.5% degree of esterification and 4.5% ash. The content of AUA indicates the high degree of purity of the pectin isolated. According to the values for methoxyl content and degree of esterification, pectin isolated fiom pomelo peels can be classified as low methoxyl pectin and are of potential use in the manufacture of low sugar products such as low sugar jam and jellies. Gelling properties of pomelo pectin such as setting time, setting temperature and gel strength were studied and compared with those of a commercially available citrus pectin. The effects of temperature, time and pH of extraction on pectin yield and optimum conditions for pectin extraction using response surfice methodology are also discussed. 1 INTRODUCTION

Pectin is a polysaccharide which can undergo gelation and can stabilise emulsions which makes it usehl in the manufacture of food, cosmetics and medicine. Its most important hnctional property is the ability to form gels and thus it is widely used in food formulations. Pectin forms two types of gels i.e. conventional high methoxyl pectin (HMP) gels with 65% or more sugar and low methoxyl pectin (LMP) gels with or without sugar using calcium or other polyvalent cations. Pomelo produces fruit throughout the year with peak seasons between January to February and August to September. Little is known about the properties of pomelo pectin which could be extracted fiom h i t unsuitable for fiesh market use and also fiom immature h i t s removed fiom the trees

28

Gums and Stabilisers for the Food Industry 10

during the trimming stage to allow for only a limited number of fruits on each tree to filly ripen. Pectin occurs commonly in most plants in the middle lamella. The two main sources of commercial pectin are apple pomace and citrus fruit rinds. Other potential sources of pectin reported in the literature include mango’, sunflower head* and lime peel3. The contents of pectin in various tropical fruits such as mango, banana, pawpaw, orange, cashew apple, guava and lime were reported to be about 1-2% ‘. Several methods of extraction of pectin have been discussed i.e. lab-scale batch extractions as well as continuous, countercurrent extraction’-’. M ~ experimental Y factors can influence the yield of pectin to be extracted. Studies on the effects of extraction variables on yield of pectin have been This paper reports three methods of extraction of pectin from pomelo peels as well as the effects of extraction variables on the yield of pectin extracted by the acetone-HCI precipitation method as a preliminary step to obtaining pectin and characterising the pectin. A response surface methodology approach to the optimization process was used. A mathematical model that can be used to predict the ‘best’ settings of temperature, time and pH to produce the optimum value for percentage yield of pectin extraction from pomelo peels was also developed. 2 MATERIALS AND

METHODS

2.1 Materials

Pomelo fruits (Citrus p u d i s ) were obtained from a pomelo fruit plantation in the northern State of Malaysia. Commercial citrus pectin from Fluka, Switzerland, was used for comparison purposes. Pomelo peels were cut into small cube sizes, treated with hot 95% ethanol for 15 min, washed with water, ressed to remove the excess water and dried following procedures described by Kertesz The dried peels were then ground in a laboratory dry blender for hrther extraction work.

P’.

2.2 Extraction Procedures

Three methods of extraction were initially performed and compared to find the best method for obtaining good yield and quality pectin. Duplicate extractions were carried out for each method. In each of the methods, the pectin precipitate collected was dried in a vacuum oven at 40 “C for a few hours to constant weight and ground finely. The yield obtained was reported as % yield (g dried pectin per 100 g dried peels). This pectin was used for firther analysis. 2.2. I Extraction using Ethanol-HCZ. A modification of the ethanol-HCI method by McCready” was employed. One hundred grams of ground and dried peel were weighed into a tared 2000 mL beaker containing 800 mL distilled water. Twelve grams of freshly ground sodium hexametaphosphatewere then added and initial pH adjusted with 3 N HCI to 2.2 f 0.1. The mixture was heated in a stirred vessel equipped with a propeller type stirrer at 1000 rpm at 75 f 5°C for 1 hour and the pH checked at intervals of 15 mins. Water loss was replaced at intervals except in the last 20 minute of extraction. The extract was vacuum filtered through muslin cloth and the residue washed with 200 mL of warm water. The washings were added to the filtrate, which was concentrated by evaporation on a hot plate to approximately one fifth of its initial volume. The

Polysaccharide Characterisation

29

concentrated pectin solution was cooled to 50 "C and poured into a volume of ethanol in a ratio of 1:3, the ethanol contained 0.5 M HCI. The mixture was stirred for 0.5hour and allowed to stand for one hour. The precipitate was vacuum filtered washed with more ethanol-HCI solution. Finally the extract was washed with acetone to remove traces of HCI and ethanol which could interfere with the methoxyl determination. The precipitate was dried to constant weight. 2.2.2 Extraction of Pectin using Acetone-HCI. Eighty grams of ground pomelo peel were extracted with water and log sodium hexametaphosphateat 75 "C in a stirred vessel as mentioned in the earlier method. The pH was adjusted to 4.0 f 0.5 when necessary using citric acid or sodium hydroxide. Water lost by evaporation was replaced at intervals except in the last 20 min of extraction. The mixture was filtered and the residue was successively extracted. The filtrate was concentrated to one third of its volume and the pectin was precipitated in a solution of acetone in 0.5M HCI. The jelly-like yellowish precipitate was collected, fbrther washed with 70% acetone until it was essentially chloride-ion fiee or the pH was above 4.0, dehydrated further in 500 mL of acetone and then dried.

The method used was a 2.2.3 Extraction of Pectin using Aluminium sulphate. modification to the method by Wiles and Smit13. Samples were treated as the ethanolHCl method except that they were alternatively treated with 1% sodium bisulphite instead of sodium hexametaphosphate and heated at 70-75OC for 1 hour with constant stirring. The extract was filtered through muslin cloth and the residue washed with 100 mL of warm water. The washings were added to the filtrate which was then cooled to 50°C and 1.75% (vh) of aluminum sulphate was added. Immediately thereafter 1.05% v/v of sodium carbonate solution was added and stirred for about 15 seconds whereupon a pectin aluminum complex separated throughout the liquor as a continuous gel which immediately coagulated into curds. The final optimum pH of the mother liquor being 3.9 to 4.2 at 30°C. The solution was centrifbged at 35OOG for 10 mins to separate the coagulated gel fiom the mother liquor. The gel was rinsed in ethanol, drained and fresh 75% alcohol added to make a fluid mixture whereby the pH of the suspension was adjusted to about 5.5 f 0.3, the pectin was then dried. 2.3 Analysis and Characterizationof Pectin

The tests were performed in triplicates and the average values are reported. Moisture was determined according to AOAC method" with 1 g of ground sample weighed and dried at 100°C for 4 hours to constant weight. To determine the ash content, 1 g of pectin was weighed in a tared crucible and ignited slowly, then heated in a muffle fbrnace at 600°C for 4 hours. Residue was cooled in a desiccator and weighed to constant weight". 2.3.I Equivalent Weight. Methoxyl and AUA contents and equivalent weights were determined by the standard methods of Owen". Values for equivalent weight of pectin are used to calculate anhydrouronic acid (AUA). Equivalent weights were done by weighing 0.5 g pectin (moisture fiee) in a 250 mL conical flask, moistened with 5 mL ethanol. 1 g of sodium chloride were added to sharpen the end point. lOOmL of carbon dioxide fiee distilled water and 6 drops of phenol red indicator were added. The pectic substances were stirred rapidly to dissolve, then titrated slowly with 0.1 N NaOH until the colour of the indicator changed (pH 7.5) and persisted for at least 30 seconds. The

30

Gums and Stabilisers for the Food Industry I0

neutralised solution was saved for methoxyl determination. 2.3.2 Methoxyl content. Methoxyl (MeO) contents were determined by adding 25 mL of 0.25 N NaOH to the neutral solution, mixing thoroughly, and allowing to stand for 30 minutes at room temperature in a stoppered flask. 25 mL, of 0.25 N HCl was then added and titrated with 0.1 N NaOH to the same end point as before.

MeO%=

meq of sodium hydroxide x 3 1 x 100 wt of sample (mg)

Where, 3 1 is the molecular weight of methoxyl. 2.3.3 Arhydouronic acid (A UA). If the equivalent weight and methoxyl content of pectin is known, its AUA can be calculated as follows",

Yo AUA =

~

176 x 100 Z

(2)

Where, 176 is the molecular weight of AUA and

Z=

wt of sample (mg) meq of alkalifor free acid + meq of alkalifor methoxyl

2.3.4 Degree of Ester$cation. The degree of esterification (DE) of pectin can be determined according to the formula given below16. Yo DE =

176 x CHaO% x 100 31xAUAYo

(4)

Where CH3O is % methoxyl content. 2.3.5 Physical Properties. Gelling properties such as setting time, setting temperature and gel strength of pomelo pectin extracted b the acetone precipitation method were determined by method described by Lodge et alh. Gels were formulated to contain 0.2 g pectin and 25 g sugar (pectin concentration of 0.4%, wlw) and adjusted to pH 2.8-3.1 with citric acid solution. The solution having a Brix value of 65" was then poured into a plastic cup and left to stand at room temperature for about 18 hrs before

PolysaccharideCharacterisation

31

being tested for gel strength. The gel strength or penetration force was measured with a stainless steel probe (P/lKSS KOBE 1 cm') forced into the gel to a depth of 4 mm using the Texture Analyzer ZTZ version 5.51 (Stable Micro System) with a cell loading of 5 kg. The gelation or setting temperature and setting time were also determined. A standard Fluka 250 grade citrus pectin (Fluka chemicals) was also tested for gel strength, setting time and gelation temperature for comparison purposes. 2.4 Effects o f Process Variables on Pectin Yield

The effects of temperature, time and pH on pectin yield by the acetone precipitation method were investigated. The temperature, pH and time of extraction were varied according to the experimental design chosen. Statistical design, analysis of variance (ANOVA) and plotting of the response surface were performed using Design Expert version 5.0 (Stat-Ease Inc.) and also Statgraphics PZus (Statistical Graphic Corp.), comprehensive graphic Window-based software for statistical experimental design, analysis and interpretation. The extraction process was assumed to be a system affected by three independent variables or factors, Zi (in this study, extraction temperature, time and pH), which were controlled and measured. It was assumed that the dependent variable, also referred to as response, Y (in this study, % pectin yield) defined the system and was experimentally determined. A mathematical finction, was assumed to describe the relationship between response, Y and the factors, Zi.

The experimental design consisted of three variables at five-levels with 20 runs. 3 RESULTS AND DISCUSSION

Three methods of extraction were initially compared to determine the yield and characteristics of pectin extracted fiom pomelo h i t peels. A summary of results is shown in Table 1. The gelling properties of pectin gels are shown in Table 2. The extraction of pectin using the acetone-HCI precipitation method gave the best yield (20.8%) and the most desirable colour. Ifthe yield of pectin is greater than lO??,then the source of pectin is considered possible for commercial use. The yield of commercial pectin from apple pomace was reported to be 17%" whereas pectin from Coorg Mandarin peel was 13.2%'*. The ash content is quite low except for pectin extracted by method C which is 16.1%. Ash content also can affect the gelling of pectin. A low ash content is more favorable for gel formation". Commercial high methox I and low methoxyl pectin has been reported to contain about 1.9-5.2% ash content. The equivalent weights obtained were used for the calculations of %AUA and % DE. In all of the extraction methods, %DE in pectin are found to be lower than 50% and the % methoxyl content are also low. This indicates that pectin extracted fiom pomelo peels is of the low methoxyl type (LMP) since LMP is classified as having % DE lower than 50% and % methoxyl content between 0.5-7.0%. The content of methoxyl plays a key role in the ability for gel formation. LMP can form gels with a lower amount of sugar or without sugar in the presence of divalent cations. Hence, this will be more usefil in making of low sugar or

r

32

Gums and Stabilisersfor the Food Industry 10

calcium pectate gels. The AUA content obtained from methods A and B are >70% . The content of AUA indicates the high degree of purity of pectin isolated. The AUA content (39.9%) of the pectin extract by aluminium precipitation method indicates that the extract is not pure. Citrus pectin obtained commercially contains about 83% AUA and 10% methoxyl content5. AUA content in pectin was suggested to be not less than 70%”.

Table 1 Yield and Chemical Characteristics of Pectin Extracted and Pectin Gel Formed

Yield of pectin, %

A. Ethanol-HCl precipitation* 19 f 0.2

Moisture, %

13.4 f 0.5

Ash content, % Equivalent weight

B. Acetone-HCI precipitation 20.8 f 0.3 12.7 f 0.2

2.5 f 0.01

9.3 f 0.3

4.5 f 0.02

391.2 f 0.6

Methoxyl content

C. Aluminium surphare precipitation* 12.6 f 1.0 16.1 f 0.6

395.6 f 0.8

738.3 f 0.6

5.3 f 0.3

2.8 f 0.1

5.1 f 0.03

AUA, %

73.9 f 0.2

74.9 f 0.9

39.9 f 1.1

DE, Yo

39.2 f 0.1

40.5 f 0.6

40.3 f 0.1

Brown

Light brown

Light green

Solubility -good, pale yellow

Solubility -good, pale yellow

Solubility - low, Slightly greenish

Colour of pectin Colour (gel)

n = triplicate runs The results given in Table 2 indicate that pectin from pomelo peel is a ‘rapid set’ type whereas the citrus pectin is of the ‘slow-set’ type. The results are consistent with reports by Owens” that pectin with DE >70% is the ‘rapid set’ type; between 50-70% is the ‘slow set’ type but is ‘rapid-set’ again when DE is lower than 50%. The lower the DE for LMP the faster it sets and at higher setting temperature. Table 2 GellingProperties of Pectin Gels Isolatedpectin gel

Citruspectin ( F L U . 250 grade)

75-80

55-60

Setting time (min)

-1

> 10

PH

2.9

2.9

Gel strength (g)

50.0

194.0

Setting temperature (‘C)

Polysaccharide Characterisation

33

Figures 1 and 2 show force vs time curves from texture analysis on the pectin gels prepared. The results show that gel strength was low for the isolated pectin from pomelo (50.0g or OSON) when compared with pectin from standard citrus grade 250 (194.0 g or 1.94 N). Citrus pectin at a concentration of 0.4% formed fm gels but pomelo pectin formed softer gels at this concentration. Thus it estimated that the jelly grade of pomelo pectin is lower than that of citrus pectin grade 250. Response surface methodology was applied for the defined experimental region and used to evaluate the effects of three main extraction variables (temperature, time and pH). A central composite rotatable second-order design with six replicates at the centre point, eight Q3) factorial points and six axial points was chosen for optimisation of extraction conditions. A predicted model for pectin yield was developed as functions of temperature, time and pH i.e. a reduced quadratic equation as shown below,

Y = 18.32 + 3.43 (XI) + 1.74(X2) - 0.27 (X3) - 1.62 (X'I)

Where, XI, XZ, X3 are coded values for temperature, time and pH of extraction respectively and Y is the response ,% yield of pectin. The effects of all three variables (temperature, time and pH) on the yield of the extracted pectin were found to be significant with pH being least significant. Reports by Aravantinos-Zafiris and Oreopoulou9on the optimisation of pectin from orange waste also showed that all three factors, temperature,time and pH had significant effects. There was no significant interaction between the independent variables. Thus the increase in time of extraction was independent of the effect of temperature and pH on pectin yield. A significant quadratic effect of temperature was observed 'which results in a decrease of pectin yield. Among the variables, regression coefficient for temperature is of a higher value indicating the effect of temperature is greater than that for time. A compromise has to be reached between obtaining maximum pectin yield and minimum pectin degradation. High heat enhances rate of reaction and pectin can undergo degradation at temperatures greater than 90°C7. Long time extraction causes more hydrolysis and pectin chains become shorter than short time extraction2'. A verification experiment at the optimum conditions on the yield of pectin was performed. The experimental value obtained is 20.8 f 0.36% yield whereas predicted value is 20.1%. The results showed that the experimental and predicted values are reasonably close. A contour plot and a response surface plot for percentage pectin yield as a function of two most important factors (temperature and time), holding pH at a value of 3.6 are drawn in Figures 3 and 4 respectively. The results indicate that a temperature not greater than 90°C associated with time not longer than 64 min and a low pH value (pH 3.6) lead to a good % yield of pectin. 4. CONCLUSION

In conclusion, pomelo peel was found to be a good source of pectin. The yield obtained is about 20% and pomelo pectin is of the low methoxyl type. The extraction procedure using acetone precipitation was also optimised in which the 'best' conditions for extraction is at a temperature of 90°C, extraction time not longer than 64 min and at low

34

Gums and Stabilisersfor rhe Food Industry 10

Force (N) 0 . m

1

Stable Micro Systems - Texture Expert

3 . 6 4.0

1

5.0

6.0

Time (sec) -0.2000 J

Figure 1 Force vs time curvefor pomelo pectin

Force (N) 2.000 -1

1.750

Stable Micro Systems - Texture Expert

-

1.5001.250 1.2501.000-

0.750 0.500 0.250

-

0.000 -0.0 -0.250’ -0.250

1

1.0

3.0

4.0

5.0

6.0

Time (sec)

Figure 2 Force vs time curvefor cilrus pectin

35

Polysaccharide Characterisation

Yield DESIGN-EXPERT Plot Actual Factors: X = Temp Y = Time Actual Constants: pH = 3.60

70.00

76.25

82.50

88.75

95.00

Temp

Figure 3 A mrjime contourplot ofyield as afunction oftemperature and time

-

DESIGKWERT Plot

Actual Factom: X = Temp Y = Time

-

Actual Constants: pH 3.60

,w.w .:.:.:...

40.00 70.00

Figure 4 A 3-a?memiomlpesentation of the response surface for the yield as a function of temperature (“C)andtime (min)

Gums and Siabilisersfor the Food Industry 10

36

pH value of 3.6. Acknowledgement The authors acknowledge the research grant provide by Universiti Sains Malaysia, Penang that has resulted in this article. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21.

A. N. Srirangarajan and A. J. Shrikhande,J. FoodSci.. 1977,42,279. A. Miyamoto and K. C. Chang, J. FoodSci., 1992,57, 1439. R. A. Padival, S.Rangana and S.P. Manjrekar, J. Food Technol., 1979, 14,333. E. C. Nwanekezi, 0. C. G. Alawuba and C. C. M. Mkpolulu, J. FoodSci. Technol., 1994, 31, 159. B. J. Francis, and J. M. K. Bell, J. TropicaZSci.,1975, 17(1), 25. B. K. Simpson, K. B. Egyankor and A. M. Martin, J. Food Processing and Preservation, 1984, 8(2), 63. L. Phatak, K. C. Chang and G. Brown, J. FoodSci., 1988,53(3), 830. J. Wang, D. P. Wiesenborn, J. G. Schwarz, and K. C. Chang, Transactions of the ASAE, 1997,40(6), 1654. G. Aravantinos-Zafiris and V. Oreopoulou, J. Sci. FoodAgric., 1992,60, 127. F. Michel, J. F. Thilbault, C. Mercier, F. Heitz and F. Pouileande, J. FoodSci.,1985, 50, 1499. Z. I. Kertesz, ‘The Pectic Substance’ Wiley (Interscience), New York, 1951. R. M. McCready, ‘Methods in Food Analysis’ Academic Press, New York and London, 1970, 2“ded., p 565. R. R. Wiles and C. J. B. Smit, U.S. Patent 3,622559, Sunkist Growers Inc, Nov 23. 1971. AOAC Association of OWcial Analytical Chemists. ‘Oficial Methods of Analysis’, 13* edition. Washington D.C., 1980. H. S. Owens, R. M. McCready, A. D. Shepard, T. H. Schultz, E. L. Pippen, H. A. Swenson, J. C. Miers, R. F. Erlandsen and W. D. Maclay, ‘Methods used at Western Regional Research Laboratory for Extraction of Pectic Materials’, USDA Bur. Agric. Ind. Chem. 1952, Report No. 340, p9. T. H. Schultz, ‘Methods in Carbohydrate Chemistry’, Academic Press, New York, USA, 1976, Vol5, p189. N. Lodge, T. T. Nguyen and D. McIntyre, A FoodSci., 1987,52(4), 1095. J. S.Pruthi, C. M. Parekh and L. Ghirdari, Food Science, 1961, lO(1 l), 327. F. W. Sosulski, M. J. Y. Lin and E. S.Humbert, Can. Znst. FoodSci. Technol., 1978, 11(3), 113. National Research Council, ‘Food Chemical Codex’, 3d ed., National Academy of Science, Washington, DC, 1987, p 215. K. C. Chang, N. Dhurandhar, X. You and A. Miyamoto, J. Food Sci., 1994, 59, 1209.

Gum Arabic

- Quality and Quantity Assured

K. A. Karamallah

Gandil Agriculture Co. Ltd, PO Box 522, Khartoum, Sudan.

SUMMARY Analytical data have shown that many qualifying indices of such a natural product, as Gum Arabic remained virtually constant over the years. Evidence obtained from the analyses of over 800 authenticated gum samples collected from 12 locations in the Gum belt of Sudan revealed that such factors as soil, rainfall, temperature, age of the tree, site of exudation and picking number had very little or no significant effect on gum quality. Data indicate that Gum Arabic from Acacia senegal is different and distinct botanically, physicochemically and functionally from other Acacia gums e.g. A. seyal gum. It is the only Acacia gum that has been tested toxicologically and proved to be safe (AD1 not specified) as a food additive. A prelimivary toxicological investigation carried out on both seyal varieties (var. seyal and var. fistula) on samples supplied by the Gum Arabic Company (GAC) has indicated some histo-pathological effects (particularly at higher dozes) on rats fed with these two gums. Good quality Gum Arabic is assured in sufficient supply at a reasonable price through good agricultural practices as offered by Gandil Agricultural Co. Ltd. through Acacia senegal plantations on very large scale.

introduction Gum Arabic, a natural exudate of the trees of Acacia senegal has been an important article of commerce for about 4000 years. This gum is widely used in the food, pharmaceutical and cosmetic industries due to its unique physico-chemical and functional properties. It has been subjected to intensive investigations, and systematic approach of research into the characteristics and properties have taken place for about four decades by many workers including Anderson and his colleagues. As a result, there is a wealth of information which clearly indicates that gum from Acacia senegal is a complex proteoglycan acidic salt (mainly Ca, K, Mg and Na) of high molecular weight (average 4 X lo5, Mv). The polysaccharide moeity is composed of D-galactose, L-arabinose, L-rhamnose, D-glucuronic acid and its 4-0-methyl ether together with a proteinaceous com onent. The protein is an integral part of the molecular structure of the gum1 2'3* 4 5,6,$3

38

Gums and Stabilisers for the Food Industry 10

Qualifying Indices of Gum Arabic Quality means identity and purity of a certain product. Therefore for each product there are certain qualifying indices or parameters, that should be assured before the product is judged suitable for use or application. In the case of gum from Acacia senegal (Gum Arabic), about twenty parameters have been proposed by different author^^^^'^"^'^' This natural product is composed mainly of carbohydrates (ca. 80%), protein (ca. 3%), minerals (ca. 4%) and moisture (ca. 11%). The five most important parameters that can be used to identify raw Gum Arabic (Acacia senegal) and distinguish it from other Acacia gums are: (1) Specific optical rotation, (2) Nitrogen content, (3) Ash content, (4) Moisture content and (5) Absence of tannins. While tannin will be discussed later in the paper, it is interesting to note that there is a distinct relationship between the nitrogen content and specific optical rotation of gum from different Acacia species. Such relationship clearly distinguishes between gums of Acacia species12. Fig. 1 shows that the different species can be clustered into two major groups based on optical rotation. Again, within each group, there is a distinct differentiation between each species based on the nitrogen content. It also shows that there are differences within variants of each Acacia species. Table 1 shows the major physico-chemical properties of different Acacia species of the Sudan. Table 2 depicts the major physico-chemical differences between Gum Arabic from Acacia senegal and gum from Acacia ~ e y a ? ' ' ~ ' ' ~ ,~ For ' ~ ' example, '~'~~ gum from Acacia senegal has a negative optical rotation (mean -31.3"), while that of gum from Acacia seyal has a positive optical rotation (mean +50.1°). Again, gum from Acacia senegal has higher content of rhamnose (12 - 14%) and lower arabinose content (24 - 29%) compared to the rhamnose (3 - 4%) and arabinose content (41 - 45%) of gum from Acacia seyal. Also the nitrogen content of gum from Acacia senegal is nearly three times that of the nitrogen content of gum from Acacia seyal. Gum from Acacia senegal has been shown to have a macromolecular structure that differs significantly in a number of ways from that of gums from Acacia seyal. C13NMR spectral studies clearly indicate the different peaks for rhamnose residues. Such peaks are virtually absent in spectra of gum from Acacia seyal. This clearly distinguishes between the two gums" (Fig. 2). In a very recent studylg, over 1500 authentic and commercial Acacia senegal var. senegal gum samples were analyzed to evaluate existing quality control parameters. This study also objectively assessed the potential of new parameters such as pH, viscosity, viscosity average molecular weight (Mv), equivalent weight and total uronic acid content as additional qualifying indices. The results from this study indicate the following mean values: moisture (10.75%), ash (3.77%), nitrogen (0.328%), specific optical rotation (-31.3'), pH (4.66), equivalent weight (1436) and total uronic acid content (13.71%) (Table 4). No significant variations have been observed in these parameters over the last forty years. The results of this study indicate wide variations in values for intrinsic viscosity and viscosity average molecular weight, indicating that such parameters cannot be used as qualifying indices.

60

30

a

4 0

0

5

+

0

0

+

EP

+

0.5

+

A. sieberana var. sieberana nubica A. tortilis subsp. tortilis 0 A. seyal var. seyal x A. nilotica subsp. nilotica A. laeta o A. mellifera

+A.

A

P 0

Nitrogen (K)

0

_-

0 A. drepanolobium 0 A. seyal var. fistula + A. nilotica subsp. tomentosa A. polyacantha

m A. tortilis subsp. raddiana

-

1.5

A. sieberana var. vennoesenii

1

+

.

-

2

A A. sieberana var. villosa A. tortilis subsp. spimcarpa A.genardii A A. ehrenbergiana 0 A. nilotica subsp. astringens E1 A. senegal

Fig. 1. Relationship between specific optical rotation and nitrogen content of Acacia gums from Sudan.

-50

-40

-30

-20

-10

0

10

i 20

2

.-0 5

40

m 50

g!

0

-

70

80

90

100

a

: Karamallah, 1964. : Khogali, 1977.

0 P

Polysaccharide Characterisation

-

1

180

-

100

1

140

41

I

120

,.

I

100

80

60

40

I

Fig. 2 13CNMR spectra of (a) Acacia senegal and (b) Acacia seyal gums”

20

0

42

Gums and Srabilisers for the Food lndusrry 10

Table 2. Comparitive analytical data for Acacia senegal and Acacia seyal gums collected between 1960 and 1999 in Sudan.

I Arabinose (%)

Rhamnose (YO) Glucuronic acid (%)

24 - 29 12 - 14 16 - 17

I

3-4 11 -12

Another study of 800 authentic samples of gum from Acacia senegal var. senegal collected from thirty two different locations of the gum producing belt of the Sudan, showed that factors such as soil, rainfall, temperature, age of trees, site of exudation and picking sequence had no significant effect on the physico-chemical properties of the gum (Tables 3,4,5,6,7 and 8). Attempts have been made using statistical analysis to show that gums from different Acacia species are the same. When the objective is to distinguish between two species, use of principal components analysis combined with some cluster analysis is highly inappropriate and incomplete. Principal component analysis leads to inherent bias towards equality hypothesis, e.g. because of the linear nature of the principal components, but also because their construction is driven solely by the data. Principal components analysis is over strongly influenced by outlying observations, any non-linearities in the data, and by the way the data are scaled. It also tends to smooth away the key conclusions to be drawn from the data2'. How Safe is Your Gum Arabic

Fifty-two samples of authenticated and commercial Acacia senegal gum (Gum Arabic), from the seasons 1995/96, 1996/97 and 1977/78, were tested for the count of total bacteria, mould, yeasts and coliforms2'. The maximum microbial loads of these samples were about lo6 CFU/gm bacteria, and lo5 CFU/gm of moulds. The microbial loads of processed gum samples were much lower than the loads of raw gum samples and ranged between 4 0 to a maximum of 7.9 X l o 4 CFUIgm. Coliforms and yeasts were not detected in any of the samples tested (Table 9).

43

Polysaccharide Characterisation

Table 3. The effect of soil type on the moisture, ash, nitrogen mntent and specific optical rotation and pH of Acacia senegal var. senegal of Sudan. Parameter

Soil type

Moisture (%)

Mixed Clay Mixed Clay Mixed Clay Mixed Clay Mixed Clay

Ash (%) Nitrogen (%)

Sp. Rotation (degrees) PH

No. of samples

Mean

18 09 18 09 18 09 18 09 18 09

10.77' 10.76' 3.88' 3.05a 0.31' 0.30' -30.91' -31.20' 4.74' 4.52'

~

Table 4. The effect of rainfall on the moisture, ash and nitrogen content and specific rotation and pH of Acacia senegal var. senegal of Sudan.

p. Rotation (degrees)

800

20

Mean values marked with (a) are not significantly different from each other.

Table 5. The effect of season (temperature) on the moisture, ash and nitrogen content, and specific rotation and pH of Acacia -

Mean values marked with (a) are not significantly different from each other.

44

Gums and Stubilisers for the Food Industry I0

Table 6. The effect of age of the trees on the physico-chemical properties of gum from Acacia senegal var. senegal of Sudan. Ash

(Yo)

Nitrogen (%)

pH

Specific rotation

Equivalent weight

1

I

Uronic acid

1

Table 7. The effect of site of exudation (branch, stem) on the physico-chemical properties of gum from Acacia senegal var. senegal of Sudan.

Mean values marked with (a) are not significantly different from each other.

The dominant bacteria contaminating Gum Arabic samples studied were gram (+ve), rod shaped, spore-forming ones. Gram (+ve) non-spore forming rods and gram (+ve) cocci types of bacteria were also detected, but in insignificant amounts. An experiment carried out to simulate the effect of sunlight and UV light on the microbial load of samples resulted in a very clear decrease in the bacteria and mould counts (Table 10). Although the microbial counts are well within the acceptable limits; efforts have been made to develop a' simple technique to reduce such counts. The proposed method of application of UV radiation leaves with the possibility of alterations in the carbohydrate composition of the gum. The possibility of changes in some of the physical characteristics (gummosis) due to elimination of some of the natural micro flora, cannot be ignored.

Polysaccharide Characterisation

45

Table 8. The effect of picking sequence on the physicochemical properties of gum from Acacia senegal var. senegal of Sudan.

Bacteria count (CFUIgm)

Mould count (CFUIgm)

Raw gum 3 X 105

Raw gum

I Processed gum I N.D.

1 . 6104 ~

1 Processed gum 1 N.D.

Table 10. Effect of sunlight* on microbial loads of raw and processed gum from Acacia senegal var. senegal from Sudan.

Gums and Stabilisersfor the Food Industry 10

46

Data clearly indicate that processing reduces microbial counts. It appears to be a result of the reduction of the moisture content in the gum. Thus, reducing the moisture content of the natural gum can be readily used as a tenable method of reducing the microbial counts. Gums from Acacia seyal var. seyal; Acacia seyal var. fistula, and Acacia polyacantha were tested for their pathological properties”. Diets of rats were substituted with these gums at various levels for twelve weeks. Preliminary data indicated atrophy of the villi and mucosal glands of the intestine. It was also found that some of the rats, fed on diet containing 50% gum from Acacia seyal var. seyal, showed focal lymphocytic infiltration in the hepatic cells and slight fatty changes. Another post-mortem observation was that the caeci were distended with gases. Rats fed with diets containing various levels of gum from Acacia senegal var. senegal indicated no pathological effects, thus confirming the clinical safety of such gum from Acacia senegal. This study confirms earlier findingsz3. A studyz4of raw gums from different Acacia species of Sudan for their taxonomic classification showed that these Acacia species could be divided into two main groups. Out of the thirteen gums tested, all but one fell into one group. The species falling in the large group showed presence of tannins in their gums. The tannin content ranged between 0.03 to 1.63%. The only gum that did not show presence of tannin was the gum from Acacia senegal var. senegal, thus distinguishing itself distinctly and distantly from other Acacia gums (Table 11). This finding is of significant importance when considering gums as food additives. It is established that tannins are anti-nutritional. Bleaching can remove the colour and off-flavour of gum from Acacia seyal. It can thus become very easy for some to pass off bleached gum from Acacia seyal as gum from Acacia senegal. The property of specific optical rotation can be used here very effectively to identify the gum. The base test, i.e. the addition of NaOH to a 20% solution of the gum can reveal if the gum is from Acacia senegal or from other Acacia species. Bleaching leads to the problem of residues of the bleaching agent staying behind in the bleached gum. Such residues of bleaching agents are harmful.

Gum Arabic

- The Future

The Sudan has been and is still the single largest producer of Gum Arabic (gum from Acacia senegal, var. senegal). The country supplies about 75% of the world needs of Gum Arabic. The other significant suppliers of Gum Arabic are Nigeria Senegal and Mauritania, each with around 5% share of the total production. The other producers, albeit small are Ethiopia, Chad, Mali and Niger. It is of prime significance to note that all of these countries fall within the agro-climatically suitable, traditional and well defined “gum producing belt” of Africa. However, this traditional gum belt has been subjected to devastating droughts, thus affecting Gum Arabic p r o d ~ ~ t i Fig. on~ ~~~ 3. ~ This has finally resulted in high

Polysaccharide Characterisation

41

Acacia species Tannin content A. senegal var. senegal 0.00 A. seyalvar. seyal 0.11 A. seyalvar. fistula 0.09 A. laeta 0.13 A. nilotica subsp. nilotica 0.05 A. nilotica subsp. tomentosa 0.14 A. drepanolobium 0.03 A. gerrardi 0.16 A. polyacantha 0.16 A. tortillis subsp. raddiana ~ _ ._.. -1.63A. mellifera 0.10 A. echrenbergiana 0.17 A. seiberana var. seiberana 0.08 ~

~

~

~

~~~

(YO)

~~

I

--

Production (000 Tons)

-Exports(000

Tons)

Fig. 3 Production and exports of Gum Arabic (Acacia senegal var. senegal) from Sudan (Source: Gum Arabic Company Ltd., Khartoum, Sudan 1998; National Forest Corporation, Khartoum, Sudan 1998)

prices of Gum Arabic. For example, in 1972, raw Gum Arabic was sold at USD 797 per '000 Kg (CIF). While in 1975, it fetched a price of USD 1,625 per '000 Kg (FOB Port Sudan) - an increase of 110% on FOB basis. The years 1973-1974 were years of severe drought in the Sahel region of Africa. Only 12,000 to 15,000 tons of Gum Arabic was produced in the drought years. Consumption of Gum Arabic until then was estimated at about 70,000 tons per annum. This kind of fluctuation in supply and prices forced the end users to come up with alternatives of Gum Arabic. Many large users re-formulated their products and replaced old machinery used for handling Gum Arabic by modern machinery to enable modified starches to be used in their production process.

48

Gums and Stabilisers for the Food Industry I0

The production of Gum Arabic picked up again in 1976/77 and such high production continued till 1982. lnspite of the higher production, sales of Gum Arabic never exceeded 40,000 tons per annum during this period. This trend can be attributed to the fact that the Sudan failed to reduce the high prices quickly enough, as a result, sales of modified starch continued to increase. The Sahel region was again subjected to severe droughts in 1983-84. Production was severely affected, and export of Gum Arabic from the Sudan in 1985 amounted to only 12,600 tons, at an average price of USD 1,600 per ton (FOB, Port Sudan). Again due to droughts in 1993-94, prices of Gum Arabic increased. In 1995, the average price was estimated at USD 4,200 per ton (FOB, Port Sudan) - an increase of 163% over 1985 prices. During these periods of limited availability, many more users modified their machinery to handle modified starches. In May 1996, the Sudan reduced the price of Gum Arabic to USD 2,200 per ton. But this had no positive effect on the sales of Gum Arabic from Sudan. Exports of Gum Arabic from the Sudan averaged around 10,000 tons per annum. At the same time, it is interesting to note that in 1997, USA, UK, Germany, France, Italy and some other Western European countries together imported about 23,000 tons of Gum Arabic from different sources including the Sudan. Official figures of the Sudan show that only about 11,000 tons of Gum Arabic was exported in 1997, while the total production in Sudan was estimated to be around 35,000 tons in 1997. This is a clear indication that a large quantity of Gum Arabic is smuggled out of the Sudan and reaches the destination countries, primarily through its neighbors. It has already been indicated that Sudan alone produces more than 70% of the world Gum Arabic. This trend is also an indication that modified starches cannot replace Gum Arabic in certain products, due to the superior intrinsic qualities of Gum Arabic. It cannot be claimed that even a near perfect substitute has been found for Gum Arabic - such are the properties and qualities of Gum Arabic. During the early 199Os, some foreign investors realized the potential of Gum Arabic, and decided to put up commercial plantations. As is known, till date, most of the Gum Arabic comes from trees growing naturally, as there were no commercial plantations. The author today is representing a foreign investment company in the Sudan whose main agricultural interest is in Gum Arabic. The company has taken large concessions (0.25 million acres) in the traditional Gum Arabic growing areas (Kordufan and Damazine) of Sudan, Till date, about 47,000 acres have been planted to pure stands of Acacia senegal, var. senegal (Gum Arabic). The company plans to plant 150,000 acres of Gum Arabic by the year 2003. The commercial production is expected from the year 2000. The company is also in the process of setting up Gum Arabic plantations in other African countries, but which fall within the "gum producing belt" of Africa, thus diversifying the production base.

Polysaccharide Characterisation

49

There are a few more investors involved in setting up pure stands of Gum Arabic in the Sudan, as well as in other African countries. It is estimated that the total production of Gum Arabic would be more than 175,000 tons starting from year 2010, due to these commercial plantations. This figure far exceeds the total world consumption of 70,000 tons/annum of the 1960s and 1970s. Production of Gum Arabic from Gandil plantations alone is expected to be around 35,000 tons per annum. Different laboratories from around the world have tested gum from Gandil Gum Arabic plantations, and the analysis results conform to the quality specifications set by CODEX. As Gandil gum is seasonly tested for identity and purity, figures 4a, 4b and 4c show identical C13NMR spectra of gum collected from lst, 2"d and 3'd picking respectively. After spending a fair amount of time on experiments, Gandil Company has developed a technique of direct seeding, thus avoiding nurseries, transplanting etc. This has resulted in huge cost savings. About 81% of the costs of setting up a plantation have been reduced using this technique. To ensure increased survival and better vigour for the plants, deep ploughing and water harvesting techniques were employed, to make maximum use of the rainfall. As a result of water harvesting, 96% survival was obtained in the field, compared to 79% survival where seeds were used under normal conditions. The general health of the plants was better and growth (length of stem and branches) was nearly 2.9 times that of trees of the same age but growing under normal conditions. In addition, weeding is also carried out on time regularly, thereby avoiding any competition for the Gum Arabic trees. It can thus be expected that yields from individual trees would also be high compared to trees in the wild. The overall result would be low cost of production of Gum Arabic. The detailed engineering for a Gum Arabic processing plant (cleaning, kibbling, spray drying) is near completion. This processing plant is expected to go on stream from 2001. While a European team is involved in designing the plant, machinery are expected to be sourced from the US and Europe, and some from Japan. It is worth mentioning here that till 1998, the Gum Arabic Company of Sudan was the only unit allowed to export Gum Arabic from the Sudan. However, since early 1999, the Government of Sudan has practically liberalized the Gum Arabic trade.

To sum up, facing problems of irregular supply of and unstable can be avoided by the large scale production of Gum Arabic in association with different companies like Gandil etc. Thus sufficient and cost effective supplies of pure and good quality Gum Arabic will be assured from Sudan.

Gums and Stabilisers for the Food Industry 10

50

(4

.I.

I.

.

11.

..

mk*r*;m .h

-oh

Y

A

Fig. 4 13C NMR spectra of Gum Arabic (Acacia senegal) from Gandil Plantation, season 1999: (a) 1st picking; (b) 2nd picking; (c) 3rd picking

Polysaccharide Characterisation

51

REFERENCES 1.

Anderson, D.M.W., Dea, I.C.M., Karamalla, K.A. and Smith, J.F. (1968). Carbohydr. ReS. 6, 97-103.

2.

Anderson, D.M.W., Bridegman, M.M.E., Farquhar, J.G.K. and McNab, C.G.A., (1983). The International Tree Crom 2,245-254.

3.

Anderson, D.M.W. (1986). Food Additives and Contaminants 3,225.

4.

Anderson, D.M.W., Brown Douglas, D.M., Morrison, N.A. and Weiping, W. (1990). Food Additives and Contaminants. 7,303-321.

5.

Lamport, D.T.A., Qui, Wand Fong, C. (1991). Plant physiol. 96,845-855.

6.

Osman, N.E., Williams, P.A., Menzies, A.R. and Phillips, G.O. (1993). J.Agric. Food Chem. 41,71-77.

7.

Osman, M.E., Menzies, A.R., Martin, B.A., Williams, P.A., Phillips, G.O. and Baldwin, T.C. (1995). Phytochemistry, 38,409.

8.

Karamalla, K.A., Siddiq, N.E., and Osman, M.E. (1 998). Food Hvdrocolloids 12,373- 378.

9.

Anderson, D.M.W., Millar, J.R.A. and Weiping, W. (1 991). Biochemistry Systamatic and Ecology 19,447-452.

10.

Jurasek, P., Kosik, M. and Phillips, G.O. (1993a). Food Hydrocolloids 7, 7385.

11.

Jurasek, P., Kosik, M. and Phillips, G.O. (1993b). Food Hvdrocolloids 7, 255280.

12.

Ishaaq, K.E.A. (1 977). M.Sc. Thesis, University ofKhartoum.

13.

Anderson, D.M.W. and Herbich. (1 963). J. Chem. Soc. (1963) 1, 1-6.

14.

Anderson, D.M.W. and Karamalla (1966 b). J.Chem. Soc. (C), 8, 762.

15.

The Gum Arabic Company Ltd. (1994).

16.

Hassan, E.A. (1 994). Unpublished results.

17.

Menzies, A.R., Osman, M.E., Malik A.A. and Baldwin T.C. Food Additives and Contaminants (1996). 13,991-999.

Gums and Stabilisersfor the Food Industry 10

52

18.

Astaud, J. (1 982). Analust, 10, 124.

19.

Karamalla (1 999). Unpublished results.

20.

Leonard, T. (1 999). Personel Communication.

21.

Osman, I.M. (1 998). M.Sc. Thesis, University of Khartoum.

22.

Wahbi 1.S. ( 1 998). M.Sc. Thesis, University of Khartoum.

23.

Anderson, D.M.W. (1986). Food Additives and Contaminants 3,225-230.

24.

Zahir, A.S. (1 998). M.Sc. Thesis, University ofKhartoum.

25.

Forest National Corporation (1 998). Khartoum, Sudan.

26.

The Gum Arabic Co. Ltd, (1998). Khartoum, Sudan.

IDENTIFICATION OF GUM ARABIC USING PAGE AND IEF

S. Motlagh', P. Ravines', Q. Ma2 and F. Jaksch3 'Bahh'i World Centre, PO Box 155, Haifa, 31001 Israel 2Archidex,431 Amapola Ave., Torrance, CA 90501 USA 'Phenomenex, 2320 W. 205" St., Torrance, CA 90501 USA

1 INTRODUCTION

The Joint Expert Committee for Food Additives of the Food and Agriculture Organization (FAO) of the United Nations' defines gum arabic as the gum exudate from Acacia senegal (L. Willdenow) or closely related Acacia species (family Leguminosae). Around 1 100 species of Acacia have been identified2. A. senegal has four varieties3:senegal, Kerensis SchweinJ rostata Brenan, and leiorhachis Brenan. Gum arabic is found in sub-Saharan Africa. The Kordofan province of Sudan is the largest producer and has the best quality gum arabic. Gum oozes from stems and branches of the tree when it is subjected to heat, drought, insect attack or wounding as a viscous sticky liquid that dries to form hard glassy nodules that are collected and sorted by color and size. Gum arabic has been used for thousands of years. In ancient Egypt it was used in the embalming of mummies and in preparing colors for painting4. Today it is used in the food industry for emulsification, texture control, stabilization and flavor or color encapsulation. It is also used in the textile, printing and pharmaceutical industries. Because of its extensive use in food products, and in order to eliminate adulteration of gum arabic by gums which do not fall under the FA0 definition, the authentication of gum arabic is of great interest to the food industry. Historically gum arabic has been used as a binder and medium for colorants and inks. It is still the medium of choice in watercolors and water-soluble carbon-based black inks. It's chemical nature and long tern stability are of interest because libraries, archives and museums contain large collections of old gum in documents, manuscripts and works of art on paper, such as watercolors, in various degrees of preservation. As in the food industry, the characterizationof gum arabic is of interest to the museum and library community in charge of preserving valuable artistic and historic collections of books and works of art. Gum arabic is a mixture of complex carbohydrates. Gum arabic interacts with Yariv reagent5and is therefore classified as an arabinogalactan-protein complex (AGP). The gum has a highly variable composition6and nodules collected from the same tree often have different chemical signatures. Gel filtration chromatographyhas been the most commonly used technique for the separation of the components of the whole gum and has shown' the molecular weights of the complex carbohydrate molecules to be between 1 . 5 ~ 1 and 0 ~ 4x106. Hydrophobic interaction chromatography (HIC), which has also often been used to fractionate gum

Gums and Stabilisers for the Food Industry 10

54

arabic, has demonstrated*that the protein content of the various complex carbohydrate molecules composing the whole gum is between 0.7% and 36.8%. Weak anion exchange chromatography using a DEAE-cellulose stationary phase has also been tested'. Several groups have used electrophoresis with varying degrees of success. In a communication, Lewis and Smith" reported a 10 hour-long electrophoretic separation of gum arabic obtained from A. Senegal using glass-fiber paper in conjunction with a 2 M NaOH buffer; however, no mention was made of the possibility of hydrolysis of the gum by the base. Pechanek, et al." encountered problems with the polyacrylamide gel electrophoresis (PAGE) separation of gum arabic. Instead, they used cellulose acetate membranes with a boric acid buffer and obtained one band. Most recently, Osman, et al." performed a SDS-PAGE separation of three HIC fractions of A. Senegal gum and obtained protein bands with molecular masses in the 30 000 to 200 000 range. We have improved the resolution of the SDS-PAGE separation of gum arabic so as to allow it to be used for the authentication and identification of the gum. We have also performed an isoelectric focusing (IEF) separation of gum arabic for the same purpose. 2 EXPERIMENTAL All instrumentation was obtained from Bio-Rad. All reagents were electrophoresis grade and were purchased from Bio-Rad except where otherwise noted. Staining was done using the Bio-Rad silver stain kit. Sarah Taylor of the National Resources Institute, University of Greenwich, UK, provided gum from A . Jischeri (Tanzania), A. nilotica (Zimbabwe) and A. senegal (Uganda). Tessa Nolan-Neylan provided gum from A. senegal (Kenya). C. Nicholson of A.F. Suter & Co., London, provided Sudanese commercial grade gum arabic. The spraydried gum arabic labeled Fluka was purchased from Sigma.

2.1 Polyacrylamide Gel Electrophoresis 2. I . 1 Native PAGE. Buffer composition: 25 mM Tris, 192 mM glycine. Samp!e buffer composition: 62.5 mM Tris-HC1 (ICN), 40 % glycerol (BDH), 0.01 % bromophenol blue. Potential: 200 V (constant). 2.1.2 SDS-PAGE. Buffer composition: 25 mM Tris, 192 mM glycine, 0.1 % SDS. Sample buffer composition: 62.5 mM Tris-HC1,2 % SDS, 25 % glycerol, 0.01 % bromophenol blue, 4% 2-mercaptoethanol, heat for 4 min at 95 O C . Standards: Bio-Rad low and high range silver stain SDS-PAGE standard. Potential: 200 V (constant). 2.2 Isoelectric Focusing 2.2.I Native IEF. Bio-Rad IEF cathode, anode and sample buffers. Standard: BioRad IEF standard. Potential: 1 h at 100 V, 1 h at 250 V, 30 min at 500 V. 2.2.2 Denaturing IEF. Bio-Rad IEF cathode and anode buffers. Sample buffer composition: 4 M urea, 0.012 YOpH 3.5-10 ampholyte solution, 1 % Triton X-100 (Sigma), 0.5 YO2-mercaptoethanol, 0.02 YObromophenol blue. Standard: Bio-Rad 2-D SDS-PAGE standard. Potential: 30 min at 150 V, 2.5 h at 200 V.

Polysaccharide Characrerisarion

55

3 RESULTS AND DISCUSSION 3.1 Polyacrylamide Gel Electrophoresis

We were unable to separate gum arabic using native conditions. Separationsof the gum were performed under denaturing conditions (SDS-PAGE) using uniform and gradient gels ranging in concentration from 5% to 20%. The best separations were obtained using uniform 15% gels. In an effort to obtain more information about the sub-unit structure of the proteinaceous part of the molecules by identifying any inter-chain disulfide linkages, parallel samples of gum arabic with and without 2-mercaptoethanolwere run. There were a few minor differences between the two sets of results, indicating the presence of a small number of disulfide linkages. Figure 1 shows a typical SDS-PAGE separation of samples of gum arabic using a 15% Tris-glycine gel. Table 1 displays the molecular weights of the bands. Note that all molecular weights were obtained from Ferguson plots of protein standards. The numbers obtained would be accurate only for protein bands. We suspect the presence of both proteins and glycoproteins. Glycoprotein bands would have to be standardized against glycoprotein standards, since highly glycosylated glycoprotein molecules would have different shapes from protein molecules.

2 2

Figure 1 Separation of several Acacia gums using SDS-PAGE

Protein molecules form narrow bands while glycoprotein and carbohydrate populations form broad bands. The following conclusions can be drawn from Figure 1 : The commercial grade Sudanese gum arabic and the spray-dried powder produced by Fluka

56

Gums and Stabilisers for the Food Industry 10

share bands at 21 kd and 28 kd. This indicates that the spray-dried gum may have been prepared from Sudanese raw gum. In contrast to the Sudanese gum arabic, both gum samples obtained from A. senegal have very few bands. This may indicate that the latter two samples were obtained from a different variety of A. senegal than the Sudanese gum. Table 1 Molecular weights (in kilodaltons, using protein standards) of Acacia gum fractions obtained by SDS-PAGE using 15% gel concentration

In order to determine if any glycoprotein bands were migrating abnormally and to confirm the homogeneity of the bands obtained on the 15% gel, the above separation was repeated using a 12% gel concentration. The results were essentially similar to those obtained using the 15% gel concentration. This implies that, unlike other glycoproteins that have been separated using PAGEI3, the bands obtained from gum arabic migrate normally. Also, unlike results obtained by other groups,14no significant advantage was seen in the use of gradient gels. A substantial fraction of gum is seen to remain in the stacking gel, this is especially significant in the case of the spray-dried gum and the Sudanese gum. This indicates that both of these samples have higher molecular weight fractions whose molecular weights are not measured. When lower density gels are used, these fractions enter the gel as very diffuse bands that are difficult to quantitate. This technique separates both protein and glycoproteins well. Based on our limited number of samples, it appears that the separations are species specific, i.e. the gums can be identified by their protein and glycoprotein bands. If this were the case, then this technique would be a simple one to use for the identification and authentication of Acacia gums. A comparison between our results and those obtained by Osman, et a1.I2shows that our range of molecular weights of the compounds that are mobile under electrophoretic conditions (i.e. proteins and glycoproteins) are comparable to their range.

3.2 Isoelectric Focusing A gel with a pH gradient of 3-10 was used to perform an IEF separation of three samples of gum arabic (Figure 2). The A. senegal (Uganda) gum sample was separated into three narrow bands at pH 4.4,4.7 and 4.8. The Sudanese and spray-dried samples have poorly focused bands at around pH 7. The Sudanese sample has bands at around pH 3.5 and 4, while the spray-dried sample has a band at around pH 4. Note that the A. senegal sample exhibits one band under SDS-PAGE conditions. Under IEF conditions, however, this sample separates into three bands. This implies that the single SDS-PAGE band has, as shown by IEF, three components of similar molecular weight but differing

PolysaccharideCharacterisation

isoelectric points. 2D SDS-PAGE would confirm this. This also demonstrates the complimentary nature of these two electrophoretic techniques. Preliminary results of denatured IEF experiments are similar to those obtained by native IEF.

Figure 2 Separation of three samples of gum arabic using IEF 4 ACKNOWLEDGEMENTS

S Motlagh and P Ravines would like to express our gratitude to the following for their support and encouragement: The Universal House of Justice, governing body of the B M i world community; Farshad Mahjoor, president of Phenomenex; Faizy Ahmed, chief of Research and Development, and his staff at Phenomenex; Arie Admon, Biochemistry Department, Technion - Israel Institute of Technology; and our families. References 1. 2. 3. 4.

5. 6.

JEFCA-FAO/WHO, ‘Food Nutrition Paper No. 52, Addendum 3’, FAO, Rome, 1995, p. 83. D.M.W. Anderson and N.A. Momson, Food Hydrocolloids, 1989,3,57. J.J.W. Coppen, ‘Gums, Resins and Latex of Plant Origin’, FAO, Rome, 1995. Herodotus, ‘The History of Herodotus’, Encyclopedia Britannica Inc, Chicago, 1990, pp. 65-66. Y. Akiyama, S.Eda and K. Kato, Agric. Biol. Chem., 1984,48,235. A.M. Islam, G.O. Phillips, A. Sljivo, M.J. Snowden and P.A. Williams, Food Hydrocolloids, 1997,11,493.

58

7. 8. 9. 10. 11. 12. 13.

14.

Gums and Stabilisersfor the Food Industry 10

O.H.M. Idris, P.A. Williams and G.O. Phillips, Food Hydrocolloids, 1998,12,379. A.K. Ray, P.B. Bird, G.A. Iacobucci and B.C. Clark Jr., Food Hydrocolloids, 1995, 9, 123. M.E. Osman, A.R. Menzies, B.A. Martin, P.A. Williams, G.O. Phillips and T.C. Baldwin, Phytochemistry, 1995,38,409. B.A. Lewis and F. Smith, J Am. Chem. SOC.,1957,79,3929. U. Pechanek, G. Blaicher, W. Pfannhauser and H. Woidich, J Assoc. 08 Anal. Chem., 1982,65,745. M.E. Osman, A.R. Menzies, P.A. Williams, G.O. Phillips and T.C. Baldwin, Carbohydr.Res., 1993,246,303. Q. Shi and G. Jackowski, in ‘Gel Electro horesis of Proteins’, ed. B.D.Hames, Oxford University Press, Oxford, UK, 3r redn., 1998, p. 32. A.T. Andrews, ‘Electrophoresis:Theory, Techniques, and Biochemical and Clinical Applications’, Clarendon Press, Oxford, UK, 2”dedn., 1986, p. 133.

RELEVANT STRUCTURAL FEATURES OF TEE GUM FROM Entemlobium cyclocurpum

G. Le6n de Pinto, M. Martinez, 0. Beltrsln, C. Clamens, F. Rinc6n and L. Sanabria. Centro de Investigaciones en Quimica de 10s Productos Naturales, Facultad de Humanidades y Educaci6n. Universidad del Zulia. Maracaibo. Venezuela. E-Mail: [email protected].

1 INTRODUCTION

Entemlobium cyclocurprun, a large tree found in the hottest area of Venezuela,' exudes gum easily, three days after the incision is made at trunk level. The average yield is very high (36 g/specimen/week).2The solubility of this gum is not as high as that from Acacia senegul but it has been observed that this parameter increases when the gum is collected

as soon as is exuded.2Analytical data of gum exudates from five Venezuelan specimens of E. cyclociupum showed that the limiting viscosity number (100 mLg-') is comparable to that exhibited for some Combretaceae gums but higher than that reported for any Acuciu gum studied so far.'". The functionality of E. cyclocurpm gum as a stabilising agent for

emulsions in pharmaceutical preparations has been demonstrated'. A comparative study showed that an aqueous solution of this gum (0.75%) has the same rheological properties of that from ~ c u c i usenegd (20%)~' Chemical and I3C-NMR studies of E. qclocurprtm gum and its degradation products have been published.8 Some relevant structural features of the plysaccharide, isolated from this gum, have beem reexamined and are discussed in this work. 2 RESULTS AND DISCUSSION

The plysaccharide, isolated from Entetolobium qclocutpum, contains galactose, arabinose, glucuronic acid and its 4-0-methyl derivative, Table 1.

Gums and Stabilisersfor the Food Industry 10

60

Table 1 Sugar composition of E. eyelocarpurn gum and its degradationproducts, Sugarsa (?A) Yield (?A)

Gal

Ara

Rha

UronicAcids

Original gum

60

49

20

10

21

Autohydrolysis polymer

80

53

19

7

21

Degraded gum A

31

67

6

27

Degraded gum B

1

100

Polysaccharide I

3

69

12

19

Polysaccharide 11

5

63

23

14

Polysacchande I11

8

100

Polymer

*Corrected for moisture.

Methylation analysis, Table 2, of the original polysaccharide shows the presence of terminal and 3-O-arabinofiuanoseresidues, terminal and 3-O-arabinopyranose, 3-0-, 6-0-, 4-0- and 3,6-di-O- substituted galactose, uronic acids and rhamnose as terminal residues. Partial hydrolysis of the polysaccharide led the isolation of the oligomer, Rgal 0.1 I(b), 0.66 (a), [ a ] ~ +94", : which was characterized by acid hydrolysis, methylation acid) analysis and 13C-NMRspectroscopy as 4-O-(4-O-methyl-a-D-glucopyranosyluroNc p-D-galactose. The separation and identification of neutral and acidic components, by column chromatography, confirmed that galactose, arabinose and rhamnose were present in the polysaccharide structure. The acidic components were characterized as 4-0-(4-0-methylacid)-Da-D-glucopyranosylunic acid)-D-galactose and 6-O-(~-D-glucopyranosyluronic galactose. Monitoring of the preparation of the autohydrolysis polymer and degraded gum A led to the removal of rhamnose and arabinose residues. The susceptibility of these

sugars to acid hydrolysis has been observed previou~ly.~*'~ Degraded gum B, obtained from periodate oxidation of degraded gum A, was a p-(1+3) galactan. The main structural features of this gum correspond to an arabinogalactan type 11." A series of three successive Smithdegradation products was prepared. The

polysaccharide I11 showed the same structure of degraded gum B. According to these results, the arabinose sidechains wereup to three units long, comparable to those

61

Polysaccharide Characterisation

Table 2 Methylation analysis of E. cyclocarpum gum Methyl ethers”

T (min)

Linkage

2,3,4-Me~Rha

0.46

L-Rhap-(l+

2,3,5-Me3-L-Ara

0.67

L-Araf-( 1-+

2,3,4-Me3-L-Ara

0.85

L-Amp-( 1+

2,5-Mez-L-Ara

1.26; (2.20)

+3) L-Amf-( 1+

2,4-Mez-L-Ara

(2.20); 2.37

+3) L-Arap-( 1-+

2,3,4,6-Me4-D-Gal

1.65

Galp-( 1-+

2,3,6-Me3-D-Gal

2.50; (3.60) (4.00)

d)-GaIp-(l+

2,4,6-Me3-D-Gal

(3.60) (4.00)

-+3)-Wp-( 1+

2,3,4-Me3-D-Gal

5.67; 6.00

4j)-Galp-(l-+

2,4-Me2-D-Gal

12.22; 13.80

+3,6)-Galp-( 1+

2,3,4-Me3-D-GlcAb

(2.20); 2.75

GlcA-( I+

qelative to methyl 2,3,4,fAetra-O-methyl-fLDglucopyranoside. bAsmethyl ester methyl glycoside.

observed for Acacia xanthophloea: Sw&?tenia mahagoni

l2

and Melicocca bijuga

lo

gums, but shorter than those reported for A. sawgal, where five degradations were

required to obtain the core of the structure.l3 Removal of uronic acid residues from the plysaccharides I and I1 was difficult, which may be related either to hydrogen linkages that involved these sugar residues, which work against periodate oxidation, andor some of them are internal residues. Signal assignments of I3C-NMR spectra was made on the basis of chemical ” ~ ”spectrum ~ of polysaccharide 11, Figure 1, shows the evidence and previous r e ~ u l t s . ~ The resonances due to a (1+3>p-galactan,

4-0-methyl-a-D-glucunic acid and 3-0-p-L

arabinopyanosyl residues, Tables 3-5. The peak at 79.0 ppm, assigned to 4-0-p-Dgalactose

residue^,'^ is according to methylation analysis and the characterization of the

acid)-D-galactose, which aldobiouronic acid 4-0-(4-0-methyl-a-D-glucopyranosyluroNc was isolated from the original gum. These results suggest that the 4-0-methyla-D-

glucuronic acid is present as internal residues in the structure of the gum polysaccharide.

62

Gums and Stabilisers for the Food Industry 10

Gd

I' I, I

Figure 1 I3C-NMR spectrum ofpolysaccaride I1 There are also resonances of 6-0-linked galactose residues, Table 3, The signals attributed to C-1 (100.5 ppm) and to C-2 (66.50 ppm) of 3-0-P-L-arabinopyranose residues,I6are supported by methylation analysis, Table 2. Signals of a-L-arabinohranose residues were not observed in this spectrum, which suggest the removal of these residues in the preparation of polysaccharide II. The spectrum of plysaccharide I, shows, in addition to the resonances that appeared in the previous spectrum, Figure I , the unequivocal signals due to C-1 (109.2

ppm), C-2 (81.2 ppm), C-3 (76.5 ppm), C-4 (83.8 ppm) and C-5 (61.2 ppm) of a-Larabinofuranose also observed.

residue^.'.'^

The resonances due to p-D-glucuronic acid residues were

63

Polysaccharide Characterisation

Table 3 13C-NMR data" of PD-galactose residues in E. cVcrocarprtrn gum and its degradationproducts. T y p of Iinkage

PoIymer

C-1

C-2

C-3

C-4

C-5

C-6

71.4

81.9

69.1

74.9

60.8

71.0

81.0

69.5

75.2

61.2

103.80 71.9

81.4

68.9

74.2

61.1

+6)P-D-Galp-(l+b~c PolysaccharideI1 103.3 70.1

72.4

67.7

72.4

68.5

DegradedgumA 102.6 70.3

72.6

66.9

72.9

68.5

73.3

67.3

73.3

68.9

4)p-D-Galp(l+h*c Polysaccharide11 104.2 DegadedgumA Originalgum

Original gum

103.5

102.9

70.2

'Values relative to 1,4-dioxane as internal reference (6 66.7 ppm). bRef.[17]. The same signals were observed in the spectrum of polysaccharideI.

The spectrum of the original gum, more complex than the previous ones, contains the resonances due to 3-0-, 6-0- and 4-O-P-D-galactose, terminal and 3-0-a-LP-D-glucuronic acid residues and its arabinofuranose residues, 3-O-P-L-arabinopyranose, 4-0-methyl derivative, Tables 3-5.

Table 4 "C-NMR data" of uronic acid residues in E. cyclocatpum gum and its degradationproducts. _ _ _ _ _ _ _ ~ ~ ~ ~

Type of Iinkage

Polymer

C-1

C-2 C-3 C-4 C-5

60.30

Degradedgum A 100.5 72.3 73.5 82.0 70.7

-

Originalgum

-

60.7

4-0-Me-a-D-Gl~A(l-+~"PolysaccharideI1 100.3 72.4

P-D-GlcA-( l-+b,c

-

C.'-6 I-0-Me

81.9 70.9

100.6 71.9 73.4 81.4 70.2

60.4

D e g r a d e d ~ A 104.0 75.8 76.5 73.5 77.1 175.4 5

-

Originalgum

-

104.6 75.5 76.1 73.3 76.1 175.1

'Values relative to 1.4dioxane as internal reference (6 66.7 ppm). k e f [14]. The Same signals wen observed in the spectrum of polysaccharideI.

64

Gums and Stabilisers for the Food Industry 10

Table 5 '3c'-NMR dutu" of L-urubinose residues in E. cyclocatpum gum and its degrudution products.

c-I

('2

(13

c-4

+3) P-L-Arap(ljb3' Polysaccharide I1

100.5

66.5

74.6

-

Degradedgum A

101.4

66.9

73.6

65.6

61.8

Original gum

100.6

67.7

74.2

-

62.3

a-L-Araf( 1+b2c

Originalgum

109.2

81.4

76.1

84.1

61.1

-+3)a-L-Araf(ljb

Orignal gum

108.5

80.8

83.5

83.5

62.3

Type of Iinkuge

Polymer

c-5

"Valuesrelative to 1.4-dioxane as internal reference (6 66.7 ppm). k e f [8].The same signals were observed in the spectrum of polysaccharideI.

Degraded gum A, obtained by mild acid hydrolysis of the original gum, shows a spectrum, Figure 2A, that contains the same resonances described in the spectrum of the original gum, except those attributed to a-L-arabinokanose residues. Expansion of the anomeric region of that spectrum, Figure 2B, shows the presence of terminal reductor sugars (92.0 ppm),17 a-D-glucuronic acid (98.3 ppm),17 4-0-methyl-a-D-glucunic acid (100.5 ppm)," 3-0-P-L-arabinopyranosylresidues (101.4 ppm)? 3-0- and 6-0-galactose

residues (103.5, 102.6 ppm)8,17and 0-D-glucuronic acid residues (104.0 ppm).14 Dearabinosylation led to a better resolution of the signals due to 3-0-p-Larabinopyranosyl and uronic acid residues. The peak of high intensity at 79.0 ppm, described previously, is resolved in two peaks (79.0,78.2 ppm). Comparison of the spectral data of the original polysaccharide and its degradation products confirmed that the gum is an arabinogalactan Type 11." The backbone structure is a p( 1+3)-galactan with some 6-0-galactose residues. Side-chains are constituted by 30-L-arabinose (furanosyl and pyranosyl) residues, up to three units long. p-D-glucuronic is probably as terminal residues while its 4-0-methyl derivative may be as internal residues. The properties and the structural features of the studied gum indicate that it may be a new source of gum for industrial application.

Polysaccharide Characterisation

65 G-6!

I

morneric linkages

Ya (

'

180

I

'

I

160

'

140

I

'

120

I

'

I

100

'

80

'

60

PPm

Figure 2A 1*7C-NMR spectrum of degraded gum A

p-Ap-1'

a-U-1'

P-u-1'

PPm Figure 2B Expansion of the anomeric region of 13C-NMRspectrum of degraded gum A

66

Gums and Srabilisers for the Food lndusrry 10

3 EXPERIMENTAL 3.1 Origin and Purification of Gum Samples. Gum from E. qclocurpum (Jacq.) Griseb was collected by the authors in MarchApril, 1990, from a tree growing in Maracaibo, Venezuela, South America. Identification

of the voucher specimen was confirmed by Dr.George S. Bilntin, a botanical taxonomist of the Botanical Garden, Maracaibo. The dissolution of gum samples, collected two weeks

after the injury was made, was carried out in a day. The solution clear brown in colour,

was passed through muslin and Whatman No41 paper and dialysed against running H20 for two days. The gum was recovered by freeze-drying (29.4 g 56%). 3.2 General Methods. General standard methods of gum analysis were used. 9~'0 The solvent systems used in paper chromatographicwere as described previou~ly.~*'~ Gas chromatographic was carried out using flame detector ionization with NZflow rates of 40 mL.min" . The glass column (168 x 0.57 cm) used contained 10% (w/w) polyethylene glycol adipate on Chromosorb WHD at 190OC. Retention times are quoted relative to that of methyl-2,3,4,6-

tetra-0-methyl-P-D-glucopyranosidefor the methyl ethers. %NMR

data points were

accumulated overnight at 37°C with complete proton 4ecoupling. Spectra were calibrated by the addition of 1,4 dioxane to the samples. The polysaccharides (100-200 mg) were dissolved in 40 (1 mL). The procedures for partial and total hydrolyses, quantitative analysis of sugars, identification of aldobiouronic acids and methylations of the different oligosaccharides and polysaccharides are described previo~sly.~~'~ 3 3 Autohydrolysis Experiments.

A solution of purified sample (5%) was heated at 100" for 120 h; portions (10 mL) were withdrawn at various intervals, concentrated and analysed by PC. The polymer was isolated by freeze-drying. 3.4 Preparation and Examination of Degraded Gums A and B.

Unless otherwise stated the experimental procedures used for the preparation and examination of degraded gums A and B were the same as those described previously. 9~10

Polysaccharide Characterisation

67

Degraded gum A (2.4 g) was obtained from purified gum (8 g). Preliminary small-scale experiments showed that 96 h were required for the preparation of degraded gum B. 3.5 Smithdegradation Studies.

A series of three sequential Smithdegradations was performed with the pure gum

as starting material (23.3 g) to obtain polysaccharide I (714.7 mg). Polysaccharide I (2.77 g) gave polysaccharide I1 (142 mg) and the latter (660 mg) yielded polysaccharide 111 (50

mg). Experimental conditions for these degradations were in general, as reported. 9 ~ 1 0The preparation of each polysaccharide was repeated three times in order to check the yields and to have enough sample to complete the Smithdegradation process.

Acknowledgement The authors thank The University of Zulia, Consejo de Desarrollo Cientifico y Humanistic0 (C.O.N.D.E.S.),Maracaibo, Venezuela, for financial support.

Referenees 1. L.C. Guevara, Rev. de Fac. Agronom. UCV, 1974,7,109. 2. G. Lehn de Pinto, M. Martinez, C. Clamens and A. Vera, Rev. Fac. Agronom, La

Plata, 1996,101,51. 3. G. Le6n de Pinto and A. Ludovic de Coredor, Acta Cientijica Venezolana, 1986,37, 92. 4. D. M. W. Anderson and P.C. Bell, Carbohydr.Res., 1977,57,215.

5 . D.M.W. Anderson, M.M.E. Bridgeman and G. M n de Pinto, Phytochemistry, 1984, 23,575. 6. D.M.W. Anderson, J.G.K. Farquhar and M.C.L. Gill, Bot. J. Linn. Soc.,1980,80,79.

7 . G. A d a de Avila, D. Attias de Galindez and G. Le6n de Pinto, Acta Cient9ca

Venezolana 1994,45,71. 8. G. Le6n de Pinto, M. Martinez, A Ludovic de Corredor, C. Rivas and E. Ocando,

Phytochemistry, 1996,37,13 1 1. 9. G. M n de Pinto, C’arbofydr. Res., 1991,220,229. 10. G. Le6n de Pinto,

S. Alvarez, M. Martinez, A. Rojas and E. Leal, Carbohydr. Res.,

1993,239,257. 11. A. Clarke, R. L. Anderson and B.A. Stone, Phytochemistry, 1979,25,2807.

Gums and Stabilisersfor the Food Industry I0

68

12. G. Leon de Pinto, M. Martinez, N. Gonzhlez de Troconis, A.C. Rojas and E.Leal, An.

Quim., 1992,88,157. 13. A.C. Street and D.M.W. Anderson, Talanta, 1983,30,887. 14. G. Leon de Pinto, M. Martinez and C. Rivas, Curbohydr. Res.,1994,260, 17. 15. N.K. Kochetkov, Chem SOC.Rev., 1990,19,29. 16. H. Joao, G. E. Jackson, N. Ravenscroft and A. M. Stephen, Carbohydr. Res., 1988, 176,300. 17. G. Le6n de Pinto, M. Martinez, 0. Gutidrrez de Gotera, C. Rivas and E. Ocando, Ciencia, 1998,6, 191.

Preliminary Study of the Gelling Properties of Polysaccharide Isolated from the Fruit of CordiaAbyssinica

M. A. N. Benhura and C. Katayi - Chidewe Department of Biochemistry, University of Zimbabwe, P.O. Box MP 167, Mount Pleasant, Harare, Zimbabwe

Summary

Cordiu ubyssinicu is a small tree that grows in riverine bush in Southern Africa. The gum, extracted fiom the fiuit of C. ubyssinicu, by precipitation with ethanol from solutions that contained 0.25 M NaCl, gelled when suspended in water at concentrationshigher than 0.15%. Optimum gel formation occurred at neutral pH and 4 "C when solutions were left overnight. Addition of sucrose or monovalent ions at concentrations up to 10% and 50 mM respectively had no effect on gel formation. Addition of calcium ions or EDTA resulted in poorly formed gels. At concentrationsof calcium ions above 20 mM, no gels were formed. The firmness of the gels depended on whether the gum had been ground into a powder or not with unground material producing firmer gels. Upon heating plain gels or those prepared in the presence of sucrose, the gels melted into viscous masses. On cooling, the melted gels formed lumpy masses with proper gels not reforming. 1. Introduction For some time we have been interested in the polysaccharides that are isolated from a variety of local plants (1). Use of the polysaccharides would be a way of using the plants in a sustainable manner. Such an approach would contribute the use of plants as renewable resources and, it is hoped, would discourage the indiscriminate harvesting of trees for firewood and accompanying degradation of the environment. Cordiu ubyssinicu is a small tree that is found in riverine bush in parts of Southern Africa The fruit, a drupe about 10 mm in diameter, turns yellow and sweet on ripening. The pulp of the fiuit, which is edible but not widely consumed by humans is used by rural people as glue for p a p products and tums semi-liquid upon ripening of the fiuit. Preliminary analysis in our laboratory has shown that the components of the acidic fraction of the polysaccharide include galactose, glucose, arabinose, rhamnose and xylose or fucose. When suspended in water, the polysaccharide from Cordiu ubyssinicu forms viscous solutions and appears to have potential for use in thickening or gelling of aqueous solutions, controlling the flow properties of liquid foods and modifying the deformation qualities of semisolid foods. While working with preparationsof the polysaccharide from Cordiu ubyssinicu, it was sometimes observed that suspensionsof the material in water formed gels.

70

Gums and Stabilisers for the Food Industry 10

Gels are continuous, rigid, two phase systems of assorted molecules, that resist flow under pressure and retain a f m distinct structural shape. The gel mass consists of a 3-dimensional network of small discrete particles. Gels are formed when segments of otherwise soluble polysaccharides are brought into association as a result of ionic

Figure 1 Maturefiwit of Cordia abyssinica

interactions, hydrogen bonding, hydrophobic interactions or Van der waals forces. The resulting three-dimensional network traps water molecules, which become more or less immobilised. In the formation of gels, polymer-polymer interactions that favour precipitation are balanced against polymer-solvent interactions that favour the formation of solutions. Enhanced interactions among polymer molecules results as the solubility of the polymer molecules decreases (2). The inter-chain associations may break and reform and are influenced by factors, such as electrolytes, that modify the aqueous environment. The cross linkages that occur in gels involve extensive segments from two or more polymer molecules usually in well defined junction zones (3). The formation of junction zones from the association of polymer segments is enhanced when the availability of water molecules for solvating the polysaccharides is restricted. Factors that reduce the availability of solvating water molecules, therefore, would be expected to enhance gel formation. Polysaccharide chains that contain uronic acid groups may be brought into association by divalent ions such as calcium ions. In this case, gelation is favoured by conditions that promote the charged form of the polyuronic polysaccharide. The strength of polysaccharide gels is influenced by substances that compete with water for binding loci, compete with the polysaccharide for water, alter the pH, interact chemically with the polysaccharide or alter the charge distribution on the polysaccharide. Gelling agents that are currently available for use in the food industry include starches, pectin, agar, alginates, carageenan and gellan gum (4). In addition, a number of thickeners can form gels when combined, as for example, when xanthan gum is mixed with locust bean gum. We decided to further investigate the conditions that are required for the formation of gel by the polysaccharides of C. abyssinica.

Polysaccharide Characterisation

71

2. Experimental 2.1 Extraction of the gum Mature but unripe Cordia abyssinica h i t was obtained h m South Eastern Zimbabwe between December and April. Harvested hit,with their calyces on, were stored at room temperature and processed within 72 hrs. The gum was extracted by squashing the h i t to release the stones, which were covered in pulp. The stones, in a strong plastic or stainless steel container, were agitated with a robust wooden rod during which process the pulp separated fiom the stones as a thick sticky mass. The stones were removed and water added to the sticky mass to make a workable mixture, which was centrifuged in a BHG Hermle ZK 401 centrifuge at 6000 rpm for 30 minutes to remove debris. To the supernatant, sodium chloride was added to make a 0.25 M solution and three volumes of ethanol added to precipitate the polysaccharide. The polysaccharide was dried in a preheated oven set at 100°C. The dry gum was broken up into small pieces, which were stored at room temperature until required.

2.2 Preparation of the gel The dry gum (1.5 g) was suspended in 80 ml of water and the mixture heated to dissolve the polysaccharide. The volume was made up to 100 ml and the solutions left overnight at 4 "C. Alternatively, flakes of dry gum were suspended in 100 ml of water and left overnight at 4 "C. The formation of gels and their quality was assessed by visual inspection. 2.3 Determination of the minimum concentration of polysaccharide required for gelation To determine the minimum concentration of polysaccharide that was required for gelation, up to 0.2 g of the polymer was mixed with 10 ml of water. The mixture was allowed to stand overnight at 4 "C before assessment of gel formation. 2.4 Effect of sucrose Solutions containing up to 10Y0w/v sucrose were prepared by dissolving the sugar in distilled water (1 0 ml). The gum (0.15 g) was suspended in the solutions of sucrose and the suspensions were allowed to stand overnight at 4 "C before assessment of gel foxmation. Alternatively sucrose (1 g) was mixed with up to 0.2 g of polysaccharide and the mixture was ground to a fine powder. The ground mixture was placed into vials and water (10 ml) was added. The mixtures were allowed to stand overnight at 4 "C before assessment of gel formation. 2.5 Effect of metal ions and EDTA The polymer (0.15 g) was suspended in solutions containing calcium, potassium and sodium chloride at concentrations up to 50 mM or 100 mM EDTA in 10 ml of distilled water. The suspensions were allowed to stand overnight at 4 "C before assessment of gel formation.

Gwns and Stabilisers for the Foodlndustry 10

12

2.6 Effect of pH The dry polysaccharide was suspended in 10 ml of buffer. The mixtures were allowed to stand overnight at 4 OC before assessment of gel formation. Citrate buffer was used between pH 3 and 6 and phosphate buffer was used at pH 7 and 8.

2.7 Effect of temperature The polymer (0.15%) was suspended in water (10 ml), in vials that were incubated at the temperatures ranging from 4 OC to 96 'C. The vials were allowed to stand overnight before assessment of gel formation. Gels in which the polysaccharide concentrations ranged from 0.75% to 2.00% w/v were prepared and heated in a boiling water bath until melted. The melted gels were allowed to stand at 4 OC overnight after which the formation of gels and their integrity was assessed. The procedure was repeated using the 1.5% gels containing up to 10% concentration of sucrose. 3. Results and discussion 3.1 Determination of the minimum concentration required for gelation As shown in table 1, the minimum concentration of Cordia abyssinica polysaccharide that was required to form a firm gel was 1.5%. The concentration of polymer at which gel formation occurred is similar to the concentrations of 0.5% to 2% at which several polysaccharides gel (2, 6). At low concentrations of polysaccharide, small gel particles were suspended in bulk liquid. As the concentration increased, the size of the gelled masses increased until, at about 0.5%, a continuous firm gelled mass was observed. Increasing the concentrationof polysaccharide resulted in increased f m e s s of the gels.

Table 1 Effect of concentration on the formation of gels by C. abyssinicapolysacchide Polysaccharide Concentration (??A)

Subjective rating of gelling

0.05

0.2

+ +

0.4

++

0.5

I-

1.o

t+H

1.5

*

2.0

I+f.H-H

Comments

Small gelled areas in bulk liquid Small gelled areas in bulk liquid Large gelled areas in a small amount of Loose gel extending throughout the container Loose gel extendingthroughout the container Firm gel extending throughout the container Firmer gel

Polysaccharidc Characterisation

73

3.2 Effect of addition of sucrose Addition of sucrose at concentrationsup to lo??had no observable effect on the formation of the gels. The observation that the addition of sugar was not required for the formation of a gel by the polysaccharide fiom C. abyssinica would be an advantage in the formulationof gelled low calorie products. 3.3 Effect of calcium, potassium, sodium and EDTA At concentratiops below 1 mM, calcium ions had no effect on the firmness of the gels formed but at higher concentrationsfirmness of the gels decreased as shown in table 2. At concentrationsof calcium ions above 20 mM, no gel was formed at all. Inclusion of the monovalent ions, sodium and potassium, at concentrations up to 20 mM had no effect on the formation of gel. At a concentrationof 50 mM, addition of both ions resulted in slightly less firm gels. With increasing concentrations of EDTA in the gel forming suspension, the firmness of the gels formed decreased as shown in table 3. EDTA was added to chelate the metal ions that were present. When EDTA was present at 50 mM or higher concentrations,no gel was formed at all with the material appearing as a hydrated mass. The monovalent ions, sodium and potassium, at concentrations up to 20 mM, had no effect on gel formation by C. abyssinica polysaccharide.

Table 2 Effect of calcium ions on the gelation of aqueous suspensions of C. abyssinica polysaccharide Concentration Subjective rating of calcium ions ( m w of gelation

Comments

Table 3 Effect of EDTA on the gelation of aqueous suspensions of C. abyssinica polysaccharide Concentration

Subjective rating of gelation

Comments

No pel

Hvdrated lumns with no pel

40 50 1nn

Gwns and Stabilisers for the Food Industry 10

74

3.4 Effect of pH In aqueous suspensions at low pH, a hard insoluble material remained in the suspending medium. Raising the pH resulted, first in the formation of a hydrated material that was not however, recognisable as a gel. At pH 6 and 7 a bonafide gel was formed as shown in table 4. Raising the pH above 7 resulted in small pieces of gel floating in a bulk of liquid. It is clear that control of pH is an important factor in the preparation of gels from C. abyssinica polysaccharide.

Table 4 Effect of pH on the gelation of aqueous suspensions of C. abyssinica polysaccharide

3.5 Effect of temperature The gels formed by the polysaccharide at room temperature were nearly as firm as those formed at 4 "C. At temperatures above 30 "C,the firmness of the gels formed decreased as shown in table 5. The observation that suspensions of the polysaccharide formed gels at room temperature would be an important factor when considering the formulation of food gels. On heating the gels at 100 OC for about 20 minutes, the gels broke up to form viscous masses. Upon cooling of the gels at 4 "C,individual lumps of gel formed but no continuous gel mass was formed. Similar results were obtained for gels that were prepared in the presence of sugar. It appears that gels formed by the polysaccharide from C. abyssinica do not reform easily after being destroyed by heat.

Temperature

eo 4

25

Subjective rating of gelation 1i-H-H

I+H-+

Comments

I Firm gel

I Gel less firmthan at 4OC

Polysaccharide Chamcteris&'on

15

3.6 Effect of physical state on gelation Optimum gel formation occurred when unground material was used and the mixture was not stirred. Grinding of the gum before suspension in water resulted in poorly formed gels. Altering particle size can influence the degree of dispersion because colloidal state depends primarily upon particle size (2). Grinding of the gum from Cordia abyssinica resulted in reduced particle size and increased dispersion of the gum in aqueous solution. The increased dispersion could result in reduced potential for interaction between the polymer chains and related reduced firmness of the gels formed. Acknowledgements

This study was generously supported by grants h m the Research Board of the University of Zimbabwe, the Swedish Agency for Research Corporation with Developing Countries (SAREC) and the International Foundation for Science (IFS). References Benhura, M. A. N. and Mavhudzi-Nyambayo, I. (1999). Depolymerisation of 1. mucilage isolated from ruredzo (Dicerocaryum zanguebarium) by ascorbic acid in the presence of catalysts, Carbohydrate polymers 38: 371-373. 2. Bowers, J. (1992). Water and food dispersions: In"Food theory and Applications", 2"dEd. Macrnillan Publishing Co. New York.pg 11-44. 3. Braek, G.S. Grasdalen, H.and Smidsrrad, 0. (1989). Inhomogenous polysaccharide ionic gels. Carbohydratepolymers 10 : 31-45. 4. Lipatov, Y.S. (1988). Gelation of solutions: Polymer gels, In "ColloidChemistry of Polymers" A.D.Jenkins, Ed., Elservier, New York. pgs 373-406. 5. Nussinovich, A. (1997). Hydrocolloid applications. Blackie Academic and Proffessional, London. 6. Oakenfull, D. (1987). Gelling agents. CRC Critical Rev.: Food Sci. and Nutri. 26(1): 1-25.

WHAT IS THE TRUE AMYLOSE CONTENT OF RICE STARCH ?

M.Ramesh, John R. Mitchell, Komelia Jumel and Stephen E. Harding Food Sciences Division, School of Biological Sciences, University of Nottingham, Sutton Bonington Campus, LE12 5RD, United Kingdom

Introduction

Starch, the major reserve polysaccharide, is of enormous nutritional importance as well as a raw material for various industries. It is used as a thickener, texturiser, stabiliser and fat-replacer in the food industry and in the manufacture of paints, adhesives, paper, textiles, pharmaceuticals and biodegradable plastics. The starch granule is composed of two glucose polymers: amylose, generally regarded as a linear chain of glucose units, and amylopectin, a highly branched and very large molecule. Among the different methods of isolation of amylose and amylopectin, gel permeation chromatography (GPC) has been extensively used to separate starch into a high molecular weight fraction, found to be branched and considered to be amylopectin and a low molecular weight fraction generally designated as amylose’“. The amylopectin was found to contribute to part of the iodine reaction with starch’ and amylose although largely linear has been shown to have some degree of branching ’-I5. Some workers have considered “branched amylose” as a third polysaccharide component’’ intermediate between amylose and amyl~pectin’~-’~. Amylose content of rice starch

GPC of rice starch on Sepharose 2BZ0and Sepharose CL-2BZ1was also reported to yield amylopectin (FrI) and amylose (FrII). Several workers showed that the long-B chains of the amylopectin molecules react with iodine to give a blue colour. Hence the am lose content, as estimated by iodine reaction, was considered to be an overestimate12.ly20.21 Subsequently, the amylose content, as estimated by the classical iodine reaction was designated as ‘apparent amylose’ and more appropriately as ’amylose-equivalent’ (AE)”. These developments lead to a general agreement that the true amylose content of rice starch may be lower than had been generally suggested. From the respective iodine affinities of whole starch, pure amylose and pure amylopectin, the true amylose content was calculated to be in the range of 15-19% for non-waxy rices”. The molecular size of enzymatically debranched starch was determined using SEC-MALLS and the high molecular weight material can be considered as the true “amylose” 22.

Polysaccharide Characterisation

I1

More recently Ramesh and his associatesn fractionated seven rice starches of graded amylose equivalent (AE) on Sepharose CL-2B column into a high molecular weight fraction (FRI) and a low-molecular weight fraction (FRII, M e r sub-divided into FR 11% IIb and IIc). They observed that the polysaccharide-iodine complex,,A of FRII was relatively lower (<630nm), except for the peak tubes (Fig.1). Which should not be the case if FRII were to be true amylose. Therefore, the FRI, IIa, IIb and IIc were debranched (using isoamylase) and fractionated on Biogel P-10. Surprisingly, all the four fractions in all the rice varieties gave a trimodal chain profile (Biogel frl, fk2 and fr3), demonstrating that the low molecular weight fraction (FRII) also contained some molecules which had a similar branching structure to amylopectin (Fig.2). When whole starch was debranched and fractionated on Sepharose CL-ZB, no material eluted at the void volume (FRI), confirming that FRI contained branched molecules alone. Assuming, the pattern of branching in the molecules of FRIIa, IIb and IIc similar to that of FRI, the following ratio (Long-B chains) / (Shorter chains) = (Biogel frl) / (Biogel fr2 + fr3) of FRI was applied to the branched molecules in FRIIa, IIb and IIc, from which the amount of long-B chains were calculated. By subtracting the long-B chain content from their total Biogel frl values the approximate amylose contents for non-waxy nces, in the range 7-1 1% (Table l), were obtained.

IS

2

--

10

::

5

0

0 LOO

'vo

800

1200

1600

LOO

800

1200

1600

2000

'v, Elution volume , ml

Figure 1. Elution profile of Type II(Co 32). IV(Basmati 370), VI(Sukanandi) & VI&T 65) rice starches on Sepharose C L - 2 bThe polysaccharide-iodine complex A- and division of diyerentfiactions are shown 23. (Reprinted with permission)

78

Gums and Srabilisers for the Food Indusrry 10

1

3 2 1

... I(

.

.._.

FRla.

..__

--.

1

-

.,_

:: I

20

I1

500 100

E C

I

I:

--

- 600

....

10

0

-

0

::

3c

2

0

,

”--.._,

FR Ilb

....

Y

- 600

-

1

-0 -400 E 4 -500

x

L

20

h 0 1

IC 0 FR UC

‘5

-. .._-....

- 600 - so0

10

- LOO

10 I I , I 0

Ir1

I

‘,

10

0

lotv

90 0

t

10

90

30

90

30

90

1LO

“I

Elulion volume. ml

Figure 2. Biogel P-10 chromatograms of the debranched Sepharose CL-2Bfi.actions FH, IIa, IIb and IIc of diflerent rice starches. ( 4 Carbohydrate; (---)AE; (...) ,I,-. The division of Biogel fi.1,fi.2 andfi.3 representing Long-B (Y amylose), intermediate-B and A and short-B chains respectively are shown 23 (Reprinted with permission).

We have recently confirmed the above by separating amylose from enzymatically debranched rice starches using SEC-MALLS-RI (Size exclusion chromatography coupled to multi angle laser light scattering and differential refra~tometer)~~, following the method of Ong and her associates22. Starch from four varieties of rice were debranched (using isoamylase) and fractionated on SEC-MALLS-RI. The following columns were used in series for fractionation : one guard column (Phenomenex; 6x40mm) followed by 5 fractionation columns (one TSK GEL G3000 PWXL; 7.8x300mm, two Asahipak GS320H; 7.8x250mm, one TSK (32500 PWXL; 7.8x300mm and one TSK G-oligo PWXL)22. In the chromatogram, the fraction I (FrI) was assumed to be true amylose (Fig.3). Similar

Polysaccharide Characterisation

79

Table 1. Approximate amount of long-B chains (LB) and amylose (Am) in rice 23 Sepharose Chains CL-2B in Fraction Biogel frl

FRI

LB

Approximate amount of LB & Am (% of total carbohydrate in rice)

I1

10.5

10.0

1.9 2.8 1.7 7.6 0.5 0.5 14.6 10.9

1.9 1.9 2.4 2.7 1.6 1.6 6.6 7.7 0.5 0.8 0.5 0.6 14.0 13.1 9.3 11.0

Am

FRIIa

LB Am

FRIIb

LB Am

FRIIc

LB Am

Total

LB Am

Rice quality type I11 IV VI VII VIII

I

8.6 6.7 6.6 5.8

-

-

1.5 1.3 3.6 1.8 1.0 0.8 5.7 4.6 0.4 0.4 0.3 0.3 9.6 9.1 9.6 6.7

2.4

1.3

0.3 2.3 1.0 0.1 4.2 0.2 0.3 0.1 0.7 0.5 8.4 3.3 7.2 0.7

0.30 0.25 0.20

Volume (mL) Figure 3. Representative elution profile of isoamylase debranched rice starch (Jhona20)fiactinated on SEC-MALLS-RI Vstem. (A) laser-light-scatteringprofie (B) Rlprojle. elution profiles has been reported for different starches, refemng the FrI as amylose and the subsequent fractions as the various chain populations of amylopectin22. It is possible that the amylose contained a mild degree of branching which was removed by the enzymatic treatment. The molecules must however, have mostly a long linear portion. In

Gums and Stabilisersfor the Food Industry 10

80

the chromatogram a baseline was established and the area under the amylose fraction was calculated from the sum of the area of several Gaussian components fitted to the data using a multiple Gaussian fit (Microcal Origin 3.0, USA). Interestingly, the amylose content, ranged from 7-1 1%, agreed well with the calculated amylose contents23 (Table 2). The amylose values observed in the recent studies23J4are not only lower than obtained by the classical iodine assay applied to whole starch but also lower than the values reported from an iodine assay for chemically separated "amylose"'2. The molecular weight and polydispersity (Mw/Mn)values of the amyloses (Table 2) are in agreement with the values reported for purified rice starch amyloses 25.

Rice

Amylose

Conclusion The above recent reports have indicated that the rice starch contains branched molecules or amylopectin of varied sizes, from vehy big to very small, and the content of amylose is much less than generally stated in literature. A primary aim of future work should be to investigate the true amylose content of starches from other botanical origin by using these approaches to understand whether the low true amylose content is a feature of other starches or is unique of rice starch.

Acknowledgement The award of Nestle Nutrition Scholarship to M.Ramesh by Nestec Ltd., Switzerland, is gratefully acknowledged.

References [I] [2] [3] [4]

R.Ebermann, & R.Schwarz, Stuerke, 27 (1975) 361-363. T.Yamada, & M.Taki, Staerke, 28 (1976) 374-377. C.G.Biliaderis, D.R.Grant, & J.R.Vose, Cereal Chem., 56 (1979) 475-480. C.D.Boyer, P.A.Damewood, & G.L.Matters, Staerke, 32 (1980) 217-222.

Polysaccharide Characterisation

81

J.Y.Yeh, D.L.Garwood, & J.C.Shannon, Sfuerke,33 (1981) 222-230. S.A.S.Craig, & J.R.Stark, Stuerke 36 (1984) 127-131. O.Kjolberg, & D.J.Manners, Biochem. J., 86 (1963) 258-262. W.Banks, & C.T.Greenwood, Arch. Biochem. Biophys. 117 (1966) 674-675. D.J.Manners, & NKMatheson, Curbohydr. Res., 90 (1981) 99-1 10. J.A.Cura, P.E.Jansson, & C.R.Krisman, Sfuerke,47 (1995) 207-209. H.Falk, R.Micura, M.Stanek, & R.Wutka, Sfuerke,48 (1996) 344-346. Y.Takeda, S.Hizukuri, & B.O.Juliano, Curbohydr. Res., 168 (1987) 79-88. Y.Takeda, N.Maruta, S.Hizukuri, & B.O.Juliano, Curbohydr.Res., 187 (1989) 287294. Y.Takeda, T.Shitaozono, & S.Hizukuri, Curbohydr. Res., 199 (1990) 207-214. S.Hizukuri, Y.Takeda, N.Maruta, & B.O.Juliano, Curbohydr.Res., 189 (1989) 227235. Y.Sugimoto, K.Yamada, S.Sakamoto, & H.Fuwa, Sfuerke,33 (1981)112-116. M.Asaoka, K.Okuno, Y.Sugimoto, J.Kawakami & H.Fuwa, Sfuerke,36 (1984)189193. M.Asaoka, K.Okuno, Y.Sugimoto, M.Yano, T.Omura, &t H.Fuwa, Sfuerke, 38(1986) 114-117. N.Inouchi, D.V.Glover, & H.Fuwa, Sfuerke,39 (1987) 259-266. R.Chinnaswamy, & K.R.Bhattacharya, Sfuerke,38 (1986) 51-57. K.R.Reddy, S.Z.Ali, & K.R.Bhattacharya, Curbohydr. Polym., 22 (1993) 267-275. M.H.Ong, K. Jumel, P.F.Tokarczuk, J.M.V.Blanshard, & S.E.Harding, Curbohydr. Res.,260 (1994) 99-117. M.Ramesh, S.Z.AIi, & K.R.Bhattacharya, Curbohydr.Polym., 38 (1999) 337-347. M.Ramesh, J.R.Mitchel1, K.Jume1, & S.E.Harding, Staerke (1999) (in press). Y.Takeda, T.Shitaozono, & S.Hizukuri, Sfuerke,40( 1988) 5 1-54.

The dispersibility of polysaccharides in water and watercadoxen mixtures Wang, Q., 1 Wood, P.J., 'Chi,W. and 2Ross-Murphy, S.B. Southern Crop Protection and Food Research Center- Food Research Program, AAFC,Guelph, Ontario, Canada. 2Division of Life Sciences, King's College London, Franklin Wilkins Building, London SEl 8WA, UK.

ABSTRACT

The effect of added cadoxen on the dispersibility of aqueous solutions of polysaccharides was studied using viscosity measurements. Three neutral polysaccharides, oat p-glucan, detarium xyloglucan and dextran were included in this study. For all three samples, the Huggins constants were found to decrease rapidly when the volume fraction of cadoxen (vcad) increased from 0 to 0.4, then consistently decreased at a much lower rate. This indicates a transition from large aggregates to unimers (or small aggregates) at solvent composition of about O
-

INTRODUCTION

When preparing aqueous solutions of polysaccharides, heating is usually needed to assist the dissolution process. However, excessive heating may induce polymer chain degradation, which is normally undesirable. The molecular weight of polysaccharides can usually be monitored by the measurement of intrinsic viscosity. However, the use -of this method is complicated by the possible aggregation behavior of polysaccharides. Energy provided by heating can disrupt the molecular aggregates, as well as cleaving the polymer

Polysaccharide Characterisation

83

chains. If a reduction in intrinsic viscosity is observed after heating, it is difficult to tell if the change is due to the disruption of the aggregates or the degradation of polysaccharide molecules. There is thus a need to develop a stable and consistent method to distinguish polymer chain degradation from disruption of supramolecularaggregates. Cadoxen is commonly applied as a solvent for cellulose and cellulose derivatives. A recent study (Grimm, 1995), using static light scattering, showed that cadoxen is also a better solvent for barley p-glucan compared to water. The degree of aggregation of barley p-glucan was about 10 times higher in water than in cadoxen. Cadoxen has also been used as solvent for a p -(1+3)-glucan (Zhang, 1997), and aggregation was significantly reduced by the addition of cadoxen to water. In this paper, we compare the dispersibility of three neutral polysaccharides in water-cadoxen mixtures and in water. The objective was to use the difference in the dispersibility between polysaccharide in water and in water-cadoxen mixture as an indicator of polymer degradation or disruption of aggregates. The polysaccharides involved were oat p-glucan, detarium gum and dextran. Oat p-glucan is a linear polysaccharide composed of mainly (1-+3)-linked cellotriosyl and (1+3)-1inked cellotetraosyl units in a molar ratio of 1:(2.1-2.4) (Wood et al., 1991; Cui, et al., 1999). Detarium gum is a xyloglucan type of polysaccharide that is believed to have a low degree of, but long, branching structure (Wang, et al., 1997). The structure of dextran varies from slightly branched to highly branched, depending on the origin. It consists mainly a-D-(1~6)-glucopyranoseas main chain with branching points at C-3 and (2-4.

MATERIALS AND METHODS Materials The p-glucan was extracted and purified by the method described previously (Wood, 1991). The detarium gum was purified by the method of Wang et al. (1996). Another polysaccharide studied in this project was a commercial sample dextran T5OO (Pharmacia Fine Chemicals AB, Sweden). Preparation of sample solutions Polysaccharide solutions were prepared by first dissolving the accurately weighed sample at 80 OC for 1 hour and then leaving overnight at room temperature under constant stirring. The concentration was selected to ensure that the relative viscosity qr fell in the range of 1.2 < q, < 2.2. The solution prepared was filtered through a 0.45 pm syringe filter before any other treatments. The cadoxen solvent was prepared as follows. A 29% aqueous solution of ethylenediaminewas saturated with CdO in an ice-water bath under vigorous stirring and kept overnight. The solution was then filtered through a sand filter and refrigerated until use. The polysaccharide solutions in different volume fractions of cadoxen was made by mixing the filtered aqueous solution with the required volume of cadoxen and stirring for at least 1 hour before measurement was made.

Gums and Stabilisersfor the Food Industry 10

84

Intrinsic viscosity measurement The intrinsic viscosity was determined by using a dilution capillary viscometer immersed in a water bath to maintain the temperature at 25 f 0.1"C. The measurement was done in a fume hood because cadoxen is a toxic solvent. RESULTS AND DISCUSSION Intrinsic viscosity can be determined according to the Huggins equation (Tanford, 1961): qs&

= [q]

+ k[q]*c + ....

using a capillary viscometer, the reduced viscosity q n d (= q&) is measured at various concentrations (c) in dilute solution and extrapolated to zero concentration to give the value of intrinsic viscosity. k is known as the Huggins constant, and is an indication of the solvent quality for a given polymer. It has been estimated to be 0.5 < k < 0.7 in 0media and to decrease with increasing polymer solvation, to a limiting value of 0.3 in good solvents (Sakai, 1968). The Huggins constant is very sensitive to the formation of molecular aggregates. If cadoxen increased polysaccharide solvation over that of pure water by reducing polymer-polymer interactions (aggregations), then intrinsic viscosity would be less. Opposing this, greater polymer expansion in a good solvent compared to a poor solvent will lead to an increase in intrinsic viscosity. Depending on the relative magnitude of these two effects, three possible results may be produced. When the first factor is predominant, the intrinsic viscosity would fall; if the second factor were predominant, it would increase. The intrinsic viscosity would remain largely unchanged if both factors were in balance.

-

Figure 1 displays the Huggins plots of detarium gum in water and water-cadoxen mixtures. The slopes of the plots were reduced consistently when the volume fraction of cadoxen (Vcd) increased. From the slopes of these plots, the Huggins constants were calculated and plotted against the volume fraction of cadoxen, as shown in Figure 2. The Huggins constants decreased from 0.61 in water to 0.40 in 0.70 (Vcd) water-cadoxen mixture. This indicates that the quality of the solvent was improved by the addition of cadoxen, i.e. the association between polysaccharide molecules was reduced. The intrinsic viscosity of detarium gum also decreased significantly in water-cadoxen mixtures. Figure 3 demonstrates that the intrinsic viscosity of detarium xyloglucan also changes with the volume fraction of cadoxen in water-cadoxen mixtures. As the volume fraction of cadoxen increased, the intrinsic viscosity decreased rapidly at the beginning from 9.3 in water to 5.5 in 0.3 (Vcad) cadoxen, then it continuously decreased slowly when v c a d > 0.3. This suggests that most of the aggregates in detarium aqueous solution break up to form unimers (or small aggregates) in a narrow range of solvent composition (o
-

-

Polysaccharide Characterisation

85

s F

0

0.02

0.04

0.06

0.08

0.1

Concentration (%) Figure 1 Huggins plots for detarium gum solutions in water and water-cadoxen mixtures with different volume fraction of cadoxen Vcd. -- H 2 0 , A -- 0.1, 0 -- 0.2, o -- 0.3, +K -0.4, A -- 0.5, and G -0.7.

+

0.2 0.1

J 0

0.2

0.4

0.6

0.0

Volume &tion of cadoxen Figure 2 Huggins constants in relation to the volume fraction of cadoxen in watercadoxen mixtures for detarium xyloglucan (+), oat p-glucan (.) and dextran ( A ).

86

Gums and Stabilisersfor the Food Industry 10

l2 n 11

i

en

3 10

W

.29 m

0

.& 2 8 > .& 0 7 m

.9 3

6

uE: 5

0

0.4

0.2

0.8

0.6

Volunr: fixtion of cadoxen Figure 3 Intrinsic viscosity in relation to the volume fraction of cadoxen in watercadoxen mixtures for detarium xyloglucan (*) and oat p-glucan (.). ~

16

5 14 F 12 10

8

4

-r

0

0.02

T

0.04

I

0.06

0.08

0.1

Concentration(%) Figure 4 Huggins plots for oat p-glucan solutions in water and water-cadoxen mixtures with different volume fraction of cadoxen Vcad. 0.4. o -- 0.6.

-- H20, A- 0.05, A -- 0.1, -- 0.2, t --

87

Polysaccharide Characterisation

In the case of oat p-glucan, the effect of added cadoxen on the solvation of the polymer was even more pronounced. The Huggins constant decreased from 0.81 in water to 0.25 in 0.60 cadoxen (Figures 2 and 4). However, the intrinsic viscosity of 0-glucan only decreased slightly on addition of cadoxen (Figures 3 and 4). It can be explained that although the disruption of the p-glucan aggregates in cadoxen leads to a decrease in intrinsic viscosity, the solvation in cadoxen caused an extensive expansion of the polymer chains which leads to an increase in [q]. As a result, the intrinsic viscosity [q] only slightly decreased. Interestingly, dextran behaved in a different way on the addition of cadoxen to the water solution. As demonstrated in Figure 5, the intrinsic viscosity of dextran instead of decreasing, increased significantly, from 0.51 in water to 0.90 in 0.5 (Vcd) cadoxen. However, the Huggins constants decreased to a lesser extent compared to the other two polysaccharides. One possible explanation is that water is a “better than 0” solvent for dextran. Dextran molecules do not aggregate extensively like other polysaccharides do, both perhaps because of the branching structure of the high molecular weight sample, and of the highly flexible nature of the saccharide chain. Consequently, the increase in intrinsic viscosity is most likely caused by the further expansion of the macromolecular chains in cadoxen.

1 n

9 0.8 .-0 B 3 0.6 + .-0m .r(

3 0.4 d

U

0.8 Volunr: fixtion of cadoxen

Figure 5 Intrinsic viscosity of dextran vs. volume fraction of cadoxen.

88

Gums and Stabilisersfor the Food Industry I0

CONCLUSIONS In conclusion, water-cadoxen mixture is a better solvent for neutral polysaccharides compared to water alone. The aggregation of polysaccharides can be significantly reduced when the volume fraction of cadoxen V,& -0.4. The difference in dispersibility of polysaccharides in water and water-cadoxen mixture is very useful for distinguishing the polymer degradation from the dissociation of aggregates, although it is only the first stage of a fuller characterisation employing, for example, static light scattering. The application of this approach in this regard will be published later.

ACKNOWLEGMENT The authors thank Mr. John Weisz and Ms Cathy Wang for technical assistance.

REFERENCE Cui, W., Wood, .P.J., Blackwell, B. and Nikiforuk, J. (1999). Physicochemical properties and structural characterization by two-dimensional NMR spectroscopy of wheat P-D-glucan - comparison with other cereal P-D-glucans, Carbohydrate Polymers, In press. G r i m , A., Kruger, E., Burchard, W.(1995). Solution properties of P-D-( 1,3)(1,4)-glucan isolated from beer. Carbohydrate Polymers, 27(3), 205-2 14. Sakai, T. (1968). Huggins constant k’ for flexible chain polymers. Journal of Polymer Science. A-2,6, 1535-1549. Tanford, C. (1961). Physical Chemistry of Macromolecules (pp.392-393). New York: John Wiley and Sons. Wang, Q., Ellis, P.R., Ross-Murphy, S.B. and Reid, J.S.G. (1996). A new polysaccharide from a traditional Nigerian plant food: Detarium senegalense Gmelin. Carbohydrate Research, 284 (2), 229-239. Wang, Q2, Ellis, P.R., Ross-Murphy S.B. and Burchard, W. (1997). Solution characteristics of the xyloglucan extracted from Detarium senegalense Gmelin. Carbohydrate Polymers 33, 115-124. Wood, P.J., Weisz, J. and Blackwell, B.A. (1991). Molecular characterization of cereal PD-glucans - Structural-analysis of oat P-D-Glucan and rapid structural evaluation of P-D-glucans from different sources by high-performance liquidchromatography of oligosaccharides released by lichenase. Cereal Chemistry, 68 (l), 31-39. Zhang, L., Ding, Q., Zhang, P.Y., Zhu, R.P., Zhou, Y.H. (1997). Molecular weight and aggregation behaviour in solution of P-D-glucan from Poria COCOS sclerotium. Carbohydrate Research, 303 (2), 193-197.

Polysaccharide Gelation

BIOPOLYMER GELATION - THE STRUCTURE-PROPERTY RELATIONSHIP

A.H.Clark Unilever Research Colworth Colworth House Sharnbrook Bedford MK44 1LQ UK

1 INTRODUCTION

Biopolymers are described as having functions in foods apart, that is, from their obvious role in nutrition. For example, they can enhance the stability of a food product against physical processes such as emulsion creaming, the breakdown of foams, the ripening of crystal networks, and the sedimentation of particles. They also contribute to mechanical properties (texture), to appearance, and to the rate at which flavours are released. Biopolymers are thus important food ingredients, functioning successfully often in rather small amounts, and sometimes specifically. The functions of biopolymer ingredients in foods are clearly related to their molecular structures, first to primary descriptors of these structures, such as residue sequence and molecular weight, but also to higher levels related to ordered chain conformations and to superstructures based on aggregation of these. For polysaccharides, primary sequences can be relatively simple and analogous to those of synthetic polymers, but for proteins, sequences are usually complex and unique, making globular proteins quite unlike any synthetic counterpart. Their only simplifying feature is perhaps their monodisperse molecular weight character, polysaccharides being usually quite polydisperse. Polysaccharide primary structure can also be complex. Polysaccharides sometimes show irregular statistical distributions of backbone residues or sidechains (e.g. alginate and locust bean gum) which are neither totally random nor totally organised: they sometimes have regular sequences broken periodically by ‘interrupting’ residues (e.g. pectin), which can play a part in their solution behaviour, and they can be highly branched (e.g. amylopectin). If they are closer to synthetic polymers than most proteins, polysaccharides are nonetheless rather complex analogues of their man-made counterparts. Structure and function are thus well-defined aspects of the study of food biopolymers, and both have been extensively examined. There remains a difficulty, however, which is basically a theoretical one, of achieving quantitative links between molecular structure and function. Such a link is clearly desirable as it should lead to more efficient molecular design, selection of ingredients, and choice of processing procedures. The fact that current links of this kind are highly imperfect and rather qualitative, reflects the complex structures of the biopolymers involved, and the nature of the physical properties that any theory or model has to address. It is the purpose of the present review to examine the progress made in this endeavour, particularly in relation to biopolymer gel formation, and to consider opportunities for future progress.

92

Gums and Stabilisers for the Food Industry 10

2 HYDROCOLLOIDS AS STATISTICAL COILS The normal environment in which hydrocolloids function, is an aqueous one, where they are often present at fairly low concentration. Low water environments are of course also possible but are not examined in the present review. Initially then, a polysaccharide can be considered much as a normal synthetic polymer, but perhaps a bit stiffer than usual, and with water as a solvent. For synthetic polymers, the main features of solution behaviour, such as high viscosity at low concentration, viscoelastic character, solution non-ideality, and a tendency to phase separate in certain conditions, have long been understood in terms of the configurational behaviour of large flexible chains, and models based on their random flight description.1 If the polymer chain is imagined.to consist of N rods, or segments, of length L connected by universal joints, a statistical model can be developed which provides an expected distribution for the end-to-end distance of the chain, and the following root mean square average value for this distance, i.e.

The picture of the polymer molecule in solution is thus one of a randomly coiled structure, occupying an effective volume much larger than its true physical volume. Of course real polymer chains differ from this ideal in that their residues show restricted rotation and can’t pass through one another (excluded volume). The influence of solvent is also important. The intrinsic stiffness of the chains can be taken into account by making L the length of a so-called statistical segment which may involve several of the basic polymer repeat units. N then becomes the number of such segments, and the stiffer the polymer, the larger is the number of residues defining L, and the smaller N. Longer range repulsions between segments, and the influence of solvent on these, are treated by introducing an expansion factor1 which allows the real end-to-end distance to differ from its idealised value. In practice, changing solvent or temperature, causes the expansion factor to vary, but when it does equal unity (as can happen) solvent-induced attraction between segments exactly counterbalances long range repulsion (excluded volume is zero), and so-called theta conditions prevail. A model for the coil expansion factor has been constructed1 by introducing a polymer-solvent interaction parameter x which quantifies an imagined equilibrium involving the weak association of polymer segments in the solvent involved. Arguments have shown that when x = 0.5 the theta condition is achieved (poor solvent conditions), with coil expansion occurring as x falls below 0.5 (good solvent). The significance of the above ideas for hydrocolloid solutions is considerable. High viscosity generation at low concentrations derives from the high effective coil volume, as coordinated rotation of the expanded coil through the solvent dissipates much more energy, than uncoordinated movement of the individual segments.1 In terms of the coil model, this behaviour is clearly related to the value of the parameter x for any biopolymer-solvent combination, and temperature, and on other factors determining the effective coil volume, such as the statistical segment length (intrinsic coil stiffness). To some extent the latter is amenable to molecular modelling calculations2 based on inter-residue interactions, butpredicting x in a similar way is a formidable challenge. Finally, it should be commented that the above is a very simple picture of hydrocolloid behaviour and often needs elaboration, particularly in relation to concentration effects and deviations of the polymer from a simple coil description. The presence of charge (i.e. polyelectrolyte character) has an important influence on chain expansion, and also where chains become really stiff (very large statistical segment length, few segments present per molecule) alternative models based on idealisations such as a ‘wormlike chain’, or extended rod, are required. This is particularly true where polymers have undergone ordering transitions of the kind now considered.

Polysaccharide Gelation

93

3 ORDERING TRANSITIONS

Coiled polysaccharides are intrinsically stiff because of restricted rotation about inter-residue bonds, but when molecular weight is large, they are often describable as random coils in the manner just outlined. In some cases, however, as temperature falls, or other conditions change, such as pH or solvent composition (addition of salts, small molecules, etc), the entropic advantage to be gained from the disordered state becomes outweighed by energy gains offered by an ordered helical state. Single helix formation would be the most obvious transformation but, for polysaccharides, winding of chains into multiple helices is also common (particularly double helices) so that the ordering transition can also be an aggregation process. The importance of ordering transitions (and the reverse process of conformational ‘melting’) in molecular biology prompted the early development of models for their description of which perhaps the most celebrated are those of Zimm and Bragg3 (uncharged coils and helices) and Zimm and Rice4 (charged molecules). In this type of approach chain residues are seen as ordered (h) or disordered (c) and all possible combinations of such states for a chain of defined length are considered, each configuration being assigned a statistical weight. This last is primarily defined in terms of an equilibrium constant ‘s’ related to the free energy difference between isolated helix and coil states. To reproduce the transition sharpness often observed, nearest neighbour correlation is introduced through a co-operativity parameter ‘u‘ which modifies the statistical weight of a helix state at the end of a helical sequence to ‘su‘,i.e. introduces an endeffect penalty. The smaller u is, the sharper is the transition as s varies (temperature change or other change in conditions). The Zimm-Bragg model is successful in describing experimental data for degree of ordering versus temperature for biopolymers and, in its more elaborate ZimmRice form, ion binding is also considered as a driving influence. Disorder-order conformational transitions in polysaccharides are extremely important for their functions in foods, partly because stiffening will promote increased effectiveness in viscosity generation, but mainly because gel network formation often requires transition to the ordered form prior to network assembly. The prediction of such behaviour from molecular structure, which is evidently desirable, is hstrated, however, by the fact that the parameters s and a are dificult to quantify in molecular terms. Some attempts to look at this issue using molecular modelling have been attempted but, at present, this problem seems to be beyond the scope of the approach, though some insights into the origins of helix stability can be obtained.5 4 VISCOSITY GENERATION

This very important property of hydrocolloid solutions is often demonstrated by plotting specific (or relative) viscosity at low shear rate (Newtonian plateau value) versus polymer concentration. When this is done on a log-log basis there is usually an apparent intersection of two straight lines at a concentration C*, the entanglement (or coil overlap) concentration.6,7 Different power law behaviour occurs below and above this threshold, the viscosity increasing more rapidly with concentration as polymer coils interpenetrate and entangle. It is found that C* is related to polymer intrinsic viscosity [q] according to C*[q] = constant, where the latter is close to 4. Since the intrinsic viscosity is actually an effective hydrodynamic volume for the polymer coil (expressed per gm of polymer) it is clear that C* is very much determined by coil expansiodcontraction, a fact confirmed by the well known Mark-Houwink relationship, [q]=KM’ where M is the molecular weight.1 In this, the exponent a is equal to 0.5 for the unperturbed random coil, rises to 0.8 for expanded coils, and can be greater than this for non-coiling extended structures.

94

Gums and Stabilisers for the Food Industry 10

4

3

2

f

-

[ rlsp

- 1011 0

, -0.5 0 0.5 1.0 1.5 2.0

c*

-41

Figure 1 Viscosity-concentrationmasterplot approach to coil thickenersfrom ref: 6 The decisive nature of the intrinsic viscosity in determining hydrocolloid viscosity generation is indicated in Figure I (data from ref. 6) which shows how experimental data for a range of coil polysaccharides can be reduced to mastercurve form by plotting specific viscosity against reduced concentration C[q]. This procedure illustrates a recurring theme of the present review in which the essential physics of a phenomenon is expressed in the mathematical form of amastercurve, while the absolute behaviour of an individual system requires extra information in the form of values for the reducing parameters (in this case the solvent viscosity and [q]). While theoretical work by Rouse78 Zimm? Graessley,lO,ll, De Gennes12713 and Doi and Edwards14 has explored the basic physics of polymer thickening, including the linear and non-Iinearviscoelastic properties of polymer solutions (e.g. shear thinning of entangled systems), and has given some explanation for the mastercurve form in Figure I , the link to detailed molecular structure remains obscure. Accurate prediction of the intrinsic viscosity from molecular structure would evidently be a main step in this, but is frustrated by difficulties in treating influences such as solvent effects, polyelectrolyte character, and the helix-coil transition (see last Section). Interestingly, the high viscosity which topologically entangled systems can generate (C>C*) often confers ‘gel-like’ properties, and indeed polymer networks of a type are being formed. To explore and define gel behaviour, however, the solid (elastic) character of such systems must be examined, and this is addressed next, through consideration of the viscoelastic character of hydrocolloid solutions, and the influence on it of more and more permanent cross-linking between chains.

Polysaccharide Gelation

95

5 TEMPORARY NETWORKS

The dynamic mechanical response (frequency spectrum) of a typical entangled polymer solution (which might be mistaken for a gel) is shown as (A) in Figure 2 over the frequency range typical of a conventional measuring instrument. Only at higher frequencies is the response ‘solid-like’ with G’ > G” and even then there is a substantial viscous component. The cross-over frequency between liquid and solid response is on short timescales, and will vary with concentration (degree of entanglement). The results reflect the very short lifetimes of topological entanglements and the fact that finite (if high) Newtonian viscosities can be measured. As gels, they are not convincing, being better described as elastic fluids. Hydrocolloid solutions can, however, show more definite ‘solid’ characteristics, hints of this appearing in viscosity plots of the type shown in Figure 1, where increased power law behaviour above C* has been noted, and described 6 7 in terms of so called ‘hyperentanglements’ (extra and longer lived cross-links). Where polysaccharides undergo conformational ordering (e.g. the polysaccharide xanthan) this effect can be pronounced, with difficulty being experienced in establishing a Newtonian viscosity plateau experimentally, and the viscosity showing strong shear-thinning character. 7915 This implies that, in quiescent conditions, a three dimensional gel network can form through chain association, with a much higher level of permanence than the topological network. In terms of the linear viscoelastic spectrum, such systems tend to behave more like Figure 2B, which now indicates almost uniform solid response over the frequency window, consistent with network bond relaxation moving to much longer timescales. The propensity of such systems to network breakdown under increasing applied stress (particularly above a critical threshold or ‘yield stress’ ) shown by their capacity to flow and shear thin, confirms, however, that the bonds present are of limited strength, and are short-ranging, and of still finite lifetime. They are sometimes called weak gels, or temporary networks, the term ‘gel’ reflecting their essentially solid character over ‘normal’ mechanical spectrometer timescales, and ‘weak’ relating to their often quite low elastic moduli, residual viscous character (significant G”) and fluid character (non-linear response).

Figure 2 Typical mechanical spectra for (A) entangled (B) temporary networks Thickeners capable of showing transient network character, and apparent yield stress behaviour, find many industrial applications, and they have attracted much attention from theorists. Models vary between those attempting analytical descriptionl6.17 (basically similar to those applied to entangled networks), to those based on computer simulation.18~19 The new ingredient in such work is of course the introduction of sites or regions of the polymer where specific attraction can occur. Off-lattice simulations~9have succeeded in reproducing much of

96

.

Gums and Stabilisersfor the Food Industry 10

the relaxational characteristics of such networks, as reflected, for example, in the frequency spectrum over a wide frequency range. Analytical approaches have had similar successes. A problem with both types of model, however, is that the component polymer chains and their interactions, are highly simplified, and it is difficult to see how results can be re-interpreted at a molecular level.

6 EXPERIMENTAL ASPECTS OF STRONG GEL FOKMATION Strong gels differ from the transient networks (weak gels) just described through their very different non-linear responses to applied stress, i.e. their tendency to stretch elastically, and show failure. Such properties are evidently a consequence of their network cross-links being much stronger, and longer lived than in temporary networks, though the distinction is not sharp. The frequency spectra of strong gels are usually significantly less frequency dependent, and have lower G” values, over comparable frequency intervals. On normal timescales little relaxation occurs and when it does, through bond breaking at large strains, the process is essentially stochastic, depends on network imperfections, and leads to failure. That even the bonds in strong gels have finite lifetimes can be shown, however, in long time creep and stress relaxation experiments.20

Time Figure 3 Typical time evolution of the dynamic modulus for a strongh gelling system showing one estimate of the gel time and the truncation time defining an apparent plateau modulus value

Strong gel formation is of course a classical property of polysaccharide solutions (and gelatin) when these are cooled, andlor specific counterions are added (amylose, agarose, carrageenan, alginate, pectin, gellan, etc), and it is also a property of many globular protein solutions when exposed to heat. For the cold-set gels, strong crosslink formation seems to have its origins in stable multiple helix formation (gelatin triple helix, carrageenan double helix) andlor in a pronounced tendency for such helices as are formed (single or multiple) to aggregate in solution. Amylose and agarose gels provide well-known examples of this latter phenomenon whilst, for the more charged carrageenans, alginates, pectin and gellan, particular counterions may be necessary

Polysaccharide Gelation

97

to achieve the same effect. Carrageenan systems in other situations may have to rely on helix formation alone to build networks, i.e where aggregation is inhibited by charge. In this respect they become more like gelatin gels. In the globular protein case (at pH’s removed from the isoelectric point) initial linear fibril formation generates stable structures, which then interact strongly enough to produce a gel network. The much stronger and permanent nature of cross-links in strong gels has important implications for experimental approaches to their study, particularly in relation to dynamic mechanical spectroscopy. Processes of bond formation in such cases require finite (and sometimes quite long) times to occur (rarely an issue for temporary networks), and delay evolution of the modulus (and other gel properties) in a quenched solution, allowing the dynamic mechanical spectrometer to measure a cure curve (G’ and G” recorded as functions of time - see Figure 3). Cure curve measurement20,21 has in fact become a generic method for characterising the kinetics of gelation, and much modelling and theory has been directed towards explaining its principal features, such as the lag period or gel time (tg), and the apparent limiting modulus value achieved at long but finite truncation times (G’ at ttr). Interest attaches to these quantities in relation to quench conditions such as the final temperature, the concentration ofbiopolymer, and its molecular weight.22 Further important properties of the gels are slow long time processes of modulus growth20 and the temperature dependence of the final modulus (gel melting).22 The frequency dependence is not of great interest over normal frequency ranges for the reason already outlined, but creep or stress relaxation experiments, sometimes attempted, allow access to longer times, and provide further information about cross-link dynamics. The results of cure experiments for different gelling systems are oftensurprisingly similar and may be summarised as follows:

Figure 4 Reduced experimental gel time (Td,and plateau modulus (G 2 - concentration data, compared to cascade theory master plot predictions flor data shown see refs. 22-28) (1) The gel time (Figure 4) appears to diverge at a certain critical concentration but shows power law behaviour above this.22 The power law varies from system to system, being quite large for

98

G u m and Stabilisers for the Food Industry 10

polysaccharides23~24,25 showing helix aggregation, and for globular proteins26327 forming fibrils, but much lower for gelatin23928 and high methoxy pectin.23 (2) The temperature dependence of the gel time is much less well characterised, but for globular proteins an Arrhenius behaviour has been observed.29 For polysaccharides, gel times23 generally become shorter on cooling, consistent with an ordering process. (3) The apparent limiting gel modulus (obtained after some chosen long truncation time) shows a C2 concentration dependence at higher concentrations, but this dependence continuously increases its power law index as concentration falls, until eventually the modulus diverges logarithmically (goes to zero) at a critical concentration (Figure 4).22

E pa

2 Ogi'

.*

iooa cascade model

500

0

1

0

4

8

Figure 5 Cascade modelling of experimental Young's modulus (E) versus weight average molecular weight (M,) for K-carrageenanfractions (datafrom ref: 30)

(4) The apparent limiting modulus for a system at constant concentration, but varying polymer molecular weight, stays at zero, until a critical molecular weight is reached, then rises rapidly to a plateau value, and then becomes independent of molecular weight (Figure 9.30

( 5 ) The apparent limiting modulus usually falls on raising temperature (Figure 6)20 (or increases on cooling) and in some cases it becomes zero at a critical upper temperature (the gel melting temperature).3 1 Sometimes this melting curve shows temporary modulus increases indicating some form of annealing,32 and for polysaccharides, the melting temperature is sometimes greater than the highest effective setting temperature (hysteresis affect).20,32

(6) The apparent limiting modulus can increase in logarithmic time indicating long-term network instability (particularly for gelatin gels).20 This seems to have little effect on the shape of the modulus-concentration relationship but it does lower the effective critical concentration (very slowly). The effect on gel melting has not been systematically examined. Some modelling exercises aimed at understanding these findings are now discussed.

Polysaccharide Gelation

99

G’P~

21

cascade f=3

_--.-__--.*__.-. _.^__“ _.-.... --*-.>i--

-*a4

.f=100

81

\

PECIW GEL MELTING

t fi

1

4.1

cascade

f=3,100

t “C 10

20

30

0

Figure 6 Cascade modelling of experimental modulus-temperature ‘meltdown’data for gelatin andpectin gels (see reJ31)

7 MEAN FIELD AND OTHER MODEL DESCRIPTIONS A number of model approaches have been developed to describe the formation of strong gels and their properties. These include the mean field model discussed below?3-35 some related approaches,17,23,24,36,37 lattice models such as percolation theory?8-40 and various off-lattice simulations.41-43 The mean field description is largely a development of Flory’s theory of rubber elasticity1 modified to include the physical nature of cross-links in biopolymer gel networks, and the large amounts of solvent within which cross-linking takes place. Further complications include the common need to introduce kinetic processes prior to cross-linking, such as the generation of ordered (or for proteins, disordered) primary gelling units, or even nonrandom clusters of these (protein fibrils, and preliminary semi-crystalline polysaccharide aggregates). A tendency for liquid-liquid demixing (see next Section) can also introduce extra kinetics, but here, where only homogeneous gel networks are considered, this is assumed absent. The presence of solvent in gelling biopolymer systems makes concentration an important physical quantity. In describing their gelation kinetics it is assumed that a starting stable solution of the biopolymer of concentration C (% w/w) is quenched rapidly, either to a higher or to a lower temperature, to set aggregation and gelling in motion. As mentioned above, any complete kinetic model, as well as describing network building, will normally have to address preliminary conformational changes, and/or non-random aggregation events. For the sake of simplification, however, such initial processes will be assumed absent, and the mean field model presented for the case where the initial molecules can begin association as they stand, and where these are essentially molecularly monodisperse.

Gums and Stabilisers for the Food Industry 10

100

To set up the mode1,33-35 each starting molecule is assigned a number f of potential binding (cross-linking) sites and, as time proceeds, a fraction a (t) of these reacts. Flory’s theory of random cross-linking] suggests that small aggregates alone will form untila = l/(f-I) when the gel point is reached, but then a very large gel molecule will suddenly appear (critical divergence of the weight average molecular weight). After this, the remaining sol fraction becomes increasingly cured into the gel and the modulus rises. The problem is to calculate this modulus as a function of time for any set of gelling conditions, a process which requires the increasing density of elastic or load-bearing chains to be estimated. Flory has provided a method for doing this for a rather restricted range of network types based on covalent cross-linking, but for biopolymer gels Clark and Ross-Murphy20933 preferred a more general approach based on the Case-Scanlan definition of the EANC (elastically-active network chain) and the cascade approach to gelation statistics.44 The approach has been described in detail elsewhere but, in essence, a final expression for the modulus at any time t is given by the expression, G = {Nfa( I-v)2(1-P)/2}aRT

(2)

in which

and

p = (f-l)av/(l-a+av)

(4).

In these expressions v is the so-called extinction probability, which is the probability that a reacted functionality of an arbitrary molecule becomes ‘extinct’, that is fails to connect to the boundaries of the gel. The expression in parentheses in (2) is then the required density of EANCS with N = C/M, the concentration in molar terms, and M the initial molecular weight. Each EANC then contributes aRT to the modulus, where a is a front factor (usually unity for ideal gel networks, but sometimes much higher for biopolymer systems, and sometimes temperature dependent). To apply the model, all that is required is to formulate a(t). Originally, it was proposed that for biopolymer gels, particularly those that were near their melting temperature, a reversible cross-linking process could be assumed, leading to an equilibrium between bond making and breaking at long times ( a dynamic gel state). According to this, some strong gels were still to be regarded as temporary networks, the difference in relation to ‘weak gels’ being one of degree (actual rate of cross-link breaking) rather than kind. This led to a(t) being formulated in terms of rate constants for a second order forward pairwise cross-linking reaction of the bonding sites, and a first order back reaction. Equilibrium was assumed at long curing times (cure curve plateaus out). In many practical situations involving strong gel formation, however, cross-linking is believed to occur irreversibly (strong driving force). Homogeneous gel formation under these conditions becomes decreasingly likely, as increased attraction will engender poor solvent conditions and segregation (see next Section). Uniform network formation in these conditions, when it occurs, is most easily explained by rapid irreversible cross-linking, aggregation outpacingdemixing, and leading to kinetic trapping. In this case further reaction between network sites is inhibited by steric hindrance arising mainly as a result of intramolecular cyclisation events. To treat this case a(t) is taken as the solution of differential equations describing competition between a second order cross-linking event, and a site wastage event assumed first order.35 In both the reversible and irreversible models gelation only begins when a exceeds l/(f-l), the critical branching limit.

Polysaccharide Gelation

101

Using equation (2), and a reversible or irreversible model for a(t), the apparent long time limiting modulus can be calculated as a function of C (constant T and M ), as a function of T (constant C and M), and as a function of M (constant C and T). At the same time, from the conditions defining the gel point, and the kinetic equations for a(t), the gel time can be similarly expressed. Success in relation to experiment may be summarised as follows: (1) Experimental modulus-concentration data can be satisfactorily described (Figure 4), including behaviour suggestive of a critical concentration.20,22,33-35 (2) Gel melting curves can be described by the reversible model (Figure 6 ) in terms of an enthalpy and entropy of cross-linking.22,31,32 Comparison of results at different concentrations for locust bean gum, however, indicate inconsistencies which can only be explained32 in terms of irreversible cross-linking during gel setting at low temperatures, which switches over to reversible cross-link melting only close to the much higher melting temperature. This problem is less apparent for pectin and gelatin gels.3 1 (3) Provided f is proportional to molecular weight, the behaviour of the modulus with respect to molecular weight for certain polysaccharides can be explained (Figure S), including the existence of a critical molecular weight.22 (4) Gel time divergence (Figure 4) can easily be explained35 at the critical concentration, but the high limiting power laws found for certain polysaccharide gels (at higher concentrations) can not. Temperature dependence of the gel time has not been examined closely. (5) Long time processes affecting the apparent limiting modulus have not been explained in detail, but are assumed to arise from various network ripening mechanisms associated (among other things) with some slow reversibility of cross-linking (even in apparently irreversible cases), allowing junction zones to change size, and phase separation to proceed slowly. A model combining reversible intermolecular cross-linking and reversible cyclisation might explain some of this, but would be highly parameterised. In relation to the failure with respect to the gel time behaviour of polysaccharides, two changes to the model can be considered, one being to generalise the cross-linking equilibrium to treat allor-none formation of groups of sites (or junction zones) rather than simply pairwise association, the other to look at more complex kinetic events prior to network formation. The first approach was explored by Oakenfull and Scott23924 using a mean field gel model very similar to the one described here, and this did indeed increase the steepness of gel time concentration log plots, the slope at higher concentrations becoming n-1, where n is the number of sites per junction. However, this result is achieved at the expense of the rather difficult concept of a high nth order forward reaction. An alternative approach is to examine the effect of introducing extra kinetic events prior to (or as part of) network building. For example, if highly aggregated junction zone formation in polysaccharides proceeds by irreversible nucleation and growth, rather than by Oakenfull and Scott’s high order reversible association process, sizeable observed power laws could relate to the number of chains in a critical nucleus. A similar result would be obtained if the crystallisation event, though still rate determining, merely produced primary ordered polysaccharide aggregates which went on to randomly associate and build the network. This mechanism has topological advantages in explaining normal polysaccharide gel formation, and certainly occurs in the formation of colloidal particle gels from maltodextrins.25 Temperature influence on gel times for polysaccharides is also better explained by nucleation and growth controlling aggregation, but here the situation is confused by the fact that chain ordering at a lower molecular level (single or

102

Gums and Stabilisersfor the Food Industry 10

double helix formation) is usually also involved and is itself a nucleated process. At present, however, little certainty exists, and one has to admit that the detailed kinetics of polysaccharide gelation, as reflected by gel times, remains somewhat obscure. The properties of the apparent limiting modulus (at longer times), on the other hand, seem to be largely unaffected (Figure 4) by details of the early kinetics, and are much better explained by the random aggregation model in its simplest form. This universality of the modulus-concentration relationship is probably a Straightforward consequence of network building over long distance scales being essentially a random event, unaffected by the exact details of assembly of the primary gelling units. Apart from mean field approaches such as that described here, and those produced by Oakenfull and Scott,23,24 Tanaka36337 and Semenov and Rubinstein,l7 the literature also contains accounts of percolation theory calculations38-40 and off-lattice simulations. 18,19,41-43 Percolation theory has focussed on description of behaviour near the critical gel point (critical exponents) but, so far, few predictions of modulus behaviour, or of gel times, have been presented. Until this is done, its contribution to the problems addressed here seems difficult to assess. Much the same thing can be said about random aggregation models (off-lattice) which have been mainly applied to particle gels (though see work by Groot and Agterofl8,19 for an exception). The fractal description of aggregates and networks which has arisen from this activity, has been related to the gel modulus, and seems to predict a constant power law dependence for this in relation to concentration$5 there being no expected critical concentration. For any given system this powerlaw is related to the appropriate fractal dimension and can vary from gel type to type. This contrasts with conclusions of the mean field approach, which predicts the existence of a critical concentration (often quite a large one), and a non power law master curve for all gel types. So far, practical experimental results for most gels seem to confirm the mean field prediction. Curved modulus-concentration data sets (log-log) are commonly found, and where they are not, the reported constant power laws are often symptomatic of very limited experimental concentration ranges. Currently, simulation approaches19 are beginning to provide values for the gel modulus, rather than relying upon the fractal argument, and it will be interesting to see what modulus-concentration relationships emerge from these. The mean field model is not without its own difficulties, however, particularly where the desire is to connect with molecular properties since, like many of the other models discussed so far, it is expressed in terms of rate constants, enthalpies and entropies of cross-linking, front factors etc, all of which have meaning, but most of which are difficult to relate quantitatively to molecular structure. Simulation approaches may overcome these difficulties, in time, but at present these are not atomistic, and a similar problem of interpreting parameters remains.

8 PHASE BEHAVIOUR OF HYDROCOLLOID SOLUTIONS/GELS Demixing in biopolymer solutions is anticipated from theory, the Flory-Huggins lattice theory 1, for example, for concentrated polymer solutions, suggesting that when the polymer-solvent x value exceeds 0.5, the solution free energy and related component chemical potentials adopt forms which make demixing (into concentrated and dilute polymer phases) inevitable. Increases in x likely to drive this are to be expected as temperature falls, molecules become less charged, or solvent repellent sidechains (e.g. hydrophobic residues) become exposed, and these are of course the conditions which also promote gelation (strong and weak). Gelation and demixing can thus occur together in a quenched system and be intimately related. Where temporary networks form (reversible weak interactions) poor solvent conditions are possible as temperature or solvent conditions vary and phase separation is an issue. Mean field,36937 percolation39 and simulation models18 all suggest an equilibrium phase diagram (T versus concentration, for example) for such systems, in which a miscibility gap is intersected by a critical sol-gel line (Figure 7). This allows homogeneous gels to split into phase-separated

103

PolysaccharideGelation

systems, or even for two gel phases (different concentrations) to equilibrate. Equilibrium conditions should lead to macroscopic phase separation but, even for temporary networks, polymer concentration in the gelled phases, and the lifetime of cross-links, will lead to trapped microstructures which ripen slowly. For the strong gels of the last Section increased cross-linking affinity will emphasise the miscibility gap regions of Figure 7 and make homogeneous gel formation less probable. Kinetic trapping becomes the only way in which high levels of homogeneity can be explained in many situations. It cannot be assumed, however, that such trapping will always be effective, particularly where cross-linking is slow. In such situations, increasing density fluctuations, or even the formation of a network of gelled liquid droplets, will occur as demixing overtakes

T

T

SOL

so L

If

SOL

QC

Figure I General form of phase behaviour of reversibly gelling single biopolymer systems as predicted by meanfield (refs. 36.3 7) and percolation theory (ref: 39). aggregation. Agarose gels, for example, become increasingly heterogeneous46if gelled slowly in conditions close to the ordering temperature, and P-lactoglobulin gels take the form of aggregated spherical droplets47 when heated near the isoelectric point. For strong gels, equilibrium phase diagrams like Figure 7 have little meaning, and are replaced by history dependent ‘state diagrams’ (often very similar, however). Liquid-liquid demixing is also a common feature of ternary biopolymer solutions (two polymers and a solvent) and has a profound effect on the gelling of such mixtures. For these, under most conditions, the Flory-Huggins concentrated solution model predicts (Figure 8) a miscibility gap defined by a binodal boundary and spanned by tielines which will move with temperature. Segregative demixing of a solution, i.e. into two phases rich in the respective polymer components, is promoted by a positive interactive x value (greater than critical) for the polymer pair, by a difference in individual polymer-solvent x values, and by high molecular weight. As Picullel has demonstrated, however, this process is inhibited when a charged and uncharged biopolymer are present, through the ion entropy effect.48 Associative demixing, i.e. formation of a concentrated phase containing both polymers, and a dilute solvent-rich phase, is also possible, this being promoted by an attractive polymer interaction, and polymerx values close to 0.5. Interestingly, the Fdmond-Ogston thermodynamic mode149 for ternary solutions, based on virial coefficients, leads to very similar conclusions, and can readily be applied to mixtures containing globular proteins where the Flory-Huggins theory is inappropriate.

104

Gums and Stabilisers for the Food Industry 10

Gels based on ternary solutions have become very important in modem food technology, e.g. in the area of fat replacement, and more generally. The phase behaviours of the solutions from which these gels form are quite crucial to the outcomes involved.

POLYMER A

30

30

20

20

10

10

I

10

20

30

POLY YER

B

POLYMER

A

10

20

30 POLYMERB

Figure 8 Calculated binodals and spinodals for ternary Polymer A-Polymer B-solvent systems using Flory-Huggins theory. Much more complete segregation occurs as the polymer-polymer interaction (xA&) is increased, and the polymer-solvent parameters (XA~,,& difler more widely. Ms, MA and MB are molecular weights in segment numbers. Complete segregation was assumed in early models for the modulus of composite gels (see below).

9 MIXED HYDROCOLLOID GELS BASED ON SOLUTION DEMIXING From the nature of typical ternary phase diagrams it is clear that starting solutions for gel formation can be chosen compositionally to lie either within or without the miscibility gap. A common situation is to start inside, and prepare a demixed water-in-water emulsion with, in the segregative case, an internal phase rich in one polymer, and a supporting phase rich in the other.

Polysaccharide Gelation

105

Figure 9 Electron micrograph of a gelled water-in-water emulsion (agar-serum albumin system 1/10% w/w) showing protein inclusions (dark) in polysaccharide gelled droplets (light)-rej34 Gelling can then proceed by temperature quenching, or applying multiple quenches, or temperature ramps. The outcome is usually a composite gel containing gelled droplets of one polymer within a matrix of the other (Figure 9). Because of incomplete initial polymer segregation (quite common) the gelled phases may contain a sub-structure of smaller droplets or heterogeneities produced by further segregation during quenching. Other morphologies are possible too, such as networks of gelled droplets interpenetrated by a liquid (or weakly gelled phase) and bicontinuous mutually interpenetrating gelled phases, if surface tension is low, and the system composition lies close to the centre of a tieline. As Brown et a1.50 have shown such morphology can also be influenced by applied shear. The water-in-water emulsion route introduces kinetic issues additional purely to gelling, associated with the stability of the emulsion to coalescence, bulk phase separation etc, during quenching, and the situation needs careful handling. An alternative procedure is to quench from a homogeneous region of the phase diagram and allow phase separation and gelling to compete. This can generate structures very similar to those just discussed, if a water-in-water emulsion forms before gelling, but factors determining drop size, and spatial distribution, may be different, and depend on quench depth and rate. Alternatively, gelling may arrest the structure in its earliest stages, and produce either a molecularly interpenetrating network (IPN), or a system trapped at the level of incipient phase separation. In this last case the mechanism by which demixing occurs may be important. Demixing from within the so-called spinodal region of the phase diagram proceeds rapidly from an unstable equilibrium position by a process of uniform density fluctuations, and when trapped by gelling, can produce uniform segregated microphase morphologies, or a collection of uniformly sized droplets.51 Demixing from outside the spinodal proceeds from a metastable equilibrium state by nucleation and growth, and is slower, and has somewhat different early characteristics. At later stages, however, the two mechanisms may be indistinguishable.

106

Gums and Stabilisers for the Food Industry 10

(Nm”) lo7

lo6 GX

t 1.o

0

Figure 10 Upper and lower bound estimates of composite modulus Gc based on series and parallel addition of component phase moduli weighted according to volumefractions (see rej 34) Modelling the above events is of course difficult. Models are available to describe individual elements of the process such as phase diagrams (including spinodals) and demixing and gelling kinetics, but where gel formation involves complex interplay between these factors, combining the models effectively is a formidable task. The extreme situations of complete segregation prior to gelation (factorisation of events) and no segregation (IPN formation) are probably the easiest to treat. The absence of phase separation is considered in the next Section. For the gelling of wellestablished highly segregated water-in-water emulsions, the principal model available34 addresses the linear shear modulus in relation to polymer composition. A phase diagram, in which the miscibility gap encompasses all compositional space, and has parallel tielines, is envisaged (limiting form of the right-hand diagram in Figure 8), and defined by a parameter p closely related to the tieline slope, and ultimately to the difference between individual polymersolvent x values. Partition of water is thus fully described by p, and this ‘avidity parameter’ is then used to calculate effective concentrations and gel moduli (based on pure component data) for the separate phases. These moduli are then added by upper and lower bound (series and parallel) additivity laws to provide an estimate of the composite modulus (Figure 10). The fully segregative model has had a number of successes, but it has serious limitations, and these should be recognised. First, the additivity laws provide only bounds for the modulus. Real microstructures (e.g. bicontinuous phases) may need more sophisticated treatment if a more precise estimate is needed. Also, importantly, segregation within the phases is unlikely to be complete, certainly initially, and tielines are rarely parallel. This makes the idea of defining concentrations by a single parameter rather doubtful, and there is every chance that gelling of the principal polymer in a phase will be affected by residual amounts of the other, making reference to pure component behaviour unsatisfactory. In addition, such gelling will generally involve processes of secondary demixing, or interpenetrating network formation, leading to more complex microstructure than envisaged originally. Finally, from the point of view of relating properties to molecular structure, such a model depends on several Flory-Huggins x values, and these, as has been said previously, are difficult to relate to molecular structure. Alternative

Polysaccharide Gelation

107

approaches to mixed gel formation based on simulation have been performed for mixed particle gels$2 and look promising, but these are still at the earliest stages of development. 10 INTERPENETRATING AND COUPLED NETWORKS

As mentioned above, the opposite extreme to the formation of fully segregated water-in-water

emulsion gels, is the formation of molecularly interpenetrating networks (IPNs to be distinguished from interpenetrating phases or even microphases). This situation is expected to occur when a homogeneous system is quenched under conditions where at least one component gels rapidly, and homogeneously, and/or the composition remains well outside the binodal phase boundary (even at the new temperature, and even when molecular weight increases commence). As was commented earlier, this last situation is likely to be promoted by combining charged and uncharged biopolymers through the ion entropy effect. If the second component gels less rapidly than the first (usually the case) it will aggregate within the pores of the supporting network, and this may involve a form of phase separation, particularly if the second component normally forms a phase separated gel structure (see gellan Paselli SA2 example in Figure 11). Alternatively, a very uniform mixed network may form where it is difficult to distinguish the components (see gellan - agarose example in Figure 1 1). Interpenetrating molecular networks raise interesting questions about modelling. There is no sense in which their polymer components compartmentalise into separate phases so, effectively, they interact with the same solvent, and feel each other’s presence. At the simplest level, one might propose that the component macromolecules will behave as though alone and at their nominal concentrations. One might then obtain a final modulus by straightforward addition of these separate contributions. In practice, however, the additivity law is likely to be more complex, and either polymer may show severely altered gelation kinetics because of the presence

Figure 11 Electron micrographs show Paselli sA2 aggregates embedded in a tenuouspreformed gellan gel network (le#) while agarose and gellan networks mutually interpenetrate (right). Confocal microscopy (le# insert) shows heterogeneous distribution of starch (light areas) within composite. Note: charged-unchargedpolymer combinations inhibit demixing

108

Gums and Stabilisersfor the Food Industry 10

of the other. This appears to be more true for the gellan-SA2 example recently studied,52 than for the gellan-agarose,53 but at present the general extent of this effect is difficult to assess. Indeed is quite difficult to study experimentally, i.e. to find techniques which can deconvolute and monitor the gelling behaviour of individual polymers in mixtures. Modelling such effects is also extremely difficult as, for any gelling system, the amount of ‘interfering’ polymer varies with system composition, and more than one of the gelling parameters (rate constants, front factor, f, etc) can be affected. Some suggestions for progress in this area have been made,52 but much remains to be done. Superficially, IPNs appear similar to what have been described as coupled or ‘synergistic’ networks. In fact there is the fundamental difference that the former are believed to contain individual polymer networks which interact minimally with each other (topological entanglement) while the latter show a large attractive interaction between the components, with even the possibility that the individual components don’t gel. The formation of coupled networks must to some extent be related to associative phase behaviour, where the polymer-polymer interaction v i a l coefficient implies attraction. Their interaction may well be more specific and stoichiometric, however, than implied by Flory-Huggins theory. Where associative demixing occurs, one might expect liquid-liquid demixing to generate highly phase-separated structures (coacervate gels) but of course synergistic gel networks are often highly uniform. The subject has not been extensively studied from a polymer compatibility point of view, but presumably the homogeneity of coupled networks reflects factors inhibiting demixing similar to those leading to IPNs.

1 1 CONCLUSIONS

The present article has considered only limited examples of biopolymer function, i.e. those mostly related to thickening and gelling. The models discussed have been confined to describing kinetic Simple mean field models have dominated the events or linear mechanical properties. discussion, but references to other lattice and simulation approaches have been given and some comments made. Although difficulties arise for all the models considered, when questions are asked at a molecular level, the approaches highlighted have been successful in describing much of the phenomenology involved, in terms, for example, of universal relationships, critical concentrations and temperatures, forms of phase diagrams, mechanisms of demixing, etc. It remains clear, however, that their underlying parameters, such as x values, rate constants, equilibrium constants, enthalpies and entropies etc, are difficult to calculate from molecular details. Computer simulations may seem more likely to succeed in this respect, by operating at a more fundamental level and being more generally versatile, but so far, the level attained has proved insufficiently atomistic to provide a real breakthrough. The problem of linking molecules to properties is a familiar one to physical chemists, and it clearly takes on particularly formidable proportions in the context of food systems where kinetic control, many chemical components, and large molecules, are the norm. Nonetheless, the search for solutions particularly via simulations, is likely to continue.

ACKNOWLEDGEMENTS The author thanks colleagues at the Colworth Laboratory and King’s College, London, for many helpful discussions and is particularly grateful to Dudley Ferdinand0 and Jenny Brigham of the Colworth Microscopy Unit for provision of electron and confocal micrographs.

Polysaccharide Gelation

109

References 1. P. J. Flory, ‘Principles of Polymer Chemistry’, Cornell University Press, New York, 1953.

S. G. Whittington, Biopolymers, 1971,10, 1481. B. H. Zimm and J. K. Bragg, J. Chem. Phys., 1959,31,476. B. H. Zimm and S. A. Rice, Molecular Physics, 1060,3,391. N. M. W. Haggett, PhD Thesis, University of Surrey, 1998. E. R. Morris, A. N. Cutler, S. B. Ross-Murphy and D. A. Rees, Carbohydr. Polym., 1981, 1,5. 7. B. Launay, J. L. Doublier and G. Cuvelier, ‘Functional Properties of Food Macromolecules’, Eds. J. R. Mitchell and D. A. Ledward, First Edition, Elsevier Applied Science, Barking, 1986,

2. 3. 4. 5. 6.

1.

8. P. E. Rouse, Jr.,J. Chem. Phys., 1953,21, 1272. 9. B. H.Zimm, J. Chem. Phys., 1956,24,269. 10. W. W. Graessley, J. Chem. Phys., 1965,43,2696. 11. W. W. Graessley, Faraday Symp. Chem. Soc.,1983,18,7. 12. P. G. De Gennes, Macromolecules, 1976,9,587. 13. P. G. De Gennes, ‘Scaling Concepts in Polymer Physics’, Cornell University Press, New York, 1985. 14. M. Doi and S. F. Edwards,J. C. S. FaradayII, 1978,74,1789: 1802: 1818. 15. S.B. Ross-Murphy, V. J. Morris and E. R. Morris, Farad. Symp. Chem. SOC.,1983,18,1 1 5 . 16. F. Tanaka and S. F. Edwards, J. Non-Newtonian Fluids, 1992,43,247:273:289. 17. A. N. Semenov and M. Rubinstein, Macromolecules, 1998,31,1373: 1386. 18. R. D. Groot and W. G. M. Agterof, J. Chem. Phys., 1994,100,1649: 1657. 19. R. D. Groot and W. G. M. Agterof, Macromolecules, 1995,28,6284. 20. A. H. Clark and S. B. Ross-Murphy, Adv. Polymer Sci., 1987,83,57. 21. S.B. Ross-Murphy, J. Texture Stud., 1995,26,391. 22. A. H.Clark and D. B. Farrer, J. Rheol., 1995,39,1429. 23. D. Oakenfull and A. Scott, ‘Gums and Stabilisers for the Food Industry 3’, Eds. G. 0. Phillips, D. J. Wedlock and P. A. Williams, Elsevier Applied , London, 1986,465. 24. D. Oakenfull and A. Scott, ‘Gums ans Stabilisers for the Food Industry 4’,Eds. G. 0. Phillips, D. J. Wedlock and P. A. Williams, IRL Press, Oxford, 1988, 127. 25. A. H.Clark, Faraday Discuss., 1995,101,77. 26. R. K. Richardson and S. B. Ross-Murphy, Int. J. Biol. Macromol.,l981,3,315. 27. G. Kavanagh, Ph. D. Thesis, King’s College London, University of London, 1998. 28. S. B. Ross-Murphy, Rheol. Acta, 1991,30,401. 29. A. Tobitani, PhD Thesis, King’s College London, University of London, 1995. 30. C. Rochas, M. Rinaudo and S. Landry, Carbohydr. Polym., 1990,12,255. 31. A. H.Clark, K. T. Evans and D. B. Farrer, Int. J. Biol. Macromol., 1994,16,125. 32. P. H.Richardson, A. H. Clark, A. L. Russell, P. Aymard and I. T. Norton, Macromolecules, 1999,32,1519. 33. A. H.Clark and S. B. Ross-Murphy, Br. Polym. J., 1985,17,164. 34. A. H. Clark, ‘Food Structure and Behaviour’, Eds. J. M. V. Blanshard and P. J. Lillford, Academic Press, London, 1987, 13. 35. A. H.Clark, Polymer Gels andNetworkr, 1993,1,139. 36. F. Tanaka, Macromolecules, 1990,23,3784:3790. 37. F. Tanaka and W. H. Stockmayer, Macromolecules, 1994,27,3943. 38. D. Stauffer, A. Coniglio and M. Adam, Adv. Polymer Sci., 1982,44,103. 39. A. Coniglio, H.E. Stanley and W. Klein, Phys. Rev. B, 1982,25,6805. 40. M. Adam and M. Delsanti, Conternporq Physics, 1989,30,203. 41. E.Dickinson, J. C. S. F a r d a y Trans., 1994,90,173.

Gums and Stabilisers for the Food Industry 10

110

42. E. Dickinson, J. C. S. Faraday Trans., 1995,91,51. 43. B. H. Bijsterbosch, M. T. A. Bos, E. Dickinson, J. H. J. Van Opheusden and P. Walstra, Faraday Discuss., 1995, 101,s 1. 44. M. Gordon and S . B. Ross-Murphy, Pure Appl. Chem., 1975,43, 1. 45. L. G. B. Bremer, T. Van Vliet and P. Walstra, J. C. S. Faraday Trans, I, 1989,85,3359. 46. M. Leone, F. Sciortino, M. Migliore, S. L. Fornili and M. B. Palma-Vittorelli, Biopolymers, 26,743. 47. M. Stading, M. Langton and A-M Hermansson, Food Hydrocolloids, 1993,7, 195. 48. L. Picullel, K. Bergfeldt and S. Nilsson, ‘Biopolymer Mixtures’, Eds. S. E. Harding, S.E. Hills and J. R. Mitchell, Nottingham University Press, Nottingham, 1995, 13. 49. E. Edmond and A. G. Ogston, Biochem. J. 1968,109,569. 50. C. R. T. Brown, T. J. Foster, 1. T. Norton and J. Underdown, ‘Biopolymer Mixtures’, Eds. S. E. Harding, S. E. Hills and J. R. Mitchell, Nottingham University Press, Nottingham, 1995,65. 5 I . A. H. Clark, ibid, 37. 52. A. H. Clark, S. C. E. Eyre, D. P. Ferdinand0 and S. Lagarrigue, Macromolecules, 1999,

manuscript submitted. 53. E. Amici and A. H. Clark, 1999, manuscript in preparation.

RHEOLOGICAL AND THERMAL PROPERllES NEAR TRANSITION OF GELLAN GUM AQUEOUS SOLUTIONS

THE

SOL-GEL

Emako Miyoshi* and Katsuyoshi Nishinari**

*Division of Development and Environment Studies, Osaka University of Foreign Studies, Minoo City, Osaka 562-8558, Japan **Department of Food and Nutrition, Faculty of Human Life Science, Osaka City University, Sumiyoshi, Osaka 558-8585, Japan

1 INTRODUCTION Food hydrocolloids have played a significant role in various fields, especially in the food industry and biotechnology, and have been long used as texture modifiers, emulsion stabilisers, water absorption agents, gelling and thickening agents.',* All these functions of food hydrocolloids affect the textural and rheological properties of a product, both immediately after manufacture and during long term storage prior to use. With increasing use of processed and simulated foods, greater efforts have been made to create unique textured food products, and the textural properties of processed and simulated foods are also mainly controlled by these food hydrocolloids.1.2 Many scientists, especially in the industrialiid countries, are now greatly interested in the potential of these polysaccharides as dietary fibre, or in the formulation of low-fat and lowcalorie foods. The biocompatibility of hydrocolloids can be used not only in drug delivery systems3 but also to supply wound dressings and replacements for human tissue,4 therefore the medical applications have been rapidly advanced as a result of the basic research. Recently, with increasing global environmental problems, there has been growing interest and demand for traditional polysaccharides, because polysaccharidesare ecological and biodegradable. Recently, microbial polysaccharides have become of growing commercial importance and are produced on a large scale by industrial fermentation. These microbial polysaccharides can be produced on demand and with consistent quality, so that availability and variability are not a concern.5 Gellan gum, one of the widely used fermentation materials, may offer a solution to many of problems encountered in current gelling agents? because it can form a transparent gel in the presence of multivalent cations, which is resistant to heat and acid. Since gellan gum can provide a wide-range of gel textures by careful

Gums and Stabi1isers.forthe Food Industry 10

112

control of added salts, these gels can give the same texture as other polysaccharide gels or create new textures. For the improvement of existing products and the development of new ones, the basic understanding of the gelation mechanism at the molecular level is required. As mentioned above, gellan gum is a superior gelling polysaccharide, therefore, many research groups have already reported6-23 on the physicochemical properties of gellan gum aqueous solutions. Consequently, the gelation mechanism of gellan gum solutions has been the subject of some controversy, but it is now accepted that gellan gum may undergo a thermallyreversible ordered helix-coil transition, and the junction zones of gellan gum gels are formed by aggregation of double helical gellan molecules, in a way analogous to the gelation of carrageenans. Although gellan gum forms a loose gel by itself, the physicochemical properties of gellan gels are influenced strongly by the presence of cations, moreover, they are related to the nature of the cations; divalent cations promote the gelation milch more strongly than monovalent cations. Gelation is thus sensitive to cation type, as well as valency. Therefore, the gelation mechanism of gellan gum solutions has not been clarified sufficiently. As is well known for polyelectrolytes,electrolyte type and concentration have a strong influence on their solution and gel properties. A sample of the sodium form of gellan gum with high purity was prepared by San-Ei Gen F.F.I. Inc (Osaka, Japan), and distributed to a number of laboratories.24 This sample was pure and was used without further purification, even to estimate the molecular weight of gellan gum by osmotic pressure25 or light scattering measurements.26 In the present work, the rheological and thermal properties near the sol-gel transition in gellan gum aqueous solutions were investigated by rheological measurement and differential scanning calorimetry (DSC). The comparison of our results with the other results obtained by our coworkers24is also discussed i n this paper. 2 MATERIALS AND MEI'HODS 2.1. Materials

Purified gellan gum in the sodium form was kindly supplied by San-Ei Gen

EEL. Inc., Osaka, Japan was used in the present work. The contents of the inorganic ions Na+, K+, Ca2+and Mg2" were determined as Na 2.59%, K 0.009%, Ca 0.02% and Mg 0.001%, respectively. LiCl, NaU, KCI, CsCl and C a q used in the present study were extra fine grade reagents (Wako Pure Chemical Industries Ltd., Osaka, Japan), and were used without further purification. The gellan gum solutions were prepared in the same way as described previously.'6,*1 The concentration of gellan gum solutions was varied from 1to 5wt%. For samples containing salts, the concentration of gellan gum solutions was fixed at l.Owt% and the concentration of Licl in solutions ranged from 5 to

Polysaccharide Gelation

113

16OmM, that of NaCl, KCl or CsCl ranged from 5 to 1OOmM and that of CaQ ranged from 0.05 to 14mM. 2.2. Rheological measurements Mechanical spectra and thermal scanning rheological measurements were performed within a linear viscoelastic regime with a Dynamic Stress Rheometer DSR from Rheometrics Co.Ltd., NJ, USA, using a parallel plate of 50mm diameter with radial grooves to avoid gel slippage.16r 18-21 The details of rheological measurements have been described previously.16.21 The steady-shear viscosity was measured at various temperatures under conditions of steady-shear on a RFSlI fluids spectrometer from Rheometrics Co., N.J.USA, using acone and plate geometry of 25mm diameter, over a shear-rate range from 0.01 to 100s-1over lOmin.23 2.3. DSC measurements

Differential scanning calorimetry (Dsc)measurements were carried out using a Setaram micro Dsc-III calorimeter, Caluire, Fran~e.17-2~ The sample and reference pans were placed inside the calorimeter and heated to 110°C and kept for lOmin to make the mixture a homogeneous solution. Then the temperature was lowered to 5°C at O.S"C/min and raised again at the same rate up to 110°C. Finally, the temperature was scanned up and down at the same rate, and the enthalpy profiles were recorded. The details of DSC measurements have been d d b e d previously.17.21 3 RESULTS AND DI!XUSSION 3.1. Rheological and thermal behaviour near the sol-gel transition

As shown in the previous studies,1*-21we have determined two transitions of gellan gum solutions (the coil-helii transition and the sol-gel transition) by thermal scanning rheological measurements. The steepest change of G" was attributed to the coil-helix transition, because this transition temperature was in good agreement with the characteristic temperature observed by circular dichroism (CD)?7 It has been suggested28P29 that the peak around 202nm in the CD spectrum reflects the optically active chemical structure of glucuronic acid unit in the random-coiled gellan. When the optically active high ordered structure of gellan gum chains (double helical structure) is formed during cooling, the CD spectrum is changed to the superposition of the spectrum for the randomcoil and that for the double-helix, so that the change of molar ellipticity at 202nm [ 8 ho2is proportional to the change of population of the random-coil. As for the present sample, the coil-helix transition temperature , T of a 1% gellan gum solution determined rheologically almost coincided with the transition temperature ,T 27' C determined by CD measurement30 at the same scan rate OS'C/min. For gellan gum solutions above 2%, the cross-over of G' and G" was observed on cooling, which was attributed to a sol-gel transition. At

Gums and Stabilisers for the Food Industry 10

114

lower temperatures than the cross-over point, the number of aggregated helices exceeded a critical value.18-21 The coil-helix transition (Tch) and the sol-gel transition (T,) temperatures during cooling as a function of concentration of gellan gum determined by the thermal scanning rheological measurement are shown in Figure 1. As illustrated in this figure, states of gellan gum solutions are dependent on the temperature and concentration, and are classified into three; (1)Sol-I (gellan gum molecules take a single coil conformation); (2)Sol-II (gellan gum molecules take a helix conformation but these ordered structure could not lead to the gel); (3) Gel (the number of aggregates of helices exceeds a critical value on cooling and could form a gel). In both Sol-I and Sol-II states, the mechanical spectra showed a liquid-like behaviour,31-35 while those in the gel state tended toward that of weak gel behaviour.32-35 Gellan gum alone 40-

0 1

30-

M

10-

0

0

1 1 2 3 4 5 Gellan Conc. I w t YO

Temperature Tch(0) at which the loss modulus G" Increased steeply, and the cross-Over temperature of G and G" TBe( 0 )for gellan gum aqueous solutions during COOUW! as a function of concentrationof gellan gum from 0 to 5%.

Figures 2(a)-(d)show the frequency dependence of the storage modulus G' and the loss modulus G" for I%, 22, 336, 3.5% gellan gum aqueous solutions at various temperatures. The data are shifted along both the horizontal and the vertical axes by shift factors a and b, respectively to avoid any overlapping. These mechanical spectra were investigated around both Tchand T, as shown i n Figure 1. In all the cases (Figures 2(a)-(d)), G' and G" at temperatures higher than T, were significantly small, however, both moduli markedly increased on cooling below T, which substantiated the results by osmotic pressure,25 light scattering,26 and SAX96 that the disordered structure was a singlecoiled chain and the ordered structure was a double-helix. At any temperature on cooling from T, the viscoelastic behaviour of a 1%gellan gum solution (Figure 2(a)) is typical of a dilute polymer solution with G'
Polysaccharide Gelation

115

indicated that a 1%gellan gum solution could not form a gel on cooling down even to O'C. Frequency dependence of moduli for this solution at 25'C tended to G'--02, G"--o, which is a characteristic feature of dilute polymer solutions.37 However, for a 1%solution at 15,5 or O'C, both moduli of this solution slightly deviated from this behaviour.3Z 38 Similar phenomena were observed for a fairly concentrated gellan gum solution (5%)at 6O'C,39 although gellan gum

b

d

Po d

+I -1

-

A +? A - u l lo-$o-5la+laSl~ZlW11@

-2

101 I d 108

Id

I@ 1 4

104 l ~

'

'

'

'

'

-2

'

/ '/... -

8"

'

'

'

~ ~ e l ~ ' l ~ ~ ~ s ~lo1 ~ 2I 1 d ~ 1 1 @I@

/rad*s-l

Frequency dependence Of the storage modulus G and the loss modulus G"for 1%. 2%. 2.5% 3% and 35% gelIan gun aqueous solutions at various temperatures. The data are shifted along both the horizontal and the vertical axes by shift factors a and b. respectively to avoid the overlapping. Shift factor;a. b For a 2% gellan gum solution (Figure 2 (b)), T, was observed around 9.4"C, so that the mechanical spectrum at 5°C showed weak gel behaviour. For a 3% gellan gum solution (Figure 2(c)) in the temperature range from TQ to TW, the mechanical spectra were typical of a concentrated polymer solution31-35 which showed a cross-over of both moduli at a certain frequency, and the cross-over point shifted to lower frequencies with decreasing temperature. For a 3%

116

Gums and Stabilisersfor the Food Industry 10

solution at 30T, G' and G" were equal over the accessible frequency range, and the slope of double logarithmic plots of both moduli against frequency was approximately 0.5, as would be expected for a critical gel.38 For 3%gellan gum solutions at temperatures lower than T , G' was larger than G" with little frequency-dependence throughout the accessible frequency range, which was classified rheologically as that of a weak ge1.32-35 The gelation mechanism for a 3.5% gellan gum solution showed complicated behaviour. As shown in Figure 1,T, and T, were observed at almost the same temperature (42 "C), so that the mechanical spectra for this solution showed a drastic change around the transition temperature. This solution at temperature slightly higher than the transition temperature showed a liquid-like behaviour, thus, it behaved as a dilute polymer solution at 45T, and. it behaved as a concentrated polymer solution at 43°C. However, at 42T, the behaviour became strikingly different; G' was much larger than G" throughout the experimentally-accessiblefrequency range and both moduli were essentially independent of frequency, as would be expected for an elastic ge1.32-35 These behaviours seem to be close to a true gel, so that these systems could be seen as almost perfect networks. Therefore, the T critical state for a 3.5% gellan gum solution could not be obtained around , Judging from these mechanical spectral change depending on temperature, gellan gum systems were classified into three types rheologically depending on the concentration of gellan gum. The first type (a) includes gellan gum solution at lower concentration (<2%) which showed only a coil-helix transition and could not form a gel (Sol-I-Sol-II). The second type (b) includes the concentrated gellan gum solution (2%-3.5%) in which the coil-helix transition and the sol-gel transition occurred separately (Sol-I-Sol-II-+Gel). The third type (c) includes gellan gum solutions at higher concentrations (>ca.3.5%) in which two transitions occurred concurrently (Sol-I-Gel). As illustrated in Figure 2, gellan gum gels formed in the third type seem to be obviously different from those in the second type. In DSC measurements, for gellan gum solutions of higher concentrations (>4.5%), an endothermic peak on heating split into two peaks, while the cooling curve showed only one main sharp exothermic peak (data not shown). This observation coincides with the result observed by the thermal scanning rheological measurement; the very concentrated gellan gum solutions (>4.5%) showed thermal hysteresis (data not shown). From rheological and Dsc results, more than two kinds of junction zones with different thermal stabilities seem to be formed above 4.5% gellan gum concentrations. It has been observed in the tilting tube method40 that over 4-5% gellan gum solutions formed a turbid gel while below 4% gellan gum solutions formed a transparent gel at any temperature (
Polysaccharide Gelation

117

for 1, 2, 2.5 and 3% gellan gum solutions at various temperatures. The flow behaviour was investigated around Tch as shown in Figure 1. In all the cases (Figures 3(a)-(d)), the gellan gum solutions showed flow behaviour close to Newtonian flow at temperatures higher than T* however, they drastically changed to shear-shinning behaviour at temperatures lower than T, The range of the Newtonian plateau at low shear rate gradually became narrower and the viscosity became more shear-rate-dependent with decreasing temperature. Intermolecular entanglements disrupted by the imposed deformation resulted i n new interactions at low shear rates, however, the rate of disruption becomes greater than the rate of formation of new entanglements, the onset of shear thinning may occur.41 The extent of re-entanglement decreaseswith increasing shear rate, so that the viscosity of the solution may gradually decrease.41 As shown in Figure 3, the steady-shear viscosity measurement indicates that gellan gum solutions tend to show more shear-shinning behaviour with the conformational change from coil to helix because the helix may be more easily oriented along the shear flow than the coil. The range of the Newtonian plateau at low shear rates gradually becomes narrower with development of an ordered structure of the gellan gum solution. 1%GelIan gum

I

6. 10, 16-C

c

2%G e h n Lnm

I

.. 2.6%GelIan sum

shear rate I 8 -3 Shear-rate dependence of the steady-shear viscosity for 1, 2, 25 and 3% gellan gum soluuons at various temperatures.

118

Gums and Siabilisersfor the Food Indusiry 10

In the analyses of SAXS,36 in terms of a molecular model, 18repeating units (72 glucose units) with the straight structure is sufficient to explain the light scattering function, because 72 residues correspond to 39nm according to the molecular weight per unit contour length MLvalue used for the single stranded chain. This value (39nm) seems to be close to the Kuhn segment length of 34nm estimated roughly from the persistence length of 17nm.26 Therefore, the model that the chains behave as a random coil over the contour length of Kuhn length might be reasonable, which fairly coincided with the results obtained Ly SAXS.36 Both the dielectric42 and IHNMR30 measurements have suggested that water is more strongly influenced by gellan gum in the coiled state compared to that i n the helical state, in other words, the interaction between gellan gum molecules was less in the coiled state than in the helical state. This observation could be understood better by the fact that gellan gum molecules behave as a fairly stiff chain even in the disordered state, and gellan gum molecules in the disordered state expand in solution.26 In our rheological results (Figure 3), in any cases, the flow behaviour of gellan gum solution showed a drastic change accompanying the conformational change of gellan gum molecules. Solutions of hyaluronate which take a coil conformation show a pronounced shear-thinning behaviour at any temperature,*3 however, gellan gum solutions in the coiled state showed a flow behaviour much closer to Newtonian flow, even at high concentrations such as a 3% (Figure 3(d)). Graessley and Segal44 have suggested that for polystyrene as the molecular weight becomes greater, the shear-thinning behaviour of solutions becomes more conspicuous. Gellan gum, gellan gum molecules even in the coiled conformation are relatively elongated, so that molecules are less entangled with comparison to flexible molecules. However, the single stranded coil of the substantial stiffness could smoothly associate with another one for the formation of double helices, therefore, the flow behaviour of gellan gum solution seems to show a drastic change over the coil-helix transition. 3.2 Effects of salt on the sol-gel transition

3.2.1 Effects of monovalent cations

In the presence of monovalent cations, for all the cases, the cooling DSC curves showed a sharp exothermic peak shifting to progressively higher temperatures with increasing concentration of added salt. However, during heating, the endothermic peak gradually developed bimodal character and eventually split into multiple peaks upon addition of a certain concentration of salt, and then the endothermic enthalpy at higher temperature gradually increased with increasing In the presence of further sufficient salt concentration of the added salt. (1OOmM KCI or CsCl),a single endothermic peak was observed at markedly higher temperature than the exothermic peak, so the thermal hysteresis was far more modified. The order of effectiveness of the monovalent cations i n

119

Polysaccharide Gelarion

promoting ordered structures (W>K+>Na+>Li+) was correlated with the dynamic hydration number.45 This order is also in good agreement with the effectiveness of the cation in increasing the cross-sectional radius of gyration R , estimated by SAXS.36 The dependence of the exothermic and endothermic peak temperatures for a 1%gellan gum solution on the salt concentration showed a common tendency. At lower concentration of salt, the differences between two transition temperatures were very small, so that the thermal hysteresis was almost negligible. However, upon addition of a certain concentration of salt, the heating Dsc curves began to split into multiple peaks and the thermal hysteresis was observed, and then the thermal hysteresis was gradually modified with increasing concentration of the added salt. Figure 4 shows the exothermic peak temperature T, in the cooling Dsc curves and endothermic peak temperature T,in the heating DSC curves (a), and T, and T, determined by the thermal scanning rheological measurement (b)for 1%gellan gum solutions as a function of the concentration of the added LiCl. As illustrated in Figure 4, the comparison of the rheological results with the DSC results indicated that the onset of detectable splitting in heating DSCarves (the onset of observation of significant thermal hysteresis) was in good agreement with the onset of the observation of two transitions occurring at the same temperature by the rheological measurement. (a) 1% Gellan with LiCl

6o

(b) l%GeUanwithLiCl

6o

mm ""r --iy I rn

I

A

50

:

1

B

I

50

1

C

I

Sol-I

u

Sol-II

'30

H%2 0

-

'"1 0

0

,

10

,

,,;,

,

,

0 20

40 60 80 1 0 0 1 2 0 1 4 0 160

LiCl Conc. / mmol-1

-I

0

20 40 60 80 1 0 0 1 2 0 1 4 0 1 6 0

LiCl Conc. / mmol-1 - I

-4

Exothermic peak temperature T, (0) in the cooling DSC curves and endothermic peak temperature T, (e) In the heating DSC curves (a).the temperature Tch (A) at which G" increased steeply and the cross-over temperature of G and G"Tsg(A)durlng cooling @) for 1% gellan gum solutions as a function of the concentration of the added LEI. Cooling and heating rate :0.5"C Imin.

Figures 5(a)-(h) show the temperature dependence of G' and G" during the cooling and heating processes for 1%gellan gum solutions containing LiCl of

Gums and Stabilisersfor the Food Industry 10

120

various concentrations, and cooling and heating DSC curves for these solutions. Upon addition of 20LiCl (Figures 4(a) and (e)), G" for a 1%gellan gum solution showed one step-like change during both cooling and heating, which was attributed to the coil-helix transition, as mentioned above. The cooling or heating Dsc curve for this solution (Figures 4(a) and (e)) showed a single exothermic or endothermic peak, and this transition temperature almost coincided with T, in the thermal scanning rheology. Upon addition of 40mM LiCl (Figures 4@) and (f)), G" for a 1%solution showed two step-like changes, and the lower temperature process at which G' and G" showed a cross-over was attributed to the sol-gel transition, as mentioned before. On further addition of LiCl, both T, and T,, shifted to higher temperatures (Figures 5(c) and (g)), however the shift in Tegwas far more pronounced than that in T,. In the presence of SOmMLiU, eventually, T, and Tsg occurred concurrently, so that during the cooling process, this solution showed one step-like change of G" and the single exothermic peak was observed (Figure 5(d)),which involved both the coil-helix transition and sol-gel transition. However, the thermal behaviour of this solution in the heating process was quite different from that of solutions with less LiU; G" showed a two step-like change and the endothermic peaks split into two peaks, which would be expected for a significant thermal hysteresis (Figure 5(h)). The endothermic peak at lower temperature corresponded with the single exothermic peak in the cooling DSC curve, however, the other one i n the heating DSC curve did not have any corresponding exothermic peak in the cooling process. UILIcl

ML a i

5oLicl

o

104

10

ao

so

40

50

eo

10

10

SO

40

5O

M)

Temperature / 'C

Fieme5 Temperature dependence of G and G" during cooling and heating processes for 1% gellan gum solutions containing LiCl of various concentrations, and cooling and heating DSC curves for these solutions. (0)G'. (amCooling and heating; rate; 05°CImln.

Polysaccharide Gelation

121

From the rheological and DSC results, the gelation mechanism for gellan gum solutions could be schematically explained as follows. At higher temperatures, gellan gum molecules exist as a single coil (Sol-I) (Figure 4(b)). On cooling, gellan gum chains associate by the formation of double helices to form a certain ordered structure, which contributes to the steepest increase of G" in the thermal scanning rheology and the main exothermic peak observed in the cooling DSC curves. However, this ordered structure, in itself, does not lead to formation of a gel network (Sol-II)(Figure4(b)). Gelation occurs with subsequent aggregation of these helices mediated by cations, so that the sol-gel transition appeared at temperatures lower than the coil-helix transition, which corresponds to the cross-over of G' and G" in thermal scanning rheology and the other small exothermic peak observed in the cooling DSC curves (Figure 5(b),(c)). Gellan gum systems, in which two transitions occurred separately (Sol-IOSol-IIOGel) (Figure 4(b):region B), do not show thermal hysteresis in cooling and heating processes (thermoreversible). Thus, the temperature at which T, or T, is observed during the cooling process is almost the same as that during the heating process. The difference between T, and Tsg decreased with increasing concentration of the added salt (or gellan gum). Eventually, for gellan gum i n the presence of sufficient salt or in sufficiently concentrated gellan gum solution, the individual helices are formed and immediately the aggregation of helices occurred in the cooling process so that T, and T, are observed at almost the same temperature (cooliig)(Figure 4@):region C). Moreover, these systems involve the thermally-stable junction zones formed by fairly aggregated helices, so that these solutions show a large thermal hysteresis (Figure 4(a)). The observations by both IHNMR and CD30 essentially coincided with our rheological results. A 1%gellan gum solution with 20mM KCl showed no thermal hysteresis while a gellan gum solution with 40mM KQ or 80mM KCl showed a large thermal hysteresis and showed the two step-like changes on heating. Judging from our results, a 1%gellan gum solution with 20mM KCl was classified as a system in which two transitions occurred separately (Sol40 Sol-IIOGel), although the difference between T, and T, was small (data not shown). However, 1%gellan gum solutions with over 35mM KQ were classified as a system which diredly changed sol to gel around the transition temperature on cooling (data not shown). In the presence of sufficient salt, more than two kinds of junction zones may be formed because G" showed the two step-liie changes during heating and the heating DSC curves showed two endothermic peaks as shown in Figure 5(h). Therefore, during heating the ordinary double helices melted around the main peak temperature and then thermally-stable aggregated helices (mediated by specific binding of cations) melted at further higher temperatures. In the rheological result, the gel structure seems to remain until the thermally-stable junction zones melt, as will be discussed later. Our rheological and Dsc results also indicated that a 1% gellan gum solution with 40mM KCl or 80mM KCl showed the two step-like

122

Gums and Stabilisersfor the Food Industry 10

change and these transition temperatures, and these peak temperatures in the heating Dsc curves and the rheological transition temperatures precisely coincided with the transition temperatures observed by both lHNMR and CD.30 In CD measurements,30 a gellan gum solution with 80mM KCI showed a large delay of the recovery in [ 6 hO2but this value completely returned to the original one during heating up to 70 ' C. However, in IHNMRP01HT2 did not show the temperature dependence during heating and could not return to the original value nevertheless the double helices seemed to disappear in CD. lHNMR results30 have suggested that in the presence of sufficient K+,the hydrogen bonds within gellan gum molecules, which are formed on cooling, are much stronger, so that the stable hydrogen bonds remain even in the random-coils. This might be explained by the possibility that in the presence of sufficient KCl, the specific interaction between K+ and hydroxyl groups in gellan gum molecules occurs i n addition to the effect of electrostatic shielding. Morris et al.46#47 have found the specific interactions of K+ with K-carrageenan by NMR. As for the present sample, in the presence of sufficient KCI, the transition temperature on cooling observed by1HNMR30 was slightly higher than that by CDPO which indicated the hydrogen bonds were formed in random-coils before the formation of doublehelices. Generally, the effect of salt on gelation is explained by the fact that salt influences the conformation and association of polymers through their effects on the solvent properties of water, but potassium ions could directly promote association of carrageenan because of their ion size.45 Therefore, both specific and non-specific ion effects seem to influence the gelation of gellan gum, and especially in the presence of sufficient salt, the specific ion effect becomes more important. Manning et al.14 have shown the schematic representation of a model on the basis of DSC, optical rotation and rheological studies of gellan in the presence of sodium chloride, and a similar model has been suggested by Morris and G ~ n n i n g . ~ 8 They14 have suggested that the lower temperature peak may be caused by the melting of unaggregated helices while the higher temperature peak may be caused by the melting of aggregated helices. Our interpretation may be essentially consistent with theirs,14 however, this interpretation is not applicable to the gelation mechanism of gellan gum in the presence of divalent cation, as will be discussed later. 3.2.2. Effects of divalent cations

Figures 6 (a)-(f) show the temperature dependence of G' and G" during cooling and heating processes for 1%gellan gum solutions containing Cac1, of various concentrations, and cooling and heating DSC curves for these solutions. The thermal behaviour in the presence of divalent cations was virtually different from that in the presence of monovalent cations. On addition of CaQ of a fairly low concentration (lmM), the cross-over of G and G" was observed around 5°C during the cooling process (Figure 6 (a)), which indicated that this

PolysaccharideGelation

123

system could form a gel on cooling. However, the clear exothermic peak corresponding to the sol-gel transition could not be observed in the cooling DSC curve. Durii the heating process, the temperature at which a steep decrease of G" was observed was slightly higher than that of the main endothermic peak observed in the heating DSC curve, moreover, many small peaks were observed at higher temperatures than the main peak (Figure 6(d)). On addition of only 1.2mMCaC4, G" showed one step-like change around 35°C on cooling, and then at the temperatures lower than this temperature, G' was significantly larger than G". On further addition of Caq, both G' and G" drastically increased, which indicated that an elastic gel could be formed by the addition of sufficient divalent cations (Figures 6 (c)). However, the height of exothermic peak in the cooling Dsc curves gradually decreased with increasing concentration of C a q (Figures 6 (a)-(c)), and this tendency was obviously different from that in the presence of monovalent cations. Moreover, the behaviour of gellan gum solutions with CacI? during the heating process became more complicated. Although the endothermic enthalpy estimated for a main peak gradually decreased with increasing concentration of Caa, the endothermic peaks even with 21nM CacI? were clearly recognized (Figure 6 (e)). However, the thermal scanning rheological results showed that for gellan gum solutions containing more than 1.2mM Cacl, (Figures 6 (e), (f)), no remarkable change of G and G" was observed during heating up to WC.

10 10

hmpaatare I 'C

-6

Temperature dependence of Gand G"durlng cooling and heating processes for 1%gellan gmn solutions containlng CaCl, of various concentrations,and cooling and heating DSC Cooling and heating rate: 0.5%Imln. curves for these solutions. (ow.(Am":

Figure 7 shows cooling and heating Dsc curves of a 1%gellan gum solution containing Cac1, of various concentrations. The concentration of CaQ was represented stoichiometrically (mN). The concentration was also expressed as a

124

Gums and Stabilisers for the Food Industry 10

percentage of the stoichiometric requirement of the carboxyl groups of gellan gum molecules. In cooling DSC curves, T, shifted slightly to higher temperatures by the addition of Caq, and on addition of 3mNCaC4, the second exothermic peak began to develop at a higher temperature and this peak shifted to higher temperatures with increasing concentration of Cac12. Although the total exothermic enthalpies in the cooling Dsc curves significantly decreased with increasing concentration of CaQ, the second exothermic enthalpy gradually developed and then upon addition of 10mNCaC4, two exothermic peaks merged into one peak on the higher temperature side. During the heating process, the endothermic enthalpy estimated for the main peak monotonically decreased with increasing concentration of the added C a q , although many other small peaks appeared at higher temperatures. Moreover, the endothermic peaks for gellan gum solution with 14mNCaC1,were too broad to be resolved from the baseline, although the exothermic peak was sharp and readily recognized. The endothermic enthalpy determined from heating Dsc curves for 1%gellan solutions as a function of concentration of CaQ is shown in Figure 8. The decrease of endothermic enthalpy with increasing concentration of Ca2+ was found to be linear, and the extrapolation to zero enthalpy approached the GI*+ concentration of stoichiometric equivalence (100%).

0

Temperature / ' c

-7

Cooling and heating DSC curves of a 1 % gellan gum solution containing CaChof various concentrations. Figures beside each curve represent the concentration of added CaCI, .

Fiplpe8

EGdothermfc enthalpy from heating DSC curves for 1% gellan solutions as a function of concentration of CaCh.

We propose the following interpretation for the gelation mechanism of gellan gum in the presence of GI2+from the present data. During the cooling process, calcium ions seem to immediately associate with the gellan gum chains at

Polysaccharide Gelation

125

temperatures higher than the conformational transition (the formation of double helices), which contributes the evidence that for gellan gum solutions containing CaQ of low concentrations (3,4,6mN), the second exothermic peak began to develop at slightly higher temperature than the one corresponding with the coil-helix transition. This ordered structure, consisting of the interactions between gellan gum segments and calcium ions, could lead to formation of an elastic gel, rheologically (Figure 6). On further addition of calcium ions, the ordered structures, which are specifically stabilized by calcium ions, stoichimetrically increase and become extremely thermally-stable. Therefore, with increasing concentration of CaQ, the endothermic enthalpy estimated for the main peak in the heating DSC curves monotonically decreased and the other numerous endothermic peaks appeared especially at higher temperatures. During the heating process up to lOO'C, these heterogeneous zones involving the specific cation-polyanion interaction were gradually melted at various temperatures, and then calcium ions may stabilize gellan gum molecules even in the partially disordered helical conformation. Therefore, this may explain the fact that in the presence of any concentration of CaQ, the total endothermic enthalpies in the heating DSC curves were essentially smaller than the total Consequently, divalent exothermic enthalpies in the cooling DSC curves. cations could form electrostatic bonds where a metal ion bridged two anions at a temperature higher than ,T so that the ordered structures stabilised by divalent cations stoichiometrically increased. The same phenomena were observed by both 'HNMR and CDPO In the presence of CaQ two components of the ordered structures in gellan gum systems exist, thus one is reversibly disintegrated on heating and another one shows an irreversibility because of the reinforcement by the electrostatic bonds between carboxyl groups and calcium ions. It has been reported49 that the gelation of alginates requires divalent cations such as Ca2+, and the gels are not thermoreversible. Morris et al.50 have suggested that in alginate gelation, the primary role of Ca*+is to combine pairs of random-coil chain segments into ordered dimers, so that the spedic metal ions are required for the formation of ordered structures of alginate. Gelation of lowmethoxyl pectins is primarily a consequence of ionic cross-linkages through calcium bridges between arrays of carboxyl groups belonging to different chains?' It is well known that two mechanisms have been proposed to explain the gelation process of carrageenan solutions, and consequently, it has been the subject of some controversy. Morris et al.52 have suggested that the crosslinks of carrageenan are formed by segments of a double helix, and then these segments are aggregated by ions such as K+. Smidrod and Grasdalen53 have proposed another mechanism that the single helices are formed, which are subsequently aggregated by K+to dimers, trimers etc. The present work indicates that the gelation characteristics of gellan gum may depend on the nature of the added salt, especially, the gelation mechanism in the

126

Gums and Stabilisers for the Food Industry 10

presence of divalent cations shows a more complicated nature compared to that in the presence of monovalent cations. From these results, the thermally s t a b i l i i structure in the presence of divalent cations was essentially different Here, we would like to from that in the presence of monovalent cations. especially emphasize that monovalent cations play a significant role in promoting the subsequent aggregation of double helices at temperatures lower than the coil-helix transition (T,,,), and that divalent cations directly interact with the gellan gum segments to form the ordered structures at slightly higher temperatures than the formation of double helices (TJ. From rheological and DSC results, it is still not possible to conclude whether the electrostatic bonds where a divalent cation bridges two anions at a temperature higher than T, were formed by the double helices or single helices. Interpretation of the gelation mechanism of gellan gum in the presence of divalent cations is extremely difficult, especially because of the lack of information from visually micrroscopic methods. More investigations by means of different methods have still to be done to clarify it further. There is also an urgent need for purified gellan gums having different molecular weights with narrow molecular weight distribution, to understand better the gelation mechanism. Recently, the advanced researches using Atomic Force Microscopy ( A M ) for investigating the gelation mechanism of gellan gum have been reported.54 These results must give some new information for crosslinking domains of gellan gum. 4.REFERENCES 1.Harris,F!(ed.) 'Food Gels', Elsevier Applied Science, London and New York, 1990, p.1. 2. Nishinari,K and D0i.E. (eds.)'Food Hydrocolloids Structures, Properties and Functions', Plenum Press,New York, 1994, p.1. 3.Uryu,T., 'Hydrocolloids 2 Fundamentals and Applications of Dispersed Systems in Food, Biology, and Medicine', Nishinari,K. (ed), Elsevier Science Publishers, 1999. 4. Phillips,G.O., 'Hydrocolloids 2 Fundamentals and Applications of Dispersed Systems in Food, Biology, and Medicine', Nishinari,K. (ed),Elsevier Science Publishers, 1999, p.3. 5.Sanderson,G.R, 'Food Gels', Harris,I?(ed.),Elsevier Applied Science, London and New York, 1990, p.201. 6. Crescenzi,V., Dentini,M., Coviello,T. and Rizzo,R., Carbohydr.Res., 1986,149, 425. 7. Grasdalen-. and Smidsrod,O., Carbohydr.Polym., 1987,7,371. 8.CTescenZi,V., Dentini,M. and Dea,I.CM., Carbohydr.Res.,l988,21,283. 9. Dentini,M., Coviello,T., Burchard,W. and Crescenzi,V., Macromolecules, 1988, 21,3312.

Polysaccharide Gelation

127

10. Shi,X., PhD Thesis, 1'Universite Joseph-Fourier, Grenoble (In French), 1990. 11.Milas,M., Shi,X. and Rinaudo,M., Biopolymers, 1990,30,451. 12. Robinson,G., Manning,C.E. and Morris,E.R., 'Food Polymers, Gels and Colloids', Dickinson,E. (ed.),Roy.,Soc.Chem., UK, 1991, p.22. 13. Moritakas., Fukubas., Kumeno,K., Nakahama,N. and Nishinari,K., Food Hydrocoll., 1991,4,495. 14. Manning,C.E., PhD Thesis, Cranfield Institute of Technology, Silsoe College, Silsoe, 1992. 15. Special Issue,Food Hydrocoll., 1993,.7,1. 16. Miyoshi,E., Takaya,T. and Nishinari,K., Food Hydrocoll., 1994,8,505. 17. Miyoshi,E., Takaya,T. and Nishinari,K., Food Hydrocoll., 1994,8,529. 18. Miyoshi,E., Takaya,T. and Nishinari,K., Makromol.Symp., 1995,99,83. 19. Miyoshi,E., Takaya,T. and Nishinari,K., Thermochim.Acta, 1995,267,269. 20. Miyoshi,E., Takaya,T. and Nishinari,K., CarbohydtPolym., 1996,36,109. 21. Miyoshi$. PhD Thesis, F a d y of Human Life Science, Osaka City University, 1996. 22. Special Issue, Carbohydr. Polym., 1996,36,1. 23. Miyoshi,E. and Nishinari,K, Colloid Polym.Sci., 1999,277,727-734. 24. 'Physical Chemistry and Industrial Application of Gellan Gum'. Special Issue, Prog.Colloid Polym.Sci., 1999, in press. 25.Ogawa,E., Prog.Colloid .Polym,Sci., 1999, in press. 26. Takahashi,R, Akutu,M, Kubotas. and Nakamura,K., Prog.Colloid Polym.Sci., 1999, in press. 27. Tanaka,Y., Sakurai,M. and Nakamurax, Food Hyrdocoll., 1996,10,133. 28. Crescenzi,V., Dentini,M., 'Gums and Stabilisers for the Food Industry, Vo1.4', Phillips,G.O., Wedlock,D.J. and Williams,F?A. (eds.),IRL Press, Oxford, 1988, p.63. 29. Rinaudo,M., 'Gums and Stabilisers for the Food Industry, Vo1.4', .PhiKps,G.O., Wedlock,D.J. and Williams,PA. (eds.),IRL Press, Oxford, 1988, p.119. 30.Matsukawa,S., Tang,Z. and Watanabe,T., Prog.Colloid Polym.Sci., 1999, in press. 31. Graessley,W. W., Adv. Polym. Sci., 1974,16,1. 32. Morris,E.R., 'Gums and Stabilisers for the Food Industry, V01.2, Phillips,G.O., Wedlock,D.J. and Williams,P.A. (eds.),Pergamon Press, Oxford and New York, 1982, p.57. 33. Clark,A.H. and Ross-Murphy,S.B., Adv.Polym.Sci., 1987,03,57. 34. te Nijenhuis,K., Adv. Polym. Sci., 1%, 130,l. 35. Almdal,K., Dyre,J., HvidtS. and Kramer,O., Polymer Gels and Networks, 1993, 1,s. 36. Yuguchi,Y., UrakawaN., Kitamura,S. and Kajiwara,K., Prog.Colloid Polym.Sci., 1999, in press. 37. Strobl,G.R., 'The Physics of Polymers', Springer, 1996,p.1. 38.WinterJ3.H. and Chambon,F., J.RheoL, 1986,30,367.

128

Gums and Stabilisers for the Food Industry 10

39. Takigawa,T, Nakajima,K. and Masuda,T., Prog.Colloid Polym.Sci., 1999, in press. 40. Izumi,Y., Saito,S. and Somas., Prog.Colloid Polym,Sci., 1999, in press. 41. Morris,E.R., Ross-Murphy,S.B., ' Techniques in Carbohydrate Metabolism., B310, p.1. 42. Shinyashiki,N., Sakai,T., Yamada,G. and Yagihara,S., Prog.Colloid Polym,Sci., 1999, in press. 43. Mo,Y., Takaya,T., Nishinari,K., Kubota,K. and Okamoto,A., Biopolymers, 1999, 50,23. 44. Graessley,W. W. and Segal,L., Macromolecules, 1969,2,49. 45. Uedaira,H. and Ohsaka,A., ' Water in Biological Systems', Kodansha, Tokyo, p.30 (In Japanese). 46. Morris,V.J. and I?S.Belton, J.Chem.Soc.Chem,Commun., 1980,983. 47. Belton,F!S., Morris,V.J. and Tanner,S.F., Int.J.Biol.Macromol., 1985,7,53. 48. Gunning,A.P. and Morris,V.J., Int.J.Biol.Macromo1, 1990,12,338 49. Sime,W.J., 'Food Gels', Harris,P (ed,), Elsevier Applied Science, London and New York, 1990, p.53. 50. Morris,E.R., Rees,D.A., Thom,D. and Boyd,J., Carbohydr.Res.,1978,66,145. 51. Ro1in.C. and De VriesJ., 'Food Gels', Harris,l? (ed.), Elsevier Applied Science, London and New York, 1990, p.361. 52. Morris,E.R., Ree,D.A. and Robinson,C, J.Mol.Biol., 1980,138,349 53. Smidsrod,O. and Grasdalen,H., Macromolecules, 1981,14,1845. 54. Morris,V.J., Kirby,A.R. and Gunning,A.P., Prog.Colloid Polym.Sci., 1999, in press.

Comparison of texture analyser and rheometer measurements on carrageenan and pectin gels

‘,

SOKOL NDONI I , BEMTA U. MARR I , HEIDI NIELSEN INGE-LISE VEDERSB and LME BORREGAARD Copenhagen Pectin A/S,a Division ofHercules Inc.. DK-4623 Lille Skenrved. Denmark Department of Chemistry, Roskilde Universiy, DK-4000. Roskilde, Denmark

ABSTRACT The viscoelastic behaviour of a number of carrageenan and pectin gels was measured in three ways: a- The so-called ’texture analysis’ consisting of stress-strain measurements on a gel confined in a stiff cylindrical dish and having a diameter larger than the probing plunger; b- Parallel plates compression stress-relaxation; c- Parallel plates oscillatory torsion shear deformation. The main scope of the study was to find a correlation between the routine quality control measurement (a) and the material properties [complex moduli (c)]. The results presented here address the low deformation limit of texture analysis, whereby an apparent compression modulus can be calculated. Depending on the diameter of the plunger and the height and diameter of the sample, the values found were up to 15 times higher than the true Young’s modulus. Qualitatively this finding is related to the contribution to the stress of the sample deformed outside the plunger’s cross-section, the lateral confinement of the sample, combined with the incompressibility of the gel. The experimental data were compared with boundary element calculations from ref. 1. The model calculations described the experimental data quantitatively, except for the thinnest samples.

1 INTRODUCTION

The continuous move from more empirical to more research-oriented activity in the quality control of food products, generates increasing interest for relating the routine measurements with fundamental material properties’”. One accepted practice for the characterisation of the mechanical properties of polysaccharide gels is based on texture analyser (TA) measurements. In this context it was of interest to study at first the correlation of the TA stress-strain results in the linear regimezT4with linear elasticity moduli such as the Young’s modulus or the shear modulus4. The basic experiment made on a TA is very similar to a stress-strain experiment made on an axial rheometer with parallel plate geometry. In both cases the raw data are force as a function of the displacement of the moving plate. But while in the case of the rheometer it is easy to translate the plate position into a sample deformation (in the sample-plate slip regime), this is not the case for the texture analyser. The reason is that the geometry is

130

Gums and Stabilisers for the Food Industry 10

more complicated in the case of the TA-sample system, because of (a) the sample confinement in a dish (typically a Bloom glass) and (b) the smaller plunger diameter compared to the sample diameter (Figure 1). From the stress-strain measurements on an axial rheometer, using small enough constant strain rates, it is possible to calculate a linear elasticity modulus (Young’s modulus). It is equal to the initial slope of the stress (force divided by the sample-plate contact area) plotted as a function of strain (the ratio between height change and the original height). This modulus is not dependent on the geometry of the sample, neither on the speed of deformation for speeds under some limiting value, and therefore it is a material property of the gel. If the same exercise is tried with the TA measurements on the same gel, by using the plate area and its displacement to calculate the stress and the deformation, the ‘modulus’ obtained is higher than the Young’s modulus. Furthermore the value of this ‘modulus’ changes with the plate diameter or the dimensions of the dish, showing that we are not calculating a material property, i.e. not a true modulus. The reason for this discrepancy is again to be related to the more complicated geometry in the TA case, which invalidates the calculation of stress and strain in the simple way, used for the axial rheometer parallel plate geometry. The ratio between the Young’s modulus measured by an axial rheometer and the shear modulus measured by a torsional rheometer, both in the linear regime and at the same time-scale, is equal to 3 for incompressible bodies’*4.The quantity of interest in the present study was the ratio between the shear modulus measured by a torsional rheometer and the apparent compression modulus measured by TA. These experimental conversion factors were compared successfully with the predictions of a model calculation from ref. 1.

-r

i

Figure 1 Geometry of the Texture Analyser measurement

131

Polysaccharide Gelation

2 EXPERIMENTAL Materials. One high methyl ester pectin and one I-carrageenan from Copenhagen Pectin AJS were used for the preparation of gels. The polysaccharides were characterised by titrimetry, viscometry, SEC with on-line multidetectors (intrinsic viscosity, right angle laser li ht scattering and refractive index). Pectin had 74.4% degree of esterification, 1.22-10 g/mol viscosity average molar mass (M,) and a polydispersity index of 3.3. The carrageenan had an iota content of at least 98% as determined by Hi h Performance Anion-Exchange Chromatography and '3C-NMR5, a M, of 7.0.10B g h o l and a polydispersity index of 5.0.

9

Preparation of the gels. The gels were prepared in Bloom glasses, Petri dishes or in a steel cylinder with a bottom of Teflon. The latter enables the height of the gel to be changed, therefore making it possible to measure at different sample heights from the same initial gel.

Carrageenan gel composition: (2% carrageenan, 12 % SS, pH 8.4) Carrageenan sugar Potassium Chloride Ion exchanped water Total 1. 2. 3. 4. 5. 6.

1o.og 50.0g 0.5g 439.5q 500.0g

Premix log carrageenan with 50g sugar and 0.5g Potassium Chloride Disperse in 460g ion exchanged water while stirring Heat to boiling while stirring Adjust final weight to 500g Fill into container Store gels overnight at 25°C

Pectin gel composition (0.25% pectin, 62% SS, pH 3.0-3.1) sugar Pectin Buffer solution Ion exchanged water Citric Acid solution (50 w/v%) Total

3 1O.Og 1.25g 112.5g 86.68 2.15ml 500.0g

Buffer solution (IPPA without Calcium)6 Dissolve 3.933g Potassium Citrate, 18ml Citric Acid, and 1.Og Sodium Benzoate in 700g ion exchanged water. Set pH to 3.4-3.5. Transfer quantitatively to a lOOOml measuring flask and fill to the mark with ion exchanged water.

132

Guls and Stabilisersfor the Food Industry 10

1. Premix 1.25g pectin with 5g sugar 2. Disperse in buffer and ion exchanged water with mechanical stirring (800rpm) for 2 minutes 3. Heat to boiling and add remaining sugar 4. Boil until sugar is dissolved 5. Adjust final weight to 5OOg 6. Add Citric Acid solution while stirring 7. Fill into container 8. Store gels overnight at 25°C Rheological characterisation.The gels were characterised on a Texture Analyser XT2 by using plungers of different diameters (Table 1) and compression speeds of 0.1 and 0.5 mm/s. The slopes from the stress-strain curves at small deformations (max. 2 %) were used to calculate an apparent compression modulus in the usual way as for uniaxial stress-strain measurements4 (see the Introduction). Stress-relaxation measurements on the free and on the confined samples were made on the TA-XT2 instrument, as well. The sample height was in the range 10-40 mm and its diameter was 75mm. A plunger of 25.4 mm diameter was used for the gels contained in a dish and a 75 mm plunger for the free-standing samples. An initial fast deformation of 2-3 % at 10 mm/s was kept constant for 100 s. These measurements were used for the estimation of the relaxation time of the gels. The same materials were characterised by dynamic-mechanical measurements with parallel plate geometry on a Rheometrics RS-2000 (stress-controlled). The plate diameter was 40 mm and the gap varied between 3 and 10 mm. The frequency range was 0.1-10 rads and the shear deformation was under 1 YO.The value of the storage modulus at 0.1 rads was taken as shear modulus. This value was estimated4to be at most 2% higher than the stress-relaxation shear modulus taken at 10 s. The temperature for all the rheological measurements was 2 3 3 1"C.

Table 1 Geometry of the samples characterised by Texture Analysis Dish [R x W(mm x mm)J

33.5 x 37.5 (Bloom glass)

Plunger (r/mm)

WR

r/R

18.5 12.7 6.35 3.175

1.12

0.552 0.379 0.190 0.0948 0.734 0.504 0.252 0.126 0.493 0.339 0.169 0.0847

25.2 x 12.5 ( Petri dish )

-11-

0.496

37.5 x 40 (steel & teflon) 37.5 x 20 37.5 x 10 37.5 x 6

-11-

1.07 0.533 0.267 0.160

133

Polysaccharide Gelarion

3 RESULTS AND DISCUSSION The calculated conversion factors f, defined as the ratio between the shear modulus and the apparent compression modulus fiom ref.1 are shown in Figure 2 (symbols) together with the relative polynomial interpolation curves. The interpolation 5* degree polynomials are also given at the inset of the figure. These polynomials are easily generated from a computer program (as f.ex. Lotus or Microsoft Excel) and the plots can be used to convert

0,25 .

0

0,l

0,2

0,3

0,4

0,5

0,6

0,7

0,8

r/R Figure 2 Boundary element calculation results’ (symbols) and interpolation lines. The curves stay for the geometrical factors UR: 0.1; A 0.2; 0.5; 0 1.0. The interpolation polynomials follow the same order from the top.

+

the TA apparent moduli to shear moduli. The results on the experimental conversion factors f a r e shown in the Figures 3a,b,c. The solid lines in the same figures are the interpolations from fig.1. Within the experimental accuracy, the experimental conversion factors of figs. 3a,b,c are the same for pectin and carrageenan. The data are well described by the interpolation lines. Table 2 contains the average values and the standard deviations of the measured moduli that are plotted in fig. 3. The assumptions of the model are not completely stated in ref.1. The gel was modelled as an incompressible elastic body, with a certain shear modulus. From our observation of the TA experiments both the plunger-sample and the walls-sample contact areas were in the stick regime during the measurement. We presume that the same was true for the model calculations, although it is not possible to exclude the possibility that a model with a slip regime could describe our experimental results, as well. No well-defined modelling of the plunger edges were given ine ref.1; for the boundaryhinite element calculations the edge might not be sharp, in order to avoid divergences in stress’. As stated above, the model assumed a perfectly elastic gel. On the other hand the real gels are viscoelasticbodies. How can we rationalise the good description of the experiment by such a drastic simplification at the bases of the theoretical model? The main reason is

Gums and Stabilisersfor the Food Industry 10

134

0,09 P Q

$L

0,05

0,Ol

A UR10.267 (Can)

-

uR=o.2 (ref 1)

Po+ (UR=O.Z (ref 1))

Q

L i

0,03

Figures 3 a, b, c Experimental conversion factors (symbols) us a function of the ratio between the radius of the plunger and the radius of the gel container The lines are interpolation @om fig.2.The error bars are standard deviations of 4-6 measurqments. (a) experimental data close to U R = 1 ; (8) experimental data close to UR=0.5; (c) experimental data close to WR=O.2.

Polysmchride Gelation

135

that the storage moduli of the samples are much higher than the loss moduli (by a factor of 15-30 or more) and therefore are almost frequency-inde endent. In this case, the quasistatic experiments are well-described by elastic models * . By taking the results of the compression experiments at equal time-scales as for the dynamic-mechanical measurements, the same procedure should be valid for gels with higher loss moduli, as well. The stress-relaxation experiments showed that the stress decreased by less than 4% in the time interval between 10 and 100 s after the rapid deformation for gels of both pectin and carrageenen. Therefore, the comparison of the storage modulus at a frequency of 0.1 rads with stress-strain measurements made at 0.5 m m / s on samples with height larger than 5 mm might not cause errors larger than 4% in the conversion factor f. This value is smaller than the experimental scatter as presented by the error bars/ standard deviation values in fig31 tab.2.

r,

Table 2 Experimental geometry parameters, moduli and conversionfactors for pectin and carrageenan gels. R/L

r/R

G’(O.1rad/s)/Pa

Pect. 1.12

0.569 0.379 0.190 0.0948

1.07

0.508

0.339 0.169

0.533

0.0847 0.508

0.339 0.169 0.0847

0.496

0.754 0.503 0.25 1

0.126 0.267

0.508

0.339 0.169

0.0847 0.160

0.508 0.339 0.169 0.0847

I cam

165f15 1360f15

EdPa Pect. I ~ a r z 4700f520 I 1004M900 338W400 I 739M650 4440f560 I8040f750 101OOf2570 I 150001t530 35903~310 I 746M80 32005290 I6050f410 404M520 I6900f740 7900f4900 I 12630f1790 2720f520 I4410f110 2080f250 3 18Of80 2870f850 I4340f330 4040f1200 13790f1790 5480f1370 I 10040f1040 2770f120 I4880GlO 20005290 I3960f14 4230f1280 I5310G50 3090f470 I4680f90 221M30013410f310 1530f160 12540f250 2210H20 I4740f450 4810k840 I 2780k1010 1 140W70 I 1590f370 I

I

faP

Pect.

I

0.0387f0.004 1 0.0538f0.0059 0,041lf0.005 1 0.0 192f0.006 1 0.0471f0.0061 0.0527f0.0057 0.0421f0.0054 0.0270f0.0135 0.0633+0.0108 0.081 7k0.0096 0.0625f0.0182 0.0480f0.0265 0.0341f0.0089 0.064 lf0.004 1 0.0901f0.0137 0.0448f0.0118 0.0552f0.0081 0.0769f0.0076 0.1 101f0.0060 0.0850+0.0283 0.0357+0.006 1 0.0680f0.0265 0.1230k0.0214 0.1 105f0.0273

Carr 0.0352f0.0030 0.0483f0.0036 0.0445+O ,0034 0.0241+0.0007 0.0483f0.0005 0.0596f0.0040 0.0525f0.0056 0.0288+0.0041 0.0817f0.0020 0.1 133f0.0030 0.0832k0.0064 0.1069f0.0504 0.0361f0.0038 0.0738f0.0032 0.0910f0.0003 0.0679f0.0032 0.0769f0.0014 0.1059f0.0095 0.1423k0.0137 0.0763+0.0072

Gums and Stabilisersfor the Food Industry 10

136

From fig.2a,b,c and from results not presented in this paper, it is possible to give an estimate of the range of the geometrical parameters that are preferable for the practical application of the conversion procedure. The range of geometric parameters producing the best reproducible experimental conversion factor f, is L/R 2 0.5 and 0.2 < r/R 5 0.6. Because of the rather large experimental scatter, the presented results were not corrected for buoyancy’; eventual corrections would increase the f-values of Figure 3 by at most 10%.

4 CONCLUSION The conversion factors of apparent compression moduli measured on a texture analyser, to true shear modulus measured on a rheometer, were determined experimentally for gels of one high ester pectin and one 1-carrageenan. Within the experimental error, the conversion factors were the same for pectin and carrageenan. For samples with heights greater than 6mm the conversion factors were quantitatively described by the boundary element calculations of ref. 1. A range of geometrical parameters in TA that gave the best reproducible experimental conversion factors, was specified. The polynomial interpolation curves that can be used to ‘translate’ the routine TA measurements to shear moduli, were presented. References 1. D. G. Oakenfull, N. S. Parker and R. I. Tanner, J. Texture Stud., 1989, 19,407-417. 2. S. B. Ross-Murphy, J. Texture Stud., 1995,26,391-400. 3. H. Nielsen, M. Kolte, B. U. Marr and S. Hvidt, Ann. Truns.NordRheoZ.Soc.,1998,6, 59-62. 4. J. D. Ferry, ‘Viscoelasticproperties of polymers’, 3d ed., John Wiley & Sons, New York, 1980. 5. C. N. Jol, T. G. Neiss, B. Penninkhof, B. Rudolph and G. A. De Ruiter, Neth.AnuI.Biochem., 1999,268(2), 213-222. 6. Private Communication from the International Pectin Producers Association. 7. R. M. Christensen, ‘Theory of viscoelasticity. An introduction’, 2”ded., Academic Press, New York, 1982.

SYNERESIS OF K-CARRAGEENAN GELS AT DIFFERENT KCl AND LBG CONCENTRATIONS

Dave E. Dunstan, Roberto Salvatore, Malin Jonsson and Ming-Long Liao Cooperative Research Centre for Bioproducts Department of Chemical Engineering and School of Botany University of Melbourne, Parkville, Vic. 3052, Australia.

ABSTRACT Measurements to quantify the degree of syneresis in K-Carrageenan gels at different KCl and LBG concentrations have been undertaken. The time evolution of static and dynamic light scattering measurements have been used during the syneresis process to examine the formation mechanisms and nature of the gel structure in these complex systems. Optical microscopy has been used to confirm the interpretation of the data using the above techniques. The K-Carrageenan gels are heterogeneous on micron length scales which vary with the added KC1 concentration. 1 mTRODUCTION

The red algal polysaccharide K-Carrageenan is an important gelling agent in a wide variety of aFplications for functionalities such as texture, water holding and stabilisation.’. K-Carrageenan is a linear polysaccharide composed of alternating 3-linked P-D-galactopyranosyl 4-sulphate and 4-linked 3,6-anhydro-a-D-galactopyranosyl residues. K-Carrageenan in aqueous solution forms thermoreversible gels whereby a conformational transition occurs, which is regulated by the temperature, ionic strength, and nature of the counter ion^.'-^ In the presence of K’, the gel formed is generally described as rigid and brittle, with a tendency to synerese. Synergistic interactions with suggesting galactomannans including locust bean gum (LBG) are widely rep~rted”~ increased gel strength and reduced s y n e r e s i ~ .The ~~’~ precise mechanism by which the gelation occurs is still debated in the literature. There is general agreement however that a two step process occurs where the random coils associate to form helices followed by aggregation of the helices to form a gel:.”-” The aim of this study is to measure and quantify the changes that occur during syneresis using gravimetric analysis. These measurements have been further complimented using conductivity, light scattering and microscopy in order to develop an understanding of the gelation mechanism and gel structure in the K-Carrageenan / KCl and K-Carrageenan / LBG systems. 2 EXPERIMENTAL SECTION 2.1 Materials

K-Carrageenan (from “Eucheumu cottonii”, presumably Kuppuphycus ulvurezii (Doty)

Doty, Sigma type 111, no. C-1263, Lot 127H1222) containing predominantly the

138

Gum and Stabilisers for the Food Industry I0

potassium ion [K, 6.8% (w/w); Ca, 2.4% (w/w); Na, 0.6% (w/w)] was used without further purification. The locust bean gum (Ceratonia siliclua, Sigma, no. G-0753, Lot 127H0527) was also used without any fbrther purification. K-Carrageenan samples were prepared by dispersing the powder in triple distilled water solutions of varying KCl electrolyte concentrations (0, lo-', 5x10", 0.1 and 0.2M) at 8OoC and stirring for 1 hr. The LBG samples were prepared by first dispersing the powder in cold triple distilled water solutions of M KCl and then heating to 90°C and stimng for 1 hr until dissolved. Mixtures of K-Carrageenan and LBG in M KCl were prepared by mixing the two polysaccharide solutions at 90°C for 15 minutes.

2.2 Syneresis Measurements The gels used in the syneresis experiments were prepared by drawing the polysaccharide solutions into 60 mL plastic syringes (O.D. -31.7mm). All the gel samples prepared had the same geometry and a sample volume of approximately 40 mL. The filled syringes were then capped to avoid evaporation and the gels were allowed to set at room temperature for 2 hrs. The gels were then transferred from the syringes into sealed jars (height -84mm,O.D. -54mm) and stored in an oven set at 25 f 0.5"C. The amount of syneresed liquid was measured by weighing the gels after removing the syneresed liquid and physically drying the surface of the gel using extra low lint paper towels. The amount of syneresis was then calculated as a gel mass loss and expressed as a percentage based on a known mass of solution weighed into each syringe. The syneresed liquid samples from the K-Carrageenan gels prepared were analysed for potassium using ICP measurements (Perkin Elmer, Optima 3000 instrument). 2.3 Conductivity Measurements

The conductivity measurements were performed on the various K-Carrageenan / LBG mixtures prepared in M KC1. A YSI Model 35 conductance meter equipped with a YSI Model 3417 platinum-iridium alloy electrode was used to perform the measurements. The measurements were made inside a glass jar that was sealed using a rubber bung with the conductance probe and a thermometer (f 0.5'C) immersed in the solution via the bung. Cooling during the temperature sweep was performed in a thermostatically controlled water bath equipped with a heater and a compressor. The solutions were allowed to equilibrate at each temperature and the conductance was measured. Each of the temperature sweeps took approximately 5 hours. 2.4 Static light Scattering

The static light scattering measurements were performed using the Malvem 4700 nm-'). Fluctuations in the system at a constant angle of 90 degrees ( q = 2.42 x scattered light intensity from the gels were measured for the scattering volume of approximately 5 . 3 ~ 1 0mm3. ~

2.5 Dynamic light scattering Dynamic light scattering was performed using a Malvern 4700 apparatus with a 10 mW Ar+ laser at 488 nm. Measurements were made at an angle of 90 degrees and a temperature of 25"C, unless otherwise stated. The dynamic light scattering experiment involves measurement of rapid time fluctuations in the intensity of the scattered light. The intensity fluctuations arise from the translation diffusion of the ensemble of particles in space which interfere due to phase shifts associated with the random particle motions. The time correlation function (TCF) of the scattered intensity is given by G, (t) = < I(0) I(t) >

(1)

139

Polysaccharide Gelation

where I is the measured intensity at time t and the brackets represent the time average. The second order autocorrelation function is defined as g, (t) = / ,

(2)

where T is the delay time. The experimental autocorrelation function is normalised by the experimental baseline intensity A to give the theoretical field autocorrelation function defined as g,(r) = A + B< gl(t) >2

(3)

where A and B are optical constants for the measuring system. The data is commonly analysed by fitting a normalised second order autocorrelation function defined as g,(r) = { a,exP(-qbfastT)

+ azexP(-q2DslowT)

I*+ 1

(4)

a, and a, are amplitudes of the intensities from the scattering population with diffusion coefficients Dht and Dslow respectively. The apparent diffusion coefficients are interpreted from the measured relaxation times of the autocorrelation function using the relationship

Where

is the relaxation time and q the scattering vector defined as

q = (4xn/h) sin ( 8/2 )

(6)

h is the wavelength of the light, n the refractive index and 8 the scattering angle. The time correlation functions were analysed using the CONTIN algorithm supplied with the instrument. This algorithm fits a multi exponential curve to the time correlation function such that the residuals are minimised over the time range measured. Second order fits were obtained for the probe particles in the polyelectrolyte solutions yielding two apparent diffusion coefficients D,, and Ddow.In all cases the residuals were random and less than 0.005 of the normalised correlation function.

2.6 Optical Microscopy Optical microscopy measurements were performed using a Dimension 3100 Series Scanning Probe Microscope (Digital Instruments). The various K-Carrageenan / LBG mixtures prepared in lo-’ M KCl as per the conductivity measurements were examined. Samples were prepared by placing a drop of each of the solutions onto freshly cleaved mica which were then left to dry in a desiccator overnight. The images were measured on a scale of 800 pm. 3 RESULTS AND DISCUSSION

The results for the gel point determination for the K-Carrageenan / LBG mixtures are presented in Figure 1. Several important features are worthy of note. The pure KCarrageenan solution shows the highest conductivities and the largest, most abrupt decrease in conductivity for the sol-gel transition. The gel set temperature as taken from the highest slope of the curves in Figure 1 is plotted for the different K-Carrageenan / LBG ratios in Figure 2. Here the gel set temperature is shown to decrease systematically with increasing LBG content. This data is similar to that observed by Ikeda and KumagaiI6 where the sol-gel transition is associated with a decrease in conductivity.

Gums and Stabilisersfor the Food Industry 10

140 6 50

6 00 5 50

7

500

I 2

450

E‘i

400

?

s

350

4 - 9 1 k-Car/ t 4 1 k-Car I -7 3 k-Car I -m-3 2 k-Car I +1 1 k-Car I

3 00 2 50

2001 20

,

,

22

24

.

.

.

26

28

30

LBG LBG LBG LBG LEG

I I

32

34

36

38

40

42

44

46

40

50

Temperature (‘C)

Figure 1. Conductivity vs temperature (cooling curves) f o r 1% (w/w) ir-Carrageenan / LBG mixtures prepared in 0.01M KCI

27

1 10

91

41

73

32

$1

KCmrrapmrun/ LBO Rlllo

Figure 2. Gel set temperature (1st derivative) vs K-Carrageenan /LBG ratio for various 1 % (wh)mixtures prepared in 0.OlM KCI

Preliminary data for the pure K-Carrageenan solutions over a range of concentrations has been obtained and shows more abrupt sol-gel transitions than for the corresponding KCarrageenan content of the K-Camageenan / LBG mixtures. Thus the presence of the LBG appears to broaden the temperature range over which the gel forms and hence affects the gelation process. Results for the degree of syneresis versus t [me are shown in Figure 3 for the 1% (w/w) K-Carrageenan gels maintained at 25°C. It hiis been generally reported in the literature that the degree of syneresis depends on the concentrations of gelling cations, so that excessive amounts of these should be avoiided.” For the gel geometry used in the

Polysaccharide Gelation

141

18

14

12 10 I)

B:

a

$ *

8

+No Added KCI -0-0 OOlM KCI

4

2

+o

O1M KCI

-0 -0

05M KCI 1M KCI 2M KCI

+O

0

2

1

3

4

5

6

7

8

Time (Days)

Figure 3 . Syneresis vs time for 1% (w/w) K-Carrageenan at various added KCI concentrations maintained at 2.5 C

0.5

0.6

07

0.4

0.3

0.8

0.9

1

YOHRKCarageena 0.5

02

0.1

0

KHlt LBG

Flgure 4. SyneaisMS ~ - C m r a g mconcentration uith and wilhout the addtion of LBGforgekjmq~rtdin0.011UKClandIejtosynemefor7daysat 25°C

experiments reported here, however the degree of syneresis decreases with increasing added KCl concentration. The degree of syneresis measured for both K-Carrageenan and K-Carrageenan / LBG gels after seven days are plotted in Figure 4. Here the degree of syneresis is observed to decrease with increasing polysaccharide concentration for the KCarrageenan systems and increasing K-Carrageenan / LBG ratio for the mixtures. Both curves show the same behaviour indicating that the LBG has very little effect on the degree of syneresis in these systems with the observed trend being due to the change in

142

Gums and Stabilisersfor the Food Industry 10

0

0 001

0.01

0.05

01

02

Gel Added KCI Concntration (M)

Figure 5. Difference between added K' conc. and syneresed liquid K ' conc. vs added

KCI concentrationfor I% (wh) K-Carrageengels lefr to syneresefor 7 days at 25" C

K-Carrageenan concentration alone. The trends observed in this work for syneresis in KCarrageenan gels as a function of KC1 concentration and the addition of LBG appears to be inconsistent with those of prior studies. These differences may be attributed to the experimental design andor samples used, which should give useful insight into the syneresis process. The concentration of K' in the syneresed fluid was analysed using ICP over the range of KCl concentrations added. Analysis of the syneresed fluid is shown in Figure 5 where the difference between the added K' concentration and the measured K' concentration in the syneresed fluid is plotted versus added KCl concentration. Interestingly, at higher added KC1 concentrations, the difference between the added K' concentration and the syneresis fluid K' concentration increases. This indicates that the gel has a K' concentration higher than that of the added K' concentration as the gel synereses fluid of a lesser K' concentration. This observation is consistent with the observed decrease in conductivity with gelation and the idea that the gel formation mechanism occurs via association of the helices with associated K' ions involved. The volume of the gels change significantly during the course of the syneresis experiments as implied by Figures 3 and 4. Changes in the internal modes of the gels should then be observed as a function of time as the gels synerese. The apparent diffusion coefficients of the polysaccharide molecules have been measured using dynamic light scattering as a function of time for the gels over the critical first week, during which time the rate of syneresis is at a maximum. The measured diffusion coefficients for the different electrolyte concentrations are shown in Figures 6a and 6b. The data shows no systematic trend with time, while the scatter in the data is significant. Reproducibility over short periods of time was not observed with significant fluctuations in the absolute intensity of the scattered light from the gel occurring. Measurement of the angular dependence of the apparent diffusion coefficients after one week equilibration shows the faster mode to increase with q2 and the slower mode to be independent of q2 (Figures 7a and 7b). These observations are consistent with the fast and slow modes post~lated,'~.'~ and those observed for concentrated polymer solutions by Adam and Delsanti.zo,z'Here the fast mode is an elastic fluctuation while the slow mode is a viscous fluctuation.

143

Polysaccharide Gelation

A

L

0

0

I

I 0

A

s

m

0

I

100

0

200

300

Time (hrs)

Figure Cia. Measured diffusion coeflicient (Dfast, vs time for I % (w/w) rc-Carrageenanprepared in a range of KCl concentrations (0 - 0.IM).

2

I

I

A

-. h

rA

3 2 X

A

-

1 -

v

m

4

a"

0

.A M

I

D

O

e m

0

W

8

8 8

0

*mm O

.

0

0

m

A

o

8

rn

A I

I

Time (hrs) Figure 6b.Measured diffusion coeflicient (Dslow) vs timefor I % (w/w) rc-Carrageenanprepared in a range of KCI concentrations (0 - 0.IM).

144

Gums and Stubilisers for the Food Industry I0 20

1

.

1

.

1

.

.

-

.

0

0

m

.

0

0

0

h

VY

“E 0 ‘p

0,

10

-

0

X

0

v

a

0 0

0

I3

.=

0.

2

L

1

-.

=

.

.

. I

’

1

9

.

1

i 0.

’

0

J

,o -

1

0

.

a

El

0

0

1 -

-

a

e l 3

0

“EV

5 X

*.

1

In

‘p

.,

0

0

E I B

e

v

a

. o

* 0-:’ 0.0000

*

o

*

‘

0.0002 0.0004

* ’

.

.

I

0.0006

0

”

.

0.0008

0.0010

.

- 0.0012

Figure l b . Dfast and D,[,, (upper and lower points respectively) vs the square of the scattering vector for 0.I % (w/w) K-Carrageenanprepared in 0.OIM KCI.

Polysacchnride Gelation

145

*

.-x

XX

X*

1

2

Time (s) (xlO3)

Figure 8 . Intensity vs timefor 1% (w/w)K-Cawageenan prepared in 0.01 M KC1 equilibrated for 19 hrs at 25OC

A typical intensity versus time measurement taken 19 hours after equilibration at 25°C is shown in Figure 8. Significant fluctuations are observed when monitoring the gel at a truly thermal equilibrium. Long time fluctuations are occurring in the gel which give rise to the intensity fluctuations and presumably also the variations in the apparent diffusion coefficients. No correlation was made between the measured intensity and the magnitude of the apparent diffusion coefficients. This should form the basis of a future study. The data suggest that the gels are heterogeneous on considerable length scales where there exist polymer-rich and polymer-depleted regions within the gel network, with syneresis occumng via a long time reorganisation of the polymers within the network. The light scattering data indicates that significant changes in the gel on the micron length scale, comparable to that of the scattering volume, must be occumng during the time course of the experiments. In order to understand the nature of the gel structures which give rise to the observed behaviour, optical microscopy measurements have been conducted on the gels in different KC1 electrolyte concentrations. The optical microscope images are shown in Figure 9 for the different added KCl electrolyte concentrations. The images are taken from gels which have been dried in air. Attempts to measure the wet gels were obscured by the small refractive index difference between the solvent and the K-Carrageenan. To some extent, dilation of the gels may have occurred upon drying, although systematic differences in the aggregate structures as a function of KCl concentration are qualitatively observed. The trend observed with increasing KC1 concentration is that of larger aggregates and more defined network structures. The primary aggregate size

Gums and Stabilisersfor rhe Food Industry 10

146

a

b

V*

U*

-600

-400

‘200

.O

C

6

400

600 y*

d

Y*

Figure 9. Optical microscope images of 1% (w/w) K-Carrageenan prepared in: a) wafer b) lo” M KCI c) M KCI d) 5 x M KCI e) 0.1 M KCI

0

2w

400

e

BOO

c*

Polysaccharide Gekztion

147

increases with a greater degree of aggregation of the primary structures at higher concentrations. The size of the aggregates is consistent with the light scattering intensity measurements which indicate that micron size structures must be present to produce the long time fluctuations in intensity. The data are also consistent with the observed syneresis data where the higher KCl concentrations produce a structure which could be expected to be more rigid. The rigid structure would be less likely to undergo a structural reorganisation of the aggregates within the gel network, hence reducing the degree of syneresis. Further work will be undertaken on these systems in order to quantify the nature of the aggregates in the presence of LBG.

4 CONCLUSIONS Gel formation temperatures have been measured using conductivity and show the presence of LBG to effect the temperature range over which gelation occurs. KCarrageenan gels are very heterogenous in nature with macroscopic structures that vary with the added KCl concentration. Syneresis occurs via reorganisation of the micron size aggregates in the gel which has been observed in the static light scattering measurements. The gel formation process appears to occur through a type of phase separation process.

References 1. N.F. Stanley, Carrageenans, in P. Harris (Ed.), ‘Food Gels’, Elsevier Applied Science, Amsterdam, 1990, pp. 79-1 19. 2. G.H. Therkelsen, Carrageenan, in R.L. Whistler and J.N. BeMiller (Eds.), ‘Inductrial Gums: Polysaccharides and Their Derivatives’, 3d ed, Academic Press, San Diego, 1993, pp. 145-180. 3. D. A. Rees, I. W. Steele, F. B. Williamson, J. Polym. Sci., 1969, 28,261. 4. S. Paoletti, F. Delben, A. Cesaro, H. Grasdalen, Macromolecules, 1985, 18, 1834. 5. E. R. Moms, D.A. Rees, G.J. Robinson,J. Mol. Biol., 1980,138,349. 6. C. Rochas, S. Landry, Carbohydr. Polym., 1987,7,435. 7. I.C.M. Dea and A. Morrison, Adv. Carbohydr. Chem. Biochem., 1975, 31,241. 8. C. Viebke, Carbohydr. Polym., 1995,28,101. 9. T. Turquois, C. Rochas, and F.R. Taravel, Carbohydr. Polym., 1992,17,236. 10. T. Turquois, J.-L. Doublier, F.R. Taravel, and C. Rochas, Znt. J. Biol. Macromol., 1994,16, 105. 1 1 . C. Viebke, L. Piculell, and S. Nilsson, Macromolecules, 1994, 27,4160. 12. I. T. Norton, D. M. Goodall, J. Chem. SOC., 1983,79,2475. 13. K. Ueda, M. Utoh, Y. Matsuzaki, H, Ochiai, et al, Macromolecules, 1998, 31, 675. 14. W. Nerdal, F. Haugen, S. Knutsen, H. Grasdalen, J. Biomol. Struct. Dyn., 1993, 10, 785. 15. K. Vanneste, M. Mandel, S. Paoletti, H. Reynaers, Macromolecules, 1994, 27, 7496. 16. S. Ikeda, H. Kumagai, J. Agric. Food Chem., 1998,46,3687. 17. K. B. Guiseley, N.F. Stanley, P. A. Whitehouse, Carrageenan, in R. L. Davidson (Ed.), ‘Handbook of water-soluble gums and resins’, McGraw-Hill, New York, 1980, pp. 5-1 - 5-30. 18. T. Tanka, L. 0. Hocker, G. B. Benedek, J. Chem. Phy., 1973,59,5 15 1 . 19. F. Brochard, P. -G. de Gennes, Macromolecules, 1977, 10, 1157. 20. M. Adam, M. Delsanti, Macromolecules, 1977,10, 1229. 21. M. Adam, M. Delsanti, Macromolecules, 1985,18, 1760.

Rheological studies of Hydroxypropylated and Cross-linked Potato Starch

K. Morikawa' and K. Nishinarib "ResearchLaboratov, Taito Co.,Ltd., Nagata, Kobe 653-0023,Japan bFacultyof Human Life Science, Osaka City University, Sumiyoshi, Osaka 5S8-8585, Japan

ABSTRACT

The characteristics of hydroxypropylated phosphate cross-linked potato starch (HPS) were studied by differential scanning calorimetry (DSC), granule size distribution and dynamic viscoelasticity measurements. The endothermic peak appeared at the temperature range 6om 48°C to 50°C for HPS, whereas it appeared at 61°C for native potato starch (NPS) in the heating DSC curves. The viscoelasticity of heated HFS dispersions was strongly dependent on the degree of modification. Dynamic viscoelastic measurements were performed for 5% dispersions of NPS and HPS heated at various temperatures. The HPS granules heated at the temperature range from 50°C to 100°C for 30 min were not ruptured, and diameters of HPS granules after being heated were strongly dependent on the degree of modification. NPS granules were fully swollen and then ruptured gradually on heating above 70°C. The storage shear modulus (G' ) and the loss shear modulus (G" ) of HPS dispersions decreased and became more 6equency dependent with increasing degree of modification, whereas the viscoelasticity of HPS dispersions heated at the temperature range from 50°C to 100°C were almost independent of heating temperature. On the other hand, G' and G" of NPS dispersions became smaller with increasing temperature of heat treatment. The retrogradation of chemically modified starch were scarcely observed in the heating DSC curves, whilst considerable retrogradation was observed in the native one. It was shown that even a slight modification retarded the retrogradation remarkably.

1. INTRODUCTION Starch is used widely in the food industry as a thickening, stabilizing and gelling agent. Starch granules are insoluble in cold water. On heating starch dispersions in water above a certain temperature, gelatinisation occurs, i.e., starch granules swell, disintegrate and are well dispersed. On cooling heated starch dispersions, a gel is formed and syneresis begins to occur. This phenomenon is called retrogradation. The viscoelasticity of starch gels increases with storage time, and it causes difficulty sometimes in the food industry. Chemical .modification of starch has been used to reduce the change in rheological properties induced by heat treatment and shear and to prevent retrogradation. Recently, there are many types of chemically modified starch, such as starches modified by acid

Polysaccharide Gelation

149

hydrolysis, oxidation, etherification, esterification and cross-linking'*. In addition to these effects, the chemical modification produces considerable change in the gelatinisation and swelling properties of the starch. Many studies of modified starch have been performed3-", but systematic studies on the rheology of chemically modified starch have not been performed sufficiently to develop further applications. The aim of the present work is to examine the effects of degree of modification on the physico-chemical properties, gelatinisat ion and retrogradation characteristics of chemically modified potato starch.

2. MATERIALS AND METHODS Materials Details of native potato starch (NPS) and hydroxypropylated phosphate cross-linked potato starches (HPS) are described elsewhere". The ratio of hydroxypropyl groups (C,H,O) and the ratio of phosphate cross linkages (P0,H) in NPS and HPS1, 2 and 3 are shown in Table 1.

2.1. DSC DSC measurements were performed with a Micro DSC 111 (Setaram, Caluire, France). Each 200mg of starch and 400mg of distilled water was directly weighed into a DSC cell. After sealing, the cell was left for one hour to equilibrate. A cell containing an equal amount of distilled water was used as a reference. The temperature was raised from 20°C to 90°C at l.O"C/min (the first run) and decreased fiom 90°C to 5°C at 2.0"c /min. After the cell was stored for 5 days at 5"c, DSC measurement was performed again (the second run).

2.2. Granule size analysis 1%dispersions of NPS and HPS1, 2 and 3 were preparcd.in distilled water using a motorized stirrer. The dispersions were stirred at 200 rpm for 30 minuies at 2Oac, and heated in an oil bath to 50,60,70,80,90 and l 0 0 O C for 10 minutes and then kept at each temperature for 30 minutes. The hot sample dispersions of NPS and all HPS were suspended with distilled water and immediately granule size distributions were determined with a computer controlled laser diffraction particle size analyzer SALD20005 (Shimadzu Co., Kyoto, Japan). These dispersions were stored at 5"c, and then time dependence of granule size were examined.

2.3. Dynamic viscoelastic measumments The details of preparation of dispersions were described previously". Dynamic viscoelastic measurements were carried out by using a Fluid Spectrometer RFSII or a Dynamic Stress Rheometer DSR 6om Rheometrics Co.,Ltd. (NJ, USA). The storage shear modulus (G') and the loss shear modulus (G") for 3% dispersions of HPS were

150

Gums and Stabilisersfor the Food Industry I0

observed as a function of frequency after being heated at 95°C for 30 minutes. Temperature dependence of G' and G" for 5 % dispersions of NPS and HPS stored at 5°C for 15 hr on the plate of the rheometer was also measured by using a R B I 1 with a parallel plate geometry (25mm diameter and 1.0 mm gap). The temperature was raised from 5°C to 95°C at the scan rate of l.S"C/min and then lowered from 95°C to 5°C at the same rate. The frequency was 1 fads and the strain was 0.01 which was within the linear viscoelastic regime. 5% Dispersions of NPS and HPS were heated at 50, 60, 70, 80, 90 and 100°C for 30 minutes. The hot sample dispersion was poured directly onto the plate of the instrument, which had been kept at 20°C. Frequency dependence of G' and G" at 20°C for these starch dispersions was measured at frequencies ranging from 0.1 to 100 rad/s by using a DSR with a cone-plate geometry (50mm diameter and cone angle 0.04 rad.) after being heated the dispersions at each temperature for 30 min. Time dependence of G' for NPS and HPS heated at various concentrations was examined. The hot sample dispersion was poured directly onto the plate of the instrument, and G' was measured immediately as a function of time at 1 rad/s and at 5°C for 15 hours by using a RFSII with a cone-plate geometry (25mm diameter and cone angle 0.1 rad). The dispersion was covered by silicone oil to prevent the evaporation of water. 3. RESULTS AND DISCUSSION

3.1. DSC Table 1 shows the ratio of hydroxypropyl groups and phosphate cross linkages, gelatinisation temperature and enthalpy (AH) in the fist and second run heating DSC curves for 33% (w/w) dispersions of NPS and HPS. Peak temperatures (T,J of HPS1,2 and 3 ranged 6om 48°C to SOC, and were about 12°C lower than that of NPS. AH, were the same for all HPS. Wootton et all' reported that the T,, decreased with increasing degree of hydroxypropylation, however, T,, of HPS was not different in this study because phosphate cross-linking content was almost equal. Quan et a120 reported that degree of phosphate cross-linking affected the AH in heating DSC curves. This shows that the effects of phosphate crosslinking are more important than those of hydroxypropylation. Generally, unmodified starch granules after gelatinising were crystallized at low temperature for a long time, and the rate of crystallization in starch was known to depend on water content and storage temperature. In the present study, any endothermic peak was scarcely observed for all HPS in the second run, whereas a broad endothermic peak was observed for NPS at a lower temperature than in the first run heating DSC curve.

3.2. Granule size analysis The granule size distribution and average granule diameter are important properties of the dispersed phase which govern the viscoelasticity of gelatinised dispersions of unruptured starch granules". The average diameters of the granules heated at various temperatures are shown in Fig.1. Granule size of NPS increased with heating temperature up to 50°C then levelled off at 70°C. Further increase in heating

151

Polysaccharide Gelation

temperature caused the rupture of granules. The granule size distribution could not be measured when heated at above 80°C because almost all granules were ruptured. This indicates that NPS granules could swell up to a certain volume, and when the granule size reached a limiting value, the granules would rupture. The granule size distributions of all HPS dispersions became similar after being heated at various temperatures, and the average diameter of these granules converged to a constant value, respectively. The granule diameter of HPS dispersions after being heated above 50°C showed a smaller value with increasing degree of modification. There should be an upper limit of granule size for HPS to swell because crosslinking and hydroxypropylation prevented starch granules to swell, which prevented the granule ffom rupture. The average granule sizes for NPS and all HPS are shown in Fig.2. The diameter of NPS decreased with storage time at 5”c,because swollen granules released water during storage. On the other hand, the diameters of all HPS were almost constant. Table 1. The ratio of hydroxypropyl groups, phosphate cross-linkages, gelatinisation temperature and enthalpy for 33% (w/w) dispersions of NPS and HPS. first m second m HP” PH” Tp1’3 TCl” A H1(J/g)’4 To203 Tp;3 Tc2.3 A Ht(J/g)’S 57.2 61.0 74.7

13.7

41.8 55.8 73.4

7.0

HPSl

0.61 0.96

42.9 48.5 66.8

10.0

ND

ND

ND

ND

HPS2 HPS3

1.28 0.88

44.3 50.0 65.6 44.1 49.3 68.0

9.3 10.5

ND

ND

ND

ND

2.51 0.86

ND

ND

ND

ND

NPS

0

0

*l; the % hydroxypropyl groups (C3H70) *2;the % phosphate cross-linkages (F‘0,H) *3; To,Tp and Tc are the onset, peak and conclusion temperatures PI,respectively. *4;A HI is the enthalpy determined from the 1st run heating DSC curves. *5;A H2 is the enthalpy of 2nd run after storage for 5 days at 5°C.

-

180

1

5160 \

1140

140

c u,

E 120 100

.!?

U

u

M

g

80

m

>

80 60

40 20

20 30 40 50 60 70 80 90 100

Heating temperature

20 L 0

/ “C

Fig.1. Average diameter (pm) of NPS and HPS granules after being heated at various temperatures for 30 minutes.

0,NPS; A,HPS1; O.HPS2;O,HPS3

1 I 10

Time

Fig.2.

/

20

30

day

Time dependence of average diameter (pm) of NPS and HPS granules after being heated at 60°C for 30 minutes.

0,NPSA,HPS1; 0,HPSZ; O.HPS3

152

Gums and Stabilisers for the Food Industry 10

3.3. Dynamic viscoelasticity Storage shear modulus G' and loss shear modulus G" for HPS dispersions were almost saturated after 24hr storage. Frequency dependence of G' and G" for HPS dispersions is shown in Fig.3. Both moduli G' and G" of 3% dispersions decreased and became more frequency dependent with increasing degree of modification. Both moduli of HPS depended on swollen granule size after being heated because these granules were nearly ruptured.

z g-4

4

-

-5 -1

*.

***

*. 0

1,

2

Log,,, o / rad-s

Fig.3. Frequency dependence of G' (closed) and G" (open) for 3% dispersions of HpSl (A), HpS2 (U), HPS3 (0)stored at 5OC for 24 hours after being heated at 95°C for 30 min. Fig.4 shows the temperature dependence of G' and G" for 10 76 dispersions of NPS and HPSl stored at 5°C for 15 hr after being heated at 95°C for 30 min. The initial values of G' and G" of NPS dispersions stored at 5°C for 15hr were larger than those before having been stored (closed symbols in Fig.4). The increase of G' and G" during storage at 5°C should be induced by the recovery of the partially ordered structure mainly by the formation of hydrogen bonds. A similar phenomenon was observed by Yoshimura et up for 20% corn starch gels stored at 5°C. G' and G" of NPS dispersions decreased drastically with increasing temperature from 5°C to 95"c, and increased slightly with lowering temperature from 95°C to 5°C because of retrogradation. G' and G" of NPS dispersions did not recover the initial value at 5°C when the temperature was lowered from 95°C to 5°C. On the hand, the values of G' and G" were almost constant at every temperature, namely, there is no temperature dependence in HPSl dispersions after being heated. G' and G" of HPSl dispersions recovered the initial value at 5°C when the temperature was lowered from 95°C to 5°C. To understand the gelatinisation properties, the frequency dependence of G' for NPS and HPS2 dispersions heated at various temperatures for 30 min was observed (Fig.5). Values of G' of NPS dispersions were larger at lower heating temperatures and became

153

Polysaccharide Gelation

smaller when heated at 100°C because almost all starch granules were ruptured. On the other hand, all HPS2 were completely gelatinised at 50°C for 30 minutes, and it showed little change by heating at higher temperatures above 50°C. This indicates that the HPS2 dispersion was completely gelatinised at 50°C for 30 minutes. 5

I

NPS

24 '

4

d

1

HPSl

'3

t

I

a

a

' 3

u

a

z

2 2

M

F2

0 J

-I

I

1 '

o

20 40 60 Temperature /

ao C

1

loo

20

0

40

60

Temperature

80

100

/C

Fig.4. Temperature dependence of G(0)and G(A) for 10%dispersions of NPS and Hpsl stored for 15 hr after being heated 95°C for 30 min. Frequency=lrad/s,Strain = 0.01 . , A ) ; before storage. Closed symbols (

3

I

NPS 2

'

2

b 1

-

b 1

$ 0

4.

$ 0 . J

-1

-1

3

x

70% 80°C 90°C 100°C

2 2 '

b ' z

2 0 1 I

-1 -1

0

1

Loglo w / rad-s-'

2

-1

-1

0 Loglo w

1

/ rad-i'

F i g 5 Frequency dependence of G' and G" for 5% dispersions of NPS and HPS2 after being heated at various temperatures for 30 minutes.

2

154

Gums and Stabilisers for the Food Industry 10

The time dependence of G' for dispersions of NPS and HPSl heated at various concentrations was observed (Fig.6). The values of G' for NPS dispersions increased with time and this tendency was more conspicuous when the concentration was higher. On the other hand, the values of G' for HPSl dispersions at any concentration did not change with time. This indicated that HPSl dispersions after being heated scarcely retrograde because the granule size of HPS did not change over a long storage time (Fig. 2).

5

a. 4

-

A

A

A A

A A

NPS

HPSl

.......

"

0 J

,LA A A A A A 1 -

A A A A A A A

"'

0 0 0 0 0 0 0 0 0 0 0 0..

5

J

1 -

"

0 - ' " " " " " " "

0

5

NPS

-

.4

HPSl

a

-

3'

0

A A A A A A A A A A A A A

J l -

o*...........". 0

* . . - ' . - * * ' . ' s . -

0

5

10

15

Time / b

0

5

10

15

Time / hour

Fig.6. Time dependence of G' at 5°C for various concentrations of NPS and HPSl dispersions after being heated at 95°C for 30 min. o = 1 rad/s. Concentrations: 0 , 5 % ; A,lO%; .,15%

+,20%; 0 , 2 5 % ; A,30%.

4 . CONCLUSION The results of the present study indicate that HPS are gelatinised at much lower temperatures than NPS. The HPS granules heated at 100°C were not ruptured. The granule diameter of HPS after being heated and the elastic moduli of HPS dispersions were strongly dependent on the degree of modification. The viscoelasticities of HPS dispersions after being heated at temperatures ranging from 50°C to 100°C were almost independent of heating temperature. It was shown that even a slight modification retarded the retrogradation remarkably. The size of heated HPS granules was constant with increasing time.

Polysaccharide Gelation

155

REFERENCES 1. M. W. Rutenberg and D. Solarek, Starch derivatives: production and uses. Starch; Chemistry and Technology 2nd ed., R. L. Whistler, J. N. BeMiller and E. F. Paschal1 Eds., Academic Press, New York, 1984,311. 2. A. Rapaille and J. Vanhemelrijck, Modified starches. Thickening and Gelling Agents for Food. 2nd e d , A. Imeson Eds., Blackie Academic and Professional, London, 1997, 199. 3. A. Yeh and S. Yeh, Cereal Chem., 1993,70,596. 4. C. G. Biliaderis, J. Agric. Food Chem., 1982,30, 925. 5 . C. Yook, U.-H. Pek and K-H. Park, J. Food Sci., 1993,58,405. 6. H. R. Kim and A-C. Eliasson, Carbohydx Polym., 1993,22,31. 7. H. R. Kim, A-C. Eliasson and K Larsson, Carbohydx Polym., 1992,19,211. 8. H. R. Kim, A-M. Hermansson and C. E. Eriksson, Starch, 1992,44, 111. 9. H. R. Kim, R Muhrbeck and A-C. Eliasson, J. Sci. FoodAgric., 1993,61,109. 1O.K Morikawa and K Nishinari, Hydrocolloids 2, K Nishinari Eds., Elsevier Science B.V., Amsterdam, 1999, in Press. ll.M. Wootton and A. Manatsathit, Starch, 1983,35,92. 12.M. Wootton and A. Manatsathit, Starch, 1984,36,207. 13.N. A. Abdulmola, M. W. N. Hember, R. K Richardson and E. R. Moms, Carbohydx Polym., 1996,31,65. 14.R K Hari, S. Garg and S. K Garg, Starch, 1989, 41, 88. 15.R. Hoover and F. W. Sosulski, Starch, 1985,37,397. 16.R. Hoover and E W. Sosulski, Starch, 1986,38,149. 17.R. Hoover, D. Hannouz and F. W. Sosulski, Starch, 1988,40,383. 18.R. Kavitha and J. N. BeMiller, Carbohydx Polym., 1998,37,115. 19.S. Takahashi, C. C. Maningat and P. A. Seib, Cereal Chem., 1988,66,499. 20.Y. Quan, M. R. Kweon and F. W. Sosulski, Starch, 1997,49,458. 21.Y. Wu and P. A. Seib, Cereal Chem., 1990,67,202. 22.P. E. Okechukwu and M. A. Rao, Gums and Stabilisers for the Food Industry 8, G. 0. Phillips, P. A. Williams and D. J. Wedlock Eds., IRL Press, Oxford, 1996, 49. 23.M.Yoshimura, T. Takaya and K Nishinari, Food Hydrocoll., 1999,13, 101.

Gelling Mechanisms of Non-Starch Polysaccharides (NSP) from Pre-Processed Wheat Bran Fraction W. Cui, P.J. Wood and Q. Wang Food Research Program Southern Crop Protection and Food Research Centre Agriculture and Agri-Food Canada

Guelph, Ontario, Canada

ABSTRACT Non-starch polysaccharide (NSP) extracted from pre-processed wheat bran contained arabinoxylan (77%) and P-D-glucan (23%), and exhibited thermally reversible gelling properties upon cooling to 4°C. Because there are at least two types of polysaccharides present in the NSP, it is important to know which polymer is the component responsible for this gelling behavior. In order to study the contribution of each polymer to the gelling property, two highly specific enzymes were used to depolymerize the individual polymers, namely xylanase (T. Viride, EC 3.2.1.8) for arabinoxylans and lichenase (EC 3.2.1.73) for P-D-glucans. It was observed that xylanase completely destroyed the gel structure formed by NSP at 2.0% (w/w). In contrast, lichenase did not destroy the gel, but instead enhanced gel strength. It was also found that purified 2.0% (w/w) wheat P-D-glucan formed a thermally reversible gel, but needed a much longer period to develop a similar gel. No gel was formed at a concentration of 0.5% (w/w), the equivalent B-D-glucan concentration as exists in 2.0% (w/w) NSP solution. The purified arabinoxylan was able to form a gel, but, the gel strength was much weaker compared to that of NSP. In conclusion, arabinoxylans are the major components responsible for the novel gelling properties of wheat bran NSP; P-D-glucan, however, assisted the gel formation, possibly through association interactions with the arabinoxylans.

INTRODUCTION The oxidative gelation of water-soluble pentosans isolated from wheat has been well documented due to its important functionality during bread making process (Hoseney and Faubion, 1981; Izydorczyk et al., 1991; Geissmann and Neukom, 1973). In contrast to the water-soluble wheat flour pentosans, aqueous solutions of non-starch polysaccharides (NSP) from pre-processed wheat bran which contained 77%

Polysaccharide Gelation

157

arabinoxylans and 23% P-D-glucan, formed gels. A strong gel can be obtained for 2.0% (w/w) NSP at 4OC for 16 h (Cui et al, 1998, 1999). The gel structure changed as the temperature increased from 4OC to 7OoC with an apparent structure breakdown temperature of -4OOC. In continuation of previous studies, the objectives of the present study were to (i) identify the component(s) responsible for the gelling properties; (ii) to elucidate the gelling mechanisms. In order to achieve the objectives, two experimental approaches were taken. The first was to study gelling properties of NSP and NSPs treated with lichenase and xylanase separately, one at a time. The second approach was to study the gelling properties of arabinoxylans and P-D-glucans in isolation and in mixtures reconstituted in the ratio of the original NSP. MATERIALS AND METHODS Materials NSP was extracted as described previously (Cui et al. 1998). P-D-xylanase (T. Viride, EC 3.2.1.8), lichenase (EC 3.2.1.73) and P-D-glucosidase were purchased from Megazyme International, Co. Wicklow, Ireland. All other chemicals were of reagent grade unless otherwise specified. Preparation of Arabinoxylan NSP solution (l%, 300 ml) was adjusted to pH 6.5 by addition of 20 ml of sodium phosphate buffer (1M). lml lichenase (lOOOu/ml) was added to the NSP solution and stirred at 50°C for 2h. The system was heated at 100°C for lOmin to deactivate the enzyme. The cooled solution was centrifuged at 5,00Og, 25OC for l0min. The Supernatant was mixed with an equal volume of 100% ethanol to precipitate the polysaccharides. The precipitate was redissolved in 300ml water, dialyzed against distilled water for 24 h (water change 3x). After dialysis, the solution was precipitated in 50% ethanol, and then suspended in 100% 2-propanol and kept at 4OC overnight. After removal of solvent the precipitate was dried with gentle warming (Wood et ul, 1989). Preparation of P-D-Glucan NSP solution (l%, 300 ml) was adjusted to pH 4.75 with 0.25 M sodium acetate buffer. Xylanase (-100 unit/lOO ml of solution) was added and kept at 5OoC under constant stirring for 2 h. The enzyme was deactivated by heating at 80°C for 30 min, and the solution centrifuged at 5,00Og, 25OC for 20 min. The Supernatant solution was made to 50% ethanol (final concentration), and the precipitate recovered by centrifugation. The resulting precipitates were washed with 50% ethanol, and dried as for arabinoxylan.

158

Gums and Stabilisersfor the Food Industry 10

Molecular Weight Determination The molecular weight distribution of P-D-glucan and arabinoxylan was determined by high performance size exclusion chromatography (HPSEC) (Beer et al., 1997) using two columns in series (Shodex Ohpak Kl3-806M, Showa Denko K.K., Tokyo, Japan; Ultrahydrogel linear, Waters, Milford, USA). Samples were filtered (0.45 pm) before injection of 100 pl. The columns were eluted with 0.1 M NaN03 (0.6 ml/min) at 40'C. P-D-glucan was monitored with both viscometric detection and post column calcoflour detection (Perkin-Elmer LS-5 Spectrofluorimeter), while the arabinoxylans were only detected by viscometry. Peak molecular weight (MWp) of P-Dglucan was calculated using a calibration curve obtained from seven P-D-glucan standards (Beer et al., 1997). The molecular weight of arabinoxylan was determined from the viscometric peak retention volume and P-D-glucan calibration. Monosaccharide Analyses Monosaccharide analysis of NSP, arabinoxylans and 0-D-glucan was carried out by hydrolysing the polysaccharides in 1M HzS04 at l0O'C for 2 hrs. The hydrolysates were diluted and filtered, then analysed by a Dionex system as described by Wood et al. (1994). Rheological Measurements and Sample Preparation Solutions of (2.0% w/w) NSP, wheat P-D-glucan and wheat arabinoxylan were prepared in distilled water by heating at 80'C for 30 min, then kept at 60°C for 2h before cooling to 25OC. Centrifugation was used to remove air bubbles if necessary. Dynamic rheological tests were performed on a Bohlin CVO stress controlled rheometer (Bohlin Instruments, N.J. USA) with a cone-plate geometry (4' angle and 40 mm diameter), gap 150 mm. Onset of gelling was monitored at 4'C over 16h at a frequency of O.1Hz. Heating of the gel was carried out at O.S"C/min from 440°C. Low viscosity silicon oil (5 centipoise at 25°C) was used to cover the sample to prevent solvent evaporation. RESULTS AND DISCUSSION Gelling Properties of NSP and Enzyme Treated NSPs In a previous paper, we reported that 2.0% (w/w) NSP solutions formed gels when stored at 4°C overnight (Cui et al., 1998 and 1999). The gel development curves, i.e. plots of storage G' and loss (G") moduli against time of 2.0% (w/w) NSP and two NSP solutions treated with different enzymes, are shown in Fig. 1. At 4"C, G' of NSP increased rapidly in the first 15 min, then increased gradually until it reached a plateau (-5-6 hr). G" remained unchanged over the whole period. An increase of gel strength was observed when the same NSP solution (2.0%, w/w) was treated with lichenase for

Polysaccharide Gelution

10 8 6

159

I

I

0

W

G

5000

0

10000

15000

20000

30000

25000

Time (sec)

Fig. 1. Plot of G’ and G” (4°C) against time of 2.0%(w/w) aqueous solutions of nonstarch plysaccharides isolated from pre-processed wheat bran fractions. NSP: original; NSP-1, lichenase treated; NSP-2: lichenase and P-glucosidase treated. 9.6

7.6

(I

b

5.6

b

9

3.6

1.6

0.4

0

10

20

30

40

50

60

70

80

SO

Temperature C

Fig. 2. Melting curves of 2.0% (w/w) aqueous solutions of non-starch plysaccharides isolated from pre- processed wheat bran fraction. NSP: original; NSP-1, lichenase treated.

160

Gums and Stabilisersfor the Food Industry 10

2h: the overall gel strength ( G ) of NSP-lwas much greater than that of the original NSP. It was expected that lichenase would have hydrolyzed 0-D-glucan into oligosaccharides; as a result, the overall gel strength would have been reduced because of the breakdown of the P-D-glucan. The increase of gel strength indicates that the P-Dglucan may not be hydrolyzed completely to oligosaccharides; instead, it may be partially hydrolyzed to a smaller molecular size P-D-glucan. Low molecular weight PD-glucan has been proved to form gels faster than high molecular weight species (Bohm and Kulicke, 1999). However, it is not clear whether the increased gel strength of NSP aqueous solutions was caused by association between the arabinoxylan and P-D-glucan or by mutual exclusion and the formation of two phases. In order to clarify the role of PD-glucan in the gel formation of lichenase treated NSP (NSP-I), P-D-glucosidase, an enzyme used in the standard analytical method to hydrolyze oligosaccharides from lichenase hydrolysis, was added to the solution (NSP-2). This enzyme further hydrolyzes lichenase released oligosaccharides to produce glucose. After treatment with P-D-glucosidase for 2h, the gel strength of NSP-2 was reduced somewhat compared to lichenase treated (NSP-l), but was still higher than the original NSP, as shown in Fig. 1. This result suggests that the lichenase may not have completely hydrolyzed the P-Dglucan to oligosaccharides which are readily accessible to the P-D-glucosidase. If that is so, the slight decrease of G after the addition of P-D-glucosidase might be caused by the removal of glucose from the end of the partially hydrolyzed P-D-glucans which either interacted with arabinoxylans or formed intermolecular junction zones between the P-D-glucans. When a 2.0% (wlw) NSP solutioddispersion was treated with xylanase (at 5OOC) for 2h, there was no gel formation (data not shown). In fact, the system became water-like due to breakdown of the arabinoxylans (77% of the mixture) into oligosaccharides. The heating curves of NSP and lichenase treated NSP-1 gels are shown in Fig. 2. NSP had a sharp melting point at -22"C, while the melting point of the lichenase treated sample (NSP-1) shifted to 25°C with a broader range of transition temperature. The reason for the increase in G' with temperature before the melting point is not clear. Changing the geometry and increasing the gap of the rheometer showed that this phenomenon was not an artifact. The increase of G' with temperature could be simply due to the increase in hydrodynamic volume, which increased the inner pressure. The build up of inner pressure might be caused by the edge effect of the geometry. When the temperature increased to a certain extent, the disruption of the gel structure outweighed the expansion of hydrodynamic volume, therefore, G' falls.

Polysaccharide Gekztion

161

0.4

0.3

b CJ

6 0.2 0.1

0 0

WOO

loo00

1MMO

zoo00

2

m

Time (Sec)

Fig. 3. Plot of G’ and G” (4°C) against time of 2.0% (w/w) aqueous solutions of arabinoxylan (AX) and their mixtures with P-D-glucan at different ratios.

Gelling Properties of Arabinoxylans, P-D-Glucans and Their Mixtures The purities and molecular weights of isolated arabinoxylan and P-D-glucan were determined by monosaccharide analysis on a Dionex system after hydrolysis with 1M sulphuric acid and by high-performance size-exclusion chromatography (Wood et al., 1991, 1994). Only arabinose and xylose were present in AX and glucose in P-Dglucans. The peak molecular weight of arabinoxylan and P-D-glucan were about 500,000 and 300,000 respectively, both decreased compared to NSP (700,000) (the value for AX is P-D-glucan based). The development of gel in 2.0% (w/w) AX and its mixtures with P-D-glucan is presented in Fig. 3. Pure AX formed a gel under identical conditions to NSP. However, the gelation rate and the overall gel strength (G’) were an order of magnitude lower than that of NSP (Fig.1 and 3). When arabinoxylan and P-D-glucan were mixed in the ratio of 9:l and 8:2, respectively, both gelation rate and gel strength further decreased. It appears that the gel strength is related to the amount of AX, and there was no observable synergistic interactions between the AX and P-D-glucan after remixing. The mechanical spectrum of NSP and AX gels are shown in Fig. 4 Isolated P-D-glucan did not form a gel within the time period examined (Fig. 5), but formed a strong gel after three days storage at 4°C. The heating curve of the gel obtained from 2.0% (w/w) wheat P-D-glucan after storing at 4°C for 72 h is shown in Fig. 6. Wheat P-D-glucan formed a much stronger gel compared to the NSP gels. with order of magnitude larger G’. The melting point of this gel was about 40”C, indicating

162

Gums and Siabilisers for the Food Indusiry 10

that it is more temperature tolerant than the NSP gel. However, at 0.5% (w/w) concentration, wheat P-D-glucan did not gel after storage at 4°C for 3 days.

100

-e (I

i?, b

10

1

0.01 0.01

0.1

1

10

100

Frequency (Hz)

Fig. 4. Mechanical spectra (4OC) of 2.0%(w/w) NSP (A) and 2.0% (w/w) arabinoxylan ( B) isolated fiom pre-processed wheat bran h t i o n .

163

Polysaccharide Gelation 0.016

b

o

O

*

-

O

0

0

5000

10000

15000

20000

25000

30000

35000

40000

Time (Sec)

Fig. 5. Plot of G‘ and G” of 2.0% (w/w) wheat P-D-glucans against time at 4°C.

0

P

40

60

80

TemperatueC

Fig. 6. Melting curve of 2.0% (w/w) wheat P-D-glucans gel obtained at 4OC for 72h.

CONCLUSIONS Arabinoxylan appears to be the component chiefly responsible for the gelling properties of the non-starch polysaccharides extracted from pre-processed wheat bran fraction. However, the role of P-D-glucan cannot be ignored. Although the recomposition of the arabinoxylan and P-D-glucan did not give the gelling properties observed for the NSP, the presence of 0-D-glucans significantly increased the gel strength of NSP. This may be particularly a property of the lower MW P-D-glucans. The

Gums and Stabilisers for the Food Industry I0

164

low melting point of NSP (lower than body temperature) may have an advantage in food applications: it may be possible to make fat mimetic products with low “melting” points. The gelation of this NSP involved physical bonding rather than the nonreversible chemical cross-links known for water-soluble wheat pentosans (Geissmann and Neukom, 1973; Hoseney and Faubion, 1981; Izydorczyk et al, 1991).

ACKNOWLEDGMENT The authors thank,Ms Cathy Wang and Mr. John Weisz for technical assistance.

REFERENCES Beer, M.U., Wood, P. J. & Weisz, J. 1997. Molecular weight distribution and (1+3)( 1+4)-P-D-glucan content of consecutive extracts of various oat and barley cultivars. Cereal Chern, 74:476-480. Bohm, N. and Kulicke,W-M. 1999. Rheological studies of barley (1+3)( 1+4)-pglucan in concentrated solution: mechanistic and kinetic investigation of the gel formation. Carbohydr. Polym. 3 15:302-311. Cui, W., Wood, P.J., Weisz, J. and Mullin, J. 1998. Unique gelling properties of nonstarch polysaccharides from pre-processed wheat bran. In “Gums and stabilisers for the food industry+”, eds Willians, P.A. and Phillips, G.O.The Royal Society of Chemistry, Cambridge, UK. p34-42. Cui, W.; Wood, P.; Weisz, J.; Beer, M. U. 1999. Non-starch Polysaccharides from PreProcessed Wheat Bran: Carbohydrate Analysis and Novel Rheological Proterties Cereal Chem, 76:129-133. Geissmann, T. and Neukom, K.1973. On the composition of the water-soluble wheat flour pentosans and their oxidative gelation. Lebensm. Wiss. Technol. 6:59-62. Hoseney, R.C. and Faubion, J.M. 1981. A mechanism for the oxidative gelation of wheat flour water-soluble pentosans. Cereal Chem., 58:421-424. Izydorczyk, M., Biliaderis, C.G. and Bushuk, W. 1991. Oxidative gelation studies of water-soluble pentosans from wheat. J. Cereal Sci., 11:153-169. Wood, P.J., Weisz and Blackwell, B.A. 1991. Molecular characterization of cereal p-Dglucans Structural analysis of oat P-D-glucan and rapid structural evaluation of pD-glucan from different sources by high-performance anion-exchange chromatography of oligosaccharides released by lichenase. Cereal Chem., 68:3139. Wood, P.J., Weisz and Blackwell, B.A. 1994. Structural studies of (1+3),(1-+4)-p-Dglucans by C’3-nuclear magnetic resonance spectroscopy and by rapid analysis of cellulose-like regions using high-performance anion-exchange chromatography of oligosaccharides released by lichenase. Cereal Chem., 71 :301-307. Wood, P.J., Weisz, J., Fedec, P., Burrows, V.B. 1989. Large scale preparation and properties of oat fractions enriched in (1+3),( 1+4)-P-D-glucan. Cereal Chem. 66:97- 103.

Mixed Biopolymer Systems

PHASE SEPARATION IN MIXED BIOPOLYMER SYSTEMS

L. Lundin', I. T. Norton', T.J. Foster', M.A.K. Williams'. A-M. Hermansson2and E. Bergstrom3. 1. Unilever Research Colworth, Colworth House, Sharnbrook, Bedford, MK44 lLQ, UK 2. SIK, The Swedish Institute for Food and Biotechnology, Box 5401,402 29 Gothenburg, Sweden. 3. Department of Chemistry, University of York, Heslington, York, YO1 5DD, UK.

1. INTRODUCTION Customers are becoming increasingly demanding and fastidious and the drive to reduce manufacturing costs of additives has led to demands for a more effective use. Biopolymers (proteins and polysaccharides) are commonly used in food products for the purpose of obtaining the desired stability, performance and consistency. To be able to control the behaviour and properties of biopolymer mixtures in multicomponent systems, we need to understand their interactions in both the solution and gelled state. Since most industrial processes involve temperature changes, it is important to have an understanding of how these interactionsare affected as the temperature is varieb. For many mixed biopolymer systems a reduction in temperature will favour phase separation, by reducing the contribution of the entropy of mixing to overall thermodynamic stability. In many cases, cooling may also cause one or more of the biopolymers in the mixture to order, aggregate and form a network. The onset of ordering may initiate phase separation as the ordering and aggregation process change the thermodynamic properties of the constituents and thus drive the demixing process'. For mixed biopolymer systems properties such as e.g. salt content, charge and molecular weight influence both the phase behaviour and the gelation process of the polymers24.Invariably gelation will slow down the phase separation process and structurally trap the system. In order to be able to control the microstructural and rheological properties of such mixed systems it is important to understand the interplay between phase separation and network formation. Earlier studies on the phase behaviour of mixed biopolymer systems have often focused on the incompatibility of the polymers at high temperatures, well above the ordering and gelation temperature of the components'. *' '. However, to be able to control the microstructure and rheology of mixed systems, it is necessary to have an understanding of how temperature, conformational ordering and gelation affect the phase behaviour of the biopolymers. In this paper work performed, in EU FAIR project "Mixed Biopolymer- Application and Mechanism of Phase Separation", on mixed biopolymer systems will be presented. The results from the project will be exemplified by discussion of the phase behaviour of systems containing maltodextrin mixed with either lime hide gelatin or 1-carrageenan.

168

Gums and Stabilisersfor the Food Industry 10

Using several different techniques, such as DSC, light scattering, polarimetry, microscopy and rheology, the effect of salt, temperature and polymer composition on the phase behaviour of these systems have been investigated.

2. MATERIALS AND METHODS Biopolymer mixtures were made using maltodextrin, SA2, (Avebe), gelatin (lime hide, Bloom 240, SKW) or t-carrageenan (sodium form, SKW). The gelatin was dispersed in cold de-ionised water and then dissolved at 60 "C for 30 min with stirring. The maltodextrin was dispersed in cold de-ionised water and then dissolved at 95 "C for 30 min with stirring. For mixtures of gelatin: SA2, the polymers were dissolved separately and, subsequently, mixed at 60 "C. For mixtures of 1-carrageenan: SA2, the two polymers were dry mixed and dispersed in cold de-ionised water and then dissolved at 95 "C for 30 min during stirring. Micrographs of the mixtures were acquired using a Leica TCS 4D confocal laser scanning microscope, CLSM, (Heidelberg, Germany), equipped with a Linkam TMS 92 heating and cooling table. For the microscopy work the SA2 was covalently labelled with rhodamine B isothiocyanate as described by de Belder and Granath9. The emission maximum of an argon-krypton laser was used as a light source. The transmittance of the mixtures was recorded using a Shimadzu 21 lOPC (Shimadzu Scientific Instrument Inc., USA) spectrophotometer in the wavelength region 400 to SOOnm, using a spectral resolution of 2nm. Using lcm quartz couvettes, the turbidity was determined as the temperature was varied linearly at 1 "C/min or quenched at rates between 20 to 30 "C/min. The gelatin: SA2 mixtures were cooled from 60 "C to different end temperatures. The t-carrageenan: SA2 mixtures were cooled from 85 "C to different end temperatures. The rheological measurements were performed using a Physica UDS 200 (Paar Physica, Stuttgart, Germany). A cup and bob geometry was used. A layer of mineral oil was applied on the surface of the sample in order to prevent evaporation. The frequency was 1Hz and the strain 0.5% during the measurements. Samples were transferred to a preheated rheometer. The viscoelastic properties were recorded as the temperature was quenched to different end temperatures and, subsequently, for 16h at the final temperature. The quench rate used was -20 "C/min. Optical rotation measurements were performed using a polarimeter based on the Faraday modulation technique. Light from an argon ion laser at 488 nm was used. The instrument monitors both the rotation in the plane of polarisation and the transmittance of the laser beam through the sample; this allows simultaneous monitoring of both optical rotation and turbidity changes in the sample. The sample is contained in a jacketed cell with a 1 cm or 0.5 cm pathlength, and is thermostatted using a water bath.

Mixed Biopolymer Systems

169

3. RESULT AND DISCUSSION

3.1. Factors affecting the phase diagram 3. I. I . Temperature. The phase behaviour of mixed biopolymer systems is very much dependent on the solution conditions, i.e. temperature, pH and salt conditions4.lo* ' I . In order to filly utilise the stabilising functionality of mixed biopolymers, it is important to understand the phase separation behaviour under different physicochemical conditions. Gelatin: SA2 mixtures are known to phase separate via segregation', with bulk phase separation (separation into two layers) being achieved after 80 min when the samples are held at elevated temperatured2. However, the temperature dependence of the phase diagram was not well understood, with the assumption often being made that with fast quenches, the thermodynamic phase equilibrium govern the system hot would hold true in the cold gel state. The phase diagram for the gelatin: SA2 system was determined using turbidity measurements, in the wavelength region of 400 to 8OOnm. A strong increase in turbidity was taken as evidence for the onset of segregative phase separation. In order to study the influence of temperature on the compatibility of the mixed systems, the turbidity of mixed samples was recorded on cooling at 1 "C/min. In figure 1 the phase diagram for gelatin: SA2 in 0.1M NaCl at different temperatures is presented, as a result of cooling at 1 "C/min. It can be observed that the binodal is shifted toward the axes as the temperature is decreased, that is the biopolymers become more incompatible as the temperature is decreased. It should be noted that measuring the temperature of phase separation, T, during cooling of the sample at 1 OC/min, which is slow when compared to industrial cooling processes, will lead to a underestimate of the temperature compared to a slower cooling rate. In figure 2, T, is plotted as a function of SA2 concentration for gelatin concentrations, in 0.1M NaCl or deionised water. These results have also been verified using CLSM". At temperatures below 30 OC a pronounced "kink" in the temperature dependence is observed, which shows that the movement of the binodal speeds up at temperatures below 30 "C. 3. 1. 2 Zonic content. In figure 2, it can also be observed that the miscibility of the biopolymers is altered, as the ionic content of the mixture is changed. The compatibility of the charged gelatin and the neutral maltodextrin is increased as the salt concentration is decreased from 0.1M NaCl to deionised water, i.e. at a constant gelatin concentration higher SA2 concentrationsare required to induce phase separation. It has been argued that this is due to the entropic effect of dissociatingcounter ionsI4. 3. 1. 3. Molecular ordering. Results from optical rotation measurements are presented in figure 3. These show that the ordering process of the lime hide gelatin, used in this study, in a mixture with SA2 is slow at 35 "C and 38 "C. DSC experiments performed at 1 "C/min support these results and indicated that the onset of ordering is at 30 "C. This ordering temperature coincides with the "kink" in the temperature dependence for phase separation. Since, the shift of the binodal becomes more pronounced, at temperatures below the onset of ordering, the driving force for phase separation appears to be different. Light scattering experiments on pure gelatin, pure S A 2 and mixtures confirm that, at a temperature above the ordering temperature of gelatin no molecular weight changes occur and that the rate of molecular ordering and crystallisation of the maltodextrin is very

Gums and Stabilisers for the Food Industry 10

170

slow". Hence, the explanation for the decreased compatibility of the biopolymers, as the temperature is decreased from 60 "C to 30 "C, is a change in the cross polymer interaction parameter, which is consistent with Flory-Huggins modelling of the phase behaviourI6.The change in the cross interaction parameter is a consequence of "stiffening"of the chains, as the energy of the system is reduced, which will lead to reduction in the degrees of freedom for the chains. Similar effects of temperature on the incompatibility of gelatin: starch mixtures have been reported". In addition to the expected decrease in miscibility due to the change in interaction parameter, demixing has previously been shown to be enhanced by conformational ordering and molecular weight changes6. At low temperatures, below the transition temperature of gelatin, the molecular weight is continuously increasing, with time, which we would expect to increase the tendencey for M e r demixing. During this process the demixing should be driven to its equilibrium end-points with time. However, the increased elastic character of the gelatin will oppose this process and kinetically trap both the microstructure and polymer composition of the phases. The extent of kinetic trapping is therefore dependent on quench rate and depth.

l4-

measured during cooling 1"C/min

5z m"1-

-doud point curve at 60°C _ _ _ _ doud point curve at 5OoC doud point curve at 4OoC

Figure 1 Gelatin: SA2 cloudpoint curve at different temperatures, in 0.IM NaCI. 3.2. Microstructure creation The understanding of how the binodal moves as a function of temperature has generated the opportunity to design experiments where it is possible to start with miscible mixtures and quench to temperatures, above and/or below the ordering temperature of the gelatin. The microstructure, turbidity and viscoelastic properties of such samples were investigated. In this paper results from 4.5% gelatin: 2.25% SA2, in 0.1MNaCl, and 4% gelatin: 4% SA2, in deionised water are presented.

171

Mixed Biopolymer Systems

1

0

2

4

a

6

10

SA2 Concentration,% WEW Figure 2 Phase separation temperature as afunction of SA2 concentrationfor 4% gelatin in deionised water and 4.5% gelatin in 0.IM NaCl, when cooled at I Wmin.

35'C

-0.7

I 0

I

I

I

I

20

40

60

80

I

time, min Figure 3 Optical rotationfollowing quenching of 4.5% gelatin 2.25% SA2 in 0.1M NaCl @om 60 to different temperatures,

172

Gums and Stabilisers for the Food Industry 10

3.3. Phase separation initiated above the ordering temperature The phase separation temperature for the mixture of 4.5% gelatin: 2.25% SA2 was measured to be 35 "C at a cooling rate of 1 "C/min, see figure 2. In order to test the kinetics of phase separation in the absence of gelatin ordering, experiments were performed, where the mixture was rapidly quenched from 60 "C to 32 "C. The phase separation process was then followed with time. In order to trap the microstructure the mixture was quenched a second time to 10 "C. The resultant microstructures were compared to a system which had been quenched directly to 10 "C. In figure 4 the turbidity as a function of time for the 4.5% gelatin: 2.25% SA2 afler the quench to 32 "C is presented. Due to phase separation, which is initiated immediately at 32 "C, the turbidity increases rapidly. As the phase separation progresses, the SA2 droplets start to coalesce and sediment, resulting in a decrease in turbidity. With time this would ultimately result in two separate biopolymer layers. The polymer compositions of these two phases are described by the equilibrium composition on the binodal. Interestingly, if the two phases were allowed to bulk phase separate, over a time period of 24h at 32 "C, the resulting low turbidity of the top gelatin phase was strongly increased, as the temperature was quenched from 32 "C to 10 "C, see figure 4. This indicates that the gelatin rich phase itself undergoes further phase separation during this cooling step. This work was verified by visualising the microstructure. Figures 5 a and b show CLSM micrographs of the 4.5% gelatin: 2.25% SA2 at 32 "C and after a double quench to 10 "C. In figure 5a a snap shot of the S A 2 droplets at 32 OC is presented. At this temperature, the droplets formed are about 3-4pm and they start to sediment as soon as they are formed, as the viscosity of the continuous gelatin phase is not sufficiently high to prevent sedimentation. If the second quench to 10 "C occurs before the SA2 droplets (created at 32 "C) have had a chance to fully sediment, then a second population of smaller SA2 droplets are produced, see figure 5b. This second population is caused by molecular ordering during the quench from 32 "C to 10 "C. Initial results from modelling of the turbidity data recorded at 32 "C, indicate that the droplet growth is diffusion limited". Microscopy performed on a bulk phase separated SA2 rich phase has shown that a similar event occurs in the SA2 phase where gelatin phase separates from the SA2. However, due to the limiting resolution of CLSM, it is not possible to observe the gelatin inclusions in the 3-4pn big S A 2 droplets. When the mixture is quenched from 60 "C to 10 "C, the SA2 droplets formed during the quench show a more homogeneous droplet size (compare figure 5 with figure 6). This is expected since quenching straight to 10 "C does not give the droplets time to grow (as in figure Sb), from figure 4, it can be seen that the second wave of phase separation occurs very soon. This suggest that' both separation events are close together, in time and temperature during a fast quench. Droplets produced in each event, therefore, experience a similar environment in terms of growth limitation, i.e. gelation, diffusion etc. This mixture did not show any sign of sedimentation of SA2, i.e. no structural inhomogeneity in z-led could be observed in the CLSM.

173

Mixed Biopolymer Systems

0

5

10

15

20

25

30

Time. min

Figure 4 Turbidity at 800nm for 4.5% gelatin: 2.25% SA2 afier quench to 32 "C and the phase separated gelatin phase afier a quenchfrom 32 "Cto I0 "C. The results from the double quench indeed show that the phase separation behaviour of a mixed biopolymer system is more complicated than expected from the equilibrium phase diagram obtained above the ordering temperatures of the biopolymers. The formation of a second population of droplet sizes, as an immiscible system is cooled down below the ordering temperature of one of the components, indicates that the included droplets are formed under conditions where the gelation kinetics of the gelatin are different in the two steps. 3.4. Phase separation initiated below the ordering temperature In order to investigate the effect gelation kinetics on the phase separation behaviour of the 4% gelatin: 4% SA2 system, in water, the turbidity and the viscoelastic properties were recorded during quench cooling from 60 "C to 20 "C or 10 "C. Figure 7 compares the

turbidity and viscoelastic response as a h c t i o n of time after quench cooling. For the mixture cooled to 10 "C, it can be observed that the turbidity started to increase as the temperature approached 10 "C, however, there is a short time lag (4 minutes) before there is a increase in the viscoelasticity of the system. Similar results were observed for the mixture quenched to 20 "C. At 20 "C the increase in turbidity was observed after 4 minutes. In this case the delay, between the increase in turbidity and viscoelasticity, was about 5 minutes.

174

Gums and Siabilisers for ihe Food Industry I0

Figure 5 CLSMmicrographsof 4.5% gelatin mixed with 2.25% SA2 in O.IMNaC1a) at 32 "Cand b) at 10 "C.

Figure 6 CLSM micrographs of 4.5% gelatin mixed with 2.25% SA2 in 0.1M NaCl afrer a quench to 10 "C.

CLSM micrographs of 4% gelatin: 4% SA2 imaged at the quench temperatures of 10 "C and 20 "C are presented in figure 8. The micrographs show that the brighter stained SA2 droplets are smaller after a quench to 10 "C than the droplets formed after a quench to 20 "C. In addition, there is a delay in time for the system to phase separate at 20 "C suggesting that the kinetics of phase separation are quicker than the kinetics of network formation.

Mixed Biopolymer Systems

175

1

'E 0

m

0.1

Figure 7 Turbidity at 800nm and G' as a function of time for 4% gelatin: 4% SA2 in deionised water after quenches *om 60 "C to 10 "C (solid and open circles) or 20 "C (solid and open triangles). The system reached the desired temperatures after approximately 2minutes.

Figure 8 CLSM micrograph showing the resulting microstructure of 4% gelatin mixed with 4% SA2 after quench to 10 "C or 20 "C. Comparison of the results from the two quenches shows that, as the gelation process is slowed down at the higher temperature, the phase separated droplets are allowed to grow to bigger sizes before the elasticity of the network formed and microstructrally traps the system. Figure 2 suggests that the Tw is approximately 20 "C, when cooled at 1 "C/min.

116

Gums and Stabilisers for the Food Industry 10

The quench to 10 "C would, therefore, be expected to phase separated. Since the quench to 20 "C does not phase separate immediately it suggests that there seems to be a critical amount of ordering required to induce phase separation. The delay in time before phase separation also suggests that, at 20 "C, the change in the amount of order, with time, will change both the molecular weight and the interaction parameters. Both of these effects are believed to facilitate phase separation and both seem, as a result of this experiment, to be inextricable linked to the degree of gelation.

3.5. Model A conceptual model illustrating our understanding of the temperature dependence of the phase behaviour for a binary polymer mixture is presented in figure 9. The model accounts for the fact that the immiscible region for a mixed biopolymer system increases with decreased temperature, i.e. the binodal moves closer towards the diagram axes. It is possible to define two regions of phase separation; i.e. demixing occurring at temperatures above 30 "C with gelatin in a disordered state, and phase separation starting at temperature below 30 "C where gelatin is undergoing a disorder-to-order transformation. We believe that the driving force for phase separation has increased between these two regions. At temperatures above the gelatin ordering temperature increasing incompatibility is thought to be due to the temperature dependence of the Flory-Huggins polymer-polymer interaction parameter for the components, a parameter which suffers a fkrther increase on conformational ordering below 30 "C. Based on theoretical modelling of polymer systems, the concept of inducing phase separation by conformational changes a n d or polymerpolymer interactions has been suggested earlier"-*'. The differences between the two regions can be exemplified with the two quench routes described in figure 9. Route 1; where a miscible system is cooled down to a temperature below the ordering temperature, phase separation is induced by the amount of ordering and increase in molecular weight. The time to achieve such ordering, which will depend on quench depth, is the reason for the delay in phase separation. During such a delay the molecular weight of the gelatin and the cross-interaction parameter will also be changing. For systems quenched via this route, the morphology of the mixture can be controlled by varying the gelation kinetics. Alternatively, phase separation can be induced above the ordering temperature and subsequently cooled down further to induce gelation (route 2 in figure 9). This route will lead to a more complex morphology depending 1) on the extent to which the two phases are allowed to ripen, at the higher temperature, 2) the increased incompatibility induced by ordering will lead to a second population of droplets, and 3) the rate of gelation will determine at what stage the structure will be trapped. Therefore, the difference in T,,, and Tgelation provides an operating window in which different microstructures can be formed. It should be pointed out that our findings indicate that the ordering process moves the binodal very close to the axes of the phase diagram, i.e. even at very low polymer compositions the ordering of gelatin will, with time, lead to phase separation.

Mixed Biopolymer Systems

1I1

Phase Separatic Temp.

Orderinj temp.

.-.

b

Polymer A conc Figure 9 Schematic model illustrating the temperture of phase separation as a function of the concentration of polymer A, at a constant polymer B concentration. 3.6. Generality of model

In order to test the generality of the model and the structural affects of using a more rapid gelling biopolymer, we have also studied mixtures of 1-carrageenan and SA2 in 0.25M NaCl. In figure 10 the temperature of phase separation, as measured by turbidity and the temperature of onset of ordering, as measured by DSC, are plotted as a function of SA2 concentration for 0.5% 1-carrageenan and 1.5% t-carrageenan. It can be observed that this mixture of ~arrageenan:SA2shows a similar behaviour to the gelatin: SA2 system. The "kink" in the temperature dependence of T, occurs at temperatures below the onset of ordering for the t-carrageenan, see mixture with 1.5% 1-carrageenan in figure 10. Both results from turbidity and CLSM experiments show that the mixtures after quenching from 85 "C to 60 "C immediately phase separated at 60 "C (results not shown). This indicates that, for the i-carrageenan: SA2 system, it does not seem possible to prevent the phase separation process by gelation of 1-carrageenan, as might have been expected by the faster gelation kinetics, when compared to gelatin. It is possible to conclude that, under the studied conditions, the 1-carrageenan: SA2 system phase separates at temperatures above the ordering temperature, but as the 1-carrageenan starts to order the incompatibility of the biopolymers increases strongly, in a similar way to gelatin.

178

Gums and Stabilisers for the Food Industry 10

F

/'

.i

I l

I

0

.

l

2

.

l

4

.

l

6

.

l

.

l

.

l

.

l

.

I

8

1

0

1

2

1

4

1

8

,

SA2 Concentration, w/w?/o

Figure 10 Temperature ofphase separation and onset of carrageenan ordering for 0.5% Icarrageenan and 1.5% t-carrageenan mixed with SA2 in 0.25M NaCI. Cooling rate I "Umin.

3.7. Designed microstructure and rheology?

In this paper we have addressed the problem of temperature dependence on the phase behaviour of mixed biopolymer systems. At temperatures where the gelatin or 1carrageenan is in a disordered state the phase separation is driven by changes in the solvent quality for the different biopolymers. If such a mixture is kept isothermally at high temperatures the biopolymers will in most cases ultimately separate into two separate layers, which have compositions described by the end of the tie-line points on the binodal. For mixtures quenched to temperatures below the ordering temperature, it was found that phase separation may be induced by the molecular weight changes associated with ordering and aggregation of the biopolymers. Inevitably, phase separation will lead to increased effective concentrations of the components in the different phases, which will influence the ordering and gelation processes. Such coupled effects make it difficult to obtain tie-lines and phase composition, i.e. establishing a phase diagram with the components in a gelled state. The results presented in this paper also show that an incompatible system, which has phase separated to is equilibrium compositions, at a certain temperature above the ordering temperature of the components, might phase separate a second time as it is cooled down further. Studying the interplay between phase separation and network formation, we found that for neither of the studied systems did the gelation process hinder phase separation. However, by controlling the rate of gelation by quenching the mixtures to different endtemperatures, it was possible to control the morphology of the mixed system. That is, by

Mixed Biopolymer Systems

179

quenching a miscible system to a lower end-temperature smaller droplet sizes are obtained compared to a quench to a higher temperature. A phenomenological understanding of how the binodal moves with temperature and changes in molecular configuration are not sufficient to be able to predict the resulting functional properties of the mixed system. Therefore, ongoing work is focused on quantification of both the biopolymer composition of the phases after quenches to low temperatures as well as image analysis of the phase volumes and connectivity of the phases, to describe the measured rheological properties.

Acknowledgement

This study has been carried out with the financial support from the Commission of the European Communities, Agriculture and Fisheries (FAIR) specific RTD programme, CT 96 1015, “Mixed Biopolymers- Mechanism and Application of Phase separation“.It does

not necessarily reflect its views and in no way anticipates the Commission’s future policy in the area. The authors would like to thank A. Clark, P. Aymard, N. Lorkn, T. K. Halstead, D. M. Goodall, M. A. Lbpez-Quintela, A. Cesliro, P. Boulenguer, C. Michon and B. Launay for helpful discussions. V. Normand, A. A l t s h , D. Fabri, M. C. Blanco, D. Leissner, C. Vbquez, F. Sussich, F. Cuppo for experimental input. References 1. A. H. Clark, ‘Bipolymer Mixtures’, Eds. S. E. Harding, S. E. Hill and J. R. Mitchell, Nottingham University Press, UK, 1995, p. 37-64. 2. V. B. Tolstoguzov, ‘Functional Properties of Food Macromolecules’, Eds. J. R. Mitchell and D. A. Ledward, Elsevier Applied Science Publisher, 1992, p. 385-445. 3. P-A. Albertsson, ‘Bipolymer Mixtures’, Eds. S. E. Harding, S. E. Hill and J. R. Mitchell, Nottingham University Press, UK, 1995, p. 1- 12. 4. L. Piculell, K. Bergfeldt and S. Nilsson. ‘Bipolymer Mixtures’, Eds. S. E. Harding, S. E. Hill and J. R. Mitchell, Nottingham University Press, UK, 1995, p. 13-35. 5. V. Y. Grindberg and V. B. Tolstoguzov, Food Hydrocoll. ,1997,11,2, 145-148. 6. M. G. Semenova and L. B. Savilova, Food Hydrocoll, 1998,12,65-75. 7. S . Kaspais, E. D. Moms I. T. Norton and M. J. Gidley, Curbohydr. Polym. , 1993,21, 249-259. 8. T. J. Foster, C. R. T. Brown and I. T. Norton, ‘Gums and Stabilisers for the Food Industry 8’, Eds. G. 0. Philipps, P. A. Williams and D.J. Wedlock, IRL Press, 1996, p 297-306. 9. A.N. De Belder and K. Granath, Curbohydr.Rex, 1973,30,375 10. V. B. Tolstoguzov, Food Hydrocoll., 1991,4,429. 11. L. I. Khomutov, N. A. Lashek, N. M. Ptitchkina and E. R. Moms, Carbohydr. Polym. , 1995,28,341 12. T. J. Foster, J. Underdown, C. R. T. Brown, D. Ferdinand0 and I. T. Norton, ‘Food

Colloids- Proteins, Lipids and Polysaccharides’, Eds. E. Dickinson and B. BergensW, The Royal Sociaty of Chemistry, 1997, p. 346. 13. N. Lor& and A-M. Hennansson. 1999, Submitted to Int. . I Bio. Macromol.

180

Gums and Stabilisersfor the Food Industry 10

14. K. Bergfeldt and L. Piculell, J. Phys. Chem., 1996,100,5935-5940. 15. M. C. Blanco, D. Leissner, C. Vhzquez and M. A. Lbpez-Quintela. Paper in preparation. 16. A. H. Clark, 1999, personal communication. 17. M. A. K. Williams, D. Fabri, C. Hubbard, D. Leisner, A. Altsktk and N. Lor&&1999 Paper in preparation. 18. H. M. J. Boots, J. G. Kloosterboer, C. Serbutoviez and F. J. Touwslager, Macromolecules, 1996,29,7683-7689. 19. C. Serbutoviez, J. G. Kloosterboer, H. M. J. Boots and F. J. Touwslager, Macromolecules, 1996,29,7690-7698. 20. G. E. Eliqabe, H. A. Larrondo and R. J. J. Williams, Macromolecules, 1998,31,81738 182.

EFFECT OF TEMPERATURE ON THE RHEOLOGICAL PROPERTIES OF STARCWCARRAGEENANMIXTURES

C. Loisel', A. Tecante2,P. Cantoni' and J.L. Doublier3 1- ENITIAA rue de la Geraudiere, BP 82225,44322 Nantes France 2- Depto. Alimentos y Biotecnologia, FQ-cc E n, UNAM, Mexico, D.F., 04510, Mexico 3- INRA, BP 7 1627,443 16 Nantes, France

ABSTRACT The aim of this study was to describe the effect of pasting procedures on the rheological properties of the starcMcarrageenan system in relation to the starch granule size, measured by laser scattering. The rheological behaviour of 3% crosslinked waxy corn starcMO.5 % kappa-carrageenan mixtures with or without KCl was studied using steady (at 60°C) and oscillatory (at 60" and 25°C) shear tests. The pasting temperatures ranged from 96°C (undercooked starch granules) to 125°C (overcooked starch granules) to obtain different degrees of starch granule swelling. All dispersions exhibited a shear-thinning behaviour. For the starch pastes, the apparent viscosity, the storage modulus and the median diameter of the swollen granules increased as the pasting temperature increased and the apparent viscosity could be related to the size of the swollen granules. In all cases it was confirmed that starch granules remained intact after thermal treatment. The viscoelastic behaviour of the starch pastes was also governed by the swollen granules leading to a solid-like behaviour at 25" or 60°C. Addition of carrageenan increased significantly the apparent viscosity compared to starch alone, while the mechanical spectra of starchkarrageenan mixtures indicated a more fluid-like system. It is noteworthy that the median diameter of the starch granules increased in the presence of carrageenan. The effect seems to be more complex than expected and would need a more thorough characterisation of the continuous phase. 1 INTRODUCTION Starchhydrocolloid mixtures are widely used in order to provide the desirable texture to prepared foodstuffs such as custards or sauces. In the present work we used kappacarrageenan and crosslinked waxy corn starch which is free of amylose and able to resist mechanical and thermal treatment due to chemical modification. Under appropriate pasting conditions starch granules swell but remain intact, while the hydrocolloid is concentrated in the continuous phase. Rheology of the starch-carrageenan combination is then related to the volume fraction of the swollen starch granules as well as the concentration of the added polysaccharide within the continuous phase. The aim of this work was to study the influence of thermal treatments on the rheological behaviour of this biphasic system under gelling (using KCl) and non gelling

182

Gums and Stabilisers for the Food Industry 10

conditions of kappa-carrageenan. Rheological measurements were combined with techniques providing particle size and size distribution of the starch granules. 2 MATERIALS AND METHODS

2.1 Materials Crosslinked corn starch (Clearam CHlO) was an adipate/acetate starch supplied by Roquette Freres (France). This starch swelled to its maximum at around 115"C, under the shear and acid conditions of the study. Kappa-carrageenan (SKW Biosystems, France) was prepared (0.5 %) either in distilled water or in KCl solution (0.02 M). This KCl concentration was chosen in order to obtain a gelled system at 25°C.

2.2 Pasting Procedure Kappa-carrageenan was first dispersed in water with or without KCl at room temperature, under mechanical stimng ; then starch was poured to reach a final concentration of 3 % under stirring to avoid lumps. Thermal treatment was then applied using a jacketed vessel (capacity 21, stirring rate 500 rpm) with gradual heating (1.5 "C/min) from 20°C to the pasting temperature (96", 112" and 125°C) ; this temperature was maintained for 10 minutes and was followed by a cooling step (1.5 "C/min) down to 70°C before sampling. The pasting temperatures were chosen to reach three swelling degrees of the starch granules : undercooked (96"C), well cooked (1 12°C) and overcooked (125°C). Preparation concentration of 3 % (w/w) was controlled by drying aliquots overnight at 102°C. Starch preparations are referred to as A, B and C for starch alone, starcldcarrageenan, starch/carrageenan/KCl, respectively. 2.3 Particle sue and size distribution Particle size determination was run at room temperature using a Malvern Master Sizer (Malvern Instruments, Ltd) laser scattering analyser with a 300 mm Fourier cell (range 0.05 to 879pm). The starch dispersion was first diluted (1/10) with water at 20°C immediately after the pasting procedure, then dispersed in the sample dispersion unit (lm1/100 ml water) and fed into the measuring cell. Volume distribution was obtained using the Mie scattering theory which requires refractive index of the media to be specified : we used 1.529 and 1.33 respectively for starch and liquid phase and 0.1 for the starch granule absorption. From each distribution a median volume diameter D(v, 0.5) was calculated.

2.4 Rheological measurements Steady shear tests were carried out at 60°C using a coaxial cylinders viscometer (Rheomat 120 ; diameters : bob= 46 mm, cup = 49 mm, height 78 mm). Two consecutive shear scans from 0 to 660 s-' and back to zero were applied for 4 minute periods each. A controlled stress rheometer (TA Instruments AR 1000) was used to perform oscillatory shear tests (4 % strain amplitude) with a cone/plate geometry (6 cd4"). The strain amplitude was chosen after determination of the linear viscoelastic range. Each sample was analysed through a three steps protocol : (1) mechanical spectrum (G' and G" as a

Mixed Biopolymer Systems

183

function of frequency) at 6OoC to characterize the hot paste, (2) gelling kinetics (at 1 Hz) after rapid cooling of the sample from 60 to 25°C using the Peltier system, (3) mechanical spectrum at 25OC of the gelled system. 3 RESULTS AND DISCUSSION 3.1 Flow curves Figure 1 shows the effect of the pasting temperature on the flow behaviour of the starch pastes. A shear-thinning behaviour is clearly exhibited with an anti-clockwise loop ; this disappeared with the second up-down scan. As expected, the pasting temperature had a strong effect on the apparent viscosity of starch pastes from the undercooked (96°C) to the well cooked state (1 12°C) ; in the meantime the shear- thinning behaviour and hysteresis were more pronounced. Figure 2 shows the effect of the pasting temperature on starch/carrageenanmixtures. We observed the same shear-thinning behaviour as for starch. However, quite surprisingly, the pasting temperature had the reverse effect with a decrease of the apparent viscosity and of the area of the hysteresis loop as the temperature increased. This effect at elevated temperature may be related to a thermal degradation of the carrageenan.

0

100

200

300

400

SO0

SO0

700

Shear rate (s-1)

Figure 1 Flow curves of 3% starch pastes (A). Effect of the pasting temperature (measurement temperature :6OoC, first scan) 120 100

2o 0

I

r 0

100

200

300

400

500

600

700

Shear rate (s-1)

Figure 2 Flow curves of starcwcarrageenan dispersions (B). Effect of the pasting temperature (measurement temperature :6OoC, first scan)

184

Gums and Stabilisers for the Food Industry 10

In Figure 3, the flow behaviour of starchkarrageenan mixtures for a pasting temperature of 112°C is compared to starch. Addition of carrageenan increased both the apparent viscosity and the hysteresis of the mixture. It is also noteworthy that the presence of KC1 strongly affected the flow behaviour of the starcldcarrageenan system. Hysteresis has been reported for modified waxy corn starch and described as an antithixotropic behaviour',2 and an indication of a flow induced structure3. 90

70

0

100

200

300

400

500

600

700

Shear rate (5-1)

Figure 3 Flow curves of 3% starch (A), starcwcarrageenan (B), and starcWcarrageenadKC1 (C) dispersions (pasting temperature I 1 2°C measurement temperature :60"C,Jirst scan)

3.2 Viscoelastic behaviour 3.2.1 At 60°C. Figure 4 shows the effect of the pasting temperature on 3% starch dispersions. The starch pastes although exhibiting a liquid aspect, had a solid-like behaviour with G'>G' 'and G' almost independent of frequency in the range investigated. After controlling the structure of starch granules through microscopic observation and solubility measurement4, we concluded that it was preserved. The viscoelastic behaviour is then governed by the swollen granules5. For starch alone an increase of the moduli with temperature was noticed. The storage modulus was maximum at 112°C. It was also more frequency dependent at 112 or 125°C than at 96"C, that indicates a more fluid-like behaviour. In Figure 5 the effect of the pasting temperature on StarcWcarrageenan mixtures is shown. We observed a reverse effect compared to starch alone : a decrease of the moduli and G' less frequency dependent with elevated temperature. As already noticed for the flow curves, increasing the pasting temperature changed the rheological behaviour in a different way, likely in relation to the composition of the continuous phase. Figure 6 shows the mechanical spectra of starch and starcWcarrageendC1 dispersions for a pasting temperature of 112°C. Adding carrageenan led to a decrease of both G' and G" which became more frequency dependent. These results suggest that the behaviour is more liquid-like. This is easily explained since carrageenan is located in the continuous phase, which therefore contributes to the overall rheology of the system.

Mixed Biopolymer Systems

185

100

-e rn

b

10

b

1

angular frequency (radls)

Figure 4 Mechanical spectra of 3% starch pastes (A). Effect of the pasting temperature. (measurement temperature :60°C ;G’ :solid ;G :hollow) ”

im

h

a e -b

10

b

1

Angular fmuencv ( W s )

Figure 5 Mechanical spectra of starchkarrageenan dispersions (B). Effect of the pasting temperature (measurement temperature :60°C ;G ’ :solid ;G :hollow) ”

100

-

b b

10

1

0,Ol

0,1

1

I0

100

1000

Angular frequency (radh)

Figure 6 Mechanical spectra of 3% starch pastes (A) and starcWcarrageenadKC1 (C) dispersions (pasting temperature 112°C ;measurement temperature : 60°C ; G’ :solid ; G ” :hollow)

186

Gums and Stabilisersfor the Food Industry 10

3.2.I A t 25°C. Figure 7 shows the mechanical spectra of 3% starch dispersions treated at 112" and 125°C. As previously stated at 60°C (Figure 4) a gel-like behaviour is clearly seen, with a lower storage modulus at 125°C than at 112°C. Figure 8 shows the mechanical spectra of the starch dispersion and starcWcarrageenan mixtures for a pasting temperature of 112°C. As expected a true gel was obtained from the starcWcarrageenanKC1 mixture : G' was about two decades higher than G" and both moduli were independent of frequency. Spectra of A and B were rather the same as those obtained at 60°C ;the storage moduli increase was merely an effect of temperature as these dispersions are non-gelling materials. lo00

0

&

b ~~

l+ 0,Ol

,

,

,

~~

~~~

~

~~

, ,

0-1

I

10

100

lo00

Angular frequency (radls)

Figure 7 Mechanical spectra of 3% starch pastes (A). Effect of the pasting temperature (measurement temperature :25°C ;G ' :solid ;G :hollow) "

i J 0,Ol

091

1

10

100

lo00

Angular frequency (rads)

Figure 8 Mechanical spectra of 3% starch pastes (A), and starchkarrageenan mixtures (B and C). (pasting temperaturell2"C ;measurement temperature : 25°C ;G ' :solid ;G" : hollow)

Table 1 shows the median volume diameters of starch granules obtained from different mixtures.

Mixed Biopolymer Systems

187

Table 1 Median volume diametersfor diflerent mixtures andpasting temperatures

Regarding the effect of the pasting temperature on starch alone, an increase was observed from 41.6 to 47.4 pm, to which the increase of the apparent viscosity and the storage modulus can be linked (Figure 1 and 4). This relationship is presumably observed because the starch granules are not disintegrated and the starch solubility is very low (< 2%). Volume fraction of the swollen granules is very high. A value of 0.75 was found for a 3 % starch paste from the same origin' heated at 96"C, which is not far from closepacking6s7. This explains the solid-like behaviour of the starch paste. For starchkarrageenan mixtures, the higher swelling of starch granules compared to starch alone was balanced by the presence of a concentrated continuous phase which led to viscoelastic behaviour which was more liquid-like. 4 CONCLUSIONS

The rheological behaviour of starch/carrageenan systems is governed both by the volume fraction of the dispersed phase and the viscosity of the continuous phase ; granule swelling yields a reduction of the volume of the continuous phase and hence should increase the viscosity. It is also noteworthy that the presence of carrageenan increased the median diameter, whatever the temperature. Other studies on starch/galactomannan mixtures*or starch/xanthan9did not mention this effect.

References 1 . A. Tecante and J. L. Doublier, Carbohydr. Polym., 1999, in press 2. M. A. Rao, P. E. Okechukwu, P.M.S. Da Silva and J.C. Oliveira, Carbohydr. Polym., 1997,33,273-283. 3 . F. R. Dintzis, E. B. Bagley and F. C. Felker, J Rheof., 1995,39, 1399-1409. 4. J. L. Doublier, G. Llamas and M. Le Mew, Carbohydr. Polym., 1987,7,251-275. 5 . M. Alloncle and J. L. Doublier, Food Hydrocolloids, 1991,5455-467. 6. E. B. Bagley and D. D. Christianson,J Texture Stud., 13, 115-126. 7 . P. A. M. Steeneken, Carbohydr.Polym., 1989,11,23-42. 8. M Alloncle, J. LeRbvre, G. Llamas and J.L. Doublier, Cereal Chem.,1991,66,455-467. 9. N. A. Abdulmola, M. W. N. Hember, R. K. Richardson and E. R. Moms, Carbohydr. Polymer., 31,65-78.

INTERACTIONS BETWEEN K-CARRAGEENAN AND P-LACTOGLOBULIN IN GELLING AND NON-GELLING AQUEOUS SYSTEMS

N. E. Hotrum'.*.', J. A. Lucey'36and H. Singhl 'Institute of Food, Nutrition and Human Health, Massey University, Private Bag 11222, Palmerston North, New Zealand *Departmentof Food Science, University of Guelph, Guelph, Ontario, N1G 2W1, Canada 'present address: Wageningen Centre for Food Sciences, P.O. Box 557, 6700 AN Wageningen, the Netherlands "resent address: Department of Food Science, University of Wisconsin-Madison, 1605 Linden Drive, Madison, WI 53706-1565, USA

1 INTRODUCTION Polysaccharides play an important role in the stability of many dairy systems. For example, the thickening and gelling properties of added polysaccharides are used to form and stabilize milk gels such as yoghurt, prevent unwanted crystal growth in ice cream, inhibit the settling of cocoa particles in chocolate milk and prevent the creaming and flocculation of emulsion droplets in infant formula. K-Carrageenan is a sulfated polysaccharide that is commonly added to dairy products. Whde its interaction with caseins is fairly well understood,'" there is relatively little information available on whey protein-K-carrageenaninteractions." When solutions of proteins and polysaccharides are mixed, three equilibrium situations may arise, miscibility, complex coacervation or thermodynamic incompatibility?.'o Mixed protein-polysaccharide systems most commonly exhibit thermodynamic incompatibility, which is due to a net repulsion between biopolymers resulting in a two-phase system where each phase is enriched by one of the biopolymer species." Thermodynamic incompatibility usually takes place at high ionic strength and at pH values greater than the isoelectric point of the protein." Complex coacervation is generally only observed for mixtures of oppositely charged biopolymers, for example, in the pH range below the isoelectric point of a protein and above the pK of an anionic polysaccharide;'o however, sulfated polysaccharides may form protein-polysaccharide complexes above the isoelectric point of the protein via electrostatic interaction between local positive charges on the protein and negatively charged sulfate groups on the polysac~haride.~-'~ Miscibility is uncommon for mixed biopolymer solutions except in dilute systems. Early research into whey protein-carrageenan interactions reported that P-lactoglobulin forms complexes with ic-carrageenan below pH 6, with maximum complex formation occurring at pH 2.5.4 More recent work showed that both whey protein isolate6 and plactoglob~lin~ exhibit thermodynamic incompatibility when mixed with ic-carrageenan at pH 7. Heat-induced denaturation of p-lactoglobulin may expose more positively charged groups available for electrostatic interaction with the sulfate groups of K-carrageenan leading to complex formation at pH values above the isoelectric point of P-lactoglobulin. The mi+n objective of this study was to determine the nature of the interactions between K-carrageenan and P-lactoglobulin in aqueous systems under both gelling and non-gelling

Mixed Biopolymer Systems

189

conditions at pH 7 and to determine the effect of heat-induced denaturation of p-lactoglobulin on these interactions. Food grade K-carrageenan was used in order to study the behaviour of a realistic food ingredient in the presence of a protein. Purified plactoglobulinwas chosen as the protein component since P-lactoglobulin is the major whey protein. 2 MATERIALS AND METHODS 2.1 Materials Food grade K-carrageenan was donated by Woods and Woods (Auburn, Australia, PO Box 98, product # = - l a ) , no further purification was canied out. The major cations present in the K-carrageenan sample were potassium, sodium, magnesium and calcium which accounted for 5.3, 0.31, 0.26 and 0.09% (w/w) of the sample, respectively. Purified plactoglobulin (Product # L-3908) and bovine serum albumin (BSA) monomer (Product # A1900) were purchased h m Sigma Chemicals (St. Louis, MO, 63178, USA); no further purification was canied out. Sodium chloride and sodium azide were purchased fiom BDH Chemicals (Palmerston North, New Zealand, PO box 1246). Toluene @PLC grade) was obtained h m Aldrich Chemicals (Milwakee, WI,53233, USA). 2.2 Methods

2.2.1 Sample preparation. Two types of samples were prepared for this study, nongelling (0.1% w/w) and gelling (1.0% w/w). For both non-gelling and gelling systems, six samples were analyzed, these samples were either unheated or heated, K-carrageenan, plactoglobulinand 1:1 K-carrageenan:p-lactoglobulin. K-Carrageenan powder was dispersed in distilled, deionized water at mom temperature by moderate stirring for 30 min. Then, 0.5 M NaCl was added in order to give a final concentrationof 0.1 M NaCl and the pH was adjusted to 7.0. The sample was heated at 75°C for 30 min with moderate stirring in order to dissolve the polysaccharide, this combinationof heating and stirring was required in order to obtain an even dispersion."-" After dissolution, non-gelling samples were cooled to room temperature. Gelling samples were cooled to 60°C at a rate of 2"C/min. p-Lactoglobulin was dissolved in 0.1 M NaCl, pH 7.0. Gelling P-lactoglobulin solutions were warmed to 60°C with moderate stirring. K-Carrageenan:p-lactoglobulinmixtures were prepared by mixing equal masses of each biopolymer solution. For non-gelling samples, the solutions were mixed for 60 min at 20°C. Gelling samples were mixed at 60°C for 10 min. 'Vnheated" samples of K-carrageenan andlor p-lactoglobulin did not undergo further heating after dissolution. "Heated" nongelling samples were prepared by heating at 80°C for 30 min with stirring. "Heated" gelling samples were heated in the rheometer as described below. 2.2.2 Size-Exclusion Chromatography with Multi-Angle Laser Light Scattering (SECUALLS). The SEC-MALLS instrumental set-up was composed of a pump (GBC model LC 1150, Pharmacia Biotech, Uppsala, Sweden), injection valve (model 7125, Rheodyne, Alltech, Deerfield, L,60015, USA), a Superose 6 HR 10130 SEC column (Pharmacia Biotech), UV-visible detector (GBC model LC 1200, Pharmacia Biotech), multi-angle laser

190

Gums and Stabilisers for the Food Industry 10

light photometer (DAWN DSP, Wyatt Technology Corp., Santa Barbara, CA, 93 103, USA), and differential refractive index (DRI) detector (Waters R401, Millipore Corp., Bedford, MA, 01730, USA). Data were analyzed using Astra (version 4.5) software (Wyatt Technology Corp.). The DRI detector signal was used as the concentration detector for data analysis, while the UV detector was used as a qualitative indicator for the presence of protein. The SEC-MALLS was calibrated and normalized using HPLC grade toluene and 0.1% (w/w) monomeric BSA, respectively. 500 pL samples were injected into the column using a 1 mL injection loop. The flow rate was 0.4 mL/min. 2.2.3 Small Deformation Rheometry. Viscoelastic measurements were carried out in a Bohlin VOR controlled strain rheometer (Bohlin Reologi, Lund, Sweden) equipped with a Couette type cup and bob (C25) geometry and 42.6 g-cm torsion bar. Experiments were carried out at a frequency of 1 Hz and an applied strain of 0.001, which was within the linear viscoelastice range. Samples that had equilibrated to 60°C were loaded into the pre-heated (60°C) rheometer sample cup, and were either cooled fiom 60 to 20°C at a rate of O.S"C/min and then held at 20°C for 60 min ("unheated") or heated from 60 to 80°C at a rate of l"C/min, held at 80°C for 20 min, then cooled from 80 to 20°C at a rate of O.S"C/min and held at 20°C for 60 min ("heated").

3 RESULTS AND DISCUSSION 3.1 SEC-MALLS

The DRI elution profiles and weight-average molar mass (M,) distributions of 0.1% Kcarrageenan, O. 1% p-lactoglobulin and 1:1 (0.1%) tc-carrageenan$-lactoglobulin are shown in Figure 1 a, b, and c, respectively. Unheated and heated K-carrageenan eluted in the same position (Figure la) as a broad peak between 6.6 and 11.2 mL. The unheated sample had a slightly higher maximum DRI signal compared to the heated sample. The DRI detector elution profiles of unheated and heated 0.1% P-lactoglobulin solutions were significantly different (Figure Ib); the unheated P-lactoglobulin eluted as a narrow peak. Whereas, the peak of the heated P-lactoglobulin had decreased in height and had broadened towards the higher M, end. K-Carrageenan in the unheated and heated mixtures of 1:l (0.1%) K-carrageenan$lactoglobulin eluted in the same position as in the 0.1% K-carrageenan solutions (Figure Ic). Again, the maximum DRI signal of the unheated K-carrageenan sample was slightly higher for the peak corresponding to K-Canrlgeenan in the unheated sample compared to the heated sample. In addition, there appeared to be a broadening of the DRI signal at the low M, end of the K-carrageenan peak for the heated mixture. The p-lactoglobulin peak in the mixture was broader and lower for the heated sample than for the unheated sample. In addition, the plactoglobulin peak in the mixture was broader than the P-lactoglobulin peak in the absence of K-carrageenan. The average M, and polydispersity coefficients for K-carrageenan and 0-lactoglobulin are given in Table 1. In general, the polydispersity coefficients for all samples were 1.1 suggesting that both the K-carrageenan and P-lactoglobulin solutions were reasonably monodisperse, confirming Figure 1, which showed only a gradual decrease in M, across the peaks. The M, profiles of heated p-lactoglobulin (Figures l b and Ic) showed an increase in

-

191

Mixed Biopolymer Systems

the amount of higher M, aggregates in heated samples compared to the unheated samples, hence the higher polydispersitycoefficients. There was a tendency for M, to increase with heating, except in the case of heated 0.1% K-carrageenan (Table 1). For p-lactoglobulin, the increased M, coincided with a broadening of the high M, part of the DRI elution profile (Figures lb and lc), which is due to the formation of non-native dimers, trimers and larger aggregate^.^"^^ The increase in M, of plactoglobulinafter heating was even greater in the presence of K-carrageenan, suggestingthat the presence of K-carrageenan accelerated the aggregation of 0-lactoglobulin.This may be 10 6 0.2 n

-L

105

rd

0.1

2

10 4

106

n t

i (b)

-2

n

.$

105

4

J-

-

---

$* 1 0 4 ;

A

- 0.2 - 0.1

Erd

.% 2 n v)

- 0.0 103 106

l@

E

104

2 n 0

5

I

1

I

10

15

20

J 25

Volume (mL)

Figure 1 Mw and DRI detector signal vs. volume for heated and unheated (a) 0.1% Kcarrageenan, (b) O. 1% plactoglobuh and (c) 1:I (0.1%) ?c-carrageenan:plactoglobulin. Mw (1,a) and DRI signals (-, ...) for heated and unheated samples, respectively. Void volume is indicated by an arrow.

192

Gums and Stabilisers for the Food Industry 10

Table 1 M, and polydispersity coeficients for non-gelling samples, data enclosed by brackets is the percent error between three replicates. Sample M, x lo4(g/mol) Unheated 0.1% K-carrageenan 30.26 (2.28) 0.1% P-lactoglobulin 2.78 (3.47) 1:1 (0.1%) mixture: K-carrageenan 30.92 (2.26) 0-lactoglobulin 2.74 (1.75) Heated 0.1% K-carrageenan 28.42 (3.15) 3.03 (6.41) 0.1% P-lactoglobulin 1:1 (0.1%) mixture: K-carrageenan 33.56 (0.95) P-lactoglobulin 3.20 (10.87)

Polydispersity Coefficient 1.15 1.04 1.15 1.03 1.17 1.12 1.08 1.32

due to phase separation, which would increase the effective concentration of P-lactoglobulin, and hence its aggregation For K-carrageenan, the M, for the K-carrageenan component of the heated mixture was significantly (P < 0.05) greater than both the M, for the K-carrageenan component of the unheated mixture and the M, for 0.1% heated Kcarrageenan. In addition to the difference in M,, there was an increase in the area under the peak of the DRI detector signal at the low M, end for the K-carrageenan peak in the heated mixture compared to the unheated mixture (Figure lc). The M, calculated for K-carrageenan in this region was similar to M, values determined for K-carrageenan molecules that eluted earlier on. This suggests a change in the conformation of K-carrageenan in the heated mixture, which may be due to an increase in the effective concentration of K-carrageenan as a result of phase separation. However, polymer concentration alone has not been reported in the literature to directly affect the coil-helix transition of K-carrageenan. Moreover, it cannot explain why a shift in the DRI signal was not observed for the K-carrageenan peak in the unheated 1:1 K-carrageenan$-lactoglobulin samples, where phase separation may have occurred, unless there was a specific effect of the presence of denatured P-lactoglobulin. Another possible explanation for the shift in the DRI signal and the increase in M, for the Kcarrageenan peak in the heated 1:1 mixtures is the occurrence of complex formation between K-carrageenan and denatured 0-lactoglobulin molecules. The M, of K-carrageenan for the heated mixture was about 20 000 g/mol greater than the M, observed for K-carrageenan in all the other samples. This increase in molar mass is approximately equal to the molar mass of a P-lactoglobulin molecule (18 000 g/mol) suggesting the formation of a complex of one Kcarrageenan molecule with one P-lactoglobulin molecule. Evidence against the formation of K-carrageenan-P-lactoglobulincomplexes is the absence of a UV signal in the volume range where K-carrageenan eluted in the heated mixture (data not shown). The UV signal for 0lactoglobulin showed that P-lactoglobulin eluted in essentially the same position in the presence and absence of K-carrageenan for both heated and unheated samples. In another study on the heat-induced aggregation and gelation of P-lactoglobulin in the presence of Kcarrageenan at pH 7, K-carrageenan and P-lactoglobulin did not form complexes.’

Mixed Biopolymer Systems

193

3.2 Controlled Strain Rheometry Unheated and heated solutionsof 1.O% P-lactoglobulin did not form gels, whereas, heated and unheated solutions of 1.O% K-carrageenan and 1:1 K-carrageenan$-lactoglobulin mixtures did. Heated and unheated K-carrageenan and unheated mixtures formed clear gels. Gels formed from heated mixtures appeared to have two phases, one clear and the other opaque. Figure 2 shows the change in storage modulus (G) over time for unheated and heated solutions of 1.0% K-carrageenan. A representative example of the gelation profile for the unheated samples is given in Figure 2a. After the gelation point (where G' > lPa), G' initially increased sharply but then decreased. The maximum G'value was 655 f 121 Pa, while the final G' value was 264 f 238 Pa. In heated 1.O% K-carrageenan solutions, G' continued to increase steadily after gelation and reached a plateau towards the end of the experiments(> 12 000 sec) at 1187 f 28 Pa. The gelation profiles for unheated and heated solutions of 1:l K-carrageenan$lactoglobulin mixtures are given in Figure 3. The unheated mixtures (Figure 3a) showed a sharp peak in G ; suggesting that a gel network was formed but it then weakened or even collapsed. In some cases, there was a partial recovery in G! The maximum G'value for the unheated mixtures was 670 f 450 Pa. For the heated mixtures, G' increased rapidly at the gelation temperature and then reached a plateau value of 3615 f 49 Pa (Figure 3b). Differences in the behaviour of heated and unheated samples of 1.O% K-carrageenan and 1:1 (1.O%) K-carrageenan:~-lactoglobulin during gelation suggest that the heated gels were more stable and more elastic than the unheated gels. This difference in behaviour between the heated and unheated samples may be related to the sensitivity of the coil-helix transition of K-carrageenan, which is dependent on the cooling rate of solutions.'*Since the gelation of K-carrageenan is a two-step process involving a coil-helix transition followed by aggregati~n,'~ a disruption of the first step, would affect the second step. Possibly, the cooling of solutions fiom 75 to 60°C at a rate of 2"C/min disrupted the coil-helix transition of the K-carrageenan molecules, resulting in an incomplete or unstable fine structure, which may have led to a weak gel network. This would explain why only unheated samples yielded unstable gels, whereas heated K-carrageenan and mixed samples, which were reheated to 8OoC after the rapid cooling step (thus erasing the previous temperature history) and followed by cooling at a rate of OS"C/min, formed stable gels. The plateau value for G'in the heated mixture was approximately three times greater than that observed for the heated K-carrageenan solutions, suggesting that denatured plactoglobulin had a synergistic effect on the gelation of K-carrageenan. This may be explainedby the occurrence of thermodynamic incompatibilitybetween the macromolecules, as was reported by Capron et al.' This was further supported by the observation that the heated mixed gels appeared to have two phases. As a result of thermodynamic incompatibility, the effective concentration of the biopolymers would be increased, hence, the strength of the K-carrageenan gel may increase.

CONCLUSIONS K-Carrageenan interacts with proteins via electrostatic interactions?o which requires the

Gums and Stabilisersfor the Food Industry 10

194

2000

n

k

1000

W

50

b

gG

a

25

0 2000

E

W

'

(b)

. ..

n

1000 -

-75

g

-50

8G

b

a

. . . . . . . ..

-25

0 I

I

.

,

,

I

I

I

,

g

F

0

.

Figure 2 G' (-) and temperature (. ..) during gelation for (a) unheated and (b) heated I , 0% K-carrageenansolutions. 4000

100 75

n

g

2000

50

I

b 0

I

25 .

I

-

4000

g

1

e 8k w

g

F

J O 100

h

2000

b

.... 0 0

5000

10000

15000

Time (s)

Figure 3 G'( j and temperature (. ..) during gelationfor (a) unheated and (b) heated 1:1 ( I . O!?!?) K-carrageenan:Plactoglobuh solutions.

Mixed Biopolymer Sysiems

195

interacting species to carry opposite charges. This would explain why little evidence for complex formation was observed. Under the experimental conditions used in our experiments, K-carrageenan and p-lactoglobulin both have a net negative charge. Although there are local positive charges on p-lactoglobulin at pH 7, the ionic strength of 0.1 M NaCl may have been strong enough to screen these local charges, thus preventing interaction between protein bound amino groups and sulfate groups on K-carrageenan. This effect of the screening of local molecular charges has been observed for the formation of complexes between K-carrageenan and BSA at pH 6.5, which only existed at ionic strengths < 0.02 M.* K-Carrageenan and p-lactoglobulin appeared to be thermodynamically incompatible in non-gelling and gelling systems. However, the heated non-gelling system did show some evidence for weak complex formation.

ACKNOWLEDGMENTS The Foundation for Research, Science and Technology (New Zealand), Natural Sciences and Engineering Research Council (Canada) and Ontario Dairy Council (Canada) are acknowledged for financial support.

REFERENCES T.H.M. Snoeren, T.A.J. Payens, J. Jeunick and P. Both, Milchwiss, 1975,30,393. D.G. Dalgleish and E.R. Moms, Food Hydrocoll, 1988,2,3 11. D.D. Drohan, A. Tziboula, D. McNulty and D.S. Home, FoodHydrocoll, 1997,11,101. J. Hidalgo and P.M.T. Hansen, J A g r Food Chem, 1969,17,1089. S . Mleko, E.C.Y. Li-Chan and S. Pikus, Food Res Znt, 1997,30,427. P.B. Femandes, ‘PolysaccharideAssociation Structures in Food’, R.H. Walter, Ed., Marcel Dekker, New York, 1998, Chapter 8, p.257. 7. I. Capron, T. Nicolai and D. Durand, Food Hydrocoll, 1999,13,1. 8. V.B. Galazka, D. Smith, D.A. Ledward and E. Dickinson, Food Chem, 1999,64,303. 9. S.K. Samant, R.S. Singhal, P.R. Kulkarni and D.V. Rege, Znt J Food Sci Technol, 1993, 28,547. 10. V.B. Tolstoguzov, ‘Food Proteins and Their Applications’, S. Damodaran, Ed., Marcel Dekker, New York, 1997, Chapter 6, p.171. 11. M. Stading and A.-M. Hermansson, Curbohydr Polym, 1993,22,49. 12. C. Viebke, J. Borgstr6m and L. Piculell, Curbohydr Polym, 1995,27, 145. 13. K. Bongaerts, H. Reynaers, F. Zanetti and S. Paoletti, Mucromol, 1999,32,683. 14. M. McSwiney, H. Singh and 0. Campanella, Food Hydrocoll, 1994,8,441. 15. E.P. Schokker, H. Singh, D.N. Pinder, G.E. Norris and L.K. Creamer, Znt Dairy J, (accepted for publication). 16. S.P.F.M. Roefs and K.G. de Kruif, Eur J Biochem, 1994,226,83. 17. M. Verheul, S.P.F.M. Roefs and K.G. de Kruif, J A g r Food Chem, 1998,46,896. 18. A.-M. Hermansson, E. Eriksson and E. Jordansson, Curbohydr Polym, 1991,16,297. 19. A.-M. Hermansson, Curbohydr Polym, 1989,10,163. 20. N.F. Stanley, ‘Food Gels’, P. Harris, Ed., Elsevier Applied Science, England, 1990, Chapter 3, p.79.

1. 2. 3. 4. 5. 6.

CASEIN MICELLES AND THEIR INTERACTION WITH EXOPOLYSACCHARIDES: TURBIDITY AND VISCOSITY

C. G . de

and R. TuinieI

'NIZO food research, Kemhemseweg 2,671 8 ZB Ede The Netherlands 2Van 't Hoff Laboratory of Physical and Colloid Chemistry, Debye Institute, Utrecht University, Padualaan 8, 3584 CH Utrecht The Netherlands

ABSTRACT We made a systematic investigation of the interaction of an exo-polysaccharide (EPS) with casein micelles as present in skim milk. The EPS was produced by the lactic acid bacterium Lactococcus luctis subsp. cremoris strain NIZO B40. The interaction of the EPS with milk components is of relevance for the consistency of fermented milk products. Low-heat skim milk was used to model a casein micelle suspension. At low EPS concentrations the mixture was stable. At higher concentrations the mixtures exhibited phase separation due to depletion interactions. In the one-phase region the strength of the segregative interaction could be determined from turbidity and rheological measurements. Then, by applying statistical mechanical methods the position of the phase boundary could be estimated. The viscosity of the mixtures was increased due to the segregative interactions in the EPS/casein micelle mixtures. The segregative interactions are of relevance for the final rheological properties of food products.

1

INTRODUCTION

In the food industry polysaccharides are used as thickening or gelling agents',*. Polysaccharides are usually extracted from plants. However, micro-organisms also excrete polysaccharides: exo-polysaccharides (EPS's). In some cases EPS's are produced in-situ in food products, notably in acidified milk products. These EPS's effectively function as food thickeners but need not be declared on the food label. As the EPS studied can be found in dairy products our focus is on the interactions and phase behaviour of EPS with the colloidal components in milk. There are three distinctly different types of particles in the colloidal size range in milk: fat globules, casein micelles and whey proteins. Smaller molecular species (over 100,000 in milk) are considered as part of the continuous phase. Here we present a study of the interaction of an EPS with casein micelles3-'. We focus on the effect of the interactions on the turbidity and the viscosity of the mixtures. The EPS does not bind to casein micelles at near neutral pH.

Mixed Biopolymer Sysrems

Therefore the interactions will be described in terms of the segregative, or depletion, type of interactions. This type of interaction arises from the excluded volume for the polymers around the particles and leads to an effective attraction between the colloidal particles. A schematic representation of depletion interactions is given in figure 1. Asakura and Oosawa' first described the strength of the interactions between two flat walls. Later, independently, Vrij lo calculated the free energy of interaction for two spherical particles as induced by the polymer coils present in solution. The depletion interaction is of an entropic nature and was put equal to the free energy which becomes available to the system when two spheres share their excluded volume for the spheres (Voverlap), so as to make it available to the polymer. Therefore AG = ll Voverlap gives the free energy of two interacting spheres, where ll is the osmotic pressure of the polymer solution.

Figure 1 Schematic picture of the depletion interaction mechamism. The osmotic pressure n o f the polymer solution pushes the two spheres together due to the fact that the pressure is unbalanced The depletion interaction induces an effective attraction and strongly affects the radial distribution function g(r) of the colloidal spheres. The quantity g(r) describes the probability of finding a particle at a distance r from the centre of another particle. It can be calculated from the interaction potential between the colloidal spheres and can be measured from scattering experiments". Light-scattering experiments on milky systems often suffer from multiple scattering, which strongly complicates the interpretation. We have therefore measured the turbidity of the mixtures of EPS and casein micelles, which also effectively measures the radial distribution function g(r) I*. In addition to affecting the equilibrium properties of the system, such as g(r), transport properties of the colloids are also affected when the attractions become strong. DhontI3 has shown that strong attractions lead to critical enhancement of the viscosity. The contributions of the individual particles in the mixtures can be taken into account into a background viscosity, while the attractions lead to a so-called anomalous viscosity. Bodnk and Dhont14 have shown that this concept works out well for model systems of colloids in the presence of non-adsorbing polymers. Here we

197

Gums and Stabilisersfor the Food Industry 10

198

apply it to our system and investigate whether we find an anomalous viscosity when the attractions become strong. In a sim le adhesive hard sphere model the rheology of a suspensions can be described by‘

P

1.9 2

qr = l + 2 . 5 $ + ( 5 . 9 + - ) $

+...

‘LB

where qr is the relative viscosity, I$ is the volume fraction and TB is the Baxter parameter, of which the inverse increases with the strength of the a t t r a ~ t i o n ~ ’ ~ ” ~ .

2

MATERIALS AND METHODS

2.1 Casein micelles Reconstituted skim milk are suspensions containing 10-13 vol% of casein micelles. Casein micelles are spherical particles consisting of an association of casein proteins and some calcium phosphates. The casein micelles have 100 nm radius with a not too polydisperse size distribution’6. Skim milk was prepared, by dissolving specially produced skim milk powder NILAC, (Ede The Netherlands) in distilled water497.Skim milk permeate (i.e. the ‘solvent’ of the casein micelles) was prepared from skim milk by filtration through a membrane with a cut-off of 0.1 pm5(so-called ultrafiltrate (UF) or permeate (PERM)).The pH of the permeate was the same as that of the skim milk (PH 6.6). The continuous phase contains up to 19’0 whey proteins, 4.6% lactose and salts with an ionic strength of about 0.1 M. It was shown that the whey proteins and the lactose exhibit no specific interaction with the EPS so the permeate can be considered as a ‘buffer’ solution.

2.2 Exo-polysaccharide The EPS from the Lactococcus Zactis subsp. cremoris NIZO B40 strain was produced on a pilot-plant scale at NIZO food research and isolated and purified as described previously Size exclusion chromatography combined with multi-angle light scattering, (SEC-MALLS)I7 analysis of the polysaccharide in aqueous 0.10 M NaN03 solutions yielded a number-averaged molar mass of (1.47 f 0.06).1O3 kg/mol, and a radius of gyration (number-averaged) of 86 f 2 nm.

’.

2.3 Turbidity measurement Turbidity experiments on casein micelle/EPS mixtures were made with a Hitachi (model U-1 100) single beam spectrophotometer and the samples were measured in quartz glass cuvettes with a path length of 2 mm (Hellma, type 110 QS). The turbidity was determined by measurement of the transmissions of skim milk (with or without EPS) and permeate. The mixtures were filtered (pore size 5 pm) before use.

199

Mixed Biopolymer System

2.4 Viscometry Ubbelohde capillary (Schott-Geriite) measurements were used to determine the viscosity of the casein micelle suspensions, the EPS solutions and the protein/EPS mixtures. The kinematic viscosity was calculated taking into account the Hagenbach correction. The solution density, needed to convert the kinetic into the dynamic viscosity, was measured using an Anton-Paar apparatus. All measurements were made at 298 K.

3

RESULTS

3.1 Turbidity measurements In figure 2 we plotted the relative turbidity of the dispersions T/TO where T is the turbidity of the solutions with EPS and TO the turbidity of skim milk without EPS as a function of L2.

1.6

cP = . 1.0 g/l A

1

0

1 .o

0

0.7 A

1

2

3

4

5

(pm'2)

Figure 2(a) Relative turbidity dro as afinction of the inverse of A2for 6 vol% of casein micelles (0 =O. 06) and various EPS concentrations as indicated in the figure. The drawn curves represent a statistical mechanical calculation.

200

Gums and Stubilisersfor the Food Industry 10

Basically the calculated curves were obtained by integrating the structure factors over the relevant Q-range. It must be realised that the light scattered from the primary beam accounts for the turbidity of the system provided that no absorption takes place. Although the calculations do not perfectly match the experimental data the consistency is quite good, taking into account that we did not use adjustable variables. For the calculations we used only independently determined parameters such as volume fractions of casein micelles, their size and the radius of gyration and molecular mass of the EPS.More details of the calculations are given elsewhere6.

1.0'

0

'

' 1

'

' 2

'

' 3

1

'

' 4

'

5

k 2 (pm-2)

Figure 2(b) '4s in 2(u) but for un 8 vol% casein rnicelle suspension

3.2 Viscosity In figure 3 we plot the relative viscosity of an EPS solution in milk UF permeate lower curve and the viscosity of the same solution but with casein micelles (CM ) which corresponds to a volume fraction of 10%. Adding EPS to the CM-solution increases the viscosity more than could be expected from the combined viscosity. The reason is that the effective attraction between the CM-particles increases the viscosity as is discussed quantitatively by Woutersen and de Kruif'* and Tuinier and de Kruif6.

20 1

Mixed Biopolymer System

2.5

0.6

2.0

0.4

1.5

0.2

r'

1 .o

0.0

0.0

0.2

0.4

0.6

c, (9/1) Figure 3 Relative viscosity of an EPS solution and a casein micelle (CM)/EPSmixture as ahnction of the EPS concentration. The data points (open symbols) give measurements and the curves representpredictions for (adhesive) hard sphere suspensions. Thefilled symbols refer to the calculated Baxter parameters.

Viscosity measurements at other CM-concentrations are quantitatively different but show exactly the same behaviour. We also measured the diffusion coefficient of the casein micelles in the presence of EPS. Again these results can be quantitatively related to a model in which the EPS induces an effective depletion type of attraction between the casein micelles. 4

CONCLUSIONS

Adding EPS to a dispersion of casein micelles as in skim milk leads to segregative interactions due to the difference in excluded volume. These interactions change the probability distribution of casein micelles in space. As a result both viscosity and turbidity are enhanced. These effects can be quantitatively understood by applying statistical mechanical models for interacting particles. The viscosity increase is in accordance with the increase of viscosity of so-called adhesive hard sphere dispersions. Not shown here, but in accordance with statistical mechanical theories the system will tend to phase separate into two phases: an EPS-rich phase and casein micelle-rich phase if the interactions are strong enough.

202

Gums and Stabilisers for the Food Industry I0

5

ACKNOWLEDGEMENT

The Netherlands Association of Biotechnology Centers (ABON) financially supported this work. Alexandra Le Roy and Blandine Oudin are thanked for performing various experiments. We thank Dr. C. Holt (Hannah research institute, UK) and Dr. P.A. Timmins (ILL, Grenoble, France) for their valuable help with the SANS experiments.

References 1 2 3 4

5 6 7 8 9 10 11 12 13 14 15 16 17 18

V. Ya, Grinberg, V. B. Tolstoguzov, FoodHydrocoll. 11 (1997) 145. W. J. Syrbe, Bauer and H. Klostermeyer. Znt. Dairy J. 8,1998,179 M. E van Marle, ‘Structure and rheological properties of yoghurt gels and stirred yoghurts’,Thesis. 1998. NIZO food researcWTwente University, The Netherlands R. Tuinier, ‘An exocellular polysaccharide and its interactions with proteins’, Thesis 1999. NIZO food researchlwageningen Agricultural University, The Netherlands R. Tuinier, E. ten Grotenhuis, C. Holt, P. A. Timmins and C. G. de Kruif, accepted for publication in Phys. Rev. E. R. Tuinier and C. G. de Kruif, J. Chem. Phys. 110 (1999) 9296. R. Tuinier, P. Zoon, C. Olieman, M. A. Cohen Stuart, G. J. Fleer and C. G. de Kruif, Biopolymers, 49 (1999) 1. R. Tuinier, J. K. G. Dhont and C. G. de Kruif, submitted. S. Asakura and F. Oosawa, J. Chem. Phys. 22 (1954) 1255. A. Vrij, Pure & Appl. Chem., 48 (1976) 47 1. D. A. McQuarrie, Statistical Mechanics, Harper & Row, New York, 1976. R. Tuinier, E. ten Grotenhuis, C. Holt, P. A. Timmins and C. G. de Kruif, accepted for publication in Phys. Rev. E. J. K. G. Dhont, An Introduction to Dynamics of Colloids, Elsevier Science, Amsterdam, 1996. I. Bodnhr, and J. .K. G. Dhont, Phys. Rev. Lett. 77 (1996) 5304. B. Cichocki, B. U. Felderhof, J. Chem. Phys. 93 (1990) 4427. C. G. de Kruif. J. Dairy Sci. 81 (1998) 3019-3028 M. A. M. H o f i a n n and P. J. J. M. van Mil, J. Agric. Food Chem. 45 (1998) 2942 A. T. J. M. Woutersen and C. G. de Kruif, J Chem. Phys. 94 (1991) 5739.

A DESCRPTION OF MICELLAR CASEINKAPPA-CARRAGEFlNANMMED SYSTEMS BY MEANS OF CALORIMETRY AND RHEOLOGY.

S.Bourriot, C. Gamier & J.-L. Doublier INRA-LPCM BP 71627 44316 Nantes Cedex 3 (France)

ABSTRACT The properties of kappaanageenan (in the K ' form; 0.5% w/w) in the presence of micellar casein in 0.25 M NaCl or 0.01 M KCl + 0.05 M NaCl have been investigated at temperatures ranging from 20°C to 50°C by combining microDSC and rheology. MicroDSC allowed one to monitor the conformational transition of the kappacarrageenan. This occurred at 29-30°C upon cooling and 4 3 4 ° C upon heating in both ionic conditions. Mixing kappa-carrageenan with micellar casein at 50"C, the carrageenan being in the coil conformation, led to a phase separation phenomenon. Demixing was ascribed to a depletion-flocculation phenomenon. Rheological measurements at W C upon cooling showed that, whatever the conditions, the system had gelled. Its viscoelastic properties depended strongly upon the medium conditions and the casein content. These results could be interpreted on the basis of microscopic observations by Phase Contrast Microscopy (PCM) and Confocal Laser Scanning Microscopy (CLSM) that have been reported recently. The systems were clearly biphasic, the casein being concentrated in one phase while the kappaanageenan was located in the other. Moreover, in the casein-enriched phase, the micelles of casein appeared aggregated and tended to form a continuous network when the casein concentration was high enough (>3%). Phase Separation phenomena appear to be a major event taking place when kappa-carrageenan is gelled in the presence of micellar casein. The way the two phases are organised with respect to each other determines the resulting rheological properties. Particularly, at a high enough casein concentration, two interpenetratedcontinuous networks are formed which both seem to contribute to the viscoelastic properties of the system. 1 INTRODUCTION.

The widespread use of dairy desserts is largely related to the fact that hydrocolloids, mostly polysaccharides, impart texture to these foodstuffs through their thickening and gelling pperties. Carrageenans are one of the most widely used of these texture agents.

This is related to the well known development of synergistic properties and hence of a

204

G u m and Stabilisers for the Food Industry 10

specific texture when they are mixed with milk at neutral pH. This specificity is This generally ascribed to the ability of carrageenan to interact with casein micelles explanation originates from studies performed on carrageenan/kappa-casein systems'.', that is when kappa-casein is not involved in the casein micelle. The micellar form is a supramolecular assembly resulting from the association of four major casein fractions with the contribution of colloidal calcium phosphate. These investigations clearly showed that carrageenan and kappa-casein strongly influence each other and modify their respective behaviour in aqueous medium. This has been explained by the formation of ionic linkages between a specific zone of the kappa-casein chain and carrageenan molecules. However, it is not clear how such interactions can take place in casein micelles. The location of kappa-casein within the casein micelle makes this interaction difficult to establish. On the other hand, investigations have been performed on carrageenan/milk systems mostly on a rheological basis4-'. Although these provided interesting descriptions of the properties of the mixed systems, clear conclusions on the role of casein-carrageenan interactions could not be reached due to the presence in the medium of other components, serum proteins, lactose, and salt. In order to improve the knowledge of the role of micellar casein in the properties of carrageendmilk systems, we have undertaken a study using micellar casein in a system devoid of the other milk proteins and in well controlled ionic conditions. Part of this work has been recently published' with a description of the phase behaviour of the mixed systems as well as of the morphology of the mixed gel. In the present study, we present another part of this work on the basis of mimDSC and rheological experiments. Two different temperatures (50°C and 25°C) were considered on both sides of the coil-helix transition temperature (Tg) of kappa-carrageenan, that is either in the coil form or in the ordered form. By a proper choice of ionic conditions (0.25M NaCl or 0.01M KCl+O.OSM NaCl), we defined Tg= 29-30°C, the helixcoil temperature Tm being at 4344°C. 2 MATERIALS AND METHODS 2.1 Materials. Micellar casein was a native calcium phosphocaseinate sample kindly supplied by Laboratoire de Recherches et de Technologie Laitik (INRA Rennes, France). It had the following characteristics : total protein content 90.7% ; non casein protein 5.0% ;lactose 0.5% ; salts 8%. The kappa-carrageenan sample was provided by SKW Biosystems (France). Its was converted to its pure potassium form by ion exchange of a hot 2% carrageenan solut$n with a commercial ion exchange resin (Amberlite IR 120) followed by freeze-drying .

2.2 Methods. Micellar casein (108, w/w) was dispersed in 0.25M NaCl or in 0.01 M KC1+ 0.05 M NaCl at 2OoC,pH 7 by stirring with a paddle at 1300 tdmin for 5 minutes and then sonicated for 8 minutes at 50 Watts. The particles size distribution was checked using a Malvern Mastersizer IP laser granulometer. The average diameter was 0.25 pm which is in agreement with literature Kappa-carrageenan solutions (18, w/w) were prepared at 90°C in 0.25M NaCl or in 0.01 M KC1 + 0.05 M NaCl under magnetic stirring during 30 minutes. CaseWcarrageenan mixtures were prepared at 70°C under magnetic stirring for 15 minutes at a fixed carrageenan concentration (0.596, w/w) and with casein content ranging from 0 to 4% (w/w).

Mired Bwpolyrner Systems

205

DSC measurements were performed using a mimDSC III from Setaram (France). 0.8 ml of the hot solution or mixture were accurately weighed in the cell which was then cooled down and sealed. Measurements were performed between 10°C and 80°C upon successiveheating and cooling cycles at 1"Umin. The DSC traces from the first cooling and the second heating steps were used. Flow measurements have been performed using a controlled strain rheometer (Rheometrics Fluid Spectrometer RFS 11) with cone-plate geometry (diameter 5 cm, at 50°C after an equilibration time of 10 minutes. angle 0.04 rad, gap 50 p) The rheological char;icterization of the gels was carried out using a controlled stress rheometer (Carrimed CLS 100) in oscillatory mode with grooved parallel plates (diameter 4 cm; gap 500 pm). The sample was placed at 70°C on the plate preheated at 70°C which was then immediately cooled down to 20°C. The mechanical spectra, G' and G ' as a function of frequency, were obtained at u)"C after a waiting period of 1 hour. The strain amplitude was chosen at 0.5% that is well within the linearity limits of the viscoelastic behaviour. Melting experiments were performed using the same measuring system by programming the temperature increase at 0.5"Uminfrom 20°C to 70°C. the measurements being made at a fmed frequency (6.3 d s ) at a strain amplitude of 0.5%.

3. RESULTS 3.1. Flow behuviour at 50°C. At this temperature, the mixtures tended to phase separate yielding a bottom phase enriched with casein and an upper phase enriched with the polysaccharide '. However, the macroscopic phase separation appeared only several hours after the preparation of the mixture. Therefore, the viscosity measurements were performed during the period between the preparation and the appearance of the phase separation, that is on macroscopically homogeneous systems. The flow behaviour of such preparations is illustrated in Fig. 1 for 0.5% kappa-carrageenan solutions in 0.25 M NaCl to which increasing amounts of micellar casein have been added. The 0.5% kappa-carrageenan solution exhibited Newtonian behaviour with a viscosity of -7.5 Pa s. Adding casein yielded dramatic changes of these properties. The behaviour became shear-thinning and thixotropic, the higher the casein content, the more pronounced was the modification. This suggests a structuring of the system upon adding micellar casein, as was also evidenced by dynamic viscoelastic measurements (not shown), from a behaviour typical of a dilute mammolecular solution with carrageenan alone to that of a gel-like system for the mixtures. These results are in agreement with our previous observations dealing with mixtures of micellar casein with neutral pol saccharides (guar gum, locust bean gum, dextran) or sulfated ones (dextran sulfate) .41#' The structuring of the system was ascribed to the formation of casein aggregates originating from the phase separation process due to a depletion-flocculation phenomenon. The casein aggregates could yield at least temporarily a continuous phase at relatively high concentration, namely >3%. Since phase separation takes place in kappa-carrageendcasein mixtures', we therefore suggest that the synergistic rheological properties of these freshly prepared systems originate from a similar mechanism. The apparent structuring of the system at 50°C is merely the result of the aggregation of casein micelles which can lead to a continuous network provided the casein concentrationis high enough.

206

Gums and Stabilisersfor the Food Industry 10

0

0

20

40

60

s h e a r rate

80

100

(s-‘)

Figure 1. Flow curves of kappa-carrageenan+micellar casein mixtures (in 0.25M NaC1). Temperature 50°C; carrageenan concentration : 0.5%; casein concentrations: I %, 3% and 5%.

-I I -

-

1oo

lo-2

lo-‘

loo

10’ angular frequency (radls)

lo2

Figure 2. Mechanical spectra of kappacarageenan+micellar casein mixtures (in 0.01M KCl+O.OSM NaCl; temperature : 20°C); lines: 0.5% carrageenan (continuous: G ’; dashed : G’Y; triangles: + 0.5% casein; squares: + 4% casein Gfilled symbols: G’; empty symbols :G’Y.

10”

’

””.‘.’

.

‘..‘’‘.I

.

.......I

.

10’‘ loo 10’ angular frequency (radls)

lo2

Figure 3. Mechanical specta of kappacarageenan+micellar casein mixtures (in 0.25M NaCl; temperature : 20°C); same legend as in Figure 2.

3.2. Viscoelasticproperties at 20°C.Upon cooling a solution of kappa-carrageenan or a freshly prepad mixhm down to 20°C, a gel was formed due to gelation of kappacarrageenan.As monitored by DSC measurements, gelation occurred at 30°C in 0.25M NaCl and 28.5"C in 0.01 M KCl + 0.05 M NaCl in the absence of casein. Adding micellar casein did not result in a sigmficant change of this gelling temperature. Examples of the resultingmechanical spectra (G'and G as a function of frequency) at u)"c are shown in Fig. 2 and 3. In 0.01 M KCl + 0.05 M NaCl (Figure 2), kappacamageenan alone exhibited the classical properties of a gel with G ' / G -10 and no variation of the storage modulus as a function of frequency within the frequency range accessed (0.1-100 d s ) . The addition of micellar casein did not change the shape of the curves although the moduli were higher. This merely reflects a reinfonxment of the kappa-carrageenan network induced by the presence of casein micelles. For the highest casein concentration however, changes appemd with G'(o) more sloping and GIG' decreasing.This reflects the contributionof casein micelles to the viscoelastic properties of the mixed system. In 0.25 M NaCl (Figure 3). the viscoelastic properties of the kappaanageenan gel were quite different. Again, the storage modulus was higher than the loss modulus but G' displayed a frequency dependence with two appamt plateaux. Furthermore, the G' values were much higher than in the previous example (-ldPa). Addition of a small amount of casein yielded a decrease in G with no change in the shape of the curves whereas a higher amount (>3%) resulted in a slight increase in G'. The difference between the two series of results is illustrated in Fig. 4 for the G' variati0nS.a~a function of casein content and in Fig. 5 for tan 6 = G / G ' . In Figure 4, the reinforcement of the gel as estimated from the compaxison of G' to G', (G', : storage modulus of the carrageenan gel at 0.5%) is much more spectacular in 0.01 M KCl + 0.05M NaCl than in 0.25 M NacI; regarding tan 6, the elastic chafacter in 0.01 M KCl + 0.05M NaCl was slightly depressed by the addition of micellar casein whereas in the case of 0.25M NaCl, the curve passed through a maximum and then returned to -0.1. Apparently, in this latter case, the network softened first with a low content of casein and then was reinforced at high concentration. These results can be paralleled to microscopic observations we recently reported9. By combining Confocal Laser Scanning Microscopy (CLSM) and Phase Contrast Microscopy (PCM),we found that the mixed systems are biphasic with a casein-rich phase and a @-carrageenan-rich one, the former one being embedded inside the latter due to carrageenan gelation. At a low casein content, typically up to 1% casein aggregates are dispersed in the carrageenan-rich continuous phase. The contribution of the casein-rich phase to the rheology of the mixed system is limited. At -3% micellar casein and beyond, the casein-rich phase tends to form a continuous network within the system and apparently the two phases can become interpenetmed. Therefore the casein network contributesby itself to the viscoelastic properties of the mixed p l although the carrageenan network still predominates. The difference between the two medium conditions is not fully explained. The strucm of the carrageenan network appamntly differs between both conditions and it is likely that the size of casein -gates as well as their distribution in the mixed system depend upon the medium.

-

3.3. Thermal behuviour. In these experiments, the viscoelastic properties of the gel were! measured while heating meanwhile the conformarional transition of carrageenan chains was monitored by micromC measurements. The helixail transition (Tm) took place at 4 3 4 ° C for kappaarrageenan in the two medium conditions investigated, that

208

Gum and Stabilisersfor the Food Industry 10

a

6 0.01M KCI+ O.05M NaCl

0

Qi34

/'

b

0.01M KC1+0.05M NaCl

C

-/----A

0.1

A3

2

B 0.2

0.05M NaCl

t

0.25M NaCl

' ' ;' 1 2 3 micellar casein (%)

0' 0

'

1

0 0

d

Figure 4. G '/G'0 variations as a function of micellar casein concentration (temperature: 20°C; angular frequency : 0.1 rads).

2 3 micellar casein (%)

1

4

Figure 5. tan Svariations as afitnction micellar casein concentration (temperature : 20°C; angular frequency :0.I rad/s). of

10 m

& 1o'r

b el

loo! 10"

10" F

o;

'

io

'

bo

'

o;

. slo

Temperature ('C)

Figure 6. G' and G" variations as a firnction of temperature (upon heating; in O.OIM KCl+O.OSM NaCl)dfiequency : 6.3 rads); 0.5% kappa-carrageenan alone (continuous line: GI; dashed line : G'Y; 0.5% carrqgeenan + 0.5% micellar casein filled triangles: GI; empty triangles : G".

1

A

O' i 0

1

30

40 50 Temperature ("C)

60

Figure 7. G' and G" variations as a firnction of temperature (upon heating; in 0.01M KCl+O.OSM NaCl)dfiequency : 6.3 rads); 0.5% carrageenan + 2% micellar casein filled squares: G'; empty squares :G'Y; 0.5% carrageenan + 4 % micellar casein Cfilled circles: G'; empty circles :G'Y.

Mired Biopolymer Systems

209

is 1415°C above the transition upon cooling (Tg). Such a hysteresis is classically observed in the case of kappa-carrageenan and is ascribed to the aggregation of kappacarrageenan double helices. When adding micellar casein up to 4 8 , this transition temperature slightly increased from 44°C to 46°C in 0.25 M NaCl and from 43°C to 47.5"C in 0.01 M KCl + 0.05 M NaCl. In Fig. 6 and 7, G' and G variations as a are plotted in the mixed ionic solvent. For kappa-carrageenan function of temperaalone (Figure 6), a sharp decrease of G' was experiencedbeyond 40°C due to melting of the gel. The melting temperature, taken as the point where G'=G', was 41"C, that is a few degrees lower than the transition temperature by DSC. However, the difference is slight and it is clear that the gel melted as a result of the conformational transition of kappa-carrageenan chains. Addition of micellar casein at a low concentration (0.5% or 2%) as illustrated in Fig. 6 and 7 resulted in a slight shift of the curves but the overall shape remained. Similarly, we could determine the melting temperature from G'=G" : this tended to increase slightly as the casein content increased. Increasing more significantly this content (4% in the example shown in Figure 7) resulted in quite different variations with no crossing of G' and G" : a drop appeared in the range 4245°C and then a plateau was experienced. Beyond this temperature, carrageenan molecules are in the disordered state. The rheological properties of the hot system are governed by the casein-rich phase and are similar to those of the fresh mixtures at 50°C as shown in Fig. 1. Therefore, dramatic changes in the viscoelastic properties of the systems always occurred over the temperature range of the conformational transition of kappacarrageenan; it is noteworthy that similar results were obtained in the other medium condition (0.25 M NaC1). This is a clear confirmation that the viscoelastic properties of the gelled system are primarily governed by the kappa-carrageenan network. However, increasing significantly the casein content (>3%) resulted in the properties of the gels to become more and more effected by the casein-rich phase. This is consistent with microscopic observations which evidenced that these two phases tended to be interpenetrated when increasing the casein content. 4. CONCLUSIONS. The present data allowed one to describe accurately the viscoelastic properties of micellar caseinkappa-carrageenan mixtures taken as a model system. These rheological data could be combined to microscopic observations we recently reprkd9 to demonstrate the respective roles of the carrageenan-rich phases and the casein-rich one. Upon m l i n g the mixtures h m 50°C to 25"C, the system phase separates until the kappa-carrageenan network is formed, that is at 30°C in the present ionic conditions. Therefore, depending on the cooling rate, the extent to which the two phases will separate macroscopically can vary to a large extent. This provides an illustration of the importance of the rate of cooling in the final properties of dairy desserts implying gelling carrageenans. Furthermore, understanding the properties of real systems would require the present work to be extended to mixtures containing the globular proteins that are present in milk. These are known to be, at least partly, complexed onto the casein micelles when a thermal treatment has been applied thus resulting in strong changes in the surface properties of the micelle. The effect of such changes in the properties of casein micelles in the presence of kappa-carrageenan can be of impoxtance in the gelling properties and hence in industrial applications.

210

Gwns and Stabilisersfor the Food Industry I0

References 1. T.H.M. Snoeren, PhD Thesis, Wageningen, Nederlands Institute voor Zuivelonderzoek, Me, The Netherlands 1976. 2. T.H.M. Snoeren, T.A.J. Payens, J. Jennince, and P. Both, Milchwissenschufi 1975, 30, 393. 3. D.G. Dalgleish, E.R. Moms, Food Hydrocolloids, 1988,2, 311. 4. M.G. Lynch.and D.M.Mulvihil1, Food Hydrocolloids, 1994,8, 3 17. 5. S.Y. Xu,D.W. Stanley, H.D. Goff and V.J. Davidson ,. J. Food Sci., 1992,57,96. 6. D.D. Drohan, A. Tziboula, D. Mc Nulty and D.S. Home,. Food Hydrocolloids, 1997, 11, 101. 7. V. Langerdorff, G. Cuvelier, B. Launay, A. Parker, Food Hydrocolloids, 1997,11, 35. 8. V. Langerdorff, G. Cuvelier, B. Launay, C. Michon, A. Parker., C. G. De Kruif, Food Hydrocolloids, 1999, 13, 21 1. 9. S. Bourriot , C. Gamier, J.L. Doublier, Carbohydr. Polym., 1999,40, 145. 10. V.J. Moms, G.P Chilvers, Curbohydr. Polym. 1983,3, 129. 11. D.G. Schmidt, In ‘Developmentsin Dairy Chemistny’, (P.F. Fox, ed.),Vol. 1, Chap. 2, Applied Science Publishers, London, New York, 1982, p. 60. 12. H.E. Swaisgood, In ‘Developments in Dairy Chemistry’, (P.F. Fox, ed.), Vol. 1, Chap 1, Applied Science Publishers, London, New York, 1982, p. 1. 13. Boumot S., Gamier C., Doublier J.L., Food Hydrocolloids. 1998, 13,43. 14. Bourriot S., Gamier C., Lefebvre J., Doublier J.L., 1997. In ‘Proceedings of the

First International Symposiumon Food Rheology and Structure’, Ziirich, Switzerland.

EFFECT OF HEAT TREATMENT ON K-CARRAGEENAN GELATION IN MILK.

A. Tziboula and D.S. Home Hannah Research Institute Ayr, KA6 5HL

ABSTRACT Carrageenans are extensively used for the stabilisation of dairy products. Heat treatment is an essential part of dairy processing to ensure the bacteriological safety of the product and also aids the hydration of the hydrocolloid when present. In this work we examine the effect of various heat treatments on the temperature dependent gelation of K-carrageenan in milk, using small deformation rheology at 0.08Hz and an applied stress 5mPa. Skim milk was heated to temperatures ranging from 70150°C for 15-300 seconds. s Carrageenan was added either to unheated skim milk or preheated milk. The hydrocolloid was added either in anhydrous form or dispersed and hydrated in skim milk ultrafiltrate (SMUF).It was found that gel strength increased with heat treatment and reached a maximum at approximately I I0 “C. Prolonged heating at higher temperatures resulted in loss of gel strength. The decrease in the pH of milk and the changes in the ionic equillibria due to heat treatment did not have a profound effect on carrageenan gelation. It was concluded that the heat treatment brought about physicochemical changes on the casein micelles, which influenced the rigidity and flexibility of the milk-carrageenan gel network. It is likely that scarrageenan forms a complex with K-casein on the surface of the micelles. I n unheated milk or milk which has been heat treated at temperatures below 90 “C, the complex remains attached on the surface of the micelles. At elevated heating temperatures ccasein dissociates from the micelles and hence the Kcarrageenadcasein complex becomes detached from the casein micelles. This results in an increase in the flexibility of the carrageenan network, which is manifested by a loss in the complex modulus of the gels. Heat treatment per se did not interfere with the interaction between carrageenan and K-casein. 1 INTRODUCTION

Carrageenans are traditionally used as stabilisers in protein systems particularly in dairy applications. The stabilising ability of carrageenans is limited to the gelling types only and has been attributed to an electrostatic interaction between K-casein and carrageenan. However, the nature of the network in milk-carrageenan gels and the relative contributions of the milk proteins in the gel structure are still subject of debate. The fact that gel structure is necessary for stabilisation, supports the notion that the gel network involves only carrageenan-carrageenan cross-linkages and not carrageenan-casein or casein-casein linkages. On the other hand, in model systems a

212

Gums and Stabilisersfor the Food Industry 10

specific interaction between K-carrageenan and K-casein is evident, as no sol-gel transition occurs with asl-or p-casein under like conditions (1- 6). Studies of K-carrageenan gelation in the presence of milk proteins using dynamic rheology, showed that gelation differed according to the milk protein composition of the dispersing medium (7). In skim milk ultrafiltrate which is devoid of any milk proteins, gelation took place in two steps: the first step corresponded to the coil-tohelix conformational transition of K-carrageenan whilst the second step signalled the onset of network formation. In the presence of milk proteins, gelation proceeded in a single step following the conformational transition of the polysaccharide (7). The effect of heat treatment on K-Carrageenan gelation also differed according to the type of milk protein present in the carrageenan dispersion. Thus, in the absence of milk proteins, for example when carrageenan was dispersed in skim milk ultrafiltrate, varying the severity of heat treatment had little effect on gelation. Conversely, heat treatment became important in the presence of casein micelles. With mild heat treatments (60"C/ for 1-10 minutes) there was no gel formation and only when the temperature of heating was raised above 70°C did gelation become evident (8). Heat induced denaturation of the whey proteins and their interaction with the casein micelles did not interfere with carrageenan gelation in milk (8). However, because carrageenan was added to milk prior to heat treatment, it was not clear whether heat treatment interfered with the interaction between the carrageenan and the milk proteins or whether the effects were a manifestation of changes occuring in the milk fraction during heating. Furthermore, heating the milk-carrageenan samples at different heating temperaturehime combinations might also be expected to affect the extent of carrageenan hydration with potential consequences on the gel properties. To elucidate these points we examined the path of carrageenan gelation in milk which was heat treated at temperatures ranging from 70- 1 50°C for 15-300sec, before the addition of carrageenan. Carrageenan was added either i) in anhydrous form with hydration achieved by re-heating the milk-carrageenan mixtures at 65°C for 20 minutes or ( i i ) dispersed and pre-hydrated in skim milk ultrafiltrate (SMUF). The SMUF-carrageenan dispersion was then combined with a preheated milk sample. The path of gel formation was followed using low frequency oscillation as the milkcarrageenan mixture was cooled from 60-5°C at a rate of 1°C /I0 minutes. The data was compared with parallel experiments in which carrageenan was added to unheated milk and the samples were then heat treated at temperatures ranging from 70-150°C for 15-300 sec. This experimental design allowed us to evaluate whether the effect of heat treatment was linked with changes in the milk fraction, whether it influenced the extent of carrageenan hydration and hence gel structure or influenced the interaction between carrageenan and milk proteins. The effect of heat treatment on the milk proteins differs with the intensity of heating; low temperatures and short holding times have a modest and often reversible effect whilst more severe heat treatments result in irreversible changes in milk (9).

2 MATERIALS AND METHODS. 2.1 Preparation of milk-carrageenan mixtures. Bulk milk from the Institute herd was

skimmed in the pilot plant using the AFM separator (S400 Marano, Italy). K-Carrageenan-milk mixtures were prepared by adding:

Mixed Biopolymer Systems

213

i ) Anhydrous K-carrageenan to unheated milk: K-Carrageenan (350ppm) was added to skimmed milk under stirring at room temperature. The sample was then homogenised for 2 min (Silverson, laboratory homogeniser). Hydration of the polysaccharide was achieved by heating at 65"C/20min in a steam-jacketed vessel. Portions of the milk-carrageenan mixture were heat treated further in the UHT plant of the Hannah Research Institute by indirect heat at 70, 90, 110, 130 and 150°C for 15, 60 and 300 sec (total of 15 treatments) in an HTST plate heat exchanger (Junior Paraflow, APV). They were then cooled to 12°C and stored at 4°C. Before the rheological test each sample was reheated at 65°C for 20 minutes (to melt any carrageenan gel formed during storage) and immediately transferred to the rheometer which was sitting at 60°C. ii) Anhydrous ccarrageenan to pre-heated milk : Skimmed milk samples were heat treated in the UHT plant as above. To portions (35 ml) of heated milk K-carrageenan (350ppm) was added as in (i). Hydration of the polysaccharide was achieved by heating at 65"C/20 min. The milk-carrageenan mixture was immediately transferred to the rheometer which was sitting at 60°C. iii) Hydrated K-carrageenan to pre-heated milk (10.3%TS): Skimmed milk was concentrated by ultrafiltration using a membrane with 30,000 MW cut off (Memtech Ltd, Swansea, UK), to a total solids level of 10.3%. The concentrated milk has then heat treated as described in (i) and stored at 4°C. K-Carrageenan (0.28% w/w) was added to skim milk ultrafiltrate (SMUF) and hydrated by heating to 90°C for 20 minutes. The K-carrageenan/SMUF dispersion was then combined with the concentrated milk at 50°C to give final K-carrageenan concentration 350ppm and total milks solids of 9%, equivalent to that of skimmed milk.

2.2 K-carrageenan was a gift from SKW Bio systems. It was extracted from Kappaphycus alvarezii (common name Euchema cottonii). It was essentially pure K-carrageenan ( > 98 %) in the sodium form. 2.3 Dialysis of heat treated milk samples When appropriate the pre-heated milk samples were dialysed against unheated milk over 48hr; 5 milk changes; ratio of dialysate to unheated milk, 1:300. 2.4 Small deformation rheology Low frequency oscillation (0.08Hz; applied stress SmPa) was performed on the milk carrageenan mixtures as they were cooled from 60" to 5°C at a cooling rate 1°C 10 m i d (CVO, Bohlin Instruments, Gloucestershire, UK, fitted with the double gap measuring geometry). The geometry consists of a hollow cylinder (diameter 45mm) which is lowered into a cylindrical groove in an outer cylinder (diameter 5Omm). The instrument calculates the phase angle (6) of the delay in the strain response of the sample. From 6 the complex modulus can be calculated (G*)= ( I G' I + I G" I2)1'2. The gelation temperature of the gels is defined as the temperature where tan 6 = 1. Evaporation in the sample was minimised by fitting the solvent trap provided on the double gap geometry. The applied stress SmPa was in the linear viscoelastic region of the structures created. The strain response varied with the viscoelastic response of the sample during the cooling cycle. The initial

214

Gum and Stabilisersfor the Food Industry 10

strain response at 60°C was constant around 1.27 whilst at 5°C it ranged from 0.090.015. 2.6 Statistical analysis The results presented in this work are the mean values of duplicates. The data was fitted with second-degree polynomial models y = bo+b,~i+bzrz+bii~i~+bzzr~~+bizri~z -here bii are the coefficients and xi , heating temperature (70-150°C); xz. heating time (15-300 sec). (Minitab version 11.2). 3 RESULTS AND DISCUSSION.

The data presented here deals with the development of the complex modulus (G*) during gelation. G* encompasses the contributions from the elastic and viscous components in gel structure. It has been previously shown that the development of the G* during carrageenan gelation follows a simple scaling behaviour, which is a function of two independent parameters, gel strength at infinite time and gelation time (GT) (7). The same scaling behaviour in gel structure development was also

I

2.5 -

5 '4

B

3

2.0 -

1

1.5 -

T 'p

1.0 -

0.5

-

0.0 1 50

r

70

90

110

130

1

150

170 50

70

90

110

130

150

170

Temperature ("C)

Figure 1. Effect of heat treatment of milk on the G*1,6(at I.6xGT) of m'lkcarrageenan gels. Carrageenan was added in previously heat treated milk in anhydrous form (a),pre-hydrated in SMUF (b,. Holding times at each heating , 300 sec. Mean values of temperature : ..........., 15 sec; -------,30 sec; duplicates. The error bars indicate standard deviation (SD). Trendlines were calculated using second degree polynomial models. observed in this work. The scaling development of gel strength was independent of the heat treatment of milk (results not shown). The existence of scaling behaviour implied that within the time scale of the present experiments, comparison of the data should be made at a time interval which is a multiple of the gelation time (GT) for each run. In this work, comparisons of the complex moduli were made at 1.6 x GT, designated G* I .6. Plots of the complex moduli (G*1,6) of the milWcarrageenan gels as a function of the heating temperatures are shown in Fig. 1. Carrageenan was added to preheated

Mixed Biopolymer Systems

215

milk either in anhydrous form (Fig la), or pre-hydrated in SMUF (cf materials and methods) (Fig. Ib). In both cases, the final carrageenan concentration in the samples was 350ppm and the total milk solids 9%. It can be seen that irrespective of the mode of carrageenan addition to milk, there was an increase in gel strength with an increase in heat treatment from 70-90°C. G*1,6 reached a maximum at heating temperatures 90-110°C and there was a gradual loss in gel strength with further increase in the severity of heat treatment. At relatively moderate heating temperatures, typically less than llO"C, increasing the holding time had an enhancing effect on gel strength. However, at temperatures greater than 110°C increasing the holding time had a detrimental effect on gel strength. The weakest gels were obtained at the extremes of heating temperaturehime combinations, ie heating at 150°C for 300 sec. A parallel experiment was run in which carrageenan was added to unheated milk and the mixtures were then subjected to heat treatments at temperatures ranging from 70150°C for 15-300sec. A similar trend of the effect of heat treatment on gel properties was observed (Fig. 2). Examination of the rheological profiles showed that the gelation temperatures (GT's) of the mixtures which were heat treated after the addition of carrageenan were about 7°C higher than in the samples where carrageenan was added to pre-heated milk (GT's 31 f1"C and 23 f 2°C respectively). A tendency for lower GT's to be associated with the weaker gels was also consistently k observed. However, the differences in E v GT's for the 16 heat treatments were not statistically significant. Overall, it b was concluded that the effect of heat treatment on carrageenan gelation was associated with heat-induced changes in the milk fraction. The effect was not influenced by the extent of 1.oo50 70 90 110 130 150 170 carrageenan hydration or the position of the heat treatment in the processing Temperature ("C) procedures. To facilitate prediction of the Figure 2 Effect of heat treatment of milk on behaviour of milk-carrageenan the G* at 7°C. K-carrageenan was added to: mixtures the data was fitted to a I unheated skim milk; - - - - -, preheated response surface using regression skim milk. Mean values for the three holding analysis techniques. The best fit was achieved with a second-degree times at each temperature. polynomial model described by the equation: y=bo+blxl+bzx2+bl1~1~+bl2~1~2, where bii are the coefficients and XI, temperature; x ~ log . heating time. Results from analysis of variance are shown on Table 1. The regression equation was statistically significant (pcO.OOO1) as were the linear terms for both temperature and log time. From the square terms only temperature was found to have a significant effect (pc 0.OOOl). As expected there was a statistically significant effect from the two way interaction of the heating temperature with log time (p
s"

''5

216

Gums and Stabilisersfor the Food Industry 10

During heat treatment there are numerous changes in milk associated with the casein micelles, ionic equillibria, lactolysation between the E-aminogroups of protein bound lysines and aldehyde groups of lactose, decomposition of lactose and formation of organic acids with concomitant decrease in pH, and the formation of covalent bonds (9). Any of these heat induced physicochemical changes could potentially have an impact on the interaction between K--carrageenan and milk proteins. Research has shown that at the natural pH of milk (around pH 6.8), heating temperatures up to 90°C have little effect on micellar size (10-13). Table 1. Response surface regression: analysis of variance and the estimated regression coefficients of the effect of heating temperature and holding time on the complex moduli at 1.6 x GT (G*1.6) of milk-carrageenan gels. Carrageenan was added in pre-heated milk. Levels of statistical significance: p
Anhydrous carragJ

Analysis of variance

Hydrated carrag./

I!!!E!@..!!!!k:

.J?!Y!.!ated.mi!k r 0.86

Regression equation Linear terms Square terms Interaction Terms Constant Heating Temperature (HT) Log Holding Time (LHTi)

(HV2 (HTx LHTi)

Coeff. -1.4 0.058 0.873 -0.o001 -0.007

**** **** **** **

r 0.79

P

Coeff. -1.4

P

.......-..

P

**** **** **

****

NS

**** ** **** **

0.059

0.988 -0.o001 -0.011

p NS

**** *** ** ****

2.4

2

.-

1.9

.-e

z

E2 1.4

4 70

90

110

130

70

90

110

130

150

Heating Temperature ("C)

Figure 3. Contour plots of the estimated logarithm complex modulus (G*l.a) of carrageenan gels at 1 . 6 GT, ~ as a function of heating temperature and holding time. The numbers in italics indicate the log values of G*1.6 (mPa). K-Carrageenan was added to preheated milk ( a ) prehydrated in SMUF by heating at 90°C for 20 minutes; (b) anhydrous and then the carrageenan-milk samples were reheated at 60 "Cfor 20 minutes for the hydration of the hydrocolloid.

Mixed Biopolymer Systems

217

A slow increase in average micellar diameter is observed with heating temperatures up to 130"C, followed by a rapid increase at more severe heat treatments (14,15). The denaturation of whey proteins is also limited at mild heat treatments. Heating of milk above 90°C results in denaturation of b-lactoglobulin and interaction with K-casein via disulphide interchange reactions (16, 17). The P-1actoglobulidK-casein complex finally dissociates from the surface of the micelles at prolonged heating. However, it has been shown previously that whey protein denaturation does not interfere with Kcarrageenan gelation in milk (8). Heating at elevated temperatures results in hydrolysis of K-casein and dephosphorylation of GI-, ~ 2 and .p-caseins. It also induces the precipitation of primary and secondary calcium phosphate (18,19) with concomitant release of H+ resulting in decrease in the pH of milk. Thermal breakdown of lactose and formation of organic acids contributes further to the pH decline during heating. This decrease in pH is likely to affect the electrostatic interactions involved in maintaining the integrity of the micelles. With heating the concentrations of soluble phosphate as well as soluble and ionic calcium decrease (20,21). Soluble calcium and phosphate precipitate as colloidal calcium phosphate on the surface of the casein micelles (22). Depending on the severity of heat treatment these changes can be reversible and on subsequent cooling, some or all of the precipitated material re-dissolves with time (23). However, when the temperature of heating is greater than llO"C, the precipitated material does not redissolve. The deposition of calcium phosphate onto the casein micelles is expected to shield the negative charges on the micelles, thus reducing their zeta potential and effectively decreasing the electrostatic repulsions. Heating has little effect on the monovalent ions, sodium, potassium and chloride. K-Carrageenan gelation is particularly sensitive to changes in the ionic composition of the dispersing medium and to a lesser extent changes in pH. In this work it was found that heat treatment up to 150°C for 15 or 30 sec had little effect on the pH of milk (Table 2). A decrease in pH was only observed after prolonged heating for 300 sec, at 130 and 150°C (Table 2). Furthermore, restoration of the pH of the heated milk samples to that of unheated milk by dialysis did not result in an increase in gel strength (Fig. 4b). Therefore, it was concluded that heat induced changes in the pH of milk were insignificant and were not associated with the effect of heat treatment on K-carrageenan gelation in milk

Table 2. pH of milk samples after heat treatment.

Gums and Stabilisersfor the Food Industry 10

218

In order to determine whether changes in the ionic equilibria of milk brought about by heating were responsible for the differences in the gel properties, the ionic equilibria of the heated milk samples were re-established to that of unheated milk by dialysis. Thus, each heat treated milk sample was dialysed against unheated skim milk. At the end of the dialysis period, pre-hydrated carrageenan was added to the dialysed milk (cf materials and methods) and the path of gel formation in the dialysed milk-carrageenan mixture was followed by low frequency oscillation as before. The effect of dialysis on K-carrageenan gelation, at the extremes of heat treatments are shown in Fig. 4. It was consistently observed that dialysis caused a slight decrease in the GT which was attributed to changes in the cation activities in milk after dialysis. The gelation temperature of carrageenan can be predicted by calculating the cation activity contributions from the solution and the carrageenan itself, reduced by their activity coefficient of -0.5 (24, 25). However, neither the gelation path nor the final gel strength were influenced by dialysis. It was therefore concluded that the effect of heat treatment on K-carrageenan gelation in milk was not linked to changes in pH or ionic equillibria in the system. 2 2

(b) 150°C/300 sec

I0-

-I-

-2-3-

-4% 0

10

20

30

40

50

I

I

I

I

I

I

60

Temperature (“C)

Figure 4 Gelation profiles of milk-carrageenan mixtures during cooling from 60‘C5°C; K-Carrageenan was added to pre-heated milk at (a), 70T/3OOsec; (b), 150 T/300sec; Thin lines correspond to undialysed heated milk; Thick lines , G ’ ; ...........,G”. correspond to dialysed heated milks;. The present and previous observations of the effect of heat treatment on Kcarrageenan gelation in milk (7,8,26) suggest that gel formation involves only carrageenan-carrageenan cross-linkages and not K-carrageenank-casein linkages. The r61e of the casein micelles is to impart gel strength and rigidity to the gel by providing additional cross-links between K-CaITageenan and K-CaSein on the surface of the micelles. Because the effect of heat treatment on carrageenan gelation is observed only in systems containing casein micelles it is concluded that an electrostatic interaction between K-Casein and K-carrageenan is most likely. This complex is not involved in the mechanism of gel formation. Gel formation is the consequence of temperature mediated conformational transition of the carrageenan molecules from random coils to helical structures. Whey protein denaturation and the interaction between P-lactoglobulin and K-casein do not interfere with the carrageenan-casein complex. This is not surprising because the P-lactoglobulink-

Mixed Biopolymer System

219

casein complex involves sulphydryl4isulphide interchange between the sulphydryl groups on P-lactoglobulin and the cysteinyl residues on the hydrophobic domain of Kcasein. On the other hand, an interaction between K-carrageenan and K-casein would involve the gl ycopeptide region (on the hydrophilic domain) of K-casein and the negatively charged sulphate groups on the carrageenan molecule. If one is to accept the supramolecular structure of the carrageenan gels, where helical dimers of carrageenan associate into fine rigid rods that aggregate into long supermolecular assemblies (27-29), then the K-CaseidCaITageenan complex would simply influence the packing of the hydrated helical strands. The attachment of the casein micelles on the carrageenan chains would decrease the hydrocolloid flexibility, hence increasing gel strength. Thus, we explain the effect of heat treatment on gel strength as follows: mild heat treatments have little effect on micellar integrity, allowing carrageenan to form complexes with casein micelles (via K-casein on the surface of the micelles) and thus strengthening the gel network over that formed in the absence of casein micelles. Heating at temperatures greater than 1 10°C causes micellar disaggregation (1 I , 30, 3l), with K-casein consisting approximately 40% of the dissociated soluble casein fraction. In these systems the K-carrageenank-casein complex does not involve the casein micelles which translates to loss in structure rigidity, greater flexibility in the gel network and hence lower complex moduli. In the intermediate temperature regions, it is possible that upon heating, side chains of some aminoacids, especially that of lysine, become very reactive and interact with carrageenan, thus contributing further to the strength of the network, producing the observed increase in gel strength with heat treatment at relatively low temperatures of heating. 4 CONCLUSIONS

In this work we examined the effect of heat treatment on K-carrageenan gelation in milk. Our conclusion is that heat treatment influenced the physicochemical properties of the casein micelles and hence through their interaction with K-carrageenan, gelation in milk. Because heat treatment influences K-carrageenan gelation only in the presence of micellar caseins it is concluded that K-carrageenan forms a complex with K-casein. In milk-carrageenan systems this complex alone is not responsible for gelation but it provides another cross-linking pathway and reinforces the strength of the network formed. Conditions favouring the complexation of carrageenan with Kcasein on the surface of the micelles result in stiffer networks and stronger gels. Conditions, such as severe heat treatment that cause the disaggregation of the micelles have a weakening effect on the network structure and softer gels. Our results also suggest that it would be possible to tailor the properties of milkcarrageenan gels by manipulating the physicochemical properties of the milk fraction by varying the heat treatment of milk.

Acknowledgements The authors wish to thank Miss G. Rokoma and W. Steele for their technical assistance. This research was funded by the Scottish Office Agriculture, Environment nad Fisheries Department.

220

Gums and Stabilisersfor the Food Industry 10

References 1. C.F Lin, “Food Colloids”, (H.D. Graham) Avi publishing Co., 1977, p. 320. 2. P.M.T. Hansen “Gums and Stabilisers for the Food Industry” (G.O. Phillips, D.J. Wedlock and P.A. Williams), Pergamon Press, 1982, p. 127. 3. Th. H.M. Snoeren, PhD Thesis, Nederlands Institute voor Zuivelonderzoek, Ede, The Netherlands, 1976. 4. G. Stainsby, Food Chemistry,1980, 6,3. 5. T.A.J. Payens, J. Dairy Sci., 1972,55, 141. 6. Th. H. M. Snoeren, T.A.J. Payens, J. Jeunink, P. Both, Milchwissenschaft, 1975, 30,393. 7. A. Tziboula and D.S. Home, Int. Dairy J., 1999,9,359. 8. A. Tziboula and D.S. Home, Colloids and Sugaces B., 1999,12,299. 9. H. Singh and L.K. Creamer “Advanced Dairy Chemistry- 1: Proteins” (P.F. Fox), Blackie Academic and Professional, 1997, p. 621. 10. H. K. Wilson and E. 0. Herreid, J. Dairy Sci., 1961,44,552. 1 1 . H. Hostettler, K. Imhof and J. Stein, Milchwissenschaft, 1965 20, 189. 12. R. V. Josephson, E. L. Thomas, C. V. Morr and S. T. Coulter, J. Dairy Sci., 1967, 50, 1376. 13. R. J. Carroll, M.P. Thompson and P. Melnychyn, J. Dairy Sci., 1971,54, 1245 14. K. S. Mohamed and P.F. Fox, N. Z. J. Dairy Sci. Technol., 1987,22, 191. 15. D. G . Dalgleish, Y. Pouliot and P. Paquin, J. Dairy Res., 1987,54,39. 16. W.H. Sawyer, S.T. Coulter, R. Jenness, J. Dairy Sci., 1963,46,564. 17. P. Smits and J.H. van Brouwershaven, J. Dairy Res., 1980, 47,313. 18. P.F. Fox, J. Dairy Sci. 198 1,64,2127. 19. P.F. Fox, “Developments in Dairy Chemistry-I. Proteins, (P.F. Fox), Applied Science publishers, 1982, p. 189. 20. D.T. Davies and J.C.D. White, Proc. 151hIntern. Dairy Cong. (London),1959, 3, 1677. 21. D. Rose and H. Tessier, J. Dairy Sci., 1959,42,969. 22. N. Evenhuis and Th. R. de Vriesflerh. Milk Dairy J., 1956,lO. 101. 23. A. Kannan and R. Jenness, J. Dairy Sci., 1961,44,808. 24. D.A. Rees, Adv. Carbohydr. Chem. Biochern.,1969,24,267. 25. C. Rochas and M. Rinaudo, Biopolymers, 1980,19,2165. 26. D. Drohan, A. Tziboula, D. McNulty, D.S. Home, Food Hydrocoll., 1997, 11,

101.

27. A.-M. Hermansson Carbohydr. Polym., 1989, 10,163. 28. J. Sugiyama, C. Rochas, T. Turquois, F. Taravel, H. Chanzy, Carbohydr. Polym., 1994,23,261. 29. J. Borgstrom, L. Piculell, C. Viebke, Y. Talmon, Int. J. Biol. Macromol. 1996, 18, 223. 30. C.V. Morr, J. Dairy Sci., 1969,52, 1 174. 3 1. C.V. Morr, J. Dairy Sci., 1973,56, 1258.

SOLVENT STRUCTURE AND THE INFLUENCE OF ANIONS ON THE GELATION OF K-CARRAGEENAN AND ITS SYNERGISTIC INTERACTION WITH LOCUST BEAN GUM

David Oakenfull, Jeffrey Naden and Janet Paterson Food Science Australia, P.O. Box 52, North Ryde, NSW 2113 and Department of Food Science and Technology, University of New South Wales, Kensington, NSW 2033, Australia.

1 ABSTRACT

The synergistic interaction of K-carrageenan with locust bean gum has been studied in the presence of potassium chloride, nitrate, acetate, sulphate, citrate and EDTA. Gelation of K-carrageenanalone was only weakly dependent on the nature of the anion but the rupture strength (RS) of the mixed gel was more strongly influenced by different anions. RS increased linearly with the viscosity B-cuefficient, a measure of the structure-making (or structure-breaking) effect of the anion. This suggests that anions that best promote structure in the surrounding water molecules are most effective in stabilising the junction zones in the mixed gel network. 2 INTRODUCTION

The synergistic interaction of K-carrageenan and locust bean gum (LBG) has numerous food applications.' Addition of LBG modifies the textural characteristics of Kcarrageenan gels, reduces syneresis' and gives gels of equivalent strength with lower total concentrations of p~lysaccharide.~A likely mechanism of interaction is shown in Figure 1. LBG is a galactomannan with a galactowlmannose ratio of approximately 1:4.4 The galactose side chains are non-randomly distributed, leaving the mannan backbone with regions free from galactose residues. These 'smooth' regions of the galactomannan chain are believed' to interact with Kcarrageenan helices as shown in Figure 1 - or with aggregated K-carrageenan superhelical rods, as recently suggested by Piculell and his colleague^.^ Gelation of K-carrageenan proceeds via a two-step mechanism in which the polysaccharide first forms double helices. These then aggregate, forming superhelical rods which further associate and form a gel Its gelation properties are strongly influenced by electrolytes - a consequence of the fact that K-carrageenan is an ionic polysaccharide, with one sulphate group per disaccharide repeat unit (Figure 2). The stability of the double helices is sensitive to the ionic radius of the cation7**and aggregation of helices also has highly specific requirements. All the monovalent cations, for example bring about gelation if at high enough concentration but K+ is particularly effective because it specifically binds to the helix, but not the ~ o i l . ~ ~ ~ Gelation is also sensitive to the nature of the anion, but less strongly

222

Gums and Stabilisers for the Food Industry I0

Figure 1

Proposed mechanism of interaction between LBG and K-carrageenan helices.

Figure 2

Structure of the disaccharide repeat unit of K-carrageenan.

The influence of electrolytes on the interaction of K-carrageenan and LBG has received much less attention. The mixed gel has the same requirement for cations as Kcarrageenan alone,' but the effects of varying the anion appear not to have been systematically investigated previously. We report here an investigation of how the anions, nitrate, chloride, acetate, sulphate, ethylenediaminetetraacetate (EDTA) and citrate influence syneresis and gel strength (rupture strength) for mixtures of LBG and the potassium-ion form of Kcarrageenan.

3 MATERIALS AND METHODS Technical grade K-carrageenan and LBG were supplied by Germantown (Australia) Pty Ltd. Analytical grade potassium salts were used and all solutions were prepared in deionised water. Gels were formed in 200 ml cylindrical cans and held at 20°C for 20

Mixed Biopolymer System

223

hrs. All measurements were made at 20°C with a fixed concentration of polysaccharide (0.7% by weight). LBG was included at percentages of 10, 20, 40, 50, 60 and 80 of the total polysaccharide. The appropriate potassium salt was added such that the potassium-ion concentration was 0.1 mol dm'3. Syneresis was determined as a weight percentage from the weight of free liquid decanted from the can immediately after the storage period. Rupture strength was measured with a TA-XT2 Texture Analyser using a probe of diameter 12.5 mm inserted at a speed of 0.5 mm/s. 4 RESULTS

4.1 Syneresis Syneresis showed the expected decrease with increasing ratio of LBG to Kcarrageenan. At lower ratios of LBG to K-carrageenan, the extent of syneresis was also sensitive to the nature of the counter anion, as shown in Figure 3. Chloride and acetate produced significantly less syneresis than the other anions.

4.2 Rupture Strength The trend of rupture strength (RS) as the ratio of LBG increased showed the synergism typical of this system (Figure 4). RS was greatest for mixtures containing 40-50% LBG, much the same as reported for a 0.75% total polysaccharide system by Damasio et aL9 For each ratio of LBG to K-Carrageenan,RS was sensitive to the type of anion. In general, nitrate and chloride gave weaker gels than the other anions. 5 DISCUSSION

5.1 Influence of Anions on the Gelation of rr-carrageenan Although highly sensitive to the nature of the gelation of Kcarrageenan is only weakly dependent on the nature of the anion. Zabic and Aldrich" measured the RS of potassium K-carrageenan gels with different anions and found only small differences, with RS decreasing in the order CH3COO- > Br- > C1- > citrate > NO;. More extensive measurements have been made on the effects of anions on the helix-coil transition."*'* Norton and studied the temperature course of conformational ordering and found that the mid-point temperature, T,, varies systematically with the nature of the anion, following the Hofmeister (lyotropic) series. Watase and colleague^'^ measured the melting temperature of K-Carrageenan gels calorimetrically and found a similar trend. The Hofmeister series is a ranking of different ions originally based on the relative effectiveness of different salts in causing the precipitation of proteins and is followed by a wide variety of phenomena taking place in aqueous salt solutions.'4 For the anions we are considering here, the Hofmeister order is NO< < C1- < CH3COO< SO:- < citrate. In some circumstances, the Hofmeister series can be related to the effects of ions on the three-dimensional hydrogen bonded structure of waterlS - this is of particular interest because addition of simple sugars enhances the rigidity of Kcarrageenan gels (and increases T,) in a manner that can be quantitatively related to

224

Gums and Stabilisersfor the Food Industry 10

=

m cn,coo

CI

NO,

80

60

50

m SO,’

EDTA

40

20

a citrate

10

LEG (% wlw)

Figure 3

Syneresis (X w/w) afier 20 hrs at 20°C of mixed gels of K-carrageenan and LBG (0.7% total polysaccharide with addition of potassium salts (as indicated) at 0.1 mol dm3 K+.

m m m cr

NO;

80

Figure 4

CH,COO

60

Ezmn so,*

EDTA

50 40 20 LEG ( ’0 wlw)

Citrm

10

Rupture strength (N) afrer 20 hrs at 20°C of mixed gels of K-carrageenan and LBG with addition of potassium salts (as indicated) at 0.1 mole dm3 K+.

how the sugar influences the structure of water.I6

5.2 Effects of Ions on the Structure of Water - the Viscosity B-coefficient The viscosity E-coefficient provides a useful quantitative measure of the structure-making or structure breaking effect of ions which has recently been given a firm theoretical foundation.” The B-coefficient is derived from the Jones-Dole equation which relates viscosity to concentration in dilute salt solutions: Is

vc = vo (1

+A&

+

BC )

qc is the viscosity of a salt solution of concentration c , qo is the viscosity of water and A and B are constants. The A term represents the influence of interionic electrostatic

225

Mixed Biopolymer Systems 60 h

P)

v

0)

e0

55 A

-_----------

50 A

c

3

e n

45

--

.

/---

A

/---

rr

40

-0.10

0.00

0.10

0.30

0.20

B-coeff icient

Figure 5

Rupture strength (g) of potassium %-carrageenangels (I X w h ) with direrent anions plotted against B-coeflcient @vm zabic and Aldrich”).

80

o^

60

0,

E t-

40

20 -0.10

0.00

0.10

0.20

0.30

B coefficient Figure 6

Mid-point transition temperature (m) for %-carrageenan (Me& salt form)from Norton, et al.I2 and gel-sol transition temperatures (A) for Kcarrageenan (NH4’ salt form) from Watase et al. l3 plotted against the Bcoeflcient of the anion.

forces on viscosity and the B term represents the influence of effects of the ions on the structure of water.l’ The contribution from anions and cations is additive’’ and the two can be separated, giving Bcoefficients for individual ions. Ions with a positive Bcoefficient are structure-making and those with a negative Bcoefficient are structurebreaking.’’ In Figure 5 we show Zabik and Aldrich’sIo RS results plotted against the Bcoefficient of the anions. There appears to be a consistent trend - RS increasing with increasing Bcoefficient. This would be consistent with the increase in rigidity of Kcarrageenan gels seen on addition of structure-promoting sugars.’6 However, when the helix-coil transition is considered in isolation, and the mid-point transition temperature (T,,,) is plotted against &coefficient, the opposite trend is seen (Figure 6). T, decreases with increasing B-coefficient, implying that structure-making ions destabilise

226

Gums and Stabilisers for the Food Industry 10

8

h

2

v

6

5 m c

-L 4 v)

L3

c,

Q

3

L

2

0 -0.05 0.05

0.15

0.25

6-coefficient Fire 7

Viscosity B-coeflcient of the anion plotted against RS (N)for mixed gels of K-carrageenan and LBG. (+ 20%; 50%; A 60%and 80%LBG.)

the helix. Nonetheless, I- ions, which strongly promote helix formation, at the same time impede aggregation and gelation.’’ This difference can be explained by the fact that helix formation is dominated by electrostatic interactions” whereas aggregation of helices and gelation appear to rely more on hydrogen bonding.21 Hydrogen bonding between molecules in aqueous solution is promoted by structure-making c ~ s o l u t e s ; ~ ~ electrostatic interactions are dominated more by factors such as charge distribution and ionic radius.23 5.4 Effects of Anions on the Synergistic Interaction of w-carrageenan and LBG

In Figure 7 we show the correlation between RS and the viscosity B-coefficient for different ratios of K-carrageenan and LBG. In each case there was a linear relationship, with RS increasing with increasing value for the B-coefficient of the anion. This suggests that these anions influence the gelation process primarily via their effects on the structure of water. The slope of the lines shown in Figure 7 can be considered as the susceptibilty of the gelation process to modification of the solvent structure. Lines of best fit were estimated by regression analysis, and the slopes and standard errors are shown in Figure 8 plotted against the ratio of K-carrageenan to LBG. Interestingly, maximum susceptibility to solvent structure occurred at a ratio of K-carrageenan to LBG of about 1:1, coinciding with the ratio of K-carrageenan to LBG at maximum synergism (see Figure 4). At this ratio there was maximum interaction between %-carrageenanand LBG chains, as shown in Figure 1. This is again consistent with the view that anions

227

Mixed Bioplymer Systems

12 c

S

10

.-Q .-0

.c u-

Q 0

8

P

m rn >

6

v)

a u0 0

4

a

-0rn

2 0 0

20 40

60

80 100

% LBG Figure 8

Slope of RS vs viscosir>,B-coeflcient for direrent ratios of K-carrageenan to LBG.

promote the K-carrageenan-LBG synergism via their effects on the structure of water. The interaction between K-carrageenan and LBG primarily involves hydrogen bonds (LBG is a nonionic polysaccharide). As hydrogen bonding between molecules in aqueous solution is promoted by structure-making co-sol~tes,~~ structure-making (or structure-breaking) cosolutes will have greater influence on the mixed gel than on Kcarrageenan alone. The more structure-making the anion, the greater the increase in the rigidity of the mixed gel. This result suggests that other structure-making solutes (such as simple sugars) should also promote the K-carrageenan-LBG synergism. Fiszman and D ~ r a nhave ~ ~ recently shown that this is indeed the case for sucrose concentrations within the range 0-4096.

Acknowledgment We thank Germantown (Australia) Pty Ltd for their interest and support.

References 1.

G.H. Therkelssen, in ‘Industrial Gums. Polysaccharides and their Derivatives’, eds. R.L. Whistler and J.N. BeMiller, Academic Press, San Diego, 1993, p.

2. 3.

T. Turquois, C. Rochas and F.R. Taravel, Carbohydr. Polym., 1992, 17, 263. V. Carroll, M.J. Miles and V. Moms, in ‘Gums and Stabilisers for the Food

145.

228

4.

5.

6. 7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Gums and Stabilisersfor the Food Industry 10

Industry - 2’, eds. G.O. Phillips, D.J. Wedlock and P.A. Williams, Pergamon Press, Paris, 1984, p. 501. I.C.M. Dea and A. Momson, Adv. Carbohydr. Chem. and Biochem., 1968, 31, 241. L. Piculell, J Borgstrom, I.S. Chronakis, P.-0. Quist and C. Viebke, Int. J. Biol Macromol., 1997, 00, OOO. D. Oakenfull, CRC Cnt. Rev. Food Sci. Nutr., 1987, 26, 1. D. Oakenfull and A. Scott, in ‘Gums and Stabilisers for the Food Industry 5 ’ , eds. G.O. Phillips, D.J. Wedlock and P.A. Williams, IRL Press, Oxford, 1990, p. 507. I.T. Norton, in ‘Gums and Stabilisers for the Food Industry 5 ’ , eds. G.O. Phillips, D.J. Wedlock and P.A. Williams, IRL Press, Oxford, 1990, p. 511. N.H. Damasio, S.N. Fiszman, E. Costell and L. Duran, Food Hydrocoll., 1990, 3, 457. M.E. Zabic and P.J. Aldnch, J. Food Sci., 1968, 33, 795. K.R.J. Austen, D.M. Goodall and I.T. Norton, Biopolymers, 1986, 27, 139. I.T. Norton, E.R. Moms and D.A Rees, Curbohydr. Polym., 1984, 134, 89. M. Watase, K. Nishinari, P.A. Williams and G . 0 Phillips, Food Hydrocoll., 1990, 4, 227. J.W McBain, ‘Colloid Science’, Reinhold Publishing Corp, New York, 1950. W.P. Jencks, ‘Catalysis in Chemistry and Enzymology’, McGraw-Hill, New York, 1969. D. Oakenfull, in ‘Confectionery Science: Proceedings of an International Symposium’, G.R. Zeigler, ed., Pennsylvania State University, 1997, p. 67. Y. Marcus, J. Solution Chem., 1994, 23, 831. H.D.B. Jenkins and Y . Marcus, Chem. Rev., 1995, 95, 2695.14. I.S. Chronakis, L. Piculell and J. Borgstrom, Curbohydr. Polym., 1996, 31, 215. S. Nilsson and L. Piculell, Macromolecules, 1991, 24, 3804. R. Chandrasekaran, Adv. Food and Nutr. Res., 1998, 42, 131. J. Israelachvili and H. Wennerstrom, Nature, 1996, 379, 219. R.W. Gurney, ‘Ionic Processes in Solution’, Dover Publications, New York, 1962. S.M Fiszman and L. Duran, Food Hydrocoll., 1989, 3, 209.

HETEROTYPIC INTERACTIONS OF DEACETYLATED XA" WITH A GALACTOMANNAN OF HIGH GALACTOSE SUBSTITUTION DURING SYNERGISTIC GELATION

Francisco M. Goycoolea", Michel Milasb and Marguerite Rinaudob "Centro de Investigacion en Alimentacion y Desarrollo, A.C. P.O. Box 1735 Hermosillo, Sonora C.P. 83000 Mexico b

Centre de Recherches sur les Macromolecules Vegetales, C.N.R.S. affiliated with Joseph Fourier University B.P. 53 38041 Grenoble, Cedex 9 France

1 ABSTRACT Physical thermoreversible gels of deacetylated xanthan (DX) mixed with varying concentrations of galactomannan extracted from mesquite (Prosops spp.) seed endosperm (MSG) (M/G 1.1; Mw 2.1 x lo6) set and melt co-operatively at 23-27°C in 5 m M NaCI. The liquid-like character of the gels at 20°C (tan 620"~)and at the gelling temperature (tan add), attained their minima values when the concentration of MSG was -0.4-0.5 g L-l, while holding fixed that of DX at -1.0 g L-l. Mechanical elasticity (G), increased progressively as the proportion of MSG incorporated in the mixture increased from 0 up to -0.5 g L-1 and beyond this concentration, G values showed little hrther change with increasing MSG concentration; this initial behaviour is related to a progressive cross-linking of galactomannan and DX leading to establishment of a temporary gel network. Both values of onset temperature of gel formation (Tg) and that of DSC midpoint thermal transition (Tm), are in good agreement (T, -23.3"C and Tm -26.0"C) and vary slightly in mixtures within this range of composition. However, as the amount of unbound MSG increased for a MSG/DX weight ratio 2 0.6, T, values increased progressively up to -27.OoC, while T, also showed a break-point in the pattern of behaviour for setting and melting as a hnction of composition. For mixtures of MSG:DX weight ratio < -0.5, the interaction in the system has the characteristics of an heterotypic association process, resulting in the creation of a coupled gel network. The optimum stoichiometric ratio seems to involve a 1:l polymer chain pair (based on the contour length) of high galactose galactomannan and disordered DX species. Greater T, and Tm values, beyond the stoichiometric composition ratio may result from the increase in solvent viscosity due to unbound surplus galactomannan.

-

-

-

2 INTRODUCTION The saga of studies focussed on interactions in solution of xanthan (X), the industrial exopolysaccharide from Xuntomonas campesbis, with p-( 1++linked plant polysaccharides,namely galacto- and glucomannans, over more than twenty years, still has controversial interpretations and unanswered questions to otherwise well recognised physical phenomena, notably synergisticphysical macroscopic gelation and enhancements in solution viscosity even in very dilute concentration.5 Several proposals have been advanced over the past decades to explain the physical behaviour of these binary systems.6'22 The major issues of controversy have circled about

230

Gums and Stabilisersfor the Food Industry 10

the following issues: the general mechanism underlying the synergistic physical phenomena (i.e. whether it arises from binding or mutual exclusion of the involved polymeric species); the nature and geometry of the heterotypic junctions, if present; the involvement of the backbone and side chains of both structures in the heterotypic complex; the role of the secondary conformation and stability of the xanthan species; what is the driving force for the interaction, and whether different forms of molecular organisation can exist for the various P-(1+4)-linked plant polysaccharides known to interact with xanthan.23 Annable et al.24 argued that the interaction of konjac glucomannan with X only occurred after X molecules had undergone a coil-to-helix conformational transition, although the most disordered helical structure produced the strongest interaction. The early model of Moms et aL6to explain the synergistic gelation of galactomannans and helix-forming xanthan or algal polysaccharides in the agar series, proposed direct association of bare mannan backbone to the ordered helix surface. This proposal was later extended to account for synergistic gelation of mixtures of xanthan with galactomannans with high galactose content (IWE1.38) as a result of attachment of unsubstituted sides of regularly spaced galactose substitutes in the mannan chain to the xanthan helix.l3?l5It has generally been accepted, that the galactomanmdxanthan interaction leading to synergistic gelation, involves preferentially unsubstituted chains or sites in the galactomannan species. While in mixtures of xanthan with high-galactose galactomannans, such as guar gum, limited proportion of such unsubstituted regions leads only to increase in solution ~iscosity.2~ Deacetylated xanthan however, has been indicated as being able to interact with guar gum leading to gel formation at 25OC at optimum DWGM ratio -2.0,26 by contrast, no gel formation was observed in these studies between xanthan and guar gum. The onset temperature of gel formation and meiting in mixtures of xanthan or deacetylated xanthan and locust bean gum has been determined by sensitive small deformation rheology and found to remain close to -60°C in the presence or absence of salt.9.27 Differences from lower gel melting temperature values reported by other workers (Tg-49 OCl1 or Tg-40-53 OCl*) may lie within experimental errors of the different methods and instruments used. Besides, in none of these previous studies, a systematic rheological method to mark the critical gel point has been utilised. In a preceding work by Bresolin el a1.,28 synergistic gel formation as monitored by both sensitive small deformation rheology and micro-DSC was documented in mixtures of native xanthan and Schizolobium galactomannan (M/G 3.0). The mid-point transition of the observed DSC peak (Tm)and onset of gel formation (Tg) temperatures of the gelling mixtures were both centred at -25°C. The involvement of disordered xanthan in this heterotypic association was probed by an experiment showing that as neutralisation of H' xanthan proceeded in the absence of salt, thus effectively inducing the transition from the ordered (acid form) to the disordered (salt fom),29 the interaction with galactomannan (GM) grew progressively. Hence, if no increase of G values was observed in the absence of GM, in the presence of GM, the experiment clearly showed that G values increased monotonically as the neutralisation degree (governing directly the proportion of disordered xanthan) increased. This was consistent with changes in CD behaviour of xanthan in the presence of GM which proved that the complex is ordered. The central aim of this study was to confirm these previous interpretations, using a galactomannan sample of nearly full galactose substitution (IWE-1. 1) in combination with deacetylated xanthan. This entailed to testing the hypothesis of direct heterotypic association as the underlying mechanism governing the interaction between both polymers.

-

3 MATERIALS AND METHODS Xanthan was purified directly from the culture broth (Rhone Poulenc, Melle, France; batch

Mixed Biopolymer Systems

231

No. I8A X E 9701207) by carehl preservation of the native structure.3oInsoluble matter was removed by centrihgation for 2 h at 20,000 x g.The supernatant was filtered through 8, 3, 1.2, 0.8, and 0.45 pm nitrate cellulose membranes succesively. Deacetylation of xanthan was carried out by hydrolysis with NaOH (60 mM; 18 h at 4°C under nitrogen). Sodium chloride was added (1M) and the solution filtrated through 3.0 pm membranes. The sodium salt of deacetylated xanthan was precipitated by addition of ethanol (50% v/v) under stirring, washed successively with ethanovwater mixtures 80/20, 85/15, 90/10, 95/5 and 100/0 and left to dry under vacuum at room temperature. Galactomannans from Peruvian mesquite tree (Prosopispullida)were extracted from the endosperm splits of the seeds in a pilot plant operation and was a kind gift of Dr.Gaston C r u (U.of Piura, Peru). Mesquite seed galactomannan was dissolved in water (6 g.L-') and clarified by centrihgation 30 min. at 10,000 rpm (20,000 x g). The supernatant was diluted with one volume of water and filtered through 3.0 and 1.2 pm membranes. MSG precipitated from the supernatant with 36% (v/v) ethanol. Successive ethanovwater washings and drying were conducted as described above for deacetylated xanthan. The mann0se:galactose (M/G) ratio was 1.12, as determined by high-field proton NMR (300 MHz) spectroscopy in a Bruker AC-300.31 Weight-average molecular weight of MSG was assessed by steric exclusion chromatography using two columns (Shodex OK-Pak 804 and 805) arranged in series, with multidetection equipment: a multi-angle laser-light-scattering detector (Dawn, Wyatt), a home-made capillary viscometer and a differential refractometer as previously described.32 The M,was -2.1 x lo6. Mixed MSG-DX solutions for physicochemical studies were prepared from stock solutions by dissolving the individual dried polymers directly in 5 m M NaCl. Under these conditions the Tmof DX is -40.3 "C, as checked by DSC of a -8 g L' solution. Full dissolution of MSG stock solution was achieved by blowing hot air into the bottle from a hair drier (at -75°C). Rheological measurements were conducted using a cone-plate (cone angle: 1"; diameter: 60 mm; truncation gap: 27 pm) on a C d - M e d CS 50, stress-controlled rheometer, fitted with a Rheo lOOOC system. Temperature was regulated by a Peltier system, and in all the cooling or heating programs the rate of change applied was 1°C mid'. In order to prevent drying of the samples during experiments at high temperature, a metal solvent trap filled up with silicon oil was used. All measurements were carried out within the linear viscoelastic region, as checked by strain sweeps at 0 = 0 . 8 H z . Differential scanning calorimetry @SC) measurements were made using a Setaram microcalorimeter (Micro DSC III) equipped with 1 cm3 batch vessels. The cooling and heating scan rate varied in the range 0.2 to 0.7 "C min-' 4 RESULTS

Figure 1 shows the temperature course evolution of storage ( G ) and loss (G') moduli during cooling from 80 to 10 "C (1 "C min-') for mixtures of DX in the presence of increasing concentrations of MSG. Notice from the individual G traces that the onset of change in mechanical response, occurs at -25-27OC for the MSG-DX series. These changes include the characteristic features of a gel formation process (GW').The increase in G with decreasing temperature for all MSG-DX mixtures, takes place at substantially lower temperature than has previous1 been reported for LBG-DX (-SOOC) and KM-DX or KM-X (40°C)synergistic co-gelsJ-11 Indeed, no indications of a process occumng at greater temperatures are present. Temperature course changes in G" are coincident with those registered in G . Visual inspection of the formed MSG-DX mixed gels at 10 "C showed that at sufficiently high concentration of added MSG, the gels were self-supporting, though with a clear yield stress behaviour (i.e. large stress is needed before flow sets in, without indicationsof failure). This behaviour is diagnostic of long relaxation processes in the

Gums and Stabilisers for the Food Industry I0

232

their minima. This would be anticipated as the 'viscous' contribution of unbound galactomannan in the sol state, to the viscoelastic response, predominates over the elastic one, thus effectively inducing an increase in tan 6 values. This is consistent with the proposal of setting up a coupled gel network involving fixed amounts of both polymers in the heterotypic junctions (associative) and in the 'solubilizing' regions, connecting the network. The rheological behaviour described so far for the co-synergistic gels of galactomannan having a ratio M/G around 1 and deacetylated xanthan, argues in favour of a direct binding process, where a coupled gel network is cross-linked by heterotypic junctions. Qualitatively, similar composition dependence of tan 6 as shown in Figure 6, has been observed previously for the dependence of tan 6 of the formed gels of mixtures of DX and KM or LBG, and taken as diagnostic evidence of direct heterotypic binding of both types of polymer^.^ Indeed, experimental evidence from independent studies via different biophysical techniques is consistent with the binding model for xanthan and IU~l.249~~ The other aspect addressed in this investigation, was to establish the precise temperature dependence of the onset of incipient gel network formation (Tg) for the MSG-DX co-gel series as a function of the proportion of MSG, while holding constant the concentration of DX at 1.0 g L-'. T, values where marked from those of tan 6&tidderived according to the criteria described above. Figure 7 illustrates computed T, values for each MSG-DX mixture, describing a state diagram of the binary system. The boundary sol-gel line, clearly distinguishes two domains of behaviour. In the first domain, given by MSG concentration in the range 0.04 to -0.5 g L', T, remains constant at a value of -23.3 "C irrespective of concentration of MSG. At MSG concentrations beyond -0.5 g L', T, increases linearly with logMSG1 up to -27°C when [MSG] = 4.2 g L-'. Thermal changes recorded using DSC follow closely the observed rheological phenomena. Indeed, cooperative DSC endothems on heating and the corresponding exotherms on cooling were observed throughout the MSG-DX mixture series coincident respectively, with gel setting and melting processes. A typical DSC cooling peak has been illustrated in Figure 3. The envelope of the DSC transition was centred at the gel point throughout the mixture series. Figure 8 illustrates apparent midpoint transition temperature (Tm) values of DSC peaks recorded during heating and cooling scans (at 0.5 "C mid') as a function of MSG concentration. In close agreement with the composition dependence displayed by T, with log[MSG] shown in Figure 7, Tmvalues also describe two distinct domains of behaviour as a hnction of MSG concentration in the mixture, both during cooling or heating scans. In the region of low MSG concentration (0.04 to -0.6gL1), Tm values seem to show a mar 'nal decrease with MSG concentration. However, at MSG concentration > -0.6 g LY, Tm values increase with log[MSG], mirroring closely the 'break-point' pattern of behaviour recorded for T, values (Figure 7). Even when the overall magnitude of the changes in apparent T, pinpointing of this 'break-point' roughly at [MSG] 0.6 g L-', can be regarded as being virtually negligible, clearly the composition ratio coincides very closely with the onset value of increase in T, at [MSG] 0.5 g L-'. This narrow MSGDX ratio matches with the composition of maximum in gel strength (Figure 4) and minima in tan 6 (Figure 6). Clearly, the Tm values registered on DSC heating scans are consistently higher by -1.6"C throughout the MSG-DX composition series than those recorded on the cooling direction. This apparent hysteresis, however, was traced back to the expected thermal lag associated with the rather high rates of temperature change (results not shown). The composition dependence of transition enthalpy (AH) will be documented in a future paper.

-

-

Mixed Biopolymer Systems

233

10

G' (Pa)

0.04 g L-I

v

0 . l O g L'I

0.30g L-'

0.1 0.04

lot

G" (Pa)l 0.1 o'040

+

0

0.50 g L'l

0

0.90 g L"

v

1.41 g L-'

w

4.209 L-l

10 20 30 40 50 60 70 80

Temperature ("C) Figure 1 Evolution of (a) G ' and (b) G (a,= 0.8 Hz;20% strain) &ring coolingfor "

&ace@lated xanthan (I g L-' in 5 mM NaCg in the presence of mesquite seed gaIactomanmn at the concentrations (SL') shown in the label.

temporary molecular network. Increasing concentration of the MSG component, results in greater mechanical elasticity and a slight increase in the onset temperature of gel formation. The temperature dependence of loss %scous- component of the viscoelastic response, namely the G ' modulus, showed a small elevation coincident with the onset of elevation of G' values, within the range of MSG/DX ratio of 4.04 to 1.41 g L" and virtually no differences associated to gel composition. The mixture containing excess concentration of MSG (4.2 g L'),however, showed marked greater G ' values than the rest in the series, thus effectively indicating the predominance of the sol fraction on the viscoelastic properties. The dependence of the G , G ' and q* mechanical moduli on fiequency for four selected MSG-DX mixtures spanning the composition range explored is shown in Figure 2. These mechanical spectra, reveal that as the proportion of MSG in the mixture increases, there is a reinforcement of the 'weak gel'-like response of xanthan (Figures 2a and 2b), which is gradually replaced by formation of a true temporary viscoelastic gel network p e . with clear GX3" and low fiequency-dependence of G'). Indeed, at MSG 0.5 gL (Figure 2c), the mechanical spectrum shows little dependence of G and G ' on frequency, as well as a slope of $(a) close to -1. It is interesting to notice, however, that when the composition changes further in favour of an excess content of MSG (4.2 g L-') over DX (1.0 g L-')(Figure 2d), G and G ' moduli show a slight dependence on u). Changes in G', G" and tan 6 observed on reheating were closely superimposable with those observed on cooling. Figure 3 shows temperature course traces of G recorded on cooling and heating for the mixtures with MSG-DX ratio of 1:4.2 (g L'),along with a Changes in G , G ' and tan 6 observed on reheating were closely superimposable with those observed on cooling. Figure 3 shows temperature course traces of G recorded on cooling and heating for the mixtures with MSG-DX ratio of 1:4.2 (g L')), along with a

-

234

Gums and Stabilisersfor the Food Industry I0

lWca;l 10

0.01

0.1

1

7

Frequency (Hz) Figure 2 Frequency dependence (IOOC. 20% strain) ofG'(9, G"( 0)and I*' (a) for deacetylared xanthan (I g L*'in 5 mM NaCo in the presence of mesquite seed galactomannan at the concentrations (g L') of (a) 0.04, 0)0.I , (c) 0.5 d ( d ) 4.2

Changes in G'. G" and tan 6 observed on re-heating were closely superimposable with those observed on cooling. Figure 3 shows temperature course traces of G' recorded on cooling and heating for the mixtures with MSG-DX ratio of 1 :4.2(g L-I), along with a to the relatively high rates of cooling and heating used in the rheological tests (-1 "C min-I). Notice that the envelope of the DSC peak encompasses perfectly the breadth of the gelation process monitored fiom the evolution of G' with temperature. The shape of the DSC peak corresponds unequivocallyto that of a highly co-operative process (2-3OC) and it is very similar to the exotherms recorded in KM-DX synergistic co-gels.33~34The mechanical rigidity (G' at o = 0.2 Hz) of the gel networks formed at 2OoC was directly quantified fiom the cooling traces. The composition dependence of G' for the mixture series, is shown in Figure 4. The plot clearly shows that the elastic response increases linearly with increasing concentration of added MSG and attains a maximum value (G'-2.7 Pa) at a MSGDX ratio 0.5. As concentration of MSG species inaeaseS beyond this maximum, it seems to be a slight dudion in G' values. H o w , the overall pattern of behaviour is indicative of a saturation type process, in which MSG and DX form a mixed network involving fixed proportions of each polymer, as expected for a direct association process. It is interesting to point out that the G' values in the MSG-DX synergistic co-gels, are at least an order of magnitude lower than those reported for mixtures of deacetylated xanthan with LBG or konjac glucomannan observed before under similar composition in lOmM NaC1.9 The general pattern of behaviour, however, is comparable in MSG-DX, KM-DX or LBG-DX synergistic c o - g e l ~ , ~ though ~ , ~ ~the optimum composition ratio of glydxanthan for 'saturation' is definitely lower than in such systems (ie. 0.5 in MSG-DX vs. -1.0 in KM or LBG/DX co-gels). The criterion adopted to mark the critical gel condition (i.e. the percolation threshold at the formation of an incipient continuous network of infinite molecular weight), was defined as the point where a similar power-law variation of the dynamic mechanical moduli versus fiequency was observed. Formally, at the gel-point G(o) a G'(@) uA,thus irrespective of 0x33934

-

tan 6 = G ' I G =constant

Mixed Biopolymer Systems

235

10

E

G'(Pa)

Y

1

4.035 0

Q=

+

4.045

g

I

t

1

0.1 ).1.1.1.1.1.1.1.1.14.065

5

1

10 15 20 25 30 35 40 45 50

Temperature ('C) Figure 3 Temperature-course&pe&nce (o= 0.8 HZ20% strain) of G 'during cooling (V), a~d heating (A)for akacetyhtedxanthan (I g 1;') and messyuile seed galactomannun (4.2 g L') in 5 m M NaCl. The solid line, shows a DSC exolhenn recarded on the same solution at 0.5 "Cmin-'.

Galactomannan concentration (g C') Figure 4 Variationof G ' (m = 0.2 Hz 20% strain) modulusfor mixed gels ojakacetylaed xanthan (1 g L- in 5 mM NaCl at IO'C) in the presence of increasing concentration (g L9 of mesquite seed galactommaan By definition, at the critical gel point, the temperature traces of tan 6 registered at varying frequency, intersect at a single point, from which the onset of the gelling temperature (T& and the Critical tan6(tan6di) values were calculated. Gel-point

236

Gums and Stubilisersfor the Food Industry 10

analysis for three representative MSG-DX gelling mixtures, is illustrated below in Figure 5 . The liquid-like character (tan 6=G"/G') of the formed MSG-DX gels was quantified as a hnction of MSG concentration in two different ways: from the mechanical spectra recorded at -10 "C and at the critical gel condition. Figure 6 shows the dependence of tan 6 1 0 and ~ tan Gmua values as a hnction of MSG concentration. Clearly, it can be ~ attain appreciated, that the co-gel composition dependencies of tan 6 1 and~ tan~GmtiUl. well defined minima values in the vicinity of a low galactomannadDX ratio in the range -0.3-0.5,indicative of maximum network connectivity. As the mixture composition changes in favour of excess of MSG,a gradual increase in tan 6 values is observed beyond

u)

c (P

c

To=23.3'C

s

-5 0.0

10

15

20

To=25.0'C 25

30 10

Temperature ("C)

15

20

To=27.0%

25

30 10

Temperature ("C)

15

20

25

30

Temperature ("C)

Figure 5 Temperaturedependence of tan 6 (20% strain; at thefrequencies shown in 1abels)jw deacetyiatedxunthan (I g L-' in 5 m M NaCI) in combination with mesquite seedgahtomannan at concentrations (g L-') of (a) 0.5, (3) 1.4 and (c) 4.2.

Galactomannanconcentration(g L-') Figure 6 Variation in tan 6 (20% strain, w = 0.2 HI) at 20 "C (-+) and at the criticat gel point (+ ) for deacetylated xanthan (I g L-' in 5 mM NaCI) in the presence of increasing concentrations (g L- of mesquite seed gaiactomannan.

9

Mixed Biopolymer Systems

231

Galactomannan concentration (g L-') Figure 7 Variationof T, with concentrationof mesquite seedgahctomam in mixedgels with deocetylatedranthan (I g L 1in 5 mM NaCo

28*b2. . - - -0.1 --'

1

. ...

* 1

10

Galactomannan concentration (g L') Figure 8 Variation of mid-point DSC transilion (Td with concentrationof mesquite seed galactomantitin in mixedgels with IIX ( I g I;' in 5 nrM N d I ) b r i n g healing (A) rud cooling (V) at 0.5 "Cmin-'.

5 DISCUSSION

In recent studies,**,36the interaction of xanthan with galactomannansof varying degree of galactose substitution (M/G 1.1 and 3.0),has been addressed. These studies show experimental evidence consistent with the proposal of a new form of interaction between xanthan and galactomannansat lower temperature (T,-2S0C) than that previously reported in xanthan-low galactose galactomannan mixtures (i.e. T, 5O-6O0C), which seems to be dependent upon galactomannan M/G ratio. By contrast, with the thoroughly studied interaction between X and LBG at high temperature (Ts 40-60 "C), the new identified process has a well defined enthalpy and high cooperati~ity.~8,36 The strong dependence of AH on the salt content has been rationalised as direct involvement of disordered xanthan

-

-

-

238

Gums and Stabilisers for the Food Industry 10

segments in the associative interaction and indeed, this process does not exist for levels of NaCl above 10 mM, as the T,,,>> T,. The interaction recognised in the present study between MSG and DX, has all the features of that recently characterised in CERMAV,28,36 France, for several galactomannans of varying M/G ratio and xanthan, under conditions that favour its disordered conformation, namely removal acetyl groups, low ionic strength and high degree of neutralisation. The major features of this ‘novel’ kind of heterotypic interaction include: 1) Gel formation and melting take place at a narrow range of temperature (-25-27OC) as a highly co-operative process, and both rheological and DSC events have sharp transitions. 2) There is optimum co-gel composition, where rheological and thermal properties are maximised, well defined at a galactomannan-xanthanweight ratio - 0 . 5 . 3) Previous studies have demonstrated that temperatures of gel formation (TB) and midpoint sol-gel @SC) transition (T,,,), are independent of ionic strength, but the interaction exists only within a very narrow range (0 to l O m M NaCI). Under such conditions, DX disordered conformation is favoured, as previously demonstrated for xanthan-galactomannansystems.28 Chandrasekaran and Rhadha,37 based on X-ray diffraction patterns in guar-xanthan systems, have recently envisaged molecular models of heterotypic structures of xanthan and galactomannans, which can account, for the phenomena and interpretations offered addressed in this investigation. At least four sterically feasible molecular models for the xanthan:galactomannan complex have been proposed, which are appropriate for a galactomannan of any general side chain composition. One of these, predicts a segment of double heterotypic helical structure, involving one chain of xanthan and one of galactomannan. Indeed, in the MSG-DX binary mixed system, if the ratio of mass per unit length (ML) of each polysaccharide species is considered, namely MSG M~=60g/A and DX M~=90g/& hence a MSGDX ratio of 0.66 is obtained for a 1:1 contour length ratio of the two polymers. The ‘novel’ synergistic interaction addressed in this paper, seems to correspond closely to the one reported previously by other authors between high galactose LBG fiactions18or guar26 and xanthan or its deacetylated derivative. The evidence presented here, is consistent with the idea that this interaction involves association of high-galactose galactomannan species and disordered xanthan (i.e. arguing for involvement of its chain backbone), which becomes ordered when complexed. In many respects this heterotypic interaction resembles the far more studied one, between xanthan and konjac glucomannan (z.e. co-operative gel setting and melting processes, maximum synergism at a well-defined stoichiometric composition and large enthalpic changes associated). Lower thermal stability and weaker mechanical response in MSG-DX may be a consequence of development of a less dense network than that in KM or LBG synergistic co-gels, which in turn display greater stoichiometry in favour the glycan component. The interaction between galactomannans of high degree of galactose substitution and deacetylated xanthan or xanthan, now firmly demonstrated to occur at -23-27OC, seems to be directly related with a thermal and rheological process observed before in LBG-DX mixtures at -27”C, in 10 mM NaCL9 This issue will also be addressed in a future paper. 6 CONCLUSION This contribution presents fbrther experimental evidence consistent with the proposal of direct (stoichiometric) association of xanthan with galactomannans of high degree of substitution to form mixed gel networks under low ionic strength conditions. This

Mixed Biopolymer Systems

239

interaction is now firmly established to occur at -25OC, under conditions that favour the disordered codormation of xanthan. Acknowledgements We are grateful to CONACyT (Mexico) and CNRS (France) for financial support to this study. References 1. P. E. Jansson, L. Kenne and B. Lindberg, Curbohy&. Res., 1975,45,275. 2. L. D. Melton, L. Mindt, D. A. Rees and G. R. Sanderson, C u r b o w . Res., 1976,46, 245. 3. H. Bjorndal, C. G. Hellerquist, B. Lindberg and S . Svensson,Angew. Chem. Znt. Ed Engl., 1970,9,610. 4. K. P. Shatwell, 1. W. Sutherland, I. C. M. Deaand S. B. Ross-Murphy, Curbolyfr. Res., 1990,206,87. 5. F. M. Goycoolea, E. R. Moms and M. J. Gidley, C u r b o w . Po,.l' 1995,28,351. 6. E. R Moms, D. A. Rees, G. Young, M. D. Walkinshaw and A. Darke, J. Mol. Biol., 1977, 110, 1. 7. P. Cairns, M. J. Miles and V. J. Morris, Nuture, 1986,322,89. 8. P. Cairns, M. J. Miles, V. J. Moms and G. J. Brownsey, Carbol@. Res., 1987,160, 411. 9. F. M. Goycoolea, R. K. Richardson, E. R. Moms and M. J. Gidley, Mucromolecules, 1995,28,8308. 10. P. A. Williams, D. H. Day, M. J. Langdon, G. 0.Phillipsand K. Nishinari, Food H@ocoll., 1991,4489. 11. D. F. Zhan, M. J. Ridout, G. J. Brownsey and M. V. I., C u r b o w . Polym., 1993, 21,53. 12. I. C. M. Dea and A. Momson, A&. C u r b o w a t e Chem. Biochem..,1972,31,241. 13. I. C. M. Dea, E. R. Moms, D. A. Rees, E. I. Welsh, H. A. Barnesand J. Price, C u r b o w . Res., 1977,57,249. 14. B. V. McCleary, R. Amado, R. Waibel and H. Neukom, C u r b o w . Res., 1981, 92,269. 15. I. C. M. Dea, A. H. Clark and B. V. McCleary, Curbohyir. Res., 1986,147,275. 16. N. W. H. Cheetham, B. V.McCleary, G. Teng, F. Lum and Maryanto, C u r b o w . Polym., 1986,6,257. 17. R. 0.Mannion, C. D. Melia, B. Launay, G. Cuvelier, S.E. Hill, S.E. Harding and J. R. Mitchell, Curbohyir. Polym.., 1992,19,91. 18. L. Lundin and A. M. Hermansson, Curbo&&. Polym., 1995,26,129. 19. P. B. Fernandes, Biopolymers, 1995,24,269. 20. C. Schorsch, C. Gamier and J. L. Doublier, Znr. J. Biol. Mucromol., 1997,34, 165. 21. C. Gomes, M. P. Goncalves and P. A. Williams In 'Gums and Stabilisersfor the Food Industry 9', Williams P.A. and Phillips G.O. Eds.; The Royal Society of Chemistry, Cambridge, 1998, p. 239. 22. M. Rinaudo, N. Milas, T. Bresolin and J. Ganter ,PofjmerPreprinrs, 1998,39,680. 23. E. R Moms In 'Biopolymer Mixtures'; Harding S.E., Hill S.E.andMitchell J.R., Eds.; Nottingham University Press, Nottingham, 1995, p. 247. 24. P. Annable, P. A. Williams and K. Nishinari, Macromolecules, 1994,27,4202. 25. B. V. McCleary, Curbohydi. Res., 1979,71,205. 26. M. Tako and S . Nakamura, C u r b o w . Rex, 1985,138,207. 27. T. J. Foster Ph.D. Thesis, University of Cranfield, 1992.

240

Gums and Stabilisers for the Food Industry I0

28. T. M. B. Bresolin, M. Milas, M. Rinaudo and J. L. M. S. Ganter, Int. J. Biol. Macromol., 1998,23,263. 29. M. Rinaudo and M. Milas In 'Solution properties of Polysaccharides', Brant D.A., Ed.; American Chemical Society, Washington D.C., 1981, ACS Symposium Series, Vol. 150, p. 367. 30. M. Milasand M. Rinaudo, Carbohydr. Rex, 1986,158, 191. 31. J. L. M. S. Ganter, A.Heyraud, C. L.0.Petkowicz, M. Rinaudo and F. Reicher, Int. J. Biol. Macromol., 1995,17, 13. 32. B. Tinland, J. Mazet and M. Rinaudo, Makromol, Chem. Rapid Commun., 1988, 9,69. 33. F. Chambon and H. H. Winter, J. Rheol., 1987,31,683. 34. K. T. Nijenhuis and H. H. Winter, Macromolecules, 1989,22,411. 35. G. J. Brownsey, P. Cairns, M. J. Miles and V. J. Moms, Carbohydr. Rex, 1988,176, 329. 36. T. M. B. Bresolin, M. Milas, M. Rinaudo, F. Reicher and J. L. M. S. Ganter Znt. J Biol. Macromol. 1999, In press. 37. R. Chandrasekaran and A.Radha, Carbohydr. Polym., 1997,32,201.

High Solid Systems

HYDROCOLLOIDS IN LOW WATER AND HIGH SUGAR ENVIRONMENTS John R Mitchell Division of Food Sciences Univeristy of Nottingham Sutton Bonington Campus Loughborough LE 12 5 R D 1. SUMMARY

Different aspects of the behaviour of biopolymers at low water contents are discussed. The gelatin state diagram is used to illustrate the dependence of glass and melting transitions on water content. Because of its strong water content dependence measurement of the glass transition can be used to determine the amount of water associated with a component in a mixture. In a mixture of gelatin and waxy maize starch at low water contents, water is preferentially taken up by the gelatin lowering the mechanically measured glass transition temperature of this component relative to the starch. This is the predicted behaviour f h m the sorption isotherms. At high water contents addition of sugar stabilises ordered regions enhancing melting temperature and junction zone strength. In starch systems some evidence is presented to show that at low water contents andor under very slow heating conditions the normally expected increase in melting temperature with sugar addition is not seen. At very high sugar levels (-70%) there is evidence that sugar acts to destabilise ordered regions resulting in more deformable structures. Kasapis and coworkers have applied a WLF analysis to master curves obtained from frequency temperature superpositionto these high sugar polysaccharide systems. This suggests that the glass transition temperature at high sugar levels can be effected by the presence of small quantities of some hydrocolloids. This effect is far greater than would be expected from DSC measurements of Tg’ on polysaccharide sugar systems. Possible reasons for this difference are discussed. For starch it is clear that crispness loss occurs at much lower water contents than that correspondingto the glass transition. In the light of the work carried out on synthetic polymers this is not surprising since a brittle to ductile transition often occurs well within the glassy state. With the exception of starch and cellulose relatively little work has been carried out on the failure properties of “dry” hydrocolloid films. This is a fertile areas for hture research though it should be appreciated that such properties are strongly influence by small quantities of impurities. 2.INTRODUCTION To a large extent work on hydrocolloids has been concerned with their behaviour in high water content environments. There are however important applications in systems where the water content is low. The most important of these is low water content foods containing starch. This includes breakfast cereals, starch based snack foods and baked products. Confectionery products contain hydrocolloids at low water contents. In this case the major components are sugars. Examples

Gums and Stabilisersfor the Food lndustry 10

244

include gums and pastilles. The non-ice phase in frozen confectionery can also be regarded as a high sugar low water content system. There are numerous non-food uses for hydrocolloids at low water contents including biodegradable packaging. This paper covers the following aspects of the low water content behaviour of biopolymers. (a) The gelatin state diagram (b) Water partition in mixed biopolymer and biopolymer sugar systems. (c) The influence of sugar on the order disorder transition and the glass transition (d) Mechanical properties below the glass transition 3. THE GELATIN STATE DIAGRAM Gelatin is of particular interest since it can be prepared at high solids relatively easily compared with most gelling polysaccharides. Figure 1 displays the glass transition and melting temperature compiled from various sources by Marin I .

140 120

1

I

d

a

CI

2 P)

n

E P)

c

Figure 1. Some aspects of the gelatin state diagram. 3-Glass Transition line. 0Levine and Slade'; 0 -Marshall and Petrie3; A-Tseretely and Smirnova4; A- Marin' 3(B)- Gel-Sol Transition Line. Concentration varied by dehydrating a preformed gel at ambient. 0 -Marshall and Petrik, 0-Tseretely and Smirnova 4; 3(A)- Gel-Sol Transition Line. Gel prepared at indicated concentration. C Tseretely and ~ m i r n o v a ~ ; Marin'.

+-

High Solid Systems

245

This illustrates some factors which are seen for all biopolymer systems. Line 3 in Figure 1 shows the glass transition temperature. This is similar although at slightly lower temperatures to that found for most biopolymers. There is an increase in the melting temperature of the gel as the solids content increases. All biopolymer orderdisorder transitions move to high temperatures at lower water contents. For example this is seen for the denaturation of globular protein^',^ and for starch gelatinisation'. For starch the two explanations for this which can be taken from the synthetic polymer world are the dependence of equilibrium melting point on the level of diluent present predicted by Flory '-lo and the non-equilibrium argument of Slade and Levine which states that amorphous regions must be in the rubbery (nonglassy state) before melting of the orderedcrystalline regions can occur. ''-I2 As less diluent is available to plasticise these regions their glass transition temperature will rise. It is of interest that the melting temperature of the low water content gelatin gel obtained by dehydrating a gel prepared at high water contents (Curve 3B) is substantially higher than a gel prepared at that concentration (Curve 3A). Presumably at high concentrations the former contains a smaller number of more stable junction zones. On dehydration previously formed junction zones will act as nuclei around which fbrther growth occurs. 4. WATER PARTITIONING IN MIXTURES

Most low water content food systems contain a mixture of ingredients. Even if these are prepared by drylng from an initially homogenous solution, phase separation generally occur. Textural properties and in some cases microbiological and chemical stability will be influenced by the partition of water between the different components. It would be expected that water will be distributed primarily as predicted by the sorption isotherms for the individual components though interactions between the components cannot be entirely ignored. Sorption isotherms for gelatin and amylopectin (waxy maize starch) obtained in our laboratory are illustrated in Figure 2. At the higher end of the water content range shown these predict that water will be preferentially partitioned to the gelatin. To show experimentally how water is distributed it is necessary to have some water dependent property of the biopolymer or sugar that can be observed in the mixture. We previously used the water content dependence of the crystallinity of wheat starch as measured by X-ray diffraction to follow water partition in starch-egg albumen mixtures at low water co.ntents,I3 and the intensity of the amide I and I1 bands observed by direct difference infra red spectroscopy to demonstrate that the presence of sugar enhances the hydration rate of gelatin at low water contentsI4. Another approach is to use the glass transition temperature (TB). As the data in Figure 1 shows this is strongly dependent on water content at low water contents. A prerequisite for the use of this approach is that it must be possible to observe two T,s in the system.

246

Gums and Stabilisers for the Food Industry 10

45X

~

0%

Amylopectin (A) =Gelatin (G)

IOU

2QX

I

3QX

4QX

SOX

8QX

IOU

8QK

SQX

1QQX

Relative Humidity

Figure 2. Sorption isothermsfor amylopectin and gelatin

Mousia et alls measured the glass transition temperature of gelatin and waxy maize starch films using dynamic mechanical thermal analysis (DMTA). The films were prepared by extrusion in the water content range 20 to 40% wwb at different gelatin starch ratios and subsequently equilibrated to the appropriate relative humidity. Infra red microscopy showed that as expected a phase separated system was obtained with starch rich and gelatin rich domains with a typical domain size -50pm. Generally analysis of the viscoelastic properties by Dynamic Mechanical Thermal Analysis showed two peaks in the tan6 response which were associated with two glass transitions (Figure 3).

9.0

a5

1

0.9

0.8 a0 0.7

;

0.6

75 0.5

70

5

0.4

65

03 60

02 55

01

Figure 3. DMTA Thermograms of a gelatin/amylopectin/waterSO:S0:32 (w/w/w) system acquired at frequencies of IHz (solids symbols) and SHz (open symbols)".

High Solid Systems

247

From experiments at different ratios of gelatin to starch it was possible to assign these transitions to the gelatin rich and starch rich regions. At the same water content the glass transition temperature of gelatin is above that of starch. As can be seen from the data in Figure 2 as the relative humidity or water content increases water will increasingly be partitioned to the gelatin phase. A consequence of this is that the glass transition of gelatin increases above that of starch. (Figure 4). By combining sorption isotherms and the dependence of the glass transition temperature on the water content it is possible to predict the T,s for the mixed systems. This can only be done successfullyif uneven water partitioning is taken into account'5.

Figure 4. DMTA tan Gresultsfor the gelatin/amylopectin 25/75 extruded containing 35.75% water ( d s b ) (solid symbols) and stored subsequently at RH of 57% and P C to reach an equilibrium water content of 19.4% (open symbols) Is.

4.THE INFLUENCE OF SUGAR ON THE BIOPOLYMER TRANSITIONS 4.1 Order Disorder Transitions There have been extensive studies on the role of low molecular weight solutes on the order-disorder transition for biopolymers. In general the incorporation of sugars raises these transition temperatures. A good illustration of this is the work of Nishinari et a1 ' 6 * ' 7 ~ h showed i ~ h that the increase in temperature depended both on the concentration of sugar and the number of equatorial hydroxyl groups that the sugar possessed. Thus the increase in setting and melting point for agarose and carrageenan per molar concentration of added sugar was linearly related to the number of equatorial hydroxyl groups. This is discussed by Oakenfull elsewhere in this volume The increase in the gelatinisation temperature of sago starch with increasing sugar content is also discussed by Ahmad and Williams in this volume. In this

248

Gums and Stabilisers for the Food Industry 10

context it is interesting to note that a replotI8 of earlier data of L e l i e ~ r e ' ~ on ~ ' ' the effect of maltose on the gelatinisation temperature of wheat starch showed that data at different sugar levels superimposed provided the results were presented at a constant water content (Figure 5). In simple terms this means that the gelatinisation temperature of 50g of starch and 50g of water would be the same as 25g of sugar 25g of starch and 50g of water, though of course if sugar was added to the first formulation the gelatinisation temperature would increase.

380

370

1

0 0 0 0 (I)

0 360 -

E c

A

OA

360

340

3331

I

1

I

0

Figure 5. Replot of data of Lelie~re".'~ showing the efJects of water and maltose content on the temperature for the disappearance of the last birefringent starch granule under slow heating conditions''

The original data was obtained under conditions aimed at making measurements as close to equilibrium as possible (a heating rate of only 2OC hour) and measuring gelatinisation fiom the loss of birefringence of the most resistant granule using a polarising microscope. It therefore seems possible that under low water content equilibrium conditions sugar has less of an effect on the starch gelatinisatiodmelting temperature, than is the case when heating is carried out more rapidly and at high water contents. In any event it is obviously important to compare results at the same water content. In the low water content high solids regions the ability of sugar to enhance polymer hydration at very low water contents but dehydrate it at higher water contents as predicted by the sorption isotherms, may be relevant. A general interpretation for the effect of sugar on raising the temperature of the order-disorder transition at high water contents where non-equilibrium effects are

High Solid Systems

249

not important, is the exclusion of sugar from the surface of the polymer. An orderdisorder transition increases the contact area with the solute molecules and becomes less favourable if the solute is depleted from the surface. This relationship between exclusion of the cosolute from the polymer surface and the transition temperature was elegantly demonstrated by Nilsson et a12’ who compared the elution volume of different solutions on an agarose column with that of radiolabelled water. Solutes which were eluted before the water can be regarded as being excluded from the agarose surface and raised the helix coil transition temperature, whereas solutes which were eluted after water decreased the helix coil transition temperature and thus destabilised ordered structures. The ability of stabilising solutes such as sugars to increase the surface tension at the air water interface is another manifestation of there depletion from a surface, relative to their concentration in bulk solution. There is scope for combining the extensive information on the effect of cosolutes on protein denaturation and solubility (salting out) with the information available on starch and other hydrocolloids. 4.2 Glass Transitions

Whereas there has been extensive studies on the effect of water on the glass transition temperature of biopolymers and sugars alone and some studies on the plasticizing effect of sugar on biopolymers, systems containing high levels of sugar and low levels of biopolymer have largely been ignored. The exception is studies on frozen systems, where T,’ has been measured, generally by DSC but in some cases mechanically. This temperature is defined as the intersection of the glass transition line and the equilibrium ice melting line. For sucrose the concentration at this point is about 80%27(gsugar/lOOg of total system). No dramatic effects on the inclusion of biopolymers on the T,’ of sugars have been reported. Some data illustrating this is shown in Table 1. It should be appreciated that there is some controversy on the precise interpretation of DSC curves and the earlier T,’ values may be slightly too high due to non-equilibrium effects and ice crystallisation events, but this does not alter the conclusion from these studies that biopolymer inclusion does not alter T,’. It is interest to contrast this with the recent work of Kasapis and colleagues2326 who have claimed vitrification phenomena in high solids systems that are strongly dependent on biopolymer type. At high sugar levels the stabilising effect of sugar on junction zones is replaced by a weakening with a decrease in the enthalpy of melting resulting in a more entropic elastic network. This network is amenable to temperature frequency superposition of viscoelastic spectra obtained using rotational rheometry. The curves obtained are then analysed by free volume theory to obtain glass transition parameters. This is discussed in more detail elsewhere in this volume. Some data from earlier work is shown in Table 2. The fractional free volume (f,).predicted is as expected for a glass transition and for sugars alone the glass transition obtained by this approach is close to what would be expected from studies using differential scanning calorimetry. 2427

Gums and Stabilisers for the Food Industry 10

250

Table 1. Influence of Added Hydrocolloids on Sucrose Solution

Hydrocolloid (Concentration before freezing) None

Tg Y°C) -321L,-32L1,-34LL

XanthaniL (0.5%) Methyl Cellulose'L (1.O%) Dextran" (8.0%) Gud' (06%) GuaP

L

Tg' Obtained on Freezing a 20%

-30

-3 1 -32 -32

-34

Gelatin"L

-33

Carrageenan"

-34

(0.6%)

Table 2. Glass Transition Temperatureand Parameters in Free Volume Theory cfg is thefi-actionalfree volume at T,and q can be identifed with the diference between the thermal expansion coeficients above and below T$ Calculatedfrom Viscoelastic Spectra Obtained in a 85% Sugar System (Glucose Syrup with or without Sucrose)"

T, ("C) Q("C') fB

Deacylated Gellan -26 5.7 x 10' 0.028

High Methoxyl Pectin -53 5.9 x 100.029

K-Carrageenan -7 6.5 x 100.032

Whereas the T, of the system containing deacetylated gellan is not very different from what would be expected for the sugar alone, the high value for carrageenan and low value for pectin are surprising and interesting. Possible explanations for the difference between this data and the results on frozen systems are:- (a) Rheological measurements provide infomation about polymer mobility whereas DSC gives information about the mobility of the sugar. (b) There is a fundamental difference between the properties of a system that has been dehydrated by freezing and those that have been dehydrated at high temperatures prior to network formation. In the former case ice crystal formation will disrupt network structures and if these are deswelled from a high water concentration the resulting structure will be different from that prepared at lower water contents. (c) At high cosolute concentrations phase separation occurs in biopolymer sugar mixtures. This

25 1

High Solid Systems

phase composition depends on the preparation method. This would complicate any analysis of the data. Even if an interpretation of the viscoelastic spectra in terms of a glass transition proves to be inappropriate there is no doubt that the rheology of very high sugar systems depends strongly on the nature of the added polysaccharide and cannot be predicted h m knowledge obtained h m the extensive high water studies which have been carried out. As such the high solids area represents a relatively unexplored area where there is opportunity for innovation based on a greater scientific understanding. 5. MECHANICAL. PROPERTIES BELOW THE GLASS TRANSITION

The failure properties of biopolymers are almost certainly more important determinants of food texture than the small deformation parameters. Incorporation of very high sugar levels can increase the strain where fiacture occurs compared with systems containing low levels of sugar. This has been shown for agarose. 24 For starch containing systems it is clear that crispness loss occurs at water contents well within the glassy region (Figure 8) which at ambient temperature correspondsto about 22% moisture (wwb). 10

,

0

-

5

10 15 water content (% wb)

20

Figure 6. Relationship Between Perceived Crispness and Water Content of a Starch Containing System” Although attempts have been made to correlate a change in fiacture behaviour to the glass transition temperature there is no reason to expect this to occur. It is

recognised from the synthetic polymer literature that the brittle to ductile transition can occur well within the glassy region29.With the exception of starch and cellulose

252

Gums and Stabilisersfor the Food Industry 10

derivatives studies, on the ultimate properties of polysaccharide films at low water contents are very limited. This area is a fruitful one for further study, though in carrying out this work it should be appreciated that results are strongly influenced by the presence of low levels of impurities. 6. CONCLUSIONS

In the authors view the high sugar low water area is one where there is potential for further thought and experimental work which could lead in some cases to innovation. The specific areas where this should take place are the following:(a) It should be possible to generalise the effects of sugars on stability of ordered biopolymer structures by combining the extensive work which has been carried out on proteins with the more limited studies on polysaccharides. (b) The view that vitrification phenomena in high sugar environments is influenced by the presence of low levels of biopolymer appears to conflict with the work carried out on fiozen systems. This can be resolved by direct measurements of T, by different techniques (DSC, rheology) and mobility measurements using ESR and NMR. (c) It is suggested that for the important starch sugar system the generally accepted view that sugar incorporation raises the melting temperature may not be true at low water contents and very slow heating rates. This is simple to test by experiment. (d) As mentioned at the end of the last section the fracture behaviour of low water hydrocolloid systems is a fruitful area for research. 7.ACKNOWLEDGEMENTS Helpful discussions with Drs Genevieve Blond, Imad Farhat, Stefan Kasapis, Gaelle Roudaut and Professors Michele Marin and Ed Moms are gratefully acknowledged. 8. REFERENCES 1. M. Marin, Personal Communication, 1999 2. H. Levine and L. Slade, 'Collapse phenomena - A unifying concept for interpreting the behaviour of low moisture foods' pp 149-180 in "Food structure - its creation and evaluation" Mitchell J.R and Blanshard J.MV. Eds., Butterworths, London, 1987 3. A.S. Marshall. and S.E.B. Petrie, J. ofPhotographic Science, 1980,28, 128-134 4. G.I. Tseretely and 0.1. Smimova, J. of Thermal Analysis, 1992,38, 1189-1201 5 . B.Hagerda1 and H. Martens, J. Food Sci., 1976, 41,933 -937

6. Y . Fujita and Y . Noda, Bulletin of the Chemical Socity of Japan, 1981,54,32333234 7. J.W. Donovan, Biopolymers, 18,263-275

High Solid Systems

253

8. P.J. Flory, ‘Principles of Polymer Chemistry’, Cornell University Press, Ithaca, 1953 9. J. Lelievre, J. Applied Polymer Science, 1973, 18,293-296 10. J. Lelievre, Polymer, 1976, 17,854-858 11. L. Slade, and H. Levine, Carbohydrate Polymers , 1988,8, 183-208 12. L. Slade and H. Levine, Critical Reviews in Food Science and Nutrition, 1991, 30, 115-350 13. L. Hartley, F. Chevance, S.E. Hill, and J.R. Mitchell, Curb. Polym., 1995, 28, 83-89 14. LA. Farhat, S.Orset, P.Moreau and J.M.V. Blanshard J Colloid and Znt Sci, 1998, 207, 200-208. 15. Z. Mousia, LA. Farhat, J.F. Blachot,. and J.R. Mitchell, J.R., Polymer (in press) 16. K. Nishinari, M. Watase, K. Kohyama, N. Nishinari, D. Oakenfull, S.Koide, K. Ogino, P.A. Williams and G.O. Phillips, Polymer Journal, 1992,.24,871-877 17. K. Nishinari, M. Watase, E. Miyoshi, T. Takaya and D. Oakenfull, 1995, Food Technology,49 (October), 90-96. 18. J.M.V. Blanshard, ‘PhysicalAspects of Starch Gelatinisation’ in “Polysaccharides in Food”, Butterworths, London, 1979,139-152 19. I.A. Farhat, S.E. Hill, J.R. Mitchell, J.M.V. Blanshard, and J.F. Blachot, ‘IrreversibleChanges in Starch Processing’in “Water Management in the Design and Distribution of Quality Foods” ed. Roos, Y.H., Leslie, R.B. and Lillford, P.J., 1999 ~~411-428. 20. S. Nilsson, L. Piculell and M Malmsten, J. of Physical Chemistry, 1990,94,5 1495154 21. G. Blond, J. Food Eng., 1994, 22,253-269 22. H.D. Goff, K.B. Caldwell, D.W. Stanley and T.J. Maurice, J. of Dairy Science, 1993,76,1268-1277 23. S. Kasapis, ‘Structuralproperties of high solids biopolymer systems’ in “Functional Properties of Food Macromolecules” 2”d Edition Edited Hill, SE, Ledward ,DA and Mitchell, JR Aspen Publishers Inc, Maryland, 1998, pp 227-25 1. 24. A. Tsoga, S. Kasapis, R.K. Richardson, Biopolymers, 1999,49,267-275 25. V. Evageliou, S. Kasapis, and M.W.N. Hember, Polymer, 1998,39,3909-3917

254

Gums and Stabilisersfor the Food Industry 10

26. S. Kasapis, P. Giannouli, M.W.N. Hember, V. Evageliou, C. Poulard, B. TortBourgeois, and G. Sworn, Curb Polym. 1999,38, 145-154 27. Y. H. Roos, “Phase Transitions in Foods”. Academic Press Inc San Diegq, 1995 ~~116-117 28. G. Roudaut, C. Dacremont and M. LeMeste, J. of Texture Studies, 1998,29, 199213 29. A.J. Kinloch and R.J. Young, “Fracture Properties of Polymers” Elsevier Applied Science Publishers, Barking, England, 1993, p 230.

EFFECT OF SUGARS ON THE GELATINISATION AND RHEOLOGICAL PROPERTIES OF SAGO STARCH

F. B. Ahmad and P. A. Williams* Faculty of Resource Science & Technology, Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia. *Centre For Water Soluble Polymers, The North East Wales Institute, Plas Coch, Mold Road, Wrexham LL11 2AW, United Kingdom.

ABSTRACT The effect of various sugars on the gelatinisation and rheological properties of sago starch have been studied using differential scanning calorimetry and small deformation oscillation techniques. Sugars were found to increase the gelatinisation temperature, T, and gelatinisation enthalpy, AH in the following order: control (water alone) < ribose < fructose < glucose < maltose < sucrose. The effect of sugars on T, and AH has been attributed to the reduced plasticizing effect of the sugars compared to water, to sugarstarch interaction and to the effect of sugars on the water structure. The swelling and amylose leaching were reduced in the presence of sugars. Due to the low amylose content in the gelatinised starch samples, the storage modulus, G' was reduced significantly in the presence of sugars. Generally G' decreased in the following order : control (water alone) > hexose > disaccharides > pentoses. 1 INTRODUCTION

Sugars have a significant effect on the gelatinisation, gelation and retrogradation properties of starches. Earlier studies showed that the presence of sugars increased the gelatinisation temperature of starch and this phenomenon has been demonstrated by various methods such as light microscopy, differential scanning calorimetry, electron spin resonance etc. The effect of various su ars on gelatinisation depends on the concentration and type of sugar. Slade and Levine reported an increase in gelatinisation temperature with sugars in the following order: water alone < galactose < xylose < fructose < mannose < glucose < maltose < lactose < maltotriose < 10 DE maltodextrin < sucrose. The addition of sugars to starch also showed varying effects on the gelation and retrogradation of starch depending on the type and concentration of sugars used. Studies on wheat starch gels containing added sugar using large deformation mechanical testing showed that the rate of retrogradation was increased significantly by fructose, slightly increased by glucose and sucrose has no effect5. Slade and Levine4 showed that the extent of retrogradation of sugar-starch systems decreased in the following order: xylose > sucrose > maltose >glucose > galactose > water alone > mannose > fructose. They suggested that sugars act as antiplasticizers and reduce the rate of retrogradation. I'Anson

8

256

Gums and Stabilisers for the Food Industry 10

et a1.6using X-ray diffraction and rheological studies on wheat strach gels showed that retrogradation was reduced in the presence of sugars. Studies by Prokopowich and Biliaderis' on various starch samples showed that sugars have different effects on gelation properties, but ribose significantly retarded the retrogradation process in all starches they studied. Based on these studies, it is clear that sugars have various effects on the gelatinization, gelation and retrogradation of starches depending on the type of starch and sugar and their concentration. In this paper we report the effect of various sugars on the gelatinisation and gelation properties of sago starch.

2 MATERIALS AND METHODS

Sago starch used was a gift from Netsei Sago Industry, Sarawak, Malaysia. The sample was used as provided. The amylose content for the sample was 3 1% and the molecular mass for the amylose fraction was 1.24 x lo6 dalton. All sugars used were analytical grade purchased from Sigma Chemicals. 2.1 Differential Scanning Calorimetry (DSC)

DSC measurements were performed using a micro DSC (Setaram, Lyon, France). Starch samples were prepared on a dry weight basis and the pH of the suspension adjusted to 5.5 by adding dilute HCI or dilute NaOH. An aliquot of the samples (about 0.95 g) was added into the DSC measurement cell, the top sealed and the weight recorded An equal mass of solvent was added into the reference cell. The cells were heated from 10" C to 99" C at a rate of O.5"/minute. Sugar solutions at the required concentration (wt/vol) were prepared and the appropriate amount of starch was added into the sugar solutions and shaken thoroughly before being weighed into the DSC cell. Starch concentrations of 10% (wt/vol) and 50% (wt/vol) were used throughout. The initial gelatinisation temperature (To), the peak temperature which corresponds to gelatinisation temperature (T,,), the conclusion temperature (T,) and gelatinisation enthalpy AH were determined. Two measurements were performed for all the samples and results are given as an average. 2.2 Swelling Factor

Sugars at the required concentration (wthol) were prepared and starch was added to give a 0.5% starch suspension The suspension was heated for 30 min at 95" C and hot distilled water was added to compensate for evaporation. The swelling factor was calculated as the ratio of the weight of swollen starch granules to the weight of the dry starch. No correction for solubles was made due to the high sugar concentration. The results reported are the mean of triplicate measurements. 2.3 Extent of amylose leaching

Starch samples at 0.5% (wt/vol) in the presence of various concentrations of sugar were heated at 95" C for 30 min. Hot distilled water was added to compensate for

High Solid Systems

251

evaporation. The samples were cooled to room temperature and then centrifuged at 2200 rpm (using MSE Centaur 2 Centrifhge) for 20 min. The supernatant liquid was withdrawn and the amylose content in the liquid was determined by the method of Chrastil'. Triplicate measurements were performed and the average result is reported. 2.4 Mechanical Spectroscopy

The storage and loss moduli (G' and G") were monitored using a Carri-Med CSLSOO controlled stress rheometer (TA Instruments, Leatherhead, UK). For all the measurements 6% starch samples were used. The concentrations of sugar used were 550% (wthol). Sugar solutions (wt/vol) were prepared at the required concentration and starch was added. The pH of the starch suspensions was adjusted to pH 5.5 using dilute HCl or NaOH. The samples were heated in a water bath for 30 minutes at 95" C with constant stirring at 400 rpm. The beaker was reweighed and the volume corrected for evaporation loss on heating using hot distilled water. Hot samples were placed immediately into the rheometer measuring systems (cone and plate, 4 cm diameter and 2" angle) which was equilibrated to 25" C. All measurements were made within the linear viscoelastic region at a frequency of 1.O Hz with an amplitude of 1 milliradian over a 6 h period without disturbing the samples. A solvent trap was used to eliminate the evaporation of water. At the end of the experiment a frequency sweep was performed in the range 0.1 and 10.0 Hz.All measurements were performed twice and agreed within f 1%. 3 RESULTS AND DISCUSSION 3.1 Gelatinisation Properties

The effect of sugars on the gelatinisation parameters of sago starch is presented in Table 1 and DSC heating curves are given in Figure 1 for starch-fructose combinations. Similar curves were obtained for the other sugars. The figure clearly shows that To, T, and T, increased in the presence of sugar. The effect was more significant at higher sugar concentrations. For all the sugars studied, T, and AH increased with sugar concentration. The effect depended on the type of sugar. Generally T, increased in the following order: control (water only) < ribose < fructose < glucose < maltose <: sucrose. The effect became more significant with increasing sugar concentration. Similarly AH increased initially with increasing sugar concentration but at higher concentrations ( > 30% sugar), a plateau value was obtained. AH increased in the following order: control (water only) < ribose < fructose < glucose < maltose < sucrose. Similar observations were noted for 50% sago starch in the presence of sugar (data not shown). Sugars have been reported to increase T, and AH and the effect depends on the botanical origin of the starch and types of s ~ g a r ~ , ' ~Various ~ ' ' . explanations have been proposed to describe the effect of sugars on gelatinisation parameters such as competition between starch and sugar for available water which will effect the water activity, the reduced effectiveness of sugar molecules to act as plasticizing agents compared to water and specific sugar-starch interaction which will have an effect on the swelling and hydration of the starch granule^^.'^*'^. The increase in T, is greater for trisaccharides > disaccharides >monosaccharides2~14.'5.

Gums and Stabilisersfor the Food Industry 10

258

Table 1 Effect of Sugars on the Gelatinisation Properties of Sago Starch

Sugar %

0 5 10 20 30 40 50

Sucrose Tg AH "c Jlg 70.1 71.3 72.4 74.7 81.4 81.4 87.8

16.5 16.8 17.2 17.9 18.2 18.2 18.2

Maltose Tg AH "C Jlg 70.1 70.8 71.9 73.8 75.7 79.9 85.2

16.5 16.8 17.5 17.8 18.1 18.2 18.2

Fructose Tg AH "c Jlg 70.1 70.5 71.3 72.8 74.4 77.7 81.4

16.5 16.6 16.9 17.2 17.4 17.5 17.6

Glucose Tg AH "C Jlg 70.1 70.8 71.6 73.3 74.9 78.6 84.4

16.5 16.7 16.9 17.2 17.4 17.5 17.6

Ribose TR AH "c Jlg 70.1 70.1 70.1 70.1 70.6 72.1 73.2

16.5 16.5 16.7 17.1 17.2 17.3 17.3

Levine and Slade4 suggested that the delay in starch gelatinisation in the presence of sugar was due to their antiplasticizing effect. The reduced plasticizing effect of sugars compared to water is due to their increased hydrodynamic volume and lower translational and rotational mobility. Besides the antiplasticizing effect, specific starchsugar interactions which will effect the swelling and hydration of starch molecules will also have effect on the gelatinisation process. It has been demonstrated by "C-NMR that sugar-starch interaction occurs before the onset of gelatinisation process'6. Sucrose and maltose are likely to have a greater effect compared to fructose, glucose or ribose since they contain a greater number of hydroxyl groups. Sucrose and maltose therefore, can be expected to increase the gelatinisation temperature and gelatinisation enthalpy to a greater extent than other sugars used in this study. A good linear relationship was reported between T, and dynamic hydration number in the presence of sugar^".'^ which indicates that sugar - water interactions are also very important. 3.2 Swelling Factor and Amylose Leaching The effect of various sugars on the swelling factor and amylose leaching of sago starch are shown in Tables 2 and 3 respectively. The swelling factor increases slightly up to 20% sugar and then decreases. The effect was more pronounced in the order disaccharides > monosaccharides. Similar finding was reported by Cheer & LeIievrel9. At low sugar concentration the increase in the swelling factor can be attributed to the decrease in granule disintegration', while at higher sugar concentration the decrease in swelling factor might be due to an osmotic effect. The presence of sugar also reduced the proportion of amylose leaching. The effect depended on the type and concentration of sugar. Ribose had the greatest effect. The effectiveness of the various sugars to decrease amylose leaching is in the following order: ribose > sucrose > maltose > glucose > fructose. The effect of sugars on amylose leaching can most probably be attributed to the reduced plasticizing effect of the sugars compared to water due to increased hydrodynamic volume and also specific starch-sugar interactions as discussed earlier. The decrease in amylose leaching suggests that sugar molecules penetrate the starch granule and interact with amylose chains in the amorphous region. The degree of accessibility of sugar molecules in the amorphous region is influenced by the extent of granular swelling. The decrease in swelling factor in the presence of higher concentrations sugar can be attributed to a combination of several factors such as

High Solid Systems

259

competition between sugar and starch for available water, reduced accessibility of water molecules into the amorphous region of the granule, changes in water structure and specific sugar-starch interaction which will restrict starch hydration.

Tmpnturr I F

Figure 1 Eflect of Fructose on DSC endotherms for 10% Sago Starch. a. 0% Fructose; 6. 5% Fructose; c. 10% Fructose; d. 20% Fructose; e. 30% Fructose; F. 40% Fructose; G. 50% Fructose.

Table 2 Efect of Various Sugars on Swelling Properties of Sago Starch at 95' C. Sugar conc. YO 0 5 10 15 20 25 30 35 40

Swelling Factor, gg-' Sucrose

Maltose

Glucose

Fructose

Ribose

31.5 31.6 32.8 35.0 35.0 31.0 29.0 26.0 22.5

31.5 31.8 33.4 35.3 32.9 26.8 24.3 22.9 20.1

31.5 31.6 35.3 38.6 38.9 35.7 30.8 28.4 26.3

31.5 31.7 34.6 40.7 41.0 35.7 33.8 30.3 29.0

31.0 31.4 32.5 33.7 35.4 35.9 31.5 29.7 26.8

260

Gums and Stabilisersfor the Food Industry 10

3.3 Gelation Properties of Sago Starch

The effect of various sugars at 30% concentration on the storage modulus of 6% sago starch in presented in Figure 2. In all cases, G was found to increase rapidly initially and Table 3 EfSect of Sugars on Amylose Leaching During Gelatinisation,forSago Starches

Sugar conc. YO

Amylose leaching, %

0 5 10 15 20 25 30 35 40

350

Sucrose

Maltose

Glucose

Fructose

Ribose

27.5 26.0 23.3 22.0 20.6 18.7 17.2 15.2 15.0

27.5 26.3 23.5 21.2 19.0 17.5 15.9 15.0 14.8

27.5 26.6 24.2 24.3 27.1 22.3 21.0 18.5 18.5

27.5 26.7 24.4 24.4 27.2 24.5 23.3 21.0 18.0

27.5 24.5 20.5 18.0 17.2 14.5 13.4 13.0 12.1

**** *****

AFructose

300

. m n

-s

u)

S

250

U

g

200

Q

01

e0

150

z 100

50

0 0

5000

10000

15000

20000

25000

Time I Seconds

Figure 2 EfSect of various sugars at 30% concentration on the storage modulus of 6% sago starch.

High Solid Systems

26 1

attain a pseudoplateau value after a few hours. The pseudoplateau values were reduced in the order ribose > maltose > sucrose> fructose > glucose. The effect depended on sugar concentration. Higher sugar concentrations (data not shown) resulted in much greater reductions in the final G value. Similar observations were reported by Prokopowich and Biliaderis’ in their study on various starches and concluded this was due to the inhibition of chain aggregation by the sugar molecules. They found that ribose has the greatest effect followed by sucrose, glucose and fructose which is in good agreement with our finding. It is clear, however, that the main reason for the observed trend is the effect of the sugars on amylose leaching as noted from the values at 30% sugar given in Table 3. 4 CONCLUSIONS

The presence of sugars significantly affected the gelatinisation process and the effect depended on the type and concentration of sugars. The T, and AH increased in the folllowing order: control < ribose < fructose < glucose < maltose < sucrose. Sugar also was found to decrease the extent of granular swelling and amylose leaching. The extent of these changes were influenced by the type and concentration of sugars and can be attributed to combination of various factor such as starch-sugar interaction, changes in the water structure in the presence of sugars, competition between starch and sugar for the available water and the antiplasticizing properties of sugars relative to water. The reduced amylose leaching in the presence of sugars significantly effected the gelation of sago starch. References E.E. Hester, A.M. Brian and C.J. Personius, Cereal Chem., 1956, 33,91. M.M. Bean and E.M. Osman, FoodRes., 1959,24,665. B.S. Miller and H.B. Trimbo, Food Technol., 1965, 19,640. L. Slade and H. Levine, Cric. Rev. in FoodSci. andNutrition, 1991, 30, 1 15. J.L. Maxwell and H.F. Zobel, Cereal Food World, 1978, 23, 124. K.J. I’Anson, M.J. Miles, V.J. Morris, L.S. Besford, D.A. Jarvis and R.A. Marsh, J. CerealSci., 1990, 11, 243. 7. D.J. Prokopowich and C.G. Biliaderis, FoodChem., 1995,52, 255. 8. J. Chrastil, Carbohydr. Res., 1987, 159, 154. 9. I.D. Evans and D.R. Haisman, Starch, 1984,34, 224. 10. K. Kohyama and A. Nishinari, J. Agric. FoodChem., 1991,39, 1406. 1 1. J.M. Johnson, E.A. Davis and J. Gordon, Cereal Chem., 1990,67,286. 12. R.I. Derby, B.S. Miller and H.B. Trimbo, Cereal Chem., 1975, 52, 702. 13. J.W. Donavan, Biopolymer, 1979, 18, 263. 14. H.L. Savage and E.M. Osman, Cereal Chem., 1978,55,447. 15. R.D. Spies and R.C. Hoseney, Cereal Chem., 1982,59, 128. 16. L.M. Hansen, C.S. Setser and J.V. Paukstelis, Cereal Chem., 1989, 66,41 I . 17. F.B. Ahmad and P.A. Williams. Biopolymers, 1999, 50,401. 18. L. Slade and H.L. Levine, Cric. Rev. FoodSci. Nutrit., 1991, 30, 322. 19. R.L. Cheer and J. Lelievre, J. Appl. Polym. Sci., 1983, 28, 1829.

1. 2. 3. 4. 5. 6.

THE RHEOLOGICAL PROPERTIES AND ENZYMATIC DIGESTIBILITY OF AMYLOSE AND AMYLOPECTIN GELS IN THE PRESENCE OF MALTITOL

E. Vesterinen, P. Forssell, P. Myllihinen and K. Autio VTT Biotechnology and Food Research P.O. Box 1500 FIN-02044 VTT, Finland

1 INTRODUCTION

Polyols are used in the con-,ctionery in-Jstry instead of saccharose in order to reduce the energy content of sweets and prevent dental caries. The use of polyols with starch can hinder their beneficial effects, if starch is simultaneously hydrolyzed to sugars by salivary a-amylase in the human mouth. Caries is a tooth disease that occurs as a result of a complex mechanism in which oral bacteria, especially Streptococcus mutuns, and fermentable carbohydrates play important roles. Gelatinized starch is rapidly hydrolyzed in the mouth by salivary a-amylase to maltose, maltotriose and dextrins, which are more diffusible into plaque than large intact starch molecules. When oral bacteria come into contact with food that contains fermentable carbohydrates, they convert them into organic acids and lower the pH. A profound fall in plaque pH after starch consumption has been demonstrated by different re~earch.'.~ In the mouth, food is submitted to several actions: lubrication with saliva, reduction to small particles and starch hydrolysis. Chewing time determines the extent of these three actions. The rate of digestion of starchy foods differs and starch hydrolysis is dependent on the initial structure of the food. It is also known that processing of starchy foods, including heat treatment, extrusion cooking, drum-drying and popping, make the product more susceptible to fermentation in saliva and dental In addition, starch stickiness and adherence to the tooth surface as well as the oral retention time are important factors when considering the cariogenic potential of different starchy f00ds.l'~ Starch granules are composed of two polysaccharides, amylose and amylopectin. The ratio of these two components varies according to the source plant. Amylose is mainly a linear component consisting of about 100-10 000 glucose monomers linked by 1-4 abindings. Amylopectin is highly branched with a-(1,4) and a-(1,6) linkages. The most widely accepted model for the arrangement of the amylopectin molecule is the cluster model in which short chains are multi-branched on longer chains which are linked together. Above all, amylopectin has the same chemical structure as amylose, except for the bran~hing.~.'.~ Starch in, for example, cereal grains or potatoes, is accumulated in granules of various sizes and shapes. These granules are semi-crystalline particles and are insoluble in water at room temperature. When starch is heated in an aqueous solution, the granules are hydrated and swollen. At a certain temperature, the starch granule loses its crystallinity, and at another temperature the anisotropic order disappears, which can be seen as a loss of

High Solid Systems

263

birefringence. Raw granules are degraded very slowly by amylase and gelatinization increases the availability for hydrolysis by amylase.'.' At cooling and storage the molecules in the gelatinized starch reassociate into an ordered structure and under favourable conditions may partly recrystallize. This phenomenon is referred to as retrogradation, and it reduces the availability of the starch for amylase. The degree of retrogradation depends on the food source and starches with a high amylose content display more retr~gradation."~.~ Gel formation during cooling and increase of rigidity during storage consist of two separate reactions with different kinetics. The short-term changes are related to the gel formation of solubilized amylose in the continuous phase. The long-term hardening is attributed to crystallization of the short chains of amylopectin within the The aim of this work was to study the effects of maltitol syrup on the rheological properties and enzymatic digestibility of amylose and amylopectin gels. Starch components gelatinized with various amounts of maltitol syrup were studied using rheological techniques and enzymatic hydrolysis. Gel formation properties during the cooling process were measured using a small deformation dynamic test and enzymatic digestibility of the cooled gels was determined using an in v i m model.

2 MATERIALS AND METHODS

2.1 Materials The components of starch, amylose and amylopectin, were studied separately. Extracted potato amylose purchased from Sigma Chemical Co., was used as a model component of amylose and the waxy-maize (Amioca), used as a model component of amylopectin, was a product of the National Starch company. Maltitol syrup was kindly donated by Huhtam& Leaf (Turku, Finland). The salivary a-amylase from Sigma Chemical Co. was used in a concentration of 1 U/pl. Water used in the solutions and measurements was deionized in an Elgastat Prima Reverse Osmosis unit combined with an Elgastat Maxima Analytical purification system.

2.2 Preparation of Samples Aqueous amylose gels (3-8% w/w) were prepared by cooking powdered amylose with deionized water in a special pressure vessel (VTT/Automation, Espoo, Finland) using a mixing blade under elevated pressure in order to the obtain the desired temperature in the range of 144-150°C. High-amylopectin gels (3-10% w/w) were prepared using waxymaize (Amioca). Amioca slumes were cooked in a boiling water bath for 10 min under atmospheric pressure or under elevated pressure like amylose gels at 120°C. Gels, containing maltitol, were prepared by adding maltitol syrup into the amylose or amylopectin slumes before cooking. The cooked solutions were collected in cylindrical tubes, which were sealed carefully, or in a preheated thermos depending on the intended usage of the gelatinized solutions.

264

Gums and Stabilisers for the Food Industy I0

2.3 Rheological Measurements Rheological measurements were camed out using a StressTech rheometer (ReoLogica Instruments AB, Lund, Sweden) with a concentric cylinder measuring geometry (CC 25 CCE). The oscillation frequency was fixed at 1 Hz and the rheological parameters were measured as a function of stress amplitude. This enabled the determination of the linear region, where the complex modulus and storage modulus are independent of applied strain at any frequency. The storage moduli (G')of the gels were measured using a fixed stress of 1 Pa and a frequency of 1 Hz.The rheometer was preheated to 90°C and after the gelatinized solution was stabilized for 5 min the oscillation measurement was started at a cooling rate of lS"C/min. The viscosities of the polymer solutions were measured as a function of shear rate at 90°C. The applied shear rate ranged from 0.1 to 600 Us.

2.4 Hydrolysis Measurements After amylose and amylopectin gels were stored for 2 days at room temperature, they were milled using an OCI Instruments Omni Mixer 17106 for 2 min. 25 ml of 0.34 wt.% NaCl solution were added into the milled sample and the pH of the sample solution was adjusted to 6.0-6.5. The hydrolysis samples were then stirred at 250 rpm and incubated in a water bath at 37°C and the hydrolysis started by adding 50 U of salivary a-amylase.'* After 5 min, a few drops of 2 M HCI were added to stop the hydrolysis. Samples were centrifuged at 14 000 rpm for 10 min. The total carbohydrate was measured in the supernatant solution using the phenolic-sulphuric acid method. The amount of reducing sugars was determined by staining them with dinitrosalicylic acid and measuring them spectrophotometrically .

3 RESULTS AND DISCUSSION

3.1 Rheological Properties Figure 1 shows the storage moduli of amylose gels of various concentrations. The G' values of all amylose samples first rised sharply upon cooling before reaching a plateaulike behaviour. All gelatinized amylose solutions seem to form a strong network at 25°C and the resulting gel strengths were dependent on amylose concentration. The rate of network formation was obviously also dependent on the amylose concentration. The association of polymer chains, and therefore, the onset of amylose gelation is faster in a concentrated solution than in a dilute solution. The level of storage modulus G' describes the number of junctions between polymer chains. The formation of networks of 5% amylose gels, with various maltitol concentrations, were studied. The G values of 5% amylose gels were at the same level despite the presence of maltitol indicating that the addition of maltitol does not prevent the association of polymer chains and the formation of cross-links. Actually a considerable maltitol content (>20%) increases the rate of gelling due to the increased viscosity of amylose pastes. The effect of maltitol on the storage modulus of the 5% amylose gels as a function of shear stress is shown in Figure 2. It can be seen that the addition of maltitol increases resistance to shear stress and these amylose gels became more elastic when maltitol was present.

265

High Solid Systems

loo00 loo0 n

0

loo

b

10

e

+-3 YOarnylose gel --b 5 %

a8 % arnylose gel

1

0,Ol

arnylose gel

4 0

1 20 2000

4ooo

6ooo

8OOO

loo00

12000

14000

Time (s)

Figure 1 Development of storage modulus G'ofgels with various amylose concentrations during the cooling process.

loo0

100 n

Q

10

+10 % rnaltitol -e- 20 % rnaltitol

0,Ol

' 0

+40 % rnaltitol 20

40 Shear stress (Pa)

60

80

Figure 2 Effect of maltitol on the storage modulus of 5% amylose gels as a function of shear stress. Rheological behaviour of amylopectin solutions upon cooling was clearly different from that of amylose solutions. It has been stated that amylopectin forms gels only at quite a high concentration, greater than 10%(w/w), and that gelation takes a much longer time. Gelation occurs only above a critical gelation concentration. At a fixed molecular weight, the branching of polymer chains reduces the hydrodynamic volume with the result that the critical concentration is shifted towards higher values than in a linear chain.6

266

Gums and Srabilisers for the Food Industry 10

The dependence of storage and loss moduli on applied frequency for a 5% amylopectin solution at 90°C was studied. The storage modulus was greater than the loss modulus at all frequencies, and G' slightly increased with the increased frequency being of the order of 10 Pa. Both moduli were measured immediately after the cooling process (lS"C/min) at 25°C and no changes were detected. However, it was shown that the storage modulus increased slightly during the cooling process indicating an association between polymer chains in dilute amylopectin solutions over a very short time period (Fig. 3). From these rheological properties it can be concluded that even dilute amylopectin solutions ( 4 0 % ) reach the critical concentration after heat treatment. It was also found that the values of the storage modulus increased with increasing amylopectin concentration, but not as distinctively as in amylose gels.

14 12 10

o 10ninatW"C

El3

b 6

lOmnat90"C

4

A under pressure at

2

0 0

1000

2000

3000

4000

5000

6000

7000

Time ( 8 )

Figure 3 Effect of heating conditions on the G'of5% amylopectin gels. It has been observed that the preparation temperature of high-amylose starches has an effect on the gelation behaviour as well as on the final properties of the starch gels. In this study also, amylopectin was found to be very sensitive to preparation temperature and time (Fig. 3). The storage moduli of amylopectin gels, treated at elevated temperature (-120°C) or longer time at 90"C, indicated lower values than gels cooked 5 min at 90°C. This phenomenon could be due to the different solubility of amylopectin.

3.2 Enzymatic hydrolysis of gels Figure 4 shows the extent of hydrolysis measured as reducing sugars of amylose gels and powder as a function of the amylose content. Hydrolysis of aqueous starch systems starts with the diffusion of the enzyme towards the solid-liquid interface. The rate of hydrolysis is essentially dependent on the porosity and the accessibility of the substrate. In this study the extent of hydrolysis was defined as the amount of reducing sugar or soluble carbohydrate present. The extent of hydrolysis was found to increase with the amount of a-amylase and hydrolysis time. Furthermore, the extent of hydrolysis naturally decreases with an

High Solid Systems

267

increase in the ratio of amylose to a-amylase. The highest amount of reducing sugar was detected by hydrolysing extracted amylose powder. The low amount of reducing sugars at high amylose powder concentration may be either due to low a-amylase activityhnylose or to limited amount of free water. This poor accessibility of polymer chains in networks can be seen clearly in Figure 4. The extent of hydrolysis decreases with an increase in the amylose gel concentration. By increasing the amylose concentration the network formed is denser, and therefore, the pore size decreases, thus making it more difficult for the enzyme to diffuse into the polymer network. Leloup et al. have studied amylose gels and they have defined the mean pore size of amylose gels decreased from 4 to 40 nm as the polymer concentration decreased from 10 to 3% (w/w)."

&

45 40

a

o Arnylose powder

35-

2

8 a

c-

30 25

o 5 % Amylore gel

-

:;: p m

A 8 % h y l O S e gel

i

10-

5 0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

AmYlOre (9)

Figure 4 Effect of amylose concentration on salivary a-amylase hydrolysis of amylase powder and gels. Figure 5 illustrates the amount of reducing sugar after 5 min hydrolysis of 5% amylose gels in the presence of maltitol and, as can be seem, the amount of reducing sugar decreases with the presence of maltitol. These results can also be explained by the diffusion phenomenon, i.e. increasing the polymer concentration or adding sticky maltitol syrup into the network makes it more difficult for the enzyme to diffuse into the polymer network and hydrolyze linkages between the glucose units.

268

Gums and Sfabilisersfor fhe Food Industry 10

0 %maltitol A 10 % maltitol 4 40 % maltitol

=

5 0

0

0,1

02

04

0.3

Amylo=

0.5

0,6

(s)

Figure 5 Effect of maltitol on the extent of hydrolysis of 5% amylose gels. In addition, the extent of hydrolysis of pure amylose and amylopectin gels was measured as the total amount of soluble carbohydrates. The results of these measurements were set against the hydrolysis results of native Amioca and amylose and Amioca powder, and presented in Figure 6. In these measurements the ratio was 0.35 g starch /50 U a-amylase. 100 A

90 80

soluble carbohydrates after amylase treatment

u)

3 70

E

r o n

60

50

3

40

3

30 20

3

v)

0soluble carbohydrates without amylase

10

0 Natiw Amioca

Amylose powder

Amioca powder

5 Yo Amylose gel

5 Yo Amioca gel

Figure 6 The amount of soluble carbohydrates a f e r a-amylase hydrolysis of amylose and amylopectin solutions and gels. Amount of starch 0.35 g/50 U a-amylase. Native Amioca granules were degraded the most slowly by a-amylase and without amylase treatment no soluble carbohydrates were detected. Furthermore, extracted amylose powder that did not receive amylase treatment contained over 20% soluble carbohydrates which were decreased in 5% amylose gel indicating that extracted amylose powder probably had soluble polymer chains which associated under the cooling process and became less soluble during gelatinization. The amount of soluble carbohydrates after

High Solid Systems

269

amylase treatment of branched amylopectin molecules was clearly higher than the amount of linear molecules. In both components, network formation decreased the amount of soluble carbohydrates, and moreover, in the case of amylose the amount of carbohydrates decreased slightly more. 4 CONCLUSIONS Amylose formed very firm gels even in dilute concentrations. The addition of maltitol syrup made the gels more elastic and resistant to shear stress. Further, the addition of maltitol increased the viscosities of hot amylose and also of amylopectin dispersions. The gel formation of amylopectin was clearly different from amylose. The gelation concentration of amylopectin was higher and gelation took longer time. The viscosities of hot gelatinized amylopectin pastes were higher than in amylose pastes and it seemed that even dilute solutions of amylopectin (5-10%)reached the critical gelation concentration and formed weak gels. The storage modulus values of these weak amylopectin gels were of the order of 10 Pa. The extent of hydrolysis increased with a higher amount of a-amylase and hydrolysis time. In the case of firm amylose gels it was found that the extent of hydrolysis decreased with increasing amylose concentration in the network. Moreover, the addition of maltitol decreased the extent of hydrolysis. The amount of soluble carbohydrates after hydrolysis treatment was highest for amylopectin gel and powder. The observed differences between these two different polymer gel systems are probably caused by different crystallinity of StNCtUR.6

References 1. 2. 3. 4.

P. Lingstrom, D. Birkhed, Y. Granfeldt and I. Bjorck, Curies Res., 1993,27, 394. P. Lingstrom, J. Holm, D. Birkhed and I. Bjorck, Scund. J. Dent. Res., 1989,97,392. S. Kashket, T. Yaskell and J. E. Murphy, Curies Res., 1994,28 (4),291. I. Bjorck, Carbohydrate in Food, ed. Ann-Charlotta Eliasson, New York, 1996 5. K. Svegmark, PhD Thesis, Chalmers University of Technology, 1992. 6. S. G. Ring, P. Colonna, K. J. I'Anson, M. T. Kalichevsky, M. J. Miles, V. J. Moms and P. D. Orford, Carbohydrate Res., 1987,162,277. 7. N.-G. Asp, I. Bjorck, J. Holm, M. Nyman and M. Siljestrom, Scund. J. Gustroenterol. Suppl., 1987, 129,29. 8. J.-I. Jane, Starch Structure anf Functionality, ed. P J. Frazier, A. M. Donald, P. Richmond, Cambridge, 1997 9. P. Cairns, V. J. Moms, R. L. Botham and S . G. Ring, J. Cereal Sci., 1996,23,265. 10. S . E. Case, T. Capitani, J. K. Whaley, Y. C. Shi, P. Trzasko, R. Jeffcoat and H. B. Goldfarb, J. Cereal Sci., 1998,27,301. 11. V. M. Leloup, P. Colonna, S . G. Ring, K. Roberts and B. Wells, Carbohydrate Polymers, 1992,18, 189. 12. A.-M. Aura, H. HhkBnen, M. Fabritius and K. Poutanen, J. Cereal Sci., 1999, 29, 139. 13. P. Parovuori, R. Manelius, T. Suortti, E. Bertoft and K. Autio, Food Hydrocolloids, 1997,ll (4), 471.

“IN VZVO” AND “IN VZTRO” ANNEALING OF STARCHES

StCphane J.J. Debon and Richard F.Tester School of Biological and Biomedical Sciences Food Research Laboratories Glasgow Caledonian University Glasgow G4 OBA, UK

1. Introduction Starch granules are synthesised in amyloplasts and stored in the major depots of seeds, tubers and roots. In this granular form, starch is semi-crystalline, cold waterinsoluble, dense and constitutes an almost universal form for packaging and storing reserve energy as carbohydrates in green plants’.’. Starch polysaccharides occur as two major high molecular weight a-glucan biopolymers; amylopectin (AP) and amylose (AM). The AP fraction contains small linear (1+4)-a-~-glucopyranose chains, with 20-25 glucose residues on average, joined by numerous (1+6)-a-~-glucopyranose linkages (amounting to 4-5% of the total linkages), to form a highly branched macromolecule. The AM fraction is an (1+4)-a-~-glucopyranan; it has the properties of an essentially linear polymer consisting of several hundreds glucose residues, with occasional side chains attached by (1+6)-a-~-glucopyranose linkage^^.^. The chemical composition and the physico-chemical properties of starches vary considerably depending on their botanical origin’, although modification of starch structure and composition in planfa by genetic mutations, and now more commonly transgenic technology, has opened a new route to create novel starches6. Against this background of variation due to genetic origin, environmental effects may also modify starch composition and functionality7.’. Indeed environmentally induced variation is usually bigger than cultivar specific variation. From this perspective, a study of the effects of growth temperature during biosynthesis and subsequently post-harvest, the combined effects of moisture, heat and time have been researched in our l a b o ~ a t o r y ~A - ’ ~review . on hydrothermal modifications of starches (annealing and heat-moisture) has been published recently by Jacobs and D e l c o ~ r and ’ ~ whilst highly informative, indicates where knowledge gaps still exist in this particular area. Hence, the need for continued work.

2. Organisation of the starch granule Wide-angle X-ray diffraction patterns have revealed the presence of two distinct allomorphs (and a hybrid one) in starch granules: A-type in cereal starches, B-type in tuber and high-amylose starches and C-type (an intermediate between A and B) in

High Solid Systems

27 1

legume ~ t a r c h e s " ' ~There . is strong experimental evidence to suggest that crystallinity of native starches is due to the AP fraction via the packing of lefthanded, parallel-stranded double he lice^'*'^^'^. Similarly, with the most commonly accepted 'cluster model' of AP, crystallinity is envisaged as clustering of doublehelices formed from AP short chains that constitute 80-90% (molar basis) of AP fraction while the clusters are linked by longer chains expanding over several c~usters~"". 13C CP-MASNMR spectroscopy is a short order range probe that has been applied successfully to quantify molecular order corresponding to double helices in native starches". The comparison of the level of crystalline order by X-ray diffraction (22-35%) in granular starches from various botanical sources (wheat, rice, maize, waxy maize, high amylose maize and potato) and the extent of molecular order by I3C NMR (40-5396) suggests, however, that many double helices are not present in crystalline arrays2'. Recent studies by small-angle X-ray scattering2'.22and neutron scattering23*"have shown a well-defined peak which is attributed to alternating stacks of crystalline (AP clusters) and amorphous (AP branch points) lamellae with a periodicity of 8.7-9.2nm for various starches". These results are in good agreement with the average of 9.91l.Onm proposed by Hizukuri'" for the amylopectin cluster dimension. Results from high resolution transmission electron microscopy suggest that the crystalline lamellae are not uniformly straight or parallel but rather between 9 to 17 double helices form a single cluster of circa lOnm surrounded by radial amorphous materialz5. At a longer distance scale, growth rings in native and acid-eroded starch granules have been identified by microscopy technique^'.^^. They represent regions of alternating high and low refractive index, which comprise the dense shells of the semi-crystalline lamellae (- 120-400nm2') interspersed by large concentric amorphous shells. With respect to our work on annealing, it is necessary to identify distinct amorphous and crystalline regions. However, in practice, this is too simplistic for starch granules where a hierarchy of chain immobilisation is more adequate to describe the level of order such as the model proposed by Gidley et al. 26: (i) crystallisation of molecularly ordered chains; (ii) regularly ordered chains (e.g. helices); (iii) irregularly ordered chains with inter-residue hydrogen bonding; (iv) irregularly ordered chains without hydrogen bonding. AP double helices arrayed in crystallites belongs to Category (i), while AP double helices not arrayed in crystallites and single helical amylose-lipid complexes present in cereal starchesz7would fit Category (ii). In opposition, free amylose and AP (1+6)-a branch points (which are fully amorphous) should be characterised as (iv). However, the chain immobilisation of the later is certainly higher due to the presence of the crystalline lamellae which act as cross-links. In conclusion, at least four main amorphous regions may be identified in starch granulesz5: (a) inter-crystalline lamellae (AP branch points); (b) between adjacent clusters (radially orientated); (c) 'soft' layer of the growth rings; (d) amorphous radial channels.

272

Gums and Stabilisers for the Food Industry 10

3. Molecular mechanism of “in vitro” annealing of starches A general definition of annealing in the context of polymer science was coined by Wunderlich”: “In polymer science, the word annealing (from the old English unaelun, to bake on) is used to describe the improvement of crystallisation by heating to temperatures below the melting point which should lead to the growing of crystalline areas, perfection of crystals, and a change to more stable crystal structures”. By analogy with semicrystalline synthetic polymers, it is postulated that the annealing of starch crystallites should occur in the region between the glass transition (T,) and the onset of gelatinisation (To)temperat~resl~,~’.”. Since sub-gelatinisation annealing of starches can only take place when the amorphous regions are in a mobile rubbery state, above the glass transition temperature, it is crucial to determine the exact location of T, in starch-water systems. However, there is still considerable controversy, not only because the glassrubber transition occurs over a wide temperature range and is relatively difficult to detect due to the presence of crystalline domains which cross-linked amorphous regions” but also because T, is dependent on other parameters such as the thermal history, the molecular weight of polymer chains, the extent of crystallinity, the composition of the sample, and the presence of plasticise?. The effect of moisture content on “in vitro” annealing of wheat starch” has shown that annealing (evident from an increase of gelatinisation temperatures by DSC) can be initiated at room temperature when the moisture content exceeded 22% on a total weight basis. Recent investigations on native and amorphous starches have shown that when a comparable critical water content is exceeded, T, is depressed below room t e m p e r a t ~ r e ~ ’ * ~ ’ - ~ ~ while solid-state I3C CP-MAS/NMR of hydrated wheat starch also suggest that the amorphous regions are mobile at room temperature due to the plasticising effect of wateP. Providing that sufficient water is present to plasticise starch granules, “in vitro” annealing is more pronounced when the incubation temperature approaches the onset . Presumably this is due to a decrease of viscosity and an of gelatini~ationl~.~~’’.~~ associated increase of molecular mobilities in amorphous regions’. The effect of incubation time suggest that the dynamic nature of annealing is a non-equilibrium process since a linear dependence on a logarithmic time scale has been rep~rted’*~’*~’. The research discussed above in general supports the view that the underlying molecular mechanism of sub-gelatinisation annealing is due to the mobility of a glucan polymer chains in plasticised amorphous regions. However, it is recognised that a more precise model is required to relate starch structure (distinct amorphous regions having different mobilities”’), the rubbery visco-elastic behaviour of aglucan polymers and WLF (Williams-Landel-Ferry) kinetics which operates in the region delimited by T, and the melting of starch crystallites T,43.

4. Effects of “in vitro” annealing on the physico-chemical properties of starches “C CP-MAS/NMR spectroscopy and wide-angle X-ray scattering of starches annealed at sub-gelatinisation temperatures have shown a constancy in terms of both . Tester et al. I ‘ have shown that this is double-helical and crystalline orders”*14,41,U against a background of constant proximate composition and a-glucan structure (as determined by gel-permeation chromatography) while extremely low levels“ or no polysaccharide l e a ~ h i n g ~during ~ . ~ ’ annealing have been reported.

High Solid Systems

273

Differential scanning calorimetry shows a displacement of annealed starches endotherms towards higher temperatures while the shapes becomes sharper due to a narrowing of the gelatinisation rangel’.14.40341.4’*46 . By contrast, there is still controversy on the effect of annealing on the enthalpy of gelatinisation where either significant increases or constancies are reported14 while a decrease implies partial gelatinisation due to an incubation temperature too close to the onset of gelatinisation. Nevertheless, the constancy of the number of double helices and crystallinity pre- and post-annealing shows that the molecular reorganisations primarily reflects9.”: (i) perfection of the registration of double helices whereby minor defaults may be eliminated via optimisation of the number of hydrogen bonds imposed by the structure of AP short chains; (ii) perfection of the registration of crystallites via improved van der Waals contacts. There is also experimental evidence that shows molecular reorganisations occur in the amorphous regions of starch granules during annealing. The extent of swelling and amylose leaching are reduced po~t-annealing”~’~.~’, it has been hypothesised that an improved ordering and polymer packing occurs within amorphous regions. This hypothesis has been confirmed by small-angle X-ray scattering (SAXS) which indicates a delay of the water uptake and a reduction of the water absorption prior to gelatinisation in annealed starches4’. It is likely that the reorganisations due to annealing occurs in growth rings where swelling is initiated4*.By contrast, the large reduction of swelling post-gelatinisation in annealed starches” would suggest changes of the interactions between amorphous and crystalline regions in the semicrystalline lamellae whereby the improved perfection at both double-helical and crystalline levels would restrict swelling. Proposed models describing the effects of “in vitro” annealing do, however, have their limitations since very different behaviour has been observed in high-amylose starches where there is a large increase in the enthalpy of gelatinisation postRecent evidence of double-helical order due to the amylose fraction in low-amylopectin starch49 suggests that improved registration of amylose double helices (and possibly amylopectin-amylose in addition to AP double helices) is likely to occur during annealing of high-amylose starches’*. 5. Effects of “in vivo” annealing on the physico-chemical properties of starches

The effects of increasing growth temperature (from 10 to 25”C, in controlled environment chambers) on the physico-chemical properties of starch granules has been investigated in two “in vivo” model sytems, potato microtubers’’ and potato tubers”. No significant correlations could be found between growth temperature and the proximate composition of the starches (AM and extent of AP phosphorylation) and the fine structure of AP as determined by gel-permeation chromatography (chain length distribution, weight-average degree of polymerisation). Similarly, ”C CPMASMMR spectroscopy and wide-angle X-ray scattering of potato starches showed that both the level of double-helical and crystalline order remained unchanged against a background of increasing growth temperature. By contrast, gelatinisation temperatures increased as growth temperature was elevated, although the enthalpy of gelatinisation remained almost constant. The extent of swelling post-gelatinisation in contrast to gelatinisation temperatures significantly decreases as growth temperature increases.

214

Gums and Stabilisers for the Food Industry I0

Table 1 Comparison between growth temperature induced “in vivo ” and laboratory generated “in vitro ” annealing of starches. Parameter

In vivo

In vitro

Regulation of biosynthetic enzymes Granule size and distribution Granule composition a-Glucan structure Number of double helices X-ray crystallinity Gelatinisation temperatures (To, T, and T,)

yes (but a parallel event to annealing) decrease

not applicable

not correlated constant constant constant increase

constant constant constant constant increase

Gelatinisation enthalpy (AH) Swelling factor

constant (or small increase) decrease

constant (or small increase) decrease

constant

When discussing the analogy we have penned to compare “in vitro” and “in vivo” annealing, we believe that the major molecular reorganisations underlying both events are: (i) improved perfection at both double-helical and crystalline registration levels (rather than formation of new double-helices and crystallites); (ii) improved ordering within amorphous regions. The molecular freedom for “in vivo” annealing due to increasing growth temperature is, presumably, related to higher molecular mobilities in the rubbery amorphous regions when amyloplasts are fully plasticised during biosynthesis. A comparison between “in vitro” and “in vivo” annealing is shown in Table 1. The effects of growth temperature during starch biosynthesis potentially have important industrial consequences such as the energy requirement for starch gelatinisation in food products or for malting and brewing”. This is also true when starch granules are exposed to annealing conditions during extraction (for example wet-milling of maize3”)or processing. Hence, there is a real need to understand these processes. In conclusion, in addition to genetic factors which affect starch composition and functionality, a knowledge of the effects of environmental conditions (especially growth temperature) and more generally the thermal history of starch granules needs to be addressed to optimise product quality. Molecular biologists must pay due regard to these effects when they develop novel starches.

High Solid Systems

215

References 1 . D. French, ‘Starch: Chemistry and Technology’, Academic Press, Orlando, 2“‘ edition, 1984, Chapter VII, p. 183. 2. W. R. Morrison and J. Karkalas, ‘Methods in Plant Biochemistry’, Academic Press, London, 1990, Vol. 2, Chapter 9, p. 323. 3. M. J. Gidley and D. Cooke, Biochem. SOC.Trans., 1991,19, 551. 4. D. J. Manners, Carbohydr. Polym., 1989,11,87. 5 . J. J. M. Swinkels, StarcWStiirke, 1985,37, 1. 6. R. G. F. Visser and E. Jacobsen, TIBTECH, 1993,11,63. 7. J. C. Shannon and D. L. Garwood, ‘Starch: Chemistry and Technology’, Academic Press, Orlando, 2“dedition, 1984, Chapter 111, p. 25. 8. R. F. Tester, Int. J. Biol. Macromol., 1997,21, 37. 9. S. J. J. Debon, “In vivo” and “in vitro” Annealing of Starches, PhD thesis, Glasgow Caledonian University, 1998. 10. S. J. J. Debon, R. F. Tester, S. Millam and H. V. Davies, J. Sci. Food Agric., 1998,76, 599. 11. R. F. Tester, S. J. J. Debon and J. Karkalas, J. Cereal Sci., 1998,28,259. 12. R. F. Tester, S. J. J. Debon and M. Sommerville, Carbohydr. Polym., 1999, submitted for publication. 13. R. F. Tester, S. J. J. Debon, H. V. Davies and M. J. Gidley, J. Sci. Food Agric., 1998, in press. 14. H. Jacobs and J. A. Delcour, J. Agric. Food Chem., 1998,46,2895. 15. H. F. Zobel, StarcWStiirke , 1988,40, 1. 16. A. Imberty, H. Chanzy, S. Perez, A. Bulton and V. Tran, J. Mol. Biol., 1988, 201,365. 17. A. Imberty and S . Perez, Biopolymers, 1988,27, 1205. 18. S. Hizukuri, Carbohydr. Res., 1986,147,342. 19. M. J. Gidley and S. M. Bociek, J. Am. Chem. SOC.,1985,107,7040. 20. M. J. Gidley, A. J. McArthur, A. H. Darke and S. Ablett, ‘New PhysicoChemical Techniques for the Characterisation of Complex Food Systems’, Blackie Academic & Professional, Chapman and Hall, London, 1995, Chapter 13, p. 296. 21. R. E. Cameron and A. M. Donald, Polymer, 1992,33,2628. 22. P. J. Jenkins, R. E. Cameron and A. M. Donald, StarcWStiirke , 1993,45417. 23. J. M. V. Blanshard, D. R. Bates, A. H. Muhr, D. L. Worcester and J. S. Higgins, Carbohydr. Polym., 1984,4427. 24. P. J. Jenkins and A. M. Donald, Polymer, 1996,37,5559. 25. D. J. Gallant, B. Bouchet and P. M. Baldwin, Carbohydr. Polym., 1997,32, 177. 26. M. J. Gidley, D. Cooke and S. Ward-Smith, “The Glassy State in Foods’, University of Nottingham Press, Nottingham, 1993, Chapter 15, p. 303. 27. W. R. Momson, R.V. Law and C. E. Snape, J. Cereal Sci., 1993,18,107. 28. B. Wunderlich, ‘Macromolecular Physics. Vol. 2, Crystal Nucleation, Growth, Annealing’, Academic Press, New York, 1976, Chapter VII, p. 348. 29. K. J. Zeleznak and R. C. Hoseney, Cereal Chem., 1987,64, 121. 30. J. Lelitvre, ‘Developments in Carbohydrate Chemistry’, AACC, St Paul, Minnesota, 1992, p. 137. 31. H. J. Thiewes and P. A. M. Steeneken, Carbohydr. Polym, 1997,32,123.

276

Gums and Stabilisers for the Food Industry 10

32. M. T. Kalichevsky, E. M. Jaroszkiewicz, S. Ablett, J. M. V. Blanshard and P. J. Lillford, Carbohydr. Polym., 1992, 18,77. 33. C. G. Biliaderis, C. M. Page, T. J. Maurice and B. 0. Juliano, J. Agric. Food Chem., 1986,34,6. 34. K. Jouppila and Y. H. Roos, Carbohydr. Polym., 1997,32,95. 35. D. Bencztdi, I. Tomka and F. Escher, Macromolecules, 1998,31,3055. 36. K. R. Morgan, R. H. Furneaux and N. G. Larsen, Carbohydr. Res., 1995,276, 387. 37. F. Nakazawa, S. Noguchi, J. Takahashi and M. Takada, Agric. Biol. Chem., 1984,48,2647. 38. B. R. Krueger, C. A. Knutson, G. E. Inglett and C. E. Walker, J. Food Sci., 1987, 52,715. 39. C. A. Knutson, Cereal Chem., 1990,67,376. 40. I. Larsson and A.-C. Eliasson, StarcWStarke , 1991,43,227. 41. R. Hoover and T. Vasanthan, J. Food Biochem., 1994,17,303. 42. C. C. Seow and C. K. Vasanti-Nair, Carbohydr. Res., 1994,261,307. 43. L. Slade and H. Levine, Carbohydr. Polym., 1993,21, 105. 44. R. Stute, StarcWStarke , 1992,44,205. 45. B. M. Gough and J. N. Pybus, StarcWStarke , 1971,23,210. 46. D. A. Yost and R. C. Hoseney, StarcWStiirke , 1986,38,289. 47. H. Jacobs, N. Mischenko, M. H. J. Koch, R. C. Eerlingen, J. A. Delcour and H. Reynaers, Carbohydr. Res., 1998,306, 1. 48. A. M. Donald, T. A. Waigh, P. J. Jenkins, M. J. Gidley, M. Debet and A. Smith, ‘Starch Structure and Functionality’, The Royal Society of Chemistry, Cambridge, 1997, p. 172. 49. Y.-C. Shi, T. Capitani, P. Trzako and R. Jeffcoat, J. Cereal Sci., 1998,27,289.

SOLVENT STRUCTURE AND GELATION OF POLYSACCHARIDES IN CONCENTRATED SOLUTIONS OF SIMPLE SUGARS

David Oakenfull Food Science Australia, P.O. Box 52, North Ryde, NSW 21 13, Australia.

1 ABSTRACT Simple sugars added at high concentration strongly influence the gelation properties of polysaccharides. Starch, K-carrageenan, gellan and low-methoxy and high-methoxy pectins have been extensively studied. With concentrations of sugar of up to about 40% gels typically become firmer (higher storage modulus, G') and the gelation temperature (Tg)increases. Sugars appear to stabilise the junction zones in the gel network and these effects are strongly dependent on the chemical structure of the added sugar, as well as its concentration. The geometric requirements of the hydrogen bonds made by water molecules impart considerable structure to the water surrounding solutes such as simple sugars. The effects of different sugars on G' and Tg can be correlated with physical parameters that are related to the effect of the sugar on the structure of the surrounding water molecules. Recent computer simulation studies of aqueous sugar solutions, using the molecular dynamics approach, have revealed more details of the solution structure than was previously accessible from spectroscopic and thermodynamic studies. This leads to a better understanding of how simple sugars stabilise the junction zones in polysaccharide gel networks and leads to some unifying principles that might be helpful in practical applications. 2 INTRODUCTION Addition of simple sugars can markedly change the properties of polysaccharide gels.' Gel strength, as measured by rupture strength or storage modulus, generally increases;2 gelation temperatures and melting temperatures also increase.293 For example, added sucrose enhances the rigidity of K-carrageenan gels, as shown in Figure l ? Curve fitting, using equations derived by Oakenf~ll,~ showed that addition of sucrose causes the junction zones to become smaller and involve fewer polysaccharide chains (Table 1). As a result, the number of junction zone per unit volume of gel increases and the rigidity modulus increases. Nishinari and Watase' have also studied the effects of sucrose on the thermal properties of K-carrageenan gels. The melting and setting temperatures were seen to increase with increasing concentration of sucrose. They also studied different sugars and found that the magnitude of the increase was sensitive to the structure of the sugar, as shown in Figure 2. They showed that the increase in the

Gums and Stabilisers for the Food Industry I0

218

melting temperature of the gel in response to added sugar was linearly related to the weighted average number of equatorial OH-groups for the conformers of the sugar found in solution. They suggested that this relationship might be explained either by the change in the structure of water as a solvent or by direct interaction between the OH-groups of the sugar and the polysac~haride.~Nilsson and colleagues6 have also argued that direct interaction between the sugar and the polysaccharide influences conformational transitions and aggregation. They showed that effects of cosolutes on conformational equilibria in polysaccharides can be explained by adsorption or depletion of the cosolute at the polysaccharide-water interface. They also found from their analysis that cosolute-water interactions are important in determining the preferential adsorption behaviour of the cosolute.6 It is argued here that effects of sugars on the structure of water are the dominant factor and that we are looking at part of a general phenomenon in which solvent structure is an important factor in controlling gelation of polysaccharides. In aqueous solution, polysaccharide structures, including the junction zones in gel networks) are stabilised by various combinations of hydrogen bonds, hydrophobic and electrostatic force^.^ Hydrogen bonds and hydrophobic forces depend very much on the unique structure of water as a solvent.839 This discussion is confined to sugar concentrations below about 40% by weight which can be considered as true aqueous solutions. At higher concentrations, aqueous sugar systems can form glasses with very different physical properties and correspondingly different effects on polysaccharide gels. lo

Junction zone parameters, estimated from the data shown in Figure 1 for gelation of K-carrageenan at 15°C in the presence of sucrose. M is the number average molecular weight of the polysaccharide; M, is the number average molecular weight of the junction zones, n is the average number of interacting segments of polysaccharide chains per junction zone, K, is the association constant for junction zones and AGOJ is the free energy ofjunction zone fonnation (aGoJ = - RT In K,). Concentration of M MJ n KJ AGO, sucrose (X w/w) (kT mol-')

Table 1

0

101 OOO

21 600

6.05

6.94

X

10"

-98.4

30

108 OOO

13 400

4.21

1.01

X

10"

-60.7

3 SUGARS AND THE STRUCTURE OF WATER Because a water molecule is both a hydrogen bond donor and a hydrogen bond acceptor, liquid water itself resembles a gel in that it is a continuous molecular network - but it is an exceedingly transient one, continually reforming and regrouping. No other liquid is held together by such strong and directional intermolecular forces. In ice, every water molecule is bonded to its four nearest neighbours. Its OH-bonds are directed towards the lone pairs of electrons on two of these nearest neighbours forming two 0-H-0 bonds; in turn, each of its lone pairs is directed towards an OH-bond of one of the neighbours, forming two 0 . - H - 0 bonds. In liquid water, this structure is incomplete and transient, in a way still not fully understood. In aqueous solution, nonpolar molecules, or molecules containing non-polar groups, are surrounded by an ordered or "structured" layer of water molecules. When these nonpolar molecules

High Solid Systems

279

4000 I

A h

L

(d

a v

0 30%

2000

CJ

0

0

Figure 1

2

4

6

K-carrageenan (g/kg) Efect of added sucrose on the shear moduIus of K-carrageenan gels at 15°C. The absolute shear modulus was measured by the method of Oakenfill, Parker and Tanner." in which a probe is inserted inro the gel formed in a cylindrical container. Gels (10 g samples) were fonned in 'scintillation vials' (radius 12.5 mm). The probe (of radius 1.5 mm) was inserted step-wise into the gel with a series of 5 s pulses at a speed of 0.0847 mm/s. The equilibriumforce exerted on the probe was measured at each step and the apparent Young's modulus (Y = stress/strain) calculated from the slope of force vs penetration. The absolute shear modulus (G) was calculated from the formula G = 0.0208.Y.

approach each other some of the ordered water molecules are "squeezed out" and the molecular rearrangements that this entails provide the thermodynamic driving force for Similarly in electrostatic interactions, the hydration hydrophobic shells of ordered water molecules surrounding the ions are important. Association of ions involves interpenetration of these hydration shells with consequent rearrangement of water mo1ecules.l2 Sugars have a strong influence on this structure. Indirect evidence, derived spectroscopy and thermodynamics, shows that sugars interact with water in a way that ~ ' ~details of solution structure have proved depends on their molecular ~ t r u c t u r e ' ~but very difficult to probe experimentally. Recently Brady and his coworker^^^*^^ have used computer simulation studies which allow solvent structuring to be examined directly. They used a molecular dynamics approach in which the sugar molecule was imagined to be in a periodic cubic box containing 502 water molecules (corresponding to three full solvation shells around the sugar solute molecule in every direction). The various intermolecular forces were expressed mathematically and initial velocities assigned randomly from a Boltzman distribution. The system was then "equilibrated" stepwise, providing a snapshot picture of the nonuniform distribution of the molecules in three dimensions. Their results confirm that a sugar molecule can impose considerable additional structure on its aqueous environment and that this structuring depends strongly on the specific topology of the sugar molecule.

Gums and Stabilisersfor the Food Industry 10

280

20

15

y^ v

10

F

a

5

0 0

2

4

6

8 1 0

n (e-OH) Figure 2

w-carrageenan gels; relationship between the eflect of added sugar on the melting temperature (ATJ and the number of equatorial@ attached OHgroups on the added sugar. Concentrations of added sugars: A, 0.2 nwl dm-3; *, 0.5 mol dmS3;0 , 1 mol ah-?.A, 2 nwl l%e sugars Gfrom left to right) are ribose, mannose, fructose, glucose, sucrose, maltose and raflnose. m e data are from Nishinari and Watase.'

"Structure making" or "structure breaking" by a solute is accompanied by a volume change.14 Consequently a useful quantitative measure of this structuring effect is provided by the dimensionless quantity:

where V, is the partial molar volume of the sugar - which includes the volume change associated with the concomitant rearrangement of water molecules as the sugar dissolves - and V, is the van der Waals volume of the water. There is an inverse relationship between { and the structure-making effect of the sugar." Back and coworkers,'* for example, found that the increase in the thermal denaturation temperature of lysozyme and ovalbumin in response to different added sugars was linearly related to {. A similar relationship is found for various gelation characteristics of polysaccharides.

28 1

High Solid Systems

23

Figure 3

24

25

26

27

28

Relationship between the storage modulus (GI) at 25°C of pea starch gels (40% whv) containing 17% whv added sugar and the water structure parameter 5 (see tat). The sugars, in order of increasing are fructose, glucose, sucrose and ribose.

r,

3 SUGARS, THE STRUCTURE OF WATER AND GELATION 3.1 Starch Added sucrose (and also glucose and fructose) increase the rigidity of starch gels.” Sucrose and other sugars also increase the gelatinisation temperature and the enthalpy of gelatinisation.m Figure 3 shows that the storage modulus (GI) of pea starch gels plotted against 5. G‘ increases with decreasing 5 suggesting that the increase in rigidity is mediated by the effect of the sugar on the structure of water.

3.2 r-carrageenan Added sucrose increases the shear modulus (G) of K-carrageen gels as shown in Figure 1. Similar results have been obtained for the sugars ribose, arabinose, glucose, galactose, fructose and rhamnase. Figure 4 shows a similar result to that for pea starch in Figure 3. The shear modulus increases with decreasing 5, again suggesting that the increase in rigidity is mediated b the effect of the sugar on the structure of water. Nishinari and Watase’s results for the effects of sugars on the melting temperature of K-carrageenan gels (Figure 2) also show a linear relationship to 5 (Figure 5).

Y

3.3 High Methoxyl Pectins High methoxyl (HM) pectins differ from other gel-forming plysaccharides in that high concentrations of sugar (typically more than about 55% w/w) are required for

*

Gums and Stabilisersfor the Food Industry 10

282

la50

20

25 lOO(V,

Figure 4

30

-

35

VW)/VW

Relationship between the shear modulus (G) at 15°C of K-carrageenan gels (0.1 whv) containing 30% whv added sugar and the water structure parameter 5 (see text). l?u? sugars, in order of increasing 5, are galactose, fructose. arabinose, glucose, ribose and rhamnose.

10 8

iz v

6

E

l-

a

4 2 0 23

24

25 lOO(V,

Figure 5

26

27

28

- V,)/VW

Relationship between the melting temperature (AT,,,)of K-carrageenan gels (2% whv) containing 0.5 mol ah" added sugar and the water structure parameter 5 (see text). The sugars, in order of increasing S; are galactose, fnrctose, glucose, sucrose and ribose.

gelation to occur at all. The junction zones in these gels are stabilised by a combination of hydrogen bonds and hydrophobic interactions between the esterified wboxylic acid groups. Oakenfull and Scottz1 showed that the hydrophobic contribution is essential and without the sugar is too weak for junction zones to be thermodynamically stable. They estimated that in a typical HM pectin (of 70% DE), the hydrophobic contribution to the free energy of formation of junction zones is -18.6 kJ mol-' compared with a contribution from hydrogen bonds Of -37.5 kJ mol-1. But

High Solid Systems

283

hydrogen bonding alone is insufficient to overcome the entropic barrier to gelation (+41.1 kJ mol-1) making the hydrophobic contribution essential. Hydrophobic interactions are enhanced by added sugars and Oakenfull and Scot?’ found that the minimum concentration of sugar (or polyol) required to form a gel with the HM pectin was also linearly related to c. 4 CONCLUSIONS Sugars added at concentrations of up to about 50% (w/w) stabilise the junction zones in a number of polysaccharide gel systems - they increase. the rigidity of the gel and they increase setting and melting temperatures. Sugars of different structure have very different effects on the gelation behaviour and these effects can be related to a physical parameter (0 which quantifies the effect of the sugar on the structure of water. It is possible that we are seeing here a general phenomenon - sugars influencing the aggregation of polysacchiuides to form a gel network via their effects on the structure and solvent properties of water. Unfortunately the structure of water and how it is influenced by added solutes is a complex area of physical chemistry that remains poorly understood. We are still a long way from understanding these complex systems. Much more research is needed both in the areas of the sugar-polysaccharide systems themselves and the underlying physical chemistry. References 1. 2. 3. 4. 5.

6. 7.

8. 9. 10. 11. 12. 13. 14. 15.

D.L. Gerdes, E.E. Bums and L.S. Harrow, Lebensm.-Wss. u. Technol., 1987, 20, 282. K. Nishinari, M. Watase, E. Miyoshi, T Takaya and D. Oakenfull, Food Technol., 1995, 10, 90. D. Oakenfull, in ‘Confectionery Science: Proceedings of an International Symposium’, G.R. Zeigler, ed., Pennsylvania State University, 1997, p. 67. D. Oakenfull, J. Food Sci., 1984, 49, 1103. K. Nishinari and M. Watase, T h e m h i m . Acta, 1992, 206, 149. S. Nilsson, L. Piculell and M. Malmsten, J. Phys. Chem, 1990, 94, 5149. D. Oakenfull, in ‘Polysaccharide Association Structures in Food’, ed. R.H. Walter, Marcel Dekker, New York, 1998, p. 15. A. Ben-Naim, ‘HydrophobicInteractions’, Plenum Press, New York, 1973. C. Tanford, ‘The Hydrophobic Effect’, John Wiley, New York, 1973. V. Evageliou, S. Kasapis and G. Sworn, in ‘Gums and Stabilisers for the Food Industry 9’, eds. P.A. Williams and G.0 Phillips, The Royal Society of Chemistry, Cambridge, 1998, p. 333. D.G. Oakenfull, N.S. Parker and R.I. Tanner, J. Tature Studies. 1989, 19, 407. J. Israelachvili and H. Wennerstrom, Nature, 1996, 379, 219. D.W. James and L. Rintoul, Aust. J. Chem., 1982, 35, 1157. F. Franks,J.R. Ravenhill and D.S. Reid, J. Solution Chem.,1972, 1, 131. Q. Liu and J.W. Brady, J. Am. Chem. Soc., 1996, 118, 12276.

284

16. 17. 18. 19. 20. 21.

Gums and Stabilisers for the Food Industry 10

Q. Liu and J.W. Brady, J . Phys. Chem. B, 1997, 101, 1317. D.G. Oakenfull and D.E. Fenwick, J . Chem. SOC.Furaduy Trans. 1, 1979, 75, 636. J.F. Back, D. Oakenfull and M.B. Smith, Biochemistry, 1979, 18, 5191. D.J. Prokopowich and C.G. Biliaderis, Food Chem., 1995, 52, 255. K. Kohyama and K. Nishinari, J. Agric. Food Chem., 1991, 39, 1406. D. Oakenfull and A. Scott, J. Food Sci., 1984, 49, 1093.

Effect of sugar concentration on the properties of gellan gum gels.

Graham Sworn and Corinne Johnson Kelco Biopolymers, Waterfield, Tadworth, Surrey, KT20 5HQ.

ABSTRACT Sugars are a major constituent of many foods and as such an understanding of their influence on the properties of other components such as hydrocolloids is important. This is a developing area of scientific interest and recent studies applying the theories of synthetic polymer rheology to polysaccharide sugar systems have greatly enhanced ow understanding. In this paper we will discuss how sugar type and concentration effect the rheological properties of the polysaccharide gellan gum. The relationship between sugar concentration and ion requirements is explored. Additionally, the effect of pH is also addressed. The results from model system studies will then be discussed in terms of food applications such as fruit fillings, jams and confectionery. Practical recommendations, that aid in the formulation of such products, are provided.

1 INTRODUCTION Gellan gum is an extracellular polysaccharide secreted by the microorganism Sphingomonus elodea. Commercially, it is manufactured using a controlled fermentation process. The primary structure of gellan gum is composed of a linear tetrasaccharide repeat unit: +3)-P-D-Glcp-( 1+4)-P-D-GlcpA-( 1+4)-P-D-Glcp-( 1+4)-a-L-Rhap-( 1+.I. The polymer is produced with two acyl substituents present on the 3-linked glucose, namely, L-glyceryl, positioned at O(2) and acetyl at O(6). On average there is one glycerate per repeat unit and one acetate per every two repeats.' Gellan is soluble in hot water and forms gels at low concentrations when hot solutions are cooled in the presence of gel promoting cations. It is available in a substituted or unsubstituted form. These are commonly referred to as high acyl (HA) and low acyl (LA) gellan gum, respectively. The high acyl form is sometimes called native gellan gum since this is the form in which it is secreted by the microorganism. Alkali treatment during manufacture removes the substituent groups to yield LA gellan gum. Gel properties depend on the degree of substitution with the substituted, HA gellan gum producing soft, elastic gels whilst the unsubstituted, LA gellan gum produces hard, brittle gels.

286

Gums and Stabilisersfor the Food Industry 10

The presence of sugars have two major effects on the properties of LA gellan gum gels. Firstly, above approximately 40% sugar gels become less firm and less brittle i.e. softer and more elastic. These effects are believed to be the result of the sugars inhibiting the aggregation step of the gelation process4.'. These effects are also influenced by the type of sugar. Sucrose has a greater inhibitory effect than glucose, fructose or corn syrups6 The differences observed between sugars mean that texture can, to some degree, be varied by manipulating the sugar composition of the system. For example, partial replacement of sucrose with fructose or corn syrup, a common practice to control crystallisation in confectionery manufacture results in firmer more brittle gels.' The second main effect of sugars is to reduce the ion requirements for optimum gel properties. For example, the presence of 40% w/w sugar approximately halves the calcium required for maximum gel modulus, from 8 - 10 mM in water gels to 4 - 5 mM in the sugar gels and the addition of 60% w/w sugar results in approximately a 10 fold reduction in the requirement for calcium, with only 0.5 - 1.O mM added calcium required for maximum gel modulus. Previous studies have provided extensive knowledge of the effect of sugars and ions on gel properties but detailed studies of the combined effects of these and the influence of pH are limited. This study can be split into three stages as outlined below: Stage 1 Determine the optimum calcium, sodium, potassium and acid @H) concentrations for gel formation of 0.5%w/w LA gellan gum gels in water (water gels) and 60% sucrose (sugar gels). Stage 2 Study the effects of pH on the properties of water and sugar gels made under optimum ion conditions. Stage 3 Study the effects of ion concentration (calcium, sodium and potassium) on the properties of water and sugar gels formed under optimum pH conditions.

2 MATERIALS AND METHODS Gels were prepared (0.5% w/w) by dispersing the LA gellan gum in deionised water and heating the dispersion with stirring to approximately 90°C to fully hydrate the gum. Sucrose was added slowly to the hot gum solution whilst maintaining the temperature above approximately 80°C. The solution was then re-heated to 90°C where necessary before addition of acid and or metal ions. Finally evaporative losses were made good by the addition of deionised water. A sample of the solution was loaded at 95°C onto a CarriMed controlled stress rheometer CSL2 fitted with a 4cm steel parallel plate with a gap of 1OOOpm. The exposed surfaces were covered with a silicone oil to minimise evaporative losses during the test. The sample was then cooled from 95°C to 5°C at 4°C mid' whilst measuring the elastic modulus (G') at a frequency of 10 rad.s-' and strain of 1%. The setting temperature was determined as the temperature at which elastic modulus first showed a significant increase. The remainder of the sample was poured into ring moulds of 14mm height and 16mm diameter, covered and stored overnight at 5°C before Texture profile analysis (TPA) testing was performed. TPA was performed using an Instron 4300 instrument fitted with a 63mm diameter compression plate and programmed to compress the gel to 85% strain twice in succession. The pH of the gel was measured using a Jenway 3020 pH meter.

High Solid Systems

287

3 RESULTS AND DISCUSSION 3.1 Optimum Ion Requirements for Gellan Gum Gelation (Stage 1) The modulus of 0.5% LA gellan gum gels as a function of ionic concentration in water (water gels) and 60% sucrose (sugar gels) are shown in Figure la. To enable the acid data to be shown on the same scale it has been plotted as mM citric acid. It can be seen that whilst metal ion requirements for optimum gel modulus are dramatically reduced in the presence of sucrose, acid requirements are the same in both water and sugar gels. Although the concentration of acid is the same in both systems the final pH was slightly lower in the sugar gels compared to the water gels. In fact, pH rather than acid concentration is the important consideration for acid gels and it was found that maximum modulus occurred between pH 2.5 and 2.8 for a number of different acids (Figure lb). These included, citric, malic, acetic, lactic and hydrochloric acid. Optimum ionic conditions to achieve maximum gel modulus and the corresponding setting temperatures are summarised in Table I. Acid gels have the highest modulus values in both water and sugar. Divalent, calcium gels are stronger than monovalent ion gels in water as has been previously reported? In this study, however it can be seen that the reverse is true in the sugar gels. In this case, sodium forms stronger gels than calcium. Whilst sodium and potassium gels are very similar in water they are quite different in the presence of sugar. LA gellan gum is less tolerant of potassium and the gels are softer compared to sodium. This is particularly important to consider when formulating fruit fillings since the ionic composition of h i t s varies considerably (Table 11). For example, switching a formulation from apple to blackcmant will result in a significant increase in the potassium ion concentration and will lead to textural changes. It is interesting to note that under optimum conditions setting temperatures are lower in the presence of sucrose than in water for the metal ion mediated gels. Conversely, the presence of sucrose dramatically increases the setting temperature of the acid gels. Previous studies have suggested that pre-gelation is linked to the release of calcium when the acid is added but in fact it may be due solely to the acid itself. The very low setting temperature of acidic water gels is important to note as these conditions are similar to many fluid gel beverage applications.” This means that in many acidic beverages it is necessary to cool to below 15OC to form the fluid gel. Table I Optimum gelling conditions for 0.5% w/w LA gellan gum gels in water and 60% sucrose. Ion (mM)

Water Modulus (Ncm-’)

8 - 10 260 - 300 240 - 260 pH 2.8 - 3.0

19.3 12.3 12.1 20.3

Ion conc.

Calcium Sodium Potassium Acid

Setting temp. (“C) 42 54 59 10

Ion conc. (mM)

60% sucrose Modulus (Ncrn-’)

0.5 - 1.0 25 - 35 8 - 10 pH 2.5 - 2.7

1.66 3.13 1.71 7.74

Setting temp. (“C) 38 47 43 64

288

Gums and Stabilisers for the Food Industry 10

A

0.1

1 10 100 Concentration (mM)

1000

1.5

1

I

I

2.0

2.5 PH

3.0

3.5

Fig.1 a) Effect of calcium (squares), potassium (triangles), sodium (circles) and acid (diamonds) on the modulus of 0.5% gellan gum gels in water (closed symbols) and 60% sucrose (open symbols) and b) Effect of citric (O),malic (V), lactic (A),acetic (0) and hydrochloric (+) acid on the modulus of 0.5%w/w LA gellan gum sugar gels. Table I1 Ionic content of raw fruits.'' Fruit Apple Blackcurrant Raspberry Strawberry Apricot Peach

Ca" (mg/lOOg) 4 60 25 16 15

7

Mg" (mg/lOOg) 3 17 19 10 11 9

Na' (mg/lOOg) 2 3 3 6 2 1

K' (mg/l OOg) 88

370 170 160 270 160

3.2 Effect of pH on Gels Prepared Under Optimum Ion Conditions (Stage 2)

Figure 2a shows the effect of pH on the modulus of 0.5% w/w LA gellan gum gels formed in water at optimum calcium and sodium concentrations. It can be seen that the modulus of the gels decreases rapidly below approximately pH 3.0 which is consistent with previous studies.*

High Solid Systems

25

289

J a)

20

0

n

-

%15

Y

A

3 10 5

0 2.0

3.0

4.0 PH

5.0

6.0

2.0

3.0

4.0 PH

5.0

6.0

Fig. 2. Effect of pH on the modulus of 0.5% w/w LA gellan gum gels prepared at optimum ion concentrations in a) water, lOmM Ca" (0),300mM Na' (A),acid only (). and b) 60% w/w sucrose, O S m M Ca" (0),lOmM K ' (A),acid only (.). A different pattern of behaviour is seen in the sugar gels (Figure 2b). Gels remain largely unaffected by pH down to around pH 4.0. Between pH 3.5 to 3.0 very weak gels are formed. Below pH 3.0 gels are formed again and show optimum strength at around pH 2.8 as noted for acid sugar gels (solid squares in Figure 2b). These acid gels are stronger than the original calcium or potassium gels. It would appear that lowering the pH below 4 gradually transforms the system from an ion mediated gel to an acid mediated gel as the acid protonates the carboxyl groups of the gellan gum. In the nodweak-gelling region between pH 3.5 to 3.0 there is probably insufficient binding sites remaining for an ion mediated gel due to displacement with H' from the acid, but too little acid to form the acid gel. Additionally, the carboxylic acids such as citric acid are also mild sequestrants. Continued reduction in pH sees the formation of the acid gel which, as we have seen from the first stage is stronger than the ion mediated gels. Addition of ions to systems in the nodweak gelling region will result in gel formation. 3.3 Effect of Ions on Gels Prepared Under Optimum Acid Conditions (Stage 3)

Addition of ions to optimum acid sugar gels results in a reduction of gel modulus and gelation temperature (Figures 3 a b d b). The onset of the reduction in gel modulus and setting temperature occurs as the ion concentration exceeds the optimum established in stage 1 of the study. This would suggest that even at these acid pH's some ion exchange occurs between the H' and the metal ions when there is an excess in terms of gellan gum requirements. Acid sugar gels with added potassium ions had the highest setting temperatures. As shown in Table I1 potassium is the predominant ion in many fruits. A formulation containing 50% blackcurrant for example, is the equivalent of adding 5 m M potassium to the system. This is close to the optimum ion requirements in a 60% sucrose gel.

290

Gums and Stabilisers for the Food Industry 10

In this study the systems are unbuffered. Since most buffers are also sequestrants it was felt that their presence would be an added complication. Addition of ions did result in a change in the pH of the gels but, although not shown here, a plot of pH versus gel modulus and setting temperature shows no specific effects associated with this small change in pH. It can therefore be concluded that the effects observed are due to the change in the ionic concentration alone. 75 70

P'

@&

!3 h

65 60 55 50

0.1

1

10

100

Concentration (mM)

1000

0.1

1

10

100

1000

Concentration (mM)

Fig.3. Effect of calcium (O),sodium (A)and potassium (0)on a) modulus and b) setting temperature of 0.5% wlw LA gellan gum sugar gels prepared at optimum acid concentrations (PH 2.7).

4 CONCLUSIONS Previous studies have shown that ion requirements for LA gellan gum gelation are reduced in the presence of sugar. This study confirms this and also shows that the relative effectiveness of the gelling ions is also changed. For example, calcium gellan gels are stronger (higher modulus) than sodium gels in water but, in the presence of 60% wlw sucrose, the reverse is observed. Similarly potassium and sodium gels have similar maximum modulus values in water but sodium gels are significantly stronger in 60% sucrose. LA gellan gum has major commercial opportunities in medium to high solids applications such as, low solids jams, fruit filling and confectionery where the products good stability to acid, excellent clarity, clean flavour and low use levels are key attributes to the manufacturer and consumer alike. These products generally contain between 30 80% total soluble solids (TSS) and have a low pH (PH 2.8 - 4.5). It is shown here that the natural ion content in h i t is ideally suited to the formulation of gellan gum gels in these systems.

High Solid Systems

29 1

5 REFERENCES 1

2 3

4

5

6 7 8

9 10 11

O’Neill, M.A., Selvendran, R.R. & Morris, V.J. Structure of the acidic extracellular gelling polysaccharide produced by Pseudomonas elodea. Curbohydr.Res., (1983) 124, 123-133. Janson, P.-E., Lindburg, B. & Sandford, P.A. Structural studies of gellan gum, an extracellular polysaccharide elaborated by Pseudomonas elodea. Curbohydr.Res., (1983) 124, 135-139. Kuo, M.-S., Mort, A.J. & Dell, A. Identification and location of Lglycerate, an unusual acyl substituent in gellan gum. Carbohydr. Res., (1986) 156, 173-187. Sworn, G. Gelation of gellan gum in confectionery systems. In Gums and Stabilisers for the Food Industry 8, (1996) eds. G.O. Phillips, P.A. Williams, & D.J. Wedlock, IRL Press, Oxford, pp 341 - 349. Sworn, G. and Kasapis, S. The use of Amhenius and WLF kinetics to rationalise the mechanical spectrum in high sugar gellan systems. Carbohydrate Research (1998) 309,353 - 361. Sworn, G and Kasapis, S. Effect of conformation and molecular weight of co-solute on the mechanical properties of gellan gum gels. Food Hydrocolloids, (1998) 12,283 - 290. Gibson, W. Gellan gum. In, Thickening and gelling agentsfor food. Ed A. Imeson, Blackie Academic & Professional, Glasgow (1992) pp 227 - 249. Sanderson, G.R. and Clark, R.C. Gellan gum, a new gelling polysaccharide. In Gums and Stabilisersfor the Food Industry 2, (1984) eds. G.O. Phillips, D.J. Wedlock, & P.A. Williams. Pergamon Press, Oxford, pp 201-210. Grazdalen, H. and Smidsrd, 0. Gelation of gellan gum. Carbohydr. Polym., (1987) 7,371-393. McCance and Widdowson’s The composition of Foodr Fifh Edition, RSC and MAFF, 1991 Sworn, G., Sanderson, G.R. and Gibson, W. Gellan gum fluid gels. Food Hydrocolloids, (1 995) 9,265 - 271.

EFFECT OF SUCROSE ON MILK PROTEIN, LBG AND THEIR INTERACTIONS

C. Schorsch, M.G. Jones and I.T. Norton Unilever Research Colworth, Shambrook, Bedford MK44 lLQ, UK

ABSTRACT Polysaccharides are widely used in the dairy industry as stabilising, thickening and gelling agents and locust bean gum (LBG) is one of the most commonly used. Together with the milk proteins, they form a pseudoternary milk protein-polysaccharide-water solution. The aim of this study was to build up a more detailed thermodynamic and kinetic understanding of molecular interactions between milk protein and LBG and to look especially at the effect of sucrose. Ternary phase diagrams in absence or in presence of sucrose (20, 30 or 40% wlw) were established and compared. For such a system, a very high degree of incompatibility has been demonstrated leading to thermodynamic separation into two distinct phases, one rich in protein and the other in polysaccharide. The similarity of a Skimmed milk (SMP/LBG and a micellar casein/LBG system suggests also that the incompatibility between skimmed milk and LBG is largely attributable to the incompatibility of micellar casein and LBG. Sucrose appears to have only a modest influence on the thermodynamics of these mixed systems, leading to a concentration effect within the protein phase and a dilution within the locust bean gum phase. Data on the effect of sucrose on the individual components suggests, however, an improved solvent effect for LBG. It has been clearly demonstrated that the microstructure of these systems is driven by the phase volume ratio.

1 INTRODUCTION The stability of emulsions involving casein and LBG is likely to be controlled by the nature of the protein-polysaccharide interactions. Any tendency towards proteinpolysaccharide incompatibility in a dairy emulsion will lead to instability. The aim of this study is to build up a more detailed thermodynamic and kinetic understanding of the molecular interactions between milk proteins and LBG and to look especially at the effect of high sucrose content. Milk contains -3.3% wt protein, of which 2.6% is in the form of casein components’.’. The casein in cows’ milk occurs in the form of a colloidal dispersion which is responsible for the high turbidity of skimmed milk. The particles of this dispersion range in size from 20 to 600 nm, and are generally referred to as casein micelles3. Their dry matter contains about 93% casein, the remainder being inorganic material of which calcium and phosphate are the main components. Although the detailed structure of casein micelles is

High Solid Systems

293

not yet well known, mainly because of their amorphous character, there is an accepted view that the particles are sterically stabilized by one of the four main casein component proteins, i.e K-casein. The hydrophobic part of this molecule functions as an anchoring block, while the more hydrophilic part provides mainly steric stabilization. In consequence casein micelles are often described as “hairy” particles4“. LBG is one of the most commonly used stabilisers. It is a linear polymer of (1-4)-linked p-D-mannose residues substituted with (1-6)-linked a-D-galactose. The average galactose-mannose ratio is close to 1:4. 2 MATERIALS AND METHODS

2.1 Materials Skimmed milk powder (SMP) used was directly provided by Express Foods Ingredients (UK). The micellar casein (PCN) used was a native calcium phosphocaseinate sample prepared from raw milk by tangential membrane microfiltration followed by purification through ultrafiltration and LBG was provided by Meyhall. Its intrinsic viscosity at 5OC was 15.4 dug. Milk buffer solution or synthetic milk ultrafiltrate (SMUF) was prepared according to the method proposed earlier by Jenness and Koops lo.

2.2 Methods 2.2.1 Preparation of systems. LBG solutions were prepared by dispersing LBG in a synthetic salt ultrafdtratel0 at 80°C while stirring with a paddle mixer for 30 minutes. Micellar casein dispersions were obtained by dispersing the appropriate powder in the milk salt buffer while stirring with a paddle mixer (20 minutes at 6OOC). The sucrose was then added, and mixing was continued at this temperature for 15 minutes. Ternary micellar casein/LBG/sucrose systems with different compositions were prepared by mixing different solutions of each biopolymer for half an hour at ambient temperature.

2.2.2 Preparation of casein micelles depleted in calcium phosphate. Micellar casein depleted in micellar calcium phosphate were prepared by the addition of EDTA (40mM) to partly dissolve the micellar calcium phosphate. A known amount of EDTA was pipetted into a dialysis bag which was then placed in a volume of PCN dispersion equal to 20*volume of EDTA solution. After stirring for 6 hours at ambient temperature, the dialysis sac was cut open and the contents allowed to dilute the micelle dispersion. These diluted subsamples containing EDTA (200 ml) were each placed in a dialysis bag, dialysed against a very much larger volume of synthetic milk ultrafiltrate (SMUF) depleted in calcium. 2.2.3 Phase diagrams. Phase diagrams of the ternary systems were established at pH 6.8, at an ionic strength of 0.08, and at 5”C, after centrifugation at 1,100 g for between 2 to 6 hours as required. The phase volume method” was used, i.e. the composition of the phases was obtained by calculating concentrations from volume considerations. 2.2.4 Confocal microscopy. Confocal Laser Scanning Microscopy was used to visualise the microstructures of the systems. Rhodamine B (0.01% w/w) was dispersed in

294

Gums and Stubilisers for the Food Industry 10

the protein solution in order to stain the casein micelles and increase the contrast. Images were acquired using a Biorad MRC 600 confocal laser microscope.

2.2.5 Average molecular weights and diameters of casein micelles were determined by turbidity. Turbidity experiments were carried out using a UV-2101 (Shimadzu) spectrometer. Spectra were recorded between 400 and 800 nm at a temperature of 5°C. The refractive index and the refractive index increments were measured using a RFM340 (Bellingham-Stanley Ltd) refractometer and, within the limits of experimental error, the refractive index increment remained unchanged in the presence of sucrose (0.160 cm3/g). The turbidity (T) per unit concentration d c is linked to the molecular weight M, of the casein article and the second virial coefficient A2 via the relationship given by Doty and SteinerP2 : H*Q*c / T = (l/M,)+ 2A2Qc for h>>Rg where H is the optical constant, expressed here-after: 32.n3.n 02.(dn/dc)2 H= 3.Na .A4 b,dn/dc, N, and h are respectively the refractive index of the solvent, the refractive increment of the particle, the Avogadro number and the wavelength of light in vacuum. M, is the mass-average molar mass (in g.mol-’) and c is the concentration expressed in g/cm3. In the present work, the wavelength dependence of the turbidity “dlogddlogh” was also used to detect the onset of phase separation in casein and LBG systems.

2.2.6 Light scattering. The effect of sucrose on the molecular characteristics of LBG was studied using a series 4700 Photon correlation spectrometer (PCS) (Malvem Instruments, UK). 3 RESULTS 3.1 Description of the micellar casein/LBG phase diagram in absence of sucrose

Figure 1 illustrates the main features of the phase diagram of micellar casein/LBG at 5”C, at 0.08 ionic strength, at pH 6.8 and in the absence of sucrose. The characteristics of this phase diagram are the following; (i) the binodal is very close to both axes and therefore the compatibility domain is very small. (ii) the tie-lines impart high asymmetry to the diagram, i.e they show that mixed solutions of these biopolymers separate into phases differing greatly in the concentrations of the two macromolecular components. These characteristics are all consistent with high micellar casein - locust bean gum incompatibility. Confocal microscopy was used to look at the microstructure of systems with varying phase volume ratio. As illustrated in Figure 1, phase separation has a clear impact on microstructure with this latter depending on the phase volume ratio. As for other system^'^-'^, on mixing, the equilibrium phases of an incompatible polymer solution behave like water and oil, i.e they form an unstable water-in-water “emulsion”. In the case of the 75/25 ProteidLBG phase volume ratio, LBG constitutes the dispersed phase

High Solid Systems

295

in a protein continuous phase. For the 25/75 ratio, proteins are dispersed in the LBG continuous phase. The 50/50 ratio coincides with a bi-continuous microstructure.

W A L B G / 50% PR

25% LBG I7SK PR

MICELLAR CASEIN(WEYO)

Figure 1: Phase diagram of micellar casein-LBG systems at 5 T and microstructure of these systems depending on the phase volume ratio @H-6.8, I-0.08, image length scale, 500p).

3.2 Effect of sucrose on micellar caseidLBG systems Figure 2 displays the different phase diagrams of micellar casein/LBG/sucrose for sucrose levels ranging fiom 0 to 40% w/w.

m9ra 0

0

2

4

8

8

m

P

-m

A

s ) % -

x

0%-

#

%

-

P

MICELLARCASEIN (WF!!4)

Figure 2: Phase diagram of micellar casein-LBG systems at 5 T depending on sucrose concentration (0, 20, 30 and 40% wh).

From a general point of view, these phase diagrams appear very similar in terms of the very high polymer incompatibility exhibited but, increasing the sucrose content fiom 0 to

296

Gums and Stabilisersfor the Food Industry 10

20% wlw, leads to a slight shift of the binodal, consistent with an increase in the compatibility domain. Further increase of sucrose has little effect on the binodal but does lead to a change of slope of the tie-lines consistent with a concentration effect on the protein phase and, in contrast, a dilution effect on the polysaccharide phase. To understand the mechanism underlying the phase separation process, different techniques such as turbidimetry but also rheology and microscopy (confocal scanning laser microscopy and transmission electron microscopy) have been combinedI6. The turbidity study at 5°C has shown that the presence of LBG in the compatible domain changes the solvent quality from good to bad. At the same time, at infinite dilution, no effect on the molecular weight, or the size of the casein micelles, was observed (see Figure 3). This demonstrates that there are no attractive interactions between the components. We have shown that the formation of large heterogeneities with different refractive index compared to the one of the surrounding medium provokes a dramatic modification of both the turbidity and its wavelength dependence. As an example, at a concentration of 0.1 % wlw LBG, formation of domains due to a phase separation process has been established in the protein concentration range investigated. The concentration of casein required for this formation appeared to be slightly higher in the presence of sucrose. Nevertheless this increase of the heterogeneity of the systems is the result of the incompatibility leading to the formation of two phases, one enriched in casein the other one enriched in LBG.

X

PRalone X

'

x

f: *

O.OOE+dw 0.00

.'

1

I

0.02

0.04

X

PR + 0.06 % LBO

I

0.00

I

0.08

1

0.10

Protein (glcm3)

Figure 3a: Evolution of the jirnction H W r at 800nm with the concentration of casein dispersion in absence of LBG and in presence of 0.05 (0) or 0.I %(w/w) (4, LBG at

(w

597

291

High Solid Systems Bulk p h u

-9

repantbn

3.3 Comparison with SMP/LBG systems The phase diagram of SMPILBG in 20 % w/w sucrose at pH 6.8 and at 5°C is illustrated in Figure 4. These experiments have been carried out to verify the possible effect of the additional components in skimmed milk as for instance the one of the whey proteins.

0

10

20 30 Skimmrd Milk Powder %

40

50

Figure 4: Phase diagram of SMP-LBG-20% w/w sucrose systems at 5 @ and microstructure of these systems depending on the phase volume ratio CDH-6.8, I-0.08, image length scale, 5 0 0 p ) .

298

Gums and Stabilisers for the Food Industry 10

Once again, due to thermodynamic incompatibility, the systems undergo liquid-liquid phase separation with the macromolecular components concentrating in different phases. This incompatibility is in agreement with previous results and for instance especially explains the fact that LBG was patented as early as 1935 for concentratingmilk proteins". If we now compare this phase diagram with the previous one established using using native phosphocaseinate in the presence of 20% sucrose the similarity is evident. This suggests that the incompatibility between skimmed milk and locust bean gum is directly attributableto the incompatibility of micellar casein and locust bean gum. Nevertheless, it seems that the use of SMP instead of PCN leads to a very slight increase in the compatibility of the system, indicating that the influence of the whey proteins has to be more carefully investigated too. Except at very low concentrations of LBG, andor micellar casein, the ternary systems are in the form of water-in-water emulsions. Three different types of microstructure have been observed, and phase inversion occurs at a 50150 phase volume ratio. 3.4 Comparison with submicelleLBG systems It has been previously demonstrated that micellar casein and locust bean gum are highly incompatible and that sucrose has only a very slight effect on the thermodynamics of the system. The alternative issue which has been investigated was to decrease quite significantly the molecular weight and the size of the casein micelle and to look at its impact on the thermodynamics behaviour. Therefore the phase diagram of submicelle/LBG has been established (see Figure 5) at a temperature of 5"C, pH 6.8 and ionic strenght of 0.08M.

PROTEIN (%WT)

Figure 5: Phase diagram of submicellar caseirdLBG system at 5 97 and microstructure of these systems depending on the phase volume ratio @H-6.8, I-0.08).

299

High Solid Systems

The shift of the phase diagram is very clear suggesting a higher degree of compatibility. Asymmetry of the phase diagram is also lower than that of micellar caseidLBG system.

3.5 Sucrose effects on the individual component Due to these small changes, it was valuable to study more carefully the effect of sucrose on the individual components of the system.

3.5 1 Sucrose eflect on the casein micelle. The effect of sucrose on casein micelles has been studied using turbidity. The molecular weight of the casein micelles was obtained through straight line extrapolation of the H*c/r function vs. c to infinite dilution in the dilute regime (Md) (concentrations below 1% w/w) and in the semi-dilute regime (Map) (concentrations between 2 and 5% w/w), as illustrated in Figure 6.

-1

P

Figure 6: Concentration dependence of the function H*c/r at 800nm of casein dispersions at 20 T

Figure 7 shows how these molecular weight estimates change with sucrose concentration. In the semi-dilute regime, the “apparent” molecular weight is mostly constant with only a slight decrease occurring as sucrose is added. This figure highlights the difference between Md and M, for the casein particles contrasting with what occurs in the 0-20 % wlw sucrose range to behaviour at higher sucrose contents. In the absence of sucrose, the two molecular weights obtained were Md-2*108 Da and MW-7*1O8 Da, while at 40 % w/w sucrose these had very similar values around 5*108 Da. The discrepancy at low sucrose concentrations is the result of a combination of two processes; (i) a dissociation of individual casein molecules from the micelle to the serum phase which is strongly temperature dependent, i.e. is particularly prevalent at low temperature where the hydrophobic bonds are quite weak and part of the casein

Gums and Stabilisers for the Food Industry I0

300

molecules, especially the p-casein, dissociates from the micelles’8-20 and (ii) an aggregation of the micelles with increasing concentration which is less temperature dependent. Thus, in the absence of sucrose and up to 20% wlw sucrose content, the micelle molecular weight increases with increasing protein concentration due to aggregation, although there is no obvious effect on the particle size. This suggests that the micelles are more compact in the aggregated form. At higher sucrose contents (30-60% wlw), the sucrose seems to reduce and ultimately inhibit this aggregation process, possibly by increasing the hydrophilicity of the proteins, by forming a hydrophilic layer around the micelle, as already suggested for other

-

I

T

T

I

.... .......................

T’......... ‘i

*

.................

1

1 0

x)

P 30 40 Sucrose % wlw

9)

B)

Figure 7 : Evolution of the molecular weight of casein dispersions at 5 97 in presence of sucrose.

3.5.2 Sucrose effect on LBG. The effect of sucrose on the molecular characteristics was studied using the Malvem PCS. Extrapolation to zero angle leads to values of the apparent M, about 4 times lower than in absence of sucrose, whereas the radius of gyration decreases to a lesser extent25.This decrease in M, is likely to be due to reduction of mannan-mannan interactions between macromolecules, these interactions involving hydrogen bonds. It is highly probable that as the sucrose content increases, the solvent quality improves and the equilibrium is substantially shifted towards the formation of sucrose-mannan interactions.

4 DISCUSSIONS Quantitative analysis of LBG-casein sucrose phase diagrams was attempted using the second virial approach of Edmond and E g ~ t o in n ~order ~ to understand better the sucrose effect. In this treatment, the sucrose was considered part of the solvent, and tie-lines

30 1

High Solid Systems

formulated in terms of the two polymer molecular weights, and the three second virial coefficientsdescribing polymer-solvent, and polymer-polymer, pairwise interactionsZS. When treating the tieline data at different sucrose levels, two procedures were adopted. One was to fix the casein parameters at the values obtained from the previous treatment, and calculate best LBG parameters, and the cross polymer interaction coefficient. The other was to hold the LBG parameters constant at the zero sucrose results already assumed, and repeat the same type of calculation for the sucrose-containing systems. The first of these approaches gave a trend towards an increase of the LBG virial coefficient (and indeed the polymer-polymer interaction virial coefficient) increasing as sucrose content increases (solvent quality improving), and a decrease for the LBG molecular weight. These effects are most clearly evident above 20% sucrose. The conclusion about improving solvent quality for LBG, and a fall in its average molecular weight, are broadly in agreement with previous light scattering results. The alternative mode of calculation, by holding the LBG parameters constant, not surprisingly, gave a different explanation for the phase diagram changes on adding sucrose, i.e. that they arise from decreasing solvent quality for the micellar casein. This, however, seems to run against the experimental findings on casein aggregation discussed earlier, and seems a less lausible option than sucrose influencing the LBG. Recently, Antipova and SemenovaZ6, in the case of proteins and polysaccharides mixtures, have shown an increase in the co-solubility of the biopolymers with increasing sucrose concentration. It was established that the increasing cosolubility of the biopolymers occurs in accordance with an increase in the protein solubility in the aqueous medium due to sucrose addition.

P',

5 CONCLUSIONS

The present results clearly show the low compatibility of milk protein5BG at 5'C, at an ionic strength of 0.08, pH 6.8 and at varying sucrose content, leading to bulk phase separation. The ternary systems SMP/LBG/sucrose or micellar casein/LBG/sucrose systems behave as water and water emulsions, the microstructure being clearly driven by the phase volume ratio. This phase separation and incompatibility is similar to that usually observed between two non interacting random coil polymers at sufficiently high concentrations. Depending on the phase volume ratio, the transient microstructure of the water-in-water system is significantly different. Sucrose appears to have only a modest influence on the thermodynamics of these mixed systems, and fits to the phase diagrams using second virial coefficient theory suggesting an improved solvent effect for the LBG (supported by independent data from light scattering). The dominant features of the phase behaviour of the present system remain, however, high polymer incompatibility and tieline asymmetry, the latter being the principal element influenced by the sucrose. Interestingly, the alternative theoretical approach of employing Flory-Huggins theory to describe the current data was not successful at all, in consequence no doubt about the non-random coil polymer character of the casein micelles. Another theory to be considered in the future is that of depletion-flocculationZ8> 29.

302

Gums and Stabilisersfor the Food Industry 10

Acknowledgements The authors would like to thank D.Ferdinando for carrying out the confocal microscopy work, A.H.Clark the quantitative analysis of the phase diagrams and P.Schuck (INRA Rennes - France) for providing the phosphocaseinate sample. References 1. P. Walstra and R. Jenness, ‘Dairy Chemistry and Physics’, J. Wiley and Sons (Eds.), New York, 1984. 2. P. Walstra, J.Duiry Sci., 1990,73, 1%5. 3. P.F. Fox and D.M. Mulvihill, Food Gels, P.Harris (Ed.), 1990, Chapter 4. 4. D.S.Home, J.Colloid and Interface Sci., 111 (1986) 250. 5. C.G. De Kruif, Th.J.M. Jeurnink and P. Zoon, Neth.MilkDairy J., 1992,46, 123. 6. C.G. De Kruif and E.B. Zhulina, Colloids and Surfaces A: Physicochemical and Engineering aspects, 1996,117, 151. 7. A. Pierre, J. Fauquant, Y. LeGraet, Y., M. Piot and J.-L. Maubois, Lait, 1992, 72, 461. 8. P. Schuck, M. Piot, S. Mejean, Y. Legraet, J. Fauquant, G. Brule and J.-L. Maubois, Lait, 1994,74,375. 9. P. Schuck, M. Piot, S. Mejean, J. Fauquant, G. Brule and J.-L. Maubois, Lait, 1994, 74,47. 10. R. Jenness and J. Koops, Neth. Milk and Dairy J., 1%2, 16, 153. 11. V.I. Polyakov, V.Ya. Grinberg and V.B. Tolstoguzov, Polymer Bulletin, 1980, 2 , 757. 12. P. Doty and R.F. Steiner, J.Chem.Phys., 1949, 17, 1211. 13. V.B. Tolstoguzov, ‘Functional properties of food macromolecules’, J.R. Mitchell and D.A. Ledwards (Eds), Elsevier Science Publishers, London, 1986, p.385. 14. A. Syrbe, P.B. Fernandes, F. Dannenberg, W. Bauer and H. Klostermeyer, ‘Food Macromolecules and Colloids’, Royal Society of Chemistry, Cambridge, 1995, p.328. 15. C.R.T. Brown, T.J. Foster, I.T. Norton and J. Underdown, ‘Biopolymer mixtures’, S.E. Harding, S.E. Hill and J.R. Mitchell (Eds.), Nottingham, 1995, p.65. 16. C. Schorsch, M.G. Jones and I.T. Norton, FoodHydrocolloidr, 1999, 13, 89. 17. A.S. Ambrose, USPatent 19, 1935, 91189. 18. D. Rose, J.Duiry Sci., 1%8,51, 1897. 19. W.K. Downey and R.F. Murphy, J.Dairy Res., 1970,37,36 1. 20. D.G. Dalgleish and A.J.R. Law, J.Dairy Res., 1988,55,529. 21. W.P. Jencks, ‘Catalysis and Chemistry and Enzymology’, McGraw-Hill (Ed.), New York, 1969, p.254. 22. I.M. Garrett, R.A. Stairs and R.G. Annet, JFoodSci., 1988,71,10. 23. P. Chinachotu and M.P. Steinberg, JFoodSci., 1988,53,932. 24. E. Edmond and A.G. Ogston, BiochemJ, 1968,109,569. 25. C.Schorsch, A.H. Clark, M.G. Jones and I.T. Norton, Colloids and Surfaces B: Biointerfaces, 1999,12,3 17. 26. A.S. Antipova and M.G. Semenova, Carbohydr. Polym., 1995,28,359. 27. A.S. Antipova and M.G. Semenova, Food Hydr., 1997,11,71. 28. P. Jenkins and M.Snowden, Advances in Colloid and Interface Science, 1996,68,57. 29. R. Hoskins, I.D. Robb, P.A. Williams and P. Warren, J. Chem. SOC.Faraday Trans., 1996,92,45 15-4520.

GLASS TRANSITIONS IN HIGH SUGAR/BIOPOLYMER MIXTURES - SOME RECENT DEVELOPMENTS

Stefan Kasapis and Insaf M. A. Al-Marhoobi Department of Food Science & Nutrition, College of Agriculture, Sultan Qaboos University, PO Box 34,Al-Khod 123, Sultanate of Oman

1 INTRODUCTION

Previously, we reported on the changing nature of polysaccharide networks induced by adding sugar to an aqueous preparation’. n u s , the commercially important polysaccharides of K-carrageenan, high methoxy pectin and low acyl gellan were mixed at normal levels of use (5 1%) with co-solute, with the total level of solids ranging between 5 and 87%. Sugar concentrations from 5 to 40% shift the onset of network formation to higher temperatures and create stronger structures than the aqueous counterparts*. The consolidation of the solid-like character was reflected in the enthalpic content of calorimetric transitions which also rose, thus arguing for enhanced macromolecular order in the mixture. Despite the increasing hydrogen bonding between sugar and water molecules, the plethora of the latter stabilise the additional polysaccharide aggregates by means of a surrounding hydration laye?. Addition of co-solute in excess of 70% sees a dramatic reduction in the rigidity of networks which, however, form at the highest experimentally accessible temperature (90°C)4. Large deformation compression testing demonstrates their elastic properties as opposed to the brittle nature of the aqueous preparations (fracture requires 0.6 and 0.25 units of strain, respecti~ely)~. The drop in mechanical strength is accompaniedby a decline in the magnitude of enthalpic transitions, a result which suggests polymeric cobwebs with reduced order6.It appears, therefore, that the saturating levels of co-solute limit the availability of water molecules to the polysaccharide thus preventing excessive order, with the chain segments remaining mainly in the disordered form’. Cooling of the samples further verifies the entropic nature of the networks, since they undergo vitrification according to the pattern recorded for amorphous synthetic polymers and diluted systems’. The free-volume theory, first proposed by Williams, Landel and Feny,was applied to the data to yield estimates for the glass transition temperature (T,), the fiactional fiee volume at T, and the thermal expansion coefficient of the sugar/polysaccharide mixturesg. The transformation from the aggregated enthalpic to lightly cross-linked, entropic type of networks was clearly demonstrated at the intermediate levels of co-solute (40 to 70%) where a ‘composite’ structure was formed with a single poly~accharide~. In the present communication, we extended the work to high sugar agarose mixtures, non-gelling polysaccharides (guar and locust bean gum), and pigskin gelatin. It seems that certain patterns of behaviour are generic and should be used as guidelines for the prediction and control of vitrification phenomena in these systems.

304

2

Gums and Stabilisers for the Food Industry I0

MONITORING THE VITRIFICATION OF HIGH SUGAFUGELATIN MIXTURES WITH THE TECHNIQUE OF DYNAMIC OSCILLATION

Over the years gelatin has proved to be one of the most popular hydrocolloids for use in the food industry, which has funded extensive research programs for the elucidation of the gelation mechanism of the protein. It has been accepted, that helix formation involves a nucleation process bringing together three strands of at least two different molecules". Once formed the helix grows rapidly, until for example the end of the strand is reached". Gelatin networks are rubbery with brilliant clarity but at subzero temperatures ice formation is unavoidable. This is, of course, an anathema for applications in the ice cream, chocolate and confectionery industries and water is replaced partially with sugar in the formulation. Typically, soft confectioneries like 'jelly babies' or 'gummy bears' contain 5% gelatin with sugar at a total level of solids that reach 85%. Within this context, therefore, we prepared a sample of 5% acid pigskin gelatin with 50% glucose syrup (dextrose equivalent of 42) and 30% sucrose, and investigated its structural properties. Figure 1 illustrates the small deformation properties of the sample recorded using the technique of dynamic oscillation. The experim.enta1 routine involved cooling of the sample to -55°C at a rate of 2OC min-', followed by heating at the same scan rate. The heating run is an approximate equilibrium process and it is the one reproduced in Figure 1. Both the solid and liquid-like components of the network are noted (G' and G", respectively), as is their ratio (tan 6 = G / G ) . At the upper range of temperature, the network shows a predominantly solid-like character, with the values of tan 6 falling short of 1. Below 5"C, however, there is a change in the state of the mixture which behaves as a very viscous solution (tan 6 > 1). The magnitude (from lo4 to lo9 Pa) and the overall form of the trace are akin to the glass transition recorded for amorphous synthetic polymers and high sugar/polysaccharide system^'^"^. Finally, there is another development at temperatures below 4 5 ° C with the solid-like component becomin dominant once more and approaching values of 10" Pa. This is known as the glassy state? ..I The development of shear moduli was also followed as a k c t i o n of frequency of oscillation which reflects changes in viscoelasticity within the experimental time of observation (1 Hz = s-'). In doing so, mechanical spectra were recorded between lo-' and 10' rad s-l at constant temperature intervals of four degrees during heating of the sample from the glassy state. Then, a reference temperature (To) somewhere within the range of the experiments was arbitrarily chosen (in this case, 9OC). The mechanical spectrum adjacent to that of the reference temperature was shifted horizontally along a logarithmic frequency axis until both of them fell into a single smooth curve. The process was repeated for all the data thus yielding the composite curve of Figure 2. Shift factors (aT) were obtained which gave the horizontal distance between adjacent curves. Clearly, a mechanical analogue of the vitrification process recorded as a function of temperature in Figure 1 has been obtained by extrapolating the frequency data at a constant temperature (9°C). This is the time-temperature superposition principle (TTS), otherwise known as the method of reduced variables, which advocates the occurrence of a second order thermodynamic transition in the ~ y s t e m ' ~ "In~ .other words, the effect of heating the high sugar/gelatin mixture is equivalent to decreasing the frequency of measurement at a constant temperature. It follows that, shear moduli measured at frequency o and temperature T are equivalent to those measured at frequency maT and temperature To. The same values of aTwere able to superimpose the traces of storage and loss modulus, a result which argues that the time-temperature effect is similar for both the long and short retardation processes in our mixture.

305

High Solid Sysrems

-55

-45

-35

-25

-15

-5

5

15

Temperature ("C) Figure 1 Heating runs of storage and loss modulus for a sample of 5% pif .skin gelatin in the presence of 5?? glucose syrup and 30% sucrose (frequency: 1 rad s- , strain: 0.1%; 2°C min-I). scan

I

-1

1

3

5

7

9

11

13

Log (fiequencyhad s-') Figure 2 Composite curve of shear moduli for the sample of Figure 1. Frequency sweeps were obtained at constant temperature intervals of four degrees, covered the temperature range between 9 and -55"C,and were superimposed at 9°C.

Gums and Stabilisers for the Food Industry 10

306

3

ESTIMATION OF THE GLASS TRANSITION TEMPERATURE IN HIGH SUGAR BIOPOLYMER MIXTURES

The single set of shift factors emanated from the composite curve of Figure 2 allowed us to monitor the temperature dependence of both shear moduli using an empirical expression which has been proved to be widely applicable': log aT= G(T)/G(T,)

=-

CP(T - T,)/(C;+ T - To)

where, CP and C; are constants. This is known as the WLF equation which in the form of (T - T,)/log aT against T - To yields values for the two constants from the slope and intercept of the resulting straight line. The Vogel temperature (Tm)is then employed since it is related to the glass transition temperature in the following form': T, - 50 = T,= T, - C;

(2)

The value of T, predicted for the high sugar/gelatin sample is -46"C, which is close to the second cross over of modulus traces in Figure 1 (- -44°C). Similarly, there has been a coincidence between the experimental and predicted values of the glass transition in high sugar polysaccharide mixtures shown in Table I. Dynamic mechanical thermal analysis (DMTA) on high solids amylopectin samples has often used the maximum value of tan 6 as a marker for T,'7*'8,which occurs early during vitrification (at -21.5"C in Figure 1). In terms of the WLF approach and our experimental observations, however, the T, is pinpointed at the onset of the glassy state where configurational rearrangements are of solid-like character (G' overtakes G ) . Recent calorimetric work sug ested that small amounts of polymeric additives have no effect on the T, of sucrose2!f . However, these were mixtures containing poly(vinylpyrrolidone) (FVP), a compound that bears no physicochemical affinity to a polysaccharide. Predictions of T, were made with the Cordon-Taylor equation which requires ingredient compatibility at the molecular level". This is not, of course, the case for the PVP-sucrose mixture, as opposed to a sugar-water-polysaccharide preparation. Judging from Figure 6 of reference 19, the molecularly incompatible ingredients phase separate and there should have been two T,s, one for PVP and another one for sucrose. This may explain the lack of fit between the single T, values reported for the various PVP-sucrose compositions and the predictions of the Cordon-Taylor equation. Dynamic oscillation measurements have now demonstrated that addition of small amounts of polysaccharide to a sugar preparation affect dramatically the vitrification properties of the mixture. Figure 3 shows that the cross over of shear moduli for 83% glucose syrup is pinpointed at -25.3OC. Replacement of 1% glucose syrup with 1% guar gum shifts the glass transition to higher temperatures with the cross over occuning at -19.7OC. We didn't observe real differencesin the T, values between guar and locust bean gum, with small deviations lying between the experimental error of the measurement (Table I). Galactomannan are non-gelling polysaccharides and were replaced in the mixture with 1% ion-exchanged K-carrageenan in the sodium form. A small amount of KCL (5 mM) was also added to introduce a lightly cross-linked cobweb in support of the glucose syrup phase. It appears that the lattice of the polysaccharide network has further immobilised the high solids system thus leading to an early vitrification (Tg= -1OC). A sample of 50% sucrose and 35% glucose syrup crystallizes at subzero temperatures, but

c

deacylated gellan; s.

3

*The cooling rate was 1°C/min

gellan I

glucose syrup

-

the composition of glucose syrup refers to dry solids

'The cooling rate was 2'C/min

sucrose; g.s.

Parameters Characterising the Temperature Dependence of aT for the Sugar/Biopolymer Systems

TABLE I

Gums and Stabilisers for the Food Industry 10

308

9.0

8.5 8.0 7.5 7.0

6.5 -35

I

I

I

I

-25

-15

-5

5

Temperature ("C) Figure 3 Temperature variation of G' (solid line) and G' (dashed line) for 83% glucose syrup (left spectrum), 82% glucose syrup plus 1% guar gum (middle spectrum), and 82% glucose syrup plus 1% K-carrageenan in the presence of 5 mM KCl (cooling rate: 2°C min-'; frequency: 1 rad s-').

the immobility imparted to the mixture on addition of polysaccharides prevents crystallisation, and vitrification ensues (e.g. 0.5% gellan in the presence of 50% sucrose and 35% glucose syrup vitrifies at -26°C on Table I). The consequences of immobility have been argued for many years by Franksz1.For sucrose crystallisation to occur, solute molecules must move out of the way but the mobility of macromolecules in the glassy state is far too low to allow this". Furthermore, phase separation between the polysaccharide and glucose syrup wouldn't alter the T, of the latter, which at 83% solids occurs at -25.3"C (Figure 3 and Table I). It appears, therefore, that there is a continuum of interactions amongst co-solute, water and polysaccharide molecules which accelerate vitrification phenomena by slowing down configurational vibrations. This is further reinforced by the formation of a macromolecular cobweb in the mixture shown for the gelling polysaccharides in Table I. 4

THE FUNCTIONS OF TEMPERATURE AND TIME IN HIGH SUGAR BIOPOLYMER MIXTURES

The cooling run of Figure 1 reflects the combined effect of temperature and frequency (time) of measurement on the viscoelastic profile of the high sugar/gelatin system. Instructive separation of the two parameters facilitates discussion of vitrification properties in terms of sample composition, chemistry and concentration of diluent, and molecular weight distribution of the ~ o l y m e ? ~Use . of the shift factors allows determination of the change in relaxation times when the temperature is changed from To to T. Figure 4 illustrates the temperature h c t i o n of sugar preparations in the presence of

309

High Solid Systems

Temperature (“C) -60

-20

-40

-20

-10

0

20

10

0

40

20

60

30

Temperature (“C) Figure 4 Development of shift factors as a function of temperature for 82.5% glucose syrup plus 0.5% guar gum (0), 82% glucose syrup plus 1% guar gum (o), 50% glucose SYNP with 30% sucrose plus 5% pigskin gelatin (A), and 82% glucose syrup plus 1% K-carrageenan with 5 mM KC1 (0). Exponentials reflect the WLF fit.

-6

-4

-2

0

2

4

6

8

Log = 6) Figure 5 Variation of the relaxation distribution function with timescale of measurement for 0.5% K-carrageenan + 10 mM KCI + 85% glucose syrup, 0.5% gellan + 7 mN Ca2’+ 50% sucrose + 35% glucose SYNP,0.8% gellan + 40% sucrose + 40% glucose SYNP, 1% high methoxy pectin + 50% sucrose + 36% glucose syrup, 0.7% agarose + 50% sucrose + 35% glucose syrup, 85% glucose syrup, and 40% sucrose + 40% glucose syrup. Reference temperature: 0°C.

Gums and Stabilisers for the Food Industry 10

310

0.5 and 1% guar gum, 1% K-carrageenan, and 5% pigskin gelatin. In the case of the nongelling polysaccharide, data are restricted within the temperature range of the glass transition region (T,s on Table I are -22.2 and -19.7OC for 0.5 and 1% guar gum, respectively), and follow the predictions of the WLF equation. The experimental temperature scale for gelatin and K-carrageenan covers both the glass transition and the glassy state. Clearly the development of shift factors in the former can be modelled by the WLF equation which predicts T, values for the protein and the polysaccharide of 4 4 and -l0C, respectively. Increasingly, however, the WLF fit deviates from the temperature dependence of shift factors in the glassy state, a result that makes T, a true threshold at the conjunction of two different spectral contributions. The molecular process in the glass transition has been rationalised with the theory of free volumez4.Holes between molecules of the order of monomeric dimensions or smaller voids due to packing irregularities are considered to constitute free volume. Volumetric measurements have unveiled a linear relationship in the change of specific volume with temperature, with the gradient being the thermal expansion coefficient of the material (a,). A graph of the above shows a discontinuity at a narrow region which is taken to be the glass transition temperature. The free volume theory gives physical significance to the WLF equation by identifjmg the constants Cp and (2; in the following form': Cp = B/2.303f,

and

C;

= fda,

(3)

where, the fractional free volume, fo, is the ratio of free to total volume of the molecule and B is usually set to one. This allows estimation of a, and f,, the fractional free volume at T,, for the high sugar biopolymer mixtures shown on Table I. The inability of the combined WLF approach to follow the curve aT(Q below T,, indicates that the free volume mechanism is not of ovemding importance in the glassy state. Evidence so far argues that the glassy state is governed by a rate process where an energetic barrier should be overcome for local rearrangements to occur'. The time function can be surmised by the composite curve of Figure 2 which plots viscoelastic data against the frequency of oscillation. It appears, however, that the distribution function of relaxation times (0) is a more usefbl concept for comparing various systems from different laboratories because it can be readily obtained from dynamic or stress relaxation measurements. It is given by the following relationshipz5: 0 (-ln O) = G' (d log G'ld log 0)

=G

(1 - d log G'/dlog 0)

(4)

The real (G') and imaginary ( G ) parts of the complex shear modulus (G*)are independent at a single frequency, but they are now connected through the distribution function. Therefore, valid experimentation should yield the same value of 0 for both shear moduli. Logarithmic plots of 0 for seven high solids systems are shown in Figure 5 . When gelation is possible (e.g. K-carrageenan with K" or gellan with Ca2'), spectra level off at long timescales indicating the formation of a soft rubbery network whose structural knots appear to be permanent within the experimental constraints. In the absence of added Ca2', the curve of the gelladsugar mixture shows a steep portion at long times (r > lo3 s) which corresponds to the relaxation of segments approaching entire molecules in length. Upon removal of gellan, the sample of 40% sucrose plus 40% glucose syrup can only exhibit the mechanical analogue of vitrification at short times. Glass transition, in the form of rising G values, is noted for all samples, a result which emphasizes the universality of the mechanism of free volume regardless of the chemical composition of the system.

311

High Solid Systems

5

MOLECULAR ORDER AS SUPRESSOR OF VITRIFICATION IN HIGH SUGARlBIOPOLYMER MIXTURES

Work so far has demonstrated that macromolecular interactions have a role to play in a high solids environment by immobilising the sugar-polysaccharidemixture and changing its phase behaviour. Thus gelling polysaccharides introduce a rubbery network between the flow and the glass transition regions which on subsequent cooling prevents crystallisation of sucrose or accelerates the vitrification of glucose syrup. It soon transpired that vitrification could be halted by increasing the macromolecular order or reducing the total amount of solids in the system. Glass transitions were engineered in Figures 3 and 5 by adding 5 mM KCI to 1% K-carrageenan or 10 mM KCI to 0.5% K-carrageenan. As shown in Figure 6, however, addition of 20 mM KCl to 1% K-carrageenan results in higher G values than those of G throughout the temperature range of observation (from 90 to -3OOC). Clearly the abundance of polymer and counterions generates additional helical associations which substantially curtail conformational freedom, with the network showing a dominant solid-like component. Thus the glass transition region ( G > G ) is not observed and the WLF approach is unable to provide a basic function of behaviour for this system. 10

,

19.5

8.5

7.5

BC

8

Iy! 6-

h

6.5

M

s

3

-i T3

5.5

4.5

-30

-10

10

30

50

70

90

Temperature ("C) Figure 6 Temperature variation of G (closed symbols) and G (open symbols) for two gelatin and a K-carrageenan sample in the presence of various levels of glucose syrup (cooling rate: l0C min"; frequency: 1 rad s-I). Figure 6 also illustrates the cooling profiles of high sugadgelatin mixtures with 1.8 and 2.1 times more water than for the preparation of Figure 1, which exhibited vitrification properties. In the case of 8% gelatin, a cohesive network is formed approaching values of

312

Gums and Stabilisersfor the Food Industry 10

storage modulus of lo5 Pa. Reduction in the protein content to 3% results in phase inversion and the formation of a glucose syrup-continuous solution ( G > G‘). In both cases, however, there is a second step of structure formation at about -1 5”C, since solute molecules move out of the way thus allowing water to freeze and ice crystals to grow. Ice crystals are suspended within the viscous phase of glucose syrup or increase by two orders of magnitude the rigidity of the mixture with 8% gelatin. Previously, it was proposed that the kinetics of network formation govern the phase continuity, phase inversion and the mechanical strength of a low-solids composite gelz6.Similarly, sugar-water interactions at a total level of solids higher than 70% are stabilised by the presence of a gelling biopolymer thus preventing crystallisation or accelerating vitrification phenomena.

References 1. V. Evageliou, S. Kasapis and G. Sworn, ‘Gums and Stabilisers for the Food Industry 9’, The Royal Society of Chemistry, Cambridge, 1998, p. 333. 2. M. Papageorgiou, S. Kasapis and R.K. Richardson, Carbohydr.Polym., 1994,25, 101. 3. K. Gekko, H. Mugishima and S. Koga, Int. J.Biol. Macromol., 1985,7,57. 4. L.E. Whittaker, I.M. Al-Ruqaie, S. Kasapis and R.K. Richardson, Carbohydr.Polym., 1997,33,39. 5. G. Sworn and S. Kasapis, Food Hydrocolloids, 1998,12,283. 6. V. Evageliou, S. Kasapis and M.W.N. Hember, Polymer, 1998,39,3909. 7. I.M. Al-Ruqaie, S. Kasapis, R.K. Richardson and G. Mitchell, Polymer, 1997, 38, 5685. 8. J.D. Ferry, ‘Viscoelasatic properties of Polymers’, John Wiley & Sons, New York, 1980, p. 264. 9. M.L. Williams, R.F. Landel and J.D. Ferry, Journal of the America1 Chemical Society, 1955,77,3701. 10. G. Stainsby, ‘The Science and Technology of Gelatin’, Academic Press, London, 1977, p. 179. 11. D.A. Ledward, ‘Functional Properties of Food Macromolecules’, Elsevier, London, 1986, p. 171. 12. W. Dannhauser, W.C. Child, Jr. and J.D. Ferry, Journal of Colloid Science, 1958, 13, 103. 13. S. Kasapis, ‘Functional Properties of Food Macromolecules’, Aspen Publishers, Gaithersburg, 1998, p. 227. 14. A.V. Tobolsky, ‘Rheology - Theory and Applications’, Academic Press, New York, 1958, p. 63. 15. A.V. Tobolsky,J. Appl. Phys., 1956,27,673. 16. G. Allen, ‘The Glassy State in Foods’, Nottingham University Press, Nottingham, 1993, p. 1 . 17. M.T. Kalichevsky and J.M.V. Blanshard, Carbohydr.Polym., 1993,20,107. 18. M.T. Kalichevsky, E.M. Jaroszkiewicz and J.M.V. Blanshard, Polymer, 1993,34,346. 19. S.L. Shamblin, E.Y. Huang and G. Zografi, Journal of Thermal Analysis, 1996, 47, 1567. 20. M. Gordon and J.S. Taylor, J.Appl. Chem., 1952,2,493. 21. F. Franks, CtyoLetters, 1986,7,207. 22. J.R. Mitchell, ‘Functional Properties of Food Macromolecules’, Aspen Publishers, Gaithersburg, 1998, p. 50.

High Solid Systems

313

23. J.D.Ferry, ‘Proceedings of the Second International Congress on Rheology’, Butterworths, London, 1954, p. 140. 24. A. Kovacs, Adv. Polym. Sci., 1964,3,394. 25. J.D. Ferry and E.R.Fitzgerald, J. Colloid Sci., 1953,8,224. 26. S. Kasapis, S. Alevisopoulos, R. Abeysekera, P. Manoj, LS. Chronakis and M. Papageorgiou, ‘Gumsand Stabilism for the Food Industry 8’, IRL Press, Oxford, 1996, p. 195.

Proteins and Emulsions

RECENT ADVANCES IN PROTEIN INTERACTIONS

Nazlin K.Howell School of Biological Sciences University of Surrey Guildford Surrey GU2 5XH.

1. INTRODUCTION

Protein-protein interactions occur in many physical and chemical processes and affect the nutritional and organoleptic quality of food products during manufacture and storage. Most studies, to date, display interesting and technologically useful properties produced by protein interactions including: enhanced gelation properties by synergistic interactions; phase separation and new textural properties and aggregation of oppositely charged proteins.' Clearly, a khowledge of protein interactions can lead to a better understanding of the biochemical changes in food products during processing and stora e such as the aggregation of proteins in fish leading to toughening on frozen storage.5.3 Furthermore, an undertanding of the effect of protein structure on protein-protein interactions, for example of smooth and skeletal muscle proteins, permits the manipulation of protein side chains in order to enhance gelation proper tie^.^ However, the exact mechanisms that govern protein interactions still require investigation. Recent advances in spectroscopic techniques have facilitated these investigations, some of which are described in this paper. The aims of this paper are to review the current status on protein interactions; to examine in detail the study of hydrophobic interactions by Raman spectroscopy and to examine the effect of free radicals on protein cross-linking. 1.1. Synergistic interactions

'

Pioneering studies by Howell et al. 69 indicated synergistic interactions between globular proteins such as egg albumen, blood plasma and whey proteins ranging in MW 14000 - 65,000. These proteins, possibly due to their close evolutionary relationships and homology, showed remarkable compatibility and ability to form a single network. With plasma-egg albumen protein mixtures, synergistic interactions in terms of viscosity and gelation were observed; this resulted in greater gel strength, as observed on the Instron compression test, for the mixed systems compared with the two proteins in isolation. The interaction between plasma, egg or whey proteins depended very much on the temperature and time of heat treatment; the concentration of each protein in the mixture and the presence of salt or sugar. During heating, each protein unfolds partially,

318

Gums and Stabilisersfor the Food Industry 10

leading to a 'molten globule' state. The buried hydrophobic, sulphydryls and other reactive groups are exposed during denaturation and can participate in protein-protein interactions, leading first, to a reversible gel with non-covalent bonds and finally to a covalently-bonded non-reversible gel at temperatures over 80°C. It was reported that for plasma-egg albumen mixtures at 4% plasma : 2% egg albumen heated at temperatures of 85OC for only 15 min produced greatest synergy compared with heating at 90 and 95OC for 30 or 60 min. At the lower time-temperature treatment, the egg albumen proteins which unfolded first, interacted with less unfolded plasma proteins or its main protein, serum albumin. At very high temperature-time treatments, the plasma proteins probably interacted with each other.'

1.2. Phase separation Similar formation of singie networks has not been observed in mixtures of other proteins including soya, meat, soluble wheat protein and whey proteins or bovine serum albumin. Instead, these protein mixtures have resulted in phase separation because they do not associate due to physical and chemical differences for example, the soya isolate proteins are larger (MW 140,000-190,000 for 7s globulins and 300,000-400,000 for the 11S globulins) compared with the whey proteins (MW 14,000 for a-lactalbumin and 18,000 for 0-lactoglobulin). Fig. 1 indicates phase separation in a 10% (w/w) soya isolate and 10% (w/w) whey protein mixture. Phase inversion occurred as the concentration of the two proteins was altered. On the whole, the phase separated systems showed gel strength values lower than those of the individual proteins.' The only exception was when a small amount of one protein was added to a larger concentration of the second protein e.g. 20% soya isolate: 1% whey protein isolate which resulted in a greater than expected gel strength value.

Figure 1. Phase separation in a mixed soya (10% w/w) ) (dispersed large aggregates) and whey (10% w/w) (continuous phase of smaller proteins) gel as shown by phase contrast microscopy. Mixtures (1 :1) of salt-soluble meat proteins (SSMP) and soluble wheat proteins (SWP) also phase-separated when autoclaved at 12OoC,formed gels that were weak, brittle and prone to syneresis. However, with the addition of only 1% (w/w) SWP to the 20% (w/w) SSMP, the shear modulus was increased considerably and the gel properties were improved. This improvement may be due to a bridging effect of the small quantity of

Proieins and Emulsions

319

soluble wheat protein which appeared to bind the meat protein fibres transversely as viewed by transmission electron microscopy. The beneficial effects of adding small amounts of proteins to a large concentration of a different protein were also observed with mixtures of other proteins including soya-whey or SWP-bovine serum albumin.g* 1.3. Aggregation of proteins A third and different type of interaction between food proteins resulting in aggregation was published for the first time by Howell et al." in relation to studies on lysozyme, alactalbumin and P-lactoglobulin. Many technological applications for the interaction of lysozyme (PI 10.1) with whey proteins (PI 5 ) have been cited. For example, lysozyrne interacts with egg ovomucin and is involved in egg white thinning during storage.'* Electrostatic interactions of lysozyme with bovine serum albumin and P-lactoglobulin were reported to enhance foaming proper tie^.'^ Lysozyme also enhances the gelation of deamidated negatively charged gluten proteins.14 However, Howell et al." reported that the combination of lysozyme with either alactalbumin or P-lactoglobulin resulted in precipitation due to electrostatic interactions as judged by turbidity measurements, chromatography and molecular modelling. A similar interaction between a-lactalbumin and P-lactoglobulin was not observed. Molecular modelling studies, using interactive docking of the crystal structures, indicated that for the P-lactoglobulin-lysozyme interaction, the optimum visual fit could involve electrostatic interactions between Glutamate 35 and Aspartate 53 in the catalytic binding site on lysozyme and Lysines 138 and 141 at the dimerization site of plactoglobulin. Similar electrostatic interaction between a-lactalbumin and lysozyme is depicted in the molecular model shown in Fig.2.

Figure 2. Molecular model to show the electrostatic interaction between a-lactalbumin (left) and lysozyme (right). (Howell and Lewis, unpublished)

Gums and Stabilisersfor the Food Industry 10

320

Direct evidence for the presence of electrostatic interactions was obtained by Raman spectroscopy, which is well-suited to the study of concentrated solutions and gels, pastes and solids. Whilst this method has been reported for the investigations of protein gels, Howell ef al .I5, l6 used it for the first time to examine protein interactions in mixed gels. Detailed analysis of the aggregate by Raman spectroscopy in the 450-1900 cm-' region suggested not only electrostatic interactions but also hydrophobic interactions. Both the lysozyme-a-lactalbumin and lysozyme-f3-lactoglobulin (mixed 1:1 weight ratio) demonstrated the involvement of hydrophobic interactions by intensification of spectral bands assigned to CH (1336 cm-l) and CH2 (1456 cm-l) bending vibrations and decrease in the intensity of bands assigned to Trp residues (759 and 1336 cm") in a non-polar environment (Fig.3 ).Is

I

I

I

I

l

l

I

,

I

I

I

1500

,

I

I

1000

500

Wavenumber,

'

CM-'

Figure 3. Raman spectra in the 450-1900 cm" region of a 9OoC heated lysozyme and plactoglobulin mixture showing a) experimental spectra b) difference (experimental theoretical) spectra. The spectra were base-line corrected, smoothed and normalised to the intensity of the phenylalanine peak at 1006 cm-' (1 .O).

Proteins and Emulsions

321

Changes in the disulphide stretching vibrations of cystine residues as well as secondary structure changes involving a lowering of helix and sheet structures was also noted. On heating at 9OoC for 30 min, gelation of the individual proteins indicated changes in the aromatic and aliphatic CH groups characteristic of hydrophobic interactions as well as changes in the disulphide bonds. For heated and unheated mixtures of lysozyme with a-lactalbumin or P-lactoglobulin the experimental spectra were different from the theoretically calculated average spectra of the individual proteins confirming hydrophobic interactions in mixed gels. 1.4. Hydrophobic interactions The importance of hydrophobic interactions in the hctionality of proteins is welldocumented." They contribute to heat gelation of proteins as indicated above and often contribute substantially to the gel strength of heated gels on cooling.' Recent evidence, arising from studies undertaken by the author, indicates that hydrophobic interactions contribute to the toughening of frozen fish on storage. We have also demonstrated that the inferior gelling properties of smooth muscle myosin compared to skeletal myosin is partly due to an imbalance of ionisable and hydrophobic residues; it was suggested that the higher level of hydrophilic groups, particularly glutamic acid in smooth muscle myosin, disrupts the water structure around the hydrophobic groups which suppressed ge~ation.~ Hydrophobicity can be measured by chromatography or with fluorescent probes to monitor the aliphatic and aromatic groups of proteins. Commonly used techniques such as ultraviolet, fluorescence, near-UV circular dichroism or resonance Raman spectroscopy are useful for investigating aromatic amino acid residues in dilute solutions. However, methodology is still lacking for monitoring changes in surface or effective hydrophobicity that may occur in concentrated samples, precipitates and gels. In particular, the role of aliphatic amino acids in hydrophobic interactions requires investigation. Recent advances in this field have shown that laser Raman spectroscopy can be used effectively to monitor such changes.I6 1.4.1. Hydrophobic Interactions by Raman Spectroscopy. In addition to the studies undertaken on the aggregates in the 450-1900 cm-' region which indicated changes in the hydrophobic groups, amino acids, peptides and proteins also exhibit C-H stretching vibrational bands in the 2800-3100 cm-' region; this is generally considered to be related to hydrophobic groups.I6 Early studies'* reported scattering in the C-H stretching region of amino acids but the implications of the C-H region in relation to protein structure have not been extensively studied. The difficulties in the interpretation of the C-H stretching region in the Raman spectrum is due to the fact that spectra are very broad and band assignments have not previously been made. In theory, discrete vibrational frequencies should be assigned for symmetrical and asymmetrical vibrations of the methylene (CH,) and methyl (CH,) groups of saturated alkyl groups and for the =CH groups of unsaturated or aromatic groups. In simple and complex molecules the large number of CH groups, conformational changes and the surrounding microenvironment lead to overlapping of vibrations and broadening of

G u m and Stabilisersfor the Food Industry 10

322

bands. However, with recent advances in mathematical deconvolution and curve-fitting techniques it is now possible to analyse spectra comprising overlapping bands. Howell et al." reported the potential of using the C-H stretching region to study hydrophobic interactions of the aliphatic and aromatic amino acid residues of proteins. Maximum likelihood techniques were used for deconvolution which objectively determine the most likely number of peaks present. The Raman spectra of amino acids showed complexity in the C-H region which was attributed to the diversity of CH, CH, and CH, groups on the side chains, ionisation state and the microenvironment. The involvement of specific amino acids in the C-H region of lysozyme, a-lactalbumin and Plactoglobulin and their binary mixtures was investigated by deconvolution using maximum likelihood techniques (Fig.4).I6

A

ii

i U

A

I

SO50

SO00

2000

2050

DIDO

2800

Wavenumber (cm-' )

3050

SO00

2050

0000

2860

2800

Wsvenumber (em-')

.

n

N m

B

sin0

so00

eon0

PO00

esao

Wavenumber (cm-' )

esoo Wavenumber (cm-' )

Figure 4. Raman spectra of the effect of heat treatment on (i) a-lactalbumin and (ii) Plactoglobulin A) original data after baseline correction and smoothing and B) spectra of P-lactoglobulin after deconvolution. Samples were unheated (U) or heated (H).

323

Proteins and Emulsions

a

6 N 0

I

J

3100 a000 moo Wavenumber (cm-' )

: e800

aioo

ado0

edoo

eioo

Wavenumber (cm-' )

b .It

Wavenumber (cm-' )

Wavenumber (cm-')

C

LD

In (Y rn

spartic acid

1b B

c1

T

3000 en00 Wavenumber (em-' )

ell00

I

3000 e900 Wavenumber (om-' )

Z8OO

Figure 5. Raman spectral analysis in the 2800-3100 cm-' region of a) aliphatic amino acids alanine and isoleucine b) aromatic amino acids tyrosine and tqtophan and c) charged amino acids aspartic acid and lysine (adapted from Howell et al.)'

Gums and Stabilisersfor the Food Industry I0

324

Interestingly, the main protein band near the 2940 cm-' was attributed not only to aromatic and aliphatic amino acids but also to many other amino acids including charged amino acids (Fig.5). Therefore relating hydrophobic amino acids to the broad peak is not justifiable. However, it is possible to assign a band near 3065 cm-' to aromatic residues, and bands near 2880 and 2900 cm-' to aliphatic amino acids. Heating at 9OoC increased the relative intensity near 2940cm-' and decreased the relative intensity at 2895-2902 cm-' for lysozyme and its mixtures with a-lactalbumin or P-lactoglobulin. Additonal bands at 2812 or 2838 and 3003 cm-l were observed after heating or in 8M deuterated urea, reflecting changes on denaturation.I6 1.5. Protein interactions via free radicals

The presence of free radicals is common in foods and arises from processing, including heating and radiation as well as oxidation during storage. Free radicals are reported to damage proteins and DNA; this results in the unavailability of essential amino acids and induces cross-linking of proteins thereby affecting the nutritional and functional properties such as texture. In addition, the new products generated may be toxic. Howell et d 9 have recently reported the effect of free radicals, generated from lipid oxidation, on protein denaturation and aggregation. In this paper extensive studies undertaken by the author's group on the production of free radicals in oxidised lipids including fish oil; the subsequent transfer of the radicals to amino acids and proteins and the resultant cross-linking are summarised. The efficacy of synthetic and natural antioxidants were also examined with a view to reducing the damaging effect of lipid oxidation. Electron spin resonance (ESR) spectroscopy is the most direct method for detecting and measuring free radicals and identifying the atoms on which the radical resides. The g value provides enough information to distinguish between the carbon, nitrogen and sulphur centred radical. In addition, a number of other techniques including fluorescence spectroscopy, nuclear magnetic resonance (NMR) spectroscopy and gas chromatography-mass spectroscopy (GC-MS) have been employed in our studies to provide a detailed picture of protein-lipid interactions. 1.5.1. Protein-Lipid Interactions. Interaction of proteins with lipids and lipid oxidation products may result in a loss of specific amino acids such as cysteine, lysine, histidine and methionine. These interactions have implications for food quality. The reaction of lipid oxidation products including free radicals with proteins (Pr) may result in cross-linking and polymerisation as shown below.

H O + PrH Pr' + Pr' Pr' + P r ~

Pr

Pr' + H20 Pr -Pr Pr - Pr - Pr'

-

An increase in fluorescence has been reported which is attributed to the formation of

certain oxidised lipid-protein complexes; for example, fluorescent compounds have been isolated from the oxidation reaction of linoleate and myosin in frozen fish. In our studies, we have employed both ESR spectroscopy and fluorescence spectroscopy to provide

Proteins and Emulsions

325

direct evidence of the transfer of free radicals from the oxidised lipid to the amino acid or protein, followed by protein cross-linking. 1.5.2.ESR Spectroscopy. An emulsion consisting of oxidised methyl linoleate (ML) or extracted oxidised fish lipid and either amino acids (arginine and lysine); or proteins (lysozyme, ovalbumin and myosin), was freeze-dried. Control systems of either amino acids or proteins in the absence of lipid were similarly prepared. The effect of antioxidants was also tested using BHT (200 ppm), Vitamin C (500 ppm), and Vitamin E (500 ppm). Samples were analysed periodically in a Jeol RE IX X-band ESR spectrometer with 100 kHz modulation. First derivative spectra were recorded everyday for the first week and once a week subsequently for five weeks. Manganese oxide was used as a reference marker to calculate the g value (1.981, 2.034).20

2.017

I

800 .

600 400 200 0 -2003

5

-400 -600 -800 -1000

Figure 5. ESR spectra of ovalbumin incubated at 37OC for seven days with (a) oxidised fish oil and b) oxidised fish oil and Vitamin E. It was interesting to note that the strong radical signal on the carbon (g 2.0048) increased for up to about seven days of incubation of amino acids or proteins with the oxidised lipid. After twenty one days the signal was considerably weaker indicating the disappearance of the radical due to interaction with other radicals or components. Antioxidants reduced the radical signals by 70-90% depending on the type; BHT was the most effective followed by Vitamin E and a combination of Vitamin E and C whereas Vitamin C and BHA on their own were not as effective (Fig. 5). 1.5.3. Cross-linking usingJIuorescencespectroscopy. Interestingly, the disappearance of the radical signal after seven days coincided with the increase in the fluorescence intensity of the organic solvent-soluble fluorescent products extracted samples due to cross-linking of amono acids and proteins either with

Gums and Stabilisersfor the Food Industry I0

326

themselves or with lipid oxidation products (Figure 6). In concurrence with the ESR spectroscopy results, BHT and Vitamin E reduced the production of fluorescent compounds whereas Vitamin C and BHA were less effective.

90 80 70 60

a

b C

50

40

30

20

d e

10

0 0

1

2

3

4

Incubation Time (weeks) Figure 6. Fluorescence formation in ovalbumin incubated with (a) oxidised fish oil (b) oxidised fish oil and Vitamin C (c) oxidised fish oil and BHA (d) oxidised fish oil and vitamin E and (e) no treatment. To conclude, protein interactions are ubiquitous in food processing and preservation. Further research is required to provide a detailed knowledge of the different types of interactions and their effect on the texture, organoleptic and nutritional properties as well as the safety of food products. References

1. N.K. Howell, (1992). ‘Protein-protein Interactions’. In Biochemistry of Food Proteins. Ed.B.J.F.Hudson. Elsevier Applied Science Publ.Ltd.,Essex, pp 35-74. 2. N.K. Howell, Y. Shavila, M. Grootveld and S.Williams, J. Sci. Food Agric. 1996 72,49-56. 3. N.K. Howell, Gelation properties and interactions of fish proteins. In: Food Hydrocolloids I: Physical Chemistry and Industrial Application of Gels, Polysaccharides and Proteins. Ed.K. Nishinari. Osaka City University Press . Japan. 1999, In press. 4. P. Somers and N.K. Howell, Gelation properties of smooth and skeletal muscle myosin proteins. In Gums and Stabilisers for the Food Industry 9. Ed. P.A.

Proteins and Emulsions

321

Williams and G.O. Philips. The Royal Society of Chemistry, Cambridge, UK. Pp 136-144, 1997. 5 . N.K. Howell and R.A. Lawrie, J.Food Technol, 1984,19,289-295. 6. N.K. Howell, and R.A. Lawrie, J. Food Technol. 1987,22, 145-151. 7. N.K. Howell, and R.A.Lawrie, Journal of Food Technology 1985 20,489-504. 8. S. Comfort, and N.K. Howell, Food Hydrocolloids. 1999a. Submitted. 9. S. Comfort, and N.K. Howell, Food Hydrocolloids. 1999b. Submitted. 10. N.K.Howel1, Food Ingredients and Analysis International, 1999, March -April, pp 23-30. 11. N.K. Howell, N.Yeboah, and D.F.V. Lewis, Int. J. Food Sci.Techno1 1995,30, 813-824. 12. D.S. Robinson and J.B. Monsey, J. Sci.Food and Agric. 1972,23,893-904. 13. S. Poole, S.I. West and J.C.Fry, Food Hydrocolloids 1987,1,301-306. 14. G.L.Friedli and N. Howell, Food Hydrocolloids 1996,10,255-261. 15. N.K. Howell and E. Li-Chan, Int. J. Food Sci. Technol. 1996,31439-452. 16. N.K. Howell, G. Arteaga, S . Nakai, and E.C.Y. Li-Chan, J. Agric. Food Chem. 1999,47,924-933. 17. S. Nakai, J.Agric.Food Biochem.31,676. 18. S.P.VermaandD.F. H. Wallach, Proc.Nat1.Acad.Sci. U.S.A. 1976,73,3558. 19. S. Saeed, S.A. Fawthrop, and N.K. Howell, J. Sci. Food Agric. 1999,79,1809. 20. N.K. Howell, and S.Saeed, The application of electron spin resonance spectroscopy to the detection and transfer of free radicals in protein-lipid systems. In: Applications of Magnetic Resonance in Food Science. Ed. P.S.Belton, B.P. Hills and G.A. Webb. The Royal Society of Chemistry, Cambridge, UK.Pp 133143. 1999. Acknowledgements The author would like to thank her PhD students and postdoctoral fellows involved in this research area as well as Dr. E.C.Y. Li-Chan and Dr.S. Nakai at the University of British Columbia, Vancouver. Financial contributions from the European Union, The Royal Society and the Canadian NSERC are gratefully acklowledged.

SURFACTANT-PROTEIN INTERACTIONS AT AIR-WATER AND OIL-WATER INTERFACES OBSERVED BY ATOMIC FORCE MICROSCOPY.

V. J. Morris, P. J. Wilde, A. R. Mackie and A. P. Gunning.

Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, UK.

ABSTRACT Both proteins and surfactants can stabilise foams or emulsions. However, the mechanisms of stabilisationare very different and, when surfactant is added to a protein stabilised foam or emulsion, the effect is antagonistic, and the foam or emulsion breaks down. A convenient model system for studying proteins or surfactants at interfaces is the Langmuir trough. If the interfacial layers are pulled onto mica using Langmuir-Blodgett techniques then it becomes possible to study the molecular structure of the films using the new technique of Atomic Force Microscopy (AFM). By combining AFM measurements with conventional interfacialbiophysical methods it has. been possible to investigate the destabilisationof dajry protein @-casein & p-lactoglobulin) films,formed at air-water and oil-water interfaces, by the non-ionic surfactant Tween 20. The mechanisms of displacement are novel involving nucleation and growth of surfatant domains within a protein network. At a given surfactant concentration the area occupied by the surfactant domains grows until fkther growth is balanced by the mechanical energy stored in the protein lattice. The reduced area of the surface occupied by the protein is compensated for by an increased protein film thickness. Rupture of the protein network permits displacement of the protein fiom the surface. The details of the breakdown process are sensitive to the type and structure of the protein, the method of film formation and the type of interface. Using the AFM it is also possible to visualise the dynamics of the surfactant displacement of protein fiom surfaces in real time. 1. INTRODUCTION

The presence of small amounts of surfiictants leads to the breakdown of interfacial protein networks and the destabilisation of protein stabilised foams and emulsions. Both proteins and surfactants can stabilise foams or emulsions but they do so by different mechanismsiJ. Stabilisationof an interface by surfactants relies on the high lateral mobility of the molecules through a process known as the Giibs-Marangonimechanism. On the other hand proteins are believed to function by forming ‘immobile’ protein networks which slow drainage and resist deformation of the interfacial film. These mechanisms are clearly incompatible and the

Proteins and Emulsions

329

presence of a small amount of surfactant leads to the displacement of protein fiom the i n t h e and instability of the foam or emulsion. Despite studies of the adsorption of proteins to various interfacesM, and the investigation of protein displacement by competitiveadsorption"*, the detailed molecular basis of this disruption and displacement of the protein by surfactant is still unclear, and forms the basis for the present studies. At present atomic force microscopy (AFM) is the only technique which offers the prospect of visualisii the molecular structure of interfacial films, and following the interactions between proteins and surfactants. Molecular interactions at interfaces can be studied using the Langmuir trough and the structures formed can be imaged by AFM if LangmuirBlodgett methods are used to pull interhcial films onto mica substrates. By combining the AFM studies with conventional interfacial biophysical methods it is possible to determine the molecular mechanisms involved in &tant displacement of proteins fiom interfaces. This study reports investigations on the displacement of the milk proteins p-casein or plactoglobulin fiom air-water or oil-water interfaces by the non-ionic surfactant Tween 20. It has been reported by Kriigel and coworkers9 that the surface shear viscosity of plactoglobulin was reduced by the addition of Tween 20. It should be possible to visualise the molecular basis of such changes by AFM. 2. MATERIALS AND METHODS The milk proteins used for the present studies were p-casein (c-6905, Lot 12H9550)and plactoglobulin (L-0130, Lot 91H7005) fiom Sigma Chemicals (Poole, UK). Protein solutions were prepared at 2 mg mL-' in phosphate buffered saline (PBS) at a pH 7.0. Surface pure water (surface tension 'yo = 72.6 mN m-' @ 20" C) cleaned using an Elga Elgastat UHQ water purification system was used in this study. The oil was n-tetradecane (Sigma Chemicals). F d y , the surfactant polyoxethylene sorbitan monolaurate (Tween 20) was obtained as a 10?! solution (Surfact-Amps20) fiom Pierce (Rockford, IL). For air-water interfaces surface tension measurements were made using a wetted ground glass Wiulelmy plate and a Langmuir trough (PTFE, 255 x 112 x 16 mm, volume 450 mL, with one fixed and one movable barrier). Both spread and adsorbed interfkial films were prepared. Spread protein films were prepared by carellly adding the required volume of protein solution dropwise via a glass rod onto the surface, and then rinsing the rod with buffer. 64.8 pg of p-lactoglobulin and 36 pg of p-casein were added to the surface. The concentrationsadded were chosen in order to produce similar surface pressures (1OmN m-') after about 30 minutes. At this point in time the Tween 20 was added to the subphase at a concentration of 0.5 pM. The surface pressure increased with increasing time and approached a quasi-equilibrium value:Further increases in surhce pressure were induced by the addition of more Tween 20 to the subphase. Langmuir-Blodgett (LB) films were pulled (driving rate 0.14 mm s-') onto fkshly cleaved mica at defined surface pressures, either as a h c t i o n of time at a fixed surfactant concentration, or as a h c t i o n of increasing surfactant concentration Adsorbed protein films were formed fiom solutions containing 1 pM protein and 0.5 pM Tween 20 which were added directly to the Langmuir trough. After 30 minutes the subphase was perfused with 2 L of 0.5 pM Tween 20 at a rate of 1.1 mL i'.LB films were d r a w at defined surface pressures in order to image the protein-surfactant interactions.

Gums and Stabilisersfor the Food Industry 10

330

For oil-water interfaces surface tension measurements were made using a platinum Du Nouy ring and a glass trough (area 26,300 mm*). In this study only data on adsorbed films will be reported. Adsorbed protein films were formed from solutions containing 1 pM protein and 1.0 pM Tween 20 as described above. After 30 minutes the subphase was perfused, again as described above. LB films were drawn (0.14 mm s-') onto hshly cleaved mica in order to monitor changes in structure of the interfkce as a function of increasing surface pressure, either as a fimction of time at a fixed surfactant concentration, or as a function of i n c d surfictant concentration in the subphase. The LB films were t r a n s f d to the liquid cell of the AFM and then imaged under redistilled butanol. The butanol dissolved the Tween 20 effectively fixing the structure of the film. In the case of the oil-water interfaces the butanol also removed any residual ntetradecane. Images were obtained using an East Coast Scientiiic (ECS Ltd, Cambridge, UK) AFM operated in the dc contact (constant force) mode. The cantilevers used were Nanoprobe silicon nitride levers with a nominal force constant of 0.38 N m-I. The forcedistance curves under butanol were reversible showing no adhesive component, allowing the imaging force to be adjusted in order to optimise contrast and eliminate damage or displacement of the sample by the probe. Imaging forces were typically of the order of 1-3 nN.Images of the surhce films were analysed in order to determine surfactant surfice area coverage and protein surface area coverage, film thickness and film volume. Histograms of grey levels within the images were generally bimodal. The separation of the two peaks provided an estimate of the mean film thickness. Choosing a suitable threshold value allowed the images to be converted into simple binary images and an evaluation of the area covered by protein and the area once occupied by surfactant. From the measured protein film area and thickness the volume of the protein film was determined.

2. RESULTS AND DISCUSSION The adsorption of surfactant onto the i n t h e could be monitored as an increase in surface pressure. AFM images provided structural information on the i n t d c i a l layer. In this article it is convenient to concentrate on the structural changes occurring close to the collapse point of the &n.

2.1. Protein layers at air-water interfnces

Figure 1 shows structural changes observed in spread p-casein protein films from an airwater interface as a function of increasing surface pressure due to the addition of Tween 20. The black regions in the image correspond to bare mica and represent regions which were originally occupied by surfactant. The white regions in the image correspond to the protein. It can be seen that initially small surfactant regions or nucleated domains appear within the protein layer (mowed in fig. la). These nuclei are small corresponding to the size of individual proteins or less. It is likely that these nuclei form due to defects in the protein lattice. These may be holes corresponding to the absence of proteins, or may be regions where the protein is more fully unfolded and can be 'displaced' by forcing refolding of the protein. At higher s h pressures the surfactant nuclei grow into progressively larger domains (fig. lb). Eventually the surfactant forms a continuous phase containing islands of protein. This is shown in fig. 2d for the protein p-lactoglobulin. For the p-casein - Tween 20 system it is noticeable that the surfactant domains are circular in shape indicating isotropic expansion. In fact the domains appear slightly elliptical: this is believed to be due

Proteins and Emulsions

331

to slight distortion of the weak P-casein films which remain on the Langmuir trough after LB film transfer.

Fipm 1AFU images of mixed LB filmsof c d $-casein and k e n - 2 0 transferred f i m an air-water intevace onto mica and imaged in butanol. (a) syrlace pressure 15.9 mN m-' , scan size 1.6 x 1.6 p. Arrows indicate Tween 20 domains. (b) Surfime p r e m 19.2 mNm" , scan size 6.4 x 6.4 pa.

It is suggested that the Tween 20 domains are actually circular at the air-water interface but become distorted to ellipses as the surface film compensates for the removal of previously drawn LB films. The oval shapes always appear to be oriented in the same direction and this would be consistent with such an explanation. Figure 2 shows the structural changes which occur during the displacement of spread Plactoglobulin films formed at the air-water interface by Tween 20. The peribsion step is important if images of the type shown in fig. 2 are to be obtained for adsorbed films. Without the peribsion step the interfacial film becomes coated with passively adsorbed protein from the subphase during the LB transfer process. This passively adsorbed protein obscures the structure of the interfacial film. However, if the perfision step is included then only the interfacial layer is transferred and the data can be analysed both qualitatively and quantitatively.The displacement process observed is superficially similar to that seen for Pcasein - Tween 20 mixtures. It involves nucleation and growth of surfactant domains which expand compressing the protein network until it fails (fig. 2a-c). After failure the protein network is broken up leading to a surfactant phase containing islands of protein (fig. 2d). The major difference between the structures shown in figures 1 & 2 is the shape of the surfactant domains. In the case of P-lactoglobulin the surfactant domains are irregular in shape. This is characteristic of redistribution and concentration of stress at the weakest points in the protein structure. Thus the domains grow by rearrangement of the protein network at the weakest linkage points. The P-lactoglobulin films fail at higher surface pressures than the 0-casein films, and this is consistent with the fact that P-lactoglobulin is better at preserving the foam structure. Studies on adsorbed films show similar behaviour although the adsorbed films fail at higher surface pressures. Figures 3a & b show displacement of p-casein and P-lactoglobulin respectively by Tween 20. Once again the surfactant domains in P-casein films are circular and those in P-lactoglobulin are irregular in shape. This suggests that the molecular mechanisms are similar for spread or adsorbed films but the adsorbed films are in general

332

Gums and Stabilisers for the Food Industry I0

more difficult to displace. This difference is probably a consequence of the different preparative regimes for spread and adsorbed films. For the spread films the protein is introduced as a concentrated domain, which is diluted as it 'spreads' across the surface, whereas the adsorbed film is formed due to a gradual build-up of surface concentration as the protein molecules migrate and adsorb from the subphase. In the latter case there is more scope for protein unfolding resulting in greater interaction between neighbouring proteins, or with the interface, leading to stiffer networks.

Figure 2 AFM images of mixed LB filmsof s m d Plactoglobulin and Tween-20 transfwedjFom an airwater interface onto mica and imaged in butanol. (a) Surface pressure 18.6 mN m-' , scan size 1.6 x 1.6 p. (b) Surface pressure 20.2 mN m-' , scan size 2.8 x 2.8 p. (c) Surface pressure 22.5 mN m-' , scan size 3.2 x 3.2 p.(4 Surface pressure 27. I mN m-' scan size 9.0 x 9.0 p. ~

Thus a qualitative inspection of the data demonstrates several important points. Firstly the AFM data provides, for the first time, direct experimental evidence for protein 'gel-like' networks at the air-water interface. Secondly, the data has demonstrated that the displacement mechanism is heterogeneous; involving nucleation and growth of surfactant domains. Thirdly the stability of the protein network, as visualised by AFM, correlates with the enhanced stability conferred by adsorbed films over spread films, and for P-lactoglobulin over P-casein. The clear differences between the behaviour of P-casein and 0-lactoglobulin are almost certainly due to the fact that, whereas P-lactoglobulin is a well structured protein

Proteins and Emulsions

333

(10% a-helix, 45% P-sheet and 2 disulphide bonds), P-casein is an essentially unstructured (10% a-helix, 13% P-sheet) protein. The simplest molecular mechanism based on the images shown in figures 1- 3 is that the surfactant displaces the protein at the protein-surfactant interface as the surfactant domain grows, i.e. an erosive process. However, quantitative analysis of the images shows that the mechanism is far more subtle. Figure 4 shows calculated protein area, thickness and volume data illustrated for spread and adsorbed P-lactoglobulin films displaced by Tween 20. The most comprehensive data is available for spread films. It can be seen that initially there is a decrease in protein surface area accompanied by a small decrease in volume but without any change in layer thickness. This is then followed by a region where the area decreases still further, the thickness increases but the volume remains approximately constant. Finally, there is a catastrophic decrease in film area and volume accompanied by an increase in film thickness.

Fipn 3 A F U images of mixed LBBfilms of a h r b e d proteins and lbeen-20 tranflerredfiom an air-water interface onto mica and imaged in butanol. (a) p a s e i n , mrface pressure 20.4 mN m-l, scan size 3.0 x 3.0 pa. (b) /h!.actoglobulin,surface pressure 25.1 mN rn-’ , scan size 3.2 x 3.2 p.

These data suggest a novel displacement mechanism. Initially it is considered that the increase in surfice pressure compresses the protein network by squeezing out defects in the ‘lattice’. This is then followed by a regime in which the protein network is compressed but there is no change in volume. The increase in film thickness suggests that either individual proteins, or the protein film, reduce the interfacial contact area by deforming or buckling. Even at this stage, where there is extensive surface coverage with surfactant, the protein network remains intact and no protein has been lost or detached from the interface into the subphase. Clearly some local rearrangement of protein-protein interactions must occur in order for the domains to expand but this does not involve loss of protein to the subphase. During these stages the quasi-equilibrium observed at given surfactant concentrations arises when the expansion of the surfactant domain is balanced by the elastic energy stored in the protein lattice. The behaviour observed for adsorbed films is similar, but the eventual failure of the protein network is more extreme, and occurs over a narrower range of surface pressures. An alternative, but not necessarily equivalent, interpretation of the behaviour at the interface is that the net decrease in surface tension, as the surfactant adsorbs, causes the protein layer to rearrange and contract, reducing its interfacial contact area by deforming or buckling. In either case the stability of the interfacial layer is determined by the

334

Gums and Stabilisersfor the Food Industry 10

distortion and eventual Mure of the protein network. The mechanism for the surfactant displacement of the protein has been termed’’’’2 an “orogenic” displacement, a process involving crushing together and the formation of faults and folds, by analogy with the geological process of plate tectonics.

:T

Figure 4 Quantijkation of protein displacement by slrrfactant as a function of s&ace pressure for $lactoglobulin and Twen-20 mixed films. (a) Protein film thickness, (b) percentage area occupied by protein, (c)protein film volume, i.e. area x thickness.

2.2. Protein layers at oil-water interfaces Figure 5 shows representative data for Tween 20 displacement of p-lactoglobulin (fig. 5a) and p-casein (Sg. 5b) fiom an oil-water interface. In both cases it can be seen that the displacement also occurs by the nucleation and growth of mrhctant domains. In the case of the p-lactoglobulin system the domains are once again irregular. However, for the p-casein system at the oil-water interhce the domains are also irregular in shape. This suggests that the presence of the oil-water interface increases the sti&ess of the protein network making it harder to displace by surfactant. Such changes are consistent with the fiict that p-casein is more effective at stabilisii emulsions than foams. The data shown in figure 5 suggests that surfactant displacement of protein fiom oil-water interfaces also occurs by an orogenic mechanism.

Proteins and Emulsions

335

3 . CONCLUSIONS

Use of AFM has led to new and important findings in the present study. Direct evidence has been obtained for the existence of protein ‘gel-like’ networks at air-water and oil-water interfaces. A new orogenic mechanism has been discovered for the displacement of proteins from such interfaces by (non-ionic) surfactants. This was only possible because of the high resolution obtainable by AFM, and could not have been deduced from other techniques, currently available for studying interfaces, which either lack the necessary resolution, or spatially average the surface structure.

Fipm 5 AFM images of mixed LBjilms of adwrbedjmteins and ?been-20 hansferredfi.om an oil-water interface onto mica and imaged in butanol. (a) ~Lmtoglobulin,surfoe premre 32.1 mN m.’ scan size 1 . 6 I~. 6 jm. (b) PCasein, surface pressure 28.7 mN m-’ , scan size 10.0 x 10.0 jm.

Recent studied2 have shown that an orogenic mechanism also applies for the surfactant displacement of protein from solid-air interfaces and that this process can also be imaged in ‘real-time’. These studies emphasise, for the first time, the need to consider protein-protein interactions, and the fracture of protein networks, in accounting for the different resistance of proteins to displacement from interfaces by surfactants. ACKNOWLEDGEMENTS The research studies presented in this article were finded by the BBSRC through core grant fbnding to the Institute of Food Research.

References 1. D. C. Clark, M. Coke, A.R. Mackie, A.C. Pinder and D. R. Wilson, J. Colloid & Interface Sci., 1990, 138, 207. 2. P. J. Wilde, J. Colloid&Interface Sci., 1996, 178, 733.

336

Gums and Stabilisersfor the Food Industry 10

3. J. C. Lee and K. J. Tynan, ‘Proceedings, Second International Conference on Bioreactor Fluid Dynamics’, Elsevier, Amsterdam/New York, 1988. 4. K. Al-Malah, J. McGuire and R. Sproull, J Colloid & Interface Sci., 1995,170,261. 5 . A. R. Mackie, J. Mingins and A. N. North, J Chem. SOC.Faraday Trans., 1991, 87, 3043. 6. C. Cao and S. Damodaran, J Agric. Food Chem., 1995,43,2567. 7. J. Chen, E. Dickinson and G. Iveson, Foodstructure, 1993, 12, 135. 8. D. C. Clark, A. R. Mackie, P. J. Wdde and D. R. Wilson, Faraday Discussions, 1994, 98,253. 9. J. Kriigel, D. C. Clark, P. J. Wilde and R. Miller, Prog. Colloid & Polym. Sci., 1995,98, 239. 10. A. R. Mackie, A. P. Gunning, P. J. Wilde and V. J. Morris, J Colloid & Inter$ace Sci., 1999,210,157. 11. A. P. Gunning, A. R. Mackie, P. J. Wdde and V. J. Morris, Sur$ace & Interface Anal., 1999,27,433. 12. A. P. Gunning, A. R. Mackie, P. J. Wdde and V. J. Morris, Langmuir, 1999,15,4636.

RHEOLOGICAL INVESTIGATIONS OF TRANSIENT GEL IN A DEPLETIONFLOCCULATED POLYDISPERSE EMULSION.

P. Manoj, A. J. Fillery-Travis, D. J. Hibberd, A. D. Watson and M. M. Robins Institute of Food Research, Norwich Research Park, Colney, NORWICH. NR4 7UA.

1. INTRODUCTION.

A large sector of the food industry uses oil-in-water emulsions where oil droplets are dispersed in the aqueous continuous phase. Because these droplets are less dense than the aqueous phase, they tend to rise to the top of thc containers (creaming). In many formulations, polymer is added to the continuous phase to slow down the creaming rate; but the presence of non-adsorbing polymer can induce the formation of a ‘space-forming’ structure or network by the depletion mechanism’. This can be followed by the collapse of the structure under the influence of gravity. Close inspection has shown a delay period exists prior to creaming which is dependent on both oil and polymer concentrations2.Two hypotheses are put forward to explain the presence of the delay period and its relationship to creaming. The hypotheses are as follows: (1) the delay originates fiom slow flocculation, hindered by the high viscosity of the continuous phase; (2) the space-spanning structure changes the dispersion rheology sufficiently to produce a yield stress effectively stopping the creaming of emulsions. To differentiate between these two hypotheses, an extensive rheological investigation using steady state shear rheology and small deformation oscillatory rheology was carried out and the findings are presented here. 1.1 Depletion Flocculation

The addition of a non-adsorbing polymer to an emulsion can lead to a weak flocculation of oil droplets by the depletion effect. When the surfaces of the droplets dispersed in a polymer solution approach to distances closer than the diameter of the polymer molecules, these molecules are excluded from the interparticle region, creating a polymer-depleted zone. The difference in the osmotic potential between the polymerdepleted zone and the bulk solution generates an inward force on the droplets (Figure 1). 1.2 Creaming and Delay Time of Emulsions.

Previous studies’ have shown that creaming of depletion-flocculated emulsions a certain period of time called the delay time or delay period. The delay

O C C U K ~after

338

Gums and Stabilisersfor the Food Industry 10

time can be easily seen by the visual observation of creaming of hexadecane (Cetane CI~HM; density 773.4 kg.m” at 20”C, Sigma Chemical Company) emulsions in presence of a non-adsorbing polymer, hydroxyethylcellulose (HEC,Natrosol 25OHR, Aqualon, mean r,(radius of gyration) = 58 nm).

!-!

b >> 2rg - depletion interaction is negligible (where rBis the radius of gyration of the polymer molecule).

b << 2rg - attraction occurs due to depletion effect.

Polymer-depleted zone. Figure 1

Schematic representation of the close approach of the depletion layers associated with the two droplets.

After preparation of a hexadecane emulsion containing HEC polymer, it was poured into a measuring cylinder which was set on a bench in a temperature-controlled room (25°C). The presence of the polymer induced the aggregation of hexadecane oil droplets by the depletion flocculation mechanism. As a consequence, the emulsion phase-separated into two layers: the top layer being colloid-rich and called the cream, and the bottom being colloid-poor and called the sub-cream. These two layers are separated by a boundary line or a meniscus, and it is this meniscus whose height was measured as a function of time for the estimation of delay time. Figure 2 shows the measurement of the height of the meniscus as a function of time for a 34% (v/v) hexadecane emulsion in presence of 0.35% (w/w) HEC polymer. As seen the delay time was approximately 380 minutes after which a meniscus formed and rose to the top of the cylinder, separating the emulsion into cream and sub-cream layers. This creaming experiment has been repeated using different oil volume fractions and different polymer concentrations. Figure 3 shows a chart of delay time plotted as a function of polymer concentration for four different oil volume fractions and as seen in the graph, the delay time increased with increasing polymer and oil concentrations. It has been suggested’ that the delay period is consistent with the formation of a spacespanning structure or network of aggregated oil droplets. The two hypotheses above have been put forward to explain the mechanism lying behind the delay period; the hypotheses have been investigated using rheological techniques and the findings are presented here. A

339

Proteins and Emulsions

better understanding of the delay period can then be applied to manipulate complex food systems, to improve their stability and quality.

-fi g 2

*a cw

3

.%

160 140 120 100 80 60 40 20

CREAM LAYER A A

h A

.c &A

I

SUB-CREAM LAYER

time (minutes)

delay time

Figure 2

A

A

Height of meniscus as a firnction of time for a 34% (vh) hexadecane emulsion in presence of 0.35% (whv) HEC polymer.

n

m

9 PI 0.04 'r( 0.03 'g z 0.05

450

c a, ,

1

-3 300 W

m

2

0.02 'h 5

-3150

Q

B

0.01

L . .

4

g

M

o g

0 0

0.1

0.2

0.3

0.4

0.5

HEC concenfrafion(why %)

(+

Figure3 Dehy times of Hexadecane-HEC emulsions 10% (v/v); 0 20% (v/v));A34% (vh) and ff 40% (vh) hexadecane oil). Dotted line represents viscosiw of continuous phase.

Gums and Stabilisers for the Food Industry 10

340

2 EXPERIMENTAL 2.1 Materials and Emulsion Preparation

The emulsions for rheological characterisation were formulated to be near densitymatched by using 1-bromohexadecane (Cetyl Bromide- (CH3(CH),5Br), density 1000.75 kg.m-3 at 20°C; Fluka) as the dispersed phase. Thus, the effects due to creaming of the droplets could be isolated from those due only to flocculation. A 50% (v/v) premix was prepared by homogenisation in a Waxing blender with a fixed series of shearing cycles. The droplets were stabilised against coalescence using the non-ionic surfactant polyoxyethylene 23-lauryl ether (Brij 35, Sigma Chemical Company) at a level of 2.7% (w/w) with respect to the continuous phase. The premix, by determination of particle size analysis, was found to be stable to coalescence and disproportionation over the complete timescale of the experiments. The final emulsions were prepared by diluting the premix with an aqueous solution of hydroxyethylcellulose (HEC). The HEC was purified by dialysis and freeze dried prior to use. Sodium azide (Sigma Chemical Company) was included within the formulation as an anti-microbial agent at a level of 0.3% (w/w) with respect to the final continuous phase. The emulsions were prepared in the range 10 - 40% (v/v) oil and 0.04 - 0.35% (w/w) HEC for the rheological investigations. 2.2 Droplet Size Distribution.

The droplet size distribution of the premix emulsion was measured using a Malvern Mastersizer laser diffraction sizer fitted with a small sample handling unit. The formation of the emulsions was found to be highly reproducible giving a size distribution of weight mean droplet diameter 1.98 pm (Figure 4). Our previous studies using the same systemsz4 have shown the primary size distribution to remain unchanged in the course of the experiments, i.e. coalescence of the droplets was negligible. 2.3 Rheology

Rheological measurements were carried out on a Bohlin Controlled Stress (CS) Rheometer. Polymer solutions, emulsions with and without polymer were prepared and poured into tfie double gap geometry preset at 25°C. The geometry was covered with solvent traps to minimise the effects of drying out and dust settlement. Viscosity measurements were carried out as a 2.3.I Viscometry Measurements. hnction of increasing shear rate (steady stress (T) - shear rate ( i ) on ) a number of emulsions in presence of the polymer. Three types of measurements were 2.3.2 D m m i c Oscillatory Measurements carried out. Time Sweep: Soon after the sample was loaded into the geometry, the sample was given a short, fast boost to break up any structure that may have been developing during the sample loading period. Immediately after the boost period, G’ (elastic) and G ’(loss) moduli were measured at a constant frequency and target amplitude as a function of time. The time period for all emulsions was 60 minutes. Measurements of a few emulsions selected for characterisation over a period of 15 - 18 hours (overnight) showed no hrther

34 1

Proteins and Emulsions

14 12 s 10 “ 8 g 6 0 + 4 2 0

&

0.1

0.3

0.9

2.4

7.3 20.7 58.8

particle diameter (microns)

Figure 4

Typical size distributionfor bromohexadecanepremix.

changes in the moduli. Preliminary exercises revealed a target amplitude of 0.003 and a frequency of 2 Hz to be the best combination for rheological measurements for all the emulsions. Frequencv Sweep: G and G ”were measured as a fbnction of frequencyfrom 0.01 Hz to 10 Hz at an amplitude of 0.003. Amplitude Sweep: G and G ’ were measured as a fbnction of increasing amplitude (strain) at the frequency of 2 Hz for all emulsions after network equilibrium had been achieved. 3 RESULTS AND DISCUSSION

The aim of the experiment was to explore the validity of the two hypotheses which explain the mechanism underlying the delay period prior to creaming. Both hypotheses are discussed below, using rheology results.

-

3.1 First Hypothesis Viscosity oftbe continuous phase

Does the delay originatefrom the viscous properties of the continuousphase? Does the viscosity of the continuous phase, which is increased by the presence of the polymer, slow down the droplet encounter rate of the oil droplets, effectively contributing to the delay period and the delay period is, therefore, a consequenceof slow flocculation? Figure 3 shows the delay time plotted as a h c t i o n of polymer concentration. The viscosity of the polymer is included: for instance, a 0.2% (w/w) polymer has a continuous phase viscOsity of approximately 0.006 P a s At this particular polymer concentration, corresponding oil volume fractions ranging from 10 to 40 % (v/v) represent increases which means an increase in the number of oil droplets. An increase in the number of oil

Gums and Stabilisersfor the Food Industry 10

342

droplets means a faster droplet encounter rate and consequently a faster flocculation rate. Therefore, theoretically, an increase in oil volume fraction would lead to a faster flocculation rate, which in turn means a shorter delay time. The experimental evidence, as presented in Figure 3, shows that the delay time increases with increasing oil volume fraction, therefore, contradicting the hypothesis of slow flocculation induced by the continuous phase. 3.2 Second Hypothesis - Yield stress of the spacespanning structure

Does a space-Spanning structure change the dispersion rheologv sujficiently to produce a yield stress effectively stopping creaming of emulsions? Have all the oil droplets aggregated by depletion flocculation into a space-spanning structure that remains stable until there comes a point when the yield stress is overcome and the network creams as a single entity? By yield stress, we adopt a working definition as given by Nguyen and Boger5 and Cheng6;that below a certain critical stress - the yield stress - a material may not flow, but deforms plastically like a solid, with a strain recovery upon the removal of the stress. Above the yield stress, the material flows as a fluid with some viscosity. Yield stress in depletion-flocculatedparticulate dispersions have been reported by several investigators” 8, 9. 10 . The existence and quantification of yield stress calls for investigation into the rheological properties of emulsions. 3.2.1 Application of Viscometryto detect YieM Stress. Figure 5 shows the viscosityshear rate profiles for 25 (v/v) bromohexadecaneemulsions in presence of 0.04%, 0.15% and 0.35% (w/w) HEC polymer.

0.001

0.0001 0.01

0.1

1

10

100

1000

Shear Rate (sl) Figure 5

Viscosity-shear rate plot for 25% (v/v) bromohexadecane emulsions in presence of 0.04% (A);0.15% (9;and 0.35% (+) (wh)HECpolymer.

343

Proteins and Emulsions

AU emulsions (10 to 40% (v/v) bromohexadecane with 0.04 to 0.35% (w/w) HEC polymer) show shear-thinning patterns from the lowest available shear rate. The fall in viscosity with increasing shear rate is indicative of a structural break down. Several models, created for the calculation of yield stress, such as Bingham, Casson and Herschley-Buckley models have been tried using our data. Such fittings were found to be unstable and highly erroneous3. The use of 3.2.2 Application of Oscilhtion Rheologv in search for Yield Stress small deformation oscillatory rheology enables us to probe the weak structure without destruction, unlike viscometry. Figures 6(a) and (b) represent the frequency spectra (measurement of storage ( G ) and loss (G’) moduli as a h c t i o n of increasing frequency) for 40% (v/v) oil alone (highest experimental oil volume fraction) and 0.35% (v/v) polymer alone (highest experimental polymer concentration) taken after one hour resting period. In both spectra, the G ’ was dominant; hence no structure was present.

n

3 b

46

F

x

k

10

I

3

1

0.1 0.01

W

?J

f

0.001 0.0001 0.00001 \ 0.01 0.1 1 10

fkquency (Hz)

Figure 6 (a) Frequency spectrum for 40% (v/v) bromohexadecane emulsion with no p o I ! e r (G’ 4 G” 0; q* A). An emulsion of 28% (v/v) bromohexadecane with 0.25% (w/w) HEC polymer was prepared and loaded into the Bohliin rheometer where a time sweep (measurement of the moduli as a h c t i o n of time) was carried out. Figure 7(a) shows the fist five minutes of the time sweep and Figure 7(b) shows the fill time sweep for 60 minutes. In the fist few seconds of the time sweep, the G ’ modulus was dominant (Figure 7 (a)), indicating liquid-like behaviour. At about 1 minute, a cross-over of the moduli occurred and the G dominated, indicating a solid-like state.

344

Gums and Stabilisers for the Food Industry 10

10

n

3 a

1

U

*

0.1

.n

n (d

b

0.01 0.001 0.0001

1

0.1

0.01

10

frequency (Hz) Figure 6 (b) Frequency spectrum for 0.35% (wh) HEC solution (G’ & G ” G T,I*A).

n

a“

W

h i3 1 0

1

2

3

4

5

time (minutes)

Figure I (a) First five minutes of the 60 minutes time sweep for 28% (v/v) bromohexadecane emulsion in presence of 0.25% (wh) HEC polymer( G ’ & G A ) at 2 Hz and target amplitude of 0.003. ”

The cross-over of the moduli may be termed the gel point and marks the onset of the development of a flocculated network of aggregated oil droplets. At the end of the time sweep, i.e. at 60 minutes, a stable and equilibrated network of aggregated droplets existed. Overnight observation revealed no changes in the moduli.

345

Proteirw and Emulsions

I

-I

GI

0

20

60

40

time (minutes)

Figure 7 (b) Full time sweep for a 28% (vh) bromoheradecane emulsion in presence of 0.25% (wh) H E C p l p e r where G’ HandG”A.

The time sweep experiments were repeated for emulsions of different oil volume fractions and different polymer concentrations. Figure 8 shows the values of G taken after 60 minutes as a finction of oil volume fraction. G increased with increasing oil and polymer concentrations, implying that stronger networks were fonned upon increasing oil and polymer concentrations.

16 12

b 4

0 20

25

30

35

40

Oil volume concentration (%; v/v) Figure 8

Concentration dependence of G ’ for emulsions containing 0.10% ( j ; 0.15% (+); 0.20% (9;0.25% (9;0.30% (Q and 0.35% (A)(wW HFC polymer

346

Gums and Stabilisersfor the Food Industry 10

Afler the time sweep, amplitude sweeps were carried out on the equilibrated networks. Figure 9 shows amplitude sweeps carried out on three different emulsions. Looking at, for example, the profile for 37% (v/v) bromohexadecane emulsion in the presence of 0.30% (w/w) HEC polymer, a stable and equilibrated network existed at very low amplitudes. However, upon increasing the amplitude, the moduli fell and the network was completely destroyed when the G’ crosses with G”.

loo

i 4

I

-I

3

0.001

0.01

0.1

Strain Figure 9

Amplitude sweep for 25% (v/v) bromohemdecane emulsion with 0.15% (whv) HEC polymer (diamond); 25% (vh) bromohexadecane emulsion with 0.25% (whv) HECpolymer (square) and 37% (vh) bromohemdecane emulsion with 0.30% (whv) HEC polymer (triangle) where filled symbols represent G ’ and unfilled symbols represent G ”.

If the data were replotted into a stress-strain chart, an interesting pattern emerges (Figure lO(a)). Extrapolations were made from zero stress-strain and from high stressstrain for each curve. The deviations from the straight l i e s are marked and joined, forming two lines which divide the chart into three regions. The high stress region is called Viscous Flow and the lines for all three emulsions are straight and parallel, indicating the same flowing behaviour for all emulsions. The low stress region is the Solid region in which the lines are straight but at different gradients, representing the different moduli of the stable networks as created by different oil and polymer concentrations. Between the Viscous Flow and Solid regions is a Transitional region which covers the breaking behaviour of the structures from stable networks to freeflowing emulsions. The high stress region is called Viscous Flow and the lines for all three emulsions are straight and parallel, indicating the same flowing behaviour for all emulsions. The low stress region is the Solid region in which the lines are straight but at different gradients, representing the different moduli of the stable networks as created by different oil and polymer concentrations. Between the Viscous Flow and Solid regions is a Transitional region which covers the breaking behaviour of the structures from stable networks to freeflowinn emulsions.

347

Proteins and Emulsions

0.04

0.03 d

0.02 v)

0.01 SOLID

0 0

0.05

0.1

0.15

0.2

Stress (Pa) Figure lqa) Stress-Strain plot for 25% (v/v) bromohexadecane emulsion with 0.15% (wh) HEC polymer (damond); 25% (vh) bromohexadecane emulsion with 0.25% (wh) HECpOrymer (square) and 37% (vh) bromohexadecane emulsion with 0.30% (wh) HEC polymer (triangle). Examination of the strain sweep data in this way prompts the following interpretation. Is it possible that the first point of deviation from the straight line extrapolated from zero stress-strain represents the first appearance of the break-down of the structure and could thus be considered as a Static Yield Stress? Would it not be possible that the first point of deviation from the straight line extrapolated from the high strain-stress behaviour represents the first appearance of flow behaviour in emulsions and be considered as a Dynarmc Yield Stress? Figure 10 @) summarises the proposal for identification and quantificationof yield stress. The curvature between the two yield stresses represents the gradual break-down of the structure between the solid state and the liquid state. SUMMARY & CONCLUSIONS The delay period that occurs prior to the onset of creaming draws attention to the stability of oil-in-water emulsions in healthwe, pharmaceutical, cosmetic and food products. Understanding the mechanism underlying the delay period would bring manipulation of the product ingredients, resulting in control over the shelf life of products. Two hypotheses have been put forward to explain the purpose of the delay period: the first based on the viscosity of the continuous phase, and the second based on the yield stress of the structure. Evidence has been shown to contradict the first hypothesis, where theoretically the delay period should decrease with increasing oil volume fraction, but experiments showed that the delay time increased with increasing oil volume hction. The second hypothesis calls for yield stress which is dehed as a point when a structure starts to flow. Therefore

348

Gums and Stabilisersfor the Food Industry I0

Dynamic Yield Stress\

0.03 d

-g 0.02 c n 0.01

/

1 Yieldstress /

0 0

0.05

0.1

0.15

0.2

Stress (Pa) Figure 10(b) Proposal for Identification and Quantification of Yield Stress for 37% (v/v) bromohexadecane emulsion in presence of 0.30% (wh)HEC polymer. (Static Yield Stress = 0.047 Pa; Qynamic Yield Stress = 0.152 Pa).

rheological techniques have been applied to investigate the flow properties of emulsions. Viscometry analysis showed firstly, the absence of zero-shear viscosity at low shear rates as provided by our rheometer and secondly, the emulsions are shear-thinning, indicating the break-down of the pre-existing structures. Application of viscosity models gave poor fits. The use of dynamic oscillatorytechniques enabled us to probe the flocculated structure non-destructively. Network strengths dependent on oil and polymer concentrations were assessed by monitoring the elastic ( G ) moduli. Amplitude sweeps have allowed us to observe the breakdown of the structure from the solid state, which was stable at low strains (viscoelastic region) to the liquid state (flow region) by monitoring the G and G ’ moduli as a hnction of increasing amplitude. Re-analysis of data into stress-strain relationship has provided a proposal for identification and quantification of not one yield stress, but two yield stresses - static and dynamic. The existence and quantification of yield stress as a hnction of oil and polymer concentrations give credence to the second hypothesis put forward to explain the delay period. Currently, we are investigating the effect of polymer and oil concentrations on the network breakdown, developing a model which should bring hrther understanding to the production, stabiity and shelflife of oil-in-water emulsions in food products. REFERENCES 1. S p e q , P. R., J. COILInt. Sci., 1982,87, 375. 2. P. Manoj, A. J. Fdlery-Travis, A. D. Watson, D. J. Hibberd and M. M. Robins, J. COIL Int.,Sci., 1998, 207, 283.

Proteins nnd Emulsions

349

3. P. Manoj, A. J. Fillery-Travis, A. D. Watson, D. J. Hibberd and M. M. Robins, J. Coll. Int. Sci,, 1998, 207,294. 4. Fillexy-Travis, A., Gunning, P. A,, Hibberd, D. J. and Robins, M. M., J. of Coll. Int. Sci., 1993, 159, 189. 5 . Nguyen, Q. D. and Boger, D. V., Annu. Rev. FluidMech., 1992,24,47. 6. Cheng, D.C.H.,RheologicuActu, 1986,25 (9,542. 7. Liang, W., Tadros, T. F. and Luckham, P. F. J. of Coll. Int. Sci., 1993, 155, 156. 8. Liang, W., Tadros,T. F. and Luckham, P. F., J. of Coll. Int. Sci., 1993, 158, 152. 9. Liang, W., Tadros, T. F. and Luckham, P. F., J. of Coll. Znt. Sci., 1993, 160, 183. 10. Buscall, R., McGowan, I. J. and Mume-Young C. A., Faraday Discuss. ChemSoc., 1990,90, 115.

EFFECT OF pH AND NaCl ON RHEOLOGICAL AND TEXTURAL, PROPERTIES OF LWIN PROTEIN EMULSIONS

Raymundo A.', Empis, J.', Sousa I.'

' Universidade Tecnica de Lisboa. Instituto Superior de Agronomia. Laboratorio Ferreira Lapa. Tapada da Ajuda. 1349-017 Lisboa. Portugal. Universidade TCcnica de Lisboa. Instituto Superior Tecnico. Centro de Engenharia Biologica e Quimica. Av. Rovisco Pais, 1049-001. Lisboa. Portugal.

'

ABSTRACT The effect of pH on textural and rheological properties of lupin protein o/w emulsions at various protein contents (2, 3,4, 5 and 6% wt) was studied. Five emulsions with different pH values ranging from 3 to 8 were prepared at each protein content. The NaCl concentration effect was also investigated on 4.5% wt protein emulsions with salt contents ranging from 0.3 to 5% wt. In both cases, the sunflower oil content was 65% wt. From the results of this work it was observed that for all protein concentrations tested, the textural and rheological properties reach a minimum at a pH value around the protein isoeletric point (pH cu. 4 . 9 , and the oil droplet size given by d,,, shows a maximum at this pH value, indicating low emulsion stability. In the study of the effect of salt, at pH 2.5 the addition of salt reduces emulsion stability, at pH 4.5 the emulsion physical properties are independent of salt concentration and at pH 6.5 the salt addition enhances structure, texture and stability of the emulsions. 1. INTRODUCTION

Vegetable proteins have been extensively used as emulsifying agents for the production of o/w food emulsions such as mayonnaise or salad dressing ensuring full or partial replacement of egg protein'.8. There are several factors and conditions that influence the emulsifying capacity of proteins, such as equipment design, rate of oil addition, temperature, pH, protein source and structure, solubility and concentration, nature of the oil used, salt (type and concentration), sugars and water content'. The optimisation of Lupinus albus protein emulsifying conditions - time and speed - was previously carried out'. The effect of thermal denaturation of lupin protein on its surface hydrophobicity and emulsion properties has been already studied' as well as the influence of the proteidoil ratio on the physical properties of the respective emulsions'. The optimisation of the chemical composition of the continuous phase (salt and acid concentrations) should be carried out bearing in mind the customers' sensory perception, as this will determine the range of variation of each variable. For commercial mayonnaise the pH value is well defined at about 4.2 and NaCl concentration is usually around 1% wt. Nevertheless, it is important to understand how the variation of these parameters can modify the physical behaviour of the emulsion. In this work the capacity of lupin protein to stabilise o/w emulsions under different pH and salt contents was investigated.

Proteins and Emulsions

35 1

The modificati.on ,of the pH. of the continuous phase affects the net charge of the protein molecule and therefore the conformation of these macromolecules, altering protein electrostatic interactions. Consequently protein functionality, namely its solubility and its capacity to form and stabilise emulsions is also rn~dified~*~-". Around the protein isoelectric point (IP), the net charge of the protein molecules is zero, attractive forces predominate and molecules tend to associate'*, reducing the interaction between protein and water and therefore its solubility. At pH values below the IP, the net charge of protein molecules is positive, enhancing their interactions with anions" and at pH values above the IP, the protein global charge is negative, inducing interactions with positive particles. The presence of salt in aqueous solutions alters water activity and also decreases the electrostatic interactions between proteins, altering their solubility and its emulsion proper tie^'^. This effect of salt is highly dependent on the pH of the solution and is not easy to predict. 2. MATERIAL AND METHODS

2.1. Preparation of the emulsions Oil in water emulsions, having a constant sunflower oil concentration of 65% wt, were prepared with lupin protein as emulsifier. The lupin protein isolates (L9020) were kindly offered by Miffex Anlangenbau GmbH (Germany). For the pH study emulsions with 2,3, 4, 5 and 6% of protein were prepared and the pH was set with acetic acid or sodium hydroxide to values ranging from 3 to 8. For the salt study lupin emulsions with 4.5 % wt of protein, 65% wt oil and NaCl contents ranging from 0.3 to 4% wt were prepared. The emulsification was processed according to a previous study (Franc0 ef ul., 1998a), using a rotor stator Ultra Turrax T-25 from Ika (Germany) at 14250 rpm, during 5 minutes.

2.2. Droplet size distribution @SD) DSD measurements were conducted in a Malvem Mastersizer X (Malvem, UK). Values of the Sauter diameter, d32rwhich is inversely proportional to the droplet specific area, were obtained as f01lows'~: d32 =

znidi3

(1)

Znidi2 where ni is the number of droplets with diameter di.

2.3. Dynamic viscoelasticity and steady state flow measurements Oscillatory measurements and steady state flow curves were performed in a controlled stress rheometer (RS 75) from Haake (Germany). Frequency sweep tests were conducted

within the linear viscoelastic range of each emulsion, using a cone and plate sensor

Gums and Stabilisers for the Food Industry 10

352

system (C35/2 - 35mm of diameter and 2" angle) in a frequency range from 0,03 to 100 rad/s. The plateau modulus, Go,, approximately given by the value of G' when tan 6 reaches the minimum value, as proposed elsewhere":

G; =[GI]

IU,,d+

,,,,,,,,,,,,,,,

was be obtained from the mechanical spectra. This parameter can be considered as a measure of the intensity of the entanglements, between protein molecules, inside and outside the o/w interface" and has also been interpreted as a measure of the complexity of the emulsion ~tructure".'~. Steady state flow curves were obtained with a serrated parallel plate system to overcome the slip effect as it was previously The emulsion flow behaviour was fitted (r' > 0,95) to the Carreau model: s

r

1

(3)

where fC is the critical shear rate for the onset of the shear-thinning behaviour and s is a parameter related to the slope of this region, qoand qmare the limiting viscosities of the first and second Newtonian region of the curve.

2.4. Textural parameters Firmness and adhesiveness were the textural parameters considered, obtained from the Textural Profile Analysis (TPA) in the texturometer TA-XT2 from Stable Micro System (UK). Penetration tests were performed with a cylindrical probe in a load cell of 5000g and 2 mm/s crosshead speed. Firmness was considered as the maximum resistance to the penetration of a cylinder with 38mm diameter at 5mm depth of the material contained in a 53 mm diameter and 45 mm height glass cylindrical flask. Adhesiveness is a characteristic of sticky materials and can be defined as a resistance of the material when the probe is recessing. This parameter is recorded as a negative area and is evaluated as the work necessary to take the probe out of the material. All measurements were replicated at least three times and conducted at 20fl"C. 3. RESULTS AND DISCUSSION

3.1. Effect of pH 3.1.1. Droplet size distribution. The effect of pH on droplet size distribution of lupin protein emulsions with 2% protein is represented in Figure 1, exemplifying the effect of pH at a low concentration of protein. For higher concentrations the pattern is similar. The variation of the subsequent Sauter diameter (d,?) with pH is represented in Figure 2 for all the protein concentrations studied.

Proteins and Emulsions

353

Figure 1: Eflect ofpH on the droplet size distribution curves of lupin protein emulsions with 2% protein content und 65% sunjlower oil.

-.-2%@&1 -0- 3%@&

-A--4%@dn 5%@& -a- 6%@en

--v-

6-

4-

21 2

. , . , . , . , . , . , . , 3

4

5

6

7

0

9

PH Figure 2: Vuriation of Suuter diameter (d32) with pH. for lupin protein emulsions with diyerent protein contents

354

Gums and Stabilisersfor rhe Food Industry 10

For the whole range of protein concentrations studied. the effect of pH on droplet size distribution was similar. Near a value of pH 4.5, i.e., at the isoelectric point (IP) of this pr~tein”.’~,the curves show high values for droplet size, which decreases for pH values far from the IP. The effect of pH on the average droplet diameter can be compared in Figure 2 , for all protein concentrations studied a negative polynomial variation of d,? with pH, reaching a maximum of d,, around the IP, was observed. In addition, a decrease of the Sauter diameter with protein content can be seen (Figure 2). This is related to the increase of emulsion stability with protein content, as was previously observed’“ as a result of more protein molecules being available for the interface, preventing droplet coalescence. This means that emulsions with pH values near the IP have larger oil droplets and would be less stable against creaming. 3.1.2. Rheological measurements. The effect of pH on the evolution of the viscoelastic functions storage (G’) and loss (G”) moduli, and on the steady state flow curves, can be observed in Figure 3 for lupin emulsions with 2% protein.

The tendency of the mechanical spectrum to show a plateau region for G’ for the whole range of frequencies studied, is typical of the flocculated This was explained as the development of the entanglement network between adsorbed and nonadsorbed macromolecules, producing a gel-like structured system”. The viscosity curves show a similar shape for all the conditions with an extended plateau region. This represents the resistance of the material to flow, which is also related to the development of the entanglement network. The emulsions prepared at the IP (figure 4) showed lower values of the plateau modulus (G”,) and low zero shear rate viscosity (17”). It can also be observed that the values of Go, and qoincrease with protein content. For all protein concentrations, G”, and qo present polynomial variation with pH, with a minimum around the IP. 3.1.3. Texture properties. The effect of pH on firmness and adhesiveness of the lupin protein emulsions with various protein contents (Figure 5) can be explained again by a polynomial variation. Both properties show a minimum value around the IP.

Clearly, at pH 4.5 all the physical parameters, either fundamental (GO, and q,) or empirical (firmness and adhesiveness) express lesser protein emulsion performance. These results are well supported by the oil droplet size measurements, which showed an increase of d,? at values of pH around IP. This can be explained by the neutral global protein charge at this pH. This neutral nature of the protein molecule reduces electrostatic repulsion with subsequent tendency for protein aggregation with a strong impact on diminishing protein s ~ l u b i l i t y ~ ~These ’ ~ . phenomena would be responsible for slow migration of protein molecules to the o/w interface, reducing the emulsifying capacity of this component. They also explain the dependence of emulsion properties on protein concentration. The smaller affinity of protein molecules for the o/w interface explains the larger size of the oil droplets as the emulsifier itself is not enough to prevent coalescence and stability is lowered. Nevertheless, the protein in the continuous phase should be responsible for some kind of three-dimensional structure as the material still shows a weak gel like structure.

355

Proteins and Emulsions

2% prctein

Figure 3: Eflect of p H on the ffequency .vweep tests (u) und on the viscosiiy curves (b) of lupin protein emulsions with 2 % protein content und 65% sunflower oil.

Gums and Stabilisersfor the Food Industry 10

356

I

I

8

6

4

PH

--m- 2%pdein .-3%p~ien

-A- 4%pctejn

-v- EPhpden

~

2

'

l

3

~

4

,

~

5

PH

l

6

.

l

7

a

.

l

r

l

9

Figure 4: Vuriuiion of (u) pluieuu modulus (GON) and (b) limiting viscosity (qO ) with pH. for lupin protein emulsions with d@ereni proiein confenis.

.

,

357

Proteins and Emulsions

a-

s .

E ":

w=

m-

I -.-

7

0 ; . 2 .

2Y-n

-@- 3Y-n

-A- 4Y-n

-'I5Y-n -

I

3

.

I

4

,

I

5

, . PH6 .

I

7

.

I

8

9

Figure 5: Vuritrtion of the firmness (uJ und udhesiveness (b) with pH, for lupin protein emulsions with dflerent protein contents.

Gums and Stabilisers for the Food Industry 10

358

3.2. Effect of salt (NaCI) concentration 3.2.1. Droplei size distribution. The effect of NaCl concentration at three different pH values of 4.5% wt protein solutions on droplet size distribution of 65% oil olw emulsions is represented in Figure 6. The effect of NaCl content on d,? values is represented in Figure 7. Different responses to salt can be observed for the three different pH values considered. At a pH value lower than IP. it is difficult to assign a common pattern to the DSD of the o/w emulsions with different salt concentrations (Figure 6a). At pH near the IP and at a higher pH value the pattern is regular, practically independent on salt (Figure 6b and c). The oil droplet Sauter diameter is represented in Figure 7 as a function of salt content for these three values of pH. At a pH value lower than the IP, d,, shows a parabolic dependence on salt with a minimum at the lowest salt concentration. For the other two pH values, there is no significant dependence of the oil droplet on the presence of salt (p< 0,05). At the values higher than IP, the oil droplets are smaller and the emulsions would be more stable. 3.2.2. Rheologicul measurements. In Figure 8, the evolution of storage (G'), loss (G") and the plateau (GoN)moduli and zero shear rate viscosity (q") with the salt content of the lupin protein emulsions for pH 6.5 are represented.

An increase of the viscoelastic functions and of the zero shear rate viscosity with the salt content can be observed showing a positive linear correlation between GoN(Y) and salt concentration (X), that can be expressed by the following: Y = 1794 + 290 X (r2 = 0.92) (4) A similar correlation could be found between qo(Y) and salt content (x):

Y = 4.8 xiOs + ~ 2 ~ x1 (& 0 ~= 0.96)

(5)

Using all the previous rheological and textural results (for the study of pH and salt), it is possible to correlate G", with firmness as: G", = 191.8 + 9 . 3 firmness ~

(r2 = 0.95)

(6)

From the previous mathematical model it will be possible to estimate the values ofthe plateau model for emulsions with known firmness. This dependence was expected since, the plateau modulus is a measure of the degree of structure and firmness and is a mechanical manifestation of this structure. This demonstrates that the fundamental rheological parameters are well correlated with the texture determinations. 3.2.3. Texturul parumeters. The textural parameters are strongly dependent on the salt concentration of the emulsion for pH values far from the IP and are not dependent on salt concentration at the IP (Figure 9).

Proteins and Emulsions

359

Figure 6: Eflect of NaCl concentration on the droplet size distribution curves of lupin prolein emulsions prepared at pH 2 5 (a), 4 5 (3) and 6 5 (c). with 4 5% protein content and 65% sunflower oil

5

1

0

360

Gums and Stabilisers for the Food Indusrry 10

Figure 6: (c)

0 Q1

I ._

I

1

I]

.

pH 2.5

-

A pH 6.5

pH 4.5

14

12

10

-Eg -

2

U

8 6

4

2

0 0.0

0,5

1,0

1.5

2,0

2.5

3.0

3.5

4.0

4.5

5,O

% NaCl (mlm)

Figure 7: Variation ofd32 with NaCI concentrution. for lupin protein emulsions prepared at pH 2.5. 3.5 und 6.5.

361

Proteins and Emulsions

2%protein

10'

1 10.2

'

'

' " ' " 1 1

10-1

'

' """I

'

".""I

100

10'

-

'

"""I

'

'

1

102

W(radls)

Figure 8: Effect of NuCI concentrution on the jkquency sweep curves (a) and on viscosig curves (b) of lupin protein emulsions prepared at pH 6.5, with 4.5% ofprotein content and 65% of sunflower oil.

.

362

,pH2.5

350

0 0,0

0,5

1,0

Gums and Stabilisers for the Food Industry 10

,pH6.5

pH4.5

1,5

2,0

2,5

3,O

A

3,5

5,O

4,5

4,O

%w ( W h )

450

400 350

Y

pc 250 al

2 200

A I

u)

al

$

150

100 1

0

0

50

0 0,O

0,5

1,0

1,5

2,O

23

3,O

33

4,O

43

5.0

%IFwT ( W M )

Figure 9: Effect of NaCl concentrution on the firmness fu) und adhesiveness (b) of lupin protein emulsions prepared at pH 2.5. 4.5 und 6.5. with 4.5% protein content and 65% sunflower oil.

363

Proteins and Emulsions

The increase of firmness and adhesiveness with salt concentration at pH 6.5 (Figure 9) is also reflected by an increase of GoNand qo (Figure 8) with salt. This can be explained by an increase of protein solubility with salt content when the continuous phase has a pH higher than the protein IP. This increase in protein solubility was well demonstrated in beef heart myofibrillar proteins”. The solubility of these proteins was investigated at pH values ranging from 1 to 10 and at salt concentrations from 0.1 to 1.O M NaCl. A similar phenomena has been previously reported for soy proteins6 and a similar behaviour was found for coconut proteins12. This increase in solubility with the addition of NaCl is due to a higher degree of unfolding of the protein, as NaCl screens electrostatic interactions between positive and negative sites in the protein molecule, improving its ability to interact with water or other protein molecules. This will promote protein migration to the o/w interface, reducing d,? values (Figure 7). Interactions between protein molecules at the interface and of these with proteins at the continuous phase will result in a better structured material (higher values of texture and rheological parameters) with an increase of gel like structure in the continuous phase. At this pH value (6.5) the protein molecule shows a negative net charge. The electrostatic layer would preferentially contain Na’ cationst4. These are relatively small ions which should not be big enough to prevent protein molecules interacting with each other. At a pH value of 2.5, lower than the protein IP, it is well apparent in Figure 9, that a decrease of texture parameters with salt content and a reduction of emulsion structure occurs. In addition, the oil droplet size of these emulsions showed a strong increase with salt concentration. This can be explained by lesser protein solubility at increasing salt contents which was already observed for other proteins, when the pH of the media is lower than the IP of the protein’. Additionally, for the lupin protein a shift of the IP towards lower pH values in the presence of NaCl was reported”. The reduced solubility will change the o/w interface as the protein migration will be reduced, allowing coalescence of the oil droplets and larger droplet diameters. In the continuous phase the interaction between protein molecules and water is reduced and protein-protein interactions are also limited. The presence of a positive protein net charge with a subsequent C1- electrostatic layer may act as a shield, as C1- is a substantially bigger ion, being another contribution to the limitation of protein-protein interactions with subsequent weakening of the emulsion structure. When the emulsion is prepared using an aqueous solution with a pH near the protein IP value, the texture expressed in terms of firmness and adhesiveness and the d,, do not change with salt concentration. This absence of any dependence on salt concentration can be explained by the fact that at the IP the protein shows minimal solubility and this was found to be near pH 4.5 for this type of protein”. This probably means that the o/w interface is not “saturated” with protein molecules and, therefore, incapable of avoiding coalescence of oil droplets leading to high values of dA2.The addition of salt, when the protein net charge is null, will not make a difference. 4. CONCLUSIONS

In conclusion, the texture measured by firmness and adhesiveness, the structure expressed in terms of G”, and q,, and the stability indirectly evaluated via the oil droplet

364

Gums and Srabilisers for the Food Industry 10

size diameter of the lupin protein o/w food emulsions are strongly dependent on pH value and salt content of the continuous phase. The lowest values scored were found for a pH value around the protein IP and for high salt concentrations at pH values lower than the protein 1P. This work points towards preferential formulations of o/w protein stabilised emulsions. Namely, food emulsions must have pH values around 3.5 - 4.2 and salt concentrations of cu. 1%” which are near the critical zone for emulsion stability and structure.

5. ACKNOWLEDGEMENTS This work was financially supported by the Portuguese PhD grant Praxis XXI BD/5754/95. 6. REFERENCES 1. Tornberg, E., J. FuodSci., 1987,29, 867-879. Rossi, M.. Paliarini. E., Peri, E., Lebens. Wiss Technol., 1985, 18,293-299. Toro-Vazquez, J. F., Regenstein, M. J., J. rood Sci., 1989,54, 1 177-1 185. Mom, C. V., JAOCS, 1990,67: 265-271. Elizalde, B. E., Bartholomai, G. B., Pilosof, A. M. R., Lehens. Wiss. Techno[, 1996.

2. 3. 4. 5.

29, 334-339. 6. McWatters. K.H., Holnies, M.R., J. FoodSci., 1979, 44, 770-781. 7. Franco. J. Raymundo, A.; Sousa, I. & Gallegos, C., J. Agric. Food Chem., 1998, 46, 3109 - 3115. 8 . Raymundo, A.; Franco, J.; Gallegos. C.;Empis, J. e Sousa, I., Nuhrung Food, 1998, 42,220-224. 9. Raymundo, A.; Empis. J. & Sousa, I., Pol. J. Nutr. Sci., 1998, 48, 127-134. lO.Eisele, T.A. and Brekke,J. FoodSci., 1981,46, 1095-1 102. 1 l.Aoky, H., Taneyama, O., Inami, M., J. Food Sci., 1980,45, 534-546. 12.Phillips, L., Whitehead, D., Kinsella, J., ‘Protein Films’ In Structure-Function Properties of Food Proteins. Eds Academic Press, 1994, pp. 1 1 1-1 30. 13.Kwon, K.S. and Rhee, K.C., JAOCS. 1996,73, 1669-1673, 14.Kinsella, J.E., JAOCS. 1979. 56, 242-258. 15.Mc Clements, D. J. Food Emulsions: Principles, Practice and Techniques. Ed. CRC Press USA, 1999. 16.Sprow. F.B.. Chem. Eng, 1967. Sci. 22,435-442. 17.Wu, S. ,J. Polym Sci., 1989,27,723-741. 18.Ferry, J.D., ‘Viscoelastic Properties of Polymers’, 3rd ed.. John Wiley, New York, 1980. 19.Franc0, J., Gerrero. A., Gallegos, C.. Rheol. Acfu., 1995,34, 513-524. 20.Franc0, J.M. Berjano, M., Gallegos, C., J. Agric. Food Chem.. 1997,45, 713-719. 21.Barnes, H. A,. J. Newtoniun FIuidMech., 1995,56,221-251. 22.Franc0, J.M., Gallegos, C. and Barnes, H. A., J. Food Eng. 1998,36, 89- 102. 23.Sgarbieri, V . and Galeazzi, M., J. Agric. Food Chem. 1978,26, 1438-1442. 24.King, J.. Aguirre, C. and Pablo, S.. ,J, Food Sci., 1985, 50, 82-87.

Proteins and Emulsions

365

25.Chang0, A., Villaume, C., Bau, H.M., Nicolas, J.P. and Mejean, L. Food Reseurch International.,1995,28,9 1-99. 26.Alamanou, S. and Doxastakis, G . ,Lehens. Wiss. u. Technol., 1995,28,641-643. 27.Dickinson. E. and Yamamoto, Y.,J. FoodSci. 1996,61,811-816. 28.Dickinson, E., Hong, S. T., J. Agric. Food Chem., 1995,43,2560 - 2566. 29.Demetriades, K., Coupland, J.N., McClements, D.J., J. FoodSci., 1997,62, 342-347. 30.Sousa I., PhD theses, Uninersity of Nottingham, 1995. 31.Sousa, I. M. N., Morgan, P. J., Mitchell, J. R., Harding, S. E., Hill, S. E., J. Agric. FoodChem., 1996,44,3018-3021. 32.Nunes, M.C., Raymundo, A., Empis, J., Sousa, I. (Submitted) Physical Characterisation of mayonnaise.

EFFECTS OF LIPID ON WHEY PROTEIN GELATION

Shinya Ikeda and E. Allen Foegeding Department of Food Science College of Agriculture and Life Sciences North Carolina State University Raleigh, NC. 27695-7624 USA

1 INTRODUCTION Whey protein isolates (WPIs) and concentrates (WPCs) are commonly used as food ingredients for controlling the quality of food products.' Heat-induced gelation is an important functional property of whey protein ingredients since gelation contributes to consumer acceptances such as the appearance, water holding, and textural properties of foods.2 Heating a WPI or WPC dispersion causes proteins to partially unfold, exposing interior reactive regions and sulphydryl groups. Denatured proteins can aggregate via intermolecular interactions, namely, hydrophobic and electrostatic forces, disulfide bonds, and hydrogen bonds. Under appropriate conditions, protein aggregates form a continuous three-dimensional structure that entraps and restricts the motion of solvent.3 WPIs and WPCs consist of several components, i.e., proteins, mineral salts, lipids, and sugars. The four major proteins in bovine whey-based ingredients are Plactoglobulin (PLG), elactalbumin, bovine serum albumin, and immunoglobulins; PLG dominates gelation properties4 Network structures and rheological properties of heat-induced gels strongly depend on composition.5 Particularly, effects of mineral salt content have been intensively investigated.2.6 At pH far from the isoelectric point of the protein, where electrostatic repulsive forces among protein molecules are substantial, the proteins show a tendency towards linear aggregate formation, leading to a transparent fine-stranded network formation. When ionic concentration is increased, a random aggregation occurs since charges on protein surfaces are shielded, forming an opaque particulate network. On the contrary, studies concerning lipid effects on gelation show somehow contradictory results. Since the amount of lipids in whey protein ingredients was negatively correlated with gel strength? residual lipids were considered to inhibit whey protein gelation by competing with hydrophobic protein interaction sites. However, adding phospholipids (lecithin) to WPI dispersions increased the elastic rigidity of heat-induced gels.* The objective of this study was to determine how lipids affect heat-induced gelation of whey proteins. We compared fundamental fracture properties, water holding capacity, and small-strain rheological properties of heat-induced WPI gels formed in the presence and in the absence of lecithin. Far UV circular dichroism (CD) was used to study the effects of phosphatidylcholine (PC) and the predominant fatty acids in whey protein ingredients9 on structural changes in PLG during gelation.

2 MATERIALS AND METHODS 2.1 Materials

A commercial WPI (92.5% w/w protein; Davisco Foods International. MN) was used. PLG (L-3908), egg yolk lecithin (p9671), L-a-phosphatidylcholine (PC) (P-7318),

Proteins and Emulsions

367

sodium butyrate (B-5887), sodium oleate (0-7501), and sodium palmitate (P-9767), were purchased from Sigma Chemicals. All other chemicals were of reagent grade quality.

2.2 Sample Preparation 2.2.1 Fracture and Water Holding Capacity Testing. WPI and lecithin were hydrated in 20 mmoYdm3 bis-tris buffer containing appropriate amounts of NaC1. The suspensions was adjusted to pH 7.0 and diluted to 10% w/v protein with appropriateNaCl solutions. After degassed under vacuum, the suspensions were poured into siliconeprecoated glass tubes (19 mm in diameter). Gels were formed by heating in an 80°C water bath for 30 min. After cooling to room temperature, gels were held overnight at 4°C. The following day gels were equilibrated to mom temperature for testing. 2.2.2 Dynamic Rheological Measurements. WPI or #MA3 suspensions containing lecithin. PC. or fatty acid were prepared similarly to those for fracture and water holding testing. 2.2.3 CD. pLG, PC, or fatty acid was hydrated independently in 10 mmoYdm' trisHC1 buffer (pH 7.0) containing appropriate amounts of NaC1. Following degassing under vacuum, the protein concentration was determined spectrophotometrically. The PC dispersions were sonicated (120 W) at 70°C until the suspensions became sufficiently clear. The pLG solutions were mixed either with the buffers or the PC/fatty acids dispersions and then diluted to a final concentration of 1 mg/cm3 protein and 2 mg/cm3 PC or a fatty acid with appropriate buffers.

2.3 Fracture Testing Gels were cut into 29 mm long cylinders and then ground into capstan shapes with a center diameter of 10 mm. Gels were twisted to fracture at 2.5 rpm using a Gelometer (Gel Consultants, NC) calculating the values of true shear stress (fracture stress) and true shear strain at fracture (fracture strain) from the torque and angular displacements.10

2.4 Water Holding Capacity (WHC) WHC was measured using the microcentrifuge-based method.10 Gels cut into cylinders (1 cm in height and 0.48 cm in diameter) were spun at a relative centrifuge force of 153 g for 10 min. The WHC values were calculated as a ratio of weight of water held in gels after centrifuge to weight of protein in the gels.

2.5 Dynamic Small-strain Rheological Measurement A Bohlin VOR rheometer (Bohlin Reologi, NJ) was used with a concentric cylinder test fixture, consisting of a oscillating cup and a fixed bob attached to a 1.82, 13.2, or 42.2 gcm torque bar. Gels were formed by heating the samples from 25 to 80°C at 2.5"C/min, holding at 80°C for 30 min. cooling to 25°C at l"C/min, and then holding at 25°C for 1530 min. During the thermal processing of protein suspensions, storage modulus (G') was recorded at a frequency of 0.05 Hz and a maximum strain of 0.01.11J2 Frequency sweep measurements were done for the final gels at the maximum strain of 0.01.

2.6 Circular Dichroism (CD) CD spectra of the samples placed in a 0.1 mm path length quartz cell were recorded on a Jasco 5-600 spectropolarimeter (JASCO, Japan). The spectra were collected in 0.2 nm steps at a rate of 50 ndmin over the wavelength range 180-260 nm and truncated at a wavelength where a transmittance became less than 1%. Each spectrum was acquired five times and the results of two independent measurements were averaged and then smoothed. Molar circular dichroism, A&, was obtained by dividing the measured ellipticity by the residue-average molecular weight of the protein. Secondary structure contents were estimated using the Fortran program Varselec.12

368

Gums and Stabilisersfor the Food Industry 10

3 RESULTS AND DISCUSSION

3.1 Fracture Properties Gels formed at 30 mm0Ydm3 NaCl without adding lecithin were translucent and showed low fracture stress (af)values and high fracture strain (R) values (Figure 1): characteristics of a fine-stranded gel network.2 High uf and low at 80 moYdm3 NaCl and low uf and low n at 500 mm0Ydm3 NaCl suggest a mixed and a particulate gel network were formed, respectively. Adding lecithin significantly increased uf at 30 mm0Ydm3 NaCl. At 80 mm0Ydm3 NaCl, ufwas almost independent of lecithin content and lecithin addition decreased of at 500 mm01/dm3 NaCl. At all NaCl concentrations examined, R decreased with increasing lecithin content (Figure l), which is a typical effect of adding filler particles to a gel network." It is also known that, if filler particles interact with the gel network, aftends to increase with increasing volume fraction of the filler.I3 Therefore, lecithin seems to act as an active filler at low salt concentration, reinforcing the mechanical strength of the gels. On the other hand, adding lecithin at high NaCl concentration appeared to cause demixing into protein- and lipid-rich phases, interfering with a continuous gel network formation.

3.2WHC Decreasing WHC with increasing salt concentration (Figure 2) is considered as a result of increasing pore size in gels.3 Lecithin addition decreased WHC values in the whole NaCl concentration range measured, suggesting that gel networks shifted toward particulate ones with larger pores. It is also possible that gel networks into which lipids are incorporated are less hydrophilic. 50

-

B3 B vl

40

30

7 1 1 °

1

-

50 20 -

rt'

10 -

0' 0.5

1

1.5

2

2.5

3

1

4' 10

3.5

'

'

' " " ' 1

'

100

loo0

NaCl [mmoVdm3]

Fracture Strain

Figure 1 Effect of lecithin on fracture properties of 10% w/v protein WPI gels containing 30 (circle),80 (triangle), or 500 mmol/dm-'NaCl (square). Numbers: lecithin content (% w/v).

'

Figure 2 Effect of lecithin on water holding capacity (WHC) of 10% w h protein WPI gels. Lecithin content: 0% w h (circle); 10% w h (square).

3.3 Rheological Transitions during Gelation Figure 3 show examples of gelation curves while heat is applied to W I dispersions. In all cases, G' values started to increase during the temperature holding processing at 80°C. A drastic increase in G' occurred during cooling to 25°C but G' values remained similar when samples were held at 25°C. Values of G' at the end of the holding at 8OoC (G'80) and following the entire thermal treatment (G'=) are summarized in Table 1.

369

Proteins and Emulsions

Adding lecithin increased G at low salt concentration and decreased G' at 500 mmoYdm3 NaCl. Similar results were obtained when PC was added to pLG (Table 2) while butyrate substantially increased G regardless of NaCl concentration. The time when G started to increase after the temperature reached 80°C was defined as gelation time t . A shorter tg in the presence of lecithin or PC at low NaCl concentration (Tabfes 1 and 2) suggests that both PC and lecithin facilitate disulfide bonding and/or hydrophobic interactions between protein molecules during heating. At low ionic concentration, the probability of forming an aggregate on a collision between protein molecules is low (reaction limited aggregation) due to the high energy barrier for aggregation produced by electrostatic repulsive forces between protein molecule^.^^ Since PC is isoelectric across a wide pH range,ls electrostatic repulsions are essentially absent between PC and proteins. Adding interactive particles such as PC would increase the overall probability of aggregate formation, accelerating aggregation during heating. 140001

I

I

I

8

I

I

I

1

1

90

c 1 70 e

loo00

i5

t

F

Figure 3 Effectof lecithin on G' development in 10% w h protein WPI samples. Solid line: temperature.

Time [s]

Negative effects of lecithin or PC addition on rheological properties of gels were observed in the presence of 500 mmoYdm3 NaCl (Tables 1 and 2). If negative charges on protein molecules are effectively shielded due to high ionic concentration,every collision of protein molecules leads to an aggregate formation (diffusion limited aggregation

Table 1 Effects of lecithin on rheological properties of 10% w h protein WPZ gels. NaCl [mmoUdm3] 30 30 100 100 500 500

Additive

G',[Pa]

-

490

10% w/v Lecithin

4580

-

6670 10900 1460 1040

10% w/v Lecithin

10% w/v Lecithin

G'= [Pa]

r"s1

1960 12100 20300 33100 7190 5210

870 150 150 90 210 390

n

0.041 0.03 1 0.042 0.034 0.045 0.054

Table 2 Effects of phosphatidylcholine (PC) and butyrate on rheological properties of 12% w h plactoglobulin gels. NaCl [mmoYdm3]

Additive

G', [Pa]

0

-

0 0 500 500 500

12% wlv Pc 12% w/v Butyrate

320 3890 6570

-

1440

12% w/v Pc 12% w/v Butyrate

620 6770

G'z [Pa]

rg [sl

1520 6910 22400 8730 4030 23100

645 270 45 270 375 60

n

0.033 0.020 <0.001 0.033 0.047 <0.001

370

Gums and Stabilisers for the Food Industry I0

process).l4 The presence of inactive fillers that are not incorporated into the gel network is expected to decrease the probability of collisions between protein molecules and thus delay the gel formation. Sodium chloride addition did not influence the effects of butyrate on the rheological properties of heat-induced pLG gels (Table 2). Adding butyrate significantly increased G' and decreased tg at all NaCl concentrations tested. Since adding sodium chloride further than 500 mmoYdm3 is known to reduce G' and gel strength,2J1 the increase in G' with an addition of 12% w/v sodium butyrate (about 1 moYdm3 sodium ions) was unexpected. Therefore, fatty acids bound to FLG appeared to act against electrostatic shielding effects of salts probably because bound fatty acids brought additional negative charges on the surface of a [pLG-fatty acid] complex. The power law exponent n of the relationship between G' and frequency (Tables 1 and 2) is regarded to be an indication of the viscoelastic nature of the gel: n is zero for purely elastic gels and increases with larger contributions by the viscous component. All n values are less than 0.06, assuring that all of the gels are predominantly elastic in nature. Adding lecithin or PC at low NaCl concentration reduced the n values of the resulting gels, indicating the formation of more elastic networks. Meanwhile, gels with added lecithin or PC at 500 mm0Ydm3 NaCl were less elastic. Adding butyrate produced very elastic gels characterized by almost frequency independent G' values at all NaCl concentrations tested.

3.4 Secondary Structural Transitions during Gelation The far UV CD is dominated by contributions due to chiral interactions between the peptide chromophore and residues in the immediate environment and is particularly suitable for secondary structural analyses in the presence of lipids or fatty acids.16 All of the fatty acids and PC induced changes in pLG CD spectra in the absence of added salts (Figure 4). The estimated secondary structure contents suggested increases in a-helical structures and decreases in psheet in the presence of PC or fatty acids (Table 3). Increased chain length was correlated with larger secondary structure changes (palmitate oleate > butyrate), suggesting that fatty acids interact with PLG mainly through hydrophobic interactions.'7 pLG is believed to have one hydrophobic binding site in the calyx formed by the /+barrel'* but is also capable of binding a large number of fatty acids weakly at the hydrophobic surface site in the region of the a-helix (residues 130-140) and pstrands A (16-27) and I (145-150).19 Formation of an amphiphilic helix was also proposed as a mechanism for interaction between oil and PLG.20 Therefore, it is probable that binding of fatty acids and PC prevents heat-induced helix unfolding or even induces helix reformation from unfolded regions. In 500 mmoYdm3 NaCl, since spectra were obtained only down to around 195 nm. secondary structure contents were not estimated. PC did not change the PLG spectrum down to 198 MI (Figure 5). At this condition, pLG suspensions tended to separate into protein- and lipid-rich phases, suggesting that PC did not bind to pLG. Therefore, the absence of changes in the CD spectrum is reasonable. In the presence of fatty acids, the shape of the spectra became featureless (Figure 5). suggesting that fatty acids substantially destroy the secondary structure. Larger molecular weight fatty acids did so more effectively (palmitate oleate > butyrate), suggesting that hydrophobic interactions

-

-

Table 3 Effects of phosphatidylcholine and fatty acids on secondary structure contents of j3-lactoglobulin with no added salt at 80°C. Additive

Phosphatidylcholine Butyrate Oleate Palmitate

*Helix 0.11 0.13 0.19 0.22 0.23

,%Sheet

Turn

Other

0.38 0.37 0.23 0.24 0.27

0.15 0.17 0.23 0.24 0.22

0.34 0.34 0.34 0.30 0.28

Proteins and Emulsions

371

4

3

-

-- & 9

I

,&lactoglobulin

+ pboSpb.tldyl&Oline

+okakA 4

+ plmltate2-

-4 180

-4. 190

200

210

220

230

240

Wavelength [nm]

Figure 4 CD spectra of Plactoglobulin with no added salt at 8OOC.

250

260 190

' 200

210

220

230

240

250

260

Wavelength [nm]

Figure 5 CD spectra of p lactoglobulin in 500 mmoUdm3 NaCl at 80°C.

occur between fatty acids and PLG. Since both fatty acids and PLG are negatively charged at pH 7, electrostatic repulsive forces between [PLGefatty acid] complexes and unbound fatty acids increase with increased number of bound fatty acids per protein molecule. Therefore, fatty acid binding would be limited to a certain degree at lower ionic strength and increase with increasing ionic strength due to electrostatic shielding effects of salts. Since both a-helix-induced and unfolded PLG (Figures 4 and 5 ) were capable of forming gels with similarly high G' values (Table 2), a specific secondary structure does not seem a critical factor for gelation. Further investigation on alternations in tertiary structure of protein and their relevance with gelation mechanisms will be required. 4 CONCLUSIONS

When phospholipids were added to whey protein dispersions, heat-induced gels were either mechanically stronger or weaker, depending on ionic concentration. A substantial positive effect was shown for the addition of egg yolk lecithin on fracture and small-strain rheological properties of heat-induced WPI gels formed at low NaCl concentration, suggesting that phospholipids act as an active filler. This hypothesis was supported by results from secondary structural analyses of PLG: PC in low ionic strength prevented heat-induced loss of a-helix structures. Positive effects of lecithin on mechanical strength of WPI gels were not observed at >lo0 mmoYdm3 NaC1. A phase separation caused by lipids in the presence of high ionic concentration seems responsible for interfering with a continuous gel network formation. Consequently, the optimal level of lipids in WPI for gel formation should be determined by relative amount of mineral ions in the system. A process for reducing the lipid content in whey protein ingredients usually causes reduction of the mineral content. While the amount of lipids in whey protein ingredients is believed to be negatively correlated with gel strength, our results suggest that a high amount of mineral salts associated with lipids is also responsible for weaker gels containing a high amount of lipids. In addition, part of lipids in whey protein ingredients is bound to proteins endogenously (lipoproteins or milk fat globule membrane) or due to processing. Therefore, lipid reduction would also cause changes in protein composition, leading to changes in gelation properties of whey protein ingmhents.

Acknowledgements This work was supported by a grant from the Southeast Dairy Foods Research Center and

Gums and Sfabilisersfor the Food Indusfry 10

372

Dairy Management, Inc. International.

Whey Protein Isolate was a gift from DAVISCO Foods

References 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15. 16.

17. 18. 19. 20.

J. N. de Wit, J. Dairy Sci., 1998,81, 597. E. A. Foegeding, E. L. Bowland, and C. C. Hardin, Food Hydrocolloids, 1995.9, 237. G. R. Zielgler and E. A. Foegeding, Adv. Food Nutr. Res., 1990,34,203. M. E. Hines and E. A. Foegeding, J. Agric. Food Chem., 1993,41,341. C . V. Morr and E. A. Foegeding, Food Technol., 1990,44.100. A. H. Clark, ‘Functional Properties of Food Macromolecules 2nd Ed.’, S. E. Hill, D. A. Ledward, and J. R. Mitchell (Eds.), Aspen Publishers, Gaitherburg, 1998, Chapter 3, p 77. D. Karleskind, I. Laye, F.-I. Mei, and C. V. Morr, J. Food Sci., 1995.60.731. E. Dickinson and Y. Yamamoto, J. Food Sci., 1996,61,811. M. Vaghela and A. Kilara, J. Dairy Sci. 1996,‘79, 1172. S. Ikeda and E. A. Foegeding, Food Hydrocollids, 1999,13,239. S . Ikeda and E. A. Foegeding, Food Hydrocollids, 1999,13,245. S . Ikeda and E. A. Foegeding, in review. G. J. Brownsey, H. S. Ellis, M. J. Ridout, and S. G. Ring, J. Rheology, 1987,31,635. S . Ikeda and E. A. Foegeding, in review. D. G. Cornell and D. L. Patterson, J. Agric. Food Chem., 1989,37, 1455. E. M. Brown, R. J. Carroll, P. E. Pfeffer, and J. Sampugna. Lipids, 1983,18, 111. N. Yuno-Ohta, T. Higasa, E. Tatsumi, H. Sakurai, R. Asano, and M. Hirose, J. Agric. Food Chem., 1998,46,4518. M. Z. Papiz, L. Sawyer, E. E. Eliopoulos, A. C. T. North, J. B. C. Findlay, R. Sivaprasadarao, T. A. Jones, M. E. Newcomer, and P. J. Kraulis, Nature, 1986,324, 383. Q. Wang, J. C. Allen, and H. E. Swaisgood, J. Dairy Sci., 1998.81.76. M. Shimizu and M. Saito, ACS Symp. Ser., 1996,650, 156.

THE INFLUENCE OF DIFFERENT CALCIUM-SEQUESTERING SALTS ON THE HYDRATION CHARACTERISTICS OF RENNET CASEIN IN A SIMPLE MODEL SYSTEM. Michael P.Ennis, Ann Thornton & Daniel M. Mulvihill. Food Chemistry, Department of Food Science & Technology, University College, Cork, IRELAND.

1 ABSTRACT

A wide range of calcium-sequestering salts is available for use in cheese analogues and processed cheese manufacture. The effects of different calcium-sequestering salts, at equimolar concentrations, on the rheological behaviour of rennet casein during hydration in a simple model system were studied. The degree of protein hydration and solubilisation varied with the calcium-sequestering abilities of the salts. Hydration and solubilisation of rennet casein in the weak calcium-sequestering salt, trisodium citrate, was very limited and a dense cheese-like mass formed. In disodium orthophosphate greater hydration to form a curd-like mass of high viscosity index was observed. More effective calcium-sequestering salts such as pentasodium polyphosphate resulted in rapid solubilisation of the protein to give a solution with a low viscosity index, while disodium pyrophosphate and tetrasodium polyphosphate-based dispersions showed intermediate behaviour. The pH of the dispersions also influenced the hydration behaviour of the casein. At the low pH (6.38) of monosodium orthophosphate-based dispersions rennet casein remained as discrete particles, while at the higher pH (>9.0)of trisodium orthophosphate-based dispersions the casein was solubilised. The cation of the calciumsequestering salt also influenced the hydration behaviour of the rennet casein. The effects of the different salts on the hydration of rennet casein in the model system are discussed in relation to their influence on finctional properties such as texture and meltability of cheese analogues and processed cheeses prepared using the salts. 2 INTRODUCTION

Calcium-sequestering salts (also known as emulsifying salts or melting salts in the industry) are used in the manufacture of processed cheese to enhance the dispersion of cheese shreds included in the formulation, to provide uniform cheese structure and to control meltability of the heated cheese.',' In the manufacture of rennet casein-based Mozzarella cheese analogues calcium-sequestering salts are also essential to disrupt calcium-mediated cross-bridges present in the casein and so allow the protein to hydrate and suitably stabilise the dispersed oil phase of the cheese analogue.394Calciumsequestering salts commonly used include trisodium citrate, sodium phosphates and

314

Gums and Stabilisersfor the Food Industry 10

sodium polyphosphates although a much wider ran e of suitable salts is available to the cheese analoguelprocessed cheese man~facturer.'~Experience in the industry and numerous studies have shown that the nature and the concentration of the calciumsequestering salt employed influence hnctional properties such as hardness, stretchability and meltability of the finished cheese product^.^-^ Study of the influence of the calcium-sequestering salt on cheese hnctionality is, however, complicated as differences in processing conditions used in the manufacture of the cheese may result in variability in the hnctional properties of the cheese, even when the same formulation has been used. It has been shown, for example, that the level of oil emulsification achieved during cheese analogue manufacture (prepared to the same formulation and using the same batches of ingredients) affects the melting properties of the cheese. lo To study the effect of the type of calcium-sequestering salt on the protein component of the cheese analogue or processed cheese requires a simplified model, where the influence of factors such as degree of oil phase emulsification are eliminated. A simple rheological method for studying the hydration behaviour of rennet casein dispersed in a solution of a calcium-sequestering salt has been developed." The influence of calciumsequestering salt anion and cation on the hydration characteristics of rennet casein and some rheological properties of the protein dispersions obtained are reported here, and these characteristics are related to the reported hnctional performances of the different salts in cheese analogue and processed cheese manufacture. 3 MATERIALS AND METHODS

The protein, ash and moisture contents of the rennet casein used in this study conformed to EU regulation No. 2921190 for edible rennet caseins." To determine the calcium content of the rennet casein approximately 0.2 g was accurately weighed, in triplicate, into acid-washed flasks and dissolved in 10 ml of concentrated nitric acid with gentle heating. The solution was made up to 50 ml with 0.1 % Lac13 (Reagecon, Shannon, Co. Clare, Ireland), an aliquot was then diluted ten-fold with 0.1 % Lac13 and the calcium content of the diluted solution determined by atomic absorption spectroscopy using a Varian SpectrAA-100 instrument (Varian Australia Pty. Ltd., Mulgrave, Victoria, Australia). Calcium standards of 1 to 10 ppm were prepared from a stock calcium solution (Reagecon) for instrument calibration. The rennet casein used in this study contained 2.14 +I- 0.05 YO, wlw, calcium. Rheological changes that occurred as rennet casein was dispersed in solutions of the calcium-sequestering salts were determined, in triplicate, using the method of Ennis et ul." Solutions of the calcium-sequestering salts were substituted for the stock disodium orthophosphate (DSP) solution in the model system, such that each dispersion comprised 60 g of a 0.035 M stock solution of the appropriate calcium-sequestering salt in deionised distilled water, sufficient rennet casein to give 6 g protein, and water to a combined weight of 75 g. (In the case of DSP this gave dispersions equivalent to the 0.4 %, wlw, DSP dispersions used by Ennis et ul."). Stock solutions of the pyrophosphates and polyphosphates were prepared immediately before use to minimise the potential for hydrolysis of the anion. Calcium-sequestering salts were obtained from BDH, Poole, Dorset, UK; Sigma-Aldrich Chemical Co, Poole, Dorset, UK; Merck, D-6100, Darmstadt, Germany; Fisons Scientific Equipment, Loughborough, UK or ADM,

Proteins and Emulsions

375

Table 1. Calcium-sequesteringsalts used in hydration of rennet casein. Ca-sequestering salt

Source

Initial pH of salt solution

Merck ADM Merck BDH Sigma Merck Sigma Merck Fisons BDH

9.23 8.49 4.67 1 1.97 4.51 10.36 9.19 9.13 7.95 11.90

pH of rennet casein dispersion 7.20 7.44 (liquid phase) 6.38 9.20 6.27 8.25 7.60 7.35 7.20 9.50

*Although not a calcium-sequesteringsalt, NaOH was included in the study to assess the effect of pH on the hydration behaviour of rennet casein.

Ringaskiddy, Co. Cork, Ireland as shown in Table 1. The initial pH of each salt solution (60 g of a 0.035 M stock solution diluted to 75 g final weight with de-ionised distilled water) and the pH of each rennet casein dispersion following shearing and cooling to ambient temperature was measured using a Radiometer PHM 240 pH meter (Radiometer Analytical S. A., Lyon, France). 4 RESULTS

The choice of calcium-sequestering salt anion was found to be crucial in influencing the hydration characteristics of the rennet casein, as shown in Figure 1 and Figure 2. As previously noted by Ennis, et al." rennet casein dispersed in DSP (0.4 %, wlw, 0.028 M) characteristically exhibited swelling, aggregation, a distinct peak in viscosity index and a subsequent decrease in viscosity index with continued shearing (Figure la). At maximum viscosity index the rennet casein dispersion had an opaque, white appearance and formed an extremely viscous macroscopically homogeneous curd-like mass, while on continued shearing the viscosity decreased and following prolonged shearing a liquid somewhat similar in appearance to milk was obtained. The pH of the DSP solution decreased from an initial value of 9.23 to a final value of 7.20 following dispersion in it of the rennet casein (Table 1). Rennet casein dispersed in monosodium orthophosphate (MSP) exhibited quite different behaviour (Figure lb). Wetting and limited swelling of the rennet casein particles occurred, as was observed when the casein was dispersed in DSP. Aggregation of the dispersed particles and subsequent network formation leading to increased viscosity index, however, were found not to occur to any appreciable extent in MSP and discrete particles remained even following a prolonged period of shearing. The rapid, large increase in viscosity index characteristicof rennet casein dispersed in DSP (Figure la) was not observed. The pH of the MSP solution increased from an initial value of 4.67 to a final value of 6.38 following dispersion of the rennet casein.

376

Gums and Stabilisersfor the Food Industry I0

100

100

80

80

i3 a .g 60 b .I

cn

40

>

.d

1

b

X

4 G

60

.I

b

.I

O" 3

40

>

.I

20

20

0

0

100

100

80

80

d X

4

-

X

.g b

60

O" .*3

40

.d

4

.Z 60 b

~

.C(

cn

>

3cn

40

-

>

.I

20

20 -

0

07

Figure 1. Hy&ation projiles of rennet casein dispersed in (a) DSP, (b) MSP, (c) TSP and (4 NaOH. Dispersion of rennet casein in trisodium orthophosphate (TSP) resulted in a hrther difference in the rheological profile (Figure lc). Following initial wetting and swelling of the particles, hydration and ultimately solubilisation of the protein proceeded more rapidly and to a much greater extent than occurred in DSP. The peak in viscosity index characteristic of rennet casein dispersed in DSP was not observed on dispersion in TSP and the opaque white milk-like appearance of rennet casein dispersed in DSP was also absent. Instead, the dispersion was a translucent viscous fluid, typical of concentrated protein solutions. The pH of the TSP solution decreased from its initial value of 11.97 to a final value of 9.20 following dispersion of the rennet casein. Dispersion of rennet casein in NaOH to a final pH of 9.50 (Figure Id) resulted in a solution of caseinate with comparable appearance and viscosity to the final TSP-based dispersion.

311

Proteins and Emulsions

100

a

80 X

4

.El 60 b

‘8 3

>

40

.I

20

v 0

20

40

60

80

100

0

20

Time (min)

40

60

80

100

Time (min)

100

C 80

80

60

8 a .5 60 b

X

4 .G 0

d

.d

.I

3 40

40

s

1

v)

T

20

20 I

0

4

0

40

80 120 160 200 240

0

40

80 120 160 200 240

Time (min)

Time (min) Figure 2. Hy&ationprojiles of rennet casein d p r s e d in (a) DSPP, (6) TSPP, (c) PSPP and(@)TSC. When rennet casein was dispersed in disodium pyrophosphate (DSPP) the pH increased from an initial dispersant pH of 4.51 to a final pH of 6.27 (Table 1). At the end of the shearing period the dispersion had an opaque grey-white appearance (although substantially less opaque than the equivalent DSP dispersion) with numerous white particles still present. The hydration profile (Figure 2a) showed that only a small increase in viscosity index occurred and the dispersion did not exhibit a peak in viscosity index characteristic of rennet casein dispersed in DSP (Figure la). Dispersion of rennet casein in tetrasodium pyrophosphate (TSPP) resulted in a rapid, but relatively small increase in viscosity index, followed by a decrease in viscosity index with prolonged shearing (Figure 2b), similar to the hydration profile obtained when rennet casein was dispersed in NaOH. The pH decreased from an initial dispersant pH of 10.36to a final pH of 8.25 following dispersion of the casein (Table 1). The macroscopic

Gums and Siabilisersfor ihe Food Indusiry 10

378

appearance of the dispersion was similar to that of rennet casein dispersed in TSP or NaOH. Rennet casein dispersed in pentasodium tripolyphosphate (PSPP) hydrated rapidly (Figure 2c) and dissolved in a manner similar to that observed when dispersed in TSP and DSPP. The pH of the PSPP solution prior to addition of the rennet casein was 9.19 and this decreased to 7.60 following hydration and solubilisation of the rennet casein (Table 1). The use of trisodium citrate (TSC) solution as the rennet casein dispersant resulted in a rheological profile (Figure 2d) markedly different from the profiles obtained when other salts were used. Swelling of the rennet casein particles and aggregation was observed, however, imbibation of all of the available liquid (as occurred in DSP) did not occur. Instead the rennet casein particles clumped and formed a dense mass, generally similar in appearance and stretchability to a melted cheese, which was surrounded by free liquid. The rennet casein mass aggregated around the stirring paddle, resulting in the apparent plateau in viscosity index, The pH of the TSC dispersant was initially 8.49 while that of the free liquid remaining following dispersion and aggregation of the rennet casein was 7.44 (Table 1). The pH of the initial dipotassium orthophosphate (DPP) solution and of the rennet casein dispersion in DPP were similar to the values obtained for DSP alone and rennet casein dispersed in DSP, respectively (Table 1). Rennet casein exhibited similar hydration behaviour on dispersion in DPP as in DSP, both resulting in almost identical rheological profiles (Figure l a and 3% respectively). The hydration behaviour of rennet caseins when dispersed in diammonium orthophosphate (DAP), however, was markedly different (Figure 3b) from the hydration behaviour in either DSP or DPP. The initial pH of the DAP solution (pH 7.95) was less than the initial pH of the DSP solution (pH 9.23), but on dispersion of the rennet casein in both, the resulting final pH values were similar (pH 7.20).Although initial swelling of the rennet casein particles occurred in DAP, aggregation or clumping of the swollen particles was greatly reduced, when compared to the behaviour in DSP solution. The liquid phase of the DAP-based dispersion appeared grey and slightly opaque but the highly viscous opaque, white curd-like mass formed on dispersion in DSP and DPP was not observed in DAP, and discrete particles remained evident in the DAP dispersion even when shearing was prolonged for 100 min (Figure 3b). 5 DISCUSSION

Rennet casein is insoluble in water, due largely to calcium-mediated cross-bridging of the para-casein micelles and the colloidal calcium phosphate present in the aggregated micelles. In order for the protein to hydrate and stabilise the dispersed oil phase of a cheese analogue the calcium-mediated cross-bridges must be disrupted using calciumsequestering salts. The calcium-sequestering ability of the salts used is consequently an important factor in determining the hydration behaviour of the rennet casein. According to Caric & Kalab,’ the calcium sequestering abilities of the salts used in the present study were in the order: TSC < MSP < DSP < DSPP < TSPP < PSPP, while according to Shimp’ “MSP, DSP and TSP all have similar calcium-binding powers”. As the salts were used at equimolar concentrations in the rheological determinations in the present study, the amount of calcium sequestered from the rennet casein would be expected to reflect

Proteins and Emulsions

379

100

b

loo 80

80 X

4

.? .-b

60

.-

40

8

> 20

'

0

0 0

20

40

60

80

100

0

20

40

60

80

100

Time (min) Time (min) Figure 3. Hydation projZes of rennet casein dispersed in (a) DPP and (3) DAP. the calcium-sequestering abilities of the salts. The calcium content of the rennet casein used in the study was such that in each dispersion the calcium concentration was approximately 0.05 M, while each calciumsequestering salt was present at a concentration of approximately 0.028 M. Assuming a 1:1 molar interaction between the divalent phosphate anion of DSP and the divalent calcium cation, sufficient phosphate anion was present in the DSP model hydration system to sequester approximately 50 % of the calcium contributed by the casein. Disruption of a proportion of the calcium-mediated cross-bridges present in the casein by the DSP would facilitate hydration of the protein molecules to some extent and allow the casein particles to swell, but with sufficient cross-linking remaining for network formation to occur. The lower calcium-sequestering ability of MSP compared to DSP5 appears to have limited the extent of disruption of calcium-mediated cross-bridges in the rennet casein dispersed in MSP to levels too low to allow appreciable hydration and network formation. In contrast, the high level of calcium sequestration from rennet casein dispersed in TSP appears to facilitate complete dissolution of the protein molecules. Based on the hydration profiles obtained in the present study MSP, DSP and TSP do not appear to have similar calcium sequestering abilities. The different pH values of the salt solutions also influenced the hydration behaviour of the rennet casein. The pH values of the sodium orthophosphate solutions ranged from 4.67 to 11.97 but due to the buffering ability of the protein molecules, the pH of the rennet casein dispersions were within a narrower range of 6.38 to 9.20 (Table 1). At the low pH of the MSP dispersion the net negative charge on the protein molecules was low and electrostatic repulsion between molecules was low. Hydration of the protein molecules was thus less extensive leading to reduced particle swelling, decreased immobilisation of water and consequently low viscosity index, when compared to dispersions in DSP. When rennet casein was dispersed in TSP the high pH of the dispersion led to greatly increased hydration of the protein, and the rennet casein was effectively converted to soluble sodium caseinate, as O C C U K ~ when ~ the casein was dispersed in NaOH.

380

Gums and Stabilisersfor rhe Food Industry 10

Dispersion in TSC resulted in limited hydration and solubilisation of the rennet casein as shown by the prolonged time taken to exhibit an increase in viscosity index on shearing and the macroscopic appearance of the dispersed rennet casein. Although the calcium-sequestering ability of TSC is lower than that of MSP,5 the higher pH of the TSC-based rennet casein dispersion compared to that of the MSP-based dispersion facilitated greater hydration of the protein molecules in the TSC and some swelling and network formation occurred. The very limited imbibitation of liquid by the dispersed rennet casein in TSC suggests a substantial proportion of the protein-protein cross-links remained intact, thereby restricting the amount of swelling possible. In contrast, when DSP was used to disrupt the calcium-mediated cross-bridges under similar pH conditions, its greater calcium-sequestering ability resulted in more extensive disruption of protein-protein cross-linking, allowing more extensive hydration and swelling of rennet casein particles than was the case when the rennet casein was dispersed in TSC. Calcium-sequestering abilities of the pyrophosphates and the polyphosphates increase with the degree of phosphate polymerisation.5 On a molar basis the divalent anion of DSPP may sequester a similar amount of calcium as sequestered by DSP; TSPP may potentially sequester up to 2 times more calcium, while PSPP may sequester up to 2.5 times more calcium than DSP. At the rennet casein and salt concentrations used in the hydration system this could lead to complete disruption of the calcium-mediated crossbridges in the rennet casein dispersed in TSPP and in PSPP. The relatively low pH of the DSPP dispersion (pH 6.27) compared to that of the DSP dispersion (pH 7.20) may account for the low level of hydration of the protein after prolonged dispersion in the DSPP. Hydration behaviours and rheological profiles of the rennet casein in TSPP and in PSPP are consistent with such extensive disruption of the calcium-mediated cross-bridges that network formation is precluded (or at least greatly reduced), and solubilisation of the protein molecules results at pH 8.25 and pH 7.60, respectively. The different hydration characteristics of rennet casein when dispersed in the different salt solutions in the model hydration system in the present study appear to reflect the influence of the salts on the properties of processed cheeses prepared by Gupta et d 6using commercially available cheeses and a range of calcium-sequestering salts (added on an anhydrous weight rather than on a molar basis). They found that acidic salts, including MSP and DSPP, resulted in cheeses with pH values of 5.2 or below, and textures described as "mealy, dry and crumbly". They suggested these textural characteristics were due to poor emulsification of fat and lack of a continuous calciummediated cross-linked protein matrix in the cheeses. The cheeses did not melt on heating. DSP and TSC were reported to give processed cheeses that were "slightly firm and gellike", with the phosphate resulting in lower meltability of the cheese than the citrate. Increased internal structure of the protein matrix was suggested as a factor contributing to the lower meltability of the DSP-based cheese.6 Use of TSP as calcium-sequestering salt resulted in cheeses with high pH values, in the range 6.6 - 7.9 depending on the concentration of salt used. The cheeses were meltable and had "firm, gel-like'' textures. PSPP resulted in non-meltable cheeses with "slightly mealy, fibrous" or "slightly mealy, gel-like" textures depending on the concentration of salt used, while TSPP produced "firm, slightly gel-like'' cheeses that showed poor meltability. The poor meltability of cheeses prepared using PSPP and TSPP was attributed to the formation of a protein matrix sufficiently rigid (and presumably continuous) to prevent flow of the cheese on heating. It is possible that differences in the degree of fat emulsification in the processed

Proteins and Emulsions

381

cheeses may have contributed to the different melting properties reported, however, the degree of fat emulsification in the different cheeses produced were not determined.6 Cavalier-Salou and Cheftel’ studied the effect of different calcium-sequestering salts on the characteristics of cheese analogues prepared using calcium caseinates and butter oil. In general agreement with the results of the present study, they found that dissociation of calcium caseinate increased with the degree of polymerisation of the phosphate anion, while in contrast to the findings of the present study, they reported that TSC led to the highest degree of calcium caseinate dissociation. They reported that on changing the salt (at 1 % concentration) in the order TSC, DSP, TSPP, PSPP, the oil droplet size on emulsification decreased. Cheeses prepared using TSC or DSP had soft textures and good meltability while firmness increased and meltability decreased with degree of phosphate polymerisation. Perhaps low levels of protein solubilisation in DSP limited the surface area of the oil-water interface that could be stabilised and so large oil droplets formed, whereas high levels of protein solubilisation, in TSPP for example, facilitated stabilisation of a greater interfacial area and so the oil phase was dispersed as smaller oil droplets. Cheese analogues where the oil phase is poorly emulsified and thus present as large oil droplets tend to be softe? and tend to show greater meltability when heated” than cheeses where the oil phase is highly emulsified and thus present as small droplets. Savello et all3 found that rennet casein-based imitation cheese prepared using TSPP or DSP showed poor meltability compared to the equivalent cheese prepared using TSC. Photo-micrographs of the cheeses showed that there was greater emulsification of the fat in cheeses pr ared using TSPP or DSP when compared to cheeses made using TSC. Thomas et al. also reported that processed cheeses prepared using TSC were less firm than those prepared using PSPP. This is consistent with the findings in the present study that rennet casein hydrates more extensively in TSPP and DSP than in TSC, thus affecting the level of emulsification of the oil phase in the cheeses and consequently cheese firmness and meltability. On varying the cation of the calcium-sequestering salt from sodium (DSP) to potassium (DPP) the hydration behaviour of the rennet casein was comparable, suggesting that substitution of potassium for sodium would have little effect on the hnctional properties of cheeses prepared using these salts. This is in agreement with the findings of Gupta et aL6 who reported that substitution of tripotassium citrate (TPC) or DPP for the equivalent sodium salts (at equivalent concentration) in processed cheeses resulted in only slightly less melting of the potassium-containing cheeses on heating. Karahadian and L i n d ~ a y prepared ‘~ 55% and 75% reduced sodium American processed cheese by substituting blends of TPC and DPP for DSP in full sodium processed cheese and found that functional properties of the reduced sodium cheeses such as sliceability, castability, and texture were comparable to the full sodium cheese. The reasons for the large differences in the hydration behaviour of the rennet caseins in solutions of DAP compared to the hydration behaviour in DSP or DPP are not clear, but do not appear to be pH-related. The ammonium cation may itself exert some influence on the hydration behaviour of the rennet casein. Girdhar and Hansen” reported that a 16 % ammonium caseinate solution, prepared by converting acid casein using ammonia gas, exhibited much lower apparent viscosity than an equivalent sodium caseinate solution at 50°C or 25OC and concluded that the ammonium cation played some role in determining the rheological behaviour of the dispersed caseinate. Although Karahadian and LindsayI4 reported that acceptable reduced sodium American processed

T

382

Gums and Stabilisers for the Food Industry 10

cheese could be prepared by substituting ammonium phosphate for potassium phosphate as the emulsifying salt, results from the present study suggest that substitution of DAF’ for DSP in formulations for rennet casein-based Mozzarella cheese analogue would probably result in failure to produce a suitable cheese. 6 CONCLUSION

The hydration behaviour of rennet casein in calcium-sequestering salt solutions is highly dependent on the salt used. Limited disruption of calcium-mediated cross-bridges by weak calcium-sequestering salts allowed network formation to occur, resulting in high values of maximum viscosity index, while extensive disruption of calcium-mediated cross-bridges by more effective calcium-sequestering salts resulted in decreased network formation, protein solubilisation and low maximum viscosity index. Low pH led to poor hydration of the casein while high pH resulted in excessive solubilisation of the protein. The differences in casein hydration in the different salts can be related to differences in hnctionality exhibited by cheese analogues and processed cheeses prepared using the different calcium-sequestering salts. TSC and DSP both permitted network formation in the rennet casein dispersions under the conditions present in the model hydration system, and these salts generally show desirable performance characteristics when used in cheese analogue manufacture. TSPP and PSPP resulted in solubilisation of rennet casein in the model hydration system, these salts produce cheeses that are firm and exhibit poor meltability due to excessive solubilisation of the protein giving rise to overemulsification of the oil phase. Substitution of potassium for sodium in DSP resulted in only minor alterations to the hydration behaviour of rennet casein, suggesting that the potassium salt could be successhlly substituted for the equivalent sodium salt in established cheese analogue manufacturing processes with minimal modification to the processing conditions. Substitution of the ammonium cation for the sodium cation in the calcium-sequestering salt for use in cheese analogue manufacture does not appear to be feasible. The simple model hydration method could be used as an effective tool to investigate the effects of using different blends of calcium-sequestering salts on the hydration behaviour of rennet caseins in cheese analogue manufacture. Evaluation of potential substitutes for currently used calcium-chelating salts in formulations may also be easily performed. The method should prove usehl in evaluating potential substitutes for currently used calcium-chelating salts in formulations and for developing new formulations of salts to impart specific hnctional characteristics to cheese analogues and processed cheeses. 7 ACKNOWLEDGEMENT

M.P.E. is supported by grant aid under the Food Sub-programme of the Operational Programme for Industrial Development which is administered by the Irish Department of Agriculture, Food and Forestry, and supported by national and EU hnds.

Proteins and Emulsions

383

8 REFERENCES

1 . Meyer, A. (1973) In Processed Cheese Mmfacture. Food Trade Press Ltd., London

2. Shimp, L. A. (1985) "Process cheese. principles." Food Technology 5(5) 63-69. 3. Aimutis, W. R.(1995) "Dairy protein usage in processed and imitation cheese." Food Technologv Europe 2(2) 30-34.

4. Ennis, M. P. and Mulvihill, D. M. (1997) "Cheese analogues" In Cogan, T. M., Fox, P. F. and Ross, R.P. (eds) Proceedings of the Sth Cheese S'posium. Teagasc, Dublin, Ireland, ppl-14.

5. Caric, M. and Kalab, M. (1993) "Processed cheese products." In Fox, P. F. (ed) Cheese: Chemistry, Physics & Microbiology. (2"" eah) Chapman and Hall, London, ~~467-505. 6. Gupta, S. K., Karahadian, C. and Lindsay, R.C. (1984) "Effect of emulsifier salts on textural and flavor properties of processed cheeses." J. Dairy Sci. 67,764-778. 7. Cavalier-Salou, C. and Cheftel, J. C. (1991) "Emulsifying salts influence on characteristics of cheese analogs from calcium caseinate." J. Food Sci. 56(6) 15421547,1551, 8. Thomas, M. A,, Newell, G., Abad, G. A. and Turner, A. D. (1980) "Effect of emulsifying salts on objective and subjective properties of processed cheese." J. Food Sci. 45,458-459,466

9. Caric, M., Gantar, M. and Kalab, M. (1985) "Effects of emulsifying agents on the microstructure and other characteristics of process cheese- a review." Food Microstructure 4,297-3 12. 10. Neville, D. P. and Mulvihill, D. M. (1995) "Meltability of Mozzarella cheese analogues." Irish J Agric. Food Res. 34, 220. 1 1 . Ennis, M. P., O'Sullivan, M. M. & Mulvihill, D. M. (1998) T h e hydration behaviour of rennet caseins in calcium chelating salt solution as determined using a rheological approach." FoodHy&ocolloih 12,45 1-457.

12. Commission of the European Communities (1990) "Commission regulation (EEC) No. 2921/90 on aid for the production of casein and caseinates from skimmed milk." Oficial Journal of the European Communities. L279,22-27. 13. Savello, P. A., Emstrom, C. A. & Kalab, M. (1989) "Microstructure and meltability of model process cheese made with rennet and acid casein." J. Dairy Sci. 72, 1 - 1 1 . 14. Karahadian, C. and Lindsay, R. C. (1984) "Flavor and textural properties of reducedsodium process American cheeses." J. Dairy Sci. 67 1892-1904.

I S . Girdhar, B. K and Hansen, P. M. T. (1974) "Soluble casein by adsorption of ammonia." J. FoodSci. 39 1237-1243.

Recent Developments, Future ' h n d s

COMMERCIAL REQUIREMENTS AND INTERESTS: AN UPDATE

P. J. Lillford Unilever Research Colworth Colworth House Shambrook Bedford MK44 1LQ

1

INTRODUCTION

At the 6th International Conference (1991)’ an “End User’s View” of the state of the Food Hydrocolloids business was presented. One of the dangers of repeating the exercise 8 years later is that the predictions made earlier become testable facts so this paper compares our view then and the changes in the market place, technology and consumer attitudes which have taken place since.

2 SALES VOLUMES and USAGE of GUMS In the early 1990’s, we examined the states of the industry which provided purified hydrocolloid ingredients, mostly polysaccharide in origin but including large volumes of gelatin. The industry was focused in the developed world (Europe and the USA) and was recovering from a major attack from consumer groups pressing for the removal of “chemical additives” from formulated foods. On the other hand, the first signs of a new market for hydrocolloids as fat replacers was emerging. Understandably, prediction of market trends was hesitant, but on balance we suggested continued growth, at least as great as for the previous decade. So what actually happened? The following Tables (1 & 2) show the actual volume change. On average, growth occurred and at levels predicted. Starch remains the dominant hydrocolloid by a high proportion. No-one predicted the emergence of Bovine Spongiform Eucephalopathy (BSE) and the possible connection with Creuzfeld-Jakobs Disease (CJD). This held back growth of gelatin sales, but even this impacted largely on Europe only and the UK in particular where gelatin replacement became an issue primarily for finished food manufacturers. The consumer concerns were focused largely on red meat consumption rather than extracted gelatins. New stabilisers were approved for food use (e.g. Gellan) but they have developed sales very slowly. The star performer continued to be Xanthan for reasons that will be discussed later. Gums derived from the developing world continued to be used but the important issue of stability of production remains and therefore influences the cost price of ingredients. These trends are shown in the following Figures (la, b & c). It is becoming clear that large scale industrialised ingredient production results in stable price structures. The cyclic and unstable nature of

388

Gums and Stabilisersfor the Food Industry 10

developing world sources is inconvenient to end users and there may be opportunities in future to counteract these effects. Table 1

Marketsfor Food Hydrocolloids

Volume (tonnes x 103

1990

1997

1990

1997

Starches

600

1000

270

600

Others

120

320

420

1900

Total

720

1320

690

2700

Table 2a Markets for Food Hydrocolloids

1990

(Id tonnes)

1998

Gum arabic

30

Xanthan

16

Carrageenan

35

Pectin

25

Alginate

30

Guar

30

389

Recent Developments, Future Trends

Table 2b Markets for Food Hydrocolloids ($x lo6)

1990

1998

% Growth

250

650

1.2

Gelatin

350

-1.0

Gum arabic

150

1.5

Xanthan

160

7.8

270

3.2

Pectin

250

2.3

Alginate

90

2.4

Guar

80

2.5

Locust Bean Gum

100

2.1

34

6.8

Starches

Carrageenan 650

BUT Modified Cellulose

Figure l a Starches

I.'

1

1.2

-

Adapted from IMR International- Webslte hnp://www.netro~m.comllmr

1 -

0.6

-

e 0.6 -

0.4

7

' - - - - - - - f Z .O

0 91

92

93

94

95 YEAR

96

97

96

Gums and Smbilisersfor the Food Industry 10

390

Figure 1b Guar Gum

0 0 el

I

m

83

2

98

95

98

a7

2

YEAR

Figure l c Xanthan Gum

-

Adapted fmm IMR lntematbnal Website http:/Ewww.netmam.codmr

'1 4-

?!

H

3-

0 91

92

93

94

95

YEAR

96

97

98

Recent Developments, Future Trends

39 1

3 FAT REPLACEMENT

In the previous paper we suggested that “hydrocolloids appear to have a major potential”. We were certainly correct. The following, (Table 3) shows just some of the proprietary ingredients which have appeared, together with their botanical source of origin. Most of these have been backed by patents, so even if the ingredient suppliers have not made fortunes, their lawyers certainly have. We identified that ingredients of three types would emerge: Simple thickeners, acting as fillers. (a) (b) Phase separated mixed gels. (c) Microparticulate suspensions. Table 3 Fat Replacers from Polysaccharides

TRADE NAME

MARKETING COMPANY

SOURCE & STRUCTURE

N-Lite

National Starch

corn starch, maltodextrin, guar

N-Oil

National Starch

tapioca maltodextrin

Slender lean

National Starch

modified tapioca starch

Stay slim

Staley

modified potato + tapioca starch

Maltrin

Grain Processing Co.

corn starch, maltodextrin

Paselli-SA2

Avebe

potato starch maltodextrin

Oatrim

Rhone Poulenc f Quaker

oat flour enzyme hydrolysate

Stellar

Staley

acid modified cornstarch particulate

Slendid

Hercules

citrus pectin - particulate

Nutricol

FMC

Konjac + xanthan - gelled

Avicel

FMC

microcrystalline cellulose

Litesse

Pfizer

polydextrose

From FOOD ADDITIVES (1992) 70 (24) p.26-44

392

Gums and Stabilisersfor the Food Industry 10

All these are identifiable in Table 3, but note that most of these are derived from the primary hydrocolloids by mixing or some other physical treatment. Manufacturing industry has not been slow in recognising this. There are now many patented applications in products where the function of fat replacement has been achieved not by the addition of a prestructured ingredient, but by the modification of existing processing regimes to produce the a ro riate microstructure in situ, using careful control of existing ingredients. (e.g.

P,P,d‘

The continuing growth of Xanthan is certainly related to its almost unique property of exhibiting a reversible yield stress which allows particulate suspensions which are pourable at low shear rates. Stronger gelling agents do not exhibit this phenomenon but the recognition that this is a generalised property of microparticulate suspensions (of which Xanthan is only one) will yield competitors in the not too distant future.

4 NUTRITION AND HEALTH As well as the drive to reduce fat in the diet, we also identified some suggestions of a positive role for hydrocolloids in the positive reduction of cholesterol. Data from the previous paper is reproduced here (Table 4). We also suggested that this should and could be investigated further. The activity in this area has been surprisingly disappointing in the last decade. We still know little about the mechanism by which ‘soluble fibre’ regulates the flow of metabolites during the digestive process and almost nothing about the sequence of structural breakdown of foods and whether different hydrocolloids can be used to selective effect. It would be interesting to know why ingredient producers and government research bodies have not collaborated more extensively. Perhaps industrial research is underway but is yet to be publicised. In the meantime, other beneficial effects of traditional edible gums have been reported. For example, edible mastic gum, derived from the leaves and bark of Pistacia lentiscus has inhibitory effects on Helicobacter pylon, Salmonella enteritidis and Staph. aureus. The active agent appears to be phenolics present in small amounts. (6) It may well be that benefits to health associated with traditional materials could be lost during refinement if products are optimised only for their rheological function.

5 SUMMARY Returning to the trends of volume growth, it is remarkable that despite various market pressures and product opportunities, the growth of hydrocolloid ingredients roughly follows the growth of GDP in the developed world. In other words, as we become individually wealthier, our demand for fabricated foods and hence the ingredients to make them appear to be linked.

393

Recent Developmenis, Future Trends

Table 4

Hydrocolloid use and Function

TYPE

SOURCE

FUNCTI 0N

EFFECTS

Alginate (ionic)

Brown Seaweed

No known risk except in excess consumption as soluble fibre

Pectin

Plant Cell Walls

Stabiliser, Suspending Thickening Emulsification Gelation Emulsifier Gelling Agent

Agar

Red Seaweed

Locust Bean Gum

Carob Tree

Guar

Leguminous tree

Arabic (acacia)

Tree exudate

Xanthan

Fermentation

Gelation Thickening S tabilising Humectant Gelling Stabilising Emulsifying Thickening Thickener Emulsion stabiliser Suspending agent Thickener Stabiliser Emulsifier Stabiliser Thickener Emulsifier Pseudoplasticiser

NOrisk Reduction blood cholesterol Flatulence Temporary flatulence Soluble fibre No risk. Possible reduction of blood cholesterol. Soluble fibre. No risk Reduction of cholesterol Sugar control in diabetics No risk. slight blood cholesterol reduction. No risk Reduced serum cholesterol.

So much for the review of the past, what will be the key drivers for the future? We will now examine the possible impacts of demographic change, wealth creation, globalisation of trade and traders, and of course, new technological options

6 DEMOGRAPHICS AND WEALTH

If we scale the current population of the World down to 100 people, and distribute current wealth as 20 bags of silver dollars, we get the picture shown in Figure 2. We can see immediately that the per capita wealth is highest in the developed world, where the food ingredient market is significant. In Asia, average wealth is some 4 to 5 times lower, and very unevenly distributed, with Japan holding approximately 3 of the money bags. Since their diet is still different to the West and food import controls are significant, ingredient markets are relatively small.

394

Gums and Stabilisersfor the Food Industry 10

Figure 2 Demographics 199%

Now let us look 20 years ahead (Figure 3). We have increased the population proportionately but scaled the money to the same level, to represent competitive buying power. Because population growth outruns economic growth, the relative purchasing power becomes MORE polarised towards the Americas, with even Europe losing some ground. This gives the most pessimistic picture of market trends. In fact, provided GDP increases at a faster rate than inflation, then the potential trade volumes increase worldwide which means an overall potential growth for the business. Figure 3 Demographics 2020

Recent Developments, Future Trends

395

7 TRADEANDTRADERS We have seen continuous opening up of global trading in food (e.g. GATT round, investment in China and Eastern Europe) and this is likely to continue. Models for the economies of Asia and Central Europe are all approximations but the consensus is that free markets rather than state controls will dominate. This means that wealth will not be distributed evenly. We therefore face a market split not simply based on developed and developing world but also by the individual wealth of consumers. This is highly significant for the ingredients business. For example if 20% of China consumes fabricated foods, the market for ingredients could be equivalent to the current market in N. America - 600M tonne, potentially worth $300M. We are also seeing global consolidation of ingredient suppliers with fewer bigger companies becoming dominant. It is these that can afford the R&D to develop and introduce new ingredients and processes. The advantage to finished food manufacturers is that supply, price and quality become more reliable, increasing the likelihood of branded products developing worldwide.

8 FUTURE TECHNOLOGIES, RAW MATERIALS AND INGREDIENTS We have identified an enlarging but significantly split market. At the high added value end, we can expect continuing growth comparable to that seen over the last decade (Tables 1 & 2). The industries involved can afford and will identify sophisticated processes for the production of better ingredients, but two other factors will drive their success.

8.1.

Raw MaterialslBiodiversity

The agricultural products of the developing world are not the same as the developed world. Whilst starches derived from maize and wheat can expect to expand, so will those based on rice. Furthermore an abundance of other food gums and thickeners are used in food throughout the world and it will be necessary to improve production and to understand the potential of these in westernised and traditional foods. Availability of local supplies will be the driver of change. 8.2. Biotechnology 8.2.1 The “Developed” World. The potential for tailoring ingredients either by post extraction enzyme modification or genetic manipulation at the cellular level, is extensive with examples already described in the literature if not yet in the market place(7). The legislation requiring testing and proof of safety of new ingredients in these markets is itself an economic hurdle, estimated at between 20 and 25 Million dollars for each ingredient (@. But we can identify another duality in attitude of consumers of “convenience” food products. In both the USA and Europe, the requirements are for less chemically modified ingredients. However, in Europe, the resistance to the use of genetic engineering, whether in food enzymes or whole plants will slow commercial development for at least 5 years. In the Americas and Asia, the attitude is totally different where the technology is regarded as either advantageous to profitability (USA) or probably necessary simply to increase overall food yields (Asia). It will be very interesting to observe whether the use of biotechnology to create or extend the quality of finished products is accepted more readily than the current targets of yield and pest resistance.

396

Gums and Stabilisersfor the Food Industry 10

8.2.2. The “Developing” World. Our demographic analysis identified a wholly different market where quite different commercial conditions apply. This is the rapidly expanding 3rd World market where food must be produced but costs must remain low. There is a challenge to use local raw materials and low cost technologies in their conversion to fabricated foods. Here again, the understanding of component hnctionality in processing and products held by sophisticated industries could be directed towards providing these foods and ingredients. The potential for biotechnology “in planta”, or by fermentation or by process modification are all possible. Furthermore, in this situation, where food supply is limited, the legal constraints on novel ingredients and process are likely to be less severe than in developed countries. 9 PREDICTIONS In a further 8 to 10 years time the forecasts for future developments in hydrocolloids are as follows:1. The volume growth will match the growth of GDP per capita as economies develop. 2. Starches will remain the dominant raw material for hydrocolloids. 3. In the “Developed” world, physical modification of raw materials, and the use of enzymic extraction and modification will overtake chemical modification. 4. The use of GMO technologies will proceed in the US and the “Developing” world. Ingredients will be imported into Europe. 5. In the “Developing” world, local raw materials will dominate but will be industrialised into new ingredients to provide stable supply and costs and incorporated into new products, offering greater convenience and less preparation time to the consumer. References

1. Gums and Stabilisers for the Food Industry 6, Eds. Phillips, G.O. et al., 3-15 (1991). 2. Kraft Foods, US Patent, 459505685593 (1997) - Bakery Products. 3. J.M. Smucker Co., US Patent, US 970895152 (1998) -Bakery Products. 4. National Starch, US Patent, US 950394929 (1998) - Ice Cream, dressings, margarine.

5. SSW (Denmark), US Patent, US 5954028 (1998) - Meat products. 6. Koutsoumanis, K. et al., J. App. Microbiology, 84, (6) 981, (1998). 7. Edwards, K.J. et al., Microbiology, 145, 1499, (1999).

8. Ouellette, J., Chem. Market Reporter, June 16, (1997).

GENETIC ENGINEERING AS A MEANS TO MODIFY POLYSACCHARIDES.

G. Tucker Nutritional Bimhemisq, School of Biological Sciences, University of Nottingham, Sutton Bonington Camps, Loughborough, Leics. LEI2 5RD. UK

Summary Polysaccharides are traditionally modified post-extraction either chemically or by the application of enzymes. The ability to modify plants genetically has now provided an alternative, in vivo, approach to the production of designer polymers. There are perhaps two basic ways in which plants may be modified to alter the structure of their polysaccharides. The first would be the addition of novel enzymes to alter the biosynthetic or degradative pathways. The second would again achieve alterations to these pathways but would involve the silencing of genes for specific enzymes. The genetic tools for the manipulation of the biosynthetic pathways are limited at present. However, those for the degradative pathways are more readily obtainable. Genes may be silenced by the use of either antisense or co-suppression technology. Tomato h i t represent a good model system for testing this approach to polysaccharide modification. The tomato plant is readily transformable and the h i t cell wall pectin undergoes extensive modification during ripening. Tomato pectin has been modified in vivo by the silencing of genes for three pecteolytic enzymes namelypolygalacturonase, pectinesterase and galactosidase. These genetic modifications result in alterations to both pectin structure and its functionality in tomato products. The effects on pectin structure are fairly predictable, the effects on functionality less so in that it would appear that there is essential enzyme interaction responsible for the viscosity of tomato pastes. Introduction Polysaccharides have been manipulated chemically, or by the use of added enzymes, for many years in order to improve their functionality for various industrial processes (Tucker and Woods 1995). Genetic modification, especially of plants, presents the prospect to manipulate polysaccharide structure actually within the organism of interest. This modification could occur via the manipulation of enzymes involved in either the synthesis or degradation of the polysaccharides (Tucker and Mitchell 1993). This could involve the introduction of novel enzymes to the organism in order to redirect either the biosynthetic or degradative pathways. It could also result in the increased expression of enzymes already present thus enhancing a beneficial step in either synthesis or degradation. Alternatively, it is also possible to silence genes and thus down-regulate enzyme activities. Manipulation of synthesis is currently difficult given the relative scarcity of biochemical, and in particular genetic, information on the enzyme systems involved. Perhaps one major exception to this general statement is the modification of starch. Several enzymes involved in starch synthesis have been modified using gene silencing

398

Gums and Stabilisersfor the Food Industry I0

techniques and the structure of the starch produced in these transgenic plants is modified (Fulton et a1 1999). However, significant advances in the area of polysaccharide biosynthesis are being made and no doubt manipulation of other polysaccharides by the modification of genes encoding biosynthetic enzymes will occur in the near future. In contrast the genetic modification of polysaccharide degrading enzymes has been achieved and has already resulted in commercial products. The modification of pectin degradation in h i t from transgenic tomato plants resulted in a fresh h i t with increased shelf life and a processed paste with increased viscosity. These commercial applications were achieved by the silencing of the gene for a single pecteolytic enzymepolygalacturonase (PG). The tomato plant provides a very good model in which to study the effect of genetic modification on polysaccharide structure and function. The changes in the cell wall polymers accompanying ripening are well documented, as are the activities of several cell wall hydrolases. Genetic information is available for several of these hydrolases and the tomato plant is relatively easy to transform. For these reasons the tomato has been used at Nottingham University over several years as a model with which to investigate the effect of gene silencing on the ripening and processing of h i t .

Silencing of cell wall hydrolase genes There are basically two methods which have been employed to silence genes namely antisense and co-suppression. It is beyond the scope of this paper to describe these mechanisms in detail however an outline of the basic process is useful. A simplified outline of how antisense technology may function is given in figure 1. Further details can be obtained from Grierson et a1 (1996).

ANTISENSE

*

ANTISENSERNA

v

PROTEIN

Figure 1

NO PROTEIN

Outline of a possible mechanism for antisense gene silencing.

Recent Developments, Future Trends

399

In normal tissue the gene, under the control of the promoter (P) and terminator (T) sequences, is transcribed by DNA polymerase to generate mRNA which is subsequently translated into protein. In antisense silencing an artificial gene construct is prepared in which a copy of the coding region of the normal gene is taken and placed under the control of a second promoter such that transcription will now produce an antisense RNA. This is achieved by placing the coding region of the gene in the inverse orientation to normal with respect to the promoter. This antisense construct is then used to transform a tomato plant. In the transgenic tissue both the normal and the antisense genes are functional. However, the two resultant RNA molecules are complimentary and as such can form double stranded RNA-RNA hybrids. This effectively blocks translation of the mRNA and hence the target enzyme is down regulated. Co-suppression is achieved by a similar mechanism. In this case a pi& of the coding region of the target gene is taken and a gene construct prepared in which this fragment of DNA is placed under the control of a suitable promoter. However, in this instance the orientation of the coding strand with respect to the promoter is retained. When this construct is transformed into the plant both the transgene and the target gene are silenced hence the term co-suppression. This technique is sometimes referred to as sense silencing since in this case any RNA produced from the transgene would be similar to normal sense mRNA. The mechanism for silencing in this case is not clear. The actual process of gene silencing by either technique is not totally straightforward. Not all, in fact relatively few, of the transformed plants actually show very high levels of silencing. This is demonstrated in figure 2 for a range of 24 independent transformants designed to silence the PG gene. It is clear that many transformed plants, despite having the transgene in their genome, fail to demonstrate any appreciable down regulation of enzyme activity. The reason for this is unclear but may be related to the site of insertion of the transgene into the genome.

Figure 2.

Typical range of gene silencing observed in independent primary transformants carrying either an antisense or sense transgene.

400

Gums and Stabilisers for the Food Industry 10

On the other hand some transformants demonstrate very effective down regulation of enzyme activity. It is interesting to note that there is a gradual transition between these two extremes and plants can readily be obtained which exhibit intermediary levels of enzyme activity. It would be useful to establish in the case of a plant showing 50% normal PG activity whether this was the result of all the cells producing only 50% of the normal enzyme level or if 50% of the cells were acting normally and the other 50% were totally silenced. It is not easy to differentiate between these two possibilities using PG as the target gene since this does not produce a visible phenotype. However, silencing of another gene in the tomato, that for phytoene synthase, would result in a clear visible phenotype. Phytoene synthase is an enzyme involved in the biosynthesis of the red pigment lycopene, in its absence tomato h i t produce no red pigmentation and thus appear yellow when ripe. #en tomato plants were transformed with an antisense construct designed to silence the phytoene synthase gene some of the resultant plants produced ripe h i t which were completely red. This reflects those transformantsin which silencing was totally ineffective. Some plants produced fiuit which were entirely yellow and these represent plants in which silencing was 100% effective. These are the two extremes identified in figure 2. However, a large number of transformed fiuit appeared sectored i.e. they contained regions of both yellow and red tissue (Jones et al 1998). #en the two types of tissue were sampled separately it was clear that expression of the phytoene synthase gene had been completely prevented in cells from the yellow regions but proceeded as normal in those from red sectors (Jones et a1 1998). This demonstrated that antisense, and presumably co-suppression since similar results were obtained in this instance, is an all or nothing process. Either the target gene is completely silenced or it works normally These gene silencing techniques have been employed by several groups to down regulate enzymes involved in pectin degradation during tomato h i t ripening (Sheehy et a1 1988; Smith et a1 1988; Tieman et a1 1992; Hall et a1 1993). In particular the enzymes polygalacturonase (PG) pectinesterase (PE) and P-galactosidase (p-gal) have been targeted at Nottingham. The resultant levels of enzyme activity are shown in table 1 .

Normal PE(as) P-gal(as) PGPE(as) Table 1.

PG activity 0.55 0.04 0.67

PE activity 0.082 0.067 0.009

0.007

0.005

p-gal activity 0.46

0.22

Enzyme activities in normal and transgenic tomato h i t down regulated for Polygalacturonase [Paas)]; Pectinesterase pE(as)]; P-galactosidase [p-@(as)] or combined polygalacturonase and pectinesterase [PGPE(as)] activities. PG activity expressed as pmole/min/gfwt; PE activity expressed as meq/min/gfwt and p-gal activity expressed as pmole/min/mg.

Recent Developments, Future Trends

401

It can be seen that the efficiency of silencing in each case was different. Fruit containing an antisense PG construct PG(as) had less than 1% of normal PG activity. The PG activity in fruit occurs as several isoforms (Pressey and Avants 1973; Tucker et a1 1980). These have been characterised and shown to result from the post-translational modification of a single gene product (Tucker et a1 1980; Mohd Ali and Brady 1982;Sheehy et a1 1987; Bird et a1 1988). This single gene is relatively easy to silence using the antisense technology. In contrast PE was only reduced to about 10% of normal activity. The PE activity in tomato fruit is due to the presence of at least three isoforms (Pressey and Avants 1972; Tucker et a1 1982; Warrilow et a1 1994). These have again been characterised but in this case the isoforms are likely to be the products of three separate genes, or gene families (Warrilow et al 1994; Tucker and Zhang 1996). One isoform (termed PE2) is predominant in fruit and accounts for about 90% of the total PE activity. It was the gene for this isoform that was targeted by the antisense construct used in the PE(as) line whose activity is reported in tablel. The resultant transgenic had less than 1% of the PE2 isoform but normal levels of both the other two isoforms of PE thus accounting for the apparent relative inefficiency of silencing in this instance (Tucker and Zhang 1996). It is presumed that the genes for the other two isoforms are sufficiently different in sequence to prevent the antisense RNA from the transgene from interacting with them. This serves to illustrate the extreme selectivity of this silencing approach. In this instance not only an individual enzyme, but a specific isoform of that enzyme, had been selectively silenced. This could be a distinct advantage over say chemical inhibition of enzymes if it were found to be beneficial to either selectively remove or retain separate isoforms of a particular enzyme. The situation with P-gal appears even more complex. If @-gal activity is monitored in tomato fruit using an artificial substrate - p nitrophenol-P-galactoside - then there are three isoforms evident (Pressey 1983; Carey et a1 1995). However, only one of these isoforms when further characterised is capable of releasing galactose fiom a P (1-4) linked galactan as found in the tomato cell wall. This isoform is in fact an ex0 acting Pgalactanase (Carey et al 1995). It is possible that there are at least four and possibly as many as six genes encoding this P-galactanase activity in the tomato fruit. The results shown for the P-&as) fruit in table 1 arise following the silencing of only one of these genes. Hence the reduction of total enzyme activity in this instance was relatively much lower than for either PG or PE. The experiments reported so far have only targeted one gene at a time. In order to manipulate polymer structure a more versatile situation would be to be able to modify two or more enzyme activities simultaneously. This can, and has, been achieved by traditional breeding methods. If a plant which had been genetically modified with an antisense gene targeted at PG is crossed with one containing an antisense gene targeted at PE then, after carehl selection, a line could finally be obtained in which both enzymes were down regulated. However, this is both time consuming and labour intensive and if the aim were to silence four or even five genes may become prohibitively so. A simple alternative would be to employ a single transgene construct. This was first achieved for PG and PE using a chimeric construct (Seymour et a1 1993). In this case the first 244bp from the 5’ end of the PG cDNA had been ligated to 1320bp from the 3’ end of the

402

Gums and Stabilisers for the Food Industry 10

cDNA for PE2. This double sense construct was then transformed into tomato plants. The resultant h i t exhibited a reduction in both PG and PE activities (Table 1) similar to that obtained for each enzyme separately in either single sense or antisense lines. The nature of the chimeric transgene can be very flexible and it has been shown that this works if both cDNA fragments are in either the antisense or sense orientations, or indeed if one is sense and the other antisense (Jones et a1 1988). Thus it appears that it is the presence of the DNA sequence homologous to the target gene that is paramount rather than the relative orientation or placing of this sequence within the transgene. From table 2 it would appear that the down regulation of PE in those lines containing the chimeric PGRE construct may be greater than that which occurs in the line containing the single PE2 antisense gene. This has been investigated firther and it has been found that in the case of the PGRE transgene two out of the three PE isoforms (PE 2 and PE3) appear to have been silenced (Simons and Tucker 1999). Thus in this line only the PEl isoform was present whereas in the single PE2 line both PE1 and PE3 isoforms were evident. The reason for this difference is unclear but may arise from the fact that a larger segment of the PE2 cDNA was employed in the chimeric construct and that the additional sequence may have sufficient homology to the PE3 gene to result in its silencing. Effect on pectin structure.

Green FZ

294

Table 2.

111

Changes in tomato h i t cell wall pectic polymers during normal ripening. *Values taken from Gross and Wallner (1979).

The down regulation of these pecteolytic enzymes has, in the main, resulted in fairly predictable changes to the pectin metabolism in the tomato h i t . During ripening of normal h i t the cell wall undergoes extensive degradation. The major changes appear to occur within the pectin rich middle lamella region of the wall. These have been well documented and are summarised in several reviews (Tucker and Grierson 1987; Fischer and Bennett 1991; Tucker 1993) and in table 2. The pectin becomes increasingly soluble and this soluble fraction becomes depolymerised. The degree of esterification of the polyuronide declines with ripening and there is a gradual loss of neutral sugars, in particular galactose. The effects of down regulating either PG or PE activity on this pectin metabolism have also been covered in several previous papers (Smith et a1 1990; Tucker 1990; Schuch et a1 1991; Tucker et a1 1999) and are summarised in table 3. It can be seen that a reduction in PG has, as expected, no effect on the degree of esterification of the pectin. It has resulted in a reduction in the extent of depolymerisation of the pectin. This would be

Receni Developments, Future Trends

403

consistent with the mechanism of PG as an endo-acting enzyme which cleaves the polyuronide backbone of the pectin. Interestingly there seemed to be little if any effect on the solubility of the pectin. It had been widely assumed that the action of PG in breaking the pectin backbone might be very significant in causing the increase in solubility associated with normal ripening. This may not be the case. It should be pointed out, however, that solubility in this instance was measured by extraction into a strongly chelating buffer. There are reports that the reduction in PG activity does result in a decline in the water solubility of the pectin (Carrington 1993). The reduction of PE activity has as predicted, resulted in a reduction in the extent of deesterification occurring during ripening. This reduction in deesterification, or looked at another way, this increase in the level of esterification of the polyuronide was evident at all stages of fruit development and ripening as shown in figure 3, the polyuronide in the PE(as) lines being around 10-20% more esterified than normal at all stages of ripening.

1

I

Solubility(%)

Esterification (%)

59

Normal Ripe P a a s ) Ripe PE(as) Ripe Table 3.

I

56 53 66

52

67

1

AverageMr (kD) 111

I

248

85

Effect of down regulation of either polygalacturonase [pG(as)] or pectinesterase [pE(as)] activities on pectin metabolism during tomato fruit ripening.

-c loo C

0

P

80

E

60

3

c 0

t

40 *O

B o

MG

B

B+7

B+14

Stage of Ripening

Figure3.

Changes in the degree of esterification of total polyuronide during the ripening of normal [O] and PE(as) [D ] tomato fruit.

404

Gums and Stabilisers for the Food Industry 10

It is interesting to note, however, that the decline in esterification during ripening still proceeds in these transgenic fruit, and this presumably is the result of the activities of either PE1 or PE3. The reduction of PE activity has not effected either solubility or depolymerisation. Again the latter may be surprising since it was again thought that the action of PE was synergistic with PG and generated sites of action for the later. However, again this role may be taken by the remaining PE1 and PE3 activities. The effect on pectin metabolism of the simultaneous down regulation of PG and PE is still under investigation. The effect of down regulating p-gal activity is also under investigation. However, in preliminary studies no effect on the loss of galactose from the cell wall has been demonstrated so far. Effects on fruit quality and processing The down regulation of these enzymes has had relatively little effect on the softening of the fruit. Reductions in PG activity did not prevent the fruit from softening to what could be considered optimal eating quality (Schuch et a1 1991; Langley et a1 1994; Errington et a1 1997). However, the subsequent over-softening of the fruit did appear to be significantly delayed resulting in an extended shelf life. In addition the resistance of ripe h i t to cracking appeared to be greatly enhanced. This is demonstrated in figure 4.

7

14

21

Days post Breaker

Figure 4.

Effect of down regulating polygalacturonase activity on fruit cracking during transportation.

Tomatoes were grown in Littlehampton on the south coast of Britain. These were then harvested at various stages of ripening, packaged and sent by rail freight to Nottingham, a journey of about 200 miles. On arrival fruit were scored for cracking. It can be seen that normal fruit harvested ripe had a very high level of cracking often reaching 50%. In contrast PG(as) h i t were much more resistant to cracking. In the US this characteristic of the PG(as) plant has been exploited in the marketing of the “Flavr

Recent Developments, Future Trends

405

Savr” variety of fresh tomato fruit. These can be harvested ripe from the vine, as opposed to green as is the normal practice. This resulted in enhanced flavour and hence improved quality. A large proportion of the tomato crop worldwide is actually processed into products such as paste and ketchup. In this instance two key quality aspects are the level of solids in the fruit and the viscosity of the resultant paste. The viscosity of tomato paste is improved if fruit are exposed to a so-called “hot-break” process during processing. The high temperaturetreatment is thought to inactivate endogenous enzymes, especially those responsible for pectin degradation, and hence result in a higher viscosity product due to the retention of polymers with greater molecular weights. It was thus likely that paste viscosity could be markedly effected in those transgenic fruit with reduced pecteolytic activities. To test this hypothesis pastes were prepared from normal and transgenic fruit without any heat treatment. The paste was passed through a sieve to remove the seeds and particles of skin and then the viscosity measured using a standard Bostwick tray (Emngton et a1 1998). This method measures the flow of a fixed volume of paste over a period of 30 seconds.

Table 4.

Physical properties of pastes as prepared using a “cold-break” process from normal or transgenic tomato fruit.

The Bostwick values for pastes from normal and transgenic fruit are shown in table 4. It can bee seen that the Bostwick value for the PG(as) paste was almost a third that of normal indicating that these fruit produce pastes with significantly improved viscosity. Fruit in which PE had been down regulated showed no significant differences in paste viscosity, interestingly fruit in which both PG and PE activities had been reduced also showed no significant improvement in Viscosity. There was an improvement in viscosity shown for pastes produced from the P-gal(as) fruit, this despite very little apparent effect on either enzyme activity or pectin metabolism in these fruit. The availability of these transgenic lines has allowed a study of pectin structurehnction relations to be carried out with respect to paste viscosity. It was possible that the paste viscosity was dependent upon the viscosity of the serum. To test this hypothesis pastes were centrifbged and the viscosity of the resultant sera determined using a standard Ostwald viscometer (Emngton et a1 1998). The sera viscosities are shown in table 4 alongside the paste viscosities and it is apparent from these two sets of values that no correlation exists. Indeed the serum from the PE(as) fruit had the highest viscosity whilst that from the PG(as) k i t was not significantly different to the control. The

406

Gums and Stabilisers for the Food Industry 10

composition of the sera in each case is shown in table 5 . It appears that whilst sera viscosity may indeed be determined by the extent of solubilisation and the nature of the pectin polymers this has little relevance for the viscosity of the paste as a whole.

Table 5.

Chemical properties of sera from pastes prepared using a “cold-break” process from normal or transgenic tomato h i t .

Following the centrifugation the solid volume fraction was recorded. Again this is shown in table 4 alongside the respective paste viscosities. In this instance a slight correlation could be seen in that the PG(as) paste had both the highest viscosity and the greatest volume fraction. It is possible therefore that it is the solid, or colloidal, fraction of the paste that has the greatest determinant effect on viscosity. The nature of the solids was hrther examined using confocal microscopy. It was apparent that in pastes from normal h i t the cells had been almost totally disrupted and the solid fraction was composed almost entirely of cell wail fragments. In contrast pastes from PG(as) hit contained many more “intact” cell structures. It is possible that these more “intact” cells are capable of greater interactions within the paste. However, retention of cell integrity cannot be the sole consideration since confocal images of the paste from PE(as) h i t also showed “intact” cells yet in this case viscosity was not improved.

1 : : ? Q, i

b loo-

/

5om 0

1

2

3

4

Time (min)

Figure 5 .

The change of viscosity with time of “cold-break” pastes prepared from normal [D], PG(as) [A]or PE(as) [O] tomato fruit.

Recent Developments, Future Trends

407

It has been pointed out earlier that from the data in table 4 it was apparent that whilst down regulation of PG activity resulted in improved paste viscosity the coupling of this with a silencing of the PE gene (PGPE(as)) actually appeared to negate any beneficial effect. This led to the suggestion that whilst the absence of PG activity was beneficial to achieve optimal paste viscosity this may also require the presence of PE activity. This could be either via an influence on pectin structure within the fruit or from activity of the enzyme in the paste following homogenisation. To examine the effect of changes occurring following homogenisation the storage modulus (G') of the paste was monitored, using a frequency of SHz, with time using a Bohlin Rheometer (Errington et a1 1998). The results are shown in figure 5 . It can be seen that the storage modulus of normal pastes, or those produced from PE(as) fruit, did not change with time. However, that for pastes from PG(as) fruit increased dramatically. This finding may suggest that for optimal paste viscosity cell walls need to remain intact, this can be achieved by a reduction in several pecteolytic enzymes such as PG or PE. However, if PE can continue to act post homogenisation then this could result in negative charges being generated in the cell wall. This negative charge on the wall may enhance cell to cell adhesion via calcium bridging or other mechanisms thus increasing the size of the colloidal particles in the paste and hence its viscosity.

Conclusion The application of genetic engineering has huge potential for the provision of tailor made polymers for industry in general. Not just polysaccharides but lipids and proteins may also be modified using this technology. The changes that can be made can be very precisely tuned and controlled and it may even be possible to generate polymers which are not found naturally but which may have enormous benefit to industry. The major bottleneck to the application of this technology is the relative scarcity of information on the biochemical and genetic basis of polymer biosynthesis and degradation. Another major problem area, at least in Western Europe at the moment, is public acceptance of this technology. The application of this technology is likely to proceed rapidly in other arms of the world and polymers produced via the application of gene technology are likely to become more prevalent in the near future.

References Bird, C.R, Smith, C.J.S., Ray, J.A., Moreau, P., Bevan, M.W., Bud, A.S., Hughes, S., Moms, P.C., Grierson, D. & Schuch, W. (1988) The tomato polygalacturonase gene and ripening specific expression in transgenic plants. Plant Molecular. Biology. 11 65 1-662. Carey,A., Holt,K.,Picard,S., Wilde,R., Tucker,G.A., Bud,C.R, Schuch.W and Seymour,G.B (1995) Tomato exo-( 1-4)-B-D-galactanase: isolation and changes during ripening in normal and mutant tomato fruit and characterisation of a related cDNA clone. Plant Physiology. 108, 1099-1107. Carrington,C.M.S., Greve, L.C and Labavitch, J.M (1993) Cell wall metabolism in ripening fruit. 6. Effect of the antisense polygalacturonase gene on cell wall changes accompanyingripening in tomato fruit. Plant Physiology. 103.429-434.

408

Gums and Stabilisers for the Food Industry I0

Emngton, N., Mitchell, J.R and Tucker, G.A (1997)Textural analysis of tomato fruit:- a comparison of stress relaxation with flat-plate compression tests. Postlmvest Physiology and Biotechnology. 11. 141-147.

Enington. N, .Tucker. G and Mitchell. J (1998) Effect of genetic Down-regulation of Polygalacturonase and Pectinesterase Activity on Rheology and Composition of Tomato Juice. Journal of the Science of Food and Agriculture. 76.5 15-519. Fischer, R.L. & Bennett, A.B. (1991) Role of cell wall hydrolases in fruit ripening. Annual. Review of Plant Physiology and Plant Molecular Biology. 42 675-703. Fulton,E.A.,Hyhon,C.M., Jobling, S.A., Gidley, M., Rossner,U., Martin$ and Smith, A.M (1999).A combined reduction in starch synthases ii and iii of potato has novel effects on the starch of tubers. Plant Journal. 17.251-261. Grierson,D., Lycett,G.W. and Tucker,G.A (Eds) (1996)Mechanisms and applications of gene silencing.Nottingham University Press. Gross, K.C and Wallner,S.J (1979).Degradation of cell wall polysaccharides during tomato fruit ripening. Plant Physiology. 63.117-120.

Hall, L. H., Tucker, G. A., Smith, C. J., Watson, C. F., Seymour, G. B., Bundick, Y., Boniwell, J. M., Fletcher, J. D., Ray, J. A., Schuch, W., Bud, C. R and Grierson,D (1993) Antisense inhibition of pectin esterase gene expression in transgenic tomatoes. The Plant Journal. 3.121-129. Jones, C.G, Scothern, G.P, Lycett,G.W and Tucker, G.A (1998)The effect of chimeric architecture on co-ordinated gene silencing. Planta 204,499-505. Langley, K.R., Martin, A., Stenning, R., Murray, A.J, Hobson, G.E., Schuch, W and Bird, C (1994)Mechanical and optical assessment of the ripening of tomato h i t with reduced polygalacturonase activity. Journal of thescience of Food and Agriculture. 66.547-554. Mohd Ali,Z. and Brady,C.J.(1982) Purification and characterisation of the polygalacturonase of tomato h i t . Australian Journal of Plant Physiology. 9,155-159. Pressey, R. (1983)bgalactosidases in ripening tomatoes. Plant Physiology. 71 132-135. Pressey,R and Avants,J.K (1972) Multiple forms of pectinesterases in tomatoes. Phytochemistry. 11.3139-3142. Pressey, R. & Avants, J.K. (1973) Two forms of polygalacturonase in tomatoes. Biochimica et Biophysica Acta. 309 363-369.

Recent Developments, Future Trends

409

Schuc~W.,Kanczler,J.,Rob~son,D.,Hobso~G.E.,Tucker,G.A., Grierson,D.,Bright,S and Bird,C (1991) Fruit quality characteristics of transgenic tomato h i t with altered polygalacturonaseactivity. Horticultural science. 26. 1517-1520. Seymour, G. B., Fray, R. G., Hill, P and Tucker, G. A (1993) Down regulation of two non-homologous endogenous tomato genes with a single chimearic gene construct. Plant Molecular Biology. 23. 1-9.

Sheehy,R.E.,Pearson,J.,Brady,C.J. and Hiatt,W.R.(1987) Molecular characterisation of tomato h i t polygalacturonase.Molecular and General Genetics. 208.30-36. Sheehy,R.E.,Kramer,M. and Hiatt,W.R. (1988) Redudion of polygalacturonase activity in tomato h i t by antisense RNA. Proceedings of the National Academy of Sciences,USA. 85. 8805-8809. Simons, H and Tucker G.A (1999) Simultanmus co-suppression of polygalacturonase and pectinesterase in tomato h i t : Inheritance and effect on isoform profiles. Phytochemistry (in press). Smith, C.J.S., Watson, C.F., Ray, J., Bud, C.J., M d s , P.C., Schuch, W. & Grimn, D. (1988) Antisense RNA inhibition of polygalacturonase gene expression in transgenic tomatoes. Nature 334 724-726. Smlth, C.J.S., Watson, C.F., Moms, P.C., Bird, C.R., Seymour, G.B., Gray, J.E., Arnold, C., Tucker, G.A., Schuch, W., Harding, S.E. & Grierson, D. (1990) Inheritance and effects on ripening of antisence polygalacturonase genes in transgenic tomatoes. Plant Molecular. Biology. 14 369-379. Tieman, D.M, Harrirnan,R W, Ramamohan,G and Handa,AK (1992) An antisense pectin methylesterase gene alters pectin chemistry and soluble solids in tomato h i t . The Plant Cell. 4. 667-679. Tucker,G.A. (1990) Genetic manipulation of h i t ripening. Biotechology and Genetic Engineering Reviews. 8, 133- 159. Tucker,G.A (1993) Fruit ripening. in The biochemistry of h i t ripening. Seymour,G.B., Taylor,J.E and Tucker,G.AEds Chapman and Hall.pp 1-51, Tucker G.A.,RobertsonN.G. and Grierson D. (1980) Changes in polygalacturonase isoenzymes during the ripening of normal and mutant tomato h i t . European Journal of Biochemistry, 112. 119-124. Tucker G.A., Robertson N.G. and Grimson D (1982) Purificationand changes in activities of tomato pectin esterase isoenzymes. Journal of the Science of Food and &culture. 33.396-400.

410

Gums and Stabilisers for the Food Industry 10

Tucker,G.A and Grierson,D (1987) Biochemistry of h i t ripening. in The biochemistry of plants:A comprehensive treatise. Volume 12 Davies,D.D(Ed) Academic Press. Tucker,G.Aand MitchelLJ (1993) Plant cell walls structure. functionand utilisation. in The manipulation of plant produ&. Volume 3. GriersoqD Ed. Blackie and sons.pp55103. Tucker,G.A and Zhang,J (1996) Expression of polygalacturonase in normal and transgenic tomatoes. in Pectins and Pectinases. Elsevier. Ed Visser.J and Voragen A.G.J Elsevier. pp 347-354. Tucker,G.A and Woods,L (Eds) (1995) Enzymes in Food Processing 2nd Editionchapman and Hall. Tucker, G., Simons. H and Errington. N (1999) Enzymic modification of pectin in pastes from transgenic tomatoes. Biotechnology and Genetic Engineering reviews. (in press). Wadow, A. G. S.,Turner, R. J and Jones (1994) A novel form of pectinesterasein tomato. Phytochemistry. 35.863-868.

SUBSTITUTION OF GELATINE IN LOW-FAT SPREAD: A RHEOLOGICAL CHARACTERISATION.

Finn Madsen

Danisco Ingredients NS Edwin Rahrsvej 38 DK-8220 Brabrand

1 ABSTRACT

In order to investigatethe possibilities for substitutinggelatine in low-fat spread, the gelling and melting properties of gelling hydrocolloidswere investigated using dynamic rheological analysis. Two hydrocolloids, low-ester amidated pectin and alginate, were tested in a 40% low-fat spread containing skimmed milk, with gelatine serving as the control. The gelling and melting behaviour of the hydrocolloidsin the water phase of the low-fat spread were related to the final quality of the low-fat spreads produced. It was found that the hydrocolloids’ gelling profile influenced the stability, appearance, texture, droplet size distribution and, thus, microbiological stability of the spread, while their melting profile had an effect on meltdown properties in the mouth and flavour release. On completing the study, it was concluded that premium low-fat spread products of a similar quality to those containing gelatine could be produced if the g e m temperature was lower than 30°C and if considerable gel softening was obtained at 37°C or less. The task of obtaining the right type of emulsion during low-fat spread production could be made easier than when gelatine is used providing the viscosity of the water phase was somewhat higher at processingtemperaturesof 10-40°C. Thus the rheological investigations have opened up the possibility of optimising the rheological behaviour of vegetable hydrocolloidsto match or even surpass the hctionality of gelatine in low-fat spread. An interaction between milk protein and hydrocolloids in the low-fat spread’s water phase cannot be ruled out and, due to the pH level and concentration of these components, incompatibility and phase separation may be expected. This needs fbrther investigation. 2 INTRODUCTION

The prevalence of excessive calories and fat in Western foods has created a pronounced interest in low-fat spreads. In Europe, the first low-fat spreads were marketed in the UK in 1968. Since then, the consumption of low-fat spread has grown to such an extent that it is now produced in most European countries, South America and North America’. Low-fat spreads should have Similar eating properties to fill-fat margarine with regard to flavour release and mouthfeel. To achieve this, protein is usually added to the water

412

Gums and Stabilisers for the Food Industry 10

phase, providing a looser emulsion with improved taste and flavour release. However, as protein favours an oil-in-water emulsion, its addition can cause instability in low-fat spread. Traditionally, low-fat spreads containing protein have been stabilised with the aid of gelatine or a high level of sodium caseinate. Due to the origin of gelatine and the cost of using sodium caseinate, alternative stabiliser systems based on vegetable raw materials are becoming increasingly popular '. The functional properties of hydrocolloids in protein-containing low-fat spreads have been investigated extensively. The functional role of the hydrocolloids is to bind water and increase the viscosity of the water phase to prevent droplet aggregation and make the product process and shelf-stable, and to provide emulsions that break down easily and give good flavour release in the mouth ', '. It is believed that the rheology of the water phase during processing and in the final product is crucial to the quality of low-fat spread4. Thermodynamic incompatibilit between hydrocolloids or hydrocolloids and protein may also play an important role 5 , 6 , 7.

',

Numerous patents cover the use of gelling hydrocolloids in low-fat spreads 9* lo, 11, although these do not generally include any in-depth rheological characterisation of the water phase. In our investigations regarding the substitution of gelatine in low-fat spread with gelling vegetable hydrocolloids, we have focused on a detailed rheological characterisation of the low-fat spread water phase during cooling (processing temperatures) and heating (organoleptic evaluation temperatures). 3 MATERIALS AND METHODS

To investigate the importance of a specific gelling and melting profile in the water phase of low-fat spread, the following five hydrocolloids were selected: Hydrocolloid 1: A low ester amidated pectin with high gelling temperature (GRINDSTEDm Pectin LA 410) 0 Hydrocolloid 2: A sodium alginate (GRINDSTEDm Alginate FD 152) Hydrocolloid 3: A blend of hydrocolloid 1 and 2 Hydrocolloid 4: A low ester amidated pectin with low gelling temperature (GRINDSTEDm Pectin LFS 120) 0 Hydrocolloid 5: A high bloom gelatine control (GRINDSTEDm Gelatine HB 240) (All the hydrocolloids tested are produced by Danisco Ingredients NS). The five hydrocolloids were tested at a dosage of 2% in 40% fat spread with 1.0% skimmed milk. They were also analysed for their gelling and melting behaviour in the water phase of the 40% fat spread. This was carried out by adding 3.3% hydrocolloid to the water phase, corresponding to 2% hydrocolloid in total spread. The quality of the low-fat spreads produced was evaluated and the quality data compared with the rheological data of the water phase.

Recent Developments, Future Trends

413

3.1 Production of low-fat spread 3.1.1 Recipe. 55.4% Water phase: Water 1 .O% Skimmed milk powder 1.5% salt 2.0% Hydrocolloid 0.1% Potassium sorbate Flavouring 9.9% Fat phase: Soya 41°C 29.6% Soya oil DIMODAN@OT Distilled Monoglyceride 0.5% b-carotene 4 PPm Flavouring 3.1.2 Procedure. The ingredients were blended into the water phase at 70°C.The water phase was subsequently cooled to 40°C and the pH adjusted to 5.5. The fat phase ingredients were blended at 40°C and poured into the emulsification tank.The water phase was then slowly added to the fat phase, while emulsiflmg. The emulsion (W/O) was processed in a Gerstenberg & Agger 2-tube lab perfector (ammonia -15"C,throughput 20 kgh, rotor speed 950rpm, outlet temperature of the spread approx. 10°C). 3.2 Evaluation of low-fat spread

3.2.1 Texture profile analysis. The low-fat spread was filled into glass dishes (6Omm in diameter and 35mm high) avoiding air incorporation. Surplus spread was removed with the back of a knife to ensure a level surface at the dish rim.The dish was then covered with plastic film and stored at 5°C for at least two days before texture measurement at the same temperature. A TA-XT2 Texture Analyser from Stable Micro Systems was equipped with a %,, probe (A05) and texture profile analysis was conducted with a penetration depth of 43%, correspondingto l5mm penetration. The penetration speed was OSmm/sec. 3.2.2 Spreading stability. Stability when spread with a knife on cardboard was evaluated, using a stability index from 0 to 10: 10 - Highly stable, very smooth 98 - Stable, but separates when worked intensively 76 - Separates when worked 54 - Separates when worked a little 32 - Separates 10 - Separates in the chiller or is an O/W product 3.2.3 Organoleptic evaluation. A trained panel evaluated the spread's appearance, meltdown properties, flavour release and taste using descriptiveterms. 3.2.4 N R droplet size analysis. The droplet size analysis was performed on a Bruker NMSlOO Minispec NMR analyser, equipped with a Pulsed Gradient Unit GU200. A

414

Gums and Stabilisersfor the Food Industry 10

modified Hahn spin-echo experiment enables the size of the droplets to be determined, assuming round droplets and log-normal size distribution. The low-fat spread was placed in a glass tube and tempered to 5°C prior to measurement. The mean droplet size (volume distribution), standard deviation and 95% confidence interval were calculated using Bruker software. 3.2.5 Microscopy. Freeze microton sections (19 pm) of the low-fat spreads were examined on a Olympus BX60 light microscope (200 x enlargement). A Peltier element secured a spread temperature of 5°C during the examination. 3.3 Dynamic rheological analysis of the low-fat spread's water phase

The measurements were performed on a Bohlin VOR Rheometer, a controlled strain instrument, with a constant strain level of 0.004 and frequency of 1 Hz,which is in the linear viscoelastic region. The low-fat spread's water phase was added to the C25 measuring system at 70°C and covered with silicone oil to prevent evaporation. The complex modulus G* and phase angle 6 were measured during the following temperature sweep: 0 Cooling from 70°C to 10°C at a cooling rate of l"C/min. 0 Heating fiom 10°C to 70°C at a heating rate of l"C/min. 3.4 RESULTS AND DISCUSSION

To distinguish the gelling profile curves, the following characteristic points on the curves were defined: Gel onset (645") 0 Strong gel build-up (6=20") To distinguish the melting profile curves, the following characteristic point was defined: 0 Gel softening (reduction of G* to 20% of the value at 10°C) 3.4.1 Inferior gelling and melting profde

The gelling and melting profile of hydrocolloid 1 is shown in figure 1. The water phase has a gel onset of around 54OC and a strong gel build-up at 37°C. Phase angle, 6.degree I

H m n a r . G*. Pa

Flnnnar, G*. Pa

0410

20

30

40

50

60

70

10

p h W w * 6 , d w =

, '0

1

20

30

40

50

60

70

415

Recent Developments, Future Trends

The quality of the low-fat spread obtained with hydrocolloid 1 is shown in table 1. The low-fat spread was very soft, pastalike, grainy and unstable when spread. The meltdown properties and flavour release were characterised as average. The handliig of the water phase during production was di5cult due to the gelling of the water phase before the emulsification step. The emulsion was grainy in appearance. 3.4.2 Optimal gelling and inferior melting profde

The water phase with hydrocolloid 2 had a gel onset of 25OC and did not remelt or soften upon heating (figure 2). The resulting low-fat spread was smooth in appearance, firmer and more brittle in texture, and the spread was very stable when spread. Organoleptically, the low-fat spread was evaluated as having slow meltdown and a rather sticky mouthfeel (see table 1).

200

I

04 10

-8

--- ff

20

30

40

50

60

200

I

, 10

70

10

Gdonta Tampmarre Y

20

30

40

50

60

70

Tcmpencure O C

E'igure 2 Gelling and meltingprofile, hyd-ocolloid2 3.4.3 Optimal gelling and melting profde

Hydrocolloids 3, 4 and 5 were all found to have relatively low gelling temperatures below 30°C.Furthermore, a melting or gel softening point was achieved at 37OC or lower while heating the water phase. The gelling and melting profiles of hydrocolloids 3 , 4 and 5 are shown in figures 3,4 and 5, respectively.

Gums and Stabilisersfor the Food Industry I0

416 Flnnnesr, G*. Pa

h

e mgle. 6.degree

250 1

Firmness, G*. R

Phase d c .

6.degree

250 1

I

I

, 10

04 10

20

Gel wet Tempentun "C

30 40 50 Gel raitenlng

60

70

Tempenturn "C

Figure 3 Gelling and melting profile, ly&ocolloid 3

"i/Fimnar. G*, Pa

250 1 1-6

Phase angle, 6,degree

Firmness. G*, Pa

h

e mgk. 6.desm

250 1

80

20060

100

I

-6 G*

--

-80

-60

150 -

-40

- 20 0

0-f 10

-r--------------___ 20

30 40 Gel softening

50

60

r

o

70

Twnpcrrarrc "C

Figure 4 Gelling and melting profile, ly&ocolloid 4 Hrmnar.G*, Pa

h

250

e angle, 6. degree

-6

b

e mgk, 6,degree

80

200

200

60

I50

I50

100

100

20

50

a

Flrmnar. G*, Pa

50

II

10 SPMlg-

0 20

cud.uponra

30

40

50

60

70

0

10

Tempemtun O C

Figure 5 Gelling and melting profik, ly&ocolloid 5

20

30 40 Gel sdtenlng

50

Tampmrurc O C

60

70

Recent Developments, Future Trends

417

Hydrocolloid 3 developed a gel at 22°C and had strong gel build-up at 10°C. When heating the water phase, gel softening was achieved at 37°C. Hydrocolloid 4 gelled at an even lower temperature - onset at 16OC - and had a strong gel build-up at 10°C or lower. The relatively steep reduction in phase angle indicates rapid gelliig, and the low value of G* in the temperature range 40°C to 16°C shows that the water phase had low viscosity at processing temperatures, compared with hydrocolloid 1, 2 and 3. When the water phase containing hydrocolloid 4 was heated, a gel softening temperature of 33°C was measured. Hydrocolloid 5 (gelatine) showed sharp gel onset at 13°C and strong gel build-up at 12°C. Gelling occurred very quickly (a very steep reduction in phase angle). The very low G* values above the gel onset temperature indicate a low viscosity water phase during a considerable part of the production process. When heated, hydrocolloid 5 showed a gel softening point of 29°C immediately followed by a complete melting of the water phase. The low-fat spreads produced with hydrocolloid 3,4 and 5 were all characterised as having a smooth appearance and a firm, brittle texture. Fast meltdown properties were accompanied by rapid flavour release. This was especially pronounced for hydrocolloid 4 and 5. While hydrocolloid 3 produced a low-fat spread with spreading stabiity, hydrocolloid 4 resulted in a spread with slightly less stability and 5 hydrocolloid resulted in a spread that was very unstable. Supplementary experiments with higher hydrocolloid dosages (e.g. 3.0% hydrocolloid 5) demonstrated that stable spreads could be obtained with hydrocolloids4 and 5 while maintaining their unique organoleptic properties. Table 1 Quality evaluation of law-farspread Hydrocolloid

Dosage

Spread *)

firmness Hydrocolloid 1 Hydrocolloid2 Hydrocolloid3 Hydrocolloid4 Hydrocolloid5 Hydrocolloid5

2.0% 2.0% 2.0% 2.0% 2.0% 3.0 YO

Stability Organolepticevaluation towards spreading Appea m c e Meltdown Flavour release

225g 330g 350g 360 g 260g 440g

I

8 10

10 8 4

10

grainy, dull smooth smooth smooth smooth smooth

average slow

average slow

fast

fast

veryfast veryfast veryfast

very fast very fast very fast I

Of the weight of the emulsion tested in: *)TA-XT2Texture M y s e r ')

3.4.4 Water droplet size distribution

The water droplet size distribution for the low-fat spreads containing the five hydrocolloids is shown in table 2. It is clear that a lowering of the gel onset temperature reduced the mean droplet size and gave a narrower droplet sue distribution. A narrow droplet sue distribution and small droplets are generally advantageous with respect to bacteriological stability. Very small droplets (2 pm or less) tend to give a slower flavour release. This is often the case in non-milk low-fat spreads, where the addition of

418

Gums and Stabilisers for the Food Industry 10

hydrocolloids increases the droplet sue and improves flavour release’. Even with very low gel onset temperatures, as with hydrocolloid 4, the droplet size is still sufficiently large to secure excellent flavour release. It is remarkable that the low-fat spread with gelatine (hydrocolloid 5 ) had relatively large water droplets, bearing in mind the low viscosity of the water phase during processing. The droplet size distribution results were further confirmed by the microscope analysis of freeze microton sections of the low-fat spread (figure 6), the low-fat spread emulsions with hydrocolloid 1 was almost bi-continuous in nature, explaining the deviating texture (table 1). The water phase droplets with gelatine (hydrocolloid 5) were less well-defined and less round in shape than those produced by the other hydrocolloids, confirming earlier observationss.Milk protein is also known to create large irregular droplets in spreads’, and the deviating behaviour of gelatine compared with the gelling vegetable hydrocolloids may have been due to insufficient water phase viscosity during processing, allowing time for the milk protein to partially destabilisethe spread before the gelatine gels and prevents further changes in droplet size.

Table 2 Droplet size distribution of low-fat spread with 2% hyd2.ocolloid (volume distributionwith mean and 95% confidnce interval. 50% < m

No. 3 No. 4

2.2

97.5% < m

6.6

* 3% dosage. The low-fat spread with 2% gelatine could not be measured. Microscope examination indicated larger droplets than with 3% gelatine.

419

Recent Developments, Future Trends

2% hydrocolloid 1

2% hydrocolloid 2

2% hydrocolloid 3

2% hydrocolloid 4

3% hydrocolloid 5 Figure 6: Photographs of low-fat spread freeze microton sections (200 x magnification)

420

Gums and Stabilisersfor the Food Industry 10

4 CONCLUSION The studies have shown that the gelling profile of gelling hydrocolloids influences the quality of low-fat spread with respect to appearance, texture, stability during spreading, droplet size, and, thus, microbiological stability. Furthermore, the studies have documented that the melting profile of gelling hydrocolloids influences the meltdown properties and flavour release of low-fat spread. The studies indicate that a low gel onset temperature below 30°C and a gel softening point below 37°C secures optimal low-fat spread quality. The processing of the low-fat spreads in the pilot plant showed that the hydrocolloids with a high viscosity in the temperature range 15-40°C (hydrocolloid 2 and 3) made the production of a W/O emulsion rather easy, whereas hydrocolloidswith a lower viscosity in the temperature range 15-40°C (hydrocolloid 4 and 5 ) did not facilitate a W/O emulsion to the same extent. The findings indicate that controlled water phase rheology during low-fat spread production may also benefit the production process. In addition to the rheological behaviour of the gelling hydrocolloid during low-fat spread production, the interaction between hydrocolloids and proteins may play a significant role in protein-containing low-fat spreads. An investigation of this interaction may brther contribute to an understanding of the hnctionality of hydrocolloids in these products. 5 REFERENCES

1. Madsen J., InternationalFood Ingredients, (1), p. 4-7, 1991. 2. Tuley L., International FoodIngredients, (2), p. 10-12, 1997. 3. Krawczyk G. R. Buliga G. S., Bertrand D. T., Humphreys W. M., INFURM, 7 , (6), p.635-639, 1996. 4. Dartey C. K., Sanderson G. R., INFORM, 7 , (6), p. 630-634, 1996. 5 . Clegg S. M., Moore A. K., Jones S. A,, J. FoodSci., 61, (5), p. 1073-1079, 1996. 6. Hermansson A. M., Altskiir A,, Jordansson E., in Gums and Stabilisersfor the Food Industry 9,ed. by Williams P. A. and Phillips G. O., The Royal Society of Chemistry, 1998, p. 107-116. 7. Walkenstrom P., Hermansson A. M., FoodHydrocolloiak, 12, p. 77-87, 1998 8. Norton I. T., European Patent Application EP 0474 299 Al, 1992. 10. Jones M. G., European Patent SpecificationEP 0 372 625 B1, 1993 11. Norton I. T., European Patent Application EP 0 369 550 A2, 1990

DESIGNING GALACTOMANNANSFOR THE FOOD INDUSTRY

M. Brooks, K. Philp, G. Cooney, L. Horgan Quest International, Kilnagleary, Carrigaline, Co. Cork, Ireland

1 ABSTRACT

The most commonly used galactomannans in the food industry are guar gum and locust bean gum. In food, guar gum is used primarily as a thickening and water binding agent whereas locust bean gum is well known for its interaction with other gums such as carrageenan and xanthan gum and this property is exploited in gelling applications where specific gel and texture characteristics are important. Guar gum is composed of a mannose backbone with galactose side chains and is isolated from the seeds of the annual leguminous plant Cyamopsis tetragonalobus. Locust bean gum, which has a reduced number of galactose residues attached to the mannose backbone compared to guar gum, is isolated from the seeds of the carob tree. Cyclical shortages in locust bean gum has attracted keen interest in supplying a hctional alternative at a steady cost. This challenge has led to the biotransformation of guar gum. Since the early modifications it has become possible to tailor guar to meet specific laboratory requirements in desired end-product applications and to produce these products on a commercial scale. Enzyme modification has allowed great flexibility in the properties of these new grades. The properties and characteristicsof a variety of modified guars are presented here. 2 INTRODUCTION

Galactomannans are reserve material of legume seeds. They are comprised of a linear (14)-P-D-mannan backbone *linkedto (1-6)-a-D-galactose residues. Natural commercially available galactomannans come from the seed endospenn of carob (Ceratoniu siliquu), guar (Cyamopsis tetrugonolobus) and less so from tara (Caesalpiniu spinosu). Guar is typically grown in India, Pakistan and the US and is used in the food industry as a water binder and thickening agent. Locust bean gum is grown in the Mediterranean and its synergy with gums such as kappa carrageenan and xanthan has been exploited for many years. Guar gum typically has been shown to contain approximately 39% galactose attached to the mannose backbone whereas locust bean gum has been shown to contain approximately 23% galactose'. Guar exhibits a limited interaction with xanthan resulting

422

Gums and Stabilisers for the Food Industry 10

in a viscosity increase rather than strong gelation and it does not have any appreciable synergy with kappa carrageenan. On the other hand locust bean gum interacts strongly with both xanthan and carrageenan yielding gels of different textures. There has been great interest in the area of enzyme-modified-guarsparticularly as a locust bean gum alternative. Locust bean gum is known to undergo cyclical shortages. The last shortage in the mid nineties saw the price of locust bean gum soar to almost thirty dollars per kilogram compared to a normal price of seven dollars per kilogram. A similar shortage occurred in the mid eighties. It is possible to modify guar gum by selective removal of galactose units to yield a modified guar gum with galactose levels similar to locust bean gum. This can be carried out using a specific a-galactosidase selective for polymeric g u d ’ . Much work has been carried out in the past on modification of guar gum by enzymatic means’-’. Here we present for the first time the properties of commercial scale modified guars (commercially available from Quest International under the trade name Sherex QSG). In addition to galactose removal it is also possible to modify the viscosity and solubility properties of enzyme-modified-guar gum. Modifications of this nature enable guar gum to be tailored to meet the needs of the food industry whether these requirements are product processing, finished product texture or effective replacement of other gums. 3 MATERIALS AND METHODS

3.1 Materials Commercial production of enzyme-modified-guar (Sherex QSG) is facilitated by the availability of purified a-galactosidase. In this study several enzyme-modified-guars with a range of galactose content (19, 21 & 25% galactose), viscosity and solubility were studied. Galactose content of guar, locust bean gum and the modified guars was analysed by Dionex ion chromatography (CarboPac PA1 column (25cm length, 4mm internal diameter), elution 0.01M NaOH at 1mYmin with pulsed amperometric detection). The locust bean gum galactose content used here is 21% as analysed by Dionex ion chromatography. 3.2 Preparation of gels

Carrageenan I enzyme-modified-guar gels were prepared with 0.4% semi-refined kappa carrageenan, 0.4% enzyme-modified-guar and 0.2% potassium chloride. The gels were made by dispersing the hydrocolloid / salt mix in deionised water before heating in a boiling water bath for twenty minutes. The hot solution was allowed to set at room temperature in lOOml gel jars (4cm height, 7cm diameter). After cooling overnight lcm was removed from the top of the gel in order to expose fresh gel surface before measurement. Xanthan I enzyme-modified-guar gels were made by firstly dissolving xanthan (0.5%) in hot deionised water. Enzyme-modified-guar (0.5%) was dispersed in deionised water before heating in a boiling water bath for twenty minutes. The individual hot solutions were mixed together in a ratio of 1:1 before pouring into gel jars. After cooling overnight lcm was removed from the top of the gel in order to expose fresh gel surface before measurement.

Recent Developments, Future Trends

423

3.3 Methods

Gel texture measurements were carried out using a Stable Micro Systems TA-XT2i Texture Analyser. Texture parameters hardness, break distance, cohesiveness, adhesiveness and chewiness were measured by texture profile analysis methods using O S d s test speed, 3 5 m m penetration distance, lcm' probe and a measurement temperature of 20°C. Texture of carrageenan / modified guar gels were evaluated by double compression of gels using an SMS-TAXT2i texture analysef. The terms used to describe the texture of the gels are defined as follows: Hardness is defined as the maximum peak force during the first compression cycle (g). Cohesiveness is defined as the ratio of the positive force area during the second compression to that during the first compression. Break distance is the distance to break of the gel in millimeters. Adhesiveness is defined as the negative force area for the first bite and represents the work required to overcome the attractive forces between the surface of a food and the surface of other material with which the food comes in contact. Chewiness is defined as the product of hardness, cohesiveness and stringiness. Gel set and melt temperature analysis was performed on a controlled stress rheometer Carri-Med CSL. A cone-plate geometry was used (cone diameter: 4.Ocm; angle: 2"; truncation: 60pm)at a frequency of 1 Hz. The cooling-heatingcycles were performed at 2°C per minute, between 80-5-80°C. Viscosity measurements were taken on a Brookfield DV-11 at 2OoC, 30 rpm (1% galactomannan). All samples were cooled to 20°C before measurement taken. Solubility is expressed as a percentage of viscosity relative to maximum viscosity. 4 RESULTS AND DISCUSSION

The effect of reducing the galactose content in guar allows a thickening agent to be transformed into a gelling agent. It is possible to selectively hydrolyse galactose from the mannose backbone by enzymatic means to a specific required level. Removing galactose from guar affects the polysaccharides' interaction with both carrageenan and xanthan. 4.1 Interaction with Carrageenan

Guar, which does not gel to any significant extent, is transformed by enzyme modification into a polymer which forms gels with carrageenan (figure 1). Gel strength is directly dependent on galactose content with modified guar containing 25% galactose gelling to a much lesser extent that modified guar containing lower amounts of galactose. Interestingly, it is possible to obtain a gelling synergy with carrageenan stronger than that currently available with LBG.

424

Gums and Stabilisers for the Food Industry I0

900

I

800 -

700 -

a 600 v

5

f d

500 400

-

300

-

200

-

100

-

18.5

19.5

20.5

21.5

22.5

23.5

24.5

25.5

Galactose (“7’0)

Figure 1. Hardness (g) of mixed kappa carrageenan / modified guar (QSG) gels versus galactose content (%). Texture profile analysis is an objective method of sensory analysis that attempts to link mouthfeel to a defined measurement. The test consists of compression of the gel twice in a reciprocating motion and attempts to mimic the action of the jaw while chewing. The texture of kappa / galactomannan gels can be varied using various modified guars (QSGs). Figure 2 illustrates how removing galactose from guar influences the texture of carrageenan / modified guar gels. In this illustration carrageenan gels with modified guar containing 21% galactose show texture closest to LBG with modified guars containing 19% and 25% galactose exhibiting texture differing to LBG. Thus removing galactose from guar provides the tools to modify traditional LBG gelling texture. Hardness

LBG --t

-

QSG 25% Gal

- -*-

QSG 21YOGal QSG 19% Gal

-+-

Figure 2. Texture profile analysis of mixed kappa carrageenan / modified guar (QSG) gels.

425

Recent Developments, Fuiure Trends

“t 100

0

m

lm,

1503

am,

2500

m

Figure 3. Viscosity (cps) of modified guar (QSG) versus hardness (g) of mixed kappa carrageenan / modified guar (QSG) gels. Guar is a highly viscous polymer with viscosities typically in the region of 5OOOcps (1% solution) but guar gum with viscosities in the region of 7OOOcps (1% solution) is also available commercially. Typically its viscosity is much greater than locust bean gum which varies around 3OOOcps. It is possible by enzyme modification of guar to produce a hydrocolloid which has the gelling characteristics of locust bean gum whilst retaining the high viscosity of guar. Viscosities of enzyme-modified-guar can range fkom 4OOcps 4ooocps. On examining kappa carrageenan / modified guar gel hardness versus modified guar viscosity (figure 3) it is clear that significant reduction in viscosity is possible without effecting gel strength. It is not until viscosities of 4OOcps are reached that gel strength is reduced. Using a lower viscosity can allow easier factory processing with galactomannans while retaining full gel strength with kappa carrageenan. 4.2 Interaction with Xanthan The gelling synergy between modified-galactomannans (QSG) and xanthan is illustrated in figure 4. Guar which does not gel to any significant extent is transformed into a polymer which now gels with xanthan above and beyond the synergy between xanthan and LBG. Interestingly even at the same galactose content QSG has stronger synergy with xanthan than LBG. This synergy is important in applications such as mayonnaise and cream cheese. This trend of increased synergy with kappa carrageenan and xanthan on reduction of galactose content of guar is also apparent in milk systems and follows the same trend regardless of viscosity and solubility of the modified-galactomannan. It has been shown that the maximum synergy of gelation for galactomannans is 1:l for kappa carrageenan whereas it is in the region of 1:3 for the xanthan / galactomannan interaction’. This also holds true for modified guars of a galactose content comparable to locust bean gum-

426

Gums and Stabilisersfor the Food Industry 10 90

,

60 ;5 0 80

70

v

40-

3020

10

-

18.5

19.5

20.5

21.5

22.5

23.5

24.5

25.5

Galactose (%)

Figure 4 Hardness (g) of mixed xanthan / modified guar (QSG) gels versus galactose content (YO). Examining the texture profiles of xanthan / galactomannans gels it can be seen (figure 5 ) that variations in texture are possible on varying galactose content of guar. Modified guar containing 21% galactose gives texture closest to locust bean gum whereas both lower and higher galactose containing modified guars give extremes of gel textures in the measured parameters. In the case of xanthan / modified-galactomannan gels the brittleness (break distance) is not effected to any great extent by galactose content, this is in contrast to kappa carrageenan / modified-galactomannan gels which show reduced brittleness with decreasing galactose content. Hardness

-c

QSC 25% Gal

-+-

QSG 21% Gal --%-

QSG 19vo Gal

-+-

Figure 5. Texture profile analysis of mixed xanthan / modified guar (QSG) gels. What other properties does removing galactose from guar effect? The melt/set profiles (figure 6) of xanthan / modified guar gels indicate that the higher galactose containing modified guar melts at a lower temperature compared to the higher galactose containing modified guar. This property is relevant in applications such as dessert jellies

Recent Developments, Future Trends

427

giving a melt in the mouth texture like gelatin or cold deposit desserts where the mix is cooled before filling into dessert pots.

M-i

I

300

zm

t3W

-

75

0

0 1 0 a 3 3 0 4 0 9 6 0 m ) 8 3 9 0

Figure 6. Melt / set profile of xanthan / modified guar (QSG) gels. 4.3 Solubility

I2O

G

10

30

50

70

=wmhm

90

110

(00

Figure 7. Solubilityof locust bean gum and modified guars (QSG). Guar gum is a cold water soluble polymer whereas locust bean gum is typically a hot water soluble polymer. Removing galactose fhm the mannan backbone of guar has the effect of increasing the solubility temperature of the polymer similar to that of locust bean gum. During production of enzyme-modified-gum, close control of processing conditions make it possible to reverse this process and to retain the original cold water soluble property of guar. In effect both hot and cold water soluble enzyme treated guars can be produced. Figure 7 illustrates the solubility of cold water soluble modified guar compared to standard modified guar and locust bean gum. Solubility is expressed as a percentage of viscosity relative to maximum viscosity and all measurements were taken at

428

Gums and Stabilisersfor the Food Industry 10

20°C. It is clear that at 20°C cold water soluble guar has reached maximum solubility whereas standard modified guar and locust bean gum both attain maximum solubility at approximately 85°C. This property has use in cold mix desserts and ice creams processed at low temperatures.

5 CONCLUSION In a world of increasing competition cost reduction is a key factor to maintaining a competitive edge. This can be done in many ways i.e. by achieving higher functionality with existing ingredients or replacing costly ingredients with more cost effective ones. It may also be necessary to replace ingredients for other reasons e.g. gelatin replacement particularly in light of the BSE scare. Other industry requirements include the need for ingredients with novel textures enabling companies to launch new products I textures into the market place and ingredients which enable ease of processing or more cost effective processing. Enzyme-modified-guar can answer many of these industry challenges. The ability to modify key properties of guar offers many exciting avenues to the food industry in terms of end product requirements and end product processing. Much study has been done in this area in the past and it is the subject of many publications'-'. Turning this research into commercial scale modified-galactomannan products has provided many challenges. In the case of enzyme-modified-guar the key, in recent years, has been a combination of innovative processing and advances in processing technology.

Acknowledgements Part of the work presented here was carried out with the help of M. Marrs and J Skinner at Leatherhead Food Research Association. I would also like to acknowledge the contribution of P. Lyons, W. Koelewijn & J. Dunlea.

References 1.

B. V. McCleary, I. C. M. Dea, J. Windust & D. Cooke, Carbohydr. Polym., 1984, 4,253.

2.

B. V. McCleary, H. Neukom, Prog. Fd. Nutr. Sci., 1982,6,109.

3.

P. V. Bulpin, M. J. Gidley, R. Jeffcoat & D. R. Underwood Carbohydr. Polym., 1990, 12, 155.

4.

B. V. McCleary, R. Amado, R. Waibel & H. Neukom, Carbohydr. Res., 1981,92, 269.

5.

I. C. M. Dea, A. H. Clark & B. V. McCleary, Carbohydr. Res., 1986,147,275.

6.

M. Pons & S . M. Fiszman, Journal of Texture Studies, 1996,27,597.

7.

I. C. M. Dea and A. Morrison, Adv. Carbohydr. Chem. Biochem., 1975,31,241.

Hydrolyzed and deodorized guar gum including other guar specialty products: functional properties and applications

Florian M. Ward, Ph.D. TIC Gums, Inc., Belcamp, Maryland MD 21017 Abstract

Guar gum is a galactomannan isolated form the seeds of a plant, Cyamopsis tetragonolobus (Family Leguminoseae). Guar gum is widely used as a stabilizer and thickening agent with a typical viscosity of 3,500 cps at 1%gum level. A specialty product such as hydrolyzed guar gum was prepared by breaking down some of the glycosidic linkages, with an enzyme, galactomannanase. This yields a low-viscosity ingredient that can be used at higher levels as a film-former and a good source of soluble dietary fiber. The molecular weight distributions of the various products were evaluated by gel permeation chromatography. The rheology and functional properties were also determined. One of the constraints in the use of guar gum is its mealy off-flavor and odor. Hence, various methods were evaluated to produce a lowcost deodorized guar gum. The reduced odor guar was subjected to sensory, chemical and functional tests. Prehydrated or agglomerated guar gum is another product which has the following features: dust-free, no lumping, faster hydration, etc. Functional properties of the various specialty products and their applications are discussed. Introduction

Guar gum, widely used in the industry as a lowcost thickening agent, is a plant polysaccharide (1, 2), isolated from the seeds of a plant, Cyamopsis tetragonolobus (Family Leguminoseae). Guar gum has a mannan backbone, with a mannose to galactose ratio close to 2:l (Figure 1). Guar gum has a molecular weight ranging from 250,000 to 2 million. It is a cold water soluble, non-ionic, and salt tolerant galactomannan. Although guar gum is commonly used as a thickening and stabilizing agent, its use in some products is limited due to its characteristic off-flavor and mealy taste. This odor and flavor may arise from the breakdown of residual lipids, proteins, as well as degraded isoflavonoid compounds. The high viscosity (typically, 3,500 cps at 1%) also precludes the addition of higher levels of guar gum as a source of soluble dietary fiber in beverages, meal replacers and similar products. In order to minimize some of these constraints in the use of guar for the food and pharmaceutical industries, some guar specialty products have been developed, processed, and made available to food manufacturers and consumers.

430

Gums and Stabilisers for the Food Industry 10

Figure 1. The partial structure of guar gum, a galactomannan with a typical mannose to galactose ratio of 2:1. Hydrolyzed guar gum: a specialty product The viscosity of guar gum can be significantly reduced by chemical or enzymatic hydrolysis of its glycosidic linkages. Some of the specialty products of guar gum from TIC Gums, Inc., include Guar MLV (medium viscosity, about 1,500 cps minimum at 2%), Guar VLV (300 cps. minimum at 2%) and the high viscosity guar gums (4,000 cps minimum) in fine and coarse mesh grades. A low viscosity product derived from enzymatic hydrolysis of the original guar gum isolated from the seeds, known as Guar HLV, has a viscosity of 300 cps maximum at 10 Om. Table 1 shows the changes in viscosity vs. increasing guar HLV levels. At 1 %, the viscosity ranges from 10 to 50 cps, as compared to 3,500 cps for the unhydrolyzed, regular guar gum. At a 15% gum level, the viscosity ranges from 600-1,000 cps. Analyses for soluble fiber using alcohol precipitation techniques, shows that the partially hydrolyzed guar retains its high fiber content.

Table 1. The effect of increasing hydrolyzed guar levels (%) on the viscositV ranges using a Brookfield rheometer. % 1

I I

Viscosity (cps) 10-50 (LV2@60rpm) 150 - 300 (LV2 60 rpm) 600 - 1,000 (RV4 @ 20 rpm) 1,700 - 2,000 (RV4 @ 20 rpm)

6

Recent Developments, Future Trends

43 1

The pseudoplastic nature of hydrolyzed guar gum is shown in Figure 2, where the viscosity of a 10% solution of hydrolyzed guar slightly decreases with increasing shear rate. The effect of increasing temperature from 20" to 80°C is shown in Figure 3. At a pH from 5 to 7, the guar gum regains the original viscosity after cooling the 10% solution to ambient temperature. 10000

-

1000

0

.-c*r$

100

I 0

M

I3 I0

so

0

I

150

100

200

RPM

+lo%

GuarNT HLV

+20%

-15%

Figure 2. Shear thinning property of hydmlyzedguar of lo%, 15%,and 20% solutions as shown by decreases in viscosity versus increasing shear rate, as measuredby a programmable Brookfieldrheometer. 150 125 100 75

so 25

0 0

10

20

30

40

50

80

70

80

90

Te m pera tu re ('C)

Figure 3. Changes in viscosity of a 10% solution of hydrolyzed guar gum at varying temperature (20" to SOOC), using a Brookfieldrheometer.

432

Gums and Stabilisers for the Food Industry 10

Using gel permeation chromatography, the molecular size distributions of the guar gum were evaluated before and after hydrolysis with galactomannanase enzyme. The molecular weight distribution of the original guar gum before hydrolysis is shown in Figure 4, with the elution times of the higher MW and

I6

14

before hydroiysls 12

Qt

s

10

g

8

6

Elution Tlme (min.) I 2

1 4

I 6

I I)

I

I

I

I

I

10

12

14

16

18

Figure 4. The molecular weight distribution of the original, unhydrolyzed guar gum showing major fractions, using gei permeation chromatography (Biorad column) with a refractive index (RI) detector,

lower MW fractions using a BiogellBiorad gel filtration column. Figure 5 shows the molecular weight distribution of the galactomannan, guar HLV after hydrolysis. The comparison of the molecular weight distribution before and after enzyme treatment can be achieved more easily by superimposing chromatograms. A significant shift to lower molecular weight fractions after hydrolysis with galactomannanase from Aspergillus sp. is evident. Applications and health benefits from hydrolyzed guar gum The cholesterol lowering effect of guar gum administered to hypercholesteremic human subjects has been reported (3, 4). The hydrolyzed guar, which retains its high soluble fiber content based on standard fiber tests, is recommended for use in products where a lower viscosity is desired in the finished formulation. Hydrolyzed guar gum, due to its relatively low viscosity may be used in beverages and meal replacers as source of soluble dietary fiber at considerably higher gum levels in nutraceutical products and functional foods.

Recent Developments, Future Trends

433

-

18

I@ -

hemicellulaseandlor I4

I*

-g-

-f

0 -a

10

8 6 -

Elution T i m (min.) 0

I 2

I

I

4

6

I 8

I 10

I I2

1

14

I

I

16

I8

Figure 5. The molecular weight distribution shift for guar HLV hydrolyzed with a galactomannanaseisolated from Aspergillus sp. (analysis as in Figure 4). Reduced odor guar: another specialty ingredient

A proprietary process that does not involve the use of organic solvents was developed by TIC Gums, Inc. to inhibit any enzymecatalyzed breakdown and retard formation of degradation products that may account for undesirable volatile flavor and odor. A series of process operations using guar splits derived from the seeds, yields a bland-tasting guar gum with improved flavor and odor. Figure 6 shows the molecular weight distribution of reduced odor guar gum, indicating that it is very similar to that of the original guar gum as shown in Figure 4. Thus, there is no significant degradation due to the deodorization process. The reduced odor guar gum is marketed as GuarNT@Bland by TIC Gums, Inc. Sensory and objective evaluation of reduced odor guar

Various methods for deodorizing guar gum processed from guar splits were evaluated using sensory as well as objective methods. In sensory tests, nine out of ten panelists preferred GuarNT@Bland to the regular guar gum as to odor and flavor. Hexanal (an index of grassy odor) was analyzed by gas chromatography in control and treated samples. A 75-90 % reduction in hexanal levels was observed (5) for treated samples. The PCA (principal component analyses) of guar gum samples were also conducted using the Aromascan, (Foss North America) to determine if there

434

Gums and Stabilisersfor the Food Industry 10

14

12

t ;

0

Ellltlon Tm (mh.) I 4

I

1

8

8

I 10

I 12

I 14

I 18

I 18

Figure 6. The molecular weight distribution of reduced odor guar gum showing similarity to the original unhydrolyzed guar gum (analysis as in Figure 4). were significant differences in the various treatments. The PCA map (Figure 7) allows sample to sample comparisons by reducing 32 parameters of data (generated by 32 sensors in the system) to a single point and is an objective tool that complements sensory data and chemical analyses of flavor indices (5).

Rheological data on reduced odor guar The viscosity changes in reduced odor guar with increasing gum levels vary from 0.1% (8 cps) to 1.0% (about 1,900 cps) solutions in water after 2 hours. At 24 hours, the viscosity of 2,400 cps is attained at a 1.O% gum level. By varying shear rate using a programmable Brookfield rheometer, the shear thinning property (pseudoplasticity) of GuarNT@Bland was noted. This property makes it suitable for handling during the manufacturing phase of a thickened finished product, e.g., beverage or salad dressings, etc. As the temperature is increased from 20" to 100°C, the viscosity of reduced odor guar decreases and regains the initial viscosity after cooling down to the original temperature.

Applications of reduced odor guar gum The application of reduced odor guar as a thickener and stabilizer has various advantages in the food and nutraceutical industries. It has significantly reduced odor and flavor and hence can be used as a lowcost, "natural" thickener and stabilizer in delicately flavored products, including dairy formulations. It contains a minimum of 80% soluble dietary fiber, which has

435

Recent Developmenis, Future Trends

been reported to help decrease serum LDL-cholesterol levels (3,4) and increase glucose tolerance. It has high water binding capacity and hence can prevent weeping or syneresis in frozen and baked products. 5

Sample 1

0 hl

a

2

o

Sample2

A

Sample 3

0

Sample4

-5 -10 -15 -8 -6

-4-2

2 4 PCA 1 0

6

8

10

Figure 7. Principal component analysis (PCA) showing clustering of repkates and the differences between products using various deodorization treatments on guar gum. [Source: (5)J Due to its solubility in cold water, it can be used in instant drinks as a cost-effective thickener and suspending agent, and fiber source. It can be added to meal replacer powders, tablets, and other similar products to enhance soluble dietary fiber. In bakery products, guar gum can be used as a component of fat-mimetic systems (6). Combining reduced odor guar gum with gum acacia or gum arabic, pectins, CMC, or xanthan gum yields stabilizers with improved and modified functional properties. A combination of GuarNT@Bland with gum arabic yields a low-cost adhesive system that acts as binder. Guar gum combined with cellulose gum (sodium carboxymethyl cellulose) is used in cappuccino or cocoa mixes to suspend the milk and whey proteins at reduced cost as compared to using CMC or carrageenan alone. Used with pectin, reduced odor guar can be applied to stabilize acidified milk with a pH ranging from 4.0 to 4.5. A very fine mesh version, known as GuarNT@Bland 200 is recommended in applications where faster hydration is desired and which requires the absence of particulate matter in the finished product, as in instant powder drink mixes or beverages. GuarNT@Bland and the other specialty products may be used as a

436

Gums and Stabilisers for the Food Industry 10

stabilizer, thickener, or suspending agent, as well as a source of dietary fiber (7) for nutraceuticals or functional foods. The maximum usage levels permitted for guar gum by the US Code of Federal Regulations are shown in Table 2, for products including baked goods, cheese, dairy, gravies and sauces, soups and soup mixes, milk products and all other food categories (8).

Table 2. Maximum usage levels permitted for guar gum by the U.S. Code of Federal Regulations (8).

Guar gum with low microbial counts Guar gum made available in the market, may be high in microbial load, and have a variable particle size range as well as seed impurities and inedible components. Hence, it is essential that the raw materials be processed under controlled conditions by employing the HACCP (Hazard Analysis Critical Control Point) system approach. Reducing bacteria, yeast, and mold counts may be achieved by heat treatment.during processing of the guar, instead of adding preservatives or other additives. Consequently, guar gum with low bacterial counts (1,000 cfulg maximum) and consistent quality specifications for food grade and pharmaceutical products are now commercially available.

Recent Developments, Future Trends

437

Agglomerated and dust free guar gum

Another problem that is associated with the use of guar gum and other hydrocolloids is the dust generated during incorporation with dry mixes and product formulations. The prehydrated PH@Guar Series, which has been subjected to an agglomeration process, includes specialty products from guar gum and is available in various grades and viscosity ranges. The regular powder is converted to a virtually dust-free product with better hydration characteristics than regular guar gum. The prehydration or agglomeration process involves the introduction of water (no additives) and subsequent drying of the agglomerates with turbulent air. The prehydrated product has a lower bulk density than regular guar gum, hydrates faster when added to water before other ingredients, and is a dust-free, free-flowing powder. Quality specifications

Based on the end application of the guar gum, the food manufacturer can choose from various guar specialty products and propose specifications for microbiological limits, viscosity ranges, particle size distribution, soluble dietary fiber content, and other parameters. These specifications may be in addition to the requirements set by the Code of Federal Regulations or the National Formulary (ash and heavy metals limit, galactomannan content, protein levels, moisture etc.) Conclusions and recommendations

Hydrolyzed guar gum as a source of soluble dietary fiber may be used in functional foods or nutraceuticals without the problem of imparting high viscosity to the finished product. Reduced odor guar samples showed a significant decrease in odor and flavor, based on the sensory and objective tests conducted on the control and on the treated samples. Thus it is recommended as a multi-functional food and pharmaceutical ingredient as well as a good source of high soluble dietary fiber. Due to its bland taste and odor, usage levels can be increased accordingly, especially in products with delicate flavor, subject to limits set by the Code of Federal Regulations. The other specialty products manufactured from guar gum offer the consumer or manufacturer a variety of alternatives and enable them to use a readily available, cost-effective food ingredient that has modified functional properties as well as health benefits when utilized as recommended.

438

Gums and Stabilisers for the Food Industry I0

Acknowledgments The author would like to thank Stephen Andon and Chris Andon, TIC Gums, Inc., Stuart Cantor, Jim Caulfield, Ricardo Pereyra, and Mar Nieto, from the R&D group, for assisting in the development and evaluation of the guar products. The author appreciates the technical help of Marcy Russell, Foss North America, for the AromaScan analyses, and Dr. Richard Ward for his assistance in preparing the graphics and editing the manuscript for publication. References 1. Whistler, R. L. and J. N. BeMiller (1993). Industrial Gums, 3'd Ed. Academic Press, San Diego California, U.S.A. 2. Glicksman, M. (1982). Food Hydrocolloids 3, 158. CRC Press, Inc. Boca Raton Florida, U.S.A. 3. Jensen, C.D. et a/. (1997). Long-term effects of water-soluble dietary fiber in the management of hypercholesterolemia in healthy men and women. Am. J. Cardiol. 79, 34-37.

4. Wilson, T. A., S. R. Behr, and R. J. Nicolas (1998). Addition of guar gum and soy protein increases the efficacy of the AHA step 1 cholesterol-lowering diet. American Society for Nutritional Sciences 1429-1433. 5. Ward, F. M. (1998). GuartNT@' Bland: reduced odor thickener and soluble fiber. Polymer Preprints %:1, 691. American Chemical Society, Washington, DC, U.S.A. 6. Ward, F. M. (1997). Hydrocolloid systems as fat mimetics. Cereal Foods World 42:5,386.

7. Dreher, M. L. (1987). Handbook of Dietary Fiber. Marcel Dekker, Inc., New York, New York, U.S.A.

8. U.S. Code of Federal Regulations (1997). Title 21, Parts 170 - 179. Federal Register Office, Washington DC, U.S.A.

MECHANICAL AND MOISTURE BARRIER PROPERTIES OF HYDROXYPROPYL RICE STARCH AND HYDROXYPROPYL RICE STARCH-POLY(ACRYL1C ACID) GRAFT COPOLYMER FILM

B.M. Noor Mohd Azemi, I. Zainal Ariffin, A.R. Mazidah, A. Abd Karim* Food Technology Divison, School of Industrial Technology, 11800 Universiti Sains Malaysia, Penang, Malaysia

ABSTRACT Physical properties of two cast films, namely, hydroxypropyl rice starch (HP) and hydroxypropyl rice starch-g-poly(acry1ic acid) (HP-g-PAA) films were investigated. Films were evaluated for water vapour permeability, tensile strength and elongation as a function of solvent composition, different types and concentrationsof plasticizers (for HP film) and acrylic acid (for HP-g-PAA film). Increasing concentrations of both glycerol and sorbitol in HP film resulted in significantly improved film extensibility but reduced film tensile strength and water vapour permeability. Water vapour permeability ranged from 0.82 to 1.58 g.rnm.m'2.day-'.mmHg" for HF-film. Sorbitol was more effective than glycerol as a plasticizer in that sorbitol-plasticized HP-film exhibited marginally lower water vapour permeability and marginally greater tensile strength. Grafting of poly (acrylic acid) onto hydroxypropyl rice starch was evidenced by FTIR analysis whereby the existence of a peak at the wavelength 1715 cm-' indicated the presence of a carbonyl group in the HP-g-PAA composite. Water vapour permeability ranged from 0.89 to 1.23 g.mm.m-2.day-'.mmHg-' for HP-g-PAA film and showed an increasing trend as the concentration of acrylic acid (010% w/w) was increased. Extensibility of the HP-g-PAA film increased linearly with increasing concentrationof acrylic acid but tensile strength of the film showed a maximum value at 2.5% w/w acylic acid. 1 INTRODUCTION

Research into edible coatings and films has been increasing in recent years, with all indications that interest will continue. An edible film coating or film is prepared from edible material that acts as a barrier to the external elements (factors like moisture, oil, vapour, gases) and thus protects the product and extends its shelf life'. Development and studies on edible coating or film have been reviewed in detail by Kester & Fennema2 and Krochta & Mulder-Johnston3. The materials that have received the greatest attention for edible film use are cellulose ethers, starch, hydroxypropylated starch, corn zein, wheat gluten, soy protein, and milk proteins. The possibility of casting commercial films from starch products has intrigued researchers for many years. Starch as a renewable and biode adable polymer is readily available in high purity and at relatively low cost. Amylose4-yhigh amylose starch6and

440

Gums and Stabilisersfor the Food Industry 10

hydroxypropylated high amylose starch7 have been used to form self-supporting films from aqueous solution. In the absence of additives, however, films made from starch or amylose are brittle and sensitive to water (hygroscopic). Starch can be made thermoplastic when a plasticizer such as water* or polyols are added. Amylose and hydroxypropylated high amylose starch can also be extruded to form films’. Grafting has been used as an important technique for modifying physical and chemical properties of polymers. Graft polymerization results fiom the formation of an active site at a point along a polymer molecule other than its end followed by exposure to a second monomer. Typically, graft polymers are prepared by fust generating free radicals on starch and then allowing these free radicals to serve as macroinitiators for the vinyl or acrylic monomer. Free radicals can be produced by chemical methods, e.g., ceric ammonium nitrate” or by high-energy radiation”, e.g., 6oCo. The discovery that starch can serve as the backbone for grafting synthetic polymers has led to a number of interesting starch copolymers. Notable among them are starchpolyacrylonitrile and starch-polyvinylalcohol. Otey et al.l2-l4 have demonstrated that aqueous dispersions of starch and poly(ethy1ene-co-acrylic acid) (EAA) can be processed to films, containing more than 50% starch and show potential as packaging materials. Otey et al.I4 have suggested that compatibility between starch and EAA might result from hydrogen bond formation between EAA carboxyl groups and the hydroxyl substituent of gelatinized starch, although interactions between the two polymers were not examined in detail. More recently, Fanta et a!l and Shogren et al.I6-l7have presented a large body of evidence to support the idea that starch and EAA form helical inclusion complexes similar to those formed from starch and fatty acids. Part of this evidence was the immediate increase in viscosity observed when solutions of starch and EAA were mixed. Viscosity did not increase when the microbial polysaccharide dextran was substituted for starch, presumably because the a-(1,6) linkage between glucose units in dextran does not permit formation of the required helical structures”. Since starch-based biodegradable packagings are of considerable potential importance, hrther studies in this area are appropriate. The technique of graft copolymerization provides a promising approach to combine synthetic polymer with natural polymer to produce versatile, tailor-made materials with the best properties of both types. For application of films to food systems, it is important to develop films possessing favourable mechanical and permeability characteristics. Therefore, combined analyses are crucial for predicting film behaviours and defining structure/functionrelationships. On this premise, we have undertaken a study to examine and compare the mechanical and moisture barrier properties of hydroxypropyl starch (HP) film and hydroxypropyl starch poly (ethylene-coacrylic acid) (HP-g-P(AA)) film based on rice starch. 2 MATERIALS AND METHODS 2.1 Materials

Rice starch was obtained fiom Sigma Chemical Co. (St. Louis, MO). Acrylic acid (anhydrous, >99%) was obtained from Fluka Chemika, melting point 12-15OC. Propylene oxide, sodium sulfate, barium chloride, sodium hydroxide and sulphuric acid, all in

Recent Developments, Future Trends

441

analytical grade, were purchased from BDH. Polyethylene glycol (PEG 400, PEG 8000) were obtained from Aldrich Chemical Co., Milwaukee, WI, and glycerol and sorbitol (ACS reagent grade) procured fiom Sigma Chemical Co., St. Louis, MO. 2.2 Preparation of Hydroxypropyl Rice Starch and Film

Hydroxypropyl rice starch was prepared according to the method of Mohd Azemi and Wootton'*. Starch (1000.0 g), sodium hydroxide (12.0 g), sodium sulfate (125 g) and water (120 ml) were weighed into screw cap flasks. The contents were mixed thoroughly and heated to 40OC. The starch suspensions were treated with 60 mi propylene oxide, sealed and mixed in a gyratory water bath for 24 h at 4OOC. The samples were then neutralized with 0.1 N HzS04 and centrifuged at 5000 x g for 20 min and supernatants discarded. The modified starch samples were then lyophilized in a fiee-dryer (Virtis Model 10-MR-TR). The dried samples were ground to 0.315 mm mesh size and kept in an airtight container for further use. Hydrox ropy1 content of the starch was determined by the spectrophotometricmethod of Johnson'Yg and expressed as molar substitution. For film preparation, hydroxypropyl rice starch was mixed with deionized water and homogenized in a homogenizer (Ultra-Turrex T25) to obtain a homogeneous slurry. The slurry was heated to 90°C for 10 rnin in a water bath with continuous stirring using a magnetic stirrer. About 25 ml solution was poured onto smooth perspex plate (12 cm x 15 cm x 3 mm) sitting on a levelled bench top. The solution was allowed to dry in an oven for approximately 16 h at the temperature studied (40' and 6OOC). Dried films could be peeled intact from the casting surface. The cast films were stored at 25OC, 50% RH, for at least one day before being used for any measurements.

2.3 Preparation of Hydroxypropyl Starch Poly (Ethylene-co-Acrylic Acid) Dispersion and Film Graft polymerization was carried out by the methods adopted by Bayazeed et al." (1989) and Heibeish et al.*' (1992). The graft polymerization reaction was carried out in a 500 ml stoppered flask containing an aqueous solution of acrylic acid and hydroxypropyl starch. The starch slurry was prepared at a ratio of 1:lO starch to water and various levels of acrylic acid (0.0, 2.5, 5.0, 7.5, 10.0% w/w dry starch). The flask was placed in a water bath (5OOC) for a certain period until the required temperature was attained, and the initiator (potassium persulfate; 8 mmol 1-') was then added. The content of the flask was shaken continuously during the polymerization reaction for 90 min. After the desired reaction time, the content of the flask was poured over 1 liter ethyl alcohol (85%) where a precipitate was formed. The mixture was stirred using a magnetic stirrer and boiled for 15 min. The pecipitate was washed twice with hot ethanol (85%), separated by centrihgation and lyophilized. Film preparation was done using sorbitol(l5% w/w dry starch) as a plasticizer, mixture of water and ethanol (70:30) as a solvent and a starch to solvent ratio of 1:20. Films were cast as described for hydroxypropyl rice film and dried at 4OOC.

Gums and Stabilisersfor the Food Industry 10

442 2.4 Film Characterizations

2.4.1 Water VapourPermeability (WVP). Water vapour permeability (g.mm. m-2day-1.mmHg-1) of the dried film was determined gravimetrically by the ASTM standard method22 E96-80 with some modifications. Before testing, all film specimens were conditioned and stabilized at the temperature of the experiment (30°C) for 6 h over silica gel desiccant. Five thickness measurements were taken on each film, one at the centre and four around the perimeter. The mean was used as the film thickness in WVP calculations. The film was sealed to the rim of test cup containing silica gel (0% RH, 0 mmHg) and the cup was then placed in a desiccator maintained at 100% RH (32.3 mmHg) with distilled water. Thus, the RH gradient across the film was 100% and water vapour pressure was 32.3 mmHg. Changes in the weight of the cup were recorded to the nearest 0.0001 g. Steady state conditions were assumed to be attained when the change in weight of the cup over time became constant. Slopes (changes in weight of the cup over time) were calculated by linear regression and the correlation coefficients for all reported data were 2 0.99. The effective surface areas were measured at the end of each experiment. Under steady state conditions, the permeability constant k is calculated from a plot of moisture (total weight of cup) gain or loss (w) versus time using the following equations: Water vapour permeability = (w.x) / A.T. (PI - P2) where w = weight gain of the cup over 24 h, g x = film thickness, mil (1 mil = 0.001 in) A = area of exposed film, m2 PI - Pz = water vapour pressure differential across the film T = time, day WVP for each type of film was determined in triplicate with individually prepared and cast films as the replicated experimental units and three subsamples (specimens) tested from each film replicate. 2.4.2 Tensile Strength and Elongation at Break. Tensile strength and percent elongation at break of the films were performed using an Instron Universal Testing Machine (Model 100, Instron Co., Canton, MA) mounted with a 10N load cell, according to ASTh4 Method" D 882-88 (1989). Films were preconditioned at 25°C at 52% RH for 7 days before testing. Rectangular strips (100 mm long and 15 mm wide) were cut for the tensile test. Film thickness was determined as an average of seven measurements with a hand-held micrometer (B.C. Ames Co., Waltham, MA). Testings were carried out at 25°C and cross head speed of 10 mm min". Tensile strength (TS) was calculated by dividing the maximum load (MPa) by the cross-sectional area of the film. Percentage elongation at break (E) was calculated by dividing film extension at the moment of rupture by 50 mm (initial gauge length of specimen) and multiplying by 100. TS and E for each type of film were determined in triplicate with individually prepared and cast films as the replicated experimental units and seven subsampies (strips) tested fkom each film replicate. 2.4.3 Fourier Transfrom Infrared Spectroscopy. h h e d spectra of hydroxypropyl starch, acrylic acid and the graft copolymer were recorded on a FTIR spectrophotometer (Perkin-Elmer Model 1600).

Receni Developments, Future Trends

443

3 RESULTS AND DISCUSSION 3.1 Effect of plasticizers on physical appearance of HP-film

Effect of plasticizers on mechanical properties of HP-film was studied using four types of plasticizers, namely polyethylene glycol (PEG 8000, PEG 400), sorbitol and glycerol. Unplasticized HP-film could not be measured as it was too brittle to be handled. HP-films containing sorbitol and glycerol were clear, indicating that these plasticizers, at the level used, were compatible with the HP starch. HP-film containing PEG 8000 and PEG 400 (at 20% level) had a white residue on the surface. The appearance of a white residue on the or surface of edible films containing plasticizers has been referred to as LLbloomin~3” This occurs when the plasticizer concentrations exceeds its compatibility limit in the polymer causing phase separation and physical exclusion of the plasticid3. Plasticizers are added to films in order to reduce brittleness and increase flexibility, toughness and tear resistance. They lead to a decrease in intermolecular forces along the polymer chains which generally produce a decrease in cohesion and tensile strength. The plasticizer must be compatible (miscible) with the polymer and if possible readily soluble in the solvent (to avoid premature separation during film drymg’. From the physical properties examined in this study, both sorbitol and glycerol appear to be the most effective and suitable plasticizers for HP-film. Thus, sorbitol and glycerol were chosen as plasticizers to further study the effect of solvent and drying temperature on the mechanical properties of the HP-films. 3.2 Water Vapour Permeability of HP-Film 3.2.I Effect of Solvent Composition. It is important to understand the effect of solvent composition because the film-forming solvent system influences characteristics of the finished film. Maximum solvation and extension of the polymer molecules will yield the most cohesive film structure^^^. Solvent systems for edible films or coatings are limited primarily to water, ethanol, or a combination of the two. Since the principal b c t i o n of a film is to impede moisture transfer, the measurement of WVP is of practical importance to predict the degree of resistance of the film towards moisture migration across the film surface.

Table 1 Water Vapour Permeabilities of Hydroxypropyl Films Prepared From Solvents of Variolls Compositiom Solvent composition, water:ethanol (%v/v)

1oo:o 90:lO 80:20 70:30 60:40 50:50

Water Va OUT PermeabiliVb (g.mm.m .day-‘.mmHg-’)

-Y

1.28 f 0.09 1.27 f 0.06 1.26 f 0.06 1.24 f 0.07 1.32 f 0.08 1.36 f 0.07

Films were prepared at 40OC; film thickness 0.085 f 0.015 mm Means of four replicates f 95% confidence interval

444

Gums and Stabilisersfor the Food Industry 10

Data in Table 1 shows the effect of solvent composition on the WVP of HP-films. Water vapour permeability of HP-film was apparently unaffected in the presence of up to 40% (v/v) ethanol. However, WVP of HP-film made from a 5050 mixture of water:ethanol exhibited significantly higher (p < 0.05) value than that of HP-films made from solvent with less than 40% ethanol. The results obtained were in agreement with that of Donhowe and FennemaZ6who found that methylcellulose films prepared from a 75% water-25% ethanol solvent exhibited smaller WVP than films prepared from water or higher ethanol composition. This was attributed to the ability of ethanol, at 25% concentration, to enhance intermolecular hydrogen bonding of methylcellulose, while higher concentrations of ethanol may have prevented complete hydration of methylcellulose. The same line of reasoning is applicable to explain the effect of ethanol on WVP of HP-film. In addition, it is interesting to note that water and ethanol form an azeotropic mixture with lower boiling point (ca 85OC) than the individual solvents; this could increase the rate of drying of the film. The disadvantage is that excessive rate of solvent evaporation may prematurely immobilize polymer molecules before they have an opportunity to coalesce into a continuous, coherent film. This may produce defects such as slipping and peeling or development of pin holes. Consequently, WVP of the film increases as a result of imperfection in the film. 3.2.2 Effect of Plasticizers and DTing Temperature. The effect of plasticizer concentration on WVP of HP-films is illustrated in Figure 1. Films dried at 40°C (without plasticizer) have WVP of approximately 0.825 g.mm.m-’.day-’.mmHg” and this value increased significantly(p < 0.05) to about 1.382 and 1.526 g.mm.m-’.day-’.mmHg-’ in the presence of 20% sorbitol and glycerol, respectively. From Figure 1, it is evident that the WVP through plasticized HP-films increased with increasing sorbitol and glycerol content. This effect is typical of plasticizers and have been reported by many workers’6-’8.This could be related to structural modifications of the hydroxypropyl starch network which might become less dense and to the hydrophilicity of the sorbitol and glycerol molecule which is favourable to adsorption and desorption of water molecules. Researchers who have examined the interaction between plasticizers and film matrix have concluded that plasticizers interact with amorphous regions of polymers because they decrease the glass transition temperature of the p~lyme?~*’’~’~. Water increases the free volume of the polymer, thereby increasing polymer mobility and permeability. Polyol plasticizers are also hypothesized to act in the same manner’’.

Water vapour permeability of the sorbitol- and glycerol-plasticized HP-films is comparable up to 15% w/w plasticizers (Figure 1) but significantly different (p < 0.05) at 20% w/w of plasticizers. At this concentration, glycerol films exhibited higher WVP values than sorbitol films. The results obtained were in agreement with those reported by Donhowe and FennemaZ6for methylcellulose, McHugh and Krochta” for whey protein films and Porte?’ for hydroxypropylcellulose, respectively. Glycerol was hypothesized to compete with water for active sites on the polymer, thus promoting water clustering and increased free volume in the polymers at low moisture levels3’. The effect of drying temperature on WVP is also shown in Figure 1. It can be seen that films dried at 6OoC exhibited higher WVP than that of films dried at 40°C. Higher drying temperature increased solvent evaporation rate and may produce non-cohesive and defective film. Evidence for this comes from microscopic examination of the film (results not shown) which shows lack of surface continuity and hairline cracks on the film surface.

445

Recent Developments, Future Trends

0

10 15 5 20 Concentra In of plasticizers (% w/w)

I 1

I

Figure 1 ESfect ofplasticizers and drying temperatures on water vapour permeability of HP-films (determined at 3OoC,0-100%RH). 3.3 Mechanical Properties of HP-Film 3.3.I Effect of Solvent Composition. Tensile strength is the maximum stress that a film can sustain. Elongation is the maximum change in length of the test specimen before breaking.Tensile strength is of importance for two reasons. First, it was not possible to measure moisture permeability if the film fell apart during the experiment. Second, the successful use of a film that is formed on the surface of a food seems to require that the film has some degree of structural intergrity. Tensile strength and elongation can be used to describe how the mechanical properties of film materials relate to their chemical structures.

The tensile strength and elongation values of HP-films dried at 4OoC are shown in Table 2. Films prepared using a 70:30water-ethanol solvent displayed the largest tensile strength whereas largest percent elongation was shown by films prepared with 80:20 water-ethanol solvent. A relatively low concentration of ethanol in the film solution could promote hydrogen bonding between hydroxypropyl molecules, thus strengthening the matrix and making it more extensible. The poor tensile strength and elongation of films prepared fiom water alone and 50:50 water-ethanol solvent is probably due to stress fiactures caused by HP solids that did not dissolve in these solvents. 3.3.2 Effects of Plasticizers and Drying Temperature.Figures 2 and 3 show the effect of plasticizers on tensile strength and elongation values of HP-films dried at 4OoC and 6OoC, respectively. As plasticizer concentration increased, tensile strength decreased significantly for both sorbitol- and glycerol-plasticized films, as compared to control (no plasticizer) film. On the other hand, percent elongation increased significantlyas the

Gums and Stabilisers f o r the Food Industry 10

446

Table 2 Tensile Strength and Percent Elongation of Hydroxypropyl Films Prepared From Solvents of Various Compositions Solvent composition, water:ethanol(Y~v/v) 1oo:o 90:lO 80:20 70:30 60:40 50:50 a

Tensile strengthqb (ma) 27.5 f 2.3 29.1 f 2.8 29.4 f 3.2 34.8 f 2.6 28.6 f 3.2 25.0 f 2.5

Elongationasb(%) 3.8 f 0.4 4.0 f 0.8 4.0 f 0.3 3.9 f 0.5 3.1 f 0.7 2.3 f 0.3

Films were prepared at 4OOC; film thickness 0.075 f 0.005 mm Means of seven replicates f 95% confidence interval

20

Concentration of plasticizers (% w/w)

Figure 2 Effect ofplasticizers on the tensile strength (TS)andpercent elongation(E) of HP-films dried at40"C (Averagefilm thickness 0.0754 0.001 mm) plasticizer concentration was increased. These effects are true at both drying temperatures (Figures 2 and 3). Sorbitol films exhibited marginally higher tensile strength than glycerol film at all plasticizer levels. Conversely, glycerol films exhibited higher elongation values than sorbitol-plasticized films. This observation suggests that glycerol has a higher plasticizing effect than sorbitol, thus making the film more extensible. On the basis of these two parameters, sorbitol was a more effective plasticizer than glycerol in HP-films with regard to tensile strength and water vapour permeability.

Recent Developments, Fuiure Trends

447

35

H

30

%

z5

25

20

9)

;15 10

5

0 5 10 15 20 Concentration of plasticizers (% w/w)

I

mGlycerol (TS) OSorbitol (TS)4 O l y c e r o l (E)-&So&itol

(E)]

Figure 3 Effect ofplasticizers on the tensile strength (TS)andpercent elongation(E) of HP-Jlms dried at 60°C(AverageJlm thickness 0.075* 0.001 mm). Tensile strength of HP-films dried at 4OoC and 6OoCis comparable (Figures 2 and 3).

This observation is consistent with the finding reported by Reading and Spring32.They found no significant difference between ultimate tensile strength and percent elongation of methylcellulose films prepared fiom a 100% water solvent and dried at either 2OoC or 6OOC. In contrast, percent elongation for films dried at 6OoC exhibited lower values than that dried at 4OOC. The reason for this is not clear; it could be attributed partly to volatilization of plasticizers at higher drying concentration which resulted in reduced effective concentrationof the plasticizers in the dried films.

3.4 Hydroxypropyl Rice Starch-g-Poly(Acry1ic Acid) (HP-g-PAA) Graft 3.4.I FTZR spectroscopy. Evidence for the graft co-polymerizationbetween starch and acrylic acid is presented in Figure 4. Grafting of poly (acrylic acid) onto hydroxypropyl rice starch was testified by FTIR analysis whereby the existence of peak at the wavelength 1715 cm-' indicated the presence of carbonyl group in the HP-g-PAA composite. The spectra of the acrylic acid (not shown) and the graft copolymer show peaks characteristic of carbohydrate systems in the range 1000-1 150 cm-'. Mared analysis has also been used to prove grafting by making use of variation in the intensity of the -OH absorption band that may occur because of &mg. It was found that the intensity of the OH band in the spectrum of HP-g-PAA is less than in the spectrum of pure acrylic acid, indicating that some of the grafted chains are linked through the hydroxyl group.

448

G u m and Stabilisersfor the Food Industry 10

5.00 4.5 4.0

3.5

A

3.0 2.5

2.0 1.5

I .o 0.5

Figure 4 Inffared absorption spectra for (A) HP-film and (B) HP-g-PAA film In a study of interactions between starch and poly(acry1ic acid) (PAA), Fanta et a1.15 presented evidence that the ability of the two polymers to form uniform blends might be due to formation of a complex similar to known helical inclusion complexes between starches and fatty acids. A significant percentage of PAA cannot be solvent-extractedfrom starch, whereas PAA is readily extractable from mixtures with dextran. This suggests formation of complex between PAA and starch, but not with dextran, presumably because the a-(1-6) linkage between glucose units in dextran does not permit formation of the required helical structure.

3.4.2 Water Vapour Permeability. The effect of acrylic acid grafting on the WVP of HP-g-PAA film is presented in Figure 5. It can be seen that WVP increased almost linearly with concentration of acrylic acid. The WVP values between control (no acrylic acid) film and film incorporated with 7% and 10% (w/w) acrylic acid were significantly different (p < 0.05). This is to be expected because the presence of carboxyl groups in acrylic acid increases the hydrophilicity of the film, hence permeability to water vapour. Acrylic acid is liquid at room temperature with a melting point of 12-13°C. According to Kester and F e ~ m e m a a~ ~compound , with low melting point has low moisture barrier properties. In addition, in the presence of more hydrophilic group, the sorption of water molecules is facilitated, thus weakening the moisture resistance. Grafting of starch with acrylic acid, therefore, does not seem to be suitable for applications which require relatively low water vapour permeability. 3.4.3 Mechanical Properties. Figure 6 shows the effects of acrylic acid on tensile strength and percent elongation of the HP-g-PAA film. It is interesting to note that tensile

Recent Developments, Future Trends

1.40

449

I

I

Concentration of acrylic acid (% w/w)

I

~~~

Figure 5 Effect of acrylic acid concentration ("/"w/wdry starch) on water permeability of HP-film [RH gradient 0-100%;film thickness 0.091f 0.022 mm)

1 50 45

40

35 30 25

20 15

10 5 0 Concentration of acrylic acid

I

Tensile strength -I) elongation

I

Figure 6 Effect of acrylic acid on the tensile strength (TS) and percent elongation(E) of HP-jlms e l m s dried at 40°C; averagefilm thickness 0.065* 0.005 mm).

450

Gums and Stabilisersfor the Food Industry 10

strength of the film was highest at 2.5% (w/w) acrylic acid and significantly higher than the control film (no acrylic acid). Increase in acrylic acid concentration showed no further increase in tensile strength. The reason for this is not clear; a plausible explanation could be attributed to the effect of grafting on the cohesion and density of the polymeric network as well as the continuity of the film matrix. In other words, at 2.5% concentration of acrylic acid, optimum grafting occurred on all possible active sites on the starch backbone resulting in maximum cohesive stength of the film. It should be noted that cohesion and adhesion are related to the polymer structure and chemistry, i.e. to the molecular weight, regularity of chain structure, branching, polarity and distribution of polar groups along the polymer chain. Cohesion and rigidity of films are favoured by a high chain order’. Percent elongation of HP-g-PAA film exhibited consistent linear increase as the concentration of acrylic acid was increased. However, the difference in extensibility of the films at different concentrations of acrylic acid were not statistically significant (p > 0.05). In a related study by Shogren et al.”, FTIR spectra of complexes of both PAA with amylose and amylopectin suggest that the carboxyl groups of PAA are weakly hydrogen bonded to the polysaccharide hydroxyls rather than them forming more stable hydrogenbonded carboxyl dimers. This could probably account for the greater extensibility of the films in the presence of higher levels of acrylic acid. 4 CONCLUSIONS

Our study demonstrated that mechanical and moisture barrier properties of hydroxypropyl rice film is affected to a different extent by solvent composition and the presence of plasticizers. Grafting of hydroxypropyl starch with acrylic acid leads to further alteration in both mechanical and moisture barrier properties, some of which desirable or undesirable, depending on intended application. References S. Gilbert, ‘Food Packaging and Preservation’, ed. M. Mathlouthi, Elsevier, Essex, England, 1986, p.371. 2. J.J. Kester and 0. Fennema, J. Food Sci.,1986,40, 47. 3. J.M. Krochta and C.D. Mulder-Johnston, Food Technology, 1977,51(2), 61. 4. J.C. Rankin, LA.Wolff, H.A. Davis, and C.E. Rist, Industr. Engr. Chem., 1958,3, 120. 5. LA.Wolf, H.A. Davis, J.E. Cluskey, L.J. Gundrum and C.E. Rist, Zndustr. Engr. Chem., 1951,43,915. 6. A.M. Mark, W.B.Roth, C.L. Mehltretter and C.E.Rist, Food Tech., 1966,20,75. 7. W.B. Roth and C.L. Mehltretter, Food Technol., 1967,21,72. 8. J.L. Willet, B.K.Jasberg and C.L. Swanson, ‘Polymers from Agricultural Coproducts’, eds. M.L. Fishman, R.B. Friedman and S.J. Huang, American Chemical Society, Washington, DC., 1994, p. 50. 9. Jokay, L., Nelson, G.E. and Powell, E.L. Food Technology, 1967,21(8), 1064. 10. M.T. Vijayakumar, C.R. Reddy and K.T. Joseph, Europan Polymer Journal, 1985, 21(4), 415. 1.

Recenr Developments, Fuhue Trends

45 1

11. Z. Reyes, M.G. Syz and M.L. Huggins, Journal ofApplied Polymer Science: Part C , 1968,23,401. 12. F.H. Otey, R.P.Westhoff and C.R. Russel Znd. Eng. Chem. Prod. Res. Dev., 1977, 16,305. 13. F.H. Otey, R.P. Westhoff and W.M. Doane, Znd. Eng. Chem. Prod. Res. Dev., 1980, 19,592. 14. F.H. Otey, R.P. Westhoff and W.M.Doane, Znd. Eng. Chem. Res., 1987,26, 1659. 15. G.F. Fanta, C.L. Swanson and W.M. Doane, Journal ofApplied Polymer Science, 1990,40,811. 16. R.L. Shogren, R.V. Greene and Y.V. Wu, J. Appl. Polym. Sci., 1991,42,1701. 17. R.L. Shogren, A.R. Thompson, R.V. Greene, S.H. Gordon and G. Cote, J. Appl. Polym. Sci., 1991,47,2279. 18. B.M.N.M. Azemi and M. Wooton, StarcWStarke, 1984,36,274. 19. D.P. Johnson, Analytical Chemistry, 1969,41(6), 859. 20. A. Bayazeed, F.E. Okieman and O.B. Said, Starch, 1989,41(6), 233. 21. A. Hebeish, M.H. Khalil El-Rafie, A. Higazy and M.A.Ramadan, Starch, 1992, 40(3), 104. 22. ASTM, Annual Book of ASTM Standards Vol. 8.01, American Society for Testing

and Materials, Philadelphia. 23. M.E. Aulton, M.H. Abdul-Razzak and J.E. Hogan, Drug Dev. Znd. Pharm., 1981, 7(6), 649. 24. R. Sakellariou, R.C. Rowe and E.F.T. White, Znt. J. Pharm., 1986,31, 55. 25. G.S. Banker, J. Pharm. Sci., 1966,55,81. 26. I.G. Donhowe and O.Fennema, J. Food Proc. Preserv., 1993,17,247. 27. N. Gontard, S. Guilbert and J.-L. Cuq, J. FoodSci., 1993,58,206. 28. T.H. McHugh and J.M. Krochta, J. Agric. Food Chem., 1994,42(4), 841. 29. C.A. Entwistle and R.C. Rowe, J. Pharm. Pharmacol., 1979,31,269. 30. S. Porter, Pharm. Technol., 1980,3,67. 31. E.R. Lieberman and S.G. Gilbert, J.Polym. Sci., 1973,41,33. 32. S . Reading and M. Spring, J. Pharm. Pharmacol., 1984,36,421. 33. J.J. Kester and O.R. Fennema, J. FoodSci., 1989,54(6), 1390.

SELECTIVITY OF VARIOUS PECTINS TO HEAVY METAL IONS

M.T. Kartel, L. A. Kupchik and I. G. Levchenko

Institute for Sorption and Problems of Endoecology Kyiv 03 164 Ukraine

1 INTRODUCTION

The problem of protecting organisms from environmental technogenic contamination (pesticides, chlorine-organic compounds, medium - and long-living radioactive nuclides, heavy metals, etc.) becomes still more and more critical for the Ukraine'. Among detoxifying drugs designed to remove harmfkl substances from organisms, special attention should be given to pectins, as harmless food supplements or drug components. Thus, a number of publications24 recommend using pure pectins or pectin-containing substances as adsorption preparations, capturing ions of heavy metals. There is however little quantitative data available for the ion exchange capacity of pectins, which hinders development of correct dosages and application tactics for natural detoxification drugs in prophylaxis and curative practice. It is stipulated by a number of the reasons. First of all, there are technologies of pectin production, which involve using various raw materials (citrus fruit and apple peel, sugar-beet roots etc.). Secondly, the selectivity of adsorption is determined by the concentration of heavy metal ions in solutions, as well as containing of Na', K', Mf, and Ca2+ions which contents are, as a rule, 2-4 orders higher than target ions. The present article states outcomes of comparative evaluation of ion exchange capability for three pectin varieties of a commercial production in relation to the most typical ions of heavy metals (Ni, Co, Cu, Zn, Cd, and Pb). 2 EXPERIMENTAL Aqueous solutions of NiC12, CoCb CuSO4, ZnS04, CdClz and Pb(N03)~of various concentrations from 0.1 to 10 mMol I-' were prepared based on standard salt solution (Ringer's solution). 0.5 g of pectin was introduced into 100 ml of the solution under examination; the suspension was mixed for 4 h, then filtered and separated. Concentration of ions examined in the liquid phase was determined by atomic-absorption spectroscopy (Automatic Analyzer crVariann, USA). Pectins used were from: apples (Bara Pectin Factory, Ukraine), beets (Krasnodar Pectin Factory, Russia) and citrus (((B.K.M. Services Ltd.)), UK). Some experiments were made on special pectin-containing mixtures with clay mineral (palygorskite) in the ratio 1:l. Th~smixture as a curative substance (named ((PECTOPAL))pharmaceutics) and is tested in medical practice in the Ukraine as an

453

Recent Developments, Future Trends

adsorption drug for removing radioactive nuclides and ions of heavy metals from patients (suffering metal poisoning, occupation diseases, e t ~ . ) ~ . Selectivity of pectins and preparation ((PECTOPAL)) on metal ions was evaluated by m1g-l) which significance of complexing capacity (C, %) and distribution coefficient (b, were determined from relations:

c = (1 - c ~ c ~ ) . 1 0 0 ,

Kd = (CdC, -l).V/m,

where Ci,, and C, are accordingly initial and equilibrium concentration of metal ions in a solution, V is the volume of such solution, m is the mass of pectin or (
Significance of complexing capacity for various pectins and preparation ((PECTOPAL)) with heavy metals ions are presented in Table 1. The results infer that the complexing capacity of pectins in contact with concentrated salt solutions (10 mMol 1-') is rather low and, as a rule, (except for cokalt) increases for less than 50%. The capacity essentially increases for more diluted solutions (1 mMol l-I). It is especially characteristic for lead, as well as for copper and nickel to a certain extent. The complexing ability of other metals (cobalt, zinc and cadmium) on decreasing concentration remains practically at the same level or even lower6. In the case of a combined preparatiotnPWTQPALn, the complexing ability only for copper and zinc is less than 50% in solutions with initial concentration 10 mMol f'. Diluted solutions show the capacity for all ions, which sharply increases and comes up to 69% (nickel) and to 91% (lead). It infers that the clay component of a preparation ((PECTOPAL, promotes the capture ability of pectin. Shown below are values of distribution coefficients Kd for ((pectin + solution of heavy metal ions)) under standard conditions (C, = 1 mMol 1-) and for concentrations, appropriate to the physiological norm of ions of mentioned metals in humans: (see Table 2).

Table 1 Complexing Cupacity of Pectins and ((PECTOPALIon Ions of Heavy Metals Ion of Initial Metal Concentrution (C,,,)/ mMol t ' Ni2+ 10 1 co2+ 10 1 cu2+ 10 1

Zn2+

10

Cd2+

10

Pb2+

10

1 1

1

Complexing Capacity (C)/ % Apple Pectin Beet Pectin Citrus Pectin PECTOPAL 38.5 48.0 50.5 48.5 36.0 73.0 31.0 23.0 48.0 16.5 44.5 85.5

43.0 53.5 52.0 48.5 45.5 83.5 30.5 22.0 40.5 48.5 32.0 95.5

39.0 52.5 68.0 50.5 30.5 68.0 30.5 24.0 22.5 23.5 38.5 88.5

50.0 69.0 54.5 70.0 38.0 70.0 45.5 75.0 74.5 77.0 79.5 91.0

454

Gums and Stabilisersfor the Food Industry I0

Table 2 Distribution Coefficientsfor Pectins and ((PECTOPALuon Ions of Heavy Metals Ion of Equilibrium Concentration Metal ( c ~ J /mMol t’ Ni” 1 1o-2 1 co2+ 1o-2 1 cu2+ 1o-2 Zn’+ 1 10-’ Cd2+ 1 lo4 Pb’+ 1 10”

Distribution Coeflcient (Kd)/ ml g-’ Apple Pectin Beet Pectin Citrus Pectin PECTOPAL

1.8.1 O2 2.0.102 4.1*102 3.1J@ 3.0*10’ 2.9.ld 0.711O2 0.3*102 0.8*10’
1.6.1 O2 1.2.10’ 2.2*102 2.6.10’ 4.4.10’ 4.7.l@ 0.6*102 0.211o2 1.5.102 0.2.102 4.0*102 5.9.104

2.01102 3.2*102 3.5*102 8.7.102 2.6.10’ 3.1.10’ 0.71lo2 0.4.102 0.611O2
5.0.1 O2 1.7Jd 4.5.1 0’ 1.3.ld 6.2*102 4.2.l@ 3.9*102 8.2*102 6.1 10’ 1.9.ld 6.5*102 6.4.104

Notably, there is no single opinion on the mechanism of capture ability of pectin. This may be qualitatively explained by formation of pectates (compounds of metal ions with hydroxylic groups of a polysaccharide matrix (oligomer of galactose) and carboxylic groups of galacturonic acid (product of polysaccharide oxidation). The compounds of the pectate type may apparently have both an ion exchange and complexing nature. Capture ability of these substances depends on the pectin origin (raw material, engineering, etc.), nature of ions and conditions of interaction (concentration and pH of solution, its multiplicity, presence of macro-cations of alkaline and alkalineearth metals, soluble chalate compounds, etc.). The analysis of the obtained results allows us to compare capture ability of pectin of various types on ions of heavy metals8. Considering the distribution coefficients for standard conditions (Ceq= 1 mMol 1-’), which actually are an affinity to metal ion, it is clear that these values for pectin of various types do not essentially differ. Maximal values of & are marked for apple pectin (A) on Co’+, for beets pectin (B) on Cu” and Cd2+,and for citrus pectin (C) on Ni”, Zn’+ and Pb2+.An order of pectin selectivity to metal ions can be presented as:

Under conditions of physiological ion concentration, the distribution coefficients for various pectins essentially differ from those for standard levels. The highest selectivity of pectins is noticed for Pb2+,Cu2’ and Co’+, and is much lower for Ni2+.In the case of Zn” and Cd”, no selectivity was found. The main pectin for Pb” and Cu2+complexation is beet pectin, for Co2+is apple pectin, and for Ni2+is citrus pectin. Thus, an order of selectivity is following: Pb’+(B>C>A)>> Cu’+(BX>A) > Co’+(A>C>B)> Ni2+(C>A>B)>>

-

>> Zn2+(C-A-B) Cd’+(B>A-C).

Recenr Developments, Future T r e d

455

In case of PECTOPAL, the selectivity of adsorption concerning ions of heavy metals undergoes substantial modifications. At first, the order of ion affinity of pectin included in structure of a combined preparation, changes: cadmium is third for adsorption efficiency, and nickel and cobalt change places. Selectivity rows for standard conditions and for level of physiological concentrations of the ions in question, are identical: Pb > Cu> Cd> Ni > Co > Zn. Notably, in absolute values, the adsorption efficiency of heavy metal ions by pectin in the structure of a preparation ((PECTOPAL))is noticeably higher. For want of it in standard conditions, the distribution coefficients for all ions are quite close and are at the level of lo3 ml g-'. Comparing the increase of Kd for pectin with pure pectin in a preparation ((PECTOPAL)),they are essentially higher, and in a number of cases they achieve 10times and even more. So, for cadmium and zinc such increase achieves values, which allow characterizing pectins as well-capturing adsorbents. The auestion of influence of a clay mineral - palygorskite on capture ability of pectin deserves special attention (see Figure 1).

.

lg K,

b3 K d

DCitruspectin

0PECTOPAL

i 5

4

41 3-

3

2-

2

1-

1

I

,

Ni Co Cu i n Cd Pb

Ni Co Cu Zn Cd Pb

Figure 1 Diagrams of K d valuesfor pure citrus pectin and citrus pectin in PECTOPAL. on various ions of heavy metals: a) standard condition (Ceq= 1 mMol t9, b) physiological concentrations of ions. Pure palygorskite has noticeable selectiviy of adsorption in relation to ions of heavy metals (increase of Kd achieves 30-50 ml g- ). Its role in a preparation is to ensure binding

456

Gums and Stabilisers for the Food Industry 10

and removing of radioactive isotopes of caesium from a solution5,where & exceeds lo4 ml g-’. The positive influence of palygorskite on pectin can be reduced to its ability to precipitate complexes of metals derived from the soluble part of pectin, which are oligomers of a medium-molecular mass. Such a coagulation effect is stipulated apparently by colloidchemical properties of the clay particles, which can adsorb these soluble complexes. The phenomenon can be explained by their special electric charge performances and ability to create local pH shifts of a medium (from acid to neutral area) on the interface of microheterogeneous aqueous suspension. The results obtained indicate that the application of pure pectins with the purpose of demetalization of liquid media including the liquid medium of organisms, is apparently not alwaysjustified, as the part of pectin in a aqueous medium results in a true solution and correspondingly creates metal soluble complexes. Use of pectin-containing products is therefore more expedient, for example specially treated apple and citrus peel, beet cake etc., and also, as it was shown in the t h s article, in special combination with clay minerals. For such products and compositions the expected selectivity of adsorption should be much higher than for the pure pectins. 4 CONCLUSIONS

The results enable us to rationally use pectins of various origin as food components and drugs intended for removal of metals from organism poisoned with ions of heavy metals. The capture ability of pectins for heavy metals ions can be essentially increased by their use in combination with some clay mineral, in particular palygorskite. Notably, beet and apple pectins, which can be manufactured in the Ukraine at the industrial scale, have, similarly to citrus pectin, an exclusively high capture ability for heavy metal ions, especially for very toxic lead and cadmium.

References 1. National Report of Ukraine at the UN Conference on Environment and Development.

Brasil-92, Eurasia-Monitoring, 1992, 7,2. 2. G. Kostenko et al, Food and Process Industry (I%)1992,1,22. , 3. B. D. Levchenko and L. M. Timonova, ‘Pectin. Pectin Prophilaxy’ (Rus), Znanie, Krasnodar, 1992. 4. I. M. Trachtenberg, Physician Business (Rus),1992,423. 5 . V. V. Strelko and N. T. Kartel, ‘Scientific Principles of Drugs Elaboration’ (Ukr), Osnova, Kharkiv, 1998, p. 490. 6. L. Kupchik, M. Kartel and B. Veisov, Foodand Process Industry (Ukr), 1998,427. 7. Yu. V. Khmelevsky and 0. K. Usatenko, ‘Basic Biochemical Constants of Human Organism in Norm and Pathology’ (Rus), Zdorov’ya, Kiev, 1987. 8. M. Kartel, L. Kupchik and B. Veisov, Chemosphere, 1999,38,2591.

List of Participants Albeek GJMW Van

Wageningen Agricultural University, Dept of Food Technology, Bomenweg 2 6703 HD, Wageningen, The Netherlands

Aune 1

Nobipol ,Dept of Biotech, NTNU, N-7034 Trondheim, Norway

Aymard P

Lu-Centre Jean Theves, 6 Rue E Vaillant, 91207 Athis-Mons France

Baddii F

School of Biological Sciences, University of Surrey, Guildford Surrey, UK

Bakker M A E

Unilever, Olivier.'v.Noortlaan 120,3 133 AT Vlaardmgen The Netherlands

Bara Herczeg 0

Jozef Attila University, College of Food Industry, H-6724 Szeged Marster 7, Hungary

Barber G A

17 Harvey Avenue, Nantwich, Cheshire, CW5 6LE, UK

Ben Zion 0

Hebrew University of Jerusalem, PO Box 12, Rehovot 76100, Israel

Benhura M

Dept of Biochemistry, University of Zimbabwe, PO Box MP167 Mount Pleasant, Harare, Zimbabwe

Bixler H J

Ingredients Solutions Inc, USA

Bjerrum K

Danisco Biotechnology, Langebrogade 1, PO Box 17, 1001 KBH Copenhagen, Denmark

Boekema R

Hercules BV, PO Box 252,3770 AG Bameveld, The Netherlands

Boerboom F

Avebe B V, Avebeweg I, (3607 PT Foxhol, The Netherlands

Boulenguer P

SKW Biosystems, Direction Systems Texturants, Centre de Recherches, 50500 Baupte, France

Brooks M

Quest International, Kilngleary, Carrigaline,Co Cork, Ireland

Brown B

British Sugar, Norwich Science Park,Colney, Norwich, UK

castro s

Universidade de Aveiro, Dept Quimica, 3810 Aveiro, Portugal

Chouard G

SIAS - MPA,I7 Av du 8 Mai 1945,77298 Mitry Mori, France

Chung K H

Nabisco Inc, Techical Center, 200 DeForest Ave, East Hanover New Jersey 07936-194, USA

Clark A H

Unilever Research, Colworth House, Shambrook, Bedfordshire MK44 ILQ, UK

Cooper G

Cerester UK Ltd, Trafford Park, Manchester, MI7 IPA, UK

458

Gums and Stabilisersfor the Food Industry 10

Cowburn P

Danisco Ingredients, Dennington Road, Wellingborough, Northants, 8" 245, UK

Cruttenden N

Kelco Alginates, Waterfield, Tadworth, Surrey, KT20 5HQ, UK

Cui S

Southern Crop Protection and Food Research Centre 43 Mc Gilvray Street, c/o University of Guelph,Guelph Ontario, Canada

Daas PJH

Wageningen Agricultural University, Bomenweg 2,6703 HD Wageningen, The Netherlands

De Bont P

Unilever Research, PO Box 114,3130 Vlaardingen, The Netherlands

De Lorgeril C

Centre de Recherche Paul Pascal, Av Dr Schweitzer, 33600 Pessac France

De Sousa I

University Technicia Lisbon, Tapada da Aguda, 1379-017 Lisbon Portugal

Debon S

Glasgow Caledonian University, School of Biological Sciences(Food Sci), Cowcaddens Road, Glasgow, UK

Demeersman M

Dera Food Technology, Rijksweg 16, B 2880 Bornem, Belgium

Doublier J - L

INRA-IPCM, BP 71627,44316 Nantes Cedex 3, France

Draget K I

Nobipol, Dept of Biotechnology, NTNU, N-7491, Trondheim Norway

Duckworth G E

Copenhagen Pectin A S, Ved Banen 16, DK 4623 Lille Skensved Denmark

Dunstan D

Dept of Chemical Engineering, University of Melbourne, Parkville 3052, Australia

Edwards WP

Bardfield Consultants, 14 Durham Close, Great Bardfield, Essex, UK

El-Fak A M

Dept of Food Science, Faculty of Agriculture, Zagazig University Egypt

Else A J

Purac Biochem, PO Box 21,4200 Gorinchem, The Netherlands

Ennis M P

Dept of Food Chemistry, University College of Cork, Ireland

Erickson J

Nestle R & D, 809 Collins Ave. Margsville, OH 43040 - 4002, USA

Fabri D

Unilever Research, Colworth House, Shambrook, Bedfordshire MK44 ILQ, UK

Fernandes P

Friskies R & D, BP 47, 80800 Aubigny, France

Fernandez 0

Ceamsa, Poligono Industrial 'Las Gandaras, 76418 Atios, Porrino (Pontevedra), Spain

Foegeding E A

Dept of Food Science, NC State University, Raleigh, NC 27695-7624 USA

459

List of Participants

Foster T

Unilever Research, Colworth House, Shambrook, Bedfordshire MK44 ILQ

Fullard J

Carbohydrate Research Foundation, Eisenhowerlaan 150,25 17 The Hague, The Netherlands

Gilsenan P

Biopolymers Group, Division of Life Sciences, Kings College London Campden Hill Road,London, W8 7AH, UK

Goron E

Meyhall AG Rhodia Food, Sonnenwigensmse 18, CH 8280, Kreuzlingen, Switzerland

Goycoolea F M

Ciad, AC, PO Box 1735, Hermosillo Sonora, 83000, Mexico

Gregory D

David Gregory Associates, The Old School, Denston, Newmarket, Suffolk, CB8 2UJ UK

Gregson C

Dept of Food Science, University of Nottingham, Sutton Bonnington Loughborough,LEI2 IRD, UK

Grondal J

Copenhagen Pectin NS, DK-4623, Lille Skensved, Denmark

Hahn I

FriskiesBJestle R & D Centre, 3916 Pettis Road, St Joesph, M 0 64503, USA

Hani N M

Food Technology Division, University Sains Malaysia, 11800 Pulau Malaysia

Haug I

Nobipol, Dept of Biotechnology, NTNU, Trondheim, N-749 1 Trondheim, Norway

Helgerud T

Pronova Biopolymers AS, PO Box 494, N-3002 Drammen, Norway

Hemar Y

Hannah Research Institute, Ayr, KA6 SHL, UK

Hendrikx P

National Starch & Chemical, hestbury Court, Greencourts Business Park, 333 Styal Road, Manchester, M22 5LW, UK

Hermasson A-M

SIK Swedish Institute of Food Research, Institute of Food Research Postbox 5401,40229 Goteborg, Sweden

Hess S

Germantown International Ltd, 200 Lawrence Drive, West Chester PA 19380,USA

Hodgson I

Kelco Alignates, Water field, Tadworth, Surrey, KTZO 5HQ, UK

Hmgh L

Danisco Ingredients, Edwin Rahrs Vej 38,8220 Braband, Denmark

Hoffman R

Zestec, Eisenhowerlaan 150,25 17 KP The Hague, The Netherlands

Hopman A

Hercules BV, PO Box 252,3770 AG Barneveld, The Netherlands

Horvath-Almassy K

Jozef Attila University, College of Food Industry Szeged, H-6725 Szeged, Marster 7, Hungary

460

Gums and Stabilisers for the Food Industry 10

Howell N K

School of Biological Sciences, University of Surrey, Guildford, Surrey, UK

Hughes H

NEWI, Mold Road, Wrexham, LLl1 2AW, UK

Ikeda S

Dept of Food Science, North Carolina State University, Raleigh NC 27695-7624, USA

Imeson A

FMC Corporation (UK) Ltd, Unit 3c, Harcourt Way, Meridan Business Park, Leicester, LE3 2WP, UK

Jenkins D

Procter and Gamble, Rusham Park Tech Park, Whitehall Lane, Egham, Surrey, TW20 9NW

Jevne G

Pronova Biopolymer AS, PO Box 494, N-3002 Drammen, Norway

Joergensen J F

Copenhagen Pectin A S, Ved Banen 16, DK 4623 Lille Skensved, Denmark

Karamallah K A

Dept of Food Science and Technology, Faculty of Agriculture, University of Khartoum, Sudan

Karg C

MRIC, The North East Wales Institute, Mold Road, Wrexham, LLI 1 2AW, UK

Kaspasis S

Dept of Food Science and Nutrition, College of Agriculture, Sultan Qaboos University, PO Box 34 Al-Khod 123, Sultanate of Oman

Kravtchenko T

Collids Naturels International, BP 415 1, F-76723, Rouen Cedex, France

Kruif C G De

Nizo Food Research, PO Box 20,6710 BA EDE, The Netherlands

Langendorff V

SKW Biosystems, Direction Systems Texturants, Centre de Recherche, 50500 Baupte, France

Larsen H

Copenhagen Pectin A S, Ved Banen 16, DK 4623 Lille Skiensved, Denmark

Le Bihan C

Best Foods, Generade Condinetaire, Rue Charles Fourier, 59760 Grande Synthe, France

Ledward D

Dept of Food Science, University of Reading, Reading, RG6 2AD, UK

Leisner D

Facultad de Quimica, Universidad de Santiago, Av das Ciencias, E15706, Spain

Lillford P J

Unilever Research, Colworth House, Shernbrook, Bedfordshire, MK44 ILQ

Limberg G

Danisco Biotechnology, Langertrogade, DK 1001 Copenhagen K, Denmark

Lloyd D

Cerester UK Ltd, Trafford Park, Manchester, M 17 1PA

461

List of Participants

Loisel C

Enitiaa, Rue de la Geraudiere, BG 82225,44322 Nates Cedex, France

Lundburg E

Nestle R + D, Box 520,26725 Bjuv, Sweden

Lundin L

Unilever Research, Colworth House, Shambrook, Bedfordshire, MK44 ILQ

Madsen F

Danisco Ingredients, Edwin Rahrs Vej 38, DK-8220 Brabrand, Denmark

MarrBU

Copenhagen Pectin A/S, DK 4000 Roskilde, Denmark

Marrs M

Leatherhead Food RA, Randles Road, Leatherhead, Surrey, KT22 7RY

Mattes F

Herbstreith & Fox KG, Turnstrasse 37,75305 Neuenburg, Germany

May C D

Jandrell Cottage, Wellington, Hereford, HR4 8AX

Michon C

ENSIA, Biophysics Laboratory, 1 Avenue des Olympiades,91744 Massy Cedex, France

Michoud L

Kelco Alginates, Waterfield, Tadworth, Surrey, KT20 5HQ

Mitchell J R

Dept of Food Science, University of Nottingham, Sutton Bonnington Campus, Lougborough, Leicestershire, LEI2 5RD

Miyoshi E

Division of Development and Environmental Studies, Osaka University of Foreign Studies, Minoo City, Osaka 562-8558, Japan

Mombarg E

Cerestar, Havenstraat 84, 1800 Vilvorde, Belgium

Moriwaka K

Research Laboratory, Taito Co Ltd, Nagata, Kobe, Japan

Morley R

Delphi Consultants Services Inc, 948 Cabot Court, Stone Mountain, Georgia 30083, USA

Morris E R

Cranfield University, SILSOE, Bedfordshire, MK45 4DR, UK

Morris G

NCMH Unit, University of Nottingham, Sutton Bonningham, Loghborough, Leics, LEI2 5 R D

Morris V J

Institute of Food Research, Norwich Science Park, Colney, Norwich NR4 7AU, UK

Morrison N

Kelco Biopolymers, 8220 Aero Drive, San Diego, USA

Motlagh S

Baha’i World Centre, 16 Golomb Avenue, Haifa 3 1001, Israel

Muhrbeck P

Orlka Foods Research Unit, SE 24181, Eslov, Sweden

Murphy P

National Starch and Chemical, Prestbury Court, Greencourts Business Park, 333 Styal Road, Manchester, M22 5LW, UK

Murray J C F

Hercules Ltd, 3 1 London Street, Reigate, Surrey, RH2 9YA

462

Gums and Stabilisers f o r the Food Industry 10

Mussa N

School of Biotechnology, University of Surrey, Guildford, Surrey, UK

Ndoni S

Copenhagen Pectin A/S, DK-4623 Lille Skensved, Denmark

Neilson H

Roskilde University, Inst I, HUS 18, DK-7000 Roskilde, Denmark

Neubeck M

Jungbunzlauer, Penhofer I, A-2064, Wulzerhofen, Austria

Nishinari K

Dept of Food and Nutrition, Faculty of Human Life Science, Osaka City University, Jumiyoshi, Osaka 558, Japan

Normand V

Unilever Research, Colworth House, Shambrook, Bedfordshire, MK44 ILQ, UK

Norsker M

Dept of Biotechnology, Technical University of Denmark, 2800 Lyngby, Denmark

Norton I

Unilever Research, Colworth House, Shambrook, Bedfordshire, MK44 ILQ, UK

Nussinovitch A

Institute of Biochemistry and Food Science, Hebrew University of Jerusalem, PO Box 12, Rehovet 76100, Israel

Oakenfill D

Food Science Australia, PO Box 224, Wahroonga, NSW 2076, Australia

Ong M

Yorkreco, Nestec York, PO Box 204, York, YO1 IXY

Onsoyen E

Pronova Biopolymers, PO Box 494, N 3002 Drammen, Norway

Osman M E

The Gum Arabic Co, Khartoum, Sudan

Ould Eleya M

Dairy Research Centre, Local 1316 Pavillion Comtois, Universite Laval, Quebec, G 1K 1PU, Canada

Paterson J

Dept Food Science and Technology, University of New South Wales, Sydney, NSW 2052, Australia

Paterson L

Dupont (UK) Ltd, Cereals Innovation, Block B The Mill Site, 40 Station Road, CBI 2UJ, UK

Pelletier E

MRIC, The North East Wales Institute, Mold Road, Wrexham, LLI 1 2AW, UK

Phillips G 0

2 Plymouth Drive, Radyr, Cardiff, CF4 8BL, UK

Philp K

Quest International, Kilngleary, Carringaline, Co Cork, Ireland

Pilman-Willers E

Orkla Foods AS, C/O Procordia Foods, SE 24181, Eslov, Sweden

Pinto G L

Centro de lnvestigacions en Quimica de 10s Productos Naturales, Univesidad del Zuila, Maraccaibo, Venezuela

Poon S

Cooperative Research Centre for Industrial Plant Biopolymers, School of Botany, University of Melbourne, Parkville 3052, Australia

Lisr of Participants

463

Puaud M

Dept of Food Science, University of Nottingham, Sutton Bonnington, Loughborough,LEI2 5RD, UK

Raising F

Dept Process Technology, ATO-DLO, PO Box 17, Wageningen, Netherlands

Ramesh M

School of Biological Sciences, Sutton Bonnington Campus, Loughborough,Leicestershire, LEI2 5RD

Ravines P

Bahai World Centre, PO Box 155, Haifa, Israel

Rayment P

Biopolymers Group, Kings College London, Dept of Life Sciences, Campden Hill, London, W8 7AH

Raymundo A

Universidade Technica De Lisboa, Lab Femra Lapa, Tupada Da Ajunda, 1399 Lisboa, Portugal

Richardson P

National Starch at Chemical Company, 10 Finderne Avenue, Bridgewater, NJ 08807, USA

Robijn G W

Snow Brand E R L, Zenikerpark 6,97 47 Groningen, The Netherlands

SakullelarasmiP

Purified Agar Co Ltd, 888 Songway Road, Bangkok 10100, Thailand

Salama M F

4 El Nabatat St, Apt No 12, Garden City Cairo, Egypt

Sanderson G

Kelco Alginates, 8220 Aero Drive, San Diego, USA

Sasaki T

National Agricultural Research Centre, Ministry of Agriculture, 3-11 Kannosdai, Tsukuba, Ibaraki 305-8666, Japan

Savage R

University of Bradford, Dept of Chemical Engineering, West Yorkshire, BD7 IDP, UK

Schmidt P

G C Hahn & Co Ltd, Maes Gwern, Mold, CH7 lXW, UK

Schols H A

Agricultural University, Food Science Group, Bomenweg 2,6703 hd Wageningen, Netherlands

Schorsch C

Unilever Research Colworth, Shambrook, Bedford, MK44 ILQ

Sharpe V

Rhodia Food UK, Pollacre Lane, Woodley, Stockport, SK6 I P Q

Sondergaard K

Danisco Inmgredients, Edwin Rahrs Vej 38, DK 8220 Braband, Denmark

Spiers C

Pedigree Masterfoods, Mill Street, Melton Mowbray, Leics, LEI3 IBB, UK Kelco Alginates, Water field, Tadworth, Surrey, KT20 SHQ, UK Technical Research Centre, Gunma University, Kiryu, Gunma 37685 15, Japan

Sworn G Takigami S Tenkanen M

VTT Biotechnology and Food Research, Po Box 1500, FM-02044, Espoo, Finland

TOK-M

Kelco Biopolymers, Waterfield, Tadworth, Surrey, KT20 5HQ, UK

464

Gums and Stabilisers for the Food Industry 10

Tongdang T

Dept of Food Science, University of Nottingham, Sutton Bonnington Loughborough, LE 12 1RD

Tucker G

School of Biological Science, University of Nottingham Loughborough, Leics, LEI2 5 R D

Turgeon S L

Universite Lava], Des Sciences De'L Agriculture L'Aldermentation Dept des Sciences alfrocents nutrition, Quebec, Canada, G l K 7P4

Tziboula A

Hannah Research Institute, Ayr, Scotland, KA6 5HL, UK

Van Aken G

NlZO Food Research, PO Box 20,6710 BA Ede, The Netherlands

Vesterinen E

V I T Biotechnology and Food Research, PO Box1500, Fin 02044 VTT, Espoo, Finland, UK

Viebke C

MRIC, The North East Wales Institute, Mold Road, Wrexham, LLI 1 2AW. UK

Vreeker R

Unilever Research, PO Box 114,3130 AC Viaardingen, The Netherlands

-

Southern Crop Protection and Food Research Centre, 43 McGilvary Street, University of Guelph, Guelph, Ontario, Canada, N l G 2W1 Ward FM

TIC Gums, Belcamp, Maryland, 21013, USA

Wareing M

Arthur Branwell and Co Ltd, 58 - 62 High Street, Epping, Essex CM 16 4AE. UK

Watson A

Institute of Food Research, Colney Lane, Norwich, NR4 7UA, UK

Weilinga W

Meyhall AG, Sonnenwregenstrasse 18, CH 8280, Kreuzlingen, Switzerland

Wheeler G D

Mertyn Downing, Whitford, Holywell, Flintshire, CH8 9EP, UK

White R

Optokem Instrements, Nercwys, Nr Mold, Flintshire

Wichmann J

Danisco Ingredients, Edwin Rahrs Vej 38, DK 8220 Braband, Denmark

Wicker L

Dept of Food Science, University of Georgia, Athens, GA 306 02 76 10. USA

Williams M A K

Unilever Research, Colworth House, Shambrook, Bedfordshire, MK44 ILQ, UK

Williams P A

MRIC, The North East Wales Institute, Mold Road, Wrexham, LLI 1 2AW, UK ETH, Institute of Food Science, Zurich, Switzerland

Windhab E J Wolf B Wulansari R

Unilever Research, Colworth House, Shambrook, Bedfordshire, MK44 ILQ, UK Dept of Food Sciences, School of Biological Science, University of Nottingham, Sutton Bonnington Campus, Loughborough, LEI2 SRD

465

List of Participants

Yaron A

The Institute of Applied Bioscience, Ben Gurion University, PO Box 653, Beer Sheva 84105, Israel

Yuno-Ohta N

Junior College At Mishima, Nihon University,BunkyoCho, Mishima City, Shizuka 41 1 8555, Japan

Zanardi E

Inst. Science and Technology Dept Alimenti, Universita di Parma, Via del Taglio 8,43100 Parma, Italy