Proquest Dissertations

STUDY OF SYNERGISTIC INTERACTIONS BETWEEN GALACTOMANNANS AND NON-PECTIC

POLYSACCHARIDES FROM WATER SOLUBLE YELLOW MUSTARD MUCILAGE

A Thesis

Presented to

The Faculty of Graduate Studies

The University of Guelph

YINGWU

In partial fulfilment of requirements

for the degree of

Doctor of Philosophy

December, 2009

© Ying Wu, 2009 Library and Archives Bibliotheque et 1*1 Canada Archives Canada Published Heritage Direction du Branch Patrimoine de I'edition 395 Wellington Street 395, rue Wellington Ottawa ON K1A 0N4 Ottawa ON K1A 0N4 Canada Canada

Your file Votre reference ISBN: 978-0-494-64480-5 Our file Notre reference ISBN: 978-0-494-64480-5

NOTICE: AVIS:

The author has granted a non L'auteur a accorde une licence non exclusive exclusive license allowing Library and permettant a la Bibliotheque et Archives Archives Canada to reproduce, Canada de reproduire, publier, archiver, publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public communicate to the public by par telecommunication ou par I'lnternet, prefer, telecommunication or on the Internet, distribuer et vendre des theses partout dans le loan, distribute and sell theses monde, a des fins commerciales ou autres, sur worldwide, for commercial or non support microforme, papier, electronique et/ou commercial purposes, in microform, autres formats. paper, electronic and/or any other formats.

The author retains copyright L'auteur conserve la propriete du droit d'auteur ownership and moral rights in this et des droits moraux qui protege cette these. Ni thesis. Neither the thesis nor la these ni des extraits substantiels de celle-ci substantial extracts from it may be ne doivent etre imprimes ou autrement printed or otherwise reproduced reproduits sans son autorisation. without the author's permission.

In compliance with the Canadian Conformement a la loi canadienne sur la Privacy Act some supporting forms protection de la vie privee, quelques may have been removed from this formulaires secondaires ont ete enleves de thesis. cette these.

While these forms may be included Bien que ces formulaires aient inclus dans in the document page count, their la pagination, il n'y aura aucun contenu removal does not represent any loss manquant. of content from the thesis.

••• Canada ABSTRACT

STUDY OF SYNERGISTIC INTERACTIONS BETWEEN GALACTOMANNANS AND NON-

PECTIC POLYSACCHARIDES FROM WATER SOLUBLE YELLOW MUSTRD MUCILAGE

Ying Wu Advisors: University of Guelph Dr. Steve Cui 2009 Dr. Douglas Goff

This thesis investigated the synergistic interactions between galactomannans and non-pectic polysaccharides from water soluble yellow mustard mucilage. A non-pectic polysaccharide (NPP) fraction was isolated from yellow mustard mucilage using pectinase hydrolysis followed by ammonium sulphate precipitation. This fraction consisted mainly of a [3-1,4 linked glucosidic backbone. Rheological tests showed that

NPP exhibited strong shear-thinning flow behaviour and a weak gel structure, acid resistance, and stable gelling properties in a wide temperature range. Detailed structure information on NPP was investigated using Nuclear Magnetic Resonance (NMR) spectroscopy.

Four types of galactomannans (GMs), namely fenugreek gum (FG), guar gum

(GG), tara gum (TG) and locust bean gum (LBG), were investigated for their emulsification and rheological properties. The M/G ratios of the four GMs were 1.2, 1.7,

3.0 and 3.7 respectively. The results revealed that the M/G ratio, along with molecular weight and intrinsic viscosity, played an essential role in emulsion and rheological properties. Synergistic interactions between the four GMs and NPP were investigated using rheological measurements. The four types of GMs were blended with NPP at various ratios. Results revealed that at a total polysaccharide concentration of 0.5% (w/w), the highest synergism occurred at the GM/NPP blending ratio of 3/7 for all four types of

GMs. The interaction between TG and NPP showed the highest synergy, followed by

LBG/NPP, FG/NPP and GG/NPP. At a higher total polysaccharide concentration (1.0% w/w), the blend of TG and NPP exhibited the highest synergy, followed by FG/NPP,

LBG/NPP and GG/NPP. At a total polysaccharide concentration of 0.5% and the

GM/NPP blending ratio of 3/7, neutral pH (pH=6.5) showed the strongest synergy

compared to that at pH 2.0 and pH 12.0.

In order to better understand the synergistic behaviour, conformations of

simulated GMs with different M/G ratios were investigated using molecular modelling

software (Insight II /Discover_3 and RIS program, Version 4.0.0). The results showed

that the insertion of galactosyl groups could cause bending of the chains. The results

could explain the synergistic interactions between GMs and cellulosic polysaccharides: a

more flexible chain could penetrate through networks in a rigid structure, while the side

groups can help with forming stronger "hyperentanglements". The overall synergistic

behaviour of GMs and NPP is a result of a combined effect of junction zones via "smooth

regions" and "hyperentanglements". ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to my advisor Dr. Steve Cui for leading me into this polysaccharide world and providing me the opportunity to get involved in the Ph.D. project. I am very thankful for his guidance, advice, patience, support and help throughout the project.

I would also like to thank my co-advisor, Dr. Douglas Goff, and my committee members Dr. Michael Eskin and Dr. Rickey Yada for their advice, guidance and support throughout my Ph.D. study. I appreciate their sincere attitude towards science and prompt help whenever I need. I have learned a lot from my committee members and what I have learned from them can benefit me for the rest of my life.

I would also like to express my appreciation to my colleagues in Agriculture and

Agri-Food Canada for the helpful discussions and input during my study. My thanks also go to the faculty and staff in the Department of Food Science for their support and help during my study.

Special thanks to my parents, Guangtai Wu and Yulan Shi, for their consistent

support, love and patience during all the years of my Ph.D. study. Thanks to my two daughters, Abigail and Elizabeth, for their understanding, love and support during the years.

Finally, thanks to my friends and community for their consistent friendship, love and help.

i TABLES OF CONTENTS

1. GENERAL INTRODUCTION 1 1.1. Galactomannan-related studies on synergistic interactions with other biopolymers2 1.2. Previous studies on yellow mustard gum 6 1.3. Studies on synergistic interactions between GMs and yellow mustard mucilage ....8 1.4. Objectives of the current project 9 2. FRACTIONATION AND PARTIAL CHARACTERIZATION OF NON-PECTIC POLYSACCHARIDES FROM YELLOW MUSTARD MUCILAGE 11 2.1. Introduction 11 2.2. Materials and Methods 13 2.2.1. Materials 13 2.2.2. Pectinase hydrolysis 13 2.2.3. Precipitation with 20% (w/v) ammonium sulphate 14 2.2.4. Monosaccharide composition 15 2.2.5. Uronic acid determination 15 2.2.6. Protein determination 15 2.2.7. Methylation analysis 15 2.2.8. Rheological tests 16 2.2.9. Effects of acid and temperature 17 2.3. Results & Discussion 17 2.3.1. Isolation of non-pectic fraction from water soluble yellow mustard mucilagel7 2.3.2. Methylation analysis 22 2.3.3. Rheological properties 23 2.4. Conclusions 28 3. STRUCTURAL ELUCIDATION OF A NON-PECTIC POLYSACCHARIDE FROM WATER SOLUBLE YELLOW MUSTARD MUCILAGE BY NMR SPECTROSCOPY 29 3.1. Introduction 29 3.2. Materials and Methods 30

u 3.2.1. Sample preparation 30 3.2.2. Methylation of NPP 30 3.2.3. NMR spectroscopy 30 3.3. Results and Discussion 31 3.3.1. Methylation of NPP 31 3.3.2. Structural elucidation of NPP-Me from NMR spectra 32 3.4. Conclusions 40 4. AN INVESTIGATION OF FOUR COMMERCIAL GALACTOMANNANS ON THEIR EMULSION AND RHEOLOGICAL PROPERTIES 42 4.1. Introduction 42 4.2. Materials and Methods 43 4.2.1. Materials 43 4.2.2. Methods 44 4.3. Results and Discussion 46 4.3.1. Emulsion Properties 46 4.3.2. Rheological Properties 51 4.4. Conclusions 59 5. RHEOLOGICAL INVESTIGATION OF SYNERGISTIC INTERACTIONS BETWEEN GALACTOMANNANS AND NON-PECTIC POLYSACCHARIDE FRACTION FROM WATER SOLUBLE YELLOW MUSTARD MUCILAGE 60 5.1. Introduction 60 5.2. Materials and Methods 61 5.2.1. Materials 61 5.2.2. Methods 62 5.3. Results and Discussions 63 5.3.1. Effect ofGM/NPP blending ratio 63 5.3.2. Effect of M/G ratio 68 5.3.3. Effect of total polymer concentration 69 5.3.4. Effect of pH 72 5.4. Conclusions 73

in 6. A MOLECULAR MODELLING APPROACH TO UNDERSTAND CONFORMATION OF GALACTOMANNANS WITH DIFFERENT MANNOSE/GALACTOSE RATIOS 75 6.1. Introduction 75 6.2. Materials and Methods 77 6.2.1 .Materials 77 6.2.2. Molecular modelling methods 77 6.3. Results and Discussion 82 6.3.1. Adiabatic contour map of epimelibiose 82 6.3.2. Effect of intra-chain interactions on the conformation of GMs 85 6.3.3. Comparison of conformational characteristics of GMs with various M/G ratios 88 6.3.4. Effect of conformation on synergistic interactions with NPP 90 6.4. Conclusion 91 7. GENERAL CONCLUSION 93 REFERENCES 99

IV LIST OF TABLES

Table 2.1. Yield, uronic acid and monosaccharide contents in samples obtained by different precipitation methods after pectinase hydrolysis 20 Table 2.2. Monosaccharide profiles (%) of samples obtained by different precipitationmethods after pectinase hydrolysis 21 Table 3.1. Assignment of resonances of H and C spectra of methylated NPP 37 Table 4.1. Chemical compositions of galactomannans 49

Table 4.2. Intrinsic viscosity and molecular weight (Mv) of galactomannans 50 Table 5.1. Characteristics of the four galactomannans 62 Table 5.2. Storage and loss modulus (G' and G") of 0.5% GM/NPP mixed gel at blending ratio of 3/7, frequency of 0.8 Hz 65 Table 6.1. Simulated galactomannans with various M/G ratios 80 Table 6.2. Comparison of torsion angles at the lowest energy level from current study and other research groups 84 Table 6.3. Molecular simulated conformational properties of galactomannans derived from different parameter settings for RMMC calculation 86

v LIST OF FIGURES

Figure 2.1. Fractionation of non-pectic polysaccharides from water-soluble yellow mustard mucilage 18 Figure 2.2. Monosaccharide profiles of precipitants after pectinase hydrolysis 19 Figure 2.3. Methylation analysis on polysaccharides precipitated with 20 % ammonium sulphate 22 Figure 2.4. Flow curve of 1% (w/v) aqueous solution of non-pectic fraction 24 Figure 2.5. Frequency sweep on aqueous solutions of non-pectic fraction 24 Figure 2.6. Frequency sweep on aqueous solutions of non-pectic fraction at various acidic conditions 26 Figure 2.7. Temperature sweep on 0.5% (w/v) aqueous solution of non-pectic fraction. .27 Figure 3.1. 1H spectrum of NPP 33 Figure 3.2. 13C NMR spectrum of methylated NPP 35 Figure 3.3. Homonuclear shift correlated spectrum (COSY) of methylated NPP 35 Figure 3.4. Total Correlation Spectroscopy (TOCSY) of methylated NPP 36 Figure 3.5. H/C Heteronuclear Multiple-Quantum Coherence (HMQC) spectrum of methylated NPP 38 Figure 3.6. H/C Heteronuclear Multiple Bond Correlation (HMBC) spectrum of methylated NPP 39 Figure 3.7. Proposed structure of methylated NPP 41 Figure 4.1. Interfacial activities of galactomannans at various concentrations 47 Figure 4.2. Emulsion capacity and stability of galactomannans 49 Figure 4.3. Flow curves of tara gum at different concentrations 52

Figure 4.4. Master curves (log(r|sp)o - logC) of galactomannans 53 Figure 4.5. Power law index (n) of galactomannans at various concentrations 54 Figure 4.6. Flow curves of 0.5% galactomannan solutions 55 Figure 4.7. Frequency sweeps of galactomannans at different concentrations 57

VI Figure 5.1. Frequency sweep of GM/NPP mixtures at total polysaccharide concentration of 0.5% 64 Figure 5.2. Comparison of synergistic effects between galactomannan and non-pectic polysaccharides from yellow mustard mucilage with various blending ratios at total polymer concentration of 0.5%, frequency of 0.8 Hz 65 Figure 5.3. Effect of total polymer concentration on synergistic interactions between galactomannan and non-pectic polysaccharides from yellow mustard mucilage 71 Figure 5.4. pH effect on synergistic interaction between galactomannans and non-pectic polysaccharides 73 Figure 6.1. Schematic repretation of epimelibiose and torsion angles 79 Figure 6.2. Contour map of epimelibiose 84 Figure 6.3. Conformation of simulated galactomannans obtained from different parameter settings 87

vn 1. GENERAL INTRODUCTION

Polysaccharides, along with proteins, are the two most important functional ingredients in food colloids (Dickenson and Galazka, 1992). In recent years, polysaccharides have drawn increased attention due to health benefits (Wood, 2007), unique functionalities in the field of pharmaceutical (Rinaudo, 2006), food industries

(Williams, Annable, Phillips and Nishinari, 1994; McClements, 2009), cosmetics

(Baldrick, 2009) and even landscape and construction industries as natural binders

(Leemann and Winnefeld, 2007; Guo, Cui, Wang and Young, 2008). Polysaccharides can be from diverse sources, including plants, animals and microorganisms. Usually, polysaccharides in plants or microorganisms are storage materials, cell wall components, exudates and extracellular substances (Izydorczyk, Cui and Wang, 2005) while polysaccharides in animals are joint substances (hyaluronic acids) or energy storage material (glycogen) (Wildman, 2001).

Synergistic interactions between polysaccharides are commercially viable because it leads to the development of new textures (Morris, 1990, Chap. 8) and reduced concentration of material, hence cost saving. When two different biopolymers are mixed in solution, their behaviour will be determined by a delicate balance of forces (Hefford,

1984). Biopolymers are limitedly co-soluble even if they are slightly different in composition, structure and/or conformation (Tolstoguzov, 2002). Cooperative interactions are mostly observed between polymers with electrostatic interactions or H- bonds via smooth regions. Segregative associations usually occur between polymers if there is no attractive electrostatic drive (Harrington and Morris, 2009) or not enough El- bonds to cooperatively associate. Segregative associations could lead to spontaneous

1 separation of the mixture into two co-existing phases. Enhanced, decreased or unchanged gel strength can be observed due to the state of the two phases (Picout et al., 2000). If too excessive aggregations occur in response to segregative associations, the overall gel strength might decrease. When one phase penetrates in the other phase, somewhat involved in the network formation, then synergistic behaviour can be observed.

1.1. Galactomannan-related studies on synergistic interactions with other biopolymers

Among many of the plant polysaccharides, galactomannans (GMs) are one of the most extensively investigated and applied materials. GMs are naturally occurring in the seeds of some legumes and consist of P-1,4 linked D-mannose backbone with a-D- galactose branches attached to the C6 position. The mannose to galactose (M/G) ratio may vary from the sources of the material, e.g. typically GMs from fenugreek gum, guar gum, tara gum and locust bean gum have M/G ratios of 1/1, 2/1, 3/1 and 4/1, respectively. GMs do not only possess significant viscosity-enhancing effects in solution, but also exhibit synergistic interactions with other biopolymers such as polysaccharides, proteins and starches (Dea et al., 1977; Fernandes, 1994; Schorsch, Gamier and Doublier,

1997; Tarvares et al., 2005; Cui, Eskin, Wu and Ding, 2006). Dea et al. (1977) indicated that mixtures of GMs and non-gelling polymers can promote gel formation. GMs can interact synergistically with a number of polysaccharides, such as xanthan gum, agar and carrageenans, resulting in viscosity increases or gel formation (Dea & Rees, 1987;

Morris, 1990). It is also reported that the presence of GM may have significant effects on the formation and viscoelastic behaviour of the gels, depending on biopolymer concentration and ionic conditions (Tavares & Silva, 2003). The ability of GMs to

2 interact synergistically with other polymers is generally enhanced by decreasing their galactose content and the distribution pattern of galactose on the mannose backbone chains (Morris, 1990; Dea, Clark and McCleary, 1986). This kind of synergistic behaviour between polysaccharides is commercially valuable because it leads to the development of new textures, for instance, biphasic gel structures (Morris, 1990).

Sudhakar, Singhal and Kulkarni (1995) investigated the interaction of corn starch and the less explored waxy Amaranthus paniculatas starch at 5% w/v with the widely used

GMs, guar gum and locust bean gum, in the concentration range of 0-0.2% w/v, with respect to changes in paste viscosity and gelatinization temperature. They observed that as the concentration of gum increased, the gelatinization temperature decreased and the cold paste viscosity increased. A higher synergism with guar gum was also observed because of its greater hydration capacity as compared to locust bean gum. According to

Alloncle and Doublier (1991), starch pastes are described as suspensions of swollen particles dispersed in a macromolecular medium. It is suggested that GMs are located within the continuous phase, and thus the volume of the phase accessible to the GMs is reduced, which causes a dramatic increase in its concentration in the continuous medium, thereby resulting in a very high viscosity. The swollen particles are mainly composed of amylopectin, while the continuous medium consists of amylose. It is probably the amylose-GM interaction that is dominating in the system. It could be clearly seen from

Sudhakar et al.'s (1995) work that no sharp differences on pasting characteristics of guar gum and locust bean gum with starch, which is devoid of amylose compared to corn starch and gum combinations. From the same study, a better freeze-thaw stability of the blend was noticed, which could be due to the interaction of these gums with amylase,

3 thereby slowing down retrogradation. An increased stability of starch-GM combinations was also observed. Brennan et al. (1996) investigated the interaction between GM from guar and starch and discovered that the in vitro hydrolysis of starch in guar bread was attenuated significantly compared with normal wheat bread. Therefore, they drew a conclusion that guar GM acts as a physical "barrier" to a-amylose-starch interactions and/or subsequent release of hydrolyzed products (e.g. maltose).

Tavares et al. (2005) studied the synergistic effects between GMs and whey protein.

The influence of the degree of branching of GMs on the heat-induced gelation of whey proteins was investigated. GMs were from different origins and/or enzymatically modified, with M/G ratio varied from 1.5 to 3.7. Whey protein gels were formed at 13% protein, pH 7 and low ionic strength. GM concentration ranged from 0-0.6%. Within the range of the concentrations used, the presence of GM decreased the gelling temperature and had a positive effect on the gel strength. These effects are more obvious as the degree of branching decreases. The effect of the original guar sample was quite different from all the other samples, either in terms of rheology or microstructure, particularly for the higher GM concentration. The mixed gels appeared as biphasic systems, with the polysaccharide enriched phase dispersed in the protein matrix at low polysaccharide concentrations, but processing to a phase inversion at higher polysaccharide concentrations, especially for the lower branched samples. The linear viscoelasticity seemed to be insensitive to some of the microstructural changes observed within the mixed gels. The branching degree of the GM did have an effect on microstructure and viscoelasticity of the gels, but this effect was limited to a short range of M/G ratio, above which the effect was insignificant.

4 Synergistic interactions of xanthan gum-GM have been extensively investigated

(Schorsch et al., 1997; Fernandes, 1995). In Fenandes' study, GM samples of different origins (tara and locust bean gum) with M/G ratios from 2.8 to 5.0, had been characterized. Mixed gels of xanthan gum and GMs were compared by means of oscillatory shear measurements. For all GMs, a synergistic maximum was observed when the ratio of xanthan gum to GM was 1:1. However, the magnitude of these maximum varied with the GM sample used. It was found that the higher the M/G ratio of GM samples, the higher was the synergistic interaction. Schorsch et al. (1997) also investigated the rheological behaviour of xanthan/guar gum systems using oscillation shear and creep-recovery measurements. In their study, the total polysaccharide concentration was kept constant at 0.5% w/w, the xanthan/GM ratio ranged from 1/99 to

90/10 and three ionic strengths were studied. As for xanthan/GM mixtures, strong synergistic phenomena were observed. It is clear that the addition of xanthan gum even at a very low level to a guar gum solution induced a transition of the system from a macromolecular solution to a structured system displaying gel-like properties. The comparison between three guar gum samples with different molecular weights evidenced a strong effect of the molecular weight: the higher this parameter, the stronger the synergistic interaction. At a low xanthan/GM ratio, (< 10/90), xanthan/guar gum mixtures and xanthan/locust bean gum mixtures resulted in similar viscoelastic behaviour. In contrast, at a higher xanthan/GM ratio, a stronger synergism was exhibited with locust bean gum. The rheological properties of the mixed systems were greatly influenced by the presence of electrolyte. For example, in xanthan/guar gum systems, the storage (G') and the loss (G") moduli were increased when electrolyte was present, in contrast to

5 xanthan/locust bean gum mixtures which exhibited a reverse tendency. Similar results were found with xanthan/guar or xanthan /locust bean gum mixtures. The differences in mechanism may exist according to the M/G ratio of the GM and also to the xanthan/GM ratio and the ionic strength. As a conclusion, xanthan gum played a major role in the rheological behaviour of the xanthan /GM systems even at a low concentration. Also, the mannose/galactose ratio of the GM and xanthan/GM ratio and the ionic strength may yield some rheological differences.

1.2. Previous studies on yellow mustard gum

In some species of plants, including members of Brassicaceae, Solanaceae, Linaceae, and Plantaginaceae, the epidermal cells of the seed coat contain a large quantity of complex polysaccharides (mucilage), a property known as myxospermy (Van Caeseele,

Mills, Sumner and Gillespie, 1981; Van Caeseele, Kovacs and Gillespie, 1987;

Boesewinkel and Bouman, 1995). Although the biological role of mucilage in plant seeds is still unclear, it is thought to aid in the dispersal and/or protection of the emerging seedling during imbibition and germination. It has been reported that the major component of mucilage is pectin, an acidic polysaccharide that forms gels in the extracellular matrix and is present in all cell walls as well as mucilage (Western, Skinner and Haughn, 2000).

6 to less than 1% in the other mustard seeds (Cui, Eskin, Wu and Ding, 2006). A recent study on yellow mustard mucilage showed that the water soluble mucilage exhibited a potent anti-cancer effect in animal models for colon cancer (Eskin, Raju and Bird, 2007).

In previous studies, the structure of water-soluble mucilage was investigated using

CTAB (cetyltrimethylammonium bromide) precipitation followed by ion exchange chromatography (Cui, 1993; Cui, Eskin and Biliaderis, 1993). Two polysaccharide fractions were obtained using CTAB: CTAB-precipitate (WSCP) and CTAB-soluble

(WSCS) fractions with WSCP containing more uronic acids (23%) and WSCP containing less uronic acids (13%). Following CTAB treatment, each of the CTAB treated fractions were further fractionated into five sub-fractions by DEAE-high capacity cellulose ion exchange chromatography. Among the ten sub-fractions, three fractions exhibited shear- thinning flow behaviour. One of the three was a pectic polysaccharide, while the other two were non-pectic polysaccharides (NPP) mainly composed of a (3-1,4 glucosidic backbone with methyl and ethyl substitutions. Recent interest in the mucilage of yellow mustard seed is attributed to its unique rheological behaviour in solutions/dispersions and its ability to interact with GMs synergistically (Cui, Eskin, Biliaderis and Mazza, 1995;

Cui et al., 2006). Understanding the composition, structure, chemical and physiological properties of the mucilage is essential for better use of this agricultural by-product. In the past, studies carried out on crude mucilage examined the extraction methods, chemical composition, physical properties, rheological properties, and synergistic effects with other components including GMs, starches and functionalities in food systems (Cui et al.,

2006). Synergistic interactions between polysaccharides are commercially valuable because it leads to the development of new textures (Morris, 1990), reduced usage of

7 materials and lowered product cost. Synergism between yellow mustard mucilage and

GMs was first reported by Weber, Taillie and Stauffer (1974). Cui et al. (1995) further identified that the non-pectic polysaccharide (NPP) from water-soluble yellow mustard mucilage consisting of P-l,4-linked D-glucan backbone chain was the functional component exhibiting synergistic interactions with GM. NPP has a similar structure to xanthan gum, which interacts with GMs synergistically. Among the many mechanisms of polysaccharide synergistic interactions as summarized by Cui (2001) and Cui et al.

(2006), they proposed the "junction zone" model (cooperative binding) for the synergistic interactions between GM and NPP assuming that the unsubstituted mannan backbone can form junction zones with the smooth region of the P-1,4 glucosyl backbone. There are several factors that may affect the magnitude of synergy in mixed polysaccharide systems, including molecular weight and intrinsic viscosity (Fernandes, Goncalves and

Doublier, 1991), salt concentration (Khouryieh, Herald, Aramouni and Alavi, 2007;

Schorsch, Gamier and Doublier, 1997), M/G ratio and fine structures (Dea and Clark,

1986; Fernades, 1995), polysaccharide-polysaccharide blending ratio (Cui et al., 1995), total polysaccharide concentration (Mannion, Melia, Launay, Cuvelier, Hill, Harding and

Mitchell, 1992; Bresolin, Sander, Reicher, Sierakowski, Rinaudo and Ganter, 1997), temperature (Mannion et al., 1992; Zhan, Brownsey, Ridout and Morris, 1993) and pH

(Morris and Foster, 1994; Ross-Murphy, Morris and Morris, 1983; Whitney, Brigham,

Darke, Reid and Gidley, 1998; Tako, 1991). In recent years, molecular modelling has become a standard tool for determination of molecular structure and conformation of polysaccharides (Tvaroska and Taravel, 1991).

1.3. Studies on synergistic interactions between GMs and yellow mustard mucilage

8 Synergism between yellow mustard mucilage and GMs was first reported by Weber,

Taillie and Stauffer (1974). Cui et al. (1995) further identified that the non-pectic polysaccharide (NPP) from water-soluble yellow mustard mucilage consisting of p-1,4- linked D-glucan backbone chain was the functional component exhibiting synergistic interactions with GM. NPP has a similar backbone structure to xanthan gum, which interacts with GMs synergistically. Several models were developed to explain the mechanisms of the synergistic interactions between the two types of polysaccharides.

Among the many mechanisms of polysaccharide synergistic interactions as summarized by Dea et al, (1977), Cui (2001) and Cui et al. (2006), they proposed the "junction zone" model (cooperative binding) for the synergistic interactions between GM and NPP assuming that the unsubstituted mannan backbone can form junction zones with the smooth region of the (3-1,4 glucosyl backbone. However, this model alone cannot explain

GMs with fully or densely substituted back bone, such as fenugreek gum and guar gum.

Therefore, an in depth understanding of the structure, conformation and their impact in the mixed system is necessary for better exploring of the mechanism of synergy.

1.4. Objectives of the current project

The objectives of the present study are to understand the mechanism of synergistic interactions between GMs and cellulose-like polysaccharides from yellow mustard mucilage. The specific objectives of the present study are to: i. Isolate and characterize non-pectic polysaccharides (NPP) from water soluble yellow mustard mucilage; ii. Elucidate the fine structure of the isolated fraction using NMR spectroscopy;

9 iii. Investigate synergistic interaction between NPP and GMs with different M/G ratios using a rheometer; iv. Study the conformation of GMs with various M/G ratios using molecular modelling software and try to understand the impact of conformation on the synergistic interactions with NPP.

There are many techniques being employed to investigate the mechanism of synergistic interactions, including rheology, molecular modeling, NMR spectroscopy,

DSC, microscopy, X-ray etc. In the present study, NMR spectroscopy, molecular modelling and rheology are the major techniques applied to carry out the current project.

10 2*. FRACTIONATION AND PARTIAL CHARACTERIZATION OF NON-

PECTIC POLYSACCHARIDES FROM YELLOW MUSTARD MUCILAGE

2.1. Introduction

Mills, Sumner and Gillespie, 1981; Van Caeseele, Kovacs and Gillespie, 1987;

Yellow or white mustard (Sinapis alba L.) has drawn great attention in the mustard family by containing much higher amounts of mucilaginous materials in the seed coat than any other type of mustards, including brown mustard, black mustard and oriental mustard: mucilage accounts for 5% in yellow mustard compared to less than 1% in the other mustards (Cui, Eskin, Wu and Ding, 2006). Understanding the composition, structure, chemical and physiological properties of the mucilage is essential for better use of this agricultural by-product. In the past, studies carried out on crude mucilage

* Published as: Fractionation and Partial Characterization of Non-Pectic Polysaccharides from Yellow Mustard Mucilage. Food Hydrocolloids, 2009, 23, 1525-1541.

11 examined the extraction methods, chemical composition, physical properties, rheological properties, and synergistic effects with other components including galactomannans

(GMs), starches and functionalities in food systems (Cui et al., 2006). A recent study on yellow mustard mucilage showed that the water soluble mucilage exhibited potent anti cancer effect in animal models for colon cancer (Eskin, Raju and Bird, 2007).

In previous studies, the structure of water-soluble mucilage was investigated by investigating the material using CTAB (cetyltrimethyllammonium bromide) precipitation followed by ion exchange chromatography (Cui, 1993; Cui, Eskin and Biliaderis, 1993).

Two polysaccharide fractions were obtained using CTAB: CTAB-precipitant (WSCP) and CTAB-soluble (WSCS) fractions with WSCP containing more uronic acids (23%) and WSCP containing less uronic acids (13%). Following CTAB treatment, each of the

CTAB treated fractions were further fractionated into five sub-fractions by DEAE-high capacity cellulose ion exchange chromatography. Among the ten sun-fractions, three fractions exhibited shear-thinning flow behaviour. One of the three was a pectic polysaccharide, while the other two were polysaccharides mainly composed of a P-1,4 glucosidic backbone with methyl and ethyl substitutions. The fraction mainly composed of a P-1,4 glucosidic backbone was found responsible for the shear-thinning and synergistic effect with other polysaccharides such as GMs. Therefore, this fraction is of special interest in the present study.

In order to further investigate the structure and functional properties of these polysaccharides, it is necessary to develop a more direct and less tedious method for obtaining this fraction. It is well reported that mucilage usually contains pectic polysaccharides and neutral polysaccharides, and that the pectin could be easily removed

12 by using pectinase (Western et al., 2000). Pectinase is extensively used by the food industry. Examples of this given by Zandleven, Beldman, Bosveld, Schols and Voragen

(2006) include depectinization of fruit juices, maceration of vegetables and fruits, and extraction of vegetable oils. Kashyap, Vohra, Chopra and Tewari (2001) reviewed the applications of pectinase in commercial sectors and covered all types of pectinases in various product areas. Pectinase is also widely used for structural analysis of polysaccharides from different plant sources (Yumaguchi, Ota and Hatanaka, 1996;

Nakamura, Maeda and Corredig, 2006). After pectinase treatment, pectic polysaccharides could be degraded, while the non-pectic polysaccharides remain in the solution. In the present study, non-pectic polysaccharides were obtained using pectinase hydrolysis followed by ammonium sulphate precipitation. Chemical composition, linkage information and rheological properties of the isolated polymer(s) are reported.

2.2. Materials and Methods

2.2.1. Materials

Yellow mustard seeds were provided by G.S. Dunn Ltd. (Hamilton, Ontario, Canada).

Pectinase was purchased from Sigma-Aldrich (P4716, Pectinase from A. niger, Sigma,

Oakvale, Ontario, Canada)

2.2.2. Pectinase hydrolysis

Water soluble mucilage (WSM) was prepared according to the procedure described by Cui et al. (1993). 0.5% (w/v) WSM solution was prepared by dissolving the sample in

50 mM phosphate buffer with pH adjusted to 5.0 containing 0.02% sodium azide (w/v).

Sample solution was kept at 80°C for 2 h with constant stirring. After cooling to room

13 temperature, pectinase (Sigma, PI746) was added into the solution (150 unit /g WSM).

Digestion was carried out at room temperature (22°C) for 40 h with constant stirring.

After the digestion, the solution was kept at 90°C for 15 min to inactivate the enzyme.

After the solution was cooled to room temperature, the pH of the solution was adjusted to

7.0 with 0.5 N NaOH.

2.2.3. Precipitation with 20% (w/v) ammonium sulphate

A 50% ammonium sulphate solution ((NH^SC^) (w/v) was slowly added to the pectinase digest stepwise with constant stirring to achieve a final solution of 20% (w/v) ammonium sulphate. The solution was left at room temperature for 2 h and centrifuged at

8,000 g for 30 min. The supernatant was decanted and the residue re-suspended in distilled water at the ratio of lg WSM to 100 ml distilled water with the aid of a polytron mixer to help with the dispersion. The solution was centrifuged, as described above, for two more times and the final residue re-suspended in distilled water at the ratio of lg

WSM to 100 ml distilled water. The dialyzed solution was then transferred to dialysis tubing (15,000 Dalton Mw Cut-off) and dialyzed against distilled water for 24 h, with water changed every 4 h. Solution was transferred to a beaker and 1 N NaOH added to the solution to achieve a final concentration of 0.1 N NaOH. The sample was stirred continuously in a fume hood for 4 h at room temperature. This step was to exchange the

NH/ residue with Na+, which was difficult to remove by dialysis. Afterwards the sample solution was dialyzed again as described earlier. Following dialysis, the sample was precipitated in 70% ethanol, centrifuged at 8,000 g for 30 min, and then decanted. The final residue was re-suspended in 100% ethanol, dried in the fume hood, and stored in desiccators for further analysis.

14 In a parallel experiment, the pectinase digest was recovered by precipitation with 3 volumes of 100% ethanol followed by centrifugation at 8,000 g for 30 min at room temperature. The polytron was used at medium speed for 1 min to re-suspend the residue in distilled water. The solution was centrifuged again and this step was repeated twice.

The final residue was re-suspended in 100% ethanol and dried in a fume hood.

2.2.4. Monosaccharide composition

Monosaccharide composition was determined by dissolving samples in 1 mL 12 M

H2SO4 at 35°C for 30 min followed by diluting the solution to 1 M H2SO4, which was kept at 100°C for 2 h. The hydrolytes were diluted to appropriate concentration and analyzed by high performance anion exchange chromatography (HPAEC) as described by Wood, Weisz and Blackwell (1994).

2.2.5. Uronic acid determination

Uronic acid was determined by m-hydroxydiphenyl method developed by

Blumenkrantz and Asboe-Hansen (1973).

2.2.6. Protein determination

Protein content was analyzed using NA2100 Nitrogen and Protein Analyzer (Thermo

Quest, Milan, Italy).

2.2.7. Methylation analysis

Methylation analysis of the samples was carried out according to the method of

Ciucanu and Kerek (1984) with slight modification. The dried sample was dissolved in anhydrous DMSO at 85°C for 2 h with constant stirring, and sonicated for 4 h to ensure a

15 complete dissolution. Dry sodium hydroxide (20 mg) was added to the mixture, and stirred for 3 h at room temperature (22°C). The mixture was stirred for an additional 2.5 h after adding 0.3 mL methyl iodide. The methylated polysaccharide was then extracted with 1 mL methylene chloride. The methylene chloride extract was passed through a sodium sulphate column (0.5x15 cm) to remove water with the solvent evaporated by a stream of nitrogen. The dried methylated polysaccharide was hydrolyzed in 0.5 mL of

4.0 M trifluoroacetic acid (TFA) in a sealed test tube at 100°C for 6h and the TFA removed by evaporation under a stream of nitrogen and dissolved in 0.3 mL distilled water. The hydrolysate was reduced by sodium borodeuteride (1-5 mg) and acetylated with acetic anhydride (0.5 mL). Aliquots of the resultant partially methylated alditol acetates (PMAA) were injected in a GC-MS system (ThermoQuest Finnigan, San Diego,

CA) fitted with a SP-2330 (Supelco, Bellefonte, PA) column (30 m*0.25 mm, 0.2 \un film thickness, 160 - 210°C at 2°C /min, and then 210 - 240°C at 5°C/min) equipped with an ion trap MS detector.

2.2.8. Rheological tests

All rheological measurements were conducted on a strain controlled ARES

Rheometer (TA Instruments, New Castle, DE). All solutions were prepared by dispersing samples in solvent at room temperature for 1 hour with constant stirring followed by dissolving in 80°C water bath for another 1 hour with constant stirring, then cooled down to room temperature in 30 min. Two concentrations, 0.1% and 0.5% (w/v) respectively, were tested to investigate the effect of concentration. Viscosity was determined by cone- and-plate geometry (4°, 50 mm) with shear rate from 0.001-1000 s"1. A 1% (w/v) sample solution was prepared for this test. For viscoelastic tests, parallel-plate geometry (50 mm)

16 was used at gap size of 0.7 mm. Storage modulus (G'), loss modulus (G") and phase angle were determined by oscillatory test at frequencies from 0.008 to 16 Hz. Temperature was kept at 25°C for all tests except otherwise specified.

2.2.9. Effects of acid and temperature

Viscoelastic tests were performed to investigate the effects of acid and temperature.

For acid treatments, 50 mM of monopotassium phosphate buffer was adjusted to pH 2 with phosphoric acid to prepare 0.5% (w/v) sample solution at pH 2. For neutral pH condition, 0.5% (w/v) sample solution was prepared in 50 mM phosphate buffer with pH adjusted to 6.5. 0.4% (w/v) sample solution was also prepared in 1 M HC1. For this treatment, rheometer plate and geometry were wrapped with parafilm to avoid corrosion of the equipment. Temperature ramp test was performed with 0.5% (w/v) aqueous solution at frequency equals to 0.0625 Hz, temperature ranged from 5 °C to 90 °C.

2.3. Results & Discussion

2.3.1. Isolation of non-pectic fraction from water soluble yellow mustard mucilage

The procedure for isolating non-pectic polysaccharide from WSM is presented in

Figure 2.1. In order to evaluate the efficiency of this procedure, two parallel precipitation methods were applied to collect the residues of enzyme digests: ammonium sulphate precipitation versus 75% ethanol precipitation. Chemical and monosaccharide compositions of the samples from the two precipitation methods were compared and results are presented in Table 2.1 and 2.2 respectively.

17 Crude extract of water soluble mucilage

1+50 mM buffer pH 5.0, +0.02% NaN3 at 0.5% solid/buffer ratio

Pectinase hydrolysis (40 h at room temperature) tInactivat e at 90°C for 15 min, cool down, and adjust pH to 7.0 20% (NH4)2S04 precipitation (slowly add 50% (NH4)2S04) <-

Leave at room temperature for 2 h

Centrifugation (8000 g, 30 min)

Decant T Repeat 3 times Residue

Re-suspend sample in distilled water, assist dispersion with polytron

Dialysis in distilled water (15,000 Dalton molecular cut tubing)

Ichan ge water every 4 h for 24 h

Solution+ NaOH to 0.1 N

iKeep stirring at room temperature for 4 h

Dialysi)ialysis (15,000 Dalton molecular cut tubing)

Change water every 4 h for 24 h

70% Ethanol precipitation

Slowly add 3 volumes of 95% ethanol and leave at room temperature for 1 h

Centrifugation (8000 g, 30 min)

Decant

Residue (Air Dry)

Figure 2.1. Fractionation of non-pectic polysaccharides from water-soluble yellow mustard mucilage

After pectinase hydrolysis and ammonium sulphate precipitation, the yield of the non-

pectic polysaccharide fraction accounted for about 15.6% of the WSM compared to

18 32.4% using ethanol precipitation. Monosaccharide profiles from the two precipitation methods are presented in Figure 2.2.

25 ;

20 ~ Galactose Hlucose 15 1 u c 10 " Rhamnose Xylose 5 " A ia 1 A/y nnose ol ^ tiU

-50o 5.0 10° 15-0 20.0 25.0 30.0 35.0

Retention Time (min)

a. 75% ethanol precipitation 120. Glucose 100 -E

75i

Galactose 25" Rhamnosinpse Mannose 0 -3

-20 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Retention Time (min)

b.20% ammonium sulphate precipitation

Figure 2.2. Monosaccharide profiles of precipitants after pectinase hydrolysis

The 20%) (NH4)2S04 precipitant material was composed of 4.5% rhamnose, 7.5% xylose, 13.P/o mannose, 61.1% glucose, 13.8% galactose and 6.9% uronic acid, with

19 glucose to galactose ratio of 4.3; the ethanol precipitant contained 11.8% rhamnose, 4.5% xylose, 10.0% mannose, 38.3% glucose, 35.3% galactose, and 13.9% uronic acid, with glucose to galactose ratio of 1.1. Protein was not detected in the 20% (NFL^SC^ precipitant but was present at 1.7% in the ethanol precipitant.

Table 2.1. Yield, uronic acid and monosaccharide contents in samples obtained by

different precipitation methods after pectinase hydrolysis

Precipitation Method 20% (NH4)2S04 75% ethanol

Yield % 15.6 ±2.4 32.4 ±2.3

Uronic acid % 6.9 ± 0.3 13.9 ± 1.8

Protein % Not detected 1.7 ± 0.3

The above result indicated that 20% (NEL^SC^ precipitation can selectively precipitate polysaccharides in the enzyme digests. This selectivity may due to the solubility of different polymers in ammonium sulphate solutions. For example, it was reported that galactans were soluble in saturated (NH^SC^ solution (Izydorczyk, 2005).

Pectinase could degrade pectic polysaccharides and leave the short chains with branches composed of neutral saccharides remaining in the solution. If the chain is long enough, it will possibly precipitate in 75% ethanol and lead to the co-precipitation with the rest of the non-degraded polysaccharides. The precipitation with ammonium sulphate is based on the molecular charges of polysaccharides.

When ammonium anions NH4"1" surrounded the negatively charged molecular surfaces, it suppressed the negative charges, thus weakened the repulsive forces among molecules

20 which could lead to the aggregation and precipitation by Vander Waals attractive forces

(Jullien and Botet, 1987) and hydrogen bonding among polymer chains. Ammonium sulphate has been successfully applied for purifying P-glucans (Wang, Wood, Huang and

Cui, 2003; Li, Cui and Kakuda, 2006). Since the fraction isolated from the current study is mainly composed of P -1, 4 linked D-glucose, a similar approach was applied in the present study. Previous studies indicated that uronic acid and galactose were also detected in the non-pectic fractions although a non-destructive separation methodology, i.e. ion exchange chromatography was utilized (Cui, 1993). This result may indicate that a portion of the uronic acids and galactose were not associated with the pectins, but with the non-pectic fraction. Therefore, we consider that the isolated fraction in the current study is homogeneous after ammonium sulphate precipitation.

Table 2.2. Monosaccharide profiles (%) of samples obtained by different precipitation

methods after pectinase hydrolysis

Precipitation Method 20% (NH4)2S04 75% ethanol

Glucose% 61.1 38.3

Galactose% 13.8 35.3

Rhamnose% 4.5 11.8

Xylose% 7.5 4.5

Mannose% 13.1 10.0

Glucose/Galactose 4.3 1.1

21 2.3.2, Methylation analysis

The chromatogram after methylation analysis showed a major peak appeared at retention time of 26.16 min (Figure 2.3), which is attributed to the 1, 4 linked glucose.

100

I 80

1 60 RT: 26.16

c* 20 \J I | I M | I I I V| I IJ I | I I I | M I | I I I | I I I | I I I | M I | I I I | II I | IH I | fl I | I I I | I I I | I I I | " I I I I I | I I I | I I I | I I I | I I 10 20 30 40 Time (min)

43 100

8 sod

§ 60

99 233 "40 87 45 71 131 173 ,§20 1 1,18, JiiLll jll...J,,Jii 0i I ii I'I'TI r] I rf [Ti ("j'Ti'T"! fTi'i'n r7iTTiTiTTTri7TTi7nT,]TTTf"rn~|Triri|TTripr'rTirrTi"i 100 200 300 400 m/z

Figure 2.3. Methylation analysis on polysaccharides precipitated with 20 % ammonium sulphate

This is in agreement with our earlier results (Cui, 1993; Cui et al., 1993) that the main backbone of the non-pectic polysaccharide was P-l, 4 linked glucose. Many small peaks on GC chromatogram (Figure 2.3) indicated that there were some other sugar linkages in this fraction, which will be identified in further structural studies.

22 Although detailed information on structure is not available yet, methylation analysis from both current study and previous studies revealed that the non-pectic polysaccharide fraction is mainly composed of (3-1,4 linked D-glucose, with small amount of uronic acids and other monosaccharides, either present in the backbone, or as branches to the backbone. The presence of these monosaccharides broke the "regularity" of the P-1,4 glucose linkage and therefore preventing the polymer chains from forming large crystallites like cellulose. The unsubstituted P-1,4 linked glucose backbone is termed the smooth region or regular region. Junction zones can be formed via hydrogen-bonding at smooth regions and the irregular region can help with extending the molecules and form three dimensional (3 D) networks. If the molecules contain more smooth regions, it will create more opportunities for forming junction zones, and therefore easier to form gels.

2,3.3. Rheological properties

2.3.3.1. Flow behaviour

The 1% (w/v) solution of isolated fraction exhibited shear-thinning flow behaviour

(Figure 2.4). This result is in agreement with Cui et al.'s results on their non-pectic fractions (Cui, 1993; Cui et al., 1993). The linear region of the flow curve was fitted to the power law model rpKy""1, where k is the consistency index of the fluid and n is the flow behaviour index. In general, when concentration or molecular weight increases, n decreases and k increases. Usually at high concentrations, n falls in the range of 0.15 to

0.25. In the current study, the parameter n equals to 0.12, and k equals to 9.89, which indicated that this polymer has a high viscosity at low shear rates and strong shear thinning characteristics.

23 =10000 1000 (Pa - >> 100 O

IS C 10 > 3 1

Figure 2.4. Flow curve of 1% (w/v) aqueous solution of non-pectic fraction

2.3.3.2. Viscoelasticproperties

The mechanical spectra of 0.1% and 0.5% (w/v) of the fraction are presented in

Figure 2.5.

100 .• G* of 0.5% 10 a a G" of 0.5% Pi aDD • an a a •••o° G'of 0.1% o i - a a A*$So G" of 0.1% a 0.1 H /v ^ O O O o o v 0.01 0.001 0.01 0.1 1 10 100 Frequency (Hz)

Figure 2.5. Frequency sweep on aqueous solutions of non-pectic fraction

24 Even at a concentration as low as 0.1%, the solution behaved like a weak gel, with G' higher than G" and phase angle less than 45° across all the frequencies examined. At concentration of 0.5% (w/v), gel structure was obviously stronger, with G' consistently larger than G" and a relatively steady low phase angle regardless of frequencies.

As discussed in the previous section, due to the abundance of smooth regions, molecules can readily form junction zones, which lead to the weak gel structure at concentration as low as 0.1%. When the concentration increases, molecules are drawing closer to one another and junction zones are more readily formed. With the increased number of junction zones, a stronger gel structure was observed. Using computer simulation, it was reported that the minimum polymer concentration for gel formation decreased with increasing stiffness of polymer chains (Okada, Koga and Tanaka, 2002).

The strong gel structure at low concentrations may suggest that the polysaccharide has a stiffer chain. This report provided some insight for understanding the gelling behaviour of the isolated polymer, which indicated that the conformation of the polymer chain could be very stiff as it can form a weak gel structure in concentration as low as 0.1%.

2.3.3.3. Effect of acid

In earlier studies, it was observed that the non-pectic fraction of WSM was highly resistant to acid (unpublished data). In the present study, the effect of pH on rheological properties was investigated. Under acidic conditions, G' at pH 2 was consistently higher than G' at pH 6.5 (Figure 2.6).

25 This could be explained by the reduced repulsive forces after introducing H+ into the

solution while enhancing hydrogen bonding between polymer chains, which essentially

favours the formation of junction zones. With the stronger acid treatment, such as 1M

HC1, the solution was heated at 80°C during sample preparation; degradation could occur

due to the hydrolysis.

100 0.5%, pH2.0 b 0.4%, 1M HC1 xn A * i4A* * 0.5%, pH6.5

1O 10

o c3

CO 0.001 0.01 0.1 1 10 100 Frequency (Hz)

Figure 2.6. Frequency sweep on aqueous solutions of non-pectic fraction at various acidic conditions

However, it was observed that a stronger gel was formed compared to a more neutral

pH (pH = 6.5). The possible explanation is that the high concentration of acid may cause

a "salting-out" effect. Even though there might be some loose areas that acid could

penetrate and degrade the molecules, the majority of the structure was not affected since

the gel structure was not weakened, instead, it was strengthened. Further investigation

will be considered using alternative techniques such as microscopic techniques to

elucidate the microstructure of the gel under this extreme acidic environment.

26 2.3.3.4. Effect of temperature

Effect of temperature on the viscoelastic property of this isolated fraction was investigated (Figure 2.7). The 0.5% sample solution exhibited gel structure within the whole temperature range, with G' »G".

14 12 co io as % B PH ^ 5 O *5 4 b G" 2 Tan8

0 20 40 60 80 100 Temperature °C

Figure 2.7. Temperature sweep on 0.5% (w/v) aqueous solution of non-pectic fraction

The temperature ramp curve can be divided into three stages. The first stage is from

5°C to 40°C, where G' decreased slightly with increase in temperature. During the second stage from 40°C to 70°C, G' was steady while the temperature was increasing. During the third or final stage from 70°C to 90°C, there was a slight increase of G'.

When temperature was increasing from 5°C to 40°C, molecules obtained more energy and their movements became more vigorous. This might lead to the breakage of some weakly associated interactions, for example, some smaller junction zones with a less amount of hydrogen bonds involved. The loss of this part of associations was reflected by the decrease of G' on the temperature ramp curve. When the temperature continued to

27 increase from 40°C to 70°C, more associations would be destroyed; meanwhile, some associations of the molecule interactions, mainly hydrophobic interactions, could be formed due to the presence of methyl or ethyl groups in the backbone chain (Cui et al.,

1993). The steady G' value reflected that the dissociation and association of molecule interactions reached an equilibrium, therefore the overall structure was not affected; at the third stage, G' increased with the temperature from 70°C to 90°C, indicating that hydrophobic interactions dominated the thermodynamic process.

2.4. Conclusions

In the present study, the combination of pectinase hydrolysis and ammonium sulphate precipitation resulted in a non-pectic polysaccharide with more glucose and less uronic acid compared to ethanol precipitation. This fraction exhibited unique gelling characteristics under extreme acidic and at very high temperature conditions. These superior rheological characteristics suggest great potentials for the utilization of this material by food or non-food industries. It was found that NPP was difficult to be dissolved and it forms aggregations due to intermolecular interactions as demonstrated by forming weak gel structure at concentration as low as 0.1%. Several attempts have been tried in the current study to improve the solubility of this fraction, including dissolving

NPP in up to 1 M sodium hydroxide, DMSO (dimethyl sulfoxide), or urea solution up to

6 M, but no success was achieved. In the future study, modification of the molecules to improve the solubility of this fraction will be considered in order to determine its molecular weight using GPC (Gel Permeation Chromatography) or Light Scattering methods. More investigations will be carried out to further understand the structure and conformation of this material.

28 3. STRUCTURAL ELUCIDATION OF A NON-PECTIC POLYSACCHARIDE

FROM WATER SOLUBLE YELLOW MUSTARD MUCILAGE BY NMR

SPECTROSCOPY

3.1. Introduction

Non-pectic polysaccharide (NPP) from water soluble yellow mustard mucilage was isolated as described in our previous paper (Wu, Cui, Eskin & Goff, 2009a). This fraction exhibited unique physical, chemical and rheological properties such as synergistic interactions with GMs, acid resistance, and stable gelling properties in wide temperature ranges (Wu et al., 2009a; Wu, Cui, Eskin & Goff, 2009b). By using DEAE column, Cui,

Eskin and Biliaderis (1994) isolated non-pectic fractions composed of 40-60% glucose, along with galactose, rhamnose, mannose, xylose, and a small amount of uronic acids.

Methylation analysis and NMR spectroscopy revealed that this fraction consisted of a

1,4-linked P-D glucose backbone chain with occasional substitutions of ethyl and/or propyl groups at the 2, 3, or 6 positions of the glucose residue. However, it is still unclear how the other monosaccharide units, e.g. mannose and galactose, link to the main chain backbone. In the present study, the NPP fraction was obtained using an alternative approach (Wu et al., 2009a) and the methylation analysis of NPP confirmed similar chemical composition and the existence of 1,4-linked J3 -D glucose backbone chain as described by Cui et al. (1995). As demonstrated in our previous studies (Wu et al.

2009a), NPP can form a weak gel even at very low concentrations, e.g., 0.1%, which caused poor resolution on NMR spectra. Therefore, a method aimed at improving the resolution of NMR spectra of NPP was developed in the present study. NPP was methylated before being subjected to the NMR examination. The methylated NPP was

29 soluble in organic solvents and suitable for NMR analysis. The objective of the current study was to produce high quality NMR data which could lead to detailed structural information on NPP.

3.2. Materials and Methods

3.2.1. Sample preparation

NPP was isolated from water soluble yellow mustard mucilage as described by Wu et al. (2009a).

3.2.2. Methylation of NPP

The dried sample was dissolved in anhydrous DMSO (dimethyl sulfoxide) at 85°C for

2 h with constant stirring, and sonicated for 4 h to ensure a complete dissolution. Dry sodium hydroxide powder was added to the mixture and stirred for 3 h at room temperature (22°C). The mixture was stirred for an additional 2.5 h after adding methyl iodide (CI13I). The methylated polysaccharide was then extracted with methylene chloride

(CH2CI2). The methylene chloride extract was passed through a sodium sulphate column

(0.5x15 cm) to remove water. Methylated NPP was recovered by evaporating the solvent under a stream of nitrogen. The resulting residue was dissolved in deuterized methylene chloride (CD2CI2) at a concentration of 2% for NMR analysis.

3.2.3. NMR spectroscopy

NMR spectra were recorded on a Bruker AMX500 spectrometer. Tetramethylsilane was used as a chemical shift reference. Homonuclear correlation (COSY) spectra were recorded with F2 time domains of 1024 points and Fl time domains of 512 points. Total

30 Correlation Spectroscopy (TOCSY) spectra were recorded at mixing time of 100 msec and with F2 time domains of 1024 points and Fl domains of 512 points. Heteronuclear

Multiple-Quantum Coherence spectra (HMQC) were recorded with F2 time domain of

1024 points and Fl time domain of 256 points. Heteronuclear Multiple Bond Correlation spectra (HMBC) were recorded with F2 time domain of 1025 points and Fl time domain of 1024 points.

3.3. Results and Discussion

3.3.1. Methylation ofNPP

NPP can form a weak gel at a low concentration of 0.1% (Wu et al. 2009a). This produced difficulties in obtaining NMR spectra with good resolutions, since usually the concentration of the polysaccharide required is at least 2%. It is worth mentioning that the NPP fraction obtained by Cui et al. (1995) could be readily dissolved in water and the solution prepared for NMR was as high as 4%. We believe that the NPP investigated in the present study shared similar properties with that obtained by Cui et al. (1995), e.g. the

P-1,4 linked glucosidic backbone and chemical composition, however their molecular weight and linkage patterns might be very different because the isolation methods were different. The NPP in the current study may possess higher molecular weight and larger amount of unsubstituted region through which hydrogen bonding could be easily formed leading to the weak gel structure under low concentrations. NPP was dissolved in DMSO

(1%) and ID 'Fl spectrum was obtained (Figure 3.1b), however, the resonance from the

ID C spectra and 2D experiments were very weak. Therefore, one of the objectives of the current study was to prevent the formation of hydrogen bonds among the molecules

31 and thus improve solubility of NPP in an appropriate solvent. Methylation is widely used for modification of polysaccharides either for structural analysis (Cui, 2001) or for production purposes (Renard and Jarvis, 1999). In the present study, modified NPP exhibited better solubility in CD2CI2 and the solution concentration could be easily prepared up to 2%. Based on this solution, ID and 2D NMR experiments were performed with sound resolution (Figures 3.1-3.6).

3.3.2. Structural elucidation of NPP-Me from NMR spectra

The lH and 13C NMR spectra of NPP-Me are shown in Figure 3.1 and Figure 3.2 respectively.

Figure 3.1a and Figure 3.1b are !H NMR spectra of NPP. Figure 3.1a is the methylated NPP and 3.1b is the non-methylated NPP respectively. One strong signal was observed at 3.48 ppm on Figure la and was identified as a methyl group, which was not shown in Figure 3.1b. The ^-NMR spectrum (Figure 3.1a) contained 3 major anomeric protons, all appearing in the range of 4.2-4.4 ppm, and labelled as A at 4.38 ppm, B at

4.33 ppm and C at 4.30 ppm respectively. The coupling constant of the B and C anomeric protons are in the same range of 7.5-8 Hz, therefore, they were identified as 0- configurations (Iwata et al., 1992). Due to the overlapping of A with B, it is hard to distinguish the coupling constant of A. Some weak signals can also be observed from the

!H-NMR spectrum, at 5.0 ppm and 5.65 ppm respectively, indicating the existence of some a-configurations and uronic acids, however, the signals are too weak for the continued identification of these residues.

32 AV

\ ViL J^- 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppn

oppm

Figure 3.1. *H spectrum of NPP

(a: Methylated NPP; b: Not methylated NPP )

Monosaccharide analysis revealed that NPP was composed of 61% of glucose,

13.8% of galactose, 13% of mannose, 7.5% of xylose, 4.5% of rhamnose and 7% of uronic acid (Wu et al. 2009a). Among the three anomeric protons, the B proton yielded the strongest signal, indicating the abundance of this proton in the sample. Since D- glucose was the most abundant residue in NPP, therefore B was identified as HI from

33 glucose. The comparison of the observed H NMR chemical shifts of the three anomeric protons with those reported in literature (Hannuksela and Penhoat, 2004) allowed the identification of A as HI from (3-Man and C as HI from P-Gal.

The *H-NMR spectrum (Figures 3.1a & b) also showed some signals in the range of 1-3 ppm, with more signals appearing in Figure 3.1b than in Figure la. Cui et al.

(1995) mentioned that some signals appeared in the similar region, which they believed were attributed to -CH3 groups of the ethers. The authors believe that the strong alkaline condition during methylation process may break down some ether linkages, therefore it is reasonable to observe less signals appeared on the NPP-Me spectrum.

The ID 13C NMR spectrum is presented as Figure 3.2. The sugar carbon signals are distributed in the range of 60-110 ppm. Other than the sugar carbon signals, some resonances appeared in the range of 0-70 ppm. The strong signal at 54 ppm was identified as the solvent signal (CD2CI2). Another strong signal appearing at 60 ppm was attributed to the -OCH3 group. Other signals in the non-sugar carbon region might be attributed to the ethers originally substituted to the NPP, which is in agreement with previous observations (Cui et al. 1995).

34 CD2:CI' 2

^ Vk*v*K^L**w*M*AA*v*^^ •Arf^Wrtrt^

1 • i''• ' I ' • '•i ' • ' • i' •• • i ' •' •i ' • • • i ' •• • i' ' ' 'i' • • ' i '' '' I' " ' I' " 'I " '' I1 " ' I ' •• •i •' • ' i • ••' i •' •' i' •' ' i ' ' •'i •'' • 103 100 95 90 85 SO 75 70 65 SO 55 50 45 40 35 30 25 20 13 10 pptt

Figure 3.2. 13C NMR spectrum of methylated NPP

COSY (Figure 3.3) was performed to obtain the information on the connectivity of protons by following the connectivity pattern H1-H2-H3-H4-H5-H6. The complete proton assignment of each sugar residue was further confirmed by TOCSY (Figure 3.4).

0 4.8 4.6 4.4 4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.8 ppm

Figure 3.3. Homonuclear shift correlated spectrum (COSY) of methylated NPP.

35 Figure 3.4. Total Correlation Spectroscopy (TOCSY) of methylated NPP

From COSY, other than the three anomeric protons, the protons from each pyranose ring (H2 to H6) are overlapping in the region of 3.0-4.0 ppm, which produced difficulty for accurately assigning protons to each sugar residue. TOCSY can overcome this disadvantage. The protons belonging to one pyranose ring can be clearly observed and identified. According to the proton assignment, !H/13C correlation of NPP-Me (HMQC)

(Figure 3.5) gave the complete assignment of proton and carbon of the three monosaccharides correspondingly (Table 3.1).

HMBC spectrum showed long range connectivity of the lWi3C correlation and the linkage information was obtained as shown in Figure 3.6. The cross peak of H1/C4

36 (4.34 ppm/77.0 ppm) of glucose residue further confirmed the |3-1,4 linked glucose in large quantity; a resonance at 3.65 ppm/103.6.0 ppm (H6 of glucose/Cl of mannose) indicated that mannose was 1-6 linked to the glucose. A resonance at 2.96 ppm/103.0 ppm (H2 of glucose/Cl of galactose) indicated galactose was 1, 2 linked to the glucose backbone. There might exit some other kinds of linkages that the present study cannot reveal due to the ambiguous resonance signals.

Table 3.1. Assignment of resonances of !H and 13C spectra of methylated NPP

Residue H/C I 2 3 4 5 6a/6b

H 4.34 2.96 3.21 3.66 3.68 3.75/3.65 Glu C 103.2 84.0 85.0 77.0 75.5 71.0 H 4.38 3.04 3.17 3.37 3.71 3.98/3.66 Man C 103.6 83.5 80.5 81.2 69.5 75.0 H 4.30 2.92 3.13 3.62 3.86 3.91/3.88 Gal C 103 84.2 80.0 75.7 74.0 73.0

From the HMQC spectrum (Figure 3.5), the methyl group formed strong resonances

1 1-5 around 3.5 ppm ( H) and 60 ppm ( C), which indicated the abundance of this group. It is obviously seen from the HMBC spectrum (Figure 3.6) that the methyl groups are densely substituted at C2, C3 and C6 positions.

37 Figure 3.5. H/C Heteronuclear Multiple-Quantum Coherence (HMQC) spectrum of methylated NPP

38 ppm

i;:50 - 0 c 55-

60-

65-

80-

85 GlcHl/GlcC4

90- U95J- 100- 105- GlcH6/ManCl—•© 4rff-^C Q

In D- —i—•—i—•—i—•—i—•—r^—r—•—i—•—i—•—i— 4.4 4.2 4.0 3,8 3.5 3.4 3.2 3.0 2.B ppm

Figure 3.6. H/C Heteronuclear Multiple Bond Correlation (HMBC) spectrum of methylated NPP

In addition to the methyl group, ethyl group (-O-CH2-CH3) was also identified from

COSY, HMQC and HMBC spectra. A small portion of the ethyl group was detected substituted at the C3 position of the glucose residue and a large portion of this group had no link to the NPP sample. As revealed by Cui et al. (1995), the non-sugar signals in the range of 1-2.5 ppm on the *H spectrum were attributed to the -OCH2CH3 and the -OCH3 groups. The methylation process may break down most of the original methyl and ethyl

39 groups and replace with methyl groups from CH3I, with a small amount of ethyl group remaining in the chain. The broken down fragments, like ethyl groups, may remain in the sample. Therefore, the ethyl groups were detected with different chemical shifts due to the difference in attachment to the NPP chain. Cui et al. (1995) reported that the methyl groups were identified mostly attached to the C2 position and the ethyl groups mostly attached to the C3 position of the main chain. At least the present study confirmed the ethyl substitution at C3 position of the main chain. This can partially explain the superior emulsification behaviour of yellow mustard mucilage reported by Cui, Eskin and

Biliaderis (1993).

The dense substitution of methyl groups to the C2, C3 and C6 positions of the sugar residues improved the solubility of NPP in CD2CI2. Meanwhile, the replacement of originally substituted groups, e.g. methyl groups and ethyl groups, with the introduced methyl group interfered with understanding of the original substitution groups and patterns. Quantitive and qualitive substitution pattern of the methyl and ethyl groups should be investigated using an alternative approach.

3.4. Conclusions

Methylation of NPP was successfully applied in the current study to obtain high quality NMR spectra. After methylation, 1 D and 2 D NMR spectra were obtained. More detailed structural information on methylated NPP was derived based on the NMR spectra. The proposed NPP-Me structure is presented in Figure 3.7.

40 H H P*>

H H

Figure 3.7. Proposed structure of methylated NPP

41 4*. AN INVESTIGATION OF FOUR COMMERCIAL GALACTOMANNANS ON

THEIR EMULSION AND RHEOLOGICAL PROPERTIES

4.1. Introduction

Galactomannans (GMs), neutral naturally occurring polysaccharides in the seeds of some legumes, consist of a mannan backbone (P-(l—>4)-D-mannose) with a-D-galactose at C6. The mannose to galactose (M/G) ratio is dependent on the source of the GM.

Typically, fenugreek gum (FG) has an M/G ratio of about 1, guar gum (GG) has an M/G ratio of about 2, tara gum (TG) has an M/G ratio of about 3, and locust bean gum (LBG) presents an M/G ratio of about 4. M/G ratio may vary even in the seeds from the same species, which can be caused by varieties or environmental factors (Dea and Morrison,

1975). Because of their high water binding capacity, GMs can form highly viscous solutions even at low concentrations without forming gels on their own, except for LBG which can form a gel at high concentrations under specific conditions (Dea, Morris, Rees,

Welsh, Barnes and Price, 1977; Dea, Clark and McCleary, 1986; Richardson, Clark,

Russell, Aymard and Norton, 1999). The high water binding capacity and highly viscous solutions make GMs effective thickeners and stabilizers in the food industry. In addition,

LBG, GG and FG are all found to exhibit some surface, interfacial and emulsification activities (Garti, Madar, Aserin and Sternheim, 1997). Synergisms of GMs with other materials, including xanthan gum, agar, K-carageenan, starch, etc. are well studied (Cui,

Eskin, Wu and Ding, 2006). The degree of the synergism is believed to be related to M/G ratio and the galactosyl distribution on the mannan backbone (Dea et al., 1977;

* Published as: An Investigation of Four Commercial Galactomannans on Their Emulsion and Rheological Properties. Food Research International, 2009, 42, 1141-1146.

42 McCleary, Clark, Dea and Rees, 1985). Although some GMs from varied sources may have similar contents of galactose, their interaction with other materials can be different due to the difference in fine structure of the GMs (McCleary et al., 1985). Many properties of GMs are related to the M/G ratio. The higher the M/G ratio, the lower the solubility; the galactose side chains can prevent the mannan backbone from forming hydrogen bonded aggregates (Garti et al., 1997). In the present study, the effect of M/G ratios of GMs from varied sources on emulsion and rheological properties was investigated.

4.2. Materials and Methods

4.2.1. Materials

4.2.1.1. Sample sources

LBG (from Ceratonia siliqua seeds) and GG were purchased from Sigma-Aldrich,

Inc. (St. Louis, U.S.A). TG was purchased from TIC Gums (Belcamp, U.S.A). FG was provided by Emerald Seeds (Canafen Gum® Saskachewan, Canada). All sample solutions were prepared based on dry matter basis. Weighed samples were dispersed in distilled water with constant stirring at room temperature for 0.5 h, and then kept in 80°C water bath for 1 h with constant stirring. All aqueous sample solutions were prepared in duplicate and in the same way unless otherwise specified.

4.2.1.2. Chemical composition

Moisture content, protein content, total ash content and M/G ratio are given in Table

4.1. Protein content was analyzed using NA2100 Nitrogen and Protein Analyzer (Thermo

Quest, Milan, Italy). Total ash content was measured by 550°C muffler oven for 5 h. M/G

43 ratio was determined by monosaccharide analysis method using Dionex High

Performance Anion Exchange Chromatography system (Dionex, Sunnyvale, CA) as described by Wood, Weisz and Blackwell (1994).

4.2.1.3. Molecular weight and intrinsic viscosity

Intrinsic viscosity and Viscosity average molecular weight (Mv) are presented in

Table 2. Intrinsic viscosity was measured by dilute solution viscometry using a Cannon

Ubbelohde Dilution B glass viscometer (size 75, Glass Atrefact Viscometers, Braintree,

Essex, UK) in a constant temperature water bath at 25.0 ± 0.1°C. The measurement was based on the concentration range that the intrinsic viscosity [n] was calculated by using the following relationship:

[n] - lim (nSp/C) =lim (In nr/C) at C -+ 0 (1)

where C is the concentration of the polymer, nris relative viscosity and nsp is specific viscosity, defined as r|r-l. Huggins-Kramer plots of nsp/C and In r)i/C versus C were used to estimate the intrinsic viscosity [n] by extrapolation to zero concentration.

Viscosity average molecular weight, Mv, was calculated using the Mark-Houwink relationship given by Doublier and Launay (1981) and modified by Gaisford, Harding,

Mitchell and Bradley (1986) taking into account the different M/G ratios of the galactomannans

6 098 [n] = 11.55-10- [(l-x)Mv] (2)

where % = 1/[(M/G) + 1] and [n] is expressed in dL/g.

4.2.2. Methods

44 4.2.2.1. Emulsion properties

4.2.2.1.1. Surface tension

Surface tension of air-water interface was measured according to a method described by Izydorczyk, Biliaderis and Bushuk (1991). The decrease of the surface tension

(Dynes/cm) of water with increased concentration of GM was measured by Du Nouy ring method using a Fisher Surface Tensiomat model 21 (Fisher Scientific, Nepean, ON). The force acting on the ring was measured as it was moved upward from an air-gum dispersion surface. Sample solutions with various concentrations, 0.01%, 0.05%, 0.10%,

0.25% and 0.50% (w/w), were placed in 50 mL glass beakers. The changes in surface tension were recorded every 5 min for 2 h at 22°C ± 0.5°C.

4.2.2.1.2. Emulsion capacity and emulsion stability

0.1 g sample was suspended in 10 mL of distilled water and well dissolved by heating for 1 h at 80°C, followed by the addition of 10 mL canola oil (Loblaws Inc., Canada). All mixtures were then emulsified using a polytron at medium speed for 1 min and centrifuged at 1,500 g for 5 min. Emulsion capacity was calculated as:

^ , . . Height of emulsion layer , „„„. Emulsion capacity = - -— = 100% (3) Total height of fluid

Emulsion stability was determined by heating the emulsion at 80 °C for 30 min, cooling with tap water for 15 min to room temperature and then centrifuging at 1,500 g for 5 min. Emulsion stability was calculated as

45 Emulsion stability = Height of remaining emulsion laye x m% (4) Total height of fluid

4.2.2.2. Rheologicalproperties

All rheological measurements were conducted on a strain controlled ARES

Rheometer (TA Instruments, New Castle, USA). Concentrations of sample solutions were 0.1%, 0.5%, 1.0%, 1.5% and 2.0% (w/v) respectively. Viscosity was determined on a cone-and-plate geometry (4750 mm) with shear rate from 0.01-1000 s"1 at 25°C.

Viscoelastic properties were measured on a parallel-plate geometry (50 mm) at gap size of 0.7 mm, and strain of 1%. Storage modulus (G'), loss modulus (G") and phase angle (8) were determined at frequencies from 0.008 to 16 Hz at 25°C.

4.2.2.3. Statistical analysis

Statistical computations were performed using the GLM procedure of the

Statistical Analysis System (SAS Release 9.1, SAS Institute Inc., Cary, NC, USA). The data of intrinsic viscosity, molecular weight, emulsion capacity, and emulsion stability of

4 types of gum was examined by analysis of variance (ANOVA). Least square means of each property factors were calculated using the option of LSMEANS and statistical differences among four types of gum were identified at P < 0.05 using the option of

PDIFF.

4.3. Results and Discussion

4.3.1. Emulsion Properties

46 4.3.1.1. Interfacialproperties

The results on interfacial activities of the four gums are given in Figure 4.1.

-o—LBG -o-TG -*-GG -«— FG 75 -j £ 73 - ( i 71 - 69 -

(Dynes/ ( 67 - 65 - 63 - tensio n 61 - 59 - 57 - Interfacia l 55 - i i i i i 0.0 0.1 0.2 0.3 0.4 0.5

Concentration (% w/v)

Figure 4.1. Interfacial activities of galactomannans at various concentrations

The surface activities of all solutions are concentration-dependent. The higher the concentration, the more active it was to reduce surface tension. FG behaved the most surface active, followed by GG, LBG and TG. At lower concentrations, very weak surface activity was observed for TG and LBG.

It was reported that at lower concentrations (less than 0.1% w/v), short chain saccharides could migrate preferentially to the surface; while at higher concentrations

(0.1%-0.5% w/v) the activity might be due to the behaviour of macromolecules at the surface (Garti et al., 1997). In the present study, at lower concentrations, the differences of interfacial activities among the four gums were similar except FG, which exhibited

47 much lower surface tension at all concentrations; at higher concentrations, the differences of interfacial activities among the gums became more obvious (Figure 4.1). Solubility of the sample will also affect the surface activity, which is related to M/G ratio, a better solubility is related to a smaller M/G ratio (Tavares, Monteiro, Moreno, Lopes and

Tavares, 2005; Doublier and Launay, 1981).

The purity of the four gums varied, with protein content from 0.71% in TG to 4.51% in LBG (Table 4.1). Garti et al. studied LBG, GG and FG and found that after removing protein contaminants from these samples, surface activity was not affected (Garti and

Reichman, 1994; Garti et al., 1997); nevertheless, Gaonkar (1991) reported that purified

GG or LBG showed no surface activity. A most recent study in our laboratory revealed that after physically removing protein from FG, surface activity still remained, but at a much lower level (Youssef, Wang, Cui and Barbut, 2009). Statistical analysis at concentration 0.25% revealed that among the four GMs, there is no significant difference on surface activity between FG and GG (p = 0.8318). The variance between TG and LBG on protein content is the largest, 0.71% and 4.56% (Table 4.1).

TG and LBG showed significant difference on surface tension (p = 0.018) but the difference is smaller than that with FG (p = 0.0001) and GG (p = 0.0001). Therefore protein content is not the main factor leading to the significant difference on surface tensions. This conclusion is in agreement with the result by Garti and Leser (2001), who also indicated that protein was not a major factor responsible for the surface activity.

Hence, we assume that the M/G ratio and fine structures are the major causes for the differences on surface tension in the present study.

48 Table 4.1. Chemical compositions of galactomannans

Fenugreek Locust Sample Gum Guar Gum Tara Gum Bean Gum

Moisture% 10.49 ±0.16 10.16 ±0.01 12.37 ±0.05 9.77 ±0.16

Protein% 2.62 ±0.13 3.46 ±0.06 0.71 ± 0.02 4.57 ±0.37

Ash% 1.50 ±0.03 0.72 ± 0.02 0.77 ±0.01 1.04 ±0.02

M/G 1.2 ±0.0 1.7 ±0.0 3.0 ±0.1 3.7 ±0.0

4.3.1.2. Emulsion capacity & stability

Results on emulsion properties demonstrated a clear trend on emulsion capacity and stability of the four GMs (Figure 4.2). Although the centrifugation procedure might not reflect the true emulsion capacity and stability, it was adequate to identify potential emulsification properties (Huang, Kakuda and Cui, 2001).

80 Capacity Stability >60 I 50 £40 1 =3 30 H lI i I 20 4J £10 I u 0 I 1 Fenugreek Guar Gum Tara Gum Locust Gum Bean Gum

Figure 4.2. Emulsion capacity and stability of galactomannans

The order of emulsion capacity and stability followed the trend: GG>FG>TG>LBG, which was in the same order as intrinsic viscosity (Table 4.2).

49 Table 4.2. Intrinsic viscosity and molecular weight (Mv) of galactomannans

Fenugreek Locust Bean Sample Gum Guar Gum Tara Gum Gum

Intrinsic viscosity 15.10±0.14 15.80 ±0.28 14.55 ± 0.07 14.20 ±0.28 (dL/g)

6 MV(X10 ) 3.23 ±0.03 2.91 ±0.05 2.23 ± 0.01 2.08 ± 0.04

Statistical analysis indicated that there are no significant differences in emulsion capacity between GG and FG (p = 0.8377) and emulsion stability between FG and GG (p

= 0.2687). All the rest comparisons are significant. Therefore, the impact of protein content, again, can be neglected.

Higher surface active polymers can reduce surface tension while at the same time form a steric layer around the droplet, forming a stable emulsion (Dickinson, 2003). It can be seen in Figure 4.1 that FG and GG exhibited higher surface activities than TG and

LBG. GG formed the best emulsion; however, TG formed a better emulsion than LBG although it exhibited the poorest surface activity. Therefore, surface activity is not the only factor determining emulsion properties when polysaccharides are involved. The main contribution of polysaccharides to an emulsion is via water-holding and thickening properties (Dickinson 2003). Once an emulsion was successfully prepared, long-term stability was determined by how well the polymer produced a barrier at the interface, therefore preventing the droplets from aggregating or coalescing (Dickinson and Galazka,

1992). Garti et al. (1997) further proposed that when a GM solution was mixed with oil phase to prepare emulsions, super-saturation occurred; as a consequence, phase separation happened and led to the gum either precipitating or adsorbing onto the oil

50 droplets and forming a gel-like film covering the oil droplets. In the current study, four kinds of GMs were well dissolved in the solution. When mixed with oil phase, super- saturation may lead to phase separation. The molecules with larger M/G ratios can form intermolecular associations between unsubstituted regions in the mannan backbone (Dea et al., 1986); while samples with smaller M/G ratios may form more gel-like layer around the oil droplets and therefore enhance the emulsion capacity and stability. After the emulsions were subjected to a higher temperature (80°C, 30 min), the "barrier"

(Dickinson et al., 1992) or the "gel-like" structure (Garti et al., 1997) around the oil droplets were destroyed, which led to the destabilisation of the emulsions. Therefore, as observed in Figure 4.2, emulsion stability was consistently lower than emulsion capacity.

In addition to the M/G ratio, the current results suggest that interfacial activities and molecular weights of the samples should also be taken into account for the overall emulsion performance since they are directly related to the interfacial property, water binding capacity and thickening property as mentioned by Dickinson (2003).

4.3.2. Rheological Properties

4.3.2.1. Flow behaviour

The four GMs exhibited similar flow behaviour. TG is given as an example (Figure

4.3).

51 •0.1% D0.5% M.0% °1.5% "2.0%

0.001 1000 Shear rate (1/s)

Figure 4.3. Flow curves of tara gum at different concentrations

At concentration of 0.1% (w/v), all four GMs showed Newtonian flow behaviour; at the other concentrations, all solutions showed shear-thinning flow behaviour, which exhibited two regions: Newtonian plateau region at lower shear rates and shear-thinning region at higher shear rates. With increase in concentration, the viscosity increased

accordingly; meanwhile, on the flow curve, the transition from Newtonian plateau region to the shear-thinning region switched to lower shear rates (Figure 4.3). At lower shear rates, the disruption of molecular entanglements by the shear was balanced by the reformation of new ones, so that the viscosity kept constant; at higher shear rates, disruption of the entanglements predominated over the reformation of new ones, therefore the viscosity decreased (Sittikijyothin, Torres and Gonqalves, 2005). Among the

four gums, GG was the most viscous, followed by FG, TG and LBG, which is analogous to the order of intrinsic viscosity (Table 4.2). Statistical analysis revealed that most GMs

showed significant differences between one another except TG and LBG (p = 0.4632).

52 Since the variance of protein content between TG and LBG is the largest (Table 4.1) but no difference was shown on viscosity between the two gums, it can be concluded that protein content is not playing a major role on viscosity.

Zero shear viscosity (r|sp)o was derived by fitting shear thinning flow curves into the

Cross model (Cross, 1965). Polymer concentration C, not C[(nsp)o], is the governing parameter when concentration is larger than critical concentration (Launay, Cuvelier and

Martinez-Reyes, 1997). Therefore, in the present study, Master curves were obtained by superposing log(nsp)o versus logC (Figure 4.4).

2.5 -i

2.0

1.5 • FG: y=4.57x+0.95

1.0 • GG:>=3.92x+1.00

o 0.5 ATG:>^4.31X+0.78 I—I 0.0 • LBG:y=4.19x+0.48 -0.5

-1.0 -0.4 -0.2 0.0 0.2 0.4 LogC

Figure 4.4. Master curves (log(r|sp)o - logC) of galactomannans

(FG: Fenugreek Gum; GG: Guar Gum; TG: Tara gum; LBG: Locust Bean Gum)

The exponents for the four Master curves are 4.57, 3.92, 4.31 and 4.19 for FG, GG,

TG and LBG respectively. For typical polymers, the exponents usually range from 3-5

(Launay et al., 1997). Higher values on GMs were obtained, for example, 6.5 for LBG

53 (Kapoor, Milas, Taravel and Rinaudo, 1994). Some authors tried to relate the high exponents to the rigidity of polymer backbone (Sittikijyothin et al., 2005). Whether the higher exponent is associated with the formation of chain "hyperentanglements" (Morris,

Cutler, Ross-Murphy, Rees and Price, 1981) is still debatable (Launay et al., 1997).

The shear-thinning region of the flow curve was fitted into the power law model n=Kyn" , where k is the consistency index of the fluid, n is the flow behaviour index. In general, when concentration or molecular weight increases, n decreases and k increases.

Usually at high concentrations, n falls in the range of 0.15 to 0.25. By plotting n versus concentration (w/v%), it is clearly evident that with increased concentration power law index n decreased (Figure 4.5), which indicated that the solutions behaved more pseudoplastic in higher concentrations.

0.7 -+-FG •GG-A-TG •LBG 0.6 x 0.5 -] -d 0.4 - I 0.3 - 0.2 - o 0.1 - 0 - 0 0 0.5 1.0 1.5 2.0 Concentration (%)

Figure 4.5. Power law index (n) of galactomannans at various concentrations

54 4.3.2.2. 0.5% (w/v) solution

Flow curves of 0.5% gum solutions in Figure 4.6 showed an interesting phenomenon.

At very low shear rates (< 0.1 s"1), gum solutions exhibited shear-thinning flow behaviour before reaching the Newtonian zone. This phenomenon was randomly mentioned in the published literatures. The revealing of this phenomenon can be attributed to the advanced technique, which could move the detection limit to a smaller range.

When Zisenis (1997) studied correlation of molecular orientation and shear viscosity of high molar mass polymers in dilute solutions, he also found the non-Newtonian

decrease of viscosity to a Newtonian viscosity plateau. Several factors should be

considered in order to explain this phenomenon. The first factor is the molecular

orientation resistance. Under dilute or semi-dilute concentration, the Brownian motion

tends to maintain a random orientation while the flow field of shear tends to orient the

particle in the direction of flow. The balance between the two forces results in the

average orientation of the particles.

0.001 0.1 10 1000 Shear rate (1/s)

Figure 4.6. Flow curves of 0.5% galactomannan solutions

(FG: Fenugreek Gum; GG: Guar Gum; TG: Tara Gum; LBG: Locust Bean Gum)

55 The viscosity produced by this orientation resistance was decreased along with the increase of the flow rate. The larger the molecule, the more resistant it is (Zisenis 1997).

This explanation was well supported in the current study. Before the Newtonian plateau of the flow curve, a greater decrease in viscosity would be noticed in the order of

FG>GG>TG>LBG (Figure 4.6), which is analogous to the order of molecular weight:

FG>GG>TG>LBG (Table 4.2). The second factor is the degree of molecular association

(Zisenis, 1997). The associations among molecules could be weak or strong, physically or chemically entangled. Mao and Chen (2006) studied the thixotropic behaviour of GMs and further confirmed the associations between polymer chains. Their results indicated that GMs with higher M/G ratio (LBG) showed obvious thixotropic behaviour while GM with lower M/G ratio (GG) exhibited no significant thixotropy. It is believed that polymer-polymer hyperentanglement also contributes to the solution viscosity (Robinson,

Ross-Murphy and Morris, 1982). With the increased shear rate, the weak associates are broken which leads to decreased viscosity. At higher concentrations, the solution is dominated by intermolecular interactions. The viscosity produced by these interactions is far greater than the viscosity produced by molecular orientation resistance and some weak associations, therefore, this phenomenon was not observed at higher concentrations.

4.3.2.3. Visicoelasticproperties

As shown in Figure 4.7, for all gum solutions with concentrations ranging from 0.5% to 2.0%, a liquid-like behaviour was observed at lower frequencies (G"> G') and a solid like behaviour was observed at higher frequencies (G'> G"). G' values follow the trend:

FG >GG >TG >LBG, which is analogous to molecular weights (Table 4.2). Crossover frequency (where G - G") moved to lower ranges when the concentration increased.

56 1000 a 0.5%G' ~o o oo o 03 100 H • 0.5%G" w b A 1.0% G' 10 A 1.0% G" 6 • . * f * o 1.5% G' o * • s :* • 1.5% G" B 2* D O 2.0% G' 0.1 • • 2.0% G" Q 0.01 0. 001 0.01 0.1 1 10 100 Frequency (Hz)

a. Fenugreek Gum

1000 • 0.5%G'

03 • 0.5% G" 100 S.«88SSSS A 1.0% G' a A 1.0% G" 10 o 1.5% G' • * S9AA A .-BH * Q* A ^ -"I0 • 1.5% G" -3 1 Q*A*A •BnD o 9 A ° O 2.0% G' v o.i i A n • 2.0% G"

0.01 0.001 0.01 0.1 1 10 100 Frequency (Hz)

b. Guar Gum

Figure 4.7. Frequency sweeps of galactomannans at different concentrations

57 1000 n n0oO°° • 0.5% G' 100 • 0.5% G"

Pi •0**o A A * * A 1.0% G' 10 A 1.0% G" o 1.5% G' H " • • 0 • 1.5% G" -8 1 ^ A " • o O 2.0% G' °o° • 2.0% G" CO

Q 0.01 0.001 0.01 0.1 1 10 100 Freauencv (Hz)

c. Tara Gum

1000 n 0.5% G' 0o° • 0.5% G" PH 100 A 1.0% G'

9 * A A • 1.0% G" 10 #» A» OA A o 1.5% G' • 1.5% G" 1 1 •!0%* A^ -" ° O 2.0% G' § § 0.1 9 ° *' • 2.0% G" Q

0.01 0.001 0.01 0.1 1 10 100 Frequency (Hz)

d. Locut Bean Gum

Figure 4.7. Frequency sweeps of galactomannans at different concentrations

58 Other than the effect of molecular weight, M/G ratio also plays a role in viscoelastic properties. It was reported that GMs did not form gels by themselves except LBG which could form gels under subzero conditions (Dea et al, 1977; Dea et al., 1986) or if given enough time (several months) under room temperature (Richardson et al., 1999). Even after a good dissolution, the solutions of GMs with higher M/G ratios are thermodynamically unstable when brought below the dissolution temperature (Pollard and Fisher, 2006). The poor solubility may contribute to the lower G' values of samples with higher M/G ratios.

4.4. Conclusions

Along with molecular weight, M/G ratio also plays a role on emulsion and rheological properties of GMs. In the present study, the surface activities of the four GMs followed the trend: FG>GG>LBG>TG. Emulsion capacities and stabilities followed the trend: GG>FG>TG>LBG. Viscosity (n) of the four GMs followed the same trend as intrinsic viscosity: GG>FG>TG>LBG; storage modulus (G1) followed the trend:

FG>GG>TG>LBG, which is in agreement with the trend of molecular weight (Table

4.2).

Of the two factors, molecular weight and M/G ratio, it is difficult to distinguish which factor attributed more to the above mentioned properties in the current study. Further investigations will be carried out to understand the effect of single factors on the overall emulsion and rheological properties.

59 5*. RHEOLOGICAL INVESTIGATION OF SYNERGISTIC INTERACTIONS

BETWEEN GALACTOMANNANS AND NON-PECTIC POLYSACCHARIDE

FRACTION FROM WATER SOLUBLE YELLOW MUSTARD MUCILAGE

5.1. Introduction

Recent interest in the mucilage of yellow mustard seed is attributed to its unique rheological behaviour in solutions/dispersions and its ability to interact with galactomannans (GMs) synergistically (Cui, Eskin, Biliaderis and Mazza, 1995; Cui,

Eskin, Wu, and Ding, 2006). Synergistic interactions between polysaccharides are commercially viable because it leads to the development of new textures (Morris, 1990) and reduce the concentration of materials, hence save costs. Synergism between yellow mustard mucilage and GMs was first reported by Weber, Taillie and Stauffer (1974). Cui et al. (1995) further identified that the non-pectic polysaccharide (NPP) fraction from water-soluble yellow mustard mucilage consisting of P-l,4-linked D-glucan backbone chain was the responsible component for the synergistic interactions with GM.

There are several factors that may affect the magnitude of synergy in mixed polysaccharide systems, including molecular weight and intrinsic viscosity (Fernandes,

Goncalves and Doublier, 1991), salt concentration (Khouryieh, Herald, Aramouni and

Alavi, 2007; Schorsch, Gamier and Doublier, 1997), M/G ratio and fine structures (Dea and Clark, 1986; Fernandes, 1995), polysaccharide-polysaccharide blending ratio (Cui et al., 1995), total polysaccharide concentration (Mannion, Melia, Launay, Cuvelier, Hill,

* Published as: Rheological Investigation of Synergistic Interactions between Galactomannans and Non- Pectic Polysaccharide Fraction from Water Soluble Yellow Mustard Mucilage. Carbohydrate Polymers, 2009,78, 112-116.

60 Harding and Mitchell, 1992; Bresolin, Sander, Reicher, Sierakowski, Rinaudo and

Ganter, 1997), temperature (Mannion et al., 1992; Zhan, Brownsey, Ridout, Morris,

1993) and pH (Morris and Foster, 1994; Ross-Murphy, Morris and Morris, 1983;

Whitney, Brigham, Darke, Reid and Gidley, 1998; Tako, 1991). In the present study,

NPP was isolated as described by Wu, Cui, Eskin and Goff (2009a), and blended with four types of GMs, namely fenugreek gum (FG), guar gum (GG), tara gum (TG) and locust bean gum (LBG). Rheological studies were carried out to investigate how the factors including GM/NPP blending ratio, M/G ratio, total polysaccharide concentration and pH affect synergistic interactions.

5.2. Materials and Methods

5.2.1. Materials

Locust bean gum (LBG) (from Ceratonia siliqua seeds) and guar gum (GG) were purchased from Sigma-Aldrich, Inc. (St. Louis, U.S.A). Tara gum (TG) was purchased from TIC Gums (Belcamp, U.S.A.). Fenugreek gum (FG) was provided by Emerald

Seeds (Canafen Gum®). Characteristics of the four GMs were determined as described by

Wu, Cui, Goff and Eskin (2009b) and are presented in Table 5.1.

NPP was obtained as described by Wu et al. (2009a). All sample solutions were prepared based on dry matter basis. Weighed samples were dispersed in distilled water with constant stirring at room temperature for 0.5 h, and dissolved in 80°C water bath for lh, then cooled to room temperature for 0.5 h with constant stirring. All aqueous sample solutions were prepared in the same way unless otherwise specified.

61 Table 5.1. Characteristics of the four galactomannans

Fenugreek Locust Bean Sample Gum Guar Gum Tara Gum Gum

Moisture% 10.49 ±0.16 10.16 ±0.01 12.37 ±0.05 9.77 ±0.16

Protein% 2.62 ±0.13 3.46 ±0.06 0.707 ± 0.02 4.57 ±0.37

Ash% 1.50 ±0.03 0.72 ± 0.02 0.77 ±0.01 1.04 ±0.02

M/G* 1.2 ±0.0 1.7 ±0.0 3.0 ±0.1 3.7 ±0.0

;ic viscosity (dl/g) 15.10±0.14 15.80 ±0.28 14.55 ±0.07 14.20 ±0.28

Av **(X106) 3.23 ± 0.03 2.91 ±0.05 2.23 ± 0.01 2.08 ± 0.04

(*M/G: mannose to galactose ratio; **Mv: viscosity average molecular weight)

5.2.2. Methods

The effects of GM/NPP ratio, M/G ratio, total polysaccharide concentration and pH were investigated in the present study. At a total polysaccharide concentration of 0.5%

(w/w), NPP was mixed with FG, GG, TG and LBG at the ratios of 10/0, 9/1, 7/3, 5/5, 3/7,

1/9 and 0/10 respectively. At a higher concentration (1% w/w), GMs and NPP were blended at a fixed GM/NPP ratio (3/7). The effect of pH was investigated by dissolving the polysaccharide mixtures at fixed GM/NPP ratio (3/7) at the total polymer concentration of 0.5% with pH adjusted to 2.0, 6.5 and 12.0 respectively.

Viscoelastic properties were measured on a strain controlled ARES Rheometer (TA

Instruments, New Castle, DE, U.S.A.) on a parallel-plate geometry (50 mm) at gap size of 0.7 mm and strain of 1%. Storage modulus (G'), loss modulus (G") were determined at frequencies from 0.008 to 16 Hz at 25°C.

62 5.3. Results and Discussions

5.3.1. Effect ofGM/NPP blending ratio

At a total polysaccharide concentration of 0.5% (w/w), four GMs were blended with

NPP at 7 ratios. Figure 5.1 presents frequency sweeps of all combinations except the four pure GM solutions (GM/NPP ratio of 10/0).

At GM/NPP ratio of 9/1, only the mixture of LBG/NPP formed a gel. At the rest

ratios all the samples showed a weak gel property, with G' constantly higher than G", but, both G' and G" are frequency dependent. Storage and loss modulus (G' and G") at

frequency 0.8 (Hz) at the highest synergy ratio (3/7) are given in Table 5.2. Figure 5.2

presents the results of storage modulus (G') at frequency 0.8 (Hz). The results showed

that three GM/NPP blending ratios, 5/5, 3/7 and 1/9, exhibited synergistic interaction

since their G' values were consistently higher than that of the two polymers in isolate

(with GM/NPP ratios of 10/0 and 0/10).

63 1000 FG/NPP o 3:7 Q 5:5 100 o00ooo*oooooooooo 2AAAAAAAAAAAAAAAA A 1:9 g 10 xxxxxxxxxx¥¥¥** x 0:10 0 + 7:3 1 - 9:1 0.1 0.001 0.01 0.1 10 100

1000 -. o 3:7 GG/NPP a 5:5 100 - O o O O O O O WnVA^11 ^ ^8 2 A A A A A ASJ A A AA+V A 1:9 + £ 10 - xxxxxxxxxxx¥¥+ x 0:10 o 1 - + 7:3 - 9:1 0.1 - 0.001 0.01 0.1 10 100

1000 TG/NPP o 3:7 100 VaVbVbVbVfaVtftftfWtrtftftftbl a 5:5 A"A~AAYAAAAAAAAAAAA A 1:9 + g 10 -] xx.: xxxxx¥¥¥¥* + x 0:10 0 + 7:3 1 - 9:1 0.1 0.001 0.01 0.1 10 100

1000 LBG/NPP o 3:7 0000 100 ooooooooooo** n 5:5 nunaauanuauuaanaa A 1:9 g AAAAAAAAAAAA|+¥ AA x 0:10 0 10 a. + + + + + X * * * * ~ ~ " + 7:3 - 9:1 1 0.001 0.01 0.1 1 10 100 Frequency (Hz)

Figure 5.1. Frequency sweep of GM/NPP mixtures at total polysaccharide concentration of 0.5%.

(GM/NPP: galactomannan to non-pectic polysaccharide blending ratio; FG: fenugreek gum; GG: guar gum; TG: tara gum; LBG: locust bean gum)

64 Table 5.2. Storage and loss modulus (G' and G") of 0.5% GM/NPP mixed gel at blending

ratio of 3/7, frequency of 0.8 Hz.

Modulus Fenugreek Guar Tara Locust Bean (Pa) Gum Gum Gum Gum

G' 100.63 89.29 151.62 138.62

G" 17.17 16.79 25.68 21.14

• 10:0 ^9:1 Dl 7:3 0 5:5 H 3:7 H 1:9 S 0:10 140 H 120 § 100 1 80 a & 60 I 40 20

FG GG TG LBG

Figure 5.2. Comparison of synergistic effects between galactomannan and non-pectic polysaccharides from yellow mustard mucilage with various blending ratios at total polymer concentration of 0.5%, frequency of 0.8 Hz.

(FG: fenugreek gum; GG: guar gum; TG: tara gum; LBG: locust bean gum)

The G' values of the GM/NPP ratios of 9/1 and 7/3 are in between those of 10/0 and

0/10, with the exception of LBG/NPP mixture, which also showed synergistic behaviour at the ratio of 7/3 with G' higher than that of NPP solution (ratio of 0/10). The largest magnitude of synergy occurred at 3/7, followed by 5/5 and 1/9. This pattern applied to all

65 the four GM/NPP mixtures regardless of M/G ratio. Fernandes (1995) investigated synergistic interactions between xanthan gum and different LBG fractions with various

M/G ratios and revealed that the highest synergy occurred at 1:1 xanthan/LBG ratio at total polymer concentration of 1% for all LBG fractions regardless of M/G ratio. A similar study carried out by Mannion et al. (1992) obtained the same conclusion that the magnitude of synergy showed little dependence on the M/G ratio of the fraction.

The "junction zone" model has been generally accepted to explain synergistic interactions between GMs and cellulose-like molecules such as xanthan (Dea and

Morrisson, 1975; Morris, 1990). Since NPP was proved possessing blocks of unsubstituted P-1,4 linked D-glucose linkages (Cui et al., 1995; Wu et al, 2009a), junction zone model was adopted for the synergistic interactions between NPP and GMs

(Cui et al., 1995). Nevertheless, the extension of this model to the highly substituted mannon backbone such as FG and GG was restricted due to the limited availability of unsubstituted mannon binding sites. Junction zone model is also limited to explain the phenomenon that the best synergy occurred at the GM/NPP blending ratio of 3/7 for all

GMs even when the G/M ratio and the available "binding sites" varied greatly.

Another mechanism, segregative associations in single phase, as proposed by

Mitchell and coworkers (Penroj, Mitchell, Hill and Ganjanagunchorn, 2005), could be an alternative mechanism to explain the present data. The segregagtive association mechanism was based on a study of synergistic interactions between kappa carrageenan and konjac glucomannan: the synergistic system was a consequence of segregative association in single phase. They suggested that kappa carrageenan can promote the gelation of konjac glucomannan by reduce thermodynamically unfavourable contacts

66 between the two polymers. In addition, association between the two polymers may also exist. Adding one non-gelling polysaccharide to promote the gelation of a second polymer was also reported by Morris and coworkers, who investigated calcium pectinate network structure induced by the addition of oxidized starch (Picout, Richardson and

Morris, 2000), potato maltodextrin (Picout, Richardson, Rolin, Abeysekera and Morris,

2000) and GMs (Giannouli, Richarson and Morris, 2004). They postulated that synergistic interactions induced by segregated association occurred regardless of the structural differences of the polymers being added into the system.

Hefford (1984) reported that the behaviour of mixed polymers in solution was clearly determined by a delicate balance of forces. After studying several pairs of polysaccharides in aqueous systems, he found that hydrogen bonding between the different polymers was the cause for the polymers being mixed rather than phase separated. The polymer interactions in the binary systems in the present study are mainly through hydrogen bonding. If NPP and GM can interact via hydrogen bonding, the phase separation would be less possible. The order of the associations, either intra or inter, may depend on the thermodynamic behaviours between the molecules. Tolstogusov (2002) stated that biopolymers are limitedly co-soluble even if they are slightly different in composition, structure and/or conformation. Thermodynamically molecules favour interactions between the same molecules. There could be two kinds of associations co exist in the current mixed gel system, self-associations and cooperative junction zones.

We assume that the primary association is probably the self association, and the secondary association is the cooperative junction zone. If the cooperative junction zones are not feasible, for example, no binding sites available, phase separation might be a

67 consequence. However, the occurrence of phase separation may depend on the mobility of the molecules in the solution. It seems difficult to happen especially in high viscous solutions. This opinion was supported by several published reports. Kohyama, Iida and

Nishinari (1993) and Kohyama, Sano and Nishinari (1996) proposed that two crystalline regions co-exist in the mixed gel of konjac glucomannan and kappa carrageenan, one region consists of kappa karrageenan alone and the other region consists of konjac glucomannan and kappa carrageenan. The latter junction zone was weaker than the former. Garcia and Andrade (1997) studied the interaction between agarose and GG. 3-D network was observed in agarose domained area in the mixed gel while GG participated in the network formation by interacting with the surface of the agarose helix.

At the present study, we propose that there are two possible mechanisms involved in the synergistic interactions: the first one is the associative interactions or formation of junction zones between the NPP backbone and the unsubstituted region of the GM backbone chains; the second mechanism is the segregative associations in single phase with NPP network dominated in the mixed system while GM distributed in the same phase with possible associations with NPP as well.

5.3.2. Effect ofM/G ratio

At a total polymer concentration of 0.5% (Figure 5.2), TG formed the strongest gel with NPP at all blending ratios tested, followed by LBG, FG and GG. This pattern applied to all GM/NPP binary solutions.

As discussed previously, the addition of a certain amount of GMs, regardless of their

M/G ratios and the fine structures, can enhance the gelling network. However, the

68 magnitude of synergy may depend on the structural differences of the GMs. If the non- substituted GM backbone can associate with NPP by forming junction zones, the magnitude of the synergy might be increased. Therefore, as observed in the present study,

TG and LBG exhibited stronger synergy with NPP, while FG and GG exhibited lower synergy (Figure 5.2).

It is interesting to see that TG showed the best synergy. Table 5.1 indicates that intrinsic viscosity and viscosity average molecular weight (Mv) of TG was greater than

LBG, while the M/G ratio of LBG was 3.7, compared to 3.0 for TG.

Consequently, there should be more unsubstituted regions in LBG than in TG.

Although TG is a relatively purer sample among the four with less protein content (Table

5.1), Bresolin, Milas, Rinaudo, Reicher and Ganter (1999) claimed that the difference in gelling temperature and G' of xanthan/GM mixtures cannot be attributed to a difference in protein content. Therefore, the improved synergy of TG compared to LBG should be attributed to the fine structure as indicated by Dea and Clark (1986), because the galactose distribution pattern on the mannan backbone will give a difference on the length and frequency of the unsubstituted region.

5.3.3. Effect of total polymer concentration

It was noticed that the order of synergy at the two polymer concentrations, 0.5% and 1.0%, switched slightly between FG and LBG (Figure 5.3). Among the two GMs, FG exhibited a better synergy with NPP at the higher concentration; while at the lower concentration, LBG showed a better synergy (Figure 5.3). It is obvious that with increased concentration, G' increased accordingly for all four blends. The degree of

69 increase varied, as shown in Figure 5.3, with FG exhibiting the largest increase in G', followed by TG, LBG and then GG. This can possibly be caused by two characteristics of

FG: water binding capacity and conformation.

Water binding capacity may play a role in the binary system, especially with increased total polymer concentration. Sudhakar, Singhal and Kullarni (1995) studied the mixed polymer system with starch (5%) and GMs (0.1%, 0.2% and 0.3%) and found GG formed a superior synergy with starch compared to LBG. The authors suggested that this could be due to the greater hydration capacity of GG. In the current study, larger amount of water was imbibed within the NPP network when the polymer concentration was increased. The consequence of binding more water by FG, due to its larger water binding capacity compared to the other three GMs, is to induce a relatively higher concentration of NPP, thus strengthen the gelling network. This might explain the large increase in storage modulus of FG/NPP binary system.

Fernandes et al. (1991) investigated the synergistic interactions between kappa- carrageenan and LBG, and their results indicated that molecular weight of polysaccharide played an important role on the synergistic behaviour. With the increased Mw, the synergistic interactions will increase. In the current study, NPP mixed with four types of

GMs, and the Mv of the four types of GMs varied. Although GMs composed of 30% of the total polysaccharide concentration, their effect on the overall synergistic behaviour cannot be ignored. In the previous study, the Mv of the GMs were investigated and followed the order of GG>FG>TG>LBG (Wu et al., 2009c). Although GG possessed the highest Mv, the synergistic interaction with NPP was the lowest (Figure 5.3). TG and

LBG exhibited better synergy with NPP at a total polymer concentration of 0.5%, and TG

70 and FG exhibited better synergy at a total polymer concentration of 1%. Therefore, the molecular weight was not the dominating factor on synergistic interactions in the mixed system in the current study. We believe that the solution properties of the GMs might play a more important role when the polymer concentration increased. As observed from their Master curves (Figure 4.4), the viscosity of FG showed the biggest increase with concentration, followed by TG, and then LBG and GG. This trend met perfectly with the trend of increase on storage modulus (G') when concentration increased from 0.5% to 1%

(Figure 5.3). As discussed earlier, "hyperentanglements" among GM molecules exhibited obvious effect on the overall network strength. It is interesting to know how the molecular weight affects the "hyperentanglement". This could be further investigated by obtaining GM fractions with different molecular weights while possessing the same M/G ratio.

Frequency 0.8 Hz, GM/NPP at 3/7

400n 350- (Pa ) b 300- OT 3 250- 200- 150- mod u 100- 50-

Storag e o- FG GG TG LBG

Figure 5.3. Effect of total polymer concentration on synergistic interactions

between galactomannan and non-pectic polysaccharides from yellow mustard mucilage

(GM/NPP: galactose to non-pectic polysaccharide blending ratio; FG: fenugreek gum; GG: guar gum; TG: tara gum; LBG: locust bean gum)

71 Conformation of fully substituted GM in the binary system may also affect the synergy. A conformational study on GM with M/G ratio of 1/1 by using modelling approach revealed that introduction of the galactosyl residues on the mannan backbone could lower the chain dimension by the attractive interactions between the galactosyl substituents and the remaining part of the molecule (Petkowicz, Milas, Mazeau, Bresolin,

Reicher, Ganter and Rinaudo, 1999). This compact conformation may help with the synergistic interactions somehow. McCleary (1979) proposed that in GM with fully substituted galactose, the galactose might align on one side of the chain leaving the other side as the "smooth region" to interact with the cellulose-like chain. All these hypotheses should be further verified in future studies.

5.3.4. Effect ofpH

The effect of pH conditions on synergistic interactions was investigated at a fixed

GM/NPP blending ratio of 3/7 and a fixed total polysaccharide concentration of 0.5%

(w/w). Three pH conditions, 2.0, 6.5 and 12.0, were investigated. Results are shown in

Figure 5.4. The order of magnitude of synergy followed the order of: pH 6.5> 2.0 > 12.0.

This order applied to all four GM/NPP mixtures.

Our previous study on NPP (Wu et al., 2009a) showed that acidic or alkaline pHs increased the gelling property of NPP, indicating an enforced gelling network within NPP molecules. Once the NPP-NPP network is enhanced, the GM-NPP interaction is depressed, and as a consequence, synergistic interactions will decrease. When pH increases, solubility of GM molecules also increases which may reduce the GM-NPP

72 interactions. The GM molecules could also be degraded under acidic or alkaline condition, hence led to the decreased G' compared to the neutral condition.

n pH 2.0 a pH 6.5 a PH 12.0 160 -,

FG GG TG LBG

Figure 5.4. pH effect on synergistic interaction between galactomannans and non-pectic

polysaccharides.

(FG: fenugreek gum; GG: guar gum; TG: tara gum; LBG: locust bean gum)

5.4. Conclusions

The authors propose that two mechanisms of synergistic interactions co-exist in the

GM/NPP binary system. One of the mechanisms is the well accepted "junction zone" model, suggesting that the unsubstituted mannan chain of GM associates with the cellulose-like backbone of NPP via hydrogen bonding to form junction zones. The second mechanism is that the polymers segregatively associated in single phase with the possible associations between the two different polymers. It is arguable on how the fully substituted mannan chain interacts with the cellulose-like chain of NPP, especially in the increased polymer concentration. One assumption is that the a-l,6-linked D-galactose fully or alternatively aligned on one side of the chain, leave the "smooth" side attached to

73 the cellulose-like backbone. This kind of conformation needs to be further verified using molecular scale techniques.

This study further confirmed that not only the M/G ratio of GM but also the fine structures of GM molecules determine the magnitude of synergistic interactions with

NPP. Due to the difficulty in obtaining the information on fine structures, it makes the elaboration of the synergy mechanism more complicated. However, the present study is able to elucidate the synergistic effects between a certain GM and NPP. The overall magnitude of the synergistic behaviour can be controlled by adjusting GM/NPP blending ratio, total polymer concentration and pH conditions.

74 6. A MOLECULAR MODELLING APPROACH TO UNDERSTAND

CONFORMATION OF GALACTOMANNANS WITH DIFFERENT

MANNOSE/GALACTOSE RATIOS

6.1. Introduction

Galactomannans (GMs), a group of neutral polysaccharides naturally occurring in the seeds of some legumes, consist of a mannan backbone, with a-glactose at C6 of mannose (Dea, Clark and McCleary, 1986). There are four primary plant sourses of GMs, i.e. locust bean or carob tree (Ceretonia siliqua L.), guar (Cyamopsis tetragonolobus L.

Tabu.), tara plant (Caesalpinia spinosa) and fenugreek (Trigonella). Galactomannans from these four plant sources give distinctive structural characteristics, generally, the mannose to galactose ratio (M/G) varies from 4:1, 3:1, 2:1 and 1:1 for locust been gum

(LBG), tara gum (TG) , guar gum (GG) and fenugreek gum (FG) respectively (Dea &

Morrison, 1975). GMs are used widely in food, pharmaceutical and cosmetic industries as stabilizers and emulsifiers due to theirs high water binding capacity, emulsification properties and easy-to-form synergistic interactions with other components such as carbohydrates or proteins thus increasinging viscosity (i.e. Bresolin et al, 1997; Bresolin et al., 1999; Cui, Eskin, Biliaderis and Mazza,1995; Sudharkar, Singhal and Kulkarni,

1996; Goff, Ferdinando and Schorsch, 1999; Tarvares, Monteiro, Moreno and Silva,

2005). The synergistic interactions of GMs with cellulosic materials such as xanthan gum and non-pectic polysaccharides (NPP) from yellow mustard mucilage are well established

(i.e. Cui, 2001; Cui, Eskin, Wu and Ding, 2006; Cui et al., 1995). In a recent study (Wu,

Cui, Eskin & Goff, 2009b) we proposed that two mechanisms may co-exist to explain the synergism between GMs and NPP: the "junction zone" model and the segregative

75 association model. The "junction zone" model has been well accepted for polymers with long unsubstituted smooth regions. However, this mechanism alone cannot explain the polymers with fully or partially substituted chains. Instead, the segregative association model can be a good alternative to explain the synergistic behaviour between fully or partially substituted polymers like FG or GG with NPP. The fine structure of GM molecules, including M/G ratio, galactose distribution pattern along the backbone, as well as its orientation, may greatly affect the conformation of the polymer, and the behaviour of its solution properties and interactions with other molecules (Bergamini et al., 1995;

Mazeau and Rinaudo, 2004; Viebke and Piculell, 1996). It still remains a major challenge to obtain information about the "fine structure"-conformation relationship due to the difficulty in obtaining information on the fine structures. However, computer simulation of molecules with specific fine structures of various distribution patterns and orientations is a superior and effective technique for comparing the conformational characteristics with the experimental data. Molecular modelling has become a unique and indispensable tool in conformational analysis by proving invaluable information complementary to the data obtained experimentally (Rao, Qasba, Balaji and Chandrasekaran, 1998).

Computer simulation of GM chains have been investigated by a number of researchers using various approaches, i.e. Bergamini et al (1995) investigated the conformational properties of oligo-GMs using MM3 force field and compared their data with NMR results. Chandrasekaran and Radha (1997) studied the geometrical characteristics of galactomannans chains and found a good agreement with data obtained from X-ray diffraction. Mazeau and Rinaudo (2004), Petkowicz, Reicher and Mazeau

(1998) and Petkowicz et al. (1999) systematically studied the conformation of GM chains

76 with different M/G ratios using MM3 force field and predicted the stiffness and the flexibility of these chains. In the present study, GMs with M/G ratios of 4, 3, 2 and 1 were built to simulate the materials used in our previous studies (Wu, Cui, Eskin and

Goff, 2009c, Wu et al., 2009b) with the conformational parameters calculated by the molecular modelling software. The objective of the current study was to investigate how

M/G ratios of GMs affect chain conformations and behaviours in solution in order to understand the mechanism responsible for synergistic interactions with other polymers.

6.2. Materials and Methods

6.2.1 .Materials

In order to simulate the materials examined in the previous studies, GM molecules with corresponding M/G ratios, 4/1, 3/1, 2/1 and 1/1 were built to represent the following materials investigated in our previous studies, locust bean gum (LBG), guar gum (GG), tara gum (TG) and fenugreek gum (FG).

6.2.2. Molecular modelling methods

6.2.2.1. Molecular models and force fields

Molecular modeling calculations were performed on Insight II /Discover_3 and RIS program (Version 4.0.0, Molecular Simulations Inc., San Diego, USA) and a silicon

Graphic 02 workstation.

Simulations were performed with a dielectric constant set to 4.0 for all calculations.

This value was believed to be most appropriate from comparisons of the D- glucopyranose ring with crystal structures (French, Rowland and Allinger, 1990). It was

77 used to obtain better agreement with experimental values in aqueous solutions (Homans,

1990). Although high dielectric constants like s = 80 were used by many researchers for better comparisons with experimental data, a low dielectric constant (e =4 in the present study) was preferred over a high dielectric constant after comparing several adiabatic maps using some sharply different dielectric constants. Stortz (1999) stated that using low dielectric constants could result in a good approximation to the true adiabatic map, while high dielectric constants could damp the electrostatic and hydrogen-bonding interactions.

The AMBER force field applied contained a correction for the anomeric and exo- anomeric effects and has been adapted for generating of adiabatic contour map of epimelibiose (a 1-6 linked galactose to mannose), and for minimization of energy of polysaccharide chains. After energy minimization of polysaccharide chains, the consistent valence force field (CVFF) was applied for RIS-Motropolis Monte Carlo

(RMMC) calculation of conformational parameters.

6.2.2.2. Nomenclature

The recommendations and symbols proposed by the Commission on Nomenclature

(IUPAC-IUB, 1997) are used throughout this paper. The following torsion angles describe the relative orientations of a pair of contiguous residues,

q> = 05'-Cl'-06-C6

V = Cl'-06-C6-C5

The orientation around the C5-C6 bond is defined by

78 co = C4-C5-C6-06

The primed atoms correspond to the non-reducing residue (a-galactose) and the unprimed atoms correspond to the reducing one (P-mannose) (Figure 6.1).

Figure 6.1. Schematic representation of epimelibiose and torsion angles

6.2.2.3. Generation of adiabatic contour map (y>-y/ map) of epimelibiose

The stable chair 4Ci conformation has been applied for the pyranose monosaccharide rings. The hydroxymethyl groups had either gauche-gauche (gg), gauche-trans (gt) or trans-gauche (tg) orientations (relative to the ring C-4 and 0-5 atoms, respectively), with their hydroxyl hydrogen atoms oriented to weakly hydrogen bond with the ring oxygen atoms. Clockwise (C) or reverse clockwise (R) orientations were chosen for the secondary hydroxyl groups. Twenty-four starting conformations of epimelibiose were applied with gt, gg or tg orientation of the hydroxymethyl groups. The conformations of epimelibiose were explored by systematically stepping the glycosidic cp and \|/ torsion angles from -180° to 180° with 20° increments. At each point, energies were calculated

79 after energy minimization with restrains for those (p and \|/ torsion angles, but while allowing the other variables to relax.

6.2.2.4 Polysaccharide chain building

Repeat units, epimelibiose and mannose, are randomly linked through mannopyranosyl residues via (3 1-4 linkages with the appropriate epimelibiose/mannose ratio to represent corresponding GMs. The chain was built using the R-Random function under Polymerizer Program. Each GM chain was composed of 120 repeating units due to the limitation of the computer capacity. The randomized results of M/G ratios are given in Table 6.1.

Table 6.1. Simulated galactomannans with various M/G ratios

Total No. of No. of No. of Galactomannan Repeatir M/G ratio Epimelibiose Mannose Unit

Fenugreek Gum 120 120 0 1.0/1

Guar Gum 120 56 64 2.1/1

Tara Gum 120 43 77 2.8/1

Locust Bean 120 25 95 4.8/1 Gum

6.2.2.5. Molecular simulation of polysaccharide chains

Molecular minimization and molecular dynamics were carried out on Discover_3 module. The overall procedure included minimization-dynamics-minimization steps described in detail in the following: The molecular minimization was performed by the initial 10,000 steps of Steepest Decent method followed by the Polak-Ribiere conjugate

80 gradient method, and then the final Quasi-Newton (BFGS) method with the default energy convergence criterion of 0.001 kcal/mol. Non-bond cutoffs were set as 9.5 A, 1.0

A and 0.5 A for group size, spline width and buffer width respectively. After molecular minimization, constant NVT Molecular Dynamic (MD) simulation was carried out using the leap-frog algorithm with a 1 fs time step by: i) increasing the temperature from 10 K to 353 K, and ii) bringing the temperature down from 353 K back to 298 K, and finally iii) keeping temperature constant at 298 K with the production run set for 10,000 for each temperature stage. The temperature was controlled by Velocity Scaling in equilibration phases. The increased temperature is to ensure that energy barriers could be jumped over from one local minimum to another. After dynamic simulation, the molecule was minimized again as described previously.

6.2.2.6. RIS (Rotational Isomeric State) Metropolis Monte Carlo (RMMC) calculation

The conformation parameters were derived by RMMC calculation using the RIS module of the MSI software. The RMMC procedure usually starts with an optimized polymer chain obtained from previous procedures. A Monte Carlo step consisted of randomly selecting a rotatable backbone bond and making a small change of the conformation by rotating between -180° and 180°. After the bond was rotated to the new torsion angle the energy of the system was computed using a reduced force field consisting of torsion and nonbonded terms, including those for Coulombic electrostatic interactions. Based on the temperature and energy of the new conformation relative to the old one, the change was either retained or not. This process was repeated many times in order to yield a set of conformations characteristics for that chain at the specified temperature. The properties were then updated accordingly. In the present study, to

81 compare the effect of H-bonds among galactosyl side groups on the conformations of each type of GMs, two sets of parameters were applied. One set of parameters,

MinBond = 3 and MaxBond = 4, indicated that when atoms were fewer than 3 bonds and larger than 4 bonds the interactions (van der Waals and Coulomb) were neglected.

The other setting, MinBond = 3 and the Cut-Off distance = 15 A, indicated that when atoms were fewer than 3 bonds and beyond the distance 15 A, the interactions (van der

Waals and Coulomb) were set to zero. The temperature of simulation was set at 298 K.

The number of equilibration steps was set at 5,000. This stage was necessary to bring the polymer chain from its initial conformation to a conformation typical for the chain to a relaxed condition. After the equilibration steps, 20,000 of simulation steps was set for each rotatable bond. The conformational parameters such as persistence length (Lp), characteristic ratio (Coo) and radius of gyration (Rg) were obtained after the RMMC calculation.

6.3. Results and Discussion

6.3.1. Adiabatic contour map of epimelibiose

The adiabatic map was generated by plotting the low energy values versus the corresponding cp and v|/ torsion angles. Different starting points may lead to different minima after geometry optimization (Stortz, 1999). It was found that the hydroxymethyl groups in monosaccharides showed a marked trend to stay at preferential positions including gauche-gauche (gg), gauche-trans (gt) and sometimes trans-gauche (tg) (Tran,

Buleon, Imberty and Perez, 1989). The true adiabatic map for a disaccharide can only be obtained after calculating 3 staggered positions (gg, gt and tg) for the 10 exocylic

82 groups, which is 59,049 (3 ) relaxed maps and determining for each cp-\|/ conformation the one with lowest energy (French et al., 1990). The secondary hydroxyl groups were likely to form a crown of cooperative hydrogen bonds oriented either clockwise (c) or reverse-clockwise (r) around the pyranose ring (Ha, Madsen and Brady, 1988). Stortz

(1999) generalized that after restricting the rotation of the hydroxymethyl groups to their preferential positions, only a few (16-36) of those 59,049 starting conformations for a disaccharide may be used after reviewing the serial works by Dowd et al. (Dowd, Zeng,

French and Reilly, 1992; Dowd, French and Reilly, 1992; Dowd, Reilly and French,

1992; Dowd, French and Reilly, 1995) and Engelsen et al. (Engelsen and Rasmussen,

1993; Engelsen et al., 1995). This approach was well applied by Dowd et al. (Dowd,

Zeng, French and Reilly, 1992; Dowd, French and Reilly, 1992; Dowd, Reilly and

French, 1992; Dowd, French and Reilly, 1995) in a systematic study of different disaccharides starting with 16 or 24 conformers, which afterwards became a standard procedure (Stortz, 2006). In the present study, 24 starting conformers were applied for generating the adiabatic contour map of epimelibiose with gg and gt for mannose residue and gg, gt and tg for galactose residue. The resulting adiabatic contour map is given in

Figure 6.2.

83 Table 6.2. Comparison of torsion angles at the lowest energy level from current study and

other research groups

Torsion from present Bergamini et al. Petkowicz et al. angles study (1995) (1998)

9 -165° 79° 79°, 87°

V -175° -168° -168°,83°,-67°

CO 65° 74° 71,74,79°

Dielectric 4 4 80 Constant

Force Field Amber MM3 MM3

Contour Map of Epimelibiose

-100

-120

-140

-160

-180

Figure 6.2. Contour map of epimelibiose

84 Two low energy wells were identified from the map, one of the low energy wells located at (p/\|/ = -1757-67° with the corresponding low energy level of-1.5 kcal/mol, the other well was corresponding to the position of (p/\|/ = -1657-175°, co = 65° with the gg- gg-cr initial conformation with the relaxed energy of -2.5 kcal/mol. By comparing this result with that from Petkowicz et al. (1998) and Bergamini et al. (1995), some similarities can be observed as listed in Table 6.2, i.e. similar co angle, 65° in the present study, 71-79° from Petkowicz et al. (1998) and 74° from Bergamini et al. (1995). The value of vj/ angle also shared some similarity with Bergamini et al. (1995), with -175° in the present study and -168° from Bergamini et al. (1995). The variations among different groups might be due to the use of different dielectric constant values, starting geometries

(Petkowicz et al., 1998) and different force fields (Stortz, 1999). In the present study, the conformer with the lower energy well (-2.5 kcal/mol), (p/\|/ = -1657-175° and co = 65° was adopted for building polysaccharide chains.

6.3.2. Effect of intra-chain interactions on the conformation ofGMs

Some results showed that the galactosyl groups occurred in blocks (Hall and

Yalpani, 1980) while others found them in pairs and triplets (Hoffman and Svensson,

1978; McCleary, Clark, Dea and Rees, 1985), or in a random arrangement (Ganter et al.,

1995). Petkowicz et al. (1998) applied a random pattern when building their polymer chains for computational simulations. In the present study, we also adopted a random distribution pattern on GG, TG and LBG.

When conducting RMMC calculations, two sets of parameters were applied to investigate the conformational properties of GMs. By comparing the results derived from

85 the two different parameter settings, the information on how the substitution rates affect the overall conformation of the substituted chains can be obtained. The conformation characteristics are shown in Table 6.3. When the intra-chain interactions were not considered (MinJBond = 3 and MaxBond = 4), the substitutions produced some bending on the chain conformation. As shown in Table 6.3, the value of Lp, Coo and Rg for FG are the smallest, followed by GG, TG and LBG. This result is in agreement with that from

Bergamini et al. (1995), who studied GM oligomers and found that the addition of a galactosyl residue always produced a kink on the chain backbone. The setting of Cut-Off distance = 15 A allowed the intra-chain interactions included in the calculation when simulating the conformation of the polymers. The results showed that, compared to the results from the first parameter settings, more compact conformations are derived as shown in Table 6.3.

Table 6.3. Molecular simulated conformational properties of galactomannans derived

from different parameter settings for RMMC calculation

Min Bond = 3, Max Bond = 4 Min Bond == 3, Cut Off=15A Galactomannans LP v^oo Rg LP v^oo Rg Fenugreek Gum 30.49 41.42 85.48 2.18 2.35 42.13

Guar Gum 36.48 50.91 93.83 4.86 6.27 31.05

Tara Gum 50.84 61.33 105.78 3.10 4.66 29.61

Locust Bean 56.30 66.49 110.68 18.46 18.24 59.18 Gum

Lp: persistence length; Ca.characteristic ratio; Rg: radius of gyration

86 Figure 6.3 demonstrates the obvious difference between the conformations using the two different parameter settings.

a: fenugreek gum a: guar gum a: tara gum a: locust bean gum

b: fenugreek gum b: guar gum b: tara gum b: locust bean gum

Figure 6.3. Conformation of simulated galactomannans obtained from different parameter settings

( a. MinJBond = 3 and Max_Bond = 4; b. Min_Bond = 3 and the Cut-Off distance =15 A.)

87 However, the degrees of flexibility of the chains were not following the order of the galactosyl substitution rate. We believe that the conformation parameters derived from the first settings (Min_Bond = 3 and Max_Bond = 4) can only resemble the geometrical orientation of the chain without considering the involvement of the surrounding residues, while the results from the second settings (Min_Bond = 3 and Cut-off = 15A) are more appropriate for representing the polymers in a real solution system. Therefore, the results based on the second setting will be discussed in detail to compare the results from other researchers and the experimental data from our previous study.

6.3.3. Comparison of conformational characteristics ofGMs with various M/G ratios

As shown in Table 6.3, LBG obviously exhibited a stiffer chain with much greater values of Lp, Coo and Rg, which are not in the same range with FG, GG and TG. These results are in good agreement with those reported by Petkowicz et al. (1998), who studied the conformational change of GM chains with different M/G ratios (from 5/1 to 1/1) and found that the insertion of galactosyl side groups greatly lowered the calculated chain extension within the M/G ratio range of 5/1 to 3/1. In their results, however, in the M/G range of 3/1 to 1/1, the lowering effect was insignificant. Compared to the extremely high values for LBG, the conformation parameters of FG, GG and TG were much closer to one another, with FG showing the most compact conformation, followed by TG and then

GG. The order of TG and GG can be confusing since if the intra-chain interactions were caused by the side groups, the more substituted chain should behave more flexible.

However, not only the galactosyl substitution rate, but also the galactosyl distribution pattern may simultaneously play a role on the overall conformation. One assumption on

TG is that the more compact conformation may be due to the combined effect of the fully

88 substituted region and the smooth region, which make TG more specific to other GMs by possessing both regions. The fully substituted region may help with making a turn on the

TG molecule and then provide opportunities for smooth regions to form junction zones.

Therefore, GG, by having a higher substitution rate, may possess a less compact conformation compared to TG due to lack of smooth regions. Interestingly, one of our previous studies (Wu et al., 2009c) examined the Master curves of the four materials, FG,

GG, TG and LBG, to investigate the viscosity-concentration relationship. The results found FG to be the most concentration dependence, followed by TG and GG, which followed the order of chain flexibility generated by the current study. Some authors believed that the higher exponent in the Master curves was associated with the formation of "hyperentanglements" (Morris, Culter, Ross-Murphy, Rees and Price, 1981). The current results suggest that the exponents might associate with the "strength" of

"hyperentanglements". The degree of the "strength" could be a consequence of the intra molecular associations via the side groups. FG may produce stronger

"hyperentanglement" compared to the less densely substituted GM chains. In addition, the chain flexibility is another factor that contributes to forming the

"hyperentanglements". As observed in the present study, TG was more flexible than GG and showed stronger concentration dependency in solutions. With respect to LBG, the

"junction zone" model could be the best explanation due to the stiffer chain property and more unsubstituted regions.

For FG, the "ordered region", as hypothesized in our previous study (Wu et al.,

2009b) was not observed from the simulated FG molecule in the present study. Doyle,

Lyons and Morris (2009) proposed that the galactosyl side groups might lie above or

89 below the mannan backbone without blocking the interactions between mannan backbones. This kind of conformation might exist in the solid state. However, we were highly doubtful about this conformation if the polymers were put in solutions with constant dynamic movements. In the current study, molecular simulation work did not reveal any "ordered" structure from FG molecules except that the galactosyl side groups may form H-bonding and thus enhance the compactness of the conformation.

6.3.4. Effect of conformation on synergistic interactions with NPP

In our previous paper (Wu et al., 2009b), we found that GMs with different M/G ratios behaved differently in synergistic interactions with NPP. Two mechanisms were proposed, including the "junction zone" model and the "segregative association" model, along with a hypothesis that there might be a "regular region" along the fully substituted mannan backbone, where galactosyl substitutions might align in an ordered format so that it could cooperatively associate with the smooth region on NPP. However, the "ordered" structure was not observed so far by molecular modeling, in stead, a more compact conformation of FG was observed, and this compact conformation provides less possibility for "cooperative association" between FG and NPP due to the lack of binding sites for the smooth region of NPP chains. Therefore, the hypothesis of "the ordered structure between the fully substituted chain and the smooth region of the NPP molecules" was not supported by the present results. As a result, "segregative association" is considered as the most possible mechanism to explain the synergistic interactions between FG and NPP. The current study further confirmed that there are intra-molecular associations along the side groups of FG, which can help to form

"hyperentanglements". Meanwhile, due to the high flexibility of the more substituted

90 chains, the molecules can penetrate the 3-D network of NPP molecules, thus strengthen the overall network. In the case of LBG, the synergistic interactions could be attributed to the "cooperative associations" like "junction zones". TG was categorized differently from LBG according to the conformational parameters in the current study (Table 6.3).

Based on this information, the synergistic behaviour between NPP and TG should follow the "segregative association" model since the more compact and flexible TG chain may not be appropriate for forming "junction zones" with the cellulosic chains. However, the behaviour of GMs in a copolymer system still needs to be further investigated.

6.4. Conclusion

In the present study, four GMs with various M/G ratios were simulated and conformational characteristics were calculated. The results indicated that M/G ratio played an important role on the chain conformation. The galactosyl side groups have an effect of lowering the chain dimension. The intra-chain associations via H-bonding along the side groups are also ascertained. However, the effect of intra-chain associations on the conformational change is not following the order of galactosyl substitution ratio. We believe that not only the M/G ratio, but also some other factors like distribution pattern are determining the chain conformations.

The "ordered" structure, as hypothesized in our previous paper (Wu et al., 2009b), was not observed in the current study. Therefore, the enhanced synergistic interactions between FG and NPP in higher concentration (1%) are believed to be caused by the intra- chain interactions along the side groups, which can form stronger "hyperentanglements".

The "hyperentanglements" themselves can increase the viscosity of the solution.

91 Meanwhile, due to the higher flexibility of the fully substituted chains, the molecules may penetrate in the NPP networks, thus increase the overall gel strength.

The results of this study obtained good agreement with similar studies from other research groups. Further investigation will be carried out on how the "fine structures" affect the chain conformations and how the different conformation will affect the synergistic interactions with other molecules.

92 7. GENERAL CONCLUSION

In the present study, synergistic interactions between non-pectic polysaccharides from water soluble yellow mustard mucilage and galactomannans (GMs) with various M/G ratios were investigated. The non-pectic-polysaccharide (NPP) fraction was isolated from the water soluble yellow mustard mucilage. Its chemical, physical and rheological properties were characterized. Its structural information was elucidated using NMR spectroscopy combined with methylation analysis coupled with GC-MS. Four types of

GMs, fenugreek gum (FG), guar gum (GG), tara gum (TG) and locust bean gum (LBG), were used to study the synergistic interactions with NPP. The chemical, physical and rheological properties of these four commercially available GMs were investigated and their emulsification properties were also determined. The synergistic interactions were studied by mixing GMs with NPP at different total polymer concentrations, GM/NPP ratios and pH values. A molecular modelling approach was applied to understand the mechanisms of the synergism between NPP and GMs by analyzing the conformations of the four types of GMs. The following results are derived from the current study: i. Isolation and partial characterization of NPP: the combination of pectinase hydrolysis and ammonium sulphate precipitation resulted in a non-pectic polysaccharide with more glucose and less uronic acid compared to ethanol precipitation. This fraction exhibited unique gelling characteristics under extreme acidic and at very high temperature conditions. These superior rheological characteristics suggest great potentials for the utilization of this material by food or non-food industries. ii. The structure elucidation of NPP: Methylation of NPP was successfully applied in the current study to obtain high quality NMR spectra. After methylation, ID !H and 13C and

93 2D WH and ^"C NMR spectra were obtained. More detailed structural information on methylated NPP was derived based on the NMR spectra. The result revealed that methylated NPP possessed a (3,1-4 linked glucosidic backbone, with P-mannose 1,2 linked to the backbone glycosidic chain, and the P-galactose 1,6 linked to the glucosidic backbone chain. The methyl groups were substituted to the 2, 3, 6 positions of the glucose residues. A minor amount of ethyl group was also detected and was found substituted at the C3 position of the glucose residue. iii. Characterization of the four GMs: The surface activities of the four GMs followed the trend: FG>GG>LBG>TG. Emulsion capacities and stabilities followed the trend:

GG>FG>TG>LBG. Viscosity (n) of the four GMs followed the same trend as intrinsic viscosity: GG>FG>TG>LBG; storage modulus (G1) followed the trend:

FG>GG>TG>LBG, which is in agreement with the trend of molecular weight. iv. Synergistic interactions between NPP and GMs: We propose that two mechanisms of synergistic interactions co-exist in the GM/NPP binary system. One of the mechanisms is the well accepted "junction zone" model, suggesting that the unsubstituted mannan chain of GM associates with the cellulose-like backbone of NPP via hydrogen bonding to form junction zones. The second mechanism is that the polymers segregatively associated in a single phase with the possible associations between the two different polymers. This study further confirmed that not only the M/G ratio of GM but also the fine structures of

GM molecules determine the magnitude of synergistic interactions with NPP. The present study is able to elucidate that the overall magnitude of the synergistic behaviour can be controlled by adjusting GM/NPP blending ratio, total polymer concentration and pH conditions.

94 It is arguable on how the fully substituted mannan chain interacts with the cellulose like chain of NPP, especially in the increased polymer concentration. One assumption is that the a-l,6-linked D-galactose fully or alternatively aligned on one side and the other side of the chain, attached to the rigid p-l,4-linked D-mannosyl cellulose-like backbone.

This kind of conformation needs to be further verified using molecular scale techniques such as molecular modelling software. v. A molecular modelling approach to understand the effect of conformation on synergistic behaviour of GM, FG, GG, TG and LBG, was carried out. The results indicated that M/G ratio played an important role on the chain conformation. The galactosyl side groups have an effect on lowering the chain dimension. The intra-chain associations via H-bonding along the side groups are also ascertained. However, the effect of intra-chain associations on the conformational change is not following the order of galactosyl substitution ratio. We believe that not only the M/G ratio, but also some other factors like distribution pattern are determining the chain conformations.

The "ordered" structure on the fully substituted chain was not observed in the current study. The enhanced synergistic interactions between FG and NPP in higher concentration (1% and above) are believed to be caused by the intra-chain interactions along the side groups, which can form stronger "hyperentanglements". The

"hyperentanglements" themselves can increase the viscosity of the solution. Meanwhile, due to the higher flexibility of the fully substituted chains, the molecules may penetrate into the NPP networks, thus increase the overall gel strength. The overall synergistic behaviour of GMs and NPP can be a result of a combined effect of "smooth region" and

"hyperentanglements".

95 The current research project has not only provided scientific evidence for better understanding of the synergistic mechanisms, but has also broadened the possible application areas based on its properties being exhibited. In additions, the completion of the project has also obtained in-depth understanding of the polysaccharides in the molecular level. The significance of the project includes the following:

A unique polysaccharide, NPP, was obtained from yellow mustard mucilage. Its rheological property is of special interest. This polysaccharide demonstrated a stable gelling property over a wide temperature range, and enhanced gel strength when subjected to acidic environment. Our study also showed that strong acids, such as 1M sulphuric acid, can even stimulate the polymer solution to form a gel. These properties are unique and are not usually seen from any other types of natural polymers. Therefore, it may have great potential in some specific applications such as exposure to both acids and high temperatures.

Methylation of polysaccharides for NMR experiments was successfully applied in the current study. Methylation has been utilized in many areas, including structure analysis and polymer synthesis. It is rarely reported as a tool to improve quality of NMR spectra.

The idea behind this application is that modification of polymers can be a useful tool for investigation of its various properties, not only limited to linkage study, like methylation coupled with GC-MASS. This idea could be used in any polysaccharides with poor solubility in solvents, or polymers easy to form aggregates, which may lead to inaccurate estimation of its properties. It can also be applied in other techniques, such as light scattering. Some polymers with poor solubility cannot be subjected to Light Scattering.

Strong alkaline solvents are usually a solution for better dispersion of single molecules

96 although having the disadvantage of degradation. Modification of molecules could be an alternative solution for carrying out similar investigations such as determination of molecular weights while the extra weights from methyl groups could be adjusted by means of calculation.

The highlight of the current project is the mechanism study of synergistic interactions in blends of GMs and cellulosic polysaccharides. By acknowledging the results based on the previous findings by many other investigators, the current study made the elaboration on mechanism of synergism into a further depth, especially with the help of conformation study on GMs with various M/G ratios. In the current study, FG was recognized as an emerging material with superior characteristic especially when used in higher concentrations. By possessing health benefits, superior functional properties, and synergistic interactions with other polymers due to its fully substituted backbone, FG can be better explored in food or non-food industries.

A molecular modelling approach was applied in the current study to derive conformation parameters of simulated GM chains. The results were related to their solution properties and their synergistic behaviours in GM/NPP blends. This approach has paved the way for further research activities with respect to conformation- functionality relationships on various polymers of interest. With the upgraded computer capacity, the polymers with much higher molecular weights could be simulated. The simulated results could be compared with experimental data from other techniques, such as light scattering, for further verification. The actual interactions between molecules could also be "observed" on screen to help with understanding of the mechanism in a vivid way.

97 The overall contributions of the current project are (1) a better understanding of synergistic mechanisms was obtained in binary polysaccharide systems with the assistance of different advanced techniques; (2) the research approach developed here could be readily applied in a wide range of polymer studies; and (3) the information and knowledge can be easily transferred to industries to better explore the potential market of the two agri-products, both GMs and yellow mustard mucilage. We hope that our research outcome will benefit the polysaccharide-related value chain, from growers to processors, distributors and consumers thus bringing enhanced profits to the Canadian agriculture sector.

98 REFERENCES

Alloncle, M. and Doublier J.-L. (1991) Viscoelastic properties of maize starch/hydrocolloid pastes and gels. Food Hydrocolloids, 5, 455-467.

Baldrick, P. (2009) The safety of chitosan as a pharmaceutical excipient. Regulatory Toxicology and Pharmacology, available online 27 September, In Press.

Bergamini, J.-F., Boisset, C, Mazeau, K., Heyraud, A. and Taravel, F.R. (1995) Conformational behaviour of oligo-galactomannan chains inferred from NMR spectroscopy and molecular modelling. New Journal of Chemistry, 19, 115-127.

Blumenkrantz, N. and Asboe-Hansen, G. (1973) New method for quantitative determination of uronic acids. Analytical Biochemistry, 54, 484-489.

Boesewinkel, F.D. and Bouman, F. (1995) The seed: Structure and function. In J. Kigel, G. Galili, Seed Development and Germination, (pp 1-24). New York: Marcel Dekker.

Bresolin, T. M. B., Milas, M., Rinaudo, M., Reicher, F. and Ganter, J. L. M. S. (1999) Role of galactomannan conformation o the binary gel formation with xanthan. International Journal of Biological Macromolecules 26, 225-231.

Bresolin, T. M. B., Sander, P. C, Reicher, F., Sierakowski, M. R., Rinaudo, M. and Ganter, J. L. M. S. (1997) Viscometric studies on xanthan and galactomannan systems. Carbohydrate polymers 33, 131-138.

Chandrasekaran, R. and Radha, A. (1997) Molecular modeling of xanthan: galactomannan interactions. Carbohydrate Polymers, 32, 201-208

Ciucanu, I. and Kerek, F. (1984) A simple and rapid method for the permethylation of carbohydrates. Carbohydrate Research, 131,209-207.

Cross, M.M. (1965) Rheology of non-Newtonian fluids: a new flow equation for pseudoplastic systems. Journal of Colloid Science, 20, 417-437.

Cui, W. S. (2001) Yellow mustard gum. Chapter 1, p. 1-59. In Polysaccharide Gums from Agricultural Products: Processing, Structure & Functionality. Ed. Cui, S. Technomic Publishing Co., Inc. Lancaster, USA.

Cui, S. W., Eskin, N. A. M., Wu, Y. and Ding, S. (2006) Synergisms between yellow mustard mucilage and galactomannans and applications in food products - A mini Review. Advances in Colloid and Interface Science 128-130, 249-256.

99 Cui, W. S. (2001) Yellow mustard gum. In Polysaccharide Gums from Agricultural Products: Processing, Structure & Functionality. Chapterl: p. 1-59.Ed. Cui, S. Technomic Publishing Co., Inc. Lancaster, USA.

Cui, W. (1993) Characterization of the structure and physical properties of yellow mustard (Sinapis alba. L.) polysaccharides. Ph.D. thesis.

Cui, W. and Eskin, N.A.M. and Biliaderis, C.G. (1993) Chemical and physical properties of yellow mustard (Sinapis alba L.) mucilage. Food Chemistry, 46, 169-176.

Cui, W., Eskin, N. A. M., Biliaderis, C. G. and Mazza, G. (1995) Synergistic interactions between yellow mustard polysaccharides and galadctomannans. Carbohydrate Polymers 27, 123-127.

Dea, I.CM. and Morrison, A. (1975) Chemistry and interactions of seed galactomannans. Advances in Carbohydrates Chemistry and Biochemistry, 31, 241-312.

Dea, I.C.M., Clark, A.H. and McCleary, B.V. (1986) Effect of galacose-substitution- patterns on the interaction properties of galactomannans. Carbohydrate Research, 147, 275-294.

Dea, I.C.M., Morris, E.R., Rees, D.A., Welsh, E.J., Barnes, H.A and Price, J. (1977) Associations of like and unlike polysaccharides: Mechanism and specificity in galactomannans, interacting bacterial polysaccharides, and related systems. Carbohydrate Research, 57, 249-272.

Dickinson, E. & Galazka, V.B. (1992) Emulsion stabilization by protein/polysaccharide complexes. In G.O. Phillips, D.J. Wedlock & P.A. Williams, (Eds.), Gums and stabilizers for the food industry, 6, 351-362. Oxford, UK: IRL Press.

Dickinson, E. (2003) Hydrocolloids at interfaces and the influence on the properties of dispersed systems. Food Hydrocolloids, 17, 25-39.

Doublier, J.L. and Launay, B. (1981) Rheology of galactomannan solutions: Comparative study of guar gum and locust bean gum. Journal of Texture Studies, 12, 151-172.

Dowd, M.K., French, A.D. and Reilly, P.J. (1992) Conformational analysis of the anomeric forms of sophorose, laminarabiose, and cellobiose using MM3. Carbohydrate Research, 233, 15-34.

Dowd, M.K., French, A.D. and Reilly, P.J. (1995) Molecular mechanics modelling of-a- (1,2)-, a -(1,3)-, and a -(1,6) linked mannosyl disaccharides with MM3 (92). Journal of Carbohydrate Chemistry, 14, 589-600.

100 Dowd, M.K., Reilly, P.J. and French, A.D. (1992) Conformational analysis of trehalose disaccharides ans analogues using MM3. Journal of Computational Chemistry, 13, 102-114.

Dowd, M.K., Zeng, J., French, A.D. and Reilly, P.J. (1992) Conformational analysis of the anomeric forms of kojibiose, nigerose, and maltose using MM3. Carbohydrate Research, 230, 223-244.

Doyle, J.P., Lyons, G. and Morris, E.D. (2009) New proposals on "hyperentanglement" of galactomannans: Solution viscosity of fenugreek gum under neutral and alkaline conditions. Food Hydrocolloids, 23, 1501-1510.

Engelsen, S.B. and Rasmussen, K. (1993). Conformations of disaccharides by empirical force field calculations. Part V: Conformational maps of P-gentiobiose in an optimized consistent force field. International Journal of Biological Macromolecules, 1993,15,56-62.

Engelsen, S.B., Koca, J., Braccini, I., Herve du Penhoat, C. and Perez, S. (1995) Travelling on the potential energy surfaces of carbohydrates: Comparative application of an exhaustive systematic conformational search with an heuristic search. Carbohydrate Research, 276, 1-29.

Eskin, N.A.M., Raju, J. and Bird, R.P. (2007) Novel mucilage fraction of Sinapis alba L. (mustard) reduces azoxymethane-induced colonic aberrant crypt foci formation in F344 and Zucker obese rats. Phytomedicine, 14, 479-485.

Fernandes, P. B. (1995) Influence of galactomannan on the structure and thermal behaviour of xanthan/galactomannan mixtures. Journal of food Engineering 24, 269-283.

Fernandes, P. B., Goncalves, M. P. and Doublier, J. L. (1991) A rheological charaterization of kappa-carrageenan/galactomannan mixed gels: a comparison of locust bean gum samples. Carbohydrate Polymers 16, 253-274.

French, A.D., Rowland, R.S. and Allinger, N.L, in French A.D. and Brady, J.W. (Eds), (1990) Computer modelling of carbohydrate Molecules, ACS Symposium Series, 430, Americam Chemical Society, Washongton, DC, 430, 120-140.

Gaisford, S.E., Harding, S.E., Mitchell, J.R. and Bradley, T.D. (1986) A comparison between the hot and cold water soluble fractions of two locust bean samples. Carbohydrate Polymers, 6, 423-442.

Ganter, J.L.M. S., Heyraud, A., Petkowicz, C. L. O. , Rinaudo, M. and Reicher F. (1995) Galactomannans from Brazilian seeds: characterization of the

101 oligosaccharides produced by mild acid hydrolysis. International Journal of Biological Macronto lecules, 17, 13-19.

Gaonkar, A.G. (1991) Surface and interfactial activities and emulsion characteristics of some food hydrocolloids. FoodHydrocolloids, 5, 329-337.

Garcia, R. B. and Andrade, C. (1997) evidence of interaction between agarose and guar gum from changes in network response to solvent perturbation. Carbohydrate Polymers 34, 157-163.

Garti, N. & Reichman, D. (1994) Surface properties and emulsification activity of galactomannans. Food Hydrocolloids, 8, 155-173.

Garti, N. and Leser, M.F. (2001) Emulsification properties of hydrocolloids. Polymers for Advanced Technologies, 12, 123-135.

Garti, N., Madar, Z., Aserin, A. and Sternheim, B. (1997) Fenugreek Galactomannans as Food Emulsifiers. Food Science and Technology, 30, 305-311.

Giannouli, P., Richardson, R. K. and Morris, E. R. (2004) Effect of polymeric cosolutes on calcium pectinate gelation. Part 1. Galactomannans in comparison with partially depolymerised starches. Carbohydrate Polymers 55, 343-355.

Goff, H. D., Ferdinando, D., Schorsch, C. (1999) Fluorescence microscopy to study galactomannan structure in frozen sucrose and milk protein solutions. Food Hydrocolloids, 13, 353-362.

Guo, Q., Cui, S.W., Wang, Q. and Young J.C. (2008) Fractionation and physicochemical characterization of psyllium gum. Carbohydrate Polymers, 73, 35-43.

Ha, S.N., Madsen, L.J. and Brady, J.W. (1988) Conformational analysis and molecular dynamics simulations of maltose. Biopolymers, 27, 1927-1952.

Hall, L.D. and Yalpani, M. (1980) A high-yielding, specific method for the chemical derivatization of D-galactose-containing polysaccharides: oxidation with D-galactose oxidase, followed by reductive amination. Carbohydrate Research, 81, C10-C12.

Hannuksela, H. and Herve du Penhoat, C. (2004) NMR structural determination of dissolved O-acetylated galactomannan isolated from spruce thermomechanical pulp. Carbohydrate Research, 339, 301-312.

Harrington, J.C. and Morris, E.R. (2009) Conformational ordering and gelation of gelatoin in mixtures with soluble polysaccharides. Food Hydrocolloids, 23, 327-336.

102 Hoffman, J. and Svensson, S. (1978) Studies of the distribution of the D-galactosyl side- chains in guaran. Carbohydrate Research, 65, Pages 65-71.

Hefford, R. J. (1984) Polymer mixing in aqueous solutions. Polymer 25, 979-984.

Homans, S. W. (1990) Oligosaccharide conformations: Application of NMR and energy calculations. Progress in Nuclear Magnetic Resonance Spectroscopy, 22, 55-81.

Homans, S.W. and Forster, M. (1992) Application of restrained minimization, simulated annealing and molecular dynamics simulations for the conformational analysis of oligosaccharides. Glycobiology, 2, 143-151.

Huang, X., Kakuda, Y., and Cui, W. (2001) Hydrocolloids in emulsions: particle size distribution and interfacial activity. Food Hydrocolloids, 15, 533-542.

IUPAC-IUB Joint Commission on Biochemical Nomenclature (1997)). Nomenclature of Carbohydrates (Recommendations 1996). Carbohydrate Research, 297, 1-92.

Iwata, T., Azuma, J-L, Pkamura, K., Muramoto,M. and Chun, B. (1992) Preparation and n.m.r. assignments of cellulose mixed esters regioselectively substituted by acetyl and propanoyl groups. Carbohydrate Research, 224, 277-283.

Izydorczyk, M. (2005) Understanding the chemistry of food carbohydrates. In Cui, S.W., Food Carbohydrates: Chemistry, Physical Properties and Applications. (Chapter 1), ppl-66. Boca Raton: CRC Press.

Izydorczyk, M.S., Biliaderis, C. G. and Bushuk, W. (1991) Physical properties of water- soluble pentosans from different wheat varieties. Cereal Chemistry, 68, 145-150.

Izydorczyk, M., Cui, S.W. and Wanf, Q. (2005)Polysaccharide gums: Structures, Functional Properties and Applications. In Cui, S.W., Food Carbohydrates: Chemistry, Physical Properties and Applications. (Chapter 6), pp263-308. Boca Raton: CRC Press.

Jullien, R. and Botet, R. (1987) Aggregation and Fractal Aggregates, pp 77-102. Singapore: World Scientific.

Kapoor, V.P., Milas, M., Taravel, F.R., Rinaudo M. (1994) Rheological properties of seed galactomannan from Cassia nodosa buch.-hem. Carbohydrate Polymers, 25, 79- 84.

Kashyap, D.R., Vohra, P.K., Chopra, S. and Tewari, R. (2001) Applications of pectinases in the commercial sector: a review. Bioresource Technology, 11, 215-227'.

103 Khouryieh, H. A., Herald, T. J., Aramouni, F. and Alavi, S. (2007) Intrinsic viscosity and viscoelastic properties of xanthan/guar mixtures in dilute solutions: Effect of salt concentration on the polymer interactions. Food Research international 40, 883- 893.

Kohyama, K., Iida, H., and Nishinari, K. (1993) A mixed system composed of different molecular weights konjac glucomannan and kappa-carrageenan - large deforemation and dynamic viscoelastic study. Food Hydrocolloids 7, 213-226.

Kohyama, K., Sano, Y., and Nishinari, K. (1996) A mixed system composed of different molecular weights konjac glucomannan and kappa-carrageenan. 2. Molecular weight dependence of viscoelasticity and thermal properties. Food Hydrocolloids 10,229-238.

Launay, B., Cuvelier, G. and Martinez-Reyes, S. (1997) Viscosity of locust bean, guar and xanthan gum solutions in the Newtonian domain: a critical examination of the

log(nj;,)0-logc[ri]0 master curves. Carbohydrate Polymers, 34 (4), 385-395.

Leemann, A. and Winnefeld, F. (2007) The effect of viscosity modifying agents on mortar and concrete. Cement and Concrete Composites, 29, 341-349.

Li, W., Cui, W. and Kakuda, Y. (2006) Extraction, fractionation, structural and physical characterization of wheat P-D-glucans. Carbohydrate polymers, 63, 408-416.

Mannion, R. O., Melia, C. D., Launay, B., Cuvelier, G., Hill, S. E., Harding, S. E. and Mitchell, J. R. (1992) Xanthan/locust bean gum interactions at room temperature. Carbohydrate Polymers 19, 91-97.

Mao, C.-F. and Chen, J.-C. (2006) Interchain association of locust bean gum in sucrose solutions: An interpretation based on thixotropic behaviour. Food Hydrocolloids, 20 (5), 730-739.

Mazeau, K. and Rinaudo, M. (2004) The prediction of the characteristics of some polysaccharides from molecular modeling. Comparison with effective behaviour. Food Hydrocolloids 18, 885-898.

McCleary, B. V. (1979) Enzymic hydrolysis, fine structure, and gelling interaction of legume-seed of D-galacto-D mannans. Carbohydrate Research 71, 205-230.

McCleary, B.V, Clark, A.H., Dea, I.C.M., Rees, D.A. (1985) The fine structures of carob and guar galactomannans. Carbohydrate Research, 139, 237-260.

McClements, D.J. (2009) Biopolymers in Food Emulsions. Modern Biopolymer Science, 129-166.

104 Morris, E. R. (1990) Mixed polymer gels, in Food Gels, Chapter 8, p.291. ed. Harris, P.. Elsevier Applied Science, London, UK,.

Morris, E. R. and Foster, T. J. (1994) Role of conformation in synergistic interactions of xanthan. Carbohydrate polymers 23, 133-135.

Morris, E.R., Cutler, A.N., Ross-Murphy, S.B., Rees, D.A. and Price, J. (1981) Concentration and shear rate dependence of viscosity in random coil polysaccharide solutions. Carbohydrate Polymers, 1, 5-21.

Nakamura, A., Maeda, H. and Corredig, M. (2006) Emulsifying properties of enzyme- digested soybean soluble polysaccharide. Food Hydrocolloids, 20, 1029-1038.

Okada, Y., Koga, T., and Tanaka, F. (2002) Network Formation of Semi-Flexible Polymers by Hydrogen Bonds. Polymer Preprints, Japan, 51, 3306-3307.

Penroj, P., Mitchell, J. R., Hill, S. E. and Ganjanagunchorn, W. (2005) Effect of konjac glucomannan deacetylation on the properties of gels formed from mixtures of kappa carrageenan and konjac glucomannan. Carbohydrate Polymers 59, 367- 376.

Petkowicz, C. L. O., Milas, M., Mazeau, K., Bresolin, T., Reicher, F., Ganter, J. L. M. S. and Rinaudo M. (1999) Conformation of galactomannan: experimental and modelling approaches. Food Hydrocolloids, 13, 263-266.

Petkowicz, C.L.O., Reicher, F. and Mazeau, K. (1998) Conformational analysis of galactomannans: from oligomeric segments to polymeric chains. Carbohydrate Polymers, 37, 25-39.

Picout, D. R., Richardson, R. K. and Morris, E. R. (2000) Co-gelation of calcium pectinate with potato maltodextrin: Parti - Network formation on cooling. Carbohydrate Polymers 43, 133-141.

Picout, D. R., Richardson, R. K., Rolin, C, Abeysekera, R. M and Morris, E. R. (2000) Ca - induced gelation of low methoxy pectin in the presence of oxidised starch: Part 1 - Collapse of network structure. Carbohydrate Polymers 43, 113-122.

Pollard, M.A. and Fisher, P. (2006) Partial aqueous solubility of low-galactose-content galactomannans-What is the quantitative basis? Current Opinion in Colloid & Interface Science, 11, 184-190.

Rao, V.S.R., Qasba, P.K., Balaji, P.V. and Chandrasekaran, R. (1998) Conformation of Carbohydrates. Hardwood academic publishers, Amsterdam, The Netherlands.

105 Renard, C.M.G.C. and Jarvis, M.C. (1999) Acetylation and methylation of homogalacturonans 1: optimisation of the reaction and characterisation of the products. Carbohydrate Polymers, 39, 201-207.

Richardson, P.H., Clark, A.H., Russell, A.L., Aymard, P., and Norton, I.T. (1999) Galactomannan gelation: A thermal and rheological investigation analysed using the cascade model. Macromolecule, 32, 1519-1527.

Rinaudo, M. (2006) Chitin and chitosan: Properties and applications. Progress in Polymer Science, 31, 603-632.

Robinson, G., Ross-Murphy, S.B. and Morris, E.R. (1982) Viscosity-Molecular weight relationships, intrinsic chain flexibility, and dynamic solution properties of guar galactomannan. Carbohydrate Research, 107, 17-32.

Ross-Murphy, S. B., Morris, V. J. and Morris, E. R. (1983) Molecular viscoelasticity of xanthan polysaccharide. Faraday Symposium of the Chemical Society 18, 115- 129.

Schorsch, C, Gamier, C. and Doublier, J. (1997) Viscoelastic properties of xanthan/galactomannan mixtures: comparison of guar gum with locust bean gum. Carbohydrate polmers 34, 165-175.

Sittikijyothin, W., Torres, D. and Gonqalves, M.P. (2005) Modelling the rheological behaviour of galactomannan aqueous solutions. Carbohydrate Polymers, 59, 339-350.

Stortz, C. A. (1999) Disaccharide conformational maps: how adiabatic is an adiabatic map? Carbohydrate Research, 322, 77-86.

Stortz, C. A. (2006) Additive effects in the modeling of oligosaccharides with MM3 at high dielectric constants: an approach to the 'multiple minimum problem'. Carbohydrate Research, 341, 663-671.

Sudhakar, V., Singhal, R. S. and Kulkarni, P. R. (1995) Starch-galactomannan interactions: functionality and rheological aspects. Food Chemistry 55, 259-264.

Tako, M. (1991) Synergistic ineraction between xanthan and tara-bean gum. Carbohydrate Polymers 16, 239-252.

Tavares, C., Monteiro, S.R., Moreno ,N., , Lopes, J.A. and Tavares, S. (2005) Does the branching degree of galactomannans influence their effect on whey protein gelation? Colloids and Surfaces A: Physicochemical and Engineering Aspects, 270-271 (1): 213-219.

106 Tavares, C. and da Silva J.A.L. (2003) Rheology of galactomannan-whey protein mixed systems. International Dairy Journal, 13, 699-706.

Tolstoguzov, V. (2002) Thermal dynamic aspects of biopolymer functionality in biological systems, foods and beverages. Critical Reviews in Biotechnology 22, 89- 174.

Tran, V., Buleon, A., Imberty, A. and Perez, S. (1989) Relaxed potential energy surfaces of maltose. Biopolymers, 28, 679-690.

Van Caeseele, L., Kovacs M.I.P. and Gillespie R. (1987) Neutral sugar analysis of polysaccharides from the seed epidermis of Brassica campestris. Journal of the American Chemists' Society, 64, 761-762.

Van Caeseele, L., Mills, J.T., Sumner, M. and Gillespie, R. (1981) Cytology of mucilage production in the seed coat of candle canola (brassica campestris). Canadian Journal of Botany, 59, 292-300.

Viebke, C. and Piculell, L. (1996) Adsorption of galactomannans onto agarose. Carbohydrate Polymers, 29, 1-5.

Wang, Q., Wood, P.J., Huang, X. and Cui, W (2003) Preparation and characterization of molecular weight standards of low polydispersity from oat and barley (1-3) (1-4) -(3-D-glucan. Food Hydrocolloids, 17, 845-853.

Weber, F. E., Taillie, S. A. and Stauffer, K. R. (1974) Functional Characteristics of mustard mucilage. Journal of Food Science 39, 461-466.

Western T.L., Skinner, D.J. and Haughn, G.W. (2000) Differenciation of mucilage secretory cells of the Arabidopsis seed coat. Plant Physiology, 122, 345-355.

Whitney, S. E. C, Brigham, J. E., Darke, A. H., Reid, G. J. S and Gidley, M. J. (1998) Structural aspects of the interaction of mannan-based polysaccharides with bacterial cellulose. Carbohydrate Research 307, 299-309.

Wildman, R.E.C. (2001) Handbook of nutraceuticals and functional foods: Health & Fitness.

Wood, P.J. (2007) Cereal P-glucans in diet and health. Journal of Cereal Science, 46, 230-238.

Wood, P.J., Weisz, J. and Blackwell, B.A. (1994) Structural studies of (1-3) (l-4)-P-D- glucans by 13C-nuclear magnetic resonance spectroscopy and by rapid analysis of cellulose-like reqions using high-performance anion-exchange chromatography of oligosaccharides released by lichenase. Cereal Chemistry, 71, 301-307.

107 Wu, Y., Cui, W.S., Eskin, M. N. A. And Goff, D. (2009a) Fractionation and partial characterization of non-pectic polysaccharides from yellow mustard mucilage. Food Hydrocolloids, 23, 1535-1541.

Wu, Y., Cui, W. S, Eskin, M. N. A. And Goff, D. (2009b) Rheological investigation of synergistic interactions between galactomannans and non-pectic polysaccharide fraction from water soluble yellow mustard mucilage. Carbohydrate Polymers, 78,112-116.

Wu, Y., Cui, W.S., Eskin, N.A.M. and Goff, H.D. (2009c) An investigation on galactomannans: Effect of mannose/galactose ratio on emulsion and rheological properties. Food Research International, 42, 1141-1146.

Youssef, M.K., Wang, Q., Cui, S.W. and Barbut, S. (2009) Purification and partial physicochemical characteristics of protein free fenugreek gums. Food hydrocolloids, 23, 2049-2053.

Yumaguchi F., Ota, Y. and Hatanaka, C. (1996) Extraction and purification of pectic polysaccharides from soybean okara and enzymatic analysis of their structures. Carbohydrate Polymers, 30, 265-273.

Zandleven, J., Beldman, G., Bosveld, M., Schols, H.A. and Voragen, A.G.J. (2006) Enzymatic degradation studies of xylogalacturonans from apple and potato, using xylogalacturonan hydrolase. Carbohydrate Polymers, et al., 2006

Zhan, D.F., Brownsey, M.J., Ridout, M.J. and Morris, V.J. (1993) Xanthan-locust bean interactions and gelation. Carbohydrate Polymer 21, 53-58.

Zisenis, M. (1997) Correlation of molecular orientation and shear viscosity of high molar mass polymers in dilute solution. European Polymer Journal, 33(5), 773-780.

108