DEVELOPMENT OF A NEW DAIRY INGREDIENT FOR THE UTILIZATION OF BUTTERMILK CONSTITUENTS
DEVELOPMENT OF A NEW DAIRY INGREDIENT FOR THE UTILIZATION OF BUTTERMILK CONSTITUENTS Thèse Maxime Saffon Doctorat en Sciences et Technologie des Aliments Philosophiæ Doctor (Ph.D) Québec, Canada © Maxime Saffon, 2013 2 RESUME L’utilisation du babeurre pour la formulation alimentaire est limitée à cause de la capacité de rétention d’eau importante de ses phospholipides. L’objectif de ce projet était de développer une nouvelle approche de valorisation des constituants du babeurre. Ce sousproduit de l’industrie laitière est riche en composants d’intérêts qui ont des propriétés nutritionnelles, santés, et fonctionnelles prometteuses comme les constituants de la membrane de globule de matière grasse (MFGM) et les phospholipides. Les deux principaux procédés de production d’agrégats protéiques laitiers ont été combinés donnant un traitement thermique intensif à pH acide des protéines du lactosérum en présence des constituants du babeurre. Les résultats ont d’abord montré qu’il était possible de substituer différentes proportions de protéines du lactosérum par des protéines du babeurre et que la présence des constituants du babeurre entrainait la formation d’agrégats protéiques variés avec une capacité de rétention d’eau plus faible. Les résultats ont révélé que des agrégats protéiques étaient préformés lors de la préparation des babeurres incluant les protéines du lactosérum, les caséines et les protéines de la MFGM. Les phospholipides sont intégrés aux agrégats par l’intermédiaire de la MFGM à des températures faibles (65°C) alors qu’ils semblent s’associer directement avec les protéines à des températures plus élevées ( 80°C). Par la suite, les agrégats du babeurre agissent à titre de noyau d’agrégation pour les protéines issues du lactosérum. Le type d’interactions formées entre les protéines a un impact significatif sur les propriétés physiques et fonctionnelles des agrégats. En dernier lieu, il a été possible d’utiliser ces agrégats variés pour la production de yaourt ferme. Les agrégats lactosérum:babeurre ont agi à titre d’agent passif plutôt que d’agent actif mais des interactions entre les agrégats et les protéines du lait écrémé ont été observées. Ces associations seraient initiées par les groupements thiols libres des agrégats présents avant le chauffage. Cependant, la mise en solution de la poudre d’agrégats doit être strictement contrôlée. Ce projet propose une nouvelle approche pour l’utilisation du babeurre ainsi qu’une meilleure compréhension du comportement à la chaleur de ses constituants. iii iv ABSTRACT The use of buttermilk in food formulation is limited due to the extensive water-holding capacity of its phospholipids. The goal of this project was to develop a new approach for the valorization of buttermilk’s constituents. This by-product is rich in valuable components with promising nutritional, healthy, and functional properties such as the milk fat globule membrane (MFGM) constituents, the phospholipids. The two main processes of production of dairy aggregates have been combined resulting of the intensive heatdenaturation of whey proteins at low pH (4.6) in presence of proteins from buttermilk. First, results showed that it was possible to substitute whey proteins by different levels of buttermilk proteins in the process and that the presence of buttermilk constituents led to the formation of mixed aggregates with new functional properties such as a low water-holding capacity. Results revealed that aggregates are pre-formed during the preparation of the buttermilk concentrates involving whey proteins, casein, and MFGM proteins. Phospholipids are integrated to the aggregates through the MFGM at low temperature (65°C), but seem to directly interact with the proteins at higher temperatures ( 80°C). These pre-formed aggregates from buttermilk can act as aggregation nucleus for the proteins from whey. The types of interactions that occur between the proteins significantly affected the properties of the aggregates such as their water-holding capacity, their size, and the solubility of the powder. Finally, it was possible to use the mixed aggregates in the production of set-type yogurt. Whey:buttermilk aggregates were acting more like a passive than a reactive filler, but some possible interactions with the proteins from the skim milk were observed due to the high concentration of thiol groups of the aggregates before heating. However, the dispersibility of the powder must be strictly controlled. Overall, this project proposed a new approach for the use of buttermilk and allowed a better understanding of the thermal behavior of its constituents. v vi TABLE OF CONTENTS RESUME ........................................................................................................................................................ III ABSTRACT ..................................................................................................................................................... V TABLE OF CONTENTS.............................................................................................................................. VII TABLE LIST ................................................................................................................................................... XI FIGURE LIST .............................................................................................................................................. XIII EQUATION LIST.......................................................................................................................................XVII ABREVIATIONS ........................................................................................................................................ XIX ACKNOWLEDGMENTS............................................................................................................................ XXI FOREWORD .............................................................................................................................................. XXV CHAPTER 1 INTRODUCTION............................................................................................................... 1 CHAPTER 2 LITERATURE REVIEW ................................................................................................... 5 2.1 BUTTERMILK AND WHEY............................................................................................................................. 5 2.1.1 Buttermilk ...................................................................................................................................................... 5 2.1.2 Whey ................................................................................................................................................................. 7 2.2 CONSTITUENTS FROM BUTTERMILK AND WHEY ..................................................................................... 8 2.2.1 Caseins ............................................................................................................................................................. 8 2.2.2 Whey proteins ........................................................................................................................................... 10 2.2.3 Milk fat globule membrane proteins .............................................................................................. 13 2.2.4 Phospholipids from buttermilk ......................................................................................................... 17 2.3 PRINCIPLES OF PROTEIN AGGREGATION ................................................................................................ 19 2.3.1 Definitions of native state, denaturation, and reversibility ................................................. 19 2.3.2 Nucleation................................................................................................................................................... 20 2.3.3 Description of the different pathways .......................................................................................... 20 2.3.4 Bonds, interactions, and exchanges in milk protein aggregation .................................... 22 2.3.5 Effects of experimental conditions .................................................................................................. 24 2.4 HEAT-INDUCED AGGREGATION OF MILK PROTEINS ............................................................................. 26 2.4.1 Aggregation mechanisms of -lactoglobulin and -lactalbumin ................................... 26 2.4.2 Heat-induced interactions with minor whey proteins ........................................................... 31 2.4.3 Heat-induced interactions involving casein micelles.............................................................. 33 2.4.4 Heat-induced aggregation mechanisms of MFGM constituents ........................................ 36 2.5 APPLICATIONS FOR DAIRY PROTEIN AGGREGATES IN FOOD FORMULATION .................................... 42 2.5.1 Role of aggregates as fat mimetic ................................................................................................... 42 2.5.2 Role of aggregates as water holder ................................................................................................ 45 CHAPTER 3 HYPOTHESIS, GOAL, AND OBJECTIVES ................................................................. 47 CHAPTER 4 THERMAL AGGREGATION OF WHEY PROTEINS IN THE PRESENCE OF BUTTERMILK CONCENTRATE .............................................................................................................. 49 4.1 RÉSUMÉ ....................................................................................................................................................... 50 4.2 ABSTRACT ................................................................................................................................................... 51 4.3 INTRODUCTION .......................................................................................................................................... 51 vii 4.4 MATERIALS AND METHODS ...................................................................................................................... 53 4.4.1 Materials...................................................................................................................................................... 53 4.4.2 Preparation of whey and buttermilk concentrates ................................................................. 53 4.4.3 Heating experiments .............................................................................................................................. 54 4.4.4 Treatments with N-ethylmaleimide (NEM) and power ultrasound ................................ 56 4.4.5 Analytical methods ................................................................................................................................. 56 4.4.6 Particle size distribution and surface area ................................................................................. 58 4.4.7 Determination of rheological properties...................................................................................... 58 4.4.8 Statistical analysis .................................................................................................................................. 59 4.5 RESULTS ...................................................................................................................................................... 59 4.5.1 Aggregation yield .................................................................................................................................... 59 4.5.2 Water-holding capacity ........................................................................................................................ 61 4.5.3 Consistency coefficient (k) ................................................................................................................... 62 4.5.4 Flow behavior index (n)........................................................................................................................ 63 4.5.5 Particle size distribution ...................................................................................................................... 64 4.6 DISCUSSION ................................................................................................................................................. 66 4.7 CONCLUSIONS ............................................................................................................................................. 68 4.8 ACKNOWLEDGMENTS ................................................................................................................................ 69 CHAPTER 5 EFFECT OF BUTTERMILK COMPONENTS ON THE HEAT-INDUCED DENATURATION OF WHEY PROTEINS ...............................................................................................71 5.1 RÉSUMÉ ....................................................................................................................................................... 72 5.2 ABSTRACT ................................................................................................................................................... 73 5.3 INTRODUCTION .......................................................................................................................................... 73 5.4 MATERIALS AND METHODS ...................................................................................................................... 76 5.4.1 Materials...................................................................................................................................................... 76 5.4.2 Preparation of whey protein concentrates.................................................................................. 76 5.4.3 Preparation of spray dried buttermilk concentrates .............................................................. 77 5.4.4 Heat-induced aggregation .................................................................................................................. 78 5.4.5 Treatment with N-ethylmaleimide (NEM) .................................................................................. 78 5.4.6 Analytical method ................................................................................................................................... 78 5.4.7 Gel electrophoresis .................................................................................................................................. 79 5.4.8 Confocal laser scanning microscopy............................................................................................... 80 5.4.9 Statistical analysis .................................................................................................................................. 80 5.5 RESULTS AND DISCUSSION........................................................................................................................ 82 5.5.1 Composition ............................................................................................................................................... 82 5.5.2 Free thiol groups concentration ....................................................................................................... 83 5.5.3 Gel electrophoresis .................................................................................................................................. 85 5.5.4 Confocal laser scanning microscopy images............................................................................... 89 5.5.5 Proposed mechanism for association of whey proteins with buttermilk constituents 93 5.6 CONCLUSIONS ............................................................................................................................................. 94 5.7 ACKNOWLEDGEMENTS.............................................................................................................................. 95 CHAPTER 6 EFFECT OF HEATING OF WHEY PROTEINS IN THE PRESENCE OF MILK FAT GLOBULE MEMBRANE EXTRACT OR PHOSPHOLIPIDS FROM BUTTERMILK .......................97 6.1 RÉSUMÉ ....................................................................................................................................................... 98 6.2 ABSTRACT ................................................................................................................................................... 99 6.3 INTRODUCTION .......................................................................................................................................... 99 6.4 MATERIALS AND METHODS ................................................................................................................... 101 6.4.1 Materials....................................................................................................................................................101 viii 6.4.2 Isolation of the milk fat globule membrane .............................................................................. 101 6.4.3 Heat treatment ....................................................................................................................................... 102 6.4.4 Analytical method ................................................................................................................................. 102 6.4.5 Confocal laser-scanning microscopy ............................................................................................ 103 6.4.6 Statistical analysis ................................................................................................................................ 104 6.5 RESULTS AND DISCUSSION ....................................................................................................................104 6.5.1 Free thiol group concentration ....................................................................................................... 104 6.5.2 Gel electrophoresis ................................................................................................................................ 108 6.5.3 Confocal laser-scanning microscopy images ............................................................................ 109 6.5.4 Thin layer chromatography ............................................................................................................. 113 6.6 CONCLUSIONS...........................................................................................................................................115 6.7 ACKNOWLEDGEMENTS ...........................................................................................................................116 CHAPTER 7 EFFECT OF FREE THIOL GROUP REACTIVITY DURING THE FORMATION OF HEAT-INDUCED AGGREGATES FROM WHEY AND BUTTERMILK ON THEIR PROPERTIES 117 7.1 RÉSUMÉ .....................................................................................................................................................118 7.2 ABSTRACT .................................................................................................................................................119 7.3 INTRODUCTION ........................................................................................................................................119 7.4 MATERIALS AND METHODS ...................................................................................................................121 7.4.1 Materials.................................................................................................................................................... 121 7.4.2 Preparation of the whey and buttermilk concentrates ....................................................... 122 7.4.3 Preparation of protein aggregates powders ............................................................................ 123 7.4.4 Particle size distribution in heated liquid mixtures .............................................................. 124 7.4.5 Protein solubility index at pH 6.8 ................................................................................................... 125 7.4.6 Exposition of free thiol groups of the mixed aggregates .................................................... 125 7.4.5 Statistical analysis ................................................................................................................................ 125 7.5 RESULTS ....................................................................................................................................................126 7.5.1 Heat-induced aggregation characterization ........................................................................... 126 7.5.2 Properties of the mixed aggregates .............................................................................................. 128 7.6 DISCUSSION ..............................................................................................................................................130 7.7 CONCLUSION.............................................................................................................................................132 7.8 ACKNOWLEDGMENTS .............................................................................................................................133 CHAPTER 8 EFFECT OF SUBSTITUTION OF SKIM MILK POWDER BY WHEY:BUTTERMILK HEAT-DENATURED PROTEIN AGGREGATES IN MODEL SET-TYPE YOGURT 135 8.1 RÉSUMÉ .....................................................................................................................................................136 8.2 ABSTRACT .................................................................................................................................................137 8.3 INTRODUCTION ........................................................................................................................................137 8.4 MATERIALS AND METHODS ...................................................................................................................138 8.4.1 Materials.................................................................................................................................................... 138 8.4.2 Preparation of whey:buttermilk heat-denatured aggregates.......................................... 139 8.4.3 Yogurt production................................................................................................................................. 139 8.4.4 Analytical methods ............................................................................................................................... 140 8.4.5 Confocal laser scanning microscopy............................................................................................. 142 8.4.6 Statistical analysis ................................................................................................................................ 143 8.5 RESULTS AND DISCUSSION .....................................................................................................................144 8.5.1 Whey:buttermilk protein aggregates composition and properties ............................... 144 8.5.2 Particle size in skim milk ................................................................................................................... 145 8.5.3 Exposure of thiol groups .................................................................................................................... 146 ix 8.5.4 Distribution of particles in enriched skim milk before acidification .............................148 8.5.5 Texture of yogurts .................................................................................................................................148 8.5.6 Water-holding capacity of yogurt gels ........................................................................................151 8.5.7 Simulation of the gel formation ......................................................................................................153 8.7 CONCLUSIONS .......................................................................................................................................... 156 8.8 ACKNOWLEDGEMENTS........................................................................................................................... 156 CHAPTER 9 GENERAL CONCLUSIONS.......................................................................................... 157 9.1 ACHIEVEMENTS AND ORIGINAL CONTRIBUTIONS .............................................................................. 157 9.2 SIGNIFICANCE OF THE RESULTS ............................................................................................................ 163 9.3 QUESTION YET TO BE ANSWERED AND PERSPECTIVES ..................................................................... 164 CHAPTER 10 x LITERATURE CITED.............................................................................................. 167 TABLE LIST Table 1.1: Proposed applications for the use of buttermilk in food formulation (Vanderghem et al., 2010). ............................................................................................. 2 Table 2.1: Comparison between the gross composition of buttermilk and skim milk (Ramachandra Rao et al., 1995; Walstra et al., 2006). ................................................... 5 Table 2.2: Comparison between the gross composition of milk, acida, and sweetb cheese whey (Britten et al., 2002; Smithers, 2008). ................................................................... 8 Table 2.3: Summary of physical and chemical properties of whey proteins adapted from Cayot and Lorient (1998), Kinsella and Whitehead (1989a), Morr and Ha (1993), and Rüegg et al. (1977)........................................................................................................ 11 Table 2.4: Summary of physical and chemical properties of bovine milk fat globule membrane proteins adapted from Cheng et al. (1988), Dewettinck et al. (2008), Heid et al. (1996), Hvarregaard et al. (1996), Pallesen et al. (2001), Singh (2006), and Stammers et al. (2000). ................................................................................................. 15 Table 2.5: Lipid composition of buttermilk (adapted from Keenan and Dylewski (1995) and Walstra et al. (2006)...................................................................................................... 18 Table 2.6: Possible positive and negative effects of fat replacers (adapted from Senanayake and Shahidi (2005). ....................................................................................................... 43 Table 4.1: Composition (% DM) of the various whey-buttermilk mixtures concentrated at 9.5% protein (w/v) used for the heating experiments ................................................... 57 Table 4.2: Summary of the significance (P values) calculated by the analysis of variance of data of each contrasts for each variables (n = 3)........................................................... 59 Table 4.3: Average particle size and span of untreated and treated whey-buttermilk mixtures acidified at pH 4.6, denatured at 90°C and homogenized at 65.50 MPa. ...... 65 Table 5.1: Summary of the significance (P value) by the analysis of variance of data of each contrast for each variable (n = 3). ......................................................................... 81 Table 5.2: Composition (% DM) of whey protein concentrate (WPC) and the different buttermilk concentrates (BC) powders used for the heating experiment with rBC: regular buttermilk concentrate, rBCSFE: regular buttermilk concentrate after supercritical fluid extraction, wBC: whey buttermilk concentrate, wBCSFE: whey buttermilk concentrate after supercritical fluid extraction. ........................................... 82 Table 6.1: Summary of the significance calculated by the analysis of variance of free thiol group concentration (n = 3). ....................................................................................... 105 Table 7.1: Summary of the significance (P value) calculated by the analysis of variance of data of each contrasts for each variables (n = 3)......................................................... 126 Table 8.1: Summary of the significance of the effects of WBAP and time on the texture and water-holding capacity of yogurt gels, as calculated by analysis of variance (n = 3).143 Table 8.2: Summary of the significance of the effects of WBAP and time on the response of milk to heating, as calculated by analysis of variance (n = 3). ................................... 144 Table 8.3: Composition of heat-denatured whey:buttermilk aggregates (WBAP), whey permeate (WP), and powdered skim milk. ................................................................. 145 Table 8.4: Summary of the mean rupture force, adhesiveness and relaxation of yogurt gels, as a function of percent substitution of powdered skim milk with whey:buttermilk xi aggregates as enrichment of the starting milk. A Duncan post-test was applied to compare the means to the control. .............................................................................. 150 xii FIGURE LIST Figure 2.1: Schematic representation of the submicelle model of the casein micelle (represented by Horne (2006). ........................................................................................ 9 Figure 2.2: Schematic representation of bovine milk fat globule membrane (from JiménezFlores’ group). .............................................................................................................. 14 Figure 2.3: Schematic representation of major protein aggregation pathways..................... 20 Figure 2.4: Schematic representation of the currently proposed pathways of formation of the heat-induced whey proteins/-casein complexes in heated skim milk from Donato et al. (2009). .................................................................................................................. 34 Figure 4.1: Experimental procedure used to prepare aggregated buttermilk and whey proteins mixtures........................................................................................................... 55 Figure 4.2: Effect of protein composition, NEM and ultrasound on aggregation yield of different mixtures of whey and buttermilk concentrates (9.5 % protein), acidified at pH 4.6 and co-denatured at 90°C. ................................................................................. 60 Figure 4.3: Effect of protein composition, application of ultrasound and addition of NEM on water-holding capacity (WHC) of mixed buttermilk-whey protein mixture (9.5 % protein) heated at pH 4.6 (90°C). .................................................................................. 61 Figure 4.4: Effect of protein composition, application of ultrasound and addition of NEM on consistency index (k) of mixed buttermilk-whey protein mixture (9.5 % protein) heated at pH 4.6 (90°C). ............................................................................................... 62 Figure 4.5: Effect of protein composition, NEM and ultrasound on flow behavior indices of different mixtures of cheese whey and buttermilk concentrates (9.5 % protein), acidified at pH 4.6 and co-denatured at 90°C. .............................................................. 63 Figure 5.1: 15% Non-reducing SDS gels of the mixtures. 1: protein standard, 2: WPC, 3: WPC + rBC, 4: WPC + rBCSFE, 5: WPC + wBC, 6: WPC + wBCSFE. ........................ 83 Figure 5.2: Effects of time (a) and composition of the mixture (b) on the accessibility of free thiol groups during heating with constant stirring at temperature up to 90°C at pH 4.6. ................................................................................................................................ 84 Figure 5.3: Evolution of the protein profile of WPC as a function of heating time (min) under non-reducing (a) and reducing (b) conditions. ................................................... 86 Figure 5.4: Evolution of the protein profile of the WPC + rBC mixture as a function of heating time (min) under non-reducing (a) and reducing (b) conditions. ..................... 86 Figure 5.5: Evolution of the protein profile of the WPC + wBC mixture as a function of heating time under non-reducing (a) and reducing (b) conditions. ............................... 86 Figure 5.6: Evolution of the protein profile of the WPC (a), WPC + rBCSFE mixture (b), and WPC + wBC (c) in presence of NEM as a function of heating time under nonreducing conditions. ...................................................................................................... 88 Figure 5.7: Confocal laser scanning miscroscope pictures taken at 100X of WPC at 0, 15, and 25 minutes of heating before (top) and after (bottom) analysis with ImageJ. ....... 89 Figure 5.8: Confocal laser scanning microscope pictures taken at 100X of WPC + rBC mixture at 0, 15, and 25 minutes of heating before (top) and after (bottom) analysis with ImageJ. .................................................................................................................. 90 Figure 5.9: Evolution of the average surface area observed with a confocal laser scanning microscope at 100X as a function of interactions time*composition (a) and xiii time*condition. In a), mixtures were presented as follows: = WPC; = WPC + rBC; = WPC + rBCSFE; = WPC + wBC = WPC + wBCSFE. In b), conditions were presented as follows: = absence of NEM; = presence of NEM. ................ 91 Figure 5.10: Three-dimensional confocal laser scanning microscope pictures of WPC + rBC mixture before heating. ......................................................................................... 92 Figure 6.1: Effect of interactions composition*temperature (a) and temperature*pH (b) on the exposition of free thiol groups during heating of WPI (W.), WPI+MFGM extract (W.+M.), and WPI+phospholipid powder (W.+P.) under constant stirring. ............. 105 Figure 6.2: Pictures of heated WPI (a), WPI+MFGM (b), and WPI+PL (c) at pH 4.6 at 80°C and after 60 minutes of sedimentation in ice-melted water bath. ..................... 107 Figure 6.3: Evolution of the protein profiles in the different mixtures at 0 (a), 10 (b), 15 (c), 20 (d) minutes of heating at 80°C using SDS-PAGE under non-reducing conditions. Letters a to d correspond to samples at pH 6.8 and letters a' to d' to pH 4.6. ............ 108 Figure 6.4: Three-dimensional confocal laser-scanning microscope pictures of WPI heated for 25 minutes at 80°C at pH 4.6. ............................................................................... 110 Figure 6.5: Three-dimensional confocal laser-scanning microscope pictures of the WPI+MFGM extract heated for 15 minutes at 80°C at pH 4.6. ................................ 111 Figure 6.6: Three-dimensional confocal laser-scanning microscope pictures of WPI+PL heated for 15 minutes at 65°C at pH 4.6 .................................................................... 112 Figure 6.7: Three-dimensional confocal laser-scanning microscope pictures of WPI heated for 15 minutes at 80°C at pH 6.8. ............................................................................... 113 Figure 6.8: Evolution of the phospholipids profiles in WPI+PL at 0 (a), 10 (b), 15 (c), 20 (d), and 25 (e) minutes of heating time at 80°C. Picture was converted in negative mode in order to make the spot easier to see.............................................................. 114 Figure 7.1: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the number-weighted average particle size distribution in the mixtures. ................................................................................ 127 Figure 7.2: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the aggregation yield. .................... 128 Figure 7.3: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the water-holding capacity of the aggregates. .................................................................................................................. 128 Figure 7.4: Effect of the composition and the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on nitrogen solubility index of the powders. Aggregates formed in the absence of NEM are represented in black and aggregates formed in the presence of NEM are represented in grey. ........ 129 Figure 7.5: Effect of the composition and the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the accessibility of free thiol groups of the mixed aggregates. ................................................................. 130 Figure 8.1: Particle size distribution profile in skim milk after 5 min at 85°C, with added powdered skim milk substituted 0% () 20% (), 40% (), 60% (), 80% () and 100% () with WBAP. .............................................................................................. 146 xiv Figure 8.2: Effect of milk composition on thiol group exposure following heating at 85°C for 25 minutes (including ramp time). A Duncan post-test was applied to compare the means to the control. ................................................................................................... 147 Figure 8.3: Images of control (enriched with powdered skim milk) skim milk (a) and skim milk + WBAP aggregates (0.6 g/100 mL; b) at pH 6.8, using the Nikon C1 confocal laser scanning microscope at 60X. Proteins were dye in green and phospholipids in red. .............................................................................................................................. 148 Figure 8.1: Firmness of control set-type yogurt, and yogurts enriched with whey:buttermilk protein aggregates. A Duncan post-test was applied to compare the means to the control. ........................................................................................................................ 149 Figure 8.5: Water-holding capacity of set-type yogurt, and yogurts made from milk enriched with whey:buttermilk protein aggregates. .................................................... 152 Figure 8.6: Confocal laser scanning microscope (Nikon C1) images taken of gels of control skim milk (a) and skim milk + WBAP (0.6 g/100 mL; b) at 60X with a zoom of 2.31X; proteins were dyed green and phospholipids are red. Images a’ and b’ were processed with ImageJ to color the cluster in red. ...................................................... 153 Figure 8.7: General trends observed in the videos of gel formation as a function of the pH and time....................................................................................................................... 155 Figure 9.1: Schematic representation of the proposed mechanism of aggregation of proteins from WPC and the properties of the heat-induced protein aggregates. ...................... 161 Figure 9.2: Schematic representation of the proposed mechanism of aggregation of constituents from WPC + BC mixtures and the properties of the heat-induced protein aggregates. .................................................................................................................. 162 xv xvi EQUATION LIST Equation 4.1: Calculation of the aggregation yield .............................................................. 57 Equation 4.2: Calculation of the water-holding capacity of protein aggregates ................... 57 Equation 4.3: Formula of the Power Law Model ................................................................. 58 Equation 7.1: Calculation of the aggregation yield ............................................................ 124 Equation 7.2: Calculation of the water-holding capacity of the aggregates ....................... 124 Equation 7.3: Calculation of the protein solubility index at pH = 6.8 ................................ 125 Equation 8.1: Calculation of the relaxation of acid set-type yogurt ................................... 141 Equation 8.2: Calculation of the water-holding capacity of the gel ................................... 141 Equation 8.3: Calculation of the theoretical concentration of thiol group.......................... 142 xvii xviii ABREVIATIONS -LA: alpha-lactalbumin -LG: beta-lactoglobulin -CN: -casein ADPH: adipophilin BC: buttermilk concentrate BSA: bovine serum albumin BTN: butyrophilin CD36: cluster of differentiation 36 CLSM: confocal laser scanning microscope FABP: fatty-acid binding protein k: Consistency coefficient n: Flow behavior index MFGM: milk fat globule membrane MUC1: Mucin1 MWP: microparticulated whey proteins NEM: N-ethylmaleimide PAS 6/7: periodic acid Shiff 6/7 PAS III: periodic acid Schiff III PC: phosphatidyl choline PE: phosphatidyl ethanolamine PI: phosphatidyl inositol PL: phospholipids xix PP: permeate powder PS: phosphatidyl serine SDS-PAGE: sodium dodecyl sulfate polyacrylamide gel electrophoresis SM: sphingomyelin SMP: skim milk powder SH/SS: Thiol/Disulfide TLC: thin layer chromatography WHC: water-holding capacity WPC: whey protein concentrate WPI: whey protein isolate XDH/XO: xanthine dehydrogenase/oxidase xx ACKNOWLEDGMENTS This work represents an achievement of four years of research during which I crossed the path of many important people that I shall therefore acknowledge. First of all, I would like to express my sincere gratitude to my advisor Dr. Yves Pouliot. When I met Dr. Pouliot in 2007, I was a kid with a head full of dreams. With his trust, advice, continual support, confidence, and his guidance I grew up quickly in my thinking, the planning of my experiments, and the discussion of my results. Most of all, Dr. Pouliot always had a challenge waiting for me, and at the end the kid was able to realize all of his dreams. I feel very fortunate and I am very thankful for all the opportunities he gave me. As the song goes “Je ne peux pas te dire ce que je ne peux pas écrire. Il faudrait que j’invente des mots qui n’existent pas dans le dico” but one day I will find a way to express all of my thankfulness towards you. I would also like to express my gratitude to my co-advisor Dr. Michel Britten. Dr Britten has always been very generous in his advice, explanations, comments, corrections, and support. His powerful brain has been the cause of few sleepless nights, but I believe that working along Dr. Britten made me stronger in my thinking, the expression of my ideas, and definitely opened my eyes. I hope that our paths will cross again in the future and I will be always ready to discuss with you the different possibilities to explain one result. I would also like to express my gratitude to my co-advisor Dr. Rafael Jiménez-Flores. We often joke by saying that working along Dr. Jiménez-Flores would be the biggest challenge I faced during this four years of grad school. I found this to be true but it was also the most enjoyable challenge of all. He welcomed me in his arms (no second degree here) in 2010 and later in his scientific team in 2011. Dr. Jiménez-Flores has been a source of inspiration for me and definitely a source of joy. I am thankful for his guidance, his support, and the trust he put in me. I am also thankful for the opportunity to manage different projects, xxi advise different candidates in their studies, meet people from industries, and present my results all over California. I am also thankful toward Dr. Jean-Christophe Vuillemard for accepting to do the prereading of this thesis. Dr. Vuillemard has always been very supportive toward this project since day one. He has been generous many times with his comments and his suggestions for the advancement of my experiments. Many thanks to Dr. Fanny Guyomarc’h for joining the committee of my thesis. Dr Guyomarc’h’s work has been a source of inspiration during the planning of many parts of this project. I am thankful toward Drs. Laurent Bazinet, Sylvie Gauthier, and Sylvie Turgeon. They have been very supportive during the last four years regarding my advancements in my scientific and personal life. Of course, nothing would have been done without the help of some person. Many many many thanks to the extraordinary Diane Gagnon form the STELA Research Group for teaching me all the lab techniques that I know, for her patience, her availability, her support, and of course her candies … Many thanks to the people from the DPTC to welcome me, for their patience, their understanding, their help, their support, and for never have laughing at my French accent. Many thanks to the interns that have worked at my side: Justine, Eva, Véronique, Sylvain, Camille et Mélanie. I would like to express all of my gratitude to Frederic Lehance. Frederic was the first person to make me appreciate Quebec City and the Quebec experience in general. He supported from day one (back in 2007) in my studies and in my personal life. He has always been here for me the last six years, listening to my problems, giving me support and advice. He is the belgio-punko big brother that I always wanted. I also must acknowledge the incredible support I had from my friends at home (particularly Dr. Alexandre). I established very strong relationships since I moved in Canada, and I xxii would like to express my gratitude to my very best friends Dr. Jeremie and Dr. Romain who have always been here for me. The expression “away from the eyes, away from the heart” is not in their vocabulary, so I never felt alone and lost during my Californian experience. Many thanks to the very good friends that I made during this journey: Benjamin, Laure, Bertrand, Cyril, Ryan, Mason, BJ, Anthony, and David. I am very grateful to have met my lovely girlfriend Melissa. Her support, comprehensiveness, understanding, encouragement gave me the strength to pursue and complete my degree. Thanks my love for being at my side, even in this scary cold French speaking part of the world. Many thanks to her family for being so supportive and for letting her come in Quebec City. I am very fortunate to have a wonderful family. I have an enormous respect for them because letting go of their little boy was probably not easy, but they always expressed their understanding and their gigantic support. I am very proud of you and would like to express all of my love to you Danièle & Lucien, Céline & Julien & Justine & Nathan, Marthe & Dominique, Marie-Rose & Hubert. Last but not least, I am thankful towards Le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT) – Novalait Inc. – Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ) for their financial support. xxiii xxiv FOREWORD The work carried out in this thesis was aimed to develop a new approach to use whey and buttermilk, and to understand the interactions that occur between components of both liquids. This thesis contains all the results obtained during the course of realization of the project. Results are presented in the form of five published, submitted, or in preparation research articles. The first article «Thermal aggregation of whey proteins in the presence of buttermilk concentrate» has been submitted and published in «Journal of Food Engineering N°103» in 2011. Authors are Maxime Saffon, Michel Britten, and Yves Pouliot. The second article «Effect of buttermilk components on the heat-induced denaturation of whey proteins» is in preparation for a submission in «Journal of Agriculture and Food Chemistry». Authors are Maxime Saffon, Rafael Jiménez-Flores, Michel Britten, and Yves Pouliot. The third article «Effect of heating of whey proteins in the presence of milk fat globule membrane extract or phospholipids from buttermilk» is pending for submission. Authors are Maxime Saffon, Rafael Jiménez-Flores, Michel Britten, and Yves Pouliot. The fourth article «Effect of buttermilk constituents on the properties of heat-induced protein aggregates» is pending for submission. Authors are Maxime Saffon, Rafael Jiménez-Flores, Michel Britten, and Yves Pouliot. The fifth article «Effect of substitution of skim milk powder by whey:buttermilk heatdenatured aggregates in model set-type yogurt» is in preparation for a submission in the Special Issue «Structure of Flavor of Dairy Products» of Foods. Authors are Maxime Saffon, Rafael Jiménez-Flores, Michel Britten, and Yves Pouliot. xxv For each article, Maxime Saffona planned and achieved the experiments, presented the results, and wrote the article. In consequence, he is the first author of all of them. Dr. Jiménez-Floresb, co-supervisor of this project, participated to the planning, the discussion of the results, and the revision of the writing from the second to the fifth articles. He is second author on these papers. Dr. Michel Brittenc, co-supervisor of this project, participated to the planning, the discussion of the results, and the revision of the writing of all articles. He is third author on all of them. Dr. Yves Pouliota, supervisor of this project, participated to the planning, the discussion of the results, and the revision of the writing of all articles. He is the last author on all of them. a STELA Dairy Research Center, Institute of Nutrition and Functional Foods (INAF), Laval University, Quebec City, QC, Canada, G1V 0A6. b Dairy Products Technology Center (DPTC), California Polytechnic State University, San Luis Obispo, CA, USA, 93405. c Food Research and Development Center (FRDC), Agriculture and Agri-Food Canada, St-Hyacinthe, QC, Canada, J2S 8E3. xxvi A Danièle, Lucien, Céline Et mes grands-parents «Les portes de l’avenir s’ouvrent à ceux qui savent les pousser» -Michel Colucci 1 CHAPTER 1 INTRODUCTION The worldwide consumption of milk or dairy products is constantly increasing. It is estimated that the production of milk increases about 1.5% every year in order to satisfy the needs of consumers. The increase of consumption of dairy products such as cheese or butter also increases the production of dairy by-products such as whey or buttermilk. As an example, the production of whey and buttermilk in Canada in 2008 has been calculated around 3,535,000 tons and 82,600 tons, respectively (Statistics Canada, 2009). Over the years, the opinion about these by-products has changed. It is now recognized that whey and buttermilk constituents have interesting biologic, nutritional properties, and health-benefits. Many applications have been developed to use whey in food formulation, such as in cheese, bakery, pastry, and delicatessen productions. Unfortunately, buttermilk has not found as many industrial applications as whey yet. Recently, Vanderghem et al. (2010) summarized ten potential applications for buttermilk or buttermilk constituents that have been tested in research labs over the last 20 years (Table 1.1). Buttermilk is a unique product due to its concentration of milk fat globule membrane (MFGM) and associated material (proteins, phospholipids, and sphingolipids) that have been associated with very promising health properties ranging from anti-viral to anticancer. Buttermilk contains other constituents with potential food applications such as caseins (75% of the total proteins), whey proteins (8 to 15% of the total proteins), and polar lipids. However, it has been demonstrated that the exploitation of the buttermilk constituents is complicated due to some irreversible changes that appear along the process. Previous work of Morin (2006) showed that the pasteurization of the cream is critical because it affects the solubility at pH 4.6 of MFGM proteins, modifies the surface of the fat membrane, and increases the accessibility of phospholipids. Phospholipids have an important capacity to hold water, so the increase of their exposure limits the use of buttermilk in dairy products such as cheese or yogurt. Above all, pasteurization of cream 1 has been correlated to the poor coagulation properties of buttermilk. Upon the treatment, MFGM proteins initiate heat-induced interactions with other proteins. Table 1.1: Proposed applications for the use of buttermilk in food formulation (Vanderghem et al., 2010). Sources Desired Properties Fresh buttermilk milk sweet Viscosity agent Cheddar cheese buttermilk Condensed Effects Agent for the heat Recombined evaporated Stability increased stability Ultrafiltered Applications Higher apparent viscosity sweet Agent for the yield, Pizza cheese cream buttermilk Higher cheese texture, meltability, moisture and coagulation Decreased Moisture retention yield, content. chewy, Reduced meltability and gel strength Buttermilk Moisture retention Bread Higher water absorption, Agent for the dough Increased resistance to and extension, sensory properties Better sensory score Whey proteins are also very sensitive to heating. A process has been developed to use whey protein aggregates in food formulation. Some of their industrial applications are, however, limited due to the water-holding capacity of denatured whey proteins. Combination of buttermilk and whey could be the key to better exploit their constituents in food formulation by reducing their individual negative effects such as extensive water retention. It is now clear that caseins, whey proteins, and MFGM proteins interact together under heating. It can be thought that it is possible to control the composition and properties of aggregates by using different ratio and/or experimental conditions. 2 The goal of this work was to develop a new approach to use buttermilk constituents based on the knowledge of the thermal behavior of its constituents in order to obtain new potential dairy ingredients with texture modification applications for food formulation. 3 4 CHAPTER 2 LITERATURE REVIEW 2.1 Buttermilk and whey 2.1.1 Buttermilk Buttermilk is the aqueous phase expulsed after the formation of large butter grains during churning of cream (Boudreau & St-Amant, 1984). As air is incorporated during the buttermaking process, proteins from the cream unfold and form unstable foam with the air bubbles. Upon mechanical stress, the foam is destabilized resulting of fat clumping and phase inversion. Butter becomes solid while the water and soluble particles form the buttermilk. Most of the protein, minerals, lactose, and water from the cream are recovered in the buttermilk (Table 2.1). A high portion of the MFGM is also present in the buttermilk. Table 2.1: Comparison between the gross composition of buttermilk and skim milk (Ramachandra Rao et al., 1995; Walstra et al., 2006). Constituents Buttermilk Total Solids (%) 9.50 – 10.50 Skim Milk 9.40 Lactose (%) 3.60 – 4.30 4.80 Proteins (%) 3.30 – 3.90 3.36 Fat (%) 0.30 - 0.70 0.07 Phospholipids (%) Ash (%) 0.07 – 0.18 0.55 – 0.90 0.015 0.88 Churning 1 kilogram of cream will typically yield 0.5 kilogram of butter and 0.5 kilogram of buttermilk. Liquid buttermilk is prone to oxidation, so this product is concentrated by evaporation and then sprays dried in order to extend the shelf life. Currently, dried buttermilk is mainly used for animal feeding (swine, bovine), but also used in the food industry for its emulsion properties and its positive impacts on the texture and taste of products. 5 Buttermilk is the richest dairy product in phospholipids. In consequence, addition of phospholipids (through substitution of milk by buttermilk) has been tested for the production of low-fat cheese (Turcot et al., 2002; Turcot et al., 2001). Addition of buttermilk increased the cheese yield, but notable changes in the texture and taste of cheese have been recorded. These results have been correlated to the high water-holding capacity of phospholipids. Authors concluded that the amount of phospholipids should be carefully calculated before addition in cheese. They suggested enriching the milk for the cheesemaking process (40 gram of proteins per kilogram of milk) with less than one-gram of phospholipids per kilogram of milk to make sure that the texture or taste will not be affected. Earlier, Mistry et al. (1996) have reported that it is possible to substitute milk with liquid buttermilk prior to the cheese-making process without changing the properties of the cheese if the level of substitution is under five percent. Later, Raval and Mistry (1999) observed that cheeses had a lower fat content (lower free oil) and harder bodies if the milk was supplemented with buttermilk (5%). However, cheese had a lower meltability, and a higher apparent viscosity comparatively to the control cheese. 2.1.1.1 Specialized «biorefinery» approach As described above, buttermilk is rich in minor components associated with the MFGM fragments (proteins, phospholipids). These components have demonstrated promising functional properties and health benefits. Morin (2006) focused on the development of an approach for the separation of MFGM material using microfiltration. Their results showed that the similitarity between the size of the casein micelle and the MFGM fragments made the separation very challenging. Later, they demonstrated that separation can be improved by addition of a cream washing step prior to the microfiltration in order to remove the casein micelles. However, Bédard ST-Amand (2009) have shown that the industrial HTST pasteurization of the cream is critical for the MFGM fragments. Their results showed that several proteins (probably located at the surface of the bilayer) have broken off the fragments. In 6 opposition, some of the available Bovine Serum Albumin and β-lactoglobulin proteins are associated with the MFGM fragments. Regarding the difficulties to separate buttermilk constituents and the modification that occur during the butter-making process (solubilization of some MFGM proteins), a global approach for the utilization of buttermilk seems more appropriate. 2.2.1.2 Global approach As described in Table 1.1, over ten applications have been proposed for the utilization of buttermilk in food formulation. The proposed applications are ranged from an agent for moisture retention in bread to a viscosity agent for cheddar cheese. It is also possible to vary the composition of buttermilk for the valorization of a specific constituent. For example, Costa et al. (2010) have proposed an approach to produce a whey buttermilk powder enriched in milk fat globule membrane phospholipids. Whey buttermilk is a very unique product that is richer in MFGM material comparatively to regular buttermilk. By combining ultrafiltration (10X) and CO2-supercritical fluid extraction (350 bars; 50°C) they obtained a powder containing 73% of whey proteins and 21% of lipids of which 61% were phospholipids. 2.1.2 Whey Whey is the aqueous phase separated from casein network formed through chymosin or mineral/organic acid during cheese-making or casein manufacture process. Depending on the cheese production, whey components represent approximately 50% of the initial milk solids with 20% of the proteins and 100% of the lactose (Table 2.2). Whey has been considered as a waste product during many years and as one of the most polluting food byproducts due to its high content in lactose (> 75% of the dry basis). 7 Table 2.2: Comparison between the gross composition of milk, acida, and sweetb cheese whey (Britten et al., 2002; Smithers, 2008). Total solids (%) Milk Acid whey Sweet whey 12.80 6.60 6.80 Lactose (%) 4.90 4.40 4.90 Total proteins (%) 3.50 0.70 0.70 Caseins (%) 2.80 < 0.1 < 0.1 Whey proteins (%) 0.70 0.70 0.80 Fat (%) 3.70 0.06 0.15 Ash (%) 0.70 0.37 0.31 However, whey represents a rich and heterogeneous content of protein with nutritional, biological and food attributes that give to it potential applications in food processing. Different technologies have been developed to concentrate (35 to 80% of protein) or isolate (> 90% of protein) whey proteins such as ultrafiltration or ultrafiltration/diafiltration. Whey protein concentrates (WPC) or isolates (WPI) have a relatively low value, and are widely produced in dairy industry. These powders have different applications in food processing industries where water binding and texturization are required, such as gelation, thermal stability, foam formation, and emulsification. Their uses vary from manufactured meats to formulated foods or reformed food products (Chatterton et al., 2006; Foegeding et al., 2002). 2.2 Constituents from buttermilk and whey 2.2.1 Caseins Caseins are the predominant proteins in buttermilk with around 75% of the total protein content (Walstra et al., 2006). Caseins are very different comparatively to other proteins. They have little secondary and tertiary structures, and are hydrophobic with a high charge. The exposed hydrophobic groups give the protein a strong ability for self-interactions or association with other proteins (Walstra et al., 2006). 8 In a milk system, caseins are present as colloidal particles named casein micelles. Caseins micelles are composed of four major proteins: S1 (33%), (33%), S2 (11%), (11%), plus calcium phosphate (8 g.100g-1 of casein), a small portion of proteose peptone and some enzymes (Brulé et al., 1997; Walstra et al., 2006). Casein micelles have an average diameter ranged from 40 to 300 nanometers and have a negative charge. The presence of casein micelles in milk is very important because it determines the physical stability of milk products during process and storage and the viscosity of products concentrated in proteins (Walstra et al., 2006). Over the years, various structural models have been proposed for the casein micelles. The most enduring model has been initiated by Schmidt (1980), and consists of submicelles connected together via hydrophobic bonds and calcium phosphate (Figure 2.1). Figure 2.1: Schematic representation of the submicelle model of the casein micelle (represented by Horne (2006). Each submicelle would be composed of 20 to 25 casein molecules, and would have an average diameter ranges from 12 to 15 nanometers. The submicelles located at the outside 9 of the micelles would contain one or two -caseins. The «hairy layer» of the casein micelles would be formed by the C-terminal end of (~ 75 amino acid residues). The hairy layer would be hydrophilic, negatively charged and would provide colloidal stability to the casein micelle (Walstra et al., 2006). -casein molecules are mostly present as oligomer of about 120 kDa. Each oligomer contains from 5 to 11 monomers. -casein differs from the other caseins, due to the presence of two cysteine residues that form an intermolecular disulfide bond (Cys11-Cys88), and because are involved in SH/SS exchanges with whey proteins. In buttermilk, caseins are also mostly found in micellar form. However, some variation in the size of the casein micelle and its properties due to the processing can be expected. Morin et al. (2008) found that the casein micelles from buttermilk have poor coagulation properties due to the cream pasteurization. As describe later, heat-induced interactions occur between -lactoglobulin and -casein located at the surface of the micelles. 2.2.2 Whey proteins Whey proteins represent between 8 to 15% of the total proteins in buttermilk (depending of the churning process) (Sodini et al., 2006), while they represent almost 100% of the proteins in whey. The two major whey proteins are -lactoglobulin (-LG) and -lactalbumin (-LA), and they represent from 70 to 80% of the total protein content. The minor proteins are serum albumin (BSA), immunoglobulin (Ig), proteose peptone (PP), lactoferrin (LF) and lactoperoxydase (LP), and they represent about 20 % of the total proteins (Table 2.3). 10 Table 2.3: Summary of physical and chemical properties of whey proteins adapted from Cayot and Lorient (1998), Kinsella and Whitehead (1989a), Morr and Ha (1993), and Rüegg et al. (1977). Proteins Concentration (g.L-1) Mol. Weight (kDa) pI S-S (SH) TU (°C) 2.0 – 4.0 18.4 5.3 2 (1) 72.8 -LG 1.0 – 1.5 14.2 4.8 4 65.2 -LA Ig 0.6 – 1.0 0.015 to1.0 5.5 – 6.8 BSA 0.1 – 0.4 66.0 5.1 17 (1) 62.2 PP 0.3 – 0.8 4.0 – 40.0 4.1 – 5.6 0 ND LF 0.1 80.0 8.4 – 9.0 17 65 / 90 LP 0.01 – 0.03 78.0 9.5 6 (3) ND pI = isoelectric point; TU = unfolding temperature; ND = not determined. Bovine -lactoglobulin represents about 50 to 55% of total whey proteins with a concentration of 2.0 to 4.0 g.L-1 of milk (Morr et al., 1993). The protein is present as a dimer of 36,400 Da at a neutral pH (Sawyer & Kontopidis, 2000), and starts to dissociate into two identical monomers at a pH above 6.5 or below 3.5 (Hambling et al., 1992; Sawyer et al., 2000; Townend et al., 1969). Monomeric -lactoglobulin is a small acidic globular protein of 162 amino acids with a molecular weight around 18,300 Da (slightly different with genetic variants). The secondary structure of the protein consists of three short helices, an -helix, and nine strands of anti- parallel -sheets (Kinsella et al., 1989a). Hydrophobic, ionic, and hydrogen-bond interactions between the peptide chains stabilize the secondary and tertiary structure of the protein. In the native form, the disulphide bridges Cys66-Cys160 and Cys106-Cys119 also stabilize its structure, and a free thiol group (free –SH) is available in position Cys121 (Kinsella et al., 1989b). In the native form of the protein, the free thiol group and the bond Cys106-Cys119 are in the hydrophobic cleft (Sawyer, 2003). The free thiol group is mainly responsible for the thermal irreversible unfolding and aggregation of the protein. -lactoglobulin is a resistant carrier of retinol, palminate, fatty acids, vitamin D and cholesterol as review by de Wit (1998) and Madureira et al. (2007). Foegeding et al. (2006) have demonstrated that -lactoglobulin has a comparable foam overrun capacity to egg white (albumin) for the formation of meringues, angel food cake or similar products. Their results brought the idea that WPC, WPI or -lactoglobulin isolate could provide a cost-effective alternative to egg albumin in food formulation. The high nutritional value and health benefits (anti-hypertensive, anti-cancer, hypocholesterolemic, opiodergic, and anti-microbial), and the variety of its functional properties has already 11 made -lactoglobulin a primary choice for modern food and beverage formulation (Chatterton et al., 2006). Bovine -lactalbumin represents about 22% of total whey proteins, and it is the second major protein with a concentration of 1.0 to 1.5 g.L-1 of milk (Morr et al., 1993). lactalbumin is also a small globular protein of 123 amino acids with a 14,200 Da molecular weight. Four disulphide bridges stabilize its structure in position Cys6-Cys120, Cys28-Cys111; Cys61-Cys77, and Cys73-Cys91 (Morr et al., 1993). -lactalbumin does not have a free thiol group. The biological function of -lactalbumin is to support the biosynthesis of lactose (de Wit, 1998). Bovine -lactalbumin is an excellent source of the essential amino acids tryptophan and cysteine, and its high homology to human -lactalbumin made it useful for the development of infant formulas. Bovine serum albumin is present at about 0.1 to 0.4 g.L-1 in cow milk (Morr et al., 1993). Bovine serum albumin is one of the biggest whey protein with 582 amino acid residues, and a molecular weight around 66,000 Da (66,267). This protein possesses seventeen intermolecular disulphide bonds, and a free thiol group at the residue Cys34. The unfolded protein (rupture of intramolecular disulphide bonds) exposes hydrophobic side groups initially buried to the aqueous phase. Bovine serum albumin binds free fatty acid for transportation in the blood, and is an important source for the production of glutathione in the liver (de Wit, 1998; Morr et al., 1993). The immunoglobulins (Ig) are an heterogenous family of glycoproteins from 15,000 to 1,000,000 Da, and are present at the concentration of 0.6 to 1.0 g.L-1 in cow milk (Morr et al., 1993). Korhonen et al. (2000) have reported that bovine serums contain three major classes of immunoglobulins that are IgG, IgM, and IgA with a similar structure: two identical light chains (23 kDa) and two identical heavy chains (53 kDa) connected together through disulphide bonds. Approximately 80% of the bovine immunoglobulins are IgG (1 or 2). Immunoglobulins are antibodies that act as a carrier of passive immunity to the newborn, and as an antimicrobial agent (de Wit, 1998; Madureira et al., 2007). 12 The proteose peptone (PP) fraction of milk is an heterogeneous group of phosphoglycoproteins formed following proteolysis of the N-terminal region in the sequence of -casein by plasmin, and possibly other milk proteins, lipoproteins, and proteolipids. These molecules are amphiphilic because of their charged phosphate groups and sequences of hydrophobic amino acid residues (Kinsella et al., 1989a). Lactoferrin (LF) is present at 0.02 to 0.35 g.L-1 in bovine milk and 2.0 to 5.0 g.L-1 in human milk. Lactoferrin is a monomeric glycoprotein of 689 amino acids residues and has a molecular weight of around 75,000 Da (76,110). The protein possesses seventeen intermolecular disulphide bonds, but no free thiol group (Kinsella et al., 1989a). Lactoferrin is a transport agent for ferric iron (Fe3+), so it makes iron more available for absorption in the gut. Lactoferrin also has antibacterial activity (de Wit, 1998). Lactoperoxidase (LP) is present at the concentration of 0.01 to 0.03 g.L-1 of milk. It is one of the most abundant enzyme in whey (0.25 to 0.5 % of the total whey proteins). Lactoperoxidase is a single peptide chain containing 612 amino acid residues. Lactoperoxidase has 4 to 5 potential sites for N-glycosylation. The protein also possesses 15 half cystines (Cals et al., 1991), and a ferric iron (Fe3+). Lactoperoxidase is a main defensive system agent in mammals because of its antimocrobial and antiviral properties. The protein provide a strong protection against invading micro-organisms (Madureira et al., 2007). 2.2.3 Milk fat globule membrane proteins The bovine milk fat globule membrane (MFGM) represents between 2% and 6% of the total mass of the fat globule and is composed of a complex mix of proteins, glycoproteins, phospholipids, triglycerides, cholesterol, enzymes, and minor constituents. The composition of the MFGM varies as a function of the size of the fat globule, the breed, the health status of the cow, or the stage of lactation (Keenan & Mather, 2006). The protein 13 content of the MFGM varies from 25% to 60% depending on the MFGM extraction techniques and the analytical methods. Over 40 proteins have been identified. For a long time, MFGM proteins have been classified according to their relative migration during polyacrylamide gel electrophoresis in presence of SDS. The nomenclature of the MFGM has been clarified by the Milk Protein Nomenclature Committee of the American Dairy Science Association and reported by Mather (2000). The main proteins of bovine MFGM are: Mucin 1 (MUC1), Xanthine dehydrogenase/oxidase (XDH/XO), Periodic acid Schiff III (PAS III), Cluster of Differentiation (CD36), Butyrophilin (BTN), Adipophilin (ADPH), Periodic acid Schiff 6/7 (PAS 6/7), and Fatty-acid binding protein (FABP) (Singh, 2006) (Table 2.4). All these proteins are located through the membrane in specific positions depending on their physical and physiological properties (Figure 2.2). The overall isoelectric point of the MFGM was determined around 4.8 (Kanno & Kim, 1990). At higher pH, the MFGM will be charged negatively. Figure 2.2: Schematic representation of bovine milk fat globule membrane (from Jiménez-Flores’ group). 14 Table 2.4: Summary of physical and chemical properties of bovine milk fat globule membrane proteins adapted from Cheng et al. (1988), Dewettinck et al. (2008), Heid et al. (1996), Hvarregaard et al. (1996), Pallesen et al. (2001), Singh (2006), and Stammers et al. (2000). Proteins Percent of the Molecular Weight MFGM proteins (kDa) pI -SS- (SH) Td (°C) (%) MUC1 n.f. 160 to 200 < 4.5 0 (0) n.f. XDH/XO 20 150 7.7 11 (38) < 60 PAS III 5 95 to 100 n.f. n.f. n.f. CD36 5 76 to 78 n.f. 3 (0) n.f. BTN 20 to 43 67 5.0 to 5.4 1 (0) 58 ADPH n.f. 52 7.5 to 7.8 n.f. n.f. PAS 6/7 n.f. 47 - 52 5.6 to 7.6 9 (0) > 80 FABP n.f. 13 n.f. n.f. n.f. pI = isoelectric point; Td= temperature of denaturation; n.f. = not found Mucin 1 (MUC1) is a mucin-type glycoprotein that is associated with the cream fraction and is found at a concentration of 40 mg.L-1 of milk. It is a polymorphic protein of 160,000 Da to 200,000 Da. Mucin 1 is strongly bound to the serum-exposed side of the MFGM, but can also become soluble after a short heat-shock. In this case, the protein is soluble in the serum. Bovine MUC1 possesses one cysteine residue in the membrane-spanning region and three cysteine residues in the cytoplasmatic tail (Pallesen et al., 2001). Snow et al. (1977) reported that the protein contains 50% of carbohydrate (w/w) including sialic acid (30.5%), N-acetylglucosamine (22.3%), galactose (15.9%), N-acetylgalactosamine (14.0%), mannose (11.1%), and fucose (5.8%). The protein is acidic with an isoelectric point under 4.5, due to its high content in sialic acid. The protein also possesses a highly glycosylated filamentous portion located outside the MFGM. The role of Mucin 1 has been uncertain for many years, however Kvistgaard et al. (2004) and Mather (2000) reported that the protein has a protective effect against physical damages and rotavirus. These effects are attributed to the filamentous portion of the protein. 15 Xanthine dehydrogenase/oxidase (XDH/XO) is the most abundant enzyme of the MFGM. The protein is present in the form of a homodimer of 300,000 Da, where each monomer is estimated to be around 148,000 Da. Twenty-two disulphide bonds reinforce the structure of the protein. XDH/XO also contain thirty eight thiol residues, four of them are available and thirty four are buried in the structure (Cheng et al., 1988). XDH/XO is a peripheral-attached protein that is generally considered as weakly bound to the membrane, because it is possible to remove approximately 60% of the total protein using appropriate buffers. In opposition, the remaining 40% are firmly attached to the membrane (Mather, 2000). The role of XDH/XO has been uncertain for many years, but it is believed that the protein could have an antimicrobial function in the gut due to production of hydrogen peroxide. Periodic acid Schiff III (PAS III) is a glycoprotein of 95,000 Da to 100,000 Da that has been poorly characterized. However, it is known that PAS III is present in the serumexposed portion of the membrane (Mather, 2000). Cluster of Differentiation (CD36) is an integral protein of the MFGM of 76,000 Da to 78,000 Da. The protein possesses ten cysteines residues (Berglund et al., 1996). Six of the ten form the disulfide bonds Cys242-Cys310, Cys271-Cys332, and Cys312-Cys321, and the other four are acylated near the intracellular side of the membrane (Rasmussen et al., 1998). The protein does not have a free thiol group. CD36 is strongly attached to the membrane; even a centrifugation treatment does not permit to remove the protein from the membrane. CD36 could act as a scavanger receptor by binding to apoptotic cells and cell fragments and precipitating their elimination by phagocytosis (Mather, 2000). Butyrophilin (BTN) is the most abundant protein in bovine MFGM and is a protein of 67,000 Da. The protein is strongly attached to the membrane and can resist to a centrifugation treatment or extraction with chaotropic agents and detergents (Mather et al., 1977). The structure of the protein is stabilized by a disulphide bond between the two antiparallel -strands (Stammers et al., 2000). BTN denatures at a temperature of 58°C and 16 forms aggregates through non-native disulfide cross-bridges at the same temperature (Appel et al., 1982). Adipophilin (ADPH) is a protein of 52,000 Da strongly attached to the membrane even after extraction with salts and non-ionic detergents (Heid et al., 1996). ADPH has not been widely studied, however it is known that the protein could conserve a hydrophobic binding region that could mediate associations with lipid droplets or eventually proteins. ADPH binds five to six molecules of fatty acids per molecule of proteins (Heid et al., 1996). Periodic acid Schiff 6/7 (PAS 6/7) are glycoproteins of 52,000 Da and 47,000 Da respectively (Hvarregaard et al., 1996). Nine disulphide bridges stabilized the structure of the proteins (Cys6-Cys17; Cys11-Cys29; Cys31-Cys40; Cys48-Cys59; Cys53-Cys76; Cys78-Cys87; Cys234-Cys238; Cys252-Cys409; Cys91-Cys247), but the proteins have no free thiol group. PAS 6/7 are poorly attached to the membrane because the proteins can be present as soluble form in skim milk or buttermilk (Mather, 2000). PAS 6/7 is probably the most heat-stable MFGM because the proteins are stable until 80°C (Ye et al., 2002). Fatty-acid binding protein (FABP) is the smallest protein of the MFGM with a molecular weight of 13,000 Da that is mainly poorly attached to the membrane (Brandt et al., 1988). 2.2.4 Phospholipids from buttermilk The lipid content of buttermilk usually ranges from 0.3% to 0.7%. Compositions of lipids in buttermilk and in milk fat globule membrane have been well characterized as shown in Table 2.5. Buttermilk is richer in polar lipids such as phospholipids, sphingolipids, and gangliosides with a content ranging from 0.07% to 0.18% of the total lipids comparatively to whole milk (0.035% of the total lipids), skim milk (0.015% of the total lipids) or whey (0.02% of the total lipids). In bovine milk, about 50% to 60% of the phospholipids are integrated to the milk fat globule membrane (fragmented or not), and represent between 26% to 31% of the total lipids of the membrane (McPherson & Kitchen, 1983; Singh, 17 2006). The most common classes of phospholipids found in the membrane are Sphingomyelin (SM), Phosphatidyl choline (PC), Phosphatidyl ethanolamine (PE), Phosphatidyl inositol (PI), Phosphatidyl serine (PS), Lysophosphatidyl choline and represent 22%, 36%, 27%, 11%, 4%, and 2% of the total phospholipids from the MFGM, respectively. These classes generally contain high levels of esterified long-chain fatty acid (C16; C18:1), but low levels of short-chain fatty acid (C4 to C10) (McPherson et al., 1983). The MFGM components and particularly the phospholipids have been extensively investigated over the last twenty years due to their possible roles in nutrition or human health. Agents against colon cancer, gastrointestinal pathogens, Alzheimer’s disease, depression, stress, support for the recovery of the liver from toxic chemical attack or chronic viral are different examples of health-benefits of the phospholipids from the MFGM (Dewettinck et al., 2008; Spitsberg, 2005). It has also been reported that SM and PC are important sources of choline that is important for the synthesis and transmission of important neurotransmitters and even maybe for the brain development. Table 2.5: Lipid composition of buttermilk (adapted from Keenan and Dylewski (1995) and Walstra et al. (2006). Constituents In Buttermilk Constituents (% of dry basis) In MFGM (% of total lipids) Total lipids 2.9 to 7.4 Triglycerides 62 Phospholipids (PL) 0.6 to 1.9 Diglycerides 9 18 PE 42.9% of PL Sterols 0.2 to 2.0 PI 8.9% of PL Free fatty acids 0.6 to 6.0 PS 8.6% of PL Phospholipids 26 to 31 PC 19.1% of PL SM 17.9% of PL CER 7.7% of PL 2.3 Principles of protein aggregation Wang et al. (2010) defined protein aggregates as a protein in a non-native state whose size is at least twice as that of the native protein. In chemistry, biochemistry, and pharmacology the term oligomers or critical oligomers is preferred (Modler et al., 2003). Individual protein molecules have the capacity to fold or (partially) unfold many times. These changes depend on environmental conditions such as intrinsic (structure of protein) or extrinsic factors (pH, temperature, pressure, etc) and lead to an energetically unfavorable unfolded state. This situation is due to the exposition of hydrophobic regions. To regain stability, protein molecules refold again or self-associate to form aggregates. Aggregation results in the absence of normal functionality of the protein. Protein aggregation has been well-studied because a large number of human disorders, ranging from type II diabetes to Parkinson’s and Alzheimer’s diseases, are associated with protein aggregation (Dobson, 2001; Mattson et al., 1999). Aggregates can also induce a cellular toxicity when they are present in organs such as the liver, heart, and brain (Varadarajan et al., 2000). 2.3.1 Definitions of native state, denaturation, and reversibility The «native state» of a protein corresponds to its operative or functional form. In their native states, most of the proteins are folded into rigid and well-defined three-dimensional structures. In protein aggregation, the term “native” is used to define the initial state of the protein before denaturation. “Denaturation” refers to a major change from the original native structure that occurs without severance of any of the primary chemical bonds (Tanford, 1968). However, this definition is subject to controversy, such as the meaning of a major change. 19 Tanford (1968) simply defined the “reversibility” as the ability of the protein to return (reversible) or to not return (irreversible) to the native conformation. It means that the reaction is reversible only if the native structure can be recovered. In most cases, the process becomes irreversible due to a secondary reaction that follows a major conformational change (i.e unfolding). These reactions can be induced by the use of temperature, acid or basic pH, etc. 2.3.2 Nucleation Most of the protein aggregation mechanisms are nucleation-dependent and, in consequence, are initiated by the formation of an aggregation nucleus (Wang et al., 2010). There is no definition of an aggregation nucleus in the literature, however, Krishnan (2003) defined a nucleus as small as the size of a dimer. In opposition, Baynes et al. (2005) observed a small multimer as a nucleus. In the case of -lactoglobulin, Schokker et al. (1999) observed the presence of non-native dimers and oligomers in the early stages of the heat-induced denaturation. Later, non-native monomers, denatured proteins, and/or small aggregates are incorporated at the surface of the nucleus to form large aggregates. This second step is named as aggregation, elongation, fibrillation, or polymerization in the literature (Wang et al., 2010). 2.3.3 Description of the different pathways Wang et al. (2010) summarized the major protein aggregation pathways as presented in Figure 2.3. Figure 2.3: Schematic representation of major protein aggregation pathways 20 Aggregation through unfolding intermediates and unfolded states is the first pathway and named (1) in Figure 2.3. Under normal conditions, a solution of proteins is a balance between proteins in a native state and a small amount of unfolding intermediate proteins (I in Figure 2.3). The latter of the two is divided into native-like intermediate or unfolded-like intermediate depending on the degree of unfolding. These intermediate proteins are considered as precursors of the physical aggregation process because they expose more hydrophobic areas and are more flexible than native proteins. Interactions of these intermediates lead to the formation of aggregates (A in Figure 2.3). Later in the process, aggregates interact with each other. As soon as aggregates reach a certain size or solubility limit, they become insoluble (P in Figure 2.3). Many proteins do not go through the intermediate state. These proteins can directly form aggregates physically from the native state through self-association (pathways 2a in Figure 2.3). These associations are due to electrostatic, hydrophobic, or van der Waals interactions. Self-association is defined as the formation of reversible aggregates and is considered as the precursor of irreversible aggregates. In fact, the distinction of the formation of intermediates and the self-association can be difficult to demonstrate. So, Wang et al. (2010) reminded that formation of intermediates is related to conformational stability and self-association is related to colloidal stability. Aggregation can also be due to chemical denaturation and chemical linkage between protein chains (pathways 2b in Figure 2.3). Formaldehyde-mediated cross-linking, dityrosine formation, oxidation, and Maillard reactions have been reported by Wang et al. (2010) as possible cross-linking pathways. Nevertheless, the most common chemical linkage is the intermolecular disulfide bond exchange. Surface-located and internally located cysteines are both involved in the formation of disulfide bonds. Protein aggregation through chemical degradation is the last major pathway (pathways 3 in Figure 2.3). The most known examples are auto-oxidation, oxidation (Rosenfeld et al., 21 2009), dimerization (Roostaee et al., 2009), hydrolysis (Van Buren et al., 2009), and glycation (Wei et al., 2009). These reactions can change protein hydrophobicity, secondary/tertiary structures, or the kinetics of protein unfolding. Theoretically, other pathways can lead to formation of protein aggregates. Proteins that are largely unfolded in natural conditions or proteins that have only two states (N or D) can directly form aggregates from their denatured state (pathways 2a’ in Figure 2.3). The denatured state can also undergo chemical degradation and form aggregates either directly (pathway 2b’ in Figure 2.3) or indirectly (pathway 3’ in Figure 2.3). Of course, the same protein can aggregate by different pathways, but one pathway often dominates and should be controlled first. Protein aggregation can be reversible or irreversible. The reversibility of aggregation is due to its ability to disaggregate and mainly depends on the stage of the aggregation process (early stage = reversible; late-stage = irreversible), environmental conditions (heat often induces irreversible aggregation), and the size of the aggregates. 2.3.4 Bonds, interactions, and exchanges in milk protein aggregation Both covalent and non-covalent interactions are involved in the heat-induced protein aggregation reactions in milk. The types of covalent bonding that occur during the heatinduced aggregation of milk proteins are thiol-thiol oxidation (also called formation of a disulfide or disulfide bonds), and mainly thiol/disulfide exchange (SH/SS). The first one appears in aerobic biological systems, and is characterized by a formal two-electron oxidation of two thiols (-SH) to a disulfide (-SS-) with the reduction of oxygen (Gilbert, 1990): 2 R-SH R-SS-R + 2e- + 2 H+ (1) Later, Gilbert (1993) defined the thiol/disulfide exchange reaction by the nucleophilic attack of a thiol on one of the two sulfurs of a disulfide bond to oxidize one thiol while 22 reducing another disulfide. This reaction occurs with an intra-molecular –SH group or an intermolecular protein free –SH group. RSH + R’SSR’ RSSR’ + R’SH (2) RSH + RSSR’ RSSR + R’SH (3) This reaction depends on the differences in stability between the two disulfides and the two thiols, and is generally reversible. The exchange will become irreversible in some processes such as strong heat treatments where the reversible oxidation of one or more -SH groups precedes the irreversible oxidation of one or more additional -SH groups (Cappel & Gilbert, 1986). It is generally accepted that the SH/SS exchanges occur during the heatinduced denaturation of -lactoglobulin. The free thiol group of the protein (Cys121) attacks one of the two-disulfide bonds of another protein (Cys106-Cys119 or Cys66-Cys160). A free thiol groups is now available in the second protein. Following this mechanism, Livney and Dalgleish (2004) have found SH/SS exchanges between the residues Cys66 and Cys160 from -lactoglobulin with residues Cys11 and Cys88 from -casein, respectively. Non-covalent bonds are critical in maintaining the three-dimensional structure of large molecules such as proteins and are also involved in the intermolecular binding of large molecules. There are four types of non-covalent bonds: hydrogen bond, ionic bond, van der Waals forces, and hydrophobic interactions. The role that plays these interactions is still uncertain in the heat-induced aggregation of milk proteins. Hydrophobic interactions seem to play an important role for whey protein aggregation, but less comparatively to SH/SS exchanges. The unfolded proteins expose both thiol and hydrophobic groups that were initially buried in the three-dimensional structure of the -lactoglobulin. Results of Oldfield et al. (2000) showed that 10 to 15% of the -lactoglobulin available in their system aggregated through hydrophobic interactions during the first step of their heat-induced denaturation at 85°C, then followed by SH/SS exchanges. Controversially, Galani and Apenten (1999) concluded that the hydrophobic interaction become important only at temperatures between 90°C and 110°C. 23 2.3.5 Effects of experimental conditions Temperature is often judged as the most critical factor that affects protein aggregation. Changing the environmental temperature (increasing or decreasing) will destabilize the proteins and will eventually promote protein aggregation. Aggregation can occur even at room temperature. In the case of milk protein aggregation, temperature and holding time will determine the level and the rate of interactions. At lower temperatures (75°C to 85°C), the rate of aggregation between whey proteins and casein micelles increases slowly. At higher temperatures (90 to 100°C) the rate increases rapidly at first and then reaches a plateau (Anema & Li, 2003a). Reducing activation energy, enhancing hydrophobic exchanges, increasing diffusion of proteins, and increasing the frequency of the molecular collisions are other examples of the effects of variation of the environmental temperatures (Wang et al., 2010). Physico-chemical environment has a strong effect on protein aggregation and particularly the ionic strength and the protein concentration. However pH is estimated the most critical one. Variations of pH will change the distribution of surface charges on proteins, and directly affect intramolecular folding interactions, intermolecular protein-protein interactions, and the rate of protein aggregation (Wang et al., 2010). At high basic or acid pH, proteins are heavily charged so the repulsion forces are important. Controlling the pH of the buffer (to reach the pI of the proteins) will reduce the repulsion forces, increase the attraction forces, or positively affect the intramolecular and intermolecular interactions between proteins. Anema and Li (2003b) showed that the interaction between whey proteins and casein micelles were more important at pH 6.5 comparatively to pH 6.7. Donato and Guyomarc’h (2009) reported later in their review that ~ 100% of the whey proteins associate with the casein micelle at pH 6.2, while (60 to 85%) associate at pH 6.4 and only (10 to 15%) at pH 6.9 during heating of skim milk for several minutes at temperatures ranging from 80 to 120°C. Wang et al. (2010) also explained that proteins can complex with both positive and negative ions present in the solution through electrostatics interactions. These associations can change the surface charge, and consequently affect the charge-charge interactions 24 during protein aggregation. Livney et al. (2003) found that a high ionic strength promotes heat-induced disulfide bonding between the two major whey proteins. The composition of the protein solution also affects the denaturation of proteins. It has been demonstrated that lactose and some other soluble non-protein components present in milk have a protective role against unfolding of -lactoglobulin. Spiegel (1999) demonstrated that the rate of thermal unfolding of -lactoglobulin (< 85°C) slows down by increasing the content of lactose. Their results showed that a denaturation degree of 0.8 was obtained after 40 seconds at 80°C at a lactose concentration of 1.5%, while 35 minutes were needed at the lactose concentration of 13.5%. Arakawa and Timasheff (1982) explained that sugars change the surface tension of water. Consequently, proteins stabilize their structure by preferentially interacting with water or soluble components. Later, Spiegel and Huss (2002) obtained the same results by varying the amount of calcium in their system. Once again, they referred the increase of the heat stability in the presence of calcium to the theory of the preferential interaction of proteins with water. Similar results and explanations were given by Anema et al. (2006). These authors found that, in general, soluble non-protein components and lactose delay the irreversible denaturation of -lactoglobulin and lactalbumin. On the other hand, they reported that increasing the protein concentration increases the rate of denaturation. However, impact of the soluble non-protein components is stronger comparatively to the concentration of proteins, but only for -lactoglobulin. This result means that in concentrated milk the denaturation rate of -lactoglobulin is still delayed. Wang et al. (2010) explained that increasing the protein concentration has three possible consequences: a decrease of aggregation phenomena due to crowding effect; an increase of aggregation phenomena due to better chance of association; or a precipitation due to solubility limit. But in most cases, increasing protein content promotes proteinprotein association as reported by Anema et al. (2006). 25 2.4 Heat-induced aggregation of milk proteins 2.4.1 Aggregation mechanisms of -lactoglobulin and lactalbumin 2.4.1.1 Thermal behavior of whey proteins The thermal behavior of whey proteins, especially -lactoglobulin, is a complex phenomenon that depends on the experimental conditions (temperature, pH, time, ionic strength). To illustrate the complexity, the following paragraph only present the result from heating experiment up to 150°C at pH > 6.8 (De Wit, 2009). At 20°C and at pH near to 6.0, -lactoglobulin is present as a dimer. The dimer starts to dissociate into two identical monomers at pH above 7.5 or below 3.5 (Townend et al., 1969), and/or at temperature up to 55°C. At pH ranging from 6.0 to 8.5, -lactoglobulin undergoes a reversible change of conformation that corresponds to a discrete change in the specific optical rotation of the protein (Tanford et al., 1959). This conformational modification of the tertiary structure appears from the native to the reversible conformation of the protein, and results in exposure of the thiol group (Cys121) (Dunnil & Green, 1965). The transition from native form to unfolded form is negligible at pH < 6.8 and at 50°C, but represents 18 % at pH 7.0 (Pantaloni, 1964). At temperatures between 60 to 70°C, more reversible conformational changes appear. Pantaloni (1964) found that irreversible modifications characterized by the exposition of the free thiol group (Cys121) appear at 70°C, but at low amount (6 %). Iametti et al. (1996) concluded that the monomer form of -lactoglobulin has a poorly defined tertiary structure at 70°C. Following this irreversible structure modification, monomers expose large hydrophobic surfaces and a reactive thiol group. They defined this conformation as “molten globule state”. Prabakaran and Damodaran (1997) reported the exposure of the buried 26 Cys121 and the presence of the disulphide bond Cys106-Cys119 at temperature from 63°C to 65°C due to a melting of the E-H -strand region of the molecule. The irreversible unfolding step of -lactoglobulin appears at temperature range from 70°C to 80°C depending of the pH value, and more specifically close to 78.5°C (Schokker et al., 1999). Results also showed that a reactive monomer (unfolded) forms a reactive dimer with another reactive monomer via thiol-disulphide exchange, thiol-thiol oxidation, and noncovalent interactions (Prabakaran et al., 1997; Schokker et al., 1999). When the concentration of reactive dimer becomes important a polymerization reaction appears. Results of Schokker et al. (1999) showed that the reactive -lactoglobulin monomers form non-native dimers and oligomers in the first step of the aggregation. Then, the formation of dimers and trimers occurs as a function of the heating time. The size of aggregates increased supposedly by addition of reactive monomers or dimers to small aggregates. Their results also showed the presence of pentamers consisting of 100 -lactoglobulin units. The accessibility of the free thiol groups increases by increasing the temperature (> 80°C). At these temperatures, protein aggregation occurs via both thiol/disulphide exchange and hydrophobic interactions (Sawyer, 1968). Residual native or reformed secondary structures of -lactoglobulin unfold completely at temperature above 125°C may be due the breakdown of disulphide bonds (Watanabe & Klostermeyer, 1976). 2.4.1.2 Mechanisms of aggregation According to their model, Roefs and de Kruif (1994) proposed that the heat-induced denaturation of β-lactoglobulin processes via a three-step mechanism: initiation, propagation, and termination. The model is based on a thiol-catalyzed aggregation mechanism where the free and active thiol groups (-SH) acts as a radical, and quickly and continuously polymerize -lactoglobulin monomers. The initiation step starts by reversible reactions, and then by a (pseudo) irreversible reaction. The reversible reactions consist of 27 the separation of dimers (B2) in monomers (B) at temperature below 70°C. They defined the irreversible reaction by the exposition of the free thiol group of the unfolded βlactoglobulin (B*): B2 … B => B* (4) A thiol/disulfide exchange is formed between a disulfide bond of a “native” monomer and the free thiol group of an unfolded β-lactoglobulin as described in 1.4. The monomer then exposes a free thiol group and can initiate a SH/SS exchange with another protein, etc. Results of Creamer et al. (2004) demonstrated that the disulfide bond Cys66-Cys160 has been broken during the heat treatment, and that Cys160 became free. In consequence, the group Cys160 probably plays an important role for the polymerization process. The propagation step is linear because it involves only one free thiol and one disulfide bond per monomer. However, the mechanism can be repeated many times until two reactive multimers (Bi* and Bj*) form a disulfide bond by thiol-thiol oxidation: B + Bi* => Bi+1* i1 (5) Bi* + Bj* => Bi+j i,j 1 (6) As recalled by Oldfield et al. (1998), the unfolding of β-lactoglobulin is also likely to expose some hydrophobic residues ([B]). The following reactions could also occur under specific environmental conditions: B2 … B => [B] (7) [B] + [B] => [Bi+1] i 1 (8) [Bi+1] => Bi+1* i 1 (9) The results of their study did not permit the confirmation of this phenomenon, but later Oldfield et al. (2000) reported that 15% to 20% of the total β-lactoglobulin were present as hydrophobic formed aggregates at 85°C. However, the portion decreased to 0% with the increase of the temperature and heating time. According to their results, they concluded that 28 hydrophobic interactions between proteins could occur during the initial step of the aggregation of β-lactoglobulin, and that aggregates became stronger by internal disulfide bonds. Hoffmann and Van Mill (1997), Sawyer (1968), and Xiong et al. (1993) have also demonstrated that it is possible to form whey protein aggregates through hydrophobic interactions if the free thiol groups are blocked by a chemical reactant such as Nethylmaleimide (NEM). Oldfield et al. (1998) also proposed a heat-induced aggregation for -lactalbumin (A). The unfolded protein does not expose free thiol groups, so aggregates are formed mainly through hydrophobic interactions. However, reactions are reversible because no aggregates are found in solution at room temperature after heating (Gezimati et al., 1997). The possible reaction involved the presence of hydrophobic sites of the unfolded protein ([A]) at temperatures between 75°C to 85°C: A [A] (10) [A] + [A] [Ai+1] (11) [Ai+1] + [A] … [An] (12) In a complex solution such as milk, -lactalbumin can form a heat-induced protein complex with β-lactoglobulin by both hydrophobic interactions ([AB]) and SH/SS exchanges (AB*). Gezimati et al. (1997) concluded that the free thiol group of βlactoglobulin can initiate SH/SS exchange reactions that incorporate both proteins in the final complex. Once again, the process could involve a propagation, and termination step: A + B2 A + B => [A] + [B] + B* (13) [A] + B* or [B] => [AB] => (AB)i* (AB)i+j* (14) Over the years, other mechanisms have been proposed. The summary of the different heatinduced aggregation mechanisms of whey proteins involves a general trend corresponding to the dissociation of dimers into monomers, exposition of a free thiol group by the unfolded β-lactoglobulin, and formation of aggregates through both SH/SS exchange reactions and non-covalent bonds. It is also important to take into consideration the 29 experimental conditions. Prior, it has been discussed that factors such as temperature, pH, ionic strength, composition, etc., strongly affect the denaturation behavior of proteins. For example, in their study Roefs and de Kruif (1994) used a buffer containing 87% of βlactoglobulin with a few percent of -lactalbumin, less that 1% of salt (< 0.1% Ca2+), 2% of lactose, 2% of soluble non-protein nitrogen compounds, and around 3% of water. Their buffers were ranged from 2 to 95 g of β-lactoglobulin per liter of double-distilled water at pH comprised between 6.75 and 6.95. They heated the buffers up to 65°C for different times. De Wit (2009) has recently reported that all the “new” proposed mechanisms are contradictory with older studies, and particularly with the «well-known chemistry of the SH/SS interchange in proteins in this pH range» explained by Cecil (1963). The author concluded that the SH/SS interchanges are induced by the presence of thiolate ions (RS-), and not based on a polymerization induced by a free thiol group (-SH) as proposed in Roefs and de Kruif’s model. The following equations represent the basis of the reduction of protein disulfides by thiolate ions and exchange reactions between disulfides. The reactions stop when the thiolates are oxidized in disulfide bonds: RS- + R’SSR’ RSSR’ + R’S- (15) RS- + R’SSR RSSR + R’S (16) De Wit (2009) has also explained that a continuous polymerization is impossible according to the two previous reactions (15 and 16). According to him, heat-induced aggregation can occur only by the oxidation of the thiolate groups of the monomers, dimers, and tetramers. Bauer et al. (2000) have proposed a model slightly deviated from Roef and de Kruif’s. Their observations indicate that the mechanism is divided into four stages such as: 1) Equilibrium between monomers and dimers; 2) «Activation» of monomers; 3) SH/SS exchanges to form dimers, trimers, and tetramers; 4) Formation of large aggregates. Their 30 results also suggest the importance of an aggregation nucleus, formed from four oligomers, as a prerequisite for the heat-induced aggregation of β-lactoglobulin. Schokker et al. (1999) have also demonstrated that non-native dimers and oligomers were formed in the early stages of the denaturation of β-lactoglobulin, presumably through SH/SS exchanges, thiol-thiol oxidation and non-covalent interactions at 78.5°C. Later, large aggregates ranging from 105 to 2x 106 Da were formed continuously by incorporation of monomers and/or small aggregates, mainly through SH/SS exchanges. Unlike other studies, their model proposes non-linear aggregates. They found that the formation of aggregates involves more than one free thiol group. According to more recent studies and reviews, models proposed by Schokker et al. (1999) and Bauer et al. (2000) seem to be the most coherent. They defined the heat-induced aggregation of β-lactoglobulin as nucleus-dependent reactions that occur really fast. Their models are also closer to the general trend reported by Wang et al. (2010) on the aggregation pathways of proteins aggregation. However, they are mainly based on thiol/disulfide exchanges or thiol-thiol oxidation. The role of non-covalent interactions is still uncertain. Galani et al. (1999) concluded that the importance of non-covalent interactions is more important at high temperature of heating ranging from 90°C to 110°C. Using a free thiol-blocking agent (N-ethylmaleimide), Kitabatake et al. (2001) demonstrated that it was possible to form heat-induced aggregates of β-lactoglobulin. 2.4.2 Heat-induced interactions with minor whey proteins It is generally accepted that -lactoglobulin dominates the formation of whey protein aggregates when all whey proteins are present. The aggregation mechanism of this protein has been studied extensively in different buffers, at a wide range of temperatures, pH, or ionic strength. The aggregation mechanisms of other whey proteins, especially minor whey proteins, are less studied. However it appears that minor whey proteins can affect the aggregation process of the major whey proteins. 31 Using one and two-dimension SDS-PAGE, Havea et al. (1998) showed that protein aggregates were formed involving both hydrophobic interactions and SH/SS exchanges during the heating of whey protein concentrate at 75°C at pH 6.9. Aggregates were composed of -lactoglobulin, -lactalbumin, bovin serum albumin (BSA), and Imunnoglobulin H and L. Gezimati et al. (1996) reported that bovine serum albumin in a solution responds in a similar way to the heat treatment than -lactoglobulin. Havea et al. (2000) also defined the aggregation mechanism of the protein as thiol catalyzed polymerization close to lactoglobulin’s, but with the exceptions that BSA begins to aggregate at a lower temperature, and without the initial separation of the dimer into two monomers. During its unfolding, BSA loses its native form for a transformed monomer form with an exposed thiol group (Cys34) that can promote inter SH/SS exchanges. Gezimati et al. (1996) also reported that BSA can be found as hydrophobic aggregates. Later, Gezimati et al. (1997) confirmed the presence of hydrophobicaly formed BSA aggregates at 70°C. They suggested that this form could be an intermediate in the formation of a gel network, and that SH/SS exchange will stabilize the aggregates. Authors have also found that lactoglobulin and BSA can interact together during heating, however, direct evidence of SH/SS inter-exchanges were found only in the second study. Havea et al. (2000) proposed a possible mechanism for the aggregation of -lactalbumin in the presence of BSA where reactive polymers of BSA would interact with unfolded -lactalbumin through hydrophobic interactions. The complexes are then stabilized by SH/SS exchanges between the free thiol group of the BSA and presumably the disulfide bond Cys6-Cys120 of lactalbumin. Then, the complex could interact with another unfolded -lactalbumin to form SH/SS exchanges with the newly exposed free thiol group of the attached -lactalbumin. However, they concluded that the SH/SS exchange process was reversible, resulting in the possible liberation of oligomers of -lactalbumin. 32 2.4.3 Heat-induced interactions involving casein micelles Heat-induced interactions of casein micelles with whey proteins have been extensively studied since the first hypothesis of Sawyer et al. (1963). It is known that whey proteins, particularly -lactoglobulin, can associate with the casein micelles through SH/SS exchanges with the -caseins (Donato et al., 2007; Morr & Josephson, 1968; Sawyer, 1968; Vasbinder et al., 2003a). Anema et al. (2003a) have observed an increase in size of the casein micelle ranging from 30 to 35 nm during the early stages of heating of skim milk at pH 6.5, while only an increase of around 10 nm was observed at pH 6.7. Earlier, Smits and van Brouwershaven (1980) revealed that 83% of the total amount of -lactoglobulin was associated with the casein micelle during heating of skim milk at 90°C at pH 5.8, while 76%, 44%, and 24% of the total -lactoglobulin was associated at the same temperature but at pH 6.6, 6.8, and 7.3, respectively. Vasbinder et al. (2003a) have reported that 65% of the total amount of -lactoglobulin and 50% of the total amount of -lactalbumin were associated with the casein micelles during heating of skim milk at 80°C at pH 6.9. However, -lactalbumin could associate with the casein micelles only in the presence of lactoglobulin. Later, Livney et al. (2004) reported evidence (but low confidence) of formation of disulfide bonds between residues Cys66 of -lactoglobulin and Cys11 of casein and between residues Cys160 of -lactoglobulin and Cys88 of -casein. Despite the new evidences concerning the heat-induced interactions of whey proteins and casein micelles, the mechanism is still unclear. Donato et al. (2009) summarized the latest and more coherent assumptions corresponding in four possible mechanisms trying to answer two major questions. The first question is to determine if whey proteins aggregate prior to the association with the casein micelles. The second question is to determine if casein dissociates from the micelle before or after association with whey proteins (Figure 2.4). 33 Figure 2.4: Schematic representation of the currently proposed pathways of formation of the heatinduced whey proteins/-casein complexes in heated skim milk from Donato et al. (2009). Corredig and Dalgleish (1999), Dalgleish et al. (1997), Guyomarc’h et al. (2003a), and Vasbinder et al. (2003a) have reported that -lactoglobulin and -lactalbumin could interact together in the serum phase during heating of milk, before interacting with the casein micelles (I.A in Figure 2.4). They defined whey protein aggregates as intermediates 34 in the heat-induced association of whey proteins with the casein micelles. Elfagm and Wheelock (1977) have already observed that two steps were involved in the complex formation between whey protein and -casein: 1) Changes in the whey protein structures that do not involve -casein: 2) Interactions between whey proteins and -casein. Later, Elfagm and Wheelock (1978) have confirmed these observations and observed that the presence of casein micelles facilitated interactions between -lactoglobulin and lactalbumin. Using a stronger heat-treatment, Guyomarc’h et al. (2003a) have isolated soluble aggregates and micelle-bound aggregates. The first type was composed of lactoglobulin and -lactalbumin. However, their results did not indicate that whey protein aggregates are present at the end of heating. In fact, their results showed that -caseins were present in all of the type of aggregates, and few s1-casein. Guyomarc’h et al. (2003b) concluded that soluble aggregates initiated the destabilization of the casein micelles because of their interactions with the micelle surface. They suggested that hydrophobic interactions might be involved if the pH of the buffer led to decrease the repulsive charge interactions. Nevertheless, Eubert and Brunner (1982) have observed in their model system that aggregation of -lactoglobulin does not seem necessary for the association with the casein located at the surface of the casein micelles and that a partial denaturation of lactoglobulin could be sufficient for the formation of micelle-bound aggregates (I.B in Figure 2.4). Guyomarc’h et al. (2009) have also observed that whey proteins integrated casein to the formation of protein aggregates with similar structure comparatively to whey protein aggregates and without a previous step. They suggested that hydrophobic bonding and/or SH/SS exchanges were involved. On the other hand, most of the studies agreed that the dissociation of -casein from the casein micelles is a major point for the understanding of the association with whey proteins. Anema and Klostermeyer (1997) have observed that at pH ranged from 6.3 to 7.1, a small proportion of casein micelles was dissociated at 20°C, but the quantity of the soluble caseins increased with the increase of the temperature. The maximum dissociation was 35 observed at 70°C, before the irreversible denaturation of whey proteins. Later, results of Anema (2008) confirmed that the dissociation of -casein from the micelles can occur before the association with whey proteins. Moreover, their results strongly indicated that heat-denatured whey proteins preferentially associate with soluble -casein (present in the serum-phase) than -casein located at the surface of the casein micelle (II.C in Figure 2.4). Oppositely, Parker et al. (2005) have reported that the addition of caseinates in skim milk hardly affected the formation of heat-induced aggregates in the serum. Later, Donato et al. (2007) have also observed that addition of -casein had little or no effect on the formation of protein aggregates during heating of milk. These results suggest that denatured whey proteins preferentially associate with the -casein located at the surface of the casein micelle rather than soluble -casein or caseinates (II.D in Figure 2.4). Aggregates may dissociate from the casein micelle in the latest stages of heating. Proposed mechanisms are difficult to compare because most of the studies have been conducted at different pH, different temperatures, different heating systems, and with different kind of proteins (purified, isolated, in milk, etc). Donato et al. (2007), for example, suggested that -casein in solution is highly charged and has less hydrophobic sites than -casein located at the surface of the casein micelles. It is most likely that the presented mechanisms co-exist more than compete with each other. 2.4.4 Heat-induced aggregation mechanisms of MFGM constituents 2.4.4.1 MFGM proteins Little information is available on the denaturation (including aggregation) mechanisms of MFGM proteins. Ye et al. (2002) were able to extract the MFGM from whole milk and to remove whey proteins. Their results showed that protein aggregates were formed during heating through intermolecular disulfide bonds between XO, BTN, and a small proportion 36 of PAS 6. Their study revealed that XO and BTN were denatured at low temperatures (60°C, 10 min), while PAS 6/7 were stable until 80°C. Appel et al. (1982) already reported that BTN forms heat-induced aggregates through SH/SS exchanges at 58°C. Heat-induced aggregates between MFGM and other dairy proteins have been found in different dairy products. Although, the exact mechanism of association remains uncertain, and different pathways could occur. Corredig and Dalgleish (1998) found protein aggregates including whey proteins and MFGM proteins in buttermilk. These interactions were induced by the pasteurization of cream even at a temperature as low as 65°C. They also observed that association between MFGM proteins and -lactoglobulin was predominant at higher temperatures (> 65°C), but not for -lactalbumin. Authors suggested that both non-covalent interactions and SH/SS exchanges may be involved, but their results did not permit to confirm their statement. It has been hypothesized that the association between whey proteins or -casein through SH/SS exchanges is probably initiated by the MFGM proteins. For example, XO has 22 disulfide bonds, 38 free thiol groups with 4 available in the native form and 34 buried in the structure (Cheng et al., 1988). Furthermore, they demonstrated that all the free thiol groups become available after the unfolding of the protein. Due to the unfolding of the MFGM proteins at a lower temperature comparatively to whey proteins, free thiol groups of the MFGM proteins are exposed during the first step of the whey protein denaturation, even before the exposition of the free thiol group of -lactoglobulin, but after the dissociation of -lactoglobulin into monomers (Ye et al., 2004b). Ye et al. (2004b) confirmed that the major whey proteins interact with MFGM proteins via SH/SS exchanges during heating of whole milk starting at 60 to 65°C for 10 min. The amount of -lactoglobulin and -lactalbumin linked to the MFGM increases with an increase of the temperature up to 80°C, and then stays constant. They determined that only 1% of the total amount of -lactoglobulin and 0.8% of the total amount of -lactalbumin were associated with the MFGM material during heating of whole cream. It appeared that 37 XO and BTN were not affected by the heat treatment, while both PAS 6 and PAS 7 were affected. They also found that a small proportion of -casein associated with the MFGM proteins via SH/SS exchanges. In another study, Ye et al. (2004a) tried to better understand the aggregation mechanisms that occur between MFGM proteins, -lactoglobulin, and -lactalbumin. The authors summarized the overall process in accordance with the literature as follows: 1) Unfolding of MFGM, -lactoglobulin, -lactalbumin; 2) Association of polypeptides of MFGM and self-association of whey proteins; 3) Interactions between unfolded proteins or small aggregates of MFGM, -lactoglobulin, -lactalbumin. Their results permitted to propose two mechanisms for the heat-induced association between MFGM proteins and lactoglobulin. At temperatures above the -lactoglobulin (B2) irreversible unfolding (60°C to 70°C), they proposed a mechanism in two steps. Under the effect of the temperature, -lactoglobulin partially unfolds (B), but does not expose its free thiol group located in the residue Cys121. In opposition, MFGM proteins (MFGM*) become irreversibly unfolded (MFGM**) and expose more free thiol groups (*). At this temperature, MFGM proteins self-associate through SH/SS exchange ((MFGM)). The MFGM protein complexes still possess available thiol groups (*): B2 … B (17) MFGM* => MFGM** (17’) MFGM* + MFGM** => (MFGMi+1)* (18) Then, partially unfolded -lactoglobulin associate with the MFGM proteins or with MFGM protein aggregates. The formation of the complex ((B*MFGM)) involves SH/SS exchanges and permits the release of one thiol group of the -lactoglobulin priorly involved in a intramolecular disulfide bond: B + MFGM** or (MFGMi)* => (B*MFGM)* 38 (19) At higher temperatures (75°C to 95°C), the mechanism is slightly different because the lactoglobulin completely unfolds and exposes its free thiol groups (B*): B2 … B => B* (20) MFGM* => MFGM** MFGM* + MFGM** => (MFGMi+1)* (20’) i1 (21) The formation of the complex between -lactoglobulin and MFGM proteins ((BMFGM)) involves SH/SS exchanges initiated by the free thiol groups of the MFGM proteins and the free thiol group of -lactoglobulin. Authors did not suggest that this association release a thiol group of the -lactoglobulin. However, it is easy to think that the complex remains reactive due to the large amount of free thiol groups of the MFGM proteins: (MFGMi)* + B* => (BMFGM)* (22) It has not been explained yet why the proportion of whey proteins that associate with the proteins of the membrane is limited and relatively low (~1.0 and ~ 0.25 mg of protein.g-1 fat for -lactoglobulin and -lactalbumin, respectively) (Ye et al., 2004a). A possible explanation is that whey proteins preferably interact with each other or with -casein during heating and that the aggregates could not interact with the MFGM proteins. Corredig and Dalgleish (1996) reported that casein micelles offer a better surface area for heat-induced interactions with whey proteins based on their diameters (50 to 300 nm) comparatively to fat globules (0.5 to 10 m). Another possible explanation is that the membrane offers a limited number of sites for interactions with other proteins upon heating. Once all the available sites have been taken, any interaction can occur between MFGM and other proteins (Ye et al., 2004a). 2.4.4.2 Involvement of minor lipids in the heat-induced aggregation of MFGM Houlihan et al. (1992) concluded that heat-induced aggregation of whey proteins with the milk fat globule membrane in whole milk is mostly due to protein-protein interactions, however they suggested that membrane lipids could also be involved. Their idea is based 39 on the fact that membrane proteins are lost during the heat treatment while the hydrophobic regions of phospholipids become accessible. As a general trend, Dufourcq and Faucon (1977) explained that the interaction between proteins and phospholipids is possible due to a ionic interaction between the lysine or arginine residues of the protein and the negative groups of the phospholipids such as phosphate or carboxylic. In a second step, hydrophobic residues (tryptophan) interact with the bilayer. Brown et al. (1983) studied the complex formation between solvent-treated -lactoglobulin and phosphatidylcholine in an aqueous solution. They concluded that proteins in their native conformation (secretory) are unable, in general, to form a complex with lipids. A certain degree of unfolding is necessary to expose the initially buried helical regions and hydrophobic groups. The complex formation can start as soon as these regions and/or groups are exposed. According to their model, lipid-protein interactions are initiated by an ionic attraction between charged amino acid residues in the protein and the polar head group of the lipid in order to position the lipid. The complex is then stabilized by hydrophobic interactions between the hydrophobic side of the helix and the hydrocarbon chain of the lipid. A similar binding mechanism has been observed with milittin to phospholipids (Dufourcq et al., 1977), and glycophorin (Ong et al., 1981). Cornell and Patterson (1989) have reported interactions between -lactoglobulin and phospholipids at pH 4.4. Their results are consistent with the idea that positively charged groups of the protein interact with negatively charged lipids. At pH 4.6, -lactoglobulin has a positive net charge and PC is isoelectric (other phospholipids are ionized at pH > 4.0) (Hauser & Phillips, 1979). Cornell et al. (1989) supposed that an electrostatic attraction between lactoglobulin and PC could be expected around this pH. Between five to seven positive charges of the protein (depending on the conformation) could welcome the negative charges of PC. Earlier, Brown et al. (1983) have also reported bondings between lactoglobulin and PC after a treatment with a helix-forming solvent. They observed that an ionic interaction permitted the position of the lipid and protein molecule, and then 40 hydrophobic interactions stabilize the complex. Later, Lefèvre and Subirade (1999; 2000) have observed using Fourier transform infrared spectrophotoscopy hydrophobic interactions between -lactoglobulin and SM and electrostatic interactions between the protein and PS. Their study also suggested that -lactoglobulin does not bond with PC. However, Spector and Fletcher (1970) concluded that -lactoglobulin offers only one highenergy binding site for long chain free fatty acids (FFA), and one weak energy-bonding site. In opposition, serum albumin offers six high-energy binding sites. Barratt et al. (1974) studied interactions between casein (-, S1-, -) fractions and synthetic lecithins (C10, C12, C14), or synthetic lysolecithins (C10-C20). Their results showed that only S1- was able to complex with dicapryl (C10) and dilauroyl (C12) lecithin, and with lysolecithins. It appeared that complexes formed with short-chain lysolecithins (C10, C12) were not stable. Conversely, long-chain lysolecithins (C14, C16, and C18) formed stable lipid-protein complexes with S1-casein. Their results showed that additional molecules of lysolecithin could interact with the formed lipoprotein (lysolecithins-S1-casein). They also found that about 30 to 35 molecules of lysolecithin (C14-C18) were bonded to one S1casein. Previously, Patrick et al. (1972) found that the fatty acid chains of the lysolecithin molecules were bounded hydrophobically to the proteins whilst the polar groups of the lysolecithin molecules remain free. It is known that the protein hydrophobicity is important in lipid binding, however, Barrat et al. (1974) judged their results surprising because - and -casein are more hydrophobic than S1-casein. Their tendency to interact with phospholipids should be consequently greater. Recently, Gallier et al. (2012) used a model system to understand the structure of the MFGM and understand lipid-protein interactions at the surface of the globule. Their atomic force microscopy images clearly demonstrated interactions between -casein and phospholipids from the bovine MFGM. They concluded that both electrostatic interactions due to the negative net charge of the protein at pH = 7 and hydrophobic interactions due to its high hydrophobicity were involved. Their study did not permit to identify which 41 phospholipids were involved in the complex formation, however based on their previous work they supposed that PE may be involved. 2.5 Applications for dairy protein aggregates in food formulation Many dairy products like cheese or yogurt are classified as protein-rich food products. In these products, proteins serve as sources of essential amino acids and as a texture agent. Coombe and Kett (2005) and Zemser et al. (1998) have shown that the structure of protein is closely correlated to its functionality. Primary structure properties have been defined as hydrophobic/hydrophilic residues, the number of disulfide bonds and free sulfhydryl groups, and the net charge of the protein (Chiti et al., 2003; Chiti et al., 2002). This information is sufficient to control texture of food during well-known process, but is insufficient to innovate with new textures, new properties, and new sensory or nutritional properties. In fact, industries use only a few different types of protein because of the poor literature about the folding of proteins, and their denaturation behavior. Moreover, thermal processes change the physical and functional properties of proteins. Some changes are desirable because they improve the texture (yogurt), but in most cases are undesirable (specially in UHT treatments). In consequence, the use of dairy protein aggregates is limited to increase the cheese yield or to substitute fat in low-fat or fat-free products. 2.5.1 Role of aggregates as fat mimetic A portion of fat can be removed in some food systems without a noticeable change in the texture of the product or the taste, but total removal is rare. Fat replacers have been developed to replace the fat in food systems without any changes in the final physical properties of the matrix and hypothetically with any changes in the flavor or savor. The definition of fat replacers and fat mimetics have been given by the American Dietetic Association (2005), the Calorie Control (2004), Glueck et al. (1994), and Omayma and Youssef (2007) as follows: 42 «The ideal fat replacer(s) recreates some of all the functional properties of fat, while also significantly reducing fat and calorie content, and without any of the undesirable properties. A fat mimetic is an ingredient that mimics one or more of the sensory and physical functions of fat in the food by holding water. They provide from 0 to 9 kcal g-1». The protein-based mimetics are made through physical aggregation of proteins from milk, whey, corn, and egg white. The aggregates from whey and corn proteins are more suitable for food preparation including heating treatments, but not for extreme heating or frying. Protein-based substitutes can replace 75 to 100% of the fat in food systems (Glueck et al., 1994) and they are partially or fully digested and absorbed, with a lower calorie level due to their low energy density. Regarding food regulation, fat mimetics are food additives or generally recognized as safe (GRAS) substances. Their use is approved only if it has been demonstrated that they are safe through long-term ingestion (American-DieteticAssociation, 2005). The common and usual name of the fat replacer must be labeled on the package in conformation of the Nutrition Labeling and Education Act of 1994. Possible positive and negative effects of fat replacers are summarized in Table 2.6. Table 2.6: Possible positive and negative effects of fat replacers (adapted from Senanayake and Shahidi (2005). Benefits Reduction of total fat in the diet Reduction of energy intake in the diet Replacement of saturated fat in the diet Reduction of intake cholesterol Reduction of absorption of cholesterol Reduction of triacylglycerol levels Reduction of coronary heart disease risk factors Increase the intake of complex carbohydrates No flavor compromise No trans-fatty acids Maintain mouth-feel, texture, and moisture Inconvenients Reduction of intake of fat-soluble vitamins Reduction of essential fatty acids intake Compensatory energy intake Increase of the cost Alteration of the function of the digestive tract Possible gastrointestinal tract disturbance The process of preparation of denatured proteins has been introduced for the first time by Walker (1970) as the Alfa-Laval Centriwhey process. Initially, the whey was taken after the 43 cheese-making process, and cooled down to 8°C to adjust its pH to 6.25. Whey was heated at 95°C for twenty minutes, and then pH was decreased to 4.6 during the heating. The denatured proteins were separated at 40°C by sedimentation or centrifugation, and the denatured whey protein concentrate was finally cooled down to 13°C. This process has been widely used in France to incorporate whey proteins in Camembert, Coulommier, Demi-Carré, Port Salut, and Bel Pease cheeses since the seventies (Lawrence, 1993). With years, the process has been modified or improved. First, Lelièvre (1990) concluded that the size of the heat-denatured whey protein is critical, and related to their incorporation in a food system. He suggested adding an homogenization treatment as the last step of the process in order to reduce the average particle size distribution between 0.1 and 10 m. This range has been judged optimal to trap the denatured whey proteins in the casein matrix. The pH can also be decreased to 4.6 at the beginning of the denaturation instead of just at the end (Lebeuf et al., 1998; Punidadas et al., 1999). Following these two major changes, the name of this ingredient changed for microparticulated and denatured whey proteins (MWP). CPKelco US Inc., and Parmalat-Ingredients have not revealed their industrial processes for the production of Simplesse and Dairy-Lo products, but were probably inspired by the Alfa-Laval Centri-whey process. Simplesse products are made through a simultaneous heating and shearing of whey proteins or a mix of milk and egg proteins. Dairy-Lo is manufactured by ultrafiltering sweet whey followed by a partial thermal denaturation of proteins by heating. Several studies have been developed to investigate the impact of the use of protein-based fat mimetics into cheese and yogurt processes. Simplesse and Dairy-Lo products are low calorie protein-based fat mimetics. Glueck et al. (1994) calculated for example that using one gram of Simpless in a food formulation really decreases the calorie content of the food product. In fact, three grams of fat (9 kcal.g-1) can be replaced by one gram of Simpless® (4 kcal.g-1), so a decrease of 5 kcal per gram of fat replaced can be expected. However, the use of this type of ingredient in dairy products is often limited due to the high water- 44 binding capacity of denatured proteins. Depending on the products, it is generally advisable to add them in proportions of 0.5% to 1.5% of the total weight. Seydim et al. (2005) have shown that Dairy Lo has very promising texture and sensory properties for the production of acid set-type yogurt. Addition of Dairy Lo at 1.7% of the final weight (w/w) did not affect the whey separation, but increased the gel strength. The firmness of yogurts supplemented with Dairy-Lo was comparable to full-fat yogurts after one and seven days of storage but significantly higher comparatively to low-fat yogurts. In opposition, Sandoval-Castilla et al. (2004) have reported that yogurts supplemented with Dairy Lo (1.05% of the final weight (w/w)) were lower in tension and firmness but higher in cohesiveness and adhesiveness comparatively to their control full fat yogurts. They have also observed that the use of protein-based fat mimetics yielded yogurt with a more open and less dense protein network. Tamine et al. (1995) have reported that a low proportion of Simplesse® 100 (1.5% of the final weight (w/w)) interacted with the milk protein during preparation of the base, but most of the interactions took place during gelation. Yogurts supplemented with Simplesse® 100 were the lowest in firmness and the most susceptible to serum separation. 2.5.2 Role of aggregates as water holder Lawrence (1993) summarized an alternative method to use proteins aggregates for the production of cheese by heating the milk at higher temperatures comparably to pasteurization in order to denature the whey proteins and form aggregates with the casein micelles. During the coagulation, the denatured whey proteins remain attached to the micelle and become a part of the gel matrix. This method has been extensively used for the yogurt-making process because it leads to the formation of gels with higher viscosity and firmness (Dannenberg & Kessler, 1988; Lucey et al., 1998b; Lucey et al., 1997; ParnellClunies et al., 1986), with more homogeneous microstructure with lower porosity of the network (Kalab et al., 1976; Parnell-Clunies et al., 1987), and with higher whey retention (Dannenberg et al., 1988; Puvanenthiran et al., 2002). The method has also been extensively investigated for the cheese-making process because it increases the production yield. However, it is often judged controversial for the cheese production because the 45 formation of aggregates limits the primary phase of the enzymatic reaction and prevents fusion of the casein micelles. Furthermore, cheeses have excessive moisture and texture defects (Guyomarc'h, 2006). It is believed that some improvements such as alkaline heattreatment can be added in order to improve this method. Later, Donato and Guyomarc’h (2009) concluded that the use of whey protein/-casein or whey proteins/casein micelle is limited due to little information available on the formation of theses complexes, their properties, the conditions to control their properties, and their nutritional properties. Buttermilk seems to be a promising candidate for the formation of whey protein//-casein aggregates. The use of buttermilk is however very challenging because this by-product is complex and is still not fully understood. Moreover, its constituents have been subjected to different rough processing steps. There is a lot of work to be done to understand the structure and properties of the proteins from buttermilk. There is even more work to be done to fully understand the microstructure and composition of dairy protein-based aggregates. Finally, it is important to determinate if the protein-based aggregates become a part of the end products or not. Tamine et al. (1995) and more recently Sandoval-Castilla et al. (2004) have concluded that the added denatured proteins were involved in the structure of the yogurt by means interactions with the casein micelles. This information is capital because it would increase the use of protein aggregates. 46 CHAPTER 3 HYPOTHESIS, GOAL, AND OBJECTIVES It is generally accepted that skim milk and buttermilk are similar in composition. Considering the literature presented, it seems obvious that caseins, whey proteins and MFGM constituents from buttermilk interact with whey proteins upon heating. However, buttermilk has been subject to intensive treatments such as skimming, pasteurization, churning. Previous studies from this research group have clearly demonstrated that buttermaking process affects the constituents of buttermilk. These modifications probably affect the sensitivity of buttermilk to heat treatment and particularly the formation of heat-induced aggregates. Examples of modification could be different pathways for aggregation or implication of new constituents. The general hypothesis of this doctoral thesis is that a better understanding of buttermilk constituents and particularly their reaction to heat treatment in presence of whey proteins can control the properties of the resulting aggregates. While using whey proteins aggregates during formulation may not be able to fulfill the texture requirements of the final product, addition of buttermilk during the heat-induced aggregation can modify the physical and functional properties of aggregates. Objectives for this work were: 1. To develop heat-induced protein aggregates by combination of whey and buttermilk concentrates. 2. To understand the effect of buttermilk constituents on the heat-induced formation of whey protein aggregates. 47 3. To characterize the contribution of phospholipids from buttermilk to the formation of heat-induced whey protein-buttermilk aggregates. 4. To understand the effect of the type of interactions that occurred during the association of proteins on the properties of the aggregates. 5. To evaluate the potential of whey:buttermilk heat-induced aggregates as replacement of skim milk powder for the production of acid-set type yogurts. 48 CHAPTER 4 THERMAL AGGREGATION OF WHEY PROTEINS IN THE PRESENCE OF BUTTERMILK CONCENTRATE A preliminary experience have been designed in order to determine the optimal processing parameters (pH, temperature, and holding time) in order to produce aggregates with lower water-holding capacity comparatively to whey protein aggregates. Statistical analysis of the data revealed that the temperature of 90°C led to a more efficient production of sedimentable protein aggregates (higher aggregation yield) as a function of the buttermilk proportion in the mixture than the temperature of 80°C and that a lower pH (4.6) led to the formation of less hydrated protein aggregates as a function of the buttermilk proportion in the mixture comparatively to higher pH (5.4 or 6.2). Statistical analysis revealed that the holding time had no significant difference on the water-holding capacity of aggregates or on the aggregation yield. The following parameters were selected: heating at pH 4.6 at 90°C and holding time of 5 minutes, The work presented in this chapter aimed at characterizing the impact of ratio whey to buttermilk on the denaturation process of cheese whey on the functional properties of denaturated protein aggregates. We also investigated in this project the impact of applying power ultrasound on the functional properties of aggregates. Results of this part of the project were published essentially in this form in Journal of Food Engineering (2011), 103:244-250. Results of this part of the project were also presented in two posters at 2010 ADSA Joint Annual Meeting in Denver (CO, USA), one poster at the Journée Scientifique GP3A in Quebec City (QC, Canada), and one poster at 2010 Journée Technologique Novalait in Drummondville (QC, Canada). 49 4.1 Résumé L’objectif de cette étude était d’évaluer l’impact de l’addition d’un concentré de protéines du babeurre sur le procédé d’agrégation des protéines du lactosérum. Des concentrés protéiques de lactosérum et de babeurre ont été mélangés puis chauffés (pH 4,6; 90°C). L’ajout de babeurre a significativement augmenté le rendement d’agrégation et diminué la capacité de rétention d’eau des agrégats. La modification des propriétés rhéologiques suggère que les protéines du lactosérum dénaturées ont interagi avec les protéines du babeurre. L’ajout d’un agent bloqueur de groupements thiols a permis de montrer que les interactions covalentes et non covalentes étaient impliquées dans la formation d’agrégats. La formation d’agrégats fut également influencée par un traitement aux ultrasons pendant le chauffage. Il apparait que sous l’influence de la cavitation et de la turbulence générée par les ultrasons, le rendement d’agrégation est augmenté et que la capacité de rétention d’eau des agrégats est diminuée. 50 4.2 Abstract The objective of this study was to assess the impact of adding buttermilk concentrate to the denaturation and microparticulation process of cheese whey protein concentrate. For this purpose, the two concentrates were mixed and co-denatured (pH 4.6; 90 °C) and homogenized. The presence of buttermilk significantly increased aggregation yield and decreased water-holding capacity of aggregates up to volumetric buttermilk - whey protein ratio of 75:25. Modifications of rheological properties suggest that denatured whey proteins interacted with caseins. A thiol blocker, N-ethylmaleimide, was added before heating to measure the role of disulphide bond formation in the aggregation process. Results showed that both covalent and non-covalent interactions were involved in the formation of aggregates. Ultrasound treatment was applied during denaturation process and was shown to influence aggregate formation. It appeared that under turbulence and cavitation conditions, aggregation yield was increased and water- holding capacity decreased. 4.3 Introduction A number of technological approaches have been developed to increase the retention of whey proteins into cheese (Lawrence, 1989; Lawrence, 1993; Lelièvre, 1990). A widely used approach consists in the addition of denatured and aggregated whey proteins to cheese milk. However, one of the most serious challenges is to optimize the size and water-holding capacity of the aggregated whey proteins. Indeed, larger aggregates readily interfere with the casein network and they are poorly retained in cheese curd (Lelièvre, 1990). Such a problem can, however, be circumvented by homogenization of the aggregates, thus facilitating retention (Mignot & Tracard, 1976; Punidadas et al., 1999). The addition of whey proteins into cheese has proved to increase yield but to reduce quality. Another problem associated with the incorporation of aggregated proteins is related to their effect on cheese moisture. Excessive cheese moisture is a common defect that limits the amount of whey proteins that can be added to cheese milk. 51 Substituting buttermilk to cheese milk at a concentration over 5–10% to milk has been shown to generate texture defects and high moisture (Joshi et al., 1994). However, other studies have shown that buttermilk could be a valuable product for increasing the moisture content and improve the texture of low fat cheese, essentially because it contains fragments of milk fat globule membrane (MFGM) including phospholipids, and whey proteins (Mistry et al., 1996; Raval et al., 1999; Turcot et al., 2002; Turcot et al., 2001). It was shown by Morin et al. (2008) that the poor coagulation properties of buttermilk were related to cream pasteurization that probably induced important modifications of the MFGM surface during the heating process. Many authors have reported that whey proteins had the potential to interact with -casein (-CN) (Donato et al., 2007; Morr et al., 1968; Sawyer, 1969; Vasbinder & de Kruif, 2003b) and particularly with serum-phase -casein after the dissociation of -CN from the micelle (Anema, 2008). In addition, Ye et al. (2004a,b) demonstrated that -lactoglobulin (-lg) and -lactalbumin (-la) associated with the MFGM by thiol–disulphide bonds in heated milk. In line with these observations, we hypothesized that co-denaturation of cheese whey and buttermilk proteins would induce the formation of aggregates between whey proteins–-CN – MFGM by means of SH/SS interchanges as well as by other non-covalent interactions such as ionic, hydrophobic or van der Waals. Also, the aggregation yield and water-holding capacity of the aggregates would be dependent to the ratio between whey and buttermilk, as well as on the availability of free-SH groups and on the possibilities of rearrangements during aggregation process. N-ethylmaleimide (NEM), a thiol blocker, can be used to prevent or slow down the thiol/disulphide interchange reactions between denatured milk proteins (Alting et al., 2000; Hoffmann et al., 1997). Havea et al. (2004) have shown that gels formed by non-covalent interactions were more rigid than those formed by disulphide bonds. We also hypothesized that applying physical treatment that would promote rearrangement of the aggregates during heating and would increase the compactness and reduce the hydration of the 52 aggregates. Ultrasound treatments are known to generate cavitation, turbulence and heat (Patist & Bates, 2008). High intensity ultrasound (20 kHz) generates shear forces that were effective in reducing viscosity of dairy ingredients (Zisu et al., 2009) and in homogenizing fat globule (Ashokkumar et al., 2009; Jambrak et al., 2008). The use of ultrasound treatments in food processing is considered as an emergent potential alternative to heat or to homogenization treatment (Patist et al., 2008). Moreover, under turbulent conditions due to ultrasound treatment, particle mobility is increased, which promotes the formation of aggregates (Walstra, 1983). The challenge remains, however, to control yield. The objective of the present work was to characterize the heat-induced aggregation process of whey protein–buttermilk concentrate mixtures with different buttermilk protein content (0%, 25%, 50%, 75% and 100%) on the aggregation yield and water-holding capacity of the aggregates. The effects of blocking free- SH by addition of NEM and applying power ultrasound (20 kHz) during aggregation process were also assessed. 4.4 Materials and methods 4.4.1 Materials Fresh Mozzarella cheese whey and buttermilk were obtained in a local cheese factory (L’Ancêtre, Bécancour, QC, Canada). N-ethylmaleimide (NEM) was obtained from Alfa Aestar (Ward Hill, MA, USA). All other reagents were from Fisher Scientific (Ottawa, ON, Canada). 4.4.2 Preparation of whey and buttermilk concentrates Fresh Mozzarella cheese whey and buttermilk were skimmed by using a pilot scale milk separator (Alfa Laval, Uppsala, Sweden). Bacterial contamination of cheese whey was 53 reduced by microfiltration (TetraPak MSF1, Lund, Sweden) through a 1.4 m membrane (Membralox®, Bazet, France). Microfiltered whey and buttermilk were concentrated by ultrafiltration (UF) through a 5 kDa membrane (Romicon, Koch Membrane Systems, Wilmington, MA, USA) to a final protein concentration of 9.5% (w/v) in the UF-retentate. The concentrates were frozen to -28 °C until further analysis. 4.4.3 Heating experiments Concentrates were mixed at the following buttermilk:whey protein ratios: 0:100; 25:75; 50:50; 75:25 and 100:0. Mixtures (200 mL) were adjusted to pH 4.6 by the slow addition of HCl 1 N and heated from 4 °C to 90 °C in a thermostatically-controlled water bath, under constant stirring for 25 min (including come-up time) at 168 rpm agitation rate. After thermal treatment, mixtures were cooled to 30 °C in an ice bath and homogenized for five passes at 65.5 MPa using an Emulsiflex-C5 (Avestin Canada, ON, Canada). The overall process used for the preparation of aggregates is summarized in Figure 4.1. 54 Figure 4.1: Experimental procedure used to prepare aggregated buttermilk and whey proteins mixtures. 55 4.4.4 Treatments with N-ethylmaleimide (NEM) and power ultrasound A concentration of 9.5 mM N-ethylmaleimide (NEM) was used in accordance with the experiments of Alting et al. (2000) in order to completely block the free thiol groups. NEM (9.5 mM) was added in whey–buttermilk concentrate (9.5% protein concentration) mixtures (200 mL) before heat treatment. Power ultrasounds were applied using a constant frequency 20 kHz probe (Virsonic-550, The Virtis Company, Gardiner, NY, USA). The power level was fixed to 275W. The titanium tip was immersed 2 cm underneath the surface during the treatment. The treatments were applied to the mixtures after 10 min of thermal denaturation, until the end of heating (25 min). 4.4.5 Analytical methods 4.4.5.1 Composition analyses Total nitrogen content was determined by the Dumas combustion method (IDF, 2002) using a LECO equipment (Protein Analyzer model FP-528 Leco Instruments Ltd., Mississauga, ON, Canada). Nitrogen values were converted into protein values using 6.38 as the nitrogen conversion factor. Fat was determined using the Mojonnier extraction method (IDF, 2008) and lactose by enzymatic method (IDF, 2001). Total solids were obtained by microwave drying (Smart System 5, CEM Corp., Matthews, NC, USA) and ash was measured by incineration in a muffle furnace at 550 °C for 20 h (AOAC, 1990). The overall composition of each mixture as determined experimentally is summarized in Table 4.1. 56 Table 4.1: Composition (% DM) of the various whey-buttermilk mixtures concentrated at 9.5% protein (w/v) used for the heating experiments % Dry Matter Proteins Lactose Fat Ashes a 100:0 73.0 ± 0.2 10.0 ± 0.2 7.7 ± 0.1 1.7 ± 0.0 75:25 64.0 ± 1.8 14.6 ± 0.5 10.6 ± 0.1 3.0 ± 0.0 50:50 59.8 ± 3.0 17.6 ± 0.6 11.7 ± 0.2 4.3 ± 0.2 25:75 54.1 ± 1.8 20.8 ± 0.4 12.6 ± 0.4 5.2 ± 0.0 0:100 52.1 ± 0.3 26.9 ± 0.3 14.3 ± 0.3 6.0 ± 0.0 a refer to the relative proportion of total proteins of cheese whey and buttermilk in the mixtures. 4.4.5.2 Determination of aggregation yield and water-holding capacity The aggregated material was separated by centrifugation (SORVALL-RC-5B, Du Pont Company, Newtown, CT, USA) at 15,000g for 20 min at 20 °C. Aggregation yield (Equation 4.1) was calculated from protein content in the supernatant: Equation 4.1: Calculation of the aggregation yield Aggregation Yield (%) = [ P ]M - [ P ]S ´100 [ P]M where [P]M was the protein concentration in mixture (%) and [P]S was the protein concentration in supernatant (%). Water-holding capacity (Equation 4.2) was estimated from the moisture of the centrifugation pellet and reported on protein content. Equation 4.2: Calculation of the water-holding capacity of protein aggregates WHC (gwater /g protein ) = [ M] P [ P]P where [M]P was the moisture content in pellet (g) and [P]P was the protein content in pellet (g). 57 4.4.6 Particle size distribution and surface area The particle size distribution of homogenized mixtures was measured using a Mastersizer 2000 laser diffraction system (Malvern Instruments, Worcestershire, UK). Samples were shaken by inversion for 1 min and dispersed directly into the recirculating cell until a 12– 19% obscuration value was obtained. Agitation rate of dispersant (deionised water) was set to 320 rpm and measurements were started after 5 min of agitation. Data was obtained from the average of five measurements by samples and expressed as the volume-weighted average particle size, also known as the D[4,3] value. 4.4.7 Determination of rheological properties Flow curves were obtained using a controlled strain rheometer (ARES 100 FR, TA Instrument, New Castle, Delaware, USA). A coaxial geometry with a cup (35 mm diameter) and a bob (35 mm length, 33 mm diameter) was used. Samples were poured directly (13 mL) into the measuring system of the rheometer at 25 °C. The shear rate and the shear stress were interpreted using the rheometric computer program. The values of the flow behavior index (n) and the consistency coefficient (k) were obtained from plots of log shear stress versus log shear rate between 5 and 500 s-1, according to the power law model (Equation 4.3). Equation 4.3: Formula of the Power Law Model log t = logk + n logg where is the shear stress (Pa), is the shear rate (s-1), n is the flow behavior index, and is the consistency index (Pa.sn) 58 4.4.8 Statistical analysis All experiments were performed in triplicate, so each value represents the mean of three measurements. The effect of ratio, addition of a thiol blocker and application of ultrasound on aggregation yield, WHC, rheological properties and particle size distribution were tested according to a factorial design. Statistical analyses were carried out using SAS software (SAS Institute Inc., Cary, NC). Table 4.2 summarizes significance for each treatment on each variable. Results were considered significantly different when P < 0.05. Table 4.2: Summary of the significance (P values) calculated by the analysis of variance of data of each contrasts for each variables (n = 3). Significance (P values) Contrast Aggregation yield WHC k n Model < 0.0001 0.0311 < 0.0001 < 0.0001 Ratio (R) < 0.0001 0.0009 < 0.0001 < 0.0001 NEM (N) 0.0045 0.0470 0.0011 < 0.0001 Ultrasound (U) < 0.0001 0.0002 0.0433 0.0017 R*N 0.0557 0.9627 0.3598 0.0019 R*U 0.2054 0.4440 0.0001 0.0009 N*U 0.6547 0.4843 0.3214 0.0007 R*N*U 0.0989 0.5816 0.5104 0.9726 a Logarithmical transformations were applied Particle sizea 0.0009 0.0040 0.0010 0.0874 0.0036 0.3023 0.9403 0.3069 4.5 Results 4.5.1 Aggregation yield The effects of buttermilk to whey protein ratio, ultrasound and NEM treatments on aggregation yield are shown in Figure 4.2. 59 Aggregation Yield (%) 100 100:0 75:25 50:50 25:75 0:100 80 60 40 Control (R) R*U R*N Figure 4.2: Effect of protein composition, NEM and ultrasound on aggregation yield of different mixtures of whey and buttermilk concentrates (9.5 % protein), acidified at pH 4.6 and co-denatured at 90°C. Aggregation yields for pure cheese whey and buttermilk concentrates were respectively 63.3 ± 1.9% and 86.1 ± 5.5%. In concentrate mixtures, the aggregation yields significantly increased (P < 0.0001) according to the proportion of buttermilk. There seems to be no deviation to the activity rule, suggesting that the components in the mixture might aggregate independently of each other. Ultrasound treatment significantly increased (P = 0.0001) aggregation yields and this effect was dependent on the proportion of buttermilk in the mixtures. Aggregation yield increased by 33% for pure whey protein concentrates, but the increase was only 3% for pure buttermilk concentrates. The aggregation yields in mixed concentrates treated with ultrasound were higher than for control concentrates, with a minimum aggregation yields observed at 75% cheese whey protein ratio (83.7 ± 1.8%) and a maximum at pure buttermilk concentrates (89.5 ± 4.0). Adding thiol blocker (NEM) significantly increased (P = 0.0045) the aggregation yields of mixed protein concentrates. On average, aggregation yields increased by 6%. Figure 4.2 60 suggests that NEM was more efficient to increase the aggregation yields in mixtures with high proportion of cheese whey protein concentrates. However statistical analysis indicates non-significant interaction with protein ratio (P = 0.0557). For protein mixtures treated with ultrasound, NEM had a significant effect (P = 0.0007) on aggregation yields (results not shown). 4.5.2 Water-holding capacity The effects of treatments on WHC are shown in Figure 4.3. Only the principal effects were presented since the second and third level statistical interactions between the factors were WHC (gwater/gprotein) not significant (Table 4.2). 3.0 3.0 3.0 2.5 2.5 2.5 2.0 2.0 2.0 1.5 1.5 1.5 1.0 0 25 50 75 100 Buttermilk protein fraction 1.0 1.0 no yes Ultrasound 0 9.5 [NEM] mM Figure 4.3: Effect of protein composition, application of ultrasound and addition of NEM on waterholding capacity (WHC) of mixed buttermilk-whey protein mixture (9.5 % protein) heated at pH 4.6 (90°C). WHC of aggregates formed in pure cheese whey and buttermilk concentrates were respectively 2.75 ± 0.08 gwater/gprotein and 2.21 ± 0.14 gwater/gprotein. In concentrate mixtures, WHC of aggregates significantly decreased (P = 0.0009) according to the proportion of buttermilk to reach minimum values at 75% buttermilk protein fraction (2.06 ± 0.16 gwater/gprotein). 61 Ultrasound treatment significantly increased (P = 0.0433) the WHC of aggregates. On average, WHC increased by 0.21 gwater/gprotein comparatively to control mixtures. Adding NEM significantly increased (P = 0.0470) the WHC of aggregates. On average, WHC increased by 0.18 gwater/gprotein comparatively to control mixtures. 4.5.3 Consistency coefficient (k) The effects of buttermilk to whey protein ratio, ultrasound and NEM on consistency Consistency Coefficients (MPa.s) coefficient are shown in Figure 4.4. 1500 1500 1500 1000 1000 1000 500 500 500 0 0 0 0 25 50 75 100 Buttermilk protein fraction no yes Ultrasound 0 9.5 [NEM] mM Figure 4.4: Effect of protein composition, application of ultrasound and addition of NEM on consistency index (k) of mixed buttermilk-whey protein mixture (9.5 % protein) heated at pH 4.6 (90°C). Consistency coefficients of pure cheese whey and buttermilk concentrates were respectively 253.2 ± 86.8 MPa.s and 705.6 ±141.6 MPa.s. In concentrate mixtures, k significantly increased (P < 0.0001) according to the proportion of buttermilk. 62 Ultrasound treatment significantly increased (P = 0.0017) k comparatively to control mixtures. On average, k increased by 100 MPa.s. The maximum values were obtained with 75% buttermilk protein ratio (988.4 ± 174.2 MPa.s). Adding NEM significantly increased (P = 0.0011) k comparatively to control mixtures. On average, k increased by 85 MPa s. The maximum values were obtained with pure buttermilk protein ratio (983.8 ± 116.2 MPa s). 4.5.4 Flow behavior index (n) The effects of buttermilk to whey protein ratio, ultrasound and NEM on flow behavior index are illustrated in Figure 4.5. A decrease of viscosity with increasing shear rate is indicative of a pseudoplastic behavior (n < 1). Flow behavior index (n) of control and treated mixtures indicated a pseudoplastic behavior. 1.0 100:0 75:25 50:50 25:75 0:100 Flow Behavior Index 0.8 0.6 0.4 0.2 0.0 Control Ultrasound NEM Figure 4.5: Effect of protein composition, NEM and ultrasound on flow behavior indices of different mixtures of cheese whey and buttermilk concentrates (9.5 % protein), acidified at pH 4.6 and codenatured at 90°C. Flow behavior indexes of pure cheese whey and buttermilk concentrates were respectively 0.751 ± 0.121 and 0.423 ± 0.072. In concentrate mixtures, n significantly decreased (P < 0.0001) according to the proportion of buttermilk in the mixture. 63 Ultrasound treatment significantly decreased (P = 0.0009) n and this effect was dependent to the proportion of buttermilk in the mixtures. Flow behavior indexes decreased by 0.124 with pure cheese whey concentrates, and, surprisingly, only by 0.008 with pure buttermilk concentrates. Flow behavior indexes of mixtures treated with ultrasound were lower than for control concentrates, with a minimum n observed at 75% buttermilk protein ratio (0.300 ± 0.0047) and a maximum at pure cheese whey concentrates (0.627 ± 0.0019). Adding NEM significantly decreased (P < 0.0001) n and this effect was dependent to the proportion of buttermilk in the mixtures. Flow behavior indexes decreased by 0.135 and 0.130 respectively for pure cheese whey and buttermilk concentrates. Flow behavior indexes of mixtures treated with NEM were lower than for control concentrates, with a minimum n observed at pure buttermilk concentrates (0.0293 ± 0.031) and maximum at pure cheese whey concentrates (0.616 ± 0.073). 4.5.5 Particle size distribution The effect of control and treatments on particle size distribution of homogenized mixtures is reported in Table 4.3. The particle size distributions of mixtures was significantly affected (P = 0.0040) by the proportion of buttermilk in the mixtures. It appeared that the size increased when whey proteins were in higher or equal proportion compared to buttermilk proteins (4.4 ± 0.1 µm with pure whey to 20.9 ± 10.2 µm with 50:50 ratios) and decreased with higher concentration of buttermilk (7.3 ± 3.9 µm with 25:75 to 5.7 ± 0.80 µm with pure buttermilk). 64 Table 4.3: Average particle size and span of untreated and treated whey-buttermilk mixtures acidified at pH 4.6, denatured at 90°C and homogenized at 65.50 MPa. Mixtures D[4,3] (µm) Span 100:0a 4.4 ± 0.1 1.6 ± 0.8 75:25 12.7 ± 1.8 3.4 ± 1.1 50:50 20.9 ± 10.2 2.2 ± 1.2 25:75 7.3 ± 3.9 2.0 ± 1.0 0:100 5.7 ± 0.8 1.2 ± 1.1 100:0 6.4 ± 4.4 6.1 ± 0.8 75:25 58.9 ± 32.1 17.4 ± 1.3 50:50 4.8 ± 2.2 2.0 ± 0.8 25:75 6.6 ± 4.4 1.5 ± 1.5 0:100 10.1 ± 8.8 1.6 ± 0.3 Absence of NEM Presence of NEM a refer to the relative proportion of total proteins form cheese whey and buttermilk in the mixtures. The aggregate size in co-denatured cheese whey–buttermilk protein mixtures was affected (P = 0.036) by both protein composition and NEM addition. Particle size decreased when mixture contained high proportion of whey proteins (6.4 ± 4.4 µm with pure whey to 4.8 ± 2.2 µm with 50:50 ratios) and decreased with higher concentration of buttermilk solids (6.6 ± 4.4 µm to 10.1 ± 8.8 µm with pure buttermilk). As a general trend, aggregates formed in presence of NEM were larger in mixture containing higher proportion of cheese whey proteins. However, a different behavior was observed for mixture with 50% buttermilk protein, where NEM had no effect on particle size. Ultrasound treatment had no significant effect (P = 0.0874) on particle size distributions of aggregates. 65 4.6 Discussion Our data show that it is possible to integrate buttermilk concentrate in a denaturation/microparticulation process of whey proteins in order to increase aggregation yields and decrease WHC of aggregates. Increasing the buttermilk protein fraction up to 75% led to higher aggregation yields and lower WHC combination. These results suggest that the presence of casein and eventually the dissociation of -casein from the micelle at low pH (4.6) before heating is more favorable to increase aggregation yield. However, it is not clear if aggregation results only from association of denatured whey proteins with each other or with -casein or both. Blond and Montupet (1989) had observed that protein WHC decreases in function of denaturated protein percentage. In line with this observation and knowing that caseins interact with denaturated whey proteins during heat treatment, the present results show that the co-denaturation of a low proportion of whey protein with a higher proportion of caseins results from the formation of condensed and low hydrated aggregates. Modifications of rheological properties and particle size distribution of aggregates caused by substitution of whey by buttermilk suggest that casein micelles participate in the aggregation process. Vasbinder et al. (2003b) showed that at low pH (<6.6) most of the whey protein aggregates are located at the surface of the casein micelles. Whey protein aggregates are larger in size than native protein and the association with -casein increases the size of the casein aggregates and led to increase viscosity of mixtures (higher k and lower n). However, an increase of viscosity could be due to some physicochemical changes in the mixture, such as the volume fraction occupied by the aggregates, a change in their size, shape or hydration, or also to an increase of attachment between particles. Considering these different possibilities, it can be thought that increasing proportion of casein in the mixture increased the number of binding sites and favored interactions between aggregated particles. 66 Hoffmann et al. (1997) have shown that the reactivity and the accessibility of -lg free thiols were higher when the pH was close to that of their pK value (8.2). In our pH condition (4.6), it is not clear if aggregation results only by thiol/disulphide bonds or noncovalent exchanges such as ionic, hydrophobic, van der Waals or both. Kitabatake et al. (2001) concluded that -lg linked with NEM interact with other -lg– NEM complexes mainly by hydrophobic interactions. Our results show that the origin of interactions between denatured proteins (disulphide bonds or non-covalent or both) does not affect aggregation yields. In accordance with Xiong et al. (1993), we hypothesized that reduced SH/SS interchange reactions would increase the molecular flexibility and that this modification could favor dehydration of aggregates by homogenization. However, it appeared that WHC of aggregates increased in presence of NEM. Havea et al. (2004) have reported that heat-induced whey protein concentrate gels formed by covalent interactions were rubbery and gels formed by non-covalent interactions were rigid. It is possible that the rigidity of these aggregates prevent their dehydration. Our observations on the changes in rheological properties and particle size distribution suggest that aggregates formed in the presence of NEM were larger (higher k, lower n) than those formed in untreated mixtures. Walstra (1983) suggested that turbulent conditions (homogenization) and heat generated by ultrasound treatment cause an acceleration of heat-induced aggregation/cross-linking and prevent the formation of disulphide bonds between denatured whey proteins. Our results show that the use of ultrasound increases protein aggregation in mixtures with high proportion of whey proteins. These results are in agreement with Villamiel and Jong (2000) who have shown that the denaturation of whey proteins (-lg and -la) was higher when the ultrasonic treatment was performed during heating. Moreover, the denatured whey proteins preferentially interact with dissociated -casein (Anema, 2008), and Taylor and Richardson (1980) had shown that sonication can generate free caseins by the disruption of casein micelle in milk treated with ultrasound treatment. Ertugay et al. (2004) demonstrated that cavitation and turbulence effects of ultrasounds are similar to conventional homogenization at a low power level (<180 W) and better at higher power levels (>180 W), 67 and that ultrasonic homogenization could reduce the size of milk fat globule (Ertugay et al., 2004; Villamiel et al., 2000). Zisu et al. (2009) and Daubert et al. (2006) have found a correlation between the decrease of whey protein size treated with power ultrasound and the decrease in viscosity. Jambrak et al. (2009) and Kresic et al. (2008) observed an increase in the apparent viscosity caused by a particle size decrease. Our results show that ultrasound treatment did not affect directly the particle size but promoted the formation of aggregates (highest yields), and especially for pure cheese whey proteins and 75% cheese whey protein ratio. In fact, the homogenization and shear stress caused by cavitation during ultrasound treatment probably modified the aggregation/cross-linking by aggregates remodeling or rearrangement during the aggregation process but it did not affect WHC of aggregates. Applying ultrasound on mixtures treated with NEM showed a significant impact for aggregation yields. We hypothesize that cavitations and turbulence could affect the interaction whey protein– NEM because yields were similar to results obtained in presence of ultrasound. 4.7 Conclusions This present study shows that co-denaturation of cheese whey and buttermilk concentrates leads to the formation of protein aggregates. Higher proportions of buttermilk protein in the mixtures increased yields and decreased water-holding capacity of aggregates. Preventing the formation of disulphide bonds between denatured whey proteins increased WHC of aggregates. This approach could be used to control WHC of aggregates. Rearrangement caused by ultrasound treatment during the denaturation increased yields and decreased WHC of aggregates. More work should be done in order to evaluate the potential use of the lowest hydrated aggregates in cheese making process and of the highest hydrated aggregates in yogurt manufacture. 68 4.8 Acknowledgments This work was funded by the Le Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT, Quebec city, Quebec, Canada) – Novalait Inc. (Sainte-Foy, Quebec, Canada) – Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ, Quebec city, Quebec, Canada). 69 CHAPTER 5 EFFECT COMPONENTS ON OF BUTTERMILK THE HEAT-INDUCED DENATURATION OF WHEY PROTEINS The work presented in the previous chapter demonstrated that 1) it is possible to integrate buttermilk constituents into the process of denaturation of whey proteins; 2) addition of buttermilk constituents led to the formation of mixed aggregates with different properties; 3) higher proportions of buttermilk proteins in the mixtures increased yields and decreased water-holding capacity of aggregates. The work presented in this chapter aimed at better understanding the heat-induced aggregation mechanisms of whey proteins that occur in presence of buttermilk constituents. Experiments of this chapter were conducted during an internship at Dairy Product Technology Center (DPTC) at California Polytechnic State University in San Luis Obispo under the supervision of Dr Rafael Jiménez-Flores. Results of this part of the project will be submitted essentially in this form in Journal of Agriculture and Food Chemistry. Results of this paper were also presented in one poster at 2012 Dairy Ingredients Symposium in San Francisco (CA, USA), one poster at 2012 Journée Technologique Novalait in Drummondville (QC, Canada), and in one oral presentation at 2012 ADSA Joint Annual Meeting in Phoenix (AZ, USA). 71 5.1 Résumé L’objectif de cette étude était d’obtenir une meilleure compréhension du phénomène d’agrégation induit par la chaleur entre les protéines du lactosérum et les constituants du babeurre. Des poudres de concentrés de babeurre ont été préparées à partir de babeurre régulier et de babeurre de lactosérum puis mises en solution dans un concentré de protéines du lactosérum avant d’être chauffées à 90°C (pH 4,6) en présence ou non d’un agent bloqueur de groupements thiols. Les gels SDS-PAGE en conditions réductrices ont montré que la majorité de protéines n’était pas présente sous formes natives avant le chauffage. Les images en trois dimensions montrent des interactions entre les protéines, la matière grasse et les phospholipides. Les résultats suggèrent que des particules réactives sont formées lors de la préparation du lactosérum et du babeurre. Ces agrégats “préformés” du babeurre agiraient à titre de noyaux d’agrégation pour les protéines du lactosérum. 72 5.2 Abstract The objective of this study was to better understand the heat-induced aggregation occurring between whey proteins and buttermilk constituents as affected by free thiol groups reactivity. Two different spray dried buttermilk concentrates were prepared from regular and whey buttermilks. A portion of these powders was treated by supercritical fluid extraction (CO2) to reduce the amount of non-polar lipids. Powders were hydrated in liquid whey protein concentrate, and mixtures were adjusted to pH 4.6 and heated up to 90°C. The dispersions were heated with or without thiol blocking N-ethylmaleimide (5 mM). The presence of buttermilk constituents decreased the accessibility of free thiol groups during heating. SDS-PAGE gels under non-reducing conditions showed a decreased in native whey and MFGM proteins during heating and the presence of large aggregates at the surface of the stacking and the resolving gels, even in presence of N-ethylmaleimide. SDSPAGE gels under reducing conditions showed that a large amount of proteins were not present in their native forms before heating. Confocal laser-scanning microscope pictures confirmed the presence of non-soluble particles corresponding to denatured proteins, fat globules, milk fat globule membrane fragments and proteins aggregates before heating. Three-dimensional images suggested interactions between proteins and fat globules or phospholipid. Overall, results suggest that reactive particles were formed during processing of whey or buttermilk powders involving proteins and fat through both thiol/disulfide exchanges and hydrophobic interactions. It appeared that these “pre-formed” particles would act as a nucleus during heat-induced denaturation of whey proteins and would control the size and number of protein aggregates. 5.3 Introduction Denatured and aggregated whey proteins have been widely used as fat mimetic in low fat cheese (Lawrence, 1989; Lawrence, 1993; Lebeuf et al., 1998; Punidadas et al., 1999), or in yogurt as a texture modifier (Sandoval-Castilla et al., 2004; Tamine et al., 1995). Authors reported an increase of the production yield in both processes. However, it is generally 73 accepted that the use of denatured whey proteins is limited due to their excessive waterholding capacity. Our previous work showed that it was possible to reduce the water-holding capacity of the whey protein aggregates by substituting whey with buttermilk before the heat treatment (Saffon et al., 2011). The lower water-holding capacities were obtained with 50% and 75% of buttermilk protein in the mixtures. The increase of the proportion of buttermilk proteins also significantly changed the physical and functional properties of the heat-induced aggregates as follows: an increase of the aggregation yield and consistency coefficient of the mixtures and a decrease of the flow behavior index of the mixtures as function of the buttermilk protein fraction. As a general trend, the average diameters of particles were higher in presence of buttermilk. The unfolding and aggregation mechanisms of whey proteins under heating conditions have been studied but are still unclear. Roefs and de Kruif (1994) proposed a mechanism for lactoglobulin denaturation following three steps (initiation, propagation, termination) based on a polymerization by the exposed free thiol groups of the unfolded -lactoglobulin. Schokker et al. (1999) observed the formation of oligomers of -lactoglobulin during the first steps of the denaturation by thiol/disulfide (SH/SS) interchange reactions, thiol/thiol oxidation, and non-covalent associations. All of the studies agreed on the importance of the SH/SS interactions between a heat-induced unfolded protein and a monomer. Oldfield et al. (1998) suggested that the -lactoglobulin also exposes hydrophobic groups during the initiation step that could participate in the mechanism. Later, the same group found that 15 to 20% of the -lactoglobulin were present as aggregates induced by hydrophobic interactions at 85°C (Oldfield et al., 2000). Several research groups have confirmed the role of the hydrophobic interactions during the heat-induced aggregation of -lactoglobulin proteins in experiments involving a thiol-blocking agent (Hoffmann et al., 1997; Sawyer, 1968; Xiong et al., 1993). The aggregation mechanism of pure -lactalbumin only involves hydrophobic interactions due to the absence of a free thiol group, but thiol/disulfide interactions are possible between -lactoglobulin and -lactalbumin when they are heated 74 together. Bauer et al. (2000) concluded that the initial presence of small oligomers such as monomers, dimers and tetramers are important for the heat-induced aggregation of lactoglobulin. In their review on protein aggregation, Wang et al. (2010) explained that the majority of protein aggregation phenomena are nucleation-dependent and that an aggregation nucleus initiates the inter-molecular interactions. In line with these observations, we hypothesized that the -casein (-CN) and the constituents of the milk fat globule membrane (MFGM) from the buttermilk act as nucleus for the aggregation of whey proteins. It has been shown that MFGM proteins denature at lower temperature (~ 60°C) comparatively to whey proteins (Ye et al., 2002). Later, the same group demonstrated that the maximal amounts of whey proteins associated with the MFGM are only 1 % and 0.8 % of the total -lactoglobulin and -lactalbumin respectively during heating of whole milk at temperatures ranging from 60 to 95°C during heat treatment of whole milk (Ye et al. 2004a,b). Several other research groups have found the same proportion during the heat treatment of skim milk or cream (Corredig et al., 1998; Kim & Jiménez-Flores, 1995). Buttermilk protein composition is globally similar to skim milk with a majority of caseins (75 % of total proteins), whey proteins (8 to 15% of total proteins) and MFGM proteins (Sodini et al., 2006; Walstra et al., 2006). Processing steps such as pasteurization of the cream and/or churning are likely to induce modifications of the buttermilk constituents. For example, Morin et al. (2008) showed that the solubility at pH 4.6 of MFGM proteins is affected by the heat treatment of the cream. They also demonstrated the poor coagulation properties of buttermilk are correlated to the pasteurization of cream. Several changes are initiated by this treatment such as the modification of the surface of the membrane and the increase of the accessibility of phospholipids. The objective of the present work was to better understand the aggregation mechanisms of whey proteins in presence of buttermilk constituents as affected by free thiol group 75 reactivity during heating. Different buttermilk powders were produced from buttermilk and whey buttermilk to understand the role of each buttermilk constituent. The importance of hydrophobic interactions during the heat-induced formation of aggregates was investigated by addition of a thiol-blocking agent (N-etylmaleimide) prior the heat treatments. The aggregate composition was characterized by confocal laser-scanning microscopy and gel electrophoresis. 5.4 Materials and methods 5.4.1 Materials Fresh whey (unpasteurized; milk pasteurized at 74.4°C for 16.1 seconds) from TILL and Gouda cheese productions and milk cream (unpasteurized; milk pasteurized at 74.4°C for 16.1 seconds) were obtained in California Polytechnic State University’s processing plant (San Luis Obispo, CA, USA). Hilmar Ingredients (Hilmar, CA, USA) donated the whey cream from cheddar cheese production (pasteurized at 82.2°C for 35 seconds; milk pasteurized at 73.3°C for 16 seconds). Lissamine Rhodamine B was obtained from Avanti Polar Lipids Inc. (Alabaster, AL, USA) and Fast green FCF from Sigma (ST Louis, MO, USA). Elman’s reagent was from Thermo Scientific (Rockford, IL, USA). Nethylmaleimide and all other reagents were from Fisher Scientific (Fair Lawn, NJ, USA). 5.4.2 Preparation of whey protein concentrates Fresh cheese whey was skimmed by using a pilot scale milk separator (Alfa Laval, Uppsala, Sweden). Bacterial contamination of whey was reduced by microfiltration using a pilot plant scale system (T-12 model, GEA-Niro Filtration, Hudson, WI, USA) through a 1.4 m membrane (Koch Membrane Systems, Wilmington, MA, USA). Microfiltered whey was concentrated using two 10 kDa spiral polymeric ultrafiltration membranes (Koch Membrane Systems, Wilmington, MA, USA) fitted in parallel up to a protein concentration of 4.0 % (w/v). Whey permeate was collected and frozen until used for protein blend formulations. The retentate was diluted sequentially in the feed tank with 3 diavolumes of 76 DI water. The diafiltration was stopped when a final concentration of 5.2% of protein (w/v) was obtained. The whey protein concentrate (WPC) was then frozen to – 35°C until further analysis. The concentration trial run was performed with whey from three different productions. 5.4.3 Preparation of spray dried buttermilk concentrates Regular and whey cream were processed at 12°C using a continuous butter-churn for pilot plant (Egli, Butschwil, Switzerland). Immediately after churning, regular and whey buttermilks were filtrated through cheesecloth, and skimmed, using a pilot scale milk separator (Alfa Laval, Uppsala, Sweden). Skim buttermilks were concentrated by UF (GEA R-12 model, Milwaukee WI) using two 10 kDa spiral polymeric membranes fitted parallel on the module, and then the retentate was diluted sequentially in the feed tank with 3 diavolumes of DI water. The DF was conducted until a concentration of 5% of protein (w/v) was reached. Buttermilk concentrates (BC) were spray-dried using a Niro Filterlab Spray-drier (Hudson, WI, USA) at 35 bars, and with inlet and outlet air temperatures of 185°C and 95°C, respectively. A portion of each powder was submitted to supercritical fluid extraction (SFE) (Thar Designs, Inc., Pittsburgh, PA, USA) to remove the non polar lipids according to Costa et al. (2010). Circulated deionized water (+ coolant) at 3°C was used for cooling different zones in the SFE apparatus. Carbon dioxide tanks were filled and inspected by A & R Welding Supply (San Luis Obispo, CA, USA). The system conditions were controlled manually by Windows 2000-based software. Approximately 125 g of each sample were submitted to an extraction cycle for maximum non-polar lipid removal using the following conditions: 7200 g of CO2 at a flow of 20 g.min-1, extraction pressure of 350 bar, and temperatures of 35°C. Powders were stored at 10°C under the following codes: regular buttermilk concentrate (rBC), whey buttermilk concentrate (wBC), regular buttermilk concentrate treated by SFE (rBCSFE), and whey buttermilk concentrate treated by SFE (wBCSFE). Gross composition of powders is presented in Table 5.2. 77 5.4.4 Heat-induced aggregation WPC was diluted with UF permeate to decrease the protein concentration from 5.2% to 2.6% (w/v). BC powder was then dispersed in diluted WPC to increase the total protein concentration to 5.2% (w/v) again. Whey-buttermilk mixtures were stirred overnight at 4°C. Denaturation process followed the previous proposed design (Chapter 4). Briefly, 250 mL blends were adjusted to pH 4.6 with 1N HCL and heated from 8°C to 90°C in a thermostatically controlled water-bath maintained at 97°C for a total time of 25 minutes. Mixtures were constantly stirred using a caframo lab mixer during heating (speed 5). Nondiluted WPC (5.2% protein) was used as a control. All heating experiments were repeated three times. 5.4.5 Treatment with N-ethylmaleimide (NEM) NEM (5 mM) was used to completely block the exposed free thiol group during heating (Alting et al., 2000). NEM was added to mixtures before overnight stirring. 5.4.6 Analytical method 5.4.6.1 Composition Gross composition was determined using recommended methods (AOAC, 2003). Nitrogen content was determined using Kjeldahl method, and converted to protein using a conversion factor of 6.38. Lipid content was determined using the Mojonnier method, total solids content by forced air oven drying, and ash content by gravimetric method of incineration in a furnace at 550°C. Lactose content was determined by enzymatic method (IDF, 2001). 5.4.6.2 Free thiol groups concentration Accessible thiol groups concentration was measured according to Ellman (1959). 250 L of mixtures were taken every five minutes during heating and mixed with 2.5 ml of 0.1M 78 sodium phosphate buffer (containing 1mM EDTA; pH 8.0), and 0.1 mL of DTNB reagent solution (4 mg in 1 mL of sodium phosphate buffer). The tubes were mixed by hand and incubated at room temperature for 30 min. Absorbance was measured at 412 nm using a Spectra Max Plus spectrophotometer (Sunnyvale, CA, USA) after setting a blank. Absorbance of every sample was also measured without the addition of DTNB. The free thiol concentration was expressed as mole per gram of proteins. 5.4.7 Gel electrophoresis Protein profiles were determined using the sodium dodecyl sulfate polyacrylamide gel electrophoresis technique (SDS-PAGE) under reducing and non-reducing conditions according to Laemmli (1970). Samples were taken every five minutes during heating and diluted 1:10 in deionized water, and then with alkaline buffer (non-reducting condition). Diluted samples were also diluted 1:2 in Laemmli buffer (95% Laemmli solution + 5% Mercaptoéthanol (v/v)) and heated for five minutes at 95°C (reducting condition). After a quick spin, supernatants were loaded in 15% acrylamide SDS gels. Analyses were conducted in a mini-Protean II system (Bio-Rad Laboratories Ltd., Hercules, Ca, USA) at 90 V. Gels were stained with Coomassie Brilliant Blue (Sigma Chemical Compagny, ST Louis, MO, USA). Protein classes were determined according to their molecular weight by comparison with a molecular weight standard (Precision Plus All Blue Protein Standard, Bio-Rad laboratories Ltd.). The aggregated material of each mixture was separated at the end of heating (25 minutes) by centrifugation (Beckman L7-35, Brea, CA, USA ) at 15,000g for 20 minutes at 20°C. Samples were taken and diluted 1:2 in deionized water, and then 1:2 in the proper buffer as described above. Samples corresponded to the soluble material were labelled “sup” in the legend of the gels. 79 5.4.8 Confocal laser scanning microscopy Samples of the mixtures were taken at 0, 15, and 25 minutes of heating and diluted 1:2 in deionized water. 50 L of the diluted samples were dyed with 2 L of Lissamine Rhodamine B and 2 L of Fast Green FCF. Samples were mixed, and then stored 20 minutes in the dark. 25 L were deposited on a slide, and fixed to the slide with agarose (50 L; 0.5% w/v in deionized water). Observations were performed using an Olympus Fluoview FV1000 inverted confocal laser scanning microscope (Olympus America Inc., Center Valleyr, PA, USA) equipped with four lasers. A 100x oil-immersion objective (UPLSAPO 100X) was used for all images. Four pictures per samples (800 x 800 pixels; 63.16 x 63.16 µm) were taken and analyzed by using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/). Briefly, pictures were transformed in 8-bit, and then made binary. The software counted particles, and the average particle surface area was expressed in m2. Three-dimensional images were obtained by scanning the sample across a defined section along the z-axis. 5.4.9 Statistical analysis All experiments were performed in triplicate, so each value represents the mean of three measurements. Three different and independent variables were selected and named as follows: Composition that expressed the effect of addition of different buttermilk concentrates, Time that expressed the effect of the holding time in the water bath, NEM that expressed the effect of blocking the free thiol groups. Data was tested according to a factorial design 5*2*2 for the number of particles, a factorial design 5*3*2 for the surface area of particle, and a factorial design 5*6 only for the concentration of free thiol groups (any results with NEM). Statistical analysis was carried out using SAS software (SAS Institute Inc., Cary, NC, USA), using the PROC GLM procedure. Mean comparisons were performed using a Duncan post-test if appropriate. Table 5.1 summarizes significance for each treatment on each variable. Results were considered significantly different when P < 0.05. 80 Table 5.1: Summary of the significance (P value) by the analysis of variance of data of each contrast for each variable (n = 3). Significance (P values) Contrast [free –SH]* Number of particles Surface area** Model < 0.0001 0.7211 < 0.0001 Composition 0.0002 0.3469 < 0.0001 Time (B) < 0.0001 0.6660 < 0.0001 NEM (C) n.a. 0.7058 0.0230 A*B 0.9670 0.4033 0.0150 A*C n.a. 0.9336 0.6037 B*C n.a. 0.6749 0.0471 A*B*C n.a. 0.3138 0.8118 (A) * The effect “blocs” was evaluated significant, so the software automatically re-adjusted the model. This result means that three whey were different. The first whey was a mix from TILL and Gouda cheeses production, while the second and third whey were from TILL cheese production only. ** The effect “blocs” was evaluated significant, so the software automatically re-adjust the model. This result means that there is a variability regarding the analysis of the pictures. Pictures were taken in order to illustrate samples in their entirety (small particles, large particles, etc) and corresponded to a mean of four pictures per samples. 81 5.5 Results and discussion 5.5.1 Composition The composition of the WPC and BC powders is shown in the Table 5.2, and the protein profile of the mixture in Figure 5.1. Table 5.2: Composition (% DM) of whey protein concentrate (WPC) and the different buttermilk concentrates (BC) powders used for the heating experiment with rBC: regular buttermilk concentrate, rBCSFE: regular buttermilk concentrate after supercritical fluid extraction, wBC: whey buttermilk concentrate, wBCSFE: whey buttermilk concentrate after supercritical fluid extraction. Liquid WPC Protein Fat Lactose Ash 5.2 0.2 0.8 0.7 4.4 4.4 0.2 0.4 % of Dry Matter Powder rBC 69.2 1.7 22.4 1.4 4.8 1.4 4.5 0.0 rBCSFE 79.1 0.8 5.7 0.4 7.5 2.7 6.4 0.0 wBC 72.1 5.1 20.3 0.8 4.8 1.4 2.7 0.2 wBCSFE 70.8 2.2 13.8 1.5 12.9 5.2 2.9 0.0 Protein content of regular buttermilk concentrate (rBC) and whey buttermilk concentrate (wBC) was similar and averaged 70.6%. The fat content decreased from 22.4 1.4% to 5.7 0.4% and from 20.3 0.8% to 13.8 1.5% after CO2-SFE treatment respectively on regular and whey buttermilk concentrates due to the removal of the non-polar lipids. The protein profile of the mixtures showed that -lactoglobulin was the main protein in the mixtures (Figure 5.1). Some proteins from the MFGM membranes were found, and are identified according to their molecular weight reported by Singh (2006) such as Periodic acid Shiff 6/7 (PAS 6/7: 48-54 kDa), Adipophilin (ADPH; 52 kDa), Butyrophiline (BTN: 82 67 kDa), Cluster of Differenciation or Periodic acid Shiff IV (CD36 or PAS IV: 76-78 kDa), Periodic acid Shiff III (PAS III: 95-100 kDa) Xanthine dehydrogenase/oxydase (XDH/XO: 150 kDa), Mucine 1 (MUC1: 160-200 kDa), and BRCA 1 (210 kDa). Fatty acid binding protein (FABP; 13 kDa) was not found. Casein was found in the mixtures WPC + rBC and WPC + rBCSFE, and was present as trace level in the WPC, WPC + wBC and WPC + wBCSFE mixtures. Figure 5.1: 15% Non-reducing SDS gels of the mixtures. 1: protein standard, 2: WPC, 3: WPC + rBC, 4: WPC + rBCSFE, 5: WPC + wBC, 6: WPC + wBCSFE. 5.5.2 Free thiol groups concentration Heating time and mixture composition had significant effect on free thiol groups concentration, but the interaction between these two variables was not significant (Table 5.1). As presented in Figure 5.2, the concentration of free thiol groups significantly increased from 0.29 ± 0.27 µmol/gprot to 8.16 ± 3.19 µmol/gprot after heating 10 minutes, and then remained constant (Figure 5.2a). It is known that whey proteins, especially lactoglobulin and Bovin Serum Albumin, are subjected to irreversible conformational changes upon heating leading to the exposition of free thiol groups (Pantaloni, 1964; Havea, 200). These changes are negligible at low temperatures (< 55°C), but increase with the temperature (~ 80°C). 83 Figure 5.2: Effects of time (a) and composition of the mixture (b) on the accessibility of free thiol groups during heating with constant stirring at temperature up to 90°C at pH 4.6. As expected, the concentration of free thiol groups was higher in the WPC comparatively to other mixtures (Figure 5.2b). In presence of caseins from rBC or rBCSFE, the exposure of free thiol groups was reduced but not significantly. The low levels detected were due to a lower proportion of -lactoglobulin in rBC, but were surprisingly higher comparatively to our expectations. Caseins do not expose free thiol groups during heating. It was expected that the 1:1 ratio between whey and buttermilk material would reduce the concentration of free thiol groups by 50% in presence of caseins. Surprisingly, the concentration of free thiol groups was lower in the WPC + wBC mixture (not significant) and WPC + wBCSFE mixture (significant) comparatively to in WPC. The reactivity of the proteins and constituents from whey buttermilk seems, in consequence, to be more affected by the butter-making (pasteurization, churning) that the constituents from regular buttermilk. Morin et al. (2008) have already reported some effects of the cream pasteurization such as the poor coagulation properties of regular buttermilk, but to our 84 knowledge no study gives examples of the impact of the processing of whey buttermilk on its constituents. Knowing that the constituents of whey buttermilk have been subjected to two pasteurizations (73.3°C for 16 sec and 82.2°C for 35 sec), the lower concentration of free thiol groups in WPC + wBC and WPC + wBCSFE mixtures could be explained by a loss of reactivity of whey buttermilk constituents during processes including oxidation of free thiol group or SH/SS interchange reactions between the different proteins. The concentration of thiol groups was equal to 0 mole/gprot in presence of N-ethylmaleimide (NEM) regardless to heating time or composition (data not shown). 5.5.3 Gel electrophoresis Figures 5.3 to 5.5 present the evolution of native proteins in WPC and mixtures as a function of heating time. 85 Figure 5.3: Evolution of the protein profile of WPC as a function of heating time (min) under nonreducing (a) and reducing (b) conditions. Figure 5.4: Evolution of the protein profile of the WPC + rBC mixture as a function of heating time (min) under non-reducing (a) and reducing (b) conditions. Figure 5.5: Evolution of the protein profile of the WPC + wBC mixture as a function of heating time under non-reducing (a) and reducing (b) conditions. 86 Content of native -lactoglobulin (-LG) and -lactalbumin (-LA) decreased after 10 minutes of heating in WPC (Figure 5.3a), but after 15 minutes in presence of caseins (WPC + rBC) (Figure 5.4a) but only after 5 minutes in WPC + wBC (Figure 5.5a). In all the mixtures the amount of native BTN decreased as a function of the heating time. The amounts of CD 36 or PAS IV, MUC 1 or BRCA 1, and PAS 6/7 remained constant in WPC, but were decreasing after 10 minutes of heating in the mixtures. Native XO was present only before heating in all the blends. In all mixtures, the amount of residual proteins in the wells seemed to increase from 0 to 15 minutes of heating and then slightly decreased until the end of heating (25 minutes). As a general trend, electrophoresis separation of proteins conducted under non-reducing conditions showed that native forms of XO and BTN were lost in the first 5 to 15 minutes of heating following by the decrease of the native forms of -LG and -LA. Ye et al. (2002) already reported that the MFGM proteins are sensitive at lower temperatures comparatively to the other dairy proteins. Later, they have proposed mechanisms of heat-induced aggregation between MFGM proteins, lactoglobulin, and -lactalbumin (Ye et al., 2004a; Ye et al., 2004b). These heat-induced associations appear after dissociation of -LG into monomers and before exposition of its free thiol groups. Free thiol groups of MFGM proteins initiate SH/SS interchange reactions. After reduction, residual proteins were observed in the wells for the mixture, but not for the WPC, and the MFGM proteins PAS III and FABP. The bands of -lactoglobulin, lactalbumin and MFGM proteins were more abundant after reduction and even for the sample before heating. This result clearly demonstrates that protein aggregates were formed previously to the heat-induced denaturation during the preparation of the WPC or powders. Gels under reducing conditions show that whey proteins, caseins, MFGM proteins were involved in the formation of aggregates during the process of preparation through SH/SS interchange reactions (mainly) and non-covalent exchanges. Residual proteins remained in the well after reduction indicating that some high molecular weight particles were formed through other types of interactions during heating such as hydrophobic interactions and van der Waals forces. 87 Figure 5.6: Evolution of the protein profile of the WPC (a), WPC + rBCSFE mixture (b), and WPC + wBC (c) in presence of NEM as a function of heating time under non-reducing conditions. Figure 5.6 shows the evolution of native proteins during heating in presence of NEM. In all mixtures, the amount of XO was the only band to visually decrease in intensity after 5 or 10 minutes. The amount of residual proteins in the wells was constant for the WPC except at 5 minutes were the amount was higher. Some particles remained in the wells after reduction (result not shown). In WPC + rBC, WPC + rBCSFE, and WPC + wBC, the amount of residual proteins in the wells stayed constant, even after reduction. For WPC + wBCSFE mixture, the amount of residual proteins in the wells was higher at the end (after 10 minutes) than at the beginning of heating. In presence of the thiol-blocking agent, high molecular weight particles were present in the well or at the beginning of the resolving gels. Sawyer (1968), Xiong et al. (1993), Hoffmann and Van Mil (1997) have observed the formation of whey protein aggregates during a heat treatment in the presence of Nethylmaleimide (NEM) via non-covalent interactions, especially hydrophobic interactions. Gels under reducing conditions (results not shown) confirmed that non-covalent interactions also contributed to the formation of protein aggregates during the preparation and the heating between whey proteins and buttermilk constituents. The role of non- 88 covalent interactions in dairy protein aggregation is, however, still not well defined. Oldfield et al. (2000) have demonstrated that 15 to 20 % of the available -lactoglobulin aggregated via hydrophobic interactions during the first steps of heat-induced denaturation followed by covalent interactions at 85°C. 5.5.4 Confocal laser scanning microscopy images Figures 5.7 and 5.8 present the average distribution of proteins (green) and phospholipids (red) at 0, 10 and 25 minutes of heating in WPC and WPC + wBC mixture respectively. Figure 5.7: Confocal laser scanning miscroscope pictures taken at 100X of WPC at 0, 15, and 25 minutes of heating before (top) and after (bottom) analysis with ImageJ. 89 Figure 5.8: Confocal laser scanning microscope pictures taken at 100X of WPC + rBC mixture at 0, 15, and 25 minutes of heating before (top) and after (bottom) analysis with ImageJ. At t0, more non-soluble proteins and phospholipids were found in the WPC + wBC mixture than in WPC. Phospholipids were not visible at t0 in whey. Phospholipids remained close to protein particles in WPC, but did not interact with them. In WPC + wBC mixture, phospholipids seemed to be part of the protein aggregates, as evidenced by close contacts between fat globules or fat globules with proteins or both. The binary transformation confirmed that interactions between phospholipids and proteins were hardly found in whey, but phospholipids were involved in the structure of the protein aggregates in the presence of buttermilk constituents and even since the beginning of heating. Statistical analysis (Table 5.1) showed no significant difference between samples for the number of particles per images observed at 100X, but the interactions composition*heating time (P = 0.0150) and heating time*NEM (P = 0.0471) had significant effect on the average surface area. Figure 5.9a represents the evolution of the surface area of particles observed at 100X as a function of composition and heating time. 90 size time*condi 4 a) b) Average area observed at 100X (µm2) Average area observed at 100X (µm2) 4 3 3 2 2 1 0 normal NEM 1 0 10 20 Time (min) 30 0 0 15 Time (min) 25 Figure 5.9: Evolution of the average surface area observed with a confocal laser scanning microscope at 100X as a function of interactions time*composition (a) and time*condition. In a), mixtures were presented as follows: = WPC; = WPC + rBC; = WPC + rBCSFE; = WPC + wBC = WPC + wBCSFE. In b), conditions were presented as follows: = absence of NEM; = presence of NEM. At every time, the average surface area of particles in WPC was higher comparatively to the blends. Figure 5.9b represents the evolution of the surface area of particles observed at 100X as a function of time and in presence or absence of NEM. As a general trend, the presence of NEM significantly reduced the average surface area of particles in the mixtures. Confocal laser-scanning microscope pictures analysis revealed that WPC and mixtures contained non-native proteins and non-soluble MFGM material before heating, especially in whey buttermilks. Small protein aggregates and protein – phospholipid/MFGM aggregates were also present. The loss of solubility and the formation of small aggregates were probably due to the pasteurization of the milk/cream or the mechanical forces occurring during the cheese-making process, butter-making process, filtration, and spray drying. Increasing the time of heating and the temperature did not increase the number of non-soluble particles in WPC or in the mixtures, while the particle size stayed constant in mixtures. Non-soluble particles already present in the mixtures became bigger by association with denatured proteins or small aggregates in mixtures. 91 Three-dimensional images were taken to visualize interactions between proteins and phospholipids/MFGM. Pictures confirmed that some non-native proteins or protein aggregates were found even at t0 in whey and mixtures. Figure 5.10 shows that some interactions between proteins and phospholipids/MFGM were also present at t0, but only in WPC + buttermilk mixtures. The presence of phospholipids (red) or interactions between proteins and phospholipids/MFGM were more important after treatment of the buttermilk powders by C02-supercritical fluid extraction (not presented). After heating, it appeared that phospholipids were mainly present at the surface of the proteins aggregates, but some small red dots were also located inside the structure of aggregates at t15 and t25. This phenomenon was observed in every mixtures, but mainly in WPC + wBC. Figure 5.10: Three-dimensional confocal laser scanning microscope pictures of WPC + rBC mixture before heating. Three-dimensional pictures confirm that the interaction between phospholipids/MFGM and proteins were present in the mixtures, including in samples treated by NEM, but were 92 limited in WPC. Surprisingly, these non-soluble particles were almost non-detected in WPC. Similar trend was observed with liquid whey buttermilk after UF/DF (results not shown), meaning that the presence of a large amount of non-soluble particles (denatured proteins, fat, and protein aggregates) in the mixture before heating is not to be attributed to poor dispersibility of buttermilk powder ingredients. Three-dimensional pictures confirmed that some proteins covered fat globules or MFGM fragments. More surprisingly, phospholipids, viewed as little red dots were present on the side of the aggregates or even buried in the structure of the protein aggregates. This phenomenon was more frequent after C02-SFE treatment of the buttermilk. Unfortunately, it is impossible to identify the nature of the small aggregates and their mechanism of interaction. They could be a small fragment of MFGM (proteins + phospholipids) or free phospholipids. Interactions between phospholipids (lecithin, lysolecithin, phosphatidylcholine) with s1-casein or lactoglobulin have been demonstrated (Barratt et al., 1974; Brown et al., 1983; Spector et al., 1970). In all cases, the complexation of phospholipids and proteins involves first an ionic attraction between the charged amino-acid residues (protein) and the polar head group (lipid), and then hydrophobic interactions to stabilize the complex. In the case of lactoglobulin, its helix formation is determinant (Brown et al., 1983). Interactions between phospholipids and proteins could also explain the presence of particles in the wells of gel after reduction. 5.5.5 Proposed mechanism for association of whey proteins with buttermilk constituents The results of this study suggest that the heat-induced association between whey proteins from WPC and buttermilk constituents involves the following steps: Preliminary reactions: Formation of small aggregates during the preparation of buttermilk powders. Regarding the temperatures involved in the process of preparation of the BC powders it appears evident that proteins from buttermilks are pre aggregated, including whey proteins, caseins, and MFGM proteins. Even if the mechanisms of heat-induced aggregation of dairy proteins are still uncertain, it has been well demonstrated that whey 93 proteins interact with -casein at the surface of the casein micelles (Livney et al., 2004; Morr et al., 1968; Sawyer, 1968; Vasbinder et al., 2003a) or with soluble -casein (Alting et al., 2003; Anema, 2008; Anema et al., 2003a) or with MFGM proteins (Ye et al., 2004a; Ye et al., 2004b) during heating. Formation of aggregates may also integrate free phospholipids. Step 1: «Activation» of the pre formed aggregates from BC. Wang et al. (2010) reminded us that most of the protein aggregation mechanisms are nucleation-dependent, and are initiated by formation of an aggregation nucleus. Krishnan et al. (2003) defined the aggregation nucleus as a small particle with a size close to a -LG dimer, while Baynes et al. (2005) observed small -LG multimers as nuclei. Schokker et al. (1999) observed the presence of non-native dimers and oligomers in the early stages of -LG denaturation. In presence of buttermilk, it appears that aggregation nucleus could be non-native whey proteins or small protein aggregates or MFGM fragments or -caseins or a combination. Step 2: Aggregation of denatured proteins from WPC with aggregated buttermilk constituents. Based on studies from Schokker et al. (1999) and Baynes et al. (2005), monomers or small aggregates of whey proteins were probably rapidly incorporated into the early-formed particles to form large aggregates through mainly SH/SS interchange reactions and non-covalent interactions. Confocal pictures revealed a higher accessibility of MFGM fragments in whey buttermilk but not in regular buttermilk, presumably due to the more complex interactions of caseins that are present in regular buttermilk. 5.6 Conclusions The present study shows that the heat-induced denaturation mechanism of whey proteins changes in the presence of buttermilk constituents. The presence of caseins, whey proteins, 94 and MFGM components from buttermilk reduced the concentration of free thiol groups, and delayed the loss of native -lactoglobulin and -lactalbumin. In presence of whey buttermilk, the concentration of free thiol groups was also reduced, but the loss of native whey proteins was accelerated. SDS PAGE confirmed that MFGM proteins seemed involved in the aggregation mechanism of whey proteins, however interaction of MFGM/phospholipids were poorly represented in WPC compared to mixtures. Buttermilk contained a small but perhaps important amount of non-soluble proteins. Small aggregates were also present, and involved both thiol/disulfide and hydrophobic bonds between whey protein, casein and MFGM components. Confocal pictures revealed a higher accessibility of MFGM fragments in whey buttermilk concentrate but not in regular buttermilk concentrate, presumably due to the more complex interactions of caseins that are present in regular buttermilk. The presence of phospholipids buried in the structure of the aggregates suggests that they could participate to the formation of the aggregates in the early stages of heating. 5.7 Acknowledgements This work was funded by the Le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT, Quebec City, Quebec, Canada) – Novalait Inc. (Sainte-Foy, Quebec, Canada) - Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ, Quebec City, Quebec, Canada). 95 96 CHAPTER 6 EFFECT OF HEATING OF WHEY PROTEINS IN THE PRESENCE OF MILK FAT GLOBULE MEMBRANE EXTRACT OR PHOSPHOLIPIDS FROM BUTTERMILK The work presented in the previous chapter demonstrated that 1) protein aggregates are preformed during the preparation of the concentrates; 2) pre-formed aggregates act as aggregation nucleuses during the heat treatment of the mixtures. Results also suggested that phospholipids from regular or whey buttermilk are integrated to the formation of the protein aggregates. The work presented in this chapter aimed at proving that phospholipids from buttermilk can act as a nucleus during aggregation of whey proteins. Experiments of this chapter were conducted during an internship at Dairy Product Technology Center (DPTC) at California Polytechnic State University in San Luis Obispo under the supervision of Dr Rafael Jiménez-Flores. Results of this part of the project will be submitted after the publication of Chapter 5. Results of this paper will be presented in a oral presentation at 2013 ADSA Joint Annual Meeting in Indianapolis (IN, USA), and in a poster form at 2013 STELA Colloque in Montreal (QC, Canada). 97 6.1 Résumé L’objectif de cette étude était de déterminer si les phospholipides interviennent lors de l’agrégation des protéines du lactosérum. Les phospholipides ont été obtenus à partir d’un isolat de membranes de globule de matière grasse (MFGM) ou d’une poudre commerciale de phospholipides laitiers (PL). Les PL ont été mis en solution (1 % p/v) avec un isolat de protéines du lactosérum (IPL) (5 % p/v) et chauffés à 65 ou 80°C (pH 4,6 ou 6,8). L’addition de phospholipides n’a pas affecté significativement la libération de groupements thiols mais un délai de 5 min a été observé pour la diminution des formes natives des protéines à pH 6,8. Les profils de chromatographie ont révélé une absence de phospholipides libres après 20 min de chauffage à 80°C à pH 4,6 dans le mélange IPL-PL. Les images en trois dimensions ont confirmé la présence d’interactions entre les protéines et les PL dans les mélanges. 98 6.2 Abstract The objective of this study was to understand if the phospholipids from buttermilk can act as heat-induced aggregation nucleus for whey proteins. For this purpose, phospholipids were obtained from a milk fat globule membrane (MFGM) extract from whey buttermilk or from a commercial dairy phospholipids powder (PL). Whey proteins isolate (WPI) was dispersed in water to a concentration of 5% protein (w/v), and the pH was adjusted to 4.6 or 6.8. Solutions were heated to 65°C or 80°C for 25 min under constant stirring. MFGM extract or PL powder were dispersed in the WPI solutions at 1.0% (w/v) before heating. The aggregate composition was characterized with the Ellman’s reagent (free thiols), onedimensional gel electrophoresis technique, thin layer chromatography (TLC), and three dimensional confocal laser-scanning microscopy. All experiments were performed in triplicate and the concentration of free thiol groups was tested according to a factorial design. Addition of phospholipids or MFGM extract did not significantly affect the liberation of free thiol group of whey proteins but delayed the loss of native form of major whey proteins by 5 min at pH 6.8. TLC profiles showed no traces of free phospholipids after 20 min of heating in WPI-PL mixtures at 80°C at pH 4.6 when phospholipids were added to the mixtures. 3D images confirmed that only few interactions occurred between whey proteins and MFGM proteins/phospholipids in the whey solution, while the interactions were frequent in presence of MFGM extract or phospholipids. Overall, our results suggest that phospholipids from buttermilk are involved in the formation of protein aggregates at pH 4.6 through the MFGM fragments at 65°C while they seem to directly interact with the proteins at 80°C. 6.3 Introduction Microparticulated whey proteins (MWP) have been developed to mimic fat globules and to increase retention of whey protein in cheese (Lawrence, 1993; Punidadas et al., 1999) or to modify the texture of yogurts (Sandoval-Castilla et al., 2004; Tamine et al., 1995). Whey proteins are concentrated by ultrafiltration, and then heat-denatured at 90-95°C for several 99 minutes at acid (4.2 to 4.6) or neutral (6.2 to 6.7) pHs (Lebeuf et al., 1998; Punidadas et al., 1999). The addition of MWP has to be controlled due to the excessive water-holding capacity of denatured whey proteins. In consequence, MWP are used in very low amounts during formulation (~ 1.0 to 1.5 % of total weight). Our previous work showed that it was possible to control the water-holding capacity of heat-denatured whey proteins by substituting whey with buttermilk before the heat treatment (Saffon et al., 2011). Our latest results demonstrated that protein aggregates were formed during butter-making process involving whey proteins, caseins, and milk fat globule membrane (MFGM) proteins. These preformed aggregates acted as an aggregation nucleus during whey protein denaturation (Chapter 5). Results also suggested that phospholipids could act as an aggregation nucleus too. Brown et al. (1983) concluded that -lactoglobulin in its native form is generally unable to form a complex with phosphatidylcholine or lipids. A certain degree of unfolding is necessary to expose initially buried helical regions and hydrophobic groups. According to their model, lipid-protein interactions are initiated by a ionic attraction between charged amino acid residues in the protein and the polar head group of the lipid in order to position the lipid. Then the complex is stabilized by hydrophobic interactions between the hydrophobic side of the helix and the hydrocarbon chain of the lipid. Similar observations have been reported by Bos and Nylander (1996) and Ong et al. (1981). Dufourcq et al. (1977) and Ong et al. (1981) observed that the charged residues involved are lysine or arginine for the protein and phosphate or carboxylic groups (negative) for the phospholipids. In their review on the binding properties of -lactoglobulin, Kontopidis et al. (2004) proposed that the monomeric form of the protein has one ligand-binding site, and potentientially another weaker site (not yet located) that can bind linear molecules such as the fatty acids hydrocarbon chains of the hydrophobic tails of the phospholipids or retinol. Phospholipids can also be attached to whey proteins through the MFGM proteins. Fat globules are disrupted during churning, so the membrane is segmented and contains a 100 complex mix of proteins, phospholipids, glycoproteins, triglycerides, cholesterol, enzymes, and different minor components. Ye et al. (2004a,b) demonstrated that whey proteins and MFGM proteins interact during heating of milk. The objective of the present work was to determine if phospholipids from buttermilk can interact with whey proteins during a heating treatment. Whey proteins were heated in solution in presence of MFGM extract from whey buttermilk or in presence of phospholipids from regular buttermilk. The aggregate composition was characterized by gel electrophoresis, thin layer chromatography, and three-dimensional confocal laser-scanning microscope images. 6.4 Materials and methods 6.4.1 Materials Hilmar Ingredients (Hilmar, CA, USA) donated the whey cream (pasteurized at 82.2°C for 35 seconds; milk pasteurized at 73.3°C for 16 seconds) from cheddar cheese. BiPro whey protein isolate (WPI) was obtained from Davisco Foods Ingredient, Inc. (Eden Prairie, MN, USA), and phospholipids concentrate 700 (the following components have been detected by HPLC: neutral lipids, lactosylceramide, sphingomyeline (SM), phosphatidylcholine (PC), phosphatidyléthanolamine (PE), phosphatidylinositol (PI), phosphatidylserine (PS)) was manufactured at Fonterra Ltd (Edgecumbe, New Zealand). Lissamine Rhodamine B was obtained from Avanti Polar Lipids Inc. (Alabaster, AL, USA), and Fast Green FCF from Sigma (St-Louis, MO, USA). Ellman’s reagent was from Thermo Scientifics (Rockford, IL, USA). All other reagents were from Fisher Scientific (Fair Lawn, NJ, USA). 6.4.2 Isolation of the milk fat globule membrane Whey cream was processed at 12°C using a continuous butter churn for pilot plant (Egli, Butschwil, Switzerland). Immediately after churning, the whey buttermilk was filtrated 101 through cheesecloth. Filtrated buttermilk was centrifuged at 25,000 rpm for 120 minutes at 4°C (Beckman L7-35 Ultracentrifuge, Beckman Coulter, Inc., USA). The supernatant was removed, and the pellet was dissolved in PBS buffer in presence of Triton X 100. The hydrated pellet was centrifuged at 25,000 rpm for 120 minutes at 4°C. The supernatant was removed, and the pellet hydrated in water to a 10% (w/v) solution. The solution was mixed in a blender, and sonicated for 1 minute, and finally stored at -20°C until further analysis. 6.4.3 Heat treatment WPI powder was dispersed in deionized water to a concentration of 5.0% protein (w/v), and the pH was adjusted to 4.6 with 1N HCl or 6.8 with 1N NaCl. The solution was then dispersed overnight at 4°C. WPI solutions were heated from 8°C to 65°C or 80°C in a thermostatically controlled water-bath maintained at 72°C or 87°C for a total time of 25 minutes. Solutions were constantly stirred using a caframo lab mixer during heating (speed 5). Heated solutions were cooled down at room temperature. One gram of MFGM extract or phospholipids powder per 100 mL of WPI solution were also added before stirring. All heating experiments were repeated three times. 6.4.4 Analytical method 6.4.4.1 Concentration of free thiol groups Accessibility of free thiol groups was measured according to Ellman (1959). 250 L of mixtures were taken every five minutes of the heating time, and mixed with 2.5 mL of 0.1M sodium phosphate buffer (containing 1mM EDTA; pH 8.0), and 0.1 mL of DTNB reagent solution (4 mg in 1 mL of sodium phosphate buffer). The tubes were mixed by hand, and incubated at room temperature for 30 minutes. Absorbance was measured at 412 nm using Spectra Max Plus spectrophotometer (Sunnyvale, CA, USA) after setting a blank. Absorbances of each sample were also measured without the addition of DTNB. The free thiol concentration was determined in mole/gprot. 102 6.4.4.2 Gel electrophoresis Protein profiles were determined using the sodium dodecyl sulfate polyacrylamide gel electrophoresis technique (SDS-PAGE) under non-reducing conditions according to Laemmli (1970). Samples were taken at 0, 10, 15, and 20 minutes of the heating time, and diluted 1:10 in deionized water, and then with alkaline buffer (non-reducing conditions). 25 L of samples were loaded in 15% acrylamide SDS-PAGE gels. Analyses were conducted in a mini-Protean II system (Bio-Rad Laboratories Ltd., Hercules, CA, USA) at 90 V. Gels were stained with Coomassie Brilliant Blue (Sigma Chemical Compagny, St Louis, MO, USA). Protein classes were determined according to their molecular weight by comparison with a molecular weight standard (Precision Plus All Blue Protein Standard, Bio-Rad laboratories Ltd.). 6.4.4.3 Thin layer chromatography Samples from the solution WPI+PL were taken at 0, 10, 15, 20, and 20 minutes of the heating time, and 5 L were loaded directly on a glass backed silica TLC plate using a glass capillary. Plates were deposed in a chamber filled with a polar solvent mixture containing 65 mL of chloroform, 25 mL of methanol, and 4 mL of deionised water. Plates were maintained in chamber for 60 to 90 minutes, and then dry into vacuum oven for 15 minutes. Dried plates were placed overnight in iodine chamber with iodine pellet. 6.4.5 Confocal laser-scanning microscopy Sample of the mixtures were taken at 0, 15, and 25 minutes of heating and diluted 1:2 in deionized water. 50 L of the diluted samples were dyed with 2 L of Lissamine Rhodamine B and 2 L of Fast Green FCF. Samples were mixed and stored 20 minutes in the dark. 25 L of the dyed solution were deposited onto a slide, and fixed with agarose (50 L; 0.5% w/v in deionized water). Three-dimensional images were taken by scanning the sample across a defined section along the z-axis using an Olympus Fluoview FV1000 inverted confocal laser-scanning microscope (Olympus America Inc., Center Valley, PA, USA). 103 6.4.6 Statistical analysis All heating treatments and experiments were performed in triplicate. Data from the concentration of free thiol groups data was tested according to a factorial design composition*time*temperature*pH (3*2*2*2). Statistical analysis was carried out using SAS software (SAS Institute Inc., Cary, NC, USA) using the PROC GLM procedure. Results were considered significantly different when P < 0.05. 6.5 Results and discussion 6.5.1 Free thiol group concentration The statistical analysis of data presented in Table 6.1 shows that only interactions composition*temperature (P = 0.0094) and temperature*pH (P = 0.0049) were significant on the concentration of free thiol groups. Figure 6.1 presents the effects of these interactions on the accessibility of thiol groups during heating. 104 Table 6.1: Summary of the significance calculated by the analysis of variance of free thiol group concentration (n = 3). Significance (P value) Contrast [Free –SH] Significance Contrast [Free –SH] Significance Composition (A) 0.5958 n.s B*C 0.1335 n.s Time (B) 0.0141 n.s B*D 0.0924 n.s Temperature (C) 0.0280 n.s C*D 0.0049 ** pH (D) 0.0626 n.s A*B*C 0.4071 n.s A*B 0.7979 n.s A*B*D 0.8675 n.s A*C 0.0094** ** A*C*D 0.6712 n.s A*D 0.4444 n.s B*C*D 0.9643 n.s a) 2.0 b) 2.5 2.0 [-SH] free (10-5 mole.gprot-1) [-SH] free (10-5 mole.gprot-1) 1.5 1.5 1.0 1.0 0.5 0.5 0.0 C pH 4. 6_ 80 ° C pH 6. 8_ 80 ° C pH 6. 8_ 65 ° C 65 ° 4. 6_ pH 0° C 0° C W .+ P. 8 .8 0° C W .+ M W .8 5° C W .+ P. 6 5° C .6 W .+ M W .6 5° C 0.0 Figure 6.1: Effect of interactions composition*temperature (a) and temperature*pH (b) on the exposition of free thiol groups during heating of WPI (W.), WPI+MFGM extract (W.+M.), and WPI+phospholipid powder (W.+P.) under constant stirring. As shown in Figure 6.1a, the average concentration of free thiol groups exposed during heating significantly increased in the mixtures between 65°C and 80°C. 105 As shown in Figure 6.1b, the average concentration of free thiol groups exposed during heating at pH 4.6 or 6.8 significantly increased between 65°C and 80°C. It has been known since Pantaloni (1964) that the exposition of the free thiol group of the -lactoglobulin (Cys121) is initiated at 70°C, but in a low proportion (6%), and defined the irreversible conformation change of the protein. However, -lactoglobulin is irreversibly unfolded at a temperature close to 78.5°C, making the free thiol groups accessible for thiol/disulfide exchanges (Schokker et al., 1999). Increasing the temperature above 80°C led to increase the liberation of free thiol groups. Hoffmann et al. (1997) demonstrated that aggregation of -lactoglobulin is related to the pH dependent reactivity of the free thiol group. Their results showed that the accessibility of the free thiol group of the protein is better at pH = 8.0 comparatively to pH < 7.0 due to the pK value of thiol group around 8.2. To summarize, the presence of milk fat globule membrane extracts (MFGM) or phospholipids from buttermilk did not affect the accessibility of thiol groups of lactoglobulin. However, the presence of MFGM extract or phospholipids affected the formation of the aggregates. Figure 6.2 illustrates differences between pellets from the sedimentation of mixtures following the heat treatment. 106 Figure 6.2: Pictures of heated WPI (a), WPI+MFGM (b), and WPI+PL (c) at pH 4.6 at 80°C and after 60 minutes of sedimentation in ice-melted water bath. At pH 4.6, the height of pellets was smaller and the texture was harder in WPI comparatively to the other mixtures. In the mixtures WPI+MFGM extract and WPI+PL the texture of the pellet was softer and easy to disturb, while in WPI it was comparable to a gel. Increasing the temperature from 65°C to 80°C, only changed the height of pellets. No pellets were observed at pH = 6.8 for both temperatures meaning that the heat treatment was not strong enough at this pH to form non-soluble aggregates. Our previous results have demonstrated that the presence of buttermilk constituents during the denaturation of whey proteins controls the size and the number of particles (Chapter 5 and 6). In general, addition of buttermilk constituents decreased the number and the size of heat-induced aggregates. The present results seem to confirm that whey proteins form a strong network of aggregates while heating alone, and that the presence of buttermilk constituents changes the physical properties (such as size) of heat-induced aggregates. 107 6.5.2 Gel electrophoresis Evolution of native proteins as a function of heating time (80°C) is presented in Figure 6.3. Figure 6.3: Evolution of the protein profiles in the different mixtures at 0 (a), 10 (b), 15 (c), 20 (d) minutes of heating at 80°C using SDS-PAGE under non-reducing conditions. Letters a to d correspond to samples at pH 6.8 and letters a' to d' to pH 4.6. At pH 6.8, content of native -lactoglobulin and -lactalbumin decrease after 15 minutes of heating time (until the end) in WPI+MFGM and WPI+PL, but only after 10 minutes in WPI. We have already reported that the presence of buttermilk constituents delays the loss of native -lactoglobulin and -lactalbumin. It is known since Sawyer et al. (1963) that the main whey proteins interact which each other during a heat treatment. The aggregation process involves mainly thiol/disulfide interchange reactions, but also thiol/thiol oxidation, and hydrophobic interactions (Oldfield et al., 1998; Schokker et al., 1999). The decrease of the amount of whey proteins at each heating time is correlated to the process of aggregation. 108 Under non-reducing conditions, only three MFGM proteins were present in the mixtures before heating such as Butyrophiline (BTN: 67 kDa) and Cluster of Differenciation or Periodic Acid Shiff IV (CD36 or PAS IV: 76-78 kDa). All the other MFGM proteins were not detected by the migration. Ye et al. (2002) have reported that MFGM proteins are sensitive at lower temperatures comparatively to whey proteins. The pasteurization treatment of the whey or whey cream used for the present work could have been sufficient to denature most of the MFGM proteins. Their results also suggested that xanthine dehydrogenase/oxydase (XDH/XO: 150 kDa) and butyrophiline form a protein complex with a higher molecular weight during heating above 60°C. This complex could be the band located around 250 kDa in our gels. However, Butyrophyline was still present after 10 minutes of heating at pH 6.8, but only in a low amount, while the two other MFGM proteins were absent. Ye et al. (2004a,b) have demonstrated that MFGM protein can interact with -lactoglobulin and -lactalbumin during heating of milk, but only in a very low proportion. Our previous work has also shown that BTN and XO (from buttermilk) lost their native form previously to whey proteins (Chapter 6). No decrease in the amount of the native whey proteins or MFGM proteins was observed at pH 4.6 or during the heating at 65°C (for both pH) probably due to the poor accessibility of free thiol groups at this temperature. 6.5.3 Confocal laser-scanning microscopy images Even if the evolution of the protein profiles was similar (Figure 6.3), the microstructures of non-soluble aggregates were different. As shown in Figure 6.4, non-soluble aggregates were composed only of proteins (green) in WPI. Phospholipids (red) if present were not integrated in the structure of the protein aggregates. Previous results showed that some interaction protein/phospholipids or protein/MFGM proteins could be observed in heated whey protein mixtures, but were isolated and limited. It appears that whey proteins do not interact with MFGM proteins or phospholipids during the heating of whey at pH 4.6. At this pH, the repulsion forces between proteins are reduced due to the absence of surface charge. Whey proteins probably preferentially interact with each other. 109 Figure 6.4: Three-dimensional confocal laser-scanning microscope pictures of WPI heated for 25 minutes at 80°C at pH 4.6. However, this theoretical conclusion does not explain the trend observed in the WPI+MFGM extract mixture. As shown in Figure 6.5, MFGM proteins or phospholipids were part of the protein aggregate structures at pH 4.6, even only after 15 minutes of heating at 80°C. Some red dots were located on the side of aggregates (arrows I), but some were also buried inside the structure of the protein aggregates (arrows II). Pictures were taken under different angles in order to confirm that the red dots and green dots were connected. Appel et al. (1982) have shown that the denaturation of butyrophiline appear at 58°C. Later, Ye et al. (2002) demonstrated that BTN and XO can form aggregates at temperature above 60°C. Interactions between whey proteins and MFGM proteins are due to the exposition of free thiol groups of MFGM proteins during the initial state of the whey proteins denaturation, and involved mainly thiol/disulfide exchanges. 110 Figure 6.5: Three-dimensional confocal laser-scanning microscope pictures of the WPI+MFGM extract heated for 15 minutes at 80°C at pH 4.6. As shown in Figure 6.6, red dots (phospholipids) were integrated to the structure of protein aggregates (green) at pH 4.6 in the mixture WPI+PL even during a heating at 65°C. Phospholipids were present on the side or buried in the structure of the protein aggregates. Our previous idea was that phospholipids could act as an aggregation nucleus during the heat-induced denaturation of whey proteins (Chapter 6). Most of the protein aggregation mechanisms (included -lactoglobulin) are defined as nucleation-dependent and are in consequence initiated by the formation of an aggregation nucleus (Wang et al., 2010). In this case, a nucleus is a small molecule often considered as the size of a dimer that is present since the first stage of denaturation. Later in the aggregation process, non-native monomers, denatured proteins, and/or small aggregates are incorporated to the nucleus to form large aggregates. 111 Figure 6.6: Three-dimensional confocal laser-scanning microscope pictures of WPI+PL heated for 15 minutes at 65°C at pH 4.6 Wang et al. (2010) defined the last state of protein denaturation as the precipitation (loss of solubility) of aggregates. As shown in Figure 6.7, this state was not reached at pH 6.8. Only few isolated non-soluble particles were observed. As a general trend, these particles were composed of phospholipids covered with few proteins. 112 Figure 6.7: Three-dimensional confocal laser-scanning microscope pictures of WPI heated for 15 minutes at 80°C at pH 6.8. 6.5.4 Thin layer chromatography The evolution of the free phospholipids (not associated with the whey proteins) as a function of the heating time is presented in Figure 6.8. Results show that at pH 6.8 the free phospholipid profile is similar to the control at each heating time with a slight decrease in dot sizes at 25 minutes. At pH 4.6, any free phospholipids were found at 20 and 25 minutes in two of the three replicates (decrease in the third repetition only). A longer time than what was expected was necessary to bind a large amount of phospholipids to whey proteins at pH 4.6. 113 Figure 6.8: Evolution of the phospholipids profiles in WPI+PL at 0 (a), 10 (b), 15 (c), 20 (d), and 25 (e) minutes of heating time at 80°C. Picture was converted in negative mode in order to make the spot easier to see. As demonstrated by the confocal images, interactions between phospholipids and whey proteins during the early stages of the heating still occurred but involved only few phospholipids or weak interactions (easily breakable by the TLC solvents). Unfortunately, our TLC plates did not permit to identify the free phospholipids involved. The conclusions that -lactoglobulin can bind long chain fatty acids in aqueous solution have been reported since Spector et al. (1970). Cornell et al. (1989) have reported interactions between lactoglobulin and phospholipids at pH 4.4. Their results are consistent with the idea that positively charged groups of the proteins interact with negatively charged lipids. At pH 4.6, -lactoglobulin has a positive net charge, and phosphatidylcholine (PC) is isoelectric (other phospholipids are ionized at pH > 4.0) (Hauser et al., 1979). As described by Cornell et al. (1989), an electrostatic attraction between PC and -lactoglobulin could be excepted, and few positive charges (between 5 to 7 depending of the conformation of the protein) could welcome the negatively charge of PC. Earlier, Brown et al. (1983) have also reported 114 interactions between PC and -lactoglobulin involving both ionic and hydrophobic interactions after a treatment with a helix-forming solvent. Ionic interactions permit to position the lipid and protein molecules, and then hydrophobic interactions stabilize the complex. Later, Lefèbre et al. (1999; 2000) used Fourier transform infrared spectroscopy to better understand interactions between -lactoglobulin and phospholipids. Their results revealed hydrophobic interactions between the proteins and SM, and electrostatic interactions with PS. Their study also suggested that -lactoglobulin does not bind with PC. 6.6 Conclusions The present study shows that the heat-induced denaturation of WPI proteins is modified in presence of milk fat globule membrane extract or in presence of phospholipids. Protein profiles as a function of heating time was similar except a delay (5 minutes) for the loss of the native form of whey proteins and the accessibility of free thiol groups of whey proteins was not affect by the presence of both components. Three-dimensional pictures confirmed that only few interactions between whey proteins and MFGM proteins/phospholipids occurred during the heating of WPI. These interactions were however frequent after the addition of MFGM extract from buttermilk or phospholipids in the WPI. Red dots corresponding to phospholipids were buried in the protein aggregates structure at pH 4.6 confirming the potential aggregation nucleus role of phospholipids. However, the interactions between proteins and phospholipids seem to be weak during the early stages of heating. A longer time (> 15 min at pH 4.6 or > 20 min at pH 6.8) was necessary to strongly bind phospholipids to whey proteins at 80°C. TLC plates did not revealed a loss of phospholipids at 65°C for both pH while three-dimensional images showed interactions. It seems that phospholipids were incorporated to the structure of protein aggregates through MFGM fragments at lower temperature (65°C), while they can directly interact with the proteins at higher temperature (80°C). 115 6.7 Acknowledgements This work was funded by the Le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT, Quebec City, Quebec, Canada) – Novalait Inc. (Sainte-Foy, Quebec, Canada) – Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ, Quebec City, Quebec, Canada). 116 CHAPTER 7 EFFECT OF FREE THIOL GROUP REACTIVITY DURING THE FORMATION OF HEATINDUCED AGGREGATES FROM WHEY AND BUTTERMILK ON THEIR PROPERTIES The work presented in Chapter 4 showed that: 1) the presence of buttermilk changed the properties of the aggregates; 2) the proportion of buttermilk in the mixture controlled the properties of the aggregates. The work presented in this chapter aimed at understanding the impact of the free thiol reactivity on the properties of heat-induced whey:buttermilk protein aggregates. The preparation of the aggregates was conducted at Dairy Product Technology Center (DPTC) of California Polytechnic State University in San Luis Obispo, but the experiments were conducted at Laval University in Quebec City. 117 7.1 Résumé L’objectif de cette étude était de mieux comprendre l’effet de la concentration en groupements thiols libres lors de la formation des agrégats sur leurs propriétés. Des poudres de babeurre issues de crème ou de crème de lactosérum ont été mises en solution dans un concentré de protéines du lactosérum. Les mélanges ont été chauffés à 90°C à pH 4,6 en présence ou non d’un agent bloqueur de groupements thiols: N-ethylmaleimide. Les agrégats ont été lyophilisés. L’analyse statistique a révélé que la présence de groupements thiols libres lors du chauffage affectait les propriétés des agrégats. Les résultats ont montré que l’utilisation du NEM augmentait significativement le rendement d’agrégation et la capacité de rétention d’eau des agrégats et diminuait significativement la taille moyenne des agrégats. En solution, les agrégats exposent des groupements thiols libres même s’ils ont été formés en présence de NEM. 118 7.2 Abstract The objective of this study was to better understand the impact of the concentration of free thiol groups during formation of the mixed aggregates on their properties. Spray dried buttermilk powders from regular and whey creams (treated with CO2-SFE or not) were hydrated in liquid whey protein concentrate and then heated up to 90°C at pH 4.6. The dispersions were heated with or without the presence of thiol blocking N-ethylmaleimide (5 mM). The aggregated material was separated after heating by centrifugation at 15,000 g for 20 minutes and then freeze-dried. Statistical analysis of the data revealed that the presence of NEM significantly affected the properties of the aggregates. The composition of the mixture significantly affected the solubility of the powders only. Results showed that the reduction of the thiol group reactivity significantly increased the aggregation yield, the water-holding capacity of aggregates before drying, but significantly reduced the numberweighted average particle size. Aggregates exposed free thiol groups when they were in solution even if they have been formed in the presence of NEM. Overall, results allow the connection between our knowledge of the formation of the mixed aggregates with their basic physical and functional properties. 7.3 Introduction The thermal behavior of the main dairy proteins have been extensively studied since the first report of Sawyer et al. (1963). Different mechanisms of aggregation have been proposed for whey proteins (Gezimati et al., 1996; Havea et al., 2000; Roefs et al., 1994; Schokker et al., 1999), whey proteins with casein micelles (Livney et al., 2004; Vasbinder et al., 2003a), milk fat globule membrane (MFGM) proteins (Appel et al., 1982; Ye et al., 2002), and MFGM proteins with whey proteins and caseins (Ye et al., 2004a; Ye et al., 2004b). Heat-induced protein aggregation has been used as a basis for the development of industrial processes to increase the retention of whey proteins in cheese matrices, or to prepare ingredients for food formulations (Lawrence, 1993.). 119 The first approach consisted of increasing the temperature of pasteurization (above 72°C – 15s) of the cheese milk in order to promote the heat-induced interactions of whey proteins with the casein micelles (Lawrence et al., 1993). This method has been extensively used for the production of yogurt (85 to 95°C – several minutes), but is not optimal for the production of cheese because it affects the enzymatic reactions and the fusion of the casein micelles. Guyomarc’h et al. (2006) also reported excessive moisture and texture defects of the cheeses. This method is, however, still under study, but limited due to the poor literature regarding the properties of the whey protein/casein complexes and the absence of industrial process for the production and isolation of the complexes on a large scale (Donato et al., 2009). The second approach consisted in pre-aggregating whey proteins (90 to 95°C – several minutes) at low pH (4.2 to 4.6) or neutral pH (6.2 to 6.7) directly in liquid whey. Later, the pre-aggregates whey proteins are incorporating in cheese milk (Lebeuf et al., 1998; Punidadas et al., 1999). These ingredients have been widely used as fat mimetics for the production of low fat cheese (Lawrence, 1989, 1993; Lebeuf et al., 1998; Punidadas et al., 1999) or as texture modifiers for yogurt (Sandoval-Castilla et al., 2004; Tamine et al., 1995). The use of these ingredients is, however, limited to 0.5 to 1.5% of the total weight due to the high water-binding capacity of denatured proteins. In both approaches, thiol/disulfide exchange reactions (SH/SS) are believed to direct the formation of protein aggregates. Gilbert (1993) defined these reactions by the nucleophilic attack of a disulfide bond (SS) of one protein by a free thiol group (SH) of another protein. The two proteins associate through thiol-thiol oxidation and a new thiol group is released. In heated milk, the protein aggregation is led by the exposition of free thiol groups of the milk fat globule membrane (MFGM) proteins around 65°C (Ye et al., 2004a,b; Corredig and Dalgleish, 1998), but mostly by the exposition of the free thiol group of β-lactoglobulin at 78.5°C (Schokker at al., 1999). These groups will initiate interactions with whey proteins (Schokker et al., 1999; Havea et al., 1998), caseins or casein micelles (Anema et al., 2003a; Donato and Guyomarc’h, 2009), and MFGM proteins (Corredig and Dagleish, 1998; Ye et al., 2004a,b). Other types of interaction such as thiol-thiol oxidation, hydrogen bonds, ionic 120 bonds, van der Waals forces, and hydrophobic interactions are less represented but seem to play an important role for the dairy protein aggregation (Oldfield et al., 2000; Schokker et al., 1999; Prabakaran et al., 1997). It has been shown that it was possible to integrate buttermilk in the process of the production of heat-induced whey protein aggregates at a low pH (Chapter 4). Our previous results demonstrated that the mixed protein aggregates were composed of whey proteins, caseins, MFGM proteins and phospholipids and were stabilized by both covalent and noncovalent interactions (Chapter 5 and 6). The presence of buttermilk concentrate also changed the physical and functional properties of aggregates such as a reduction of their water-holding capacity as a function of the level of substitution of whey concentrate by buttermilk concentrate (Chapter 4). In line with these observations, we hypothesized that the free thiol group reactivity of buttermilk constituents has a direct impact on the physical and functional properties of the mixed aggregates. The objective of the present work was to characterize the contribution of the thiol group reactivity during heating on some of the basic physical and functional properties of the mixed aggregates. 7.4 Materials and methods 7.4.1 Materials Fresh whey (unpasteurized; milk pasteurized at 74.4°C for 16.1 seconds) from TILL and Gouda cheese production and milk cream (unpasteurized; milk pasteurized at 74.4°C for 16.1 seconds) were obtained in California Polytechnic State University’s processing plant (San Luis Obispo, CA, USA). Whey cream (pasteurized at 82.2°C for 35 seconds; milk pasteurized at 73.3°C for 16 seconds) from cheddar cheese production was donated by 121 Hilmar Ingredients (Hilmar, CA, USA). Elman’s reagent was from Thermo Scientific (Rockford, IL, USA). N-ethylmaleimide and all other reagents were from Fisher Scientific (Fair Lawn, NJ, USA). 7.4.2 Preparation of the whey and buttermilk concentrates Fresh cheese whey was skimmed by using a pilot scale milk separator (Alfa Laval, Uppsala, Sweden). Bacterial contamination of whey was reduced by microfiltration using a pilot plant scale system (T-12 model, GEA-Niro Filtration, Hudson, WI, USA) through a 1.4 m membrane (Koch Membrane Systems, Wilmington, MA, USA). Microfiltered whey was concentrated using two 10 kDa spiral polymeric ultrafiltration membranes (Koch Membrane Systems, Wilmington, MA, USA) fitted in parallel up to a protein concentration of 4.0 % (w/v). Whey permeate was collected and frozen until used for protein blend formulations. The retentate was diluted sequentially in the feed tank with 3 diavolumes of DI water. The diafiltration was stopped when a final concentration of 5.2% of protein (w/v) was obtained. The whey protein concentrate (WPC) was then frozen to – 35°C until further analysis. The concentration trial run was performed with whey from three different productions. Regular and whey cream were processed at 12°C using a continuous butter-churn for pilot plant (Egli, Butschwil, Switzerland). Immediately after churning, regular and whey buttermilks were filtrated through cheesecloth, and skimmed, using a pilot scale milk separator (Alfa Laval, Uppsala, Sweden). Skim buttermilks were concentrated by UF (GEA R-12 model, Milwaukee WI) using two 10 kDa spiral polymeric membranes fitted parallel on the module and then the retentate was diluted sequentially in the feed tank with 3 diavolumes of DI water. The DF was conducted until a concentration of 5% of protein (w/v) was reached. Buttermilk concentrates (BC) were spray-dried using a Niro Filterlab Spray-drier (Hudson, WI, USA) at 35 bars, and with inlet and outlet air temperatures of 185°C and 95°C, respectively. A portion of each powder was submitted to supercritical fluid extraction (SFE) (Thar Designs, Inc., Pittsburgh, PA, USA) to remove the non polar lipids according to Costa et al. (2010). Circulated deionized water (+ coolant) at 3°C was 122 used for cooling different zones in the SFE apparatus. Carbon dioxide tanks were filled and inspected by A & R Welding Supply (San Luis Obispo, CA, USA). The system conditions were controlled manually by Windows 2000-based software. Approximately 125 g of each sample were submitted to an extraction cycle for maximum non-polar lipid removal using the following conditions: 7200 g of CO2 at a flow of 20 g.min-1, extraction pressure of 350 bar, and temperatures of 35°C. Powders were stored at 10°C under the following codes: regular buttermilk concentrate (rBC), whey buttermilk concentrate (wBC), regular buttermilk concentrate treated by SFE (rBCSFE), and whey buttermilk concentrate treated by SFE (wBCSFE). 7.4.3 Preparation of protein aggregates powders WPC was diluted with UF permeate to decrease the protein concentration from 5.2% to 2.6% (w/v). BC powders were then dispersed in diluted WPC to increase the total protein concentration to 5.2% (w/v). Whey-buttermilk mixtures were stirred overnight at 4°C. Denaturation process followed the previous proposed design (Chapter 4). Briefly, 250 mL blends were adjusted to pH 4.6 with 1N HCL and heated from 8°C to 90°C in a thermostatically controlled water-bath maintained at 97°C for a total time of 25 minutes. Mixtures were constantly stirred using a caframo lab mixer during heating (speed 5). Nondiluted WPC (5.2% protein) was used as a control. The aggregated material of each mixture was separated at the end of heating (25 minutes) by centrifugation (Beckman L7-35, Brea, CA, USA ) at 15,000g for 20 minutes at 20°C, and then freeze-dried. Heating experiments have also been conducted in presence of 5.0 mM of N-etylmaleimide (NEM) in order to completely block free and exposed thiol groups during heating (Alting et al., 2000). Nitrogen content was determined using Kjeldahl method and converted to protein using a conversion factor of 6.38 (AOAC, 2003). Aggregation yield was calculated before freezedrying from the protein content in the mixtures and in the supernatants after centrifugation as follows: 123 Equation 7.1: Calculation of the aggregation yield Aggregatio n yield (%) = PM PS 100 PM where [P]M was the protein concentration in mixtures (%), and [P]S was the protein concentration in supernatant (%). Water-holding capacity (WHC) of aggregates was calculated before freeze-drying from the moisture content of the centrifugation pellets, the moisture content of the powders, and the protein content of the powders as follow: Equation 7.2: Calculation of the water-holding capacity of the aggregates WHC(gwater /g protein ) MP PP where [M]P was the moisture content in the centrifugation pellet (g), and [P]P was the protein content in the centrifugation pellet (g). 7.4.4 Particle size distribution in heated liquid mixtures The particle size distribution of mixtures (5.2% proteins (w/v)) after heating was measured using a laser diffraction particle size analyzer in the polarization intensity differential scattering optical mode (Beckman Coulder, Brea, CA, USA). Samples were taken after 25 minutes of heating in the thermostatically controlled water bath (set up at 97°C), shaken by inversion for 1 minute, and diluted in the analysis chamber with one liter of DI water (automatically filled-up by the instrument). The instrument was allowed to adjust for electrical offsets and align of the laser prior to the measuring background. Measurements were started after 5 minutes of agitation. Data was obtained from the average of three measurements per sample and expressed as the number-weighted average particle size (D[4,3]n). 124 7.4.5 Protein solubility index at pH 6.8 Protein aggregates powders were rehydrated in SMUF solution (pH = 6.8) (Jenness and Koops, 1962) to obtain 1% (w/v) protein solutions. The pH was adjusted to 6.8 using NaOH 1N, and solutions were stirred for an hour before to be adjusted to pH 6.8 again using NaOH 1N or HCl 1N. Solutions were centrifuged after 2 hours of stirring (total time) or 26 hours of stirring (total time) at ~ 3,000g for 10 min at room temperature. The protein solubility index (PSI) was calculated from the protein concentration in solutions ([P] i) and protein content in supernatants ([P]sup) determined by BCA microplate method as follow: Equation 7.3: Calculation of the protein solubility index at pH = 6.8 PSI(%) Psup 100 Pi 7.4.6 Exposition of free thiol groups of the mixed aggregates Accessible thiol group concentration of the soluble aggregates (from the supernatants) was measured according to Ellman (1959). 250 µL of the solutions of soluble aggregates in SMUF at pH 6.8 were taken and mixed with 2.5 mL of 0.1M sodium phosphate buffer (containing 1mM EDTA; pH = 8.0), and 0.05 mL of DTNB reagent solution (4 mg in 1 mL of sodium phosphate buffer). The tubes were shaken by inversion for 1 minute and incubated at room temperature for 30 min. Absorbance was measured at 412 nm using a Multiskan Spectrum (Thermo labsystems, Romeoville, IL, USA). The absorption values of the standards (concentration of cysteine ranging from 1.5 mM to 0 mM) were plots and the equation was extracted from the figure. The free thiol concentration of solution was expressed as mole per gram of protein. 7.4.5 Statistical analysis All experiments were performed in triplicate, so each value represents the mean of three measurements. Two different and independent variables were selected and named as follows: Composition that expressed the effect of addition of different buttermilk 125 concentrates, NEM that expressed the effect of blocking the free thiol groups during the preparation of the powders. Data was tested according to a factorial design 5*2 except for the nitrogen solubility index (5*2*2). In this case, a third variable named Hydration was added in order to express the effect of a longer time of hydration of the powders (2 h vs 26h). Statistical analysis was carried out using SAS software (SAS Institute Inc., Cary, NC, USA). Mean comparisons were performed using a Duncan post-test if appropriate. Table 7.1 summarizes significance for each treatment on each variable. Results were considered significantly different when P < 0.05. Table 7.1: Summary of the significance (P value) calculated by the analysis of variance of data of each contrasts for each variables (n = 3). Significance (P values) Contrast Aggregation yield* WHC* D[4,3]n* PSI [-SH]free Model 0.0232 0.0042 0.0011 0.0002 0.0882 Composition (A) 0.1838 0.3056 0.5053 0.0074 0.8067 NEM (B) 0.0470 0.0105 0.0075 < 0.0001 0.3528 Hydration (C) n/a n/a 0.0316 n/a A*B 0.1565 0.7025 0.6818 0.0354 0.4442 A*C n/a n/a n/a 0.3908 n/a B*C n/a n/a n/a 0.3388 n/a A*B*C n/a n/a n/a 0.9263 n/a n/a * The effect “blocs” was evaluated significant, so the software automatically re-adjusted the model. This result means that three whey were different. The first whey was a mix from TILL and Gouda cheeses production, while the second and third whey were from TILL cheese production only. 7.5 Results 7.5.1 Heat-induced aggregation characterization The presence of NEM had a significant effect (P = 0.0075) on number-weighted average particle size distribution (D[4,3]n) of aggregates before drying (Table 7.1). The D[4,3]n was 126 1.50 0.56 µm in the absence of NEM and 0.23 0.04 µm in the presence of NEM (Figure 7.3). Figure 7.1: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the number-weighted average particle size distribution in the mixtures. The presence of N-ethylmaleimide (NEM) had a significant effect (P = 0.0470) on the aggregation yield (Table 7.1). The aggregation yield was 74.03 1.38% in the absence of NEM and 78.36 1.91% in the presence of NEM (Figure 7.2). 127 Figure 7.2: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the aggregation yield. 7.5.2 Properties of the mixed aggregates The presence of NEM had a significant effect (P = 0.0105) on the water-holding capacity (WHC) of aggregates (Table 7.1). The WHC was 2.97 0.72 gwater.gprot-1 in the absence of NEM and 3.59 0.70 gwater.gprot-1 in the presence of NEM (Figure 7.3). Figure 7.3: Effect of the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the water-holding capacity of the aggregates. 128 The interaction composition*NEM had a significant effect (P = 0.0354) on the protein solubility index (NSI) (Table 7.1). As shown in Figure 7.4, the presence of NEM did not affect the solubility of the aggregates powders from the heating of WPC (14.25 1.14% in the presence of NEM vs 15.44 3.32% in the absence) and the mixtures WPC + rBC (14.15 0.89% vs 14.97 1.49%) and WPC + wBM-SFE (11.60 2.50% vs 13.18 1.34%). Oppositely, the presence of NEM affected the solubility of the aggregate powders from the mixtures WPC + BM-SFE (6.52 0.46% vs 18.40 1.26%) and WPC + wBM (6.32 0.46% vs 13.30 1.22%). Protein Solubility Indice (%) 25 20 15 10 5 0 C WP rBC w.+ _S BC w.+ FE wB w.+ C w w.+ E _SF C B Figure 7.4: Effect of the composition and the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on nitrogen solubility index of the powders. Aggregates formed in the absence of NEM are represented in black and aggregates formed in the presence of NEM are represented in grey. Table 7.1 shows that none of the contrasts or interactions between contrasts had a significant effect on the accessibility of thiol groups. The concentration of thiol groups in the soluble fraction was ranged from 1.1 ± 0.2 µmole/gprot to 1.7 ± 0.4 µmole/gprot (Figure 7.5). 129 2.5 [-SH] free (µmole.gprot-1) 2.0 1.5 1.0 0.5 0.0 WP C E E E rBC _SF _SF _S F C C C w.+ B B B w w w.+ w.+ w.+ Figure 7.5: Effect of the composition and the presence of NEM during formation of mixed aggregates from heating (90°C – 5min) at low pH (4.6) mixtures containing 50% of the protein from WPC and 50% from buttermilk concentrates on the accessibility of free thiol groups of the mixed aggregates. 7.6 Discussion Statistical analysis revealed that the whey samples taken from three different productions were significantly different. It has been reported for a long time that whey protein concentrate (WPC) are inconsistent in their composition and functionality due to differences in the composition of the milk, differences in the cheese-making process, and differences in the manufacture of the WPC (Morr et al., 1993; Schmidt et al., 1984). Later, and for the first time de la Fuente et al. (2002) investigated the structural changes in whey proteins during the different steps of the manufacture of WPC. Their results suggested that the differences in protein functionalities were to be attributed to the WPC production process rather than to the cheese manufacture. Results also showed the presence of protein aggregates likely formed during the pasteurization of milk and whey. Earlier, Schmidt et al. (1984) suggested that factors affecting the cheese yield such as heat treatment, starter 130 culture, cooking conditions would be expected to alter the composition of the whey and WPC. The same factors would also alter the properties of the whey proteins. Some variations in these factors have been observed during the sampling of the whey. N-ethylmaleimide (NEM) has been used for the understanding of the heat-induced aggregation kinetics or mechanisms of whey proteins and especially β-lactoglobulin (Hoffman and Van Mill, 1997; Oldfield et al., 1998, 2000; Sawyer, 1968; Xiong et al., 1993). By its binding with the exposed free thiol groups of the protein, NEM prevents the SH/SS exchange reactions between proteins. In consequence, aggregation occurs through other interactions such as hydrogen bonds, ionic bonds, van der Waals forces, and hydrophobic interactions. However, it is not evident to find clues on the properties of the non-covalent based protein dairy aggregates because they are rarely found in normal heating conditions (without a thiol blocking agent). Oldfield et al. (2000) reported that “hydrophobic formed” aggregates of β-lactoglobulin were found at the beginning (under 15% of the total protein) but close to none at the end of heating at 85°C. First, results showed that the solubility of aggregates formed in presence of NEM was lower comparatively to aggregates formed in normal conditions except for aggregates formed from buttermilks treated by CO2-SFE. The solubility has been measured on nonhomogenized aggregates. However, a lower solubility of the powders of aggregates formed in the presence of NEM could be correlated to the higher level of denaturation. Results showed that preventing the SH/SS exchange reactions significantly increased the proportion aggregates (precipitable or sedimentable). Havea et al. (2004) observed similar results by using DTT (dithiothreitol). Their results showed that the prevention of the SH/SS exchange reactions offer a higher number of reactive sites and in consequence a faster aggregation process. They explained that the lack of restriction due to movements of disulfide bonds deeply increased the chance of proteins to associate upon collision. Later, they found that gels formed through non-covalent interactions were strong and rigid comparatively to gels formed in normal conditions (weak gels). Controversially, Alting et 131 al. (2000) reported that preventing the SH/SS exchange reactions significantly decreased the hardness of cold protein gels. The weakness of gels is often attributed to the extensive capacity of denatured proteins to hold water. Results demonstrated that aggregates were less hydrated in normal conditions than in presence of NEM. This result could be explained by the average particle size distribution. Results showed that the prevention of the SH/SS exchange reactions significantly reduced the average diameter of particles. It is easy to think that large aggregates hold less water because the hydrophilic groups of the individual proteins become buried in the more complex structure. In normal conditions, aggregates become larger due to addition of denatured proteins to the complex through SH/SS exchange reactions. By preventing the formation of the SH/SS exchange reactions, the polymerization sites are reduced so the formation of large aggregates is impossible. Results from Alting et al. (2000) also showed that large aggregates were present after gelation only in normal conditions. Surprisingly, results showed that accessible thiol groups were found after hydration of the powders of mixed aggregates. Alting et al. (2000) explained that the heat-induced aggregation mechanism that Roefs and De Kruif (1994) proposed for the whey proteins anticipates the presence of free thiol groups at the surface of the protein aggregates. This result gives a strong clue for answering the question on the reactivity of protein aggregates in a new system and the potential association with other proteins during a heat treatment. It is, however, difficult to explain the presence of accessible thiol groups at the surface of aggregates formed in the presence of NEM. 7.7 Conclusion The present study showed that the reactivity of free thiol groups during the formation of protein aggregates from whey and buttermilk has a direct impact on the properties of the aggregates. Thiol/disulfide exchange reactions (SH/SS) are suspected to direct the formation of the protein aggregates, but results presented showed that they also control the size of the protein aggregates, their water-holding capacity, and their solubility. Results 132 suggest that the prevention of the SH/SS exchange reactions can be interesting for the control of the properties of protein aggregates and offers new potential applications. Previous results demonstrated that addition of buttermilk constituents or decreasing the pH before heating (4.6 instead of 6.8) reduced the free thiol group reactivity. However, their total prevention requires the use of a strong chemical that is not acceptable for food formulation. 7.8 Acknowledgments This work was funded by the Le Fond Québécois de la Recherche sur la Nature et les Technologies (FQRNT, Quebec City, Quebec, Canada) – Novalait Inc. (Sainte-Foy, Quebec, Canada) – Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ, Quebec City, Quebec, Canada). 133 CHAPTER 8 EFFECT OF SUBSTITUTION OF SKIM MILK POWDER BY WHEY:BUTTERMILK HEATDENATURED PROTEIN AGGREGATES IN MODEL SET-TYPE YOGURT The work presented in the previous chapter showed that: 1) the thiol group reactivity during formation of the mixed aggregates has a direct impact on the properties of the aggregates; 2) thiol groups are accessible at the surface of the mixed aggregates. Results suggest that the mixed aggregates can act as active fillers in a process involving heating due to their accessible free thiol groups. The work presented in this chapter aimed at understanding the behavior of whey:buttermilk heat-denatured protein aggregates during the yogurt-making process. The effect of different level on incorporation of aggregates on texture of yogurt was tested. Water-holding capacity of gels, particle size distribution, liberation of thiol groups during heating, and repartition of clusters were investigated to better understand the texture results. Results of this part of the project will be submitted essentially in this form to the Special Issue of Foods «Structure and Flavour of Dairy Products». Results of this paper were also presented in poster form at 2010 STELA Colloque in Quebec City (QC, Canada) and at 2010 ADSA Join Annual Meeting in New Orleans (LA, USA). 135 8.1 Résumé L’objectif de cette étude était d’évaluer l’impact de l’addition des agrégats formés à partir d’un mélange protéique lactosérum (25 %) - babeurre (75 %) sur la production de yaourt ferme. Les yaourts ont été préparés à partir de laits écrémés standardisés à 15 % (p/v) de solides et 4,2 % (p/v) de protéines par l’ajout de différents taux de poudre de lait écrémé et de poudre d’agrégats. L’ajout d’agrégats protéiques a affecté significativement la texture des yaourts mais n’a pas eu d’impact sur la capacité de rétention d’eau des gels. Les images de microscopie confocale ont montré que la présence de particules de tailles importantes dans les laits enrichis avec les agrégats affectait directement l’homogénéité de la matrice protéique en formation. L’analyse des groupements thiols libres des agrégats en solution dans de l’eau suggèrent que les agrégats pourraient interagir avec les protéines du lait écrémé lors du chauffage. 136 8.2 Abstract The objective of this study was to assess the impact of adding heat-induced wheybuttermilk protein aggregates to acid-set type yogurt production. Whey and buttermilk protein mixture (25:75) was adjusted to pH 4.6 and maintained at 90°C for 5 min. Set-type yogurts were prepared from skim milks standardized to 15% (w/v) total solids and 4.2% (w/v) proteins using different substitution levels of skim milk powder and aggregates powders. Addition of protein aggregates significantly modified texture properties of yogurts but did not affect water-holding capacity of gels. Confocal laser-scanning microscope pictures showed the presence of large particles in milk fortified with protein aggregates that directly affected homogeneity of the clusters of the protein matrix. Free thiol groups were accessible during heating of the protein aggregates powder in a water solution suggesting that the aggregates could interact with the milk protein during the heat treatment of milk. 8.3 Introduction The use of microparticulated whey proteins (MWP) as a fat replacer in yogurt systems has been investigated. This dry dairy product was developed initially in order to provide means of increasing the retention of whey proteins by casein during cheese curd formation (Lawrence, 1989; Lawrence, 1993; Lelièvre, 1990). It is obtained by heating ultrafiltered whey at 90-95°C for several minutes with the pH at acidic (4.2–4.6) or neutral (6.2–6.7) values (Lawrence, 1993; Lebeuf et al., 1998; Punidadas et al., 1999) and spray drying. It is generally accepted that the product does not interact with milk proteins during the cheesemaking process and that the micro-particles are only trapped within the casein network. the cooking step. It is still unclear whether MWP acts as active or inert particles in yogurt. Tamine et al. (1995) described the role of a fat replacer as simulating fat globules without interacting with other milk proteins. Sandoval-Castilla et al. (2004) and Tamine et al. (1995) concluded that MWP not only becomes part of the protein matrix, but also modifies the microstructure of the gel The gels thus formed are lower in tension and in firmness, but 137 higher in cohesiveness (Barrantes et al., 1994; Sandoval-Castilla et al., 2004; Tamine et al., 1995). The resulting yogurt is also more susceptible to syneresis. The major drawback associated with MWP is the high water-holding capacity of the microparticles, which contribute directly to increasing the moisture content of the curd. Saffon et al. (2011) proposed reducing the water-holding capacity of MWP by combining the whey starting material with buttermilk. They showed that this significantly decreased the waterholding capacity of the resulting micro-particles, as well as modifying their properties. Once viewed as a by-product of butter-making, buttermilk is now considered as a valuable product because of its high content in fragments of milk fat globule membrane (MFGM), in addition to phospholipids and whey proteins (Jiménez-Flores & Brisson, 2008; Singh, 2006). However, studies by Mistry et al. (1996), Raval et al. (1999), and Turcot et al. (2001, 2002) indicated that the moisture content of cheese supplemented with buttermilk remained high, due largely to phospholipids. In the present work, less-hydrated aggregates of heat-denatured whey:buttermilk protein were added as a substitute for skim milk powder to skim milk used for yogurt production. The main objective of this study was to understand the behavior of these aggregates during the yogurt-making process. The effect of different levels of skim milk powder substitution (0 to 100%) on the textural properties of acid set-type yogurt was tested. The water-holding capacity of the resulting gels, the particle size distribution after heating, the liberation of thiol groups during heating, and the average size of the aggregates were investigated in order to gain better understanding of the texture results. 8.4 Materials and methods 8.4.1 Materials Fresh whey from mozzarella cheese production and fresh buttermilk were both obtained from a local cheese producer (L’Ancêtre, Bécancour, QC, Canada). Powdered whey 138 permeate was obtained from Maple Leaf Foods International (Toronto, ON, Canada) and powdered skim milk was obtained from Dairytown Products Limited (Sussex, NB, Canada). Pasteurized skim milk was purchased at a local grocery outlet and yogurt starter was obtained from Yogotherm (Saint-Hyacinthe, QC, Canada). Oregon Green 488 was obtained from Life Technologies Corporation (Burlington, ON, Canada). All other reagents were obtained from Fisher Scientific (Ottawa, ON, Canada). 8.4.2 Preparation of whey:buttermilk heat-denatured aggregates Heat-denatured aggregates of whey:buttermilk protein were formed using our method as described previously (Chapter 4). Briefly, cheese whey and buttermilk were skimmed using a pilot-scale milk separator (Alfa-Laval, Uppsala, Sweden) and then concentrated separately by ultrafiltration (UF) through a 5 kDa membrane (Romicon, Koch Membrane Systems, Wilmington, MA, USA) to a final protein concentration of 9.5% (w/v). Concentrates were mixed at a whey:buttermilk protein ratio of 25:75. Mixtures were adjusted to pH 4.6 by slow addition of HCl 6N and heated from 4°C to 90°C for 16 min (including ramp time, about 11 min at 90°C) in a stirred cooker with a steam jacket. The heated liquid was then homogenized three times at 69 MPa using an Emulsiflex C-500 device (Avestin Canada, ON, Canada). The homogenates were freeze-dried in order to obtain powdered whey:buttermilk aggregate (WBAP). 8.4.3 Yogurt production Skim milk containing 8.4% (w/v) total solids and 3.6% (w/v) protein was standardized to 15% total solids and 4.2% protein using powdered skim milk and powdered permeate. WBAP was added to replace 0, 20, 40, 60, 80 or 100% of the added powdered skim milk. The milk thus enriched was stirred for 30 minutes at 4°C and then heated to 85°C in a Stephan cooker with thermostatic control (Sympak France, Lognes, France), with constant stirring for 25 minutes including ramp time. After cooling to 42°C in an ice water bath, the milk was inoculated with a commercial yogurt starter (Streptococcus thermophilus and Lactobacillus bulgaricus) and incubated at 42°C until a pH of 4.6 was obtained. For each enrichment level, sixteen yogurts were analyzed after 18 hours at 4°C, eight were analyzed 139 after 7 days, and eight after 14 days. Yogurts were also prepared in falcon tubes (25 mL). Each production was repeated three times. 8.4.4 Analytical methods 8.4.4.1 Composition analysis Overall composition of powdered ingredients was determined using standard methods. Total nitrogen content was determined by the Dumas combustion method (IDF, 2002) using a LECO device (Protein Analyzer method FP-528 Leco Instruments Ltd., Mississauga, ON, Canada). Nitrogen values were converted into protein values using 6.38 as the nitrogen conversion factor. Fat was determined using the Mojonnier extraction method (IDF, 2008) and lactose by phenol – sulfuric acid method. Total solids were obtained by microwave drying (Smart System 5, CEM Corp., Matthews, NC, USA) and ash was measured by incineration in a refractory oven at 550°C for 20 hours (AOAC, 1990). 8.4.4.2 Texture properties The texture of the set-type yogurts was evaluated by penetration using a TA-XT2 Texture Analyzer (Texture Technologies Corp., Scarsdale, NY, USA) connected to a computer running the Stable Micro System (Stable Micro Systems Limited, Surrey, United kingdom), with a cylindrical probe (diameter 12.5 mm) and 5 kg load cell. Yogurts were analyzed in their containers (5.6 cm diameter with yogurt depth of 3.5 cm). The following parameters were set: speed 60 mm/min, distance 20 mm, hold-up time 30 seconds, force (threshold) 0.01 Newton (N). Rupture force corresponded to the first peak, firmness to the maximum force, and adhesiveness to the resistance to withdrawal of the penetration probe from the sample. Relaxation was calculated from the firmness force and force after 30 seconds of holding. Tests were carried out immediately after moving the samples from the cold room. 140 Equation 8.1: Calculation of the relaxation of acid set-type yogurt Relaxation (%) = Firmness (N) - Force at 30 s (N) 100 Firmness (N) 8.4.4.3 Water-holding capacity of yogurt Yogurt prepared in Falcon tubes (25 mL) was centrifuged at 222 x g for 10 minutes at 4°C. The clear supernatant was poured off and weighed. Water-holding capacity was calculated from the weight of the supernatant and that of the yogurt: Equation 8.2: Calculation of the water-holding capacity of the gel WHC (%) = (1 - Supernatant Weight (g) ) 100 Yogurt Wei ght (g) 8.4.4.4 Particle size distribution of milk Skim milk was enriched with WBAP in the same proportions as for yogurt production and stirred for 30 minutes at 4°C before analysis. Particle size distribution before and after heating was measured using a Mastersizer 2000 laser diffraction system (Malvern Instruments, Worcestershire, UK). Samples were shaken by inversion for 1 minute and dispersed directly into the recirculating cell until a 12–19% obscuration value was reached. The dispersant (deionized water) was stirred at 950 rpm and measurements were started after 5 minutes. Data were expressed as the volume-weight average particle size, also known as the D[4,3] value. Five measurements per sample were averaged. 8.4.4.5 Exposition of free thiol groups upon heating Exposure of free thiol groups was measured according to Ellman (1959). A 250 L sample of WBAP-enriched skim milk was taken every five minutes during heating and mixed with 2.5 mL of 0.1M sodium phosphate buffer (containing 1 mM EDTA; pH 8.0) and 0.1 mL of DTNB reagent solution (4 mg in 1 mL of sodium phosphate buffer). Ammonium persulfate (0.2000 0.0050 g) was added after 30 minutes of incubation at room temperature. 141 Absorbance was measured at 412 nm using a Multiskan Spectrum (Thermo Labsystems, Vantaa, Finland) relative to a blank. The theoretical exposure was calculated as follows: Equation 8.3: Calculation of the theoretical concentration of thiol group SH theo mole/g prot SM -SH PSM WBAPSH PWBAP PP Where [-SH]theo is the concentration of exposed free thiol groups in moles per gram of protein in the sample containing 100% WBAP substitution of the skim milk powder supplement, SM-SH is moles of thiol group contributed by the skim milk, PSM is the protein content in skim milk (3.6 g), WBAP-SH is moles of thiol group contributed by WBAP, PWBAP is the protein content of powdered WBAP added (0.6 g) and PP is the protein content of the final mixture (4.2 g). 8.4.5 Confocal laser scanning microscopy The imaging procedure was adapted from the procedure developed by Hennessy (2011) and was carried out using a Nikon C1 laser scanning confocal microscope 60X, water objective (Nikon Corporation). Briefly, skim milk with 0 or 100% of the enrichment replaced with WBAP was placed for 11 minutes in boiling water with constant stirring. After cooling to room temperature, 10 mL aliquots were treated with 1.000 0.050 g of glucano-deltalactone and then immediately with 3 L of Oregon Green 488 (0.1 g.L-1 in dimethylformamide) to stain proteins and with 3 L of Nile red (0.01 g.L-1 in dimethyl sulfoxide) to stain phospholipids. A 75 L droplet of the mixture was deposited on a cover slide placed on top of the laser, and images were captured every 15 seconds for 20 minutes beginning at 2 min after adding glucono-delta-lactone. Files were transformed into sixframe-per-second videos using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/). The videos are available at the following internet link: https://www.dropbox.com/sh/c9gmyrv56kznu45/R1GICKO4if. Single pictures were taken at the end of the gel formation using the following parameters: resolution = 1024/1024; quality = 13 db; zoom = 2.31X. The dark level was set at 0 in order to limit background 142 noise and to define the clusters more clearly. Images were loaded into ImageJ, transformed to 8-bit, and their threshold was adjusted to cover all the dark areas. 8.4.6 Statistical analysis All experiments were performed in triplicate, each value thus representing the mean of three measurements. All data were tested as a factorial experiment with powdered skim milk substitution level*time. Time corresponded to storage time at 4°C for texture and WHC analysis, and to heating time for particle size distribution and thiol group exposure in the enriched milk. Statistical analysis was carried out using the SAS PROC GLM procedure (SAS Institute Inc., Cary, NC, USA). Mean comparisons were performed using a Duncan post-test. Tables 1a and 1b summarize the significance of each treatment for each variable. Results were considered significantly different when P < 0.05. Table 8.1: Summary of the significance of the effects of WBAP and time on the texture and waterholding capacity of yogurt gels, as calculated by analysis of variance (n = 3). Significance (P values) Contrast Firmness Rupture Force Adhesiveness Relaxation WHC Model < 0.0001 0.0004 < 0.0001 < 0.0001 0.5114 Substitution Level (A) < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.4348 Time (B) 0.9781 0.2021 0.0604 0.7538 0.3903 A*B 0.4579 0.7150 0.2156 0.1394 0.5124 143 Table 8.2: Summary of the significance of the effects of WBAP and time on the response of milk to heating, as calculated by analysis of variance (n = 3). Significance (P values) Contrasts Particle size [-SH]free Model 0.0044 < 0.0001 Level of substitution (A) 0.0002 < 0.0001 Time (B) 0.8507 0.0174 A*B 0.6332 0.1394 8.5 Results and discussion Heat-denatured whey protein can reinforce the yogurt gel matrix, but is also a major contributor to increased gel water-holding capacity, firmness, and apparent/complex viscosity (Cho et al., 1999; Sodini et al., 2004). Previous observations indicate decreased firmness in yogurts made with MWP compared to controls (Barrantes et al., 1994; Sandoval-Castilla et al., 2004; Tamine et al., 1995). However, Britten and Giroux (2001) showed that the heat denaturing of the proteins can be adjusted in order to control the texture and syneresis of yogurt 8.5.1 Whey:buttermilk protein aggregates composition and properties The composition of the dried dairy products is given in Table 8.3. The protein content of WBAP was 57.3 0.1%, compared to 35.8 0.1% for powdered skim milk. Powdered whey permeate was used to increase total solids, due to its high lactose content (85.9 3.1%) and low protein content (2.2 0.1%). The fat content of WBAP was 12.7 0.3%. We hypothesized that this content made little difference because of the homogenization step (69 MPa) during the preparation of the powder. Our previous results showed that 25% whey/75% buttermilk based on protein mass gave the least hydrated aggregates. The water144 holding capacity of WBAP aggregates was calculated at 2.04 0.18 gwater/gprotein, and the percent of non-soluble protein at 78.5 3.4%. Table 8.3: Composition of heat-denatured whey:buttermilk aggregates (WBAP), whey permeate (WP), and powdered skim milk. % Dry Matter Protein Lactose Fat Ash WBAP 57.3 0.1 18.8 0.0 12.7 0.3 4.9 0.0 PP 2.2 0.1 85.9 3.1 0.2 0.3 6.2 0.0 SMP 35.8 0.1 56.6 3.4 1.0 0.3 8.0 0.0 The particle size distribution of whey:buttermilk aggregate during preparation (before homogenization) varied between 0.8 m and 700 m with an average of 50.5 m (results not shown). Lelièvre (1990) concluded that incorporation of larger heat-denatured whey protein particles during the cheese-making process interferes with the casein network. To facilitate particle retention and minimize product defects, aggregate particle size distribution should range between 0.1 m and 10 m. Homogenizing at a pressure of 69 MPa prior to freeze-drying reduced WBAP particle size to an average of 2.7 m (results not shown). 8.5.2 Particle size in skim milk Particle size distribution, after 5 minutes at 85°C, in enriched skim milk containing various levels of WBAP replacing added powdered skim milk, is shown in Figure 8.1. As shown in Table 8.2, only the substitution level had a significant effect on particle size (P = 0.0002). Average size was 22.9 19.1 m for skim milk and 159.3 36.5 m, 90.5 42.8 m, 129.6 61.6 m, 114.7 27.4 m, and 170.7 50.9 m respectively for 20%, 40%, 60%, 80%, and 100% of substitution of powdered skim milk by WBAP (results not shown). Comparison of the means indicates that average particle size was lower in skim milk than in skim milk + WBAP. 145 The higher average size could be attributed to possible interactions between skim milk proteins and aggregates present in WBAP. To the best of our knowledge, no conclusion has been reached in the literature about the properties of heat-denatured milk protein aggregates, and particularly their ability to participate in subsequent formation of heatinduced aggregates. Some researchers have concluded that added heat-denatured whey proteins are incorporated into other protein matrices, but this idea remains a hypothesis (Sandoval-Castilla et al., 2004; Tamine et al., 1995). The present results suggest that added heat-denatured whey proteins do interact with milk proteins during heating. Another plausible explanation would be that WBAP is less dispersible than skim milk protein, especially after only 30 min of stirring. 15 Volume (%) 10 5 0 0.01 0.1 1 10 Particle Size (µm) 100 1000 Figure 8.1: Particle size distribution profile in skim milk after 5 min at 85°C, with added powdered skim milk substituted 0% () 20% (), 40% (), 60% (), 80% () and 100% () with WBAP. 8.5.3 Exposure of thiol groups Exposed thiol group concentration as a function of milk composition is presented in Figure 8.2. As shown in Table 8.2, both the level of substitution (P < 0.0001) and the heating time 146 (P = 0.0174) had a significant effect on the exposure of thiol groups. The concentration was 4.3 1.6 µmole/gprot in milk enriched with powdered skim milk. When heated alone in water, WBAP had 38.3 5.3 µmole of exposed thiol groups per g of protein. The concentration in the skim milk + WBAP mixture (17.5 2.6 µmole/gprot) was unexpectedly higher than the calculated theoretical concentrations, 9.3 1.6 µmole/gprot. The calculated theoretical concentrations were not significantly different from the level measured in the control skim milk. 5 c [-SH]free (10-5 mole/gprot) 4 3 b 2 1 a a a P A B W + sm et ic al T he or Sk im m ilk + W B A P W B A P C on tr ol 0 Figure 8.2: Effect of milk composition on thiol group exposure following heating at 85°C for 25 minutes (including ramp time). A Duncan post-test was applied to compare the means to the control. These results suggest that re-suspended WBAP aggregates could interact with skim milk proteins. As reported in Chapter 5, the exposition of free thiol group increased during the first 10 minutes of heating (from 10.0 6.04 µmole/gprot to 12.1 9.97 µmole/gprot) and then stayed constant. 147 8.5.4 Distribution of particles in enriched skim milk before acidification Confocal laser scanning microscope images of enriched skim milk heated to 100°C, mixed with glucano-delta-lactone and stained are presented in Figure 8.3. These show that protein aggregates (diameter ~ 30 m) after heating were fewer in milk enriched with powdered skim milk than in milk enriched with WBAP, which contained larger aggregates and many small aggregates. The amount of free phospholipids or MFGM fragments was also greater in the WBAP-enriched milk. Figure 8.3: Images of control (enriched with powdered skim milk) skim milk (a) and skim milk + WBAP aggregates (0.6 g/100 mL; b) at pH 6.8, using the Nikon C1 confocal laser scanning microscope at 60X. Proteins were dye in green and phospholipids in red. 8.5.5 Texture of yogurts 8.5.5.1 Appearance of yogurts Lucey and Singh (1998a) defined the appearance of set-type gel as smooth with no cracks and no surface whey Control yogurts and those enriched with WBAP both had a smooth consistency with no whey separation. All gels were consistent and free of cracks or holes. At all levels of substitution of powdered skim milk with WBAP, a slight sedimentation of particles was observed. At 100% substitution, the gels were grainy. Schmidt et al. (1980) attributed the grainy texture to the heat treatment, and particularly to treatments at 90°C for 30 minutes. Since our yogurt-making process involved heating to 85°C, these defects are 148 likely due to insufficient hydration time (30 min) for the denatured proteins in the added WBAP. 8.5.2.2 Firmness The effect of WBAP as a substitute for powdered skim milk on the firmness of the acid-set type yogurt gels is presented in Figure 8.4. Table 8.1 shows that only the substitution factor was significant (P < 0.0001). The firmness was 0.170 0.007 N for yogurts made from milk enriched with powdered skim milk, and decreased from 0.143 0.019 N at 20% substitution to 0.096 0.009 N at 100% substitution. Comparison of means shows that the firmness resulting from 0% substitution was significantly higher than was obtained even in the case of low levels of substitution. Figure 8.4: Firmness of control set-type yogurt, and yogurts enriched with whey:buttermilk protein aggregates. A Duncan post-test was applied to compare the means to the control. We believe that the decrease in firmness was due mainly to the post-heating particle size. During gelation, casein chains shrink, and the density of the matrix increases by reduction of the pore dimension (Sodini et al., 2004). It can be hypothesized that the presence of large aggregates prevented the contraction of the casein network, and consequently weakened the 149 gel at certain points. This idea is consistent with the trend of the particle distribution in skim milk and enriched skim milk. 8.5.5.3 Rupture force, adhesiveness, and relaxation The effect of WBAP as a substitute for powdered skim milk on the rupture force, adhesiveness, and relaxation of the acid-set type gels is shown in Table 8.1. Table 8.4: Summary of the mean rupture force, adhesiveness and relaxation of yogurt gels, as a function of percent substitution of powdered skim milk with whey:buttermilk aggregates as enrichment of the starting milk. A Duncan post-test was applied to compare the means to the control. Level of Substitution Rupture Force (N) Adhesiveness (N/s) Relaxation (%) 0 0.093a 0.005 -0.614a 0.083 60.2a 0.6 20 0.072b 0.010 -0.374b 0.190 56.9a,b 2.5 40 0.067b 0.005 -0.274b,c 0.071 55.7a,b 1.8 60 0.067b 0.005 -0.239b,c 0.098 52.3b,c 3.9 80 0.057c 0.011 -0,105c 0.033 49.8c,d 2.7 100 0.051c 0.009 -0.094c 0.032 47.2d 2.7 (%) *Means within a column sharing the same letter were not significantly different (P > 0.05). Table 8.1 again shows that only the substitution factor was significant (P < 0.0001). The rupture force was 0.093 0.005 N for yogurts made from milk enriched with powdered skim milk, and decreased from 0.072 0.010 at 20% substitution to 0.051 0.009 N at 100% substitution. The comparison of means showed that the yogurts made using enrichment with WBAP were more susceptible to rupture than were those made using enrichment with powdered skim milk only. However, no significant difference was observed between 20%, 40% and 60% substitution. Gel relaxation was 60.2 0.6% for 150 yogurts made from milk enriched with powdered skim milk, and decreased from 56.9 2.5% at 20% substitution to 47.2 2.7 at 100% substitution. Comparison of means shows that the relaxation resulting from 0% substitution did not differ significantly from that obtained with 20 or 40% substitution. Adhesiveness was -0.614 0.083 N/s for yogurts made from milk enriched with powdered skim milk, and increased from -0.374 0.190 N.s-1 at 20% substitution to -0.094 0.032 N/s at 100% substitution with WBAP. Comparison of means shows that increasing the level of substitution increased the tendency of the gels to adhere to the probe surface. Similar results were obtained using fat replacers (Sandoval-Castilla et al., 2004). 8.5.6 Water-holding capacity of yogurt gels The effect of substituting powdered skim milk with WBAP on the water-holding capacity (WHC) of the yogurt gels is shown in Figure 8.5. Table 8.2 shows no significant differences (P = 0.5114) between the samples. The WHC was 99.1 0.1% for yogurt made from milk enriched with powdered skim milk, and varied from 98.6 0.1% to 98.2 0.2% in yogurts containing WBAP. 151 Figure 8.5: Water-holding capacity of set-type yogurt, and yogurts made from milk enriched with whey:buttermilk protein aggregates. Gel water-holding capacity is defined as the resistance of the gel to compaction (if force > 500 g) or to whey separation (if force < 500 g). Centrifugation parameters (temperature, time, force) differ from one study to the next. However, it is an accepted view that heating the milk increases the water-holding capacity of the resulting set-type yogurt (Sodini et al., 2004). Adding large amounts of denatured whey protein led to the formation of a network composed of casein micelle chains able to immobilize large amounts of water (Dannenberg et al., 1998). It was later confirmed that adding more whey protein before heating led to increased immobilization of free water in the yogurt gel, due mainly to increased gel compactness (Puvanenthiran et al., 2002). By replacing skim milk proteins with pre-formed aggregates, the amount of denatured protein in the system was changed. This could explain the small (not significant) decrease in the water-holding capacity of the gels. The textural properties of the gels showed that the size of the added aggregates affects the strength and the distribution of the protein network. Large pores allow a more open structure, facilitating the expulsion of water. 152 8.5.7 Simulation of the gel formation Confocal micrographic images taken when gel formation in milk was completed are presented in Figure 8.6. The structure of the gel formed when the milk was enriched with powdered skim milk featured small protein aggregates linked together. Heating led to the formation of large aggregates, and fat globules were entrapped in the matrix. The images transformed via ImageJ show that the clusters were homogenous in size, and distributed evenly throughout the gel structure. The structure of the WBAP-enriched milk gel featured small and large protein aggregates linked together. Fat globules were also entrapped in the gel. The presence of yellow/orange particles suggests that the structure of these aggregates could include both protein and fat. The size of the clusters was less homogenous than in the gel made from milk enriched with powdered skim milk. Small clusters were present around the large aggregates, and large clusters were present around small aggregates. Figure 8.6: Confocal laser scanning microscope (Nikon C1) images taken of gels of control skim milk (a) and skim milk + WBAP (0.6 g/100 mL; b) at 60X with a zoom of 2.31X; proteins were dyed green and phospholipids are red. Images a’ and b’ were processed with ImageJ to color the cluster in red. 153 Figure 8.7 summarizes the general trends observed in videos of gel formation as a function of pH and time in milk enriched with powdered skim milk or WBAP. As noted above, few large particles remained in the case of powdered skim milk after the heat treatment. However, all particles were moving for the first approximately 5 minutes (pH ranged from 5.50 to 5.29). The movement was intense, meaning that the initially observed particles moved quickly out of the field of view and were replaced by other particles. The first gel structure appeared about one minute after immobilization of the particles (pH ranged from ~ 5.22 to 5.16). The structure of the gel became more apparent as the pH decreased. The matrix started to contract around pH 4.65. The trend observed was slightly different in the WBAP-milk. As noted above, numerous large particles and many small ones were present after heating. Unlike in the case of powdered skim milk enrichment, these large particles were not very mobile. Small particles were mobile but the flow was slower than in the powdered skim milk case. Particle movement also stopped sooner (at pH ~ 5.60), and the first gel structure appeared almost instantaneously. The matrix started to contract when the pH reached 4.91. For no obvious reason, emission by the fluorophores (especially Oregon 488) decreased rapidly, and the images ended up dark. However, it was still possible to see the gel using ImageJ software. 154 Figure 8.7: General trends observed in the videos of gel formation as a function of the pH and time. 155 8.7 Conclusions The present study shows that heat-denatured whey:buttermilk protein aggregate behaves more like a passive filler than an active filler in yogurt. However, heated skim milk proteins may interact because of their high thiol group content. Replacing powdered skim milk with whey:buttermilk aggregate as an enrichment for milk used for yogurt production significantly affected certain textural properties of the gels. Fracture force, firmness, and relaxation decreased as the level of substitution increased, while adhesive force increased. The water-holding capacity of yogurt gels is slightly but not significantly affected. Overall, the results show that it is possible to use heat-denatured whey:buttermilk protein aggregate in the production of set-type yogurt, but the dispersibility of the powdered aggregate must be strictly controlled in order to limit the presence of large particles. 8.8 Acknowledgements The authors thank Alexandre Bastien of the department of Animal Science at Laval University for technical support and assistance with the confocal laser scanning microscope. This work was funded by Le Fonds Québécois de la Recherche sur la Nature et les Technologies (FQRNT, Quebec city, Quebec, Canada) – Novalait Inc. (Sainte-Foy, Quebec, Canada) – Ministère de l’Agriculture et de l’Alimentation du Québec (MAPAQ, Quebec city, Quebec, Canada). 156 CHAPTER 9 GENERAL CONCLUSIONS The goal of the work presented in this thesis was to develop a new dairy ingredient for the utilization of buttermilk constituents. The general hypothesis of this doctoral thesis was that a better understanding of buttermilk constituents and especially their thermal behavior offers new possibilities for the use of buttermilk in food systems. Additionally, the combination of whey and buttermilk can offer new ingredients with new functional properties. The work carried out allowed us to validate our hypotheses. The substitution of whey proteins by buttermilk constituents during an intensive heat treatment at low pH led to the formation of mixed aggregates including whey proteins, caseins, MFGM proteins, and phospholipids. The physical and functional properties of aggregates can be controlled by increasing the level of substitution of whey by buttermilk or by incorporating different constituents from buttermilk or by promoting some interactions (covalent or non-covalent). Finally, the mixed aggregates were integrated to the formation of the casein network during acid set type yogurt production. 9.1 Achievements and original contributions To our knowledge, the work presented in this thesis will be the first investigation of the thermal behavior of constituents from buttermilk, of the formation of mixed aggregates by combination of whey and buttermilk concentrates, and of use of mixed protein aggregates for the production of acid set-type gels. The first objective of this work was to develop a new approach to form protein aggregates by combination of whey and buttermilk concentrates (Chapter 4). A process has been developed by a combination of the widely used processes to form dairy protein aggregates 157 such as intensive heat treatment and a low pH. For the first time, buttermilk concentrate has been used for the preparation of mixed aggregates. By studying the effect of the level of substitution of whey concentrate by buttermilk concentrate, we developed a knowledge that can guide the control of the physical and functional properties of the aggregates. Essentially, we have shown that the water-holding capacity of aggregates significantly decreased as a function of the proportion of buttermilk concentrate in the mixture. Furthermore, rheological properties of the mixtures after heating suggested that the proteins from whey interacted with the constituents of buttermilk. This study was critical for the continuation of this doctoral thesis. Results of this part gave clear evidence to support our hypothesis that a combination of whey and buttermilk offers new ingredients with new properties and set the parameters for following works (preparation of the concentrates, volumetric protein concentrations, level of substitution, pH, temperature, time of heating). The second objective of this work aimed at the better understating of the effects of the presence of buttermilk constituents on the heat-induced formation of protein aggregates (Chapter 5). The hypothesis that constituents of buttermilk play an important role in the formation of the mix aggregates and their properties was imposed by our previous observations (Chapter 4). Results of this work allowed us to verify our hypothesis that buttermilk constituents initiate the formation of aggregates and gave us new knowledge on the thermal behavior of proteins from buttermilk. Practically, results showed that caseins, MFGM proteins, and whey proteins from buttermilk interacted together during the preparation of the concentrate through both covalent and non-covalent interactions. Furthermore, confocal laser scanning microscopy images and SDS-PAGE gels showed that whey proteins and MFGM proteins from the mixtures interacted with the pre-formed aggregates. Results suggested that free phospholipids could be involved in the formation of both pre-formed aggregates in buttermilk and mix aggregates. Results of this part gave clear evidence of the heat-induced aggregation reactions that occur between whey proteins and buttermilk constituents at a low pH. The third objective of this work was to characterize the contribution of phospholipids from buttermilk to the formation of the heat-induced whey-buttermilk protein aggregates 158 (Chapter 6). The hypothesis that free phospholipids are involved in the formation of protein aggregates was imposed by our previous observations (Chapter 5). Results of this work allowed us to verify our hypothesis that free phospholipids are involved to the aggregate formation, but results demonstrated that a longer time than expected was necessary to bind the phospholipids with the proteins during heating. Practically, results showed that phospholipids were integrated to the aggregates at a low pH through the MFGM at a low temperature (65°C), but suggested that phospholipids can directly interact with whey proteins at a higher temperature (80°C). Results did not permit to conclude on the interaction phospholipids – proteins at a higher pH (6.8). Results of this part gave evidences on the role of a non-protein constituent of buttermilk on the formation of heatinduced protein aggregates at low pH. The fourth objective of this work aimed at understating the effect of the free thiol groups reactivity during heating on the properties of the mixed protein aggregates (Chapter 7). The hypothesis that the properties of the aggregates can be controlled by varying the thiol group reactivity was imposed by our previous observations (Chapter 4 to 6). Results of this work permit to confirm our hypothesis. Practically, results demonstrated the physical and functional properties of the mixed aggregates are controlled by the type of interactions (SH/SS vs non-covalent) that occur during the formation of the mixed aggregates rather than by the composition of the mixtures. Results of this part gave strong evidence on the role of free thiol groups during formation of aggregates and their properties. Finally, the fifth objective of this work was to evaluate the potential of the mixed aggregates for the production of acid-set type yogurts (Chapter 8). The hypothesis that the mixed aggregates will be integrated to the formation of a casein network after a heat treatment was imposed by our previous observations (Chapter 7) and the literature. Results of this work allow us to verify our hypothesis that the mix aggregates are still reactive and can initiate the formation of a protein network. Practically, results showed that the mix aggregates acted more like a passive than active fillers during the formation of the gels, but some possible interactions with the proteins from the skim milk have been observed. However, the dispersibility of the powder must be strictly improved. Results of this part 159 gave strong evidence on the potential application of the mixed aggregates for food formulation. Overall, according to our results the relationship between the free thiol groups reactivity in the WPC and the properties of the heat-induced whey protein aggregates has been summarized in Figure 9.1. First, results showed that a low portion of proteins were denatured or aggregated during the preparation of the WPC. These pre-formed aggregates exposed free thiol groups at their surface before the heat treatment (Chapter 5). Upon heating, the proteins from the WPC denature and interact with the pre-formed particles to form larger aggregates. The MFGM proteins or phospholipids if present in the WPC are rarely involved in the formation of aggregates. At the end of heating, the protein aggregates have an important water-holding capacity, and have some free thiol groups accessible at their surface (Chapter 4 and 7). The prevention of the SH/SS exchange reaction led to the formation of smaller aggregates with a higher water-holding capacity. Surprisingly, aggregates still expose free thiol group at their surface as evidenced by the free thiol concentration of the aggregates after dispersibility in SMUF buffer. However, non-covalent based aggregates are rarely found in the heated WPC (Chapter 5). 160 Figure 9.1: Schematic representation of the proposed mechanism of aggregation of proteins from WPC and the properties of the heat-induced protein aggregates. The properties of the heat-induced aggregates formed by the mix of whey and buttermilk concentrate (protein ratio of 50:50) are, however, probably different due to the differences in the proportion of the major dairy proteins (caseins, whey proteins) and minor constituents (MFGM fragments, free phospholipids). The relationship between the free thiol reactivity in the mixture WPC + buttermilk concentrate and the properties of the heatinduced protein aggregates has been summarized in Figure 9.2. First, addition of buttermilk constituents slightly decreased the accessibility of free thiol groups during heating in the mixture (Chapter 5). This result is to be attributed to the addition of the casein micelles or a more complex formation of pre-formed aggregates during the preparation of the buttermilk concentrates. However, the composition of the preformed aggregates has no significant impact on the properties of the mixed aggregates. These pre-formed particles exposed free thiol groups at their surface before the heat 161 treatment (Chapter 5). Upon heating, the proteins from the WPC and the buttermilk denature and interact with the pre-formed particles (from both) to form larger aggregates. At the end of heating, the protein aggregates have a lower water-holding capacity (WHC) than aggregates formed in the heated WPC (Chapter 4). Moreover, the WHC of mixed aggregates can be controlled by variation of the proportion of buttermilk constituents in the mixtures. The aggregates are also smaller. The proportion of buttermilk constituents added can also control the size because results from Chapter 5 showed that the size of the aggregates is related to the number of pre-formed aggregates. The mixed aggregates also expose free thiol groups at their surface. The prevention of the SH/SS exchange reaction led to the formation of smaller aggregates with a higher water-holding capacity. Surprisingly, aggregates still expose free thiol group at their surface. That kind of aggregates is found during heating of the WPC mixed with the different buttermilks, at any time (Chapter 5). Figure 9.2: Schematic representation of the proposed mechanism of aggregation of constituents from WPC + BC mixtures and the properties of the heat-induced protein aggregates. 162 9.2 Significance of the results The work presented in this thesis is expected to have impact in various fields of dairy science and technology. This thesis work led to several advancements in our understanding of the thermal behavior of both whey and buttermilk constituents. For years, they have been considered to be by-products with limited applications. Results presented in this thesis proposed a new approach for their use and potential applications. The thermal behavior of whey proteins has been well studied in model systems (WPI, βlactoglobulin isolate, etc). However, to our knowledge very few studies used whey directly from cheese production. The results we have presented demonstrated that protein aggregates are formed during the preparation of the WPC, including cheese-making process, skimming, and filtration. The pre-formed aggregates are composed mainly from whey proteins and involve mostly SH/SS exchanges. Upon heating, whey proteins are denatured and associated with the pre-formed aggregates mostly through SH/SS exchanges but also through non-covalent interactions. MFGM material is rarely involved in the formation of the aggregates. Results also showed that aggregates formed in the WPC hold a lot of water. Previous results from this research group have shown that the butter-making process, specifically the pasteurization of the cream, affects the buttermilk constituents. To our knowledge, no study has focused at the thermal behavior of buttermilk constituents. The results we have presented demonstrated that aggregates are formed during the preparation of the buttermilk concentrates or powders, including pasteurization of the cream, churning, skimming, filtration, drying, etc. The pre-formed aggregates are composed of whey proteins, caseins, MFGM proteins and phospholipids and involve both SH/SS exchanges and non-covalent interactions. Upon heating, the pre-formed aggregates associate with the denatured proteins and phospholipids through both SH/SS exchanges and non-covalent interactions. MFGM material is often involved in the formation of the aggregates. Results have shown that the mixed aggregates hold less water comparatively to WPC aggregates. 163 It is usually admitted that dairy protein aggregates formed during heating at a low pH act more like a passive than active filler during the production of cheese or yogurts. The results we have presented showed that mixed aggregates are still reactive due to their high concentration of free thiol groups. Results also identified some interactions between the mixed aggregates and the proteins from the skim milk during a heat treatment. The properties of the aggregates can be controlled by increasing the proportion of buttermilk constituents in the mixture and by promoting some type of interactions. These results led to visualize a large number of options for the use of our mix aggregates. 9.3 Question yet to be answered and perspectives According to results presented in chapters 5 and 7, the proposed preparation of the WPC was sufficient to standardize the composition of the concentrates (amount of protein, fat, etc), but was insufficient to standardize the properties of the protein (size, exposure of free thiol groups, etc.). It appears that a single change in the cheese-making process significantly affects the properties of the subsequent. Results presented in chapters 7 and 8 showed that dispersibilities of the powder containing aggregates were critical and significantly affected the formation of the yogurt gels. A homogenization step is necessary before drying in order to reduce the size of the mix aggregates. We also suggest adding another homogenization step in the final product, especially if the formulation involves a heating. The results presented in this thesis demonstrate the complexity of regular and whey buttermilk but also show promising applications for its use. This project did not focus on incorporation of the aggregates in cheese formulation (cheddar, mozzarella, etc). This is being performed in another research project. 164 Regarding the formation of aggregates and among the unresolved issues, the understanding of the butter-making process and its effects on the constituents is a key issue regarding the understanding of the properties of the buttermilk constituents and their potential uses. Further information on the denaturation of the MFGM proteins and the release of phospholipids from the membrane could help to understand the formation of the preformed aggregates. Further information on the denaturation of dairy protein at low pH and the type of interactions involved is capital to understand the formation of aggregates and their properties. Regarding the use of aggregates and among the unresolved issues, the optimization of the dispersibility of the aggregates powders is a key issue. Understanding the behavior of the aggregates in a new system is necessarily to make their incorporation efficient. Understanding the effect of addition of the mixed aggregates in a new system is important for the control of the properties of the aggregates. The confocal laser-scanning microscope has been a very powerful tool for this project. The uses of this tool are, however, often inadequate and the results over discussed. There is still a lot of work do be done for the analysis of the pictures and extraction of data. Standardization of the techniques is a key issue regarding to the exploitation of the pictures. With the advances in computer science and the availability of new dye and equipment for the microscope, future work should emphasize on the use of the microscope directly connected to the water-bath for the understanding of the heat-induced aggregation of protein or on the formation of gels directly on top of the microscope. Last but not least, the work presented in this thesis gave a new look regarding the use of dairy constituents. Every year, dairy powders are produced from the surplus of milk solids such as skim milk powders. The frame built during this project can be easily adapted to the other milk solids products. Each major constituent of milk play a role on the formation and on the properties of the aggregates by varying their composition and controlling their properties. In other source of proteins (skim milk for example), the constituents may be less 165 affected by the preparation. In consequence, the thermal behavior may be slightly different, but results of this project and the most recent literature give sufficient strong clues for the understanding of the aggregation mechanism and the properties of the aggregates. Protein aggregates have been mostly used as fat replacers or as water holder agents. The work presented suggests that they can also be used for the incorporation of phospholipids in the food matrix. The mixed aggregates also show very promising properties for the development of products with new textures, properties, and sensory or nutritional properties. 166 CHAPTER 10 LITERATURE CITED Alting, A. C., Hamer, R. J., De Kruif, C. G., Paques, M., & Visschers, R. W. (2003). Number of thiol groups rather than the size of the aggregates determines the hardness of cold set whey protein gels. Food Hydrocolloids, 17, 469-479. Alting, A. C., Hamer, R. J., De Kruif, C. G., & Visschers, R. W. (2000). Formation of disulfide bonds in acid-induced gels of preheated whey protein isolate. Journal of Agricultural and Food Chemistry, 48, 50001-55007. Amatayakul, T., Sherkat, F., & Shah, N. P. (2006). Physica characteristics of set yoghurt made with altered casein to whey protein ratios and EPS-producing starter cultures at 9 and 14 % total solids. Food Hydrocolloids, 20, 314-324. American-Dietetic-Association (2005). Position of the American Dietetic Association: fat replacers. Journal of American Dietetic Association, 105, 266-275. Anema, S. G. (2008). On heating milk, the dissociation of k-casein from the casein micelles can precede interactions with the denatured whey proteins. Journal of Dairy Research, 75, 415-421. Anema, S. G., & Klostermeyer, H. (1997). Heat-induces, pH-dependent dissociation of casein micelles on heating reconstituted skim milk at temperature below 100°C. Journal of Agricultural and Food Chemistry, 45, 1108-1115. Anema, S. G., Lee, S. K., & Klostermeyer, H. (2006). Effect of protein, nonproteinsoluble components, and lactose concentrations on the irreversible thermal denaturation of Beta-Lactoglobulin and Alpha-Lactalbumin in skim milk. Journal of Agricultural and Food Chemistry, 54, 7339-7348. Anema, S. G., & Li, Y. (2003a). Association of denatured whey proteins with casein micellesin heated reconstituted skim milk and its effect on casein micelle size. Journal of Dairy Research, 70(73-83). Anema, S. G., & Li, Y. (2003b). Effect of pH on association of denatured whey proteins with casein micelles in heated reconstituted skim milk. Journal of Agricultural and Food Chemistry, 51, 1640-1646. AOAC (1990). Official method of analysis, 15th ed. Association of Official Analytical Chemists, Whashington DC, USA. AOAC (2003). Official methods of analysis (17th ed.). Washington, DC: USA: AOAC. 167 Appel, K. C., Keenan, T. W., & Low, P. S. (1982). Differential scanning calorimetry of milk fat globule membranes. Biochimica et Biophysica Acta, 690, 243-250. Arakawa, T., & Timasheff, S. N. (1982). Stabilization of protein structure by sugars. Biochemistry, 21, 6536-6544. Ashokkumar, M., Bhaskaracharya, R., Kentish, S., Lee, J., Palmer, M., & Zisu, B. (2009). The ultrasonic processing of dairy products - An overview. Dairy Science and Technology, 1-22. Barrantes, E., Tamine, A. Y., Muir, D. D., & Sword, A. M. (1994). The effect of substitution of fat by microparticulate whey protein on the quality of set-type, natural yogurt. Journal of the Society of Dairy Technology, 47(2), 61-68. Barratt, M. D., Austin, J. P., & Whitehurst, R. J. (1974). The influence of the alkyl chain length of Lecithins and Lysolecithins on their interaction with αs1-casein. Biochimica et Biophysica Acta, 34, 126-135. Bauer, R., Carrotta, R., Rischel, C., & øgendal, L. (2000). Characterization and isolation of intermediates in beta-lactoglobulin heat aggregation at high pH. Biophysical Journal, 79, 1030-1038. Baynes, B. M., Wang, D. I. C., & Trout, B. L. (2005). Role of arginine in the stabilization of proteins against aggregation. Biochemistry, 44, 4919-4925. Bédard ST-Amand (2009). Effet de la pasteurisation haute du babeurre sur sa composition minérale et celle des constituants protéiques des fragments de la membrane du globule gras du lait. Department of Food Science and Nutrition vol. M.Sc. (p. 119). Quebec City: Université Laval. Berglund, L., Petersen, T. E., & Rasmussen, J. T. (1996). Structural characterization of bovine CD36 from the milk fat globule membrane. Biochemica et Biophysica Acta, 1309, 63-68. Blond, G., & Montupet, I. (1989). Capacité de rétention d'eau des protéines laitières. Lait, 69(2), 151-154. Bos, M. A., & Nylander, T. (1996). Interaction between β-lactoglobulin and phospholipids at the air/water interface. Langmuir, 12, 2791-2797. Boudreau, A., & St-Amant, L. (1984). Beurre. Sciences et technologie du lait, 213240. Brandt, R., Pepperle, M., Otto, A., Kraft, R., Boehmer, F.-D., & Grosse, R. (1988). A 13-kilodalton protein purified from milk fat globule membranes is closely related to a mammary-derived growth inhibitor. Biochemistry, 27(5), 1420-1425. 168 Britten, M., & Giroux, H. J. (2001). Acid-induced gelation of whey protein polymers: effects of pH and calcium during polymerization. Food Hydrocolloids, 15, 609617. Britten, M., Pouliot, Y., & Gauthier, S. (2002). Science et technologie des aliments: Transformation du lait, Chp9. Ingrédients laitiers. Brown, E. M., Carrol, R. J., Pfeffer, P. E., & Sampugna, J. (1983). Complex formation in sonicated mixtures of beta-lactoglobulin and phosphatidylchiline. Lipids, 18(2), 111-118. Brulé, G., Lenoir, J., & Remeuf, F. (1997). La micelle de caséine et la coagulation du lait. Tec doc. Cals, M.-M., Mailliart, P., Brignon, G., Anglade, P., & Dumas, B. R. (1991). Primary structure of bovine lactoperoxidase, a fourth member of a mammalian heme peroxidase family. European Journal of Biochemistry, 198, 733-739. Cappel, R. E., & Gilbert, H. F. (1986). Cooperative behavior in the thiol oxidation of rabbit muscle glycogen phosphorylase in Cysteamine/Cystamine redox buffers. The Journal of Physical Chemistry, 261(25), 15378-15384. Cayot, P., & Lorient, D. (1998). Structure et technofonctions des protéines du lait. Tec & Doc. CC, C. C. (2004). Fat replacers: Food ingredients for healthy eating. In: http://www.caloriecontrol.org/fatreprint.html: May 5 2004. Cecil, R. (1963). Intramolecular bonds in proteins. 1. The role of sulfur in proteins. In: H. Neurath, The proteins, composition, structure, and function. New York: Academic Press. Chatterton, D. E. W., Smithers, G., Roupas, P., & Brodkorb, A. (2006). Bioactivity of beta-lactoglobulin and alpha-lactalbumin - Technological implications for processing. International Dairy Journal, 16, 1229-1240. Cheng, S. G., Koch, U., & Brunner, J. R. (1988). Characteristics of putrified cows ' milk Xanthine oxydase and its submolecular characteristics. Journal of Dairy Science, 71(4), 901-916. Chiti, F., Stefani, M., Taddei, N., Ramponi, G., & Dobson, C. M. (2003). Rationalization of the effects of mutations on peptide and protein aggregation rates. Nature, 424(14), 805-808. Chiti, F., Taddei, N., Baroni, F., Capanni, C., Stefani, M., Ramponi, G., & Dobson, C. M. (2002). Kinetic partitioning of protein folding and aggregation. Nature Structural Biology, 9(2), 137-143. 169 Cho, Y., Lucey, J. A., & Singh, H. (1999). Rheological properties of acid milk gels as affected by the nature of the fat globule surface material and heat treatment of milk. International Dairy Journal, 9, 537-545. Coombe, D. R., & Kett, W. C. (2005). Heparan sulfate-protein interactions: therapeutic potential through structure-function insights - Review. Cellular and Molecular Life Sciences, 62, 410-424. Cornell, D. G., & Patterson, D. L. (1989). Interaction of phospholipids in monolayers with beta-lactoglobulin adsorbed from solution. Journal of Agricultural and Food Chemistry, 37(6), 1455-1459. Corredig, M., & Dalgleish, D. G. (1996). Effect of temperature and pH on the interactions of whey proteins with casein micelles in skim milk. Food Research International, 29(1), 49-55. Corredig, M., & Dalgleish, D. G. (1998). Effect of heating of cream on the properties of milk fat globule membrane isolates. Journal of Agricultural and Food Chemistry, 46, 2533-2540. Corredig, M., & Dalgleish, D. G. (1999). The mechanisms of the heat-induced interaction of whey proteins with casein micelles in milk. International Dairy Journal, 9, 233-236. Costa, M. R., Elias-Argote, X. E., Jiménez-Flores, R., & Gigante, M. L. (2010). Use of ultrafiltration and supercritical fluid extraction to obtain a whey buttermilk powder enriched in milk fat globule membrane phospholipids. International Dairy Journal, 20, 598-602. Creamer, L. K., Bienvenue, A., Nilsson, H., Paulsson, M., Van Wanroij, M., Lowe, E. K., Anema, S. G., Boland, M. J., & Jiménez-Flores, R. (2004). Heat-Induced Redistribution of Disulfide Bonds in Milk Proteins. 1. Bovine β-Lactoglobulin. Journal of Agricultural and Food Chemistry, 52, 7660-7668. Dalgleish, D. G., Van Mourik, L., & Corredig, M. (1997). Heat-induced interactions of whey proteins and casein micelles with different concentrations of α-lactalbumin and βLactoglobulin. Journal of Agricultural and Food Chemistry, 45, 4806-4813. Dannenberg, F., & Kessler, H. G. (1988). Effect of denaturation of betalactoglobulin on texture properties of set-type nonfat yoghurt. 1. Syneresis. Milchwissenschaft, 45(10), 632-634. Daubert, C. R., Hudson, H. M., Foegeding, A. E., & Prabhasankar, P. (2006). Rheological characterization and electrokinetic phenomena of charged whey protein dispersions of defined size. LWT, 39, 206-215. 170 de la Fuente, M. A., Hemar, Y., Tamehana, M., Munro, P. A., & Singh, H. (2002). Process-induced changes in whey proteins during the manufacture of whey protein concentrates. International Dairy Journal, 12, 361-369. de Wit, J. N. (1998). Nutritional and functional characteristics of whey proteins in food products. Journal of Dairy Science, 81, 597-608. De Wit, J. N. (2009). Thermal behaviour of bovine beta-lactoglobulin at temperatures up to 150°C - A review. Trends in Food Science & Technology, 20, 27-34. Dewettinck, K., Rombaut, R., Thienpont, N., Le, T. T., Messens, K., & Van Camp, J. (2008). Nutritional and technological aspects of milk fat globule membrane material. International Dairy Journal, 18, 436-457. Dobson, C. M. (2001). The structural basis of protein folding and its links with human disease. Philosophical Transactions of The Royal Society B, 356, 133-145. Donato, L., & Guyomarc'h, F. (2009). Formation and properties of the whey protein/K-casein complexes in heated skim milk -A review. Dairy Science and Technology, 89, 3-29. Donato, L., Guyomarc'h, F., Amiot, S., & Dalgleish, D. G. (2007). Formation of whey protein-k-casein complexes in heated milk : Preferential reaction of whey protein with k-casein in the casein micelles. International Dairy Journal, 17, 1161-1167. Dufourcq, J., & Faucon, J. F. (1977). Intrinsic fluorescence study of lipid-protein interactions in membrane models. Binding of melittion, an amphipathic peptide, to phospholipid vesicles. Biochimica et Biophysica Acta, 467, 1-11. Dunnil, P., & Green, D. W. (1965). Sulphydryl groups and the N <=> R conformational cahnge in beta-lactoglobulin. Journal of Molecular Biology, 15, 147-151. Elfagm, A. A., & Wheelock, J. V. (1977). Effect of alpha-lactalbumin and betalactoglobulin in bovine milk. Journal of Dairy Research, 44, 367-371. Elfagm, A. A., & Wheelock, J. V. (1978). Heat interaction between alphalactalbumin, beta-lactoglobulin and casein in bovine milk. Journal of Dairy Science, 61, 159-163. Ellman, G. L. (1959). Tissue sulfhydryl groups. Archives of biochemistry and biophysics, 82, 70-77. Ertugay, M. F., Sengül, M., & Sengül, M. (2004). Effect of ultrasound treatment on milk homogenisation and particle size distribution of fat. Turkish Journal of Veterinary and Animal Sciences, 28, 303-308. 171 Eubert, J. R., & Brunner, J. R. (1982). Interaction of kappa-casein with immobilized beta-lactoglobulin. Journal of Dairy Science, 65, 2384-2387. Foegeding, E. A., Davis, J. P., Doucet, D., & McGuffey, M. K. (2002). Advances in modifying and understanding whey protein functionality. Trends in Food Science & Technology, 13, 151-159. Foegeding, E. A., Luck, P. J., & Davis, J. P. (2006). Factors determining the physical properties of protein foams. Food Hydrocolloids, 20, 284-292. Galani, D., & Apenten, R. K. O. (1999). Heat-induced denaturation and aggregation of Beta-Lactoglobulin: kinetics of formation of hydrophobic and disulphide-linjed aggregates. International Journal of Food Science and Technology, 34, 467-476. Gallier, S., Gragson, D., Jiménez-Flores, R., & Everett, D. W. (2012). β-Caseinphospholipids monolayers as model system to understand lipid-protein interactions in the milk fat globule membrane. International Dairy Journal, 22, 58-65. Gezimati, J., Creamer, L. K., & Singh, H. (1997). Heat-induced interactions and gelation of mixtures of beta-lactoglobulin and alpha-lactalbumin. Journal of Agricultural and Food Chemistry, 45, 1130-1136. Gezimati, J., Singh, H., & Creamer, L. K. (1996). Heat-induced interactions and gelation of mixtures of bovine β-lactoglobulin and serum albumin. Journal of Agricultural and Food Chemistry, 44, 804-810. Gilbert, H. F. (1990). Molecular and cellular aspects of thiol-disulfide exchange. Advances in Enzymology and Related Areas of Molecular Biology, 63, 69-172. Gilbert, H. F. (1993). Thiol/Disulfide exchange equilibria and disulfide bond stability. Methods in Enzymology, 251, 8-28. Glueck, C. J., Streicher, P. A., Illig, E. K., & Weber, K. D. (1994). Dietary fat substitutes. Nutrition Research, 14(10), 1605-1619. Guyomarc'h, F. (2006). Formation of heat-induced protein aggregates in milk as a means to recover the whey protein fraction in cheese manufacture, and potential of heattreating milk at alkaline pH values in order to keep its rennet coagulation properties. A review. Lait, 86, 1-20. Guyomarc'h, F., Law, A. J. R., & Dalgleish, D. G. (2003a). Formation of soluble and micelle-bound aggregates in heated milk. Journal of Agricultural and Food Chemistry, 51, 4652-4660. Guyomarc'h, F., Nono, M., Nicola, T., & Durand, D. (2009). Heat-induced aggregation of whey proteins in the presence of kappa-caasein or sodium caseinate. Food Hydrocolloids, 23, 1103-1110. 172 Guyomarc'h, F., Queguiner, C., Law, A. J. R., Horme, D. S., & Dalgleish, D. G. (2003b). Role of soluble and micelle-bound heat-induced protein aggregates on network formation in acid skim milk gels. Journal of Agricultural and Food Chemistry, 51, 77437750. Guzman-Gonzalez, M., Morais, F., Ramos, M., & Amigo, L. (1999). Influence of skimmed milk concentrate replacement by dry dairy products in a low fat set-type yoghurt model system. I: Use of whey protein concentrates, milk protein concentrates and skimmed milk powder. Journal of the Science of Food and Agriculture, 79, 1117-1122. Hambling, S. G., McAlpine, A. S., & Sawyer, L. (1992). Beta-Lactoglobulin In: P. F. Fox, Proteins. Advanced dairy chemistry, vol. 1 (pp. 141-190). London: Elsevier Applied Science. Hauser, H., & Phillips, M. C. (1979). Interactions of the polar groups of phospholipids bilayer membranes. Progress in surgace and membrane science, vol. 13 (p. 297). New York: Cadenhead, D.A., Dianielli, J.F.,Eds. Havea, P., Carr, A. J., & Creamer, L. K. (2004). The roles of disulphide and noncovalent bonding in the functional properties of heat-induced whey protein gels. Journal of Dairy Research, 71, 330-339. Havea, P., Singh, H., & Creamer, L. K. (2000). Formation of new protein structures in heated mixtures of BSA and α-lactalbumin. Journal of Agricultural and Food Chemistry, 48, 1548-1556. Havea, P., Singh, H., Creamer, L. K., & Campanella, O. H. (1998). Electrophoretic characterization of the protein products formed during heat treatment of whey protein concentrate solutions. Journal of Dairy Research, 65, 79-91. Heid, H. S., Schnölzer, M., & Keenan, T. W. (1996). Adipocyte differentiationrelated protein is secreted into milk as a constituent of milk lipid globule membrane. Biochemistry Journal, 320, 1025-1030. Hennessy, R. J. (2011). Studying milk coagulation kinetics with laser scanning confocal microscopy, image processing, and computational modeling. Biomedical Engineering, vol. M.Sc. San Luis Obispo: California Polytechnic State University. Hoffmann, M. A. M., & Van Mil, P. J. J. M. (1997). Heat-induced aggregation of βlactoglobulin: Role of the free thiol group and disulfide bonds. Journal of Agricultural and Food Chemistry, 45, 2942-2948. Horne, D. S. (2006). Casein micelle structure: Models and muddles. Current Opinion in Colloid & Interface Science, 11, 148-153. 173 Houlihan, A. V., Goddard, P. A., Nottingham, S. M., Kitchen, B. J., & Masters, C. J. (1992). Interactions between the bovine milk fat globule membrane and skim milk components on heating whole milk. Journal of Dairy Research, 59, 187-195. Hvarregaard, J., Andersen, M. K., Berglund, L., Rasmussen, J. T., & Petersen, T. E. (1996). Characterization of glycoprotein PAS-6/7 from membranes of bovine milk fat globules. European Journal of Biochemistry, 240, 628-636. Iametti, S., de Gregori, B., Vecchio, G., & Bonomi, F. (1996). Modifications occur at different structural levels during the heat denaturation of β-lactoglobulin. European Journal of Biochemistry, 237, 106-112. IDF (2001). Milk and milk products. Detrmination of lactose in milk and dairy products by enzymatic method (Routine method). International Dairy Federation, Brussels, Belgium, Standard 79B. IDF (2002). Determination of nitrogen content routine method using combustion according to the Dumas principle. IDF 186, Brussels(Belgium). IDF (2008). Skim milk, whey and buttermilk. Determination of fat content Gravimetric method (Reference method). International Dairy Federation, Brussels, Belgium, Standard 22. Jambrak, A. R., Lelas, V., Mason, T. J., Kresic, G., & Badanjak, M. (2009). Physical properties of ultrasound treated soy proteins. Journal of Food Engineering, 93, 386-393. Jambrak, A. R., Mason, T. J., Lelas, V., Herceg, Z., & Herceg, I. L. (2008). Effect of ultrasound treatment on solubility and foaming properties of whey protein suspensions. Journal of Food Engineering, 86, 281-287. Jiménez-Flores, R., & Brisson, G. (2008). The milk fat globule membrane as an ingredient: why, how, when ? Dairy Science and Technology, 88, 5-18. Joshi, N. S., Thakar, P. N., & Jana, A. H. (1994). Utilization of buttermilk in cheese making - A review. Indian Food Packer, March-April, 59-65. Kalab, M., Emmons, D. B., & Sargant, A. G. (1976). Milk gel structure. V. Microstructure of yoghurt as related to the heating of milk. Milchwissenschaft, 31, 402-408. Kanno, C. M., & Kim, D. H. (1990). A simple procedure for the ppreparation of bovine-milk fat globule membrane and a comparison of its composition, enzymaticactivities, and electrophoretic properties with those prepared by other methods. Agricultural and Biological Chemistry, 54, 2845-2854. 174 Keenan, T. W., & Dylewski, D. P. (1995). Intracellular origin of milk lipid globules and the nature of structure of milk fat globule membrane. In: P. F. Fox, Advanced dairy chemistry Lipids, vol. 2. London: Chapman & Hall. Keenan, T. W., & Mather, I. H. (2006). Intracellular origin of milk fat globules and the nature of the milk fat globule membrane. In: Birkhäuser, Advanced Dairy Chemistry: Lipids, vol. 2 (pp. 137-170): P.F.Fox, P.L.H. McSweeney. Kim, H. H. Y., & Jiménez-Flores, R. (1995). Heat-induced interactions between the proteins of milk fat globule membrane and skim milk. Journal of Dairy Science, 78, 24-35. Kinsella, J. E., & Whitehead, D. M. (1989a). Proteins in whey: chemical, physical, and functional properties. Advances in Food and Nutrition Research, 33, 343-438. Kinsella, J. E., Whitehead, D. M., Brady, J., & Bringe, N. A. (1989b). Milk proteins: possible relationships of structure and function. In: P. F. Fox, Developments in dairy chemistry, vol. 4 (pp. 55-95). London/New York: Elsevier Applied Science. Kitabatake, N., Wada, R., & Fujita, Y. (2001). Reversible conformational change in beta-Lactoglobulin A modified with N-Ethylmaleimide and resistance to molecular aggregation on heating. Journal of Agricultural and Food Chemistry, 49(4011-4018). Kontopidis, G., Holt, C., & Sawyer, L. (2004). Invited Review: β-Lactoglobulin Binding Properties, Structure and function. Journal of Dairy Science, 87, 785-796. Korhonen, H., Marnila, P., & Gill, H. S. (2000). Milk immunoglobulins and complement factors. British Journal of Nutrition, 84(suppl.1), S75-S80. Kresic, G., Lelas, V., Jambrak, A. R., Herceg, Z., & Brncic, S. R. (2008). Influence of novel food processing technologies on the rheological and thermophysical properties of whey proteins. Journal of Food Engineering, 87, 64-73. Krishnan, S., Chi, E. Y., Wood, S. J., Kendrick, B. S., Li, C., Garzon-Rodriguez, W., Wypych, J., Randolph, T. W., Narhi, L. O., Biere, A. L., Citron, M., & Carpenter, J. F. (2003). Oxidative dimer formation is the critical rate-limiting step for Parkinson's disease alpha-synuclein Fibrillogenesis. Biochemistry, 42, 829-827. Kvistgaard, A. S., Pallesen, L. T., Arias, C. F., Lopez, S., Petersen, T. E., Heegaard, C. W., & Rasmussen, J. T. (2004). Inhibitory effects of human and bovine milk constituents on rotavirus infections. Journal of Dairy Science, 87, 4088-4096. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227(5259), 680-685. Lawrence, R. C. (1989). The use of ultrafiltration technology in cheese making. Bulletin of the International Dairy Federation, 240, 1-15. 175 Lawrence, R. C. (1993). Incorporation of whey protéins in cheese. Factors affecting the yield of cheese, IDF Special Issu 9301, 79-87. Lebeuf, Y., Lacroix, C., & Paquin, P. (1998). Effet de l'incorporation des protéines du lactosérum dénaturées et microparticulées dans le cheddar jeune. Lait, 78, 303-318. Lefèvre, T., & Subirade, M. (1999). Structural and interaction properties of βlactoglobulin as studied by FTIR spectroscopy. International Journal of Food Science and Technologie, 34, 419-428. Lefèvre, T., & Subirade, M. (2000). Interaction of β-lactoglobulin with phospholipid bilayers: a molecular level elucidation as revealed by infrared spectroscopy. International Journal of Biological Macromolecules, 28, 59-67. Lelièvre, J. (1990). Incorporation of size-reduced dispersions of whey proteins into Mozzarella cheese. Proceedings, 23rd IDF Congress (Montreal), 517. Livney, Y. D., & Dalgleish, D. G. (2004). Specificity of disulfide bond formation during thermal aggregation in solution of beta-lactoglobulin B and kappa-casein A. Journal of Agricultural and Food Chemistry, 52, 5527-5532. Livney, Y. D., Verespej, E., & Dalgleish, D. G. (2003). Steric effects governing disulfide bond interchange during thermal aggregation in solutions of beta-lactoglobulin and alpha-lactalbumin. Journal of Agricultural and Food Chemistry, 51, 8098-8106. Lucey, J. A., & Singh, H. (1998a). Formation and physical properties of acid milk gels: a review. Food Research International, 30(7), 529-542. Lucey, J. A., Tamehana, M., Singh, H., & Munro, P. A. (1998b). Effect of interactions between denatured whey proteins and casein micelles on the formation and rheological properties of acid skim milk gels. Journal of Dairy Research, 65, 555-567. Lucey, J. A., Tet Teo, C., Munro, A., & Singh, H. (1997). Rheological properties at small (dynamic) and large (yield) deformations of acid gels made from heated milk. Journal of Dairy Research, 64, 591-600. Madureira, A. R., Pereira, C. I., Gomes, A. M. P., Pintado, M. E., & Malcata, F. X. (2007). Bovine whey proteins - Overview on their main biological properties. Food Research International, 40, 1197-1211. Mather, I. H. (2000). A review and proposed nimenclature for major proteins of the milk-fat globule membrane. Journal of Dairy Science, 83, 203-247. Mather, I. H., Weber, K., & Keenan, T. W. (1977). Membranes of mammary gland. XII. Loosely associated proteins and compositional heterogeneity of bovine milk fat globule membrane. Journal of Dairy Science, 60(3), 394-402. 176 Mattson, M. P., Pedersen, W. A., Duan, W., Culmsee, C., & Camandola, S. (1999). Cellular and molecular mechanisms underlying perturbed energy metabolism and neuronal degeneration in Alzheimer's and Parkinson's diseases. Annals New York Academy of Sciences, 893, 154-175. McPherson, A. V., & Kitchen, B. J. (1983). Review of the progress of dairy science: the bovine milk fat globule membrane- its formation and secretion of fat globules and origin of MFGM. Journal of Dairy Research, 50, 107-133. Mignot, B. M., & Tracard, H. (1976). Procédé d'incorporation aux fromages, sous forme colloïdale, de lactoprotéines thermocoagulables du lait. vol. FR2301180. Mistry, V. V., Metzger, L. E., & Maubois, J. l. (1996). Use of Ultrafiltered sweet buttermilk in the manufacture of reduced fat cheddar cheese. Journal of Dairy Science, 79, 1137-1145. Modler, A. J., Gast, K., Lutsch, G., & Damaschun, G. (2003). Assembly of amyloid protofibrils via critical oligomers- A novel pathway of amyloid formation. Journal of Molecular Biology, 325(135-148). Modler, H. W., & Kalab, M. (1983a). Microstructure of yogurt stabilized with milk proteins. Journal of Dairy Science, 66, 430-437. Modler, H. W., Larmond, M. E., lin, C. S., Froehlich, D., & Emmons, D. B. (1983b). Physical and sensory properties of yogurt stabilized with milk proteins. Journal of Dairy Science, 66, 422-429. Morin, P. (2006). On the fractionation of buttermilk by microfiltration membranes. Department of Food Science and Nutrition vol. Ph.D (p. 155). Quebec City: Université Laval. Morin, P., Pouliot, Y., & Britten, M. (2008). Effect of buttermilk made from creams with different heat treatment. Histories on properties of rennet gels and model cheeses. Journal of Dairy Research, 91, 871-882. Morr, C. V., & Ha, Y. W. (1993). Whey protein concentrates and isolates: Processing and functional properties. Critical Reviews in Food Science and Nutrition, 33(6), 431-476. Morr, C. V., & Josephson, R. V. (1968). Effect of calcium, N-ethylmaleimide and casein upon heat-induced whey protein aggregation. Journal of Dairy Science, 51, 13491355. Oldfield, D. J., Singh, H., Taylor, M. J., & Pearce, K. N. (1998). Kinetics of denaturation and aggregation of whey proteins in skim milk heated in an Ultra-high temperature (UHT) pilot plant. International Dairy Journal, 8, 311-318. 177 Oldfield, D. J., Singh, H., Taylor, M. J., & Pearce, K. N. (2000). Heat-induced interactions of beta-lactoglobulin and alpha-lactalbumin with the casein micelle in pHadjusted skim milk. International Dairy Journal, 10, 509-518. Omayma, E. S., & Youssef, M. M. (2007). Fat replacers and their applications in food products: a review. Alex. Journal of Food Science & Technology, 4(1), 29-44. Ong, R. L., Marchesi, V. T., & Prestegard, J. H. (1981). Small unilamellar vesicles containing Glycophorin A. Chemical characterization and proton nuclear magnetic resonance studies. Biochemistry, 20(15), 4283-4292. Pallesen, L. T., Andersen, M. H., Nielsen, R. L., Berglund, L., Petersen, T. E., Rasmussen, L. K., & Rasmussen, J. T. (2001). Purification of MUC1 from bovine milk-fat globules and characterization of a corresponding full-length cDNA clone. Journal of Dairy Science, 84, 2591-2598. Pantaloni, D. (1964). Etude de la transformation R => S de la beta-lactoglobuline par spectropolarimétrie et par spectrophotométrie de différences. Comptes Rendus Academic Science, 258, 5753-5756. Parker, E. A., Donato, L., & Dalgleish, D. G. (2005). Effects of added sodium caseinate on the formation of particles in heated milk. Journal of Agricultural and Food Chemistry, 53, 8265-8272. Parnell-Clunies, E. M., Kakuda, Y., & Deman, J. M. (1986). Influence of heat treatment of milk on the flow properties of yoghurt. Journal of Food Science, 51(6), 14591462. Parnell-Clunies, E. M., Kakuda, Y., & Smith, A. K. (1987). Microstructure of yogurt as affected by heat-treatment of milk. Milchwissenschaft, 42, 413-417. Patist, A., & Bates, D. (2008). Ultrasonic innovations in the food industry: From the laboratory to commercial production. Innovative Food Science and Emerging Technologies, 9, 147-154. Patrick, B., Barratt, M. D., Davis, M. A. F., & Rayner, L. (1972). Lysolecithincasein interactions: II. Gel filtration, gel electrophoresis and density-gradient ultracentrifugation of the lysolecithin-αsa-casein system. Biochimica et Biophysica Acta, 255, 981-987. Prabakaran, S., & Damodaran, S. (1997). Thermal unfolding of beta-lactoglobulin: Characterization of initial unfolding events responsible for heat-induced aggregation. Journal of Food Chemistry, 45, 4303-4308. Punidadas, P., Feirtag, J., & Tung, M. A. (1999). Incorporating whey proteins into mozzarella cheese. International Journal of Dairy Technology, 52(2), 51-55. 178 Puvanenthiran, A., Williams, R. P. W., & Augustin, M. A. (2002). Structure and visco-elastic properties of set yoghurt with altered casein to whey proteins ratios. International Dairy Journal, 12, 383-391. Ramachandra Rao, H. G., Lewi, M. J., & Grandison, A. S. (1995). Effect of pH on flux during ultrafiltration of sweet whey and buttermilk. Journal of Dairy Research, 62, 441-449. Rasmussen, J. T., Berglund, L., Rasmussen, M. S., & Petersen, T. E. (1998). Assignment of disulfide bridges in bovine CD36. European Journal of Biochemistry, 257, 488-494. Raval, D. M., & Mistry, V. V. (1999). Application of ultrafiltered sweet buttermilk in the manufacture of reduced fat process cheese. Journal of Dairy Science, 82, 2334-2343. Roefs, P. F. M., & de Kruif, K. G. (1994). A model for the denaturation and aggregation of beta-lactoglobulin. European Journal of Biochemistry, 226, 883-889. Roostaee, A., Côté, S., & Roucou, X. (2009). Aggregation and amyloid fibril formation induced by chemical dimerization of recombinant prion protein in physical-like conditions. Journal of Biological Chemistry, 284(45), 30907-30916. Rosenfeld, M. A., Leonova, V. B., Konstaninova, M. L., & Razumovskii, S. D. (2009). Self-assembly of fibrin monomers and fibrinogen aggregation during ozone oxidation. Biochemistry (Moscow), 74(1), 41-46. Rüegg, M., Moor, U., & Blanc, B. (1977). A calorimetric study of the thermal denaturation of whey proteins in simulated milk ultrafiltrate. Journal of Dairy Research, 44, 509-520. Saffon, M., Britten, M., & Pouliot, Y. (2011). Thermal aggregation of whey proteins in the presence of buttermilk concentrate. Journal of Food Engineering, 103, 244-250. Sandoval-Castilla, O., Lobato-Calleros, C., Aguirre-Mandujano, E., & VernonCarter, E. J. (2004). Microstructure and texture of yogurt as influenced by fat replacers. International Dairy Journal, 14, 151-159. Sawyer, L. (2003). Beta-Lactoglobulin. In: Fox P.F., & P. L. H. McSweeney, Proteins (pp. 319-386). New York: Kluwer Academics/Plenum Publishers. Sawyer, L., & Kontopidis, G. (2000). The core lipocalin, bovine BetaLactoglobulin. Biochimica et Biophysica Acta, 1482, 136-148. Sawyer, W. H. (1968). Heat denaturation of bovine beta lactoglobulins and relevance of disulfide aggregation. Journal of Dairy Science, 51(3), 323-329. 179 Sawyer, W. H. (1969). Complex Between ß-Lactoglobulin and -Casein. A Review. Journal of Dairy Science, 52, 1347-1355. Sawyer, W. H., Coulter, S. T., & Jenness, R. (1963). Role of sulfhydryl groups in the interaction ok kappa- casein and beta- lactoglobulin. Journal of Dairy Science, 46(6), 564-565. Schmidt, D. (1980). Colloidal aspects of casein. Netherlands milk ans dairy journal, 34(1), 42-64. Schmidt, R. H., Packard, V. S., & Morris, H. A. (1984). Effect of processing on whey protein functionality. Journal of Dairy Science, 67(11), 2723-2733. Schmidt, R. H., Sistrunk, C. P., Richter, R. L., & Cornell, J. A. (1980). Heat treatment and storage effects on texture characteristics of milk and yogurt systems fortified with oilseed protein. Journal of Food Science, 45, 471-475. Schokker, E. P., Singh, H., Pinder, D. N., Norris, G. E., & Creamer, L. K. (1999). Characterization of intermediates formed during heat-induced aggregation of betalactoglobulin AB at neutral pH. International Dairy Journal, 9, 791-800. Senanayake, S. P. J. N., & Shahidi, F. (2005). Dietary fat substitutes. Bailey's industrial oil and fat products: John Wiley & Sons, Inc. Seydim, Z. B. G., Sarikus, G., & Okur, O. D. (2005). Effect of inulin and DairyLo® as fat replacer on the quality of set-type yogurt. Milchwissenschaft, 60(1), 1-5. Singh, H. (2006). The milk fat globule membrane - A biophysical system for food applications. Current Opinion in Colloid & Interface Science, 11, 154-163. Smithers, G. W. (2008). Whey and whey proteins- From "gutter-to-gold". International Dairy Journal, 18, 695-704. Smits, P., & van Brouwershaven, J. H. (1980). Heat-induced association of betalactoglobulin and casein micelles. Journal of Dairy Research, 47, 313-325. Snow, D. L., Colton, D. G., & Carraway, K. L. (1977). Purification and properties of the major sialoglycoprotein of the milk fat globule membrane. Archives of biochemistry and biophysics, 179, 690-697. Sodini, I., Morin, P., Olabi, A., & Jiménez-Flores, R. (2006). Compositional and Functional Properties of Buttermilk : A comparison between sweet, sour, and whey buttermilk. Journal of Dairy Science, 89, 525-536. Sodini, I., Remeuf, F., Hadda, S., & Corrieu, G. (2004). The relative effect of milk base, starter, and process on yogurt texture: A review. Critical Reviews in Food Science and Nutrition, 44, 113-137. 180 Spector, A. A., & Fletcher, J. E. (1970). Binding of long chain fatty acids to betalactoglobulin. Lipids, 5(4), 403-411. Spiegel, T. (1999). Whey protein aggregation under shear conditions - effects of lactose and heating temperatures on aggregate size and structure. International Journal of Food Science and Technology, 34, 523-531. Spiegel, T., & Huss, M. (2002). Whey protein aggregation under shear conditions effects of pH-value and removal of calcium. International Journal of Food Science and Technology, 37, 559-568. Spitsberg, V. L. (2005). Invited Review: Bovine milk fat globule membrane as a potential nutraceutical. Journal of Dairy Science, 88(7), 2289-2294. Stammers, M., Rowen, L., Rhodes, D., Trowsdale, J., & Beck, S. (2000). BTL-II: a polymorphic locus with homology to the butyrophilin gene family, located at the border of the major histocompatibility complex class II and class III regions in human and mouse. Immunogenetics, 51(373-382). Statistics Canada (2009). Production laitière mondiale. http://www.dairyinfo.gc.ca/index_f.php?s1=dff-fcil&s2=world-monde&s3=prod. Tamine, A. Y., Kalab, M., Muir, D. D., & Barrantes, E. (1995). The microstructure of set-type, natural yogurt made by substituting microparticulate whey protein for milk fat. Journal of the Society of Dairy Technology, 48(4), 107-111. Tanford, C. (1968). Protein denaturation. Advances in Protein Chemistry, 23, 121282. Tanford, C., Bunville, L. G., & Nozaki, Y. (1959). The reversible transformation of beta-lactoglobulin at pH 7.5. Journal of American Chemical Society, 81, 4032-4036. Taylor, M. J., & Richardson, T. (1980). Antioxidant activity of skim milk: Effect of sonication. Technical Notes. Journal of Dairy Science, 63, 1938-1942. Townend, R., Herskovits, T. T., & Timasheff, S. N. (1969). The state of amino acid residues in beta-lactoglobulin. Archives of biochemistry and biophysics, 129, 567-580. Turcot, S., St.Gelais, D., & Turgeon, S. L. (2002). Affinage de fromages allégés de type Cheddar fabriqués à partir de laits enrichis en phospholipides. Lait, 82, 209-223. Turcot, S., Turgeon, S. L., & St.Gelais, D. (2001). Effet de la concentration en phospholipides de babeurre dans le lait de fromagerie sur la production et la composition de fromages allégés de type Cheddar. Le lait, 81, 429-442. 181 Van Buren, N., Rehder, D., Gadgil, H., Matsumura, M., & Jacob, J. (2009). Elucidation of two major aggregation pathways in an IgG2 antibody. Journal of Pharmaceutical Sciences, 98(9), 3013-3030. Vanderghem, C., Bodson, P., Danthine, S., Paquot, M., Deroanne, C., & Blecker, C. (2010). Milk fat globule membrane and buttermilks: from composition to valorization. Biotechnology, Agronomy, Society and Environment, 14(3), 485-500. Varadarajan, S., Yatin, S., Aksenova, M., & Butterfield, D. A. (2000). Review: Alzheimer's amyloid beta-peptide-associated free radical oxidative stress and neurotoxicity. Journal of Structural Biology, 130, 184-208. Vasbinder, A. J., Alting, A. C., & De Kruif, K. G. (2003a). Quantification of heatinduced casein-whey protein interactions in milk and its relation to gelation kinetics. Colloids and Surfaces, B: Biointeraces(31), 115-123. Vasbinder, A. J., & de Kruif, C. G. (2003b). Casein-whey interactions in heated milk: The influence of pH. International Dairy Journal, 13, 669-677. Villamiel, M., & de Jong, P. (2000). Influence of high-intensity ultrasound and heat treatment in continuous flow on fat, proteins, and native enzymes of milk. Journal of Agricultural and Food Chemistry, 48(472-478). Walker, Y. (1970). The centriwhey process for increasing yield in the manufacture of cheese. XVIII International Dairy Congress, vol. 1E (p. 327). Sydney. Walstra, P. (1983). Formation of emulsion. In: M. Decker, Encyclopedia of Emulsion Technology: Basis Theory, vol. 1 (pp. 57-127). New York. Walstra, P., Wouters, J. T. M., & Geurts, T. J. (2006). Dairy Science and Technology Handbook Second Edition. Wang, W., Nema, S., & Teagarden, D. (2010). Protein aggregation - Pathways and influencing factors. International Journal of Pharmaceutics, 390, 89-99. Watanabe, K., & Klostermeyer, H. (1976). Heat-induced changes in sulphydryl and disulphide levels of beta-lactoglobulin A and the formation of polymers. Journal of Dairy Research, 43, 411-418. Wei, Y., Chen, L., Chen, J., Ge, L., & He, R. Q. (2009). Rapid glycation with Dribose induces globular amyloid-like aggregation of BSA with high cytotoxicity to SHSY5Y cells. BMC Cell Biology, 10(10). Xiong, Y. L., Dawson, K. A., & Wan, L. (1993). Thermal aggregation of betalactoglobulin. Effect of pH, ionoc environment, and thiol reagent. Journal of Dairy Science, 76(70-77). 182 Ye, A., Singh, H., Oldfield, D. J., & Anema, S. (2004a). Kinetics of heat-induced association of β-lactoglobulin and α-lactalbumin with milk fat globule membrane in whole milk. International Dairy Journal, 14, 389-398. Ye, A., Singh, H., Taylor, M. J., & Anema, S. (2002). Characterization of protein components of natural and heat-treated milk fat globule membranes. International Dairy Journal, 12, 393-402. Ye, A., Singh, H., Taylor, M. W., & Anema, S. (2004b). Interactions of whey proteins with milk fat globule membrane proteins during heat treatment of whole milk. Lait, 84, 269-283. Zemser, M., Gorinstein, S., & Friedman, M. (1998). The structure-function relationship of ovalbumin matrix as the result of protein denaturation - Short communication. Nahrung, 42(3/4), 252-253. Zisu, B., Bhaskaracharya, R., Kentish, S., & Ashokkumar, M. (2009). Ultrasonic processing of dairy systems in large scale reactors. Ultrasonic Sonochemistry, doi:10.1016/j.ultsonch.2009.10.014. 183
© Copyright 2024