Langmuir 2007, 23, 4155-4166
4155
Whey Protein Soluble Aggregates from Heating with NaCl: Physicochemical, Interfacial, and Foaming Properties Christophe Schmitt,*,† Claudine Bovay,† Martine Rouvet,† Sabrina Shojaei-Rami,‡ and Eric Kolodziejczyk† Department of Food Science and Department of Bioanalytical Science, Nestle´ Research Center, Vers-chez-les-Blanc, CH-1000 Lausanne 26, Switzerland ReceiVed NoVember 7, 2006. In Final Form: January 23, 2007 Whey protein isolate was heat-treated at 85 °C for 15 min at pH ranging from 6.0 to 7.0 in the presence of NaCl in order to generate the highest possible amount of soluble aggregates before insolubility occurred. These whey protein soluble aggregates were characterized for composition, hydrodynamic diameter, apparent molecular weight, ζ-potential, surface hydrophobicity index, activated thiol group content, and microstructure. The adsorption kinetics and rheological properties (E′, ηd) of these soluble aggregates were probed at the air/water interface. In addition, the gas permeability of a single bubble stabilized by the whey protein soluble aggregates was determined. Finally, the foaming and foamstabilizing properties of these aggregates were measured. The amount of whey protein soluble aggregates after heat treatment was increased from 75% to 95% from pH 6.0 to pH 7.0 by addition of 5 mM to 120 mM NaCl, respectively. These soluble aggregates involved major whey protein fractions and exhibited a maximum of activated thiol group content at pH > 6.6. The hydrodynamic radius and the surface hydrophobicity index of the soluble aggregates increased from pH 6.0 to 7.0, but the molecular weight and ζ-potential decreased. This loss of apparent density was clearly confirmed by microscopy as the soluble aggregates shifted from a spherical/compact structure at pH 6.0 to a more fibrillar/elongated structure at pH 7.0. Surface adsorption was faster for soluble aggregates formed at pH 6.8 and 7.0 in the presence of 100 and 120 mM NaCl, respectively. However, interfacial elasticity and viscosity measured at 0.01 Hz were similar from pH 6.0 to 7.0. Single bubble gas permeability significantly decreased for aggregates generated at pH > 6.6. Furthermore, these aggregates exhibited the highest foamability and foam liquid stability. Air bubble size within the foam was the lowest at pH 7.0. The coarsening exponent, R, fell within predicted values of 1/3 and 1/ , except for very dry foams where it was 1/ . 2 5
Introduction Whey proteins are widely used as an ingredient in manufactured foods. They exhibit a high nutritional profile (protein efficiency ratio is around 120 compared to 100 for milk caseins), but additionally a wide range of functional properties.1,2 The variety of functional groups present in whey proteins (especially alkoyl and thiol groups) confers them a high level of secondary structure (R-helix, β-sheet). This enables whey proteins to stabilize waterbased structures (gels), fat-based structures (emulsions), and airbased structures (foams) upon denaturation (chemical, physical, enzymatic).3-5 Whey protein isolates (WPI, protein content >90%) obtained by microfiltration/ultrafiltration of milk or ionic chromatography/ ultrafiltration of liquid whey are often used to stabilize air/water interfaces in foams.6 These protein isolates are mainly composed by β-lactoglobulin (more than 50% of the total whey proteins) and R-lactalbumin (10-15% of total whey proteins), which are well-known for their surface activity and foam-stabilizing properties.7,8 * E-mail:
[email protected]. Phone: 41 21 785 89 36. Fax: +41 21 785 85 54. † Department of Food Science. ‡ Department of Bioanalytical Science. (1) Kinsella, J. E. Crit. ReV. Food Sci. Nutr. 1984, 21, 197. (2) de Wit, J. N. J. Dairy Sci. 1998, 81, 597. (3) Mulvihill, D. M.; Kinsella, J. E. Food Technol. 1987, 41, 102. (4) Rodriguez Patino, J. M.; Rodriguez Nino, M. R.; Carrera Sanchez, C. J. Agric. Food Chem. 1999, 47, 3640. (5) Zhu, H.; Damodaran, S. J. Agric. Food Chem. 1994, 42, 846. (6) Foegeding, E. A.; Luck, P. J.; Davis, J. P. Food Hydrocolloids 2006, 20, 284. (7) Davis, J. P.; Foegeding, E. A. J. Food Sci. 2004, 69, 404. (8) Damodaran, S. J. Food Sci. 2005, 70, 54.
Interfacial and foaming properties of whey proteins have been widely tested on “native” proteins at varying pH, ionic strength, and protein concentration.9 However, as most of the food products are submitted to heat treatment to ensure microbiological safety during their processing, it is very likely that a fraction of the whey proteins are no longer in their native state, but in an aggregated state.10 The resulting interfacial and foaming properties will thus be a combination of that of the native whey proteins and those of the aggregates. Several papers have investigated the effect of whey protein thermal aggregation on such functional properties. The coupled effect of pH and heat treatment of a WPI led to the improvement of the foam stability by 65% at pH 5.0.11 From a study on the effect of heat treatment on the foamability and foam stability of WPI at neutral pH, Zhu and Damodaran5 emphasized the importance of the monomeric/polymeric ratio of protein for either improved foamability (ratio 60:40) or foam stability (ratio 40:60). These results were confirmed by Davis and Foegeding7 who mixed native whey protein together with whey protein polymers generated by heat treatment at 10% protein content. Here again, a foam stability improvement was observed for a ratio of 40:60, whereas foamability decreased. Bals and Kulozik12 demonstrated that a denaturation protein yield of 70% was required to reduce drainage in whey protein foams. Very recently, it was shown that the presence of 80% soluble aggregates generated upon heat treatment of a β-lactoglobulin-enriched WPI at pH 7.0 in the presence of small ligands (NaCl, arginine, (9) Wilde, P. J.; Clark, D. C. Foam formation and stability. In Methods of testing protein functionality; Hall, G. M., Ed.; Chapman and Hall: London, 1996; p 110. (10) Kazmierski, M.; Corredig, M. Food Hydrocolloids 2003, 17, 685. (11) Phillips, L. G.; Schulman, W.; Kinsella, J. E. J. Food Sci. 1990, 55, 1116. (12) Bals, A.; Kulozik, U. Int. Dairy J. 2003, 13, 903.
10.1021/la0632575 CCC: $37.00 © 2007 American Chemical Society Published on Web 03/07/2007
4156 Langmuir, Vol. 23, No. 8, 2007
guanidinuim) enabled a marked improvement of foamability and foam stability.13 The most likely explanation for this effect resided in the combination of the fast diffusion of monomeric protein at the interface upon foaming in order to decrease surface tension coupled with the subsequent formation of a viscoelastic network (mainly driven by disulfide bond formation and hydrophobic interactions) at the interface by the soluble aggregates.7,9,13-15 Recently, several studies have shown that air bubble stabilizing properties of proteins could not reach those of nanoparticles with the exception of egg ovalbumin.16 A likely reason is that ovalbumin is able to “coagulate” at an interface to form a viscoelastic network around air bubbles, which was not the case for proteins like β-lactoglobulin, sodium caseinate, or soy globulin.16,17 However, it has been shown that the generation of soluble soy protein aggregates upon heat treatment could improve the foaming properties of an isolate due to the specific physicochemical properties (hydrophobicity) of the formed aggregates.18 In this context, the aim of this study was first to generate the highest possible amount of whey protein aggregates with controlled physicochemical properties (size, charge, molecular weight, specific surface hydrophobicity, thiol reactivity) by heat treatment of WPI at 1 wt % in the presence of NaCl at pH ranging from 6.0 to 7.0. The main idea is to generate WPI aggregates which could form the viscoelastic network that is required to reduce air bubble shrinkage in foams and reduce drainage more easily than native proteins. Thus, in a second step, the interfacial, foaming, and foam-stabilizing properties of these WPI soluble aggregates were investigated. Materials and Methods Material. Whey protein isolate, Bipro, (lot JE 030-3-420, from 30-01-2003) was obtained from Davisco Foods International, Inc. (Le Sueur, MN). The protein was purified from sweet whey by ion exchange chromatography followed by ultrafiltration and spraydrying. The powder was stored in sealed aluminum bags gassed nitrogen at 10 ( 2 °C. The composition of the WPI powder (wet basis) was 93.6 wt % protein (Kjeldahl, N × 6.38), 4.4 wt % moisture, 0.3 wt % fat, 6.4. This is an indication of the formation of larger soluble aggregates at low pH, as the turbidity is directly proportional to the molar mass and radius of the scattering objects.32 The critical sodium chloride concentration leading to the formation of the maximum amount of WPI soluble aggregates was investigated as a function of pH (Figure 1). The turbidity values obtained were significantly higher than those obtained without NaCl addition, leading to the conclusion that soluble aggregates formed were probably larger. It clearly appeared also that the critical salt concentration increased almost linearly with the pH, ranging from 5 mM at pH 6.0 to 120 mM NaCl at pH 7.0 (data not shown). This linear dependence may to be explained by considering the linear variation of the negative charges of the β-lactoglobulin (which represents almost 80% of the total WPI used in this study) between pH 6 and 8. This behavior was described early on by Tanford and co-workers33 and confirmed by recent calculations based on NMR measurements.34 In other terms, the formation of the WPI soluble aggregates up to the insolubilization is mainly governed by the neutralization of the negative charges present on the denatured protein by the positive sodium ions. Very similar salt behavior has been recently described upon denaturation aggregation of β-lactoglobulin in neutral pH conditions in the presence of NaCl, arginine HCl, or guanidinium HCl.35 The amount of soluble aggregates formed at the different pH values with or without salt addition was investigated by RPHPLC analysis (Figure 2). The conversion yield of native whey proteins was found to vary from 72% at pH 6.6 to 65% at pH 7.0. Croguennec and co-workers36 reported very similar values upon heating β-lactoglobulin dispersions at 85 °C for 15 min at
pH 6.6. The slight decrease of the yield of soluble aggregate formation upon pH increase can be attributed to the increased negative charge carried by the unfolded proteins, preventing further aggregation by repulsive interactions. In addition, the increased dissociation yield of the thiol groups, especially for β-lactoglobulin, at higher pH values might result in an increased probability that the activated non-native intermediate will react and form disulfide bonds that will terminate the propagation of aggregate growth.37 Upon addition of sodium chloride, the amount of WPI soluble aggregates markedly increased from 75% at pH 6.0 to 95% at pH 7.0. This increased soluble aggregate amount might result from the charge neutralization of unfolded and reactive whey proteins that could therefore continue to aggregate to larger sizes (as could be explained by the higher turbidity values found in the presence of salt). Hence, whey protein denaturation/aggregation close to neutral pH is mainly characterized by the formation of non-native reactive aggregates via thiol/ disulfide group activation/interchange, whose activated aggregates could further aggregate in the presence of salts due to charge screening.38-41 Electrophoresis of the WPI soluble aggregates in nonreduced and reduced conditions confirmed the RP-HPLC results (Figure 3A). It can be clearly seen that the intensity of the spots corresponding to the native major proteins R-lactalbumin and β-lactalbumin remained higher without salt than in the presence of sodium chloride, explaining the lower amount of soluble aggregates formed. Interestingly, bovine serum albumin was involved in the formation of the soluble aggregates with or without salt. A marked difference between the two systems was also the difference in the profile of the soluble aggregates that were formed.
(32) Pouzot, M.; Durand, D.; Nicolai, T. Macromolecules 2004, 37, 8703. (33) Tanford, C.; Nozaki, Y. J. Biol. Chem. 1959, 234, 2874. (34) Fogolari, F.; Ragona, L.; Licciardi, S.; Romagnoli, S.; Michelutti, R.; Ugolini, R.; Molinari, H. Proteins: Struct., Funct., Genet. 2000, 39, 317. (35) Unterhaslberger, G.; Schmitt, C.; Sanchez, C.; Appolonia-Nouzille, C.; Raemy, A. Food Hydrocolloids 2006, 20, 1006. (36) Croguennec, T.; O’Kennedy, B. T.; Mehra, R. Int. Dairy J. 2004, 14, 399.
(37) Hoffmann, M. A. M.; van Mil, P. J. J. M. J. Agric. Food Chem. 1999, 47, 1898. (38) Aymard, P.; Durand, D.; Nicolai, T. Int. J. Polym. Anal. Charact. 1996, 2, 115. (39) Bauer, R.; Rischel, C.; Hansen, S.; Øgendal, L. Int. J. Food Sci. Technol. 1999, 34, 557. (40) Le Bon, C.; Nicolai, T.; Durand, D. Int. J. Food Sci. Technol. 1999, 34, 451. (41) Schokker, E. P.; Singh, H.; Creamer, L. K. Int. Dairy J. 2000, 12, 843.
4160 Langmuir, Vol. 23, No. 8, 2007
Schmitt et al.
Table 1. Physicochemical Properties of Soluble Whey Protein Aggregates Obtained by Heat Treatment (85 °C, 15 min) of a 1 wt % Protein Dispersion in the Presence or Absence of NaCl pH/NaCl content (mM)
z-averaged hydrodynamic diameter (nm)
polydispersity index
molecular weight Mw (×106 g.mol-1)
ζ-potential (mV)
protein surface hydrophobicity index (µg.mmol-1 ANS)
accessible SH groups (nmol SH.mg-1 prot.)
6.0/0 6.2/0 6.4/0 6.6/0 6.8/0 7.0/0 6.0/5 6.2/10 6.4/40 6.6/70 6.8/100 7.0/120
120.3 ( 9.1 86.8 ( 3.3 71.6 ( 2.7 60.4 ( 3.9 56.2 ( 4.6 53.5 ( 1.8 179.6 ( 12.3 293.9 ( 30.4 260.8 ( 41.0 300.0 ( 87.4 296.3 ( 26.2 327.1 ( 41.1
0.056 ( 0.012 0.076 ( 0.017 0.130 ( 0.004 0.187 ( 0.014 0.267 ( 0.063 0.311 ( 0.041 0.085 ( 0.030 0.188 ( 0.039 0.214 ( 0.026 0.300 ( 0.060 0.372 ( 0.111 0.362 ( 0.096
27.02 ( 8.1 nda nd nd nd 0.64 ( 0.01 19.5 ( 3.5 3.0 ( 1.3 2.8 ( 0.8 3.3 ( 0.9 2.4 ( 0.9 1.1 ( 0.5
-31.8 ( 0.8 -31.1 ( 1.2 -31.7 ( 3.0 -26.8 ( 1.3 -27.9 ( 1.2 -28.2 ( 1.6 -27.1 ( 1.3 -23.9 ( 0.9 -21.4 ( 1.1 -15.4 ( 1.3 -14.1 ( 1.2 -13.9 ( 1.2
105 217 223 205 201 207 85 130 143 169 140 130
3.5 ( 0.4 7.1 ( 0.5 7.1 ( 0.1 7.3 ( 0.1 6.8 ( 0.5 7.1 ( 0.1 3.4 ( 0.1 5.2 ( 0.3 5.7 ( 0.0 6.6 ( 0.7 6.8 ( 0.1 6.8 ( 0.9
a
nd: not determined.
Figure 3. Coomassie blue stained SDS Tris-glycine electrophoresis gels of the WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 in the presence or absence of NaCl. Protein content was reduced to 0.2 wt % before deposition in the gel: nonreduced gel (A), reduced gel (B). Molecular makers (lane M), nonheated native WPI (lane 1), pH 6.0 (lane 2), pH 6.0 - 5 mM NaCl (lane 3), pH 6.2 (lane 4), pH 6.2 - 20 mM NaCl (lane 5), pH 6.4 (lane 6), pH 6.4 - 40 mM NaCl (lane 7), pH 6.6 (lane 8), pH 6.6 - 70 mM NaCl (lane 9), pH 6.8 (lane 10), pH 6.8 - 100 mM NaCl (lane 11), pH 7.0 (lane 12), pH 7.0 - 120 mM NaCl (lane 13).
as much larger soluble aggregates that could just enter the gel (Figure 3A). No intermediate spots were clearly visible when the samples were heated in the presence of NaCl, and soluble aggregates were mainly excluded from the electrophoresis gel (either because their Mw > 116.3 kDa or because their physicochemical properties were incompatible with the gel used). Nevertheless, these large aggregates were still mainly disulfidebridged, which tended to prove that they were essentially formed by the aggregation of the non-native protein monomers. Similar results on the absence of intermediate species upon heating β-lactoglobulin close to neutral conditions had already been reported by Renard,43 who described the formation of soluble aggregates with molecular weight higher than 1 × 106 g.mol-1 upon heating β-lactoglobulin at pH 7 in the presence of 100 mM NaCl. The lowest amount of soluble aggregates with intermediate molecular weight might be explained by their faster aggregation into larger aggregates when NaCl was present.13,36 Table 1 summarizes some physicochemical properties of the soluble aggregates that were formed. In the absence of NaCl, the z-averaged hydrodynamic diameter clearly decreased with increase of the pH. Concomitantly, the polydispersity index evolved in the opposite direction, indicating that the size distribution of the soluble aggregates was broader. In addition, the apparent Mw of the aggregates decreased markedly from 27 × 106 g.mol-1 to 0.6 × 106 g.mol-1. These features are consistent with the turbidity results presented in Figure 1. As already discussed, the likely explanation for the larger aggregates formed at pH 6.0 is the reduction of the electrostatic repulsions, leading to favored aggregation of the unfolded protein molecules. On the contrary, increased electrostatic repulsions together with formation of covalent disulfide bonds lead to a limitation of the aggregate size at pH 7.0. Similar trends have been described in the literature on β-lactoglobulin systems or WPI.10,37,40,41,44-47 Unfortunately, as the protein concentration, initial mineral content/composition in the powders, and applied heat treatment used differed from those used in this study, it is difficult to compare the absolute values of hydrodynamic diameter or molecular weight. Nevertheless, in the closed conditions of protein concentration (1 wt %, pure β-lactoglobulin) and pH 6.8, Hoffmann and co-workers44 reported a z-average hydrodynamic radius of 20 nm and a
Without salt, a strong band appeared around 31-36 kDa with increasing pH. This band could correspond to either β-lg dimer or to β-lg/R-la dimer. To a certain extent, these dimers could not be reduced. As these dimers were also present in the nonheated WPI (lane 1), they were not formed via disulfide bonds but presumably by hydrophobic interactions or another type of covalent cross-link such as dityrosine (Figure 3B).41,42 In addition, several oligomeric spots were present up to 116.3 kDa, as well
(42) Aeschbach, R.; Amado, R.; Neukom, H. Biochim. BioPhys. Acta 1976, 439, 292. (43) Renard, D. Etude de l’agre´gation et de la ge´lification des prote´ines globulaires: application a` la beta-lactoglobuline. Ph.D. Dissertation. University of Nantes, France, 1994. (44) Hoffmann, M. A. M.; Sala, G.; Olieman, C.; de Kruif, C. G. J. Agric. Food Chem. 1997, 45, 2949. (45) Le Bon, C.; Durand, D.; Nicolai, T. Int. Dairy J. 2002, 12, 671. (46) Schokker, E. P.; Singh, H.; Pinder, D. N.; Norris, G. E.; Creamer, L. K. Int. Dairy J. 1999, 9, 791. (47) Shimada, K.; Cheftel, J. C. J. Agric. Food Chem. 1989, 37, 161.
Whey Protein Soluble Aggregates
Langmuir, Vol. 23, No. 8, 2007 4161
Figure 4. Negative-staining transmission electron micrographs of WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 in the presence or absence of NaCl: pH 6.0 (A), pH 6.0 - 5 mM (B), pH 6.6 (C), pH 6.6 - 70 mM (D), pH 7.0 (E), pH 7.0 - 120 mM (F). Scale bars represent 0.5 µm.
molecular weight of less than 2.5 × 106 g.mol-1 after 24 h heating at 68 °C. Heat treatment of 1.7 wt % β-lactoglobulin at pH 7.0 for 10 min at 78 °C resulted in the formation of soluble aggregates larger than 2 × 106 g.mol-1 with radii of gyration larger than 20 nm.46 These results seem to fit reasonably with the order of magnitude of our results, considering the presence of ions in the WPI powder as well as the R-lactalbumin and bovine serum albumin that are known to impact the heat aggregation of β-lactoglobulin.35 The addition of sodium chloride in the system markedly impacted the size and apparent molecular weight of the WPI aggregates (Table 1). Here, an important difference was between pH 6.0, where the z-averaged diameter was around 180 nm, and pH g6.2, where it was almost constant around 300 nm. In addition,
the polydispersity index increased when pH increased toward neutrality. The molecular weight of the soluble aggregates formed at pH 6.0 was much larger than for aggregates formed at other pH (10-fold increase), which was a similar trend to that without salt. In other words, there is a clear tendency for the WPI aggregates to get denser when NaCl is added. This fact was recently described and was explained by the facilitated aggregation of the primary WPI denatured aggregates into larger ones under low electrostatic repulsion conditions.48 The specific effect of pH and ionic strength on the morphology of the WPI soluble aggregates is highlighted in the TEM pictures presented on Figure 4. Without salt, soluble aggregates were mainly compact and spherical at low pH (Figure 4A) and
4162 Langmuir, Vol. 23, No. 8, 2007
transformed into thinner “banana-shaped” aggregates at high pH (Figure 4E). An intermediate mixture of both morphologies was observed at the intermediate pH of 6.6 (Figure 4C). With addition of NaCl, WPI soluble aggregates formed at pH 6.0 exhibited a compact and spherical shape, with an apparent diameter of about 140 nm for the larger particles (Figure 4B). Upon increasing the pH to 6.6, the dispersion turned into wormlike particles, with no more spherical objects present (Figure 4D). These particles were characterized by a length of about 50 to 60 nm and seemed to be more contrasted than the equivalent sample without salt. At pH 7.0, only fibrillar aggregates with a length of about 50 nm could be distinguished. Here again, these aggregates were denser than the corresponding sample without salt (Figure 4F). Details of the microstructure of the soluble aggregates formed in the presence of NaCl are presented on Figure 5. It is important to note that the contrast between the particles and the background in this negative staining mode was much higher at low pH, revealing a denser structure of the WPI soluble aggregates. Such local densification of the WPI aggregates in the presence of salt was recently reported for β-lg. The authors explained that this was mainly due to the extensive charge neutralization on the primary soluble aggregates, preventing head-to-tail aggregation that is found at low ionic strength.48 The transition from spherical/ compact protein aggregates to linear ones upon pH increase has been already been described using TEM for β-lactoglobulin and whey protein concentrate (WPC) gels formed at a protein concentration of 10 wt %.49 However, as the protein concentration used was larger than in the present study, the sizes of the aggregates constituting the gel were slightly larger, i.e., 300 nm in diameter. Table 1 summarizes the surface charge properties of the WPI soluble aggregates (ζ-potential) as well as the specific surface hydrophobicity and amount of accessible thiols. It is interesting to note that, in the absence of sodium chloride, the apparent ζ-potential of the soluble aggregates was constant around -30 mV. This was somewhat surprising, as one would expect an increase of the ζ-potential because of the increase in total net charge of the whey proteins upon pH increase (pI is around 4.6). However, it must be recalled that the shape, hydrodynamic properties, and molecular weight of the soluble aggregates were different at pH 6.0 and 7.0. These parameters might impact the measured ζ-potential values, as the model used for calculation assumes a spherical shape for the particles.25 It must also be taken into account that only 60-70% of soluble WPI aggregates are formed without salt addition. Therefore, the measured ζ-potential is probably a combination of those of the nonaggregated protein and the soluble aggregates.50 In the presence of salt, ζ-potential obtained at pH 6.0 was close to the value obtained without salt, i.e., -30 mV. Increasing pH and salt content, however, produced a progressive decrease of the ζ-potential to about -14 mV at pH 7.0-120 mM NaCl. The interpretation of the latter results has to be made cautiously because of the differences in the types of aggregates that were formed. However, as the contribution of the nonaggregated proteins became negligible, it is fair to conclude that salt addition led to a decrease of the overall surface charge carried by the WPI soluble aggregates.35 The likely explanation could be that the more compact spherical structure of the aggregates formed at pH 6.0 (48) Pouzot, M.; Nicolai, T.; Visschers, R. W.; Weijers, M. Food Hydrocolloids 2004, 19, 231. (49) Langton, M.; Hermansson, A.-M. Food Hydrocolloids 1992, 5, 523. (50) Schmitt, C.; Sanchez, C.; Despond, S.; Renard, D.; Thomas, F.; Hardy, J. Food Hydrocolloids 2000, 14, 403.
Schmitt et al.
Figure 5. Negative-staining transmission electron micrographs of WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 in the presence of NaCl: pH 6.0 5 mM (A), pH 6.6 - 70 mM (B), pH 7.0 - 120 mM (C). Scale bars represent 200 nm.
produced a significantly lower interfacial area for adsorption of cations compared to the more extended structure formed at pH 7.0, where more positive charges (Na+ ions) could presumably bind.51,52
Whey Protein Soluble Aggregates
Figure 6. Time evolution of air/water surface tension of WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 without NaCl. Protein concentration was reduced to 0.1 wt % for measurements. Inset: initial variation of surface tension during first 50 s. pH 6.0 ([), pH 6.2 (]), pH 6.4 (2), pH 6.6 (4), pH 6.8 (b), pH 7.0 (O).
Interesting features on the specific ANS surface hydrophobicity of the soluble WPI aggregates were revealed in the presence or absence of NaCl as a function of the pH of heat treatment (Table 1). The lowest protein surface hydrophobicity index was obtained with or without NaCl at pH 6.0. Interestingly, this corresponded to the denser aggregates with the lower polydispersity index and a spherical shape. Thus, interfacial area was highly reduced in these conditions, and one may infer that the ANS specific hydrophobic patches exposed upon heating (generally described as being the tryptophan residues) have been buried in the interior of the aggregate.53,54 Protein surface hydrophobicity doubled with the increase of the pH of heat treatment up to 7.0. Here again, this implied that the specific ANS binding sites were more accessible and was fairly correlated with the more extended of the WPI soluble aggregates that are produced, resulting in an increase of their surface area. Moreover, it should be kept in mind that, without NaCl, hydrophobicity values correspond to the average contribution of the WPI soluble aggregates and the non-native monomeric or oligomeric proteins, which are known to expose hydrophobic residues upon heat treatment at temperature >70 °C close to neutral pH.43,54 Upon heat treatment in the presence of NaCl, the increase in specific ANS surface hydrophobicity was limited, with the highest value being obtained for pH 6.6 and 70 mM NaCl. The likely explanations are that (i) less intermediate non-native proteins are present, leading to a reduction of the overall protein surface hydrophobicity, and (ii) even if aggregates formed at high pH in the presence of salt were not spherical, counterions screened their surface charges, reducing the number of available ANS binding sites. Finally, the content of accessible SH groups was determined within the WPI soluble aggregates (Table 1). A marked difference was visible (51) de Lint, W. B. S.; Benes, N. E.; Lyklema, J.; Bouwmeester, H. J. M.; van der Linde, A. J.; Wessling, M. Langmuir 2003, 19, 5861. (52) da Silva, F. L. B.; Linse, S.; Jo¨nsson, B. J. Phys. Chem. B 2005, 109, 2007. (53) Manderson, G. A.; Hardmann, M. J.; Creamer, L. K. J. Agric. Food Chem. 1999, 47, 3617. (54) Palazolo, G.; Rodriguez, F.; Farrugia, B.; Pico, G.; Delorenzi, N. J. Agric. Food Chem. 2000, 48, 3817.
Langmuir, Vol. 23, No. 8, 2007 4163
Figure 7. Time evolution of air/water surface tension of WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 with NaCl. Protein concentration was reduced to 0.1 wt % for measurements. Inset: initial variation of surface tension during first 50 s. pH 6.0 - 5 mM NaCl ([), pH 6.2 - 20 mM NaCl (]), pH 6.4 - 40 mM NaCl (2), pH 6.6 - 70 mM NaCl (4), pH 6.8 - 100 mM NaCl (b), pH 7.0 - 120 mM NaCl (O).
among the WPI aggregates formed in the presence of salt. Hence, the exposure of the accessible SH was more progressive than that observed without salt, as intermediate values of 5.2 and 5.7 nmol SH.mg-1 of protein were measured at pH 6.2 and 6.4. Thus, in addition to the already reported exposure of the SH in heated whey proteins with the increase of the pH, it seemed here that the aggregation pattern induced by the presence of NaCl is also controlling the accessibility of the thiol groups with the soluble aggregates.44,47,53,55 Heat treatment of WPI at various pH values in the presence of NaCl enabled us to produce a wide range of soluble aggregates at yields >80%. These aggregates exhibited very different sizes, molecular weights, charges, surface hydrophobicity, and thiol accessibility, which supposedly will influence their interfacial and foaming properties.22,56,57 Interfacial and Foaming Properties of the WPI Soluble Aggregates. WPI soluble aggregates were tested for their ability to decrease air/water surface tension at a concentration of 0.1 wt % (Figures 6 and 7). When no salt was present during the generation of the aggregates, surface tension decreased from about 72 mN.m-1 to 52 mN.m-1 after 1800 s (Table 2). This value obtained after 1800 s was in the range of reported data for native β-lactoglobulin or WPI.7,58,59 Interestingly, the decrease of the surface tension exhibited three distinct phases: a very sharp decrease within the first 50 s, followed by a less sharp decrease until 700 s and a leveling off until 1800 s. These different phases in the kinetics of surface tension decrease could be (55) Laligant, A.; Dumay, E.; Casas-Valencia, C.; Cuq, J.-L.; Cheftel, J. C. J. Agric. Food Chem. 1991, 39, 2147. (56) Martin, A. H.; Cohen-Stuart, M. A.; Bos, M. A.; Van Vliet, T. Langmuir 2005, 21, 4083. (57) Cornec, M.; Kim, D. A.; Narsimhan, G. Food Hydrocolloids 2001, 15, 303. (58) Bos, M. A.; Van Vliet, T. AdV. Colloid Interface Sci. 2001, 91, 437. (59) Wang, Z.; Narsimhan, G. Langmuir 2005, 21, 4482.
4164 Langmuir, Vol. 23, No. 8, 2007
Schmitt et al.
Table 2. Interfacial and Foaming Properties of Soluble Whey Protein Aggregates Obtained by Heat Treatment (85 °C, 15 min) of a 0.1 wt % Protein WPI Dispersion in the Presence or Absence of NaCl pH/ionic strength (mM NaCl)
a/w surface tensiona (mN.m-1)
surface elasticity E′ a (mN.m-1)
surface viscosity ηd a (mN.m-1.s)
single bubble permeabilityb (cm.s-1)
foam capacity
foam liquid stability (s)
6.0/0 6.2/0 6.4/0 6.6/0 6.8/0 7.0/0 6.0/5 6.2/10 6.4/40 6.6/70 6.8/100 7.0/120
53.4 ( 1.6 52.7 ( 1.6 52.4 ( 1.0 52.4 ( 0.5 52.4 ( 0.5 51.6 ( 1.5 57.3 ( 2.3 53.9 ( 0.5 52.8 ( 0.5 52.8 ( 1.6 52.3 ( 1.0 52.5 ( 1.2
22.6 ( 0.6 22.9 ( 0.5 23.9 ( 0.6 27.7 ( 0.5 27.1 ( 0.8 26.3 ( 0.4 30.1 ( 0.7 17.1 ( 0.6 15.4 ( 0.4 16.0 ( 0.5 15.7 ( 0.4 20.2 ( 0.7
133.9 ( 4.3 131.9 ( 3.6 137.3 ( 4.3 165.3 ( 2.8 177.2 ( 7.4 164.5 ( 3.1 129.3 ( 9.7 113.7 ( 3.8 96.8 ( 4.4 112.2 ( 4.3 138.5 ( 6.7 139.9 ( 6.3
0.006 ( 0.001 nd 0.031 ( 0.006 nd nd 0.033 ( 0.005 0.048 ( 0.004 0.040 ( 0.003 0.038 ( 0.006 0.010 ( 0.004 0.014 ( 0.004 0.006 ( 0.006
1.23 ( 0.02 nd 1.22 ( 0.01 nd nd 1.23 ( 0.01 1.14 ( 0.06 1.06 ( 0.06 1.17 ( 0.00 1.19 ( 0.01 1.24 ( 0.00 1.19 ( 0.01
28.0 ( 16.9 nd 21.5 ( 13.4 nd nd 21 ( 1.41 27.3 ( 12.1 57.3 ( 6.5 68.0 ( 5.6 95.5 ( 9.2 121.0 ( 5.6 98.0 ( 8.5
a
measured after 1800 s equilibration. b film was equilibrated for 1800 s.
interpreted by the presence of native whey proteins together with soluble aggregated proteins. As the initial surface tension decrease is mainly due to the diffusion of the species having the highest surface activity and highest diffusion coefficient and considering that with salt the amounts of native whey proteins were similar in all samples (around 30%), it seems reasonable that the surface tension decrease kinetics were very similar on small time scales (inset Figure 6).7,60 However, on larger time scales, the decrease was more important for soluble WPI aggregates formed at high pH (>6.6). It is likely that the physicochemical properties of the soluble aggregates formed in these conditions lead to better surface activity. Here, the hydrodynamic diameter probably played an important role, as the diffusion coefficient of the corresponding aggregates would be higher (decrease with pH increase), but the combination with the increased specific surface hydrophobicity might be an additional reason for the differences between the different types of WPI aggregates. Hence, it has been shown that, for proteins to adsorb at interfaces, energetic barriers (mainly electrostatic) have to be overcome (in this study, ζ-potential values were close), and this could be achieved in the case of increased surface hydrophobicity.57,61 When aggregates were formed in presence of sodium chloride, clear differences in the surface tension behavior could be seen, compared to previous experiments (Figure 7). Although the values of the surface tension obtained after 1800 s were still in the range of 52 mN.m-1, WPI soluble aggregates generated at pH 6.0 with 5 mM NaCl lead to a significantly higher surface tension of 57 mN.m-1. By contrast, WPI aggregates generated with salt at pH >6.2 exhibited a similar, and sometimes faster, surface tension decrease than without salt even if the dispersion contained a significantly lower amount of native proteins. Note the special behavior at short time scales for pH 6.0 and 6.2 (Figure 7, inset). The reported differences on interfacial behavior are interesting, especially comparing pH 6.0 with and without salts where the amount of aggregates and native protein were close. These differences revealed that a critical amount of native protein is required to have rapidly decreasing surface tension in these two systems but WPI soluble aggregates produced with salt were not highly surface active. A partial explanation could reside in the very low protein surface hydrophobicity index, coupled with a high ζ-potential and a large z-average hydrodynamic radius (Table (60) Miller, R.; Aksenenko, E. V.; Kra¨gel, J.; O’Neil, M.; Makievski, A. V.; Fainerman, V. B. Dynamics of protein adsorption layers at liquid interfaces. In Food Colloids: biopolymers and materials; Dickinson, E., Van Vliet, T., Eds.; Royal Society of Chemistry: Cambridge, 2003; p 207. (61) MacRitchie, F.; Alexander, A. E. J. Colloid Sci. 1963, 18, 464.
1). On the other hand, to explain how larger WPI soluble aggregates generated with salt at pH >6.2 could decrease surface tension to the same extent and at the same rate as a mixture of native proteins and smaller soluble aggregates, one should consider that the ζ-potential and surface hydrophobicity index were optimum at these conditions. In that case, repulsive barriers upon protein adsorption were likely reduced by the 3-fold decrease in ζ-potential and were seemingly overcome by the surface hydrophobicity of the aggregates. The dilational viscoelastic properties of the interface were probed after 1800 s at a frequency of 0.01 Hz (Table 2). In the absence of NaCl, dilational surface elasticity increased upon pH increase, which can be explained by the increase of the accessible thiol content enabling the formation of an elastic network upon disulfide bridging.5,62 In addition, dilational surface viscosity increased by 30% from pH 6.0 to 7.0, which could be interpretated not only by an increased amount of material at the interface, but also by an increased ability for some of the protein segments to desorb upon interface compression/expansion. Likely, the more open and fibrillar structure of the aggregates formed at pH 7.0 might be responsible for this observation. It should be mentioned that the elasticity values reported here were slightly lower than those reported for native β-lactoglobulin or WPI, which can be explained by the fact that fewer structural rearrangements could occur at the interface once the proteins were aggregated.7,56,63 Upon heat treatment in the presence of salt, the evolution of the dilational viscoelastic parameters was more complex. E′ was first higher than that without salt at pH 6.0 and about 30% lower afterward. The surface dilational viscosity was lower than that without salt, but increased upon pH increase (Table 2). It is not clear here why the measured elasticity at pH 6.0 was so high, considering the high value of the surface tension and the low hydrophobicity index of the WPI soluble aggregates. It might be that this result could be explained by a particle-like adsorption that could lead to such viscoelastic properties.16 Upon increasing the pH, elasticity values remained almost constant, implying that the structure of the film led to the formation of covalent bonds, as would be the case for native proteins. However, dilatational viscosity increased, revealing more reorganization of the protein segments at the interface, maybe again explained by the transition from a globular to a fibrillar structure of the aggregates. Here, it is worth mentioning that these interfacial results were slightly different from those reported on a β-lactoglobulin-enriched WPI, where it was found that aggregates (62) Croguennec, T.; Renault, A.; Bouhallab, S.; Pezennec, S. J. Colloid Interface Sci. 2006, 302, 32. (63) Graham, D. E.; Phillips, M. C. J. Colloid Interface Sci. 1980, 76, 227.
Whey Protein Soluble Aggregates
formed at pH 7.0 in the presence of 100 mM NaCl led to an E′ of 45 mN.m-1 and the formation of a very elastic network.13 However, the heating conditions were less drastic than ours, i.e., 80 °C for 10 min, which might lead to less extended changes in the secondary structure of the whey proteins within the aggregates that might more easily form viscoelastic interfacial networks. These differences highlighted the fact that small differences in the heating conditions in order to obtain WPI could result in large differences in their functionality.5,14,18 The gas permeability of single bubbles formed after 1800 s in a 0.1% dispersion of WPI soluble aggregates was determined in the various test conditions (Table 2). In the absence of sodium chloride and at low pH, the gas permeability was very low (K ) 0.006 cm.s-1) and might be explained by the dense packing of the nonaggregated WPI, as was demonstrated for nondenatured β-lactoglobulin close to its isoelectrical pH (K ) 0.022 cm.s-1).64,65 Upon pH change to 7.0, gas permeability of the single bubble permeability increased to 0.033 cm.s-1, which might be due to increased lateral repulsive interactions within the thin film. Thus, even if the nonaggregate content in the dispersion was similar at pH 6.0 and 7.0, the physicochemical properties of the WPI soluble aggregates were sufficiently different (ζ-potential, charge, thiol activation) to modify gas permeability. For WPI aggregates generated in the presence of NaCl, a general trend in the reduction of the single-bubble gas permeability could be seen as K decreased from 0.048 cm.s-1 at pH 6.0 to 0.006 cm.s-1 at pH 7.0 (Table 2). A marked transition was observed once pH increased above 6.6 with 70 mM NaCl, where gas permeability dropped by a factor of 4. Interestingly, this transition corresponds to a soluble WPI aggregate content of >90% in the dispersion. In parallel, these experimental conditions also corresponded to a sharp decrease in the ζ-potential of the aggregates to around -14 mV but a maximum in surface hydrophobicity index and thiol activation (Table 1). The decreased gas permeability of the single bubble at pH >6.6 might thus be related to a more favorable organization of the protein interfacial layer, owing to the specific physicochemical properties of the soluble aggregates.56 The foam capacity (FC) of the WPI soluble aggregates was determined upon gas sparging to generate a constant foam volume so that foam formation was mainly diffusion-limited (Table 2).9 No significant difference was found for the soluble aggregates generated without salt, FC ≈ 1.22, which could be explained by the constant amount of “native” proteins present in the dispersion which are mainly responsible for the surface tension lowering and the resulting gas incorporation. This result fitted clearly with the similar results found for the surface tension decrease at small time scales (within 0-50 s), which correspond to the time scale of foam formation (about 0-120 s). For the soluble WPI aggregates generated in the presence of NaCl, a clear increase of the FC could be noted upon increase of pH/NaCl content. Thus, aggregates generated at pH 6.0 formed a foam significantly less easily in comparison to those formed at pH >6.6. Here again, results are fairly well correlated with the results obtained for the decrease of surface tension, where it was shown that for pH >6.6, the decrease rate was faster. Figure 8 shows the number of bubbles present in the 38 mm2 of foam surface that was analyzed by image analysis for a set of selected WPI soluble aggregates. Useful information in relation to FC and initial rate of surface tension decrease could be found by looking at the initial number (64) Petkova, V.; Sultanem, C.; Nedyalkov, M.; Benattar, J.; Leser, M. E.; Schmitt, C. Langmuir 2003, 19, 6942. (65) Sultanem, C. Films et bulles de prote´ines solubles: Structure, interactions et perme´abilite´ au gaz. Ph.D. Dissertation. University of Paris XI, Orsay, France, 2004.
Langmuir, Vol. 23, No. 8, 2007 4165
Figure 8. Time evolution of the number of bubbles determined by image analysis in foams stabilized by WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 with or without NaCl. Protein concentration was reduced to 0.1 wt % for foam generation. pH 6.0 (]), pH 6.0 - 5 mM NaCl ([), pH 6.4 (4), pH 6.4 - 40 mM NaCl (2), pH 7.0 (0), pH 7.0 - 120 mM NaCl (9).
of bubbles formed upon stopping gas injection. Without salt, the initial number of bubbles was not very different, about 150 to 180 bubbles. On the contrary, in the presence of NaCl, many more bubbles were created at pH 7.0/120 mM NaCl (400 bubbles) compared to pH 6.4/40 mM NaCl (200 bubbles) or pH 6.0/5 mM NaCl (90 bubbles). It turns out that the foam generated with soluble aggregates at pH 7.0/120 mM contained smaller bubbles than the others, and consequently, a higher amount of liquid could be entrapped in the more numerous thin films and Plateau borders.9,66 Thus, foam produced using WPI soluble aggregates generated at pH 7.0/120 mM NaCl was initially wetter than those generated without NaCl or for lower pH/NaCl combinations. These results on foam capacity somewhat contradict previous results obtained upon membrane foaming of WPI soluble aggregates obtained at 10 wt % protein concentration (pH 7.0) after membrane foaming or mechanical whipping of mixtures of native WPI proteins and aggregates generated at 10 wt %.7,12 Hence, these studies reported an increase in the average air bubble size upon increasing the amount of soluble aggregates in the dispersion higher than 60-80 wt %. These discrepancies in the results really substantiate the importance of the physicochemical properties of the WPI aggregates as well as the experimental conditions used to produce them, and could lead to very different foamability results. The stability of the foams produced has been investigated by the measurement of the time for the foam to drain by 50% of its liquid, the number of bubbles, and the variation of the average air bubble diameter over time. The liquid stability within the foams formed without NaCl was almost constant for all pH values; however, it should be noted that it became more stable upon increasing the pH (see standard deviation values in Table 2). The relatively low foam-liquid stability compared to that of WPI soluble aggregates formed in the presence of salt might be explained by the larger air bubbles present in the foam, reducing (66) Weaire, D.; Hutzler, S. The Physics of Foams; Clarendon Press: Oxford, 1999.
4166 Langmuir, Vol. 23, No. 8, 2007
Figure 9. Time evolution of the mean bubble diameter (d) determined by image analysis in foams stabilized by WPI dispersions at 1 wt % after heat treatment at 85 °C for 15 min at pH ranging from 6.0 to 7.0 with or without NaCl. Protein concentration was reduced to 0.1 wt % for foam generation. pH 6.0 (]), pH 6.0 - 5 mM NaCl ([), pH 6.4 (4), pH 6.4 - 40 mM NaCl (2), pH 7.0 (0), pH 7.0 - 120 mM NaCl (9). Lines represent power-law fit of the experimental data according to d ∼ tR, with d being bubble diameter and R being the coarsening exponent. Regression coefficients R2 are indicated for each set of foam.
the amount of water trapped in the plateau borders but also increasing the drainage rate.67 By contrast, the foam-liquid stability increased tremendously if NaCl was used to form the aggregates, reaching a maximum value of 121 s at pH 6.8/100 mM NaCl. This improved liquid stability of the foam can be related to the smaller air bubble size within the foams or by the properties of the interfacial layers that were formed depending on the physicochemical characteristics of the WPI soluble aggregates. Hence, for pH >6.6, the combination of the low surface hydrophobicity index with the reduced electrostatic charge and the fibrillar shape of the aggregates may have led to the formation of thicker interfacial networks. Even if the dilational viscoleastic modules of these interfaces did not reach the very high values reported in other studies,7,13 this combination of interfacial properties was optimum for liquid entrapment. The variation of the number of air bubbles within the foam was mainly characterized by an exponential decay for all foams, revealing an increase of the air bubble size during foam coarsening (Figure 8). The mean air bubble diameter was plotted against time, and the coarsening exponents were determined (Figure 9). For the very unstable foams at pH 6.0/5 mM, the power-law fit of air bubble coarsening was very poor and the scaling behavior hardly interpretable. However, for more stable foams, the coarsening exponent was between two limiting values of 1/3 and 1/2. Such values have been reported for the two extremes cases of (i) 1/3 for a wet foam composed of spherical gas cells where the coarsening is mainly driven by homogeneous and spherically distributed gas diffusion and can be described by the Lifshitz and Slyozov theory68 and (ii) 1/2 for a dry foam composed of polyhedral gas cells where the coarsening is mainly driven by inhomogeneous and linearly distributed gas diffusion.66,69 Here, it is worth mentioning that the coarsening exponent of 1/3 calculated for the foams stabilized with WPI aggregates formed (67) Saint-Jalmes, A.; Langevin, D. J. Phys.: Condens. Matter 2002, 14, 9397. (68) Lifshitz, I. M.; Slyozov, V. V. J. Phys. Chem. Solids 1961, 19, 35. (69) Durian, D. J.; Weitz, D. A.; Pine, D. J. Phys. ReV. A 1991, 44, R7902.
Schmitt et al.
at pH 6.4 with 40 mM NaCl was different from the value of 1/2 found for the foam formed with WPI soluble aggregates formed at pH 7.0 with 120 mM NaCl. As the liquid stability of the latter foam was higher (98 s vs 68 s), the foam was wetter. Consequently, one would have expected opposite results. Partial explanation for these results might reside in the inhomogeneous distribution of water within the foam formed at pH 7.0/120 mM NaCl, leading to a faster bubble coarsening through linear gas diffusion. Another interesting feature was found for the driest foams (foam liquid stability of less than 30 s; Table 2) stabilized by WPI soluble aggregates generated without salt at pH 6.4 and 7.0. Indeed, the coarsening exponent was found to be 1/5, i.e., lower than predicted values. Very close values of 1/5 have been recently reported in liquid crystal foams.70 The authors explained that the reduced gas loss could be due to the formation of interfacial liquid crystal defects that were able to modify locally the surface tension. In the case of WPI soluble aggregates, one may hypothesize that such a “liquid crystal-like” phase might occur due to the possible local overconcentration in proteins due to drainage. Hence, local organization of WPI soluble aggregate dispersions at concentrations higher than 2.5 wt % has been already described using X-ray scattering,48 but this assumption would require further investigation. Nevertheless, even if it is not possible to definitely draw conclusions on the type of coarsening of the probed foams within the theoretical limiting cases, foam stabilized with WPI soluble aggregates generated in the presence of NaCl at pH >6.6 led to significantly smaller air bubble diameters after a given coarsening time if compared to WPI aggregates generated without salts, which generally matched identical air bubble diameters with NaCl at pH 6.4.
Conclusions The combination of pH adjustment and NaCl addition to WPI allowed the production of soluble aggregates upon heating at 85 °C for 15 min with very specific physicochemical properties and morphologies that cannot be obtained by pH adjustment alone. These soluble WPI aggregates could be formed at yields up to 95% of the initial protein. Aggregates generated at pH >6.6 in the presence of salt were characterized by a fibrillar structure, a reduced ζ-potential, a lower surface hydrophobicity index than the aggregates formed without salt, and a high level of thiol group accessibility. They exhibited a viscoelastic interfacial behavior and a reduced single air bubble gas permeability, which in turn resulted in different foaming and foam stabilization properties. Wet foams with small air bubbles were obtained by combining pH >6.6 and salt content >70 mM with heat treatment. More compact WPI soluble aggregates that were formed at pH
