134
Langmuir 2005, 21, 134-139
Influence of Protein Concentration and Order of Addition on Thermal Stability of β-Lactoglobulin Stabilized n-Hexadecane Oil-in-Water Emulsions at Neutral pH H-J. Kim, E. A. Decker, and D. J. McClements* Biopolymers and Colloids Research Laboratory, Department of Food Science, University of Massachusetts, Amherst, Massachusetts 01003 Received August 5, 2004. In Final Form: October 5, 2004 The influence of protein concentration and order of addition relative to homogenization (before or after) on the extent of droplet flocculation in heat-treated oil-in-water emulsions stabilized by a globular protein were examined using laser diffraction. n-Hexadecane (10 wt%) oil-in-water emulsions (pH 7, 150 mM NaCl) stabilized by β-lactoglobulin (β-Lg) were prepared by three methods: (1) 4 mg/mL β-Lg added before homogenization; (2) 4 mg/mL β-Lg added before homogenization and 6 mg/mL β-Lg added after homogenization; (3) 10 mg/mL β-Lg added before homogenization. The emulsions were then subjected to various isothermal heat treatments (30-95 °C for 20 min), with the 150 mM NaCl being added either before or after heating. Emulsion 1 contained little nonadsorbed protein and exhibited extensive droplet aggregation at all temperatures, which was attributed to the fact that the droplets had a high surface hydrophobicity, e.g., due to exposed oil or extensive protein surface denaturation. Emulsions 2 and 3 contained a significant fraction of nonadsorbed β-Lg. When the NaCl was added before heating, these emulsions were relatively stable to droplet flocculation below a critical holding temperature (75 and 60 °C, respectively) but showed extensive flocculation above this temperature. The stability at low temperatures was attributed to the droplets having a relatively low surface hydrophobicity, e.g., due to complete saturation of the droplet surface with protein or due to more limited surface denaturation. The instability at high temperatures was attributed to thermal denaturation of the adsorbed and nonadsorbed proteins leading to increased hydrophobic interactions between droplets. When the salt was added to Emulsions 2 and 3 after heating, little droplet flocculation was observed at high temperatures, which was attributed to the dominance of intra-membrane over inter-membrane protein-protein interactions. Our data suggests that protein concentration and order of addition have a strong influence on the flocculation stability of proteinstabilized emulsions, which has important implications for the formulation and production of many emulsionbased products.
Introduction Globular proteins are used as emulsifiers in a variety of industries because of their ability to facilitate emulsion formation and improve emulsion stability.1-6 An oil-inwater emulsion is usually formed by homogenizing an oil and aqueous phase together using a mechanical device capable of generating intense pressure and shear gradients within the fluids.1 Globular proteins rapidly accumulate around the oil droplets formed during this process, where they facilitate further droplet disruption by lowering the interfacial tension and retard droplet coalescence within the homogenizer by forming protective membranes around the droplets.2,3,7 After homogenization, the adsorbed proteins contribute to the long-term stability of emulsions by generating repulsive interactions (e.g., steric and electrostatic) that prevent droplets from coming into close contact.7-9 In many industrial applications, it is important that a product is capable of withstanding some form of thermal * Corresponding author. Tel: +1-413-545-1019. Fax: +1-413545-1262. E-mail:
[email protected]. (1) Dickinson, E. Introduction to Food Colloids; Oxford University Press: Oxford, 1992. (2) Damodaran, S. In Food Chemistry, 3rd ed.; Fennema, O. R., Ed.; Marcel Dekker: New York, 1996; p 321. (3) Dalgleish, D. G. In Emulsions and Emulsion Stability; Sjoblom, J., Ed.; Marcel Dekker: New York, 1996. (4) Nakai, S.; Modler, H. W. Food Proteins: Properties and Characterization; VCH Publishers: New York, 1996. (5) Phillips, L. G.; Whitehead, D. M.; Kinsella, J. E. StructureFunction Properties of Food Proteins; Academic Press: San Diego, 1994. (6) Wilde, P. J. Curr. Opin. Colloid Int. 2000, 5, 176-181. (7) Dickinson, E. Colloids Surf., B. 1999, 15, 161-176.
processing after it has been produced, e.g., sterilization, pasteurization, drying, cooking, etc.10,11 Under certain circumstances, globular-protein-stabilized emulsions are susceptible to droplet flocculation when they are heated above a particular temperature.12-17 This phenomenon has been attributed to thermal denaturation of adsorbed globular proteins, which leads to the exposure of amino acid residues originally located within their hydrophobic interiors.18-19 Exposure of certain amino acids, such as those containing nonpolar or sulfhydryl groups, increases (8) Bergensthanl, B. A.; Claesson, P. M. In Food Emulsions, 3rd ed.; Friberg S. E., Larsson, K., Eds.; Marcel Dekker: New York, 1997; pp 57-109. (9) McClements, D. J. Food Emulsions: Principles, Practice and Techniques; CRC Press: Boca Raton, 1999. (10) Sliwinski, E. L.; Lavrijsen, B. W. M.; Vollenbroek, J. M.; van der Stege, H. J.; van Boekel, M. A. J. S.; Wouters, J. T. M. Colloids Surf., B 2003, 31, 219-229. (11) Sliwinski, E. L.; Roubos, P. J.; Zoet, F. D.; van Boekel, M. A. J. S.; Wouters, J. T. M. Colloids Surf., B 2003, 31, 231-242. (12) Hunt, J. A.; Dalgleish, D. G. J. Food Sci. 1995, 60, 1120-1123; 1131. (13) Monahan, F. J.; McClements, D. J.; German, J. B. J. Food Sci. 1996, 61, 504-510. (14) Demetriades, K.; Coupland. J. N.; McClements, D. J. J. Food Sci. 1997, 62, 462-467. (15) Kim, H. J.; Decker, E. A.; McClements, D. J. Langmuir 2002, 18, 7577-7583. (16) Euston, S. R.; Finnigan, S. R.; Hirst, R. L. J. Agr. Food Chem. 2001, 49, 5576-5583. (17) Chen, J. S.; Dickinson, E. Food Hydrocolloids 1999, 13, 363369. (18) Corredig, M.; Dalgleish, D. G. Colloids Surf., B 1995, 4, 411420. (19) Fang, Y.; Dalgleish, D. G. J. Colloid Interface Sci. 1997, 196, 292-298.
10.1021/la048019t CCC: $30.25 © 2005 American Chemical Society Published on Web 11/18/2004
Droplet Flocculation by Protein Surface Denaturation
intermolecular protein interactions through hydrophobic attraction and thiol-disulfide interchange reactions.13,20-23 Intermolecular interactions between protein molecules adsorbed onto different droplets can occur when the droplets are in close proximity during protein thermal denaturation, e.g., when the repulsive interactions are relatively weak or the attractive interactions are relatively strong. Under these circumstances, extensive droplet flocculation may occur,12-16 which is usually undesirable because flocculation can lead to a substantial increase in creaming instability and emulsion viscosity.9 On the other hand, intramolecular interactions between protein molecules adsorbed onto the same droplets dominate when droplets are prevented from coming into close proximity during protein thermal denaturation, e.g., when the repulsive interactions are relatively strong or the attractive interactions are relatively weak.15,24 Under these circumstances, extensive droplet flocculation can be avoided since the protein molecules tend to interact with neighboring protein molecules adsorbed onto the same droplet, rather than on different droplets. Recent studies have shown that the flocculation stability of oil-in-water emulsions stabilized by globular proteins at room temperature and during thermal processing is influenced by the concentration of free protein in the aqueous phase.11,16,25 Nonadsorbed protein could alter emulsion stability by adsorbing to the surface of hydrophobic patches on the droplet surfaces or by inducing bridging or depletion flocculation.9,16,25-26 The molecular conformation and association of the nonadsorbed proteins may also influence the stability of emulsions to droplet flocculation.16,25-27 The conformation and association of nonadsorbed globular proteins may be altered during the homogenization process,28 hence, the time when the proteins are added to an emulsion relative to homogenization may also influence emulsion stability.25 The objective of the present study is to examine the influence of protein concentration and time of addition (before or after homogenization) on the thermal stability of oil-in-water emulsions stabilized by a model globular protein (β-lactoglobulin, β-Lg) at neutral pH. Knowledge obtained from this study could lead to the development of new strategies for designing protein-stabilized emulsions that can be used in a variety of industries, e.g., foods, pharmaceuticals, and cosmetics. Materials and Methods Materials. Analytical grade sodium chloride (NaCl), hydrochloric acid (HCl), sodium hydroxide (NaOH), sodium azide (NaN3), Tween 20, and n-hexadecane were purchased from the Sigma Chemical Company (St. Louis, MO). Powdered β-lactoglobulin was obtained from Davisco Foods International (LOT # JE 001-1-922, Le Sueur, MN). As stated by the manufacturer, the β-Lg content of the powder determined by electrophoresis was 98% (the remainder being mostly ,globulins). The decrease in mass of the protein powder upon drying was 2.6% and the (20) Sawyer, W.H. J. Dairy Sci. 1968, 51, 323-329. (21) De Wit, J. N. J. Dairy Sci. 1990, 73, 3602-3612. (22) Hoffmann, M. A. M.; van Mill, P. J. J. M. J. Agric. Food Chem. 1999, 47, 1898-1905. (23) Kitabatake, N.; Wada, R. Fugita, Y. J. Agric. Food Chem. 2001, 49, 4011-4018. (24) Kim, H. J.; Decker, E. A.; McClements, D. J. J. Ag. Food Chem. 2002, 50, 7131-7137. (25) Kim, H. J.; Decker, E. A.; McClements, D. J. Langmuir 2004, submitted for publication. (26) Dickinson, E.; Golding, M. Colloids Surf., A 1998, 144, 167177. (27) Srinivasan, M.; Singh, H.; Munro, P. A. J. Food Sci. 2001, 66, 441-446. (28) Rampon, V.; Riaublanc, A.; Anton, M.; Genot, C.; McClements, D. J. J. Agric. Food Chem. 2003, 51, 5900-5905.
Langmuir, Vol. 21, No. 1, 2005 135 nitrogen content of the powder was 15.6%. Distilled and deionized water was used for the preparation of all solutions. Solution Preparation. Emulsifier solutions were prepared by dispersing powdered β-Lg into 5 mM sodium phosphate buffer solution at pH 7 containing 0.04 wt% NaN3 (as an antimicrobial agent) and stirring for at least 2 h to ensure complete dispersion. Emulsion Preparation. An oil-in-water emulsion was prepared by homogenizing 20 wt% n-hexadecane oil and 80 wt% emulsifier solution (8 or 20 mg/mL β-Lg in 5 mM phosphate buffer, pH 7.0, 0 mM NaCl) at room temperature. The oil and emulsifier solution were blended using a high-speed blender for 2 min (Model 33BL79, Warring, Inc., New Hartford, CT) and then passed through a high-pressure valve homogenizer five times at 7500 psi (Rannie High Pressure, APV-Gaulin, Model MiniLab 8.30H, Wilmington, MA). The emulsions were diluted to 10 wt% oil with aqueous solutions containing salt and/or protein to give final aqueous phase concentrations of 0 or 150 mM NaCl, and 4-10 mg/mL total protein. Dilution was carried out either before or after heat treatment. Heat treatment involved placing emulsions in glass test tubes and holding them isothermally at selected temperatures (ranging from 30 to 95 °C) for 20 min then placing them in water for 20 min to cool them to room temperature. Emulsions containing added salt (150 mM NaCl) exhibited some creaming during heating at the higher temperatures (which suggested that some flocculation occurred), but all the other emulsions were stable to creaming during heating. Emulsions were then stored in a temperature-controlled water bath at 30 °C for 24 h, after which their particle size was measured. The final pH of the emulsions was measured to be 7.0 ( 0.2 (pH Meter 320, Corning, Inc., Corning, NY). In summary, three different 10 wt% n-hexadecane oil-in-water emulsions were prepared: • Emulsion 1 was prepared by homogenizing 20 wt% oil with 80 wt% emulsifier solution containing 8 mg/mL β-Lg. The emulsion was then diluted with buffer solution to give a final composition of 10 wt% oil, 4 mg/mL β-Lg (added before homogenization), and 0 or 150 mM NaCl with a buffer solution. • Emulsion 2 was prepared by homogenizing 20 wt% oil with 80 wt% emulsifier solution containing 8 mg/mL β-Lg. The emulsion was then diluted with a buffered protein solution to give a final composition of 10 wt% oil, 4 mg/mL β-Lg added before homogenization, 6 mg/mL β-Lg added after homogenization, and either 0 or 150 mM NaCl. • Emulsion 3 was prepared by homogenizing 20 wt% oil with 80 wt% emulsifier solution containing 20 mg/mL β-Lg. The emulsion was then diluted with buffer solution to give a final composition of 10 wt% oil, 10 mg/mL β-Lg (added before homogenization), and 0 or 150 mM NaCl. The protein concentration used in Emulsion 1 was just sufficient to cover the droplet surfaces produced by homogenization, leaving little free protein in the continuous phase (see below). The total protein concentrations in Emulsions 2 and 3 were similar, but part of the protein was added after homogenization in Emulsion 2, whereas all of the protein was added before homogenization in Emulsion 3. The salt (0 or 150 mM NaCl) was added to each type of emulsion either before or after heating since salt concentration and order of addition relative to thermal processing have previously been found to have a large impact on the flocculation stability of whey-protein-stabilized emulsions.15 Particle Size Determination. The particle size distribution of the emulsions was measured using a laser diffraction instrument (LS230, Coulter Corporation, FL). This instrument measures the angular dependence of the intensity of light scattered from a stirred dilute emulsion and then indicates the particle size distribution that gives the closest fit between theoretical calculations and experimental measurements. A refractive index ratio of 1.08 was used in the particle size calculations. Emulsions were vortexed for 3 s, inverted, and then vortexed for another 3 s to ensure that they were homogeneous. They were then diluted (∼1:1000) with pH-adjusted distilled water (pH 7) prior to making the particle size measurements to avoid multiple scattering effects. The emulsions were stirred continuously (at an instrument stirring speed of 50%) throughout the measurements to ensure the samples were homogeneous. Vortexing, dilution, and stirring may have partially disrupted weakly flocculated droplets, although it is unlikely that they will have disrupted any strongly
136
Langmuir, Vol. 21, No. 1, 2005
Kim et al.
flocculated droplets. The theory used to calculate the particle size distribution assumes that the particles are spherical and homogeneous, and therefore the data obtained on emulsions that contained flocs should be treated with caution because they are nonspherical and nonhomogeneous. Particle size measurements are reported as weight-average mean diameters, d43 ()Σnidi4/ Σnidi3, where ni is the number of particles with diameter di). Particle diameters are reported as the average of measurements made on at least two freshly prepared samples, with standard deviations being less than 5% for nonflocculated emulsions and less than 15% for flocculated emulsions. Free Protein and Surface Load Measurements. The free protein concentration in the aqueous phase of the emulsions was determined using a modified Lowry method.29 The aqueous phase was collected from the emulsions by placing them in Eppendorf tubes and centrifuging them for 15 min using a benchtop centrifuge. After centrifugation, the emulsions separated into a creamed layer and a serum layer, and the serum layer was carefully collected using a syringe and placed into a fresh Eppendorf tube. After recentrifugation, the serum layer was collected and filtered using a syringe filter (0.2 µm pore size, Regenerated, Alltech, IL). A portion of the serum phase was then diluted with distilled water (1:2 or 1:40) to obtain an absorbance in the range of the calibration curve. The protein concentrations in the serum phases were determined using a calibration curve prepared using β-Lg solutions of known concentration. The concentration of adsorbed protein, expressed as the surface load (Γ in mg/m2), was calculated using the following expression:9
Γ)
Cad32 6φ
(1)
where Ca is the mass of emulsifier adsorbed to the surface of the droplets per unit volume of emulsion (mg/m3), which is equal to the initial emulsifier concentration minus that remaining in the aqueous phase after homogenization, φ is the disperse phase volume fraction, and d32 (m) is the volume-surface mean droplet diameter (measured for nonflocculated droplets, e.g., by adding surfactant to disrupt flocs before making light scattering measurements).
Results and Discussion Initial Free Protein Concentration and Surface Load. In the absence of salt, the free protein concentration in the emulsions after storage at room temperature for 24 h was 0.06 ( 0.03, 3.6 ( 0.05, and 1.4 ( 0.2 mg/mL for Emulsions 1, 2, and 3, respectively. The measured initial mean droplet diameters of the same three emulsions were d32 ) 0.20, 0.20, and 0.14 ( 0.01 µm, respectively. The smaller initial droplet diameter in Emulsion 3 was presumably because there was a higher protein concentration in the aqueous phase during homogenization, so that droplet disruption and stabilization were facilitated.9 The protein surface loads (Γ) calculated from the free protein concentration and mean droplet diameter values are 1.04 ( 0.02, 1.63 ( 0.02, and 1.62 ( 0.03 mg/m2, for Emulsions 1, 2, and 3, respectively. These values are in good agreement with previously reported values for whey proteins.30-31 The fact that the surface load in Emulsion 3 was appreciably more than that in Emulsion 1 suggested that the droplets were covered by a thicker and/or denser interfacial layer. The tendency for globular proteins to form thicker interfacial membranes when present in the continuous phase at higher concentrations prior to homogenization has been reported recently.31 The fact that the surface load in Emulsion 2 was significantly higher (29) Markwell, M. A.; Haas, S. M.; Bieber, L. L.; Tolbert, N. E. Anal. Biochem. 1978, 87, 206-210. (30) Tcholakova, S.; Denkov, N. D.; Ivanov, I. B.; Campbell, B. Langmuir 2002, 18, 8960-8971. (31) Tcholakova, S.; Denkov, N. D.; Ivanov, I. B.; Campbell, B, Langmuir 2003, 19, 5640-5649.
than that in Emulsion 1 suggests that some of the native protein added to the emulsions after homogenization adsorbed to the droplet surfaces, either by binding directly to exposed patches of oil or by binding to already adsorbed proteins.30 Droplet Aggregation in Emulsions Containing No Salt. In the absence of NaCl, each type of emulsion exhibited very similar behavior regardless of whether it was heated first and then diluted with buffer or diluted with buffer first and then heated (compare parts a and b of Figure 1). We will therefore discuss the behavior of the emulsions containing no salt together. At 30 °C, the emulsions containing 4 mg/mL β-Lg added before homogenization (Emulsions 1 and 2) had a slightly larger mean diameter than those containing 10 mg/mL β-Lg added before homogenization (Emulsion 3), presumably because there was less protein present to effectively cover all of the oil-water interfaces created in the homogenizer.9,30 In the absence of salt, Emulsions 2 and 3 were stable to thermal processing across the entire temperature range studied. This stability to droplet aggregation can be attributed to the fact that the droplets had a relatively low surface hydrophobicity (which would have generated a hydrophobic attraction between the droplets) and that there was a strong electrostatic repulsion between the droplets due to the low ionic strength. We postulate that the droplets in Emulsions 2 and 3 had a low surface hydrophobicity because there was sufficient free protein present to adsorb to the droplet surfaces and cover any nonpolar patches (e.g., exposed oil or unfolded proteins). On the other hand, Emulsion 1 was only stable to thermal processing at holding temperatures below 70 °C, after which there was a pronounced increase in mean particle diameter, suggesting that at least some of the droplets became aggregated at higher temperatures. Adsorbed β-Lg is known to have a thermal transition temperature (Tm) around 70 °C,19 which is believed to increase the number of exposed nonpolar and sulfhydryl groups at the droplet surfaces.15 We propose that in Emulsion 1 there was either insufficient protein present to completely cover the droplet surfaces (leading to exposed patches of oil) or their was more-extensive surface denaturation of the adsorbed proteins (leading to a greater exposure of nonpolar amino acid side groups). As a result, there was an appreciable hydrophobic attraction between the droplets.9 At low ionic strengths, this hydrophobic attraction was not sufficient to overcome the strong electrostatic repulsion between the droplets. Nevertheless, once the emulsion was heated above the thermal denaturation temperature of the adsorbed proteins, they became extensively unfolded, thereby increasing the droplet surface hydrophobicity. The strength of the hydrophobic attraction between the droplets was then sufficiently strong to at least partly overcome the electrostatic repulsion. Consequently, the rate of droplet aggregation was increased at holding temperatures exceeding the thermal denaturation temperature of the adsorbed proteins. Measurements of the free protein concentration and surface load after thermal treatment of the emulsions containing no salt are shown in Figures 3-6. In Emulsion 1, the free protein concentration (Tm when NaCl was added after heating (∼14 µm at 30 °C and 300 µm at 95 °C). Emulsions 2 and 3 were relatively stable to droplet aggregation across the whole range of holding temperatures studied when salt was added after heating (Figure 2b), in contrast to the large degree of droplet aggregation observed at higher temperatures when the salt was added before heating (Figure 2a). The observed differences in the thermal stabilities of the emulsions depending on the order of addition of the salt relative to the thermal treatment were similar to those observed in our previous study.15 When the emulsion droplets are heated in the absence of salt, there is a relatively large electrostatic repulsion between them, so
Droplet Flocculation by Protein Surface Denaturation
Langmuir, Vol. 21, No. 1, 2005 139
the higher NaCl concentration for the reasons stated above. Nevertheless, in Emulsion 2 it appeared that the amount of protein that adsorbed to the droplet surfaces above Tm was less when the salt was added to the emulsions after heating, rather than before heating (Figures 5 and 6). It is possible that there was a relatively strong electrostatic repulsion between the free protein molecules and the adsorbed protein molecules when the emulsions were heated in the absence of salt, which may have retarded adsorption during thermal processing. Consequently, the adsorbed proteins may have interacted with their neighbors, reducing the number of active nonpolar and sulfhydryl groups available for reaction. Conclusions Figure 6. Influence of holding temperature (30-95 °C, 20 min) on the surface load when 0 or 150mM NaCl added after heating.
that their surfaces cannot come into close proximity. Consequently, when adsorbed proteins become thermally denatured, they tend to interact with neighboring proteins adsorbed onto the same droplet (or free proteins), rather than with protein molecules adsorbed onto different droplets. This leads to the formation of an interfacial membrane containing proteins that are extensively crosslinked by hydrophobic and disulfide bonds, and there are few free nonpolar and sulfhydryl groups remaining at the outer surface of the membrane capable of forming bonds with proteins on other droplets. On the other hand, when emulsions are heated in the presence of salt, the electrostatic interactions are strongly screened, which means that the droplets can come into close proximity during heating. Consequently, extensive interactions can occur between proteins adsorbed onto different emulsion droplets, which leads to appreciable droplet aggregation. The nonadsorbed protein concentrations and surface loads measured for Emulsion 1 were similar in the absence and presence of NaCl (Figures 3 to 6), i.e., there was very little free protein present and the surface load (Γ ) 1.04 ( 0.02 mg/m2) remained constant at all holding temperatures. The dependencies of the free protein concentrations and surface loads of Emulsions 2 and 3 on holding temperature when salt was added after heating (Figures 4 and 6) followed the same general trends as observed when salt was added before heating (Figures 3 and 5). In general, the free protein concentration decreased and the surface load increased with increasing holding temperature, and the amount of adsorbed protein was greater at
We have shown that the concentration of free protein in protein-stabilized oil-in-water emulsions has a major impact on their stability to droplet aggregation during thermal processing. The time when the protein was added to the emulsion relative to homogenization was also important. We found that the emulsions were highly unstable to droplet aggregation at all holding temperatures (30-95 °C) when the initial protein concentration was just sufficient to cover all of the droplets formed during homogenization, so that there was little free protein present in the continuous phase. On the other hand, we found that emulsions containing excess protein in the aqueous phase had much better stability to droplet aggregation during thermal treatment, particularly if the protein was added after, rather than before, heat treatment. These results may have important consequences for the development of protein-stabilized emulsions with improved stability to thermal treatment. We are currently examining alternative strategies for improving the thermal stability of protein-stabilized emulsions in our laboratory, e.g., by increasing the repulsion between the droplets by adsorbing a layer of polysaccharide to the surfaces of protein-coated emulsion droplets. Acknowledgment. This material is based upon work supported by the Cooperative State Research, Extension, Education Service, United States Department of Agriculture, Massachusetts Agricultural Experiment Station (Project No. 831) and by a United States Department of Agriculture, CREES, IFAFS Grant (Award No. 20014526). We also thank Davisco Foods International for kindly donating the protein ingredients used in this study. LA048019T
