Langmuir 2005, 21, 5931-5939
5931
Effect of r-Lactalbumin on the Phase Behavior of AOT-Brine-Isooctane Mixtures: Role of Charge Interactions Justin W. Shimek,‡ Catherine M. Rohloff,† Jessica Goldberg,‡ and Stephanie R. Dungan*,†,‡ Department of Chemical Engineering and Materials Science and Department of Food Science and Technology, University of California, One Shields Avenue, Davis, California 95616 Received October 14, 2004. In Final Form: February 28, 2005 We have found that both electrostatic and hydrophobic interactions are involved in the ability of the protein R-lactalbumin (R-LA) to affect the self-assembly of the anionic surfactant sodium bis(ethylhexyl) sulfosuccinate (AOT, 3.5 wt %) in equivolume mixtures of organic and aqueous solutions. The composition and size of AOT phase structures that form in the presence of 0.35 wt % protein were evaluated as a function of pH and ionic strength. In the absence of protein, AOT forms water-in-oil microemulsion droplets for all pH and salt concentrations studied here. The presence of the protein in the water-in-oil microemulsion phase boosts water solubilization and droplet size, as the spontaneous curvature of the surfactant interface becomes less negative. Aggregates of protein, surfactant, and oil also form in the water-continuous phase. The size and composition of structures in both phases can be tuned in the presence of protein by varying the pH and ionic strength. R-LA induces the appearance of an anisotropic surfactant phase at pH 5.8 reached equilibrium within 24 h, and their properties did not change over several months of storage. Samples at somewhat lower pH exhibited a distinctly different behavior, forming a clear upper organic phase in contact (32) Rohloff, C. M. Ph.D. Dissertation. University of California, Davis, 2001. (33) Johnson, W. C. Biochemistry 1981, 20, 1085. (34) Compton, L. A.; Johnson, W. C. Anal. Biochem. 1986, 155, 155.
Langmuir, Vol. 21, No. 13, 2005 5933
Figure 1. Appearance of R-lactalbumin-AOT-brine-isooctane microemulsions at equilibrium. Each two-phase system contains 0.1 mol of AOT per L of isooctane, 0.1 mol of NaCl per L of water, and 7 g of protein per L of water. pH varies as indicated.
Figure 2. Concentration of protein (b) and water (9) in the upper (organic) phase; protein concentration in the lower (aqueous) phase (O). Conditions same as in Figure 1.
with a lower aqueous phase with a bluish iridescent quality (Figure 1). This equilibrium state evolved slowly. After 24 h the upper organic phase was completely turbid while the lower aqueous phase was clear. At 72 h the top of the w/o microemulsion phase began to clear, an opaque middle phase formed between the bulk organic and aqueous phases, and another turbid layer appeared at the bottom of the vial. The middle phase diminished over a period of weeks as the lower layer grew. At 1 month changes in visual appearance and composition ceased. The composition and structures in the upper and lower phases shown in Figure 1 represent systems at the twophase boundary in the overall system phase diagram. Their compositions, which we measure experimentally, are connected by tielines on that boundary. They do not characterize the existence or structure of single phases outside the two-phase envelope. The concentration of R-LA in the upper and lower phases and the water concentration in the organic phase were measured at various pH values (Figure 2). The amount of protein solubilized in the upper, w/o microemulsion phase increased monotonically as the pH decreased to pH 5.8. The amount of protein solubilized in this microemulsion phase increased from 0.09 g/L (less than 2% of the total protein) at pH 11.2 to 5.84 g/L R-LA (95% of total protein) at pH 5.8. Below pH 5.8 there was a dramatic decrease of the total protein in the w/o microemulsion phase, with protein emerging in the lower aqueous phase. The protein content of the aqueous phase complemented that of the organic phase: mass balance closures were within 5% at all conditions, and we conclude that all the protein was contained in either the organic or aqueous phase. No third phase such as a solid precipitate was formed at any pH.
5934
Langmuir, Vol. 21, No. 13, 2005
Figure 3. Solubilization of AOT (2) and isooctane ([) in the lower (aqueous) phase as a function of pH. Conditions same as in Figure 1.
Figure 2 shows the water content of the upper organic phase. The water content of this w/o microemulsion followed qualitatively the organic concentration of protein as a function of pH. At pH 11.2, where almost no protein was found in the organic phase, water solubility in the w/o microemulsion was 85% of that in the microemulsion in the absence of protein. As the pH decreased, the water content increased until a maximum was reached at pH 5.8. At this condition the water content of the w/o microemulsion was approximately double that of an AOT w/o microemulsion without R-LA. At pH values below 5.8, the w/o microemulsion water content was significantly decreased. The concentration of R-LA in the aqueous phase at various pH values affected the surfactant and oil in that lower phase. In the absence of protein, AOT and isooctane are found almost exclusively in the upper, organic phase; their solubilities (in the absence of protein) in 0.1 M NaCl solution are 2 × 10-4 35 and 2 × 10-5 M,36 respectively. Figure 3 shows the concentration of AOT and isooctane in the lower phase as a function of pH in the presence of protein. AOT concentration in this water phase was 0.06 M at pH 11.2. At this pH no isooctane was present in the lower phase. As the pH decreased, the aqueous AOT composition increased and then plateaued for pH 6.69.2. At lower pH values AOT in the water decreased sharply at pH 5.8, and then again partitioned strongly to the water-continuous phase at more acidic pH values. Aqueous concentrations of isooctane were qualitatively similar to that of AOT (Figure 3). The significant quantities of isooctane (1-2 orders of magnitude greater than the solubility limit in water) solubilized within the aqueous phase indicate that that phase contains hydrophobic nanodomains such as those found in micelles or in o/w microemulsion droplets. Figure 4 presents the ratio of isooctane to AOT in the aqueous phase on a mole per mole basis. As the pH decreased, the number of isooctane molecules solubilized per AOT molecule present in the o/w microemulsion phase generally increased. It appears that the o/w microemulsion aggregates swell with oil and range from one to three oil molecules per AOT molecule. Since normal micelles of anionic, water-soluble surfactants typically solubilize far fewer than one molecule of oil per surfactant,37 it appears that the structures observed in our system are similar to oil-in-water microemulsion assemblies. At the lowest pH (35) Aveyard, R.; Binks, B. P.; Clark, S.; Mead, J. J. Chem. Soc., Faraday Trans. 1 1986, 82, 125. (36) McAuliffe, C. J. Phys. Chem. 1966, 70, 1267. (37) Chaiko, M. A.; Nagarajan, R.; Ruckenstein, E. J. Colloid Interface Sci. 1984, 99, 168.
Shimek et al.
Figure 4. Mole ratio of isooctane to AOT in the lower aqueous phase as a function of pH.
Figure 5. Correlation lengths of w/o microemulsion droplets (b) or o/w aggregate structures ([) as a function of pH. Estimated hydrodynamic radii (O) in the w/o microemulsion phase are also shown at pH 6.2 and 9.2. Conditions same as in Figure 1.
values, where the phase behavior of the aqueous phase shifted to a new type of structure, the amount of oil solubilized in the water-continuous phase was lower (∼1 mol of oil per mol of surfactant). Effect of pH on Phase Structures. Dynamic light scattering measurements were able to detect phase structures within the aqueous and organic phases for most of the pH range studied. Results for the effective correlation length ξ of these structures as a function of pH are presented in Figure 5. The correlation lengths were obtained by substituting the measured diffusion coefficient D into the Stokes-Einstein equation D ) kT/(6πξη). The viscosities used in this equation were η ) 0.47 cP for the isooctane medium and 0.89 cP for the aqueous medium. Since the results are presented at finite concentration, correlation lengths include information on the size as well as any interparticle interactions that may be in effect. For these two-phase equilibrium studies, it is not possible to vary directly and independently the number concentration of the particles in light scattering experiments, and thus to quantify separately interparticle interactions. However, by separating the organic phase and diluting it successively with isooctane, we obtained an estimate for Do, the diffusion coefficient at infinite dilution, at two different pH values.38 Values for the hydrodynamic radius from this estimate are given by the open circles in Figure 5. At the most alkaline pH, where the protein was found primarily in the aqueous phase (Figure 2), the correlation length of the droplets in the organic phase corresponded (38) Shimek, J. W. Ph.D. Dissertation. University of California, Davis, 2003; pp 76-77.
Charge Interactions in R-LA-AOT Phase Behavior
Langmuir, Vol. 21, No. 13, 2005 5935
Figure 6. Concentration of protein (b) in the upper (organic) phase; protein (O), AOT (0), and isooctane (4) concentration in the lower (aqueous) phase. Each two-phase system contains 0.1 mol of AOT per L of isooctane and 7 g of protein per L of water.
Figure 7. Correlation lengths of w/o microemulsion droplets containing R-lactalbumin (b) or without protein (O) as a function of sodium chloride concentration. Correlation lengths for aqueous phase aggregates ([) also shown. Conditions same as in Figure 6.
closely to the radius of AOT microemulsion droplets formed in the absence of R-LA (11 nm).14 As the pH decreased, correlation lengths in the w/o microemulsion increased. Comparison of Figures 2 and 5 shows that, as pH decreased in the range 5.8 < pH < 11.2, protein concentration in the organic phase went up, the w/o droplets increased in correlation length, and they took up more water. In the aqueous phase, light scattering measurements of diffusion coefficients at various volume fractions yielded values of D that were independent of concentration. Thus, the effect of interparticle interactions on the reported correlation lengths in this phase can safely be neglected. At the most basic pH values, the size of the aqueous structures was at a minimum, corresponding to a radius of approximately 3.3 nm (Figure 5). This radius is larger than the Stokes radius of approximately 2 nm that has previously been reported for R-LA.39 At pH 11.2, 12 molecules of AOT and no detectable isooctane were measured per molecule of R-LA in the aqueous phase (Figure 3). Thus, particles consisting of protein decorated with 12 AOT molecules seem likely at this condition. As the pH decreased, the size of the aggregates increased and isooctane was incorporated into the aggregate in increasing proportions (Figures 4 and 5), as oil-in-water assemblies formed. Effect of Ionic Strength on Phase Behavior. At low (0.8 M (2.6 wt %)) salt concentrations, the system appeared visually as two clear immiscible liquids. At intermediate salt concentrations (1.3-2.6%, or 0.4-0.8 M NaCl) the formation of a third, transparent phase was visually apparent. This third phase was more viscous than isooctane or water and held its shape under the force of gravity, similar to a gel. It formed initially at the bulk oil-water interface, but eventually fell to the bottom of the aqueous phase, presumably due to a very slightly greater density than brine. Reports of nonthermal protein-surfactant gels are not common in the literature: More´n and Khan state that the formation of the SDS-lysozyme gel they observed in water was previously unknown.40 Concentrations of R-LA in the organic and aqueous phases were measured as the concentration of sodium chloride was increased (Figure 6). R-LA concentration in the organic phase was only a weak function of ionic strength. Protein concentration in the aqueous phase was, however, greatly affected by changes in ionic strength,
largely due to the emergence of the third, viscous phase. As the aqueous sodium chloride concentration was increased from 0.1 to 0.4 M (0.3 to 1.3 wt %), the concentration of R-LA in the aqueous phase decreased substantially, as more than half of the protein went into the viscous third phase. Increasing the sodium chloride concentration to 1.0 M caused the aqueous protein content to rise to 2 g/L, as the third phase disappeared. The quantity of AOT and isooctane in the aqueous phase was highest at low salt concentrations (Figure 6), and then decreased sharply as the salt concentration approached 0.4 M (1.3 wt %) from below. At higher salt concentrations corresponding to the formation of the third viscous phase, the aqueous phase contained constant and low levels of AOT (∼0.001 M) and isooctane (∼0.003 M). Measurements in both the organic and aqueous phases indicated that, as salt concentration increased, AOT and isooctane moved from the aqueous phase into the viscous phase. At the highest salt concentrations ([NaCl] g 0.8 M (2 wt %)), surfactant and oil moved from the third viscous phase into the organic phase, as the former disappeared. At salt concentrations of 0.4 M (1.3 wt %) and 0.6 M (2.0 wt %), we estimate that 55-60% of the protein and 1014% of the AOT were found in the third, viscous phase. We estimate that the viscous phase contained approximately 1 mol of isooctane per mol of surfactant. Effect of Ionic Strength on Phase Structures. Dynamic light scattering was used to explore the effect of R-LA on the size of w/o microemulsion droplets formed in the organic phase over a range of ionic strengths. The size of the droplets is expected to decrease with increasing ionic strength due to screening of repulsive electrostatic forces between charged surfactant headgroups, causing a decrease in the spontaneous curvature.35,41 This was indeed the case in both the presence and absence of protein (Figure 7); the latter was in good agreement with previous results.42,43 More significantly, for all ionic strengths the presence of protein caused the droplets to increase in correlation length (Figure 7), as well as to take up more water (data not shown). The assemblies in the aqueous phase were also found to be strongly dependent on the concentration of sodium chloride (Figure 7). The size of the aqueous aggregates grew rapidly with increasing salt as the boundary for
(39) Gast, K.; Zirwer, D.; Muller-Frohne, M.; Damaschun, G. Protein Sci. 1998, 7, 2004. (40) More´n, A. K.; Khan, A. Langmuir 1995, 11, 3636.
(41) Kellay, H.; Meunier, J.; Binks, B. P. Phys. Rev. Lett. 1993, 70, 1485. (42) Rahaman, R. S.; Hatton, T. A. J. Phys. Chem. 1991, 95, 1799. (43) Shioi, A.; Harada, M.; Obika, M.; Adachi, M. Langmuir 1998, 14, 5790.
5936
Langmuir, Vol. 21, No. 13, 2005
Figure 8. Lower, birefringent phase at pH 5.4 at 10× magnification. Overall system contains 0.1 mol of AOT per L of isooctane, 7 g of protein per L of water, and 0.1 mol of NaCl per L of water.
formation of the third viscous phase was approached. No complexes were detected in the aqueous phase for sodium chloride concentrations between 0.25 and 1.0 M (0.833.3%). However, at 1.0 M NaCl, particles of 2.5 nm radius were measured. This size is similar to, but somewhat larger than, the hydrodynamic radius of R-LA.39 Optical Properties. The optical properties of the oilcontinuous and water-continuous phases were evaluated by examining the sample visually between crossed polarizers and through a polarizing microscope. The aqueous phase at low pH (
