- Email: [email protected]
Electrophoretic Mobilities of Protein-Coated Hexadecane Droplets at Different pH
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
205, 185–190 (1998)
CS985669
Electrophoretic Mobilities of Protein-Coated Hexadecane Droplets at Different pH H. C. van der Mei,1 S. Meijer, and H. J. Busscher Materia Technica, University of Groningen, Bloemsingel 10, 9712 KZ Groningen, The Netherlands Received March 17, 1998; accepted May 18, 1998
Electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate solutions (pH between 3.0 and 7.0) and in phosphate buffered saline (pH 7.0) were measured during adsorption of bovine (BSA) and human (HSA) serum albumin, immunoglobulin G (IgG), and fibrinogen (Fg) from single-protein solutions as well as during protein adsorption from binary HSA/IgG and HSA/Fg mixtures and from diluted plasma. Electrophoretic mobilities became less negative upon adsorption of proteins within 1.5 min after the initiation of adsorption. Only for IgG, was a timedependent change of the electrophoretic mobilities of the protein– hexadecane complex observed. In phosphate buffered saline, less negative electrophoretic mobilities were measured than in the potassium phosphate solution. Iso-electric points of the protein– hexadecane complexes in 10 mM potassium phosphate were located at pH 5.0 for albumin, at pH 5.5 for Fg, and at pH 6.6 for IgG, i.e., about the iso-electric points of the pure unadsorbed proteins. This confirms that the net charge addition upon protein adsorption, which is positive below the iso-electric point of the proteins, at low protein concentrations determines the effects on the final electrophoretic mobilities of the protein– hexadecane complexes. As a name for the methodology applied, we propose PATH (protein adsorption to hydrocarbons), in analogy to the well-known MATH (microbial adhesion to hydrocarbons) method. The major advantage of PATH is that it represents an in situ method to study protein adsorption, without artifactual rinsing steps, while furthermore the hydrocarbon phase can be replaced by organic solvents to study the role of acid– base interactions in protein adsorption. In combination with drop-shape analysis techniques, PATH also enables us to determine in situ effects of protein adsorption on interfacial tensions. © 1998 Academic Press Key Words: protein adsorption; electrophoretic mobility; hexadecane.
INTRODUCTION
Protein adsorption is extensively studied both from an applied as well as from a fundamental point of view (1). Adsorption of structurally stable proteins is governed by the hydrophobicity of the sorbent surface and electrostatic interactions, whereas less stable proteins may show unexpected adsorption behavior due to their tendency to unfold upon adsorption (1, 2). 1
To whom correspondence should be addressed.
Adsorption of proteins to sorbent surfaces is still frequently studied by methods that require removal of “loosely-bound proteins” by rinsing, dipping, or other washing steps, although it is realized more and more that in situ methods can be used with a greater accuracy to study protein adsorption as they avoid these irreproducible steps. Frequently employed in situ methods already include ellipsometry and other reflectometric methods (3, 4), streaming potential measurements (5), and axisymmetric drop shape analysis (6 – 8). Often adsorption of proteins is accompanied by changes in the surface free energy and electrostatic properties of the sorbent surface. Contact angle measurements on dried sorbent surfaces after protein adsorption demonstrated that surface free energies of different materials converged to one value (9, 10). The rinsing and drying steps in such measurements are admittedly hard to avoid, but nevertheless make the reliability of the results subject to debate (11). Changes in electrostatic properties during protein adsorption have been assessed with in situ streaming potential measurements (5). Recently, axisymmetric drop shape analysis by profile (ADSA-P) (6) has been employed to determine the changes in surface tension and contact angle of a protein solution droplet on a sorbent surface. Subsequently, without any washing steps or controversial assumptions regarding the surface thermodynamics involved, the surface tension and contact angle measured can be inserted in the Young equation to yield the solid–liquid interfacial tension. Analyses of the solid–liquid interfacial tension changes over time during adsorption of proteins (12) indicated that the major decrease in surface energetics is accomplished by adsorption of proteins, while conformational changes of the adsorbed protein have less influence (13, 14). By far all studies on protein adsorption have been carried out on real-life surfaces, under the assumption that they are ideally smooth and homogeneous. Clearly, all surfaces are rough at the nanometer-scale and chemical heterogeneities are impossible to avoid. Protein adsorption on hydrophobicity gradient surfaces, obtained for instance by silane diffusion, is strongly influenced by the chemical heterogeneity existing along the length of the gradient, and detergent-induced desorption of albumin is most difficult from the heterogeneous transition region and not from the ends of the gradient with extreme
185
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
186
VAN DER MEI, MEIJER, AND BUSSCHER
hydrophobicities (15). Hydrocarbon interfaces against an aqueous solution represent an almost ideally smooth and homogeneous surface for the study of biological interactions at an interface, and the use of such systems has been pioneered for microbial adhesion (16). The hydrocarbon interface is hydrophobic due to its inability to exert acid– base interactions (17) and is negatively charged above pH 3 (18), as established by particulate microelectrophoresis on hydrocarbon emulsions (19). In fact, microbial adhesion to hydrocarbons (MATH) has yielded an elegant description of microbial adhesion in terms of the DLVO-approach, with only hydrophobic microorganisms with little negative surface charge adhering well to the hydrocarbon interface (20 –22). Bellon-Fontaine et al. (23) has demonstrated the involvement of acid– base interactions in microbial adhesion by comparing adhesion to hydrocarbons with microbial adhesion to organic solvents (MATS), able to exert acid– base interactions. Cheng et al. (24) used the ideal properties of hydrocarbons in aqueous solutions to study protein adsorption (PATH). By analyzing the shape of a pendant decane droplet in a 20 mM Tris buffer (pH 5.6) solution of human serum albumin, “equilibrium” interfacial tensions were obtained. The interfacial tension remained almost constant at approximately 52 mJ/m2 for the two lowest (0.0001 and 0.001 mg/ml) albumin concentrations, implying a negative interfacial pressure of 2 mJ/m2. With further increases in the bulk protein concentration, the interfacial tension dropped sharply to 20 mJ/m2 and remained constant at 20 mJ/m2 at protein concentrations above 0.05 mg/ml. Measurements of the interfacial tension as a function of pH for the lowest concentrations suggested that the negative interfacial pressures found were caused by electrostatic repulsion between the negatively charged hydrocarbon interface and the albumin. The aim of this study is to determine the electrophoretic mobility changes upon protein adsorption to hexadecane (PATH) during particulate microelectrophoresis. All experiments were carried out in 10 mM potassium phosphate solutions with pH adjusted to vary between 3 and 7. Both bovine (BSA) and human (HSA) serum albumin as well as immunoglobulin G (IgG) and fibrinogen (Fg) were used as singleprotein solutions. Experiments were also done with diluted human plasma and binary protein mixtures of HSA/IgG and of HSA/Fg. MATERIALS AND METHODS
Proteins The proteins used in this study were Bovine serum albumin (BSA, 96 –99% albumin, Fraction V), Human serum albumin (HSA, 96 –99% albumin, Fraction V), Immunoglobulin G (IgG, human), and Fibrinogen (Fg, from human plasma, fraction I, type III). All proteins were obtained from Sigma Chemical Company, St. Louis, and used as received. Furthermore,
pooled human plasma anticoagulated with sodium citrate was used. Preparation of Hexadecane Emulsions Hexadecane employed in this study was obtained from Merck (Darmstadt, Germany) and was of highest purity grade. Emulsification in 10 mM potassium phosphate (0.87 g/liter K2HPO4 and 0.68 g/liter KH2PO4) solutions with pH adjusted to vary between 3 to 7 through the addition of HCl or KOH was done by sonicaton (Sonics & Materials, Dombury, USA) at 300W for 20 s at a concentration of 40,000 ppm. The particle size of the hydrocarbon droplets obtained varied from 1 to 1.5 mm. Subsequently these emulsions were diluted to a concentration of 2500 ppm and supplemented with proteins immediately prior to electrophoretic mobility measurements. Proteins dissolved in potassium phosphate solutions were added to a final concentration between 0.0001 and 0.1 mg/ml for HSA and BSA, between 0.0005 and 0.01 mg/ml for IgG, and between 0.00005 and 0.005 mg/ml for Fg. Plasma was diluted 400 times with phosphate solution before use. For comparison, some experiments were done in phosphate buffered saline, PBS (per liter: 0.87 g K2HPO4, 0.68 g KH2PO4, and 8.76 g NaCl, pH 7) as well. Microelectrophoresis Electrophoretic mobilities of the hexadecane droplets in the different solutions with the appropriate pH’s were measured at room temperature with a Lazer Zee Meter 501 (PenKem, Bedford Hills, NY), which uses scattering of incident laser light to enable detection of the hexadecane droplets at relatively low magnifications (25). Measurements were done automatically as a function of the protein adsorption time with a homemade tracking routine based on image analysis (26). However, preparation of the electrophoresis chamber required 1.5 min, after which it appeared that electrophoretic mobilities of the protein-coated hexadecane droplets were stable in time (except for IgG). All electrophoretic mobilities reported are averages of experiments carried out in triplicate with separately prepared hexadecane emulsions. RESULTS
Figure 1 presents the electrophoretic mobilities of hexadedane droplets in solutions of BSA, HSA, IgG, and Fg as a function of concentration and for different pH values of the potassium phosphate solution. With the exception of IgG (see Fig. 2), the effect of protein adsorption on the electrophoretic mobilities of hexadecane occurred within 1.5 min, and no changes over time were measured after the first measurement, taken 1.5 min after the injection of the protein solution in the hexadecane emulsion. Electrophoretic mobilities generally increased upon adsorption of the proteins, with the largest in-
ELECTROPHORETIC MOBILITIES
187
Iso-electric points of protein-coated hexadecane complexes depend clearly on the type of protein adsorbed. For both types of albumin, an iso-electric point of 5.0 is observed, while Fg-coated hexadecane has an iso-electric point of 5.5. The most elevated iso-electric point is found after IgG adsorption to the hexadecane, i.e., at pH 6.6. Interestingly, after protein adsorption from the binary mixtures, the iso-electric point for the HSA/IgG mixture is also highest at pH 5.7, while for the HSA/Fg mixture the iso-electric point after adsorption to hexadecane is located at pH 4.8. After adsorption of plasma proteins to hexadecane, an iso-electric point in between is measured at pH 5.0. The effect of the ionic strength and composition upon the electrophoretic mobilities of hexadecane after protein adsorption can be seen in Fig. 4 for selected protein concentrations. In PBS, electrophoretic mobilities are considerably less negative
FIG. 1. Electrophoretic mobilities of hexadecane droplets after BSA, HSA, IgG, and Fg adsorption in 10 mM potassium phosphate as a function of protein concentration and for different pH values. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
creases occurring for the highest protein concentrations. Adsorption of BSA and HSA had similar effects on the hexadecane electrophoretic mobilities, but for pH 3 and 4 electrophoretic mobilities of BSA-coated hexadecane were significantly more positive than those of HSA-coated hexadecane (see Fig. 1). Adsorption of IgG yielded a significant change in the electrophoretic mobilities of hexadecane toward more positive values, and in fact the hexadecane electrophoretic mobilities become positive already for low IgG concentration below pH 6. The increase of hexadecane electrophoretic mobilities upon Fg adsorption toward positive values is extremely gradual at pH 6 and 7, but for pH 5 and below positive electrophoretic mobilities are found readily for low-protein concentrations. Figure 3 shows the electrophoretic mobilities of hexadecane after protein adsorption as a function of pH. This figure is complemented with electrophoretic mobilities of binary HSA/ IgG and HSA/Fg protein mixtures as well as of diluted plasma.
FIG. 2. Electrophoretic mobilities of hexadecane droplets during IgG adsorption (0.005 mg/ml) in 10 mM potassium phosphate as a function of adsorption time and for different pH values. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
188
VAN DER MEI, MEIJER, AND BUSSCHER
use of an almost ideally smooth and chemically homogeneous hexadecane interface for adsorption. The results obtained on charge compensation in PATH as described here are complimentary to the results by Cheng et al. (24) on interfacial tension changes during PATH obtained by ADSA-P. In a slightly modified form, by replacing the hydrocarbon phase by an organic solvent such as chloroform or di-ethylether, the influence of acid– base interactions in microbial adhesion has been directly assessed (MATS) (23). Replacing the hydrocarbon phase by an organic solvent is feasible for PATH too. Makino et al. (27) did a similar study on the reversal of electrophoretic mobilities upon adsorption of plasma proteins from single-protein solutions to poly(L-lactide) microcapsules. However, their experiments were done in a phosphate buffer (10 –154 mM) at pH 7.4 only, did not involve protein mixtures, and were conducted over a protein concentration range of up to 5 mg/ml, whereas our concentrations did not exceed 1 mg/ml.
FIG. 3. Electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate as a function of pH after adsorption of BSA (0.1 mg/ml), HSA (0.1 mg/ml), IgG (0.05 mg/ml), Fg (0.005 mg/ml), and binary protein mixtures of HSA/IgG and HSA/Fg as well as after protein adsorption from diluted plasma. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
than in 10 mM potassium phosphate solutions at pH 7, except for IgG. For IgG-coated hexadecane, either from a single IgG protein solution or from IgG adsorption from binary HSA/IgG solutions and from plasma, the difference between electrophoretic mobilities in both solutions is less than 0.70 1028m2V21s21. DISCUSSION
In this paper we employ a standard technique to study microbial adhesion (MATH) in a somewhat modified form to study the effect of protein adsorption to hydrocarbons (PATH) on charge compensation of hexadecane droplets in aqueous emulsions. PATH in combination with particulate microelectrophoresis is an easily applicable in situ technique that makes
FIG. 4. A comparison of the electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate (KPi) and in PBS (pH 7) after adsorption of BSA (0.1 mg/ml), HSA (0.1 mg/ml), IgG (0.05 mg/ml), Fg (0.005 mg/ml), and binary protein mixtures of HSA/IgG and HSA/Fg as well as after protein adsorption from diluted plasma. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
189
ELECTROPHORETIC MOBILITIES
Comparing our 10 mM data at pH 7 with those from Makino et al. (27), a correspondence is observed in the sense that electrophoretic mobilities become less negative upon albumin adsorption, but we do not see a minimum in electrophoretic mobility at an albumin concentration of 0.5 mg/ml as this is far higher than our highest concentration used. Adsorption of IgG to poly(L-lactide) microcapsules yielded a rapid increase in electrophoretic mobility to 20.70 1028m2V21s21, similar to the electrophoretic mobility changes upon IgG adsorption to hexadecane here. Also the data for Fg adsorption are comparable, although at the lower concentrations used here we observe a very gradual increase in electrophoretic mobility of Fg-coated hexadecane droplets with protein concentration. The minimum in electrophoretic mobility observed by Makino et al. (27) at an albumin concentration of 0.5 mg/ml and a pH of 7.4 was interpreted to be due to an increase in the amount of negative charge to the poly(L-lactide) microcapsule surface upon albumin adsorption, compared to the increase in amount of positive charge (the iso-electric point of albumin is around pH 4.7). IgG and Fg did not show such an electrophoretic mobility minimum because they possess considerably less negative charge at pH 7.4 (their iso-electric points are around pH 6.8 and 5.5, respectively). For elevated protein concentrations, it was suggested that the plane of slipping that determines the electrophoretic mobilities measured may shift into the outer adsorbed protein layers, containing the positively charged groups. Consequently, they proposed that adsorbed proteins rearrange their structure so that their hydrophobic and negatively charged groups contact the sorbent surface, while the positive charge-rich groups form the outer adsorbed layer. Our results, albeit obtained at significantly lower protein concentrations, confirm this model. Likely, the model also explains why only IgG adsorption to hexadecane shows time-dependent effects (Fig. 2), as the other, much smaller proteins, may rearrange their structure more rapidly than IgG. In addition to the work by Makino et al. (27), the present study is done as a function of pH. Interestingly, the iso-electric points of the protein-coated hexadecane (see Fig. 3) nearly coincides with the iso-electric points of the pure, unadsorbed proteins. It is a confirmation of the model by Makino et al. (27) in the sense that the net charge addition upon protein adsorption, which is positive below the iso-electric points of the proteins, at low concentrations determines the effects on the final electrophoretic mobilities measured. Generally, for cationic sorbent surfaces, the iso-electric points of protein–sorbent complexes do not coincide with the iso-electric point solute protein (28). Elgersma et al. (29) emphasized the importance of the iso-electric points of protein–sorbent complexes rather than the iso-electric point of solute protein as the pH value where protein adsorption is maximal. The iso-electric point of protein– hexadecane complexes after adsorption from binary mixtures is for HSA/Fg close to the iso-electric point of pure, unadsorbed HSA, while for HSA/IgG it is in between the iso-electric points of the protein–sorbent
complexes after adsorption of the single proteins. Also for proteins adsorbed from plasma, the iso-electric point of the plasma– hexadecane complex is close to the iso-electric point of albumin. Tentatively, this might enable the calculation of the surface coverage of the hexadecane by each of the adsorbing proteins and suggests that albumin is adsorbed to the hexadecane surface in higher amounts than the other proteins. This study does not indicate any anomalous adsorption behavior of albumin at the lowest concentration as found by ADSA-P (24) but does demonstrate that electrostatic repulsion between hexadecane and albumin may be a cause for the negative interfacial pressures at concentrations below 0.001 mg/ml. Finally, both with respect to electrophoretic mobilities as well as with respect to interfacial pressures, PATH offers several more possibilities for studying physico-chemical effects of protein adsorption in situ than presented here and by Cheng et al. (24). Desorption phenomena, either detergent stimulated or not, the use of other hydrocarbons and solvents as sorbent surface, and protein displacement interactions are feasible to study by PATH. REFERENCES 1. Haynes, C. A., and Norde, W., Colloids Surf. B 2, 517 (1994). 2. Norde, W., Arai, T., and Shirahama, H., Biofouling 4, 37 (1991). 3. Elwing, H., Ivarsson, B., and Lundstrom, I., Eur. J. Biochem. 156, 359 (1986). 4. Lin, Y. S., and Hlady, V., Colloids Surf. B 2, 481 (1994). 5. Norde, W., and Rouwendal, E., J. Colloid Interface Sci. 139, 169 (1990). 6. Rothenberg, Y., Boruvka, L., and Neumann, A. W., J. Colloid Interface Sci. 93, 169 (1983). 7. Miller, R., Policova, Z., Sedev, R., and Neumann, A. W., Colloids Surf. A 76, 179 (1993). 8. Busscher, H. J., Van der Vegt, W., Noordmans, J., Schakenraad, J. M., and Van der Mei, H. C., Colloids Surf. 58, 239 (1991). 9. Van Dijk, L. J., Goldsweer, R., and Busscher, H. J., Biofouling 1, 19 (1988). 10. Schneider, R. P., and Marshall, K. C., Colloids Surf. B 2, 387 (1994). 11. Van der Vegt, W., Van der Mei, H. C., and Busscher, H. J., Langmuir 10, 1314 (1994). 12. Serrien, G., Geeraerts, G., Ghosh, L., and Joos, P., Colloids Surf. 68, 219 (1992). 13. Van der Vegt, W., Van der Mei, H. C., and Busscher, H. J., J. Colloid Interface Sci. 156, 129 (1993). 14. Van der Vegt, W., Norde, W., Van der Mei, H. C., and Busscher, H. J., J. Colloid Interface Sci. 179, 57 (1996). 15. Go¨lander, C.-G., Lin, Y.-S., Hlady, V., and Andrade, J. D., Colloids Surf. 49, 289 (1990). 16. Rosenberg, M., Gutnick, D., and Rosenberg, E., FEMS Microbiol. Lett. 9, 29 (1980). 17. Van Oss, C. J., Chaudhury, M. K., and Good, R. J., Chem. Rev. 88, 927 (1988). 18. Busscher, H. J., Van de Belt-Gritter, B., and Van der Mei, H. C., Colloids Surf. B 5, 111 (1995). 19. Medrzycka, K. B., Colloid Polym. Sci. 269, 85 (1991). 20. Geertsema-Doornbusch, G. I., Van der Mei, H. C., and Busscher, H. J., J. Microbiol. Methods 18, 61 (1993). 21. Van der Mei, H. C., De Vries, J., and Busscher, H. J., Appl. Environ. Microbiol. 59, 4305 (1993).
190
VAN DER MEI, MEIJER, AND BUSSCHER
22. Van der Mei, H. C., Van de Belt-Gritter, B., and Busscher, H. J., Colloids Surf. B 5, 117 (1995). 23. Bellon-Fontaine, M. N., Rault, J., and Van Oss, C. J., Colloids Surf. B 7, 47 (1996). 24. Cheng, P., Lahooti, S., Policova, Z., Cabrerizo-Vilchez, M. A., and Neumann, A. W., Colloids Surf. B 6, 279 (1996). 25. James, A. M., “Microbial Cell Surface Analysis—Structural and Physicochemical Methods,” VCH, New York, 1991.
26. Wit, P. J., Noordmans, J., and Busscher, H. J., Colloids Surf. A 125, 85 (1997). 27. Makino, K., Oshima, H., and Kondo, T., J. Colloid Interface Sci. 115, 65 (1987). 28. Galisteo-Gonzalez, F., Puig, J., Martin-Rodriguez, A., Serra-Domenech, J., and Hidalgo-Alvarez, R., Colloids Surf. B 2, 435 (1994). 29. Elgersma, A. V., Zsom, R. L. J., Norde, W., and Lyklema, J., Colloids Surf. 54, 89 (1991).
205, 185–190 (1998)
CS985669
Electrophoretic Mobilities of Protein-Coated Hexadecane Droplets at Different pH H. C. van der Mei,1 S. Meijer, and H. J. Busscher Materia Technica, University of Groningen, Bloemsingel 10, 9712 KZ Groningen, The Netherlands Received March 17, 1998; accepted May 18, 1998
Electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate solutions (pH between 3.0 and 7.0) and in phosphate buffered saline (pH 7.0) were measured during adsorption of bovine (BSA) and human (HSA) serum albumin, immunoglobulin G (IgG), and fibrinogen (Fg) from single-protein solutions as well as during protein adsorption from binary HSA/IgG and HSA/Fg mixtures and from diluted plasma. Electrophoretic mobilities became less negative upon adsorption of proteins within 1.5 min after the initiation of adsorption. Only for IgG, was a timedependent change of the electrophoretic mobilities of the protein– hexadecane complex observed. In phosphate buffered saline, less negative electrophoretic mobilities were measured than in the potassium phosphate solution. Iso-electric points of the protein– hexadecane complexes in 10 mM potassium phosphate were located at pH 5.0 for albumin, at pH 5.5 for Fg, and at pH 6.6 for IgG, i.e., about the iso-electric points of the pure unadsorbed proteins. This confirms that the net charge addition upon protein adsorption, which is positive below the iso-electric point of the proteins, at low protein concentrations determines the effects on the final electrophoretic mobilities of the protein– hexadecane complexes. As a name for the methodology applied, we propose PATH (protein adsorption to hydrocarbons), in analogy to the well-known MATH (microbial adhesion to hydrocarbons) method. The major advantage of PATH is that it represents an in situ method to study protein adsorption, without artifactual rinsing steps, while furthermore the hydrocarbon phase can be replaced by organic solvents to study the role of acid– base interactions in protein adsorption. In combination with drop-shape analysis techniques, PATH also enables us to determine in situ effects of protein adsorption on interfacial tensions. © 1998 Academic Press Key Words: protein adsorption; electrophoretic mobility; hexadecane.
INTRODUCTION
Protein adsorption is extensively studied both from an applied as well as from a fundamental point of view (1). Adsorption of structurally stable proteins is governed by the hydrophobicity of the sorbent surface and electrostatic interactions, whereas less stable proteins may show unexpected adsorption behavior due to their tendency to unfold upon adsorption (1, 2). 1
To whom correspondence should be addressed.
Adsorption of proteins to sorbent surfaces is still frequently studied by methods that require removal of “loosely-bound proteins” by rinsing, dipping, or other washing steps, although it is realized more and more that in situ methods can be used with a greater accuracy to study protein adsorption as they avoid these irreproducible steps. Frequently employed in situ methods already include ellipsometry and other reflectometric methods (3, 4), streaming potential measurements (5), and axisymmetric drop shape analysis (6 – 8). Often adsorption of proteins is accompanied by changes in the surface free energy and electrostatic properties of the sorbent surface. Contact angle measurements on dried sorbent surfaces after protein adsorption demonstrated that surface free energies of different materials converged to one value (9, 10). The rinsing and drying steps in such measurements are admittedly hard to avoid, but nevertheless make the reliability of the results subject to debate (11). Changes in electrostatic properties during protein adsorption have been assessed with in situ streaming potential measurements (5). Recently, axisymmetric drop shape analysis by profile (ADSA-P) (6) has been employed to determine the changes in surface tension and contact angle of a protein solution droplet on a sorbent surface. Subsequently, without any washing steps or controversial assumptions regarding the surface thermodynamics involved, the surface tension and contact angle measured can be inserted in the Young equation to yield the solid–liquid interfacial tension. Analyses of the solid–liquid interfacial tension changes over time during adsorption of proteins (12) indicated that the major decrease in surface energetics is accomplished by adsorption of proteins, while conformational changes of the adsorbed protein have less influence (13, 14). By far all studies on protein adsorption have been carried out on real-life surfaces, under the assumption that they are ideally smooth and homogeneous. Clearly, all surfaces are rough at the nanometer-scale and chemical heterogeneities are impossible to avoid. Protein adsorption on hydrophobicity gradient surfaces, obtained for instance by silane diffusion, is strongly influenced by the chemical heterogeneity existing along the length of the gradient, and detergent-induced desorption of albumin is most difficult from the heterogeneous transition region and not from the ends of the gradient with extreme
185
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
186
VAN DER MEI, MEIJER, AND BUSSCHER
hydrophobicities (15). Hydrocarbon interfaces against an aqueous solution represent an almost ideally smooth and homogeneous surface for the study of biological interactions at an interface, and the use of such systems has been pioneered for microbial adhesion (16). The hydrocarbon interface is hydrophobic due to its inability to exert acid– base interactions (17) and is negatively charged above pH 3 (18), as established by particulate microelectrophoresis on hydrocarbon emulsions (19). In fact, microbial adhesion to hydrocarbons (MATH) has yielded an elegant description of microbial adhesion in terms of the DLVO-approach, with only hydrophobic microorganisms with little negative surface charge adhering well to the hydrocarbon interface (20 –22). Bellon-Fontaine et al. (23) has demonstrated the involvement of acid– base interactions in microbial adhesion by comparing adhesion to hydrocarbons with microbial adhesion to organic solvents (MATS), able to exert acid– base interactions. Cheng et al. (24) used the ideal properties of hydrocarbons in aqueous solutions to study protein adsorption (PATH). By analyzing the shape of a pendant decane droplet in a 20 mM Tris buffer (pH 5.6) solution of human serum albumin, “equilibrium” interfacial tensions were obtained. The interfacial tension remained almost constant at approximately 52 mJ/m2 for the two lowest (0.0001 and 0.001 mg/ml) albumin concentrations, implying a negative interfacial pressure of 2 mJ/m2. With further increases in the bulk protein concentration, the interfacial tension dropped sharply to 20 mJ/m2 and remained constant at 20 mJ/m2 at protein concentrations above 0.05 mg/ml. Measurements of the interfacial tension as a function of pH for the lowest concentrations suggested that the negative interfacial pressures found were caused by electrostatic repulsion between the negatively charged hydrocarbon interface and the albumin. The aim of this study is to determine the electrophoretic mobility changes upon protein adsorption to hexadecane (PATH) during particulate microelectrophoresis. All experiments were carried out in 10 mM potassium phosphate solutions with pH adjusted to vary between 3 and 7. Both bovine (BSA) and human (HSA) serum albumin as well as immunoglobulin G (IgG) and fibrinogen (Fg) were used as singleprotein solutions. Experiments were also done with diluted human plasma and binary protein mixtures of HSA/IgG and of HSA/Fg. MATERIALS AND METHODS
Proteins The proteins used in this study were Bovine serum albumin (BSA, 96 –99% albumin, Fraction V), Human serum albumin (HSA, 96 –99% albumin, Fraction V), Immunoglobulin G (IgG, human), and Fibrinogen (Fg, from human plasma, fraction I, type III). All proteins were obtained from Sigma Chemical Company, St. Louis, and used as received. Furthermore,
pooled human plasma anticoagulated with sodium citrate was used. Preparation of Hexadecane Emulsions Hexadecane employed in this study was obtained from Merck (Darmstadt, Germany) and was of highest purity grade. Emulsification in 10 mM potassium phosphate (0.87 g/liter K2HPO4 and 0.68 g/liter KH2PO4) solutions with pH adjusted to vary between 3 to 7 through the addition of HCl or KOH was done by sonicaton (Sonics & Materials, Dombury, USA) at 300W for 20 s at a concentration of 40,000 ppm. The particle size of the hydrocarbon droplets obtained varied from 1 to 1.5 mm. Subsequently these emulsions were diluted to a concentration of 2500 ppm and supplemented with proteins immediately prior to electrophoretic mobility measurements. Proteins dissolved in potassium phosphate solutions were added to a final concentration between 0.0001 and 0.1 mg/ml for HSA and BSA, between 0.0005 and 0.01 mg/ml for IgG, and between 0.00005 and 0.005 mg/ml for Fg. Plasma was diluted 400 times with phosphate solution before use. For comparison, some experiments were done in phosphate buffered saline, PBS (per liter: 0.87 g K2HPO4, 0.68 g KH2PO4, and 8.76 g NaCl, pH 7) as well. Microelectrophoresis Electrophoretic mobilities of the hexadecane droplets in the different solutions with the appropriate pH’s were measured at room temperature with a Lazer Zee Meter 501 (PenKem, Bedford Hills, NY), which uses scattering of incident laser light to enable detection of the hexadecane droplets at relatively low magnifications (25). Measurements were done automatically as a function of the protein adsorption time with a homemade tracking routine based on image analysis (26). However, preparation of the electrophoresis chamber required 1.5 min, after which it appeared that electrophoretic mobilities of the protein-coated hexadecane droplets were stable in time (except for IgG). All electrophoretic mobilities reported are averages of experiments carried out in triplicate with separately prepared hexadecane emulsions. RESULTS
Figure 1 presents the electrophoretic mobilities of hexadedane droplets in solutions of BSA, HSA, IgG, and Fg as a function of concentration and for different pH values of the potassium phosphate solution. With the exception of IgG (see Fig. 2), the effect of protein adsorption on the electrophoretic mobilities of hexadecane occurred within 1.5 min, and no changes over time were measured after the first measurement, taken 1.5 min after the injection of the protein solution in the hexadecane emulsion. Electrophoretic mobilities generally increased upon adsorption of the proteins, with the largest in-
ELECTROPHORETIC MOBILITIES
187
Iso-electric points of protein-coated hexadecane complexes depend clearly on the type of protein adsorbed. For both types of albumin, an iso-electric point of 5.0 is observed, while Fg-coated hexadecane has an iso-electric point of 5.5. The most elevated iso-electric point is found after IgG adsorption to the hexadecane, i.e., at pH 6.6. Interestingly, after protein adsorption from the binary mixtures, the iso-electric point for the HSA/IgG mixture is also highest at pH 5.7, while for the HSA/Fg mixture the iso-electric point after adsorption to hexadecane is located at pH 4.8. After adsorption of plasma proteins to hexadecane, an iso-electric point in between is measured at pH 5.0. The effect of the ionic strength and composition upon the electrophoretic mobilities of hexadecane after protein adsorption can be seen in Fig. 4 for selected protein concentrations. In PBS, electrophoretic mobilities are considerably less negative
FIG. 1. Electrophoretic mobilities of hexadecane droplets after BSA, HSA, IgG, and Fg adsorption in 10 mM potassium phosphate as a function of protein concentration and for different pH values. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
creases occurring for the highest protein concentrations. Adsorption of BSA and HSA had similar effects on the hexadecane electrophoretic mobilities, but for pH 3 and 4 electrophoretic mobilities of BSA-coated hexadecane were significantly more positive than those of HSA-coated hexadecane (see Fig. 1). Adsorption of IgG yielded a significant change in the electrophoretic mobilities of hexadecane toward more positive values, and in fact the hexadecane electrophoretic mobilities become positive already for low IgG concentration below pH 6. The increase of hexadecane electrophoretic mobilities upon Fg adsorption toward positive values is extremely gradual at pH 6 and 7, but for pH 5 and below positive electrophoretic mobilities are found readily for low-protein concentrations. Figure 3 shows the electrophoretic mobilities of hexadecane after protein adsorption as a function of pH. This figure is complemented with electrophoretic mobilities of binary HSA/ IgG and HSA/Fg protein mixtures as well as of diluted plasma.
FIG. 2. Electrophoretic mobilities of hexadecane droplets during IgG adsorption (0.005 mg/ml) in 10 mM potassium phosphate as a function of adsorption time and for different pH values. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
188
VAN DER MEI, MEIJER, AND BUSSCHER
use of an almost ideally smooth and chemically homogeneous hexadecane interface for adsorption. The results obtained on charge compensation in PATH as described here are complimentary to the results by Cheng et al. (24) on interfacial tension changes during PATH obtained by ADSA-P. In a slightly modified form, by replacing the hydrocarbon phase by an organic solvent such as chloroform or di-ethylether, the influence of acid– base interactions in microbial adhesion has been directly assessed (MATS) (23). Replacing the hydrocarbon phase by an organic solvent is feasible for PATH too. Makino et al. (27) did a similar study on the reversal of electrophoretic mobilities upon adsorption of plasma proteins from single-protein solutions to poly(L-lactide) microcapsules. However, their experiments were done in a phosphate buffer (10 –154 mM) at pH 7.4 only, did not involve protein mixtures, and were conducted over a protein concentration range of up to 5 mg/ml, whereas our concentrations did not exceed 1 mg/ml.
FIG. 3. Electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate as a function of pH after adsorption of BSA (0.1 mg/ml), HSA (0.1 mg/ml), IgG (0.05 mg/ml), Fg (0.005 mg/ml), and binary protein mixtures of HSA/IgG and HSA/Fg as well as after protein adsorption from diluted plasma. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
than in 10 mM potassium phosphate solutions at pH 7, except for IgG. For IgG-coated hexadecane, either from a single IgG protein solution or from IgG adsorption from binary HSA/IgG solutions and from plasma, the difference between electrophoretic mobilities in both solutions is less than 0.70 1028m2V21s21. DISCUSSION
In this paper we employ a standard technique to study microbial adhesion (MATH) in a somewhat modified form to study the effect of protein adsorption to hydrocarbons (PATH) on charge compensation of hexadecane droplets in aqueous emulsions. PATH in combination with particulate microelectrophoresis is an easily applicable in situ technique that makes
FIG. 4. A comparison of the electrophoretic mobilities of hexadecane droplets in 10 mM potassium phosphate (KPi) and in PBS (pH 7) after adsorption of BSA (0.1 mg/ml), HSA (0.1 mg/ml), IgG (0.05 mg/ml), Fg (0.005 mg/ml), and binary protein mixtures of HSA/IgG and HSA/Fg as well as after protein adsorption from diluted plasma. All electrophoretic mobilities are averages of three experiments with an average SD over the experiments of 0.21 1028m2V21s21.
189
ELECTROPHORETIC MOBILITIES
Comparing our 10 mM data at pH 7 with those from Makino et al. (27), a correspondence is observed in the sense that electrophoretic mobilities become less negative upon albumin adsorption, but we do not see a minimum in electrophoretic mobility at an albumin concentration of 0.5 mg/ml as this is far higher than our highest concentration used. Adsorption of IgG to poly(L-lactide) microcapsules yielded a rapid increase in electrophoretic mobility to 20.70 1028m2V21s21, similar to the electrophoretic mobility changes upon IgG adsorption to hexadecane here. Also the data for Fg adsorption are comparable, although at the lower concentrations used here we observe a very gradual increase in electrophoretic mobility of Fg-coated hexadecane droplets with protein concentration. The minimum in electrophoretic mobility observed by Makino et al. (27) at an albumin concentration of 0.5 mg/ml and a pH of 7.4 was interpreted to be due to an increase in the amount of negative charge to the poly(L-lactide) microcapsule surface upon albumin adsorption, compared to the increase in amount of positive charge (the iso-electric point of albumin is around pH 4.7). IgG and Fg did not show such an electrophoretic mobility minimum because they possess considerably less negative charge at pH 7.4 (their iso-electric points are around pH 6.8 and 5.5, respectively). For elevated protein concentrations, it was suggested that the plane of slipping that determines the electrophoretic mobilities measured may shift into the outer adsorbed protein layers, containing the positively charged groups. Consequently, they proposed that adsorbed proteins rearrange their structure so that their hydrophobic and negatively charged groups contact the sorbent surface, while the positive charge-rich groups form the outer adsorbed layer. Our results, albeit obtained at significantly lower protein concentrations, confirm this model. Likely, the model also explains why only IgG adsorption to hexadecane shows time-dependent effects (Fig. 2), as the other, much smaller proteins, may rearrange their structure more rapidly than IgG. In addition to the work by Makino et al. (27), the present study is done as a function of pH. Interestingly, the iso-electric points of the protein-coated hexadecane (see Fig. 3) nearly coincides with the iso-electric points of the pure, unadsorbed proteins. It is a confirmation of the model by Makino et al. (27) in the sense that the net charge addition upon protein adsorption, which is positive below the iso-electric points of the proteins, at low concentrations determines the effects on the final electrophoretic mobilities measured. Generally, for cationic sorbent surfaces, the iso-electric points of protein–sorbent complexes do not coincide with the iso-electric point solute protein (28). Elgersma et al. (29) emphasized the importance of the iso-electric points of protein–sorbent complexes rather than the iso-electric point of solute protein as the pH value where protein adsorption is maximal. The iso-electric point of protein– hexadecane complexes after adsorption from binary mixtures is for HSA/Fg close to the iso-electric point of pure, unadsorbed HSA, while for HSA/IgG it is in between the iso-electric points of the protein–sorbent
complexes after adsorption of the single proteins. Also for proteins adsorbed from plasma, the iso-electric point of the plasma– hexadecane complex is close to the iso-electric point of albumin. Tentatively, this might enable the calculation of the surface coverage of the hexadecane by each of the adsorbing proteins and suggests that albumin is adsorbed to the hexadecane surface in higher amounts than the other proteins. This study does not indicate any anomalous adsorption behavior of albumin at the lowest concentration as found by ADSA-P (24) but does demonstrate that electrostatic repulsion between hexadecane and albumin may be a cause for the negative interfacial pressures at concentrations below 0.001 mg/ml. Finally, both with respect to electrophoretic mobilities as well as with respect to interfacial pressures, PATH offers several more possibilities for studying physico-chemical effects of protein adsorption in situ than presented here and by Cheng et al. (24). Desorption phenomena, either detergent stimulated or not, the use of other hydrocarbons and solvents as sorbent surface, and protein displacement interactions are feasible to study by PATH. REFERENCES 1. Haynes, C. A., and Norde, W., Colloids Surf. B 2, 517 (1994). 2. Norde, W., Arai, T., and Shirahama, H., Biofouling 4, 37 (1991). 3. Elwing, H., Ivarsson, B., and Lundstrom, I., Eur. J. Biochem. 156, 359 (1986). 4. Lin, Y. S., and Hlady, V., Colloids Surf. B 2, 481 (1994). 5. Norde, W., and Rouwendal, E., J. Colloid Interface Sci. 139, 169 (1990). 6. Rothenberg, Y., Boruvka, L., and Neumann, A. W., J. Colloid Interface Sci. 93, 169 (1983). 7. Miller, R., Policova, Z., Sedev, R., and Neumann, A. W., Colloids Surf. A 76, 179 (1993). 8. Busscher, H. J., Van der Vegt, W., Noordmans, J., Schakenraad, J. M., and Van der Mei, H. C., Colloids Surf. 58, 239 (1991). 9. Van Dijk, L. J., Goldsweer, R., and Busscher, H. J., Biofouling 1, 19 (1988). 10. Schneider, R. P., and Marshall, K. C., Colloids Surf. B 2, 387 (1994). 11. Van der Vegt, W., Van der Mei, H. C., and Busscher, H. J., Langmuir 10, 1314 (1994). 12. Serrien, G., Geeraerts, G., Ghosh, L., and Joos, P., Colloids Surf. 68, 219 (1992). 13. Van der Vegt, W., Van der Mei, H. C., and Busscher, H. J., J. Colloid Interface Sci. 156, 129 (1993). 14. Van der Vegt, W., Norde, W., Van der Mei, H. C., and Busscher, H. J., J. Colloid Interface Sci. 179, 57 (1996). 15. Go¨lander, C.-G., Lin, Y.-S., Hlady, V., and Andrade, J. D., Colloids Surf. 49, 289 (1990). 16. Rosenberg, M., Gutnick, D., and Rosenberg, E., FEMS Microbiol. Lett. 9, 29 (1980). 17. Van Oss, C. J., Chaudhury, M. K., and Good, R. J., Chem. Rev. 88, 927 (1988). 18. Busscher, H. J., Van de Belt-Gritter, B., and Van der Mei, H. C., Colloids Surf. B 5, 111 (1995). 19. Medrzycka, K. B., Colloid Polym. Sci. 269, 85 (1991). 20. Geertsema-Doornbusch, G. I., Van der Mei, H. C., and Busscher, H. J., J. Microbiol. Methods 18, 61 (1993). 21. Van der Mei, H. C., De Vries, J., and Busscher, H. J., Appl. Environ. Microbiol. 59, 4305 (1993).
190
VAN DER MEI, MEIJER, AND BUSSCHER
22. Van der Mei, H. C., Van de Belt-Gritter, B., and Busscher, H. J., Colloids Surf. B 5, 117 (1995). 23. Bellon-Fontaine, M. N., Rault, J., and Van Oss, C. J., Colloids Surf. B 7, 47 (1996). 24. Cheng, P., Lahooti, S., Policova, Z., Cabrerizo-Vilchez, M. A., and Neumann, A. W., Colloids Surf. B 6, 279 (1996). 25. James, A. M., “Microbial Cell Surface Analysis—Structural and Physicochemical Methods,” VCH, New York, 1991.
26. Wit, P. J., Noordmans, J., and Busscher, H. J., Colloids Surf. A 125, 85 (1997). 27. Makino, K., Oshima, H., and Kondo, T., J. Colloid Interface Sci. 115, 65 (1987). 28. Galisteo-Gonzalez, F., Puig, J., Martin-Rodriguez, A., Serra-Domenech, J., and Hidalgo-Alvarez, R., Colloids Surf. B 2, 435 (1994). 29. Elgersma, A. V., Zsom, R. L. J., Norde, W., and Lyklema, J., Colloids Surf. 54, 89 (1991).