Langmuir 2002, 18, 4653-4658
4653
Protein Transport in Nanoporous Membranes Modified with Self-Assembled Monolayers of Functionalized Thiols Kyoung-Yong Chun and Pieter Stroeve* NSF Center on Polymer Interfaces and Macromolecular Assemblies, CPIMA, Department of Chemical Engineering and Materials Science, University of California, Davis, 1 Shields Avenue, Davis, California 95616 Received August 6, 2001. In Final Form: February 16, 2002 Control of external pH and ionic strength is used to separate proteins with surface-modified, nanoporous polycarbonate track etched (PCTE) membranes. The porous PCTE membranes were modified with monolayers of self-assembled thiols (HSC10H20COOH) on electroless gold. The hydraulic radius of the pores in the surface-modified membranes was 8.7 nm. Two proteins of nearly identical molecular weight, bovine serum albumin (BSA) and bovine hemoglobin (BHb), were used as the permeants. The fluxes of BSA and BHb through the membranes show maximum values at the isoelectric points (pI) of the proteins. At pH values above and below the pI, charge interactions between the proteins, their counterions, and the pore surface leads to a decrease in flux. The imposition of a difference in ionic strength across the membrane causes osmotic flow and leads to a significant increase in the protein fluxes and an enhancement of the selectivity of BSA over BHb. In protein separation experiments, the BSA and BHb fluxes are nearly 3 times larger than those observed with no ionic strength difference.
Introduction Separation of proteins can be effectively carried out with membrane separation processes.1-7 In general, proteins are easily separated with porous membranes if the proteins have significantly different molecular sizes. However, membrane separation of similar size proteins, such as hemoglobin and albumin, is more difficult, and these separations have received renewed interest. For example, Zydney and Pujar have studied fundamental colloidal interactions that govern hemoglobin and albumin transport through porous membranes.8 Musale and Kulkarni used ultrafiltration studies to study the relative rates of hemoglobin and albumin transport through poly(acrylonitrile) homopolymer and copolymer membranes.9 Van Eijndhoven et al. reported the effect of pH on the separation of albumin and hemoglobin in poly(sulfone) membranes at different solution ionic strengths.10 Kontturi and co-workers studied separation of albumin and hemoglobin by convective electrophoresis at high ionic strength.11 In pores modified by self-assembled monolayers (SAMs), the transport of species inside the pore can be influenced * To whom correspondence should be addressed. E-mail:
[email protected]. Telephone: (530) 752-8778. Fax: (530) 7521031. (1) Torres, M. R.; Ramos, A. J.; Soriano, E. Bioprocess Eng. 1998, 19, 213-215. (2) Li, Q. Y.; Cui, Z. F.; Pepper, D. S. J. Membr. Sci. 1997, 136, 181190. (3) Koehler, J. A.; Ulbricht, M.; Belfort, G. Langmuir 1997, 13, 41624171. (4) Saksena, S.; Zydney, A. L. J. Membr. Sci. 1997, 125, 93-108. (5) Pujar, N. S.; Zydney, A. L. J. Chromatogr., A 1998, 796, 229-238. (6) Nakao, S.; Osada, H.; Kurata, H.; Tsuru, T.; Kimura, S. Desalination 1988, 70, 191. (7) Fane, A. G.; Fell, C. J. D.; Waters, A. G. J. Membr. Sci. 1983, 16, 211-224. (8) Zydney, A. L.; Pujar, N. S. Colloids Surf., A 1998, 138, 133-143. (9) Musale, D. A.; Kulkarni, S. S. J. Membr. Sci. 1997, 136, 13-23. (10) van Eijndhoven, R. H. C. M.; Saksena, S.; Zydney, A. L. Biotechnol. Bioeng. 1995, 48, 406-414. (11) Kontturi, A.-K.; Kontturi, K.; Vuoristo, M. Acta Chem. Scand. 1996, 50, 102-106.
by the functional headgroups of the SAM.12 In our previous work, we have shown that SAMs formed from thiols with functional carboxylic acid groups on electroless gold impart a pH-dependent flux of ions across the porous membrane.13,14 Recently, Lee and Martin also have reported on nanoporous membranes with pH-switchable ion transport due to SAMs.15,16 Transport studies on ultrafiltration membranes are often focused on pressure-driven convective flow.4-10 Transport studies on SAM-modified porous membranes are very new, and current attention has been focused on diffusive transport phenomena. A recent article has appeared on protein transport in poly(ethylene glycol)derivatized gold nanotubule membranes.17 Protein transport has not been studied with surface-modified, porous membranes in which the pore surface charge can be controlled by pH. For protein transport inside pores lined with charged headgroups, the situation is complex since the charge of the protein and the SAM functional headgroups are sensitive to the value of the pH in the solution. Such studies are important to give insight on what type of surface modification may give unusual transport properties in porous membranes. In this work, we have studied the diffusive transport and separation of bovine serum albumin (BSA) and bovine hemoglobin (BHb) in nanoporous, polycarbonate track etched (PCTE) membranes modified with a SAM of HSC10H20COOH thiol on electroless gold. In addition, we have conducted experiments in the presence of flow caused by an osmotic pressure gradient. The functional headgroup of the thiol that we used is carboxylic acid whose charge is pH dependent. Bovine serum albumin (66 000 kD) is of similar molecular weight as bovine hemoglobin (65 000 kD) but has a lower isoelectric point (pI ) 4.7) than BHb (12) Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Anal. Chem. 1999, 71, 4913-4918. (13) Hou, Z.; Abbott, N. L.; Stroeve, P. Langmuir 2000, 16, 24012400. (14) Chun, K. Y.; Stroeve, P. Langmuir 2001, 17, 5271-5275. (15) Lee, S. B.; Martin, C. R. Anal.Chem. 2001, 73, 768-775. (16) Lee, S. B.; Martin, C. R. Chem. Mater. 2001, 13, 3236-3244. (17) Yu, S.; Lee, S. B.; Kang, M.; Martin C. R. Nano Lett. 2001, 1, 495-498.
10.1021/la011250b CCC: $22.00 © 2002 American Chemical Society Published on Web 05/15/2002
4654
Langmuir, Vol. 18, No. 12, 2002
Chun and Stroeve
Figure 1. Sink concentration versus time (a) and flux (b) of BHb in a PCTE/Au/HS(CH2)10COOH membrane as a function of external pH (lines in all figures in this article are to aid the eye).
(pI ) 7.0). At the isoelectric point of a protein, electrostatic interactions between the protein and the membrane surface are minimized and subsequently the flux of the protein should be larger than that at other pH values. Thus, the pH should have a profound effect on the transport of BSA and BHb inside the surface-modified pores which should lead to selectivity between these proteins of similar molecular weight. We change the surface charge of the pores in the PCTE membranes and the charge of the proteins by controlling the external pH. We demonstrate that electrostatic interactions between proteins and pore walls are important factors in protein transport across porous membranes and that these interactions are complex. Experimental Section Materials. Commercial PCTE membranes (Poretics, Inc.) were used as substrates for electroless deposition of gold. The membranes have a hydraulic pore radius of 28 nm, a pore density of 6 pores/(µm)2, and a thickness of 6 µm. The chemicals SnCl2 (98%), AgNO3 (99+%), Na2SO3 (98+%), NH4OH, trifluoroacetic acid (99%), formaldehyde, methanol (HPLC grade), and ethanol (HPLC grade) were obtained from Aldrich. A commercial solution of Na3Au(SO3)2 (Oromerse Part B, Technic Inc., Na3Au(SO3)2 concentration is 7.9 × 10-3 M) was used after dilution in water (40 times). Thiol 11-mercaptoundecanoic acid (HS(CH2)10COOH, Aldrich) was used for forming SAMs on the gold surfaces. Ultrapure water (18 MΩ) was used for preparation of all solutions and for rinsing. Bovine serum albumin and bovine hemoglobin were obtained from Sigma, and no further purification was carried out. Potassium dihydrogen phosphate and disodium hydrogen phosphate (Fluka) were used in buffer solutions for maintaining the external pH. Preparation of Membranes. We deposited gold on the inner pore walls and both faces of the PCTE membranes using the procedure reported by Martin and co-workers.18,19 The temperature during the deposition process was fixed at 1 °C, and the pH of the gold solution was set to around 10 to prevent bottleneck structure formation near the ends of the pore.18,19 After gold plating, the membranes were thoroughly rinsed with water. The gold-coated PCTE membranes were immersed overnight in an (18) Hulteen, J. C.; Jirage, K. B.; Martin, C. R. J. Am. Chem. Soc. 1998, 120, 6603-6604. (19) Jirage, K. B.; Hulteen, J. C.; Martin, C. R. Science 1997, 278, 655-658.
aqueous solution of 25% HNO3 and then rinsed with copious amounts of pure water. To form the SAMs on the gold surface, the gold-coated membrane was first thoroughly rinsed with ethanol and then immersed in an ethanol solution of the acid alkane thiol (5 mM) for 24 h. After deposition of the SAM on the gold membrane, the hydraulic pore radius of the membranes was reduced to 8.7 nm, as determined from hydraulic water flow experiments.13 Protein Transport. We used a batch diffusion cell with two compartments (reservoir and sink). The reservoir contained a buffer solution of protein, and the sink contained a blank buffer solution initially without protein. The effective permeation area of the membrane was 1.77 cm2, and the volumes of the reservoir and sink were 35.0 mL. Vigorous stirring (400 rpm) was present in both compartments by using two magnetic stirrers and a Corning stirring plate. For a desired pH value, we used 0.1 M buffer solutions in the sink and reservoir at 25 °C. In the diffusion experiments with an ionic strength difference, we used buffer solutions of 0.15 and 0.05 M in the reservoir and sink compartment, respectively. The initial concentration of protein in the reservoir (CR0) was 1.0 g/L. The protein concentration is sufficiently small (1.5 × 10-5 M) that it has no effect on the total ionic strength. The protein concentration in the sink (CS) was measured with an UV-visible spectrophotometer (Varian, Cary 3) at 280 nm for BSA and 408 nm for BHb. Both single protein transport experiments and mixed protein separation experiments were carried out. In mixed protein separation, the initial concentrations of BHb and BSA in the reservoir compartment were 1.5 × 10-5 M each. The fluxes of the proteins across the membrane (JBSA and JBHb) were obtained from the slopes of the protein concentrations in the sink (CS) versus time (t).
Results and Discussion Single Protein Transport. We first investigated the effect of external pH on single protein transport through the porous membranes. The hydraulic pore radius (R) of the PCTE/Au/HSC10H20COOH membranes was 8.7 nm. The buffer solution pH varied from 3.76 to 8.35, and the ionic strength of phosphate buffer solution in the compartments was 0.1 M. Figure 1a,b shows the sink concentration versus time and the flux of BHb across a PCTE/Au/HSC10H20COOH membrane as a function of the external pH. The highest flux of BHb is obtained at pH ) pI ) 7, due to a decrease of electrostatic interactions between the charged pore wall and neutral BHb. At pH values above 7, both BHb and the pore surface are
Protein Transport in Nanoporous Membranes
Langmuir, Vol. 18, No. 12, 2002 4655
Figure 2. Sink concentration versus time (a) and flux (b) of BSA in a PCTE/Au/HS(CH2)10COOH membrane as a function of external pH.
negatively charged. At these pH values, the electrostatic repulsion between the negatively charged protein and the negatively charged pore walls is large, leading to a significant decrease in protein flux. For pH values below the pI, the decrease of the BHb flux is also significant but not as profound as for pH values above the pI. Below the pI, the protein is positively charged while the pore surface is partially negatively charged. In this case, BHb is attracted to the pore. However, the BHb counterions are repelled from the pore, slowing their transport. A reduction of the counterion transport should reduce the BHb transport. There is another plausible explanation for the reduction in the flux that has been discussed in the literature. For example, Burns and Zydney20 have attributed the reduction in protein flux below the pI to the energetic penalty associated with the distortion of the electric double layer around the charged protein by the pore wall. This distortion leads to a net repulsive interaction, even when the protein and pore have opposite charge. The repulsive interactions are proportional to the square of the protein charge.20 It will be possible to determine which explanation is more likely to be correct by conducting transport experiments with several protein species with different ionic strengths and pH values. Figure 2a shows the concentration of BSA in the sink for transport through a PCTE/Au/HSC10H20COOH membrane as a function of time and external pH. Similar to the BHb experiments, the flux of BSA through the membrane decreases with pH above and below the pI of BSA (Figure 2b). The flux of BSA is more reduced for pH values above the pI. At the pI of BSA, electrostatic interactions between the neutral BSA and the negatively charged wall are weakest. At pH values above 4.7, the pore wall and the protein are negatively charged, leading to a reduction of the BSA flux. Below pH ) 4.7, the reduction in BSA flux is probably due to either electrostatic repulsion of the counterions of the BSA from the pore or the energetic penalty associated with bilayer distortion,20 leading to a retardation of the flux of BSA. At a pH of 4.7, the pore surface has a negligible negative charge. For lower pH (pH < 4.7), the weak acid groups on the pore wall become uncharged and electrostatic interactions between protein and pore surface disappear. Thus, the reduction of BSA flux below the pI is less prominent at low pH than at high pH values. (20) Burns, D. B.; Zydney, A. L. Biotechnol. Bioeng. 1999, 64, 27-37.
Comparison of the fluxes of the two proteins in Figures 1 and 2 show that the flux of BSA in the porous membranes is larger than the flux of BHb at all pHs except at pH ) 7. Although the molecular weights of BSA and BHb are similar, other parameters such as the mobility, molecular shape, isoelectric point, and the hydrophilic/hydrophobic ratio of amino acids are quite different,9,21-25 and these factors play a role in transport inside the narrow, charged pores. The diffusion coefficient of BSA in dilute bulk solution is lower than that of BHb in dilute solution, 5.9 × 10-11 m2/s for BSA versus 6.4 × 10-11 m2/s for BHb. The shape of BSA is 4 × 4 × 14 nm and resembles a prolate ellipsoid, while the shape of BHb is 6.4 × 5.5 × 5 nm and thus is more spherical than that of BSA. In bulk diffusion, the cigarlike BSA diffuses in all orientations, while BHb diffuses more like a sphere. This shape effect explains the lower mobility in the bulk for BSA compared to BHb. However, diffusion of BSA and BHb in narrow pores is quite different. First, the shapes of the protein molecules influence hindered diffusion for BSA and BHb in narrow pores. When there is electrostatic repulsion between the pore wall and the BSA, the prolate ellipsoid should be more aligned with the axis of the pore. This alignment of BSA within the pore should cause a reduction in hindered diffusion and cause the BSA to diffuse faster compared to BHb. Second, the hydrophilic/hydrophobic ratio of amino acids within the two proteins may play a role in pore transport. The hydrophilic amino acid content for BSA is 56% (hydrophobic 44%), while the hydrophilic amino content for BHb is 44% (hydrophobic 56%). Since the pores are lined with the hydrophilic weak acid groups, this difference in amino acid content should make BSA more compatible inside the pore than BHb. The above factors influence transport of protein molecules through a pore, in addition to the electrostatic interactions between protein molecules themselves and protein interactions with the pore surface. In addition, the ionic strength is important (21) Blank, M. Biomembrane Electrochemistry; American Chemical Society: Washington, DC, 1994; p 432. (22) Dawson, R. W. C.; Elliott, D. C.; Elliott, W. C.; Jones, K. M. Data for Biochemical Research, 3rd ed.; Oxford University Press: New York, 1986. (23) Kupku, S.; Margit, S.; Sleytr, U. B. Desalination 1993, 90, 6576. (24) van den Berg, G. B.; Smolders, C. A. J. Membr. Sci. 1989, 47, 1-24. (25) Barisas, B. G. In Thermodynamic Data for Biochemistry and Biotechnology; Hinz, H.-J., Ed.; Springer: Berlin, 1986.
4656
Langmuir, Vol. 18, No. 12, 2002
Chun and Stroeve
Figure 3. Comparison of BSA (a) and BHb (b) fluxes as a function of ionic strength difference and external pH in single protein diffusion.
in protein transport in pores since it influences the value of the Debye length. In some publications dealing with protein separation, it has been observed that pH effects are more pronounced at lower ionic strengths.10,26 Generally, with high ionic strength, protein separation is decreased due to the reduction of the thickness of the electrical double layer. Our flux measurements were repeatable to within (3% for the same protein. Further, the flux measurements of BHb or BSA in a membrane could be reproduced within (3% before and after diffusion experiments with the other protein. From the former and the latter results, it appears that irreversible protein adsorption does not take place inside the pores. If irreversible protein adsorption took place, the adsorbed protein would partially block the pore and cause a change in protein flux. This is consistent with previous results where we have shown that self-assembled monolayers of hydrophilic thiols on electroless gold do not show significant irreversible adsorption of protein.27 However, it is possible that reversible adsorption of positively charged protein to the negatively charged pore surface plays a role in protein transport. To determine the effect of an ionic strength difference across the membrane on protein transport, the ionic strengths of the buffer solutions were fixed at 0.15 M for the reservoir and 0.05 M for the sink. Experiments were done with the same external pH in both compartments and the same initial protein concentration in the reservoir as before. The 0.10 M difference in ionic strength causes an osmotic flow from the reservoir to the sink. In addition, the change of ionic strength across a membrane pore will lead to an increase in the Debye length along the inside of the pore, from the reservoir to the sink. Both osmotic flow in the pores and the change in Debye length along the pore should have a significant impact on protein transport. Figure 3a,b shows the fluxes for single protein diffusion of BSA (a) and of BHb (b) in the presence (circles) and absence (squares-line) of the 0.10 M ionic strength difference across the membrane. For every pH, the flux in the presence of the ionic strength difference is always larger than the flux obtained in the absence of the ionic strength difference. The increase in flux of the proteins in the presence of the ionic strength difference is ap(26) Ghosh, R.; Cui, Z. F. J. Membr. Sci. 1998, 139, 17-28. (27) Dubrovsky, T. B.; Hou, Z.; Stroeve, P.; Abbott, N. L. Anal. Chem. 1999, 17, 327-332.
Figure 4. Comparison of BSA and BHb fluxes in mixed protein diffusion as a function of external pH.
proximately 10%, except for the case of BSA at pH values less than or equal to 4.7 where the difference is about 40%. The reason for this larger difference at these pH values is not clear. The protein transport is due to both diffusion and convection. In addition, the salt flux can give rise to a diffusion potential that would influence the protein transport. Diffusion potentials are affected by salt concentrations. It is expected that at the 0.15 M ionic strength, which is near the upstream side of the pore, the diffusion potential is smaller than at the downstream side of the pore where the ionic strength is 0.05 M.28-30 There exists a gradient in the diffusion potential across the pore, which makes analysis of the protein transport complicated. Mixed Protein Separation. Figure 4 shows the membrane fluxes of both proteins during mixed protein separation as a function of the external pH. The transport processes taking place in the case of the mixed protein diffusion are more complex than in single protein transport. In the pH range between 4.7 and 7, the two proteins are attracted by electrostatic attraction since BHb is (28) de Koning, J.; Stroeve, P.; Meldon, J. H. Adv. Exp. Med. Biol. 1974, 94, 183-188. (29) Meldon, J. H.; de Koning, J.; Stroeve, P. Bioelectrochem. Bioenerg. 1978, 5, 77-87. (30) Hoofd, L. J. C.; Tong, R. R.; Stroeve, P. Ann. Biomed. Eng. 1986, 14, 493-511.
Protein Transport in Nanoporous Membranes
Langmuir, Vol. 18, No. 12, 2002 4657
Figure 5. Separation selectivity (JBSA/JBHb) as a function of external pH.
positive and BSA is negative. In this pH range, the pore surface is negatively charged. Above a pH of 7, both BSA and BHb are negatively charged, giving rise to electrostatic repulsion while the proteins diffuse through the negatively charged pore. Below a pH of 4.7, the proteins are positively charged and repel each other, while the pore surface is neutral (pH , 4.7). From Figure 4, the flux of BHb changes less than the flux of BSA in the whole pH range. The flux change of BHb from minimum to maximum values is 40%, while that of BSA is about 100% over the pH range. Further, the BHb fluxes are lower than the fluxes of BSA. At pH 4.7, the pI of BSA, the fluxes of both proteins and the difference in the two fluxes are largest. The flux of BSA has a maximum value similar to single protein diffusion at pH 4.7. However, the BSA flux is 20% lower than that for single BSA diffusion and the flux versus pH response is broader than that seen in the single protein experiments. At pH ) 7, the BHb flux does not exhibit a maximum flux, while BSA diffuses faster. It is clear that in the mixed protein experiments there must be significant interactions between the two proteins. Even at the isoelectric point, neutral proteins can interact electrostatically. At the isoelectric point, there usually are positive and negative charges present, even though the net charge of the protein is zero. The presence of positive and negative charges imparts a dipole moment to the protein, which will cause the protein to interact with other charged protein species, leading to attractive interactions that inhibit transport of the proteins. A comparison of the protein fluxes in the mixed protein experiments (Figure 4) with the protein fluxes in the single protein experiments (Figures 1 and 2) shows that in the pH range from 3.7 to 7.5 the fluxes for the mixed protein experiments are lower. These results suggest that protein interactions between BSA and BHb in mixed protein separation cause a decrease in protein fluxes. It is possible that protein clusters form (e.g., dimers, trimers) and these clusters will diffuse slower than single molecules. The data from Figure 4 are shown in Figure 5 in terms of the separation selectivity for BSA and BHb as a function of external pH. The separation selectivity S is defined as
S)
JBSA JBHb
(1)
Here, JBSA and JBHb represent the fluxes of BSA and BHb, respectively. As a comparison, the apparent selectivity
Figure 6. Comparison of BSA and BHb fluxes in mixed protein transport with an ionic strength difference (reservoir, 0.15 M; sink, 0.05 M) as a function of external pH.
Figure 7. Separation selectivity (JBSA/JBHb) with an ionic strength difference (reservoir, 0.15 M; sink, 0.05 M) as a function of external pH.
for the single protein transport from Figures 1 and 2 is also shown. For the mixed protein diffusion, the highest selectivity is 4.3 at pH 4.7 and the lowest selectivity is 2.8 at pH 7. It is clear that the selectivity for the mixed protein experiments is quite different compared to the apparent selectivity of the single protein experiments. Comparison of the fluxes of BSA and BHb for mixed protein transport is given in Figure 6 for the case of a 0.10 M ionic strength difference (the same ionic strength conditions that were used in single protein transport). In Figure 6, the variation of the maximum flux of BHb compared to the minimum flux is about 100% and the same is true for the flux of BSA. Further, the fluxes of BHb and BSA are a factor of 2-3 times larger than the fluxes of the mixed protein separation in the absence of an ionic strength difference shown in Figure 4. These differences in the fluxes observed with and without an ionic strength difference suggest that osmotic flow inside the pore reduces the protein interactions between BHb and BSA inside the pore. It may be that the shear stress due to osmotic flow in the pore is sufficient to break up protein clusters. From the results in Figures 3 and 6, we can obtain the separation selectivity for BSA and BHb in the presence of an ionic strength difference as shown in Figure 7. Again, the selectivity for mixed protein separation is quite different compared to the single protein experiments.
4658
Langmuir, Vol. 18, No. 12, 2002
Further, in a comparison of the results in Figure 7 with the results obtained for uniform ionic strength (Figure 5), the selectivity in mixed protein separation, in the presence of an ionic strength gradient, is increased and the highest separation selectivity is about 4.6 at pH ) 4.7. The lowest selectivity for mixed protein separation in the presence of an ionic strength gradient is found at pH values near 8, where both proteins repel each other. Qualitatively, the behavior of the selectivity results in Figures 5 and 7 is similar, except that the selectivities are larger in the presence of an ionic strength difference. Thus, the imposition of osmotic flow causes a significant increase in protein fluxes with enhanced selectivity of BSA over BHb. The control of pH, ionic strength, and pore size in porous membranes modified with SAMs can lead to unusual transport conditions for the separation of proteins of similar molecular weight. Change of pH leads to change in protein and surface wall charge. Decrease of pore size and/or the overall ionic strength should improve separation of similar-sized proteins because it increases the ratio of Debye length to pore radius. An increase in the ratio of Debye length to pore radius increases the region of the negative voltage inside the pore lumen.31 Simultaneously, there are diffuse double layers surrounding the charged protein molecules and an increase in Debye length or a decrease in pore radius will cause increased electrostatic interactions between proteins and pore wall. In our experiments, the ionic strength was of the order of 0.1 M which leads to a Debye length of about 1 nm. The protein diameters are of the order of 6 nm and the pore radius is of the order of 9 nm, so the Debye length is of modest size in this case. Increased protein-protein and protein-wall interactions will be experienced at lower ionic strength, since the Debye length will increase. For example, at an (31) Adamson, A. W.; Gast, A. P. Physical Chemistry of Surfaces, 6th ed.; John Wiley and Sons: New York, 1997.
Chun and Stroeve
ionic strength of 0.001 M, the Debye length is 10 nm, which is bigger than the pore radius or the protein diameters. At this ionic strength, the electrostatic interactions are larger and one would expect significant increases in protein selectivity and decreases in protein fluxes. We will conduct transport experiments at lower ionic strength in future work. Conclusions Protein transport through surface-modified, porous membranes has been studied as a function of external pH with two proteins of similar molecular weight, BSA and BHb. Polycarbonate track etched (PCTE) membranes were modified with SAMs of alkane thiol (HSC10H20COOH) on gold. The hydraulic pore radius of the membranes was 8.7 nm. For experiments of single protein permeation, BHb and BSA show maximum fluxes at their respective isoelectric pH points (pI), pH ) 7.0 and pH ) 4.7. In mixed protein separation, the fluxes of the proteins are affected by electrostatic interactions between the proteins, the counterions, and the pore surface. Overall, the flux of BSA was larger than that of BHb by a factor of 4.2 at pH 4.7 for mixed protein separation. Imposition of osmotic flow across the membrane causes a significant increase in the protein fluxes and enhances selectivity of BSA over BHb. Chemisorption of functionalized thiols on pore surfaces greatly impacts protein transport through pores and involves an interplay of ionic strengths upstream and downstream of the membrane, protein size and charge, pore surface charge, and pore size. Self-assembled monolayers on pore surfaces can also be used to eliminate irreversible adsorption of proteins on pore walls.17,27 Acknowledgment. This work was supported in part by the MRSEC program of the National Science Foundation (DMR-9808677). LA011250B
