10148
Langmuir 2005, 21, 10148-10153
Mechanisms of Competitive Adsorption of Albumin and Sodium Myristate at the Silicon Oxide/Aqueous Interface† Scott J. McClellan and Elias I. Franses* School of Chemical Engineering, Purdue University, 480 Stadium Mall Drive, West Lafayette, Indiana 47907-2100 Received March 3, 2005 The competitive adsorption of proteins and surfactants has applications to chromatographic systems and biological materials. Adsorption for systems of bovine serum albumin (BSA) and sodium myristate (SM) was investigated with in-situ ATR-IR spectroscopy and ex-situ ellipsometry. The results were used to determine quantitatively the surface densities of the adsorbates at the surface. For a mixture of SM and BSA at 25 °C in water, the adsorbed density of BSA is 0.3 mg/m2, which is much less than the value of 3.1 mg/m2 for BSA alone. Sodium myristate, some of which is protonated to myristic acid (MA) when adsorbed because of a pH decrease from 9.0 to 8.2, adsorbs to a surface density of 4.0 × 10-6 mol/m2, which is greater than the value of 1.7 × 10-6 mol/m2 from a solution of SM alone. Adsorbed SM and MA are removed, or desorbed, when the bulk mixture is replaced with water, with only a slight amount of SM remaining. When placed in contact with a layer of BSA, SM can displace most of the adsorbed protein, even when BSA is present in the bulk solution, with some BSA at 0.3 mg/m2 remaining adsorbed. Allowing BSA to adsorb to a layer of SM results in ΓBSA ) 2.3 mg/m2, with little displacement of the SM layer. These results indicate that SM can remove some BSA from the surface by displacement, and that some BSA remains adsorbed in patches.
1. Introduction An understanding of the adsorption equilibria and dynamics of protein/surfactant systems at solid/aqueous interfaces is important for chromatography, dental applications, biomaterials design, and other fields.1-3 In chromatography, small molecules are often used to displace protein molecules if they are strongly or “irreversibly” adsorbed. A fundamental understanding is needed of the mechanisms by which surfactants displace proteins at solid surfaces or compete with proteins for adsorption sites. Little has been reported on the direct probing of competitive adsorption of such systems. In previous studies, sodium myristate (SM), which is an anionic surfactant with a carbon chain length of 14, has shown the ability to exclude and expel bovine serum albumin (BSA) or fibrinogen (Fb) from the air/water interface.4,5 BSA is a serum protein responsible for the transport of fatty acids for metabolism within the body.6 Up to six molecules of myristic acid (MA), the protonated form of SM, have been shown to bind strongly per BSA molecule as a result of this biological function; additional molecules may adsorb weakly.7,8 Complexes formed between SM and BSA may have a different adsorption affinity from that of BSA. Moreover, these solutions have †
Part of the Bob Rowell Festschrift special issue. * Corresponding author. E-mail:
[email protected]. Tel: (765) 494-4078. Fax: (765) 494-0805. (1) Josic, D.; Reuter, W. In High-Performance Liquid Chromatography of Peptides and Proteins: Separation, Analysis and Conformation; Mant, C. T., Hodges, R. S., Eds.; CRC Press: Boca Raton, FL, 1991; p 231. (2) Arnebrant, T. Biopolymers at Interfaces, 2nd ed.; Malmsten, M., Ed.; Marcel Dekker: New York, 2003; p 811. (3) Baier, R. E.; Dutton, R. C. J. Biomed. Mater. Res. 1969, 3, 191. (4) McClellan, S. J.; Franses, E. I. Colloids Surf., B 2003, 30, 1. (5) Hernandez, E. M.; Phang, T. L.; Wen, X. Y.; Franses, E. I. J. Colloid Interface Sci. 2002, 250, 271. (6) Peters, T. J. All about Albumin: Biochemistry, Genetics, and Medical Applications; Academic Press: San Diego, CA, 1996. (7) Spector, A. A.; John, K.; Fletcher J. E. J. Lipid Res. 1969, 10, 56. (8) Curry, S.; Brick, P.; Franks, N. P. Biochim. Biophys. Acta 1999, 1441, 131.
different pH values, which may affect the protein and surface charges and hence the electrostatics of adsorption. The pH of a 2 mM solution of SM is 9.0, and that of 0.1 wt % BSA is 5.2. (These concentrations are used primarily in this article.) At the mixture pH of 8.2, SM is protonated to MA, which is less soluble than SM and has a different surface activity for either the air/water or solid/aqueous interface. BSA has been shown to adsorb on hydrophilic or hydrophobic surfaces,9-13 which are important for normal or reverse-phase chromatography.14 SM or MA may also adsorb on such surfaces and may modify their hydrophilicity or hydrophobicity. In a previous study, the adsorbed density of BSA was found to be higher on a hydrophilic silicon oxide surface than on that surface covered with a hydrophobic layer of adsorbed dipalmitoylphophatylcholine (DPPC).9 This suggests that the order by which BSA and SM are introduced into a solution and how they interact with the surface may affect the ultimate adsorbed densities because of changes in the surface hydrophilicity. In this article, we use in-situ ATR-IR and ex-situ ellipsometry to study adsorption onto surfaces of silicon oxide/silicon from aqueous BSA/SM systems. Specifically, we investigate the ability of the smaller SM molecules to compete for the surface with the BSA. After we report the results for competitive adsorption from mixtures, we examine the effect of the order of introducing the molecules on the effectiveness of SM, which can displace BSA from the surface or reduce the adsorbed amount of the protein compared to that for BSA alone. (9) McClellan, S. J. Ph.D. Thesis, Purdue University, 2005. (10) Su, T. J.; Lu, J. R.; Thomas, R. K.; Cui, Z. F. J. Phys. Chem. B 1999, 103, 3727. (11) Giacomelli, C. E.; Norde, W J. Colloid Interface Sci. 2001, 233, 234. (12) Vermette, P.; Gauvreau, V.; Pezolet, M.; Laroche, G. Colloids Surf., B 2003, 29, 285. (13) Kim, J.; Somorjai, G. A. J. Am. Chem. Soc. 2003, 125, 3150. (14) Snyder, L. R.; Kirkland, J. J.; Glajch, J. L. Practical HPLC Method Development, 2nd ed.; John Wiley and Sons: New York, 1997; Chapter 6.
10.1021/la050579k CCC: $30.25 © 2005 American Chemical Society Published on Web 07/19/2005
Albumin/Myristate Adsorption at Silicon Oxide/Aqueous Interface Table 1. Summary of Conditions for Adsorption Experiments for Albumin and Sodium Myristate pH
CBSA wt %
CSM mM
9000
8.2 6.5
0.1 0.0
2.0 0.0
4, 5 4, 5 4, 5
9000
5.2 9 6.5
0.1 0.0 0.0
0.0 2.0 0.0
6, 7 6, 7 6, 7
16 000
5.2 8.2 6.5
0.1 0.1 0.0
0.0 2.0 0.0
8, 9 8, 9 8, 9
7500
9 5.2 6.5
0.0 0.1 0.0
2.0 0.0 0.0
expt
figures
1
1-3 1-3
2
3
4
time s
2. Materials and Methods 2.1. Materials and Sample Preparation. Sodium myristate (SM) (>99%) was purchased from Fluka Chemical Co. (Ronkonkoma, NY); bovine serum albumin (BSA) (96-99% pure by gel electrophoresis) was purchased from Sigma Chemical Company (St. Louis, MO). Both were used as received. Solutions were prepared by weight using Millipore water with a resistivity of 18 MΩ cm and were used the same day to minimize possible denaturation. 2.2. In-Situ ATR-FTIR Spectroscopy. A Nicolet Prote´ge´ 460 Fourier transform infrared spectrometer equipped with an MCT detector cooled with liquid nitrogen was used to obtain all FTIR spectra. Spectral contributions from water vapor and carbon dioxide were minimized by continuously purging the instrument’s sample chamber with dry air from a Balston purge gas generator. Attenuated total reflection (ATR) spectra were collected with unpolarized incident light at 25 °C using a custom-made accessory. In-situ spectra were taken of the films on the IRE in contact with aqueous solution using a liquid cell from Harrick Scientific (Ossining, NY). The incident angle θ was typically 50°, with the resulting number of reflections N being 7, as determined from a calibration procedure involving the water band.9,15 All spectra were taken at a resolution of 4 cm-1 with one level of zero filling (resulting in the same data spacing as spectra taken at 2 cm-1) using Happ-Genzel apodization. For dynamic measurements, spectra from 0 to 10 min were collected using 64 scans. Some data points at time less than 10 min have been omitted for clarity. All other spectra were collected using 256 scans. A summary of the experiments conducted and the conditions for each can be found in Table 1. Because of the strong H-O-H bending vibration peak at 1640 cm-1, which overlaps with the amide I band of the protein, the water band was subtracted by using the method outlined by Chittur.16 In the region from 1900 to 1740 cm-1, a straight baseline was created by adjusting the subtraction fraction for the water band. The spectrum of water vapor in the same region was also subtracted. The spectra were finally baseline corrected from 1720 to 1480 cm-1. Surface densities from ATR absorbance data were calculated on the basis of the following equations for a flat isotropic film with unpolarized light.17,18 The concentration profile away from the surface in the z direction is assumed to be stepwise: C(z) ) Ci + Cb for 0 < z < d, and C(z) ) Cb for z > d; Ci is the interfacial concentration, Cb is the bulk concentration, and d is the film thickness. For a two-layer system in which the refractive index of the layer is ignored, the absorbance per reflection is
( )
( )
2de 2de A ) Cbde + (Cid) ) Cbde + Γ N dp dp
(1)
Γ ) (Ci + Cb)d ≈ Cid
(2)
where
(15) Chittur, K. K.; Fink, D. J.; Leininger, R. I.; Hutson, T. B. J. Colloid Interface Sci. 1986, 111, 419.
Langmuir, Vol. 21, No. 22, 2005 10149
n21 ≡ n2/n1, where n2 is the real refractive index of the sample (for water at 1550 cm-1, n2 ) 1.33719) and n1 is that of the substrate (n1 ) 3.4 for silicon); is the molar absorptivity of the solute in solution in cm2/g, as determined from transmission IR data using Beer’s law; and θ is the incident angle. For adsorbed SM, a literature value of 1.7 × 106 cm per mol of CH2 groups was used for , whereas 1.1 × 106 cm/mol was used for bulk SM.20 The error in the surface density of calculation for SM is higher in the presence of BSA because of the uncertainty from the subtraction of the protein side-chain absorbance. The depth of penetration dp of the electric field at wavelength λ is defined as21
dp )
λ 2πn1(sin2 θ - n212)1/2
(3)
Typical values of dp at 1550 cm-1 range from 0.5 to 0.6 µm. The effective thickness de (nm), which is the sample thickness resulting in the same absorbance in transmission, is21
de ) de⊥ )
de| )
de⊥ + de| 2
(4)
2n21dp cos θ
(5)
(1 - n212)
2n21(2 sin2 θ - n212)dp cos θ
.
(1 - n212)[(1 + n212) sin2 θ - n212]
(6)
The surface density Γ can be determined if A/N, , n21, θ, de, dp, and Cb are known. If Cb ) 0 when the solution is replaced by water or small (Cb , dp/2Γ), then
( )
2de A ) Γ N dp
(7)
where Γ is the same quantity as that determined from eq 1 if adsorption is irreversible and if there is no loss of protein from the film upon replacement of the solution by water. 2.3. Ellipsometry. Ellipsometry angles ∆ and Ψ of BSA layers (from dip coating or adsorption) at the air/solid interface were measured using an auto ELII automatic null ellipsometer (Rudolph Research Corp., Flanders, NJ). Measurements were taken at three wavelengths (633, 546, and 405 nm) and two angles (70 and 60°). The average for five or more measurements was used in the calculations. After determining the refractive index (n - ki) of silicon and the refractive index and thickness (n and d) of the overlaying native silicon dioxide using our optimization algorithm,8 film thicknesses df were determined using a four-layer system (Si/ SiO2/film/air). Because the exact solution of this system usually resulted in unrealistic values of nf (∼1.7), for films consisting of only BSA or SM the values of df were found by using a film refractive index nf equal to that of the pure substance (1.55 for BSA, 1.43 for SM22). The surface density Γ is then found from the following equation23
Γ ) df
( )( ) 2 M nf - 1 Ar nf2 + 2
(8)
where M is the molecular weight (g/mol) and Ar is the molar (16) Chittur, K. K. Biomaterials 1998, 19, 357. (17) Sperline, R. P.; Muralidharan, S.; Freiser, H. Langmuir 1987, 3, 198. (18) Jang, W. H.; Miller, J. D. Langmuir 1993, 9, 3159. (19) Downing, H. D.; Williams, D. J. Geophys. Res. 1975, 80, 1656. (20) Kumar, V.; Krishnan, S.; Steiner, C.; Maldarelli, C.; Couzis, A. J. Phys. Chem. B 1998, 102, 3152. (21) Harrick, N. J. Internal Reflection Spectroscopy, 2nd ed.; Harrick Scientfic Corp.: Ossining, NY, 1979; Chapter 2. (22) Jamshed, J. In CRC Handbook of Chemistry and Physics; Weast, R. C., Ed.; CRC Press: Boca Raton, FL, 1980. (23) Cuypers, P. A.; Corsel, J. W.; Janssen, M. P.; Kop, J. M. M.; Hermens, W. T.; Hemker, H. C. J. Biol. Chem. 1983, 258, 2426.
10150
Langmuir, Vol. 21, No. 22, 2005
McClellan and Franses
Figure 2. Dynamic absorbances for (0) amide II for 0.1 wt % BSA alone; (4) νa-CH2 for 2 mM SM alone; (9) amide II for a mixture of 0.1 wt % BSA and 2 mM SM; (2) νa-CH2 for a mixture of 0.1 wt % BSA and 2 mM SM; and (b) CdO for MA from a mixture of 0.1 wt % BSA and 2 mM SM. Table 2. Infrared Spectral Data for Adsorbed Layers from a Mixture of SM and BSA amide I
Figure 1. ATR-IR absorbances for (a) a BSA/SM mixture after 30 s; (b) a BSA/SM mixture after 5 min; (c) a BSA/SM mixture after 20 min; (d) a BSA/SM mixture after 60 min; (e) a BSA/SM mixture after 150 min; and (f) a layer from the mixture after a water rinse. (Top) C-H stretching region; (bottom) carbonyl/ amide region. For a and b, the scale was enlarged because of the small signal. refractivity (cm3/mol). For BSA and SM, M/Ar values of 4.12 and 3.35 g/cm3 were used, respectively.23 For mixed films of SM and BSA, the surface densities were calculated first with a fourlayer model and for nf ranging from 1.43 (that of pure SM) to 1.55 (that of pure BSA), which resulted in maximum values for the surface densities of each molecule. The use of an intermediate value of 1.50 for nf combined with the Bruggeman effective media approximation24 results in the estimate that 61% of the film is made up of BSA and 39% is SM. Then, a five-layer model (Si/ SiO2/SM/BSA/air) was used by assuming that the BSA and SM layers are stratified and then setting the refractive indices equal to 1.43 and 1.55 for each separate layer.
3. Results and Discussion 3.1. Competitive Adsorption of BSA and SM. The in-situ infrared spectra of a mixture of 0.1 wt % BSA and 2 mM SM adsorbing competitively were obtained at various times after contact with the surface (Figure 1). For times shorter than 5 min (spectra a and b), the amide I and II bands of the protein can be resolved better than the CH2 peaks of SM at 2920 and 2850 cm-1 (which are quite noisy because of the low absorbance values) and have higher absorbances, indicating that the surface contains mainly BSA. The amide I peak position (Table 2) for spectrum a (at a pH of 8.2) of 1638 cm-1 is quite different from the value of 1654 cm-1 observed for BSA adsorbing alone (at a pH of 5.2). This indicates that the protein conformation has changed because of the interaction with the surfactant or the pH difference.25 Only minor differences were observed in the amide II band, which is (24) Bruggeman, D. A. G. Ann. Phys. 1935, 24, 636. (25) Ananthapadmanabhan, K. P. In Interactions of Surfactants with Polymers and Proteins; Goddard, E. D., Ananthapadmanabhan, K. P., Eds.; CRC Press: Boca Raton, FL, 1993; p 319.
amide II
a-CH2
s-CH2
spectrum in
ν ∆ν ν ∆ν ν ∆ν ν ∆ν cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1 cm-1
Figure 1a Figure 1b Figure 1c Figure 1d Figure 1e Figure 1f BSA ref SM ref
1638 1641 1642 1650 1650 1643 1654 N/A
56 51 45 38 48 46 49 N/A
1545 1545 1544 1544 1544 1545 1547 N/A
59 43 39 32 53 47 45 N/A
2921 2915 2918 2918 2918 2917 2930 2918
17 15 15 29 23
2850 2854 2850 2850 2850 2849 2855 2849
11 10 9 25 14
less sensitive than the amide I band to changes in the secondary structure of the protein.9,26 For longer times (spectra c-e), the peak position of amide I shifts closer to that of the value for BSA alone. The shift indicates that at these times BSA is less affected by interactions with SM, and the band intensities change little. The absorbances of the CH2 peaks of SM, however, become larger, which shows that SM continues to adsorb. Moreover, in spectra c-e, the peak at 1725 cm-1 is assigned to the carbonyl group of myristic acid (MA). Evidently, at this pH of 8.2, SM adsorbs much more than at a pH of 9.0 for SM alone. After the bulk solution is replaced with water (spectrum f, pH ∼6.5), the amide peaks remain and resemble more closely those of the first spectrum. The CH2 bands become broader than those of the previous spectra, they match those of SM alone more closely (Table 2), and their absorbances decrease sharply. Because the carbonyl peak is no longer observed, indicating that no MA remains at the surface, and because CH2 is seen, we infer that some SM remains adsorbed. For BSA adsorbing alone, as shown for reference in Figure 2, the absorbance of the amide II band (AII) reaches a value of 0.0035 within seconds, indicating fast protein adsorption, and reaches a steady value of 0.0060 after 1 h. At a pH of 5.2 for BSA alone, which is close to the isoelectric point of 4.7,5 the protein has a small negative charge, and electrostatic repulsion from the negatively charged silicon oxide surface is expected to be minimal. The initial AII value for the mixture is significantly lower than for BSA alone (0.0006 to 0.0035, respectively). At longer times, a slight increase in AII to 0.0010 is observed, indicating some additional adsorption. The lower values of AII may result from the presence of SM at the surface (26) Susi, H. In Structure and Stability of Biological Macromolecules; Timasheff, S. N., Fasman, G. D, Eds.; Dekker: New York, 1969; p 575.
Albumin/Myristate Adsorption at Silicon Oxide/Aqueous Interface
Figure 3. Dynamic surface densities (Γ) for adsorption from (0) 0.1 wt % BSA alone; (4) 2 mM SM alone; (9) BSA from a mixture of 0.1 wt % BSA and 2 mM SM; and (2) SM/MA from a mixture of 0.1 wt % BSA and 2 mM SM.
Langmuir, Vol. 21, No. 22, 2005 10151
Figure 5. Dynamic absorbances for (9) amide II and (2)νaCH2 for a layer from 0.1 wt % BSA followed by the injection of 2 mM SM.
Figure 4. Schematic representation of partial displacement of BSA from the silicon oxide/water interface by SM. Our hypothesis is that the BSA is in patches on the surface directly and on the SM monolayer.
or from increased electrostatic repulsion from the surface by the more negatively charged protein at a pH of 8.2. After the bulk mixture is replaced with water, there is a slight drop in AII, showing that the majority of BSA remains adsorbed, apparently irreversibly. For SM alone at a pH of 9.0, the headgroup of the molecule is negatively charged and may create an electrostatic barrier to adsorption. The absorbance of the CH2 band (ACH2) is initially small, 0.0001, and increases slowly to 0.0013 after 2.5 h. Because there is no carbonyl peak at 1725 cm-1, all of the surfactant should be in the SM form. In the mixture, ACH2 is much higher than those values observed for SM alone, reaching a maximum of 0.0040. The adsorbed amount of MA is also significant, as evidenced by the absorbance of the carbonyl peak (ACdO) reaching 0.0018. After replacement with water, ACH2 remains at a value of 0.0009, whereas no carbonyl peak is observed, as shown in Figure 1. The surface densities (Γ) were calculated from the absorbance data as described in section 2.2 (Figure 3). For BSA adsorbing alone, it was found that Γ ) 3.1 mg/ m2, which corresponds to a layer thickness of df ) 2.3 nm for a complete, homogeneous layer with nf ) 1.55. For adsorption from a mixture, ΓBSA reaches a value of 0.3 mg/m2, or only about 10% of that of BSA alone. The maximum value of Γ observed for SM is 1.7 × 10-6 mol/ m2. For the mixture, the total surfactant surface density (ΓSM/MA) reaches a value of 4.0 × 10-6 mol/m2. After replacing the bulk solution with water, the SM surface density drops to ΓSM ) 2.0 × 10-6 mol/m2. Our hypothesis is that the surface contains a layer of SM and patches of BSA adsorbed either on the silicon oxide surface or on the SM-covered surface (Figure 4). As SM adsorption proceeds, some of it becomes protonated to MA, which can be completely desorbed during the rinsing stage. 3.2. Displacement of Adsorbed BSA Layers by SM. To examine the ability of SM to remove an adsorbed layer of BSA from the interface, BSA was allowed to adsorb to the silicon oxide surface and was followed with in-situ
Figure 6. Dynamic adsorbed densities for (9) BSA and (2) SM for a layer from 0.1 wt % BSA followed by the injection of 2 mM SM.
ATR-IR (Figures 5 and 6). As with the results reported in section 3.1 for BSA alone, AII quickly reaches a value of 0.0025 and then increases to a value of 0.0044. The intensity of the protein side-chain band (at ∼2930 cm-1) overlaps slightly with the CH2 stretch of the SM (Figure 5). After the injection of SM, there is a sharp drop in AII of the protein to 0.0026 with a continuous decrease to 0.0005, indicating that most of the BSA is removed from the surface. After the protein is removed, ACH2 initially decreases and then starts increasing as more SM adsorbs to the surface. After the bulk solution is replaced by water, there is little change in either ACH2 or AII. The final value of ΓBSA is 0.2 mg/m2, which is comparable to that of adsorption from the BSA/SM mixture. The final value of ΓSM is 0.3 × 10-6 mol/m2, which is significantly smaller than the value of 2.0 × 10-6 mol/m2 for the mixture. A question arises as to what happens when BSA is first allowed to adsorb and then the bulk BSA solution is replaced not with SM only (pH 9.0), but with an SM and BSA mixture (pH 8.2). The results of such an experiment are reported in Figures 7 and 8. As with the injection of SM alone, a large decrease is observed in AII from 0.0069 to 0.0011. The corresponding values of ΓBSA are 3.5 and 0.4 mg/m2, respectively. The value of ACH2 again decreases initially and then increases to 0.0034, which is less than for SM alone. The band at 1725 cm-1 is again observed, indicating the presence of MA on the surface. The total surfactant surface density is 3.4 × 10-6 mol/m2 before replacement of the solution with water. After water replacement, there is little change in AII. There are large decreases, however, in ACH2, 0.0039-0.0007, and ACdO, 0.0015-0.0000. These results show that SM is able to
10152
Langmuir, Vol. 21, No. 22, 2005
Figure 7. Dynamic absorbances for (9) amide II, (2)νa-CH2, and (O) CdO for a layer from 0.1 wt % BSA followed by the injection of a 0.1 wt % BSA/2 mM SM mixture.
McClellan and Franses
Figure 10. Dynamic adsorbed densities for (9) BSA and (2) SM for a layer from 2 mM SM followed by the injection of 1000 ppm BSA. Table 3. Film Thicknesses and Surface Densities of Dip-Coated Films from Solutions of SM and BSA Determined by Ex-Situ Ellipsometry expt 1 2
4
Figure 8. Dynamic adsorbed densities for (9) BSA and (2) SM/MA for a layer from 0.1 wt % BSA followed by the injection of a 0.1 wt % BSA/2 mM SM mixture.
Figure 9. Dynamic absorbances for (9) amide II and (2)νaCH2 for a layer from 2 mM SM followed by the injection of 1000 ppm BSA.
displace BSA from the surface, with or without BSA present in the bulk solution. The main difference is that more surfactant (SM and MA) adsorbs from the mixture than from the pure solution (SM alone). 3.3. Adsorption of BSA onto an Adsorbed Layer of SM. Because SM shows the ability to displace BSA from the surface, we tested whether BSA would adsorb on an already-formed SM monolayer. SM was first allowed to adsorb at the surface from a solution with a pH of 9.0 (Figures 9 and 10). The final measured value of ACH2 before BSA injection is 0.0004, which corresponds to ΓSM ) 0.3 × 10-6 mol/m2. After BSA is introduced, the value of ACH2 increases to 0.0014, which is mainly attributed to additional adsorption of SM during the time in which measurements were not taken. The small increase of ACH2 with time is due to absorbance contributions from BSA.
4 a
system mixture
3
BSA BSA + SM SM SM + BSA BSA, five
dfa IR data ΓBSA 106 × ΓSM in figure nf,assumed nm mg/m2 mol/m2
layerb
5 5 9 9 9
1.55 1.50 1.43 1.55 1.55 1.50 1.43 1.43 1.55 1.50 1.43 1.55
2.8 2.9 3.1 3.3 1.4 1.5 1.6 1.6 4.5 4.7 4.8 3.1
3.7 2.1 4.3 1.8 1.1 5.9 3.5
4.4 10.7 2.3 5.5 5.5 7.1 16.6
4.0 b
Four-layer model, unless otherwise stated. Five-layer model assumes that there is no loss of SM upon exposure to the BSA solution and that the layers are stratified. (See section 2.3).
The value of ΓSM remains fairly constant at 1.0 × 10-6 mol/m2, indicating that the protein adsorbs without removing SM from the surface. There is more uncertainty in these calculated surface densities because the BSA absorbance must be subtracted from the total. The value for AII quickly reaches a value of 0.0040, with ΓBSA ) 1.7 mg/m2. The absorbance increases rapidly to a plateau at about 0.0050, with ΓBSA ) 2.3 mg/m2. Upon water replacement, the absorbance shows little change, indicating that all BSA adsorption was irreversible. 3.4. Ex-Situ Characterization of Dip-Coated Layers of BSA/SM Systems. Ellipsometric angles ∆ and Ψ (not shown) were obtained for each film at an angle of incidence of 70° and for three wavelengths (633, 546, and 405 nm). These data were used to determine film thicknesses and surface densities for assumed values of the refractive index of the film (Table 3). For the film from the mixture, if one assumes that only BSA is present and nf ) 1.55, then one obtains df ) 2.8 nm and ΓBSA ) 3.7 mg/ m2. If one assumes that only SM is present and nf ) 1.43, then one finds df ) 3.1 nm and ΓSM ) 10.7 × 10-6 mol/m2, which seems to be quite unrealistic. Because the ex-situ IR spectrum of this film shows that both SM and BSA are present, one has to use an intermediate value for nf. A value of nf ) 1.50 results in df ) 2.9 nm, ΓBSA ) 2.1 mg/m2, and ΓSM ) 4.4 × 10-6 mol/m2, which are plausible. The results for an SM solution contacting a layer of BSA show that the effective value of df decreases from 3.3 to about 1.5 nm if nf ) 1.50 is assumed. The calculation of ΓBSA based on the assumptions outlined in section 2.3 results in the decrease in the surface density of BSA from 4.3 to 1.0 mg/m2, showing that SM can displace adsorbed
Albumin/Myristate Adsorption at Silicon Oxide/Aqueous Interface
BSA from the surface. Conversely, when a BSA solution contacts a layer of SM, df increases from 1.6 to 4.7 nm. Because the substrate was not exposed to any additional SM and hence ΓSM is expected to remain unchanged or decrease, it follows that the increase in df implies that BSA adsorbs on the SM monolayer. If one assumes nf ) 1.50, then the resulting values for ΓBSA and ΓSM would be 3.5 mg/m2 and 7.1 × 10-6 mol/m2, respectively. A slight change of nf ) 1.51 changes the resulting values for ΓBSA and ΓSM to 4.0 mg/m2 and 5.5 × 10-6 mol/m2, respectively. If one uses a five-layer model (silicon/silicon oxide/SM/ BSA/air) with the assumption of no loss of SM, then one obtains dBSA ) 3.1 nm and ΓBSA ) 4.0 mg/m2, which are the same values obtained for a four-layer model with nf ) 1.51. Although the surface densities determined from ellipsometry tend to be up to four times greater than the values determined from the in-situ data, consistent trends are observed. The differences can result from errors in the parameters used (, nf, M/A, etc.) or probably from the model assumptions of a flat, homogeneous surface and a flat, uniform, isotropic film. Another difference is that the ATR results represent a global average of the film for many reflections, whereas the ellipsometry results represent a local average at the spot measured. 4. Conclusions The adsorption of BSA/SM systems to a hydrophilic native silicon oxide on a silicon surface was examined using quantitative in-situ ATR-IR. For the competitive adsorption of a mixture of BSA/SM, the surface density of BSA quickly reaches a steady value of 0.5 mg/m2, which is much less than 3 mg/m2 for BSA alone, indicating that the presence of SM reduces BSA adsorption. The adsorbed density of SM reaches a value of 4 × 10-6 mol/m2 with some being in the form of MA, which arises from the
Langmuir, Vol. 21, No. 22, 2005 10153
protonation of SM at a pH of 8.2. The adsorption of SM/ MA is faster than for SM alone, indicating some electrostatic nature in the adsorption. Replacement of the mixture by water results in little change in the amount of ΓBSA, whereas ΓSM is reduced to 0.9 × 10-6 mol/m2 with no evidence of the presence of MA. Changes in the amide I band of BSA adsorbed from a mixture with SM, compared to that adsorbed from a solution of BSA alone, imply that the presence of SM may lead to changes in the protein conformation. The introduction of a solution of SM in contact with an adsorbed layer of BSA results in the reduction of ΓBSA from 2.0 to 0.2 mg/m2. The surface density of SM reaches a value of 0.2 × 10-6 mol/m2 before water replacement, which then results in little change in AII or ACH2. An adsorbed layer of BSA in contact with a solution of BSA and SM is also reduced from 3.4 to 0.4 mg/m2. At a pH of 8.2 for the mixture, some adsorbed SM is protonated to MA; the total surface density of SM/MA reaches a value of 3.4 × 10-6 mol/m2, with only SM with ΓSM ) 0.6 × 10-6 mol/m2 remaining after water replacement. BSA adsorbs irreversibly onto a layer of SM, without displacing it, to a surface density of 2.3 mg/m2. Ex-situ ellipsometry results are consistent with the in-situ ATR conclusions. We conclude that SM can partially displace BSA from the hydrophilic silicon oxide surface and reduce protein adsorption from a mixture. The results have implications for designing chromatographic systems and for understanding or designing biomaterial surfaces for specific interactions with proteins, surfactants, or protein/surfactant mixtures. Acknowledgment. This work was supported in part from a grant by the Indiana 21st Century Research and Technology Fund. LA050579K
