Lipid Membrane Formation by Vesicle Fusion on Silicon Dioxide

Yujia Jing , Hana Trefna , Mikael Persson , Bengt Kasemo , Sofia Svedhem ... and intraneuronal HMW-AbetaOs represent a molecular basis of memory loss ...
1 downloads 0 Views 298KB Size
Download PDF
7526

Langmuir 2004, 20, 7526-7531

Lipid Membrane Formation by Vesicle Fusion on Silicon Dioxide Surfaces Modified with Alkyl Self-Assembled Monolayer Islands Ryugo Tero,*,† Morio Takizawa,‡ Yan-Jun Li,† Masahito Yamazaki,§ and Tsuneo Urisu*,†,‡ Institute for Molecular Science, Okazaki, 444-8585 Japan, Department of Structural Molecular Science, The Graduate University for Advanced Studies, Okazaki, 444-8585 Japan, and Department of Physics, Faculty of Science, Shizuoka University, Shizuoka, 422-8529, Japan Received February 23, 2004. In Final Form: June 8, 2004 Using atomic force microscopy, we have investigated the formation of the dipalmitoylphosphatidylcholine (DPPC) membrane by the vesicle fusion method on SiO2 surfaces modified with self-assembled monolayer (SAM) islands of octadecyltrichlorosilane (OTS) with sizes comparable to those of the vesicles. OTS-SAM islands with various sizes and coverages can be constructed on the SiO2 surfaces prepared by thermal oxidation followed by partial hydroxylation in a H2O2/H2SO4 solution. When vesicles are sufficiently smaller than the SiO2 domains, DPPC bilayers and DPPC/OTS layers form on the SiO2 and OTS domains, respectively. However, the adhesion of larger vesicles onto SiO2 is prevented by the OTS islands; therefore only DPPC/OTS layers form without formation of DPPC bilayers on the SiO2 domains. On surfaces with domains on the scale of tens to hundreds of nanometers, the relative size between the hydrophilic domains and the vesicles becomes an important factor in the membrane formation by the fusion of vesicles.

1. Introduction Self-assembled monolayers (SAMs) and supported membranes have attracted enormous attention as functionalizing methods on inorganic solid surfaces with biomaterials.1-5 Many technical applications are expected for surface modification by SAMs, such as protective coats, surface patterning, nonlinear optics, and control of mechanical and electrical properties.1,2 In bioelectronics fields, SAMs have been applied to the modification of chemical and physical surface properties, for example, hydrophilicity and electric charge, and anchoring other molecules such as lipids and proteins. Supported membranes have been studied as an active medium for biosensors and as systems for mimicking plasma membranes due to the character that membrane proteins can be deposited on solid surfaces, retaining the biological functions as well as controlling the orientations.3-5 The combination of the SAM and supported membrane techniques brings big advancement to the development of new biosensors and biofunctionalized devices.6,7 On SiO2 surfaces, SAMs of alkylchlorosilane8 have been extensively studied. The formation mechanism and the kinetics have been investigated by various experimental techniques8-25 and theoretical simulations.9,10,26 The reac* To whom correspondence should be addressed. E-mail: [email protected], [email protected]. † Institute for Molecular Science. ‡ The Graduate University for Advanced Studies. § Shizuoka University. (1) Ulman, A. Chem. Rev. 1996, 96, 1533. (2) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151. (3) Sackmann, E. Science 1996, 271, 43. (4) McConnell, H. M.; Watts, T. H.; Weis, R. M.; Brian, A. A. Biochim. Biophys. Acta 1986, 864, 95. (5) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, 105. (6) Bieri, C.; Ernst, O. P.; Heyse, S.; Hofmann, K. P.; Vogel, H. Nat. Biotechnol. 1999, 17, 1105. (7) Terrettaz, S.; Mayer, M.; Vogel, H. Langmuir 2003, 19, 5567. (8) Angst, D. L.; Simmons, G. W. Langmuir 1991, 7, 2236. (9) Schwartz, D. K.; Steinberg, S.; Israelachvili, J.; Zasadzinski, J. A. N. Phys. Rev. Lett. 1992, 69, 3354.

tion mechanism is understood as the hydration of alkylchlorosilane to form alkylsilanol followed by the formation of chemical bonds between the alkylsilanol and surface hydroxyl groups. Two-dimensional diffusion-limited aggregation of the physisorbed alkylsilanol leads to the growth of large dendritical islands, and grafting of alkylsilanol clusters formed in the solvent to the substrate results in small dot-type islands.9 These processes are highly sensitive to experimental conditions such as reaction time,9-16 temperature,15-19 the amount of surface water layers,8,19-22 age of the solution,13 and the roughness of the substrate.23 This sensitivity to many parameters makes it difficult to control the growth of the alkylsilaneSAM islands and to obtain good reproducibility. (10) Bautista, R.; Hartmann, N.; Hasselbrink, E. Langmuir 2003, 19, 6590. (11) Balgar, T.; Bautista, R.; Hartmann, N.; Hasselbrink, E. Surf. Sci. 2003, 532-535, 963. (12) Wasserman, S. R.; Tao, Y.-T.; Whitesides, G. M. Langmuir 1989, 5, 1074. (13) Leitner, T.; Friedbacher, G.; Vallant, T.; Brunner, H.; Mayer, U.; Hoffmann, H. Mikrochim. Acta 2000, 133, 331. (14) Bierbaum, K.; Grunze, M. Langmuir 1995, 11, 2143. (15) Goldmann, M.; Davidovits, J. V.; Silberzan, P. Thin Solid Films 1998, 327-329, 166. (16) Carraro, C.; Yauw, O. W.; Sung, M. M.; Maboudian, R. J. Phys. Chem. B 1998, 102, 4441. (17) Brzoska, J. B.; Shahidzadeh, N.; Rondelez, F. Nature 1992, 360, 719. (18) Brzoska, J. B.; Azouz, I. B.; Rondelez, F. Langmuir 1994, 10, 4367. (19) Kumar, N.; Maldarelli, C.; Steiner, C.; Couzis, A. Langmuir 2001, 17, 7789. (20) Britt, D. W.; Hlady, V. J. Colloid Interface Sci. 1996, 178, 775. (21) Tripp, C. P.; Hair, M. L. Langmuir 1992, 8, 1120. (22) Allara, D. L.; Parikh, A. N.; Rondelez, F. Langmuir 1995, 11, 2357. (23) Komeda, T.; Namba, K.; Nishioka, Y. J. Vac. Sci. Technol., A 1998, 16, 1680. (24) Fujii, M.; Sugisawa, S.; Fukada, K.; Kato, T.; Shirakawa, T.; Seimiya, T. Langmuir 1994, 10, 984. (25) Wasserman, S. R.; Whitesides, G. M.; Tidswell, I. M.; Ocko, B. M.; Pershan, P. S.; Axe, J. D. J. Am. Chem. Soc. 1989, 111, 5852. (26) Witten, T. A., Jr.; Sander, L. M. Phys. Rev. Lett. 1981, 47, 1400.

10.1021/la0400306 CCC: $27.50 © 2004 American Chemical Society Published on Web 07/30/2004

Lipid Membrane Formation by Vesicle Fusion

The vesicle fusion method is a feasible and prevalent technique to form supported membranes on solid surfaces.4 When substrates are immersed in an aqueous solution of lipid vesicles, the vesicles adhere to the surface, break up, and spread to form a bilayer on hydrophilic surfaces and a monolayer on hydrophobic surfaces. Another popular technique to fabricate supported membranes is the Langmuir-Blodgett (LB) method.5,27-29 A solution of lipids in a volatile organic solvent is spread at the air-water interface, and the lipid layer is transferred onto the sample surface after the organic solvent evaporates. The vesicle fusion method has several advantages over the LB method: the area over which the membrane is deposited can easily be selected by limiting the area where the suspension is dropped on the substrate, homogeneous membranes form even on surfaces that are not flat, and denaturing due to the organic solvents is avoided for biofunctionalizing molecules such as proteins and enzymes. Membrane structures, formation mechanisms, and the influences of the vesicle size and buffer composition in the vesicle fusion method have been studied.29-48 Recent atomic force microscopy (AFM) studies have particularly given valuable information about the mechanism of the membrane formation by the fusion of vesicles.40-48 These previous studies, however, were carried out only on homogeneous surfaces30-46 or on surfaces with micrometerorder patterns,47,48 and no result was reported about the behavior of vesicles on nanopatterned surfaces. In the present study, we have investigated the formation of the lipid bilayer membrane by the vesicle fusion method on coexisting hydrophilic and hydrophobic domains with the size of the order of hundreds of nanometers. Dipalmitoylphosphatidylcholine (DPPC) membranes on SiO2 surfaces modified by SAM islands of octadecyltrichlorosilane (OTS) with varied coverage were observed by means of AFM. A combination of thermal oxidation and chemical treatment to prepare the SiO2 surfaces made it possible to obtain OTS-SAM islands with various sizes and coverage. We found that the membrane formation is greatly affected by OTS islands and that the relative size of the lipid vesicles and the hydrophilic SiO2 regions is the critical factor. (27) Zasadzinski, J. A.; Viswanathan, R.; Madsen, L.; Garnaes, J.; Schwartz, D. K. Science 1994, 263, 1726. (28) Furuikea, S.; Hirokawa, J.; Yamada, S.; Yamazaki, M. Biochim. Biophys. Acta 2003, 1615, 1. (29) Dufreˆne, Y. F.; Lee, G. U. Biochim. Biophys. Acta 2000, 1509, 14. (30) Puu, G.; Gustafson, I. Biochim. Biophys. Acta 1997, 1327, 149. (31) Nollert, P.; Kiefer, H.; Ja¨hnig, F. Biophys. J. 1995, 69, 1447. (32) Lingler, S.; Rubinstein, I.; Knoll, W.; Offenha¨usser, A. Langmuir 1997, 13, 7085. (33) Parikh, A. N.; Beers, J. D.; Shreve, A. P.; Swanson, B. I. Langmuir 1999, 15, 5369. (34) Pierrat, O.; Lechat, N.; Bourdillon, C.; Laval, J.-M. Langmuir 1997, 13, 4112. (35) Winger, T. M.; Chaikof, E. L. Langmuir 1998, 14, 4148. (36) Seifert, U. Adv. Phys. 1997, 46, 13. (37) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397. (38) Williams, L. M.; Evans, S. D.; Flynn, T. M.; Marsh, A.; Knowles, P. F.; Bushby, R. J.; Boden, N. Langmuir 1997, 13, 751. (39) Kalb, E.; Frey, S.; Tamm, L. K. Biochim. Biophys. Acta 1992, 1103, 307. (40) Muresan, A. S.; Lee, K. Y. C. J. Phys. Chem. B 2001, 105, 852. (41) Egawa, H.; Furusawa, K. Langmuir 1999, 15, 1660. (42) Reviakine, I.; Brisson, A. Langmuir 2000, 16, 1806. (43) Jass, J.; Tja¨rnhage, T.; Puu, G. Biophys. J. 2000, 79, 3153. (44) Leonenko, Z. V.; Carnini, A.; Cramb, D. T. Biochim. Biophys. Acta 2000, 1509, 131. (45) Mou, J.; Yang, J.; Huang, C.; Shao, Z. Biochemistry 1994, 33, 9981. (46) Mou, J.; Yang, J.; Shao, Z. Biochemistry 1994, 33, 4439. (47) Jenkins, A. T. A.; Bushby, R. J.; Evans, S. D.; Knoll, W.; Offenha¨usser, A.; Ogier, S. D. Langmuir 2002, 18, 3176. (48) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651.

Langmuir, Vol. 20, No. 18, 2004 7527

2. Experimental Section As-delivered Si(111) wafers were RCA-cleaned by successive treatment with the following solutions: concentrated H2SO4 + H2O2 (30%) (4:1 in the volume ratio), NH4OH + H2O2 (30%) + H2O (1:1:3), HF (5%), and concentrated HCl + H2O2 (30%) + H2O (1:1:4). After this cleaning, the surface was covered with a chemically oxidized SiO2 layer. The surface roughness of this SiO2 layer was ∼1 nm. The wafers were then further oxidized thermally by heating in the air for 2 h at 700 °C, and then the surface hydrophilicity was adjusted by immersing the wafer into a boiling solution of H2O2 (30%) and concentrated H2SO4 (1:4) for either 3 or 5 s. Very flat surfaces with a roughness about 1/5 that of the chemically oxidized SiO2 were obtained. OTS-SAM islands were deposited by immersing the SiO2/Si(111) substrate in a 10 mM solution of OTS (Aldrich) in water-saturated toluene for 5 s at room temperature (RT). The OTS/SiO2 sample was sonicated in toluene, acetone, ethanol, and pure water in order to remove the excess OTS molecules. The sample treatment and the OTS deposition are described in detail elsewhere.49 Unilamellar vesicles of DPPC (Avanti) were prepared by agitating vacuum-dried DPPC films in a buffer solution (150 mM NaCl, 1.0 mM CaCl2, 10 mM HEPES/NaOH (pH 7.0)) at a concentration of 0.02-0.05 mg/mL, freeze-and-thaw cycles by liquid nitrogen for three times, and extruding the suspension through a 100 or 200 nm polycarbonate filter (Liposo-Fast, Avestin, Inc.). The diameter of the extruded vesicles distributes around the pore size of the filter.39-41,47,50 The suspension after the extrusion was used in the following processes without further dilution. Smaller vesicles were obtained by sonicating the suspension of the 200nm-filtered vesicles in a water-bath sonicator for 40 min in total. Sonication for 5-10 min was repeated to avoid too much increase in the suspension temperature. The sonicated vesicles have been reported to measure 12-50 nm in diameter.37,38,42 The suspension was kept above the transition temperature from the gel to the liquid crystal (Tm) (41 °C for DPPC) during the preparation processes. For the deposition of DPPC membranes, the substrate was incubated under 80 µL of the vesicle suspension for 2 h at 45 °C. Previous studies have shown that the incubation time of 2 h is long enough to assemble a full-coverage bilayer membrane.30-32,37-39,41,47 Then the sample was washed by exchanging 200 µL of the suspension for the buffer solution for four times and finally sunk under 1.5 mL of buffer in an AFM cell. The sample surface was not exposed to the air during the preparation. Fabrication of the bilayer membrane by this scheme was confirmed on mica surfaces by AFM. AFM observations were performed in the buffer solution using a SPI3800 scanning probe microscopy system (Seiko Instruments Inc.) in the dynamic-force mode (tapping mode) using a Si cantilever. The force constants of the cantilevers used for the observation in air and water were 16 and 1.5 N/m, respectively. When the OTS/SiO2 surfaces before and after the DPPC deposition were observed by AFM, the irregularity of the OTS islands and the tip condition was compensated by successively observing two OTS/SiO2 pieces, one with and the other without DPPC, that had been cut from the same OTS/SiO2 sample, using the same cantilever. The averaged values and the standard deviations of the height in AFM images were obtained from the values measured at various places on the same sample using the same cantilever.

3. Results 3.1. Modification of SiO2 Surfaces with OTS-SAM Islands. First we have prepared OTS-SAM islands with varied size and coverage on SiO2 surfaces. Figure 1a shows an OTS-SAM deposited on the chemically oxidized SiO2 surface. High hydrophilicity of the SiO2 surface before the OTS deposition was indicated by a water-contact angle (WCA) of ∼0°. Large dendritical islands and small dot- type islands are observed as in previous studies.9,11,13-16,19,20,23 The formation of the former is (49) Takizawa, M.; Kim, Y.-H.; Urisu, T. Chem. Phys. Lett. 2004, 385, 220. (50) MacDonald, R. C.; MacDonald, R. I.; Menco, B. Ph. M.; Takeshita, K.; Subbarao, N. K.; Hu, L. Biochim. Biophys. Acta 1991, 1061, 297.

7528

Langmuir, Vol. 20, No. 18, 2004

Figure 1. AFM images (5.0 × 5.0 µm2) of SiO2 surfaces modified with OTS islands obtained by various preparation of the SiO2 substrates and reaction time in the reagent solution of OTS. (a) Preparation of SiO2 layer: chemical oxidation in a mixed solution of HCl, H2O2 (30%), and H2O (1:1:4) (just after the RCA cleaning). Reaction time in the solution of OTS: 5 s. (b) Preparation of SiO2 layer: the same as in image a followed by thermal oxidation and H2O2/H2SO4 treatment for 5 s. Reaction time in the solution of OTS: 5 s. (c) Preparation of SiO2 layer: the same as in image b. Reaction time in the solution of OTS: 15 s. (d) Preparation of SiO2 layer: almost identical to that for image b, but the time of the H2O2/H2SO4 treatment is 3 s. Reaction time in the solution of OTS: 15 s. All images were observed in the air.

explained by the two-dimensional diffusion-limited aggregation model,9,26 and that of the latter results from the adsorption of silanol polymers, which form in the liquid phase, on the rough surface.9,51 The shapes and the sizes of the OTS islands on the chemically oxidized SiO2 surface are multiform and quite hard to control. In contrast, circular OTS islands with a uniform size are obtained on the SiO2 surface prepared by thermal oxidation at 700 °C followed by hydrophilicity control with the H2O2/H2SO4 solution (Figure 1b). The WCA of the SiO2 substrate before the OTS deposition was ∼30°. The total coverage of the OTS (θOTS) is 0.29. There are no dot-type islands between the circular islands. The protrusions on the OTS islands are probably aggregated OTS. They will be covalently bonded since the sample was carefully washed by sonication after the deposition of OTS. It is not clear what causes the uniform size and semicircular shape of these protrusions, but it may relate to the first nucleation site of the OTS islands. As shown in Figure 1c (θOTS ) 0.79), the OTS islands keep their circular shape and grow larger with increase of the immersion time in the OTS reagent solution. The size and morphology of the OTS islands depend on the hydrophilicity of the surface. Figure 1d shows OTS islands deposited on a thermally oxidized SiO2 surface with a shorter pretreatment time in the H2O2/ H2SO4 solution (3 s, rather than 5 s used for the surface in Figure 1b,c). Less hydrophilicity of the surface restrains the growth of each OTS island, and the surface is covered with a networklike OTS layer composed of small islands which linked with each other. The heights of the islands (51) Takizawa, M. Doctoral Thesis. 2004.

Tero et al.

Figure 2. AFM images (3.0 × 3.0 µm2) and line profiles of the OTS/SiO2 surfaces (θOTS ) 0.21) obtained in the buffer solution (a) before and (b) after the deposition of 100-nm-filtered vesicles. (c) The magnified image of the square area in image b (800 × 800 nm2). (d) Schematic illustration for the line profile in image b.

in Figure 1a-d are 2.1 ( 0.2 nm, 2.1 ( 0.2 nm, 2.0 ( 0.2 nm, and 2.0 ( 0.2 nm, respectively. These are reasonable values in view of the length of the alkyl chain of OTS (2.62 nm),12 the tilt of the alkyl chain (∼10° from the surface normal22), and the compression by the AFM cantilever. Values ranging from 2.0 to 2.5 nm have been reported for the heights of condensed OTS-SAM in previous AFM studies.11,14,15,19,20 The uniform height of the OTS islands in Figure 1a-d indicates that closely packed uniform OTS layers form independent of the island shape and coverage. The control of the OTS island growth by the combinative pretreatment of the thermal oxidation and H2O2/H2SO4 treatment is not completely precise due to the changes in ambient humidity and temperature, but we can successfully obtain the surface with the coexisting OTS and SiO2 domains with various morphologies and coverages with good reproducibility. 3.2. Formation of DPPC Membranes on OTSModified SiO2 Surfaces. We have deposited DPPC vesicles on the SiO2 surfaces modified with OTS islands and investigated the effect of the OTS islands in the order of hundreds of nanometers in diameter on the formation of lipid bilayer membranes by fusion of vesicles. Figure 2 shows the AFM images of an OTS/SiO2 surface observed in the buffer solution before and after the DPPC deposition by the fusion of 100-nm-filtered vesicles. The average diameter of the OTS islands in Figure 2a is approximately 322 nm, and θOTS is 0.21. The large protrusions are probably polymerized OTS aggregations, which are mainly positioned on the OTS islands. The observed height of the OTS-SAM islands in the buffer solution is 3.7 ( 0.3 nm, which is larger by 1.7 nm than that observed in the air. The exact difference between the heights of hydrophilic and hydrophobic domains (here, SiO2 and OTS) cannot be determined by AFM in an aqueous solution, because the solvent exclusion effect results in an additional force between the cantilever and the hydrophobic surface.52 It also affected the quality of the AFM image. It was quite difficult to stably observe (52) Sinniah, S. K.; Steel, A. B.; Miller, C. J.; Reutt-Robey, J. E. J. Am. Chem. Soc. 1996, 118, 8925.

Lipid Membrane Formation by Vesicle Fusion

the OTS/SiO2 surfaces in the buffer solution when the hydrophilic and hydrophobic domains were mixed as in Figure 2a. We observed the OTS/SiO2 surfaces with and without DPPC successively and repeatedly, and the image quality was always poor only on the surface without DPPC. After the deposition of 100-nm-filtered vesicles, the surface is covered by a flat membrane (Figure 2b). Neither the multilayer membranes nor adsorbed vesicles are observed. As previously reported, the washing with buffer solutions removes excess vesicles and multilayer membranes almost completely and leaves only the first bilayer membrane.39,41 Three regions are observed on the membrane surface in Figure 2c: predominant brighter ()higher) regions (A), gray ()lower) domains (B), and dark defects (C). There were only small and shallow defects existing on the membrane just after the DPPC deposition, and correct measurement of the membrane thickness was not achieved. The defects C are intentionally introduced by exposing the deposited membrane surface to the air for a moment in order to measure the thickness of the membrane from the depth of the defect C. The brighter region A measures 4.6 ( 0.5 nm in height (Figure 2b), which corresponds to the thickness of a lipid bilayer membrane.40-46 We regard that the SiO2 surface is exposed in the defects C and thus the interaction between the cantilever and the surface on C is almost the same as that on A, since both of the DPPC bilayer membrane and SiO2 surfaces are sufficiently hydrophilic. The gray domain B, which is lower than the region A by 0.7 ( 0.2 nm (Figure 2b), is assigned to the OTS region for the following reasons. First, the shapes and the density of the domains B are similar to those of the OTS islands before DPPC deposition, although it is difficult to recognize all of the gray areas in the z-range of Figure 2b. Second, almost all the OTS aggregations, which have been positioned on the OTS islands before the DPPC deposition (Figure 2a), are observed in the domain B. Some of the OTS aggregations were buried in the DPPC monolayer. These OTS domains will be covered by DPPC monolayers since the hydrophobic OTS islands exposed to the aqueous solution are unstable. Previous studies have shown that hybridized lipid/SAM bilayers form after the fusion of vesicles on hydrophobic SAMs, such as thiol/Au32,34,37,47 and OTS/glass systems.35 The result in Figure 2b is summarized in the illustration in Figure 2d: the SiO2 regions are covered by a DPPC bilayer, and the OTS islands are covered by a DPPC monolayer. The structure of the membrane edge is not considered here. When the SiO2 surface has been premodified by OTS islands with higher coverage, a completely different result is obtained. Figure 3a,b shows the AFM images of an OTS/ SiO2 surface (θOTS ) 0.81) before and after the deposition of the 100-nm-filtered vesicles, respectively. The observed height of the OTS islands in Figure 3a is 4.0 ( 0.3 nm, similar to the results in Figure 2a. A remarkable difference from the result shown in Figure 2 is that the shape of the OTS islands is clearly observed even after the deposition of DPPC (Figure 3b). This indicates that a regular DPPC bilayer membrane does not form on the SiO2 regions between the OTS islands. The OTS islands will be covered with a DPPC monolayer similar to Figure 2b. The formation of the DPPC/OTS layer is also indicated from the surface roughness of the OTS domains as shown in the insertions in Figure 3a,b. It decreases from 0.33 nm (root mean square (rms)) to 0.09 nm (rms) after DPPC deposition. The surface of the lipid membrane reported in the previous AFM studies is quite smooth.28,30,40-42,44-46 The monolayer of DPPC obscures the roughness and defects on the OTS islands in Figure 3b. Some circular

Langmuir, Vol. 20, No. 18, 2004 7529

Figure 3. AFM images (5.0 × 5.0 µm2) and line profiles of the OTS-modified SiO2 surfaces (θOTS ) 0.81) obtained in the buffer solution (a) before and (b) after the deposition of 100-nm-filtered vesicles. The insertions show the surface roughness of the OTS domains (1.0 × 1.0 µm2). (c) Schematic illustration for the line profile in image b.

protrusions (indicated by circles in Figure 3b) higher by 0.5 ( 0.1 nm are observed in the OTS domains, but their assignment is not evident in these experiments. The height difference between the DPPC/OTS domains and the SiO2 regions is 1.2 ( 0.1 nm (Figure 3b). This value seems too small for the height of a DPPC/OTS layer if it is measured from the SiO2 surface (see Figure 2b). The DPPC in the SiO2 region could not be removed by scratching with the cantilever applying higher forces, and the precise thickness of the layers was not obtained. We suppose that the amount of DPPC molecules on the SiO2 regions is not sufficient to form a regular bilayer; therefore a smaller height is observed because of its irregular structure or compression by the cantilever. These interpretations are illustrated in Figure 3c: a DPPC monolayer forms on the OTS islands, but a regular DPPC bilayer does not form on the SiO2 regions. The results in Figures 2 and 3 reveal that the domain size on the surface largely affects the formation of the DPPC bilayer on the SiO2 regions. We have also investigated the effect of the vesicle size. SiO2 surfaces modified with networklike OTS domains (Figure 4a, θOTS ) 0.71) were incubated in the suspensions of vesicles with varied sizes. Figure 4b shows the OTS/SiO2 surface after the deposition of 200-nm-filtered vesicles. The regular DPPC bilayer membrane clearly does not form since the shape of the OTS islands remains clearly like the results in Figure 3b. The higher region is probably the DPPC/OTS layer, and its coverage is 0.67, which is almost the same as θOTS before the DPPC deposition. Some of the large protrusions (163 nm in diameter on average) are likely unruptured vesicles adsorbed on the surface. Figure 4c shows an OTS/SiO2 surface on which the sonicated vesicles (10-50 nm in diameter)37,38,42 are deposited. In stark contrast to the result of 200-nm-filtered vesicles in Figure 4b, a flat membrane covers the surface almost completely. This membrane is probably composed of the DPPC bilayers on the SiO2 regions and the hybridized DPPC/OTS layers on the OTS islands. The thickness of the membrane, which is acquired from the depth of the defects, is 4.3 ( 0.2 nm. The results in Figure 4 reveal that the size of the vesicles as well as that of the hydrophilic surface domains has a considerable effect on the bilayer formation on the SiO2 regions.

7530

Langmuir, Vol. 20, No. 18, 2004

Tero et al.

Figure 4. AFM images (5.0 × 5.0 µm2) of the OTS-modified SiO2 surfaces (θOTS ) 0.71) obtained in the buffer solution (a) before and after the deposition of (b) 200-nm-filtered vesicles and (c) sonicated vesicles.

4. Discussion 4.1. Morphology of OTS Islands. The most effective factor in the formation of OTS islands shown in Figure 1 is the degree of surface hydroxylation, which is controlled by the thermal oxidation and the H2O2/H2SO4 treatment. Effects of the amount of adsorbed water on OTS-SAM growth have been previously studied on dehydrated SiO2 surfaces.8,19-22 The adsorbed water layer on the surface works as a reactive medium for chemical reactions between silanol molecules and hydroxyl groups (OHs) on the surface as well as a medium for the diffusion of physisorbed silanol molecules. Dehydration of the SiO2 substrates by annealing at 150 °C prevented the growth of the closely packed OTS-SAM.8,21,22 An AFM study has shown that the OTS islands decrease in size with dehydration but retain the dendritical shape.19 It was attributed to the shortening of the diffusion length of the physisorbed OTS in the diffusion-limited island growth process caused by the diminution of the surface water layer. The decrease in the diffusion length due to the surface roughness has been also proposed to suppress the size and density of the dendritical OTS islands.23 In the dehydration experiments, annealing at 150 °C removes only the physisorbed water from the SiO2 surface and causes little change in the surface OHs.53,54 In the present study, however, the SiO2 substrate has been dehydroxylated irreversibly and almost completely by thermal oxidation at 700 °C53,54 and then partially rehydroxylated by H2O2/H2SO4 treatment. There has been no report on the distribution of surface OHs during the chemical oxidation by H2O2/H2SO4 to our knowledge, but the coverage and deployment of these OHs will certainly differ from those on a fully hydroxylated surface. These differences may lead to the circular shape and uniform size of the OTS islands. The absence of the small dot-type islands in Figure 1b-d indicates that the density of the reactive sites is low. This would increase the diffusion length of the physisorbed OTS molecules and change the island formation from a diffusion-limited process to a reaction-limited one, therefore leading to the circular island shape. The distribution and density of the surface OHs and their effects cannot be determined only from the AFM measurements reported here, but these factors will be quite important for the SAM formation. The amount of OTS molecules on the SiO2 regions between the OTS islands should also be discussed. Adsorption of OTS in the SiO2 region was not recognized by AFM, but a small amount of OTS molecules can change the hydrophilicity of the SiO2 surface and affect the adhesion energy of the vesicles. We conclude that almost bare SiO2 surface is exposed in the regions from the following reasons. The condition of the OTS/toluene solution is the same among all the samples in this study. (53) Chuang, I.-S.; Maciel, G. E. J. Phys. Chem. B 1997, 101, 3052. (54) Zhuravlev, L. T. Colloids Surf., A 2000, 173, 1.

Figure 5. Schematic drawings of the membrane formation from vesicles on (a) bare or sufficiently wide SiO2 surfaces, (b) a narrow SiO2 area between OTS islands, and (c) OTS-SAMs.

If the OTS molecules adsorbed on the SiO2 region, the small dot-type islands in Figure 1a, which originated from the polymerized OTS formed in the solution, would also be observed in Figures 1b-d, 2a, 3a, and 4a. The absence of the dot-type OTS islands indicates that the OTS molecules have rarely adsorbed on the region. We recently reported the results of the adhesion-mode imaging of an OTS/SiO2 surface prepared similarly to this study.49 Sharp contrast in the adhesive force mapping revealed that the amount of OTS molecules in the regions between the SAM islands is negligible or too small to change the surface hydrophilicity. Physisorbed OTS was probably removed completely by the repeated sonication after the deposition of OTS. 4.2. Size Effects of Surface Domains and Vesicles on Bilayer Formation. The formation of DPPC bilayer membranes is highly affected by the size of the surface domains (Figures 2 and 3) and vesicles (Figure 4). The DPPC bilayer membranes do not form on the SiO2 surfaces modified with the OTS islands when the exposed SiO2 regions are small (Figure 3b) or the size of the deposited vesicle is large (Figure 4b). These results indicate that the interrelation between the sizes of the vesicles and the surface domains is an influential factor. The membrane formation by the fusion of vesicles on a hydrophilic surface is a three-step process comprising adhesion, rupture, and spreading (Figure 5a).36,42 The important point is the adhesion process. Stable adhesion is necessary for the membrane formation. The bilayer membrane also forms through the process shown in Figure 5a on a SiO2 surface modified with OTS islands when sufficiently wide SiO2 regions are exposed as in the case of Figures 2b and 4c. When the exposed SiO2 regions are too narrow, on the other hand, OTS islands prevent stable adhesion of vesicles

Lipid Membrane Formation by Vesicle Fusion

due to the hydrophilic-hydrophobic repulsion (Figure 5b), and consequently the bilayer membrane does not form like the result in Figures 3b and 4b. Jass et al. reported that edge-to-edge contact of adsorbed vesicles accelerates the bilayer formation.43 Interruption of vesicle adhesion by the OTS islands would not only prevent stable adhesion but also cause the intervesicle process to occur less frequently. It has been reported that a smaller vesicle transforms to a bilayer membrane more easily than a larger one because of a greater surface tension.36,42 However, we have confirmed on a SiO2 surface without an OTS-SAM that the 200-nm-filtered vesicles form the bilayer membrane after the same incubation time (2 h) as Figure 4b and that its morphology is not largely different from that made from the sonicated vesicles. Therefore we eventually attributed the size effect observed in Figure 4 to the influence of the OTS islands, not to the surface tension of the vesicles. The OTS islands will be covered with a DPPC monolayer to form a DPPC/OTS hybridized bilayer during the incubation in the vesicle suspension, since it is unstable to expose the hydrophobic OTS island surfaces to the aqueous solution. Lipid monolayers form on hydrophobic SAM surfaces in several systems, as mentioned in the previous section.32,34,35,37,47 In the present study, the change in the roughness on the OTS domains induced by the DPPC deposition (Figure 3a,b) also implies the DPPC monolayer formation on the OTS islands, although the direct detection of the monolayer is not achieved. The DPPC monolayer on the OTS islands may form irrespective of the size of OTS islands, since it is proposed that spreading of the lipid monolayer simultaneously proceeds with the adhesion process on a hydrophobic surface (Figure 5c).39 The height of the DPPC/OTS hybridized bilayer is smaller by 0.7 ( 0.2 nm than that of the DPPC bilayer (Figure 2), while the length of the OTS alkyl chain (C18H37) is almost equivalent to that of the DPPC acyl chain (OCOC15H31). This is mainly due to the large headgroup of the DPPC molecule, which corresponds to 0.9 nm thickness.55 In addition, there is a thin water layer between the DPPC bilayer and the substrate surface,5,45,46 whereas the OTSSAM is directly bound on the substrate. It has also been reported that DPPC molecules incline ∼36° from the surface normal in the monolayer on the OTS-SAM deposited by the LB method.33 The height of the AFM images is largely affected by the cantilever and applied force, especially on soft materials such as lipid membranes. In this study, we used the same type of cantilever to observe all the images and the uppermost layer of both area A and area B in Figure 2b is assumed to be a DPPC layer. Therefore, it is qualitatively reasonable that the DPPC/OTS layer is observed to be lower than the DPPC bilayer. Maybe the difference in the interaction between the OTS-DPPC and DPPC-DPPC is also related. The result in Figure 3 denotes that a small amount of DPPC exists on the SiO2 regions even though a regular bilayer membrane does not form. If we assume the thickness of the DPPC/OTS layer in Figure 3b is the same as that in Figure 2b (3.9 nm), the thickness of the thin DPPC layer on the SiO2 region is estimated at 2.7 nm. This value is too small for a regular lipid bilayer membrane. We suppose that excess DPPC molecules slide (55) Nagle, J. F.; Tristram-Nagle, S. Biochim. Biophys. Acta 2000, 1469, 159.

Langmuir, Vol. 20, No. 18, 2004 7531

on during the DPPC/OTS layer formation or that part of the DPPC molecules in the vesicle adsorb and remain on the SiO2 surface when the vesicles touch on the SiO2 surface. They will assemble not to expose the acyl chains to the water phase. Accurate determination and control of the size and amount of vesicles and the size of surface domains will be necessary to avoid the adsorption of the excess DPPC. We temporarily assign this layer to the disordered DPPC layer in Figure 3c. It seems strange that DPPC does not form a normal gel phase bilayer at RT since the Tm of DPPC is 41 °C. However, if an insufficient amount of lipid molecules on the SiO2 region gather to form a normal bilayer, defects would appear in the membrane. The uniform and thinner layer observed in Figure 3b is inconsistent with the formation of a normal gel phase bilayer. We cannot clarify the structure and the orientation of DPPC in the layer at present, but lateral interaction from the OTS islands might extend the DPPC layer on the narrow SiO2 regions as in Figure 3. Jenkins et al. have investigated the fusion of vesicles on microcontact-printed mercaptoethanol-SAMs on gold substrates and reported that the vesicles on the hydrophilic areas show high reactivity within the 2 µm region from the edge of the hydrophobic SAM areas.47 The hydrophilic domain sizes in the present study are of submicron scale, and no difference in the membrane thickness or morphology is observed between the edge and the center of the domains. In the previous studies, membrane formation by the vesicle fusion method has been investigated on homogeneous surfaces5,30-35,37-46 or sufficiently large areas even on micropatterned surfaces.47,48 In this study, we have clearly revealed that surface domains with sizes comparable to those of vesicles considerably affect the membrane formation processes by the vesicle fusion. The size of vesicles effective to the membrane formation is of the order of tens to hundreds of nanometers, which is just the target of the nanofabrication and patterning. The relative size of vesicles and hydrophilic surface domains becomes one of the important factors in the fusion process of vesicles in this scale. 5. Conclusion We have investigated the formation of DPPC bilayer membranes on SiO2 surfaces modified with OTS islands on the scale of hundreds of nanometers. Thermal oxidation and subsequent H2O2/H2SO4 treatment of the SiO2 substrate make it possible to prepare the OTS islands with circular shape and uniform size. The formation of the DPPC bilayer membrane by the vesicle fusion method on the OTS/SiO2 surfaces largely depends on the relative size between the vesicles and surface domains. Vesicles that are sufficiently smaller than the SiO2 regions stably adhere on the SiO2 surface and form a bilayer, whereas the adhesion of larger vesicles is prevented by the OTS islands. On nanopatterned surfaces, the relative size between the vesicles and surface domains is involved as an important factor in the process of membrane formation by the fusion of vesicles. Acknowledgment. This research was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture and by a Collaboration Program of the Graduate University for Advanced Studies. LA0400306