pubs.acs.org/Langmuir © 2009 American Chemical Society
Nondisruptive Micropatterning of Fluid Membranes through Selective Vesicular Adsorption and Rupture by Nanotopography Sang-Wook Lee, Yu-Jin Na, and Sin-Doo Lee* School of Electrical Engineering #032, Seoul National University, Kwanak P.O. Box 34, Seoul 151-600, South Korea Received December 13, 2008. Revised Manuscript Received March 16, 2009 We report on a nondisruptive method of patterning fluid membranes into micrometer-scale arrays through a selective vesicular rupture pathway by nanotopography. The site- and pathway-selective formation of supported lipid bilayers (SLBs) was achieved by different vesicular adsorption and rupture processes between nanocorrugated and nanosmooth topographies. The SLBs were first developed in the nanocorrugated region due to fast vesicular adsorption and then grew into the nanosmooth region through bilayer edge-induced vesicular rupture. Our topographic approach provides a viable scheme, yet unattainable in conventional ways, of actively controlling the position and the coverage of the SLBs on a variety of substrates without disrupting two-dimensional fluidity for highly integrated membrane devices.
Lipid bilayers, possessing the intrinsic two-dimensional (2-D) fluidity, resemble cell membranes that serve as essential ingredients for transport of biological components such as ions and molecules as well as for cellular communications in living organisms. The formation of such lipid bilayers on various substrates, so-called supported lipid bilayers (SLBs), has been a primary route to the fundamental studies of membrane activities1,2 including intermembrane interactions.3 Recently, as important applications, membrane-based biocompatible devices have been demonstrated for analyzing specific binding events at membrane interfaces in either an optical or an electrical detection scheme with high sensitivity.4,5 For highly integrated membrane devices in use for cell-based assay6 and high throughput drug screening,7 a nondisruptive way of patterning fluid membranes is inevitably required on a variety of substrates. Several methods of directly patterning the SLBs, grown entirely on flat substrates, are known to be mechanical scratching,8 microcontact printing,9 photolithography,10 and photopolymerization.11 One interesting approach is based on the modification of the surface chemistry with a metal,12 a photoresist,13 and a self-assembled monolayer14 having hydrophilic/hydrophobic *To whom correspondence should be addressed. E-mail: sidlee@ plaza.snu.ac.kr. Telephone: +82-2-880-1823. Fax: +82-2-874-9769. (1) Parthasarathy, R.; Yu, C.; Groves, J. T. Langmuir 2006, 22, 5095–5099. (2) Yoon, T.-Y.; Jeong, C.; Lee, S.-W.; Kim, J. H.; Choi, M. C.; Kim, S.-J. Kim, M. W.; Lee, S.-D. Nat. Mater. 2006, 5, 281–285. (3) Parthasarathy, R.; Groves, J. T. Proc. Natl. Acad. Sci.U.S.A. 2004, 101, 12798–12803. (4) Zhou, X.; Moran-Mirabal, J. M.; Craighead, H. G.; McEuen, P. L. Nature Nanotechnol. 2007, 2, 185–190. (5) Brozell, A. M.; Muha, M. A.; Sanii, B.; Parikh, A. N. J. Am. Chem. Soc. 2006, 128, 62–63. (6) Mossman, K. D.; Campi, G.; Groves, J. T.; Dustin, M. L. Science 2005, 310, 1191–1193. (7) Fang, Y.; Frutos, A. G.; Lahiri, J. J. Am. Chem. Soc. 2002, 124, 2394–2395. (8) Groves, J. T.; Boxer, S. G.; McConnell, H. M. Proc. Natl. Acad. Sci.U.S.A. 1997, 94, 13390–13395. (9) Hovis, J. S.; Boxer, S. G. Langmuir 2000, 16, 894–897. (10) Yee, C. K.; Amweg, M. L.; Parikh, A. N. Adv. Mater. 2004, 16, 1184–1189. (11) Morigaki, K.; Kiyosue, K.; Taguchi, T. Langmuir 2004, 20, 7729–7735. (12) Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651–653. (13) Groves, J. T.; Ulman, N.; Cremer, P. S.; Boxer, S. G. Langmuir 1998, 14, 3347–3350. (14) Han, X.; Critchley, K.; Zhang, L.; Pradeep, S. N. D.; Bushby, R. J.; Evans, S. D. Langmuir 2007, 23, 1354–1358.
Langmuir 2009, 25(10), 5421–5425
properties prior to the SLB formation. These existing methods may involve physical disruption of the SLB at edges or chemical disturbances to the SLB structure on the substrates. Moreover, it is not simple for them to produce micropatterned membranes with undulation, mimicking biological membranes5,15 for structural organization of biochemical components. The membrane curvature is known to play a critical role in the phase separation between liquid-ordered (lo) and liquid-disordered domains1,16 as well as spatial reorganization of the lo domains2 and lipid molecules.17 Therefore, a simple and reliable method of producing micropatterned, undulated membranes in an array format is extremely important for in vitro studies of various curvaturemediated membrane activities including cell-cell interactions and cell signal pathways. In this work, we demonstrate a nanotopographic approach to patterning of fluid membranes into a micrometer-scale array through a selective vesicular rupture pathway on a single solid support. The key concept behind this is to generate different vesicular adsorption and rupture pathways of the vesicles between two different, nanocorrugated and nanosmooth, topographies that result in different elastic energy landscapes. Figure 1a shows the schematic diagram demonstrating our idea of producing micropatterned SLBs with undulation through different pathways of vesicular adsorption and rupture on the two surface topographies. In principle, the vesicular adsorption and rupture pathways are dictated by the competition between the vesicular curvature elastic energy and the adhesion energy on the substrate.18 Consider that a morphological domain in the nanocorrugated region (left in Figure 1b) has lateral (l ) and vertical (h) dimensions on the order of a vesicle size (d ) while that in the nanosmooth region (middle and right in Figure 1b) is flat in relative to d on a nanometer scale. Under these circumstances, individual vesicles will be preferentially adsorbed onto the nanocorrugated region without experiencing large elastic distortions at two edges (Figure 1b), and therefore, the vesicular adsorption rate (15) Sanii, B.; Smith, A. M.; Butti, R.; Brozell, A. M.; Parikh, A. N. Nano Lett. 2008, 8, 866–871. (16) Rozycki, B.; Weikl, T. R.; Lipowsky, R. Phys. Rev. Lett. 2008, 100, 098103. (17) Liang, Q.; Ma, Y.-Q. J. Phys. Chem. B 2008, 112, 1963–1967. (18) Seifert, U. Adv. Phys. 1997, 46, 13–137.
Published on Web 4/15/2009
DOI: 10.1021/la8041082
5421
Letter
Figure 1. Schematic diagrams showing (a) a conceptual representation of producing micropatterned SLBs with undulation, (b) vesicular adsorption in a nanocorrugated region (left) and nanosmooth regions with different adhesion energies (middle and right), and (c) rupture pathways of vesicles in nanocorrugated (left) and nanosmooth (right) regions. The characters of V and B represent the vesicle and the bilayer, respectively. The size of a single vesicle is denoted by d. The lateral and vertical dimensions of a morphological domain are l and h, respectively.
will become high. In contrast, the vesicular adsorption rate in the nanosmooth region will be relatively low, since the adhesion area is relatively small (middle in Figure 1b) or high elastic distortions at the two edges are not energetically favorable (right in Figure 1b). This difference in the vesicular adsorption rate between the two topographies will lead to different time scales and pathways in vesicular rupture. Among several possible pathways,19 a plausible scenario is that the SLB formation in the nanocorrugated region is most likely initiated through the cooperative effects of neighboring vesicles resulting from intervesicular (V-V) interactions (left in Figure 1c) at the critical vesicle coverage.19,20 In the nanosmooth regions, the bilayer edge effects21 resulting from the vesicle-bilayer edge (V-B) interactions (right in Figure 1c) are expected to primarily contribute to the SLB formation below the critical vesicle coverage. As a result, the SLB formation will be first completed in the nanocorrugated region and then promoted into the nanosmooth region. Note that a simple geometrical argument of l ∼ d (preferably h ∼ d) provides a criterion for the site-selective formation of the SLB as shown in Figure 1. We first determine the relative adsorption of phospholipidbased vesicles (d = 50 nm) and the resultant SLB formation in three different types of topographies (see Experimental Section): nanosmooth (l and h , d, Figure 2a), microporous (l ≈ 0.5 μm and h ≈ 0.1 μm, Figure 2b), and nanocorrugated (l ≈ 50 nm and h ≈ 5-10 nm, Figure 2c). In fact, l can be treated as a primary geometrical parameter as long as h is not zero. From the epifluorescence micrographs taken at the same time of duration (t = 1 min) for the substrate exposure to the vesicle solution (Figures 2d-f), it is clear that the intensity of red fluorescence in the nanosmooth substrate (Figure 2d) is the weakest, indicating (19) Richter, R. P.; Brisson, A. R. Biophys. J. 2005, 88, 3422–3433. :: (20) Keller, C. A.; Glasmastar, K.; Zhdanov, V. P.; Kasemo, B. Phys. Rev. Lett. 2000, 84, 5443–5446. (21) Hamai, C.; Cremer, P. S.; Musser, S. M. Biophys. J. 2007, 92, 1988–1999.
5422
DOI: 10.1021/la8041082
that the vesicular adsorption rate is the lowest. At t = 1 min, the relative intensities of the background-subtracted fluorescence in Figure 2e and f to the intensity in the nanosmooth topography in Figure 2d were 2.24 ( 0.18 and 2.38 ( 0.20, respectively. At t = 10 min, although the relative fluorescence intensity was increased by about twice (2.07 ( 0.15) in the nanosmooth substrate, the fluorescence recovery after photobleaching22 (FRAP, enclosed by white circles) study still showed no 2-D fluidity. This implies that ruptured vesicles are negligible below the critical vesicle coverage. In the microporous substrate (Figure 2e), 2-D fluidity was observed from FRAP at t = 5 min. The relative fluorescence intensity and the diffusion coefficient (D) were found to be 2.73 ( 0.18 and 0.98 ( 0.41 μm2/s, respectively. In the nanocorrugated substrate (Figure 2f), 2-D fluidity appeared even at t = 1 min (D = 0.81 ( 0.12 μm2/s). These experimental results definitely support that morphological domains in the nanocorrugated topography, comparable to the vesicle size (Figure 2c), will facilitate the vesicular adsorption and the resultant SLB formation at high vesicle coverage. It was found that a lipid mixture having net positive charges shows no qualitative difference in the vesicular adsorption (Supporting Information, Figure S1). This suggests that the topographic effect plays a primary role in vesicular adsorption in our case. Based on the above findings, we prepared binary patterns of the nanocorrugated (label 1 in Figure 3a) and nanosmooth (label 2 in Figure 3a) topographies on a (100) quartz wafer whose top surface was covered with a SiO2 layer of 1.5 μm thick (Figure 3a; scanning electron microscopy (SEM) image). The SiO2 layer was produced by chemical vapor deposition of tetraethoxysilane (TEOS). Atomic force microscopy (AFM) measurements show that the bare SiO2 surface gives the nanocorrugated topography23 (Figure 3b) as described in Figure 2c. The nanosmooth region (Figure 3c) was obtained by chemically etching the nanocorrugated region by 1 μm in depth. The root-mean-square values of the surface roughness in the nanocorrugated region and that in the nanosmooth region were 2.7 and 0.5 nm, respectively. The prepared substrate was exposed to a phospholipid-based vesicle solution to promote SLB formation through vesicular adsorption and rupture. Let us now examine how the SLB is formed in the nanocorrugated region and is dynamically grown into the nanosmooth region. Due to the difference in the vesicular adsorption and rupture pathway between the two topographies, the temporal growth of the SLB needs to be precisely controlled for constructing a variety of membrane microarrays with high resolution. Figure 3d shows the SLB growth into each nanosmooth region, having a square shape (label 2), at a step of 1 min from the surrounding nanocorrugated region (label 1) where the SLB was fully covered for t = 1 min. The SLB formation through the V-V interactions in nanocorrugated region until t = 1 min is discussed in the Supporting Information, Figure S2. Clearly, the SLB began to grow from the four bilayer edges with increasing exposure time. After about 5 min of exposure, the SLBs were fully grown into the nanosmooth region. This definitely supports that, in the nanosmooth regions, the vesicular rupture takes place mostly at the bilayer edges in the regime of low vesicular coverage. During the SLB growth, there exists a density gradient of fluorescencelabeled lipid near the SLB boundary which was seen in Figure 3. After the equilibration of the lipid density, the fluorescence intensity is higher in region 1 than in region 2 due to the larger (22) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Biophys. J. 1976, 16, 1055–1069. :: (23) Elsholz, F.; Scholl, E.; Scharfenorth, C.; Seewald, G.; Eichler, H. J.; Rosenfeld, A. J. Appl. Phys. 2005, 98, 103516.
Langmuir 2009, 25(10), 5421–5425
Letter
Figure 2. Vesicular adsorption and resultant SLB formation on three different types of topographies: (a) AFM image of the nanosmooth topography having l , d (h ≈ 0), (b) SEM image of the microporous topography having l ≈ 0.5 μm (h ≈ 0.1 μm), and (c) AFM image of the nanocorrugated topography having l ≈ 50 nm (h ≈ 5-10 nm). Inset in (c): typical edge form of a morphological domain. Epifluorescence micrographs showing the FRAP data taken at substrate exposure of (d) t = 1 and 10 min in the nanosmooth substrate, (e) t = 1 and 5 min in the microporous substrate, and (f) t = 1 min in the nanocorrugated substrate. Photobleached spots are enclosed by white circles.
surface area of the SLB in region 1 (Supporting Information, Figure S3). The real-time observation of the SLB formation in this regime allows us to determine the circumference change and the SLB coverage resulting from the V-B interactions. Since the SLB coverage is directly related to the change of the edge circumference l(t), the time-dependent coverage of the SLB in region 2 can be obtained by integrating l(t) over time under the condition that the initial coverage is zero at t = 1 min. From the SLB growth in Figure 3a, the measured edge circumference in time was found to be represented as a straight line, l(t) = -gt + l0 (t > 1 min), as shown in Figure 3b. Here, g and l0 denote the rate of the circumference change and a constant at t = 0, respectively. The least-squares fit of the experimental data to a straight line gives g = 224 μm/min and l0 = 1110 μm. For t = 1 min, l(1) = 886 μm, which corresponds to the initial circumference of a single nanosmooth pattern (about 800 μm). The gradual coverage of the SLB in region 2 agrees well with the theoretical prediction, made by integrating l(t) over time under the constraint of the full coverage at t = 5 min, as shown in Figure 3c. Note that, in our case, the exposure time is the only parameter to control the SLB coverage in region 2 for given topographic configuration. Based on two different vesicular rupture pathways in the nanocorrugated and nanosmooth topographies described above, we present a general approach to construction of micropatterned membrane arrays having undulated SLB patterns. Basically, a crossover between two different regimes, having different vesicular rupture pathways, exists in our topographic configuration. The crossover comes from the interplay between the V-V and V-B interactions for the SLB formation. More specifically, for a given nanocorrugated region of 40 μm 200 μm (a rectangle enclosed by a white line in Figure 4a), the SLB was developed through the intervesicular rupture at the crossover exposure time of tc = 60 s, and it grew into a nanosmooth region (an ellipse enclosed by a white line in Figure 4b) through the bilayer edge induced vesicular rupture on further exposure (t = 100 s) beyond tc. During the subsequent exposure time of 40 s, in each nanosmooth pattern, the additional coverage of the SLB in an elliptical Langmuir 2009, 25(10), 5421–5425
shape was measured as approximately 7000 μm2 (Figure 4b) using an image analyzing program. Although the boundary of the bilayer edge is irregular (Supporting Information, Figure S4), the SLB takes an elliptic shape macroscopically due to the uniform growth from the four edges.24 With the help of g = 224 μm/min obtained from Figure 3b, the predicted value is given as 8138 μm2 under the assumption that the SLB grows uniformly along each side of the nanosmooth pattern. A theoretical model taking into account the line tension at bilayer edges can describe more quantitatively the SLB coverage which results from the V-B interactions. For implementing the concept of our approach into potential applications including biosensors and protein chips, we fabricated a membrane microarray with several different types of resolution ranging from 10 to 40 μm. Specific binding events of proteins on the patterned SLB are demonstrated in our microarray. Figure 4c shows a patterned membrane microarray where the SLB with undulation was selectively grown in nanocorrugated regions (label 1) during the crossover exposure time tc. The width of each rectangle was varied from 10 to 40 um, being able to reach a few micrometers. In Figure 4d, the strong green fluorescence indicates that specific binding events occur between the streptavidin molecules, conjugated with Alexa Fluor 488, and the biotylated SLBs formed mostly in region 1. In summary, we have presented a nanotopographic methodology of micropatterning fluid membranes through a selective vesicular adsorption and rupture pathway for use in fabricating diverse membrane arrays such as biochips and biosensors. In the binary, nanosmooth, and nanocorrugated topographic configuration, the interplay between the V-V and V-B interactions dictates the SLB formation in the site- and pathway-selective manner. The SLB was found to be developed first in the nanocorrugated region through the V-V interactions at the critical vesicle coverage owing to the fast vesicular adsorption. Recently, an interesting study on the effect of the phospholipid (24) Furukawa, K.; Sumitomo, K.; Nakashima, H.; Kashimura, Y.; Torimitsu, K. Langmuir 2007, 23, 367–371.
DOI: 10.1021/la8041082
5423
Letter
Figure 4. Membrane microarrays showing two regimes of SLB formation and specific binding events of proteins: (a) SLB pattern grown in a rectangular shape in the nanocorrugated region for the crossover exposure time of tc = 60 s, (b) SLB pattern grown in an elliptical shape into the nanosmooth region for t = 100 s, (c) membrane microarray with lateral resolutions of 40, 20, and 10 μm for tc, and (d) selective binding of streptavidin molecules with biotylated SLBs in the array (c).
Fabrication of Substrates. For producing the microporous topography, a thin aluminum (Al) layer of 100 nm in thickness was first coated on a (100) quartz wafer. Using plasma etching equipment (AOE, STS, U.K.), plasma of C4F8 and O2 was produced at a radio frequency of 13.56 MHz. During plasma etching for 3 min, the Al-covered substrate was etched irregularly due to the Al clusters. The porous topography used in our study was finally obtained by further chemically etching with a buffer solution of HF (7:1 (v/v) NH4F/HF) for 30 s. For producing the nanocorrugated topography, a SiO2 layer of 1.5 μm was prepared by chemical vapor deposition of TEOS (P-5000, Applied
Material, Korea). The deposition rate was 12.5 nm/s. For obtaining binary topographic patterns, the nanosmooth topography was produced by chemically etching the nanocorrugated region by 1 μm in depth using a buffer solution of HF (7:1 (v/v) NH4F/HF). The etching rate was approximately 100 nm/min. For selective etching, the photolithographic process using a positive photoresist (AZ1512, AZ Electronic Materials) was used for producing a variety of patterns. The nanotopographic surfaces were characterized by SEM (XL30FEG, Philips) and AFM (XE-150, PSIA and SPA-400, SEIKO instrument) measurements. Substrate Cleaning. The substrate was cleaned with piranha solution (3:1 (v/v) H2SO4/H2O2) at 120 °C for more than 10 min, followed by ultrasonication in deionized (DI) water for 10 min before the formation of the SLBs. Forming and Imaging the SLBs. The phospholipid, used as a base for the formation of the SLBs, was 1,2-dioleoyl-sn-glycero3-phophocholine (DOPC, Avanti Polar Lipids, Birmingham, AL). For imaging the SLBs, a red fluorescent dye labeled lipid, 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (Texas Red-DHPE, Molecular Probes, Eugene, OR) was mixed with DOPC at 1 mol %. All lipids were dissolved in chloroform. The rapid solvent exchange method26 was employed for evaporation of chloroform and hydration with Tris buffer (100 mM NaCl and 10 mM Tris at pH 8.0) simultaneously. The total concentration of a lipid mixture was 0.2 mg/mL. Small unilamellar vesicles (SUVs) were produced by extruding (Mini-Extruder, Avanti Polar Lipids) 20 times through a 50 nm filter. When the cleaned substrate was incubated in the SUV solution, the formation of the SLBs was developed via vesicle adsorption, rupture, and fusion. The SUVs were removed with DI water after the SLB was formed for the duration of the substrate exposure time. The Texas red channel was monitored using an epifluorescence microscope (Eclipse E600-POL, Nikon) during the formation of the SLBs. The measurements of the edge circumference and the SLB coverage were carried out using an image analyzing program of i-Solution (IMT, Korea). Streptavidin-Biotin Binding. For the streptavidin-biotin binding process, a phospholipid mixture of DOPC and Texas Red-DHPE was additionally doped with 1,2-dipalmitoyl-snglycero-3-phosphoethanolamine-N-(cap biotinyl), sodium salt
(25) Hamai, C.; Yang, T.; Kataoka, S.; Cremer, P. S.; Musser, S. M. Biophys. J. 2006, 90, 1241–1248.
(26) Buboltz, J. T.; Feigenson, G. W. Biochim. Biophys. Acta 1999, 1417 232–245.
Figure 3. Time evolution of the SLB formation in binary topographic patterns. (a) SEM micrograph showing the cross section of the nanotopographic substrate. Labels 1 and 2 represent the nanocorrugated and nanosmooth regions, respectively. AFM measurements in (b) the nanocorrugated topography and (c) the nanosmooth topography. (d) Epifluorescence micrographs in realtime observation of the SLB grown from the bilayer edges. The scale bar is 200 μm. (e) Edge circumference of the SLB (filled circles) determined from (d) and least-squares fit to a straight line l(t) as a function of time. (f) SLB coverage in region 2 (filled circles) measured from (d) and theoretical curve obtained by integration of l(t) over time. Error bars represent one standard deviation.
geometry within a vesicle on the SLB formation has been carried out to discuss the importance of elasticity-based, vesicular interactions with substrates in the SLB formation.25 The topographic approach presented here would provide a robust platform for fundamental studies of curvature-mediated membrane activities as well as the development of highly integrated biocompatible devices.
Experimental Section
5424
DOI: 10.1021/la8041082
Langmuir 2009, 25(10), 5421–5425
Letter (biotin-DPPE, Avanti Polar Lipids) at 1 mol %. The SLBs in the microarray shown in Figure 4c were incubated in a buffer solution (phosphate buffered saline at pH 7.2) mixed with Alexa Fluor 488 conjugated streptavidin (Molecular Probes) at 5 μg/mL for 30 min in the dark environment. The temperature was maintained at 36.5 °C during the incubation.
Acknowledgment. This work was supported in part by Zema International and Jein Medical Center, Korea. We would
Langmuir 2009, 25(10), 5421–5425
like to thank W. Lee for his technical assistance in the AFM measurements.
Supporting Information Available: Adsorption of vesicles having net positive charges, FRAP studies on the SLB formation at t = 1 min in the nanocorrugated region, fluorescence intensity change with the lipid density in region 2, and boundary of the bilayer edge during the SLB growth. This material is available free of charge via the Internet at http://pubs.acs.org.
DOI: 10.1021/la8041082
5425
