Micropatterned Composite Membranes of Polymerized and Fluid Lipid Bilayers

Transcription

1 Langmuir 2004, 20, Micropatterned Composite Membranes of Polymerized and Fluid Lipid Bilayers Kenichi Morigaki,*, Kazuyuki Kiyosue, and Takahisa Taguchi Special Division for Human Life Technology, National Institute of Advanced Industrial Science and Technology (AIST), Ikeda , Japan Received March 15, In Final Form: May 19, 2004 Micropatterned composite membranes of polymerized and fluid lipid bilayers were constructed on solid substrates. Lithographic photopolymerization of a diacetylene-containing phospholipid, 1,2-bis(10,12- tricosadiynoyl)-sn-glycero-3-phosphocholine (DiynePC), and subsequent removal of nonreacted monomers by a detergent solution (0.1 M sodium dodecyl sulfate (SDS)) yielded a patterned polymeric bilayer matrix on the substrate. Fluid lipid bilayers of phosphatidylcholine from egg yolk (egg-pc) were incorporated into the lipid-free wells surrounded by the polymeric bilayers through the process of fusion and reorganization of suspended small unilamellar vesicles. Spatial distribution of the fluid bilayers in the patterned bilayer depended on the degree of photopolymerization that in turn could be modulated by varying the applied UV irradiation dose. The polymeric bilayer domains blocked lateral diffusion of the fluid lipid bilayers and confined them in the defined areas (corrals), if the polymerization was conducted with a sufficiently large UV dose. On the other hand, lipid molecules of the fluid bilayers penetrated into the polymeric bilayer domains, if the UV dose was relatively small. A direct correlation was observed between the applied UV dose and the lateral diffusion coefficient of fluorescent marker molecules in the fluid bilayers embedded within the polymeric bilayer domains. Artificial control of lateral diffusion by polymeric bilayers may lead to the creation of complex and versatile biomimetic model membrane arrays. 1. Introduction Since the seminal work by Irvine Langmuir, artificial model membranes have played important roles in the development of our understanding on the structure and function of the biological membrane. 1-4 Such model systems include Langmuir monolayers, planar bilayers (black lipid membranes), and liposomes (lipid vesicles). Substrate-supported planar bilayers (SPBs) are a relatively new breed of model membranes that were introduced in the 1980s. 5-7 They typically comprise a single lipid bilayer adsorbed on the solid surface by physical interactions or chemical bonds. In some cases, lipid monolayers are adsorbed on the hydrophobic surface of self-assembled monolayers (alkane thiol or alkylsilane), forming hybrid bilayer membranes SPBs have some unique features compared with other formats of model membranes, including mechanical stability (in contrast to black lipid membranes) and the accessibility to various optical and electrochemical analytical techniques that can detect interfacial events with an extremely high sensitivity. These features render SPBs highly attractive for the development of devices that utilize artificially mimicked cellular functions. 7,11 * Corresponding author. Fax: morigaki-kenichi@aist.go.jp. Current affiliation: Research Institute for Cell Engineering, AIST. (1) Langmuir, I. J. Am. Chem. Soc. 1917, 39, (2) Ringsdorf, H.; Schlarb, B.; Venzmer, J. Angew. Chem., Int. Ed. 1988, 27, (3) Chapman, D. Langmuir 1993, 9, (4) Edidin, M. Nat. Rev. Mol. Cell Biol. 2003, 4, (5) Brian, A. A.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, (6) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, (7) Sackmann, E. Science (Washington, D.C.) 1996, 271, (8) Plant, A. Langmuir 1993, 9, (9) Plant, A. Langmuir 1999, 15, (10) Parikh, A. N.; Beers, J. D.; Shreve, A. P.; Swanson, B. I. Langmuir 1999, 15, Another important feature of SPBs is the potential to generate micropatterned membranes on the substrate. This aspect allows the creation of designed microarrays of model membranes and should facilitate various new applications, such as high throughput drug screening There have been several lines of micropatterning approaches reported to date, including the use of mechanical scratching, 13 prepatterned substrates, 12,14-17,20 microcontact printing, 19 microfluidics, 18,21 inkjet printers, 22 and liftoff of prepatterned polymer films. 25 Photolithographic deep UV decomposition of SPBs has also been utilized to create micropatterned lipid-free voids. 26 We have recently reported a novel micropatterning method of SPBs based on the lithographic photopolymerization of a diacetylene (11) Cornell, B. A.; Braach-Maksvytis, V. L. B.; King, L. G.; Osman, P. D. J.; Raguse, B.; Wieczorek, L.; Pace, R. J. Nature (London) 1997, 387, (12) Groves, J. T.; Ulman, N.; Boxer, S. G. Science (Washington, D.C.) 1997, 275, (13) Groves, J. T.; Boxer, S. G.; McConnell, H. M. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, (14) Wiegand, G.; Jaworek, T.; Wegner, G.; Sackmann, E. Langmuir 1997, 13, (15) Heyse, S.; Ernst, O. P.; Dienes, Z.; Hofmann, K. P.; Vogel, H. Biochemistry 1998, 37, (16) Jenkins, A. T. A.; Boden, N.; Bushby, R. J.; Evans, S. D.; Knowles, P. F.; Miles, R. E.; Ogier, S. D.; Schönherr, H.; Vancso, G. J. J. Am. Chem. Soc. 1999, 121, (17) Cremer, P. S.; Yang, T. J. Am. Chem. Soc. 1999, 121, (18) Kam, L.; Boxer, S. G. J. Am. Chem. Soc. 2000, 122, (19) Hovis, J. S.; Boxer, S. G. Langmuir 2000, 16, (20) Srinivasan, M. P.; Ratto, T. V.; Stroeve, P.; Longo, M. L. Langmuir 2001, 17, (21) Künneke, S.; Janshoff, A. Angew. Chem., Int. Ed. 2002, 41, (22) Fang, Y.; Frutos, A. G.; Lahiri, J. J. Am. Chem. Soc. 2002, 124, (23) Groves, J. T.; Boxer, S. G. Acc. Chem. Res. 2002, 35, (24) Orth, R. N.; Kameoka, J.; Zipfel, W. R.; Ilic, B.; Webb, W. W.; Clark, T. G.; Craighead, H. G. Biophys. J. 2003, 85, (25) Rehfeldt, F.; Tanaka, M. Langmuir 2003, 19, (26) Yee, C. K.; Amweg, M. L.; Parikh, A. N. Adv. Mater., in press /la049340e CCC: $ American Chemical Society Published on Web 08/04/2004

2 7730 Langmuir, Vol. 20, No. 18, 2004 Morigaki et al. bilayer. After the lithographic photopolymerization, nonreacted monomeric DiynePC was removed by a detergent solution (0.1 M sodium dodecyl sulfate (SDS)) and new lipid bilayers of phosphatidylcholine from egg yolk (egg- PC) containing a fluorophore were incorporated by the vesicle fusion technique. 5,32-35 The experimental observations showed unambiguously that fluid and polymerized bilayer domains were forming an integrated composite bilayer membrane. Depending on the applied UV irradiation dose, the incorporated lipid bilayers were either confined within the defined geometries or penetrating into the polymeric bilayer domains; that is, whereas DiynePC bilayers polymerized with a large UV irradiation dose blocked lateral diffusion of the fluid bilayers, those polymerized with a smaller UV dose allowed egg-pc and marker molecules to penetrate. These results suggest that it is basically possible to modulate the lateral mobility of membrane associated molecules by using polymeric bilayers with a defined degree of polymerization. This possibility, in turn, should open a new avenue for the fabrication of complex biomimetic membrane systems that can model the lateral organization of lipids and proteins within the cellular membrane. We briefly describe the concept at the end of this report. Figure 1. (A-D) Schematic outline of the patterning procedure. (E) The polymerizable diacetylene phospholipid DiynePC and the polymerization scheme. phospholipid (1,2-bis(10,12-tricosadiynoyl)-sn-glycero-3- phosphocholine (DiynePC)) The fabrication process is composed of four steps (a schematic illustration is given in Figure 1): (A) formation of a monomeric bilayer on a solid substrate, (B) lithographic photopolymerization by UV light, (C) removal of the protected monomeric bilayers, and (D) refilling the lipid-free wells with new lipid bilayers. The lipid bilayers incorporated in the last step retain some characteristic features of native cellular membranes (e.g., lateral fluidity) and are intended to be used for further biological applications. One distinctive feature of the current micropatterning strategy is the fact that the patterns are imprinted in the bilayer membrane as polymeric bilayer domains. The integration of polymerized and fluid bilayers as a single composite membrane has various potential advantages for the functional incorporation of membrane proteins. For example, one can separate the bilayer membrane from the substrate by using a thin layer of hydrophilic polymer spacer. 30,31 One could also expect certain mechanical stabilization of otherwise unstable lipidic membranes by the matrix of polymeric bilayers. Herein, we report on another unique feature of the current micropatterning strategy, namely, the control of the bilayer photopolymerization process by varying the applied UV irradiation dose that results in the controlled lateral diffusion of lipid molecules within the polymeric (27) Morigaki, K.; Baumgart, T.; Offenhäusser, A.; Knoll, W. Angew. Chem., Int. Ed. 2001, 40, (28) Morigaki, K.; Baumgart, T.; Jonas, U.; Offenhäusser, A.; Knoll, W. Langmuir 2002, 18, (29) Morigaki, K.; Schönherr, H.; Frank, C. W.; Knoll, W. Langmuir 2003, 19, (30) Spinke, J.; Yang, J.; Wolf, H.; Liley, M.; Ringsdorf, H.; Knoll, W. Biophys. J. 1992, 63, (31) Wagner, M. L.; Tamm, L. K. Biophys. J. 2000, 79, Materials and Methods 2.1. Materials. Diacetylene phospholipid (1,2-bis(10,12-tricosadiynoyl)-sn-glycero-3-phosphocholine (DiynePC)) and phosphatidylcholine from egg yolk (egg-pc) were purchased from Avanti Polar Lipids (Alabaster, AL). Texas Red 1,2-dihexadecanoyl-sn-glycero-phosphoethanolamine (TR-PE) and N-(7-ni- trobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3- phosphoethanolamine (NBD-PE) were purchased from Molecular Probes (Eugene, OR). Sodium dodecyl sulfate (SDS) was purchased from Nacalai Tesque (Kyoto, Japan). All chemicals were purchased as the reagent grade and used without further purification Substrate Cleaning. For the substrate of SPBs, microscopy glass slides were used (Matsunami, Osaka, Japan). The substrates were cleaned first with a commercial detergent solution (0.5% Hellmanex/water, Hellma, Mühlheim, Germany), rinsed with deionized water, treated in a warm solution of 28% NH 4OH/30% H 2O 2/H 2O (1:1:5) (65 C for 15 min), rinsed again with deionized water extensively, and then dried in a vacuum oven at 110 C. This protocol resulted in clean and hydrophilic surfaces for the adsorption of lipid bilayer membranes Preparation of Monomeric DiynePC SPBs. Bilayers of monomeric DiynePC were deposited onto solid substrates from the air/water interface by the Langmuir-Blodgett (LB) and subsequent Langmuir-Schaefer (LS) methods using a Langmuir trough (HBM-AP, Kyowa Interface Science, Asaka, Japan). Diacetylene lipid DiynePC was spread from chloroform solution. The lipids formed stable monolayers at the air/water interface up to a surface pressure of 40 mn/m. The monolayers were transferred onto solid substrates at 35 mn/m (fully condensed state). The first monolayer was deposited by dipping and withdrawing the substrate vertically (LB method). The second leaflet was deposited onto the hydrophobic surface of the first monolayer by pressing the substrate horizontally through the monolayer at the air/water interface and dropping it into the subphase (LS method). After the deposition of the second monolayer, the samples were collected from the trough and stored in deionized water (in the dark) for the photopolymerization Photopolymerization of DiynePC Bilayers. As the light source, we used a mercury lamp (SP-V Deep UV lamp, Ushio, Tokyo, Japan) that has strong emission in the deep UV region (<300 nm). UV light around 250 nm was selectively irradiated by using a Kr*F laser mirror (TFM-30C05-248, Sigma (32) Bayerl, T. M.; Bloom, M. Biophys. J. 1990, 58, (33) Koenig, B. W.; Krueger, S.; Orts, W. J.; Majkrzak, C. F.; Berk, N. F.; Silverton, J. V.; Gawrisch, K. Langmuir 1996, 12, (34) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, (35) Reviakine, I.; Brisson, A. Langmuir 2000, 16,

3 Micropatterned Composite Membranes of Bilayers Langmuir, Vol. 20, No. 18, Koki, Saitama, Japan). The applied intensity was typically 25 mw/cm 2 (254 nm). The polymerization was conducted at room temperature in a closed system that was comprised of a water reservoir, a pump, and a cell ( 4 ml volume). The water reservoir was depleted of oxygen by purging it with argon. 36 Oxygen-free water was circulated continuously by the pump through the cell where polymerization of the bilayers was conducted. The cell had two walls on opposite sides, one being the sample (the SPB was inside the cell) and the other being a quartz window through which UV light was illuminated. Desired patterns were transferred to the SPB in the photopolymerization process by illuminating the sample through a mask (a quartz slide with a patterned gold coating) which was placed directly on the SPB. After sufficient circulation of deaerated water (typically for 15 min), the pump was stopped and the photopolymerization was started. The applied UV dose was adjusted by changing the irradiation time between 10 s and 5 min. After the UV irradiation, nonpolymerized DiynePC molecules were removed from the substrate surface with 0.1 M SDS solution (at 25 C for 30 min) Incorporation of Fluid Bilayers into the Wells Surrounded by Polymerized Bilayers. To incorporate new lipid bilayers into the lipid-free wells, the vesicle fusion method was applied. 5,32-35 Vesicle suspensions of 1 mm egg-pc containing either 1 mol % TR-PE or 2 mol % NBD-PE fluorescence marker (0.01 M phosphate buffer with 0.15 M NaCl (ph 6.8)) were prepared by the following protocol. Lipids were mixed in a chloroform solution and the solvent was evaporated by a stream of nitrogen and vacuum, and the lipids were then allowed to hydrate in buffer. The resulting multilamellar vesicles were put through five freeze/thaw cycles and then extruded through a polycarbonate filter with pores of a diameter of 50 nm (LiposoFast, Avestin, Ottawa, Canada). The vesicle suspension was brought into contact with the patterned DiynePC bilayer and incubated typically for 30 min. The samples were then rinsed extensively with deionized water Microscopy Observation. For fluorescence microscopy, an upright microscope (AX70, Olympus, Tokyo, Japan) equipped with a mercury lamp (HBO-100, Olympus) and a 60 waterimmersion objective (NA 0.90) was used. Polymeric DiynePC bilayers and NBD-PE were observed by the Olympus U-MNIBA2 filter set (excitation wavelength, nm; emission wavelength, nm; abbreviated as B/G filter set), and TR-PE was observed by the Olympus U-MWIY2 filter set (excitation wavelength, nm; emission wavelength, >610 nm; abbreviated as Y/R filter set). Fluorescence microscopy images were obtained with a CCD camera (DP-50, Olympus) and processed by Adobe Photoshop (Adobe, San Jose, CA) Fluorescence Recovery after Photobleaching (FRAP) Experiments. For the fluorescence recovery after photobleaching (FRAP) experiments, a commercial confocal scanning laser microscope (FV-300, Olympus) equipped with an argon laser (488 nm, 10 mw) and a 60 water-immersion objective (NA 0.90) was used. The lateral diffusion coefficient of NBD-PE in fluid bilayer samples (2 mol %) was measured by a programmed sequence of scans. After a rectangular area (32 23 µm 2 ) was scanned first at the full power of the laser, the fluorescence recovery of the bleached spot was regularly imaged with an attenuated laser beam (1%). To avoid additional photobleaching, only a limited number of images were acquired. The fluorescence intensity was integrated for the whole bleached area and plotted versus the elapsed time. The diffusion coefficient of the fluorescent probe was obtained by using the FRAPFIT program (Jürgen Worm, Max-Planck-Institute for Polymer Research, Mainz, Germany) that applied the theoretical framework originally proposed by Axelrod et al. 37 and modified by Soumpasis. 38 (36) Oxygen had to be removed from the aqueous solution prior to photopolymerization, since the presence of oxygen inhibited the polymerization, presumably by quenching diacetylene radicals. Day, D.; Ringsdorf, H. J. Polym. Sci., Polym. Lett. Ed. 1978, 16, (37) Axelrod, D.; Koppel, D. E.; Schlessinger, J.; Elson, E.; Webb, W. W. Biophys. J. 1976, 16, (38) Soumpasis, D. M. Biophys. J. 1983, 41, Figure 2. Fluorescence micrographs of a patterned bilayer on a glass substrate. (A) Green fluorescence from the bilayer of polymerized DiynePC (UV irradiation dose, 4.4 J/cm 2 ). (B) Red fluorescence from TR-PE doped egg-pc bilayers incorporated in the wells between polymerized bilayers (corrals). (C) Local photobleaching of TR-PE fluorescence in the lower part of the central corral. (D) 4 min after the fluorescence micrograph was obtained for part C. The partially photobleached corral became homogeneously dark due to lateral diffusion of lipid molecules, whereas the neighbor corrals retained the original fluorescence intensity. The scale bar corresponds to 50 µm. 3. Results 3.1. Fluorescence Microscopy Observation of the Patterned Bilayers. Micropatterning of the bilayers was achieved by placing a contact lithography mask on the bilayer surface during the photopolymerization. After the UV irradiation, nonreacted monomers were removed selectively by immersing the sample in a 0.1 M SDS solution for 30 min (at 25 C) and new lipid bilayers (egg- PC containing 1 mol % TR-PE) were introduced into the wells surrounded by polymeric bilayers by the vesicle fusion technique. Figure 2 shows fluorescence micrographs of a patterned bilayer in which a large UV irradiation dose (4.4 J/cm 2 ) was applied for the polymerization. In Figure 2A, the polymerized bilayer is observed to be green due to the fluorescence from the conjugated backbone. Lipid bilayers of egg-pc containing 1 mol % TR-PE were incorporated selectively into the square-shaped areas where monomers had been protected with the mask during the lithographic UV exposure and selectively removed (Figure 2B). The bilayers are continuous and fluid within the areas surrounded by the polymeric bilayers (corrals), as has been clearly demonstrated by the local photobleaching of TR-PE with a focused intense illumination. The partial photobleaching initially created a concentration gradient of intact TR-PE molecules (Figure 2C), but the lateral diffusion of lipid molecules made the fluorescence intensity in the whole corral homogeneous (Figure 2D). The other corrals retained the initial fluorescence intensity, indicating that the polymerized bilayer of DiynePC acted as an effective barrier for the lateral diffusion of lipid molecules. Reduction of the UV irradiation dose in the photopolymerization had a significant influence on the distribution of newly incorporated lipid bilayers in the patterned polymeric bilayers. Figure 3 shows the case where the poymerization of DiynePC was made with a UV irradiation dose of 0.9 J/cm 2. The polymerized bilayer of DiynePC showed weaker fluorescence (Figure 3A). On the other hand, fluorescence of TR-PE was observed not only in the corrals but also within the polymerized DiynePC bilayer domain (Figure 3B), suggesting that fluid lipid bilayers

4 7732 Langmuir, Vol. 20, No. 18, 2004 Morigaki et al. Figure 3. Fluorescence micrographs of a patterned lipid bilayer on a glass substrate. A smaller UV dose (0.9 J/cm 2 ) was applied for the polymerization of DiynePC compared with the sample in Figure 2. (A) The polymerized bilayer of DiynePC could be recognized with green fluorescence. (B) Fluorescence of egg- PC/TR-PE bilayers was observed both in the corrals and within the polymeric bilayer domains. (C and D) The lateral mobility of egg-pc and TR-PE molecules within the polymeric bilayer domain was demonstrated by locally photobleaching TR-PE and observing the dissolution of the bleach spot. The image in part D was obtained 6 min after the image was obtained for part C. The scale bar corresponds to 50 µm. (39) Vesicles that were adsorbed on the polymeric bilayer (bright dots in Figure 4B-D) also retained their fluorescence. The energy transfer from poly-diynepc to TR-PE was apparently restricted within the adsorbed planar bilayer. Figure 4. Fluorescence micrographs of a patterned lipid bilayer (the same sample as that in Figure 3). (A) Fluorescence observation of poly-diynepc using the B/G filter set (B) resulted in photobleaching of TR-PE embedded within the polymeric bilayer domain. TR-PE fluorescence was photobleached only in the polymeric bilayer region. TR-PE fluorescence in the central corral remained intact, indicating that photobleaching is the result of energy transfer from excited poly-diynepc to embedded TR-PE molecules. Fluorescence of the bleached area recovered by the diffusion of intact TR-PE molecules from the central corral and the surrounding area. (C) 4 min after the fluorescence micrograph was obtained for part B. (D) 54 min after the fluorescence micrograph was obtained for part B. The scale bar corresponds to 50 µm. of egg-pc/tr-pe were penetrating into the region of polymerized DiynePC. Lipid molecules were mobile within the polymeric bilayers, as shown by local photobleaching of the fluorophore (Figure 3C and D). The edge of the bleached spot became unclear with time, and the fluorescence recovered due to the influx of intact TR-PE molecules. The penetration of fluid lipid bilayers into the polymeric domains was indicated by another set of experiments where TR-PE molecules were effectively photobleached by energy transfer from the excited poly- DiynePC bilayer (Figure 4). A patterned bilayer was locally illuminated through an aperture at a wavelength of 488 nm (Figure 4A). Subsequent fluorescence microscopy observation using the Y/R filter set revealed that TR-PE molecules within the polymeric bilayer domain were photobleached by the previous illumination (Figure 4B). Since polymeric DiynePC was selectively excited at this wavelength, the photobleaching of TR-PE must have occurred through the mechanism of energy transfer from poly-diynepc. This conclusion was corroborated by the fact that the fluorescence of TR-PE remained intact within the central corral where no poly-diynepc was present (Figure 4B). 39 As we kept the sample in the dark for a prolonged period, TR-PE molecules diffused laterally from the nonbleached areas (both the corral and surrounding bilayer domains) and the bleached area recovered fluorescence with time (Figure 4C and D) Fluorescence Recovery after Photobleaching (FRAP). The fluidity of lipid bilayers embedded within the polymeric bilayers has been studied more quantitatively by estimating the lateral diffusion coefficients of the lipids. For this purpose, the fluorescence recovery after photobleaching (FRAP) technique was applied using a confocal scanning laser microscope (CSLM). A small rectangular area (32 23 µm 2 ) was scanned with an intense laser beam to photobleach incorporated fluorescence dye molecules, and subsequent recovery of fluorescence was monitored with an attenuated laser beam. For this investigation, egg-pc bilayers containing 2 mol % NBD-PE were used. The average diffusion coefficient was found to be 1.54 ( 0.28 µm 2 /s for samples without polymeric bilayers (homogeneous bilayers of egg-pc/nbd-pe). The presence of polymeric bilayers retarded the lateral diffusion of NBD-PE. Figure 5 compares the fluorescence recovery profiles in the patterned bilayer samples, where DiynePC bilayers were polymerized with various UV irradiation doses. The samples had the same pattern as those in Figures 2 and 3 (corrals of fluid lipid bilayers were surrounded by grid-shaped polymeric bilayers), and the fluorescence recovery profiles were compared in the corrals (homogeneous bilayers of egg-pc/nbd-pe) and in the polymeric bilayer domains where fluid bilayers penetrated. In the polymerized bilayer that was produced with a small UV dose, the photobleached spot recovered its fluorescence rather rapidly, almost as fast as it did within the corral (Figure 5A). The fluorescence recovery in the polymer became progressively retarded, if a larger UV dose was applied for the photopolymerization (recovery in the polymer became slower compared with that in the corral, Figure 5B), and there was a threshold UV irradiation dose above which practically no fluorescence recovery was observed (Figure 5C). The photobleached spots in the corrals recovered fluorescence rapidly for all samples. The final fluorescence intensity after recovery was lower than the initial level because of the limited amount of fluorophore molecules within the same corral (Figure 5B and C). In the case of the smallest UV dose (Figure 5A), gradual fluorescence recovery continued, presumably due to the supply of new fluorophore molecules that diffused through the surrounding polymeric bilayers. The diffusion coefficients within the polymer region were estimated by fitting the recovery curves to the model originally proposed by Axelrod et al. 37 and modified by Soumpasis 38 and were

5 Micropatterned Composite Membranes of Bilayers Langmuir, Vol. 20, No. 18, Figure 5. Fluorescence recovery curves of egg-pc/nbd-pe bilayers incorporated in the micropatterned bilayers. Polymeric DiynePC bilayers were created with various UV irradiation doses. Photobleaching was made either at the center of the corrals (fluid bilayers of egg-pc/nbd-pe, closed symbols) or at the center of the grid-shaped polymeric bilayers with embedded fluid bilayers (open symbols). The vertical axis is the raw fluorescence intensity (in relative units) integrated for the whole bleached area of the CSLM (32 23 µm 2 ). The experiments consisted of initial scanning with an attenuated laser beam, photobleaching (full power laser beam, no signal acquisition), and successive observation of the fluorescence recovery. All the experimental steps are shown. Figure 6. Relative diffusion coefficients of egg-pc/nbd-pe incorporated within the polymerized bilayers of DiynePC as a function of the UV irradiation dose applied for the photopolymerization. The diffusion coefficients were normalized to the value obtained in an SPB of pure egg-pc/nbd-pe. plotted as a function of the UV dose (Figure 6). 40 The diffusion coefficients decreased with the increase of the UV dose and reached zero at a UV dose of 1.3 J/cm Discussion (40) The currently applied geometrical configuration is not consistent with the assumption made in the theoretical frameworks of FRAP analyses, where an infinitely large reservoir of intact fluorophore is assumed for the estimation of the lateral diffusion coefficients. Nevertheless, we have employed this configuration because of the experimental advantages that we can directly compare diffusion coefficients in the fluid membranes and in the polymeric bilayers. Consequently, the obtained values are relative rather than absolute. This, however, does not affect the main conclusions that we drew from these measurements. From the above-described experimental results, we infer that the current micropatterning approach can generate composite membranes of polymerized and fluid lipid bilayers on the substrate with defined geometries. The UV dose applied for the photopolymerization of DiynePC strongly affected the distribution of incorporated fluid bilayers. For a large UV dose (typically over 1.5 J/cm 2 in the current system), the polymerized bilayer was impermeable to the newly incorporated fluid bilayers and confined them in the region where monomeric DiynePC was protected during the photopolymerization and removed subsequently by the SDS treatment (Figure 2). On the other hand, in the case where the UV dose for the photopolymerization was smaller (typically J/cm 2 ), we observed that lipids in the fluid bilayers were laterally penetrating into the bilayer region of polymerized DiynePC (Figure 3). Polymeric bilayers of DiynePC and fluid bilayers of egg-pc were forming closely integrated membranes. Local photobleaching of embedded fluorophores (TR-PE or NBE-PE) and observation of the fluorescence recovery have shown that lipid molecules are laterally mobile within these regions. Furthermore, the efficient energy transfer from polymerized DiynePC to TR-PE observed in Figure 4 suggested strongly that the polymerized bilayer of DiynePC and the fluid bilayer containing TR-PE were spatially integrated with molecular proximity. In the previous atomic force microscopy (AFM) study, we showed that polymerized bilayers of DiynePC formed solubilization-resistant domains, even if the applied UV dose was small. 29 These domains remained on the substrate as dispersed islands after the removal of monomers with SDS solution. The thickness corresponded to a bilayer ( 5.0 nm), indicating that the basic bilayer structure was preserved. As the UV irradiation dose was increased, these islands grew in size and number and a larger area became covered by the polymerized bilayers. From this observation, it is plausible that the newly incorporated bilayers of egg-pc are filling the void between these islandlike domains of polymeric DiynePC bilayer. The current fluorescence microscopy observations also support that polymerized DiynePC bilayers were not removed completely from the substrate by the SDS treatment, even if the UV irradiation dose was small. However, separate experiments (UV-visible absorption measurements, see the Supporting Information) have indicated that considerable conformational changes of DiynePC molecules within bilayers may be taking place during the SDS treatment. More detailed molecular structures of the polymeric bilayers have to be clarified in further investigations. 41 (41) There are some previous AFM studies of polymerized SPBs which reported rather heterogeneous morphologies for poly(diacetylene) lipid bilayers. ((a) Ross, E. E.; Rozanski, L. J.; Spratt, T.; Liu, S. C.; O Brien, D. F.; Saavedra, S. S. Langmuir 2003, 19, (b) Morigaki, K.; Schönherr, H.; Frank, C. W.; Knoll, W. Langmuir 2003, 19, ) However, it is becoming clear that the obtained bilayer structures depend very much on the fabrication procedures and conditions.

6 7734 Langmuir, Vol. 20, No. 18, 2004 Morigaki et al. The diffusion coefficient of lipid molecules embedded in the polymeric DiynePC bilayer was suppressed progressively with the increased UV dose applied for the photopolymerization (Figure 6). The increased coverage of the substrate surface with the polymeric bilayer domains, as shown by the AFM measurements, has most likely acted as an effective obstacle for the long distance diffusion of lipid molecules. There was a critical UV irradiation dose where the lateral diffusion of lipid molecules was completely hindered by the polymer domains. This obstructed diffusion can be interpreted by using percolation theory. At low obstacle concentrations, the obstacles form islands in the connected conducting phase. As the concentration of the obstacles is increased, they connect to each other and separate the conducting phase. Eventually, the conducting phase forms isolated pools in a continuous phase of the obstacle. The point at which the conduction is abolished is called the percolation threshold. Percolation analyses have been applied extensively to the retarded lateral diffusion of membrane associated lipids and proteins in cellular membranes The observed decrease of the lateral mobility in Figure 6 fits in well with the expected percolation behavior, although we are currently not able to apply the percolation analysis in terms of the obstacle area fraction because we do not have quantitative data on the surface area coverage of polymerized and fluid bilayers. Furthermore, our previous AFM study indicated that the shapes of polymerized DiynePC bilayers are highly branched and extended. 29 Extended obstacles have been shown to be more effective than compact obstacles in hindering diffusion. 43 Therefore, the shape of polymerized domains should also play a critical role in the restriction of lateral diffusion. The possibility to integrate polymerized and fluid lipid bilayers as a composite membrane and regulate lateral diffusion of membrane associated molecules by polymeric bilayers has further implications for the construction of micropatterned lipid bilayers. We previously proposed the construction of microarrays of fluid lipid bilayers (model cellular membranes) separated from each other by lithographically polymerized bilayers. The current results suggest that one could modulate the lateral mobility of membrane associated molecules by purposefully designing the geometry and degree of polymerization. In this way, one should be able to construct arrays of lipid bilayers that are not completely isolated but partially connected. A simplified conceptual illustration is given in Figure 7. Figure 7A schematically depicts the procedure for generating a patterned bilayer film with spatially defined polymerization profiles. Either by using a contact mask with varied transmission profiles or by conducting successive exposure with multiple protection patterns, one can modulate the UV dose and therefore the progress of UV polymerization. After removing monomers (and oligomers) by a detergent solution, one obtains three distinctive regions, that is, (1) a lipid-free area, (2) an area partially covered by the polymer, and (3) an area completely covered by the polymer. Lipid-free areas are subsequently refilled with fluid lipid bilayers. The areas partially covered by the polymer also incorporate fluid bilayer with partial lateral mobility of membrane associated molecules. Figure 7B is an example of the configuration where patches of fluid lipid bilayers are connected by a membrane channel. The monomeric DiynePC bilayer in the channel was irradiated with 40% UV light compared with the surrounding area upon polymerization so that Figure 7. The proposed strategy for controlling lateral diffusion of membrane associated molecules by using lithographically defined polymerization profiles. (A) Schematic outline of the procedure for generating a patterned bilayer film with spatially defined polymerization profiles. The UV dose of lithographic exposure is controlled either by using a mask with varied transmission profiles or by successive exposure with multiple protection patterns. (B) A membrane channel of permeable membrane domain consisting of polymerized DiynePC and fluid bilayers connects two fluid lipid bilayer membranes: a schematic illustration (left) and an experimental demonstration (right). The channel connecting two square areas has been created by polymerizing DiynePC with 40% UV light dose compared with the surrounding area. The fluorescence images of polymerized DiynePC (green) and TR-PE (red) were superimposed. The scale bar corresponds to 50 µm. the formed polymerized bilayers allow penetration of fluid lipid bilayers. As a result of this controlled UV photopolymerization, the channel connects two fluid membrane domains with a reduced lateral mobility of molecules. It is currently not clear whether there is a certain molecular selectivity in the lateral diffusion within the membrane channel, but it should be basically possible to impose molecular size dependent obstruction by optimizing the shape and area fraction of polymeric bilayer domains. 45 Micropatterning of SPBs using lithographically polymerized lipid bilayers is unique compared with other approaches in that polymeric and fluid bilayers are integrated as a composite membrane. As demonstrated in this report, this feature enables geometrically controlled obstruction for the lateral diffusion of membrane associated molecules by locally varying the degree of photopolymerization. Such obstructed lateral diffusion is commonly observed in cellular membranes, and in fact, nonrandom distribution of membrane components and confinement of them in microdomains are regarded as a prerequisite for numerous vital functions such as signal (42) Saxton, M. J. Biophys. J. 1982, 39, (43) Saxton, M. J. Curr. Top. Membr. 1999, 48, (44) Ratto, T. V.; Longo, M. L. Biophys. J. 2002, 83, (45) Saxton, M. J. Biophys. J. 1993, 64,

7 Micropatterned Composite Membranes of Bilayers Langmuir, Vol. 20, No. 18, transduction and trafficking. 46,47 Therefore, micropatterned model membranes with modulated lateral mobility should be able to reproduce more precisely the situations in cellular membranes and provide new tools to study the cellular functions in an artificially controlled platform. Acknowledgment. This work has been supported by the Mitsubishi Chemical Corporation Fund and the Promotion Budget for Science and Technology (AIST Upbringing of Talent in Nanobiotechnology Course) from the Ministry of Education, Science, Culture and Sports (MEXT). We thank Dr. Jürgen Worm (Max-Planck- Institute for Polymer Research) and Dr. Tobias Baumgart (46) Simons, K.; Ikonen, E. Nature 1997, 387, (47) Kusumi, A.; Sako, Y. Curr. Opin. Cell Biol. 1996, 8, (Cornell University) for their assistance in the FRAP data analysis. Dr. Junji Nishii and Dr. Kenji Kintaka (Photonics Research Institute, AIST) are gratefully acknowledged for their support in fabricating the photolithography masks, and Dr. Shunsuke Yuba, Dr. Keiko Tawa (Special Division for Human Life Technology, AIST), and Dr. Kenji Kamada (Photonics Research Institute, AIST), for allowing us to use the upright light microscope, the UV lamp, and the UV-visible spectrophotometer, respectively. Supporting Information Available: Information on the UV-visible absorption measurements and a figure showing UV-visible absorption spectra of polymerized DiynePC. This material is available free of charge via the Internet at LA049340E