Integrated model membrane for biophysical studies and biomedical applications

Patterned SLBs were generated on a microscopy coverslip by photolithographic polymerization of 1,2-bis(10,12-tricosadiynoyl)-sn-glycero-3-phos-phocholine (DiynePC) (Fig. 2). ⁷⁾ Microscopy coverslips were cleaned with 0.1 M sodium dodecyl sulfate (SDS) for 20 min under sonication, rinsed with deionized water, treated in a solution of NH₄OH (28%)/H₂O₂ (30%)/H₂O (0.05:1:5) for 10 min at 65 °C, rinsed extensively with deionized water, and then dried under nitrogen blow. The substrates were further cleaned by the UV/ozone treatment for 20 min (PL16−110, Sen Lights Corporation, Toyonaka, Japan). Bilayers of monomeric DiynePC were deposited onto substrates by the spontaneous transformation of lipid vesicles into SLB (vesicle fusion). The substrate was cooled on ice to induce rapid formation of crystalline domains. ⁸⁾ Polymerization of DiynePC bilayers was conducted by UV irradiation using a Mercury lamp (SP9, Ushio, Tokyo, Japan) as the light source. The applied UV intensity was typically 170 mW cm⁻² at 254 nm, and the UV dose was modulated with the illumination time. After UV irradiation, nonpolymerized DiynePC molecules were removed from the substrate surface by immersing in 0.1 M SDS at 30 °C and rinsing extensively with deionized water. Natural lipid bilayers were incorporated into the patterned framework of polymeric bilayer by vesicle fusion or spontaneous spreading of phospholipids from lipid aggregates (self-spreading). ^9,10)

Zoom In Zoom Out Reset image size

Membrane proteins were reconstituted into the patterned membrane either from purified/solubilized micelles or crude native membranes. The photoreceptor rhodopsin (Rh) was purified from the rod outer segments (ROS) of bullfrogs (Rana catesbeiana), and was reconstituted into a pre-formed SLB by the rapid dilution of solubilized Rh molecules. ¹¹⁾ In a different set of experiment, native membranes such as thylakoids from spinach chloroplast were reconstituted into the preformed polymeric bilayer framework by applying them together with phospholipid vesicles. ¹²⁾ The phospholipid vesicles acted as a catalyst that induced spontaneous reorganization of thylakoid membranes into a planar SLB surrounded by polymeric bilayers.

To realize biomimetic 3D architecture based on a patterned membrane, the surface of polymeric bilayer was functionalized using a polymerizable phospholipid bilayers having a reactive ethanolamine head group 1,2-bis(10,12-tricosadiynoyl)-sn-glycero-3-phosphoethanolamine (DiynePE) (Fig. 2, Fig. 3). Although DiynePE does not form a stable bilayer structure because the head group is significantly smaller than the hydrophobic tails, ¹³⁾ stable bilayers could be formed by mixing DiynePE with a polymerizable phospholipid having a choline head group (DiynePC). Bilayers composed of DiynePC and DiynePE were deposited onto 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). The LB/LS process was employed, since formation of SLBs by vesicle fusion is unfavorable for DiynePC/DiynePE due to the intrinsic curvature of PE-containing membrane. ¹⁴⁾ Lithographic polymerization of DiynePC/DiynePE bilayers was conducted by UV irradiation and subsequent SDS treatment as described above. Natural lipid bilayers were incorporated into the patterned framework of polymeric bilayer by vesicle fusion.

Zoom In Zoom Out Reset image size

The surface of DiynePC/DiynePE bilayer was functionalized with biotin using a conjugate of N-Hydroxysuccinimide (NHS), polyethylene glycol (PEG), and biotin (NHS-PEG₄-biotin). A nanometer-sized gap-junction was generated by attaching a patterned bilayer with a polydimethylsiloxane (PDMS) slab using an adhesion layer having a defined thickness (Fig. 4). As the adhesion material, we used polymer-coated silica nanoparticles (SiNPs). Polymer brush consisting of 2-methacryloyloxyethyl phosphorylcholine (MPC) and 2-aminoethyl methacrylate (AEMA) (poly(MPC-co-AEMA) was formed on SiNPs by the surface-initiated atom transfer radical polymerization (ATRP). ¹⁵⁾ To use the polymer-coated SiNPs as the adhesion material, AEMA in the polymer brush was biotinylated. Patterned polymeric bilayer, SiNPs, and PDMS surfaces were bound with the biotin-streptavidin linkage.

Zoom In Zoom Out Reset image size

Lithographic polymerization of DiynePC bilayer and subsequent incorporation of natural lipid bilayers result in a patterned hybrid membrane (Fig. 2). The polymeric bilayer acts as a framework for defining the area of natural lipid bilayers, whereas the natural lipid bilayers retain the physicochemical properties of biological membrane such as lateral mobility of membrane-bound molecules (fluidity). Natural lipid bilayers were incorporated into the patterned polymeric bilayer scaffold by two spontaneous processes, i.e. vesicle fusion and self-spreading (Fig. 5). ^9,10) In vesicle fusion, lipid vesicles adsorb onto a hydrophilic surface, rupture, and lipid patches fuse with each other. Currently, SLBs are most commonly generated by vesicle fusion owing to its technical simplicity. ^16,17) On the other hand, self-spreading is a spontaneous wetting phenomenon of a hydrophilic surface by a lipid bilayer that extends from a reservoir of liquid crystalline lipid bilayers. ^18–20) This technique is potentially useful for the formation of supported membranes that are more resistant towards defects formation, since lipid molecules are continually supplied from the lipid reservoir. Heightened lipid packing was observed in self-spread bilayers compared with those formed by vesicle fusion [Fig. 5(c)]. These features give self-spreading important advantages for preparing patterned model membranes for reconstituting membrane proteins, where detergent solutions are applied. Importantly, the scaffold of polymeric bilayer can stabilize incorporated fluid lipid bilayers. Formation of SLBs was accelerated in the presence of polymeric bilayers both for vesicle fusion and self-spreading, indicated that the hydrophobic edges of polymeric bilayer reduced the energetic barriers of SLB formation. ^9,10)

Zoom In Zoom Out Reset image size

In the patterned hybrid membrane, polymeric and fluid lipid bilayers form a continuous membrane. A partially polymerized bilayer, in which nanoscopic polymeric bilayer domains only partially cover the substrate surface, can be generated by modulating the UV dose for polymerization [Figs. 6(a), 6(b)]. Fluid lipid bilayers can penetrate into the partially polymerized region, forming a nanometer-sized composite membrane of polymeric and fluid bilayers [Figs. 6(c), (d)]. ^21,22) Lateral diffusion of membrane-bound molecules is obstructed by polymeric bilayer domains [Fig. 6(e)]. The degree of obstruction increases linearly with the polymeric bilayer coverage, and leads to percolation. An important manifestation of the partially polymerized bilayer region is spontaneous phase separation of lipid domains. As we introduced a lipid bilayer that can separate into liquid-ordered (L_o) and liquid-disordered (L_d) domains, we observed that L_o and L_d domains preferentially accumulated in the polymer-free and partially polymeric regions, respectively [Figs. 6(f)–6(h)]. ^22,23) Two possible mechanisms are involved in the patterned phase separation. First, L_d domains preferentially accumulate at the boundary between polymeric and fluid bilayers due to the lower bending elasticity of L_d domains, which results in a reduced energetic cost of bilayer deformation. ²⁴⁾ The second factor is the Ostwald ripening of domains. L_o domains in the matrix of a continuous L_d phase tend to grow in size in order to minimize the line tension. ²⁵⁾ However, the domain growth is hindered in the partially polymerized regions by polymeric bilayer domains. Patterned L_o/L_d domains have cholesterol-enriched and cholesterol-depleted domains in a predefined geometry. They can be used as a model for lipid micro-/nano-domains in the biological membrane (e.g. lipid raft) for studying their physicochemical and functional roles.

Zoom In Zoom Out Reset image size

Membrane proteins play essential roles in the cell, including signal transduction and energy conversion, and they represent important targets of drugs. ²⁶⁾ We reconstituted a G protein coupled receptor (GPCR) of vertebrate retina (Rh) and its cognate G protein transducin (G_t) into a patterned membrane composed of L_o/L_d domains and observed their lateral distribution (Fig. 7). ¹¹⁾ We determined the partition and diffusion coefficients of Rh and G_t in the L_o-rich and L_d-rich regions. Both Rh and G_t were predominantly localized in the L_d phase, suggesting their low affinity to lipid rafts (raftophilicity). ¹¹⁾ The raftophilicity was presumably influenced by the lipid chains attached to proteins by the post-translational modifications. Measuring the raftophilicity for a variety of molecules in the signal transduction should reveal the functional significance of protein localization in the photosignal transduction.

Zoom In Zoom Out Reset image size

In a different approach, we incorporated native biological membranes into the framework of polymeric bilayer. Thylakoid membranes purified from spinach leaves could be introduced into the corrals surrounded with polymeric bilayer together with synthetic phospholipid (DOPC) vesicles, spontaneously forming a laterally continuous and fluid 2D membrane. ^12,27) Chlorophyll fluorescence arising from photosystem II (PSII) recovered after photobleaching, suggesting that the membrane components are laterally mobile. We confirmed the electron transfer activity of photosystem I (PSI) by observing the generation of NADPH in the presence of water-soluble ferredoxin and ferredoxin-NADP⁺ reductase. The lateral mobility of membrane-bound molecules and the functional reconstitution of major photosystems provide evidence that the hybrid thylakoid membranes could be an excellent experimental platform to study the 2D molecular organization and machinery of photosynthesis. In a different line of study, we directly transferred a mammalian GPCR, dopamine D2 receptor (D2R), from Chinese hamster ovary (CHO) cells into the patterned membrane using vesicles that bud off from the cell membrane (blebs). ²⁸⁾ D2R molecules were effectively reconstituted in a nanometer-sized thin cleft formed between a patterned SLB and PDMS. The density of D2R in the model membrane was high enough for the transient dimerization similar to the live plasma membrane. Although the mechanism of this enhanced reconstitution in the PDMS cleft is not fully understood, this finding points to a new possibility to use the PDMS cleft as a platform for reconstituting and studying membrane proteins under the quasi-physiological conditions.

Although SLBs have a 2D membrane structure, the use of 3D micro-/nano-fluidics significantly enhances the utility of the model membrane. Therefore, it is important to establish a means to couple SLBs with 3D structures. We developed a methodology to chemically functionalize the surface of polymeric bilayer for attaching it with molecules, cells, and micro-fabricated structures (Fig. 3). ⁷⁾ Polymeric bilayer with a reactive moiety was generated using a mixed bilayer of DiynePC and DiynePE. The amine group in DiynePE allowed to conjugate functional moieties (e.g. biotin) onto the membrane surface. Bonding PDMS and a polymeric bilayer with a molecular glue allows us to combine PDMS microstructures and substrate supported biomimetic membranes. We confined a membrane bound enzyme (cytochrome P450) with its water-soluble substrate in micro-compartments formed between PDMS and micro-patterned model membrane, and evaluated its functions. ⁷⁾

We combined the 2D patterned membrane and a nanometric detection volume by forming a very thin aqueous space between the fluid bilayer and a PDMS sheet (nanogap-junction) using an adhesion material with a pre-defined thickness (e.g. silica nanoparticles coated with hydrophilic polymer). The fabrication process of a nanogap-junction is outlined in Fig. 4. The adhesion material was attached onto the surface of the polymeric bilayer by the specific biotin-streptavidin linkage. The polymeric bilayer and the adhesion material were functionalized with biotin moiety. Then, streptavidin having four binding sites to biotin molecules was used to link them. In the same way, PDMS having the biotin-modified surface was attached onto the surface of the adhesion layer. The thickness of the nanogap-junction was determined by the adhesion layer, and we could generate a gap thickness of less than 100 nm. ²⁹⁾ The combination of a nanometric gap structure and a fluid lipid bilayer enabled to selectively transport and detect target molecules that could bind to the membrane surface by specific molecular recognition, whereas the nonspecific penetration of coexisting molecules was suppressed [Fig. 8(a)]. In Fig. 8(b), the model target molecule (cholera toxin subunit B: CTB) could bind to a fluid bilayer containing its ligand (G_M1) and penetrate into the nanogap-junction by lateral diffusion (green fluorescence). In contrast, the model coexisting molecule (bovine serum albumin: BSA) could not bind to the membrane surface, and mostly remained in the solution in the microchannel. The selective transport of CTB into the nanogap-junction contributed to enhance the signal-background (S/B) ratio (i.e. the ratio of fluorescence intensities from specificcally bound CTB and the non-specific background noise measured in the polymeric bilayer regions). The use of silica nanoparticles as the adhesion layer enabled to control the thickness of the nanogap-junction accurately due to the uniform sizes of silica nanoparticles. ³⁰⁾ By using silica nanoparticles having a smaller size, we could generate a thinner nanogap-junction, which realized a higher S/B ratio [Fig. 8(c)]. In the micro-channel, the background fluorescence of BSA was very intense, and the fluorescence from membrane-bound CTB in the fluid bilayer was hardly observable. The signal from CTB was more clearly visible in the nanogap-junction due to the suppressed background fluorescence, because BSA molecules were mostly excluded. The background fluorescence intensity was more effectively suppressed for 30 nm nanoparticles compared with 100 nm nanoparticles, because a thinner gap is more effective in eliminating BSA, thus enhancing the S/B ratio even more drastically [Fig. 8(c)]. Furthermore, silica nanoparticles are mechanically robust, and the nanogap-junctions formed had a long life time (>week). The stability allows one to store a pre-assembled biochip having a nanogap-junction, thus eliminating the necessity for elaborate onsite assembly process. Compared to other techniques used for ultrasensitive detection, the nanogap-junction is unique in the following points. First, one does not need the total internal reflection of the incident light, significantly simplifying the illumination setup and enabling the use of cost-effective light sources, such as LEDs. Second, unlike zero-mode-waveguide, which measures single molecules at the bottom of nanoscopic wells, ³¹⁾ a nanogap-junction offers an open 2D space where the lateral movement of single molecules can be tracked. It enables to monitor the lateral positions, movement, and binding/unbinding of membrane-bound molecules in real time. These features should provide a uniquely versatile platform for biosensing and analytical applications. Nanogap-junction can also be used for observing reconstituted membrane proteins. Single molecules of Rh could be observed in a nanogap-junction with a heightened S/B ratio compared with SLB facing a bulk aqueous solution. ³²⁾

Zoom In Zoom Out Reset image size