Substrate-supported planar lipid bilayers (SPBs) are being utilized as a versatile model system of the biological membrane. However, the proximity between the solid support and membrane limits utility of SPBs for the functional analyses of membrane proteins. Here, we present a model membrane that can enlarge the distance between the substrate surface and the membrane by combining a stable scaffold of polymerized lipid bilayer with a hydrophilic polymer brush. A micropatterned SPB was generated by the lithographic polymerization of diacetylene lipids and subsequent incorporation of natural (fluid) lipid bilayers. Hydrophilic polymer brush of poly-2-methacryloyloxyethyl phosphorylcholine (poly(MPC)) was formed on the surface of polymeric bilayer by the in situ atom transfer radical polymerization (ATRP) in aqueous solution, in the presence of embedded fluid lipid bilayers. A model membrane protein (Haloquadratum walsbyi bacteriorhodopsin: HwBR) could be reconstituted into the polymer brush-supported bilayers with significantly reduced immobile molecules. Furthermore, the polymer brush terminals could be functionalized by successively polymerizing MPC and 2-aminoethyl methacrylate (AMA). The reactive amine moiety of poly(AMA) enables to conjugate a wide range of biological molecules and surfaces to the membrane. The combination of micropatterned bilayer and polymer brush mimics the two- and three-dimensional structures of the biological membrane, providing a platform to assay membrane proteins in a truly biomimetic environment.

Phospholipid bilayers spontaneously spread on a hydrophilic substrate such as glass in aqueous solution due to the energetic gain of surface wetting. This process (selfspreading) was utilized to form a patterned model biological membrane containing reconstituted membrane proteins. A mechanically stable framework of polymerized lipid bilayer was first generated by the lithographic polymerization of a diacetylene phospholipid. Then, natural lipid membranes (fluid bilayers) were introduced into the channels between polymeric bilayers by the self-spreading from a phospholipid reservoir. The spreading velocity could be fitted into a slope of-0.5 in a double logarithmic plot versus time, due to the balance between spreading force and resistive drag. The preformed polymeric bilayer accelerated the spreading by the energetic gain of covering hydrophobic edges with lipid bilayer. At the same time, the domains of polymeric bilayer obstructed spreading, and the spreading velocity linearly decreased with their fractional coverage. Above the critical coverage of ca. 50%, self-spreading was completely blocked (percolation threshold), and fluid bilayer was confined in the polymer-free regions. Nonspecific adsorption of lipids onto the surface of polymeric bilayers was negligible, which enabled a heightened signal-to-background ratio in the reconstitution and observation of membrane proteins. Self-spread bilayers had a higher density of lipids than those formed by the spontaneous rupture of vesicles (vesicle fusion), presumably due to the continual supply of lipid molecules from the reservoir. These features give the selfspreading important advantages for preparing patterned model membranes with reconstituted membrane proteins.