Bilayer edges catalyze supported lipid bilayer formation

Abstract

Supported lipid bilayers (SLB) are important for the study of membrane-based phenomena and as coatings for biosensors. Nevertheless, there is a fundamental lack of understanding of the process by which they form from vesicles in solution. We report insights into the mechanism of SLB formation by vesicle adsorption using temperature-controlled time-resolved fluorescence microscopy at low vesicle concentrations. First, lipid accumulates on the surface at a constant rate up to approximately 0.8 of SLB coverage. Then, as patches of SLB nucleate and spread, the rate of accumulation increases. At a coverage of approximately 1.5 x SLB, excess vesicles desorb as SLB patches rapidly coalesce into a continuous SLB. Variable surface fluorescence immediately before SLB patch formation argues against the existence of a critical vesicle density necessary for rupture. The accelerating rate of accumulation and the widespread, abrupt loss of vesicles coincide with the emergence and disappearance of patch edges. We conclude that SLB edges enhance vesicle adhesion to the surface and induce vesicle rupture, thus playing a key role in the formation of continuous SLB.

Figure 1

Schematics of sample flow cell drawn to scale. The flow chamber is defined by a Teflon channel pressed against a borosilicate coverslip and secured with four screws to a brass platform that rests on the stage of an inverted microscope. Sample solution drips into one reservoir and is removed at the rim of the other reservoir, allowing gravity to drive flow through the 1 mm-deep channel. Sample temperature is monitored via a thermocouple inserted downstream from the observed region. Temperature is maintained by cartridge heaters in the brass platform on either side of the channel using a PID temperature controller and RTD.

Figure 2

Fluorescence video microscopy of vesicles adsorbing and rupturing to form a supported lipid bilayer (see Movie S1). (A–H) Sample frames taken at successive time points marked by corresponding letters in Fig. 3. The field of view brightens as vesicles adsorb to the glass surface (A and B). Dark patches appear where vesicles rupture to form SLB whereas, elsewhere, vesicles continue to adsorb (C). SLB patches continue to nucleate and spread (D–G) until the entire surface is covered in SLB (H).

Figure 3

Mean intensity versus time for a representative sample. Lettered points correspond to images in Fig. 2. Mean intensity increases quickly at first, as buffer is replaced by fluorescent vesicle suspension, then more slowly, as vesicles adsorb on the substrate, then quickly again. It is in this late phase of accelerating adsorption that resolvable patches of SLB first appear (see Fig. 2, C and D). As patches spread, the mean intensity reaches a maximum and drops rapidly to a stable value. The fluorescent vesicle suspension is then replaced by buffer and the associated decrease in intensity provides an in situ measure of the bulk fluorescence. Finally, the sample is photobleached to test the mobility and continuity of lipid on the surface.

Figure 4

Mean intensity (corrected for background and normalized to I_SLB; see the Supporting Material) versus time (normalized to t_max) for all 16 concentrations (inset) and for one, representative sample (same as in Fig. 3). All samples exhibit a linear increase in intensity at early times (the least-squares linear fit for I/I_SLB < 0.8 is highlighted and extended in yellow). This is followed by a period of accelerating adsorption that ends abruptly. From its maximum, intensity decreases rapidly as the SLB becomes continuous and vesicles desorb. At the highest concentrations, some vesicles remain adsorbed to the bilayer (until rinsed away with lipid free buffer) causing intensity levels to hover above I_SLB toward the end of the time range depicted.

Figure 5

Quantitative characteristics of the time course of mean intensity, I(t), as a function of bulk lipid concentration. (A) The initial rate of adsorption (slope of the linear fit for I/I_SLB < 0.8) is directly proportional to concentration at all but the highest concentrations. (B) The time to maximum mean intensity is inversely dependent on concentration. (C) The characteristic time of an exponential fit to the rapid decrease after maximum intensity is weakly dependent on concentration. (D) The maximum intensity varies, but is not sensitive to concentration. (See the Supporting Material for a comparison of the variability of I_max within a sample and across samples at different concentrations.) Horizontal error bars are based on reproducibility of bulk fluorimetry measurements (M.S.E. for n = 9). Vertical error bars are based on the video sampling rate for time measurements and the uncertainty in the measured background for intensity measurements.

Figure 6

Spatial variation in the time evolution of surface fluorescence. (A) Successive intensity traces along a horizontal line (taken at 1-min intervals) are arrayed vertically, revealing triangular dark (bilayer-covered) regions along the bright-dark interface, which attest to the tendency for SLB patches to spread via edge-induced rupture. (B) Despite similar initial adsorption rates, neighboring regions (1 μm diameter, centers indicated by symbols in A) reach different maximum intensities before forming SLB.

Figure 7

Proposed stages of SLB formation from vesicles (labeled in correspondence with Figs. 2 and 3). (B) Vesicles accumulate on the substrate and, in isolation, rupture slowly into patches of SLB. (C) Vesicles accumulate preferentially along edges of SLB patches, where they rupture more quickly and fuse with the adjacent SLB patch. (D and E) SLB coverage reaches a critical point, or percolation limit, where liquid-like coalescence of patches occurs rapidly (at constant bilayer area), abruptly decreasing the total edge length and releasing vesicles into solution. Modeling the associated kinetics requires six rate constants: three reflecting vesicles' higher affinity for bilayer edge than for glass than for bilayer surface (k₃′ < k₁′ < k₄′), and three more for adsorption (k₁), isolated rupture (k₂), and edge-bound rupture (k₅ > k₂) of vesicles.