Biomembrane Fabrication by the Solvent-assisted Lipid Bilayer (SALB) Method

Fabrication of supported lipid bilayers on hydrophilic substrates.

The VF and SALB formation methods were attempted on silicon dioxide and gold and the formation processes were monitored in real-time by the QCM-D measurement technique. The QCM-D instrument measures changes in the resonance frequency (Δf) of an oscillating piezoelectric quartz crystal upon mass adsorption onto the surface of the crystal. In addition, the QCM-D instrument measures the dissipation of the oscillation energy in order to characterize the viscoelastic properties (rigidity and softness) of the adlayer. Vesicle fusion experiments were performed as previously described³¹. Briefly, a baseline was first established for the frequency and dissipation signals in aqueous buffer solution [10 mM Tris, 150 mM NaCl, pH 7.5; (Figure 2A and B)]. Next, small unilamellar DOPC lipid vesicles in the same buffer were injected at t = 10 min (arrow 1) onto silicon dioxide (Figure 2A) and gold (Figure 2B). For vesicle fusion on silicon dioxide, two-step adsorption kinetics were observed with final changes in frequency and energy dissipation of −26 (± 1) Hz and 0.3 (± 0.2) × 10⁻⁶, respectively. These values are consistent with the formation of a supported lipid bilayer²⁹.

As shown in Figure 2B, the addition of the vesicle solution to the gold surface led to a simultaneous decrease and increase in Δf and ΔD signals, respectively, until their values reached −150 (± 10) Hz and (7.5 ± 2) × 10⁻⁶, respectively. These values correspond to the formation of an adsorbed vesicle layer. Thus, as expected, an SLB was not formed on gold via the vesicle fusion method.

QCM-D analysis for bilayer formation on silicon dioxide and gold by the SALB method is presented in Figure 2C and D. On silicon dioxide, final Δf and ΔD shifts of −25.6 (± 0.55) Hz and 0.4 (± 0.27) × 10⁻⁶, respectively, were achieved and these values indicate formation of an SLB. Similar ranges of Δf and ΔD (Δf_Au: −27.3 ± 2.7 Hz, ΔD_Au: 0.48 (± 0.26) × 10⁻⁶) were observed on gold. These results support that the SALB method enables formation of supported lipid bilayers on surfaces which prevent vesicle rupture.

Effect of solvent-exchange flow rate on the quality of supported lipid bilayers.

To determine the optimal conditions for fabrication of high quality supported lipid bilayers via the SALB method, the influence of lipid concentration, solvent-exchange rate and the choice of organic solvent were examined.

Figure 3 shows the change in QCM-D frequency during the final step of the SALB protocol, as performed on silicon dioxide at two different flow rates (100 and 600 µl/min) and two different lipid concentrations (0.125 and 0.5 mg/ml).

When 0.5 mg/ml DOPC lipid was used (Figure 3A), bilayer formation was not affected by the flow rate and a final Δf shift of about −26 Hz was obtained at both flow rates.

By contrast, when a lower lipid concentration was used, the amount of adsorbed lipid after complete solvent exchange was significantly affected by the flow rate (Figure 3B). At an average flow rate of 100 µl/min, bilayer formation was complete (Δf around −26 Hz). However, at a 6-fold higher flow rate (600 µl/min), a complete bilayer was not formed (Δf around −17 Hz). These results provide a guideline for choosing the right experimental conditions for successful bilayer formation using the SALB method with isopropanol as the organic solvent of choice.

Characterization of supported lipid bilayers formed from different lipid concentrations in various alcohol solutions.

Lipid concentration is another parameter that affects the quality of SLBs obtained by using the SALB method. Fluorescence microscopy revealed that, at 0.05 mg/ml lipid concentration, only isolated, sub-microscopic lipid structures were formed (Figure 4A). With increasing lipid concentration used in the SALB procedure, the fluorescence intensity of the lipid structures became more homogenous. At 0.1 mg/ml lipid concentration, there were microscopic lipid patches although the structures did not span across the entire field of view (Figure 4B). However, when 0.25 mg/ml lipid concentration was used, a homogenous lipid bilayer was formed (Figure 4C). Therefore, there is a minimum lipid concentration required to form a complete, full-spanning SLB.

The influence of lipid concentration on the final outcome of the SALB experiments was also examined over a wider lipid concentration range (0.01 to 5 mg/ml) and in different organic solvents (isopropanol, ethanol, and n-propanol). The Δf and ΔD values corresponding to the final step in the SALB procedure are presented in Figure 5.

We defined bilayer formation based on final changes in frequency and energy dissipation between -25 and -30 Hz and less than 0.5 x 10^-6, respectively. Based on these criteria, the optimal lipid concentration range to form a supported lipid bilayer was determined to be between 0.1 and 0.5 mg/ml largely independent of the type of organic solvent. The deviations in the acquired Δf and ΔD shifts outside of the aforementioned range is due to the presence of additional mass (e.g., bilayer stacks),the occurrence of non-bilayer morphologies (e.g., vesicles, worm-like micelles), and the presence of fragmented bilayer islands with incomplete morphology across the substrate.

While the SALB procedure to obtained supported lipid bilayers is relatively robust, the optimal lipid concentration for high-quality bilayer formation may require tuning depending on the specific experimental configuration. Basically, there is not only a minimum lipid concentration but also a maximum lipid concentration required for optimal bilayer formation via the SALB method. The optimal concentration range depends on the flow rate and may be also be affected by the substrate and lipid composition. Empirically, in many cases, we have found that a lipid concentration of 0.5 mg/ml and a flow rate of 100 µl/min is the optimal set of conditions for the formation of a homogenous supported lipid bilayer. However, depending on the lipid composition and flow-cell geometry—the latter of which affects the flow profile during solvent exchange—further optimization of the lipid concentration might be necessary. Thus, we recommend performing pilot SALB experiments using a 0.5 mg/ml lipid concentration and assess the bilayer quality using the QCM-D or fluorescence microscopy techniques. If the bilayers appear incomplete, then the lipid concentration should be increased in 10% increments until satisfactory outcomes are achieved. If the bilayer appears to co-exist with additional lipid structures, then the lipid concentration should be decreased in 10% increments until satisfactory outcomes are achieved.

Fabrication of supported lipid bilayers with various lipid compositions and different cholesterol fractions.

Next, the SALB and VF methods were employed in order to form cholesterol-enriched SLBs. Figure 6 shows representative fluorescence images (100 × 100 µm) of cholesterol-containing supported membranes prepared by the SALB method. The membranes consist of circular shaped dye-excluded domains surrounded by a continuous phase characterized by uniform fluorescent brightness. The dark domains increased in area with increasing cholesterol fraction in the precursor lipid mixture. Next, FRAP measurements were performed in order to examine the fluidity of the lipid bilayer films. The FRAP measurements revealed almost complete fluorescence recovery in the surrounding phase, indicating the lateral mobility of lipids and thus formation of a single lipid bilayer. Since Rh-PE partitions preferentially into the fluid phase, the dark domains are most likely composed of dense cholesterol-enriched structures.

For comparison, the fabrication of DOPC/Chol bilayers using the VF method was also attempted. DOPC vesicles with increasing cholesterol fractions (10 – 40 mol%) were prepared by the vesicle extrusion method. Rh-PE lipid (0.5 wt%) was used as the fluorescent label for imaging. Figure 7 shows representative fluorescence images (100 × 100 µm) of structures created upon incubation of the glass substrates with cholesterol-containing vesicles. FRAP analysis showed the formation of a fluidic lipid bilayer using vesicles that contained 20 mol% Chol or less. However, samples prepared by using vesicles with higher cholesterol fractions did not exhibit recovery, indicating the presence of adsorbed but unruptured vesicles.

The fraction of cholesterol which was ultimately incorporated into the supported lipid bilayers was quantified as a function of the cholesterol fraction which had been included in the precursor lipid mixture in organic solvent in the SALB method or in the vesicles in aqueous solution in the VF method. Using the QCM-D technique, bilayer formation was monitored and then MβCD was added in order to specifically extract Chol from the supported lipid bilayers³². The mass loss due to the removal of cholesterol led to a decrease in the absolute value of the frequency shift (|Δf|) associated with the SLB. The relative positive frequency shifts caused by the MβCD treatment step are shown in Figure 8A. The mole fraction of cholesterol was calculated based on the frequency shift, as presented in Figure 8B.

The cholesterol fraction incorporated into supported lipid bilayers prepared by the SALB method was almost linearly proportional to the cholesterol contents in the precursor lipid mixture. Interestingly, the cholesterol fraction in bilayers prepared by the VF method (vesicles containing up to 20 mol% Chol) was substantially lower than that contained in the precursor vesicles. In fact, the highest fraction of cholesterol obtained by the VF method was only about 10 mol%.

Observation of stripe superstructure in the β-two-phase co-existence region of cholesterol-phospholipid supported lipid bilayers.

SALB experiments were further performed using a lipid mixture with an even higher fraction of cholesterol. When a 4:6 mixture of DOPC and cholesterol was used, a gradual demixing of a uniform liquid phase into two coexisting phases visualized as bright stripe-shaped domains on a dark (dye-excluding) background was observed (Figure 9). It is established that Rh-PE is excluded from cholesterol-rich domains³³, and hence the predominant dark domains which appeared as background are cholesterol-enriched regions. The formation of micron-sized bright domains on a dark background at high cholesterol fraction (>50 mol%) is consistent with the β region in the monolayer phase diagram of cholesterol/phospholipids mixtures^34,35. Moreover, the formation of stripe domains, which arise from a weak line tension, suggests that the mixture is near a miscibility critical point.

Figure 1. Microfluidic chamber for SALB formation in a suitable configuration for epifluorescence microscopy. (A) Commercial microfluidic chamber, (B) Glass coverslip attached onto the adhesive side of the chamber, (C) Complete setup on a microscope holder with tubing connected into the inlet and outlet ports of the chamber, and (D) Peristaltic pump used to control the rate of solvent exchange. Lipids dissolved in isopropanol are injected into the measurement chamber with the aid of the peristaltic pump. Please click here to view a larger version of this figure.

Figure 2. QCM-D analysis of vesicle fusion and SALB experiments on silicon dioxide and gold substrates. QCM-D frequency (Δf, blue) and dissipation (ΔD, red) responses for the third overtone (n = 3) were recorded as a function of time during lipid adsorption onto (A and C) silicon dioxide, (B and D) gold. Panels a and b present the vesicle fusion method. DOPC lipid vesicles were injected at t = 10 min (arrow 1). Panels c and d correspond to the SALB formation method. Arrows indicate the injection of buffer (1), isopropanol (2), lipid mixture [0.5 mg/ml DOPC lipid in isopropanol; (3)] and buffer exchange (4). The dashed curve in panel B corresponds to a control experiment in which lipid was not injected. The final values of Δf and ΔD for each surface are specified. The schematics show the proposed assembled lipid structures as inferred from the final frequency and dissipation shifts. Adapted from reference 24 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 3. Influence of the solvent-exchange rate on the SALB formation process. QCM-D frequency shifts (Δf) corresponding to the final step (see Arrow 4 in Figure 2C) in the SALB method were measured on silicon dioxide at two different exchange rates, 100 and 600 µl/min, using (A) 0.5 mg/ml and (B) 0.125 mg/ml DOPC lipid in isopropanol. The final Δf values are also specified, as compared to the measurement baseline in aqueous buffer solution. Adapted from reference 24 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 4. Threshold of Lipid Concentration for Complete SALB Formation. Epifluorescence microscopy of lipid layers on silicon dioxide prepared by SALB using (A) 0.05 mg/ml ; (B) 0.1 mg/ml; and (C) 0.25 mg/ml lipid concentration. Adapted from reference 26 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 5. Influence of lipid concentration and organic solvent on supported lipid bilayer formation by the SALB method. The final changes in QCM-D (A) frequency and (B) energy dissipation for SALB experiments using various organic solvents, as a function of lipid concentration. The dashed green lines correspond to the expected frequency and dissipation shifts for a complete bilayer (-30 Hz < Δf < -25 Hz and ΔD < 1 x 10^-6). Adapted from reference 26 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 6. Fluorescence recovery after photobleaching analysis of supported lipid bilayers with varying fractions of cholesterol prepared by the SALB method on a glass substrate. (A-E) Fluorescence micrographs of bilayers prepared using various cholesterol fractions in the precursor mixture. Images were recorded immediately (top) and 1 min (middle) after photobleaching. The dark spot in the image center corresponds to the photobleached region. The scale bars are 20 µm. Surface area histograms of individual dye-excluded domains within each sample are also presented (bottom). Adapted from reference 36 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 7. FRAP analysis of cholesterol-containing supported bilayers prepared by the vesicle fusion method. (A-D) Fluorescence micrographs of bilayers prepared using various cholesterol fractions (10 to 40 mol%), in the precursor vesicles. Images were recorded immediately (top) and 1 min (bottom) after photobleaching. The dark spot in the image center corresponds to the photobleached region. The scale bars are 20 µm. Adapted from reference 36 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 8. Quantification of cholesterol fraction in supported lipid bilayers. (A) The positive QCM-D frequency shift upon injection of 1 mM MβCD onto supported lipid bilayers with varying mole fractions of cholesterol in the precursor mixture (between 0 and 50 mol%). (B) Mole percent of cholesterol depleted from the bilayers prepared by the SALB and vesicle fusion methods as a function of the cholesterol fraction in the precursor mixtures or vesicles. Adapted from reference 36 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.

Figure 9. Time dependent evolution of fluorescence microstructure in a supported bilayer of DOPC/Chol (4:6 molar ratio containing 0.5% Rhodamine-PE) prepared by the SALB method. A uniform phase gradually phase separates into a liquid-liquid coexistence region upon complete solvent exchange. Adapted from reference 37 and used with permission of the American Chemical Society. Please click here to view a larger version of this figure.