Fig 1.
Representative confocal fluorescence images of lipid layers formed on PDMS, ox-PDMS, and glass immediately after preparation.
A uniformly fluorescent lipid layer was present on all the surfaces after the fusion of SUVs. The layer on PDMS appeared to have a qualitatively lower intensity compared to the layers on ox-PDMS and glass. Unfused small unilamellar vesicles (SUV) appeared as bright compact spots on the lipid layers. The lower images are schematic diagrams of the configuration of the lipid molecules on the respective surfaces. Lipids form a monolayer on the PDMS, bilayers on ox-PDMS, and bilayers on glass. Scale bar 10 μm.
Fig 2.
Representative confocal fluorescence images of the lipid layers on PDMS, ox-PDMS, and glass over time.
Twelve hours after formation, all the lipid layers appeared similar to when they were first prepared. A) The monolayer on PDMS remained homogeneous and unchanged for 132 hours. B) The lipid bilayer on ox-PDMS appeared mottled at 60 hours. C) The homogeneity of the lipid bilayer on glass remains largely unchanged over the course of 132 hours. A large number of adsorbed SUVs form on the surface. Scale bar 10 μm.
Fig 3.
Correction for the effects of surface position on the values of intensity.
A) Three dimensional representation of raw pixel intensities from a confocal image of a lipid bilayer on ox-PDMS. The x- and y-axis correspond to the pixel numbers while the z-axis corresponds to the intensity of the pixels. The image was 50.7 μm × 50.7 μm. It is clear that there are pixel-to pixel variations and a large-scale gradient that makes the plane of intensity values appear ‘tilted’. The tilt angle is relatively small ~ O(0.01) radians. Large peaks significantly above the mean fluctuation of the pixel intensities correspond to adsorbed SUVs. B) The result of a plane fit to the pixel intensities. The plane is a good representation of the position z(x,y) of the surface. C) A three-dimensional representation of the corresponding matrix of correction factors C(x,y) to account for the non-planar position of the substrate. D) The intensities after multiplication with the correction factors, demonstrating the removal of the large-scale gradient due to the varying position of the surface, while preserving the pixel-to-pixel fluctuations. S1 Fig.
Fig 4.
Histogram of pixel intensities of lipid bilayers on ox-PDMS and lipid monolayers on PDMS.
Each histogram is obtained from N = 10 images, corresponding to a total count of ~ 2.6 × 106 pixels per histogram. The orange continuous line is a fit of a normal distribution to the histogram for the monolayers. The green continuous line is a fit of a normal distribution to the histogram for the bilayers. The mean intensity of the monolayer, μm = 1360 is half the mean intensity of the bilayer, μb = 2591. The histogram of pixel intensities between the bilayers and monolayers do not overlap up to ~ 2.5σ from the mean. The clear separation of the distribution of pixel intensities allows classification of pixels as being monolayers or bilayers in images that potentially have mixed compositions of monolayers and bilayers.
Table 1.
Table of intensity values employed to classify pixels in lipid layers.
Fig 5.
Quantification of the evolution of lipid bilayers on ox-PDMS as a function of time.
A) The normalized area fraction of pixels with intensities consistent with lipid bilayers are represented by solid squares (■), the normalized area fraction of pixels with intensities consistent with lipid monolayers are represented by solid triangles (▲), and the normalized area fraction of pixels with indeterminate intensities are represented by solid circles (●). Standard deviations of the mean are represented by the corresponding error bars. The area fraction of bilayer was close to 100 percent for the duration of the 132-hour experiment. B) Plot of the histogram of pixel intensities (from N = 10 images at each time point) of lipid layers on ox-PDMS at 0 hours, 60 hours, and 132 hours after preparation. The peak of the distribution shifts with time signifying that the layer as a whole becomes less bright. The distributions show a longer right tail. The combination of movement of the peak of the distribution towards lower values and the development of a right tail should result in regions that are brighter when compared to the background, i.e. the mottled appearance of the layer. Note however, that the majority of pixels had intensities consistent with bilayers. The number of pixels classified as indeterminate increases slightly, and none of the pixels had intensities consistent with monolayers.
Fig 6.
Measurement of the diffusive mobility of lipids on the lipid layers.
Plot of the fluorescence recovery curves on, A) ox-PDMS, and B) glass, immediately after preparation of the lipid bilayers. Each plot shows recovery curves from FRAP measurements on ten different locations. The recovery curves are highly reproducible. Insets are representative images before the bleach pulse, the first image after the bleach pulse, and the final image of the time series at 24 seconds. Scale bars 2 μm. C) Plot of the diffusion coefficient versus the time since preparation of the lipid layer. The black bars are for the bilayers on ox-PDMS, the gray bars are for bilayers on glass, and the cross-hatched bars are for monolayers on native PDMS. The diffusion coefficient does not change significantly for the lipid monolayers over the course of 132 hours. The diffusion coefficient of the lipids in the bilayer on ox-PDMS and glass were similar and higher than the diffusion coefficient of the lipids in monolayers on native PDMS for the first 36 hours. The diffusion coefficient however, dropped by 20 percent at 60 hours and then matched those of the lipid monolayers prepared on PDMS. Each bar is a mean of N = 10 measurements. The error bars are standard deviations of the mean.
Fig 7.
Confocal fluorescence images of the evolution of lipid bilayers on ox-PDMS not subjected to solvent sonication.
A) The bilayer looks uniformly fluorescent immediately after the fusion of SUVs. B) At 60 hours, large areas of lipids seemed to have peeled away, revealing the non-fluorescent PDMS substrate (black patches). C) The dark regions became more extensive at 132 hours. Scale bar 10 μm.