Principle of bimFCS analysis, FCS curves and bimFCS analysis of differently anchored membrane proteins.

<p>(A) Normalized FCS curves for several pixel-bin sizes (bin 2, 4 and 6) from LissRhod PE-lipids freely diffusing in a supported lipid bilayer (SLB) with LissRhod PE:DOPC ratio of 1:10000. (B) Apparent diffusion coefficients (<i>D</i>_<i>app</i>, top) and the corresponding transit time <i>t</i>_<i>D</i> (bottom) are plotted versus observation area <i>ω</i>² for LissRhod PE in a SLB (black, filled triangles, <i>N</i> = 9) and overlaid with results from two Monte-Carlo Simulations. One simulation considers the microscope optics by convoluting the pixels with the PSF (red, empty squares), while the other one is analyzed as super-resolution (SR) without convoluting with the PSF (green, empty circles). (C) Comparison of the normalized FCS curves (<i>ω</i>² = 0.087 <i>μm</i>²) obtained from mGFP-GPI (green, empty circles), mGFP-GPI with COase treatment (blue, empty triangles) and mGFP fused LDL-R transmembrane domain (mGFP-TM; black, empty squares) in PtK2 cells at 37°C. Data from GPI-anchored and LDL-R anchored mGFP proteins is fitted with one and two diffusion-component square-pinhole FCS functions respectively. (D) <i>D</i>_<i>app</i> (top) and the corresponding <i>t</i>_<i>D</i> (bottom) versus <i>ω</i>²relationships of mGFP-GPI (green, empty circles), mGFP-GPI with COase treatment (blue, empty triangles) and mGFP-TM (black, empty squares) in PtK2 cells. A linear fit through the <i>t</i>_<i>D</i> versus <i>ω</i>²relationship obtained from mGFP-GPI (<i>N</i> = 9) yields a time-axis intercept <i>t</i>₀ of 12±4 <i>ms</i> whereas mGFP with transmembrane anchor (<i>N</i> = 3) gave a negative time-axis intercept, <i>t</i>₀ = -60±10 <i>ms</i>. In cells whose cholesterol has been oxidized using COase (<i>N</i> = 11), <i>t</i>₀ is decreased to be 7±2 <i>ms</i>. (E) Normalized FCS curves (<i>ω</i>² = 0.053 <i>μm</i>²) of Kinetic Monte-Carlo simulations for diffusers transiently trapped in submicroscopic domains (green, empty circles, <i>d</i> = 16%,<i>l</i> = 30<i>nm</i>), weakly interacting with such domains (blue, empty triangles) or interacting with and hopping over fences (black, empty squares, <i>a</i> = 75<i>nm</i>). Data for fence simulation is fit globally using two-component square-pinhole FCS function with held free diffusion coefficient for the fast component of the curve. (F) <i>D</i>_<i>app</i> (top) and the corresponding <i>t</i>_<i>D</i> (bottom) versus <i>ω</i>² relationships obtained from SR analysis of simulations for diffusers transiently trapped in submicroscopic domains (green, empty circles), weakly interacting with such domains (blue, empty triangle) or interacting with fences (black, empty squares) (<i>N</i> = 3). Linear fit of the <i>t</i>_<i>D</i> versus <i>ω</i>² relationship beyond optically accessible size (red, dashed line) for transiently trapping simulations yields a time-axis intercept <i>t</i>₀ of <i>27 ms</i> for strong and 15 <i>ms</i> for weak interaction simulations. Linear fit for the results of fence simulations showed a negative time-axis intercept t₀ = -16 <i>ms</i>.