JCS001404 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure S1 -

  Fig. S1. (A) Histogram of cA₉₅%-values of all FRET-pairs, showing at which acceptor concentrations 95% of E_max were reached. The cA₉₅%-values were analytically derived, using parameters obtained from monoexponential fits, which gave very similar fit results for E_max. Using cA=σR₀², R₀=4.7 nm, we can recalculate the range from cA=2-8 10^-3 to σ=90-360 acceptors/μm². (B) Table of x_micro-values, which give the percentage of indicated proteins being co-clustered in a common microdomain, with densest square packing, while assuming that the remainder is randomly distributed. Values are calculated using Eqn 2, with the corresponding E_max of the FRET-pairs, E_micro=87% and E_bulk=10%. For comparison with matrices in the main text, we used bold characters for values of the same FRET-pairs as in the main text Figs. 1E, 3C, 4C, 5A. However, unlike in those matrices, no statistical tests were performed and therefore no statistical inferences can be made on the basis of the values of bold/regular characters. (C) Calculation of approximated microdomain diameters. Having a constant total area of the microdomains, smaller individual microdomains would lead to a higher offset, cA₀, as they would be populated by a single fluorophore only at low concentrations, which would not lead to FRET. Based on this assumption, we calculated their diameters, d_μ, assuming a circular surface of the microdomains (A_μ=πd_μ²/4). This area was calculated, using A_μ=1/(2σ_cA0 α), with 2σ_cA0: surface density of fluorophores at the offset, in number of fluorophores/μm², i.e. 2cA₀/R₀; α: scaling factor as determined by fits using Eqn 12 as described in Meyer et al., 2006 (data not shown). Thus we calculated typical diameters of around 30 nm. However, comparatively high domain sizes were found for N_GAP-43-mCFP/mCit-tH, N_i2Cγ-mCit/mCFP-tR and N_qCγ-mCit/mCFP-tR (bold), while on the other hand the complementary G-protein anchor-construct/ microdomain-marker pairs had approximately two to three-times smaller domain diameters. These size relationships corresponded to the inverse relationships found for the FRET-efficiencies. Values are rounded to the nearest five.

• Supplemental Figure S2 -

  Fig. S2. Validation of cytometer FRET data, using sensitized acceptor emission FRET-imaging and a FRET-assay for FRET from internal membranes (A) G-protein anchor-construct/microdomain-marker FRET-imaging confirms distinct FRET-levels on the plasma membrane. From left to right showing the FRET-channel images, the FRET-efficiency images and the dependencies of the FRET-efficiency of indicated FRET-pairs on the acceptor-intensity at approximately 500-2000 acceptors/μm². Only regions of interest on plasma membranes with donor mole fractions x_D=0.5±0.17 were analysed, unless otherwise shown by insets of E vs x_D-plots. (B) Subcellular distribution of the internal membrane-marker, mCit-γ, imaged by confocal microscopy. Bar, 10 μm. (C) The histogram of the average cA₀-values reveals two populations, separated at cA<0.0015 (arrow). The cA₀>cA₀=0.0015 are mostly associated with the FRET-pairs containing the internal membrane-marker as one partner (data not shown). (D) The E(0.0015)-matrix was calculated for cA=0.0015, which corresponds to approximately 50’000 acceptor molecules per cell, using the individual fit parameters determined for the FRET-pairs. The differences between the efficiencies of FRET-pairs having the internal membrane-marker as one partner and the remaining FRET-pairs are more enhanced than in the corresponding E_max-matrix (Figs 1E, 3C, 4C, 5A and data not shown). The calculated efficiency is given in percent, errors correspond to the relative errors of the corresponding E_max-values. Importantly, the FRET-values that we calculated for FRET-pairs of any of our anchor constructs with the internal membrane-marker are the maximal values that could be reached, as for any other pair the internal membrane localized fraction would be much lower than for the internal membrane-marker at the same total acceptor expression level. Importantly, all of the FRET-signatures and FRET-relationships described in the main text are preserved. These experiments therefore strongly support our assumption that the plasma membrane is the primary origin of our measured FRET-signal. For comparison with matrices in the main text, we used bold characters for the same FRET-pairs as in the main text Figs. 1E, 3C, 4C, 5A. However, unlike in those matrices, no statistical tests were performed and therefore no statistical inferences can be made on the basis of the values of bold/regular characters (E) Plot of the FRET-efficiency, E, against the apparent, normalized acceptor surface concentration, cA, at a constant donor mole fraction (x_D=0.50±0.17) for cells co-expressing soluble mCFP and mCit demonstrate that the contribution of soluble proteins to the FRET is negligible.

• Supplemental Table S1 -

  Adobe PDF