Supplementary Figure S1: Schematic view of the confocal laser scanning STED microscope used for STED-RICS. For a detailed description of our

Transcription

1 Supplementary Figure S1: Schematic view of the confocal laser scanning STED microscope used for STED-RICS. For a detailed description of our home-built STED microscope used for the STED-RICS experiments, see Methods.

2 Supplementary Figure S2: Dependence of the correlation amplitude on the STED beam power and wavelength in STED-FCS measurements. Static FCS measurements were performed on a sample consisting of a supported DOPC bilayer on a microscope slide containing Atto647N-DPPE as a fluorescence marker. The fluorescence emission was recorded with a time-correlated single photon counting (TCSPC) system (Simple-Tau 152, Becker und Hickl GmbH, Berlin, Germany) and autocorrelated with the software supplied by the company. A linear dependence of the autocorrelation amplitude, G(0), which is proportional to the inverse average number of molecules, 1/N, in the imaging volume, on the STED power, I, is expected from a conversion of the resolution scaling squareroot law 20 : 1 I 1 N I The proper resolution scaling behavior is only observed with a STED wavelength of 780 nm. At the lower wavelengths, saturation occurs, presumably due to excitation rather than depletion by the STED beam. S

3 Supplementary Figure S3: Resolution of the STED microscope as a function of the STED beam power. The point spread function of the STED microscope was determined by imaging individual Atto647N dye molecules attached to a glass surface, using exc = 640 nm and STED = 780 nm. Square regions ( nm 2 ), each containing only a single molecule, were chosen manually from larger frames. The intensity distributions were fitted with a two dimensional Gaussian using Matlab (Matlab, The Math Works, Natick, MA, United States). Their full widths at half maximum (FWHM), plotted as solid squares, were calculated as an average over 20 molecules, the standard error of the mean is indicated by the error bars. We performed a fit (red line) according to the resolution scaling square root law, d, 2NA 1 I / I with resolution, d, emission wavelength,, STED beam intensity, I, and characteristic saturation intensity of the dye, S I, representing the STED beam intensity that reduces the dye emission by a factor of two 25. S

4 Supplementary Figure S4: Comparison of diffusion times determined by STED-FCS and STED-RICS at various STED intensities. To compare the diffusion times measured by the STED-RICS method to a set of static STED-FCS measurements on the same region of the lipid bilayer, the FCS data were fitted according to a model that takes into account diffusional dynamics, triplet state population and a kinetic term representing changes in the fluorescence brightness, 1 G GD GT GK, N 1 with diffusional term G D, 1 / D triplet term 1 T G T exp / T 1 T, and kinetic term 1 K exp G. K / The characteristic time due to triplet state population, T, was fixed at 5 μs and the characteristic time for changes in the fluorescence brightness, K, was limited to a range of μs 13. The diffusion coefficient of (3.4 ± 0.9) µm 2 s -1 resulting from these fits (dashed red line) is in good agreement with the one obtained by the RICS measurements, i. e., (3.0 ± 0.4) µm 2 s -1 (solid black line). K

5 Supplementary Figure S5: FRAP measurement on a lipid bilayer. FRAP reference measurements were performed on a spinning disk confocal microscope (Andor Revolution XD, BFI Optilas GmbH, Gröbenzell, Germany). A circular spot (diameter 3.5 µm) within the lipid bilayer was bleached with 640 nm light. The fluorescence recovery was measured with a time resolution of 86 ms. The recorded intensity signal was background corrected and fitted by a simple diffusion model 26. Results from several measurements were combined to yield an average diffusion coefficient of (2.9 ± 0.4) µm 2 s -1.

6 Supplementary Figure S6: The effect of the pixel size on the correlation of a STED-RICS measurement. An important consideration for both conventional RICS and STED-RICS measurements is the use of appropriate pixel sizes. To be able to temporarily resolve the particle motion, the pixel dwell time, p, must be smaller than the diffusion time, p < D. In laser scanning microscopy, the pixel size, δ p, is related to the scanning velocity of the focus, v s, and the pixel dwell time, p, by δ p = v s p. For two-dimensional diffusion with coefficient D, the particle displacement, Δ, is determined by Δ = (4D D ) 1/2. To be able to spatially resolve the motion, the pixel size should be chosen such that δ p < Δ. As an example, the pixel-to-pixel correlation of a STED-RICS measurement is shown for a lipid bilayer labeled with DPPE-Atto647N, using 200 mw of STED power and various pixel sizes. The correlation amplitude is observed to increase with decreasing pixel size and saturates at 10 nm. Choosing an even smaller pixel size will not improve the precision of the measurement because the problem is oversampled; so 10 nm is the optimal pixel size under the applied conditions.

7 Supplementary Figure S7: Screen shot of Matlab routine for (STED-)RICS analysis. After loading and executing main.m within the Matlab environment, a graphical user interface appears as shown. To start a new session, click the new button on the upper left. The script will prompt for a location to save the results. Later, the session can be continued via the load button. Click browse on the upper right to point to a TIF stack containing the image data for correlation analysis. By clicking the load button, the content of this file will be loaded into memory. Also, a figure called DATA will appear, displaying the image averaged over all frames. Note: For testing purposes, a model dataset called testdata.tif is supplied with this script. If the entire image should be analyzed, keep the selection full image ; usually, however, a region of interest (ROI) is selected by clicking the select button. A rectangle can be drawn on the DATA window, double clicking on the rectangle selects the region and writes the coordinates to the edit fields left of the select button. Alternatively, coordinates can be entered into the corresponding edit fields and displayed on the image in the DATA window by clicking the display button. Note: When performing correlation analysis within a ROI, the coordinates in the edit fields will be used, and not the region displayed in the DATA window.

8 Select if subtraction of an immobile fraction should be performed. Note: Subtraction of an immobile fraction can alter the correlation amplitude. Select if bleaching correction should be performed. Note: Bleaching correction can alter the correlation amplitude. Select the range of pixels and lines to be used in the correlation analysis. Usually the number of pixels and number of lines are chosen to be identical. The frame range will not affect the correlation range itself, since no frame-to-frame correlation is performed. Multiple frames are used only to improve statistics and to subtract an immobile fraction. Note: If the frame range is set to 0, all frames will be used in the analysis. Correlation analysis is started by clicking the correlate button. A window named CORR will open and display the result of the image correlation. Note: Image correlation may take a while, depending on the settings and size of the input dataset. Furthermore, the result window on the lower right will be updated, displaying all settings used for the analysis. Each correlation analysis will be added as an individual item to the listbox on the left. All results will automatically be written to the session file selected in the beginning. The result of each correlation can be viewed and/or deleted by selecting it from the item list and clicking the load or delete button on the lower left. For the image correlation analysis, a model equation can be entered into the fit eq field. To specify which parameters are varied and which are held fixed during the fit, they need to be specified in the corresponding edit fields, separated by commas. Note: The independent variables always have to be named x for image pixel and y for image line. By clicking the set button, a dialog will appear and ask for the start values and boundaries of the variable parameters as well as the values of the fixed parameters. The equation can be saved via the save button and opened later using the load button. Note: For testing purposes, a model equation called eq_diff2d_kin.mat is supplied with this script. Fitting starts by clicking the fit button. A window named FIT will open and display the correlation data as circles overlaid with the fit result as a surface plot. Also, the result window on the lower right will be updated, displaying all settings as well as the result of the fit. The result of each fit can be viewed by selecting it from the item list and clicking the load button on the lower left. Along with the correlation data, all fit results will be automatically written to the session file selected in the beginning.

9 Supplementary References 25. Wildanger, D., Medda, R., Kastrup, L., & Hell, S. W., A compact STED microscope providing 3D nanoscale resolution. J. Microsc. 236, (2009). 26. Yguerabide, J., Schmidt, J. A., & Yguerabide, E. E., Lateral mobility in membranes as detected by fluorescence recovery after photobleaching. Biophys. J. 40, (1982).