Dynamic clustered distribution of hemagglutinin resolved at 40 nm in living cell membranes discriminates between raft theories

Hess et al. 10.1073/pnas.0708066104.

Supporting Information

Files in this Data Supplement:

SI Text
SI Movie 1
SI Movie 2
SI Figure 5
SI Figure 6
SI Figure 7
SI Figure 8
SI Figure 9
SI Figure 10
SI Figure 11
SI Figure 12
SI Figure 13
SI Table 1

SI Figure 5

Fig. 5. Distance histograms for a simulated random distribution of 5,000 molecules within a 40 mm ´ 40 mm square area localized with 28-nm precision. (A) The histogram for immobile molecules (i.e., simulating fixed cells) shows a sharp decline as a function of distance at small distances followed by an approximately linear increase at larger distances. The amplitude of the peak declines, but the linear profile is invariant, as a function of time delay between frames (curves in green, red, blue, purple, and yellow correspond to time delays of k = 1, 2, 3, 4, and 5 frames, respectively). (B) For molecules mobile with D = 0.03 mm²/s, the distance histogram at short distances (<1 mm) changes significantly as the time delay between frames increases, shown by the direction of the black arrow. (C) For larger diffusion coefficient D = 0.06 mm²/s, the peak in the distribution of distances occurs at a larger distance than for D = 0.03 mm²/s, comparing curves with the same time delay (indicated by the same color). Molecules were simulated to have a 1% chance of activation and ~30% chance of photobleaching, per frame.

SI Figure 6

Fig. 6. Simulated distance histograms are well described by an analytical function (Eq. S10A). The distance histogram for 5,000 molecules with a 1% chance of activation and ~30% chance of photobleaching per frame. (A) Immobile molecules. (B) Molecules mobile with D = 0.03 mm²/s. (C) Same histogram and fit shown in B replotted as a function of distance squared, including distances greater than 40 nm.

SI Figure 7

Fig. 7. Quantification of simulated photobleaching using the distance histogram. The effect of photobleaching on the histogram of immobile molecules is to decrease the amplitude of the histogram at short distances. (A) The number of events (points) in the peak of the distribution (summed over distances from 0-0.18 mm) decays exponentially (red curve) as a function of k, the delay in frames. (B) The decay constant j (black points) was determined from fitting the exponential decay of the distance distribution and is equal to the photobleaching probability per frame (i.e., the value chosen explicitly in each simulation) in the limit as j << 1. For j > 0.1, the value is approximately equal to 1.1 to 1.2 times the photobleaching probability. For reference, the red curve shown is a linear fit with slope 1.17 and zero intercept.

SI Figure 8

Fig. 8. Quantification of photobleaching using the measured distance histogram. The area under the peak of the measured distance histogram (points) for fixed cells (summed over distances from 0 to 0.18 mm) was fitted with an exponential decay (red curve). The best fit by least squares resulted for j= 0.35 ± 0.03 frames^-1 and corresponds to an average time before photobleaching of 0.54 ± 0.05 s.

SI Figure 9

Fig. 9. Analysis of measured distance histograms for HA in live cells using an analytical function (Eq. S10A). The measured distance histograms for all live cell data taken at a single frame rate (0.19 s per frame) are shown for various time delays (points) along with fits using Eq. S10A (lines). Modest agreement was observed between the measured histograms and the analytical function. The fitting parameter s² was used to estimate the diffusion coefficient of HA.

SI Figure 10

Fig. 10. The distance squared as a function of time delay between frames (points) was determined from fitting the measured distance histograms for HA in live cells using Eq. S10A. The distance squared increases approximately linearly with time delay. Each value of distance squared as a function of time delay was used to calculate the diffusion coefficient for HA using Eq. S13 (points).

SI Figure 11

Fig. 11. Comparison of rendering methods for positions of molecules localized by FPALM. (Left) Molecules are plotted as Gaussian spots with integrated area proportional to the estimated number of photons collected from that molecule, and a width equal to the localization precision [calculated from the diffraction-limited resolution, the number of photons, and the measured background noise per pixel using the formula of Thompson et al. 2002 (Thompson RE, Larson DR, Webb WW (2002) Biophys J 82:2775-2783)]. (Right) Molecules are plotted as Gaussian spots with amplitude proportional to the number of photons, and a fixed width of 40 nm. (Upper) Data from a whole HAb2 fibroblast expressing PA-GFP-HA. (Lower) Zoom-in of the portion of the image contained in the green box.

SI Figure 12

Fig. 12. Additional examples of elongated clusters of HA imaged by transmission electron microscopy after fixation, labeling with anti-HA primary and 10-nm colloidal gold tagged secondary, followed by critical-point drying. Selected elongated clusters are marked by red arrows.

SI Figure 13

Fig. 13. Additional examples of elongated clusters of HA imaged by FPALM in a live HAb2 cell expressing PA-GFP-HA. The FPALM image from SI Fig. 11 with fixed spot width of 40 nm was inverted and its brightness and contrast linearly adjusted. Selected elongated clusters are marked by red arrows.

SI Movie 1

SI Movie 1. Positions of PA-GFP-HA molecules localized in living cell imaged by FPALM. The focal plane was positioned near the lower (coverslip-proximal) surface of the cell. The white dots are molecules localized in the given frame, which fade from gray to black after eight frames. The red structures show the positions of all molecules localized during the entire acquisition, to provide a context for the motions of the molecules observed for a few successive frames. The number in the upper left of the image is the time in seconds. The short yellow bar is 1 mm.

SI Movie 2

SI Movie 2. Example of build-up of FPALM image data from a live fibroblast cell expressing PA-GFP-HA. The data analysis routine identifies single molecules, localizes (green boxes) or rejects (yellow and red boxes) them, and builds up a map of plotted positions of molecules, weighted by the number of photons detected from each molecule. Red boxes indicate that the observed object within the box was not large or bright enough (too few pixels above a threshold) to decisively be a single molecule. Yellow boxes indicate that the object within the box was too large (too many pixels above a threshold) to be a single molecule. Frames are approximately 0.19 sec apart in real time. The full width of each window is 53.2 mm, including a 5-pixel pad on either side of the actual image.