Extended Data Fig. 2: NPC positional stability.
From: Overlapping nuclear import and export paths unveiled by two-colour MINFLUX
a,b, Confocal images of the EGFP emission from NPCs in a permeabilized U2OS NUP96-EGFP cell (EX = 488 nm) before (a) and after (b) MINFLUX localizations of NbGFP-HMSiR and Imp α-JF549. Typical results from N = 18 cells. c, NPC positions before and after MINFLUX imaging. Confocal spots were fit with a 2D Gaussian distribution to determine the approximate center of the NPC before (red x’s) and after (green circles) the MINFLUX data acquisition. This process was used to identify scaffolds that were sufficiently stable during image acquisition to justify further analysis of the HMSiR localizations. Only well-formed scaffolds identified by HMSiR localizations (see Methods) were analyzed further. The boxed area approximates the image areas of a and b. d, NPC movements during imaging. The distance moved by confocal spot centroids during MINFLUX imaging indicates that ~10% of NPCs (data from 11 cells) were sufficiently stable (shift <40 nm; considered within the error range of the precision for the conditions) for potential analysis. These stable NPCs were invariably near the center of the bottom of the nuclear envelope. e-g, Drift estimated from MINFLUX measurements. To assess NPC scaffold drift using higher precision data, the estimated centers of clusters acquired during the 1st and 2nd half of a MINFLUX HMSiR acquisition were compared. Data were fit using a double-circle algorithm (Extended Data Fig. 3c). e, Scatter plot of HMSiR localizations (cyan for 1st half; magenta for 2nd half) overlapped with the fitted centers (orange circle for 1st half; blue star for 2nd half). Black circles identify sufficiently stable scaffolds [deviations ≤ 8 nm; see g]. f, Distances between early (1st half) and late (2nd half) centroids for all clusters (N = 254) identified in 4 nuclei. g, Early/late centroid distances for those clusters identified in e that were selected for Imp α-JF549 track analysis. Note that the centroid positional accuracy for the two estimates using split datasets is reduced from the single estimate from complete datasets for each pore.
