Figure 1

Super-resolution reveals that Norbin does not colocalize with PSD-95. A) Confocal (left) and 3D-SIM image of Norbin-GFP distribution in a dendrite with spines. B) STED image of immunolabeled Norbin in dendrite and spines. C) Line profiles taken from B) (dots) fitted with a Gaussian function (red) indicate a resolution of ~40 nm in the Norbin channel. In D) two emitters 116 nm apart have been clearly separated. For the profiles, a patch was taken from the image as indicated in B) and integrated along the shorter edge. E) Confocal recordings of Norbin-GFP and PSD-95-mCherry show a complete colocalization in spines. F) Using 3D-SIM, the distributions of Norbin-GFP and PSD-95-mCherry are resolved. G) Immunolabeled endogenous Norbin and PSD-95 recorded using stimulated emission depletion (STED) show the same overall distribution as 3D-SIM images of exogenous expressed Norbin and PSD-95. H-J) Intensity profiles along the dashed lines over spine heads in A-C illustrate that confocal microscopy does not have the resolving power of 3D-SIM and STED to identify the discrete localizations of Norbin (green) and PSD-95 (magenta). K) Norbin is non-uniformly distributed in spines and has a low degree of colocalization with PSD-95, reflected in a median Pearson’s correlation coefficient (PCC) of 0.49 (n = 27) for 3D-SIM and 0.50 (n = 17) for STED. Compared to the 3D-SIM and STED PCCs, the confocal recordings gave a significantly higher median PCC of 0.81 (n = 15) (***p < 0.00001, Kruskal-Wallis test). The PCCs between the 3D-SIM and STED images are not different (n.s p > 0.5, Kruskal-Wallis test). n = number of analyzed spines.