Figure 3. MT invasion of spines is correlated with BDNF-induced increase of PSD-95. A, An example of two hippocampal dendritic spines transfected with GFP-PSD-95 and mCherry-α-tubulin. The spine on the right is shown in a series of time-lapse images during MT entries; these time points correspond to purple arrows in A′. A′, A kymograph of the PSD-95 punctum in the spine shown in A. B, C, Two other examples of MT-invaded spines after BDNF treatment. D, A section of dendrite from a neuron transfected as in A. D′, A series of images showing transport of a PSD-95 punctum (top, purple arrowheads) into a spine head. E, Spines show no enlargement 20 min after BDNF treatment. F, G, No significant change in PSD-95 level was detected at 20 min after BDNF treatment in spines devoid of MTs or in spines invaded by MTs only during the control time period. H, I, PSD-95 levels increased significantly both in spines that had MT entry before and after BDNF (H) and only after BDNF (I) (two-tailed Wilcoxon signed rank test). Numbers in black boxes (F–I) show the average value of each dataset. Symbol colors in F–H correspond to line colors in J. J, Percentage change in PSD-95 positively correlates with the percentage chance of an MT being present in a spine only in spines that experience post-BDNF MT invasions. (nonparametric correlation Spearman's r = 0.6845, p < 0.01). Scale bars: A, 2 μm; D, D′, 1 μm.