After 15 minutes of washing with an external solution that blocks most NMDAR-mediated synaptic Ca2 influx (5 mM Mg2 /0.6 mM Ca2 )

VR23 Recurring steps ANOVA was executed (p,.0001), followed by Bonferroni post hoc examination. (D) In DIV7 (light gray) and DIV21 (��-Amatoxin darkish gray) neurons, FRET performance decreases after 1 min Glu/Gly stimulation and 5 min 0Mg2+/Gly stimulation. doi:10.1371/journal.pone.0112170.g002 One of the hallmark features of the NMDAR in developing postnatal hippocampal and cortical neurons is the increasing ratio of GluN2A/GluN2B [103]. This developmental switch was also seen in our cultures we measured a two-fold increase in the GluN2A/GluN2B ratio from DIV7 to DIV21 (Figure S5 in file S1). This raises the question of whether such change in ratio would affect the NMDAR activity-dependent disruption of the NMDARPSD95 interaction. We thus compared two developmental stages, DIV7 and DIV21, and found no significant difference in the stimulus-induced changes in FRET efficiencies between the two ages (Figure 2D). It is probable that PSD95 can interact with the NMDAR via both GluN2 subunits [5,15,16,31]. We tested whether Ca2+ influx in HEK cells can disrupt the NMDARPSD95 interaction through either GluN2 subunit. The results indicated that i) both GluN2A and GluN2B could support FRET between GluN1-GFP and PSD95-mCherry and that ii) this interaction was disrupted by Ca2+ influx (Figure 2E). Taken together, these results show that NMDAR/PSD95 interaction is transiently disrupted upon NMDAR stimulation both in young and mature hippocampal neurons, a process that likely reflects dynamic interactions via both GluN2A and 2B. It should be noted that the measured decrease in FRET between PSD95 and the NMDAR could also involve a conformational change.PSD95-S73D mutant had little basal interaction with the NMDAR, and no change was seen upon stimulation (Figure 3A). However, PSD95-S73A did interact with the receptor in unstimulated neurons and stimulation did not disrupt the interaction (Figure 3A). This suggests that CaMKII phosphorylation of PSD95 at S73 regulates the NMDAR/PSD95 interaction in DIV21 cultured neurons. In contrast, in young neurons (DIV7), CaMKII-dependent phosphorylation of PSD95-S73 did not seem to regulate NMDAR/PSD95 interaction in the same manner. Indeed, both PSD95-S73D and S73A mutants interacted with the NMDAR to a similar extend as PSD95-WT in basal conditions, and a disruption of the interaction was caused by 1 min Glu/Gly stimulation for the S73D, but not for S73A (Figure 3B). One possible interpretation of these results is that the changing ratio of GluN2 subunits during development alters the mechanisms of PSD95-NMDAR interaction. Indeed, phosphorylation of PSD95S73 is unlikely to impact on it binding to GluN2B [16], which is the dominantly expressed GluN2 subunit at DIV7. Overall, these results suggest that while CaMKII can impact on NMDARPSD95 interaction, other NMDAR-dependent signaling processes are likely involved, since for instance, the pre-phosphorylated PSD95 (S73D) still dissociates during NMDAR stimulation.We next sought to understand the mechanisms by which the interaction between the two proteins is regulated upon NMDAR stimulation. Since CaMKII is known to play a role in regulating PSD95 trafficking [5], we used KN93 to test whether CaMKII influences NMDAR/PSD95 interaction.

After fifteen minutes of washing with an external solution that blocks most NMDAR-mediated synaptic Ca2+ influx (5 mM Mg2+/.six mM Ca2+), GluN1-GFP lifetime lowered, suggesting a partial recovery of the interaction between PSD95 and the NMDAR.