An attractive explanation for the impairment of long-term contextual memory after strong training in Paip2a−/− mice is excessive activity-induced translation in the absence of PAIP2A. It is conceivable that partial reduction of PAIP2A, as in Paip2a+/− mice, might have a smaller effect on translation and thus lead to a salubrious effect on memory.

Reduction in the PAIP2A protein levels in Paip2a+/− mice was confirmed by western blotting ( Figure 3A). It is striking that, while similar freezing was observed 1 hr after strong contextual training, it was enhanced 24 hr after training in Paip2a+/− relative to WT mice ( Figures 3B and 3C). These data are consistent with the idea that complete removal of PAIP2A might cause memory impairment via excessive translation in response to strong training. In accordance with these results, L-LTP elicited by TBS was not impaired in Paip2a+/− relative to WT hippocampal Trichostatin A in vitro slices ( Figure 3D). Because adult neurogenesis contributes to fear memory extinction (Pan et al., 2012), which is impaired in Paip2a−/− mice, we examined neurogenesis in WT, Paip2a−/−, and Paip2a+/− mice.

Progenitor cell proliferation within the subgranular zone of the dentate gyrus was assessed using systemic injection of BrdU followed by immunostaining or by staining for Ki-67, a marker of proliferating progenitor cells. It is surprising that the number of BrdU- and Ki-67-positive cells was reduced in Paip2a−/− but not in Paip2a+/− mice as compared to WT mice ( Figure S4A), suggesting that impaired memory BMS-354825 ic50 extinction in Paip2a−/− of mice might result from reduced adult neurogenesis.

We also examined the memory phenotype of Paip2b−/− mice, although PAIP2B expression in the brain is lower than PAIP2A ( Berlanga et al., 2006). No differences were found in contextual fear conditioning task between Paip2b−/− mice and their WT littermates 1 hr and 24 hr after training ( Figures S4B and S4C, respectively), suggesting that PAIP2B is not involved in translational regulation of learning and memory. Next, it was pertinent to determine whether PAIP2A is controlled in an activity-dependent manner. No phosphorylation of PAIP2A has been reported. However, PAIP2A levels are homeostatically controlled by proteasome-mediated degradation upon PABP depletion in cell cultures (Yoshida et al., 2006). Therefore, cultured neurons were depolarized with KCl for 5 min to study the effect on PAIP2A levels. PAIP2A protein levels decreased to 69.6% ± 3.3% of baseline 1 min after KCl-induced depolarization and were further reduced to 59.3% ± 4.0% after 10 min. PAIP2A levels returned to normal after 30 min (Figure 4A). Similarly, activation of NMDA receptors with NMDA resulted in the reduction of PAIP2A to 71.8% ± 2.2% of prestimulation levels (Figure 4B). We reasoned that the rapid downregulation of PAIP2A is mediated by proteolytic activity.