Hippocampal N-methyl-D-aspartate receptor (NMDAR) is required for spatial working memory. Although evidence from genetic manipulation mice suggests an important role of hippocampal NMDAR NR2B subunits (NR2B-NMDARs) in spatial working memory, it remains unclear whether or not the requirement of hippocampal NR2B-NMDARs for spatial working memory depends on the time of spatial information maintained. Here, we investigate the contribution of hippocampal NR2B-NMDARs to spatial working memory on delayed alternation task in T-maze (DAT task) and delayed matched-to-place task in water maze (DMP task). Our data show that infusions of the NR2B-NMDAR selective antagonists, Ro25-6981 or ifenprodil, directly into the CA1 region, impair spatial working memory in DAT task with 30-s delay (not 5-s delay), but severely impair error-correction capability in both 5-s and 30-s delay task. Furthermore, intra-CA1 inhibition of NR2B-NMDARs impairs spatial working memory in DMP task with 10-min delay (not 30-s delay). Our results suggest that hippocampal NR2B-NMDARs are required for spatial working memory in long-delay task, whereas spare for spatial working memory in short-delay task. We conclude that the requirement of NR2B-NMDARs for spatial working memory is delay-dependent in the CA1 region.

Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels are widely expressed in neurons in the central nervous system. It has been documented that HCN channels regulate the intrinsic excitability of pyramidal cells in the medial prefrontal cortex (mPFC) of rodents. Here, we report that HCN channels limited GABAergic transmission onto pyramidal cells in rat mPFC. The pharmacological blockade of HCN channels resulted in a significant increase in the frequency of both spontaneous and miniature inhibitory postsynaptic currents (IPSCs) in mPFC pyramidal cells, whereas potentiation of HCN channels reversely decreases the frequency of mIPSCs. Furthermore, such facilitation effect on mIPSC frequency required presynaptic Ca2+ influx. Immunofluorescence staining showed that HCN channels expressed in presynaptic GABAergic terminals, as well as in both soma and neurite of parvalbumin-expressing (PV-expressing) basket cells in mPFC. The present results indicate that HCN channels in GABAergic interneurons, most likely PV-expressing basket cells, constrain inhibitory control over layer 5–6 pyramidal cells by restricting presynaptic Ca2+ entry.

The number and subtype composition of N-methyl-d-aspartate receptor (NMDAR) at synapses determines their functional properties and role in learning and memory. Genetically increased or decreased amount of GluN2B affects hippocampus-dependent memory in the adult brain. But in some experimental conditions (e.g., memory elicited by a single conditioning trial (1 CS-US)), GluN2B is not a necessary factor, which indicates that the precise role of GluN2B in memory formation requires further exploration. Here, we examined the role of GluN2B in the consolidation of fear memory using two training paradigms. We found that GluN2B was only required for the consolidation of memory elicited by five conditioning trials (5 CS-US), not by 1 CS-US. Strikingly, the expression of membrane GluN2B in CA1was training-strength-dependently increased after conditioning and that the amount of membrane GluN2B determined its involvement in memory consolidation. Additionally, we demonstrated the increases in the activities of cAMP, ERK and CREB in the CA1 after conditioning, as well as the enhanced intrinsic excitability and synaptic efficacy in CA1 neurons. Up-regulation of membrane GluN2B contributed to these enhancements. These studies uncover a novel mechanism for the involvement of GluN2B in memory consolidation by its accumulation at the cell surface in response to behavioral training.