Slice orientation and muscarinic acetylcholine receptor activation determine the involvement of N-methyl D-aspartate receptor subunit GluN2B in hippocampal area CA1 long-term depression

Abstract

Background: The contribution of different GluN2 subunits of the N-methyl D-aspartate (NMDA) receptor to the induction of bidirectional hippocampal synaptic plasticity is a controversial topic. As both supporting and refuting evidence for the hypothesis of subunit specialization in opposing directions of plasticity has accumulated since it was first proposed a few years ago, we hypothesize that differences in experimental conditions may have in part contributed to some of the inconsistent results from these studies. Here we investigate the controversial hypothesis that long-term depression (LTD) is preferentially induced by GluN2B-containing NMDA receptors in area CA1 of hippocampal slices.

Results: We find that brain slices from 2-3 week old rats prepared in the sagittal orientation have GluN2B-independent LTD whereas slices prepared in the coronal orientation have GluN2B-dependent LTD. There was no difference between the orientations in the fraction of the NMDAR EPSC sensitive to a GluN2B-selective antagonist, leading us to believe that the intracellular signaling properties of the NMDARs were different in the two preparations. Coronal slices had greater association of LTD-related intracellular signaling protein RasGRF1 with GluN2B relative to sagittal slices. Antagonism of muscarinic acetylcholine receptors (mAChRs) in the sagittal slices returned LTD to a GluN2B-dependent form and increased the association of GluN2B with RasGRF1.

Conclusions: These results suggest a novel form of NMDAR modulation by mAChRs and clarify some disagreement in the literature.

Figure 1

LTD in sagittal and coronal slices has different GluN2B-dependence. Hippocampal slices (400 μM) prepared in either sagittal (A) or coronal (B) orientations, both with CA3 attached, were used for extracellular recordings of field excitatory postsynaptic potentials (fEPSPs) in area CA1. After a stable baseline recording period (at least 30 minutes) the GluN2B-selective antagonist Ro 25-6981 (5 μM, Ro) was applied and stimulation at 0.03 Hz continued for another 30 minutes until a 1 Hz, 15 minute low frequency stimulation (LFS) at baseline intensity, also in the presence of Ro. Drug application was stopped at the end of LFS and the magnitude of LTD was quantified in the last ten minutes of the hour after LFS. In the slices of either orientation that were not treated with Ro, LFS induced LTD. Likewise, in the sagittal slices, Ro did not prevent the induction of LTD (A) although D-AP5 (50 μM) did (data not shown). However, in the coronal slices treated with Ro, LFS did not induce LTD (B). Scale bars represent 10 ms and 0.5 mV.

Figure 2

The GluN2B-mediated fraction of the NMDAR EPSC is not significantly different between coronal and sagittal orientations. A) Whole-cell patch clamp recordings from CA1 pyramidal cells in either sagittal or coronal slices were made and the NMDAR EPSC was isolated at -40 mV in the presence of bicuculine (10 μM) and NBQX (5 μM). After a 10 minute stable baseline recording period at stimulation frequency 0.03 Hz, the GluN2B-selective antagonist Ro was applied for 40 minutes while stimulation continued. Scale bars represent 200 ms and 200 pA. B) The level of depression in the NMDAR EPSC induced by Ro, when quantified in the 40-50 minute epoch, was similar in sagittal and coronal slices. This indicates activation of a similar number of synaptic GluN2B-containing NMDARs during baseline synaptic stimulation in either slice orientation.

Figure 3

GluN2B-dependence of LTD is created in the sagittal slice by a muscarinic acetylcholine receptor antagonist. A) Sagittal slices were again used for extracellular recordings of synaptic activity. After a stable baseline stimulation period (0.03 Hz, >30 minutes), the muscarinic receptor antagonist scopolamine (Scop, 20 μM) was added to the aCSF with or without GluN2B-selective antagonist Ro (5 μM) and stimulation continued at baseline frequency for another 30 minutes. LFS was then begun while drug application continued. Drug washout began at the termination of LFS and the magnitude of LTD was quantified in the final 10 minutes of the hour after LFS. B) In the sagittal slices, application of scopolamine did not change the induction of LTD or the baseline fEPSP. However, when co-applied with GluN2B antagonist Ro there was a significant inhibition of LTD compared to the slices treated with scopolamine alone. Scale bars represent 10 ms and 0.5 mV.

Figure 4

Greater association of GluN2B with RasGRF1 occurs under conditions that show GluN2B involvement in LTD. A,B) Sagittal hippocampal slices were prepared with the same methodology as for the electrophysiological experiments and treated with control aCSF (C) or 20 μM scopolamine in aCSF (S) for 2 hrs, followed by homogenization in RIPA buffer. 500 μg of each lysate was used for immunoprecipitation with a polyclonal GluN2B antibody bound to protein A sepharose. After washing in RIPA, the coIPs were subjected to SDS PAGE and Western blot, alongside 30 μg per lane of the crude lysates. The blots were sequentially probed with RasGRF1 and GluN2B antibodies using HRP- and AP-conjugated secondary antibodies respectively. In each coimmunoprecipitate the level of RasGRF1 was normalized to GluN2B and then a scopolamine/control ratio was calculated and averaged across experiments. C, D) Coronal (C) or sagittal (S) slices were prepared and recovered in the absence of scopolamine. The lysates were processed and taken forward into immunoprecipitation as described above.