Journal: Neuron

Article Title: Postsynaptically silent synapses in single neuron cultures.

doi: 10.1016/s0896-6273(00)80662-5

Figure Lengend Snippet: Figure 1. In the Autapse, Detectable NMDAR-Only Events Do Not Figure 2. The NMDAR EPSC Has a Greater Mean2/Variance than Derive from Spillover the AMPAR EPSC With multiple postsynaptic cells, NMDAR-only events may derive (A) Representative average traces of pharmacologically isolated from direct activation of synapses that have only functional NMDARs NMDAR and AMPAR EPSCs superimposed on the action potential (A2) or from spillover of glutamate from one synapse to another, artifact (A1) and after subtraction of the action potential artifact (A2). resulting in the activation of NMDARs but not AMPARs due to the Traces are the average of 15 responses in each pharmacological higher affinity of NMDARs for glutamate (B1). Traces 1 and 2 depict condition: NBQX (5 mM) to isolate the NMDAR EPSC, D-APV (100 the response in the recorded cell to glutamate release at synapses mM) to isolate the AMPAR EPSC, and both to isolate the action 1 and 2, respectively. In the autapse, NMDAR-only events due to potential artifact. synapses expressing functional NMDARs but not AMPARs should (B) Typical experiment to measure mean2/variance of NMDAR and be visible as discrete independent events (C2), but NMDAR-only AMPAR EPSCs. A minimum of 30 trials was collected in each phar- events due to spillover would be masked by the concomitant activa- macological condition. tion of synaptic AMPARs and NMDARs at the releasing synapse (D). (C) Summary graph of the AMPAR and NMDAR mean2/variance for 13 cells. The NMDAR mean2/variance is greater than the AMPAR mean2/variance for the majority of the cells and is greater on average (p , 0.0l). The average mean2/variance ratio of NMDAR EPSCs to were done in the absence of external Mg21 so that the AMPAR EPSCs is 1.6 6 0.2. A line of slope l (dotted line) illustrates NMDAR EPSCs could be recorded at the resting mem- the predicted values for equal mean2/variance of AMPAR and brane potential. NMDAR EPSCs. We first compared the variation, calculated as the mean2/variance, of the evoked AMPAR EPSC to that of the NMDAR EPSC. If there are synapses that have only rising AMPAR component of the response. An average of these AMPAR-containing mEPSCs reveals a slow tail, NMDARs in addition to synapses that have both AM- PARs and NMDARs, then the variation of the NMDAR which is due to the activation of NMDARs presumably colocalized at the same synapse (Figure 3A). When the EPSC should be less (mean2/variance will be greater) than that of the AMPAR EPSC (Kullmann, 1994). Evoked relative contribution of the NMDAR component and AMPAR component is compared between the mEPSC autaptic responses from a typical cell are shown in Fig- ure 2A in which the selective AMPAR antagonist NBQX and the evoked dual component response in the same cell, it is clear that the slow NMDAR EPSC contributes and the selective NMDAR antagonist APV are applied sequentially. The stimulus artifact recorded in the pres- considerably less to the mEPSC than it does to the evoked response (Figure 3A). A summary of a number ence of both antagonists has been subtracted from the records shown in Figure 2A2. The entire experiment is of cells with such a within cell comparison is graphed in Figure 3B. This discrepancy was also seen when am- plotted in Figure 2B, and it is clear that the variation of the size of the AMPAR EPSCs is greater than that of plitudes of the two components were compared rather than charge transfer. The NMDAR/AMPAR amplitude the NMDAR EPSCs. The relative fluctuation of the two components is plotted for a number of cells in Figure ratio was 0.35 for the mEPSCs but was 0.65 for evoked EPSCs, corresponding to a proportion of NMDAR-only 2C, and for the majority of cells (10 of 13), the points fall above the line of unity, indicating that the mean2/ events of approximately 44% 6 5% to make up the difference. Interestingly, there was a direct correlation variance is greater for the NMDAR EPSC (p , 0.01). The average mean2/variance ratio of NMDAR EPSCs to between the size of the AMPAR and the NMDAR compo- nents of the AMPAR-containing mEPSCs. Thus, mEPSCs AMPAR EPSCs is 1.6 6 0.2 (n 5 13), which corresponds to 38% of synapses containing only NMDARs. These having large AMPAR EPSCs on average have large NMDAR EPSCs (Figure 3C). This was true for all cells results are consistent with the observations made in hip- poccampal slice preparations (Kullmann, 1994; Selig et examined (n 5 5). This finding makes it unlikely that the failure to identify and include some mEPSCs with small al., 1995; but see Niu et al., 1998) and suggest that there is a substantial population of synapses that have AMPAR EPSCs could explain the difference between the mEPSC and the evoked EPSC, since such events NMDARs but no detectable AMPARs. Further evidence suggesting the existence of NMDAR- are likely to have correspondingly small NMDAR compo- nents. Thus, the larger contribution of NMDARs to the only EPSCs comes from the comparison of AMPAR- containing mEPSCs to evoked EPSCs in the same cell. evoked response when compared to dual component mEPSCs suggests the existence of a population of For this analysis mEPSCs were identified by the fast

Article Snippet: These sources of noise triple stained either as described previously or using goat anti-NR1 were removed from the calculated mean2/variance (see Kullmann, (Santa Cruz Biotechnology, 1:2000), rabbit anti-GluR1 (Upstate Bio1994) but contributed ,5% to the total variation.

Techniques: Isolation, Activation Assay, Functional Assay, Expressing, Comparison