A key issue, therefore, is whether the NMDAR content is altered at individual synapses. We first addressed this functionally, by collecting mixed spontaneous AMPAR- and NMDAR-mediated High Content Screening currents at −70 mV in the absence of external Mg2+, then washing on APV and collecting the pure AMPAR-mediated currents. The pure AMPAR currents were then subtracted from the mixed currents to give a pure NMDAR-mediated spontaneous current. We performed these experiments using simultaneously recorded NLGN1 miR-expressing neurons and neighboring control cells in the dentate gyrus and collected both evoked and spontaneous currents, using the evoked currents to assess the validity

of the technique. The stimulation-evoked, subtracted NMDAR-mediated currents in NLGN1 miR expressing cells were reduced, as expected, compared to control cells (Figures 2A and 2B).

Moreover, the magnitude of the reduction was identical to that found when NMDAR currents were measured at +40 mV in the previous experiment (as percent selleck chemicals llc of control, +40 mV, 32.12 ± 5.26; subtracted 23.4 ± 4.92; p > 0.05), thus providing validation of the technique. Furthermore, neither the charge transfer of the NMDAR current as a percent of the total charge transfer of the mixed AMPAR/NMDAR current nor the kinetics of the NMDAR current were altered in the evoked Amisulpride response (Figures 2C and 2D). We next analyzed the spontaneous currents in these same cells (Figure 2E) and found a dramatic reduction in the frequency of spontaneous events (Figure 2F), but no

change in amplitude of either the mixed current, the pure AMPAR current, or the pure, subtracted NMDAR current (Figure 2G). Like the evoked current, knockdown did not affect the percentage of spontaneous charge transfer accounted for by NMDA current (Figure 2H). We consequently conclude that the reduction in evoked NMDAR currents is functionally due to an all-or-none loss of synapses, while the remaining synapses have normal numbers of NMDARs. To complement the functional evidence for an all-or-none loss of synapses following neuroligin knockdown, we examined spine density. Following knockdown of NLGN1, we filled transduced dentate granule cells and neighboring control cells with fluorescent dye and imaged their dendrites (Figure 2I). We observed a reduction in spine density in NLGN1 miR expressing cells as compared to control (Figure 2J) of a similar magnitude to the reduction in evoked currents. Spine density in dentate granule cells following the knockdown of NLGN3 was also reduced, confirming that synaptic loss is a general response to neuroligin knockdown (Figures S2A and S2B). Finally, we performed a coefficient of variation analysis on the paired evoked recordings following neuroligin knockdown.