To test this predication, we infected hippocampal cultures with a lentivirus expressing an shRNA against LPHN3 (shLphn3; Figure S3D) such that the preponderance (>90%) of neurons

in the culture was transduced and recorded mEPSCs from neurons in LPHN3 knockdown and control cultures (Figure 3J). Culture-wide knockdown of LPHN3 reduced the frequency of mEPSCs (Figure 3K) without affecting mEPSC amplitude (Figure 3L), suggesting that loss of LPHN3 decreases the number of excitatory synapses in these cultures. These experiments in dissociated find more hippocampal cultures show that three distinct manipulations designed to perturb LPHN3-FLRT3 complexes—competition with ecto-LPHN3-Fc (Figures 3A–3C), shRNA knockdown of FLRT3 (Figures 3D–3I), and shRNA knockdown of LPHN3 (Figures 3J–3L)—all lead to a reduction in the number of glutamatergic synapses, strongly suggesting that FLRT3 and LPHN3 serve to positively regulate synapse number. To test whether endogenous FLRT3 contributes to synapse development

in vivo, we first used in utero electroporation to sparsely label and manipulate DG GCs for anatomical analysis. In fixed slices from P14 electroporated mice, GFP-filled http://www.selleckchem.com/products/LBH-589.html GC dendrites in the middle molecular layer were imaged on a confocal microscope, and dendritic spines were counted (Figure 4A). Knockdown of FLRT3 resulted in a highly significant reduction in dendritic protrusion density relative to controls (Figure 4B). The density of spines on the apical dendrites of electroporated CA1 pyramidal neurons, which do not express Flrt3 Bay 11-7085 ( Figure 2A), did not differ

between shFlrt3 and control cells ( Figures 4C and 4D), functionally confirming the specificity of the shRNA used. To determine whether the reduction in GC dendritic spine density reflects a decrease in the strength of synaptic input onto these cells, we stereotaxically injected shFlrt3 or control lentivirus into the hippocampus of P5 rat pups and cut acute slices for electrophysiology between P13 and P16. Infected GCs were identified by GFP epifluorescence, and simultaneous whole-cell voltage-clamp recordings were made from nearby infected and uninfected cells while perforant path synaptic inputs were evoked from the middle molecular layer. We observed that AMPAR-mediated EPSCs onto shFlrt3-infected neurons were strongly reduced in amplitude relative to simultaneously recorded uninfected control cells (Figures 4E and 4F). NMDAR-mediated EPSCs, measured 50 ms after the stimulus at a holding potential of +40 mV, were similarly reduced by FLRT3 knockdown (Figures 4G and 4H). This reduction was proportional, because the ratio of AMPAR EPSC to NMDAR EPSC for each input was not affected by FLRT3 knockdown (Figure 4I), consistent with a reduction in number of synapses rather than a selective loss of certain glutamate receptors.