At P10 and thereafter, muscimol decreased the firing rate in most PCs (Figures 1E, 1G, and 1H). The action of muscimol was abolished by bicuculline (10 μM) (Figure 1E), confirming that the changes in PC firing rates resulted from GABAA receptor activation. The analyses so far indicate that overall, GABAergic transmission is attenuated in GAD67+/GFP mice during the second postnatal week and

that GABA inhibits PC activity after P10 in both control and GAD67+/GFP mice. Since CF synapse elimination is known to proceed in this period (Hashimoto et al., 2009b and Kano and Hashimoto, 2009), we examined CF innervation of PCs in GAD67+/GFP mice to clarify whether and how GABAergic transmission LY2157299 concentration contributes to CF synapse elimination. We first checked gross anatomy of the cerebellum, morphology of PCs Panobinostat and PF-PC synaptogenesis (Figure S2). We found that foliation and layer structure of the

cerebellum (Figures S2A and S2B), morphology of PC (Figures S2C–S2F), and morphology and density of PF-PC synapse (Figures S2G–S2J) were normal in GAD67+/GFP mice. We then examined basic electrophysiological properties of PF-mediated EPSC (PF-EPSC). The 10%–90% rise time (control: 1.08 ± 0.04 ms, n = 42; GAD67+/GFP: 0.99 ± 0.04 ms; n = 57, p = 0.142), the decay time constant (control: 6.01 ± 0.38 ms; GAD67+/GFP: 5.91 ± 0.28 ms; p = 0.985), and stimulus-response curves of PF-EPSCs at P10-P13 (Figure S2K) were similar between the two mouse strains. These results indicate that PF-PC synapses possess normal morphology and function in the GAD67+/GFP cerebellum. Normal morphology of the cerebellum has also been shown in the cerebellum-specific GAD67 knockout mice next (Obata et al., 2008). Thus, reduction of GAD67 activity does not cause severe changes in the morphology and synaptic wiring of the cerebellum. We then analyzed CF innervation of PCs in mature GAD67+/GFP mice at P21–P52. In slices prepared from control or GAD67+/GFP mice, CFs were stimulated in the granular layer and evoked responses in single PCs were recorded.

In most PCs (93/114, 81.6%), large CF-EPSCs were elicited in an all-or-none fashion as the stimulus intensity was gradually increased (Figure 2A, upper left). By contrast, in 50.4% (61/121) of GAD67+/GFP PCs, CF-EPSCs had two or three discrete steps (Figure 2A, lower left). The frequency distribution histogram of PCs in terms of the number of CF-EPSC steps (Figure 2A) shows significant difference between control and GAD67+/GFP mice (p < 0.001). In both control and GAD67+/GFP mice, each PC was either monoinnervated by a strong CF (termed “CF-mono”) or multiply-innervated by a strong CF (termed “CF-multi-S”) plus one or two weaker CFs (termed “CF-multi-W”) (Hashimoto et al., 2009a and Hashimoto and Kano, 2003).