In conclusion, dye-filling experiments in combination with post-hoc immunohistochemistry provide independent evidence for a synapse elimination deficit at the calyx of Held synapse of Robo3 cKO mice. To investigate the presynaptic defects underlying the impaired synaptic transmission, we performed simultaneous pre- and postsynaptic recordings (Figure 5). The presynaptic Ca2+ currents in response to a 50 ms depolarization to 0 mV were significantly smaller in Robo3 cKO mice (0.82 ± 0.10 nA; n = 26) as compared to control mice (1.40 ± 0.10 nA, n = HSP inhibitor 14; p < 0.001) (Figures 5A and 5B). The basal presynaptic membrane capacitance

(Cm), a proxy of the membrane surface of the calyx, was smaller in Robo3 cKO mice (15.4 ± 1.4 pF) than in control (22.4 ±

1.4 pF; p < 0.001; Figure 5B). This agrees well with the smaller calyx surface found in the three-dimensionally rendered calyces (Figure 4C). The Ca2+ current density, calculated by normalizing the maximal Ca2+ current by the Cm value http://www.selleckchem.com/products/MDV3100.html of each recording, was unchanged on average (p = 0.35), but was more variable in Robo3 cKO mice (Figure 5B). The EPSCs in response to pool-depleting presynaptic depolarizations were smaller and had slower rise times in Robo3 cKO mice (Figure 5A), indicating smaller pool sizes and less synchronized transmitter release. Deconvolution analysis of EPSCs indeed showed a strong reduction of the fast release component in Robo3 cKO mice. In the example of a Robo3 cKO recording in Figure 5A3, release was very slow and the cumulative release trace could be fitted with a single exponential with a time constant of 26 ms. Overall, n = 8 out of 20 synapses recorded in Robo3 cKO mice showed similarly slow release, with time constants of 10 ms or more. Over the entire population of synapses, the release time constant was significantly slower in Robo3cKO as compared to control mice (Figure 5C). Furthermore, the number of vesicles released in the fast component was significantly lower in Robo3 cKO mice (772 ± 98; n = 12 cells) as compared to control calyces (1,602 ± 196; n = 10; p < 0.001) (Figure 5C). Thus, the vesicle release

kinetics were slowed, and there were fewer vesicles in the fast-releasable subpool, FRP (Sakaba and Neher, 2001). Previous Phosphoprotein phosphatase work has shown that phasic transmitter release in response to presynaptic APs is mainly contributed by FRP vesicles (Sakaba, 2006). Therefore, we would expect that a lower number of FRP vesicles in Robo3 cKO mice should translate into a similar decrease in the number of fast-releasable vesicles available for AP-evoked release. To test this prediction, and to investigate possible changes in release probability, we used 100 Hz trains of brief AP-like presynaptic depolarizations, and back-extrapolation of cumulative EPSC amplitudes as a pool size estimate (Schneggenburger et al., 1999; Figures 5D and 5E).