We then tested whether Munc13 binding by the RIM Zn2+ finger domain is required for RIM-dependent vesicle priming by expressing rescue proteins in RIM-deficient neurons. Wild-type RIM1α and RIM1β reversed the

decrease in spontaneous minirelease in RIM-deficient neurons; in fact, RIM1α appeared to even enhance spontaneous release (Figure 3D). The Zn2+ finger domain mutation in RIM1α and RIM1β, however, impaired rescue. Moreover, RIM1α and RIM1β both rescued the impairment in sucrose-induced release in Raf pathway RIM-deficient neurons; again, the Zn2+ finger mutation partly blocked this rescue in RIM1α and completely in RIM1β (Figure 3E and Figure S3C). Overall, these experiments indicate that in RIM proteins, the Zn2+ finger domain is the major effector domain for priming; moreover, the experiments show that RIM1α may mediate rescue more efficiently than RIM1β, consistent with the notion that the N-terminal Rab3-binding activity of RIM1α (which is absent from RIM1β; Kaeser et al., 2008) contributes to release.

We next asked whether the RIM Zn2+ finger requires the context of other C-terminal domains of RIM to promote priming, as would be expected for a scaffolding protein, or whether it acts autonomously. We examined rescue with RIM1α fragments composed of either only its N-terminal Rab3- and Munc13-binding sequences (referred to as the RIM-RZ fragment), or of its C-terminal fragment containing the PDZ, C2A, and C2B domains and the RIM-BP-binding sequence (referred to as the RIM-PASB fragment; Figure 4A). Surprisingly, the N-terminal

RIM-RZ fragment was sufficient to GDC-0068 nmr rescue vesicle priming in RIM-deficient neurons, whereas the C-terminal PASB-fragment had no rescue effect (Figures 4B–4D; note that the RIM-PASB fragment efficiently rescues the Ca2+ influx impairment in RIM-deficient neurons [Kaeser et al., 2011]). Importantly, the N-terminal RIM-RZ fragment did not significantly alter vesicle priming when overexpressed in wild-type Electron transport chain neurons (Figure S4). Unlike release induced by hypertonic sucrose, both the N-terminal and the C-terminal RIM1α fragment increased release stimulated by a 10 Hz train of action potentials (Figure 4E). This result is consistent with completely separated roles of the N-terminal RIM domains in vesicle priming and of the C-terminal RIM domains in boosting local Ca2+ influx (Kaeser et al., 2011). The rescue of priming in RIM-deficient neurons by the RIM-RZ fragment alone is surprising because it suggests that RIM does not act as a classical scaffolding protein that functions by recruiting multiple other proteins via its N- and C-terminal domains to the same subcellular location. However, the RIM-RZ fragment still binds to two proteins in a trimeric complex—Rab3 and Munc13 (Dulubova et al., 2005). Thus, its rescue activity could either be mediated by coupling Rab3 on synaptic vesicles to Munc13 in the active zone or it could be because of autonomous functions of each of its binding activities.