RIM proteins activate vesicle priming by reversing autoinhibitory homodimerization of Munc13

Abstract

At a synapse, the presynaptic active zone mediates synaptic vesicle exocytosis. RIM proteins are active zone scaffolding molecules that--among others--mediate vesicle priming and directly or indirectly interact with most other essential presynaptic proteins. In particular, the Zn²+ finger domain of RIMs binds to the C₂A domain of the priming factor Munc13, which forms a homodimer in the absence of RIM but a heterodimer with it. Here, we show that RIMs mediate vesicle priming not by coupling Munc13 to other active zone proteins as thought but by directly activating Munc13. Specifically, we found that the isolated Zn²+ finger domain of RIMs autonomously promoted vesicle priming by binding to Munc13, thereby relieving Munc13 homodimerization. Strikingly, constitutively monomeric mutants of Munc13 rescued priming in RIM-deficient synapses, whereas wild-type Munc13 did not. Both mutant and wild-type Munc13, however, rescued priming in Munc13-deficient synapses. Thus, homodimerization of Munc13 inhibits its priming function, and RIMs activate priming by disrupting Munc13 homodimerization.

Figure 1. Conditional deletion of RIM1α, 1β, 2α, 2β, and 2γ dramatically decreases the RRP size

All experiments in this and the following figures employ hippocampal neurons cultured from conditional RIM1/2 DKO mice that were infected with lentiviruses expressing inactive (Control) or active cre-recombinase (cDKO) A & B, Sample traces (left) and summary graphs (right) of the frequency and amplitude of spontaneous miniature excitatory (mEPSCs) and inhibitory postsynaptic currents (mIPSCs), monitored in control (gray) and RIM-deficient cDKO hippocampal neurons (black). C, Sample traces (left) and summary graphs (right, as charge transfer) of inhibitory postsynaptic currents induced by hypertonic sucrose (0.5 M) in control (gray) and RIM-deficient cDKO (black) neurons. Hypertonic sucrose was puffed onto the patched neuron for 30 s in the presence of 1 μM TTX, 10 μM CNQX and 50 μM APV. Charge transfer during the initial (0-10 s) and the steady-state responses (15-30 s of application) were quantified to estimate the RRP size and the RRP recovery rate, respectively. D, Cumulative charge transfer (left) and kinetic analyses of the cumulative charge transfer (right) of sucrose-induced postsynaptic currents in control (gray) and RIM-deficient cDKO neurons (black). The integrated charge transfer was fitted by a double exponential function to determine kinetic parameters for the fast and the slow component (τ_fast, A_fast and τ_slow, A_slow, respectively). E, Sample immunoblots (left) and protein quantitations (right) of control and cDKO neurons. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by Student’s t-test (*, p<0.05; ***, p<0.001). For short-term synaptic plasticity, see Fig. S1.

Figure 2. Differential effects of the RIM deletion on RRP capacity and refilling rates

A, Representative IPSC traces (left) and summary graphs (right, showing absolute and normalized data) of the RRP recovery as a function of the inter-puff interval between two sucrose applications, applied at increasing inter-stimulus intervals (10, 20, 60 and 120 s). RRP recovery was calculated by dividing the second sucrose-induced charge transfer by the first, and normalizing the resulting ratio to that observed at the 120 s interval. B, Sample traces (left) and summary data (right, showing absolute and normalized data) of the recovery rate of single evoked IPSCs after sucrose-induced RRP depletion. IPSC amplitudes were measured at multiple time points after RRP depletion (10, 30, 60, 120 s), and normalized to control IPSC amplitudes before RRP depletion. C, Sample traces (top) and summary graph of the IPSC amplitudes (bottom) measured in cDKO and control neurons during depletion of the RRP by a 50 Hz stimulus train applied for 1 s, and during the recovery from the RRP depletion after the train. The stimulation protocol is outlined on top; neurons were stimulated with 10 pulses at 0.2 Hz before the high frequency train. For the responses in RIM-deficient cDKO neurons, both absolute and normalized data are shown. For expanded views of IPSCs during the high-frequency train and an analysis of the delayed release kinetics, see Fig. S2. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 2-way ANOVA (**, p<0.01; ***, p<0.001). P-values and n’s are listed in Supplementary Table S2.

Figure 3. RIM N-terminal domains mediate synaptic vesicle priming

A, Diagram of full-length wild-type and mutant RIM1α and RIM1β rescue proteins expressed in RIM-deficient cDKO neurons via an IRES sequence from the same mRNA as cre-recombinase (asterisk = K144/6E substitution). B, Localization of wild-type or K144/6E-mutant RIM1α and ubMunc13-2 in transfected HEK293 cells. RIMs and ubMunc13-2 were transfected either alone (top) or in combination, and their colocalization was studied after translocation of Munc13 by phorbol ester (PMA) to the plasma membrane. Note that without Munc13 binding, RIM1α is sequestered into the nucleus in non-neuronal cells, but recruited to the plasma membrane in the presence of ubMunc13-2 and PMA (N, nucleus, scale bar 5 μm). C, Crosslinking of wild-type or K144/6E-mutant RIM1α with Munc13s. HEK293 cells expressing the indicated proteins were treated in control solution or solution containing 0.008% glutaraldehyde, and proteins were analyzed by SDS-PAGE and immunoblotting with antibodies against the HA-tag of RIM1α, or against the central domains of RIM (R809). ubMunc13-2 monomer was detected with antibodies against the mVenus tag (*, unspecific band). Note that we included as a control the K32E-mutant of ubMunc13-2 that does not homodimerize but retains RIM-binding (see Lu et al., 2006). The high-molecular weight bands are not clearly resolved on the blot, probably because of their large size and variable degrees of crosslinking. D, Sample traces (left) and summary graphs of the frequency (right) of spontaneous mIPSCs monitored in control and RIM-deficient cDKO neurons with or without RIM rescue. E, Sample traces (left) and summary graphs of the RRP size as measured by the initial sucrose response (right, top) and of the steady-state refilling size (right, bottom) in control neurons, and cDKO neurons with or without rescue. Rescue efficacies are shown in Fig. S3 and Table S3. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (*, p<0.05; **, p<0.01; ***, p<0.001).

Figure 4. RIM N-terminal domains are necessary and sufficient for priming

A, Diagram of full-length RIM1α and of RIM1α fragments expressed as rescue proteins in RIM-deficient cDKO neurons via an IRES sequence from the same mRNA as cre-recombinase. The single-letter code above the RIM1α diagram identifies the various domains (R, Rab3-binding α-helical region; Z, Munc13-binding Zn²⁺-finger region, P, PDZ-domain; A, C₂A-domain; S, proline-rich SH3-binding PxxP motif; B, C₂B-domain); H marks the presence of a human influenza hemagglutinin (HA)-tag. B-D, Sample traces (left) and summary graphs of initial and steady-state RRP sizes in control or RIM-deficient cDKO neurons without or with rescue with RIM1α (B), RIM-RZ (C), or RIM-PASB (D). E, Analysis of IPSCs evoked by a 10 Hz stimulus train in control neurons, RIM-deficient cDKO neurons without rescue, or RIM-deficient cDKO neurons rescued with the RIM-RZ or the RIM-PASB fragment. Note that different from sucrose-evoked release, action-potential induced release is partly rescued by both RIM fragments. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (**, p<0.01; ***, p<0.001). For an analysis of the effect of RIM-RZ expression in wild-type neurons, see Fig. S4; numerical values are shown in Table S4.

Figure 5. Munc13-binding by the N-terminal RIM Zn²⁺-finger domain autonomously activates vesicle priming

A, Diagram of N-terminal wild-type and mutant RIM fragments expressed as rescue proteins in cDKO neurons via an IRES sequence from the same mRNA as cre-recombinase (asterisk = K144/6E substitution). B, Sample traces (left) and summary graph (right) of the frequency of spontaneous mIPSCs monitored either in control neurons and cDKO neurons with or without rescue. C, Sample traces (left) and summary graphs (right) of RRP size in control neurons, and cDKO neurons with or without rescue. Rescue efficacies are indicated in Fig. S5 and Table S5. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (*, p<0.05; **, p<0.01; ***, p<0.001).

Figure 6. Constitutively monomeric mutant ubMunc13-2 bypasses the loss of priming in RIM-deficient neurons

A, Diagram of wild-type and mutant ubMunc13-2 rescue proteins expressed in cDKO neurons. Key domains of Munc13 (C₁- and C₂-domains and the MUN-domain; V = mVenus-tag) and their interactions are indicated (DAG, diacylglycerol). Cre-recombinase and Munc13s were expressed from separate lentiviruses by consecutive infection (cre and control viruses at DIV3, Munc13 viruses at DIV5). Note that the K32E-mutation of the Munc13 C₂A-domain (Munc13^K32E) renders Munc13 constitutively monomeric (Lu et al., 2006). B, Sample traces (left) and summary graph (right) of the frequency of spontaneous mIPSCs monitored in control and RIM-deficient cDKO neurons with or without wild-type or mutant Munc13 expression (top 4 traces and left panels), or in wild-type neurons with Munc13 expression (bottom two traces and right panels). C, Sample traces (left) and summary graphs (right) of RRP size in control and cDKO neurons with or without rescue, or in wild-type neurons with Munc13 expression. D, Confocal sections of cDKO neurons with or without rescue stained for all Munc13 isoforms (top) or for ubMunc13-2 (bottom), and counterstained with synapsin antibodies. Note that wild-type and mutant ubMunc13-2 rescue proteins exhibit similar levels and localizations. E, Expression of K32E-mutant but not wild-type ubMunc13-2 partially rescues the amplitude of action-potential evoked IPSCs in RIM-deficient cDKO neurons, but does not rescue their impaired Ca²⁺-responsiveness. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (*, p<0.05; **, p<0.01; ***, p<0.001). For additional analyses and numerical data, see Fig. S6 and Table S6.

Figure 7. The ubMunc13-2 C₂A-domain is dispensable for rescuing the RRP in RIM-deficient neurons

A, Diagram of ubMunc13-2 rescue proteins expressed in cDKO neurons via an IRES sequence from the same mRNA as cre-recombinase (V, mVenus-tag). Key interactions of various domains are indicated (DAG, diacylglycerol). B, Co-localization of full-length RIMs and Munc13s in HEK293 cells depends on the Munc13 C₂A-domain. RIM1α and wild-type or ΔC₂A-mutant ubMunc13-2 were transfected into HEK293 cells, and their co-localization was studied before (−PMA) or after phorbol ester-induced translocation (+PMA) of Munc13 to the plasma membrane. Note that without Munc13 binding (i.e., with ΔC₂A-mutant ubMunc13-2), RIM is largely localized to nuclei in non-neuronal cells, whereas in the presence of wild-type ubMunc13-2, it is either cytosolic (−PMA) or on the plasma membrane (+PMA)(N, nucleus; scale bar 5 μm). C, Crosslinking of wild-type or ΔC₂A-mutant ubMunc13-2 with RIM1α. HEK293 cells expressing the indicated proteins were treated in control solution or solution containing 0.008% glutaraldehyde, and proteins were analyzed by SDS-PAGE and immunoblotting with antibodies against the mVenus tag of Munc13 and the central domains of RIM (R809). D, Sample traces (left) and summary graph (right) of miniature mIPSCs monitored in control neurons and cDKO neurons with or without ubMunc13^ΔC2A expression. E, Sample traces (left) and summary graphs (right) of RRP size in control neurons, and cDKO neurons with or without ubMunc13^ΔC2A rescue. Initial (left) and steady-state RRP sizes (right) are indicated. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (***, p<0.001). For additional data, see Fig. S7.

Figure 8. Wild-type and constitutively monomeric mutant ubMunc13-2 rescue priming in Munc13-deficient neurons

A, Immunoblotting for Munc13-1 and control proteins in neurons from Munc13-2 KO mice infected with control lentiviruses or with the Munc13-1 KD virus. Neurons were consecutively infected with the Munc13 rescue or control lentiviruses at DIV3, and with the Munc13-1 KD or control lentiviruses at DIV5. B, Quantitative real-time rt-PCR measurements of the mRNA levels of Munc13-1 in control and Munc13-1 KD neurons. C and D, Sample traces (left) and summary graph (right) of the spontaneous mIPSC frequency monitored in control neurons and M13 KD neurons either with or without ubMunc13 or ubMunc13^K32E rescue (C) or ubMunc13^ΔC2A rescue (D). Data of mIPSC amplitudes are shown in Fig. S7. E, Sample traces (left) and summary graphs (right) of RRP size in control neurons and Munc13 KD neurons with or without ubMunc13 rescue. Initial (left) and steady-state RRP sizes (right) are indicated. F, Sample traces (left) and summary graphs (right) of RRP size in control neurons, and Munc13 KD neurons with or without ubMunc13^ΔC2A rescue. Initial (left) and steady-state RRP sizes (right) are indicated. G, Model of the RIM priming switch. We propose that RIM activates priming by disrupting auto-inhibitory Munc13 C₂A-domain homodimers. The inactive Munc13 homodimer is physiologically activated by the RIM Zn²⁺-finger domain which converts the homodimer into a RIM/Munc13 heterodimer. This activation can be bypassed with constitutively monomeric mutant Munc13^K32E or Munc13^ΔC2A, but not with wild-type Munc13, in RIM-deficient neurons. Data shown are means ± SEMs (numbers in bars indicated numbers of cells/cultures analyzed). Statistical significance was assessed by 1-way ANOVA (*, p<0.05; **, p<0.01; ***, p<0.001). For additional data, see Fig. S8.