RIM Tether Ca2+ Channels to Presynaptic Active Zones via a Direct PDZ-Domain Interaction

Pascal S. Kaeser,1,3,6 Lunbin Deng,1,3,6 Yun Wang,3 Irina Dulubova,4,7 Xinran Liu,3 Josep Rizo,4 and Thomas C. Su¨dhof1,2,3,5,* 1Department of Molecular and Cellular Physiology 2Howard Hughes Medical Institute Stanford University, Lorry Lokey Building, 265 Campus Dr., Stanford 94305-5453, USA 3Department of Neuroscience 4Department of Biochemistry 5Howard Hughes Medical Institute UT Southwestern Medical Center, Dallas, TX 75390, USA 6These authors contributed equally to this work 7Present address: Reata Pharmaceuticals, 2801 Gateway Dr., Irving, TX 75063, USA *Correspondence: [email protected] DOI 10.1016/j.cell.2010.12.029

SUMMARY membrane that docks and primes vesicles for exocytosis (Wojcik and Brose, 2007). Fast synchronous release requires At a synapse, fast synchronous colocalization of Ca2+ channels with the release machinery at release requires localization of Ca2+ channels to the active zone (Llinas et al., 1992; Meinrenken et al., 2002). presynaptic active zones. How Ca2+ channels are re- Voltage-gated Ca2+ channels consist of a pore-forming a1 cruited to active zones, however, remains unknown. subunit and accessory b and a2d subunits (Catterall et al., 2005). Presynaptic neurotransmitter release almost exclusively Using unbiased two-hybrid screens, we here 2+ identify a direct interaction of the central PDZ domain depends on N- and P/Q-type Ca channels containing CaV2.1 and Ca 2.2 a1 subunits, respectively; R-type Ca2+ channels con- of the active-zone RIM with the C termini V 2+ taining CaV2.3 a1 subunits may also contribute, whereas L-type of presynaptic N- and P/Q-type Ca channels but 2+ 2+ and T-type Ca channels containing CaV1 and CaV3 a1 subunits not L-type Ca channels. To test the physiological do not (Castillo et al., 1994; Dietrich et al., 2003; Luebke et al., significance of this interaction, we generated 1993; Poncer et al., 1997; Regehr and Mintz, 1994; Takahashi conditional knockout mice lacking all multidomain and Momiyama, 1993; Wu et al., 1999). However, how N- and RIM isoforms. Deletion of RIM proteins ablated P/Q-type Ca2+ channels are specifically localized to active zones most neurotransmitter release by simultaneously and coupled to the release machinery is unknown. impairing the priming of synaptic vesicles and by Active zones are composed of evolutionarily conserved decreasing the presynaptic localization of Ca2+ proteins, including Munc13s, RIMs, RIM-BPs (RIM-binding pro- channels. Strikingly, rescue of the decreased Ca2+- teins), ELKS’s, and a-liprins (Wojcik and Brose, 2007). Of these channel localization required the RIM PDZ domain, proteins, RIMs are likely the central organizers because they directly or indirectly interact with all other known active-zone whereas rescue of vesicle priming required the RIM proteins and with synaptic vesicles (Mittelstaedt et al., 2010). N terminus. We propose that RIMs tether N- and P/ 2+ RIM proteins are expressed in three principal isoforms (Kaeser Q-type Ca channels to presynaptic active zones et al., 2008; Wang et al., 1997, 2000): RIM1a and RIM2a that via a direct PDZ-domain-mediated interaction, contain all RIM domains (i.e., N-terminal Rab3- and Munc13- thereby enabling fast, synchronous triggering of binding sequences, central PDZ domains, and C-terminal C2A neurotransmitter release at a synapse. and C2B domains with an intercalated PxxP sequence that binds to RIM-BPs); RIM1b and RIM2b that are identical to RIM1a and INTRODUCTION RIM2a but lack only the N-terminal Rab3-binding sequences (RIM1b) or both the N-terminal Rab3- and Munc13-binding se- At a synapse, action potentials induce Ca2+ influx into a presyn- quences (RIM2b); and RIM2g, RIM3g, and RIM4g that are aptic terminal, which triggers rapid synchronous neurotrans- composed only of C2B domains and are not considered here mitter release, thereby initiating synaptic transmission (Katz, further. Genetic experiments in C. elegans and mice revealed 1969). Release is mediated by synaptic vesicle exocytosis at that RIMs are essential for synaptic vesicle docking and priming the active zone, a specialized region of the presynaptic plasma and for presynaptic plasticity (Castillo et al., 2002; Fourcaudot

282 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. et al., 2008; Gracheva et al., 2008; Kaeser et al., 2008; Koushika (DxWC). Among 84 and 134 independent prey clones obtained et al., 2001; Schoch et al., 2002, 2006), but their mechanism of in P/Q- and N-type Ca2+-channel screens, 33 and 16 prey action remains unclear. clones, respectively, represented RIM-BPs (Figure 1B), consis- Several presynaptic proteins were shown to interact with Ca2+ tent with earlier studies (Hibino et al., 2002). In addition, 2 and channels. However, none of the reported interactions is selective 3 independent prey clones, respectively, contained RIM1 frag- 2+ for N- and P/Q-type Ca channels. For example, (1) the RIM C2A ments, whose only overlapping sequence encoded the PDZ 2+ and C2B domains bind to a1 subunits of L- and N-type Ca domain (Figure 1C). channels (Coppola et al., 2001), (2) the RIM C2B domain interacts The direct interaction of the RIM1 PDZ domain with P/Q- and with the b4Ca2+ channel subunit (Kiyonaka et al., 2007) that N-type Ca2+ channels was unexpected, prompting us to quantify associates with all Ca2+ channels subtypes but is not required it using liquid yeast two-hybrid assays. We found that the RIM1 for neurotransmitter release (Qian and Noebels, 2000), and (3) PDZ domain strongly and specifically bound to P/Q- and N-type the proline-rich sequences of RIMs bind to RIM-BPs (Wang Ca2+ channels, whereas no other RIM1 domain tested exhibited et al., 2000) that in turn bind to L-, N-, and P/Q-type Ca2+ chan- Ca2+-channel binding activity (Figures 1D–1F). Mutations in nels (Hibino et al., 2002). Thus, no plausible hypothesis at the three conserved motifs of the cytoplasmic sequences of present suggests how N- and P/Q-type Ca2+ channels are selec- N- and P/Q-type Ca2+ channels showed that only the C-terminal tively recruited to presynaptic active zones. More importantly, no sequence motif was essential for binding to RIM1 PDZ domains, presynaptic protein has been identified that is essential for re- as predicted for a PDZ-domain interaction (Figures 1E and 1F). cruiting N- and P/Q-type Ca2+ channels to presynaptic terminals. To validate the interaction of RIM PDZ domains with Ca2+ Only a-, which are presynaptic cell-adhesion mole- channels by an independent method, we employed NMR spec- cules, were found to be required for presynaptic Ca2+-channel troscopy. We produced a recombinant 15N-labeled RIM1 PDZ function (Missler et al., 2003). a-neurexins, however, are also domain and acquired 1H-15N heteronuclear single quantum essential for the organization of other components of the presyn- coherence (HSQC) spectra in the absence or presence of unla- aptic release machinery, and no molecular mechanism is known beled peptide from the C terminus of the P/Q-type Ca2+ channel that links a-neurexins to active zones or Ca2+ channels. (residues RHDAYSESEDDWC). Previous studies of the 1H-15N Using an unbiased yeast two-hybrid screen, we here identify HSQC spectrum of the RIM1 PDZ domain (Lu et al., 2005) a direct interaction of P/Q- and N-type Ca2+ channels with RIM allowed us to assign cross-peak shifts induced by the Ca2+- PDZ domains. To test whether RIMs act to localize Ca2+ chan- channel peptide to specific residues in the RIM PDZ domain nels to active zones, and whether this function requires the (see Extended Experimental Procedures, experimental ratio- PDZ domains of RIMs, we generated conditional double nale 1). We found that the C terminus of P/Q-type Ca2+ channels knockout (KO) mice of all RIM isoforms that contain PDZ directly bound to the -binding pocket of the RIM1 domains. Using these mice, we show by electrophysiological PDZ domain (Figures 1G and 1H and Figure S1B). Additional recordings, Ca2+ imaging, and quantitative immunostaining of 1H-15N HSQC experiments revealed analogous binding of the Ca2+ channels that RIMs are essential for localizing Ca2+ chan- C terminus of N-type Ca2+ channels to the RIM1 PDZ domain, nels to release sites. Moreover, we show that only two RIM and of Ca2+ channels to the RIM2 PDZ domain (Figures S1C sequences are required for localization of Ca2+ influx to active and S1D). zones: their PDZ domains that bind to Ca2+ channels and their An atomic model suggests that the Ca2+-channel sequence proline-rich sequences that bind to RIM-BPs, which in turn fits well into the PDZ-domain-binding pocket (Figures 1H and bind to Ca2+ channels. Thus, we propose that RIMs tether Ca2+ 1I). The binding envisioned in this model agrees with the shifts channels to active zones via two parallel essential interactions, we observed in the HSQC spectra in cross-peaks from residues direct binding of Ca2+ channels to the RIM PDZ domains that such as K651, K653, and K694, which contribute to the binding is specific for N- and P/Q-type Ca2+ channels and indirect pocket of the RIM1 PDZ domain and move upon P/Q-type Ca2+- binding of Ca2+ channels to RIMs via RIM-BPs that is shared channel peptide binding (Figure 1G). Furthermore, we confirmed among different types of Ca2+ channels. these interactions by isothermal calorimetry with the RIM1 PDZ domain (Figures S1E and S1F), yielding dissociation constants similar to those of other PDZ-domain interactions (P/Q-type RESULTS Ca2+ channel, 10.3 ± 0.6 mM; N-type Ca2+ channel, 23.4 ± 1.7 mM; Wiedemann et al., 2004). Thus, RIM PDZ domains stoi- A Screen for Synaptic Proteins Binding to Ca2+ Channels chiometrically interact with the C-terminal sequences of N- and We performed yeast-two hybrid screens for proteins that P/Q-type Ca2+ channels, which are not found in L- and T-type interact with the cytoplasmic C-terminal sequences of P/Q- Ca2+ channels. and N-type Ca2+ channels. We chose these Ca2+-channel baits because P/Q- and N-type Ca2+ channels mediate nearly all Conditional KO Mice for Presynaptic RIM Isoforms presynaptic Ca2+ influx, and their C termini have been impli- The binding of RIM PDZ domains to Ca2+ channels was unex- cated in targeting Ca2+ channels to the active zone (Catterall pected because RIM PDZ domains interact with the active- et al., 2005). The C termini of N- and P/Q-type Ca2+ channels zone protein ELKS (a.k.a., Rab6-binding protein, CAST, or contain three motifs (Figure 1A and ERC; Ohtsuka et al., 2002; Wang et al., 2002). To test whether Figure S1A available online): an SH3 domain-binding motif the RIM PDZ domain physiologically binds to presynaptic Ca2+ (RQLPGTP), a PNGY motif, and a C-terminal sequence motif channels, we generated conditional double KO mice in which

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 283 A outside

inside PxxP PNGY I II III IV DDWC (P/Q-type) DHWC (N-type) +H N 3 Ca2+-channel bait B RIM-BP2 SH3 FN3 FN3 FN3 SH3 SH3 Ca2+- P/Q-type channel prey bait clones N-type

RIM1α Zn PDZ C ACPxxP B C   2 2 Ca2+- 350 P/Q-type 386 prey channel 100 aa 48 bait 56 clones N-type 65

P/Q-type D Ca2+-channel bait E N-type Ca2+-channel bait F RIM1 bait wild-type pLexN PxxPM PNGYM ΔCterm pLexN A B AB 2 2 2 PDZ C C C PDZ 4 RIM1 48 clone 1.6 1.0 RIM1 56 clone RIM1 65 clone

2 0.5 0.8

n.d. n.d. n.d. n.d. n.d. n.d. n.d. interaction strength (a.u.) 0 interaction strength (a.u.) 0 interaction strength (a.u.) 0 wild-type M M term Ca2+-channel SH3 ΔC control control control control control control control prey PNGY wild-type RIM1 prey RIM1 prey RIM1 prey RIM1 prey RIM1 prey RIM1 prey clones clones clones clones clones clones

GHRIM1 PDZ RIM1 PDZ + P/Q-type Ca2+-channel RIM1 PDZ 102 G662 N

G675 107 C

G680 S695 112 G646 C G667 D659 αB Q698 N 117 βB N (ppm) 15 T650 K651 K653 122 N637 K653 E696 I L629 A681 K651 L671 K694 K694 127 W673sc K633 I701 C K653

G631 N674 L691 132

K654 N Y687

11 10 9 8 7 1H (ppm)

Figure 1. Direct Interaction of P/Q- and N-Type Ca2+ Channels with RIM PDZ Domains (A) Structure of the a1 subunits of P/Q- and N-type Ca2+ channels. Following the 4 3 6 transmembrane regions (I–IV), P/Q-type and N-type Ca2+ channels contain a C-terminal cytoplasmic tail with conserved SH3-domain binding sequences (PxxP), PNGY motifs, and C-terminal sequence motifs (DxWC).

284 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. all RIM isoforms containing PDZ domains (RIM1a,1b,2a, and 2b) strongly desynchronized release (Figures 3L and 3M). These can be deleted by cre-recombinase (Figures 2A–2C and Fig- data suggest that RIMs are essential not only for vesicle priming, ure S2), thereby enabling us to avoid the lethality of RIM-deficient a previously identified RIM function that partly accounts for the mice (Kaeser et al., 2008; Schoch et al., 2006). We cultured decrease in release (Koushika et al., 2001; Schoch et al., from newborn conditional double KO mice and infected 2002), but also for the synchronous timing of fast release that these neurons with lentiviruses expressing EGFP-tagged active is not accounted for by a priming deficit. Given that RIMs directly or inactive cre-recombinases. Rescue experiments were per- bind to Ca2+ channels (Figure 1), and that a loss of Ca2+ channels formed by coexpressing various RIM proteins from the same from presynaptic terminals would explain the impaired synchro- lentiviruses via an IRES sequence (Kaeser et al., 2009). nous timing of release, we hypothesized that RIM binding to N- Immunoblotting demonstrated that after 10 days in vitro and P/Q-type Ca2+ channels may tether Ca2+ channels to the (DIV10), neurons expressing active cre-recombinase (referred active zone, thereby increasing the efficiency, speed, and to as cDKO neurons) lack RIM proteins, whereas neurons synchrony of release. expressing inactive cre-recombinase (referred to as controls) retain RIM expression (Figure 2D). Despite lacking RIM proteins, RIM Deletion Alters the Ca2+ Dependence of Release cDKO neurons exhibited an overall normal morphology with To explore the hypothesis that RIMs localize Ca2+ channels to unchanged synapse size and density (Figures 2E and 2F). Elec- active zones, we first examined the speed with which addition tron microscopy revealed that in cDKO neurons, the number of of a membrane-permeable Ca2+ buffer (EGTA-AM) decreases docked synaptic vesicles per active zone was decreased nearly Ca2+-triggered release, measured as the inhibitory postsynaptic 2-fold (Figures 2G and 2H), consistent with a role for RIM in current (IPSC) amplitude (see Extended Experimental Proce- vesicle docking (Gracheva et al., 2008; see also Han et al., dures, experimental rationale 2). EGTA-AM caused a significantly 2011). All other measured parameters were unchanged, and faster rate of decline in IPSC amplitude in RIM-deficient cDKO removal of RIMs did not alter the structure of presynaptic neurons than in control neurons (Figures 3N–3P). Thus, EGTA- dense projections visualized by phosphotungstic acid staining dependent chelation of Ca2+ inhibits release more effectively in (Figure 2G). RIM-deficient than in control neurons, consistent with a longer average distance between Ca2+ channels and release sites in RIM Deletion Severely Impairs Neurotransmitter RIM-deficient synapses. Release We next examined the dependence of release in RIM-deficient 2+ 2+ Electrophysiologically, deletion of RIM proteins caused a 3- to 4- synapses on the extracellular Ca concentration ([Ca ]ex; Fig- fold decrease in the frequency of spontaneous ‘‘minis,’’ in the ure 4). If RIM-deficient presynaptic terminals contain fewer amplitude of postsynaptic currents evoked by isolated action tethered Ca2+ channels, presynaptic Ca2+ influx should be 2+ potentials, and in the size of the readily releasable pool mea- decreased, and more [Ca ]ex should be required for equivalent 2+ sured by application of hypertonic sucrose (Figures 3A–3D and amounts of release—i.e., release should exhibit a higher [Ca ]ex Deng et al., 2011). Moreover, deletion of RIMs significantly dependence without a change in apparent Ca2+ cooperativity decelerated and desynchronized release, as evidenced by an (see Extended Experimental Procedures, experimental ratio- increase in rise times and in rise-time variability (Figures 3E– nale 3). However, because Ca2+ influx through Ca2+ channels 2+ 3G) but did not change the relative contributions of P/Q- and saturates at high [Ca ]ex (Church and Stanley, 1996; Schneg- 2+ 2+ N-type Ca channels to evoked synaptic responses (Figures genburger et al., 1999), [Ca ]ex titrations underestimate the S3A and S3B). Similarly, the RIM deletion massively decreased change in Ca2+ dependence of release in mutant synapses, release induced by stimulus trains (30 stimuli at 10 Hz; Figures and only relative changes are interpretable. 3H and 3I), decelerated release as manifested by a relative RIM-deficient cDKO neurons exhibited a large reduction in 2+ increase in delayed release (Figures 3J and 3K), and again neurotransmitter release at all [Ca ]ex (Figures 4B and 4C) and

(B and C) Summary of the RIM-BP (B) and RIM prey clones (C) isolated in yeast two-hybrid screens with the C-terminal sequences of N- and P/Q-type Ca2+ channels. (D–F) Liquid yeast-two hybrid assays with baits containing wild-type C-terminal sequence of the P/Q-type Ca2+ channel and the three indicated RIM prey clones (D); baits containing the C-terminal sequence of the N-type Ca2+ channel with point mutations in the PxxP (PxxPM) or the PNGY sequence (PNGYM), or with a deletion of the 4 C-terminal residues (DCterm), and the indicated RIM prey clones (E); and baits containing the indicated RIM1 domains (PDZ, PDZ domain only;

C2AorC2B, C2AorC2B domains only; C2AB, both C2 domains and the intercalated PxxP motif) and preys consisting of the wild-type C-terminal sequence of the N-type Ca2+ channel (F, left bars), or the indicated mutants of this Ca2+ channel (F, right bars). For all assays, pLexN served as a control; a.u. = arbitrary units; n.d. = not detectable (means ± standard error of the mean [SEM]). (G) Analysis of P/Q-type Ca2+ channel binding to the RIM1 PDZ domain by NMR spectroscopy. 1H-15N HSQC spectra of the 15N-labeled RIM1 PDZ domain (38 mM) were acquired in the absence (black contours) and presence (blue contours) of unlabeled P/Q peptide (0.1 mM). Selected cross-peak assignments from residues on the periphery of the binding site are indicated; cross-peaks from three lysine residues in the binding pocket that shift upon peptide binding are labeled in bold, underlined typeface (K651, K653, and K694). (H) Model of the RIM1 PDZ domain (blue ribbon diagram; orange indicates residues corresponding to shifted cross-peaks assigned in panel G) bound to the six C-terminal residues of the P/Q-type Ca2+ channel peptide represented as a stick model with color-coded atoms (carbon, yellow; oxygen, red; nitrogen, blue; sulfur, orange). Strand bB and helix aB, the two structural elements that line the peptide-binding site (Lu et al., 2005), are indicated. (I) Close-up view of the surface of the RIM1 PDZ domain peptide-binding pocket with the bound P/Q-type Ca2+ channel peptide (colors are identical to panel H). For additional 1H-15N HSQC spectra and affinity measurements by isothermal titration calorimetry, see Figure S1.

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 285 Figure 2. Conditional Deletion of RIM Proteins in Mice RIM2αβγ floxed or KO A (A) Structure of the RIM2 gene (a.k.a., Rims2). Exons are shown as black 1’ 2-4 571’’ 6 -251’’’ 26-30 boxes and numbered, positions of exons containing the initiator codons for RIM2a, RIM2b, and RIM2g are labeled 10,100, and 1000, respectively. The γ 2α 2β 100 kb 2 first exon that is shared by all RIM2 isoforms (exon 26) was used for gene targeting in the conditional RIM2abg KO mice (shaded blue area). 1’’’ 26 27 - 29 B wild-type allele (B) RIM2abg targeting strategy. The diagram shows (from top to bottom) an expanded map of the RIM2 gene surrounding exon 26; the targeting targeting construct N C DT vector (C = ECFP-tetracysteine tag in exon 26; blue triangles = loxP sites; N = neomycin resistance cassette; green circles = frt recombination original KI allele N C sites; DT = diphtheria toxin gene cassette); the knockin allele (KI); the RIM2abgfloxed allele (neomycin resistance cassette was removed by flp- RIM2αβγ floxed C recombination); and the KO allele (cre recombination deleted exon 26, creating a nontranslated, unstable mRNA). RIM2αβγ KO (C) Domain structures of RIM1a,1b,2a,2b, and 2g that are deleted in the RIM1/RIM2 conditional double KO neurons. Coils surrounding the 2 kb N-terminal Zn2+-finger domain (Zn) signify Rab3-binding sequences. αβfloxed C RIM1 or KO (D) Representative immunoblots of RIM1 and RIM2 proteins in cultured hippocampal neurons from RIM1/RIM2 double conditional KO mice in- 1α  PxxP fected with lentiviruses expressing inactive (control) or active cre- Zn PDZ C ACB 1β 2 2 recombinase (cDKO). Neurons were infected on DIV3 and analyzed at the indicated times (DIV6–DIV14). αβγ floxed RIM2 or KO 200 amino acids (E) Representative images of cDKO and control neurons stained with antibodies to MAP2 (green) and synapsin (red). Scale bar = 5 mm, applies to all images. 2α Zn PxxP  (F) Quantitations of the size and density of synapses analyzed as shown in PDZ C2A 2β C B (E) (control, n = 18 neurons/3 independent cultures; cDKO neurons, 17/3). γ 2 2 (G) Electron micrographs of osmium tetroxide- (top) or phosphotungstic D DIV 6 8 10 12 14 acid-stained (bottom) control and cDKO neurons (scale bars, 200 nm). (H) Quantitations of synaptic ultrastructure in electron micrographs. Docked vesicles are defined as vesicles touching the plasma membrane control cDKO control cDKO control cDKO control cDKO control cDKO (DP, dense presynaptic projections; PSD, ). ± RIM1/2 Data in (F) and (H) show means SEM. Statistical significance by Student’s t test: ***p < 0.001; for additional detailed information, see Figure S2 and Table S1. VCP β-

E control cDKO F ) 2 0.6 4 μ m

MAP-2 0.4

2 m dendrite m

0.2 μ

Syn

puncta per per puncta 10 puncta size ( 0.0 0

cDKO merge cDKO control control

G control cDKO H 40 4 * * 2 20

per active zone

docked vesicles

0 vesicles per bouton 0

3 m)

μ 0.4

m) μ 2 gth ( gth 0.2

bouton bouton 1 DP DP

0 len PSD 0.0 circumference ( circumference

PSD PSD cDKO cDKO control control

286 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. A EPSCs B EPSCs Figure 3. RIM Deletion Decreases, Decelerates, and Desynch- control control ronizes Neurotransmitter Release Sucrose 6 (A and B) Excitatory synaptic responses in cultured hippocampal control and 0.4 cDKO neurons evoked by an (A) or hypertonic sucrose 4 application (B) (left, representative traces; right, summary graphs of ampli- tudes or charge transfers; A: control, n = 8 neurons/3 independent neuronal cDKO 0.2 * * * cultures; cDKO, n = 9/3, B: control, n = 9/3; cDKO, n = 10/3). * cDKO 2 * * Sucrose 0.1 nA (C and D) Inhibitory synaptic responses evoked by an action potential (C) or charge transfer (nC)

EPSC amplitude (nA) 0 hypertonic sucrose application (D) (C: control, n = 21/4; cDKO, n = 18/4; D: 10 ms 0.0 0.25 nA control, n = 11/3; cDKO, n = 11/3). 5 s cDKO cDKO control control (E–G) Analysis of the kinetics of isolated IPSCs (E, superposed representative C IPSCs D IPSCs traces from control and cDKO neurons; F, 20%–80% IPSC rise times; and G, 4 control 16 rise time variability as expressed by the standard deviation (SD) of the 20%– control Sucrose 80% rise time [control, n = 31/6; cDKO, n = 36/6]). 3 12 (H–K) Synaptic responses elicited by a 10 Hz stimulus train in cDKO and 2 8 * control neurons (H, representative IPSCs; I–K, summary graphs of the synaptic * charge transfer for the first IPSC [I] and for delayed release [J; release starting cDKO * * 1 cDKO 4 100 ms after the last stimulus] and of the ratio of delayed release/first * Sucrose response) (K; control, n = 20/4; cDKO, n = 21/4). IPSC amplitude (nA)

0.5 nA charge transfer (nC) 0 0 (L and M) Analysis of the kinetics of IPSCs during 10 Hz stimulus trains (L, 100 ms 0.5 nA superposed representative traces of the first 10 IPSCs during a 10 Hz stimulus cDKO 5 s cDKO control control train applied for 3 s [top, first response indicated as a thick line, later responses E control F G *** represented as thin lines], and 20%–80% rise times for three sample trains 3 *** [bottom]; M, standard deviation (SD) of the 20%–80% rise times during the 0.6 10 Hz stimulus train as a measure of synchrony; control, n = 7/3; cDKO, n = 9/3). 2 (N–P) Time course of the decrease in IPSCs induced by addition of the 0.5 nA 0.4 2+ 5 ms membrane-permeable Ca -chelator EGTA-AM (10 mM; N, sample traces; O, cDKO summary graphs; P, decay time constants; control, n = 8/3; cDKO, n = 9/3). 1 0.2 Decay time constants t were calculated by fitting individual experiments to rise time (ms) SD of 20-80%

20-80% rise time (ms) 0 0.0 a single exponential function. 0.1 nA All data are means ± SEMs; *p < 0.05, **p < 0.01, ***p < 0.001 as determined by 5 ms cDKO cDKO Student’s t test. Numerical values of electrophysiology results are in Table S2, control control further analysis of synaptic responses elicited at 10 Hz in Figure S3. H control IJK 0.3 20 ** 1.2 2+ a major shift in the [Ca ]ex dependence of neurotransmitter 0.2 0.8 * release (Figure 4D). Both phenotypes were equally observed in * 10 * * inhibitory and excitatory synapses and fully rescued by full- 0.1 * 0.4 length wild-type RIM1a (Figures 4A–4D and Figures S4A–S4N).

cDKO delayed release/ 2+ Fitting the [Ca ]ex-response curve of individual experiments 1st peak charge ratio delayed release (nC) first peak charge (nC) 0.0 0.0 0 to a Hill function showed that the RIM deletion increased the 0.5 nA [Ca2+] requirement for release almost 2-fold, without changing 1 s cDKO ex cDKO cDKO control control control the apparent Ca2+ cooperativity of release (Figure 4E and LMcontrol cDKO ** Figure S4H). Note that in the Hill function fits, the experimentally #2-10 #1 5 2+ measured amplitudes suggest near saturation at 10 mM [Ca ]ex, #2-10 4 allowing direct comparison of the fitted parameters (Figure S4I). #1 2+ 1 nA 0.2 nA 3 The [Ca ]ex titration provided us with a facile assay to 10 ms 10 ms 2 examine which RIM sequences determine vesicle priming and 20 20 the [Ca2+] dependence of release. We first tested rescue of rise time (ms) ex

SD of 20-80% 1 10 10 the RIM double KO phenotype by the N-terminal RIM-RZ frag- 0 2+ 0 0 ment that contains the Rab3-binding (‘‘R’’) and Zn -finger rise time (ms) IPSC 20-80% rise time (ms) 15 30 IPSC 20-80% 15 30 AP number AP number cDKO control domains (‘‘Z’’) of RIM1a and that had been previously implicated N OPin vesicle priming (Dulubova et al., 2005), and the C-terminal EGTA-AM 1.0 RIM-PASB fragment that contains its PDZ (‘‘P’’), C2A (‘‘A’’), 0 min 6 min 12 min control 6 control 0.8 PxxP (‘‘S’’ for SH3 domain binding), and C2B domains (‘‘B’’) (Fig- 2+ ure 4A). At 2 mM [Ca ]ex, the RIM-RZ and RIM-PASB fragments 0.6 * 4 each partially rescued the decrease in release and alleviated the 0.4 cDKO 2 previously described decrease in Munc13 protein levels in cDKO 0.2 neurons (Figures 4F and 4I; Figure S4D; Schoch et al., 2002). cDKO 0 However, whereas the RIM-PASB fragment completely reversed normalized IPSC ampl. 0.0 04 812 decay time const. (min) 2+ 0.5 nA the impairment of the [Ca ]ex dependence of release in cDKO 100 ms time (min) cDKO control neurons, the RIM-RZ fragment did not (Figures 4F–4K). RIM- RZ, conversely, rescued the readily releasable pool, whereas

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 287 A R Z PA S B RIM-PASB did not (Deng et al., 2011). Note that different from RIM1α Zn PDZ C2A PxxP C2B mutations (Shin et al., 2009), none of the RIM RIM-RZ ZnH manipulations changed the apparent Ca2+ cooperativity of

RIM-PASB PDZ C2A PxxP H C2B release (Figure S4H; see also Table S3).

B [Ca2+] (mM): 0.5 1 2 5 10 We next tested whether RIM-RZ or RIM-PASB rescued the ex 2+ control speed and synchrony of release. Consistent with the [Ca ]ex titrations, the RIM-PASB fragment fully reversed the decelera- cDKO tion and desynchronization of release in cDKO neurons, as as- cDKO + 2+ RIM1α sessed at 2 mM [Ca ]ex, whereas the RIM-RZ fragment did cDKO + not (Figures 4L and 4M and Figures S4O–S4Q). Together, these RIM-RZ data indicate that the N-terminal RIM domains function in vesicle cDKO + docking and priming (Betz et al., 2001; Dulubova et al., 2005; RIM-PASB 2 nA Gracheva et al., 2008; Koushika et al., 2001; Schoch et al., 250 ms 2002), whereas the C-terminal RIM domains function in the C D E 2+ 1.0 Ca dependence and synchrony of release. 3 *** control 0.8 4 (mM) RIM Deletion Impairs Presynaptic Ca2+ Influx cDKO + ex

] 2 α 0.6 control 2+

RIM1 2+ The increased [Ca ]ex dependence, decreased synchrony, and 2 0.4 cDKO + 1 lowered speed of release in RIM-deficient neurons support the α 2+ cDKO 0.2 RIM1 of [Ca hypothesis that RIMs localize Ca channels to the active normalized IPSC cDKO 50

IPSC amplitude (nA) 0 0.0 0

EC zone. However, these measurements are indirect. Their results 110 110 α 2+ 2+ could be explained by other hypotheses, for example that RIM [Ca ]ex (mM) [Ca ] ex (mM) cDKO control cDKO + 2+ RIM1 directly regulates Ca triggering of release by binding to synap- F G H totagmin (Coppola et al., 2001; Schoch et al., 2002). To address 1.0 *** 3 *** this issue with an independent approach, we monitored action- control 0.8 2+ 2+ 4 control (mM) potential-induced Ca transients by Ca imaging in presyn- ex 0.6 ] 2 cDKO + cDKO + 2+ aptic boutons and postsynaptic dendrites. 2 RIM-RZ 0.4 RIM-RZ We engineered new active and inactive EGFP-tagged cre- cDKO 1 0.2 of [Ca recombinase proteins that exhibit a tight nuclear localization

cDKO 50 normalized IPSC

IPSC amplitude (nA) (Figures S5A–S5C), and expressed them in neurons from condi- 0 0.0 EC 0 110 110 2+ 2+ tional KO mice. We then filled single neurons via a patch pipette [Ca ] ex (mM) [Ca ] (mM) cDKO ex control 2+ cDKO + RIM-RZ with Alexa 594 and the Ca indicator Fluo5F, identified presyn- I J K aptic axonal boutons and second-order dendrites by imaging 1.0 6 cDKO + 3 *** Alexa 594, elicited isolated action potentials by brief somatic control RIM-PASB 0.8 2+

(mM) current injections, and monitored the resulting Ca transients

cDKO + control ex 4 RIM-PASB 0.6 ] 2 in boutons and dendrites by imaging Fluo5F in line scans (at 2+

0.4 2 cDKO 1 0.2 of [Ca cDKO 50 2+ normalized IPSC (B) Representative traces of IPSCs evoked at the indicated extracellular Ca EC IPSC amplitude (nA) 0.0 0 2+ 0 concentrations [Ca ]ex in control neurons, cDKO neurons without rescue, 110 110 2+ 2+ cDKO neurons with full-length RIM1a rescue, and cDKO neurons with rescue [Ca ] (mM) [Ca ] (mM) cDKO ex ex control cDKO + with the RIM-RZ or the RIM-PASB fragments. Each rescue experiment was RIM-PASB LM performed with independent control groups. *** *** ** (C–K) Summary plots of absolute (C, F, and I) and normalized IPSC amplitudes 3 ** 2+ * ** ** (D, G, and J; normalized to the 10 mM [Ca ]ex response) evoked at the indi- 0.8 2+ 2+ *** cated [Ca ]ex, and summary graphs of the Ca dependence of release (E, H, 2 2+ and K; expressed as the [Ca ]ex producing a half-maximal IPSC amplitude 0.4 [EC ], as determined by fitting in individual experiments the [Ca2+] depen- 1 50 ex dence of the IPSC amplitude [C, F, and I] to a Hill function). Control neurons rise time (ms) 0 0.0 and cDKO neurons were analyzed in comparison with cDKO neurons rescued SD of rise time (ms) with RIM1a (C–E), RIM-RZ (F–H), or RIM-PASB (I–K). (C and D): n = 8 neurons/3

cDKO cDKO cDKO cDKO control control cDKO control control control cDKO control independent batches of culture in control, 6/3 in cDKO, 9/3 in cDKO + RIM1a; cDKORIM1 + α cDKO + cDKO + cDKORIM1 + α cDKO + cDKO + RIM-RZ RIM-RZ RIM-PASB RIM-PASB (E and F): n = 7/3 in control, 7/3 in cDKO, 7/3 in cDKO + RIM-RZ; (G and H): n = 6/3 in control, 5/3 in cDKO, 8/3 in cDKO + RIM-PASB. Figure 4. Mutational Dissection of the RIM KO Phenotype (L and M) Summary graphs of 20%–80% rise times (L) and rise time variability 2+ (A) Diagram of RIM rescue proteins expressed in cDKO neurons via an IRES (M) for the indicated rescue experiments at 2 mM [Ca ]ex (for sample traces, sequence from the same mRNA as cre-recombinase. The single-letter code see Figure S4O; for number of neurons/independent cultures analyzed, see above the RIM1a diagram identifies the various domains (R, Rab3-binding panels C–K). a-helical region; Z, Zn-finger region, P, PDZ domain; A, C2A domain; S, Data shown are means ± SEM, ***p < 0.001 by one-way ANOVA; detailed 2+ proline-rich SH3-binding PxxP motif; B, C2B domain); H marks the presence of statistical analysis for all data points can be found in Table S3.ForCa co- a human influenza hemagglutinin (HA)-tag. operativity and Imax, see Figure S4.

288 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. A lortnoc OKDc cDKO + RIM-PASB Figure 5. RIM Deletion Decreases Presynaptic Ca2+ Transients (A) Representative fluorescence images of control neurons, cDKO neurons, and cDKO neurons rescued with the C-terminal RIM-PASB fragment. Neurons were filled

Alexa594 via a patch pipette with Fluo5F and Alexa 594 (red); nuclear EGFP-fluorescence (produced by the active and inactive cre-recombinase EGFP-fusion proteins; see Figures S5A– S5C) is shown in green; and coincident Alexa 594 GFP /Fluo5F/ and EGFP- or Fluo5F signals are shown in yellow. Insets (bottom right) show areas in dotted rectangles containing a sample axonal bouton (gray lines = positions B of the patch pipette; white lines = positions of line scans for 20 mV 2+ 50 ms the Ca transients shown in B). Scale bar (bottom left) = -70 mV 20 mm. (B) Representative action potentials (top); line scans of Ca2+ transients in presynaptic boutons induced by these Fluo5F action potentials and monitored via Fluo5F fluorescence 50% ΔG/G 0 (middle; colored white for better visibility); and quantita- 50 ms tions of Ca2+ transients (bottom; averaged across the bouton). (C and D) Summary plots of action potential-induced bouton dendrite bouton changes in Ca2+-indicator fluorescence monitored in C n.s. D E 80 presynaptic boutons from control neurons, cDKO neurons, control cDKO 120 *** 0 60 202 1.0 RIM1/2 and cDKO neurons rescued with the C-terminal RIM-PASB 40 G/G 2+

Δ fragment (C, time course of the Ca -indicator fluorescence 100 20 *** 202 Ca 2.1-A 0 0.8 V [inset: the same plot for dendrites]; D, the cumulative 2+ 80 0 200 400 600 Ca 2.1-S probability of the peak Ca -indicator fluorescence, ex- time (ms) 202 V 0.6 pressed as DG/G0). Data in (C) are means (line) ± SEM 60 control 202 CaV2.2

(% change) cDKO (shaded area);***p < 0.001 as assessed by two-way ANOVA 0 Ca α /δ 40 cDKO + 0.4 116 V 2 1 for peak amplitudes during the first 60 ms after action RIM-PASB G/G control Ca β4 potential induction (C) or by Kolmogorov-Smirnov test (D); Δ V

20 cumulative probability 0.2 cDKO 50 control, n = 45 boutons/10 neurons/4 independent cultures; cDKO + Liprin-α3 RIM-PASB 116 0 87 cDKO, n = 46/11/4; cDKO + RIM-PASB, n = 44/11/4. 0.0 GDI (E) Immunoblot analysis of Ca2+-channel subunit levels in 0 100 200 300 400500 600 0.0 0.5 1.0 1.5 2.0 50 β-actin control and cDKO neurons. Blots were probed with anti- time (ms) ΔG/G peak amplitude 0 bodies to the indicated Ca2+-channel proteins (P/Q-type

[CaV2.1-A and CaV2.1-S] and N-type [CaV2.2] a subunits, and a2/d1 and b4 subunits) and control proteins (GDI, GDP-dissociation inhibitor); numbers indicate positions of molecular weight markers. For analysis of dendritic Ca2+ transients, statistical values, and quantitative assessment of mRNA levels, see Figure S5 and Table S4.

333 Hz, 100–150 mm from the cell body). To ensure that the decreased presynaptic Ca2+ influx in hippocampal neurons. observed Ca2+ transients were not due to passive depolariza- A parallel study extends this finding by direct measurements tions in response to the somatic current injections, we decreased of Ca2+ currents in RIM-deficient calices of Held (Han et al., in control experiments the injected current to a threshold level 2011). that only sometimes evoked an action potential. In these exper- iments, Ca2+ transients in boutons and dendrites strictly de- RIM PDZ Domain Is Required for Localizing Ca2+ Influx pended on the induction of action potentials, confirming that We next tested whether the RIM PDZ domain mediates the RIM- we were monitoring action-potential-induced Ca2+ transients dependent presynaptic localization of Ca2+ influx, using system- 2+ (Figures S5D and S5E). atic rescue experiments combined with [Ca ]ex titrations as In RIM-containing control boutons, isolated action potentials an assay. induced a brief, 100% increase in Fluo5F Ca2+-indicator Deletion of either the PDZ domain or the PxxP motif from the fluorescence, whereas in RIM-deficient cDKO boutons, action RIM-PASB fragment (Figure 6A, top) blocked rescue of the 2+ potentials induced only an 50% increase (Figures 5A–5D). impaired [Ca ]ex dependence in RIM-deficient cDKO neurons, 2+ Similar to the impaired [Ca ]ex dependence of release (Figure 4), whereas deletion of the C2A or the C2B domains had no effect the decreased Ca2+ influx in RIM-deficient cDKO neurons was (Figures 6A–6E and Figures S6A–S6C). Moreover, expression fully rescued by the C-terminal RIM-PASB fragment (Figures of a RIM fragment composed only of the PDZ domain and 5A–5D). Deletion of RIM proteins did not alter dendritic Ca2+ tran- PxxP motif (‘‘RIM-PS’’) in cDKO neurons fully rescued the 2+ sients (Figure 5C, inset and Figures S5F–S5H), suggesting that impaired [Ca ]ex dependence of release (Figures 6F–6H and the RIM deletion did not generally impair Ca2+-channel function Figures S6D–S6F). In contrast, expression of a RIM fragment 2+ or Ca buffering. Moreover, we detected no change in expres- composed of its two C2 domains (‘‘RIM-AB’’) did not rescue 2+ 2+ sion levels of various Ca -channel subunits in cDKO neurons the [Ca ]ex dependence of release but partially reversed the (Figure 5E and Figure S5I). Thus, RIM deletions selectively decrease in IPSC size in RIM-deficient cDKO neurons,

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 289 [Ca2+] (mM): Figure 6. RIM PDZ Domain and PxxP Motif A RIM1α B ex 2+ 0.5 1 2 5 10 Restore Normal [Ca ]ex Dependence to control R Z PA S B RIM-Deficient Synapses Zn PDZ C2A PxxP C2B   cDKO (A) Domain structures of rescue proteins.

RIM-ASB C2A PxxP H C2B cDKO + (B) Sample traces of IPSCs in control neurons, RIM-ASB cDKO + cDKO neurons, and cDKO neurons rescued with RIM-PSB PDZ PxxP H C2B RIM-PSB the indicated proteins. 2+ RIM-PAB PDZ C A H C B cDKO + (C–E) Systematic rescue analyses of the [Ca ]ex 2 2 RIM-PAB dependence of release in RIM-deficient cDKO neu- PDZ C A PxxP H cDKO + RIM-PAS 2 RIM-PAS rons with RIM fragments containing three of the four PDZ PxxP H cDKO + RIM domains present in the RIM-PASB fragment. RIM-PS RIM-PS cDKO + Absolute IPSC amplitudes (C), IPSC amplitudes RIM-AB C2A H C2B 2+ RIM-AB 2 nA normalized to the response at 10 mM [Ca ]ex (D), 2+ 250 ms and apparent Ca affinities (EC50 values; E) are CDE indicated (control, n = 9 cells/3 independent batches 6 1.0 3 *** control *** of cultures; cDKO, n = 9/3; cDKO + RIM-ASB, 0.8 *** n = 9/3;cDKO+ RIM-PSB,n = 8/3;cDKO+RIM-PAB, RIM-PSB control (mM) 4 RIM-PSB ex n = 8/3; cDKO + RIM-PAS, n = 10/3). ] 2 RIM-PAB 0.6 2+ RIM-PAS 2+ RIM-PAS (F–H) Rescue analyses of the Ca dependence of RIM-ASB release with RIM fragments containing either only RIM-ASB 0.4 RIM-PAB 2 1 cDKO of [Ca the PDZ domain and PxxP motif (RIM-PS), or only

cDKO 50 normalized IPSC 0.2 the C2AandC2B domains (RIM-AB) of RIM1 IPSC amplitude (nA) EC (control, n = 8/3; cDKO, n = 7/3; cDKO + RIM-PS, 0 0.0 0 110 110 n = 9/3; cDKO + RIM-AB, n = 9/3). 2+ 2+ [Ca ] (mM) [Ca ] (mM) cDKO Data shown are means ± SEM; ***p < 0.001 by ex ex control RIM-ASBRIM-PSBRIM-PABRIM-PAS one-way ANOVA. Cooperative factor n and Imax cDKO + rescue can be found in Figure S6; all numerical data are in Table S5. F G 1.0 H control RIM-PS *** *** 0.8 4

(mM) 2 ex

0.6 ]

RIM-AB control 2+ RIM-PS 0.4 cDKO 2 1 The RIM PDZ Domain Localizes cDKO RIM-AB 2+ of [Ca P/Q-Type Ca Channels

0.2 50 normalized IPSC IPSC amplitude (nA)

EC to Presynaptic Terminals 0 0.0 0 2+ 110 110 The decrease in presynaptic Ca influx in [Ca2+] (mM) [Ca2+] (mM) ex ex cDKO control RIM-deficient terminals could be due to RIM-PSRIM-AB 2+ cDKO + a loss of presynaptic Ca channels or rescue to a decrease in their activity. To address this question, we measured by quantita- tive immunofluorescence presynaptic 2+ suggesting that the C2 domains are dispensable for localizing levels of P/Q-type Ca channels, which mediate >80% of the presynaptic Ca2+ influx but enhance the efficacy of release. synaptic responses in our preparation (Figures S3A–S3C). Strik- To directly probe whether the RIM PDZ domain is actually ingly, the RIM deletion reduced presynaptic P/Q-type Ca2+ essential for localizing presynaptic Ca2+ influx to active zones, channel levels 40% but had no effect on the active-zone we next performed rescue experiments with full-length wild-type protein bassoon (Figures 8A and 8B and Figures S8A and S8B). RIM1a or RIM1a lacking the PDZ domain (RIM-DPDZ). Whereas Finally, to test whether the presynaptic localization of the former rescued all RIM cDKO phenotypes (Figures 4C–4E), P/Q-type Ca2+ channels depends on the RIM PDZ domain, we 2+ 2+ the latter was unable to reverse the impairment in [Ca ]ex depen- examined rescue of the Ca -channel localization deficit in dence of release in RIM-deficient neurons, although it completely RIM-deficient cDKO neurons with full-length wild-type RIM1a 2+ rescued the decrease in evoked IPSC amplitude at high [Ca ]ex or RIM1a lacking the PDZ domain (Figures 8A and 8B). In agree- concentrations (Figures 7A–7E and Figures S7A–S7C). Moreover, ment with the electrophysiological and Ca2+-imaging results full-length but not PDZ-domain-deleted RIM1a rescued the decel- described above (Figure 7), deletion of the PDZ domain rendered eration and desynchronization of release in RIM-deficient neurons RIM1a unable to rescue the decrease in Ca2+-channel levels in (Figures 4L and 4M and Figures 7F–7H). Possibly most importantly, presynaptic terminals (Figures 8A and 8B), suggesting that the RIM1a lacking the PDZ domain did not restore normal Ca2+ influx RIM PDZ domain is essential for tethering of Ca2+ channels to into presynaptic nerve terminals in RIM-deficient neurons as presynaptic terminals. measured by Ca2+ imaging, whereas full-length RIM1a fully rescued Ca2+ influx (Figures 7I and 7J and Figures S7Dand DISCUSSION S7E). Thus, the RIM-PDZ domain is critical for localizing Ca2+ influx to presynaptic active zones, thereby promoting the normal Ca2+ Based on protein/protein interaction studies (Figure 1), genera- dependence, speed, and precision of neurotransmitter release. tion of conditional KO mice (Figure 2), electrophysiological

290 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. recordings (Figure 3, Figure 4, Figure 6, and Figure 7), Ca2+ A RIM1α Zn PDZ C2A PxxP C2B imaging (Figure 5 and Figure 7), and quantitative immunofluores- RIM-ΔPDZ Zn C2A PxxP C2B cence (Figure 8), we here propose that the PDZ domains of RIM 2+ B 2+ proteins stoichiometrically interact with N- and P/Q-type Ca [Ca ]ex (mM): 0.5 125 10 control channels in vitro, that RIM proteins are essential for tethering Ca2+ channels to presynaptic terminals in vivo, and that the cDKO RIM PDZ domain is required for this function. In addition, RIMs cDKO + indirectly interact with Ca2+ channels via their RIM-BP binding RIM-ΔPDZ 1 nA 0.1 s sequence (Hibino et al., 2002), which we show is also essential. C DE 2+ 6 1.0 Thus, RIMs perform two parallel interactions with Ca channels: cDKO + cDKO + 3 *** RIM-ΔPDZ RIM-ΔPDZ 5 a direct interaction via their PDZ domains that is specific for 0.8 *** 2+ (mM) N- and P/Q-type Ca channels and an indirect interaction via

4 ex 2 0.6 ] 2+ 2+ RIM-BPs that is not specific for N- and P/Q-type Ca channels. 3 control control 0.4 2 1 Our results suggest a physiologically validated mechanism

of [Ca 2+

cDKO 50 by which Ca influx is localized to the active zone, as required 1 normalized IPSC 0.2

IPSC amplitude (nA) cDKO EC 0 for fast, synchronous triggering of neurotransmitter release (Fig- 0 0.0 110 110 2+ 2+ ure 8C) (also see Han et al., 2011). [Ca ] (mM) [Ca ] (mM) control cDKO ex ex cDKO + RIM-ΔPDZ

F control GH* 1 nA **** 0.6 5 ms 2 * The RIM PDZ-Domain/Ca2+-Channel Interaction 0.4 cDKO 0.5 nA 1 We found that the C-terminal sequences of N- and P/Q-type

SD of 0.2 2+ 5 ms Ca channels specifically bind to RIM PDZ domains. R-type

cDKO + 0 rise time (ms) 0.0 2+ RIM-ΔPDZ rise time (ms) Ca channels have a similar C-terminal sequence and may 0.5 nA 5 ms 2+ control cDKO control cDKO cDKO + cDKO + also interact, whereas L- and T-type Ca channels do not (Fig- Δ Δ 2+ RIM- PDZ RIM- PDZ ure 1 and Figure S1). The PDZ-domain/Ca -channel interac- I control cDKO cDKO + cDKO + RIM1α RIM-ΔPDZ tion was surprising because the PDZ-domain proteins Mints and CASK were previously shown to bind to Ca2+ channels 20 mV -70 mV 50 ms (Maximov et al., 1999), and because the RIM PDZ domain is known to bind to ELKS proteins (Ohtsuka et al., 2002; Wang et al., 2002). However, it remains unclear whether these previ- Fluo5F Δ 50% G/G0 ously described interactions are physiologically important; in 50 ms fact, Mint- and CASK-deficient synapses exhibit multiple abnor- malities that do not resemble a Ca2+-influx impairment (Atasoy ns J *** bouton dendrite et al., 2007; Ho et al., 2006), whereas ELKS2-deficient synapses 120 80 display increased release at inhibitory synapses (Kaeser et al.,

60 2009), and ELKS levels are not detectably changed in RIM

40 cDKO neurons (data not shown). Thus, it seems unlikely that % change

80 0 20 the effects we observe here are indirectly mediated via Mints, Δ G/G 0 CASK, or ELKS. It is possible, however, that different PDZ- (% change) 0 40 0 200 400 600 domain-binding reactions compete with each other at the active time (ms)

Δ G/G control zone. For example, ELKS binding to RIM PDZ domains may cDKO inhibit Ca2+-channel binding and thereby attenuate neurotrans- cDKO + RIM1α 0 cDKO + RIM-ΔPDZ mitter release; this inhibitory role of ELKS binding could be 0 200 400 600 regulated during plasticity, which might account for the central time (ms) role of RIM in short- and long-term plasticity (Castillo et al., Figure 7. RIM-Dependent Localization of Presynaptic Ca2+ Influx 2002; Fourcaudot et al., 2008; Kaeser et al., 2008; Schoch Requires the RIM PDZ Domain et al., 2002). (A) Domain structures of rescue proteins. 2+ (B–E) Sample traces and quantitative analysis of [Ca ]ex dependence of release of IPSCs in control neurons, cDKO neurons, and cDKO neurons rescued with the PDZ-domain-deficient RIM-DPDZ fragment. Absolute IPSC cDKO neurons rescued with RIM1a or RIM-DPDZ. (Boutons: control, n = 40 2+ amplitudes (C), IPSC amplitudes normalized to the response at 10 mM [Ca ]ex boutons/6 neurons/5 independent cultures; cDKO, n = 52/7/5; cDKO + RIM1a, 2+ (D), and apparent Ca affinities (EC50 values; E) are indicated (control, n = 10/ n = 51/7/5; cDKO + RIM-DPDZ, n = 52/7/5; dendrites: control, n = 22/6/5; 3; cDKO, n = 9/3; cDKO + RIM-DPDZ, n = 10/3). cDKO, n = 22/7/5; cDKO + RIM1a, n = 19/7/5, cDKO + RIM-DPDZ, n = 22/7/5.) (F–H) Speed and synchrony of neurotransmitter release in control neurons, For cumulative peak amplitudes and statistical values, see Figure S7 and cDKO neurons and cDKO neurons rescued with RIM-DPDZ (control, n = 7/3; Table S6. Statistical analyses: *p < 0.05; **p < 0.01; ***p < 0.001; (E, G, and H) cDKO, n = 10/3; cDKO + RIM-DPDZ, n = 12/3). One-way ANOVA; (J) Two-way ANOVA for peak amplitudes during the first (I and J) Sample line scans (I) and summary data (J) of action potential-evoked 60 ms after action potential induction. Ca2+ transients in presynaptic boutons of control neurons, cDKO neurons, and All data shown are means ± SEM (error bars in C–H, shaded areas in J).

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 291 cDKO + cDKO + Figure 8. RIM-Dependent Ca2+-Channel Tethering A control cDKO RIM1α RIM-ΔPDZ Linked to Synaptic Vesicle Docking and Priming (A and B) Immunofluorescent stainings (A) and quantitative immunolocalization analyses (B) of P/Q-type Ca2+ channels Syn (top panel in A) and presynaptic bassoon (bottom panel in A) in control and RIM-deficient cDKO neurons and in cDKO 2.1 V neurons rescued with RIM1a or RIM-DPDZ (means ± SEM,

Ca n=3 cultures per condition, *p < 0.05; **p < 0.01 by Student’s t test compared to control; a second, independent experiment is found in Figure S8 and Table S7).

merge (C) Model of the presynaptic release machinery. The drawing illustrates the structures of major active-zone proteins (RIMs, Munc13s, and RIM-BPs), P/Q- or N-type Ca2+ channels,

Syn a partially assembled SNARE complex (composed of syn- aptobrevin/VAMP on synaptic vesicles and SNAP-25 and syntaxin-1 on the plasma membrane), Munc18-1, complexin,

Bsn and key synaptic vesicle proteins (Rab3 and synaptotagmin-1 [Syt1]). Domain identification is provided on the top right. We propose that RIMs determine the specific localization of P/Q- and N-type Ca2+ channels at the active zone via a direct Ca2+- merge channel/PDZ-domain interaction and via indirect binding of Ca2+ channels to RIMs via RIM-BPs (Hibino et al., 2002). In B control 1.0 1.0 addition, RIMs form an N-terminal priming complex with Rab3 cDKO 0.8 * 0.8 and Munc13, in which Munc13 likely acts by binding to SNARE ** cDKO + RIM1α complexes (not depicted due to restrictions of the two- 0.6 0.6 cDKO + dimensional presentation). Synaptotagmin-1 on the vesicles 0.4 0.4 RIM-ΔPDZ serves as the Ca2+ sensor for exocytosis. With this architec- 2+ 2+

normalized normalized ture, Ca channels and Ca sensors are in close proximity, 0.2 0.2 synaptic levels synaptic levels accounting for the speed, synchrony, and extent of release. 0 0 Ca 2.1 bassoon V

C presynaptic -domain PDZ-domain PxxP motif 2+ x x non-Ca -binding C2-domain 2+

2+ 2+ Ca -binding C -domain Ca Ca 2 synaptic cleft MUN-domain SH3-domain FN3 repeat postsynaptic neuron

synaptic vesicle

Munc13 x Rab3/27

x GTP RIM Ca2+ x Syt1 Ca2+ x Munc18 Ca2+ Ca2+ 2+ Ca 2+ Synaptobrevin 2+ Ca Ca 2+ 2+ Ca 2+ Ca Ca Ca2+ 2+ RIM-BP Ca2+ Ca 2+ 2+ Ca SNAP-25 Ca2+ Ca Ca2+ Ca2+ Complexin Syntaxin Ca2+

2+ 2+ Ca 2+ Ca -channel Ca2+ Ca 2+ Ca2+ Ca 2+ Ca2+ Ca2+ Ca 2+ Ca2+ Ca 2+ 2+ Ca 2+ Ca Ca2+ Ca2+ Ca synaptic cleft 2+ 2+ Ca Ca Ca2+ Ca2+

292 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. The RIM PDZ Domain Is Essential for Localizing Ca2+ formed with the various Ca2+ channel or RIM1 bait vectors and the Ca2+ Channels to the Active Zone channel or RIM1 prey vectors. HSQC spectroscopy was performed with rat Using rescue experiments in RIM-deficient neurons, we found RIM1 (residues 596–704, expressed as described; Lu et al., 2005), and nonla- beled P/Q-type Ca2+-channel peptides. 1H–15N HSQC spectra were acquired that the RIM PDZ domain was invariably required in various RIM in a Varian Inova500 spectrometer at a 40–200 mM protein concentration. rescue constructs to reverse the impairment in presynaptic Ca2+ influx in RIM-deficient neurons (Figure 4, Figure 5, Figure 6, and Generation of Double Conditional RIM KO Mice Figure 7) and for localizing P/Q-type Ca2+ channels to presynaptic The RIM2abg targeting vector was constructed from a l-phage DNA clone iso- boutons (Figure 8). In addition, loss of RIM-BP-binding sequences lated from a genomic library, and conditional RIM2abg KO mice were gener- blocked rescue of Ca2+ influx (Figure 6). These experiments ated by homologous recombination in R1 embryonic stem cells. The recom- bined stem cells were used for blastocyst injections to obtain chimeric mice. suggest that RIMs tether Ca2+ channels to the active zone via 2+ After germline transmission of the mutant allele, the newly generated condi- two parallel interactions: directly by binding to Ca channels tional RIM2abg KO mice were crossed to conditional RIM1ab KO mice (Kaeser via their PDZ domains and indirectly by binding to RIM-BPs, which et al., 2008). in turn bind to Ca2+ channels (see model in Figure 8C and recon- stitution of the tripartite complex in Figure S8C). Our data also Electrophysiology account for the specificity of N- and P/Q-type Ca2+ channels in Whole-cell patch-clamp recordings were performed in cultured hippocampal release, as the RIM PDZ domain/Ca2+-channel interaction is neurons at DIV13–15. Synaptic responses were elicited by a local stimulation electrode and were acquired with a multiclamp 700B amplifier. The extracel- specific for these Ca2+-channel types (Figure 1). lular solution contained (in mM) 140 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2,10 HEPES-NaOH (pH 7.3), and 10 glucose, with 315 mOsm, and either 50 mM Active-Zone Functions of RIMs picrotoxin (excitatory postsynaptic currents, EPSCs) or 10 mM CNQX and By showing that, besides their role in vesicle docking and 50 mM D-APV (IPSCs). For all electrophysiological experiments, the experi- priming, RIMs are essential for tethering Ca2+ channels to active menter was blind to the condition/genotype of the cultures analyzed. zones, our findings corroborate the notion that RIM proteins are Ca2+ Imaging central organizers of active zones (Betz et al., 2001; Gracheva All experiments were performed in a Zeiss LSM 510 confocal microscope. et al., 2008; Kaeser et al., 2008; Koushika et al., 2001; Schoch Cultured hippocampal neurons were examined at DIV14–18 in whole-cell et al., 2002, 2006)(Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, patch-clamp configuration after filling with Fluo5F and Alexa 594 dyes for and Figure 8) (see also Deng et al., 2011 and Han et al., 2011). 10 min. Action potentials were induced by short, somatic current injections 2+ Moreover, RIM proteins perform additional functions, as indi- through the patch pipette (typically 5 ms, 600 pA); Ca transients were measured with line scans through presynaptic boutons and second-order cated by the fact that although the RIM C2 domains had no 2+ dendrites at a frequency of 333 Hz, typically 100–150 mm away from the detectable role in Ca influx, they boosted neurotransmitter neuronal cell body. Fluorescent signals were quantified as mean region of 2+ release, possibly by binding to a-liprins, the b4Ca channel interest and plotted as GG0/G0 (G = average green emission in a given subunit, or other proteins (Kiyonaka et al., 2007; Schoch et al., line; G0 = average of 20 line scans before action potential induction). The 2002). Moreover, in a parallel study we found that RIM proteins experimenter was blind to the condition/genotype until all recordings and anal- activate priming by binding to Munc13 proteins, thereby disrupt- yses were completed. ing their homodimerization and reversing the autoinhibitory Immunofluorescence Staining of Cultured Neurons effects of the Munc13 homodimerization (Deng et al., 2011). Cultured neurons were fixed in 4% paraformaldehyde, permeabilized in 0.1% Thus, RIMs occupy the center of an interaction network in the Triton X-100/3% bovine serum albumin, and incubated overnight with anti- molecular anatomy of the active zone and influence all aspects CaV2.1 rabbit polyclonal antibodies (Alomone labs, 1:100) or anti-bassoon of neurotransmitter release (Figure 8C). However, our data also rabbit polyclonal antibodies (Synaptic Systems, 1:250) and anti-synapsin mouse monoclonal antibodies (Synaptic Systems, 1:1000). Alexa Fluor 546 raise new questions. How do RIM C2 domains boost release without altering Ca2+ influx? Why does deletion of just one RIM anti-mouse and Alexa Fluor 633 anti-rabbit secondary antibodies were used for detection with a confocal microscope. Single sections were acquired isoform, RIM1a, which has only a partial effect on neurotrans- with identical settings applied to all samples in an experiment and were mitter release due to its redundancy with other RIM isoforms, used to quantify levels of P/Q-type Ca2+ channels and bassoon in ImageJ block multiple forms of presynaptic long-term synaptic plasticity (NIH); the data were normalized to synapsin staining and expressed relative (Castillo et al., 2002; Fourcaudot et al., 2008; Kaeser et al., to control cultures. The experimenter was blind to the condition/genotype in 2008)? Is the RIM/Ca2+-channel interaction a mechanism by all experiments. which synaptic strength can be regulated? With the availability of double conditional KO mice described here, these questions Data Analysis All data are shown as means ± standard errors of the mean (SEM). Statistical can now be addressed. significance was determined by one-way ANOVA (some electrophysiological recordings), two-way ANOVA (Ca2+-imaging peak amplitude), Kolmogorov- EXPERIMENTAL PROCEDURES Smirnov test (cumulative distribution of peak amplitudes in Ca2+ imaging), c-test (mouse survival analysis), or Student’s t test (all other experiments). 2+ In Vitro Protein-Binding Assays Ninety-five percent confidence intervals for [Ca ]ex-titration data and fitting Two yeast two-hybrid screens of a rat brain cDNA library using bait vectors en- parameters were calculated based on the covariance matrix. All numerical 2+ coding the CaV2.2 N-type Ca channel C terminus (residues 2163–2339) or and statistical values and the tests used can be found in Table S1, Table S2, 2+ CaV2.1 P/Q-type Ca channel C terminus (residues 2213–2368) and liquid Table S3, Table S4, Table S5, Table S6, and Table S7. yeast two-hybrid assays were performed as described (Wang et al., 1997). Of the 134/84 isolates with the N-type/P/Q-type bait, 8/16 clones corre- Miscellaneous sponded to RIM-BP1, 8/17 to RIM-BP2, and 3/2 to RIM1. For mapping of Mixed hippocampal cultures and lentiviruses generated in transfected the interaction region and the relative strength, yeast strain L40 was cotrans- HEK293T cells expressing EGFP-tagged active or inactive cre-recombinases

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 293 followed by an IRES sequence for expression of rescue constructs were Dulubova, I., Lou, X., Lu, J., Huryeva, I., Alam, A., Schneggenburger, R., Sud- produced as described (Kaeser et al., 2009). SDS-PAGE gels, immunoblotting, hof, T.C., and Rizo, J. (2005). A Munc13/RIM/Rab3 tripartite complex: from and electronmicroscopic analyses were done according to standard methods priming to plasticity? Embo J. 24, 2839–2850. described in the Extended Experimental Procedures. All experiments Fourcaudot, E., Gambino, F., Humeau, Y., Casassus, G., Shaban, H., Poulain, were performed according to institutional guidelines. A detailed methods B., and Luthi, A. (2008). cAMP/PKA signaling and RIM1alpha mediate presyn- section can be found in the Supplemental Information. aptic LTP in the lateral amygdala. Proc. Natl. Acad. Sci. USA 105, 15130– 15135. SUPPLEMENTAL INFORMATION Gracheva, E.O., Hadwiger, G., Nonet, M.L., and Richmond, J.E. (2008). Direct interactions between C. elegans RAB-3 and Rim provide a mechanism to Supplemental Information includes Extended Experimental Procedures, eight target vesicles to the presynaptic density. Neurosci. Lett. 444, 137–142. figures, and seven tables and can be found with this article online at doi:10. Han, Y., Kaeser, P.S., Su¨ dhof, T.C., and Schneggenburger, R. (2011). 1016/j.cell.2010.12.029. RIM determines Ca2+ channel density and vesicle docking at the presyn- aptic active zone. Neuron 69, in press. Published online January 26, ACKNOWLEDGMENTS 2011. 10.1016/j.neuron.2010.12.014. Hibino, H., Pironkova, R., Onwumere, O., Vologodskaia, M., Hudspeth, A.J., We thank E. Borowicz, I. Kornblum, L. Fan, J. Mitchell, H. Ly, and I. Huryeva for and Lesage, F. (2002). RIM binding proteins (RBPs) couple Rab3-interacting technical assistance, Dr. R.E. Hammer for blastocyst injections, Dr. N. Brose molecules (RIMs) to voltage-gated Ca(2+) channels. Neuron 34, 411–423. 2+ for Munc13-1 antibodies, Dr. C. Acuna-Goycolea for advice on Ca -imaging Ho, A., Morishita, W., Atasoy, D., Liu, X., Tabuchi, K., Hammer, R.E., Malenka, experiments, Dr. Z. Ma for assistance with yeast two-hybrid screening, Drs. Z. R.C., and Sudhof, T.C. (2006). Genetic analysis of Mint/X11 proteins: essential Pang, T. Bacaj, and C. Fo¨ ldy for help with data analysis, and Dr. R. Schneggen- presynaptic functions of a neuronal adaptor protein family. J. Neurosci. 26, burger for comments. This work was supported by grants from the NIH (NINDS 13089–13101. 33564 to T.C.S., NS37200 to J.R., DA029044 to P.S.K.), a Swiss National Kaeser, P.S., Deng, L., Chavez, A.E., Liu, X., Castillo, P.E., and Sudhof, T.C. Science Foundation Postdoctoral Fellowship (to P.S.K.), and a NARSAD (2009). ELKS2alpha/CAST deletion selectively increases neurotransmitter Young Investigator Award (to P.S.K.). release at inhibitory synapses. Neuron 64, 227–239.

Received: April 1, 2010 Kaeser, P.S., Kwon, H.B., Chiu, C.Q., Deng, L., Castillo, P.E., and Sudhof, T.C. Revised: September 2, 2010 (2008). RIM1alpha and RIM1beta are synthesized from distinct promoters of Accepted: November 15, 2010 the RIM1 gene to mediate differential but overlapping synaptic functions. Published: January 20, 2011 J. Neurosci. 28, 13435–13447. Katz, B. (1969). The Release of Neural Transmitter Substances (Liverpool: REFERENCES Liverpool Univ. Press). Kiyonaka, S., Wakamori, M., Miki, T., Uriu, Y., Nonaka, M., Bito, H., Beedle, Atasoy, D., Schoch, S., Ho, A., Nadasy, K.A., Liu, X., Zhang, W., Mukherjee, K., A.M., Mori, E., Hara, Y., De Waard, M., et al. (2007). RIM1 confers sustained Nosyreva, E.D., Fernandez-Chacon, R., Missler, M., et al. (2007). Deletion of activity and neurotransmitter vesicle anchoring to presynaptic Ca2+ channels. CASK in mice is lethal and impairs synaptic function. Proc. Natl. Acad. Sci. Nat. Neurosci. 10, 691–701. USA 104, 2525–2530. Koushika, S.P., Richmond, J.E., Hadwiger, G., Weimer, R.M., Jorgensen, Betz, A., Thakur, P., Junge, H.J., Ashery, U., Rhee, J.S., Scheuss, V., Rosen- E.M., and Nonet, M.L. (2001). A post-docking role for active zone protein mund, C., Rettig, J., and Brose, N. (2001). Functional interaction of the active Rim. Nat. Neurosci. 4, 997–1005. zone proteins Munc13-1 and RIM1 in synaptic vesicle priming. Neuron 30, Llinas, R., Sugimori, M., and Silver, R.B. (1992). Microdomains of high 183–196. concentration in a presynaptic terminal. Science 256, 677–679. Castillo, P.E., Weisskopf, M.G., and Nicoll, R.A. (1994). The role of Ca2+ chan- Lu, J., Li, H., Wang, Y., Sudhof, T.C., and Rizo, J. (2005). Solution structure of nels in hippocampal mossy fiber synaptic transmission and long-term poten- the RIM1alpha PDZ domain in complex with an ELKS1b C-terminal peptide. tiation. Neuron 12, 261–269. J. Mol. Biol. 352, 455–466. Castillo, P.E., Schoch, S., Schmitz, F., Sudhof, T.C., and Malenka, R.C. (2002). Luebke, J.I., Dunlap, K., and Turner, T.J. (1993). Multiple calcium channel RIM1alpha is required for presynaptic long-term potentiation. Nature 415, types control glutamatergic synaptic transmission in the hippocampus. 327–330. Neuron 11, 895–902. Catterall, W.A., Perez-Reyes, E., Snutch, T.P., and Striessnig, J. (2005). Inter- Maximov, A., Sudhof, T.C., and Bezprozvanny, I. (1999). Association of national Union of Pharmacology. XLVIII. Nomenclature and structure-function neuronal calcium channels with modular adaptor proteins. J. Biol. Chem. relationships of voltage-gated calcium channels. Pharmacol. Rev. 57, 274, 24453–24456. 411–425. Meinrenken, C.J., Borst, J.G., and Sakmann, B. (2002). Calcium secretion Church, P.J., and Stanley, E.F. (1996). Single L-type calcium channel conduc- coupling at calyx of held governed by nonuniform channel-vesicle topography. tance with physiological levels of calcium in chick ciliary ganglion neurons. J. Neurosci. 22, 1648–1667. J. Physiol. 496, 59–68. Missler, M., Zhang, W., Rohlmann, A., Kattenstroth, G., Hammer, R.E., Gott- Coppola, T., Magnin-Luthi, S., Perret-Menoud, V., Gattesco, S., Schiavo, G., mann, K., and Sudhof, T.C. (2003). Alpha-neurexins couple Ca2+ channels and Regazzi, R. (2001). Direct interaction of the Rab3 effector RIM with to synaptic vesicle exocytosis. Nature 424, 939–948. Ca2+ channels, SNAP-25, and synaptotagmin. J. Biol. Chem. 276, 32756– Mittelstaedt, T., Alvarez-Baron, E., and Schoch, S. (2010). RIM proteins and 32762. their role in synapse function. Biol. Chem. 391, 599–606. Deng, L., Kaeser, P.S., Xu, W., and Su¨ dhof, T.C. (2011). RIM proteins activate Ohtsuka, T., Takao-Rikitsu, E., Inoue, E., Inoue, M., Takeuchi, M., Matsubara, vesicle priming by reversing autoinhibitory homodimerization of Munc13. K., Deguchi-Tawarada, M., Satoh, K., Morimoto, K., Nakanishi, H., et al. (2002). Neuron 69, in press. Published online January 26, 2011. 10.1016/j.neuron. Cast: a novel protein of the cytomatrix at the active zone of synapses that 2011.01.005. forms a ternary complex with RIM1 and munc13-1. J. Cell Biol. 158, 577–590. Dietrich, D., Kirschstein, T., Kukley, M., Pereverzev, A., von der Brelie, C., Poncer, J.C., McKinney, R.A., Gahwiler, B.H., and Thompson, S.M. (1997). Schneider, T., and Beck, H. (2003). Functional specialization of presynaptic Either N- or P-type calcium channels mediate GABA release at distinct hippo- Cav2.3 Ca2+ channels. Neuron 39, 483–496. campal inhibitory synapses. Neuron 18, 463–472.

294 Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. Qian, J., and Noebels, J.L. (2000). Presynaptic Ca(2+) influx at a mouse central Wang, Y., Okamoto, M., Schmitz, F., Hofmann, K., and Sudhof, T.C. (1997). synapse with Ca(2+) channel subunit mutations. J. Neurosci. 20, 163–170. Rim is a putative Rab3 effector in regulating synaptic-vesicle fusion. Nature Regehr, W.G., and Mintz, I.M. (1994). Participation of multiple calcium channel 388, 593–598. types in transmission at single climbing fiber to Purkinje cell synapses. Neuron Wang, Y., Sugita, S., and Sudhof, T.C. (2000). The RIM/NIM family of neuronal 12, 605–613. proteins. Interactions with Rab3 and a new class of Src homology 3 Schneggenburger, R., Meyer, A.C., and Neher, E. (1999). Released fraction domain proteins. J. Biol. Chem. 275, 20033–20044. and total size of a pool of immediately available transmitter quanta at a calyx synapse. Neuron 23, 399–409. Wang, Y., Liu, X., Biederer, T., and Sudhof, T.C. (2002). A family of RIM-binding Schoch, S., Castillo, P.E., Jo, T., Mukherjee, K., Geppert, M., Wang, Y., proteins regulated by alternative splicing: Implications for the genesis of Schmitz, F., Malenka, R.C., and Sudhof, T.C. (2002). RIM1alpha forms a protein synaptic active zones. Proc. Natl. Acad. Sci. USA 99, 14464–14469. scaffold for regulating neurotransmitter release at the active zone. Nature 415, Wiedemann, U., Boisguerin, P., Leben, R., Leitner, D., Krause, G., Moelling, K., 321–326. Volkmer-Engert, R., and Oschkinat, H. (2004). Quantification of PDZ domain Schoch, S., Mittelstaedt, T., Kaeser, P.S., Padgett, D., Feldmann, N., Cheva- specificity, prediction of ligand affinity and rational design of super-binding leyre, V., Castillo, P.E., Hammer, R.E., Han, W., Schmitz, F., et al. (2006). peptides. J. Mol. Biol. 343, 703–718. Redundant functions of RIM1alpha and RIM2alpha in Ca(2+)-triggered neuro- transmitter release. EMBO J. 25, 5852–5863. Wojcik, S.M., and Brose, N. (2007). Regulation of membrane fusion in synaptic excitation-secretion coupling: Speed and accuracy matter. Neuron 55, 11–24. Shin, O.H., Xu, J., Rizo, J., and Sudhof, T.C. (2009). Differential but convergent functions of Ca2+ binding to synaptotagmin-1 C2 domains mediate neuro- Wu, L.G., Westenbroek, R.E., Borst, J.G., Catterall, W.A., and Sakmann, B. transmitter release. Proc. Natl. Acad. Sci. USA 106, 16469–16474. (1999). Calcium channel types with distinct presynaptic localization couple Takahashi, T., and Momiyama, A. (1993). Different types of calcium channels differentially to transmitter release in single calyx-type synapses. J. Neurosci. mediate central synaptic transmission. Nature 366, 156–158. 19, 726–736.

Cell 144, 282–295, January 21, 2011 ª2011 Elsevier Inc. 295