Synergistic roles of Synaptotagmin-1 and complexin in calcium-regulated neuronal exocytosis

Calcium (Ca2+)-evoked release of neurotransmitters from synaptic vesicles requires mechanisms both to prevent un-initiated fusion of vesicles (clamping) and to trigger fusion following Ca2+-influx. The principal components involved in these processes are the vesicular fusion machinery (SNARE proteins) and the regulatory proteins, Synaptotagmin-1 and Complexin. Here, we use a reconstituted single-vesicle fusion assay under physiologically-relevant conditions to delineate a novel mechanism by which Synaptotagmin-1 and Complexin act synergistically to establish Ca2+-regulated fusion. We find that under each vesicle, Synaptotagmin-1 oligomers bind and clamp a limited number of ‘central’ SNARE complexes via the primary interface and introduce a kinetic delay in vesicle fusion mediated by the excess of free SNAREpins. This in turn enables Complexin to arrest the remaining free ‘peripheral’ SNAREpins to produce a stably clamped vesicle. Activation of the central SNAREpins associated with Synaptotagmin-1 by Ca2+ is sufficient to trigger rapid (<100 msec) and synchronous fusion of the docked vesicles.


Introduction
The controlled yet rapid (sub-millisecond) release of neurotransmitters stored in synaptic vesicles (SVs) is central to all information processing in the brain (Südhof, 2013;Kaeser and Regehr, 2014;Rizo, 2018). Synaptic release of neurotransmitters relies on the efficient coupling of SV fusion to the triggering signal -action potential (AP)-evoked Ca 2+ influx into the pre-synaptic terminal (Kaeser and Regehr, 2014;Südhof, 2013). SV fusion is catalyzed by synaptic SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins, VAMP2 on the vesicles (v-SNAREs) and Syntaxin/SNAP25 (t-SNAREs) on the pre-synaptic membrane (Sö llner et al., 1993;Weber et al., 1998). On their own, SNARE proteins are constitutively active and intrinsically trigger fusion in the range of 0.5-1 s Ramakrishnan et al., 2019;Xu et al., 2016).
To achieve the requisite speed and precision of synaptic transmission, the nerve terminals maintain a pool of docked vesicles that can be readily released upon Ca 2+ influx (Südhof, 2013;Kaeser and Regehr, 2014;Südhof and Rothman, 2009). The prevailing theory is that in a 'releaseready' vesicle, multiple SNARE complexes are firmly held ('clamped') in a partially assembled state. These 'SNAREpins' are then synchronously released by Ca 2+ to drive fusion dramatically faster than any one SNARE alone (Rothman et al., 2017;Brunger et al., 2018;Volynski and Krishnakumar, 2018). Several lines of evidence suggest that two synaptic proteins, Synaptotagmin-1 (Syt1) and Complexin (Cpx) play a critical role in establishing Ca 2+ -regulated neurotransmitter release (Geppert et al., 1994;Xu et al., 2007;Bacaj et al., 2013;Yang et al., 2013;Huntwork and Littleton, 2007).
Synaptotagmin-1 is a SV-localized protein with a large cytoplasmic part containing tandem C2A and C2B domains that bind Ca 2+ (Fuson et al., 2007;Sutton et al., 1995). It is well established that fast AP-evoked synchronous release is triggered by Ca 2+ binding to Syt1 C2 domains (Brose et al., 1992;Geppert et al., 1994;Littleton et al., 1993). Genetic analysis shows that Syt1 also plays a crucial role in 'clamping' spontaneous release and Ca 2+ -evoked asynchronous release to ensure high fidelity of the Ca 2+ -coupled synchronous transmitter release (Xu et al., 2009;Bacaj et al., 2013;Littleton et al., 1993). Recently it has been demonstrated that C2B-driven self-oligomerization of Syt1 provides the structural basis for its clamping function (Wang et al., 2014;Bello et al., 2018;Tagliatti et al., 2019). Syt1 is also involved in initial stages of docking of SVs to the plasma membrane (PM), in part mediated by its interaction with the anionic lipids, phosphatidylserine (PS) and phosphatidylinositol 4, 5-bisphosphate (PIP2) on the PM (Honigmann et al., 2013;Parisotto et al., 2012;Park et al., 2012). Besides the membrane interaction, Syt1 also binds the neuronal t-SNAREs on the PM, both independently (primary binding site) and in conjunction with Complexin (tripartite binding site) (Zhou et al., 2015;Zhou et al., 2017;Grushin et al., 2019). The Syt1-SNARE interactions are important for Syt1 role in SV docking, clamping fusion and triggering Ca 2+ -dependent neurotransmitter release (Zhou et al., 2015;Zhou et al., 2017).
It is broadly accepted that Syt1 and Cpx are both involved in clamping spurious or delayed fusion events and synchronize neurotransmitter release to Ca 2+ -influx. However, there is presently no coherent understanding how these proteins assemble and operate together. To gain mechanistic insights into this process, a reductionist approach where the variables are limited, and components can be rigorously controlled is required. We recently described a biochemically-defined fusion system based on a pore-spanning lipid bilayer setup that is ideally suited for this purpose . This reconstituted setup is capable of precisely tracking individual vesicle docking, clamping (delay from docking to spontaneous fusion) and Ca 2+ triggered release at tens of milliseconds timescale Ramakrishnan et al., 2019). Critically, this setup allows us to examine these discrete sub-steps in the vesicular exocytosis process, independent of alterations in the preceding or following stages Ramakrishnan et al., 2019;Ramakrishnan et al., 2018).
Using the in vitro fusion system, we recently reported that under artificially low-VAMP2 conditions (~13 copies of VAMP2 and~25 copies of Syt1 per vesicle), Syt1 oligomerization is both necessary and sufficient to establish a Ca 2+ -sensitive fusion clamp . Here we extend this study to more physiologically-relevant conditions, using SV mimics reconstituted with~25 copies of Syt1 and~70 copies of VAMP2. We report that under these conditions, both Cpx and Syt1 oligomers are needed to stably clamp all SNARE complexes and the reversal of the Syt1 clamp is sufficient to achieve fast, Ca 2+ -triggered synchronized fusion.

Results
Synaptotagmin and complexin co-operate to clamp vesicle fusion With the goal of approximating the physiological context, we chose a reconstitution condition for small unilamellar vesicles (SUV) resulting in an average of 74 copies and 25 copies of outward-facing VAMP2 and Syt1, respectively ( Figure 1-figure supplement 1). We employed pre-formed t-SNAREs (1:1 complex of Syntaxin1 and SNAP-25) in the planar bilayers (containing 15% PS and 3% PIP2) to both simplify the experimental approach and to bypass the requirement of SNARE-assembling chaperones, Munc18 and Munc13 (Baker and Hughson, 2016;Rizo, 2018). In most of the experiments, we used fluorescently-labelled lipid (2% ATTO647-PE) included in the SUVs to track the docking, diffusion and fusion of individual SUVs ( Figure 1A). Figure 1. Syt1 and Cpx co-operatively clamp SNARE-mediated vesicle fusion under physiologically-relevant conditions. The effect of Syt1 and Cpx on SNARE-driven fusion was assessed using a single-vesicle docking and fusion analysis with a pore-spanning bilayer setup Ramakrishnan et al., 2018). (A) Representative fluorescence (ATTO647N-PE) traces showing the behavior of small unilamellar vesicles (SUV) containing VAMP2 (vSUV) or Syt1 and VAMP2 (Syt1-vSUV) on t-SNARE containing bilayer in the presence or the absence of Cpx. (B) The time between docking and fusion was measured for each docked vesicle and the results for the whole population are presented as a survival curve. vSUVs (black curve) are diffusively mobile upon docking (Figure 1-figure supplement 2) and fuse spontaneous with a half-time of~1 s. Addition of soluble Cpx (2 mM) does not change this behavior (green curve). Inclusion of Syt1 in the v-SUV (red curve) does not block fusion but increases the time from docking-to-fusion (~5 s half-life), in effect delaying the kinetics of fusion. When included together Syt1 and Cpx (blue curves) fully arrest fusion to produce stably docked SUVs that attach and remain in place during the entire period of observation. (C) Syt1 and Cpx, both individually and collectively increase the number of docked vesicles. In all cases, a mutant form of VAMP2 (VAMP2 4X ) which eliminated fusion was used to unambiguously estimate the number of docked vesicles after the 10 min interaction phase. The average values and standard deviations from three to four independent experiments are shown for each condition. In sum, 500-1000 vesicles were analyzed for each condition. The online version of this article includes the following figure supplement(s) for figure 1:   We initially focused on the kinetics of constitutive fusion to assess the ability of Syt1 and Cpx to 'clamp' SNARE-driven fusion in the absence of Ca 2+ . We monitored large ensembles of vesicles to determine the percent remaining unfused as a function of time elapsed after docking and quantified as 'survival percentages' ( Figure 1B). Docked immobile vesicles that remained un-fused during the initial 10 min observation period were defined as 'clamped' . Vesicles containing VAMP2 only (vSUV) that docked to the t-SNARE containing bilayer surface were mobile and a majority (>95%) spontaneously fused typically with a t 1/2~1 sec post-docking ( Figure 1B Table 1, Table 2, Video 1).
Inclusion of wild-type Syt1 in the vesicles (Syt1-vSUVs) enhanced the vesicle docking rate, with~8 fold increase in total number of docked vesicles ( Figure 1C). The majority (~80%) of docked Syt1-vSUVs remained mobile on the bilayer surface and fused on an average~5-6 s after docking ( Figure 1B Table 1, Table 2, Video 2). The remaining small fraction (~20%) were immobile and stably clamped. This is in stark contrast to our earlier finding under low-copy VAMP2 conditions wherein the bulk of the Syt1-vSUVs (>90%) were stably clamped . The pronounced docking-to-fusion delay introduced by Syt1 (t 1/2~5 sec for Syt1-vSUV compared to~1 s for vSUV) suggests that under physiologically-relevant ('normal' VAMP copy number) conditions, Syt1 alone can meaningfully delay but not stably clamp fusion.
This unstable Syt1 clamp was stabilized by addition of Cpx ( Figure Table 1, Table 2, Video 3). In fact, these vesicles remained stably docked up to 1 hr without fusing (Figure 1-figure supplement 3). Furthermore, Syt1 and Cpx together significantly increased (~18 fold) the total number of docked vesicles ( Figure 1C). Thus, we find that Syt1 and Cpx act synergistically to increase the rate of vesicle docking and serve to effectively block fusion of the docked vesicles under resting conditions. Addition of soluble Cpx (2 mM) alone produced a~10 fold increase in the number of docked vesicles ( Figure 1C) but did not change the behavior of the docked vSUVs ( Figure 1B, Figure 1- Table 1. Survival probabilities at specific points in time post-docking (Kaplan Meier estimators) for the VAMP2-containing vesicles (vSUV) in the presence of Cpx, Syt1 or both. The corresponding survival curves are shown in Figure 1B Time ( Table 2. The survival curves in Figure 1B were compared pair-wise using the log-rank test to determine the statistical significance (pvalues) of the observed effects of Syt1 and Cpx on v-SUV fusion In the presence of Cpx alone, virtually all docked vSUVs fused spontaneously typically within 1-2 s ( Figure 1A,B). This meant that Syt1 somehow synergizes with Cpx to form the overall fusion clamp.
Synaptotagmin and complexin establish fast, Ca 2+ -triggered vesicle fusion We then investigated the effect of Ca 2+ on the stably Syt1/Cpx-clamped vesicles ( Figure 2). We estimated the time of arrival of Ca 2+ at/near the docked vesicles using a lipid-conjugated Ca 2+ indicator (Calcium green C24) attached to the planar bilayer (Figure 2-figure supplement 1). Influx of free Ca 2+ (1 mM) triggered simultaneous fusion of >90% of the docked vesicles ( Figure 2B, Video 5). These vesicles fused rapidly and synchronously, with a characteristic time-constant (t) of~110 msec following the arrival of Ca 2+ locally ( Figure 2C). This estimate is constrained by the temporal resolution limit (150 msec per frame) of our imaging experiment. Indeed, most of Ca 2+ -triggered fusion occurs within a single frame ( Figure 2C). We thus suspect that the true Ca 2+ -driven fusion rate is <100 msec. Notably, the Syt1/Cpx clamped vesicles remained Ca 2+ -sensitive even 1 hr post-docking (Figure 1-figure supplement 3). Our data indicate that Syt1 and Cpx acting together synchronize vesicle fusion to Ca 2+ influx and greatly accelerate the underlying SNARE-mediated fusion process (which typically occurs at a rate of~1 s). We also tested and confirmed these findings with a content-release assay using sulforhodamine B-loaded SUVs (Figure 2-figure supplement 2) under similar experimental conditions. Overall, we find that Syt1 and Cpx act co-operatively to clamp the SNARE assembly process to generate and maintain a pool of docked vesicles that can be triggered to fuse rapidly and synchronously upon Ca 2+ influx.

Synaptotagmin and complexin clamp different sets of SNARE complexes
We next examined if Syt1 and Cpx act on the same SNARE complexes sequentially or if they function separately to produce molecularly-distinct clamped SNAREpins under the same docked vesicles. To this end, we employed an accessibility-dependent competition assay ( Figure 3). We washed Cpx out from stably clamped Syt1/Cpx vesicles (by dilution) in the absence or the presence of excess inhibitory soluble cytoplasmic domain of the t-SNARE complex (CDT). Cpx binds to half-zippered (clamped) SNAREpins with a K d~0 .5 mM Kümmel et al., 2011) and is therefore expected to freely dissociate when the bulk concentration of CPX is reduced well below that level. CDT will bind and sequester/inactivate any free VAMP2 on the vesicles and is also expected to effectively compete out bilayer-anchored t-SNAREs. This is because CDT can form fullyzippered SNARE complexes (stabilized by~70 k B T) as compared to the half-zippered SNAREpins (~35 k B T) formed by bilayer-anchored t-SNAREs (Li et al., 2016). We reasoned that if Syt1 and Cpx act on the same SNARE complex, then CDT treatment (with Cpx wash-out) would irreversibly block all vesicle fusion. However, if Syt1 and Cpx clamp different SNARE complexes, then some SNAREpins might be sequestered/protected from CDT by Syt1, thereby keeping the vesicles clamped yet sensitive to Ca 2+ influx.
We used fluorescently-labeled Cpx to test and confirm the near-complete washout of Cpx from the clamped Syt1/Cpx vesicles following the extensive (40X) buffer wash (Figure 3-figure supplement 1). Without CDT, the docked vesicles proceeded to fuse spontaneously following the buffer wash ( Figure 3A, Figure 3-figure supplement 2). This further confirmed that both Syt1 and Cpx are needed to produce a stably clamped state. In the presence of CDT, most of the vesicles remained docked even after the removal of previously-bound Cpx by the buffer wash ( Figure 3B, Video 1. Representative video showing the docking, diffusion and spontaneous fusion of a vSUV to freestanding t-SNARE containing lipid bilayer containing t-SNARE complex. The movie was acquired using a Leica confocal scanning microscope at a speed of 7 frames/sec and the ATTO647N-PE fluorescence was used to track the fate of the vesicle. For clarity, a single hole corresponding to 5 mm suspended bilayer marked by the white circle is shown. https://elifesciences.org/articles/54506#video1  Figure 2C). This implied that there are at least two types of clamped SNAREpins under a docked vesicle -those clamped by Syt1 (which are shielded from CDT) and others arrested by Cpx (which become accessible to CDT following the buffer wash-out). It further indicated that even though both Syt1 and Cpx are required to produce a stably 'clamped' vesicle, the activation of the Syt1-associated SNAREpins is sufficient to elicit rapid, Ca 2+ -synchronized vesicular fusion.

Molecular mechanism of synaptotagmin clamp and Ca 2+ activation of fusion
Considering these findings, we sought to establish the molecular determinants of the Syt1 clamp and its reversal by Ca 2+ . To focus on the Syt1 component of the clamp, we tested the effect of specific Syt1 mutations using low copy VAMP2 conditions, i.e. SUVs containing~13 copies of VAMP2 and~25 copies of Syt1 (wild type or mutants) in the absence of Cpx. Consistent with our earlier report , wild-type Syt1 (Syt1 WT ) alone was sufficient to produce stablyclamped vesicles under these conditions ( Figure 4A, Table 3, Table 4). Selective disruption of Syt1-SNARE 'primary' binding using the previously described (Zhou et al., 2015;Zhou et al., 2017) mutations in Syt1 C2B domain (R281A/E295A/Y338W/R398A/R399A; Syt1 Q ) and t-SNARE SNAP25 (K40A/D51A/E52A/E55A/D166A; SNARE Q ) abolished the Syt1 clamp ( Figure 4A, Table 3, Table 4,), with >99% of the docked Syt1 Q vesicles fusing constitutively in the 10 min observation period (Note: From this point onwards, the 'primary' site mutation is simply referred as Syt1 Q ). On the other hand, Syt1 mutations (L387Q/L394Q; Syt1 LLQQ ) that disrupt the hydrophobic interaction that is an integral part of the SNARE-Cpx-Syt1 'tripartite' interface (Zhou et al., 2017) had no effect on the Syt1 clamp ( Figure 4A, Table 3, Table 4). Destabilization of the Syt1 C2B oligomers with a point-mutation (F349A) (Bello et al., 2018) also abrogated the Syt1 clamp, wherein~85% of docked Syt1 F349A vesicles proceeded to fuse spontaneously ( Figure 4A, Table 3, Table 4). However, disrupting Ca 2+ binding to the C2B domain (D309A, D363A, D365A; Syt1 DA ) (Shao et al., 1996) had no effect on Syt1 clamping function, with all vesicles remaining un-fused ( Figure 4A, Table 3, Table 4). This meant that Syt1 ability to oligomerize and bind SNAREpins via the primary binding site is key to its clamping function.
Addition of Ca 2+ (1 mM) triggered rapid and synchronous fusion of all of the docked Syt1 WT and Syt1 LLQQ vesicles and the remaining minority fraction (~15%) of the 'clamped' Syt1 F349A containing vesicles ( Figure 4B). In fact, the docked Syt1 F349A vesicles were indistinguishable in their behavior from Syt1 WT vesicles, suggesting the Syt1 oligomerization is not critical for the Ca 2+ -activation mechanism. In contrast, docked vesicles containing the Ca 2+ -binding mutant (Syt1 DA ) never fused even after Ca 2+ -addition ( Figure 4B). Similarly, Ca 2+ - influx failed to trigger the fusion of the residual (~1%) Syt1 Q vesicles. However, the relatively small number of docked Syt1 Q prior to Ca 2+ -influx precludes any meaningful quantitative analysis. Nonetheless, our data suggest that Ca 2+ -binding to the Syt1 C2B domain and its simultaneous interaction with the t-SNARE protein via the primary binding site is required for Ca 2+ -triggered reversal of the fusion clamp. We also tested the effect of the Syt1 mutants using vesicles containing physiological VAMP2 and Syt1 copy numbers in the presence of 2 mM Cpx ( Figure 5A, Table 5, Table 6). For all mutations To visualize this, fluorescence was simultaneously monitored in a circular region of interest (ROI) encompassing the docked vesicle (vesicle ROI, green and blue circles) and in a surrounding annular ROI (outer ROI, yellow circle). Corresponding to actual fusion events, we observed a sudden decrease of fluorescence intensity in the vesicle ROI with a concomitant increase of fluorescence in the annular outer ROI. Note that the two docked vesicles fuse synchronously in response to Ca 2+ -influx. (B) End-point analysis at 1 min post Ca 2+ -addition shows that >90% of all clamped vesicles fuse following Ca 2+ addition. (C) Kinetic analysis shows that the Syt1/Cpx clamped vesicles fuse rapidly following Ca 2+ -addition with a characteristic time constant of 0.11 s. This represents the temporal resolution limit of our recordings and the true Ca 2+ -triggered fusion rate is likely well below 0.1 s. Thus, Syt1 and Cpx synchronize vesicle fusion to Ca 2+ -influx and accelerate the underlying fusion process. The average values and standard deviations from three independent experiments (with~1000 vesicles in total) is shown. The online version of this article includes the following figure supplement(s) for figure 2:  tested, we observed immobile, stably docked vesicles ( Figure 5A, Table 5, Table 6). Surprisingly, the Cpx-dependent clamp was observed even under conditions wherein the Syt1 clamp was absent (i.e. Syt1 Q and Syt1 F349A ). Survival analysis revealed that both Syt1 Q and Syt1 F349A mutants introduce a meaningful delay in the overall fusion process albeit less than that observed with a stable Syt1 clamp ( Figure 5-figure supplement 1). This implies that the kinetic delay introduced by Syt1, independent of its ability to clamp, is sufficient to enable Cpx to function as a fusion clamp, perhaps by providing time for Cpx to bind and block all SNAREpins.
However, the clamped Syt1 Q and Syt1 F349A vesicles were insensitive to Ca 2+ and did not fuse following Ca 2+ (1 mM) addition as opposed to the rapid and synchronous fusion observed with the majority of the Syt1 WT and Syt1 LLQQ vesicles ( Figure 5B). Notably, a significant partial fraction (~25%) of the Syt1 LLQQ vesicles remained un-fused even following Ca 2+ addition ( Figure 5B). This suggests that while the Syt1-Cpx-SNARE tripartite interface is not essential for establishing the fusion clamp, it is likely important in the Ca 2+ -triggering of fusion.
Taken together, our data indicate that (i) the Syt1 clamp and associated SNAREpins are critical for Ca 2+ -triggered fusion and (ii) Cpx, on its own, irreversibly blocks vesicle fusion. These conclusions are corroborated by Cpx washout (±CDT) experiments on the Syt1 Q vesicles (Figure 5-figure supplement 2). Without CDT, Syt1 Q vesicles fused spontaneously following the Cpx washout. In the presence of CDT, the Syt1 Q vesicles remained docked following the buffer wash, but were insensitive to Ca 2+ and failed to fuse following the addition of 1 mM Ca 2+ ( Figure 5-figure supplement  2). This denotes that in Syt1 Q vesicles (and presumably in Syt1 F349A vesicles), all SNAREpins are clamped by Cpx alone and become accessible to CDT block following the Cpx wash-out.

Discussion
Here we report that Syt1 and Cpx act concomitantly to clamp SNARE-driven constitutive fusion events (Figure 1). We find there are at least two distinct sets of clamped SNAREpins under every docked vesicle -a small population that is reversibly clamped by Syt1 oligomers and the remainder that is irreversibly blocked by Cpx. (Figure 3). These results taken together with the known structural and biochemical properties of Syt1 and Cpx prompts a novel 'synergistic clamping' mechanism.
We posit that the Syt1 C2B domain binds PIP2 (via the poly-lysine motif) on the PM and assembles into ring-like oligomeric structures (Bello et al., 2018;Wang et al., 2014;Zanetti et al., 2016). The Syt1 oligomers concurrently bind the t-SNAREs via the 'primary' interface (Zhou et al., 2015;Zhou et al., 2017;Grushin et al., 2019). The Syt1-t-SNARE interaction, which likely precedes the engagement of the v-and the t-SNAREs, positions the Syt1 such that it sterically blocks the full assembly of the associated SNAREpins . The Syt1 oligomers in addition to creating a stable steric impediment could also radially restrain the assembling SNAREpins Video 4. Representative video showing the docking, diffusion and spontaneous fusion of a vSUV to freestanding t-SNARE containing lipid bilayer in the presence of 2 mM Cpx. The movie was acquiredusing a Leica confocal scanning microscope at a speed of 7 frames/sec and the ATTO647N-PE fluorescence was used to track the vesicle. For clarity, a single hole corresponding to 5 mm suspended bilayer marked by the white circle is shown. https://elifesciences.org/articles/54506#video4 Video 5. Representative video showing the rapid and synchronous fusion of multiple Syt1/Cpx clamped vesicles to free-standing t-SNARE containing lipid bilayer triggered by Ca 2+ -influx (1 mM free). The ATTO647N-DOPE introduced in the SUV was used to track the fate of the vesicle and the movie was acquired using a Leica confocal scanning microscope at a speed of 7 frames/sec. For clarity, a single hole corresponding to 5 mm suspended bilayer marked by the white circle is shown.
The Syt1 oligomers can bind and clamp only a small sub-set of SNAREpins (which we refer to as 'central SNAREpins') as the number of potential SNAREpins per SV far exceeds the Syt1 density (~70 copies of VAMP2 vs.~20 copies of Syt1). We suggest that the remaining 'peripheral' SNAREs would then be unrestrained and can fully zipper to catalyze vesicle fusion, though at an impeded rate (Figure 1). It is tempting to speculate that the limited set of Syt1-clamped 'central' SNAREpins correspond physiologically to the six symmetrically organized protein densities that underlie each docked synaptic-like vesicle visualized by cryo-electron tomography analysis (Li et al., 2019). Indeed, this symmetrical organization depends on Syt1 oligomerization (Li et al., 2019).
Cpx potentially binds the SNARE complex only when the Syntaxin and VAMP2 are partiallyassembled (Chen et al., 2002;Kümmel et al., 2011) and competitively blocks the complete zippering of the C-terminal portion of the VAMP2 SNARE motif (Kümmel et al., 2011;Li et al., 2011;Giraudo et al., 2009). It is well-established that the C-terminal portion of VAMP2 assembles with extraordinarily high energy and rate acting as the major power stroke to drive membrane fusion (Gao et al., 2012). Consequently, Cpx, on its own, is ineffective in clamping SNARE-driven vesicle fusion (Figure 1).
We envision that under physiological conditions, Syt1 sets the stage for Cpx to bind and clamp the 'peripheral' SNARE complexes. Syt1, in addition to fully-arresting the central SNAREpins, also  kinetically hinders the assembly of the peripheral SNAREpins enabling Cpx to effectually arrest their complete zippering. In this manner, the sequential action of Syt1 and Cpx on different populations of SNAREpins would be involved in the generation of the stably docked, release-ready vesicles. Under each vesicle, Syt1 oligomers bind and clamp a small sub-set of available SNAREpins at the early stages of docking, which in turn enables Cpx to block the remainder of the SNAREpins. This molecular model would readily explain the importance of both Syt1 and Cpx in establishing the To focus on the Syt1 component of the fusion clamp, all fusion analysis was carried out using vesicles containing low copy VAMP2 (~13 copies) with normal number (~25 copies) of Syt1 molecules (wild-type or targeted mutations) in the absence of Cpx. (A) Survival analysis shows that disrupting the Syt1-SNARE primary interface (Syt1 Q , green curve) or destabilizing Syt1 oligomerization (Syt1 F349A , purple curve) abrogates the Syt1 clamp, whilst the Syt1-Cpx-SNARE tripartite interface (Syt1 LLQQ , yellow curve) and the Ca 2+binding motif on the Syt1 C2B domain (Syt1 DA , red curve) are not involved in establishing the fusion clamp. (B) Addition of Ca 2+ triggered rapid fusion of the majority (>90%) of Syt1 LLQQ and the remainder (~15%) of the docked Syt1 F349A vesicles, very similar to the behavior of the Syt1 WT vesicles. Predictably, blocking Ca 2+ -binding to C2B domain rendered the vesicle Ca 2+ -insensitive, with the majority of Syt1 DA remaining un-fused. We did not have sufficient number of docked Syt1 Q vesicles to do a quantitative analysis, but qualitatively, the few that remained docked failed to fuse following Ca 2+ -addition. This implies that both C2B binding to Ca 2+ and SNAREs are required for the Ca 2+ -activation, but the ability to form oligomers or Syt1-Cpx-SNARE tripartite interface are not crucial for the Ca 2+ -triggered reversal of the Syt1 clamp. The average values and standard deviations from four independent experiments are shown for each condition. In total,~250 vesicles were analyzed for each condition. Table 3. Survival probabilities at specific points in time post-docking (Kaplan Meier estimators) for the different Syt1 mutants under low-copy VAMP2 conditions in the absence of Cpx. The corresponding survival curves are shown in Figure 4A Time ( fusion clamp as evidenced in several genetic deletion studies (Bacaj et al., 2013;Geppert et al., 1994;Cho et al., 2014;Martin et al., 2011;Littleton et al., 1993;Yang et al., 2013).
We also report that the reversal of the Syt1 clamp is sufficient to drive rapid Ca 2+ -triggered fusion of the docked vesicles ( Figure 2). It involves Syt1 C2B domain binding both Ca 2+ and the SNARE complex ( Figure 4). We have recently demonstrated that Ca 2+ -binding to Syt1 C2 domains induce a large-scale conformational rearrangement of the Syt1-SNARE complex on the lipid membrane surface, disrupting the pre-fusion clamped architecture . This, in effect, reverses the Syt1 clamp, allowing the associated SNAREs to complete zippering and drive fusion. This implies that only a small fraction of available SNAREpins per vesicle (i.e. only those associated with Syt1) are involved in the Ca 2+ -activation process. This is consistent with the earlier reports that 2-3 SNARE complexes can be sufficient to facilitate Ca 2+ -evoked synchronous neurotransmitter release (Sinha et al., 2011;Mohrmann et al., 2010). Indeed, recent modeling studies considering the concept of mechanical coupling have predicted that an optimum of 4-6 SNAREpins is required to achieve sub-millisecond vesicular release (Manca et al., 2019).
There is a long-standing debate over the role of Cpx in establishing a fusion clamp Ló pez-Murcia et al., 2019). We find that Cpx is an integral part of the overall clamping mechanism and Syt1 and Cpx play distinct roles in clamping different pools of SNAREpins. However, Cpx requires a kinetic delay (likely introduced by Syt1) to block vesicle fusion and this inhibition is not released by Ca 2+ . This raises the intriguing possibility that the Cpx clamp is not necessarily reversed during the Ca 2+ -activation process. The molecular details of the observed Cpx 'clamp' and its physiological relevance remains to be determined.
Overall, our data are in alignment with the emerging view that Syt1 plays a pivotal role in orchestrating Ca 2+ -regulated neurotransmitter release. Syt1 functions both as a fusion clamp and the principal Ca 2+ -sensor to establish Ca 2+ -regulation of vesicular fusion (Geppert et al., 1994;Littleton et al., 1993). Furthermore, Syt1, by virtue of self-assembling into oligomeric structures, also provides the molecular framework to organize the exocytic machinery into a co-operative Table 4. The survival curves in Figure 4A were compared pair-wise using the log-rank test to determine the statistical significance (pvalues) of the effects of the Syt1 mutants under low-copy VAMP2 conditions in the absence of Cpx.   Figure 5A. structure to enable ultra-fast fusion (Li et al., 2019;Rothman et al., 2017;Volynski and Krishnakumar, 2018).
We have articulated the simplest hypothesis, considering discrete 'central' and ''peripheral' SNAREpins associated with Syt1 and Cpx, respectively. However, it is easy to imagine that Cpx also binds the central SNAREpins. Indeed, recent X-ray crystal structure revealed that both Syt1 and Cpx can bind the same pre-fusion SNARE complex (Zhou et al., 2017). Cpx binding is also predicted to create a new Syt1 binding site (i.e. tripartite interface) on SNAREs. In our reconstituted setup, we find that this SNARE-Cpx-Syt1 'tripartite' interface is not required to produce the fusion clamp but is likely involved in Ca 2+ activation of fusion from the clamped state ( Figure 5). (A) Survival analysis shows neither the disruption of the Syt1-SNARE primary interface (Syt1 Q , green curve) and the Syt1-Cpx-SNARE tripartite interface (Syt1 LLQQ , yellow curve) nor destabilizing Syt1 oligomerization (Syt1 F349A , purple curve) has any effect on the fusion clamp in the presence of Cpx, with stably docked vesicles observed under all conditions. (B) End-point analysis indicates that the Syt1 Q and Syt1 F349A vesicles are insensitive to Ca 2+ and do not fuse upon addition of 1 mM Ca 2+ . In contrast, rapid and synchronous Ca 2+ -triggered fusion was observed with the majority of the docked Syt1 WT and Syt1 LLQQ vesicles. Notably, the LLQQ mutation does have a small but significant effect on Ca 2+ -triggered release. This implies that the Syt1-Cpx-SNARE tripartite interface is not necessary for clamp, but might be involved in the Ca 2+ -activation process. Overall, the data suggest that the Syt1 clamped SNAREpins are required for Ca 2+ -triggered exocytosis and in the absence of Syt1 clamp, Cpx irreversibly blocks vesicle fusion. The average values and standard deviations from three independent experiments are shown for each condition. In total,~200 vesicles were analyzed for each condition. The online version of this article includes the following figure supplement(s) for figure 5:    Figure 5A were compared pair-wise using the log-rank test to determine the statistical significance (p-values) of the effects of the Syt1 mutants under normal VAMP2 conditions in the presence of 2 mM Cpx. Since the SNARE-Syt1-Cpx tripartite interface is created only when the SNAREs are partiallyassembled (i.e. when Cpx binds), it is possible that this interaction plays an auxiliary role in establishing the fusion clamp. As such, it can be bypassed in our biochemically-defined experimental setup. The tripartite interface might become more relevant in the pre-synaptic terminals (Zhou et al., 2017) where~30% of Syt1 is present in the PM (Wienisch and Klingauf, 2006) and other Synaptotagmins could also participate in the tripartite interface (Rothman et al., 2017;Volynski and Krishnakumar, 2018;Zhou et al., 2017).
Indeed, we have postulated that such a 'dual-clamp' arrangement with Syt1 (from SV) occupying the 'primary' site and with Syt1 and Syt7 (from PM) competing for the 'tripartite' site (in conjunction with Cpx) could potentially explain the synergistic regulation of neurotransmitter release by different Syt isoforms (Rothman et al., 2017;Volynski and Krishnakumar, 2018). Such an arrangement could also potentially explain how Cpx could regulate Ca 2+ -evoked synchronous neurotransmitter release.
Physiologically, all three synaptic SNAREs (Syntaxin, SNAP-25 and VAMP2) are assembled into a SNAREpin in a concerted reaction involving chaperones Munc13 and Munc18 (Baker and Hughson, 2016;Rizo, 2018), the need for which is by-passed in our system by pre-assembled t-SNARE complexes. We imagine that these chaperones also function in a concerted manner with Syt1 and Cpx and have recently advanced a speculative model outlining how these proteins can co-operate to template SNAREpin assembly (Rothman et al., 2017). Moreover, lipid membranes may contribute to synergism or cooperativity between Syt1 and Cpx in both clamping the un-initiated fusion events and triggering rapid and synchronous fusion in response to Ca 2+ -influx.
As such, further studies with high temporal resolution involving detailed mutational and topological analysis of Syt1 and Cpx, along with the chaperones Munc13 and Munc18, is needed to establish pointillistic details of fast Ca 2+ -triggered SV fusion. Nonetheless, our data demonstrate that Syt1 and Cpx, along with SNARE proteins, form the minimal protein machinery that is necessary and sufficient to establish rapid Ca 2+ -regulated exocytosis.

Protein expression and purification
All proteins (v-and t-SNAREs, Cpx, Syt1 wild type and mutants) were expressed and purified as described previously (Mahal et al., 2002;Krishnakumar et al., 2013;Krishnakumar et al., 2011;Weber et al., 1998). In brief, proteins were expressed in E. coli strain Rosetta2(DE3) (Novagen, Madison, WI) using 0.5 mM IPTG for 4 hr. Cells were pelleted and lysed using a cell disruptor (Avestin, Ottawa, Canada) in HEPES buffer (25 mM HEPES, 400 mM KCl, 4% Triton X-100, 10% glycerol, pH 7.4) containing 0.2 mM Tris(2-carboxyethyl) phosphinehydrochloride (TCEP), and 1 mM matlabcentral/fileexchange/66521-fusion-analyzer-fas. Note: We excluded vesicles bound to the edge of the holes as they may not be representative of vesicles bound to the free-floating membrane. We thus used only the centrally-docked vesicles for analysis. We did not observe any change in ATTO-647-PE fluorescence for the vesicles that remain docked and un-fused during the observation period or post Ca 2+ -addition. Thus, we can rule out hemi-fusion diaphragm formation as a possible explanation for the observed 'clamped' or 'un-fused' state.

Single-vesicle docking analysis
To get an accurate count of the docked vesicles, we used VAMP2 protein with mutations in the C-terminal half (L70D, A74R, A81D and L84D; VAMP2 4X ) that eliminates fusion without impeding the docking process (Krishnakumar et al., 2013). For the docking analysis, VAMP2 4X containing SUVs (vSUV 4X ) were introduced into the chamber and allowed to interact with the t-SNARE bilayer. After a 10 min incubation, the bilayer was thoroughly washed with running buffer (3x minimum) and the number of docked vesicles were counted. For an unbiased particle count, we employed a customwritten algorithm to count particles from top-left to bottom-right that ensures every spot is counted only once.

Calcium dynamics
To quantify the kinetics of Ca 2+ -triggered vesicle fusion, we used a Ca 2+ -sensor dye, Calcium Green conjugated to a lipophilic 24-carbon alkyl chain (Calcium Green C24) introduced in the suspended bilayer to directly monitor the arrival of Ca 2+ . Calcium green is a high-affinity Ca 2+ -sensor (K d of~75 nM) and exhibits a large fluorescent increase (at 532 nm) upon binding Ca 2+ . To accurately estimate the arrival of Ca 2+ to/near the vesicles docked on the bilayer, we used confocal microscopy equipped with resonant scanner focused at or near the bilayer membrane and acquired movies at a speed of up to 36 msec per frame. We typically observed the fluorescence signal increase at the bilayer surface between about three frames (~100 msec) after Ca 2+ addition (Figure 2-figure supplement 1). We therefore used 100 msec as the benchmark to accurately estimate the time-constants for the Ca 2+ -triggered fusion reaction.