Two recent papers by Mehta et al. and Zhu et al. in this issue (https://doi.org/10.1083/jcb.202311191) discover that synaptotagmin-1, the primary calcium sensor at the synapse, forms biomolecular condensates, identifying a new layer of regulation in calcium-triggered synaptic vesicle exocytosis.
Synaptic vesicle (SV) fusion is arguably the fastest membrane remodeling reaction in nature, occurring in just a few milliseconds. This remarkable pace of membrane bilayers fusing together is mediated by the SNARE complex, yet only upon calcium signaling (1). Tight regulation is accomplished by synaptotagmin-1, the main calcium sensor at the synapse (2). While genetics, biochemistry, and structural biology revealed detailed insights on the trajectory and the regulation of the fusion reaction, whether and how synaptotagmin-1 (re)organizes at fusion sites upon depolarization remains controversial (3, 4).
In two recent papers, combining computational and experimental approaches, Mehta et al. and Zhu et al. discovered that synaptotagmin-1 forms biomolecular condensates (5, 6). Synaptotagmin-1 is an integral SV protein consisting of a luminal region at the N-terminus (aa. 1–57), followed by a transmembrane domain (aa. 58–79) and a cytoplasmic tail (aa. 80–421), which compromises a juxtamembrane linker (aa. 80–141), a tandem C2AB domain (C2A, aa. 142–261; C2B, aa. 273–406), and a C-terminal region (aa. 407–421).
Most efforts focused on calcium sensing by the C2 domains, overlooking the juxtamembrane linker, at least partially due to difficulties in expressing and purifying the full cytoplasmic tail of synaptotagmin-1. Instead, studies often used truncated versions missing a lysine-rich motif in the juxtamembrane linker. The two new studies successfully purified the entire cytoplasmic tail, yielding novel insights into the function of synaptotagmin-1 (5, 6). Computational analyses suggested the presence of an intrinsically disordered region (IDR)—a primary sequence lacking a stable secondary or tertiary structure—in the juxtamembrane linker. IDRs can drive the condensation of associative biomolecules by multivalent interactions (7). Indeed, coarse-grained molecular dynamic (MD) simulations by Mehta et al. predicted synaptotagmin-1 to form condensates (5).
Both groups discovered that synaptotagmin-1 forms fluid-like condensates under near-physiological conditions, showing fusion, fission, and rapid fluorescence recovery upon photobleaching. A careful analysis of truncation variants revealed that the lysine-rich N-terminal segment within the IDR drives condensation, suggesting an important role of electrostatic interactions in this process. This conclusion is further supported by the dissolution of synaptotagmin-1 condensates at high salt concentrations and the inability of mutants, lacking the lysine residues in the juxtamembrane linker, to form condensates in vitro (5, 6) and in simulations (5). MD simulations suggested that the lysine residues in the juxtamembrane linker are implicated in forming long-living hydrogen bonds with negatively charged side chains.
Classically, calcium binding to the C2 domains is perceived to regulate synaptotagmin-1 binding to both SNAREs and membranes by inducing a conformational change (8, 9). However, new data indicate that the condensates can buffer the response of synaptotagmin-1 to calcium. Furthermore, without calcium, these condensates stiffen into a gel-like state, a shift reversible by calcium in a concentration-dependent manner (6). The combination of isothermal titration calorimetry and computational analyses indicated that synaptotagmin-1 condensation decreases the affinity of the protein for calcium, adding another layer of regulation to the interplay between calcium binding and phase separation (5). Synaptotagmin-1 condensates selectively recruited the preformed SNARE complex (consisting of syntaxin-1, SNAP-25, and VAMP2), syntaxin 1, and SNAP-25, whereas VAMP2 and complexin-1 remained excluded unless part of the SNARE complex (6). This strongly suggests the existence of chemical complementarity that renders the presynapse as a complex emulsion of multiple condensates (10).
To verify their biochemical findings in a cellular system, both groups ectopically expressed different synaptotagmin-1 variants in mammalian cell lines. Similar to the in vitro data, the cytosolic tail of synaptotagmin-1 readily formed condensates as a function of the lysine-rich motif within the IDR. Additionally, Zhu et al. employed an OptoDroplet system by fusing synaptotagmin-1 to a fluorescently tagged Cry2 (6), which weakly self-oligomerizes upon blue light activation, boosting local concentration (11). Indeed, within seconds, mCherry-Cry2-synaptotagmin-1 formed condensates, unlike the mCherry-Cry2 vector. On the other hand, Mehta et al. applied the calcium ionophore calcimycin on HEK cells overexpressing synaptotagmin-1 to determine whether the acute increase of calcium affects synaptotagmin-1 condensates in cells (5). The rise in the cytosolic calcium concentration led to an increase in droplet size, recapitulating the in vitro observations.
The synaptotagmin-1 C2AB domain inserts itself into the plasma membrane during SV exocytosis, emphasizing the crucial role membranes play in synaptotagmin-1 function. In an elegant series of experiments, Zhu et al. examined the interplay between synaptotagmin-1 condensation and membranes (6). The authors prepared supported lipid bilayers containing a modified lipid that binds His-tags (i.e., DGS-NTA-Ni(II)) and reconstituted the N-terminal His-tagged cytosolic tail on the membrane surface, thereby mimicking its physiological localization. Synaptotagmin-1 attachment to the membrane decreased the concentration necessary for condensation compared with the solution. The condensation was further accelerated by including negatively charged phospholipids, highlighting the importance of electrostatic interactions. Similarly, Mehta et al. reported that the presence of soluble anionic phospholipids promoted condensation in solution (5).
Furthermore, Zhu et al. prepared giant unilamellar vesicles (GUVs) and reconstituted them together with preformed condensates, finding that the droplets wet the membrane in a charge-dependent manner (6). Using the DGS-NTA-Ni(II)/His-tag system, the authors attached the cytosolic tail of synaptotagmin-1 to small unilamellar vesicles (SUVs), mimicking the organization of SVs, and added these to GUVs, mimicking the plasma membrane. The synaptotagmin-1-bound SUVs were efficiently attached to GUVs, unlike protein-lacking SUVs. Notably, an IDR-lacking synaptotagmin-1 variant, unable to form condensates, drastically reduced SUV recruitment, supporting the role of synaptotagmin-1 condensates in vesicle docking.
Both groups validated different GFP-tagged synaptotagmin-1 variants in rat hippocampal neurons of either synaptotagmin-1 KD (6) or WT background (5), respectively. While both full-length (FL) and the cytosolic tail of synaptotagmin-1 led to the formation of condensate-like structures in living neurons, only the FL construct retained synaptic localization, indicating an essential role for the intraluminal region and the transmembrane domain in targeting to synapses. The expression of synaptotagmin-1 variants lacking the IDR or the lysine residues within the IDR resulted in a more diffuse signal, articulating that the lysine-rich motif is required for the condensation in living neurons.
Together, Mehta et al. and Zhu et al. provide a new layer of insights into synaptotagmin-1-mediated exocytosis (Fig. 1). At rest, synaptotagmin-1 oligomerization, presumably into ring-like structures, triggers the formation of liquid-like condensates mediated by the positively charged IDR in the juxtamembrane region, which allows anchoring of SVs to the plasma membrane via calcium-independent electrostatic interactions (step 1). SNARE proteins and other regulatory factors concentrate within the condensates, facilitating their interactions (step 2). The trans-SNARE complex assembles with synaptotagmin-1 binding to its primary interface, while another synaptotagmin molecule engages with the tripartite interface via C2B domains. Simultaneously, the C2B domain interacts with the inner leaflet of the plasma membrane. Over time, condensates transition into a gel-like state, slowing down the diffusion kinetics within the condensate (step 3).
Once an action potential reaches the synaptic bouton, depolarization of the plasma membrane leads to the opening of voltage-gated calcium channels, resulting in a local influx of calcium, reversing the liquid-to-gel-like transition (step 4). Additionally, calcium binds to synaptotagmin-1’s C2AB domain, resulting in dissociation from the SNARE complex and membrane insertion of the C2B domain (step 5). Then, SNARE zippering reduces the distance between the two membranes, rendering lipid splaying more likely, thereby lowering the energy barrier for membrane fusion.
This new layer of regulating exocytosis at synapses raises intriguing questions. How does calcium drive synaptotagmin-1 conformational changes within condensates to facilitate membrane binding? How is coupling between the synaptotagmin-1 domains and active zone components that also form viscoelastic phases (12) achieved, and how is this influenced by activity-induced phosphorylation (13)? What is the impact of synaptotagmin-1 condensation on SV diffusion in synapsin-rich condensates upstream of fusion sites (14)? The selective recruitment of the preformed SNARE complex into synaptotagmin-1 condensates, while excluding VAMP2, an SV protein integral to the SNARE complex, highlights a complex network of interactions yet to be clarified. Understanding these molecular mechanisms remains challenging, but these studies shed fresh light on the regulation of exocytosis and promise to spark further research into this vital cellular process.
