Calcium-sensitive synaptotagmin 11-lipid interaction modulates exo-endocytosis

Abstract

Synaptotagmins (Syts) are the primary Ca²⁺-sensors for synaptic vesicle exocytosis, while most mammalian Syts are non-Ca²⁺-affinitive and play critical roles in neurotransmission and synaptic plasticity with unclear mechanisms. Here, we show that high-alkaline non-Ca²⁺-binding Syt11 exhibits higher affinity for acidic phospholipids and Ca²⁺-inhibited liposome-binding, thereby competing with the Ca²⁺-binding Syt1. Physiological levels of Ca²⁺ eliminate this competition by promoting Ca²⁺-dependent membrane insertion of Syt1 while suppressing Syt11’s binding through electrostatic shielding of the membrane surface. Site-directed mutagenesis reveals a dual-regional lipid-binding mode (a lysine-rich motif for Ca²⁺-independent binding and Ca²⁺-binding loops for Ca²⁺-facilitation) for Syt1, and a redundant multi-point lipid-binding interface for Syt11. Consistent with the Ca²⁺-dependent competition, Syt11 inhibits both the early stages of exocytosis and endocytosis in neurons, while the maximal rate of exocytosis remains intact. This Ca²⁺-sensitivity of Syt11 proposes Syt1-Syt11 inter-switching in membrane-occupancy as a critical step precisely controlling exocytosis and endocytosis during synaptic transmission.

Syt1 is a primary Ca²⁺ sensor for secretion but the function of non-Ca²⁺ -affinitive Syt11 remains unknown. Here, authors identify strong but Ca²⁺ -inhibited lipid affinity of Syt11, and a Syt11-1-11 inter-switch during excitation-coupled exo-endocytosis.

Introduction

Ca²⁺ is the key messenger triggering neural transmission, yet how Ca²⁺ signals are faithfully translated into transmission events is not fully understood. Synaptotagmins (Syts) are the primary Ca²⁺ sensors responsible for triggering synaptic vesicle exocytosis by forming fusion machinery with the soluble N-ethylmaleimidesensitive factor (NSF) attachment protein receptor (SNARE) complex. However, the molecular architecture and assembly mechanisms of this fusion machinery remain largely unknown¹. Mammalian Syts constitute a large gene family with 17 members (Syt 1–17), each sharing a conserved structural framework: (1) a luminal (or extracellular) domain; (2) a single-helix transmembrane domain (TMD, or membrane-anchoring sites in Syt 17); (3) a juxtamembrane linker; and (4 and 5) two tandem C2 domains (C2A and C2B). These C2 domains bind 2–4 Ca²⁺ ions each through aspartate residues (D_Ca) in their loop regions. Notably, only 8 of them (Syt 1, 2, 3, 5, 6, 7, 9, and 10) are canonical Ca²⁺-affinitive Syts (Ca☑-Syts), while the remaining 9 isoforms are non-Ca²⁺-affinitive (Ca⌧-Syts) due to evolutionary mutations in these key aspartate or related residues, e.g., the aspartate-to-serine substitutions in the C2A domains of Syt4 and Syt11². Despite their lack of Ca²⁺-affinity, several Ca⌧-Syts are abundantly expressed in the central nervous system (CNS) and have been implicated in neurodegenerative and psychiatric disorders, e.g., Syt11 has been defined as a risk factor mediating the pathogenesis of Parkinson’s disease (PD) and schizophrenia^3–6. However, whether and how these Ca⌧-Syts directly participate in vesicular exocytosis remain unclear.

Ca☑-Syts, despite their distinct binding affinities, serve as the primary Ca²⁺ sensors initiating vesicular exocytosis and synaptic transmission^7–9. They play critical roles in vesicle docking, priming, and fusion pore dynamics during exocytosis^10–13, as well as the coupling of endocytosis to exocytosis during synaptic transmission. Our previous work has depicted a bidirectional regulatory role of Syt1 in promoting clathrin-mediated endocytosis (CME) but inhibiting activity-dependent bulk endocytosis during sustained synaptic activity¹⁴. Meanwhile, Syt3 has been reported to localize post-synaptically to drive the off-membrane trafficking of α-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor (AMPAR) through Ca²⁺-dependent endocytosis¹⁵. In contrast, Ca⌧-Syts have been shown to function opposite to Ca☑-Syts. Ca⌧-Syt4 has been shown to inhibit exocytosis and endocytosis^16–18. We have also demonstrated that Syt11 exerts inhibitory effects on both CME and bulk endocytosis in dorsal root ganglion, hippocampal, and midbrain dopaminergic neurons^4,19. However, the molecular basis and mechanisms underlying the inhibitory roles of Ca⌧-Syts remain poorly understood. Furthermore, whether and how Ca☑-Syts and Ca⌧-Syts cooperate to regulate exocytosis and endocytosis during synaptic transmission remain virtually unknown.

The molecular mechanism by which Syt proteins regulate exocytosis and endocytosis remains a subject of debate¹. In the case of Syt1, crystallographic studies have revealed its multiple distinct interaction interfaces with SNARE complex, including a conserved R-apex (residues R398/R399)^20,21, which is essential for Syt1 to initiate Ca²⁺-dependent exocytosis. Syt1 also adopts a clamping conformation with SNARE to prevent spontaneous membrane fusion before Ca²⁺ influx^22–24. Meanwhile, the interactions of Syt1 with phospholipids and Ca²⁺ are equally critical for vesicle docking, fusion initiation, and pore dilation^25–27. Syt1 is proposed to bind acidic phosphatidylserine (PS) liposomes in a Ca²⁺-dependent manner^28–31, while a conserved K-rich motif (also known as the polybasic patch) enables Ca²⁺-independent binding to phosphatidylinositol 4,5-bisphosphate (PIP2)^32–34. Lipid-binding enhances the Ca²⁺ affinity of Syt1³⁵, forming a Syt-Ca²⁺-lipid complex that enables a deep penetration of the Ca²⁺-binding loops into phospholipids bilayer and causes the deforming of the plasma membrane to facilitate fusion^36,37. While some hypotheses suggest that Syt-Ca²⁺-lipid interactions may be more critical than Syt-SNARE interactions in exocytosis³⁸. According to our model, the Ca²⁺-dependent membrane remodeling effects of Syt1 also provide a compelling explanation for its bidirectional role in endocytosis¹⁴. Therefore, elucidating the molecular basis of Syt-lipid interactions and their Ca²⁺-sensitivity is essential for understanding the dual roles of Syt proteins in coordinating exocytosis and endocytosis during neurotransmission.

In the present work, by using a combination of biochemical assays, in silico analyses, in vitro liposome experiments, and pHlourin imaging in primary cultured neurons, we show that the C2 domains of Ca☑-Syts exhibit distinct dynamics in Ca²⁺/lipid binding, define the Ca²⁺-sensitive binding of Ca⌧-Syts to liposomes, and propose a Ca²⁺-dependent Ca☑-Syt/Ca⌧-Syt inter-switch working model for the excitation-coupled exocytosis and endocytosis during synaptic transmission.

Results

Non-Ca²⁺-affinitive Syts display higher alkaline surfaces

The structures of all 34 rat Syt-C2 domains predicted by AlphaFold^39,40 were used to estimate their lipid-binding potential. We found that the C2 domains of Ca☑-Syts shared a common characteristic of highly basic surfaces, e.g., the lysine (K)-rich motifs in the C2B domains of Syt1 and Syt7, which is critical for establishing strong electrostatic interactions, thereby conferring high affinity for the acidic surface of the plasma membrane. Based on the full-length amino acid sequence similarity of Syts, the Ca⌧-Syts 12 ~ 17 were distinguished from Ca☑-Syts, while the Ca⌧-Syts 4, 8, and 11 clustered with Ca☑-Syts 1, 2, 5, and 7 (Fig. 1a), indicating that not all Ca⌧-Syts are evolved independently with Ca☑-Syts. Thus, Syts 4, 8, and 11 are likely to exhibit greater molecular and functional consistencies rather than differentials with Ca☑-Syts, except for Ca²⁺-affinity. Although the C2A domains of Ca☑- and Ca⌧-Syts showed diverse isoelectric points (pIs), most C2B domains had a higher pI (> 8.5), suggesting that the C2B domains of Ca⌧-Syts also exhibit highly alkaline, positively-charged surfaces (Fig. 1b). Consistent with this, the electrostatic potential mapping of C2 domains revealed that most C2B domains of both Ca☑- and Ca⌧-Syts harbored a large region of high alkalinity (Fig. 1c, red region), centered around the K-rich motif. In contrast, such alkaline region was absent in the C2A domains.

Fig. 1. Non-Ca²⁺-affinitive Syts exhibit high alkalinity.

a Phylogenetic tree of amino acid sequence similarities among all human and rat Syts, showing Ca⌧-Syts 4, 8, and 11 are clustered within Ca☑-Syts. Ca☑-Syts: gray; Ca⌧-Syts: red; x-axis: the distance values represent the number of substitutions as a proportion of the alignment length (excluding gaps), thereby indicating the extent of genetic change. b Calculated pIs of all rat Syt-C2 domains, highlighting the high pI of most C2B domains. c AlphaFold-predicted structures of all rat C2 domains, illustrating surface electrostatic potential. Top left: Syt11-C2 domains aligned with a Ca²⁺-bound Syt1 crystal structure (PDB: 5CCH); top right: Ca☑-Syt C2 domains with positive (red) and negative (blue) electrostatic potentials; bottom: Ca⌧-Syt C2 domains showing extensive alkaline surfaces. d Electrostatic potential profiles of Ca☑- and Ca⌧-C2 domains plotted from N- to C-terminus, showing regional similarity and difference in alkalinity. e Averaged electrostatic potential of residues across all C2 domains (Number of residues is indicated by N). f Calculated solvent-accessible surface area (SASA) of highly basic regions (potential > 75 mV) for all C2 domains, demonstrating higher alkalinity of Ca⌧-Syts (Number of Syt isoforms is indicated by N). Data are presented as the mean ± 90% CI for (d); Box-and-whisker for (e, f): the median represented by the central line inside each box, the 25th and 75th percentiles represented by the edges of the box, and the whiskers extending to the most extreme data points; two-tailed Mann–Whitney test, ***P < 0.001, n.s. no significant difference. Source data are provided as a Source Data file.

The electrostatic profiles of Ca☑- and Ca⌧-Syt C2 domains were largely similar when analyzed from the N- to C-terminus, except for a short fragment around residue-25 in the C2B domains where Ca☑-Syts exhibited a negative and Ca⌧-Syts displayed a positive potential (Fig. 1d). Most C2A domains in both Ca☑- and Ca⌧-Syts showed a negative electrostatic potential, though this potential was higher in Ca⌧-Syts compared to Ca☑-Syts. In contrast, the C2B domains had a positive potential, which was also higher in Ca⌧-Syts (Fig. 1e). To further quantify these differences, we then analyzed the solvent-accessible surface area (SASA) of highly basic regions (defined as atoms with an electrostatic potential > 75 mV). Most C2A domains had a smaller basic SASA than C2B domains, which was higher in Ca⌧-C2s than that in Ca☑-C2s in average (Fig. 1f). Collectively, these findings indicate a high potential of Ca⌧-Syts for lipid interactions, which is even stronger than that of Ca☑-Syts.

Syt11 binds acidic phospholipids with Ca²⁺-sensitivity

To investigate the phospholipid binding capabilities of Ca☑- and Ca⌧-Syts, we purified Syt1 (a representative Ca☑-Syt) and Syt11 (a representative Ca⌧-Syt) for the in vitro co-liposome precipitation (co-LP) analysis (Fig. 2a). The TMD and most of the juxtamembrane linker region were removed from both proteins to minimize their interference. We used two distinct lipid compositions: (1) the tubulation assay (TA) lipids, which contain a higher concentration of acidic phospholipids (45% PS + 5% PIP2), as used in our previous research¹⁴; and (2) the brain extract total (BET) lipids, which contain a lower concentration of acidic phospholipids (~10.6% PS + 1% PIP2). As a Ca²⁺ sensor, Syt1 exhibited a clear Ca²⁺-enhancement in binding to BET liposomes. However, when TA lipids were used, Ca²⁺ did not enhance the binding (Fig. 2b), suggesting that the interaction is predominantly electrostatic in nature and that the higher negative charge density of the TA membrane compensates for the Ca²⁺-promotion effect on Syt-lipid interactions. In contrast, Syt11, as a Ca⌧-Syt, bound to both TA and BET liposomes in the absence of Ca²⁺, and this binding was suppressed by increasing Ca²⁺ concentrations. The inhibitory effect became apparent at 0.1 mM Ca²⁺ and a complete blockage occurred at ~5 mM Ca²⁺. This Ca²⁺-dependent inhibition was not due to changes in protein stability or ionic strength, as Ca²⁺ did not alter Syt11’s behavior in the absence of liposomes, and higher concentrations of KCl (up to 100 mM) did not affect the binding as CaCl₂ did (Fig. 2c). These findings validate the Ca²⁺-sensitive nature (facilitation vs inhibition) of Syt1/11–lipid interactions, most probably mediated by direct electrostatic forces.

Fig. 2. Ca²⁺ inhibits Syt11-acidic membrane interaction and blocks its competition with Syt1.

a Schematic of the co-LP assay. b Soluble tandem C2A-C2B domain of Syt1 binds to liposomes, with significant Ca²⁺-promotion effect dependent on acidic phospholipid levels (n = 3 independent biological replicates). c Syt11-liposome binding is suppressed at high-Ca²⁺ concentrations. The “Lipo (-)” condition, which contains the protein in solution in the absence of liposomes, is used as a negative control, and the results from other groups are quantified and presented as a percentage relative to this control condition (left: n = 5 independent biological replicates; right: n = 3 independent biological replicates). d Ca²⁺-inhibition persists in Ca⌧-Syts (Syt4 and Ca-deficient Syt1 mutants, n = 3 independent biological replicates) but is absent in the Ca²⁺-affinity-restored Syt11-S247D mutant (n = 4 independent biological replicates). e Syt11 (0, 0.5, 1, 2, 4, and 8 μg) competes with Syt1 (5 μg) for liposome-binding in Ca²⁺-free conditions but not in 2 mM Ca²⁺ (n = 3 independent biological replicates). GST-Syt1 (amino acids 96-421) was used to distinguish the protein bands of Syt1 and Syt11. f To eliminate the potential influence of GST dimerization on the competition experiments, tag-free Syt1 (96-421) was generated from Thrombin digestion (n = 3 independent biological replicates). g The Syt11-S247D mutant partially regains competitive ability in the presence of Ca²⁺ (n = 3 independent biological repeats). h 20 μM Ca²⁺ abolishes competitive effects (n = 6), suggesting physiological relevance. Data are presented as the mean ± s.e.m.; two-tailed paired Student’s t test for two-group comparison (b–h), one-way ANOVA for multi-group comparison (c–h), and two-way ANOVA for multi-group comparison (e–g). *P < 0.05, **P < 0.01, ***P < 0.001, n.s. no significant difference. TA tubulation assay, BET brain extract total. Source data and full scan gels are provided as a Source Data file.

An intermediate lipid composition (40% PS + 1.5% PIP2) was used in subsequent experiments to compensate the charge effect. Various lipid compositions have been used to study Syt–lipid interactions in vitro over the past 30 years. For instance, the TA lipids, which contain higher levels of PIP2, can attenuate the Ca²⁺-promotion effect of Syt1–lipid binding (Fig. 2b). In contrast, BET lipids, sourced from Avanti as whole brain lipids from porcine brains, contain ~10.6% PS but lack the specific presentation of the synaptic active zone (AZ). Meanwhile, human brain PS levels vary between 11.4 and 20.5%, depending on individual differences and brain regions⁴¹; and PIP2 in whole-cell lipids accounts for ~1%^42,43. On the other hand, since both PS and PIP2 are predominantly located in the inner leaflet of the plasma membrane, their concentrations can reach 22.8-41% and ~2%, respectively. Furthermore, PIP2 is enriched in clusters at AZs, with local concentrations reaching up to ~5%^44–46 or even higher⁴⁷. Therefore, to maintain the potential Ca²⁺-promoting effects, the lipid composition of 40% PS and 1.5% PIP2 are used to approximate physiological relevance.

To explore the molecular basis for the Ca²⁺-sensitivity of Syt-lipid interactions, the evolutionarily-substituted serine residue in the C2A domain of Syt11 was mutated back to aspartate (Syt11-S247D) to restore Ca²⁺-binding ability. Interestingly, Ca²⁺ failed to inhibit the liposome binding of the Syt11-S247D mutant, suggesting that the Ca²⁺-neutralized loop region facilitates lipid binding. Meanwhile, Ca²⁺ did not show promotion either, likely because the C2B domain of this mutant remains Ca²⁺-insensitive. On the contrary, the Syt1-D2N (D230/232/363/365 to asparagine) and D2E (D230/232/363/365 to glutamic acid) mutants, in which D_Ca residues in the C2AB were mutated to abolish Ca²⁺-affinity, replicated the Ca²⁺-inhibition in lipid binding like that observed in Ca⌧-Syts. Consistently, the Ca⌧-Syt4 also exhibited Ca²⁺-suppressed lipid-binding (Fig. 2d). Collectively, these results validate a conserved Ca²⁺-inhibition mechanism in the binding of Ca⌧-Syts to acidic phospholipids, and propose both facilitatory and inhibitory roles of Ca²⁺ in modulating Syt-lipid interactions.

The Ca²⁺-sensitive Syt1/11 competition for lipid binding

Given the opposing effects of Ca²⁺ on lipid-binding properties of Ca☑- and Ca⌧-Syts, we next tested whether these proteins compete for lipid binding. Notably, we observed a dose-dependent competition between Syt1 and Syt11 for lipid binding in the absence of Ca²⁺. However, this competitive effect was abolished under 2 mM Ca²⁺ conditions, and Syt11 was mostly removed from liposomes by Ca²⁺ (Fig. 2e, f). This outcome indicates a dual effect of Ca²⁺ on Syt1 (facilitation) and Syt11 (inhibition). In contrast, the Syt11-S247D mutant restored the competition with Syt1 in liposome-binding to some extent in 2 mM Ca²⁺ conditions. This might be due to the Ca²⁺-facilitation of lipid-binding of the S247D mutant, although its C2B domain remains Ca²⁺-insensitive, which was supported by that the liposome attachment of S247D mutant was much stronger than that of wild type Syt11 (Fig. 2g). These findings also suggest that Ca²⁺-binding generates additional membrane attachment sites for Ca☑-Syts, thereby reinforcing their Ca²⁺-independent lipid interactions. Thus, we propose that the competitive effect of Syt11 on Syt1-lipid binding can be effectively negated by Ca²⁺. Consistent with this, 20 μM Ca²⁺ was sufficient for Syt1 to overcome competition from Syt11 (Fig. 2h). Considering that the activation of voltage-gated calcium channels (VGCCs) induces a transient (lasting ∼1–2 ms) local Ca²⁺ concentrations of ∼10–100 μM within nano/microdomains^13,48–50, our results reveal a fascinating inter-switch mechanism between Syt1 and Syt11 on membrane occupancy.

Lipid charge shielding by Ca²⁺ increases surface Zeta potential

Given the well-established Ca²⁺-insensitivity of Syt4 and Syt11, the universal Ca²⁺-inhibition effect observed here is unlikely attributable to the protein itself, but rather suggests a lipid-centered mechanism. We proposed that the Coulomb force between Syts and acidic lipids plays a central role in their interactions. The highly concentrated phosphate groups of PIP2 have been shown to facilitate Ca²⁺ chelation⁵¹, which may be responsible for the Ca²⁺-modulated Syt-membrane interactions. We performed electrophoretic light scattering (ELS) to measure the Zeta potential of 400-nm liposomes in different Ca²⁺ conditions (Fig. 3a). Strikingly, 0.05 mM CaCl₂ was sufficient to increase the Zeta potential of the liposomes, indicating an effective shielding of negative charges of PIP2. This effect is not attributed to altered physical properties of the solution, such as conductivity or viscosity, because 10 mM KCl did not produce similar effect (Fig. 3b–d). These results are consistent with previous reports that high concentrations of Mg²⁺ inhibit Syt1-mediated liposome fusion³⁸. Collectively, these findings suggest that the shielding of surface negative charges is a key contributor to the Ca²⁺-dependent inhibition of Syt-liposome interactions.

Fig. 3. Ca²⁺ shields electrostatic surface potential of acidic membranes.

a Schematic of Zeta potential measurement via electrophoretic light scattering (ELS). b Ca²⁺ increases the surface potential of PC:PS:PIP2 liposomes (400 nm), with near-maximal effects at ~0.1 mM Ca²⁺. c 50 μM Ca²⁺ significantly alters membrane electrostatic states. Numbers of independent biological replicates are labeled on each bar. d Phase plot for Zeta potential calculation of liposomes at varying Ca²⁺ concentrations. Data are presented as the mean + s.e.m., one-way ANOVA and two-tailed Mann–Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. no significant difference. Source data are provided as a Source Data file.

Double-regional interactions confer Ca²⁺-sensitive Syt1-membrane associations

The highly basic surface of C2 domains may provide a large interface for association with acidic membrane. The K-rich motif of Syt1 is composed of 4 consecutive lysine (K) residues in both the C2A and C2B domains. However, in the C2B domain, this motif can be extended to an 8-residue fragment containing ~6 lysine/arginine (K/R) residues, forming a conserved feature shared by both Ca☑- and Ca⌧-Syts (Fig. 4a). To examine the role of these motifs in Ca²⁺-independent lipid binding, we mutated key residues in the K-rich motifs of Syt1-C2A and C2B domains and assessed their binding capabilities using co-LP assays (Fig. 4b). The Syt1-AKA (K189/190/191 to alanine in C2A) showed a minimal reduction in liposome-binding in Ca²⁺-free conditions, while Syt1-BKA (R322N-K324/325A in C2B) exhibited a dramatic decrease in binding. Furthermore, Syt1-ABKA (AKA + BKA), in which both motifs were mutated, completely lost its ability to bind liposomes. Importantly, Ca²⁺-dependent enhancement of lipid binding was retained in all mutants (Fig. 4c), suggesting a double-regional interaction (Ca²⁺-dependent and Ca²⁺-independent) between Syt1 and liposomes. The Ca²⁺-independent lipid-binding was mediated primarily by the K-rich motif in the C2B domain, but not the C2A domain, which is consistent with the electrostatic analyses of Ca☑-Syts (Fig. 1). On the other hand, the Ca²⁺-dependent lipid-binding was mediated by Ca²⁺-binding loops in both C2A and C2B domains, independent of K-rich motifs (Figs. 2d, 4c). Collectively, these findings define both Ca²⁺-independent (through the K-rich motif in the C2B) and Ca²⁺-facilitated (Ca²⁺-binding loops in both C2 domains) Syt1-liposome interactions, highlighting the dual regulatory roles of Syt1 in lipid interactions.

Fig. 4. Syt11 employs a multi-regional redundant membrane-binding mode.

a Conserved K-rich motifs in Syt-C2 domains. b K-rich motif of Syt1 displayed on structures. c Mutations in the K-rich motif of Syt1-C2B dramatically reduce liposome affinity, while C2A plays a minor role, indicating a primarily single-region membrane-binding mode for Syt1 in Ca²⁺-free conditions (n = 3 independent biological repeats). d The C2B, but not C2A, domain of Syt11contributes to liposome binding (n = 3 independent biological replicates). e The C2B, but not C2A, domain of Syt11 contributes to the competition with Syt1 for liposome binding (n = 4 independent biological replicates). f The K-rich motif of Syt11-C2B, and basic regions in loops 1, 2, and 3, displayed on structures. g Mutations in the K-rich motif of Syt11-C2B or deletions of loops 1, 2, 3, or 1 + 3 do not block liposome binding, suggesting redundant interactions (Syt11-BKA: n = 4 independent biological replicates; others: n = 3 independent biological replicates). h–j Systematic single- (h), double- (i), and triple- (j) regional mutations reveal loss of liposome-binding, confirming a multi-regional membrane-binding mode of Syt11 (WT and BLp1_m: n = 4 independent biological replicates; others: n = 3 independent biological replicates). k Mutated residues and regions displayed on Syt11-C2B structures. l Opposite-potential fragments in Ca☑- and Ca⌧-C2B domains highlight a alkaline region in the loop1 of Ca⌧-C2Bs for acidic membrane interaction. Data are presented as the mean + s.e.m.; two-tailed paired Student’s t test (d–e, g) for two-group comparison, and two-way ANOVA (c, e) and one-way ANOVA (g) for multi-group comparison. *P < 0.05, **P < 0.01, n.s. no significant difference. Source data and full scan gels are provided as a Source Data file.

Multi-regional interactions confer redundant binding of Syt11-C2B to the membrane

As for Ca⌧-Syt11, we found that the C2B domain, rather than the C2A domain, is responsible for liposome-binding and Ca²⁺-inhibition. No associations between the C2A and C2B domains were observed (Fig. 4d). Competition assay further demonstrated that only Syt11-C2B, but not Syt11-C2A, competes with Syt1 for liposome binding in Ca²⁺-free conditions (Fig. 4e), underscoring a central role of the C2B domain in Syt11-membrane interactions.

Strikingly, the mutations in the K-rich motif (K340N-R341N-K344/346A) of Syt11-C2B (Syt11-BKA) failed to abolish Ca²⁺-inhibition of its lipid-binding ability (Fig. 4f, g), suggesting additional binding interfaces in Syt11. Given that loop-1 and loop-3 are responsible for Ca²⁺-coordination and membrane insertion of the C2 domains in Ca☑-Syt1, we next investigated analogous basic residues in the loop regions of Syt11-C2B, despite their lack of Ca²⁺-binding ability. Systematic deletion of loops-1, 2, and 3 individually or in combination revealed that none of the Syt11-C2B mutants, including Δloop1 (C2B: aa319-325 to “GS”), Δloop2 (C2B: aa348-357 to “GSG”), Δloop3 (C2B: aa382-390 to “GGS”), or Δloop1 + 3 (Δloop1+Δloop3), showed abolished liposome-binding either (Fig. 4f, g), and all of them retained Ca²⁺-inhibition (Fig. 4g). These results suggest a redundant multi-regional interaction mode for Syt11-C2B in liposome binding.

To comprehensively define the membrane-binding mode of Syt11, we generated 14 Syt11-C2B mutants by permuting and combining the K-rich motif (Kmotif_m) and 3 loop regions (BLp1_m, BLp2_m, and BLp3_m). To preserve chemical properties of the protein, K/R/histidine (H) residues were mutated to N/glutamine (Q) instead of A, D, E, or glycine (G). These mutants were categorized into 3 groups: single-regional, double-regional, and triple-regional mutations. We found that the lipid-binding ability gently decreased but the Ca²⁺-inhibition remained in single-regional mutantss (Fig. 4h). Some double-regional mutants showed a significant decrease in lipid-binding (Fig. 4i), while most triple-regional mutants lost binding capability entirely (Fig. 4j). This defines a redundant multi-regional membrane-binding hierarchy for Syt11: C2B K-rich motif (primary) > C2B loop-1 > C2B loop-3 ≈ C2B loop-2 ≫ C2A (negligible).

Key sites for Syt11-C2B-lipid binding and the mutations examined in this study are summarized in Fig. 4k. Notably, Ca☑-Syt1 and Ca⌧-Syt11 share similar lipid-binding sites, except for a basic “KARH” fragment that confers sub-regional positive potential in loop-1 of Syt11-C2B, whereas the corresponding “EAKN” fragment in Syt1 exhibits an overall negative potential (Fig. 4l). This motif is exactly located in the short fragment representing a major electrostatic potential difference between the C2B domains of Ca☑- and Ca⌧-Syts (Fig. 1d). This suggests a universal yet distinct membrane-binding mode for Ca⌧-Syts, which may confer higher lipid-binding affinity in the absence of Ca²⁺, enabling competitive membrane binding. Together with the universal mechanism of Ca²⁺-inhibition of the K-rich motif, and the Ca²⁺-facilitation of the loop region in Ca☑-Syts, these findings provide a molecular basis for the activity-dependent Ca☑/Ca⌧-Syts inter-switch model mediating precise control of synaptic transmission.

PInP2 docks to specific regions of Syt11-C2B

To investigate the atomic mechanism underlying the interaction between Syt11 and acidic lipids, we applied in silico docking of the headgroup of PIP2 (PInP2) to predict the lipid-binding structure of Syt11-C2B (Fig. 5a). Previous studies have validated that the K-rich motif in the C2 domains of rabphilin and PKC serves as a basic binding pocket for the acidic headgroup of PIP2^52,53. Consistent with these findings, we found that PInP2 can dock into the same pocket of Syt11-C2B with various conformations, forming tight interactions mediated by electrostatic forces (Fig. 5b). We then generated a Syt11-Kmotif_m mutation to neutralize the electric interference of the K-rich motif, and found that PInP2 could still dock to the basic regions of loop-1 and loop-2 in Syt11-C2B through electrostatic interactions (Fig. 5c). However, PInP2 failed to dock automatically to the loop-3 region of Syt11-Kmotif_m or Kmotif_mBLp2_m unless a predefined binding box was used to enforce interaction in this region (Fig. 5d). This suggests a synergistic effect between the K-rich motif and loop-3 in lipid-interactions, which is supported by the observed interaction between the glycerol backbone of PInP2 and K388 in loop-3 (Fig. 5b). Thus, loop-3 may play a more significant role in liposomal or in vivo conditions, where PIP2 is more abundant or clustered. These results define an atomic model of Syt11-lipid interactions, in which PIP2 binds to multiple regions of Syt11-C2B to form a large multi-point interface (Fig. 5e).

Fig. 5. In silico docking of PInP2 to Syt11-C2B.

a Atomic view of Syt11-C2B and PInP2 docking. b PInP2 docks into the K-rich motif’s basic pocket in Syt11-C2B via electrostatic interactions. c PInP2 docks to basic regions on loops 1 and 2 of the Syt11-Kmotif_m mutant. d PInP2 requires a predefined box to dock to the basic region of loop3, suggesting synergistic effects between loop3 and other regions, such as the K-rich motif. e Model of Syt11-C2B interacting with abundant PInP2s in the active zone of the plasma membrane.

Syt11-membrane interaction regulates exocytosis and endocytosis at different stages

We have previously established that Syt11 inhibits both clathrin-mediated endocytosis (CME) and activity-dependent bulk endocytosis during synaptic transmission^4,5,14,19. To investigate the role of Syt11-lipid interactions in synaptic exocytosis and endocytosis, we next performed live synaptophysin-pHluorin (Syp-pH) imaging in cultured primary neurons. Cortical neurons from homozygous floxed Syt11-null mice were infected with synapsin-driven Cre-expressing (Syn-Cre) recombinant adeno-associated virus (rAAV) to generate a knockout (KO) genotype, and full-length wild-type (WT) Syt11 or mutants were over-expressed to rescue the phenotypes (Fig. 6a). A sustained 70 mM KCl depolarization was applied to trigger exocytosis and endocytosis of synaptic vesicles, represented by the rising and decay of Syp-pH fluorescence signals (Fig. 6b, c). The differential of each trace (dF/dt) was calculated to assess the rate of exo-endocytosis (Fig. 6d).

Fig. 6. Syt11-membrane interaction regulates synaptic exocytosis and endocytosis.

a Schematic of Syp–pH imaging in the synaptic bouton of a primary cortical neuron. b 70 mM K⁺ depolarization-triggered vesicle exo-endocytosis events, reflected by Syp-pH fluorescence. Scale bar: 20 μm. c Syp-pH signals show rising and decay phases across genotypes. d Derivative fluorescence signals (dF/dt) indicate the speed of exo-endocytosis. e Syt11-KO shortens acceleration time (from 0.1V_max to V_max) in the rising phase, indicating an inhibitory role of Syt11 in exocytosis. For normalization, only traces with similar F₀ were calculated. f Syt11-KO does not affect V_max. g Syt11-KO reduces τ_decay, indicating endocytosis inhibition. Lipid-binding-deficient mutants fail to rescue Syt11-KO effects, confirming that the regulatory role of Syt11 in exo/endocytosis is mediated by protein-lipid interactions. h Hypothetical model of exo/endocytosis uncoupling and the cytoplasmic Ca²⁺ dynamics. Immediate post-stimulation decay reflects reduction of exocytosis and cytoplasmic Ca²⁺ levels. A Syt1-Syt11 inter-switch may mask Syt11’s influence on V_max, depending on Ca²⁺ fluctuations during neuronal activity. Data are presented as the mean ± 90% CI for (c, d) and mean + s.e.m. for (e–g); number of biological replicates were labeled as N in (e–g); two-tailed Mann–Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. no significant difference. Source data are provided as a Source Data file.

In our experiments, fluorescence changes reflect the collective activity of all vesicles per synapse rather than transient dynamics of individual vesicles. While the rising time depends on stimulation duration, the V_max of the rising phase reflects the maximal exocytosis. To define the early stage of exocytosis, we calculated the rising time from 0.1·V_max to V_max as the acceleration time (t_accel), during which interference from Ca²⁺ accumulation and endocytosis is minimized, allowing assessment of exocytosis driven by the fusion machinery. Strikingly, Syt11-KO reduced the t_accel of Spy-pH, which was fully rescued by WT-Syt11 but not Syt11-Kmotif_m or Syt11-BLp123_m mutant (Fig. 6e), suggesting that Syt11 inhibits exocytosis via membrane-binding during the early stage of synaptic transmission. However, the V_max of the rising phase remained unchanged in Syt11-KO neurons (Fig. 6f), suggesting that the inhibitory role of Syt11 is alleviated or abolished during sustained exocytosis when accumulated Ca²⁺ weakens Syt11-lipid interactions and promotes Syt1 competition. In the decaying phase, the exponentially fitted τ_decay was dramatically reduced by Syt11-KO, which was rescued by WT-Syt11 but not the mutants (Fig. 6g), suggesting that the inhibitory role of Syt11 in endocytosis is also mediated via membrane binding. Therefore, we proposed a hypothetic working model of the molecular processes during Syp-pH imaging (Fig. 6h). The fluorescence change rate (dF/dt) is composed of a positive exocytosis component and a negative endocytosis component. Vesicular exocytosis mainly depends on Ca²⁺, while endocytosis depends on both Ca²⁺ and exocytic extent. Syt11 inhibits exocytosis through competitive binding to the plasma membrane at the early stage of Ca²⁺ rising, but not the high-activity stage following sustained stimulation, when high Ca²⁺ promotes a probable Syt1-Syt11 displacement to temporarily retire Syt11. The immediate and rapid decay of Syp-pH after 70 mM KCl treatment indicates an immediate and rapid restore of cytoplasmic Ca²⁺, removing the Ca²⁺-inhibition in Syt11-lipid binding and thus enabling Syt11 as a brake of endocytosis at this stage.

To investigate the impact of the altered Ca²⁺-affinity of Syt11-S247D on vesicle recycling, we conducted membrane capacitance (Cm) recordings in dorsal root ganglion (DRG) neurons (Fig. S1a). A 200-ms depolarization stimulus induced a Cm jump and decay, validating that knock-down (KD) of Syt11 increased both the speed and the amount of endocytosis (Fig. S1b, c). Notably, the expression of the Syt11-S247D mutant led to a further increase in endocytic speed, but it did not affect the amount of endocytosis. The Cm decay in Syt11S247D-rescued neurons was initially very rapid, reaching a stationary phase more quickly than in control conditions, and the overshoot phenomenon observed in KD neurons was effectively eliminated (Fig. S1a). This behavior can be rationalized by considering the dual-stage function of Syt11-S247D. In the early phase, when Ca²⁺ levels are high, Syt11-S247D can sense Ca²⁺ and accelerate endocytosis similarly to a canonical Ca☑-Syt, such as Syt1. However, as Ca²⁺ levels decline at the later phase, Syt11-S247D reverts to functioning like WT Syt11, thereby rescuing the KD effect. This switch is likely attributed to that the S247D mutant selectively restores the Ca²⁺ affinity of the C2A domain, while leaving the C2B domain unaffected. Consequently, its overall Ca²⁺ sensitivity remains lower than that of canonical Ca☑-Syts like Syt1, rendering it inactive under low Ca²⁺ conditions.

Discussion

Collectively, our findings reveal a multiple-point lipid-binding interface for both Ca☑- and Ca⌧-Syts, but with quite different membrane interaction modes (Fig. 7). Ca☑-Syt1 primarily binds the acidic membrane via the highly basic K-rich motif in the C2B domain in the absence of Ca²⁺, and Ca²⁺ binding to loops 1-3 of both C2 domains dramatically enhances membrane insertion into phospholipid bilayer, leading to membrane deformation. In contrast, Ca⌧-Syt11 adopts a multi-regional interaction involving the K-rich motif and basic sites on loops 1, 2, and 3 of the C2B domain, in which the “KARH” fragment on loop-1 may confer higher lipid-binding affinity in Ca²⁺-free conditions. The surface distribution of these regions might enable Syt11 to prefer curved membranes or stabilize curvature. However, Ca²⁺ does not bind Syt11 but interacts strongly with the acidic headgroups of lipids, effectively weakening the membrane-binding of Syt11 via electrostatic shielding. This mechanism facilitates the Syt1-Syt11 inter-switch in membrane occupancy during synaptic transmission.

Fig. 7. Working model of Syt11 function in neural secretion.

In molecular terms, Syt11 engages in C2B-driven, multi-regional, and Ca²⁺-sensitive acidic membrane interactions. Ca²⁺ inhibits the lipid-binding of Syt11 but promotes Syt1-membrane interactions. Physiological Ca²⁺ levels block the competitive effect of Syt11 with Syt1 in membrane occupancy. Synaptically, the inhibitory role of Syt11 in exo/endocytosis, mediated by membrane interactions, is allocated via a Syt1-Syt11 inter-switch in high-activity, high-Ca²⁺ conditions (e.g., continuous neuronal excitation). This model underscores Syt proteins as direct regulators for vesicle recycling, extends Ca²⁺-sensitivity to Ca⌧-Syts, and highlights Syt family similarity, diversity, and functional connections.

At the synaptic level, the strong membrane-binding ability of Syt11 may contribute to the vesicle docking in resting states (given its strong membrane affinity and our previous report showing that Syt11-KO led to a reduced Cm jump by a 200 ms depolarization pulse stimulus in DRG neurons¹⁹). During low-activity exocytosis, action potential (AP)-induced Ca²⁺ influx triggers membrane penetration of Syt1 and the subsequent vesicle-plasma membrane fusion. Syt11 remains membrane-bound and inhibits early exocytosis dynamics. However, sustained excitation leads to Ca²⁺-accumulation and weakens Syt11’s membrane affinity, enabling free Syt1 to displace Syt11 and thus accelerating fusion. Syt1, a major synaptic vesicle protein (7–8% of total vesicle proteins)³⁰, is well-suited for such displacement under physiological conditions. Syt11 re-establishes its inhibitory roles in endocytosis by re-displacing Syt1 when cytoplasmic Ca²⁺ decays from the steady state. Therefore, the dynamic competitive membrane-binding of Syt1 and Syt11 cooperatively controls both precise excitation-secretion coupling and tight exocytosis-endocytosis coupling during synaptic transmission (Fig. 7).

Although Syt proteins are well-established as primary Ca²⁺ sensors mediating exocytosis and endocytosis, the coexistence of numerous Ca☑- and Ca⌧-Syts in synaptic transmission remains enigmatic. The molecular properties and regulatory mechanisms of Ca⌧-Syts are particularly poorly understood. Here we elucidate the lipid-binding properties and Ca²⁺-sensitivity of both Ca☑- and Ca⌧-Syts, define a multi-point Syt-lipid interface, and propose a Syt1-Syt11 inter-switch model governing the precise control of excitation-secretion coupling and vesicle recycling during synaptic transmission.

Non-Ca²⁺-affinitive but highly alkaline

While the functions of Ca☑-Syts are well-documented in neuronal and non-neuronal cells, Ca⌧-Syts have been considered evolutionarily functional-deficient. However, the greater number of Ca⌧-Syts suggests their underappreciated roles in the brain. For instance, Syt4, one of the earliest identified Ca⌧-Syts, has been reported to inhibit exocytosis and endocytosis with unknown mechanisms^17,18. Syt11, linked to Parkinson’s disease and schizophrenia^4,5,19,54, is gaining attention. We reveal that the high alkalinity of the C2B domain is a common feature of most Syts. Sequence similarity analyses indicate that some Ca⌧-Syts (Syt 4, 8, and 11) share more similarities with Ca☑-Syts, suggesting functional consistency between them in certain aspects. Consistent with this, we found that Syt11 inhibits exo/endocytosis via the membrane interaction with its strongly positively-charged surface, which is similar as Syt1 but with opposing biological effects. Remarkably, Ca⌧-Syts exhibit a higher average surface electrostatic potential and larger basic surface area than Ca☑-Syts, suggesting that they evolved to lose their Ca²⁺-binding ability while gaining higher resting-state membrane affinity.

C2A vs C2B

Most C2B domains are highly alkaline, whereas C2A domains are not, often being acidic. The co-LP assays show weak lipid-binding ability of C2A domains of both Syt1 and Syt11 in Ca²⁺-free conditions. Although the C2A domain is required for Syt1 to support fast synchronous release^55,56 and enhance lipid affinity with Ca²⁺, it is redundant for membrane-binding. In contrast, the C2B domain of Syt1 is capable of self-assembling into ring-like oligomers and plays a key role in vesicle docking, clamping-to-fusion switch, and fusion pore dilation^57–63. However, the tandem C2AB domains of Syt 1/7 bind membranes more strongly than individual domains⁶⁴, and the orientation of C2A and C2B, modulated by the poly-proline linkers, is critical for exocytosis⁶⁵. Moreover, Syt1-C2A prefers vesicle membranes to the plasma membrane, while C2B does the opposite, leading to a Ca²⁺-activated bridging effect^66–70. For Syt11, the C2A domain is dispensable for membrane-binding but may interact with proteins like Vti1a to inhibit spontaneous neurotransmission⁷¹. Some Syt C2A domains (Syts 7, 14, 15, and 16) remain alkaline (Fig. 1), suggesting a higher Ca²⁺-independent lipid-binding probability. In agreement with this, the membrane affinity and Ca²⁺-sensitivity of Syt7-C2A is much higher than that of Syt1^72–74.

Non-Ca²⁺-affinitive but Ca²⁺-sensitive

Neuronal activity is modulated by intracellular Ca²⁺. Ca⌧-Syts, though lacking Ca²⁺-binding ability, remain sensitive to neural activity and thus may play essential roles in synaptic transmission. The non-Ca²⁺-affinity of Syt11 is determined by its amino acid sequence and has been validated by thermal denaturation and nuclear magnetic resonance studies^2,75. Here, we have defined the Ca²⁺-sensitive lipid-binding of Syt11 in vitro and in situ, which is suppressed by high levels of Ca²⁺. A similar inhibitory phenomenon has been observed in early studies of Ca☑-Syts, where GST-Syts were used to pull-down liposomes⁷⁶. This Ca²⁺-sensitivity is shared by Syt4 and Ca²⁺-deficient Syt1 mutants. Membrane-binding and Ca²⁺-sensitivity together define Ca⌧-Syts as fine-tuned regulators during synaptic transmission.

Multi-regional lipid-binding

Screening of the lipid-binding sites on Syt1/11-C2B identify distinct Ca²⁺-dependent binding-modes between them. Syt1 binds to lipid via a K-rich motif in Ca²⁺-free conditions, transitioning to a dual-regional binding mode (K-rich motif and Ca²⁺-binding loops) in the presence of Ca²⁺. In contrast, Syt11 adopts a multi-regional mode (K-rich motif and basic loop regions) in resting states, which is abolished by high Ca²⁺. The conserved K-rich motif, the primary lipid-binding region of both Ca☑- and Ca⌧-Syts in resting states, has been termed as “polybasic strand” or “polylysine patch”. However, only Ca☑-Syts (except Syt7) and the Ca⌧-Syt8 contain the canonical pattern of consecutive four basic residues, while a more general pattern is the basic residue-enriched short strand (~6 K/R enriched 8 aa) when extended to Ca⌧-Syts. Moreover, we use “motif” to hint at a correspondence to the biological function of this fragment, e.g., acidic membrane binding, instead of “strand” or “patch” which are just structural descriptions.

Interplay between Ca☑- and Ca⌧-Syts

We found that Ca²⁺ not only enhances membrane-affinity of Ca☑-Syts but also dose-dependently inhibits the membrane-binding of Ca⌧-Syts. This bi-directional promotion-inhibition effect drives a membrane-occupying switch between Ca☑- and Ca⌧-Syts, in which Syt11 competes with Syt1 under resting conditions but is excluded by Syt1 upon strong neural excitation, providing a fine-tune control of vesicular exocytosis by microdomain Ca²⁺ during synaptic transmission. Post-exocytosis, we have demonstrated the bidirectional role of Syt1 in endocytosis via its membrane remodeling ability¹⁴, as well as the constitutively inhibitory role of Syt11 in endocytosis, with mechanisms remained unknown^4,5,8,19,77. Here, we found that the Ca²⁺-independent membrane-remodeling enables Syt11 as a constitutive brake of endocytosis, while the Ca²⁺-sensitivity of this binding-ability establishes the dynamic modulation of Syt11-membrane attachment and thus the activity-dependent removal of this endocytic brake. Collectively, these findings suggest a Syt1–Syt11 inter-switch working model mediating both the excitation-secretion coupling and the exocytosis-endocytosis coupling during synaptic transmission, emphasizing the cooperation and competition within the Syt family, especially that between Ca☑- and Ca⌧-Syts, in neural or non-neural cells.

Ca²⁺-affinitive phospholipid membrane

The interaction between Ca²⁺/Mg²⁺ and PIP2 has been extensively documented across various chemical studies^51,78–81, in which the phenomenon of PIP2 shielding provides a compelling explanation for the Ca²⁺-mediated inhibition of Syt11-membrane binding observed in the present work. Similar Mg²⁺ shielding effects have been reported to abrogate Syt1-SNARE interactions and thus inhibit vesicle fusion³⁸. Ca²⁺ has been shown to increase the surface potential of PC:PIP2 liposomes more effectively than Mg²⁺, but the concentrations below 0.1 mM show only minimal effects⁸². Therefore, the role of Ca²⁺-lipid interplay has long been under-estimated during synaptic transmission. Here, we found that 0.05 mM Ca²⁺, a physiological level of free intracellular Ca²⁺ during neural activity, is sufficient to induce a dramatic increase in the Zeta potential of PC:PS:PIP2 liposomes. Moreover, the microdomain PIP2-Ca²⁺ coordination could be facilitated by that: (1) PIP2 is concentrated in exo/endocytic hot-spots; (2) the tightly-coupling between Ca²⁺ channel and fusion machinery⁸³ can generate a nanodomain of >100 μM Ca²⁺ near the fusion sites; (3) the transmembrane potential (TMP) inversion post-depolarization also drives the influent Ca²⁺ to diffuse toward the inner membrane. Collectively, we propose a PIP2-Ca²⁺ chelation model for the dynamic modulation of Ca☑-Syts/Ca⌧-Syts inter-switches in membrane occupancy (Fig. 7).

Efficiency of Ca²⁺ inhibition

High concentrations of Ca²⁺ are required to inhibit Syt11 to a significant extent, and the average intracellular Ca²⁺ levels rarely exceed 100 μM. However, it is important to recognize that Ca²⁺ signaling is highly compartmentalization, with local Ca²⁺ concentration vary significantly⁸⁴. This compartmentalization is particularly pronounced in synapses, where voltage-gated calcium channels (VGCCs) directly interact with Syt and SNARE proteins, positioning them close to the fusion site and forming Ca²⁺ micro- and nano-domains⁸⁵. For instance, a 120 μs depolarization simulating an AP in Xenopus neuromuscular junction (NMJ) synapse can generate local Ca²⁺ concentration >150 μM⁸⁶. Similarly, the giant squid synapse can produce Ca²⁺ microdomains ranging from 200 to 300 μM⁸⁷, and local Ca²⁺ concentrations in anuran hair cell synapse has been measured to vary between 10 μM and 1 mM⁸⁸. Therefore, it is reasonable to conclude that Syts, including Syt11, frequently encounter high Ca²⁺ levels, especially under conditions of high synaptic activity.

A primary concern regarding the Ca²⁺-shifted competition between Syt1 and Syt11 is that Syt1 is well-known for its highly Ca²⁺-sensitive membrane binding, which exceeds that of Syt11. To assess the relative contributions of Syt11 and Syt1, we used Ca²⁺-binding deficient Syt1 mutants in Ca²⁺-sensitive competition assays. However, Syt1-D2N and Syt1-D2E, while losing Ca²⁺-promoted binding, exhibited Ca²⁺-inhibition in lipid binding similar to Syt11 (Fig. 2d), complicating the interpretation of our results. Despite these challenges, we propose that the Ca²⁺-promoted binding of Syt1 and the Ca²⁺-inhibited binding of Syt11 work synergistically to drive Ca²⁺-sensitive competition. Even if Syt11’s contribution is relatively smaller, the observed Ca²⁺-sensitive competition supports our model of Syt1-Syt11 inter-switching.

Collectively, this work reveals diverse membrane-binding abilities of both Ca☑- and Ca⌧-Syts, defines a multi-point Syt-membrane interface, pinpoints both Ca²⁺-inhibition and Ca²⁺-facilitation of Ca☑-Syt membrane interactions, demonstrates the Ca²⁺-sensitive membrane-binding abilities of Ca⌧-Syts, and proposes the activity-dependent Syt1-Syt11 inter-switch membrane-occupancy as a promising and fine-tuned working mode controlling both excitation-secretion coupling and exocytosis–endocytosis coupling during synaptic transmission. This model extends the understanding of the Ca²⁺-sensitivity of non-Ca²⁺-affinitive Syts, and highlights the kinetic interplay between Ca☑- and Ca⌧-Syts during neural secretion. Given pathogenic roles of Syt11 in both PD-like neurotoxicity (via the impaired vesicle recycling)^4,14,19,54 and schizophrenia (via the accelerated vesicle recycling)¹⁴, this work underscores critical roles of fine-tuned vesicle recycling in the homeostasis of neurotransmission.

Methods

Ethics statement

For all animal studies, the use and care of animals were conducted in accordance with the guidelines and regulations approved by the Animal Care and Use Committee of Xi’an Jiaotong University (NO.2016-10).

Structural analyses

All structures of C2 domains used in this study were predicted using AlphaFold 2^39,40. Amino acid sequence similarity analyses of all Syts were performed using Clustal Omega⁸⁹. The sequences used from humans are (hSyt 1-17): P21579, Q8N9I0, Q9BQG1, O00445, Q5T7P8, O43581, Q86SS6, Q6XYQ8, Q9H2B2, Q8NBV8, Q9BT88, Q8IV01, Q7L8C5, Q8NB59, Q9BQS2, Q17RD7, and Q9BSW7. The sequences used from rat are (rSyt 1-17): P21707, P29101, P40748, P47861, Q62746, Q62747, Q925C0, O08625, P50232, Q925B4, O08835, P97610, Q925B5, M0R7W7, P59926, A0A0G2K644, and Q62807. The isoelectric points (pI) of C2 domains were calculated using the ExPASy Server⁹⁰. Electrostatic potential calculations for each C2 domain were performed using Protein-Sol⁹¹. Solvent-accessible surface area (SASA) calculations and protein structure mutagenesis were performed using PyMol. In silico docking of the PIP2 headgroup (PInP₂, CID 125105) to Syt11-C2B was performed using DockThor⁹².

Plasmids

For the expression of cytoplasmic tandem C2A-C2B domains of Syt for co-liposome precipitation (co-LP) experiments, the rat Syt1 (rSyt1, aa139-421) and Syt11 (rSyt11, aa154-430) sequences were cloned into the pET-His vector. The isolated C2A and C2B domains of rSyt11 were defined as aa154-282 and aa290-430, respectively, and cloned into the pET-His and pRSFDuet-sumo vectors. For the Syt11-Syt1 competition assays, the rSyt1-C2AB (aa96-421) was expressed from the pGEX-KG vector described in our previous research¹⁴. All mutants were generated using the QuikChange site-directed mutagenesis kit (Stratagene). The mutated sites are as follows: Syt1-AKA (C2A domain: K189/190/191-to-A), Syt1-BKA (C2B domain: R322-to-N and K324/325-to-A), Syt1-ABKA (combination of AKA and BKA mutations), Syt11-BKA (C2B domain: K340-to-N, R341-to-N, and K344/346-to-A), Syt11-Δloop1 (C2B domain: aa319-325 replaced with “GS”), Syt11-Δloop3 (C2B domain: aa382-390 replaced with “GGS”), Syt11-Δloop1 + 3 (combination of Δloop1 and Δloop3 mutations), Syt11-Δloop2 (C2B domain: aa348-357 replaced with “GSG”), and Syt11-Kmotif/BLp1/2/3 m (see Fig. 4k for details). For the K-motif of Syt11, 4 of the 6 K/R residues were randomly selected for mutation. The H/K/R residues mutated in loop1/2/3 were chosen based on their alkaline properties. The primers used to generate these mutants are listed in Table S1. They are available for reuse from the corresponding author upon request.

Protein purification

Constructs were transformed into E. coli BL21(DE3) and cultured in lysogeny broth at 37 °C to an optical density of ~1.0. Protein expression was induced with 0.2 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) at 18 °C for 10 h. Cells were harvested by centrifugation at 8000 × g for 5 min at 4 °C and resuspended in lysis buffer (50 mM HEPES pH 7.4, 500 mM NaCl, 20 mM imidazole, 5% glycerol, 0.2% Triton X-100, 5 mM β-mercaptoethanol, and 1 mM PMSF). After sonication softly on ice and centrifugation at 20,000 × g for 30 min, the supernatant was filtered through a 0.45 μm pore-size membrane. For pGEX-KG-expressed proteins, the supernatant was further incubated with Glutathione Sepharose beads (GE Healthcare) at 4 °C for 2 h. Beads were collected by centrifugation at 400 × g for 3 min and washed thoroughly with 10 μg/mL DNase I and 10 μg/mL RNase I in the binding buffer (50 mM HEPES pH 7.4, 100 mM NaCl, and 1 mM β-mercaptoethanol), then eluted with additional 10 mM reduced glutathione (GSH). For the preparation of tag-free Syt1, thrombin digestion was carried out on the beads at room temperature for 30 min, and then the flow-through containing the cleaved protein was collected. For pET-His and pRSFDuet-sumo-expressed proteins, Ni-NTA Sepharose 6FF beads were used for purification. Proteins on column were washed with additional 40 mM imidazole, then eluted with 400 mM imidazole. The purified proteins were checked via Nanodrop for a 280:260 nm absorbance ratio >1.8 and concentrated to ~10 mg/mL.

Liposome preparation

For co-LP assays, phospholipids (Avanti) were mixed in chloroform with indicated compositions and dried under vacuum. Lipids were resuspended in binding buffer to a final concentration of 0.5 mg/mL, vortexed thoroughly for 5 min, and subjected to 5 freeze-thaw cycles to form large multilamellar vesicles. For Zeta potential measurements, liposomes containing either 2% or 10% PIP2 were used to assess their coordination with Ca²⁺. The liposomes were prepared in ELS buffer (10 mM HEPES pH 7.4, 40 mM NaCl; the lower salt concentration was chosen to facilitate electrophoresis), with 15 freeze-thaw cycles, followed by extrusion through 400-nm polycarbonate filters for another 15 cycles to generate homogenous unilamellar vesicles.

Co-LP assays

Co-LP assays were performed at room temperature (RT) in binding buffer with additional amounts of EDTA, CaCl₂, or KCl. Syt proteins (5 μg) were pre-incubated in binding buffer for 10 min, centrifuged to remove unstable precipitates, and incubated with 10 μg liposomes for 30 min. Liposomes were collected by centrifugation at 20,000 × g for 20 min, and co-precipitated Syt proteins were analyzed via SDS-PAGE and Coomassie blue staining. Band intensities were quantified using Fiji⁹³. Full gel scans and quantified data were supplied in the Source Data file.

Zeta potential measurements

The surface potential of liposomes was measured via electrophoretic light scattering using a Zetasizer Nano ZSE (Malvern). Liposomes were diluted to 0.1 mg/mL in buffers with indicated concentrations of CaCl₂ or KCl, and Zeta potential was calculated using phase analysis light scattering.

Animals and viruses

Floxed Syt11-null (syt11^f/f) transgenic mice were obtained from The Jackson Laboratory (strain B6.129-Syt11tm1Sud/J). C57BL/6J (B6) mice were sourced from Charles River Laboratories. The use and care of animals were approved and directed by the Animal Care and Use Committee of Xi’an Jiaotong University. All animals were housed in the animal facility of Xi’an Jiaotong University, maintained under a 12-h light/dark cycle at 22 ± 2 °C and 40–60% humidity, and provided with ad libitum access to food and water. The animals were finally euthanized with CO₂, followed by cervical dislocation. For each experiment, the postnatal day 0 (P0) to P2 C57 or syt11^f/f mice of both sexes (overall number of 20) were used for primary cortical neuron culture. Wistar rats (male, ~60 g, overall number of 6) were used for DRG culture. AAV2/9 used for primary neuron transfection included: AAV2/9-Psyn-Cre produced by Shanghai Heyuan Biotech, AAV2/9-Pcag-DIO-Syp-pHluorin produced by Hanbio Biotechnology, and AAV2/9-Pef1α-DIO-rSyt11-P2A-mCherry and its mutants (Kmotif_m, BLp123_m) produced by BrainVTA.

Cortical neuron culture and transfection

Primary cortical neurons were isolated and cultured as previously described with few modifications⁹⁴. Anterior cortex tissues from P0 to P2 mice were dissected and cut into ~1 mm pieces in the Hank’s balanced salt solution (HBSS). The tissue fragments were digested with 0.25% trypsin at 37 °C for ~12 min, followed by three rounds of mechanical dissociation using the 1 mL tips on pipettes. Neurons were plated on poly-L-lysine-coated glass coverslips and maintained in DMEM (Gibco) + 10% FBS medium for 4 h, which was then replaced with Neurobasal (Gibco) + 2% B27 + 1% GlutMAX medium. Five micromolar cytosine arabinoside was added at day 2 in vitro (DIV2) to suppress glia growth. Medium were half-changed every 3–4 days. rAAV transfection was performed at DIV6, and experiments were conducted on mature neurons after DIV14.

DRG neuron culture and Cm recording

The culture of dorsal root ganglion (DRG) neurons and the membrane capacitance (Cm) recordings were performed as previously described (PMID: 26589353). Adult Wistar rats (male, ~60 g) were used to isolate DRGs in ice-cold L15 medium (Gibco). The ganglia were treated with a mixture of trypsin (0.3 mg/ml) and collagenase (1 mg/ml) for 40 min at 37 °C. Subsequently, cells were dissociated, collected, and transfected using the Neon^TM transfection system (MPK10096, Invitrogen). Cells were then plated on poly-L-lysine-coated coverslips and maintained in Dulbecco’s modified Eagle’s medium (DMEM)-F12 supplemented with 10% FBS (Gibco). For gene silencing experiments, 24 h post-transfection, the culture medium was replaced with Neurobasal-A medium supplemented with 2% B27 (Gibco), 0.5 mM L-glutamine (Gibco), 10 ng/ml nerve growth factor, and 5 μM cytosine arabinoside. Experiments were performed at day 2 in vitro (DIV2).

To silence the expression of Syt11, the nucleotide target sequences G GCT GAG ATC ACA AAT ATA CG (shSyt11) was synthesized and inserted into the plasmid pRNAT-H1.1-GFP-shSyt11. A random sequence (TTC TCC GAA CGT GTC ACG T), which was predicted to target no genes in human, rat, and mouse cells, was used as a negative control. For rescue experiments, an RNAi-resistant form of rSyt11 was generated by introducing the following silent mutations: G GCG GAAATTACCAAT ATACG.

Cm recordings were performed under the whole-cell configuration using an EPC10/2 amplifier controlled by Pulse software (HEKA Elektronik). The membrane potential was clamped at −70 mV, and pipette resistance was maintained between 3 and 4 MΩ. The external solution contained (in mM) 150 NaCl, 5 KCl, 2.5 CaCl₂, 1 MgCl₂, 10 H-HEPES, and 10 D-glucose (pH 7.4). The intracellular pipette solution contained (in mM) 153 CsCl, 1 MgCl₂, 10 H-HEPES, and 4 Mg-ATP (pH 7.2). A 200 ms membrane depolarization was used to induce exocytosis and endocytosis. Data were collected with PatchMaster v2x90.2 (HEKA EPC10). Offline data analysis was performed using Igor Pro 6.22 software (Wavemetrics), with capacitance traces low-pass filtered at 30 Hz.

Syp-pH imaging

The real-time synaptic fluorescence imaging was performed using total internal reflection fluorescence (TIRF) microscopy. An inverted microscope equipped with a 100 × TIRF objective lens (Nikon ECLIPSE Ti-U; numerical aperture 1.45) was used for image-capturing via an Andor EMCCD camera with NIS-Elements BR software, with an exposure time of 100 ms per frame. Neurons were bathed in resting solution (in mM: 140 NaCl, 2.5 KCl, 2 CaCl₂, 2 MgCl₂, 10 D-glucose, and 10 HEPES pH 7.4), and high K⁺ solution (changed to 72.5 mM NaCl, 70 mM KCl) was locally applied to induce neuronal activity. Time-series signals were analyzed using Fiji 1.54f and Graphpad Prism 8.0 for smoothing and differentiation.

Statistics

All experiments were performed with side-by-side controls and in random order, and were replicated at least three times. No samples that provided successful measurements were excluded from analysis. All the data are presented as the mean ± s.e.m. Statistical comparisons were made with the one/two-way ANOVA, two-tailed Mann-Whitney test or paired Student’s t test as indicated. Normality of the data was tested with the Shapiro–Wilk test, and the equality of variance was determined with Lenene’s test. All the statistical analyses were made using GraphPad Prism 8.0 (Graphpad Software, San Diego, CA), and significant differences were accepted at P < 0.05. Numbers of cells or biological repeats analyzed are indicated in the figures and figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

H.X., C.W., Q.S., and L.J. conceived the study and designed the experiments with the help of J.H., Y.L., C.M., X.K., R.H., and S.Z. X.W., J.Y., J.H., S.H., B.W., Z.L., Y.P., and H.F. performed the experiments and analyses. H.X., C.W., Q.S., L.J., and X.W. wrote the manuscript. All authors reviewed the manuscript and approved its submission.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All data presented in this study are either included in this article and its Supplemental Information or are available upon request to the corresponding author. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-67320-4.