Complexin Suppresses Spontaneous Exocytosis by Capturing the Membrane-Proximal Regions of VAMP2 and SNAP25

Abstract

The neuronal protein complexin contains multiple domains that exert clamping and facilitatory functions to tune spontaneous and action potential-triggered synaptic release. We address the clamping mechanism and show that the accessory helix of complexin arrests assembly of the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex that forms the core machinery of intracellular membrane fusion. In a reconstituted fusion assay, site- and stage-specific photo-cross-linking reveals that, prior to fusion, the complexin accessory helix laterally binds the membrane-proximal C-terminal ends of SNAP25 and VAMP2. Corresponding complexin interface mutants selectively increase spontaneous release of neurotransmitters in living neurons, implying that the accessory helix suppresses final zippering/assembly of the SNARE four-helix bundle by restraining VAMP2 and SNAP25.

Keywords: SNARE; complexin; cross-linking; membrane fusion; neurotransmission; synaptotagmin; syntaxin.

Figure 1. Cpx-Syt1 Primed SNAREpins and Docked Vesicles

(A) Crystal structure of the membrane-distal Cpx-Syt1-SNARE interface (accession number PDB: 5W5C; modified from Zhou et al., 2017). The positions of the hydrophobic layers and the central ionic 0 layer within the SNARE motifs are indicated below. The second C2B binding site facing SNAP25 is not shown. (B) Slice through a cryo-tomogram of docked SUVs, whose fusion can be triggered by Ca²⁺. GUVs containing full-length t-SNARE complexes (syntaxin1/SNAP25) were mixed with SUVs containing Syt1 and the v-SNARE VAMP2 in the presence of WT CpxII. Samples were pre-incubated for 30 min on ice to efficiently accumulate the Cpx-stabilized prefusion intermediate. (C) Magnification of the area outlined in (B). Arrows indicate the position of putative SNAREpins and associated proteins.

Figure 2. Mapping Interactions of the CpxII Accessory Helix with the Fusion machinery by Site- and Stage-Specific Cross-linking

GUVs containing the full-length t-SNARE complex (syntaxin1/SNAP25) were mixed with SUVs containing Syt1 and the v-SNARE VAMP2 in the presence of WT CpxII or CpxII BPA mutants. Samples were preincubated for 30 min on ice to accumulate docked SUVs linked to the GUVs by Syt1 and trans-SNARE complexes. UV irradiation of the reaction mix was performed before (pre-fusion) and after triggering fusion with 100 μM calcium at 365 nm for 15 s on ice. Cross-link products were analyzed by western blotting using the indicated antibodies. Colored numbers indicate CpxII BPA mutants that show prominent cross-links to SNAP25 (blue), VAMP2 (red), and synaptotagmin (gray). Positions of identified cross-link products are indicated by the molecular identity. (A) Cross-link products identified at the prefusion stage (left panels). A view along the axis of the CpxII accessory helix shows interactions with SNAP25 (blue) and VAMP2 (red) (right panel). The membrane-proximal regions of VAMP2 and SNAP25 are depicted as cylinders because their prefusion structures are not known but may form α-helical structures. (B) Cross-link products identified at the postfusion stage and structural model.

Figure 3. Site-Specific Arrest of the Fusion Machinery by Scanning BPA Cross-linking of the CpxII Accessory Helix

Syntaxin1/SNAP25 GUVs were mixed with Syt/VAMP2 SUVs in the presence of WT CpxII or the indicated CpxII BPA mutants and incubated for 30 min on ice to dock vesicles. The reaction mixes were irradiated at 365 nm for 15 s on ice (control reactions without UV irradiation). Subsequently, fusion kinetics were recorded at 37°C for 1 min in the absence of Ca²⁺, and the measurement was continued for another 2 min after injection of 100 μM free Ca²⁺ to trigger fusion. (A) Lipid-mixing kinetics of WT CpxII and distinct CpxII BPA mutants with or without UV irradiation. CpxII R48BPA impairs SNARE complex binding, resulting in loss of the clamp and in an elevated starting signal. Error bars indicate SEM (n = 3). (B) BPA cross-link scan of the complete accessory helix and the effect on Ca²⁺-triggered fusion (signal change 10 s after Ca²⁺trigger as illustrated in A WT). Blue and red bar graphs indicate crosslinks to SNAP25 and VAMP2, respectively. Error bars indicate SEM (n = 3).

Figure 4. CpxII Quadruple Mutants of the SNAP25- and VAMP2-Binding Regions Selectively Increase Ca²⁺-Independent Fusion of Proteoliposomes and Spontaneous but Not Evoked Neurotransmitter Release in Neurons

(A) t-SNARE-GUVs were mixed with Syt1/VAMP2 SUVs in the absence or presence of WT CpxII or the indicated mutants (CpxII-SNAP25-binding mutant: D27R, A30R, Q31E, E34R; CpxII-VAMP2 binding mutant: K33E, R37E, A40K, Q44E). Samples were pre-incubated for 5 min on ice, and then fusion kinetics were recorded at 37°C for 2 min in the absence of Ca²⁺. Fusion was triggered by injection of 100 μM free Ca²⁺. Error bars indicate SEM (n = 3). (B) Spontaneous release activity as determined by mean mEPSC frequency. n = 43 (Cpx TKO), n = 44 (WT CpxII), n = 43 (CpxII-VAMP2-binding mutant), n = 42 (CpxII-SNAP25-binding mutant) Error bars indicate SEM. **p < 0.01, ***p < 0.001, ****p < 0.0001. (C) Cpx TKO glutaminergic neurons were transduced with a lentivirus containing WT CpxII or the mutants. Shown is analysis of evoked responses by mean EPSC amplitudes. n = 47 (Cpx TKO), n = 45 (WT CpxII), n = 47 (CpxII-VAMP2-binding mutant), n = 42 (CpxII-SNAP25-binding mutant).

Figure 5. Model of How the Cpx Accessory Helix Clamps Fusion at the Synapse

The central helix (CH) of Cpx stabilizes a partially zippered SNARE complex, and the accessory helix (AH) binds the membrane-proximal C-terminal ends of SNAP25 and VAMP2, preventing further SNARE complex zippering/assembly/membrane fusion. SNAP25-N and SNAP25-C indicate the first and second SNARE motif, respectively.