Different regions of synaptic vesicle membrane regulate VAMP2 conformation for the SNARE assembly

Abstract

Vesicle associated membrane protein 2 (VAMP2/synaptobrevin2), a core SNARE protein residing on synaptic vesicles (SVs), forms helix bundles with syntaxin-1 and SNAP25 for the SNARE assembly. Prior to the SNARE assembly, the structure of VAMP2 is unclear. Here, by using in-cell NMR spectroscopy, we describe the dynamic membrane association of VAMP2 SNARE motif in mammalian cells, and the structural change of VAMP2 upon the change of intracellular lipid environment. We analyze the lipid compositions of the SV membrane by mass-spectrometry-based lipidomic profiling, and further reveal that VAMP2 forms distinctive conformations in different membrane regions. In contrast to the non-raft region, the membrane region of cholesterol-rich lipid raft markedly weakens the membrane association of VAMP2 SNARE motif, which releases the SNARE motif and facilitates the SNARE assembly. Our work reveals the regulation of different membrane regions on VAMP2 structure and sheds light on the spatial regulation of SNARE assembly.

Fig. 1. Membrane association of VAMP2 extravesicular domain in mammalian cells by in-cell NMR spectroscopy.

a Domain organization of VAMP2. P-rich NT: proline-rich N-terminal domain. JMD: juxta-membrane domain. TMD: transmembrane domain. R56 is conserved in the VAMP family which forms the zero ionic layer in the SNARE complex. b Scheme of in-cell NMR sample preparation. NMR-visible ¹⁵N-VAMP2(1–96) was delivered into mammalian cells by electroporation. Un-delivered proteins were washed off before NMR signal acquiring. c Estimation of VAMP2 quantity in cells by immunoblotting. Cells prepared for NMR were diluted 10 times before loading on gels. Known concentrations of VAMP2(1–96) proteins were loaded as standards. d Sub-cellular localization of VAMP2(1–96) by fractionation and immunoblotting. Fractions of the total lysate (Lys), cytosol (Cyto) and membrane (Mem) were validated by immunoblotting with antibodies of GAPDH, IRE1α, ACSL4, and VDAC to indicate cytosol, endoplasmic reticulum membrane, plasma membrane-associated membrane, and mitochondrial membrane, respectively. e Sub-cellular localization of VAMP2(1–96) by immunofluorescence staining. F-actin filaments beneath cell membranes were stained by FITC-phalloidin. Nuclei were stained by DAPI. f Overlay of 2D ¹H-¹⁵N NMR spectra of VAMP2(1–96) in solution (black), SH-SY5Y cells (blue) and HEK-293T cells (red). Disappeared crosspeaks in cells were denoted. *The peak was not able to be assigned. g Residue-resolved relative NMR intensity ratios (I/I₀) of VAMP2 (1–96) in cells to that in solution. Domain organization of VAMP2 extravesicular domain is indicated. h Residue-resolved ratios of ¹⁵N transverse (R₂, s⁻¹) to longitudinal (R₁, s⁻¹) relaxation rates of VAMP2(1–96) in HEK-293T cells (red) and in solution (black). Source data are provided as a Source Data file.

Fig. 2. Residue-specific structural change of VAMP2(1–96) in cholesterol up- and down-regulated cells.

a Scheme of experimental design for studying the conformations of VAMP2 in cholesterol-level-regulated cells. Treatments 1 and 2 for increasing and decreasing cellular cholesterol levels were noted in the Methods section. b Overlay of 2D ¹H-¹⁵N NMR spectra of VAMP2(1–96) in cholesterol up-regulated (Chol-up, blue), down-regulated (Chol-down, red) and control (Ctrl, black) HEK-293T cells at the same contour levels. Representative crosspeaks were enlarged on the right. c Relative NMR signal intensity changes of VAMP2(1–96) in Chol-up (blue) and Chol-down (red) cells to those in control cells. Details of data processing are described in the Methods section. Source data are provided as a Source Data file.

Fig. 3. Conformational transition of VAMP2 on membrane subdomains of synaptic vesicles.

a Scheme of experimental design for the structural study of VAMP2 on SV membrane. Upper: SVs isolated from mouse brain were used for NMR titration of VAMP2(1–96) at the physiological ratio. Middle: natural SV membranes were divided into lipid-raft and non-raft membranes by sucrose gradient sedimentation which were analyzed by quantitative lipidomic profiling. Lower: lipid-raft- and non-raft-mimicking vesicles were reconstituted with natural-sourced lipids to titrate VAMP2(1–96). b Residue-resolved NMR signal intensity ratios (I/I₀) of VAMP2(1–96) titrated by SVs to that in solution. The molar ratio of SV to VAMP2(1–96) is indicated. c Left: distribution of endogenous VAMP2 on SV membrane. SV membranes were fractionated into 13 layers which were collected as lipid raft (layer 3) and non-raft (layers 9–12) according to flotillin2 (lipid raft membrane protein) and rabphilin3A (non-raft membrane protein). Right: lipid compositions of lipid-raft and non-raft membranes by MS-based lipidomic profiling. Chol: cholesterol; PC: phosphatidylcholine; PE: phosphatidylethanolamine; PS: phosphatidylserine; PI: phosphatidylinositol; PG: phosphatidylglycerol; PA: phosphatidic acid; Cer: ceramide; SM: sphingomyelin; TG: triacylglycerol. d Residue-resolved NMR signal intensity ratios (I/I₀) of VAMP2(1–96) titrated by lipid-raft-mimicking (blue) or non-raft-mimicking (red) vesicles to that in solution at indicated lipid/protein molar ratios. e Scheme of single-vesicle docking assay. A saturated layer of DiD-labeled (red) t-SNARE vesicles carrying syntaxin-1a and SNAP25 was immobilized on the imaging surface. Free DiI-labeled (green) v-SNARE vesicles, reconstituted with full-length VAMP2, were injected into the system. Green laser illumination imaged the v-vesicles that docked on t-vesicles through SNARE complex formation. f Images on the right are representative fluorescence images of the single-vesicle docking assay. The bar graph on the left shows the numbers of lipid-raft- and non-raft-mimicking v-vesicles that docked on t-vesicles. Error bars are standard deviations from 20 random imaging locations in the same sample channel. *** indicates p-value < 0.001 by One-way analysis of variance (ANOVA) with Tukey Test. Source data are provided as a Source Data file.

Fig. 4. VAMP2 SNARE motif sensors the electrostatic property of membrane surface.

a Residue-resolved NMR signal intensity ratios (I/I₀) of VAMP2(1–96) titrated by acidic liposomes (DOPS, DOPG, liver PI, and egg PA) to that in solution at indicated lipid/protein molar ratios. b Residue-resolved NMR signal intensity ratios (I/I₀) of VAMP2(1–96) titrated by neutral liposomes (DOPC, cholesterol/DOPC (1/1, mol/mol) and DOPE/DOPC (1/1, mol/mol)) to that in solution at indicated lipid/protein molar ratios. c Comparisons of individual lipid classes in lipid-raft and non-raft SV. Error bars are standard deviations from three biological replicates. The data were analyzed by Student’s t-test. The p-values of all lipid subclasses, except for PC and PA, are less than 0.001. n.s. represents not significant. * indicates p-value < 0.05. d Schematic presentation of VAMP2 extravesicular domain on lipid-raft and non-raft membranes. Source data are provided as a Source Data file.

Fig. 5. Hypothetic models of VAMP2 conformations and engagement in SNARE complex assembly for neurotransmitter release.

The extravesicular domain of VAMP2 adopts distinctive conformations in different membrane regions. On the cholesterol-rich microdomains, it is less associated with SV membrane and thus more active to engage in calcium-evoked SNARE assembly and neurotransmitter release. In contrast, in other regions of SV membrane, it tends to hibernate on the membrane with an increased content of α-helical conformation which is relatively inactive in SNARE assembly.