Calmodulin and lipid binding to synaptobrevin regulates calcium-dependent exocytosis

Abstract

Neurotransmitter release involves the assembly of a heterotrimeric SNARE complex composed of the vesicle protein synaptobrevin (VAMP 2) and two plasma membrane partners, syntaxin 1 and SNAP-25. Calcium influx is thought to control this process via Ca(2+)-binding proteins that associate with components of the SNARE complex. Ca(2+)/calmodulin or phospholipids bind in a mutually exclusive fashion to a C-terminal domain of VAMP (VAMP(77-90)), and residues involved were identified by plasmon resonance spectroscopy. Microinjection of wild-type VAMP(77-90), but not mutant peptides, inhibited catecholamine release from chromaffin cells monitored by carbon fibre amperometry. Pre-incubation of PC12 pheochromocytoma cells with the irreversible calmodulin antagonist ophiobolin A inhibited Ca(2+)-dependent human growth hormone release in a permeabilized cell assay. Treatment of permeabilized cells with tetanus toxin light chain (TeNT) also suppressed secretion. In the presence of TeNT, exocytosis was restored by transfection of TeNT-resistant (Q(76)V, F(77)W) VAMP, but additional targeted mutations in VAMP(77-90) abolished its ability to rescue release. The calmodulin- and phospholipid-binding domain of VAMP 2 is thus required for Ca(2+)-dependent exocytosis, possibly to regulate SNARE complex assembly.

Fig. 1. Calmodulin and phospholipid binding to VAMP. (A) GST– VAMP_1–96 was immobilized on the sensor chip of an SPR (BIAcore) apparatus, and 2.5 µM calmodulin was injected into the running buffer in the presence of the indicated concentrations of free Ca²⁺. Calmodulin binding sensorgrams are shown in resonance units (RU). (B) Calcium dependency of calmodulin binding to VAMP determined at plateau levels in (A). (C) [³H]Liposomes containing PS and/or PC as indicated were incubated with GST–VAMP_1–96 or GST immobilized on agarose beads. After washing and centrifugation, bound lipid was evaluated by scintillation counting. Mean of assays in triplicate from three independent experiments. (D) [³H]Liposomes (25% PS, 75% PC) were incubated with GST–VAMP_1–96 in the presence of the indicated concentrations of calmodulin and 1 mM Ca²⁺. 100% = binding in the absence of calmodulin, 0% = binding to GST. Representative of two independent experiments with assays in triplicate.

Fig. 2. Effect of mutations in VAMP_77–90 on calmodulin and lipid binding. (A) Domain structure of VAMP indicating the mutations used throughout this study (CAM, calmodulin-binding domain; TM, transmembrane region). Mutations Q₇₆V,F₇₇W confer resistance to TeNT. The effects of mutations (in bold) within the consensus calmodulin-binding domain (i.e. VAMP residues 77–90) were evaluated in (B) and (C) by SPR using synthetic peptides. (B) Calmodulin was immobilized on the sensor chip, and the ability of peptides to inhibit binding to GST–VAMP_1–96 was measured. The figure represents the mean of data from between two and five independent experiments with each peptide (for K_I values, see text). (C) Liposomes (25% PS, 75% PC) were immobilized on the sensor chip, and direct binding of VAMP_77–90 peptides was evaluated. The figure represents a typical experiment from several independent assays (for mean data, see text).

Fig. 3. Effect of VAMP_77–90 peptides on catecholamine release from chromaffin cells. (A) Cultured bovine chromaffin cells were injected with H₂O (Control), VAMP_77–90 (wild-type) or mutated VAMP_77–90 (W₈₉A,W₉₀A) peptides. Evoked catecholamine release was measured by carbon fibre amperometry following local application of 100 mM KCl. (B) Individual spikes in the 15 s following KCl stimulation were counted in 26 cells injected with the indicated peptides. The illustrated experiment is representative of three totally independent experiments.

Fig. 4. Effect of mutations on tetanus toxin cleavage and SNARE complex stability. (A) In vitro translated [³⁵S]VAMP with the indicated mutations was incubated with 100 nM TeNT light chain for 0 or 2 h. The band corresponding to intact VAMP was monitored by SDS–PAGE and autoradiography. (B) SNARE complexes were assembled on glutathione–agarose beads using in vitro translated [³⁵S]SNAP-25, bacterially expressed syntaxin 1A and GST–VAMP_1–96 with the indicated mutations. After washing, 3% SDS was added and the stability of the SDS-resistant SNARE complex (120 kDa) during a 2 min incubation at the indicated temperatures was monitored by SDS–PAGE and autoradiography.

Fig. 5. Subcellular distribution of VAMP-myc expressed in PC12 cells. The distribution of transfected VAMP-myc was examined in PC12 cells by confocal immunofluorescence microscopy. (A) Wild-type VAMP and (B) W₈₉A,W₉₀A mutants stained with a polyclonal anti-myc antibody (green). (C and D) Staining of the cells illustrated in (A) and (B), respectively, with a monoclonal anti-synaptotagmin antibody (red). Co-localization is shown in yellow. Scale bar = 10 µm.

Fig. 6. Effects of a calmodulin antagonist and mutations in VAMP_77–90 on calcium-dependent exocytosis in PC12 cells. (A) PC12 cells, transfected with a plasmid encoding hGH, were permeabilized with digitonin. Following the indicated pre-incubation times (±300 nM TeNT light chain), Ca²⁺-dependent hGH release was evoked by 10 µM Ca²⁺, and compared with basal release (0 µM Ca²⁺). The hGH released over 4 min was assayed and is represented as a percentage of total cellular hGH content. The graph is representative of three independent experiments with duplicate data points. (B) Intact cells were treated with the indicated concentrations of the calmodulin antagonist ophiobolin A. After washing and permeabilization, hGH release was measured as in (A) and plotted as Ca²⁺-dependent release (per cent evoked minus basal release). The graph is representative of three independent experiments with duplicate data points. (C) PC12 cells were transfected with a plasmid encoding hGH (vector) or a single plasmid encoding both hGH and VAMP (wild-type or with the indicated mutations). After permeabilization, cells were treated for 10 min with or without TeNT (as in A) and hGH release was assayed. Results are presented as the percentage of Ca²⁺-dependent release that persists after TeNT treatment. Data were pooled from three or four independent experiments with duplicate assays.