A new life for an old pump: V-ATPase and neurotransmitter release.

Neurons fire by releasing neurotransmitters via fusion of synaptic vesicles with the plasma membrane. Fusion can be evoked by an incoming signal from a preceding neuron or can occur spontaneously. Synaptic vesicle fusion requires the formation of trans complexes between SNAREs as well as Ca(2+) ions. Wang et al. (2014. J. Cell Biol. http://dx.doi.org/jcb.201312109) now find that the Ca(2+)-binding protein Calmodulin promotes spontaneous release and SNARE complex formation via its interaction with the V0 sector of the V-ATPase.

Evoked release of synaptic vesicles occurs after an action potential (or series of action potentials), whereas spontaneous release occurs in the absence of a presynaptic action potential (Kochubey et al., 2011). Both forms of release require the formation of trans complexes between SNARE (“soluble NSF attachment receptor”) proteins, one in the vesicle and the other anchored in the plasma membrane. SNARE proteins exist in two cognate forms: R-SNAREs (e.g., Synaptobrevin, which is located in the synaptic vesicles) and Q-SNAREs, which can contain up to three subunits (e.g., Syntaxin1A and SNAP-25, located in the plasma membrane). Both forms of neurotransmitter release also require Ca2+ ions. Synaptotagmins act as a Ca2+ sensor for evoked release. Little is known about how the Ca2+ dependence of spontaneous release is generated.

The V-ATPase is a proton pump consisting of two subcomplexes that can reversibly dissociate from each other: The membrane-integral V0 sector and the peripheral V1 sector, which carries the ATPase activity. V0 contains a cylinder of proteolipids (consisting of highly homologous isoforms of subunit c); the large, membrane-integral subunit a; and the peripheral subunit d and subunit e (Fig. 1 A). V-ATPase provides the electrochemical potential that contributes to loading of the secretory vesicles with neurotransmitters. V-ATPase–dependent lumenal acidification can influence protein trafficking, e.g., between endolysosomal compartments (Wada et al., 2008; Huotari and Helenius, 2011). In addition, in vivo evidence points to a physical role of V0 in exocytosis and membrane fusion, which is independent of proton pumping. Observations that have uncovered a function for V0 in exocytosis and membrane fusion stem from regulated secretion in both Drosophila melanogaster (Hiesinger et al., 2005) and mammalian cells (Sun-Wada et al., 2006; Di Giovanni et al., 2010), secretion of multivesicular bodies in Caenorhabditis elegans (Liégeois et al., 2006), phagosome–lysosome fusion in the zebrafish Danio rerio (Peri and Nüsslein-Volhard, 2008), and vacuole fusion in yeast (Bayer et al., 2003; Strasser et al., 2011). A recent study challenged a physical role for V0 in vacuole fusion in vivo (Coonrod et al., 2013); however, the main assay in that study suffered from a conceptual flaw, i.e., it could not measure vacuole fusion but only biosynthetic transport of the indicator proteins to the vacuoles. At the same time, another study provided a compelling in vivo demonstration that a lack of vesicle acidification impairs exocytosis, but that it does so by reducing the pool of free V0 sectors and thus impeding their physical function in the fusion process (Poëa-Guyon et al., 2013).

Figure 1.

The V-ATPase V (A) The V-ATPase is composed of a peripheral sector V1 (blue, 10 subunits) and a membrane integral sector V0 (red and yellow). V0 contains a cylinder of proteolipids (subunits c, yellow), the proton-conducting subunit a, and subunits d and e (red). V1 and V0 can dissociate from each other in a regulated and reversible fashion. This equilibrium can be influenced by SNAREs. B and C show two roles of V0 in membrane fusion, which are not mutually exclusive. The emphasis is on subunit a or subunits c, respectively. (B) Regulation of SNARE complex formation by subunit a. Binding of subunit a to Ca2+–Calmodulin (CaM, green) alleviates the block of SNARE complex assembly caused by the interaction of subunit a with the Q-SNARE. (C) Model of how proteolipids (yellow) might enhance the capacity of SNAREs to stimulate lipid mixing. Stimulated by their interaction with trans-SNARE complexes, the proteolipids undergo a conformational change, which might expose hydrophobic surfaces between the proteolipid subunits and facilitate lipid reorientation and membrane fusion (Strasser et al., 2011). V0 subunits e and d are not shown because there are no data implicating them in fusion. V0 subunit a is not shown for the sake of clarity.

V0 subunits interact with Calmodulin (Peters et al., 2001; Zhang et al., 2008) and with Q- and R-SNAREs (Galli et al., 1996; Peters et al., 2001; Takeda et al., 2008; Di Giovanni et al., 2010). In this issue, Wang et al. provide compelling evidence that Calmodulin regulates SNARE complex assembly via V0. They used the V0 subunit a from Drosophila (v100) with point mutations in its Calmodulin binding site to selectively disrupt the interaction of Calmodulin and v100. This allele (v100WFI) rescues most defects resulting from the loss of v100 (Hiesinger et al., 2005; Williamson et al., 2010), notably endolysosomal acidification and endolysosomal protein sorting. Thus, v100WFI retains its functionality as part of the V-ATPase proton pump. v100WFI rescues evoked neurotransmitter release but it diminishes spontaneous transmitter release by >90%. An extensive biochemical characterization revealed that v100 disrupts the assembly of Q-SNARE complexes by competitively binding to Syntaxin1A and SNAP-25. Ca2+–Calmodulin can disrupt the competitive interactions of v100 with the SNAREs, permit the Q-SNAREs to form a complex, and also incorporate the R-SNARE VAMP and catalyze fusion. These results indicate that v100, when associated with Calmodulin, can serve as a regulator of SNARE complex formation (Fig. 1 B).

This finding adds a novel aspect to the role of V0 in exocytosis. However, other known effects and interactions of V0 subunits illustrate that regulating SNARE complex formation cannot be the only way in which V0 influences membrane fusion and exocytosis. If V0 subunit a were only a negative regulator of SNARE complex formation, as suggested by the results of Wang et al. (2014), its deletion should leave fusion intact or even stimulate it. The opposite is the case, which suggests that V0 also serves to promote fusion. This is also evident from the effects of mutations in other V0 subunits, such as the central ring of proteolipids (subunit c), which interfere with fusion. Proteolipid rings can adopt at least two conformations (Clare et al., 2006) and they can form Ca2+-inducible pores in the membrane that are permeable to hydrophilic molecules (Morel, 2003). Single amino acid substitutions in proteolipid transmembrane domains and proteolipid fusion proteins impede lipid mixing, but, unlike v100WFI, they permit the formation of normal levels of trans-SNARE complexes (Strasser et al., 2011). These effects could only be explained by a conformational change in V0. Therefore, the current working model postulates that V0 proteolipids can exist in at least two conformations: one conductive to fusion and one supporting V-ATPase assembly and proton pumping (Strasser et al., 2011; Poëa-Guyon et al., 2013).

According to this model, the molecular interactions of V0 with SNAREs (Peters et al., 2001; Di Giovanni et al., 2010) might influence V0 conformation and/or compete with the attachment of V1 to V0 (Fig. 1 A). The recruitment of V0 into the V1–V0 V-ATPase holoenzyme can inhibit exocytosis by depleting the pool of free V0 sectors that are necessary to support fusion (Poëa-Guyon et al., 2013). Experimental observations from yeast support a role for SNAREs in regulating this pool of free V0: Deletion of the vacuolar R-SNARE NYV1 (a Synaptobrevin homologue in yeast) increases V0–V1 association, and NYV1 overexpression reduces it (Strasser et al., 2011). Furthermore, structural data suggest that there is substantial space between the subunits of the proteolipid cylinder that might be invaded by lipids (Clare et al., 2006). These data gave rise to a hypothesis explaining how V0 might promote membrane fusion. In this hypothesis, it is assumed that SNAREs destabilize the V0–V1 interaction and favor a V0 conformation that supports fusion by allowing lipids to invade the space between subunits of the proteolipid cylinder, facilitate their reorientation, and thereby promote the merger of the membrane leaflets (Fig. 1 C). The role of v100 in SNARE complex assembly that Wang et al. (2014) describe is not mutually exclusive with a role of V0 in lipid reorientation (Fig. 1, B and C). It adds an interesting new aspect of V0 function and suggests that V0 is intimately linked with the membrane fusion apparatus. This illustrates that much remains to be discovered before we will fully understand the mechanisms by which V0 subunits regulate and promote the different stages of the fusion process.