Is it Zippered? Does it Flare? That Darn Complexin Clamping SNARE

SNARE proteins form the molecular machine that drives vesicle-membrane fusion within cells, which is critical for exocytosis and neurotransmitter release. The name “SNARE” originally came from four layers of acronyms: SNAp REceptor (SNARE); Soluble NSF Attachment Protein (SNAP); NEM-Sensitive Fusion (NSF); and N-Ethylmaleimide (NEM). Thus, it is easy to see why SNARE has become simply a name for a family of fusion proteins rather than the term “soluble n-ethylmaleimide sensitive fusion protein attachment protein receptor”.

The structures of SNARE proteins are more dynamic than most proteins because the very function of SNARE proteins (bringing membranes together for exocytosis) requires major structural changes that must be undone during a recycling step. SNARE proteins are characterized by all having SNARE domains of ∼60 amino acids which can form an α-helix with hydrophobic faces (1). The helix has at least 16 loops (numbered layer −7 through layer +8) that typically interact with hydrophobic amino acids in corresponding layers of other SNAREs. When four SNARE domains come together, favorable hydrophobic interactions quickly zipper them together, forming a coiled-coil known as the SNARE complex (2,3). Before zippering, SNARE domains are less structured. The four helixes come from different proteins, synaptobrevin in the vesicle membrane (v-SNARE) and two other proteins in the target membrane (t-SNAREs). The t-SNAREs include members of the syntaxin family (containing one SNARE and one trans-membrane domain), and the SNAP-25 family (with two SNARE domains and no trans-membrane domain).

Normally, there is a strong repulsive force preventing two membranes from approaching closer than 5 nm; SNARE proteins drive fusion by using the energy released during complex formation. The topography of the proteins and the stability of the SNARE complex have led to the SNARE-hypothesis of vesicle-membrane fusion. Namely, as SNAREs zipper together to form the SNARE complex, opposite membranes are brought close together and forced to fuse (4).

SNARE-driven fusion is a ubiquitous process that may be regulated or constitutive. Constitutive exocytosis occurs in such common cellular processes as depositing additional membrane for cell growth and filopodia extension. For regulated exocytosis, a different set of SNARE and accessory proteins are used (e.g., SNAP-25, not SNAP-23) (1). However, it is the regulated version, specifically the regulation by calcium for neurotransmitter release in nerve cells, which initially received most attention by scientists. After vesicles dock, regulation requires additional proteins that block or stall formation of the full coiled-coil SNARE complex. Synaptotagmin, complexin, and SM proteins are widely accepted as key proteins in the regulation of SNARE-driven fusion (5–7). Synaptotagmin is a calcium-binding protein and thus links fusion with calcium entry into cells. The role of complexin (originally called synaphin (8)) is incompletely understood. Studies in mouse (9), zebra fish (10), Drosophila (11), and Caenorhabditis elegans (12) show an increase in nonregulated (spontaneous) fusions when complexin is mutated or absent, suggesting that complexin helps with regulated release.

A popular model for complexin function is that it blocks fusion by clamping the partly formed SNARE complex (11–13). In this model, complexin binds to the pre-complex (formed of three SNARE domains from syntaxin and SNAP-25) and thus interferes with the binding of the v-SNARE, synaptobrevin, at layer 2. The clamp leaves the N-terminal of synaptobrevin’s SNARE domain wound with the pre-complex, but the C-terminal layers are displaced and flared out. Thus, the vesicle is tethered to the membrane, but not brought close enough to fuse. An alternate version of the clamp model posits that complexin clamps SNARE complex formation by cross-linking two adjacent pre-complexes (14). This model was inspired and is consistent with crystallography data showing stable cross-links. Both versions of this clamp model stipulate that complexin adds a barrier to the formation of the SNARE complex (Fig. 1, left).

Figure 1.

Three models for regulation of SNARE-driven exocytosis. The slide represents energy released during formation of the SNARE complex that brings the vesicle close to the cell membrane where it fuses (shadow, bottom of slide). In one model (left), complexin regulates fusion by partly blocking (clamping) synaptobrevin from contributing to the complex. The model of Jahn and Fasshauser (middle) argues that complex formation cannot be blocked midway and that regulation is by a gating protein before synaptobrevin engages the pre-complex (15). Bykhovskaia et al. (16) propose a variation on these models. In their model (right), there is not a block in the middle, but, instead, complexin stabilizes a state near the end of SNARE zippering.

Jahn and Fasshauer (15) recently proposed an alternate model. They argue that the binding of synaptobrevin to the pre-complex is strong enough to displace any clamping agent and that regulation of fusion must occur by a gating protein at an earlier step. In this model, some protein, such as synaptotagmin, holds the vesicle away from the membrane to prevent synaptobrevin from interacting with the pre-SNARE complex. After calcium entry, synaptotagmin changes conformation and allows initiation of SNARE interaction leading directly to fusion (Fig. 1, middle).

In this issue, Bykhovskaia et al. (16) consider and test these models using molecular dynamics simulations (MD) and electrophysiology. Their MD data suggest a late-clamp model; where complexin stabilizes a partly zippered SNARE complex with just layers 7 and 8 flared out (see Fig. 4 in Bykhovskaia et al. (16)). This would act as a dip in the energy profile near the end of zippering that could stabilize the nearly formed complex (Fig. 1, right). This is in contrast to the original clamp model where complexin blocks formation in the middle stages of zippering, between layers 1 and 2. Their MD data are also consistent with part of the model of Jahn and Fasshauer (15), by showing that it is unlikely that electrostatic forces (associated with membrane-membrane repulsion) are large enough to hinder SNARE zippering before layer 7. Bykhovskaia et al. (16) test their model electrophysiologically by measuring regulated versus spontaneous neurotransmitter release in Drosophila. Wild-type release is compared to release with a t-SNARE mutation that alters the amino acid at layer 7. MD modeling suggest that such a mutation would weaken interactions between complexin and synaptobrevin and so would be expected to have a similar phenotype to a complexin null mutant (increased spontaneous release), which they observed. However, no data are presented that eliminates either of the other models; instead, the late-clamp model contains aspects of both models, and it is more consistent with most data. Distinguishing between contemporary models will require additional experimentation including effects of point mutations in SNAREs, complexin, and synaptotagmin.