Securing SNAREs for assembly

Abstract

SNARE proteins are essential for exocytosis, mediating the fusion of vesicles with their target membrane. Tethering factors play a key role in chaperoning SNARE assembly; however, the underlying molecular mechanisms are not well-understood. Here, Travis et al. report two crystal structures of a yeast tethering factor, the Dsl1 complex, bound with two SNARE proteins, revealing new insights into how tethering factors bridge vesicles to target membranes, recruit multiple SNARE proteins, trigger their conformational changes, and facilitate SNARE assembly.

Eukaryotic cells use vesicles to deliver molecules to the plasma membrane for secretion, uptake materials from the extracellular environment, and transport cargoes among different membrane compartments intracellularly (1). When these vesicles approach their target membranes, they are first tethered to the membranes by specific tethering factors (2). Once linked to a membrane, two sets of soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) proteins, one located on vesicles (v-SNAREs) and the other on the target membrane (t-SNAREs), couple into four-helix bundles to draw the vesicle membrane close to the target membrane, inducing their fusion. Tethering factors and other specialized proteins regulate vesicle fusion and allow it to take place at a specific time, location, and speed. However, the question of how these proteins cooperate to drive fusion remains a mystery.

Tethering factors are key proteins at the interface of vesicles and their target membranes, and their role in SNARE assembly has been of great interest to researchers studying membrane fusion (2). They are categorized into two classes in eukaryotes: large homodimeric coiled-coil proteins spanning a few hundred nm in length and smaller multisubunit tethering complexes (MTCs) up to tens of nm in length. In 2009, the Hughson group (3) determined the atomic structure of the yeast Dsl1 complex, a relatively small and simple MTC, and suggested that it may mediate the fusion of Golgi-derived vesicles coated with coat protein complex I (COPI) to the ER via binding of ER-associated SNAREs. Since then, studies have shown that MTCs, such as the yeast HOPS complex, bind SNARE and Sec1/Munc18 (SM) proteins (4), but high-resolution structures of MTC bound in complex with SNAREs have been rare and are needed to directly decipher the role of MTCs in SNARE assembly.

The Dsl1 MTC is comprised of three subunits: two parallel legs ∼20 nm in length, formed by Tip20 and Sec39, and Dsl1, the central unit that bridges the two legs (3). In the present study by Travis et al. (5), the Hughson group shows that the Dsl1 “legs” (Tip20 and Sec39) bind the ER t-SNAREs Sec20 and Use1, respectively, by solving the crystal structures of these MTC-SNARE complexes. These structures represent, to our knowledge, the first high-resolution structural model of an MTC bound with two SNARE proteins and provide new insights into how MTCs coordinate SNARE assembly. Separate work revealed the atomic structure of the MTC-SNARE complex yeast Exocyst subunit Sec3 while bound with a single SNARE, Sso2 (6). It is compelling to compare these structures to understand how tethering factors target membranes, bind SNAREs, and promote SNARE assembly.

Many SNAREs contain an N-terminal autonomously folded three-helix bundle called the Habc domain (see Fig. 1). This domain turns to associate with the C-terminal SNARE domain, forming an autoinhibitory conformation (7), and how these closed SNAREs open to allow SNARE assembly is largely unknown. Both the Dsl1 and Sec3 structures reveal unique SNARE binding modes with respect to SNARE assembly. The new structures solved by Travis et al. (5) show that both Tip20 and Sec39 only bind to the Habc domains of their cognate SNAREs, leaving their SNARE domains free to assemble with other SNAREs or SM-family proteins, which chaperone SNARE assembly. Particularly, the binding completely blocks the interface on the Habc domain that the SNARE domain usually binds. Consequently, Dsl1 may remain bound to both SNAREs during or even after SNARE assembly, mediating multiple rounds of SNARE assembly and disassembly. In contrast, Sec3 binds both the Habc domain and SNARE domain, which maintains the overall closed conformation of Sso2 alone (6). However, Sec3 binding induces a small conformational change around the N-terminal SNARE domain, partially opening Sso2. Further SNARE assembly likely displaces Sec3 from the SNARE complex. Therefore, these tethering factors use different strategies to bind and overcome SNARE autoinhibition to facilitate SNARE assembly.

Figure 1.

The Dsl1 complex tethers the COPI-coated vesicle to the ER membrane by simultaneously binding to COPI and two t-SNAREs on their N-terminal Habc domains (depicted as solid ovals). The conformations of the two t-SNAREs prior to Dsl1 binding are not known but are likely autoinhibitory, as is shown for the closed Sec20. Dsl1 binding opens the t-SNAREs, allowing Dsl1 to cooperate with the SM protein Sly1 to chaperone SNARE assembly.

The Dsl1 complex structures also offer a new mechanism by which tethering factors cooperate with SM proteins to chaperone SNARE assembly. SNARE motifs are classified into Qa-, Qb-, Qc-, and R-SNARES, and SM proteins bind their cognate Qa- and R-SNAREs to partially align both SNARE domains on their surfaces as a nucleation site (8, 9). Other SNAREs (Qb- and Qc-SNAREs) then bind the template complex to complete SNARE assembly. It was recently found that the MUN domain of Munc13-1, a tethering factor required for synaptic vesicle fusion, serves as a functional template to coordinate neuronal SNARE assembly (10), in part by interacting with both the cognate Qa- and R-SNAREs. Travis et al. (5) show that the binding of the Qb- and Qc-SNAREs at the feet of Dsl1 well-positions the SNAREs to bind to the potential template complex of Sly1-Ufe1-Sec22 (see Fig. 1). Therefore, the two Dsl1-SNARE structures demonstrate a potentially new means by which tethering factors cooperate with SM proteins to promote SNARE assembly.

Finally, this work reveals an important strategy for tethering factors to bridge membranes. Whereas many tethering factors rely on Rab GTPases to recognize their target membranes (2), the strong and multivalent interactions between Dsl1 and the two t-SNAREs provide a tethering mechanism by binding to SNARE proteins.

Taken together, the new data strongly support that tethering factors orchestrate key fusion machineries to mediate intracellular membrane fusion. This work also paves the way to delineate other open questions. It is interesting to test whether Sec20 and Use1 form autoinhibitory conformations as other Habc-containing SNAREs and, if they do, how their opening transition triggered by Dsl1 promotes SNARE assembly and membrane fusion. Finally, Dsl1 may serve as a model system to elucidate the conformational changes of tethering factors, such as flexing of Dsl1 legs, in membrane tethering and SNARE assembly.