View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

R258 Dispatch

Membrane transport: Take your fusion partners Michael J. Clague

Recent studies of how vesicles are targeted to fuse with SNARE complex can be formed in trans — that is, specific membranes inside cells highlight a role for between cognate v-SNAREs and t-SNAREs on distinct extended coiled-coil proteins in tethering partner membranes — the SNARE families are implicated in the membranes prior to formation of the ‘SNARE complex’ selection of fusion partners. The extent to which SNAREs that mediates the fusion reaction. The tethering protein are responsible for the initial recognition and specificity of is recruited to membranes by a Rab family GTPase. interactions between vesicles has, however, been contro- versial. Doubts have arisen because some SNAREs, like Address: Physiological Laboratory, University of Liverpool, Crown Street, Liverpool, L69 3BX, UK. some Rabs, can act at more than one transport step. E-mail: [email protected] Furthermore, some individual SNAREs are promiscuous in their partnerships, for example the yeast v-SNARE Vti1 Current Biology 1999, 9:R258–R260 has been shown to pair with five of the eight yeast http://biomednet.com/elecref/09609822009R0258 t-SNAREs [3]. © Elsevier Science Ltd ISSN 0960-9822 Recent studies using a cell-free assay for homotypic The distinct identity of the various subcellular compart- endosome fusion, and a permeabilised cell assay for fusion ments is dynamically maintained by a balance of vesicle of yeast endoplasmic reticulum (ER)-derived vesicles with formation and consumption. These processes occur in cis-Golgi elements [4], have highlighted a role for two pro- many different contexts inside a cell, and together they teins, early endosomal antigen 1 (EEA1) and Uso1, play a crucial role in sustaining cellular organization. It is respectively, in mediating the initial interaction between generally believed that the basic mechanisms of vesicle the partner membranes. These proteins both have exten- formation and consumption are conserved, with organelle- sive coiled-coil regions, which form extended flexible-rod specific molecular components being drawn from large structures, similar to the tail region of myosin. Owing to protein families. Recent investigations have implicated their extended shape, EEA1 and Uso1 have been pro- extended coiled-coil proteins in mediating the initial posed to act as membrane ‘tethers’, which may confine attachment between membranes destined to fuse. In the search for trans-SNARE partners to the tethered concert with small GTPases of the Rab family and mem- partner membranes. In each case, this role has been estab- brane receptors known as SNAREs, the coiled-coil pro- lished by experimental approaches that allow the dissec- teins determine the specificity of vesicle fusion events. tion of membrane attachment from membrane fusion.

Most intracellular fusion events are known to require the Barlowe and co-workers [4] developed a simple centrifu- common components ‘N-ethyl maleimide-sensitive gation assay for vesicle attachment on the yeast ER-to- factor’ (NSF) and ‘α-soluble NSF attachment protein’ (α- Golgi transport pathway, based on the fact that SNAP) [1]. These factors have been shown to interact ER-derived vesicles produced from semi-intact cells are with a 7S membrane receptor complex comprising ‘SNAP freely diffusible, so that following incubation under condi- receptors’ (SNAREs) which can be disassembled as a tions where fusion is inhibited, only vesicles attached to result of ATPase activity of the NSF component. This target membranes will pellet. They found that vesicle interaction was initially described for SNAREs involved attachment could still be observed, even when fusion was in the consumption — or exocytosis — of synaptic vesi- inhibited by withholding LMA1, an essential cytosolic cles: synaptobrevin, a vesicle-associated or v-SNARE, fusion factor, or using strains of yeast carrying loss-of-func- and the two target-membrane-associated or t-SNAREs, tion mutations of SNARE proteins. This indicates that syntaxin and SNAP-25, which interact to form a four- SNARE complex formation is not required for initial helix bundle. These are now known to be representatives vesicle attachment. Vesicle attachment was found to of protein families for which individual members act at require the cytosolic factor Uso1 and the small GTPase discrete sites. Ypt1, a member of the Rab family [4] (that these proteins have a close relationship on the vesicle consumption As well as breaking pre-assembled SNARE complexes, pathway had already been indicated by genetic analysis NSF and α-SNAP serve to reconfigure t-SNAREs in some [5]). Removal of Ypt1 from membranes led to a corre- way, such that they can participate in fresh SNARE sponding loss of Uso1, although a direct interaction complex assembly. This action, referred to as ‘priming’, is between the two proteins has not so far been demon- essential for vesicle fusion and, at least in some systems, strated [4]. Ypt1 may either bind directly to Uso1 or regu- can be completed prior to vesicle attachment [2]. As a late availability of an alternative receptor binding site. Dispatch R259

Zerial and colleagues [6] have developed a simple Figure 1 microscopic assay for endosomal aggregation. They prepared endosomes specifically labelled with fluorescently-tagged ATP ADP + Pi transferrin, and then incubated the labelled endosomes under conditions in which fusion does not occur (specifi- cally, SNARE complex assembly was impaired by the NSF presence of an excess of a dominant-negative mutant form of α-SNAP). They found that, under these conditions, SNAP (a) EEA1-dependent aggregation of labelled vesicles could still be observed, indicating that it is EEA1, rather than SNAREs, that mediates initial endosome attachment [6]. Vesicle Vesicle GDP GDP EEA1 is a specific early endosomal marker protein, which equilibrates between membrane and cytosolic fractions and is required for endosome fusion [7,8]. Structurally the GTP GTP protein consists of an extended coiled-coil, flanked by two exchange hydrolysis zinc-finger domains. The carboxy-terminal zinc finger belongs to the family of FYVE domains, which bind (b) specifically to the lipid phosphatidylinositol 3-phosphate GTP GTP (PI(3)P) [9]. EEA1 additionally contains two binding sites for GTP–Rab5, one at each end of the protein [7]. Both PI(3)P binding and Rab5 binding play a part in the local- ization of EEA1 inside cells.

In addition to EEA1, Zerial and colleagues [6] have (c) identified 21 other proteins that bound specifically to a GTP–Rab5 affinity column. Fractionation of this protein pool, however, revealed that EEA1 is not only required to support both endosome attachment and fusion, but that it can do so in the absence of other cytosolic proteins. This raises the question, what are the other 21 proteins good for? Some may be regulators of the Rab5 nucleotide cycle, whilst others may be effectors of further possible Rab5 functions, such as controlling endosome–microtubule inter- Membrane fusion actions. It is a reasonable bet that some may be alternative ‘tethering factors’ that mediate endosome fusion events v-SNARE n-SEC1 homologue distinct from the ‘homotypic’ fusion governed by EEA1. t-SNARE Tethering protein Rabaptin-5, for example, also contains an extended coiled- coil region and has a Rab5-binding site at one end and a Rab proteins Rab4-binding site at the other [10]; perhaps Rabaptin-5 Current Biology mediates tethering between a subset of early endosomes. A generalised scheme for the tethering and fusion of intracellular So where do the SNAREs fit in? Certainly trans-SNARE vesicles. The model is a composite, based on data obtained with a complex formation is required for fusion and lies down- number of experimental systems, principally ER-to-Golgi transport, trans-Golgi-to-endosome transport, homotypic vacuole fusion in yeast stream of vesicle tethering. That SNAREs interact and homotypic endosome fusion in mammalian cells (see text for directly with a tethering protein is suggested by recent details). (a) α-SNAP recruits NSF to a syntaxin homologue (t-SNARE), studies of genes required for vesicle transport between the which is then ‘primed’ in an ATP-dependent reaction; Rab proteins trans-Golgi network and endosomes in budding yeast. undergo nucleotide exchange to provide the active GTP-bound form. Characterisation of the proteins that these genes encode (b) Rab–GTP recruits, either directly or indirectly, an extended coiled- coil protein, which tethers partner membranes and can also interact has thrown up some striking similarities with homotypic with both components of a t-SNARE–n-SEC1 complex, possibly endosome fusion. Thus, as recently reported in Current promoting the availability of the t-SNARE for the next step. Biology [11], vesicle transport in yeast requires PI(3)P (c) Assembly of a trans-SNARE complex, leading to further irreversible generation by the phosphatidylinositol 3-kinase Vps34 and steps that culminate in membrane fusion. The time for cognate-SNARE interaction is limited by the GTP hydrolysis rate of the Rab proteins, as also the FYVE-domain-containing protein Vac1, which hydrolysis before the trans-SNARE complex is properly assembled binds to the yeast Rab5 homologue Vps21. On the basis of leads to breaking of the tether. these findings, Peterson et al. [11] suggest that Vac1 fulfils R260 Current Biology, Vol 9 No 7

the function of EEA1 in this system. They further found 5. Sapperstein SK, Lupashin VV, Schmitt HD, Waters MG: Assembly of the ER to Golgi SNARE complex requires Uso1p. J Cell Biol 1996, that Vac1 also binds to the t-SNARE Pep12 and to Vps45 132:755-767. [11]. Vps45 is related to the n-SEC1 family of proteins, 6. Christoforidis S, McBride HM, Burgoyne RD, Zerial M: The Rab5 which have been proposed to be negative regulators of effector EEA1 is a core component of endosome docking. Nature 1999, 397:621-625. SNARE complex assembly through a t-SNARE interac- 7. Simonsen A, Lipp R, Christoforidis S, Gaullier J-M, Brech A, Callaghan tion [12,13]. This suggests a more complex function for J, Toh B-H, Murphy C, Zerial M, Stenmark H: EEA1 links PI(3)K Vac1, beyond initial attachment of the vesicle, perhaps in function to Rab5 regulation of endosome fusion. Nature 1998, 394:494-498. regulating the availability of Pep12 for participation in 8. Mills IG, Jones AT, Clague MJ: Involvement of the endosomal SNARE complex assembly. autoantigen EEA1 in homotypic fusion of early endosomes. Curr Biol 1998, 8:881-884. 9. Burd CG, Emr SD: Phosphatidylinositol(3)-phosphate signaling In the simplest model of vesicle fusion that is consistent mediated by specific binding to RING FYVE domains. Mol Cell with the currently available data, a Rab family protein in the 1998, 2:157-162. 10. Vitale G, Rybin V, Christoforidis S, Thornqvist P-O, McCaffrey M, GTP-bound form recruits an extended coiled-coil ‘tether- Stenmark H, Zerial M: Distinct Rab-binding domains mediate the ing’ protein to one or both interacting membranes; the interaction of Rabaptin-5 with GTP bound Rab4 and Rab5. EMBO J 1998, 17:1941-1951. coiled-coil protein then promotes membrane attachment for 11. Peterson MR, Burd CG, Emr SD: Vac1p coordinates Vps21 Rab a time limited by the rate of Rab nucleotide hydrolysis GTPase and Vps34 PtdIns 3-kinase signaling, essential for (Figure 1). If a trans-SNARE complex is able to form during Pep12- and Vps45p-dependent vesicle docking/fusion at the endosome. Curr Biol 1999, 9:159-162. this period, then irreversible SNARE-dependent steps on 12. Pevsner J, Hsu SC, Braun JE, Calakos N, Ting AE, Bennett MK, the fusion pathway can proceed. A number of aspects of this Scheller RH: Specificity and regulation of a synaptic vesicle process help to ensure the fidelity of fusion. These are the docking complex. Neuron 1994, 13:353-361. 13. Lupashin VV, Waters GM: t-SNARE activation through transient interaction of a tethering protein with membrane-specific interaction with a Rab-like guanosine triphosphatase. Science factors; the Rab-based timer that limits the time allowed for 1997, 276:1255-1258. trans-SNARE complex formation; cognate SNARE proteins that can complex within the time set by the Rab protein; and perhaps also direct interactions of the tethering protein with t-SNAREs and/or n-SEC1 homologues, which may serve to couple the availability of individual t-SNAREs to membrane attachment by the specific tethering factor.

Uso1 and EEA1 provide the first direct examples of a tethering function for coiled-coil proteins on a vesicle fusion pathway. In both cases, Rab proteins are implicated in recruitment of the proteins to target membranes. Other proteins might have similar functions and form a family of tethering proteins; this is variously suggested by their predicted secondary structure, intracellular localisation and Rab-dependent recruitment, as well as the results of some lower resolution functional studies. These include the yeast protein Imh1, a multi-copy suppressor of ypt6 mutations, Vac1, Rabaptins and p115, a Golgi protein which can tether vesicles to Golgi-cisternae. At this point, the functional assignment for these proteins is only tentative; further high resolution transport assays that are capable of dissecting vesicle attachment from vesicle fusion will be required to directly test whether these proteins have tethering functions.

References 1. Rothman JE: Mechanisms of intracellular protein transport. Nature 1994, 372:55-63. 2. Mayer A, Wickner W, Haas A: Sec18p (NSF)-driven release of sec17p (α-SNAP) can precede docking and fusion of yeast vacuoles. Cell 1996, 85:83-94. 3. Holthuis JCM, Nichols BJ, Dhruvakumar S, Pelham HRB: Two syntaxin homologues in the TGN/endosomal system of yeast. EMBO J 1998, 17:113-126. 4. Cao X, Ballew N, Barlowe C: Initial docking of ER-derived vesicles requires Uso1p and Ypt1p but is independent of SNARE proteins. EMBO J 1998, 17:2156-2165.