BI-DIRECTIONAL PROTEIN TRANSPORT BETWEEN

Annu. Rev. Cell Dev. Biol. 2004. 20:87 123 doi: 10.1146/annurev.cellbio.20.010403.105307 Copyright c 2004 by Annual Reviews. All rights reserved First published online as a Review in Advance on April 21, 2004 BI-DIRECTIONAL PROTEIN TRANSPORT BETWEEN THE ER AND GOLGI Marcus C.S. Lee, 1 Elizabeth A. Miller, 1 Jonathan Goldberg, 2 Lelio Orci, 3 and Randy Schekman 1 1 Howard Hughes Medical Institute and Department of Molecular and Cell Biology, University of California, Berkeley, California; 2 Howard Hughes Medical Institute and Memorial Sloan-Kettering Cancer Center, New York; 3 Department of Morphology, University of Geneva Medical School, Switzerland; email: mcslee@uclink4.berkeley.edu; e miller@uclink4.berkeley.edu; jonathan@ximpact4.ski.mskcc.org; Lelio.Orci@medecine.unige.ch; schekman@uclink4.berkeley.edu These authors contributed equally to this work Key Words vesicle transport, COPI, COPII, coat proteins, cargo selection Abstract The endoplasmic reticulum (ER) and the Golgi comprise the first two steps in protein secretion. Vesicular carriers mediate a continuous flux of proteins and lipids between these compartments, reflecting the transport of newly synthesized proteins out of the ER and the retrieval of escaped ER residents and vesicle machinery. Anterograde and retrograde transport is mediated by distinct sets of cytosolic coat proteins, the COPII and COPI coats, respectively, which act on the membrane to capture cargo proteins into nascent vesicles. We review the mechanisms that govern coat recruitment to the membrane, cargo capture into a transport vesicle, and accurate delivery to the target organelle. CONTENTS INTRODUCTION TO THE SECRETORY PATHWAY... 88 GAINING TRANSPORT COMPETENCE... 90 Folding and Assembly... 90 Accessing an Active Transport Zone... 91 Bulk Flow Exit from the ER... 92 Enrichment of Cargo Proteins... 93 GENERATING A VESICLE... 97 Coat Recruitment and Assembly... 97 Cargo Capture...105 DELIVERY OF VESICLES: MOTORS, TETHERING, AND FUSION...109 Motor Proteins and Vesicle Delivery...109 Vesicle Tethering...110 SNARE Assembly and Fusion...111 NON-CANONICAL TRANSPORT...112 1081-0706/04/1115-0087$14.00 87

88 LEE ET AL. Anterograde Transport of Large Cargo...113 COPI-Independent Retrograde Transport...113 CONCLUSIONS...114 INTRODUCTION TO THE SECRETORY PATHWAY Eukaryotic cells possess an elaborate endomembrane system that makes up the secretory pathway (Figure 1). This network consists of a number of independent organelles that function sequentially to effect protein secretion to the extracellular environment. Each compartment provides a specialized environment that facilitates the various stages in protein biogenesis, modification, sorting, and, ultimately, secretion. The ER is the entry point into the secretory pathway for newly synthesized proteins: Ribosomes dock onto a protein pore in the ER membrane, thereby releasing the nascent polypeptide into the lumen of the ER. The primary role of the ER is to provide a milieu that facilitates protein folding. This is achieved by the presence of chaperones that massage a newly synthesized protein into its correct conformation, sometimes through the catalysis of disulfide bond formation. Post-translational modification of nascent chains first occurs in the ER, including addition of N-linked glycan chains and hydroxylation of proline residues. The next station that a folded secretory protein visits is the Golgi apparatus, a series of cisternae housing enzymes that function in glycan side chain modification and proteolytic cleavage. The modification of carbohydrate moieties can be extensive and proceeds sequentially, with distinct glycosyltranferases specifically located to cis-, medial-, or trans-compartments of the Golgi. The trans-golgi network (TGN) serves as a sorting station that either sends proteins to the cell surface or diverts them to additional compartments of the endomembrane system, Figure 1 Transport in the early secretory pathway. (a) Thin section electron micrograph of an insulin-secreting cell. The ribosome-studded rough endoplasmic reticulum (ER) is visible, giving rise at one site to a budding profile (arrow). Such exit sites, also known as the transitional ER, produce numerous transport vesicles (TV) that will ultimately be delivered to the stacked cisternae of the Golgi apparatus (Orci 1982). (b) Diagrammatic representation of ER-Golgi transport. COPII coat proteins mediate anterograde vesicle formation, selecting anterograde cargo and SNAREs. COPII vesicles in plants and yeast fuse directly with the Golgi, but in mammalian cells they seem to undergo homotypic fusion to generate a pleiomorphic structure known both as the ER-Golgi intermediate compartment (ERGIC) and vesicular tubular clusters (VTC). The ERGIC/VTC is a site for concentrating retrograde cargo into COPI vesicles for delivery back to the ER. The ERGIC/VTC is delivered en bloc to the Golgi in a microtubule-dependent manner. Additional retrograde traffic from the Golgi proper is also mediated by COPI vesicles.

ER-GOLGI TRANSPORT 89 including the vacuole/lysosome, which is responsible for protein degradation, or to a variety of endosomal compartments that broadly function in protein sorting events between the plasma membrane, TGN, and lysosome/vacuole. Transport of proteins between these various compartments of the secretory pathway largely occurs via small vesicles that are generated at a donor compartment

90 LEE ET AL. and fuse with a downstream acceptor compartment. Cytoplasmic coat proteins sculpt vesicles by locally deforming the donor membrane. Distinct sets of coats function at different steps in the secretory pathway: Clathrin and its various adaptors function in a number of transport events between the TGN, endosome, lysosome/vacuole, and plasma membrane; the COPI coatomer complex functions in retrograde transport between the Golgi and ER; and the COPII coat delivers proteins from the ER to the Golgi. In addition to their role in vesicle formation, coat proteins also drive the selective capture of proteins into vesicles by interacting with specific signals present on the cytoplasmic domains of proteins. The combination of coat recruitment to the correct donor membrane and signal-specific interactions between the coat and cargo proteins contributes to the directionality and fidelity of vesicular transport. Further transport specificity is imparted by specific combinations of tethering complexes and fusion assemblies known as SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptors), both of which facilitate fusion of vesicles with acceptor organelles. This review focuses on the mechanisms of vesicular transport between the first two compartments of the secretory pathway, the ER and the Golgi. Transport between these organelles occurs in two directions: anterograde (ER-Golgi) transport delivers newly synthesized cargo proteins to the Golgi, whereas retrograde (Golgi-ER) transport retrieves ER residents and other machinery that constantly cycle between these two compartments (Figure 1). We consider the requirements for access of cargo proteins to nascent vesicles, as well as the mechanisms of vesicle formation, with particular emphasis on the detailed molecular interactions driving this process that have been elucidated through recent structural analysis of the various coat proteins. GAINING TRANSPORT COMPETENCE An important feature of vesicular transport is that, for the most part, only designated cargo proteins are packaged into a nascent vesicle; organelle resident proteins fail to be incorporated or are included at their prevailing concentration (Malkus et al. 2002). Designated cargo proteins need not only refer to newly made biosynthetic cargo; a variety of cellular machinery proteins that facilitate various aspects of vesicle formation and downstream fusion are also specifically recruited. In any case, a number of factors govern whether a bona fide cargo protein will be recognized and subsequently captured into a newly forming vesicle. Some of these factors are common to both anterograde and retrograde pathways, whereas others are more specific. Folding and Assembly Newly synthesized proteins entering the anterograde COPII pathway must escape the clutches of the ER quality control system, which monitors the folding status of

ER-GOLGI TRANSPORT 91 proteins and targets for degradation those that are terminally misfolded. The precise mechanisms that mediate this assembly checkpoint remain poorly understood, but the process is thought to be important in preventing proteotoxicity associated with aberrant secretion of improperly assembled proteins (Kim & Arvan 1998). The observation that misfolded proteins such as a temperature-sensitive mutant of the vesicular stomatitis virus G protein (ts045 VSVG) show prolonged interactions with ER chaperones led to the suggestion that extensive association with an immobile matrix of resident proteins would preclude incorporation of the aberrant cargo protein into a COPII vesicle (Hammond & Helenius 1994). Fluorescence microscopy of a green fluorescent protein (GFP)-tagged version of ts045 VSVG revealed that the mobility of the misfolded protein within the ER was identical to that observed for the correctly folded form, suggesting that the mechanism of retention is not immobilization (Nehls et al. 2000). However, depletion of ATP in these cells caused a decrease in both the mobile fraction of ts045 VSVG and the diffusion rate of a soluble ER marker. Hence, under conditions where chaperonesubstrate dynamics are perturbed, the ER matrix may be more viscous and thereby impede the mobility of a misfolded substrate. How these dynamics might influence the availability of a cargo protein for packaging into a vesicle is unclear; however, the mobility of a substrate protein within the ER may regulate access to a privileged site of COPII budding (see below). It seems likely that the vast majority of cargo proteins that transit the retrograde pathway in COPI vesicles are by nature fully folded; this pathway functions to retrieve escaped ER residents and to recycle the machinery that mediates vesicle formation and fusion. However, recent studies examining the fate of misfolded soluble proteins have indicated that some misfolded proteins travel to the Golgi before being retrieved to undergo ER-associated degradation (ERAD) (Caldwell et al. 2001, Taxis et al. 2002, Vashist et al. 2001). In this respect, the inverse of a folding requirement is likely to act in recruitment of these cargoes into COPI vesicles; only misfolded proteins will gain access to a vesicle and thereby be prevented from further transport through the secretory pathway. Presumably, this recruitment acts via an unknown transmembrane receptor that couples the misfolded substrate to the cytoplasmic coat. Accessing an Active Transport Zone In most eukaryotes, the endoplasmic reticulum is not a homogeneous environment. Fluorescence and immunoelectron microscopy of COPII proteins reveals a restricted localization to domains of the ER that have been variously designated as ER exit sites (ERES) or transitional ER (ter), which represent regions dedicated to generating COPII vesicles (Kuge et al. 1994, Orci et al. 1991, Pagano et al. 1999). Precisely how these distinct zones are maintained is unclear, and the repertoire of proteins that mark these sites is not fully characterized. Furthermore, the functional significance of these privileged budding sites is not known, although the degree of organization of these ER zones may influence the morphology of newly forming

92 LEE ET AL. Golgi elements (Rossanese et al. 1999). An additional role for these sites in regulating the quality control of cargo selection is implicated by immunofluorescence microscopy examination of the colocalization of ts045 VSVG with COPII proteins under permissive and restrictive conditions (Mezzacasa & Helenius 2002). When ts045 VSVG is misfolded, it fails to gain access to the ERES, which are also devoid of ER resident chaperones. Upon shift to the permissive temperature, ts045 VSVG rapidly folds and becomes associated with ERES, suggesting that the folding status of this cargo protein is monitored such that only the fully folded protein gains access to a privileged site of COPII vesicle formation. The localization of COPI proteins on the Golgi apparatus does not appear to show the same degree of organization. If the main function of these specialized zones in the anterograde pathway is to act in quality control, this apparent lack of organization in the retrograde pathway could merely reflect the absence of a specific checkpoint that cargo proteins must pass. However, in mammalian cells the selective and efficient capture of cargo proteins into COPI vesicles may make use of an additional membrane-bound compartment, the ER-Golgi intermediate compartment (ERGIC), also known as vesicular tubular clusters (VTC). This compartment is thought to arise from the homotypic fusion of COPII vesicles but is also marked by COPI proteins. Immunogold electron microscopy of cargo and coat proteins suggests that the ERGIC is a main sorting station for the retrieval of escaped ER resident proteins (Martinez-Menarguez et al. 1999) and therefore may represent a functionally distinct budding zone similar to that of the ter/eres in the anterograde pathway. Bulk Flow Exit from the ER Once proteins destined to leave the ER are correctly folded and have escaped the retention mechanisms described above, they may enter transport vesicles either at their prevailing concentrations (bulk flow) or at concentrations significantly higher than that in the general ER. Early support for rapid bulk flow transport of soluble cargo was obtained by measuring the rate of secretion of a glycosylated acyltripeptide from mammalian cells (Wieland et al. 1987). In a recent study, however, a comparison of three separate bulk flow markers (a glycosylated acyltripeptide, ER-lumenal GFP, and total phospholipid) revealed that only 0.6 2% of each was captured into COPII vesicles generated from a cell-free reaction using yeast microsomes (Malkus et al. 2002). Thus passive sampling of the lumenal contents of the ER appears to be an inefficient mode of transport, but one that is nonetheless used by some secreted proteins. Exocrine pancreatic cells secrete the soluble enzymes amylase and chymotrypsinogen, which are concentrated during passage through the secretory pathway and accumulate in secretory granules prior to regulated exocytosis. Immunoelectron microscopy analysis of amylase and chymotrypsinogen in the early secretory pathway demonstrated that enrichment did not occur in COPII-coated ER buds but in the VTC, where they were concentrated 1.5-fold and 16-fold, respectively (Martinez-Menarguez et al. 1999).

ER-GOLGI TRANSPORT 93 Enrichment of these proteins in the VTC most likely reflects exclusion from retrograde COPI vesicles rather than selective capture into anterograde COPII vesicles (Martinez-Menarguez et al. 1999). It remains to be seen whether this process of concentration by exclusion is used more widely or is limited to especially abundant proteins or proteins that do not require rapid delivery to their final destination. Enrichment of Cargo Proteins Specific enrichment of both membrane and soluble cargo proteins in transport vesicles occurs at concentrations 3- to 50-fold higher than the bulk flow markers described above. This enrichment is achieved by interaction of the cytoplasmic coat with distinct sorting signals on cytoplasmic segments of membrane cargo proteins. To be recognized by the coat machinery, these signals must be accessible and in an appropriate conformation that may be influenced by a number of factors, including the folding status of the protein (discussed above), interactions with accessory proteins or transport receptors, or an oligomerization state that may influence the presentation of positive sorting signals or masking of retention signals. TRANSPORT RECEPTORS Soluble cargo proteins and GPI-anchored membrane proteins that do not present cytoplasmic signals must instead interact with specific transmembrane receptors that serve to link these lumenal substrates to the cytoplasmic coat. The yeast membrane protein Erv29 forms a complex with a precursor of the soluble pheromone, α-factor, and is required for the efficient packaging of pro-α-factor into a COPII vesicle (Belden & Barlowe 2001b). Deletion of Erv29 results in a reduction of the α-factor precursor packaging to levels similar to bulk flow markers (Belden & Barlowe 2001b, Malkus et al. 2002). In addition to α- factor, Erv29 is also likely to package other soluble proteins, including the vacuolar hydrolases, carboxypeptidase Y, and proteinase A; however, other secretory proteins such as invertase are packaged normally in Erv29-depleted cells, suggesting additional transport receptors remain to be identified (Caldwell et al. 2001). In mammalian cells, a lectin, ERGIC-53, is required for the secretion of a number of glycoproteins, including the clotting proteins factor V and factor VIII (Appenzeller et al. 1999). Mutations in either ERGIC-53 (also called LMAN1) or an associated protein, MCFD2, cause an autosomal recessive bleeding disorder as a result of a combined deficiency of the clotting factors (Zhang et al. 2003). Although not well characterized, the interactions between soluble cargo proteins and transmembrane sorting receptors are likely to be influenced by the folding state of the cargo protein, providing an additional level of quality control. In the case of ERGIC-53, this quality control function may be facilitated by sequential interaction of the cargo substrate with the lectin chaperones calnexin/calreticulin and ERGIC-53. Furthermore, the mechanism by which these transport receptors release their cargo upon deposition in the Golgi remains to be determined, although by analogy with other transport steps in the secretory pathway, a change in ph or another

94 LEE ET AL. physiologic condition may cause dissociation of the cargo and its receptor and thereby allow retrieval and reuse of the receptor. In the retrograde pathway, a different transmembrane protein fulfills the function of transport receptor. Soluble ER resident proteins bear a specific retrieval signal (KDEL in mammals, HDEL in yeast) that mediates interaction with the KDEL receptor (Erd2 in yeast). The KDEL receptor itself possesses a non-canonical cytoplasmic di-lysine retrieval motif that contributes to interaction with the COPI coat in conjunction with a phosphoserine residue that is also important for ER- Golgi retrieval (Cabrera et al. 2003). Although the precise mechanisms by which the KDEL receptor binds its ligands are not known, a number of point mutations on the lumenal surface of the receptor abrogate binding to KDEL peptides, implicating these residues as a ligand-binding pocket (Scheel & Pelham 1998). Furthermore, ligand binding to the KDEL receptor is thought to trigger uptake of the assembled complex into COPI vesicles (Lewis & Pelham 1992). This regulated transport may be driven by the ligand-induced oligomerization of the receptor, which in turn may stimulate interaction with the COPI coat (Majoul et al. 2001). In vitro binding assays have demonstrated that the affinity of the KDEL receptor for its ligands shows a striking ph dependency, suggesting that subtle differences in the luminal environment will allow release of KDEL-containing proteins upon fusion with the ER (Wilson et al. 1993). Thus many aspects of the KDEL retrieval pathway mirror similar ligand-induced endocytic events at the plasma membrane: ligandinduced dimerization, specific recruitment to a budding site, and ligand release upon a change in luminal ph. Similar to soluble cargoes, GPI-anchored proteins project no signals into the cytoplasm, suggesting a requirement for a transmembrane receptor. Yeast Emp24 may function as the receptor for the GPI-anchored membrane protein Gas1; Emp24 can be cross-linked to Gas1 in transport vesicles, and packaging of Gas1 into transport vesicles is reduced in Emp24-depleted cells (Muniz et al. 2000). A further specialized role for Emp24 in ER export is suggested by the observation that Gas1 and other GPI-anchored proteins enter transport vesicles that are distinct from those that carry other cargo proteins such as α-factor and amino acid permeases (Muniz et al. 2001). Yet another role for the p24 family of proteins (of which Emp24 is a member) in the fidelity of protein sorting has been implicated by the observation that ER resident proteins are secreted from cells harboring deletions in p24 proteins (Elrod-Erickson & Kaiser 1996). Likewise, mutation of a Caenorhabditis elegans p24 protein, sel-9, causes the secretion of an aberrantly folded plasma membrane receptor that is normally retained in the ER (Wen & Greenwald 1999). Various models explaining the mechanism of p24 function in sorting fidelity remain to be fully tested (Kaiser 2000), and the possibility remains that this family of proteins has multiple functions, perhaps in both anterograde and retrograde traffic. PACKAGING CHAPERONES In some cases, cargo receptors also function as molecular chaperones. The Drosophila protein, NinaA, likely functions as a peptidylprolyl cis-trans isomerase and is required for proper ER export of rhodopsin.

ER-GOLGI TRANSPORT 95 NinaA is also associated with rhodopsin in secretory vesicles and is thought to act both as chaperone and escort, accompanying rhodopsin in its journey through the secretory pathway (Baker et al. 1994, Colley et al. 1991). However, a direct role for NinaA as a cargo receptor for uptake into ER-derived vesicles has not yet been demonstrated. Whereas NinaA seems to accompany its substrate out of the ER, another putative chaperone, yeast Shr3, seems to facilitate packaging of its substrates yet is retained in the ER (Kuehn et al. 1996). Cells lacking Shr3 accumulate amino acid permeases in the ER. Shr3 interacts with the COPII coat and through this interaction facilitates packaging of the permeases without being captured into avesicle itself (Gilstring et al. 1999). Additional examples of membrane proteins that seem to facilitate ER export of transmembrane substrates include yeast Chs7, the mutation of which causes ER accumulation of a plasma membrane chitin synthase (Trilla et al. 1999); yeast Erv14, which accompanies Axl2 out of the ER (Powers & Barlowe 1998); yeast Pho86, deletion of which causes the phosphate transporter Pho84 to accumulate in the ER (Lau et al. 2000); and C. elegans odr-4, which is involved in secretion of odorant receptors (Dwyer et al. 1998). Exactly how these packaging chaperones function in ER export remains to be characterized. The requirement for accessory factors to mediate transport of transmembrane proteins is common to both the anterograde and retrograde pathways. The yeast Golgi membrane protein, Rer1, facilitates retrieval of a number of escaped ER resident membrane proteins, including Sec12 and Sec71 (Sato et al. 1996, 1997). Interactions between Rer1 and its substrates within transmembrane domains govern retrieval, although separate regions in Rer1 seem to interact with Sec12 and Sec71 (Sato et al. 2003). Whether Rer1 itself is taken up into COPI vesicles remains to be determined; however, it interacts directly with subunits of the COPI coat and contains an essential cytoplasmic signal that resembles the canonical COPI di-lysine retrieval motif, as well as a separate tyrosine signal (Sato et al. 2003). The mechanism by which Rer1 releases its ligands upon deposition in the ER remains to be fully characterized. However, the observation that many of the substrates that use Rer1-mediated retrieval are oligomeric proteins suggests that Rer1p functions to retrieve escaped monomers that have a higher affinity for their ER-localized partners. Retrieval to the ER should result in a displacement of Rer1 as the cargo monomer is restored to an oligomeric form. OLIGOMERIC ASSEMBLY Instead of using accessory proteins to facilitate transport, many cargo molecules rely on assembly cues to drive uptake into a transport vesicle. This seems to be true for both biosynthetic cargo and the transport machinery that cycles between the ER and Golgi. The mammalian lectin ERGIC-53 contains a cytoplasmic sorting signal that directs interaction with the COPII coat. However, homo-oligomerization of ERGIC-53 is a prerequisite for efficient ER export: Mutation of cysteine residues that participate in intermolecular disulfide bonds disrupts ER-Golgi transport (Nufer et al. 2003). Similar oligomerizationdependent ER export has been reported for the yeast homologs of ERGIC-53,

96 LEE ET AL. Emp46, and Emp47, where disruption of coiled-coil domains that mediate assembly impairs uptake of these proteins into COPII vesicles but has little effect on cargo-coat interactions (Sato & Nakano 2003). Although the mechanism by which oligomerization might affect ER export remains unclear, one could imagine that presenting an oligomeric signal might dramatically increase the affinity of the coat for a specific cargo protein, thereby enriching the assembled complex in a nascent budding zone. Alternatively, the coat may not recognize a signal unless it is presented in a specific conformation that is achieved only through oligomerization. Such a contextual arrangement of ER export signals has been proposed for the yeast ER/Golgi proteins Erv41/Erv46 (Otte & Barlowe 2002). Yet another requirement for oligomerization may result from the absence of ER export signals on specific partners within a heteromeric complex. For example, a subfamily of mammalian inwardly rectifying potassium channels, Kir3, possesses four members that can combine in different permutations to yield active channels. Homotetramers of one member, Kir3.1, are not functional because these assemblies are retained in the ER. Kir3.1 lacks any ER export signals and therefore relies on signals found on its partners for efficient transport from the ER (Ma et al. 2002). Such reliance on heteromeric assembly may reflect an additional layer of quality control, such that only properly assembled complexes are selected for forward transport, leaving unassembled monomers behind to search for an appropriate partner or to interact further with the ER folding machinery. Similar assembly-dependent transport has been described in retrograde retrieval, although this represents an inverse assembly requirement such that fully assembled complexes are not a substrate for COPI retrieval. Individual subunits of a number of oligomeric cell surface receptors and channels possess ER retrieval motifs that are masked when the holocomplex is correctly assembled (Letourneur et al. 1995, Mallabiabarrena et al. 1992, Margeta-Mitrovic et al. 2000). This masking of retention signals is presumed to be a steric effect of the interaction of adjacent cytoplasmic domains to prevent access to the COPI coat. In contrast to intersubunit masking of retrieval motifs, recent studies on the secretion of oligomeric potassium channels implicate an external component in regulating forward transport: 14-3-3 proteins. These potassium channels possess di-arginine motifs that function largely as retrieval signals, mediating COPI-dependent retrieval to the ER. Correct assembly into an oligomeric structure obstructs this ER retrieval motif by recruiting specific members of the 14-3-3 protein family (O Kelly et al. 2002, Yuan et al. 2003). Binding of 14-3-3 thus competes with the COPI coat for interaction with these substrates, preventing uptake into the retrograde pathway. Again, the function of this assembly-mediated retrieval seems to be in denying access to the later secretory pathway for proteins that are improperly assembled and thereby giving these complexes a second chance for interaction with the folding apparatus of the ER. A common theme in this discussion of the requirements for specific cargo capture into budding vesicles is that the process seems tightly linked to quality control. Whether through the use of specific cargo receptors that might monitor the folding

ER-GOLGI TRANSPORT 97 state of a protein or the correct presentation of oligomeric signals or masking of retrieval motifs that indicates correct assembly, cells have multiple checkpoints that ensure only legitimate cargo substrates are delivered to the appropriate compartment. Furthermore, the retrograde retrieval pathway provides an extra layer of quality security in giving inappropriately assembled substrates a second chance at achieving the correct conformation. Our next challenge lies in understanding how these different checkpoints are mechanistically linked to the process of vesicle biogenesis. GENERATING A VESICLE The process of constructing a vesicle follows a relatively prescribed course of action that is common to all vesicle budding events (Figure 2). The most fundamental of these is the localized curvature of the membrane that sculpts a vesicle out of the donor compartment. Generally, this curvature is driven by the coat subunits that define the vesicle type or by additional effectors that may locally deform a membrane at specific sites. To this end, coat components must be specifically recruited to appropriate sites of vesicle budding and interact there with bona fide cargo proteins to generate a functional vesicle. Generally, small GTP-binding proteins that act as molecular switches to turn on coat assembly drive the initial recruitment of the coat. Additional coat components are subsequently recruited, cargo molecules are recognized, and coat polymerization drives release of the nascent vesicle from the donor membrane. Our understanding of the molecular mechanisms that drive these processes has been greatly enriched by the structural characterization of individual coat components. Although many of the structural details of the anterograde and retrograde coat complexes are remarkably distinct, these two pathways share many broad similarities. Coat Recruitment and Assembly ACTIVATION OF THE SMALL G PROTEINS Coat assembly begins through activation of the small G proteins; Sar1 for COPII and Arf1 for COPI (Figure 3). Both proteins cycle between cytosolic and membrane-associated pools, with GTP binding generating the active, membrane-bound state. As with other small G proteins such as Ras and Rho, exchange of GDP for GTP is catalyzed by guanine nucleotide exchange factors (GEFs), but Sar1 and Arf1 are unusual in that the exchange reaction can proceed only in the presence of a membrane surface (Antonny et al. 2001, Paris et al. 1997). This requirement for membranes upon GTP binding results from the exposure of an N-terminal amphipathic α-helix that is unique to the ARF family. In Arf1 this helical domain extends for 17 residues and is capped at the N terminus by a myristoyl group (Antonny et al. 1997), whereas Sar1 is not lipid modified but has a longer domain of approximately 23 residues (Huang et al. 2001). In the GDP state the helix is retracted into a surface pocket, and truncation

98 LEE ET AL.

ER-GOLGI TRANSPORT 99 of this helical region permits nucleotide exchange to proceed in the absence of membranes (Antonny et al. 1997, Bi et al. 2002). Structural information now available for both the GDP- and GTP-bound states of Sar1 and Arf1 illustrate how GTP binding induces conformational changes that couple membrane binding with coat recruitment (Amor et al. 1994, Bi et al. 2002, Goldberg 1998, Huang et al. 2001). As with other Ras-like G proteins, the classical switch I and switch II regions are rearranged to accommodate the γ phosphate of GTP (Bi et al. 2002, Goldberg 1998). Uniquely to the ARF family, however, this rearrangement is coupled to a 7 Å displacement of a β-hairpin, consisting of the β2 β3 strands, which connect the switch regions, into the pocket occupied by the N-terminal helix (Bi et al. 2002, Goldberg 1998). Thus insertion of the N- terminal helix into the membrane bilayer is a prerequisite for GTP exchange, and retraction of this membrane anchor can occur only upon GTP hydrolysis (Figure 3). This requirement for membranes coupled with localized GEF activity may help to restrict coat recruitment to the appropriate compartment. Despite similarities in the conformational changes adopted by Sar1 and Arf1 upon GTP binding, the exchange factors that act on each protein appear to be unrelated and are likely to catalyze nucleotide exchange by different mechanisms. Sec12, the GEF for Sar1, is a type II membrane protein located in the ER (Barlowe & Schekman 1993). In Saccharomyces cerevisiae, Sec12 appears to be distributed throughout the ER. However, in the related budding yeast Pichia pastoris, Sec12 is found in punctate spots that colocalize with other COPII subunits, suggesting Sec12 restricts COPII recruitment to distinct transitional ER budding zones (Rossanese Figure 2 Coat assembly and vesicle budding. (a) COPII coat assembly is initiated by the ER resident, Sec12, which serves as a guanine nucleotide exchange factor (GEF) for the small GTPase, Sar1 (1). GTP binding by Sar1 exposes an amphipathic α-helix that facilitates association with the ER membrane. Membrane-associated Sar1 recruits the Sec23-24 heterodimer (2), and this complex interacts with cargo proteins via specific sorting signals (3). The Sar1-Sec23-Sec24 complex then recruits the Sec13-31 heterotetramer (4), which is thought to polymerize the coat and drive membrane deformation to yield a COPII vesicle, shown in thin section (Schekman & Orci 1996). (b) COPI coat assembly is similar in that coat recruitment is initiated by GDP-GTP exchange on Arf1, mediated by the ARF GEF, Gea1 (1). Membrane-bound Arf1 then recruits the preassembled coatomer complex, which contains seven subunits: the α/β /ε complex and the β/γ /δ/ζ complex (2). The COPI coatomer complex likely contains multiple cargo recognition sites on separate subunits that mediate recruitment of cargo proteins (3). Ultimately, the coat is polymerized by an unknown mechanism, and membrane curvature may facilitate the recruitment of an Arf GTPase-activating protein (ARF GAP), stimulating GTP hydrolysis by Arf and subsequent dissociation from the membrane (4). A purified COPI vesicle is shown in thin section (Schekman & Orci 1996).

100 LEE ET AL. Figure 3 Nucleotide-dependent conformational changes in Arf1. A comparison of the structures of Arf1 bound to GDP or a nonhydrolyzable analog of GTP shown in identical orientation. The amphipathic N-terminal helix of Arf1 (red) issequestered in ahydrophobic groove in the GDP-bound state. Exchange of GDP for GTP is associated with the insertion of this N-terminal helix into the membrane bilayer and the loss of the hydrophobic binding pocket. Subsequent retraction of the helix is contingent on GTP hydrolysis. et al. 1999). The cytosolic domain of Sec12 is predicted to form WD40 repeats, folding as a seven-bladed β-propeller (Chardin & Callebaut 2002). In this respect it is reminiscent of the GEF for Ran, RCC1, which also forms a seven-bladed β-propeller (Renault et al. 2001). In contrast to Sar1, several different exchange factors exist for Arf1, likely reflecting the role of Arf1 in multiple vesicle budding events. Although widely different in size, all Arf GEFs contain a catalytic domain of 200 amino acids, known as the Sec7 domain, and thus are likely to mediate nucleotide exchange by the same mechanism (Donaldson & Jackson 2000). Structural and biochemical analyses of the Sec7 domain reveal an extensive interaction with the switch I and II regions of Arf1, positioning a catalytic glutamine residue on the Sec7 domain into the GTPase active site. Displacement of the bound nucleotide occurs through disruption of the nucleotide β-phosphate group and the Mg 2+ ion (Goldberg 1998, Mossessova et al. 1998). The complex between the Sec7 domain and Arf1-GDP is the target of the fungal metabolite Brefeldin A, which is commonly used to disrupt Arf1-dependent trafficking (Mossessova et al. 2003b, Robineau et al. 2000). ARF GEFs fall broadly into two classes. The large GEFs such as Gea1, Gea2, and Sec7 in yeast, and GBF1 and BIG1 in humans, are all larger than 100 kd. The

ER-GOLGI TRANSPORT 101 small GEFs (<100 kd), such as ARNO and cytohesin-1, do not have homologs in yeast (Jackson & Casanova 2000). The Gea1 and Gea2 proteins, which have overlapping functions, are largely responsible for stimulating the formation of retrograde COPI vesicles in yeast, whereas Sec7 is located mainly at the TGN (Peyroche et al. 1996, Spang et al. 2001). The mammalian ortholog of Gea1/2, GBF1, is located at the cis-golgi and VTCs and is likely to perform the equivalent function (Zhao et al. 2002). Unlike Sec12, the exchange factor for Sar1, all ARF GEFs are cytosolic proteins. How then is Arf1 activation directed to the appropriate membrane? One likely explanation is that additional lipid- and protein-binding domains could localize different GEFs to their target membranes. Recently Gea1/2 recruitment to the Golgi membrane was found to be stimulated by interaction with a transmembrane protein, Gmh1 (Garcia-Mata et al. 2003). In addition, small GEFs such as ARNO have a pleckstrin homology domain that mediates binding to phosphoinositides, although specific lipid-binding domains on the larger GEFs have yet to be identified (Jackson & Casanova 2000). In addition, recruitment of Arf-GDP to the Golgi by the cytosolic tails of some p24 cargo proteins suggests a second mechanism to restrict Arf1 activation (Gommel et al. 2001). COAT ASSEMBLY ON THE MEMBRANE Membrane-bound Sar1-GTP and Arf1-GTP are now primed to recruit additional coat proteins. The COPI coat consists of seven subunits, a subset of which appears to be structurally related to the adaptor protein (AP) complex found in clathrin-coated vesicles. The COPII coat has four subunits that bear no apparent relation to either the COPI or AP coat complexes. Both the COPI and COPII coat proteins are conserved between yeast, plants, and mammals. Despite the lack of similarity between the COPI and COPII subunits, the common function of both coats is to promote membrane curvature and cargo protein capture. COPII coat recruitment At the ER, Sar1-GTP sequentially recruits two cytosolic complexes, the Sec23-Sec24 heterodimer and the Sec13-Sec31 heterotetramer (Figure 2). On synthetic liposomes, these five components are sufficient to deform the membrane and generate coated vesicles (Matsuoka et al. 1998). Sec23 is the GTPase-activating protein (GAP) for Sar1, and binds to Sar1-GTP via the switch regions, on the side opposite to the membrane anchor (Bi et al. 2002, Yoshihisa et al. 1993). The structure of the Sar1-Sec23 complex indicates Sec23 may stimulate the intrinsically slow GTPase activity of Sar1 by supplying an arginine finger into the catalytic site (Bi et al. 2002). Sec24 is responsible for the majority of the cargo recognition decisions (see below). Despite low sequence identity (<14%), Sec23 and Sec24 are structurally similar and form a bowtie-shaped complex approximately 15 nm in length (Figure 4). A concave inner face enriched in basic residues may act to enhance membrane association and facilitate membrane curvature (Bi et al. 2002). Sec13-31 forms the outer layer of the coat and cannot be cross-linked to the lipid membrane, unlike the other subunits (Matsuoka et al. 2001). Both Sec13

102 LEE ET AL. Figure 4 Structure of the Sar1-Sec23-Sec24 complex. Sar1 is shown in red, Sec23 in yellow and Sec24 in green. The three panels illustrate successive 90 rotations of the complex, with the central panel showing the predicted membrane-proximal surface facing forward. This concave inner face is shown in relation to the curved membrane of a 60-nm vesicle (gray line, left and right panels). On Sec24, peptides encompassing ER export signals are shown bound to the A-site (pink) and B-site (blue), and the critical arginine residue in the C-site is highlighted in red. ( 33 kd) and Sec31 ( 140 kd) are predicted to contain WD40 repeats, and a stable complex of two molecules each of Sec13 and Sec31 can be isolated from yeast cytosol (Lederkremer et al. 2001). Electron microscopy images show that Sec13-Sec31 forms an elongated molecule 30 nm long with multiple globular domains (Lederkremer et al. 2001, Matsuoka et al. 2001). Although the exact nature of the interaction between Sec13-31 and Sec23-24 is not known, Sec13-31 is likely to function as a structural scaffold to cross-link adjacent Sec23-24 complexes, forming a coat lattice that propagates membrane curvature. Prior to Sec13-Sec31 recruitment, the lifetime of the Sar1-Sec23-24 prebudding complex ( 30 s on synthetic liposomes) is likely to be sufficiently long as to allow lateral diffusion on the ER membrane and cargo capture by Sec24. Subsequent binding of Sec13-31 accelerates the Sec23-mediated GAP activity by an order of magnitude (Antonny et al. 2001). Thus the internal timer for Sar1 release by GTP hydrolysis is controlled by the stepwise assembly of the COPII coat proteins. This sequential assembly is in contrast to that of the COPI coat complex, which comes as a preassembled heptamer; the GAP activity associated with simultaneous recruitment of all COPII subunits would preclude stable coat formation. Instead, this two-gear mechanism of GAP stimulation may allow propagation of the coat lattice and membrane curvature to proceed sufficiently to withstand loss of the Sar1

ER-GOLGI TRANSPORT 103 anchor from the epicenter. Additionally, multivalent interactions of the Sec23-24 complex with cargo molecules could further stabilize the outer coat. In support of this, gently isolated vesicles generated in the presence of GTP retained substantially more Sec23-24 and Sec13-31 than Sar1 (Barlowe et al. 1994). An additional factor that may promote coat stability is the large peripheral ER protein, Sec16 (Espenshade et al. 1995). Temperature-sensitive alleles of SEC16 block ER to Golgi transport, and are lethal in combination with mutations in SEC12, SEC13, and SEC23 (Kaiser & Schekman 1990). In addition, physical interactions have been observed between Sec16 and all of the COPII components except Sec13 (Espenshade et al. 1995, Gimeno et al. 1996, Supek et al. 2002). In vitro budding reactions using ER-enriched membranes stripped of Sec16 revealed that vesicle formation was significantly reduced in reactions performed with GTP but not with a nonhydrolyzable GTP analog. Efficient vesicle formation in the GTP reactions was restored by addition of purified Sec16. The effect of Sec16 on coat dynamics is not mediated by a suppresion of GTP hydrolysis on Sar1, but may instead reflect a role as a platform to stabilize coat proteins on the membrane despite Sar1-GTP hydrolysis (Supek et al. 2002). COPI coat recruitment Activated Arf1 is essential to the recruitment of the 700 kd-heptameric COPI complex from the cytosol (Figure 2; Orci et al. 1993), although additional interactions between the COPI subunits and cargo proteins such as the p24 family may facilitate membrane association. Under high salt conditions the heptamer dissociates into two functionally distinct subcomplexes, the F-COPI subcomplex (β, γ,, ζ ) and the B-COPI subcomplex (α, β, ε) (Fiedler et al. 1996). The four components of the F-COPI subcomplex are related in sequence and structure to the clathrin-binding AP complexes (Boehm & Bonifacino 2001, Eugster et al. 2000, Hoffman et al. 2003, Watson et al. 2004), and by analogy, the B-COPI subcomplex may function as a structural scaffold in a manner similar to clathrin. In the F-COPI subcomplex, the N-terminal regions of β-cop and γ -COP share sequence homology with the trunk domain of the large AP subunits (e.g., α and β in AP2), and recent structural analysis revealed similarities between the γ -COP C-terminal domain and the α- and β-ap2 appendage domains (Eugster et al. 2000, Hoffman et al. 2003, Watson et al. 2004). The two other subunits of the F-COPI subcomplex, -COP and ζ -COP, share homology with the µ- and σ -AP2 subunits, respectively (Cosson et al. 1996). Although no structural information is available for the B-COPI subcomplex, both the α-cop and β -COP subunits are predicted to possess β-propeller WD40 repeats at the N-terminus, which function in cargo recognition (see below). In yeast, loss of the WD40 domain from either subunit is tolerated; however, WD40 deletion from both α-cop and β -COP is lethal (Eugster et al. 2004). Interactions between Arf1 and several of the COPI subunits have been documented by photolabile cross-linking (β- and γ -COP) (Zhao et al. 1997, Zhao et al. 1999) and yeast two-hybrid analysis (β- and ε-cop) (Eugster et al. 2004). Site-specific cross-links between the switch regions of Arf1 and β- and γ -COP are consistent with a requirement for GTP-dependent membrane binding of COPI

104 LEE ET AL. (Zhao et al. 1999) and may be structurally analogous to Arf1 binding to the β and γ subunits of AP1 (Austin et al. 2000). Unlike the COPII coat, where the Sar1 GAP is an integral part of the coat, stimulation of GTP hydrolysis on Arf1 to promote coat dissassembly is not mediated by a subunit of the coat per se but by a separate ARF GAP. Several different ARF GAPs have been identified and all share an essential 70-residue zinc finger domain and a conserved arginine residue, although many participate in non- COPI processes (Donaldson & Jackson 2000). In yeast, the Golgi-localized GAPs Glo3 and Gcs1 appear to have overlapping functions in COPI coat dissassembly (Dogic et al. 1999, Poon et al. 1999). The equivalent function in mammalian cells is performed by ARF GAP1, the first ARF GAP to be identified, and possibly other members of the ARF GAP1 family (Nie et al. 2003, Watson et al. 2004). The mechanism by which ARF GAP proteins catalyze GTP hydrolysis on COPI vesicles remains controversial. Structural analysis of a complex between the catalytic domain of ARF GAP1 and Arf1-GDP lacking the N-terminal membrane anchor unexpectedly revealed that ARF GAP1 did not supply a catalytic arginine finger at the GTPase active site. Rather, ARF GAP1 appears to stabilize the switch II region in a more catalytically favorable configuration (Goldberg 1999). Addition of stoichiometric amounts of COPI further stimulated ARF GAP1 activity by three orders of magnitude, although it is not known if this is mediated by an arginine finger on COPI (Goldberg 1999). However, under different experimental conditions, using full-length myristoylated Arf1 in the presence of ARF GAP1 and synthetic membranes, addition of COPI did not stimulate GTP hydrolysis (Szafer et al. 2000), although similar experiments with the yeast GAP Glo3p and Golgi membranes yielded a 50-fold stimulation (Szafer et al. 2001). In all cases, the presence of ARF GAP is an essential component to the reaction, as COPI alone has no effect on GTP hydrolysis by Arf1. Furthermore, Arf1, ARF GAP, and COPI are all required for the formation of an aluminum fluoride complex (B. Antonny, personal communication). Aluminum fluoride conventionally acts in a G protein GAP complex to coordinate the arginine finger of the GAP by mimicking the γ -phosphate of GTP. Thus additional studies that reveal how ARF GAP and the COPI coat may simultaneously interact with Arf1 should be informative. In essence, however, the potential effect of COPI on the rate of GTP hydrolysis is broadly similar to the two-gear mechanism described for the COPII system in that the greatest GTPase activity is observed only upon assembly of all the coat components. Given that stable assembly of the COPI coat in the presence of ARF GAP is challenged by the high rate of GTP hydrolysis on Arf1, how is the coat maintained to yield productive vesicles? Two potential mechanisms have been discovered for COPI. Recently, the activity of ARF GAP1 was shown to be sensitive to the degree of membrane curvature (Bigay et al. 2003). When myrisotylated Arf1 and COPI were recruited to synthetic liposomes of defined size, Bigay et al. (2003) observed that the effect of ARF GAP1 on the rate of GTP hydrolysis, and subsequently coat

ER-GOLGI TRANSPORT 105 dissassembly, increased by two orders of magnitude as the size of the liposomes was reduced to that of a COPI vesicle. This suggests an elegant mechanism to restrict ARF GAP activity to late in the budding process, preventing premature uncoating. In particular, ARF GAP would be excluded from regions of negative curvature such as the vesicle neck. This could potentially result in a vesicle bud with a stable ring of Arf1 and COPI coat at the neck and repeated cycles of Arf1 GTP hydrolysis at the bud tip (Lippincott-Schwartz & Liu 2003). Localized cycles of binding and release by Arf1 and coat proteins may play a role in cargo concentration at the bud site (Lanoix et al. 1999). An additional mechanism to regulate coat stability may be mediated by the cargo proteins themselves. The cytosolic tail of the human hp24a (p24β 1 ) cargo protein significantly inhibited GTP hydrolysis by Arf1, although other p24 members or di-lysine-retrieval signals had no effect. Slowed GTP hydrolysis may be induced by hp24a binding to COPI (Goldberg 2000) or ARF GAP1 (Lanoix et al. 2001). The net effect would be to increase uptake of the inhibitory cargo protein and potentially provide a time window for additional coat-coat interactions and cargo capture to occur. Cargo Capture The sorting signals that determine the cellular location of a given protein work by interacting directly with specific subunits of vesicle coats. By nature, the interaction between cargo and coat must be relatively weak because the coat must subsequently be released from the vesicle to allow membrane fusion with downstream compartments. Direct interaction between sorting signals and coat components has been demonstrated for many different transport steps, including those between the ER and Golgi. The canonical ER retrieval motif, KKXX, has long been known as the sorting signal that interacts with the COPI coat to cause ER deposition (Cosson & Letourneur 1994). However, a universal ER export signal has not been described. Instead, a number of different signals seem to govern interaction with the COPII coat, including di-acidic motifs and hydrophobic/aromatic motifs (Barlowe 2003). Our understanding of how these cytoplasmic coats mediate selective cargo capture has been greatly facilitated by recent structural, biochemical, and genetic characterization of the interactions between coat and cargo. CARGO RECOGNITION BY THE COPII SUBUNIT, Sec24 A significant body of evidence has implicated the Sec24 subunit of the COPII coat in cargo selection. The Sar1-Sec23-Sec24 complex is the minimal machinery required to interact with cargo proteins contained in ER membranes (Aridor et al. 1998, Kuehn et al. 1998), and inclusion of the Sec24 subunit is required for incorporation of cargo (Miller et al. 2002). The yeast Sec24 homolog, Lst1, is required for ER export of an abundant plasma membrane protein (Roberg et al. 1999, Shimoni et al. 2000), and vesicles generated with Lst1 in place of Sec24 package a distinct subset of cargo (Miller et al. 2002). Furthermore, Sec23-24, and in some cases Sec24 alone,

106 LEE ET AL. binds in vitro to the cytoplasmic sorting signals of a number of cargo proteins (Mossessova et al. 2003a, Peng et al. 1999, Springer & Schekman 1998). Molecular details of the interaction between Sec24 and its cargo substrates were revealed by structural analysis of yeast Sec24 bound to peptides that correspond to specific sorting signals (Mossessova et al. 2003a). Two sites on Sec24 form independent cargo-binding pockets that recognize distinct sorting signals (Figure 4). The A-site interacts with a YNNSNPF-containing peptide, one of two binding motifs on the SNARE, Sed5. The B-site recognizes signals found on two different peptides: the di-acidic sorting signal of the Golgi protein Sys1 (Votsmeier & Gallwitz 2001) and the LXXLE motif of the SNARE, Bet1. Furthermore, the second motif of Sed5, LXXME, competes for interaction with the Bet1 peptide, suggesting it also employs the same site. The YNNSNPF motif of Sed5 represents a novel class of ER export signal, although its function in Sed5 transport remains to be tested. The peptide binds in a pocket of Sec24 that is membrane-proximal and corresponds spatially to the site occupied on Sec23 by Sar1 (Bi et al. 2002, Mossessova et al. 2003a). The capacity for additional proteins to bind in this pocket remains to be tested. In contrast, the B-site, located on the opposite side of Sec24, clearly interacts with multiple sequences. Although the same electrostatic interactions are involved in binding of both the Sys1 and Bet1 motifs, the two peptides show subtly different conformations (Mossessova et al. 2003a), suggesting that this pocket accommodates substrates in multiple binding modes. The multiplicity of cargo interactions with the B-site was also demonstrated by mutagenesis of this site (Miller et al. 2003). Point mutations in residues that line this pocket or make electrostatic contact with the Bet1 and Sys1 sorting signals not only disrupt packaging of Bet1 and Sys1, but also affect additional cargo proteins to varying degrees, including the p24 family of proteins and the SNAREs, Bos1 and Sec22. Of particular interest is the observation that packaging of a second cargo protein Gap1, which employs a di-acidic sorting signal, is unaffected by these mutations, suggesting that similar classes of sorting signals may in fact be recognized by distinct sites on the COPII coat. Other experiments to disrupt the B- site on the Sec24 homolog (Lst1) demonstrated that this site is conserved as a cargo interaction domain even though the substrate specificities of the two proteins differ (Miller et al. 2002, 2003). Additional mutagenesis and genetic screening identified a third cargo-binding site on Sec24 (Miller et al. 2003). Mutation in this C-site specifically impairs packaging of the SNARE, Sec22, which is also affected by mutations in the B-site, suggesting a bipartite mode of interaction similar to that of Sed5 (Mossessova et al. 2003a). The molecular details of the interactions between Sec22 and Sec24 remain to be determined. The observation that a number of cargo proteins are packaged into vesicles normally in the presence of the mutant forms of Sec24 suggests that additional sites on the COPII coat remain to be identified (Miller et al. 2003). Although it is possible that additional cargo interaction sites will also map to Sec24, a role for either Sec23 or Sar1 cannot be ruled out in some instances. Several cargo proteins that

ER-GOLGI TRANSPORT 107 are packaged by both Sec24 and its homolog, Lst1, are unaffected by mutation of the B-site in either protein (Miller et al. 2003). These cargoes are good candidates for substrates that may interact with Sar1 or Sec23. Indeed, a direct role for Sar1 in cargo selection has been implicated by its recruitment to cytoplasmic domains of cargo proteins in vitro (Belden & Barlowe 2001a, Springer & Schekman 1998) and by its interaction with some cargo proteins within ER membranes (Aridor et al. 2001, Otte & Barlowe 2002). Mammalian glycosyltransferases contain a C-terminal di-arginine motif that is required for ER export; mutation of this motif disrupts interaction with Sar1, suggesting a major role for Sar1 in selective capture of these proteins (Giraudo & Maccioni 2003). There is less evidence for a role for Sec23 in cargo selection, although its structural similarity to Sec24 and moderate conservation of residues important for cargo binding in the Sec24 B-site make it a prime candidate for further investigation. Together, the multiplicity of cargobinding sites on Sec24, the presence of Sec24 homologs that expand the cargo repertoire recognized by the COPII coat, and the potential for additional cargo recognition domains on COPII subunits give an indication of the diversity of cargo substrates that may be accommodated by this coat. KKXX MOTIFS INTERACT WITH THE COPI COAT Superficially, interaction of the COPI coat with its cargo substrates seems more straightforward because of the apparent simplicity of the signals involved. However, despite the recognition many years ago that KKXX motifs interact directly with the COPI coat, the molecular details of this interaction remain obscure (Cosson & Letourneur 1994). Indeed, there is still debate over exactly which subunit of the COPI coat interacts with KKXX motifs. Initial experiments demonstrated that the α/β /ε subcomplex binds to the cytoplasmic tail of a KKXX-containing protein (Cosson & Letourneur 1994). Futhermore, coatomer from sec27-1 (β -COP) or ret1-1 (α-cop) mutant yeast cells fails to bind KKXX signals in vitro (Letourneur et al. 1994). These results are consistent with further in vitro binding studies that examined coatomer binding to the cytoplasmic domains of the p24 family of proteins (Fiedler et al. 1996). Binding of the α-, β -, and ε-cop subunits is dependent on an intact KKXX motif, whereas the β-, γ -, and ζ -COP subunits bind to a different subset of p24 proteins and require an aromatic residue. Thus the KKXX-mediated retrieval function was thought to involve the α/β /ε subcomplex while the β/δ/γ /ζ subcomplex would direct a separate transport event. Contradictory evidence came from cross-linking experiments that used a radiolabeled photoreactive KKXX-containing peptide to identify γ -COP as the sole binding partner (Harter et al. 1996, Harter & Wieland 1998). Furthermore, γ - COP also binds to the Rho-related G protein, Cdc42, via a di-lysine motif, an interaction that may link vesicles to the actin cytoskeleton (Wu et al. 2000). One explanation for these disparate views of cargo binding by the COPI coat is that two independent sites recognize the same motif. In support of this, the COPI coat can be precipitated by neomycin, an antibiotic that contains multivalent pairs of amino groups (Hudson & Draper 1997). The high degree of cross-linking suggests that

108 LEE ET AL. each coatomer complex contains two binding sites for the drug. Drug-dependent precipitation was inhibited by di-lysine, indicating two di-lysine recognition sites on the COPI coat. Indeed, different di-lysine motifs may interact distinctly with the COPI coat (Schroder-Kohne et al. 1998), reminiscent of the multiple interactions of di-acidic motifs with Sec24. In addition to the two putative KKXX-binding sites, further cargo recognition domains remain to be identified on the COPI coat. A yeast two-hybrid screen for interacting partners of the δ-cop subunit identified a peptide sequence that can serve as an ER retrieval motif (Cosson et al. 1998). A similar aromatic motif is found on the ER protein Sec71 and may also correspond to the signal on the p24 proteins recognized by the β/δ/γ /ζ -COP subcomplex. Furthermore, the COPI-binding signals on the ER/Golgi SNAREs that cycle between these two compartments also remain to be identified. Some yeast COPI mutants show defects in anterograde transport, thought to be an indirect effect of failure to retrieve anterograde machinery (Gaynor & Emr 1997). These subunits are prime candidates for interacting with the ER-Golgi SNAREs such as Bet1, Bos1, and Sec22. PRIMING VESICLES FOR FUSION The fundamental importance of the inclusion of SNAREs in a vesicle to mediate downstream fusion has led to the hypothesis that these crucial cargo proteins influence their own uptake by initiating coat polymerization (Springer et al. 1999). The observation that the vesicle-associated SNARE, Bet1p, binds in vitro to Sar1p and that this complex subsequently recruits Sec23-24 suggests that this essential cargo protein could nucleate coat assembly and thereby ensure its incorporation into a nascent vesicle, priming that vesicle for fusion with the Golgi (Springer & Schekman 1998). However, reconstitution of COPII budding from synthetic liposomes demonstrated that this priming event is not absolutely required to initiate coat assembly (Matsuoka et al. 1998). Furthermore, in vitro ER budding reactions with the Sec24 homolog, Lst1, generate COPII vesicles that lack Bet1, confirming that association with Bet1 is not required to generate a vesicle (Miller et al. 2002). However, the process of cargo selection may still be mechanistically coupled to budding such that any abundant cargo protein may nucleate coat assembly. Such priming of coat recruitment would explain the relatively high levels of COPII proteins required to generate vesicles from cargo-free synthetic liposomes (Matsuoka et al. 1998). Such a model is supported by the observation that overexpression of a specific cargo protein rescues defects associated with the COPI coat and increases the efficiency of vesicle formation (Sandmann et al. 2003). Although SNAREs are not absolutely required to generate vesicles in vitro, the process of vesicle budding may still be mechanistically coupled to vesicle fusion through incorporation of specific fusion factors in vivo. Further evidence in favor of the vesicle priming hypothesis has come from detailed biochemical analysis of the accessibility of the sorting signals of the ER-Golgi SNAREs Bet1 and Sed5 (Mossessova et al. 2003a). When Bet1 is assembled in a trans-snare bundle, its LXXLE-binding motif is buried, suggesting that monomeric Bet1 is

ER-GOLGI TRANSPORT 109 preferentially recruited into a COPII vesicle. Conversely, monomeric Sed5 contains an N-terminal regulatory domain that masks the YNNSNPF motif, although the LXXME motif is still available. When Sed5 is assembled into a t-snare complex with Bos1 and Sec22, the N-terminal domain is rearranged such that the YNNSNPF motif is now accessible, suggesting that this assembly is the preferred COPII substrate. These data suggest a variation of the vesicle priming hypothesis whereby the COPII coat preferentially selects the fusogenic forms of the SNAREs, thereby programming vesicles for fusion either with each other or the Golgi (Mossessova et al. 2003a). Priming of vesicles for fusion may not be restricted to SNARE incorporation. Mammalian p115 is one of the tethering factors that helps facilitate vesicle fusion with the Golgi (see below) and is recruited to COPII vesicles in a manner dependent on the G protein, Rab1. Failure to incorporate p115 results in vesicles that are reduced in their ability to fuse with the Golgi, suggesting that Rab1 primes vesicles for fusion by preassembling a tethering factor with its interacting SNARE partners (Allan et al. 2000). Similar priming of COPI vesicles with either SNAREs or tethers to facilitate retrograde transport has not been described, although the observation that in vitro binding of yeast COPI to Bet1 requires a catalytic activity supplied by the ARF GAP, Glo3, suggests that a version of priming may also occur in this pathway (Rein et al. 2002). DELIVERY OF VESICLES: MOTORS, TETHERING, AND FUSION To complete their journey and fulfill their role as intracellular shuttles, cargo-laden vesicles must be correctly delivered to and fuse with the appropriate acceptor compartment. Two separate processes ensure the specificity of this delivery process: vesicle tethering and SNARE assembly. Vesicle tethering anchors the vesicle to the donor compartment and may prime the SNAREs for subsequent action. Specific combinations of SNAREs then form a four-helix bundle that drives membrane fusion. As described above, the coat proteins may program vesicles for fusion by selecting the appropriate tethers and SNAREs during vesicle budding. Prior to the action of tethers and SNAREs, vesicles may undergo directed cytoskeleton-driven movement to deliver them to the site of fusion. Motor Proteins and Vesicle Delivery Although there is little evidence for a role for the cytoskeleton in yeast ER-Golgi transport, this step in mammalian cells is sensitive to microtubule disruption (Malaisse & Orci 1979, Presley et al. 1997). In mammalian cells, COPII vesicles that bud from the ER may be first delivered to the ERGIC/VTC, most likely through a combination of homotypic and heterotypic fusion. This intermediate compartment then relocates to the perinuclear Golgi ribbon in a manner that is

110 LEE ET AL. dependent on the microtubule cytoskeleton and the motor protein dynein/dynactin (Presley et al. 1997, Scales et al. 1997). Thus the microtubule dependence of ER-Golgi transport may lie both in the movement of COPII vesicles away from the ER, and in the delivery of the ERGIC to the Golgi proper. Perhaps the presence of an intermediate compartment between the ER and Golgi in mammalian cells is the result of the long-range transport that must occur between dispersed ER exit sites and the central peri-nuclear Golgi. Yeast cells, being much smaller, may rely instead on simple diffusion to generate encounters between ER-derived vesicles and single dispersed Golgi elements (Preuss et al. 1992). By contrast, plant cells, which contain a mobile pool of Golgi stacks, may employ the coordinated movement of both organelles such that ER exit sites are juxtaposed with an individual Golgi stack, precluding the requirement for microtubules in ER-Golgi transport (Brandizzi et al. 2002). The actin cytoskeleton seems likely to play a role in retrograde traffic of COPI vesicles to the ER. Numerous actin-binding proteins and actin itself are located on the Golgi and are further recruited in response to activation of ARF (Fucini et al. 2000, Godi et al. 1998). Additional interactions between the COPI coat and the actin cytoskeleton are likely to involve the Rho-related G protein, Cdc42, which binds to γ -COPI (Wu et al. 2000) and is thought to modulate actin dynamics through the Arp2/3 complex to effect Golgi-ER transport (Luna et al. 2002). However, whether this coat-dependent actin assembly is involved in transport of vesicles to the ER, or instead serves to regulate vesicle fission, remains to be determined (Fucini et al. 2002, Stamnes 2002). Vesicle Tethering The first interaction that a vesicle has with its acceptor compartment is likely to be mediated by a large protein assembly that tethers the vesicle to the membrane. The mechanistic details of how tethers contribute to vesicle fusion remain obscure; they may merely increase the probability of SNARE encounter or may function catalytically to rearrange SNARE complexes to facilitate fusion. Indeed, different sets of tethering complexes may perform each of these functions (Whyte & Munro 2002). In the anterograde pathway, diffusible vesicles become anchored to Golgi membranes by the addition of Uso1, the yeast homolog of p115 (Barlowe 1997, Cao et al. 1998). Uso1/p115 is a large coiled-coil protein that forms an extended oligomeric structure (Sapperstein et al. 1995), possibly facilitating SNARE complex assembly (Sapperstein et al. 1996, Shorter et al. 2002). In addition to Uso1, a second large oligomeric complex is also likely to function in vesicle tethering. The TRAPP I (transport protein particle I) complex contains seven subunits and is thought to be the landmark on the Golgi that attracts COPII vesicles (Sacher et al. 1998, 2001). TRAPP also functions in vitro as a GEF for the yeast Rab1 homolog, Ypt1 (Jones et al. 2000, Wang et al. 2000), which is responsible for the recruitment of p115/uso1 to membranes (Allan et al. 2000, Cao et al. 1998). Thus the tethering

ER-GOLGI TRANSPORT 111 of COPII vesicles to Golgi membranes is likely to involve initial binding of COPII vesicles via TRAPP, which would activate Ypt1 to recruit Uso1. Binding of Uso1 would stabilize the association of the vesicle with the membrane and also facilitate SNARE assembly, thereby leading to membrane fusion and delivery of the vesicle contents. An additional putative tethering assembly located on the Golgi, the conserved oligomeric Golgi (COG) complex, has been implicated in early transport steps in the secretory pathway. However, the vesicles that dock to this complex are unknown, and the ER-Golgi transport defects associated with mutations in this complex are likely to be indirect effects of disrupting Golgi function (Whyte & Munro 2002). The tethering factors that link COPI vesicles to the ER membrane remain poorly characterized. The best candidate is the yeast Dsl1, a peripheral membrane protein that localizes to the ER. Dsl1 exists in an oligomeric complex that also contains Tip20, which in turn interacts with Sec20 and Ufe1, two of the retrograde SNAREs, (Lewis et al. 1997, Reilly et al. 2001). Dsl1 is an essential protein and certain dsl1 mutants show specific retrograde transport defects (Reilly et al. 2001). Further characterization of the Dsl1-containing complex and its interaction with the Golgi- ER SNAREs should shed further light on how this putative tether may facilitate membrane fusion. SNARE Assembly and Fusion The final step in vesicle-mediated transport between organelles is fusion of the vesicle with the acceptor compartment. This process is mediated by transmembrane SNARE proteins that interact in specific combinations to bring the vesicle and acceptor membranes into close proximity and drive fusion. The cytoplasmic domains of SNAREs assemble into a four-helix bundle, generally with each helix contributed by a single transmembrane protein. Thus a combination of four different SNAREs is required to drive membrane fusion. In vitro fusion experiments with SNAREs reconstituted into synthetic liposomes have been instrumental in defining the specific combinations of SNAREs that mediate membrane fusion (Fukuda et al. 2000, McNew et al. 2000, Parlati et al. 2000). This reconstitution has yielded a model for SNARE assembly whereby the target membrane SNARE (t-snare) has three subunits: a syntaxin-like heavy chain and two light chains composed of either one or two additional SNAREs, depending on the number of coiled-coil domains. The vesicle-snare (v-snare) is a monomeric protein that must be on the membrane opposite from the t-snare assembly (Fukuda et al. 2000). Testing all possible combinations of the yeast SNAREs, only the specific combination of the t-snare Sed5/Bos1/Sec22 and the v-snare Bet1 resulted in liposome fusion (Parlati et al. 2000). This is somewhat contradictory to in vitro budding and fusion experiments that reconstitute transport between the ER and Golgi. In these experiments, Sed5 was required on the Golgi acceptor membranes, whereas both Bet1 and Bos1 were required on the ER-derived vesicle (Cao &

112 LEE ET AL. Barlowe 2000, Spang & Schekman 1998). In the retrograde Golgi-ER pathway, both Sec22 and Bet1 were implicated as the v-snares (Spang & Schekman 1998). It is possible that the mutant membranes used in these in vitro budding and fusion assays are subject to pleiotropic effects, making the precise role of each SNARE difficult to parse out. The SNAREs of the retrograde pathway have been characterized by genetic and biochemical analysis. Ufe1 is the syntaxin-like SNARE of the ER membrane, required for both homotypic ER fusion and retrograde transport (Lewis et al. 1997, Patel et al. 1998). The other SNARE components that interact with Ufe1 include Sec22, Sec20, and Use1/Slt1 (Burri et al. 2003, Dilcher et al. 2003, Lewis et al. 1997), although the topology of these SNAREs remains to be tested using the liposome fusion assay. In addition to the SNAREs themselves, a class of regulatory proteins plays an important but ill-defined role in membrane fusion. Proteins of the Sec1/Munc18 family (SM) interact with syntaxin-like SNARE proteins to modulate SNARE activity. It is still controversial whether SM proteins act as positive or negative regulators of membrane fusion, and it is possible that different mechanisms function at various steps of membrane transport in different systems (Toonen & Verhage 2003). The yeast protein, Sly1, is an SM protein that binds to both Ufe1 and Sed5 via a short motif upstream of the N-terminal regulatory domain of these syntaxin-like SNAREs (Yamaguchi et al. 2002). Sly1 seems to promote the assembly of Sed5 into the correct v-t-snare complex (Peng & Gallwitz 2002). Structural analysis of the binding of Sly1 to the Sed5 motif suggests that the interaction between these two proteins occupies a very small binding surface on Sed5, leaving substantial room for additional protein-protein interactions that could contribute to SNARE assembly, membrane fusion, and disassembly of SNAREs after fusion (Bracher & Weissenhorn 2002). Thus membrane fusion is mediated by numerous proteins including the Rabs, which act with tethering complexes to dock vesicles to the target membrane, SNAREs, which interact in specific combinations to drive membrane fusion, and the SM proteins, which also contribute to specificity. NON-CANONICAL TRANSPORT Given the diversity of traffic through the secretory pathway, it is perhaps not surprising that additional mechanisms exist to deal with particular cargo molecules. Although the majority of secreted proteins are small enough to be comfortably enclosed in a 60 80-nm COPII vesicle, several cell types secrete especially large protein and lipid macromolecular complexes. In particular, the transport of lipoparticles and procollagen fibers would challenge the dimensions of a typical vesicle. In addition, some lipid species such as ceramide can reach the Golgi through a nonvesicular route (Funato & Riezman 2001, Hanada et al. 2003). Finally, a subset of Golgi resident proteins, as well as some bacterial toxins, may reach the ER via a COPI-independent mechanism.

ER-GOLGI TRANSPORT 113 Anterograde Transport of Large Cargo Secretion of both apob-containing very low density lipoproteins (VLDL) from hepatic cells and chylomicrons from intestinal cells is of considerable interest because of their relevance to cardiovascular disease. Hereditary mutations in an isoform of Sar1 (SARA2) are associated with intracellular accumulation of chylomicron particles (Jones et al. 2003), suggesting a role for COPII despite the large size of chylomicrons (150 500 nm; Siddiqi et al. 2003) and VLDL particles (80 200 nm; Gusarova et al. 2003). In addition, expression of a dominant-negative form of Sar1 (Sar1T39A) in VLDL-secreting rat hepatoma cells resulted in apob retention in the ER. In vitro reconstitution of apob100-vldl export from the ER of hepatic cells suggests that full lipidation to the mature VLDL particle may occur in post-er compartments, at least partially relieving the size burden on ER export (Gusarova et al. 2003). However, a separate study that used a cell-free assay to examine the formation of prechylomicron transport vesicles from the ER of intestinal epithelial cells observed the release of large (350 500 nm) membraneenclosed particles (Siddiqi et al. 2003). Although COPII proteins were present on the purified prechylomicron vesicles, depletion of Sar1 or Sec31 unexpectedly stimulated prechylomicron release but prevented fusion with the Golgi (Siddiqi et al. 2003). Thus the precise role of COPII in the generation of large lipoparticle carriers remains unclear. Procollagen assembles in the ER into fibers >300 nm in length, and procollagen secretion has been used as a tool to examine the role of small vesicles in anterograde intra-golgi and ER-Golgi transport (Bonfanti et al. 1998, Mironov et al. 2003, Stephens & Pepperkok 2002). As with lipoprotein export, microinjection of a dominant-negative form of Sar1 blocked exit of procollagen from the ER (Mironov et al. 2003, Stephens & Pepperkok 2002). However, it is not known whether a potentially flexible COPII coat directly contributes to the formation of specialized larger transport carriers (Stephens & Pepperkok 2002) or acts only indirectly at sites adjacent to procollagen protrusion (Mironov et al. 2003). Development of a cell-free export assay for procollagen export may delineate the exact contribution of cytosolic coat proteins to procollagen carrier formation. COPI-Independent Retrograde Transport Retrieval of proteins with di-lysine and KDEL-sorting signals is clearly dependent on COPI function. However, the existence of a COPI-independent retrieval pathway is suggested by the fact that certain protein toxins, which lack any obvious retrieval motifs, can be delivered to the ER (Storrie et al. 2000). Inhibition of COPI function by microinjection of anti-copi antibodies or expression of a dominantnegative form of Arf1 fails to block the delivery of Shiga toxin and resident Golgi glycosyltransferases to the ER. Under the same conditions, the retrieval of the di-lysine-containing cargo protein, ERGIC53, and the KDEL receptor is inhibited as expected (Girod et al. 1999). Furthermore, the converse effect is observed upon

114 LEE ET AL. expression of a GDP-restricted Rab6, which inhibits retrieval of Shiga toxin and glycosyltransferases but not ERGIC53 and KDEL receptor (Girod et al. 1999). Rab6 may act to recruit the dynein-dynactin motor proteins through interaction with BICD1 and 2, the mammalian orthologs of Drosophila Bicaudal-D (Matanis et al. 2002). In comparison with the kinetics of COPI transport, COPI-independent recycling occurs slowly (Girod et al. 1999), and the nature of the transport vehicle is not known. This alternate pathway may function to return long-lived Golgi glycosylation enzymes back to the ER for quality surveillance or may be important for lipid recycling (Storrie et al. 2000). CONCLUSIONS One of the ongoing challenges in cell biology is to understand the molecular details that allow cytosolic coat proteins to fulfill their dual roles in membrane deformation and cargo capture. Structural details of the coat machinery have already provided many insights and will continue to be particularly informative. The similarities between the COPI coat and the clathrin AP complexes, which mediate trafficking in the late secretory pathway, are becoming increasingly clear, and it will be interesting to see if the resemblance extends to cargo- and effector-binding sites as these are uncovered. For both COPI and COPII, the contribution of the outer coat, i.e., the α/β /ε subcomplex for COPI and Sec13-Sec31 for COPII, is still somewhat mysterious, and awaits biochemical and structural clues about their role in propagating the coat lattice on the membrane. The efficient capture of most cargo proteins requires interactions with the vesicle coat, either directly or through adaptors. Identification of new protein export and retrieval signals and adaptors, in combination with an analysis of cargo-binding site mutants on the coat, will help to classify these sorting signals. It will also be important to understand the exceptions to the rule. What role does the COPII machinery play in the anterograde transport of large macromolecular complexes, and what is the machinery required to direct cargo proteins into the COPI-independent retrograde route? Identification of specific mutants that disrupt these processes in combination with cell-free assays and microscopic imaging should be informative. Finally, much is yet to be discovered on how transport is regulated: in particular, how the transition between protein folding and export is sensed; how sites of coat recruitment and vesicle budding are marked; how progress during membrane deformation is monitored; and how vesicle uncoating prior to fusion progresses. Although many new molecules that perform these functions no doubt await discovery, many others may already be familiar players with as yet unappreciated roles. ACKNOWLEDGMENTS We are grateful to Pierre Cosson for critical comments on the manuscript and to Bruno Antonny for helpful discussions and communicating unpublished results.

ER-GOLGI TRANSPORT 115 Work in the authors laboratories is sponsored by the Howard Hughes Medical Institute (R.S. and J.G.) and by the Swiss National Science Foundation (L.O.). The Annual Review of Cell and Developmental Biology is online at http://cellbio.annualreviews.org LITERATURE CITED Allan BB, Moyer BD, Balch WE. 2000. Rab1 recruitment of p115 into a cis-snare complex: programming budding COPII vesicles for fusion. Science 289:444 48 Amor JC, Harrison DH, Kahn RA, Ringe D. 1994. Structure of the human ADPribosylation factor 1 complexed with GDP. Nature 372:704 8 Antonny B, Huber I, Paris S, Chabre M, Cassel D. 1997. Activation of ADP-ribosylation factor 1 GTPase-activating protein by phosphatidylcholine-derived diacylglycerols. J. Biol. Chem. 272:30848 51 Antonny B, Madden D, Hamamoto S, Orci L, Schekman R. 2001. Dynamics of the COPII coat with GTP and stable analogues. Nat. Cell Biol. 3:531 37 Appenzeller C, Andersson H, Kappeler F, Hauri HP. 1999. The lectin ERGIC-53 is a cargo transport receptor for glycoproteins. Nat. Cell Biol. 1:330 34 Aridor M, Fish KN, Bannykh S, Weissman J, Roberts TH, et al. 2001. The Sar1 GTPase coordinates biosynthetic cargo selection with endoplasmic reticulum export site assembly. J. Cell Biol. 152:213 29 Aridor M, Weissman J, Bannykh S, Nuoffer C, Balch WE. 1998. Cargo selection by the COPII budding machinery during export from the ER. J. Cell Biol. 141:61 70 Austin C, Hinners I, Tooze SA. 2000. Direct and GTP-dependent interaction of ADP-ribosylation factor 1 with clathrin adaptor protein AP-1 on immature secretory granules. J. Biol. Chem. 275:21862 69 Baker EK, Colley NJ, Zuker CS. 1994. The cyclophilin homolog NinaA functions as a chaperone, forming a stable complex in vivo with its protein target rhodopsin. EMBO J. 13:4886 95 Barlowe C. 1997. Coupled ER to Golgi transport reconstituted with purified cytosolic proteins. J. Cell Biol. 139:1097 108 Barlowe C. 2003. Signals for COPII-dependent export from the ER: what s the ticket out? Trends Cell Biol. 13:295 300 Barlowe C, Orci L, Yeung T, Hosobuchi M, Hamamoto S, et al. 1994. COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum. Cell 77:895 907 Barlowe C, Schekman R. 1993. SEC12 encodes a guanine-nucleotide-exchange factor essential for transport vesicle budding from the ER. Nature 365:347 49 Belden WJ, Barlowe C. 2001a. Distinct roles for the cytoplasmic tail sequences of Emp24p and Erv25p in transport between the endoplasmic reticulum and Golgi complex. J. Biol. Chem. 276:43040 48 Belden WJ, Barlowe C. 2001b. Role of Erv29p in collecting soluble secretory proteins into ER-derived transport vesicles. Science 294:1528 31 Bi X, Corpina RA, Goldberg J. 2002. Structure of the Sec23/24-Sar1 pre-budding complex of the COPII vesicle coat. Nature 419:271 77 Bigay J, Gounon P, Robineau S, Antonny B. 2003. Lipid packing sensed by ArfGAP1 couples COPI coat disassembly to membrane bilayer curvature. Nature 426:563 66 Boehm M, Bonifacino JS. 2001. Adaptins: the final recount. Mol. Biol. Cell 12:2907 20 Bonfanti L, Mironov AA Jr, Martinez- Menarguez JA, Martella O, Fusella A, et al. 1998. Procollagen traverses the Golgi stack

116 LEE ET AL. without leaving the lumen of cisternae: evidence for cisternal maturation. Cell 95:993 1003 Bracher A, Weissenhorn W. 2002. Structural basis for the Golgi membrane recruitment of Sly1p by Sed5p. EMBO J. 21:6114 24 Brandizzi F, Snapp EL, Roberts AG, Lippincott-Schwartz J, Hawes C. 2002. Membrane protein transport between the endoplasmic reticulum and the Golgi in tobacco leaves is energy dependent but cytoskeleton independent: evidence from selective photobleaching. Plant Cell 14:1293 309 Burri L, Varlamov O, Doege CA, Hofmann K, Beilharz T, et al. 2003. A SNARE required for retrograde transport to the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA 100:9873 77 Cabrera M, Muniz M, Hidalgo J, Vega L, Martin ME, Velasco A. 2003. The retrieval function of the KDEL receptor requires PKA phosphorylation of its C-terminus. Mol. Biol. Cell 14:4114 25 Caldwell SR, Hill KJ, Cooper AA. 2001. Degradation of endoplasmic reticulum (ER) quality control substrates requires transport between the ER and Golgi. J. Biol. Chem. 276:23296 303 Cao X, Ballew N, Barlowe C. 1998. Initial docking of ER-derived vesicles requires Uso1p and Ypt1p but is independent of SNARE proteins. EMBO J. 17:2156 65 Cao X, Barlowe C. 2000. Asymmetric requirements for a Rab GTPase and SNARE proteins in fusion of COPII vesicles with acceptor membranes. J. Cell Biol. 149:55 66 Chardin P, Callebaut I. 2002. The yeast Sar exchange factor Sec12, and its higher organism orthologs, fold as beta-propellers. FEBS Lett. 525:171 73 Colley NJ, Baker EK, Stamnes MA, Zuker CS. 1991. The cyclophilin homolog ninaa is required in the secretory pathway. Cell 67:255 63 Cosson P, Demolliere C, Hennecke S, Duden R, Letourneur F. 1996. Delta- and zeta- COP, two coatomer subunits homologous to clathrin-associated proteins, are involved in ER retrieval. EMBO J. 15:1792 98 Cosson P, Lefkir Y, Demolliere C, Letourneur F. 1998. New COPI-binding motifs involved in ER retrieval. EMBO J. 17:6863 70 Cosson P, Letourneur F. 1994. Coatomer interaction with di-lysine endoplasmic reticulum retention motifs. Science 263:1629 31 Dilcher M, Veith B, Chidambaram S, Hartmann E, Schmitt HD, Fischer von Mollard G. 2003. Use1p is a yeast SNARE protein required for retrograde traffic to the ER. EMBO J. 22:3664 74 Dogic D, de Chassey B, Pick E, Cassel D, Lefkir Y, et al. 1999. The ADP-ribosylation factor GTPase-activating protein Glo3p is involved in ER retrieval. Eur. J. Cell Biol. 78:305 10 Donaldson JG, Jackson CL. 2000. Regulators and effectors of the ARF GTPases. Curr. Opin. Cell Biol. 12:475 82 Dwyer ND, Troemel ER, Sengupta P, Bargmann CI. 1998. Odorant receptor localization to olfactory cilia is mediated by ODR-4, a novel membrane-associated protein. Cell 93:455 66 Elrod-Erickson MJ, Kaiser CA. 1996. Genes that control the fidelity of endoplasmic reticulum to Golgi transport identified as suppressors of vesicle budding mutations. Mol. Biol. Cell 7:1043 58 Espenshade P, Gimeno RE, Holzmacher E, Teung P, Kaiser CA. 1995. Yeast SEC16 gene encodes a multidomain vesicle coat protein that interacts with Sec23p. J. Cell Biol. 131:311 24 Eugster A, Frigerio G, Dale M, Duden R. 2000. COP I domains required for coatomer integrity, and novel interactions with ARF and ARF-GAP. EMBO J. 19:3905 17 Eugster A, Frigerio G, Dale M, Duden R. 2004. The α- and β -COP WD40 domains mediate cargo-selective interactions with distinct di-lysine motifs. Mol. Biol. Cell 15:1011 23 Fiedler K, Veit M, Stamnes MA, Rothman JE. 1996. Bimodal interaction of coatomer with the p24 family of putative cargo receptors. Science 273:1396 99

ER-GOLGI TRANSPORT 117 Fucini RV, Chen JL, Sharma C, Kessels MM, Stamnes M. 2002. Golgi vesicle proteins are linked to the assembly of an actin complex defined by mabp1. Mol. Biol. Cell 13:621 31 Fucini RV, Navarrete A, Vadakkan C, Lacomis L, Erdjument-Bromage H, et al. 2000. Activated ADP-ribosylation factor assembles distinct pools of actin on Golgi membranes. J. Biol. Chem. 275:18824 29 Fukuda R, McNew JA, Weber T, Parlati F, Engel T, et al. 2000. Functional architecture of an intracellular membrane t-snare. Nature 407:198 202 Funato K, Riezman H. 2001. Vesicular and nonvesicular transport of ceramide from ER to the Golgi apparatus in yeast. J. Cell Biol. 155:949 59 Garcia-Mata R, Szul T, Alvarez C, Sztul E. 2003. ADP-ribosylation factor/copidependent events at the endoplasmic reticulum-golgi interface are regulated by the guanine nucleotide exchange factor GBF1. Mol. Biol. Cell 14:2250 61 Gaynor EC, Emr SD. 1997. COPI-independent anterograde transport: cargo-selective ER to Golgi protein transport in yeast COPI mutants. J. Cell Biol. 136:789 802 Gilstring CF, Melin-Larsson M, Ljungdahl PO. 1999. Shr3p mediates specific COPII coatomer-cargo interactions required for the packaging of amino acid permeases into ERderived transport vesicles. Mol. Biol. Cell. 10:3549 65 Gimeno RE, Espenshade P, Kaiser CA. 1996. COPII coat subunit interactions: Sec24p and Sec23p bind to adjacent regions of Sec16p. Mol. Biol. Cell 7:1815 23 Giraudo CG, Maccioni HJ. 2003. Endoplasmic reticulum export of glycosyltransferases depends on interaction of a cytoplasmic dibasic motif with Sar1. Mol. Biol. Cell 14:3753 66 Girod A, Storrie B, Simpson JC, Johannes L, Goud B, et al. 1999. Evidence for a COP-Iindependent transport route from the Golgi complex to the endoplasmic reticulum. Nat. Cell Biol. 1:423 30 Godi A, Santone I, Pertile P, Devarajan P, Stabach PR, et al. 1998. ADP ribosylation factor regulates spectrin binding to the Golgi complex. Proc. Natl. Acad. Sci. USA 95:8607 12 Goldberg J. 1998. Structural basis for activation of ARF GTPase: mechanisms of guanine nucleotide exchange and GTPmyristoyl switching. Cell 95:237 48 Goldberg J. 1999. Structural and functional analysis of the ARF1-ARFGAP complex reveals a role for coatomer in GTP hydrolysis. Cell 96:893 902 Goldberg J. 2000. Decoding of sorting signals by coatomer through a GTPase switch in the COPI coat complex. Cell 100:671 79 Gommel DU, Memon AR, Heiss A, Lottspeich F, Pfannstiel J, et al. 2001. Recruitment to Golgi membranes of ADP-ribosylation factor 1 is mediated by the cytoplasmic domain of p23. EMBO J. 20:6751 60 Gusarova V, Brodsky JL, Fisher EA. 2003. Apolipoprotein B100 exit from the endoplasmic reticulum (ER) is COPII-dependent, and its lipidation to very low density lipoprotein occurs post-er. J. Biol. Chem. 278:48051 58 Hammond C, Helenius A. 1994. Folding of VSV G protein: sequential interaction with BiP and calnexin. Science 266:456 58 Hanada K, Kumagai K, Yasuda S, Miura Y, Kawano M, et al. 2003. Molecular machinery for non-vesicular trafficking of ceramide. Nature 426:803 9 Harter C, Pavel J, Coccia F, Draken E, Wegehingel S, et al. 1996. Nonclathrin coat protein gamma, a subunit of coatomer, binds to the cytoplasmic dilysine motif of membrane proteins of the early secretory pathway. Proc. Natl. Acad. Sci. USA 93:1902 6 Harter C, Wieland FT. 1998. A single binding site for dilysine retrieval motifs and p23 within the gamma subunit of coatomer. Proc. Natl. Acad. Sci. USA 95:11649 54 Hoffman GR, Rahl PB, Collins RN, Cerione RA. 2003. Conserved structural motifs in intracellular trafficking pathways: structure of the gammacop appendage domain. Mol. Cell 12:615 25

118 LEE ET AL. Huang M, Weissman JT, Beraud-Dufour S, Luan P, Wang C, et al. 2001. Crystal structure of Sar1-GDP at 1.7 Å resolution and the role of the NH 2 terminus in ER export. J. Cell Biol. 155:937 48 Hudson RT, Draper RK. 1997. Interaction of coatomer with aminoglycoside antibiotics: evidence that coatomer has at least two dilysine binding sites. Mol. Biol. Cell 8:1901 10 Jackson CL, Casanova JE. 2000. Turning on ARF: the Sec7 family of guanine-nucleotideexchange factors. Trends Cell. Biol. 10:60 67 Jones B, Jones EL, Bonney SA, Patel HN, Mensenkamp AR, et al. 2003. Mutations in a Sar1 GTPase of COPII vesicles are associated with lipid absorption disorders. Nat. Genet. 34:29 31 Jones S, Newman C, Liu F, Segev N. 2000. The TRAPP complex is a nucleotide exchanger for Ypt1 and Ypt31/32. Mol. Biol. Cell 11:4403 11 Kaiser C. 2000. Thinking about p24 proteins and how transport vesicles select their cargo. Proc. Natl. Acad. Sci. USA 97:3783 85 Kaiser CA, Schekman R. 1990. Distinct sets of SEC genes govern transport vesicle formation and fusion early in the secretory pathway. Cell 61:723 33 Kim PS, Arvan P. 1998. Endocrinopathies in the family of endoplasmic reticulum (ER) storage diseases: disorders of protein trafficking and the role of ER molecular chaperones. Endocr. Rev. 19:173 202 Kuehn MJ, Herrmann JM, Schekman R. 1998. COPII-cargo interactions direct protein sorting into ER-derived transport vesicles. Nature 391:187 90 Kuehn MJ, Schekman R, Ljungdahl PO. 1996. Amino acid permeases require COPII components and the ER resident membrane protein Shr3p for packaging into transport vesicles in vitro. J. Cell Biol. 135:585 95 Kuge O, Dascher C, Orci L, Rowe T, Amherdt M, et al. 1994. Sar1 promotes vesicle budding from the endoplasmic reticulum but not Golgi compartments. J. Cell Biol. 125:51 65 Lanoix J, Ouwendijk J, Lin CC, Stark A, Love HD, et al. 1999. GTP hydrolysis by ARF-1 mediates sorting and concentration of Golgi resident enzymes into functional COP I vesicles. EMBO J. 18:4935 48 Lanoix J, Ouwendijk J, Stark A, Szafer E, Cassel D, et al. 2001. Sorting of Golgi resident proteins into different subpopulations of COPI vesicles: a role for ArfGAP1. J. Cell Biol. 155:1199 212 Lau WT, Howson RW, Malkus P, Schekman R, O Shea EK. 2000. Pho86p, an endoplasmic reticulum (ER) resident protein in Saccharomyces cerevisiae, is required for ER exit of the high-affinity phosphate transporter Pho84p. Proc. Natl. Acad. Sci. USA 97:1107 12 Lederkremer GZ, Cheng Y, Petre BM, Vogan E, Springer S, et al. 2001. Structure of the Sec23p/24p and Sec13p/31p complexes of COPII. Proc. Natl. Acad. Sci. USA 98:10704 9 Letourneur F, Gaynor EC, Hennecke S, Demolliere C, Duden R, et al. 1994. Coatomer is essential for retrieval of dilysine-tagged proteins to the endoplasmic reticulum. Cell 79:1199 207 Letourneur F, Hennecke S, Demolliere C, Cosson P. 1995. Steric masking of a dilysine endoplasmic reticulum retention motif during assembly of the human high affinity receptor for immunoglobulin E. J. Cell Biol. 129:971 78 Lewis MJ, Pelham HR. 1992. Ligand-induced redistribution of a human KDEL receptor from the Golgi complex to the endoplasmic reticulum. Cell 68:353 64 Lewis MJ, Rayner JC, Pelham HR. 1997. A novel SNARE complex implicated in vesicle fusion with the endoplasmic reticulum. EMBO J. 16:3017 24 Lippincott-Schwartz J, Liu W. 2003. Membrane trafficking: coat control by curvature. Nature 426:507 8 Luna A, Matas OB, Martinez-Menarguez JA, Mato E, Duran JM, et al. 2002. Regulation

ER-GOLGI TRANSPORT 119 of protein transport from the Golgi complex to the endoplasmic reticulum by CDC42 and N-WASP. Mol. Biol. Cell 13:866 79 Ma D, Zerangue N, Raab-Graham K, Fried SR, Jan YN, Jan LY. 2002. Diverse trafficking patterns due to multiple traffic motifs in G protein-activated inwardly rectifying potassium channels from brain and heart. Neuron 33:715 29 Majoul I, Straub M, Hell SW, Duden R, Soling HD. 2001. KDEL-cargo regulates interactions between proteins involved in COPI vesicle traffic: measurements in living cells using FRET. Dev. Cell 1:139 53 Malaisse WJ, Orci L. 1979. The role of the cytoskeleton in pancreatic B-cell function. Meth. Achiev. Exp. Pathol. 9:112 36 Malkus P, Jiang F, Schekman R. 2002. Concentrative sorting of secretory cargo proteins into COPII-coated vesicles. J. Cell Biol. 159:915 21 Mallabiabarrena A, Fresno M, Alarcon B. 1992. An endoplasmic reticulum retention signal in the CD3 epsilon chain of the T-cell receptor. Nature 357:593 96 Margeta-Mitrovic M, Jan YN, Jan LY. 2000. A trafficking checkpoint controls GABA(B) receptor heterodimerization. Neuron 27:97 106 Martinez-Menarguez JA, Geuze HJ, Slot JW, Klumperman J. 1999. Vesicular tubular clusters between the ER and Golgi mediate concentration of soluble secretory proteins by exclusion from COPI-coated vesicles. Cell 98:81 90 Matanis T, Akhmanova A, Wulf P, Del Nery E, Weide T, et al. 2002. Bicaudal-D regulates COPI-independent Golgi-ER transport by recruiting the dynein-dynactin motor complex. Nat. Cell Biol. 4:986 92 Matsuoka K, Orci L, Amherdt M, Bednarek SY, Hamamoto S, et al. 1998. COPII-coated vesicle formation reconstituted with purified coat proteins and chemically defined liposomes. Cell 93:263 75 Matsuoka K, Schekman R, Orci L, Heuser JE. 2001. Surface structure of the COPIIcoated vesicle. Proc. Natl. Acad. Sci. USA 98:13705 9 McNew JA, Parlati F, Fukuda R, Johnston RJ, Paz K, et al. 2000. Compartmental specificity of cellular membrane fusion encoded in SNARE proteins. Nature 407:153 59 Mezzacasa A, Helenius A. 2002. The transitional ER defines a boundary for quality control in the secretion of tso45 VSV glycoprotein. Traffic 3:833 49 Miller E, Antonny B, Hamamoto S, Schekman R. 2002. Cargo selection into COPII vesicles is driven by the Sec24p subunit. EMBO J. 21:6105 13 Miller EA, Beilharz TH, Malkus PN, Lee MC, Hamamoto S, et al. 2003. Multiple cargo binding sites on the COPII subunit Sec24p ensure capture of diverse membrane proteins into transport vesicles. Cell 114:497 509 Mironov AA, Mironov AA Jr, Beznoussenko GV, Trucco A, Lupetti P, et al. 2003. ERto-Golgi carriers arise through direct en bloc protrusion and multistage maturation of specialized ER exit domains. Dev. Cell 5:583 94 Mossessova E, Bickford LC, Goldberg J. 2003a. SNARE selectivity of the COPII coat. Cell 114:483 95 Mossessova E, Corpina RA, Goldberg J. 2003b. Crystal structure of ARF1 Sec7 complexed with Brefeldin A and its implications for the guanine nucleotide exchange mechanism. Mol. Cell 12:1403 11 Mossessova E, Gulbis JM, Goldberg J. 1998. Structure of the guanine nucleotide exchange factor Sec7 domain of human arno and analysis of the interaction with ARF GTPase. Cell 92:415 23 Muniz M, Morsomme P, Riezman H. 2001. Protein sorting upon exit from the endoplasmic reticulum. Cell 104:313 20 Muniz M, Nuoffer C, Hauri HP, Riezman H. 2000. The Emp24 complex recruits a specific cargo molecule into endoplasmic reticulumderived vesicles. J. Cell Biol. 148:925 30 Nehls S, Snapp EL, Cole NB, Zaal KJ, Kenworthy AK, et al. 2000. Dynamics and retention of misfolded proteins in native ER membranes. Nat. Cell Biol. 2:288 95

120 LEE ET AL. Nie Z, Hirsch DS, Randazzo PA. 2003. Arf and its many interactors. Curr. Opin. Cell Biol. 15:396 404 Nufer O, Kappeler F, Guldbrandsen S, Hauri HP. 2003. ER export of ERGIC-53 is controlled by cooperation of targeting determinants in all three of its domains. J. Cell Sci. 116:4429 40 O Kelly I, Butler MH, Zilberberg N, Goldstein SA. 2002. Forward transport. 14 3 3 binding overcomes retention in endoplasmic reticulum by dibasic signals. Cell 111:577 88 Orci L. 1982. The macro- and micro-domains in the endocrine pancreas. Diabetes 31:538 65 Orci L, Palmer DJ, Ravazzola M, Perrelet A, Amherdt M, Rothman JE. 1993. Budding from Golgi membranes requires the coatomer complex of non-clathrin coat proteins. Nature 362:648 52 Orci L, Ravazzola M, Meda P, Holcomb C, Moore HP, et al. 1991. Mammalian Sec23p homologue is restricted to the endoplasmic reticulum transitional cytoplasm. Proc. Natl. Acad. Sci. USA 88:8611 15 Otte S, Barlowe C. 2002. The Erv41p-Erv46p complex: multiple export signals are required in trans for COPII-dependent transport from the ER. EMBO J. 21:6095 104 Pagano A, Letourneur F, Garcia-Estefania D, Carpentier JL, Orci L, Paccaud JP. 1999. Sec24 proteins and sorting at the endoplasmic reticulum. J. Biol. Chem. 274:7833 40 Paris S, Beraud-Dufour S, Robineau S, Bigay J, Antonny B, et al. 1997. Role of proteinphospholipid interactions in the activation of ARF1 by the guanine nucleotide exchange factor Arno. J. Biol. Chem. 272:22221 26 Parlati F, McNew JA, Fukuda R, Miller R, Sollner TH, Rothman JE. 2000. Topological restriction of SNARE-dependent membrane fusion. Nature 407:194 98 Patel SK, Indig FE, Olivieri N, Levine ND, Latterich M. 1998. Organelle membrane fusion: a novel function for the syntaxin homolog Ufe1p in ER membrane fusion. Cell 92:611 20 Peng R, Gallwitz D. 2002. Sly1 protein bound to Golgi syntaxin Sed5p allows assembly and contributes to specificity of SNARE fusion complexes. J. Cell Biol. 157:645 55 Peng R, Grabowski R, De Antoni A, Gallwitz D. 1999. Specific interaction of the yeast cis-golgi syntaxin Sed5p and the coat protein complex II component Sec24p of endoplasmic reticulum-derived transport vesicles. Proc. Natl. Acad. Sci. USA 96:3751 56 Peyroche A, Paris S, Jackson CL. 1996. Nucleotide exchange on ARF mediated by yeast Gea1 protein. Nature 384:479 81 Poon PP, Cassel D, Spang A, Rotman M, Pick E, et al. 1999. Retrograde transport from the yeast Golgi is mediated by two ARF GAP proteins with overlapping function. EMBO J. 18:555 64 Powers J, Barlowe C. 1998. Transport of Axl2p depends on Erv14p, an ER-vesicle protein related to the Drosophila cornichon gene product. J. Cell Biol. 142:1209 22 Presley JF, Cole NB, Schroer TA, Hirschberg K, Zaal KJ, Lippincott-Schwartz J. 1997. ERto-Golgi transport visualized in living cells. Nature 389:81 85 Preuss D, Mulholland J, Franzusoff A, Segev N, Botstein D. 1992. Characterization of the Saccharomyces Golgi complex through the cell cycle by immunoelectron microscopy. Mol. Biol. Cell 3:789 803 Reilly BA, Kraynack BA, VanRheenen SM, Waters MG. 2001. Golgi-to-endoplasmic reticulum (ER) retrograde traffic in yeast requires Dsl1p, a component of the ER target site that interacts with a COPI coat subunit. Mol. Biol. Cell 12:3783 96 Rein U, Andag U, Duden R, Schmitt HD, Spang A. 2002. ARF-GAP-mediated interaction between the ER-Golgi v-snares and the COPI coat. J. Cell Biol. 157:395 404 Renault L, Kuhlmann J, Henkel A, Wittinghofer A. 2001. Structural basis for guanine nucleotide exchange on Ran by the regulator of chromosome condensation (RCC1). Cell 105:245 55

ER-GOLGI TRANSPORT 121 Roberg KJ, Crotwell M, Espenshade P, Gimeno R, Kaiser CA. 1999. LST1 is a SEC24 homologue used for selective export of the plasma membrane ATPase from the endoplasmic reticulum. J. Cell Biol. 145:659 72 Robineau S, Chabre M, Antonny B. 2000. Binding site of brefeldin A at the interface between the small G protein ADP-ribosylation factor 1 (ARF1) and the nucleotide-exchange factor Sec7 domain. Proc. Natl. Acad. Sci. USA 97:9913 18 Rossanese OW, Soderholm J, Bevis BJ, Sears IB, O Connor J, et al. 1999. Golgi structure correlates with transitional endoplasmic reticulum organization in Pichia pastoris and Saccharomyces cerevisiae. J. Cell Biol. 145:69 81 Sacher M, Barrowman J, Wang W, Horecka J, Zhang Y, et al. 2001. TRAPP I implicated in the specificity of tethering in ER-to-Golgi transport. Mol. Cell 7:433 42 Sacher M, Jiang Y, Barrowman J, Scarpa A, Burston J, et al. 1998. TRAPP, a highly conserved novel complex on the cis-golgi that mediates vesicle docking and fusion. EMBO J. 17:2494 503 Sandmann T, Herrmann JM, Dengjel J, Schwarz H, Spang A. 2003. Suppression of coatomer mutants by a new protein family with COPI and COPII binding motifs in Saccharomyces cerevisiae. Mol. Biol. Cell 14:3097 113 Sapperstein SK, Lupashin VV, Schmitt HD, Waters MG. 1996. Assembly of the ER to Golgi SNARE complex requires Uso1p. J. Cell Biol. 132:755 67 Sapperstein SK, Walter DM, Grosvenor AR, Heuser JE, Waters MG. 1995. p115 is a general vesicular transport factor related to the yeast endoplasmic reticulum to Golgi transport factor Uso1p. Proc. Natl. Acad. Sci. USA 92:522 26 Sato K, Nakano A. 2003. Oligomerization of a cargo receptor directs protein sorting into COPII-coated transport vesicles. Mol. Biol. Cell 14:3055 63 Sato K, Sato M, Nakano A. 1997. Rer1p as common machinery for the endoplasmic reticulum localization of membrane proteins. Proc. Natl. Acad. Sci. USA 94:9693 98 Sato K, Sato M, Nakano A. 2003. Rer1p, a retrieval receptor for ER membrane proteins, recognizes transmembrane domains in multiple modes. Mol. Biol. Cell. 14:3605 16 Sato M, Sato K, Nakano A. 1996. Endoplasmic reticulum localization of Sec12p is achieved by two mechanisms: Rer1p-dependent retrieval that requires the transmembrane domain and Rer1p-independent retention that involves the cytoplasmic domain. J. Cell Biol. 134:279 93 Scales SJ, Pepperkok R, Kreis TE. 1997. Visualization of ER-to-Golgi transport in living cells reveals a sequential mode of action for COPII and COPI. Cell 90:1137 48 Scheel AA, Pelham HR. 1998. Identification of amino acids in the binding pocket of the human KDEL receptor. J. Biol. Chem. 273:2467 72 Schekman R, Orci L. 1996. Coat proteins and vesicle budding. Science 271:1526 33 Schroder-Kohne S, Letourneur F, Riezman H. 1998. Alpha-COP can discriminate between distinct, functional di-lysine signals in vitro and regulates access into retrograde transport. J. Cell Sci. 111:3459 70 Shimoni Y, Kurihara T, Ravazzola M, Amherdt M, Orci L, Schekman R. 2000. Lst1p and Sec24p cooperate in sorting of the plasma membrane ATPase into COPII vesicles in Saccharomyces cerevisiae. J. Cell Biol. 151: 973 84 Shorter J, Beard MB, Seemann J, Dirac-Svejstrup AB, Warren G. 2002. Sequential tethering of Golgins and catalysis of SNAREpin assembly by the vesicle-tethering protein p115. J. Cell Biol. 157:45 62 Siddiqi SA, Gorelick FS, Mahan JT, Mansbach CM 2nd. 2003. COPII proteins are required for Golgi fusion but not for endoplasmic reticulum budding of the pre-chylomicron transport vesicle. J. Cell Sci. 116:415 27 Spang A, Herrmann JM, Hamamoto S, Schekman R. 2001. The ADP ribosylation factornucleotide exchange factors Gea1p and Gea2p have overlapping, but not redundant

122 LEE ET AL. functions in retrograde transport from the Golgi to the endoplasmic reticulum. Mol. Biol. Cell 12:1035 45 Spang A, Schekman R. 1998. Reconstitution of retrograde transport from the Golgi to the ER in vitro. J. Cell Biol. 143:589 99 Springer S, Schekman R. 1998. Nucleation of COPII vesicular coat complex by endoplasmic reticulum to Golgi vesicle SNAREs. Science 281:698 700 Springer S, Spang A, Schekman R. 1999. A primer on vesicle budding. Cell 97:145 48 Stamnes M. 2002. Regulating the actin cytoskeleton during vesicular transport. Curr. Opin. Cell Biol. 14:428 33 Stephens DJ, Pepperkok R. 2002. Imaging of procollagen transport reveals COPIdependent cargo sorting during ER-to-Golgi transport in mammalian cells. J. Cell Sci. 115:1149 60 Storrie B, Pepperkok R, Nilsson T. 2000. Breaking the COPI monopoly on Golgi recycling. Trends Cell Biol. 10:385 91 Supek F, Madden DT, Hamamoto S, Orci L, Schekman R. 2002. Sec16p potentiates the action of COPII proteins to bud transport vesicles. J. Cell Biol. 158:1029 38 Szafer E, Pick E, Rotman M, Zuck S, Huber I, Cassel D. 2000. Role of coatomer and phospholipids in GTPase-activating proteindependent hydrolysis of GTP by ADPribosylation factor-1. J. Biol. Chem. 275: 23615 19 Szafer E, Rotman M, Cassel D. 2001. Regulation of GTP hydrolysis on ADP-ribosylation factor-1 at the Golgi membrane. J. Biol. Chem. 276:47834 39 Taxis C, Vogel F, Wolf DH. 2002. ER-Golgi traffic is a prerequisite for efficient ER degradation. Mol. Biol. Cell 13:1806 18 Toonen RF, Verhage M. 2003. Vesicle trafficking: pleasure and pain from SM genes. Trends Cell Biol. 13:177 86 Trilla JA, Duran A, Roncero C. 1999. Chs7p, anew protein involved in the control of protein export from the endoplasmic reticulum that is specifically engaged in the regulation of chitin synthesis in Saccharomyces cerevisiae. J. Cell Biol. 145:1153 63 Vashist S, Kim W, Belden WJ, Spear ED, Barlowe C, Ng DT. 2001. Distinct retrieval and retention mechanisms are required for the quality control of endoplasmic reticulum protein folding. J. Cell Biol. 155:355 68 Votsmeier C, Gallwitz D. 2001. An acidic sequence of a putative yeast Golgi membrane protein binds COPII and facilitates ER export. EMBO J. 20:6742 50 Wang W, Sacher M, Ferro-Novick S. 2000. TRAPP stimulates guanine nucleotide exchange on Ypt1p. J. Cell Biol. 151:289 96 Watson PJ, Frigerio G, Collins BM, Duden R, Owen DJ. 2004. gamma-cop appendage domain structure and function. Traffic 5:79 88 Wen C, Greenwald I. 1999. p24 proteins and quality control of LIN-12 and GLP-1 trafficking in Caenorhabditis elegans. J. Cell Biol. 145:1165 75 Whyte JR, Munro S. 2002. Vesicle tethering complexes in membrane traffic. J. Cell Sci. 115:2627 37 Wieland FT, Gleason ML, Serafini TA, Rothman JE. 1987. The rate of bulk flow from the endoplasmic reticulum to the cell surface. Cell 50:289 300 Wilson DW, Lewis MJ, Pelham HR. 1993. phdependent binding of KDEL to its receptor in vitro. J. Biol. Chem. 268:7465 68 Wu WJ, Erickson JW, Lin R, Cerione RA. 2000. The gamma-subunit of the coatomer complex binds Cdc42 to mediate transformation. Nature 405:800 4 Yamaguchi T, Dulubova I, Min SW, Chen X, Rizo J, Sudhof TC. 2002. Sly1 binds to Golgi and ER syntaxins via a conserved N-terminal peptide motif. Dev. Cell 2:295 305 Yoshihisa T, Barlowe C, Schekman R. 1993. Requirement for a GTPase-activating protein in vesicle budding from the endoplasmic reticulum. Science 259:1466 68 Yuan H, Michelsen K, Schwappach B. 2003. 14-3-3 dimers probe the assembly status of

ER-GOLGI TRANSPORT 123 multimeric membrane proteins. Curr. Biol. 13:638 46 Zhang B, Cunningham MA, Nichols WC, Bernat JA, Seligsohn U, et al. 2003. Bleeding due to disruption of a cargo-specific ERto-Golgi transport complex. Nat. Genet. 34: 220 25 Zhao L, Helms JB, Brugger B, Harter C, Martoglio B, et al. 1997. Direct and GTPdependent interaction of ADP-ribosylation factor-1 with coatomer subunit beta. Proc. Natl. Acad. Sci. USA 94:4418 23 Zhao L, Helms JB, Brunner J, Wieland FT. 1999. GTP-dependent binding of ADP-ribosylation factor to coatomer in close proximity to the binding site for dilysine retrieval motifs and p23. J. Biol. Chem. 274:14198 203 Zhao X, Lasell TK, Melancon P. 2002. Localization of large ADP-ribosylation factor-guanine nucleotide exchange factors to different Golgi compartments: evidence for distinct functions in protein traffic. Mol. Biol. Cell 13:119 33