GTP hydrolysis by Sar1 mediates proof-reading for protein sorting into COPII vesicles

Transcription

1 JBC Papers in Press. Published on November 19, 2003 as Manuscript C GTP hydrolysis by Sar1 mediates proof-reading for protein sorting into COPII vesicles Ken Sato 1,2,4 and Akihiko Nakano 1,3 1 Molecular Membrane Biology Laboratory, RIKEN Discovery Research Institute, 2 PRESTO, Japan Science and Technology Corporation, Hirosawa, Wako, Saitama , and 3 Department of Biological Sciences, Graduate School of Science, University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan 4 To whom correspondence should be addressed Phone: Fax: kensato@postman.riken.go.jp Copyright 2003 by The American Society for Biochemistry and Molecular Biology, Inc.

2 Abstract Secretory proteins are transported from the endoplasmic reticulum (ER) in vesicles coated with coat protein complex II (COPII). We have developed an in vitro budding reaction comprising purified COPII proteins and cargo-reconstituted proteoliposomes. When proteoliposomes were mixed with COPII and guanine nucleotide, reconstituted cargos were concentrated into COPII vesicles. We show here that reconstituted cargo accelerates the GTP hydrolysis by Sar1p as stimulated by its GTPase-activating protein, Sec23/24p, both of which are components of the COPII coat. Furthermore, this GTP hydrolysis decreases the error of cargo sorting. We suggest that GTP hydrolysis by Sar1p promotes exclusion of improper proteins from COPII vesicles.

3 In eukaryotic cells, intracellular protein transport between the organelles of the secretory pathway is mediated by vesicular carriers that are released from a donor organelle and fuse with an appropriate acceptor organelle. The starting point of the secretory route is the ER. Transport of newly synthesized secretory proteins from the ER are sorted from ER-resident proteins and packaged into COPII-coated vesicles 1. COPII coat consists of the small GTPase Sar1p 2, the Sec23/24p complex, and the Sec13/31p complex that sequentially bind on the ER membrane 3. Budding from the ER involves activation of Sar1p-GDP to Sar1p-GTP by the ER-resident protein Sec12p, a Sar1p-specific guanine-nucleotide exchange factor 4, 5. Sar1p-GTP bound to the ER membrane promotes Sec23/24p and Sec13/31p assembly, which drives vesicle budding 6. Reconstitution experiments demonstrate that this minimal set of COPII proteins is sufficient to form COPII vesicles from synthetic liposomes 7. Several lines of evidence suggests that certain cargo possesses a binding affinity for COPII subunits, resulting selective packaging into COPII vesicles. In this process, the pre-budding complex consisting of Sar1p-GTP and Sec23/24p bound to cargo is formed on the ER membrane 8, 9. This pre-budding complex is then gathered by the Sec13/31p complex to form COPII vesicles. Conversely, the Sec23p subunit stimulates the GTP hydrolysis of Sar1p 10, which reverses the coat assembly process. Thus, the Sar1p GTPase cycle is thought to regulate both coat assembly and cargo sorting. To study precise molecular mechanisms of cargo sorting into COPII vesicles, it is important to examine GTP hydrolysis of Sar1p during coat assembly and cargo sorting. However, such studies have so far been difficult with in vitro budding assays using the fractionated organelle. We have overcome this and developed a new in vitro budding assay comprising purified coat proteins and cargo-reconstituted proteoliposomes. Emp47p, a type-i membrane protein, is specifically required for the transport of an integral membrane protein, Emp46p, from the ER 11. The C-terminal tail of Emp47p includes COPII binding signals and this protein was shown to be packaged into COPII vesicles 12. We chose these Emp46/47p and resident ER membrane protein, Ufe1p, for this study. To express Emp47p and Emp46p on E. coli membrane, the signal sequences of Emp46/47p were replaced with that of E. coli OmpA followed by two tandem repeats of Strep-tag for affinity purification. Ufe1p was tagged with a maltose-binding protein (MBP) at its N terminus to express on E. coli membrane. These proteins were separately

4 expressed on E. coli membranes and purified by affinity chromatography. The fulllength proteins were obtained in a detergent solution and the purity was analyzed by SDS-PAGE. To reconstitute the Emp46/47p and MBP-Ufe1p into liposomes, phospholipids whose composition resembles yeast ER membranes (major-minor mix), previously defined as the optimal composition for coat recruitment 7, were mixed with purified proteins in the presence of the octylglucoside. After detergent removal by dialysis, the resulting proteoliposomes were isolated by flotation in a density gradient (Fig. 1a). A quantitative assessment of the protease accessibility to the N terminus of reconstituted Emp46/47p and MBP-Ufe1p revealed that 20-30% of Emp46/47p and 100% of MBP-Ufe1p were inserted with the same relative orientation as in the ER membrane (data not shown). Both dynamic laser light scattering and transmission electron microscopy indicated a mean diameter of the resulting proteoliposomes of about 200 nm and no vesicles smaller than 100 nm were observed (Fig. 1b). We have shown before that Emp47p alone forms 600 kda homooligomer through the coiled-coil region of the luminal domain in the ER, which is essential for the exit of Emp47p from the ER. In contrast, Emp46p alone exists as a monomer in the ER and 600 kda heterooligomer formation with Emp47p is required for the Emp46p to be incorporated into COPII vesicles 11. Although no such heterooligomer formation was observed when separately purified Emp47p and Emp46p were mixed in octylglucoside solution before reconstitution, they assembled efficiently into a 600 kda heterooligomer when they were reconstituted into liposomes as judged by gel filtration (Fig. 1c). When proteoliposomes were prepared with the coiled-coil deleted mutant, Emp47p-D , or Emp46p, these proteins were eluted at lower molecular mass positions, most consistent with their monomer sizes. Thus, it seems unlikely that the 600 kda complex is nonspecific aggregates or monomeric forms of each molecule surrounded by detergent micelles. We conclude that Emp47p and Emp46p spontaneously form heterooligomers on the reconstituted membrane. We tested the ability of COPII components to bind to the proteoliposomes (Fig. 1d). Assembly of Sec23/24p and Sec13/31p onto synthetic liposomes (major-minormix) was observed in a Sar1p-GMP-PNP-dependent manner as has been shown before 7. Similarly, Sec23/24p and Sec13/31p formed a Sar1p-GMP-PNP dependent complex on Emp47p reconstituted proteoliposomes. We next addressed whether COPII components might bind directly to reconstituted Emp47p. It has been shown previously that Sar1p-

5 GMP-PNP binds to neutral liposomes (PC/PE), but that the subsequent binding of Sec23/24p requires the presence of acidic phospholipids 7. Emp47p was reconstituted into the neutral liposomes that do not support the Sec23/24p recruitment and COPII recruitment was assessed. When Emp47p alone was reconstituted into the neutral liposomes, the efficiency of incorporation was ~10% compared with the major-minor mix liposomes. Successful reconstitution of Emp47p into the neutral liposomes required the co-existence of MBP-Ufe1p for unknown reasons. The presence of Emp47p in the neutral liposomes supported the recruitment of Sec23/24p and Sec13/31p in a Sar1p- GMP-PNP dependent manner, although the recruitment efficiency was quite low (less than 5% of the major-minor mix). Because the carboxy-terminal region of Emp47p in the neutral liposomes is the only portion that could interact directly with Sec23/24p, we conclude that COPII directly binds to reconstituted Emp47p. Since COPII binding to proteoliposomes behaved similarly to plain liposomes, we assessed the real-time coat assembly using a light-scattering assay (Fig. 1e). Binding of COPII proteins to liposomes leads to an increase in mass that can be detected by monitoring changes in light scattering 13. The addition of GMP-PNP to the Sec23/24p, Sec13/31p, and Sar1p-loaded liposomes generated a large increase in light scattering within 350 sec. In contrast, the kinetics of Sec23/24p and Sec13/31p recruitment to the Emp47p-reconstituted proteoliposomes was increased (150 sec). Inclusion of monomeric Emp47p-D in liposomes, which we have previously shown to bind COPII but not allow COPII vesicle incorporation 11, yielded a rapid increase in light scattering, which was almost identical to the case of Emp47p. Our data demonstrate that the presence of cargo on membranes enhances the rate of COPII recruitment and suggest that the cargo molecules actively participate in the recruitment of the COPII coat. This high affinity binding of COPII subunits to cargo might significantly improve the capture of cargo molecules as does the binding to acidic phospholipids. Based on the finding that COPII coat can bind directly to reconstituted Emp47p with an increased rate, we next asked whether the presence of cargo in the membrane affects COPII budding from proteoliposomes. It has been shown before that synthetic COPII vesicles bud from protein-free liposomes in the presence of Sec23/24p, Sec13/31p, and Sar1p-GMP-PNP 7. Proteoliposomes containing Emp46/47p and MBP- Ufe1p were incubated with purified Sec23/24p, Sec13/31p, and Sar1p in the presence of GMP-PNP or GDP, and then the reaction was centrifuged on a linear sucrose-density

6 gradient to separate the coated vesicles from the parental proteoliposomes. The recovery of lipids in each fraction was monitored by measuring the fluorescence of NBDphospholipids included in the proteoliposomes. When incubated with GMP-PNP, the high-density peak (fractions 5-9) was detected with a corresponding decrease in the low-density peak (fractions 1-3), whereas the proteoliposomes incubated with GDP remained at the top of the gradient (Fig. 2a). The high-density fractions expected to contain coated vesicles were further analyzed by electron microscopy to examine their identity and we confirmed that this high-density peak indeed consisted of coated vesicles (Fig. 2c). The fractions were also analyzed by immunoblotting to establish the distribution of Emp46/47p. Most of Emp46/47p was found (~80%) in the high-density peak (fractions 5-8) along with the phospholipid peak when incubated with GMP-PNP, whereas almost all of Emp46/47p remained at the top of the gradient in the presence of GDP. In contrast, when proteoliposomes containing monomeric Emp47p-D and Emp46p were tested, both of which were shown to be deficient in the ER export, a decreased level of coated vesicle budding was observed (Fig. 2b). Considering that monomeric forms of Emp47p and Emp46p can bind the COPII coat, our data indicate that the simple binding of COPII components to the membrane via monomeric cargo is not sufficient for efficient COPII vesicle production, but binding to oligomerized cargo is required. In other words, COPII binding to oligomerized cargo with a particular spatial arrangement on the membrane probably promotes efficient coat polymerization which drives vesicle formation. We next monitored the extent of incorporation of Emp46/47p and MBP-Ufe1p into vesicles generated from proteoliposomes (Fig. 2d). Starting proteoliposomes were compared with gradient-purified COPII vesicles by quantitative immunoblotting normalized to the intensity of NBD phospholipid fluorescence. A range of fold concentration of Emp46/47p was detected in several replications of this experiment. This efficiency of sorting is not what can be achieved with microsomes and purified COPII components (3~4-fold enrichment of Emp47p relative to phospholipid; unpublished data). One reason for this might be that the parental proteoliposomes (mean diameter 200 nm) used in this assay can vesiculate into only a small number of COPII vesicles (50~60 nm). Furthermore, our system includes only 20~30% of correctly oriented Emp46/47p, thereby the opposite oriented species provide low estimation of cargo enrichment. On the other hand, control protein, MBP- Ufe1p, was neither concentrated nor depleted in proteoliposome-derived COPII vesicles

7 (~1.1-fold enrichment). Although it is not clear whether the ER resident protein Ufe1p has a nature to prevent its ER export, the above result showed that our system comprising the minimal set of COPII proteins do not includes active exclusion mechanisms of ER residents. Sec23/24p interacts with membrane-bound Sar1p-GTP and the export motifs of transmembrane cargos to form the pre-budding complex, which is prerequisite for Sec13/31p recruitment to yield COPII vesicles. The fact that the GTPase activating protein (GAP) of Sar1p, Sec23p, is a subunit of the coat complicated this event and if and how this step is regulated on the membrane by cargo molecules remains unknown. This issue is now testable with our budding-competent proteoliposomes. We examined a single round of real-time coat assembly and disassembly in the presence of GTP by light scattering as reported before (Fig. 3a) 13. On Sar1p-GTP pre-loaded plain liposomes, addition of Sec23/24p yielded an instant increase in light scattering that persisted for ~200 sec. Strikingly, addition of Sec23/24p to Emp47p- or Emp47p-D reconstituted proteoliposomes also generated an immediate binding signal, but this signal was more transient, continuing for ~120 sec. We next examined whether this accelerated pre-budding complex disassembly in the presence of the cargo might be due to activated GTP hydrolysis by Sar1p (Fig. 3b). We directly tested the GTP hydrolysis by Sar1p as stimulated by Sec23/24p using a tryptophan fluorescence assay that perceives Sar1p-GTP and Sar1p-GDP as established before 13. In the presence of Emp47p or Emp47p-D , the GTP hydrolysis by Sar1p was ~1.4-fold more efficient than without cargo, in agreement with the more transient binding in the presence of cargo detected by light scattering. To study the role of GTP hydrolysis by Sar1p more precisely in terms of cargo sorting, we performed in vitro budding assays with microsomes in the presence of GMP-PNP or GTP to determine whether GTP hydrolysis influenced the cargo sorting into COPII vesicles (Fig. 3c). COPII vesicle fractions were subjected to SDS-PAGE followed by immunoblotting. The signals were then compared with a standard curve of known amounts of microsomal membranes. When a budding reaction was performed in the presence of GTP, a lower level (0.3%) of monomeric Emp47p-D was detected in the COPII vesicle fraction as compared to wild-type oligomeric Emp47p (3.8%) as we reported before. In marked contrast, when vesicles were generated in the presence of GMP-PNP, the level of Emp47p-D was significantly higher (1.6%).

8 The levels of other cargos found in the vesicle fractions such as wild-type Emp47p, Sec22p, and Sed5p were similar under GTP and GMP-PNP conditions, indicating a preferential incorporation of monomeric Emp47p-D in the absence of Sar1p GTP hydrolysis. The Emp47p-D mutant lacking the ability to assemble into an oligomer might mimic the transient state of newly synthesized unassembled Emp47p in the ER. Such an unassembled cargo has an ability to bind coat subunits via the C- terminal signal but should be efficiently excluded from COPII vesicles. From these data, we suggest that Sar1p selectively promotes exclusion of improper proteins from COPII vesicles by GTP hydrolysis. It has been shown that GTP hydrolysis by Sar1p triggers uncoating of the COPII coat from the vesicles, in which case Sar1p may not play a direct role for cargo sorting. This scenario is based on the facts that in vitro generated COPII vesicles in the presence of GMP-PNP or GTPgS remain coated and could be a functional intermediate of transport after coat removal As GTP hydrolysis of Sar1p leads to the release of COPII coat from the ER membrane, i.e. COPII coat has a short intrinsic lifetime on membranes 13, a counteracting mechanism was thought to be needed to support vesicle formation The proposed model was that the GTP hydrolysis by Sar1p in the prebudding complex is temporally restricted to ensure cargo capture and coat polymerization. In the light of our new findings, an alternative model should be developed to explain the cargo-dependent stimulation of Sar1p GTPase during COPII vesicle formation, and we favor the existence of an additional role for GTP hydrolysis by Sar1p besides that shown in uncoating. We propose here a kinetic proof-reading model outlined in Figure 4. The pre-budding complex with assembled cargo has a high rate to polymerize into COPII vesicles before GTP hydrolysis by Sar1p (Fig. 2a). By contrast, the pre-budding complex with unassembled cargo also has an ability to bind COPII coat (Fig. 1e), but polymerization proceeds slowly (Fig. 2b). The high GTP hydrolysis rate of Sar1p counteracts the pre-budding complex polymerization with unassembled cargo, ensuring that only the pre-budding complex with correct cargo is successfully incorporated into COPII vesicles. COPI-dependent sorting of Golgi resident proteins has been reported to require GTP hydrolysis by ARF1 16, 17. Although similar components are required for both COPII- and COPI-dependent transport and similar basic mechanisms operate commmonly, they may adopt different ways for protein sorting. COPI binds to the

9 membrane proteins that are retrieved back from the Golgi to the ER via the C-terminal K(X)KXX motifs in their cytosolic tails 18, 19. However, no such a simple transplantable motif for COPII binding has yet been identified, and multiple signals including oligomer formation are required in many cases 20, 21, 22, 23, 11. A requirement for multiple signals in ER cargos might be important in terms of the ER quality control. ER contains a certain amount of newly synthesized unassembled cargo, which should be segregated from secretory proteins. A requirement for combinational signals could operate efficient exclusion of unassembled cargo from COPII vesicles, thereby the ER might require different mechanisms from Golgi in protein sorting. Our results indicating stimulation of the Sar1p GTPase activity by cargo do not exclude the contribution of other factors in this process. Nevertheless, the unpredicted function of Sar1p during cargo sorting shown in this study provides an avenue toward the investigation of molecular mechanisms of cargo sorting during vesicle formation.

10 Acknowledgments We are grateful to Randy Schekman for providing strains and antibodies. This work was supported by PRESTO, Japan Science and Technology Corporation, by Grants-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan, and by a fund from the Bioarchitect Research Project of RIKEN.

11 Methods Plasmids To construct pkse131, pksy132, and pkse133 carrying the Strepx2-myc- EMP47, Strepx2-HA-EMP46, and Strepx2-myc-EMP47-D , respectively, these fragments were generated by PCR and subcloned into pask-iba4 (Sigma). The plasmid pkse136 encoding MBP-UFE1 was constructed by cloning the PCR fragment of UFE1 into pmal-p2 (New England Biolabs). Protein expression and purification Strepx2-myc-Emp47p, Strepx2-HA-Emp46p, and Strepx2-myc-Emp47p-D were purified from E. coli JM109 cells containing the respective plasmids. The cells were cultured at 30 C and induced with tetracycline (200 mg/l). After 3hr, the cells were collected and washed once with buffer A (20 mm HEPES-KOH, ph 7.4, 150 mm KOAc). The total membrane fraction was prepared by sonication in buffer A with 0.5 mm PMSF and centrifugation at 42,000rpm (Hitachi P45AT rotor). The membrane was solubilized with 2.5% (w/v) N-octyl-b-D-glucopyranoside in buffer B (20 mm HEPES- KOH, ph 6.8, 160 mm KOAc) containing 0.5 mm PMSF at 4 C for 1hr. After the removal of insoluble materials by centrifugation, supernatant was applied to a Streptactin-Sepharose (Sigma) column equilibrated with buffer B containing 1.25% (w/v) N-octyl-b-D-glucopyranoside. The column was washed with the same buffer, and then Strep-tagged protein was eluted with buffer B containing 2.5 mm desthiobiotin. MBP-Ufe1p was purified from JM109 cells carrying the plasmid pkse136. After 3hr induction with IPTG, the total membrane fraction was solubilized by the same method as above. The solubilized membrane was applied to an Amylose resin (New England Biolabs), washed with the same buffer, and then MBP-Ufe1p was eluted with buffer B containing 10 mm maltose. Preparation of proteoliposomes Lipid composition of the major-minor mix and PC/PE has been described 7. All proteoliposomes used contained 80% phospholipids and 20% of cholesterol. Aliquots of lipid solutions in chloroform (0.99 mg lipids) were placed in a glass tube and solvents were removed by a gentle stream of nitrogen to yield a lipid film. Lipids were hydrated with buffer B containing 1.25% (w/v) N-octyl-b-D-glucopyranoside by bath sonication,

12 and then protein solution was added (total 1 ml). After incubation at room temperature for 30 min, the solution was dialyzed against 2 liters of buffer B containing Biobeads SM-2 at room temperature for 3 hr and then overnight at 4 C. Proteoliposomes were recovered by flotation in a Nycodenz (Sigma) step gradient. Each 1 ml dialysate was mixed with the same amount of 80% (w/v) Nycodenz in buffer B and then overlaid with 2 ml of 30% (w/v) Nycodenz in buffer B followed by 1 ml of buffer B. The sample was then centrifuged in an RPS65T rotor (Hitachi) at 60,000 rpm for 1 hr at 4 C. The proteoliposomes were collected from the 0/30% Nycodenz interface followed by ultracentrifugation to recover proteoliposomes, which were resuspended in buffer B and stored at -80 C. Size distribution profiles of proteoliposomes were measured by dynamic light scattering using LB-500 (Horiba) according to the manufacturer s instructions. Solubilization of proteoliposomes followed by size exclusion chromatography of reconstituted Emp46/47p was performed as described previously 11. Assay of COPII binding to proteoliposomes Reactions contained 1.6 mg Sar1p, 1.7 mg Sec23/24p, 2mg Sec13/31p, 0.1 mm GMP-PNP or GDP, and 9.9 mg (lipids) proteoliposomes in buffer B containing 1 mm MgOAc (total volume 50 ml). After incubation at 30 C for 15 min, reactions were mixed with the equal volume of 80% (w/v) Nycodenz in buffer B and then overlaid with 75 ml of 30% (w/v) Nycodenz in buffer B followed by 10 ml of buffer B. The resulting step gradient was centrifuged at 100,000 rpm in a Hitachi 100AT3 rotor for 1hr. Samples (25 ml) were collected from the top and the recovery of NBD-labeled phospholipids were determined by the fluorescence. After normalization for lipid recovery, proteins in the fractions were separated by SDS-PAGE followed by immunoblotting. Vesicle budding assay For proteoliposome budding assays, proteoliposomes (corresponding to 9.9 mg phospholipids) were incubated with 8.4 mg Sar1p, 10 mg Sec23/24p, 11 mg Sec13/31p, and 0.1 mm GMP-PNP or GDP in buffer B containing 1 mm MgOAc (total volume 100 ml) at 30 C for 30 min. The reaction was loaded on a 20-40% sucrose gradient in buffer B with 1 mm MgOAc, and centrifuged for 22 hr in an RPS40T rotor at 35,000 rpm at 4 C. Fractions (0.5 ml) were collected from the top of the gradient, and the fluorescence peak fraction was used as the vesicle fraction. Proteins were separated by SDS-PAGE

13 followed by immunoblotting. For microsome budding assays, microsomes expressing indicated proteins were incubated with saturate amount of COPII proteins with 0.1 mm guanine nucleotides, 1 mm ATP, and ATP regeneration system as described before 11. Electron microscopy Vesicles were generated from proteoliposomes as described above, and the COPII vesicle peak was collected by ultracentrifugation and processed for thin-section electron microscopy as described previously 24. Light scattering and tryptophan fluorescence assays The scattering of light (l=350 nm) was measured at 90 in a Hitachi fluorescence spectrophotometer (F-2500) equipped with a thermostatically controlled cell holder. The cuvette initially contained a suspension of proteoliposomes (corresponding to 9.9 mg phospholipids) in buffer B containing 1 mm MgCl 2. Sar1p, Sec23/24p, Sec13/31p, and GMP-PNP or GTP were added at the indicated times. Tryptophan fluorescence was recorded at 340 nm upon excitation at 298 nm. Free Mg 2+ concentration in the reaction was temporarily decreased by the addition of EDTA to accelerate GDP/GTP exchange as described before. All experiments were performed at 25 C.

14 References 1. Schekman, R. & Orci, L. Coat proteins and vesicle budding. Science 271, (1996). 2. Nakano, A. & Muramatsu, M. A novel GTP-binding protein, Sar1p, is involved in transport from the endoplasmic reticulum to the Golgi apparatus. JCellBiol 109, (1989). 3. Barlowe, C. et al. COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum. Cell 77, (1994). 4. Nakano, A., Brada, D. & Schekman, R. A membrane glycoprotein, Sec12p, required for protein transport from the endoplasmic reticulum to the Golgi apparatus in yeast. JCellBiol107, (1988). 5. Barlowe, C. & Schekman, R. SEC12 encodes a guanine-nucleotide-exchange factor essential for transport vesicle budding from the ER. Nature 365, (1993). 6. Springer, S., Spang, A. & Schekman, R. A primer on vesicle budding. Cell 97, (1999). 7. Matsuoka, K. et al. COPII-coated vesicle formation reconstituted with purified coat proteins and chemically defined liposomes. Cell 93, (1998). 8. Kuehn, M. J., Herrmann, J. M. & Schekman, R. COPII-cargo interactions direct protein sorting into ER-derived transport vesicles. Nature 391, (1998). 9. Aridor, M., Weissman, J., Bannykh, S., Nuoffer, C. & Balch, W. E. Cargo selection by the COPII budding machinery during export from the ER. JCell Biol 141, (1998). 10. Yoshihisa, T., Barlowe, C. & Schekman, R. Requirement for a GTPaseactivating protein in vesicle budding from the endoplasmic reticulum. Science 259, (1993). 11. Sato, K. & Nakano, A. Oligomerization of a Cargo Receptor Directs Protein Sorting into COPII-coated Transport Vesicles. Mol Biol Cell 14, (2003). 12. Sato, K. & Nakano, A. Emp47p and Its Close Homolog Emp46p Have a Tyrosine-containing Endoplasmic Reticulum Exit Signal and Function in Glycoprotein Secretion in Saccharomyces cerevisiae. Mol Biol Cell 13, (2002).

15 13. Antonny, B., Madden, D., Hamamoto, S., Orci, L. & Schekman, R. Dynamics of the COPII coat with GTP and stable analogues. NatCellBiol3, (2001). 14. Oka, T. & Nakano, A. Inhibition of GTP hydrolysis by Sar1p causes accumulation of vesicles that are a functional intermediate of the ER-to-Golgi transport in yeast. JCellBiol124, (1994). 15. Antonny, B. & Schekman, R. ER export: public transportation by the COPII coach. Curr Opin Cell Biol 13, (2001). 16. Lanoix, J. et al. GTP hydrolysis by arf-1 mediates sorting and concentration of Golgi resident enzymes into functional COP I vesicles. Embo J 18, (1999). 17. Lanoix, J. et al. Sorting of Golgi resident proteins into different subpopulations of COPI vesicles: a role for ArfGAP1. JCellBiol155, (2001). 18. Cosson, P. & Letourneur, F. Coatomer interaction with di-lysine endoplasmic reticulum retention motifs. Science 263, (1994). 19. Lewis, M. J. & Pelham, H. R. SNARE-mediated retrograde traffic from the Golgi complex to the endoplasmic reticulum. Cell 85, (1996). 20. Doms, R. W., Keller, D. S., Helenius, A. & Balch, W. E. Role for adenosine triphosphate in regulating the assembly and transport of vesicular stomatitis virus G protein trimers. JCellBiol105, (1987). 21. Emery, G., Rojo, M. & Gruenberg, J. Coupled transport of p24 family members. JCellSci113 ( Pt 13), (2000). 22. Otte, S. & Barlowe, C. The Erv41p-Erv46p complex: multiple export signals are required in trans for COPII-dependent transport from the ER. Embo J 21, (2002). 23. Malkus, P., Jiang, F. & Schekman, R. Concentrative sorting of secretory cargo proteins into COPII-coated vesicles. JCellBiol159, (2002). 24. Matsuoka, K. & Schekman, R. The use of liposomes to study COPII- and COPIcoated vesicle formation and membrane protein sorting. Methods 20, (2000).

16 Legends to figures Figure 1. Reconstitution of proteoliposomes from purified Emp46/47p and MBP-Ufe1p (a) Purified Emp46p, Emp47p, and MBP-Ufe1p (Input, lanes1-3) were analyzed by SDS-PAGE and stained with Coomassie blue. Reconstituted proteoliposomes with indicated proteins are shown in lanes 4-6. The bands labeled by asterisks are SDS resistant dimer(*) and tetramer(**) of Emp47p as revealed by immunoblotting. (b) Size distribution of proteoliposomes. Proteoliposomes reconstituted with Emp46/47p and MBP-Ufe1p were subjected to dynamic light scattering measurement. The result is shown as the calculated vesicle size. (c) Size exclusion chromatography of reconstituted Emp46/47p. Proteoliposomes reconstituted with indicated proteins were solubilized with n-dodecylmaltoside and then applied to a G3000 column. Emp47p and Emp46p in each fraction was detected with anti-myc and anti-ha antibodies, respectively. (d) Binding of COPII proteins to the Emp47p reconstituted proteoliposomes. The majorminor mix (lanes 1-4) or neutral (PC/PE) (lanes 5-8) liposomes were reconstituted with purified Emp47p (lanes 3, 4, 7, and 8) and binding of Sar1p, Sec23/24p, and Sec13/31p in the presence of GDP or GMP-PNP was assessed by floatation. Membrane-associated proteins were resolved by SDS-PAGE and detected by immunoblotting. (e) Real-time binding of the COPII coat to the proteoliposomes. The light scattering of a suspension of major-minor mix liposomes (black trace), Emp47p- (final 5.0 mg/ml) (red trace) or Emp47p-D (final 5.5 mg/ml) (blue trace) reconstituted major-minor mix liposomes was continuously monitored in the presence of 1 mm Sar1p, 135 nm Sec23/24p, 250 nm Sec13/31p, and 0.1 mm GMP-PNP. GMP-PNP was added at the specific time point as indicated. Figure 2. COPII vesicle formation from proteoliposomes Proteoliposomes reconstituted with Emp46p/47p/MBP-Ufe1p (a) or Emp46p/47p-D /MBP-Ufe1p (b) into major-minor mix liposomes were incubated with COPII proteins in the presence of GMP-PNP (closed box) or GDP (open box) at 30 C for 30 min. The mixture was sedimented to equilibrium on a sucrose-density gradient. Upper panel, distribution of lipids on a sucrose-density gradient measured by the fluorescence of NBD-phospholipids. Lower panel, distribution of reconstituted proteins detected by immunoblottiong. (c) Morphology of COPII-coated proteoliposomes. A view of the

17 budded proteoliposomes in fractions 5-8 from (A). The bar represents 100 nm. (d) Concentration of Emp46/47p into synthetic COPII vesicles. Reconstituted proteins in the donor proteoliposomes and in the vesicle fraction were analyzed by immunoblotting. COPII vesicle fractions 5-8 from (a) were used as a vesicle fraction. The amounts of fluorescent lipids in the vesicle fraction is the same as in proteoliposome lane x1. The loaded amounts of proteoliposomes in different lanes are multiplied by the x1 lane as indicated. Figure 3. Cargo accelerates the GTP hydrolysis by Sar1p as stimulated by Sec23/24p, which influenced the cargo sorting into COPII vesicles. (a) Emp47p accelerates disassembly of pre-budding complex after the addition of Sec23/24p. The reaction initially contained major-minor mix liposomes (black trace), Emp47p- (final 5.5 mg/ml) (red trace) or Emp47p-D (final 5.0 mg/ml) (blue trace) reconstituted majorminor mix liposomes with 1 mm Sar1p and 0.1 mm GTP. After 10 min incubation, Sec23/24p (135 nm) was added and the light scattering of the suspension was monitored. (b) Measurements of the Sar1p GTPase activity as stimulated by Sec23/24p by tryptophan fluorescence. Sar1p (0.5 mm) was activated by the addition of GTP (30 mm) in the presence of major-minor mix liposomes (black trace), Emp47p- (final 11 mg/ml) (red trace) or Emp47p-D (final 10 mg/ml) (blue trace) reconstituted major-minor mix liposomes. Sec23/24p (67 nm) was then added to promote GTP hydrolysis. The conformational change of Sar1p upon GDP/GTP exchange was monitored by tryptophan fluorescence and only the deactivation phase is shown. (c) Packaging of Emp47p-D into COPII vesicles is influenced by GTP hydrolysis. In vitro budding reactions were performed in the presence of GTP (lane 2) or GMP-PNP (lane 3) with microsomes expressing Emp47p-D or wild-type Emp47p. Total reactions (5%, 3%, and 1%), budded COPII vesicles isolated after incubation with (lanes 2 and 3) or without (lane 1) COPII proteins were analyzed by SDS-PAGE followed by immunoblotting. Figure 4. A model for the role of GTP hydrolysis by Sar1p during vesicle formation.

18 Sato & Nakano Fig. 1abc a MBP-Ufe1p Emp47p Emp46p Emp47p Emp47p/MBP-Ufe1p Emp46/47p/MBP-Ufe1p b Input c Reconstituted ** * MBP-Ufe1p Emp47p Emp46p Frequency (%) Diameter (nm) Vo 669k 158k 43k 13.7k Emp47p Emp46p Emp46p Emp47p- D Emp46/47p Proteoliposomes Emp46p Proteoliposomes Emp47p-D Proteoliposomes

19 Sato & Nakano Fig. 1de d MBP-Ufe1p Emp47p GMP-PNP GDP Sec31p e Sec23p Sar1p 2.3 Major/Minor mix GMP-PNP PC/PE Light scattering Liposome Emp47p Emp47p-D Time (sec)

20 a 30 Sato & Nakano Fig. 2ab Lipid recovery (%) GDP GMP-PNP Sucrose% b Emp47p Emp47p Emp46p Lipid recovery (%) 0 1 top GDP GMP-PNP bottom Fraction GDP GMP-PNP Fraction Emp47p-D GDP Emp47p -D Emp46p top bottom GMP-PNP

21 Sato & Nakano Fig. 2cd c d MBP-Ufe1p proteoliposomes x4 x3 x2 x1 vesicle fraction Emp47p Emp46p Major-Minor mix

22 Sato & Nakano Fig. 3ab a 2.4 Sec23/24p Liposome Emp47p Emp47p-D Light scattering 2.1 b Fluorescence change Time (sec) Liposome Emp47p Emp47p-D Sec23/24p Time (sec) 600

23 Sato & Nakano Fig. 3c c Emp47p-D Emp47p Sec22p Sed5p Sec61p

24 Assembled cargo GTP- Sar1p GDP- Sar1p COPII coat FAST GDP GTP FAST FAST ER SLOW Sato & Nakano Fig. 4 Unassembled cargo GDP GTP ER

25 Reconstitution of COPII vesicle formation from cargo reconstituted proteoliposomes reveals the potential role of GTP hydrolysis by Sar1p in protein sorting Ken Sato and Akihiko Nakano J. Biol. Chem. published online November 19, 2003 Access the most updated version of this article at doi: /jbc.C Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts