Sfb2p, a Yeast Protein Related to Sec24p, Can Function as a Constituent of COPII Coats Required for Vesicle Budding from the Endoplasmic Reticulum*

The COPII coat is required for vesicle budding from the endoplasmic reticulum (ER), and consists of two heterodimeric subcomplexes, Sec23p/Sec24p, Sec13p/Sec31p, and a small GTPase, Sar1p. We characterized a yeast mutant, anu1 (abnormalnuclear morphology) exhibiting proliferated ER as well as abnormal nuclear morphology at the restrictive temperature. Based on the finding that ANU1 is identical to SEC24, we confirmed a temperature-sensitive protein transport from the ER to the Golgi in anu1-1/sec24-20 cells. Overexpression ofSFB2, a SEC24 homologue with 56% identity, partially suppressed not only the mutant phenotype ofsec24-20 cells but also rescued theSEC24-disrupted cells. Moreover, the yeast two-hybrid assay revealed that Sfb2p, similarly to Sec24p, interacted with Sec23p. InSEC24-disrupted cells rescued by overexpression ofSFB2, some cargo proteins were still retained in the ER, while most of the protein transport was restored. Together, these findings strongly suggest that Sfb2p functions as the component of COPII coats in place of Sec24p, and raise the possibility that each member of the SEC24 family of proteins participates directly and/or indirectly in cargo-recognition events with its own cargo specificity at forming ER-derived vesicles.

In eukaryotic cells, protein transport along the secretory pathway is mediated by vesicle budding from a donor membrane and by specifically fusing the formed vesicle to an acceptor organelle. Vesicle budding is driven by the recruitment of specific coat proteins to a donor membrane (1,2). The COPII coat, that is required for the vesicle budding from endoplasmic reticulum (ER), 1 consists of heterodimeric protein complexes, Sec23p/Sec24p, Sec13p/Sec31p, and a small GTPase Sar1p.
COPII-coated vesicle formation begins with recruitment of Sar1p to the ER membrane where Sar1p-GDP is converted to Sar1p-GTP by a specific guanine nucleotide exchanging factor, Sec12p. Subsequently, Sec23p/Sec24p binds a membranebound Sar1p-GTP to form Sar1p-Sec23p/Sec24p prebudding complex. Finally, Sec13p/Sec31p binds the prebudding complexes to cross-link them, resulting in vesicle budding (3)(4)(5)(6). An additional protein, Sec16p, is also required for COPII-coated vesicle budding. Sec16p is a peripheral membrane protein tightly associated with the ER, and can bind COPII coat components, Sec23p, Sec24p, and Sec31p via its distinct domains. Thus Sec16p is considered to serve as a scaffold to recruit and/or assemble COPII coat components (5).
In addition to secretory and membrane proteins (cargo proteins), vesicle machinery proteins, such as v-SNAREs (vesicle targeting proteins) and cargo receptors are also specifically concentrated into COPII-coated vesicles (7)(8)(9)(10)(11). Cargo receptors are membrane proteins cycling between the ER and Golgi that interact with specific cargo molecules and facilitate their uptake into transport vesicles (12). For instance, several yeast proteins belonging to the p24 family are known to be involved in the transport of a subset of cargo proteins (9,10,13,14). At least eight members of the yeast p24 family have been identified as putative cargo receptors, i.e. EMP24, ERV25, and ERP1-ERP6, all encoding type I integral membrane proteins (9,10,14). Genetic and biochemical studies revealed that Emp24p, Erv25p, Erp1p, and Erp2p function in a heteromeric complex (10,14), and are responsible for the efficient and selective export of invertase and/or Gas1p out of the ER (9,10,13,14). However, clear cut evidence for a direct interaction between these cargo proteins and p24 heteromeric complexes has not been observed yet.
Several lines of evidence strongly suggest that selective export from the ER and COPII-coated vesicle formation should be coupled. Matsuoka et al. (15) demonstrated that v-SNAREs but not ER resident proteins are concentrated in synthetic COPIIcoated vesicles in a reconstituted liposome budding assay. It was also demonstrated in both yeast and mammals that some membrane cargo and vesicle machinery proteins, such as v-SNAREs, Emp24p, and vesicular stomatitis virus glycoprotein (VSV-G), specifically interact with Sar1p-Sec23p/Sec24p prebudding complex (16 -18).
Although Sec24p has binding domains for Sec23p, Sec31p, and Sec16p (19 -21) and a zinc finger-like domain essential for its function (21), the precise role of Sec24p itself in vesicle formation and cargo selection remains to be elucidated. The Saccharomyces genome data base (22) shows that there are two additional genes related to SEC24, designated SFB2 (YNL049c; 56% identity) and SFB3 (YHR098c; 23% identity). While neither SFB2 nor SFB3 is essential for growth (23), the zinc finger-like domain is conserved among the SEC24 family of proteins. Recently, SFB3 was identified as LST1, the gene exhibiting synthetic lethal interactions with SEC13 and all of other COPII genes. Lst1p has an ability to bind Sec23p, and chromosomal deletion of LST1 specifically inhibits export of the plasma membrane proton-ATPase (Pma1p) from the ER. Thus Lst1p is considered to be a specialized form of the COPII subunit required for efficient packaging of Pma1p into COPIIcoated vesicles (24). On the other hand, very little is known about the role of SFB2. Analysis of SFB2 function may provide us with further information about the roles of the SEC24 family of proteins in vesicle formation and/or cargo selection.
We have obtained four temperature-sensitive mutants, designated anu (abnormal nuclear morphology), that exhibited abnormal nuclear morphology at the restrictive temperature. All four genes designated ANU1-ANU4 were identical to SEC24, SEC28, SEC13, and SEC18, respectively, genes involved in vesicle transport between the ER and Golgi (25). anu1-1/sec24-20 mutant cells are severely defective in vesicle transport from the ER to the Golgi at the restrictive temperature, resulting in accumulation of both nascent proteins and membranes in the ER. We examined whether the overexpression of SEC24 homologues suppresses sec24-20 phenotypes.
Here we report that only the overexpression of SFB2 suppresses defective phenotypes of both sec24-20 and ⌬sec24 mutants. Moreover, Sfb2p has an ability to bind Sec23p, suggesting that Sfb2p may participate in both COPII-coated vesicle formation and cargo selection events in concert with other constituents of COPII-coated vesicles. We also report the difference between Sec24p and Sfb2p in their cargo specificity. Studies on Sfb2p (this study) and Lst1p (24) suggest that the ER-to-Golgi protein transport is mediated by vesicles with heterogeneous COPII coats.
Strains, Media, and Growth Conditions-Yeast strains were grown in rich (YPD) or synthetic dextrose (SD) media, and standard genetic manipulations were performed as described previously (32). For some experiments, cells were grown in synthetic complete (SC) media (32) buffered to pH 7.0 with 0.1 M potassium phosphate buffer or adjusted to pH 4.0 with HCl. The Escherichia coli strain DH5␣ (33) was used for manipulation of recombinant DNA and was grown in LB media (1% NaCl, 1% peptone, and 0.5% yeast extract) containing 100 g/ml ampicillin. All yeast strains used in this study were listed in Table I, and yeast strains except for PJ69-4A were derived from FY strains (generous gifts from Dr. F. Winston). Diploid strain DFY24 was sporulated to obtain a KFY1 strain. To obtain YKH1 (⌬sfb2) and YKH2 (⌬sfb3) strains from TKO1, gene deletions were performed by PCR amplification of the URA3 gene with oligonucleotides that encoded 45 base pairs of the gene-specific sequences near each end of the open reading frames according to the method of Baudin et al. (34). The diploid strain DFY24 replaced one of its SEC24 genes with the LEU2 gene, as described above, was transformed with pAN1 (SEC24, URA3, CEN), and then sporulated to obtain a YKH3 (⌬sec24) strain. Strains YKH4 and YKH6 were obtained from the experiment represented in Fig. 4 (see figure legend of Fig. 4).
Electron and Immunofluorescence Microscopy-Electron microscopic analysis was performed as follows: preparation of a thin section of yeast cells was carried out by the freeze-substituted fixation method as described previously (35), except that Reichart KF80 was used to freeze the cells. Thin sections were viewed on a JEOL100CX electron microscope (JEOL, Tokyo, Japan) at 80 kV. For Kar2p staining, cells were fixed by direct addition of formaldehyde to the culture to a final concentration of 3.7% and incubated at room temperature for 2 h. Fixed cells were washed with potassium phosphate buffer (0.1 M potassium phosphate, pH 6.5) and converted into spheroplasts by incubation with 50 g/ml zymolyase-100T (Seikagaku Corp., Tokyo, Japan) in 1.2 M sorbitol, 0.1 M potassium phosphate, pH 6.5, and 0.2% ␤-mercaptoethanol at 30°C for 1-2 h. Immunofluorescence was performed essentially as described previously (36). The rabbit anti-yeast Kar2p antiserum prepared as described previously (37) was used as the primary antibody and rhodamine-conjugated goat anti-rabbit IgG antibody (Cappel Research Products, ICN, Inc., Tokyo, Japan) was used as the secondary antibody. DNA was stained with 1 g/ml 4,6-diamino-2-phenylindole for 3 min at room temperature before mounting. Preparations were viewed on a Axiophot fluorescence microscope (Zeiss, Jena, Germany).
␤-Galactosidase Assay-Assays of ␤-galactosidase activity in yeast extracts were carried out as described previously (38). ␤-Galactosidase activity was expressed as units defined as (A 420 ϫ 1000)/(A 600 ϫ t ϫ v), where A 420 is the absorbance at 420 nm of the reaction mixture, after t minutes, A 600 the turbidity of the culture at the time of harvest, t the number of minutes for which the reaction mixture was incubated, and v the volume of the sample in milliliters. The values are averages of four independent yeast transformants, and error bars are presented (Figs. 2C and 5E).
Pulse-Chase Experiments-Pulse-chase experiments were performed as described previously (39) with the following modifications. Cells were grown exponentially in SC, pH 4.0, lacking methionine and cysteine. For invertase transport assay, cells were transferred to SC, pH 4.0, media containing 0.1% glucose and incubated for 30 min to induce the expression of invertase before pulse labeling. Immunoprecipitations were performed using anti-carboxypeptidase Y (CPY) antibody (Rockland Immunochemicals, Inc., Gilbertsville, PA), anti-invertase antibody (kindly provided by Dr. K. Mihara, Kyushu University, Japan), and anti-Gas1p antibody (kindly provided by Dr. H. Riezman, University of Basel, Switzerland) at 1:500 dilution. Proteins were dissolved in Laemmli sample buffer containing 0.1 M dithiothreitol, incubated at 95°C for 3 min, and then subjected to SDS-PAGE (8%). Assays for general secretion competence were performed as described (40) with the following modifications. Before labeling, cells were grown exponentially in SC, pH 4.0 or 7.0, lacking methionine and cysteine. Cells were pulse-labeled with 4 MBq/ml [ 35 S]methionine ϩ cysteine, at 5 A 600 /ml in SC, pH 4.0 or 7.0, media containing 100 g/ml ␣ 2 -macroglobulin and 250 g/ml bovine serum albumin, and then chased for 45 min by adding excess amounts of unlabeled methionine and cysteine (final concentrations of 1 and 0.8 mM, respectively). Proteins equivalent to 0.25 A 600 were loaded on SDS-PAGE (8%), and autoradiographed with a BioImage BAS2000 analyzer (Fuji Photo Film, Tokyo, Japan).
Yeast Two-hybrid System-The interaction between Sec23p and Sfb2p, Sec24p, or sec24-20 gene product was tested by the yeast twohybrid method as described (31). Open reading frames of SFB2, SEC24, anu1, and SEC23 were cloned into vectors pGAD-C1 and pGBD-C1 in-frame with the GAL4 activation or binding domain as described under "Plasmid Construction." RNA Preparation and Northern Analysis-Cells were grown exponentially in SC, pH 4.0, medium and then harvested by centrifugation, washed, frozen by liquid nitrogen and stored at Ϫ80°C. Total RNA of cells was isolated by hot phenol extraction (41). RNA was quantified by absorbance at 260 nm and the integrity of the RNA was confirmed by ethidium bromide staining of RNA in agarose gels. For Northern blot-  2. sec24-20 cells show temperature-sensitive growth and ER chaperone-related defects. A, 10-fold serial dilution (starting from A 600 of 0.01) of wild-type and sec24-20 cells containing 2 -based plasmids, pSF1 (SFB2), pSF2 (SFB3), or control vector (pRS426), were spotted on SC, pH 4.0, plates and incubated at 23°C for 3 days, or at 30 and 37°C for 2 days. B, wild-type and sec24-20 cells were grown exponentially at 23°C in SD followed by 3 h incubation at the indicated temperatures. Immunofluorescent microscopic analysis was performed with anti-Kar2p antibody. C, ␤-galactosidase activities of wild-type, sec24-20, and sec18 (TKA33; anu4) cells containing a UPRE-CYC1 promoter-LacZ reporter plasmid (pCZY1). The growth conditions were as in B.
ting, 5 or 10 g of each RNA sample was subjected to electrophoresis in 1.0% formaldehyde/agarose gel, followed by transferring onto a nylon membrane (Hybond N ϩ , Amersham Pharmacia Biotech) and hybridization analysis as described (42). Probes used for hybridization were the open reading frames of SEC24 and SFB2, obtained by PCR amplification of chromosomal DNA using the primer sets TH24-5 and TH24-3 and TH2-5 and TH2-3, respectively. Probes were labeled by Random primer labeling kit (TaKaRa, Kyoto, Japan). The probed membranes were autoradiographed and analyzed with a BioImage BAS2000 analyzer.

RESULTS
Accumulation and Severe Extention of the ER in anu1-1/ sec24-20 Mutants-In our previous study on nuclear morphology, we obtained four mutants, anu1-anu4, exhibiting abnormal nuclear morphology at the restrictive temperature (25). Unexpectedly, ANU1 was identical to SEC24, which is known to be involved in the early secretory pathway rather than the construction of nuclei (25). Electron microscopic analysis revealed that anu1-1/sec24-20 cells cultured at the restrictive temperature of 37°C for 2 h contained morphologically aberrant and apparently fragmented nuclei (Fig. 1B, D, and E). Accumulated membranes having appearances like an extension of the ER because of their continuity with the outer nuclear membrane were also observed in sec24-20 cells (Fig. 1, B and D). In addition, irregular invagination of the ER was often observed in sec24-20 cells cultured at the restrictive temperature (Fig. 1C). These results suggested that some abnormalities of the ER led to the aberrant nuclear morphology in sec24-20 cells.
Abnormal Phenotype in the ER at the Restrictive Temperature-sec24-20 cells barely grew at 30°C, and died at 37°C (Fig. 2A). To identify some other abnormalities related to the ER in sec24-20 cells, we investigated the distribution of Kar2p/ BiP, a well characterized ER resident chaperone, by indirect immunofluorescence microscopic analysis. Contrary to wildtype cells, punctate staining of Kar2p was observed in 90 -95, 60 -70, and Ͻ5% of sec24-20 cells incubated at 37, 30, and 23°C, respectively (Fig. 2B). The Kar2p distributions in sec24-20 cells resembled those termed BiP bodies (43) observed in various mutants defective in the transport between ER and Golgi. Since the unfolded protein response (UPR) activates transcription of ER resident chaperones in response to the accumulation of un-or mal-folded proteins in the lumen of the ER accompanied by the blockade of ER-to-Golgi vesicle transport (44 -48), we next investigated whether the UPR was activated in sec24-20 cells. To assess the activation of the UPR, ␤-galactosidase activity was determined in wild-type, sec24-20, and sec18 (anu4) cells containing a UPRE (UPR element of KAR2)-CYC1 promoter-LacZ reporter plasmid (pCZY1). LacZ expression from pCZY1 was under the control of UPRE, the 22-base pair cis-acting element necessary and sufficient for transcriptional induction of Kar2p by the UPR (30,38). At the semipermissive temperature of 30°C, but not at 23°C, ␤-galactosidase activity of sec24-20 and sec18 cells was 3-4-fold higher than that of wild-type cells, indicating that the UPR was activated in these two mutants (Fig. 2C).
To further confirm secretory defects in sec24-20 cells, we first compared the general profiles of protein secreted in the media of sec24-20 cells with that of wild-type cells by a pulse-chase experiment. In contrast to wild-type cells, a temperature-sensitive secretion was observed in sec24-20 cells (Fig. 3A). In sec24-20 cells, some proteins appeared to be secreted in more reduced kinetics than other proteins at 30°C (Fig. 3A, lane 6,  arrows), whereas the overall secretion was almost blocked at 37°C. We further investigated whether sec24-20 cells had ERto-Golgi vesicle transport defects by a transport assay of CPY. CPY is detected in the ER as the p1 form (67 kDa), further modified to the p2 form (69 kDa) in the Golgi, and then processed proteolytically to the mature form (m; 61 kDa) in the vacuole (49). Unlike wild-type cells, sec24-20 cells exhibited a temperature-sensitive maturation of CPY to accumulate the p1 form even after the 60-min chase at 37°C (Fig. 3B). Thus, sec24-20 cells are defective in ER-to-Golgi vesicle transport at the restrictive temperature causing the activation of the UPR.
Suppression of sec24-20 Phenotypes by the Overexpression of a SEC24 Homologue, SFB2-The Saccharomyces genome data base contains two genes related to SEC24, SFB2 (56% identity) and SFB3 (23% identity) (22). We therefore investigated whether the overexpression of either gene suppressed growth and/or vesicular transport defects of sec24-20 mutant. As shown in Fig. 2A, the overexpression of SFB2 suppressed the growth retardation of sec24-20 cells at 30°C, but not the growth inhibition at 37°C. On the contrary, the overexpression of SFB3 did not affect the growth of sec24-20 cells (Fig. 2A). In addition, at 30°C but not at 37°C, the overexpression of SFB2 also restored the secretion competence of sec24-20 cells to the same extent as that of sec24-20 cells complemented with a SEC24 plasmid (pAN1; URA3, CEN) (Fig. 3A). Consistent with this observation, the overexpression of SFB2 similarly improved the maturation rate of CPY in sec24-20 cells (Fig. 3B). These results indicated that both growth and ER-to-Golgi vesicle transport defects of sec24-20 cells were partially suppressed by the overexpression of SFB2.
Two-hybrid Interaction between SFB2 and SEC23-Sec24p was first identified as a complex with Sec23p (55). Because of the similarity to Sec24p, Sfb2p is supposed to bind Sec23p and act as a component of COPII coat. To investigate whether Sfb2p interacts with Sec23p by the yeast two-hybrid assay, SEC23 and each of SFB2, SEC24, and sec24 -20 were fused to the GAL4-DNA binding domain (pGBD-SEC23) and to the GAL4activation domain (pGAD-SFB2, pGAD-SEC24, pGAD-sec24-20), respectively. Each pGAD-plasmid was co-transformed with pGBD-SEC23 into host strain PJ69-4A, and the resultant clones were replica-plated onto SD plate lacking histidine and adenine (Fig. 6) or subjected to ␤-galactosidase assay (Table II). When the fusion proteins interact in host cells, reporter genes HIS3, ADE2, and LacZ are activated in GAL-promoter dependent manner, resulting in the disappearance of both histidine and adenine auxotrophies and in the elevated levels of ␤-galactosidase. As shown in Fig. 6, cells coexpressing the SFB2and SEC23-fusions grew on SD (-His, -Ade) as well as those coexpressing the SEC24-and SEC23-fusions or the sec24-20and SEC23-fusions. Consistent with this observation, ␤-galactosidase activity of these cells was considerably higher than that of others (Table II). These results suggested that Sfb2p and sec24-20 gene product could bind to Sec23p as well as Sec24p.
Effect of sfb2 Disruption on Growth or Protein Transport-Another SEC24 homologue, SFB3, was identified as LST1 (lethal with sec-thirteen). LST1 disrupted cells (⌬lst1) were defective in export of plasma-membrane proton-ATPase (Pma1p) from the ER, and thereby their growth was sensitive to an acidic environment at high culturing temperature (24). We knocked out two SEC24 homologues, SFB2 and SFB3, to eval-uate the growth of the resultant null mutants. Unlike ⌬sfb3 (YKH2) cells, ⌬sfb2 (YKH1) cells could grow as well as the wild-type cells at any culturing temperature or pH of media (Fig. 7A). When the general secretion competence of ⌬sfb2 cells was examined, all of the secretion profiles at 37°C (Fig. 7B,  lanes 9 -12) were dramatically different from those presented in Fig. 3A because of culturing cells in SC, pH 7.0, a neutral medium suitable for ⌬sfb3 cells. The secretion profile of ⌬sfb2 cells, however, was indistinguishable from that of wild-type or ⌬sfb3 cells at any culturing temperature (Fig. 7B). This was consistent with the observation (23) that chromosomal deletion of SEC24 homologues did not affect the transport of either CPY or invertase. Finally, to obtain the information about the expression of SFB2, Northern blot analyses were carried out with total RNA isolated from wild-type cells. To examine the transcriptional level of SFB2 in comparison to that of SEC24, the RNA blots were hybridized with SFB2 and SEC24 probes of comparable specific activity and were exposed for the same period of time. As shown in Fig. 7C, the transcription level of SFB2 is much lower (approximately 1/8, at any culturing temperature) than that of SEC24, consistent with the fact that chromosomal deletion of SEC24 is lethal (note the expression level of SFB2 that compensates the chromosomal deletion of SEC24).

DISCUSSION
In this study, we describe some phenotypes of a temperaturesensitive mutant anu1-1, previously screened for abnormal nuclear morphology (25), and the partial suppression of anu1 phenotypes by overexpression of the SEC24 homologue, SFB2. At the restrictive temperature, anu1 cells exhibited the proliferated ER and the punctate distribution of Kar2p. Consistent with the fact that ANU1 is identical to SEC24 (25), anu1-1/ sec24-20 cells exhibited the temperature-sensitive ER-to-Golgi vesicle transport. Both temperature-sensitive growth and transport were partially restored by overexpression of SFB2. Moreover, the overexpression of SFB2 suppressed the lethality derived from the chromosomal deletion of SEC24, depending on culturing conditions. The resultant ⌬sec24(SFB2,2) cells exhibited some cargo-specific ER to Golgi transport defects, while their secretion competence appeared to be recovered to the same extent as that of wild-type cells. A yeast two-hybrid assay revealed that like Sec24p, Sfb2p could bind Sec23p.
The abnormal nuclear morphology of sec24-20 cells (Fig. 1) has also been observed in various mutants defective in vesicular transport between the ER and the Golgi such as sec28/ anu2, sec13/anu3, sec18/anu4, sar1, sec12, sec16, and uso1 (25,56). 2 Transport defects between the ER and the Golgi are considered to trigger a proliferation of ER, perturbing the nuclear structure because of the continuity of the nucleus with the ER (25). Consistent with the retardation of the ER to Golgi protein transport (Fig. 3), the BiP body-like distribution of Kar2p and the activation of the UPR were also observed in sec24-20 cells at 30°C (Fig. 2, B and C). Since the UPR was also activated in sec18 cells (Fig. 2C), the activation of the UPR in sec24-20 cells is considered to represent the accumulation of proteins in the ER owing to the reduced ER-to-Golgi vesicle transport. On the other hand, the punctate distribution of Kar2p in sec24-20 cells is probably identical to the BiP body, the site where secretory proteins accumulate when ER to Golgi vesicle transport is blocked (43). It is thought that BiP may escort secretory proteins to keep them transport-competent until the transport between the ER and Golgi is resumed (43).
The precise role of Sec24p itself in vesicle formation and cargo selection yet remains unclear. The suppression of the ⌬sec24 mutation by overexpression of SFB2 (Figs. 4 and 5) in addition to the two-hybrid interaction between Sfb2p with Sec23p ( Fig. 6 and Table II), strongly suggest that Sfb2p can replace Sec24p in the COPII-coated vesicle formation and in the transport of most proteins. This is supported by the difference in the growth rate between ⌬sec24(SFB2,2) and sec24-20(SFB2,2) cells. sec24-20(SFB2,2) cells grew slower than ⌬sec24(SFB2,2) cells at 30°C, but faster at 23°C (see "Results"). Moreover, the overexpression of SFB2 could rescue the growth defect of ⌬sec24 cells but not that of sec24-20 cells at 37°C ( Figs. 2A and 4B). Like Sec24p, Sfb2p and the sec24-20 gene product can bind Sec23p (Fig. 6 and Table II). Taken together, these results suggest that Sfb2p competes with the functionally impaired sec24-20 gene product for the binding to Sec23p in sec24-20(SFB2,2) cells at high culturing temperature.
The secretion assay in ⌬sec24(SFB2,2) cells revealed that most of the protein transport appeared to be recovered to the same extent as that of ⌬sec24(SEC24,CEN) cells (Fig. 5A). This implies that vesicle machinery proteins such as v-SNAREs, are packaged or at least included in the ER-derived vesicles of ⌬sec24(SFB2,2) cells. On the other hand, transport assays of several cargo proteins revealed that not all of the secretory proteins were efficiently exported from the ER in ⌬sec24(SFB2,2) cells (Fig. 5, B-D). In particular, ⌬sec24(SFB2,2) cells exhibited markedly slow transport kinetics of invertase and Gas1p at any culturing temperature (Fig. 5, C and D), suggesting that Sec24p, rather than Sfb2p, is indirectly responsible for concentrating them into ER-derived vesicles. It is known that the efficient ER to Golgi transport of invertase and Gas1p requires p24 proteins, the putative cargo receptor expected to serve as an adaptor linking the vesicle-forming machinery to soluble cargo recruitment (9,10,(12)(13)(14). In the study using ER-derived microsomes and purified COPII components, Emp24p, a member of the p24 family, was found to associate with the Sar1p-Sec23p/Sec24p prebudding complex, suggesting that the prebudding complex is responsible for sorting Emp24p into COPIIcoated vesicles (17). According to this observation, p24 proteins appear to interact efficiently with the Sar1p-Sec23p/Sec24p rather than the Sar1p-Sec23p/Sfb2p complex, emerging the difference between ⌬sec24(SEC24,CEN) and ⌬sec24(SFB2,2) cells in the kinetics of invertase and Gas1p transport. Alternatively, a reduced packaging of machinery proteins required for the Golgi to ER retrograde vesicle transport (such as Golgi to ER v-SNAREs) into ER-derived vesicles of ⌬sec24(SFB2,2) cells may retard the retrieval of p24 proteins from the Golgi to the ER, resulting in the reduced kinetics of invertase and Gas1p transport. On the other hand, the cold-sensitive transport of CPY in ⌬sec24(SFB2,2) cells (Fig. 5B) suggests that COPII-like-coated vesicles including Sfb2p are cold-sensitive in the selection and/or the concentration of CPY. Since the UPR was activated in ⌬sec24(SFB2,2) cells at both 23 and 30°C (Fig. 5E), some other cargo proteins such as invertase and Gas1p might be retained in the ER independently of culturing temperature. The conditional growth of ⌬sec24(SFB2,2) cells may be due to the temperature-dependent and/or -independent ER retention of proteins affecting the cell growth, such as those required for the growth under neutral conditions. 2 A. Hirata, unpublished observations. FIG. 6. Two-hybrid interaction between SFB2 and SEC23. Either pGBD-SEC23 or control plasmid (pGBD-C1) was co-transformed with pGAD-SFB2, pGAD-SEC24, pGAD-sec24-20, or control plasmid (pGAD-C1), into host strain PJ69-4A. The resulting clones containing each plasmid combinations were replica-plated onto SD (ϪHis, ϪAde) and control SD (ϩHis, ϩAde) plate, and incubated at 30°C for 4 days. The fact that some cargo proteins (such as invertase and Gas1p) require Sec24p but not Sfb2p for their efficient export from the ER suggests that Sec24p itself is responsible for the cargo recognition. Another Sec24p homologue, Lst1p, is involved in the export of restricted cargo proteins (only Pma1p is known) from the ER (24). Based on ⌬sfb2 phenotypes and the transcriptional level of SFB2 in wild-type cells (Fig. 7), Sfb2p may be involved in the efficient transport of specific cargo molecules like Lst1p rather than the efficient transport of many cargo proteins required for their growth like Sec24p. Hence, each of the SEC24 family of proteins may function as a component of COPII coats and serve its own cargo selectivity.
We could not find the unique role of Sfb2p in the cargo selection, but demonstrated that in the ER-to-Golgi vesicle transport Sfb2p could function as the component of COPII coats. The identification of Sfb2p (this study) and Lst1p (24) as the component of COPII coats suggests that the coats of ERderived vesicles may be heterogeneous. We are interested in whether one ER-derived vesicle contains one member of the SEC24 family of proteins, or alternatively plural members of the family act together as the components of coats. If all members of the SEC24 family of proteins participate in the cargo selection event in concert with Sec23p, Sar1p, and others, the composition of the SEC24 family of proteins in the cell may affect that of cargo proteins transported by ER-derived vesicles. Further experiments are required to elucidate the roles of Sec24p and Sfb2p in cargo selection.