The beta subunit of the Sec61p endoplasmic reticulum translocon interacts with the exocyst complex in Saccharomyces cerevisiae.

The exocyst is a conserved protein complex proposed to mediate vesicle tethering at the plasma membrane. Previously, we identified SEB1/SBH1, encoding the beta subunit of the Sec61p ER translocation complex, as a multicopy suppressor of the sec15-1 mutant, defective for one subunit of the exocyst complex. Here we show the functional and physical interaction between components of endoplasmic reticulum translocon and the exocytosis machinery. We show that overexpression of SEB1 suppresses the growth defect in all exocyst sec mutants. In addition, overexpression of SEC61 or SSS1 encoding the other two components of the Sec61p complex suppressed the growth defects of several exocyst mutants. Seb1p was coimmunoprecipitated from yeast cell lysates with Sec15p and Sec8p, components of the exocyst complex, and with Sec4p, a secretory vesicle associated Rab GTPase that binds to Sec15p and is essential for exocytosis. The interaction between Seb1p and Sec15p was abolished in sec15-1 mutant and was restored upon SEB1 overexpression. Furthermore, in wild type cells overexpression of SEB1 as well as SEC4 resulted in increased production of secreted proteins. These findings propose a novel functional and physical link between the endoplasmic reticulum translocation complex and the exocyst.

The targeting and fusion of transport vesicles with plasma membrane is mediated by a molecular machinery highly conserved in evolution. Prior to the fusion at the plasma membrane the secretory vesicles are recognized by the exocyst complex, which is proposed to function as a tethering factor for the vesicles at plasma membrane in Saccharomyces cerevisiae. This protein complex, composed of Sec3p, Sec5p, Sec6p, Sec8p, Sec10p, Sec15p, Exo70p, and Exo84p (1,2), has a central role in establishing cell polarity in yeast (3,4). The mammalian homologue of the exocyst, the octameric Sec6-Sec8 complex, is essential for epithelial cell polarization and for synapse formation (5,6). The interaction of small GTPases of the Ras and Rho family proteins, Sec4p, Cdc42p, Rho1p, and Rho3p in yeast (7)(8)(9)(10)(11) and RalA in mammalian cells (12)(13)(14), with the exocyst complex suggests that the function of the exocyst is highly regulated (15). The exocyst-like tethering function may not be restricted to the exocytosis process as a homologous octameric Sec34-Sec35 complex was recently characterized in the ER 1 -Golgi interface membranes in yeast (16,17). Based on sequence homology an additional exocyst-like complex, the Vps53-Vps54-Vps55p complex, might be functional at the yeast Golgivacuole interface (16,18). The precise function of these exocystlike complexes is presently unclear. Mammalian equivalents for these novel complexes have not yet been identified. However, homologues of their subunits are found in mammalian cells. The finding that Exo84p protein is functional in mRNA splicing in yeast cells (19) suggests that exocyst complex or its individual subcomponents may also display functions at least apparently unrelated to secretion.
In a genetic screen we previously identified SEB1/SBH1 as a multicopy suppressor of the sec15-1 exocyst mutant (20). SEB1 (SBH1) and its homologue SEB2 (SBH2) encode the ␤ subunits of the trimeric Sec61p and Ssh1p ER translocation complexes, respectively (20 -22). The ␣ subunits of the Sec61p and the Ssh1p complexes are Sec61p and Ssh1p, respectively, which have been shown to form the protein conducting channel (23) and the ribosome binding site at the ER (24). The ␥ subunit in both complexes is Sss1p (22,25). These complexes have been shown to function in cotranslational translocation and the Sec61p complex also in the posttranslational translocation (21,22,26). The exact functions of the ␤ subunits of the ER translocation complexes are not known.
We have now extended our previous finding on the genetic link between SEB1 and sec15-1 mutation and provide biochemical and functional evidence for protein-protein interactions between the ER translocation complex ␤ subunit and components of the exocyst complex and Sec4p.

EXPERIMENTAL PROCEDURES
Strains, Media, and Culture Conditions-The yeast strains are presented in Table I. Disruptions of SEB1 and SEB2 genes in NY179 were done as described (20). To obtain strain H1982, the double seb disruptant strain, H1239, was converted to Ura Ϫ by transformation with a SacI-NsiI fragment from pBUF and selecting for growth in the presence of 5-fluoroorotic acid. The strain NY1427 (1) was similarly converted to Ura Ϫ (H2647) by disrupting the URA3 with SpeI-XhoI fragment from pJL164 (27). To obtain H1256 the DBY746 strain was transformed with the integrating plasmid YIplac204␣a linearized by BstXI digestion within the TRP1 and selected for Trp prototrophy. Yeast cells were grown in either YPD (yeast extract peptone dextrose) or SCD (synthetic complete dextrose) (28). For plasmid selection, SCD lacking leucine, tryptophan, or uracil was used. For temperature shift-up experiments, the transformants were grown in selective liquid medium to the early logarithmic growth phase. The cultures were divided in half, and cells were pelleted and resuspended into fresh growth medium. One culture was incubated at 24°C and the other at the restrictive temperature.
Plasmids and DNA Methods-Plasmids YEpHA-SEB1 and pBSEB1 have been described (20). The SEB1 cDNA covering the open reading frame was transferred from pBSEB1 under TPI1 promoter in pYX212 multicopy vector (R&D Systems) as an EcoRI-XhoI fragment yielding plasmid YEpT-SEB1U. YEpSEB2U is pRS426 (29) containing the SEB2 expression cassette as a BamHI-HindIII fragment from YEpSEB2 (20). YEpSEC1aU overexpressing SEC1 and YEp24H (30) as well as YEpSSO1U and YEpSSO2U overexpressing SSO1 and SSO2, respectively, have been described (31). For disruption of the URA3 gene, pBUF containing a 331-base pair (bp) deletion within the 1135-bp-long URA3 coding region was constructed as follows. Two URA3 fragments were amplified by PCR and cloned into the multiple cloning site of pBluescript SK(-). The first fragment comprising bp 71-450 of the URA3 was cloned as a SacI-XbaI fragment and the second (bp 781-1135) as a XhoI-KpnI fragment. pBW65 containing the SEC61 gene in YEp351 vector (32) and FKp53 (33) containing the SSS1 gene in YEp352 were obtained from Colin Stirling. The control plasmids for pBW65 and FKp53 were pRS425 (29) and YEp24H (30), respectively, and for YEpHA-SEB1 was pVT102U (34). To generate a His 6 -tagged Sec4p, the coding region was amplified by PCR with BamHI-XhoI sites added and cloned into a version of pGAT (provided by Johan Perä nen) from which the GST coding region was removed by SpeI digestion and religation. This resulted in pGATHis-Sec4p. To generate a His 6 -tagged version of the first 241 amino acids of Sec15p, the coding region for these amino acids was amplified by PCR and cloned as a BamHI-EcoRI fragment to pRSETA (Invitrogen) to create pSET-His-Sec15 (amino acids 1-241). The fragment encoding Seb1p without transmembrane domain (amino acids 1-54) was created by PCR with oligonucleotide primers containing XbaI and XhoI sites and ligated into XbaI-XhoI cut pVT102U (34) yielding YEpSEB1-MA. The plasmid YIplac204␣a for ␣-amylase gene integration was obtained by inserting the ␣-amylase expression cassette (Bacillus amyloliquefaciens ␣-amylase gene between ADH1 promoter and terminator) from YEp␣a6 (35) as BamHI-SalI fragment into the multiple cloning site of YIplac204 (36). SEC3 in YEp352 (37) was obtained from Brian Haarer, YEpSEC4 (NRB524) and pNB148 overexpressing SEC4 (38) and SEC15 (39), respectively, from Peter Novick, and pRD6 and pDF11 overexpressing SEC62 (40), and SEC63 (41), respectively, from Randy Schekman. Standard DNA methods were used for cloning. The yeast cells were transformed as described (20).
Multicopy Suppression-The mutant strains were transformed with the multicopy plasmid overexpressing the gene under study or with the pertinent vector control. The multicopy suppression was first screened by replica plating freshly grown patches on selective plates and on YPD and growing them at different temperatures. Growth was monitored for 3 successive days. The suppression test was then done by dotting 10 5 cells and three 10-fold dilutions thereof on selective plates and incubating the plates at different temperatures for 3 days. The patching method was slightly more sensitive and rescued the growth at a slightly higher temperature or detected in some cases weak suppression that was not detected by dotting.
Determination of ␣-Amylase and Invertase Activity-Yeast cell growth, preparation of the periplasmic and cytoplasmic fractions, and determination of the ␣-amylase activity were done as described previously (42). Invertase activity was determined as described (42) except that the reaction was terminated with addition of 1 M Tris, pH 8.5, after which the cells were pelleted instead of filtering and the glucose liberated from sucrose by invertase was measured from the supernatant.
Antibodies and Protein Methods-Antibodies against an N-terminal peptide of Seb1p have been described (20). The antibodies against Sec4p and Sec15p were raised in rabbits injected with bacterially (Escherichia coli BL-21de3) expressed and purified His 6 -Sec4p and His 6 -Sec15p (amino acids 1-241). HA (12CA5) and Myc tag antibodies (9E10) were purchased from Roche Applied Science and Santa Cruz Biotechnology, respectively. For detection of Sec4p in Western blots, an IgG heavy chain-specific secondary antibody (Jackson ImmunoResearch Laboratories Inc.) was used to reduce signals from closely migrating IgG light chain specific bands. Other antibodies were generous gifts from Howard Riezman (antibodies against Gas1p) and Romano Serrano (anti-Pma1p antibodies). The proteins were separated in SDS-PAGE using the buffer system of either Laemmli (43) or Schä gger and von Jagow (44) with acrylamide concentrations from 7.5 to 15% as appropriate for the proteins analyzed and transferred onto Hybond-P filter (Amersham Biosciences). The protein bands were visualized with the ECL detection system (Amersham Biosciences).
Preparation of Yeast Cell Lysates and Immunoprecipitations-Transformants of strains H2647 and NY15 were grown overnight in SCD-Ura to early logarithmic growth phase of A 600 ϭ 1 and pelleted. The cells were washed with 10 mM NaN 3 and resuspended in spheroplasting  (45). Seb1p was coprecipitated with ␣Sec15p antibodies with equal efficiency from cell lysates prepared either in homogenization buffer or buffer B. The cell homogenates were centrifuged at 10,000 ϫ g for 10 min at 4°C, and the supernatant was used for immunoprecipitations. 1.1 mg of total protein of cell lysates in homogenization buffer or buffer B was precleared with protein G-Sepharose (Amersham Biosciences) for 30 min at 4°C and centrifuged at 20,000 ϫ g for 15 min. Precleared lysate containing 1 mg of total protein was subjected to overnight immunoprecipitation with 3 l of antiserum. Immunoprecipitations were centrifuged (17,000 ϫ g for 1 min); the supernatant was transferred to new tubes, and 20 l of 50% protein G-Sepharose in the same buffer was added. The immunocomplexes were bound to the beads in an end-over-end mixer for 2 h at room temperature. The beads were washed five times with 1 ml of the lysis buffer and once with 1 ml of 10 mM Tris, pH 6.8. The Complete Protease inhibition mixture (Roche Applied Science) and 4 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (Roche Applied Science) were present during homogenization and preparation of the lysates and immunoprecipitations but were omitted from the washes. The immunocomplexes were released into 20 l of Laemmli sample buffer, heated at 95°C for 5 min, and analyzed by SDS-PAGE and Western blotting. For detection in Western blots, anti-Seb1p was used in 1:2000 to 1:8000 dilutions. Sec8p-Myc was detected with c-Mycspecific antibody (1:500). Anti-Sec4p was used in a 1:80000 dilution and IgG heavy chain-specific antibody in a 1:100000 dilution. Pma1p was detected with anti-Pma1p diluted 1:100000 and Gas1p with anti-Gas1p diluted 1:10000. The bound antibodies in Western blots were visualized with the ECL detection system (Amersham Biosciences). The quantitations were done from Kodak Biomax MR films using a GS-710 imaging densitometer and Quantity One software (Bio-Rad). The immunoprecipitates were quantitated by comparing the amount of a given protein precipitated to the amount of that protein in the cell lysate that was used for precipitation.

Overexpression of SEB1 Suppresses Temperature-sensitive (ts) Defects of Exocyst Mutants-SEB1
encoding the ␤ subunit of the Sec61p complex was isolated as a multicopy suppressor of the temperature sensitive sec15-1 mutation (20). As Sec15p is a component of the exocyst complex, it was of interest to see whether overexpression of SEB1 can suppress temperaturesensitive mutations in the other components of the complex. The exocyst sec mutant strains were transformed with a multicopy plasmid encoding SEB1 or with an empty vector, and their ability to grow at the nonpermissive temperature was tested. Growth of all the exocyst sec mutants was rescued at the restrictive temperature at least partially by SEB1 overexpression (Fig. 1). The growth of sec8-9 and sec15-1 was rescued near or at the level of growth at the permissive temperature, whereas partial rescue was obtained for the rest of the mutants. Possible suppression of the secretion defect of sec8-9 and sec15-1 cells at the restrictive temperature by SEB1 overexpression was also tested. Overexpression of SEB1 ameliorated defective secretion of invertase by these mutants but did not fully restore it to the level of wild type cells (data not shown).
Exocytosis Mutations Are Suppressed by Overexpression of the Sec61p Complex Components but Not Vice Versa-As multiple genetic interactions were observed for SEB1 and the exocytosis machinery, we studied whether overexpression of the other ER translocation components could suppress mutations in the late-acting SEC genes. The exocyst mutant strains were transformed with plasmids harboring SEC61, SSS1, SEC62, SEC63, or SEB2 or with their vector controls, and the growth of the transformants was tested at different temperatures. SEC61 rescued the growth of sec3-2 efficiently and the growth of sec5-24 and sec6-4 partially ( Fig. 2A). Overexpression of SSS1 rescued the growth of sec8-9 and sec15-1 efficiently and growth of sec3-2 and sec5-24 partially (Fig. 2B). In contrast, overexpression of SEC62, SEC63, or SEB2 did not suppress the ts defects in any of the exocytosis mutants tested (data not shown).
Suppression of the other late-acting secretory mutants (sec1-1, sec2-41, sec4-8, sec9-4, and sso2-1), mutants of the general components (sec17-1, sec18-1, and sec19-1), as well as mutants of SEC1 homologues (sly1-1 and slp1⌬) were also analyzed in this screen. Weak suppression by SEB1 overexpression was observed in sec2-41 and in sec19-1 mutants (data not shown). No interactions were detected with any of the other genes studied. The effect of SEB1 overexpression in kar2-133, kar2-159, and sec13-1 mutants was also tested, but their growth at the restrictive temperature was not rescued. Thus, the genetic interactions observed were most prominent between the genes encoding components of the two multiprotein complexes, the Sec61p translocation complex and the exocyst.
To test whether the observed genetic interactions function also in the opposite direction, the possible suppression of de- fects in the ER translocation components by overexpression of genes encoding components of the exocytosis machinery was studied. The sec61-1 and sec61-2 mutants are temperaturesensitive at 37°C (46,47), and the sec61-32 and sec61-41 mutants are cold-sensitive at 17°C (48). Their growth at these temperatures was not rescued by overexpression of SEC1, SEC3, SEC15, SSO1, or SSO2 gene (data not shown). The double deletion of both SEB1 and SEB2 results in impaired growth at 38°C (20,22). The growth of this double mutant was not rescued by overexpression of any of the late-acting secretion genes tested. Thus, the suppression is unidirectional; only defects in the exocytosis machinery are corrected by increased amounts of the ER translocon components and not vice versa.
The Seb1p Transmembrane Domain Is Required for sec15-1 Suppression-Seb1p is a membrane protein with a single Cterminal transmembrane domain (20). To test whether the suppression activity of Seb1p requires the transmembrane domain and thus tight membrane association, a mutant version of Seb1p lacking the hydrophobic C-terminal membrane anchor was created. The sec15-1 cells were transformed with the plasmid encoding the mutant, YEpSEB1-MA, the plasmid encoding the wild type SEB1, or the empty vector, and the growth of the transformants was monitored at different temperatures. The Seb1p-MA did not suppress the temperature sensitivity of sec15-1 even at 35°C, the lowest temperature restrictive for sec15-1 cells (Fig. 3A). The Seb1p-MA mutant did not affect the cell growth at the permissive temperature (24°C), indicating that its expression is not harmful (Fig. 3A). The lack of a phenotype for the Seb1p-MA overexpression could be because of instability of the mutant protein. To test this possibility, lysates were prepared from the transformants used in the suppression experiment, and the amounts of Seb1p and Seb1p-MA were studied by Western blotting with Seb1p-specific antibodies. Quantitation of the blots showed that the wild type Seb1p was ϳ14 times and the Seb1p-MA 10 times overexpressed in comparison to the endogenous Seb1p level (Fig.  3B). This indicates that Seb1p-MA is stably expressed in sec15-1 cells but fails to multicopy suppress this mutant.
Seb1p Coimmunoprecipitates with the Exocyst Components Sec8p and Sec15p in Yeast Cell Lysates-The genetic interactions between SEB1 and the exocytosis machinery prompted us to study protein-protein interactions between Seb1p and the exocyst proteins. For detection of Sec8p, a yeast strain, H2647, was used in which the only chromosomal copy of SEC8 gene encodes a protein tagged with three Myc epitopes at its C terminus (1). This strain was transformed either with a multicopy plasmid expressing HA-tagged SEB1 from the ADH1 promoter (20) or with an empty vector. In parallel, wild type strain NY15 was transformed with the same plasmids. The HA-tagged form of Seb1p is fully functional such that its overexpression suppresses the ts growth defects of sec15-1, sec10-2, and sec8-9 like the wild type gene (data not shown) and was utilized in these experiments to detect the overexpressed Seb1p. Antibodies against the N-terminal peptide of Seb1p coprecipitated the Myc-tagged Sec8p (Fig. 4, top panel). The Seb1p antibodies precipitated 1.6% of the Sec8-Myc protein present in the cell lysate used for precipitation. The coprecipitation of Sec8p-Myc was genuinely due to the presence of the Seb1p antibodies as evidenced by the absence of Sec8p-Myc in the mock precipitations carried out in the absence of the Seb1p antibodies (data not shown) or in the presence of the preimmune serum (Fig. 4, top panel, pis IP). In the Western blots the anti-Myc antibodies also detected a slightly faster migrating protein band, which may be a proteolytic cleavage product of the tagged Sec8p. The Seb1p antiserum immunoprecipitated both the HA-tagged and the wild type Seb1p (Fig. 4). The antibodies against Sec15p coprecipitated the endogenous Seb1p rather inefficiently (Fig. 4, middle panel). Sec15p antibodies brought down 0.6% of the Seb1p present in the input of the immunoprecipitation. More of the HA-tagged form of Seb1p was detected, as it was overexpressed in these transformants. However, the endogenous Seb1p also was precipitated more efficiently in the overexpressing cells, suggesting increased stability of the complex containing Seb1p and Sec15p (Fig. 4,  middle panel). In the Sec15p immunoprecipitations, the Myctagged Sec8p was precipitated from lysates of H2647 transformants as expected because both of these proteins are present in the exocyst complex.
Coimmunoprecipitation of Seb1p and Sec4p-Sec15p binds the small GTPase Sec4p (49). To study whether Sec4p is present in the complex containing Seb1p, Sec8p, and Sec15p, we performed immunoprecipitations with Sec4p-specific antibodies. H2647 or wild type cells transformed with a plasmid overexpressing HA-SEB1 or with the empty vector were lysed and subjected to immunoprecipitations with antibodies against Sec4p. Seb1p coimmunoprecipitated with Sec4p as did the Sec8p-Myc, indicating that the complex that is formed by the exocyst components and Seb1p also contains Sec4p (Fig. 4, bottom panel). The preimmune serum did not precipitate Seb1p, Sec8p-Myc, or Sec4p.
To rule out the possibility that Seb1 protein is brought down nonspecifically with immunocomplexes formed in the yeast cell lysates, two cell surface proteins not related to the exocytosis machinery were used as controls. Thus, immunoprecipitations with antibodies against Pma1p, an abundant plasma membrane protein (50), and Gas1p, a glycosylphosphatidylinositolanchored plasma membrane/cell wall protein (51), were performed. Neither one was able to coprecipitate Seb1p (data not shown), which further indicates that Seb1p immunoprecipitation together with the exocyst components and Sec4p is not because of nonspecific cross reactions.
Overexpression of SEB1 Recovers Seb1p-Sec15p Interaction in sec15-1 Cells-As SEB1 was isolated originally as a multicopy suppressor able to rescue the growth of sec15-1 cells at the restrictive temperature, we next tested whether the putative Seb1p-Sec15p complex is present in the sec15-1 mutant cells. The sec15-1 cells overexpressing SEB1 or those carrying the empty vector plasmid were grown to early logarithmic growth phase at 24°C, and the cultures were divided in half. One-half was shifted to 38°C for 30 min before cell lysis, while the other remained at 24°C. Immunoprecipitations were performed with polyclonal Sec15p antibodies. Interestingly, Seb1p could not be coprecipitated with Sec15p antibodies from the cells harboring the control plasmid even at the permissive temperature (24°C) (Fig. 5). From transformants overexpressing SEB1, efficient precipitation of Seb1p was obtained with the Sec15p antibodies. More importantly this precipitation was obtained also from cells that had been incubated at the restrictive temperature, 38°C. The Sec15p preimmune serum did not precipitate Seb1p from any of the transformants (Fig. 5). These results suggest that the Sec15-1 mutant protein is partially defective even at the permissive temperature, that an increased level of Seb1 protein restores the formation of the putative Seb1p-Sec15p complex with the mutant protein, and that this happens also at the restrictive temperature.
Overexpression of SEB1 or SEC4 Increases Production of Secreted Proteins in Wild Type Yeast-The above genetic inter-  action results suggest that SEB1 overexpression facilitates exocytosis. We have shown previously that overexpression of the SSO1 and SSO2 genes encoding plasma membrane t-SNAREs (target-soluble N-ethylmaleimide-sensitive factor attachment protein receptors), essential for exocytosis, increases the production of secreted proteins (42). We therefore proceeded in studying the effect of SEB1 overexpression on production of secreted proteins in wild type cells. Secretion of the yeast endogenous enzyme invertase was studied in DBY746 cells grown in 2% sucrose medium to allow expression of the glucose repressed SUC2 gene. Secretion of invertase to the cell surface was analyzed at several time points during growth in liquid medium, and reproducibly, up to 40% higher invertase activity was detected with SEB1 overexpression. This is the same level that is obtained with SSO2 overexpression from the endogenous SUC2 gene (42).
We also studied the effect of SEB1 overexpression on secretion of a heterologous reporter protein, Bacillus ␣-amylase (42). The ␣-amylase gene was integrated into the yeast genome at the TRP1 locus. This strain, H1256, was transformed with YEpHA-SEB1 or the empty vector, pVT102U. The transformants were grown in shake flasks in selective medium, and the amount of ␣-amylase secreted into the culture medium was determined. In comparison with the control transformant, up to 3-fold more ␣-amylase activity was found in the medium of SEB1 overexpression strain depending on the cultivation time point analyzed (Fig. 6A). The increased ␣-amylase activity in the medium was due to increased amount of the enzyme as determined by Western blotting (data not shown). To verify whether the increased production of ␣-amylase was due to increased transcription of the gene, the ␣-amylase mRNA levels were determined by Northern blotting. The ␣-amylase transcript level was increased slightly in the SEB1 overexpressing strain, being about 1.5-fold compared with that in the control strain. We next studied the effect of overexpression of SEC4 on ␣-amylase production. Sec4p is a small GTPase that is essential for regulation of exocytosis and binds to the post-Golgi secretory vesicles and Sec15p. Similar to SEB1, overexpression of SEC4 resulted in 3-fold increase in production of secreted ␣-amylase (Fig. 6B).
␣-Amylase is secreted efficiently by the wild type yeast. About 75% of the enzyme activity is found in the culture medium and in the periplasmic fraction/cell wall (42). Fractionation of the yeast cells to the periplasmic and cytoplasmic fractions was performed for SEB1-overexpressing and control cells to quantitate the total amount of exocytosed ␣-amylase. The enhancement of the production of secreted ␣-amylase was 2.3-fold (Table II). The overall ␣-amylase level in the SEB1 transformant was about 2-fold higher, and 90% of the enzyme was secreted instead of 75% of the enzyme present in the control strain.

DISCUSSION
In this report we present evidence for an intriguing connection between the synthesis and exocytosis of secreted proteins in yeast. This is based on the observed functional and physical interactions between components of the ER translocon and the exocyst complex functioning at the first and the last step of the secretory pathway, respectively. SEC15 gene encodes a component of the exocyst complex essential for targeting/fusion of the Golgi-derived secretory vesicles at the plasma membrane. In a screen for new genes functionally linked to exocytosis, we have isolated multicopy suppressors of the temperature sensitive sec15-1 mutation. One of the genes thus identified was SEB1, which encodes the ␤ subunit of the Sec61p ER translocation complex (20). This unexpected finding prompted us to study the genetic interactions in more depth and to search for potential physical interactions between the protein components functioning at these early and late steps of secretion.
Here we show that overexpression of SEB1 suppresses the ts mutations in all SEC genes encoding components of the exocyst complex. A functional link between the ER translocon and the exocyst complex is further supported by the increased temperature sensitivity detected in seb1⌬ seb2⌬ sec10-2 and in seb1⌬ seb2⌬ sec15-1 triple mutants. Although this synthetic effect was not strong, it may be important, because Sec10p and Sec15p form a subcomplex in vivo suggested to bridge the transport vesicle with the remaining exocyst components (49). Overexpression of the other two genes encoding components of the Sec61p complex, SEC61 and SSS1, also suppressed the ts defects in several of the exocyst mutants. The other late-acting mutations suppressed by SEB1 overexpression were sec2-41 and sec19-1 (data not shown). The Sec2p and Sec19p are regulators of the Sec4p, a Rab GTPase that interacts physically with Sec15p (49). These genetic interactions strongly suggest FIG. 6. Increased production of secreted ␣-amylase in SEB1 and SEC4 overexpressing cells. The H1256 strain carrying the ␣-amylase gene integrated in the TRP1 locus was transformed with multicopy plasmids expressing SEB1 or SEC4 or with the pertinent vector controls. The transformants were grown in SCD-Ura, pH 6.0, supplemented with 10 mM CaCl 2 . Cell growth (open symbols) and ␣-amylase activity present in the culture medium (filled symbols) were monitored as described under "Experimental Procedures." A, YEpHA-SEB1 transformant (circles) and pVT102U transformant (squares). B, YEpSEC4 transformant (triangles) and YEp24H transformant (diamonds). The mean values of two independent SEB1 and SEC4 transformants are shown. closer functional connection between the ER translocation complex and the exocytosis machinery than was surmised before.
The role of the nonessential ␤ subunits of the Sec61p and Ssh1p complexes in translocation is not known. Disruption of both SEB1 and SEB2 results in only a slight reduction in protein translocation and Gas1p maturation (22). 2 The mammalian Sec61␤ is not essential for translocation but has been shown to facilitate this process (52). In yeast, overexpression of SEB1 suppresses the sec61 mutations that destabilize the Sec61p structure (20,33,53). However, mutations that do not affect the stability of Sec61p, but render it nonfunctional in translocation, are not suppressed by SEB1 (54). This suggests an accessory or stabilizing role for the ␤ subunit in the complex.
The overexpression suppressions reported here show strict directionality; only the ER translocon components can rescue the function of the mutated exocyst components and not vice versa. Thus, the observed genetic interactions and the rescued physical interaction of Seb1p with Sec15-1 mutant protein in SEB1 overexpressing cells (Fig. 5) suggest that increased Seb1p level facilitates directly or indirectly the function of the destabilized (2,55,56) exocyst complex in the mutant cells. Coprecipitation of Seb1p with exocyst components Sec8p and Sec15p, and the Sec15p-interacting protein Sec4p, implies that these proteins can be present in the same protein complex within the yeast cell. Notably, these coprecipitations were detected also in cells expressing the endogenous level of Seb1 protein. Under the experimental conditions used here the proportion of Seb1p coimmunoprecipitating with Sec8p or Sec15p was low. The interactions were, however, specific as shown by the control experiments performed. The low coimmunoprecipitation efficiencies may be due in part to the fact that Seb1p may not need to exist in a stable complex with exocyst components to fulfill its biological function. We favor the view that the observed interaction reflects the existence of a regulatory signal between the two endpoints of the secretory pathway. Such interactions are likely to be transient by nature and do not necessarily require simultaneous association of a large proportion of the cellular Seb1p with exocyst subcomponents. We would also like to stress that the fact that Seb1p overexpression rescued the interaction with the mutant Sec15p and the cell growth indicates a direct or indirect functional interaction between these molecules. Additional support for the interaction is provided by the extensive genetic data and the observed secretion enhancement, which collectively point to an in vivo interaction of these processes.
The interactions between the ER translocon and the exocyst components are intriguing because functionally the Sec61p translocation complex and the exocyst are located at opposite ends of the secretory pathway. It has, however, been shown that the cortical ER membrane and the plasma membrane frequently run in parallel in yeast cells (57,58). Thus, at least a portion of the Sec61p and the exocyst complexes can be in close proximity with each other within a cell. Lipschutz et al. (Ref. 65; accompanying article) demonstrate that ER translocation machinery components can exist in a complex with components of the exocyst also in Madin-Darby canine kidney cells. Interactions between the exocyst and ER resident proteins are supported also by the observed coimmunoprecipitation and partial colocalization of the mammalian Sec8p with the inositol 1,4,5,-trisphosphate receptor, IP 3 R (59).
The physical interactions with the exocyst complex both in yeast and mammalian cells were studied only with the ␤ subunit of the ER translocon. In these experiments several components (Sec6, Sec8, Sec10, and Sec15) of the exocyst complex were detected. Although our genetic data suggest that all components of these complexes may have roles in these interactions, the ␤ subunit clearly seems to play the most important role. Increased production of secreted proteins was achieved by overexpression of SEB1 only, not with that of SEC61 or SSS1, 3 and SEB1 was most efficient also in the suppression of the exocyst mutations. Overexpression of a mutant form of Seb1p lacking the transmembrane domain failed to suppress sec15-1 temperature-sensitive growth. This suggests that association of Seb1p in the ER membrane may be required to establish the functional link with the exocyst complex. Presently, we do not know the proportion of Seb1p in the translocating Sec61p complexes in SEB1 overexpressing cells. As the amount of Sec61p or Sss1p does not increase in an Seb1p-overexpressing strain (data not shown), a portion of Seb1p may exist outside of these complexes under these conditions. The possible contribution of the unassembled Seb1p to the interaction remains to be studied. It may be that Seb1p interaction with the exocyst is not direct but is mediated by other, presently unknown factors.
In addition to the observed physical and genetic interactions, an additional functional link was observed between the early and late processes of protein secretion. Both overexpression of Seb1p and the Sec15p-interacting protein, Sec4p, resulted in enhanced protein secretion in yeast. Increased production of secreted proteins in yeast can occur as a result of ER proliferation caused by overexpression of a heterologous ER membrane protein (60,61). On the basis of previous morphological experiments, even the strongly overexpressed Seb1p is localized in the ER (20). When analyzed by electron microscopy, SEB1 overexpression does not cause ER membrane proliferation nor does overexpression of Seb1p increase the levels of Sec61p or Sss1p (data not shown). Lipschutz and colleagues (62) have shown that hsec10 overexpression in Madin-Darby canine kidney cells increases the amount of secreted and basolateral membrane proteins. Similarly, we show here that overexpression of Sec4p, a Sec15p-interacting protein, results in enhancement of protein secretion. The overexpression of SEC15 is toxic to yeast cells (39). Thus its possible enhancing effect on secretion could not be studied. In an accompanying article, Lipschutz et al. (65) now show an unaltered transcript level of proteins with increased secretion levels. Likewise, enhanced production of ␣-amylase in SSO1 or SSO2-overexpressing cells is not because of increased transcription of the ␣-amylase gene (42). Taken together, these results suggest that the increased production of secreted proteins in the overexpression strains occurs at post-transcriptional level.
In mammalian cells the Sec61␤ subunit has been shown to bind nontranslating ribosomes (63). Thus, it is possible that increased amount of the ␤ subunit facilitates ribosome binding to ER and thereby increases the number of ribosomes that are available for binding and translation of the messengers of secreted proteins in SEB1 overexpressing cells. This perhaps explains the observed small increase in the ␣-amylase transcript level in SEB1 overexpressing yeast cells. Efficient ribosome binding and ER targeting have been shown to increase the half-life of mRNAs in yeast (63,64). Although a link between the ER translocation machinery and exocyst is observed both in yeast and in mammalian cells, differences in forming and maintaining this link are likely to exist because of the different cellular signals that these cells encounter in their growth environments.
The functional and physical interactions between the ER translocon and the exocyst complex indicate the existence of a regulatory circuit that may be used for adjusting the level of translation to the secretory capacity of the cell. Increasing the level of a component involved in this regulation could lead to enhanced production of secreted proteins in wild type cells. Along the same lines, this cross-talk would be compromised by a decreased amount of (or defective) regulatory component, e.g. destabilized exocyst complex in the mutant cells. Increasing the amount of an interacting component, in this case Seb1p, could restore sufficient communication even in mutant cells and allow growth at the nonpermissive temperature. This view is supported by the fact that Sec15-1p and Seb1p were found in the same protein complex only when SEB1 was overexpressed (Fig. 5).
Taken together, our data reported here reveal novel links between molecular complexes of the ER translocon and the exocyst complex, implying new regulatory circuits between the ER and plasma membrane in the control of protein secretion in yeast.