Distinct Domains within Yeast Sec61p Involved in Post-translational Translocation and Protein Dislocation*

The translocation of secretory polypeptides into and across the membrane of the endoplasmic reticulum (ER) occurs at the translocon, a pore-forming structure that orchestrates the transport and maturation of polypeptides at the ER membrane. Recent data also suggest that misfolded or unassembled polypeptides exit the ER via the translocon for degradation by the cytosolic ubiquitin/proteasome pathway. Sec61p is a highly conserved multispanning membrane protein that constitutes a core component of the translocon. We have found that the essential function of the Saccharomyces cerevisiaeSec61p is retained upon deletion of either of two internal regions that include transmembrane domains 2 and 3, respectively. However, a deletion mutation encompassing both of these domains was found to be nonfunctional. Characterization of yeast mutants expressing the viable deletion alleles of Sec61p has revealed defects in post-translational translocation. In addition, the transmembrane domain 3 deletion mutant is induced for the unfolded protein response and is defective in the dislocation of a misfolded ER protein. These data demonstrate that the various activities of Sec61p can be functionally dissected. In particular, the transmembrane domain 2 region plays a role in post-translational translocation that is required neither for cotranslational translocation nor for protein dislocation.

The translocation of secretory polypeptides into and across the membrane of the endoplasmic reticulum (ER) occurs at the translocon, a pore-forming structure that orchestrates the transport and maturation of polypeptides at the ER membrane. Recent data also suggest that misfolded or unassembled polypeptides exit the ER via the translocon for degradation by the cytosolic ubiquitin/proteasome pathway. Sec61p is a highly conserved multispanning membrane protein that constitutes a core component of the translocon. We have found that the essential function of the Saccharomyces cerevisiae Sec61p is retained upon deletion of either of two internal regions that include transmembrane domains 2 and 3, respectively. However, a deletion mutation encompassing both of these domains was found to be nonfunctional. Characterization of yeast mutants expressing the viable deletion alleles of Sec61p has revealed defects in post-translational translocation. In addition, the transmembrane domain 3 deletion mutant is induced for the unfolded protein response and is defective in the dislocation of a misfolded ER protein. These data demonstrate that the various activities of Sec61p can be functionally dissected. In particular, the transmembrane domain 2 region plays a role in post-translational translocation that is required neither for cotranslational translocation nor for protein dislocation.
Secretory protein translocation across the ER 1 membrane occurs at gated channels containing an aqueous interior (1)(2)(3), which together with other components that function at these channels constitute the translocon. The primary functions of the translocon are to facilitate protein transport from the cytosol to the ER lumen and to act in the assembly of integral membrane proteins into the ER bilayer. However, a range of other functions involved in the maturation of translocating polypeptides and the degradation of ER polypeptides are also linked to the translocon. Direct visualization of the core translocon and fluorophore quenching experiments have identified a macromolecular structure with a regulated pore diameter of 15-60 Å that forms the channel through which the nascent polypeptide traverses the membrane (4 -6). The reconstitution of translocation reactions in vitro using purified components (7,8), together with a wealth of other data, has demonstrated that the core translocon in both yeast and mammalian ER is formed by the highly conserved heterotrimeric Sec61 complex (for review see Refs. 9 and 10).
Characterization of the yeast Sec61 complex has shown that it consists of a polytopic membrane protein Sec61p and two small C-terminal anchor proteins, Sss1p and Sbh1p (8,(11)(12)(13). Sec61p is a 53-kDa integral membrane protein that spans the ER bilayer 10 times (14). A functional heptameric complex consisting of the Sec63 complex (containing the Sec63p, Sec62p, Sec71p, and Sec72p polypeptides) and Sec61 complex has been purified (8). Genetic data suggest that the yeast Sec61 complex participates in both cotranslational and post-translational translocation, whereas the Sec63 complex is primarily involved in post-translational translocation (10).
The ER has to guarantee proper folding of nascent membrane and secretory proteins before transport to their site of action. Recent studies have shown that several soluble and integral membrane ER substrates are degraded via the cytosolic ubiquitin/proteasome pathway (15)(16)(17), indicating the necessity for export of misfolded proteins from the ER to the cytosol. Studies carried out in yeast and mammalian systems have identified an export pathway (16,18,19), in which the translocon is implicated as the route of export (19 -22).
The roles of any specific amino acids or domains of Sec61p are largely undefined, despite the number of functions attributed to this translocon component. A region of yeast Sec61p including transmembrane domain (TM) 6, TM7, and TM8 (residues Leu 232 -Arg 406 ) has been shown to interact with Sss1p (23). Cross-linking studies have implicated TM2 and TM7 in recognition of pp␣F signal sequence during its interaction with the translocon (24). In addition, some amino acid residues around the lumenal ends of TM3 and TM4 have been identified as being important for both translocation and dislocation (20,25). Here, we describe two novel deletion alleles, sec61-⌬2 and sec61-⌬3, in which TM2 or TM3 containing regions of Sec61p have been deleted, respectively. Yeast strains expressing these alleles were viable and displayed defects in post-translational translocation. In addition the sec61-⌬3 mutant strain was induced for the unfolded protein response and had a defect in ER dislocation. These data suggest that the TM2/TM3 region of Sec61p plays a role in the post-translational translocation of preproteins and that in addition the TM3 region is required for the efficient dislocation of misfolded proteins. Media and Growth Conditions-Yeast strains were grown at 30°C in YP medium (2% peptone, 1% yeast extract) containing either 2% glucose (YPD) or 2% galactose (YPGAL) or in minimal medium (0.67% yeast nitrogen base) with 2% glucose or galactose, plus appropriate supplements for selective growth. Solid media were supplemented with 2% Bacto-agar. All media were from Difco Laboratories. Growth rates were determined in YPD cultures at 30°C with cell density monitored by A 600 nm using a CE292 spectrophotometer (Cecil Instruments, UK). Yeast transformations and 5-FOA counter selection of Ura3 ϩ cells were carried out as described previously (23).

Materials
Yeast Strains-Saccharomyces cerevisiae strains used in this study are listed in Table I. BWY46 was derived by tetrad dissection of diploid W303. Diploid strain BWY100 was made by mating BWY46 with BWY47 followed by counter selection of pBW62 on 5-FOA medium. An IRE1 null allele strain (JTY20) was made by transformation of TR3 with a linear ire1::kanMX4 cassette replacing IRE1 codons 2-1108 with kanMX4. The cassette was generated by polymerase chain reaction amplification of a kanMX4 module from pFA6-kanMX4 (33) using primers 5Ј-ACAGCATATCTGAGGAATTAATATTTTAGCACTTTGAAAAA-TGCGCGTACGCTGCAGGTCGAC-3Ј and 5Ј-ATGCAATAATCAACCA-AGAAGAAGCAGAGGGGCATGAACATGTTAATCGATGAATTCGAG-CTCG-3Ј with 5Ј homology to the IRE1 open reading frame flanking regions and 3Ј homology with kanMX4. TR3 was transformed with 1 g of this template with selection on YPD containing 200 g/ml Geneticin. Correct integration of the ire1::kanMX4 cassette was confirmed by polymerase chain reaction analysis of the integrated locus, and the resulting strain JTY20 displayed the growth phenotypes previously reported for IRE1 mutant strains (28).
A SEC61/sec61::HIS3, SEC65/sec65-1 heterozygous diploid was constructed by mating CSY126 with BWY97. Extensive subculture on YPD medium was used to generate a Leu Ϫ strain (BWY308) that had lost pBW132. The same approach was used to generate a SEC61/sec61:: HIS3, SEC62/sec62-1 heterozygous diploid by crossing RDM50 -94C with BWY67, with subculture on YPD medium to produce a Leu Ϫ strain (BWY341) lacking pBW11. A sec61::HIS3, sec62-1 haploid was generated by transformation of BWY341 with pBW7 to uracil prototrophy, followed by sporulation and dissection. Tetrads containing four viable spores were scored for temperature sensitivity (37°C), Ura ϩ /Ura Ϫ and the ability to remain Ura ϩ upon subculture on YPD medium. A temperature-sensitive, His ϩ spore that was stable for the Ura ϩ phenotype indicative of the presence of pBW7 in a sec61::HIS3 background was named BWY343.
SEC61 alleles were combined with the prc1-1 allele encoding CPY* by transforming W303-1C with a SEC61 plasmid, pBW7 and subsequently replacing the genomic SEC61 with the sec61::HIS3 (null mutant) cassette as described previously (11). His ϩ transformants were confirmed as correct by their inability to grow on 5-FOA medium and hence the inability to survive without the SEC61 plasmid. The wild type and deletion allele plasmids pBW11, pBW132, and pBW134 (see below) were transformed into one such transformant (BWY402) and shuffled into the null mutant background on 5-FOA medium to produce strains BWY403-405.
The UPRE-lacZ reporter cassette present in strain CS165 (Table I) was recovered by gap repair. Nae1/BamHI-digested pRS314 was transformed into CS165, and Trp ϩ transformants were selected. Plasmid isolates recovered in E. coli (36) from these transformants contained the expected 4.2-kilobase pair repaired insert. HindIII digestion of one of these isolates yielded a fragment containing the UPRE-lacZ cassette, which was ligated into HindIII-digested pRS316 (CEN, URA3) to make pJT30.
Radiolabeling and Immunoprecipitation-Preparation of radiolabeled yeast cell extracts was carried out as described previously (11). 2.0 A 600 nm unit equivalents of cell extract were added to saturating amounts of antiserum followed by with rotation for 4 h at room temperature. Protein A-Sepharose CL4B beads (Sigma) at 20% suspension in IP buffer (1.25% Triton X-100, 187.5 mM NaCl, 6.25 mM EDTA, 62.5 mM Tris⅐HCl, pH 7.4) were added at 5 l/1 l of antiserum with further incubation at room temperature for 2 h. Beads were pelleted in a microcentrifuge and washed three times with 1 ml of IP buffer. The addition of 40 l of SDS sample buffer containing 5% ␤-mercaptoethanol and heating to 95°C for 5 min was used to dissociate antigen from the beads.
DPAPB antiserum was raised using an hexahistidine-tagged C-terminal portion of DPAPB. For this purpose, a 3.4-kilobase pair SspI/XbaI fragment derived from pRG1 (37) was subcloned into pET16b (Novagen), to create an in-frame fusion between 10 His residues and residues 200 -642 of DPAPB under the control of the T7 promoter. The fusion protein was expressed and purified as described previously for Histagged proteins (14). Eluted protein fractions were pooled and subjected to 10% SDS-PAGE. A band corresponding to the 77-kDa His-tagged protein was excised, crushed, and provided an antigen to inoculate sheep at the Scottish Antibody Production Unit (Lanark, Scotland). Antisera were tested by immunoblotting whole cell extracts detecting a ϳ124-kDa species and a 97-kDa species after tunicamycin treatment consistent with both the mature and unglycosylated forms of the protein (37). The CPY antiserum will be described in another manuscript. 2 ␤-Galactosidase Assays-Yeast cells were grown at 30°C in minimal medium containing 2% glucose and appropriate supplements. Cultures were diluted to A 600 nm of 0.2 and grown for a further 4 h. 0.5 A 600 nm units of cells were removed to a 15-ml tube, and the volume was adjusted to 1.35 ml with minimal medium. 150 l of 10ϫ Z buffer (38) containing 2% Sarcosyl NL-30 was added, and the sample was shaken gently at 30°C for 30 min. 375 l of O-nitrophenyl ␤-D-galactopyranoside solution (4 mg/ml in 1ϫ Z buffer) was added with further incubation at 30°C for 30 min. 1 ml Na 2 CO 3 solution was added to stop the reaction, and 1 ml was removed to a 1.5-ml microcentrifuge tube and cleared by centrifugation for 2 min at 14,000 ϫ g. The A 420 nm of 0.9 ml of cleared sample was recorded using a CE292 spectrophotometer (Cecil Instruments).

Transmembrane Domains 2 and 3 Are Not Essential for
Sec61p Function-Several observations support the suggestion that TM2 may have a crucial role in Sec61p function. Firstly, multiple alignment of all eukaryotic Sec61p sequences and the more distantly related bacterial SecY homologues reveals the overall conservation of just three residues corresponding to Gly 81 , Pro 84 , and Gln 93 of S. cerevisiae Sec61p (39, 40) all located in TM2 (Fig. 1A). Secondly, during Sec61p assembly into the ER membrane, TM2 appears to be unable to act as a stop-transfer domain, possibly because of its limited hydrophobicity, and requires a downstream region for its membrane integration (14). These findings suggest that specific residues that contribute to the limited hydrophobicity of TM2 have been conserved because they are crucial for some aspect of Sec61p function. In addition, cross-linking studies suggest a role for TM2 in binding the signal sequence of pp␣F at an early sep in its translocation (24).
To investigate the role of TM2 in Sec61p function, we used site-directed mutagenesis to change the conserved residue Gln 93 to Leu 93 . When this mutation was shuffled into a SEC61 null mutant (⌬sec61) background on a low copy plasmid, the cells grew without any discernible growth phenotype (data not shown). Given this rather surprising finding, we used BamHI sites previously introduced into the SEC61 sequence to con-struct a deletion allele (sec61-⌬2) lacking 35 residues (Asn 73 -Ser 107 ) of Sec61p including TM2 (Fig. 1A). A low copy plasmid encoding sec61-⌬2 (pBW134) was introduced into a wild type strain to examine the expression of the mutant protein (Sec61⌬2p). The Sec61⌬2p was identified as a relatively poorly expressed species by immunoblot analysis (Fig. 2A, lane 3). When the plasmid was introduced into BWY47, a haploid strain permitting the analysis of sec61 mutations by complementation of the sec61::HIS3 (⌬sec61) null allele (14), the sec61-⌬2 allele was found to complement the growth defect of the null mutant as efficiently as SEC61 expressed from a low copy plasmid (Fig. 2B). Plasmid pBW134 was subsequently 2 W. Cui and C. J. Stirling, manuscript in preparation.
FIG. 1. The structure of yeast Sec61p internal deletions. A, amino acid sequence and transmembrane topology of the N-terminal 230 residues of Sec61p including transmembrane domains 1-5 shown as rectangular boxes. This structure has been previously reported (14) and includes a putative amphipathic helix at the N terminus designated by the shaded box. The three residues conserved between Sec61p/ SecY (Gly 81 , Pro 84 , and Gln 93 of Sec61p) are located in TM2 and are shown in shaded circles. The residues removed upon constructing the deletions encompassing TM2 and TM3 are indicated by arrows: sec61-⌬2 (Asn 73 -Ser 107 ), sec61-⌬3 (Lys 108 -Ser 143 ), and sec61-⌬2/3 (Asn 73 -Ser 143 ). B, schematic representation of the Sec61p deletion series included in this report. The following residues were deleted: shuffled into the ⌬sec61 background by passage of the BWY47 transformant on 5-FOA medium to counter-select the URA3 plasmid pBW62. The expression of Sec61⌬2p was examined in the resulting strain, BWY98,by immunoblot analysis. The mu-tant protein was now found to be expressed at a level comparable with wild type Sec61p ( Fig. 2A, lanes 6 and 7), suggesting that it is stabilized when present as the only functional source of Sec61p.
We considered the possibility that TM3 may act in concert with TM2 not only during their membrane assembly but also in Sec61p function. We addressed this by taking the same approach as described above for TM2 and constructed a sec61-⌬3 allele that removed 36 residues (Lys 108 -Ser 143 ) including TM3 immediately downstream of the TM2 deletion (Fig. 1A). The Sec61⌬3p species was also expressed at a low level in a wild type background (Fig. 2A, lane 4). The sec61-⌬3 allele could complement the null mutant (Fig. 2B), and after plasmid shuffling to generate BWY97, the expression level of Sec61⌬3p was also dramatically raised ( Fig. 2A, lane 8). Because either the TM2 or the TM3 region of Sec61p can be deleted without loss of function, we went on to test a deletion of residues Asn 73 -Ser 143 equivalent to a combination of both deletions. The level of Sec61⌬2/3p expressed by this sec61-⌬2/3 allele was again poor relative to wild type ( Fig. 2A, lane 5). However, the sec61-⌬2/3 allele could not complement the ⌬sec61 mutant (Fig. 2B).
We have previously shown that sec61 alleles expressing the following deletions that removed the TMs shown in parentheses are nonfunctional: ⌬Phe 13 -Ser 72 (TM1); ⌬Phe 13 -Ser 107 (TM1/TM2); ⌬Asp 437 -Phe 468 (TM10); and ⌬Ser 409 -Phe 468 (TM9/TM10) (23). To extend this study, we constructed a series of deletion alleles individually encompassing the remaining transmembrane domains and adjacent pairs of domains (see "Experimental Procedures" and Fig. 1B). All of these deletion alleles expressed species smaller than wild type Sec61p as expected, but none of them were able to complement the null mutation when introduced into strain BWY47 followed by incubation at either 30 or 24°C (data not shown). In addition, all of the deletion alleles described in this report were tested for function by tetrad dissection into a ⌬sec61 background after transformation of the SEC61/sec61::HIS3 heterozygous strain BWY100 (Table I). Only the sec61-⌬2 and sec61-⌬3 alleles could support germination and growth of ⌬sec61 spores.
Analysis of Growth and Translocation Phenotypes of the Viable Deletion Mutants of Sec61p-The sec61-⌬2 (BWY98) and sec61-⌬3 (BWY97) mutant strains had moderate growth defects with doubling times of 115 and 140 min, respectively, compared with 95 min for the isogenic wild type (BWY67). Testing for temperature (37°C) and cold sensitivity (17°C) on YPD medium revealed that BWY97 was cold-sensitive for growth (not shown).
The effect of the sec61 deletion mutations on translocation was examined by assaying the biogenesis of CPY and DPAPB, which are translocated via post-and co-translational mechanisms, respectively. The wild type strain did not accumulate the unprocessed ppCPY (Fig. 3A, lanes 2), whereas both the sec61-⌬2 and sec61-⌬3 strains were substantially defective in the translocation of this precursor (Fig. 3A, lanes 3 and 4). No significant defects in the translocation of DPAPB were apparent in either wild type or sec61-⌬2 cells (Fig. 3A, lanes 2 and 3). However, the presence of some unglycosylated DPAPB in sec61-⌬3 cells (Fig. 3A, lane 3) indicates an additional defect in the SRP-dependent co-translational translocation pathway. The translocation of pp␣F (post-translationally translocated) was analyzed by immunoblotting (Fig. 3B). ER forms of this precursor are absent from the wild type sample (Fig. 3B, lane 3) because it is rapidly processed to smaller peptides in the Golgi. Tunicamycin treatment resulted in the accumulation of p␣F, and both mutants accumulated the slightly larger pp␣F (Fig. 3B, lanes 4  and 5) also detectable in the sec61-3 strain (Fig. 3B, lane 1).
The targeting of ppCPY and pp␣F to the ER can occur inde-  sec61-⌬2 and sec61-⌬3 alleles. A, aliquots (0.1 A 280 nm equivalents) of membranes were solubilized in sample buffer at 50°C and resolved by 12% SDS-PAGE before immunoblotting with an antiserum specific for the C terminus of Sec61p (11). Membrane preparation and immunoblotting were performed as described previously (23). Lane 1 represents BWY46 transformed with vector (pRS315), and lanes 2-5 are from strains BWY89, BWY90, BWY91, and BWY92 derived from BWY46 by transformation with pBW11 (SEC61), pBW134 (sec61-⌬2), pBW132 (sec61-⌬3), and pBW140 (sec61-⌬2/3), respectively. Lanes 6 -8 are from BWY67 (SEC61), BWY98 (sec61-⌬2), and BWY97 (sec61-⌬3) derived from BWY47 after passage on 5-FOA medium to shuffle the wild type and deletion allele plasmids into the ⌬sec61 background. B, plasmids encoding the deletion alleles were transformed into haploid BWY47, which contains a ⌬sec61 null mutation (sec61::HIS3) that is complemented by the presence of a low copy plasmid (pBW62) encoding a GAL1-SEC61 allele. This confers the ability of this strain to grow on galactose medium but the inability to grow on glucose medium (14). All transformants grew on galactose medium, but only the wild type, sec61-⌬2 and sec61-⌬3 alleles could complement the null mutation upon plating to glucose medium. pendently of SRP, and translocation of ppCPY and pp␣F is dependent upon both the Sec61 and Sec63 complexes (11,30,41,42), whereas DPAPB is targeted to the ER by the SRP-dependent pathway and then cotranslationally translocated across the membrane via the Sec61 complex (11,29). DPAPB translocation appears to be unaffected by mutations in components of the Sec63 complex, suggesting that this complex is exclusively involved in the post-translational translocation reaction (42). The sec61-⌬2 and sec61-⌬3 strains exhibited much stronger defects in the translocation of ppCPY and pp␣F than was apparent for DPAPB, suggesting that the TM2/3 region of Sec61p may be primarily involved in the post-translational translocation mechanism. To further test this hypothesis we have examined the effects of the sec61-⌬2 and sec61-⌬3 alleles in combination with either the sec62-1 or sec65-1 temperaturesensitive mutations that are specifically defective in posttranslational and cotranslational translocation, respectively (29,30). Yeast strains BWY67 (SEC61), BWY97 (sec61-⌬3), and BWY98 (sec61-⌬2) were mated to RDM50 -94C (sec62-1) and CSY126 (sec65-1), and the resulting diploids were sporulated and subjected to tetrad analysis. Spores were incubated at 24°C, and analysis of auxotrophic markers and temperature sensitivity among the progeny showed that viable spores lacking notable growth defects were obtained for the deletion alleles in combination with sec65-1 (data not shown). Despite extensive tetrad dissection of several independent diploids, viable spores combining the sec61-⌬ alleles with sec62-1 were not obtained, whereas the wild type SEC61 plasmid (pBW11) was successfully combined with the sec62-1 mutation in a ⌬sec61 background. This indicated that the double mutant combination might be synthetically lethal. To verify this, a haploid strain was constructed (BWY343) containing sec62-1 and the sec61::HIS3 allele plus a complementing SEC61 plasmid (pBW7; URA3). Wild type and deletion derivative plasmids were used to transform BWY343 to leucine prototrophy, and the resulting transformants were then plated to 5-FOA medium to counter-select pBW7 (Fig. 4). The strains containing vector or plasmids expressing the sec61-⌬ alleles were sensitive to growth on 5-FOA medium, indicating that any cells losing plasmid pBW7 through mis-segregation were inviable, thus confirming the synthetic lethality with sec62-1.
The sec61-⌬3 Mutant Is Induced for the Unfolded Protein Response-The Sec61 complex has been implicated as the route across the ER bilayer during the of export of misfolded or unassembled proteins to the cytosol (19 -21) commonly referred to as ER dislocation. A defect in this export pathway would be expected to result in the accumulation of misfolded proteins in the ER, leading to the induction of the UPR. This response is induced under conditions of protein misfolding, such as the inhibition of disulphide bond formation by reducing agents or the inhibition of N-glycosylation by tunicamycin. It follows that mutants defective in dislocation might display enhanced sensitivity to growth under such stress conditions. We tested the

FIG. 3. Analysis of protein translocation phenotypes in sec61-⌬ mutants.
A, wild type and mutant cells were pulse labeled with [ 35 S]Lmethionine for 5 min (ppCPY) and 10 min (DPAPB) and processed as described under "Experimental Procedures." Extracts were immunoprecipitated with DPAPB-or ppCPY-specific antiserum, and 2.0 A 600 nm equivalents were resolved by 7.5% SDS-PAGE (DPAPB) and 10% SDS-PAGE (ppCPY). Tunicamycin treatment was performed by addition of the drug to 10 g/ml 1 h before harvesting cells. CPY is synthesized as an inactive prepro-protein (ppCPY) whose signal sequence is removed upon entering the ER lumen. The signal cleaved proCPY (pCPY) is N-glycosylated in the ER to the p1 form (p1CPY), which undergoes further processing in the Golgi and the vacuole to generate the final mature form. The ppCPY, pCPY, and p1CPY forms are indicated. DPAPB lacks a cleavable signal sequence but assembles in the ER membrane with type II topology and N-glycosylation to its C-terminal domain (37). The nonglycosylated (pDPAPB) and the mature glycosylated (mDPAPB) forms are indicated. Lane 1, BWY67ϩTun; lane 2, BWY67; lane 3, BWY98; lane 4, BWY97. B, immunoblot analysis of pp␣F translocation. Whole cell yeast extracts were prepared from YPD cultures grown at 30°C as described previously except with 95°C incubations (14). 0.5 A 600 nm equivalents were resolved by 14% SDS-PAGE and immunodetected with antiserum against pp␣F. Concomitant with translocation into the ER, the signal sequence of pp␣F is cleaved, yielding pro-␣-factor (p␣F), which is rapidly core glycosylated and subsequently processed to smaller peptides in the Golgi.  4. The sec61-⌬2 and sec61-⌬3 alleles are synthetically lethal in combination with sec62-1. Strain BWY343 was transformed to leucine prototrophy with pRS315 (Vector), pBW11 (SEC61), pBW134 (sec61-⌬2), and pBW132 (sec61-⌬3). The transformants were plated to minimal medium containing 2% glucose and appropriate supplements but lacking leucine and uracil to maintain selection for pBW7 and the SEC61/sec61 encoding plasmids or to the same medium containing uracil and 5-FOA to counter-select pBW7.
possibility that the sec61-⌬2 and sec61-⌬3 mutants were sensitive to these conditions by plating to media containing tunicamycin or dithiothreitol (Fig. 5A). The IRE1 gene encodes Ire1p, which is essential for the induction of the UPR, and ⌬ire1 strains are very sensitive to reducing agents and tunicamycin (28). The sec61-⌬3 strain was almost as sensitive as ⌬ire1 cells, whereas sec61-⌬2 cells were sensitive to a lesser degree (Fig. 5A). There was no apparent sensitivity when the deletion alleles were expressed in a wild type background (Fig. 5A).
The UPR results in the co-ordinate transcriptional induction of a set of genes involved in protein chaperone and folding activities within the ER (43). The UPRE has been identified in the promoters of these genes and is required for their transcriptional induction in the yeast UPR (Ref. 44 and references therein). We analyzed the induction of the UPR by introducing a low copy plasmid (pJT30) containing a reporter allele of lacZ under the control of a yeast UPRE element into the deletion mutants. The level of constitutive UPR induction was then monitored by measurement of ␤-galactosidase activity. The absence of UPR induction in the wild type demonstrated that the response was not induced under the growth conditions used, and similarly there was no induction in the ⌬ire1 strain because it cannot elicit the UPR (Fig. 5B). Both of the sec61-⌬ mutants were induced for the UPR, with the sec61-⌬3 mutant having a 3-fold higher level of ␤-galactosidase activity than sec61-⌬2. This is consistent with the greater sensitivity of the sec61-⌬3 mutant to dithiothreitol and tunicamycin. The level of induction in sec61-⌬3 cells is approaching the maximal level of induction observed with tunicamycin treatment of the wild type. Expression of the deletion alleles in a wild type background did not elicit the UPR (Fig. 5B).
The sec61-⌬3 Mutant Has a Defect in ER Dislocation-The prc1-1 allele encodes a form of CPY (CPY*) that misfolds in the ER lumen accumulating in the p1 form before dislocation to the cytosol for degradation by the proteasome (16). Haploid strains combining the sec61-⌬ alleles and the ⌬der3 degradation mutant with prc1-1 were analyzed for CPY * degradation by pulsechase analysis (Fig. 6). The sec61-⌬2 mutant displayed only a very minor defect in degradation with a CPY* t1 ⁄2 of 22 min compared to a t1 ⁄2 of 20 min for the wild type. The sec61-⌬3 mutant, however, displayed a delay in degradation with a CPY* t1 ⁄2 of 45 min, compared with 88 min for the ⌬der3 dislocation mutant (Fig. 6). The half-life of CPY * was calculated from the second chase point (30 min) for the sec61-⌬ strains, by which time only the p1CPY * was apparent (Fig. 6). The appearance of both p1 and ppCPY * forms in the sec61-⌬ strains demonstrates that at least a proportion of the ppCPY accumulated upon a shorter pulse labeling (Fig. 3A) chases into the ER. This suggests that TM2/TM3 promotes the efficiency of post-translational translocation.
FIG . 6. The sec61-⌬3 mutant has a defect in ER dislocation. The BWY403 (SEC61), BWY404 (sec61-⌬3), BWY405 (sec61-⌬2), and the dislocation mutant W303-1C⌬3 (⌬der3) strains containing the prc1-1 mutation were analyzed for CPY* degradation by pulse-chase. Cells were grown at 30°C, harvested and labeled with [ 35 S]L-methionine for 20 min. The chase was initiated by the addition of cold methionine to 2.0 mM with incubation continuing at 30°C. 2.0 A 600 nm equivalents were removed at the indicated time points, and cell extracts were prepared and immunoprecipitated with CPY antiserum. 2.0 A 600 nm equivalents were resolved by 10% SDS-PAGE, and the fixed gels were subject to detection and quantification using a Fujix BAS 2000 Bioimager. The p1 and ppCPY forms are shown. plexes was analyzed in both wild type and sec61-⌬ digitoninsolubilized membranes, using a lectin Con-A binding assay described by Pilon et al. (25). The Sec complex binds to lectin Con-A because of the presence of oligosaccharides on Sec71p. In agreement with previously reported data, we found that ϳ80% of the Sec61p in wild type membranes solubilized in digitonin was found in the high speed supernatant (not shown). The fractionation of wild type Sec61p was found to be comparable with that reported previously (25) with the majority of digitonin soluble Sec61p found in the Con-A binding fraction (Fig.   7A). The majority of Sec63p in wild type membranes was also found in the Con-A binding fraction, indicating that Sec61p and Sec63p co-fractionated as part of the heptameric Sec complex (Fig. 7A). With sec61-⌬2 membranes, a higher proportion of Sec61⌬2p was found in the Con-A-binding fraction versus the Con-A-free fraction in comparison with wild type. This apparent increase in the level of heptameric complex may reflect a role for TM2 in complex disassembly but could also be explained by a decrease in the stability of any free pool of the mutant Sec61⌬2p protein compared with wild type. In contrast Sec61⌬3p  (14) derived from wild type (BWY67), sec61-⌬2 (BWY98), and sec61-⌬3 (BWY97) strains. ER membrane proteins were fractionated after solubilization in digitonin. Total, total microsomal protein; Free, fraction of digitonin-soluble proteins not binding to Con-A; Con-A, proteins binding to concanavalin A; RAMP, digitonin soluble ribosomal-associated membrane protein fraction. Equal aliquots (0.1 A 280 nm equivalents) were analyzed by SDS-PAGE and immunoblotting with antisera against Sec63p and the C terminus of Sec61p. B, protease protection of Sec61p, Sec61-⌬2p, and Sec61-⌬3p in microsomal membranes. Membranes prepared from BWY67, BWY98, and BWY97 at a concentration of 0.1 A 280 nm equivalents/10 l were treated with 500 g/ml (BWY67, BWY97) or 200 g/ml (BWY98) proteinase K (PrK) in the absence or presence of 0.2% Nonidet P-40 on ice for 30 min. Reactions were terminated by the addition of 15% trichloroacetic acid. 0.2 A 280 nm equivalents were resolved by 12.5% SDS-PAGE and immunoblotted with antiserum against the C terminus of Sec61p. Untreated membranes prepared from strains expressing the TM 1-2/3-10 and TM1-3/4 -10 complementary fragment alleles of SEC61, respectively, were also included. The C 8 (TM3-10) fragment (lane 1) and the C 7 (TM4 -10) and N 3 (TM1-3) fragments (lane 2) are indicated. C, the topological frustration of TM domains and protease-sensitive regions in Sec61⌬2p and Sec61⌬3p. The wild type shows the residues important in the construction of the Sec61p deletion derivatives. Sec61⌬2p has a deletion of residues Asn 73 -Ser 107 and Sec61⌬3p has a deletion of residues Lys 108 -Ser 143 . The C 8 (TM3-10) fragment was derived by fusion of residues 1-13 to Lys 108 and likewise C 7 (TM4 -10) by fusion of 1-13 to Asp 144 (23). The topological alteration in Sec61⌬2p is proposed to result in the frustration of TM3 such that it no longer spans the ER membrane but instead is embedded in the bilayer at the lumenal side. As a consequence, that part of the TM 2-3 hydrophilic loop remaining in this mutant protein is translocated to the ER lumen (shown in black) as a hybrid with the remaining portion of the TM 1-2 loop. Proteinase K cleavage in this hybrid loop close to TM3 would yield a fragment slightly smaller than C 8 (TM3-10) as detected. The same phenomenon is proposed for Sec61⌬3p, except that in this case TM2 is frustrated as an embedded domain and the remaining part of the TM 2-3 loop (shown in black) is translocated to the ER lumen as a hybrid with the remaining portion of the TM 3-4 loop. Proteinase K cleavage in this hybrid loop close to TM4 would yield a fragment slightly smaller than C 7 (TM4 -10) as detected.
was found predominantly in the Con-A-free fraction, but the fractionation of Sec63p was largely unaffected, suggesting that the interaction between the Sec63 complex and the Sec61⌬3p complex is disrupted compared with wild type. Low levels of Sec61p and Sec61⌬2p but not Sec61⌬3p were also found in the ribosomal-associated membrane protein fractions (Fig. 7A).
Perturbation of membrane topology in Sec61⌬2p and Sec61⌬3p was analyzed by proteinase K treatment (Fig. 7B). The deletion of an internal TM domain in a polytopic membrane protein is expected to result in a topological alteration. Native Sec61p appeared to be completely resistant to cleavage with proteinase K (Fig. 7B, lanes 3-5) irrespective of the presence of 0.2% Nonidet P-40, which permeabilizes membranes to proteases (14). However, Sec61⌬2p was sensitive to cleavage in the presence of Nonidet P-40, yielding a product (Fig. 7B, lane 8, asterisk) that migrated slightly smaller than the TM3-10 (C 8 ) Sec61p fragment (Fig. 7B, lane 1). This suggests an altered structure on the luminal side of the membrane, which is cleaved at a sensitive site either at the TM1-3 fusion region or within TM3. Treatment of Sec61⌬3p membranes in the presence of Nonidet P-40 yielded a proteolytic product slightly smaller than the TM4 -10 (C 7 ) Sec61p fragment (Fig. 7B, lane 11, asterisk), indicating an altered structure on the lumenal face that is cleaved close to TM4. A lesser amount of this fragment detected in the absence of detergent (Fig. 7B, lane 10) was probably due to the presence of some broken or inverted membranes. The C 8 and C 7 fragments consist of Sec61p residues 1-13 fused to Lys 108 and Asp 144 of Sec61p, respectively (23), and the presence of the N-terminal residues would explain why they migrate slightly larger than the proteolytic fragments. This suggests protease cleavage close to residues 108 and 144 in Sec61⌬2p and Sec61⌬3p, respectively, consistent with a local alteration of Sec61p structure, namely the topological frustration of TM3 in Sec61⌬2p and also TM2 in Sec61⌬3p on the luminal face of the membrane. DISCUSSION Remarkably, the essential function of Sec61p is retained upon deletion of internal 35 (⌬73-107) or 36 (⌬108 -143) amino acid residue regions, both of which include transmembrane domains. In contrast to Sec61p function in the absence of TM2 or TM3, the other 13 deletion alleles examined in this study were lethal, indicating that the majority of Sec61p sequences are crucial for its structure and function. Each deletion removed part of hydrophilic loops neighboring each transmembrane domain (Fig. 1B), and thus it is possible that the phenotypes observed are due to loss of the transmembrane regions and/or the perturbation of extramembranous loops. The retention of Sec61p function in the absence of either TM2 or TM3 and the loss of function upon deleting both are surprising in terms of Sec61p transmembrane topology. The deletion of a single internal TM domain must be accompanied by some topological rearrangement of the mutant protein compared with the native structure, although regions of the protein critical for function must presumably maintain the correct topology in assembled Sec61 complexes. The minimal structural alteration is predicted to be the topological frustration of TM3 in the absence of TM2 and vice versa as suggested by the protease sensitivity data (Fig. 7B). In the simplest case, the frustrated domain might adopt a membrane-embedded conformation that fails to span the bilayer, thus maintaining the correct topology for TM1, TM4, and the rest of the molecule (Fig. 7C). The phenomenon of topological frustration of individual transmembrane domains has some precedent in eukaryotic ER membranes (45).
The finding that TM2 is not essential for function is particularly surprising given that this domain contains the only residues absolutely conserved throughout the Sec61/SecY pro-tein family (39,40,46). It seems unlikely that the conservation of these residues is the result of a conserved intramolecular interaction, because there is no correspondingly conserved region elsewhere within the Sec61p/SecY sequences. Homologues of Sss1p, a component of the Sec61 complex have also been identified across phyla (13), but again there is no obvious sequence conservation within the transmembrane domain(s) that might account for an interaction with the conserved residues of Sec61p TM2. However, the finding that the Sec61 complex can form oligomeric structures (4, 5) raises the possibility that the conserved residues are required for the optimal intermolecular interaction between Sec61 complexes. Alternatively, the conserved region within TM2 may interact directly with precursor and may be required for the initiation and/or propagation of the translocation reaction.
Post-translational translocation has been shown to occur in two distinct phases. The first of these involves the signal sequence-dependent binding of precursor to the cytosolic face of the translocon (47,48). During this initial phase, the signal sequence of pp␣F can be cross-linked to the TM2 and TM7 regions of Sec61p, suggesting that these domains may be involved in signal sequence recognition (24). The deletion of either TM2 or TM3 resulted in a dramatic defect in the translocation of both pp␣F and ppCPY. These results would be consistent with a role for both TM2 and TM3 in signal sequence recognition. In the mammalian systems it has been reported that the Sec61p homologue, Sec61␣, is required for a signal recognition event involved in the SRP-dependent targeting of preprolactin to the translocon (49,50). However the TM2 deletion had no effect on the translocation of an SRP-dependent precursor, DPAPB, suggesting that it is not involved in the recognition of such precursors. This apparent contradiction might be explained by the nature of the signal sequence. In yeast, the signal sequences of SRP-dependent precursors are significantly more hydrophobic that those of precursors such as pp␣F and ppCPY that are translocated post-translationally (42). We have previously noted that TM2 contains several polar residues, and we now propose that these may be required to accommodate the less hydrophobic signal sequences targeted by the post-translational mechanism.
The effect of the sec61-⌬3 deletion is more complex with mutant cells exhibiting a severe defect in post-translational translocation (of ppCPY and pp␣F) but also significant defects both in the cotranslational translocation of DPAPB and in the dislocation of a malfolded form of p1CPY. From this it might appear that Sec61-⌬3p is generally defective in channel formation and thus deficient in all aspects of protein transit through the bilayer. Yet the extent of the defect observed for DPAPB translocation remains slight, a finding that is consistent with the absence of any genetic interaction between sec61-⌬3 and the sec65-1 mutation that is defective in a component of SRP. This latter observation is in striking contrast to the synthetic lethality observed between either sec61-⌬2 or sec61-⌬3 with the sec62-1 mutation, suggesting that the roles of both TM2 and TM3 regions are intimately linked to that of the Sec63 complex (comprising Sec62p/Sec63p/Sec71p/Sec72p). The interaction between Sec61p and this complex appears to be at least as stable in the sec61-⌬2 mutant as wild type but less stable in the sec61-⌬3 mutant (Fig. 7A). An inability to efficiently recruit the Sec63 complex offers an explanation of the post-translational translocation defect seen in the sec61-⌬3 mutant and also the observed defect in dislocation given that Sec63p has also been implicated in the degradation of CPY * (21).
It is interesting to speculate on the lethal nature of the combined TM2/TM3 deletion. Either this mutant is topologically irretrievable or the observed synthetic phenotype may arise either because the two deletions have incremental effects on the same process or because they each perturb different aspects of Sec61p function. The TM2 deletion appeared to be solely defective in post-translational translocation, and yet the TM3 deletion was at least as defective in this process. However, one must note that the observed defects are kinetic in nature, with at least some of the post-translational precursors being translocated in time (Fig. 6). Therefore, it would appear that the simplest explanation would be that TM2 and TM3 play distinct roles in the post-translational mechanism, e.g. signal sequence recognition (TM2) and Sec63 complex recruitment or channel formation (TM3). The loss of either one of these activities results in a severe kinetic defect, but their combined loss is then lethal.
The observed induction of the UPR in sec61-⌬2 and sec61-⌬3 mutant cells, coupled with their sensitivities to dithiothreitol and tunicamycin, suggests some defect in protein folding and/or quality control within the ER lumen. The mutant Sec61 proteins might themselves be recognized as being misfolded, leading directly to an induction of the UPR, but this would appear unlikely given that their expression in wild type cells does not produce a similar response. The UPR induction must therefore relate to the functional expression of Sec61 deletion derivatives in cells lacking wild type Sec61p. This effect might relate to some defect in protein folding during the translocation reaction, perhaps because of the perturbation of some direct role of Sec61p in providing a folding environment or because the Sec61 deletion derivatives fail to efficiently recruit folding chaperones (e.g. BiP/Kar2p) or modifying enzymes such as oligosaccharyltransferase to the translocon. However, the strong induction of the UPR in the sec61-⌬3 mutant (Fig. 5B) is perhaps most likely due to the accumulation of misfolded proteins in the ER as a result of the dislocation defect apparent in this mutant (Fig. 6).
This study has indicated a role for the TM2 region (residues 73-107) of Sec61p that is specific to the post-translational translocation pathway and also identified roles for TM3 (residues 108 -143) in both translocation and in ER dislocation. Two novel conditional sec61 mutants defective in both ER translocation and dislocation with corresponding mutations in TM3 and TM4 have been reported (20). Taken together these data suggest that the TM2-TM4 region of Sec61p may interact with a component(s) required for post-translational translocation, whereas the TM3-TM4 region has an additional role in dislocation. This distinction would be consistent with a model in which TM2 is specifically involved in the interaction of Sec61p with the post-translational class of signal sequences, because such an activity would be required neither for translocation of SRP-dependent precursors nor for the dislocation of signal sequence cleaved polypeptides from within the ER lumen.