Inter-Species Complementation of the Translocon Beta Subunit Requires Only Its Transmembrane Domain

In eukaryotes, proteins enter the secretory pathway through the translocon pore of the endoplasmic reticulum. This protein translocation channel is composed of three major subunits, called Sec61α, β and γ in mammals. Unlike the other subunits, the β subunit is dispensable for translocation and cell viability in all organisms studied. Intriguingly, the knockout of the Sec61β encoding genes results in different phenotypes in different species. Nevertheless, the β subunit shows a high level of sequence homology across species, suggesting the conservation of a biological function that remains ill-defined. To address its cellular roles, we characterized the homolog of Sec61β in the fission yeast Schizosaccharomyces pombe (Sbh1p). Here, we show that the knockout of sbh1 + results in severe cold sensitivity, increased sensitivity to cell-wall stress, and reduced protein secretion at 23°C. Sec61β homologs from Saccharomyces cerevisiae and human complement the knockout of sbh1 + in S. pombe. As in S. cerevisiae, the transmembrane domain (TMD) of S. pombe Sec61β is sufficient to complement the phenotypes resulting from the knockout of the entire encoding gene. Remarkably, the TMD of Sec61β from S. cerevisiae and human also complement the gene knockouts in both yeasts. Together, these observations indicate that the TMD of Sec61β exerts a cellular function that is conserved across species.


Introduction
Protein secretion is an essential process in all organisms. In eukaryotes, entry of proteins into the secretory pathway requires passage across the endoplasmic reticulum (ER) lipid bilayer. Translocation of proteins into the ER is accomplished through a protein conduction channel (PCC) [1,2]. The PCC is composed of a highly conserved heterotrimeric core called the Sec61abc complex in eukaryotes, and the SecYEG complex in bacteria [3,4,5]. The PCC associates with several soluble proteins to form the translocon, the ER translocation machinery [6,7]. The translocon functions in co-translational translocation of nascent proteins, as well as in the retro-translocation of misfolded polypeptides for degradation in the cytosol by the ER-associated degradation (ERAD) mechanism [3,8,9,10,11]. In addition, the association of the translocon with the tetrameric Sec62/63 complex and the BiP molecular chaperone forms the Sec complex, which is involved in post-translational translocation [12,13,14]. The genomes of all organisms sequenced to date encode at least one homolog of each Sec61abc/SecYEG subunits, suggesting that these proteins have a universal cellular role [15].
So far, the functions of the a and c subunits of the PCC are the best understood. The a subunit consists of ten transmembrane segments assembled in an hourglass shape, forming the pore of the channel [16,17] and acting as the main ribosome receptor [18,19]. The c subunit is a single-span membrane protein that binds and stabilizes the translocon pore by linking the two halves of the a subunit [16,20]. By contrast, the precise function of the b subunit, a 10 kD C-tail-anchored transmembrane protein, remains elusive.
Interestingly, Sec61b and its prokaryotic counterpart SecG are the only translocon subunits that are not essential for cell viability in yeast and bacteria [21,22,23,24]. Moreover, mammalian Sec61b is not required for protein translocation in vitro, but has a kinetic effect on co-translational translocation [25]. Likewise, bacterial SecG is dispensable for translocation in reconstituted membrane vesicles, but has a stimulatory effect in this process [26]. In S. cerevisiae, the double knockout of the genes encoding the two paralogs of the translocon b subunits (SBH1 and SBH2) has only a moderate effect on translocation of reporter proteins [22,24,27,28]. Regarding protein secretion, the relevance of the function of Sec61b in co-and posttranslational translocation remains unclear as the knockout of the translocon b subunits reduces a-amylase secretion in S. cerevisiae, but has no effect in Yarrowia lipolytica [23,27].
Based on the phenotypes resulting from its knockout, several cellular roles for the translocon b subunit have been proposed. In prokaryotes, SecG facilitates the membrane insertion/de-insertion cycle of the SecA ATPase, the bacterial translocation motor [29,30]. In S. cerevisiae, Sbh1p was proposed to act as the guanine nucleotide exchange factor for the signal recognition particle receptor (SR) [31], whereas Sbh2p was suggested to mediate recognition of vacant translocons by functional interactions with the SR [28]. Sbh1p also rescues several exocyst mutants in yeasts and coimmunoprecipitates with Sec10p, Sec6p and Sec8p in mammalian cells, three components of the exocyst complex, suggesting a role in vesicular transport [24,32,33]. The translocon b subunit was also shown to be recruited to the ER-derived quality control compartment in ERAD [34], and to have an accessory function in the unfolded protein response in different organisms [23,35,36,37]. The diversity of functions proposed for the translocon b subunit suggests that different homologs of this protein have species-specific roles.
Recently, Feng et al. [27] demonstrated that the transmembrane domain (TMD) of S. cerevisiae Sbh1p or Sbh2p are necessary and sufficient to suppress the phenotypes of sbh1Dsbh2D cells, including heat sensitivity at 38uC, and co-translational translocation defects. This was further confirmed in another study, where N-terminal truncation mutants of Sbh2p restored normal integration of DPAPB and translocation of CPY in the sbh1Dsbh2D strain [28]. Furthermore, the TMD of Sbh1p is sufficient for co-immunoprecipitation with the translocon a and c subunits, indicating correct integration and interaction at the Sec61 translocon site [27]. The TMD of Sbh1p also interacts with Rtn1p, a protein involved in structuring the cortical ER, in absence of other translocon subunits, suggesting a role for the b subunit outside the translocation apparatus [27]. Taken together, these results suggest that the TMD is the critical region responsible for the undefined function of Sec61b.
Here, we analyzed the functional conservation of the translocon b subunit and its TMD. We show that the translocon b subunit is not essential in fission yeast at 30uC. However, it is required for growth at low temperature, and its deletion causes SDS sensitivity and reduced protein secretion at 23uC. Overexpression of sbh1 + also diminishes protein secretion of a model protein, increases sensitivity to cell-wall stress, and causes morphological abnormalities. The phenotypes associated with sbh1 + deletion are suppressed by the 26 amino acids of the TMD. Remarkably, the TMDs from both S. cerevisiae homologs of the translocon b subunit and that from human also complement the deletion phenotypes in fission and budding yeast. These results demonstrate that the TMD function of Sec61b is conserved and sufficient to exert most biological roles of the full-length protein. These observations pose Sec61b as a model to study alternative functions of TMDs apart from the anchoring to the ER membrane.

S. pombe encodes a single Sec61b homolog
Based on Wu-Blast2 analyses of the fission yeast proteome [38], S. pombe encodes a single Sec61b homolog at ORF SPBC2G2.03c (sbh1 + ), hereafter designed as SP_Sec61b for the sake of clarity. The deduced primary sequence of the protein is presented in Figure 1. As expected for a translocon subunit, the predicted protein displays a high level of sequence conservation with both paralogs of S. cerevisiae (SC) and that of human (HS). SP_Sec61b is 35% identical (66% similar) to SC_Sec61b1 (Sbh1p), 43% identical (65% similar) to SC_Sec61b2 (Sbh2p), and 29% identical (58% similar) to HS_Sec61b (human Sec61b). Multiple sequence alignment using Clustal W2 [39] revealed that SP_Sec61b is evolutionary closer to HS_Sec61b than its S. cerevisiae counterparts (not shown), thus validating our choice of S. pombe as a model organism to study the translocon b subunit of higher eukaryotes. Analyses of SP_Sec61b using various ExPASy proteomic tools [40] allowed the identification of a single domain, constituted of a continuous stretch of 26 amino acids near the C-terminus. This hydrophobic tail is predicted to form an alpha helix crossing the lipid bilayer once, with the N-terminus of the protein exposed to the cytoplasm. Thus, as in S. cerevisiae and human, SP_Sec61b is predicted to be a tail-anchored protein. Interestingly, the TMDs of Sec61b homologs are much more conserved than their N-terminal domains, 53-61% identical from yeast to human as compared to only 18-41% for the cytosolic part. In comparison, the tail anchor of the translocon c subunit is only 40-50% identical from yeast to human, a level of sequence homology comparable to the entire protein, which is 44-48% identical. This conservation suggests a critical function for the TMD region of Sec61b homologs, not solely explained by its role as a membrane anchor, especially as almost any stretch of hydrophobic amino acids can perform this function [41].

Overexpression of SP_Sec61b induces morphological changes and cell-wall sensitivity
To explore its cellular roles in S. pombe, we overexpressed the translocon b subunit expecting to exacerbate its function and induce a revealing phenotype. The ORF of SP_Sec61b was cloned under the control of the strong thiamine-repressible nmt promoter in a pREP2 vector [42]. As shown in Figure 2A (upper panel), overexpression of SP_Sec61b resulted in a elongated phenotype with a single nucleus per cell (see DAPI staining, middle panel). This phenotype was specific to the translocon b subunit because the overexpression of S. pombe c subunit (SP_Sec61c) had no effect on cell morphology. Moreover, the elongated phenotype was reverted by addition of thiamine, which represses the promoter; and the phenotype was not observed when expression was driven by a weaker nmt* promoter (not shown). The elongated phenotype caused by overproduction of SP_Sec61b was further confirmed by global population analysis using flow cytometry (FACS). As shown in Figure 2A (lower panel), cells overexpressing SP_Sec61b had an average size (FSC-H) value 1.9 higher than WT. The wider size distribution can be explained by the presence of cells with various altered morphologies, which were either more granulated, had bigger vesicles or displayed an aberrant number of septa in ,10% of cases (not shown). To assess if the elongated phenotype is associated with membrane and/or cell-wall defects, we performed an SDS resistance assay at various temperatures ( Figure 2B). SDS is a strong detergent that disturbs the cytoplasmic membrane when the cell-wall is compromised, thus resulting in a dead-cell halo that can be measured [43]. In these conditions, cells overexpressing SP_Sec61b were up to 1.8 more sensitive to SDS as compared to the vector control at 30uC. In comparison, cells overexpressing SP_Sec61c were only 1.3 times more sensitive at this temperature. At 23uC, the strains overproducing the b or c subunits were both 1.4 times more sensitive than the vector control. As addition of 1.2 M sorbitol, an osmotic stabilizer [44], did not rescue the SDS Three exponential cultures of WT and Dsbh1 cells bearing the indicated plasmids were spread on top-agar MM+AL, following which a drop of 10 mL of SDS 10% was added on the center of each plate. After 3-5 days of incubation at the indicated temperatures, the area of the dead-cell halo was measured and reported to the WT strain at 30uC (100%). Data shown are mean6standard deviation of three or more independent experiments. *** indicates p,0.001 for Student's t-test. (C) Growth of WT and Dsbh1 strains devoid or overexpressing SP_Sec61b or SP_Sec61c at different temperatures. Exponential cultures were serially-diluted (10 21 -10 24 ) and spotted on EMM+AL plates. Growth was monitored for 7 days at the indicated temperatures. Results are representative of three or more independent experiments. doi:10.1371/journal.pone.0003880.g002 sensitivity nor the elongated cell size (not shown), it is unlikely that these phenotypes are only due to an incorrect cell-wall formation. These results demonstrate that the overexpression of SP_Sec61b alters cellular morphology, accompanied by a higher sensitivity to membrane stress, suggesting a potential role of this protein in the coordination between cell-membrane growth and the cell-cycle in fission yeast.

Cells depleted of SP_Sec61b are cold-sensitive
To investigate whether SP_Sec61b is essential for viability, a sbh1 + /Dsbh1::kan R diploid strain was sporulated to analyze its meiotic progeny. All spores were viable on YE medium, with a 2:2 segregation of the kan R marker. Deletion of sbh1 + was confirmed by Northern blotting, Southern blotting and PCR, demonstrating that SP_Sec61b is not essential for viability (not shown). In contrast to the overexpression of SP_Sec61b, the deletion of sbh1 + did not induce morphological changes ( Figure 2A). As it was shown in S. cerevisiae that deletion of the genes encoding the two translocon b subunits (SBH1 and SBH2) results in heat sensitivity at 38uC [22,24], we measured the growth of S. pombe Dsbh1 cells at different temperatures. In contrast to S. cerevisiae, Dsbh1 cells from S. pombe were not heat-sensitive (not shown), but the deletion resulted in severe cold sensitivity at 23uC ( Figure 2C). Normal growth was not restored by replacing the cells at 30uC after incubation at 23uC for 7 days, indicating that growth is permanently arrested or that most Dsbh1 cells are dead (not shown). In addition, Dsbh1 cells were 2.2 times more sensitive to SDS than WT cells at 23uC, while sensitivity was similar at 30uC ( Figure 2B). Transformation of sbh1 + on a plasmid under the medium strength nmt* promoter was sufficient to rescue to WT levels the cold-and SDS-sensitivity ( Figure 2B-C). Globally, these results demonstrate that while SP_Sec61b is not essential for cell viability at 30uC, it is critical at low temperature.

The levels of SP_Sec61b modulate secretion
Because the knockout of Sec61b was reported to lower protein secretion in S. cerevisiae, but not in Y. lipolytica [23,27], we wanted to assess the importance of SP_Sec61b in protein secretion. To this end, we used a previously characterized genomic cassette expressing the reporter protein cellulase from Aspergillus aculeatus [45]. The activity of secreted cellulase was measured during 7 days at different temperatures to evaluate the rate of secretion of each strain. As shown in Figure 3B, the secretion rate was not reduced in Dsbh1 cells at 30uC, but was reduced to 82% of the control strain at 23uC. Thus, the kinetic effect of the translocon b subunit on secretion is only observed in conditions where secretion is already considerably diminished, such as at low temperature (secretion rate 2.6-fold lower than at 30uC, Figure 3A).
In S. cerevisiae, overexpression of the endogenous translocon b subunit or the Kluyveromyces lactis homolog increases the secretion efficiency of a-amylase [46]. By contrast, in fission yeast, we show that overproduction of the translocon b subunit reduces the secretion rate of cellulase ( Figure 3B). Again, this effect was more pronounced at 23uC (72% of the WT value) than at 30uC (91% of the WT value). This reduction is not an unspecific effect caused by saturation of the ER, as overexpression of the translocon c subunit did not alter the secretion rate. These results support the idea that even though SP_Sec61b is not essential for growth at normal temperature, a correct SP_Sec61b level is important for efficient protein secretion, especially at low temperature.
Sec61b is functionally conserved between species As described above, the knockout of sbh1 + results in cold-sensitive phenotypes in S. pombe. Interestingly, in budding yeast, the double deletion of both paralogs of Sec61b, hereafter designated as SC_Sec61b1 and SC_Sec61b2, results in heat sensitivity instead of cold sensitivity [22,24]. Nevertheless, we reasoned that if the function of Sec61b is conserved there should be cross-species complementation in spite of the opposite temperature phenotypes found in S. pombe versus S. cerevisiae. As shown in Figure 4A, both Sec61b paralogs of S. cerevisiae complemented the cold sensitivity of the S. pombe Dsbh1 strain as efficiently as the endogenous translocon b subunit. Moreover, SC_Sec61b1 and SC_Sec61b2 also fully reversed the SDS sensitivity of the S. pombe Dsbh1 strain ( Figure 4C). Similarly, SP_Sec61b complemented the heat sensitivity of the sbh1Dsbh2D S. cerevisiae strain ( Figure 4B). This effect was specific to the translocon b subunit, as the c subunit homologs of S. pombe (SP_Sec61c) and S. cerevisiae (SC_Sec61c) were unable to complement the temperaturedependent phenotypes in either yeast.
To investigate the possibility of a functional conservation with higher eukaryotes, we transformed the human homolog of Sec61b (HS_Sec61b) in the Sec61b knockout strain of both S. pombe and S. cerevisiae. Strikingly, HS_Sec61b also complemented the cold sensitivity and SDS sensitivity of the S. pombe Dsbh1 strain, as well as the heat sensitivity of the S. cerevisiae sbh1Dsbh2D strain ( Figure 4, line 5). Hence, despite distinct phenotypes resulting from the deletion of translocon b subunit in two different yeast species, the function of Sec61b appears to be conserved from yeast to human.

The transmembrane domain of Sec61b is sufficient for heterologous complementation
Next, we wanted to define the domain of Sec61b responsible for the inter-species complementation. Bioinformatic analyses revealed that the most conserved region of Sec61b in all species studied is mapped to its TMD [47]. Moreover, the TMD of SC_Sec61b1 is sufficient to rescue the heat sensitivity of sbh1Dsbh2D cells, and to associate with the translocon a and c subunits [27]. Therefore, we tested if the TMD of different Sec61b homologs could rescue the temperature-sensitive phenotypes in S. pombe and S. cerevisiae knockout strains. We found that the 26 amino acids of the tail anchor of SP_Sec61b are sufficient to complement the cold sensitivity ( Figure 5A), SDS sensitivity ( Figure 5C) and secretion defects ( Figure 3B) of the S. pombe Dsbh1 strain to levels equivalent to the full-length protein. The TMDs of the other Sec61b homologs tested were also able to fully complement the defects of the Dsbh1 strain, indicating that the fundamental function of the TMD is conserved between species ( Figure 5A-C). Likewise, all Sec61b TMDs complemented the heat sensitivity of the sbh1Dsbh2D S. cerevisiae strain ( Figure 5B). Again, complementation was specific to the TMD of Sec61b, as the TMD from Sec61c ( Figure 5, line 5) and calnexin (not shown) failed to suppress the phenotypes of the sbh1Dsbh2D S. cerevisiae strain. Collectively, these results indicate that the tail anchor of Sec61b is sufficient to complement the knockout of the entire protein, and that this can be indistinctively accomplished by homologous transmembrane domains ranging from yeast to human. This further confirms that the diverse phenotypes associated with the deletion of Sec61b in distinct species are not due to divergence in the amino acid sequences, but rather they reflect different downstream effects of a conserved Sec61b function supported by the TMD.

Discussion
We have characterized for the first time the Sec61b homolog from fission yeast. We describe S. pombe-specific phenotypes resulting from the overproduction or the lack of this protein, revealing the involvement of SP_Sec61b in cellular morphology, stress resistance and protein secretion. These S. pombe phenotypes are distinct from those observed in S. cerevisiae [22,23,24,27]. Nevertheless, the Sec61b homologs from S. pombe, S. cerevisiae and human complement the phenotypes of knockout strains of both fission yeast and budding yeast, to a level comparable to that of the endogenous proteins. Remarkably, the transmembrane domains of Sec61b homologs are sufficient for heterologous complementation, demonstrating that the TMD region supports a fundamental biological function that is conserved across species. Sec61b is not essential for survival in microorganisms, but is required for development in Drosophila [22,23,24,48]. In S. pombe, we show that Sec61b is dispensable for growth at optimal temperature, but becomes critical at low temperature (Figure 2-3). Cold sensitivity was also observed for the Escherichia coli SecG deletion [21], although it was dependent on a second mutation in glpR, involved in glycerol phosphate regulation [49]. In bacteria, the translocation process itself is cold-sensitive, presumably because membrane insertion of the SecA ATPase motor, which is mediated by SecG, is reduced at low temperature [30,50]. In eukaryotes, a role for Sec61b in cold resistance was proposed in the Gymnadenia conopsea psychrophile plant [51]. Thus, the heatsensitive phenotype observed in sbh1Dsbh2D cells of S. cerevisiae [22,24] does not reflect a universal phenotype. A possibility is that the differences between the two organisms reflect the prevalence of co-vs post-translational translocation processes in each yeast.
Interestingly, the SDS sensitivity of Dsbh1 cells in S. pombe contrasts with the SDS resistance of Dsbh1 cells observed in Y. lipolytica [23]. In addition, in Y. lipolytica, cells genetically depleted of Sec61b form colonies displaying morphological changes, while in S. pombe, only the overexpression of Sec61b affected cellular morphology and not colony formation. As morphology and SDS resistance are linked to cell-wall biosynthesis, these opposite effects could be due to differences in the composition of the cell-wall of these two yeasts [52]. Since we observed that the translocon b subunit levels also affect cell size, it remains possible that SP_Sec61b is involved in the coordination of cell-membrane growth in fission yeast.
Regarding the effect of Sec61b on secretion, it appears correlated with the efficiency of this process in each yeast. As S. cerevisiae is somewhat inefficient for secretion [53], the impact of Sec61b deletion is more apparent in this yeast (47% diminution) [27], while in Y. lipolytica, which is more competent for secretion, no difference are observed [23]. In fission yeast, the importance of Sec61b is revealed at 23uC, when the secretion rate of the WT strain is already reduced by 62% as compared to 30uC (Figure 3). The different phenotypes observed for the Sec61b-knockout in S. pombe as compared to S. cerevisiae and Y. lipolytica suggested that different homologs could have distinct functions. However, the lack of the translocon b subunit could affect more severely the secretion of certain proteins, thus providing an alternative explanation for the diverse, and sometimes contrasting, phenotypes associated with deletion or overexpression of Sec61b among species.
Importantly, we show that despite these phenotypic differences, SP_Sec61b is able to complement sbh1Dsbh2D in budding yeast. Similarly, both SC_Sec61b paralogs complement the cold and SDS sensitivity in S. pombe. This is in accordance with previous reports that Sec61b homologs from Yarrowia lipolytica, Kluyveromyces lactis, Antonospora locustae and Encephalitozoon cuniculi rescue the heat sensitivity of S. cerevisiae, indicating that they are functional homologs [23,46,54]. Here we demonstrate that this functional complementation is possible even with higher eukaryotes, as the human homolog of Sec61b also complements the knockout in S. pombe and S. cerevisiae. These results further suggest that the distinct phenotypes associated with the Sec61b deletion are caused by downstream effects resulting from loss of a common function of the translocon b subunit that can be performed by any homolog.
Phylogenetic studies on Sec61b/Sbh1p/SecG revealed that the consensus region for this protein is located at the hydrophobic peak corresponding to its TMD, while the N-terminal domain is more divergent and disordered [15,16]. In many archaea, the Nterminal portion of Sec61b homologs is almost twice smaller than their mammalian counterparts. In some eukaryotes, such as parasitic microsporidia, the N-terminal region of Sec61b is practically as short as in archaea, and the protein still complements heat sensitivity of S. cerevisiae sbh1Dsbh2D cells [54]. Domain mapping of SC_ Sec61b showed that the N-terminal part is dispensable for normal rates of co-translational translocation, heat resistance, association with other translocon subunits and interaction with the SR [27,28]. This suggests that the TMD of Sec61b is the critical region for protein function, while the larger N-terminal cytosolic part plays secondary roles. In this paper, we show that the fundamental function of the TMD of Sec61b is conserved among distant species. Indeed, the S. pombe, S. cerevisiae and human TMD homologs complemented various defects caused by Sec61b deletion in fission and budding yeasts, to levels comparable to the entire protein. This property is inherent to the TMD of Sec61b, since other transmembrane sequences did not suppress the phenotypes associated with the Sec61b-knockout in fission and budding yeast.
To date, the only precise function ascribed to the TMD of Sec61b is its role as a tail anchor and ER localization signal. Disruption of this domain results in accumulation of Sec61b in large cytoplasmic puncta [55,56], while N-terminal fragments of Sec61b devoid of the TMD are unable to complement any known function of the protein in S. cerevisiae [27]. Insertion of Sec61b into the ER is not spontaneous, but requires recognition of the TMD by a post-translational targeting machinery, TRC40/Asna1 in mammals or the GET complex in yeast [55,56]. Thus, it is conceivable that specific residues of the TMD are essential to recruit this complex to the ER membrane. Yet, other tailanchored proteins, including the essential translocon c subunit, display lower conservation of the TMD than the b subunit, suggesting a key function for this domain in Sec61b. Indeed, apart from its adaptor role in the translocon, the TMD of SC_Sec61b1 was also found in a distinct complex including Rtn1p, an exocystassociated protein involved in ER reticulation [57], suggesting additional localizations of Sec61b in the ER [27]. Nevertheless, the TMD of SC_Sec61b2, which is part of the second translocon complex in budding yeast (the Ssh1 complex) [22,58,59], also complements cold-and SDS-sensitivity in S. pombe ( Figure 5). This suggests a redundant and conserved role for this domain. Considering the small size of these hydrophobic sequences, it is unlikely that it exerts a catalytic function, but interactions with conserved transmembrane proteins could mediate the central TMD function. As Sec61b was also shown in crosslinking experiments to form dimers [25], a role of the TMD in translocon oligomerization cannot be excluded [60]. The demonstration of a scaffolding or a recruiting function for the TMD of Sec61b awaits further experimentation.
In sum, our results demonstrate a conserved biological role for the TMD of Sec61b. Our observations support the notion that single-span TMDs have other functions than anchoring in the lipid bilayer. Indeed, many TMDs have been implicated in protein complex assembly, oligomerization, helix packing, transmembrane interactions and signal transduction [61]. Sec61b from S. pombe appears as an interesting model for elucidation of the cellular roles of a highly conserved TMD present in every living kingdom.

Materials and Methods
Strains and media S. pombe strains were cultured in Edinburgh Minimal Medium (EMM) supplemented with required nutrients [62]. S. cerevisiae strains were grown in synthetic defined media (SD) lacking appropriate amino acids [63]. All strains were grown aerobically at 30uC unless stated otherwise. G418 was used at a concentration of 150 mg/mL. Yeast strains used in this study are listed in Table 1.

Plasmid constructions and transformation
Plasmids used in this study are described in Table 2. All oligonucleotide sequences are available upon request. The coding sequences of sbh1 + and sss1 + were amplified by PCR from a S. pombe cDNA library. Forward primers contained a NdeI restriction site followed by a sequence encoding the V5 epitope tag. Reverse primers contained a BamHI restriction site. The entire ORFs were cloned into the pREP42 and pREP2 vectors [42], at the NdeI and BamHI sites. pREP42 contains a medium strength thiaminerepressible nmt* promoter, while pREP2 contains a strong nmt promoter for overexpression [64]. As previously described in S. cerevisiae [27], addition of a N-terminal tag on Sec61b had no effect on overexpression phenotypes nor complementation phenotypes (not shown). Amplification of the SBH1 and SBH2 genes was similarly achieved except that S. cerevisiae genomic DNA was used and no V5 tag was added. Amplification of SEC61B was carried out using a human cDNA bank as template. No V5 tag was added, and a BglII site was used instead of BamHI. Cloning of SBH1, SBH2 and SEC61B in pREP42 was performed as for sbh1 + . For TMD constructs, the sequences encoding the 26 amino acids of the conserved transmembrane domains of each protein, as mapped by HMMTOP [65], were cloned (see Figure 1 for amino-acid sequences). All TMD sequences were amplified to encode a V5 tag on the N-terminal side of the protein, and inserted into pREP42 using NdeI and BamHI restriction sites. For cloning in p415, amplification of sbh1 + , SBH1, SBH2, SEC61B and TMD constructs was carried out as before but with primers containing BamHI and XhoI restriction sites instead of NdeI and BamHI, respectively. SSS1 and SSS1_TMD were amplified from S. cerevisiae genomic DNA, adding a c-MYC tag instead of a V5 tag. The amplified sequences were inserted into the BamHI/XhoI restriction sites of p415. All constructs were verified using standard DNA sequencing methods. Transformations were done by the PEG (Polyethylene glycol)/lithium acetate procedure [66]. Plasmids derived from pREP42 were transformed into SP15039 and SP15074 (see below), while plasmids derived from pREP2 were transformed into SP556 and SP15073. Plasmids derived from p415 were transformed into SC3232. Transformations were confirmed by prototrophy and Western blotting.

Microscopy imaging
Differential interference contrast (DIC) Nomarski imaging and DAPI-staining microscopy were carried out as previously described [68]. Microscopy analyses were performed with a fluorescence inverted microscope Nikon TE2000U. Images were acquired using a motion-picture camera CCD coolSnapFX MH 12 bit, and treated with UIC MetamorphH software. Flow cytometry analyses (FACS) were done using a FACS Calibur device (Becton Dickinson Biosciences) on 10,000 cells after dilution of exponential cultures to OD 595 = 0.5 in PBS buffer (136 mM NaCl, 25 mM KCl, 12 mM NaHPO 4 , 18 mM KH 2 PO 4 pH 7.4). Phenotypic assays The sodium dodecyl sulfate (SDS) sensitivity measurements were carried out as described previously [69]. Briefly, exponentially growing cells were spread on EMM+AL top agar (0.7%), following which 10 mL of 10% SDS was added on a small 3 M paper at the center of each plate. After 3-5 days of incubation at 30uC or 23uC, area of the sensitivity halo was measured and reported on the WT value at 30uC. For temperature sensitivity assays, exponential cultures of each strain were diluted to OD 595 = 0.5, serially diluted (10 21 -10 24 ), spotted on solid media (EMM+AL for S. pombe or SD-L for S. cerevisiae), and incubated at the indicated temperatures (7 days for S. pombe or 5 days for S. cerevisiae). All results are representative of at least 3 independent experiments.

Cellulase secretion assays
Cellulase secretion [45] was quantified by activity halo measurements at 30uC or 23uC. An equivalent of 0.2 OD 595 of exponentially growing cells derived from strains SP15073 (WT) or SP15074 (Dsbh1) were spotted on EMM+ALH supplemented with 0.1% (w/v) AZCL-HE-Cellulose. The area of the blue halo created by secreted cellulase activity was measured each day during 7 days after time zero, which was adjusted for every strain as the moment when the halo exceeds area of the colony (all colonies were of equal size). Each measure is the mean of 3 plates. Secretion rate was calculated for each strain by averaging the slope of independent experiments and divided them by the control strain value (100%) at the corresponding temperature.