Subunits of the Translocon Interact with Components of the Oligosaccharyl Transferase Complex*

Following initiation of translocation across the membrane of the endoplasmic reticulum via the translocon, polypeptide chains are N-glycosylated by the oligosaccharyl transferase (OT) enzyme complex. Translocation and N-glycosylation are concurrent events and would be expected to require juxtaposition of the translocon and the OT complex. To determine whether any of the subunits of the OT complex and translocon mediate interactions between the two complexes, we initiated a systematic study in budding yeast using the split-ubiquitin approach. Interestingly, the OT subunit Stt3p was found to interact only with Sec61p, whereas another OT subunit, Ost4p, was found to interact with all three components of the translocon, Sec61p, Sbh1p, and Sss1p. The OT subunit Wbp1p was found to interact very strongly with Sec61p and Sbh1p and weakly with Sss1p. Other OT subunits, Ost1p, Ost2p, and Swp1p were found to interact with Sec61p and either Sbh1p or Sss1p. Ost3p exhibited a weak interaction with Sec61p and Sbh1p, whereas Ost5p and Ost6p interacted very weakly with Sec61p and failed to interact with Sbh1p or Sss1p. We were able to confirm these split-ubiquitin findings by a chemical cross-linking technique. Based on our findings using these two techniques, we conclude that the association of these two complexes is stabilized via multiple protein-protein contacts. Based on extrapolation of the structural parameters of the crystal structure of the prokaryotic Sec complex to the eukaryotic complex, we propose a working model to understand the organization of the translocon-OT supercomplex.

translational mode of translocation requires the Sec63 complex, which is made up of Sec62p, Sec63p, Sec71p, and Sec72p, in addition to the Sec61 complex and BiP (3). Genetic studies in yeast have identified a second set of translocon homologs (4), which has been shown to mediate co-translational translocation of a subset of proteins (5,6). Although the Sec61 complex is the protein-conducting channel, several other proteins associate with the translocon. The heterodimeric signal recognition particle receptor is required for the targeting of a ribosomesignal sequence-nascent chain-signal recognition particle complex to the ER membrane. Following translocation of the nascent chain, the signal peptidase (SP) cleaves the signal sequence as soon as the cleavage site is exposed on the lumenal side of the membrane. The SP is anchored to the Sec61 channel via the ␤ subunit (mammalian homolog of Sbh1p) (7). In the mammalian system, two other proteins associated with the translocon are the translocon-associated membrane protein and translocon-associated protein, the functions of which are not clearly understood (8 -12).
Following translocation of polypeptides across the ER membrane, certain -Asn-X-Ser/Thr-(-NX(S/T)-) sites are N-glycosylated. The oligosaccharyl transferase (OT) enzyme complex catalyzes the N-glycosylation reaction in the lumen of the ER (for a recent review on OT, see Ref. 13). The N-glycosylation reaction involves the transfer of a high mannose oligosaccharide chain from a dolichol pyrophosphate carrier to the Asn residue of the -NX(S/T)-acceptor site in the nascent polypeptide chain, with X being any amino acid except proline (14). Glycosylation of ribosome-bound nascent chains is observed when the glycosylation site is 65-75 residues away from the ribosomal P site (15). Thus glycosylatable -NX(S/T)-sites are glycosylated in the lumen of the ER as soon as they emerge from the translocon, well in advance of completion of the synthesis of the polypeptide chain (16). Translocation and N-glycosylation are believed to be concomitant events, because Nglycosylation sites can occur almost anywhere along the length of the newly synthesized polypeptide chain. It appears that coordination between the two processes is necessary; otherwise the glycosylatable sites might be buried during the process of protein folding. For this reason, OT is believed to be a part of the functional translocon, although it does not contribute to the formation of the structure of the channel (1).
In yeast the oligosaccharyl transferase enzyme complex is composed of 9 subunits (Ost1p, Ost2p, Ost3p, Ost4p, Ost5p, Ost6p, Stt3p, Wbp1p, and Swp1p), five of which (Ost1p, Ost2p, Stt3p, Wbp1p, and Swp1p) are encoded by essential genes. Homologs of all the OT subunits have been identified in higher eukaryotic organisms (17)(18)(19)(20)(21)(22). It is unclear why the catalysis of the N-glycosylation reaction requires 9 subunits. In fact, with the exception of Stt3p, the precise function(s) of the OT subunits is unknown. Besides participating in the oligosaccharyl group transfer, it is possible that some of the subunits may mediate complex formation among different OT subunits and with components of the translocon complex. It has been shown that when rough microsomes are solubilized with certain nonionic detergents, membrane-bound polysomes co-sediment as a complex containing both the translocon and OT (12). Ribophorins I (the mammalian homolog of yeast Ost1p) and II (the mammalian homolog of yeast Swp1p) have been shown to be chemically cross-linked to the 60 S ribosomal subunit (23), and polyclonal as well as antipeptide antibodies against ribophorins I and II inhibit the translocation of nascent polypeptide chains (24). Nascent chains undergoing translocation or integration at the ER membrane were found to cross-link to the STT3 subunit of the OT complex, suggesting a close proximity of the OT complex to the translocon (25).
Clearly, several lines of evidence suggest that the OT complex is closely associated with the translocon complex. However, other than one report indicating the interaction between Wbp1p and Sss1p (26), the issue of the molecular basis underlying the association of the translocon machinery and the OT complex has not been addressed. In this study we have carried out a systematic analysis of the organization of the translocon-OT supercomplex using chemical cross-linking and splitubiquitin analysis. Based on the protein-protein interactions identified, we provide a working model of the translocon-OT assemblage. This is the first detailed study providing evidence for the organization of the translocon-OT supercomplex, the existence of which has been hypothesized for at least a decade.

MATERIALS
Strains used in this study are listed in Table I. Standard yeast media and genetic techniques were used. T4 DNA ligase and shrimp alkaline phosphatase were obtained from Roche Diagnostics. Restriction enzymes and primers were obtained from Invitrogen (San Diego, CA). Dithiobis(succinimidyl propionate) (DSP) was obtained from Pierce (Rockford, IL). Anti-hemagglutinin (HA) monoclonal antibody HA.II was obtained from Covance (Richmond, CA). Anti-Myc polyclonal antisera and anti-HA polyclonal antisera was obtained from Santa Cruz Biotechnologies (Santa Cruz, CA). Horseradish peroxidase (HRP)-labeled anti-rabbit IgG raised in goat was obtained from Chemicon International (Temecula, CA). Antibodies to Sec61p raised in rabbit were generously provided by Dr. Tom Rapoport (Harvard Medical School). Protein G-agarose beads were purchased from Invitrogen. Triple Master DNA polymerase was obtained from Eppendorf-Brinkman (Westbury, NY) and Pfu polymerase was obtained from Stratagene (La Jolla, CA). Nitrocellulose membrane was obtained from Schleicher and Schull (Keene, NH).

Methods
Cross-linking Experiments-Yeast microsomes were prepared as described earlier except that dithiothreitol was completely eliminated from the protocol (27). The protocol used for cross-linking experiments has been described earlier (28). Following cross-linking, microsomes were solubilized in buffer containing 10 mM HEPES, 0.15 M NaCl, 5 mM MgCl 2 , 1% (v/v) Triton X-100, and 0.2% (w/v) SDS. The supernatant fraction was used for immunoprecipitation with monoclonal HA anti-body. The immune complex was immobilized using protein G-agarose beads. After incubation for 2 h with rocking at ambient temperature, the beads were recovered by centrifugation and washed 5 times with the co-immunoprecipitation buffer and once with Tris-buffered saline. Immunoprecipitated proteins were incubated for 30 min at 60°C in SDS-PAGE sample buffer containing 50 mM dithiothreitol, resolved on SDS-PAGE, and transferred to nitrocellulose membranes. The blots were probed with appropriate primary antibodies followed by secondary antibodies coupled to HRP. The HRP-conjugated secondary antibodies were visualized by the Lumiglo kit from KPL (Gaithersburg, MD). For immunoblotting, primary and HRP-coupled secondary antibodies were used at a dilution of 1:3000. The primary antibodies for immunoprecipitation were used at a dilution of 1:500.
Other Protocols-The protocol used for the ␤-galactosidase filter assay was carried out as described in the Clontech manual. The quantitative ␤-galactosidase liquid assay was performed as per the manufacturer's protocol (Pierce). The spotting assay was carried out as described earlier (29). Protein estimation was carried out using a BCA protein estimation kit from Pierce. The lithium acetate protocol was used for all yeast transformations (30). Escherichia coli strain DH5␣ was used for the recombinant DNA procedures. Luria-Bertani medium was used for bacterial growth. Standard recombinant DNA procedures were carried out as described earlier (31). Tagging of the genomic copy of SBH1 and SSS1 was performed using the protocol as described earlier (28).

Preparation of Plasmids
pRS315-SEC61-CUB-PLV-A PCR product corresponding to the promoter region and open reading frame of SEC61 was amplified using wild type yeast genomic DNA and 5Ј-AGATAAGCTTCGG-CGCGGCAGGCTCCAG and 5Ј-GATCCTGCAGCATCAAATCAGAA-AATCCTGGAAC as primers. The PCR products were digested with HindIII and PstI and ligated to pRS315-CUB-PLV digested with the same two enzymes to generate pRS315-SEC61-CUB-PLV.
pRS314-NUB(G)-SSS1-A PCR product corresponding to the promoter region and sequence corresponding to NUB(G) was amplified using pRS314-NUB(G)-ALG5 as template and 5Ј-GATCGAGCTC-CCGACATTTGGGCGCTATACG and 5Ј-CATGCCGCGGAGGGATC-CCTTCCTTGTCTTG as primers. The PCR product was digested with SstI and SstII, and ligated into pRS314 digested with the same two enzymes to generate pRS314-NUB(G). The open reading frame corresponding to the SSS1 gene was amplified using wild type yeast genomic DNA and 5Ј-CATGCCGCGGATGGCTAGAGCTAGTGAAAA-AGGTGAAGAG and 5Ј-GATCCTGCAGTTAAACAATAACGTATCT-GATTGGAATATG as primers. The PCR product was digested with SstII and PstI and ligated into pRS314-NUB(G) digested with the same two enzymes to generate pRS314-NUB(G)-SSS1.
pRS315-SBH1-CUB-PLV-A PCR product corresponding to the promoter region and open reading frame of SBH1 was amplified using wild type yeast genomic DNA and 5Ј-CATGAAGCTTCCGAGACCGAACCT-TCTTCC and 5Ј-CATGCTGCAGAAATAACTTACCGGCAAC as primers. The PCR products were digested with HindIII and PstI and ligated to pRS315-CUB-PLV digested with the same two enzymes to generate pRS315-SBH1-CUB-PLV.

Split-Ubiquitin Approach to Study the Interaction between
the Subunits of the Translocon and OT-Using the split-ubiquitin approach, which is now being widely used to study protein-protein interactions (26, 32-36), we analyzed the interac- tion of OT subunits with the translocon components. The principle of the split-ubiquitin technique have been described in detail elsewhere (37)(38)(39). Knowledge of the topology of the membrane proteins under study is a prerequisite in the utilization of the split-ubiquitin technique. The topology of the protein enables us to decide whether the protein should be fused at the C terminus or the N terminus with the N-terminal half of ubiquitin bearing a I13G mutation (NUB(G)) or the C-terminal (CUB) of ubiquitin. In addition, the Cub domain is fused with the protein A-LexA-VP16 (PLV) reporter (37). The topology of some of the OT subunits (Ost1p, Ost4p, and Wbp1p) has been known for sometime (40 -42), whereas that of others (Ost2p, Ost3p, Ost5p, Ost6p, Stt3p, and Swp1p) has been recently described (43). Based on this information, the appropriate OT constructs described earlier (43) were selected for this work. We designed the 2 constructs, pRS315-SEC61-CUB-PLV and pRS314-NUB(G)-SSS1 based on the topology of Sec61p and Sss1p described earlier (44,45). Although the topology of Sbh1p is predicted to be N cyt -C lum on the basis of the x-ray structure of the bacterial protein conducting channel (46), in yeast, the SBH1 construct bearing CUB-PLV at the C terminus (pRS315-SBH1-CUB-PLV) exhibited interactions with certain OT subunits (see below). All the constructs utilized in this study are listed in Table II and diagrammatically indicated in Fig. 1.
Interaction of OT Subunits with Sec61p-Except for Ost2p and Stt3p, all 7 subunits of OT possess a cytosol-oriented C terminus (40 -43). Therefore we used constructs in which NUB(G) was fused to the C terminus of Ost1p, Ost3p, Ost4p, Ost5p, Ost6p, Wbp1p, and Swp1p (see Fig. 1A) (OT-NUB(G)). In the case of Stt3p and Ost2p, we used the constructs in which NUB(G) was fused to the N terminus (see Fig. 1A) (NUB(G)-OT). We fused the C-terminal half of ubiquitin-PLV (CUB-PLV) to the C terminus of Sec61p known to be oriented toward the cytosol depicted in Fig. 1B. Following co-transformation of the SEC61-CUB-PLV and NUB(G)-OT (or OT-NUB(G)) in the LY40 strain, we determined the His phenotype of the transformants by spotting assay on -Leu-Trp-His plates. When we carried out ␤-galactosidase filter assay for the transformants (data not shown), we found that instant color development was observed for some interactions under study (for e.g. Sec61p and Stt3p), whereas for others, it took about 4 h for color development (e.g. Sec61p and Ost1p). Therefore we performed quantitative ␤-galactosidase assay to analyze the interactions between OT subunits and the translocon components. The results for the quantitative ␤-galactosidase and spotting assay are summarized in Fig. 2. Four OT subunits, Ost2p, Stt3p, Wbp1p, and Swp1p, encoded by essential genes, and Ost4p were found to interact strongly with Sec61p in the split-ubiquitin system. The other four OT subunits (Ost1p, Ost3p, Ost5p, and Ost6p) interacted to a much lesser degree with Sec61p as compared with the other 5 OT subunits, a finding that was in agreement with their growth phenotype on the -His plate.
Interaction of the OT Subunits with Sbh1p-To study this interaction, we used the same set of OT plasmids NUB(G)-OT (or OT-NUB(G)) ( Fig. 1A) that had been used in combination with the SEC61-CUB-PLV. CUB-PLV was fused to the C terminus of Sbh1p (depicted in Fig. 1B). Following co-transformation of the SBH1-CUB-PLV and NUB(G)-OT (or OT-NUB(G)) in the LY40 strain, we examined growth of the transformants on the -His plate and performed a quantitative ␤-galactosidase assay. Ost2p, Ost4p, and Wbp1p were found to interact with Sbh1p (Fig. 3), but much more weakly than with Sec61p. Ost3p was also found to interact with Sbh1p weakly.
Interaction of the OT Subunits with Sss1p-Because Sss1p is a single transmembrane protein with its N terminus oriented to the cytosol (45), we fused the NUB(G) at the N terminus of Sss1p (NUB(G)-SSS1), indicated in Fig. 1D. Ost2p and Stt3p possess cytosol-oriented N-terminal and lumen-oriented C-terminal domains (43); because CUB-PLV cannot be fused to the N terminus of the protein under study, we could not analyze the interaction of Sss1p with Ost2p and Stt3p using the splitubiquitin assay. The remaining 7 OT subunits bear a cytosoloriented C-terminal domain, and therefore OT-CUB-PLV constructs, indicated in Fig. 1C were used to study this interaction. Ost1p, Ost4p, and Swp1p were found to interact with Sss1p; Wbp1p was found to interact with Sss1p to a lesser extent (Fig.  4). Once again, the interaction of Ost4p, Wbp1p, and Swp1p with Sec61p was stronger than that with Sss1p (see scales in Figs. [2][3][4]. None of the other OT subunits were found to interact with Sss1p. Thus we identified specific interactions between the OT subunits and components of the translocon using the split-ubiquitin approach. However, a quantitative estimation of the ␤-galactosidase activity revealed a much stronger interaction between some of the OT subunits and Sec61p than that with Sss1p or Sbh1p (see length of bars in Fig. 2-4). It is unclear whether this difference in the level of ␤-galactosidase activity is a measure of the avidity of the interaction.
Chemical Cross-linking Approach to Analyze the Proximity of the OT Subunits and Translocon-The split-ubiquitin approach evaluates the potential interactions between OT subunits and translocon components (37). To confirm that various OT subunits are in fact, in close proximity to the translocon components, we incubated the microsomes with a cleavable chemical cross-linking reagent, DSP. This cross-linking agent bears homobifunctional amino-reactive groups with a 12-Å hydrophobic spacer arm between them. Following quenching of the unreacted DSP, the microsomes were solubilized in harsh denaturing buffer containing 1% Triton X-100 and 0.2% SDS. This buffer, which is known to disrupt physical, but not covalent interactions between proteins was used in all the crosslinking studies (28). This was followed by immunoprecipitation using monoclonal anti-HA antibody and protein G-agarose beads. SDS-PAGE sample buffer containing 50 mM dithiothreitol was used to elute bound proteins from the protein Gagarose beads. The samples were then analyzed by SDS-PAGE and Western blotting. To study which OT subunits are in close proximity to Sec61p or Sss1p, we used a battery of strains in which the genomic copy of Sss1p was Myc tagged and one OT subunit was HA tagged. We confirmed that Myc tagging of Sss1p does not alter the growth phenotype of the yeast cells (data not shown). Yeast lysates of these strains were electrophoresed and electroblotted. The blots were probed with polyclonal antibodies against the HA epitope, Sec61p, or Myc epitope (Fig. 5A, i, ii, and iii, respectively). All the OT subunits were easily visualized with the exception of Ost2p, which appears as a faint band. The Sec61 protein was found to migrate at its expected size (36 kDa) and the Sss1 protein following Myc tagging was found to migrate at 39 kDa. Using the microsomes of these strains, we carried out cross-linking experiments using DSP as the cross-linker and co-immunoprecipitation using monoclonal HA antibody. Microsomes not treated with the cross-linker were also analyzed (ϪDSP). First we confirmed that the OT subunit was immunoprecipitated (Fig. 5B, i). On probing a similar blot with anti-Sec61p antibody, it was observed that Sec61p cross-links to Ost1p, Ost2p, Ost4p, Stt3p, Wbp1p, and Swp1p (Fig. 5B, ii). It is interesting that although Ost1p interacted only very weakly with Sec61p in the splitubiquitin assay (Fig. 2), it was clearly seen to cross-link to Sec61p. Previously, Ost1p was shown to cross-link with all the OT subunits, and also to Alg1p (28).
Finally to analyze the interactions between OT subunits and Sbh1p, we Myc-tagged the C terminus of the genomic copy of Sbh1p in strains in which one OT subunit was HA-tagged. We confirmed that Myc tagging of Sbh1p did not alter the growth phenotype of the yeast cells (data not shown). Yeast lysates of the OT-HA, SBH1-Mycp strains were analyzed by SDS-PAGE and electroblotting, and the resulting blots were probed with polyclonal anti-HA (data not shown) and anti-Myc antibodies. Sbh1-Myc protein was found to migrate at 45 kDa as expected (Fig. 5A, iv). Cross-linking and coimmunoprecipitation experiments were performed as described above. We found that Ost2p, Ost4p, and Wbp1p cross-linked to Sbh1p. Ost3p was also found to cross-link Sbh1p to a lesser extent. This was in agreement with the results obtained by the split-ubiquitin approach. It was observed that Ost1p and to a lesser extent Stt3p also cross-linked to Sbh1p, interactions that we failed to detect by the split-ubiquitin assay. None of the other OT subunits cross-linked to Sbh1p. DISCUSSION In the biosynthesis of N-linked glycoproteins, glycosylation is temporally coupled to translocation. Several lines of evidence suggest the physical juxtaposition of the OT complex and the translocon (12, 15, 23-26). In Saccharomyces cerevisiae, nine FIG. 1. Constructs used to study the interaction of the OT subunits with the components of the translocon in the split-ubiquitin approach. Pictorial representation of the OT subunit constructs with NubG fused at the N or C terminus (A), Sec61p and Sbh1p construct with Cub-PLV at the C terminus (B), OT subunit constructs with Cub-PLV fused at the C terminus (C) and Sss1p with NubG fused at the N terminus (D). The constructs in A were used in combination with one of the constructs in B to study the interaction of OT subunits with Sec61p or Sbh1p. The constructs in C were used in combination with construct in D to analyze the interaction of the OT subunits with Sss1p. Ost2p and Stt3p possess a lumenally oriented C terminus and cytosol-oriented N terminus. Because Cub-PLV cannot be fused to the N terminus of the protein under study, the interaction of Stt3p and Ost2p with Sss1p could not be analyzed in the split-ubiquitin approach. N and C stand for the N and C terminus of the respective protein.
OT subunits have been identified using biochemical and genetic means. Whereas the precise function of none of the OT subunits is completely clear, it is hypothesized that one or more of them may aid in association with the translocon. To study the potential interactions between the translocon and the OT complex, we have primarily used the split-ubiquitin approach, which is a well accepted technique to study membrane proteinprotein interaction (37,39). Usually the protein interactions identified using the split-ubiquitin approach are believed to be physical in nature. However, in rare cases, the split-ubiquitin screen has incorrectly reported interaction of two membrane proteins (47). Therefore it was important that interaction(s) observed via the split-ubiquitin approach be confirmed by independent physical methods like in vitro binding assays. However, with respect to associations observed between OT subunits and translocon components, it is important to note that all the proteins under study are highly hydrophobic transmembrane proteins. An additional factor that may play an important role in the OT-translocon interaction is the presence of a nascent chain provided by an actively translating ribosome, which would be obviously absent during in vitro studies. Because two interacting proteins would be expected to cross-link to each other using a suitable cross-linker, we used chemical cross-linking as an alternative in vitro approach to study the translocon-OT organization. Although chemical cross-linking techniques may report indirect cross-linking of two proteins via another partner, we found that results from the chemical cross-linking technique for the most part agree with those from the split-ubiquitin approach. Therefore, we believe that the observed interactions are real and are likely to have functional significance.
Based on our findings using the split-ubiquitin and the  Fig. 1A. The transformants were selected on -Leu-Trp plates. Three independent colonies were tested for growth on -Leu-Trp-His plates. Spotting assay of one of the colonies on -Leu-Trp or -Leu-Trp-His plates is shown. A quantitative ␤-galactosidase assay was carried out using the yeast ␤-galactosidase assay kit as per the manufacturer's protocol (Pierce). To measure ␤-galactosidase activity, the equation used was 1000 A 420 /tVOD 660 , where A is the absorbance measured at 420 nm after the reaction, t is the time in minutes of reaction incubation, V is the volume of cells (ml) used in the assay, and OD is the optical density of the culture used for the reaction. Average of three estimations for each cell culture is plotted in the figure and error bars indicate S.E.  Fig. 1A. The transformants were selected and processed as described in the legend to Fig. 2. chemical cross-linking techniques, and the extrapolation of the crystal structure of the prokaryotic Sec complex to the eukaryotic system, we propose the model shown in Fig. 6 for the organization of the translocon-OT supercomplex. Interestingly, the OT subunit, Stt3p was found to interact only with Sec61p, whereas Ost4p was found to interact with all three components (Sec61p, Sbh1p, and Sss1p) of the translocon. Three other OT subunits encoded by essential genes (Ost1p, Ost2p, and Swp1p) were found to interact with Sec61p and also either with Sbh1p or Sss1p. Ost3p was found to interact weakly with Sbh1p, the significance of which will be discussed in detail elsewhere. 2 Ost5p and Ost6p interacted very weakly with Sec61p and failed to interact with Sbh1p or Sss1p. The OT subunits that were found to interact with Sec61p, Sbh1p, or Sss1p in the splitubiquitin approach were found to cross-link to each other. Based on the crystal structure of the prokaryotic Sec complex, it has been proposed that one side of the ␣-subunit of the translocon (Sec61p in yeast) is open and the remaining three sides are surrounded by ␤ (Sbh1p in yeast) and ␥ (Sss1p in yeast) subunits (46). We propose that the OT complex is associated with the translocon via the open side of the ␣-subunit (Sec61p). It is known that the OT complex may exist in translocon-associated and -unassociated form (12). In this study, we have focused on a subset of the OT complex that is associated with the translocon. It is unclear whether the transloconassociated and -unassociated OT complexes differ in their organization.
As mentioned earlier, Ost4p was found to interact with all three components of the translocon. This is surprising because Ost4p is a mini-membrane protein of only 36 amino acids (48). It is possible that multiple copies of Ost4p may be present in the OT complex or its kinked structure (49) may facilitate its interaction with multiple partners. Ost4p is known to interact very strongly with Stt3p (42,50), which was found to interact only with Sec61p. Recent studies using a photoprobe attached to an acceptor peptide indicates that Stt3p contains the glycosylatable peptide-binding and/or the active site of the OT reaction (51). These findings were corroborated by studies carried out in the mammalian system in which the STT3 protein was found to be the only OT subunit cross-linked to the glycosylation consensus sequence of the translocating nascent chain (25). A homolog of STT3 has been identified in a bacterium and has been shown to be essential for glycosylation observed in this species (52). Homologs of none of the other OT subunits exist in this species. A very high degree of conservation between various STT3 homologs ranging from S. cerevisiae to Homo sapiens also suggests that this protein may indeed be central to the OT reaction. Therefore a strong direct association between Stt3p and Sec61p would be expected to improve the efficiency of glycosylation following polypeptide translocation.
Based on earlier biochemical and genetic studies, it has been proposed that the OT complex is composed of 3 subcomplexes, Ost3p-Ost4p-Stt3p, Ost1p-Ost5p, and Ost2p-Wbp1p-Swp1p (50). The model shown in Fig. 6 not only provides support for the existence of the two of these subcomplexes, Ost3p-Ost4p-Stt3p and Ost5p-Ost1p, but also is consistent with previous observations such as interaction of the lumenal domains of Ost1p and Swp1p (53) and physical association of Ost3p and Wbp1p (50). It has been proposed that the 3 OT subunits Ost2p, Wbp1p, and Swp1p exist in a subcomplex within the OT complex based on several observations: 1) suppression of the temperature-sensitive phenotype of wbp1-2 mutation by overexpression of OST2 or SWP1 (41,54); 2) chemical cross-linking of Wbp1p and Swp1p (55); 3) identification of cross-linked products of the mammalian homologs of Ost2p (DAD1) and Wbp1p (OST48) and Swp1p (ribophorin II) (19); and 4) chemical crosslinking of Ost2p to Wbp1p (28). Surprisingly, a genetic or biochemical interaction has never been observed between Ost2p and Swp1p. Our model supports the proximity between Ost2p and Wbp1p, and Wbp1p and Swp1p, however, it does not support the existence of the third subcomplex, Ost2p-Wbp1p-Swp1p. We observed an interaction of Sss1p with Wbp1p, which has been previously noted (26). However, we found that Sss1p interacts much more strongly with Ost1p, Ost4p, and Swp1p, but only weakly with Wbp1p.
In the present study, Ost1p was found to interact with Sss1p and to a minor extent Sec61p in the split-ubiquitin assay, whereas it was found by cross-linking to be close to all the three components of the translocon. In an earlier study, Ost1p was found to cross-link with all the other OT subunits and Alg1p, an enzyme in the Dol-P-P-oligosaccharide biosynthesis (28). Interestingly, the mammalian homolog of Ost1p (ribophorin I) was found to cross-link to the ribosomes (23). In another recent study, ribophorin I was found to cross-link to a subset of membrane proteins, irrespective of their glycosylation status, after their integration into the Sec61 translocon (56). In another study, because Ost1p was able to cross-link to the glycosylatable substrate-based peptide photoaffinity probe, it was concluded that Ost1p bears the peptide-binding site of the OT complex (57). However, extensive mutagenesis studies later disproved this hy-2 A. Yan and W. J. Lennarz, manuscript in preparation.  Fig. 1C. The transformants were selected and processed as described for Fig. 2. YG0673, transformed with pRS314-OST1-NUB(G) (2 m plasmid) or pRS314-NUB(G)-ALG5, were used as positive and negative controls, respectively, for the split-ubiquitin assay (37). The growth phenotype and ␤-galactosidase activity of the positive and negative control is indicated in the figure. pothesis (58). All these observations indicate that Ost1p (or its mammalian homolog) possesses an extraordinary ability to be cross-linked. It is important to note that the functional domain of Ost1p is only its membrane-anchored lumenal domain (58). One could speculate that Ost1p may be involved in quality control of newly synthesized transmembrane proteins, because it is found to cross-link predominantly transmembrane proteins but not lumenal proteins like protein disulfide isomerase (28).
The OT complex is localized to the ER membrane, and its subunits are translocated and integrated into the ER membrane via the Sec61 complex, with which it interacts to synthesize glycoproteins. Therefore one specific concern arises: are the interactions observed during the course of this study an artifact because of the translocation of the nascent OT subunits themselves, or are these real functional interactions between mature OT and the translocon? The former possibility seems unlikely, because it would suggest that every membrane or secreted protein should cross-link and exhibit interactions with the translocon components. This was not found to be the case when we tested if Alg1p cross-links or interacts with any of the components of the translocon (data not shown). Besides, the process of translocation and integration of the membrane protein would require transient interaction with the translocon; prolonged association will result in a highly inefficient protein biosynthetic apparatus, which is not the case. Therefore we believe that the interactions observed during the course of this study are functionally relevant.
Thus it is clear that a direct association between Sec61p and Stt3p as well as other multiple interactions between the OT complex and the translocon have evolved to generate an effi-FIG. 5. Cross-linking experiment to identify the OT subunits in close vicinity of the translocon components. The chromosomal copy of Sbh1p or Sss1p was Myc-tagged in yeast strains in which one OT subunit was HA-tagged. Myc tagging of Sbh1p or Sss1p did not result in any obvious growth defects (data not shown) as compared with the wild-type strain. To identify the OT subunits in proximity to Sec61p or Sss1p, the strains bearing the Myc tag on Sss1p and HA tag on one of the OT subunit were used. To analyze the OT subunits in close proximity to Sbh1p, the strains bearing the Myc tag on Sbh1p and HA tag on any one of the OT subunits was used. A, following tagging, yeast lysates were prepared from strains indicated in A as described earlier (59). The yeast lysates were separated on SDS-PAGE and electroblotted on nitrocellulose membrane. The membrane was probed with polyclonal antibodies indicated in the figure, followed by treatment with HRP-coupled anti-rabbit IgG. All OT subunits could be easily visualized with the exception of Ost2-HAp, which was a faint band. Degradation products of certain OT subunits were observed and are indicated with an asterisk. B, yeast microsomes of the OT-HA, Sss1-Mycp were treated with the cross-linking agent, DSP (ϩDSP). Following quenching of the excess cross-linker, the microsomes were solubilized in buffer containing 1% Triton X-100 and 0.2% SDS. Microsomes not treated with DSP (ϪDSP) were also analyzed. Immunoprecipitation of the OT subunits was carried out using with monoclonal anti-HA antibody and protein G-agarose beads. Following SDS-PAGE and electroblotting, the blot was probed with polyclonal anti-HA antibody (i) and HRP-coupled anti-rabbit IgG to confirm that equal amounts of the OT-HA subunit were immunoprecipitated. An identical blot was probed with polyclonal anti-Sec61p antisera (ii) or anti-Myc antibodies (iii) and HRP-coupled anti-rabbit IgG. Yeast microsomes of the OT-HA, Sbh1-Mycp strain were analyzed similar to the OT-HA, Sss1-Mycp strains. The results using blots probed with poly-HA antibodies were similar to B (i), therefore these results are not shown. The blot probed with poly-Myc antibodies to detect Sbh1-Mycp is indicated in B (iv). WB, Western blot. cient glycoprotein biosynthetic machine. What remains to be established is a detailed picture of the structure of the translocon-OT supercomplex! FIG. 6. Top (pore) view of a working model for the organization of the translocon-OT supercomplex. Based on the x-ray structure of the bacterial protein-conducting channel (46), the structure of the eukaryotic translocon complex has been predicted. Based on the interactions identified using the split-ubiquitin approach, and proximities determined by the cross-linking studies, the organization of the translocon-OT super-complex is depicted. The OT subunits encoded by 5 essential and 3 non-essential genes are indicated in green and blue, respectively. The three components of the translocon are designated in purple. Ost2p, Wbp1p, Stt3p, Ost4p, and Swp1p were found to interact strongly with Sec61p, whereas Ost3p, Ost6p, and Ost5p were found to interact weakly with Sec61p. Ost6p failed to interact with Sss1p or Sbh1p, and is believed to exist in place of Ost3p in a subset of OT complexes (60). Therefore Ost6p is not indicated in the model. Ost4p, Ost1p, and Swp1p were found to associate with Sss1p. Although Ost5p was not found to interact with any of the translocon proteins, Ost5p is placed next to Ost1p based on previous genetic studies (61). Ost2p, Wbp1p, and Ost3p were found to associate with Sbh1p. Wbp1p was found to interact with Sec61p, Sbh1p, and Sss1p. Wbp1p possesses only one transmembrane domain and therefore it is not possible to position Wbp1p in the model with enough confidence. Although Ost1p was found to cross-link Sec61p, Sbh1p, and Sss1p, it was found to interact with only Sss1p and Sec61p by split-ubiquitin assay and has been depicted in the model accordingly. Although the transmembrane proteins indicated in the figure are depicted to be oriented vertically, it is possible that their transmembrane helices may not traverse the membrane vertically, as seen for other membrane proteins (46).