Mammalian Sec61 is associated with Sec62 and Sec63.

In yeast, efficient protein transport across the endoplasmic reticulum (ER) membrane may occur co-translationally or post-translationally. The latter process is mediated by a membrane protein complex that consists of the Sec61p complex and the Sec62p-Sec63p subcomplex. In contrast, in mammalian cells protein translocation is almost exclusively co-translational. This transport depends on the Sec61 complex, which is homologous to the yeast Sec61p complex and has been identified in mammals as a ribosome-bound pore-forming membrane protein complex. We report here the existence of ribosome-free mammalian Sec61 complexes that associate with two ubiquitous proteins of the ER membrane. According to primary sequence analysis both proteins display homology to the yeast proteins Sec62p and Sec63p and are therefore named Sec62 and Sec63, respectively. The probable function of the mammalian Sec61-Sec62-Sec63 complex is discussed with respect to its abundance in ER membranes, which, in contrast to yeast ER membranes, apparently lack efficient post-translational translocation activity.

The mammalian Sec61 complex consisting of Sec61␣, Sec61␤, and Sec61␥ has been identified as a crucial membrane component involved in the signal recognition particle (SRP) 1dependent co-translational protein translocation across the endoplasmic reticulum (ER) membrane (for review see Ref. 2). The Sec61 complex forms the hydrophilic pore in the membrane through which the nascent polypeptide is translocated (3)(4)(5)(6), and it is responsible for the tight binding of the ribosome to the ER membrane during the co-translational transport process (7). Moreover, the Sec61 complex is involved in the recognition of the signal sequence regulating the insertion of the nascent polypeptide chain into the translocation channel (8,9). The ribosome-bound Sec61 complex is in spatial proximity to several other membrane components that interact with the nascent polypeptide chain during its co-translational translocation. These components include the translocating chain-associated membrane protein (TRAM) (10), the signal peptidase complex (SPC) (11)(12)(13), the oligosaccharyltransferase complex (14), the translocon-associated protein (TRAP) complex (15), and the ribosome-associated membrane protein 4 (RAMP4) (4,16). Whereas the functions of the SPC and the oligosaccharyltransferase complex are well established, the role of the other components is at best poorly understood or completely unknown. Vectorial co-translational protein translocation into the ER can be reconstituted in the absence of chaperones using proteoliposomes consisting exclusively of the SRP receptor (essential for the SRP-dependent targeting step (17)), the Sec61 complex, and TRAM (4). However, other data suggest that chaperones in the ER lumen play a stimulatory role during translocation in vitro (18,19). Moreover, the Hsp70 homolog BiP is likely involved in the formation of a tight seal that blocks ion transport across the Sec61 complex in the absence of protein translocation (20).
In the yeast Saccharomyces cerevisiae, in the absence of tightly bound ribosomes, the trimeric Sec61p complex is found associated with other polypeptides (Sec62p, Sec63p, Sec71p, and Sec72p) (21,22), which together form the Sec complex (23). The yeast Sec complex is essential and sufficient for the posttranslational protein translocation into the ER (24). Sec63p has a DnaJ-like domain located in the ER lumen (25), which recruits the Hsp70 homolog Kar2p to the translocation sites. The DnaJ domain of Sec63p and Kar2p form a molecular ratchet that is responsible for the ATP-dependent vectorial movement of the polypeptide into the ER lumen (26). In contrast to Sec62p, Sec63p, and Kar2p, neither Sec71p nor Sec72p are essential for post-translational translocation (27)(28)(29). It is possible that components of the Sec complex are also involved in other cellular processes. Mutations in Kar2p, Sec71p, Sec72p, and Sec63p affect karyogamy in yeast (30,31), and there is genetic evidence that Sec61p, Sec63p, and Kar2p are involved in the ubiquitin-and proteasome-dependent degradation of proteins at the ER (32)(33)(34).
In mammalian cells it is also the case that not all Sec61 complexes are tightly associated with ribosomes (35). To date, however, homologs of neither the yeast Sec62-Sec63 subcomplex nor other membrane components found preferentially associated to ribosome-free Sec61 complexes have been identified. We therefore set out to identify and characterize such proteins of the mammalian ER.
We show here, that the mammalian ER contains proteins that display structural homology to the yeast Sec62p and Sec63p. Both proteins are expressed ubiquitously, and their abundance is similar to that of known components of the ER translocation machinery. To gain an indication as to the function of these proteins, we analyzed their molecular environment in the ER using different biochemical methods. We found that Sec62 and Sec63 are associated with Sec61 complexes, indicating that these proteins might be involved in transport processes similar to those performed by the yeast Sec complex. The association was only detectable in the absence of tightly bound ribosomes. This suggests that the function of Sec62 and Sec63 is most likely not directly linked to the co-translational protein translocation across the mammalian ER.

EXPERIMENTAL PROCEDURES
cDNA Cloning-Human Sec63 cDNA was recovered by polymerase chain reaction using the oligonucleotides GGTGGTACCAAGGCAC-CGCCACTGCCTAG and ATGGCCGGGCAGCAGTTCCAG and HeLa cDNA as template. Design of the oligonucleotide sequences was based on information from two human expressed sequence tags, GenBank TM accession numbers N79940 and A227342, respectively. Two of the obtained clones were sequenced on both strands.
Partial clones of human Sec62 were isolated by screening a HeLa cDNA library (Stratagene) with the oligonucleotide GTCCACCAATAT-GATGGGTCACC from a human expressed sequence tag (GenBank TM accession number H01748) using standard protocols. The missing 5Јends were obtained by rapid amplification of cDNA ends using a fetal brain Marathon-cDNA (CLONTECH). Clones covering the entire open reading frame were subsequently obtained by polymerase chain reaction using the same cDNA, and two of them were sequenced.
Sequence analysis was performed using software from PCGENE and from the web page of the National Center for Biotechnology Information. The sequences of human Sec62 and Sec63 were deposited in the GenBank TM data base under the accession numbers U93239 and AF100141, respectively.
Cell Fractionation-Cell fractionation was performed (37). 10 g of bovine liver were homogenized in buffer H (50 mM HEPES-KOH (pH 7.8), 25 mM potassium acetate, 5 mM magnesium acetate, 5 mM ␤-mercaptoethanol, 0.8 mM phenylmethylsulfonyl fluoride, and protease inhibitor mix) with 250 mM sucrose and centrifuged for 13 min at 1,750 rpm (Sigma 3K12 centrifuge, 4°C). The supernatant was recovered and centrifuged for 13 min at 3,850 rpm (Sigma 3K12 centrifuge, 4°C). To remove mitochondrial material the supernatant was centrifuged a further 10 min at 17,000 rpm (Ti-60 rotor, 4°C). The post-mitochondrial supernatant was adjusted to 1.35 M sucrose and layered over a step gradient containing buffer H with 1.5 M sucrose and 2.0 M sucrose. The gradient was overlaid with buffer H containing 1 M sucrose and centrifuged for 17 h at 45,000 rpm (Ti45 rotor, 4°C). The membranes that concentrated at the 1.5-2.0 M interface (RM) were collected, diluted with 1 volume of buffer H without sucrose, and recovered by centrifugation for 2 h at 45,000 rpm (Ti45 rotor, 4°C). Similarly, the smooth membranes were recovered from the 1.0 -1.35 M interface.
Sucrose gradient centrifugation was performed essentially as described (35). 50 eq bovine rough microsomes were solubilized at 0.4 eq/ml in 2% digitonin, 50 mM HEPES-KOH (pH 7.8), 450 mM potassium acetate, 8 mM magnesium acetate, and protease inhibitor mix. After centrifugation for 3 min at 14,000 rpm in a microcentrifuge the supernatant was layered on top of a 25-50% (w/v) sucrose gradient containing 2% digitonin, 50 mM HEPES-KOH (pH 7.8), 500 mM potassium acetate, 10 mM magnesium acetate, and protease inhibitor mix and centrifuged at 55,000 rpm for 1 h (TLS55 rotor, 4°C). Fractions containing ribosomes were identified by the presence of the ribosometypical protein pattern after separation by SDS-PAGE and staining with Coomassie Blue.
Alternatively, bovine RM pre-washed with saponin in the presence of 0.8 M potassium acetate (24) or PK-RM were solubilized in 2.5% digitonin, 50 mM HEPES-KOH (pH 7.8), 500 mM potassium acetate, 10 mM magnesium acetate, 10% (w/v) glycerin, 5 mM ␤-mercaptoethanol, and protease inhibitor mix and centrifuged for 30 min at 100,000 rpm (TLA-100.3 rotor, 4°C). The supernatant was bound to the respective antibody columns, and the bound material was eluted as indicated in the text.
Purification of the Sec61-Sec63 Complex-PK-RM corresponding to about 30,000 eq were resuspended in 2.5% digitonin, 50 mM HEPES-KOH (pH 7.8), 500 mM potassium acetate, 10 mM magnesium acetate, 10% (w/v) glycerin, 5 mM ␤-mercaptoethanol, and protease inhibitor mix to a final concentration of 0.75 eq/l. After a centrifugation step (1.5 h at 4°C and 70,000 rpm, Ti70 rotor), the supernatant was applied to a HiTrap Q column (Amersham Pharmacia Biotech). The column was washed with buffer W (0.5% digitonin, 50 mM HEPES-KOH (pH 7.8), 10% (w/v) glycerin, 5 mM ␤-mercaptoethanol, 10 mM magnesium acetate) supplemented with 500 mM potassium acetate. The elution was performed with increasing concentrations of salt in step fractions from 0.6 to 1.2 M potassium acetate in buffer W. The eluate at 1.0 M salt was then diluted with 2 volumes of buffer W containing protease inhibitor mix and passed over an anti-Sec61␤ antibody column. The column was washed with buffer W containing 300 mM potassium acetate and protease inhibitor mix. The bound material was eluted at room temperature (5 ml per h) with 1 mg/ml of the peptide against which the antibodies were raised in buffer W supplemented with 200 mM potassium acetate and protease inhibitor mix.
Miscellaneous-Immunoblotting and immunoprecipitation were carried out as described (10). Partial protein sequences were obtained from purified proteins as described (24). The final concentration of protease inhibitor mix was 10 g/ml leupeptin, 5 g/ml chymostatin, 2 g/ml pepstatin, and 10 g/ml aprotinin. Protein concentrations were determined by quantitative immunoblotting using ECL peroxidase (NEN Life Science Products) and a CSC chemiluminescence camera (Raytest). Digitonin was purchased from Sigma, deoxy-BIGCHAP from Calbiochem, and saponin from Roth.

Identification and Cloning of Human Homologs of Sec63p
and Sec62p-In order to identify proteins that are associated to the Sec61 complexes in the absence of ribosomes, bovine rough microsomes were treated with puromycin in the presence of 500 mM salt (PK-RM) and solubilized with digitonin. The detergent extract was bound to anti-Sec61␤ antibodies (Fig. 1). The affinity-purified fraction contained the components of the trimeric Sec61 complex as well as two proteins approximately 80 and 97 kDa in size (Fig. 1, lane 4). Peptides derived from both proteins were subjected to Edman degradation. The 80-kDa protein was identified as BiP. The peptide sequences of the 97-kDa protein ( Fig. 2A) corresponded to a group of human expressed sequence tags in the GenBank TM data base. An analysis of these sequences revealed that they belonged to a cDNA that has some homology to yeast Sec63p (about 20% identical amino acids). The entire coding region of this cDNA was cloned and sequenced. The deduced protein sequence of human Sec63 contains all peptides obtained from the 97-kDa protein ( Fig. 2A). Similar to the yeast Sec63p, the human Sec63 and the homologs from Arabidopsis thaliana and Caenorhabditis elegans have a DnaJ domain and three membrane-spanning domains ( Fig. 2A). However, several residues proven to be critical for the interaction of the DnaJ domain of yeast Sec63p with Kar2p are not strictly conserved (38). At the primary structure level the most conserved part is predicted to be located in the ER lumen, spanning the DnaJ domain through to the end of the third proposed membrane anchor (about 40% identity between yeast and human). Remarkably, the carboxylterminal 500 amino acids of Sec63 proteins from higher eukaryotes form a sequence motif that is found twice in the carboxyl terminus of 200-kDa proteins of U5 small nuclear ribonucleoprotein.
We postulated that mammals may also contain homologs of the other proteins present in the yeast Sec62p-Sec63p subcomplex. Therefore we screened the GenBank TM data base and identified several partial human cDNA sequences that displayed homology to Sec62p. Based on this information we cloned and sequenced cDNAs that contained the entire open reading frame of the gene. While this work was in progress, a complete human Sec62 cDNA was published (39). Fig. 2B shows the alignment of the human protein with homologous proteins from the invertebrate Drosophila melanogaster and C. elegans and from the yeast species Schizosaccharomyces pombe, Yarrowia lipolytica, and S. cerevisiae. In all cases, Sec62 is predicted to have two membrane-spanning segments. The domains flanking the membrane anchors, including the intervening luminal domain, display a high degree of conservation in their primary structure among all proteins analyzed (34% identity between yeast and human Sec62). Regions that are closer to the termini of Sec62, show a striking similarity exclusively to the homologous animal proteins. Based on the sequence information peptides were designed to raise antibodies against Sec62 and Sec63 (Fig. 2).
Sec62 and Sec63 Are Not Associated with Membrane-bound Ribosomes-In order to gain an indication of the possible function of the two proteins, we next analyzed their molecular environment. First we tested whether or not Sec62 and Sec63 are associated with membrane-bound ribosomes characteristic of Sec61␣, TRAP␣ and other ribosome-associated membrane proteins (RAMPs) (4). Rough microsomes were solubilized with digitonin in a buffer containing 450 mM potassium acetate and separated by sucrose gradient centrifugation (Fig. 3). Most of Sec61␣ and Sec61␤ and nearly 50% of TRAP␣ were found in fractions 1-10 co-migrating with the ribosomes as has been reported previously (35). In contrast, Sec62, Sec63, TRAM, and the 25-kDa subunit of the SPC remained in the ribosome-free fractions [11][12][13][14][15][16][17][18][19]. Similar results were obtained if membranes were solubilized with the detergent deoxy-BIGCHAP (not shown and Fig. 5B, lane 3), with the exception that TRAP␣ was predominantly found in the ribosome-free fractions (not shown).
Sec62 and Sec63 Associates with Ribosome-free Sec61 Complexes-We subsequently chose to investigate the molecular environment of Sec62 and Sec63 in the membrane. The purification of mammalian Sec63 by an anti-Sec61␤ antibody column indicated that the two proteins are in a complex (Fig. 1). To confirm this result and to identify further proteins of the ER membrane that are associated with mammalian Sec62 or Sec63, we performed immunoprecipitation experiments using anti-Sec62 and anti-Sec63 antibodies. Bovine RM were first treated with saponin in the presence of 0.8 M salt to obtain membranes enriched in integral membrane proteins. These membranes were solubilized with digitonin, and the extract was bound to the antibody column (Fig. 4A). The proteins that eluted from the anti-Sec63 antibody column were analyzed by peptide sequencing (Fig. 4A, lane 1). Sec63, Sec61␣, Sec61␤, Sec61␥, and a contamination with immunoglobulins were detectable, thus confirming the association between Sec63 and the Sec61 complex. No other proteins were present in significant amounts in the eluate. Material that eluted from the anti-Sec62 antibody column contained exclusively the Sec62 protein (Fig. 4A, lane 2). To find proteins that interacted with Sec62, we repeated the immunoprecipitation experiments using the detergent deoxy-BigCHAP. RM were solubilized; the RAMPs were separated by centrifugation, and the ribosomefree supernatant was applied to the antibody columns. Under these conditions, the anti-Sec61␤ antibodies did not only precipitate Sec61␣ and Sec63 but also Sec62 (Fig. 4B, lane 5). Binding of this membrane extract to an anti-Sec63 antibody column also precipitated about 10% of Sec62, in addition to Sec61␣ and Sec61␤ (Fig. 4B, lanes 7 and 8). In both cases other proteins of the translocation site such as TRAM or SRP receptor ␣ were not co-precipitated.
Together these data indicate that both Sec62 and Sec63 form a complex with Sec61 complexes. In agreement with the results of the sucrose gradient centrifugation (Fig. 3), these complexes could be purified from ribosome-free supernatants of RM (Fig.  4). A caveat in these experiments was that the amount of Sec62 precipitated was very sensitive to the salt concentration used for the solubilization of the membranes (not shown). To confirm the association of Sec62 with the Sec61 complex, we therefore performed cross-linking experiments using canine RM and the chemical cross-linker bismaleimidohexane. We observed several cross-linked products between Sec62 and other proteins in immunoblots (Fig. 5, lane 4). Peptide sequencing of the main FIG. 1. Identification of mammalian Sec63. Bovine PK-RM were solubilized with digitonin, and non-solubilized material was removed by centrifugation. The supernatant (extract) was applied to an anti-Sec61␤ antibody column. After washing (wash), the bound material was eluted with the peptide against which the antibodies were raised (eluate). Samples corresponding to 20 eq (extract and flow-through) or 450 eq (wash and eluate) of the starting material were separated by SDS-PAGE and stained with Coomassie Blue. Protein bands were analyzed by Edman degradation of fragments obtained after digestion with trypsin. The peptides obtained from BiP were TKPYIQVDVG and AVEEKI; peptides obtained from Sec63 are indicated in Fig. 2. product, which contained about 30% of the Sec62 present in the membranes, identified the ␤-subunit of the Sec61 complex as the cross-linked partner. To test whether or not the Sec61␤ cross-linked to Sec62 belongs to a ribosome-bound Sec61 complex, we separated the membrane proteins by centrifugation after solubilization with digitonin into a ribosome-free supernatant and a pellet fraction containing the ribosomes and the RAMPs. The cross-linked product between Sec62 and Sec61␤ remained in the supernatant (Fig. 5, lanes 5 and 6). An analysis of the same samples by immunoblotting using anti-Sec61␤ antibodies revealed that the other Sec61␤-containing crosslinked products were found in the pellet fraction (Fig. 5, lane 9), suggesting that they were ribosome-associated. Among them was a cross-linked product between Sec61␤ and SPC25, a protein that in the absence of cross-linker does not behave like a RAMP (see Fig. 3 (13)). This demonstrates that cross-linking of a non-RAMP to a ribosome-associated Sec61␤ can identify a protein such as a RAMP. Only one band, which according to its mobility in the SDS-PAGE corresponds to the Sec62-Sec61␤ cross-linking product, was entirely found in the supernatant, indicating that it is not ribosome-bound.
Further Characterization of the Sec63-Sec61 Subcomplex-To analyze the Sec61-Sec63 complex in more detail an alternative purification protocol was developed. PK-RM were solubilized with digitonin. The detergent extract was passed over a HiTrap Q column at 0.5 M salt, and the bound material was eluted stepwise with increasing salt concentrations. As reported previously (4), the bulk of the Sec61 complex did not bind and was therefore found in the flow-through (Fig. 6A, lane  2). However, about 5% of the Sec61 complex eluted at 1.0 M salt together with the bulk of the Sec63 (Fig. 6, A, lane 6, and B,  lane 1). This fraction was passed over an anti-Sec61␤ antibody column. About 30% of the Sec63 was found to bind to the Sec61 complex (Fig. 4C, lane 3). Vice versa, all of the Sec61a in this fraction bound to an anti-Sec63 column (not shown). A Coomassie Blue staining of the recovered Sec61-Sec63 complex after its separation by SDS-PAGE revealed that the preparation did not contain significant amounts of other proteins (Fig. 6B, lane 3). Sec62 remained either in the flow-through or eluted at 0.6 M salt (Fig. 6A, lanes 2 and 4). Immunoprecipitation of these fractions using anti-Sec62 antibodies did not detect proteins associated to Sec62 (not shown). The amount of Sec63 and of Sec61␣ in four independently purified complex preparations was determined by semi-quantitative immunoblotting (not shown). The molar ratio between Sec63 and Sec61␣ was in the range between 1.2 and 1 and 1.9 and 1.
Sec62 and Sec63 Are Ubiquitously Expressed in the Endoplasmic Reticulum-Finally we wanted to explore the expression pattern of Sec62 and Sec63. First, we performed immunoblot experiments using crude membrane fractions from different rat tissues and rough microsomes derived from different bovine tissues (Fig. 7A). Both proteins were identified in all tissues examined with the highest abundance in samples that also have a high level of Sec61␤. Next, we performed a cell fractionation using bovine liver as starting material (Fig. 7B). Nearly all Sec62 and Sec63 was found in the post-mitochondrial supernatant. A separation of this fraction into RM and smooth membranes revealed that both Sec62 and Sec63 were present in the rough ER. Both proteins were also found in the membrane fraction that contained the smooth ER. This fraction, which was essentially free of RM, contained more than 50% of the Sec62. The relative amount of Sec63 in the smooth membranes was significantly lower. Similar results were obtained in fractionation experiments using mouse liver (not shown). The pattern of the intracellular distribution of Sec62 and Sec63 in HepG2 cells observed by immunofluorescence was indistinguishable from that obtained with antibodies against proteins of the ER lumen (not shown). We concluded that most of Sec62 and Sec63 found in the smooth membranes was actually located to the smooth ER.

DISCUSSION
The data presented here demonstrate that the mammalian Sec61␣ can be found in a protein complex with structural similarity to the yeast Sec complex. In addition to the components of the trimeric Sec61 complex, Sec61␣, Sec61␤, and Sec61␥, this larger complex contains at least two other membrane proteins. These proteins, Sec62 and Sec63, display homology to the yeast proteins Sec62p and Sec63p, respectively. . After washing, the bound material was eluted using either 1% Triton X-100 in 100 mM glycine at pH 2.2 (lane 1) or at pH 7.8 using a solution of 1 mg/ml of the peptides against which the antibodies were raised (lane 2). The precipitated material was separated by SDS-PAGE and stained with Coomassie Blue. Protein bands were analyzed by Edman degradation of fragments obtained after digestion with trypsin. B, co-immunoprecipitation of Sec62, Sec63, and the Sec61-complex. RM were solubilized in deoxy-BIGCHAP, and the unsolubilized material was removed by low speed centrifugation. The extract (total) was cleared of ribosomes (pellet), and the supernatant (supernatant) was applied to the indicated antibody columns. Samples corresponding to 2.5 eq of RM were separated by SDS-PAGE and analyzed by immunoblotting. Lane 8 (eluate 2 ϫ) contains material corresponding to 5 eq RM.
Remarkably, for both proteins, regions with significant homology were those exposed to the ER lumen or located close to the cytosolic surface of the membrane. The more distal cytosolic parts were very divergent. While this manuscript was in preparation, Skowronek et al. (40) also published the existence of the mammalian Sec63 and showed that it has the same membrane topology as the yeast protein. Both Sec62 and Sec63 were ubiquitously expressed in the rough ER of mammals, and the expression level of Sec62 in a particular tissue is roughly the same in all species tested. However, we cannot exclude that the abundance of Sec63 differs between species, because the epitope recognized by our anti-Sec63 antibodies is not conserved among mammals. Sec62 was also very abundant in smooth membranes that are essentially free of Sec61 complex. This is in agreement with our observation that Sec62 is expressed at high levels in the adrenal gland 2 and that the mRNA is abundant not only in liver and pancreas but also in muscle tissues (39).
Although Sec62 and Sec63 were abundantly expressed, the actual concentration of Sec61-Sec62-Sec63 complexes in the ER appears to be relatively low. Regardless of the detergent used, and whether the purification started with RM or with PK-RM, only about 5% of Sec61 and about 30% of Sec63 were found in a complex. In each case the molar ratio between Sec61␣ and Sec63 in the complex appeared to be 1:1 to 1:2. The binding of Sec63 to the complex was stable, whereas the association of Sec62 with this complex was much weaker. Under optimized purification conditions, we found about 15% of the Sec62 copurified with Sec61␤. However, in the cross-linking experiment 30% of the Sec62 was linked to Sec61␤. In the same samples less than 5% of the Sec61␤ was in proximity to Sec62, similar to that found with Sec63. Therefore the amount of Sec62 and Sec63 in these complexes was likely to be the same. If one assumes that 3 to 4 pentameric Sec61-Sec62-Sec63 units form a translocation pore as it has been shown for the trimeric complex (5), then not more than 5% of all pores in the mammalian rough ER have a Sec-like structure. What could be the function of the mammalian Sec61-62-63 complexes? Despite the clear differences between the cytosolic domains of this complex and the yeast Sec complex, one may 2 H. Grau and E. Hartmann, unpublished information.
FIG. 5. The cross-linked Sec62-Sec61␤ is not ribosome-associated. Canine RM were treated with 50 M bismaleimidohexane. Aliquots of the cross-linked RM and of the untreated control were solubilized in 2.0% digitonin, 50 mM HEPES-KOH (pH 7.8), 500 mM potassium acetate, 10 mM magnesium acetate 10% (w/v) glycerin, 5 mM ␤-mercaptoethanol, and protease inhibitor mix and centrifuged at 14,000 rpm for 5 min in a microcentrifuge. The supernatant (T) was centrifuged at 100,000 rpm for 10 min (TLA100 rotor, 4°C). Supernatant (S) and pellet (P) of the second centrifugation were collected and analyzed by SDS-PAGE and subsequent immunoblotting using antibodies directed against Sec62 or Sec61␤. BMH, bismaleimidohexane.
FIG. 6. Purification of the Sec63-Sec61 complex. A, digitonin-solubilized PK-RM were cleared of unsolubilized material (total) and applied to a HiTrap Q column. After a washing step (wash) the bound material was eluted stepwise at increasing salt concentrations. Aliquots of each purification step were separated by SDS-PAGE and analyzed by Coomassie Blue staining and by immunoblotting using the antibodies indicated. KAc, potassium acetate. B, the material eluted at 1.0 M salt (total) was applied to an anti-Sec61␤ antibody column. After washing, the bound material was eluted with the peptide against which the antibodies were raised (eluate). Aliquots of the purification steps were separated by SDS-PAGE and analyzed by Coomassie Blue staining and by immunoblotting using the antibodies indicated. Lane 3 of the SDS-PAGE contains 20 times more material than lanes 1 and 2. The immunoblot contains 2.5 times more material in lane 3 than in lanes 1 and 2.
speculate that the mammalian Sec61-62-63 may also perform post-translational protein translocation into the ER. However, so far no efficient post-translational translocation has been observed in mammals. Proteins that translocate post-translationally across yeast membranes in vitro, like the prepro-␣factor or the prepro-OmpA, do not do so across mammalian RM (46). The only natural substrates known to enter the mammalian ER after termination of their synthesis in a signal sequencedependent manner are short membrane protein precursors such as M13 procoat (41) or short precursors of hydrophobic secretory proteins like prepro-cecropin (42,43). It has been suggested that these proteins may require membrane proteins for their translocation (42,44). However, it remains possible that their transport is independent of membrane components as has been found for M13 procoat in Escherichia coli (45).
We were not able to identify Sec62 or Sec63 in association with ribosome-bound Sec61 complexes. However, this does not exclude that the Sec61-Sec62-Sec63 complexes translocate nascent polypeptide chains that are in the process of being synthesized by ribosomes not tightly bound to the translocation site. Another possibility is that the Sec61-Sec62-Sec63 complex performs the backward transport of ER proteins that are subject to the ubiquitin-proteasome-dependent degradation pathway as it has been suggested for the yeast Sec complex (32-34).
One should bear in mind that the majority of the Sec62 and the Sec63 were not found to be complexed. These populations could represent a pool that under appropriate conditions form Sec61-Sec62-Sec63 complexes. Alternatively, these molecules may perform functions while loosely associated with a subset of co-translational translocation sites. For example, they could recruit BiP molecules to resting trimeric Sec61 complexes in order to seal the pores, in alignment with a previous suggestion (20). They may also assist in the release of the translocational pausing of polypeptides such as the apolipoprotein B (47).
The identification of a Sec-like complex in mammals again demonstrates the high degree of evolutionary conservation of the translocation machinery in the ER among eukaryotic organisms. However, the extent to which this structural similarity results in functional homology remains to be determined. FIG. 7. Mammalian Sec62 and Sec63 are ubiquitously expressed in the endoplasmic reticulum. A, ubiquitous expression of Sec62 and Sec63. Crude A membranes obtained from rat tissues corresponding to 70 g of protein or 3 eq of RM purified from bovine tissues were separated by SDS-PAGE and analyzed by immunoblotting using the indicated antibodies. B, Sec62 and Sec63 are located to the rough ER. Bovine liver was fractionated as described. Samples of the different fractions were separated by SDS-PAGE and analyzed by immunoblotting using the antibodies indicated. Smooth microsomes were tested for the absence of rough ER by TRAP␣ antibodies and for the presence of smooth ER by antibodies against a NADPH cytochrome P450 (CytP450). Sup., supernatant.