(Received for publication, January 13, 1997, and in revised form, March 17, 1997)
From the Max-Delbrück-Center for Molecular Medicine, Robert-Rössle-Strasse 10, 13125 Berlin-Buch, Germany
In eucaryotic cells signal sequences of secretory and membrane proteins are cleaved by the signal peptidase complex during their transport into the lumen of the endoplasmic reticulum. The signal peptidase complex in yeast consists of four subunits. To date, three of these subunits have been functionally characterized. One of them, the Sec11p, is essential for viability of yeast cells. It shows significant homology to the mammalian SPC18 and SPC21 as well as to bacterial leader peptidases. Two other subunits, Spc1p and Spc2p, have been shown to be homologous to mammalian SPC12 and SPC25, respectively, and are not essential for protein translocation or signal peptide cleavage.
We have purified and analyzed the fourth subunit of yeast signal peptidase, Spc3p. The protein is essential for viability of yeast cells. Depletion of SPC3 leads to accumulation of precursors of secretory proteins in vivo and to the loss of the signal peptidase activity in vitro. Therefore, in contrast to the bacterial leader peptidases, yeast signal peptidase requires a second subunit for its function.
Protein translocation into the endoplasmic reticulum
(ER)1 is triggered by signal sequences (1)
that direct precursor proteins to the translocation sites at the ER
membrane (for review, see Ref 2). The trimeric Sec61 complex, the
central constituent of these translocation sites (3), is involved in
various functions directly linked to the transport of the nascent
chain. Its -subunit is a major part of the protein conducting
channel, and during the co-translational mode of translocation, the
majority of ribosome binding sites are provided by the Sec61 complex
(4, 5). In addition, the Sec61 complex is reportedly involved in a
second signal sequence recognition event that takes place in the
membrane during the protein translocation (6). Moreover, recent
electronmicroscopic data demonstrate that the Sec61 complex forms
pore-like structures in the ER membrane (7).
In addition to trimeric Sec61 complexes, yeast contains a heptameric Sec complex consisting of the subunits of the Sec61p complex (Sec61p, Sbh1p, and Sss1p) as well as of the subunits of the Sec62/63 subcomplex (Sec62p, Sec63p, Sec71p, and Sec72p) (8, 9). This Sec complex is sufficient to translocate proteins post-translationally into reconstituted proteoliposomes in the presence of the luminal chaperone Kar2p (9).
Two other protein complexes that are part of the translocation site modify the nascent chain while being translocated, the oligosaccharyl transferase complex and the signal peptidase complex (SPC). The SPC cleaves the signal peptides of most secretory and many membrane proteins as soon as the lumenal domain of the translocating polypeptide is large enough to expose its cleavage site to the enzyme (10). Mammalian signal peptidase has been purified from dog pancreas microsomes as a complex of five different polypeptide chains (11). The cDNAs of all five subunits have been cloned and sequenced (12-15). Remarkable differences exist between the topologies of these proteins within the ER membrane. Three of them, SPC18, SPC21, and the glycoprotein SPC22/23 are single-spanning membrane proteins with their amino termini facing the cytosol. The major part of these proteins is located within the lumen of the ER, and they all contain a second only moderately hydrophobic sequence close to their carboxyl termini (11, 16). SPC18 and SPC21 show high homology to each other (12). Moreover, their sequences are related to leader peptidase (17), the enzyme responsible for signal sequence cleavage during translocation of proteins across the plasma membrane of bacteria (18). Therefore, these polypeptides may function as catalytic subunits. The functions of the SPC22/23, however, remains unclear.
In contrast, SPC12 and SPC25 contain two membrane spanning segments with their amino and carboxyl termini facing the cytosol and almost no residues in the lumen of the ER. Therefore, it has been speculated, that SPC12 and SPC25 are involved in processes other than substrate binding or the actual enzymatic reaction. This assumption was supported by the fact that an active signal peptidase that comprises only an SPC18 homolog and an SPC22/23 homolog could be purified from hen oviduct (19, 20).
In yeast, the SPC contains four subunits (21, 22). Three of them have been characterized to date. The Sec11p is homologous to SPC18 and SPC21 (23) as well as to leader peptidases of bacteria (17, 24). SEC11 is essential for cell viability. The essential nature of Sec11p is seen when haploid cells containing a temperature-sensitive allele of SEC11 are shifted to nonpermissive temperature. These cells rapidly accumulate precursors of secretory proteins and stop growing (25). Spc1p and Spc2p, the other two subunits that were analyzed until now, are homologs of the canine SPC12 and SPC25, respectively. Both yeast proteins are nonessential for cell viability and signal peptide cleavage under normal growth conditions (22, 26). However, they modulate the activity of yeast SPC as overproduction of Spc1p suppresses the sec11 temperature-sensitive phenotype and depletion of Spc2p results in a defect in signal peptidase activity in cells incubated at high temperature.
In this paper, we analyze the function of the fourth subunit of yeast SPC, the Spc3p, which is revealed as the homolog of the canine SPC22/23. SPC3 is essential for cell viability. Its depletion leads to accumulation of precursors of secretory proteins in vivo and to the loss of the signal peptidase activity in vitro. Overexpression of Sec11p does not suppress this phenotype. Therefore, in contrast to bacterial leader peptidases, the yeast enzyme requires a second subunit for its function.
Media and cell growth conditions have been described elsewhere (9, 27). Specific modifications to these procedures are indicated in the text.
Yeast StrainsYeast strains used were: DF5, mat
/a, trp1-1(am)/trp1-1(am),
his3-D200/his3-D200, ura3-52/ura3-52,
lys2-801/lys2-801, leu2-3,-112/leu2-3,-112; HMY1 (this study), mat a,
spc3,
trp1-1(am), his3-D200, ura3-52, lys2-801, leu2-3,-112; HMY2 (this study),
mat a,
sec11::LEU, trp1-1(am), his3-D200, ura3-52,
lys2-801, Leu2-3,-112.
The disruption of the SPC3 was
performed according to the method of Schneider et al. (28).
Two 60-mer DNA-oligos were designed with 40 bp complementary to the
start or the end region of the SPC3 gene and 20 bp
complementary to the sequence of a hemagglutinin tag. Using pMPY-3xHA
as a template in a polymerase chain reaction (PCR), a product
containing a URA3 gene flanked by 3xHA tags and the 40-bp
sequence complementary to the SPC3 gene regions was amplified. The diploid yeast strain DF5 was transformed with this PCR
product and directly selected on ura dropout plates.
Correct integration of the URA3 gene was confirmed by PCR.
Plating of cells onto 5-FOA selects for strains where recombination
between flanking epitope sequences removed the URA3 gene.
The disruption of the SEC11 was performed according to Böhni et al. (25). The pCBsec11::LEU plasmid is a generous gift from Dr. F. Kepes.
Plasmid ConstructsThe plasmid pHM1 used for
complementation of the spc3 mutation was constructed as
follows. SPC3 was amplified by a PCR reaction using a
forward oligonucleotide (ATGTTCTCCTTTGTCCAAAGATTC) beginning with the
start codon of SPC3 and a reverse primer
(AAAGGTACCGAGGATAAGACGCAGTCCAG), which is located 420 bp downstream of
the SPC3 stop codon. The PCR fragment was digested with
KpnI, and the resulting fragment was cloned between the
SmaI-KpnI site of plasmid pRS414 (29) A 500-bp
BamHI fragment containing the Gal1/Gal10 (30) promoter was
inserted into the BamHI site upstream of the
SPC3.
Two plasmids were used for Sec11p overexpression. For pHM3, SEC11 was amplified by PCR using a primer beginning at the start codon of the gene (ATGAATCTAAGATTTGAATTGCAG) and a reverse primer (ATATGGTACCCATCGCGATGGATATTATATG) 400 bp downstream of the stop codon. The resulting fragment was digested with KpnI and cloned into pRS414 at SmaI-KpnI sites. A Gal1/Gal10 promoter was inserted using the BamHI site described above. For pHM4, the above-described PCR fragment containing the Sec11 gene was subcloned into the pUC19 SmaI-KpnI sites. A 0.65-kb HindIII-SacI fragment was excised and then inserted into a pRS426 (31) HindIII-SacI site. A 500-bp SalI-KpnI fragment containing the Met3 (32) promoter was inserted into the SalI-KpnI site upstream of the SEC11.
Purification of ER Signal Peptidase and Amino-terminal Sequence Analysis of Spc3pIsolation of the ER signal peptidase complex from yeast membrane was performed as described previously (26). The identity of the Spc3p was confirmed by Edman degradation of the purified protein.
Sequence AnalysisSequence analysis was performed as described earlier (15). The GenBankTM accession numbers of the homolog data base entries used for the alignment were: X60796[GenBank], chicken; L14331[GenBank], Caenorhabditis elegans; L06072[GenBank], Saccharomyces cerevisiae; and Z69728[GenBank], Schizosaccharomyces pombe.
N-Glycosidase F Assay15 µl of purified signal peptidase
were denatured for 5 min at 62 °C in 1% SDS. 85 µl of buffer F
(20 mM Hepes, pH 7.6, 12 mM EDTA, 1% Triton
X-100, 4 mM -mercapthoethanol) and 1 unit of
N-glycosidase F (Boehringer Mannheim) were added. The sample was incubated at 37 °C for 8 h, separated by SDS-PAGE, and
analyzed by immunoblotting.
Aliquots of cells growing in SD minimal medium were removed at distinct time points and were labeled with 200 µCi of [35S]methionine for 10 min at 30 °C (pulse). After incubation, the cells were stopped with 20 mM sodium azide, washed, and homogenized with glass beads in 100 µl of lysis buffer (1% SDS, 50 mM Tris, pH 7.5, 10 µl/ml leupeptin, 5 µl/ml chymostatin) using a Vortex mixer. The homogenate was diluted with 4 volumes of IP dilution buffer (1.25% Triton X-100, 190 mM NaCl, 60 mM Tris, pH 7.4, 6 mM EDTA, 10 µl/ml leupeptin, and 5 µl/ml chymostatin), and labeled proteins were precipitated with anti-Kar2p antibodies. The immunoprecipitated material was analyzed by SDS-PAGE and fluorography. The ratio of the radioactivity in the preKar2p and Kar2p bands was determined by a phosphoimager (Fuji Bas 2000).
In Vitro Assay for Signal Peptidase ActivityThe
assay was performed according to YaDeau et al. (21) with the
following modifications. Yeast cells of induced cultures were shifted
to glucose (YPD or SD minimal medium) and were grown for various times.
After harvesting the cells, membranes were isolated. To obtain
comparable amounts of membranes, the samples were normalized to same
amounts of Sec62p. Digitonin extract corresponding to 10 eq (membrane
equivalent, see Ref. 3) membranes were prepared, and 210 µl of buffer
was added. The final conditions of the assay were 50 mM
triethanolamine, pH 8.0, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 0.4 mg/ml
phospatidylcholine, 0.01% SDS (from the SDS pre-treatment of the
prepro--factor), 120 mM KAc, and 0.2% digitonin. The
added prepro-
-factor was synthesized in a reticulocyte system in the
presence of [35S]methionine. After 2 h at 25 °C,
the samples were precipitated and separated by SDS-PAGE and then
analyzed by a phosphoimager (Fuji Bas 2000).
Antibodies against the amino termini of Sec11p and Spc1p as well as the antibody against the carboxyl terminus of Spc2p have been described earlier (26). Anti-Spc3p antibodies were raised by the same methods using a recombinant expressed carboxyl-terminal part of the protein as antigen. Immunoblots were performed as described earlier (9) using the ECL system (Amersham Corp.).
For immunoprecipitation, digitonin extracts corresponding to 50 eq of membranes were used and diluted with LD buffer (9) to a final concentration of 1.5% digitonin, 100 mM KAc, 50 mM Hepes, pH 7.6, 2 mM MgAc, 1 mM phenylmethylsulfonyl fluoride, 10 µl/ml leupeptin, 5 µl/ml chymostatin. 5 to 10 µl of purified antibodies were added and incubated overnight at 4 °C in a rotary shaker. After washing the beads, the proteins were eluted with 30 µl of SDS-sample buffer.
To identify a possible yeast homolog of
canine SPC22/23, we searched the GenBankTM data base for
similar proteins. In addition to the already described chicken gp23
(20), we identified open reading frames in C. elegans, S. pombe, and S. cerevisiae (Fig.
1) as well as expressed sequence tags of human, mouse,
Drosophila melanogaster, Brugia malayi, Schistosoma mansoni,
and Arabidopsis thaliana with significant homology to
SPC22/23. Analysis of the yeast gene, hereafter designated as
SPC3, with different computer programs predicted that it
codes for a glycoprotein with 26% identity to canine SPC22/23. Spc3p has a membrane anchor close to its amino terminus like the other proteins of the family (Fig. 1). Because Spc3p is a glycoprotein (see
below) that has potential sites for N-linked glycosylation only in the carboxyl-terminal part (Fig. 1), it is very likely that
Spc3p has the same membrane topology as canine SPC22/23.
To determine whether Spc3p is part of the yeast signal peptidase, we
purified the SPC from S. cerevisiae as described earlier (26). Antibodies raised against the luminal part of Spc3p recognize a
protein on Western blots that co-purifies with the already known yeast
SPC subunits, Sec11p, Spc1p, and Spc2p (Fig.
2A). This immunoreactive band corresponds to
a prominent protein in a fraction highly enriched in yeast SPC (Fig.
2B). To confirm its identity with Spc3p, the purified
protein was subjected to Edman degradation. The obtained peptide
sequence matches perfectly with the predicted protein sequence (Fig.
1).
To demonstrate the integrity of the complex by an independent method, we performed immunoprecipitation experiments with anti-Spc1p antibodies using yeast membranes solubilized in digitonin as starting material. The analysis of the precipitates by immunoblotting showed that under these conditions, Spc2p, Sec11p, and Spc3p are co-precipitated but the control protein Sec62p was not (Fig. 2C, lanes 1-3).
Spc3p contains two potential sites for N-linked glycosylation (see Fig. 1). To determine whether Spc3p is indeed a glycoprotein, we analyzed a Western blot of the purified SPC which was then incubated with concanavalin A-peroxidase. As depicted in Fig. 2D, lane 1, a concanavalin A-binding protein co-migrates with Spc3p. To confirm this result and to check whether both potential sites of Spc3p might be used, purified yeast SPC was treated with the N-glycosidase F and then analyzed by immunoblotting (Fig. 2D, lane 3). Only Spc3p, but not Spc1p showed a shift in mobility during SDS-PAGE compared with mock-treated membranes (Fig. 2D, lanes 2 and 3). The size of the mobility shift of about 5 kDa indicates that both glycosylation sites are likely to be used. The double band in lane 3 is probably a result of uncomplete digestion.
Spc3p Is Essential for Cell Viability and Signal Peptidase ActivityNext we tested the effect of the in vivo depletion of Spc3p. First, we disrupted one allele of SPC3 in the diploid yeast strain DF5 according to the protocol described under "Experimental Procedures." Tetrade analysis of this strain revealed that the function of SPC3 is essential for cell viability. The cells could be rescued by introducing a plasmid that expresses Spc3p under control of the Gal10 promoter in the presence of galactose, confirming that the loss of SPC3 function is the reason for the observed phenotype.
To analyze the function of Spc3p in more detail, we then cultivated the
rescued strain in liquid culture and determined its growth rate under
conditions that either induce (presence of galactose) or repress
(presence of glucose) the promoter (see "Experimental Procedures").
In different experiments, we found that the growth rate of the
repressed culture starts to decrease after about 10 h (Fig.
3A).
In a second set of experiments, cells were grown in glucose for various times and then analyzed for the occurrence of unprocessed Kar2p after 10 min of labeling in vivo (Fig. 3B). After an 8-h repression of SPC3, precursors of Kar2p accumulated in the cells. As this strong defect in protein processing is preceding the decrease in growth rate, it is likely to be a direct effect of Spc3p depletion.
To test this assumption further, we analyzed cells harvested at
different time points after shifting to glucose for signal peptidase
activity using an in vitro assay (see "Experimental Procedures"). The amount of material used for the assay was
calibrated by immunoblotting using anti-Sec62p-antibodies. Signal
peptidase activity started decreasing after 6 h, a time point
where repressed cells still have about the same growth rate as the
control. After 15 h, signal peptidase activity of the repressed
cells was about 10 times lower than that of DF5 control cells (Fig.
4A). In addition, the integrity of the yeast
SPC during the Spc3p depletion was monitored by immunoblotting (Fig.
4B). Surprisingly, we found that, parallel to the loss of
Spc3p also, the amount of Spc1p and Sec11p decreased below the
detection level. The content of Spc2p decreased also, but in contrast
to the other subunits, the amount of Spc2p appears to stabilize at a
lower expression level. It should be noted that we never observed a
complete loss of signal peptidase activity. Whether this is due to
residual amounts of SPC that could not be detected by our antibodies or
whether another protease with a substrate specificity similar to SPC
was present in the samples (e.g. the mitochondrial inner
membrane protease, a homolog of bacterial leader peptidase) remains
open.
Phenotypes Caused by the Depletion of Spc3p Cannot Be Rescued by Sec11p Overexpression
Sec11p is known to be an essential yeast
protein. Therefore, one may speculate that the loss of signal peptidase
activity in response to the down-regulation of Spc3p is mainly due to
the simultaneous loss of Sec11p. To test this, we tried to rescue the
spc3 mutant by overexpression of Sec11p. SEC11
was put under control of the Met3 promotor and introduced into an HMY1
yeast strain, which contained the plasmid pHM1 with SPC3
under control of a Gal10 promoter. As a control, the Met3Sec11
containing plasmid was also introduced into the
SEC11-deficient strain HMY2. Growth of these strains was
then tested under conditions that induce the Met-promotor of Sec11p on
agar plates either containing galactose or glucose (Fig.
5A). We found, that Sec11p alone is not able to rescue the depletion of Spc3p although Sec11p is strongly
overproduced in these cells (Fig. 5B).
Another possibility is that Sec11p alone is sufficient to function as signal peptidase. In this scenario, Sec11p cannot suppress the depletion of Spc3p because Spc3p performs an essential function, different from the processing of signal peptides. To investigate this possibility, we set up the following experiment. HMY1 cells that contain the two plasmids pHM1 (Gal10Spc3) and pHM4 (Met3Sec11) were cultivated in liquid culture under conditions that were selective for both plasmids and and that induced both the Gal10 and the Met3 promotors. Cells were then shifted to a similar medium, which contained no galactose but glucose as carbon source, to repress the Gal10 promoter. At various time points, samples were taken and analyzed for in vitro signal peptidase activity, and the amount of the different yeast SPC subunits was monitored by immunoblotting (see "Experimental Procedures"). As a control, HMY1 cells containing only the pHM1 were grown using the same protocol with the only exception that uracil was added to the media. Under these conditions, the cells behaved similarly as in the experiments using complete media (Figs. 3 and 4) with respect to growth rate, signal peptidase activity, and stability of the different SPC subunits (data not shown).
As depicted in Fig. 6, we found that also in the Sec11p
overexpressing strain, signal peptidase activity dropped in parallel to
the loss of Spc3p. After 10 h, the cells of this strain contained still as much Sec11p as wild-type cells (Fig. 6B), but the
signal peptidase activity was about 5 times lower (Fig. 6A),
reaching an expression level that is almost identical to the level
observed in the control strain that does not overexpress Sec11p (data
not shown). This indicates that Sec11p alone is not sufficient to perform signal peptide cleavage.
We have identified the yeast homolog of canine SPC22/23, Spc3p, which is identical with the last uncharacterized subunit of yeast SPC. Homologs of the other subunits of canine SPC have been identified in yeast earlier (22, 25, 26), and it was shown that they are part of a heterotetrameric complex (21, 22). Therefore, the principle structure of the eucaryotic SPC turns out to be conserved from yeast to mammals.
Spc3p is essential for cell viability. Its depletion results in
accumulation of precursors of secretory proteins in vivo and loss of signal peptidase activity in vitro, a phenotype that
is very similar to that observed in sec11 mutants
(25).2 Moreover, the depletion of Spc3p
destabilizes the SPC. While the Spc1p and the Sec11p are not detectable
in the ER in the absence of Spc3p, Spc2p seems to be more stable.
The Spc3p may be directly involved in the cleavage of signal peptides.
This is mainly indicated by the tight coupling of the loss of signal
peptidase activity with the loss of Spc3p. Although Spc3p depletion
leads to a simultaneouse loss of Sec11p, the overexpression of Sec11p
is not able to rescue spc3 cells. Moreover, depletion of
Spc3p in a strain overexpressing Sec11p revealed that, even in the
presence of normal amounts of Sec11p, cells lose signal peptidase
activity if Spc3p is absent. Therefore, it is unlikely that the
decrease in signal peptidase activity observed during Spc3p depletion
is caused by the parallel loss of Sec11p.
It should be noted, that a simultaneous overexpression of Sec11p and Spc3p did not result in a significantly higher signal peptidase activity as measured in vitro. Although Spc1p and Spc2p are not essential in vivo, a limiting amount of one or both of these proteins could be the reason for this observation, as the dimeric subcomplex between Spc3p and Sec11p might be less stable under the conditions used for the in vitro assay than the complete SPC. Alternatively, it may well be that other parts of the translocon are needed for a correct assembly and for stabilization of the SPC.
These findings shed a new light on the question of to what extent the eucaryotic signal peptidase of the ER and bacterial leader peptidase may be similar to each other. Signal peptidase activity of the bacterial plasma membrane is comprised of a single polypeptide, the leader peptidase. It has been shown that in contrast to this, the eucaryotic signal peptidase consists of several subunits that substantially differ from each other. However, the identification of regions in the bacterial leader peptidase and the eucaryotic SPC18, SPC21, and Sec11p, which display primary sequence similarity, raised the question of whether all of the eucaryotic subunits are directly involved in the cleavage of signal peptides (17, 24). This notion has been supported by two other facts. 1) Enzymatically active signal peptidase may be purified from hen oviduct as a complex of only two subunits (SPC18 and gp23) (19), and 2) neither the SPC12 homolog, Spc1p, nor the SPC25 homolog, Spc2p, is essential for viability or signal peptidase activity in S. cerevisiae. The identification of Spc3p as a second essential component of yeast SPC described above demonstrates that there must be indeed a substantial difference in the mode of signal peptide cleavage in eucaryotes and procaryotes.
What could be the precise function of Spc3p in this process? It could well be, that the protein provides only a backbone for Sec11p to fold correctly. However, Spc3p could also be more directly involved in the proteolytic process. First, experiments indicate that the hydrophobic region at the carboxyl terminus that is conserved among all SPC22/23 homologs may be essential for the in vivo function of Spc3p.2 This domain could be involved in substrate binding, either alone or in combination with the similarly located second hydrophobic domain of the SPC18/SPC20/Sec11p subunit as proposed earlier by Shelness et al. (12). It may even be that the SPC22/23 homolog is part of the catalytic center of the enzyme. One major way to answer this question will be the determination of the crystal structure of the Sec11p/Spc3p subcomplex.
Why did this complex structure of SPC evolve in eucaryotes? One possibility is that the ER signal peptidase has a broader substrate specificity than the bacterial enzyme. For instance, in contrast to data reported for E. coli leader peptidase (33), we did not find that yeast SPC activity is blocked in vivo if the substrate contains a proline immediatly after the cleavage site.3 Another reason could be that the SPC is involved in other proteolytic processes like the digestion of signal peptides (34) or the degradation of membrane proteins (35). It is also possible that protein translocation across the ER is more regulated than the related process in bacteria and that some SPC subunits are involved in this function. In this regard, it will be of interest to investigate also Archea signal peptidase as the possible ancestor to eucaryotic signal peptidase. To date, a gene having higher homology to eucaryotic SPC18/SPC20/Sec11p than to bacterial leader peptidases has been identified in Methanococcus jannaschii (36),4 and whether homologs of the other SPC subunits are also present in Archea is still open.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) U92975[GenBank].
We thank Francoise Kepes, Thomas Sommer, and Steffen Panzner, for providing antibodies, strains, and plasmids, Brigitte Nentwig and Angelika Wittruck, for technical assistance, and T. Sommer and the members of the Hartmann lab, for critical reading of the manuscript.