(Received for publication, January 13, 1997, and in revised form, March 17, 1997)
From the Department of Microbiology and Immunology, School of Medicine, Vanderbilt University, Nashville, Tennessee 37232-2363
Among the three characterized subunits comprising the signal peptidase complex of the yeast Saccharomyces cerevisiae (Sec11p, Spc1p, and Spc2p), only Sec11p is essential for cell growth, signal peptide cleavage, and signal peptidase-dependent protein degradation. Here we report the cloning of the SPC3 gene encoding the homolog to mammalian signal peptidase subunit SPC22/23. We find that Spc3p is also required for cell growth and signal peptidase activity within the yeast endoplasmic reticulum.
Amino-terminal signal sequences of proteins targeted to the endoplasmic reticulum (ER)1 are cleaved by a membrane-bound endoprotease termed signal peptidase (1, 2). This enzyme has also been shown to catalyze the proteolytic fragmentation of some abnormal membrane proteins, thereby leading to their degradation (3, 4). Isolated from the yeast Saccharomyces cerevisiae, ER signal peptidase (SP) contains four nonidentical protein subunits (5-7). Three of the subunits of this signal peptidase complex (SPC) have been functionally examined. Sec11p (17 kDa) is required for signal peptide cleavage (8) and signal peptidase-dependent protein degradation (4). Sec11p is related to two subunits of the mammalian SPC (SPC18 and SPC21) (9, 10). In contrast to Sec11p, the Spc1p (11 kDa) and Spc2p (18 kDa) subunits of the yeast SPC are nonessential for cell growth and enzyme activity (6, 7). Spc1p and Spc2p, however, perform auxiliary and nonredundant roles. Spc1p facilitates signal peptide cleavages in cells burdened with high levels of a membrane protein substrate of the signal peptidase-dependent protein degradation pathway (6). Spc2p is important for enzyme activity and cell viability at elevated temperatures (7). Spc1p and Spc2p are homologous to mammalian subunits SPC12 and SPC25, respectively (6, 7, 11, 12).
Despite the fact that a multisubunit signal peptidase has been purified from yeast and mammalian cells, enzymes exhibiting less subunit complexity have been identified in other systems. Leader peptidase from the inner membrane of E. coli consists of a single polypeptide chain (13). This protein exhibits limited homology to Sec11p and to mammalian SPC18 and SPC21 (14, 15). Signal peptidase purified from hen oviduct contains two subunits, one related to Sec11p and a second related to SPC22/23 of the mammalian SPC (16-19).
In the present study, we have cloned and characterized the SPC3 gene encoding the yeast homolog to mammalian SPC22/23. We find that, as with Sec11p, Spc3p is essential for signal peptide cleavage and signal peptidase-dependent protein degradation. Our data are thus in agreement with in vitro studies, using an avian system, which show that a two-subunit complex functions to cleave signal peptides. The idea that both Sec11p and Spc3p are related to the prokaryotic signal peptidase is also discussed.
Plasmid pSPC3 bearing SPC3 was isolated from a high copy (2 µm) plasmid library marked with LEU2 (American Type Culture Collection ATCC No. 37323). Among 5,500 transformants of strain CMYD1 (sec11-7) (genotypes of yeast strains are listed in Table I), pSPC3 was isolated from one colony that grew at 32 °C. The SPC3 gene was amplified from pSPC3 by a polymerase chain reaction (PCR) using the forward oligonucleotide primer CGGGATCCACACGTGAATACTACC, which is located 204 bp upstream of the SPC3 start codon, and the reverse primer CGGAATTCAATAAATGGGAACAG, which is located 189 bp downstream of the SPC3 stop codon. The amplified fragment was restricted with BamHI and EcoRI and inserted into low-copy (CEN) plasmid pRS314 (TRP1) (21). The resulting plasmid was named pHF332. A 1.5-kb HindIII-XbaI restriction fragment containing the SPC3 gene was excised from plasmid pSPC3 and inserted into 2 µm plasmid pRS426 (URA3) (22). The resulting plasmid was named pHF331.
|
The 1.5-kb HindIII-XbaI restriction fragment containing the SPC3 gene was inserted into the polylinker of pUC19 (23). The resulting plasmid was restricted with NheI, which cuts 30 bp upstream of the SPC3 open reading frame, and SpeI, which cuts 3 bp downstream of the stop codon of SPC3, and the SPC3 gene was replaced with a 1.6-kb NheI DNA fragment containing the LEU2 gene. This NheI fragment was obtained from a modified YDp-L plasmid (24) that contains LEU2 flanked by two NheI sites (a NheI linker was inserted at the SmaI site of YDp-L). This construct containing a replacement of SPC3 with the LEU2 gene was restricted with HindIII and XbaI and transformed into diploid strain SEY6210.5 with selection for leucine prototrophs. Transformants bearing a replacement of the SPC3 gene with the LEU2 gene were confirmed by genomic PCR using oligonucleotide primers described above.
Construction of Mutations in SPC3For construction of a frameshift mutation in SPC3, the single BglII site of pHF331 (located at codons 80-82 of SPC3) was restricted and then filled in with the Klenow fragment of DNA polymerase and ligated.
For construction of temperature-sensitive mutations, the plasmid pSPC3
was used as a template for PCR under conditions that favor
misincorporation of dNTPs by Taq polymerase (25). The above
described oligonucleotide primers located upsteam and downstream of the
SPC3 gene were used. After 30 cycles of DNA amplification, the PCR product was restricted with BamHI and
EcoRI and cloned into the polylinker of plasmid pRS314
(CEN6 TRP1). The resulting plasmids carrying a pool of
mutagenized SPC3 genes were transformed into strain HFY405
(spc3)/pHF331 (2 µm SPC3 URA3), and
transformants were subjected to the plasmid shuffle technique (26).
From 200 transformants, four temperature-sensitive mutants containing
plasmids pHF333 (CEN6 spc3-1 TRP1), pHF334 (CEN6
spc3-2 TRP1), pHF335 (CEN6 spc3-3 TRP1), and pHF336
(CEN6 spc3-4 TRP1) were isolated. To confirm that the
temperature-sensitive phenotype was plasmid linked, each of the
plasmids was purified and transformed into strain HFY405/pHF331, using
plasmid shuffling to replace pHF331 with each of the transformed
plasmids.
The spc3-4 mutation was integrated into the SPC3
chromosomal locus using methods described previously (27). The
BamHI-EcoRI fragment containing the
spc3-4 gene from pHF336 was inserted into plasmid pRS306
(URA3) (22). The resulting construct was restricted with
NheI. Linear DNA was transformed into haploid strains
SEY6210 and SEY6211. A Ura+ transformant of each strain was
placed on a minimal agar plate supplemented appropriately (28) and
containing 5-fluoroorotic acid (1 mg/ml). After several days of
incubation at 23 °C, colonies were picked and tested for growth at
37 °C. From this analysis, temperature-sensitive strains were
detected at a frequency of ~50%. Plasmid pSPC3 was introduced into
representative temperature-sensitive strains to test for
complementation of the temperature-sensitive growth defect. The
resulting temperature-sensitive strains used in this study were named
HFY406 (MAT spc3-4) and HFY407 (MATa spc3-4).
In a previous study, we observed that overexpression of
Spc1p suppressed a mutational defect in Sec11p (6). A similar approach was thus employed in a search for the yeast homolog to mammalian SPC
subunit SPC22/23 (Experimental Procedures). In this analysis, pSPC3 was
isolated from a high-copy (2 µm) plasmid library through the ability
of this plasmid to suppress the growth defect of strain CMYD1
(sec11-7) (Fig. 1A). As shown
below, the plasmid pSPC3 bears the gene encoding the yeast homolog of
SPC22/23. Fig. 1A also depicts the growth of strain CMYD1
carrying high copy plasmid pSPC1, which contains the SPC1
gene (6). The data show that pSPC1 suppressed the sec11
mutation at a higher temperature than pSPC3.
DNA sequencing at one end of the chromosomal fragment contained within pSPC3 revealed a match to a sequence on chromosome XII. The identified sequence was that of PET309, which encodes a protein required for stability and translation of COX1 mitochondrial mRNA (29). Among the open reading frames present near PET309, we identified an uncharacterized gene that we have named SPC3 (accession no. Z73238[GenBank]).2 To determine whether SPC3 suppressed sec11-7, a frameshift mutation was constructed at codon 82 of the SPC3 gene (see "Experimental Procedures"). The high copy plasmid bearing frame-shifted SPC3 did not suppress the growth defect of strain CMYD1 (sec11-7) at 32 °C. The control plasmid containing wild-type SPC3 suppressed the sec11-7 mutation.
To determine if SPC3 also suppresses the signal peptidase defect seen in sec11 mutant cells, strains CMYD1 (sec11) and CMYD1/pSPC3 were subjected to a pulse-chase analysis using methods described previously (4, 28). Cells were grown to log phase at 23 °C, shifted to 32 °C for 30 min, pulse labeled for 5 min with 35S-EXPRESS, and chased for 30 min in the presence of excess unlabeled methionine and cysteine, and proteins were precipitated from cell extracts with antibodies directed against ER resident Kar2p (30, 31). As shown in Fig. 1B, the precursor of Kar2p was converted to mature Kar2p to a greater extent in sec11 mutant cells containing pSPC3 than in sec11 cells lacking pSPC3 (compare lanes 1 and 3). Overexpression of the SPC3 gene thus partially corrected the signal peptide cleavage defect in the sec11 mutant.
Primary amino acid sequence analysis revealed that the product of the
SPC3 gene, Spc3p, exhibited 55% similarity and 22%
identity to SPC22/23 of the mammalian SPC (Fig. 2).
These measurements are similar to the homologies shared between Spc1p
and Spc2p and their mammalian counterparts (6, 7). Spc3p is also
related to the SPC22/23 homolog (18) present in the two-subunit avian SPC (54% similarity, 24% identity). Spc3p is nonhomologous, however, to Sec11p, Spc1p, and Spc2p. Hydropathy analysis of Spc3p and SPC22/23
(32) identified a single hydrophobic stretch of amino acids located
near the amino termini of the respective proteins. This region in
SPC22/23 has, indeed, been shown to span the ER membrane (33). The
calculated molecular masses of SPC22/23 (20.3 kDa) and Spc3p (21.3 kDa)
as well as the isoelectric points, 8.8 for SPC22/23 and 8.6 for Spc3p,
are similar.
SPC3 Is Required for Cell Growth
Heterozygous disruption of the SPC3 gene was accomplished in diploid strain SEY6210.5 through replacement of SPC3 with the LEU2 marker (see "Experimental Procedures"). Genomic PCR analysis was used to confirm the heterozygous disruption of SPC3 in Leu+ transformants of strain SEY6210.5 (data not shown). The heterozygote was sporulated, and the spores generated were examined using tetrad analysis. Dissection of 10 tetrads revealed only two viable spores from each tetrad. All recovered strains were leucine auxotrophs, suggesting cells bearing the spc3 null mutation were inviable.
The requirement of SPC3 for cell growth was confirmed by introducing plasmid pHF331 (SPC3 URA3) into above described strain SEY6210.5 containing a heterozygous disruption of the SPC3 gene. Tetrad analysis of this transformant yielded four viable spores from each of five tetrads examined. Two spores from each tetrad were Leu+ Ura+. Colonies were not recovered upon placement of the Leu+ Ura+ cells on agar plates containing 5-fluoroorotic acid, a chemical used to select for yeast cells cured of plasmids bearing the URA3 gene (34).
Spc3p Appears to Function Directly in the Signal Peptide Cleavage ReactionThe requirement of SPC3 for cell growth was
circumvented in further biochemical studies through the generation of
temperature-sensitive (ts) mutations in the SPC3 gene. Using
a PCR-based method of nucleotide misincorporation (see "Experimental
Procedures"), mutant alleles spc3-1, spc3-2, spc3-3,
and spc3-4 were generated. The spc3-3 mutant
failed to grow at 32 °C, whereas the nonpermissive temperature for
the remaining mutants was 37 °C. We screened cells containing these
mutations in an attempt to identify at least one allele that rapidly
inhibited Spc3p function following a shift of mutant cells to their
nonpermissive temperatures. Strain HFY405 (spc3) bearing
a series of plasmids containing these mutations (plasmids are described
in "Experimental Procedures") was grown to log phase at 23 °C,
then shifted to 37 °C for the time intervals indicated in Fig.
3. Cell extracts were examined by Western blotting using anti-Kar2p antibodies. Significant amounts of preKar2p accumulated in
the spc3-3 and spc3-4 mutants after a 30-min
incubation at their nonpermissive temperature (lanes 8-9).
Increasing levels of preKar2p appeared in these cells after 1 h
(lanes 12 and 13) and 2 h (lanes
16 and 17). In contrast, the spc3-1 and
spc3-2 mutants exhibited a modest defect in the cleavage of
preKar2p even after 2 h at the nonpermissive temperature
(lanes 14 and 15). Furthermore, the nonpermissive
temperature for the spc3-4 mutant (37 °C) was higher
than the sec11-7 mutant (32 °C), a difference that made
it possible to follow both mutations in genetic crosses. For these
reasons, the spc3-4 mutation was chosen for further study.
A pulse labeling analysis (4, 28) was employed to more accurately
measure signal peptide cleavage in the spc3-4 mutant. Strain HFY405 (spc3) bearing pHF336 (CEN6 spc3-4
TRP1) or control plasmid pHF332 (CEN6 SPC3 TRP1) was
grown to log phase at 23 °C and then shifted to 37 °C for 5 min.
Cells were pulse labeled, and proteins were precipitated from cell
extracts with anti-Kar2p antibodies. The sec11 mutant was
examined for use as a control. The data in Fig. 4
demonstrate that cells bearing the wild-type SPC3 gene
displayed efficient cleavage of preKar2p (lane 1). In contrast, the spc3 (lane 2) and sec11
(lane 3) mutations strongly inhibited preKar2p cleavage at
the nonpermissive temperature. The fact that this defect was present in
the spc3 mutant after shifting cells to 37 °C for only 5 min argues in favor of a direct effect of the spc3 mutation
on the signal peptidase reaction.
The results presented thus far do not eliminate the possibility that
the spc3-4 mutation inhibited protein translocation, thereby resulting in an accumulation of preKar2p in the cytoplasm. To
test this possibility, strain HFY405 (spc3)/pHF336
(spc3-4) was grown to log phase at the permissive
temperature and shifted to 37 °C for 1.5 h, and membranes were
prepared using methods described previously (4). The membrane fraction
was treated with proteinase K in the presence and absence of detergent
and subjected to Western blotting with anti-Kar2p antibodies. The results depicted in Fig. 5A demonstrated that
preKar2p was protected from proteolytic digestion in extracts lacking
detergent (lanes 3-7) but exposed to proteolytic attack in
extracts incubated with detergent (lanes 8-12). A
proteolytic fragment, Kar2f, was also present in this analysis. Kar2f
has been seen previously in proteinase K digests of Kar2p (4).
Together, these data indicate that the spc3-4 mutation did
not inhibit protein translocation.
To corroborate this result, vacuolar protein carboxypeptidase Y (CPY)
(35) was examined. CPY was chosen for this analysis because
untranslocated preproCPY (ppCPY) migrates differently from the
translocated form on SDS-PAGE gels due to core glycosylation of this
protein in the ER lumen. The Western blot shown in Fig. 5B
reveals significant amounts of ppCPY in the temperature-sensitive sec61 mutant (lane 5), which is known to exhibit
a defect in protein translocation (36). In contrast, mutant strain
HFY405 (spc3)/pHF336 (spc3-4) exhibited ppCPY
levels that were similar to the levels present in cells containing
wild-type SPC3 (compare lanes 1 and 2). A distinct form of CPY (designated CPY*) was present in
the spc3-4 mutant at the nonpermissive temperature
(lane 1). CPY* comigrated with a form of CPY found in the
sec11 mutant (lane 3). In addition, CPY* migrated
similarly to a form of CPY found in the sec23 mutant
(lane 4). The sec11 (8) and sec23 (37) mutations inhibit signal peptide cleavage and protein transport to the
Golgi apparatus, respectively. CPY*, therefore, probably represents
aberrant glycosylation of ppCPY resulting from a defect in signal
peptidase activity.
We have previously demonstrated that chimeric
membrane protein AHDK2 is fragmented in a Sec11p-dependent
manner and that the fragmentation of AHDK2 by signal peptidase is
necessary for its proteolytic elimination (4). AHDK2 is anchored to the
ER membrane by a transmembrane segment from arginine permease. Fused to
the luminal side of this transmembrane segment is a fragment of
histidinol dehydrogenase (HD), which is followed by the Kar2p moiety.
The presence of Kar2p at the C terminus of AHDK2 permits detection of
proteolytic intermediate f1, which is produced through a cleavage near
the transmembrane segment of AHDK2. To determine whether Spc3p is
important for cleavage at the f1 site of AHDK2, expression plasmid
pAHDK2 (URA3) (4) was introduced into strain HFY405 (spc3)/pHF336 (CEN6 spc3-4 TRP1). This strain
and control strain HFY405 (
spc3) bearing plasmids pAHDK2
and pHF332 (CEN6 SPC3 TRP1) were grown to log phase at
23 °C, shifted to 37 °C for 30 min, then analyzed by pulse-chase
using anti-HD antibodies. As expected, cells containing the wild-type
SPC3 gene displayed AHDK2 and proteolytic fragment f1 during
the pulse-chase analysis (Fig. 6, lanes
1-3). In contrast, no apparent f1 was produced in spc3
mutant cells (lanes 4-6).
The spc3-4 Mutation Interacts in a Synthetic Manner with sec11-7 and
The sec11-7 mutation
displays a genetic interaction, termed synthetic lethality, with the
spc1 and
spc2 mutations (6, 7).
Specifically, the growth of cells containing the sec11 mutation and either the spc1 or spc2 null
mutation is inhibited more severely than the growth of cells containing
these mutations individually. We therefore asked whether the
sec11-7 mutation is synthetically lethal with the
spc3-4 mutation. To avoid the problem of loss of plasmid
pHF336 (spc3-4) in the analysis of yeast tetrads, the
chromosomal SPC3 gene locus was replaced by the
spc3-4 mutation (see "Experimental Procedures").
Analysis of a spc3 sec11 double mutant was initiated by
crossing strains HFY406 (spc3-4) and CMYD1
(sec11-7). These mutations were followed in the analysis of
tetrads by the ability of the spc3-4 and
sec11-7 mutations to inhibit cell growth at 37 and
32 °C, respectively. The distribution of progeny in 18 tetrads that
were examined at 23 °C was as follows: 14 tetratype (one
spc3 mutant spore, one sec11 mutant spore, one
wild-type spore, and one inviable spore), three nonparental ditype (two
wild-type and two inviable spores), and one parental ditype (two
sec11 and two spc3 mutant spores). An additional
three tetrads containing only two viable spores were produced.
Phenotypic analysis of these spores was inconsistent with the
nonparental ditype. Therefore, some of the spores in these tetrads
probably succumbed to random spore death. The data derived from this
analysis demonstrated that the spc3-4 sec11-7 double
mutant was inviable at 23 °C, temperature permissive for the
spc3-4 and sec11-7 single mutants.
We next questioned whether the spc3-4 mutation was
synthetically lethal with the spc1 and
spc2
mutations. Random spore analysis of diploid cells derived from a cross
between strains HFY407 (spc3-4) and HFY401
(
spc1::TRP1) produced viable spores that were
both tryptophan prototrophs and temperature-sensitive (frequency of 4/14). This indicated the spc3-4
spc1 double
mutant grew at 23 °C. The double mutant also grew well at 32 °C,
indicating no apparent genetic interaction between the
spc3-4 and
spc1 mutations under the tested
conditions. In contrast, analysis of diploid cells derived from strains
HFY407 (spc3-4) and CMY195
(
spc2::URA3) yielded progeny with the following
distribution: 11 tetratype (one Ura+, one
Ura
, one ts/Ura
, and one inviable spore),
two nonparental ditype (two Ura
and two inviable spores),
and two parental ditype (two ts/Ura
, and two
Ura+ spores). Random spore death probably led to one tetrad
containing one Ura+, one Ura
and two inviable
spores and a second tetrad containing 2 ts/Ura
, one
Ura+, and one inviable spore. Importantly, the absence of
any spore from this dissection that was both ts and Ura+
indicates that the spc3-4
spc2 double mutant
was inviable at 23 °C.
These results thus make a distinction between Spc1p and Spc2p, both of
which are nonessential for signal peptidase activity (6, 7). The
spc2 mutation but not the
spc1 mutation is synthetically lethal with the spc3-4 mutation. In addition,
these data suggest a difference may exist between Sec11p and Spc3p, both of which are essential for signal peptidase activity. The sec11-7 mutation is synthetically lethal with the
spc1 and
spc2 mutations, whereas the
spc3-4 mutation is synthetically lethal with
spc2 but not
spc1. One interpretation of
these results is that Spc3p may physically interact with Spc2p but not
Spc1p; however, it is also plausible that allele-specific mutations in the SPC3 gene may be found that synthetically interact with
the
spc1 mutation.
Sec11p has been shown previously to function directly in the cleavage of signal peptides and the fragmentation of abnormal membrane proteins in the yeast Saccharomyces cerevisiae (4, 8). Results described here indicate that Spc3p is also required for these enzymatic activities. Moreover, Spc3p and Sec11p interact genetically in that overexpression of Spc3p suppresses the sec11-7 mutation, and the spc3-4 and sec11-7 mutations are synthetically lethal. Importantly, the fact that the conditional spc3-4 mutation inhibits signal peptidase activity after a brief shift to its nonpermissive temperature supports the notion that Spc3p also directly participates in the signal peptide cleavage reaction.
In contrast to Spc3p and Sec11p, Spc1p and Spc2p are nonessential for enzyme activity (6, 7). The mammalian homologs of Spc1p and Spc2p contain two transmembrane segments, an orientation that positions the majority of their sequences on the opposite side of the ER membrane from the active site of signal peptidase (12). Conversely, the mammalian homologs of Sec11p and Spc3p are type II single spanning membrane proteins containing a large luminal domain (33). Since Spc3p and Sec11p function directly in the signal peptide cleavage reaction, it seems likely that the SPC subunits predominately exposed to the lumen comprise a two-subunit core enzyme that is required for signal peptidase activity. This conclusion is supported by the fact that a two-subunit avian signal peptidase containing homologs to Spc3p and Sec11p has been shown to function in the signal peptide cleavage reaction in vitro (16).
The conservation of the signal peptide cleavage reaction throughout
evolution is revealed by the fact that signal sequences are cleaved
correctly when expressed in either eukaryotic or prokaryotic cells.
Despite this fact, signal peptides are cleaved by a single polypeptide
chain, leader peptidase, in E. coli (13). While the overall
similarity between Sec11p and leader peptidase is rather weak, it has
been noted that the periplasmic domain of leader peptidase contains
three distinct regions of homology to the proposed luminal domain of
Sec11p (termed Box I, II, and III) (14, 15). Box I contains a serine
amino acid found in all proteins of the signal peptidase I family (Fig.
7A). This residue is the only serine in
leader peptidase required for enzyme activity (39). Box II contains a
lysine residue that is essential for leader peptidase function (40)
although the corresponding amino acid in Sec11p is histidine. These
serine and lysine residues are proposed to constitue a catalytic dyad
within leader peptidase, which is similar to the active site of
LexA-type proteases (14, 38, 40). Only four amino acids in Box III of
leader peptidase are present in Sec11p (39, 40).
Considering that Spc3p and Sec11p are nonhomologous to each other, we suggest that these subunits function together in a manner analogous to leader peptidase. We have performed a comparison of leader peptidase and Spc3p sequences, using the computer-based algorithm described previously (41, 42). As shown in Fig. 7B, this analysis reveals a similarity between the luminal domain of Spc3p and the carboxyl-terminal half of the periplasmic domain of leader peptidase. The Box I and Box II amino acids are located in the amino-terminal half of leader peptidase, thus suggesting the alignment of Sec11p and Spc3p sequences to distinct regions of leader peptidase. Moreover, the number of amino acids present in the luminal domain of leader peptidase (~250 amino acids), which contains the catalytic site, is similar to the number of amino acids in the proposed luminal domains of both Sec11p (~135 amino acids) and Spc3p (~150 amino acids). These correlations raise the intriguing possibility that Sec11p and Spc3p evolved from a common ancestor. However, in light of the modest homologies between these proteins, we cannot exclude the possibility that leader peptidase and the SPC utilize fundamentally distinct mechanisms for cleaving signal peptides. Future experiments to test these two models are in progress.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) Z73238[GenBank].
We thank Dean Ballard (Vanderbilt University) for critical reading of this manuscript and Randy Sheckman (University of California, Berkeley) for the gift of anti-CPY antibodies.