The Catalytic Mechanism of Endoplasmic Reticulum Signal Peptidase Appears to Be Distinct from Most Eubacterial Signal Peptidases*

Many type I signal peptidases from eubacterial cells appear to contain a serine/lysine catalytic dyad. In contrast, our data show that the signal peptidase complex from the endoplasmic reticulum lacks an apparent catalytic lysine. Instead, a serine, histidine, and two aspartic acids are important for signal peptidase activity by the Sec11p subunit of the yeast signal peptidase complex. Amino acids critical to the eubacterial signal peptidases and Sec11p are, however, positioned similarly along their primary sequences, suggesting the presence of a common structural element(s) near the catalytic sites of these enzymes.

Many type I signal peptidases from eubacterial cells appear to contain a serine/lysine catalytic dyad. In contrast, our data show that the signal peptidase complex from the endoplasmic reticulum lacks an apparent catalytic lysine. Instead, a serine, histidine, and two aspartic acids are important for signal peptidase activity by the Sec11p subunit of the yeast signal peptidase complex. Amino acids critical to the eubacterial signal peptidases and Sec11p are, however, positioned similarly along their primary sequences, suggesting the presence of a common structural element(s) near the catalytic sites of these enzymes.
Cleavable signal sequences, which are usually located at the N termini of precursor proteins, function in the delivery of their protein cargo to specific destinations within both eukaryotic and eubacterial cells. A large number of signal sequences possess a common motif consisting of an n-region that is often positively charged and a hydrophobic core (the h-region) followed by a polar c-region containing the cleavage site (1). Small uncharged amino acids are usually present at the Ϫ1 and Ϫ3 positions from the cleavage site. Signals exhibiting this motif are recognized and cleaved by type I signal peptidases. Type I signal peptidases are found within the endoplasmic reticulum (ER) 1 membrane, the mitochondrial inner membrane, and the cytoplasmic membrane of eubacterial cells (reviewed in Ref. 2). There is a strong functional conservation of the signal sequence cleavage reaction, revealed by the fact that signal sequences of proteins targeted normally to the ER can be cleaved by eubacterial signal peptidase (3), and eubacterial signal sequences can be cleaved by ER signal peptidase (4).
Site-directed mutagenesis studies suggest that many eubacterial signal peptidases contain a serine/lysine dyad with which to catalyze the cleavage reaction (5)(6)(7)(8)(9). A similar catalytic dyad is thought to be present in the LexA and UmuD proteins of Escherichia coli (10,11) and in both catalytic subunits of mitochondrial signal peptidase (12). Recent x-ray crystallographic analysis confirms the role of serine as a nucleophile and is consistent with a lysine acting as a general base in catalysis by E. coli leader peptidase (13).
At least one eubacterial signal peptidase, SipW of Bacillus subtilus, may exhibit a catalytic site more like that of Sec11p of the ER signal peptidase. B. subtilus contains five distinct chromosomally encoded signal peptidases, SipW being the only one like Sec11p (14,15). As shown in Fig. 1, Sec11p contains a serine residue that aligns to the catalytic serine of leader peptidase; however, within the limited regions of homology that exist between Sec11p and leader peptidase, Sec11p contains a histidine that has been aligned to the catalytic lysine of leader peptidase (16,17). From this, the type I signal peptidase family may contain a subgroup, represented by Sec11p and SipW, that uses a distinct catalytic mechanism (14). It has been noted previously, however, that an alignment of Sec11p to the E. coli leader peptidase sequence is of low statistical significance (5). In addition, two subunits, Sec11p and Spc3p, of the ER signal peptidase are essential in the yeast Saccharomyces cerevisiae (18 -22), thus raising the possibility that Spc3p may contribute amino acids to the catalytic site.
To identify the probable catalytic site residues of ER signal peptidase and thus explore the evolutionary relationship of the type I signal peptidase family, all serine, lysine, histidine, and aspartic acid residues that are conserved among Sec11p-and Spc3p-type subunits have been changed using site-directed mutagenesis. We have found that Ser-44, His-83, Asp-103, and Asp-109 are essential for signal peptidase activity by Sec11p. Because all conserved and partially conserved lysines were nonessential, this work strengthens the notion that the type I signal peptidase family is comprised of two groups, one containing a serine/ lysine dyad and one that appears to lack a catalytic lysine.
Site-directed Mutagenesis-All site-directed mutagenesis reactions were performed using the Muta-Gene® in vitro mutagenesis kit (Bio-Rad) with the following modifications. Oligonucleotides were phospho-rylated following the directions from the Sculptor® in vitro mutagenesis kit (Amersham Pharmacia Biotech) except that T4 DNA ligase buffer was added instead of T4 polynucleotide kinase buffer, and the reaction time was extended to 30 min. After production of phage, the steps for single-stranded DNA synthesis outlined in the Sequenase™ Version 2.0 kit (U. S. Biochemical Corp.) were followed. Oligonucleotides used for mutagenesis of SEC11 (Table I) and SPC3 (Table II) are listed.
Enrichment of ER Membranes-Preparation of ER membranes was based on work done previously (27,28). Two thousand A 600 yeast cell equivalents were resuspended in distilled water (50 ml), pelleted, resuspended in 100 ml of zymolase buffer (1.4 M sorbitol, 50 mM potassium phosphate (pH 7.5), 40 mM ␤-mercaptoethanol, and 50 g/ml zymolase), and agitated gently for 65 min. The resulting spheroplasts were subjected to centrifugation at 3000 ϫ g for 5 min and resuspended gently in 50 ml of 1.4 M sorbitol. The rest of the manipulations were performed on ice. The spheroplasts were pelleted at 3000 ϫ g for 5 min and resuspended in 3 ml of HS buffer (20 mM KOH-HEPES (pH 7.5), 500 mM sucrose, 2 g/ml phenylmethylsulfonyl fluoride, and a protease inhibitor mixture consisting of 1 g/ml each of pepstatin A, chymostatin, aprotinin, leupeptin, and antipain). Homogenization was performed in the cold room using a Potter homogenizer (10 strokes). The homogenized spheroplasts were subjected to centrifugation at 10,700 ϫ g for 10 min. The supernatant was saved, and the pellet was subjected to a second homogenization and centrifugation using 3 ml of HS buffer. The supernatants from the two homogenizations were combined and subjected to centrifugation at 10,700 ϫ g. The final supernatant was overlaid on a discontinuous sucrose gradient containing 1 ml of 70% sucrose, 2 ml of 48% sucrose, 2 ml of 39% sucrose, and 1 ml of 30% sucrose. The sucrose was dissolved (w/w) in a solution containing 50 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 25 mM KCl. The gradient was subjected to centrifugation at 40,000 rpm for 2.5 h in a Beckman SW 41 rotor. ER membranes at the 48/39 interface were collected, mixed with HS (2 ml), and pelleted at 40,000 rpm for 45 min in a Beckman Ti50 rotor.
Epitope Tagging-Spc3p was tagged with the HA epitope (29) as follows. SPC3 was polymerase chain reaction-amplified using upstream primer CGGGATCCATATGTTCTCCTTTGTCCAAAGA, which contains BamHI and NdeI sites, and downstream primer GACAAGGGTA-AATAACTGAATTC, which contains an EcoRI site. The product was inserted between the BamHI and EcoRI sites of pRS314 (CEN6 TRP1) (30). A 1.5-kilobase BamHI fragment containing the ADH1 promoter (31) and a 100-base pair NdeI fragment encoding the HA tag were then inserted into the construct. To generate a plasmid overproducing HA-Spc3p, an EcoRI-SacI fragment encoding HA-Spc3p under control of the ADH1 promoter was removed from the above construct and inserted into pRS426 (32), generating pCV102 (2 m, SPC3(HA) URA3).
Sec11p was tagged with the FLAG epitope (33) as follows. The SEC11 gene and in vitro mutagenized SEC11 were polymerase chain reaction-amplified using the upstream primer GGCGGATCCATGGAC-TACAAGGACGACGATGACAAGGACTACAAGGACGACGATGACAA-GAATCTAAGATTTGAATTGCAGAAACTATTGAAC, which encodes two FLAG epitopes in tandem, and downstream primer CCGGAAT-TCGGCGAACTACTCGCCCCCCAG. The polymerase chain reaction products were digested with BamHI and EcoRI and ligated to identically digested pHF454 (2-m TRP1 ADH1-promoter) (4) to create plasmid pCV101.
Pulse Labeling, Immunoprecipitation, and Western Analyses-Pulse labeling of yeast cells was performed as described (23,26) with specific changes noted in the text and legends. To detect FLAG-tagged Sec11p by immunoprecipitation, enriched ER membranes from 1000 A 600 yeast cell equivalents were sonicated in 60 l of a solution containing 50 mM Tris-HCl (pH 7.5), 50 mM EDTA, and 5 g/ml phenylmethylsulfonyl fluoride. The material was mixed with SDS-PAGE sample buffer (4ϫ concentrate) to a final volume of 80 l and boiled. 30 l of this mixture was diluted with 1 ml of phosphate-buffered saline containing Triton X-100 (1%) as described (26), and proteins were immunoprecipitated using 25 l of a suspension of anti-FLAG antibodies conjugated to agarose. For Western blotting, proteins were transferred to HybondECL nitrocellulose membranes (Amersham Pharmacia Biotech), incubated with anti-FLAG antibodies at 1/6000 dilution for 55 min, then incubated with anti-mouse horseradish peroxidase-conjugated secondary antibody at 1/8000 dilution for 50 min.
To coimmunoprecipitate Sec11p-FLAG and Spc3p-HA from yeast cells, ER membranes from 2000 A 600 units of cells were prepared as described above. Membrane pellets were resuspended via sonication in 125 l of sonication buffer (50 mM KOH-HEPES (pH 7.5), 10% glycerol, 7 mM ␤-mercaptoethanol, and above-described protease inhibitor mixture). This mixture was diluted to 500 l to achieve a final concentration of 0.5 M potassium acetate, 50 mM KOH-HEPES (pH 7.5), 10% glycerol, 16 mM magnesium acetate, 7 mM ␤-mercaptoethanol, and 1% digitonin. The mixture was placed on ice for 30 min, then subjected to centrifugation for 45 min at 40,000 rpm in a Beckman Ti-50 rotor. The supernatant containing solubilized membranes was diluted to 0.2 M potassium acetate by adding 750 l of dilution buffer (50 mM KOH-HEPES, 1% digitonin). Agarose-conjugated anti-FLAG antibodies (30 l) were added to the solution, which was then mixed for 3 h at 4°C. The agarose was sedimented and washed as described (26). Proteins were analyzed by Western blotting using a 1/6000 dilution of anti-HA antibodies and a 1/8000 dilution of anti-rat-conjugated secondary antibodies (Roche Molecular Biochemicals).

One Conserved Serine, One Conserved Histidine, and Two
Conserved Aspartic Acids Are Essential in Sec11p-To identify amino acids important for ER signal peptidase activity, we employed a site-directed mutagenesis approach. We reasoned that catalytic amino acids are probably conserved among various Sec11p-and Spc3p-type subunits. We therefore aligned the sequence of Sec11p to the sequences of the two related subunits from the canine signal peptidase complex, SPC21 and SPC18 (34), and to B. subtilus SipW (14). As shown in Fig. 2A, two serines, one histidine, one lysine, and two aspartic acid residues were found to be conserved among these Sec11p-type proteins. Among the Spc3p-type proteins, one lysine and two FIG. 1. Sequence homology between leader peptidase and Sec11p. Box I, Box II, and Box III represent sequences that are homologous in E. coli leader peptidase and the Sec11p subunit of the yeast ER signal peptidase. The positions of the catalytic serine and lysine of leader peptidase and the corresponding serine and histidine of Sec11p and the aspartic acids important in Sec11p and their corresponding amino acids in leader peptidase are indicated (*). Gaps in these sequences are indicated by dashes. Identical residues are indicated by stacked dots, and similar residues are indicated by single dots. aspartic acids were conserved upon alignment of yeast Spc3p to its canine, chicken, Caenorhabditis elegans, and Schizosaccharomyces pombe counterparts (Fig. 2B) (21).
Mutations that alter amino acids critical for signal peptidase activity are lethal to yeast cells (18,19). A cell growth assay, based on the plasmid shuffle technique (35), was therefore employed to determine which of the amino acids was important for cell viability. In this assay, yeast cells bearing a chromosomal disruption of an essential signal peptidase gene and a plasmid bearing both a wild-type signal peptidase gene and the URA3 gene were used. The URA3 gene is toxic to yeast cells incubated in the presence of 5-fluoroorotic acid (36); however, by introducing a plasmid bearing a mutated form of the signal peptidase gene and a marker other than URA3, cell growth in the presence of 5-fluoroorotic acid indicates that the URA3based plasmid has been cured and that the protein encoded by a mutated signal peptidase gene is functional.
Mutations altering the conserved amino acids in Sec11p were constructed by site-directed mutagenesis and introduced into a TRP1-based plasmid. This series of plasmids and the control vector lacking the SEC11 gene were transformed into yeast strain CMY710 (⌬sec11)/pCM112 (SEC11 URA3). Cells were then subjected to plasmid shuffling. Mutations that alter amino acids Ser-44, His-83, Asp-103, and Asp-109 were found to be lethal (Table III), whereas mutations altering the conserved amino acids Ser-42 and Lys-101 were nonlethal. As expected, the vector control that lacked SEC11 was unable to support growth of strain CMY710 (⌬sec11). To determine whether the amino acids conserved in Spc3p were essential, a series of TRP1-based plasmids carrying the appropriate spc3 mutations and the vector control were transformed into strain HFY405 (⌬spc3)/pHF331 (URA3 SPC3), and transformed cells were subjected to plasmid shuffling. Results from monitoring growth of cells bearing each of these mutations indicated that the conserved amino acids Lys-91, Asp-81, and Asp-115 were nonessential (Table IV).
The fact that the conserved lysines in Sec11p and Spc3p were nonessential (Tables III and IV) suggests that a fundamental difference may exist in the catalytic mechanism of ER signal peptidase relative to E. coli leader peptidase, which contains a serine/lysine catalytic dyad. We therefore constructed additional mutations in the SEC11 and SPC3 genes, paying particular attention to lysine residues that were conserved in only a subset of the sequences presented in Fig. 2, A and B. In the SEC11 gene, mutations altering Lys-122, Lys-124, Lys-148, Asp-53, Ser-33, and Ser-35 were found to be nonlethal (Table  III). In the SPC3 gene, mutations affecting Lys-106, Lys-116, Lys-134, Lys-144, Lys-148, Lys-181, Lys-183, and Ser-103 were nonlethal (Table IV).
Most of the nonlethal mutations listed in Tables III and IV produced no detectable growth defect. The only exception was the mutations altering Ser-42 of Sec11p. The S42T mutation inhibited cell growth ϳ10-fold, whereas the S42A mutation inhibited the growth rate of mutant cells ϳ20-fold as judged by colony size on agar plates at 30°C. None of the mutants tested exhibited a temperature-sensitive or cold-sensitive phenoytpe at 37 and 18°C.
Conserved Serine, Histidine, and Aspartic Acid Residues Are Essential for Signal Peptidase Activity by Sec11p-We next monitored signal peptidase activity in mutant cells using an in vivo assay. Because the temperature-sensitive sec11-7 mutation inhibits signal peptidase activity almost completely at the nonpermissive temperature (18,23), the sec11-7 mutant serves as a suitable host to assess enzyme function in the presence of the mutations constructed by site-directed mutagenesis. Before using this mutant, we chose to sequence the sec11-7 mutation. This mutation was found to produce a single amino acid change, Gly to Asp, at position 67 of Sec11p (Table III).
To monitor signal peptidase activity, a series of plasmids bearing the above-described lethal mutations (see Table III) was introduced into strain HFY409 (sec11-7). Cells were grown to early log phase at the permissive temperature, incubated at the nonpermissive temperature (37°C) for 75 min, and pulselabeled for 10 min. Proteins were precipitated from cells extracts using anti-Kar2p antibodies. Kar2p, an ER lumenal heat shock protein, provided a simple assay for signal peptidase activity because its precursor, preKar2p, was distinguishable from Kar2p on sizing gels (37). As can be seen in Fig. 3, little or no mature Kar2p was present in the sec11-7 mutant control at the nonpermissive temperature, whereas sec11 cells carrying the wild-type SEC11 gene exhibited processing of preKar2p. In contrast, sec11 mutations S44A, H83K, H83F, D103E, D103N, D109E, and D109N inhibited the conversion of preKar2p to Kar2p.

The Stability of Sec11p Is Affected by Some Amino Acid Substitutions at the His-83, Asp-103, and Asp-109 Positions-
Having identified a number of new sec11 mutations that abolish signal peptidase activity, we next asked whether these mutations affect the stability of Sec11p, as a mutation may alter an amino acid important for folding of the enzyme (and stability in yeast cells) but not its catalytic site. Sec11p was epitope-tagged immediately after its N-terminal methionine using the FLAG epitope and shown to function in a ⌬sec11 mutant. However, our initial attempts to detect this protein above background proteins using Western blotting were unsuccessful, even though enriched ER membranes were used. We therefore employed an approach based on the immunoprecipitation of FLAG-tagged Sec11p followed by Western blotting with anti-FLAG antibodies ("Experimental Procedures"). Strain CMY710 (⌬sec11)/pCM112 (SEC11) was transformed with plasmids bearing wild-type SEC11 or the above-described mutations that were lethal to yeast cells (see Table III). The plasmids used allow for overproduction of FLAG-tagged Sec11p because gene expression is under control of a strong promoter (ADH1) (31), and each plasmid contains a 2-m (high copy) origin of replication. We did not, however, determine the extent of overexpression, and we relied on endogenous levels of Spc3p   (19). Conserved residues are indicated using bold letters. Arrows point to amino acids that were changed as listed in Tables III and IV. for complexing to Sec11p, not overexpressed levels.
ER membranes from 1000 A 600 units of log phase (A 600 ϭ 1-4) cells were prepared, and proteins were precipitated from solubilized membranes using agarose-conjugated anti-FLAG antibodies as described under "Experimental Procedures." The precipitate was resolved by SDS-PAGE and subjected to Western blotting using anti-FLAG antibodies. All of the precipitate from 1000 A 600 units of cells was used for each lane of an SDS-PAGE gel, and because the same anti-FLAG antibody was used for the immunoprecipitation and Western blot analyses, the light chain of the anti-FLAG antibody preparation was present in this analysis (Fig. 4). In cells bearing FLAG-tagged Sec11p (see lane denoted wild-type), two forms of the protein were found, one exhibiting an apparent molecular mass of 20 kDa (labeled as Sec11p) and another of 23 kDa (labeled with *). The 20-and 23-kDa species were absent from cells lacking FLAG-tagged Sec11p (see Fig. 4; lane denoted control). The 20-kDa species had the size expected of FLAG-tagged Sec11p. The 23-kDa form was probably ER-glycosylated Sec11p, as its sequence contains one Asn-linked glycosylation site (18). Indeed, two forms of Sec11p were seen previously in ER signal peptidase purified from yeast cells (22).
We observed the absence of Sec11p in yeast cells carrying the H83A, H83K, D103E, and D109N mutations (Fig. 4). In contrast, mutations S44A, H83F, D103N, and D109E did not have a dramatic effect on the steady state levels of Sec11p. One interpretation of the fact that the H83A, H83K, D103E, and D109N mutations caused a loss of detectable Sec11p is that His-83, Asp-103, and Asp-109 are important for the structure and, thus, stability of Sec11p. In contrast, the presence of Sec11p in yeast cells containing the H83F, D103N, and D109E mutations suggests one of two possibilities: (i) His-83, Asp-103, and/or Asp-109 are important structurally but not catalytically, and therefore specific alterations of these amino acids inhibit the normal folding of Sec11p (causing enzyme inactivity as indicated in Fig. 3) without affecting Sec11p stability or (ii) His-83, Asp-103, and/or Asp-109 are important catalytically; however, specific alterations of these residues can lead to protein instability.
Binding of Inactive Sec11p Mutant Proteins to Spc3p-Because Sec11p and Spc3p associate with each other in a complex, a more direct assay of Sec11p structural integrity was to determine whether Sec11p coimmunoprecipitated with Spc3p. To this end, we devised a method to monitor the association of HA-tagged Spc3p with FLAG-tagged Sec11p. Spc3p was tagged immediately after its N-terminal methionine, and the tagged protein was shown to function in a ⌬spc3 mutant. For determinations of binding of FLAG-tagged Sec11p to HA-tagged Spc3p, plasmids pCV101 (SEC11⅐FLAG TRP1) and pCV102 (SPC3⅐HA URA3) were introduced into cells of strain CVY1 (wild type). The SEC11 and SPC3 genes carried by these high copy (2 m) plasmids were under control of a strong (ADH1) promoter and, thus, overexpressed in yeast cells. ER mem- FIG. 3. Pulse-labeling analysis. Strains HFY409 (sec11-7), HFY409/pCV101 (SEC11) (wild-type), and HFY409 harboring a series of plasmid derivatives of pCV101 containing the sec11 mutations indicated at the top of this figure were grown to early log phase at 23°C. Cells were pulse-labeled for 10 min following a 75-min shift to 37°C. Labeled proteins were precipitated with anti-Kar2p antibodies (37) as described previously (23) and resolved on a 7% SDS-PAGE gel.
FIG. 4. Stability of Sec11p. Strains CMY710 (⌬sec11)/pCM112 (SEC11) (control), CMY710/pCV101 (SEC11⅐FLAG) (wild type), and CMY710 harboring a series of plasmid derivatives of pCV101 containing the sec11 mutations indicated at the top of this figure were grown to log phase at 30°C. Enriched ER membranes were generated, solubilized, and subjected to immunoprecipitation using anti-FLAG beads ("Experimental Procedures"). Proteins were resolved on a 12% SDS-PAGE gel and subjected to Western blot analysis using anti-FLAG antibodies. The light chain of the anti-FLAG antibody is a contaminant of the preparation. The positions of Sec11p and a glycosylated form of Sec11p (*) are indicated. Glycosylated Sec11p is present in the D109E and H83F mutants but is not detectable in this figure. Temperature-sensitive a Constructed by site-directed mutagenesis ("Experimental Procedures").

TABLE IV Mutations examined in Spc3p
Mutation produced a Mutant cells are Viable a Constructed by site-directed mutagenesis ("Experimental Procedures").
b Produced by an error-proned polymerase chain reaction in a previous study (19). branes from 2000 A 600 equivalents of the transformed cells were extracted with digitonin, then incubated with agaroseconjugated anti-FLAG antibodies ("Experimental Procedures"). The agarose was sedimented, and proteins were examined by Western blotting using anti-HA antibodies derived from rat. The use of a rat-derived anti-HA antibody was needed to eliminate the murine-derived anti-FLAG light chain band, which was found to obscure Spc3p on the Western blot. As a control, strain CVY1/pCV102 (SPC3⅐HA URA3) that contains Spc3p⅐HA but lacks Sec11p⅐FLAG was treated identically. Derivatives of pCV101 bearing the S44A, H83F, D103N, and D109E mutations, which did not dramatically affect the steadystate levels of Sec11p (Fig. 4), were also introduced into cells of strain CVY1/pCV102 (SPC3⅐HA). All of the material from 2000 A 600 equivalents of cells was represented on a single lane of an SDS-PAGE gel.
In the Western blot probed with anti-HA antibodies, two major forms of Spc3p co-immunoprecipitated with FLAGtagged Sec11p (Fig. 5, see lane denoted wild type). The 27-kDa form was the size expected for HA-tagged Spc3p. The 35-kDa form (denoted by *) probably represents Spc3p that had been transported out of the ER (because of its overproduction) and hyper-glycosylated by Golgi-derived enzymes. In the lane denoted control, the 27-and 35-kDa species were absent, indicating that, as expected, these proteins were not precipitated by anti-FLAG antibodies from cells lacking FLAG-tagged Sec11p yet containing HA-tagged Spc3p. In contrast, Sec11p containing the S44A, H83F, D103N, or D109E mutation was found to coimmunoprecipitate with both the 27-and 35-kDa species, although variable amounts of the 35-kDa form were detected in cells containing the D103N mutation. The data thus reveal that the S44A, H83F, D103N, and D109E mutations do not alter the overall structure of Sec11p, determined through the binding of mutant Sec11p to wild-type Spc3p. DISCUSSION In this study, we have taken a site-directed mutagenesis approach aimed at identifying amino acids that may serve catalytic roles within the two essential subunits, Sec11p and Spc3p, of ER signal peptidase from the yeast S. cerevisiae. The following conclusions are presented. (i) Among the conserved serines, histidines, lysines, and aspartic acids in Sec11p and Spc3p, Sec11p contains one serine, one histidine, and two aspartic acids necessary for signal peptidase activity, and (ii) none of the lysines changed in Sec11p and Spc3p are essential for enzyme function. The absence of an apparent catalytic lysine suggests that, although related functionally, the mechanism of proteolysis by ER signal peptidase differs fundamentally from that of E. coli leader peptidase, which contains a serine/lysine catalytic dyad (13). Our data suggest that Ser-44, His-83, Asp-103, and Asp-109 of Sec11p are candidate catalytic residues, because certain amino acid substitutions at these positions abolish signal peptidase activity without inhibiting the overall structure of Sec11p as judged by the stability of Sec11p and the binding of Sec11p to Spc3p. A subset of changes at the His-83, Asp-103, and Asp-109 positions interfere, however, with the stability of Sec11p (Fig. 4). This finding is unexpected, as conservative changes of catalytic site amino acids do not typically result in protein instability (7,9). The instability detected here suggests that the folding and/or assembly of Sec11p may be unusually sensitive to alterations of at least a subset of catalytic site residues. As with any site-directed mutagenesis approach, however, we cannot eliminate the possibility that the structure of Sec11p has been changed locally by the subset of mutations that inhibit Sec11p function without affecting its stability or its binding to Spc3p. Further analyses will be necessary to establish the identity of the catalytic site amino acids in ER signal peptidase, especially because it is unlikely that both Asp-103 and Asp-109 serve catalytic roles.
Ser-44 and His-83 of Sec11p have been aligned previously to serine and lysine residues, respectively, that make up the serine/lysine dyad of leader peptidase (16). These similarities suggest a commonality in the positioning of catalytically important amino acids and may portend a structural similarity in Sec11p and leader peptidase at or near their catalytic sites. The two essential aspartic acids of Sec11p (Asp-103 and Asp-109) align to Asp-273 and Asp-280, respectively, which are positioned near the catalytic site of leader peptidase (13). Asp-273 and Asp-280 seem to play important structural roles in leader peptidase (7,13). Indeed, Asp-280 forms a salt bridge with Arg-282, an amino acid conserved among many of the leader peptidase-like proteins from eubacterial cells (2,13). Interestingly, sequence alignments suggest that this arginine residue has been replaced by Ser-111 in Sec11p (Fig. 1) 36 -184) was aligned to the sequence of leader peptidase that resides between Box II and Box III (amino acids 154 -271) (see Fig. 1 for a description of Boxes II & III) using the computer-based analysis described previously (19). Gaps in these sequences are indicated by dashes. Identical residues are indicated by stacked dots, and similar residues are indicated by single dots. role in Sec11p.
We have observed here that some changes at the Asp-103 and Asp-109 positions abolish Sec11p function but not its binding to Spc3p. This suggests that one of these aspartic acids (along with Ser-44 and His-83) may be important catalytically, a scenerio that is consistent with the more typical serine proteases that contain a catalytic triad consisting of serine, histidine, and aspartic acid (38). That one of these aspartic acid residues may be catalytic suggests that changing the other essential aspartic acid residue alters only a local structure of Sec11p. Alternatively, both of these aspartic acid residues may be important structurally, leading to the idea that Sec11p uses a serine/histidine dyad for catalysis, such as that found in an esterase from Streptomyces scabies (39). Because a subset of amino acid changes at the His-83 position results in instability of Sec11p, we cannot rule out the possibility that His-83 is important not for catalysis but for the structural integrity of the protein. However, the fact that His-83 has been aligned previously to the catalytic lysine of leader peptidase (16,17) and that substitution of His-83 with a phenylalanine residue leads to a Sec11p that is stable in yeast cells, binds to Spc3p, and lacks detectable enzymatic activity (Figs. 3, 4 and 5) argues for the importance of histidine in catalysis.
The recently solved crystal structure of leader peptidase (13) reveals that this enzyme contains two distinct domains, one possessing both amino acids of the serine/lysine dyad, and a second domain, which we refer to here as the noncatalytic domain. From sequence alignments, we suggested previously that Sec11p and Spc3p may correspond to distinct regions of leader peptidase (19). The strongest homology between the sequences of Sec11p and leader peptidase resides within Boxes I, II, and III (Fig. 1). However, a notable difference exists in the number of amino acids located between Box II and Box III. In Sec11p, this intervening stretch consists of 10 amino acids, whereas the corresponding region of leader peptidase contains 118 amino acids. Most of this 118-amino acid stretch is placed within the noncatalytic domain of leader peptidase (13). Moreover, a portion of these 118 amino acids exhibits detectable homology to the lumenal domain of Spc3p (Fig. 6) (refer also to Ref. 19). These similarities and the finding of probable catalytic amino acids in Sec11p thus raise the possibility that the lumenal portions of Sec11p and Spc3p correspond to the catalytic and noncatalytic domains, respectively, of leader peptidase. Considering the apparent catalytic differences between Sec11p and leader peptidase, this model suggests an unusual evolutionary history among members of the type I signal peptidase family.