Identification of the Amino Acid Residues Essential for Proteolytic Activity in an Archaeal Signal Peptide Peptidase*

Signal peptide peptidases (SPPs) are enzymes involved in the initial degradation of signal peptides after they are released from the precursor proteins by signal peptidases. In contrast to the eukaryotic enzymes that are aspartate peptidases, the catalytic mechanisms of prokaryotic SPPs had not been known. In this study on the SPP from the hyperthermophilic archaeon Thermococcus kodakaraensis (SppATk), we have identified amino acid residues that are essential for the peptidase activity of the enzyme. ΔN54SppATk, a truncated protein without the N-terminal 54 residues and putative transmembrane domain, exhibits high peptidase activity, and was used as the wild-type protein. Sixteen residues, highly conserved among archaeal SPP homologue sequences, were selected and replaced by alanine residues. The mutations S162A and K214A were found to abolish peptidase activity of the protein, whereas all other mutant proteins displayed activity to various extents. The results indicated the function of Ser162 as the nucleophilic serine and that of Lys214 as the general base, comprising a Ser/Lys catalytic dyad in SppATk. Kinetic analyses indicated that Ser184, His191 Lys209, Asp215, and Arg221 supported peptidase activity. Intriguingly, a large number of mutations led to an increase in activity levels of the enzyme. In particular, mutations in Ser128 and Tyr165 not only increased activity levels but also broadened the substrate specificity of SppATk, suggesting that these residues may be present to prevent the enzyme from cleaving unintended peptide/protein substrates in the cell. A detailed alignment of prokaryotic SPP sequences strongly suggested that the majority of archaeal enzymes, along with the bacterial enzyme from Bacillus subtilis, adopt the same catalytic mechanism for peptide hydrolysis.

Secretion proteins and membrane proteins in many cases harbor signal sequences at their extreme N termini that are cleaved during translocation by signal peptidases (1). The released signal peptides are subsequently cleaved into smaller fragments by signal peptide peptidases (SPPs) 2 (2). It had long been presumed that the signal peptides, after their removal from the precursor protein, have no active function in the cell and are simply degraded to free amino acids. However, it is now known that in human cells, peptide fragments generated after cleavage by SPP exhibit vital regulatory functions in immune surveillance (3,4). The function of SPP has also been reported to be necessary for the proper development of Drosophila larvae (5).
In bacteria, SPP was first identified in Escherichia coli. A previously identified cytoplasmic membrane protein, protease IV (6), was found to exhibit SPP activity toward the signal peptide of the outer membrane lipoprotein (2). The gene was cloned (sppA Ec ) (7), and gene disruption studies strongly implied the involvement of SppA Ec in signal peptide degradation (8). Other cytoplasmic proteases including oligopeptidase A have also been found to participate in this degradation pathway (9). It is now presumed that in E. coli, SppA initiates the degradation by introducing endoproteolytic cuts into the signal peptide, whereas the other cytoplasmic proteases are responsible for complete degradation of the smaller fragments into free amino acids (10). SppA has also been identified and genetically characterized in the Gram-positive Bacillus subtilis. The enzyme, along with a cytoplasmic peptidase TepA, has been found to play an important role in signal peptide degradation in this organism (11).
In the Archaea, much remains to be understood on the mechanisms of protein secretion (12)(13)(14) and the fate of the signal peptide after its release from the precursor protein. In terms of signal peptidases, the type I signal peptidase from Methanococcus voltae has been characterized, and a catalytic triad comprised from Ser 52 , His 122 , and Asp 148 has been determined to be critical for its peptidase activity (15,16). FlaK, an aspartic protease essential for preflagellin signal cleavage has also been studied from this organism (17). In the Crenarchaeota, the homologue of bacterial type IV prepilin peptidases from Sulfolobus solfataricus (PibD) has been characterized, and residues on the substrate that are important for recognition by PibD have been examined (18).
As for SPPs, we have recently carried out the first examination of an archaeal SPP (SppA Tk ) (19) from the hyperthermophilic archaeon, Thermococcus kodakaraensis KOD1 (20). SppA Tk (334 residues) was much smaller in size compared with its bacterial counterpart SppA Ec (618 residues). A single, putative membrane-spanning domain was present in the N-terminal region of the protein. We found that ⌬N54SppA Tk , a truncated protein without the N-terminal 54 residues, was a soluble protein exhibiting peptidase activity and stable against autoproteolysis. The substrate specificity of ⌬N54SppA Tk examined with a FRETS peptide library was consistent with its presumed role as an SPP in T. kodakaraensis.
From the primary structures, archaeal and bacterial homologues of SppA Tk and SppA Ec are all members of the S49 family of peptidases included in the Clan SK (MEROPS, the peptidase data base, merops. sanger.ac.uk/) (21). Although the eukaryotic SPPs have been determined to be aspartic proteases (22,23), the effects of various inhibitors on SppA Tk and SppA Ec strongly suggest that the enzymes are serine proteases (6,19). Although sequence comparisons among SppA homologues reveal the presence of several conserved serine residues, experimental evidence identifying the nucleophilic serine has not yet been obtained. Moreover, even with the sequence comparisons, it is still difficult to estimate what other residues might be involved in the catalytic mechanism of SPPs. In particular, His and Asp/Glu residues that com-prise the well known catalytic triad of serine proteases are not clearly conserved among the bacterial and archaeal SppA sequences.
To gain insight on the residues involved in the catalytic mechanism of prokaryotic SPPs, here we have performed a detailed site-directed mutagenesis study on ⌬N54SppA Tk . Through the analyses of various mutant proteins, we have been able to determine multiple residues that are essential or important for the activity of this protein. Our results strongly suggest that ⌬N54SppA Tk and other SppA homologues from the Archaea utilize a Ser/Lys dyad mechanism in peptide cleavage.
DNA Manipulation, Sequence Analysis, and Site-directed Mutagenesis-Plasmid DNA was isolated with the plasmid mini kit from Qiagen. DNA sequencing was performed using a BigDye terminator cycle sequencing kit v.3.1 and a model 3100 capillary DNA sequencer (Applied Biosystems, Foster City, CA). Sequence comparisons and alignments were performed with the ClustalW program provided by the DNA Data Bank of Japan. Site-directed mutagenesis was carried out using a QuikChange XL site-directed kit (Stratagene). The template used was an expression plasmid previously constructed for wild-type ⌬N54SppA Tk , a protein with an N-terminal truncation of 54 amino acid residues. An artificial Met residue was inserted directly before Cys 55 (19). The primers used to incorporate each mutation are shown in Table 1. After sequence confirmation, the plasmids were introduced into E. coli BL21-CodonPlus(DE3)-RIL cells.
Expression and Purification of Wild-type and Mutant ⌬N54SppA Tk Proteins-The recombinant E. coli cells were grown in LB medium, and gene expression was induced with 0.1 mM isopropyl-␤-D-thiogalactopy-ranoside at the mid-to late exponential growth phase. After 6 h, the cells were collected, washed with 50 mM Tris-HCl (pH 8.0), and resuspended in the same buffer. The cells were sonicated on ice, and the supernatant after centrifugation (20,000 ϫ g, 30 min at 4°C) was applied to heat treatment at 85°C for 15 min, immediately cooled on ice, and then centrifuged (20,000 ϫ g, 30 min at 4°C). The soluble protein sample was brought to 35% saturation with ammonium sulfate, and the precipitate was dissolved in 50 mM Tris-HCl (pH 8.0). This was applied to anion exchange chromatography (ResourceQ; Amersham Biosciences) equilibrated with 50 mM Tris-HCl (pH 8), 0.2 M NaCl and eluted with a linear gradient (0.2-1.0 M) of NaCl. After desalting with a HiPrep26/10 column (Amersham Biosciences), the sample was applied to gel filtration chromatography (Superdex 200 HR 10/30; Amersham Biosciences) equilibrated with 50 mM Tris-HCl (pH 8.0), 0.15 M NaCl, and the obtained fractions were used for enzyme analysis. The protein concentration was determined with a protein assay kit (Bio-Rad) using bovine serum albumin as a standard.
Enzyme Activity Measurements-Standard activity measurements were performed at 60°C in 1 ml of 50 mM CHES (pH 10.0) with 0.1 g of purified protein and 200 M Ala-Ala-Phe-MCA available from Peptide Institute (Osaka, Japan). Release of 7-amino-4-methyl-coumarin was monitored consecutively with a fluorescence spectrophotometer capable of maintaining the cuvette at desired temperatures between 30 and 100°C. ex and em were 380 and 460 nm, respectively. The final concentration of Me 2 SO used to dissolve the substrate was constant at 3% of the reaction mixture. Kinetic parameters were calculated with IGOR Pro version 5.0 (WaveMetrics, Lake Oswego, OR).
Circular Dichroism Spectroscopy of Wild-type and Mutant Enzymes-Each protein sample was prepared in 25 mM Tris-HCl (pH 8.0), 75 mM NaCl at a protein concentration of ϳ0.1 mg ml Ϫ1 . A J-820 spectropolarimeter (Jasco, Tokyo, Japan) was used to measure ellipticity as a function of wavelength from 250 to 200 nm in 0.2-nm increments using a 0.1-cm cylindrical quartz cuvette. The samples were scanned one hundred times and averaged. The mean molar ellipticity [] (deg cm 2 dmol Ϫ1 ) was calculated from the equation [] ϭ /10 nCl, where is the measured ellipticity in millidegrees, C is the molar concentration of enzyme subunits, l is the path length in centimeters, and n is the number of residues/subunit.
Examination of the Substrate Preference of Wild-type and Mutant Enzymes-The substrate preferences of the wild-type, S128A, and Y165A proteins were examined with a FRETS peptide library (25Xaa series, Peptide Institute) (19,24). These peptide substrates harbor a highly fluorescent 2-(N-methylamino)benzoyl group linked to the side chain of the N-terminal D-2,3-diamino propionic acid residue (D-A 2 pr), along with a 2,4-dinitrophenyl group (quencher) linked to the ⑀-amino group of a Lys residue. In between the D-2,3-diamino propionic acid residue and the Lys residue lies the peptide Gly-Zaa-Yaa-Xaa-Ala-Phe-Pro, where Zaa is a mixture of Phe, Ala, Val, Glu, or Arg; Yaa is a mixture of Pro, Tyr, Lys, Ile, or Asp; and the Xaa residue is a defined single amino acid of choice. ex and em were 340 and 440 nm, respectively. To examine the preference for residues at the P-1 position, 1 g of purified enzyme was added to the reaction mixture with a final volume of 1 ml containing 30 M substrate in 50 mM CHES (pH 10.0). The final concentration of Me 2 SO used to dissolve the substrate was constant at 3% of the reaction mixture.

Highly Conserved Amino Acid Residues among the Signal Peptide
Peptidases from Archaea and Bacteria-To determine which residues should be selected for site-directed mutagenesis and subsequent

Mutation
Primer sequences biochemical analyses, we aligned all SppA homologues found in the archaeal genomes, along with the biochemically and/or genetically characterized bacterial SppA from E. coli and B. subtilis. Among the 21 archaeal genomes that have been sequenced, we could identify 19 SppA homologues in 16 organisms (see legend of Fig. 1). Although a number of other open reading frames have been annotated as putative SPPs, they displayed clearly lower degrees of similarity with SppA Tk and SppA Ec and were therefore not selected for further examination. The sequence comparison indicated that SppA homologues were not present in Aeropyrum pernix, S. solfataricus, and Sulfolobus tokodaii from the Crenarchaeota and Methanopyrus kandlerii and Archaeoglobus fulgidus from the Euryarchaeota. A region corresponding to residues between Met 115 and Lys 291 in SppA Tk was fairly conserved among all sequences and presumably includes the catalytic core regions of these enzymes. An alignment of the sequences spanning this region, along with the mutations introduced in this study, is shown in Fig. 1.
Although the eukaryotic SppA has been demonstrated to be an aspartate peptidase (22), inhibitor studies on the bacterial SppA Ec and the archaeal SppA Tk have indicated that both of these enzymes are serine peptidases (6,19). We therefore initially focused on the presence of conserved serine residues. Ser 128 , Ser 162 , and Ser 184 of SppA Tk were completely conserved in all sequences, with the only exceptions being a replacement of the first Ser to Thr in one of the three homologues from Haloarcula marismortui and a replacement of the third Ser to His in another homologue from this strain. In terms of the His residue in the well known Ser/His/Asp catalytic triad, only three His residues were found in this region of SppA Tk and were not highly conserved among the SppA sequences. His 140 and His 213 were not conserved even in the closely related Pyrococcus spp., whereas His 191 , although shared by SppA Tk and the three Pyrococcus enzymes, was not found in the other SppA sequences. As an alternative, we found that the basic residue Arg 250 was highly conserved among the SppA proteins. As for the Asp residue, Asp 277 was the only acidic residue that was present without exception in all sequences (Fig. 1).
Production and Purification of Mutant SppA Tk Proteins via Site-directed Mutagenesis-Based on the sequence alignment described above, we first constructed the following six mutants; S128A, S162A, S184A, H191A, R250A, and D277A. The primers used to incorporate the mutations are shown in Table 1. The mutations were incorporated into the expression vector for ⌬N54SppA Tk , a truncated protein without the N-terminal 54 residues of SppA Tk . Sequence analysis confirmed that only the intended mutations were introduced into the genes. Recombi- nant mutant proteins were produced in E. coli BL21-CodonPlus(DE3)-RIL cells and purified from the cell-free extracts by heat treatment at 85°C for 15 min, followed by ammonium sulfate fractionation, anion exchange chromatography, and gel filtration chromatography. Although we were able to obtain each mutant protein in a soluble form, the amount of protein produced in the E. coli cells varied with each mutant. We previously reported that the recombinant ⌬N54SppA Tk was obtained in an octameric or hexadecameric form (19). We also observed various quaternary structures in the mutant proteins, but only the fraction corresponding to the octameric form of the protein was used for further analysis. The apparent homogeneity of each protein after the purification procedure was examined by SDS-PAGE (Fig. 2).
Peptidase Activity of the Wild-type and Mutant ⌬N54SppA Tk Proteins-We have previously reported that the synthetic peptide Ala-Ala-Phe-MCA is a good substrate for the wild-type ⌬N54SppA Tk (19). The activity levels of each mutant protein were measured at 60°C in the presence of a fixed concentration of this substrate (200 M) and compared with that of the wild-type enzyme (Fig. 3). We found that the mutation of Ser 162 had the most detrimental effect. No activity could be observed with the standard assay procedures. We could estimate that the activity of the protein was only 0.004% of that of the wild-type enzyme by using a 100-fold higher protein concentration. The S184A and H191A mutations also brought about a decrease in enzyme activity, but significant levels of activity, 24 and 45% of that of the wild-type enzyme, respectively, were still observed in each mutant protein. Interestingly, the S128A, R250A, and D277A mutations did not have any negative effect but rather led to an unexpected increase in peptidase activity. It has been shown that mutations in the catalytic residues of serine proteases with a catalytic triad result in 10 4 -10 6 -fold reduction in activity (25). Our results clearly indicate that Ser 162 is essential for the peptidase activity of SppA Tk and is most likely the nucleophilic serine of the enzyme. The results also reveal that Arg 250 and Asp 277 , two highly conserved residues in both archaeal and bacterial SPPs, have little, if any, role in the peptide hydrolyzing mechanism of SppA Tk .
Mutations to Identify the General Base Residue of SppA Tk -The absence of conserved histidine residues among SppA proteins and the fact that His 191 was not essential for the catalytic activity suggested that SppA Tk was not dependent on the well known Ser/His/Asp catalytic triad. We therefore searched for other basic and acidic residues as targets for site-directed mutagenesis. Intriguingly, there were no other charged residues that were completely conserved in all of the sequences. We therefore chose those that were the most highly conserved among the sequences, Lys 150 , Lys 209 , Lys 214 , Asp 215 , Arg 221 , Glu 226 , and Glu 227 . Each residue was replaced by Ala with appropriate primers (Table 1). Expression and purification were carried out as described for the initial mutant proteins, and the apparent homogeneity of each mutant protein is displayed in Fig. 2.
The K150A and E226A mutations did not have a negative effect on the activity of the enzyme. All other residue replacements led to mutant proteins with lower levels of activity than that of the wild-type enzyme. Moderate effects were observed in K209A (59% activity retained), E227A (35%), and D215A (20%), whereas the effect of the R221A (4%) was much more significant. The K214A mutation had the most dramatic effect, and as in the case of the S162A mutation, activity could not be detected under standard procedures. By applying increased protein concentration, activity of K214A was estimated at 0.01% of that of the  wild-type protein. The results indicate that Lys 214 is essential for activity of SppA Tk and that the protein most likely utilizes a Ser-Lys catalytic dyad for peptide cleavage. The results also indicate that Arg 221 , although not essential, plays an important role in the peptidase activity of SppA Tk .
Effects of Mutations on Other Highly Conserved Residues-We next focused on conserved, noncharged residues. Several glycine residues were highly conserved among the homologues, and in particular, a consecutive Gly-Gly sequence was conserved in all SppA proteins. In some lipases and esterases, a consecutive glycine sequence provides the main chain amino group(s) that stabilize the oxyanion intermediate that is formed after the nucleophilic attack on the carbonyl group by serine (26). We therefore constructed the mutant proteins G130A and G131A. We also selected a Tyr residue in the near vicinity of the nucleophilic Ser 162 that is completely conserved among Spp proteins and introduced a Y165A mutation. As a result, we observed a moderately positive effect with the G130A mutation and a surprisingly high increase in activity in the Y165A mutant protein. The G131A mutation had a drastic effect on activity, and the protein exhibited only 2% of the activity levels of the wild-type enzyme, indicating an important role for this glycine residue.
Circular Dichroism Spectra of Wild-type and Mutant ⌬N54SppA Tk Proteins-To examine whether the single-residue replacements had unintended broad effects on the protein structure, we analyzed the CD spectrum of each mutant protein. In all mutant proteins, including the two that hardly exhibited activity (S162A and K214A), the CD spectra were indistinguishable from that of the wild-type protein (data not shown). We would like to note that in the case of G130A and G131A, we observed a slight increase in ellipticity between 225 and 230 nm, raising the possibility of protein-protein interaction. Overall, the results indicate that the single-residue replacements introduced did not lead to significant changes in the secondary structures of the proteins. This allows us to interpret the changes in activity as direct consequences brought about by residue exchange.
Kinetic Analysis of the Mutant Proteins-We next performed kinetic analyses for all mutant proteins that exhibited sufficient levels of activity. Most mutant proteins were analyzed by standard procedures with various concentrations of Ala-Ala-Phe-MCA. The S162A and K214A mutant proteins could not be examined because of their extremely low levels of activity. Analyses of the G131A and R221A proteins were carried out with 20-and 10-fold higher amounts of enzyme, respectively.
We have previously reported that the cleavage reaction of Ala-Ala-Phe-MCA by the wild-type ⌬N54SppA Tk followed Michaelis-Menten kinetics (19). However, in the process of analyzing the activity levels of mutant proteins at high substrate concentrations, we observed a decrease in activity levels. Further examination of all active proteins, including the wild-type ⌬N54SppA Tk , confirmed the occurrence of substrate inhibition. By considering several equations, we found that our data fit very well (R 2 Ͼ 99.4) to one of the typical substrate inhibition models, expressed as v ϭ V max [S]/(K s1 ϩ [S] ϩ 1/K s2 [S] 2 ), where v is reaction velocity, V max is maximum velocity, [S] is substrate concentration, K s1 is the dissociation constant between enzyme and the first substrate, and K s2 is the dissociation constant between the enzyme-substrate complex and the second, inhibitory substrate. Representative [S]-v plots with the respective curves are shown for the wild-type and K150A, Y165A, and H191A mutant proteins in Fig. 4. The kinetic parameters of each protein are indicated in Table 2.
When compared with the wild-type protein, a number of mutant proteins exhibited similar K s1 values, indicating that these residue replacements did not greatly alter the affinity of the enzyme toward the substrate Ala-Ala-Phe-MCA. Significant increases in the K s1 values were observed for S128A, G130A, G131A, D215A, R221A, and E227A.
Because these mutations, other than G130A and G131A, result in a decrease in the size of the side chain, the increase in K s1 values may be due to the abolishment of favorable interactions between the binding pocket and substrate. The increase may also be due to the hydrophobic property of the Ala residue, which may lead to a repulsive effect on a hydrophilic region of the substrate. As for G131A, reaction velocity continued to increase with increases in substrate concentration up to at least 3,000 M, and therefore reliable curve fitting could not be performed. Fitting the equation with the limited data indicated that the K s1 value was higher than 10 6 M. The G130A mutant also displayed K s1 values Ͼ10 4 M, and reliable values for the kinetic parameters could not be obtained. In these two proteins, the additional methyl group may directly hinder the binding between enzyme and substrate. On the other hand, the glycine residues may be important in maintaining the optimized structure of the protein, because consecutive glycine residues would allow dramatic bends in the main chain backbone. As mentioned above, we did observe subtle changes in the CD spectra of these two glycine mutant proteins. On the SDS-PAGE gels, we also noticed a slight difference in the mobility of G130A along with the appearance of a high molecular weight band in the case of G131A, both of which are most likely due to incomplete denaturation or dissociation of the oligomeric  a Accurate estimations of the kinetic parameters were not possible because of reasons described in the text.
protein (Fig. 2). Although the CD spectra rule out drastic changes in protein conformation, mutations of these two residues may result in subtle changes in the enzyme conformation. A rather unexpected result was the relatively large number of mutations that led to increases in the V max or k cat /K s1 values compared with the wild-type enzyme. Significant increases in the V max value were observed in S128A (690%), K150A (148%), Y165A (232%), E226A (135%), R250A (146%), and D277A (159%). Even in the k cat / K s1 values, which should better represent the functional capacity of the enzyme in the cell, notable increases were observed in S128A (155%), K150A (163%), Y165A (222%), E226A (135%), and R250A (169%). The fact that these conserved residues seem to suppress the activity levels of SppA Tk may be important for the proper function of the enzyme in vivo (see "Discussion").
Examination of the Activity of S128A and Y165A toward Various Substrates-Two mutant enzymes, S128A and Y165A, which displayed particularly large increases in both V max and k cat /K s1 values, were selected for examination of their activities against a FRETS peptide library as described under "Experimental Procedures." We have previously examined the activity of the wild-type enzyme against these substrates and observed higher cleavage rates toward peptides with small amino acid residues at the Xaa position, corresponding to the P-1 position. Activity toward substrates with charged residues or aromatic residues at the Xaa position was low (19). We found that the wild-type enzyme exhibited the same tendencies in this study also (Fig. 5). When we performed the same activity measurements with the S128A and Y165A mutant proteins, we observed that activity levels toward substrates with positively charged His, Arg, or Lys residues at the Xaa position greatly increased. An increase in activity levels against substrates with the aromatic residues Phe and Tyr was also observed. Even when normalizing the activity levels of each protein by designating their activities against the glycine substrate as 100, we observed nearly a 3-fold increase in preference toward the Lys substrate in the S128A protein and a more than 2-fold increase in preference toward the Arg substrate in Y165A. Preferences toward the Tyr and Phe substrates also increased ϳ3-fold in both enzymes. The results clearly indicate that in addition to limiting the activity levels of SppA Tk , Ser 128 and Tyr 165 also play a role in restricting the substrate preference of the enzyme toward peptide substrates.

DISCUSSION
In this study, we have performed a detailed site-directed mutagenesis study on an archaeal SPP from T. kodakaraensis. The results have revealed multiple residues that are critical for the peptidase activity of SppA Tk and provide the first insight into the catalytic mechanism of prokaryotic SPPs. Our analyses strongly indicate that the catalytic center of SppA Tk is comprised of a Ser-Lys dyad and not the Ser/His/Asp catalytic triad that is present in the majority of serine proteases.
The nucleophilic serine residue of SppA Tk has been clarified to be Ser 162 . Along with the inhibitor studies that indicated the protein was a serine peptidase (19), activity was completely abolished in the S162A protein, whereas mutations in the other well conserved serine residues (S128A and S184A) had relatively smaller or no effects on the peptidase activity. Because Ser 162 is completely conserved in all of the archaeal SppA homologues as well as the SppA from E. coli and B. subtilis, this serine residue most likely serves as the nucleophile in all prokaryotic SPPs. Besides the three residues examined, no other Ser residue is conserved in the core region of the enzyme, even among the closely related proteins from Thermococcales. Our results also indicate that the residue acting as the general base in SppA Tk is Lys 214 . This residue is clearly conserved in the archaeal enzymes from Pyrococcus spp., the methanogens, Thermoplasma spp., Picrophilus, and one homologue from Haloarcula. The residues surrounding the Lys residues are also highly conserved among these proteins, and the sequences can be clearly aligned without any gaps from the nucleophilic Ser 162 to a conserved Glu 227 . Although to a lower extent, the enzyme from Nanoarchaeum equitans also displays similarity and harbors most of these residues. It can therefore be presumed that the residue acting as the general base in these enzymes (Fig. 1, upper group of sequences), corresponds to the Lys 214 in SppA Tk . Although a lysine residue is found in some of the remaining homologues from the halophiles and the Crenarchaeon Pyrobaculum (Fig. 1, lower group of sequences), the relatively lower sequence similarity makes it difficult for us to estimate the general base in these enzymes. Interestingly, the bacterial SppA from B. subtilis also harbors the basic Lys residue, and surrounding sequences are particularly well conserved between the enzyme and SppA Tk , including the charged residues Lys 209 , Lys 214 , Asp 215 , Arg 221 , Glu 226 , and Glu 227 . This strongly suggests that the bacterial SppA from B. subtilis also utilizes Lys 214 as the general base. On the other hand, Lys 214 is not conserved in SppA Ec , and Lys residues are not found in the near vicinity. This indicates that the enzyme from E. coli utilizes distinct residues for catalysis.
Besides Ser 162 and Lys 214 , which were essential for activity, we found that the S184A, H191A, K209A, D215A, and R221A mutations led to decreases in V max or k cat values. The results suggest that these residues, although not essential for activity, may also play a role in the catalytic mechanism. In particular, the mutation of Arg 221 had severe effects. Because our results have clearly indicated the presence of a Ser-Lys catalytic dyad in SppA Tk , one important question to be solved would be FIGURE 5. Peptidase activity of wild-type ⌬N54SppA Tk and its mutants S128A and Y165A toward various substrates from a FRETS peptide library. Amino acid residues at the Xaa position are indicated with single-letter abbreviations. Each substrate was examined at a concentration of 30 M. The activity of the wild-type (WT) protein toward the Xaa ϭ Gly substrate was designated as 100%.
how the lysine residue is maintained in a deprotonated state, which would be necessary to increase the nucleophilicity of the serine side chain O␥. The pKЈ value of the lysine side chain is ϳ10.8, and the microenvironment surrounding the catalytic center would have to be extremely alkaline. One feasible explanation taking our results into consideration would be that the side chain of Arg 221 is positioned in the very near vicinity of the Lys 214 side chain. Because the pKЈ value of an arginine side chain is much higher (ϳ12.5), Arg 221 may act to sequester protons that would otherwise tend to protonate Lys 214 . The positive charge of the protonated Arg side chain may also act to repulse other protons from approaching the Lys 214 ⑀-amino group. Another possibility would be that Lys 214 is surrounded by a neutral or hydrophobic environment, which would also result in an apparent decrease in the pKЈ value of the ⑀-amino group, as reported in LexA (27) and the signal peptidase from E. coli (25,28).
The S184A mutation also had significant effects on the activity of SppA Tk . From a comparison of the three-dimensional structures of the bacterial type I signal peptidase (28), Lon protease (29), LexA (27), UmuDЈ (30), and the repressor C-terminal domain (31), which are all Ser/Lys proteases, possibilities of a third residue interacting with the Ser/Lys dyad have been raised (29). The residue is conserved as either a serine or threonine residue, with the side chain O␥ forming a hydrogen bond with the catalytic Lys ⑀-amino group. Because no other serine or threonine residues are conserved among the SPPs and because the S128A mutation led to an increase in activity, Ser 184 may represent the corresponding residue in SppA Tk . Future studies on the three-dimensional structure of the protein should clarify the roles of Ser 184 and Arg 221 .
As mentioned under "Results," it is intriguing that the replacements of so many highly conserved residues, usually presumed to contribute in maintaining enzyme function, actually lead to increases in peptidase activity. One possibility is that the substrate or the reaction conditions we have applied in our activity measurements differ so much from the actual environment of the enzyme that our results do not reflect at all the actual signal peptide degrading activity of the individual proteins. However, another tempting possibility is that these residues are deliberately present to limit the functional capacity of SppA. In bacteria, SppA is presumed to be anchored to the cell membrane with its catalytic core facing the cytoplasm. Although peptidase activity is necessary for the breakdown of free signal peptides, it should also be important for the cell that the enzyme does not cleave other proteins, whether they are soluble proteins in the cytoplasm or proteins integrated or anchored to the cell membrane. This can be achieved by a variety of strategies on the protein itself or toward its environment. A strict confinement of the substrate specificity of the enzyme would surely lead to a decrease in SPP acting on proteins other than free signal peptides. Limiting the activity levels of the enzyme would also contribute to prevent unintended protein degradation. Preventing the access of SPPs to proteins other than signal peptides would be another practical strategy. Because we have found that the conserved Ser 128 and Tyr 165 not only place a limitation on activity levels but also act to restrict the substrate specificity of SppA Tk , the presence of these residues may well be to suppress activity of the enzyme toward unintended substrates.
Although still a minority among the serine proteases, the number of enzymes known to utilize the Ser/Lys catalytic dyad is steadily increasing. In addition to the bacterial type I signal peptidase, Lon protease, LexA, UmuDЈ, and the repressor C-terminal domain mentioned above, other examples include the archaeal Lon protease from Methanocaldococcus jannaschii (32), PH1510 from Pyrococcus horikoshii (33), the periplasmic tail-specific protease from E. coli (34), and the C-terminal endoprotease from cyanobacteria (35). This study has revealed that the SPPs from Euryarchaeota are also members of this enzyme family. There are still no indications for a reason why certain serine proteases utilize the Ser/His/Asp catalytic triad, whereas others use the Ser/Lys dyad. Although presumed to execute identical functions, the signal peptidase from the archaeon M. voltae utilizes a catalytic triad (16), whereas the signal peptidase from E. coli harbors a dyad (28). The results of our study also raise the possibility that the SPPs from E. coli and T. kodakaraensis function through distinct catalytic mechanisms. Discovery of novel peptidases/proteases that utilize the Ser/Lys dyad and further structure-function studies on both families of enzymes will be necessary for an understanding of this intriguing distinction in catalytic mechanisms.