INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Members of the sanguis group of streptococci are highly successful colonizers of both hard and soft surfaces in the human oral cavity and thus represent a significant portion of the indigenous microflora (24). Members of this group of bacteria are important in the development of oral diseases because they are the primary colonizers of newly erupted or freshly cleaned tooth surfaces (7, 35). This colonization initiates the formation of dental plaque, the biofilm associated with caries and periodontal disease. In addition, these organisms are major etiological agents of subacute bacterial endocarditis (5). The successful establishment of streptococci in the host appears to be dependent on the ability of these organisms to adhere to oral as well as to other host surfaces. These surfaces include epithelia, salivary pellicle, platelets, other bacteria in plaque, and extracellular matrix components, particularly fibronectin (22, 28). Adherence is facilitated by multiple adhesins expressed on the surface of the bacterial cell (27, 28).

Adhesion of the oral bacterium Streptococcus parasanguis FW213 to an in vitro tooth surface model, saliva-coated hydroxyapatite (SHA), is mediated by peritrichous surface fimbriae (15, 21, 22). The genes encoding two fimbria-associated adhesins, Fap1 and FimA, in FW213 have been identified. The fap1 gene encodes a high-molecular-weight glycoprotein that is important for adhesion to SHA and appears to be involved in the assembly of fimbriae (56). The fimA gene encodes a 36-kDa lipoprotein (20, 23) that is an important virulence factor in S. parasanguis endocarditis and appears to promote adherence to fibrin in cardiac vegetations (6). fimA is part of an operon that encodes components of an ATP-binding cassette-type membrane transport system. In addition to fimA, the fim locus includes fimB (formerly orf1), which encodes an ATP-binding protein, and fimC (formerly orf5), which encodes a hydrophobic transmembrane protein (20). Homologues of FimA have been identified, and sequence analysis of FimA and its homologues has suggested that this group of proteins belongs to a new subfamily of solute-binding proteins that potentially bind metals (13, 29). Just downstream of the fimA operon is the gene tpx (formerly orf3), which encodes a 20-kDa protein (20) with significant homology to a periplasmic thiol peroxidase (Tpx) of Escherichia coli (8). In E. coli, Tpx is an enzyme that prevents the formation of reactive oxygen species toxic for many bacteria. Recently, it has been shown that S. parasanguis Tpx functions in a similar manner (54). Tpx may play a role of particular importance in the survival of bacteria, such as S. parasanguis, that lack catalase.

In this study, we isolated and sequenced DNA fragments upstream of the fimA operon in order to identify additional factors that may be associated with adherence to host surfaces, virulence, or survival in the host environment. We identified an open reading frame (ORF) that is divergently transcribed from the fimA operon and that encodes an endopeptidase (PepO) with high sequence identity to known bacterial and mammalian endopeptidases.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and media. The bacterial strains and plasmids used in this study are listed in Table 1. The wild-type S. parasanguis (formerly S. sanguis) strain used was FW213 (11). Strain VT1346 pepO::Km^r was constructed by insertional inactivation of pepO as described below. Streptococcal strains were grown statically in Todd-Hewitt (TH) broth (Difco Laboratories, Detroit, Mich.) in the presence of 5% CO₂ at 37°C. Tetracycline (15 µg/ml) or kanamycin (120 µg/ml) was added as required. E. coli JM109 (Promega, Madison, Wis.) served as the host for cloning experiments and for plasmid propagation. Strain LE392 (Promega) served as the host strain for the propagation of recombinant bacteriophage. E. coli was grown in Luria-Bertani medium or NZYM medium (40) and, when required for plasmid selection, ampicillin (100 µg/ml), tetracycline (15 µg/ml), or kanamycin (25 µg/ml) was included. Agar was added to a final concentration of 1.5% to prepare solid medium.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Physical map of pepO and the fimA operon region of S. parasanguis. The diagram shows a simplified restriction map of a 5.1-kb DNA fragment containing pepO, three genes (fimC, fimB, and fimC) of the fimA operon, and tpx. Arrows indicate the locations and orientations of the ORFs. The lines below the map indicate the sizes and locations of probes and plasmid inserts. The Kmr marker, shown by an open arrowhead, was inserted at the KpnI restriction site of pepO. Restriction sites were as follows: B, BamHI; E, EcoRI; G, BglII; K, KpnI; P, PstI; S, SalI; and T, SstI.

DNA manipulations. Molecular biology techniques were carried out essentially as described previously (40). Plasmids were isolated from E. coli with Mini or Midi column plasmid purification kits (Qiagen, Valencia, Calif.). S. parasanguis genomic DNA was isolated with a Puregene DNA isolation kit (Gentra Systems, Inc., Minneapolis, Minn.). Phage DNA was isolated as described previously (40) or with a Qiagen lambda DNA isolation kit. DNA probes used in Southern or plaque hybridization analyses were radiolabeled with [-³²P]dCTP by nick translation with a commercially available kit (Gibco BRL, Grand Island, N.Y.). Southern and plaque hybridizations were performed as suggested by the membrane manufacturer (Amersham Corp., Arlington Heights, Ill.) and included high-stringency washes.

RNA extraction and Northern hybridization analysis. S. parasanguis cells were harvested from 10-ml TH broth cultures during mid-logarithmic-phase growth, and the cell pellet was resuspended in water to 200 µl. RNA was isolated from cells with the FastRNA BLUE protocol and a FastPrep instrument according to the manufacturer's (Bio 101, Inc., Vista, Cal.) directions. Cells were processed in the FastPrep instrument at a speed rating of 6 two times for 30 s each time, for a total of 60 s. For Northern analysis, RNA (8 µg) was separated on a 1.2% agarose-0.44 M formaldehyde gel, blotted, and hybridized to an [-³²P]dCTP-labeled DNA probe as previously described (25). A New England Biolabs (Beverly, Mass.) 0.46- to 4.06-kb RNA ladder was used as a molecular weight marker. The probe used was a PCR-amplified 1,062-bp fragment from pepO (Fig. 1).

Electrotransformation of E. coli and S. parasanguis. E. coli cells were prepared for electroporation as previously described (4). Cells (50 µl) and approximately 7 µg of DNA were electroporated in a 0.2-cm-gap-size cuvette in a Bio-Rad Gene Pulser set at 2.5 kV, 25 µF, and 200 . Immediately following electroporation, 360 µl of SOC medium (40) was added to the cuvette. Cells were transferred to a 1.5-ml microcentrifuge tube and incubated at 37°C for 1 h. Cells were then spread on Luria-Bertani agar plates supplemented with appropriate antibiotics and incubated overnight at 37°C.

PCR. PCR was performed with a model 9600 thermocycler (Perkin-Elmer Cetus, Norwalk, Conn.) and a GeneAmp PCR reagent kit (Perkin-Elmer Cetus) according to the manufacturer's directions. Genomic DNAs of S. parasanguis FW213 and pepO allelic replacement mutant VT1346 were amplified with upstream (P1, 5'-TTCTCCCCACAAGACATTC-3') and downstream (P2A, 5'-GAGTTTGCTCCTTTAC-3') primers bracketing the pepO KpnI site. The PCR conditions were as follows: one cycle at 94°C for 3 min followed by 35 cycles of 1 min at 94°C, 1 min at 50°C, and 1 min at 72°C. The amplification reaction was completed with an additional cycle of 3 min at 72°C. The PCR amplification products were analyzed by agarose gel electrophoresis.

DNA sequencing and analysis. DNA sequencing was carried out at the Vermont Cancer Center DNA Analysis Facility at the University of Vermont. Nucleotide and protein similarity searches were done with the BLAST programs (1, 2) via the National Center for Biotechnology Information (NCBI) BLAST server (2). Protein similarity searches were conducted with both the NCBI nonredundant protein sequence database and the NCBI database of unfinished genomes. Pairwise and multiple amino acid sequence alignments were obtained with the BLAST and CLUSTAL W (48) programs, respectively, both with the default parameters.

Assay for endopeptidase activity. Endopeptidase activity was assayed by a procedure modified from that of Tan et al. (47). S. parasanguis strains were harvested from 500-ml TH broth cultures during late logarithmic growth, washed once with 40 ml of 20 mM potassium phosphate buffer (pH 6.0), and resuspended in 5 ml of cold 20 mM potassium phosphate buffer (pH 6.0). Cells were disrupted with a French press (Aminco, Silver Spring, Md.), with the pressure being maintained at 15,000 lb/in². Extracts were obtained from disrupted cells by centrifugation at 17,210 × g for 20 min at 4°C. The protein concentration in each cell extract was determined with the bicinchoninic acid protein assay reagent (Pierce, Rockford, Ill.). Reaction mixtures contained 1.25 mM metenkephalin (Tyr-Gly-Gly-Phe-Met; Sigma Chemical Co., St. Louis, Mo.) in 20 mM Tris-HCl at pH 7.0 and 15 µg of cell extract in a total volume of 50 µl. Reaction mixtures were incubated at 30°C, and reactions were stopped by adding 10 µl of 30% acetic acid and cooling the mixtures to 4°C.

Adhesion of S. parasanguis to SHA. Preparation of hydroxyapatite beads, clarification of saliva, preparation of SHA beads, and the adhesion assay used have been described previously (15). Data were obtained from two independent one-point adhesion assays (22) performed in triplicate.

Phylogenetic analysis. Numerous S. parasanguis PepO amino acid sequence homologues were identified from the BLAST search and analyzed. Several hypothetical nematode proteins and the KELL blood group glycoprotein were not included in the final analysis (see Fig. 6). Inclusion of the nematode and KELL sequences in the analysis did not change the results, but the topologies of the resulting trees were otherwise uncertain. The final data set used to construct phylogenetic trees consisted of the 22 amino acid sequences listed in Table 2.

Nucleotide sequence accession number. The sequence data for pepO have been submitted to the DDBJ, EMBL, and GenBank databases under accession no. AF029677.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation and characterization of the S. parasanguis pepO gene. The operon encoding the S. parasanguis FW213 fimbria-associated adhesin, fimA, was cloned and sequenced previously (18, 19). Initially, pVT573, containing an internal fragment of fimA (Fig. 1), was used to screen a genomic S. parasanguis phage EMBL3 library for additional sequences linked to the fimA operon. A recombinant phage containing fimA and an incomplete upstream ORF was identified. Subsequently, both pVT573 and a 1.9-kb EcoRI fragment upstream of fimA (Fig. 1) were used to screen the EMBL3 library again. Of approximately 3,000 plaques screened, 8 that hybridized to both probes were isolated. DNA was prepared from three of the recombinant phage, and restriction enzyme analysis revealed that all three harbored identical inserts. One phage, termed VT1331, was chosen for further investigation. Phage VT1331 contained an insert of approximately 11.0 kb, and Southern hybridization analysis confirmed that the insert contained fimA on a 6.0-kb EcoRI restriction fragment. A restriction site map of the 11.0-kb insert revealed a 1.9-kb EcoRI fragment adjacent to the 6.0-kb EcoRI fragment (Fig. 1). Subsequently, this 1.9-kb EcoRI fragment, as well as an overlapping 4.4-kb SalI-KpnI fragment, was subcloned from phage VT1331, and the nucleotide sequence of 2.3 kb of DNA was determined.

The product of the ORF upstream of the fimA operon displays homology to bacterial and eukaryotic zinc metallopeptidases. Nucleotide sequence analysis of the 2.3-kb region revealed the presence of a single ORF 148 bp upstream and divergently transcribed from the fimA operon (Fig. 1). The ORF was comprised of 1,893 nucleotides and was predicted to encode a protein of 631 amino acids and with a molecular mass of approximately 70.6 kDa. No typical signal sequence or regions of hydrophobicity were found in the predicted protein, suggesting that it is a soluble cytoplasmically located protein.

View larger version (143K): [in this window] [in a new window]   FIG. 2.   Sequence of the deduced S. parasanguis PepO protein and alignment with related enzymes. Amino acid residues, indicated by the standard one-letter code, are numbered on the left. Highlighted residues were identical to those of the S. parasanguis PepO protein. Gaps are indicated by dashes. Arrowheads indicate the HEXXH motif and residues in the mammalian enzymes known to be critical for enzyme activity. The sequences shown are as follows: PepO(Sp), S. parasanguis PepO (GenBank accession no. AF029677); PepO(Ll), L. lactis PepO (53); ECE(Hs), H. sapiens ECE isoform 1 (41); and NEP(Hs), H. sapiens NEP (32).

pepO is transcribed as a monocistronic message. Based on DNA sequence analysis, it was reasonable to predict that pepO was transcribed as a monocistronic message. Northern hybridization analysis was performed to verify this assumption. A 1,062-bp PCR product (Fig. 1) corresponding to an internal coding region of pepO was hybridized to total RNA from S. parasanguis FW213 that had been grown in TH broth. Under high-stringency conditions, the probe hybridized to an abundant RNA transcript with an apparent molecular size of approximately 2.1 kb (Fig. 3). A transcript of this size is consistent with the hypothesis that pepO is transcribed as a monocistronic mRNA. The pepO probe also hybridized to a second, slightly larger and less abundant transcript that is obscured in reproductions of autoradiographs because of its close proximity to the major band. At present, the physiological significance of these two transcripts is unknown.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Northern analysis. RNA (8 µg) was fractionated on a 0.44 M formaldehyde-1.2% agarose gel, blotted, and hybridized to an -32P-labeled 1,062-bp internal PCR product from pepO. Molecular size markers (in kilobases) are shown at the left. The 16S rRNA runs at approximately 1.6 kb.

Construction of an S. parasanguis pepO mutant strain. A disrupted allele of pepO was constructed by inserting aphA-3 (50), a gene for kanamycin resistance (Km^r), into the KpnI restriction site of pepO (Fig. 1). The mutated allele was cloned into a plasmid unable to replicate in streptococcal cells, and the chromosomal mutation was obtained by the method of allelic replacement (19). The supercoiled plasmid pVT1333, carrying the disrupted pepO gene, was transformed into FW213 by electroporation. Transformants arising by allelic replacement were expected to be Km^r and tetracycline sensitive (Tc^s). Therefore, transformants were initially selected for Km^r and then screened for Tc^s. Ten transformants with the expected allelic replacement phenotype were screened by Southern hybridization analysis. The DNA from 9 of 10 transformants gave hybridization patterns indicative of allelic replacement (Fig. 4A). The pepO probe hybridized with both 6.0- and 1.9-kb EcoRI restriction fragments in the wild-type strain (Fig. 4A, lane 1). The 1.9-kb fragment was absent in the putative pepO mutants and replaced by a 3.4-kb fragment (Fig. 4A, lanes 2 and 3). The Southern blot was then stripped and probed with the Km^r determinant aphA-3 (Fig. 4B). The aphA-3 probe hybridized with the new 3.4-kb fragment in the pepO mutants (Fig. 4B, lanes 2 and 3) but not with fragments in the wild-type strain (lane 1). These data confirm that the shift in size of the fragment from 1.9 to 3.4 kb was due to insertion of the Km^r gene into pepO. Proper replacement of the wild-type pepO allele with the mutagenized allele was also confirmed by PCR analysis (data not shown). One pepO mutant, designated VT1346, was chosen for further study.

View larger version (76K): [in this window] [in a new window]   FIG. 4.   Southern hybridization analysis of the pepO region of wild-type S. parasanguis FW213 and two putative pepO mutant derivatives. Chromosomal DNA from each strain was digested with EcoRI and fractionated on a 0.75% agarose gel. DNA fragments were transferred to Hybond-N+ membranes and hybridized to either pVT1327 (A) or the Kmr determinant aphA3 (B) that had been -32P labeled. Lane 1, FW213 DNA; lanes 2 and 3, DNAs from two putative pepO mutant derivatives. Sizes in kilobases are noted at the left.

The pepO mutant adheres to SHA. It is unlikely that PepO acts directly as an adhesin in S. parasanguis FW213, but because of the close linkage of pepO to the operon encoding the fimbria-associated adhesin, fimA, we examined whether PepO plays a role in adhesion. No difference was found between wild-type and mutant strains in an SHA adhesion assay (data not shown), indicating that the inactivation of pepO did not affect the ability of S. parasanguis cells to adhere to SHA. We also found no obvious differences in growth rate or viability between wild-type and pepO mutant cells under standard growth conditions (data not shown).

FW213 PepO has activity similar to that of mammalian NEP. In the human brain, NEP is a major inactivator of endogenous opioid peptides, the enkephalins (38). NEP cleaves substrates at the amino side of hydrophobic amino acids, such as phenylalanine; in the case of the enkephalins, it cleaves at Gly³-Phe⁴. Based on the similarity between S. parasanguis PepO and mammalian NEP, we reasoned that S. parasanguis FW213 would hydrolyze physiological substrates of NEP but that the pepO mutant strain VT1346 would not. Endopeptidase activities of wild-type and pepO mutant strains were compared (Fig. 5) with an in vitro assay that included the substrate metenkephalin (Tyr-Gly-Gly-Phe-Met). Endopeptidase activity was detected in protein extracts from the wild-type strain but not from the pepO mutant strain. Wild-type cell extracts cleaved metenkephalin at the Gly³-Phe⁴ peptide bond, producing the products Tyr-Gly-Gly and Phe-Met (Fig. 5A). These products were not detected when extracts of the pepO mutant strain were incubated with the substrate (Fig. 5B). These data indicate that S. parasanguis FW213 pepO encodes an enzyme with activity similar to those of known mammalian endopeptidases.

View larger version (51K): [in this window] [in a new window]   FIG. 5.   TLC of metenkephalin hydrolysis by extracts from S. parasanguis FW213 and pepO mutant VT1346. Protein (15 µg) from FW213 (A) or VT1346 (B) was incubated with 1.25 mM metenkephalin for 0 to 10 min (corresponding to lanes) before termination of the reaction. Endopeptidase activity was detected by TLC. Abbreviations: F-M, phenylalanylmethionine; F, phenylalanine; M, methionine; Y, tyrosine; Y-G-G, tyrosylglycylglycine; Y-G-G-F-M, metenkephalin.

Phylogenetic analysis of PepO and its homologues. Phylogenetic analysis was included in this study because of the striking similarity between bacterial PepO sequences and their eukaryotic homologues. The S. parasanguis PepO sequence and 7 bacterial and 14 eukaryotic sequences (Table 2) were aligned and then analyzed by distance methods and by maximum parsimony. A phylogenetic tree (Fig. 6) was constructed from the multiple sequence alignment. Similar results were obtained with different methods of tree construction and with a variety of sequence sets (see Materials and Methods). A clear distinction between eukaryotic and prokaryotic species was seen in 100% of the bootstrap replicates, suggesting that the PepO homologues fall into two monophyletic groups. Within the bacterial branch of the tree, the phylogeny shown was generally consistent with the taxonomic classification given by the NCBI taxonomy browser (47a). What was exceptional about the phylogeny is that PepO homologues were not detected in completely sequenced genomes of organisms representing a broad range of taxa. These included archaea, fungi, early branching bacteria (Aquificales), proteobacteria, spirochetes, cyanobacteria, chlamydia, mycoplasmas, and Bacillus subtilis. These results suggested lateral gene transfer of pepO sequences between bacteria and eukaryotes.

View larger version (12K): [in this window] [in a new window]   FIG. 6.   Phylogenetic tree based on bacterial and eukaryotic PepO sequence homologues. The phylogenetic tree was constructed based on CLUSTAL W alignments of the amino acid sequences listed in Table 2. Numbers on the branches indicate the frequency (percentage) with which the corresponding cluster occurred during bootstrap analysis. ECE-1 and ECE-2, ECE isoforms 1 and 2, respectively; NEP, NEP 24.11; PEX, zinc metallopeptidase homologue (52).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The purpose of this study was to determine if any factors immediately upstream of the fimA operon were associated with adherence or survival within the host. Just 148 bp upstream and divergently transcribed from the fimA operon, a 1,893-bp ORF (pepO) that encodes a protein with many structural features in common with the M13 (neprilysin) family of metallopeptidases was identified. This family of metallopeptidases includes bacterial members as well as mammalian neuropeptide-processing, immunoregulatory, and peptide hormone-processing zinc metallopeptidases (36). While the role of the bacterial enzymes is uncertain, the mammalian enzymes have been shown to play essential roles in events such as inflammatory response phenomena and in pain and cardiovascular regulation (45, 51, 52).

FW213 PepO contains the essential HEXXH consensus motif typical of this family of peptidases as well as numerous positionally conserved residues known to be critical for the activity of the mammalian enzymes. Some of the most important residues in the active sites of NEP and ECE are included in the HEXXH consensus motif. In mammalian NEP, the two histidines (His⁵⁸³ and His⁵⁸⁷) are zinc-coordinating residues, and the glutamate (Glu⁵⁸⁴) plays an important role in catalysis (38, 44). In PepO, each of these important residues as well as many other residues surrounding the critical HEXXH consensus motif are positionally conserved (Fig. 2). In addition to the conserved HEXXH motif, FW213 PepO contains other positionally conserved residues known by site-directed mutagenesis to be important in zinc binding, substrate binding, or hydrolytic activity in NEP (Fig. 2). These include (in NEP) an asparagine (Asn⁵⁴²), an alanine (Ala⁵⁴³), and a valine (Val⁵⁸⁰) that are involved in substrate binding; a third Zn-coordinating ligand (Glu⁶⁴⁶); an aspartic acid that is critical for hydrolytic activity (Asp⁶⁵⁰); and a histidine (His⁷¹¹) that may be involved in transition state binding (14, 30, 38, 55).

pepO appears to be transcribed as a monocistronic message. One major transcript and one minor transcript, differing in size by about 200 bases, were consistently detected by Northern hybridization. The 2.1-kb major transcript is within the size range expected from the ORF that is predicted by the pepO sequence; the minor transcript is slightly larger. It is possible that these two transcripts are the result of multiple transcription initiation sites or differential RNA processing. The physiological relationship between the two transcripts in S. parasanguis is currently unknown. Three distinct NEP mRNAs that differ only in their 5' noncoding regions have been identified for humans and rats, and data suggest that their expression is regulated developmentally and in a tissue-specific manner (31, 38). The expression of S. parasanguis pepO may also be regulated, and we suggest that this regulation may be related to the expression of the fimA operon. The tight physical arrangement of pepO and the fimA operon, as well as the overlapping arrangement of their 5' regulatory regions, suggests the possible cotranscription of these genes. Further analysis is necessary to determine if specific sequences within the 148-bp region separating pepO and the fimA operon are involved in their coregulation.

An investigation of the physiological role of PepO in S. parasanguis was initiated by constructing a mutant strain of FW213. Initially, the adhesion and growth properties of the mutant were examined. PepO does not appear to play a role in the adhesion of S. parasanguis cells in an in vitro SHA tooth surface model. This finding is not surprising, since it is known that FimA is not required for binding to SHA (17). An important, but as-yet-unanswered question, is whether PepO plays a role in S. parasanguis endocarditis, as does FimA (6). In this regard, it is of interest that a putative zinc metallopeptidase, MEP1, which has homology to the M13 family of metallopeptidases (including mammalian NEP and now S. parasanguis PepO), has been identified for the highly pathogenic blood-feeding nematode Haemonchus contortus (37). H. contortus infects the fourth division of the stomach of sheep, goats, and other small ruminants and can induce anemia, weight loss and, in severe cases, death of the animal. The putative zinc metallopeptidase MEP1 is developmentally regulated, with expression being limited to the blood-feeding stages of H. contortus. Whereas we found that the disruption of pepO did not affect the growth characteristics of FW213 in TH broth under standard conditions, we observed recently that wild-type S. parasanguis FW213 forms larger colonies on blood agar plates than the pepO mutant. S. parasanguis PepO may also play a role during growth in the presence of blood, such as when streptococcal cells gain access to the bloodstream.

Within the M13 family of metallopeptidases, NEP is the prototype and the most thoroughly characterized. The biochemical activity of PepO was examined by testing the ability of wild-type and mutant strains to hydrolyze metenkephalin, a physiological substrate of mammalian NEP. Wild-type cell extracts cleaved metenkephalin at the same site as does NEP, producing the products Tyr-Gly-Gly and Phe-Met. Cell extracts from the pepO mutant failed to cleave metenkephalin at the Gly³-Phe⁴ peptide bond. These data indicate that the pepO gene in S. parasanguis FW213 encodes an enzyme with activity similar to that of NEP. Furthermore, these data emphasize the close correspondence between the active sites of S. parasanguis PepO and mammalian NEP. The similarity of PepO to NEP, in terms of conservation of critical amino acid residues and substrate specificity, demonstrates that PepO is a member of the M13 family of metallopeptidases.

Homologues of the PepO endopeptidase occur in many animals but in only a limited group of bacteria, the majority of which are gram positive and pathogenic for humans. We wonder about the origin of this endopeptidase in these diverse organisms. Since the phylogenetic trees constructed in this study are consistent with species trees, it is possible that the ultimate ancestor of PepO and its animal homologues was present in an organism that gave rise to both modern bacteria and eukaryotes. However, PepO sequence homologues were not detected by the BLAST search of completely sequenced genomes representing a broad range of taxa, including those from groups that diverged early. While this finding could be explained by postulating the widespread loss of PepO sequences from this broad range of taxa, it is highly unlikely. A more plausible explanation is that PepO sequences were acquired through horizontal genetic exchange between bacteria and eukaryotes. Given the distribution of PepO homologues that we found, this event probably took place after the divergence of the eubacterial, archaeal, and eukaryotic lineages. Rawlings and Barrett (36) have proposed that the genes for a number of bacterial metallopeptidases were transferred into eukaryotic cells by endosymbiotic organisms that gave rise to organelles. Transfer of PepO probably did not occur by this mechanism, given that PepO sequences have not been found in proteobacteria and recent data indicating that the -proteobacterium Rickettsia prowazekii and mitochondria are specific evolutionary relatives (3). Whereas the direction of genetic transfer cannot be determined from the information currently available, it is clear from the comparison of PepO homologues that the protein, particularly in the region encompassing the active site, is under strong evolutionary constraints.

The PepO homologues play critical roles in eukaryotic physiology, and it is likely that PepO plays an important role in the relatively small group of bacteria that have maintained or acquired similar sequences. Future studies are aimed at elucidating the physiological role of PepO in S. parasanguis. PepO may be useful as a model for understanding the regulation of this family of enzymes.