The AprV5 Subtilase Is Required for the Optimal Processing of All Three Extracellular Serine Proteases from Dichelobacter nodosus

Dichelobacter nodosus is the principal causative agent of ovine footrot and its extracellular proteases are major virulence factors. Virulent isolates of D. nodosus secrete three subtilisin-like serine proteases: AprV2, AprV5 and BprV. These enzymes are each synthesized as precursor molecules that include a signal (pre-) peptide, a pro-peptide and a C-terminal extension, which are processed to produce the mature active forms. The function of the C-terminal regions of these proteases and the mechanism of protease processing and secretion are unknown. AprV5 contributes to most of the protease activity secreted by D. nodosus. To understand the role of the C-terminal extension of AprV5, we constructed a series of C-terminal-deletion mutants in D. nodosus by allelic exchange. The proteases present in the resultant mutants and their complemented derivatives were examined by protease zymogram analysis, western blotting and mass spectrometry. The results showed that the C-terminal region of AprV5 is required for the normal expression of protease activity, deletion of this region led to a delay in the processing of these enzymes. D. nodosus is an unusual bacterium in that it produces three closely related extracellular serine proteases. We have now shown that one of these enzymes, AprV5, is responsible for its own maturation, and for the optimal cleavage of AprV2 and BprV, to their mature active forms. These studies have increased our understanding of how this important pathogen processes these virulence-associated extracellular proteases and secretes them into its external environment.


Introduction
Dichelobacter nodosus is a slow growing gram-negative anaerobic bacterium that is the principal causative agent of footrot in ruminants. Ovine footrot is characterized by the separation of the keratinous hoof from the underlying tissue, leading to lameness, loss of body weight, and reduced wool growth and quality, which results in significant economic losses [1]. Clinical disease is dependent upon the virulence properties of the D. nodosus isolate and the presence of warm wet climatic conditions. There are two major clinical forms of footrot, benign and virulent, with benign footrot presenting as an interdigital dermatitis with no further disease progression whilst virulent disease leads to a severe underrunning of the horn of the hoof. The virulence factors of D. nodosus include its extracellular subtilisin-like serine proteases (or subtilases) [2][3][4][5][6], type IV fimbriae [7,8] and potentially, the vrl and vap genomic islands, which are preferentially associated with virulent strains [9][10][11][12].
Subtilases are produced by a wide variety of archaea, bacteria, fungi and eukaryotes. They mostly are synthesized as pre-pro-enzyme precursors that are subsequently translocated across the cell membrane and then activated by cleavage of the 27-280 residue pro-domains [13]. The relatively conserved catalytic domains contain 268-511 residues, with some members of the subtilase family also having C-terminal extensions with variable lengths and less sequence conservation [13]. The function of the C-terminal extensions is also variable. Studies on aqualysin I from Thermus aquaticus YT-1 [14], the serine protease SSP from Serratia marcescens [15] and the IgA protease from Neisseria gonorrhoeae [16] showed that the C-terminal extensions of these proteins are required for extracellular secretion of the proteases, while the C-terminal extension of the subtilase SISBTS from the tomato plant is required for both pro-domain processing and secretion [17]. By contrast, it has been shown that the C-terminal extension of MCP-03, a subtilase from Pseudoalteromonas sp. SM9913, is not crucial for secretion, but instead negatively affects catalytic efficiency and is essential for protease thermostability [18]. Studies on the Lactococcus lactis protease suggested that some residues in the C-terminal extension affect catalytic activity [19].
Three closely related subtilases are secreted by virulent strains of D. nodosus: two acidic proteases, AprV2 and AprV5, and the basic protease BprV. The equivalent proteases in benign strains are AprB2, AprB5 and BprB [20,21]. AprV5 has been shown to be responsible for most of the extracellular protease activity in a virulent strain of D. nodosus, VCS1703A, although AprV2 is responsible for most of the extracellular elastase activity and is essential for virulence [6]. These proteases are encoded by separate genes and are synthesized as precursors with an Nterminal pre-pro-region, a serine protease domain and a Cterminal extension. Based on the cleavage sites determined previously [20,21], the estimated sizes of the mature domains of AprV5, AprV2 and BprV are 36.0 kDa, 36.4 kDa and 35.8 kDa, respectively; their respective C-terminal extensions are 13.9 kDa, 14.0 kDa and 14.1 kDa in size. The amino acid sequences and structures of the catalytic protease domains are highly conserved (,65% identity) [6,22], but the sequences of the C-terminal extensions are less conserved, with only approximately 35% similarity. The C-terminal extension of AprV5 contains a P-domain, which is typically associated with eukaryotic pro-protein convertases that belong to the subtilisinlike superfamily [13,23]. The active proteases are produced by cleavage of the pre-pro region and the C-terminal extension [2,24,25]. The mechanism of protease processing and secretion is unknown, although type IV fimbriae are required for optimal secretion [7,8,26].
In this study, we aimed to examine the mechanism of processing of the extracellular proteases from D. nodosus and to determine the role of their C-terminal extensions. To achieve these aims we constructed a series of C-terminal truncated aprV5 mutants in D. nodosus by allelic exchange. The results showed that AprV5 was required for efficient processing of AprV2 and BprV and that the C-terminal extension of AprV5 was required for maximal AprV5 activity.

DNA Manipulation and Molecular Techniques
Unless otherwise stated, standard procedures were used [27]. D. nodosus genomic DNA was prepared using a QIAGEN DNeasy kit, according to the manufacturer's instructions. Southern hybridizations were performed as described previously [26,29], with probes specific for the relevant antibiotic resistance genes or the target genes. Reverse transcriptase (RT)-PCR was carried out as previously described [8]. DNA sequencing was performed using an Applied Biosystems 3730S Genetic Analyser. Sequence data were compiled using Sequencher version 3.0 (Gene Codes Corporation) or Vector NTI advance TM 11 (Invitrogen).

Construction of Suicide Vectors and Mutants of D. nodosus
All of the suicide vectors used in this study were constructed in pUC18K [30], which does not replicate in D. nodosus. They contained two 1.2-kb to 1.9-kb D. nodous-derived DNA fragments that were located on either side of an aphA-3 kanamycin-resistance cassette, with a bla ampicillin-resistance gene located on the plasmid vector. Therefore, D. nodosus transformants that were derived from double crossovers could be obtained by selecting for kanamycin resistance and screening for susceptibility to ampicillin.
To construct an aprV5D478-595 mutant, the suicide vector pJIR3520 was constructed as follows. A 1.5-kb PCR product that contained the aprV5 gene region that encoded AprV5 residues 1-477 was cloned into the SacI/KpnI sites of pUC18K followed by the cloning of a 1.5-kb PCR fragment located downstream of aprV5 into the BamHI/XbaI sites of the resultant plasmid. To construct the aprV5D503-595 suicide vector pJIR3708, the 1.5-kb aprV5D478-595 fragment of pJIR3520 was deleted by SacI/KpnI digestion and replaced with a 1.4-kb PCR product containing aprV5D503-595. The aprV5D560-595 suicide vector pJIR3729 was constructed in a similar way. It contained the 1.5-kb aprV5 downstream region inserted into BamHI/XbaI sites and a 1.8-kb fragment containing aprV5D560-595 that was obtained by EcoRI/ SacI digestion from pJIR3695, a plasmid containing a full-length aprV5 gene cloned into the KpnI site of pGEM7zf (2). Finally, an aprV5His 6 vector (pJIR3661) was constructed in a similar manner by cloning a 1.2-kb PCR product containing an N-terminal truncated aprV5 plus a hexahistidine-encoding sequence into the KpnI site and cloning a 1.2-kb PCR product downstream of aprV5 into the XbaI/PstI sites of pUC18K.
To complement the truncated aprV5 mutants, a 1.2-kb PCR product from the region downstream of aprV5 was cloned into the XbaI/PstI sites of pUC18K and then a 1.9-kb fragment containing the region encoding AprV5 C-terminal extension was cloned into the KpnI site of the resultant plasmid to form pJIR3711. To enable an antibiotic selection for the complemented strains, aphA-3 was deleted from pJIR3711 by SmaI digestion and replaced by an erm(B) cassette from pJIR2412 [26] to form the complementation vector pJIR3730. The complemented strains were selected on erythromycin and screened for susceptibility to kanamycin and ampicillin.
Plasmid DNA was introduced into D. nodosus by natural transformation and mutants were constructed by homologous recombination between the double crossover suicide vectors and the D. nodosus chromosome, as described previously [7,8]. Transformants were initially screened by their resistance or susceptibility to the appropriate antibiotics as described above. Potential mutants were analysed by PCR using relevant antibioticresistance gene and target gene primers. Southern hybridization was used to confirm that the C-terminal extension encodingregion had been deleted or that a 6xHis-encoding tag had been incorporated onto the C-terminus of aprV5 by double crossover events.

Extracellular Protease Assays
Caseinase activity was detected qualitatively by growing D. nodosus colonies on EYE agar containing 2% (wt/vol) skim milk [31]. Total protease activity in the culture supernatant was determined quantitatively using azocasein (Sigma) as the substrate, as previously described [7]. One unit of protease activity is defined as the amount of enzyme required to digest 1 mg of azocasein per min [7]. Protease zymograms were carried out as previously described [32], except that they were run on a BioRad Mini-Protean RIII apparatus. D. nodosus culture supernatant (10 ml) was mixed with 5 ml of sample buffer (50% (wt/vol) sucrose, 0.1% (wt/ vol) bromophenol blue) and subjected to electrophoresis at 100 V for approximately 2 h, until the dye reached the bottom of the gel.
After electrophoresis, the polyacrylamide gel was placed on a glass plate, overlaid with gelatin/agarose and incubated for 75 min at 37uC in a moist chamber. The gelatin/agarose gel was removed and developed by submerging it in saturated ammonium sulphate solution that had been heated to approximately 60uC. The gel was washed in tap water prior to photography.

Mass Spectrometry and Protein Identification
Protein bands of interest were excised from Coomassie Brilliant Blue stained gels and prepared for MS analysis as described previously [33]. After the washing and dehydration steps the gel pieces were rehydrated with a solution containing 0.5 mg of sequencing grade trypsin, incubated at 37uC overnight and sonicated, Digested samples were co-spotted onto the target plate with matrix solution (10 mg/ml of a-cyano-4-hydroxycinnamic acid) in 50% acetonitrile-0.1% TFA) and analysed on an Applied Plasmids pUC18K pUC18 SmaIVaphA-3, nonpolar base vector [30] pGEM7Zf (2) A p r , lacZ + cloning vector Promega Biosystems 4700 Proteomics Analyser MALDI TOF/TOF in reflectron mode with a mass range of 800 to 3500 Da, with a focus mass of 1400 Da at 1500 shots per spectra. The 4700 Series Explorer software automatically selected the 15 most intense peptides as precursor masses for MS/MS analysis, acquired in the order of decreasing intensity. MS/MS analysis was carried out in reflectron mode with a relative precursor mass window of 50 resolution with metastable ion suppression on and spectra summed at 2500 shots/spectrum. PMF and MSMS data were compiled by GPS explorer software Ver. 3 (build 311) (Applied Biosystems) and searched using the MASCOT search engine (version 1.9, Matrix Science Inc.,) against an in-house database compiling the FASTA formatted Dichelobacter nodosus genome obtained from the National Centre for Biotechnology Information (NCBI) FTP site. The following search parameters were used: missed cleavages, 1; peptide mass tolerance, 650 ppm; peptide fragment tolerance, 60.1 Da; peptide charge, 1+; Variable modification, oxidation (Methionine). Positive matches were accepted based on the protein score above a significance threshold as defined by MASCOT (www. matrixscience.com). Identifications below this threshold were manually verified by interpretation of the MSMS spectra of matched peptides and matched MS peaks.

The C-terminal Extension of AprV5 is Required for Processing of AprV2
In studies aimed at determining the role of these extracellular proteases in virulence we previously constructed mutations in each of the three D. nodosus extracellular protease genes [6]. To determine if mutation of any of these proteases altered the production of the other two enzymes the extracellular protease activity of these insertionally-inactivated aprV5, aprV2 and bprV mutants and their complemented derivatives was characterized by protease zymogram analysis after native gel electrophoresis. The results ( Figure 1) showed that, as expected, in the aprV2 mutant there was no AprV2 activity, as represented by the two missing isoenzyme activity bands; these bands were restored by complementation. Mutation of aprV2 had no effect on AprV5 activity (the band of highest electrophoretic mobility). By contrast, in the aprV5 mutant the AprV5 band was absent as expected and the two AprV2 bands were either not present (Figure 1), or in some preparations were very greatly reduced in intensity; all of these protease bands were restored in the complemented aprV5 derivative (Figure 1). Based on these studies we postulated that AprV5 was required for the correct processing of AprV2. Finally, there was no difference in the active AprV2 or AprV5 bands in the bprV mutant ( Figure 1). However, an additional protease activity band, of slightly greater electrophoretic mobility than the faster of the AprV2 bands, was observed in this strain. This band disappeared in the bprV-complemented derivative, but although difficult to see in this specific preparation; both of the AprV2derived protease bands were always present in supernatants from this strain. Note that the basic protease BprV can not be observed by this method, as a result of its basic pI. Attempts to visualise BprV activity by carrying out zymogram analysis under different electrophoretic conditions were unsuccessful.
To determine if the C-terminal domain of AprV5 was involved in the processing of AprV2 we again used a genetic approach. The aprV5 sequence encoding the previously determined C-terminal extension, residues 478-595 [25], was deleted by allelic exchange between the suicide vector pJIR3520 and the wild-type chromosome. The genotype of the C-terminal-truncated strain aprV5D478-595, which was designated as aprV5DC1 (Figure 2), was confirmed by PCR analysis and Southern hybridization. To ensure that the C-terminal deletion had no effect on protease gene expression, RT-PCR analysis using aprV5-, aprV2and bprVspecific primers was performed. The results showed that aprV5, aprV2 and bprV were all transcribed in this mutant (data not shown). We subsequently constructed two additional AprV5-Cterminal truncated strains, aprV5D503-595 (aprV5DC2) and aprV5D560-595 (aprV5DC3) (Figure 2), which also were confirmed by PCR and Southern hybridization. All three mutants were complemented by using homologous recombination to reconstitute the deleted C-terminal-encoding aprV5 regions.
Qualitative protease analysis on skim-milk agar was carried out on the C-terminal deletion strains and their complemented derivatives. The results showed that after 24 h incubation, all of the aprV5DC mutants had reduced protease activity compared to the wild-type strain; this activity was restored in each of the complemented derivatives (data not shown). These results were confirmed by quantitative protease assays carried out on culture supernatants of 16 h (mid-logarithmic growth phase), 25 h (late logarithmic/early stationary phase) and 40 h (stationary phase) TAS broth cultures of each strain, using azocasein as the substrate. An aprV5 null mutant [6] was included as a negative control in this experiment. The 16 h and 25 h samples, which represented late logarithmic and stationary growth phases, respectively, of each of the aprV5DC mutants, had significantly lower levels (p,0.05; student's t-test) of extracellular protease activity than the wild-type strain (Figure 3). There was no significant difference in activity in the 16 h culture supernatants between any of the mutants, Gelatin was used as a substrate overlay of supernatants separated by native polyacrylamide gel electrophoresis [32] to screen AprV5 and AprV2 activity in 25 h TAS broth cultures from the wild-type strain VCS1703A (WT), the aprV5 mutant JIR3756 (aprV5), the aprV2 mutant JIR3743 (aprV2), the bprV mutant JIR3928 (bprV) and the complemented derivatives JIR3883 (aprV5(aprV5 + )), JIR3900 (aprV2(aprV2 + )) and JIR3930 (bprV(bprV + )), respectively. doi:10.1371/journal.pone.0047932.g001 although they all had significantly less protease activity than the wild type (p,0.05). However, the protease activity in 25 h culture supernatants of the aprV5 and aprV5DC1 mutants was significantly lower (p,0.05) than that of the aprV5DC2 and aprV5DC3 mutants. By contrast, at 40 h there was no significant difference between the aprV5DC mutants and the wild-type strain. The reduced protease activity of 16 h and 25 h cultures of each aprV5DC strain was restored to the wild-type level upon reconstitution (Figure 3). These results implied that deletion of the C-terminal domain of AprV5 delays the appearance of fully functional extracellular protease in the culture supernatant of D. nodosus.
To determine if the C-terminal deletions of AprV5 affected the processing of AprV2 in a similar manner to the aprV5 null mutant (Figure 1), protease zymogram analysis was carried out. The results showed that the AprV5 activity in the 16 h supernatants from each aprV5DC mutant was significantly reduced ( Figure 4); moreover, AprV2 activity also was reduced in these samples. In addition, extra bands of lower electrophoretic mobility than AprV2 were present in these samples, presumably unprocessed forms of the proteases. At 25 h the AprV5 and AprV2 activity bands in the aprV5DC mutants were not discernibly different to those of the wild-type strain (Figure 4), although in the aprV5DC1 and aprV5 null mutant, the extra lower mobility bands were still present, which suggested that there was a delayed processing of proteases in these two mutants. Note that in each of the mutants all of these effects were reversed when the wild-type aprV5 gene was reconstituted.

AprV5 is Required for Optimal Processing of AprV2 and BprV
To further examine the effect of the C-terminal extension of AprV5 on each of the three individual proteases, equal amounts of the concentrated supernatants from the aprV5DC mutants were analysed on 4-15% gradient SDS-PAGE gels and blotted with specific antisera ( Figure 5). Note that in these experiments the results were complicated by unavoidable immunological crossreaction between these closely related proteases. Accordingly, the individual aprV5, aprV2 and bprV null mutants isolated previously [6] were used as negative controls alongside the wild-type positive control.
Western immunoblotting with AprV5-specific antisera showed that the mature AprV5 band produced by the aprV5DC1 mutant was slightly smaller in size than the equivalent band in the wild type. This band was missing in the aprV5 null mutant. By contrast, the AprV5 band in the aprV5DC2 and aprV5DC3 mutants was the same size (approx. 40 kDa) as that in the wild-type strain (Figure 5a). These results indicated that the AprV5 C-terminal extension cleavage site was located between residues 478-502, not at residue 478 as previously suggested [25]. In an attempt to determine the C-terminal cleavage site of AprV5, we introduced codons encoding a 6xHis tag onto the 39-end of aprV5 in D. nodosus by allelic exchange, with the objective of carrying out N-terminal sequence analysis on the cleaved C-terminal extension-6xHis fragment. The resultant aprV5 His6 strain (JIR3953) was confirmed as before and protease analysis on skim-milk agar showed that introduction of the 6xHis tag did not affect overall protease production (data not shown). However, Western immunoblotting using 6xHis antibody could not detect a 6xHis-tagged protein from either culture supernatants or whole cell extracts after repeated attempts, suggesting that the C-terminal extension of AprV5 was degraded after protease processing.
The results also showed that in the aprV5DC2 and aprV5DC3 strains there was a subtle reduction of the amount of mature AprV5 at 16 h and 25 h, compared to the wild-type strain and the complemented derivatives. In addition, there were protein bands of higher molecular size (approx. 50 kDa) that specifically reacted with the AprV5 antisera. We postulated that these bands represented pro-AprV5, which was supported by the observation that in aprV5DC1, these proteins were smaller in size than in aprV5DC2 and aprV5DC3, which was consistent with the fact that the mature AprV5 band in this strain was smaller in size. Note that all of the complemented derivatives showed a similar profile to the wild-type strain and that at 40 h all of the deletion mutants had a wild-type profile, with no larger immunoreactive bands. These data support the hypothesis that the C-terminal extension of AprV5 is required for efficient protease processing. Loss of the Cterminal extension delays processing to the mature form.
Western immunoblotting using AprV2-specific antiserum confirmed that AprV5 had a role in the maturation of AprV2. The results showed that the amount of the mature AprV2 band, which migrated at approx. 38 kDa, was greatly reduced in the 16 h and 25 h cultures of the aprV5, aprV5DC1, aprV5DC2, and aprV5DC3 mutants and that two groups of larger proteins (approx. 65 kDa and 50 kDa) were present in these samples (Figure 5b), these were presumably less processed forms of AprV2. Similar results were obtained when BrpV-specific antiserum was used, with less processed forms of BprV also detected (Figure 5c). Once more, reconstitution of the intact aprV5 gene reversed this effect and no effects on protease processing were observed at 40 h. These data provided clear evidence that AprV5 was involved in the processing of both AprV2 and BprV. In the aprV2 and bprV mutants there was no reduction in the mature forms of the other two proteases or any evidence for the accumulation of larger proteins (Figure 5a, b & c). Therefore, it appeared highly unlikely that AprV2 and BprV processed each other or AprV5. Western immunoblotting also was carried out on whole cell lysates of each strain using AprV2 and BprV antisera. The results showed that there were no AprV2 or BprV bands present in the whole cell lysates (data not shown), confirming that AprV2 and BprV are processed after secretion.
To confirm that the novel proteins accumulated in the aprV5 mutants were unprocessed forms of AprV2 and BprV MALDI-TOF-TOF (MS-MS) analysis was performed on in-gel proteins excised from the aprV5, aprV5DC1 and aprV5DC3 lanes. The results (Table 2) revealed that in the extracted 65 kDa bands both pro-AprV2-CTE and pro-BprV-CTE were identified, whereas the 50 kDa bands contained AprV2-CTE and BprV-CTE. Unfortunately, no unprocessed AprV5-derived peptides were detected, despite repeated analyses.

Discussion
Virulent isolates of D. nodosus produce three homologous extracellular subtilases, AprV5, AprV2 and BprV, each of which is synthesized as a precursor molecule that is subsequently processed to form the mature active enzyme by an unknown mechanism. In this study, we have shown that AprV5 is responsible for its own maturation and for optimal processing of AprV2 and BprV. Our results have also revealed that the Cterminal extension of AprV5 is required for efficient processing of all three enzymes, presumably because it is required for the optimal processing of AprV5. In the absence of this domain, Figure 4. Extracellular protease zymograms of the aprV5 C-terminal deletion strains. Cultures from 16 h, 25 h and 40 h TAS broth cultures of the wild-type strain VCS1703A (WT), the aprV5D478-595 strain JIR3947 (aprV5DC1) and its complemented derivative JIR3968 (aprV5DC1(V5 + )), the aprV5D503-595 strain JIR3956 (aprV5DC2) and its complemented derivative JIR3965 (aprV5DC2(V5 + )) and the aprV5 mutant JIR3756 (aprV5) were examined. The profile of the aprV5D560-595 strain JIR3969 (aprV5DC3) is not shown but was identical to that of aprV5DC2. doi:10.1371/journal.pone.0047932.g004 protease processing is delayed. Finally, we have provided evidence that the cleavage of the pro-domain and the C-terminal extensions of the AprV2 and BprV precursors occurs after secretion.
It has been reported that P-domain in some bacterial subtilases is required for both the folding and regulation of pH dependence of the catalytic domain [23,34,35]. In this study we have shown that AprV5 proteins lacking the C-terminal P-domain exhibited delayed production of the mature enzyme as a result of delayed pro-protein-C-terminal extension processing. This effect was not due to a secretion defect. We postulate that the C-terminal extension may function as an intramolecular chaperon that optimizes the folding of the precursor protein prior to proteolytic cleavage.
The processing of serine proteases involves autocatalytic cleavage of each of the pro-and C-terminal domains after membrane translocation [13,36]; a similar process appears to occur for AprV5. By contrast, our results showed that the optimal maturation of AprV2 and BprV, including cleavage of the proregion and the C-terminal extension, requires a functional AprV5 enzyme; processing was delayed in the absence of the P-domain of AprV5. There are precedents for these observations. In eukaryotes, subtilisin-like pro-protein convertases are responsible for the maturation of hormones and other proteins [37]. In bacteria, the subtilisin-like autotransporter protease SphB1 is essential for the maturation of the precursor of the secreted protein FhaB in B. pertussis [36]. We conclude that in D. nodosus the maturation of these proteases must occur very rapidly after secretion since the wild-type strain did not accumulate unprocessed forms of AprV2 and BprV at any growth stage and we could not detect a Histagged C-terminal extension fragment in the aprV5 His6 derivative. The fact that we observed pro-mature-CTE and mature-CTE forms, but not pro-mature forms, of AprV2 and BprV in the aprV5 mutants suggests that the C-terminal extensions of AprV2 and BprV are processed after pro-region cleavage, which provides further support for the hypothesis that the C-terminal extension plays an important role in the formation of the mature enzyme. It appears that AprV2 and BprV do not process each other or AprV5 since no unprocessed proteases were detected in the aprV2 or bprV mutants, although an active band with a different electrophoretic mobility was observed in the bprV mutant, suggesting that BprV may have some role in protease processing.
Studies on the C-terminal extensions of other subtilases suggested that they formed a b-barrel structure that played a role in the protease secretion process [14,16,36,38]. In agreement with these studies, it previously was suggested that AprV5, AprV2 and BprV were potential autotransporters, in which the C-terminal extension forms a b-barrel structure on the outer membrane to allow the proteases to be secreted [39]. However, our analysis indicates that their C-terminal extensions are not large enough to form a b-barrel structure. To determine if the C-terminal extension played a role in secretion, we constructed three separate C-terminal AprV5 deletion mutants. We found that there was no significant difference in the levels of AprV5 secreted into the culture supernatants in these mutant strains and the wild-type strain, indicating that the C-terminal extension was not essential for secretion. These results are in agreement with our previous finding that protease secretion involves the type IV fimbrial system [7,8,26].
Finally, previous studies provided evidence that the cleavage sites of the C-terminal extensions of AprV5 was at amino acid residues V477 [25]. By constructing strains expressing C-terminal truncated proteins of AprV5 we observed that the actual Cterminal cleavage site of AprV5 was different to that previously reported and is slightly closer to the C-terminus of AprV5, between residues 478 and 502.
In conclusion, we have demonstrated that the C-terminal extension of AprV5, which contains a P-domain similar to that of eukaryotic subtilisin-like proteases, is required for optimal processing and active enzyme production. We have also shown that AprV5 is responsible for the maturation of the precursors of AprV2 and BprV into their mature forms. These studies have made a major contribution to our understanding of how this pathogenic bacterium, which unusually produces three closely related subtilisin-like proteases, processes these proteases into their mature active forms.