A novel secreted protease from Pseudomonas aeruginosa activates NF-κB through protease-activated receptors

Authors


*E-mail kuwano@med.kurume-u.ac.jp; Tel. (+81) 942 31 7548, Fax: (+81) 942 31 0343

Summary

The Pseudomonas aeruginosa-derived alkaline protease (AprA), elastase A (LasA), elastase B (LasB) and protease IV are considered to play an important role in pathogenesis of this organism. Although the sequence analysis of P. aeruginosa genome predicts the presence of several genes encoding other potential proteases in the genome, little has been known about the proteases involving in pathogenesis. Recently, Porphyromonas gingivalis gingipains and Serratia marcescens serralysin have been shown to activate protease-activated receptor 2 (PAR-2), thereby modulating host inflammatory and immune responses. Accordingly, we hypothesized that unknown protease(s) from P. aeruginosa would also modulate such responses through PARs. In this study, we found that P. aeruginosa produces a novel large exoprotease (LepA) distinct from known proteases such as AprA, LasA, LasB and protease IV. Sequence analysis of LepA showed a molecular feature of the proteins transported by the two-partner secretion pathway. Our results indicated that LepA activates NF-κB-driven promoter through human PAR-1, -2 or -4 and cleaves the peptides corresponding to the tethered ligand region of human PAR-1, -2 and -4 at a specific site with exposure of their tethered ligands. Considered together, these results suggest that LepA would require PARs to modulate various host responses against bacterial infection.

Introduction

The pathogen Pseudomonas aeruginosa is responsible for a variety of diseases, including lung infection with cystic fibrosis, burn wound infections, hospital-acquired pneumonia and keratitis (Lyczak et al., 2000). Major virulence factors produced by this opportunistic pathogen include secreted proteases that damage host tissues. Several P. aeruginosa proteases have been isolated and shown to be involved in pathogenesis. Of the proteases analysed, alkaline protease, elastase A, elastase B and protease IV have been characterized extensively. Alkaline protease (also termed AprA or aeruginolysin), a metalloprotease with a molecular mass of approximately 50 kDa, degrades complement components C1q and C3, as well as cytokines and chemokines (Horvat and Parmely, 1988; Parmely et al., 1990; Hong and Ghebrehiwet, 1992; Leidal et al., 2003), suggesting that AprA could potentially function as an immunomodulatory agent during infection (Kharazmi, 1991). Elastase A (also termed LasA or staphylolysin), a metalloprotease with a molecular mass of approximately 27 kDa, has been shown to play a role in the pathogenesis of corneal and lung infections (Coin et al., 1997; Preston et al., 1997). Elastase B (also termed LasB or pseudolysin), a metalloprotease with a molecular mass of approximately 33 kDa, is suspected to be important for virulence in burn and respiratory infections (Pavlovskis and Wretlind, 1979; Suter, 1994; Sawa et al., 1998). Protease IV, a serine protease with a molecular mass of approximately 26 kDa, has been shown to be a virulence factor in keratitis (Traidej et al., 2003; Caballero et al., 2004).

Protease-activated receptors (PARs) were discovered members of the seven-transmembrane domain superfamily of G-protein-coupled receptors and are characterized by a novel mechanism of activation (Macfarlane et al., 2001). Unlike classical receptor–agonist interactions where each molecule is a separate entity, PARs are tethered to their activating ligand. In the resting state, the tethered ligand is unable to induce receptor activation because it is bound to an inhibitory amino-terminus. Activation occurs when protease activity cleaves the amino-terminus, allowing the newly exposed tethered ligand to dock with the second extracellular loop of the PAR (Macfarlane et al., 2001; Hollenberg and Compton, 2002). To date, four PARs have been identified; three of them (PAR-1, -3 and -4), although PAR-1 and -4 can be activated by trypsin, are activated mainly by thrombin, and the fourth (PAR-2), although unresponsive to thrombin, can be activated by a large number of other proteases, including trypsin, tryptase, factor Xa, neutrophil protease 3, membrane-tethered serine protease 1 and house dust mite proteases (Ossovskaya and Bunnett, 2004; Reed and Kita, 2004). In addition to enzymatic activation, PAR-1, -2 and -4 can be activated selectively by exogenous hexapeptide amides that share the putative tethered ligand sequence of each PAR subtype. These peptides can activate PARs independently of proteolysis (Ossovskaya and Bunnett, 2004).

The PARs activation induces the G-protein-coupled signal transduction [e.g. activation of phospholipase C (PLC), increased intracellular Ca2+ and activation of protein kinase C (PKC)] (Macfarlane et al., 2001; Ossovskaya and Bunnett, 2004). The activation of PLC leads to increment of intracellular Ca2+, thereby activating PKC, which leads to activation of mitogen-activated kinase cascades (extracellular signal-regulated kinase 1/2 pathway, c-Jun N-terminal kinase pathway and p38 kinase pathway). Alternatively, the PKC is shown to lead to activation of nuclear factor-κB (NF-κB) (Kanke et al., 2001).

Protease-activated receptors are expressed in various tissues and cells, including circulatory, gastrointestinal, respiratory, immune, haematopoietic and central nervous systems (Macfarlane et al., 2001; Hollenberg and Compton, 2002; Ossovskaya and Bunnett, 2004). Therefore, PARs activation contributes to a variety of physiological and pathophysiological roles, including development, growth, immunity, inflammation, tissue repair and pain. Interestingly, bacterial proteases such as Porphyromonas gingivalis gingipains and Serratia marcescens serralysin, as well as trypsin and also several other trypsin-like serine proteases, activate PAR-2, thereby modulating host inflammatory and immune responses (Lourbakos et al., 1998; 2001a,b; Abraham et al., 2000; Oido-Mori et al., 2001; Holzhausen et al., 2006; Inomata et al., 2007; Kida et al., 2007). Thus, the activation of PAR-2 appears to play a crucial role for host inflammatory and immune responses in bacterial infection (Shpacovitch et al., 2007). Although P. gingivalis gingipains has been shown to activate PAR-1 and -4 (Lourbakos et al., 2001a,b), not much is available on the activation of other PARs by bacterial proteases and the role of other PARs in bacterial infection.

As described above, AprA, LasA, LasB and protease IV are considered to play an important role in pathogenesis of P. aeruginosa infection. Although P. aeruginosa protein database predicts the presence of several other potential proteases, little has been known about the proteases involving in pathogenesis. Recently, LasB has been shown to disarm PAR-2 in respiratory epithelial cells (Dulon et al., 2005). In contrast, serralysin, which has been reported to reveal the similar features to AprA with respect to enzymatic properties, substrate specificity and primary structure (Maeda and Morihara, 1995), activates the critical transcription factors activator protein 1 (AP-1), CCAAT/enhancer-binding protein β (C/EBPβ), and NF-κB for host inflammatory and immune responses through PAR-2 (Kida et al., 2007). Accordingly, we hypothesized that like LasB and serralysin, unknown protease(s) from P. aeruginosa would also have an ability to activate or disarm PARs, thereby modulating host inflammatory and immune responses such as cytokine production. In this study, we found that P. aeruginosa produces a novel large exoprotease (LepA) distinct from known proteases such as LasA, LasB, AprA and protease IV. LepA as well as trypsin and also thrombin is shown to activate NF-κB-driven promoter through PAR-1, -2 or -4 and to cleave the peptides of PARs tethered ligand region. These results suggest that LepA would require PAR-1, -2 or -4 to activate the critical transcription factor NF-κB for host inflammatory and immune responses.

Results

Casein zymography of culture supernatant from P. aeruginosa clinical isolates

To elucidate whether P. aeruginosa clinical isolates produce unknown secreted protease(s) distinct from LasA, LasB, AprA and protease IV, we initially examined cell-free culture supernatant from P. aeruginosa clinical isolates by casein zymography. P. aeruginosa PAO1, which has been shown to produce LasB and AprA, was used as a positive control (Twining et al., 1993). Although LasA failed to be detected on the casein zymogram, LasA, LasB, AprA and protease IV are observed on the casein or gelatin zymogram as the bands of 160, 160, 50 and 200 kDa respectively (Caballero et al., 2001). As shown in Fig. 1, the protease with a molecular mass of ∼100 kDa on the casein zymogram was observed from P. aeruginosa clinical isolates KU2, α05-6 and α05-43 culture supernatants (lanes 2–4). The molecular mass of the protease was different from those of known proteases such as LasB, AprA and protease IV. In contrast, P. aeruginosa PAO1 showed the detectable bands of protease activity with the molecular mass of 160 and 50 kDa corresponding to LasB and AprA respectively (lane 1). Therefore, theses results indicate that P. aeruginosa would have an ability to produce unknown secreted protease(s) distinct from characterized proteases such as LasA, LasB, AprA and protease IV.

Figure 1.

Casein zymography of cell-free culture supernatant from P. aeruginosa clinical isolates. Cell-free culture supernatants (10 μl) of P. aeruginosa grown to stationary phase were mixed with equal volume of sample buffer and electrophoresed under non-reducing conditions using an 8% SDS-PAGE with 0.1% casein. The gel was washed in 2.5% Triton X-100, followed by incubation for 16 h in reaction buffer, and processed for visualization as described in Experimental procedures. Lane M, molecular weight markers; lane 1, strain PAO1; lane 2, strain KU2; lane 3, strain α05-6; lane 4, strain α05-43.

Identification of a novel secreted protease from P. aeruginosa as LepA

To identify the protease, which reveals a molecular mass of ∼100 kDa on the casein zymogram, we tried to purify the protease from P. aeruginosa KU2 strain as described in Experimental procedures. As shown in Fig. 2, the protease purified from strain KU2 appeared as a single band at 100 kDa (lane 2). We then determined the amino-terminal amino acid sequence of the purified protease. The protease had the amino-terminal amino acid sequence of HMVIDQXSHXXITNWXEFXXXADER (in which X represents undefined amino acid). The determined amino-terminal sequence was compared with the protein database of P. aeruginosa, indicating that the protease is 72% identical (expect value, 8e-10) to the sequence of the hypothetical protein PA4541 [P. aeruginosa PAO1: 1417 amino acids (aa), ∼140 kDa], PaerPA_01000896 (P. aeruginosa PACS2: 1408 aa, ∼139 kDa), PACG_04965 (P. aeruginosa C3719: 1357 aa, ∼134 kDa), PA2G_05515 (P. aeruginosa 2192: 1345 aa, ∼133 kDa) and PA140 (P. aeruginosa UCBPP-PA14: 1417 aa, ∼140 kDa). We have thus designated these proteins as LepA (large exoprotease from P. aeruginosa) and its coding gene as lepA.

Figure 2.

SDS-PAGE of LepA purified by anion-exchange and affinity column chromatography. The gel was stained with Coomassie blue G-250. Lane M, molecular weight markers; lane 1, fraction containing LepA purified by Q-Sepharose anion-exchange column; lane 2, purified LepA.

To confirm whether lepA encodes the 100 kDa protease, double-cross-over mutation was created in the lepA. Elaborately constructed was the allelic exchange vector that harboured the wild-type P. aeruginosa lepA gene disrupted by the insertion of a cassette encoding kanamycin resistance. Isogenic mutant of P. aeruginosa KU2 was constructed by introducing the allelic exchange vector into wild-type strain and selecting for subsequent double homologous recombination events between the lepA DNA flanking the antibiotic cassette in the vector and the wild-type lepA locus in the genome. The resulting mutant was assayed for the ability to produce protease by casein zymography. As shown in Fig. 3, wild-type KU2 strain exhibited a detectable band of protease activity with a molecular mass of ∼100 kDa on the casein zymogram (lane 1). In contrast, KU2 lepA mutant failed to produce a band of protease activity corresponding to that seen with wild-type KU2 (lane 2). Thus, these results suggest that lepA encodes for a novel secreted protease from P. aeruginosa.

Figure 3.

Casein zymography of cell-free culture supernatant of P. aeruginosa strains KU2 and KU2lepA-. Zymography was performed as described in Experimental procedures. Lane M, molecular weight markers; lane 1, strain KU2; lane 2, strain KU2lepA-.

Sequence analysis of LepA

We next performed sequence analysis of PA4541 as a model of LepA, because P. aeruginosa PAO1 is a well-characterized strain (Stover et al., 2000). The molecular features of LepA are shown as a schematic representation in Fig. 4A. SignalP3.0 (http://www.cbs.dtu.dk/services/SignalP) predicted that the first 53 residues of LepA constitute a signal peptide. An analysis of the first 200 amino acids of LepA identified the two-partner secretion (TPS) domain, which is a distinct feature of the TpsA proteins transported by the TPS pathway (Jacob-Dubuisson et al., 2004; Mazar and Cotter, 2007). The TPS domain of LepA contains a motif NPNGV (residues 137–141). This motif is similar to the conserved secretion motif NPNGI of TPS exoproteins (Jacob-Dubuisson et al., 2001). Further Entrez protein database searches (http://www.ncbi.nlm.nih.gov/) revealed the presence of a carbohydrate-dependent haemagglutination activity domain (Protein Families Databese codename, pfam05860) between residues 53 and 166. Haemagglutination activity domains are found in a number of proteins, including the adhesions FhaB (Bordetella pertussis) (Locht et al., 1993) and HMW1/HMW2 (Haemophilus influenzae) (St Geme and Grass, 1998; Dawid et al., 2001), the S. marcescens haemolysin ShlA (Schonherr et al., 1993), and the large secreted proteins LspA1 and LspA2 of Haemophilus ducreyi (Ward et al., 1998; 2003; 2004). These proteins belong to the TpsA family of molecules, which are exoproteins secreted in a TPS manner (Jacob-Dubuisson et al., 2001). In addition, PROSITE scan (http://www.expasy.org/prosite/) demonstrated that LepA contains a trypsin-like serine protease motif YQFSGDSGATVS (residues 171–182; S177 is the putative serine protease active site) and a cell attachment motif RGD (residues 680–682).

Figure 4.

Schematic representation of molecular features of the LepA from P. aeruginosa PAO1 and genetic organization of the lepBA two-partner secretion locus.
A. Molecular features of the LepA from P. aeruginosa PAO1.
B. Genetic organization of the lepBA operon. The Kanr cassette insertion in lepA deletes the region corresponding to the motifs of two-partner secretion and trypsin-like serine protease, resulting in the absence of protease activity observed in strain KU2lepA- (Fig. 3, lane 2).

A TPS system consists of two proteins, an outer membrane transporter/activator protein, TpsB protein, and an exoprotein, TpsA protein that is secreted by the transporter. Genes encoding the transporter and substrate proteins are usually found in an operon or within the same locus (Jacob-Dubuisson et al., 2001). DNA sequence analysis of the lepA flanking regions revealed the presence of a second open reading frame (ORF) beginning at 1686 nt upstream from, and in the same orientation as the lepA ATG start codon (Fig. 4B). This ORF, which is the hypothetical protein PA4540, was found to be 1638 nt in length and predicted to encode a protein of ∼60 kDa (545 aa). An analysis of the amino acid sequence of PA4540 using SignalP3.0 revealed that a putative signal sequence cleavage site is detected between residues 13 and 14. Further database searches demonstrated that PA4540 contains one polypeptide transport-associated domain (PORTA_2; pfam08479), localized between aa 59 and 134. The PORTA_2 domain is present in a number of porin-like proteins responsible for the transport of polypeptide across the outer membrane of Gram-negative bacteria and also is a molecular feature of the Omp85 family, of which the TpsB proteins are members (Gentle et al., 2005). Moreover, the analysis with PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/) revealed that PA4540 contains 18–22 β-strands, suggesting a porin conformation. Therefore, these observations suggest that PA4540 would be a TPS transporter/activator protein and also organized with LepA in an operon. We have thus designated PA4540 as lepB and its gene product as LepB.

PARs activation by LepA

Recently, some bacterial proteases have been shown to modulate host inflammatory and immune responses through PAR-1, -2 and -4 (Lourbakos et al., 2001a,b; Chung et al., 2004; Tancharoen et al., 2005; Uehara et al., 2005; Dommisch et al., 2007; Giacaman et al., 2007; Kida et al., 2007; Yun et al., 2007). In contrast, PAR-3 does not signal but rather can act as an accessory receptor for either PAR-1 or -4 (Nakanishi-Matsui et al., 2000; Sambrano et al., 2001; McLaughlin et al., 2007). Therefore, especially PAR-1, -2 and -4 are considered to play an important role for host inflammatory and immune responses. We then tested whether LepA activates NF-κB-driven promoter through PAR-1, -2 and -4. To address this question, COS-7 cells without functional expression of endogenous PARs were used for transfection experiments. As shown in Fig. 5A, co-transfection of COS-7 cells with increasing amounts of human PAR-1, -2 and -4 expression plasmids enhanced LepA-induced NF-κB-driven promoter activity in a dose-dependent manner. By contrast, LepA failed to activate NF-κB-driven promoter in COS-7 cells co-transfected with a mock plasmid. Transactivation of NF-κB-driven promoter was observed in response to stimulation with LepA in a dose-dependent manner with a plateau at 0.5–8 nM (Fig. 5B). On the other hand, thrombin, used as a positive control for activation of PAR-1 and -4, was able to augment NF-κB-driven promoter activity through PAR-1 or -4 (Fig. 5C). In addition, trypsin, used as a positive control for activation of PAR-2 and -4, could induce transactivation of NF-κB-driven promoter through PAR-1, -2 or -4 (Fig. 5D). Thus, these results indicate that like thrombin and trypsin, LepA can also activate NF-κB-driven promoter through human PAR-1, -2 or -4.

Figure 5.

LepA-induced transactivation of NF-κB-driven promoter in COS-7 cells transfected with human PAR-1, -2 or -4. COS-7 cells were transfected with 25–200 ng of the appropriate plasmid (phPAR-1, phPAR-2 or phPAR-4 as human PARs expression plasmid), 95 ng of pNF-κB-Luc as a reporter plasmid and 5 ng of phRG-TK as an internal control plasmid. The pcDNA3.1(+) was used to ensure that all transfection mixtures had a total of 300 ng of DNA. After 48 h, transfected cells were stimulated with or without the indicated concentrations of LepA, thrombin or trypsin. After a further 6 h of incubation, the cells were lysed and assayed for luciferase activity. The data are presented as the relative luciferase activity. The results for each set of transfections were normalized for Renilla luciferase activity and to the unstimulated samples. Values represent the mean ± standard deviation (SD) from triplicate determinations of a representative experiment. Similar results were obtained in three independent experiments. hPAR, human PAR.
A. LepA-induced NF-κB-driven promoter activity in response to co-transfection with increasing amounts of human PAR-1, -2 or -4 expression plasmid in COS-7 cells.
B. Transactivation of NF-κB-driven promoter in response to stimulation with increasing amounts of LepA in human PAR-1, -2 or -4-transfected COS-7 cells.
C and D. Like LepA, thrombin (C) and trypsin (D) activate NF-κB-driven promoter in human PARs-transfected COS-7 cells.
*P < 0.05; **P < 0.01 (compared with unstimulated cells).

Cleavage of PARs tethered ligand region by LepA

The above observation suggests that LepA is able to cleave human PAR-1, -2 and -4 at a specific site with exposure of their tethered ligands, as in the cases of thrombin and trypsin. To examine this possibility, the peptides corresponding to the region surrounding the tethered ligand of human PAR-1, -2 and -4 were incubated with LepA and proteolytic fragments were analysed. Thrombin, an agonist for PAR-1 and -4, and trypsin, an agonist for PAR-2 and -4, were used as positive controls. As shown in Fig. 6A, the PAR-1 peptide was cleaved at the R41–S42 site by LepA, thrombin or trypsin. The two peptide fragments were identified as NATLDPR and SFLLR, and the measured molecular masses (786.5 and 635.5 for LepA; 786.7 and 635.4 for thrombin; 786.5 and 635.7 for trypsin) were in good agreement with the calculated values of 785.4 and 634.4. The analysis also demonstrated that LepA as well as trypsin can cleave the PAR-2 peptide at the R36–S37 site, generating the two major peptide fragments of SSKGR and SLIGKVDGT (Fig. 6B). The measured molecular mass (534.3 and 889.5 for LepA; and 534.1 and 889.4 for trypsin) was compatible with the calculated values of 533.3 and 888.5. In addition, LepA cleaved the PAR-4 peptide at the R47–G48 site, as in the cases of thrombin and trypsin (Fig. 6C). The two major peptide fragments were identified as SILPAPR and GYPGQV, and the molecular masses (753.6 and 620.4 for LepA; 753.8 and 620.3 for thrombin; 753.6 and 620.6 for trypsin) were in good agreement with the calculated values of 752.4 and 619.3. Therefore, these results indicate that LepA has the protease activity required to activate human PAR-1, -2 and -4.

Figure 6.

Analysis of the LepA-cleaved human PARs tethered ligand region peptides. The peptide (50 μM) was incubated for 30 min at 37°C with 5 nM LepA, 25 nM thrombin or 10 nM trypsin and the products were separated by reversed-phase HPLC, then each peak was analysed by MALDI-TOF MS.
A. Treatment of hP1-35/46 with LepA. P1, uncleaved hP1-35/46; P2, SFLLR; P3; NATLDPR.
B. Treatment of hP2-32/45 with LepA. P1, uncleaved hP2-32/45; P2, SSKgr; P3; SLIGKVDGT.
C. Treatment of hP4-41/53 with LepA. P1, uncleaved hP4-41/53; P2, GYPGQV; P3; SILPAPR.
AUFS, absorbance units at full scale. The results shown are representative of those for three different experiments.

LepA induces IL-8 production in a human bronchiole epithelial cell line, EBC-1 cells

We previously reported that S. marcescens serralysin induces inflammatory responses through PAR-2 in a human bronchiole epithelial cell line, EBC-1 cells (Kida et al., 2007). We then examined whether LepA is able to induce IL-8 production in EBC-1 cells. As shown in Fig. 7A, LepA as well as thrombin and also trypsin induced IL-8 production in EBC-1 cells. Moreover, pre-treatment with PAR-2 antagonist peptides, FSLLRY-NH2 and LSIGRL-NH2, significantly reduced IL-8 production in response to stimulation with LepA (Fig. 7B). The peptides were shown not to inhibit the proteolytic activity of trypsin but to block trypsin-induced activation of PAR-2 by a mechanism that they possibly interact with a tethered ligand receptor-docking site (Al-Ani et al., 2002). We observed that neither FSLLRY-NH2 nor LSIGRL-NH2 inhibits the proteolytic action of LepA in the protease activity assay (data not shown). Thus, these results suggest that LepA has an ability to induce IL-8 production through PAR-2 in EBC-1 cells.

Figure 7.

LepA induces IL-8 secretion in EBC-1 cells. EBC-1 cells were stimulated with or without LepA (1 nM), thrombin (10 nM) or trypsin (50 nM) for 24 h and then IL-8 levels in the culture supernatants were determined. To antagonize PAR-2, human PAR-2 antagonist peptides were added directly to the culture medium at a final concentration of 200 μM at 1 h prior to stimulation.
A. LepA as well as thrombin and also trypsin induces IL-8 production.
B. Human PAR-2 antagonist peptides reduce IL-8 production in response to stimulation with LepA.
Values represent the mean ± standard deviation (SD) from triplicate determinations of a representative experiment. Similar results were obtained in three independent experiments. *P < 0.05; **P < 0.01 (compared with unstimulated cells). #P < 0.05; ##P < 0.01 (compared with stimulated cells in the absence of the antagonists).

Discussion

In this article, we reported that P. aeruginosa produces a novel large exoprotease (LepA) distinct from well-characterized proteases such as LasA, LasB, AprA and protease IV. Sequence analysis of LepA showed that LepA contains the TPS domain, which is a distinct feature of the TpsA proteins transported by the TPS pathway. Our results indicated that LepA induces transactivation of NF-κB-driven promoter through human PAR-1, -2 or -4, and that LepA cleaves human PAR-1, -2 and -4 at a specific site with exposure of their tethered ligands, as in the cases of thrombin and trypsin.

We isolated the protease with a molecular mass of ∼100 kDa from P. aeruginosa KU2 strain (Fig. 2). Double-cross-over mutation of lepA revealed lack of an ability to produce the protease (Fig. 3). These results indicate that lepA encodes for the 100 kDa protease. However, the sequence analysis of LepA demonstrated that the predicted molecular mass of pre-protein and mature protein are ∼140 kDa and ∼134 kDa respectively. In addition, SignalP3.0 revealed that the predicted amino-terminal amino acid is 54L. In fact, however, our experimental data indicated that the molecular mass of the mature protein and the amino-terminal amino acid are ∼100 kDa and 75H respectively. Therefore, it is conceivable that these differences may be caused by autoproteolysis of LepA. In contrast, LepA could not be detected in P. aeruginosa PAO1 (Fig. 1). It is possible that the sequence diversity in the regulatory region of lepBA operon might lead to the failure of LepA expression, because the organization of the P. aeruginosa genome is very diverse, with signs of insertions, deletions and other genome rearrangements (Romling et al., 1997).

Interestingly, P. aeruginosa keratitis isolates have been reported to reveal two general production patterns of protease on casein or gelatin zymograms (Lomholt et al., 2001; Zhu et al., 2006). One type (type I) was characterized by a combination of major activity bands at > 200 (protease IV), 145–163 or 116 (LasB) and 50 (AprA) kDa. Another type (type II) was characterized by a combination of activity bands at > 200 (protease IV), 98 or 66 [modified LasB or P. aeruginosa small protease (PASP)] and 50 (AprA) kDa. All ocular isolates harbouring only exoS gene encoding a ADP-ribosylating enzyme showed type I protease profile, whereas the strains carrying only exoU gene encoding a cytolytic factor exhibited type II protease profile. The type III secretion system directs secretion and translocation of several exotoxins such as ExoS and ExoU into a target cell. ExoS plays a role in invasion into non-phagocytic cells and modulation of bacterial phagocytosis by phagocytes, and induces apoptosis of a variety of cell types, including epithelial cells, fibroblasts and lymphocytes (Fleiszig et al., 1997; Frank, 1997; Frithz-Lindsten et al., 1997). ExoU mediates rapid lysis (cytotoxicity) of a variety of cell types, including macrophages, epithelial cells and fibroblasts (Finck-Barbancon et al., 1997; Hauser et al., 1998; Coburn and Frank, 1999). Indeed, the strains that harbour exoS are reported to reveal an invasive phenotype, and those that harbour exoU are referred to as having a cytotoxic phenotype (Roy-Burman et al., 2001; Hauser et al., 2002). Thus, it is conceivable that the association between the production of type I or type II protease and the possession of an exotoxin gene encoding ExoS or ExoU would participate in the pathogenesis of P. aeruginosa infection. However, the question of which strains harbouring exoS or exoU genes can produce a novel protease LepA remains to be elucidated.

The exoproteins secreted by the TPS pathway are synthesized as pre-proteins or pre-proproteins. All are large proteins that range in size from 100 kDa to more than 500 kDa, and many of them are associated with virulence (Jacob-Dubuisson et al., 2001; 2004; Mazar and Cotter, 2007). Although a large number of genes encoding potential TPS systems have been identified through DNA sequencing of microbial genomes, only a limited number of TPS molecules have been characterized so far. Of the few TPS systems that have been characterized, most have been shown to play a role in adhesion or haemolysis/cytolysis, the others appear to be involved in iron acquisition (Jacob-Dubuisson et al., 2001; 2004; Mazar and Cotter, 2007). In this study, we reported that LepA contains the TPS domain, which is a distinct feature of the TpsA proteins transported by the TPS pathway. Moreover, LepA contains a trypsin-like serine protease motif and a cell attachment motif (Fig. 4A). Indeed, we demonstrated that LepA serves as protease and activates NF-κB-driven promoter through human PAR-1, -2 or -4 (Figs 3, 5 and 6). Nevertheless, the attachment function of the LepA remains to be clarified.

Although the sequence analysis of P. aeruginosa genome identifies five genes encoding potential TPS systems in the genome (Yen et al., 2002; Ma et al., 2003), to our knowledge, there have been no reports describing the biological functions of them. In the present study, we first described a biological function of LepA, which is one of them. The other potential TpsA–TpsB proteins of P. aeruginosa PAO1 are PA0041 (3535 aa, ∼362 kDa)–PA0040 (562 aa, ∼63 kDa), PA0690 (4180 aa, ∼430 kDa)–PA0692 (544 aa, ∼60 kDa), PA2462 (5627 aa, ∼573 kDa)–PA2463 (565 aa, ∼63 kDa) and PA4625 (2154 aa, ∼220 kDa)–PA4624 (568 aa, ∼63 kDa). The sequence analysis shows that all of these potential TpsA–TpsB proteins have the hallmarks of the TPS systems. Furthermore, like PA4541, these TpsA proteins contain some RGD–cell attachment motifs. Therefore, it is possible that the TpsA proteins of P. aeruginosa might play a role in adhesion like other bacteria-derived TpsA proteins, because among the few TpsA proteins that have been characterized extensively, most have been shown to function as adhesin or haemolysin/cytolysin (Jacob-Dubuisson et al., 2001; Jacob-Dubuisson et al., 2004; 2007). Further analysis needs to clarify whether these molecules participate in pathogenesis of P. aeruginosa infection.

Recently, P. aeruginosa LasB (elastase B) has been shown to disarm PAR-2 in respiratory epithelial cells (Dulon et al., 2005). Intriguingly, a neutrophil elastase, a member of neutrophil-derived serine protease, has been shown not to disarm but to activate PAR-2 expressed by non-epithelial cells (Uehara et al., 2003). A possible explanation for the opposite functions is that the susceptibility of PAR-2 to cleavage and/or activation by protease is dependent on the glycosylation pattern of its amino-terminal exodomain. The pattern of glycosylation of PAR-2 is different by cell type (Compton et al., 2001; 2002). On the other hand, PAR-1, -3 and -4 have been shown to contain the putative glycosylation sites in their amino-terminal exodomains or extracellular loop 2 or 3 (Macfarlane et al., 2001; Hollenberg and Compton, 2002; Ossovskaya and Bunnett, 2004). To date, although there have been no reports describing the relevance of glycosylation in the activation of PAR-1, -3 and -4, it is possible that the pattern of glycosylation of PAR-1, -3 and -4 might be different by cell type, leading to the change of the susceptibility of the receptors to activation or disarming by various proteases. However, whether LepA-induced PARs activation is affected by cell type remains to be elucidated.

In summary, we found P. aeruginosa produces a novel large exoprotease, LepA, and investigated whether LepA activates NF-κB-driven promoter through PAR-1, -2 or -4. Our results indicated that LepA as well as thrombin and also trypsin has an ability to induce transactivation of NF-κB-driven promoter through human PAR-1, -2 or -4. Furthermore, we demonstrated that LepA has the protease activity required to activate human PAR-1, -2 and -4. Thus, these results suggest that bacterial proteases such as LepA would require PARs to modulate various host responses against bacterial infection. Further characterization of LepA with respect to the proteolytic mechanism, secretion mechanism via TPS, and its role in virulence is now in progress in our laboratory.

Experimental procedures

Bacterial strains and plasmids

The bacterial strains and plasmids used in this study are listed in Table 1. All bacterial strains were grown in Luria–Bertani (LB) medium (LB-Miller; Nacalai tesque, Kyoto, Japan) unless otherwise noted. The growth medium was supplemented with antibiotics at the following concentrations: ampicillin, 100 μg ml−1 (Escherichia coli); carbenicillin, 500 μg ml−1 (P. aeruginosa); and kanamycin, 1 mg ml−1 (P. aeruginosa).

Table 1.  Bacterial strains and plasmids.
Strain or plasmidDescriptionSource or reference
Strains
P. aeruginosa
 PAO1LasB- and AprA-producing strainTwining et al. (1993)
 KU2Clinical isolateThis study
 α05-6Clinical isolateThis study
 α05-43Clinical isolateThis study
 KU2lepA-Isogenic lepA mutant, KanrThis study
E. coli
 DH5αCloning strainToyobo
 S17-1λpirMobilizer strainBiomedal
Plasmids
 pNF-κB-LucReporter vector, containing five repeats of the binding site (GGGGACTTTCC) for the NF-κBStratagene
 pcDNA3.1(+)Mammalian expression vectorInvitrogen
 phPAR-1pcDNA3.1(+) with a human PAR-1 cDNAThis study
 phPAR-2pcDNA3.1(+) with a human PAR-2 cDNAThis study
 phPAR-4pcDNA3.1(+) with a human PAR-4 cDNAThis study
 pDNR-1rsacBAmpr; Source of sacBClontech
 pUC18NotAmpr; pUC18 with two NotI sitesBiomedal
 pUTmini-Tn5 KmAmprKanroriT; sources of Kanr cassette and oriTBiomedal
 pYK1sacBAmpr; a 1.0 kb NotI deletion of pDNR-1rThis study
 pYK1-TSuicide vector, sacBoriTAmpr; pYK1 with a 1.7 kb MluI–BglII fragment containing oriTThis study
 pYK2lepAAmpr; pUC18Not with a 5.9 kb PCR fragment containing lepAThis study
 pYK2-KmpYK2 with 1.4 kb EcoRV deletion in lepA and insertion of Kanr; ΔlepA::KanrThis study
 pYK3pYK1-T with a 5.3 kb NotI fragment containing ΔlepA::Kanr of pYK2-KmThis study

Reagents

Bovine pancreas-derived endotoxin-free trypsin (endotoxin contents, < 0.2 ng per milligram of protein) and human plasma-derived thrombin (endotoxin contents, < 0.1 ng per milligram of protein) were purchased from Calbiochem (San Diego, CA). The peptides of human PARs tethered ligand region were synthesized by solid-phase methods at Invitrogen (Carlsbad, CA): hP1-35/46 (corresponding to human PAR-1 residues 35–46, N35ATLDPRSFLLR46), hP2-32/45 (corresponding to human PAR-2 residues 32–45, S32SKGRSLIGKVDGT45) and hP4-41/53 (corresponding to human PAR-4 residues 41–53, S41ILPAPRGYPGQV53) (tethered ligand sequences are underlined). The peptides of human PAR-2 antagonist (FSLLRY-NH2 and LSIGRL-NH2) were synthesized with amidated carboxyl-terminus by solid-phase methods at Sigma-Aldrich (St Louis, MO). All peptides were > 95% purity by HPLC and mass spectrometry analysis. Stock solutions were prepared in 50% DMSO and stored at −20°C.

Cell culture

A human lung squamous cell carcinoma, EBC-l cells (Cell Resource Center for Biomedical Research Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan), was maintained in RPMI 1640 medium (Nissui Pharmaceutical, Tokyo, Japan) supplemented 10% heat-inactivated fetal bovine serum (FBS) (Gibco, Rockville, MD), 2 mM l-glutamine, 0.2% sodium bicarbonate, 100 U ml−1 penicillin and 100 μg ml−1 streptomycin (Kida et al., 2006). An African green monkey kidney fibroblast-like cell line, COS-7 cells (Health Science Research Resources Bank, Osaka, Japan), which has no functional expression of endogenous PARs, was maintained in Dulbecco's modified Eagle's medium (DMEM; Nissui Pharmaceutical) supplemented 10% heat-inactivated FBS, 4 mM l-glutamine, 0.35% d-glucose, 0.37% sodium bicarbonate, 100 U ml−1 penicillin and 100 μg ml−1 streptomycin (Molino et al., 1997). Cells were passaged without the use of trypsin by non-enzymatic cell dissociation solution (Gibco) to minimize the proteolytic activation of the PARs.

Zymography

Pseudomonas aeruginosa were grown to stationary phase in LB broth containing 1% d-glucose at 35°C. Supernatants were harvested by centrifugation at 13 000 g for 10 min followed by filtration through a 0.45 μm filter. Cell-free culture supernatants (10 μl) were mixed with equal volume of sample buffer [125 mM Tris-HCl (pH 6.8), 20% glycerol, 4% sodium dodecyl sulfate (SDS), 0.01% bromophenol blue] and electrophoresed under non-reducing conditions using an 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with 0.1% casein. To remove SDS, the gels were then soaked three times in 2.5% Triton X-100 for 20 min and incubated for 16 h at 37°C in reaction buffer [50 mM Tris-HCl (pH 8.0), 200 mM NaCl, 5 mM CaCl2, 1 μM ZnCl2, 0.02% Brij 35). The gels were stained with 0.1% Coomassie blue G-250 in acetic acid : methanol : water (10:20:70) for 1 h and destained in the same solution without dye. Protease-containing bands were visualized as clear zones against a dark background.

Purification of LepA

Pseudomonas aeruginosa KU2, a clinically isolated strain, was grown in LB broth containing 1% d-glucose for 20 h at 35°C with rotary shaking at 100 r.p.m. (AT-12R shaker; Thomas, Tokyo, Japan). Bacterial cells were removed from the medium by centrifugation (8000 g, 30 min, 4°C) and LepA was purified from the supernatant. Briefly, proteins in the supernatant were precipitated with ammonium sulfate (80% saturation). The precipitate was dissolved in 20 mM Tris-HCl (pH 8.0), dialysed, and loaded onto a Q-Sepharose anion-exchange column (Amersham-Pharmacia Biotech, Uppsala, Sweden) equilibrated with the same buffer. The column was washed with 20 mM Tris-HCl (pH 8.0) containing 50 mM NaCl until unbound material was removed. LepA were eluted with 20 mM Tris-HCl (pH 8.0) containing 100–200 mM NaCl. Then, the pooled fractions containing LepA were dialysed against 20 mM Tris-HCl (pH 8.0) containing 1 mM CaCl2 and applied to a benzamidine-Sepharose affinity column (Amersham-Pharmacia Biotech) equilibrated with the same buffer. The column was washed with 20 mM Tris-HCl (pH 8.0) containing 1 mM CaCl2 until unbound material was removed. LepA was eluted with 100 mM benzamidine in 20 mM Tris-HCl (pH 8.0) and 1 mM CaCl2. Fractions containing LepA were pooled and dialysed against 20 mM Tris-HCl (pH 8.0) containing 1 mM CaCl2. The purity of LepA was determined by SDS-PAGE. The endotoxin level of purified LepA was determined with a Limulus amebocyte lysate QCL-1000 (Cambrex, Walkersville, MD) and was revealed to be < 0.5 pg ml−1 when suspended in 20 mM Tris-HCl (pH 8.0) containing 1 mM CaCl2 at a protein concentration of 1 nM. Protein concentration was determined with a Coomassie Protein Assay Reagent (Pierce, Rockford, IL) using bovine serum albumin as a standard. The amino-terminal amino acid sequence of LepA was determined using an automated protein sequencer (PSQ-1; Shimadzu, Kyoto, Japan) at Hipep Laboratories (Kyoto, Japan).

Replacement of the lepA region of KU2 strain

Allele replacement of the lepA was performed by a modified method of Schweizer (1992). In brief, a 5.9 kb PCR fragment containing lepA was amplified from P. aeruginosa PAO1 genomic DNA by using primers delPA100-F (5′-GGAAACAAGCTTATGGGAATGGCCATGTTGTCGTCCTGCCTTGCC-3′, underline indicates HindIII restriction site) and delPA100-R (5′-GGAAACGGATCCTCACAGGTTGTAGTGGCGACCTATCATTGGAAC-3′, underline indicates BamHI restriction site). After digestion with HindIII and BamHI, the resulting fragments were cloned into pUC18Not, producing plasmid pYK2. A 1.5 kb PCR fragment containing a Kanr cassette was amplified from pUTmini-Tn5 Km by using primers Kan-F (5′-AGCTTCACGCTGCCGCAAGCACTCAGGGCGC-3′) and Kan-R (5′-GTCGACCAAAGCGGCCATCGTGCCTCCCCAC-3′). After phosphorylated with T4 polynucleotide kinase (Toyobo), the resulting fragments were cloned into pYK2, which was digested with EcoRV and dephosphorylated with calf intestine alkaline phosphatase (Toyobo) to yield plasmid pYK2-Km. The ΔlepA::kanr fragment was then subcloned into the NotI site of the suicide vector pYK1-T, which has the oriT for conjugative transfer and the counter-selectable marker sacB, producing plasmid pYK3. This plasmid was used for allelic exchange and conjugated from E. coli S17-1λpir into P. aeruginosa KU2 on LB agar using filters.

Merodiploid single-cross-over mutants were selected from the conjugation mixture by plating on LB agar containing 1 mg ml−1 kanamycin. Purified single-cross-over mutants were cultured overnight in LB broth without antibiotics. This culture was then serially diluted in PBS, and plated on LB agar containing 1 mg ml−1 kanamycin and 7% sucrose to select against the sacB marker present on the pYK3 vector, and hence select for strains which had undergone a second homologous recombination event resulting in loss of the pYK3 vector. This was confirmed by the loss of the vector-encoded carbenicillin resistance. In addition, the double-cross-over mutants were confirmed by PCR using the primers delPA100-F and delPA100-R and decreased zones on LB-skim milk agar (data not shown).

Construction of human PARs expression plasmids

EBC-1 cells maintained as described above were detached from culture flask with non-enzymatic cell dissociation solution and washed twice in phosphate-buffered saline (PBS). Total RNA was then purified using an RNeasy mini kit (Qiagen, Chatsworth, CA) and treated with an RNase-free DNase set (Qiagen) to remove contaminated DNA according to the instructions provided by the manufacturer. Synthesis of cDNA was performed with an RNA PCR kit (Takara, Ohtsu, Japan), according to the manufacturer's protocol. PCR was performed with a KOD Plus DNA polymerase (Toyobo, Osaka, Japan) according to the protocol recommended by the manufacturer. The following human PAR-1-specific primers were used: sense, 5′-GGAAACAAGCTTCCACCATGGGGCCGCGGCGGCTGCTGCTG-3′ (underline indicates HindIII restriction site); and antisense, 5′-GGAAACGGATCCCTAAGTTAACAGCTTTTTGTATATGC-3′ (underline indicates BamHI restriction site). The following human PAR-2-specific primers were used: sense, 5′-GGAAACAAGCTTCCACCATGCGGAGCCCCAGCGCGGCGTG-3′ (underline indicates HindIII restriction site); and antisense, 5′-GGAAACGGATCCTCAATAGGAGGTCTTAACAGTGG-3′ (underline indicates BamHI restriction site). The following human PAR-4-specific primers were used: sense, 5′-GGAAACAAGCTTCCACCATGTGGGGGCGACTGCTCCTGTG-3′ (underline indicates HindIII restriction site); and antisence, 5′-GGAAAC GAATTCTCACTGGAGCAAAGAGGAGTGGG-3′ (underline indicates EcoRI restriction site). The PCR profile included denaturation at 96°C for 3 min, followed by 35 cycles of denaturation at 96°C for 30 s, annealing at 60°C for 30 s, extension at 68°C for 90 s and a final extension at 68°C for 5 min. The PCR products were digested with HindIII and BamHI (for human PAR-1 and -2) or EcoRI (for human PAR-4) and cloned into a pcDNA3.1(+) (Invitrogen). The resultant plasmids were named phPAR-1, phPAR-2 and phPAR-4. The plasmids were purified with the Qiagen plasmid kit (Qiagen) and used for transient transfection.

Transient transfection of COS-7 cells and luciferase assay

COS-7 cells maintained as described above were detached from culture flask with non-enzymatic cell dissociation solution and washed three times in serum-free DMEM/F-12 (Gibco). Then, the cells were seeded into 24-well plates (Costar, Cambridge, MA) at a density of 5 × 104 cells per well in serum-free DMEM/F-12 and incubated for 24 h. One hour before transfection, the growth medium was replaced by fresh serum-free DMEM/F-12. Transient transfections were performed with 25–200 ng of the appropriate plasmids (phPAR-1, phPAR-2 and phPAR-4 as human PARs expression plasmids), 95 ng of pNF-κB-Luc as a reporter plasmid and 5 ng of phRG-TK (Promega, Madison, WI) as an internal control plasmid using the FuGENE6 transfection reagent (Roche, Basel, Switzerland), according to the manufacturer's protocol. After 48 h, transfected cells were stimulated with or without the indicated concentrations of LepA, thrombin or trypsin. After a further 6 h of incubation, cells were lysed and assayed for luciferase activity using a Dual-Luciferase Reporter Assay System (Promega). Both firefly and Renilla luciferase activities were monitored with a Lumat LB9507 luminometer (Berthold, Wildbad, Germany). Normalized reporter activity was expressed as the firefly luciferase value divided by the Renilla luciferase value. Relative fold induction was calculated as the normalized reporter activity of the stimulated samples divided by the unstimulated samples.

Cleavage of human PARs tethered ligand region by LepA

The peptides of human PARs tethered ligand region (50 μM): hP1-35/46, hP2-32/45 and hP4-41/53 were incubated for 30 min at 37°C with 5 nM LepA in 20 mM Tris-HCl (pH 8.0) containing 1 mM CaCl2, then the digested materials were separated by reversed-phase HPLC on a Cosmosil 5C18-AR-300 column (4.6 × 250 mm; Nacalai tesque) using a linear gradient of acetonitrile (5–45% for hP1-35/46, 0–40% for hP2-32/45 and 5–45% for hP4-41/53) in 0.1% TFA for 40 min at a flow rate of 1 ml min−1 and each peak was analysed by a matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS) (Autoflex; Bruker Daltnics, Bremen, Germany). As controls, the peptides were incubated under the same experimental conditions with 25 nM thrombin, which cleaves human PAR-1 and -4 tethered ligand region, or with 10 nM trypsin, which cleaves human PAR-1, -2 and -4 tethered ligand region.

Measurement of IL-8 secretion

EBC-1 cells prepared as described above were washed three times in serum-free DMEM/F-12 (Gibco). The cells were then seeded into 24-well plates (Costar) at a density of 2 × 105 cells per well in serum-free DMEM/F-12 and incubated for 24 h. One hour before stimulation, the growth medium was replaced by fresh serum-free DMEM/F-12. Cells were stimulated with or without LepA (1 nM), thrombin (10 nM) or trypsin (50 nM) for 24 h and the culture supernatants were collected and stored at −80°C until assayed. To antagonize PAR-2, human PAR-2 antagonist peptides were added directly to the culture medium at a final concentration of 200 μM at 1 h prior to stimulation. The levels of IL-8 in the culture supernatants were measured using enzyme-linked immunosorbent assay (ELISA) kit (R&D systems, Minneapolis, MN) according to the manufacturer's protocol.

Statistical evaluation

Data were analysed by using the Student's paired t-test. Data with a P-value of < 0.05 were considered significant.

Acknowledgements

We thank Professor Hiroshi Watanabe, Kurume University School of Medicine for kind donation of P. aeruginosa clinical isolates.

Ancillary