SGM Journals
PATHOGENICITY AND VIRULENCE

Aerolysin is activated by metalloprotease in Aeromonas veronii biovar sobria

Tianyan Song, Claudia Toma, Noboru Nakasone, Masaaki Iwanaga

  • Division of Bacterial Pathogenesis, Department of Microbiology, Graduate School of Medicine, University of the Ryukyus, 207 Uehara, Nishihara, Okinawa 903-0215, Japan
  • Correspondence Tianyan Song k018763med.u-ryukyu.ac.jp

    • Journal of Medical Microbiology 2004; 53(6):477–482 · https://doi.org/10.1099/jmm.0.05405-0

    View at publisher PubMed

    Abstract

    Aeromonas veronii is an opportunistic human pathogen that causes diarrhoea and extraintestinal infections, such as wound infection and septicaemia. An A. veronii protease (AVP) from a biovar sobria strain, AeG1, was partially purified and characterized. Mature AVP hydrolysed casein but not elastin, and protease activity was inhibited by metalloprotease inhibitors. Nucleotide sequence analysis showed that AVP belongs to the thermolysin family of proteases. An AVP-deficient mutant was constructed to investigate the role of AVP in aerolysin activation. Western blot analysis using anti-aerolysin antisera revealed that proaerolysin (52 kDa) in the AVP-deficient mutant was not completely activated to mature aerolysin (47 kDa) as seen in the wild-type strain. The AVP-deficient mutant showed lower cytotoxic and haemolytic activities than wild type. AVP and proaerolysin had no haemolytic activity; however, activity appeared after incubating both proteins. Taken together, these results suggested that AVP plays an indirect role in virulence through activating aerolysin, which is an essential step for cytotoxic activity.

    • Abbreviations: AerA, aerolysin; AVP, Aeromonas veronii protease.

    • The GenBank/EMBL/DDBJ accession numbers for the avp and aerA sequences of A. veronii bv. sobria strain AeG1 are AB103464 and AB109093, respectively.

    INTRODUCTION

    Aeromonas species are Gram-negative, facultative anaerobic bacteria that can be isolated from many sources, such as food, drinking water, sewage, environmental water and human clinical specimens (Janda & Abbott, 1998). Substantial evidence points to aeromonads as causative agents of sporadic diarrhoea, dysentery and extra-intestinal infections that may be life-threatening (Janda & Abbott, 1998; Wang et al., 2000). The three major pathogenic aeromonads are Aeromonas hydrophila [HG 1, hybridization group 1 (Borrell et al., 1997)], Aeromonas caviae (HG 4) and Aeromonas veronii biovar (bv.) sobria (HG 8), which account for more than 85 % of all clinical isolates (Janda & Abbott, 1998). Despite the increasing number of reports that implicate A. veronii bv. sobria as a cause of human diseases (Janda & Abbott, 1998; Wang et al., 2000; Vila et al., 2003), few studies have been carried out on A. veronii, and its virulence mechanisms are poorly understood.

    Proteases are secreted by various human pathogenic bacteria (Häse & Finkelstein, 1993). It has been suggested that proteolytic enzymes secreted by Aeromonas spp. play an important role in invasiveness and establishment of infection by overcoming initial host defences and providing nutrients for cell proliferation (Leung & Stevenson, 1988). Three proteases have been reported for A. hydrophila: (i) a thermostable 38-kDa metalloprotease (Rivero et al., 1990), (ii) a 19-kDa zinc protease (Loewy et al., 1993) and (iii) a 68-kDa temperature-labile serine protease (Rivero et al., 1991). The 38-kDa metalloprotease (AhyB) has elastolytic activity, and an isogenic mutant showed decreased virulence in rainbow trout (Cascón et al., 2000). A. caviae produces at least two proteases, a 34-kDa metalloprotease (AP34) and a 19-kDa protease (Toma et al., 1999). Western blot analyses showed that AP34 related metalloproteases are widely distributed in Aeromonas species (Toma et al., 1999); however, the role of this metalloprotease in pathogenicity has not yet been studied, and purification of an A. veronii protease has not been described so far.

    Aerolysin is a well-known pore-forming toxin that was first purified from A. hydrophila and later from A. sobria (Fujii et al., 1998). The toxin is secreted to the culture supernatant as a precursor (named proaerolysin), which has no biological activity. A C-terminal peptide must be removed to convert proaerolysin into active aerolysin (Nomura et al., 1999). Evidence has shown that proaerolysin can be activated by protease(s) produced by host cells (Abrami et al., 1998) or by Aeromonas itself (Howard & Buckley, 1985). Activation of proaerolysin by mammalian proteases has been studied by Abrami et al. (1998) and furin was demonstrated to be the major convertase involved in proaerolysin processing. No studies have addressed activation of aerolysin by bacterial proteases.

    The objective of the present study was to characterize the major extracellular A. veronii bv. sobria protease (AVP) and to determine whether AVP is involved in activation of proaerolysin.

    METHODS

    Bacterial strains, plasmids and media.

    Bacterial strains and plasmids used in this study are listed in Table 1. The AeG1 strain was identified as A. veronii bv. sobria (HG 8) by biochemical tests and RFLP of the PCR-amplified 16S rRNA gene (Borrell et al., 1997). Organisms were cultured in Luria–Bertani (LB) medium supplemented, as necessary, with 100 μg ampicillin ml−1, 25 μg kanamycin ml−1 and 5 μg chloramphenicol ml−1. YEP broth (5 g NaCl, 10 g peptone and 5 g yeast extract per litre distilled water) was used for cultivation of Aeromonas strains.

    Table 1. Strains and plasmids used in this study

    DNA manipulation.

    Chromosomal DNA was isolated by using Qiagen Genomic tips (Qiagen). Plasmid DNA was isolated according to Birnboim & Doly (1979), or by using Qiagen resin columns. Restriction enzyme digestion, ligation, gel electrophoresis and transformation of DNA were carried out as described by Bagdasarian & Bagdasarian (1994).

    AVP purification and N-terminal amino acid sequence.

    A. veronii bv. sobria strain AeG1 was precultured in brain heart infusion broth at 37 °C for 4 h. Precultured bacteria were then inoculated into 500 ml YEP broth, in a 3-l Erlenmeyer flask, and incubated for 24 h with shaking at 30 °C. The cell-free culture supernatant was salted out with 60 % saturated ammonium sulfate. After centrifugation, the pellet was resuspended in 50 mM Tris/HCl buffer (pH 8.0) and dialysed against the same buffer. Insoluble material was removed by centrifugation at 16 000 g for 20 min. Dialysed materials were applied to a DEAE-Sephadex A25 column equilibrated with 50 mM Tris/HCl buffer (pH 8.0). The column was washed with 500 ml 50 mM Tris/HCl buffer (pH 8.0) and bound material was eluted with 500 ml of a linear gradient of 0−0.5 M NaCl in the same buffer. Fractions with protease activity were pooled and regarded as partially purified AVP protease. The N-terminal amino acid sequence was determined by automated Edman degradation on an Applied Biosystems 477A protein sequencer.

    Protease and elastase assays.

    Proteolytic and elastolytic activities were detected by using a single-diffusion technique in agar containing 1.5 % (w/v) skimmed milk or 1 % (w/v) insoluble elastin as substrate. Sample solution (20 μl) was added to 3-mm diameter wells and incubated overnight. For quantitative analysis of protease activity, a test tube method using azocasein as a substrate was performed (Honda et al., 1989). To test for inhibitory effects, the following compounds were added to purified protease or culture filtrates in the azocasein assay: 10 mM EDTA, 1 or 4 mM PMSF, 1 mM 1,10-phenanthroline (OPA), 10 mM N-α-p-tosyl-l-arginine methyl ester (TAME) and 10 mM l-leucine methyl ester (LEME). After incubation of protease with these compounds for 30 min at 37 °C, residual activity was measured. To examine protease heat stability, protease was incubated at 60, 70 or 80 °C for 15 min and hydrolytic activity was compared with that of the control.

    Sequencing of avp.

    Primers used in this study are shown in Table 2. On the basis of the N-terminal amino acid sequence of AVP, degenerate primer APS1 was designed. Primer APS8 was designed on the basis of the A. caviae metalloprotease sequence previously reported by Kawakami et al. (2000). PCR was carried out with primers APS1 and APS8, using chromosomal DNA of the AeG1 strain as a template. The nucleotide sequence of the resulting product (about 950 bp) was determined. Upstream and downstream regions flanking the 950-bp core region were sequenced using the inverse PCR technique (Ochman et al., 1988). Chromosomal DNA was digested with BamHI or SphI restriction enzymes. Inverse PCR for the upstream region was performed by using a circularized 900-bp BamHI fragment as a template. Primers, APS18 toward the 5′ end and APS15 toward the 3′ end, were synthesized based on the 950-bp core sequence. The amplified product (about 450 bp) was sequenced. The downstream region was sequenced by the same method using a circularized 1.8-kb SphI fragment as a template. Primers used were APS22 toward the 5′ end and APS17 toward the 3′ end. Nucleotide sequencing was performed on both strands with BigDye Terminator Cycle Sequencing FS ready reaction kits (Applied Biosystems) and analysed with an ABI PRISM 310 Genetic Analyzer (Perkin-Elmer). The program blast 1.4.9 was used to search for related sequences in the database.

    Table 2. Oligonucleotide primers used in PCR or protein expression Underlined letters indicate restriction endonuclease sites. F, Forward primer; R, reverse primer. Primers were designed in this study.

    Construction of the AVP-deficient mutant.

    The avp gene in A. veronii AeG1 was inactivated by insertional mutagenesis. A DNA fragment (625 bp) encoding an internal region of AVP was amplified from the AeG1 genome by PCR using primers APS1 and 1304. The PCR product was initially cloned into the pCR2.1 vector and then subcloned into the SmaI site of the suicide vector pKY719 (Whayeb et al., 1996). The resulting plasmid, pM8, was transformed into Escherichia coli SM10λpir (Simon et al., 1983). Bacterial conjugation was then accomplished by mating the resulting E. coli SM10λpir bearing pM8 as the donor and AeG1 as the recipient. avp transconjugants were selected on LB agar containing 100 μg ampicillin and 5 μg chloramphenicol ml−1. The AVP-deficient mutant was identified by PCR using primer PUCR, which is specific for the multiple cloning site of pKY719, and primer APS8, which is specific for the avp sequence downstream of primer 1304. From the DNA sequence, the size of the PCR product was predicted to be 986 bp.

    Cloning, sequencing and expression of recombinant proaerolysin.

    A DNA fragment (1392 bp) containing the aerA gene was amplified from the AeG1 genome by PCR using primers Aehly-1 and Aehly-2 (Table 2), which were designed based on the published aerA nucleotide sequence of A. sobria strain 33 (GenBank accession no. X65046). The purified PCR product was ligated into the cloning vector pCR2.1 and transformed into E. coli XL-1 Blue. The sequence of the aerA gene was determined by primer walking. The aerA fragment was then digested with restriction enzymes BamHI and HindIII and subcloned into the expression vector pQE30. The resulting plasmid pAerA was then transformed into E. coli M15. The transformant was cultured in LB broth supplemented with 100 μg ampicillin and 25 μg kanamycin ml−1. Histidine-tagged proaerolysin was induced with 1 mM IPTG and purified on a His Trap nickel column (Amersham) as described previously (Miyazato et al., 2003).

    Antiserum preparation.

    The antiserum to aerolysin was raised in a rabbit (Japanese White) by intramuscular injection of 50 μg purified recombinant proaerolysin emulsified with TiterMax Gold (CytRx Corporation).

    SDS-PAGE and Western blot analysis.

    SDS-PAGE and Western blotting were carried out according to the methods of Laemmli (1970) and Towbin et al. (1979). Biotin-labelled goat anti-rabbit IgG (Chemicon) was used as a secondary antibody. Colour development was performed with an avidin–biotin peroxidase conjugate (ImmunoPure ABC peroxidase staining kit; Pierce). Pre-stained molecular markers (New England Biolabs) were used as size standards.

    Cytotoxic assay.

    Cytotoxicity of culture filtrates and purified proteins was assayed using Vero, HEp-2 and CHO cell monolayers. Vero and HEp-2 cells were grown in Dulbecco's modified Eagle's medium (DMEM) and CHO cells were grown in minimal essential medium (MEM). Media were supplemented with 10 % fetal calf serum. Cells were seeded in 96-well tissue culture plates at 104 cells per well. Samples were diluted twofold with DMEM or MEM and added to the Vero, HEp-2 or CHO cell monolayers. After 24 h of incubation at 37 °C in the presence of 5 % CO2, cytotoxic activities were recorded. The cytotoxic titre was defined as the reciprocal of the highest dilution of the sample that demonstrated 100 % destruction of the cells.

    Haemolysis assay.

    The haemolysis assay was carried out against human erythrocytes by a previously described procedure (Sha et al., 2002) with some modifications. Briefly, a serial twofold dilution of sample (100 μl) was mixed with an equal volume of 1 % human erythrocytes on a microdilution plate. The mixture was incubated at 37 °C for 1 h and subsequently at 4 °C overnight. The haemolytic titre was defined as the reciprocal of the highest dilution that showed complete haemolysis.

    Neutralization test.

    Culture supernatant was preincubated with an equal volume of 1 : 8-diluted anti-aerolysin antisera at 37 °C for 30 min. The mixture was then assayed for haemolytic and cytotoxic activities as described above. As a control, pre-immune sera was used instead of anti-aerolysin antisera.

    RESULTS AND DISCUSSION

    Characterization of AVP

    Partially purified AVP showed a single band at 34 kDa on SDS-PAGE (Fig. 1, lane 2). The protease activity of AVP was stable at 60 °C, but inactivated completely by heating at 80 °C for 15 min. The protease activity was significantly inhibited by metalloprotease inhibitors such as EDTA and OPA, but was not affected by serine protease inhibitors such as PMSF, TAME and LEME, suggesting that AVP is a metalloprotease (Table 3). AVP did not show cytotoxic or haemolytic activity (Table 5).

    Figure image not available in archive

    Fig. 1. Coomassie brilliant blue-stained SDS-12 % PAGE gel. Lanes: 1, molecular mass markers; 2, partially purified AVP; 3, purified recombinant proaerolysin.

    Table 3. Effect of protease inhibitors on AVP

    Table 5. Haemolytic and cytotoxic activities of purified proteins and culture filtrates produced by wild-type and AVP-deficient strains Haemolytic and cytotoxic activities are defined in Methods. The mixture of proaerolysin and AVP was incubated at room temperature for 5 min before adding to the cells. Cell-free supernatant (s/n) was obtained from organisms cultured in YEP broth at 37 °C for 24 h.

    The 20 N-terminal amino acid residues of AVP were KDATGPGGNIKTGKYVYGTD. Genetic analysis revealed that the predicted N-terminal amino acid sequence matched that obtained from purified AVP. The two motifs of the zinc-binding sites were revealed as HEVSH and GINEAFSD, which are conserved regions in the thermolysin family of bacterial metalloproteases (Miyoshi & Shinoda, 2000). The deduced amino acid sequence of AVP showed about 90 % identity to other Aeromonas metalloproteases (Izawa & Hayashi, 1996; Cascón et al., 2000; Kawakami et al., 2000). AVP lacks elastolytic activity, in contrast to the A. hydrophila metalloprotease (AhyB) reported by Cascón et al. (2000).

    Characterization of the AVP-deficient mutant

    A chloramphenicol-resistant A. veronii transconjugant (AVP8) was isolated, which has pM8 integrated at the chromosomal avp region. The AVP-deficient mutant showed weak protease activity, compared with the wild-type strain (Table 4). No significant difference in growth rate was found between the wild-type and AVP-deficient mutant strains (data not shown). Protease activity of the AVP-deficient mutant was approximately 45 % of that in the wild-type strain (Table 4). The observation that the mutant still exhibited protease activity suggested that other proteases were produced by the AVP-deficient mutant, a suggestion supported by the fact that PMSF (a serine protease inhibitor) and EDTA (a metalloprotease inhibitor) inhibited the proteolytic activity of the AVP-deficient mutant (Table 4). Protease activity in the mutant may be due to the 68-kDa serine protease (Rivero et al., 1991) and the 19-kDa metalloprotease (Toma et al., 1999) that have been reported for other Aeromonas. We have confirmed production of a 68-kDa serine protease and a 19-kDa metalloprotease by the wild-type strain (data not shown).

    Table 4. Effect of protease inhibitors on culture filtrates Residual activities are shown as percentages relative to protease activity of the wild-type strain without inhibitor. Cell-free supernatants were obtained from the organisms cultured in YEP broth at 37 °C for 24 h.

    Cytotoxic and haemolytic titres of the AVP-deficient mutant were more than eightfold lower than those of the wild-type strain (Table 5). The cytotoxic and haemolytic activities were neutralized completely by anti-aerolysin antisera (Table 5), but not by pre-immune rabbit sera, indicating that the toxic activities were caused by aerolysin.

    Production of aerolysin in the culture supernatant was detected by Western blot analysis. In the wild-type strain, proaerolysin (52 kDa) was activated completely to mature aerolysin (47 kDa; Fig. 2, lane 1). In contrast, proaerolysin in the AVP-deficient mutant was only partially activated (Fig. 2, lane 2). Western blot results supported the data obtained in the cytotoxic and haemolytic assays. In the AVP-deficient mutant, proaerolysin was less activated, which led to lower cytotoxic and haemolytic activities.

    Figure image not available in archive

    Fig. 2. Western blot reacted with anti-aerolysin antisera, showing the production of aerolysin in supernatants of wild-type (lane 1) and AVP-deficient (lane 2) strains (lane 3, purified proaerolysin). Strains were grown in YEP broth at 37 °C for 24 h with shaking and cell-free supernatants were concentrated 10-fold with 60 % saturated ammonium sulfate. Each lane was loaded with 20 μl sample.

    AVP activates proaerolysin

    Recombinant proaerolysin (Fig. 1, lane 3) did not show haemolytic activity, but had slight cytotoxic activity (Table 5). Treatment of proaerolysin with AVP resulted in an increase in haemolytic and cytotoxic activities (Table 5). Activity of proaerolysin following AVP treatment was lower than that in the culture supernatant. The reason for the higher cytotoxic activity in supernatant was that the AVP concentration in supernatant was higher than 0.04 μM (data not shown). Cytotoxic activity of recombinant proaerolysin can be interpreted with reference to the data of Abrami et al. (1998), who reported that a protease produced by the cells activates proaerolysin. In CHO cells, the responsible protease has been determined as furin (Abrami et al., 1998). HEp-2 and Vero cells may have a similar, as yet undefined, mechanism.

    Activation of proaerolysin in the AVP-deficient mutant may be due to the serine protease. A role of serine protease (AspA) in processing glycerophospholipid : cholesterol acyltransferase (GCAT) from Aeromonas salmonicida has been demonstrated (Vipond et al., 1998). The contribution of the 19-kDa metalloprotease also needs to be investigated; however, our results showed that the 34-kDa metalloprotease (AVP) is the major bacterial proaerolysin activator.

    The deduced amino acid sequence of A. veronii bv. sobria (strain AeG1, HG 8) aerolysin is 99 % identical to that of the A. sobria strain 357 aerolysin-like toxin (Fujii et al., 1998). Only four amino acid residues differ, Ala247, Asn454, Glu458 and Gly459 in A. veronii aerolysin, which are Val247, Ser454, Asp458 and Glu459 in A. sobria aerolysin. Studies on the cleavage site of A. sobria aerolysin have been carried out by Nomura et al. (1999). By determining the C-terminal region of mature aerolysin from the culture supernatant, the cleavage site of the toxin was determined to be between Ser446 and Ala447. Members of the thermolysin family of metalloproteases hydrolyse the peptide bond at the amino group side of the P1′ amino acid residue, usually a hydrophobic amino acid residue (Matthews, 1988). Although we have not determined the cleavage site of AVP, it is plausible that AVP, as a member of the thermolysin family, cleaves between Ser446 and Ala447 of proaerolysin. In addition, an AVP homologue, A. caviae PA protease (91 % similarity), cleaves the Ser−Ala site of pro-aminopeptidase (Izawa & Hayashi, 1996).

    Virulence factors of A. veronii are poorly understood. In this study, we have shown that the pathogenicity of A. veronii can be attributed to aerolysin, as in related aeromonads. Furthermore, we demonstrated that the 34-kDa metalloprotease is needed for high haemolytic and cytotoxic activities. Strains expressing only aerolysin may be not as virulent as strains expressing both aerolysin and the 34-kDa metalloprotease. This is the first report to demonstrate that Aeromonas metalloprotease is involved in cytotoxic activity through activating proaerolysin. Our data provide a better understanding of the interaction of multiple virulence factors and the role of metalloprotease in Aeromonas virulence.

    Acknowledgments

    We thank Yoshio Ichinose at the Institute of Tropical Medicine, Nagasaki University, for the N-terminal amino acid sequence.

    References

    1. Abrami, L., Fivaz, M., Decroly, E., Seidah, N. G., Jean, F., Thomas, G., Leppla, S. H., Buckley, J. T. & van der Goot, F. G. (1998). The pore-forming toxin proaerolysin is activated by furin. J Biol Chem 273, 32656–32661.
    2. Bagdasarian, M. & Bagdasarian, M. M. (1994). Gene cloning and expression. In Methods for General and Molecular Bacteriology, pp. 409–412. Edited by P. Gerhardt, R. G. E. Murray, W. A. Wood & N. R. Krieg. Washington, DC: American Society for Microbiology.
    3. Birnboim, H. C. & Doly, J. (1979). A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res 7, 1513–1523.
    4. Borrell, N., Acinas, S. G., Figueras, M.-J. & Martínez-Murcia, A. J. (1997). Identification of Aeromonas clinical isolates by restriction fragment length polymorphism of PCR-amplified 16S rRNA genes. J Clin Microbiol 35, 1671–1674.
    5. Cascón, A., Yugueros, J., Temprano, A., Sánchez, M., Hernanz, C., Luengo, J. M. & Naharro, G. (2000). A major secreted elastase is essential for pathogenicity of Aeromonas hydrophila. Infect Immun 68, 3233–3241.
    6. Fujii, Y., Nomura, T., Kanzawa, H., Kameyama, M., Yamanaka, H., Akita, M., Setsu, K. & Okamoto, K. (1998). Purification and characterization of enterotoxin produced by Aeromonas sobria. Microbiol Immunol 42, 703–714.
    7. Häse, C. C. & Finkelstein, R. A. (1993). Bacterial extracellular zinc-containing metalloproteases. Microbiol Rev 57, 823–837.
    8. Honda, T., Lertpocasombat, K., Hata, A., Miwatani, T. & Finkelstein, R. A. (1989). Purification and characterization of a protease produced by Vibrio cholerae non-O1 and comparison with a protease of V.cholerae O1. Infect Immun 57, 2799–2803.
    9. Howard, S. P. & Buckley, J. T. (1985). Activation of the hole-forming toxin aerolysin by extracellular processing. J Bacteriol 163, 336–340.
    10. Izawa, N. & Hayashi, K. (1996). Cloning and nucleotide sequencing of the aminopeptidase gene from Aeromonas caviae T-64. J Ferment Bioeng 82, 544–548.
    11. Janda, J. M. & Abbott, S. L. (1998). Evolving concepts regarding the genus Aeromonas: an expanding panorama of species, disease presentations, and unanswered questions. Clin Infect Dis 27, 332–344.
    12. Kawakami, K., Toma, C. & Honma, Y. (2000). Cloning, sequencing and expression of the gene encoding the extracellular metalloprotease of Aeromonas caviae. Microbiol Immunol 44, 341–347.
    13. Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685.
    14. Leung, K. Y. & Stevenson, R. M. W. (1988). Tn5-induced protease-deficient strains of Aeromonas hydrophila with reduced virulence for fish. Infect Immun 56, 2639–2644.
    15. Loewy, A. G., Santer, U. V., Wieczorek, M., Blodgett, J. K., Jones, S. W. & Cheronis, J. C. (1993). Purification and characterization of a novel zinc-proteinase of Aeromonas hydrophila. J Biol Chem 268, 9071–9078.
    16. Matthews, B. W. (1988). Structural basis of the action of thermolysin and related zinc peptidases. Acc Chem Res 21, 333–340.
    17. Miyazato, T., Toma, C., Nakasone, N., Yamamoto, K. & Iwanaga, M. (2003). Molecular analysis of VcfQ protein involved in Vibrio cholerae type IV pilus biogenesis. J Med Microbiol 52, 283–288.
    18. Miyoshi, S. & Shinoda, S. (2000). Microbial metalloproteases and pathogenesis. Microbes Infect 2, 91–98.
    19. Nomura, T., Fujii, Y. & Okamoto, K. (1999). Secretion of haemolysin of Aeromonas sobria as protoxin and contribution of the propeptide region removed from the protoxin to the proteolytic stability of the toxin. Microbiol Immunol 43, 29–38.
    20. Ochman, H., Gerber, A. S. & Hartl, D. L. (1988). Genetic applications of an inverse polymerase chain reaction. Genetics 120, 621–623.
    21. Rivero, O., Anguita, J., Paniagua, C. & Naharro, G. (1990). Molecular cloning and characterization of an extracellular protease gene from Aeromonas hydrophila. J Bacteriol 172, 3905–3908.
    22. Rivero, O., Anguita, J., Mateos, D., Paniagua, C. & Naharro, G. (1991). Cloning and characterization of an extracellular temperature-labile serine protease gene from Aeromonas hydrophila. FEMS Microbiol Lett 81, 1–8.
    23. Sha, J., Kozlova, E. V. & Chopra, A. K. (2002). Role of various enterotoxins in Aeromonas hydrophila-induced gastroenteritis: generation of enterotoxin gene-deficient mutants and evaluation of their enterotoxic activity. Infect Immun 70, 1924–1935.
    24. Simon, R., Priefer, U. & Pühler, A. (1983). A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in gram negative bacteria. Bio/technology 11, 784–791.
    25. Toma, C., Ichinose, Y. & Iwanaga, M. (1999). Purification and characterization of an Aeromonas caviae metalloprotease that is related to the Vibrio cholerae hemagglutinin/protease. FEMS Microbiol Lett 170, 237–242.
    26. Towbin, H., Staehelin, T. & Gordon, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci U S A 76, 4350–4354.
    27. Vila, J., Ruiz, J., Gallardo, F., Vargas, M., Soler, L., Figueras, M. J. & Gascon, J. (2003). Aeromonas spp.and traveler's diarrhoea: clinical features and antimicrobial resistance. Emerg Infect Dis 9, 552–555.
    28. Vipond, R., Bricknell, I. R., Durant, E., Bowden, T. J., Ellis, A. E., Smith, M. & MacIntyre, S. (1998). Defined deletion mutants demonstrate that the major secreted toxins are not essential for the virulence of Aeromonas salmonicida. Infect Immun 66, 1990–1998.
    29. Wang, J.-T., Fang, C.-T., Hsueh, P.-R., Chang, S.-C. & Lun, K.-T. (2000). Spontaneous bacterial empyema caused by Aeromonas veronii biotype sobria. Diagn Microbiol Infect Dis 37, 271–273.
    30. Whayeb, S. A., Yamamoto, K., Castillo, M.-E. Y., Tojo, H. & Honda, T. (1996). Lysophospholipase L2 of Vibrio cholerae O1 affects cholera toxin production. FEMS Immunol Med Microbiol 15, 9–15.