ABSTRACT
Two different representative recombinant clones encodingAeromonas hydrophila lipases were found upon screening on tributyrin (phospholipase A1) and egg yolk agar (lecithinase-phospholipase C) plates of a cosmid-based genomic library of Aeromonas hydrophila AH-3 (serogroup O34) introduced into Escherichia coli DH5α. Subcloning, nucleotide sequencing, and in vitro-coupled transcription-translation experiments showed that the phospholipase A1 (pla) and C (plc) genes code for an 83-kDa putative lipoprotein and a 65-kDa protein, respectively. Defined insertion mutants ofA. hydrophila AH-3 defective in either pla orplc genes were defective in phospholipase A1 and C activities, respectively. Lecithinase (phospholipase C) was shown to be cytotoxic but nonhemolytic or poorly hemolytic. A. hydrophila AH-3 plc mutants showed a more than 10-fold increase in their 50% lethal dose on fish and mice, and complementation of the plc single gene on these mutants abolished this effect, suggesting that Plc protein is a virulence factor in the mesophilic Aeromonas sp. serogroup O:34 infection process.
Mesophilic motileAeromonas species are opportunistic and primary pathogens of a variety of aquatic and terrestrial animals, including humans; the clinical manifestations range from gastroenteritis to soft tissue infections, including septicemia and meningitis (19). Serogroup O:34 strains of mesophilic Aeromonas spp. have been recovered from moribund fish (36) and from clinical specimens (40); O:34 is the most common Aeromonasserogroup (31), accounting for 26.4% of all infections. Previous investigations have documented O:34 strains as important causes of infections in humans (30, 31). The varied clinical picture of Aeromonas infections, and gastroenteritic illness in particular, suggests that complex pathogenic mechanisms occur in aeromonads.
Most aeromonads elaboratore a variety of extracellular enzymes: proteases, DNase, RNase, elastase, lecithinase, amylase, lipases, gelatinase, and chitinases; some of them are now confirmed as toxins (32, 38, 42): the cytotoxic/cytolytic enterotoxin (10), three different hemolysins (24, 25, 27) and cytotonic enterotoxins (34, 46, 48). Some of these toxins (for instance the aerolysin) are involved in septicemic infection (8). However, no clear information is available, to our knowledge, about the possible role of other extracellular enzymes (elastase, amylase, gelatinase, and chitinases) in Aeromonaspathogenesis. Clearly, it seems that the extracellular lipases play an important role in pathogenesis, for instance the glycerophospholipid-cholesterol acyltransferases from A. hydrophila and A. salmonicida (15, 53), which are implicated in the pathogenesis of this bacterium.
Phospholipases (PL) produced by bacteria are involved in different pathogenic process (14, 52) and are often associated with intestinal damage (5, 22, 54). Members of the familyVibrionaceae produced secreted PL, some of which act as hemolysins and some of which act as glycerophospholipid-cholesterol acyltransferases (47, 50, 53). Some of these PL have been cloned and sequenced (18, 49, 51), for instance the alpha-hemolysin (glycerophospholipid-cholesterol acyltransferase) ofA. hydrophila (53). We report here the cloning, sequencing, identification of gene product, and role in virulence of two different genes of A. hydrophila AH-3 (serogroup O:34 [33]) encoding two different PL (PLA1 and PLC [lecithinase]).
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions.The bacterial strains and plasmids used in this study are listed in Table1. Escherichia coli strains were grown on Luria-Bertani LB Miller broth and LB Miller agar (6), while Aeromonas strains were grown on tryptic soy broth or agar (TSB and TSA) (37). Tributyrin-agar and egg yolk-agar were prepared as described in reference 6. Ampicillin (50 μg/ml), chloramphenicol (50 μg/ml), kanamycin (30 μg/ml), and/or tetracycline (20 μl/ml) was added to the different media when needed.
Bacterial strains, cosmids, and plasmids used in this study
General DNA methods.DNA manipulations were carried out essentially as previously described (44). DNA restriction endonucleases, T4 DNA ligase, E. coli DNA polymerase (Klenow fragment), and alkaline phosphatase were used as recommended by the suppliers.
DNA sequencing.The primers used for DNA sequencing were purchased from Pharmacia LKB Biotechnology. Double-stranded-DNA sequencing was performed by the dideoxy-chain termination method of Sanger et al. (45) with the Abi Prism dye terminator cycle-sequencing kit (Perkin-Elmer).
DNA and protein sequence analysis.The DNA sequence was translated in all six frames, and all open reading frames longer than 100 bp were inspected. Deduced amino acid sequences were compared with those of DNA translated in all six frames from the nonredundant GenBank and EMBL databases by using the BLAST network service at the National Center for Biotechnology Information (2). Multiple-sequence alignments and determination of putative terminator sequences were done with the PileUp and Terminator programs from the Genetics Computer Group package (GCG, Madison, Wis.) on a VAX 4300.
Construction of pla and plc mutant strains.To obtain mutants with defined insertion mutations in thepla and plc genes, a method based on the use of the suicide plasmid pFS100 was applied (43). Plasmid pBR-PLA2 was BglII digested and blunt ended by treatment with the Klenow fragment of DNA polymerase, and a plainternal DNA fragment (1,230 bp) was isolated, ligated toEcoRV-digested and dephosphorylated pFS100, and transformed into E. coli MC1061(λpir) to generate plasmid pFS-PLA.
Plasmid pBR-PLC was double digested with SalI andKpnI and blunt ended, and a plc internal DNA fragment (697 bp) was isolated, ligated to EcoRV-digested and dephosphorylated pFS100, and transformed into E. coliMC1061(λpir) to generate plasmid pFS-PLC. Plasmid pFS-PLA and plasmid pFS-PLC were isolated, transformed into E. coli SM10(λpir), and transferred by conjugation toA. hydrophila AH-405 (rifampin resistant) to obtain mutants with defined insertion mutations in the pla andplc genes, respectively.
Substrate specificity and enzyme activity measurements.The substrate specificity with neutral glycerides and glycerophospholipids was determined as previously described (16, 17). Lipase activity was initially determined with tributyrin by the method of Ihara et al. (28). PLC activity was tested initially withp-nitrophenylphosphorylcholine as described by Ingham and Pemberton (29).
Determination of extracellular activities.Hemolysin and cytotoxin assays were performed as previously described (37). Briefly, hemolysin activity was assayed with a 1% suspension of sheep, bovine, or rainbow trout erythrocytes (4) and cytotoxin activity was assayed on Vero cell monolayers (7) and EPC (epithelioma papulosum of carp,Cyprinus carpium) monolayers (57). Enterotoxin activity was assayed by the rabbit ligated ileal loop assay as described by Nishibuchi et al. (41). In some E. coli strains, these activities were also assayed with the periplasmic proteins released by osmotic shock (56).
Virulence for fish and mice.The virulence of the strains grown at 20°C was measured by monitoring their 50% lethal dose (LD50) by the method of Reed and Muench, as previously described (37).
(i) Fish.Rainbow trout (12 to 20 g) were maintained in 20-liter static tanks at 17 to 18°C. The fish were injected intraperitoneally with 0.05 ml of the test samples (approximately 109 viable cells). Mortality was recorded up to 2 weeks; all the deaths occurred within 2 to 8 days.
(ii) Mice.Albino Swiss female mice (5 to 7 weeks old) were injected intraperitoneally with 0.25 ml of the test samples (approximately 5 × 109 viable cells). Mortality was recorded up to 1 week; all the deaths occurred within 2 to 5 days.
RESULTS AND DISCUSSION
Mesophilic Aeromonas strains from different O serogroups, but mainly from serogroup O:34, seem to produce more than one enzyme with lipopolytic activity when grown on egg yolk medium according to Matos et al. (35). It is clear that in this medium these strains produce a precipitate zone (lecithinase reaction) and an iridescent film or “pearly layer,” visible by reflected light, with lipase activity (35). Furthermore, O:34 strains were able to degrade trybutirin in a solid medium (lipolytic activity). We decided initially to call PLA the lipase activity on tributyrin medium and PLC the lecithinase activity on egg yolk medium. We decided to clone the corresponding genes from a single O:34 strain (AH-3) and see if they encoded two different PL activities.
Cloning of two A. hydrophila AH-3 genomic regions encoding PLA and PLC activities. A. hydrophila AH-3 (serogroup O:34) produces PLA and PLC activities. To determine the basis for these activities, a cosmid-based genomic library of A. hydrophila AH-3 was constructed and introduced into E. coli DH5α (21). Tetracycline-resistant (20 μg/ml) clones were screened independently on tributyrin and egg yolk-agar plates. We found two representative recombinant clones, COS-PLA and COS-PLC, with lipase activity on tributyrin and lecithinase activity on egg yolk-agar plates, respectively. It is important to note that COS-PLC showed only the precipitation zone, not the possible lipase activity (iridescent film) on egg yolk-agar plates.
To localize the genes involved in PLA (COS-PLA) and PLC (COS-PLC) activities, different subcloning experiments with some plasmid vectors were performed and the recombinant transformants were screened for the lipase (tributyrin) and lecithinase (egg yolk) activities, respectively. The strains and plasmids obtained are described in Table 1.
Sequencing of the DNAs conferring PLA and PLC activities.The nucleotide sequences of 2,602 and 2,868 bp were determined in both directions from plasmid pSK-PLA (pla) and pSK-PLC (plc), respectively; oligonucleotides M13 universal, reverse M13, SK, and other sequence-derived oligonucleotides were used to complete the nucleotide sequence.
Analysis of the deduced sequence of pSK-PLA showed a potential ORF (pla) (nucleotides 176 to 2591), encoding a putative protein of 805 amino acid residues with a predicted molecular mass of 82.7 kDa. Sequence analysis of pSK-PLC showed a potential ORF (plc) (nucleotides 885 to 2691), encoding a putative protein of 572 amino acid residues with a predicted molecular mass of 64.8 kDa. Upstream of the pla and plc genes, sequences similar to the ribosomal binding site were found. Sequences similar to the −10 and −35 consensus sequences of E. coli promoter were found upstream of plc gene, and a palindromic sequence, which could be involved in transcription termination, was found downstream from plc gene.
Analysis of the pla and plc deduced amino acid sequences.The deduced 805-amino-acid sequence frompla showed amino acid similarities to three extracellular lipases (LipE, Lip, and Apl-1) and a heat-labile cytotonic enterotoxin (Alt) from A. hydrophila (Table2). All five proteins show the lipase substrate binding signature sequence VHFLGHSL (13). Analyses of the Pla-1 amino acid sequence showed a putative lipoprotein signal sequence (residues 1 to 17) and a putative lipoprotein signal sequence cleavage site between residues 17 and 18. Another putative signal peptidase cleavage site was found between residues 48 and 49. Taken together, these two features strongly suggest that the Pla-1 is a secreted protein. Similar signal sequences and cleavage sites were previously found in the three extracellular lipases similar to Pla-1 but not in the heat-labile cytotonic enterotoxin. These results are in agreement with the previously reported similarity (47 to 51%) between Apl-1 and Alt (11). Furthermore, Alt also showed 54 and 52% of similarity to Lip (12) and LipE (3) extracellular lipases, respectively.
Identity and similarity of the amino acid sequence of Pla1 and Plc proteins from A. hydrophila AH-3 to other proteins
The putative Pla-1 lipoprotein had an apparent molecular mass of 83 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of plasmid pBR-PLA2 gene products with an E. coli S30 coupled transcription-translation system (Amersham) and [35S]methionine (Amersham) (Fig.1).
In vitro E. coli S30 coupled transcription-translation system with [35S]methionine performed as specified by the supplier (Amersham). Lanes: 1, pBR-PLC plasmid DNA (plc gene); 2, pBR328 plasmid DNA (cloning vector); 3, pBR-PLA2 plasmid DNA (pla gene). ∗, chloramphenicol acetyltransferase (25 kDa) from the Cmrdeterminant of pBR328 plasmid DNA; ∗∗, β-lactamase (31.5 kDa) from the Apr determinant of pBR328 plasmid DNA. Molecular mass standards are shown in kilodaltons.
The deduced 572-amino-acid Plc protein was found to be nearly identical to hemolysin ASH1 from A. salmonicida ATCC14174 (26), and furthermore there were only four different nucleotides between the plcand this hemolysin structural gene, showing that the plcgene is practically identical in both Aeromonas species. The Plc protein was also found to be similar to the unpublished PLD fromVibrio damsela (Table 2). However, no significative similarity was detected between Plc and a previously reported PLC fromA. hydrophila (Apl-1), which is similar to Pla-1 (Table 2). Analyses of the Plc amino acid sequence showed a putative signal sequence (residues 1 to 19) and a putative signal sequence cleavage site between residues 19 and 20. In vitro coupled transcription-translation experiments on plasmid pBR-PLC showed a polypeptide of about 65 kDa, similar in size to the expectedplc mature product (Fig. 1).
PLA and PLC are two different enzymes with PL activity. E. coli strains harboring the pla gene in the COS-PLA, pBR-PLA1, pBR-PLA2, and pSK-PLA plasmids were able to degrade tributyrin but unable to show any lecithinase activity on egg yolk medium. E. coli DH5α and the same strain carrying the same plasmids used as vectors without the DNA insert from AH-3 showed no activity in either trybutirin or egg yolk media. E. colistrains harboring COS-PLA, pBR-PLA1, pBR-PLA2, pACYC-PLA, and pSK-PLA plasmids were able to degrade other neutral glycerides (di- or triolein) or natural glycerophospholipids carrying a 1-acyl bond (phosphatidylcholine, phosphatidylinositol, phosphatidylethanolamine, or phosphatidylglicerol) but unable to degrade neutral glycerides like cholesteryloleate or p-nitrophenylacetate or substituted (at position 1-acyl) glycerophospholipids like 1-alkyl-2-acyl-sn-glyero-3-phosphocholine. For all these reasons, we concluded that this activity found in tributyrin plates is a PLA1 activity (15, 16).
E. coli strains harboring the plc gene in the Cos-PLC, pBR-PLC, pACYC-PLC, and pSK-PLC plasmids were able to precipitate the egg yolk medium (lecithinase activity) but unable to degrade tributyrin. Also, these strains were able to rapidly degrade the typical synthetic substrate for PLC,p-nitrophenylphosphorylcholine, at pH 7.2 with 2 mM CaCl2, and this activity was inhibited by spermine tetrahydrochloride (Calbiochem). The PL activity was dramatically reduced (<5% of the activity) on the same substrate at pH = 5.7 and 0.5 M CaCl2 (typical conditions for measuring PLD activity [20]). Two different assays were performed to differentiate PLC and PLD activities, and the PL activity determined by the plc gene showed (i) the presence of the diglyceride when incubated with lecithin as determined by thin-layer chromatography (petroleum ether-diethyleter-acetic acid [90:10:1, vol/vol]) and visualized by UV light as described by Grossman et al. (20) and (ii) hydrolysis of phosphatidylcholine and release of Pi in a low-Pi medium (33). The release of Pi was measured spectophotometrically by the method of Chen et al. (9), in either concentrated supernatants (AH-3 = 18 μg of Pi; AH-3PLC2 [plc insertion mutant] = 0.15 μg of Pi) or periplasmic fractions (DH5α = 0.3 μg of Pi; DH5α with the pACYC-PLC plasmid [plc] = 22 μg of Pi). From these results, we concluded that the lecithinase activity in these strains on egg yolk medium is mainly a PLC activity.
Construction of defined pla and plcinsertion mutants.Plasmid pFS-PLA, a replicationpir-dependent plasmid, carrying an internal fragment of thepla gene was transferred by mating to a rifampin-resistantA. hydrophila strain, AH-3 (AH-405 [39]), and Rifr and Kmr colonies were selected. We obtained mutants AH-3PLA1 and AH-3PLA4 unable to degrade tributyrin but still able to show lecithinase (PLC) activity. The insertion of plasmid pFS-PLA in these mutants was confirmed by Southern blot analysis with appropriate DNA probes. Complementation of these mutants with COS-PLA or pACYC-PLA restored the PLA1 activity on tributyrin medium.
Plasmid pFS-PLC carrying a plc gene internal fragment was used in identical way to that mentioned above to generate mutants AH-3PLC2 and AH-3PLC3 (also confirmed by Southern blot analyses), which were unable to show any lecithinase (PLC) activity on egg yolk medium but were able to degrade tributyrin (PLA1 activity). Complementation of these mutants with COS-PLC or pACYC-PLC restored the lecithinase activity on egg yolk medium. All these results indicate thatpla and plc are different genes in A. hydrophila AH-3 and also are unique genes for PLA1 and PLC activities in this strain.
PLC (lecithinase) is cytotoxic.As we previously shownA. hydrophila AH-3, as well as other strains from serogroup O:34, are hemolytic, cytotoxic, and enterotoxic (37, 38). As shown in Table 3, neither E. coli strains carrying pla nor A. hydrophilaAH-3 pla insertion mutants were altered in their hemolytic or cytotoxic activities (no activities found for E. colistrains in either supernatants or periplasmic cellular fractions). Also, nonenterotoxic activity was found in E. coli strains carrying the pla gene and reduced enterotoxicity was found in pla insertion mutants of A. hydrophila AH-3, despite the high homology between Alt (heat-labile cytotonic enterotoxin [11]) and Pla in 33% of the last protein. Then, we concluded that Pla is an extracellular lipase with PLA1 activity but is nonhemolytic, noncytotoxic, and nonenterotoxic.
Determination of some extracellular and periplasmic activities in different E. coli and A. hydrophilastrains as described in Materials and Methods
However, E. coli strains carrying the plc gene showed a clear cytotoxic activity on Vero and EPC cells (by assaying either supernatant or periplasmic fraction). Furthermore,plc insertion mutants of A. hydrophila AH-3 were reduced in their cytotoxic activity, and this activity was fully restored in these mutants by complementation with COS-PLC or pACYC-PLC (Table 3). These results for the cytotoxicity of Plc suggested that the wild-type A. hydrophila strain has several cytotoxic factors, with Plc being one of them. The results observed for the hemolytic activity of E. coli strains carrying theplc gene, plc insertion mutants of A. hydrophila AH-3, and the mutants complemented with COS-PLC or pACYC-PLC (Table 3) suggested the following. (i) Plc is nonhemolytic on sheep or bovine erythrocytes and only slightly on rainbow trout erythrocytes. These results are mainly in agreement with those reported for Hly1 of A. salmonicida (26), which has 99% identity to Plc (Table 2). The authors concluded that Hly1 (termed “hemolysin” despite its poorly hemolytic activity) found inA. salmonicida ATCC 14174 did not originate from this bacterium because of its low GC content, the codon usage frequency, and the negative results obtained by a DNA probe for this gene; unfortunately, they did not test if this protein showed any lecithinase activity (26). We suggest that this gene originates fromA. hydrophila. (ii) Plc is not a major hemolytic factor onA. hydrophila AH-3, because no major differences could be observed between the plc insertion mutants and the wild-type strain.
Finally, nonenterotoxic activity was found in E. colistrains carrying the plc gene and reduced enterotoxicity was found in plc insertion mutants of A. hydrophilaAH-3. We therefore concluded that Plc is nonenterotoxic.
PLC (lecithinase) is an important virulence factor.As pointed out by Vipond et al. (55), the major A. salmonicida secreted proteins (toxins) are not essential for the virulence of this bacterium, as they demonstrated with defined deletion mutants. We therefore decided to study the contribution to A. hydrophila pathogenesis of our defined pla andplc insertion mutants. We tested the virulence of the wild-type strain and the corresponding pla andplc insertion mutants (LD50), as shown in Table4. As can be observed, no differences were found in mortality between the wild-type strain (or they rifampin-resistant mutant) and mutants AH-3PLA1 and AH-3PLA4, which suggests that the PLA1 activity is not essential for virulence on these strains or that the mutation is unstable in vivo. However, mutants AH-3PLC2 and AH-3PLC3 showed a higher LD50 (an increase of 1 to 2 log units) in both fish and mice than the wild-type strain did. Complementation of these insertion mutants with COS-PLC or just with pACYC-PLC (carrying the singleplc gene) completely restored their virulence for fish or mice (similar LD50 to the wild-type strain [Table 4]). These results suggested that Plc (lecithinase) is an important virulence factor for mesophilic Aeromonas (serogroup O:34) pathogenesis.
Virulence for rainbow trout and mice of A. hydrophila AH-3 (serogroup O:34), defined insertion mutants, and mutants complemented by pla or plc
ACKNOWLEDGMENTS
This work was supported by grants from DGICYT and Plan Nacional de I+D (Ministerio de Educación y Cultura, Madrid, Spain). A.A. and M.M.N. are predoctoral fellows from the same institution and Generalitat de Catalunya, respectively.
We thank Maite Polo for her technical assistance.
Notes
Editor: J. T. Barbieri
FOOTNOTES
- Received 29 March 1999.
- Returned for modification 20 April 1999.
- Accepted 13 May 1999.
- Copyright © 1999 American Society for Microbiology
