INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Aeromonas spp. are enteropathogens (11, 22) and have been increasingly found in chlorinated drinking water, poultry, meat, seafood, cheese, milk, and produce (3, 14, 15, 25, 31). The majority of human isolates of Aeromonas spp. are hemolytic; soft-tissue necrosis is characteristic of extraintestinal Aeromonas infections, suggesting that aerolysin is an important virulence factor. In 1991 Carnahan et al. (5) described a new species belonging to the genus Aeromonas. The unique profile of this species includes negative reactions for esculin hydrolysis, arabinose fermentation, and the Voges-Proskauer test, positive reactions for cellobiose fermentation, lysine decarboxylation, and citrate utilization, and susceptibility to ampicillin (4, 5). The isolates of this species, placed in hybridization group 13, were described as members of the new species Aeromonas trota (5). A. trota strains have been found recently in southeast Asia, Europe, and the United States and are known to cause diarrhea in children (33) and adults (5). The mechanism of pathogenicity of Aeromonas hydrophila and the role of aerolysin have been studied previously (8); however, little is known about the A. trota aerolysin.

In this report, we describe molecular cloning and the complete nucleotide sequence of the hemolytic toxin (aerolysin) gene aerA of A. trota (HG 13) and its expression in Escherichia coli. In addition, in this study we demonstrated the application of the mini-Tn5Km1 transposon in Aeromonas species by isolating a mutant of A. trota that was unable to produce aerolysin.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. A. trota ATCC 49659 was obtained from the American Type Culture Collection, Rockville, Md. The authenticity of this strain was confirmed by biochemical methods (4, 5) and PCR (24). The E. coli strains and plasmids used in this study are listed in Table 1. Luria-Bertani (LB) broth consisted of 1% tryptone (Difco Laboratories, Detroit, Mich.), 0.5% yeast extract (Difco), and 0.5% NaCl (Sigma Chemical Co., St. Louis, Mo.). For LB agar plates, 1.5% agar (Difco) was added to LB broth. Organisms containing plasmids were grown and maintained on LB agar supplemented with appropriate antibiotics. For solid media, agar (Difco) was added at a concentration of 1.5%. Organisms were grown overnight at 37°C in LB broth or on tryptic soy agar (TSA) (Difco) plates supplemented with 5% defibrinated sheep blood (Remel, Lenexa, Kans.). Ampicillin, kanamycin, rifamycin, and tetracycline at concentrations of 100, 100, 50, and 15 µg/ml, respectively, were added to the medium when needed. The mutant strain A. trota AK253 was grown on TSA containing blood and 50 µg of kanamycin per ml. Long-term storage of bacteria was done in LB broth containing 20% glycerol at 70°C.

Generation of transposon mutants. Transposon mutagenesis of A. trota AK2 was performed by the method of de Lorenzo et al. (12). A. trota AK2 (Rif^r) and E. coli S17- pir, harboring the pUT (12) (derived from plasmid pGP704) derivative with mini-Tn5Km1, were grown overnight in 5 ml of LB broth. Rifamycin (20 µg/ml) was added to the A. trota culture, and kanamycin (100 µg/ml) was added to the E. coli culture. Samples (250 µl) of each culture were mixed in 5 ml of 10 mM MgSO₄ and filtered through a 13-mm-diameter cellulose acetate filter (Millipore Corp., Bedford, Mass.). The filter was placed on the surface of an LB agar plate and incubated for 24 h at 30°C. The cells grown on the filter surface were resuspended in 5 ml of sterile 10 mM MgSO₄, and 0.1 ml of the suspension was plated onto LB agar containing kanamycin and rifamycin. The plates were incubated at 37°C. The colonies growing on plates were transferred to TSA plates containing sheep blood (Remel) and incubated at 37°C overnight. The nonhemolytic mutants were picked from the plates, and their identities were confirmed by repeating the steps.

Cell fractionation and hemolysin assays. A. trota AK2 and AK253 and E. coli strains harboring recombinant plasmids encoding the aerA gene were grown in brain heart infusion broth (Difco) modified with appropriate antibiotics at 37°C with shaking for 16 h. Cells were removed by centrifugation, and the supernatants were used to determine extracellular hemolysin contents. Periplasmic fractions were obtained by the sucrose-EDTA method of Willis et al. (38). Briefly, the cells were washed twice with 10 mM Tris-hydrochloride (pH 7.0) and then suspended in 1/10th the original culture volume of buffer containing 10 mM Tris-hydrochloride (pH 7.4), 25% sucrose, 40 mM sodium EDTA, and 100 µg of lysozyme (Sigma Chemical Co.) per liter. After 30 min on ice, the cells were pelleted, the supernatant (designated the periplasmic fraction) was removed, and the cell pellet was suspended in 10 mM sodium phosphate buffer (pH 6.8). The cells were lysed by several (8 to 12) 15-s bursts with a Branson ultrasonifier (Branson Ultrasonic Corp., Danbury, Conn.) and were used directly for the hemolysin assay. Aerolysin hemolytic activity was assayed in 1-ml tubes containing 2% (vol/vol) (final concentration) defibrinated sheep blood (Remel) and appropriate volumes of various extracts containing hemolysin as described by Wagner et al. (37). The volumes of extracts were always adjusted in such a way that they were comparable (based on cell number). The reaction mixtures were kept at 37°C for 60 min, and the unlysed erythrocytes were removed by centrifugation for 1 min in an Eppendorf centrifuge (Brinkmann Instruments, Inc., Westbury, N.Y.). The absorbance of released hemoglobin was read at 540 nm. The activities were scored as either plus, which indicated that the optical density of the released hemoglobin was more than 0.60, or minus, which indicated that the optical density was less than 0.15. No intermediate values were found.

Molecular techniques. Total genomic DNA from A. trota AK253 was isolated as previously described (23). Plasmid DNA was isolated by the alkaline sodium dodecyl sulfate procedure of Birnboim and Doly (2) or was purified by the QIAprep spin column procedure (Qiagen, Inc., Chatsworth, Calif.). Restriction enzymes, T4 DNA ligase, calf intestine alkaline phosphatase, and a random primer labeling kit were purchased from Bethesda Research Laboratories (Gaithersburg, Md.). Standard methods for analysis of DNA, such as restriction endonuclease digestion, T4 DNA ligation, vector dephosphorylation with calf intestine alkaline phosphatase, and agarose gel electrophoresis, were performed as described previously (27). Plasmid DNA and ligation mixtures were transformed into E. coli cells by the CaCl₂ method, and transformants were selected by plating preparations onto media containing appropriate antibiotics (27). DNA was transferred from agarose gels to nylon membranes (NEN, Life Science Products, Boston, Mass.) with a Vacugene apparatus (Pharmacia LKB, Alameda, Calif.) as recommended by the supplier. DNA restriction fragments that were to be used as probes in Southern blotting experiments were separated by gel electrophoresis and were purified from agarose gels by using a DNA extraction kit (Qiagen). Purified DNA fragments were labeled by a random priming method (13). Southern hybridizations (35) were performed as recommended by the supplier (NEN). A chromosomal library of strain AK253 was constructed in E. coli JM109 with two plasmids, pGEM3Z and pGEM5Zf(), by digesting the chromosomal DNA with restriction enzymes KpnI and NotI, respectively. The transformed cells with a kanamycin-resistant plasmid were tested for hemolytic activity on TSA plates containing blood. The plasmids from hemolysin-positive E. coli clones were purified by using a plasmid minikit (Qiagen), were digested with various restriction enzymes, and were subcloned into pGEM3Z, pRK415, and pGEM11Zf() cloning vectors.

Determination of nucleotide sequence and analysis. The nucleotide sequence of the aerA gene was determined with a model 377 DNA sequencer (Perkin-Elmer, Foster City, Calif.). Both strands were sequenced with universal T7 and SP6 primers and by primer walking both strands with synthetic oligonucleotide primers. DNA sequence analysis, translation, and alignment with other related genes and proteins were done by using a computer program, Lasergene (DNASTAR, Inc., Madison, Wis.).

In vitro transcription-translation reaction. The protein encoded by plasmid pAK80, which included the entire coding region of the aerA gene, was identified by using a linked SP6 transcription translation kit (Amersham Life Sciences, Arlington Heights, Ill.). [³⁵S]methionine was used to label the protein, and the reactions were done under conditions recommended by the supplier (Amersham). The reaction mixtures were separated on sodium dodecyl sulfate-12% polyacrylamide gels and autoradiographed. A prestained, low-range molecular weight marker (Bio-Rad Laboratories, Hercules, Calif.) was used as a standard.

Nucleotide sequence accession number. The nucleotide sequence of the A. trota ATCC 49659 aerA gene has been deposited in the EMBL nucleotide sequence database under accession no. AF064068.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation of mini-Tn5Km1 insertion mutant of A. trota. A rifampin-resistant spontaneous mutant of A. trota ATCC 49659 was isolated and designated strain AK2 (Table 1). A. trota AK2 was mutagenized with a conjugative suicide delivery plasmid carrying the transposable element mini-Tn5Km1 (12) to produce another mutant strain, AK253. Southern blotting experiments performed with genomic DNA from strain AK253 digested with EcoRI, KpnI, SstI, XbaI, SphI, and NotI demonstrated that a single restriction fragment hybridized with the kanamycin resistance gene from mini-Tn5Km1 (1.7-kb EcoRI fragment) used as a probe (Fig. 1). Mini-Tn5Km1 does not have a restriction site for these enzymes. This means that mutant AK253 has only one copy of the transposable element in the genome and thus rules out the possibility that there are multiple insertions. A. trota AK2 mutants that were unable to secrete hemolytic aerolysins were initially detected with TSA plates containing blood; mutant strain AK253 completely lacks the hemolytic phenotype. The mutant strain was characterized by performing a hemolysin assay with culture supernatants, including the periplasmic contents, as well as whole-cell lysates. Further analysis of periplasmic and cell lysate fractions with the hemolysin assay showed no activity.

View larger version (54K): [in this window] [in a new window]   FIG. 1.   Agarose gel electrophoresis of A. trota AK253 (Tn5Km1) total DNA digested with various restriction enzymes and corresponding Southern blot hybridization. (A) Lane 1, molecular weight marker DNA (lambda phage DNA digested with HindIII and X174 DNA digested with HaeIII); lane 2, EcoRI; lane 3, KpnI; lane 4, SstI; lane 5, XbaI; lane 6, SphI; lane 7, NotI; lane 8, blank; lane 9, plasmid pUT-Tn5Km1 digested with EcoRI. (B) Autoradiogram of a Southern blot of panel A after hybridization in which the 32P-labeled 1.7-kb EcoRI fragment of mini-Tn5Km1 was used as the probe (15).

Cloning and subcloning of A. trota DNA encoding the aerolysin gene in E. coli. The Southern hybridization data (Fig. 1) indicated that the 18.0-kb KpnI and 8.7-kb NotI fragments contain the kanamycin resistance gene of mini-Tn5Km1. The DNA fragments from NotI and KpnI chromosome digests were excised and cloned into pGEM5Zf() and pGEM3Zf(), respectively, with selection for kanamycin resistance. The clones from both digests were analyzed for aerolysin expression on blood agar plates. The KpnI digest clone pAK60 was unable to express aerolysin on blood agar plates or in culture supernatants (data not shown). The NotI digest clone pAK71 was consistently hemolysin positive on blood agar plates. These data indicate that the structural gene for aerolysin, aerA, was adjacent to the transposon insertion and that it was intact. On the basis of these results, we concluded that either the transposon insertion in mutant AK253 interrupted the regulatory gene, which is required for expression of aerolysin in A. trota, or that there was a polar effect of Tn5. Figure 2 shows the locations of Tn5Km1 and the aerolysin gene in clone pAK71.

View larger version (9K): [in this window] [in a new window]   FIG. 2.   Restriction map of the cloned region in pAK71 and localization of the aerolysin gene. A box indicates the deduced position of the kanamycin (Km) resistance gene (from mini-Tn5Km1). The activity column indicates whether each clone expressed hemolysin. The arrows indicate the directions of transcription from the lac promoters on the cloning vectors [pGEM5Zf() for pAK71; pGEM3Z for pAK72 and pAK75; pRK415 for pAK76; pGEM11Zf() for pAK79 and pAK80]. The dashed arrows indicate the direction of aerA transcription. Abbreviations: A, ApaI; E, EcoRI; SI, SalI; S, SstI; K, KpnI.

Location of aerolysin in E. coli clones. E. coli JM109 cells containing pAK71, pAK72, and pAK80 showed no extracellular hemolytic activity in the culture supernatant after 16 h, indicating that the activity was completely intracellular. We looked for the presence of hemolytic activity in various fractions of osmotically shocked cells. Overnight cultures of strains harboring the recombinant plasmids pAK71, pAK72, and pAK80 were subjected to lysozyme treatment to release the periplasmic fractions. The cells were lysed by sonication, and both types of fractions were assayed for hemolytic activity. The periplasmic fractions showed hemolytic activity; however, there was no activity in the sonicated fractions of these three clones.

Identification of the product of the cloned gene as aerolysin. To identify the gene product encoded by the cloned aerolysin determinant, recombinant plasmid pAK80 and vector pGEM11Zf() were subjected to an in vitro transcription-translation reaction. The translated radiolabeled products were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and detected by autoradiography. The results are shown in Fig. 3. Plasmid pAK80 directed the synthesis of a 54-kDa protein, whereas the vector did not produce any detectable polypeptide.

View larger version (91K): [in this window] [in a new window]   FIG. 3.   In vitro transcription-translation of pAK80 and pGEM11Zf(). Lane 1, molecular size marker; lane 2, pAK80; lane 3, pGEM11Zf().

Nucleotide sequence of the aerolysin gene. To characterize the aerolysin gene in more detail, the nucleotide sequence of the aerA gene was determined by using clones expressing aerolysin. The aerA aerolysin gene sequence and the predicted amino acid sequence are shown in Fig. 4. The restriction endonuclease-mediated deletions indicated the location of the aerolysin gene. Both strands of the 1.83-kb EcoRI-ApaI DNA fragment were sequenced in their entirety. Within the region sequenced, there was only one long open reading frame, which was 1,479 bp long and included positions 397 through 1875 (Fig. 4). This open reading frame produced a protein with a molecular mass of 54.38 kDa, which is in agreement with in vitro transcription-translation data (Fig. 3). The protein was acidic, with a predicted pI of 5.83. The composition of the protein favored acidic residues, with 53 of the 492 amino acids being acidic. The predicted secondary structure of the aerA gene translated protein had a predominantly hydrophilic profile (data not shown). A typical ribosome-binding sequence, 5'-AAGGG-3' (Shine-Dalgarno sequence [34]), was found 5 bp upstream of the ATG aerolysin initiation methionine codon (Fig. 4). A 10 box (TGATAT) and a 35 region (TTGAGT) of the putative promoter were assigned to positions 318 to 323 and 297 to 302, respectively (Fig. 4). The G+C content of the aerolysin gene is 59%, which indicates that the gene is endogenous to A. trota.

View larger version (80K): [in this window] [in a new window]   FIG. 4.   Nucleotide sequence of the aerA gene region of A. trota and the predicted amino acid sequence of its product. The sequence shown includes the 1,479-bp open reading frame and flanking regions. The putative ribosome-binding site (RBS) is underlined. The dot indicates the position of the stop codon.

Sequence similarities and phylogenetic analysis. All of the aerA gene and protein sequences available from the GenBank library were searched and compared with the A. trota aerA gene and aerolysin sequences. Figure 5 is a phylogenetic tree based on aerolysins. A comparison of the deduced amino acid sequence of the aerolysin from A. trota with the amino acid sequences of other aerA gene products confirmed the designation of the A. trota aerolysin as an aerA gene product (Fig. 5). We found 99% similarity between it and the Aeromonas sobria AB3 aerolysin protein sequence; however, it showed only 57 to 78% sequence homology with other aerolysin protein sequences.

View larger version (10K): [in this window] [in a new window]   FIG. 5.   Phylogenetic tree produced by comparison of amino acid sequences of different aerolysins. The dendrogram was constructed with the Megalign program of the DNASTAR Lasergene biocomputing software package.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Aeromonas spp. produce a number of extracellular proteins that are known to play a vital role in pathogenicity (11, 22). Aerolysin is a hydrophilic protein which exhibits both hemolytic and cytolytic properties (1, 9, 19, 20, 26). Mutant AK253 did not show hemolytic activity in either extracellular or intracellular fractions; this may have been due to the transposon Tn5 insertion in the regulatory region of the aerolysin gene. Chakraborty et al. (7) reported that insertion of a transposon within a region on the cloned aerC gene of A. hydrophila led to reduction of the hemolytic phenotype, because the promoter for the aerolysin gene was involved in modulating the expression of the aerA gene product. In contrast, we observed that when the NotI fragment from AK253 was cloned into the pGEM5Zf() vector, aerA transcription was not affected in E. coli. Alternatively, the location of the insertion relative to the operon could be different.

Not all exotoxins cloned and introduced into E. coli K-12 are excreted efficiently; e.g., phospholipase C and the exotoxin of Pseudomonas aeruginosa (10), the aerolysin of A. hydrophila (7), the hemolysins of Vibrio cholerae and Vibrio parahaemolyticus (36), and cholera toxin of V. cholerae (29, 30) are not excreted efficiently. A similar situation was observed with the aerolysin gene of A. trota. The expressed active protein was localized in the periplasmic region of E. coli harboring the aerA gene. This implies that there may be fundamental differences in the secretory and excretory mechanisms of E. coli and A. trota.

The direction of transcription of the aerA gene was established by an in vitro coupled transcription-translation method by using the SP6 promoter of the vector. Potomski et al. (32) purified the cytotoxic enterotoxin of A. sobria by using monoclonal antibodies. The molecular mass of the purified enterotoxin was 63 kDa. This protein may represent proaerolysin, an immature aerolysin. The deduced molecular mass of the translated protein, 54 kDa, agrees closely with previously published values for aerolysins, which range from 50 to 65 kDa (6, 16-18, 32).

The nucleotide sequence of the aerolysin gene was determined, and the gene was located in the 1.8-kb EcoRI-ApaI region (Fig. 2). This region also contains the regulatory sequences of the aerolysin gene aerA. Upstream from the promoter is aerC, a gene involved in regulation of aerolysin activity (21). The G+C content of the aerA gene reading frame was 59%. A relatively high G+C content of the aerA gene is a characteristic of many aeromonads.

The similarities between the aerolysin sequence and the sequences of other hemolytic toxins revealed that the aerA gene of A. trota is 57 to 99% similar to other aeromonad aerA genes. The phylogenetic tree based on the deduced amino acid sequences of the aerolysin genes from Aeromonas spp. shows that there are three groups of aerolysin genes. In general, the members of each subfamily exhibit greater homology to each other than to the members of the other subfamilies. Most of the aligned aerA proteins showed a close relationship between the species and the aerA gene product except for A. sobria 33 (Fig. 5). The two A. sobria aerA gene products showed only 64% homology. In contrast, Aeromonas salmonicida 17-2 aerA was 98% homologous to A. sobria aerA. Most of the A. sobria AB3 and A. trota aerA amino acids are conserved; the only exceptions are the amino acids at positions 21, 22, and 44. It is possible that one of the A. sobria strains was misidentified or that more aerA gene sequence data are necessary for alignment. The phylogenetic interrelationships revealed by the 16S ribosomal DNA of Aeromonas species have shown that A. trota and A. sobria are in different clusters with 79% similarity (28).

In summary, the aerolysin gene was cloned from A. trota and was expressed in E. coli in an active form. The molecular mass of the aerolysin was 54 kDa. The nucleotide sequence of the entire aerA gene was determined, and this sequence had several conserved regions that may be responsible for activity. We are currently investigating the role of the aerC region in regulation of the aerA gene.