|Home | About | Journals | Submit | Contact Us | Français|
A novel multiplex PCR method using three sets of specific primers was developed for the detection of the cytotoxic (act), heat-labile (alt), and heat-stable (ast) enterotoxin genes in Aeromonas spp. This assay was used to characterize 35 reference strains as well as 537 food-borne isolates. A total of seven gene pattern combinations were encountered, including act, alt, act/alt, act/alt/ast, act/alt/148-bp amplicon, alt/ast, and alt/148-bp amplicon. The alt gene was detected with 34 reference strains (97%) and occurred singly in 14% of these strains. The frequency of occurrence of the act/alt, act/alt/ast, and alt/ast gene patterns in reference strains was 14 (40%), 2 (6%), and 2 (6%), respectively. An unpredicted amplicon was detected in 11 reference strains (31%). Characterization of this amplicon showed that its size was 148 bp, as generated by the AHLF and AHLR primers, and that it uniquely aligned with the Aeromonas salmonicida A449 genome sequence (GenBank accession number CP000644). This amplicon was named Aeromonas salmonicida subsp. salmonicida hypothetical protein amplicon (AssHPA). In the 537 food-borne isolates, the act and alt genes were most dominant and were detected in 349 (65%) and 452 (84%) isolates, respectively, either alone or in combinations. The act and alt genes occurred singly in 30 (6%) and 128 (24%) of these strains, respectively. The act/alt gene pattern occurred in 315 isolates (59%), whereas the ast gene was always linked to strains exhibiting the act/alt/ast and alt/ast gene combinations in 4 (0.7%) and 5 (0.9%) isolates, respectively. The uniplex amplification of three enterotoxin genes separately confirms the specificity of the unique selected primers. This multiplex PCR is rapid and simple and can detect the presence of three Aeromonas enterotoxin genes in a single assay.
Aeromonas spp. are widely distributed in aquatic environments and are isolated from a wide range of food of animal and plant origin (6). Aeromonas spp. can produce several putative virulence factors, including extracellular enzymes, siderophores, cytotoxic and cytotonic enterotoxins, Shiga-like toxins, endotoxins, invasins, and adhesins (18). Aeromonas spp. can grow and produce toxins in foods at refrigeration temperatures (17). The recent isolation of enterotoxigenic aeromonads from drinking water in the United States (21) reiterates the potential human health hazard of waterborne Aeromonas spp. The observation that strains harboring the alt and ast genes were more prevalent in children with diarrhea than in healthy controls underlines the importance of enterotoxins in the pathogenicity of aeromonads (3).
Exotoxins are major virulence factors of Aeromonas spp. that include a cytotoxic heat-labile enterotoxin (Act), also known as aerolysin/hemolysin (8); a cytotonic heat-labile enterotoxin (Alt), also known as lipase, extracellular lipase, or phospholipase (4, 10, 24); and a cytotonic heat-stable enterotoxin (Ast) (9). These toxins are encoded by the genes act (GenBank accession number M84709), alt (GenBank accession number L77573), and ast (GenBank accession number AF419157), respectively (22).
Although conventional uniplex PCR techniques have been used successfully to detect individual virulence factors in clinical, food, and environmental strains of Aeromonas spp. (1, 5, 6), a multiplex PCR approach would allow for simultaneous detection of these virulence genes. In a recent study, waterborne Aeromonas isolates were screened for different virulence genes in three series of multiplex PCRs targeting the elastase (ahyB), lipase (pla/lip/lipH3/alp-1), flagella A and flagella B (flaA and flaB), and enterotoxin act, alt, and ast genes (21).
The present work reports the development of a rapid, practical, and simple single-test tube assay multiplex PCR system for the simultaneous detection of the act, alt, and ast enterotoxin genes in food-borne Aeromonas spp. The assay was further used to assess the distribution of these genes in 537 food-borne Aeromonas isolates. In addition, this multiplex PCR could help to detect specifically A. salmonicida subsp. salmonicida by its ability to generate an amplicon named Aeromonas salmonicida subsp. salmonicida hypothetical protein amplicon (AssHPA).
Known hybridization groups (HGs) of 35 reference strains of Aeromonas spp. (14) (and other molecular and biochemical characteristics) were used to optimize the multiplex PCR method and included A. bestiarum HG2, A. caviae HG4, A. caviae/A. media HG5, A. eucrenophila HG6, A. hydrophila HG3, A. salmonicida HG3, A. sobria HG7, and A. veronii HG8/HG10 (see Table Table2).2). For the uniplex PCR, the following selected reference strains were used: A. bestiarum HG2 (LMG 13447) and A. hydrophila HG3 (LMG 13450) for the act gene, A. bestiarum HG2 (LMG 13446) and A. hydrophila HG3 (LMG 13450) for the alt gene, and A. hydrophila HG1 (LMG 13439) and A. hydrophila HG1 (LMG 13659) for the ast gene. Twenty reference strains of non-Aeromonas spp. were also used to assess the specificity of the method (5).
Of 459 food samples tested in this study, 330 (72%) were found positive for Aeromonas spp. by use of a standard culture method described here. Positive samples included beef (n = 14), fish (n = 21), lamb (n = 5), poultry (chicken and turkey; n = 44), pork (n = 18), salmon swabs (n = 48), sausages (n = 9), and seafood (n = 53). A total of 118 samples of fresh vegetables and fruits, including alfalfa sprouts, bean sprouts, broccoli, cauliflower, celery, cilantro, egg plant, papaya, curly parsley, fiddlehead, green onion, lettuce, spinach, parsnip, rapini, radish, and tofu, were also found to contain Aeromonas spp. Raw and processed meats and fresh vegetables were obtained from local retail outlets. Swabs of fresh eviscerated aquaculture and wild salmon, as well as homogenates of fresh oysters and clams, originated from Canadian Pacific coastal waters. A total of 537 Aeromonas strain isolates were selected from the positive samples and used to screen for the presence of enterotoxin genes by use of the novel multiplex PCR method.
A food sample (100 g) was suspended in 900 ml of buffered peptone water and incubated for 18 h at 28°C. A loopful of the buffered peptone water enrichment culture was inoculated onto ampicillin dextrin agar (13) and on Aeromonas agar (CM 833; Oxoid Inc., Nepean, Ontario, Canada), and plates were incubated for 18 h at 28°C and 35°C, respectively. Three suspect colonies were screened using determinant biochemical tests, including triple sugar iron agar, resistance to the vibriostatic agent 0129, the presence of oxidase, hemolysis on sheep blood agar, growth on McConkey agar and in nutrient broth with and without 6% NaCl, and API 20 NE profiling. Isolates were stored on beads (CryoStor; Innovatek Medical Inc., Vancouver, British Columbia, Canada) at −80°C.
At least three isolates of each Aeromonas sp., 25 in total, were randomly selected for molecular characterization, which included gyrB and 16S rRNA gene sequencing as previously described (21).
A bead stored at −80°C was inoculated into 3 ml of tryptone soya broth and incubated for 24 h at 35°C. A portion (100 μl) of the tryptone soya broth culture was added to 1 ml of sterile double-distilled water, vortexed, and then centrifuged (12,000 × g) for 3 min at 4°C. The supernatant was decanted, and the pellet was mixed with 200 μl InstaGene matrix (Bio-Rad Laboratories, Ltd., Mississauga, Ontario, Canada) and incubated for 30 min at 56°C. The mixture was vortexed for 10 s and incubated in a heating block (Eppendorf Thermomixer; Eppendorf AG, Hamburg, Germany) for 10 min at 99°C. The reaction mixture was then centrifuged (12,000 × g) for 3 min at 4°C, and an aliquot (2.5 μl) of the resulting supernatant was used as the DNA template. The remaining supernatant was stored at −20°C for future use. The nucleic acid concentration in the DNA template was quantified using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Wilmington, DE). The average amount of nucleic acid template per PCR (2.5 μl) extracted from 30 isolates was 150 ng or 60 ng/μl.
Primers (AHCF and AHCR) for the amplification of the cytotoxic enterotoxin gene (act; GenBank accession number M84709) have been described previously (5). Novel primers were designed using Clone Manager 9 professional edition software (Scientific & Educational Software, Cary, NC) as follows. For the amplification of the heat-labile enterotoxin (alt) gene (11), the design of the unique primers (AHLF [5′-TGC TGG GCC TGC GTC TGG CGG T-3′] and AHLR [5′-AGG AAC TCG TTG ACG AAG CAG G-3′]) was based on the large open reading frame (ORF) sequence (position 106 to 1212) of the A. hydrophila alt gene (GenBank accession number L77573). For the amplification of the heat-stable enterotoxin (ast) gene (9), the design of the unique primers AHSF ([5′-GAC TTC AAT CGC TTC CTC AAC G-3′] and AHSR [5′-GCA TCG AAG TCA CTG GTG AAG C-3′]) was based on ORF1 of the heat-stable cytotonic enterotoxin ast gene in A. hydrophila (GenBank accession number AF419157). The characteristics of the primers used in this multiplex PCR are described in Table Table1.1. All primers were synthesized by the Biotechnology Research Institute, University of Ottawa, Ottawa, Ontario, Canada.
The master mixtures for the uniplex and multiplex PCR were made using the HotStartTaq master mix kit (Qiagen Inc., Mississauga, Ontario, Canada). The uniplex PCR, as well as the single-tube multiplex PCR, was based on a 25-μl total volume of a master mixture containing 2.5 μl of DNA (approximately 150 ng) as a template. The master mixture formula was prepared in a large volume in a sterile laminar airflow environment to produce approximately 100 tubes of 22.5 μl each containing all of the ingredients except the DNA template. The master mixture formulas were as follows: 250 μl of 10× PCR buffer containing 15 mM MgCl2; 4 μl of MgCl2 at 25 mM; 125 μl of individual primers AHCF and AHCR for the act gene, AHLF and AHLR for the alt gene, and AHSF and AHSR for the ast gene for the uniplex PCR at 0.2 μM each and AHCF, AHCR, AHLF, AHLR, AHSF, and AHSR at 0.2 μM each, all together for the multiplex PCR; 50 μl of a mixture containing each deoxynucleoside triphosphate at 200 μM each; 25 μl of Taq polymerase of 5 U/μl; 50 μl of Tween 20; and 1,121 μl of sterile distilled water. Tubes were stored at −20°C until needed. Following an initial activation step of 15 min at 95°C, the reaction mixture was amplified for 35 cycles using an iCycler system (Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) set for denaturation for 15 s at 95°C, annealing for 30 s at 69°C, and extension for 30 s at 72°C; this amplification program was concluded with a final elongation for 10 min at 72°C. The consistency of the PCR assay was ascertained in triplicate using the reference strains listed in Table Table2.2. Each amplicon preparation (13 μl, including 3 μl of stop solution) was loaded onto a 2.5% agarose gel (pulsed-field-certified agarose; Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) and separated by electrophoresis in 0.5× Tris-borate-EDTA buffer (Roche Diagnostics, Montreal, Quebec, Canada) for 1 h at 100 V. Separated amplicons on gels stained for 15 min in ethidium bromide (0.5 μg/ml) (Bio-Rad, Mississauga, Ontario, Canada) were visualized with UV light. The sizes of amplicons were determined using molecular size marker VIII (Roche Diagnostics, Montreal, Quebec, Canada).
To confirm the specificity of the amplified amplicons, the alt and ast amplicons of the 25 molecularly characterized food isolates and the seven 148-bp amplicons of the fish moribund isolates were sequenced and analyzed. Amplicons were purified from the cut band of MetaPhor agarose (FMC BioProducts, Rockland, ME) using the KingFisher PCR and gel cleanup kit (ThemoLab Systems, Vantaa, Finland) according to the manufacturer's instructions. Purified PCR products were concentrated using a SpeedVac concentrator (Thermo Fisher Scientific Inc., Nepean, Ontario, Canada). The purified and concentrated amplicon products were cloned using the TOPO TA cloning kit (Invitrogen Corporation, Carlsbad, CA) according to the manufacturer's instructions. From each cloning plate two white colonies were selected and inoculated in 2 ml of Luria broth supplemented with 100 μg/ml of ampicillin and grown overnight at 37°C. The plasmids were purified using the GenElute miniprep kit (Sigma, Oakville, Ontario, Canada). Sequencing was accomplished using the BigDye version 3.1 sequencing kit, POP-7 polymer gel, and a 36-cm capillary (Applied Biosystems, Foster City, CA), and the M13F and M13R primers (Sigma, Oakville, Ontario, Canada). The sequencing reaction products were cleaned using Seq cleanup plates (Millipore, Nepean, Ontario, Canada), and the electrophoresis was done using the genetic analyzer system 3130 XL (Applied Biosystems, Foster City, CA). The sequences of the 25 food isolates were subject to multialignment analysis using Clone Manager 9 professional edition software (Scientific & Educational Software, Cary, NC), and the consensus sequence of the seven fish isolate amplicons generated by DNA alignment using Clone Manager 9 professional edition software (Scientific & Educational Software, Cary, NC) was subject to a BLAST N search for its similarity to available Aeromonas gene sequences in GenBank, including the Aeromonas hydrophila subsp. hydrophila ATCC 7966 NC 008570 and Aeromonas salmonicida subsp. salmonicida A449 NC 009348 sequences.
Biochemical characterization identified to species level the 537 Aeromonas strain isolates into four species complexes, including the A. caviae complex (n = 191; 36%), A. hydrophila complex (n = 206; 38.4%), A. sobria complex (n = 13; 2.4%), and the nonassigned Aeromonas sp. complex (n = 127; 24%), which originated mainly from salmon swabs and seafoods (Table (Table33).
The specificity of the designed primers was ascertained by searching the GenBank repository for similar sequences using the BLAST N system (Table (Table1).1). The screening method used to confirm the specificity of primers AHCF1 and AHCR1 was described previously (5). In a search of Aeromonas and other bacterial genera in GenBank, sequences 100% homologous to AHLF and AHLR were encountered only in nine and 10 gene sequences, respectively, of Aeromonas spp., including the alt, pla, and lip genes (Table (Table1).1). Similarly, a search of GenBank identified sequences homologous to primers AHSF and AHSR only in the ast gene (accession number AF419157) and in the genome sequence of Aeromonas hydrophila ATCC 7966T (accession number CP000462) (23) (Table (Table11).
The six reference strains used for the uniplex PCR produced the predicted amplicons (Fig. (Fig.1).1). Examination of reference strains by the multiplex assay showed that 34 strains (97%) harbored one or more enterotoxin genes corresponding to one of the six enterotoxin gene profiles alt, act/alt, act/alt/ast, alt/ast, act/alt/AssHPA, and alt/AssHPA (Table (Table2).2). The act gene always occurred in combination with other enterotoxin genes in reference strains, whereas the ast gene was restricted in A. hydrophila HG1 to gene profiles act/alt/ast and alt/ast. The alt gene was detected in all strains of A. bestiarum (HG2), in which it occurred alone or in combination with act and AssHPA (Table (Table2).2). For strains of HG3, all A. hydrophila isolates from clinical or environmental materials presented a single act/alt gene pattern. In contrast, the two ATCC A. salmonicida fish isolates contained the act/alt or alt gene patterns, whereas the act/alt/AssHPA profile was encountered in nine (82%) A. salmonicida isolates from moribund fish. A. caviae (HG4) and A. caviae/A. media (HG5) were positive either for the alt (three isolates [60%]) or the alt/AssHPA (one isolate [20%]) gene profile. Enterotoxin genes and AssHPA in one strain of A. caviae HG4 could not be detected. A single gene pattern, act/alt, was encountered with A. veronii (HG8/10).
The BLAST N system similarities of PCR product sequences of the 25 molecularly characterized food isolates are shown in Table Table4.4. The similarities of the alt amplicons to the A. hydrophila subsp. hydrophila ATCC 7966 complete gene sequence in GenBank (NC_008570) ranged from 86 to 97%. The similarities of the alt amplicons to the Aeromonas salmonicida subsp. salmonicida complete gene sequence in the GenBank (NC_009348) ranged from 87 to 97%. The results of the sequencing of the unpredicted amplicons showed that the size of the amplicons was 148 bp. The multiway and global DNA alignments of the amplicon sequences using Clone Manager 9 professional edition software (Scientific & Educational Software, Cary, NC) showed similarities of 98 to 100% compared to the consensus sequence. Using the same programs with the six primer sequences used in the multiplex PCR showed that the amplicons were generated by primers AHLF and AHLR. A search for the sequences similar to the 148-bp amplicon consensus sequence using the GenBank repository BLAST N system revealed a unique alignment from site 2124904 to site 2124780 and a similarity of 97% to A. salmonicida subsp. salmonicida A449 GenBank accession number NC_009348. This fragment is part of a gene encoding a hypothetical protein of Aeromonas salmonicida subsp. salmonicida. Accordingly, we named the amplicon, as a novel finding, Aeromonas salmonicida subsp. salmonicida hypothetical protein amplicon (AssHPA).
The 537 strain isolates of food-borne Aeromonas spp. were tested using the novel multiplex PCR assay for the presence of the act, alt, and ast enterotoxin genes. Of these, 191 (36%) belonged to the A. caviae complex, 206 (38%) to the A. hydrophila complex, and 13 (2.4%) to the A. sobria complex, and 127 (23%) isolates which could not be identified to species level were assigned empirically to the Aeromonas sp. complex based on determinant biochemical tests and API 20NE reactions (Table (Table3).3). Five enterotoxin gene patterns, including act, alt, act/alt, act/alt/ast, and alt/ast (Table (Table3),3), were associated with food-borne aeromonads, of which 482 (90%) isolates carried one or more enterotoxin genes.
The alt gene was detected in 164 (86%) A. caviae complex isolates, in which it occurred either alone (101 isolates [53%]) or in combination with the act gene (62 isolates [33%]) or ast gene (1 isolate [0.5%]). The act gene was encountered with 68 A. caviae complex isolates (36%), either alone (n = 6; 3%) or in combination with the alt gene (n = 62; 32.5%).
The act and alt genes were prominent among members of the A. hydrophila complex. The act gene occurred in 171 (83%) isolates, either alone (n = 9; 4.4%), in combination with alt (n = 162; 79%), or in combination with ast (n = 4; 2%). The alt gene was detected in 181 isolates (88%), either alone (n = 15; 7%) or in combination with the act gene (n = 162; 77%) or the ast gene (n = 4; 2%).
The act and alt genes occurred either alone or in combination in 3 of the 13 strains (23%) of A. sobria isolated and tested.
The 127 strain isolates within the Aeromonas sp. complex positive for enterotoxin genes were recovered only from salmon swabs and seafood samples. The act gene was encountered with 104 (82%) of these strain isolates alone (9.5%) or as an act/alt combination (92 isolates [72%]). The alt gene was detected in 101 (79.5%) of the Aeromonas sp. isolates either alone (n = 9; 7.1%) or in combination with act (n = 92; 72.4%).
In summary, the act (65%) and alt (84%) genes showed the highest frequency of occurrence in test strain isolates. In contrast, detection of the ast gene was restricted to reference strains of A. hydrophila HG1, to a single food-borne isolate of A. caviae, and to eight food-borne isolates belonging to the A. hydrophila complex. The ast gene in food-borne aeromonads never occurred singly but always in the act/alt/ast (n = 4; 0.74%) and alt/ast (n = 5; 0.93%) gene combinations.
The uniplex amplification of reference strains of different Aeromonas HGs showed and confirmed the specificity of the designed primers used in this study (Fig. (Fig.1).1). Also, the amplicon sequences of the food isolates that were characterized by molecular methods exhibited high similarity to reference sequences of Aeromonas hydrophila subsp. hydrophila and Aeromonas salmonicida subsp. salmonicida in the GenBank database (Table (Table4),4), thus confirming the specificity of the method.
PCR-restriction fragment length polymorphism and DNA amplicon sequence analysis of the 232-bp amplicons derived from the AHCF1 and AHCR1 primers have previously confirmed the diversity of the 232-bp amplicons, including the hemolytic, cytotoxic, and enterotoxic aspects of the A. hydrophila cytolytic enterotoxin gene (5). It appears that the act gene may fulfill a multifunctional role in Aeromonas spp. by encoding proteins with hemolytic, cytotoxic, and enterotoxic activities (8). Several studies (6, 11, 12, 20, 21, 25) have confirmed the usefulness of AHCF1 and AHCR1 primers for the detection of the act gene. For instance, the detection of act genes with AHCF1 and AHCR1 primers in a uniplex PCR assay was reported for 65% of Aeromonas strains isolated from different sources (5): from 70% of water isolates from the United States (21), 77% of isolates from different materials from Libya (1), and 67% of clinical isolates from southern Taiwan (25). In this study, the use of these two primers in this novel multiplex PCR system detected the act gene alone or as a member of different enterotoxin gene combinations with 65% of food-borne isolates, similar to the results reported in an earlier study (5). Accordingly, the inclusion of the AHCF1 and AHCR1 primers in the novel multiplex PCR was essential.
The species-dependent distribution of the act gene is of interest. The aerolysin/hemolysin genes (7) and the act gene (19) were detected in 12 and 19%, respectively, of the A. caviae strains examined. In the present study, the detection of the act gene alone or in combination with the alt gene in food-borne A. caviae (36%) (Table (Table3)3) was significantly higher than previously reported (7, 25). The act genes were not detected with reference strains of A. caviae (HG4) or A. media (HG5) in an earlier uniplex PCR study (5) or in the reference strains examined in the present multiplex study, possibly because species identification was based on the more reliable fatty acid analysis (14) and amplicon fragment length polymorphism (15) techniques rather than the conventional biochemical tests and API 20 NE used in this study. Similarly, strains of A. caviae characterized by PCR-restriction fragment length polymorphism of the 16S rRNA genes lacked the act gene (1). It is also accepted that methods used for isolate species identification, such as DNA-DNA hybridization, can highly improve the characterization of some Aeromonas isolates (16). Reports of the presence of the act gene in A. caviae are inconsistent and could be related to the method used for species identification (Tables (Tables22 and and3),3), to the primers used in the PCR assay (7), or to the geographic variation in the distribution of the act gene in A. caviae spp. (5).
The occurrence of the act gene in food isolates was highest in isolates of A. hydrophila (83%) compared to A. caviae (36%) and A. sobria (46%) isolates but was similar to the incidence level encountered with Aeromonas isolates that were not identified to species level (82%). The act gene was detected in 65% of the A. hydrophila isolates originating from clinical, food, and environmental samples (1, 5, 25). These findings contrast with the reported rare occurrence of the act gene in Aeromonas isolates from a drinking water plant in the United States (21).
The isolation of only 13 strains of A. sobria from foods was insufficient to reliably measure the occurrence of enterotoxin genes in this species (Table (Table3).3). Interestingly, the act gene was detected in six (68%) of the enterotoxin gene-positive A. sobria isolates, as observed in other studies (1, 5, 25).
The act gene was detected in 82% of the Aeromonas isolates that were not identified to species level (Table (Table3).3). These strains which originated exclusively from salmon swabs or seafood may more accurately belong to A. salmonicida. Currently, the differentiation of A. hydrophila (HG3) from A. salmonicida (HG3) using a traditional biochemical test, such as the API 20 NE diagnostic panel, fatty acid analysis (14), or DNA-based methods (1), is limited. The limited occurrence of the 148-bp amplicon of uncertain function in the A. bestiarum HG2 (ATCC 14715) fish isolate, the A. caviae/A. media HG5 (LMG13461) clinical isolate, and A. hydrophila HG3 moribund fish isolate (Table (Table2)2) may prove to be of taxonomic significance. However, as a new finding, further investigation is needed to determine the nature and role of this DNA fragment in Aeromonas spp., especially in Aeromonas salmonicida subsp. salmonicida.
For the development of this novel multiplex PCR method, the alt and ast genes encoding cytotonic enterotoxins were attractive as targets. They encode heat-labile (non-cholera toxin cross-reacting Shiga-like toxins) and heat-stable (56°C, 10 to 20 min, cross-reacting and non-cross-reacting cholera toxin) enterotoxins, respectively (18). In the event of food poisoning, an Aeromonas isolate harboring any of these genes may be a hazard, and its detection is essential.
The alt gene codes for a heat-labile toxin which cross-reacts with Shiga-like enterotoxin (10). A search of GenBank for the alt gene sequence of A. hydrophila (GenBank accession number L77573) showed 84 to 95% identity with related alt, pla, and lip gene sequences (Table (Table1).1). This homology was used to design two unique primers, AHLF and AHLR, targeting the conserved region within the main ORF1 of alt-pla-lip, which spans from site 106 to 1212 and generates a DNA fragment of 1,106 bp. This selective design strategy enabled amplification of any of the listed genes (Table (Table1)1) in the sample community bacterial DNA. The greater intensity of the 361-bp amplicons shown in Fig. Fig.11 (lanes 4 and 5) and Fig. Fig.22 (lanes 2, 3, 6, and 7) putatively indicates the presence of more than one amplicon amplified from these isolates compared to the isolates represented in Fig. Fig.22 (lanes 4 and 5). The restriction enzyme analysis of the 361-bp amplicons with HpaII confirmed the presence of other distinct amplicons (data not shown).
As previously reported (7), the alt gene occurs commonly in Aeromonas spp. Food-borne A. caviae isolates which were identified using 20NE and other biochemical tests were positive for the alt gene alone (53%) or in combinations with other enterotoxin genes (86%). The apparent discrepancy between the incidence of the alt gene alone in reference strains of A. caviae and A. caviae/A. media (Table (Table2)2) and the occurrence of different alt enterotoxin gene combinations in food-borne isolates of the same species (Table (Table3)3) may stem from differences in the accuracies of biochemical and molecular techniques that were used for identification to species level of food-borne and reference strains of Aeromonas, respectively.
The observed high frequency of occurrence of the alt gene and its combinations with the act gene in food-borne Aeromonas caviae and A. hydrophila isolates and Aeromonas isolates that were not identified to species level concurs with earlier findings (1, 21) (Table (Table33).
The ast gene codes for a cytotonic heat-stable enterotoxin. Although this gene occurs widely in clinical (2, 3, 23) and in environmental isolates of Aeromonas spp. (2, 21), the data on the presence of the ast gene in food-borne isolates of Aeromonas spp. are lacking. A search of GenBank for the ast gene sequence using the BLAST N system showed a 96% homology with the ast gene in the complete genome of A. hydrophila subsp. hydrophila ATCC 7966 (accession number CP000462 ) (Table (Table2).2). This strain had been found to be positive for the act gene (5), and the novel primers for the alt and ast genes generated virtual amplicons from the ATCC 7966 gene sequence using Clone Manager 9 professional edition software (Scientific & Educational Software, Cary, NC). The food-borne isolate identified as the A. hydrophila subsp. hydrophila ATCC 7966 strain was isolated in the United States from milk presenting a fishy odor. We speculate that an Aeromonas isolate harboring the molecular enterotoxin profile of the A. hydrophila ATCC 7966 strain, an ast-positive strain, may induce not only an infection but also intoxication due to the presence of the heat-stable enterotoxin, even if the contaminated food vehicle was heat treated (tin of milk). Interestingly, with the exception of one isolate of A. caviae (Table (Table1),1), all reference and food-borne strains of Aeromonas in this study carrying the ast gene belonged to A. hydrophila HG1 or to the A. hydrophila complex, which always carried alt and/or act genes (Tables (Tables11 and and3).3). These results are in agreement with earlier studies (2, 21, 25) and may suggest or confirm that the ast gene might be specific for A. hydrophila HG1 and might represent a potential taxonomic tool for the identification of A. hydrophila HG1.
The low prevalence (2%) of the ast gene in food-borne Aeromonas isolates in the present study is similar to that reported in earlier studies involving human cases of diarrhea in India (23) but contrasts with the reported high prevalence of the ast gene in waterborne isolates (30%) from the United States (22), in isolates from healthy (48%) and sick (71%) children and in environment isolates (66%) from Bangladesh (3), and in clinical human isolates (13%) from southern Taiwan (25). The common occurrence of the ast gene in A. caviae, A. veronii biotype sobria, and A. trota and the single carriage of this gene in some human isolates from Bangladesh (3) may be due to the method used (16) or to the geographic difference in virulence gene carriage, as observed previously (5).
Another unexpected and novel finding of this study is the detection and amplification of a DNA amplicon of 148 bp that we named AssHPA from 11 reference strains of Aeromonas isolated from fish and feces in Switzerland and France (Table (Table22 and Fig. Fig.2).2). One of these strains, ATCC 14715, was isolated from a juvenile silver salmon (Oncorhynchus kisutch) intestine in France and was identified as A. bestiarum. An additional nine strains of A. salmonicida labeled as UniBern strains (Table (Table2)2) were isolated from moribund fish at the Institute of Veterinary Bacteriology, University of Bern, Bern, Switzerland, and identified as A. salmonicida, and the remaining one (LMG 13461) was isolated from human feces in Switzerland (Table (Table2).2). Molecular characterization of AssHPA indicates clearly that it is amplified from a gene coding for a hypothetical protein; further characterization of this protein is required. The existence of AssHPA and its occurrences in a large number of clinical fish isolates could be exploited as a taxonomical and specific tool for the detection of A. salmonicida subsp. salmonicida.
In conclusion, we have developed a multiplex PCR method for the detection of three enterotoxin genes in Aeromonas spp. The predicted amplicons produced by uniplex PCR (Fig. (Fig.1)1) were confirmed by the multiplex PCR (Fig. (Fig.2).2). The data indicate that some enterotoxin genes, for example, alt in A. caviae HG4 and A. caviae/A. media HG5 and ast in A. hydrophila HG1, may be species or HG group specific. Also, the potential of using AssHPA amplified mostly from fish isolates as a specific detection tool for A. salmonicida subsp. salmonicida is proposed.
We thank Catherine Carrillo, John Austin, Franco Pagotto, Ashton Hughes, and Sarah McIlwham of Health Canada, Bureau of Microbial Hazards, Food Directorate, Health Products and Food Branch, Ottawa, Ontario, Canada, for reviewing the manuscript. We thank also Makonnen Abebe of Health Canada, Healthy Environments and Consumer Safety Branch, Ottawa, Ontario, Canada; Enrico Bonaventura of Canadian Food Inspection Agency, Ottawa, Ontario, Canada; and Marion Lyœn, Claire Perrin, Audrey Massip, and Marjorie Marechal of the University of St. Etienne, Institute of Technology, Biological and Environmental Engineering Department, France, for their valuable technical contributions to this work.
Geert Huys is a postdoctoral fellow of the Fund for Scientific Research-Flanders (Belgium) (F.W.O.-Vlaanderen).
Published ahead of print on 20 November 2009.