Aquariums as Reservoirs for Multidrug-resistant Salmonella Paratyphi B

Multidrug-resistant Salmonella enterica serovar Paratyphi B dT+ isolates from patients with gastroenteritis were identical with isolates from their home aquariums. Matched isolates had identical phage types, XbaI and IS200 profiles, and Salmonella genomic island 1 (SGI1). Ornamental fish tanks are reservoirs for SGI1-containing S. Paratyphi B dT+.

in different countries are clonally related is also not known.
Although a few epidemiologic studies suggest that antimicrobial drug-susceptible S. Paratyphi B dT+ may be linked to aquacultural practices (9,10), no molecular data confirm this. However, the first reported SGI1-containing S. Paratyphi B dT+ isolate with drug-resistance phenotype ApCmSmSpSuTc was isolated in 1997 from a tropical fish in Singapore (11), raising the possibility that tropical fish and aquariums are a reservoir. The aim of this study was to determine if domestic aquariums are reservoirs for SGI1containing, multidrug-resistant S. Paratyphi B dT+ that infect humans.

The Study
S. Paratyphi B dT+ with the resistance phenotype ApCmSmSpSuTc had been isolated sporadically in various states of Australia since 1997, and initial surveys showed a potential association with ownership of home aquariums (D. Lightfoot, unpub. data). In 2000, multidrugresistant S. Paratyphi B dT+ with an identical phage type (reaction does not conform [RDNC]), designated here as Aus2, and the same drug-resistance profile (ApCmSm SpSuTc) was isolated from humans with gastroenteritis and from fish tanks in the homes of 2 infected patients (   (Table). One isolate (SRC50) characterized previously (7) was used as a control (Table).
To determine if the resistance phenotype of these strains was due to SGI1 (4,5,7,8,11), polymerase chain reaction (PCR) with primer pairs shown in Figure 1 was used as previously described (7). The left and right junctions of SGI1 with the chromosome and of In104 with SGI1 were present in all cases. Regions containing the gene cassettes were amplified by using standard primers (L1 and R1) in the 5′-and 3′-conserved segments of class 1 integrons. Fragments of 1.0 and 1.2 kb were amplified from all isolates, and digestion of these amplicons with RsaI generated a profile (data not shown) that was indistinguishable from the pattern for the 2 amplicons containing the aadA2 and blaP1 cassettes found in In104 and S. Paratyphi B dT+ isolates SRC49 and SRC50 from 2001 (7). The aadA2 gene cassette was linked to SO26 in the SGI1 backbone, which indicates that it is on the left, as in In104, and the expected 1.8-kb PCR fragment was generated by using primers in groEL and blaP1 (Figure 1), which places the blaP1 cassette on the right. Southern hybridization of XbaI-digested whole-cell DNA with a probe for the floR gene as described previously (7) identified a band of ≈12 kb, which is consistent with an SGI1 structure identical to that reported previously (7,8,11) and the groEL-blaP1 amplicon linked this 12-kb XbaI fragment with the adjacent 4.3-kb XbaI fragment (Figure 1).
To obtain further evidence for the identity of the matched human and fish tank isolates, macrorestriction analyses of XbaI-digested whole-cell DNA by pulsed-field gel electrophoresis (PFGE) were performed as previously described (12). Several studies (3-6,13) suggest that S. Paratyphi B dT+ isolates possess considerable genetic heterogeneity. However, the SGI1-containing isolates appear to be homogeneous. The band patterns for all SGI1-containing S. Paratyphi B dT+ were identical from humans and fish tanks with phage type RDNC Aus3 ( Figure 2A) and Aus2 (data not shown). IS200 profiles were also analyzed by hybridization of an IS200 probe with Pst I-digested whole-cell DNA as described elsewhere (6). Again, all strains showed identical profiles ( Figure 2B and data not shown) that differed by 1 band from profile IP1 recently described (6). Thus, matched isolates from humans and their fish tanks were indistinguishable from each other.
An unusual observation in this study was that isolates with different phage types showed identical PFGE and IS200 profiles, indicating that they represented a clonal cluster. The control strain SRC50 (RDNC) also displayed the same patterns, demonstrating that it also is a member of the same clone. Thus, variation in phage type (Table) appears to have occurred within a single clone. Variation in phage type has also been reported in other studies of multidrug-resistant S. Paratyphi B dT+ strains (4-6), although a number of related but slightly different XbaI PFGE patterns were observed in those studies. This finding suggests that all multidrug-resistant S. Paratyphi B dT+ found globally have a single origin, but that variations, possibly because of acquisition of other temperate phages or plasmids, have arisen over time. However, direct comparisons of strains from different countries will be needed to confirm this hypothesis.

Conclusions
This is the first definitive report showing that ornamental fish tanks are a reservoir for multidrug-resistant S. Paratyphi B dT+ (ApCmSmSpSuTc phenotype) containing SGI1 that causes severe disease in humans, particularly young children. In addition to containing SGI1, the matched isolates from humans and their fish tanks had the same phage type and the same XbaI macrorestriction digest pattern and IS200 profile. These findings identify home aquariums containing tropical fish as the most important, although not necessarily the only, source of multidrug-resistant S. Paratyphi B dT+. The fact that 12%-14% of Australian households have ornamental fish (14) and as many as 12 million American and 1 million Canadian families own domestic aquariums (9), together with the young age of most affected patients, indicate that multidrug-resistant S. Paratyphi B dT+ in home aquariums is a risk factor for Salmonella infection and thus becomes a public health issue. Figure 2. Pulsed-field gel electrophoresis (PFGE) and IS200 profiles of Salmonella enterica serovar Paratyphi B dT+ isolates positive for Salmonella genomic island 1. A) PFGE profiles. XbaIdigested whole-cell DNA was separated by PFGE as previously described (12). Molecular mass markers (lane M) are low-range PFGE markers (New England BioLabs, Beverly, MA, USA) composed of concatamers of bacteriophage lambda DNA. The band absent in lane 147 was present in other runs. B) IS200 profiles. PstI digests of whole-cell DNA were separated and hybridized with an IS200 digoxigenin (DIG)-labeled probe. Molecular mass markers (lane M) are DIG-labeled bacteriophage lambda DNA digested with HindIII (Roche Diagnostics, Castle Hill, New South Wales, Australia). Primers and polymerase chain reaction conditions used to generate the IS200 probe have been previously described (6).