Characterization of genetic determinants of extended-spectrum cephalosporinases (ESCs) in Escherichia coli isolates from Danish and imported poultry meat

Sir, The predominant cause of resistance towards cephalosporins in Escherichia coli is production of plasmid-encoded extended-spectrum β-lactamases (ESBLs) and AmpC-type β-lactamases, also referred to as extended-spectrum cephalosporinases (ESCs). Most studies have focused on description of ESCs in bacteria from humans as well as production animals, while fewer studies have focused on the prevalence of ESCs in bacteria from imported food products. In 2006, 1650 E. coli isolates were sampled from Danish and imported (15 different countries) retail broiler chicken (68.7%) and broiler turkey meat (31.3%) all across Denmark. In total, 1.3% (n = 21; Table 1) had ceftiofur MICs of ≥8 mg/L. This breakpoint was adopted from the CLSI guidelines for bacteria isolated from animals.1 Although this breakpoint refers to testing of E. coli isolated from bovine mastitis, it seems reasonable to assume that the breakpoint can be applied to E. coli from food products in general. Based on this, all the 21 isolates were considered resistant. These isolates were derived from chicken meat from Germany (n = 19; 3.5% of the 546 German isolates), France (n = 1; 0.3% of the 299 French isolates) and Brazil (n = 1; 16.7% of the 6 Brazilian isolates). All 19 German isolates came from the same slaughterhouse (36.6% of the 546 German isolates). No ceftiofur-resistant isolates were detected among the 543 Danish isolates. Phenotypic and genotypic characteristics of the 21 ceftiofur-resistant isolates BR, Brazil; DE, Germany; F, France; ampC-up; up-regulation of the ampC promoter. Phenotypic and genotypic characteristics of the 21 ceftiofur-resistant isolates BR, Brazil; DE, Germany; F, France; ampC-up; up-regulation of the ampC promoter. Nineteen of the 21 isolates produced distinct XbaI patterns in PFGE. The two isolates showing indistinguishable PFGE profiles were isolated from different chicken meat samples on the same date at the same regional laboratory. PCR amplification and sequencing determined a blaCMY-2 gene in 12 isolates, a blaCTX-M-1 gene in 7 isolates and a blaTEM-52 gene in 1 of the isolates. Eight isolates carried a blaTEM-1b gene in addition to the blaCMY-2 or the blaCTX-M-1 gene. Amplification of the blaampC promoter showed the most frequently described promoter mutations resulting in hyperproduction of the blaampC gene2 in the only isolate (isolate 7825 from Brazil) where no β-lactamase genes except a blaTEM-1b gene could be identified. Plasmids carrying the detected β-lactamase genes were transferred to a competent, plasmid-free E. coli recipient (Genehogs®, Invitrogen) by electroporation. The presence of single plasmids in transformants was verified by S1 nuclease digestion3 and subsequent PFGE analysis. The 12 plasmids harbouring a blaCMY-2 gene produced four different EcoRV restriction fragment length polymorphism (RFLP) profiles (designated A1, A2, B and C; Table 1) with A1 and A2 differing by only three fragments. The two isolates with indistinguishable PFGE profiles contained plasmids that exhibited RFLP patterns of type C (Table 1). PCR-based replicon typing4 demonstrated that the plasmids of type A1 and A2 belonged to the IncK group while types B and C belonged to the IncI1 group. Plate-mating experiments with transformants as donors and plasmid-free E. coli MT102RN and Salmonella Dublin JEO66RN as recipients demonstrated plasmids of types A and C (blaCMY-2) to be successfully transferred by conjugation to the E. coli recipient. No transconjugants were obtained using transformants carrying the type B plasmid (blaCMY-2) as donors. Six of the seven plasmids encoding blaCTX-M-1 had indistinguishable EcoRV RFLP profiles (D1) and were from strains originating from Germany. The seventh plasmid (isolate 5405, D2), originating from a French poultry meat sample, had several fragments in common with the other six German plasmids. All plasmids belonged to the IncI1 group, had a size of ∼105 kb and were able to transfer by conjugation to the E. coli recipient only. Comparison of the seven plasmids with a previously published blaCTX-M-1-encoding plasmid from an E. coli (TK-3) isolated from healthy poultry in France,5 revealed that the plasmid present in isolate 5405 had an RFLP profile indistinguishable from that of the plasmid isolated from TK-3. The blaTEM-52-harbouring plasmid (RFLP type E) was ∼45 kb, belonged to the IncX1 group and successfully transferred to both recipients by conjugation. The observed differences in prevalence of β-lactamase genes among broiler meat type, countries as well as slaughterhouses within Germany in this study suggest a very uneven distribution of these genes in poultry products sold in Denmark in 2006. None of the ceftiofur-resistant E. coli isolates originated from turkey meat, suggesting a lower prevalence of these in turkey meat sold in Denmark compared with chicken meat. The fact that no ceftiofur-resistant isolates were isolated from Danish broiler meat indicates a very low prevalence of ceftiofur resistance among isolates from Danish poultry. The higher prevalence of resistant bacteria in imported meat is of special concern taking into consideration the increasing proportion of imported meat sold in retail stores.6 Since only two of the isolates were clonally related and different β-lactamases were detected among the isolates, bacterial contamination by a ‘house strain’ at the German slaughterhouse seems less likely. Nevertheless, the presence of a pool of clonally unrelated E. coli carrying different ESC resistance genes within the slaughterhouse cannot be entirely excluded. The fact that the plasmid in the TK-3 isolate from healthy poultry in France5 was indistinguishable from the blaCTX-M-1-carrying plasmid isolated from the French isolate 5405 demonstrates that poultry and poultry products can share E. coli strains with indistinguishable resistance plasmids and indicates a transfer of such strains from animal to meat products, e.g. during slaughtering. The characterization and comparison of plasmids in this study support the theory of epidemic spread of plasmids being the primary mechanism for dissemination of β-lactamase genes and the existence of a gene reservoir in food products, especially imported food products, that potentially can be transferred via the food chain to humans. This work was supported by grant 274-05-0117 from the Danish Research Agency. None to declare. We thank Patrice Nordmann for providing strain TK-3 and Berith Kummerfeldt for excellent help.