Comparative analysis of extended-spectrum- -lactamase-carrying plasmids from different members of Enterobacteriaceae isolated from poultry, pigs and humans: evidence for a shared -lactam resistance gene pool?

Sir, β-Lactam antibiotics are extensively used in human and veterinary medicine. The detection rate of extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae isolated from humans has increased rapidly worldwide.1 In addition, ESBLs have been increasingly described in bacterial populations circulating in animals.2,3 Recently, a high diversity of ESBLs in Escherichia coli was reported in Belgian poultry farms. In that instance, CTX-M enzymes were the predominant ESBL family.4 CTX-M-2-producing Salmonella enterica serovar Virchow strains and TEM-52-producing S. enterica serovar Infantis strains have also been isolated from Belgian poultry.2,3 This raises a potential public health concern. Moreover, the presence of ESBLs in the microbiota of food-producing animals may pose a human health hazard since these bacteria may represent a reservoir of resistance genes for pathogens causing disease in humans and animals.4 Therefore, to demonstrate whether a common ESBL gene pool exists among isolates in different hosts, we characterized the plasmids and determined the location and transfer possibilities of the ESBLs blaTEM-52, blaCTX-M-2 and blaCTX-M-15 that were present in different members of Enterobacteriaceae isolated from humans, broilers and pigs. Fourteen blaTEM-52-, blaCTX-M-2- or blaCTX-M-15-carrying clonally unrelated strains were used in this study (Table 1). These strains were isolated in Belgium from humans, pigs and broilers. The human E. coli strains were isolated from patients hospitalized at the Ghent University Hospital. All isolates from poultry were obtained from the faeces of healthy broilers. The E. coli and Klebsiella pneumoniae isolates were obtained as described previously.4 The S. enterica isolates from poultry were collected in the framework of mandatory Salmonella monitoring programmes in Belgium. The two porcine E. coli isolates originated from pigs with diarrhoea. The ESBL gene of each isolate was characterized as described previously by isoelectric focusing, PCR and sequencing.4 Plasmid transfer experiments were carried out as described previously.2 The antimicrobial susceptibility of the parental strains and their E. coli transconjugants was determined by the Kirby–Bauer disc diffusion test (Neo-Sensitabs, Rosco Diagnostica, Taastrup, Denmark) (Table 1).4 For the parental strains and their E. coli transconjugants, plasmid profiles were determined and the size of each ESBL-carrying plasmid was estimated.2 The incompatibility (Inc) group of each ESBL-carrying plasmid was defined by the PCR-based replicon typing method.5 Restriction fragment length polymorphism (RFLP) fingerprint analysis and Southern blot hybridization were performed as described previously.2 Characteristics of the parental strains and the ESBL-carrying plasmids analysed in this study aAll strains were isolated from faeces, with the exception of BUT 10 that was isolated from the human throat. The RFLP fingerprint pattern of the ESBL-carrying plasmid of the 10101-1 isolate and the 142-1 isolate have been described previously2,3 and were taken into account for RFLP fingerprint analysis to compare with the ESBL-carrying plasmids of the isolates used in this study. bAntimicrobial drugs used were the following: chloramphenicol (CHL), enrofloxacin (ENR), florfenicol (FFC), gentamicin (GEN), kanamycin (KAN), nalidixic acid (NAL), neomycin (NEO), tetracycline (TET), trimethoprim (TMP), streptomycin (STR), sulphonamides (SULPH). Characteristics of the parental strains and the ESBL-carrying plasmids analysed in this study aAll strains were isolated from faeces, with the exception of BUT 10 that was isolated from the human throat. The RFLP fingerprint pattern of the ESBL-carrying plasmid of the 10101-1 isolate and the 142-1 isolate have been described previously2,3 and were taken into account for RFLP fingerprint analysis to compare with the ESBL-carrying plasmids of the isolates used in this study. bAntimicrobial drugs used were the following: chloramphenicol (CHL), enrofloxacin (ENR), florfenicol (FFC), gentamicin (GEN), kanamycin (KAN), nalidixic acid (NAL), neomycin (NEO), tetracycline (TET), trimethoprim (TMP), streptomycin (STR), sulphonamides (SULPH). In order to better understand the spread and persistence of mobile β-lactam resistance plasmids among different members of Enterobacteriaceae isolated from different reservoirs, a closer look at the pool of conjugative plasmids was appropriate and timely. All isolates tested here contained high-molecular-weight ESBL-carrying plasmids (∼150 kb) and, for all these isolates, E. coli transconjugants were obtained. The blaCTX-M-2-, blaTEM-52- and blaCTX-M-15-carrying plasmids belonged to IncHI2, IncI1 and IncI1, respectively (Table 1), as has already been demonstrated in previous reports.2,3,6 RFLP analysis of plasmid DNA from the transconjugants revealed, in most cases, closely related fingerprints for plasmids carrying the same ESBL gene. All blaTEM-52-carrying plasmids showed the same fingerprint pattern analysis, suggesting that this is a rather stable plasmid circulating in different members of the Enterobacteriaceae, present in different animal reservoirs and in humans. Southern blot hybridization with a blaTEM-52 probe revealed two PstI fragments of 2.75 and 2.9 kb, as has already been shown in a previous report.3 The spread of a blaCTX-M-2-carrying multiresistant plasmid among E. coli and S. enterica isolates from pigs and broilers was demonstrated. Only the plasmid from the human E. coli isolate differed in RFLP fingerprint pattern from the other blaCTX-M-2-carrying plasmids. Southern blot hybridization with a blaCTX-M-2 probe revealed a >10 kb EcoRI fragment in the plasmids from the porcine and broiler E. coli isolates and from the Salmonella Virchow CODA-1 isolate. Two EcoRI fragments of 6 and 10 kb were found in the plasmid from the human E. coli isolate and in the plasmid from the Salmonella Virchow 142-1 isolate.2 For the blaCTX-M-15-carrying plasmids, results of RFLP analysis were identical except for the plasmid from the human E. coli strain. For the animal strains, Southern blot hybridization with a blaCTX-M-15 probe revealed two EcoRI fragments of 6.5 and 7 kb and one 5 kb PstI fragment. The plasmid from the human E. coli strain showed 5 and >10 kb EcoRI fragments and a 5.5 kb PstI fragment. The Southern blot results suggest the existence of two copies of the tested ESBLs on their ∼150 kb plasmids. The differences seen in the RFLP analyses for the blaCTX-M-2- and blaCTX-M-15-carrying plasmids may possibly reflect the rapid evolution of these plasmids as they were exposed to different environmental stresses. The human, porcine and poultry environments may be experienced by bacteria in different ways. In summary, ESBL resistance plasmids appear to move readily between different microorganisms and different ecosystems. TEM-52-carrying cephalosporin-resistant organisms may have been transmitted from food animals to humans, or vice versa.3 For the CTX-M-2- or CTX-M-15-carrying cephalosporin-resistant organisms, exchange between food animals and humans, however, remains unclear, mainly due to the unknown plasticity and evolutionary speed of the plasmids carrying them. This work was supported by a grant from the Federal Public Service of Health, Food Chain Safety and Environment (grant number RT 06/3 ABRISK). None to declare. We are grateful to A. Carattoli for providing the plasmid incompatibility group controls. We thank Hanne Vereecke and Danielle Vandergheynst for their skilled technical assistance.