Characterization of bla_TEM-52-Carrying Plasmids of Extended-Spectrum-β-Lactamase-Producing Salmonella enterica Isolates from Chicken Meat with a Common Supplier in Japan

TEXT

Salmonella is a zoonotic pathogen and one of the leading causes of food-borne diseases worldwide. Extended-spectrum cephalosporins can be a choice for the treatment of severe salmonellosis, especially in children and the elderly. Recently, Salmonella isolates producing extended-spectrum β-lactamases (ESBLs) have been identified in livestock and foods (1–3). Poultry are regarded as a major source of ESBL producers among livestock. A variety of bla genes, bla_CTX-M, bla_TEM, and bla_SHV, have been identified so far in Salmonella isolates from poultry and also from humans (2–4). We conducted a survey on ESBL-producing salmonellae in retail chicken meat in Yokohama City, Japan, from 2004 to 2012. Here, we report the results of the survey to finally identify 38-kb plasmids harboring bla_TEM-52, which were unexpectedly identical with a plasmid that originated from Denmark (5).

A total of 524 Salmonella isolates from 700 retail chicken meat specimens were used in this study.

The screening of ESBL-producing Salmonella isolates was done by a disk diffusion method. Disks of cephalothin (30 μg) and cefotaxime (30 μg) (BD Diagnostics) were used. Escherichia coli strain ATCC 25922 was used as a quality control strain. Interpretation was performed according to the Clinical and Laboratory Standards Institute guidelines (6). Isolates that were resistant to cephalothin and intermediate or resistant to cefotaxime were chosen for further analyses. MICs were determined by Etest (bioMérieux). The antimicrobials used were ampicillin, cephalothin, cefuroxime, cefotaxime, ceftriaxone, ceftazidime, cefoperazone, cefoxitin, and aztreonam. The detection of ESBLs was performed using the Etest ESBL (cefotaxime-clavulanic acid and ceftazidime-clavulanic acid). The bla_SHV, bla_TEM, bla_CTX-M, bla_CTX-M-2, and bla_AmpC genes were screened by PCR, according to a previous study (7). The sequences of the resulting PCR products were determined by the Sanger method using the BigDye Terminator cycle sequencing kit (Applied Biosystems). Pulsed-field gel electrophoresis (PFGE) was performed according to the PulseNet protocol using Salmonella enterica serovar Braenderup strain H9812 as a standard (8). Plasmids were extracted using the Wizard Plus SV minipreps DNA purification system (Promega) and then subjected to transformation of E. coli strain DH10B cells (Invitrogen). The entire sequences of plasmids extracted from one isolate each of S. enterica serovars Infantis and Manhattan were analyzed by next-generation sequencing (NGS). The total DNA of the transformant was separated in a PFGE gel, and the plasmid DNA was extracted from a corresponding band. A DNA library was constructed using the Nextera XT kit and analyzed by the MiSeq sequencer (Illumina). The resulting data were assembled with ABySS and the CLC bio Genomics Workbench and then confirmed by read mapping.

Five hundred twenty-four isolates from 700 specimens were screened by a disk diffusion method using cefotaxime and cephalothin, and 24 were chosen as ESBL-producing candidates. The ESBL Etest confirmed that they produce ESBLs (data not shown). The 24 isolates were further screened for resistance genes by PCR and direct sequencing. The results are shown in Table 1. We identified bla_TEM-52 in 10 isolates, bla_CTX-M-2 in 11 isolates, bla_TEM-20v in one isolate, and bla_CTX-M-14 in one isolate. One S. enterica serovar Heidelberg isolate was positive for bla_TEM-1, but no putative gene for ESBL has been identified so far.

All 10 isolates with bla_TEM-52 were found to have originated from specimens of the common supplier A at prefecture W between 2005 and 2012, while the others originated from different suppliers. Therefore, we focused on the 10 isolates in more detail to uncover the genetic backgrounds of the isolates and resistance profiles. The 10 isolates consisted of two serotypes, S. Infantis (n = 3) and S. Manhattan (n = 7). The PFGE analyses of the isolates revealed that the 3 S. Infantis isolates show 2 similar profiles and that all 7 S. Manhattan isolates show indistinguishable profiles (Table 2).

The resistance profiles of the 10 isolates are summarized in Table 2. All were resistant to ampicillin, cephalothin, cefuroxime, cefotaxime, ceftriaxone, ceftazidime, and cefoperazone but sensitive to cefoxitin and aztreonam. The resistances to cefotaxime and ceftazidime were inhibited by clavulanic acid, indicating ESBL production.

A plasmid with a size of 40 kb was detected in each of the 10 isolates. A transformation experiment with the E. coli DH10B strain resulted in the successful transfer of the resistance phenotype (Table 2). This indicates that the plasmids are responsible for the ESBL phenotype of the isolates (Table 2). The plasmid DNA extracted from the transformant of each isolate was subjected to restriction fragment length polymorphism (RFLP) analyses. All the plasmids displayed indistinguishable RFLP patterns generated by EcoRI, SpeI, NdeI, or NruI (data not shown). This suggests that the plasmids of the 10 isolates are quite similar to each other. Therefore, the representative plasmids from one isolate each of S. Infantis and S. Manhattan were analyzed by NGS. The size of the plasmid from the S. Manhattan isolate was 38,612 bp. This plasmid, pYM4, was completely identical to another plasmid, pDKX1. pDKX1 is an IncX plasmid, identified in 2011 in an E. coli strain from beef in Denmark, and it harbors bla_TEM-52 (GenBank accession no. JQ269336.1) (5). The plasmid from S. Infantis was also almost identical, with only one insertion of a T at nucleotide 11833 within a putative gene encoding a hypothetical protein. Another plasmid, pE001, also was found to show high similarity with pYM4, with 38,602/38,614 nucleotides found to be identical (with 10 gaps) in a BLASTn search. pE001 was identified in 2006 in an E. coli strain from broiler meat in Denmark (9). A comparison of the sequences indicated that the bla gene of pYM4 belongs to the TEM-52b allele, and the Tn3 transposon is located upstream of the bla gene. pYM4 probably belongs to the IncX1a replicon type, since the pir sequence shares 100% similarity, and the flanking region also shares high similarity with that of pE001.

In this study, we analyzed 10 isolates of ESBL-producing Salmonella obtained from retail chicken meat that originated from the common supplier A. They comprised two serovars, S. Infantis and S. Manhattan. The PFGE profiles were similar or indistinguishable within each serovar, which was consistent with them being from the common supplier A. Intriguingly, plasmids responsible for the ESBL phenotype were also common among the isolates, according to the RFLP analysis, even though the plasmids exist in different serovars. This suggests that the resistance genes were transmitted by a type of plasmid among Salmonella serovars, as described in Cloeckaert et al. (3). Genetic analyses, including NGS, revealed that the plasmid is 38 kb in size, harbors the TEM-52b allele of bla with the Tn3 transposon upstream, and belongs to a replicon type of IncX1a. bla_TEM-52 was initially identified from Klebsiella pneumoniae in France, 1996 (10). Several studies also reported bla_TEM-52–positive Salmonella from livestock, meat, and humans in European countries, suggesting the dissemination of the resistance gene (1–4, 9).

The plasmid is a major vehicle of resistance. In addition to plasmid-mediated transmission of resistance among Salmonella serovars, NGS in this study suggests a close association of the resistance plasmid with Europe because of the sequence identity between pYM4 of Japan and pDKX1 of Denmark. It is speculated that the breeding stock of broilers might be a route of dissemination, since most of the farms, including company A, import breeding stock (parental and/or grandparental stock) from foreign countries. It is also interesting that the resistances were kept in supplier A during 2005 to 2011. It is unknown whether they were maintained in the farm after the initial introduction or they had been repeatedly introduced to the farm. We need to analyze a more extensive collection of plasmids in order to evaluate the rates of mutation and relationships among the plasmids. In addition, more extensive epidemiological investigation is required to provide robust evidence for the spread of resistance.

We previously reported bla_CTX-M-14-positive S. enterica serovar Enteritidis strains isolated from chicken meat imported from China (11) and indicated the importance of imported food as a possible route of disseminating resistance. In this study, we identified resistance plasmids even from domestic chicken meat and revealed an unexpected commonness of resistance plasmids beyond countries, and we suggest an alternative route of the dissemination of resistance.