Description of anerm(B)-carryingCampylobacter coliisolate in Europe

Sir, Campylobacter jejuni and Campylobacter coli are the principal cause of foodborne zoonoses in humans. Campylobacteriosis is caused by C. jejuni and C. coli, and the infection is produced by ingesting contaminated food. The main reservoir is the intestinal microbiota of birds, especially poultry. Although the treatment for campylobacteriosis is generally oral rehydration therapy, the drugs of choice are macrolides, such as erythromycin. Thus, macrolide resistance poses a serious public health threat. Bacteria carrying the erm(B) gene are resistant to macrolides, lincosamide and streptogramin B (MLSB phenotype) via methylation of the 23S rRNA gene. The erm(B) gene has been described in highly erythromycin-resistant Campylobacter isolates in China, located in an MDR genomic island (MDRGI). Transference to erythromycin-susceptible C. jejuni by transformation has been observed. The aims of our study were to detect erm(B) in erythromycin-resistant isolates of our collection and to compare these findings with the presence of other known mechanisms of erythromycin resistance (target mutations of the 23S rRNA gene and ribosomal proteins L4 and L22, and antibiotic efflux pumps). Of 555 isolates obtained from food animals in Spain during 2008–11, 74 isolates with an erythromycin MIC ≥32 mg/L (broth microdilution by UNE-EN ISO20776 –1) were randomly selected (broiler: n1⁄429; fattening pigs: n1⁄434; young cattle: n1⁄411). Additionally, 14 Campylobacter isolates from urban effluents (MIC ≥32 mg/L) were also selected. Isolates were not epidemiologically related. Agar dilution susceptibility testing (CLSI methodology) was performed to increase the antimicrobial range (0.25– 1024 mg/L). The assay was carried out using Mueller –Hinton 2 agar supplemented with erythromycin and 5% lysed horse blood. Isolates with the highest level of resistance (MIC ≥1024 mg/L) were tested to detect the erm(B) gene as well as target mutations in the 23S gene, target modifications in ribosomal proteins L4 and L22 and the effect of PABN as an efflux pump inhibitor. – 9 WGS (Ion Torrent PGM) was performed to determine the location of the methylase gene erm(B) and its genetic environment. Raw data were aligned to C. coli RM4661 (NZ_CP007181) using CLC Genomics Workbench 7.5.1. The BLAST application was used to study similarities. Five isolates of 88 revealed a high level of resistance to erythromycin (MIC ≥1024 mg/L), all of them C. coli isolated from broilers. Only the C. coli ZTA09/02204 isolate was positive for the erm(B) gene; additional erythromycin-resistance mechanisms were not detected. The C. coli ZTA09/02204 isolate presented resistance to nalidixic acid, ciprofloxacin, tetracycline and streptomycin, and was susceptible to gentamicin (data not shown). The MLSB phenotype could not be confirmed, as lincosamide and streptogramin B were not included in our panel. No effect of PABN was observed on the MIC for isolate ZTA09/ 02204, which suggests that the active efflux mediated by CmeABC makes no contribution to the phenotype. Although overexpression of cmeABC could have led to the results, previous studies describe no synergy between methylase erm(B) gene and efflux pumps. The erm(B) gene is 738 bp and shows 100% nucleotide identity to erm(B) of Enterococcus faecium e389 (JN899587), Clostridium difficile 200785596 (FN665654) and Streptococcus suis 2 – 22 (EU047808). When the erm(B) gene from this study is compared with the erm(B) gene described in Campylobacter, the nucleotide identity is 99% (G299A). The erm(B) gene is located in the chromosome of C. coli ZTA09/ 02204 within a gene cluster with other antimicrobial resistance genes constituting an MDRGI (Figure 1). The MDRGI (GenBank accession no. KT953380) is 11769 bp, and it was inserted in one hypothetical protein (YSS_00750 gene). The average guanine and cytosine content is 38.5%, higher than the guanine and cytosine content in the genome (31.1%). The MDRGI contains 12 ORFs, five of them antibiotic resistance genes (Table S1, available as Supplementary data at JAC Online). The MDRGI of C. coli ZTA09/ 02204 seems to be a new type, as the organization of genes and its location in the chromosome differ from those in other types. Therefore, the MDRGI of this study might be designated type VIII. When the MDRGI C. coli ZTA09/02204 is compared with those already described, its similarity is higher to types V, VI and VII, as they present a complete tet(O) coding region. However, our isolate contains a full-length pnp gene, in contrast to the truncated forms carried by types V to VII (Figure 1). The different structures of the ZTA09/02204 MDRGI and the MDRGI previously published likely reflect a different origin. According to the guanine and cytosine percent content, the MDRGI have been divided into three different regions (Figure 1 and Table S1). Region B had 99% identity with Eggerthella sp. YY918 (AP012211) and Bacteroides uniformis WH207 (AY345595). Regions A and C presented 99% identity with plasmid pN29710– 1 from C. coli CVM N29710 (CP004067) (Figure 1), which would suggest a plasmid as the insertion vehicle of the erm(B) gene