Detection of Plasmid-Mediated Colistin Resistance,mcr-1 and mcr-2 Genes,in Salmonella spp. Isolated from Food at Retail in Belgium from 2012 to 2015

Until recently it was believed that the only mechanism of resistance to colistin was chromosomally mediated, therefore, not transferable from cell to cell and restricted to clonal transmission. However in 2015, a new mechanism of resistance, referred as plasmid mediated, was described in China in Escherichia coli isolates from chickens and the mcr-1 gene was assigned as the responsible gene conferring resistance (Liu et al., 2016). Latter the mcr-2 gene that has 76.7% nucleotide identity to mcr-1 was reported in Belgium porcine E. coli (Xavier et al., 2016). The mechanism of resistance of the mcr genes is a phosphatidylethanolamine transferase. The enzyme transfers a phosphoethanolamine residue to lipid A of the cell membrane, the initial site of action of colistin. The altered lipid A has much lower affinity for colistin and related polymyxins resulting in reduced activity of the antimicrobial (Hinchliffe et al., 2017). Plasmid-mediated colistin resistance gene mcr-1 has been identified and reported worldwide mostly in E. coli isolated from sick and healthy animals, food, the environment, and humans (Skov and Monnet, 2016), leading to the recognition of its global distribution and spread. In Belgium, mcr-1 has been detected in E. coli from sick pigs and calves (Malhotra-Kumar et al., 2016), and recently a new gene has been reported, mcr-2, which exhibits a higher prevalence in porcine colistin-resistant E. coli in this country (Xavier et al., 2016). The detection of mcr-1 in Salmonella has been reported in chicken farms in France (Skov and Monnet, 2016), in pigs in Japan (Skov and Monnet, 2016), and from a pig farm in Great Britain (Anjum et al., 2016); however, mcr-2 has never been reported before in any country in Salmonella. Salmonella carrying either the mcr-1 or mcr-2 gene in livestock in Belgium is so far not documented. The aim of this study is to determine the presence of the plasmid-mediated colistin resistance in Salmonella spp. isolated from food in Belgium.

According to the national and European antimicrobial surveillance programs, antimicrobial susceptibility testing on Salmonella isolated from food is performed by broth microdilution according to ISO 20776-1:2006 using EUVSEC and EUVSEC2 sensititre plates (Trek Diagnostic Systems; Thermo Scientific). Minimum inhibitory concentration (MIC, mg/L) was determined and interpreted according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) epidemiological cutoff values (ECOFF) as described in the Commission Implementing Decision 2013/652/EU (European Commission, 2013).

Isolates with a MIC of colistin ≥2 mg/L were selected for screening by PCR for the presence of mcr-1 and mcr-2 genes, using a protocol previously described (Liu et al., 2016; Xavier et al., 2016). In 2012, 32 out of 398 (8.04%) Salmonella isolates were found to be colistin resistant, 18 out of 296 in 2013 (6.08%), 38 out of 294 (12.92%) in 2014, and 17 out of 427 (3.74%) in 2015. In total, 105 Salmonella isolates belonging to 13 different serovars were isolated [Autoagglutinable (1), Derby (4), Enteritidis (80), Give (1), Idikan (1), Infantis (1) Livingstone (1), Monophasic Typhimurium (7), Paratyphi B var L(+) tartrate (+) (2), Rissens (1), Species (1), Typhimurium (2), and Typhimurium var Copenhagen (3)]. Isolates recovered mainly from poultry (n = 54) and pork (n = 18), but also from cut poultry meat (n = 10), beef minced meat (n = 6), poultry meat preparation (n = 4), and other food matrices (n = 13), were subjected to PCR.

Of the total of 105 isolates screened by PCR, 2 tested positive for mcr-1 and 1 tested positive for mcr-2. Interestingly, the presence of mcr-1 and mcr-2 was found in pork carcasses in 2012, the single year in which both genes were detected. The presence of the genes in all three Salmonella isolates has been confirmed by whole genome sequencing (WGS). Genomic DNA was extracted using the genomic Tip 20/G (Qiagen) following the manufacturer's instructions. DNA was quantified using a Qubit 3.0 fluorometer (ThermoFisher Scientific). Sequencing libraries were prepared with a Nextera XT DNA sample preparation kit (Illumina) and sequenced on an Illumina MiSeq instrument (Illumina) with a 300-bp paired-end protocol (MiSEv3 chemistry) according to the manufacturer's instructions. Raw sequence reads were trimmed by using the Trimmomatic v0.32 with the ‘SLIDINGWINDOW:4:20’ option (Bolger et al., 2014) and assembled using the SPAdes v3.8 (Bankevich et al., 2012) (www.genomicepidemiology.org). The number of contigs produced is described in Table 1. Acquired antibiotic resistance-encoding gene prediction was done in silico using ResFinder 2.1 (https://cge.cbs.dtu.dk/services/ResFinder) by performing a blast of the genomes assembled with SPAdes to a curated database, using a percentage identity of 98% and minimum length of 60% (Zankari et al., 2012). The presence of plasmid was identified by PlasmidFinder 1.3 (https://cge.cbs.dtu.dk/services/PlasmidFinder) (Carattoli et al., 2014), and the sequence type (ST) of Salmonella isolates was predicted by using MLST (multilocus sequence typing) 1.8 service (https://cge.cbs.dtu.dk/services/MLST) with default settings available.

In this study, we have detected and confirmed the presence of the mcr-1 gene in two Salmonella isolates belonging to different STs and isolated in different years from meat (pork and poultry origin) in the Belgian food chain, and report for the first time, to the best of our knowledge, the presence of mcr-2 in Salmonella species isolated from retail meat. In the screening analysis, the plasmid-related genes were found in 3 out of 105 (2.85%) colistin-resistant Salmonella isolated from 2012 to 2015. Results from the WGS confirmed the presence of plasmid-mediated colistin-resistant mechanism (Table 1). WGS analysis found that both genes, mcr-1 and mcr-2, showed 100% sequence similarity with those found in the E. coli strain SHP45 (Genbank accession no. KP347127) (Liu et al., 2016) and the porcine E. coli strain (Genbank accession no. LT598652.1) (Xavier et al., 2016), respectively. Genes encoding for acquired antimicrobial resistance were predicted by ResFinder 2.1. Isolates harbored a coresistant profile, however, they did not harbor any mechanisms of resistance to third-generation cephalosporins nor to carbapenems, which is in concordance with earlier findings in Salmonella from chicken and turkey meat (Veldman et al., 2016) and in E. coli isolates from livestock in Belgium (Malhotra-Kumar et al., 2016). The presence of plasmid of type IncX4 was identified by PlasmidFinder in all three Salmonella isolates. We screened for the absence of the mobile element ISApl1 upstream of the mcr-1 gene, which has been reported not to be present in mcr-1 carrying plasmid IncX4 (Veldman et al., 2016). PCR conditions and primers were described elsewhere (Liakopoulos et al., 2016). The absence of this element was also confirmed by IS finder (https://www-is.biotoul.fr/blast.php) by using the assemblies generated by WGS.

Other plasmid replicons [IncQ1, IncI1, Col (BS512), and ColpVC] not previously associated with the plasmid-mediated colistin resistance were also identified by the bioinformatics analysis. To provide further evidence that the mcr-1 and mcr-2 genes were located in the plasmid replicon type IncX4, the contigs containing the genes mcr-1 and mcr-2, respectively, were compared against the NCBI database using the BLAST algorithm blastn. Although the lineage of the plasmid is not annotated in NCBI, results show that the genes were located in the same contig as the plasmid. BLASTn analysis found a best matching of 100% of query coverage and maximal 99% identity with previous sequenced plasmids IncX4 type harboring mcr-1, namely pHNSHp45 (GenBank accession no. KP347127), and mcr-2, namely pKP37-BE (GenBank accession no. LT598652.1).

In conclusion, this is the first report of detection of mcr-1 in Salmonella isolated from the food chain in Belgium at a slaughterhouse line and in retail in two unrelated matrices, that is, pork carcasses and poultry meat, and the first report of plasmid-mediated mcr-2 gene conferring colistin resistance in Salmonella species isolated at retail. To the best of our knowledge, these genes have not yet been detected in Belgium in colistin-resistant Salmonella from livestock, pigs, or chicken and the mcr-2 gene has not even been reported in other countries. It seems that this gene is less transferable and is confined to Belgium. Whether the origin of the genes has been acquired through plasmid transfer from other reservoirs of mcr-1 or mcr-2 as suggested previously (Xavier et al., 2016), or resulted from human contamination during handling of the meat, is currently unknown. Although its presence is sporadic and low, continuous monitoring of plasmid-mediated colistin resistance in food is desirable to survey the potential risk to humans.