Findings

Salmonella enterica serovar Typhimurium is prevalent in many animal species [1–3] including food-producing animals that are considered to be reservoirs for human infection. S. Typhimurium was the top 5 serovar found most frequently in cases of Salmonella foodborne illness in Japan between 2006 and 2010 https://hasseidoko.mhlw.go.jp/Byogentai/Pdf/data48e.pdf. Multidrug-resistant S. Typhimurium definitive phage type 104 (DT104) causes human salmonellosis in Japan [3]. S. Typhimurium DT104 was first isolated in the late 1980 s, and has spread widely among food-producing animals across Japan [3–5]. Although a decreased proportion of DT104-related isolates among the animals was found between 2002 and 2005, multidrug-resistant S. Typhimurium remains prevalent among food-producing animals in Japan [6].

In Japan, fluoroquinolone drugs were approved in veterinary fields in 1991 and are commonly used for treatment of bacterial diseases such as enteritis and pneumonia in food-producing animals [7]. In 2001, fluoroquinolone resistance was found in S. Choleraesuis from pigs [8] and S. Typhimurium from cattle [9]. In addition, a fluoroquinolone-resistant S. Typhimurium was identified in bovine isolates in 2005 [6]. The mechanism of fluoroquinolone resistance in these isolates is the mutation of quinolone resistance-determining regions (QRDRs) in DNA gyrase and topoisomerase IV [8, 9]. In 2006, qnrS1 was identified in two S. Typhimurium isolates (including one DT104 isolate) from dairy cows and beef cattle, and S. Thompson from poultry in Japan [10]. The report identified the potential risk of foodborne infections of Salmonella conferring the gene from food-producing animals to humans in Japan.

Quinolone resistance mechanisms mediated by plasmids are responsible for target protection such as the qnr genes, active efflux such as qepA, and enzymatic modifications such as aac(6')Ib-cr [11]. The plasmid-mediated quinolone resistance (PMQR) genes contribute to a reduction of quinolone susceptibility. In Japan, qnrS was first identified in human isolates of Shigella flexneri in 2003 [12]. qepA-harboring clinical isolates of Escherichia coli were found in 2002 in Japan [13]. qnrB in Klebsiella oxytoca, Pseudomonas mirabilis, and P. fluorescens, and qnrS in E. coli and Enterobacter cloacae were found in zoo animal isolates in 2006 [14]. In addition, the presences of qnrS1 and qnrS2 in Salmonella isolated from fecal samples of overseas travelers were reported in Japan [15]. These reports provided an infectious source of Enterobacteriaceae conferring plasmid-mediated quinolone resistance in Japan. We examined the prevalence of plasmid-mediated quinolone resistance in S. Typhimurium isolated from food-producing animals.

A total of 225 isolates of S. Typhimurium from food-producing animals collected between 2003 and 2007 were derived from 156 cattle, 62 pigs and 7 poultry: includes 42 isolates of DT104, 8 of DT104B, and 2 of U302 (Table 1). Bacteriophage typing was performed according to the methods of the Health Protection Agency, London, United Kingdom [16]. Of the isolates, 132 S. Typhimurium isolates collected between 2003 and 2005 [6] were subjected to detection of the PMQR genes. The remaining 93 isolates between 2006 and 2007 were investigated for the presence of the PMQR genes and antimicrobial susceptibility. The presence of qnrA, qnrB and qnrS genes was determined by PCR [17]. The qnrC and qnrD genes were detected using the primers as previously described [18, 19], respectively. The qepA and acc(6')-Ib-cr genes were examined as previously described [20, 21]. Nucleotide sequences of both strands were determined directly on PCR products. The DNA alignments and deduced amino acid sequences were examined using the BLAST program (National Center for Biotechnology Information, USA). Minimum inhibitory concentrations (MICs) of antimicrobial agents were determined using the agar dilution methods according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [22]. The following 11 antimicrobials were tested: ampicillin (ABPC), cefazolin, colistin, chloramphenicol (CP), dihydrostreptomycin (DSM), gentamicin, kanamycin, oxytetracycline (OTC), nalidixic acid, enrofloxacin (ERFX), and trimethoprim. The MICs of each antimicrobial agent were interpreted using the recommendations of the CLSI [23]. The breakpoints not seen in the CLSI were defined in a previous study [1]. Staphylococcus aureus ATCC 29213, Enterococcus faecalis ATCC29212, E. coli ATCC 25922 and P. aeruginosa ATCC 27853 were used as quality control strains.

Table 1 Salmonella Typhimurium isolates used in this study
Full size table

Of 225 S. Typhimurium isolates, two isolates of DT104, 18-PLS-16 and 19-PLS-45, from different dairy cows on a single farm in 2006 and 2007 showed qnrS positive results. The sequencing of amplicons showed complete identity to qnrS1 previously identified on pAH0376 from a S. flexneri strain. None of the S. Typhimurium isolates had qnrA, qnrB, qnrC, qnrD, qepA and acc(6')-Ib-cr. The two isolates exhibited ERFX resistance (ERFX MIC, 2 mg/L) with resistances to ABPC, DSM, OTC and CP (Table 2).

Table 2 Susceptibility for several fluoroquinolones
Full size table

The QRDR of gyrA, parC and parE was examined in ERFX-resistant isolates by PCR amplification and sequencing using primers as described elsewhere [24]. In addition, susceptibility of ERFX-resistant isolates to fluoroquinolones was examined using the micro broth dilution methods according to CLSI guidelines [22]. For evaluation of active efflux of the ERFX-resistant bacteria, the MIC of ERFX was determined by the agar dilution method in the presence of carbonyl cyanide m-chlorophenylhydrazone (CCCP) (100 μM). They had no mutations in the QRDR of GyrA, ParC and ParE. The MIC of ERFX was not changed in the presence of CCCP (100 μM). The two isolates with qnrS1 exhibited almost the same MIC observed for each fluoroquinolone, which is relative low compared with the MIC for isolate (17-PLS-75) with mutations in the QRDR of GyrA and ParC.

Plasmid DNA was isolated from the qnrS1-positive isolates by the alkaline lysis method [25]. Extracted plasmids were transferred to Hybond-N+ membrane (Amersham Biosciences, Buckinghamshire, UK) using capillary blotting apparatus. The qnrS1 PCR product was labeled with DIG-11-dUTP by PCR using a DIG High Prime DNA Labeling Kit (Roche Diagnostics Ltd, East Sussex, UK). After hybridization with the qnrS1 probe, hybridized DNA was detected using a DIG Nucleic Acid Detection Kit (Roche Diagnostics Ltd). Using a plasmid profiling test, an approximately 93-kbp plasmid (virulence plasmid) was found in all four isolates, whereas there was also an approximately 9.6-kbp plasmid found in the qnr-conferring isolates. Hybridization tests revealed that qnrS1 was located on the 9.6-Kbp plasmid (Figure 1).

Figure 1
figure 1

Plasmid profile of S. Typhimurium isolates on a farm. Lane 1: 17-PLS-27, Lane 2: 17-PLS-28, Lane 3: 18-PLS-16, Lane 4: 19-PLS-45, Lane 5: Super coiled DNA ladder.

Full size image

The appearance of S. Typhimurium DT104 conferring qnrS1 on the farm is caused either by the introduction of S. Typhimurium DT104 conferring qnrS1 or the transfer of the qnrS1 plasmid to S. Typhimurium DT104 already existing on the farm. According to the CDC PulseNet protocol [26], genetic relatedness of isolates were analyzed by PFGE with XbaI and BlnI restriction enzymes. The isolates tested included two qnrS1-negative isolates of S. Typhimurium DT104 isolated in 2005 on a farm in which qnrS1-conferring isolates were found. In the present study, it was difficult to precisely distinguish between the two S. Typhimurium DT104 isolates without qnrS1 and the two S. Typhimurium DT104 isolates with qnrS1 by PFGE analysis (Figure 2). Our previous study showed that there is a variation in the BlnI-digested PFGE profiles of S. Typhimurium DT104 isolated from food-producing animals in Japan [5]. These results suggested that the S. Typhimurium DT104 already present on the farm fortuitously acquired the qnrS1 plasmid. Previous studies showed that qnrS1 in Typhimurium isolated in the UK was present on plasmids of 10,066 bp, which were transferable by the conjugation test and carry an IncN replicon [27, 28]. Further study need to clarify the source of plasmid bearing qnrS1.

Figure 2
figure 2

PFGE profile of S. Typhimurium isolates on a farm. Electrophoresis was performed using a CHEF-DR III System (Bio-Rad Laboratories, Hercules, CA, USA) with running conditions of 1 phase from 2.2 to 63.8 s at 180 V for 19 h. Lane1, 6, 11: Salmonella Braenderup H9812 digested with XbaI. Lane 2 and 7: 17-PLS-27, Lane 3 and 8: 17-PLS-28, Lane 4 and 9: 18-PLS-16, Lane 5 and 10: 19-PLS-45. Lane 2 to 5: XbaI digestion, Lane 7 to 10: BlnI digestion.

Full size image

This study demonstrated that the two isolates of S. Typhimurium collected from different cattle on a farm in 2006 and 2007 harbored qnrS1 on a 9.6-Kbp plasmid. At present in Japan, dissemination of qnrS1 among S. Typhimurium isolates from food animals may remain restricted. The spread of plasmids carrying qnr among Salmonella isolates of animal origin could have serious consequences for fluoroquinolone treatment of non-typhoid Salmonella infection in humans and animals. Previously, qnrS1 and qnrS2 were found in serovars Typhimurium, Corvallis, Montevideo, Agona, Braenderup and Alacua of Salmonella isolates from fecal samples of overseas travelers who had visited Thailand, Malaysia, Vietnam, Indonesia and Singapore, between 2001 and 2007 [15]. PMQR is identified in human isolates of Enterobacteriaceae but is likely to be rare in isolates from food-producing animals [29]. However, in China, plasmid-mediated quinolone resistance is frequently found in the isolates from food-producing animals [20]. Thus it would be difficult to prevent the invasion of resistance genes from foreign countries to Japan. The monitoring of fluoroquinolone use and quinolone resistance in bacteria of food-producing animal origin is essential to assess the level of risk of resistance in food-borne bacteria in the animals.