Salmonella enterica (56 isolates) and Escherichia coli (24 isolates) were selected from the HPA collection based on their resistance to amikacin (breakpoint concentration routinely used in HPA Salmonella Reference Unit = 4 µg/mL). Because 16S rRNA methyltransferases confer high-level resistance to amikacin, 13 S. enterica isolates were selected on the basis of ability to grow on Isosensitest agar containing 500 µg/mL amikacin, whereas none of the E. coli isolates grew under these conditions. All isolates belonged to serotype Virchow. Further antimicrobial susceptibility testing by microdilution by using dehydrated Sensititer plates following the CLSI guidelines confirmed high-level resistance to 4,6-disubstituted 2-deoxystreptamines (Table 1). PCR screening of the 13 isolates for armA, rmtA, rmtB, rmtC, and rmtD (8) identified rmtC. Nucleotide sequencing of the amplicons confirmed an rmtC gene with 100% identity with those originally identified in Proteus mirabilis strain ARS68 isolated from an inpatient in Japan (6) and P. mirabilis strain JIE273 from Australia (7). To our knowledge, this is the third report of rmtC-bearing bacteria. Class one integrons were amplified (9), and sequenced. Isolates resistant to neomycin bore the aac(6′)-Ib gene cassette, whereas the dfrA1 gene was responsible for resistance to trimethoprim.

Figure 1

Figure 1. Pulsed-field gel electrophoresis patterns of rmtC-positive Salmonella enterica serovar Virchow isolates. Lanes: B, S. Braenderup H9812 size standard; 1, H0 5164 0340; 2, H0 5366 0426; 3, H0 6018 0151; 4,...

Figure 2

Figure 2. Map of United Kingdom showing geographic location of the 13 Salmonella enterica serovar Virchow isolates bearing rmtC. Each number represents 1 isolate in chronologic order of isolation as shown in <>

Twelve of the 13 S. enterica strains were originally isolated over a 4-year period from patients with clinical infection; 1 strain was obtained from frozen produce. Seven of 12 strains were obtained from patients with histories of foreign travel; 4 of the 7 patients had reported recent travel to India (Table 2). P. mirabilis strain JIE273 was also isolated from a patient who had recently returned from India (7). Investigations to ascertain the presence of rmtC genes in India are under way. To identify a possible link between the isolates, chromosomal DNA was embedded in agarose plugs prepared according to the pulsed-field gel electrophoresis (PFGE) protocol of PulseNet Europe (10). PFGE patterns showed only 1–2-band differences (Figure 1) and correlated with phage typing data (Table 1). All clinical isolates were recovered from feces, except a blood isolate recovered from a patient with invasive salmonellosis (Table 2). The temporal and geographic distribution of the isolates suggested independent acquisition of infections in most cases and possibly epidemiologically linked cases, e.g., strains 9 and 10 (Table 2; Figure 2).

PCR with primers ISEcp1R-F and rmtC-down (7) showed that the rmtC gene and immediate upstream sequences (GenBank accession nos. FJ984623–FJ984634 for human isolates and GQ131574 for the food isolate) were identical to those previously identified in P. mirabilis (6,7), in which ISEcp1 has been shown to play a role in the expression and transposition of the rmtC gene (11). However, the complete ISEcp1 element could not be amplified by using primers ISEcp1 5′ and ISEcp1 reverse, which suggests either partial deletion of this element or involvement of a different ISEcp1-like element in spread of rmtC in Salmonella (6,12). Attempts to isolate rmtC by conjugal transfer to rifampin-resistant E. coli 20R764 were unsuccessful, as was electroporation into E. coli LMG194 and ElectroMAX DH10B cells (both Invitrogen, Paisley, UK) by using plasmid preparations. An ≈100-kb rmtC-bearing plasmid was previously transferred from P. mirabilis ARS68 by electroporation but could not be mobilized by conjugation (6), and attempts to transfer the rmtC plasmid from P. mirabilis JIE273 by electroporation and conjugation failed (7). This finding contrasts with some qualities of the other methyltransferases, such as armA and rmtB, which are mostly located on conjugative plasmids (8,13).

The location of the rmtC gene was determined with PFGE by using I-CeuI nuclease treatment. Agarose plugs were digested with 9.5 U I-CeuI nuclease (New England Biolabs, Beverly, MA, USA). Separated DNA fragments were transferred onto a nylon membrane (GE Healthcare, Madrid, Spain) and hybridized with 16S rDNA and rmtC probes labeled with DIG-11-dUTP. Hybridization, labeling, and detection were performed according to the manufacturer’s recommendations (Roche Applied Science, Mannheim, Germany). A DNA band hybridized with both probes, showing that the rmtC gene was located on the chromosome. Results of hybridization of plasmid extractions (Plasmid Midi kit; QIAGEN, Inc., Chatworth, CA, USA) with the rmtC probe were negative (data not shown).

We describe the occurrence of 16S rRNA methyltransferase rmtC in Salmonella isolates and the rmtC gene in Europe. We also report that a producer of 16S rRNA methyltransferase was isolated from food.

The overall isolation frequency of 16S rRNA methyltransferase–producing S. enterica is low (13/81,632 strains) in the United Kingdom, and these genes were absent in E. coli. However, spread of multidrug-resistant isolates that express 16S rRNA methyltransferases, amplified by the association of these genes with the ISEcp1 element, raises clinical concern that further spread is likely. Ongoing surveillance of 16S rRNA methyltransferases in isolates found in food products and in humans and animals is crucial to delay the spread of resistance to these classes of antimicrobial agents.

While this manuscript was under revision, an S. enterica ser. Virchow isolate bearing the rmtC gene isolated from a child with a history of travel to India was reported in the United States (14).