Sir,
Members of the genus Bacteroides are among the most abundant members of the intestinal microbiota and are important opportunistic pathogens. They have the most numerous antibiotic resistance mechanisms and the highest percentage of resistance of all pathogenic anaerobic bacteria.1 During the last decade, strains of Bacteroides spp. with reduced susceptibility to metronidazole, the drug of choice for preventing and treating anaerobic infections, have been frequently described.2–4 Despite the fact that several mechanisms of metronidazole resistance have been described, reduced susceptibility to metronidazole is more often associated with the presence of a nitroimidazole reductase encoded by a nim gene.3,5,6 This enzyme seems to convert 4- or 5-nitroimidazole into 4- or 5-aminoimidazole, thus avoiding the formation of toxic nitroso radicals that are essential for antimicrobial activity.5 Currently, nine nim genes (nimA/nimG, nimI and nimJ) have been described, mainly in Bacteroides spp. and more rarely in other anaerobes.1–3,5,6 They were either plasmid or chromosomally encoded and associated with highly variable metronidazole MICs (from 0.25 to >128 mg/L). Transcription may be activated by various ISs found upstream of the nim genes (IS1168, IS1169, IS1170, IS612 and ISBf6).2,3,5,6 We describe here a 10th variant of the metronidazole resistance gene.
Two Bacteroides fragilis clinical isolates, BF2545 and BF5114, were recovered 3 months apart (February and May 2005) in polymicrobial cultures from hip fluid samples of a patient with a 19 year history of post-accident paraplegia, presenting sacral and large trochanteric pressure ulcers, as well as infection of the exposed bone (chronic osteitis). The patient was hospitalized at the University Hospital of Montpellier (South of France). PFGE comparison of XbaI-restricted DNA of both isolates showed identical pulsotypes indicating that they belonged to the same clone. In vitro susceptibility testing7,8 showed that strains BF2545 and BF5114 displayed low-level resistance to metronidazole (MICs = 8 mg/L), as well as high-level resistance to amoxicillin (MICs >64 mg/L), and were susceptible to imipenem, co-amoxiclav and clindamycin (MICs of 2, 2 and 1 mg/L and 4, 2 and 2 mg/L, respectively). PCR amplification with NIM-3/NIM-5 primers5 yielded, in both isolates, a 458 bp product sharing 100% sequence identity (accession numbers KX576455 and FJ969397). The highest sequence identity was obtained with nimG (83.6%), followed by nimD (82.5%), nimB (81.7%), nimF (76.5%), nimE and nimA (74.6%), nimC (73.4%) and nimI and nimJ (70.4%). This new nitroimidazole reductase gene was named nimH. Predicted amino acid sequence alignment of NimH with homologous proteins from other anaerobic bacteria showed 54.6%–90.0% identity. Maximum likelihood and neighbour-joining phylogenetic trees based on partial Nim amino acid sequences (151 amino acids) showed NimH to be closest to NimG and NimB. Of note, nim genes were very scarcely detected in our institution (15 detected since 2003).
IS-targeting PCR5 led to an 840 bp product in both isolates with primers targeting IS612 only. PCR mapping led to ∼1710 bp band showing that this IS was located upstream of the nimH gene (Figure 1). For both strains, the 1690 bp sequence (accession numbers KX576456 and KX576457) of this PCR product showed the presence of a 1214 bp sequence encoding a partial transposase gene presenting 99.8% identity with the IS614B element sequence. This is the first known description of an association between a nim gene and an IS614 transposase. IS614B was previously described as a putative mosaic or hybrid of IS612, IS614 and IS942, all members of the IS4 family.9 It was currently only associated with the imipenem or cephalosporin resistance genes cfiA and cfxA.9,10 Although both strains presented in vitro susceptibility to imipenem, the absence of the cfiA gene was confirmed by PCR.9 The putative promotor sequence of nimH seemed to be located within IS614B where we identified sequences homologous to consensus sequences of B. fragilis promotors (Figure 1).
Schematic diagram showing nimH and IS614B genetic organization, as well as the PCR strategy for genetic mapping by using primers targeting IS1168, IS1169, IS1170, ISBf6 and IS612.5,6 Of note, the latter primer pair (IS6141i and IS6142i) is able to detect IS612, IS614, IS614B and IS614C.6 P indicates the putative promotor sequence of nimH within IS614B, based on sequences homologous to consensus sequences of B. fragilis promotors, TAnnTTTG in the −7 region and TnTG in the −33 region. IRR indicates the right inverted repeats of IS614B. Positions of primers are presented by arrows. NIM3 and NIM5 primers were used for amplification of the resistance gene. IS6141i and IS6142i primers were used for amplification of the IS. IS6141i and NIM3 reverse (5′-CGCTTACGCCGCATTTCTCTGAACAT-3′; this study) primers were used to determine whether IS614B was located upstream of nimH or not. IS6141i, NIM3 reverse and NIM5 primers were used for sequencing.
The location of nim genes in either plasmid or chromosomal DNA was determined by plasmid extraction or DNA EcoRI restriction followed by Southern hybridization.2,5 Chromosomal hybridization was observed for nimH-positive strains, as well as nimB-positive B. fragilis strain BF8 (used as positive control). Plasmid extraction revealed the presence of ∼4 kb plasmid in the two B. fragilis isolates (BF2545 and BF5114), as well as ∼8 kb plasmid in the B. fragilis 638R(pIP417) nimA-positive strain. Hybridization of plasmid DNA was positive for the nimA-positive control strain only. Altogether, these results supported the chromosomal location of the nimH gene in both strains BF2545 and BF5114.
In this patient, we were unable to trace previous treatment with nitroimidazole derivatives as a selective pressure for the emergence of this metronidazole-resistant B. fragilis clone. The medical file of the patient documented treatment with co-amoxiclav and silver sulphadiazine. Patient follow-up included additional surgery for persistent discharge, treatment with ciprofloxacin and trimethoprim/sulfamethoxazole and no isolation of B. fragilis from the specimens sampled by the end of 2005.
This work describes the 10th variant of the metronidazole resistance gene, i.e. nimH. Its chromosomal location suggests a likely acquisition due to the IS614B transposase, as we did not find it in other B. fragilis genomes available in sequence databases. The increasing diversity of nim genes potentially encoding metronidazole resistance is worrisome in the context of the emergence of MDR Bacteroides isolates and because this antimicrobial agent is traditionally used in empirical therapy of infections where anaerobes are highly suspected.4
Nucleotide sequence accession numbers
Sequences determined in this work have been deposited in the EMBL sequence database with accession codes from KX576455 to KX576457 and FJ969397.
Acknowledgements
These results were presented during the Twelfth National Congress of the Société Française de Microbiologie, Paris, France, 2016 (Poster number: P2).
We are very grateful to J. Sóki for kindly supplying the following strains used as positive controls: Bacteroides thetaiotaomicron BT13 (nimC, IS1170), B. fragilis BF388 (nimE, ISBf6) and B. fragilis BF6712 (IS612). We also thank Agnes Masnou, Raymonde Devine and Laurent Calvet for technical assistance, Corinne Teyssier for help with hybridization experiments and Christophe Merlin for interesting discussions on mobile genetic elements.
Funding
This study was supported by internal funding.
Transparency declarations
None to declare.
