LETTER
Staphylococcus warneri, one of the emerging coagulase-negative staphylococci, may be responsible for bone and joint infections (1–3), for which linezolid (LZD) seems to be a good alternative treatment (4, 5). LZD resistance is related to mutations in domain V of 23S rRNA (rrl), in ribosomal proteins L3 and L4, and to cfr and optrA genes (6, 7). In this report, we explored the possibilities and consequences of the emergence of LZD resistance in S. warneri, since it has not been previously published.
Three independent LZD resistance selections were performed in vitro (8) with an S. warneri isolate (SW0; LZD MIC = 0.5 mg/liter) from a case of monomicrobial diabetic foot osteitis. We obtained LZD-resistant mutants for all selections (MIC = 64 mg/liter). Resistance-associated mutations occurred in the three selections with identical timelines, first with G152D in L3 and then with G2576T in rrl (Table 1) (9). The mutation G2576T is the most frequently described mutation in clinical LZD-resistant staphylococci, and G152D has also been previously reported (6, 10). Small differences occurred, with a K68N mutation in L3 in the third selection and the reversion of the G152D mutation in the first selection (Table 1). Mutations in ribosomal proteins were not associated with a significant LZD resistance (MIC ≤ 4 mg/liter) (8, 11). A specific PCR described for Staphylococcus aureus was used to determine the rRNA operon (rrn) copy number (12), and the number of mutated rrl copies was inferred from pyrosequencing (Pyromark; Qiagen) (13) or whole-genome sequencing (MiSeq; Illumina). The MIC increase to 64 mg/liter was associated with an increase in mutated rrl copies and with the deletion of 1 of the 6 rrn copies (Table 1; Fig. 1). This loss always involved the third operon (12).
LZD resistance-associated mutations, number of SNPs, and generation time increase among sequenced mutants, in comparison with parental strain SW0a
Correlation of LZD resistance-associated mutations and increase in generation time. Gray ovals represent L3 and L4 ribosomal protein mutations, and black ovals represent the rrl mutation.
Various biological costs (BCs) were observed throughout the selections. In the first and second selections, a weak BC was observed, in contrast to the third selection, where an important increase in generation time of up to 40% appeared when 60% of the rrl copies harbored the G2576T mutation (Fig. 1). A small-colony variant (SCV) phenotype also appeared concomitantly. These differences among selections could be related to the L4 K68N mutation in the third selection, in association with perturbation of the three-dimensional structure of 23S rRNA (14) or to additional mutations. A whole-genome single-nucleotide polymorphism (SNP) analysis was performed after de novo assembly of the parental LZD-susceptible strain (SW0) (BioNumerics; Applied Maths). We observed one or two supplementary SNPs (apart from resistance-associated mutations) in mutants SW1-16 and SW2-17 without a BC and four SNPs in mutant SW3-36 with a BC (Table 1). Therefore, the BC and SCV phenotype may be due to additional mutations. We observed the T70K mutation in RelA/SpoT, which is a hydrolase/synthase enzyme for (p)ppGpp (Table 2). Gao et al. have shown in S. aureus that the F128Y mutation, localized in the N-terminal hydrolase domain of the RelA hydrolase, like our mutation, was also associated with reduced growth and an SCV phenotype (15). Moreover, a mutation leading to a codon stop in the sigB gene encoding the alternative sigma factor that is involved in bacterial metabolism regulation was observed (Table 2). The possible relationship with the observed phenotype should be further explored.
Amino acid change analysis among sequenced mutants, in comparison with parental strain SW0
We described for the first time the genome modification of S. warneri under LZD pressure without BC and with the systematic loss of the third rrn operon associated with the increase in LZD resistance level. These features might require caution when using LZD in monotherapy.
Accession number(s).The sequencing raw data for each isolate were submitted to GenBank under BioProject record number PRJNA473123.
ACKNOWLEDGMENTS
We thank Delphine Guirlich for technical assistance and Anita Owens for English editing.
We declare no conflicts of interest.
- Copyright © 2018 American Society for Microbiology.
