|Home | About | Journals | Submit | Contact Us | Français|
The ribosomal L3 protein was identified as a novel target in linezolid (LZD)-resistant Mycobacterium tuberculosis strains. Next-generation sequencing confirmed rplC T460C as the sole mutation in an LZD-resistant M. tuberculosis H37Rv strain selected in vitro. Sequencing analysis revealed the rplC T460C mutation in eight further LZD-resistant isolates (three in vitro-selected mutants and five patient isolates, including isolates from three different patients that developed LZD resistance during treatment) but in none of the susceptible control strains (n = 84).
Mycobacterium tuberculosis, the causative agent of tuberculosis (TB), still represents a huge challenge for infectious disease control worldwide. Since the appearance of multidrug-resistant (MDR; resistant to at least isoniazid and rifampin) (8) and extensively drug-resistant (XDR; additionally resistant to one fluoroquinolone and one injectable drug) (4) M. tuberculosis strains, the application of reserve antibiotics has become more important. Linezolid (LZD), which belongs to the oxazolidinone group (1), has been used primarily for the treatment of methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococcus infections (10) and has been used off-label to treat MDR and XDR TB patients (9, 12). LZD acts on the 50S ribosomal subunit, specifically, the peptidyl-transferase center (PTC), by blocking the binding of tRNA and thus inhibiting bacterial cell growth (5).
The first occurrence of LZD-resistant M. tuberculosis clinical isolates was described by Richter et al. (11) As potential target genes, the 23S rRNA gene, rplV and rplD (which encode ribosomal proteins L4 and L22), erm-37 (which encodes methyltransferase), and whiB7 (which encodes a putative regulator) were sequenced. However, since no mutations were detected in these genes, the mechanism of LZD resistance in clinical M. tuberculosis isolates has remained unclear. To gain further insight into the mechanisms involved, LZD-resistant M. tuberculosis strains were selected in vitro (2). Sequence analysis of the 23S rRNA gene revealed mutations in 50% of the selected mutants, primarily at nucleotide position 2061. However, no further polymorphisms were found in the remaining selected clones.
Recently, mutations in rplC were described as a novel LZD resistance mechanism in MRSA (7). To determine if rplC is also involved in LZD resistance in M. tuberculosis, we sequenced rplC in (i) LZD-resistant, in vitro-selected mutants; (ii) LZD-resistant clinical isolates; (iii) follow-up isolates from three different patients; and (iv) a reference collection comprising all of the major phylogenetic lineages. The primers used for amplification were rplC_5′(+249) (5′-GCTGCGGCTGGACGACTC-3′) and rplC_3′(+668) (5′-CTCTTGCGCAGCCATCACTTC-3′). The conditions used were (i) initial denaturation at 95°C for 15 min, (ii) denaturation at 94°C for 30 s, (iii) annealing at 65°C for 30 s, (iv) elongation for 30 s at 72°C, and (v) terminal elongation for 10 min at 72°C. Steps ii to iv were performed 35 times. The PCR products thus obtained were sequenced using an ABI 3130xl genetic analyzer (Applied Biosystems) and an ABI BigDye Terminator cycle sequencing kit (version 3.1) according to the manufacturer's instructions. Analysis of sequence data was performed using the DNAStar Lasergene (version 8.0) software suite.
First, the rplC genes of strains from the aforementioned selection experiments were sequenced (2) (selected on 4 μg/ml LZD; MICs for LZD of 4 to 16 μg/ml). As references, the genes of the parental strains used for in vitro selection were sequenced. All but one of the five strains sequenced carried a mutation at nucleotide 460 (T/C) compared to the respective reference strain, resulting in an amino acid exchange from cysteine to arginine at codon 154 (Table 1).
To verify the identified rplC T460C mutation as the only polymorphism in the clones selected, and thus responsible for the resistance phenotype, whole-genome sequencing of one of the in vitro-selected LZD-resistant mutants (9679/00 1.1.2) was carried out. Isolated genomic DNA was sequenced by GATC Biotech in a paired end run on the Illumina platform. Derived reads had a length of 46 bp for strain 9679/00 1.1.2. Reads were mapped to the M. tuberculosis H37Rv genome (GenBank accession no. NC_000962.2) using the CLC Genomics Workbench software (CLC bio) with default settings. For strain 9679/00 1.1.2, 11,431,954 of the 12,400,138 reads were matched to the reference genome, resulting in an average coverage of 119 reads. CLC Genomics Workbench was also employed for single nucleotide polymorphism (SNP) detection using default settings with a minimum coverage of four reads and a minimum variant frequency of 75%. A total of 30 SNPs were detected compared to the H37Rv reference sequence (see Table S1 in the supplemental material). Of these SNPs, the only ones considered were those not located in repetitive elements (possible mapping artifacts) and absent from the genomes of three further H37Rv variants selected in vitro for resistance to other antibiotics (possible failures in the H37Rv reference sequence). The rplC T460C mutation was the only SNP detected in the in vitro-selected LZD-resistant strain (9679/00 1.1.2) under these criteria.
To confirm the clinical significance of the rplC T460C mutation, the rplC DNA sequence was analyzed in two XDR TB patient isolates also showing LZD resistance (Table 1). Both of the strains analyzed showed the rplC T460C mutation as the only variation in the sequence analyzed.
To further investigate the significance of rplC T460C for LZD resistance development in clinical isolates, follow-up isolates from three different patients were sequenced (Table 2). The isolates were described in detail in a previous report (11). All patients were infected with M. tuberculosis strains which acquired LZD resistance during treatment. Samples were taken over a time span of several years. IS6110 fingerprint genotyping was used to confirm that the isolates obtained from one patient were the same over the entire treatment period (no reinfection). Furthermore, IS6110 fingerprinting confirmed that the three patients do not belong to one cluster, as the patterns of the strains showed marked differences (data not shown).
rplC sequence analysis of the follow-up isolates showed a clear correlation between the occurrence of the LZD resistance phenotype and the emergence of the rplC T460C mutation (Table 2). All of the susceptible isolates obtained before the emergence of LZD resistance had no mutation in rplC (n = 13).
Besides mutations in rplC, alterations in the rplD gene, which codes for ribosomal protein L4, have been described as mediating LZD resistance in Staphylococcus aureus (7), Clostridium perfringens (3), and pneumococci (13). To further elucidate previously unknown resistance mechanisms in the strains analyzed in this study, the rplD genes of all of the strains were sequenced by using primers rplD_5′(−153) (5′-CCGGGCGGATGGGCAATGACC-3′) and rplD_3′(+782) (5′-GGAATCCGGGCGCACCAAAAAC-3′).
The wild-type (WT) rplD sequence was determined for all of the in vitro-selected mutant strains, all of the resistant clinical isolates (Table 1), and the follow-up isolates collected from patients 1 and 2 (Table 2). A synonymous SNP (Asn214Asn [AAC/AAT]) was detected in all of the follow-up isolates from patient 3 (Table 2). However, as this polymorphism occurs in both susceptible and resistant isolates, it obviously does not play a role in mediating LZD resistance. Whole-genome sequencing might be interesting to determine putative previously unknown mutations responsible for the resistance phenotype of in vitro-selected, LZD-resistant mutant 9679/00 4.4.1 and the elevated MIC for 9679/00 1.4.1 (16 μg/ml).
To investigate if mutations in rplC, especially T460C, occur in a phylogenetic-lineage-dependent manner, we sequenced the rplC genes of 71 LZD-susceptible strains of the M. tuberculosis complex, comprising all of the major phylogenetic lineages (see Table S2 in the supplemental material). The collection comprises 14 different M. tuberculosis genotypes, as well as 3 M. africanum genotypes, M. bovis, M. canettii, M. caprae, M. microti, and M. pinnipedii. For the majority of the genotypes, three different strains were included. One phylogenetic SNP was detected in all of the M. canettii strains (A459G). Although the polymorphism detected is directly adjacent to the T460C SNP, it is highly unlikely to play a role in LZD resistance development, as the A459G mutation is a synonymous SNP which does not lead to an amino acid exchange. The rplC T460C mutation has not been found in any of the strains analyzed and can therefore definitely be excluded as a phylogenetic SNP.
In conclusion, this study reveals a novel mechanism of LZD resistance in M. tuberculosis complex strains. The same mutation in the rplC gene (T460C) was identified in in vitro-selected mutants and in clinical isolates. Mutations in rplC have previously been identified in LZD-resistant Staphylococcus strains at positions Gly152Asp, Gly155Arg, Ala157Arg, and Met169Leu, located in the PTC region (6, 7). An amino acid alignment of the L3 proteins of M. tuberculosis and S. aureus illustrates that the rplC T460C mutation (Cys154Arg) corresponds to the Ala157Arg position and is therefore probably also located directly in the ribosome and drug binding region. However, to definitely prove the influence of the rplC T460C mutation on LZD resistance, the generation of an isogenic mutant strain via homologous recombination is mandatory.
The MICs for the in vitro-selected mutants analyzed ranged from 4 to 8 μg/ml (16 μg/ml for one strain); higher MICs (32 μg/ml) were detected for strains displaying mutations in the 23S rRNA gene (2). As was observed in S. aureus (7), mutations in rplC could be responsible for lower-level LZD resistance, whereas mutations in the 23S rRNA gene might, alone or in combination with rplC mutations, mediate higher levels of resistance in M. tuberculosis.
This study highlights the clinical relevance of the rplC T460C mutation for LZD resistance, which should be incorporated into molecular assays for the detection of second-line drug resistance in M. tuberculosis.
We thank I. Razio, Research Center Borstel, for excellent technical assistance and C. Rückert, Bielefeld University, for help with the CLC Genomics Workbench Software.
S. Niemann and S. Feuerríegel were supported by the European Union TB-PAN-NET (FP7-223681) project.
Published ahead of print 27 February 2012
Supplemental material for this article may be found at http://aac.asm.org/.