Validation of Novel Mycobacterium tuberculosis Isoniazid Resistance Mutations Not Detectable by Common Molecular Tests

Abstract

Resistance to the first-line antituberculosis (TB) drug isoniazid (INH) is widespread, and the mechanism of resistance is unknown in approximately 15% of INH-resistant (INH-R) strains. To improve molecular detection of INH-R TB, we used whole-genome sequencing (WGS) to analyze 52 phenotypically INH-R Mycobacterium tuberculosis complex (MTBC) clinical isolates that lacked the common katG S315T or inhA promoter mutations. Approximately 94% (49/52) of strains had mutations at known INH-associated loci that were likely to confer INH resistance. All such mutations would be detectable by sequencing more DNA adjacent to existing target regions. Use of WGS minimized the chances of missing infrequent INH resistance mutations outside commonly targeted hotspots. We used recombineering to generate 12 observed clinical katG mutations in the pansusceptible H37Rv reference strain and determined their impact on INH resistance. Our functional genetic experiments have confirmed the role of seven suspected INH resistance mutations and discovered five novel INH resistance mutations. All recombineered katG mutations conferred resistance to INH at a MIC of ≥0.25 μg/ml and should be added to the list of INH resistance determinants targeted by molecular diagnostic assays. We conclude that WGS is a useful tool for detecting uncommon INH resistance mutations that would otherwise be missed by current targeted molecular testing methods and suggest that its use (or use of expanded conventional or next-generation-based targeted sequencing) may provide earlier diagnosis of INH-R TB.

Keywords: Mycobacterium tuberculosis; drug resistance evolution; isoniazid; tuberculosis.

FIG 1

Selection and exclusion criteria for clinical strains included in this study.

FIG 2

Genetic impact of large insertions and deletions at the furA-katG-Rv1907c operon in clinical MTBC strains. Of the 14/52 study set clinical strains that encoded large (>10-bp) insertions or deletions at the furA-katG-Rv1907c operon, 7/14 were predicted to impact katG activity, leading to INH resistance. The insertions/deletions present at this locus in the other 7/14 strains were not predicted to impact INH resistance based on the literature but were generally accompanied by mutations in katG (6/7 strains), likely to confer the INH-R phenotype. These seven strains also shared an identical 14-bp deletion in Rv1907c (marked with an asterisk; deleted sequence, TCATCCCCGTCTCG) and were of the same lineage (L2). One strain presented here, RLT-5, did not encode any mutations clearly related to INH resistance. The dotted-line boxes indicate deleted sequences. Carrots indicate insertion sites. Hatched open reading frames are cotranscribed in M. tuberculosis (86). Promoters are indicated by bent arrows (48).

FIG 3

INH resistance determinants in M. tuberculosis. Illustration (not to scale) of the two primary loci involved in INH resistance, furA-katG and fabG1-inhA. Promoters are shown as bent arrows (a novel promoter generated by the fabG1 L203L silent mutation is depicted with a dashed outline). Shown in parentheses are mutations previously confirmed by functional genetics in M. tuberculosis (16–18, 21–23). Underlined mutations are targeted by conventional rapid molecular tests (Hain GenoType MTBDRplus v2 [29, 87], Nipro NTM+MDRTB detection kit 2 [30], and CDC's MDDR service). Windows of detection for CDC's MDDR service Sanger sequencing assay, per current primer binding sites, are shown as gray bubbles (Jeff Driscoll, personal communication). Recombineered INH resistance mutations from this study are in larger text and boldfaced. Mutations in unformatted text were observed among INH-R clinical MTBC strains analyzed by WGS in this study and were generally considered likely to confer INH resistance due to previous reports in the literature (Table 1). ahpC promoter mutations have not yet been definitively demonstrated to independently confer INH resistance in MTBC and thus are not depicted here. This illustration omits numerous mutations observed in clinical isolates of MTBC or evaluated in other organisms or biochemically and is not intended to be comprehensive.