Catalase activity deficiency sensitizes multidrug-resistant Mycobacterium tuberculosis to the ATP synthase inhibitor bedaquiline

Abstract

Multidrug-resistant tuberculosis (MDR-TB), defined as resistance to the first-line drugs isoniazid and rifampin, is a growing source of global mortality and threatens global control of tuberculosis disease. The diarylquinoline bedaquiline has recently emerged as a highly efficacious drug against MDR-TB and kills Mycobacterium tuberculosis by inhibiting mycobacterial ATP synthase. However, the mechanisms underlying bedaquiline's efficacy against MDR-TB remain unknown. Here we investigate bedaquiline hyper-susceptibility in drug-resistant Mycobacterium tuberculosis using systems biology approaches. We discovered that MDR clinical isolates are commonly sensitized to bedaquiline. This hypersensitization is caused by several physiological changes induced by deficient catalase activity. These include enhanced accumulation of reactive oxygen species, increased susceptibility to DNA damage, induction of sensitizing transcriptional programs, and metabolic repression of several biosynthetic pathways. In this work we demonstrate how resistance-associated changes in bacterial physiology can mechanistically induce collateral antimicrobial drug sensitivity and reveal druggable vulnerabilities in antimicrobial resistant pathogens.

Fig. 1. Catalase activity deficiency sensitizes drug-resistant Mycobacterium tuberculosis to bedaquiline.

a Normalized count distributions of BDQ MICs for MDR (n = 3,958) and non-MDR (n = 7,761) clinical strains characterized by the CRyPTIC Consortium. b Normalized count distributions of BDQ MICs for INH-susceptible (n = 5,986) and INH-resistant (n = 5,078) clinical strains characterized by the CRyPTIC Consortium. c Left: MDR (orange: TDR-TB-19, TDR-TB-31, TDR-TB-193, TDR-TB-198) and non-MDR (red: TDR-TB-42) INH-resistant clinical strains from the TDR-TB strain bank exhibit greater susceptibility to 2.7 μg/mL BDQ than INH-susceptible (gray: TDR-TB-77, TDR-TB-81, TDR-TB-126, TDR-TB-164) clinical strains in 30-day time-kill experiments. Wild-type H37Rv is included for reference (black). Right: Relative 30-day survival following BDQ treatment was significantly worse in INH-resistant strains than in INH-susceptible strains as determined by two-sided Mann–Whitney statistical testing (p = 0.0317). 4 TDR-TB strains were MDR (TDR-TB-19, TDR-TB-31, TDR-TB-193, TDR-TB-198). 4 INH-resistant and MDR TDR-TB strains possessed katG S315 mutations (TDR-TB-19, TDR-TB-31, TDR-TB-42, TDR-TB-198). n = 1 biological replicate for each strain. d H37Rv ∆katG cells are sensitized to BDQ relative to wild-type cells in 8-day growth inhibition dose-response experiments. katG complementation reduces BDQ sensitivity in ∆katG cells. n = 3 biological replicates for H37Rv and ∆katG. n = 2 biological replicates for ∆katG:pkatG. e H37Rv ∆katG cells are hyper-susceptible to 2.7 μg/mL BDQ relative to wild-type cells in 30-day time-kill experiments from n = 3 biological replicates. f A furA over-expression mutant is hyper-susceptible to 2.7 μg/mL BDQ relative to empty vector control cells in 30-day time-kill experiments. Time-kill data quantified as % CFUs normalized to CFUs on Day 0 from n = 3 biological replicates. Data depicted as mean ± SEM. Source data are provided in the Source Data file.

Fig. 2. Bedaquiline potentiates reactive oxygen species formation in cells with deficient catalase activity.

a H37Rv ∆katG cells are hypersensitive to H₂O₂ relative to wild-type cells in 8-day growth inhibition dose-response experiments. katG complementation reduces H₂O₂ sensitivity in ∆katG cells. n = 3 biological replicates for H37Rv and ∆katG. n = 2 biological replicates for ∆katG:pkatG. b MDR and non-MDR INH-resistant clinical strains are hypersensitive to H₂O₂ relative to INH-susceptible clinical strains in 8-day growth inhibition dose-response experiments. n = 4 biological replicates for non-MDR INH-resistant TDR-TB-42 (red). n = 1 biological replicate for all other TDR-TB clinical strains. Error bars for INH-susceptible and MDR clinical strains computed across INH-susceptible and MDR clinical strains, respectively. c BDQ treatment increases devR, katG, and oxyS expression in wild-type H37Rv cells as measured by RNA sequencing. Expression data reported as smooth quantile normalized log₂ sequencing counts. n = 3 biological replicates. d KatG-deficient mc²8245 cells are hypersensitive to BDQ relative to KatG-replete mc²7902 cells in 8-day growth inhibition dose-response experiments. n = 2 biological replicates for mc²7902 and n = 3 biological replicates for mc²8245. e Left: BDQ induces ROS accumulation in mc²7902 and mc²8245 cells as reported by CellROX fluorescence over 4 days treatment with 0.68 μg/mL BDQ. Right: ROS accumulation was further enhanced in mc²8245 cells than in mc²7902 cells. n = 3 biological replicates. p = 0.0279 as determined by two-sided Welch’s t-test. Error bars depict standard errors of the mean after error propagation. f Protein carbonylation is potentiated in mc²8245 cells relative to mc²7902 cells after 16 hours treatment with 2.7 μg/mL BDQ as reported by ELISA. n = 3 biological replicates. g Deoxyguanosine oxidation is potentiated in mc²8245 relative to mc²7902 cells after 16 hours treatment with 2.7 μg/mL BDQ as reported by ELISA. n = 3 biological replicates. Brown-Forsythe and Welch ANOVA statistical tests were performed on RNA expression, CellROX, protein carbonylation, and deoxyguanosine oxidation experiments, with comparisons between BDQ-treated and untreated cells wild-type and ∆katG cells or between untreated wild-type and ∆katG as indicated with FDR correction. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Data depicted as mean ± SEM. Source data are provided in the Source Data file.

Fig. 3. Transcriptional programs induced by catalase activity deficiency sensitize Mtb to bedaquiline.

a Expression of the transcriptional regulators Rv3160c (p = 0.0042), kmtR (p = 0.0022), and prpR (p ≤ 0.0001) is higher in untreated ∆katG cells than in H37Rv cells as measured by RNA sequencing. Expression data reported as smooth quantile normalized log₂ sequencing counts. Two-sided Welch’s t-tests were performed on RNA expression data with comparisons between untreated wild-type and ∆katG cells. b Rv3160c over-expression sensitizes cells to BDQ-inhibited ATP synthesis and growth in 7-day dose-response experiments as determined by BacTiter-Glo and optical density. c kmtR over-expression sensitizes cells to BDQ-inhibited ATP synthesis in 7-day dose-response experiments. n = 3 biological replicates for each experiment. **p ≤ 0.01. ****p ≤ 0.0001. Data depicted as mean ± SEM. Source data are provided in the Source Data file.

Fig. 4. Catalase activity deficiency sensitizes Mtb to DNA damage.

a Expression of several DNA repair enzymes is higher in BDQ-treated ∆katG cells relative to BDQ-treated H37Rv cells as measured by RNA sequencing. alkA, radA, recG, and ung are members of gene Cluster 7. Expression data reported as smooth quantile normalized log₂ sequencing counts. n = 3 biological replicates. b ∆katG cells are sensitized to the DNA damaging agent phleomycin relative to wild-type cells in 8-day growth inhibition dose-response experiments. n = 3 biological replicates. c KatG-deficient mc²8245 cells are sensitized to phleomycin in 8-day dose-response experiments. n = 3 biological replicates. d INH-resistant clinical isolates are sensitized to phleomycin relative to INH-susceptible clinical isolates in 8-day dose-response experiments. n = 4 biological replicates for non-MDR INH-resistant TDR-TB-42 (red). n = 1 biological replicate for all other TDR-TB clinical strains. Error bars for INH-susceptible and MDR clinical strains computed across INH-susceptible and MDR clinical strains, respectively. Brown-Forsythe and Welch ANOVA tests were performed on RNA expression data with comparisons between BDQ-treated and untreated cells wild-type and ∆katG cells or between untreated wild-type and ∆katG with Dunnett’s T3 multiple comparisons test FDR correction, as indicated. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Data depicted as mean ± SEM. Source data are provided in the Source Data file.

Fig. 5. Catalase activity deficiency sensitizes Mtb to inhibition of folate biosynthesis.

a Simulated chorismate synthase (CHORS reaction), dihydropteroate synthase (DHPS2 and FOLD3), dihydrofolate synthase (DHFS), and dihydrofolate reductase (DHFR) activities for BDQ-treated and untreated wild-type and ∆katG cells from the iEK1011 Mtb genome-scale metabolic model. Reactions involved in folate biosynthesis are significantly decreased in BDQ-treated ∆katG cells relative to BDQ-treated wild-type cells. b Catalase activity-deficient ∆katG, mc²8245, and non-MDR INH-resistant TDR-TB-42 cells are sensitized to TMP relative to their wild-type, mc²7902, and INH-susceptible controls in 8-day growth inhibition dose-response experiments. MDR clinical strains were not sensitized to folate biosynthesis inhibition relative to INH-susceptible clinical strains. c Catalase activity-deficient ∆katG, mc²8245, and non-MDR INH-resistant TDR-TB-42 cells are sensitized to SMX relative to their wild-type, mc²7902, and INH-susceptible controls in 8-day growth inhibition dose-response experiments. MDR clinical strains were not sensitized to folate biosynthesis inhibition relative to INH-susceptible clinical strains. n = 10,000 flux samples were collected for each metabolic simulation. n = 3 biological replicates for experiments involving wild-type, ∆katG, mc²7902, and mc²8245 cells. n = 4 biological replicates for experiments involving non-MDR INH-resistant TDR-TB-42. n = 1 biological replicate for all other TDR-TB clinical strains. Error bars for INH-susceptible and MDR clinical strains computed across INH-susceptible and MDR clinical strains, respectively. Brown-Forsythe and Welch ANOVA statistical tests were performed on RNA expression data with comparisons between BDQ-treated and untreated cells wild-type and ∆katG cells or between untreated wild-type and ∆katG as indicated with FDR correction. **p ≤ 0.01, ****p ≤ 0.0001. Data depicted as mean ± SEM. Source data are provided in the Source Data file.