Loss of glycerol catabolism confers carbon-source-dependent artemisinin resistance in Mycobacterium tuberculosis

Abstract

In view of the urgent need for new antibiotics to treat human infections caused by multidrug-resistant pathogens, drug repurposing is gaining strength due to the relatively low research costs and shorter clinical trials. Such is the case of artemisinin, an antimalarial drug that has recently been shown to display activity against Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis. To gain insight into how Mtb is affected by artemisinin, we used RNAseq to assess the impact of artemisinin on gene expression profiles, revealing the induction of several efflux pumps and the KstR2 regulon. To anticipate the artemisinin resistance-conferring mutations that could arise in clinical Mtb strains, we performed an in vitro evolution experiment in the presence of lethal concentrations of artemisinin. We obtained artemisinin-resistant isolates displaying different growth kinetics and drug phenotypes, suggesting that resistance evolved through different pathways. Whole-genome sequencing of nine isolates revealed alterations in the glpK and glpQ1 genes, both involved in glycerol metabolism, in seven and one strains, respectively. We then constructed a glpK mutant and found that loss of glpK increases artemisinin resistance only when glycerol is present as a major carbon source. Our results suggest that mutations in glycerol catabolism genes could be selected during the evolution of resistance to artemisinin when glycerol is available as a carbon source. These results add to recent findings of mutations and phase variants that reduce drug efficacy in carbon-source-dependent ways.

Keywords: Mycobacterium tuberculosis; artemisinin; drug resistance mechanisms; glpK; glycerol kinase; glycerol metabolism.

Fig 1

Disruption of glpK impairs growth and reduces susceptibility to artemisinin in glycerol-supplemented media. (A) Scheme of the evolution experiment for acquiring artemisinin-resistant mutants. Mtb log phase cultures were brought to an OD_600nm= 0.1 and artemisinin was added at a concentration of 150, 300, or 600 µg/mL. Those cultures in which growth was observed were diluted in fresh 7H9 to an OD_600nm= 0.01, grown to OD_600nm= 0.8 and the cycle was repeated three times. (B) Growth kinetics of artemisinin-resistant isolates in 7H9 medium. Colors denote slower (pink), slightly lower (red), or similar (dark red) growth rates compared to the parental strain (black). (C) Growth of artemisinin-resistant isolates and the parental strain after 6 days of incubation with 150 µg/mL artemisinin. Error bars show standard deviation among replicates. (D) Scheme showing the mutations detected by WGS within the glpK gene. The region containing the homopolymeric tract of guanines is highlighted in gray. (E) CFUs relative to day 0 for WT, ΔglpK, and ΔglpK::glpK in 7H9 or minimal medium (MM) with the indicated carbon sources, with and without 150 µg/mL artemisinin (AN).