A Novel Small-Molecule Inhibitor of the Mycobacterium tuberculosis Demethylmenaquinone Methyltransferase MenG Is Bactericidal to Both Growing and Nutritionally Deprived Persister Cells

Abstract

Active tuberculosis (TB) and latent Mycobacterium tuberculosis infection both require lengthy treatments to achieve durable cures. This problem has partly been attributable to the existence of nonreplicating M. tuberculosis "persisters" that are difficult to kill using conventional anti-TB treatments. Compounds that target the respiratory pathway have the potential to kill both replicating and persistent M. tuberculosis and shorten TB treatment, as this pathway is essential in both metabolic states. We developed a novel respiratory pathway-specific whole-cell screen to identify new respiration inhibitors. This screen identified the biphenyl amide GSK1733953A (DG70) as a likely respiration inhibitor. DG70 inhibited both clinical drug-susceptible and drug-resistant M. tuberculosis strains. Whole-genome sequencing of DG70-resistant colonies identified mutations in menG (rv0558), which is responsible for the final step in menaquinone biosynthesis and required for respiration. Overexpression of menG from wild-type and DG70-resistant isolates increased the DG70 MIC by 4× and 8× to 30×, respectively. Radiolabeling and high-resolution mass spectrometry studies confirmed that DG70 inhibited the final step in menaquinone biosynthesis. DG70 also inhibited oxygen utilization and ATP biosynthesis, which was reversed by external menaquinone supplementation. DG70 was bactericidal in actively replicating cultures and in a nutritionally deprived persistence model. DG70 was synergistic with the first-line TB drugs isoniazid, rifampin, and the respiratory inhibitor bedaquiline. The combination of DG70 and isoniazid completely sterilized cultures in the persistence model by day 10. These results suggest that MenG is a good therapeutic target and that compounds targeting MenG along with standard TB therapy have the potential to shorten TB treatment duration.IMPORTANCE This study shows that MenG, which is responsible for the last enzymatic step in menaquinone biosynthesis, may be a good drug target for improving TB treatments. We describe the first small-molecule inhibitor (DG70) of Mycobacterium tuberculosis MenG and show that DG70 has characteristics that are highly desirable for a new antitubercular agent, including bactericidality against both actively growing and nonreplicating mycobacteria and synergy with several first-line drugs that are currently used to treat TB.

FIG 1

PcydAB induction identifies DG70 as a candidate respiratory inhibitor. (A) The relative induction of the PcydAB::mWasabi reporter as measured by determining the relative fluorescence units of each drug-treated culture compared to the DMSO-only control is shown for several anti-TB drugs as well as DG70. Each was tested in triplicate in a 96-well format. Isoniazid (INH), rifampin (RIF), and ethambutol (EMB) were used as negative controls; thioridazine (THZ), bedaquiline (BDQ), and PA824 were used as positive controls (21–23). Means and standard deviations from three independent experiments are shown. DG70 consistently induced Pcyd 1.7-fold over DMSO. The Student t test was used to determine statistical significance between DMSO and DG70. *, P ≤ 0.05. (B) Molecular structure of DG70.

FIG 2

Dose response of DG70 against M. tuberculosis mc²6206 in J774A.1 murine macrophages. J774A.1 macrophages were infected with 10⁴ CFU/ml of M. tuberculosis (T0, red bar) and treated with DG70 at final concentrations of 0.5, 1, 2.5, 5, and 10× its MIC. Macrophages treated with either bedaquiline (BDQ) or isoniazid (INH) at 10× MIC were used as positive controls. CFU were determined after 3 days of treatment and were back calculated to CFU per milliliter. T0 shows the initial CFU used for infecting macrophages. Means and standard deviations from three independent experiments are shown. The Student t test was used to determine statistical significance between DMSO and different treatments. **, P ≤ 0.01.

FIG 3

Predicted binding modes of DG70 with models of MenG. (A and B) Multitemplate-based homology model of MenG from MODELLER. (C and D) Model made by threading MenG onto the known class, architecture, topology, and homology domains in PsiPRED. (A and D) Ribbon representation of these two MenG models. The residues that mutate in M. tuberculosis upon DG70 treatment are shown according to the Corey, Pauling, and Koltun scheme with blue carbon atoms, and residues that mutate in BCG are displayed with purple carbon atoms. Two representative docked modes of DG70 are presented in panels B and D as ball-and-stick models with dark green carbon atoms.

FIG 4

Growth inhibition of mycobacteria overexpressing wild-type and mutant menG. BCG strains constitutively overexpressing wild-type M. tuberculosis menG and DG70-resistant M. tuberculosis menG mutants containing either F118L or V20A substitutions were incubated with DG70 at the indicated concentrations for 7 days, after which each culture was plated and CFU were counted after 4 weeks of incubation at 37°C. CFU counts were back calculated to CFU per milliliter. Means and standard deviations from three independent experiments are shown. The Student t test was used to determine statistical significance between DMSO and different treatments. ***, P ≤ 0.01; ns, no significant difference.

FIG 5

Effect of DG70 on menaquinone biosynthesis. (A) Schematic showing conversion of demethylmenaquinone (DMK9) to menaquinone (MK9) by MenG. SAH, S-adenosyl-l-homocysteine. (B) TLC analysis of neutral lipids isolated from BCG labeled with l-[methyl-¹⁴C]methionine after DG70, isoniazid (INH), or bedaquiline (BDQ) treatment. Menaquinone [MK-9(II-H₂)] was identified by comigration with an authentic standard as observed under UV light. (C and D) Ratios of MK9/DMK9 present in neutral lipid extracts of wild-type BCG (C) and the DG70-resistant BCG mutant (BCG 70B1 MenG-A60V) (D) cultures treated with the indicated compounds as detected by high-resolution mass spectrometry. Means and standard deviations from three independent experiments are shown. The Student t test was used to determine statistical significance between DMSO and different treatments. **, P ≤ 0.01; ns, no significant difference.

FIG 6

DG70 inhibits oxygen consumption and ATP synthesis in mycobacteria. (A) Effect of DG70 on oxygen consumption. Mycobacterial cells containing 0.01% methylene blue were treated with thioridazine (THZ), isoniazid (INH), or DG70 at the indicated concentrations (×MIC). Cultures were supplemented with MK4 (400 µM) where indicated. Decolorization of methylene blue, as an indicator of oxygen consumption, was monitored at 665 nm. (B) The ATP levels of 1 ml of cellular lysate of BCG (OD₅₉₅ = 0.4) after 3 days of treatment with the indicated compounds are shown. MK4 (400 µM) was added as a supplement as indicated.

FIG 7

Activity of DG70 against nonreplicating M. tuberculosis. M. tuberculosis starved for 6 weeks in PBS (nutritionally deprived persistence model) was exposed to various concentration of DG70 and 10× MIC of isoniazid (INH), bedaquiline (BDQ), PA824, thioridazine (THZ), and rifampin (RIF) for 7 days and then plated to determine viability. Menaquinone (MK4) at a 400 µM concentration was used to test for reversal of the killing mediated by DG70 in the nutrient-deprived M. tuberculosis. Experiments were carried out in duplicate, and each dilution was plated in triplicate. CFU counts were back calculated to CFU per milliliter. Results are shown as means and standard deviations.

FIG 8

In vitro bactericidal activity of DG70 and other anti-TB drugs. (A) Mid-log-phase cultures of BCG were incubated with the indicated compounds and compound combinations at 10× MIC of each compound. Aliquots of each culture were plated for CFU determination at the indicated time point and back calculated to CFU per milliliter. Dashed lines correspond to combination treatments with two drugs. (B) Highlight of panel A showing the time-kill kinetics of DG70 and PA824 alone and in combination. (C) Highlight of panel A showing the time-kill kinetics of DG70 and bedaquiline (BDQ) alone and in combination. (D) Highlight of panel A showing the time-kill kinetics of DG70 and isoniazid (INH) alone and in combination. Dimethyl sulfoxide (DMSO) was used as a vehicle to dissolve the compound. Means and standard deviations from three independent experiments are shown.