Metabolic regulation of mycobacterial growth and antibiotic sensitivity

Abstract

Treatment of chronic bacterial infections, such as tuberculosis (TB), requires a remarkably long course of therapy, despite the availability of drugs that are rapidly bacteriocidal in vitro. This observation has long been attributed to the presence of bacterial populations in the host that are "drug-tolerant" because of their slow replication and low rate of metabolism. However, both the physiologic state of these hypothetical drug-tolerant populations and the bacterial pathways that regulate growth and metabolism in vivo remain obscure. Here we demonstrate that diverse growth-limiting stresses trigger a common signal transduction pathway in Mycobacterium tuberculosis that leads to the induction of triglyceride synthesis. This pathway plays a causal role in reducing growth and antibiotic efficacy by redirecting cellular carbon fluxes away from the tricarboxylic acid cycle. Mutants in which this metabolic switch is disrupted are unable to arrest their growth in response to stress and remain sensitive to antibiotics during infection. Thus, this regulatory pathway contributes to antibiotic tolerance in vivo, and its modulation may represent a novel strategy for accelerating TB treatment.

Figure 1. Triglyceride synthesis mutants continue to replicate under growth-limiting conditions.

(A) The predicted TAG biosynthetic pathway of M. tuberculosis and its relationship to the TCA cycle. Mutations in the underlined genes were predicted by Transposon Site Hybridization to result in overrepresentation after hypoxia. OAA, oxaloacetate; MAG, monoacylglycerol; DAG, diacylglycerol. (B) Δtgs1 bacteria grow to a higher cell density in hypoxic cultures. (C) Δtgs1 mutants continue to replicate in hypoxic culture. The replication dynamics of the indicated strains were assessed by quantifying the rate at which unstable plasmid pBP10 was lost (right axis, open symbols). The “cumulative bacterial number” (left axis, closed symbols) represents the total number of organisms that would have been present if cell death was negated. Arrows in (B) and (C) indicate the initiation of hypoxia based on methylene blue decolorization. (D and E) Growth of M. tuberculosis strains at an initial pH of 5.5 (D) and in low iron medium (E). Optical density measurements are shown (similar data were obtained by quantifying CFU). Means ± SD of two independent experiments each performed in duplicate or triplicate are shown. Insets demonstrate the lack of TAG accumulation (upper species) in Δtgs1 bacteria, as assessed by thin layer chromatography. Each TLC was developed independently. In inset, “a,” H37Rv; “b,” Δtgs1; and “c,” complemented strain Δtgs1+pTGS1.

Figure 2. TAG synthesis modulates growth by consuming acetyl CoA.

(A) Oxaloacetate (OAA) stimulates bacterial growth in low iron medium. Growth of H37Rv expressing gfp was assessed in 384-well plates by fluorometry. Wells contained medium alone, succinate (SUC), fumarate (FUM), pyruvate (PYR), or oxaloacetate (OAA). Each metabolite was added at increasing concentrations (0.1, 0.5, 1, 2, and 5 mM). Fluorescence intensity of the plates was measured after 10 d of growth and normalized to control wells containing a Sybr green standard. (B and C) Growth of the indicated strains was assessed in hypoxic (B) or low iron cultures (C). “citA*” indicates citrate synthase overexpressing strain. The inset in (B) shows TAG accumulation by H37Rv (a) and the citA* strain (b) under hypoxic conditions. (D) Addition of tetrahydrolipostatin (THL) to low iron cultures inhibits growth in a tgs1-dependent manner. Growth inhibition was determined from the optical density cultures after 21 d. Means ± SD of two independent experiments each performed in duplicate or triplicate are shown. The inset shows TAG accumulation by H37Rv (a) and the tgs1 strain (b) with the highest tested concentration of THL. Lipid extracts were normalized to represent the same bacterial mass. (E) Radiolabeled acetate (¹⁴C_1,2) was introduced into hypoxic vials after 7 d of sealed culture. The left four bars indicate the cpm of CO₂ sampled from the headspace after 6 h. The right four bars indicate relative abundance of ¹⁴C-labeled TAG on the same time point. Inset shows a representative TLC plate that was quantified. Means ± SD of triplicate experiments are shown for CO₂ measurements.

Figure 3. Metabolic modulation reverses the antibiotic tolerance induced by low iron and hypoxic conditions.

Bacterial survival in the presence of the indicated antibiotics under hypoxic conditions (A, C, E, G) and in low iron media (B, D, F, H). Isoniazid (“INH”, 2 and 0.25 µg ml⁻¹, A and B), streptomycin (“SMP”, 2 and 1 µg ml⁻¹, C and D), ciprofloxacin (“CIP”, 4 and 1 µg ml⁻¹, E and F), and ethambutol (“EMB”, 5 and 3 µg ml⁻¹, G and H) were introduced into each culture. Antibiotics were added to the hypoxic vials after 14 d of culture. Means ± SD of two independent experiments each performed in duplicate are shown.

Figure 4. Modulating carbon fluxes reverses the antibiotic tolerance induced during infection.

Mice were infected via the aerosol route with the indicated bacterial strains. Total bacterial burden in the spleens (A, C, E) and lungs (B, D, F) is shown. Mice were treated at the indicated times with isoniazid (“INH”, A, B), ethambutol (“EMB”, C, D), or isoniazid plus pyrazinamide (“INH+PZA”, E, F). Dotted line represents the detection limit of the experiment. “ND” indicates no colonies detected. ND* indicates two colonies were detected but neither retained the citA overexpression plasmid. Means ± SD from three to five mice are shown.