The protonmotive force is required for maintaining ATP homeostasis and viability of hypoxic, nonreplicating Mycobacterium tuberculosis

Respiratory Functions Are Essential for Maintaining the Viability of Hypoxic Nongrowing Mycobacteria.

Previous work suggested that in vitro grown nonreplicating tubercle bacilli have a reduced susceptibility to the current anti-TB drugs (6). However, a systematic comparison of the cidal activity of anti-TB drugs against exponentially growing vs. quiescent bacilli has never been carried out. The minimum bactericidal concentration 90% (MBC_90s; concentrations that kill 90% of an exponentially growing population) and the Wayne Cidal Concentration 90% (WCC_90s; concentrations that kill 90% of a quiescent population) were then determined for the major first- and second-line anti-TB drugs. Table 1 shows that all anti-TB drugs tested were indeed much less active against hypoxic nonreplicating Mtb cells, compared with replicating bacteria. Rifampicin and moxifloxacin were 50-fold less cidal, whereas isoniazid (INH) and ethionamide were totally inactive against nonreplicating Mtb. The observed drastic loss in cidal activities against quiescent bacilli is consistent with the concept that the macromolecular synthesis pathways targeted by the current TB drugs are not essential in the nongrowing form of the bacilli (6).

Recent microarray studies showed that key complexes of the electron transport chain (ETC) are down-regulated in nongrowing hypoxic or nutrient-starved Mtb (8, 9, 12, 14). We reasoned that, because the respiratory functions and the factors involved in energy production are maintained at a low level, they may represent a point of metabolic vulnerability in nonreplicating Mtb cells. Using various inhibitors, we have addressed the requirement of the protonmotive force (PMF), the F₀F₁ ATP synthase, and the electron donors NADH-dehydrogenase (ndh)-1 and -2 in quiescent hypoxic Mtb. The MBC_90s and WCC_90s were determined for the recently described F₀F₁ ATP synthase inhibitors R207910 (15) and N,N′-dicyclohexylcarbodiimide (DCCD), for the ionophores valinomycin and nigericin, which cause dissipation of membrane potential (ΔΨ) and proton gradient (ΔpH), respectively, and for rotenone and thioridazine, which inhibit ndh-1 and -2, respectively (16–19). Table 1 shows that all of the compounds but rotenone, killed quiescent Mtb. Similar results were obtained by using M. bovis bacillus Calmette–Guérin (data not shown). More importantly, the cidal activities of these compounds were found to be higher against quiescent bacilli than against growing organisms (Table 1). These observations indicate that the respiratory functions and the mechanisms involved in ATP production represent a point of metabolic vulnerability in nongrowing mycobacteria.

De Novo ATP Synthesis Is Essential for Survival of Hypoxic Nongrowing Mtb.

To address directly the energetic status of hypoxic nongrowing mycobacteria, the intracellular ATP concentration was measured and compared with growing aerobic mycobacteria. The ATP content was found to be 5-fold lower in hypoxic nonreplicating mycobacteria (0.44 ± 0.22 attomoles/cfu and 2.50 ± 0.40 attomoles/cfu, respectively). To gain further insight into the maintenance of the ATP homeostasis in hypoxic nonreplicating Mtb, we monitored the ATP concentration over a 25-day period in the Wayne dormancy model. The intracellular ATP concentration progressively reduced as the oxygen tension dropped (Fig. 1). The ATP content was at a minimum in early NRP phase (15 days of the Wayne model) and then remained constant throughout the stationary phase (Fig. 1). To determine whether the respiratory F₀F₁ ATP synthase is required to maintain this reduced ATP level, the intracellular ATP content was measured in quiescent mycobacteria exposed to the F₀F₁ ATP synthase inhibitor R207910 (15, 20). We found that R207910 reduced the ATP level in a dose-dependent manner (Fig. 2A), and that this decrease in intracellular ATP level correlated with an increased killing activity (Fig. 2B). A similar effect was observed when hypoxic nongrowing Mtb was incubated with DCCD, another respiratory ATP synthase inhibitor (Fig. 2). The drop in ATP observed with R207910 and DCCD was specific, because none of the first- and second-line TB drugs tested in this model had any significant effect on the ATP level [Fig. 2A and supporting information (SI) Table S1, SI Text]. These results show that a reduced ATP level is maintained in hypoxic nonreplicating mycobacteria, and that de novo ATP synthesis is required for the maintenance of Mtb survival.

Hypoxic Nongrowing Bacilli Maintain Their Membrane Energized.

The F₀F₁ ATP synthase utilizes the PMF to drive ATP production. The two components contributing to the PMF are the membrane potential (ΔΨ) and the transmembrane proton concentration gradient (ΔpH). The chemical probing shown in Table 1 demonstrated that the ionophores dissipating the ΔΨ (valinomycin) and ΔpH (nigericin) are cidal on quiescent bacilli, suggesting that the membrane of nonreplicating bacilli is energized and that this potential energy (the PMF) is required for ATP production and survival of the bacilli. To confirm the membrane energization, the actual ΔΨ and ΔpH were measured in growing and quiescent organisms.

The ΔΨ was first detected by using the cationic fluorescent dye DiOC₂ (3, 21) in aerobic replicating and hypoxic nongrowing Mtb. Results showed that ΔΨ was similar in intensity in both cell populations (Fig. 3). To confirm this result, the ΔΨ was then measured with [³H]tetraphenylphosphonium bromide (TPP⁺), as described (22–24). Results showed that ΔΨ in growing and nonreplicating hypoxic cells was identical (Table 2), confirming that quiescent Mtb are able to maintain their membranes polarized to the same extent as growing cells. In addition, it was possible to depolarize the membranes of hypoxic nongrowing cells by using the ionophore valinomycin (known to dissipate the ΔΨ but not ΔpH), suggesting that the maintenance of the ΔΨ is an active process (Fig. 2A and Table 2). Results obtained with valinomycin were specific, because the ionophore nigericin (dissipates the ΔpH but not the ΔΨ) or rifampicin did not affect the ΔΨ value of the cells.

The ΔpH was measured by using [¹⁴C]benzoic acid as a pH probe (23, 24). We found a comparable ΔpH in replicating and hypoxic nongrowing bacilli (Table 2). The dissipation of the ΔpH was observed when the cells were exposed to nigericin but not to rifampicin or valinomycin (Table 2). Because the PMF is usually used by the F₀F₁ ATP synthase for ATP production, the ATP level was measured in hypoxic nonreplicating Mtb in the presence of nigericin or valinomycin. The results show a reduction of ATP level in a dose-dependent manner that was associated with a loss of viability (Fig. S1). Altogether, our observations demonstrated that the membrane of hypoxic nonreplicating Mtb is fully energized and that both components of the PMF, ΔΨ and ΔpH, are required for ATP synthesis.

ndh-2 Acts as an Electron Donor for the Anaerobic ETC and Is Essential for the Maintenance of the PMF.

The PMF is usually generated and maintained by the ETC. The ndh is described as one of the electron donors to initiate ETC through NADH oxidation (25). We have observed that the ndh-2 inhibitor thioridazine displayed a significant cidal activity against quiescent bacilli, whereas the ndh-1 inhibitor rotenone was not active (Fig. 4B). The ATP level was also significantly reduced upon thioridazine treatment whereas rotenone had no effect (Fig. 4A and Table S1). These results suggested that ndh-2, but not -1, acts as the electron donor for the anaerobic ETC in Mtb. The essentiality of ndh-2 in hypoxic nonreplicating Mtb was confirmed by using trifluoperazine, another ndh-2-specific inhibitor (17). As observed for thioridazine, trifluoperazine was cidal for quiescent bacilli, with a WCC₉₀ of 20 μM (data not shown). In contrast, exposure of nongrowing bacilli to another ndh-1 inhibitor, Piericidin A, did not show any effect on intracellular ATP level and viability (data not shown). To rule out the possibility that Rotenone and Piericidin A were not active because of a limited cellular penetration, a Mtb strain deficient for the expression of ndh-1 activity was constructed in which the entire nuoA-N operon was deleted. No survival defect was observed for this mutant when grown in hypoxic conditions, demonstrating that ndh-1 is not essential in quiescent mycobacteria (data not shown).

We further tested the role of ndh-2 in the generation of the PMF. We showed that inhibition of ndh-2 with thioridazine resulted in a dose-dependent dissipation of the ΔΨ, therefore linking the ndh-2 activity to the generation of the PMF (Fig. 5). When tested under similar experimental conditions, rifampicin and the R207910 compound had no effect on the PMF (Fig. 5).

ndh-2 Is Required to Replenish the [NAD⁺] Pool in Hypoxic Nonreplicating Mtb.

Because [NAD⁺] is the main cellular oxidant, the restoration of the [NAD⁺] pool is essential for the maintenance of cellular functions. The mechanisms by which hypoxic nonreplicating mycobacteria replenish their [NAD⁺] pool are still unknown. The involvement of isocitrate dehydrogenase and a putative glycine dehydrogenase has been proposed (3) but not yet demonstrated. Because ndh-2 initiates the anaerobic ETC, we hypothesized this dehydrogenase might be involved in the restoration of the [NAD⁺] pool in hypoxic quiescent mycobacteria. [NAD⁺] and [NADH] concentrations were measured in hypoxic nongrowing mycobacteria 3 h after addition of the ndh-2 inhibitor thioridazine. Results show that, compared with the control group, the [NAD⁺] concentration significantly decreased, whereas the [NADH] concentration increased in thioridazine-treated mycobacteria (Table 3), resulting in a [NADH]/[NAD⁺] ratio 2.65-fold higher than that measured in the nontreated group (Table 3). The effect was specific to ndh-2-inhibitors because treatment with rifampicin or with R207910 compound did not change the [NADH]/[NAD⁺] ratio significantly (Table 3). These results therefore demonstrate the importance of ndh-2 in replenishing the pool of [NAD⁺] in hypoxic nonreplicating mycobacteria.

Chemicals.

Moxifloxacin was obtained from Chempacific. All other compounds were obtained from Sigma-Aldrich. The compound R207910 was synthesized as described earlier (15).

[³H]water, [³H]tetra phenyl phosphonium (TPP⁺), [¹⁴C]benzoic acid, and [¹⁴C]sucrose were obtained from GE Healthcare radiochemicals division.

Strains and Media.

Mtb H37Rv (ATCC #27294) was maintained in Dubos liquid medium. Hypoxic nonreplicating Mtb were generated under conditions of slow stirring in sealed tubes, as described (6, 7).

Determination of Bactericidal Concentrations.

The MBC₉₀ was defined as the minimal concentration of antibiotics that killed 90% of the initial inoculum in 5 days. Approximately 1 × 10⁶ cfu/ml were exposed to varying concentrations of chemicals in sterile 96-well flat bottom plates for 5 days aerobically. cfu was estimated on days 0 and 5 of drug exposure by plating the bacteria on 7H11-agar plates. A 90% reduction in cfu on day 5 compared with day 0 was the MBC₉₀.

The WCC₉₀ was defined as the concentration of antibiotics that killed 90% of hypoxic nonreplicating Mtb in 5 days. Hypoxic nonreplicating mycobacteria were exposed to varying concentration of antibiotics in sterile 96-well flat bottom plates for 5 days under anaerobic conditions. cfu were enumerated at days 0 and 5. Metronidazole (MTZ) at 400 μM and INH at 25 μM were used in all experiments as positive and negative controls, respectively. A 90% reduction in cfu on day 5 of drug exposure compared with day 0 was the WCC₉₀.

Quantification of Intracellular ATP.

Intracellular ATP was quantified by using the BacTiter-Glo Microbial Cell Viability Assay Kit (Promega). Aliquots of 100 μl of bacteria were collected at various time points and immediately heat-inactivated. Twenty-five microliters of cell lysates were transferred into 96-well white plates, mixed with an equal volume of the BacTiter-Glo reagent and incubated for 5 min in the dark. The emitted luminescence was detected by using a luminometer (Safire², Tecan Instruments) and was expressed as relative luminescence units. ATP standards ranging from 0.1 to 100 nM were also included in all of the experiments as internal control.

Detection and Measurement of the Membrane Potential.

The ΔΨ was detected by using BacLight bacterial membrane potential kit (Invitrogen). Briefly, Mtb cells were diluted to an OD₆₀₀ of ≈0.005 (≈10⁶ cfu/ml) in fresh Dubos liquid medium. Hypoxic nonreplicating Mtb were diluted in degazed Dubos liquid medium to maintain anaerobiosis. One milliliter of diluted cells was stained with 10 μl of 3 mM DiOC₂ (3) (3,3′-diethyloxa-carbocyanine iodine, fluorescent dye) for 15–30 min at 37°C. After staining, the cells were centrifuged and resuspended with an equal volume of 2% paraformaldehyde. The stained cells were analyzed by FACS by using 488-nm excitation and emission filters suitable for fluorescein and Texas red dye (21).

The ΔΨ was measured as described (22–24). Briefly, similar concentrations of aerobic and hypoxic Mtb cells (≈1 × 10⁹ cfu/ml) were incubated with 10 μM [³H]TPP⁺ (380 mCi/mmol) for 30 min at 37°C. The bacteria were then centrifuged at 9,300 × g for 5 min. The supernatant was transferred to an Eppendorf tube and the pellet resuspended in 100-μl aliquot of perchloric acid 20% (vol/vol). Aliquots of 100 μl of both cell pellet and supernatant were added with equal volume of scintillation mixture and the radioactivity was measured with a scintillation counter (GE Healthcare). The membrane potential was calculated according to the method of Rottenberg (23).

Measurement of the Intracellular pH.

Mtb cells were centrifuged and resuspended in Dubos liquid medium with varying pH to a cell density of ≈1 × 10⁹ cfu/ml. [¹⁴C]benzoic acid was used as a pH probe, [³H] water and [¹⁴C] sucrose were used to calculate the internal cell volume. Mycobacteria were incubated with 1 μCi/ml [¹⁴C] benzoic acid for 30 min at 37°C. The cells were centrifuged, and the radioactivity in pellet and supernatant was measured as described above. The internal pH was calculated as described (23, 24).

Determination of [NADH] and [NAD⁺] Concentrations.

[NADH] and [NAD⁺] were extracted and quantified as described (37). Mtb cells were preincubated either with or without compounds for 3 h before extraction of [NAD⁺] and [NADH] under anaerobic conditions.

The intracellular [NADH] and [NAD⁺] concentrations were measured with a cycling assay, as described (38, 39).

Genetic Deletion Mycobacteria.

The nuoA-N gene was disrupted by allelic exchange method by using the pYUB854 plasmid (40). A sacB-lacZ cassette was excised from the pGOAL17 (41) and ligated into the PacI site of pYUB854 containing 5′ and 3′ flank of the nuoA-N locus. The final plasmids were UV-irradiated (42) before electroporation into Mtb.