The anti-mycobacterial activity of the cytochrome bcc inhibitor Q203 can be enhanced by small-molecule inhibition of cytochrome bd

Mycobacterial energy metabolism currently attracts strong attention as new target space for development of anti-tuberculosis drugs. The imidazopyridine Q203 targets the cytochrome bcc complex of the respiratory chain, a key component in energy metabolism. Q203 blocks growth of Mycobacterium tuberculosis at nanomolar concentrations, however, it fails to actually kill the bacteria, which may limit the clinical applicability of this candidate drug. In this report we show that inhibition of cytochrome bd, a parallel branch of the mycobacterial respiratory chain, by aurachin D invoked bactericidal activity of Q203. In biochemical assays using inverted membrane vesicles from Mycobacterium tuberculosis and Mycobacterium smegmatis we found that inhibition of respiratory chain activity by Q203 was incomplete, but could be enhanced by inactivation of cytochrome bd, either by genetic knock-out or by inhibition with aurachin D. These results indicate that simultaneously targeting the cytochrome bcc and the cytochrome bd branch of the mycobacterial respiratory chain may turn out as effective strategy for combating M. tuberculosis.

a promising candidate TB drug and this compound currently is evaluated in phase 1 clinical trials. However, it has been reported that disabled assembly of cytochrome bcc in M. tuberculosis or genetic knock-out of cytochrome bcc in Mycobacterium smegmatis did not completely abolish bacterial growth 14,15 . In these mutants, network adaptation in the respiratory chain can lead to induction of cytochrome bd 14 , which constitutes an alternative branch of the respiratory chain and has been implied in the bacterial defense against a variety of stresses [16][17][18][19][20][21][22] . In mycobacteria, cytochrome bd is involved in the defense against hypoxia 23 , cyanide 23 , hydrogen peroxide 15,24 , nitric oxide 15,25 , and a variety of antibacterials including BDQ 24,[26][27][28] . Cytochrome bd also facilitates metabolic adaptation of certain M. tuberculosis laboratory strains, including the reference strain H37Rv, to imidazopyridine-type cytochrome bcc inhibitors 29 . These adapted strains displayed considerably elevated minimal inhibitory concentrations (MICs) for Q203, effectively evading growth inhibition by these drugs 29 . Upon knock-out of cytochrome bd the susceptibility for growth inhibition by Q203 was restored 29 . In in vitro time kill kinetics experiments Q203 acted bacteriostatic against M. tuberculosis H37Rv, even when applied at concentrations of 200-300 × MIC 30,31 . However, a recent report showed that Q203 exhibited bactericidal activity against an M. tuberculosis bd-KO strain in vitro and in a mouse infection model 31 . The adaptability of M. tuberculosis strains and the lack of bactericidal activity may significantly diminish the suitability of the cytochrome bcc complex as antibiotic target and restrict the clinical applicability of Q203 as TB drug. It has been proposed that simultaneously targeting both branches of the mycobacterial respiratory chain might be required to effectively disrupt respiration in M. tuberculosis 15,24,31 .
In this report, we explore if small-molecule inhibition of cytochrome bd can enhance the activity of a cytochrome bcc inhibitor, Q203, against M. tuberculosis.

Results
Small-molecule inhibition of cytochrome bd stimulates Q203. In line with previously reported results 30 , treatment of the M. tuberculosis H37Rv strain used in our laboratory with Q203 resulted in only a marginal decrease of colony forming units (Supplementary Figure 1). We also confirmed that Q203 acted bactericidal against an M. tuberculosis strain lacking cytochrome bd (Supplementary Figure 1), as described recently 31 . Next, we set out to explore if inactivation of cytochrome bd and concomitant enhancement of Q203 activity can also be achieved by a small-molecule inhibitor. For this purpose we determined the activity of aurachin D against M. tuberculosis. Aurachin D has previously been described as inhibitor of cytochrome bd in isolated cytoplasmic membranes from Escherichia coli 32 , Synechocystis PCC6803 33 and M. smegmatis 24 . Aurachin D did not effectively inhibit growth of M. tuberculosis when applied alone, with a minimal inhibitory concentration for inhibition of growth (MIC 90 ) >100 μM (Table 1), likely reflecting the non-essentiality of cytochrome bd in M. tuberculosis under standard culture conditions. However, addition of aurachin D considerably enhanced growth inhibition of M. tuberculosis by Q203. The MIC decreased from 10 nM for Q203 when applied alone to 1.25 nM for Q203 in combination with aurachin D (25 μM) ( Table 1). The impact of aurachin D on growth inhibition by Q203 mirrored the effect achieved by genetic inactivation of cytochrome bd (Table 1).
Next, we characterized if addition of aurachin D can invoke bactericidal activity of Q203. In kill kinetics experiments, aurachin D alone did not decrease bacterial counts within 21 days. However, addition of aurachin D converted the bacteriostatic activity of Q203 (30 × MIC) into bactericidal activity (Fig. 1A). The enhancement of Q203 activity by aurachin D was dose-dependent, with >2 log 10 units decrease of colony forming units (cfu) triggered by 25 μg/ml aurachin D (Fig. 1B). These results demonstrate that a cytochrome bd inhibitor can considerably stimulate the impact of a cytochrome bcc-targeting companion drug.

Inhibition of respiratory chain activity by Q203 is incomplete but can be enhanced by aurachin D.
Next, we evaluated the ability of Q203 to inhibit its target, the cytochrome bcc complex. For safety reasons these experiments were performed with the strongly attenuated M. tuberculosis strain mc 2 6020 34 . Q203 inhibited oxygen consumption activity of inverted membrane vesicles (IMVs) from M. tuberculosis strain 6020 in a dose-dependent manner, with an IC 50 of ~20 nM ( Fig. 2A). However, inhibition of respiratory chain activity by Q203 was incomplete, with ~60% inhibition observed at the highest Q203 concentration tested (10 μM) ( Fig. 2A). These results reveal that Q203 has high affinity for its target, but indicate that a considerable part of respiratory electron flow can be re-directed away from the cytochrome bcc complex. We then evaluated if aurachin D can complement inhibition of respiratory chain activity by Q203. Aurachin D alone displayed dose-dependent inhibition of respiratory chain activity, with maximal inhibition of ~60% at 25 μM and an IC 50 of ~400 nM (Fig. 2B). The combination of Q203 (10 μM) with aurachin D (400 nM) displayed significantly higher inhibition than 10 μM Q203 alone (Fig. 2C). Inhibition by this Q203/aurachin D combination was also significantly higher than the maximal inhibition achievable with Q203 alone under these conditions (assessed by one-site saturation-binding model y = 70.93×/0.009 + x, data not shown)(P < 0.050). This enhanced effect found for the Q203/aurachin D combination in the biochemical assay may explain why genetic or chemical inactivation of cytochrome bd can augment inhibition of bacterial growth and trigger bacterial killing by Q203. Inhibition of respiratory chain activity in M. smegmatis by Q203. Interestingly, the respiratory chain activity of IMVs isolated from M. smegmatis, a fast-growing mycobacterial strain that is not sensitive to growth inhibition by Q203 (MIC > 50 μM) 12 , was also efficiently blocked by Q203 (Fig. 3A, black bars). The affinity of Q203 for the cytochrome bcc complex in M. smegmatis (IC 50 ~ 20 nM) was comparable to the affinity for M. tuberculosis cytochrome bcc in the employed assay system, although M. tuberculosis and M. smegmatis vastly differ in drug susceptibility. These results demonstrate that the lack of growth inhibition found for M. smegmatis is not caused by insufficient affinity of Q203 for the cytochrome bcc complex in M. smegmatis. As observed for M. tuberculosis, inhibition of respiratory chain activity of M. smegmatis IMVs by Q203 was incomplete (max. inhibition 50-80%, depending on membrane batch). IMVs isolated from an M. smegmatis strain lacking the cytochrome bcc complex 14 did not show significant inhibition of respiratory chain activity by Q203 (Fig. 3A, red bars), demonstrating the drug's target specificity. In contrast, IMVs from a M. smegmatis strain lacking cytochrome bd 23 displayed 100% maximal inhibition by Q203 (Fig. 3A, grey bars), further supporting the interpretation that cytochrome bd can partially compensate for inactivation of the cytochrome bcc complex. As observed for M. tuberculosis, the 10 μM Q203/400 nM aurachin D combination showed significantly stronger inhibition of M. smegmatis wild-type respiratory chain activity than 10 μM Q203 alone (Fig. 3C). Consistent with these results, we found that genetic inactivation of cytochrome bd in M. smegmatis 23

Discussion
Development of new anti-tuberculosis drugs is urgently needed in order to combat multi-drug resistant strains of M. tuberculosis. For effective antibacterial compounds bactericidal instead of bacteriostatic activity is highly desirable. Absence of bactericidal activity can be regarded as an argument against further consideration of a drug candidate. As an example, the development of a new imidazopyridine sub-class, the imidazopyridine ethers, was not further pursued, in part based on lack of bactericidal activity of these compounds 35 . Q203 efficiently blocks growth of M. tuberculosis at nanomolar concentrations, however, this drug acts bacteriostatic on M. tuberculosis and lacks bactericidal activity. Upon chemical inhibition or genetically knockout of the cytochrome bcc branch, respiratory electron transport through the alternate cytochrome bd terminal oxidase alone may be sufficient to maintain mycobacterial viability. Targeting both branches of the respiratory chain may be required for effective shutdown of mycobacterial energy conversion and concomitantly for killing M. tuberculosis 15,31 . As proof-of-concept for this strategy, we here showed that inactivation of cytochrome bd by a small-molecule inhibitor or by genetic modification can turn the bacteriostatic activity of Q203 into bactericidal activity.
Our results reveal high affinity of Q203 for the cytochrome bcc complex, but inhibition of respiratory chain activity in M. tuberculosis is incomplete. We regard it as likely that imid(azo)pyridines and structurally related compounds 12,29,[35][36][37][38][39][40][41][42] as well as structurally not related compounds such as lansoprazol 43 that share the cytochrome bcc complex as target also share incomplete respiratory chain inhibition and lack of bactericidal activity against M. tuberculosis. One study published during the peer-review/revision process of our work revealed incomplete growth inhibition and bacteriostatic activity for the phenoxoyalkyl benzimidazoles, which are hypothesized to target the cytochrome bcc complex in M. tuberculosis 44 .
Inhibiting the catalytic activity of cytochrome bd can contribute to efficient killing of M. tuberculosis. Interestingly, the opposite approach, killing M. tuberculosis based on activation of cytochrome bd, has recently been suggested for the triple drug combination Q203/BDQ/clofazimine. The strong bactericidal activity of this combination was attributed to increased cytochrome bd-mediated respiratory electron flux upon inhibition of cytochrome bcc by Q203, thereby facilitating production of reactive oxygen species by clofazimine 30 . Dysregulation of cytochrome bd function represents an efficient strategy for weakening the defense of M. tuberculosis and apparently can be achieved either way, by inhibition or by activation of this survival factor.

Materials and Methods
Chemicals. Aurachin D was synthesized as described earlier 45   Valerie Mizrahi (University of Cape Town). Replicating bacterial cultures were grown in Middlebrook 7H9 broth (Difco) supplied with 0.05% Tween-80 and 10% Middlebrook albumin dextrose catalase enrichment (BBL) at 37 °C with shaking. If applicable, 50 μg/mL kanamycin or 50 μg/mL hygromycin was added to the medium to select for mutant strains. The attenuated M. tuberculosis strain mc 2 6020 34 was kindly provided by Dr. William R. Jacobs, Jr. (Albert Einstein College of Medicine). Bacterial culture were grown in 7H9 medium (Difco) supplemented with 10% (vol/vol) OADC enrichment (oleic acid-albumin-dextrose-catalase, Difco), 0.05% Tween-80, 0.2% Casaminoacids, 0.24 mg/ml Pantothenate and 0.8 mg/ml L-lysine. Culture and handling of strains were done in a Biological Safety Level 3 laboratory for M. tuberculosis H37Rv and at Biological Safety Level 2 for M. tuberculosis mc 2 6020 and the M. smegmatis strains.
Determination of MICs. The resazurin microtiter assay (REMA) plate method was performed in 7H9 medium containing 10% (vol/vol) OADC enrichment (oleic acid-albumin-dextrose-catalase, Difco), 0.05% Tween-80. If applicable, 50 μg/mL hygromycin was added to the medium to select for mutant strains. Q203 and aurachin D solutions were thawed and diluted in the 7H9 medium. Serial two fold dilutions of each drug in 100 µl of 7H9 medium were prepared directly in 96-well plates. Growth controls containing no antibiotic and sterility controls without inoculation were also included. The inoculum was prepared from exponential growth mycobacterial culture adjusted to OD 0.3 then diluted 1:100, and 100 µl was used as an inoculum. The plates were covered, sealed in plastic bags, and incubated at 37 °C in the normal atmosphere. After 7 days of incubation, 40 µl of fresh prepared 0.1 mg/ml resazurin was added to each well, incubated 48 hours at 37 °C, and assessed for color development. A change from blue to pink indicates reduction of resazurin and therefore bacterial growth. The MIC was defined as the lowest drug concentration that prevented this color change.
Time-kill kinetics assay. Fast-growing (log phase) mycobacterial cultures were grown to OD 0.8 to 1.
Inoculum culture was prepared by diluting original culture to OD 0.4 then dilute 1:150 to achieve colony forming units around 1*10 6 . The tested concentrations of Q203 and aurachin D were added to inoculum culture and incubated at 37 °C without shaking. At indicated antibiotic exposure time, samples were collected, serially diluted (10-fold, 10 0 -10 6 ) and subcultured onto 7H11 agar plates supplemented with 10% (vol/vol) OADC enrichment (oleic acid-albumin-dextrose-catalase, Difco), 0.5% glycerol, 0.4% activated charcoal. Plates were sealed in plastic bags and incubated for 28 days at 37 °C to determine colony forming units (cfu) counts. The lower limit of detection was 10 cfu/mL. All experiments were performed in duplicate.

Preparation of inverted membrane vesicles.
Inverted membrane vesicles (IMVs) from the bacterial strains were prepared as described previously 46,47 . Briefly, M. smegmatis and M. tuberculosis mc 2 6020 were grown in a pre-culture to late-exponential phase. Cells were sedimented by centrifugation at 6000 × g for 20 minutes. The pellet was washed with phosphate buffered saline (PBS, pH 7.4) and centrifuged at 6000 × g for 20 min. Each 5 g of cells (wet weight) was re-suspended in 10 mL of ice-cold lysis buffer (10 mM HEPES, 5 mM MgCl 2 and 10% glycerol at pH 7.5) including protease inhibitors (complete, EDTA-free; protease inhibitor cocktail tablets from Roche). Lysozyme (1.2 mg/mL), deoxyribonuclease I (1500 U, Invitrogen) and MgCl 2 (12 mM) were added and cells were incubated with shaking for one hour at 37 °C. The lysates were passed three times through a One Shot Cell Disruptor (Thermo Electron, 40 K) at 0.83 kb to break up the cells. Unbroken cells were removed by three centrifugation steps (6000 × g for 20 min at 4 °C). The membranes were pelleted by ultracentrifugation at 222,000 × g for one hour at 4 °C. The pellet was re-suspended in lysis buffer and snap-frozen until use. The protein concentration was measured using the BCA Protein Assay kit (Pierce) as described by the manufacturer.
Oxygen consumption activity assay. Oxygen respiration and the effect of inhibitors on oxygen respiration were measured by polarography using a Clark-type electrode. The electrode was fully aerated (212 μM O2 at 37 °C) and calibrated with sodium hydrosulfite. The inverted membrane vesicles were pre-incubated for three minutes with the inhibitors in a pre-warmed (37 °C) buffer containing 50 mM MES and 2 mM MgCl 2 (pH 6.5). NADH was added as electron donor to a final concentration of 500 μM and oxygen respiration was measured for 3 minutes. Data were normalized relative to solvent (DMSO) control for full activity and to a sample with 10 mM potassium cyanide for complete inhibition. Statistical analysis (t-test) to determine P values and fitting of experimental results with one-site binding curves was done with GraphPad Prism software. Data availability. All data generated or analyzed during this study are included in this published article.