Arylvinylpiperazine Amides, a New Class of Potent Inhibitors Targeting QcrB of Mycobacterium tuberculosis

New drugs against Mycobacterium tuberculosis are urgently needed to deal with the current global TB pandemic. We report here on the discovery of a series of arylvinylpiperazine amides (AX-35 to AX-39) that represent a promising new family of compounds with potent in vitro and in vivo activities against M. tuberculosis. AX compounds target the QcrB subunit of the cytochrome bc1 terminal oxidase with a different mode of interaction compared to those of known QcrB inhibitors. This study provides the first multifaceted validation of QcrB inhibition by recombineering-mediated allelic exchange, gene expression profiling, and bioenergetic flux studies. It also provides further evidence for the compensatory role of cytochrome bd oxidase upon QcrB inhibition. In the absence of cytochrome bd oxidase, AX compounds are bactericidal, an encouraging property for future antimycobacterial drug development.

M ycobacterium tuberculosis, the causative agent of tuberculosis (TB), is at the origin of a severe global health problem. TB is the leading cause of death due to an infectious agent, with an estimated death toll of 1.6 million in 2016 (1). Drugsusceptible TB is typically treated over the course of 6 months with a combination of the first-line drugs isoniazid (INH), rifampin (RIF), pyrazinamide, and ethambutol; however, this regimen has been rendered ineffective nowadays in many cases due to the emergence of drug-resistant TB (1). Treatment of multidrug-resistant (MDR) TB, whose defining characteristics are resistance to both rifampin and isoniazid, requires a regimen of second-line drugs, including a fluoroquinolone and injectable agents, but treatment success rates are lower than 50% (1)(2)(3). In addition to resistance to rifampin and isoniazid, extensively drug-resistant (XDR) TB strains are resistant to at least one of the fluoroquinolones and an injectable. XDR TB cure rates are low at 30% (1). Long treatment duration, toxicity issues, drug resistance, and the expense of the second-line regimens are important factors driving the search for new anti-TB drugs with novel mechanisms of action.
To generate new anti-TB candidates, drug screening efforts have employed phenotypic or whole-cell screening approaches recently, whereby compound libraries of new chemical entities are screened against M. tuberculosis for inhibition of bacterial growth in various settings that, ideally, should recapitulate pathophysiological conditions (4)(5)(6).
Attempts are subsequently made to identify and validate the target(s) of hits to enable and facilitate further compound optimization. The diarylquinoline bedaquiline (BDQ), which targets subunit c of mycobacterial ATP synthase (7,8), emerged from such a compound-to-target approach, has been hailed as a milestone for TB drug discovery, and with its approval in 2012, became the first anti-TB drug to be approved in more than 40 years. The imidazopyrimidine amide Q203 and lansoprazole sulfide (LPZS) both target the QcrB subunit of cytochrome bc 1 oxidase and were identified from intracellular screens with M. tuberculosis-infected macrophages and fibroblasts, respectively (9,10). The approval of BDQ for the treatment of MDR TB, inclusion of BDQ as a component of the highly promising NIX-TB combination trial for XDR TB treatment (11), and the fact that Q203 is currently in phase I clinical development (2,12) altogether validate mycobacterial respiration as a new, relevant, and attractive target for an obligate aerobic pathogen.
In 2013, GlaxoSmithKline published the results of a large phenotypic screening campaign against Mycobacterium bovis BCG and M. tuberculosis, which yielded a total of 177 low-molecular-weight hits (13). One of these compounds was GW861072X (referred to here as AX-35) (Fig. 1), which exhibited potent activity against both M. bovis BCG and M. tuberculosis H37Rv (MIC of 0.3 M against both species).
Due to its structural simplicity, we considered AX-35 as an attractive starting point for lead optimization and prepared a series of analogs for structure-activity relationship (SAR) studies, the full details of which will be published in a forthcoming report. In this article, we describe the characterization of AX-35 and four of the most potent analogs that have emerged from our SAR work (AX-36 to AX-39) and are active against M. tuberculosis replicating in vitro and in infected macrophages. Subsequent work validated the target of these compounds as QcrB, the b subunit of the cytochrome bc 1 oxidase, based on the isolation of resistant mutants, cross-resistance studies, transcriptome sequencing (RNA-seq), and bioenergetic flux assays. Finally, further characterization of three AX compounds in an acute mouse model of TB demonstrated in vivo efficacy against M. tuberculosis.
with thiophene derivatives. For the same reason, the thiophene moiety in AX-35 was also replaced by a phenyl ring, as an established bioisostere (AX-38). As shown in the data in Table 1 (Table 1). The IC 50 s for M. tuberculosis H37Rv-infected THP-1 macrophages ranged from 0.1 to 1.8 g/ml (Table 1), reflecting the potent ex vivo activity of these molecules. The metabolic stability of the compounds in mouse and human liver microsomes was moderate to low (see Table S1 in the supplemental material).
Evidence of AX compounds targeting QcrB of M. tuberculosis. Initial attempts to raise mutants in M. tuberculosis H37Rv spontaneously resistant to AX-35 and AX-36 on solid 7H10 medium failed, even at exposures of up to 100ϫ MIC. An alternative method of continual passaging of H37Rv in complete 7H9 liquid medium containing AX-35 or AX-36 at concentrations above the MIC, beginning at 2ϫ MIC and increasing to 100ϫ MIC over five passages, proved to be more successful. Increasing MICs reflected the gradual selection of a resistant subpopulation in the culture on constant exposure to the compounds, especially after the fourth and fifth passages. Single clones isolated on 7H10 plates at passage 5 were subsequently tested for resistance to AX-35 or AX-36 by MIC determinations using the resazurin microtiter assay plate (REMA) method ( Fig. 2A).
To identify mutations associated with resistance to AX-35 or AX-36, whole-genome sequencing (WGS) analysis was performed on six clones with a 50-fold increase in MIC compared to that of the parental strain H37Rv ( Fig. 2A). Three missense mutations in qcrB were revealed, leading to nonsynonymous substitutions in QcrB, the b subunit of cytochrome bc 1 oxidase. S182P and M342V mutations were associated with AX-35 resistance and M342I with AX-36 resistance. The specific contribution of these substitutions to AX-35 and AX-36 resistance was confirmed by the generation of the same missense mutations in chromosomal qcrB of wild-type H37Rv by recombineering and MIC determination (see Fig. S1 in the supplemental material).
Since mutations in qcrB have also been identified in mutants of M. tuberculosis H37Rv resistant to Q203 and LPZS, two compounds known to target QcrB, crossresistance studies were performed to investigate if AX-35 and AX-36 behaved in a similar manner. Q203-and LPZS-resistant strains, harboring single-nucleotide polymorphisms (SNPs) leading to T313A and L176P in QcrB, respectively, are cross-resistant to AX-35 and AX-36 (Fig. 2B). These results indicate that the binding mode of the AX compound with QcrB shares some similarities with Q203 and LPZS. However, not all AX-resistant mutants were fully resistant to Q203 and LPZS (Fig. 2B), thus, suggesting a distinct mode of binding for these QcrB inhibitors. On inspection of a model of M. tuberculosis QcrB, residues S182, M342, T313, and L176 were found to cluster around the quinol oxidation site of the enzyme (Fig. 2C), implying that AX-35 and AX-36 bind to QcrB at this pocket, similar to Q203 and LPZS.
As QcrB is one of the respiratory subunits of the cytochrome bc 1 oxidase, a consequence of its inhibition is depletion of ATP levels in the bacterium (9, 10). The intracellular ATP level Arylvinylpiperazine Amides Target M. tuberculosis QcrB was measured in wild-type H37Rv after 24 h of exposure to AX-35 or AX-36 and was found to be ϳ90% lower than that for the untreated control, similar to the effect of treatment with the ATP synthase inhibitor BDQ or with Q203 under the same conditions (Fig. 2D), thus, indicating that AX compounds do indeed affect ATP levels. ATP was also depleted in BDQor Q203-exposed AX-resistant mutants, with ATP levels similar to those in BDQ-or Q203exposed wild-type M. tuberculosis. However, ATP levels were essentially unaffected in the AX-resistant strains in the presence of AX-35 or AX-36 (Fig. 2D). This demonstrates that the mutations in QcrB associated with AX resistance do not greatly impact the activity of BDQ  or Q203, indicating that AX compounds do not have off-target effects on the ATP synthase and that they interact in a different manner with QcrB compared to Q203. Taken together, these findings indicate that QcrB is the direct target of the AX compounds. Transcriptional response of M. tuberculosis to AX-35 treatment. To gain insight into the initial adaptive response of M. tuberculosis to AX-35 treatment, the transcriptomes of wild-type H37Rv exposed to AX-35 at 10ϫ and 30ϫ MIC for a duration of 4 h were examined. Of 81 genes significantly downregulated (fold change [FC] of ՅϪ2, Benjamini and Hochberg's adjusted P value [P adj ] of Յ0.05), leuC and leuD (involved in leucine biosynthesis) were the most extensively downregulated ( Fig. 3A and B). Thirteen genes were significantly upregulated (FC of Ն2; P adj Յ 0.05) in the presence of AX-35, seven of which were involved in intermediary metabolism and respiration ( Fig. 3A and B). The most highly upregulated gene was lipU, a gene implicated in lipid hydrolysis. Notably, two main operons in the M. tuberculosis transcriptome were upregulated, namely the cyd and mymA operons. The cyd operon consists of cydA, cydB, and cydD (cydA and cydB were close to the cutoff with FCs of 1.97 and 1.99, respectively), with cydA and cydB coding for subunits I and II, respectively, of the cytochrome bd oxidase, while cydDC encodes an ABC transporter involved in cytochrome bd Arylvinylpiperazine Amides Target M. tuberculosis QcrB assembly. The mymA operon from rv3083 to rv3089 (14) includes tgs4 and rv3088 encoding triacylglycerol (TAG) synthases (15), lipR, rv3085, and rv3086 encoding shortchain dehydrogenases, and an acyl coenzyme A (acyl-CoA) synthase gene, fadD13.
To validate the RNA-seq data, real-time quantitative PCR (qPCR) was performed targeting lipU, cydB, and tgs4 in both wild-type H37Rv and its AX-resistant mutants after exposure to AX-35 at 30ϫ MIC (Fig. 3C). The upregulation of all three genes in wild-type H37Rv upon AX-35 treatment detected by qRT-PCR is consistent with the RNA-seq data, with lipU having the highest relative expression of around 6-fold compared to the control (i.e., H37Rv treated with dimethyl sulfoxide [DMSO] alone). No significant differences in the expression levels of these three genes were measured for the AX-resistant mutant when treated with DMSO or AX-35, therefore establishing that their upregulation and, by extension, the deregulation of genes observed by RNA-seq are indeed a consequence of AX-35 treatment in M. tuberculosis.
Respiratory response of M. tuberculosis to AX-35 treatment. Bioenergetic flux studies were carried out to assess the respiratory response of M. tuberculosis to AX-35, with Q203 used as a control. The oxygen consumption rate (OCR) was first measured The increase in OCR upon the addition of AX-35 or Q203 can be attributed to cytochrome bd oxidase, since this increase was no longer observed in the presence of either compound in the ΔcydAB strain ( Fig. 4A and B). This switch in M. tuberculosis respiration to the bd oxidase has previously been validated as a respiratory signature of cytochrome bc 1 oxidase inhibitors such as Q203 (16), and our results indicate a bioenergetic profile of AX-35 similar to that of Q203.
Interestingly, a difference can be observed in the bioenergetic profiles on exposure of the cytochrome bd oxidase knockout (KO) strain harboring a QcrB(A317V) mutation to these two QcrB inhibitors. Addition of AX-35 decreases the OCR of this strain, generating a profile similar to that of H37Rv ΔcydAB (Fig. 4A), implying that respiration is via the bd oxidase branch and that QcrB remains fully inhibited despite the A317V mutation. The basal OCR remains unaffected by the addition of Q203, however (Fig. 4B), which would suggest a partial inhibition of QcrB, and thus the inability of the drug to fully inhibit QcrB due to the mutation. This therefore reveals additional differences in the binding modes of AX-35 and Q203 to QcrB.
Taken together, these results demonstrate the capability of respiratory adaptation of M. tuberculosis via the bd oxidase upon treatment with AX-35, generating a respiratory signature similar to that of another established QcrB inhibitor, Q203, albeit highlighting different modes of interaction of the two compounds within the binding pocket.
Cidality, drug interactions, and in vivo efficacy of AX compounds. The QcrB inhibitors described to date are bacteriostatic due to the compensatory role of the cytochrome bd oxidase. To determine the mode of action of AX compounds against M. tuberculosis, minimum bactericidal concentrations (MBCs) were determined in the M. tuberculosis H37Rv and H37Rv ΔcydAB strains and the complemented H37Rv ΔcydAB:: cydAB strain (see Fig. S2 in the supplemental material). In H37Rv, AX-35 is bacteriostatic, with an MBC up to a concentration 32-fold higher than the MIC. Conversely, AX-35 was bactericidal against the H37Rv ΔcydAB strain at a concentration 4-fold in excess of its MIC. Cidality of AX-35 can be specifically attributed to the lack of the cytochrome bd Since synergistic interactions have been previously reported between the cell wall inhibitor PBTZ169 and BDQ (17), as well as between compounds targeting the mycobacterial respiratory chain (16,18), checkerboard assays for two-drug combinations were performed to determine the nature of interactions between AX-35 with PBTZ169, BDQ, or clofazimine (CFM) in M. tuberculosis H37Rv. The ⌺ fractional inhibitory concentration (⌺FIC) indices obtained for all three combinations range from 0.8 to 1.6 for concentrations of AX-35 of Յ0.5-fold MIC (see Table S2 in the supplemental material), indicating nonantagonistic, additive interactions.
The activity of AX compounds in vivo was determined in mouse models of chronic and acute TB. None of the AX compounds reduced the bacterial burden in mice at the dose of 100 mg/kg of body weight in the chronic model (see Fig. S3 in the supplemental material). In the acute model, however, AX-35 at an oral dose of 200 mg/kg and AX-37 and AX-39 at 100 mg/kg significantly reduced the bacterial loads in mouse lungs by 0.4, 0.6, and 0.9 log 10 , respectively, compared to the D-␣-Tocopherol polyethylene glycol 1000 succinate (TPGS) vehicle control (Fig. 5).

DISCUSSION
The compounds investigated in this study are chemically distinct and different from the diverse QcrB inhibitors described in recent years (9,(19)(20)(21)(22)(23)(24)(25). Of the three mutations in QcrB identified here in spontaneous mutants resistant to AX-35 and AX-36, namely S182P, M342V, and M342I, the first two have been previously reported (20). Crossresistance studies and bioenergetic flux assays highlight influential residues for the interaction of AX-35 and AX-36 with QcrB, which occurs in the same binding site as Q203 and LPZS. In particular, mutations S182P, M342V/I, T313A, and L176P impair the interaction of AX-35 and AX-36 with QcrB, resulting in resistance of M. tuberculosis to the compounds. The substitution A317V, however, does not interfere with this interaction. This highlights the different modes of interaction of AX-35 and AX-36 with QcrB compared with Q203 or LPZS, and this information will be useful for further optimization of the compounds through structure-based drug design.
In addition to cross-resistance studies, transcriptomic and bioenergetic flux studies tuberculosis with AX-35 results in the upregulation of the cydABD operon, but not the dosR regulon, which is consistent with the transcriptomic signature of respiratory inhibitors specifically inhibiting the cytochrome bc 1 -aa 3 terminal oxidase (26). In line with this, exposure of M. tuberculosis to AX-35 results in an increase in OCR due to activity of the cytochrome bd oxidase, generating a bioenergetic profile similar to that of the respiratory signature of cytochrome bc 1 oxidase inhibitors such as Q203 (16). Cytochrome bd oxidase is non-proton translocating and, thus, the less bioenergetically efficient of the two terminal oxidases of M. tuberculosis (27). It has been demonstrated that the bd oxidase is upregulated under conditions in which the function of the cytochrome bc 1 -aa 3 is compromised (20,(28)(29)(30). In addition, inhibitors of respiration such as BDQ and QcrB inhibitors have markedly improved activity upon deletion or inhibition of bd oxidase (20,(31)(32)(33). The pronounced compensatory role of this alternate oxidase in the respiratory adaptation of M. tuberculosis for its survival is also apparent in our study, where the absence of bd oxidase is associated with AX-mediated bactericidal activity. Targeting both terminal oxidases simultaneously may therefore be a novel and effective strategy against M. tuberculosis. Although the increase in respiration, upregulation of cytochrome bd oxidase, and enhanced killing of ΔcydAB mutants upon inhibition of ATP synthase by BDQ in Mycobacterium smegmatis (34) have been attributed to the ionophoric/off-target effect of BDQ (35), it is unlikely that AX compounds behave in a similar manner to uncouplers, based on the bioenergetic flux data.
It is also of note that AX treatment of M. tuberculosis results in the upregulation of tgs genes (tgs4 and rv3087) and lipases (lipU and lipR). tgs genes are involved in the biosynthesis of triacylglycerol (TAG), the predominant energy source for M. tuberculosis in host macrophages (15,36), while lip genes encode lipolytic enzymes thought to be involved in the utilization of host TAG (37). Such upregulation of tgs and lip genes has been observed in various models of host stress (38,39), and similarly AX-induced stress remodels central carbon metabolism in M. tuberculosis toward lipid metabolism.
The compounds investigated here show promising properties as anti-TB agents. AX-35 to AX-39 demonstrated potent activity against M. tuberculosis in vitro and in macrophages, while having mild to no cytotoxicity. When tested in an acute mouse model of TB, AX-35, AX-37, and AX-39 were capable of reducing the bacterial burden in the lungs of mice. This in vivo activity can likely be improved further by addressing metabolic stability issues, such as those observed in mouse microsomes, although the compounds appear to be more stable in human microsomes.
The fact that activity of the AX compounds was only observed in the acute rather than the chronic mouse model of TB could be attributed to the difference in expression levels of the terminal oxidases found between these two stages of M. tuberculosis lung infection (29). The cytochrome bc 1 -aa 3 oxidase is more highly expressed than the bd oxidase during the exponential phase of growth of M. tuberculosis in mice, whereas it is downregulated as a response to host immunity and remains downregulated as the bacteria enter a nonreplicating state in the chronic phase (29). Thus, targeting QcrB during the exponential phase of M. tuberculosis growth would have a bigger impact on bacterial viability despite any compensatory response by bd oxidase. Another potential application of the AX inhibitors could be in the treatment of the human diseases Buruli ulcer and leprosy as their causative agents, M. ulcerans and Mycobacterium leprae, respectively, lack the cytochrome bd oxidase.
In conclusion, our SAR data prove that structural changes to AX-35 (GW861072X) can lead to improved in vivo activity against M. tuberculosis and that the arylvinylpiperazine amide series target the QcrB subunit of the cytochrome bc 1 -aa 3 oxidase. As part of the mycobacterial respiratory chain, QcrB is an attractive target for combination therapy. Furthermore, the biochemical information obtained from this study could guide future lead optimization work to enhance potency and stability of the compounds.

MATERIALS AND METHODS
Drugs used in this study. Q203 was synthesized as described before (9,40), LPZS was purchased from Santa Cruz Biotechnology, Toronto Research Chemicals, Inc., BDQ was a gift from Janssen Pharmaceutica NV, and PBTZ169 (macozinone) was provided by Innovative Medicines for Tuberculosis, while CFM, INH, and RIF were from Sigma-Aldrich.
Determination of MICs. MICs were determined using the resazurin reduction microplate assay (REMA) as previously described (41). Strains were grown to log phase (optical density at 600 nm [OD 600 ] of 0.4 to 0.8) and diluted to an OD 600 of 0.0001. Two-fold serial dilutions of each test compound were prepared in 96-well plates containing 100 l of bacteria per well (3 ϫ 10 3 cells per well). Plates were incubated at either 30 or 37°C as required with appropriate incubation times (e.g., 6 days for M. tuberculosis at 37°C). Ten microliters of resazurin (0.0025% wt/vol) was added to each well, after which the fluorescence intensity of the resorufin metabolite (excitation/emission, 560/590 nm) was read using an Infinite F200 Tecan plate reader. MIC values representing 90% growth inhibition were determined by a nonlinear fitting of the data to the Gompertz equation using GraphPad Prism.
Cytotoxicity for HepG2 cells. Human HepG2 cells (4,000 cells/well) were incubated for 3 days with 2-fold serially diluted compounds at 37°C under an atmosphere of 5% CO 2 . Cell viability was determined by the addition of resazurin (0.0025% wt/vol) for 4 h at 37°C, and the fluorescence intensity was measured as in REMA.
Assessment of drug activity in THP-1 macrophages. THP-1 human monocytic cells (10 5 /well) were seeded into 96-well plates and incubated with 100 nM phorbol-12-myristate-13-acetate (PMA) overnight to stimulate macrophage differentiation. Differentiated macrophages were infected with M. tuberculosis H37Rv grown to log phase (OD 600 of 0.4 to 0.8) at a multiplicity of infection (MOI) of 5. Extracellular bacteria were removed after 3 to 4 h of incubation at 37°C with 5% CO 2 by removing the RPMI medium and washing with phosphate-buffered saline (PBS). The compounds to be tested were prepared in separate 96-well plates by 2-fold serial dilutions in a final volume of 100 l RPMI, which was then transferred to the plates of infected THP-1 macrophages. Plates were sealed and incubated for 48 h at 37°C with 5% CO 2 . Ten microliters of PrestoBlue (Thermo Fischer Scientific) was added, and plates were incubated for up to 1 h at 37°C with 5% CO 2 before the fluorescence intensity (excitation/emission, 560/590 nm) was measured using the Infinite F200 Tecan plate reader. Dose-response curves were plotted, and 50% inhibitory concentrations (IC 50 s) were obtained using a nonlinear regression fit equation (log[inhibitor] versus response, variable slope) in GraphPad Prism.
Microsomal stability studies. Metabolic stability of the compounds was measured based on intrinsic clearance (Cl int ) in mouse and human liver microsomes as previously described (42). The final compound concentration in the mixture of microsomes and NADPH regeneration system was 2 g/ml, and a mixture without NADPH regeneration was also prepared for each compound as a control of the stability of the compound with time. Carbamazepine and nifedipine at 2 g/ml were used as low-and high-Cl int controls, respectively.
Mouse studies. To assess compound activity in an acute mouse model of TB, female BALB/c mice from Charles River Laboratories (7 to 8 weeks old, 20 g, 5 mice per group) were aerosol infected with M. tuberculosis H37Rv at a low dose and treated by oral gavage the following day with either vehicle controls (20% TPGS or 20% TPGS containing 1% DMSO) or test compounds once daily for 10 days. AX compounds were prepared in 20% TPGS, Q203 in 20% TPGS containing 1% DMSO (as described in reference 32), and INH in double-distilled water (ddH 2 O). Compounds were ground with a pestle and mortar followed by sonication at room temperature for 30 min. Compound solutions were stored at 4°C and prepared freshly after 5 days. The day after the final treatment, mice were sacrificed and serial dilutions of lung homogenates were plated on 7H10 agar containing 10 g/ml cycloheximide and 50 g/ml ampicillin. Experiments were approved by the Swiss Cantonal Veterinary Authority (authorization no. 3082).
Isolation and characterization of AX-resistant mutants. AX-resistant mutants of M. tuberculosis H37Rv were isolated from 7H9 cultures over 5 passages with increasing concentrations of AX-35 or AX-36 starting from 2ϫ, 5ϫ, and 10ϫ MIC to final concentrations of 50ϫ and 100ϫ MIC. Single colonies were obtained from three independent cultures by streaking on 7H10 agar plates, and resistance to AX was measured by REMA. Genomic DNA extraction was performed using the QiaAMP UCP pathogen minikit (Qiagen) as per the manufacturer's instructions. Whole-genome sequencing was performed using Illumina technology with sequencing libraries prepared using the KAPA HyperPrep kit (Roche) and sequenced on an Illumina HiSeq 2500 instrument. All raw reads were adapter and quality trimmed with Trimmomatic v0.33 (43) and mapped onto the M. tuberculosis H37Rv reference genome (RefSeq no. NC_000962.3) using Bowtie2 v2.2.5 (44). The bamleftalign program from the FreeBayes package v0.9.20-18 (45) was used to left-align indels. Reads with a mapping quality below 8 and duplicate reads were omitted. Variant analysis. Variant calling was done using VarScan v2.3.9 (46) using the following cutoffs: minimum overall coverage of 10 nonduplicated reads, minimum of 5 nonduplicated reads supporting the SNP, base quality score of Ͼ15, and an SNP frequency above 30%. The rather low thresholds, especially the SNP frequency, were deliberately chosen to avoid missing potential variants in regions where alignment was difficult or in case of a mixed population. All putative variants unique to the mutant strains were manually checked by inspecting the alignments.
Recombineering for target confirmation. Insertion of mutations associated with AX resistance in chromosomal qcrB of M. tuberculosis was done based on a recombineering method (47). The desired mutations were centered in lagging-strand oligonucleotides of 70 nucleotides. M. tuberculosis H37Rv containing plasmid pJV53 was grown to an OD 600 of 0.8 in 7H9 liquid medium containing 25 g/ml kanamycin and exposed to 0.2% acetamide for 24 h and 0.2 M glycine for 16 h. Competent cells were prepared and cotransformed with 100 ng of lagging-strand oligonucleotide and 50 ng of pYub412 carrying a hygromycin selection marker. After rescue for 3 days at 37°C, cells were plated on 7H10 plates containing hygromycin at 50 g/ml, and transformants were evaluated for resistance to AX by REMA. Desired mutations associated with AX resistance were subsequently confirmed by Sanger sequencing across qcrB, and the absence of other mutations was confirmed by whole-genome sequencing.
Modeling of M. tuberculosis QcrB. The QcrB protein of M. tuberculosis was modeled using the homology modeling web server (SWISS) (48) and chain A of the crystal structure of the mutant Rhodobacter sphaeroides cytochrome bc 1 oxidase (PDB code 2QJK). Illustrations were made using Pymol software (49).
Quantification of intracellular ATP. Log-phase cultures of wild-type H37Rv or AX-resistant mutants (about 10 6 CFU/ml) were exposed to test compounds at 2.5ϫ MIC for 24 h in a final volume of 100 l and incubated with BacTiter Glo reagent (Promega) (4:1 vol/vol) for 5 min in the dark. Luminescence was measured on a TECAN Infinite M200 in relative light units (RLU) with an integration time of 1 s.
Total RNA extraction and RNA-seq. Wild-type and AX-resistant H37Rv cultures were grown to mid-log phase and exposed to DMSO (vehicle control) or compounds for 4 h at 37°C. Cells were harvested by centrifugation and pellets were stored with 1 ml of TRIzol reagent (Thermo Fisher Scientific) at Ϫ80°C until further processed. Cells were lysed by bead-beating, and total RNA was extracted by phenol-chloroform with DNase treatment (RQ1 RNase-free DNase; Promega). Library preparation was done using the Ribo-zero rRNA removal kit (Illumina) for Gram-positive bacteria to deplete rRNA from total RNA. Two biological replicates for each strain were prepared for RNA-seq. Reads were adapter and quality trimmed with Trimmomatic v0.33 (43) and mapped onto the M. tuberculosis H37Rv reference genome (RefSeq no. NC_000962.3) using Bowtie2 v2.2.5 (44) Counting of reads over features was done with featureCounts from the Subread package v1.4.6 (50). DESeq2 (51) was used to infer differentially expressed genes.
qPCR. cDNA was prepared from total RNA using the SuperScript III first-strand synthesis kit (Invitrogen) and analyzed by qPCR for targeted gene expression in duplicates using Power SYBR Green PR master mix (Applied Biosystems) on a QuantStudio 5 real-time PCR system (Thermo Fisher Scientific). sigA was used as a housekeeping gene for normalization, and the threshold cycle (ΔΔC T ) method was used for quantification. Extracellular flux analysis. All strains of M. tuberculosis used were cultured in Middlebrook 7H9 medium (Difco) supplemented with 10% OADC (oleic acid-albumin-dextrose-catalase [Difco]) and 0.01% tyloxapol (Sigma) at 37°C to an OD 600 of ϳ0.6 to 0.8. M. tuberculosis H37Rv was obtained from BEI Resources (NR-123), and M. tuberculosis H37Rv ΔcydAB (52) and H37Rv Δcyd KO A317V (20) were gifts from Digby Warner and Helena Boshoff, respectively. The M. tuberculosis oxygen consumption rate (OCR) was measured using the Seahorse XF96 analyzer (Agilent) as previously described (16). In short, M. tuberculosis bacilli were adhered to the bottom of a XF96 cell culture microplate (Agilent) at a density of 2 ϫ 10 6 bacilli/well using Cell-Tak cell adhesive (Corning). Extracellular flux analysis was carried out in unbuffered 7H9 medium at pH 7.35 containing 0.2% glucose. Basal OCR was measured for ϳ19 min before the automatic addition, through the drug ports of the XF96 sensory cartridge (Agilent), of either  Fig. 4. OCR data points are representative of the average OCR after 3 min of continuous measurement, with the error calculated automatically by the Seahorse Wave Desktop 2.3.0 software (Agilent) from the OCR measurements from at least four replicate wells. OCR plots are representative of two independent experiments performed, and data representation was done using GraphPad Prism 7.02.