Inducing vulnerability to InhA inhibition restores isoniazid susceptibility in drug-resistant Mycobacterium tuberculosis

ABSTRACT Of the approximately 10 million cases of Mycobacterium tuberculosis (Mtb) infections each year, over 10% are resistant to the frontline antibiotic isoniazid (INH). INH resistance is predominantly caused by mutations that decrease the activity of the bacterial enzyme KatG, which mediates the conversion of the pro-drug INH to its active form INH-NAD. We previously discovered an inhibitor of Mtb respiration, C10, that enhances the bactericidal activity of INH, prevents the emergence of INH-resistant mutants, and re-sensitizes a collection of INH-resistant mutants to INH through an unknown mechanism. To investigate the mechanism of action of C10, we exploited the toxicity of high concentrations of C10 to select for resistant mutants. We discovered two mutations that confer resistance to the disruption of energy metabolism and allow for the growth of Mtb in high C10 concentrations, indicating that growth inhibition by C10 is associated with inhibition of respiration. Using these mutants as well as direct inhibitors of the Mtb electron transport chain, we provide evidence that inhibition of energy metabolism by C10 is neither sufficient nor necessary to potentiate killing by INH. Instead, we find that C10 acts downstream of INH-NAD synthesis, causing Mtb to become particularly sensitive to inhibition of the INH-NAD target, InhA, without changing the concentration of INH-NAD or the activity of InhA, the two predominant mechanisms of potentiating INH. Our studies revealed that there exists a vulnerability in Mtb that can be exploited to render Mtb sensitive to otherwise subinhibitory concentrations of InhA inhibitor. IMPORTANCE Isoniazid (INH) is a critical frontline antibiotic to treat Mycobacterium tuberculosis (Mtb) infections. INH efficacy is limited by its suboptimal penetration of the Mtb-containing lesion and by the prevalence of clinical INH resistance. We previously discovered a compound, C10, that enhances the bactericidal activity of INH, prevents the emergence of INH-resistant mutants, and re-sensitizes a set of INH-resistant mutants to INH. Resistance is typically mediated by katG mutations that decrease the activation of INH, which is required for INH to inhibit the essential enzyme InhA. Our current work demonstrates that C10 re-sensitizes INH-resistant katG-hypomorphs without enhancing the activation of INH. We furthermore show that C10 causes Mtb to become particularly vulnerable to InhA inhibition without compromising InhA activity on its own. Therefore, C10 represents a novel strategy to curtail the development of INH resistance and to sensitize Mtb to sub-lethal doses of INH, such as those achieved at the infection site.

in TB disease surveillance and global control efforts (1,2).A major obstacle in the treatment of Mtb infections is that the sterilizing activity of antibiotics is slow and sometimes incomplete at the site of infection due to several contributing factors.The penetration of antibiotics into the Mtb lesion can be limited, causing Mtb to be exposed to fluctuating and often subinhibitory concentrations of antibiotics (3).In addition, Mtb has the propensity to develop phenotypically drug-tolerant populations in the host, which allows a population of Mtb to persist despite exposure to antibiotics (3)(4)(5)(6).Long treatment regimens are required to overcome the issues of drug penetration and bacterial drug tolerance to ultimately clear the infection.The standard of care for the treatment of active TB lasts 6 months, including a 2-month intensive phase of regular doses of isoniazid (INH), rifampicin, pyrazinamide, and ethambutol (EMB) followed by a 4-month continuation phase of INH and rifampicin (7).Recently, the World Health Organization approved the recommendation for a shortened 4-month treatment regimen that can be made available to some patients, which includes a 2-month intensive phase of INH, rifapentine, pyrazinamide, and moxifloxacin, followed by 2 months and 1 week of INH, rifapentine, and moxifloxacin (7).As a component of the intensive and continuation phases of both longer and shorter treatment regimens, INH is a critical frontline antibiotic that is the cornerstone of our current anti-TB regimens.
The utility of INH for the treatment of Mtb infections is threatened by the emergence and prevalence of INH-resistant mutant strains of Mtb.An estimated 10.7% of newly infected and 27.2% of previously treated cases are INH-resistant (8).INH is a prodrug, and resistance is most commonly caused by mutations in the gene katG, which encodes the sole bifunctional catalase-peroxidase enzyme in Mtb that is also responsible for converting INH to its active form within the bacteria (8)(9)(10).KatG is an oxidative defense enzyme, and its typical substrates are H 2 O 2 and other peroxides (11).However, KatG acts on INH as a non-canonical substrate to generate a radical intermediate of INH (12), which spontaneously reacts with and attaches to the abundant cofactor NAD (H), forming INH-NAD (10,13).The INH-NAD adduct, the activated form of INH, inhibits the enzyme InhA (10,13,14), which is the enoyl-acyl carrier protein reductase enzyme that functions in the fatty acid synthase II (FAS-II) system (15,16).InhA is required for the FAS-II system to elongate the shorter fatty acids synthesized by FAS-I to generate long lipid precursors that are subsequently converted to mycolic acids (MAs) through a multi-step process.MAs are an essential structural component of the outermost layer of the Mtb cell envelope and, therefore, by inhibiting InhA, INH-NAD compromises the integrity of the Mtb cell envelope, leading to growth inhibition and death (17,18).
Identifying ways to enhance the antibacterial activity of INH has the potential to greatly improve the standard of care for TB.To this end, we recently reported the identification of the bicyclic 2-pyridone compound C10 as a potentiator of INH activity in Mtb (19).At concentrations that on their own do not inhibit growth, C10 promotes the killing of Mtb by INH and prevents the emergence of spontaneous INH-resistant mutants (19).Whereas high-level resistance to INH mediated by mutations in katG is generally considered to render INH ineffective (20), we discovered that C10 was able to re-sensitize multiple INH-resistant katG mutants to inhibition by INH, which had previously not been thought to be possible.The ability of C10 to potentiate INH activity in both WT and INH-resistant katG mutant strains of Mtb demonstrates that there is a vulnerability in the bacteria that can be exploited to enhance the antimicrobial activity of INH and even circumvent INH resistance (19).Understanding the target and mechanism of action of C10 could lead to the discovery of novel therapeutic approaches that can be used in the clinic to disarm INH resistance in katG mutants.
In previous work, we performed RNA-sequencing on C10-treated Mtb and found that C10 induces a transcriptional signature consistent with inhibition of respiration (19).We subsequently demonstrated that C10 blocked Mtb oxygen consumption and decreased bacterial ATP levels, suggesting that C10 disrupts Mtb energy metabolism (19).In the current study, we aimed to determine how C10 potentiates killing by INH and whether this potentiation is linked to inhibition of Mtb energy metabolism.We used a combination of forward genetics and chemical biology to reveal that while some of the effects of C10 are mediated through its impact on energy homeostasis, the inhibition of respiration by C10 is not required for C10 to potentiate INH, thereby uncoupling these two effects of C10 on Mtb physiology.Instead, we present evidence that C10 restores INH susceptibility in a subset of resistant mutants by enhancing the bacterial vulnerability to InhA inhibition.Our findings reveal potential strategies to improve the efficacy of INH as well as other antibiotics that target mycolic acid metabolism in Mtb.

Isolation of mutants that are resistant to C10 growth inhibition and ATP depletion
To better understand the mechanism of action of C10 (Fig. 1A), we chose a forward genetic approach and isolated mutants that are resistant to C10, with the goal of identifying mutations in genes linked to the mechanism of action of C10.Previous studies used 25 µM C10 to disrupt Mtb energy homeostasis and deplete bacterial that harbored the depicted mutation.Representative strains GHTB136, GHTB146, and GHTB149 were selected for follow-up studies.(E and F) The indicated strain of Mtb was cultured in the presence of increasing concentrations of (E) C10 or 17h (F) for 1 week, and the % inhibition of Mtb growth and metabolism was determined using the resazurin assay, n = 6.(G) The indicated strain of Mtb was cultured in Sauton's liquid medium containing 0 or 25 µM C10 for 24 h before ATP levels were measured by the BacTiter Glo assay.The relative luminescence units (RLUs) were normalized to the optical density (OD 600 ) of the culture to control for differences in cell density.Fold change in ATP levels were calculated relative to the 0 µM C10 control for each strain, n = 4-7.A one-way ANOVA (E and F) or a two-way ANOVA (G) with Tukey's post test was performed to determine statistically significant differences across samples.IC 50 values in panels E and F were log-transformed before statistical significance testing was performed.Selected comparisons are depicted in the figure.ns, not significant; ****P < 0.0001.For all pairwise comparisons, please see Table S1.ATP (19).However, 25 µM C10 only results in a modest decrease in Mtb growth (19).Therefore, to select for C10-resistant mutants, we first determined the concentration of C10 that was sufficient to inhibit the growth of wild-type (WT) Mtb.By increasing the C10 concentration above 25 µM, we found that C10 caused dose-dependent inhibition of Mtb growth both in liquid media and on agar plates (Fig. 1B and C), consistent with our previous studies (19).We found that 200 µM C10 completely inhibited Mtb growth in both conditions, so we used this concentration to select for resistant mutants.By spreading Mtb on agar containing 200 µM C10 and allowing these agar plates to incubate for a total of 9 weeks, we eventually observed the emergence of spontane ous resistant colonies.We isolated 11 resistant mutants and performed whole-genome sequencing to identify mutations that confer the C10 resistance.We found that each resistant strain harbored one of 3 different mutations (Fig. 1D).To probe how these mutations impact C10 sensitivity, we chose representative strains that each harbor one of these three mutations as the sole nucleotide change that could be identified by whole-genome sequencing.Strain GHTB136 harbors a C to A substitution in the intergenic region 69 base pairs (bp) from the start of the putative S-adenosylmethionine (SAM)-methyl transferase Rv0731c and 92 bp upstream of the secY-adk-mapA operon.To determine if this mutation impacts the expression level of the neighboring genes, we performed quantitative real-time PCR (qRT-PCR) and found that the GHTB136 strain exhibited >100-fold upregulation of the Rv0731c gene and 2-to 6-fold upregulation of the secY-adk-mapA operon compared to WT (Fig. S1A).The GHTB146 strain harbors an A to G substitution in the intergenic region 9 bp upstream of the lpdA-glpD2 operon and >100 bp upstream from the uncharacterized gene Rv3304.The Mtb LpdA enzyme is a dehydrogenase that can oxidize or reduce NAD (H) or NADP (H) with concomitant oxidation or reduction of quinones (21).GlpD2 has not been characterized in Mtb but based on homology to the GlpD enzyme of Escherichia coli (22), it is predicted to be a glycerol-3-phosphate dehydrogenase that interconverts glycerol-3-phosphate and dihydroxyacetone phosphate with concomitant oxidation or reduction of mena quinone/menaquinol in the membrane.Since lpdA is a leaderless transcript (23), this mutation is likely located within the RNA polymerase binding region.Indeed, qRT-PCR analysis revealed that the GHTB146 mutant exhibited 4-to 8-fold upregulation of the lpdA-glpD2 operon, but no change in expression of the Rv3304 gene compared to WT (Fig. S1B), suggesting that this mutation enhances lpdA-glpD2 promoter activity.The third representative strain, GHTB149, harbors a missense mutation within the putative SAM-methyl transferase Rv0830 that results in the substitution of valine for a leucine residue at position 292 (L292V).Out of all of the genes identified by this approach, secY, adk, mapA, lpdA, and glpD2 have been characterized in Mtb or Escherichia coli, and their functions are listed in Fig. S1C (21,22,(24)(25)(26).However, Rv0731c and Rv0830, the putative SAM-methyltransferases affected in GHTB136 and GHTB149, remain hypothetical with a predicted catalytic activity but no characterized function in Mtb (Fig. S1C).
We quantified the level of C10 resistance in the GHTB136, GHTB146, and GHTB149 strains using a resazurin microplate assay.This assay takes advantage of the redox-sen sitive dye resazurin, which is blue in its oxidized form but becomes reduced to the fluorescent pink product resorufin as a result of bacterial growth and metabolism.Therefore, fluorescence can be monitored as a proxy for bacterial growth and metab olism.C10 inhibits WT Mtb in this assay with a half-maximal inhibitory concentration (IC 50 ) of 25 µM (Fig. 1E) (19).In contrast, the C10 IC 50 in both GHTB136 and GHTB146 was >100 µM and the IC 50 of C10 in GHTB149 was 59 µM (Fig. 1E).We also confirmed that both GHTB136 and GHTB146 exhibited high-level resistance to the recently published more potent C10 analog, 17h (Fig. 1A) (27), whereas GHTB149 was unable to significantly suppress the effects of 17h (Fig. 1F).Therefore, although all three mutants are resistant to C10, the mutations in GHTB136 and GHTB146 confer a higher level of resistance than the GHTB149 strain.
We previously showed that C10 inhibits respiration and depletes ATP levels in WT Mtb (19).Therefore, to directly determine how the mutations in the C10-resistant strains affected ATP-depletion by C10, we cultured WT, GHTB136, GHTB146, and GHTB149 in the presence and absence of 25 µM C10 for 24 h and quantified bacterial ATP levels using a luciferase-based BacTiter Glo assay (Fig. 1G).Similar to WT, the GHTB149 strain exhibited a decrease in bacterial ATP in response to C10, suggesting that the low level of resistance conferred by the Rv0830 L292V mutation is not sufficient to overcome the depletion of ATP by C10.In contrast, C10 did not significantly decrease the ATP levels in the GHTB136 and GHTB146 strains (Fig. 1G), demonstrating that these mutants are resistant to the ATP-depleting effects of C10.Notably, none of the mutants exhibited an altered level of ATP at baseline (Fig. S1D), indicating that these strains do not overcome the effects of C10 by harboring increased pools of ATP but instead are able to maintain ATP levels during C10 treatment.Collectively these findings demonstrate that C10-mediated growth inhibition is linked to ATP depletion, as mutants that maintain ATP levels in the presence of C10 are able to overcome the toxicity of C10.To determine if the GHTB136 and GHTB146 strains are more generally resistant to ATP depletion, we quantified the sensitivity of the C10-resistant mutants to the ATP synthase inhibitor bedaquiline (BDQ) or the protonophore CCCP, compounds known to inhibit ATP synthesis by targeting the electron transport chain (ETC) (28,29).We found that GHTB146 and GHTB149 exhibited a modest but significant 2.3-and 1.4-fold increase in the BDQ IC 50 , respectively (Fig. S1E), and GHTB136, GHTB146, and GHTB149 all exhibited a subtle but significant <1.5-fold increase in the CCCP IC 50 compared to WT (Fig. S1F).While statistically significant, these minor changes in sensitivity to BDQ and CCCP indicate that GHTB136, GHTB146, and GHTB149 are especially resistant to the effects of C10, and the mechanism of resistance in these strains does not confer the same degree of cross-resistance to respiration inhibitors in general.

The effects of C10 on ATP levels are associated with sensitization to the respiration inhibitor Q203 but not INH
Since the GHTB149 mutant only had low-level resistance to C10 (Fig. 1E), was not cross-resistant to the more potent C10 analog 17h (Fig. 1F), and was not resistant to the ATP-depleting effects of C10 (Fig. 1G), we focused our subsequent analysis on GHTB136 and GHTB146.In addition to decreasing ATP levels, C10 treatment potentiates the bactericidal effects of the cytochrome bc 1 inhibitor Q203, low pH, and INH in WT Mtb (19).Therefore, we sought to determine if the mutations in GHTB136 and GHTB146 affected these additional consequences of C10 treatment.In particular, for Q203, we hypothesized that the ability of C10 to potentiate killing by Q203 was dependent on its effect on Mtb energy metabolism because respiration inhibitors often exhibit synergistic activity with other inhibitors of bacterial respiration (30)(31)(32).Indeed, when we exposed GHTB136 and GHTB146 to C10 and/or Q203 and monitored survival after 15 days of treatment, we found that C10 was unable to potentiate killing of GHTB136 and GHTB146 by Q203 (Fig. 2A).This is in contrast to WT Mtb, where C10 in combination with Q203 causes a significant 2 orders of magnitude decrease in the number of colony forming units (CFUs)/mL compared to Q203 alone (Fig. 2A) (19).These findings support that the ability of C10 to disrupt Mtb energy homeostasis is required for C10 to potentiate the bactericidal activity of Q203.Sensitivity to low pH is also a phenotype that is associ ated with inhibition of mycobacterial energy homeostasis (33,34).However, when we examined the effect of C10 on GHTB136 and GHTB146 survival in pH 5.5 media, we found that C10 still mediated decreased survival of GHTB136 and GHTB146 Mtb strains in low pH media, to a similar magnitude as WT Mtb (Fig. 2B).These data suggest that the mechanisms that allow GHTB136 and GHTB146 to resist C10-mediated growth inhibition in pH 7.0 media are not sufficient to resist killing by C10 in low pH media.
In addition to its effects on energy homeostasis, we previously showed that C10 potentiates killing by INH (19).However, whether the disruption of energy homeostasis by C10 contributes to its ability to enhance the bactericidal activity of INH remains unknown.We used GHTB136 and GHTB146 to test if disruption of energy homeostasis is required for C10 to potentiate INH by culturing WT Mtb, GHTB136, and GHTB146 in

Isolation of mutants that are resistant to the combination of C10 and INH
Our findings that C10 still mediates bacterial killing in low pH and potentiates the bactericidal activity of INH in GHTB136 and GHTB146 (Fig. 2B-C) supports that the mutations in these C10-resistant strains specifically overcome the effects of C10 on energy homeostasis but are unlikely to represent the direct targets of C10.These findings also imply that C10 must impart another effect on Mtb that is not reversed in the GHTB136 and GHTB146 mutants to elicit the increased sensitivity to low pH and to INH.
To specifically address how C10 potentiates INH activity in Mtb, we sought to identify genes that are required for C10 to potentiate INH by selecting for spontaneous Mtb mutants that can grow in the presence of C10 and INH.We had previously been unable to select for mutants that grew in the presence of 25 µM C10 and 0.5 µg/mL INH (19).Therefore, we decreased the selective pressure by lowering the concentration of INH.We inoculated WT Mtb onto agar media containing 25 µM C10 and 0.2 µg/mL INH, and incubated the bacteria for 4 months at 37°C.We isolated three spontaneous mutants that could grow on agar media containing 25 µM C10 and 0.2 µg/mL INH and performed whole-genome sequencing to identify the genetic basis for resistance.Two of the mutant strains harbored large genomic deletions.One strain, GHTB089, was deleted for 27.9 kilobases (kb) of its genome (Δ2145809-2173696), which disrupted 27 annotated genes, including deleting the first 1183 bp of katG.The second isolate, GHTB092, was deleted for 38.6 kb of its genome (Δ2132215-2170824), which included the entire katG gene and 36 additional annotated genes.In contrast, the third strain harbored a single point mutation when compared to the parental WT strain, a nucleotide change in katG that results in an early stop codon at tryptophan 198.We designated this strain katG W198* .KatG is a 740-amino acid protein with several residues that are critical for catalaseperoxidase activity, including a heme-coordinating histidine at position 270 and catalytic residues at R104, H108, and W321.The katG W198* mutation is predicted to result in truncation of over two-thirds of the KatG protein, including several of these essential residues.The selection of the katG W198* mutant on media containing INH and C10 was perplexing because in our earlier study, we had shown that C10 potentiates INH in multiple INH-resistant katG mutants, including a katG mutant that harbors a frameshift mutation at amino acid 6 that results in an early stop codon (katG FS6 ) and a katG mutant with a single amino acid substitution of a leucine for a tryptophan at position 328 (katG W328L ) (19).To determine if the katG W198* mutant was truly unique from the previously studied INH-resistant katG mutants, we compared the ability of the katG W198* , katG FS6 , and katG W328L mutants to grow on agar media containing 0.5 µg/mL INH and/or 25 µM C10 (Fig. 3A).As expected based on our previous data (19), the katG FS6 and katG W328L mutants grow well in the presence of either C10 or INH alone, but are re-sensitized to INH in the presence of C10 such that their growth is inhibited on agar containing both C10 and INH together (Fig. 3A).In contrast, the katG W198* mutant grew on agar containing C10 alone, INH alone, and the combination (Fig. 3A), demonstrating that this strain is resistant to INH even in the presence of C10.
We monitored the effects of C10 and INH on the viability of each katG mutant strain compared to WT by culturing the bacteria in liquid media with or without C10 and/or INH for 10 days and enumerating CFU (Fig. 3B through E).Similar to our previously reported results, while the katG FS6 mutant can grow in the presence of INH or C10 alone, 25 µM C10 in combination with INH caused a significant decrease in the number of viable bacteria compared to C10 or INH alone (Fig. 3C) (19).Similar to the katG FS6 mutant, the katG W328L mutant can grow in the presence of INH or C10 alone, but in combination with INH, 10 or 25 µM C10 was able to decrease the number of CFU/mL two to four orders of magnitude below the inoculum (Fig. 3D), demonstrating that C10 restores the bacterici dal activity of INH against this mutant.In contrast, the CFU/mL of the katG W198* mutant increased from days 0 to 10 in cultures treated with C10 alone, INH alone, and the combination (Fig. 3E).We found that 25 µM C10 significantly decreased the CFU/mL on day 10 compared to the untreated controls, but this was not enhanced by the addition of INH, indicating that C10 inhibits the growth of the katG W198* mutant but is unable to potentiate INH in this strain.
The ability of the katG W198* mutant to grow in the presence of C10 and INH could be due to resistance to the C10/INH combination or resistance to C10 specifically.To determine if the katG W198* mutant was resistant to C10 activity, we examined the effect of C10 treatment on ATP levels in the katG FS6 , katG W328L , and katG W198* mutants by treating with 10 µM or 25 µM C10 for 24 h and then measuring ATP levels using the luciferasebased BacTiter Glo assay.C10 treatment caused a similar dose-responsive decrease in ATP in WT Mtb and the katG mutants as compared to the DMSO-treated cultures (Fig. 3F; Fig. S2).C10 also inhibited the katG mutants to a similar or even greater extent compared to WT in the resazurin microplate assay, with the IC 50 of C10 in the WT, katG FS6 , katG W328L , and katG W198* strains being 22 µM, 12 µM, 7.4 µM, and 14 µM, respectively (Fig. 3G).These findings demonstrate that C10 is still able to disrupt Mtb energy homeostasis in katG W198* and, therefore, the loss of INH potentiation in katG W198* is not due to insensitivity to C10 activity.

KatG activity is required for C10 to enhance INH inhibitory activity in Mtb
Given the finding that C10 was able to re-sensitize both katG FS6 and katG W328L mutant strains but not katG W198* to INH, we hypothesized that there may be a functional difference between the KatG protein variant expressed in the katG W198* mutant compared to katG FS6 and katG W328L .To begin to investigate this possibility, we probed the expression of KatG protein in each strain by western blot using a monoclonal α-KatG antibody (Fig. 4A).katG W328L Mtb harbored similar or higher levels of KatG protein compared to WT Mtb, whereas katG W198* Mtb did not express any full-length KatG protein, similar to a ΔkatG strain.In contrast, the katG FS6 strain expressed a low abundance protein species that was recognized by the α-KatG antibody and migrated slightly faster than the full-length KatG protein expressed in WT Mtb (Fig. 4A).Upon closer examination of the katG mRNA sequence, we noted that there is a putative alternative GUG start codon downstream from the early stop codon introduced by the frameshift mutation in katG FS6 that may reinitiate translation in the correct frame beginning at codon 23.Therefore, we postulate that the faint band detected by western blot from katG FS6 cell lysate represents low-level expression of a variant KatG protein with a small N-terminal truncation (Δ1-22) that still contains all necessary catalytic residues (Fig. 4A and B).
Western blot analysis indicates that the katG FS6 and katG W328L mutants express KatG protein variants that contain all necessary catalytic residues, while katG W198* does not (Fig. 4A and B).To determine if the katG FS6 and katG W328L mutants retained KatG catalytic activity, we cultured each strain in liquid media with or without 0.25 µg/mL INH and monitored the growth of the culture by measuring OD 600 over time.We found that while WT Mtb was completely inhibited by 0.25 µg/mL INH (Fig. 4C), the katG FS6 and katG W328L mutants were able to grow in the presence of INH, eventually reaching an OD 600 over 10fold above the inoculum (Fig. 4D and E).However, both mutants exhibited a significant decrease in the OD 600 compared to the untreated controls, indicating that although the katG FS6 and katG W328L mutants exhibit decreased sensitivity to INH, some KatG activity was retained to impart this modest growth inhibition in the presence of INH (Fig. 4D and E).In contrast, the katG W198* and ΔkatG mutants were completely resistant and grew uninhibi ted in the presence of INH (Fig. 4F and G), supporting that both of these mutations result in katG-null alleles.Therefore, the katG FS6 and katG W328L mutants that are re-sensitized to INH by C10 are hypomorphic for katG, exhibiting decreased KatG activity leading to INH resistance, but retaining a residual level of KatG enzymatic activity (Fig. 4D and E).In contrast, the katG W198* mutant that is not re-sensitized to INH by C10 exhibits no KatG activity (Fig. 4F), similar to a katG-null strain (Fig. 4G).Based on these data, we hypothe sized that some KatG activity is required for C10 to enhance INH sensitivity.To test this hypothesis, we examined whether C10 could sensitize a ΔkatG mutant to INH.Deletion of katG phenocopied the katG W198* mutant and enabled Mtb to grow on agar and in liquid media containing both C10 and INH (Fig. 4H and I), demonstrating that some KatG expression and activity is required for C10-mediated sensitization to INH.

C10 induces vulnerability to inhibition by INH without altering KatG activity or INH-NAD levels
Most strategies that renew the sensitivity of bacterial pathogens to antibiotics do so by increasing the levels of active antibiotics.Key examples of this are β-lactamase inhibitors that prevent the degradation of β-lactam antibiotics (38) and the recent discovery of SMARt-420, which increases the conversion of ethionamide (ETH) to its active form, ETH-NAD, in Mtb (39).In addition, Mtb encodes the enzyme CinA that cleaves NAD-drug adducts to promote tolerance to antibiotics like INH and ETH (40), highlighting that regulation of active drug concentration is a major mechanism of modulating drug efficacy.Therefore, we investigated whether C10 sensitizes Mtb to INH by promoting KatG activity to enhance the conversion of INH to INH-NAD, thus increasing the levels of INH-NAD in the cell.We monitored the effects of C10 on KatG enzymatic activity in vitro by incubating purified KatG protein with H 2 O 2 , a natural substrate of KatG, and monitor ing KatG catalase activity as measured by the H 2 O 2 degradation rate.We found that C10 did not enhance the rate or kinetics of H 2 O 2 degradation by purified KatG protein in vitro, indicating that C10 did not directly affect KatG catalase activity (Fig. S3A).We  S1.
protein in buffer containing Mn 2+ with INH and NAD + in the presence or absence of C10 and monitored the levels of C10, INH, NAD + , and INH-NAD by liquid chromatographymass spectrometry (LC-MS).We found that while the level of C10 remained unchanged over the course of the experiment, INH and NAD + were depleted from the reaction with a concomitant increase in INH-NAD, as expected since INH and NAD + are consumed to produce INH-NAD (Fig. 5A through D).C10 did not impact the rate or level of INH-NAD produced in these conditions.In addition, although we were able to detect the interac tion between purified KatG and INH in vitro (Fig. S3B), we were unable to detect a direct interaction between C10 and KatG using a thermal shift assay (Fig. S3C), together supporting that C10 does not directly bind KatG or promote its enzymatic activity in vitro.
To determine if C10 affects KatG activity within the bacterium, we examined whether C10 promotes KatG expression, which could explain how C10 sensitizes WT and katG  S1.
hypomorphic strains to INH but does not sensitize katG-null strains.We treated both WT and katG FS6 Mtb with C10 for 3 days and collected whole-cell lysate to perform western blot analysis for KatG.We found that C10 did not increase the protein levels of KatG in these conditions (Fig. 5E).We also monitored the effect of C10 on KatG catalase activity in Mtb by treating the bacteria with C10 for 6 days, collecting whole-cell lysate, and measuring the H 2 O 2 degradation rate of the lysate.The ΔkatG mutant exhibited no H 2 O 2 degradation in this assay, demonstrating that the assay is specific for KatG (Fig. 5F).While the lysate from katG FS6 Mtb had significantly more activity than the ΔkatG mutant, this strain exhibited a greater than 10-fold reduction in catalase activity compared to WT, consistent with the katG FS6 strain being hypomorphic for katG.We found that C10 did not change the H 2 O 2 degradation rate of WT or katG FS6 Mtb (Fig. 5F), confirming that C10 treatment does not enhance the expression or activity level of KatG within Mtb.However, this finding did not rule out whether C10 could enhance the levels of INH-NAD within the bacteria without affecting KatG activity.To determine if C10 promotes INH-NAD accumulation in Mtb, we cultured WT, katG FS6 , and katG W328L Mtb with and without C10 and/or INH for 3 days, collected aqueous metabolite extracts from the bacteria, and measured the amount of activated INH-NAD in the bacterial extract using LC-MS.We included the katG-null mutants katG W198* and ΔkatG as controls.We found that upon treatment with INH alone, WT, katG FS6 , and katG W328L strains all accumulated INH-NAD, and the katG FS6 and katG W328L mutants produced significantly decreased levels of INH-NAD compared to WT Mtb, confirming that these strains are indeed hypomorphic for katG (Fig. 5G).As expected, the katG-null mutants katG W198* and ΔkatG were deficient in INH-NAD synthesis (Fig. 5G).C10 did not significantly increase the amount of INH-NAD in any of the strains tested (Fig. 5G) and, therefore, does not affect INH-NAD synthesis or degradation.These data demonstrate that C10 sensitizes Mtb to INH through a novel mechanism of action without impacting the levels of INH-NAD in the bacteria.

C10 sensitizes Mtb to direct inhibition of InhA
Our data indicate that although the effect of C10 on INH sensitivity relies on KatG activity, C10 does not increase the expression or enzymatic activity of KatG.Based on these findings, we hypothesized that C10 acts downstream of KatG to enhance the antibacterial activity of INH-NAD after it is produced, such that even the katG hypomorphic strains that produce lower levels of INH-NAD become inhibited by this lower concentration in the presence of C10.INH-NAD inhibits Mtb growth by binding the NADH binding pocket in the enoyl-acyl carrier protein reductase InhA, leading to inhibition of mycolic acid biosynthesis (13,14,18,42,43).Therefore, it is possible that C10 renders Mtb more sensitive to inhibition of InhA.To test this possibility, we cultured WT Mtb with C10 and/or NITD-916, a direct inhibitor of InhA that does not require KatG or any other known enzyme for activation (44), and quantified the surviving CFU after 10 days of treatment.Similar to the effect of C10 on INH sensitivity, we found that treating Mtb with C10 in combination with NITD-916 caused a significant decrease in the number of surviving bacteria compared to NITD-916 alone, leading to an additional 2-3 orders of magnitude decrease in survival (Fig. 6A).Therefore, C10 increases Mtb sensitivity to direct InhA inhibition.
Given the ability of C10 to potentiate killing by NITD-916, we selected for mutants that could grow on agar containing both C10 and NITD-916 with the goal of identifying mutations in genes that are required for C10 to potentiate killing during direct InhA inhibition.We isolated two mutant colonies that emerged on agar containing both C10 and NITD-916 after 4 weeks and performed whole-genome sequencing.We found that each isolate harbored a mutation within the inhA gene that resulted in a single amino acid substitution, inhA I202S and inhA G205C .Both the inhA I202S and inhA G205C mutants exhibited a significant increase in the IC 50 for NITD-916 compared to WT Mtb in a microplate absorbance-based growth inhibition assay (Fig. 6B), confirming that they were more resistant to NITD-916.In contrast, the inhA I202S and inhA G205C mutants were significantly more sensitive to INH or C10 alone (Fig. 6C and D).When we cultured the inhA I202S and inhA G205C mutants in the presence of 10 ng/mL NITD-916 and enumerated the number of surviving bacteria after 10 days of treatment, we observed that while both inhA mutants were resistant to NITD-916, the inhA G205C mutant was significantly more resistant to killing by NITD-916 than inhA I202S (Fig. 6E).In addition, the degree of resistance to NITD-916 alone correlated with the level of increased resistance to the combination of C10 and NITD-916, where the inhA G205C mutant was significantly more resistant to the C10 and NITD-916 combination than the inhA I202S mutant (Fig. 6E).These data support that the potentiation of NITD-916 by C10 occurs through InhA inhibition, where NITD-916mediated inhibition of InhA is required for C10 to potentiate its bactericidal activity.Thus, these data support the model that C10 sensitizes Mtb to inhibition of InhA.

C10 sensitizes Mtb to inhibition of InhA without inhibiting InhA itself
One possible mechanism by which C10 could enhance the susceptibility of Mtb to InhA inhibition would be to decrease the expression or activity of InhA, thereby lowering the amount of INH-NAD or NITD-916 required to inhibit this target.When we analyzed our previously published RNA-sequencing data from C10-treated cultures, we found that after treatment with 25 µM C10 for 48 h, inhA was expressed at 1.1-fold relative to the untreated control (19), demonstrating that C10 does not decrease inhA expression at the transcriptional level.We next examined whether C10 compromises InhA activity by culturing Mtb with 14 C-acetate and measuring de novo mycolic acid biosynthesis by thin-layer chromatography (TLC).In the final steps of mycolic acid biosynthesis, the mycolic acid moiety is coupled to trehalose to form trehalose monomycolate (TMM), which can be transported out of the cell through the activity of the TMM flippase MmpL3 (45,46).The mycolic acid can then be trans-esterified from TMM to the arabinose moieties of arabinogalactan to form the inner leaflet of the mycolic acid layer that is covalently attached to the underlying cell wall layers (47)(48)(49).Alternatively, the mycolic acid can be trans-esterified to a second TMM molecule to form trehalose dimycolate (TDM) (50), which comprises a major component of the freely associated lipid layer that is intercalated within the covalently mycolic acids.Free mycolic acid, TMM, and TDM are the primary forms of mycolic acid that are not covalently linked to the cell wall, making them readily extractable and easy to separate by TLC, so we focused our analysis on these mycolic acid species as a read-out for de novo biosynthesis.Cultures of WT and katG W328L Mtb were labeled with 14 C-acetate for 20 hours in the presence or absence of C10 and/or INH before extracting whole cell lipids and monitoring the incorporation of the 14 C into mycolic acids by TLC (Fig. 7A through C; Fig. S4).As standards, we used TDM purified from H37Ra (Invivogen), free mycolic acid saponified and extracted from the H37Ra TDM, the representative fatty acid oleic acid, and free 14 C-acetate.Although we did not have a standard for TMM, we identified a band that we predict corresponds to TMM in our samples because it migrated slower than TDM due to the overall increased polarity and based on the migration pattern reported in published studies (51).As expected, INH treatment of WT Mtb decreased the intensity of bands correspond ing to free mycolic acid, TDM, and TMM (Fig. 7A through C; Fig. S4A through C), and significantly increased the intensity of a band that co-migrates with oleic acid (Fig. S4B).We believe this oleic acid co-migrating band represents FAS-I-generated fatty acids that serve as substrates for mycolic acid biosynthesis and, thus, it is not surprising that inhibition of InhA would lead to their accumulation (15)(16)(17).In contrast, INH treatment C) WT, inhA I202S , or inhA G205C Mtb was exposed to increasing concentrations of (B) NITD-916 or (C) INH for 1 week and OD 600 was quantified as a measure of growth.
The % inhibition was quantified relative to untreated controls (100% growth) and streptomycin-treated controls (0% growth) for each strain.(D) WT, inhA I202S , or inhA G205C Mtb was exposed to increasing concentrations of C10 and the % inhibition of Mtb growth and metabolism was determined using the resazurin assay, n = 3. Best fit curves were determined in GraphPad Prism.(E) WT, inhA I202S , or inhA G205C Mtb was exposed to 25 µM C10 and/or 10 ng/mL NITD-916 and CFU/mL was enumerated on days 0 and 10 of the treatment.A two-way ANOVA with Tukey's post test was used to identify statistically significant differences in panels A and E. IC 50 values were log-transformed before performing a one-way ANOVA with Tukey's post test to identify statistically significant differences in panels B-D.
of katG W328L Mtb did not affect the levels of free mycolic acids, TDM, or TMM over the 20-h period, supporting less efficient InhA inhibition due to the mutation in katG and decreased INH-NAD levels (Fig. 7A through C; Fig. S4).Addition of 25 μΜ C10 on its own or in combination with INH did not decrease the synthesis of free mycolic acid, TDM, or TMM in the WT or katG W328L Mtb (Fig. 7A through C; Fig. S4A through C).Since 0.25 µg/mL INH causes nearly complete inhibition of mycolic acid biosynthesis in WT Mtb, it may be difficult to detect a further decrease in synthesis with this concentration of INH.We titrated INH to identify a concentration that only caused partial inhibition of mycolic acid biosynthesis and found that 0.0625 µg/mL INH significantly decreased free mycolic acid, TDM, and TDM biosynthesis in WT Mtb, but to a lesser degree than 0.25 µg/mL INH (Fig. S4D through G).Using 0.0625 µg/mL INH, we again tested if C10 could enhance the ability of INH to block mycolic acid biosynthesis in WT Mtb.In addition to 25 µM C10, we also included 50 µM C10 to investigate if a higher concentration of C10 is required to observe effects on mycolic acids.While 25 µM C10 did not significantly decrease free mycolic acid, TDM, or TMM biosynthesis on its own, 50 µM C10 caused a subtle but significant decrease in the labeling of free mycolic acids and TDM and a significant increase in the labeling of TMM (Fig. S4E through G).However, 50 µM C10 is also growth inhibitory (Fig. 1B) and, therefore, the effects on mycolic acids at this concentration may merely reflect inhibition of Mtb growth.In addition, including 25 µM and 50 µM C10 in combination with 0.0625 µg/mL INH did not cause a further decrease in the 14 C-labeling of free mycolic acid, TDM, or TMM in WT Mtb (Fig. S4D through G).Monitoring TDM, and/or AU1235, and CFU/mL was enumerated on days 0 and 10, n = 6.Statistically significant differences were determined by two-way ANOVA and selected pairwise comparisons are depicted in the figure.ns, not significant; ***P < 0.001; ****P < 0.0001.For all pairwise comparisons, see Table S1.
TMM, and free mycolic acids only reflects changes in non-covalently attached mycolic acids, but the cell wall also harbors mycolic acids that are covalently attached to the arabinogalactan.In addition, mycolic acids produced by Mtb belong to three classes depending on their modifications, α-, methoxy-, and keto-mycolic acids, the relative proportions of which can impact the Mtb cell surface and physiology (52)(53)(54).To account for the covalently attached mycolic acids and to examine whether C10 alters the relative amounts of α-, methoxy-, or keto-mycolic acids, we derivatized whole cell extracts of Mtb to generate mycolic acid methyl esters, which liberates the covalently attached mycolic acids and also facilitates the separation of α-, methoxy-, and keto-mycolic acids.We found that treating WT Mtb with 25 µM C10 for 48 h did not significantly alter the relative amounts of α-, methoxy-, or keto-mycolic acid species (Fig. S4H and I).These findings suggest that 25 μΜ C10 does not alter the mycolic acid profile of Mtb.Since C10 enhances the bactericidal effect of direct InhA inhibition without enhanc ing the inhibition of mycolic acid biosynthesis in the presence of INH, we wondered whether the potentiating effects of C10 are specific for InhA inhibitors, or if C10 generally sensitizes Mtb to cell envelope targeting antibiotics.We examined whether C10 impacted the sensitivity of Mtb to killing by the arabinogalactan-targeting antibiotics EMB and BTZ-043 (55,56) and the MmpL3 inhibitor AU1235 (45).In combination with 2 µg/mL EMB or 0.05 µg/mL BTZ-043, 25 µM C10 increased Mtb survival (Fig. 7D).The finding that C10 antagonizes killing by EMB and BTZ-043 could be explained by the effect of C10 on energy homeostasis, since ETC inhibitors BDQ and CCCP also antagonize the bactericidal activity of EMB in Mycobacterium bovis (35).Consistent with the previous reports in M. bovis, treating Mtb with BDQ in combination with EMB recapitulates the antagonistic effect we observed with C10 (Fig. S4J), demonstrating that ATP depletion is sufficient to antagonize killing by EMB.Treatment of Mtb with 25 µM C10 in combination with the MmpL3 inhibitor AU1235 did not impact the number of surviving bacteria as compared to treatment with 0.25 µg/mL AU1235 alone (Fig. 7D).Since AU1235 inhibits mycolic acid metabolism downstream of InhA by blocking export of TMM to the cell envelope, these data suggest that the effects of C10 may be specific for inhibitors of mycolic acid synthesis that act upstream of MmpL3.While these studies do not rule out that C10 impacts the mycobacterial cell envelope in other ways, our data indicate that C10 specifically potentiates the bactericidal effect of InhA inhibitors without causing general susceptibility to cell envelope inhibitors.

DISCUSSION
C10 represents a new strategy to improve the efficacy of antibiotic therapy against TB, where it sensitizes Mtb to killing by the frontline antibiotic INH as well as the cytochrome bc 1 inhibitor Q203 that is in development for use in the clinic (19,57).However, the mechanism by which C10 potentiates the bactericidal effect of these compounds remained unknown.Using two mutants resistant to growth-inhibitory concentrations of C10, we determined that the ability of C10 to potentiate Q203 is linked to its effects on ATP homeostasis.In contrast, both C10-resistant mutants were still susceptible to C10-mediated INH potentiation, demonstrating that the mechanism of INH potentiation is not dependent on depletion of ATP by C10, uncoupling the effect of C10 on ATP production from the sensitization to INH.Therefore, these C10-resistant mutants allowed us to differentiate between the mechanisms underlying C10-mediated potentiation of INH and Q203.Through our work to understand the mode of action of C10, we have revealed that there exist novel mechanisms to enhance the bactericidal activity of existing antibiotics, highlighting the utility of using C10 as a chemical tool to dissect Mtb physiology and antibiotic susceptibility.
The mechanism by which the C10-resistant mutants suppressed the ATP-depleting effects of C10 and allowed for growth in the presence of C10 remains an open ques tion.An intergenic mutation in one of the C10-resistant mutants results in >100-fold upregulation of expression of the predicted SAM-methyltransferase Rv0731c and >2-fold upregulation of the secY-adk-mapA operon, while an intergenic mutation in the other C10-resistant mutant results in >4-fold upregulation of the lpdA-glpD2 operon.We also isolated a mutant with low-level C10 resistance that harbors a missense mutation in the predicted SAM-methyltransferase Rv0830.When we examine our previously published RNA-sequencing data set to determine if C10 causes dysregulated expression of any of these genes, we find that Rv0830, adk, and mapA are each significantly upregulated in response to 25 µM C10 by 1.64-fold, 1.55-fold, and 1.35-fold, respectively, whereas the other genes are unchanged (19).Therefore, these genes are part of the Mtb transcrip tional response to C10, perhaps to compensate for the inhibition of energy metabolism pathways during exposure to C10.Further supporting that the genes affected in the C10-resistant mutants represent a compensatory mechanism Mtb employs during C10 treatment and not the direct C10 target, the C10-resistant mutants maintain sensitivity to C10-mediated bactericidal activity in the presence of low pH and INH.In addition, these findings suggest that it is unlikely that the mechanism of resistance to C10 in these strains is mediated by inactivation of C10 by the enzymes that are upregulated, since we would expect that mutations that enable Mtb to degrade C10 or otherwise decrease its concentration would lead to resistance to all of the effects of C10.Further work is required to understand how these genes are involved in the ability of C10 to decrease ATP levels and inhibit growth in Mtb.
Instead of being associated with C10-mediated decreases in ATP, our data support a model where C10 potentiates INH by making Mtb particularly vulnerable to the inhibition of InhA, even in the INH-resistant katG hypomorphs that accumulate a significantly lower concentration of INH-NAD.Our findings show that C10 sensitizes Mtb to INH without changing the concentration of INH-NAD or decreasing the activity of its target InhA, which are the two predominant mechanisms of potentiating INH reported in the literature thus far (40,(58)(59)(60).In contrast to the potentiation strategies that increase the INH-NAD concentration or decrease InhA expression, which would be specific for INH or InhA inhibitors, respectively, it remains possible that C10 could sensitize Mtb more generally to additional inhibitors of mycolic acid biosynthesis.We found that C10 did not potentiate killing by the MmpL3 inhibitor AU1235, but it still remains to be tested whether C10 affects the sensitivity of Mtb to compounds that block steps in mycolic acid biosynthesis other than InhA that are upstream of MmpL3.Mycolic acid biosynthesis is essential in mycobacteria and not conserved in non-actinobacteria or eukaryotes, making this process a very attractive target for the development of specific antimycobac terials to treat TB and increasing the value of understanding how C10 sensitizes Mtb to inhibition of InhA.
The clinical utility of INH is currently being threatened by the increasing rates of INH-resistant TB cases.While resistance to INH can occur through multiple mechanisms, the predominant cause of INH resistance is mutation of katG, which accounts for an estimated 78.6% of INH-resistant strains (8).Clinically, mutations in katG are consid ered to confer a high level of resistance, often necessitating the use of an alternative treatment regimen.However, the overwhelming majority of these katG mutations are not null alleles.The most common resistance variant is an S315T amino acid substitution in KatG that decreases the enzyme's affinity for INH (8,61,62).The S315T mutation and other single amino acid substitutions that are commonly identified in resistant isolates significantly impair the ability of KatG to activate INH, but these mutations do not completely abolish INH-NAD synthesis by the KatG enzyme in vitro (62,63).The complete inactivation of KatG is likely detrimental to Mtb survival in the host due to the role of KatG in the oxidative stress response, which could explain why the majority of INH-resistant clinical isolates harbor single amino acid substitutions as opposed to more deleterious mutations (64)(65)(66).We found that C10 selectively potentiates killing by INH in katG mutants that retain some KatG enzymatic activity, indicating that it is possible to rescue the utility of INH against most clinically relevant INH-resistant isolates since most isolates will retain the low level of KatG activity.While the precise mechanism by which C10 induces sensitivity to InhA inhibition remains unclear, by deciphering how C10 promotes susceptibility to INH and NITD-916, we will reveal cryptic vulnerabilities in Mtb that can be exploited to enhance our current antimicrobial regimen.
In addition to resistance, INH efficacy can be limited by subinhibitory concentrations of antibiotics at the site of infection.For instance, a clinical study that quantified the distribution of antibiotics across lung lesions from TB patients within 24 h of dosing showed that approximately 35% of lesions harbored sub-inhibitory concentrations of INH, likely due to a combination of drug diffusion and host metabolism (3).Therefore, the Mtb within these lesions likely experience fluctuating concentrations of antibiotics that can be sub-inhibitory.C10 represents a possible strategy to sensitize Mtb to even sub-inhibitory concentrations of antibiotic, suggesting that in addition to circumventing INH resistance, C10 could enhance the efficacy of INH at the site of infection, although this remains to be tested.During the course of infection, Mtb is exposed to a myriad of host-derived stresses, including hypoxia within Mtb-containing lesions in the lung (67), as well as low pH within phagolysosomes in infected phagocytes (68,69).C10 mediates the killing of Mtb during low pH stress and more potent C10 analogs, such as 17 h, decrease Mtb growth in a cultured murine macrophage cell line (19,27), providing further evidence that C10 sensitizes Mtb to intracellular host-derived stresses.Thus, a better understanding of C10's mechanism of action could provide new opportunities for targeting Mtb at the site of infection.

Bacterial strains and growth conditions
Mtb Erdman strains (Table 1) were inoculated from a freezer stock into Middlebrook 7H9 liquid medium supplemented with 60 µL/L oleic acid, 5 g/L BSA, 2 g/L dextrose, 0.003 g/L catalase (OADC), 0.5% glycerol, and 0.05% Tween 80 and cultured at 37°C.Actively growing Mtb was then inoculated into Sauton's liquid medium [0.5 g/L KH 2 PO 4 , 0.5 g/L MgSO 4 , 4.0 g/L L-asparagine, 6% glycerol, 0.05 g/L ferric ammonium citrate, 2.0 g/L citric acid, and 0.01% (wt/vol) ZnSO 4 , pH 7.0] supplemented with 0.05% Tween 80 and grown to late-log phase before use in growth curve and survival experiments.The ΔkatG strain was generated using specialized transduction with the tempera ture-sensitive phage phAE87 engineered to harbor sequence homologous to regions upstream (Erdman nucleotides 2144027-2144776) and downstream (Erdman nucleoti des 2146989-2147705) of katG and selected on 50 µg/mL hygromycin as previously described (70).For Mtb growth and survival experiments, Mtb was inoculated into roller bottles containing Sauton's medium supplemented with Tween 80 at a starting OD 600 of 0.1, and growth was measured by OD 600 and survival was monitored as CFU/mL.Viable CFU from bacterial cultures were enumerated on Middlebrook 7H11 agar medium supplemented with OADC and 0.5% glycerol and plates were incubated at 37C with 5% CO 2 for 2-3 weeks.To select for mutants resistant to both C10 and INH, the equivalent of 0.5 mL of OD 600 = 1.0 of Mtb growing in 7H9 + OADC media was spread on Sauton's agar containing 25 µM C10 and 0.2 µg/mL INH and incubated at 37°C and 5% CO 2 for 4 months.Isolated colonies were passaged on agar containing 25 µM C10 and 0.2 µg/mL INH to ensure that they were resistant before performing whole-genome sequencing.For agar growth assays, 0.003 g/L bovine catalase was included in the Sauton's agar, and the equivalent of 2.5 mL of OD 600 = 1.0 of Mtb was spread on the agar surface to enhance the reproducibility of Mtb growth on the Sauton's agar plates.

Whole-genome sequencing
Genomic DNA was isolated using cetyltrimethylammonium bromide-lysozyme lysis, followed by phenol-chloroform-isoamyl alcohol extraction and isopropanol precipita tion, as previously described (71).Whole-genome sequencing was performed by use of an Illumina NovaSeq 6000.The identification of single nucleotide polymorphisms was done using SeqMan NGen software (DNASTAR).The genomes were assembled and compared to the genomic DNA from the WT parental control strain, using Integrative Genome Viewer to visualize and confirm changes within regions of interest (72).

Preparation of compounds
C10 was synthesized using previously described methods (73,74) and prepared as an imidazole salt as described previously (19).Stocks of C10-imidazole were resuspen ded in DMSO.INH (Sigma) was dissolved in water, and NITD-916, BDQ, EMB (Sigma), BTZ-043, and AU1235 (MedChem Express) were dissolved in DMSO.In all experiments, the concentration of both DMSO and imidazole was normalized across all samples to ensure that any differences were due to the effect of the indicated compounds and not due to DMSO or imidazole.

Detection of ATP
Mtb growing in Sauton's media was treated with the indicated concentration of C10 for 24 h before the OD 600 of each culture was measured and samples were removed.Mtb samples were inactivated at >95°C for 20 min and stored at −20°C until analyzing the ATP levels using the BacTiter Glo assay (Promega) as previously described (19).Samples were diluted 1:10, then mixed 1:1 with the BacTiter Glo reagent in a white, opaque 96-well dish, and the luminescence was read on a Synergy HT plate reader with a 1-s integration.The relative luminescence units (RLUs) were calculated by subtracting the luminescence of a media-only control from the luminescence value of each sample.The RLU/OD 600 was determined to account for differences in bacterial density, and the fold change in each sample was calculated relative to the average of the 0 µM C10 control from that experiment to facilitate the combining of multiple experiments onto a single graph.

Resazurin IC 50 assays
Logarithmically growing Mtb was inoculated into Sauton's medium in 96-well plates with wells containing increasing concentrations of the indicated compounds.Mtb was inoculated at an OD 600 of 0.0025 in 200 µL per well.The plates were incubated at 37°C in 5% CO 2 for 1 week, at which point 32.5 µL of a mixture containing an 8:5 ratio of 0.6 mM resazurin (Sigma) dissolved in 1× phosphate-buffered saline (PBS) to 20% Tween 80 was added, and the production of fluorescent resorufin was measured on a Synergy HT plate reader with excitation λ ex = 530 nm and emission λ em = 590 nm after incubation at 37°C in 5% CO 2 overnight.For each assay, medium alone served as a negative control, and untreated Mtb was included as a positive control.The percent inhibition was calculated

Growth inhibition IC 50 assays
Logarithmically growing Mtb was inoculated into Sauton's medium in 96-well plates with wells containing increasing concentrations of the indicated compounds.Mtb was inoculated at an OD 600 of 0.05 in 200 µL per well.The plates were incubated at 37°C in 5% CO 2 for 1 week, and the absorbance at wavelength λ = 600 nm was measured on a Synergy HT plate reader.For each assay, wells containing media alone served as a blank, 400 µg/mL streptomycin was included as a negative control, and untreated Mtb was included as a positive control.The percent inhibition was calculated similarly to that described for the resazurin assay above.

Quantitative reverse transcription PCR (qRT-PCR)
RNA was isolated from WT, GHTB136, and GHTB146 Mtb growing in Sauton's medium using TRIzol, and purified by chloroform extraction followed by isopropanol precipita tion.cDNA was prepared using the SuperScript III first-strand synthesis kit (Invitrogen), and qPCR was performed using an SYBR green kit (Bio-Rad) with gene-specific primers (Table 2) on a CFX96 Real-Time System (Bio-Rad).The relative expression of genes was calculated using the 2 −ΔΔCt method, with sigA serving as an internal reference control gene.

Western blot for KatG protein
Mtb samples were pelleted and resuspended in buffer containing 10 mM sodium phosphate pH 8.0, 150 mM NaCl, 2 mM EDTA, 1 mM PMSF, 0.1% NP-40, and a 1× protease inhibitor cocktail (Roche), then lysed by bead beating, and filtered two times through a 0.22-µm Spin-X column (Costar) to remove unlysed Mtb.SDS-polyacrylamide gel electrophoresis was performed and samples were transferred to a nitrocellulose membrane, after which KatG was detected using a mouse monoclonal α-KatG anti body used at 1:500 dilution (clone IT-57; BEI Resources).Either CarD or RpoB served as a loading control, using a mouse monoclonal α-CarD antibody at 1:2,000 dilution (clone 10F05; Memorial Sloan-Kettering Cancer Center) or a mouse monoclonal α-RpoB antibody at 1:1000 dilution (clone 8RB13; Neoclone).The membrane was probed with a goat anti-mouse antibody conjugated to horseradish peroxidase and bands were visualized using the Western Lighting Plus-ECL reagent (PerkinElmer).When performing the KatG expression analysis in response to C10 treatment, the amount of protein in each sample was measured by BCA (Pierce) and the amount of protein loaded in each lane was normalized to 67 ng to facilitate comparisons between samples.

Purification of Mtb KatG protein
The Mtb katG coding region was cloned into NotI and EcoRI sites in the pGEX-6P-1 expression vector to translationally fuse glutathione-S-transferase to the N-terminus of the KatG protein and the expression of the fusion protein was induced in logarithmically growing E. coli BL21-DE3 cells by treating 1 L of cells with 0.1 mM IPTG for 4 h.Cells were pelleted, resuspended in 20 mL 1× PBS containing a 1× protease inhibitor cocktail, and lysed two times in a cell disruptor.Lysate was treated with 9 U/mL benzonase (Sigma) and clarified by centrifugation.GST-KatG was purified from the supernatant by incubating lysate overnight with glutathione agarose resin (Goldbio), washed with 300 mL 1× PBS, and eluted from the resin by cleaving the KatG protein from the GST tag using PreScission protease in buffer containing 50 mM Tris-HCl pH 7.0, 150 mM NaCl, 1 mM EDTA, and 1 mM DTT.

Thermal shift assay on KatG protein
The melting temperature (T m ) of KatG in the presence of C10 or INH was determined by differential scanning fluorimetry.Purified Mtb KatG protein (1.1 µM) was incubated with INH or C10 in 50 mM potassium phosphate buffer pH 7.0 containing 50 µM MnCl 2 before mixing samples with SYPRO Orange (ThermoFisher) at a final concentration of 1× in a 96-well PCR plate.The plate was incubated for 5 s at increasing temperatures in 0.5°C increments from 10 to 95°C and fluorescence was monitored over time in the HEX channel on a CFX96 Real-Time System (Bio-Rad).The T m was calculated by fitting curves with a Boltzmann sigmoidal equation in GraphPad Prism, and the ΔT m was calculated as the difference between the T m of each sample and the average of the untreated control samples.

Hydrogen peroxide degradation assay
The catalase activity of either purified KatG or Mtb cell lysates was measured in a UV/Vis spectrophotometer using quartz cuvettes.For in vitro assays of purified KatG activity, 1.9 mL of 50 mM potassium phosphate buffer pH 7.0 containing 25 nM purified KatG was incubated with or without C10 and/or INH at 30°C for 5 min and the sample was used to blank the spectrophotometer before the indicated concentration of H 2 O 2 was added to initiate the reaction, bringing the final volume of the reaction to 2 mL.For samples containing Mtb lysate, the samples were bead beat as described above and the protein concentration in the lysate was measured by BCA (Pierce).Samples were normalized such that 25 µg of total cell protein was present in each assay sample in 1.9 mL of 50 mM potassium phosphate buffer pH 7.0, and the reactions were initiated with the addition of H 2 O 2 to a concentration of 5 mM in 2 mL final volume.The absorbance at 240 nm was read every 10 s for 2 min, and the negative slope of the curve was used to calculate the rate of H 2 O 2 degradation.The absorbance was converted to molarity using the extinction coefficient for H 2 O 2 , ε 240 = 43.6M −1 cm −1 .

INH-activation assay and detection of INH-NAD in Mtb
The  and the indicated concentration of C10 in a final volume of 1 mL.At the indicated time points, 100 µL was removed from the reaction and inactivated in 100 µL ice cold methanol and stored at −20°C before liquid chromatography/mass spectrometry (LC/ MS).To monitor INH activation in live Mtb, 50 mL cultures of Mtb in Sauton's media without Tween 80 were treated with the indicated concentration of C10 and/or INH for 3 days.To extract polar metabolites, the cultures were pelleted, washed twice in H 2 O, and resuspended in 1.5 mL of 2:1 chloroform:methanol in glass conicals.Samples were kept on ice and vortexed each for 1 min in 20-s intervals, and stored at 4°C overnight.Then 375 µL of H 2 O was added, the samples were vortexed for 1 min in 20-s intervals, keeping the samples on ice.Samples were incubated at room temperature for 1 h with constant agitation, and then centrifuged for 10 min at 1000 RPM.The top aqueous layer was transferred to a fresh 1.5 mL microcentrifuge tube, stored overnight at −20°C, centrifuged for 5 min to pellet any insoluble material, and the supernatant was transferred to a fresh tube.INH-NAD was detected using methods similar to those previously described (40,59,75) with some modifications.Ultra high-performance LC (UHPLC)/MS was performed with an Agilent 1290 Infinity UHPLC system interfaced with an Agilent 6530 QTOF mass spectrometer.Hydrophilic interaction liquid chromatography (HILIC) analysis was performed by using a HILICON iHILIC-(P) Classic column with the following specifications: 100 mm × 2.1 mm, 5 µm.Mobile-phase solvents were composed of A = 20 mM ammonium bicarbonate, 0.1% ammonium hydroxide (adjusted to pH 9.2) and 2.5 µM medronic acid in water:acetonitrile (95:5) and B = 2.5 µM medronic acid in acetonitrile:water (95:5).The column compartment was maintained at 45°C for all experiments.The following linear gradient was applied at a flow rate of 250 µL/min: 0-1 min: 90% B, 1-12 min: 90-35% B, 12-12.

Measurement of de novo lipid synthesis by 14 C labeling and TLC
To monitor de novo mycolic acid biosynthesis, Mtb growing in Sauton's liquid medium was adjusted to OD 600 of 0.5, treated with the indicated concentrations of C10 and/or INH in 1 mL final volume, and immediately exposed to 2 µCi/mL 14 C-acetate (Perki nElmer).After incubation at 37°C for 20 h, the cells were pelleted, and resuspended in 2:1 chloroform:methanol, vortexed, then samples were pelleted to remove insoluble cell debris, and the supernatant was transferred to a fresh vial.To separate lipid species by TLC, 40 µL of the sample was added dropwise to an HPTLC plate coated with silica gel 60 matrix (Sigma).TLC plates were developed in 75:10:1 chloroform:methanol:H 2 O and imaged by phosphorimaging on a Typhoon laser-scanner (Cytiva).The intensity of each band was quantified in ImageJ, normalized to the total intensity in the whole lane, and the fold change was quantified relative to the untreated control for each replicate.To assign putative identities to relevant bands, standards for 14 C-acetate, TDM, free mycolic acid, and oleic acid were run on each plate.TDM purified from H37Ra (Invivogen) was derivatized to generate a free mycolic acid standard using methods similar to those previously described (76).Briefly, 100 µL of 0.5 mg/mL TDM in isopropanol was subjected to an alkaline ester hydrolysis by mixing with 2 µL H 2 O and 5 µL of 10M KOH, and heating to 90°C for 1 h to ensure efficient saponification.The resulting mycolic acids were purified from the trehalose by neutralizing the reaction with 50 µL 1.2M HCl, adding 100 µL chloroform, 100 µL H 2 O, vortexing, and separating the organic phase from the aqueous layer to obtain free mycolic acids in chloroform.
To monitor α-, methoxy-, and keto-mycolic acid biosynthesis, we adapted previously published methods (54,77,78), with some alterations.Mtb growing in Sauton's liquid medium was adjusted to OD 0.5 and treated with the indicated concentrations of C10 or INH in 1 mL final volume of media containing 2 µCi/mL 14 C-acetate.After incubation at 37°C for 48 h, cells were pelleted, resuspended in 1 mL H 2 O, transferred to 15 mL glass conicals containing 1 mL of 40% tetrabutylammonium hydroxide (Sigma), and then heated at 100°C overnight.This was followed by the addition of 2 mL of dichloromethane (Sigma) and 100 µL iodomethane (ThermoFisher) followed by constant mixing for 1 h.The upper aqueous phase was then discarded and the remaining liquid in the organic layer was allowed to evaporate overnight.The resulting mycolic acid methyl esters were resuspended in 1 mL diethyl ether (Sigma) and then transferred to 2 mL glass vials.Diethyl ether was then evaporated and resulting mycolic acid methyl esters were resuspended in 200 µL dichloromethane.Equal counts (2,500 CPM) were loaded onto TLC plates and developed three times in 85:15 petroleum ether:diethyl ether.

FIG 1
FIG 1 Isolation and characterization of C10-resistant mutants.(A) Chemical structures of C10 and 17h.(B) WT Mtb was cultured in Sauton's medium containing the indicated concentration of C10, and growth was measured by OD 600 over time.(C) WT Mtb was spread on Sauton's agar medium containing the indicated concentration of C10 and incubated at 37°C for 3 weeks.(D) Whole-genome sequencing of 11 C10-resistant mutants revealed three groups of mutants.The mutant loci are depicted along with the resultant nucleotide or amino acid change, and the GHTB strain numbers indicate the mutant isolates

21 FIG 3
FIG 3 Forward genetic selection results in isolation of a katG W198* mutant that is resistant to INH even in the presence of C10.(A) The indicated strain of Mtb was spread on Sauton's agar medium containing 25 µM C10 and/or 0.5 µg/mL INH and incubated at 37°C for 3 weeks.(B) WT, (C) katG FS6 , (D) katG W328L , and (E) katG W198* Mtb were cultured in Sauton's liquid medium containing the indicated concentrations of C10 and INH, and the CFU/mL was enumerated.The legend in panel B applies to panels B-E.(F) Mtb harboring the indicated katG allele was cultured in Sauton's liquid medium containing 0, 10, or 25 µM C10 for 24 h before ATP levels were measured by the BacTiter Glo assay.The RLU was normalized to the OD 600 of the culture to control for differences in cell density.Fold change in (Continued on next page)

FIG 4
FIG 4 KatG activity is necessary for C10 to potentiate INH.(A) Western blot against whole-cell lysate from the indicated strain of Mtb.CarD was used as a loading control in this experiment.(B) Diagram of KatG protein indicating the required catalytic residues in the protein as well as the variant proteins expressed in each katG mutant.(C-G) The indicated strain of Mtb was treated with or without 0.25 µg/mL INH and growth was monitored over time by the optical density (OD 600 ), n = 4. (H) ΔkatG Mtb was spread on Sauton agar medium containing 25 µM C10 and/or 0.5 µg/mL INH and incubated at 37°C for 3 weeks before images were taken.(I) ΔkatG Mtb was cultured in Sauton's liquid medium containing 0, 10, or 25 µM C10 with or without 0.25 µg/mL INH and the CFU/mL were enumerated on 7H11 agar with no antibiotics to determine the number of viable bacteria in each culture at days 0 and 10 of treatment.A two-way ANOVA with Tukey's post test was performed to determine statistically significant differences.ns, not significant; ****P < 0.0001.Relevant comparisons are depicted in the figure, but for all pairwise comparisons, please see TableS1.

FIG 5
FIG 5 C10 enhances Mtb sensitivity to INH-NAD without changing its activation.(A-D) Purified KatG protein (200 nM) was incubated in 50 mM potassium phosphate buffer pH 7.0 containing INH, NAD + , and 50 µM MnCl 2 with 0, 10, or 25 µM C10, and samples were taken at the indicated time points and analyzed via LC/MS.For each ion, the peak intensity was calculated as the area under the curve for (A) INH-NAD, (B) INH, (C) NAD + , and (D) C10, n = 3.The legend in panel D applies to panels A-D.(E) Either WT (top) or katG FS6 Mtb (bottom) was cultured in Sauton's liquid medium in the presence of 0, 5, or 25 µM C10 for 3 days, andwhole-cell lysate was collected for a Western blot using a monoclonal α-KatG antibody or an α-RpoB antibody as a loading control, n = 3.Note that the KatG blot from the katG FS6 samples has a high amount of background since the exposure time for this blot is substantially longer than for WT samples due to the very low level of KatG protein in this strain.(F) WT or katG FS6 Mtb was cultured in Sauton's liquid medium in the presence of 0, 10, or 25 µM C10 for 6 days, and whole-cell lysate was collected to measure the catalase activity of the sample as a proxy for KatG activity, n = 2. Catalase activity is normalized to the amount of total protein in the whole-cell lysate sample.Note that lysate from the ΔkatG mutant, n = 3, harbors no detectable amount of catalase activity compared to buffer alone, n = 3, demonstrating that the assay is specific for KatG.(G) Mtb with the indicated katG allele was cultured for 3 days in Sauton's liquid medium in the presence and absence of 25 µM C10 and/or 0.25 µg/mL INH, after which polar metabolites were extracted from the culture and analyzed by LC/MS.The area under the curve was calculated by the integration of the peak to determine the peak intensity of INH-NAD in each bacterial sample, n = 3. Statistically significant differences were identified by Brown-Forsyth/Welch's one-way ANOVA with Dunnet's T3 post test in panel F and a two-way ANOVA with Tukey's post test in panel G, and the relevant comparisons are indicated on the graph.ns, not significant; ***P < 0.001; ****P < 0.0001.For all pairwise comparisons, see TableS1.

FIG 6
FIG 6 C10 enhances the vulnerability of Mtb to direct InhA inhibition.(A) WT Mtb was cultured in Sauton's liquid medium with the indicated concentrations of C10, INH, and/or NITD-916, and CFU/mL was enumerated on days 0 and 10 of treatment to determine the number of viable bacteria in each sample, n = 6.(B and (Continued on next page)

FIG 7
FIG 7 C10 sensitizes Mtb to InhA inhibition without altering bacterial InhA activity or making Mtb generally susceptible to cell envelope inhibitors.(A) WT or katG W328L Mtb was cultured in Sauton liquid medium, treated with 25 µM C10 and/or 0.25 µg/mL INH, and immediately exposed to 2 µCi/mL of 14 C-labeled acetate.After 20 h, lipids were extracted and analyzed by TLC to measure the de novo synthesis of mycolic acids and other lipids.The TLC plate was developed with 75:10:1 Chloroform:Methanol:H 2 O and radioactivity was analyzed by phosphorimaging.Bands corresponding to free mycolic acid (MA), fatty acid, trehalose dimycolate (TDM), and 14 C-acetate were identified by comigration with a standard.Trehalose monomycolate (TMM) is indicated with a * to emphasize that this lipid is putatively identified, and not correlated with a standard.The plate in panel A is representative of three biological replicates, and the standards and additional replicates are shown in Fig. S4.(B and C) The intensity of bands corresponding to (B) TDM or (C) free MA were quantified in ImageJ, and normalized to the DMSO sample, with each WT sample being normalized to WT DMSO and each katG W328L sample being normalized to katG W328L DMSO in order to compare across replicates on separate plates, n = 3. (D) WT Mtb was cultured in Sauton's liquid medium with the indicated concentrations of C10, INH, EMB, BTZ-043, as the {[(fluorescence of the positive control − fluorescence of the negative control) − (fluorescence of the sample − fluorescence of the negative control)]/(fluorescence of the positive control − fluorescence of the negative control)} × 100.

TABLE 1
Mtb strains used in this study G205CErdman selected on 10 ng/mL NITD-916 and 25 µM C10, harbors a G to T single nucleotide substitution in inhA resulting in a G205C amino acid change This study activation of INH was monitored in vitro using 200 nM purified KatG protein in 50 mM potassium phosphate buffer pH 7.0, 50 µM NAD + , 50 µM INH, 50 µM MnCl 2 ,

TABLE 2
Primers used in this study