Impaired Succinate Oxidation Prevents Growth and Influences Drug Susceptibility in Mycobacterium tuberculosis

ABSTRACT Succinate is a major focal point in mycobacterial metabolism and respiration, serving as both an intermediate of the tricarboxylic acid (TCA) cycle and a direct electron donor for the respiratory chain. Mycobacterium tuberculosis encodes multiple enzymes predicted to be capable of catalyzing the oxidation of succinate to fumarate, including two different succinate dehydrogenases (Sdh1 and Sdh2) and a separate fumarate reductase (Frd) with possible bidirectional behavior. Previous attempts to investigate the essentiality of succinate oxidation in M. tuberculosis have relied on the use of single-gene deletion mutants, raising the possibility that the remaining enzymes could catalyze succinate oxidation in the absence of the other. To address this, we report on the use of mycobacterial CRISPR interference (CRISPRi) to construct single, double, and triple transcriptional knockdowns of sdhA1, sdhA2, and frdA in M. tuberculosis. We show that the simultaneous knockdown of sdhA1 and sdhA2 is required to prevent succinate oxidation and overcome the functional redundancy within these enzymes. Succinate oxidation was demonstrated to be essential for the optimal growth of M. tuberculosis, with the combined knockdown of sdhA1 and sdhA2 significantly impairing the activity of the respiratory chain and preventing growth on a range of carbon sources. Moreover, impaired succinate oxidation was shown to influence the activity of cell wall-targeting antibiotics and bioenergetic inhibitors against M. tuberculosis. Together, these data provide fundamental insights into mycobacterial physiology, energy metabolism, and antimicrobial susceptibility.

IMPORTANCE New drugs are urgently required to combat the tuberculosis epidemic that claims 1.5 million lives annually. Inhibitors of mycobacterial energy metabolism have shown significant promise clinically; however, further advancing this nascent target space requires a more fundamental understanding of the respiratory enzymes and pathways used by Mycobacterium tuberculosis. Succinate is a major focal point in mycobacterial metabolism and respiration; yet, the essentiality of succinate oxidation and the consequences of inhibiting this process are poorly defined. In this study, we demonstrate that impaired succinate oxidation prevents the optimal growth of M. tuberculosis on a range of carbon sources and significantly reduces the activity of the electron transport chain. Moreover, we show that impaired succinate oxidation both positively and negatively influences the activity of a variety of antituberculosis drugs. Combined, these findings provide fundamental insights into mycobacterial physiology and drug susceptibility that will be useful in the continued development of bioenergetic inhibitors.
KEYWORDS CRISPR interference, Mycobacterium tuberculosis, succinate dehydrogenase T uberculosis (TB) is a leading cause of infectious disease morbidity and mortality globally, with 10 million new cases and 1.5 million deaths reported in 2020 (1). Successful treatment of infections with Mycobacterium tuberculosis, the causative agent of the disease, requires 6 months of combination therapy involving 2 months of intensive treatment with ethambutol (EMB), isoniazid (INH), rifampicin (RIF), and pyrazinamide (PZA), followed by a further 2 months of treatment with INH and RIF (2). Infections with multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains of M. tuberculosis require up to 2 years of chemotherapy with a combination of 8 to 10 antibiotics and have cure rates as low as 39% for XDR-TB infections (1). Thus, global TB control urgently depends on the development of new drugs and regimens that can quickly and effectively treat both drug-sensitive and drug-resistant TB infections.
Succinate oxidation is a major focal point in mycobacterial energy generation and directly couples central carbon metabolism to the respiratory chain (29). The oxidation of succinate to fumarate is catalyzed by succinate dehydrogenase (SDH) enzymes, which couple succinate oxidation in the tricarboxylic acid (TCA) cycle to the reduction of menaquinone in the electron transport chain (ETC) (succinate 1 menaquinone $ fumarate 1 menaquinol) (30,31). Fumarate reductase (FRD) enzymes catalyze the reverse reaction: fumarate reduction coupled to menaquinol oxidation. SDH and FRD enzymes encoded by bacteria are highly similar and are often functionally interchangeable (31,32). M. tuberculosis encodes three different SDH/FRD enzymes which are all predicted to be capable of catalyzing the oxidation of succinate to fumarate (33). This includes two different succinate dehydrogenase enzymes (Sdh1, Rv0249c to Rv0247c, and Sdh2, Rv3316 to Rv3319), as well as a separate fumarate reductase with possible bidirectional behavior (Frd, Rv1552 to Rv1555) (33). All three enzymes have distinct phylogenies, prosthetic groups, and predicted biochemistries (31,(34)(35)(36)(37) and are differentially expressed in M. tuberculosis (38)(39)(40), suggesting that these enzymes perform distinct, but overlapping, roles.
Previous attempts to elucidate the function of these individual enzymes in M. tuberculosis have focused on the characterization of single-gene deletion mutants (38,41). However, this leaves the possibility that the remaining enzymes can carry out succinate oxidation in the absence of the other. Consistent with this, single deletion mutants in M. tuberculosis only have minor (i.e., for Dsdh1 and Dsdh2) (41) or no (i.e., for Dfrd) (38) phenotypes, while the construction of double deletion mutants was unsuccessful (41). Similarly, two recent whole-genome CRISPR interference (CRISPRi) studies showed that the depletion of either sdhA1, sdhA2, or frdA alone had no effect on the growth or viability of M. tuberculosis (42,43). Moreover, while high-throughput transposon hybridization (TraSH) screens have hinted at the essentiality of SDH enzymes for mycobacterial growth and persistence (3,44,45), they have also been unable to overcome the functional redundancy in SDH catalysis. Consequently, the essentiality of succinate oxidation in M. tuberculosis is incompletely understood.
Here, we used mycobacterial CRISPR interference (CRISPRi) (46)(47)(48) to transcriptionally repress the expression of sdhA1, sdhA2, and frdA in M. tuberculosis alone and in combination, allowing us to overcome the functional redundancy within these enzymes. We show that succinate oxidation is essential for the optimal growth of M. tuberculosis, with the simultaneous knockdown of sdhA1 and sdhA2 (sdhA1 1 sdhA2) significantly impairing the activity of the respiratory chain and preventing growth on a range of carbon sources. Moreover, we demonstrate that impaired succinate oxidation both positively and negatively affects the susceptibility of M. tuberculosis to a variety of anti-TB drugs.

RESULTS
Succinate oxidation is essential for the optimal growth of M. tuberculosis. To elucidate the roles and essentialities of the SDH and FRD enzymes in M. tuberculosis, we used CRISPR interference (46) to knock down the expression of Sdh1, Sdh2, and Frd alone and in combination. Three guide RNAs (i.e., single guide RNA [sgRNA]) were designed to target the catalytic A subunit of Sdh1, Sdh2, and Frd ( Fig. 1A to C), which resulted in high-level transcriptional repression (50-to 100-fold) of target genes in single and multiplexed constructs ( Fig. 1D and E). Interestingly, expression of the sgRNA targeting sdhA1 also resulted in reduced expression of frdA in multiplexed repression constructs, regardless of the sgRNA used ( Fig. 1E and see also Fig. S1A in the supplemental material), suggesting a shared regulatory mechanism governing sdh1 and frd gene expression.
Transcriptional repression of either sdhA1, sdhA2, or frdA alone did not affect the growth of M. tuberculosis strain mc 2 6230 in 7H9 medium supplemented with OADC (i.e., with glucose and oleic acid as carbon sources) (Fig. 1F), consistent with the previously published single deletion mutants (38,41) and transcriptional knockdowns (42,43). Simultaneous knockdown of both sdhA1 1 sdhA2 with or without frdA (sdhA1 1 sdhA2 6 frdA) was required to prevent growth in 7H9-OADC medium ( Fig. 1G and Fig. S1B) and had bacteriostatic consequences on cell viability ( Fig. 1H and Fig. S1C). The combined knockdown of frdA with either sdhA1 or sdhA2 displayed no growth defect in the same media (Fig. 1G).
To further understand the contribution of each enzyme to mycobacterial growth, we investigated the ability of the M. tuberculosis SDH/FRD knockdown strains to grow on a range of fermentable and nonfermentable carbon sources. When provided with succinate as the sole carbon and energy source, the knockdown of sdhA1 alone, but not sdhA2 or frdA, had a minor (approximately 10%) defect in growth rate ( Fig. 2A), consistent with its proposed role as the primary aerobic SDH (41). For all single transcriptional knockdowns, no significant difference in growth compared to a nontargeting control was observed in media containing glucose, glycerol, acetate, or a mixture of glucose and acetate ( Fig. 2B to E). This demonstrates that these enzymes are individually dispensable for the growth of M. tuberculosis, regardless of carbon source.
The combined transcriptional repression of sdhA1 1 sdhA2 6 frdA prevented the optimal growth of M. tuberculosis across all carbon sources tested ( Fig. 2A to E). M. tuberculosis repressing sdhA1 1 sdhA2 6 frdA was unable to grow with the nonfermentable carbon source succinate or acetate as the sole carbon and energy source ( Fig. 2A and B). The growth impairment of the sdhA1 1 sdhA2 6 frdA double or triple knockdown strain was partially rescued when using fermentable carbon sources such as glycerol or glucose, or a combination of acetate and glucose, although growth was still significantly impaired ( Fig. 2C to E). Overall, these data demonstrate that a functional SDH enzyme is essential for the optimal growth of M. tuberculosis, regardless of carbon source.
SDH and FRD enzymes are dispensable for the survival of M. tuberculosis under nutrient starvation. We next sought to investigate the contribution of SDH and FRD enzymes to the survival of M. tuberculosis under nutrient-starved nonreplicating persistence (39). We first depleted cells of SDH/FRD enzymes by growing strains for 8 days in the presence of anhydrotetracycline (ATc) (100 ng/mL), before placing knockdown strains under nutrientstarved conditions (i.e., phosphate-buffered saline [PBS]). Individual depletion of either Sdh1, Sdh2, or Frd did not affect the survival of M. tuberculosis over 8 weeks of nutrient starvation (Fig. 2F). Moreover, all double and triple transcriptional knockdown strains had no survival defect under nutrient-starved conditions (Fig. 2G). Together, this suggests that SDH and FRD enzymes are dispensable for the survival of M. tuberculosis under nutrient starvation.
Impaired succinate oxidation significantly disrupts the activity of the mycobacterial respiratory chain. The oxidation of succinate to fumarate results in two electrons being donated to the ETC (31). Therefore, we sought to determine the bioenergetic consequences of impaired succinate oxidation for the activity of the M. tuberculosis respiratory chain. To achieve this, we predepleted cells of SDH/FRD enzymes by inducing transcriptional repression for 8 days prior to measuring oxygen consumption rates (OCR) as a proxy for respiratory activity. We measured the OCR when cultures were grown with OADC, or with glycerol or succinate as the sole carbon and energy source, as these were the growth conditions under which the SDH/FRD knockdown strains were the least and most impaired, respectively ( Fig. 2A and C).
When grown in media supplemented with OADC, only the combined knockdown of sdhA1 1 sdhA2 was able to overcome the functional redundancy in succinate oxidation and  (Fig. 3A). All other combinations of transcriptional repression did not influence the OCR of M. tuberculosis when grown with OADC (Fig. 3A). In contrast, with succinate as the sole carbon source, the knockdown of sdhA1 alone reduced the OCR of M. tuberculosis by approximately 20% (Fig. 3B). This is consistent with the growth impairment of this strain when grown on succinate ( Fig. 2A) and suggests that Sdh1 is the dominant SDH enzyme in M. tuberculosis. No significant difference in OCR was observed for the sdhA1 knockdown strain in media supplemented with glycerol (Fig. 3C). The combined repression of sdhA1 1 sdhA2 6 frdA significantly reduced the respiration rate of M. tuberculosis across all three carbon sources tested ( Fig. 3A to C). The double and triple sdhA1 1 sdhA2 6 frdA knockdown strains had a 50% reduction in OCR compared to their respective no-ATc controls in media containing glycerol and a 70% reduction in OCR when energized with OADC or succinate ( Fig. 3A to C). To further investigate the contribution of FIG 2 Succinate oxidation is essential for the optimal growth of M. tuberculosis but is dispensable for survival under nutrient starvation. (A to E) Growth profiles of M. tuberculosis single and multiplexed sdhA1, sdhA2, and frdA knockdown strains in 7H9 medium supplemented with 30 mM succinate (A), 0.2% acetate (B), 0.2% glycerol (C), 20 mM glucose (D), or a combination of 0.1% acetate and 10 mM glucose (E). Cultures were grown in 10-mL volumes from a starting OD 600 of 0.005, and gene knockdown was induced at time zero with ATc (100 ng/mL). An mmpl3 knockdown strain was included as a positive inhibition control. The means and standard deviation for three replicates are shown. Each experiment was repeated twice. (F and G) Consequences of the single (F) or multiplexed (G) depletion of Sdh1, Sdh2, or Frd for the survival of M. tuberculosis under a nutrient starvation model of nonreplicating persistence. Before entering nutrient starvation, cells were predepleted of SDH and FRD enzymes by inducing transcriptional repression of target genes for 8 days. Cultures were then harvested, washed twice in PBS, and nutrient starved by inoculation into inkwells containing PBS plus tyloxapol. Viability was measured by enumerating CFU per milliliter over 8 weeks. Error bars represent the standard deviation from three replicates.
Essentiality of Succinate Oxidation in M. tuberculosis mBio succinate oxidation to total respiratory flux, we used Q203 and ND-011992 to inhibit the activity of both terminal oxidases (cytochrome bc 1 :aa 3 and cytochrome bd, respectively) and abolish respiration (49,50). The complete inhibition of aerobic respiration highlighted an additional ;30% respiratory capacity in the endogenous respiration rate of M. tuberculosis in the absence of succinate oxidation (Fig. 3D). Combined, these results demonstrate that succinate oxidation is the master driver of M. tuberculosis respiration, regardless of carbon source. This may occur directly (i.e., through succinate being the major contributor of electrons to the ETC) or indirectly through SDH activity regulating the activity of the TCA cycle. Essentiality of Succinate Oxidation in M. tuberculosis mBio Impaired succinate oxidation influences the drug susceptibility of M. tuberculosis. TB is treated with multidrug therapies, and as such, any new drug needs to be effective in combination. While bioenergetic inhibitors are proving promising at reducing treatment times, the consequences of inhibiting respiration for the activity of other antibiotics are not clearly understood. Several studies have linked antibiotic killing to the dysregulation of respiration and the production of reactive oxygen species (ROS) (51)(52)(53)(54)(55)(56)(57), while others have demonstrated that inhibiting respiration can attenuate the activity of bactericidal drugs (56,(58)(59)(60)(61)(62). Given that impaired succinate oxidation reduced the activity of the respiratory chain by approximately 70% (Fig. 3), we sought to investigate how the simultaneous knockdown of sdhA1 1 sdhA2 affected the activity of anti-TB drugs against M. tuberculosis. We profiled the . We induced gene knockdown and treated cultures with antibiotics simultaneously on day 0 as previously described (47,48) and monitored killing (rather than MIC) as a measure of susceptibility as M. tuberculosis is significantly impaired for growth when transcriptionally repressing sdhA1 1 sdhA2 ( Fig. 1 and 2).
The combined knockdown of sdhA1 1 sdhA2 resulted in increased killing of M. tuberculosis by INH and PA-824 after 10 days of incubation ( Fig. 4A and B), while the susceptibility to all other drugs tested remained unchanged (Fig. S2). For example, concentrations above 1.5 mM INH resulted in a 3-log 10 reduction in CFU per milliliter to below the limit of detection for the sdhA1 1 sdhA2 double knockdown strain, whereas in the no-knockdown control the same level of killing was achieved only at the highest concentration tested (40.5 mM INH) (Fig. 4A). Note that the large "jump" in CFU per milliliter for the no-knockdown control at 13.5 mM INH was due to the emergence of INH resistance in one well. Similarly, 1.8 mM and 5.4 mM PA-824 resulted in a 0.6-and 2.5-log 10 reduction in CFU per milliliter, respectively, when M. tuberculosis was repressing sdhA1 1 sdhA2, but neither concentration killed the no-knockdown control (Fig. 4B). Combined, these data suggest that impaired succinate oxidation synergizes with INH and PA-824 when knockdown is induced simultaneously with antibiotic challenge.
We investigated this interaction further by monitoring killing by INH and PA-824 over time. Interestingly, there was no change in the rate of killing by either drug, but instead the simultaneous knockdown of sdhA1 1 sdhA2 prevented the emergence of resistance to both INH and PA-824 ( Fig. 4C and D). After 3 weeks, no INH-or PA-824-resistant mutants were isolated when both Sdh1 and Sdh2 were depleted, while INH and PA-824 resistance emerged by day 10 in the no-knockdown controls ( Fig. 4C and D). This could indicate that normal respiratory chain function is required for the development of INH and PA-824 resistance as previously described (14,63,64), or it could be the result of the reduced replication rate of M. tuberculosis when repressing sdhA1 1 sdhA2. Either way, these results demonstrate that preventing the emergence of INH and PA-824 resistance underlies the synergy observed at day 10 ( Fig. 4A and B).
Predepletion of SDH enzymes reveals additional interactions with cell wall inhibitors and bioenergetic inhibitors. Interestingly, and in contrast to our results, impaired respiration through inhibition of the cytochrome bc 1 :aa 3 terminal oxidase or ATP synthase has previously been shown to antagonize INH activity (58)(59)(60). To address this discrepancy, we considered whether depleting cells of CRISPRi target enzymes prior to antibiotic challenge affected the interactions observed. We exploited the known synthetic lethality between the two terminal oxidases (the cytochrome bc 1 :aa 3 complex and cytochrome bd) (13, 49, 65, 66) as a positive control. To do this, we predepleted M. tuberculosis of the cytochrome bd oxidase using CRISPRi for 6 days before treating cultures with the bc 1 inhibitors Q203 and TB47 (67) which are bactericidal against cells lacking a functional cytochrome bd but bacteriostatic against wild-type (WT) cells (13,49,65,66). Killing of M. tuberculosis was observed only when cells were predepleted of cytochrome bd, resulting in cell death up until day 22 (the endpoint of the experiment), but not when knockdown was induced simultaneously with antibiotic challenge (Fig. S3). Given the prolonged time frame of this experiment and the propensity for suppressor mutants when using ATc-inducible systems (48), we further confirmed the sustained knockdown of CRISPRi target enzymes when using this experimental setup by demonstrating that the predepleted sdhA1 1 sdhA2 strain was inhibited for growth over 10 days to a similar level as that when gene knockdown was induced on day 0 (Fig. S4). Taken together, these results demonstrate that there is a requirement for the predepletion of CRISPRi target enzymes when screening for chemical-genetic interactions in M. tuberculosis and show that target depletion is maintained over long time periods at sufficient levels to detect interactions (68).
Consequently, we predepleted M. tuberculosis of SDH enzymes by inducing CRISPRi for 6 days before challenging cultures with the TB drugs listed above. Under these conditions, the dual depletion of Sdh1 and Sdh2 attenuated the bactericidal activity of INH against M. tuberculosis at concentrations above the MIC, while no difference in susceptibility was seen at concentrations below the MIC (Fig. 4E). As in Fig. 4A, the large "jumps" in CFU per milliliter for the noknockdown control were due to the emergence of INH resistance in one well. In contrast, depletion of Sdh1 and Sdh2 sensitized M. tuberculosis to growth inhibition by concentrations of PA-824 that were below the MIC of the no-knockdown control (i.e., 0.06 to 0.6 mM PA-824) but, conversely, attenuated the bactericidal activity of PA-824 at concentrations above the MIC (i.e., 1.8 and 5.4 mM PA-824) (Fig. 4F). Interestingly, this attenuation was overcome at higher concentrations of PA-824 (i.e., 16.2 and 48.6 mM PA-824), resulting in the sterilization of cultures and preventing the emergence of resistance seen in the no-knockdown control (Fig. 4F). As with INH, the depletion of Sdh1 and Sdh2 attenuated the activity of all three other cell wall inhibitors tested: EMB, ETH, and SQ109 ( Fig. 5A to C). Combined, these data demonstrate that impaired succinate oxidation attenuates the bactericidal activity of cell wall inhibitors against M. tuberculosis when cells are first depleted of SDH enzymes.
In contrast, the dual depletion of Sdh1 and Sdh2 increased the susceptibility of M. tuberculosis to the bioenergetic inhibitors BDQ, TB47, and Q203, which target the ATP synthase and cytochrome bc 1 :aa 3 terminal oxidase, respectively (Fig. 5D to F). M. tuberculosis with both Sdh1 and Sdh2 depleted was hypersusceptible to growth inhibition and/or killing by BDQ at concentrations that were below the MIC of the no-knockdown control (i.e., 0.06 mM BDQ prevented the growth of the SDH-depleted cultures and 0.18 mM BDQ resulted in an ;0.5-log 10 reduction in CFU per milliliter while neither concentration affected the growth of the noknockdown control) (Fig. 5D). Likewise, concentrations of TB47 that were below the MIC of the no-knockdown control were bacteriostatic against M. tuberculosis when Sdh1 and Sdh2 were depleted, while concentrations above the MIC (i.e., 0.1 to 1 mM TB47) resulted in an ;0.5to 0.9-log 10 reduction in CFU per milliliter in SDH-depleted cells compared to the bacteriostatic effects on WT cells (Fig. 5E). Moreover, the depletion of Sdh1 and Sdh2 rendered M. tuberculosis susceptible to killing by the normally bacteriostatic Q203, resulting in a 1.2-log 10 reduction in CFU per milliliter at the highest concentration tested (Fig. 5F). Depletion of Sdh1 and Sdh2 also resulted in slight increases in susceptibility to the two other bioenergetic inhibitors tested, CFZ and THZ, at concentrations around the MIC of the respective compound, although no differences in killing were observed above the MIC (Fig. 5G and H). Taken together, these findings demonstrate that impaired succinate oxidation through the depletion of Sdh1 and Sdh2 broadly synergizes with other bioenergetic inhibitors in M. tuberculosis.
Further, impaired succinate oxidation synergized with RIF, LZD, and STREP at sub-MICs with the depletion of Sdh1 and Sdh2, resulting in the growth inhibition or killing of M. tuberculosis at concentrations that do not affect the growth of the no-knockdown control (Fig. 5I to K). However, there were no differences in the magnitude of killing above the respective MIC of each compound (Fig. 5I to K). Finally, as with the cell wall inhibitors, impaired succinate oxidation attenuated the bactericidal activity of the fluroquinolone antibiotic LEV, resulting in reduced killing against M. tuberculosis with Sdh1 and Sdh2 depleted (Fig. 5L).

DISCUSSION
Succinate oxidation is a major focal point in both the central carbon metabolism and respiratory chain of M. tuberculosis, and yet its essentiality remains poorly characterized. Here, (Continued on next page) Essentiality of Succinate Oxidation in M. tuberculosis mBio we address this by reporting on the roles and essentialities of three different enzymes (Sdh1, Sdh2, and Frd) predicted to be capable of catalyzing the oxidation of succinate to fumarate in M. tuberculosis and on the consequences of impaired succinate oxidation. We show that Sdh1 is the primary aerobic SDH utilized by M. tuberculosis but that the simultaneous transcriptional repression of both sdhA1 and sdhA2 is required to overcome the functional redundancy within the SDH/FRD enzymes and prevent growth. Our results demonstrate that succinate oxidation is the major contributor of electrons to the respiratory chain and is required for the optimal growth of M. tuberculosis on both fermentable and nonfermentable carbon sources. Finally, we show that impaired succinate oxidation both positively and negatively affects the susceptibility of M. tuberculosis to a variety of anti-TB drugs. Sdh1 has previously been proposed to be the primary aerobic SDH in M. tuberculosis, with its deletion resulting in a minor growth impairment in media containing succinate as the sole carbon and energy source (41). Likewise, the transcriptional repression of sdhA1 impaired the growth and OCR of M. tuberculosis in succinate-containing media, validating our CRISPRi approach to studying the essentiality of SDH and FRD enzymes in M. tuberculosis. Notably, the deletion of sdh1 was reported to result in an increased rate of respiration in M. tuberculosis (41), which may appear to contradict our finding that the single sdhA1 knockdown strain consumes oxygen more slowly in succinate-containing media. However, the previously described increased respiration rate in the Dsdh1 strain was observed only when oxygen levels fell below ;40% dissolved oxygen, while oxygen consumption rates were similar at higher oxygen tensions (41). In our experiments using the same media (containing glycerol as an energy source), we observed no difference in OCR between the single sdhA1 knockdown strain and the noknockdown control, thus reconciling the two findings.
In addition to validating Sdh1 as the primary aerobic SDH in M. tuberculosis, our work elucidates some of the compensatory mechanisms within succinate oxidation by demonstrating that Sdh2 can catalyze succinate oxidation in the absence of Sdh1. However, the growth and oxygen consumption impairments observed when Sdh2 is the only functional SDH enzyme present (i.e., when sdhA1 and frdA are transcriptionally repressed) suggest that Sdh2 is less efficient at catalyzing succinate oxidation than Sdh1. The molecular mechanisms underlying this are unclear, although both enzymes appear to use different reaction mechanisms to drive succinate oxidation (35)(36)(37). Moreover, these findings are in contrast to Mycobacterium smegmatis, where Sdh2 is proposed to have a higher affinity for succinate oxidation and is essential, while Sdh1 can be deleted or transcriptionally silenced with no identifiable phenotype (34,69).
We noted that the transcriptional repression of sdhA1 resulted in a concomitant reduction in frdA expression in single and multiplexed repression constructs, regardless of the sgRNA used, suggesting that a shared regulatory mechanism governs sdh1 and frd gene expression in M. tuberculosis. Despite this, no difference in frdA expression was observed in the Dsdh1 strain (41), meaning that this finding requires additional investigation. Most significantly, the linked repression of sdhA1 and frdA made it difficult to rule out a role for Frd in M. tuberculosis succinate oxidation as Sdh1 and Sdh2 could not be depleted independently of Frd. Therefore, additional experimentation is required to determine whether Frd contributes to succinate oxidation in M. tuberculosis. Previous bioinformatic analysis suggests that the M. tuberculosis Frd enzyme contains the residues required to covalently bind a flavin adenine dinucleotide (FAD) cofactor [including the conserved histidine needed to make the 8a-N(3)-histidyl linkage] (70), which is required for succinate oxidation (70)(71)(72)(73). However, the high degree of similarity between SDH and FRD enzymes (including at the FAD site)   (38), and we therefore suggest that Frd predominately functions in the reverse TCA cycle, which is essential for survival under hypoxia (3,38,78), rather than contributing to aerobic succinate oxidation. However, this still requires experimental validation. While individually dispensable, the dual depletion of both Sdh1 and Sdh2 demonstrates the essentiality of succinate oxidation in M. tuberculosis. The double sdhA1 1 sdhA2 knockdown strain was significantly impaired for growth across a range of carbon sources, including the fermentable carbon sources glycerol and glucose, as well as the nonfermentable carbon sources succinate and acetate. The improved growth on glucose and glycerol is likely due to substrate-level phosphorylation augmenting ATP levels. Importantly, mycobacterial growth and persistence in vivo are believed to be fueled by fatty acids (4,40,79), rather than glycerol and glycolytic carbon sources (80). Thus, the inability of M. tuberculosis to grow on acetate (a fatty acid surrogate [78]) when succinate oxidation is impaired suggests that inhibition of SDH activity may prevent growth in vivo. This is supported by the previously reported survival defect of the M. tuberculosis Dsdh1 strain in mouse lungs (41). However, this needs to be investigated further using the double sdhA1 1 sdhA2 knockdown strain in M. tuberculosis infection models.
The simultaneous knockdown of sdhA1 and sdhA2 reduced the OCR of M. tuberculosis by approximately 70%, demonstrating that succinate oxidation is essential for the normal function of the respiratory chain. This is interesting as the ETC of M. tuberculosis is known for its significant plasticity, with the bacterium encoding multiple other primary dehydrogenases in addition to its two SDH enzymes (33,81) that could theoretically compensate for reduced SDH activity. The simplest explanation for the observed reduction in OCR is that the majority of electrons are donated to the respiratory chain of M. tuberculosis through the oxidation of succinate to fumarate. However, NADH is generally considered the primary electron donor for aerobic growth in mycobacteria (81). Therefore, it is possible that the reduced OCR in the double sdhA1 1 sdhA2 knockdown strain is reflective of an overall reduction in central carbon metabolism and respiration to circumvent the loss of SDH activity. Succinate oxidation appears to be an invariant component of mycobacterial central carbon metabolism with few options to reroute around it (29), and we hypothesize that cells downregulate metabolism and respiration to prevent an accumulation of succinate, which may disrupt the metabolic homeostasis of replicating cells. Regardless, our data support the previous proposal that SDH is the master regulator of respiration in M. tuberculosis (41).
The essentiality of succinate oxidation for the optimal growth and ETC activity of M. tuberculosis complements previous reports on the requirement for SDH/FRD enzymes for mycobacterial persistence (3,34,38,41,44,78). While not investigated in this study due to technical limitations of using CRISPRi under hypoxia, several studies have reported the essentiality of succinate metabolism for the survival of M. tuberculosis under low oxygen tensions (3,34,38,41,44,78). In particular, inhibition of SDH activity using the suicide inhibitor 3-nitropropionate (3-NP) resulted in time-dependent killing of M. tuberculosis during adaptation to hypoxia (78). In contrast, our results suggest that the mycobacterial SDH/FRD enzymes are dispensable for survival under nutrient starvation, despite Frd being strongly upregulated under this condition (39). However, a potential limitation of using transcriptional repression without concomitant protein degradation (82,83) is that there may be residual SDH/FRD enzymes in the cell that could catalyze succinate oxidation. This is particularly relevant under nonreplicating conditions (i.e., nutrient starvation) where the lower energetic requirement of cells (i.e., maintenance energy) (10, 11) means that even a small amount of residual enzyme may be sufficient to sustain persistence. Therefore, while our data argue against a role for SDH/FRD under nutrient starvation, this needs additional validation.
In addition to providing fundamental insights into the physiology of mycobacterial energy metabolism, the findings described in this study add to the complex narrative surrounding the interaction between modulating respiration and antimicrobial efficacy. Here, we showed that inhibition of succinate oxidation through the depletion of Sdh1 and Sdh2 attenuated the bactericidal activity of several cell wall inhibitors against M. tuberculosis, including INH, ETH, EMB, and SQ109, as well as the fluroquinolone LEV. These results are consistent with previous findings that the inhibition of respiration by cotreating mycobacteria with respiratory inhibitors such as Q203 or BDQ attenuates the activity of INH, EMB, ETH, and moxifloxacin (another fluoroquinolone) by preventing a drug-induced lethal ATP burst (58)(59)(60). Similarly, disruptions to the NADH/NAD 1 ratio by inhibition of NADH dehydrogenase activity also protect against INH killing (68,(84)(85)(86)(87). Interestingly, other studies have demonstrated that inhibiting respiration through alternative mechanisms can potentiate INH killing in M. tuberculosis (14,64), while both enhancing and inhibiting respiration can prevent the emergence of INH resistance (14,63,64). Despite this complexity, our results demonstrate a requirement for succinate oxidation in the susceptibility of M. tuberculosis to INH and other cell wall biosynthesis inhibitors under in vitro conditions. Paradoxically, and in contrast to the above, when gene knockdown was induced simultaneously with antibiotic challenge, the knockdown of sdhA1 and sdhA2 prevented the emergence of INH and PA-824 resistance. Interestingly, for PA-824 this phenotype was maintained in predepleted cells at high concentrations of PA-824, despite the attenuation of killing seen at lower concentrations. While our results indicate that predepletion of CRISPRi target enzymes is required to observe known chemical-genetic interactions, we suggest that the observation that impaired succinate oxidation prevents the emergence of INH and PA-824 resistance is relevant for several reasons, all of which may support our observations: (i) INH and PA-824 are both prodrugs whose mechanism of activation is linked to mycobacterial metabolism and respiration (86,88,89), (ii) both INH and PA-824 have relatively high frequencies of resistance compared to other TB drugs (89), and (iii) the development of both INH and PA-824 resistance can be prevented by inhibiting respiration in M. tuberculosis (14,64). It is possible that the timing of the disruption to the ETC (i.e., prior to or simultaneously with INH or PA-824 treatment) affects the interactions observed (90), although this finding requires additional investigation.
Furthermore, our results highlight novel chemical-genetic interactions between ETC components that could be exploited for drug development. M. tuberculosis with both Sdh1 and Sdh2 transcriptionally depleted was hypersusceptible to growth inhibition and/or killing by other bioenergetic inhibitors, such as BDQ, TB47, and Q203. These findings add to previous reports of synergistic interactions between ETC complexes (13,61,62,66,69,(91)(92)(93) and support the proposal that targeting multiple components of mycobacterial energy generation could result in efficacious drug regimens (94).
Lastly, it is worth noting that while our results may appear to mimic the antibiotic susceptibility profile of M. tuberculosis under nonreplicating conditions (i.e., tolerant to conventional antibiotics but sensitive to respiration inhibitors), there are several key differences that demonstrate that this is not the case. Importantly, we did not observe attenuated killing for all bactericidal antibiotics tested (e.g., CFZ, THZ, STREP, and LZD), arguing against a general inhibition of killing by inducing bacteriostasis as has previously been reported (56). Further, all four of these antibiotics lose activity against nutrient-starved or hypoxic cultures (10,95), whereas our data show equal susceptibility to killing by CFZ, THZ, STREP, and LZD in the presence or absence of Sdh1 and Sdh2. Additionally, our results showed a synergistic interaction with RIF at sub-MIC levels whereas RIF has reduced activity against nutrient-starved nonreplicating cultures (10,39). These differences distinguish our findings from being a generic response to cells being in a nonreplicating state and instead argue that the interactions we observed are a specific response to impaired succinate oxidation in M. tuberculosis.
In summary, our work demonstrates that succinate oxidation is required for the optimal growth of M. tuberculosis and further defines the roles and essentiality of SDH/FRD enzymes in mycobacteria. The simultaneous transcriptional repression of sdhA1 and sdhA2 significantly reduced the activity of the respiratory chain, had bacteriostatic consequences for cell viability, and influenced the susceptibility of M. tuberculosis to a variety of antibiotics. Combined, these findings provide useful insights into the physiology of energy metabolism of M. tuberculosis and its utility as a nascent area for antimicrobial development.

MATERIALS AND METHODS
Bacterial strains and culture conditions. Escherichia coli strain MC1061 was used for the construction of all CRISPRi expression plasmids. E. coli MC1061 was grown at 37°C in liquid Luria-Bertani broth (LB) with agitation (200 rpm) or on solid Luria-Bertani agar (LBA) supplemented with 1.5% (wt/vol) agar. CRISPRi plasmids were selected for and maintained with 50 mg/mL kanamycin.
Construction of CRISPRi knockdown plasmids and strains. For transcriptional knockdowns in M. tuberculosis, a 20-to 25-bp sequence downstream of a permissible protospacer-adjacent motif (PAM) sequence was identified in the nontemplate strand of sdhA1 (Rv0248c), sdhA2 (Rv3318), and frdA (Rv1553). The 20-to 25-bp sequences from the template and nontemplate strand were ordered as oligonucleotides with 59 GGGA and AAAC overhangs, respectively. Information for individual sgRNAs (i.e., target sequence, PAM sequence, and predicted PAM strength) is provided in Table S1 in the supplemental material. Oligonucleotides used for plasmid construction are listed in Table S2. Oligonucleotides were annealed and cloned into CRISPRi plasmids (i.e., pJLR965) using BsmB1 Golden Gate cloning, sequence verified, and transformed into M. tuberculosis strain mc 2 6230 as previously described (48). Plasmids are listed in Table S3.
Multiplexed plasmids that express two or three sgRNAs targeting a combination of sdhA1, sdhA2, and frdA were constructed by cloning additional sgRNAs into a SapI-based Golden Gate cloning site (46). Plasmids that express either a total of two or three sgRNAs were constructed using specific sets of primers to ensure that overhangs generated following SapI digestion would facilitate plasmid reassembly in a specific sgRNA order. In brief, multiplex plasmids that express a total of two sgRNAs were constructed by amplifying the additional sgRNA module (transcriptional promoter, sgRNA scaffold, and transcriptional terminator) from the appropriate CRISPRi plasmid using Phusion polymerase and the indicated primer pair (sgRNA-2 = MMO120 1 MMO121) as described in Table S4. The amplified sgRNA module was purified and cloned into the CRISPRi plasmid that expresses the appropriate sgRNA partner using Golden Gate cloning with SapI, as described in Table S5. Multiplexed plasmids that express a total of three sgRNAs were constructed by amplifying the additional sgRNA modules from the appropriate CRISPRi plasmids as described above using the indicated primer pairs (sgRNA-2 = MMO120 1 MMO123 and sgRNA-3 = MMO122 1 MMO121). The amplified sgRNA modules were purified and cloned into the CRISPRi plasmid that expresses the appropriate sgRNA partner as described above. Constructed plasmids were sequence verified and transformed into M. tuberculosis strain mc 2 6230 as previously described (48). All constructed plasmids are listed in Table S3.
RNA extraction and mRNA quantification by quantitative PCR (qPCR). M. tuberculosis single and multiplexed sdhA1, sdhA2, and frdA CRISPRi strains were inoculated at an optical density at 600 nm (OD 600 ) of 0.1 in 10 mL 7H9-OADC medium containing 100 ng/mL ATc. Cultures were harvested for RNA extraction after 3 days of gene knockdown as previously described (48). RNA was extracted using TRIzol reagent (Invitrogen) and purified using the RNA Clean and Concentrator kit (Zymo) as previously described (48). Samples were DNase treated using the Turbo DNA-free kit (Invitrogen) and confirmed to be DNA free by PCR using 1 mL of extracted RNA as a template with the primer combination of MMO200 and MMO201.
cDNA was synthesized from the RNA using the SuperScript IV Vilo master mix (Invitrogen) according to the manufacturer's instructions. qPCRs were performed in 384-well plates in a ViiA7 Thermocycler using the Invitrogen PowerUp SYBR green master mix as previously described (48). Primer sequences are listed in Table S2. All qPCR primer pairs and cDNA masses tested were experimentally validated to be within the linear range of the assay. Signals were normalized to the housekeeping gene sigA and quantified by the threshold cycle (2 DDCT ) method.
Nutrient starvation experiments. The survival of M. tuberculosis knockdown strains under conditions of nutrient starvation was determined using previously published protocols (39). Briefly, M. tuberculosis knockdown strains were inoculated at a starting OD 600 of 0.005 in complete medium (7H9 plus OADC-PAN-KAN) in the presence of 100 ng/mL ATc to induce CRISPRi and deplete cells of SDH and FRD enzymes. After 8 days, cultures were harvested and washed twice in PBS plus 0.05% tyloxapol. Cells were resuspended in PBS-tyloxapol and nutrient starved by inoculation into inkwells containing PBS-tyloxapol, KAN, and 100 ng/mL ATc at an OD 600 of 0.01 (i.e., approximately 10 5 CFU/mL). PAN was not included as M. tuberculosis mc 2 6230 has previously been Essentiality of Succinate Oxidation in M. tuberculosis mBio shown not to require PAN under conditions of nutrient starvation (97,98). Cultures were incubated at 37°C without shaking for 8 weeks, and viability was measured by plating for CFU every 7 days. Oxygen consumption rate measurements. Oxygen consumption rate measurements of exponentially growing M. tuberculosis cells were performed as previously described using an Oroboros O2k FluoRespirometer (61,67). In brief, M. tuberculosis strains expressing sgRNA targeting sdhA1, sdhA2, and frdA in single or multiplexed constructs were inoculated at a starting OD 600 of 0.005 in 10 mL 7H9 medium supplemented with either OADC, glycerol, or succinate and 0 or 100 ng/mL ATc. Cultures were grown for 8 days until the sdhA1 1 sdhA2 double knockdown strain reached an OD 600 of ;0.2. All cultures were adjusted to an OD 600 of 0.2 in supplemented 7H9 medium on the day of experimentation from log-phase cultures. Two milliliters of culture was added to each measurement chamber, which contains an O 2 -sensing Clark-type electrode. Oxygen consumption rates were monitored for 15 min once oxygen consumption reached a steady state. All measurements were made at 37°C with 750 rpm in a closed chamber and a data recording interval of 2 s 21 . The combined inhibition of both terminal oxidases with Q203 (400 nM) and ND-011992 (100 mM) was used to demonstrate complete inhibition of oxygen consumption. Chemicals were added after 10 min of respiration at a steady state through the injection port of the stoppers using Hamilton syringes. Oxygen consumption rates were calculated using the Oroboros Data Lab software.
Antibiotic susceptibility testing. The MICs and minimum bactericidal concentrations (MBCs) of the M. tuberculosis sdhA1 1 sdhA2 CRISPRi strain were determined as previously described (48). In brief, M. tuberculosis was inoculated into a 96-well plate containing 7H9 plus OADC-PAN-KAN medium at an OD 600 of 0.005 in a final volume of 150 mL with 0 or 100 ng/mL ATc. Antibiotics were dispensed from a 9-point, 3-fold dilution gradient into each well, with a maximum of 2% DMSO. Cultures were incubated at 37°C for 10 days without agitation. On day 10, OD 600 values were measured in a Varioskan Lux microplate reader and MIC values for the no-knockdown controls were calculated using the Gompertz equation (99). For MBC determination, cultures were removed from rows containing 0 and 100 ng/mL ATc at day 10 and 10-fold serially diluted in 7H9 and 5 mL of each dilution was spotted onto 7H11 agar. Colonies were counted after 5 weeks of incubation at 37°C.
For preknockdown experiments, the M. tuberculosis sdhA1 1 sdhA2 CRISPRi strain was grown in 7H9-OADC-PAN-KAN medium in 10-mL volumes from a starting OD 600 of 0.005 for 6 days prior to antibiotic challenge. Cultures were treated with 0 or 100 ng/mL ATc on day 0 to induce gene silencing. After 6 days, cultures were inoculated at an OD 600 of 0.005 into a 96-well plate containing 7H9 plus OADC-PAN-KAN medium, 0 or 100 ng/mL ATc, and a 9-point, 3-fold dilution gradient of the respective antibiotic. Plates were incubated for 10 days before plating for CFU per milliliter as described above.
Time-kill experiments. M. tuberculosis cultures were inoculated into inkwells containing 10 mL 7H9 medium supplemented with OADC-PAN-KAN at a starting OD 600 of 0.005. Cultures were treated with 0 or 100 ng/mL ATc and with selected antibiotics at the specified concentrations. Bacterial viability was determined by plating onto 7H11 agar plates at specified time points and enumerating CFU after 5 weeks of incubation at 37°C. For preknockdown experiments, strains were incubated in the presence of 0 or 100 ng/mL ATc for 6 days as described above before being inoculated into inkwells.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.