Cysteine desulfurase (IscS)–mediated fine-tuning of bioenergetics and SUF expression prevents Mycobacterium tuberculosis hypervirulence

Iron-sulfur (Fe-S) biogenesis requires multiprotein assembly systems, SUF and ISC, in most prokaryotes. M. tuberculosis (Mtb) encodes a complete SUF system, the depletion of which was bactericidal. The ISC operon is truncated to a single gene iscS (cysteine desulfurase), whose function remains uncertain. Here, we show that MtbΔiscS is bioenergetically deficient and hypersensitive to oxidative stress, antibiotics, and hypoxia. MtbΔiscS resisted killing by nitric oxide (NO). RNA sequencing indicates that IscS is important for expressing regulons of DosR and Fe-S–containing transcription factors, WhiB3 and SufR. Unlike wild-type Mtb, MtbΔiscS could not enter a stable persistent state, continued replicating in mice, and showed hypervirulence. The suf operon was overexpressed in MtbΔiscS during infection in a NO-dependent manner. Suppressing suf expression in MtbΔiscS either by CRISPR interference or upon infection in inducible NO-deficient mice arrests hypervirulence. Together, Mtb redesigned the ISC system to “fine-tune” the expression of SUF machinery for establishing persistence without causing detrimental disease in the host.


INTRODUCTION
The Mycobacterium tuberculosis (Mtb) life cycle inside the human host relies upon successful adaptation to commonly encountered stresses such as reactive oxygen species (ROS), reactive nitrogen species (RNS), iron starvation, low pH, and hypoxia (1).Ironsulfur (Fe-S) clusters, which are the most ancient protein prosthetic groups, are sensitive targets for ROS and RNS, and their biogenesis is adversely affected by iron starvation (2,3).Mtb contains more than 50 Fe-S cluster proteins that carry out diverse functions within central metabolism, gene regulation, drug resistance, and persistence (table S1) (4,5).Therefore, knowledge of the biogenesis and repair of Fe-S clusters is critical to understanding the basis of persistence for this human pathogen.
Many organisms express multiple Fe-S assembly systems (e.g., NIF, ISC, and SUF), wherein the SUF system is largely restricted to organisms that are frequently exposed to Fe-S cluster-damaging conditions (6).In line with this, Mtb expresses a complete SUF system (Rv1461-Rv1466) that not only is essential under standard growth conditions but also protects Mtb from stresses [e.g., nitric oxide (NO) and iron starvation] by repairing damaged Fe-S clusters (7)(8)(9).Furthermore, depletion of the SUF system impairs Mtb's ability to maintain redox balance, central carbon metabolism (CCM), respiration, and persistence when in animals (10,11).Consistent with these findings, the SUF system was induced under stressful conditions such as ROS, reactive nitrogen intermediate (RNI), low iron, antibiotics, macrophage milieu, and sputum of tuberculosis (TB) patients (9,(12)(13)(14)(15)(16). Mtb also encodes an iscS gene (Rv3025c; cysteine desulfurase) that is not a part of the suf locus and is not surrounded by other isc genes, such as iscA and iscU, and by chaperones hscA and hscB (17).Although these findings indicate that the SUF machinery is the primary Fe-S biogenesis system in Mtb, a previous report suggested the involvement of Mtb IscS in building a 4Fe-4S cluster of the transcription factor WhiB3 in vitro (18).Moreover, a mutant of IscS (MtbΔiscS) was hypersensitive to oxidative stress and displayed impaired activity of the Fe-S cluster-dependent enzymes, aconitase and succinate dehydrogenase (17).These findings indicate that IscS is involved in maintaining Fe-S cluster homeostasis and protects Mtb from oxidative stress.However, important questions remain unanswered: What are the mechanisms by which IscS contributes to oxidative stress resistance and what is the consequence of an IscS loss in the backdrop of a fully functional SUF system on the persistence and virulence of Mtb.
In this study, we used a redox biosensor and extracellular flux (XF) analyzer to compare the redox balance and bioenergetics of wild-type (WT) Mtb and MtbΔiscS.We also examined metabolomics and transcriptomics of MtbΔiscS and analyzed the survival phenotype of the mutant under diverse stresses in vitro and in mice.Last, we found that the interplay between IscS and SUF system is crucial for adjusting the virulence of Mtb in mice.We anticipate that understanding the IscS-linked metabolic and regulatory events will contribute toward knowledge of how Mtb maintain cellular homeostasis for persistence.

IscS is required to maintain redox homeostasis in Mtb
A previous study reported that depletion of IscS results in a slightly slower growth rate of Mtb and diminished activity of Fe-S clustercontaining enzyme, aconitase, under aerobic culture conditions (17).We confirmed this observation using MtbΔiscS (fig.S1, A  and B).Aerobic metabolism inevitably generates highly deleterious superoxide (O 2 -• ) and hydrogen peroxide (H 2 O 2 ) due to the univalent reduction of molecular oxygen (O 2 ) by redox enzymes (Fig. 1A).ROS-scavenging enzymes, along with Fe-S cluster biogenesis systems, protect cells from the adverse consequences of oxidative radicals and maintains redox homeostasis in diverse organisms (2,5,(19)(20)(21).Therefore, we asked whether IscS is required to maintain redox balance under aerobic growth conditions in Mtb.To do this, we used a genetically encoded redox biosensor, Mrx1-roGFP2, to measure the redox potential of the antioxidant buffer mycothiol [reduced mycothiol (MSH)/oxidized mycothione (MSSM); E MSH ] as a proxy for cytoplasmic redox state of Mtb (22).The readout of Mrx1-roGFP2 can be analyzed by measuring the fluorescence intensity at 510 nm emission and excitation at 405 and 488 nm (fig.S2).An increase in the 405/488 ratio indicates a shift in the MSH/ MSSM ratio toward MSSM due to either enzymatic oxidation of MSH in response to ROS or depletion of the total mycothiol pool (22).Treatment of Mtb-expressing Mrx1-roGFP2 to increasing concentrations of H 2 O 2 gradually increased the biosensor ratio (fig.S3).
Exponentially growing MtbΔiscS expressing the biosensor exhibited a ~1.5-to 2.0-fold higher 405/488 ratio than WT Mtb and iscScomp (Fig. 1B), indicating that the mutant suffers from oxidative stress under aerobic conditions.We have previously reported that a twofold increase in the biosensor ratio is similar to the oxidative stress induced by 500 μM H 2 O 2 (fig.S3).These changes in the biosensor ratio corresponds to slightly oxidized E MSH of −260 mV for MtbΔiscS as compared to WT Mtb (E MSH = −280 mV) and iscScomp (−275 mV).Using CellROX Deep Red dye, which becomes fluorescent upon intracellular oxidation by ROS, we confirmed that MtbΔiscS stained with CellROX displayed 1.5-and 3.0-fold greater fluorescence than WT Mtb and iscS-comp, respectively (Fig. 1C).An equivalent increase in CellROX fluorescence was observed upon treatment of Mtb with 50 to 100 μM cumene hydroperoxide (CHP) (fig.S4).ROS are known to damage Fe-S clusters and increase the pool of labile iron (19).We confirmed this by showing higher ROS in Mtb growing under iron-excess conditions, whereas iron limitation decreases ROS levels (fig.S5).Consistent with this finding, we observed an ~1.5-fold increase in free iron in MtbΔiscS compared to WT Mtb (Fig. 1D).Collectively, these data suggest that aerobically grown MtbΔiscS displays disruption of redox and iron homeostatic mechanisms.
Because most of the cytoplasmic pool of labile iron exists in the ferrous form (23) and catalyzes the Fenton reaction to generate deleterious hydroxyl radicals, we investigated whether ROS accumulation and slow growth of MtbΔiscS are due to elevated iron in the mutant.We treated exponentially grown WT Mtb, MtbΔiscS, and iscS-comp with 250 μM of the cell-permeable iron chelator 2,2-bipyridyl (Bip) followed by measurement of ROS and viability at 24 hours after treatment.Iron deprivation by Bip induces the IdeR-dependent expression of esx-3 and suf operons (13).We detected increased expression of esx-3 (esxH) and suf genes (sufR, sufS, and sufB) in Bip-treated Mtb, confirming iron deprivation (fig.S1C).Surprisingly, CellROX staining revealed higher ROS in all three strains upon treatment with Bip (Fig. 1E).While Bip treatment increased ROS in WT Mtb (~4-fold) and iscS-comp (~6-fold), a >10fold increase was observed in MtbΔiscS as compared to WT Mtb untreated (Fig. 1E).Consistent with this, Bip treatment reduced the viability of MtbΔiscS by 70%.In contrast, Bip treatment reduced viability by only 10 to 20% in WT Mtb and iscS-comp (Fig. 1F).We determined the minimal inhibitory concentration at 90% (MIC 90 ) of Bip for WT Mtb, MtbΔiscS, and iscS-comp using Microplate Alamar Blue Assay (MABA).A twofold lower MIC 90 of Bip was observed in the case of MtbΔiscS compared to WT Mtb and iscS-comp (fig.S1D).While surprising, our findings correlate with work on obligate aerobes and photosynthetic organisms such as Anabaena sp. and Caulobacter cresentus that display heightened ROS under iron starvation-a phenomenon not detected in facultative anaerobes/aerobes, such as Escherichia coli and Bacillus subtilis (24,25).
Given that a deficiency of IscS leads to accumulation of ROS in Mtb, treatment with antioxidants could potentially rescue the slow growth phenotype of MtbΔiscS.To test this, we evaluated the effect of thiourea (Thio), an ROS scavenger (20), on ROS accumulation using CellROX and survival of MtbΔiscS.As expected, treatment with 10 mM Thio for 24 hours reduced ROS levels in MtbΔiscS (Fig. 1G).All three strains showed comparable levels of ROS upon Thio treatment (Fig. 1G).Surprisingly, despite lowering ROS levels, Thio treatment further inhibited the growth of MtbΔiscS as revealed by two-to fourfold lower MIC 90 of Thio against MtbΔiscS (3.125 mM) than WT Mtb (6.25 mM) and iscS-comp (12.5 mM) (fig.S1D).
Intracellular ROS generation takes place when flavoenzymes inadvertently relocate a fraction of their electron flux directly to molecular oxygen (26).Thus, under oxygen-deficient conditions, we expect a reduction in ROS generation.We hypothesize that if ROS accumulation is the cause of slow growth of MtbΔiscS, we could rescue the growth phenotype of mutant under conditions that limit ROS formation such as oxygen depletion.We used an in vitro Wayne model of hypoxia and reaeration (27).In this model, Mtb respiration gradually uses oxygen such that hypoxia and bacteriostasis is achieved by day 12 (Fig. 1H).We examined the growth phenotype of WT Mtb, MtbΔiscS, and iscS-comp after establishment of hypoxia in the Wayne model (27).We monitored oxygen depletion by examining decoloration of the dye methylene blue (MB).As expected, the blue color of MB dye decolorized by day 12 for WT Mtb and iscS-comp.However, and consistent with the slow growth phenotype of the mutant, the color of MB dye faded by day 16 for MtbΔiscS.We followed the viability of three strains for 10 days after establishment of hypoxia (i.e., after dye decolorization).As previously reported (28), WT Mtb retained 100% viability under hypoxic conditions for the entire duration of the experiment (Fig. 1I).In contrast, MtbΔiscS showed gradual loss of viability over time under hypoxic conditions (Fig. 1I).We conclude that increased ROS accumulation during aerobic metabolism might not be the primary cause of slow growth of MtbΔiscS.Because IscS coordinates Fe-S clusters of metabolic (e.g., aconitase) and respiratory (e.g., succinate dehydrogenase) enzymes (17), impaired bioenergetics could lead to ROS accumulation and slow growth of MtbΔiscS.Aberrant respiration could trap components of the electron transport chain in the reduced state that can directly transfer electron to molecular oxygen leading to ROS formation (29,30).
These events are possible as enough molecular oxygen might be present even under hypoxia for the reduction and ROS formation (31,32).

IscS deficiency affects CCM in Mtb
Fe-S cluster-dependent enzymes play key roles in CCM, branchedchain amino acid synthesis, and nucleotide biosynthesis (33).To examine whether the growth defect exhibited by MtbΔiscS correlates with a consequence of impaired metabolism, we used targeted, quantitative liquid chromatography-mass spectrometry (LC-MS/ MS).We quantified the steady-state amounts of glycolytic, pentose phosphate pathway (PPP), and tricarboxylic acid (TCA) intermediates.We uniformly observed decreased amounts of several glycolytic and PPP intermediates in MtbΔiscS as compared to WT  Mtb (Fig. 2A).The influence of IscS deficiency on glycolysis and PPP was surprising, as none of the glycolytic and PPP enzymes contains Fe-S clusters.However, several glycolytic and gluconeogenesis enzymes contain cysteine thiols, which are sensitive to inhibition by oxidation and S-mycothiolation in mycobacteria (34,35).Further, glucose-6-phosphate (G6P) levels are regulated by endogenous ROS (36); therefore, diminished G6P levels in the mutant could be a consequence of ROS accumulation in MtbΔiscS.Efficient metabolism of G6P is also essential for generating nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) through PPP (36), which serves as reductive power for antioxidant buffers (e.g., mycothiol and thioredoxin systems) and cell wall lipid biogenesis.Consistent with diminished PPP intermediates, levels of NADPH were slightly reduced, and nicotinamide adenine dinucleotide phosphate (NADP + ) levels were increased in MtbΔiscS, resulting in an overall increase in NADP + /NADPH ratio as compared to WT Mtb (fig.S6, A to C).
Impaired glycolysis and PPP are also reflected in a lower abundance of sugar metabolites uridine 5 0 -diphosphate glucose (UDP-Glc) along with a modest reduction in guanosine diphosphate mannose (GDP-MAN) and uridine diphosphate N-acetyl glucosamine (UDP-Glc-NAC) required for the biosynthesis of peptidoglycan (Fig. 2A) (37).UDP-Glc-NAC is an essential substrate for mycothiol glycosyltransferase, which catalyzes the first step in the biosynthesis of mycothiol in Mtb (38).As expected, the total mycothiol (MSH + MSSM) content was substantially diminished in MtbΔiscS as compared to WT Mtb (Fig. 2B; P ≤ 0.001), correlating well with the ratiometric increase in the mycothiol-specific biosensor (Mrx1-roGFP2) observed in the mutant.Mycothiol is also essential for detoxifying a highly reactive dicarbonyl by-product, methylglyoxal (MG) (38).We found that MG levels accumulated in MtbΔiscS as compared to WT Mtb, and this could in part be linked to the lack of mycothiol-linked remediation (Fig. 2B).MtbΔiscS has been previously reported to display diminished activities of the Fe-S cluster-containing TCA cycle enzymes aconitase and succinate dehydrogenase (17).Inactivation of aconitase is expected to raise citrate levels and reduce downstream TCA cycle metabolites such as α-ketoglutarate (α-KG) (11).Citrate showed only a marginal increase.Likewise, α-KG was unaffected in MtbΔiscS (Fig. 2A).Lowered succinate dehydrogenase activity indicates that the conversion of succinate to fumarate would be deficient in MtbΔiscS.Surprisingly, rather than accumulating succinate, MtbΔiscS displayed succinate depletion along with a concomitant increase in fumarate as compared to WT Mtb (Fig. 2A).These results suggest impairment of activity or expression of fumarate reductase in MtbΔiscS.Fumarate reductase is a 4Fe-4S cluster enzyme encoded by frdABCD and catalyzes the conversion of succinate to fumarate (39).Along with fumarate accumulation, a notable depletion of malate was observed in MtbΔiscS (Fig. 2A), indicating weakened fumarate hydratase activity.In line with the defective TCA cycle, the generation of nicotinamide adenine dinucleotide hydrogen (NADH) was marginally reduced, leading to higher nicotinamide adenine dinucleotide (NAD + ) and NAD + /NADH ratio in MtbΔiscS as compared to WT Mtb and IscS-comp (fig.S6, D to F).A massive accumulation of the transsulfuration pathway metabolite cystathionine (Cth; 20-fold) and minor albeit nonsignificant decrease in methionine (Met) and S-adenosyl methionine (SAM) suggests an impairment of sulfur metabolism in MtbΔiscS (Fig. 2B).The enzyme adenosine 5 0 -phosphosulfate reductase (CysH) that catalyzes reductive assimilation of inorganic sulfate into Met, SAM, and cysteine (Cys) in Mtb contains a 4Fe-4S cluster (5); the activity of this enzyme may also be insufficient in MtbΔiscS.Under such conditions (abnormal reduction of inorganic sulfate), Mtb relies upon reverse transsulfuration (RTS) enzymes Cth-β-synthase (CBS) that converts homocysteine to Cth and Cth-γ-lyase (MetB) that generates Cys for mycothiol biosynthesis from Cth (40,41).MetB is also a bifunctional enzyme that can catalyze a γ-replacement reaction using Cys and O-succinyl/acetyl-homoserine to generate Cth via forward transsulfuration (FTS) reaction for regenerating methyl cycle intermediates [Met, SAM, and S-adenosyl homocysteine (SAH)].While MtbΔiscS accumulates Cth and maintains methyl cycle intermediates, it shows depletion of mycothiol.These findings suggest that in the absence of IscS, Mtb switches from synthesizing Cys via the RTS pathway to using Cys for generating methyl cycle intermediates via the formation of Cth intermediate through the FTS pathway.We recently showed that metabolic switching from RTS to FTS pathways ensured Mtb survival by avoiding excess oxidative stress (40).Future experiments are required to clarify the link between IscS, RTS, and FTS pathways.
We observed a notable accumulation of various amino acids (Phe, Tyr, Trp, Glu, Arg, and His) in the mutant (Fig. 2C), suggesting attempts of Mtb to balance the loss of TCA cycle intermediates by anaplerosis in the absence of IscS.For example, elevated Glu could support the replenishment of α-KG by glutamate dehydrogenase in the background of defective glycolysis and aconitase activity in MtbΔiscS.Despite the reduction in levels of 2-phosphoenol pyruvate (PEP) and PPP intermediates, which are the initial substrates for the biosynthesis of aromatic amino acids (shikimate pathway), MtbΔiscS showed an accumulation of Phe, Trp, and Tyr.This could come from an impairment in the catabolic pathways for these amino acids, for instance, utilization into TCA or cofactor generation (e.g., molybdopterin biosynthesis) (42), along with increased utilization of PPP intermediates for amino acid synthesis.Complementation of IscS expression readjusted abundance of ~70% of the metabolites to levels comparable to those detected in WT Mtb (fig.S7, A and B).Last, the highlighted changes in the abundance of glycolysis, PPP, TCA, and amino acid pathways metabolites are also observed in the WhiB3 mutant of Mtb (43).IscS directly interacts with WhiB3 to coordinate 4Fe-4S cluster in vitro, indicating that IscS can regulate the Fe-S cluster-dependent regulatory activity of WhiB3 (18).Overall, the data indicate an important interplay between IscS, CCM, sulfur metabolism, and redox homeostasis in Mtb.

IscS is required to maintain bioenergetic homeostasis of Mtb
The altered pool of CCM in MtbΔiscS and the dependency of mycobacterial respiration on Fe-S cluster homeostasis (10,11) motivated us to measure the bioenergetics of the mutant.We used the XF analyzer and noninvasively tracked the extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) of MtbΔiscS.ECAR is an accurate indication of proton (H + ) translocation due to glycolytic and TCA cycle activity, whereas OCR is a measure of respiration linked to oxidative phosphorylation.
We starved Mtb cultures overnight, followed by seeding in the XF microchamber, exposed them to glucose at a specific time, and then treated cells twice with the uncoupler carbonyl cyanide m-chlorophenyl hydrazine (CCCP).The injection of CCCP stimulates maximal respiration by Mtb, which provides an estimate of the bioenergetic reserves of the cells (44).The data are normalized to an equal number of viable cells (2 × 10 6 per well) in each case.We noticed that the basal OCR (before the addition of glucose) of MtbΔiscS is twofold (P = 0.0001) reduced as compared to WT Mtb (Fig. 3, A and C) with a change of 1.6-fold (P = 0.005) in ECAR (Fig. 3B).In the presence of glucose, MtbΔiscS displayed a substantial reduction in OCR (~2.5-fold,P = 0.008) and ECAR (~2.6-fold,P = 0.00007) relative to WT Mtb (Fig. 3, A to C).As expected, two sequential additions of CCCP stimulated OCR and ECAR in WT Mtb (Fig. 3, A and B).In contrast, a negligible increase in OCR and ECAR was observed in the case of CCCP-treated MtbΔiscS (Fig. 3, A and B).As a result, MtbΔiscS showed a 2.4fold reduction in the spare respiratory capacity as compared to WT Mtb (Fig. 3C).The IscS complemented strain largely restored the bioenergetic profile to that of WT Mtb (Fig. 3, A to C).We conclude that IscS is required for maintaining the optimal bioenergetics of Mtb.

IscS modulates the expression of DosR, WhiB3, and SufR regulons in Mtb
We next asked whether any transcriptional response was associated with the redox, metabolic, and bioenergetic changes we described in MtbΔiscS.We analyzed the global transcriptome of MtbΔiscS compared to WT Mtb.Total bacterial RNA was isolated from logarithmically growing cultures at an absorbance of ~0.4,sequenced, and analyzed using the EdgeR platform.Multidimensional scaling analysis showed that the samples clustered by their biological replicates (fig.S8A).Compared to the transcriptome of WT Mtb, the expression of 547 genes was altered in MtbΔiscS [false discovery rate (FDR) ≤ 0.05, absolute fold change ≥ 1.5], of which 181 genes were up-regulated, whereas 366 genes showed down-regulation (table S2).
As expected from a defective adaptation of MtbΔiscS to O 2 limitation, the hypoxia-responsive Dos-dormancy regulon was uniformly suppressed in the mutant (Fig. 4D).Furthermore, we compared the expression of the Dos-dormancy regulon under hypoxic conditions by performing quantitative reverse transcription polymerase chain reaction (qRT-PCR) of a set of genes (hspX, fdxA, devR, devS, and narK2) representing the Dos regulon.As expected, hypoxia induces the expression of Dos regulon genes in WT Mtb.However, the expression of Dos genes is either less induced (hspX and fdxA) or down-regulated (devR, devS, and nark2) in MtbΔiscS compared to WT Mtb or iscS-comp (fig.S9A).We also found that the expression of eight genes encoding Fe-S cluster proteins was repressed, while two genes (rubA and whiB6) were induced in MtbΔiscS as compared to WT Mtb (Fig. 4B).We reasoned that the observed transcriptional changes are not a direct consequence of iscS loss but rather due to defective occupancy of Fe-S clusters on transcription factors.Consistent with this idea, the expression of several Fe-S cluster-containing transcription factors, such as whiB1, whiB6, sirA, and sufR, was deregulated in MtbΔiscS (Fig. 4B).Moreover, the transcriptome of MtbΔiscS overlapped with that of previously reported regulons of Fe-S cluster-containing transcription factors: sufR and whiB3 (Fig. 4E) (10,45).Because whiB1 is essential and its transcriptome is not reported, we depleted whiB1 in Mtb (whiB1-KD), using   S2).
tetracycline-based CRISPR interference (CRISPRi) (fig.S10, A and  B), and performed RNA sequencing (RNA-seq).An ~50% overlap between the transcriptomes of whiB1 and iscS was observed (Fig. 4E).We also compared the overlap between DEGs in MtbΔiscS with DEGs of Mtb grown under oxygen limitation, H 2 O 2 , and Fe − starvation.We considered that DEGs overlapped irrespective of whether the genes are similarly regulated (induced or repressed) between MtbΔiscS and WT Mtb exposed to different stress conditions.On this basis, the RNA-seq data of MtbΔiscS overlapped with Mtb grown under oxygen limitation and H 2 O 2 .Similarly, the transcriptome of MtbΔiscS did not coincide with iron starvation (13), a finding consistent with an increased fraction of free iron in the iscS mutant (Fig. 4E).The iscS-comp strain notably complements the IscS deregulated transcripts (table S2).

IscS protects Mtb from lethality caused by anti-TB drugs
Redox homeostasis, CCM, and respiration modulate Mtb's response to anti-TB drugs (44).Furthermore, the ISC system in E. coli promotes killing by bactericidal antibiotics by providing iron for Fenton chemistry (46).In addition, ISC-mediated maturation of Fe-S clusters on respiratory complexes I and II promotes aminoglycoside killing by enabling respiration-dependent drug uptake (47).In the absence of ISC, cells switch to a less efficient SUF system for the maturation of Fe-S clusters on respiratory complexes, leading to defective uptake of aminoglycoside (47).However, a knowledge gap exists in our understanding of the role of Fe-S cluster biogenesis pathways in influencing lethality caused by anti-TB drugs.To begin understanding this, we measured the inhibition of MtbΔiscS growth by the clinically relevant anti-TB drugs rifampicin (Rif ), isoniazid (Inh), moxifloxacin (Mox), and bedaquiline (Bdq).Using MABA, we determined MIC for Rif, Inh, Mox, and Bdq with WT Mtb, MtbΔiscS, and iscS-comp (Fig. 5A).MtbΔiscS exhibited two to fourfold lower MIC (therefore higher sensitivity) for Mox and Bdq (Fig. 5A).While MtbΔiscS showed no change in MIC for Rif, mutant growth was lowered by 84% at a subinhibitory concentration (8.8 nM) of Rif as compared to 37 to 67% inhibition of WT Mtb and iscS-comp (Fig. 5B).
We also measured the kill kinetics for Mox and Rif by colonyforming unit (CFU) analysis at various times after treatment with 1× MIC for WT Mtb.Compared to WT Mtb, MtbΔiscS displayed ~20-and 600-fold greater killing for Mox and Rif, respectively, at 10 days after exposure (Fig. 5, C and D).In contrast, MtbΔiscS exhibited higher MIC and resistance toward Inh (Fig. 5, A and E).One likely possibility is diminished levels of mycothiol in MtbΔiscS, which is known to cause Inh resistance in diverse mycobacterial species (48).Together, our data are in contrast with the current paradigm in bacteria, which indicates that Fe-S cluster biogenesis pathways promote killing by antibiotics (46).In Mtb, IscS reduces susceptibility toward anti-TB drugs.

IscS protects Mtb from oxidative stress but not nitrosative stress
Fe-S cluster biogenesis systems also contribute to protection from oxidative and nitrosative stress in bacteria (49).In general, the Suf system maintains Fe-S cluster homeostasis during high demand (e.g., NO and iron starvation), whereas IscS coordinates housekeeping requirements for Fe-S clusters in diverse bacteria (3,33,50,51).The Mtb Suf system is required for Fe-S cluster biogenesis/repair both under low (standard growth conditions) and high demand (nitrosative stress) for Fe-S clusters (10,11).Although IscS seems important for optimal growth, its contribution in assisting Mtb to counteract stresses remains uncertain.Thus, we next examined the requirement for IscS in providing tolerance to stresses such as organic hydroperoxide (CHP; 80 μM), superoxide [menadione (Mnd); 60 μM], and NO donor [diethylenetriamine (DETA)/NO adduct (1 mM)].WT Mtb and MtbΔiscS were treated with CHP and Mnd for 4 hours (Fig. 6A) and 24 hours (Fig. 6B), and survival was measured by CFU analysis.WT Mtb was unaffected at 4 hours, while 24-hour exposure led to a 20% decrease in survival (Fig. 6B).In contrast, 4-hour CHP and Mnd treatment was sufficient to reduce survival by 80 and 25% in MtbΔiscS relative to WT Mtb, respectively (Fig. 6A).At 24 hours after exposure, CHP and Mnd treatment resulted in ~99% lower survival as compared to WT Mtb (Fig. 6B).Using the Mrx1-roGFP2 biosensor, we confirmed that sensitivity to CHP or Mnd is associated with excessive oxidative stress in MtbΔiscS.We observed nearly complete oxidation of the biosensor in CHP-or Mnd-treated MtbΔiscS relative to WT Mtb within 4 hours of exposure (Fig. 6C).The iscS-comp strain showed a survival phenotype comparable to that of WT Mtb (Fig. 6, A to C).In contrast to oxidative stress, treatment of 1 mM DETA NO uniformly induces bacteriostasis in all three strains without any adverse effect on MtbΔiscS (Fig. 6D).Because NO's ability to kill Mtb depends upon concentration and exposure time (9), we repeatedly challenged Mtb strains to 1 mM DETA NO every 24 hours for 4 days, and viability was assessed.We observed that WT Mtb and iscS-comp maintained survival until day 2 but displayed a gradual loss in viability at day 3 (10-fold) and day 4 (100-fold) after treatment (Fig. 6E).In contrast, MtbΔiscS retained viability throughout the experiment, indicating that the mutant resists nitrosative stress (Fig. 6E).A similar trend was observed upon repeated exposure to 2 mM DETA NO.In the case of WT Mtb and iscS-comp, 2 mM DETA NO resulted in a complete loss of viability at day 4 after treatment (Fig. 6F), corresponding to more than 5-log reduction in bacterial numbers (no colonies detected).In contrast, we observed only 100-fold killing of MtbΔiscS at day 4 after treatment (Fig. 6F).
To understand the relevance of in vitro findings to infection settings, we assessed survival of the mutant in macrophages.We infected resting and immune-activated RAW264.7 murine macrophages with Mtb strains at a multiplicity of infection (MOI) of 2 and monitored survival over time.The WT Mtb showed unrestricted growth in resting RAW264.7 over time.MtbΔiscS displayed a marginal (twofold) reduction in growth as compared to WT Mtb and iscScomp at day 3 after infection (Fig. 6G).As expected, activation of macrophages prevents the multiplication of Mtb strains.MtbΔiscS showed survival comparable to WT Mtb and iscS-comp in activated macrophages (Fig. 6H).Because NO generated via inducible nitric oxide synthase (iNOS) is one of the main contributors of redox stress in Mtb inside immune-activated murine macrophages (1), these data agree with comparable survival of the mutant and WT Mtb in response to a bacteriostatic concentration of NO (singledose 1 mM DETA NO) in vitro.We further verified the association between IscS and NO by treating RAW264.7 with the iNOS inhibitor 1400W (52) and examined the growth phenotype of MtbΔiscS.All three strains resumed proliferation in the presence of 1400W; however, the growth was marginally slow for MtbΔiscS compared to WT Mtb and iscS-comp (Fig. 6I).

Disruption of IscS resulted in a hypervirulent mutant of Mtb
Depletion of Suf components (SufR and SufT) attenuated the survival and persistence of Mtb in vivo (10,11).However, the contribution of IscS to the pathogenesis of Mtb remains unknown.We examined the requirement for IscS in coordinating Mtb's virulence in a mice model of experimental TB.BALB/c mice were infected with WT Mtb, MtbΔiscS, and iscS-comp via the aerosol route, and survival was determined by CFU counts at days 28 and 56 after infection.WT Mtb having a functional IscS showed increased bacterial burden until 28 days after infection, after which the lung bacillary load stabilized (Fig. 7A).In contrast with the poor survival of MtbΔiscS in vitro, the mutant showed higher survival and persistence in mice (Fig. 7, A and B).Lung bacterial burden in mice with MtbΔiscS was significantly higher (~50-fold) than that in those infected with the WT Mtb at 28 days after infection (Fig. 7A).The bacillary load in the lungs reached a maximum of ~100-fold (P ≤ 0.0001) higher for MtbΔiscS as compared to WT Mtb.A similar trend of elevated bacterial burden in the spleen of mice infected with MtbΔiscS as compared to WT Mtb was also detected (Fig. 7B).
Gross organ examination revealed that the lung surface of mice infected with MtbΔiscS showed discrete, well-circumscribed lesions (Fig. 7C) compared with poorly circumscribed and diffuse lesions in lungs of WT Mtb-infected mice.Histopathologic examination of the lungs of mice infected with WT Mtb revealed moderate granulomatous pneumonia with relatively smaller granulomas as compared to mild to severe granulomatous pneumonia characterized by multiple foci of granulomas with varying degrees of severity in the case of MtbΔiscS (Fig. 7D and fig.S11A).The iscS-comp strain showed partial attenuation of the hypervirulence phenotype in mice (Fig. 7A).Overall, the disruption of iscS in Mtb led to hypervirulence and abnormal granuloma formation in mice.
Hypervirulence of MtbΔiscS is reduced by blocking NOdependent suf induction ROI, RNI, iron starvation, and an intraphagosomal environment increase expression of the suf operon but not IscS in Mtb (9,13,16).We confirmed this observation by qRT-PCR and found that 5 mM H 2 O 2 marginally induced (1.5-to 2.0-fold) sufS and sufB in WT Mtb and MtbΔiscS (fig.S12A).Notably, treatment with a bacteriostatic concentration of NO (0.5 mM DETA NO) induces an ~40-to 60-fold higher expression of suf genes (sufS and sufB) in WT Mtb and MtbΔiscS (fig.S12B).In contrast, NO reduces the expression of iscS (fig.S12B).However, at a bactericidal concentration of NO (2 mM DETA NO), the expression of suf genes was consistently higher in MtbΔiscS than in WT Mtb and iscS-comp (fig.S13A).We also performed qRT-PCR of suf genes on bacterial RNA isolated from immune-activated RAW264.7 infected with WT Mtb and MtbΔiscS in the presence or absence of iNOS inhibitor 1400W.Immune activation of macrophages induces the expression of suf genes by 100to 300-fold in WT Mtb and 300-to 500-fold in MtbΔiscS (Fig. 8A).Inhibition of iNOS by 1400W uniformly represses the expression of suf genes to comparable levels in both strains (Fig. 8B).
On the basis of the above findings, we hypothesized that NO could be an environmental cue that induces suf genes, and the absence of IscS results in a compensatory, albeit unregulated increase in the suf operon, leading to a higher bacillary load with MtbΔiscS in mice.To test the above possibilities, we measured the expression of sufS as a proxy for the suf operon in bacterial RNA extracted from the lungs of mice infected with Mtb strains at 56 days after infection.We observed that the expression of sufS was 6.5-fold induced in WT Mtb derived from animal lungs compared to in vitro grown Mtb (Fig. 8C).Consistent with our hypothesis, the transcript of sufS was ~600-fold greater in MtbΔiscS than in vitro grown mutant bacilli (Fig. 8C).The iscS-comp strain showed a partial reduction in sufS transcript as compared to MtbΔiscS (Fig. 8C), which is consistent with partial complementation noted above.Induction of NO production was evident in mice after 14 to 24 days of infection with Mtb (53).Therefore, we quantified sufS expression in the lungs of infected mice at 21 days after infection.The expression of sufS was induced twofold in WT Mtb, whereas a 75-fold increase was detected in MtbΔiscS compared to in vitro grown conditions (Fig. 8D).To confirm that NO leads to the higher expression of the Suf system in mice, we infected a mouse strain lacking iNOS (iNOS −/− ) with WT Mtb and MtbΔiscS, extracted bacterial RNA from lungs at 4 weeks after infection, and examined the expression of sufS.We found that the expression of sufS was 40-and 90-fold down-regulated in WT Mtb and MtbΔiscS derived from the lungs of iNOS −/− mice as compared to in vitro grown bacteria, respectively (Fig. 8E).Consistent with these findings, we found that both the strains proliferated in mouse lungs without iNOS with no sign of hypervirulence exhibited by MtbΔiscS (Fig. 8F).The gross pathology of the lungs in both cases was comparable (fig.S14).While there was no difference in the bacillary burden for WT Mtb and MtbΔiscS in the lungs of iNOS −/− mice, we observed that the colonies of MtbΔiscS recovered from the lung homogenates of iNOS −/− mice were notably smaller than those of WT Mtb.The mutant colonies appeared only after 8 weeks of incubation as opposed to 2 to 3 weeks for WT Mtb.The reasons behind the delayed resumption of growth and small colony size of MtbΔiscS isolated from iNOS −/− mice need future experimentation.Our data suggest that Mtb preferentially mobilizes the Suf system under Fe-S cluster-damaging conditions such as NO in vivo and the hypervirulence of MtbΔiscS is associated with NO-dependent overexpression of the suf operon during infection in mice.

MtbΔiscS during infection
We next sought to determine the impact of the suf operon on MtbΔiscS.Because the suf operon (sufBCDSUT) is essential (7), we interrupted its expression by anhydrotetracycline (ATc)-dependent CRISPRi-mediated depletion of sufS in WT Mtb and MtbΔiscS (fig.S15A) (54).Coexpression of dCas9 (fig.S15B) and sufS-specific guide RNA (sgRNA) in response to ATc treatment uniformly  reduced (~10-fold) the expression of sufS and downstream genes sufU and sufT in WT Mtb (sufS-KD) and MtbΔiscS (ΔiscS-sufS KD) (fig.S15, C and F).A ~10-fold reduction in sufS did not affect the survival of sufS-KD under aerobic growth conditions (fig.S15D) or in response to 5 mM H 2 O 2 (fig.S15E).In contrast, ΔiscS-sufS KD grew slower than MtbΔiscS under aerobic conditions, and treatment with 5 mM H 2 O 2 for 24 hours reduced the survival of ΔiscS-sufS KD by ~150-and 20,000-fold as compared to MtbΔiscS and WT Mtb, respectively (fig.S15, D and E).These data suggest that in the absence of IscS, Mtb uses the Suf system to sustain aerobic growth and survive H 2 O 2 challenge.Expression of Suf system depends on a 4Fe-4S cluster containing the transcription factor SufR (10).Under Fe-S cluster-deficient conditions (e.g., NO and low iron), the cluster-less form of SufR (apo-SufR) derepresses the expression of the Suf system (10).We substantiated the compensatory role of the Suf system in the absence of IscS by examining the relative expression of the suf operon in WT Mtb, MtbΔiscS, sufS-KD, and ΔiscS-sufS KD by RNA-seq.Defects in both iscS and sufS (ΔiscS-sufS KD) induced the expression of the suf operon, which is higher than MtbΔiscS and marginally supersedes sufS-KD as compared to WT Mtb (fig.S15F).These findings demonstrate that ΔiscS-sufS KD is severely impaired in building Fe-S clusters, which results in defective aerobic growth and hypersensitivity to H 2 O 2 .
We infected BALB/c mice with the sufS-KD and ΔiscS-sufS KD strains by aerosol.As previously reported (11), we initiated SufSUT depletion in both strains at day 7 after infection to assess acutephase phenotype and for the chronic phase at day 21 after infection by feeding doxycycline (Fig. 9A), a tetracycline derivative commonly used for gene silencing in vivo (11).The sufS-KD strain that retained IscS, but depleted SufSUT, showed the lowest lung bacillary load during both the acute and chronic phases of infection (Fig. 9, B  and C).The ΔiscS-sufS KD mutant that lacked IscS and had depleted SufSUT exhibited a lung bacillary load intermediate between that of the severely attenuated sufS-KD and hypervirulent MtbΔiscS (Fig. 9,  B and C).The bacterial burden of ΔiscS-sufS KD was similar with that of WT Mtb (Fig. 9, B and C).The gross and histopathological changes observed in the lungs of chronically infected mice were proportionate with the bacillary load observed (Fig. 9, D and E).The extent of pulmonary tissue destruction was highest in animals infected with MtbΔiscS, intermediate in WT Mtb and ΔiscS-sufSKD (score, ~15), and lowest in sufS-KD (score, 3.5) (fig.S16A).Thus, overinduction of the suf operon is likely responsible for the hypervirulence of MtbΔiscS in mice.

DISCUSSION
Mtb expresses a nearly complete Suf system, thereby satisfying the demand for de novo Fe-S cluster assembly under normal growth conditions and repair in response to stress (10,11).Why Mtb retained a single gene (IscS) of the ISC system was initially unclear.The genetic, biochemical, and infection experiments described above show that IscS is required for maintaining redox balance, bioenergetics, antibiotic susceptibility, resistance to ROS, and survival inside macrophages.In the context of infection, deletion of IscS led to hypervirulence in mice, a phenotype linked to uncontrolled induction of the Suf system in response to NO.Our data indicate that Mtb uses the IscS and Suf systems to attain an intermediate degree of virulence that is critical for persistence.
Earlier transposon mutagenesis studies indicated that IscS is essential for Mtb (55).However, this is not the case, although IscS promotes optimal overall growth of Mtb.Previous studies were done at a subsaturating level, requiring statistical tools for downstream data processing to estimate essentiality.This affected the ability to assess the essentiality of many genes, including IscS, reliably.Consistent with this, a follow-up study used a completely saturated transposon library more authoritative cataloging of essential genes in Mtb and confirmed that Mtb IscS is not essential (56).Generally, Fe-S cluster biogenesis occurs during de novo assembly on the nascent apoproteins and repair of damaged Fe-S clusters (57).Similar to the IscS mutant of E. coli (19), assessment of MtbΔiscS indicated that most of the phenotypes exhibited by the mutant are likely due to inadequate de novo assembly rather than an inability to repair oxidized clusters.For example, when cells grow aerobically, the Fe-S clusters appear to be more sensitive to oxidative damage than in anaerobic conditions.Therefore, if IscS is functioning through the repair of oxidatively damaged Fe-S clusters, then the general growth defect of MtbΔiscS should show a greater dependence on oxygen.However, MtbΔiscS exhibited impaired survival under both aerobic and hypoxic conditions.Furthermore, MtbΔiscS accumulates ROS, which is known to convert [4Fe-4S] 2+ clusters to [3Fe-4S] 1+ clusters (19,33).The 3Fe-4S clusters can be repaired back to 4Fe-4S clusters upon scavenging of ROS without a requirement for either cysteine desulfurase or a sulfur source (19).In contrast, the reduction in ROS levels of MtbΔiscS by Thio did not restore the aerobic growth defect of the mutant.In the absence of IscS, Mtb likely accumulates incomplete Fe-S clusters that lack sulfur atoms, which differ from oxidatively damaged Fe-S clusters.Consistent with this idea, the pool of unincorporated intracellular iron was elevated in MtbΔiscS.Free iron could cause ROS accumulation in MtbΔiscS via the Fenton reaction (26).However, in contrast to our expectations, ROS accumulation and the slow growth defect were accentuated upon treatment of MtbΔiscS with an iron chelator.These results suggest that ROS increase in MtbΔiscS is not a consequence of free iron.We have recently shown that increased NADH/NAD + ratio and NADHreductive stress are critical for iron-mediated ROS surge (58).MtbΔiscS showed neither an increased NADH pool nor higher NADH/NAD + poise, which suggests that iron accumulation is unlikely to drive ROS generation in the mutant.Because iron is a cofactor for several antioxidant enzymes (e.g., superoxide dismutase and catalase) (59,60), ROS accumulation could be due to disruption of their activities upon iron limitation.A possible explanation for the elevated oxidative stress upon iron chelation in MtbΔiscS is an impairment of the metabolic and respiratory functions, mainly due to reduced synthesis of Fe-S clusters for enzymes of the TCA cycle and respiratory chain.Similarly, dysfunction of mechanisms that prevented the assembly of mitochondrial Fe-S cluster led to ROS accumulation and DNA damage (61).
The slow-growth phenotype of MtbΔiscS was expected because several metabolic enzymes and respiratory complexes contain Fe-S clusters.Consistent with this, MtbΔiscS had diminished glycolytic, PPP, and TCA cycle intermediate pool.These results aligned well with impaired oxygen consumption, reduced glycolysis, and reductant levels (NADH/NADPH) in MtbΔiscS.Surprisingly, despite a previous study showing ~50% reduction in aconitase and succinate dehydrogenase activities in MtbΔiscS (17), the respective substrates and products of these enzymes remain unaltered in the mutant.One possibility is that the residual activity of aconitase and succinate dehydrogenase is due to the Suf system.Alternatively, in the absence of IscS, Mtb could attempt to restore metabolism by operating the TCA cycle in the reverse direction.This appears to be partially successful and could have resulted in the restoration of citrate, α-KG, and oxaloacetate in MtbΔiscS.However, accumulation of fumarate and reduction in succinate and malate indicate the impaired activity of a reverse TCA cycle enzyme fumarate reductase, which contains an Fe-S cluster (39), and a canonical TCA enzyme fumarate hydratase in MtbΔiscS, respectively.Our RNA-seq data confirm that genes expressing fumarate reductase are down-regulated in Mtb, which could be another reason why fumarate reductase activity is likely more affected than succinate dehydrogenase.
The gene encoding fumarate hydratase was unaffected in the mutant.The fumarate hydratase catalyzes the stereospecific hydration of fumarate to malate and is essential for Mtb's survival in vitro and in vivo (62).At least two possible mechanisms could link fumarate hydratase with redox balance and Fe-S cluster biogenesis in Mtb.First, fumarate hydratase maintains redox homeostasis by preventing fumarate accumulation and subsequent covalent modification of mycothiol and thiol-containing proteins via succination in Mtb (62).Second, in mammalian cells, proteins involved in Fe-S cluster biogenesis and respiration are sensitive targets of succination, directly linking fumarate hydratase functionality with Fe-S cluster biogenesis and bioenergetics (63).Additional experiments are needed to determine whether IscS activity is regulated by fumarate hydratase and succination in Mtb.
While any of the above Fe-S cluster-dependent mechanisms could have contributed to the slow growth rate of MtbΔiscS, other cellular functions requiring IscS as a sulfur donor, such as the thiolation of tRNA, might contribute to these metabolic phenotypes.In this context, IscS exists in an operon with trmU, which encodes 2-thiouridylase.This enzyme is known to catalyze the formation of 2-thiourdine in E. coli tRNA (64).While IscS physically associates with TrmU, deletion of trmU neither reduced aconitase activity nor affected the survival of Mtb under H 2 O 2 stress (17).However, our transcriptomics data suggest that trmU expression remained downregulated in MtbΔiscS and showed only partial restoration in iscScomp, which could explain incomplete complementation of the mutant phenotypes in some of our assays.A more careful examination is needed to decipher the role of IscS and TrmU in the sulfur metabolism of Mtb.
Our metabolomics and transcriptomics data provide insight into why Mtb harbors a truncated ISC system in the backdrop of a complete SUF pathway.Both IscS and Suf systems are expressed and contribute to enzymatic activities of the Fe-S cluster proteins aconitase and succinate dehydrogenase under normal growth conditions (11,17).However, the absence of IscS mainly affected metabolites unrelated to Fe-S cluster enzymes (e.g., glycolysis and PPP), except for the TCA cycle enzyme fumarate reductase.In contrast, depletion of SufT resulted in an altered pool of many metabolites directly or indirectly dependent on Fe-S cluster enzymes (11).These observations explain why the Suf system is essential for the growth of Mtb, whereas disruption of IscS only slows growth under aerobic conditions.Because the expression of the Suf system was maintained in MtbΔiscS, the residual metabolic, redox, and bioenergetic activities in the mutant could be due to the action of the Suf system.
The connection between the Suf system and IscS becomes clearer under conditions that highly induce the suf operon.NO treatment induces an elevated and long-lasting expression of the suf operon in Mtb (9), which remained comparable in MtbΔiscS (Fig. 8B).In contrast, hypoxia represses the suf operon (65).In line with this, nitrosative stress did not affect MtbΔiscS, whereas hypoxic conditions reduced survival of MtbΔiscS.In general, mild oxidizing conditions, such as low concentrations of H 2 O 2 or O 2 , oxidize a [4Fe-4S] 2+ cluster to a [3Fe-4S] + cluster that can be repaired back to the original [4Fe-4S] 2+ under reducing conditions in the presence of ferrous ion (6).For these environmental situations, basal expression of both IscS and Suf is sufficient.However, for proteins whose clusters are degraded beyond the [3Fe-4S] + state or to apo-form or dinitrosyl-iron dithiol complex (DNIC), wherein the sulfide ligands of the 4Fe-4S clusters were displaced by NO to form [Fe-(NO) 2 ], sulfur atoms must be supplied by inducing de novo-style Fe-S cluster biogenesis pathways for rebuilding the Fe-S clusters (6,10).This seems to be the case with Mtb, wherein the expression of the suf operon is repressed by a 4Fe-4S cluster containing SufR (holo-SufR) under standard aerobic growth conditions.At the same time, cluster-less SufR (apo-SufR) derepresses Suf expression (10).The 4Fe-4S cluster of SufR is resistant to mild oxidizing conditions (O 2 ) (10).NO or high molar concentrations of H 2 O 2 rapidly damage 4Fe-4S of SufR to DNIC or apo-form, respectively, resulting in the de-repression of the suf operon in vitro and in vivo (10).In this context, data using iNOS inhibitor and iNOS −/− mice confirm the induction of the suf operon in response to NO during infection of macrophages and mice.The absence of a survival defect of MtbΔiscS when inside the immune-activated murine macrophages suggests that the NO/ H 2 O 2 -driven induction of suf operon might compensate for the loss of IscS.Consistent with this observation, we showed that residual survival of MtbΔiscS under aerobic or H 2 O 2 stress is due to the expression of the suf operon.
Our phenotypic data suggest that IscS and Suf systems work in distinct ways.Mtb requires IscS system more under hypoxic conditions where the demand for Fe-S clusters is minimal.As the demand for Fe-S cluster increases, such as during growth in an oxygen-rich environment, H 2 O 2 , and upon exposure to NO stress, Mtb gradually shifts its reliance from IscS to Suf for maintaining Fe-S cluster homeostasis and survival.The E. coli Isc system is generally poisoned with as low as 1 μM H 2 O 2 , whereas the Suf system remains resistant to oxidative stress (6).Similar differences in sensitivity to oxidants are expected for Mtb IscS and Suf systems (10,11,17).However, the artificial overproduction of IscS suppresses the growth defect of MtbΔiscS and renders the mutant more resistant to H 2 O 2 than WT Mtb (17).Thus, in principle, the induction of SUF/ISC pathways elevates the total cellular capacity for cluster assembly during high-demand conditions rather than being intrinsically more resistant to H 2 O 2 or NO.Agreeing with this, overexpression of the Suf system in MtbΔiscS inside murine lungs increased bacillary load and pathological defects more than WT Mtb.
During infection in mice, Mtb is exposed to a more potent RNIperoxynitrite that primarily damages Fe-S clusters with a high-rate constant (~1.0 × 10 10 M −1 s −1 ) (66).While NO largely induces the Suf operon to the same level in WT Mtb and MtbΔiscS in vitro, peroxynitrite exposure during murine infection could have caused additional damage to Fe-S cluster proteins including SufR in the mutant, leading to more induction of Suf system in MtbΔiscS and hence hypervirulence.This agrees with our findings showing both the loss of suf inducibility and hypervirulence phenotype of MtbΔiscS in iNOS −/− mice lacking the ability to produce peroxynitrite.Consistent with a dominant role of the Suf system during stress, organisms that are routinely exposed to H 2 O 2 (e.g., lactic acid bacteria, Enterococcus faecalis, Xylella fastidiosa, and chloroplasts) rely upon Suf rather than ISC for cluster assembly (6,67,68).While true for other organisms, Mtb's persistence depends upon an adequate response toward oxidative/nitrosative stress and hypoxia.Therefore, the importance of IscS in mediating the response to hypoxia in vitro and suppressing hypervirulence by adjusting Suf expression in vivo provides the biological significance of IscS in the persistence of Mtb.This notion must be tested in animal models (e.g., guinea pigs and nonhuman primates) (69), exposing Mtb to hypoxic lesions during infection.

Measurement of E MSH using the Mrx1-roGFP2 redox biosensor
The intra-mycobacterial E MSH (defined as the standard reduction potential of the MSH reduced /MSSM oxidized redox couple) determination during in vitro growth of Mtb, MtbΔiscS, and iscS-comp was performed as in (22).Briefly, bacterial cultures expressing Mrx1-roGFP2 were treated with 5 mM N-ethylmaleimide (Sigma-Aldrich, St. Louis, MO) for 5 min at room temperature followed by 4% paraformaldehyde (PFA) fixation (Himedia, Mumbai, India) for 1 hour at room temperature.Bacteria were analyzed using a FACSVerse flow cytometer (BD Biosciences, San Jose, CA).The E MSH was calculated using the Nernst equation as described previously in (22).The biosensor response was quantified by measuring the fluorescence ratio at a fixed emission (510 nm) on excitation at 405 and 488 nm.The data obtained were analyzed with the BD FACSuite software.These ratiometric data were normalized to measurements of cells treated with 10 mM CHP (Sigma-Aldrich, St. Louis, MO), giving maximal oxidation of the biosensor, and 20 mM dithiothreitol (Sigma-Aldrich, St. Louis, MO), yielding a readout of maximal reduction of the biosensor.Ten thousand events per sample were analyzed.Biosensor response was measured in a similar way for bacterial cells exposed to ROS (80 μM CHP or 60 μM Mnd) after 4-hour treatment.

ROS measurement using CellROX deep red
Exponentially growing cultures of Mtb, MtbΔiscS, and iscS-comp were taken for measurement of intracellular ROS.According to the manufacturer's instructions, CellROX Deep Red (Invitrogen, Waltham, MA) was added to a final concentration of 5 μM, and cells were stained by incubating on a rocker for 30 min at 37°C.The cells were then washed with 1× phosphate-buffered saline (PBS; pH 7.4) to remove residual dye by centrifugation (5000 rpm for 5 min).Last, cells were resuspended in 300 μl of PBS and fixed using 4% PFA for 1 hour at room temperature.Fluorescence was measured using a BD FACSVerse flow cytometer (BD Biosciences, San Jose, CA) at a fixed emission maxima of 670 nm allophycocyanin (APC) channel) after excitation with a red laser (640 nm) for 10,000 events per sample.No autofluorescence was observed.
For determination of intracellular ROS in the presence of Bip and Thio, exponentially grown cultures at an initial OD 600 (optical density at 600 nm) of 0.6 (10 ml) were treated with 250 μM Bip and 10 mM Thio separately and incubated for 24 hours in a shaker incubator (180 rpm, 37°C).This was followed by the above procedure for CellROX Deep Red staining and ROS measurement.

Cellular iron estimation
Intracellular iron levels were measured using the ferrozine-based colorimetric assay described previously (70,71).Briefly, exponentially grown cultures of Mtb strains (OD 600 ~0.8) were harvested by centrifugation and washed twice with ice-cold PBS.The cell pellets were resuspended in 1 ml of 50 mM NaOH and lysed using a bead beater (MP Biomedicals, Santa Ana, CA).HCl (10 mM, 300 μl) was added to the cell lysate samples (300 μl), followed by the addition of the Fe-detection reagent (6.5 mM ferrozine, 6.5 mM neocuproine, 1 M ascorbic acid, and 2.5 M ammonium acetate in water) (90 μl).This reaction mix was incubated for 30 min at 37°C, followed by reading the absorbance of the samples at 562 nm using a microplate reader (Versa-Max; Molecular Devices, San Jose, CA).The cellular free iron concentration was equated by plotting the absorbance values against a standard curve of FeCl 3 concentration gradient and readouts normalized to the protein content of the respective samples.Protein concentration was estimated using a Pierce bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, Rockford, IL).

Determination of bacterial survival
The number of viable bacilli was determined after treatment with 250 μM iron chelator Bip or 10 mM ROS scavenger Thio by removing aliquots from the cultures followed by preparation of 10-fold dilutions.Twenty microliters of dilutions was trailed on 7H11 agar plates.Plates were incubated for 3 to 5 (for MtbΔiscS) weeks at 37°C, followed by CFU enumeration.

Establishment of hypoxia
To determine the viability of Mtb, MtbΔiscS, and iscS-comp under hypoxic conditions, bacterial cultures (OD 600 = 0.1) were injected into vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ) followed by incubation at 37°C for 12 to 16 days (58,72).A marker for the achievement of hypoxia was the decoloration of MB (final concentration of 1.5 μg/ml) (Sigma-Aldrich, St. Louis, MO) in the culture medium.Once hypoxia was established, cells were harvested for determination of bacterial survival by CFU.

Growth curves for Mtb strains
Freezer stocks of Mtb strains were revived in 7H9-OADC broth with 1:10 dilution and grown for 1 week.These cultures were subcultured and inoculated in 7H9-ADS at OD 600 ~0.05 and incubated at 37°C at 180 rpm, and growth was tracked for 8 to 9 days by measuring the OD 600 until the cultures reached a stationary phase.

Aconitase activity
The aconitase activity was measured by monitoring the disappearance of cis-aconitate at wavelength (λ) 240 nm in an ultraviolet spectrophotometer (Thermo Fisher Scientific, Biomat 3S, USA) as described (11).One unit of aconitase activity is defined as 1 μmol of cis-aconitate formed or converted per minute.Reaction mixtures for aconitase assay contained 25 mM tris-HCl (pH 8.0), 100 mM NaCl, and 50 μg of Mtb cell lysates in 1 ml of reaction volume.Reaction was initiated by adding 0.15 mM cis-aconitate and monitored by following the disappearance of cis-aconitate at λ 240 nm after every 15 s for 30 min.Absorbance at λ 240 nm was plotted against time.Aconitase activity was calculated from linear portion of the curve in initial 5 min.An extinction coefficient of 3500 M −1 cm −1 was used to calculate the rates.

Metabolite extraction and analysis
Metabolite extraction from all Mtb strains was performed as outlined in (11).Briefly, exponentially growing Mtb, MtbΔiscS, and iscS-comp cultures (OD 600 ~0.6) were quenched with four volumes of 60% methanol for 5 min (maintained at −45°C) in a dry ice-methanol bath, followed by centrifugation at 4000 rpm for 5 min (at −5°C).The pellet was resuspended in 700 μl of 60% methanol (maintained at −45°C) and centrifuged at 4000 rpm for 5 min.The pellet obtained was resuspended in 1 ml of 75% ethanol and incubated at 80°C for exactly 3 min, with intermittent mixing at 1.5-min intervals, followed by incubation on ice for 5 min and centrifugation at 13,000 rpm for 15 min.The final supernatant was lyophilized and then stored at −80°C until further analysis.
Steady-state levels of metabolites were analyzed using methods described in (73).Briefly, extracted metabolites were first separated using a Synergi 4-μm Fusion-RP 80 Å (150 × 4.6 mm, Phenomenex) LC column on the Shimadzu Nexera ultrahigh-performance LC system.For TCA intermediates, derivatization was done before separation.The following solvent systems were used: 0.1% formic acid in water (solvent A) and 0.1% formic acid in methanol (solvent B) for amino acids, nucleotides, and TCA metabolites and 5 mM ammonium acetate in water (solvent A) and 100% acetonitrile (solvent B) for sugar phosphates.The flow parameters used were as described in (73).The mass spectrometer used was AB Sciex Qtrap 5500, and data acquisition was made by Analyst 1.6.2software (Sciex).Amino acids and TCA intermediates were detected in positive polarity, while sugar phosphates were detected in negative polarity mode.The parent and daughter ion masses for each metabolite are given in table S3.The data were analyzed by calculating the area under the obtained peaks by using Multi Quant (version 3.0.1)software.

Estimation of NAD + , NADH, NADP + , and NADPH
Mtb strains were grown to OD 600 ~0.8 and harvested for the detection of pyridine nucleotide levels by a redox-cycling assay as described in (74,75).NADPH and NADP + concentrations were normalized to the protein contents of NADH and NAD + extracts using BCA protein assay.

OCR and ECAR measurements
To estimate the basal OCR and ECAR, Mtb, MtbΔiscS, and iscScomp cultures of OD 600 ~0.6 were briefly (24 hours) incubated in 7H9 medium containing the nonmetabolizable detergent tyloxapol (MP Biomedicals, Santa Ana, CA) and devoid of ADS or any carbon source.These cultures were then harvested, and single-cell suspensions of bacteria were prepared by passing five times through a 26gauge syringe needle followed by centrifugation at 100g for 1 min.Cells (2 × 10 6 per well) were seeded into the wells of an XF cell culture microplate (Agilent/Seahorse Biosciences, Santa Clara, CA) coated with Cell-Tak (Corning, Corning, NY).Measurements were carried out using a Seahorse XFp analyzer (Agilent Technologies, CA) with unbuffered 7H9 as the assay media (pH 7.35; lacking disodium phosphate and monopotassium phosphate).Basal readings were taken for an initial 18 min, followed by an injection of D-glucose (10 mM), giving glucose-induced respiration rates.The maximum rate of respiration was achieved by the addition of two consecutive doses of 2 and 8 μM CCCP (Sigma-Aldrich, St. Louis, MO).Readouts were normalized to the number of bacteria seeded.

RNA isolation, amplification, and library preparation for RNA-seq
Mtb strains grown to an OD 600 of 0.4 were harvested in 5 M guanidinium thiocyanate (GTC) buffer (containing 1% 2-mercaptoethanol, 0.5% sarcosyl, and 0.5% Tween 80) for total RNA extraction.Total RNA was extracted using a FastRNA Pro Blue kit (MP Biomedicals, USA) and further purified with RNeasy spin columns (Qiagen, USA) as described (14).The concentration and quality of extracted RNA were checked spectrophotometrically using NanoDrop ND-1000 (Thermo Fisher Scientific).Total RNA was then enriched for mRNA by depletion of 16S and 23S rRNA using the MICROBExpress Kit (Life Technologies, USA), and the concentration of the ribo-depleted RNA was quantified by the QuBit RNA HS Assay Kit (Life Technologies).Fifteen nanograms of mRNA per sample was taken for fragmentation and random priming, followed by first and second cDNA synthesis and library preparation using the NEBNextUltra II Directional RNA Library Prep Kit for Illumina (New England Biolabs), according to the manufacturer's protocol.The library size distribution and quality were assessed using a high-sensitivity DNA chip (Agilent Technologies).Last, equimolar quantities of all libraries were pooled and sequenced in a high-throughput run on the Illumina HiSeq 2500 sequencer (14).

Differential gene expression and statistical analysis for RNA-seq
The reference genome sequence (.fna) and annotation (.gff ) files for the Mtb H37Rv strain (accession number: NC_000962.3)were downloaded from the National Center for Biotechnology Information (NCBI) ftp website ("ftp.ncbi.nlm.nih.gov").The annotation file format (.gff ) was changed to .bed using an in-house Python script.Sequencing-based raw reads obtained (as .fastq)were checked for quality using FastQC software (version v0.11.5; http://www.bioinformatics.babraham.ac.uk/projects/fastqc). BWA (version 0.7.12-r1039) (76) was used to index the reference genome.Reads having raw read quality of ≥20 were aligned using the BWA aln -q option.SAMTOOLS (version 0.1.19-96b5f2294a)(77) was used to filter out the multiple mapped reads.Read count per gene was calculated using BEDTOOLS (version 2.25.0)(78) with the annotation (.bed) file.The normalization and DEG analysis for the data were carried out using edgeR, as mentioned previously (79).DEGs were determined on the basis of the cutoff: absolute fold change ≥ 1.5 and FDR ≤ 0.05.
For overlap analysis of DEGs with the processed data of other studies, the significance of gene number overlap was determined by Fisher's exact test on a two-by-two contingency table.The universal set of the genes was calculated on the basis of the intersection of total genes analyzed in the studies with our study.If this information was not available, then we took 3975 genes as the universal set (table S2).

Bacterial stress survival assay
Exponentially growing Mtb, MtbΔiscS, and iscS-comp (OD 600 ~0.6) cultures were diluted to an OD 600 of ~0.15 and exposed to 80 μM CHP, 60 μM Mnd (Sigma-Aldrich, St. Louis, MO), 1 mM DETA-NO (Cayman Chemical, USA) and incubated at 37°C in a shaker incubator.After 4 and 24 hours of treatment, cells were serially diluted and plated on 7H11-ADS plates.Colonies were enumerated after an incubation of 3 to 5 weeks at 37°C.Macrophage preparation and Mtb infection RAW264.7 murine macrophage cell line was used for the ex vivo infection study.This cell line was acquired from the American Type Culture Collection (Manassas, VA) and tested negative for mycoplasma contamination using the DE-MycoX Mycoplasma PCR Detection Kit (CELLclone, catalog no.GX-E-250).Mtb infection was performed in either naïve or activated macrophages.Macrophage activation was achieved by treating RAW264.7 cells with interferon-γ (IFN-γ) (30 ng/ml; PeproTech, catalog no.315-05) and E. coli lipopolysaccharide (LPS) (10 ng/ml; Sigma-Aldrich, catalog no.L2630) for 16 hours.These activated macrophages were separately treated with a specific inhibitor of iNOS, 1400W dihydrochloride (25 μM) (Sigma-Aldrich, catalog no.W4262).Macrophage cells were infected with Mtb, MtbΔiscS, and iscS-comp at an MOI of 1:2 for 4 hours, followed by washing thoroughly with warm PBS (137 mM NaCl + 2.7 mM KCl + 10 mM Na 2 HPO 4 + 1.4 mM KH 2 PO 4 ) and Dulbecco's modified Eagle's medium (DMEM; CELLclone) to remove all extracellular bacteria and finally added fresh media [DMEM + 10% fetal bovine serum (MP Biomedicals, catalog no.092916754)] with or without IFN-γ and incubated at 37°C with 5% CO 2 .To determine bacterial survival after infection at different time points, infected cells were lysed in 0.06% SDS in PBS.Lysates were serially diluted and plated on 7H11-OADC agar plates.Plates were incubated at 37°C until colonies appeared for CFU enumeration.

Determination of MIC and antibiotic susceptibility assay
MIC was evaluated by a MABA in a 96-well plate format (58).Exponentially growing Mtb strains (OD 600 ~0.6) were harvested to prepare 1 × 10 6 cells/ml density culture.Approximately 1 × 10 5 bacteria were added per well in a total volume of 200 μl of 7H9 + ADS medium containing a gradient of drug concentrations.Controls for the assay consisted of wells lacking Mtb for a media background and wells devoid of a drug for maximum bacterial growth.Following an incubation period of 5 days at 37°C, 30 μl of 0.02% resazurin (Sigma-Aldrich, catalog no.R7017) was added, and plates were incubated for another 24 hours at 37°C.Fluorescence intensity was detected using a SpectraMax M3 plate reader (Molecular Devices, San Jose, CA) in a bottom-reading mode with excitation at 530 nm and emission at 590 nm.Percent inhibition for respective antibiotics was derived from the relative fluorescence values compared to the untreated control.MIC was considered the minimum antibiotic concentration that yielded at least 90% reduction in fluorescence compared to the untreated growth control.Antibiotic concentration ranges were as follows: Bdq (1.8 to 0.005 μM), Mox (1 to 0.003 μM), Rif (1.13 to 0.003 μM), and Inh (7.3 to 0.021 μM).

Antibiotic kill kinetic assay
Exponentially growing cultures of Mtb strains were brought to a density of 1 × 10 6 CFU/ml and incubated with or without 1× MIC 90 (for Mtb) of antibiotics: Mox, Rif, and Inh in 10-ml cultures for a period of 10 days in a shaker incubator at 37°C.The survival kinetics of the bacteria was monitored by plating for viable cells on days 0, 1, 3, 5, 7, and 10 after treatments.Following incubation at 37°C for 3 to 5 weeks, viable colonies were enumerated.

In vivo infection experiments
For the mice infection model, 5-to 6-week-old female BALB/c mice (n = 5 per group) were infected via an aerosol route with approximately 100 bacilli per mouse with Mtb, MtbΔiscS, and iscS-comp strains using a Madison chamber by aerosol generation.At indicated times after infection, mice were sacrificed, and the lungs and spleens were harvested for determination of bacillary load and tissue histopathology analysis as described (14).The bacillary load was quantified by plating serial dilutions of tissue homogenates on 7H11-OADC agar plates supplemented with lyophilized BBL MGIT PANTA antibiotic cocktail (polymyxin B, amphotericin B, nalidixic acid, trimethoprim, and azlocillin; as supplied by BD Biosciences, USA).Colonies were enumerated after 4 weeks of incubation at 37°C.Pathological analyses were represented as granuloma scores as described previously (14).
To determine the role of NO in Mtb infection and pathogenesis, 5-to 7-week-old female iNOS −/− mice (n = 6) were infected with approximately 100 bacilli per mouse with Mtb and MtbΔiscS strains using a Madison chamber by aerosol generation.Mice were humanely sacrificed, and the lungs were harvested for determination of bacillary load.The bacillary load was quantified by plating serial dilutions of tissue homogenates as described above.All mice for the above experiments were obtained from the institute Central Animal Facility (CAF).

Generation of conditional knockdowns using CRISPRi strategy
Construction of SufS-KD strains was carried out using the CRISPRi technology as described in (54).Briefly, an inactive form of Streptococcus pyogenes Cas9 enzyme harboring mutations D10A and H820A (dCas9) resulting in nuclease inactivity offers the ability to control targeted gene expression transiently or stably without altering the genomic sequence.dCas9 was expressed in the integrative plasmid system pRH2502 from a TetR-regulated uvtetO promoter.sufS was targeted by expressing a gene-specific sgRNA in the episomal vector pRH2521 under the control of a TetR-regulated smyc promoter (Pmyc1tetO).For sufS depletion, sgRNAs were designed for the region 52 to 75 bp of the sufS ORF and cloned in pRH2521.Both the plasmids pRH2502-dCas9 and pRH2521-sgRNA were sequentially electroporated into Mtb and MtbΔiscS strains, followed by the selection of positive clones on KAN + HYG plates.The SufS − KD was achieved by culturing the KD

Fig. 1 .
Fig. 1.IscS is required to maintain redox balance of Mycobacterium tuberculosis (Mtb).(A) Diagrammatic representation of endogenous reactive oxygen species (ROS) generation (O 2 −• , H 2 O 2 , and • OH) due to the univalent reduction of O 2 via electron leak from redox-active enzymes.O 2 −• and H 2 O 2

Fig. 4 .
Fig. 4. MtbΔiscS displayed altered expression of genes regulated by iron-sulfur (Fe-S)-containing WhiBs, SufR, and DosR/S/T system.Total RNA was isolated from Mycobacterium tuberculosis (Mtb) cultures, grown to 0.4 OD 600 , and subjected to RNA sequencing (RNA-seq) analysis.(A) Mycobrowser-based classification of the genes/ pathways [1.5-fold change; false discovery rate (FDR) ≤ 0.05] deregulated in MtbΔiscS.(B to D) Heatmaps showing gene expression changes belonging to various functional categories.Heatmaps were constructed with row z-score on normalized logCPM values.(E) The table summarizes the overlap between the IscS transcriptome with the whole-genome expression under different stress conditions and transcriptomes of Mtb-whiB1 KD, MtbΔwhiB3, and MtbΔsufR.Fisher's exact test with *P < 0.05 as a cutoff for significance (tableS2).

Fig. 5 .
Fig. 5. MtbΔiscS shows enhanced sensitivity to anti-tuberculosis (TB) drugs.(A) Mtb, MtbΔiscS, and iscS-comp grown until OD 600 values of 0.6 were treated with various concentrations of Rif, Mox, Bdq, and Inh, and MIC 90 was determined using Alamar blue assay.Data are representative of two independent experiments done in duplicate.(B) Alamar blue dataset was used to calculate percent inhibition of MtbΔiscS by Rif compared to Mtb and iscS-comp.The strains Mtb, MtbΔiscS, and iscS-comp were treated with 1× MIC 90 of (C) Mox (D) Rif, and (E) Inh, and survival over time was monitored by enumerating colony-forming units (CFUs).Data are presented as means ± SEM. (B to E) *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001 calculated by unpaired two-tailed t test.ns, not significant.

Fig. 7 .
Fig. 7. MtbΔiscS displays hypervirulence in mice.BALB/c female mice (n = 5) were given aerosol challenge with Mycobacterium tuberculosis (Mtb) , MtbΔiscS, and iscScomp and assessed for survival in (A) lungs and (B) spleen at indicated time points.(C) The gross pathology of the infected lungs is shown after 4 and 8 weeks of infection.Black arrows show the granulomatous lesions formed upon Mtb infection.The white circles highlight the patches of lung consisting of multiple granulomatous lesions in the MtbΔiscS-infected lungs.(D) Hematoxylin and eosin-stained lung sections were imaged (at ×4 magnification) and analyzed for histopathology on Mtb infection.Changes in lung morphology are shown with formation of granuloma (G) and the normal alveolar spaces (AS).(A) P ≥ 0.05, *P ≤ 0.05, ***P ≤ 0.001, and ****P ≤ 0.0001 calculated by two-way analysis of variance (ANOVA) with Bonferroni's multiple comparisons test.(B) *P ≤ 0.05 and ***P ≤ 0.001 calculated by unpaired two-tailed t test.

Fig. 8 .
Fig. 8. Nitric oxide (NO) induces the expression of Suf system and contributes to hypervirulence of MtbΔiscS in mice.Gene expression analysis by quantitative reverse transcription polymerase chain reaction (qRT-PCR) of bacterial Suf genes (sufS and sufC) in Mtb, MtbΔiscS, and iscS-comp isolated from infected RAW264.7 murine macrophages (A) Lipopolysaccharide (LPS) + interferon-γ (IFN-γ) activated and (B) activated + iNOS inhibitor (1400W) after 48 hours of infection.(C) Gene expression of sufS was analyzed in Mtb, MtbΔiscS, and iscS-comp isolated from mouse lung after 8 weeks of infection.(D) Similarly, transcript levels of sufS were determined in Mtb and MtbΔiscS isolated from 3-week-infected mouse lung.(A and B) Fold change in transcript levels is compared to that of respective strains grown under standard in vitro conditions and compared to in vitro Mtb (C to D).Data are presented as means ± SEM. (E and F) iNOS −/− female mice (n = 6) were given aerosol challenge with Mtb and MtbΔiscS and assessed for survival in the lung after 4 weeks of infection.(F) Bacterial burden was determined by plating lung homogenates and colony-forming unit (CFU) enumeration.(E) The transcript levels of sufS were estimated in Mtb and MtbΔiscS isolated from 4-week-infected lungs of iNOS −/− mice.(A, B, and F) ****P ≤ 0.0001, calculated by two-way analysis of variance (ANOVA) with Bonferroni's multiple comparisons test.(C to E) Statistical significance was analyzed over untreated control by paired two-tailed t test (**P < 0.01 and ****P < 0.0001).ns, not significant.

Fig. 9 .
Fig. 9. Overexpression of Suf system contributes to hypervirulence of MtbΔiscS in mice.(A) BALB/c mice (n = 5) were given aerosol challenge with Mtb-sufSKD and ΔiscS-sufSKD and divided into three groups of (i) no doxycycline treatment (−Dox), (ii) acute phase [doxycycline started at 7 days after infection (+Dox acute)], and (iii) chronic phase [doxycycline started at 21 days after infection (+Dox chronic)] (credit: BioRender.com).Post-infection animals were sacrificed from the (B) acute group at days 7 and 21 and (C) at days 21 and 42 from the chronic group, and colony-forming unit (CFU) per lung was measured.(D) Gross pathology of Mtb-sufSKD and ΔiscS-sufSKD-infected lungs at 42 days after infection.White arrows and white circles show the granulomatous lesions formed upon Mycobacterium tuberculosis (Mtb) infection.(E) Hematoxylin and eosin-stained lung sections were imaged (at ×40 magnification) and analyzed for histopathology on Mtb infection.Changes in lung morphology are shown with formation of granuloma (G) and the normal alveolar spaces (AS).(B and C) ****P ≤ 0.0001, calculated by two-way analysis of variance (ANOVA) with Bonferroni's multiple comparisons test.