Redox-dependent condensation of the mycobacterial nucleoid by WhiB4.

Oxidative stress response in bacteria is mediated through coordination between the regulators of oxidant-remediation systems (e.g. OxyR, SoxR) and nucleoid condensation (e.g. Dps, Fis). However, these genetic factors are either absent or rendered non-functional in the human pathogen Mycobacterium tuberculosis (Mtb). Therefore, how Mtb organizes genome architecture and regulates gene expression to counterbalance oxidative imbalance is unknown. Here, we report that an intracellular redox-sensor, WhiB4, dynamically links genome condensation and oxidative stress response in Mtb. Disruption of WhiB4 affects the expression of genes involved in maintaining redox homeostasis, central metabolism, and respiration under oxidative stress. Notably, disulfide-linked oligomerization of WhiB4 in response to oxidative stress activates the protein's ability to condense DNA. Further, overexpression of WhiB4 led to hypercondensation of nucleoids, redox imbalance and increased susceptibility to oxidative stress, whereas WhiB4 disruption reversed this effect. In accordance with the findings in vitro, ChIP-Seq data demonstrated non-specific binding of WhiB4 to GC-rich regions of the Mtb genome. Lastly, data indicate that WhiB4 deletion affected the expression of ~ 30% of genes preferentially bound by the protein, suggesting both direct and indirect effects on gene expression. We propose that WhiB4 structurally couples Mtb's response to oxidative stress with genome organization and transcription.

Introduction

Mycobacterium tuberculosis (Mtb) is the causative agent of tuberculosis (TB), which is one of the major global human health problems. Upon infection, Mtb is phagocytosed by alveolar macrophages, where it is exposed to redox stresses such as reactive oxygen and nitrogen species (ROS and RNS) [1], [2]. To overcome host-generated redox stress, Mtb exploits several resistance mechanisms including alkyl hydroperoxidase, catalase, superoxide dismutases, thioredoxins, glucose-6-phosphate dehydrogenase, methionine-sulfoxide reductase, and the membrane-associated oxidoreductase complex [3], [4], [5], [6], [7], [8], [9].

Apart from these response mechanisms, the bacterial nucleoid also undergoes dynamic changes in its architecture in response to stresses such as ROS, RNS, and changes in osmolarity and pH [10]. Nucleoid-associated proteins (NAPs) including Dps (DNA-binding protein from starved cells) and Fis (Factor for inversion stimulation) in E. coli, MrgA (Metallo regulated genes A) in Staphylococcus aureus, DR0199 (the homolog of EbfC) in Deinococcus radiodurans, HU (the homolog of Hup protein) in Helicobacter pylori, and Lsr2 in Mtb protect the cells from Fenton-mediated oxidative damage by physically shielding and compacting genomic DNA [11], [12], [13], [14], [15]. Whereas regulated condensation of the nucleoid protects the cells against oxidative stress, long-lasting condensation has been shown to perturb normal cellular processes such as replication and transcription and stimulate oxidative stress-mediated death [16]. Despite the recognized importance of bacterial NAPs in modulating the oxidative stress response, the effector proteins that monitor the changes in the cytoplasmic redox potential and sculpt the genome architecture in response to oxidative stress remain unknown. Although Mtb NAPs such as HupB (HU) and EspR impact pathogen’s survival and virulence, respectively [17], [18], whether these proteins organizes chromosome under physiologically-relevant stress conditions is not known. In Mtb, several studies have demonstrated the role of redox-responsive WhiB proteins (WhiB1 to WhiB7) in regulating gene expression and controlling a plethora of functions including antibiotic resistance, oxidative/nitrosative/acidic stress response, immune-response, cell division, and secretion [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29]. However, several functional aspects of WhiB proteins in Mtb have not been investigated. Previously, we have shown that one of the WhiB family member, WhiB4, is required to regulate survival in response to oxidative stress in vitro and to regulate virulence in vivo [20]. A fundamentally important question remains yet unanswered; what is the mechanism by which WhiB4 coordinates oxidative stress response in Mtb? More importantly, while electrophoretic mobility shift assays (EMSAs) revealed DNA binding activity of some of the Mtb WhiB proteins [30], [27], [29], an in vivo evidence for redox-dependent DNA binding at a genomic scale using chromatin immunoprecipitation sequencing (ChIP-seq) is still lacking. In this study, we employed multiple analytical techniques to show that WhiB4 dynamically manipulates both DNA architecture and gene expression to control oxidative stress response in Mtb.

Methods

Bacterial strains and growth conditions

E. coli cultures were grown in Luria-Bertani (LB) medium (BD Biosciences). When required, the culture medium was supplemented with hygromycin (50 µg ml-1 for mycobacteria, 150 µg ml-1 for E.coli), kanamycin (25 µg ml-1 for mycobacteria and 50 µg ml-1 for E.coli), and ampicillin (100 µg ml-1). For cumene hydroperoxide (CHP, Sigma Aldrich) stress, strains were grown to an exponential phase and exposed to different CHP concentrations. Survival was monitored by enumerating colony forming units (CFU) at 0, 6, 24, and 48 h post-treatment. To examine the influence of WhiB4 over-expression on stress tolerance, whiB4-OE strain was grown aerobically till O.D.600 of 0.3, induced with 200 ng ml-1 Anhydrotetracycline (Atc, Cayman Chemicals) for 24 h at 37oC or as indicated, and exposed to (i) normal aerobic environment, (ii) acidic stress (pH 4.5), and (iii) heat stress (42oC). The growth kinetics was monitored over time by measuring absorbance at 600 nm. Various Mtb strains and primers used in this study were listed in Supplementary information Table S6 and S7 and were cultured as described previously [20].

Microarray experiments

The wt Mtb, MtbΔwhiB4, and whiB4-Comp strains were grown in 7H9 medium supplemented with 1× ADS (Albumin Dextrose Saline enrichment) to an O.D.600 of 0.4 and exposed to 0.25 mM CHP for 2 h at 37oC. The experiment was carried out with two biological replicates. At 2 h post-treatment with CHP, total RNA was purified as described [20]. The total RNA was processed and hybridized to the Mtb Whole Genome Gene Expression Profiling microarray- G2509F (AMADID: G2509F_034585, Agilent Technologies PLC). DNA microarrays were provided by the University of Delhi South Campus MicroArray Centre (UDSCMAC). RNA amplification, cDNA labeling, microarray hybridization, scanning, and data analysis were performed at the UDSCMAC as described previously [31]. Slides were scanned on a microarray scanner (Agilent Technologies) and analyzed using the GeneSpring software. Results were analyzed in MeV and data was considered significant at p ≤ 0.05. The normalized data has been submitted to NCBI's Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/); accession number GSE73877 and GSE114045.

qRT-PCR

Total RNA was isolated from logarithmically grown cells of wt Mtb, MtbΔwhiB4 and whiB4-OE strains and qRT-PCR was performed using gene specific primers (Supplementary information Table S7) as described [20].

Electrophoretic Mobility Shift Assays (EMSA)

Histidine-tagged version of WhiB4 and its various cysteine mutants were purified by as described elsewhere [20]. The oxidized apo-form of WhiB4 was prepared as described previously [20]. To assess the non-specific DNA binding ability of WhiB4, oxidized apo-WhiB4 was incubated with various DNA molecules in a buffer containing 89 mM Tris, 89 mM boric acid and 1 mM EDTA (pH 8.4) for 30 min. The protein-DNA complexes were resolved on a 1% agarose gel in 1× TBE buffer, stained with ethidium bromide and visualized under UV light. As a control, EMSAs were repeated with a known NAP (histidine tagged HU protein; a kind gift from V. Nagaraja , IISc Bangalore) and a non-NAP transcription factor (histidine-tagged TcrX protein; a kind gift from Deepak K. Saini , IISc Bangalore).

In vitro transcription assays

Approximately 350 ng of oxidized apo-WhiB4 was incubated with 150 ng of a pGEM plasmid in a final volume of 20 µl, followed by in vitro transcription using the Riboprobe in vitro Transcription Kit according to the manufacturer’s instructions (Promega). Samples were treated with 6% SDS (Amresco Inc.) and 4 mg ml-1 proteinase K (Amresco Inc.) for 30 min at 37oC and the reaction products were analyzed on a 1% agarose gel.

In vivo thiol-trapping analysis

Histidine-tagged WhiB4 was expressed under its native promoter in MtbΔwhiB4. At an O.D.600 of 0.4, MtbΔwhiB4 expressing histidine-tagged WhiB4 was treated with 0.1 and 0.5 mM CHP for 24 h. Cultures were treated with 10 mM NEM for 5 min to block the thiol-state of WhiB4. Bacterial pellets were resuspended in lysis buffer [300 mM NaCl, 20 mM Na-Phosphate, 10% Glycerol and 1X protease inhibitor (Amresco Inc.), pH 7.5] and lysed using bead beater (MP Biomedicals). Approximately 30 µg of total cell lysate was resolved by 12% non-reducing SDS-PAGE. Proteins were transferred on to 0.2 µm PVDF membrane and used for Western blot. Western blot analysis was achieved using 1:20,000 dilution of anti-His antibody (GE Life Sciences) for 12 h. The blotted membrane was developed with a 1:20,000 dilution of peroxidase-conjugated anti-mouse IgG (Cell Signaling) and an enhanced chemiluminescence substrate (ECL, Bio-Rad).

Immuno-blot analysis

The expression of FLAG-tagged version of wt WhiB4 or WhiB4-cys3 variant was induced using various concentrations of Atc for 16 h. For the immuno-blot assay, bacterial cells were processed as described earlier. WhiB4 was detected using anti-FLAG primary antibody. The blot was developed with enhanced chemiluminescence (ECL, Bio-Rad).

Atomic force microscopy (AFM)

The oxidized apo-WhiB4 was incubated for 1 min with 7 ng ml-1 of supercoiled or relaxed forms of plasmid DNA (pEGFP-C1) in a concentration range from 1:2.5–1:20 (DNA: WhiB4; w/w) at RT and 10 µl of this solution was loaded onto freshly cleaved mica surface. The unbound material was washed with deionised water and the bound surface allowed to air dry. Imaging was carried out using the 5500 scanning probe microscope (Agilent Technologies, Inc.) and the PicoView software. Images were obtained in tapping mode in the air with 225-mm-long silicon cantilevers (Agilent Technologies) that have a resonance frequency of 75 kHz and a force constant of 2.8 Newton m-1. Scan speed used was one line s-1. Minimum image processing (first order flattening and brightness contrast) was employed. Image analysis was performed using Pico Image software v1.4.4. A similar procedure was adopted for WhiB4-Cys3, Mtb HU and Mtb TcrX proteins.

Confocal microscopy

Various Mtb strains were grown to exponential phase (O.D.600 of 0.4) in 7H9 medium and WhiB4 expression was induced as described [20]. Cells were fixed with paraformaldehyde (PFA) and washed with 1X PBS and stained with 4’,6-diamino-2-phenylindole (DAPI, 1 µg ml-1, Invitrogen). The bacterial cells were visualized for DAPI fluorescence (excitation at 350 nm and emission at 470 nm) in a Leica TCS Sp5 confocal microscope under a 63X oil immersion objective. Quantification of nucleoid size (larger axis) of 100–150 independent cells was carried out using public domain program OBJECT-IMAGE J. Using these images, relative nucleoid size (RNS) was measured by determining the ratio between length of the nucleoid(s) and the length of the cell by relying on the end points of their larger axes. We only considered cells with bilobed or irregular shaped nucleoid for measurements. For this measurement, we assumed that the width of the nucleoid and that of Mtb cells are similar with no significant differences. To calculate RNS, each bacterium was zoomed in and brightness/contrast was adjusted to clearly visualize the contour of nucleoids. After scaling the image to appropriate size using scaling tool, line tool was used to measure the length of the larger axes of nucleoids and cells. Since the images are two-dimensional and Mtb cells are small, the value of RNS are merely approximations indicating the trend of compactions or expansion of the nucleoid under conditions studied.

Immunofluorescence microscopy

The MtbΔwhiB4 cells overexpressing either FLAG-tagged WhiB4 (whiB4-OE strain) or histidine-tagged WhiB4 (His-WhiB4) were fixed with 4% PFA, washed with 1X PBS and permeabilized using 0.1% of Triton-X 100. Cells were blocked with 2% bovine serum albumin (BSA) and incubated with anti-FLAG or anti-His primary antibody. The cells were incubated with anti-mouse IgG Alexa Fluor® 594-conjugated secondary antibody (Invitrogen) followed by staining with DAPI as described earlier. A Leica TCS Sp5 confocal microscope under a 63X oil immersion objective was used for imaging.

Live dead cell staining

CHP treated cells were washed with 1X PBS and stained using the Live/Dead BacLight Bacterial Viability Kit (Invitrogen). The cells were fixed with 4% PFA and imaged for SYTO9 dye (excitation at 480 nm and emission at 500 nm) and propidium iodide (excitation at 490 nm and emission at 635 nm) in a Leica TCS Sp5 confocal microscope.

ChIP-Seq

DNA–protein interactions were characterized by cross-linking 100 ml of culture of wt Mtb or whiB4-OE strains (O.D.600 of 0.4–0.5) with 1% formaldehyde while agitating cultures at 37 °C for 30 min. Crosslinking was quenched by the addition of glycine to a final concentration of 125 mM. Cells were pelleted, washed in 1X PBS+1X protease inhibitor cocktail (Sigma), and resuspended in IP Buffer (20 mM K-HEPES pH 7.9, 50 mM KCl, 0.5 mM dithiothreitol and 10% glycerol) + the 1 X protease inhibitor cocktail. Samples were sonicated at high amplitude in a water bath sonicator (Bioruptor, Diagenode) set for 30 s ON & 30 s OFF at 4 °C. Shearing of DNA to an average size of 100–400 bp was confirmed by running a de-crosslinked aliquot on 1% agarose gel. Cell debris was removed by centrifugation and the supernatant was used in the IP experiment. The samples were mixed with buffer IPP150 (10 mM Tris-HCl—pH 8.0, 150 mM NaCl and 0.1% NP40) and immunoprecipitation of FLAG-tagged proteins was initiated by incubating samples overnight rotating at 4 °C with Anti-FLAG® M2 Magnetic Beads (Sigma: M8823). Beads were washed twice with IP buffer and once with Tris-EDTA buffer pH 7.5. Elution was performed in 50 mM Tris–HCl pH 7.5, 10 mM EDTA, 1% SDS for 40 min at 65 °C. Samples were finally treated with RNAse A for 1 h at 37 °C, and cross-links were reversed by incubation for 2 h at 50 °C and for 8 h at 65 °C in elution buffer with Protease K. DNA was purified by phenol-chloroform extraction and quantified. The concentration of the immunoprecipitated DNA was measured using the Qubit HS DNA kit. The resulting ChIP-DNA was subjected to qRT PCR analysis to determine the enrichment in the immunoprecipitated sample (whiB4-OE) over the control sample (wt Mtb) after normalization with input. Truseq ChIP kit was used for the library preparation. DNA was end-repaired and adapters were ligated. After ligation, products were purified and amplified. Samples were sequenced on the HiSeq. 2500 Sequencing System.

ChIP- Sequencing analysis

The single end reads obtained were aligned to the reference genome H37Rv (ASM19595v2) using the bwa-0.7.12 version [32]. The tab delimited .sam files as output were then converted to .bam format using samtools-1.2 [33]. We then used bedtools2 to convert .bam into .bed files, and both the resultant ChIP and input files in replicates were used in MACS peak calling algorithm (version2) [34], [35]. The p value cutoff for the peaks was set to be 0.01. Peaks obtained for each condition were visualized in the UCSC microbial genome browser [36]. MEME-ChIP suite (web interface) was used to search for motifs in the WhiB4 peaks [37]. The analysis was also performed using the Z-score approach [38]. This approach assumes normal distribution of the read counts and firstly normalizes the ChIP samples and the corresponding input controls individually with the mode of the distribution. Mode was calculated using the Short function in Genefilter package in R (R Core Team (2016). R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/) [39] and genefilter: methods for filtering genes from high-throughput experiments. R package version 1.60.0 (https://rdrr.io/bioc/genefilter/). We compared the peaks obtained from both the methods and found ~ 90% similarity. The GC% was calculated using bedtools2 nuc command, with the default parameters [34]. All the related statistical analyses were performed in R.

Peak coordinates were then intersected with the annotated H37Rv genome. Peaks that were found within 300 base pairs upstream and 50 base pairs downstream of the start sites of the genes were inferred as bound. In parallel, genes were categorized as up-regulated/down-regulated (>= 1.5 log2 fold change) from the microarray datasets. These bound genes were then overlapped with the up- and down-regulated genes to understand the effect of WhiB4 binding on the regulatory regions and the downstream effects on gene expression.

Macrophage infections

IFN-Ɣ+LPS activated RAW 264.7 macrophages were infected with wt Mtb and whiB4-OE strains at a multiplicity of infection 10 (MOI 10) of 2 for 4 h, followed by treatment with amikacin to remove extracellular bacteria. After infection, cells were washed thoroughly with warm DMEM medium and resuspended in the same containing 10% FBS. Samples were collected at the indicated time points, lysed using 0.06% SDS-7H9 medium and plated on OADC-7H11 agar medium for CFU determination.

Transmission electron microscopy

Cells were processed for transmission electron microscopy (TEM), as described previously [40], [41]. Cells were prefixed with 1% (w/v) osmium tetroxide buffered in 0.15 M cacodylate buffer (pH-7.2) (Sigma) for 1 h at room temperature. The prefixed cells were then washed once with the same buffer and post fixed for 2 h at room temperature in 0.15 M sodium cacodylate (Sigma) buffer containing 2% (w/v) tannic acid (Sigma) and 2% (v/v) glutaraldehyde (Sigma). Subsequently, the cells were subjected to washing with the same buffer, and were re-fixed in 1% (w/v) osmium tetroxide overnight at 4ºC. Cells were dehydrated in a graded series of ethanol solutions (Merck Millipore) ranging from 20% to 100% with an incubation period of 10 min at each step and finally embedded in LR White resin (Electron Microscopy Sciences) overnight. The embedded samples were then cut with a glass knife using an ultramicrotome by maintaining the section thickness at 70 nm. The sections were stained with 0.5% uranyl acetate (Sigma) and 0.04% lead citrate (Fluka), and observed using FEI Tecnai™ G2 Spirit electron microscope at 120 kV.

EMSH measurements

Measurements of intrabacterial E during growth in vitro and upon exposure to CHP stress were performed as described previously [42].

Statistical analysis

The overlap was assessed using the GeneOverlap package within R package version 1.6.0 (http://shenlab-sinai.github.io/shenlab-sinai/) and statistical significance was calculated using the Fisher’s exact test. GeneOverlap: Test and visualize gene overlaps. R package version 1.6.0.). Chip-Seq data has been submitted to NCBI's Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/); accession number GSE100440. Unless otherwise stated all data were graphed and analyzed with Prism v7.0 (GraphPad ). Statistical analyses were performed using Student’s t-tests (two-tailed). Where comparison of multiple groups was made either one-way or two-way ANOVA with Bonferroni multiple comparisons was performed. Differences with a p value of < 0.05 were considered significant.

Results

Mtb WhiB4 regulates gene expression in response to oxidative stress

Previously, we had shown that WhiB4 functions as an autorepressor and marginally influences the expression of 23 genes in Mtb under unstressed conditions [20]. Furthermore, a whiB4 mutant survived better under oxidative stress in vitro, in immune-activated macrophages, and was hypervirulent in animals [20]. In agreement with these findings, exposure to increasing concentrations of an oxidant, cumene hydroperoxide (CHP) induced a dose dependent decrease in whiB4 expression as compared to that of 16s rRNA (Fig. 1A). These findings indicate that the expression of whiB4 and its regulatory function thereof could be dependent on oxidative stress.

Fig. 1

Oxidative stress downregulatesexpression in a dose dependent manner. (A) Total RNA was isolated from wt Mtb upon treatment with the indicated concentrations of CHP for 2 h. qRT-PCR for whiB4 was done using whiB4-specific oligonucleotides and fold change was normalized to 16 s rRNA expression. Data are represented as mean ± standard deviation (SD). In each experimental condition n = 3; two tailed Student’s t-test was used to determine statistical significance and is indicated where differences were significant. ** p value ≤ 0.01, *** p value ≤ 0.001 (as compared with 0.1 mM CHP treated wt Mtb). (B-D) Generation ofknockout in. (B) A step-wise illustration for construction of an unmarked strain of MtbΔwhiB4. Upper right panel denotes wt Mtb whiB4 loci. After the allelic exchange, entire whiB4 ORF in Mtb genome was replaced by right and left flanking regions of whiB4 along with the loxP-hyg-gfp-loxP cassette. This knock out strain was unmarked by expressing Cre recombinase to remove the hyg-gfp cassette. (C) Genomic DNA was isolated from putative MtbΔwhiB4 colonies (7, 8, 9, 11, 14, 16, 19 and 23) which were KanSHygRGFP+ and replacement of whiB4 allele with the Hyg/GFP cassette was confirmed using PCR with F1 and R1 primers (Supplementary information Table S7). An increase in amplicon size from 1.2 kb to 3.3 kb due to insertion of the loxP-hyg-gfp-loxP cassette was observed in case of mutant clones, confirming the double crossover event. (D) RNA was isolated from logarithmically grown wt Mtb and the putative MtbΔwhiB4 clones. qRT-PCR for whiB4 was done using whiB4 specific oligonucleotides (Supplementary information Table S7) and Ct values were plotted to assess the expression. Amplification was detected based on fluorescence emitted by Sybergreen upon interaction with DNA and plotted as Relative Fluorescence Units (RFU).

Since the whiB4 mutant reported earlier retained expression from initial 96 bases [20], we first deleted the entire open reading frame (ORF) encoding WhiB4 in Mtb (MtbΔwhiB4) (Fig. 1B–D). Next, we performed global transcriptome profiling of wt Mtb, MtbΔwhiB4, and whiB4-complemented (whiB4-Comp) strains upon exposure to 0.25 mM CHP. The whiB4-Comp strain was constructed by restoring the expression of whiB4 from its native promoter in MtbΔwhiB4 [20]. We minimized any influence of oxidative stress-induced cell death on gene expression by performing microarrays at an early time point (2 h) post-CHP treatment. Treatment with CHP influenced the expression of 614 and 429 genes in wt Mtb and MtbΔwhiB4, respectively, as compared to untreated controls (Supporting information Table S1). First, we analyzed the effect of CHP on wt Mtb [2-fold up- and down-regulated, p value ≤ 0.05]. Genes encoding proteins implicated in mitigating oxidative stress such as thioredoxins (trx), catalase (katG), alkyl hydroperoxidase (ahpCD), protein kinase G (pknG), isocitrate lyase (icl), and Fe-S biogenesis/repair operon (suf system) were induced by CHP (Fig. 2 and Supporting information Table S1). Since oxidative stress damages DNA, protein, and lipids, a significant fraction of genes mediating DNA repair, lipids biogenesis, and protein quality control were induced by CHP (Fig. 2). Genes encoding energetically efficient respiratory complexes such as NADH dehydrogenase I (nuo operon) were down-regulated, whereas the energetically less favored NADH dehydrogenase type II (ndh) and fumarate reductase (frdB) were induced in response to CHP (Fig. 2 and Supporting information Table S1). These changes along with the induction of genes associated with the pentose phosphate pathway, glyoxylate shunt, and nicotinamide metabolism involved in maintaining NAD(P)H/NAD+(P) poise implicate a strategic shift from energy generation to redox balance in response to CHP stress. As expected, oxidative stress triggers the expression of several transcription factors controlling iron homeostasis (furA and ideR), protein turn-over (hspR), and redox homeostasis (clgR, nrdR, sigE, sigB, and sigH) (Fig. 2). It is known that NAPs, DNA gyrases, and topoisomerases modulate the nucleoid architecture of bacteria in response to oxidative stress [13], [43], [44]. However, microarray analysis showed that topoisomerase 1 (topA) was down-regulated and NAPs such as espR, lsr2, and hupB were unaffected upon oxidative stress in Mtb (Fig. 2 & Supporting information Table S1). It has been reported that Mtb displays massive transcriptional changes in response to non-lethal concentrations of another oxidant hydrogen peroxide (H2O2; 5–10 mM) [45]. We compared gene expression changes published in response to H2O2 (10 mM; 40 min exposure) [45] with the CHP transcriptome. Approximately, ~ 48% of genes induced by CHP also displayed induction upon H2O2 treatment, indicating significant similarities between the oxidative stress response induced by H2O2 and CHP (Supplementary information Table S1).

Fig. 2

WhiB4 mediated regulation of gene expression in response to oxidative stress in. Wt Mtb, MtbΔwhiB4, and whiB4-Comp strains were grown and exposed to 250 μM CHP for 2 h. Total RNA was isolated and subjected to microarray analysis. Heat maps depicting the comparison of gene expression (Log2 fold-change, p value ≤ 0.05) coordinating respiration, CCM, DNA damage, PE-PPE and redox balance between wt Mtb [T] vs wt Mtb [UT]; Mtb∆whiB4 [T] vs wt Mtb [T] and whiB4-Comp [T] vs wt Mtb [T] from two biological samples. UT- untreated, T- CHP treated.

Thereafter, we examined the role of WhiB4 in regulating oxidative stress-induced changes in gene expression. To identify genes whose regulation under CHP conditions is influenced by WhiB4, we compared gene expression changes between CHP-treated MtbΔwhiB4 and CHP-treated wt Mtb. We found that the expression of 275 genes was induced and 198 genes was repressed in MtbΔwhiB4 as compared to that in wt Mtb upon CHP-treatment (Supplementary information Table S1). Genes playing overlapping roles in redox-metabolism and central carbon metabolism (CCM) such as dlaT, bkdC, aceE, ahpE, and ahpCD were significantly induced in MtbΔwhiB4 in response to CHP as compared to that in wt Mtb (Fig. 2) [46], [47], [48], [49], [50], [51]. While nuo operon was down-regulated in MtbΔwhiB4 in response to CHP treatment (Supplementary information Table S1), the compensatory increase in the expression of ndh, frdA, and cytochrome bd oxidase (cydAB) was notably higher in MtbΔwhiB4 than in wt Mtb (Fig. 2 and Supplementary information Table S1). In bacteria, including Mtb, cytochrome bd oxidase also displays catalase and/or quinol oxidase activity [52], [53], which confers protection against oxidative stress. This is consistent with the higher potential of MtbΔwhiB4 to tolerate oxidative stress [20], which is further supported by a substantially lesser number of DNA repair genes induced by CHP in MtbΔwhiB4 as compared to wt Mtb (Fig. 2). Several members of the PE-PGRS gene family involved in maintaining cell wall architecture and protection from oxidative stresses were highly expressed in MtbΔwhiB4 (Fig. 2 & Supplementary information Table S1) [54], [55]. Further, whiB4 deletion increased the expression of transcription factors involved in metal sensing (kmtR, smtB, and zur), antibiotic tolerance (blaR), and virulence (virS), whereas the expression of whiB6 was reduced (Fig. 2). In Mtb H37Rv, expression and activity of ESX-1 secretion system is linked to regulatory mechanisms involving PhoP, WhiB6 and EspR [21], [56], [57]. We observed that genes encoding ESX-1, ESX-3, and ESX-4 family members were down-regulated in MtbΔwhiB4 upon CHP treatment (Fig. 2), indicating a cross-talk between ESX-associated regulators (e.g., whiB6/phoP/espR) and whiB4.

Lastly, we determined how many of genes induced or repressed in MtbΔwhiB4 upon CHP treatment were complemented by the expression of whiB4. To do this, we compared wt Mtb CHP expression profile with the expression profile of CHP-treated whiB4-Comp. We found that expression of ~ 75% of genes showing differential regulation in MtbΔwhiB4 was normalized to nearly wt Mtb levels in the whiB4-Comp strain under CHP stress (Fig. 2 and Supplementary information Table S1). For example, differential expression of genes associated with redox metabolism, CCM, PE-PGRS, ESX-systems, and alternate respiration in MtbΔwhiB4 was significantly stabilized in whiB4-comp (Fig. 2 and Supplementary information Table S1). However, expression of ~ 25% of genes remained altered in both MtbΔwhiB4 and whiB4-Comp, indicating the involvement of other genetic factors in regulating their expression upon oxidative stress (Supplementary information Table S1). Lastly, we validated microarray results by measuring the expression of a select set of genes affected by CHP in both wt Mtb and MtbΔwhiB4 by qRT-PCR. As a control, we also examined the expression of genes which were not affected by CHP. As shown in the Table S2, qRT-PCR confirmed differential expression of only CHP-responsive genes in wt Mtb and MtbΔwhiB4.

Mtb WhiB4 possesses NAP-like DNA binding properties

Having established that WhiB4 influences gene expression upon oxidative stress, we next sought to determine the underlying mechanism. It is known that despite containing redox active 4Fe-4S cluster, WhiB4 binds DNA in the apo-form. Furthermore, the DNA binding of apo-WhiB4 is regulated by the redox state of its cysteine (Cys) residues [20]. The thiol-reduced form of apo-WhiB4 exists as a monomer and lacks DNA binding, whereas oligomeric form of oxidized apo-WhiB4 binds at a specific locus (ahpCD) with a poor sequence specificity [20]. The findings, along with the low molecular weight (13.1 kDa) and a highly basic pI (10.28) of WhiB4, are reminiscent of various nucleoid-associated proteins (NAP) (e.g., HNS, HU, IHF and Lrp) [10]. Since oxidative stress caused by CHP exposure is likely to generate oxidized apo-WhiB4 in vivo, it is possible that under oxidative conditions WhiB4 influences gene expression by interacting non-specifically with the nucleoid. Using in vivo thiol-trapping experiment, we confirmed that WhiB4 largely exists as a monomer of thiol-reduced WhiB4 under normal growing conditions in Mtb (Fig. 3B–C and Supplementary information SI Note I). In contrast, exposure to oxidative stress (0.1 and 0.5 mM of CHP) significantly increased the proportion of disulfide-linked dimer (10,000–14,000 molecules) and trimer (4000–6000 molecules) of oxidized apo-WhiB4 per Mtb cell (Fig. 3B–C and Supplementary information SI Note I). Next, we investigated whether oxidized apo-WhiB4 possesses a NAP-like DNA binding properties in vitro and inside Mtb. Similar to other bacterial NAPs [11], [10], the oxidized apo-WhiB4 formed a high molecular weight complex with a range of DNA substrates (e.g., supercoiled DNA, linearized DNA, and 1 kb λ DNA ladder) and inhibited transcription from a standard T7-promoter of pGEM plasmid in vitro (Fig. 4A–C). The inclusion of the thiol-reductant, dithiothreitol (DTT), or replacement of any cysteine residues with alanine in WhiB4, reversed the DNA binding and transcriptional inhibitory activities of oxidized apo-WhiB4 (Fig. 4D-E). As controls, we showed that a well-established NAP (Mtb HU, Rv2986c [17]) forms a high molecular weight complex with a supercoiled DNA or a 1 kb λ DNA ladder, whereas a non-NAP transcription factor (Mtb TcrX, Rv3765c [58]) remained ineffective (Fig. 4F–G). In contrast to WhiB4, addition of DTT did not influence the activities of HU or TcrX in these assays (Fig. 4H–I).

Fig. 3

CHP stress induces the generation of oxidized apo-WhiB4 oligomerso. (A) Indicated concentrations of purified histidine-tagged WhiB4 were subjected to immuno-blot analysis using anti-His antibody. Band intensities were quantified by Image J and plotted to generate a standard curve. (B)MtbΔwhiB4 expressing chromosomally-integrated histidine-tagged WhiB4 from native promoter were grown till an OD600 nm of 0.8, harvested, and cell free extract was analyzed for monitoring expression and oligomerization of WhiB4 under normal growing conditions and upon exposure to 0.1 and 0.5 mM of CHP for 24 h. To minimize the possibility of O2-induced thiol oxidation and subsequent oligomerization of apo-WhiB4 during cell-free extract preparation, cells were pre-treated with the thiol-alkylating agent NEM as described [20]. Approximately 30 µg of the cell-free extract was resolved on a 12% non-reducing SDS-PAGE and immuno-blotted using anti-his. (C) The band intensities were calculated using Image J and compared with the standard curve to calculate the amount of WhiB4 expressed inside a single Mtb cell. These values were converted to WhiB4 copies per cell as described in Materials and Methods and plotted. Data shown are the representatives of two independent experiments. The asterisk (**P < 0.01/****P < 0.0001) represents a significant difference in the band intensity of monomer and dimer between UT and treated using One-way ANOVA with Bonferroni's post hoc test. M: monomer, D: dimer, T: trimer.

Fig. 4

WhiB4 exhibits non-specific DNA binding activities. (A) 150 ng of either linear (L) or supercoiled (SC) plasmid DNA was incubated with oxidized apo-WhiB4 (250 ng [lanes 2, 6]; 300 ng [lanes 3, 7]; 350 ng [4, 8]). Lanes 1 and 5 represent DNA without WhiB4. (B) Binding of apo-WhiB4 (100, 150, 200, 250, 300 and 350 ng) with 1 kb λ-DNA ladder (λ) (150 ng). (C) WhiB4 represses transcription. In vitro transcribed pGEM plasmid in the absence of apo-WhiB4 (lane 1), oxidized apo-WhiB4 in complex with pGEM (lane 2), in vitro transcription of pGEM in the presence of oxidized (lane 3) or reduced (lane 4) apo-WhiB4. The reactions in lane 3 and 4 were treated with 6% SDS and 4 mg ml-1 protease K to remove WhiB4 from the mixture before separation. C: pGEM plasmid DNA alone. Cysteine residues regulate DNA binding of WhiB4 (D) DTT-treatment abolished DNA binding of WhiB4. Oxidized apo-WhiB4 (350 ng) was treated with DTT (0, 200, 400, 800 mM) for 2 h followed by binding to the supercoiled plasmid DNA. C: represents DNA without WhiB4. (E) Purified wt apo-WhiB4 and Cys mutant variants were treated with thiol oxidant, diamide, and the gel-shift assay was performed as described earlier. DNA binding assays using Mtb HU (50–350 ng) and Mtb TcrX (150–350 ng) with (F) supercoiled DNA and (G) 1 kb λ-DNA ladder. (H) 350 ng of Mtb HU or (I) Mtb TcrX were treated with increasing concentrations of DTT (0, 200, 400, 800 mM) for 2 h followed by binding to the supercoiled plasmid DNA. The samples were analyzed on 1% agarose gel.

Further, we examined apo-WhiB4 DNA binding properties using Atomic Force Microscopy (AFM). A super-coiled plasmid DNA (pEGFP-C1) was pre-exposed to increasing concentrations of oxidized apo-WhiB4 and subjected to AFM. The oxidized apo-WhiB4 and plasmid DNA were taken as the experimental controls. AFM images in the absence of WhiB4 showed uniform structure for the negatively supercoiled plasmid DNA (Fig. 5A). Interestingly, at lower amounts of oxidized apo-WhiB4, a gradual relaxation of supercoiled plasmid DNA with initial opening at the ends, followed by a mixed population of partially or fully opened DNA circles was detected (Fig. 5B–C). Further increase in oxidized apo-WhiB4 resulted in the formation of multiple loops, bends and condensation within a single DNA molecule (Fig. 5D–F). A further increase in the protein concentrations may lead to nucleoprotein filament formation. However, higher concentrations of oxidized apo-WhiB4 forms large aggregates on the mica surface, thus precluding any further AFM imaging. An assessment of various geometric properties of the imaged protein: DNA complexes confirmed an initial increase in diameter and length followed by a decrease in both parameters at higher protein concentrations (Fig. 5G). This dual compactional ability of WhiB4 indicates a stoichiometry-dependent regulatory role for WhiB4, as shown for other chromosomal architectural proteins HU and RdgC in E.coli [59], [60], [61]. The Cys3-WhiB4 (replacement of the third cysteine to alanine) mutant protein was unable to exhibit a similar effect on supercoiled pEGFP-C1 (Supplementary information Fig. S1), indicating that WhiB4 Cys residues function as a thiol-based redox switch that modulates DNA condensing activity. Lastly, similar to apo-WhiB4, Mtb HU initially relaxed supercoiled pEGFP-C1 at a lower concentration followed by extensive DNA compaction at higher concentrations (Supplementary information Fig. S2). As expected, Mtb TcrX did not influence DNA architecture (Supplementary information Fig. S2).

Fig. 5

AFM analysis of WhiB4-mediated DNA condensation. (A) Supercoiled plasmid DNA, (B-F) Oxidized apo-WhiB4 incubated with supercoiled plasmid in increasing concentration (DNA: protein w/w =1:2.5, 1:5,1:10,1:15, and 1:20). (B and C) Low protein to DNA ratio gradually opens up the DNA until a fully relaxed circular DNA molecule is formed as observed in the case of a moderate protein to DNA ratio. (D and E) Further increase in protein to DNA ratio reverses the phenomenon when lateral compaction occurs. (F) High protein to DNA ratio further causes condensation of the DNA molecules thereby leading to rigid filamentous structures. The scale of images is 0.8 µm × 0.8 µm, Scale bar is 100 nm. (G) Geometric properties of the imaged protein: DNA complexes. Geometrical parameters of supercoiled DNA in the absence and presence of oxidized apo-WhiB4. (A- DNA only, B- low protein to DNA ratio, C-E- moderate protein to DNA ratio, F- high protein to DNA ratio; n = 70 independent DNA molecules measured in each case). Data are represented as mean ± SEM.

WhiB4 condenses the mycobacterial nucleoid

To further understand the link between DNA condensation and WhiB4, we examined the nucleoid morphology of wt Mtb, MtbΔwhiB4, whiB4-Comp, and whiB4-OE strains. The whiB4-OE strain overexpresses FLAG-tagged WhiB4 from an anhydrotetracycline (Atc)-inducible promoter system [20]. As controls, we similarly overexpressed Mtb HU (HU-OE) and Mtb TcrX (TcrX-OE). We stained the nucleoids of various strains with 4’,6-diamidino-2-phenylindole (DAPI) and visualized the cells by confocal microscopy. DAPI stained cells from exponentially grown cultures of Mtb frequently showed the presence of expanded nucleoids with a fewer bilobed nucleoid. Infrequently, single cells containing more than 3 distinct DAPI stained regions were also observed. Using these images we measured relative nucleoid size (RNS) by determining the ratio between the length of the nucleoid(s) and the length of the cell by relying on the end points of their larger axes (region with see Methods). Since the images are two-dimensional and cells are small, the values of RNS are merely approximations indicating the trend of compactions or expansion of the nucleoid under conditions studied. The differences in RNS values across multiple samples and assay conditions were comprehensively analyzed for significance using ANOVA followed by multiple post-hoc adjustment tests (Bonferroni, Sidak, and Tukey’s). Please refer to Supplementary information Table S3 for a complete list of p values.

Wt Mtb, MtbΔwhiB4, and whiB4-Comp showed expanded nucleoids with a slightly higher RNS value for MtbΔwhiB4 (Fig. 6 and Supplementary information Fig. S3). Overexpression of WhiB4 resulted in a significant compaction of the nucleoid (Fig. 6 & Supplementary information Fig. S3). The mean RNS value reduced by ~ 50% (0.125 ± 0.002 µm) in case of whiB4-OE as compared to wt Mtb (0.261 ± 0.005 µm, p value < 0.0001) or MtbΔwhiB4 (0.291 ± 0.005 µm, p value < 0.0001). Overexpression of HU induced compaction similar to WhiB4 (mean RNS = 0.120 ± 0.002 µm), whereas the nucleoid remained expanded upon TcrX overexpression (mean RNS = 0.265 ± 0.007 µm) (Fig. 6 and Supplementary information Fig. S3). Over-expression of untagged WhiB4 also condensed the mycobacterial nucleoid, ruling out the influence of the FLAG-tag on genome compaction (Supplementary information Fig. S4). Further, our intracellular localization studies using FLAG-specific antibody followed by indirect immuno-fluorescence revealed that WhiB4 is associated with the DAPI-stained clumped nucleoids of whiB4-OE (Fig. 7A). In contrast, over-expression of the cysteine mutant of WhiB4 (whiB4-cys3-OE) did not induce DNA condensation (mean RNS = 0.255 µm ± 0.046, p value < 0.001 as compared to whiB4-OE), and the mutant protein was found scattered across the length of the cell (Fig. 7A–B). Overall expression of WhiB4 was maintained in the whiB4-cys3-OE strain, indicating that the loss of nucleoid condensation is likely due to disruption of the thiol-disulfide redox switch (Fig. 7C).

Fig. 6

WhiB4-mediated nucleoid compaction in. (A) Nucleoid of wt Mtb, MtbΔwhiB4, whiB4-OE, whiB4-Comp, HU-OE and TcrX-OE strains were stained with DAPI (pseudo colored green) and visualized under confocal microscopy (63X). (B) Relative Nucleoid Size (RNS) (larger axis; inset) of ~ 100–150 independent cells of various Mtb strains. For overexpression, 100 ng ml-1 of Atc was added to cultures. As a control, the same amount of Atc was added to other strains. Scale bar = 2 µm. Data shown are the average of two independent experiments done in triplicate. Results are expressed as Mean ± SEM. One-way analysis of variance (ANOVA) with Bonferroni's post hoc test was employed to determine statistical significance between the RNS of different strains. Comparisons in which p value is < 0.05 were indicated with the symbol (). Symbol: comparison to wt Mtb.

Fig. 7

Cysteine residues of WhiB4 are important for its DNA condensation activity. (A) Confocal microscopic images (63X) of MtbΔwhiB4 strain over-expressing WhiB4- and WhiB4-cys3 FLAG tagged proteins. Nucleoids were stained with DAPI (pseudo colored green), wt WhiB4-FLAG and WhiB4-cys3-FLAG mutant were stained with the AF594 secondary antibody (red) against the anti-FLAG primary antibody, and co-localization was indicated by yellow in the merge panel. For inducing WhiB4, 100 ng ml-1 of Atc was added to cultures of whiB4-OE and whiB4-cys3 strains. The uninduced (UI) control lacks immuno-fluorescence due to the absence of WhiB4 expression. The scale bar = 3 µm. (B) The RNS values of ~ 100-150 independent cells of various Mtb strains. Results are expressed as mean ± SEM. The significance of differences between the strains were determined by one-way ANOVA with Bonferroni's post hoc test. Comparisons whose P value is <0.05 were indicated with different symbols. Symbols: (): comparison to wt Mtb; (): comparison to whiB4-OE. (C) 30 µg of cell free extract from either whiB4-OE or whiB4-cys3-OE strains was analyzed for the WhiB4 expression by immuno-blotting using the antibody against FLAG tag. The figure shown is the representative image of two independent experiments.

We substantiated WhiB4-induced nucleoid condensation by performing ultrastructure imaging of the mycobacterial nucleoid using Transmission Electron Microscopy (TEM). Analysis of mid-logarithmic (log) phase cells overexpressing WhiB4 showed a highly condensed nucleoid, unlike the well spread out nucleoid in the wt Mtb cells, MtbΔwhiB4, and whiB4-Comp (Fig. 8, compare C-D with A-B, E-F, & G-H, respectively). The nucleoid morphology of the wt Mtb cells was as reported previously [62], [40], [63]. Similar to whiB4-OE, TEM of HU overexpressing strain showed a highly compacted nucleoid, whereas TcrX-OE displayed nucleoid architecture comparable to wt Mtb (Fig. 8I-L).

Fig. 8

Transmission electron microscopic (TEM) analysis ofnucleoids. TEM images of the longitudinal and transverse sections of: (A, B) wt Mtb from mid-log phase culture showing a well spread out nucleoid in the middle of the cell; (C, D)whiB4-OE cells showing highly condensed nucleoid in the middle of the cell; (E, F)MtbΔwhiB4 showing unorganized and well spread out nucleoid throughout the cell; (G, H)whiB4-comp showing nucleoid like wt Mtb (A, B). (I, J)HU-OE showing highly condensed nucleoid, and (K, L)TcrX-OE showing a well spread nucleoid. Arrowhead indicates nucleoid. For induction, 100 ng ml-1 of Atc was used. As a control, the same amount of Atc was added to other strains. In each case, ~250–300 independent cells were visualized and representative images are shown.

A broadly comparable nucleoids of wt Mtb, MtbΔwhiB4, and whiB4-Comp suggests a negligible effect of WhiB4 on DNA condensation under normal growing conditions. This is in agreement with our in vivo thiol-trapping data showing that only a minor fraction of WhiB4 exists in a DNA-condensing redox form (i.e. oxidized apo-WhiB4 dimer/trimer) in vivo under native conditions. In contrast, a highly compacted nucleoid of whiB4-OE indicates the accumulation of DNA condensing oligomeric forms of oxidized apo-WhiB4 upon WhiB4 overexpression. To explore this, we performed in vivo thiol-trapping experiment upon Atc-mediated WhiB4 induction. We found that induction with 100 ng ml-1 of Atc resulted in ~ 14,000 dimers and 4000 trimers of disulfide-linked oxidized apo-WhiB4 per cell (Fig. 9 and Supplementary information SI Note II). The proportion of dimers and trimers were comparable to the oxidized apo-WhiB4 forms generated in Mtb upon treatment with 0.1–0.5 mM CHP. Since oxidized apo-WhiB4 functions as an autorepressor and also as a repressor of antioxidant systems [20], it is conceivable that the overexpression of WhiB4 by Atc circumvents this autoregulatory loop to amplify the negative effect of WhiB4 on antioxidants expression. Consequently, oxidative stress increases inside Mtb, resulting in oxidation of WhiB4 thiols to generate disulfide-linked oxidized apo-WhiB4 oligomers. Altogether, using multiple analytical techniques, we confirmed that WhiB4 participates in mycobacterial nucleoid compaction in a redox-dependent manner.

Fig. 9

Ectopic expression of WhiB4. (A)whiB4-OE harboring FLAG-tagged WhiB4 under Atc inducible promoter was cultured till OD600nm of 0.4. Expression of WhiB4 was induced by addition of indicated concentration of Atc for 16 h. Approximately, 30 µg of cell free extract from whiB4-OE strain was analyzed for WhiB4 expression by immuno-blotting using the antibody against FLAG tag. To minimize the possibility of O2-induced thiol oxidation and subsequent oligomerization of apo-WhiB4 during cell-free extract preparation, cells were pre-treated with the thiol-alkylating agent NEM as described [20]. Immuno-blot clearly revealed the appearance of monomer (M), dimers (D), and trimers (T) of WhiB4. (B) The band intensities were quantified by ImageJ software. Amount of WhiB4 expressed at a single cell level was quantified and plotted as described in Fig. 3.

WhiB4 regulates DNA condensation in response to oxidative stress

Having shown that WhiB4 modulates nucleoid condensation, we wanted to understand the influence of oxidative stress on mycobacterial nucleoids and the role of WhiB4 in this outcome. We treated wt Mtb, MtbΔwhiB4, and whiB4-OE strains with 0.5 mM CHP and monitored intramycobacterial oxidative stress, nucleoid condensation, and survival at various time points. To image oxidative stress in Mtb, we measured the redox potential of its most abundant cytoplasmic thiol (mycothiol; MSH) using Mrx1-roGFP2. The biosensor shows an increase in fluorescence excitation ratio at 405/488 nm upon oxidative stress, whereas a ratiometric decrease is associated with reductive stress [5]. The ratiometric changes (405/488 nm) in the fluorescence of the biosensor can be fitted to the modified Nernst equation to precisely determine the millivolt (mV) changes in the redox potential of mycothiol (E) [42].

Under unstressed conditions, both wt Mtb and MtbΔwhiB4 displayed a steady-state E of ~ −276 mV. However, overexpression of WhiB4 induces an oxidative shift in E of whiB4-OE (~ −250 mV), consistent with the antioxidant repressor function of WhiB4. Treatment with 0.5 mM CHP induces a significant oxidative shift in E of wt Mtb (-220 to −225 mV) at 6 h and 24 h post-treatment (Fig. 10A). Importantly, the nucleoids of wt Mtb underwent a noteworthy condensation at 6 h (mean RNS: 0.14 µm ± 0.02, p value < 0.001 as compared to 0 h) and 24 h (mean RNS: 0.10 µm ± 0.01, p value < 0.001) post-treatment (Fig. 10A). At 48 h post CHP-treatment, cells showed loss of DAPI fluorescence and appeared rounded/irregular, thus precluding any measurements of nucleoid length. These morphological alterations are indicative of DNA fragmentation and complete killing, which we confirmed by CFU analysis (Fig. 10B) and SYTO9-propidium iodide staining (Fig. S5). We expressed whiB4 fused with histidine-tag (His-tag) and examined colocalization of HIS-WhiB4 with DAPI stained nucleoids under CHP stress. Using His-tag antibody and indirect immunofluorescence, we confirmed that WhiB4 remained associated with the condensed nucleoid upon CHP exposure at 6 h and 24 h post-treatment (Fig. 11). Notably, the skew towards condensed DNA and oxidative E was activated at a time point where toxicity was modest (i.e. 6 h), indicating that genome hypercompaction likely precedes oxidative stress-induced killing of Mtb.

Fig. 10

Oxidative stress leads to DNA condensation, skewed redox homeostasis and bacterial killing in a WhiB4-dependent manner. (A) wt Mtb, MtbΔwhiB4 and whiB4-OE were exposed to 0.5 mM CHP for 6, 24 and 48 h. Nucleoids were stained with DAPI (pseudo colored green) and visualized by confocal microscopy (63X). The scale of images is 1 µm. For inducing WhiB4, 100 ng ml-1 of Atc was added to cultures of whiB4-OE. As a control, the same amount of Atc was added to other strains. The relative RNS of ~ 100-150 independent cells was measured and represented as scatter plots. Intramycobacterial E of various Mtb strains at each time point was measured using flow cytometer and shown along with a corresponding image of a representative DAPI stained cell visualized by confocal microscopy under similar conditions. Data are represented as mean ± SEM. One-way ANOVA with Bonferroni's post hoc test was employed to compare RNS at different time points between all the strains. Comparisons whose p value is <0.05 were indicated with different symbols. Symbols: (): comparison to wt Mtb 0 h; (): comparison to wt Mtb 6 h; (): comparison to wt Mtb 24 h; (): comparison to MtbΔwhiB4 0 h; (): comparison to MtbΔwhiB4 6 h; (): comparison to MtbΔwhiB4 24 h; (): comparison to MtbΔwhiB4 48 h; (): comparison to whiB4-OE 0 h. (B) Plots showing the survival of Mtb strains in response to CHP stress as assessed by enumerating CFUs (n=3). Error bars represent mean ± SD. Symbols indicate significant differences using two-way ANOVA with Bonferroni's post hoc test between strains at different time interval as indicated. Symbols: (): comparison to wt Mtb; (): comparison to MtbΔwhiB4.

Fig. 11

WhiB4 associates with condensed nucleoid upon oxidative stress. Confocal microscopic images (63X) of MtbΔwhiB4 cells expressing histidine-tagged WhiB4 (His-WhiB4) from native promoter. The intramycobacterial E of MtbΔwhiB4 was shifted towards oxidative using 0.5 mM of CHP and WhiB4 association with nucleoids was visualized at various time points post-treatment. Nucleoids were stained with DAPI (pseudo colored green), WhiB4 was marked using anti-His primary antibody and AF594 conjugated secondary antibody (red), and yellow merge indicated co-localization of WhiB4 with DAPI stained nucleoids. The scale bar is 2 µm.

In comparison to wt Mtb, nucleoids of MtbΔwhiB4 underwent a slower transition towards a condensed state, which corresponds to a moderate shift towards oxidative E (E: -247 to − 239 mV) and a slower killing by CHP (Fig. 10A–B). The degree of condensation displayed by MtbΔwhiB4 at 48 h of CHP treatment (RNS=0.118 ± 0.02 ) was similar to that observed in case of wt Mtb (RNS = 0.104 ± 0.01) at 24 h post CHP-treatment. Similarly, while MtbΔwhiB4 was successful in maintaining intermediate degree of oxidative E (-239 mV) and survival, wt Mtb displayed acute oxidation and killing at 48 h post-CHP treatment (Fig. 10A–B). However, the fact that MtbΔwhiB4 exhibited nucleoid compaction (albeit slower) in response to CHP suggests that WhiB4 is involved in mediating appropriate degree of DNA condensation perhaps with the involvement of one or more partner NAPs under oxidative conditions. In whiB4-OE, 0.5 mM CHP treatment for 6 h was sufficient to induce substantial DNA condensation (mean RNS= 0.06 µm ± 0.01), overwhelm redox balance (E: −225 mV ± 0.98) and induce substantial killing (Fig. 10A–B). At 24 and 48 h, CHP induces complete killing of whiB4-OE, precluding further measurements of RNS (Fig. 10B).

It can be argued that the high concentration of CHP (0.5 mM) can adversely affect Mtb’s physiology to influence genome topology. To address this issue, we reassessed nucleoid condensation, intrabacterial E, and survival upon exposure to a non-toxic concentration of CHP (0.1 mM) over time. As expected the effect of 0.1 mM CHP was moderate, however, the relative differences in RNS, E, and survival between wt Mtb, MtbΔwhiB4, and whiB4-OE followed a pattern similar to 0.5 mM CHP (Supplementary information Fig. S6 & SI Note III). Finally, to examine if the hypersensitivity to oxidative stress is unique to WhiB4 overexpression, we monitored survival of HU-OE and TcrX-OE in response to 0.1 mM CHP. In contrast to whiB4-OE, survival of HU-OE and TcrX-OE was not significantly affected by CHP (Supplementary information Fig. S7), indicating that WhiB4 directed changes in genome condensation and expression calibrate survival upon CHP stress.

Since our results indicate an inverse relationship between genome condensation and survival under oxidative stress, we examined the state of nucleoids in a Mtb strain completely devoid of MSH antioxidant (MtbΔmshA) and in a mshA complemented strain (mshA-comp) [64]. Similar to whiB4-OE, MtbΔmshA grows normally under aerobic conditions but shows oxidative E and acute sensitivity towards oxidative stress [65]. In agreement to findings with whiB4-OE, nucleoids of MtbΔmshA cells showed hypercondensation (mean RNS: 0.11 µm ± 0.04) as compared to the mshA-comp cells (mean RNS: 0.25 µm ± 0.02) (Supplementary information Fig. S8) under normal growing conditions. These results reinforce a functional association between DNA condensation and tolerance to oxidative stress in Mtb. In summary, our data indicate that deletion of WhiB4 reduces, and overexpression potentiates, the adverse impact of DNA condensation on oxidative stress survival.

ChIP-Seq demonstrates a nearly-uniform association of WhiB4 with Mtb chromosome

To determine genome-wide distribution of oxidized apo-WhiB4, we performed ChIP-Seq. The oxidized apo-WhiB4 can be generated by CHP treatment in vivo, however this form of WhiB4 functions as an autorepressor and therefore can pose challenges in identifying transient binding sites. To circumvent this, we utilized whiB4-OE, which consistently generates oxidized apo-WhiB4 to perform ChIP-Seq. We considered the possibility that ectopic expression might result in nonphysiological DNA binding. However, our data showed that over-expression of WhiB4 using 100 ng ml-1 of Atc produced nearly physiological concentrations of oxidized apo-WhiB4 molecules/cell (comparable to CHP treated Mtb). Furthermore, using qRT-PCR we confirmed that overexpression of FLAG-tagged or untagged WhiB4 influences the expression of only WhiB4-specific genes (Supplementary information Table S2). Further, whiB4-OE displayed survival defects exclusively under oxidative stress conditions in vitro and inside immune-activated macrophages (Fig. 12 and SI Note IV). Lastly, recent studies have demonstrated remarkable consistency between the DNA binding profile of several ectopically overexpressed FLAG-tagged transcription factors and ChIP-Seq studies performed under native conditions in Mtb [66]. Thus, while we cannot exclude the possibility that overexpression could lead to nonphysiological DNA binding, we conclude that WhiB4 overexpression did produce physiologically-relevant redox-form(s) necessary for DNA binding in Mtb.

Fig. 12

WhiB4 overexpression specifically reverses phenotypes displayed by. We have earlier shown that MtbΔwhiB4 survived better upon oxidative stress in vitro, inside immune-activated macrophages, and displayed hypervirulence in animals [20]. To investigate the functional relevance of WhiB4 overexpression, we determined the phenotype of WhiB4-OE under defined conditions specific to the WhiB4 function. The whiB4-OE strain was either subjected to 200 ng ml-1 of Atc induction or left uninduced. Cells were monitored for growth at different time intervals by measuring absorbance at 600 nm under (A) aerobic conditions, (B) 0.1 mM CHP, (C) 42ºC, and (D) pH 4.5. Data shown is the average of two experiments performed in triplicates. Error bars represent SD from the mean. (E) IFN-Ɣ-LPS activated RAW 264.7 macrophages were infected with wt Mtb, and whiB4-OE strains at an MOI of 10 and survival was monitored by enumerating CFU. Infected macrophages were maintained in 200 ng ml-1 of Atc for sustained induction of WhiB4. The data is shown as Mean ± SD from two independent experiments having triplicate wells each. Two tailed Student’s t-test was used to determine statistical significance and is indicated where differences were significant *P < 0.05.

We induced the expression of WhiB4 using 100 ng of Atc for 16 h in duplicate, harvested chromatin samples for ChIP-Seq using anti-FLAG antibody conjugated to magnetic beads, and sequenced the cross-linked DNA using the Illumina Genome Analyzer system. As a negative control, we sequenced the input chromatin sample prior to immunoprecipitation (IP) with the anti-FLAG antibody. In parallel, we performed the analysis of a published genome-wide DNA binding study that was conducted using a sequence-specific transcription factor CRP in Mtb [67]. Since the ChIP-seq data for Mtb HU are not available, we analyzed genome-wide DNA binding data of E.coli HU as an exemplar of a near-uniform DNA binding protein [68]. For each sample, we examined read count distribution by measuring the number of reads mapped to each base on the chromosome and then plotted the densities of read count distributions. The numbers of reads mapping to each position on the chromosome were first normalized to Z-scores as described previously [68], and represented as density distribution curves (Fig. 13). The normalization is based on the assumption that the noise in the ChIP experiment follows a normal distribution. The Z-score for the signal at any position on the chromosome is defined by the difference between the signal at that position and the mean of the normal noise distribution, divided by the standard deviation of the normal noise distribution. The distribution for data generated from input DNA sequencing shows a peak, whose position depends on the sequencing depth, and a short right hand tail (Fig. 13). The equivalent distribution for a highly specific DNA binding protein such as Mtb CRP shows a sharp peak to the left, corresponding to the background, and a long right handed tail corresponding to strong DNA binding signals (Fig. 13). A near-uniform DNA binding protein, such as HU in E. coli, displays a peak that is largely similar to the input, but with a slight right tail corresponding to regions of weak binding specificity (Fig. 13). Analysis of WhiB4 ChIP-seq resulted in a density distribution curve that showed similarity to input/HU with a slightly more prominent right tail corresponding to regions of weak DNA binding (Fig. 13). The length of the tail can be visually appreciated, and can be quantified using a standard statistical parameter, skewness; higher the skewness, more prominent the tail (Fig. 13 showing higher skewness for CRP than WhiB4). For clarity, we also plotted the Z-score distribution (Z-score > 3) within the right hand tail region (enriched regions of DNA binding) for ChIP experiments for WhiB4, CRP, and E. coli HU. In comparison to corresponding inputs, Z-score distributions were weakly skewed to the right for WhiB4 and E. coli HU, whereas distributions were noticeably skewed towards right in case of CRP (Fig. 14A). Consistent with this, the input normalized signal for peaks identified by the software Model-based Analysis of ChIP-Seq (MACS) for WhiB4 is significantly weaker than peaks identified for Mtb CRP (p value < 10-10), and comparable to those for E. coli HU (Fig. 14B). These observations can be clearly visualized in the UCSC genome browser view showing near-uniform binding of WhiB4 and HU across the genome as opposed to specific-regions of strong binding for CRP (see later [Fig. 15]). Finally, ChIP experiments for WhiB4 were consistently successful and obtained high concentrations of DNA (8.54 ng µl-1 and 8.52 ng µl-1) as compared to mock-IP (2.72 ng µl-1and 3.12 ng µl-1) or the strain overexpressing non-DNA binding FLAG-tagged WhiB4-Cys3 variant (3.31 ng µl-1and 2.98 ng µl-1). Altogether, the data support a more uniform binding of WhiB4 to the chromosome of Mtb.

Fig. 13

Density distribution plots for the Z-score distribution forWhiB4,HU andCRP. ChIP signal is shown in blue and the input signal is shown in black. Compared to CRP ChIP, the distribution of the WhiB4 and E. coli HU distribution does not have the characteristic heavy right end tail. This is also evident from the skewness measured for each distribution. Positive skewness values depict right-handed distribution as in the case of CRP ChIP and slightly in the WhiB4 ChIP.

Fig. 14

WhiB4 exhibits largely uniform binding to thechromosome. The left panel shows (A) boxplots displaying the distribution above the threshold Z-score (Z-score > 3) in the right-hand tail region (enriched DNA binding) for ChIP samples and the corresponding Z-score values in the input for WhiB4 replicate 1, HU and CRP. The difference between the ChIP enriched regions is significant against the input control in the case of CRP, a specific binding transcription factor (Wilcoxon test, P < 10 -10) as well as WhiB4 and HU (Wilcoxon test, P < 10 -10), (B) Fold enrichment distribution as observed for WhiB4, HU and CRP using MACS peak calling. On the y-axis is the fold enrichment denoted for each peak identified in the analyses. WhiB4 fold enrichment is less and similar to HU, whereas CRP shows higher values for fold enrichment in its peaks. (C) Boxplot showing GC content distribution between the obtained peaks for WhiB4 versus randomized peaks from the Mtb H37Rv genome. The WhiB4 peaks in comparison with the randomized input regions have higher GC content (Wilcoxon test, P < 10 -10). (D) GC-rich motif obtained from the WhiB4 peaks using MEME-ChIP suite (web interface).

Fig. 15

Genome wide mapping of WhiB4 binding sites. Input normalized HU, CRP, and WhiB4 binding on the respective chromosome as visualized on the UCSC genome browser with the corresponding GC% for WhiB4 and CRP. On the y-axis is the fold enrichment of the peaks. WhiB4 peaks in majority of the cases are observed in the genomic regions rich in GC content as seen in the zoomed in examples.

Since above results indicate the presence of weak specificity of WhiB4 to some loci across the Mtb genome, we identified these loci by performing peak identification using MACS. This analysis revealed that WhiB4 occupancy positively correlates with the G+C content of the bound DNA (Fig. 14C), which is in agreement with our earlier work on in vitro specificity of WhiB4 interaction with the ahpCD locus [20]. As expected, a highly GC-rich motif was identified by MEME within these WhiB4 binding peaks (Fig. 14D, E-value: - 6.4 × 10-103). For two independent WhiB4 samples, approximately 700 regions of enriched signals with a remarkable (> 90%) overlap between the binding regions were obtained (Supplementary information Table S4). Approximately, 7–10% of the Mtb genome encodes PE and PPE proteins (58, 59). Of the 90 PE proteins, 63 were encoded by GC-rich genes belonging to the PE-PGRS sub-class. Expectedly, WhiB4 bound overwhelmingly to PE-PGRS genes (55 of 63), but marginal to PE/PPE genes that do not belong to the PGRS sub-class (Fig. 15). We confirmed that detection of peaks at the PE-PGRS loci is not an artifact of ambiguous read mapping due to repeat elements present in the PGRS regions. Our peak calling analysis using only reads that can be unequivocally aligned to a single locus on the genome also recovers peaks at the PE-PGRS genes. This result is consistent with the observed preference of WhiB4 for GC rich regions (Fig. 15 and Supplementary information Table S4) and WhiB4-mediated regulation of PE-PGRS expression in Mycobacterium marinum [69]. Additionally, genes involved in respiration (e.g., cydAB), redox homeostasis (mshA/B/D, trxB2, nadA/B/C), and Fe/Fe-S hemostasis (sufR, sufB, bfrB, mbtB) were also bound by WhiB4 (Supplementary information Table S4). Interestingly, WhiB4 binds to regions encoding several transcription factors, sigma factors, and NAPs (lsr2, espR, IHF) (Supplementary information Table S4). To validate WhiB4 binding, we performed ChIP followed by quantitative PCR (ChIP-qPCR) on 9 target DNA sequences (5 within intergenic regions, 3 within ORFs and 1 non-peak region). Consistent with the ChIP data, all of the selected WhiB4-binding regions showed enrichment (Supplementary information Table S5). As a negative control, we over-expressed the FLAG-tagged WhiB4-Cys3 variant and performed ChIP-qPCR of the WhiB4 targets on immuno-precipitated genomic DNA as described earlier. ChIP-qPCR of WhiB4-specific regions showed no enrichment in the case of WhiB4-Cys3 as compared to corresponding input control (Supplementary information Table S5). Thus, the results indicated that all peaks were likely to be genuine WhiB4-targets.

Finally, we investigated whether genes bound by WhiB4 showed differential expression in Mtb during CHP stress. By comparing microarray of CHP-treated wt Mtb with the ChIP-seq data, we found that expression was marginally lower for genes bound by WhiB4 than those that are not (p value < 2.2 × 10-16; Supplementary information Fig. S9 and Table S4). Next, we compared ChIP-seq data with the microarray data for CHP-treated Mtb∆whiB4 (over CHP-treated Mtb). Approximately 30% of genes differentially expressed in CHP-treated Mtb∆whiB4 were bound by WhiB4, indicating both a direct and an indirect influence of WhiB4 on gene expression (Supplementary information Fig. S9 and Table S4). Similar to our findings, studies on FIS and CRP in E. coli have found little association between FIS/CRP DNA binding and differential gene expression in ∆fis/∆crp strains [70], [38]. It was proposed that the primary function of E. coli CRP and FIS proteins is to sculpt the chromosome, while their role in regulating transcription is likely to be circumstantial [70], [38]. A similar explanation may be suited for WhiB4 as well.

Discussion

In this study, we identified a link between DNA condensation, WhiB4, and oxidative stress response in Mtb. WhiB proteins are proposed to be Fe-S cluster-containing transcription factors; however, the exact molecular mechanism of action remained poorly understood. Studies revealed that some of the WhiB family members might regulate gene expression by binding to promoter sequences [71], [20], [27]. Furthermore, ChIP-Seq analysis of WhiB in Streptomyces coelicolor (Sco) seems to indicate DNA binding wherein a specific regulatory effect may be achieved by its association with a transcription factor WhiA [72]. Here, we confirmed that WhiB4 condenses DNA by binding more uniformly to the Mtb chromosome, with a notable preference for GC-rich DNA. Over-expression of WhiB4 induced hyper condensation of nucleoids in vivo, indicating that the protein, in sufficient quantity, can function independently in regulating DNA compaction. Notably, over-expression of NAPs (e.g. Dps) does not always result in nucleoid condensation owing to the negative influence of other NAPs (e.g. Fis) [43]. While overexpression of WhiB4 can condense nucleoid in vivo, how it contributes to DNA condensation in conjunction with other mycobacterial NAPs remains to be determined. Moreover, a large majority of WhiB4-binding events are inconsequential from the transcriptional perspective, suggesting that WhiB4-binding does not necessarily affect transcription locally, but may serve as focal points to organize the genome into domains and thereby influence transcription indirectly. In line with this, a genome-wide DNA binding study suggested that ~ 30% of Mtb transcription factors binds DNA with a much smaller impact on transcription [73]. Association of WhiB4 with genomic regions encoding Lsr2, EspR, and IHF suggest the possibility of an interplay between different NAPs to alter chromosome structure and organization, thereby influencing patterns of gene expression in response to oxidative stress. Consistent with this, CHP-treated Mtb∆whiB4 displayed nucleoid compaction (albiet slower), indicating that WhiB4 may collaborate with other NAPs (e.g., Lsr2, EspR, and IHF) under oxidatively hostile conditions.

Several NAPs are known to play important role in influencing nucleoid condensation under specific stress conditions such as iron starvation, hypoxia, and acidic pH [10]. However, the importance of DNA condensation and NAPs in controlling oxidative stress response remains controversial. For example, while S. aureus resists H2O2 by rapidly condensing DNA through a Dps homolog (MrgA) [13], oxidative stress simply did not lead to DNA condensation in other bacteria including Sco, Dickeya dadantii, and E. coli [74], [43], [75]. Interestingly, a recent study elegantly showed that MrgA-mediated protection of S. aureus from oxidative stress is mainly due to its ferroxidase activity rather than its ability to shield DNA [76]. However, the role of Lsr2 in protecting Mtb from oxidative damage by shielding DNA is controversial [11], [77]. Therefore, it is unclear as of now how mycobacteria remodel their DNA condensation during oxidative stress. To begin understanding this, we demonstrated that the loss of WhiB4 delays DNA condensation, maintains redox homeostasis, and protects the cells from oxidative stress, whereas its over-expression reversed these phenotypes. Our data indicate that the detrimental effect of WhiB4 overexpression is likely to be due to the repression of antioxidant machinery and the diminished capacity to counteract cytoplasmic redox stress. It is also likely that hyper-condensation of the nucleoid inhibits other metabolic processes such as DNA replication/repair, transcription, and translation to exert efficient killing. We conclude that resistance to oxidative stress in Mtb is unlikely to be mediated by nucleoid hypercompaction. In this context, it has been suggested that certain bacteria (e.g., gamma-proteobacteria) have evolved genetic mechanisms (e.g. Fis, TopA, and GyrA) to block DNA condensation and promote the expression of OxyR-dependent antioxidant genes as a major mechanism to guard genomic DNA against oxidative stress [43]. Moreover, similar to Mtb, a higher degree of oxidative stress and killing were associated with nucleoid hypercondensation in E. coli [78], [16]. Although, unlike Mtb where WhiB4 overexpression is sufficient to induce both DNA condensation and killing under oxidative stress, a similar consequence of oxidative challenge seems to be mediated through a combined action of multiple OxyR-regulated NAPs (e.g, H-NS, Hup, Him, MukB, and Dps) in E.coli [16]. In the absence of OxyR, Fis and Dps activities in Mtb, WhiB4, with its redox-active cysteines can be an important regulator of both nucleoid condensation and expression of oxidative stress responsive genes in Mtb.

We have previously shown that the 4Fe−4S cluster of WhiB4 is exceptionally sensitive to oxygen in vitro [20]. Consistent with this, we showed the presence of a minor fraction of disulfide-linked apo-WhiB4 oligomer under aerobic growing condition. This indicates that despite the reduced environment of mycobacterial cytosol, WhiB4 Fe-S cluster can be targeted during aerobic metabolism to generate apo-WhiB4. Since Fe-S clusters are acutely sensitive to superoxide generated during aerobic respiration [79], it is likely that WhiB4 Fe-S cluster is oxidatively damaged by superoxide during aerobic growth. However, this requires future experimentation. Further oxidation of WhiB4 thiols would require two-electron oxidation by free radicals as triggered upon exposure to CHP. This form of oxidized apo-WhiB4 oligomer is involved in nucleoid condensation and gene expression in Mtb. Interestingly, the four redox-active cysteine (Cys) residues involved in Fe-S cluster coordination in WhiB4 are surrounded by several basic amino acids (Lys36–Cys37–Arg38, Cys59–Arg60), which may decrease Cys residues pKa, resulting in the formation of disulfide bonds in the cytosol of mycobacteria. We propose that under aerobic growing conditions, oxidation of holo-WhiB4 to oxidized apo-WhiB4 activates DNA binding and repressor function to preclude the unnecessary expression of oxidative stress responsive pathways. A gradual increase in oxidative stress on the one hand increases generation of oxidized apo-WhiB4 oligomers while on the other hand down-regulates WhiB4 expression. Both of these activities ensure levels of intracellular oxidized apo-WhiB4 oligomers required to induce appropriate degree of nucleoid condensation and calibrated activation of antioxidants/stress pathways. The steady increase in antioxidants such as MSH and Fe-S biogenesis machinery (Suf operon) in response to sustained oxidative stress can create a regulatory feedback loop by reducing oxidized thiols of apo-WhiB4 to regenerate monomeric reduced apo-WhiB4 and/or holo-WhiB4. This would allow WhiB4 to loose chromosomal binding and completely derepress antioxidants, maintaining both redox and topological balance (Fig. 16). Besides this, other factors (e.g., NAPs, sigma factors etc.) which directly or indirectly cooperate with WhiB4 could play a role in maintaining topological homeostasis and adaptation to oxidative stress. The fact that Mtb∆whiB4 showed higher resistance to oxidative stress and whiB4-OE displayed hypersensitivity as compared to wt Mtb indicates that a gradual decrease in WhiB4 expression upon oxidative stress is a cellular decision to induce the oxidative stress response and reduce DNA hypercondensation.

Fig. 16

Proposed model of WhiB4 function in oxidative stress response. Under mild oxidative conditions, a minor fraction of oxidized apo-WhiB4 generated via oxidation of holo-WhiB4 activates DNA binding and repressor function to preclude unnecessary overexpression of oxidative stress responsive pathways. Further elevation in oxidative stress levels result in higher thiol-linked apo-WhiB4 oligomers, which would have further repressed antioxidants expression and induced genome hypercondensation. To circumvent this, Mtb downregulates expression of WhiB4 (via autorepression) to control the levels of oxidized apo-WhiB4, which ensures appropriated degree of nucleoid condensation and antioxidants expression. A gradual increase in antioxidants such as MSH and the Fe-S repair pathways (Suf operon) can create a regulatory feedback loop, which reduces oxidized thiols of apo-WhiB4 to eventually regenerate monomeric reduced apo-WhiB4/holo-WhiB4. All of this ultimately results in complete derepression of antioxidants to mitigate oxidative stress and to also restore topological homeostasis. Ectopic over-expression of WhiB4 by Atc leads to sustained accumulation of oxidized apo-WhiB4 without any feedback control. This eventually represses antioxidants/stress pathways and induces long lasting nucleoid hypercondensation. Both of these activities of WhiB4 adversely affect mycobacterial ability to tolerate oxidative stress leading to enhanced susceptibility of whiB4-OE strain towards CHP.

Conclusion

In conclusion, the combined results indicate that WhiB4 is a candidate for a dual-function NAP that could integrate environmental signals with DNA conformation and transcription. The focus of the present investigations has been mainly on the nucleoid condensing function of WhiB4 and how it can modulate oxidative stress response in Mtb. As far as we are aware, this is the first example of a redox-dependent NAP in bacteria.