Contents lists available at ScienceDirect

Tuberculosis

journal homepage: www.elsevier.com/locate/tube

Molecular Aspects

Distinct properties of a hypoxia specific paralog of single stranded DNAbinding (SSB) protein in mycobacteria

Amandeep Singha, M. Vijayana,∗∗, Umesh Varshneyb,c,∗

a Molecular Biophysics Unit, Indian Institute of Science, Bangalore 560012, Indiab Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore 560012, Indiac Jawaharlal Nehru Centre for Advamced Scientific Research, Jakkur, Bangalore 560064, India

A R T I C L E I N F O

Keywords:DNA repairRecombination repairHypoxiaRecASSBbStress response

A B S T R A C T

In addition to the canonical Single Stranded DNA Binding (SSBa) protein, many bacterial species, includingmycobacteria, have a paralogous SSBb. The SSBb proteins have not been well characterized. While in B. subtilis,SSBb has been shown to be involved in genetic recombination; in S. coelicolor it mediates chromosomal segre-gation during sporulation. Sequence analysis of SSBs from mycobacterial species suggests low conservation ofSSBb proteins, as compared to the conservation of SSBa proteins. Like most bacterial SSB proteins, M. smegmatisSSBb (MsSSBb) forms a stable tetramer. However, solution studies indicate that MsSSBb is less stable to thermaland chemical denaturation than MsSSBa. Also, in contrast to the 5–20 fold differences in DNA binding affinitybetween paralogous SSBs in other organisms, MsSSBb is only about two-fold poorer in its DNA binding affinitythan MsSSBa. The expression levels of ssbB gene increased during UV and hypoxic stresses, while the levels ofssbA expression declined. A direct physical interaction of MsSSBb and RecA, mediated by the C-terminal tail ofMsSSBb, was also established. The results obtained in this study indicate a role of MsSSBb in recombinationrepair during stress.

1. Introduction

Mycobacterium tuberculosis is known to modulate its gene expressionto adapt to the stress conditions encountered in the host cells and it mayremain dormant in a non-replicating persistent (NRP) state, for decadesin the host [1]. M. tuberculosis NRP is characterized by slow metabolismupon encountering hypoxia, starvation or acidic pH in the host cells[2–4]. DNA damages such as oxidation of nucleotides and DNA breaks,by host generated reactive oxygen species and reactive nitrogen inter-mediates threaten the genomic integrity during this dormant phase,necessitating a robust DNA repair in the early stages of exit from dor-mancy [5,6]. Recombination repair forms a major part of this repairsystem. Upon DNA break, RecA binds to the single stranded DNA(ssDNA) and stimulates autocatalytic cleavage of LexA, which regulatesgenes for damage repair and tolerance [7,8].

Single Stranded DNA Binding (SSB) proteins are highly conservedproteins implicated in protection of ssDNA from nucleases, and preventformation of secondary structures during major DNA transactions[9,10]. During DNA break repair, SSBs play a crucial role in nucleo-protein filament formation and RecA loading on DNA [11–13]. Mostbacterial SSBs are homo-tetramers, where tetramerization is mediated

by the conserved N-terminal oligo-nucleotide binding (OB) domainfollowed by a less conserved, and highly disordered C-terminal domainor tail. The C-terminal tail of SSBs is known to interact with many otherproteins involved in DNA repair, replication and recombination[14–18]. It has also been reported to have a role in modulating the DNAbinding affinity [19,20].

Apart from the well-studied SSB (SSBa), several naturally competentbacteria possess a second SSB paralog (referred to as SSBb). SSBhomologs are often encoded by nearly all conjugative plasmids found inbacteria. While SSBa (homologous to E. coli SSB) is attributed to per-form all the canonical functions, the biological functions of the para-logous SSBb proteins are still ambiguous. In Bacillus subtilis, SSBb(BsSSBb) shares the load of genetic recombination with SSBa and aids innatural-competence associated recombination [21]. In Streptococcuspneumoniae, SSBb (SpSSBb) protects the naturally internalized ssDNA toimprove the likelihood of crossover events [22]. Interestingly, deletionof ssbB gene in Streptomyces coelicolor leads to irregular chromosomalsegregation during sporulation [23]. The members of genus myco-bacteria, except for M. leprae, are also found to possess ssbB genes [24].

The interaction of DNA with the OB domain has been studied indetail in E. coli and is known to be sequence independent. However,

http://dx.doi.org/10.1016/j.tube.2017.10.002Received 12 July 2017; Received in revised form 18 September 2017; Accepted 1 October 2017

∗ Corresponding author. Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore 560012, Karnataka, India.∗∗ Corresponding author.E-mail address: varshney@iisc.ac.in (U. Varshney).

Tuberculosis 108 (2018) 16–25

1472-9792/ © 2017 Elsevier Ltd. All rights reserved.

T

depending on the concentration of the protein and ions in the solution,SSBs interact with DNA in either SSB35 (in which 35 nucleotides wraparound two subunits of the tetramer), or SSB56/65 (in which 56/65nucleotides wrap around all four subunits of the tetramer) modes[25,26]. The DNA binding affinities of the paralogous SSBs were foundto be different. The SpSSBb and ScSSBb proteins were reported to havehigher DNA binding affinity compared to their SSBa counterparts; al-though in the case of B. subtilis proteins the converse was observed[21,22].

On account of slow growth of M. tuberculosis and requirement ofcontained laboratory conditions, M. smegmatis, a fast growing myco-bacterium, has often been used as a surrogate to study fundamentalmolecular processes in mycobacteria [27]. Sequence analysis of M. tu-berculosis and M. smegmatis genomes suggests that the latter retainsmost of the genes involved in adaptation to hypoxia [28,29]. SSBa hasbeen characterized from multiple mycobacterial species [30–33]. Thethree-dimensional structure of M. smegmatis SSBb (MsSSBb), solvedrecently in our lab, revealed a classical homo-tetrameric structure [34].The quaternary structure of MsSSBb is highly similar to that observedfor mycobacterial SSBa and ScSSBa structures, retaining the 'clamp' likeinter-subunit strand exchange [35,36]. To address the question of whythere are two SSBs in mycobacteria, we have now studied DNA bindingproperties of MsSSBb. We also performed the same experiments onMsSSBa, for comparison of properties. Transcriptional regulation of thetwo ssb genes was tested and a possible role of SSBb in stress responsehas been established.

2. Materials and methods

2.1. Sequence analysis and DNA

The sequences of SSB proteins were obtained from Uniprot [37].Multiple sequence alignment (MSA) of proteins was generated usingthree combined iterations, while keeping others settings as default, inClustal Omega [38]. ESPript was used to generate the MSA [39]. M13phage ssDNA was obtained from New England Biolabs (NEB). DNAoligomers, including 6FAM-tagged (fluorescein) were obtained fromSigma-Aldrich.

2.2. Cloning of SSBb open reading frames

Cloning of the gene encoding MsSSBb (MSMEG_4701) has beendescribed previously [34]. The gene encoding MsSSBa (MSMEG_6896)was also cloned using a similar strategy. The details of the primers usedare provided in Table S1. SSBbΔ7 and SSBbΔ27 constructs, with 7 and27 residue truncations respectively, from the C-terminal tail of MsSSBbwere also made in pET14b vector. MsRecA (MSMEG_2723) was clonedin pET His6 Sumo TEV LIC cloning vector (1S) (Addgene) using man-ufacturer's protocol. Cloning was confirmed by primer-based sequen-cing (Xcleris genomics).

2.3. Protein expression and purification

The expression constructs were transformed into E. coli BL21 (DE3)(Novagen Inc.). All SSBs, including the shorter constructs were purifiedas described previously [34]. For purification of RecA, cultures in Luria-Bertani (LB) medium supplemented with ampicillin (100 μg/ml) weregrown for 8 h after induction, at 0.5 OD600, with 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 30 °C. The cells were harvested bycentrifugation at 6000g for 15 min, re-suspended in buffer A (30 mMTris-acetate, pH 7.5, 300 mM NaCl, 10 mM imidazole and 10% (v/v)glycerol) and lysed by sonication. The lysate was then centrifuged at15,000 g for 60 min. The supernatant was loaded onto Ni-NTA column(GE Healthcare) equilibrated with buffer A, washed with the samebuffer supplemented with 20 mM imidazole and eluted with a lineargradient of imidazole (30 mM–500 mM). The protein was dialyzed

against 50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 0.5 mM Na2EDTA and10% (v/v) glycerol for 12 h. The His-SUMO tag was cleaved using TEVprotease (Sigma-Aldrich) for 6 h at 25 °C, in the same buffer. Theprotein was further purified by passing through Ni-NTA beads followedby size-exclusion chromatography using High load 16/600 Superdex200 column (GE Healthcare) in buffer B (30 mM Tris-HCl, pH 7.5,250 mM NaCl and 10% (v/v) glycerol). The purity was established bySDS-PAGE [Fig. S1]. M. smegmatis Ung (Uracil DNA glycosylase FamilyI) was purified as described previously [40].

2.4. Electrophoretic mobility shift assay (EMSA)

Standard reaction mixtures containing 200 ng M13 phage DNA or2 μM fluorescein tagged DNA were incubated with increasing con-centrations of SSBs (MsSSBa or MsSSBb) at 4 °C for 30 min. Differentreaction buffers used are specified in the figure legends. Reaction wasmixed with 4 μl loading dye [containing 0.12% (w/v) each of bromo-phenol blue and xylene cyanol FF in 20% glycerol], and resolved on 1%agarose gel, containing ethidium bromide, or 4–8% native PAGE in 1XTBE buffer. Agarose gels were visualized under UV light. The PAGE gelswere visualized using fluorescein filter in the BIORAD gel doc system.

Interaction with Ung: Standard reaction mixtures (20 μl) containing200 ng M13 ssDNA in 20 mM Tris-HCl, pH 8.0 and 50 mM NaCl wereincubated with 500 nM MsSSBa or MsSSBb and increasing concentra-tions (500 nM–4 μM) of MsUng at 4 °C for 30 min. The reaction mix-tures were mixed with 4 μl of loading dye and individual samples wereloaded and resolved on a 0.8% agarose gel, containing ethidium bro-mide, in 1X TBE buffer.

2.5. Fluorescence measurements

Fluorescence measurements were made using a FP-6300 spectro-meter from JASCO analytical instruments. A 10 μM solution of MsSSBaor MsSSBb was titrated against increasing concentrations of 35-merDNA (poly-dT) and 76-mer DNA (poly-dT) and the quenching offluorescence was recorded. Titrations were carried out in differentbuffers, as specified in figure legend and Results section. An excitationwavelength of 295 nm was used and emission spectra at 343 nm wererecorded. After each titration, the solution was allowed to equilibratefor 1 min before the fluorescence was measured. Experiments weresetup in duplicates and an average of 3 readings for each data pointwere used to plot the graph.

2.6. Thermal melt assay

Protein unfolding as a function of temperature was monitored usingfluorophore SYPRO Orange (Sigma, S5692) in iQ5, BioRad iCyclerMulticolor Real-Time PCR detection system. Increase in fluorescence ofdye upon binding to the hydrophobic regions, upon thermal dena-turation, was measured. Protein samples (10 μM) were mixed withbuffer (20 mM Tris-HCl, pH 7.5 and 200 mM NaCl) and 1x SYPROorange dye. The readings were taken in 96-well PCR microplates(BioRad) in the RT-PCR device. Samples were heated at 0.5 °C per min,ranging between 10 °C and 95 °C and the fluorescence intensity wasmeasured using Cy5 filter with red-orange color intensity detection.

2.7. Chemical melt assay

A decrease in intrinsic tryptophan fluorescence was recorded as afunction of chemical denaturation by guanidine hydrochloride (GndCl).Protein samples (10 μM) were mixed with increasing concentrations ofGndCl (0.0, 0.2, 0.4, 0.6, 0.8, 1.0, 1.2, 1.4, 1.6, 1.8, 2.0, 2.4, 2.8, 3.2,3.6, 4.0, 4.5, 5.0, 5.5 and 6 M) in 10 mM Tris-HCl, pH 7.5 and incubatedfor 30 min at 37 °C.

A. Singh et al. Tuberculosis 108 (2018) 16–25

17

2.8. RT-PCR

M. smegmatis mc2155 was grown to 0.6 OD600 and exposed to UVradiation (150 J/m2) for 5 min. The cultures were allowed to growfurther for 12 h before harvesting cells. Hypoxic cultures were estab-lished in duplicates [41]. Cells from one flask were allowed to recoverunder aerobic growth for 90 min before harvesting. Cell pellets fromuntreated and treated cultures were re-suspended in 2 ml buffer(20 mM Tris-HCl, pH 7.4, 0.45 M sucrose, 8 mM Na2EDTA) with 4 mg/ml lysozyme and incubated for 45 min at 37 °C. The suspension wassubjected to centrifugation (3000 g) for 10 min. RNA was isolated usingRNeasy Mini Kit (Qiagen). No-RT (reverse transcription) control PCR,for ssbA and ssbB genes were performed to check for any DNA con-tamination. Maxima First strand cDNA synthesis kit (Thermo Scientific)was used to synthesize cDNA with 2 μg of total RNA template. Real timePCR assays (20 μl) were conducted in triplicates using 200 ng of cDNAtemplate, 10 μl of GeneSureTM SYBR Green qPCR Master Mix (2X)(Puregene), 1 μl of forward and reverse primers (5 μM stock) in iQ5,BioRad iCycler Multicolor Real-Time PCR detection system. Details ofprimers used are provided in Table S2.

2.9. Far-western blotting

Increasing amounts of purified RecA in buffer A (20 mM Tris-HCl,pH 7.4, 100 mM NaCl and 5 mM MgCl2) were spotted on nitrocellulosemembranes. Bovine serum albumin (BSA) was used as negative control.The membrane was blocked for 2 h at room temperature using buffer Asupplemented with 5% (w/v) skimmed milk. The membranes were thenincubated with a 2 μM MsSSBb solution in buffer A for 8 h at 4 °C,followed by 5 washing steps of 10 min each. Following this, membraneswere incubated with HRP conjugated anti-His antibody, as the His-tagon SSBb was retained after purification, for 1 h at room temperatureand developed using 20 μl of 30% H2O2 and 1 mg/ml of DAB in100 mM citrate phosphate buffer pH 5.5 (10 ml).

3. Results

3.1. Sequence analysis: SSBa is more conserved than SSBb

ssbB genes are typically distantly located from the ssbA genes inmost mycobacterial genomes. According to ProOPDB, ssbB gene is en-coded within an operon containing a putative ABC transporter ATPbinding protein [42]. While the SSBa sequences are highly conserved inmycobacteria (sequence identity of 86% between M. tuberculosisMtSSBa and MsSSBa), SSBb proteins are relatively less conserved (se-quence identity of 54% between MtSSBb and MsSSBb). The paralogousSSBs from M. smegmatis share an overall sequence identity of 37% and a52% similarity. Both the MsSSBa and MtSSBa harbor eleven and fourglycine residues, respectively at their C-termini. However, SSBb pro-teins possess an alanine rich region. While the MsSSBb harbors twoalanine residues, MtSSBb possesses ten [Fig. 1]. The computed iso-electric points of M. smegmatis SSBs are between 4.5 and 5.5.

The mycobacterial SSBa and SSBb are of comparable lengths, 165and 170 amino acids, respectively in M. smegmatis. However, the C-terminal tails of SSBb proteins characterized so far, are shorter thanthose of the respective SSBa proteins [Table 1]. In SSBa proteins, theconserved acidic C-terminal D-D-D-(I/P)-P-F motif, where Asp (D) canbe replaced with Glu (E), was found to be critical for mediating inter-actions with other proteins [15]. A similar acidic tail motif is present inSpSSBb and NgSSBb. This motif has also been shown to be crucial forssDNA processing in S. pneumoniae [22]. No such motif is present in anyof the ScSSBb, BsSSBb or the mycobacterial SSBb proteins [Fig. 1].However, the C-terminal stretch of mycobacterial SSBb proteins ishighly acidic in nature. While the MsSSBb harbors 18 acidic aminoacids, MtSBBb possesses 6 acidic amino acids in their C-terminal do-mains. Such unusual differences further fueled the need to study these

proteins. In EcSSB, W40, W54, F60 and W88 are the base stacking re-sidues. The relative positions of corresponding residues, R39, Y53, W59and Y87 in MsSSBb remain similar. With the exception of Arg, the otherthree amino acids are aromatic in nature and may be important inbinding of MsSSBb with DNA. The amino acid residues corroborate tothe base stacking residues identified in BsSSBb as well [21].

3.2. Solution properties of mycobacterial SSB proteins

The SSBa, SSBb and RecA proteins from M. smegmatis were over-produced in E. coli and purified to apparent homogeneity [Fig. S1].MsSSBb, like most bacterial SSBs, forms a stable tetramer in the solution[34]. Chemical denaturation assays employing fluorescence spectro-scopy were used to investigate differences in MsSSBa and MsSSBb.Quenching of intrinsic fluorescence was recorded with respect to theincreasing concentrations of guanidine hydrochloride (GndCl). About95% fluorescence quenching was observed at 2.4 M GndCl for MsSSBa,and the corresponding concentration for MsSSBb was 1.6–1.8 M GndCl[Fig. 2A]. Also, in thermal denaturation, while the MsSSBa revealed aTm of about 65 °C, the corresponding value for MsSSBb was 56 °C[Fig. 2B]. Thus, both the assays suggest higher structural stability ofMsSSBa than the paralogous SSBb.

3.3. Binding of MsSSBb to DNA is modulated by salt and Mg2+ ionconcentration

The ssDNA binding activity of MsSSBa has been examined pre-viously [13,43]. To examine the binding properties of MsSSBb, theprotein was incubated with the naturally occurring M13 ssDNA andanalyzed by agarose gel electrophoresis. As observed in Fig. 3A, themobility of M13 ssDNA decreased progressively upon increasing con-centration of the protein. Also, a gradual quenching of ethidium bro-mide fluorescence was observed, probably owing to its occlusion due todenaturation of the secondary structures in M13 ssDNA.

The DNA binding properties of previously characterized SSBs arereported to be modulated by the presence of salts and divalent cationsin the solution [44,45]. We incubated M13 ssDNA with increasingconcentration of MgCl2 and Na2EDTA separately and resolved thecomplexes on agarose gel. As seen from Fig. 3B, MsSSBb remainedbound to M13 ssDNA even in the presence of 10 mM Na2EDTA in-dicating that divalent ions are not necessary for DNA binding byMsSSBb. However, for the same amount of M13 ssDNA andMsSSBb, themobility of the complex decreased on increasing the MgCl2 con-centration.

3.4. MsSSBa has a higher DNA binding affinity than that of MsSSBb

To further explore the DNA binding properties of MsSSBb, fluores-cence spectroscopy was employed. When bound to ssDNA, a gradualdecrease in the intrinsic tryptophan fluorescence of SSBs has been ob-served and values obtained from titration plots have been used tocharacterize binding affinities and occluded binding site of SSBs [46].Upon addition of saturating quantities of dT35 and dT76 (poly-dT,subscript denotes the length), intrinsic fluorescence was quenched byapproximately 95 (± 1)%. Fluorescence was measured under 3 dif-ferent buffer conditions, namely, H10 (10 mM HEPES, pH 7.5), H10N200

(10 mM HEPES, pH 7.5 and 200 mM NaCl) and H10Mg5 (10 mM HEPES,pH 7.5 and 5 mM MgCl2). The quench was observed to be salt in-dependent. However, a consistent change in the intrinsic fluorescenceand decrease in quench to 90 (± 2)% was observed on addition ofMgCl2 in the solution. DNA binding constants of 5.97 (± 0.45) μM and2.81 (± 0.21) μMwere observed for dT35 and dT76 respectively, in H10.The presence of Mg2+ or NaCl did not significantly change the bindingaffinities, except in the case of dT35 where the binding affinity wasslightly affected in the presence of 200 mM NaCl. However, the bindingmodes of SSBs were affected substantially (discussed below). In the

A. Singh et al. Tuberculosis 108 (2018) 16–25

18

current range of buffers, the occluded binding site was estimated to be32 (± 2) nucleotides, similar to highly cooperative EcSSB35 bindingmode [Fig. 4A and B].

The binding of dT35 and dT76 with MsSSBa and MsSSBb in Tris-HClbuffer (HEPES buffer was not suitable for MsSSBa) was compared usingfluorescence spectroscopy. DNA binding constants of 4.01 (± 0.38)and 1.43 (± 0.12) μM were observed for binding of MsSSBa with dT35

and dT76 respectively [Fig. 4C and D]. For MsSSBb, these values were7.36 (± 0.59) and 2.76 (± 0.30) μM; suggesting that MsSSBb hasabout two fold lower DNA binding affinity relative to that ofMsSSBa forboth the DNA substrates. The property of DNA binding is regulated bythe C-terminal tail of SSBa. Thus, we carried out titration experimentswith SSBbΔ7 and SSBbΔ27, C-terminal truncated MsSSBb. Each trun-cation had a higher affinity for both the DNA substrates, suggesting therole of tail, in regulating DNA binding [Fig. S2].

Because MsSSBa and MsSSBb bound to DNA with different bindingaffinities, an additional experiment was carried out to observe compe-titive binding of these proteins. We incubated M13 phage ssDNA withsaturating amounts of either MsSSBa or MsSSBb individually, or si-multaneously in the same tube. Slow moving complexes were observedin the presence of MsSSBa, alone or in combination with MsSSBb, as

Fig. 1. Multiple sequence alignment of characterized SSBb sequences. Sequences of SSBb from six organisms have been aligned with EcSSBa. Most of the similarity is observed in the N-terminal OB fold domain of SSBs. The alanine residues are indicated in grey. Triangles represent the base stacking residues, as observed in the EcSSBa structure. The dotted box outlinesthe acidic C-terminal motif observed on the SSBs.

Table 1Properties of paralogous SSBs characterized in all organisms.

SSB Length offullprotein(aa)

Length of C-terminalregion (aa)

MW ofmonomer(kDa)

IdentitywithEcSSBa (%)

Sequence oflast 6 aminoacids

MsSSBa 165 57 17.4 33.78 DDEPPFMsSSBb 170 62 18.5 20.51 GLPLTAMtSSBa 164 57 17.3 32.65 DDEPPFMtSSBb 161 62 16.7 19.05 PLPISAScSSBa 199 89a 19.9 33.13 SDEPPFScSSBb 156 50a 16.8 23.08 DPVPVGBsSSBa 172 66a 18.7 35.85 DDDLPFBsSSBb 113 7a 12.5 34.58 REKAADSpSSBa 156 53 17.4 35.42 DDDLPFSpSSBb 131 29 14.9 34.15 EEELPFNgSSBa 174 69 19.3 47.90 DDDIPFNgSSBb 143 35 16.4 24.03 DDDIPL

SSB_OBF boundaries were generated using CDD. The length of the C-terminal region isdefined as the total number of residues after the OB fold. For MtSSBa, MsSSBa, MtSSBband ScSSBa, the residues forming the extra hook found in the structure are not included.

a The values have been taken as specified in the publications [21,23].

A. Singh et al. Tuberculosis 108 (2018) 16–25

19

compared to MsSSBb alone. The resulting complexes were resolved onan agarose gel, and excised to further analyze by SDS-PAGE.Irrespective of the ratio of MsSSBb to MsSSBa in the reaction, theamounts of MsSSBa bound to DNA was found to be higher than MsSSBb[Fig. 5]. However, even at relatively higher MsSSBa concentration,considerable amounts of MsSSBb were always observed in gel, sug-gesting that both proteins can simultaneously bind to the DNA.

3.5. Differential DNA binding modes of paralogous SSBs

The purified M. smegmatis SSB proteins were further analyzed fordifferential ssDNA binding properties. A fixed concentration of 5′fluorescein tagged dTn [n = 12, 15, 18, 21, 25, 35, 76 and 120 nu-cleotides in length] were incubated with increasing concentration ofeither MsSSBa or MsSSBb and analyzed on native PAGE. While a smearwas observed for dT12, sharp bands appeared for dT15 or higher sizedDNA oligomers in case of MsSSBa [Fig. 6A]. However, sharp bandsappeared for size ≥ dT18, in case of MsSSBb [Fig. 6B], suggesting SSBacan bind to shorter lengths of DNA than SSBb.

To study the binding modes, increasing concentrations of proteinwere incubated with labeled dT120 in different buffers and resolved on4% native PAGE. The concentration of MsSSBa was kept at half theconcentration of MsSSBb, owing to its higher binding affinity observedduring previous experiments. Three distinct complexes, SSB1, SSB2 andSSB3 could be observed in the gel owing to one, two or three molecules

(tetramers) of SSB bound to DNA, respectively [Fig. 7]. At high proteinto DNA ratio, SSB3 complex was predominant irrespective of the bufferconditions, suggesting imposition of SSB35 DNA binding mode. Atmoderate protein concentrations, presence of high salt or Mg2+ favoredSSB65 binding mode for both SSBa and SSBb. For instance, in MsSSBb[Fig. 7B], a SSB3 complex is formed in lane 3, while at equal con-centration of protein, an SSB2 complex is observed in the presence ofboth Mg2+ (lane 8) or high salt (lane 13). Similar observations weremade for MsSSBa [Fig. 7A, lanes 4, 9 and 14]. This suggests that at leastone SSB tetramer is bound in SSB65 binding mode such that only twotetramers of SSB can be accommodated on 120 nucleotides. However,at very low protein to DNA ratio, both SSB65 and SSB35 binding modeswere observed for MsSSBb [Fig. 7B, lanes 2, 7 and 12], while MsSSBanotably favored SSB65 binding mode [Fig. 7A lanes 3, 8 and 13].

3.6. Insights into the cellular role of SSBb

To assess the involvement of SSBb during hypoxia and DNA da-maging conditions, we examined the ssbA and ssbB gene expressionlevels following UV treatment, and 14 days of hypoxic culturing of M.smegmatis. Re-aeration was allowed for 1.5 h in an aliquot of hypoxiccultures. Upregulation of hspX validated the establishment of hypoxicculture. The expression levels of ssbB gene increased by approximately2 and 7 fold in UV and hypoxic stress, while simultaneously the levels ofssbA gene expression declined [Fig. 8].

Fig. 2. Solution properties of mycobacterial SSB proteins. (A) Standard reaction mixtures (200 μl) containing 10 μM of MsSSBa or MsSSBb with increasing concentrations of GndClincubated at 37 °C for 30 min. Intrinsic fluorescence was recorded and normalized using the minimum value set as 0 and maximum value as 1. (B) Thermal melt curve of 10 μMMsSSBa orMsSSBb in 200 mM NaCl and 20 mM Tris-HCl, pH 7.5. Graphs were plotted using Graph Pad Prism 5 software.

Fig. 3. Electrophorectic mobility shift assays of MsSSBb. (A) Analysis of M13 ssDNA binding to MsSSBb. M13 ssDNA (250 ng) was incubated with MsSSBb (0, 10, 20, 40, 80, 160, 320,500, 1000, 3000 nmol, lanes 1–10, respectively) in a binding buffer (B) Increasing concentrations of MgCl2 (0, 0.5, 2, 5 and 10 mM, lanes 1–5, respectively) and Na2EDTA (1, 3, 5 and10 mM, lanes 6–9, respectively) were added to a fixed concentration of MsSSBb (1 μM), pre-incubated with 200 ng of M13 ssDNA. All reactions were electrophoresed on 1% agarose gel(containing EtBr) in 1X TBE. The gel was visualized under UV light.

A. Singh et al. Tuberculosis 108 (2018) 16–25

20

SSBa proteins are known to interact with a plethora of other pro-teins, including RecA [13]. Thus, we investigated the possibility of di-rect interaction of MsSSBb with MsRecA using far-western techniqueunder non-denaturing conditions [47]. Purified native RecA wasspotted in increasing concentrations on membranes followed byblocking, washing and incubation with MsSSBb. Membranes were thenwashed and probed with anti-His antibodies (for SSBb). We observed

binding between MsRecA and MsSSBb, indicating a direct physical in-teraction [Fig. 9A]. To observe the effect of C-terminal tail, we repeatedthe experiment with SSBbΔ7 and SSBbΔ27. While, SSBbΔ7 retained theinteraction, SSBbΔ27 was found not to show a detectable interactionwith RecA [Fig. 9B and C], indicating an important role of C-terminaltail in the interaction. MsSSBa is also known to affect uracil excision byMsUng (Uracil DNA glycosylase - family I) [43]. To investigate the in-teraction, we used an EMSA based assay. M13 ssDNA-MsSSBa or M13ssDNA-MsSSBb complexes were incubated with increasing concentra-tions of MsUng. The reaction mixture containing MsSSBa showed re-duced mobility suggesting interaction with MsUng. However, a similareffect on DNA-MsSSBb complexes was not observed, suggesting lack ofa detectable interaction between SSBb and MsUng [Fig. 9D].

4. Discussion

The work described in this report encompasses the biochemicalcharacterization of SSBb from M. smegmatis. In an attempt to under-stand mycobacterial SSBs better, we also purifiedMtSSBb. However, theprotein showed higher order oligomerization or aggregation as estab-lished by DLS analysis [Fig. S3]. Even in its aggregated form, MtSSBbwas able to bind M13 phage ssDNA [Fig. S4]. However, all furtheranalysis were carried out with MsSSBb. For comparison, similar ex-periments with MsSSBa were also carried out. Sequence analysis sug-gested that the MsSSBb possesses a canonical oligo-nucleotide bindingfold (OB-fold) similar to the previously characterized bacterial SSBs.

Fig. 4. Fluorescence studies on SSB-DNA binding. Titration of 10 μM MsSSBb with increasing concentrations of poly-dT35 (A), or poly-dT76 (B) in three different buffers. [dT35]nucleotidesor [dT76]nucleotides represent the concentration in terms of nucleotides. Ka35 and Ka76 represent the equilibrium DNA association/binding constant of poly-dT35 and poly-dT76 and werecalculated with [DNA]molecule vs. quenching plot (similar to shown below). (C) Comparison of fluorescence titration of poly-dT35 with 10 μM of MsSSBa and 10 μM of MsSSBb in T20N200.(D) Comparison of fluorescence titration of poly-dT76 with 10 μM of MsSSBa and 10 μM of MsSSBb T20N200. The buffer composition are as follows: H10 (10 mM HEPES pH 7.5), H10N200

(10 mM HEPES pH 7.5 and 200 mM NaCl), H10Mg5 (10 mM HEPES pH 7.5 and 5 mM MgCl2) and T20N200 (20 mM Tris-HCl, pH 7.5 and 200 mM NaCl).

Fig. 5. Binding of MsSSBa and MsSSBb to M13 ssDNA. M13 ssDNA (150 ng) was in-cubated with saturating amounts of either MsSSBa or MsSSBb alone, or in combination.The total amount of protein in the reaction was kept constant, while changing the ratio ofproteins. After 15 min incubation at 37 °C, the complexes were resolved on a 0.8%agarose gel. The complexes were then excised from agarose gel and boiled in SDS-loadingdye at 96 °C, until the agarose melted. The samples were then analyzed on a 14% SDS-PAGE gel, and visualized with Coomassie blue staining.

A. Singh et al. Tuberculosis 108 (2018) 16–25

21

Fig. 6. Binding of different size DNAs to MsSSBa and MsSSBb. Standard reaction mixtures (10 μl) containing 2 μM of fluorescein tagged DNA (dTn, where n denotes the length of the DNAoligomer) were incubated with an increasing concentrations of (A) MsSSBa, and (B) MsSSBb (0.3, 0.9 and 3 μM, lanes 2–4, respectively) at 4 °C for 30 min, and resolved on 8% nativePAGE in 1X TBE buffer. The gel was visualized using fluorescein filter in BIORAD gel doc system.

Fig. 7. Binding modes of MsSSBa and MsSSBb. Standardreactions (10 μl) containing 2 μM fluorescein tagged dT120

were incubated with an increasing concentrations of (A)MsSSBa (0.15, 0.3, 0.75 and 1.5 μM, lanes 2–5, 7–10, and12–15, respectively), or (B) MsSSBb (0.3, 0.6, 0.9 and 3 μM,lanes 2–5, 7–10, and 12–15, respectively) at 4 °C for 30 min.The reactions were resolved on 4.5% native PAGE in 1X TBEbuffer. DNA-SSBn refers to complex of SSB molecules onDNA, where n = 1, 2 or 3, is the number of SSB molecules(tetramers) bound to DNA. The gel was visualized usingfluorescein filter in BIORAD gel doc system.

A. Singh et al. Tuberculosis 108 (2018) 16–25

22

Mycobacterial SSBa proteins demonstrate relatively higher levels ofidentities and similarities to EcSSB than the SSBb proteins. Multiplesequence alignments have led to identification of putative base stackingresidues Y53, W59 and Y87 in SSBb. The binding site size suggests nomajor differences in the binding modes compared to those of MsSSBaand EcSSB. No obligatory dependence on the presence of divalent ionsin the solution was observed for DNA binding.

In an effort to further investigate the relative properties of MsSSBaand MsSSBb, we carried out thermal and chemical denaturation assay.Both the assays suggest higher stability of MsSSBa, towards chemicaland thermal denaturation to that of MsSSBb. The DNA binding affinityof MsSSBa was about two-fold better than that of MsSSBb. In contrast todrastic differences in the DNA binding affinities of paralogous SSBs inother organisms, ranging between 5 and 20 fold [21–23], the values arenot substantially different in M. smegmatis. We observed that the C-terminal tail is involved in negatively modulating the DNA bindingaffinity in MsSSBb. Previously, it has been reported that the disorderedC-terminal extends out laterally from the OB-domain and folds back to

compete for the same surface required for DNA binding [48,49]. Thedifferences in the DNA binding affinity of paralogous SSBs can beconceivably attributed to the differences in lengths of the spacer regionof the tail. As noted earlier, the lengths of paralogous SSBs are similar inmycobacteria, as opposed to considerable differences in other organ-isms [Table 1]. However, the surface charge and hydrophobic proper-ties of the OB-domains also appear to play a decisive role.

A consensus of cellular function or of structural features cannot beestablished with the current data available for SSBb proteins from dif-ferent species. While a role in natural-competence associated re-combination was attributed in B. subtilis and S. pneumoniae, a role inchromosomal segregation during sporulation was shown in S. coelicolor.The inverse relationship in regulation of mycobacterial SSBs duringhypoxia and UV stress suggests the protein to be of importance in stressresponse. We examined the previously reported microarray data for M.tuberculosis H37Rv proteins MtSSBa (Rv0054) and MtSSBb (Rv2478c) inTB database [50]. ssbB gene was found to be over-expressed during non-replicating persistence stage of dormancy, while at the same time ssbAgene was under-expressed [51,52]. Also, exposure to DNA damagingagents [52], growth on long chain fatty acids such as palmitic acid oroleic acid and starvation lead to under-expression of ssbA gene and si-multaneous over expression of ssbB gene (unpublished results, Yang Liuet al.). Though the changes in the level of expression are different, thepattern is consistent with other types of stress induced responses andwith strains other than H37Rv.

To gain insights into cellular properties of MsSSBb, we checked thepossibility of SSBb to interact with known interaction partners of SSBa.MsSSBa is known to interact with Ung and impact uracil excisionproperties [43]. No interaction was established between MsSSBb andMsUng, shown by agarose gel assay [Fig. 9D] or far-western blot (datanot shown). However, a detectable interaction with RecA was observed.The physical interaction of SSBb with RecA suggests a probable in-volvement of the protein to load RecA onto the ssDNA generated forrecombination event during the repair processes. The loss of interactionwith RecA on C-terminal truncation of SSBb, suggests a similar me-chanism of interaction as with SSBa. Role of SSBb proteins in naturalcompetence associated recombination of the exogenous DNA has beenestablished [21]. Since the paralogous SSBs can bind the DNA si-multaneously, MsSSBb alone or in tandem with MsSSBa can probablyplay an accessory role in recombination events in mycobacteria. Also, a

Fig. 8. Expression levels of M. smegmatis ssbA and ssbB genes in response to exposure toUV or under growth conditions of hypoxia or recovery from hypoxia.

Fig. 9. Interaction between MsRecA and MsSSBb. Increasing amounts of MsRecA (10, 20 and 30 pmol, spots 3–5 respectively) were spotted on to nitrocellulose membrane. BSA (min50 pmol) was spotted (spot 1) as negative control. After blocking with non-fat milk, membrane were incubated with a 2 μM solution of MsSSBb (A), SSBbΔ7 (B) or SSBbΔ27 (C) andprobed with anti-His antibody (to probe for MsSSBb), developed using DAB as described in Materials and Methods. Spot 2, in each blot, is the probing protein, as positive control for anti-His antibody. (D) M13 ssDNA (200 ng) was taken alone (lanes 1 and 6) or pre-incubated with a fixed concentration of 500 nM MsSSBa (lanes 2–5) or 500 nM MsSSBb and increasingconcentration ofMsUng (0.5, 2 and 4 μM, lanes 3–5 and lanes 8–10, respectively). Reactions mixtures were incubated on ice for 15 min and electrophoresed on 1% agarose gel (containingethidium bromide) in 1X TBE. The gel was visualized under UV.

A. Singh et al. Tuberculosis 108 (2018) 16–25

23

difference in the C-terminal sequence of MsSSBb may increase the re-pertoire of proteins with which SSBs can interact during DNA trans-actions in the cell. Together, these data suggest a probable role of SSBbin DNA repair during the persistence stage, when DNA damage to thebacillus is one of the major host defense mechanisms. The results re-ported here will provide the groundwork for further exploration of therole of paralogous SSBs in mycobacteria.

Acknowledgments

Financial support from the Department of Biotechnology, New Delhi(DBT) is acknowledged. AS is a CSIR research fellow. MV is AlbertEinstein Professor of the Indian National Science Academy. UV is a J. C.Bose fellow of the Department of Science and Technology, New Delhi(DST). The authors acknowledge the DBT-IISc partnership programme;University Grants Commission, New Delhi for the Centre of AdvancedStudies, and the DST-FIST level II infrastructure supports to carry outthis work.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.tube.2017.10.002.

References

[1] Namouchi A, Gomez-Munoz M, Frye SA, Moen LV, Rognes T, Tonjum T, et al. TheMycobacterium tuberculosis transcriptional landscape under genotoxic stress. BMCGenomics 2016;17(1):791.

[2] Muttucumaru DGN, Roberts G, Hinds J, Stabler RA, Parish T. Gene expressionprofile of Mycobacterium tuberculosis in a non-replicating state. Tuberculosis2004;84(3–4):239–46.

[3] Via LE, Lin PL, Ray SM, Carrillo J, Allen SS, Eum SY, et al. Tuberculous granulomasare hypoxic in Guinea pigs, rabbits, and nonhuman primates. Infect Immun2008;76(6):2333–40.

[4] Wayne LG, Sohaskey CD. Nonreplicating persistence of Mycobacterium tubercu-losis. Annu Rev Microbiol 2001;55:139–63.

[5] Dos Vultos T, Mestre O, Tonjum T, Gicquel B. DNA repair in Mycobacterium tu-berculosis revisited. FEMS Microbiol Rev 2009;33(3):471–87.

[6] Adams LB, Dinauer MC, Morgenstern DE, Krahenbuhl JL. Comparison of the roles ofreactive oxygen and nitrogen intermediates in the host response to Mycobacteriumtuberculosis using transgenic mice. Tuber Lung Dis 1997;78(5–6):237–46.

[7] Little JW. Mechanism of specific LexA cleavage: autodigestion and the role of RecAcoprotease. Biochimie 1991;73(4):411–21.

[8] Davis EO, Dullaghan EM, Rand L. Definition of the mycobacterial SOS box and useto identify LexA-regulated genes in Mycobacterium tuberculosis. J Bacteriol2002;184(12):3287–95.

[9] Meyer RR, Laine PS. The single-stranded DNA-binding protein of Escherichia coli.Microbiol Rev 1990;54(4):342–80.

[10] Marceau AH. Functions of single-strand DNA-binding proteins in DNA replication,recombination, and repair. Methods Mol Biol 2012;922:1–21.

[11] West SC. Enzymes and molecular mechanisms of genetic recombination. Annu RevBiochem 1992;61(1):603–40.

[12] Reddy MS, Vaze MB, Madhusudan K, Muniyappa K. Binding of SSB and RecAprotein to DNA-containing stem loop structures: SSB ensures the polarity of RecApolymerization on single-stranded DNA. Biochemistry 2000;39(46):14250–62.

[13] Reddy MS, Guhan N, Muniyappa K. Characterization of Single-stranded DNA-binding Proteins from Mycobacteria: the carboxyl-terminal domain of ssb is es-sential for stable association with its cognate RecA protein. J Biol Chem2001;276(49):45959–68.

[14] Murzin AG. OB(oligonucleotide/oligosaccharide binding)-fold: common structuraland functional solution for non-homologous sequences. EMBO J 1993;12(3):861–7.

[15] Shereda RD, Kozlov AG, Lohman TM, Cox MM, Keck JL. SSB as an organizer/mo-bilizer of genome maintenance complexes. Crit Rev Biochem Mol Biol2008;43(5):289–318.

[16] Handa P, Acharya N, Varshney U. Chimeras between single-stranded DNA-bindingproteins fromEscherichia coli and Mycobacterium tuberculosisReveal that their C-terminal domains interact with uracil DNA glycosylases. J Biol Chem2001;276(20):16992–7.

[17] Antony E, Weiland E, Yuan Q, Manhart CM, Nguyen B, Kozlov AG, et al. Multiple C-Terminal tails within a single E. coli SSB homotetramer coordinate DNA replicationand repair. J Mol Biol 2013;425(23):4802–19.

[18] Costes A, Lecointe F, McGovern S, Quevillon-Cheruel S, Polard P. The C-Terminaldomain of the bacterial SSB protein acts as a DNA maintenance hub at activechromosome replication forks. PLoS Genet 2010;6(12):e1001238.

[19] Kozlov AG, Cox MM, Lohman TM. Regulation of single-stranded DNA binding bythe C Termini of Escherichia coli single-stranded DNA-binding (SSB) protein. J BiolChem 2010;285(22):17246–52.

[20] Marintcheva B, Marintchev A, Wagner G, Richardson CC. Acidic C-terminal tail ofthe ssDNA-binding protein of bacteriophage T7 and ssDNA compete for the samebinding surface. Proc Natl Acad Sci U. S. A 2008;105(6):1855–60.

[21] Yadav T, Carrasco B, Myers AR, George NP, Keck JL, Alonso JC. Genetic re-combination in Bacillus subtilis: a division of labor between two single-strand DNA-binding proteins. Nucleic Acids Res 2012;40(12):5546–59.

[22] Attaiech L, Olivier A, Mortier-Barrière I, Soulet A-L, Granadel C, Martin B, et al.Role of the single-stranded DNA–binding protein SsbB in pneumococcal transfor-mation: maintenance of a reservoir for genetic plasticity. PLoS Genet2011;7(6):e1002156.

[23] Paradzik T, Ivic N, Filic Z, Manjasetty BA, Herron P, Luic M, et al.Structure–function relationships of two paralogous single-stranded DNA-bindingproteins from Streptomyces coelicolor: implication of SsbB in chromosome segre-gation during sporulation. Nucleic Acids Res 2013;41(6):3659–72.

[24] Singh A, Bhagavat R, Vijayan M, Chandra N. A comparative analysis of the DNArecombination repair pathway in mycobacterial genomes. Tuberculosis2016;99:109–19.

[25] Lohman TM, Ferrari ME. Escherichia coli single-stranded DNA-binding protein:multiple DNA-binding modes and cooperativities. Annu Rev Biochem1994;63:527–70.

[26] Purnapatre K, Varshney U. Cloning, over-expression and biochemical character-ization of the single-stranded DNA binding protein from Mycobacterium tubercu-losis. Eur J Biochem 1999;264(2):591–8.

[27] Jacobs Jr. WR, Kalpana GV, Cirillo JD, Pascopella L, Snapper SB, Udani RA, et al.Genetic systems for mycobacteria. Methods Enzymol 1991;204:537–55.

[28] Tyagi, J.S. and D. Sharma. Mycobacterium smegmatis and tuberculosis. TrendsMicrobiol 10(2): 68–69.

[29] Mayuri Bagchi G, Das TK, Tyagi JS. Molecular analysis of the dormancy response inMycobacterium smegmatis: expression analysis of genes encoding the DevR–DevStwo-component system, Rv3134c and chaperone α-crystallin homologues. FEMSMicrobiol Lett 2002;211(2):231–7.

[30] Saikrishnan K, Jeyakanthan J, Venkatesh J, Acharya N, Purnapatre K, Sekar K, et al.Crystallization and preliminary X-ray studies of the single-stranded DNA-bindingprotein from Mycobacterium tuberculosis. Acta Cryst D 2002;58(2):327–9.

[31] Saikrishnan K, Jeyakanthan J, Venkatesh J, Acharya N, Sekar K, Varshney U, et al.Structure of Mycobacterium tuberculosis single-stranded DNA-binding protein.Variability in quaternary structure and its implications. J Mol Biol2003;331(2):385–93.

[32] Saikrishnan K, Manjunath GP, Singh P, Jeyakanthan J, Dauter Z, Sekar K, et al.Structure of Mycobacterium smegmatis single-stranded DNA-binding protein and acomparative study involving homologus SSBs: biological implications of structuralplasticity and variability in quaternary association. Acta Cryst D 2005;61(Pt8):1140–8.

[33] Kaushal PS, Singh P, Sharma A, Muniyappa K, Vijayan M. X-ray and molecular-dynamics studies on Mycobacterium leprae single-stranded DNA-binding protein andcomparison with other eubacterial SSB structures. Acta Cryst D2010;66(10):1048–58.

[34] Singh A, Varshney U, Vijayan M. Structure of the second single stranded DNAbinding protein (SSBb) from Mycobacterium smegmatis. J Struct Biol2016;196(3):448–54.

[35] Saikrishnan K, Manjunath GP, Singh P, Jeyakanthan J, Dauter Z, Sekar K, et al.Structure of Mycobacterium smegmatis single-stranded DNA-binding protein and acomparative study involving homologus SSBs: biological implications of structuralplasticity and variability in quaternary association. Acta Cryst D2005;61(8):1140–8.

[36] Arif SM, Vijayan M. Structural diversity based on variability in quaternary asso-ciation. A case study involving eubacterial and related SSBs. Methods Mol Biol2012;922:23–35.

[37] Consortium TU. UniProt: a hub for protein information. Nucleic Acids Res2015;43(D1):D204–12.

[38] Sievers F, Wilm A, Dineen D, Gibson TJ, Karplus K, Li W, et al. Fast, scalablegeneration of high-quality protein multiple sequence alignments using ClustalOmega. Mol Syst Biol 2011;7. 539–539.

[39] Robert X, Gouet P. Deciphering key features in protein structures with the newENDscript server. Nucleic Acids Res 2014;42(W1):W320–4.

[40] Purnapatre K, Varshney U. Uracil DNA glycosylase from Mycobacterium smegmatisand its distinct biochemical properties. Eur J Biochem 1998;256(3):580–8.

[41] Kurthkoti K, Varshney U. Detrimental effects of hypoxia-specific expression of ur-acil DNA glycosylase (Ung) in Mycobacterium smegmatis. J Bacteriol2010;192(24):6439–46.

[42] Taboada B, Ciria R, Martinez-Guerrero CE, Merino E. ProOpDB: prokaryotic operonDataBase. Nucleic Acids Res 2012;40(Database issue):D627–31.

[43] Acharya N, Varshney U. Biochemical properties of single-stranded DNA-bindingprotein from Mycobacterium smegmatis, a fast-growing mycobacterium and itsphysical and functional interaction with uracil DNA glycosylases. J Mol Biol2002;318(5):1251–64.

[44] Bujalowski W, Overman LB, Lohman TM. Binding mode transitions of Escherichiacoli single strand binding protein-single-stranded DNA complexes. Cation, anion,pH, and binding density effects. J Biol Chem 1988;263(10):4629–40.

[45] Bochkareva E, Korolev S, Bochkarev A. The role for zinc in replication protein A. JBiol Chem 2000;275(35):27332–8.

[46] Kozlov AG, Galletto R, Lohman TM. SSB-DNA binding monitored by fluorescenceintensity and anisotropy. Methods Mol Biol 2012;922:55–83.

[47] Wu Y, Li Q, Chen XZ. Detecting protein-protein interactions by Far western blotting.Nat Protoc 2007;2(12):3278–84.

[48] Marintcheva B, Marintchev A, Wagner G, Richardson CC. Acidic C-terminal tail of

A. Singh et al. Tuberculosis 108 (2018) 16–25

24

the ssDNA-binding protein of bacteriophage T7 and ssDNA compete for the samebinding surface. Proc Nat Acad Sci U. S. A 2008;105(6):1855–60.

[49] Savvides SN, Raghunathan S, Fütterer K, Kozlov AG, Lohman TM, Waksman G. TheC-terminal domain of full-length E. coli SSB is disordered even when bound to DNA.Protein Sci 2004;13(7):1942–7.

[50] Reddy TBK, Riley R, Wymore F, Montgomery P, DeCaprio D, Engels R, et al. TBdatabase: an integrated platform for tuberculosis research. Nucleic Acids Res2009;37(Database issue):D499–508.

[51] Voskuil MI, Visconti KC, Schoolnik GK. Mycobacterium tuberculosis gene expressionduring adaptation to stationary phase and low-oxygen dormancy. Tuberc (Edinb)2004;84(3–4):218–27.

[52] Boshoff HI, Myers TG, Copp BR, McNeil MR, Wilson MA, Barry 3rd CE. The tran-scriptional responses of Mycobacterium tuberculosis to inhibitors of metabolism:novel insights into drug mechanisms of action. J Biol Chem2004;279(38):40174–84.

A. Singh et al. Tuberculosis 108 (2018) 16–25

25