Mycobacterium tuberculosis DinG Is a Structure-specific Helicase That Unwinds G4 DNA

Background: The G4 helicase in M. tuberculosis is so far unidentified. Results: M. tuberculosis DinG is a 5′ → 3′ helicase/translocase that unwinds G4 structures. Conclusion: M. tuberculosis DinG is a structure-specific helicase that unwinds G4 structures and replication/recombination/repair intermediates. Significance: G4-forming sequences exist in M. tuberculosis, and DinG is a G4 helicase. We implicate G4 structures and DinG as targets for potential therapy. The significance of G-quadruplexes and the helicases that resolve G4 structures in prokaryotes is poorly understood. The Mycobacterium tuberculosis genome is GC-rich and contains >10,000 sequences that have the potential to form G4 structures. In Escherichia coli, RecQ helicase unwinds G4 structures. However, RecQ is absent in M. tuberculosis, and the helicase that participates in G4 resolution in M. tuberculosis is obscure. Here, we show that M. tuberculosis DinG (MtDinG) exhibits high affinity for ssDNA and ssDNA translocation with a 5′ → 3′ polarity. Interestingly, MtDinG unwinds overhangs, flap structures, and forked duplexes but fails to unwind linear duplex DNA. Our data with DNase I footprinting provide mechanistic insights and suggest that MtDinG is a 5′ → 3′ polarity helicase. Notably, in contrast to E. coli DinG, MtDinG catalyzes unwinding of replication fork and Holliday junction structures. Strikingly, we find that MtDinG resolves intermolecular G4 structures. These data suggest that MtDinG is a multifunctional structure-specific helicase that unwinds model structures of DNA replication, repair, and recombination as well as G4 structures. We finally demonstrate that promoter sequences of M. tuberculosis PE_PGRS2, mce1R, and moeB1 genes contain G4 structures, implying that G4 structures may regulate gene expression in M. tuberculosis. We discuss these data and implicate targeting G4 structures and DinG helicase in M. tuberculosis could be a novel therapeutic strategy for culminating the infection with this pathogen.

Tuberculosis, resulting from infection with Mycobacterium tuberculosis, continues to be a leading cause of deaths world-wide. It is estimated that one-third of the world's population is latently infected with this pathogen, and ϳ10% of those latently infected will develop active disease (1,2). The emergence of multidrug-resistant and extremely drug-resistant strains coupled with the HIV/AIDS pandemic has further exacerbated the risk of a tuberculosis resurgence (2), underscoring the development of novel therapeutic interventions. Thus, it is important to understand the pathways and the mechanisms that M. tuberculosis adapts for its survival, persistence, and pathogenesis for the identification of novel drug targets.
In addition to the canonical B-form structure, DNA can fold into various other inter-and intramolecular secondary structures, including G-qudruplexes (G4 structures) (3). Mounting evidence points toward these structures existing under physiological conditions and playing functional roles in vivo. The G-rich sequences that fold into G-qudruplex structures (G4 DNA) are conserved throughout evolution and are abundantly present in telomeres, mitotic and meiotic double strand break sites, and transcriptional start sites (3)(4)(5). The human genome contains Ͼ375,000 G-rich sequence motifs that have the potential to form G4 DNA (6). During replication, unwinding of DNA duplexes provides the opportunity for G-rich sequences to assume G4 DNA structures. In addition, G4 structures that play roles in transcription could be present during DNA replication. Pre-existing G-quadruplexes or those formed during DNA replication block DNA replication and may lead to DNA breakage and cause genome instability (4). Thus, G4 structures need to be resolved for faithful completion of DNA replication, and helicases are implicated in unwinding G4 DNA structures (3,7). Indeed, mutations in RecQ helicases BLM and WRN as well as FANCJ, CHLR1, PIF1, and RTEL1 helicases that have been shown to resolve G4 DNA are associated with various chromosome instability and cancer susceptibility genetic disorders (4,7,8). Similar to the human genome, G4 DNA-forming sequences exist in prokaryotes. A bioinformatics study identified that the Escherichia coli genome contains ϳ3,000 G4 DNAforming sequences (9). E. coli RecQ family helicases, such as Saccharomyces cerevisiae-encoded Sgs1 and human BLM and and loaded onto the Sephacryl S-200 column that was preequilibrated with buffer C. The peak EcDinG fractions were pooled and concentrated by Amicon ultracentricon (30 kDa cut-off) using buffer C containing 100 mM NaCl and 5 mM DTT; later, the aliquots were stored at Ϫ80°C.
UV-visible Spectrum-Purified MtDinG protein was analyzed for the iron-sulfur cluster by absorbance at 403 nm using 1 mg/ml protein in Tris-HCl buffer, pH 7.5, in a 1-cm optical path length quartz cuvette by UV-visible spectrophotometer between 250 and 700 nm.
Iron Chelation Assay-The bound iron in purified MtDinG was quantified by the bathophenanthroline method. Briefly, 100 l of purified MtDinG in Tris-HCl (pH 8.0) buffer was mixed with 30 l of concentrated HCl and heat-denatured at 100°C for 15 min. The denatured protein mix was centrifuged for 5 min at 13,000 rpm at room temperature. The supernatant (100 l) was transferred to a fresh tube and mixed thoroughly with 1.3 ml of Tris-HCl (pH 8.5). Freshly prepared 5% ascorbic acid (100 l) and 0.4 ml of 0.1% bathophenanthroline were added sequentially and thoroughly mixed. The reaction was incubated for 1 h at room temperature, and later, absorbance was measured at 535 nm using a 1-cm optical path length cuvette in a UV-visible spectrophotometer (Shimadzu UV-160). Using the molar extinction co-efficient of bathophenanthroline 22 (369 mol Ϫ1 cm Ϫ1 ), the number of moles of iron bound to MtDinG was calculated.
Preparation of DNA Substrates-All oligonucleotides used in this study were purchased from Sigma-Aldrich and IDT (Table  1). Oligonucleotides were labeled at the 5Ј end with [␥-32 P]ATP using T4 polynucleotide kinase in the buffer provided by the manufacturers. The radiolabeled oligonucleotides were purified through a 1-ml Sephadex G50 column. Oligonucleotide substrates ( Table 2) were prepared by annealing 1:1.5 molar amounts of 5Ј-␥-32 P-phosphorylated and non-phosphorylated oligonucleotides in a 50-l reaction volume containing 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, and 1ϫ SSC (3.0 M sodium chloride and 0.3 M sodium citrate). The reaction mixture was heated to 95°C for 5 min and then cooled slowly in the thermal cycler. The assembled substrates were purified by resolving on 8% native PAGE. All of the G4 DNA substrates were prepared as described previously (11,12,24). Briefly, 2 M or 30 pmol (for tailed G4 structures) of radiolabeled DNA was boiled for 5 min and then allowed to cool gradually over a period of 2 h in a reaction containing 10 mM Tris-HCl, pH 7.5, 1 mM EDTA (TE) and 100 mM KCl. For tailed G4 structure preparation, in addition to 100 mM KCl, 100 mM NaCl was used in the reaction buffer. For the preparation of unimolecular G4 DNA from the M. tuberculosis genome, the sequences were retrieved from the PROQUAD G4 DNA database (9), and the formation of G4 DNA was verified using the Non-B DNA database (25) with the criteria of stem size 3-5 and loop length 1-5 nucleotides. From these sequences, G4 DNA was formed using 0.5 M radiolabeled oligonucleotide in a reaction containing 50 mM Tris-HCl, pH 7.5, 100 mM KCl. The reaction mixture was heated to 95°C and allowed to cool in thermal cycler over a period of 2 h. The intramolecular G4 formation was identified by the rapid migration on 15% non-denaturing polyacrylamide gels in 1ϫ TBE relative to the unstructured oligonucleotide in the presence or absence of KCl. All of the G4 DNA substrates were resolved and purified from 15 and 18% non-denaturing polyacrylamide gels in 1ϫ TBE containing 100 mM KCl.
Mass Spectrometry and Peptide Mass Fingerprinting-Purified MtDinG (2 mg/ml) was dialyzed in a buffer containing 20 mM Tris-HCl, 1 mM DTT, and 50 mM NaCl and spotted on a MALDI-TOF MS plate using 4-chloro-␣-cyanocinnamic acid as matrix. MALDI-TOF analysis was performed in an Ultraflex TOF mass spectrometer (Bruker Daltonics) using the positive ion detection mode. Peptide mass fingerprinting of trypsin-digested MtDinG was performed as described (26).
Electrophoretic Mobility Shift Assay-The assays were carried out as described previously (27). Briefly, the reaction mixture (20 l) containing 25 mM Tris-HCl, 1 mM DTT, 0.1 mg/ml BSA, 5 mM MgCl 2 , and 0.5 nM 32 P-labeled DNA substrates was incubated with increasing concentrations of MtDinG for 15 min at 37°C. The DNA-protein complex was resolved on 6% non-denaturing polyacrylamide gels using 0.25ϫ TBE (22.5 mM Tris borate buffer, pH 8.3, and 0.25 mM EDTA) and visualized by a phosphor imager (Fuji FLA-5000, Tokyo, Japan). Reaction products were quantified using Fujifilm MultiGauge software (version 3.0); data were subjected to nonlinear regression analysis and plotted using GraphPad prism (version 5.0; San Diego, CA).
Helicase Assay-The reaction mixture (20 l) contained 50 mM Tris-HCl, pH 7.5, 1 mM DTT, 0.1 mg/ml BSA, 0.5 nM 32 Plabeled DNA substrate, 5 mM ATP, and 5 mM MgCl 2 . The reactions were initiated by the addition of either EcDinG or MtDinG and incubated at 37°C for 30 min. Later, the reactions were terminated by the addition of 5 l of stop buffer (2% SDS, 0.1% (w/v) bromphenol blue, xylene cyanol in 20% glycerol and 5 mg/ml proteinase K) and further incubated for 20 min. The products were separated on 8 -10% non-denaturing polyacrylamide gels in 0.5ϫ TBE at 150 V for 2.5 h at room temperature. For G4 unwinding assays, 10 mM KCl was included in the reaction mixtures as well as in the 1ϫ TBE for resolving the G4 unwound products using 10% non-denaturing polyacrylamide gels. The products were quantified by phosphor imager (Fuji FLA-5000) and analyzed using Fujifilm MultiGauge software.
Streptavidin Displacement Assay-The reaction mixture (20 l) contained 0.5 nM 32 P-labeled biotinylated oligonucleotide in a helicase assay buffer (50 mM Tris-HCl, pH 7.5, 1 mM DTT, 0.1 mg/ml BSA) with 5 mM ATP, 5 mM MgCl 2 , and 100 nM streptavidin. The reaction mixture was preincubated for 5 min at 37°C to facilitate streptavidin binding to biotin. The reactions were initiated by the addition of EcDinG or MtDinG and incubated for 15 min at 37°C. Later, 1 M free biotin was added before terminating the reaction with 5 l of stop buffer (2% SDS, 0.1% (w/v) bromphenol blue, xylene cyanol in 20% glycerol and 5 mg/ml proteinase K) and further incubated for 20 min. The products were separated on an 8% non-denaturing polyacrylamide gels in 0.5ϫ TBE at 150 V for 2.5 h at room temperature. The products were quantified by a phosphor imager (Fuji FLA-5000) and analyzed using Fujifilm MultiGauge software.
DNase I Footprinting-DNase I footprinting was carried out as described previously (28). Reaction mixtures (40 l) con-tained 5 nM 32 P-labeled 3Ј overhang substrate in 25 mM Tris-HCl (pH 7.5), 1 mM DTT, 100 g/ml BSA, 5 mM MgCl 2 , 5 mM CaCl 2 , and increasing concentrations of EcDinG and MtDinG. Samples were incubated for 15 min at 4°C. Reactions were initiated by the addition of DNase I to a final concentration of 0.05 units and incubated for 1 min at 25°C. The reactions were terminated by the addition of 50 mM EDTA. DNA was precipitated with the addition of one-tenth volume of 3 M sodium acetate, pH 5.2, and 3 volumes of 95% ethanol with 100 g/ml yeast tRNA. The pellet was washed with 70% ethanol, dried under vacuum, and resuspended in loading dye (80% (v/v) formamide, 0.1% (v/v) BPB, and 0.1% (v/v) xylene cyanol) and analyzed on a denaturing 15% polyacrylamide gel containing 7 M urea. The gel was dried, and the bands were visualized with a Fuji FLA-5000 phosphor imager. Maxam and Gilbert A ϩ G ladder and T ladder were prepared as described previously (29).
Circular Dichroism-For CD analysis, 30 M G4 DNA prepared using the sequences from the M. tuberculosis genome was taken in a 20-l reaction volume. The DNA samples were boiled for 10 min in the presence or absence of 100 mM KCl and gradually allowed to cool. The reaction was diluted to 200 l using TE (pH 7.5) containing 100 mM KCl. The CD spectra were recorded from 220 to 320 nm at room temperature using a JASCO J-810 spectropolarimeter at a scanning speed of 50 nm/min, 1-s response time, in a 2-mm optical path length quartz cuvette. The spectrum represents the average of three scans. The spectrum for buffer was recorded separately, and it was subtracted from the spectrum obtained for all the G4 DNA substrates.
DMS Protection Assay-The reaction was carried out as described previously (30). Briefly, the band corresponding to the G4 DNA was excised from the gel and eluted in Tris-HCl, pH 7.5, containing 100 mM KCl for 6 h at 4°C. Later, the DNA was concentrated to 50 l, and 10 5 cpm/reaction was diluted in buffer containing 10 mM Tris-HCl, pH 7.5, 100 mM KCl, 100 g/ml yeast tRNA to a volume of 80 l. The reaction was started with the addition of 1 l of 1:100 diluted dimethyl sulfate solution (DMS/ethanol, 4:1 (v/v)) for 15 min at room temperature. The reaction was stopped by adding 18 l of stop buffer (3 M ␤-mercaptoethanol/water/sodium acetate, 1:6:7 (v/v/v)) and precipitated by ethanol. The dried pellet was subjected to cleavage by piperidine. The products were resolved on denaturing PAGE (15%) and visualized by a phosphor imager.

Cloning and Purification of M. tuberculosis DinG-E. coli
DinG helicase belongs to the iron-sulfur family of helicases that are implicated in DNA replication and repair (17). Pairwise comparison of amino acid sequence between MtDinG and EcDinG reveals that MtDinG is 34% identical and 48% similar to EcDinG. Although the overall primary sequence of MtDinG is less conserved to that of E. coli, the seven helicase signature motifs are highly conserved (Fig. 1A). Notably, MtDinG possesses the four conserved cysteines that form the iron-sulfur domain as that of EcDinG. To understand the biochemical functions of MtDinG, the ORF encoding for full-length DinG was PCR-amplified using forward and reverse primers as described under "Experimental Procedures." PCR amplifica-tion yielded a 1995-bp product corresponding to the predicted length of MtdinG (data not shown). Similarly, PCR was carried out for an MtDinG ATPase-deficient mutant (Walker A motif K55R). These PCR products were cloned into pET28a (ϩ) E. coli expression vector to obtain expression of MtDinG with an N-terminal histidine (His 6 ) tag. Positive clones were identified by restriction analysis and confirmed by nucleotide sequencing. WT MtDinG and MtDinG K55R were expressed in E. coli C43 (DE3) cells and purified as described under "Experimental Procedures." However, our attempts to obtain the native protein by cleaving the N-terminal His tag using tobacco etch virus protease were unsuccessful. Fig. 1B shows the MtDinG purification profile using different chromatography steps. The purified WT (Fig. 1E) and mutant MtDinG were yellowish brown in appearance, indicating the iron-bound status of the purified MtDinG. The purity, mass, and identity of the purified MtDinG were verified by MALDI electrospray ionization-MS, MALDI MS-MS, and peptide mass fingerprinting ( Fig. 1C) (data not shown). The purified MtDinG was Ͼ95% homogeneous and was devoid of any contaminating helicase from the E. coli host. For a comparative study, we also purified EcDinG with a modified protocol as described under "Experimental Procedures." The purified protein was Ͼ95% homogeneous (Fig. 1D). To validate the iron-bound status of the purified MtDinG, it was subjected to UV-visible absorption spectrophotometric analysis. Similar to purified EcDinG, MtDinG displayed peak absorption at ϳ403 nm (Fig. 1E), a characteristic absorption exhibited by iron-sulfur domain-containing proteins. In addition, a colorimetric assay with bathophenanthroline to determine the iron concentration revealed that each monomer of MtDinG and EcDinG contained 3.1 and 3.3 Fe 2ϩ ions, respectively, which is similar to what has been reported for EcDinG (31).
The DNA Binding Activity of M. tuberculosis DinG-E. coli DinG is a structure-specific helicase that catalyzes unwinding of branched DNA molecules (21). To investigate whether MtDinG can bind to synthetic substrates that resemble intermediates of various DNA metabolic pathways, such as replication, repair, and recombination, we carried out electrophoretic mobility shift assays. The sequences of oligonucleotides and the different substrates used in this study are shown in Tables 1 and 2. Each substrate was prepared using one of the oligonucleotides that was labeled with [␥-32 P]ATP as described under "Experimental Procedures." The indicated radiolabeled DNA substrates were incubated individually with increasing concentrations of purified MtDinG. The DNA-protein complexes were separated on 6% polyacrylamide gel and analyzed as described under "Experimental Procedures." MtDinG exhibited robust DNA binding activity with ssDNA; the DNA-protein complex was evident at the tested lowest concentration of protein (25 nM), and the maximum binding was observed at 500 nM MtDinG ( Fig. 2A). Similarly, we tested the DNA binding activity of MtDinG with overhang substrates, flap duplexes, and other types of branched DNA molecules. However, compared with ssDNA, MtDinG exhibited relatively weak DNA binding activity with 3Ј or 5Ј overhang substrates, 3Ј or 5Ј flap structures, and forked duplex as well as three-strand junctions (Fig.  2, B-G and K). Notably, MtDinG failed to exhibit a significant binding with replication fork, three-way junction, and blunt end duplex DNA substrates that lack ssDNA tails, indicating that MtDinG requires ssDNA overhangs to bind to different DNA substrates (Fig. 2, H-J). The quantitative data clearly show that MtDinG preferentially binds to ssDNA in contrast to other types of replication/recombination intermediate substrates that lack ssDNA overhangs (Fig. 2K). The data in Fig. 2A show that at 500 nM MtDinG, it was able to form ϳ100% DNAprotein complexes with 83-mer ssDNA. We also compared the binding affinity of purified K55R MtDinG with ssDNA substrates and overhang substrates and found that Walker motif mutant binds to the DNA as efficiently as WT protein (data not shown).
The Helicase Activity of M. tuberculosis DinG-E. coli DinG has been shown to unwind branched DNA substrates with a 5Ј 3 3Ј polarity (20). To investigate the helicase activity of MtDinG and the substrate preference as well as the polarity of unwinding, we assembled various branched DNA molecules, and MtDinG helicase activity was measured. Radiolabeled blunt-ended double-stranded DNA (dsDNA), 5Ј or 3Ј ssDNA overhangs, and 5Ј or 3Ј flap structures were incubated with increasing concentrations of MtDinG (5-250 nM) in the presence of 5 mM ATP and 5 mM MgCl 2 . The unwinding activity was measured as described under "Experimental Procedures." MtDinG failed to unwind 22-mer and 50-mer dsDNA (data not shown) but exhibited robust unwinding activity with 3Ј overhang substrates (Fig. 3, B and D). In a parallel study, MtDinG was proficient in unwinding 5Ј overhangs, although maximum activity was found at the highest tested concentration of protein (Fig. 3, A and D). The purified MtDinG Walker A motif K55R mutant was devoid of unwinding activity at the highest tested concentration (250 nM), suggesting that MtDinG-mediated helicase activity requires ATP hydrolysis. In parallel reactions, we compared the helicase activity of EcDinG. Consistent with previous studies (20,21), EcDinG was able to unwind 5Ј overhangs efficiently (Fig. 3, A and C). However, in contrast to ear-    lier studies (20,21), EcDinG exhibited significant unwinding activity with 3Ј overhangs (Fig. 3, B and C). Subsequently, we tested the ability of MtDinG to unwind forked duplex substrates as well as 5Ј flap and 3Ј flap structures. MtDinG and EcDinG displayed significant unwinding activity with forked duplex DNA (data not shown). Notably, consistent with overhang structures, WT MtDinG exhibited more efficient unwinding of 3Ј flap structures than 5Ј flap structures (Fig. 3, E, F, and H). Similar to overhang structures, the ATPase-deficient mutant of MtDinG failed to unwind flap structures and forked duplex substrates (Fig. 3, E and F) (data not shown). In agreement with overhang structures, EcDinG was proficient in unwinding 5Ј flap structures compared with 3Ј flap structures (Fig. 3, E-G). These results suggest that MtDinG preferentially unwinds 3Ј overhangs or flap structures rather than 5Ј substrates. However, at higher concentration, its unwinding activity with 5Ј overhangs or flap structures was comparable with that of 3Ј substrates.
To further understand the polarity of MtDinG-catalyzed unwinding, we used two different previously characterized types of substrates (20). The forked duplex substrate was mod-ified to switch the polarity such that the unpaired singlestranded ends contain either 5Ј or 3Ј ends. Consistent with overhang substrates and flap structures, MtDinG was able to proficiently unwind substrates with 3Ј ends more robustly than 5Ј ends (Fig. 4, A, B, and D). However, at a higher concentration of protein, MtDinG was equally proficient in unwinding substrates with 5Ј ends. In agreement with previous data (20,21), EcDinG displayed robust unwinding with substrates containing 5Ј ends, suggesting the preferential unwinding activity with a 5Ј 3 3Ј polarity (Fig. 4, B and C). However, it exhibited a significant unwinding with modified substrates containing 3Ј ends (Fig. 4, A and C). Together, these data suggest that MtDinG and EcDinG exhibit helicase activities with substrates that contain ssDNA with either 5Ј or 3Ј ends. In contrast to DinG helicase, H. influenzae UvrD (HiUvrD) helicase was able to unwind modified forked duplex substrates only with 3Ј ends and not with 5Ј ends (Fig. 4E), suggesting that HiUvrD unwinds DNA substrates in a 3Ј 3 5Ј polar fashion, which is in agreement with an earlier study (23). In all of our helicase assays, we found that EcDinG was able to display optimal activity between 10 and 25 nM protein, but MtDinG required Ͼ100 nM protein (substrate/ protein ratio of 1:200). However, it has been shown that mycobacterial enzymes require Ͼ100 nM protein for its optimal activity (32)(33)(34)(35).
M. tuberculosis as Well as E. coli DinG Helicases Exhibit ssDNA Translocase Activity with a 5Ј 3 3Ј Polarity-To understand whether a relation exists with respect to the helicase activity of MtDinG with a translocation on ssDNA, we performed a streptavidin displacement assay using biotinylated oligonucleotides. We used two different types of biotinylated oligonucleotides. One set had a biotin at the 5Ј end of ssDNA, and the second had a biotin at the 3Ј end. Streptavidin forms a strong complex with biotin with high affinity (K d ϳ10 Ϫ15 M). Using these substrates, streptavidin displacement was monitored with an incubation of increasing concentrations of EcDinG and MtDinG. Interestingly, both EcDinG and MtDinG failed to displace streptavidin from the 5Ј biotin-streptavidin complex (Fig. 5, A, C, and D). In contrast, these proteins were efficient in displacing streptavidin from the 3Ј biotin-streptavidin complex, suggesting that EcDinG and MtDinG translocates on ssDNA in the 5Ј 3 3Ј direction (Fig. 5, B-D). Notably, the ATPase-deficient mutant of MtDinG failed to exhibit displacement of streptavidin, indicating that the translocase activity of MtDinG requires ATP hydrolysis. In a parallel experiment, we compared the ssDNA translocase activity of HiUvrD helicase and found that in contrast to DinG, HiUvrD was able to translocate only in the 3Ј 3 5Ј direction (Fig. 5, A and B).
Mechanistic Insights into the Helicase Activity of M. tuberculosis DinG-Our data with helicase activity and ssDNA translocation with DinG proteins were striking, such that ssDNA translocation was observed only in the 5Ј 3 3Ј direction, but helicase action was observed with substrates containing either 5Ј or 3Ј ends. Moreover, our DNA binding as well as unwinding studies clearly indicated that MtDinG was able to bind ssDNA with high affinity and also unwind model substrates that con- tain either 5Ј or 3Ј ssDNA. These results imply that MtDinG requires ssDNA for loading and unwinding by translocating on ssDNA. These observations led us to investigate the minimum length of ssDNA that is required by MtDinG and EcDinG for its helicase activity. To address this, we prepared 5Ј and 3Ј overhang substrates with varying lengths of ssDNA overhangs (0, 5, 15, 25, 35, and 45 nucleotides). With our earlier data (Fig. 3, A  and B), we noticed that MtDinG was able to unwind 3Ј overhangs more proficiently than 5Ј overhangs at lower concentrations of proteins. This could be because MtDinG exhibited a higher affinity for 3Ј overhang or 3Ј flap structures than 5Ј sub-strates (Fig. 2, B, C, E, F, and K). To rule out any sequence bias in the loading of DinG molecules to ssDNA, we designed oligonucleotide substrates to assemble 5Ј or 3Ј overhangs with different lengths of thymine (poly(dT)) residues in the ssDNA tail region. In all of these substrates, the 20-bp duplex region was identical. These substrates were incubated with increasing concentrations of MtDinG (0 -100 nM), and the products were analyzed as described under "Experimental Procedures." MtDinG failed to exhibit helicase activity with dsDNA or 5Ј or 3Ј overhang structures containing a 5-nucleotide overhang (Fig. 6, A-D). However, it was able to show significant unwinding activity with both 5Ј and 3Ј overhang structures that contain 15-nucleotide overhangs, and this activity was maximal with 25-nucleotide overhangs (Fig. 6, A-D). Interestingly, a further increase in unwinding was not observed with the substrates that had 35or 45-nucleotide overhangs (data not shown). These results are in agreement with our DNA binding as well as ATPase activities of MtDinG with these substrates (data not shown), suggesting that 25-nucleotide overhangs provide an optimal length of ssDNA region for efficient binding of MtDinG to 5Ј or 3Ј overhang structures and for its helicase activity. Notably, in contrast to earlier data with 3Ј overhangs, 3Ј flap structures, and forked duplex substrates, MtDinG exhibited more robust unwinding with 5Ј overhangs than 3Ј overhangs, which is consistent with its 5Ј 3 3Ј translocation polarity.
In a parallel study, we also examined unwinding activity of EcDinG with various 5Ј and 3Ј overhang substrates. Similar to MtDinG, EcDinG failed to unwind dsDNA as well as 5-nucleotide overhang structures. However, EcDinG exhibited significant helicase activity with 5Ј overhangs that contain 15-nucleotide ssDNA tails, and this activity was maximal with 25-nucleotide overhangs (Fig. 6, E and G). Interestingly, EcDinG was less efficient with 15-nucleotide 3Ј overhangs but displayed significant unwinding when ssDNA tail was increased to 25 nucleotides (Fig. 6, F and H). Together, these results suggest that both MtDinG and EcDinG helicases require ϳ25-nucleotide ssDNA tails for their efficient loading as well as optimal unwinding activity with 5Ј 3 3Ј polarity.
To gain mechanistic insights into the helicase activity of MtDinG, we carried out DNase I footprinting experiments with 3Ј overhang substrates containing a 25-nucleotide (T-25) ssDNA tail. The longer strand was radiolabeled with [ 32 P]ATP, and overhang substrate was prepared as described under "Experimental Procedures." In the absence of MtDinG, DNase I hypersensitive sites were found at the ssDNA-dsDNA junction. Interestingly, in the presence of MtDinG, hypersensitive sites of DNase I were protected (Fig. 7A), suggesting that MtDinG specifically binds to the ssDNA-dsDNA junction. Our parallel examination with EcDinG revealed that, similar to MtDinG, EcDinG was also able to protect the hypersensitive sites of DNase I (Fig. 7A). To understand whether binding of nucleo- tide cofactor can influence the binding pattern of DinG proteins to overhang structures, we carried out a DNase I footprinting assay in the presence of ATP␥S and found that the pattern of the DNase I footprint was very similar compared with the data obtained in the absence of ATP␥S (data not shown). These results clearly suggest that MtDinG and EcDinG bind at the junction of ssDNA and dsDNA. Interestingly, our comparative study showed that binding of MtDinG but not EcDinG to the ssDNA-dsDNA junction resulted in specific cleavage at thymine in the ssDNA-dsDNA junction (Fig. 7A,  asterisk), implying that the orientation and binding pattern of MtDinG and EcDinG to overhang structures appears to be different. Based on our observation, we propose a model in which binding of MtDinG and EcDinG to the ssDNA-dsDNA junction of 3Ј overhangs causes destabilization of the shorter com-plementary strand from the 5Ј end, which provides a nucleation site for loading of MtDinG as well as EcDinG. Subsequent to nucleation from the 5Ј end, DinG protein can translocate on the same strand in a 5Ј 3 3Ј polar fashion and can unwind the opposite strand (Fig. 7B). However, further studies are required to gain more insight into the mechanism of helicase and translocase activities of DinG.
M. tuberculosis DinG but Not EcDinG Unwind Replication Fork Structures-Spontaneous or damage-induced replication stalling presents a serious threat to genome stability and cell viability (36 -43). In E. coli, RuvABC, RecG, and UvrD helicases have been shown to remodel the stalled replication forks to resume normal DNA replication (44 -47). Three-way junctions with a gap represent a stalled fork, and we examined whether MtDinG can unwind these structures. We prepared two differ- ent replication fork structures, and in one case, we labeled leading strand, and in the second case, we labeled lagging strand. Incubation of increasing concentrations of MtDinG (25-750 nM) with a leading strand-labeled replication fork structure resulted in unwinding of lagging strand, and the 5Ј flap structure generated was further unwound by MtDinG (Fig. 8A). Similarly, MtDinG was proficient in unwinding lagging strand-labeled replication fork structures (Fig. 8B). In a parallel assay, we measured the ability of EcDinG to unwind replication fork structures. Interestingly, in contrast to MtDinG, EcDinG failed to unwind these structures (Fig. 8, A and B). We also tested the activity of MtDinG with a three-strand junction and a threeway junction without a gap, and MtDinG was proficient in unwinding three-strand junctions but exhibited a weak activity with three-way junctions (data not shown). In a parallel comparison, EcDinG was able to unwind three-strand junctions but failed to exhibit activity with three-way junctions (data not shown). These data suggest that MtDinG but not EcDinG possesses branch migration activity to remodel replication fork structures by unwinding leading as well as lagging strands. Notably, although MtDinG was able to efficiently unwind overhang substrates, flap structures, and forked duplexes in the low-  1, 6, 11, and 16) and in the presence of 12.5, 25, 50, and 100 nM MtDinG (A and B) (lanes 2-5, lanes 7-10, lanes 12-15, and lanes 17-20, respectively) or 2.5, 5, 10, and 20 nM EcDinG (E and F) (lanes 2-5, lanes 7-10, lanes 12-15, and lanes 17-20, respectively). The filled triangle at the top of the gel image denotes increasing concentrations of MtDinG or EcDinG. A and E, 5Ј overhang substrates containing 5, 15, and 25 nucleotides (poly(dT)) at the ssDNA tails. B and F, 3Ј overhang substrates containing 5, 15, and 25 nucleotides (poly(dT)) at the ssDNA tails. C and D, quantitative data for the ability of MtDinG to unwind 5Ј and 3Ј overhang structures, respectively. G and H, quantitative data for the ability of EcDinG to unwind 5Ј and 3Ј overhang structures, respectively. Data represent the mean of at least three independent experiments with S.D. values indicated by error bars. est range of concentration (5-250 nM), it exhibited proficient unwinding activity of branched substrates at 3-fold higher concentration (25-750 nM) (Fig. 8, A and B).
M. tuberculosis DinG, but Not E. coli DinG, Promotes Holliday Junction Branch Migration-HJs are the intermediate structures that are generated during replication fork restart pathways and during recombinational repair processes (40,41,48). Because MtDinG was able to efficiently unwind replication fork structures, we tested whether it can unwind HJs. We assembled radiolabeled mobile HJ substrate that contains 12-bp homology at the core of HJ. Interestingly, MtDinG exhibited significant unwinding activity with mobile HJs (Fig. 8C). In a parallel reaction, we examined the ability of EcDinG to unwind HJs. Consistent with a previous study (21), EcDinG failed to exhibit HJ branch migration activity even at the highest tested concentration (Fig. 8C). Recently, we have shown that MtRecG helicase robustly unwinds mobile HJs at a concentration ranging from 5 to 50 nM (28). To gain insight into the HJ branch migration activity of MtDinG, we compared the efficiency of HJ unwinding with MtRecG. Consistent with our previous study (28), MtRecG exhibited HJ branch migration activity at the tested lowest concentration (5 nM), and this activity was maximal with 10 nM concentration (Fig. 8D). However, MtDinG displayed weak HJ unwinding activity in the range of ϳ50 nM, and this activity increased modestly at higher concentrations (250 -750 nM) (Fig. 8D). In order to characterize the rate of HJ unwinding by MtDinG and MtRecG, we measured the kinetics of HJ branch migration by MtDinG in comparison with MtRecG. As shown in Fig. 8, E and F, MtRecG was able to unwind ϳ40% of HJ within 5 min, and by 30 min, it was able to unwind ϳ70% of substrates. In a parallel assay, MtDinG exhibited an ϳ50% lower rate of unwinding compared with MtRecG (Fig. 8, E and F). These data suggest that MtDinG is a less potent HJ-processing helicase compared with MtRecG.
M. tuberculosis DinG Unwinds G-quadruplex (G4) DNA-Germ line mutations in FANCJ cause Fanconi anemia, a rare chromosome instability and cancer susceptibility genetic disorder (49 -52). Biochemical studies show that FANCJ and E. coli DinG helicases resolve G4 DNA (22,24). MtDinG is an ortholog of human FANCJ; however, it is unknown whether MtDinG can unwind G4 DNA structures. To test whether MtDinG can resolve G4 structures, we radiolabeled an oligonucleotide and assembled previously characterized intermolecular TP-G4 structures (11). We incubated TP-G4 structures with increasing concentrations of MtDinG (5-250 nM). Strikingly, MtDinG exhibited a robust unwinding activity with TP-G4 structures, and 100 nM protein was able to unwind almost all of the TP-G4 structures (Fig. 9, A and F). In a parallel reaction, EcDinG also displayed efficient G4-unwinding activity (Fig. 9, A and E). To validate our observation, we have tested the ability of MtDinG as well as EcDinG to unwind other types of G4 DNA substrates that have been reported (11,24). Similar to TP-G4 structures, both MtDinG and EcDinG were proficient in unwinding TPB-G4 and OX-1T-G4 G4 structures (Fig. 9, B, C, E, and F). In addition to parallel G4 structures, MtDinG was proficient in unwinding anti-parallel OX-1T-G2 tetraplex structures (Fig. 9,  D and F). To understand the rate of G4 DNA resolution by MtDinG in comparison with EcDinG, we measured the kinetics of G4 unwinding at different time intervals using TP-G4, TPB-G4, and OX-1T-G4 parallel G4 structures as well as OX-1T-G2 anti-parallel tetraplex structures. With TP-G4 and TPB-G4 substrates, both MtDinG and EcDinG exhibited a robust G4 unwinding with an efficiency of Ͼ80% unwinding by 1 min (data not shown). Similarly, EcDinG and MtDinG were able to efficiently unwind OX-1T-G4 with a similar kinetics (data not shown). However, with anti-parallel OX-1T-G2 tetraplex structures, MtDinG exhibited a slower rate of unwinding compared with EcDinG (data not shown).
The G4-stabilizing ligands TMPyP4 and N-methyl mesoporphyrin IX (NMM) have been shown to inhibit the G4-unwinding activity of E. coli RecQ and RecQ orthologs Sgs1 and BLM helicases (13,53). To determine the specificity of G4 resolution activity of MtDinG, we incubated increasing concentrations of TMPyP4 and compared the TP-G4-and TPB-G4-unwinding activity of MtDinG. Compared with controls, 50 nM TMPyP4 abolished ϳ50% of the TP-G4-unwinding activity of MtDinG, and this activity was suppressed to ϳ100% at 500 nM TMPyP4 (Fig. 10, A and B). Similarly, 50 M NMM caused ϳ100% inhibition of TP-G4-unwinding activity by MtDinG (data not shown). In a parallel study, we examined the ability of TMPyP4 to mediate inhibition of TPB-G4 unwinding by MtDinG. Consistent with TP-G4 structures, TMPyP4 abrogated the G4 resolution activity of MtDinG (Fig. 10, C and D). Similar to MtDinG reactions, G4 inhibitors were able to suppress the G4-unwinding activity of EcDinG (Fig. 10, A-D). Notably, both TMPyP4 and NMM did not interfere with MtDinG-and EcDinG-promoted forked duplex unwinding (Fig. 10, E and F) (data not shown). These results further suggest that MtDinG resolves G4 structures in vitro, and this activity can be specifically inhibited by the G4-stabilizing TMPyP4 and NMM ligands.
M. tuberculosis DinG Unwinds Both 5Ј-and 3Ј-Tailed G4 Structures Efficiently-To test whether 5Ј or 3Ј tails influence the G4 resolution activity of MtDinG, we measured the G4-unwinding activity with previously characterized G4 structures that contain either 5Ј or 3Ј tails (12,24). In parallel reactions, we compared the activity of EcDinG. Interestingly, EcDinG and MtDinG were able to efficiently unwind 5Ј-tailed G4 structures (Fig. 11, A and B). Strikingly, MtDinG, but not EcDinG, was proficient in unwinding 3Ј-tailed G4 DNA substrates (Fig. 11, C  and D). These data suggest that, unlike EcDinG, MtDinG can efficiently unwind both 5Ј-and 3Ј-tailed G4 structures. To understand the rate of G4 unwinding with tailed substrates, we performed the kinetics of G4 resolution at different time intervals. Compared with EcDinG, MtDinG exhibited a slower kinetics of G4 unwinding with 5Ј-tailed G4 structures (data not shown). Notably, in contrast to MtDinG, EcDinG failed to unwind 3Ј-tailed G4 structures (data not shown).
G4 DNA-forming Sequences Exist in the M. tuberculosis Genome-The M. tuberculosis genome is GC-rich (65%), and a bioinformatics study identified that there are Ͼ10,000 GC-rich sequence motifs that have the potential to form G4 structures (9). Notably, G4 DNA-forming sequences are conserved and enriched at the promoter regions of human genes (54). However, experimental evidence for whether GC-rich sequences in the M. tuberculosis genome can fold into G4 structures is lacking. To test this, using a bioinformatics tool, we identified many G4-forming motifs in the promoters of M. tuberculosis genes. To validate some of these sequences experimentally, sequences at the PE_PGRS2, mce1R, and moeB1 genes were used (Fig.  12A) (Table 3). Oligonucleotides corresponding to these sequences were analyzed by CD. Parallel G-quadruplexes show a peak at ϳ260 -265 nm and a trough at 240 nm, whereas antiparallel quadruplexes exhibit a peak at 290 nm and a dip at 260 nm. CD spectra were recorded for all of the three oligomers in the presence of 100 mM KCl, where K ϩ is known to stabilize the G-quadruplex structures. A characteristic peak at ϳ265 nm and dip at 240 nm were evident with mce1R and PE_PGRS2 sequences (Fig. 12B), indicating that these sequences form a parallel G-quadruplex structure. In contrast, moeB1 sequence exhibited peak at 290 nm as well as 265 nm (Fig. 12B), suggesting the existence of a G-quadruplex with parallel and anti-parallel conformation.
To validate our observation obtained with CD as well as to determine whether GC-rich sequences located at promoters of PE_PGRS2, mce1R, and moeB1 form intra-or intermolecular G4 DNA structure, these sequences were analyzed by native polyacrylamide gel electrophoresis in the presence or absence of KCl. The mobilities of these oligonucleotides before and after forming G4 structures were compared with known oligonucleotides that do not form G4 DNA. The electrophoresis was carried out in the absence or presence of 100 mM KCl in 1ϫ TBE buffer. As can be seen in Fig. 12C, in the absence of KCl, all of the three gene promoter sequences migrated closely with the control non-G4 DNA-forming sequences. Interestingly, the GC-rich oligonucleotides migrated faster than the control oligonucleotides in the presence of KCl, suggesting the formation of intramolecular G4 structures (Fig. 12C).
The G4 structures also can be analyzed by a DMS protection assay. The N7 positions of guanines that are present in ssDNA and dsDNA are susceptible for modification by DMS. However, if the guanines are involved in Hoogsteen base pairing, which is required for G-quadruplex formation, the N7 position is protected from methylation by DMS. To validate further, we performed a DMS protection assay for the three G4-forming sequences in the presence or absence of KCl. Consistent with our CD spectra and data from gel electrophoresis, in the presence of KCl, the guanine residues of PE_PGRS2, mce1R, and moeB1 sequences were protected from DMS modification, whereas in the absence of KCl, these residues were unprotected from methylation by DMS (Fig. 12D). Together, these data clearly show that G4 structure-forming sequences exist in the M. tuberculosis genome.

DISCUSSION
The role of DinG helicase in various DNA metabolic transactions and genome maintenance is poorly understood. Here, we demonstrate that M. tuberculosis DinG possesses structurespecific helicase activity. Although MtDinG did not exhibit structure-specific DNA binding activity, it was proficient in unwinding a wide variety of model substrates that resemble intermediates of DNA repair, recombination, and replication. For the first time, our study demonstrates that, unlike E. coli DinG, MtDinG unwinds replication fork and HJ structures, indicating the likely role of MtDinG in the rescue of stalled replication forks and in the recombinational repair processes. E. coli RecQ and its eukaryotic orthologs unwind G4 structures (8,55). Notably, RecQ is absent in M. tuberculosis, and for the first time, our study shows that MtDinG unwinds intermolecular G4 structures, implying that MtDinG could replace RecQ at least in part with G4 DNA unwinding activity. Together, these data suggest that MtDinG possesses multiple functions, including G4 DNA resolution, to maintain genome integrity. Finally, our in vitro studies demonstrate that G4 DNA-forming sequences exist in the M. tuberculosis genome and implicate that these structures can be targeted for therapy using small molecule inhibitors.
We find that MtDinG exclusively binds to ssDNA and ssDNA-containing substrates and fails to bind linear duplexes or other structures that lack ssDNA. Interestingly, MtDinG exhibited efficient unwinding with 3Ј overhangs and 3Ј flap structures compared with 5Ј substrates, probably due to its higher affinity for these substrates. However, at higher concentrations, it was proficient in unwinding 5Ј overhangs or 5Ј flap structures. Interestingly, with 5Ј or 3Ј overhang structures that contained poly(dT) in the ssDNA region, MtDinG displayed proficient unwinding with 5Ј overhangs, suggesting that MtDinG unwinds DNA substrates with a 5Ј 3 3Ј polarity. Earlier reports showed that EcDinG unwinds DNA substrates with a 5Ј 3 3Ј polarity (20,21). However, in addition to 5Ј substrates, we found a significant unwinding activity with 3Ј overhangs, 3Ј flap structures, and 5Ј-5Ј-linked forked duplex structures. Interestingly, with our ssDNA translocation assay, both MtDinG and EcDinG displayed translocation activity in the 5Ј 3 3Јdirection, which is in agreement with the XPD family of helicases (56 -59). These data clearly suggest that MtDinG and EcDinG unwind DNA substrates with a 5Ј 3 3Ј polarity.
Our studies also provided mechanistic insights into translocation and unwinding activity of MtDinG with 3Ј overhangs. First, both MtDinG and EcDinG failed to unwind linear duplex DNA. Second, DinG proteins required a minimum of 15-nucleotide 5Ј or 3Ј overhang structures for its loading onto the ssDNA and helicase activity. Third, DNase I footprinting data showed that DinG proteins bind to ssDNA-dsDNA junctions of 3Ј overhangs. Based on these observations, we propose a model in which binding of DinG proteins to ssDNA-dsDNA junctions causes destabilization of the shorter complementary strand from the 5Ј end. This destabilization provides the ssDNA platform for the nucleation of DinG proteins from the 5Ј end, allowing them to translocate on the same strand in the 5Ј 3 3Ј polarity to unwind the opposite strand (Fig. 7B). Because MtDinG and EcDinG failed to bind and unwind duplex DNA, it is likely that thermal fraying of ends may not be sufficient for the DinG proteins to access the duplex ends for its nucleation. However, our attempts to study the destabilization of the complementary shorter strand at the ssDNA-dsDNA junction with KMnO 4 footprinting failed. This could be because T residues at the destabilized end were not accessible for KMnO4 modification, possibly due to the nucleation of DinG at the 5Ј end. Nonetheless, the ssDNA translocation and unwinding activities of helicases are determined by various factors, such as the oligomeric nature of the protein, rate of unwinding, processivity, step size, and molecules of ATP utilized (59 -63). This information is obtained by a combination of structural data, pre-steady-state kinetics, thermodynamic, and single molecule studies. Further studies are required with DinG protein to understand the detailed mechanism of its translocation and helicase functions.
Our biochemical data with replication fork structures and HJ substrates revealed novel functions of MtDinG. Unlike EcDinG, MtDinG was able to proficiently unwind replication fork structures and HJs, albeit at ϳ3-fold higher concentrations compared with the helicase activity with overhangs and flap structures. Surprisingly, although MtDinG did not exhibit binding to replication fork structures, three-way junctions, and HJ substrates (data not shown), it was able to unwind these structures efficiently. These data suggest that MtDinG can recognize replication, repair, and recombination intermediates in a structure-specific manner and unwind these structures, possibly by its dsDNA translocase activity. It is likely that MtDinG interaction with branched DNA substrates is dynamic and transient. The MtDinG K55R Walker motif mutant failed to unwind all of the substrates tested, suggesting that ATP hydrolysis is essential for the translocase as well as helicase activity of MtDinG.
In response to spontaneous or induced DNA damage, replication fork stalls and helicases have been implicated in remodeling the stalled forks to resume replication (38,40,41,64,65). Our study shows that MtDinG, but not EcDinG, unwinds replication fork structures. It is likely that MtDinG may participate in the fork reversal activity to restart the replication in vivo. Formation of "chicken foot" structures from the stalled forks is one of the mechanisms by which a cell resumes replication (40,41,65,66). E. coli RuvAB, RecG, and UvrD helicases remodel stalled replication forks to restart the replication (40,41,47,66). In mammalian cells, RecQ helicases are implicated in the recovery of stalled replication forks by their fork reversal activity (7,38,55). In the absence of RecQ, it is likely that DinG helicase could have a similar function in M. tuberculosis. Nonetheless, it will be interesting to test whether MtDinG exhibits fork rever-sal activity to generate chicken foot structures. E. coli RecQ helicase unwinds replication forks to generate gaps and to facilitate RecA loading for SOS signaling and to promote recombination at stalled forks (67,68). It will be interesting to study whether MtDinG participates in DNA damage signaling through remodeling the replication forks.
Unlike EcDinG, we also noted that MtDinG unwinds HJ substrates, suggesting the possible involvement of MtDinG in recombinational repair through its branch migration activity. We recently showed that MtRecG is a potent helicase that binds and unwinds HJ substrates (28). MtDinG may overlap with RecG to unwind replication fork structures and HJ intermediates in vivo. However, compared with MtRecG, MtDinG displayed reduced activity with HJ substrates. E. coli SSB modulates the activity of RecQ, PriA, RecG, and DinG helicases (69 -72). It is possible that MtDinG may require stimulation by SSB for its HJ branch migration activity.
E. coli RecQ and its orthologs unwind G4 DNA structures (10 -12), and a recent study shows that EcDinG also resolves intermolecular G4 structures (22). Our study clearly demonstrates that MtDinG unwinds intermolecular G4 structures robustly even at 50 nM concentrations, indicating that MtDinG could be a functional ortholog of E. coli RecQ at least in part with G4 DNA unwinding activity. Moreover, our data revealed that, unlike EcDinG, MtDinG proficiently unwound G4 structures with both 5Ј and 3Ј tails. Recent studies reveal that S. cerevisiae as well as human Pif1 helicases also resolve G4 structures (73,74). Notably, many prokaryotes that encode the Pif1 family of helicases lack both RecQ and DinG (75). The M. tuberculosis genome is devoid of both RecQ-and Pif1-encoding genes. In the absence of these helicases, DinG in M. tuberculosis could be the sole enzyme that processes G4 structures in vivo. Nonetheless, in addition to G4 resolution, MtDinG was proficient in unwinding replication fork and HJ structures at a ϳ3-fold higher concentration. Conceivably, G4 structures are the more physiological substrates for MtDinG.
Targeting G4 structures that are present in the promoters of oncogenes and telomeres is actively under consideration for cancer therapy (76,77). Interestingly, the M. tuberculosis genome is 65% GC-rich with Ͼ10,000 G-rich sequences having the potential to form G4 structures. Notably, many of these G4-forming motifs are present in the promoters or in the proximity of promoters of several genes (9). Our study demonstrates the existence of G4 structures in M. tuberculosis cells. The genes that are encoded by the mce1R operon play a crucial role in the persistence of M. tuberculosis in the murine model (78). We demonstrate that G-rich sequences that are present in the promoter region of the mce1R gene indeed fold into G4 structure, suggesting that the G4 structure may regulate the expression of mce1R. The genes that encode PE_PGRS family proteins are so far found only in the genome of mycobacteria and largely restricted to pathogenic mycobacteria. These proteins have been implicated in M. tuberculosis persistence and pathogenesis (79). Our data show that G-rich promoter sequences of PE_PGRS2 folds into G4 structure. The moeB1 encodes for an enzyme that is involved in synthesis of molybdopterin, which is required as a cofactor for many enzymatic reactions (80). Interestingly, disruption of moeB1 in M. tuberculosis causes growth attenuation in macrophages (81). Our study shows that promoter sequences of the moeB1 gene also contain G4-forming sequences. These data indicate that G4 structure may play an important role in the mycobacterial gene regulation. Interestingly, a recent study shows that the G4 structure regulates expression of recF, recO, and recR genes in Deinococcus radiodurans, which has implications for its radioresistance (82). However, further studies are required to understand the mechanism of G4 DNA-mediated regulation of gene expression in M. tuberculosis and also whether such regulation may contrib- ute to virulence, persistence, and survival of this pathogen in humans.
The abundance of G4-forming sequences in M. tuberculosis provides a great opportunity for identifying inhibitors that target these structures for therapy. Indeed, M. tuberculosis H37Ra cells are sensitive to G4-specific ligand TMPyP4 and exhibit reduced growth. 6 It will be interesting to identify the genes that are essential for virulence, persistence, and survival of this pathogen in humans. Targeting such genes that contain G4 structures in their promoters could be a potential therapeutic strategy for culminating the infection with this pathogen. In addition to G4 structures, it will be interesting to identify inhibitors that specifically target DinG helicase such that G4 resolution can be potentially blocked at the genome level. Targeting G4 structures in parallel with DinG helicase using specific inhibitors can be a potential synergistic strategy for an effective killing of this pathogen in humans.