Apurinic/Apyrimidinic Endonucleases of Mycobacterium tuberculosis Protect against DNA Damage but Are Dispensable for the Growth of the Pathogen in Guinea Pigs

In host cells, Mycobacterium tuberculosis encounters an array of reactive molecules capable of damaging its genome. Non-bulky DNA lesions are the most common damages produced on the exposure of the pathogen to reactive species and base excision repair (BER) pathway is involved in the repair of such damage. During BER, apurinic/apyrimidinic (AP) endonuclease enzymes repair the abasic sites that are generated after spontaneous DNA base loss or by the action of DNA glycosylases, which if left unrepaired lead to inhibition of replication and transcription. However, the role of AP endonucleases in imparting protection against DNA damage and in the growth and pathogenesis of M.tuberculosis has not yet been elucidated. To demonstrate the biological significance of these enzymes in M.tuberculosis, it would be desirable to disrupt the relevant genes and evaluate the resulting mutants for their ability to grow in the host and cause disease. In this study, we have generated M.tuberculosis mutants of the base excision repair (BER) system, disrupted in either one (MtbΔend or MtbΔxthA) or both the AP endonucleases (MtbΔendΔxthA). We demonstrate that these genes are crucial for bacteria to withstand alkylation and oxidative stress in vitro. In addition, the mutant disrupted in both the AP endonucleases (MtbΔendΔxthA) exhibited a significant reduction in its ability to survive inside human macrophages. However, infection of guinea pigs with either MtbΔend or MtbΔxthA or MtbΔendΔxthA resulted in the similar bacillary load and pathological damage in the organs as observed in the case of infection with wild-type M.tuberculosis. The implications of these observations are discussed.


Introduction
Despite the potentially intimidating environment of the macrophage, M.tuberculosis persists in the host as a result of several defense mechanisms [1,2,3,4,5,6]. The repair of damaged DNA has been considered as an important mechanism for the survival of M.tuberculosis in the host [7]. The damages, that frequently occur to DNA as a consequence of reactive oxygen intermediates (ROI) and reactive nitrogen intermediates (RNI) produced by the macrophages, include the base modifications, generation of abasic sites and DNA strand breaks. The DNA damaged in such a manner represents the most common substrate for the Base excision repair (BER) pathway [8]. BER pathway is initiated by DNA glycosylases, a highly specialized class of enzymes, that specifically recognize and excise modified bases in DNA by hydrolyzing the N-glycosidic bond between the base and the sugar [9]. This step leads to the creation of abasic (also known as apurinic/apyrimidinic or AP) sites [9]. Abasic sites can also arise in DNA spontaneously [10]. The accumulation of AP sites in DNA is detrimental as they daunt essential processes such as replication and transcription [10]. For this reason, class II AP endonucleases are considered important enzymes that cleave the phosphodiester backbone on the 59 end of the AP site leaving a 39-hydroxyl group. DNA repair is completed by the actions of a DNA polymerase that fills in new base and DNA ligase that finally seals the gap [11,12].
AP endonucleases have been classified into two families, the exonuclease III (ExoIII or Xth) and endonuclease IV (EndoIV or Nfo) families, based on their homology to the two Escherichia coli enzymes. In E.coli, XthA, which is the major AP endonuclease, represents 90% of the cellular AP endonuclease activity, while Nfo accounts for the remaining 10% activity [13]. In addition to AP endonuclease and 39R59 exonuclease activities, the E.coli AP endonucleases also exhibit additional 39 phosphatase and 39 phosphodiesterase activities that are responsible for removing a multitude of blocking groups, including 39 phosphate and 39 phosphoglycolate, that are present at single-stranded breaks in DNA, induced by oxidative agents [14,15]. Saccharomyces cerevisiae also possesses two AP endonucleases, the Apn1 and Apn2 proteins that represent the EndoIV and the ExoIII family, respectively [16]. However, the major AP endonuclease in this organism is Apn1, that exhibits a strong AP endonuclease activity in yeast cells while the Apn2 protein, is a weak AP endonuclease that exhibits strong 39R59 exonuclease and 39 phosphodiesterase activities [16,17,18,19]. The human AP endonucleases, Ape1 and Ape2, are both members of the ExoIII family where Ape1 is the major human AP endonuclease. The EndoIV homologs are not known to be present in humans [20]. Although neither of the two AP endonuclease genes is universal, all species encode at least one of these genes, suggesting that AP endonuclease activity is required for all species [21].
E.coli mutant deficient in both the AP endonuclease genes (nfo and xth) have been demonstrated to display increased lethality to ionizing radiation and chemical oxidants implicating Nfo and Xth in the repair of such damages [22]. AP endonucleases have been evaluated for their importance in the pathogenesis of several bacteria such as Salmonella typhimurium, Neisseria meningitides and Brucella abortus [23,24,25]. To determine whether the repair of oxidatively damaged DNA is involved in the growth and pathogenesis of S.typhimurium, mutants of the AP endonuclease genes were generated that were deficient in one (Dxth, Dnfo) or both (Dxth/nfo) the AP endonucleases [23]. S.typhimurium Dxth and Dxth/nfo were significantly impaired for survival in RAW 264.7 murine macrophages and in C57BL/6 primary murine macrophages activated with IFN-c [23]. In addition, Dxth/nfo was 12fold attenuated when compared with the wild type in the murine typhoid fever model [23]. Two AP endonuclease paralogues namely NApe and NExo (both belonging to the Xth family) have been identified and characterized in the human pathogen N.meningitides [24]. By employing the mutants of these genes (Dnexo, Dnape and Dnexo:nape), it has been demonstrated that these enzymes are necessary for the survival of N.meningitidis under oxidative stress [24]. In addition, the Dnexo and Dnape were recovered from the bloodstream of infected infant rats at significantly lower levels than the wild-type strain, the most significant reduction being observed in the case of double mutant (Dnexo:nape) demonstrating that both NApe and NExo are required for the full virulence of N.meningitides [24]. The Brucella abortus possesses two xthA homologs (xthA-1 and xthA-2) but have no homolog of nfo [25]. The B.abortus xthA-1 mutant exhibited increased sensitivity to oxidative and alkylation stress when compared with the parental strain [25]. However, the mutant and the parental strains displayed equivalent spleen colonization profiles in BALB/c mice through 8 weeks post-infection and equivalent intracellular survival and replication profiles in the macrophages from C57BL/6 mice [25]. These authors suggested that residual AP endonuclease activity provided by XthA-2 may be responsible for the lack of attenuation of the xthA-1 mutant in the murine model [25].
M.tuberculosis, is amongst the few bacteria to possess homologues of all known BER genes [21,26]. The sequencing of M.tuberculosis genome revealed the presence of E.coli AP endonuclease homologs-XthA and Nfo namely, Exonuclease III (XthA) and Endonuclease IV (End) that are encoded by the genes xthA (Rv0427c) and end (Rv0670), respectively [27]. The biological importance of xthA in M.tuberculosis is highlighted by the fact that no variations in xthA have been observed in clinical strains [28]. Furthermore, a second AP endonuclease gene, encoding the endonuclease IV (End), is present in M.tuberculosis [27]. We have previously carried out biochemical characterization of the annotated M.tuberculosis AP endonucleases, namely Endonuclease IV (End) and Exonuclease III (XthA) and demonstrated that these proteins are functional AP endonucleases [29]. Given the importance of AP endonucleases in other bacteria, these proteins appear to be good targets to design anti-tubercular molecules.
Besides, in a study by Sassetti and Rubin, these AP endonucleases have been implicated in the in vivo growth of the pathogen in mouse model of infection when the spleen colonization profiles of mice infected intravenously with an M.tuberculosis transposon mutant library were evaluated [30]. However, as suggested by the authors, this study had several limitations [30]. For example, the study employed a library of transposon mutants in M.tuberculosis to intravenously infect C57BL/6J mice, hence, the effect of single gene mutation in M.tuberculosis was not evaluated in isolation [30]. The phenotypes of individual AP endonuclease mutants observed in this study could have been influenced by the presence of fully virulent bacteria that predominated the pool which could have led to an accentuation of mutant phenotypes caused by competition. Besides, this study employed intravenous route of infection rather than the aerosol route which is the common route of infection for M.tuberculosis. Moreover, the animal model employed for these experiments may also have influenced the results of the identification of in vivo essential genes as the disease in mice differs markedly from human illness. Hence, though End and XthA seem to be important drug targets, these proteins need to be validated for their in vivo essentiality in a relevant model of experimental tuberculosis before their addition to the drug development pipeline. In the present study, we have deleted the AP endonuclease gene(s)-end or/and xthA from the M.tuberculosis genome to evaluate their biological significance to this pathogen and describe the phenotypes of these mutants in the context of their importance in defense against DNA damage, intracellular growth in macrophages and in the pathogenesis of M.tuberculosis by employing the guinea pig model of experimental tuberculosis.

Bacterial strains and growth conditions
The details of bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli strains XL-1 Blue (Stratagene, Heidelberg, Germany) and HB101 (Life Technologies, CA, USA) were used for cloning and were grown in Luria-Berteni (LB) broth or on LB agar. Mycobacterial strains were grown on Middlebrook (MB) 7H11 agar supplemented with 10% OADC (oleic-acid albumin dextrose catalase) and 0.2% glycerol or in MB7H9 broth supplemented with 10% ADC (albumin dextrose catalase), 0.2% glycerol and 0.05% Tween 80 at 37uC with shaking at 200 rpm. For the generation of mutants, M.tuberculosis H37Rv transformed with pJV53 was employed, as described previously [31]. Kanamycin and chloramphenicol were used at concentrations of 25 mg/ ml and 30 mg/ml, respectively. Hygromycin was used at a concentration of 50 mg/ml for mycobacteria or at 150 mg/ml for E.coli.

Disruption of AP endonuclease genes and genetic complementation of the mutants
Oligonucleotides (Table 2) were designed to amplify (i) a ,700 bp amplicon (amplicon I) comprising of ,200 bp of the 59 proximal region of the gene (end or xthA) in addition to a ,500 bp region immediate upstream to it, and (ii) a ,700 bp amplicon (amplicon II) comprising of ,200 bp of the 39 distal region of the gene (end or xthA) in addition to a ,500 bp region immediate downstream to it. The amplicons I and II, specific for end, were PCR amplified by employing the oligonucleotides End-F1/End-R1 and End-F2/End-R2, respectively and cloned into the vector pYUB854 flanking the hygromycin cassette at XhoI/SpeI and KpnI/ XbaI, respectively, for the generation of pYUBDend. For the generation of pYUBDxthA the amplicons I and II, specific for xthA, were PCR amplified by employing the oligonucleotides XthA-F1/ XthA-R1 and XthA-F2/XthA-R2, respectively and cloned into the vector pYUB854 flanking the hygromycin cassette at KpnI/ XbaI and XhoI/SpeI, respectively. The linear allelic exchange substrates (AES), end::hyg and xthA::hyg were excised from pYUBDend and pYUBDxthA, respectively, by using the restriction enzymes KpnI/SpeI and electroporated into wild-type M.tuberculosis carrying pJV53 separately, to generate the end (MtbDend) and xthA (MtbDxthA) mutants of M.tuberculosis [31]. For the generation of a double mutant of M.tuberculosis, xthA was disrupted in MtbDend. Briefly, a hygromycin resistance cassette in pYUBDxthA was removed by using XbaI and XhoI and was replaced with the chloramphenicol resistance gene expressed under the M.tuberculosis ribosomal RNA promoter to generate pYUBDCATxthA. This vector was digested with KpnI/SpeI and the resulting AES (DxthA::CATtrrn) was electroporated into MtbDend to generate the double mutant of M.tuberculosis (MtbDendDxthA). For the generation of mutants by employing recombineering method, wild-type M.tuberculosis carrying pJV53 (for the generation of MtbDend and MtbDxthA) and MtbDend (for the generation of MtbDendDxthA) were cultured separately in MB7H9, 4% OADC, 0.05% Tween 80 and 25 mg/ml kanamycin at 37uC with shaking at 200 rpm, sub-cultured in MB7H9, 0.05% Tween 80, 25 mg/ml kanamycin and 0.2% succinate, grown to mid-log phase and induced by the addition of 0.2% acetamide. After growth for 48 h at 37uC, electrocompetent cells were prepared and transformed with AES.
To facilitate genetic complementation, a 1233 bp PCR product (proend) encompassing end gene (759 bp) along with its promoter region (474 bp) was PCR amplified by using M.tuberculosis H37Rv genomic DNA as template by employing the oligonucleotides End-F3/End-R3. The above PCR product was cloned at EcoRV site in pLITMUS38 plasmid and the clone was confirmed by DNA sequencing. proend was excised out from plit38.proend by using BglII, end-repaired by Klenow polymerase and cloned in pVR1 plasmid (digested by using MluI/XbaI and end-repaired by Klenow polymerase) [32]. The resulting plasmid pVRend, was electroporated into MtbDend to generate the complemented strain, MtbDendComp. For the complementation of MtbDxthA mutant, two separate PCR products were amplified-a 876 bp region of xthA gene was PCR amplified by employing oligonucleotides XthA-F3/XthA-R3 and a 441 bp promoter region of xthA (immediately upstream to Rv0429c) was PCR amplified by employing oligonucleotides XthA-F4/XthA-R4. By employing these two PCR products as template, PCR was performed by employing the oligonucleotides XthA-F4/XthA-R3 by the overlap extension method as described previously [33]. The resulting PCR product, proxthA (1317 bp) was digested with EcoRV and cloned at EcoRV site in pLITMUS38 plasmid and the clone was confirmed by DNA sequencing. proxthA was excised out from plit38.proxthA by using EcoRV and cloned in pVR1 plasmid (digested by using MluI/XbaI and end-repaired by Klenow polymerase). The resulting plasmid pVRxthA, was electroporated into MtbDxthA to generate the complemented strain, MtbDxthA-Comp.

Sensitivity to DNA damaging agents
To evaluate the ability of the AP endonuclease mutants to withstand DNA damage, the wild-type and the mutant strains were subjected to different concentrations of DNA damaging agents, such as methyl methane sulfonate (MMS), hydrogen peroxide (H 2 O 2 ) and mitomycin C (MMC) for 24 h and the effect of the deletion of the gene(s) on the survival of M.tuberculosis was evaluated by CFU enumeration. Briefly, early log phase cultures of MtbWT, MtbDend, MtbDxthA and MtbDendDxthA were diluted to an A 600 nm of 0.1. The cultures were pelleted and the cells were washed once with equal volume of PBS (pH 7.4) and resuspended in MB7H9 media (without OADC supplementation) containing the desired concentration of stress agent. Each M.tuberculosis strain was also resuspended in MB7H9 media alone (without any stress agent) which was used as a control for that strain. The cultures were incubated at 37uC for 24 h. Based on the earlier studies on these DNA damaging agents in other bacterial systems such as E.coli and S.typhimurium [13,34,35], the initial standardization experiments were carried out in M.tuberculosis by employing a broad range of concentration in order to select the final concentration range for the studies. After a 24 h exposure, cells were diluted and plated on MB7H11 supplemented with OADC. Colonies were counted after 3 weeks of incubation at 37uC. The survival of M.tuberculosis strains was expressed as percent survival by using the formula: Percent survival = 1006(Number of bacteria obtained post stress/Number of bacteria obtained without addition of stress agent). Experiment was performed in duplicates and was repeated twice.

Intracellular growth in human THP-1 macrophages
Human monocytic THP-1 cells were cultured in RPMI-GlutaMAX medium [containing 10% heat inactivated fetal bovine serum (FBS) and 1% antibiotic-antimycotic mix] (Gibco Invitrogen Life Technologies, NY, USA) at 37uC in 5% CO 2 and infected as described previously [36]. Briefly, THP-1 macrophages were seeded at 5610 5 cells per well in 24-well tissue culture plates and differentiated to macrophages by using 30 nM phorbol 12myristate 13-acetate (PMA) (Sigma, MO, USA) for 16 h at 37uC in 5% CO 2 . Cells were washed with RPMI medium and rested for 2 h before infection in 1 ml RPMI medium supplemented with 10% FBS. The monolayers were infected with wild-type M.tuberculosis, MtbDend, MtbDxthA and MtbDendDxthA separately, at a multiplicity of infection of 1 bacterium per 5 THP-1 macrophages for 4 h at 37uC in triplicates, following which the monolayers were washed twice with the medium. Subsequently, extracellular bacteria were removed by treatment with 200 mg/ml amikacin for 2 h at 37uC. On days 0 (6 h), 2, 4 and 6, the infected macrophage monolayers (three wells per strain) were lysed with 1 ml of 0.025% SDS (Sigma, MO, USA) to release intracellular mycobacteria, which were then enumerated by plating serial dilutions on MB7H11 agar. Colonies were counted after 4 weeks of incubation at 37uC and the data were expressed as CFU/ml. The extent of THP-1 infection by M.tuberculosis strains was calculated by employing the formula21006(Number of intracellular bacteria obtained on Day 0 (i.e. 6 h post infection) per well/ Number of THP-1 macrophages seeded per well).

In vivo guinea pig experiments
Pathogen-free out-bred female guinea pigs of the Duncan-Hartley strain in the weight range of 250 to 350 g were obtained from the Disease Free Small Animal House Facility, Chaudhary Charan Singh Haryana Agricultural University, Hisar, India. The animals were maintained in a biosafety level 3 facility and routinely cared for according to the guidelines of the CPCSEA (Committee for the Purpose of Control and Supervision on Experiments on Animals), India. To study the influence of end or/ and xthA disruption on the growth and pathogenesis of M.tuberculosis, guinea pigs were infected by the aerosol route with 10 to 30 bacilli of either wild-type M.tuberculosis, MtbDend, MtbDendComp, MtbDxthA, MtbDxthAComp or MtbDendDxthA. Animals (n = 5) were euthanized at 4 weeks and 10 weeks post-infection by CO 2 asphyxiation. After dissecting the animals, lungs, liver and spleen were scored for gross pathological damage such as the extent of involvement of the organ, number and size of tubercles, areas of inflammation and damage due to necrosis. The gross pathological scores were graded from 1-4 based on the modified Mitchison scoring system [37,38]. Left caudal lung lobe and caudal segment of spleen from the infected animals were aseptically removed for bacterial enumeration. The specific segments of lung and spleen were weighed and homogenized separately in 5 ml saline by using a polytron homogenizer. Appropriate dilutions of the homogenates were plated on MB7H11 agar plates in duplicates and incubated at 37uC for 3-4 weeks. Colonies were counted and expressed as mean log 10 CFU/organ. For histopathological evaluation, the right lung and a portion of left dorsal lobe of liver from the infected animals were removed and fixed in 10% buffered formalin. 5 mm thick sections from the formalin fixed, paraffin embedded tissues were stained with haematoxylin and eosin (H&E). The tissues were coded and the coded samples were analysed by a certified pathologist having no knowledge of the experimental groups.

Statistical analysis
For comparing the i) sensitivity of M.tuberculosis strains against DNA damaging agents and ii) growth of mycobacterial strains in THP-1 cells, two-way analysis of variance (ANOVA) with the Bonferroni multiple comparison test was employed. For the comparison of bacillary load in the lungs or spleen of infected guinea pigs, one-way ANOVA with the Tukey post test was employed. For comparison of the gross pathological scores of various groups, the nonparametric Kruskal-Wallis test followed by the Mann-Whitney U test was employed. Differences were considered significant when P,0.05. For the statistical analysis and generation of graphs, Prism 5 software (version 5.01; GraphPad Software Inc., CA) was used.

Disruption of the AP endonuclease gene(s) in M.tuberculosis and characterization of the mutants
To investigate the importance of AP endonucleases in M.tuberculosis, we employed recombineering method to generate three deletion mutants of M.tuberculosis lacking end (MtbDend), xthA (MtbD xthA) and both the genes (MtbDendDxthA) (Fig. 1A) [31]. The frequency of mutant isolation was observed to range between 63% and 88% and was calculated by using the formula2 1006(Number of legitimate recombinants confirmed by PCR)/ (Number of transformants screened). The mutations were confirmed by three approaches. (i) We confirmed all the deletion mutations by PCR analysis. Primers employed for the PCR confirmation were designed 100 bp external to the linear AES and are indicated as blue arrows in figure 1A ( Table 2). To confirm the disruption of end, we carried out PCR amplification by employing the oligonucleotides End-F and End-R (Fig. 1B). DNA from wildtype M.tuberculosis yielded a PCR product of 1.90 kb while the DNA from MtbDend or MtbDendDxthA resulted in a PCR product of 3.48 kb corresponding to replacement of 381 bp region in end by the hygromycin resistance cassette (1.95 kb) (Fig. 1B). Similarly, to confirm the disruption of xthA, PCR amplifications were carried out by employing the oligonucleotides XthA-F and XthA-R (Fig. 1C). DNA from wild-type M.tuberculosis yielded a PCR product of 2.0 kb. The DNA from MtbDxthA yielded a PCR product of 3.41 kb corresponding to replacement of 554 bp region in xthA by the hygromycin resistance cassette. The DNA from MtbDendDxthA resulted in a PCR product of 2.50 kb, corresponding to replacement of 554 bp region of xthA by chloramphenicol resistance gene expressed under the M.tuberculosis ribosomal RNA promoter (1.05 kb) (Fig. 1C). (ii) The above amplification products were subjected to DNA sequencing that further confirmed the disruption of end in MtbDend, xthA in MtbDxthA and both the genes in MtbDendDxthA. (iii) The disruption of AP endonuclease genes was further confirmed by immunoblot analysis by employing polyclonal antibodies raised against the respective proteins ( Fig. 1D  and 1E). In the case of cell free protein extracts from wild-type M.tuberculosis, the presence of 27 kDa and 32 kDa bands confirmed the presence of End (Fig. 1D) and XthA (Fig. 1E), respectively. These End and XthA bands (27 kDa and 32 kDa, respectively) were absent in the immunoblot analysis carried out by employing the cell free protein extracts from the MtbDend and MtbDxthA, respectively, while either protein was absent in the cell free protein extract of MtbDendDxthA. Complementation of end in the case of MtbDendComp or xthA in the case of MtbDxthAComp was confirmed by the restoration of 27 kDa End band or 32 kDa XthA band, respectively, when the cell free protein extracts from these strains were subjected to immunoblot analysis. When the growth characteristics of M.tuberculosis strains were assessed under standard culture conditions in MB7H9 media, the AP endonuclease mutants displayed no significant difference in growth when compared with wild-type M.tuberculosis (data not shown). Thus, end and xthA are dispensable for the in vitro growth of M.tuberculosis.
Influence of the disruption of AP endonuclease genes on the susceptibility of M.tuberculosis to DNA damaging agents The influence of disruption of AP endonuclease(s) on the ability of M.tuberculosis to withstand alkylation stress was measured by exposure to methymethane sulfonate (MMS) (Figure 2A). At a concentration of 0.5 mM, MMS displayed no effect on the survival of any of the M.tuberculosis strains. However, when the concentration of MMS was further increased, the differences between the single gene mutants and double gene mutant became apparent. At a concentration of 1 mM MMS, the survival of MtbWT (75.63%), MtbDend (74.39%) and MtbDxthA (74.49%) strains was comparable, suggesting that the function of any of these two enzymes can be compensated by the other under these experimental conditions. However, a clear indication of the importance of these enzymes became apparent when the sensitivity of MtbDendDxthA towards MMS was evaluated. The double gene mutant exhibited a significantly enhanced sensitivity to the toxic effect of MMS at a concentration of 1 mM as demonstrated by a significantly reduced survival (35%), when compared with the wild-type strain, indicating thereby that these AP endonucleases play important role in protecting the pathogen against the alkylation damage (Figure 2A). Similarly, at a concentration of 2 mM MMS, while the survival of the parental (60.63%), MtbDend (54.39%) and MtbDxthA (62.49%) strains was comparable, the survival of MtbDendDxthA significantly reduced by 100 folds (0.63%) when compared with the parental strain ( Figure 2A). These results signify an important role of AP endonucleases in withstanding alkylation stress in M.tuberculosis.
The influence of disruption of AP endonuclease(s) on the ability of M.tuberculosis to withstand oxidative stress was measured by exposure to hydrogen peroxide (H 2 O 2 ) (Figure 2B). At a concentration of 0.5 mM H 2 O 2 , the survival of MtbWT and MtbDxthA was 79.64% and 92.91%, respectively, which was not significantly different from each other. However, the survival of MtbDend (53.84%) was significantly reduced in comparison to the parental strain. These observations indicate that at this concentration of H 2 O 2 , End protects bacteria from oxidative stress whereas XthA does not. However, at this concentration of H 2 O 2 , the highest sensitivity to peroxide damage was exhibited by the double gene mutant i.e. MtbDendDxthA which exhibited a significant reduction in its survival (34.57%) in comparison to the parental strain indicating the importance of AP endonucleases in protecting the pathogen against oxidative damage ( Figure 2B). Similarly, at a concentration of 1 mM H 2 O 2 , it was observed that the survival of MtbWT (70.66%) and MtbDxthA (75.27%) was not significantly different from each other, while the survival of MtbDend (39.08%) was significantly reduced in comparison to the wild-type strain. The double mutant was maximally sensitive at this concentration with a significantly reduced survival of 6.85% in comparison to the wild-type strain. At a concentration of 2 mM H 2 O 2 , all the strains exhibited a markedly reduced survival with no statistical differences in their sensitivity to peroxide stress. Taken together, these observations point out that while both the AP endonucleases are important in protecting M.tuberculosis from the damage caused by oxidative radicals, their contribution may not necessarily be comparable and End appears to be playing a more important role than XthA because of its higher capacity to withstand oxidative damage.
The AP endonuclease mutants were also examined for their sensitivity against another DNA damaging agent, mitomycin C (MMC) along with the wild-type strain ( Figure 2C). The survival of Similarly, the spleens of the guinea pigs infected with various M.tuberculosis strains, exhibited a comparable bacillary load (5.7 to 6.4 log 10 CFU) (Fig. 4A). Similarly, at 10 weeks post-infection also, the bacillary load in the lungs of animals infected with any of the M.tuberculosis strains was comparable and corresponded to a range of 6.2 to 7.2 log 10 CFU, while the bacillary load in the spleens of these animals corresponded to a range of 5.8 to 7.1 log 10 CFU (Fig. 4B). This data indicates that the absence of AP endonucleases does not influence the growth of M.tuberculosis in the guinea pig infection model.

Discussion
In the host, Mycobacterium tuberculosis is exposed to an environment that is rich in reactive oxygen and nitrogen intermediates capable of damaging the genome of the pathogen [39,40]. It has been thought that the BER pathway, which repairs the cytotoxic and mutagenic AP sites in DNA, may play a major role in maintaining the integrity of DNA in mycobacteria [26,41,42]. However, the experimental evidence gathered by us in this study suggests that though under in vitro conditions, the AP endonucle- In response to the alkylating agent MMS, AP sites are generated via spontaneous and enzymatic hydrolysis of glycosidic bonds [22]. Our results demonstrate that End and XthA are able to compensate the absence of each other to repair the DNA damaged in response to alkylation stress. However, a simultaneous disruption of both the AP endonucleases in M.tuberculosis (MtbDendDxthA) significantly reduces the ability of the pathogen to withstand alkylation stress when compared with wild-type M.tuberculosis, thereby indicating the importance of AP endonucleases in protecting the pathogen against alkylation damage. In the previous studies, it has been demonstrated that the AP endonuclease mutants of E.coli and B.abortus exhibit a significantly enhanced sensitivity to MMS when compared with the parental strain, indicating the involvement of AP endonucleases in the repair of MMS induced DNA lesions [22,25]. H 2 O 2 generates nicks and breaks in DNA with blocked 39 termini apart from giving rise to abasic sites [43,44]. Our results on the role of these AP endonucleases against oxidative damage by employing H 2 O 2 demonstrate that while both the AP endonucleases make an important contribution in the repair of DNA damage in M.tuberculosis, End plays a more important role than XthA in protecting this pathogen against oxidative assault.
The intra-and interstrand cross-linking of DNA resulting from the damage caused by MMC is repaired by nucleotide excision and recombination repair pathways rather than BER in bacteria [45]. However, based on microarray and quantitative RT-PCR analysis Rand et al., reported that addition of MMC to M.tuberculosis induces the expression of xthA indicating a plausible role of XthA in the repair of DNA damage induced in response to MMC [46]. This tempted us to evaluate the role of these AP endonucleases in the repair of MMC induced DNA damage. However, the experimental evidence gathered in our study does not support any apparent role of XthA or End in the repair of such damage to DNA.
Our studies demonstrate that the disruption of any one of the AP endonucleases in M.tuberculosis (MtbDend or MtbDxthA) did not affect the growth of the pathogen in THP-1 macrophages when compared with the parental strain. However, disruption of both the AP endonucleases in M.tuberculosis (MtbDendDxthA) significantly reduced the growth of the pathogen in THP-1 cell line. These observations may be attributable to the inability of M.tuberculosis to repair the DNA damage inflicted upon by ROI and RNI produced by the THP-1 macrophages, in the absence of both the AP endonucleases. However, in the guinea pig infection model, we observed that disruption of either one or both the AP endonucleases did not affect the growth of M.tuberculosis apparently indicating that these AP endonucleases may not be indispensable for the growth and pathogenesis of M.tuberculosis. Our observations were substantiated by gross pathological and histopathological damage.
The difference in the influence of mutations in vitro and in the host, we think, may be related to the fact that the autonomous in vitro response of the bacterium in the macrophages may be limited. However, in vivo, the outcome may depend upon the hostpathogen interaction resulting in the induction of gene expression that is vital for the bacterium to survive in the face of assault mounted by the host. In this context, it is important to note that M.tuberculosis has ada operon responsible for what is called the Ada response, which is an adaptive response to alkylation damage [47,48]. This ada operon encodes a composite protein of AdaA and AlkA and a separate AdaB/Ogt protein [48]. By employing an M.tuberculosis mutant lacking the ada operon and the parental strain, Durbach et. al. demonstrated that the mutant grew normally and was hypersensitive to the alkylation damage in vitro but displayed no attenuation in vivo in murine model of infection [49]. Such apparent differences in the function of the proteins responsible for Ada response, or AP endonucleases such as End or XthA in vitro and in vivo in M.tuberculosis, may be attributable to either variations in the DNA damage inflicted upon the pathogen during its growth in vitro and in the host or due to redundancy in such activities evolutionarily developed by the pathogen to safeguard its genome. The phenotype of MtbDendDxthA also resembles the phenotype of a recA mutant of M.bovis BCG, that has also been demonstrated to exhibit hypersensitivity to DNA damage in vitro but displayed no growth impairment in mice model of infection [50]. Besides, there is an increasing evidence of the existence of alternative pathways that cater to damages generally repaired by AP endonucleases and such enzymes are found across different species from bacteria to fungi [51,52,53,54].
The DNA polymerase X of B.subtilis (PolX Bs ), another gram positive organism, has been shown to possess an intrinsic AP endonuclease activity that enables it to recognize, incise, and further repair abasic sites in a AP-endonuclease-independent manner [54]. This persuaded us to search for AP endonuclease analogues in M.tuberculosis genome. We found a conserved hypothetical protein in M.tuberculosis (Rv3856c) that displayed 39% identity and 57% similarity to the C-terminal polymerase and histidinol phosphatase (PHP) domain of PolX Bs . In B.subtilis, the AP endonuclease activity of PolX Bs maps at the PHP domain [54]. Multiple sequence alignment of the PHP domain of bacterial/ archaeal PolXs has resulted in the identification of highly conserved residues assembling a catalytic core required to coordinate the metal ions important for the activity of this protein [55]. All these residues (H101, H103, H133, E172, H199, H227, E262, D292 and H294) were found to be conserved in the M.tuberculosis protein encoded by Rv3856c [56]. Based on these observations, it appears possible that M.tuberculosis may also possess repair pathways or proteins that overlap with AP endonucleases in order for it to protect its DNA from damage during its survival in the host.
We demonstrate that M.tuberculosis AP endonucleases are important in protecting the pathogen against alkylation and oxidative damage in vitro as well as for its growth in human macrophage cell line. However, M.tuberculosis may induce overlapping repair pathways and proteins to protect its DNA from damage during its survival in the host that make the role of these AP endonucleases in vivo dispensable.