d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ier .com/ locate /dnarepai r

A distinct role of formamidopyrimidine DNA glycosylase(MutM) in down-regulation of accumulation of G, Cmutations and protection against oxidativestress in mycobacteria

Ruchi Jain1, Pradeep Kumar1, Umesh Varshney ∗

Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore 560012, India

a r t i c l e i n f o

Article history:

Received 11 February 2007

Received in revised form

24 June 2007

Accepted 26 June 2007

Published on line 16 August 2007

Keywords:

Fpg

DNA polymerase

MutY

M. smegmatis

8-oxoG

a b s t r a c t

Reactive oxygen species produced as a part of cellular metabolism or environmental agent

cause a multitude of damages in cell. Oxidative damages to DNA or the free nucleotide pool

result in occurrence of 7,8-dihydro-8-oxoguanine (8-oxoG) in DNA, and failure to replace it

with the correct base results in a variety of mutations in the genome. Formamidopyrimidine

DNA glycosylase (Fpg/MutM), a functionally conserved repair enzyme initiates the 8-oxoG

repair pathway in all eubacteria. DNA in mycobacteria with G + C rich genomes is particularly

vulnerable to the oxidative damage. In this study, we disrupted fpg gene in Mycobacterium

smegmatis to generate an Fpg deficient strain. The strain showed an enhanced mutator

phenotype and susceptibility to hydrogen peroxide. Analyses of rifampicin resistance deter-

mining region (RRDR) revealed that, in contrast to Fpg deficient Escherichia coli where C to A

mutations predominate, Fpg deficient M. smegmatis shows a remarkable increase in accu-

mulation of A to G (or T to C) mutations. Interestingly, exposure of the mutant to sub-lethal

level of hydrogen peroxide results in a major shift towards C to G (or G to C) mutations.

Biochemical analysis showed that mycobacterial Fpg; and MutY (which excises misincorpo-

rated A against 8-oxoG) possess substrate specificities similar to their counterparts in E. coli.

However, the DNA polymerase assays with cell-free extracts showed preferential incorpora-

tion of G in M. smegmatis as opposed to an A in E. coli. Our studies highlight the importance

and the distinctive features of Fpg mediated DNA repair in mycobacteria.

(∼65%) are at a greater risk of these damages. Another source

1. Indroduction

Mycobacterium tuberculosis, one of the most successfulpathogens, infects the host macrophages where it is exposedto reactive oxygen and nitrogen species (ROS and RNI) pro-

duced as a part of the host’s innate immune response. ROS andRNI are highly reactive and cause severe damages to intracel-lular components such as lipids, proteins and DNA [1–3]. Some

∗ Corresponding author. Tel.: +91 80 2293 2686; fax: +91 80 2360 2697.E-mail address: varshney@mcbl.iisc.ernet.in (U. Varshney).

1 Authors contributed equally to this work.1568-7864/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.dnarep.2007.06.009

© 2007 Elsevier B.V. All rights reserved.

of the most common damages among these are deaminationof cytosine to uracil, and oxidation of guanine to 7,8-dihydro-8-oxoguanine (8-oxoG) in DNA and the free nucleotide pool[4–6]. The mycobacterial genomes with high G + C contents

of the occurrence of uracil and 8-oxoG in DNA is through uti-lization of dUTP and 8-oxoGTP present in the dNTP pool, assubstrates during DNA synthesis [7,8]. However, this is kept to

d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785 1775

Table 1 – List of plasmids and strains

Strain/plasmid Relevant details Reference

pYUB854 (hygR) A mutlicopy plasmid containing a hygromycin cassette between two muliplecloning sites.

26

pYUB-MsmfpgLF pYUB854 derivative harboring ∼1kb region (corresponding to a region upstream offpg ORF) in the muliple cloning site upstream to hygromycin cassette.

This study

pYUB-MsmfpgRF pYUB-MsmfpgLF harboring ∼1kb region (corresponding to a region downstream offpg ORF) in the multiple cloning site downstream of hygromycin cassette.

This study

pPR27 (GmR) An E. coli-mycobacteria shuttle vector containing a temperature-sensitive origin ofreplication (mycobacteria) and a sacB marker.

27

pPR-Msmfpg::hyg pPR27 harboring an insert wherein M. smegmatis fpg gene ORF has been replacedby hygromycin resistance cassette.

This study

pTKmx (KanR) A multicopy plasmid having E. coli and mycobacterial origin of replications. 32pTK-Mtufpg pTKmx containing M. tuberculosis fpg gene. This StudyM. smegmatis mc2155

(M. smegmatis)A high efficiency transformation strain of M. smegmatis. 24

M. smegmatis (fpg-) M. smegmatis mc2155 strain wherein the fpg gene has been disrupted with hygR This Study

ene h

ad

nprrlMeDcOis8tottroaaetG8Fpo

aattmoDt

face, Middlebrook 7H10 or LBT supplemented with 1.5% agarwere used. When required, the media were supplementedwith hygromycin and gentamycin, at 50 and 5 �g ml−1 con-centrations, respectively.

Table 2 – List of oligodeoxyribonucleotides

Name Sequence (5′–3′)

MsmfpgLF-Fp ACGGGTACCCCGACAGCACCAC (KpnI)MsmfpgLF-Rp ACGATCTAGACCTCGGGAAGCTC (XbaI)MsmfpgRF-Fp CGGAAGCTTCGACCGGTCGTTG (HindIII)MsmfpgRF-Rp CGCCTCGAGCGGGTATTCCC (XhoI)306-rpoB-Fp CGACCACTTC GGCAACCG306-rpoB-Rp CGATCAGACC GATGTTGGMsm-rpoB-seq-Fp GTCTGCGCAC CGTCGGTGG-oligo CGATCATGGAGCCACGAGCTCCCGTTACAG8-oxoG(*)-oligo CGATCATGGAGCCACG*AGCTCCCGTTACAGComp-G-oligo CTGTAACGGGAGCTGGTGGCTCCATGATCGComp-A-oligo CTGTAACGGGAGCTAGTGGCTCCATGATCGComp-T-oligo CTGTAACGGGAG CTTGTGGCTCCATGATCGComp-C-oligo CTGTAACGGGAGCTCGTGGCTCCATGATCG

cassetteE. coli TG1 An E. coli K strainE. coli (fpg-) E. coli TG1 strain wherein the fpg g

minimum by the action of Dut and MutT which hydrolyzeUTP and 8-oxoGTP, respectively.

Although DNA repair mechanisms in mycobacteria areot well understood, these are vital for the survival andersistence of the pathogenic species in the host, and theeactivation of the persistent bacilli [9,10]. The pathways thatepair uracil and 8-oxoG are initiated by uracil DNA glycosy-ase (Ung) and formamidopyrimidine DNA glycosylase, Fpg (or

utM), respectively, the two important members of the basexcision repair pathway [11,12]. Failure of uracil repair prior toNA replication leads to accumulation of C to T (G to A on theomplementary strand) mutations in the genome [4,8,11–13].n the other hand, failure to repair 8-oxoG, is known to result

n G to T (or C to A) mutations in E. coli because of the propen-ity of the DNA polymerases to insert A (in place of C) against-oxoG during replication [14–16]. The frequency of G to T (or Co A) mutations has been reported to increase upon treatmentf E. coli with H2O2 [17]. Notably, Fpg does not remove 8-oxoGo any significant level when paired with A [18,19]. However,he cells possess yet another DNA glycosylase, MutY whichemoves A from the 8-oxoG:A mispairs, increasing the chancesf incorporation of the correct base (C) against 8-oxoG, as wells the probability of further repair of 8-oxoG by Fpg. Thus, thection of MutY prior to second round of replication provides anfficient mechanism to decrease fixation of A in place of C inhe genome [7,14,20]. MutY also excises A when found against

or C [16,21]. Fpg, on the other hand, is also known to excise-oxoG from DNA when paired against G or T. This property ofpg helps repair misincorporation of 8-oxoG (from the dNTPool) by DNA polymerase(s). Fpg is also active against severalther modified purines [14].

Earlier studies using M. smegmatis, showed that Ung medi-ted base excision repair is important in mutation preventionnd residency of mycobacteria in macrophages [22]. Valida-ion of these observations in M. tuberculosis [23] highlightedhe significance of the use of this fast-growing nonpathogenic

ycobacterium with an established advantage of carryingut genetic manipulations, as model organism to understandNA repair mechanisms in M. tuberculosis. To understand

he significance of Fpg in dealing with the oxidative dam-

30as been disrupted with CmR cassette 34

age to DNA in mycobacteria, we developed a knockout strainof M. smegmatis deficient in Fpg activity. Our studies show adistinct mechanism of Fpg mediated DNA repair in mycobac-teria.

2. Materials and methods

2.1. Oligodeoxyribonucleotides, media and growthconditions

Bacterial strains, plasmids and DNA oligonucleotides used inthis study are listed in Tables 1 and 2. M. smegmatis mc2155 [24]was grown in LB supplemented with 0.2% Tween 80 (LBT) orMiddlebrook 7H9 (Difco) as indicated. For growth on solid sur-

Comp-14-oligo CTGTAACGGGAGCTFpKRS AGCUCGCCGACGAATATGATACAGGAG

G*: 8-oxo-7,8-dihydro-2′-deoxyguanosine, Comp-G, -A, -T, -C and -14 oligomers are complimentary to G and G*-oligo.

( 2 0

1776 d n a r e p a i r 6

2.2. Generation of M. smegmatis fpg gene knockout

Approximately 3.0 kb DNA sequence corresponding tofpg locus of M. smegmatis (Msmfpg, 858 bp ORF) wasretrieved from The Institute of Genome Research website (http://www.tigr.org) based on its homology with fpg gene(Rv2924c) of M. tuberculosis [25]. DNA encompassing ∼1 kbregion (left flank) was amplified from M. smegmatis by PCRusing a forward primer (MsmfpgLF-Fp) and a reverse primer(MsmfpgLF-Rp), complimentary to the initial 29 nucleotidesin the fpg open reading frame (ORF). PCR (50 �l) consisted of∼300 ng of the genomic DNA, 200 �M dNTPs, 20 pmol each ofMsmfpgLF-Fp and MsmfpgLF-Rp primers and 2.5 U of Taq DNApolymerase (Promega) in the supplied reaction buffer. PCRconditions included a step of initial heating at 94 ◦C for 4 min,30 cycles of incubations at 94 ◦C for 1 min, 56 ◦C for 30 s and70 ◦C for 1 min, followed by incubation at 70 ◦C for 10 min. ThePCR product (∼1 kb) was gel purified, cloned into pGEM-T-Easyvector (Promega), and then mobilized into pYUB854 vector([26], kindly provided by Dr. W.R. Jacobs Jr., Albert EinsteinCollege of Medicine, New York, USA) between KpnI and XbaIsites resulting in pYUB-MsmfpgLF construct. Similarly, theright flank of Msmfpg gene was amplified using a forwardprimer (MsmfpgRF-Fp, derived from sequence 708 bp fromstart codon of the 858 bp Msmfpg ORF) and a reverse primer(MsmfpgRF-Rp, complimentary to a region 898 nucleotidesdownstream of Msmfpg ORF). The right flank (∼1.0 kb) wasPCR amplified using conditions identical to those used forleft flank, cloned into pGEM-T-Easy vector and mobilized intopYUB-MsmfpgLF between HindIII and XhoI sites to give riseto pYUB-MsmfpgLF-RF construct. This procedure resulted ingeneration of a construct wherein ∼0.68 kb of fpg ORF wasswapped with ∼1.8 kb DNA harboring hygR cassette. The con-struct was digested with StuI and the fragment harboring hygR

gene flanked on either side by ∼1 kb sequences of the fpg locusfrom M. smegmatis (MsmfpgLF-hyg-MsmfpgRF) was mobilizedinto BamHI cut and blunt ended pPR27 vector to generatepPR-Msmfpg::hyg used to replace fpg gene in M. smegmatis[22,27]. Briefly, pPR-Msmfpg::hyg (GentamycinR, harboringa thermosensitive mycobacterial origin of replication anda sacB gene for counter-selection) was introduced into M.smegmatis mc2155 by electroporation, and the transformantswere selected at 30 ◦C on gentamycin plates. Transformantswere grown at 30 ◦C in 7H9 in presence of hygromycin.The culture was then plated on 7H10 medium containing10% sucrose and incubated at 39 ◦C. Colonies appearingon the plate were screened for allelic exchange of fpglocus.

2.3. Screening of fpg gene knockout in M. smegmatis

The genomic DNA was prepared and analyzed by PCR usingMsmfpgLF-Fp and MsmfpgRF-Rp primers. It may be noted thatthe sequences complimentary to MsmfpgRF-Rp were lost fromthe knockout construct, pPR-Msmfpg::hyg as a consequenceof its subcloning from pYUB-MsfpgLF-RF to pPR27. Hence this

primer annealed to the genomic sequences not contributed bypPR-Msmfpg::hyg. The PCR (50 �l) for screening of disruption ofchromosomal allele of fpg included 1.5 U DyNAzyme EXT DNApolymerase, 1X DyNAzyme EXT reaction buffer, 5% DMSO,

0 7 ) 1774–1785

10 pmol of each primer, 200 �M dNTPs and 200 ng genomicDNA (WT or putative fpg- isolate). After initial heating at 94 ◦Cfor 4 min, 30 cycles of incubations were carried out at 94 ◦Cfor 1 min, 56 ◦C for 45 s, and 72 ◦C for 4 min followed by 72 ◦Cfor 10 min. The PCR products were analyzed on 1% agarosegel.

2.4. Genomic blot analysis of M. smegmatis (fpg-)isolate

Genomic DNAs (∼2.5 �g) were digested with excess of restric-tion enzymes (20 U), separated on a 0.7% agarose gel using TBEbuffer system, transferred [28] to nylon membrane (BiodyneB, Pall Gelman Laboratory) and subjected to hybridization [29]with radiolabeled probe that annealed downstream of fpg ORFin M. smegmatis. The probe was prepared by PCR [30] using�-32P[dCTP] and, MsmfpgRF-Fp and MsmfpgRF-Rp primers.

2.5. Fpg activity assay

The synthetic oligomer containing 8-oxoG (10 pmol) was 5′32P end labeled, purified on Sephadex G50 spin column [30]and annealed to either Comp-C, -A, -G or -T oligomers in ∼1:2molar ratio by heating at 75 ◦C for 10 min in a buffer (10 mMTris–HCl pH 8.0, 250 mM NaCl and 1 mM Na2EDTA), followedby slow cooling [31]. The duplex substrates (20,000 cpm) wereused for assays with 10 or 20 �g of cell-free extracts fromM. smegmatis (WT), M. smegmatis (fpg-), E. coli TG1 and E. coli(fpg-) in 20 �l reactions containing 70 mM HEPES pH 7.5, 2 mMNa2EDTA and 100 mM KCl. Reactions were incubated at 37 ◦Cfor 30 min, extracted with chloroform, vacuum dried, takenup in 20 �l of 80% formamide dye, heated at 90 ◦C for 5 minand analyzed on 15% polyacrylamide gel containing 8 M urea[30] and visualized by BioImage Analyzer (FLA 2000, Fuji Film,Japan).

2.6. Generation of M. smegmatis (fpg-) straincomplemented with extrachromosomal fpg

M. tuberculosis fpg gene (Rv2924c) was isolated fromMTCY338.13C cosmid (kindly provided by Dr. S. Cole, Pas-teure Institute, Paris) by PstI digestion as ∼2.0 kb fragmentand subcloned into PstI site of a pTrc99c derivative to gen-erate pTrc-Mtufpg, released from it by EcoRI and HindIIIdigestion, and mobilized into the same sites of pTKmx [32]to generate pTK-Mtufpg. The pTK-Mtufpg was introducedinto M. smegmatis (fpg-) by electroporation, and Fpg expres-sion in the transformants was confirmed by gain of Fpgactivity.

2.7. Assay of susceptibility to hydrogen peroxide

Isolated colonies of various M. smegmatis strains from the7H10-agar plates were grown for 48 h in 7H9 broth media con-taining appropriate antibiotic. The cultures were inoculatedfor growth curve experiment with 1% of primary culture in

50 ml LBT (without antibiotics) under shaking at 37 ◦C. After 6 hof growth, hydrogen peroxide was added and the growth wasmonitored by taking absorbance (595 nm) of samples (1 ml)drawn at regular time intervals.

0 0 7

2

Ipmwrdcva

2r

Mecranat2hbfwB(

2e

T(w3

se8fdMA2lu

2

Cemc2t2nt

d n a r e p a i r 6 ( 2

.8. Determination of mutator phenotype

solated colonies of M. smegmatis strains from LBT-agarlates containing suitable antibiotic were inoculated in brothedia and grown for 48 h. The cells from 1 ml culturesere spread on solid media (100 �l × 10 plates) containing

ifampicin (50 �g ml−1). Total viable counts in the culture wereetermined by dilution plating. Mutation frequencies werealculated by dividing the number of RifR colonies by the totaliable counts of the bacteria plated. Number of replicates wass indicated in the legends.

.9. Analysis of rifampicin resistance determiningegion (RRDR)

. smegmatis cultures grown for 48 h (in the absence or pres-nce of 2 mM hydrogen peroxide) were plated on LBT-agarontaining rifampicin (50 �g ml−1). The isolated colonies wereesuspended in 20 �l water, incubated at 90 ◦C for 5 min, spunt 13,000 rpm for 5 min in a table-top centrifuge and the super-atant used as template to amplify RRDR using 306-rpoB-Fpnd 306-rpoB-Rp primers. Reactions (50 �l) containing 20 �lemplate DNA, 2.5 U Pfu DNA polymerase, 1× reaction buffer,00 �M dNTPs, 5% DMSO and 20 pmol of each primer wereeated at 94 ◦C for 4 min and subjected to 30 cycles of incu-ations at 94 ◦C for 1 min, 55 ◦C for 30 s, and 70 ◦C for 1 minollowed by incubation at 70 ◦C for 10 min. The PCR productsere analyzed on 1% agarose gel and eluted (Geneclean II Kit,IO101). DNA sequence analysis was done by Macrogen Inc.

Seoul, S. Korea), using Msm-rpoB-seq-Fp.

.10. DNA polymerase activity assay in cell-freextracts

he cell-free extracts from M. smegmatis (fpg-) and E. colifpg-) strains were prepared as described [33]. The substrateas prepared by annealing 8-oxoG or G-oligomers with 5′

2P end labeled Comp-14, and used to verify the propen-ity of DNA polymerase in M. smegmatis and E. coli cell-freextracts to incorporate various deoxyribonucleotides against-oxoG residues. The reaction (40 �l) contained 5 �g of cell-ree extract, 50,000 cpm of labeled duplex, 1 pmol of unlabeleduplex, 100 �M dNTP(s) in 10 mM Tris–HCl (pH 7.9), 10 mMgCl2, 50 mM NaCl and 1 mM DTT, and incubated at 37 ◦C.liquots (10 �l) were drawn at indicated times and mixed with0 �l 80% formamide dye, heated at 90 ◦C for 5 min and ana-yzed on 15% polyacrylamide sequencing gels containing 8 Mrea and visualized by BioImage Analyzer (FLA 2000, Fuji Film).

.11. MutY activity assay

omp-G, -A, -T and –C oligomers (10 pmol each) were 5′ 32Pnd labeled, purified and annealed to 8-oxoG-oligomer in ∼1:2olar ratio by heating at 75 ◦C for 10 min, followed by slow

ooling. Such duplexes (20,000 cpm) were used in assays with5 �g of purified M. smegmatis cell-free extract in 20 �l reac-

ions containing 20 mM Tris–HCl (pH 7.6), 1 mM Na2EDTA,0 mM NaCl, 100 �g ml−1 BSA and 2 pmol FpKRS oligomer asonspecific DNA. Reactions were incubated at 37 ◦C for 30 min,erminated by addition of 5 �l of 0.4 M NaOH, extracted with

) 1774–1785 1777

chloroform and vacuum dried, taken up in 20 �l of 80% for-mamide dye, heated at 90 ◦C for 5 min and analyzed on 15%polyacrylamide gel containing 8 M urea and visualized byBioImage Analyzer (FLA 2000, Fuji Film).

3. Results

3.1. Generation and characterization of fpg geneknockout in M. smegmatis

To generate fpg gene knockout in M. smegmatis, we replacedessentially entire ORF of fpg with hygR cassette (Fig. 1A and B)and introduced the disrupted fpg gene into the chromosomeby allelic exchange method employing the pPR-Msmfpg::hygconstruct (Table 1). The PCR screening showed products of∼2.7 and 4.0 kb sizes, respectively, from the wild type andthe knockout strains (Fig. 1C, lanes 2 and 3). This observationagreed with the expected pattern (Fig. 1A and B, lower panels)and indicated a successful fpg gene knockout. Genomic blotanalysis resulted in detection of ∼2.9 and 4.8 kb size bandswith FspI (Fig. 1D, lanes 1 and 2) and ∼2.2 kb and 2.1 kb sizebands with Ecl136II (lanes 3 and 4) digests of the wild type andknockout strains, respectively. Agreement of these sizes withthe expected pattern (Fig. 1A and B) confirmed the authentic-ity of the fpg gene loci in the wild type and knockout strainsof M. smegmatis. Furthermore, while the cell-free extracts pre-pared from the wild type strain resulted in processing of the30-mer duplex harboring an 8-oxoG lesion (Fig. 1E, lanes 1 and2) same as the control reaction with E. coli Fpg (lane 11), thisactivity was largely missing from the extracts prepared fromthe knockout strain (lanes 3 and 4). The activity in the mixedextracts of the wild type and the knockout strains (lane 9) ruledout the presence of inhibitors in the extracts of the knock-out strain. Taken together, these experiments confirmed thatthe replacement of fpg ORF with hygR cassette resulted in Fpgdeficiency and that the major activity of excision of 8-oxoGwas encoded by the fpg. Similar observations with the extractsprepared from E. coli TG1 and the corresponding fpg knockoutstrain (lanes 5–8, and 10) [34] confirmed the desired phenotypeof the E. coli strains used as control.

3.2. Effect of treatment with H2O2 on survival of M.smegmatis

To assess the contribution of Fpg towards protection of M.smegmatis from the damages inflicted on DNA by oxidativestress such as upon treatment of the cells with H2O2, weanalyzed growth of M. smegmatis (harboring pTKmx), and itsderivative M. smegmatis (fpg-) harboring the pTKmx plasmidor pTK-Mtufpg containing an extrachromosomal copy of fpggene from M. tuberculosis. As seen in Fig. 2, in the absence ofany added H2O2 (Fig. 2A), all the strains followed through sim-ilar lag, exponential and stationary phases of growth. Uponaddition of H2O2 to 3 mM, while the wild type strain showeda slightly extended lag phase prior to entering into log and

stationary phases, the fpg-strain did not show any significantgrowth. However, the presence of pTK-Mtufpg complementedthe strain to show growth kinetics similar to that of the wildtype strain even in the presence of H2O2 (Fig. 2B). Further-

1778 d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785

Fig. 1 – Generation and characterization of M. smegmatis (fpg-) strain. Schematic diagrams of wild type, WT (A) and thedisrupted (B) fpg loci. The expected sizes of PCR products and the sites for the restriction enzymes used in this analysis areas shown. (C) Agarose gel analysis of PCR products obtained from the wild type (WT) and the knockout (fpg-) strains. Lane Mshows � DNA HindII/HindIII size markers. (D) Genomic blot of M. smegmatis strains. Lanes: 1 and 3, M. smegmatis (WT); 2and 4, M. smegmatis (fpg-). The right flank of fpg gene was labeled with �-32P[dCTP] by PCR using MsmfpgRF-Fp andMsmfpgRF-Rp (Table 1) primers used for genomic blot analysis. Sizes of DNA bands are as indicated. (E) Fpg activity assayson duplex DNA containing 8-oxoG:C pair. Fpg assays were performed using indicated amounts of cell extracts of M.smegmtis (lanes 1 and 2, WT; and 3 and 4, fpg-) and E. coli (lanes 5 and 6, WT; and 7 and 8, fpg-) and analyzed on 15% PAGEcontaining 8 M urea. M (lane 9), assay with a mixture of cell-free extracts of M. smegmatis WT and fpg-strains (10 �g each). E(lane 10), assays with a mixture of cell-free extracts from E. coli WT and fpg-strains (10 �g each). Lanes 11 and 12 arecontrols wherein 8-oxoG containing DNA duplex was either supplemented (+) or not (−) with pure EcoFpg. The bands

d S,

corresponding to product and substrate are indicated as P an

more, Fpg activity assays (Fig. 2C) in the cell-free extracts ofM. smegmatis (fpg-) harboring either the empty vector (lane 2)or the pTK-Mtufpg plasmid (lane 3) revealed that the presenceof pTK-Mtufpg plasmid restored the Fpg activity in M. smegma-tis (fpg-). Taken together, these observations showed that thelack of Fpg in M. smegmatis rendered it susceptible to oxidative

stress. More importantly, rescue of M. smegmatis (fpg-) growthin the presence of H2O2 (Fig. 2B) by an extrachromosal copy offpg, ensured that the susceptibility of the strain was primar-ily because of the deficiency of Fpg and not because of some

respectively.

indirect consequences of alterations in the fpg locus on thechromosome.

3.3. Analysis of mutation frequencies

The loss of a DNA repair function invariably results in an

increased mutator phenotype of an organism. Hence, to inves-tigate the consequence of the loss of Fpg on the mutatorphenotype of the M. smegmatis strain, we plated bacte-rial cultures on media containing rifampicin and scored for

d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785 1779

Fig. 2 – Analysis of growth. Cultures of M. smegmatis (WT)harboring pTKmx, M. smegmatis (fpg-) harboring eitherpTKmx or pTK-Mtufpg were inoculated with 1% of a 48 hculture derived from a single colony. The cultures weregrown at 37 ◦C in the absence of added H2O2 (panel A), orpresence of 3.0 mM H2O2 (panel B). Fpg activity assays(panel C) with cell-free extracts (20 �g) prepared from M.smegmatis (fpg-) harboring pTKmx (lane 2) or pTK-Mtufpg(lane 3) were carried out using duplex DNA containing8-oxoG:C base pair and analyzed on 15% PAGE containing8w

rlia

3d

Amr

Table 3 – Effect of deficiency of Fpg on mutationfrequencies in M. smegmatis

Strain aFrequency of RifR Fold increase

M. smegmatis 1.0 (0.94–1.2) × 10−7 1M. smegmatis (fpg-) 3.8 (3.3–4.4) × 10−7 3.8

a The numbers shown are median values from a sample size of 6.The numbers shown within parentheses are the lower and uppervalues calculated for 95% confidence limits. Mutation frequen-cies were calculated as the number of colonies on the rifampicinplate divided by total viable counts of the plated culture. Theactual mutation frequencies obtained in the 6 independent cul-

−7

T (or G to A) mutations was common to both the wild typeand the fpg-strains. However, the deficiency of fpg in M. smeg-matis resulted in an increase in A to G (or T to C) mutations

Table 4 – Spectrum of mutations in RRDR locus ofrifampicin resistant isolates of different strains of M.smegmatis mc2155 (Information on the positions ofchanges has been summarized in Fig. 3)

Strain Mutation Numbers

Untreated H2O2 treated

aM. smegmatis C → T/G → A 36% (12/33) 40% (18/45)C → G/G → C 21% (7/33) 24% (11/45)C → A/G → T 6% (2/33) Not detectedA → T/T → A Not detected 2% (1/45)A → G/T → C 24% (8/33) 29% (13/45)A → C/T → G 12% (4/33) 4% (2/45)cNomutations inRRDR

42/75 12/57

bM. smegmatis(fpg-)

C → T/G → A 26% (11/42) 41% (25/61)C → G/G → C 17% (7/42) 40% (24/61)C → A/G → T 2% (1/42) Not detectedT → A/A → T 2% (1/42) 3% (2/61)A → G/T → C 45% (19/42) 11% (7/61)A → C/T → G 7% (3/42) 5% (3/61)

cNomutations inRRDR

21/63 6/67

a A total of 75 and 57 samples were sequenced in the untreated andH2O2 treated experiments; of these 33 and 45 samples, respec-tively, showed mutations in RRDR.

M urea. Lane with (+) symbols indicate control reactionith pure EcoFpg.

ifampicin resistant (RifR) colonies. As shown in Table 3, theoss of Fpg activity, in this assay, resulted in a ∼3.8-foldncrease in mutation frequency over the wild type strain when48 h old culture was used.

.4. Characterization of rifampicin resistanceetermining region (RRDR) in rpoB gene in M. smegmatis

lthough resistance to rifampicin (RifR) can arise because ofutations elsewhere in the genome, a majority of mutations

esulting in RifR have been mapped to the rpoB gene in E.

tures were (0.98, 1.026, 1.11, 1.17, 1.02, 0.938) × 10 , and (3.29, 3.7,4.0, 4.45, 4.28, 3.45) × 10−7 for the wild-type and the fpg-strains,respectively.

coli and M. tuberculosis [35,36]. A comparison of M. tuberculo-sis RRDR sequence (27 amino acids, 81 nucleotides) showedthat this region is 100% conserved in M. smegmatis. Thus, toexamine the nature of mutations, we analyzed DNA sequenceof the 81-nucleotide stretch (RRDR) of rpoB encoding theseamino acids from 57 to 75 isolates of M. smegmatis strains.Results of these analyses are summarized in Table 4 and Fig. 3.As expected of a G + C rich genome, predominance of C to

b A total of 63 and 67 samples were sequenced in the untreated andH2O2 treated experiments; of these 42 and 61 samples, respec-tively, showed mutations in RRDR.

c Mutations elsewhere gave resistance to rifampicin.

1780 d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785

Fig. 3 – DNA (81 nucleotides) and the amino acid sequences corresponding to RRDR locus of M. smegmatis (426–452 aminoacids). Total number of mutations observed in the RRDR; nucleotide and amino acid changes and the frequencies of the

areation

changes in RRDR locus in untreated or H2O2 treated samplesin the wild type H2O2 treated sample indicates that the mut

[24% (wild type) to 45% (fpg-), untreated samples, Table 4] asopposed to the expected C to A (or G to T) mutations. To fur-ther our understanding of this phenomenon, we exposed M.smegmatis strains to sub-lethal level (2 mM) of hydrogen per-oxide prior to analyzing RRDR from RifR isolates (Table 4 andFig. 3). A predominant category of mutations still belongedto C to T (or G to A). Interestingly, now M. smegmatis (fpg-) showed an increase in C to G (or G to C) mutations [24%(wild type) to 40% (fpg-) H2O2 treated samples, Table 4]. TheC to A mutations accounted for 6 and 2% of all changes in theuntreated samples of wild type and fpg-strains, respectively.However, the C to A mutations remained undetectable in theH2O2 treated samples. These observations suggest that in M.smegmatis, Fpg plays a major role in down-regulating accumu-lation of A to G (or T to C) and C to G or (G to C) mutations in thegenome.

3.5. Determination of activity of E. coli and M.smegmatis cellular extracts for introducing dNMPsagainst 8-oxoG in DNA

To understand possible role of DNA polymerase(s) in the G, Cbiased spectrum of mutations in Fpg deficient M. smegmatis,we made use of a primer extension assay wherein oligomerictemplates containing either a G or 8-oxoG at position 15 (fromthe 3′ end) were annealed to a 14-mer 5′ 32P-labeled primerwhich paired one nucleotide short of the G or 8-oxoG in the

templates, and extended it with the DNA polymerase activityin the cell-free extracts prepared from E. coli and M. smegmatis(deficient in Fpg) in the presence of dNTPs (in singles or mix-tures). The cell-free extracts from E. coli (fpg-) extended the

indicated. An asterisk (*) at codon number 450 (Ser to Phe)occurred as consequence of TCG to TTC change.

primer by incorporating C against G in the template (Fig. 4A,lanes 5–7). Incorporation of other dNTPs was not detectable(lanes 2–4 and 8–13). However, in the template with 8-oxoG,incorporation of A (in addition to C) was clearly seen (Fig. 4B,lanes 2–4). This observation is as expected of dNTP bias of DNApolymerase(s) against 8-oxoG in E. coli, and which explains apredominant occurrence of C to A mutations in Fpg deficientE. coli. In similar experiments using extracts from M. smeg-matis (fpg-), we noted that while the primer was extendedwith C as +1 nucleotide in a template with G (Fig. 5A, lanes5–7), extension occurred with either C, A or G when the tem-plate possessed 8-oxoG (Fig. 5B, lanes 2–10). Incorporation ofT was not detectable (lanes 11–13). Notably, while the incor-poration of C or A was seen to increase with increasing timeof the reaction, incorporation of G was very rapid and reachedits maximal level even at the earliest time point, suggestinga preferential incorporation of G against 8-oxoG. Interest-ingly, a difference of mobility of the product of the +1 primerextension with either C, A or G (Fig. 5B), allowed us to inferpreferential incorporation of dNMPs against 8-oxoG from mix-tures of dNTPs. In the reaction with a dNTP mix lacking G,predominant incorporation was that of C (lanes 14–16). How-ever, when all 4 dNTPs were present (lanes 17–19 and lanes21–22), the +1 band at the earliest time point corresponded toG (compare lane 17 with lane 8). Notably, with increase in thereaction time, this G was replaced predominantly by C (com-pare lanes 18 and 5), which is better seen in a longer run of

the reaction products (lanes 20–22). These results suggest thatwhile mycobacterial DNA polymerase(s) are capable of insert-ing A against 8-oxoG, this activity is severely curtailed to favorG and C when all dNTPs are present in the reaction.

d n a r e p a i r 6 ( 2 0 0 7

Fig. 4 – DNA polymerase assays in cell-free extracts from E.coli (fpg-). The duplex DNA was prepared from syntheticoligomers as described in Section 2, where either G (panelA) or 8-oxoG (panel B) containing oligomers were annealedwith 5′ 32P end labeled 14-mer primer. The assays werecarried out in the presence of 100 �M dATP (lanes 2–4),dCTP (lanes 5–7), dGTP (lanes 8–10), dTTP (lanes 11–13),dNTP mix without dGTP (lanes 14–16) or all four dNTPs mix(lanes 17–19). Cell-free extract (0.5 �g) was added to eachreaction, aliquots were drawn at ∼1/2, 2 or 5 min andanalyzed on 15% PAGE containing 8 M urea. Lane 1 iscontrol (C) from the reaction mixture taken prior to addingcell extract.

) 1774–1785 1781

3.6. Substrate specificity of mycobacterial MutY

In E. coli, when A is incorporated against 8-oxoG, MutY isknown to remove this A for the polymerase(s) to make addi-tional attempts to insert C against the 8-oxoG [14,15]. Thisproperty of MutY provides yet another opportunity to Fpgto excise the 8-oxoG from DNA. Analysis of primer exten-sion with mycobacterial extracts showed (Fig. 5B) that themycobacterial polymerase(s) could insert G against 8-oxoG. Tounderstand role of mycobacterial MutY in removing mispairednucleotides against 8-oxoG, we analyzed substrate specificityof MutY in M. smegmatis. As is known for MutY from otherorganisms, M. smegmatis extracts possessed a predominantactivity of processing A against 8-oxoG (Fig. 6, lane 2). Exci-sion of C or T from pairs with 8-oxoG was not detectable (lanes4 and 8). However, a minor activity of processing G against8-oxoG was present in the extract (lane 6). Identical resultswere obtained in assays with purified recombinant MutY fromM. tuberculosis (which shows 70% identity, and 78% similarityto M. smegmatis MutY) suggesting that the activities of exci-sion of A and G present against 8-oxoG in DNA are intrinsic tomycobacterial MutY (Srinath, T. and Varshney, U., unpublished).

3.7. Removal of 8-oxoG from various base mispairs bycellular extracts of M. smegmatis and M. smegmatis (fpg-)

To further our understanding of the G, C bias of mutations inRRDR in Fpg deficient mycobacteria, we assayed for processingof 8-oxoG from its pairs with A, C, G or T (Fig. 7). These assaysshowed that 8-oxoG was processed when present against C,G or T (lanes 6, 9 and 12). A control reaction with purified Fpgfrom E. coli (lane 13) shows that the product bands seen in lanes6, 9 and 12 correspond to processing of 8-oxoG. As the 8-oxoGprocessing activity was largely absent in the extracts from theFpg deficient M. smegmatis (lanes 5, 8 and 11) it can be inferredthat the 8-oxoG processing activity in the extracts was due toFpg. Further, like the reported property of E. coli Fpg, mycobac-terial Fpg also does not show any detectable processing of8-oxoG when paired with A (lane 3).

4. Discussion

Mycobacteria possess G + C rich genomes, and are, therefore,at high risk of accumulating 8-oxoG in DNA. Fpg (MutM) isthe chief enzyme that initiates repair of this damaged base inDNA. However, there have been no earlier reports on Fpg fromthis important group of organisms. The present study wasdirected towards understanding the role of Fpg in mycobac-teria. As assessed by an increase in the appearance of RifR

colonies, a deficiency of Fpg in M. smegmatis resulted in a mod-erate increase (∼3.8-fold) in the mutator phenotype of thebacterium. Another expected phenotypic change associatedwith the Fpg deficiency was an increase in sensitivity of themutant to H2O2. However, we observed that when the test cul-tures were started with inoculum from the primary cultures

grown to stationary phase (55–70 h) as opposed to the log orlate log phase (48 h, used in Fig. 2), the susceptibility of themutant to H2O2 was significantly decreased (data not shown).Although the reasons for this differential susceptibility are

1782 d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785

Fig. 5 – DNA polymerase assays in cell-free extracts from M. smegmatis (fpg-). The duplex DNA was prepared from syntheticoligomers as described in Section 2, where either G (panel A) or 8-oxoG (panel B) containing oligomers were annealed with32P 5′ end labeled 14-mer primer. The assays were carried out either in the presence of 100 �M dATP (lanes 2–4), dCTP (lanes5–7), dGTP (lanes 8–10 and 20), dTTP (lanes 11–13), dNTP mix without dGTP (lanes 14–16, 23 and 24) or all four dNTPs mix(lanes 17–19, 21 and 22). Cell-free extract (5 �g) was added to each reaction and aliquots were drawn at ∼1/2, 2 or 5 min andanalyzed on 15% PAGE containing 8 M urea. Lane 1 is control (C) from the reaction mixture taken prior to adding cell extract.

d n a r e p a i r 6 ( 2 0 0 7 ) 1774–1785 1783

Fig. 6 – Assays for MutY on DNA duplexes containing 8-oxoG. The duplexes were prepared from DNA oligomers asdescribed in Section 2 wherein 8-oxoG containing oligomer was annealed with 5′ 32P end labeled complementary strandcontaining A (lanes 1 and 2), C (lanes 3 and 4), G (lanes 5 and 6) or T (lanes 7 and 8) against 8-oxoG. Lanes with (+) or (−)symbols indicate addition or not, respectively, of 25 �g cell-free extract of M. smegmatis mc2. MtuMutY (200 ng) was used asp PAG(

nrcotstt

tmtFGt

fcagm8

Fi(ecr

ositive control (lane 9). The reactions were analyzed on 15%P) and substrate (S) are indicated.

ot understood, the observation is consistent with the earliereports that growth dependent changes in the composition ofell wall, the very first line of defence against any insult to therganism, alter its potential in detoxifying a variety of oxida-ive stress [37,38]. The products of several other genes such asodA, sodC, katG, aphC, oxyR, sigJ, hsp16.3 etc. are also knowno make crucial contributions to the capacity of mycobactriao tolerate oxidative stress [39–45].

Fpg deficient E. coli accumulates C to A mutations [16]. Onhe contrary, while analysis of mutations in RRDR in M. smeg-atis (wild type for Fpg) showed diverse mutations, two of

hese (A to G and C to G) showed a remarkable association withpg deficiency (Fig. 3 and Table 4). Interestingly, while the C tomutations predominated in the cultures treated with H2O2,

he A to G mutations predominated in the untreated cultures.The primer extension experiments carried out with cell-

ree extracts of M. smegmatis (fpg-) showed that either A, C or Gould be incorporated against 8-oxoG (Fig. 5B). However, when

ll four dNTPs were present, the initial incorporation of G wasradually replaced, predominantly with C pointing to involve-ent of different polymerases in inserting nucleotides against

-oxoG. Based on these assays, as well as earlier reports from

ig. 7 – Assays for Fpg activity on duplex DNA. The duplexes wen Section 2 wherein 5′ 32P end labeled 8-oxoG oligomer was pailanes 1–3), C (lanes 4–6), G (lanes 7–9) or T (lanes 10–12) againstxtracts from M. smegmatis (lane 3, 6, 9 and 12) and M. smegmationtaining 8 M urea. Lanes with (−) or (+) symbols indicate contrespectively, with pure EcoFpg. The bands corresponding to prod

E containing 8 M urea. The bands corresponding to product

other organisms [15,46–48], we believe that use of differentpolymerases under normal and oxidative stress conditionscould be primarily responsible for this phenomenon.

While the precise reasons for the change in the mutationpattern seen between untreated and H2O2 treated M. smeg-matis (fpg-) remain largely unclear at present, the biochemicalassays (Fig. 5B) provide some insight into our understanding ofthe accumulation of specific mutations in M. smegmatis (fpg-).Fpg being the major protein involved in excision of 8-oxoG(Fig. 1E), any occurrence of 8-oxoG in DNA in Fpg deficientstrains will persist. And, a bias of mycobacterial DNA poly-merases to insert G (in place of C) against 8-oxoG would resultin occurrence of 8-oxoG:G mispair. As the mycobacterial MutYdoes not efficiently excise the mispaired G (Fig. 6), Fpg initiatedexcision repair of 8-oxoG from this mispair (Fig. 7) would resultin fixation of G:C to C:G mutation (subsequent to filling in ofthe gap left behind by Fpg, by DNA polymerase). The G:C to C:Gmutations may also arise solely as a consequence of DNA repli-

cations prior to the MutY or Fpg mediated repair. On the otherhand, A to G (or T to C) mutation could arise if 8-oxo-dGTPpresent in the dNTP pool was utilized by DNA polymerase inplace of dTTP resulting in a 8-oxoG:A mispair, which is a sub-

re prepared by annealing synthetic oligomers as describedred with different complementary strands containing A8-oxoG. The assay was carried out with 20 �g of cell-frees (fpg-) (lane 2, 5, 8 and 11) and analyzed on 15% PAGEols wherein the DNA was either supplemented or not,uct (P) and substrate (S) are indicated.

( 2 0

r

1784 d n a r e p a i r 6

strate of MutY [7,21,49]. Excision of A residue by MutY followedby incorporation of G by DNA polymerase(s) against 8-oxoGwould result in A to G mutations. Alternatively, but not mutu-ally exclusive, incorporation of 8-oxo-dGMP against T wouldgive rise to 8-oxoG:T pair. As the 8-oxoG:T pair is not a sub-strate for MutY (Fig. 6), deficiency of Fpg in the cell would allow8-oxoG in the DNA to be used as template by the DNA poly-merases, allowing for incorporation of C against it, resultingin T to C (or A to G) mutations.

C to A mutations occurred at a very low frequency (twoin the wild type and one in fpg-strains, untreated samples,Table 4). Similarly, earlier analyses in mycobacteria have eitherfailed to observe accumulation of C to A mutations [50] orobserved it at very low frequency [51]. It is, therefore, reason-able to hypothesize that either the occurrence of A against8-oxoG is efficiently eliminated by MutY (Fig. 6) and/or eventhough incorporation of A against 8-oxoG is seen when noother nucleotides are present in the reaction; this is severelycurtailed in the presence of other dNTPs (Fig. 5B). However,given that understanding of the DNA repair mechanisms inmycobacteria is in its early stages, alternate mechanisms suchas the presence of backup activities [9,52] that remove 8-oxoGand, thus, prevent C to A mutations can not be ruled out.

Finally, the results presented here clearly show that Fpgis an important enzyme in mycobacteria in preventing occur-rence of A to G (or T to C) and, G to C (or C to G) mutations. Thispattern of mutations in Fpg deficient mycobacteria is funda-mentally different from the one known for Fpg deficient E. coli.However, considering that, to a first approximation, the sub-strate specificities of mycobacterial Fpg and MutY are similar(Figs. 1E, 6 and 7), we believe that the distinct pattern of muta-tions, in Fpg deficient M. smegmatis is more a consequenceof the interface of the mycobacterial DNA polymerase(s) withthe 8-oxoG repair machinery. M. tuberculosis and M. smegmatisare known to possess two functional DNA polymerases, DnaE1and DnaE2, the major replicating and repair polymerases,respectively [51]. It would be interesting to study the inter-play between these enzymes and the 8-oxoG repair pathway.Also, from the predominant occurrence of A to G mutationsin the Fpg deficient M. smegmatis (Table 4, untreated), whichwould occur as a consequence of direct incorporation of 8-oxoG from the dNTP pool, it is evident that the levels of thelatter could be quite high in the intracellular pool of dNTPs. Infact, the bioinformatics analyses have revealed the presenceof at least three homologs of mutT (product of which preventsaccumulation of 8-oxo-dGTP in the cell) in the genomes M.tuberculosis and M. smegmatis, as opposed to a single gene in E.coli. Thus, the distinctive features of the 8-oxoG repair pathwayin mycobacteria make it an exciting area for further researchto understand the DNA repair mechanisms in this importantclass of bacteria; to further our understanding of developmentof multiple drug resistance and to design novel strategies tocontrol the growth of the pathogenic mycobacteria.

Acknowledgements

We thank our laboratory colleagues for their suggestions onthe manuscript. This work was supported by grants fromDepartment of Biotechnology, Council of Scientific and Indus-

0 7 ) 1774–1785

trial Research, and Indian Council of Medical Research, NewDelhi.

e f e r e n c e s

[1] A.O. Fels, Z.A. Cohn, The alveolar macrophage, J. Appl.Physiol. 60 (1986) 353–369.

[2] J.R. Lancaster Jr., Diffusion of free nitric oxide, MethodsEnzymol. 268 (1996) 31–50.

[3] J.A. Imlay, Pathways of oxidative damage, Annu. Rev.Microbiol. 57 (2003) 395–418.

[4] B.K. Duncan, DNA Glycosylases, in: P. Boyer (Ed.), TheEnzymes, Academic Press, Orlando, 1981, pp. 565–586.

[5] T. Lindahl, DNA repair enzymes, Annu. Rev. Biochem. 51(1982) 61–87.

[6] D.A. Wink, K.S. Kasprzak, C.M. Maragos, R.K. Elespuru, M.Misra, T.M. Dunams, T.A. Cebula, W.H. Koch, A.W. Andrews,J.S. Allen, et al., DNA deaminating ability and genotoxicity ofnitric oxide and its progenitors, Science 254 (1991)1001–1003.

[7] H. Maki, M. Sekiguchi, Mut T protein specifically hydrolysesmutagenic substrates for DNA synthesis, Nature 355 (1992)273–275.

[8] D.W. Mosbaugh, S.E. Bennett, Uracil-excision DNA repair,Prog. Nucleic Acid Res. Mol. Biol. 48 (1994) 315–370.

[9] V. Mizrahi, S.J. Andersen, DNA repair in Mycobacteriumtuberculosis. What have we learnt from the genomesequence? Mol. Microbiol. 29 (1998) 1331–1339.

[10] R. Manganelli, E. Dubnau, S. Tyagi, F.R. Kramer, I. Smith,Differential expression of 10 sigma factor genes inMycobacterium tuberculosis, Mol. Microbiol. 31 (1999)715–724.

[11] H.E. Krokan, R. Standal, G. Slupphaug, DNA glycosylases inthe base excision repair of DNA, Biochem. J. 325 (1997) 1–16.

[12] L.H. Pearl, Structure and function in the uracil-DNAglycosylase superfamily, Mutat. Res. 460 (2000) 165–181.

[13] T. Lindahl, DNA glycosylases, endonucleases forapurinic/apyrimidinic sites, and base excision-repair, Prog.Nucleic Acid Res. Mol. Biol. 22 (1979) 135–192.

[14] M.L. Michaels, J.H. Miller, The GO system protects organismsfrom the mutagenic effect of the Spontaneous lesion8-Hydroxyguanine (7,8-dihydro-8-oxoguanine), J. Bacteriol.174 (1992) 6321–6325.

[15] A.P. Grollman, M. Moriya, Mutagenesis by 8-oxoguanine: anenemy within, Trends Genet. 9 (1993) 246–249.

[16] M.L. Michaels, J. Tchou, A.P. Grollman, J.H. Miller, A repairsystem for 8-oxo-7,8-dihydrodeoxyguanine, Biochemistry 31(1992) 10964–10968.

[17] A. Sakai, M. Nakanishi, K. Yoshiyama, H. Maki, Impact ofreactive oxygen species on spontaneous mutagenesis inEscherichia coli, Genes Cells 11 (2006) 767–778.

[18] M.D. Leipold, J.G. Muller, C.J. Burrows, S.S. David, Removal ofhydantoin products of 8-oxoguanine oxidation by theEscherichia coli DNA repair enzyme, FPG, Biochemistry 39(2000) 14984–14992.

[19] J.C. Fromme, G.L. Verdine, Structural insights into lesionrecognition and repair by the bacterial 8-oxoguanine DNAglycosylase MutM, Nat. Struct. Biol. 9 (2002) 544–552.

[20] J. Tchou, H. Kasai, S. Shibutani, M.H. Chung, J. Laval, A.P.Grollman, S. Nishimura, 8-oxoguanine (8-hydroxyguanine)DNA glycosylase and its substrate specificity, Proc. Natl.

Acad. Sci. USA 88 (1991) 4690–4694.

[21] T. Tajiri, H. Maki, M. Sekiguchi, Functional cooperation ofMutT, MutM and MutY proteins in preventing mutationscaused by spontaneous oxidation of guanine nucleotide inEscherichia coli, Mutat. Res. 336 (1995) 257–267.

0 0 7

d n a r e p a i r 6 ( 2

[22] J. Venkatesh, P. Kumar, P.S. Krishna, R. Manjunath, U.Varshney, Importance of uracil DNA glycosylase inPseudomonas aeruginosa and Mycobacterium smegmatis,G + C-rich bacteria, in mutation prevention, tolerance toacidified nitrite, and endurance in mouse macrophages, J.Biol. Chem. 278 (2003) 24350–243508.

[23] C.M. Sassetti, D.H. Boyd, E.J. Rubin, Genes required formycobacterial growth defined by high density mutagenesis,Mol. Microbiol. 48 (2003) 77–84.

[24] S.B. Snapper, R.E. Melton, S. Mustafa, T. Kieser, W.R. JacobsJr., Isolation and characterization of efficient plasmidtransformation mutants of Mycobacterium smegmatis, Mol.Microbiol. 4 (1990) 1911–1919.

[25] S.T. Cole, R. Brosch, J. Parkhill, T. Garnier, C. Churcher, D.Harris, S.V. Gordon, K. Eiglmeier, S. Gas, C.E. Barry III, F.Tekaia, K. Badcock, D. Basham, D. Brown, T. Chillingworth, R.Connor, R. Davies, K. Devlin, T. Feltwell, S. Gentles, N.Hamlin, S. Holroyd, T. Hornsby, K. Jagels, A. Krogh, J. McLean,S. Moule, L. Murphy, K. Oliver, J. Osborne, M.A. Quail, M.A.Rajandream, J. Rogers, S. Rutter, K. Seeger, J. Skelton, R.Squares, S. Squares, J.E. Sulston, K. Taylor, S. Whitehead, B.G.Barrell, Deciphering the biology of Mycobacterium tuberculosisfrom the complete genome sequence, Nature 393 (1998)537–544.

[26] S. Bardarov, S. Bardarov Jr., M.S. Pavelka Jr., V.Sambandamurthy, M. Larsen, J. Tufariello, J. Chan, G. Hatfull,W.R. Jacobs Jr., Specialized transduction: an efficient methodfor generating marked and unmarked targeted genedisruptions in Mycobacterium tuberculosis, M. bovis BCG andM. smegmatis, Microbiology 148 (2002) 3007–3017.

[27] V. Pelicic, M. Jackson, J.M. Reyrat, W.R. Jacobs Jr., B. Gicquel,C. Guilhot, Efficient allelic exchange and transposonmutagenesis in Mycobacterium tuberculosis, Proc. Natl. Acad.Sci. USA 94 (1997) 10955–10960.

[28] K.C. Reed, D.A. Mann, Rapid transfer of DNA from agarosegels to nylon membranes, Nucleic Acids Res. 13 (1985)7207–7221.

[29] M. Vasanthakrishna, N.V. Kumar, U. Varshney,Characterization of the initiator tRNA gene locus andidentification of a strong promoter from Mycobacteriumtuberculosis, Microbiology 143 (1997) 3591–3598.

[30] J. Sambrook, E.F. Fritsch, T. Maniats, Molecular Cloning: ALaboratory Manual, Cold Spring Harbor Laboratory, ColdSpring Harbor, NY, 1989.

[31] G. Dianov, A. Price, T. Lindahl, Generation ofsingle-nucleotide repair patches following excision of uracilresidues from DNA, Mol. Cell Biol. 12 (1992) 1605–1612.

[32] T.J. Kenney, G. Churchward, Genetic analysis of theMycobacterium smegmatis rpsL promoter, J. Bacteriol. 178(1996) 3564–3571.

[33] N. Acharya, P. Kumar, U. Varshney, Complexes of theuracil-DNA glycosylase inhibitor protein, Ugi, withMycobacterium smegmatis and Mycobacterium tuberculosisuracil-DNA glycosylases, Microbiology 149 (2003) 1647–1658.

[34] P. Kumar, DNA repair in mycobacteria, Ph.D. Thesis, IndianInstitute of Science, Bangalore, India (2005).

[35] J.M. Musser, Antimicrobial agent resistance in mycobacteria:molecular genetic insights, Clin. Microbiol. Rev. 8 (1995)496–514.

[36] M. Heep, B. Brandstatter, U. Rieger, N. Lehn, E. Richter, S.Rusch-Gerdes, S. Niemann, Frequency of rpoB mutations

inside and outside the cluster I region in rifampin-resistantclinical Mycobacterium tuberculosis isolates, J. Clin. Microbiol.39 (2001) 107–110.

[37] M. Rawat, G.L. Newton, M. Ko, G.J. Martinez, R.C. Fahey, Y.Av-Gay, Mycothiol-deficient Mycobacterium smegmatis

) 1774–1785 1785

mutants are hypersensitive to alkylating agents, freeradicals, and antibiotics, Antimicrob, Agents Chemother. 46(2002) 3348–3355.

[38] N.A. Buchmeier, G.L. Newton, T. Koledin, R.C. Fahey,Association of mycothiol with protection of Mycobacteriumtuberculosis from toxic oxidants and antibiotics, Mol.Microbiol. 47 (2003) 1723–1732.

[39] Y. Zhang, R. Lathigra, T. Garbe, D. Catty, D. Young, Geneticanalysis of superoxide dismutase, the 23 kilodalton antigenof Mycobacterium tuberculosis, Mol. Microbiol. 5 (1991)381–391.

[40] D.R. Sherman, P.J. Sabo, M.J. Hickey, T.M. Arain, G.G.Mahairas, Y. Yuan, C.E. Barry III, C.K. Stover, Disparateresponses to oxidative stress in saprophytic and pathogenicmycobacteria, Proc. Natl. Acad. Sci. USA 92 (1995) 6625–6629.

[41] S. Dhandayuthapani, Y. Zhang, M.H. Mudd, V. Deretic,Oxidative stress response and its role in sensitivity toisoniazid in mycobacteria: characterization and inducibilityof ahpC by peroxides in Mycobacterium smegmatis and lack ofexpression in M. aurum and M. tuberculosis, J. Bacteriol. 178(1996) 3641–3649.

[42] T.C. Zahrt, J. Song, J. Siple, V. Deretic, Mycobacterial FurA is anegative regulator of catalase-peroxidase gene katG, Mol.Microbiol. 39 (2001) 1174–1185.

[43] A. Abulimiti, X. Qiu, J. Chen, Y. Liu, Z. Chang, Reversiblemethionine sulfoxidation of Mycobacterium tuberculosis smallheat shock protein Hsp16.3 and its possible role inscavenging oxidants, Biochem. Biophys. Res. Commun. 305(2003) 87–93.

[44] Y. Hu, S. Kendall, N.G. Stoker, A.R. Coates, The Mycobacteriumtuberculosis sigJ gene controls sensitivity of the bacterium tohydrogen peroxide, FEMS Microbiol. Lett. 237 (2004) 415–423.

[45] A. Koshkin, X.T. Zhou, C.N. Kraus, J.M. Brenner, P.Bandyopadhyay, I.D. Kuntz, C.E. Barry III, P.R. Ortiz deMontellano, Inhibition of Mycobacterium tuberculosis AhpD,an element of the peroxiredoxin defense against oxidativestress, Antimicrob. Agents Chemother. 48 (2004) 2424–2430.

[46] S. Shibutani, M. Takeshita, A.P. Grollman, Insertion ofspecific bases during DNA synthesis past theoxidation-damaged base 8-oxodG, Nature 349 (1991) 431–434.

[47] E. Efrati, G. Tocco, R. Eritja, S.H. Wilson, M.F. Goodman,“Action-at-a-distance” mutagenesis.8-oxo-7,8-dihydro-2′-deoxyguanosine causes basesubstitution errors at neighboring template sites whencopied by DNA polymerase beta, J. Biol. Chem. 274 (1999)15920–15926.

[48] L. Haracska, S.L. Yu, R.E. Johnson, L. Prakash, S. Prakash,Efficient and accurate replication in the presence of7,8-dihydro-8-oxoguanine by DNA polymerase eta, Nat.Genet. 25 (2000) 458–461.

[49] R.G. Fowler, S.J. White, C. Koyama, S.C. Moore, R.L. Dunn,R.M. Schaaper, Interactions amoug the Escherichia coli mutT,mutM and mutY damage prevention pathways, DNA Repair 2(2003) 159–173.

[50] G.P. Morlock, B.B. Plikaytis, J.T. Crawford, Characterization ofspontaneous, In vitro-selected, rifampin-resistant mutantsof Mycobacterium tuberculosis strain H37Rv, Antimicrob.Agents Chemother. 44 (2000) 3298–3301.

[51] H.I. Boshoff, M.B. Reed, C.E. Barry III, V. Mizrahi, DnaE2polymerase contributes to in vivo survival and theemergence of drug resistance in Mycobacterium tuberculosis,

Cell 113 (2003) 183–193.

[52] J.O. Blaisdell, Z. Hatahet, S.S. Wallace, A novel role forEscherichia coli endonuclease VIII in prevention ofspontaneous G → T transversions, J. Bacteriol. 181 (1999)6396–6402.