Portal de Periódicos da CAPES

The Role of the Escherichia coli Mug Protein in the Removal of Uracil and 3,N 4-Ethenocytosine from DNA

1999; Elsevier BV; Volume: 274; Issue: 43 Linguagem: Inglês

10.1074/jbc.274.43.31034

1083-351X

Eugene A. Lutsenko, Ashok S. Bhagwat,

DNA and Nucleic Acid Chemistry

The human thymine-DNA glycosylase has a sequence homolog in Escherichia coli that is described to excise uracils from U·G mismatches (Gallinari, P., and Jiricny, J. (1996)Nature 383, 735–738) and is named mismatched uracil glycosylase (Mug). It has also been described to remove 3,N 4-ethenocytosine (εC) from εC·G mismatches (Saparbaev, M., and Laval, J. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 8508–8513). We used a mug mutant to clarify the role of this protein in DNA repair and mutation avoidance. We find that inactivation of mug has no effect on C to T or 5-methylcytosine to T mutations in E. coli and that this contrasts with the effect of ung defect on C to T mutations and of vsr defect on 5-methylcytosine to T mutations. Even under conditions where it is overproduced in cells, Mug has little effect on the frequency of C to T mutations. Because uracil-DNA glycosylase (Ung) and Vsr are known to repair U·G and T·G mismatches, respectively, we conclude that Mug does not repair U·G or T·G mismatches in vivo. A defect inmug also has little effect on forward mutations, suggesting that Mug does not play a role in avoiding mutations due to endogenous damage to DNA in growing E. coli. Cell-free extracts frommug + ung cells show very little ability to remove uracil from DNA, but can excise εC. The latter activity is missing in extracts from mug cells, suggesting that Mug may be the only enzyme in E. coli that can remove this mutagenic adduct. Thus, the principal role of Mug in E. coli may be to help repair damage to DNA caused by exogenous chemical agents such as chloroacetaldehyde. The human thymine-DNA glycosylase has a sequence homolog in Escherichia coli that is described to excise uracils from U·G mismatches (Gallinari, P., and Jiricny, J. (1996)Nature 383, 735–738) and is named mismatched uracil glycosylase (Mug). It has also been described to remove 3,N 4-ethenocytosine (εC) from εC·G mismatches (Saparbaev, M., and Laval, J. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 8508–8513). We used a mug mutant to clarify the role of this protein in DNA repair and mutation avoidance. We find that inactivation of mug has no effect on C to T or 5-methylcytosine to T mutations in E. coli and that this contrasts with the effect of ung defect on C to T mutations and of vsr defect on 5-methylcytosine to T mutations. Even under conditions where it is overproduced in cells, Mug has little effect on the frequency of C to T mutations. Because uracil-DNA glycosylase (Ung) and Vsr are known to repair U·G and T·G mismatches, respectively, we conclude that Mug does not repair U·G or T·G mismatches in vivo. A defect inmug also has little effect on forward mutations, suggesting that Mug does not play a role in avoiding mutations due to endogenous damage to DNA in growing E. coli. Cell-free extracts frommug + ung cells show very little ability to remove uracil from DNA, but can excise εC. The latter activity is missing in extracts from mug cells, suggesting that Mug may be the only enzyme in E. coli that can remove this mutagenic adduct. Thus, the principal role of Mug in E. coli may be to help repair damage to DNA caused by exogenous chemical agents such as chloroacetaldehyde. thymine-DNA glycosylase 3,N 4-ethenocytosine uracil-DNA glycosylase mismatched uracil glycosylase isopropyl-β-d-thiogalactoside polymerase chain reaction Cytosine is the most unstable of the four bases in DNA and deaminates hydrolytically to create U·G mismatches. If unrepaired, uracil can pair with an adenine during replication causing a C to T mutation. For this reason, cells contain uracil-DNA glycosylase (Ung), an enzyme that removes the uracil and initiates its replacement with cytosine. The importance of Ung in mutation avoidance is evidenced by the observation that ung strains of Escherichia coli (1Duncan B.K. Weiss B. J. Bacteriol. 1982; 151: 750-755Crossref PubMed Google Scholar) and yeast (2Burgers P.M. Klein M.B. J. Bacteriol. 1986; 166: 905-913Crossref PubMed Google Scholar) accumulate C to T mutations. Cytosines methylated at position 5 similarly deaminate to create T·G mismatches, which are not subject to repair by Ung. In E. coli, a specialized mismatch correction process called very short patch repair corrects these mispairs to C·G (3Lieb M. Bhagwat A.S. Mol. Microbiol. 1996; 20: 467-473Crossref PubMed Scopus (83) Google Scholar). The key enzyme in this repair pathway is a sequence-specific, mismatch-specific endonuclease, Vsr, which hydrolyzes the phosphodiester linkage preceding the mismatched T (4Hennecke F. Kolmar H. Bründl K. Fritz H.-J. Nature. 1991; 353: 776-778Crossref PubMed Scopus (113) Google Scholar). No eukaryotic sequence homologs of this enzyme have been reported; instead a DNA glycosylase is thought to serve the same function (5Wiebauer K. Neddermann P. Hughes M. Jiricny J. Exs (Exper. Suppl.). 1993; 64: 510-522PubMed Google Scholar). This enzyme excises thymines from T·G mismatches (6Wiebauer K. Jiricny J. Nature. 1989; 339: 234-236Crossref PubMed Scopus (152) Google Scholar) and prefers mismatches that are followed by a G·C pairs (7Sibghat U. Gallinari P. Xu Y.Z. Goodman M.F. Bloom L.B. Jiricny J. Day 3rd., R.S. Biochemistry. 1996; 35: 12926-12932Crossref PubMed Scopus (82) Google Scholar, 8Waters T.R. Swann P.F. J. Biol. Chem. 1998; 273: 20007-20014Abstract Full Text Full Text PDF PubMed Scopus (137) Google Scholar). This enzyme, thymine-DNA glycosylase (TDG),1 could prevent mutations when 5-methylcytosines within CG dinucleotides deaminate to thymine. The cDNA for TDG was cloned and its sequence was determined (9Neddermann P. Gallinari P. Lettieri T. Schmid D. Truong O. Hsuan J.J. Wiebauer K. Jiricny J. J. Biol. Chem. 1996; 271: 12767-12774Abstract Full Text Full Text PDF PubMed Scopus (235) Google Scholar). Remarkably, a sequence homolog of this protein was found in E. coli and Serratia marcescence (10Gallinari P. Jiricny J. Nature. 1996; 383: 735-738Crossref PubMed Scopus (184) Google Scholar). The investigators who made this observation suggested that the bacterial homolog was a uracil-DNA glycosylase specific for U·G mismatches and named it mismatch-specific uracil-DNA glycosylase (Mug). Their conclusions were based on properties of truncated forms of TDG, and biochemical assays done using E. coli cell-free extracts. They further suggested that it may act as a backup enzyme for Ung and may be important in avoiding mutations during stationary phase of cell growth (10Gallinari P. Jiricny J. Nature. 1996; 383: 735-738Crossref PubMed Scopus (184) Google Scholar). Because there was no evidence in previous work by any research group that such a backup enzyme existed, we tested this possibility. For this purpose, we studied the repair of U·G and T·G mismatches inmug + and mug strains. Our results clearly indicate that Mug plays no role in the repair of U·G or T·G mismatches and may repair 3,N 4-ethenocytosine·G mismatches as suggested recently by Sapabaev and Laval (11Saparbaev M. Laval J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8508-8513Crossref PubMed Scopus (164) Google Scholar). JM253.140 (F- araD139 Δ(argF-lac)U169 rpsL150 relA1 deoC1 rbsR22 flhD5301 fruA25 mug::mini-Tn10) was kindly provided by J. Reiss (Princeton University). GM31 (dcm-6 thr-1 hisG4 leuB6 rpsL ara-14 supE44 lacY1 tonA31 tsx-78 galK2 galE2 xyl-5 thi-1 mtl-1) and RP4182 {(flaD-flaP)DE4trp gal rpsL} are from our collection. BH156 is GM31 withung-1 tyrA::Tn10 and was obtained from M. Lieb (University of Southern California School of Medicine, Los Angeles, CA). BH161 (BH156 with λ DE3 lysogen) was constructed by A. Beletskii (Wayne State University) using a kit from Novagen (Madison, WI). Construction of BH157 and BH158 is described below. P1vir phage is from our collection. Plasmid pET11-d was purchased from Stratagene (La Jolla, CA). BH157 and BH158 were constructed by the P1 transduction ofmug::mini-Tn10 from JM253.140 into GM31 and BH156, respectively. LB medium supplemented with 0.2% glucose and 5 mm CaCl2 was inoculated with an overnight culture of JMR253.140 and was shaken vigorously at 37 °C for 30 min. P1vir phage was added to the culture at a multiplicity of infection of 0.1, and the infected cells were further incubated 37 °C for 3 h with continued shaking. The cells were lysed with chloroform, the culture was centrifuged to clear the cell debris, and the supernatant was removed. The phage in the supernatant was titered and was used to infect GM31 or BH156. For the infection, 10 ml of overnight cultures were centrifuged, and the cell pellets were resuspended in 1 ml of a buffer containing 5 mm CaCl2 and 10 mm MgSO4. One hundred μl of the suspensions were infected with 10, 50, or 100 μl of the P1 phage, and the cultures were incubated at 37 °C for 30 min without shaking. One hundred μl of 1 m sodium citrate and 1 ml of LB were added to each tube, and the cells were incubated for 1 h at 37 °C with shaking. Following concentration by centrifugation, the cells were spread on LB plates containing 12 μg/ml tetracycline. The plates were incubated overnight at 37 °C, and three colonies from each transduction were studied further. The DNA from these colonies was amplified by PCR using the following primers: primer 1, 5′-GATCACCTATCTGCTGGAACAGTACGATCGTG-3′; and primer 2, 5′-CTGTATGTCTGCGATGAATCCGGAATG-3′. The colonies that gave rise to a larger PCR product than mug + cells were chosen for further analysis. The successful transfer ofmug::mini-Tn10 allele was confirmed by Southern hybridization. The region flanking the mug gene in BH156 chromosomal DNA was amplified by PCR using the following primers 1 and 2 mentioned above. The 2.6-kilobase pair PCR product was excised from a low melting point agarose gel and labeled with nonradioactive digoxigenin using the digoxigenin High Prime DNA labeling and detection starter kit II (Roche Molecular Biochemicals). Chromosomal DNAs from BH156, BH157, and BH158 were digested with PvuII, and the fragments were separated in 0.7% agarose gel. The DNA was blotted onto a nylon membrane and hybridized with probe labeled with digoxigenin. After treatment with alkaline phosphotase-conjugated antibodies raised against digoxigenin, the membrane was incubated with solution containing the chemiluminescence substrate for the enzyme CSPD and exposed to x-ray film for 1 h. The autoradiograph showed that themug + strain BH156 contains a 3.0-kilobase pairPvuII fragment containing mug, whereas the corresponding fragment in BH157 and BH158 is 6 kilobase pairs. This confirms the disruption of mug in BH157 and BH158. The open reading frame of mug + gene was amplified from chromosomal DNA of RP4182 using the following primers: primer 3, 5′-CCCGCTCTATCGCGGATCAGGCGCGCA-3′; and primer 4, 5′-CCCCCCCATGGTTGAGGATATTTTGGCTCCAGGG-3′. The amplification was done with the Pfu DNA polymerase, and the ∼500-base pair PCR product was isolated from a low melting agarose gel. The DNA was digested withNcoI and MboI (compatible with BamHI), and ligated to pET11-d expression vector digested with NcoI and BamHI. Two clones with the expected inserts were picked for further analysis. The plasmids were transformed into BL21(DE3), and one transformant from each plasmid was grown in LB medium containing 1 mm IPTG. The cells were broken by sonication, and the cell debris was removed by centrifugation. The supernatant was subjected to SDS-polyacrylamide gel electrophoresis and stained with Coomassie Brilliant Blue. Cell-free extract from one of the two clones showed the presence of a new protein of the expected size. This plasmid clone, pF168, was used in further studies. To confirm that pF168 bears a wild-type copy of the mug gene, both strands of the insert were sequenced using the following primers: primer 5, 5′-CTAGTT ATTGCTCAGCGGTGGCAGC-3′; and primer 6, 5′-TATAGGGGAATTGTGAGCGGATAAC-3′. The sequence of the cloned mug + was compared with the sequence in the GenBankTM data base (GenBankTM accession number U28379), and the two sequences were identical. The genetic reversion assays were performed as described previously (12Lutsenko E. Bhagwat A.S. Mutat. Res. 1999; 437: 11-20Crossref PubMed Scopus (72) Google Scholar), except in the case where the cells carried the plasmid pF168. In this case, E. colistrain (BH161), carrying the overproducer pF168, was electroporated with pAKS2. The latter plasmid contains the kan allele cloned in pACYC184 (13Wyszynski M. Gabbara S. Bhagwat A.S. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1574-1578Crossref PubMed Scopus (64) Google Scholar). Following plating, three independent colonies were picked and grown in 10 ml of LB containing 50 μg/ml carbenicillin and 20 μg/ml chloramphenicol at 37 °C until theA 550 reached 0.3. One hundred μl of each culture was used to inoculate 10 ml of prewarmed LB containing appropriate antibiotics and 30 μm IPTG. The cells were again grown till A 550 reached 0.3. The cells were centrifuged at 3000 × g for 10 min, and the cell pellet was resuspended in 1 ml of LB. Appropriate dilutions of these cultures were spread on LB plates to determine the number of viable cells, and the remaining culture was spread on kanamycin plates to determine the number of revertants. The principal source of variation in mutation frequency data is the existence of mutational “jackpot” (14Luria S.E. Delbrück M. Genetics. 1943; 28: 491-511Crossref PubMed Google Scholar). We eliminated such data points from our data sets using the following procedure: the data point suspected of being from a jackpot was set aside and the mean and S.D. of the remaining data points were calculated. If the suspected data point was greater than 3 times the S.D. away from the mean, it was declared to be a jackpot and eliminated from the data set. If the data point was within 3 S.D. of the mean, it was included in the set, and the mean and S.D. of the complete data set were used in further analysis. Appropriate E. colistrains were grown from single colonies in 5 ml of LB for 24 h at 37 °C with shaking. To determine the frequency of 5-fluorocytosine-resistant cells, the cultures were centrifuged to pellet the cells, and the pellets were washed twice in 5 ml of M63 minimal medium. The cells were ultimately resuspended in 1 ml of M63 medium and were spread on LB plates to determine the total number of viable cells. They were also spread on M63 minimal plates supplemented with 0.1% of casamino acids, 10 μg/ml 5-fluorocytosine, and 20 μg/ml each leucine, threonine, and histidine to determine the number of 5-fluorocytosine-resistant cells. The plates were incubated for 24 h at 37 °C. The mutant frequency is the number of 5-fluorocytosine resistant cells divided by the total number of viable cells. To determine frequency of rifampicin-resistant cells, the overnight cultures were centrifuged to pellet the cells, and the pellets were resuspended in 1 ml of LB. The cells were spread on LB plates or LB plates containing 100 μg/ml rifampicin. The mutant frequency is the number of rifampicin-resistant cells divided by the total number of viable cells. Five ml of overnight cultures of the appropriate E. coli strain were used to inoculate 250 ml of LB and grown at 37 °C to an A 550 of 1.0. The cells were harvested by centrifugation, and the pellet was washed with 50 ml of a buffer containing 25 mm HEPES, pH 7.6, and 1 mm ETDA. The cells were again harvested by centrifugation, and the pellet was resuspended in 20 ml of the lysis buffer (25 mm HEPES, pH 7.6, 0.5 mm ETDA, 10% glycerol, 1 mm phenylmethylsulfonyl fluoride, and 1 mm dithiothreitol). Cells were kept on ice for 30 min and sonicated five times with 30-s pulses. Between pulses, they were chilled on ice for 30 s. The cell debris was removed by centrifugation for 30 min at 12,000 rpm 4 °C, and the supernatant was divided in aliquots. The aliquots were stored at −70 °C. The DNA oligonucleotides containing uracil (oligo vsr-U) or 3,N4-ethenocytosine (εC1) were labeled with32P at the 5′ end and hybridized to the unlabeled oligomer, oligo dcm-94b, at a molar ratio 1:10 to form the duplexes with a single U:G or etheno-C:G mismatch. The duplexes were subjected to treatments with different dilutions of cell-free extracts for 30 min at 37 °C in a treatment buffer (20 mm Tris-HCl, 10 mmEDTA) and then were treated with 0.1 m NaOH or FA-PY glycosylase to cleave at the AP sites for another 30 min at 37 °C. The FA-PY glycosylase protein was purified in collaboration with B. Taffe (Wayne State University). The products were separated in a 20% sequencing gel and identified by autoradiography and by scanning with a phosphorimager. The sequences of the oligos used in these experiments were as follows: vsr-U, 5′-GACTGGCTGCTACUAGGCGAAGTGCC-3′; εC1, 5′-GACTGGCTGCTAC(εC)AGGCGAAGTGCC-3′; and dcm-94b, 5′-GGCACTTCGCCTGGTAGCAGCCAGTC. Duplex DNA containing a U·G mismatch was treated with cell-free extract prepared from ung mug + cells. The DNA was further treated with NaOH to convert the abasic sites created by the extract to strand breaks, and the products were separated on a denaturing gel. The reactions were carried out in pairs, one with the Ung inhibitor UGI (15Wang Z. Mosbaugh D.W. J. Biol. Chem. 1989; 264: 1163-1171Abstract Full Text PDF PubMed Google Scholar) and the other without the inhibitor. The use of UGI in one reaction assures that any uracil excision activity seen in that reaction must be due to an enzyme other than Ung. The cell extract contained a weak uracil glycosylase activity (Fig.1 A, lane 4) that was resistant to UGI (lane 5). This activity was reproducible and is probably the same activity reported by Gallinari and Jiricny (10Gallinari P. Jiricny J. Nature. 1996; 383: 735-738Crossref PubMed Scopus (184) Google Scholar). Compared with this weak activity, the uracil excision activity in extracts prepared from ung + mug + cells was much easier to detect. Whereas 8 μg of cell-free extract from ung cells had barely detectable activity, 38 ng of extract from ung + cells cleaved all the uracil in DNA (Fig. 1 B, lane 7). The latter activity was completely inhibited by UGI confirming that it was due to Ung. Based on these and other results we conclude that uracil excision activity in E. coli due to Ung is at least 400 times greater than uracil excision due to any other enzyme including Mug. We used a genetic reversion system involving a defective kanamycin resistance gene (13Wyszynski M. Gabbara S. Bhagwat A.S. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1574-1578Crossref PubMed Scopus (64) Google Scholar) to assess the role of Mug in avoiding spontaneous C to T mutations. The reversion to kanamycin resistance results exclusively from C to T mutations at a site for cytosine methylation and either a 5-methylcytosine to T change or C to U to T change can be studied with this system in appropriate genetic backgrounds. T·G and U·G mismatches are, respectively, the intermediates in these mutagenic pathways, and hence any excision of T or U by Mug should reduce C to T mutations. We compared the antimutagenic effects of Mug to those of Ung and Vsr using this assay. The presence of Mug did not affect mutations by either pathway (Table I). Whereas the presence of Ung reduced the reversion frequency by a factor of ∼10, Mug did not significantly affect the frequency. We have previously shown that in a mug + strain, Vsr reduces the frequency of 5-methylcytosine to T mutations by a factor of about 4 (16Bandaru B. Wyszynski M. Bhagwat A.S. J. Bacteriol. 1995; 177: 2950-2952Crossref PubMed Google Scholar). In contrast, Mug reduced these mutations only slightly, and this reduction was not statistically significant (Table I).Table IRepair of U·G and T·G mismatchesGenetic backgroundaThe following strains were used in these assays- GM31 (dcm ung +), BH157 (dcm ung+ mug), BH156 (dcm ung), BH158 (dcm ung mug), GM31 with pDCM72 (dcm+ ung+ vsr), and BH157 with pDCM72 (dcm+ ung+ vsr mug).GrowthbThe number of independent cultures used in each experiment is shown in parentheses.Revertant frequencycMean ± S.D.mug +mugdcm ung +Exponential (8)(1.4 ± 0.6) × 10−7(1.1 ± 0.6) × 10−7Stationary (4)(4.2 ± 1.2) × 10−7(4.9 ± 2.1) × 10−7dcm ungExponential (8)(1.7 ± 0.2) × 10−6(2.6 ± 1.3) × 10−6Stationary (6)(5.7 ± 2.4) × 10−6(7.0 ± 2.7) × 10−6dcm+ ung+ vsrExponential (3)(2.1 ± 0.8) × 10−6(1.3 ± 0.6) × 10−6a The following strains were used in these assays- GM31 (dcm ung +), BH157 (dcm ung+ mug), BH156 (dcm ung), BH158 (dcm ung mug), GM31 with pDCM72 (dcm+ ung+ vsr), and BH157 with pDCM72 (dcm+ ung+ vsr mug).b The number of independent cultures used in each experiment is shown in parentheses.c Mean ± S.D. Open table in a new tab In the experiments discussed above, cells were dividing at the time of their selection for kanamycin resistance. To assess the role of Mug in stationary phase of cell growth, cells were shaken at 37 °C for ∼24 h and then plated to select for kanamycin-resistant revertants. Growing the cells to stationary phase increased the reversion frequency by a factor of ∼2 compared with growing cells, but there was no significant effect of Mug on the mutant frequencies (Table I). In contrast, the Ung defect again increased the mutant frequency by a factor of approximately 10. Based on these results, we conclude that Mug does not play a significant role in the repair of U·G or T·G mismatches in E. coli. It seemed possible that the inability of mug + ung cells to repair U·G mismatches was due to inadequate expression of Mug in the cells. To see whether overexpression of Mug in the cells can reduce C to T mutations, the mug +gene was cloned in a multicopy plasmid and expressed from a bacteriophage T7 promoter. A plasmid carrying this construct was introduced into mug + ung cells along with the tester plasmid containing the kan gene. The expression of Mug was optimized by varying the concentration of IPTG used to induce the promoter, and the level of Mug in the cells was monitored by gel electrophoresis. The amount of free Mug in the cells increased with IPTG concentration, reaching a maximum between 20 and 100 μm of the inducer (Fig.2). The level of soluble Mug did not increase at higher concentrations of the inducer, probably because Mug tends to aggregate at high concentrations. 2E. Lutsenko and A. S. Bhagwat, unpublished results. The cell viability is also low at concentrations above 40 μm. Consequently, the assays were done at 30 μm IPTG. Cell extracts containing high levels of Mug are able to excise uracils from DNA (Fig. 3, lane 4). In this case, nearly all the uracil was excised from U·G mismatches regardless of the presence of UGI in the reaction (Fig. 3, lane 5). Purified Mug is also able to excise uracil from U·G mismatches (not shown). Surprisingly, induction of mug + had little effect on the frequency of C to T mutations. In one data set there was a slight decrease in mutant frequency as a result of Mug overproduction, but this effect was not reproducible (TableII). The mutant frequencies reported in this table are lower than that in Table I, because in this case thekan gene was on a low copy number plasmid. Induction of the T7 promoter with lower concentrations of IPTG or with 40 μm IPTG also did not affect the mutant frequency (not shown), suggesting that even at high concentrations, Mug cannot effectively substitute Ung to repair U·G mismatches in DNA.Table IIU·G repair in Mug overproducerGenetic backgroundaStrain BH161. The number of independent cultures used in each experiment is shown in parentheses.Revertant frequencybMean ± S.D.No IPTGWith IPTGung mug +1 (3)(3.8 ± 2.1) × 10−7(1.4 ± 0.9) × 10−72 (3)(3.5 ± 0.7) × 10−7(4.0 ± 2.2) × 10−7a Strain BH161. The number of independent cultures used in each experiment is shown in parentheses.b Mean ± S.D. Open table in a new tab We wondered whether Mug could repair endogenous DNA damage other than U·G and T·G mismatches. If this were true, a mug mutation would have a mutator phenotype. We tested this possibility by comparing the frequencies of rifampicin-resistant and 5-fluorocytosine-resistant mutants in mug + and mug strains. The results were largely negative (TableIII). The frequency of rifampicin-resistant mutants was slightly higher in a mugstrain, but the the frequency of 5-fluorocytosine-resistant mutants was the same in the two genetic backgrounds. It is possible that Mug does prevent a small number of mutations and that this is not evident in the 5-fluorocytosine-resistant mutation assay because the background frequency of mutations is high in this assay (Table III). In any case,mug is at best a very weak mutator and hence is unlikely to be important for correcting endogenous damage to DNA in E. coli.Table IIIForward mutation frequenciesGenetic backgroundMutant frequencyaMean ± S.D.Rifampicin resistance5-Fluorocytosine resistancemug +bStrains BH156 (mug +) and BH158 (mug).mug bStrains BH156 (mug +) and BH158 (mug).mug +mugung (3)cThe number of independent cultures used in each experiment is shown in parentheses.(2.4 ± 1.0) × 10−7(6.5 ± 2.9) × 10−7(4.4 ± 1.3) × 10−4(3.4 ± 1.8) × 10−4a Mean ± S.D.b Strains BH156 (mug +) and BH158 (mug).c The number of independent cultures used in each experiment is shown in parentheses. Open table in a new tab A wide range of chemicals react with bases in nucleic acids to form ethenobases. These chemicals include vinyl compounds, haloaldehydes, α-haloketones, haloalkanes, halothioketenes, halo- ketenes, and products of lipid peroxidation (17Bartsch H. Barbin A. Marion M.J. Nair J. Guichard Y. Drug Metab. Rev. 1994; 26: 349-371Crossref PubMed Scopus (151) Google Scholar, 18Guengerich F.P. Drug Metab. Rev. 1994; 26: 47-66Crossref PubMed Scopus (18) Google Scholar). After the work presented above was completed, Saparbaev and Laval purified a protein from E. coli that removes 3,N 4-ethenocytosine (εC) from DNA and showed that it was Mug (11Saparbaev M. Laval J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8508-8513Crossref PubMed Scopus (164) Google Scholar). Purified Mug excised εC from duplex DNA with εC·G pairs, but not from single-stranded DNA. It also excised uracils, but its catalytic efficiency for the removal of εC excision was 50 times higher than for the removal of uracil from U·G mismatches (11Saparbaev M. Laval J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8508-8513Crossref PubMed Scopus (164) Google Scholar). We have confirmed the ability of Mug to excise εC from DNA. Cell-free extracts from cells containing overproduction of Mug and purified Mug excised εC from DNA (not shown). We also wanted to find out whether εC removal activities other than Mug existed in E. coli. For this purpose, we used the mug strain described above. When duplex DNA containing a εC·G pair was treated with various cell-free extracts in a manner similar to that described for U·G mispairs, removal of εC was readily detected inmug + extracts (Fig.4). This treatment converted the labeled substrate to shorter products of expected length or products that were ∼1 nucleotide longer or shorter (Fig. 4, lane 2). Although the shorter product is likely to have resulted from the action of an AP endonuclease in the extract to the 5′ side of the abasic site created by Mug, the source of the longer product is not known. Regardless, these results confirm the existence of εC removal activity inE. coli (11Saparbaev M. Laval J. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8508-8513Crossref PubMed Scopus (164) Google Scholar). Furthermore, extract prepared frommug cells did not possess the ability to create abasic sites at the εC (Fig. 4, lane 3). Thus mug appears to code for the principal εC excision activity in E. coli. We have shown here that although concentrated cell extracts containing Mug can be used to show a small amount of excision of uracils from DNA, this enzyme does not act as a U·G correction enzyme in growing or stationary E. coli. Furthermore, Mug does not appear to play a significant role in repairing any other DNA damage that occurs spontaneously in the cells. Mug is clearly more efficient at excising εC from εC·G pairs, and this may be the only activity of the kind in E. coli. It is surprising that despite its ability to excise uracils from U·G mispairs, overproduction of Mug from a strong T7 promoter does not result in the reduction of C to T mutations (Table II). A possible reason for this apparent inactivity of Mug in vivo is that Mug may aggregate to form inclusion bodies. Consistent with this hypothesis we have found that purified Mug rapidly aggregates to form stable high molecular weight complexes. As a result, very little Mug may be available to repair the mismatches despite overproduction. An alternate possibility is that C to U deaminations mostly occur in single-stranded regions of the genome and that Mug is unable to excise these uracils because of its strict requirement for U·G mismatches. In contrast, Ung acts preferentially on uracils in single-stranded DNA (19Krokan H. Wittwer C.U. Nucleic Acids Res. 1981; 9: 2599-2613Crossref PubMed Scopus (152) Google Scholar, 20Leblanc J.P. Martin B. Cadet J. Laval J. J. Biol. Chem. 1982; 257: 3477-3483Abstract Full Text PDF PubMed Google Scholar) and should efficiently repair such uracils. If so, Mug is poorly suited to be a backup enzyme for Ung. At this time, the biological role of Mug in E. coli remains a matter of speculation. If the role is in the removal of εC from DNA—and this is very likely—then the lack of a strong mutator phenotype for mug is not surprising. Defects in genes that code for enzymes that repair alkylated bases also do not have a mutator phenotype. This has been interpreted to mean that there is little alkylation damage to DNA bases in exponentially growing E. coli. By analogy, it is likely that there is very little endogenous εC, or any other damaged base that may be removed by Mug, in E. coli. It is important to note that some of the enzymes involved in the repair of DNA damage caused by exogeneous agents are induced by the damaging treatment. These include the the nucleotide excision proteins and the 3-methyl adenine glycosylase II, AlkA. It would be interesting to know whether mug + is similarly inducible in response to damage to cellular DNA. We are grateful to J. Reiss (Princeton University) for providing a bacterial strain and to A. Beletskii (Wayne State University) for constructing a phage lambda lysogen.