Proc. Natl. Acad. Sci. USA
`Vol. 89, pp. 7022-7025, August 1992
`Biochemistry
`
`Evidence that MutY and MutM combine to prevent mutations by an
`oxidatively damaged form of guanine in DNA
`(8-oxo-7,8-dihydroguanlne/8-hydroxyguanine/DNA glycosylase/Fpg protein/mismatch repair)
`MARK LEO MICHAELS*, CHRISTINA CRUZ*, ARTHUR P. GROLLMANt, AND JEFFREY H. MILLER*t
`*Department of Microbiology and Molecular Genetics and the Molecular Biology Institute, University of California, Los Angeles, CA 90024; and tDepartment
`of Pharmacological Sciences, State University of New York, Stony Brook, NY 11794-8651
`Communicated by Charles Yanofsky, May 6, 1992 (received for review January 30, 1992)
`
`It has been previously shown both in vivo and
`ABSTRACT
`in vitro that DNA synthesis past an oxidatively damaged form
`of ganine, 7,8-dihydro-8-oxoguanine (8-oxoG), can result in
`the nusincorporation of adenine (A) opposite the 8-oxodG. In
`this study we show that MutY glycosylase is active on a
`site-specific, oxidatively damaged A/8-oxoG mispair and that
`it removes the undamaged adenine from this mispair. Strains
`that lack active MutY protein have elevated rates of G-C T-A
`transversions. We find that the mutator phenotype of a inut)
`strain can be fully complemented by overexpressing MutM
`protein (Fpg protein) from a plasmid clone. The MutM protein
`removes 8-oxoG lesions from DNA. In addition, we have
`isolated a strain with a chromosomal mutation that suppresses
`the mutY phenotype and found that this suppressor also
`overexpresses MutM. Finally, a mutY mutM double mutant has
`a 25- to 75-fold higher mutation rate than either mutator alone.
`The data strongly suggest that MutY is part of an intricate
`repair system directed against 8-oxoG lesions in nucleic acids
`and that the primary function of MutY in vivo is the removal
`of adenines that are misincorporated opposite 8-oxoG lesions
`during DNA synthesis.
`
`All organisms that use molecular oxygen need to defend
`themselves from the reactive byproducts of oxygen metab-
`olism. Reactive oxygen species such as superoxide, hydro-
`gen peroxide, and hydroxyl radical can be produced by the
`incomplete reduction of oxygen during aerobic metabolism
`(1). These oxidants can also be generated from lipid perox-
`idation, from phagocytic cells, and by exposure to radiation
`(1). Although organisms have developed systems to control
`active oxygen species, those that escape the primary de-
`fenses can damage nucleic acids and other cellular macro-
`molecules.
`Oxidative damage to DNA has been estimated at 104
`lesions per cell per day in humans and an order of magnitude
`higher in rodents (2). Not surprisingly, organisms have de-
`veloped a second line of defense to repair the oxidative
`damage to DNA. One of the lesions that is actively removed
`is 7,8-dihydro-8-oxoguanine (8-oxoG, also termed 8-hydroxy-
`guanine). The lesion is removed by the glycosylase/apurinic
`endonuclease activity of the MutM protein [Fpg protein or
`8-oxoG glycosylase] (3, 4). Recently, it has been shown that
`DNA synthesis past 8-oxoG can result in the misincorpora-
`tion of adenine opposite the damaged guanine (5-7) and that
`inactivation of the mutM (_fpg) gene leads specifically to G-C
`T-A transversions (4, 8).
`In this paper we present evidence that Escherichia coli has,
`at least in the case ofthe 8-oxoG lesion, a third line ofdefense
`against oxidative damage to DNA. Unlike the first two forms
`of protection, which target the active oxygen species or the
`damage it inflicts on nucleic acids, this activity leads to the
`
`correction of errors that are induced by replication of tem-
`plates containing 8-oxoG lesions. We find that MutY protein,
`originally identified as an adenine glycosylase active on A/G
`mispairs (9), can also remove the undamaged adenine from
`A/8-oxoG mispairs. Strains that lack active MutY protein
`have elevated rates of G-C -+ TEA transversions (10). Over-
`expressing the MutM protein from a plasmid clone can
`completely complement the mutator phenotype of a mutY
`strain. We have isolated a strain with a chromosomal muta-
`tion that suppresses mut Yand found that it has 15-fold greater
`MutM activity than the parental strain. Finally, a mutM mutY
`double mutant shows a 20-fold higher rate of G-C -* TEA
`transversions than the sum of G-C -* T-A transversions in
`either mutant alone. These observations suggest that the
`primary function of MutY glyrosylase in vivo is the removal
`of adenines that have been misincorporated opposite 8-oxoG
`lesions during DNA replication, and that MutY and MutM
`are part of a multiple line of defenses against oxidative
`damage to DNA.
`
`MATERIALS AND METHODS
`Ofigodeoxynucleotides. The 23-mer oligodeoxynucleotides,
`including the one containing the site-specific 8-oxoG lesion,
`were synthesized and purified as described (6, 11). The
`purified oligomers (4 pmol) were 32P-labeled at the 5' termi-
`nus with 10 units of T4 polynucleotide kinase in the presence
`of 7 pmol of [P32P] ATP (6 Ci/humol; New England Nuclear;
`1 Ci = 37 GBq) for 10 min at 370C. After heat inactivation at
`900C for 7 min, the labeled primers were annealed with an
`excess of unlabeled complementary oligodeoxynucleotide at
`900C in 50 mM NaCl for 2 min, cooled to room temperature,
`and then incubated on ice. The sequences of the oligodeox-
`ynucleotides are shown below (nucleotides involved in mi-
`spair formation are underlined; GO, 8-oxoG).
`(1) 5'-CTCTCCCTTCQQCTCCTllCCTCT-3'
`
`(2) 5'-CTCTCCCTTCGCTCCTTTCCTCT-3'
`
`(3) 5'-AGAGGAAAGGAGAGAAGGGAGAG-3'
`
`(4) 5'-AGAGGAAAGGAGCGAAGGGAGAG-3'
`MutY Glycosylase Assay. Glycosylase reactions were car-
`ried out in a solution (10 ILI) containing 20 mM Tris HCl (pH
`7.6), 0.5 pug of bovine serum albumin, 10 mM EDTA, 45 ng
`of MutY glycosylase [purified essentially as previously de-
`scribed (9)], and 20 fmol of 23-mer duplex with the indicated
`mispair. After incubation at 370C for 30 min, reactions were
`terminated by the addition of2 jul of 1 M NaOH and incubation
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement"
`in accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`Abbreviation: 8-oxoG, 7,8-dihydro-8-oxoguanine.
`tTo whom reprint requests should be addressed.
`
`7022
`
`GDX 1006
`
`

`

`Biochemistry: Michaels et al.
`at 900C for 4 min. This also served to cleave any apurinic/
`apyrimidinic sites generated by the glycosylase. Four micro-
`liters of loading dye (95% formamide/20 mM EDTA/0.05%
`bromophenol blue/0.05% xylene cylanol FF) was added to the
`mix and aliquots were run in denaturing (7 M urea) 15%
`polyacrylamide gels.
`Strains and Plasmids. E. coli strain GT100 is ara, A(gpt-
`lac)S, rpsL, [F'1acIqL8, proA+B+]. Strain CC104 is ara,
`A(gpt-4ac)S, [F'lacI378, lacZ461, proA+B+] (12). CSH115 is
`ara, A(gpt-lac)5, rpsL, mutS::mini-TnlO and CSH116 is ara,
`A(gpt-4ac)5, rpsL, mutDS zaeSO2::TnlO; these strains are
`further described by Miller (13). Plasmid pKK223-3 was
`purchased from Pharmacia, and pKK-Fapy2 carries a wild-
`type mutM (fpg) gene (4). All media and genetic manipula-
`tions, unless otherwise stated, were as described (14).
`Complementation Tests. Eight independent cultures con-
`taining plasmid pKK223-3 or pKK-Fapy2 in strains GT100,
`GT100 mutY::mini-TnJO, GT100 mutM, or GT100
`mutS::mini-TnJO were grown overnight in LB medium sup-
`plemented with ampicillin (100lg/ml), and 50-,lI samples
`were plated onto LB with rifampicin (200 ,g/ml). After
`overnight incubation at 37°C, Rifr colonies were counted.
`MutM Protein Assay. Crude extracts were obtained by
`growing Supl7 [a strain carrying a suppressor of mutY(C.C.,
`M.L.M., and J.H.M., unpublished work)] in LB medium.
`Cells (30 OD6co
`units) were collected by centrifugation,
`resuspended in 2 ml 50 mM Tris HCl, pH 7.5/20 mM f3-mer-
`captoethanol/1 mM phenylmethylsulfonyl fluoride, and fro-
`zen at -70°C. The solution was thawed and sonicated on a
`Fisher Sonic Dismembrator model 300 on maximum output
`for two 30-sec bursts. The cell debris was removed by
`centrifugation (27,000 x g) for 20 min at 4°C. The supernatant
`was taken and glycerol was added to a concentration of 5%
`(vol/vol). The protein content of the extracts was measured
`with the BioRad protein assay kit using bovine serum albu-
`min as the standard. Various amounts of extract were incu-
`bated with 20 fmol of 23-mer duplex containing a C/8-oxoG
`pair with a 32p label in the 8-oxoG-containing strand (oligo-
`nucleotides 4 and 1; see above). The reaction was incubated
`at 37°C for 30 min in 10 ,ul of 20 mM Tris-HCl, pH 7.6/10 mM
`EDTA containing bovine serum albumin at 50 ,g/ml. The
`reaction was stopped by the addition of 3 1l of loading dye
`and by heating at 1000C for 2 min. An aliquot of the reaction
`was loaded onto a denaturing 15% polyacrylamide gel. Au-
`toradiographs were quantitated by measuring transmittance
`of substrate and product bands with a Bio-Rad 620 video
`densitometer.
`Mutational Specificity Tests. Four or more independent
`cultures of CC104 and its derivatives were grown overnight,
`and 50 IlI was plated onto LB with rifampicin (100 ,ug/ml) and
`100 1l onto minimal lactose medium. After overnight incu-
`bation at 37°C, Rifr colonies were counted. Lac+ colonies
`were counted after 2 days of incubation at 37°C. For the
`double mutant (mutYmutM), 10-2 dilutions were done before
`plating. Lac+ colonies were somewhat difficult to count for
`the double mutant due to Lac+ colonies arising on the plate
`during incubation. Incubation times were reduced to combat
`this problem. This problem does not occur with the Rifr
`mutants.
`
`RESULTS
`MutY Glycosylase Activity on A/8-oxoG Mispairs. Du-
`plexes (23-mers) containing an undamaged A/G mispair or a
`site-specific A/8-oxoG mispair were tested as substrates for
`MutY glycosylase. The glycosylase was active on the duplex
`containing the A/8-oxoG mispair and removed the undam-
`aged adenine from the duplex (Fig. 1, lane 7). MutY acted
`strictly as a glycosylase in this reaction-no endonuclease
`activity was detected. As previously observed (9), MutY was
`
`Proc. Natl. Acad. Sci. USA 89 (1992)
`
`7023
`
`-MutY
`+MutY
`Mispair Ai Ab* 1GO bo A A* A;0A/o*
`2
`3
`4
`5
`6
`8
`7
`1
`10
`40
`
`m
`
`Substrates
`
`Products
`
`40
`
`4w
`
`Activity of MutY glycosylase on mispairs containing a
`FIG. 1.
`site-specific 8-oxoG lesion. Glycosylase reactions were carried out
`as described in Materials and Methods. Blank reaction mixtures
`(lanes 1-4) contained everything except the glycosylase. Reactions
`were terminated by the addition of 2 ,ul of 1 M NaOH and incubation
`at 90°C for 4 min. This also served to cleave any apuninic sites
`generated by the glycosylase. The progress of the glycosylase
`reaction can therefore be monitored by a change in migration of the
`cleaved products. The A oligomer is sequence no. 3 (see Materials
`and Methods); the G and GO (8-oxoG) oligomers are nos. 2 and 1,
`respectively. The star indicates which strand in the duplex is
`32P-labeled.
`
`also active on the undamaged adenine of an A/G mispair and
`did not cause strand cleavage at the apurinic site (lane 5).
`All of the other possible combinations of mispairs with
`8-oxoG were tested as potential substrates for MutY glyco-
`sylase. MutY does not removecytosine, guanine, or thymine
`opposite 8-oxoG, nor does it remove 8-oxoG from any of the
`damaged duplexes (data not shown). MutY was active only
`on the A/8-oxoG and A/G mispairs.
`The relative activity of MutY glycosylase on an A/8-oxoG
`or A/G mispair duplex substrate was determined. MutY
`removed the undamaged adenine from the A/G and A/8-
`oxoG mispairs at approximately the same rate (Fig. 2). The
`glycosylase specific activity on these substrates was >26
`nmol/hr per mg of protein. This is comparable to the glyc-
`osylase specific activity observed for purified MutY glyco-
`sylase on a phage fl-derived heteroduplex containing an A/G
`mispair (1.35 nmol/hr per mg; ref. 9).
`Complementation Tests. Previous work has shown that
`mutY (10) and mutM (fpg) (8) have nearly identical mutation
`spectra in vivo and that both specifically increase G-C -- T-A
`transversions. Therefore, we suspected that the two proteins
`could be part of a common repair pathway. The above
`biochemical assay suggested that 8-oxoG might be common
`to both repair proteins. In the proposed scheme, the MutM
`protein would remove 8-oxoG lesions in DNA, and MutY
`A*/GO
`Mispair
`5.0 1.0 0.5 0.1'
`MutY (ng)
`8
`9
`6
`7
`
`A*/G
`5.0 1.0 0.5 0.1'
`2
`3
`4
`5
`
`0
`1
`*M'
`
`Substrates
`
`Products
`
`4* _ sQr_
`iw. now ki_. 4m
`
`Relative activity of MutY glycosylase on A/G and
`FIG. 2.
`A/8-oxoG substrates. The blank reaction (lane 1) and glycosylase
`reactions (lanes 2-9) were performed on duplex 23-mers containing
`the indicated mismatch and analyzed as described in the legend to
`Fig. 1 except that the amount of MutY glycosylase was varied as
`shown in the figure. The A* oligomer in each duplex is no. 3
`(Materials and Methods); the G and GO (8-oxoG) oligomers are nos.
`2 and 1, respectively. The blank reaction (no glycosylase) was
`performed on A*/G.
`
`GDX 1006
`
`

`

`7024
`
`Biochemistry: Michaels et al.
`
`Complementation of a mutY strain with a
`Table 1.
`MutM-overproducing plasmid
`
`No. of Rifr colonies per 108
`cells
`
`Strain
`
`pKK-Fapy2
`pKK223-3
`1±1
`it 1
`GT100
`57 ± 13
`GT100 mutY:: mini-Tn)O
`2 + 1
`1± 1
`13 ± 5
`GT100 mutM
`102 ± 21
`71 ± 18
`GT100 mutS:: mini-TnIO
`Eight independent cultures were grown to saturation and plated
`onto LB medium with rifampicin. The average number of Rifr
`colonies (±SD) per 108 cells is reported. Plasmid pKK223-3 is the
`vector and pKK-Fapy2 is a MutM overproducer (4).
`glycosylase would remove any adenines that were misincor-
`porated opposite 8-oxoG lesions that were not removed
`before replication.
`The results of complementation tests support this theory.
`By overexpression of MutM protein in a mutY strain, the
`mutation rate of mutY was reduced to wild-type levels (Table
`1). Overexpression of MutM protein in a mismatch repair-
`deficient strain (mutS) had no effect on its mutation rate. This
`control shows that the overproducer does not affect a mu-
`tator strain that causes mutations by an unrelated pathway.
`As expected, the clone complemented a mutM strain, and the
`vector alone had no effect on the mutation rate of any strain.
`mutY Double Mutant. Consis-
`Characterization of a mut
`tent with the idea that MutY glycosylase and the MutM
`protein protect the cell from the potentially mutagenic
`8-oxoG lesion, a strain that lacks both defense systems has an
`extremely high rate of GC -- TA transversions. We used P1
`phage to transfer a mutY::mini-TnlO marker into a mutM
`strain and found that, like mutM and mutY, the double mutant
`was very specific for stimulating G-C -- T-A transversions
`(Table 2). However, the mutation rate of the double mutant,
`as judged by both the generation of Rifr colonies and the rate
`of GC -. T-A transversions in a Lac+ reversion assay, is
`20-fold higher than the sum of the separate mutators, sug-
`gesting that they are involved in a related repair system
`(Table 3).
`The importance of the repair system is illustrated by
`comparing the mutation rate of a mutM mutY double mutant
`to the mutation rates obtained when other well-characterized
`error-avoidance systems are disabled. Strains lacking the
`polymerase III editing subunit (mutD) or the methyl-directed
`mismatch repair system (mutS) are the most potent mutator
`strains in E. coli. Asjudged by the formation ofRifr colonies,
`the mutM mut Y double mutant is just as strong a mutator as
`mutD and about an order of magnitude stronger than a strain
`lacking the mismatch repair system (Table 3).
`Characterization of a Suppressor of mutY. We have char-
`acterized a suppressor of mutYthat maps near to but distinct
`Mutational specificity of mutM mutY strains
`Table 2.
`No. of
`No. of
`Lac+
`Lac+
`Reversion
`revertants
`Strain
`Strain
`revertants
`event
`CC101 mutM mutY
`GCG CC101
`1
`1
`A-T
`G-C -AT CC102
`2
`CC102 mutM mutY
`2
`GCO CC103
`<1
`CC103 mutM mutY
`<1
`G-C
`5300
`CC104 mutM mutY
`G-C TA CC104
`2
`1
`1
`CC105 mutM mutY
`A-T -- T-A CC105
`A-T -. GC CC106
`3
`CC106 mutM mutY
`<1
`Four or more independent cultures of each strain were grown to
`saturation in LB medium and plated onto minimal lactose medium.
`Average number of Lac+ colonies per 108 cells is recorded above.
`Strains in this table are streptomycin-resistant (rpsL) versions of the
`CC101-106 series (12).
`
`Proc. NatL Acad Sci. USA 89 (1992)
`
`Mutation frequency of mutM, mutY, and mutM
`Table 3.
`mutY strains
`
`Strain
`
`No. of Rifr
`No. of Lac+
`colonies
`revertants
`5-10
`3
`CC104
`151
`25
`CC104 mutM
`290
`62
`CC104 mutY
`8200
`1900
`CC104 mutM mutY
`760
`ND
`CSH115 (mutS)
`4900
`ND
`CSH116 (mutD)
`Mutation frequency of the double mutant is compared with those
`of the separate mutM and mutY mutants as well as with the mutation
`frequency of strains lacking the polymerase III editing function
`(mutD) or the mismatch repair system (mutS). Cultures were grown
`to saturation in LB medium and plated onto minimal lactose medium
`and LB with rifampicin. Average numbers of Lac+ and Rjfr colonies
`per 108 cells are recorded. ND, not determined.
`from the mutM gene (C.C., M.L.M., and J.H.M., unpub-
`lished work). We suspected that this antimutator candidate,
`Supl7, might overexpress MutM protein based upon our
`previous complementation results. In fact, when the extracts
`were tested for glycosylase/apurinic endonuclease activity
`on a 23-mer duplex containing a site-specific C/8-oxoG pair,
`the Supl7 extract had 15-fold greater activity than the parent
`strain. These results will be described in detail elsewhere.
`
`DISCUSSION
`The attack of reactive oxygen species on DNA poses a
`substantial threat to the cell. Endogenous oxidants generated
`by the incomplete reduction of oxygen or by lipid peroxida-
`tion can damage DNA and may be a major cause of the
`physiological changes associated with aging and cancer (2).
`One of the products of oxidative attack on DNA is the
`8-oxoG lesion. It has been determined that rat liver cells have
`a steady-state level of over 4 x 105 8-oxoG lesions per cell,
`yet it is estimated that 8-oxoG lesions may represent only 5%
`of the total oxidative damage to DNA (2). Even more
`significant is the finding that 8-oxoG lesions can lead to
`replication errors. In vivo and in vitro studies have shown that
`adenine is frequently misincorporated opposite oxidatively
`damaged guanine residues in DNA (5-7). In a mutM strain the
`protein that removes 8-oxoG lesions in DNA is not active (3,
`4). The elevated level of G-C -- TA transversions in a mutM
`strain is thus due to the misinsertion of adenine residues
`opposite the accumulated 8-oxoG lesions in the parental
`strand during DNA replication. MutM (Fpg protein) can also
`excise formamidopyrimidine lesions in DNA (20).
`The 8-oxoG lesion is not only present in chromosomal
`DNA but is also found in the nucleotide pool as 8-oxo-dGTP.
`This damaged nucleotide is potentially mutagenic, as 8-oxo-
`dGTP can be frequently misincorporated opposite template
`adenines (15). The MutT protein hydrolyzes 8-oxo-dGTP to
`8-oxo-dGMP, thus eliminating the mutagenic substrate from
`the nucleotide pool (15). A strain that lacks active MutT
`C-G transversions (16).
`protein has elevated levels of A-T
`These transversions are presumably due to the misincorpo-
`ration of 8-oxo-dGTP opposite template adenines. Strains
`deficient in MutM or MutT protein have increased frequen-
`cies of opposite transversions. In a mutM strain, 8-oxoG
`lesions accumulate in the chromosome and lead to the
`misincorporation of adenine opposite template 8-oxoG le-
`sions and produce G-C -) TA transversions, while in a mutT
`strain 8-oxo-dGTP nucleotides are misinserted opposite tem-
`plate adenines, resulting in A-T -+ CG transversions.
`Cells have multiple lines of defense against oxidative
`damage to DNA. The primary line of defense guards against
`the active oxygen species themselves. Oxidants can be
`
`GDX 1006
`
`

`

`Biochemistry: Michaels et al.
`
`O
`
`H
`
`NH2
`
`dR
`
`B r,^,
`
`+ Oxidative Stress
`1 1X 1 11
`r I
`I
`G111GIII
`Replicatio
`
`MUMGO
`
`I
`
`XI
`
`I
`
`I
`
`1,I
`
`I
`
`1
`
`1
`
`1
`
`1
`
`utY
`
`I
`
`I I~I
`
`I
`Repair
`
`I
`
`\
`
`~~~II IG01 I
`Repair
`
`I
`
`,I
`
`I
`
`I D
`
`Role of MutM and MutY in the 8-oxoG repair system. (A)
`FIG. 3.
`Structure of the predominant tautomeric form of the 8-oxoG lesion.
`dR, deoxyribose. (B) Oxidative processes can lead to 8-oxoG lesions
`in DNA. The MutM protein removes 8-oxoG lesions and subsequent
`repair restores the original G-C base pair. If the 8-oxoG lesion (GO)
`is not removed before replication, translesion synthesis can be
`accurate or inaccurate. Accurate translesion synthesis results in a
`C-8-oxoG pair-a substrate for the MutM protein. However, inac-
`curate translesion synthesis leads to the misincorporation of dAMP
`opposite the 8-oxoG lesion (5-7). MutY glycosylase removes the
`misincorporated dA from the A/8-oxoG mispairs that result from
`error-prone replication past the 8-oxoG lesion. Repair polymerases
`are less error-prone during translesion synthesis (5) and can lead to
`a C-8-oxoG pair-a substrate for MutM.
`
`eliminated by various enzymatic and nonenzymatic systems
`such as superoxide dismutase, catalase, ascorbic acid, and
`j-carotene (17). However, active oxygen species that escape
`these primary defenses can damage nucleic acids and other
`cellular macromolecules. The second line ofdefense works to
`remove the oxidative damage from nucleic acids. The MutM
`and MutT proteins are examples of this type of defense.
`Exonuclease III, endonuclease IV, and the excision nuclease
`and exonuclease activities associated with UvrAB are further
`examples of enzymes that can repair oxidized DNA (18-20).
`The MutY protein represents a third level of protection
`against oxidative damage to DNA. Unlike the other defenses,
`which seek to neutralize the reactive oxygen species or repair
`the damage those species cause to nucleic acids, MutY
`glycosylase helps to correct the errors that result from the
`replication of DNA containing oxidative damage. It removes
`undamaged adenines that are misincorporated opposite tem-
`plate 8-oxoG lesions in DNA.
`Although MutY glycosylase can remove the undamaged A
`from both an A/8-oxoG mispair or an undamaged A/G
`mispair duplex in DNA (Fig. 1), our results suggest that its
`primary function in vivo is the removal of the A from the
`oxidatively damaged mispair, A/8-oxoG. First, overexpres-
`sion of MutM protein, which removes 8-oxoG lesions, com-
`pletely complements a mutY strain (Table 1). Similarly, a
`
`Proc. Natl. Acad. Sci. USA 89 (1992)
`
`7025
`
`chromosomal suppressor of the mutator phenotype of mutY
`had 15-fold greater glycosylase/apurinic endonuclease activ-
`ity on an 8-oxoG substrate than the parent strain. Finally, a
`mutM mutYdouble mutant has an extremely high G-C -- T-A
`mutation rate (Tables 2 and 3). The mutation rate is about
`20-fold higher than would be expected if the genes were
`involved in unrelated repair mechanisms. In fact, the muta-
`tion rate of the double mutant suggests that although the only
`mutagenic substrate this system handles is 8-oxoG, it is
`nonetheless one of the more important error-avoidance sys-
`tems in E. coli.
`The A/8-oxoG mispair represents an oxidatively damaged
`purine/purine mispair that is not efficiently removed by the
`proofreading function of polymerase III and appears to
`require a devoted repair system to prevent the oxidatively
`damaged guanine from introducing mutations into the chro-
`mosome. An intricate mechanism involving the MutM,
`MutT, and MutY proteins has evolved to protect the cell from
`the mutagenic effect of 8-oxoG (see Fig. 3).
`
`We thank V. Bodepudi and R. Rieger for preparing the oligonu-
`cleotides containing the 8-oxoG lesion. We thank J. Tchou for helpful
`discussions and H. Hsiao for technical assistance. This work was
`supported by Grant GM32184 from the National Institutes of Health
`to J.H.M., Grant CA47995 from the National Cancer Institute to
`A.P.G., and a California Institute for Cancer Research Fellowship to
`M.L.M.
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`Halliwell, B. & Gutteridge, J. M. C. (1989) Free Radicals in
`Biology and Medicine (Clarendon, Oxford), 2nd Ed.
`Fraga, C. G., Shigenaga, M. K., Park, J.-W., Degan, P. &
`Ames, B. N. (1990) Proc. Nati. Acad. Sci. USA 87,4533-4537.
`Tchou, J., Kasai, H., Shibutani, S., Chung, M.-H., Grollman,
`A. P. & Nishimura, S. (1991) Proc. Natl. Acad. Sci. USA 88,
`4690-4694.
`Michaels, M. L., Pham, L., Cruz, C. & Miller, J. H. (1991)
`Nucleic Acids Res. 19, 3629-3632.
`Shibutani, S., Takeshita, M. & Grollman, A. P. (1991) Nature
`(London) 349, 431-434.
`Moriya, M., Ou, C., Bodepudi, V., Johnson, F., Takeshita, M.
`& Grollman, A. P. (1991) Mutat. Res. 254, 281-288.
`7. Wood, M. L., Dizdaroglu, M., Gajewski, E. & Essigman, J. M.
`(1990) Biochemistry 29, 7024-7032.
`Cabrera, M., Nghiem, Y. & Miller, J. H. (1988) J. Bacteriol.
`170, 5405-5407.
`Au, K. G., Clark, S., Miller, J. H. & Modrich, P. (1989) Proc.
`Natl. Acad. Sci. USA 86, 8877-8881.
`Nghiem, Y., Cabrera, M., Cupples, C. G. & Miller, J. H. (1988)
`Proc. Natl. Acad. Sci. USA 85, 2709-2713.
`Bodepudi, V., Iden, C. R. & Johnson, F. (1991) Nucleotides
`Nucleosides 10, 755-761.
`Cupples, C. G. & Miller, J. H. (1989) Proc. Natl. Acad. Sci.
`USA 86, 5345-5349.
`Miller, J. H. (1992) A Short Course in Bacterial Genetics (Cold
`Spring Harbor Lab., Cold Spring Harbor, NY).
`Miller, J. H. (1972) Experiments in Molecular Genetics (Cold
`Spring Harbor Lab., Cold Spring Harbor, NY).
`Maki, H. & Sekiguchi, M. (1992) Nature (London) 355, 273-
`275.
`Yanofsky, C., Cox, E. C. & Horn, V. (1966) Proc. Natl. Acad.
`Sci. USA 55, 274-281.
`Degan, P., Shigenaga, M. K., Park, E.-M., Alperin, P. E. &
`Ames, B. N. (1991) Carcinogenesis 12, 865-871.
`Demple, B., Johnson, A. & Fung, D. (1986) Proc. Natl. Acad.
`Sci. USA 83, 7731-7735.
`Lin, J. J. & Sancar, A. (1989) Biochemistry 28, 7979-7984.
`Czeczot, H., Tudek, B., Lambert, B., Laval, J. & Boiteux, S.
`(1991) J. Bacteriol. 173, 3419-3424.
`
`8.
`
`9.
`
`10.
`
`11.
`
`12.
`
`13.
`
`14.
`
`15.
`
`16.
`
`17.
`
`18.
`
`19.
`20.
`
`GDX 1006
`
`