Oxidative Nucleotide Damage: Consequences and Prevention

Oxidative nucleotide damage: consequences and prevention

Mutsuo Sekiguchi*,1 and Teruhisa Tsuzuki2

1Biomolecular Engineering Research Institute, Suita, Osaka 565-0874, Japan; 2Department of Medical Biophysics and Radiation Biology, Faculty of Medical Sciences, Kyushu University, Fukuoka 812-8582, Japan

8-Oxoguanine (8-oxo-7,8-dihydroguanine) is produced in Nishimura, 1984; Fraga et al., 1990). Unlike other DNA, as well as in nucleotide pools of cells, by reactive oxidative DNA damage, such as thymine glycol and 5’, oxygen species normally formed during cellular meta- 8-purine cyclodeoxynucleoside (Evans et al., 1993; bolic processes. 8-Oxoguanine nucleotide can pair with Brooks et al., 2000; Kuraoka et al., 2000), 8- cytosine and adenine nucleotides with an almost equal oxoguanine does not block DNA synthesis, rather it efficiency, then transversion mutation ensues. MutT induces base mispairing. 8-Oxoguanine can pair with protein of Escherichia coli and related mammalian both cytosine and adenine during DNA synthesis, and protein MTH1 specifically degrade 8-oxo-dGTP to 8- this mispairing could contribute significantly to oxo-dGMP, thereby preventing misincorporation of 8- spontaneous mutations in genomic DNA (Shibutani oxoguanine into DNA. The bacterial and mammalian et al., 1991; Smith, 1992). enzymes are close in their size and share a highly Studies on Escherichia coli mutator mutants revealed conserved region consisting of 23 residues with 14 that cells possess elaborate mechanisms that can identical amino acids. Following saturation mutagenesis prevent mutations caused by oxidation of guanine of this region, most of these residues proved to be residues of DNA. 8-Oxoguanine residues in DNA can essential to exert 8-oxo-dGTPase activity. Gene target- be removed by an enzyme that is coded by the mutM ing was done to establish MTH1-deficient cell lines and gene (Cabrera et al., 1988; Chung et al., 1991; mice for study. When examined 18 months after birth, a Michaels et al., 1991; Bessho et al., 1992), while greater number of tumors were formed in the lungs, MutY protein removes adenine from adenine : 8- livers, and stomachs of MTH17/7 mice, as compared oxoguanine pairs, which are produced as a result of with findings in wild-type mice. These proteins protect misincorporation of adenine opposite to 8-oxoguanine genetic information from untoward effects of threats of (Au et al., 1988, 1989; Nghiem et al., 1988; Michaels et endogenous oxygen. al., 1992). Thus, two proteins, MutM and MutY, act Oncogene (2002) 21, 8895 – 8904. doi:10.1038/sj.onc. consecutively at the site of the oxidized guanine 1206023 residue in DNA to prevent the occurrence of specific transversion mutation (Tchou and Grollman, 1993). In Keywords: DNA replication fidelity; oxygen radicals; higher organisms, similar enzyme activities have been oxidized guanine base; mutagenesis; carcinogenesis; detected, and may account for the rapid elimination of genomic stability 8-oxoguanine from chromosomal DNA. These include two types of MutM-related proteins, OGG1 and AtMMH, which carry 8-oxoguanine-DNA glycosylase activity (van der Kemp et al., 1996; Nash et al., 1996; Introduction Ohtsubo et al., 1998; Boiteux and Radicella, 1999), and MYH protein that excises adenine paired with 8- Reactive oxygen species, such as superoxide, hydrogen oxoguanine in DNA (Yeh et al., 1991; Slupska et al., peroxide, hydroxyl radical and singlet oxygen, are 1996). produced through normal cellular metabolism, and Oxidation of guanine proceeds also in the cellular formation of such radicals is further enhanced by nucleotide pool, and 8-oxo-dGTP, the oxidized form of ionizing radiations and by various chemicals (Ames dGTP, is the mutagenic substrate for DNA synthesis. and Gold, 1991; Henle and Linn, 1997). Nucleic acids It can be incorporated opposite adenine or cytosine exposed to oxygen radicals generate various modified residues of template DNA, the result being A : T to bases, and more than 20 different types of oxidatively C : G and G : C to T : A transversions (Maki and altered purines and pyrimidines have been detected Sekiguchi, 1992; Cheng et al., 1992). However, in (Gajewski et al., 1990; Demple and Harrison, 1994). normally growing cells, the frequency of these types of Among them, 8-oxo-7, 8-dihydroguanine (8-oxogua- mutations remains low, owing to the action of enzymes nine) is the most abundant, and seems to play critical degrading such mutagenic substrates (Schaaper et al., roles in mutagenesis and in carcinogenesis (Kasai and 1986; Mo et al., 1991). The MutT protein of E. coli hydrolyzes 8-oxo-dGTP to 8-oxo-dGMP, thereby preventing misincorporation of 8-oxoguanine into *Correspondence: M Sekiguchi; E-mail: [email protected] DNA (Maki and Sekiguchi, 1992). A similar enzyme Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8896 activity has been detected in mammalian cells, and the anti and syn forms, the former pairing with cytosine protein responsible was named MTH1 (Mo et al., 1992; and the latter pairing with adenine. When 8-oxo-dGTP Sakumi et al., 1993; Furuichi et al., 1994). was added to an in vitro DNA replication system, 8- It has been proposed that one early step in the oxo-dGMP was indeed incorporated opposite cytosine progression of human cancers is elevation of the rate and adenine residues of the template, with almost equal of spontaneous mutation, that is, acquisition of a frequencies. The MutT protein possesses enzyme mutator phenotype (Ionov et al., 1993; Jackson and activity which hydrolyzes 8-oxo-dGTP to 8-oxo- Loeb, 1998; Stoler et al., 1999). If changes in dGMP, thereby preventing the misincorporation of 8- spontaneous mutation rates are indeed involved in oxoguanine into DNA. carcinogenesis, it is expected that defectiveness in the Here, one must emphasize the importance of eliminat- MTH1 or related genes would increase the likelihood ing the oxidized form of guanine base from DNA. 8- of tumor occurrence. To address this question, targeted Oxoguanine was first described to be a minor modified disruption of the MTH1 as well as the OGG1 gene has base found in DNA treated with heated glucose (Kasai been done (Klungland et al., 1999; Minowa et al., and Nishimura, 1984). This modified base was later 2000; Tsuzuki et al., 2001). These studies are revealed to be present in untreated DNA, albeit at levels significantly important in terms of multiple defence no higher than approximately three molecules of 8- mechanisms against induction and progression of oxoguanine/106 guanine residues in the chromosomal tumors in animal bodies. DNA (Tajiri et al., 1995). It thus seems that the level of reactive oxygen species produced by cellular metabolic intermediates may be sufficient to oxidize the guanine MutT-related error avoidance mechanism for DNA base of the nucleotide pool as well as that of DNA, even synthesis in normally growing cells. There are at least three steps required to prevent errors Spontaneous mutagenesis caused by the oxidation of during DNA replication; (1) selection of a nucleotide guanine nucleotides can be separated into two path- complementary to the template by DNA polymerase, ways: one starting from oxidation of guanine in DNA (2) removal of a misincorporated non-complementary and the other from oxidation of guanine nucleotide in nucleotide by an editing nuclease, associated with the precursor pool (Figure 1). The oxidation of DNA DNA polymerase, and (3) correction of a misincorpo- forms the 8-oxoG : C pair which would induce a rated nucleotide by the post-replicational mismatch G:C?T : A transversion if the 8-oxoG is not corrected repair system (Kornberg and Baker, 1992). Recently, by MutM (Fpg) protein. When 8-oxoG remains another process for preventing replication errors was unrepaired until DNA polymerase arrives at the lesion, revealed. This error-avoiding mechanism functions at dAMP would be inserted opposite the mutagenic the pre-replicational step by degrading a naturally lesion. However, in most cases, the resulting 8-oxoG : A occurring mutagenic substrate for DNA synthesis pair is reversed back to 8-oxoG : C by the action MutY (Maki and Sekiguchi, 1992). Here, MutT protein plays protein which removes the adenine base from the 8- a major role while MutM and MutY participate in the oxoG : A mispair. These dual defence mechanisms process for providing a unique mutational specificity to against mutagenesis were first proposed by Michaels the mutT7 mutant (Tajiri et al., 1995). et al. (1992), based on findings that the combination of Among the many mutators found in E. coli, mutT mutM and mutY mutations resulted in synergistic has drawn particular attention. MutT is one of the first mutator effects specific for G : C?T : A transversion mutators found in organisms (Treffers et al., 1954) and and that a multi-copy plasmid carrying the mutM gene specifically induces A : T to C : G transversion muta- suppressed the mutator effect of the mutY7 mutant. tions (Yanofsky et al., 1966). As a consequence of this The high rate of G : C to T : A mutation in the mutM unidirectional mutator activity, mutT7 cells have mutY double mutator strain likely reflects the high level increased GC content in the chromosomal DNA of 8-oxoG content in the DNA (Tajiri et al., 1995). (Cox and Yanofsky, 1967). Akiyama et al. (1987) While the mutation caused by the oxidation of DNA cloned the mutT gene and, based on sequence analysis, is unidirectional, incorporation of 8-oxo-dTP into identified a protein with 129 amino acid residues. The DNA would result in base substitutions in two MutT protein was purified to physical homogeneity directions. In E. coli cells, the MutT protein seems to and was shown to have nucleoside triphosphatase prevent both A : T to C : G and G : C to T : A activity (Bhatnagar and Bessman, 1988). Using an in transversions by eliminating 8-oxo-dGTP from the vitro DNA synthesis system, Akiyama et al. (1989) nucleotide pool. For the spontaneous mutagenesis demonstrated that the MutT protein specifically caused by 8-oxo-dGTP, MutM and MutY proteins prevented misincorporation of dGMP onto the also have a role, but their functions in the A : T to poly(dA)/oligo(dT)20 template-primer. Subsequently C : G pathway differ from those in the G : C to T : A Maki and Sekiguchi (1992) found that the nucleotide pathway. The repair enzymes, especially the MutY misincorporated opposite the adenine residue of the protein, promote the fixation of A : T to C : G template is not dGMP but rather its oxidized form, 8- transversion whereas the MutM and MutY proteins oxo-dGMP. Although the guanine base assumes almost cooperate to suppress the G : C to T : A transversion, in exclusively anti conformation in DNA, in the case of 8- the same manner as they do for oxidative DNA, as oxoguanine, there exists almost the same quantity of illustrated in Figure 1c.

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8897

Figure 1 8-Oxoguanine-related mutagenesis. (a) Base substitution mutations, classified into transition and transversion according to orientations of the base changes. Among two types of transition and four types of transversion mutations, mutT deficiency causes A:T?C : G transversion while deficiencies in mutM and mutY induce G : C?T : A transversion. (b) Distribution of types of mutations, as measured in the rpsL gene of E. coli (Tajiri et al., 1995). Blue bars represent G : C?T : A transversion and yellow bars represent A : T?C : G transversion. All six types of base substitutions appearing though the mutation frequency of wild-type cells is 100 times less than those for mutant strains (Mo et al., 1991). (c) 8-Oxoguanine-related mutagenesis pathways leading to induction of specific types of mutations. The guanine residue in DNA as well as dGTP in the nucleotide pool can be oxidized to 8-oxoguanine. As shown in the left side surrounded by a blue line, 8-oxoguanine produced in DNA induces G : C?T : A transversion, which is prevented by the actions of MutM and MutY proteins. Incorporation of 8-oxoguanine-containing nucleotide into DNA could induce A : T?C : G and G : C?T : A transversions, both of which can be prevented by MutT protein, by hydrolyzing 8- oxo-dGTP. In mutT7 mutant cells, occurrence of G : C?T : A transversion is specifically prevented by functions of MutM and MutY proteins, resulting in a preferential occurrence of A : T?C : G transversion, as surrounded by a yellow line

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8898 Mammalian MTH1 protein with 8-oxo-dGTPase Northern blot analysis revealed that MTH1 mRNA activity is present in almost all tissues of adult mouse, among which expression levels are exceedingly high in thymus In search of an activity similar to E. coli MutT in and testis (Oda et al., 1997). Fetal tissues (brain and mammalian cells, Jurkat cells (a human T-cell leukemia liver) contained large amounts of MTH1 mRNA. In cell line) were found to contain a high level of such peripheral blood lymphocytes, the level of MTH1 activity. Taking advantage of this high level of activity, mRNA was significantly increased after concomitant the enzyme was purified to apparent physical homo- treatment with phytohemagglutinin and interleukin-2. geneity (Mo et al., 1992; Sakumi et al., 1993). Substrate Analyses of the 5’ regions of the MTH1 transcripts specificity of the enzyme was examined using a-32P- revealed that seven types of MTH1 mRNA, which may labeled dNTPs. Although dGTP and dATP were also be produced by transcription initiation at different sites hydrolyzed to the corresponding nucleoside monopho- and/or alternative splicing (Figure 2). In most or all of sphates, product yields were about 5% of those with 8- human cells and tissues, type 1 mRNA which lacks the oxo-dGTP. Neither TTP nor dCTP was hydrolyzed by exon 2 sequence is present as a major form. More than the enzyme. The apparent Km for hydrolysis of 8-oxo- one ATG initiation codons in-frame were found in the dGTP was 70 times lower than that for the degradation 5’ regions of some of the MTH1 mRNAs and, of dGTP, whereas the maximal reaction rates observed moreover, there is a polymorphic alteration at the 5’ with both substrates were similar. splicing site located in exon 2. Thus, regulation of Based on the partial amino acid sequence determined expression of the human MTH1 gene is complex and with the purified human 8-oxo-dGTPase protein, a cDNA for human enzyme was cloned and the nucleotide sequence determined (Sakumi et al., 1993). The molecular mass of the protein, as calculated from the predicted amino acid sequence, was 17.9 kDa, a value close to that estimated from analysis of SDS – PAGE. By placing the cDNA under control of the lac promoter, a high level of 8-oxo-dGTPase activity was induced in the presence of isopropyl-1-thio-b-D- galactopyranoside. On expression of cDNA in E. coli mutT7 cells the increased level of spontaneous mutation frequency was considerably reduced. Similar but more striking suppressive effects were observed when mouse or rat cDNA was expressed in the mutT7 cells (Kakuma et al., 1995; Cai et al., 1995). Thus, mammalian 8-oxo-dGTPase functions in E. coli cells to prevent mutations caused by the accumulation of 8- oxo-dGTP in the nucleotide pool. The mammalian gene for 8-oxo-dGTPase has been named MTH1 for mutT homolog 1. Transfection of human MTH1 cDNA brought about a significant reduction in 8-oxoguanine content of DNA in mouse embryonic fibroblasts as well as in tumor cells, with or without H2O2 treatment (Colussi et al., 2002). As MTH1 protein decreases both steady-state and oxidant-induced 8-oxoguanine levels in DNA, endogenous oxidation of the deoxynucleotide pool is a definite source of DNA damage and the deoxynucleotide pool is a significant target for exogenous oxidative damage. To elucidate structure of the gene, human genomic libraries were screened using cDNA as a probe (Furuichi et al., 1994). The gene which spans approximately 8 kb consists of five exons, and the coding sequence resides on exons 2, 3, 4 and 5 (Oda et al., 1997, 1999). As will be described below, some of the exons are composed of differentially processed segments. Using fragments of genomic DNA as probes, Figure 2 Genomic structure and alternative splicing of the hu- chromosal assignment of the gene was made. These man MTH1 gene (Oda et al., 1997). In the upper part, the overall fragments derived from different regions, consistently structure of the gene is shown, in which exons 1 to 5 are indicated by filled boxes. In the lower part, seven types of MTH1 mRNA hybridized to a single locus on the short arm of are shown, which can be produced by transcription initiation at chromosome 7 at p22. Thus, the MTH1 gene is located different sites and alternative splicing. Hatched boxes represent on human chromosome 7p22 (Furuichi et al., 1994). exons

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8899 may be subjected to cell type-specific modification (Oda 1993, Frick et al., 1995) and the conserved region was et al., 1999). found to consist of a loop – helix – loop motif which In eukaryotic cells, a pool of dNTP for nuclear may contribute to nucleotide binding and nucleotide DNA replication is present mainly in the cytosol pyrophosphohydrolase activity. NMR studies were (Bestwick et al., 1982). Mitochondria, which preserve further extended to determine the roles of conserved a pool of dNTP for mitochondrial DNA synthesis, amino acid residues in the possible active site of the consist of more than 10% of the total intracellular MutT protein, in conjunction with site-directed dNTP. The mitochondrial respiratory chain located on mutagenesis analyses (Lin et al., 1996, 1997). Recently, inner membranes is a major site for the initiation of NMR analysis of human MTH1 protein has been lipid peroxidation, which can lead to oxidation of the made (Mishima et al., unpublished result). Structure of guanine to 8-oxoguanine. In addition, the mitochon- MutT protein, both free and Mn-associated forms, was drial respiratory chain produces superoxide, which can resolved by X-ray crystallographic analysis (Yamagata be converted to hydroxyl radicals via hydrogen et al., unpublished result). peroxide. The hydroxyl radical is the main species of The 23-residue sequence is a sole conserved sequence reactive oxygen that attacks the guanine base. Thus, among all MutT and MTH1 homologs with 8-oxo- DNA and dNTP in the mitochondrial pool may be dGTPase, and of the many other proteins with the exposed to a greater oxidative stress than is the case in MutT signature so far identified, some hydrolysing the nucleus. In this context, the intracellular location of various nucleotide derivatives, such as dATP, diade- 8-oxo-dGTPase was determined (Kang et al., 1995). In nosine oligophosphates, NADH, ADP-ribose, and human Jurkat cells, most of the enzyme activity was GDP-mannose (O’Handley et al., 1996, 1998; Sheikh found in cytosolic and the mitochondrial fractions. The et al., 1998). Furthermore, a diphosphoinositol poly- specific activity of 8-oxo-dGTPase in the mitochondrial phosphate phosphohydrolase that hydrolyzes a non- fraction was about 17% of that in the cytosolic related polyphosphate also contains the 23-residue fraction; almost no enzyme activity was found in (Safrany et al., 1998). A chimeric protein MTH1-Ec, nuclear and microsomal fractions. Electron micro- in which the 23-residue sequence of MTH1 was scopic immunocytochemistry, using a specific replaced with that of MutT, retains its capacity to antibody against MTH1 protein, showed that MTH1 hydrolyze 8-oxo-dGTP (Fujii et al., 1999), thereby protein was localized in the mitochondrial matrix. An indicating that the 23-residue sequences of MTH1 and identical molecular form of MTH1 protein appears to MutT are functionally and structurally equivalent and be present in the cytosol and in the mitochondria. constitute functional modules (Figure 3b). Among these 23 residues in the putative active site region, 10 amino acids are conserved through bacteria Structural features of MTH1-related proteins to human. To establish the functional significance of these residues, saturation mutagenesis of the 10 Human MTH1 and E. coli MutT proteins are similar conserved residues of MutT and MTH1 proteins were in size and there is a certain degree of sequence done, in combination with negative and positive homology in these proteins. Genes for analogous mutant screening, using mutagenesis primers that functions were isolated from Proteus vulgaris and contain a completely degenerated codon for each Streptococcus pneumoniae, bacteria distantly related to residue (Figure 3c). Shimokawa et al. (2000) found E. coli (Kamath and Yanofsky, 1993; Bullions et al., that Gly37, Gly38, Glu44, Arg52, Glu53, and Glu57 in 1994). The products of the latter two genes carry MutT protein could not be replaced by any other enzyme activity specifically degrading dGTP to dGMP residue without losing activity. and are structurally and functionally related to the E. Replacement of the corresponding residues of MTH1 coli MutT protein. Most of the identical residues are in protein also caused loss of function (Fujii et al., 1999). a region corresponding to the 23 residues from Gly37 In the latter case two additional residues were found to to Gly59 of E. coli MutT, known as the MutT be essential, though there is the possibility that signature (Bessman et al., 1996). Homologs of human formation of the human protein in heterologous cell MTH1 protein were identified in mice and rats, based (E. coli) might cause a more severe effect on expression on isolated cDNAs (Kakuma et al., 1995; Cai et al., of the function. It seems that these essential amino acid 1995). Both proteins comprise 156 amino acid residues, residues interact with the substrate nucleotide and as was the case for the human MTH1 protein, and Mg++ to exert catalytic functions. amino acid sequences are highly conserved. Alignment of the sequences of these six proteins shows that all carry a highly conserved sequence in nearly the Tumorigenesis and mutagenesis in mice lacking MTH1 same region, corresponding to amino acids 36 to 58 for human MTH1 (Figure 3a). Ten of 23 amino acid residues To obtain insight into the role of MTH1 protein in in this region are identical, hence this probably terms of spontaneous tumorigenesis as well as constitutes an active center for the enzyme. mutagenesis caused by the oxygen-induced DNA The secondary structure of MutT protein has been damage, it is necessary to establish cell and mouse determined by heteronuclear multi-dimensional NMR lines defective in the MTH1 gene. This was achieved (Abeygunawardana et al., 1993, 1995; Weber et al., recently using gene targeting techniques.

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8900

Figure 3 Comparison of structures of MutT family proteins. (a) The 23-residue modules of MutT-related proteins from various organisms, as shown by hatched boxes. Conserved amino acid residues are indicated by bold letters. (b) Secondary structures of E. coli MutT, human MTH1 and a chimeric protein MTH1-Ec. In the chimeric MTH1 protein, the 23-residue module is replaced with that of E. coli MutT protein. (c) Targeted mutagenesis analyses of E. coli MutT and human MTH1 protein. Amino acid residues conserved through the six MutT-related proteins are indicated by bold letters and those that could not be replaced by any other residue without losing function are placed in the yellow boxes. The data were taken from Fujii et al. (1999) and Shimokawa et al. (2000)

The mouse MTH1 gene is composed of five exons MutT-related functions, differs considerably in mouse and spans about 10 kb (Igarashi et al., 1997). Using and E. coli cells. Increases in Hprt mutations detected the isogenic genomic DNA fragment, the entire area of in mouse MTH17/7 cells is twofold compared with the third exon containing the initiation codon and the that in MTH1+/+ cells, whereas the value of mutation adjacent intron regions were replaced with a neo frequency for E. coli mutT7 cells is 100 times over that cassette (Tsuzuki et al., 2001). The resulting construct in wild-type cells (Tajiri et al., 1995). Several was electroporated into ES cells, and cells showing hypotheses to explain this difference may be consid- resistance to both G418 and ganciclovir were selected ered, among which the most plausible is that to obtain MTH1+/7 cells. MTH17/7 cells were mammalian cells may possess an enzyme(s) capable generated from the MTH1+/7 cells by growing in the of degrading 8-oxo-dGTP in addition to MTH1. This presence of 1.5 mg/ml of G418, a concentration which notion was strengthened by the recent finding that even is six times higher than that used for the isolation of E. coli, despite its predominant MutT activity, MTH1+/7 cells. possesses an additional enzyme activity that degrades MTH17/7 cells thus obtained were used to examine 8-oxo-dGTP. GTP cyclohydrase II, encoded by the the effect of MTH1 deficiency on spontaneous ribA gene of E. coli, can hydrolyze 8-oxo-dGTP to the mutagenesis. Mutations in the Hprt gene, located on corresponding nucleoside monophosphate. In the the X chromosome in the mouse genome, render cells mutT7 background, ribA7 cells showed two times resistance to 6-thioguanine and, in this forward muta- higher spontaneous mutation frequencies as compared tion assay, two independently isolated MTH17/7 cell with ribA+ cells (Kobayashi et al., 1998). It is unlikely lines exhibited an approximately twofold higher that a GTP cyclohydrase II-type enzyme is present in mutation rate, compared with the value of MTH1+/+ mammalian cells because the biosynthesis of riboflavin, cells. Thus, MTH1 may have the potential to prevent in which this type of enzyme functions, does not take the occurrence of mutations under normal growth place in animals. Recently, a mouse cDNA clone that conditions. considerably suppresses the high mutability of E. coli It should be noted, however, that degree of increase mutT7 cells was isolated (Cai et al., unpublished in spontaneous mutation frequency, due to the loss of result). The protein encoded by this cDNA carried

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8901 activity to degrade 8-oxo-dGTP, and could act as an sign of development of malignancy during about 12 MTH1 redundancy factor. months after birth. Alternative mechanisms might MTH1 homozygous mutant mice were generated by function to minimize the effects of an increased load microinjecting MTH1+/7 ES cell clones into blasto- of 8-oxoguanine in mammalian genomes. cysts and subsequently crossing the resulting chimera and heterozygous mice. MTH17/7 mice have a normal physical appearance, but have a high susceptibility for Metabolism of oxidized nucleotides spontaneous tumorigenesis (Tsuzuki et al., 2001). Around 18 months after birth, many tumors were Figure 4 summarizes enzymatic reactions for inter- found in lungs and livers of MTH17/7 mice, but there conversion of 8-oxoguanine-containing nucleotides in were few in MTH1+/+ mice. The elevated incidence of mammalian cells. Enzymatic conversion of ribonucleo- tumor formation in the liver of MTH17/7 mice tides to deoxyribonucleotides occurs at the level of correlated well with the highest content of MTH1 nucleoside diphosphate, and ribonucleotide reductase, protein in this organ of the wild-type mouse (Kakuma the enzyme responsible, has a relatively broad substrate et al., 1995). As generally observed in spontaneous and specificity. Four types of naturally occurring ribonu- carcinogen-induced hepatocarcinogenesis in rodents cleotides, ADP, GDP, CDP and UDP, are converted (Poole and Drinkwater, 1996), there is a high to the corresponding deoxyribonucleotides by a single susceptibility of male mice to liver tumorigenic events, species of reductase enzyme (Thelander and Reichard, compared with their female counterparts. It has been 1979). However, this enzyme is inactive on the 8- suggested that the hormonal environment of the host oxoguanine-containing nucleotide, as revealed with affects the development of many types of tumors, mouse ribonucleotide reductase (Hayakawa et al., especially those in the liver. More tumors tend to form 1999). This implies that 8-oxoguanine-containing in the stomach of MTH17/7 mice, as compared with deoxyribonucleotides must be generated in the cellular MTH1+/+ mice, although the statistical significance is deoxyribonucleotide pool. weak. Indeed, as there are different profiles of Human cells contain nucleoside diphosphate kinase, antioxidant enzyme, such as superoxide dismutases an enzyme activity which phosphorylates various and catalases, different susceptibility to tumorigenesis nucleoside diphosphates to the corresponding nucleo- among organs may reflect the metabolic balance of side triphosphates (Kornberg and Baker, 1992). This oxidative stress, inducing lesions in DNA as well as in enzyme can convert 8-oxo-dGDP to 8-oxo-dGTP, the nucleotide pool. Altogether, it can be concluded although the rate of phosphorylation of 8-oxo-dGDP that the intracellular level of MTH1 protein is an was one-third that of dGDP (Hayakawa et al., 1995). important factor in determining susceptibility of mice Thus, in addition to direct oxidation of dGTP, 8-oxo- to tumor induction by endogenous oxidative damage. dGTP may be generated by the phosphorylation of 8- To examine in vivo mutation events due to the oxo-dGDP. MTH1-deficiency, a reporter gene, rpsL of E. coli, was Once 8-oxo-dGTP is produced, this can be incorpo- introduced into MTH17/7 mice (Egashira et al., 2002). rated into DNA. Various DNA polymerases from Interestingly, the net frequency of rpsL7 forward eukaryotes and prokaryotes have the potential to mutants showed no apparent increase in MTH17/7 utilize 8-oxo-dGTP as substrate (Cheng et al., 1992; mice as compared to findings in MTH1+/+ mice. Maki and Sekiguchi, 1992; Pavlov et al., 1994; Minnick However, there are differences between these two et al., 1994). Thus, the action of 8-oxo-dGTPase is a genotypes in class- and site-distributions of the rpsL7 pre-requisite for high fidelity of DNA replication. mutations recovered from the mice. Unlike MutT- 8-Oxo-dGMP, produced by the action of 8-oxo- deficient E. coli, an increase in frequency of dGTPase, cannot be re-phosphorylated by cellular A:T?C : G transversion was not evident in MTH1 enzymes. Human guanylate kinase, which phosphor- nullizygous mice. Nevertheless, the frequency of single- ylates both GMP and dGMP to the corresponding base frameshifts at mononucleotide runs was 5.7-fold nucleoside diphosphates, is totally inactive for 8-oxo- higher in spleens of MTH17/7 mice than in those of dGMP (Hayakawa et al., 1995). This would provide wild-type mice. Since the elevated incidence of single- another basis for excluding this mutagenic substrate base frameshifts at mononucleotide runs is a hallmark from the DNA precursor. of the defect in MSH2-dependent mismatch repair 8-Oxo-dGMP is dephosphorylated to yield the system, this weak site-specific mutator effect of corresponding nucleoside, 8-oxodeoxyguanosine. MTH17/7 mice could be attributed to a partial Nucleosides are readily transported through the cell sequestration of the mismatch repair function that membrane, and extracellular nucleosides can be may act to correct mispairs with the oxidized excreted into the urine. Dephosphorylation of 8-oxo- nucleotides. dGMP, therefore, may be an essential step for Studies of OGG1-defective mice would provide much excretion of 8-oxoguanine-containing compounds. The more insight into the matter. OGG17/7 mice accumu- enzyme that catalyzes this reaction, 8-oxo-dGMPase, late abnormally high levels of 8-oxoguanine in their was partially purified from an extract of human Jurkat genomes and exhibit a moderately elevated sponta- cells, and the mode of action was elucidated neous mutation rates in their tissues (Klungland et al., (Hayakawa et al., 1995). 8-Oxo-dGMP is the preferred 1999; Minowa et al., 2000). Nevertheless, there was no substrate of the enzyme, and other nucleoside mono-

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8902

Figure 4 Interconversion of 8-oxoguanine-containing nucleotides in mammalian cells. This scheme is based on the results of Haya- kawa et al. (1995, 1999). (1) Ribonucleotide reductase; (2) nucleoside diphosphate kinase; (3) 8-oxo-dGTPase (MTH1); (4) guanylate kinase; (5) DNA polymerase; (6) 8-oxo-dGMPase. O. denotes oxidative reaction

phosphates are cleaved albeit at significantly lower whereas mammalian RNA polymerase II incorporates rates. little of 8-oxoguanine into RNA (Hayakawa et al., As discussed above, MTH1 and MutT are key 1999). enzymes for sustaining high fidelity of DNA replica- The second point is that the mammalian enzyme has tion, in oxygenated stated of living systems. These a broader substrate specificity as compared with the enzymes of diverged origins share a common character bacterial enzyme. MTH1 hydrolyzes 8-oxo-dATP, 2- to degrade 8-oxo-dGTP to 8-oxo-dGMP, thereby hydroxy-dATP and 2-hydroxyATP as well as 8-oxo- eliminating the mutagenic substrate from the DNA dGTP (Sakai et al., 2002). This is in contrast with precursor pool. In spite of this common feature, they MutT, which acts on 8-oxoguanine-containing nucleo- exhibit certain significant differences. First, MTH1 and tides alone. Recognition of 8-oxo-dGTP and 2- MutT differ in the capacity to degrade 8-oxoguanine- hydroxy-dATP appears to depend on different residues containing ribonucleotides. E. coli MutT protein of MTH1 protein. Transgenic mice expressing engi- cleaves 8-oxoGTP as efficiently as does 8-oxo-dGTP neered MTH1 proteins lacking each one of the (Taddei et al., 1997). On the other hand, human activities would be useful to reveal the biological MTH1 is hardly capable of hydrolyzing 8-oxoGTP; its significance of these two activities. potential to degrade 8-oxoGTP is only 2% that for 8- oxo-dGTP (Hayakawa et al., 1999). These apparent differences in efficiency of degrading an error-evoking Acknowledgements substrate may be counterbalanced by the capacity of We thank Drs H Hayakawa, H Maki, Y Nakabeppu, Y RNA polymerases to discriminate the unfavorable Yamagata, Y Nakatsu and H Shimokawa for discussion substrate. E. coli RNA polymerase can utilize 8- and helpful advice, and M Ohara for useful comments on oxoGTP as substrate at a rate 10% that of GTP the manuscript.

References

Abeygunawardana C, Weber D, Frick DN, Bessman MJ and Akiyama M, Horiuchi T and Sekiguchi M. (1987). Mol. Gen. Mildvan AS. (1993). Biochemistry, 32, 13071 – 13080. Genet., 206, 9 – 16. Abeygunawardana C, Weber D, Gittis AG, Frick DN, Lin J, Akiyama M, Maki H, Sekiguchi M and Horiuchi T. (1989). Miller A-F, Bessman MJ and Mildvan AS. (1995). Proc. Natl. Acad. Sci. USA, 86, 3949 – 3952. Biochemistry, 34, 14997 – 15005. Ames BN and Gold LS. (1991). Mutat. Res., 250, 3 – 16.

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8903 Au KG, Cabrera M, Miller JH and Modrich P. (1988). Proc. Kang D, Nishida J, Iyama A, Nakabeppu Y, Furuichi M, Natl. Acad. Sci. USA, 85, 9163 – 9166. Fujiwara T, Sekiguchi M and Takeshige K. (1995). J. Biol. Au KG, Clark S, Miller JH and Modrich P. (1989). Proc. Chem., 270, 14659 – 14665. Natl. Acad. Sci. USA, 86, 8877 – 8881. Kamath AV and Yanofsky C. (1993). Gene, 134, 99 – 102. Bessho T, Tano K, Kasai H and Nishimura S. (1992). Kasai H and Nishimura S. (1984). Nucleic Acids Res., 12, Biochem. Biophys. Res. Commun., 188, 372 – 378. 2137 – 2145. Bessman MJ, Frick DN and O’Handley SF. (1996). J. Biol. Klungland A, Rosewell I, Hollenbach S, Larsen E, Daly G, Chem., 271, 25059 – 25062. Epe B, Seeberg E, Lindahl T and Barns DE. (1999). Proc. Bestwick RK, Moffett GM and Mathews CK. (1982). J. Biol. Natl. Acad. Sci. USA, 96, 13300 – 13305. Chem., 257, 9300 – 9304. Kobayashi M, Ohara-Nemoto Y, Kaneko M, Hayakawa H, Bhatnagar SK and Bessman MJ. (1988). J. Biol. Chem., 263, Sekiguchi M and Yamamoto K. (1998). J. Biol. Chem., 8953 – 8957. 273, 26394 – 26399. Boiteux S and Radicella JP. (1999). Biochimie, 81, 59 – 67. Kornberg A and Baker TA. (1992). DNA Replication 2nd Brooks PJ, Wise DS, Berry DA, Kosmoski JV, Smerdon MJ, edn. New York: W. H. Freeman and Company,. Somers RL, Mackie H, Spoonde AY, Ackerman EJ, KuraokaI,BenderC,RomieuA,CadetJ,WoodRDand Coleman K, Tarone RE and Robbins JH. (2000). J. Biol. Lindahl T. (2000). Proc. Natl. Acad. Sci. USA, 97, 3832 – Chem., 275, 22355 – 22362. 3837. Bullions LC, Mejean V, Claverys JP and Bessman MJ. Lin J, Abeygunawardana C, Frick DN, Bessman MJ and (1994). J. Biol. Chem., 269, 12339 – 12344. Mildvan AS. (1996). Biochemistry, 35, 6715 – 6726. Cabrera M, Nghiem Y and Miller JH. (1988). J. Bacteriol., Lin J, Abeygunawardana C, Frick DN, Bessman MJ and 170, 5405 – 5407. Mildvan AS. (1997). Biochemistry, 36, 1199 – 1211. Cai JP, Kakuma T, Tsuzuki T and Sekiguchi M. (1995). Maki H and Sekiguchi M. (1992). Nature, 355, 273 – 275. Carcinogenesis, 16, 2343 – 2350. Michaels ML, Pham L, Cruz C and Miller JH. (1991). Cheng KC, Cahill DS, Kasai H, Nishimura S and Loeb LA. Nucleic Acids Res., 19, 3629 – 3632. (1992). J. Biol. Chem., 267, 166 – 172. Michaels ML, Cruz C, Grollman AP and Miller JH. (1992). Chung MH, Kasai H, Jones DS, Inoue H, Ishikawa H, Proc. Natl. Acad. Sci. USA, 89, 7022 – 7025. Ohtsuka E and Nishimura S. (1991). Mutat. Res., 254, 1– Minowa O, Arai T, Hirano M, Monden Y, Nakai S, Fukuda 12. M, Itoh M, Takano H, Hippou Y, Aburatani H, Colussi C, Parlanti E, Degan P, Aquilina G, Barnes D, Masumura K, Nohmi T, Nishimura S and Noda T. Macpherson P, Karran P, Crescenzi M, Dogliotti E and (2000). Proc. Natl. Acad. Sci. USA, 97, 4156 – 4161. Bignami M. (2002). Current Biol., 12, 912 – 918. Minnick DT, Pavlov YI and Kunkel TA. (1994). Nucleic Cox EC and Yanofsky C. (1967). Proc. Natl. Acad. Sci. USA, Acids Res., 22, 5658 – 5664. 58, 1895 – 1902. Mo JY, Maki H and Sekiguchi M. (1991). J. Mol. Biol., 222, Demple B and Harrison L. (1994). Annu. Rev. Biochem., 63, 925 – 936. 915 – 948. Mo JY, Maki H and Sekiguchi M. (1992). Proc. Natl. Acad. Egashira A, Yamauchi K, Yoshiyama K, Kawate H, Katsuki Sci. USA, 89, 11021 – 11025. M, Sekiguchi M, Sugimachi K, Maki H and Tsuzuki T. Nash HM, Bruner SD, Scha¨ rer OD, Kawate T, Addona TA, (2002). DNA Repair, 75, 1–13. Spooner E, Lane WS and Verdine GL. (1996). Curr. Biol., Evans J, Maccabee M, Hatahet Z, Courcelle J, Bockrath R, 6, 968 – 980. Ide H and Wallace S. (1993). Mutat. Res., 299, 147 – 156. Nghiem Y, Cabrera M, Cupples CG and Miller JH. (1988). Fraga CG, Shigenaga MK, Park JW, Degan P and Ames BN. Proc. Natl. Acad. Sci. USA, 85, 2709 – 2713. (1990). Proc. Natl. Acad. Sci. USA, 87, 4533 – 4537. Oda H, Nakabeppu Y, Furuichi M and Sekiguchi M. (1997). Frick DN, Weber DJ, Abeygunawardana C, Gittis AG, J. Biol. Chem., 272, 17843 – 17850. Bessman MJ and Mildvan AS. (1995). Biochemistry, 34, OdaH,TaketomiA,MaruyamaI,ItoR,NishiokaK, 5577 – 5586. Yakushiji H, Suzuki T, Sekiguchi M and Nakabeppu Y. Fujii Y, Shimokawa H, Sekiguchi M and Nakabeppu Y. (1999). Nucleic Acids Res., 27, 4335 – 4343. (1999). J. Biol. Chem., 274, 38251 – 38259. O’Handley SF, Frick DN, Bullions LC, Mildvan AS and Furuichi M, Yoshida MC, Oda H, Tajiri T, Nakabeppu Y, Bessman MJ. (1996). J. Biol. Chem., 271, 24649 – 24654. Tsuzuki T and Sekiguchi M. (1994). Genomics, 24, 485 – O’Handley SF, Frick DN, Dunn CA and Bessman MJ. 490. (1998). J. Biol. Chem., 273, 3192 – 3197. GajewskiE,RaoG,NackerdienZandDizdarogluM. Ohtsubo T, Matsuda O, Iba K, Terashima I, Sekiguchi M (1990). Biochemistry, 29, 7876 – 7882. and Nakabeppu Y. (1998). Mol. Gen. Genet., 259, 577 – Hayakawa H, Hofer A, Thelander L, Kitajima S, Cai Y, 590. Oshiro S, Yakushiji H, Nakabeppu Y, Kuwano M and Poole TM and Drinkwater NR. (1996). Carcinogenesis, 17, Sekiguchi M. (1999). Biochemistry, 38, 3610 – 3614. 191 – 196. Hayakawa H, Taketomi A, Sakumi K, Kuwano M and Pavlov YI, Minnick DT, Izuta S and Kunkel TA. (1994). Sekiguchi M. (1995). Biochemistry, 34, 89 – 95. Biochemistry, 33, 4695 – 4701. Henle ES and Linn S. (1997). J. Biol. Chem., 272, 19095 – Safrany ST, Caffrey JJ, Yang X, Bembenek ME, Moyer MB, 19098. Burkhart WA and Shears SB. (1998). EMBO J., 17, 6599 – Igarashi H, Tsuzuki T, Kakuma T, Tominaga Y and 6607. Sekiguchi M. (1997). J. Biol. Chem., 272, 3766 – 3772. Sakai Y, Furuichi M, Takahashi M, Mishima M, Iwai S, Ionov Y, Peinado MA, Malkhosyan S, Shibata D and Shirakawa M and Nakabeppu Y. (2002). J. Biol. Chem., Perucho M. (1993). Nature, 363, 558 – 561. 277, 8579 – 8587. Jackson AL and Loeb LA. (1998). Genetics, 148, 1483 – 1490. Sakumi K, Furuichi M, Tsuzuki T, Kakuma T, Kawabata S, Kakuma T, Nishida J, Tsuzuki T and Sekiguchi M. (1995). J. Maki H and Sekiguchi M. (1993). J. Biol. Chem., 268, Biol. Chem., 270, 25942 – 25948. 23524 – 23530.

Oncogene Oxidative nucleotide damage M Sekiguchi and T Tsuzuki 8904 Schaaper RM, Danforth BN and Glickman BW. (1986). J. Tchou J and Grollman AP. (1993). Mutat. Res., 299, 277 – Mol. Biol., 189, 273 – 284. 287. Sheikh S, O’Handley SF, Dunn CA and Bessman MJ. (1998). Thelander L and Reichard P. (1979). Annu.Rev.Biochem., J. Biol. Chem., 273, 20924 – 20928. 48, 133 – 158. Shibutani S, Takeshita M and Grollman AP. (1991). Nature, Treffers HP, Spinelli V and Belser NO. (1954). Proc. Natl. 349, 431 – 434. Acad. Sci. USA, 40, 1064 – 1071. Shimokawa H, Fujii Y, Furuichi M, Sekiguchi M and Tsuzuki T, Egashira A, Igarashi H, Iwakuma T, Nakatsuru Nakabeppu Y. (2000). Nucleic Acids Res., 28, 3240 – Y, Tominaga Y, Kawate H, Nakao K, Nakamura K, Ide 3249. F, Kura S, Nakabeppu Y, Katsuki M, Ishikawa T and Slupska MM, Baikalov C, Luther WM, Chiang JH, Wei YF Sekiguchi M. (2001). Proc. Natl. Acad. Sci. USA, 98, and Miller JH. (1996). J. Bacteriol., 178, 3885 – 3892. 11456 – 11461. Smith KC. (1992). Mutat. Res., 277, 139 – 162. van der Kemp PA, Thomas D, Barbey R, de Oliveira R and Stoler DL, Chen N, Basik M, Kahlenberg MS, Rodriguez- Boiteux S. (1996). Proc.Natl.Acad.Sci.USA,93, 5197 – Bigas MA, Petrelli NJ and Anderson GR. (1999). Proc. 5202. Natl. Acad. Sci. USA, 96, 15121 – 15126. Weber D, Abeygunawardana C, Bessman MJ and Mildvan Taddei F, Hayakawa H, Bouton M, Cirinesi A, Matic I, AS. (1993). Biochemistry, 32, 13081 – 13088. Sekiguchi M and Radman M. (1997). Science, 278, 128 – Yanofsky C, Cox EC and Horn V. (1966). Proc. Natl. Acad. 130. Sci. USA, 55, 274 – 281. Tajiri T, Maki H and Sekiguchi M. (1995). Mutat. Res., 336, Yeh YC, Chang DY, Masin J and Lu AL. (1991). J. Biol. 257 – 267. Chem., 266, 6480 – 6484.

Oncogene