The formation of methylated bases in DNA by dimethylnitrosamine and its relation to differences in the formation of tumours in the livers of GR and C3HF mice

Metabolism of and DNA methylation by dimethylnitrosamine (DMNA) were measured in the livers of GR male and C3Hf male and female mice which showed widely different susceptibilities to tumour formation by this hepatocarcinogen.It was previously shown that continuous DMNA administration results in vascular tumours in the livers of C3Hf female mice, whereas C3Hf males develop a high incidence of hepatomas both after continuous treatment and after a single injection of DMNA to adult animals. GR males showed a low susceptibility to the formation of liver tumours under these conditions.N-demethylation of DMNA by liver microsomes showed similar activity for both C3Hf sexes; but GR males were significantly more active.At 5 and 48 h after a single injection of [¹⁴C]DMNA, the amounts of O⁶-methylguanine (O⁶-MeGua), 7-methylguanine (7-MeGua), 1-methyladenine (1-MeAde) and 3-methyladenine (3-MeAde) were similar for C3Hf males and females, with the possible exception of 7-MeGua which seemed to be slightly higher in the female. O⁶ MeGua disappeared from C3Hf liver DNA with an apparent half-life time of about 24 h. Especially at 48 h after injection, GR liver DNA was methylated to a higher extent than was C3Hf liver DNA. This result, which antiparallels the tumour incidences, may be explained by the differences in rate of N-demethylation of DMNA. where higher 7-MeGua values were found for fasted animals under otherwise identical conditions.The general conclusior to be drawn is that neither the metabolism of DMNA nor DNA methylation by this carcinogen in the livers of male GR and C3Hf male and female mice correlates With the formation of hepatomas after DMNA administration. A possible explanation of the absence of such a correlation between DNA methylation and tumour formation might be that there exists no causal relationship between both events. However, a complicating factor is that the eventual development of a tumour may be influenced by a number of—sometimes decisive—secondary factors like hormonal²⁵ or immunological²⁶ status or the presence of cellular proliferation in target organs^27,28. Evidence from other systems suggests a relationship between inactivating, mutagenic or carcinogenic effects of alkylating agents and their ability to interact with nucleic acids, especially DNA^29,30.     

Intermolecular transacylation of phosphatidylethanolamine by a Butyrivibrio sp.

The alkylation of purine bases in DNA of several rat tissues was determined during weekly injections (10 mg/kg) of N-[3H]methyl-N-nitrosourea, a dose schedule known to selectively induce tumours of the nervous system. Each group of animals was killed 1 week after the final injection, and the DNA hydrolysates were analysed by chromatography on Sephadex G-10. After five weekly applications, O6-methylguanine had accumulated in brain DNA to an extent which greatly exceeded that in kidney, spleen and intestine. In the liver, the final O6-methylguanine concentration was less than 1% of that in brain. Between the first and the fifth injection, the O6-methylguanine/7-methylguanine ratio in cerebral DNA increased from 0.28 to 0.68. In addition, 3-methylguanine was found to accumulate in brain DNA whereas in the other organs no significant quantities of this base were detectable. The results are compatible with the hypothesis that O6-alkylation of guanine in DNA plays a major role in the induction of tumours by N-methyl-N-nitrosourea and related carcinogens. The kinetics of the increase of O6-methylguanine in cerebral DNA suggest that there is no major cell fraction in the brain which is capable of excising chemically methylated bases from DNA. This repair deficiency could be a determining factor in the selective induction of nervous-system tumours by N-methyl-N-nitrosourea and other neuro-oncogenic compounds.     

O6-methylguanine-DNA methyltransferase and alkylation of liver DNA in mice exposed to dimethylnitrosamine during dietary deficiency of essential amino acids

Male NMRI mice were fed a diet containing a complete mixture of amino acids or a mixture deficient in methionine-cysteine or lysine (30% of the control level) for a period of 6 days. During the feeding period all mice received dimethylnitrosamine in the drinking water ad libitum. The exposure averaged 1 mg dimethylnitrosamine/kg body weight and day. The concentration of O6-methylguanine-DNA methyltransferase was measured in liver extracts. It decreased significantly in the methionine-cysteine deficient mice. When DNA from the liver was analyzed for alkylated purine bases the mice received a single dose of 14C-labeled dimethylnitrosamine (0.5 or 1 mg/kg body weight) at 120 min before sacrifice. The concentration of O6-methylguanine increased significantly over the control level upon feeding the deficient diets and was restored to the concentration of the controls by refeeding lysine for 2 days following 6 days of lysine deficiency. The increased ratio of O6-methylguanine to N-7-methylguanine indicated that methylation of guanine in the N-7 position was not subject to variation by the intake of dimethylnitrosamine during the dietary deficiencies. The results demonstrate the requirement for a balanced composition of amino acids in the diet to maintain a sufficient concentration of O6-methylguanine-DNA methyltransferase in the cells and thus to permit efficient removal of the methyl group from the O-6 position of guanine in DNA after exposure to dimethylnitrosamine.     

Molecules involved in cell–cell adhesion during development

A peculiarity of nitrosamines is the high degree of cell and organ specificity in inducing tumors. There is substantial evidence that the initiation of the carcinogenesis process by carcinogens of this group is linked to the metabolic competence of the target tissue or cell to convert these carcinogens into mutagenic metabolites and to the binding of those metabolites to cellular DNA. Alkylation occurs in the DNA at the N-1, N-3, and N-7 positions of adenine; the N-3, N-7, and O⁶ of guanine; the N-3, and O² of cytosine; and the N-3, O⁴, and O² of thymine; and the phosphate groups. The initial proportion of each DNA adduct depends upon the alkylating agent used. The various DNA adducts are lost to a variable extent from DNA in vivo by spontaneous release of bases and Or by specific DNA repair processes. Studies conducted in vitro and in vivo indicate that alkylation at the oxygen atoms of DNA bases is more critical than alkylation at other positions in the mutagenesis and carcinogenesis induced by N-nitroso compounds. In particular, tissues in which tumors occur more frequently after a pulse dose of nitrosamine are those in which O⁶-alkylguanine persists longest in DNA, presumably resulting in an increased probability that a miscoding event (mutation) will take place during DNA synthesis. The more rapid removal of O⁶-methylguanine from the DNA of liver (as compared with cxtrahepatic tissues) of rats has been associated with the absence of tumor production in this organ by a single dose of dimethylnitrosamine; however, a significant incidence of liver tumors is observed if the same dose is given 24 hr after partial hepatectomy, and tumors arc induced by such a dose of dimethyl-nitrosamine in the liver of hamsters, which has a low capacity to remove O⁶-methylguanine from its DNA. These data also indicate that the rate of disappearance of 7-methylguanine from the liver or cxtrahepatic tissues is independent of the dose of dimethylnitrosamine; whereas O⁶-methylguanine is lost from DNA more rapidly after a low dose of this nitrosamine. It has been shown that in liver the removal of O⁶-methylguanine, but not of other DNA adducts, from DNA can be affected by pretreating the animals with N-nitroso compounds. The modulation of DNA repair processes observed after a single dose and after chronic treatment with nitrosamines is discussed in relation to the tissue-specific carcinogenic effect of this group of carcinogens.     

Chronic or acute administration of various dialkylnitrosamines enhances the removal of O6-methylguanine from rat liver DNA in vivo

The effect of pretreatment of rats with various symmetrical dialkylnitrosamines on the repair of O⁶-methylguanine produced in liver DNA by a low dose of [¹⁴C]dimethylnitrosamine (DMN) has been examined. DMN, diethylnitrosamine (DEN), dipropylnitrosamine (DPN) or dibutylnitrosamine (DBN) were administered to rats for 14 consecutive weekdays at a daily dose of 5% of the LD₅₀. Animals were given [¹⁴C]DMN 24 h after the last dose and were killed 6 h later. DNA was extracted from the liver and analysed for methylpurine content after mild acid hydrolysis and Sephadex G-10 chromatography. While the amounts of 3-methyladenine and 7-methylguanine were only slightly different from controls, the amounts of O⁶-methylguanine in the DNA of the dialkylnitrosamine pretreated rats were about 30% of those in control rats, indicating a considerable increase in the capacity to repair this base. Liver ribosomal RNA from control and dialkylnitrosamine pretreated rats contained closely similar amounts of O⁶-methylguanine suggesting that the induced enzyme system does not act on this base in ribosomal RNA in vivo. Pretreatment with these dialkylnitrosamines also enhanced the repair of O⁶-methylguanine in liver DNA when they were given as a single dose (50% of the LD₅₀) either 3 or 7 days before the [¹⁴C]DMN. In addition, single low doses of DMN or DEN (5% of the LD₅₀) given either 1 or 6 days before [¹⁴C]DMN increased O⁶-methylguanine repair and the magnitude of the effect after DEN was similar to that produced by the other pretreatment schedules. The possible mechanism(s) of the induction of O⁶-methylguanine repair and its relation to hepatotoxicity, DNA alkylation, carcinogenesis and the adaptive response in Escherichia coli are discussed.     

Methylated purines in the deoxyribonucleic acid of various Syrian-golden-hamster tissues after administration of a hepatocarcinogenic dose of dimethylnitrosamine.

1. DNA was extracted from livers, kidneys and lungs of Syrian golden hamsters at various times (up to 96h) after injection of a hepatocarcinogenic dose of [14C]dimethylnitrosamine. Purine bases were released from the DNA by mild acid hydrolysis and separated by Sephadex G-10 chromatography. 2. At 7h after dimethylnitrosamine administration liver DNA was alkylated to the greatest extent, followed by that of lung and kidney, the values for which were 8 and 3% respectively of those for liver. 3. The O6-methylguanine/7-methylguanine ratios were initially the same in all three organs and in the liver DNA of rats under similar conditions of dose. 4. O6-Methylguanine was the most persistent alkylated purine in all three hamster tissues. There was evidence for excision of 7-methyl-guanine, the highest activity for this being present in the liver. 5. Detectable amounts of the minor products 3-methyladenine, 1-methyladenine, 3-methylguanine and 7-methyladenine were present in most hamster tissues, and their individual rates of loss from liver DNA were determined. 6. Ring-labelling of the normal purines in DNA was highest in the liver, followed closely by the lung (80% of that in liver) whereas the kidney had very low incorporation (3% of that in liver). 7. The results are discussed with respect to the hepatotoxicity of dimethylnitrosamine, the miscoding potential of the various alkylation products and the induction of liver tumours in hamsters.     

Regulation of the capacity for O6-methylguanine removal from DNA in human lymphoblastoid cells studied by cell hybridization.

Hybrids were made between a ouabain-resistant, thioguanine-resistant human lymphoma line able to remove O6-methylguanine from its DNA (Mex+) and human lymphoblastoid lines deficient in this capability (Mex-). The formation of hybrids was confirmed by chromosomal analysis. Hybrid cells had an O6-methylguanine removal capacity per mole of guanine about one third to one half that of the Mex+ parents, i.e., about the same per cell. Cell hybrids removed the same amount of the alkylation adduct 3-methyladenine as did their parents per mole of guanine, i.e., about twice as much per cell. Although the cell hybrids had intermediate resistance to the cytotoxic action of N-methyl-N'-nitro-N-nitrosoguanidine used to induce O6-methylguanine and 3-methyladenine, there is evidence that the ability to remove O6-methylguanine and resistance to the cytotoxic effect of N-methyl-N'-nitro-N-nitrosoguanidine are dissociable characteristics.     

Evaluation of the DNA-repair host-mediated assay. I. Induction of repairable DNA damage in E. coli cells recovered from liver, spleen, lungs, kidneys, and the blood stream of mice treated with methylating carcinogens

The DNA-repair host-mediated assay was further calibrated by determining the genotoxic activities of 4 methylating carcinogens, namely, dimethylnitrosamine (DMNA), 1,2-dimethylhydrazine (SDMH), methyl nitrosourea (MNU) and methyl methanesulphonate (MMS) in various organs of treated mice. The ranking of the animal-mediated genotoxic activities of the compounds was compared with that obtained in DNA repair assays performed in vitro. The differential survival of strain E. coli K-12/343/113 and of its DNA-repair-deficient derivatives recA, polA and uvrB/recA, served as a measure of genotoxic potency. In the in vitro assays and at equimolar exposure concentrations, MMS and MNU are the most active chemicals, followed by DMNA, which shows a slight genotoxic effect only in the presence of mouse liver homogenate; SDMH has no activity under these conditions. In the host-mediated assays, the order of genotoxic potency of the compounds was quite different: those carcinogens which require mammalian metabolic activation, namely, DMNA and SDMH, show strong effects in liver and blood, a lesser effect in the lungs and kidneys and the least effect in the spleen. The activity of MNU, a directly acting compound, is similar in all organs investigated, but it is clearly lower than that of DMNA and SDMH. MMS, also a directly acting carcinogen, causes some (barely significant) effect at the highest dose tested. A similar order of potency was observed when the compounds were tested in intrasanguineous host-mediated assays with gene mutation as an endpoint. DMNA and SDMH induce comparable frequencies of L-valine-resistant mutants in E. coli K-12/343/113 recovered from liver and spleen of treated mice, the effect in the liver being the strongest. MNU is mutagenic only at a higher dose, while MMS shows no effect. The results are discussed with respect to the literature data on organ-specific DNA adduct formation induced by the compounds. It is concluded that qualitatively there is a good correlation between the degree of genotoxic activity found in the DNA repair host-mediated assay and DNA adduct formation in the animal's own cells. This is exemplified by the finding that the relative order of genotoxic activity of the 4 methylating agents in bacteria recovered from various organs (DMNA approximately equal to SDMH greater than MNU greater than MMS) is reflected by the same order of magnitude in DNA alkylation in corresponding mammalian organs. Quantitatively, the indirectly acting agents DMNA and SDMH seem to induce fewer genotoxic effects in bacteria present in the liver than would be expected on the basis of DNA-adduct formation data.     

Methylated purines formed in DNA by dimethylnitrosamine in rats previously exposed to hepatotoxic and hepatocarcinogenic regimes: effects on the repair of O6-methylguanine

Studies of mammalian systems for the repair of O6-methylguanine in DNA have revealed large differences in the capacities of tissues and cells to perform this function and in the case of rat liver it has been shown that the O6-methylguanine repair system can be stimulated by exposure to hepatotoxic and hepatocarcinogenic regimes. In this report an assessment is made of possible relationships between toxic liver injury, DNA synthesis, cell proliferation and DNA repair by treating Wistar rats with agents selected to provide differing degrees of liver involvement. The effects of long-term (20 week) treatments with acetylaminofluorene (15 mg/kg/day), quinoxaline 1,4-dioxide (10 mg/kg/day), 4-aminobiphenyl-HCl (15 mg/kg/day) and pronethalol (20 mg/kg/day) were assessed, using the same strain of animals in which the original toxicity and carcinogenicity data were obtained. Repair of O6-methylguanine produced in liver DNA by a low, non-toxic dose (2 mg/kg) of [14C]dimethylnitrosamine was increased 3-4-fold throughout the period of treatment with acetylaminofluorene, to a lesser extent by quinoxaline 1,4-dioxide and 4-aminophenyl-HCl and not at all in the case of pronethalol. No evidence was obtained to indicate a direct relationship between O6-methylguanine repair and either the induced hepatotoxicity or the ensuing increased rates of DNA synthesis which occur following exposure to these agents.     

Alternative pathways for the in vivo repair of O6-alkylguanine and O4-alkylthymine in Escherichia coli: the adaptive response and nucleotide excision repair.

The in vivo removal of three different O-alkylated bases from DNA was measured in Escherichia coli. Using monoclonal antibodies specific for O6-methylguanine, O6-ethylguanine and O4-ethylthymine we have monitored the removal of these lesions from six different strains to assess the relative contributions of the adaptive response and of nucleotide excision repair. During the first hour after DNA alkylation, O6-methylguanine, O6-ethylguanine and O4-ethylthymine lesions were repaired almost exclusively by nucleotide excision, except when the adaptive response was being constitutively expressed. In wild-type E. coli the adaptive response began to contribute to O6-methylguanine repair about one hour after alkylation, the time required for the full induction of the ada DNA methyltransferase. In contrast, the adaptive response did not play such a large role in the repair of O6-ethylguanine and O4-ethylthymine in wild-type E. coli, presumably because DNA ethylation damage is a poor inducer of the adaptive response; possible reasons for this poor induction are discussed. The repair of all three O-alkylated lesions was virtually absent in ada- uvr- bacteria suggesting that no alternative pathway is available for their repair, at least during the first two hours after alkylation. When the repair of O-alkylated bases was compromised by an ada- or by a uvr- mutation, the bacteria became more sensitive to alkylation induced killing and mutation.     

Nitrosamine-induced carcinogenesis. The alkylation of nucleic acids of the rat by N-methyl-N-nitrosourea, dimethylnitrosamine, dimethyl sulphate and methyl methanesulphonate

1. N[(14)C]-Methyl-N-nitrosourea, [(14)C]dimethylnitrosamine, [(14)C]dimethyl sulphate and [(14)C]methyl methanesulphonate were injected into rats, and nucleic acids were isolated from several organs after various time-intervals. Radioactivity was detected in DNA and RNA, partly in major base components and partly as the methylated base, 7-methylguanine. 2. No 7-methylguanine was detected in liver DNA from normal untreated rats. 3. The specific radioactivity of 7-methylguanine isolated from DNA prepared from rats treated with [(14)C]dimethylnitrosamine was virtually the same as that of the dimethylnitrosamine injected. 4. The degree of methylation of RNA and DNA produced in various organs by each compound was determined, and expressed as a percentage of guanine residues converted into 7-methylguanine. With dimethylnitrosamine both nucleic acids were considerably more highly methylated in the liver (RNA, about 1% of guanine residues methylated; DNA, about 0.6% of guanine residues methylated) than in the other organs. Kidney nucleic acids were methylated to about one-tenth of the extent of those in the liver, lung showed slightly lower values and the other organs only very low values. N-Methyl-N-nitrosourea methylated nucleic acids to about the same extent in all the organs studied, the amount being about the same as that in the kidney after treatment with dimethylnitrosamine. In each case the RNA was more highly methylated than the DNA. Methyl methanesulphonate methylated the nucleic acids in several organs to about the same extent as N-methyl-N-nitrosourea, but the DNA was more highly methylated than the RNA. Dimethyl sulphate, even in toxic doses, gave considerably less methylation than N-methyl-N-nitrosourea in all the organs studied, the greatest methylation being in the brain. 5. The rate of removal of 7-methylguanine from DNA of kidneys from rats treated with dimethylnitrosamine was compared with the rate after treatment of rats with methyl methanesulphonate. No striking difference was found. 6. The results are discussed in connexion with the organ distribution of tumours induced by the compounds under study and in relation to the possible importance of alkylation of cellular components for the induction of cancer.     

Organ-specific effects of DNA methylation by alkylating agents in the inbred Swiss mouse.

Young adult inbred Swiss mice given single or repeated equitoxic doses of N-methyl-N-nitrosourea (MNUA) or methyl methanesulphonate (MMS) develop thymomas and pulmonary adenomas only following MNUA in spite of nearly identical overall alkylation of DNA of tumour target tissues by both agents due mainly to the biologically ineffective product 7-methylguanine. The main difference in DNA alkylation was the production of O6-methylguinine, a known pre-mutagenic product, by MNUA in amounts 10 or more times as large as following MMS. This supports the possibility that somatic mutations are a part of the process of carcinogenesis.     

Repair studies on methylated DNA of CHO cells

DNA methylation pattern was investigated on Chinese hamster ovary (CHO) cells after treatment with N-(14C)-methyl-N-nitrosourea (14C-MNU). The main target was the N-7 position of guanine, exceeding the methylation in the O6 position of guanine by a factor of 8 and that in the N-3 position of guanine and adenine by a factor of 20. No DNA repair could be observed within 2 hours after methylation. Pretreatment of cells with gamma irradiation (7 rad) before application of MNU induced repair of N-7-methylguanine. This methylation product was decreased to about 50% within two hours, whereas the repair of the other methylated bases was not influenced. The analysis was carried out by high performance liquid chromatography after acid hydrolysis of isolated DNA. 14C-methylated products were determined by liquid scintillation counting.     

A system in mouse liver for the repair of O6-methylguanine lesions in methylated DNA.

An activity from mouse liver with catalyzes the disappearance of O6-methylguanine from DNA methylated with methylnitrosourea has been partially purified by ammonium sulfate fractionation and DNA-cellulose chromatography. The activity does not require divalent metal ions and is not affected by EDTA. It is specific for the repair of O6-methylguanine lesions and does not affect the removal of 7-methylguanine, 7-methyladenine or 3-methyladenine. The disappearance of O6-methylguanine is linear with respect to the concentration of protein and is dependent on incubation temperature. The kinetics and substrate dependence experiments suggest that the protein factor is product-inactivated. Amino acid analysis of hydrolysates of protein obtained after incubation of methylated DNA with the protein factor indicates the presence of radiolabeled S-methyl-L-cysteine, suggesting that during the repair of O6-methylguanine from methylated DNA, the methyl group is transferred to a sulfhydryl of a cysteine residue of a protein. This represents the first such demonstration in a mammalian system.     

Mutant Escherichia coli Ada proteins simultaneously defective in the repair of O6-methylguanine and in gene activation.

The activated Ada protein triggers expression of DNA repair genes in Escherichia coli in response to alkylation damage. Ada also possesses two distinct suicide alkyltransferase activities, for O6-alkylguanines and for alkyl phosphotriesters in DNA. The mutant Ada3 and Ada5 transferases repair O6-methylguanine in DNA 20 and 3000 times more slowly, respectively, than the wild-type Ada protein, but both exhibit normal DNA phosphotriester repair. These same proteins also exhibit delayed and sluggish induction of the ada and alkA genes. Since the C-terminal O6-methylguanine methyltransferase domain of Ada is not implicated in the direct binding of specific DNA sequences, this part of the Ada protein is likely to play an alternative mechanistic role in gene activation, either by promoting Ada dimerization, or via direct contacts with RNA polymerase.     

Identification of C8-methylguanine in the hydrolysates of DNA from rats administered 1,2-dimethylhydrazine. Evidence for in vivo DNA alkylation by methyl radicals.

C8-Methylguanine was identified in the neutral hydrolysates of DNA isolated from the liver or colon tissue of rats administered 1,2-dimethylhydrazine. In all the samples examined, the biologically isolated adducts were characterized by co-elution with synthetic C8-methylguanine under different high pressure liquid chromatography conditions. The sample isolated from liver DNA was also identified by UV spectroscopy at different pH values and by mass spectrometry. The estimated yields of C8-methylguanine obtained in hydrolysates of DNA from the liver or colon tissue were comparable to those of O6-methylguanine. C8-Methylguanine was not detected when the spin trap alpha-(4-pyridyl-1-oxide)-N-tert- butylnitrone was administered together with 1,2-dimethylhydrazine. The spin trap also inhibited N7-methylguanine and O6-methylguanine yields, although to a lesser extent. These results constitute the first evidence that DNA alkylation by carbon-centered radicals can occur in vivo.     

Investigation of the specificity of O6-alkylguanine-DNA-alkyltransferase

Comparison of the abilities of alkylated RNA and DNA to serve as substrates for the O6-alkylguanine-DNA-alkyltransferase have been carried out. It was found that the O6-methylguanine in tRNA was much less active as a substrate for the protein than O6-methylguanine in double stranded DNA. The difference in rates of repair was such that it is unlikely that the alkyltransferase would act on RNA in vivo and, therefore, the reaction with RNA should not contribute towards the exhaustion of its repair capacity.     

Suicide inactivation of the E. coli O6-methylguanine-DNA methyltransferase

The O6-methylguanine-DNA methyltransferase of Escherichia coli acts rapidly and stoichiometrically to convert a mutagenic O6-methylguanine residue in DNA to unsubstituted guanine. Even at low protein concentrations and in the absence of any cofactors, the transfer of a methyl group to one of the protein's own cysteine residues occurs in less than 2 s at 37 degrees C. The entire kinetic process can be followed experimentally at 5 degrees C. Formation of S-methylcysteine in the protein is accompanied by loss of activity and accounts for the exceptional suicide kinetics of this enzyme as well as for the sharp saturation of O6-methylguanine repair observed in vivo. The enzyme can remove greater than 98% of the methyl groups from O6-methylguanine present in alkylated DNA, but leaves N-alkylated purines untouched. Single-stranded DNA containing O6-methylguanine is a poor substrate, with the methyl transfer occurring at approximately 0.1% of the rate for duplex DNA. This latter observation may explain the high frequency of mutations induced by alkylating agents at DNA replication forks.     

Mouse embryonic stem cells are hypersensitive to apoptosis triggered by the DNA damage O(6)-methylguanine due to high E2F1 regulated mismatch repair

Exposure of stem cells to genotoxins may lead to embryonic lethality or teratogenic effects. This can be prevented by efficient DNA repair or by eliminating genetically damaged cells. Using undifferentiated mouse embryonic stem (ES) cells as a pluripotent model system, we compared ES cells with differentiated cells, with regard to apoptosis induction by alkylating agents forming the highly mutagenic and killing DNA adduct O(6)-methylguanine. Upon treatment with N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), ES cells undergo apoptosis at much higher frequency than differentiated cells, although they express a high level of the repair protein O(6)-methylguanine-DNA methyltransferase (MGMT). Apoptosis induced by MNNG is due to O(6)-methylguanine DNA adducts, since inhibition of MGMT sensitized ES cells. The high sensitivity of ES cells to O(6)-methylating agents is due to high expression of the mismatch repair proteins MSH2 and MSH6 (MutSalpha), which declines during differentiation. High MutSalpha expression in ES cells was related to a high hyperphosphorylated retinoblastoma (ppRb) level and E2F1 activity that upregulates MSH2, causing, in turn, stabilization of MSH6. Non-repaired O(6)-methylguanine adducts were shown to cause DNA double-stranded breaks, stabilization of p53 and upregulation of Fas/CD95/Apo-1 at significantly higher level in ES cells than in fibroblasts. The high apoptotic response of ES cells to O(6)-methylguanine adducts may contribute to reduction of the mutational load in the progenitor population.