O1-methylguanine (O1mG) produced in DNA by such SN1 methylating agents as N-methyl-N-nitrososurea and N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) has been suggested by some to be the lesion that leads to certain biological endpoints in mammalian cells: cell killing, sister chromatid exchange (SCE) production, mutagenesis and cellular transformation. Other evidence is interpreted as inconsistent with this point of view. The finding of Karran & Williams (1985) that O1mG delivered to cells in culture resulted in the depletion of the activity of the protein responsible for repair of O1mG in DNA (O1mG-DNA methyltransferase, O1MT) provided a tool for the assessment of the role of O1mG in producing biological endpoints. In this paper we review much of the literature on human cells pertinent to this question. In addition we present our survival data obtained using the depletion technique of Karran & Williams as well as data supporting a model invoking a mismatch and excision response to O1mG proposed by Sklar & Strauss (1980). Although data linking O1mG to causation are inconclusive, it is premature to conclude that O1mG is not a lesion lethal to certain cultured cells.

The ideal system: one base change/one lethal hit

It is important to know the identity of the DNA lesions that produce biological endpoints such as sister chromatid exchange (SCE) induction, mutation and loss of reproductive capacity, because this knowledge will lead to the experimental determination of the molecular mechanism by which these endpoints are caused. Identification of the initiating lesion is merely the first step. The fact that the production of about one pyrimidine dimer per genome correlated with the production of one ‘lethal hit’ in a population of uvrA recA Escherichia coli (Howard- Flanders & Boyce, 1966) demonstrated the possibility of assigning specific DNA lesions to the production of biological endpoints in genomes having of the order of 101 nucleotides. Bacteria that are wild-type with respect to repair contained 3700 dimers at one hit lethality (37 % survival). In uvrA mutants, unable to repair dimers, a lethal hit correlated with the production of roughly 50 dimers per genome and in uvrA recA mutants with 1·5 dimers per genome. It is likely that one to two dimers are lethal to uvrA recA mutants, but it is impossible to distinguish without further knowledge whether a dimer or another lesion, possibly a lesion produced 50-fold less frequently but repaired like a dimer, is the lesion lethal to uvrA mutants. We believe dimers are lethal to uvrA mutants for two reasons. The first is that photoreactivation, which reverses only dimers among known DNA photoproducts, reverses the lethal effects due to much of the ultraviolet light (u.v.) dose to uvrA mutants. The second is that restoring to the uvrA mutants the ability to remove dimers increases post-u.v. survival. (See Harm (1980) for review.) Eight years ago a spokesman from the DNA- repair community charged it with the responsibility of determining the biological effects of individual lesions rather than of DNA damaging agents (Cerutti, 1978). This has not proved to be an easy task.

Lesions due to methylating agents: strategies to determine their effects

We are concerned with considering the identity of the lesions that direct the biological changes in human cells treated with the DNA methylating agents MNNG (N-methyl-N′-nitro-N-nitrosoguanidine) andMNU (N-methyl-N-nitrosourea). The genome size of human cells is about 2 m, 2000 times that of E. coli, but the basic considerations are the same as used for E. coli. We find mutants that are unable to repair given lesions and determine what differences in methylation-produced biological phenomena occur; or we produce changes in the number of one kind of lesion and observe the corresponding changes in biological effects. If we find cases in which there are fewer than one of a given lesion per genome but more than one biological event per biological unit, that lesion is unlikely to have caused the effect.

MNNG or MNU produce many alterations in DNA components, methylating nitrogen and oxygen atoms (Fig. 1). MMS produces 25-fold less O1-methylguanine (O1mG). Comparing biological responses after treatment with MMS with those obtained after treatment with MNNG or MNU can thus give information regarding the relative roles of methylated nitrogens and methylated oxygens in producing the effects.

Fig. 1.

Products of the reaction of MNNG or MMS with DNA. (From Beranek et al. 1980; see also Lawley & Thatcher, 1970; Swenson & Lawley, 1978.)

Fig. 1.

Products of the reaction of MNNG or MMS with DNA. (From Beranek et al. 1980; see also Lawley & Thatcher, 1970; Swenson & Lawley, 1978.)

Recognition of O1-methylguanine: altered base pairing

The question of the biological effects due to O1mG begins with the paper of Loveless (1969), who when reporting the discovery of O1mG in DNA treated with N-methyl-N-nitrosourea pointed to the expected ‘atypical base pairing’ properties of O1mG as a possible explanation of the high mutagenicity and carcinogenicity of compounds like MNU (Fig. 2). Overall, pairing of O1mG with thymine appears better than pairing of O 6mG with cytosine. The question is whether and in what ways the cell is able to recognize such mispairing in DNA. The presence of misrepaired O 6mG might be monitored during DNA replication, or there might be a mismatch repair system involved in O 6mG recognition. (For a short review, see Discussion.)

Fig. 2.

Fig. 2. Base pairing of guanine with cytosine and of O1-methylguanine with cytosine and thymine. (Based on discussion by Loveless, 1969).

Fig. 2.

Fig. 2. Base pairing of guanine with cytosine and of O1-methylguanine with cytosine and thymine. (Based on discussion by Loveless, 1969).

A lesion having a methylated oxygen is lethal to adenovirus when infecting cells deficient in repair of O6mG

MNNG-treated adenoviruses show greater survival when infecting cells of some lines (defined as Mer+) than when infecting cells of others (Mer; and the differential survival correlates with an ability or inability of the cells to repair O1mG in DNA; such ability or inability is defined as Mex+ or Mex, respectively) (Day et al. 1980b; Sklar & Strauss, 1981; Watatani et al. 1985). A similar differential inactivation is seen with adenoviruses treated with MNU (Day et al. 1980a) or MTIC (5-(3′-methyl-l-triazeno)imidazole-4-carboximide (Hayward & Parsons, 1984). By contrast, adenoviruses treated with MMS show a survival independent of the ability of the cells infected to repair O1mG (Scudiero et al. 1984&). It would therefore seem likely that a lesion produced by MNNG, but not by MMS, is a lesion lethal to the virus. The lesion would be repaired by a mechanism lacking in Merstrains, possibly similar to that which repairs O1mG. It was determined that one lethal hit to the virus population correlates with the production of 2·3 O1mG, 1·4 3mA, 16 7mG, and about 5 methyl phosphotriesters per viral genome (Fig. 3; and Day et al. 1984). All other lesions as O1mT, O1mT, 3mG and 7mA are too few to be present at one/genome for one lethal hit per virion. 3mA is repaired both by cells able and by cells unable to repair O1mG (Day et al. 1984). 7mG (Day et al. 1980a, 1984) and methyl phosphotriesters (Yarosh et al. 1985) appear to be repaired slowly or not at all and independently of the ability of the cell to repair O1mG. Thus, if one were pressed for an answer, one would guess that O1mG is the lesion lethal to adenovirus infecting cells deficient in ability to repair O1mG.

Fig. 3.

Survival of [1H]MNNG treated adenovirus 5 as a function of O1-methylguanines (○, •) or 3-methyladenines (▴) produced per viral genome. (•) and (○): respectively, survival as measured in strains incapable and capable of repairing O1mG in their DNA; (–––) the survival expected if one methylated purine per genome were lethal.

Fig. 3.

Survival of [1H]MNNG treated adenovirus 5 as a function of O1-methylguanines (○, •) or 3-methyladenines (▴) produced per viral genome. (•) and (○): respectively, survival as measured in strains incapable and capable of repairing O1mG in their DNA; (–––) the survival expected if one methylated purine per genome were lethal.

Consideration of O1mG as a lesion lethal to cells

Scudiero et al. (1984a) studied inactivation by MNNG of 23 human cell strains prepared from normal and tumorous human tissue. Their sensitivities, assayed by cell killing, correlated with their ability to repair O1mG. Sensitivities fell into one of three groups (Fig. 4). The first group showed proficient repair of MNNG-treated adenovirus (Mer+), repaired O1mG in their DNA (Mex+) and had cell killing sensitivities in the range of normal human fibroblasts (Rem+; Scudiero, 1980; Scudiero et al. 1981). Strains belonging to the second group were also able to repair MNNG-treated adenoviruses (Mer+), but repaired O1mG (and so are Mex+) but less than the first, and showed three- to fourfold more sensitivity to MNNG killing (Rem). The third group (most sensitive to killing by MNNG, Rem) showed lack of ability to repair MNNG-treated adenoviruses (Mer) and inability to repair O1mG (Mex). Sensitivities ranged from 0·9 lethal hits per μM-MNNG in an osteosarcoma cell line to 18·0 hits per μM-MNNG for an ovarian cancer line, approximately 100-fold more sensitive to killing. As a group, the Mer Rem strains were 20-fold, and Mer+ Rem strains threefold, more sensitive to killing than was the Mer+ Rem+ group. Thus a good correlation between O1mG repair and sensitivity to MNNG was established. Evidence along similar lines was published by Day et al. (19806) and by Domoradski et al. (1984).

Fig. 4.

Survival of human cell strains as a function of MNNG concentration. Mer+ Rem+ strains: (○) KD; HT29; (◊) TE85; (▫) HuTu80; A673; (▿) U138MG; U373MG; A427F; (▵) CRL 1187. Mer+ Rem strains: (◑) A2182;A704; (◧) A549; (◭) A498; A388. Mer Rem strains: (•) A172; (▴) A1235; (▾) U105MG; (◆) U87MG; (▪) A2095; A1336;, A875; A427; BE.

Fig. 4.

Survival of human cell strains as a function of MNNG concentration. Mer+ Rem+ strains: (○) KD; HT29; (◊) TE85; (▫) HuTu80; A673; (▿) U138MG; U373MG; A427F; (▵) CRL 1187. Mer+ Rem strains: (◑) A2182;A704; (◧) A549; (◭) A498; A388. Mer Rem strains: (•) A172; (▴) A1235; (▾) U105MG; (◆) U87MG; (▪) A2095; A1336;, A875; A427; BE.

Scudiero et al. (19846) studied many of the same strains in an attempt to understand the mode of cell killing by other chemical agents, methylmethane sulphonate (MMS), l,3-bis-(2-chloroethyl)-l-nitrosourea (BCNU), l-(2-chloro- ethyl)-3-(2-hydroxyethyl)-l-nitrosourea (HECNU) and N-ethyl-N′-nitro-N-nitro- soguanidine (ENNG) (Fig. 5). With MMS, which produces primarily 7mG and 3mA in DNA, there is less of a range of cell killing than with MNNG. On average the strains deficient in O1mG repair were about two to four times more sensitive to MMS killing than the repair-proficient ones. MMS produces about 25-fold less O1mG than does MNNG (Lawley et al. 1975) or MNU (Beranek et al. 1980) relative to total methylated DNA adducts produced. The different sensitivities of the strains to inactivation by MNNG and MMS are therefore consistent with a heightened lethality of 01mG in Mer- cells. Sensitivities of the strains to inactivation by MNNG and MMS would be understood.

Fig. 5.

Survival of human fibroblasts and tumour cell strains after treatment with MMS (A), BCNU (B), HECNU (C), ENNG (D), MNNG (E) or u.v. (F). Cells were treated and assayed for remaining colony-forming ability. Symbols are the same as for Fig. 4.

Fig. 5.

Survival of human fibroblasts and tumour cell strains after treatment with MMS (A), BCNU (B), HECNU (C), ENNG (D), MNNG (E) or u.v. (F). Cells were treated and assayed for remaining colony-forming ability. Symbols are the same as for Fig. 4.

Events other than DNA methylation in treated cells should be mentioned. We are assuming that it is DNA alteration by MNNG, MNU or MMS that produces the lesions lethal to cells. This is a reasonable assumption because in the field of DNA repair, DNA-repair defects, not other defects, have been identified in cells abnormally sensitive to DNA damaging treatments. However, it is known that MNNG and MNU methylate primarily the lysines and arginines of cellular nuclear proteins, whereas MMS methylates cysteine and histidine preferentially (Boffa & Bolognesi, 1985a,b), so that methylation of protein could conceivably contribute to the differential lethal effects of MMS and MNNG. Because this would appear to be a less-attractive possibility, very little work has been done in this area. In addition, glutathione depletion due to MNNG-activation (Lawley & Thatcher, 1970; Sedgwick & Robins, 1980) may be important in the cellular response to that agent.

MerRem+strains: fly in the ‘O1 in tment’ or proof of the pudding?

It was shown that transformation of apparently normal Mer+ Mex+ fibroblasts by simian virus 40 (SV40) often results in the production of Mer Mex transformed cell lines (Day et al. 1980). Indeed 8 of 12 SV40-transformed cell lines have now been shown to be Mer- in our laboratory. The response of these Mer lines to cell killing by MNNG is greatly different from that expected on the basis of the extreme sensitivity shown by Mer human tumour lines (Day et al. 1980; Scudiero & Day, unpublished results). The Mer- SV40-transformed lines are as resistant to killing by MNNG as are both Mer+ SV40-transformed lines (Fig. 6) and normal human fibroblasts, and are therefore called Mer Rem+. Although these strains are resistant to killing by MNNG, we believe that other assays may reflect their inability to repair methylated-oxygen lesions in DNA: (1) they fail to repair MNNG-damaged adenovirus (Day et al. 19806); (2) they show as sensitive induction of SCE due to MNNG treatment as seen in the tumour Mer cell lines (Day et al. 19806; (3) they are as hypersensitive to killing by HECNU as are the tumour Mer- lines (Scudiero et al. 19846); (4) they are hypermutable by MNNG compared with a Mer+ SV40- transformed line (Baker et al. 1980; these authors were using cells differing in repair of O1mG, see Day et al. 1980b).

Fig. 6.

Survival of SV40-transformed human fibroblast lines after alkylation treatment. Mer+ strains: (○) W98VA1; (▫) WI26VA4. Mer strains: (•) W18VA2; (▴) GM638; (▾) IMR90-830; (▪) WI38VA13.

Fig. 6.

Survival of SV40-transformed human fibroblast lines after alkylation treatment. Mer+ strains: (○) W98VA1; (▫) WI26VA4. Mer strains: (•) W18VA2; (▴) GM638; (▾) IMR90-830; (▪) WI38VA13.

Thus, if O1mG is a lesion lethal to the tumour Mer- lines, it is certainly not a lesion lethal to the SV40-derived Mer lines. And one could certainly remain within the guidelines drawn by the data should one propose that O1mG is not a lethal lesion at all. The most direct rationale for this is that two groups of cell lines (Mer Rem+, Mer Rem), showing widely different cellular sensitivities to MNNG killing, both fail to repair O1mG. Therefore, how could O1mG be implicated in lethality? It is more difficult to build a good argument against the involvement of an alkylated DNA oxygen in MNNG-produced SCE formation and mutagenesis, and in HECNU- produced cell killing. In these endpoints, lines that fail to repair O1mG are uniformly hypersensitive, independently of their sensitivity to MNNG-produced cell killing. We know that Mer cells lack repair of O1mG. It is possible that they all lack another repair mechanism, e.g. for 3mG, 3mA or another methylated DNA nitrogen. Were this the case, MNNG and MMS would be expected to cause the same magnitude of differential response in Mer+ and Mer cells. For adenovirus inactivation this is not the case (Scudiero et al. 19846). MNNG produced many more SCE than did MMS in Mer than in Mer+ lines (Wolff et al. 1977; the lines used were later identified as Mer and Mer+; Day et al. 1980). We are not, however, aware of a study comparing MMS and MNNG with regard to mutagenesis of human cells. Certainly, there is among human cells a correlation between ability to repair O1mG and resistance to MNNG-induced mutagenesis (Baker et al. 1979, 1980; the cells used were Mer and Mer+; see Day et al. 1980b; Domoradski et al. 1984).

In summary, the very existence of MNNG-resistant lines that lack repair of O1mG casts considerable doubt as to the role of oxygen-methylated DNA bases in lethality to Mer Rem+ cells. On the other hand, these lines have sensitivities much like MNNG-sensitive O1mG repair-deficient (Mer Rem) lines in terms of repair of methyl-damaged virus, susceptibility to MNNG-produced SCE and mutations, and to killing by HECNU. This latter encourages the idea that O1mG (or another lesion both repaired by a mechanism lacking in Mer- cells and produced by agents that produce O1mG) leads to these endpoints.

Depletion of O1mG-DNA methyltransferase by free O1mG: the ultimate test?

If one were able to alter cellular repair of O1mG in DNA preferentially over other lesions, one would have a very good tool for determining the effects due to O1mG in DNA. Karran (1985) and Karran & Williams (1985) found that free O1mG, when delivered in culture to cells proficient in repair of O1mG, would result in the depletion of O1MT (O1mG-DNA methyltransferase), the protein responsible for repairing O1mG in Mer+ cells and absent from Mer cells (Yarosh et al. 1983, 1984). This finding provided a tool for reducing cellular capacity to repair O1mG in a presumably selective fashion. Karran & Williams (1985) reported that depletion of O1MT by 0·1 mM free O1mG failed to sensitize Mex+ Raji cells to inactivation either by MNNG or by CNU (l-(2-chloroethyl)-l-nitrosourea). Were all O1MT activity fully depleted during the period when DNA lesions produced by MNNG became lethal, then the conclusion of the authors, i.e. that O1mG in DNA has nothing to do with cell killing by either agent, would be correct. However, Dolan et al. (1985b) showed that a greater dose of O1mG (0·4 mM) sensitized HeLa cells to killing by CNU in a major way, and to killing by MNNG in a minor way. Mutagenesis due to MNNG treatment of human fibroblasts was enhanced somewhat by a pretreatment with 0’4mM-O1mG (Domoradski et al. 1985). Our own results, obtained with a 2mM-O1mG pretreatment (Fig. 7) are consistent with those of Dolan et al. (1985b). As a control, we detect little sensitizing effect of the pretreatment on the killing of Mer cells by CNU. This would be expected if a lesion not repaired by Mer cells were the lesion whose lethality was enhanced by the O1mG pretreatment. We have found that protein fractions containing O1MT activity had little if any capacity to repair methylphosphotriesters or O1mT in DNA (Yarosh et al. 1985) so it is unlikely that depletion of O1MT alters repair of these lesions. Thus, the O-4 position of Thy and DNA phosphate can in all likelihood be excluded from consideration as target sites for lethality of CNU.

Fig. 7.

Survival of Mer+ (HT29, A549, KD) and Mer (A1336, IMR90-830) strains with (filled symbols) and without (open symbols) a 3-h pretreatment with O1mG just prior to a 1-h treatment with CNU.

Fig. 7.

Survival of Mer+ (HT29, A549, KD) and Mer (A1336, IMR90-830) strains with (filled symbols) and without (open symbols) a 3-h pretreatment with O1mG just prior to a 1-h treatment with CNU.

We do not detect a great sensitization by free O1mG of Mer+ cells to killing by MNNG (Fig. 8). However, treatment of HT29 cells with even 2mM-O1mG fails to block completely the repair of O1mG produced in HT29 DNA by a 1-h treatment of cells with 12·5 μM-MNNG (Fig. 9). Therefore, significant recovery of the cells due to repair of O1mG in DNA is expected. We agree with the explanation of Dolan et al. (1985a) that accounts for the difference between sensitization to CNU and to MNNG. In the case of CNU, if O1-choroethylguanine is not repaired rapidly, DNA crosslinks that are not repairable by O1MT would accumulate within hours (Kohn, 1977) (see Fig. 10; for discussion of repair of O1-haloethylguanine adducts and the DNA crosslinks they produce, see Scudiero et al. (1984b).) The lethal lesions, the DNA crosslinks (Sasaki (1977) has shown that crosslinking agents are five- to tenfold more lethal than their non-crosslinking congeners), would become fixed. In the case of MNNG, where little sensitization due to depletion with O1mG occurs, DNA lesions relevant to survival would be repairable over a longer time before their lethality becomes fixed. Several experimental results would be expected according to this hypothesis: (1) Mer Rem+ strains, to which we do not believe O1mG should be lethal, should not be sensitized at all to killing by MNNG by the free O1mG treatment; and (2) cells whose O1MT is depleted by free O1mG prior to MNNG treatment, and which are thereby sensitized somewhat to MNNG killing, should be sensitized further by a post-MNNG treatment with free O1mG. However, O1mG may not be lethal. O1mT, O1mC and O1mT are repairable lesions (Den Engelse et al. 1986), and these are logical candidates for being lesions lethal to Mer Rem cells. It is conceivable that O1mG could be excised from DNA. Such removal would not be likely to be blocked by the free base O1mG so that sensitization of colony-forming ability by free O1mG would not be observed. If O1mG were excised from DNA, Mer cells would have to lack the excision process as well as O1MT because there is no loss of O1mG from MNNG-treated Mer cells.

Fig. 8.

Survival of Mer+ (HT29 and A549) strains with (•) and without (○) a 3-h pretreatment with O1mG just prior to a 1-h treatment with MNNG.

Fig. 8.

Survival of Mer+ (HT29 and A549) strains with (•) and without (○) a 3-h pretreatment with O1mG just prior to a 1-h treatment with MNNG.

Fig. 9.

Persistence of O1mG and 3mA (relative to 7mG) in HT29 cells that were treated for 1 h with 12·5 μM-[METHYL-1H]MNNG just after a 3-h pretreatment with 2mM-O1mG (•) or after a mock pretreatment (○). A, O1mG; B, 3mA.

Fig. 9.

Persistence of O1mG and 3mA (relative to 7mG) in HT29 cells that were treated for 1 h with 12·5 μM-[METHYL-1H]MNNG just after a 3-h pretreatment with 2mM-O1mG (•) or after a mock pretreatment (○). A, O1mG; B, 3mA.

Fig. 10.

Reaction of chloroethylating nitrosoureas with guanine in DNA: crosslink formation in Mer cells; and repair by Mer+ cells. For reactivity of chloroethylisocyanate see Kann et al. (1974). For crosslink formation by chloroethylating agents by the pathways marked by the *, see Tong et al. (1982).

Fig. 10.

Reaction of chloroethylating nitrosoureas with guanine in DNA: crosslink formation in Mer cells; and repair by Mer+ cells. For reactivity of chloroethylisocyanate see Kann et al. (1974). For crosslink formation by chloroethylating agents by the pathways marked by the *, see Tong et al. (1982).

The depletion of O1MT produced by treating cells with the free base O1mG sensitizes cells much more to killing by CNU than to killing by MNNG. Although both results can be understood in terms of O1mG (or related lesion) as lethal, explanations based on the lethality of other lesions are not excluded.

Increased DNA repair synthesis in Mer strains: a key to the lethal lesion?

Because of its altered base pairing, it is not difficult to imagine why O1mG might be mutagenic. It is far more difficult to divine the mechanism by which O1mG might be lethal. We suspect that a type of mismatch repair operating improperly on the O1mG: C mismatch may trigger lethality.

Altamirano-Dimas et al. (1979), Day et al. (1980a) and Scudiero et al. (1984a) have shown that MNNG-treated Mer (Mex) and Mer+ Rem strains show more incorporation of [1H]dThd into double-stranded DNA than do MNNG-treated Mex+ strains. Together with the fact that this damage-stimulated incorporation (DNA repair synthesis) was less with MMS (Altamirano-Dimas et al. 1979), the evidence points to a DNA lesion methylated at oxygen as causing the increase. Mer Mex strains show no removal of the 01mG produced in their DNA. Were O1mG the lesion triggering increased DNA repair synthesis it would not, therefore, be reasonable to suppose that the increased DNA repair synthesis is associated with removal of the DNA lesion, as in the case of removal of pyrimidine dimers from the DNA of u.v.-irradiated normal fibroblasts (Cleaver, 1968). However, repair may occur in the strand opposite the lesion (Sklar & Strauss, 1980), in which case either dCMP or dTMP might be incorporated opposite O1mG. Because neither Cyt nor Thy pairs exceptionally well with O1mG (see Discussion) a recognizable mismatch might still exist. The mismatch would continue to stimulate incorporation of dTMP into the DNA of Mer- cells (Fig. 11), whereas Mer+ cells would be expected to cease repair DNA synthesis once the lesions were removed by O1MT. The results of an experiment in which cells were treated with 40 μg ml−1 MNNG for 1 h and incubated with [1H]dThd (in the presence of hydroxyurea to block semiconservative DNA synthesis) for 1-h periods are shown in Fig. 12. The A172 and A427 strains (both Mer Rem ) show such an effect as does the A549 strain (Mer+ Rem; the O1MT is severely depleted at this MNNG dose; Scudiero et al. 1984a). Further experiments need to be done.

Fig. 11.

Scheme to account for increased MNNG-induced incorporation of [1H]dThd into double-stranded DNA by Mer compared with Mer+ cells after MNNG treatment.

Fig. 11.

Scheme to account for increased MNNG-induced incorporation of [1H]dThd into double-stranded DNA by Mer compared with Mer+ cells after MNNG treatment.

Fig. 12.

Accumulated MNNG-stimulated DNA synthesis in double-stranded DNA as a function of post-MNNG incubation time. Cells were blocked in DNA synthesis by 10mM-hydroxyurea (HU) for 30 min, then treated from 0–1 h with 40μgml−1 MNNG and 10 mM-HU. At 1 h, the medium was removed and replaced with medium containing 10mM-HU. During the 1-h prior to a data point the cells were exposed to the same medium, but containing 5 μCiml−1 [1H]dThd in order to label repaired regions of DNA. Separation of double-stranded DNA from DNA with single strands was done by BND- cellulose chromatography. (See Day et al. 1980a; Scudiero et al. 1984b, for details.) Cell strains: Mer+ Rem+: (○) KD; (▵) CRL 1187; (▿) U138MG; (▫) TE85. Mer+ Rem: (◒) A388; AS49. Mer Rem: (•) A172; (▴) A427N; (▪) A123S.

Fig. 12.

Accumulated MNNG-stimulated DNA synthesis in double-stranded DNA as a function of post-MNNG incubation time. Cells were blocked in DNA synthesis by 10mM-hydroxyurea (HU) for 30 min, then treated from 0–1 h with 40μgml−1 MNNG and 10 mM-HU. At 1 h, the medium was removed and replaced with medium containing 10mM-HU. During the 1-h prior to a data point the cells were exposed to the same medium, but containing 5 μCiml−1 [1H]dThd in order to label repaired regions of DNA. Separation of double-stranded DNA from DNA with single strands was done by BND- cellulose chromatography. (See Day et al. 1980a; Scudiero et al. 1984b, for details.) Cell strains: Mer+ Rem+: (○) KD; (▵) CRL 1187; (▿) U138MG; (▫) TE85. Mer+ Rem: (◒) A388; AS49. Mer Rem: (•) A172; (▴) A427N; (▪) A123S.

Lethal lesions: cell and virus

We have seen that the group of MNNG-produced lesions possibly lethal to cells that fail to repair O1mG is narrowed to those lesions not made in any great quantity by MMS, i.e. almost certainly to lesions containing methylated oxygen. O1mG is very probably the lesion lethal to MNNG-treated adenovirus infecting cells incapable of repairing O1mG, encouraging the idea that O1mG can be lethal. In addition CNU or HECNU, both chloroethylating agents believed to produce crosslinks in DNA through an O1-chloroethylguanine intermediate, probably produce lethal lesions primarily by attacking the O-6 of guanine.

Methylated DNA components, leading to SCE and mutations

The lesion responsible for the production of SCE and mutagenesis is one that is repaired by neither Mer Rem nor Mer Rem+ strains. It may be O1mG or another lesion that is not lethal to Mer” Rem” strains. In the case of production of SCE, the lesion appears not to be produced by MMS. In the case of mutagenesis of human cells, comparative data obtained with MMS are not available to these authors’ knowledge. (They would certainly like to be informed if there is.) Were O1mG again considered to be the mutagenic lesion, the available data would be accommodated, including the increase in mutability seen after O1MT depletion (Domoradski et al. 1985).

O1mG in DNA is detected by cells: polymerases and mismatch recognition

In this section we consider the known ways in which in vitro or in vivo systems are known to respond to O1mG to provide a basis for the hypothesis in Fig. 11. Polymerases. Using synthetic templates with DNA polymerase I from bacteriophages T4 and T5, and E coli, Snow et al. (1984) found that O1mG in template DNA directed the in -vitro incorporation of dTMP in preference to dCMP by ratios of 6 (T4), 10 (T5) and 100 (E. coli). O1mG in the template slowed down but did not terminate incorporation. Stronger preference for TMP incorporation was accompanied by a high turnover ratio (dNMP released per dNMP incorporated). Blocks to DNA synthesis introduced by MNNG into DNA used as an in vitro template for primed synthesis by the polymerase I Klenow fragment occurred most frequently before the positions of Ade residues (Larson et al. 1985), indicating no particular tendency for O1mG to cause chain termination. There are conflicting reports concerning the ability of polA mutants of E. coli (which lack polymerase I) to repair O1mG in DNA (Warren & Lawley, 1980; Karran et al. 1982), but it appears thatpolA mutants do not show the adaptive response (Samson & Cairns, 1977) to the lethal effects of alkylating agents (Jeggo et al. 1977, 1978), indicating involvement of polymerase I in this response.

Mismatch repair

One mismatch repair mechanism in E. coli is believed to work on newly replicated DNA. Mismatches are detected, and the daughter strand, distinguished from the m6Ade-methylated parental strand by its lack of m6Ade, is repaired. The m6Ade-methylated parental strand is used as a template, and replacement of DNA occurs over a stretch of some 3000 nucleotides (Wagner & Meselson, 1976).

A site-specific mismatch repair mechanism that restores CC(A/C)GG sequences in phage lambda operates over a shorter distance (roughly 5–10 base pairs) and is independent of Jam-dependent Ade (or Jem-dependent Cyt) methylation (Lieb, 1985). A similar but not identical system operates in Streptococcus pneumoniae (Lacks et al. 1982; Lefevre et al. 1984).

On the basis of their data on MNNG-induced multiple site mutations in E. coli, Sklar & Strauss (1980) suggested that the uvrE product may be involved in correction of a mismatch between an O1mG (parental strand) paired with a Cyt (daugher strand) near the replication fork. The resolution of the mismatch would, however, not be repair, but error production: the removal of Cyt and replacement with Thy. E. coli dam mutants are sensitive to MNNG (but not dimethylsulphate, DMS)-produced killing, and this sensitivity was suppressed by mutL or mutS, indicating involvement of a mismatch-triggered response in determining resistance to a lesion made by MNNG but not by DMS, presumably a lesion having a methylated oxygen (Karren & Marinus, 1982). Eadie et al. (1984) used the polymerase I Klenow fragment to incorporate O1mG during in vitro DNA synthesis across from a single-stranded bacteriophage f1/pBR322 based Amp+ vector used as template. Amp mutants were produced when transfecting the double-stranded product into E. coli mutH deficient in mismatch repair, or when the replicated vector was methylated with dam methylase (removing the signal triggering mismatch repair) prior to transfection of cells competent to repair mismatches. Such mutants were not observed on transfection of repair-proficient hosts. Solely AT to GC transitions were observed as expected if O1mGMP had been incorporated across from Thy. Apparently mismatch repair processes resulted in the loss of mutant molecules of the vector. In summary, 01mG in DNA is a lesion that slows but does not block DNA replication. Its presence in DNA is detected by mismatch repair systems in E coli. It would not be surprising if mammalian cells responded similarly because these also may carry mismatch repair systems (Lai & Nathans, 1975; Miller et al. 1976; Wake & Wilson, 1980; Folger et al. 1985). Hare & Taylor (1985) provided evidence that the strand selection mechanism may be 5mCyt methylation of the parental strand and/or proximity to a nick. The involvement of mismatch recognition in leading to loss of genetic material has been supported in bacterial transformation systems (see Day & Rupert, 1971). Such an event could be lethal.

Altamirano-Dimas
,
M.
,
Sklar
,
R.
&
Strauss
,
B.
(
1979
).
Selectivity of the excision of alkylation products in a xeroderma pigmentosum-derived lymphoblastoid line
.
Mutât. Res.
60
,
197
206
.
Baker
,
R. M.
, Van
Voorhis
,
W. C.
&
Spencer
,
L. A.
(
1979
).
HeLa cell variants that differ in sensitivity to monofunctional alkylating agents, with independence of cytotoxic and mutagenic responses
.
Proc. natn. Acad. Sci. U.S A.
76
,
5249
5253
.
Baker
,
R. M.
,
Zuernodorfer
,
G.
&
Mandel
,
R.
(
1980
).
Enhanced susceptibility of a xeroderma pigmentosum cell line to mutagenesis by MNNG and EMS
.
Environ. Mutagen.
2
,
269
270
.
Beranek
,
D. T.
,
Weis
,
C. C.
&
Swenson
,
D. H.
(
1980
).
A comprehensive quantitative analysis of methylated and ethylated DNA using high pressure liquid chromatography
.
Carcinogenesis
1
,
595
606
.
Boffa
,
L. C.
&
Bolognesi
,
C.
(
1985a
).
Nuclear proteins damaged by alkylating agents wtih different degrees of carcinogenicity
.
Chem.-Biol. Interact.
55
,
235
245
.
Boffa
,
L. C.
&
BOLOGNESI
,
C.
(
1985b
).
Methylating agents: their target amino acids in nuclear proteins
.
Carcinogenesis
6
,
1399
1401
.
Cerutti
,
P. A.
(
1978
).
Repairable damage in DNA
. In
DNA Repair Mechanisms
(ed.
P. C.
Hanawalt
, E. C. Friedberg &
C. F.
Fox
), pp.
1
-
14
. New York, London: Academic Press.
Cleaver
,
J. E.
(
1968
).
Defective repair replication of DNA in xeroderma pigmentosum
.
Nature, Lond.
218
,
652
656
.
Day,
R. S.
Ill
&
Rupert
,
C. S.
(
1971
).
Ultraviolet sensitivity of Haemophilus influenzaetransforming DNA. I. Effects of genetic mismatch and target size
.
Mutât. Res.
11
,
293
311
.
Day
,
R. S. Ill
,
Yarosh
,
D. B.
&
Ziolkowski
,
C. H. J.
(
1984
).
Relationship of methyl purines produced by MNNG in adenovirus 5 DNA to viral inactivation in repair-deficient (Mer) human tumor cell strains
.
Mutât. Res.
131
,
45
52
.
Day
,
R. S. Ill
,
Ziolkowski
,
C. H. J.
,
Scudiero
,
D. A.
,
Meyer
,
S. A.
,
Lubiniecki
,
A. H. J.
,
Girardi
,
A. J.
,
Galloway
,
S. M.
&
Bynum
,
G. D.
(
1980a
).
Defective repair of alkylated DNA by human tumour and SV40-transformed human cell strains
.
Nature, Lond.
288
,
724
727
.
Day
,
R. S. Ill
,
Ziolkowski
,
C. H. J.
,
Scudiero
,
D. A.
,
Meyer
,
S. A.
&
Mattern
,
M. A.
(
1980b
).
Human tumor cell strains defective in the repair of alkylation damage
.
Carcinogenesis
1
,
21
32
.
Den Engelse
,
L.
,
Menkveld
,
G. J.
, De Bru, R.-J. &
Tates
,
A. D.
(
1986
).
Formation and stability of alkylated pyrimidines and purines (including imidazole ring-opened 7-alkylguanine) and alkylphosphotriesters in liver DNA of adult rats treated with ethylnitrosourea or dimethylnitrosamine
.
Carcinogenesis
7
,
393
403
.,
Dolan
,
M. E.
,
Corsico
,
C. D.
&
Pegg
,
A. E.
(
1985b
).
Exposure of HeLa cells to O(6)- alkylguanines increases sensitivity to the cytotoxic effects of alkylating agents
.
Biochem. biophys. Res. Commun.
132
,
178
185
.
Dolan
,
M. E.
,
Morimoto
,
K.
&
Pegg
,
A. E.
(
1985a
).
Reduction of O6-alkylguanine-DNA alkyltransferase activity in HeLa cells treated with O6-alkylguanines
.
Cancer Res.
45
,
6413
6417
.
Domoradski
,
J.
,
Pegg
,
A. E.
,
Dolan
,
M. E.
,
Maher
,
V. M.
&
McCormick
,
J. J.
(
1984
).
Correlation between O6-methylguanine-DNA-methyltransferase activity and resistance to the cytotoxic and mutagenic effect of N-methyl-N’-nitro-N-nitrosoguanidine
.
Carcinogenesis
5
,
1641
1647
.
Domoradski
,
J.
,
Pegg
,
A. E.
,
Dolan
,
M. E.
,
Maher
,
V. M.
&
McCormick
,
J. J.
(
1985
).
Depletion of O6-methylguanine-DNA-methyltransferase in human fibroblasts increases the mutagenic response to N-methyl-N’-nitro-N-nitrosoguanidine
.
Carcinogenesis
6
,
1823
1826
.
Eadie
,
J. S.
,
Conrad
,
M.
,
Toorchen
,
D.
&
Topal
,
M. D.
(
1984
).
Mechanism of mutagenesis by O6-methylguanine
.
Nature, Lond.
308
,
201
203
.
Folger
,
K. R.
,
Thomas
,
K.
&
Capecchi
,
M. R.
(
1985
).
Efficient correction of mismatched bases in plasmid heteroduplexes injected into cultured mammalian cells
.
Molec. cell. Biol.
5
,
70
74
.
Hare
,
J. T.
&
Taylor
,
J. H.
(
1985
).
One role for DNA methylation in vertebrate cells is strand discrimination in mismatch repair
.
Proc. natn. Acad. Sci. U.S.A.
82
,
7350
7354
.
Harm
,
W.
(
1980
).
Biological Effects of Ultraviolet Irradiation. Cambridge, New York: Cambridge University Press
.
Hayward
,
I. P.
&
Parsons
,
P. G.
(
1984
).
Comparison of virus reactivation, DNA base damage, and cell cycle effects in autologous human melanoma cells resistant to methylating agents
.
Cancer Res.
44
,
55
58
.
Howard-Flanders
,
P.
&
Boyce
,
R. P.
(
1966
).
DNA repair and genetic recombination: studies on mutants of Escherichia coli defective in these processes
.
Radiat. Res. Suppl.
8
,
156
184
.
Jeggo
,
P.
,
Défais
,
M.
,
Samson
,
L.
&
Schendel
,
P.
(
1977
).
An adaptive response of E. coli to low levels of alkylating agent: comparison with previously characterized DNA repair pathways
.
Molec. gen. Genet.
157
,
1
9
.
Jeggo
,
P.
,
Défais
,
M.
,
Sansom
,
L.
&
Schendel
,
P.
(
1978
).
The adaptive response of E. coli to low levels of alkylating agent: the role of polA in killing adaptation
.
Molec. gen. Genet.
162
,
299
305
.
Kann
,
H. E.
,
Kohn
,
K. W.
&
Lyles
,
J. M.
(
1974
).
Inhibition of DNA repair by the 1,3-bis(2- chloroethyl)-l-nitrosourea breakdown product, 2-chloroethylisocyanate
.
Cancer Res.
34
,
398
402
.
Karran
,
P.
(
1985
).
Possible depletion of a DNA repair enzyme in human lymphoma cells by subversive repair
.
Proc. natn. Acad. Sci. U.S.A.
82
,
5285
5289
.
Karran
,
P.
&
Marinus
,
M. G.
(
1982
).
Mismatch correction at O6-methylguanine residues in E. coli DNA
.
Nature, Lond.
296
,
868
869
.
Karran
,
P.
,
Stevens
,
S.
&
Sedgwick
,
B.
(
1982
).
The adaptive response to alkylating agents: the removal of O6-methylguanine from DNA is not dependent on DNA polymerase I
.
Mutat. Res.
104
,
67
73
.
Karran
,
P.
&
Williams
,
S. A.
(
1985
).
The cytotoxic and mutagenic effects of alkylating agents on human lymphoid cells are caused by different DNA lesions
.
Carcinogenesis
6
,
789
—792.
Kohn
,
K. W.
(
1977
).
Interstrand cross-linking of DNA by l,3-bis-(2-chloroethyl)-l-nitrosourea and other l-(2-haloethyl)-l-nitrosoureas
.
Cancer Res.
37
,
1450
1454
.
Lacks
,
S.
,
Dunn
,
J. J.
&
Greenberg
,
B.
(
1982
).
Identification of base mismatches recognized by the heteroduplex-DNA-repair system of S
.
pneumoniae. Cell
31
,
327
336
.
Lai
,
C.
&
Nathans
,
D.
(
1975
).
A map of temperature-sensitive mutants of simian virus 40
.
Virology
66
,
70
81
.
Larson
,
K.
,
Sahm
,
J.
,
Shenkar
,
R.
&
Strauss
,
B.
(
1985
).
Methylation-induced blocks to in vitroDNA replication
.
Mutât. Res.
150
,
77
84
.
Lawley
,
P. D.
(
1976
).
Carcinogenesis by alkylating agents
. In
Chemical Carcinogens, ACS Monograph 173
(ed.
C. E.
Searle
), pp.
83
-
244
. Washington, DC: Am. Chem. Soc.
Lawley
,
P. D.
,
Orr
,
D. J.
&
Jarman
,
M.
(
1975
).
Isolation and identification of products from alkylation of nucleic acids: ethyl- and isopropyl-purines
.
Biochem.J.
145
,
73
84
.
Lawley
,
P. C.
&
Thatcher
,
C. J.
(
1970
).
Methylation of deoxyribonucleic acid in cultured mammalian cells by N-methyl-N’-nitro-N-nitrosoguanidine
.
Biochem. J.
116
,
693
707
.
Lefevre
,
J. C.
,
Gasc
,
A. M.
,
Burger
,
A. C.
,
Mostachfi
,
P.
&
Sicard
,
M.
(
1984
).
Hyperrecombination at a specific DNA sequence in pneumococcal transformation
.
Proc. natn. Acad. Sci. USA.
81
,
5184
5188
.
Lieb
,
M.
(
1985
).
Recombination in the lambda repressor gene: evidence that very short patch (VSP) mismatch correction restores a specific sequence
.
Molec. gen. Genet.
199
,
465
470
.
Loveless
,
A.
(
1969
).
Possible relevance of 0-6 alkylation of deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides
.
Nature, Lond.
223
,
206
207
.
Miller
,
L.
,
Cooke
,
B.
&
Fried
,
M.
(
1976
).
Fate of mismatched base-pair regions in polyoma heteroduplex RNA during infection of mouse cells
.
Proc. natn. Acad. Sci. U.S.A.
73
,
3073
3077
.
Samson
,
L.
&
CAIRNS
,
J.
(
1977
).
A new pathway for RNA repair ‘mEscherichia coli. Nature, Lond.
267
,
281
282
.
Sasaki
,
M. S.
(
1977
). Role of DNA repair in the susceptibility to chromosome breakage and cell killing in cultured human fibroblasts. In
Biochemistry of Cutaneous Epidermal Differentiation
(ed. M. Seiji &
I. A.
Bernstein
), pp.
167
180
.
Tokyo
:
University of Tokyo Press
.
Scudiero
,
D. A.
(
1980
).
Decreased repair synthesis and colony forming ability of ataxia telangiectasia fibroblast cell strains treated with N-methyl-N’-nitro-N-nitrosoguanidine
.
Cancer Res.
40
,
984
990
.
Scudiero
,
D. A.
,
Meyer
,
S. A.
,
Clatterbuck
,
B. E.
,
Mattern
,
M. R.
,
Ziolkowski
,
C. H. J.
& Day,
R. S.
Ill
(
1984a
).
Relationship of DNA repair phenotypes of human fibroblast and tumor strains to killing by N-methyl-N’-nitro-N-nitrosoguanidine
.
Cancer Res.
44
,
961
969
.
Scudiero
,
D. A.
,
Meyer
,
S. A.
,
Clatterbuck
,
B. E.
,
Mattern
,
M. R.
,
Ziolkowski
,
C. H. J.
& Day,
R. S.
Ill
(
1984b
).
Sensitivity of human cell strains having different abilities to repair O6- methylguanine in DNA to inactivation by alkylating agents including chloroethylnitrosoureas
.
Cancer Res.
44
,
2467
2474
.
Scudiero
,
D. A.
,
Meyer
,
S. A.
,
Clatterbuck
,
B. E.
,
Tarone
,
R. E.
&
Robbins
,
J. H.
(
1981
).
Hypersensitivity to N-methyl-N’-nitro-N-nitrosoguanidine in fibroblasts from patients with Huntington disease, familial dysautonomia, and other primary neurological degenerations
.
Proc, natn. Acad. Sci. U.SA.
78
,
6451
6455
.
Sedgwick
,
B.
&
Robins
,
P.
(
1980
).
Isolation of mutants of Escherichia coli with increased resistance to alkylating agents: mutants deficient in thiols and mutants constitutive for the adaptive response
.
Molec. gen. Genet.
180
,
85
90
.
Sklar
,
R.
&
Strauss
,
B.
(
1980
).
Role of the uvrE gene product and of inducible O6-methylguanine removal in the induction of mutations by N-methyl- N ′ -nitro- N -nitrosoguanidine
.
J. molec. Biol.
143
,
343
362
.
Sklar
,
R.
&
Strauss
,
B.
(
1981
).
Removal of O6-methylguanine from DNA of normal and xeroderma pigmentosum-derived lymphoblastoid lines
.
Nature, Lond.
289
,
417
419
.
Snow
,
E. T.
,
Foote
,
R. S.
&
Mitra
,
S.
(
1984
).
Base-pairing properties of 06-methylguanine in template DNA during in vitro DNA replication
.
J’, biol. Chem.
259
,
8095
8100
.
Swenson
,
D. H.
&
Lawley
,
P. D.
(
1978
).
Alkylation of deoxyribonucleic acid by carcinogens dimethyl sulfate, ethyl methanesulfonate, N-ethyl-N-nitrosourea and N-methyl-N-nitrosourea
.
Biochem.J.
171
,
575
587
.
Tong
,
W. P.
,
Kirk
,
M. C.
&
Ludlum
,
D. B.
(
1982
).
Formation of the crosslink 1-[N3- deoxycytidyl]. 2-[N1-deoxyguanosinyljethane in DNA treated with N,N ′ -bis(2-chloroethyl)-N- nitrosourea
.
Cancer Res.
42
,
3102
3105
.
WAGNER
,
R.
Jr
&
Meselson
,
M.
(
1976
).
Repair tracts in mismatched DNA heteroduplexes
.
Proc, natn. Acad. Sci. U.SA.
73
,
4135
4139
.
Wake
,
C. T.
&
Wilson
,
J. H.
(
1980
).
Defined oligomeric SV40 DNA: a sensitive probe of general recombination in somatic cells
.
Cell
21
,
141
148
.
Warren
,
W.
&
Lawley
,
P. D.
(
1980
).
The removal of alkylation products from the DNA of Escherichia coli cells treated with the carcinogens N-ethyl-N-nitrosourea and N-methyl-N- nitrosourea: influence of growth conditions and DNA repair defects
.
Carcinogenesis
1
,
24
78
.
Watatani
,
M.
,
Ikenaga
,
M.
,
Hatanaka
,
T.
,
Kinuta
,
M.
,
Takai
,
S.
,
Mori
,
T.
&
Kondo
,
S.
(
1985
).
Analysis of N-methyl-N’-nitro-N-nitrosoguanidine (MNNG)-induced DNA damage in tumor cells strains from Japanese patients and demonstration of MNNG hypersensitivity of Mer xenografts in athymie nude mice
.
Carcinogenesis
6
,
549
553
.
Wolff
,
S.
,
Rodin
,
B.
&
Cleaver
,
J. E.
(
1977
).
Sister chromatid exchanges induced by mutagenic carcinogens in normal and xeroderma pigmentosum cells
.
Nature, Lond.
265
,
347
349
.
Yarosh
,
D. B.
,
Foote
,
R. S.
,
Mitra
,
S.
& Day,
R. S.
III
(
1983
).
Repair of O6-methylguanine in DNA by demethylation is lacking in Mer human tumor cell strains
.
Carcinogenesis
4
,
199
205
.
Yarosh
,
D. B.
,
Fornace
,
A.
& Day,
R. S.
Ill
(
1985
).
Repair of O4-methylthymine and methyl phosphotriesters in DNA by human O6-alkylguanine alkyltransferase
.
Carcinogenesis
6
,
949
953
.
Yarosh
,
D. B.
,
Rice
,
M.
, Day,
R. S.
Ill
,
Foote
,
R. S.
&
Mitra
,
S.
(
1984
).
O6-methylguanine- DNA methyltransferase in human cells
.
Mutât. Res.
131
,
27
36
.