Alkylating agents can produce a variety of biological effects in mammalian cells and organisms including toxicity, mutagenicity and malignant transformation. These agents react with oxygen and nitrogen atoms in DNA resulting in 12 products some of which are known to be eliminated from DNA by repair systems. One method of assessing the relative importance of a specific product in any of the biological effects of DNA alkylation would be to convert a cell line that is deficient in a particular repair function into a repair-proficient cell line and to determine whether this influences the magnitude of the effect. The cloning and expression in mammalian cells of the Escherichia coli DNA repair gene coding for the O1-alkylguanine-alkylphosphotriester dual alkyltransferase will be described. The E. coli gene product acts on damage produced in host cell DNA by treatment with methylnitrosourea, and reduces the toxicity and mutagenicity of this agent. The effects on the toxicity of a variety of other mono and bifunctional alkylating agents have also been assessed.

The mechanism by which alkylating agents induce tumours in experimental animals has been the subject of much research since the first demonstration of liver tumour induction in rats by dimethylnitrosamine (DMN; Magee & Barnes, 1956), which was later shown to alkylate cellular molecules such as protein (Magee & Hultin, 1962) and DNA (Magee & Farber, 1962). The alkylating agents are now known to be a large family of chemical compounds that includes N-nitrosamines such as DMN and also nitrosamides, sulphates, sulphonates, triazenes and others, and they can produce a wide range of biological effects including not only malignant transformation but also mutagenicity, cytotoxicity, chromosome aberrations and teratogenicity (reviewed by Magee et al. 1976; Preussmann & Stewart, 1984).

The N-nitroso compounds may have particular relevance to human cancer since these agents are present in the environment and can be generated from ingested inactive precursors (see Bartsch & Montesano, 1984). It has long been known that human liver is capable of the appropriate metabolic activation processes from DMN that give rise to DNA alkylation products (Montesano & Magee, 1970) and these products have been detected in liver DNA after known human exposure to DMN Herron & Shank, 1980).

Most of the nitrogen and oxygen atoms in DNA are potential sites for reaction of alkylating agents but the possible relevance of most of these has essentially been ignored since 1969 when it was first suggested that O1-alkylation of guanine might be responsible for the mutagenic and possibly carcinogenic effects of alkylating agents (Loveless, 1969). The interest in this product was polarized by several factors: (1) the observation that the extent of formation of the major base alkylation product 7-methylguanine (7MeG) did not correlate with the carcinogenicity of a series of methylating agents (Swann & Magee, 1968). (2) The demonstration that O1-methylguanine (O1MeG) miscoded when present in oligonucleotides (Gerchman & Ludlum, 1973). (3) Another major alkylation product, 3-alkyladenine was unstable in DNA in vivo and in vitro, and its biological relevance was difficult to investigate (Margison & O’Connor, 1973). (4) Alkylphosphotriesters, though being the major product with some ethylating agents and a minor product with the methylating agents, were relatively difficult to quantify (Warren et al. 1979; O’Connor et al. 1975) and also more difficult to envisage as giving rise to mutagenic events. (5) At that time most of the other products, principally pyrimidine adducts in DNA, were present in such small amounts as to be unmeasurable for practical purposes (e.g. see Lawley, 1972).

The probable adverse biological effects of O1-alkylguanine (O1-AG) were also suggested by the early observation that Escherichia coli (Lawley & Orr, 1970) and rat liver (O’Connor et al. 1973) contained systems for the repair of such damage in DNA. The importance of this repair process in reducing the susceptibility of a tissue to tumour induction was demonstrated in numerous experiments in which the effectiveness of the process was compared in target and non-target tissues and organisms (reviewed by O’Connor et al. 1979).

Work with bacteria and cultured mammalian cells strongly implicated the mutagenic and in some cases toxic effects of O1-AG in DNA (reviewed by Yarosh, 1985; Saffhill et al. 1985) and, more recently, evidence that O1MeG is responsible for cell transformation by methylating agents has been produced (Doniger et al. 1985).

In order to examine further the role of O1-AG in the various biological effects of alkylating agents, we have cloned the E. coli gene that codes for this repair function and produced vectors that permit its expression in cultured mammalian cells. Initially, a transient expression vector was used, but later a retrovirus-based drug- selectable plasmid was employed. Clones of repair-proficient cells have been generated and, by comparing their susceptibility to the toxic and mutagenic effects of alkylating agents with those of the parent line, the contribution of O1-AG to these effects can be deduced.

The repair of O1-AG in DNA in E. coli takes place by the transfer of the alkyl group to a cysteine residue within the 18K (K represents 101Mr) repair protein itself (Olsson & Lindahl, 1980; Demple et al. 1982). This is probably true in mammalian cells, although in this case the protein has not been purified to homogeneity (e.g. see Bogden et al. 1981). Exposure of E. coli to low doses of alkylating agents increases their resistance to the toxic and mutagenic effects of these agents and this, the ‘adaptive’ response (Samson & Cairns, 1977), was found to be under the control of a gene, ada, which was cloned by mutant rescue (Sedgwick, 1983). Since the relationship between the ada gene and E. coli O1-AG AT (alkyltransferase) activity had not then been shown (Teo et al. 1984), we decided to attempt the cloning of the AT gene using a functional assay for the gene product that we had devised for the rapid assay of activity in mammalian cell extracts.

An E. coli genomic DNA library was made by partial digestion with SawIII Al, ligation into the BamHI site of pUC8 and transformation into E coli JM83. Colonies harbouring plasmids containing inserts were selected and pools of eight colonies were grown up. Extracts of these were screened for AT activity as described (Margison et al. 1985) using JM83 as a negative control. From two slightly positive pools three highly positive colonies were identified and plasmids from these were transformed into E. coli DH1 to avoid recombinational loss. The two resulting colonies were called 061 and 062. By comparing the stoichiometry of the transfer of methyl groups from O1MeG in the substrate to the protein it was concluded that the former expressed only O1-AG AT activity. Transfer of methyl groups to protein in 062 extracts was higher than could be attributed to O1MeG alone and we concluded that these extracts contained an additional AT function (Fig. 1). This was shown to act on both methyl and ethyl phosphotriesters (Margison el al. 1985), and was probably the function identified earlier in extracts of adapted E. coli (McCarthy et al. 1983). It has recently been shown that this function acts only on the S stereoisomer of methylphosphotriesters (MeP) (McCarthy & Lindahl, 1985).

Fig. 1.

Methyltransferase activity in extracts of E. coli harbouring various plasmids: (○ — ○) p061; (• — •) p062; (▫ — ▫) P062SX; (▾ — ▾) p062H; (▵ — ▵) P062C (see Fig. 3B). Extracts were prepared and assayed as described by Margison et al. (1985) except that the substrate was prepared using [1H]methylnitrosourea (23 Ci mmol l−1) and an amount of this corresponding to 25 fmol of O1MeG was used.

Fig. 1.

Methyltransferase activity in extracts of E. coli harbouring various plasmids: (○ — ○) p061; (• — •) p062; (▫ — ▫) P062SX; (▾ — ▾) p062H; (▵ — ▵) P062C (see Fig. 3B). Extracts were prepared and assayed as described by Margison et al. (1985) except that the substrate was prepared using [1H]methylnitrosourea (23 Ci mmol l−1) and an amount of this corresponding to 25 fmol of O1MeG was used.

We carried out further characterization of extracts of 062 and its plasmids p062 and various subclones thereof (Fig. 1), and concluded that we had isolated a gene from E. coli that codes for a 37K protein containing both O1-AG (3′ region) and alkylphosphotriester (AP) (5′ region) AT activities (Fig. 2). This was later shown to be identical to the earlier cloned ada gene the sequence of which has since been published (Demple et al. 1985; Nakabeppu et al. 1985).

Fig. 2.

Arrangement of AP and O1-AG (06) AT activities in the p062HX subclone. Hatched area indicates approximate position of protein coding sequence. H, HindIII; S, SalI-,X,XhoI.

Fig. 2.

Arrangement of AP and O1-AG (06) AT activities in the p062HX subclone. Hatched area indicates approximate position of protein coding sequence. H, HindIII; S, SalI-,X,XhoI.

Initially, we thought that p061 contained the O1-AG AT section of this dual function gene (SX fragment, see Fig. 2). However, a limited restriction endonuclease map showed no common sites and Southern analysis has shown no sequence homology. Other evidence (P. M. Potter, J. Brennand & G. P. Margison, unpublished results) also suggests that this is a separate gene and sequence analysis is in progress to establish if this is the case.

EXPRESSION OF THE AT GENE IN COS7 MONKEY FIBROBLASTS

In order to show whether the bacterial gene could be expressed in mammalian cells we initially chose a simian virus 40 (SV40)-based transient expression vector, which had been used for the expression of the rabbit β-globin gene in COS7 monkey fibroblasts (Mulligan et al. 1979). The β-globin sequences were removed and a section of p062 containing the E. coli dual function AT gene was ligated into the gap to produce pSV206 as shown in Fig. 3A. Since this E. coli sequence contained an ATG upstream from the known dual AT protein initiation ATG, in a separate series of experiments (Fig. 3B) this was removed and the resulting fragment (062C) was used as above to generate pSV206C (Fig. 3A). COS7 monkey fibroblasts express an endogenous AT gene, which codes for a 24K protein acting on O1MeG in substrate DNA. In pSV2/3g-transfected cells the AT specific activity was 140 fmol mg−1 but in pSV206- and pSV206C-transfected cells it was increased to 460 and 810 fmol mg−1, respectively. The increase varied between transfections and was time-dependent (Brennand & Margison, 1986a). It was essential to show that this increase was not due to upregulation of the endogenous gene, and this was achieved by preparing extracts of pSV2βg- and pSV206-transfected cells, and by sodium dodecyl sulphate— polyacrylamide gel electrophoresis (SDS-PAGE)-fluorography following incubation with substrate DNA. Scanning densitometry of the fluorographs shows that the pSV206-transfected population contained, in addition to the endogenous 24K protein, 37K and 18K proteins which correspond to the F. coli dual function and O1- AG AT function proteins, respectively (Fig. 4).

Fig. 3.

Manipulations of p062HX to generate p062C, and the construction of pSV206 and pSV206C from p062HX or p062C and pSV2βg. A, AluI; B, BamHI; Bg, BglII; H, HindIII; R, EcoRI; S, SmaI.

Fig. 3.

Manipulations of p062HX to generate p062C, and the construction of pSV206 and pSV206C from p062HX or p062C and pSV2βg. A, AluI; B, BamHI; Bg, BglII; H, HindIII; R, EcoRI; S, SmaI.

Fig. 4.

Scanning densitométrie traces of fluorographs of extracts of COS7 cells harbouring pSV2βg (top) or pSV206 (centre and bottom, representing two different experiments) following incubation with substrate DNA and SDS—PAGE (see Margison et al. 1985, for details). The peak around 4·5 cm corresponds to an Mr of around 37K. Those at 7 and 9 cm correspond to approx. 24K and 18K, respectively.

Fig. 4.

Scanning densitométrie traces of fluorographs of extracts of COS7 cells harbouring pSV2βg (top) or pSV206 (centre and bottom, representing two different experiments) following incubation with substrate DNA and SDS—PAGE (see Margison et al. 1985, for details). The peak around 4·5 cm corresponds to an Mr of around 37K. Those at 7 and 9 cm correspond to approx. 24K and 18K, respectively.

In E. coli we found that the extent to which the 37K protein was broken down to subfragments was host-dependent (Margison et al. 1985). In further experiments with pSV206-transfected COS7 cells we found that a similar effect could take place giving rise in extreme cases to 18K and 16K bands exclusively (Fig. 4). Whether this degradation occurs before (giving rise to active subfragments) or after reaction with substrate DNA has not been elucidated.

Although we could have used these pSV206C-transfected cells to study the repair of AP, the presence of the endogenous O1-AG AT activity and the relatively small increase in specific activity of the combined function in transfected cells obviated investigation of any changes in the biological effects of alkylating agents in this system. However, the feasibility of the approach had been adequately demonstrated and we turned to vectors with a less restricted host range so that we could achieve E. coli AT gene expression in AT-deficient cell lines.

The plasmid pJCB06C was constructed as shown in Fig. 5 and transfected into O1-AG AT-deficient V79 Chinese hamster fibroblasts (Warren et al. 1979; Brennand & Margison, 1986a). Cells incorporating this plasmid into their genome should be capable of producing two mRNA molecules (Fig. 5). One is a full-length mRNA starting transcription in the 5′LTR (long terminal repeat) and terminating in the 3′LTR, which, by initiation at the first AUG encountered, directs E. coli AT translation. The other is a spliced version of this, which does not contain the E. coli sequence and which directs translation from the initiation AUG of the neo gene, the product of which confers resistance to the antibiotic G418.

Fig. 5.

Construction from p062C and pZipneoSV(X)l and putative transcription in mammalian cells of pJCB06C.

Fig. 5.

Construction from p062C and pZipneoSV(X)l and putative transcription in mammalian cells of pJCB06C.

Of the extracts of the 12 clones of G418-resistant pJCB06C-transfected cells that were assayed, only four expressed high levels of AT activity (see Fig. 6A) and when corrected for protein concentration, extracts of clone 8 had the highest level (Fig. 6B, l600 fmol mg−1 AT protein extracted). Transfection with control plasmid pZipneoSV(X) 1 did not increase endogenous gene expression (Fig. 6B) and one clone (clone 2, AT sp. act. 4 fmol mg−1 protein) was used as the control in subsequent experiments. Southern analysis of DNA from clone 2 and clone 8 cells using 06C as a probe showed that the E. coli AT gene had been incorporated into the genome of clone 8 cells (data not shown).

Fig. 6.

A. Methyltransferase (MT) activity in clones of G418-resistant V79 Chinese hamster cells following transfection with pJCB06C (3–12) or pZipneo SV(X)1 (1 and 2). B. Substrate-limiting assay for methyltransferase activity in extracts of E. coli harbouring p062SX (○ – – – ○) or p062HX (• – – – •), or V79 cells harbouring pZipneoSV(X)l (○ — ○) or pJCB06C (• — •).

Fig. 6.

A. Methyltransferase (MT) activity in clones of G418-resistant V79 Chinese hamster cells following transfection with pJCB06C (3–12) or pZipneo SV(X)1 (1 and 2). B. Substrate-limiting assay for methyltransferase activity in extracts of E. coli harbouring p062SX (○ – – – ○) or p062HX (• – – – •), or V79 cells harbouring pZipneoSV(X)l (○ — ○) or pJCB06C (• — •).

Substrate-limiting experiments showed that extracts of pJCB06C-transfected cells contained the same AT activities as extracts of E. coli harbouring the dual-function gene plasmid p062HX (Fig. 6B). In most experiments this protein had a molecular weight of approx. 37K as shown by PAGE and fluorography (Fig. 7, lane A). In some cases, however, there was evidence for degradation to 18K and 20K proteins (Fig. 7, lanes B, C) as had been found in COS7 extracts (see above). Again, whether this breakdown occurs before or after incubation with substrate DNA is not known.

Fig. 7.

Fluorographs of extracts of pJCB06C-transfected G418-resistant V79 cells following incubation with substrate DNA and SDS-PAGE. Lane A, a sample undergoing little degradation; lane B, bands around 18K prominent, but 37K material still present; lane C, no 37K material detectable.

Fig. 7.

Fluorographs of extracts of pJCB06C-transfected G418-resistant V79 cells following incubation with substrate DNA and SDS-PAGE. Lane A, a sample undergoing little degradation; lane B, bands around 18K prominent, but 37K material still present; lane C, no 37K material detectable.

Although the E. coli AT gene was clearly present in clone 8 cells and was being expressed in an active form, it was necessary to show that the protein was capable of penetrating the nucleus and acting on alkylation damage in host cell DNA. Clone 2 and clone 8 cells were exposed to radiolabelled N-methyl-N-nitrosourea (MNU) for 1 h and harvested up to 22 h later. DNA was extracted and the amounts of the major product 7MeG and one of the substrates for the dual-function protein, O1MeG, were determined by radiochromatography of acid hydrolysates of the DNA. Fig. 8 shows that the amounts detected immediately after treatment and the loss of 7MeG were identical in both clones, whereas the amounts of O1MeG were much lower in clone 8 cells. This indicated that the E. coli gene product was capable of repairing host cell DNA damage very rapidly since even by the end of 1-h treatment with MNU (the ‘zero’ time) the O1MeG level in clone 8 cells was only about 20 % of that in clone 2 cells.

Fig. 8.

Levels of 7-methylguanine (7MeG, left) and O1-methylguanine (O1MeG, right) in DNA of clone 8 (• — •) or clone 2 (○ — ○) cells various times after exposure to 1H-labelled N-methyl-.N-nitrosourea.

Fig. 8.

Levels of 7-methylguanine (7MeG, left) and O1-methylguanine (O1MeG, right) in DNA of clone 8 (• — •) or clone 2 (○ — ○) cells various times after exposure to 1H-labelled N-methyl-.N-nitrosourea.

Exposure of clone 2 cells to increasing doses of MNU resulted in the generation of thioguanine(TG)-resistant cells with a frequency of about 1 in 101 at a survival rate of about 35 % (see Fig. 9). In contrast, in clone 8 cells there was no increase in TG- resistant clones above the spontaneous level of 0·5×101 to 1·0 × 101. E. coli AT gene expression therefore protects these cells against the mutagenic effect of MNU.

Fig. 9.

Frequency of HPRT mutants (thioguanine-resistant cells) in relation to the % of cells surviving after exposure to increasing doses of N-methyl-N-nitrosourea; (○ — ○) clone 2; (• — •) clone 8.

Fig. 9.

Frequency of HPRT mutants (thioguanine-resistant cells) in relation to the % of cells surviving after exposure to increasing doses of N-methyl-N-nitrosourea; (○ — ○) clone 2; (• — •) clone 8.

The survival curves of clone 2 and clone 8 cells after exposure to increasing doses of MNU or methylmethanesulphonate (MMS) are shown in Fig. 10. Clone 8 cells are more resistant than clone 2 cells to the toxic effects of MNU and this suggests that E. coll AT-repairable DNA damage is responsible for part of this effect. Further work is needed to decide if the toxicity in clone 8 cells is a consequence of saturation of the repair function or the toxic effects of another lesion. This latter possibility appears to be more likely since both clones of cells are (equally) susceptible to killing by MMS, an agent that produces only a very small proportion of its total DNA damage in the form of AT-repairable lesions.

Fig. 10.

Survival of clone 2 (○ — ○) or clone 8 (• — •) cells following exposure to increasing doses of MNU (A) or MMS (B).

Fig. 10.

Survival of clone 2 (○ — ○) or clone 8 (• — •) cells following exposure to increasing doses of MNU (A) or MMS (B).

Clone 8 cells were much less susceptible than clone 2 cells to the toxic effects of four chloroethylating agents, bis-chloroethylnitrosourea (BiCNU), taurine chloroethylnitrosourea (TCNU), chlorozotocin and nitrozolamide (Fig. 11). It has been suggested that these kinds of agents are toxic to cells because of their ability to form crosslinks that occur slowly via an initial mono addition to guanine to form O1- chloroethylguanine (Tong et al. 1982). Circumstantial evidence of this is that the mono adduct can be dealkylated by the AT and this prevents crosslinking (Robins et al. 1983), and that cells that contain low levels of this AT protein either normally (Erickson et al. 1980), or after depletion by low doses of methylating agents (Zlotogorski & Erickson, 1983), are more susceptible to the toxic effects of these agents than cells containing high levels. Our results suggest that E. coli AT- repairable damage is almost exclusively responsible for cell killing by these agents, with the possible exception of BiCNU, which though much more toxic for clone 2 cells did show extensive killing of clone 8 cells (Fig. 11). The higher toxicity of BiCNU may be related to its carbamoylating activity.

Fig. 11.

Survival of clone 2 (○ — ○) or clone 8 (• — •) cells following exposure to increasing doses of bis-chloroethylnitrosourea (BiCNU, top left), taurine chloroethylnitrosourea (TCNU, top right), chlorozotocin (bottom left) or nitrozolamide (bottom right).

Fig. 11.

Survival of clone 2 (○ — ○) or clone 8 (• — •) cells following exposure to increasing doses of bis-chloroethylnitrosourea (BiCNU, top left), taurine chloroethylnitrosourea (TCNU, top right), chlorozotocin (bottom left) or nitrozolamide (bottom right).

Both clones were equally susceptible to the toxic effects of nitrogen mustard (Brennand & Margison, 19866), which is thought to crosslink DNA via the N7 atoms of guanine residues (Kohn étal. 1966). This result further indicates that AT-repairable damage is not a cause of death in nigrogen-mustard-treated cells (see also Robinsei al. 1983; Gibson et al. 1985).

From these experiments we can conclude that damage in DNA that can be repaired by the E. coli dual AT is responsible for the mutagenic effects of MNU and to a greater or lesser extent responsible for the toxic effects of MNU and chloroethylating agents but not nitrogen mustard or MMS. This conclusion is possible because the two clones we have generated differ only in their capacity to repair oxygen atom alkylation products in DNA: in all other respects they are identical.

The E. coli AT expressed in clone 8 cells has been shown to act on O1-AG (see above) and this same repair function has also been shown to act on O1-methyl- thymine in DNA (McCarthy et al. 1984). Another function in the protein acts on the S stereoisomers of MeP (McCarthy & Lindahl, 1985). By generating clones that express only the former or latter function we can now indicate precisely which effects can be attributed to these lesions. Initial results (Brennand & Margison, 1986c) show that clones expressing only the O1-AG AT function have the same resistance to MNU as clone 8 cells, indicating that MeP S isomers are not toxic lesions. The same cells are, however, more susceptible to the toxic effects of chlorozotocin and nitrozolamide (but more resistant than clone 2 cells). There are several possible explanations for this observation, one of them being that chloroethylphosphotriesters may form lethal crosslinks if not repaired. Clones expressing only the AP AT function, which are now being generated, should help to resolve these questions.

Other biological effects of alkylating agents in these cells are being investigated. Initial results (White et al. 1986) indicate that expression of the E. coli O1-AG AT protects V79 cells against MNU-induced sister chromatid exchanges, chromatid damage and micronucleus formation, indicating a role for O1MeG in these effects. This approach is being extended to other classes of agents, and experiments intended to integrate E. coli DNA repair genes into cells and organisms that are susceptible to the transforming effects of alkylating agents are in hand.

We thank Sean Baker, Geoffrey Clarke, Martin Greaves, Brian C. Keenan, Yemisi Obisesan and Shrilene Oh for assistance in various aspects of this work, which was supported by grants from the Cancer Research Campaign. The saintly patience of Linda Evans deserves special mention! Mitozolamide was kindly provided by Dr E. Lunt of May and Baker Ltd.

Bartsch
,
H.
&
Montesano
,
R.
(
1984
).
Relevance of nitrosamines to human cancer
.
Carcinogenesis
5
,
1381
1393
.
Bogden
,
J. M.
,
Eastman
,
A.
&
Bresnick
,
E.
(
1981
).
A system in mouse liver for the repair of O6- methylguanine lesions in methylated DNA
.
Nucl. Acids Res.
9
,
3089
3103
.
Brennand
,
J.
&
Margison
,
G. P.
(
1986a
).
Expression of the E. coli O6-methylguanine- methylphosphotriester methyltransferase gene in mammalian cells
.
Carcinogenesis
7
,
185
188
.
Brennand
,
J.
&
Margison
,
G. P.
(
1986b
).
Reduction in the toxicity and mutagenicity of alkylating agents in mammalian cells harbouring the Escherichia coli alkyltransferase gene
.
Proc, natn. Acad. Sci. U.SA. (in press)
.
Brennand
,
J.
&
Margison
,
G. P.
(
1986c
).
Expression in mammalian cells of a truncated E coligene coding for O6-alkylguanine alkyltransferase reduces the toxic effects of alkylating agents
.
Carcinogenesis (in press)
.
Demple
,
B.
,
Jacobsson
,
A.
,
Olsson
,
M.
,
Robins
,
P.
&
Lindahl
,
T.
(
1982
).
The repair of alkylated DNA in Escherichia coli. Physical properties of O6-methylguanine-DNA methyltransferase
.
J. biol. Chem.
257
,
13 776
13 780
.
Demple
,
B.
,
Sedgwick
,
B.
,
Robins
,
P.
,
Totty
,
N.
,
Waterfield
,
M. D.
&
Lindahl
,
T.
(
1985
).
Active site and complete sequence of the suicidal methyltransferase that counters alkylation mutagenesis
.
Proc. natn. Acad. Sci. U.S.A.
82
,
2688
2692
.
Doniger
,
J.
,
Day
,
R. S.
&
DiPaolo
,
J. A.
(
1985
).
Quantitative assessment of the role of O6- methylguanine in the initiation of carcinogenesis by methylating agents
.
Proc. natn. Acad. Sci. U.S.A.
82
,
421
425
.
Erickson
,
L. C.
,
Bradley
,
M. O.
,
Ducore
,
J. M.
,
Ewig
,
R. A. G.
&
Kohn
,
K. W.
(
1980
).
DNA crosslinking and cytotoxicity in normal and transformed human cells treated with antitumour nitrosoureas
.
Proc. natn. Acad. Sci. U.S.A.
78
,
6766
6770
.
Gerchman
,
L. L.
&
Ludlum
,
D. B.
(
1973
).
The properties of O6-methylguanine in templates for RNA polymerase
.
Biochim. biophys. Acta
308
,
310
316
.
Gibson
,
N. W.
,
Zlotogorski
,
C.
&
Erickson
,
L. C.
(
1985
).
Specific DNA repair mechanisms may protect some human tumour cells from DNA interstrand crosslinking by chloroethylnitrosoureas but not from crosslinking by other antitumour alkylating agents
.
Carcinogenesis
6
,
445
450
.
Herron
,
D. C.
&
Shank
,
R.
(
1980
).
Methylated purines in human liver DNA after probable dimethylnitrosamine poisioning
.
Cancer Res.
40
,
3116
3117
.
Kohn
,
K. W.
,
Spears
,
C. L.
&
Doty
,
P.
(
1966
).
Interstrand crosslinking of DNA by nitrogen mustard
.
J. molec. Biol.
19
,
266
273
.
Lawley
,
P. D.
(
1972
). The action of alkylating agents on nucleic acids: N-methyl-N-nitroso compounds as methylating agents. In
Topics in Chemical Carcinogenesis
(ed.
W. N.
Kakahara
et al. ), pp.
237
256
.
Tokyo
:
University of Tokyo Press
.
Lawley
,
P. D.
&
Orr
,
D. J.
(
1970
).
Specific excision of methylation products from DNA of Escherichia coli treated with A’-methyl-A’’-nitro-A’-nitrosoguanidine
.
Chem.-Biol. Interact.
2
,
154
.
Loveless
,
A.
(
1969
).
Possible relevance of O6-alkylation of deoxyguanosine to the mutagenicity and carcinogenicity of nitrosamines and nitrosamides
.
Nature, Lond.
223
,
206
207
.
Magee
,
P. N.
&
Barnes
,
J. M.
(
1956
).
The production of malignant primary hepatic tumours in the rat by feeding dimethylnitrosamine
.
Br.J. Cancer
10
,
114
122
.
Magee
,
P. N.
&
Farber
,
E.
(
1962
).
Toxic liver injury and carcinogenesis. Methylation of rat liver nucleic acids by dimethylnitrosamine
.
Biochem. J.
83
,
114
124
.
Magee
,
P. N.
&
Hultin
,
T.
(
1962
).
Toxic liver injury and carcinogenesis. Methylation of proteins of rat liver slices by dimethylnitrosamine in vitro.
Biochem. J.
83
,
106
114
.
Magee
,
P. M.
,
Montesano
,
R.
&
Preussmann
,
R.
(
1976
).
N-nitroso compounds and related carcinogens
.
Chemical Carcinogens. ACS Monograph 173
(ed.
C. E.
Searle
), pp.
491
-
625
. Washington, DC: Am. Chem. Soc.
Margison
,
G. P.
,
Cooper
,
D. P.
&
Brennand
,
J.
(
1985
).
Cloning of the E. colt O6-methylguanine and methylphosphotriester methyltransferase gene using a functional DNA repair assay
.
Nucl. Acids Res.
13
,
1939
1952
.
Margison
,
G. P.
&
O’Connor
,
P. J.
(
1973
).
Biological implications of the instability of the N-glycosidic bond of 3-methyldeoxyadenosine in DNA
.
Biochim. biophys. Acta
331
,
349
—356.
McCarthy
,
J. G.
,
Edington
,
B. V.
&
Schendel
,
P. F.
(
1983
).
Inducible repair of phosphotriesters in Escherichia coli.
Proc. natn. Acad. Sci. U.S.A.
80
,
7380
—7384.
McCarthy
,
T. V.
,
Karran
,
P.
&
Lindahl
,
T.
(
1984
).
Inducible repair of O-alkylated DNA pyrimidines in Escherichia coli.
EMBOJ.
3
,
545
550
.
McCarthy
,
T. V.
&
Lindahl
,
T.
(
1985
).
Methylphosphotriesters in alkylated DNA one repaired by the ada regulatory protein of E. coli. Nucl. Acids Res.
13
,
2683
2698
.
Montesano
,
R.
&
Magee
,
P. N.
(
1970
).
Metabolism of dimethylnitrosamine by human liver slices in vitro.
Nature, Lond.
228
,
173
174
..
Mulligan
,
R. C.
,
Howard
,
B. H.
&
Berg
,
P.
(
1979
).
Synthesis of rabbit β-globin in cultured monkéy kidney cells following infection with a SV40- β-globin recombinant genome
.
Nature, Lond.
277
,
108
114
.
Nakabeppu
,
Y.
,
Kondo
,
H.
,
Kawabata
,
S.
,
Iwanaga
,
S.
&
Sekiguchi
,
M.
(
1985
).
Purification and structure of the intact ada regulatory protein of Escherichia coli K12, O6-methylguanine- DNA methyltransferase
.
J, biol. Chem.
260
,
7281
7288
.
O’Connor
,
P. J.
,
Capps
,
M. J.
&
Craig
,
A. W.
(
1973
).
Comparative studies of the hepatocarcinogen N,N-dimethylnitrosamine in vivo’, reaction sites in rat liver DNA and the significance of their relative stabilities
.
Br.J. Cancer 2.1, 153-166
.
O’Connor
,
P. J.
,
Margison
,
G. P.
&
Craig
,
A. W.
(
1975
).
Phosphotriesters in rat liver DNA after the administration of the carcinogen N,N-dimethylnitrosamine in vivo.
Biochem. J.
145
,
475
482
.
O’Connor
,
P. J.
,
Saffhill
,
R.
&
Margison
,
G. P.
(
1979
). N-nitroso compounds: biochemical mechanisms of action. In
Environmental Carcinogenesis
(ed. P. Emmelot &
E.
Kriek
), pp.
73
96
.
Amsterdam
:
Elsevier/North Holland Biomedical Press
.
Olsson
,
M.
&
Lindahl
,
T.
(
1980
).
Repair of alkylated DNA in E. coli:. methyl group transfer from O6-methylguanine to a protein cysteine residue
.
J. biol. Chem.
255
,
10569
10571
.
PREUSSMANN
,
R.
&
Stewart
,
B. W.
(
1984
).
N-nitroso compounds and related carcinogens
. In
Chemical Carcinogens
(ed. C. E. Searle), 2
nd
edn
, pp.
643
-
828
. Washington, DC: Am. Chem. Soc.
Robins
,
P.
,
Harris
,
A. L.
,
Goldsmith
,
I.
&
Lindahl
,
T.
(
1983
).
Cross-linking of DNA induced by chloroethylnitrosourea is prevented by O6-methylguanine-DNA methyltransferase
.
Nucl. Acids Res.
11
,
7743
7758
.
Saffhill
,
R.
,
Margison
,
G. P.
&
O’Connor
,
P. J.
(
1985
).
Mechanisms of carcinogenesis induced by alkylating agents
.
Biochim. biophys. Acta
823
,
111
145
.
Samson
,
L.
&
Cairns
,
J.
(
1977
).
A new pathway for DNA repair in Escherichia coli.
Nature, Lond.
267
,
281
283
.
SEDGWICK
,
B.
(
1983
).
Molecular cloning of a gene which regulates the adaptive response to alkylating agents in Escherichia coli.
Molec. gen. Genet.
191
,
466
472
.
Swann
,
P. F.
&
Magee
,
P. N.
(
1968
).
Nitrosamine induced carcinogenesis: The alkylation of nucleic acids of the rat by N-methylW-nitrosourea, dimethylnitrosamine, dimethylsulphate and methylmethanesulphonate
.
Biochem. J.
110
,
39
47
.
Teo
,
I.
,
Sedgwick
,
B.
,
Demple
,
B.
,
Li
,
B.
&
Lindahl
,
T.
(
1984
).
Induction of resistance to alkylating agents in E. coli: the ada+ gene product serves both as a regulatory protein and as an enzyme for repair for mutagenic damage
.
EMBOJ.
3
,
2151
2157
.
Tong
,
W. P.
,
Kohn
,
K. W.
&
Ludlum
,
D. B.
(
1982
).
Modifications of DNA by different haloethylnitrosoureas
.
Cancer Res.
42
,
4460
4464
.
Warren
,
W.
,
Crathorn
,
A. R.
&
Shooter
,
K. V.
(
1979
).
The stability of methylated purines and of methylphosphotriesters in the DNA of V79 cells after treatment with N-methyl-N- nitrosourea
.
Biochim. biophys. Acta
563
,
82
88
.
White
,
G. R. M.
,
Ockey
,
C. H.
,
Brennand
,
J.
&
Margison
,
G. P.
(
1986
).
Chinese hamster cells harbouring the E. coli O6-alkylguanine alkyltransferase gene are less susceptible to SCE induction and chromosome damage by methylating agents
.
Carcinogenesis (in press)
.
Yarosh
,
D. B.
(
1985
).
The role of O6-methylguanine DNA methyltransferase in cell survival, mutagenesis and carcinogenesis
.
Mutat. Res.
145
,
1
16
.
Zlotogorski
,
C.
&
Erickson
,
L. C.
(
1983
).
Pretreatment of normal human fibroblasts and human colon carcinoma cells with MNNG allows chloroethylnitrosourea to produce DNA interstrand crosslinks not observed in cells treated with chloroethylnitrosourea alone
.
Carcinogenesis
4
,
759
763
.