3-Methyladenine-DNA glycosylase activities have been identified in all eukaryotic cell systems studied. Some of the results from these studies are reviewed here. The enzymes possess molecular weights between 24 ×101 and 34 ×101, they have a broad pH optimum at approximately pH 8, require double-stranded DNA and act in the absence of any cofactors. The enzyme can excise several different methylated bases from DNA such as 3-methyladenine, 7-methylguanine and 3-methylguanine.

The specific activity of this DNA glycosylase in mouse L-cells was found to be a function of the proliferative state of the cell. In vitro quantification of this DNA repair activity in synchronized mouse L-cells suggests that it is regulated within a defined temporal sequence prior to the onset of DNA replication.

Using DNA fragments of defined sequences it was observed that the efficiency of removal of the methylated bases is sequence-dependent.

Following exposure to alkylating agents DNA is modified at many sites, in particular at the ring nitrogen and the exocyclic oxygen atoms of the bases. The oxygen atoms of the phosphate internucleotide linkages are also attacked by alkylating agents (Singer & Kusmierek, 1982). In both bacteria and higher organisms, the alkylated bases or phosphate groups may either be repaired by a direct transfer of the modifying alkyl group from DNA to the repair protein or by direct cleavage of the glycosylic bond of the alkylated base by a specific DNA glycosylase (for a review, see Friedberg, 1985). In Escherichia coli several DNA glycosylases have evolved to deal with specific lesions (Lindahl, 1982). Two distinct 3-methyladenine-DNA glycosylases (types I and II) have been purified from this organism. The genetics and biochemistry of these two DNA glycosylases have been thoroughly studied (Friedberg, 1985). While 3-methyladenine-DNA glycosylase I releases only 3- methyladenine from DNA, 3-methyladenine-DNA glycosylase II has a much wider substrate specificity (Thomas et al. 1982; Karranet al. 1982). One major problem in the study of similar DNA repair activities in eukaryotic cells is that such cells are poorly characterized genetically with regard to DNA repair function. However, a number of observations made in different laboratories suggest that an activity similar to the type II DNA glycosylase from E. coli exists in eukaryotic cells.

In order to elucidate the role of DNA repair in the maintenance of a biological system one has to identify the different enzymic activities responsible for the removal of damaged regions of the DNA. Furthermore, it is important to examine how these different enzymes are regulated throughout the cell cycle and during differentiation of the eukaryotic cell.

This article reviews studies on 3-methyladenine-DNA glycosylases in higher organisms and compares the properties of these enzymes with the well-characterized bacterial counterpart. Some recent data from our laboratory are presented concerning the influence of DNA sequence on the removal of methylated purines by the 3-methyladenine-DNA glycosylase purified from calf thymus, and the cell cycle-dependent regulation of the same activity in cultured mouse cells.

3-Methyladenine-DNA glycosylases have been purified from several different mammalian systems. Some of the properties and characteristics of the different enzymes isolated are listed in Table 1. For comparison, the properties of the two 3-methyladenine-DNA glycosylases present in E coli are also provided.

Only the enzyme isolated from calf thymus has been obtained in an apparent homogeneous form, as judged by SDS-polyacrylamide gel electrophoresis and silver staining (Male et al. 1985a).

All the reported enzyme activities are associated with proteins of rather low molecular weights of around 30 ×101. In calf thymus and mouse cells we have found that multiple molecular weight species are present. Only the low molecular weight form has been purified to homogeneity. The presence of multiple species indicates that mammalian cells either have several gene products for these repair enzymes, that the enzymes are processed in the cell or that the enzymes are degraded by proteases during the purification procedure. Since the smaller molecular weight species is the only one present in the nucleus in calf thymus cells we conclude that the larger molecular weight form found only in the cytoplasm is a precursor of the chromatin- associated enzyme (Male et al. 1985a). Experiments verifying this hypothesis are in progress in our laboratory.

A striking difference is observed between the pl for the calf thymus enzyme and the two bacterial enzymes. The rather high pl for the calf thymus enzyme indicates a stronger affinity for DNA by the mammalian enzyme as compared to the E. coli enzyme.

The two 3-methyladenine-DNA glycosylases in A. coli differ in specificity. While the 3-methyladenine-DNA glycosylase I only releases 3-methyladenine, the type II enzyme catalyses the release of 3-methyladenine, 3-methylguanine, 7-methyl- guanine, O1-methylcytosine and O1-methylthymine from alkylated DNA (Thomas et al. 1982; McCarthy et al. 1984). The eukaryotic 3-methyladenine-DNA glycosylases are similar to the E. coli type II DNA glycosylase in this respect (Table 1). Gallagher & Brent (1984) have reported evidence that these different activities are associated with a single DNA glycosylase in human placenta. Our data on the purified calf thymus enzyme confirm that the wide specificity observed for the other mammalian 3-methyladenine-DNA glycosylases is associated with one protein.

A general property of all known DNA glycosylases is that no divalent cations are required for the enzyme-catalysed release of the modified bases. The mammalian 3-methyladenine-DNA glycosylases seem to be dependent on a cysteine residue in the reduced form since p-hydroxymercuribenzoate totally abolishes the activity (Gallagher & Brent, 1984; Male et al. 1985a).

We have shown that the rate of release of 3-methyladenine is much faster than that for the other methylated bases. In the case of the enzyme from calf thymus the ratio of release of 3-methyladenine versus 7-methyladenine residues was approximately 10:1 when methylated calf thymus DNA was used as a substrate. This ratio is, however, dependent on the structure of the DNA (Male et al. 1986). The interpretation of these observations is that following alkylation of DNA in the cell some sites are less efficiently repaired than others by the 3-methyladenine-DNA glycosylase. In order to elucidate further the influence of different DNA structures and base sequences on the rate of removal of methylated bases from DNA, we have used as substrates DNA fragments of defined sequences.

A 186 base-pair (bp) Sall-Haell fragment from the plasmid pUC18 was labelled at the 5’ end by bacteriophage T4 polynucleotide kinase, followed by methylation according to the Maxam-Gilbert procedure (Maxam & Gilbert, 1980), and then subsequently treated with the 3-methyladenine-DNA glycosylase. To induce nicks at the depurinated sites the DNA was then incubated with the Mg2+-independent apurinic/apyrimidinic (AP)-endonuclease associated with the redoxy endonuclease from calf thymus (Doetsch et al. 1986). Determination of the sites of incision in the DNA fragment was carried out by analysis of the products on a 12% denaturing polyacrylamide sequencing gel alongside the Maxam-Gilbert sequencing reactions. To quantify the incision sites made by the enzyme, the autoradiograms of the sequencing gels were subjected to densitometer scanning. The results from one analysis are shown in Fig. 1. The efficiency of incision is indicated by the size of the histogram at a certain purine site. Some variations in the alkylation of DNA are known for the guanine-specific reaction. However, these variations are minor compared to the enzyme-induced nicking found in the DNA fragment. To minimize the effect of possible sequence preference by the redoxy endonuclease (Doetsch et al. 1986; Helland et al. 1986), a great excess of this enzyme was included in the assay mixtures to ensure that all AP sites generated by the action of the 3-methyladenine- DNA glycosylase were nicked.

Fig. 1.

DNA sequence preference of 3-methyladenine-DNA glycosylase. The graph represent the results of a densitometer scan of the autoradiogram of the 5′ end-labelled fragments generated from the 186bp SalI-HaeII fragment of the plasmid pUC18. The DNA had been methylated by dimethyl sulphate according to the A+G reaction in the Maxam-Gilbert sequencing protocol (Maxam & Gilbert, 1980) before being treated with the 3-methyladenine-DNA glycosylase (13 milli units) purified from calf thymus (Male et al. 1985a). Nicks in the DNA at the apurinic sites generated by the action of the glycosylase were introduced by adding an excess of the Mg2+-independent AP-endo- nuclease associated with the calf thymus redoxyendonuclease (Doetsch et al. 1986) to the reaction mixture. The bases are numbered from the 1P-labelled 5′ end of the restriction enzyme fragment.

Fig. 1.

DNA sequence preference of 3-methyladenine-DNA glycosylase. The graph represent the results of a densitometer scan of the autoradiogram of the 5′ end-labelled fragments generated from the 186bp SalI-HaeII fragment of the plasmid pUC18. The DNA had been methylated by dimethyl sulphate according to the A+G reaction in the Maxam-Gilbert sequencing protocol (Maxam & Gilbert, 1980) before being treated with the 3-methyladenine-DNA glycosylase (13 milli units) purified from calf thymus (Male et al. 1985a). Nicks in the DNA at the apurinic sites generated by the action of the glycosylase were introduced by adding an excess of the Mg2+-independent AP-endo- nuclease associated with the calf thymus redoxyendonuclease (Doetsch et al. 1986) to the reaction mixture. The bases are numbered from the 1P-labelled 5′ end of the restriction enzyme fragment.

Table 1.

Properties of mammalian 3-Methyladenine-DNA glycosylase

Properties of mammalian 3-Methyladenine-DNA glycosylase
Properties of mammalian 3-Methyladenine-DNA glycosylase

Some of the alkylated purine sites are apparently seldom or never nicked by the combined action of these enzymes. These purine residues are frequently found in a sequence of several pyrimidines; in particular, adenine residues 5′ to another adenine were rarely removed. The most efficient substrates appear to be a base residue with a purine next to it on the 5 ′ side, or bases situated within a sequence of several purines. Further experiments are required to determine the relationship between sequence specificity and the removal of methylated bases in vivo.

In addition to the differences in substrate specificity, the two 3-methyladenine- DNA glycosylase activities present in E. coli also differ in the way they are regulated. Thus the E. coli 3 -methyladenine-DNA glycosylase I is constitutively expressed, while the glycosylase II is inducible (Karran et al. 1982). The biological consequence of this phenomenon is that bacteria may respond to exposure to alkylating agents by increasing the DNA repair capacity for the damage caused by these agents (Friedberg, 1985).

Several studies with mammalian cells have shown that both replicative DNA synthesis and the activities of both uracil-DNA glycosylase and 3-methyladenine- DNA glycosylase are increased following exposure to DNA damaging agents. The increase, however, is always found to be linked to an enhanced proliferation of the cells (Sirover & Gupta, 1985).

In a previous study we reported the effect of exposure to subtoxic doses of alkylating agents on C3H/10T1/2 cells and mouse L-cells (Male et al. 1985b). No significant differences between treated and control cells were observed when the survival of the cells was measured as a function of different concentrations of dimethylsulphate (from 2·5 μM to 25 μM). However, a twofold increase in the specific activity of 3-methyladenine-DNA glycosylase was found in L-cells following exposure to 130 μM-methyl methane sulphonate for 24 h. This increase could be due to a general stimulus to cell proliferation caused by the alkylating agent, similar to the response in bacteria.

Sirover & Gupta (1985) have documented a general increase in the activity of individual DNA repair enzymes as quiescent cells are stimulated to proliferate. Their primary approach has been to examine the regulation of DNA repair in human fibroblasts, growth-arrested by serum depletion, and then stimulated to proliferate by the readdition of serum. From these results the authors suggest that normal cells regulate excision repair genes in a defined sequence with respect to the induction of DNA replication, such that DNA repair is stimulated prior to DNA synthesis.

In order to study the regulation of the 3-methyladenine-DNA glycosylase activity as a function of the cell cycle phase L-cells in logarithmic suspension growth (Fig. 2) were sorted according to size by means of the Beckman elutriation system. The DNA content of the different populations obtained was analysed by flow cytometry and the distribution of cells in the cell cycle established. Gi cells were allowed to progress through the cycle and their 5-phase index (measured by incorporation of [1H]thymidine) and viability were assessed.

Fig. 2.

Cell cycle-dependent induction of 3-methyladenine-DNA glycosylase in mouse L-cells. L-cells from suspension culture in logarithmic growth were separated on the basis of their size in a Beckman elutriation system. The DNA content and the cell cycle distribution of the different populations from the elutriation were established by means of flow cytometry. G1 cells were incubated further and the cell number was corrected for the presence of dead cells by means of vital staining. The onset of DNA synthesis was followed by measuring the incorporation of [1H]thymidine. At different times after the start of the synchronous growth, samples of the cells were collected for enzyme extraction. Enzyme activity of the 3-methyladenine-DNA glycosylase was measured as cts min−1 released from [1H]methyl methane sulphonate-treated calf thymus DNA per 101 cells. The procedure for measuring 3-methyladenine-DNA glycosylase activity has been described elsewhere (Male et al. 1985a).

Fig. 2.

Cell cycle-dependent induction of 3-methyladenine-DNA glycosylase in mouse L-cells. L-cells from suspension culture in logarithmic growth were separated on the basis of their size in a Beckman elutriation system. The DNA content and the cell cycle distribution of the different populations from the elutriation were established by means of flow cytometry. G1 cells were incubated further and the cell number was corrected for the presence of dead cells by means of vital staining. The onset of DNA synthesis was followed by measuring the incorporation of [1H]thymidine. At different times after the start of the synchronous growth, samples of the cells were collected for enzyme extraction. Enzyme activity of the 3-methyladenine-DNA glycosylase was measured as cts min−1 released from [1H]methyl methane sulphonate-treated calf thymus DNA per 101 cells. The procedure for measuring 3-methyladenine-DNA glycosylase activity has been described elsewhere (Male et al. 1985a).

At different times during synchronous growth cells were harvested and analysed for enzyme activity. It is evident from Fig. 2 that the activity of 3-methyladenine- DNA glycosylase in these cells increases substantially, reaching its maximum before the onset of DNA synthesis. During S-phase the 3-methyladenine-DNA activity drops to the basal level. It may be concluded from these observations that 3-methyladenine-DNA glycosylase activity in mammalian cells is regulated in a cell cycle-dependent manner. These results also indicate that no specific signal, such as alkylation of the DNA, is likely to increase the specific activity of this enzyme in the cells.

In order to understand better the function and biological role of mammalian DNA repair enzymes, these enzymes will have to be purified to homogeneity so that the amino acid sequence can be determined. On the basis of the amino acid sequence it should be possible to identify the genes coding for these activities and elucidate their structure and regulation. This approach, which is now pursued in several laboratories, will open up a new area in the knowledge of the role of DNA repair in eukaryotic systems. So far the main problem in this approach has been to obtain enough of the pure enzymes.

This study was supported by grants from the Norwegian Cancer Society (Landsforeningen mot Kreft). R.M. and D.E.H. are research fellows of this society. The authors thank Mrs Kirsten Selvik for excellent technical assistance.

Brent
,
T. P.
(
1979
).
Partial purification and characterization of a human 3-methyladenine-DNA glycosylase
.
Biochemistry
18
,
911
916
.
Doetsch
,
P. W.
,
Helland
,
D. E.
&
Haseltine
,
W. A.
(
1986
).
Mechanism of action of a mammalian DNA repair endonuclease
.
Biochemistry
25
,
2212
2220
.
FRIEDBERG
,
E. C.
(
1985
).
DNA Repair. New York: W. H. Freeman and Company
.
Gallagher
,
P. E.
&
Brent
,
T. P.
(
1982
).
Partial purification and characterization of 3-methyladenine-DNA glycosylase from human placenta
.
Biochemistry
21
,
6404
6409
.
Gallagher
,
P. E.
&
Brent
,
T. P.
(
1984
).
Further purification and characterization of human 3-methyladenine-DNA glycosylase
.
Biochim. biophys. Acta
782
,
394
401
.
Helland
,
D. E.
,
Doetsch
,
P. W.
&
Haseltine
,
W. A.
(
1986
).
Substrate specificity of a mammalian DNA repair endonuclease that recognizes oxidative base damage
.
Molec. cell. Biol.
6
,
1983
1990
.
Karran
,
P.
,
Hjelmgren
,
T.
&
Lindahl
,
T.
(
1982
).
Induction of a DNA glycosylase for N-methylated purines is part of the adaptive response to alkylating agents
.
Nature, Lond.
296
,
770
773
.
Lindahl
,
T.
(
1982
).
’ DNA repair enzymes. A
.
Rev. Biochem.
51
,
61
87
.
Male
,
R.
,
Haukanes
,
B. L
,
Helland
,
D. E.
&
Kleppe
,
K.
(
1987
).
Substrate specificity of 3-methyladenine DNA glycosylase from calf thymus
.
Eur.J. Biochem. (in press)
.
Male
,
R.
,
Helland
,
D.
&
Kleppe
,
K.
(
1985a
).
Purification and characterization of 3-methyladenine-DNA glycosylase from calf thymus
.
J. biol. Chem.
260
,
1623
1629
.
Male
,
R.
,
Helland
,
D. E.
,
Lillehaug
,
J. R.
&
Kleppe
,
K.
(
1985b
). Properties of mammalian 3-methyladenine-DNA glycosylases. In
Repair of DNA Lesions Introduced by i-nitroso Compounds
(ed. B. Myrnes &
H.
Krokan
), pp.
230
241
.
Oslo
:
Norwegian University Press
.
Male
,
R.
,
Nés
,
I. F.
&
Kleppe
,
K.
(
1981
).
Purification and properties of 3-methyladenine-DNA glycosylase from L-cells
.
Eur.J. Biochem.
121
,
243
248
.
Margison
,
G. P.
&
Pegg
,
A. E.
(
1981
).
Enzymatic release of 7-methylguanine from methylated DNA by rodent liver extracts
.
Proc. natn. Acad. Sci. U.SA.
78
,
861
865
.
Maxam
,
A. M.
&
Gilbert
,
W.
(
1980
)., Sequencing end-labelled DNA with base-specific chemical cleavages. In
Methods in Enzymology,
vol.
65
, part I, (ed. L. Grossman &
K.
Moldave
), pp.
499
560
.
New York
:
Academic Press
.
McCarthy
,
T. V.
,
Karran
,
P.
&
Lindahl
,
T.
(
1984
).
Inducible repair of O-alkylated DNA pyrimidines tn Escherichia coli.
EMBOJ.
3
,
545
550
.
Reiazudin
,
S.
&
Lindahl
,
T.
(
1978
).
Properties of 3-methyladenine-DNA glycosylase from Escherichia coli.
Biochemistry
17
,
2110
2118
.
Singer
,
B.
&
Brent
,
T. P.
(
1981
).
Human lymphoblasts contain DNA glycosylase activity exising N-3 and N-7 methyl and ethyl purines but not O6-alkylguanines or 1-alkyladenines
.
Proc. natn. Acad. Sci. U.SA.
78
,
856
860
.
Singer
,
B.
&
Kusmierek
,
J. T.
(
1982
).
Chemical mutagenesis. A
.
Rev. Biochem.
52
,
655
693
.
Sirover
,
M. A.
&
Gupta
,
P. K.
(
1985
). Regulation of DNA repair in human cells. In
Human Carcinogenesis
(ed. H. Autrup &
C.
Harris
), pp.
255
280
.
New York
:
Plenum Press
.
Thomas
,
L.
, Yang, C.-H. &
Goldwait
,
D. A.
(
1982
).
Two DNA glycosylases inEscherichia coliwhich release primarily 3-methyladenine
.
Biochemistry
21
,
1162
1169
.