ABSTRACT
Aziridine (ethyleneimine) reacts with DNA
in vitro,
mainly at the N7 position of guanine and N3 of adenine, then imidazole ring
opening of the modified guanine results in formation of formamidopyrimidine
(FaPy) residues. The
Escherichia coli
fpg
gene encodes a DNA glycosylase that removes FaPy residues from DNA. To
determine whether aziridine produces FaPy lesions in mammalian cells we have
expressed the
E.coli
fpg
gene in CHO cells. The transfected cells, expressing high levels of the
bacterial protein, are more resistant to the toxic and mutagenic effects of aziridine than the control
population. Less DNA damage was measured by quantitative PCR analysis in transfected than in control cells
treated with equimolar concentrations of aziridine. The results suggest that
aziridine produces
in vivo
FaPy residues that could account for the deleterious effects of this compound.
Chemical compounds induce a wide variety of lesions in DNA that can interfere
with replication and transcription and, therefore, result in mutagenesis or
cell death. Aziridine (ethyleneimine) is an alkylating agent produced in large
amounts in industry (
1
). This compound has been shown to be toxic and mutagenic in various biological
systems and to produce chromosome abberations and sister chromatid exchanges in
human cells (reviewed in
2
). However, there are few data in the literature concerning the nature of the
DNA damage produced by this compound
in vivo
and, hence, concerning the mechanisms implicated in repair of these lesions.
It has been shown that aziridine reacts
in vitro
with guanine and guanosine to form an
N
7-aminoethyl derivative (
3
). This product exhibits an unusual tendency to undergo imidazole ring opening (
4
) and the reaction of guanosine with aziridine results in the formation of two
main products: imidazole ring-opened
N
7-alkylguanosine and
N
1-alkylguanosine, accounting for 80 and 14% of all adducts respectively (
4
). Protonated aziridine also reacts with the N3 nitrogen of adenine, although
this site is predicted to be less reactive than N7 of guanine (
5
). In
Escherichia coli
ring-opened guanine and ring-opened adenine, formed by alkylating agents , are repaired by a specific formamidopyrimidine-DNA glycosylase, the Fpg protein (
6
). This protein is encoded by the
fpg
(or
MutM
) gene (
7
), which, besides its glycosylase activity, also possesses AP-nicking (
6
) and dRPase activities (
8
).
In order to check whether the aziridine-induced DNA damage formed in mammalian cells could be recognized and
repaired by formamidopyrimidine-DNA glycosylase activity we have overexpressed the bacterial Fpg protein
in Chinese hamster ovary (CHO) cells. The lethal and mutagenic effects of
aziridine were measured in CHO cells expressing or not the Fpg protein. The
number of lesions formed in a specific genomic region was measured by
quantitative PCR (
9
,
10
) in control and transfected cells. The results show that overexpression of the
bacterial Fpg protein protects the cells against the deleterious effects of
aziridine and decreases the lesion frequency produced by this coupound.
CHO cells were grown in Dulbecco's medium supplemented with 5% fetal calf serum
and 5% horse serum. They were transfected by electroporation, using a BioRad
gene pulser apparatus as described (
11
). The transfected cells were grown in G418-containing medium (750 [mu]g/ml) until appearance of clones.
The survival of aziridine-treated cells was measured by incubating exponentially growing cells for
30 min at 37oC in culture medium supplemented with 2% serum and increasing
concentrations of the drug. They were then rinsed, trypsinized and aliquots of
the suspension were cultured in normal medium until appearance of clones.
For mutagenicity determination the cells were incubated with aziridine as
described above, then grown in fresh medium for 8-11 days to allow expression of the mutant phenotype. They were then
subcultured in normal medium for survival determination or in the presence of 6-thioguanine (2.5 [mu]g/ml) to measure the mutation frequency.
To measure DNA synthesis the cells were incubated with various aziridine
concentrations for 30 min, rinsed and then grown for 2 h in normal medium. They
were then grown for 30 min in the presence of [
3
H]thymidine and the specific activity of the cellular DNA determined as already
described (
12
).
The coding sequence of the
E.coli
fpg
gene (809 bp) was excised with
Hin
dIII and
Pvu
II from plasmid fpg220 (
6
). After purification by gel electrophoresis the gene was ligated in the
Hin
dIII site of the psV2-neo vector. The plasmid carrying the insert was designated psV2-fpg. A psV2-neo vector carrying
APDG
cDNA (encoding rat
N
3-methyladenine-DNA glycosylase) (
13
) was constructed and designated psV2-APDG.
Cells were suspended (10
8
cells/ml) in a buffer containing 70 mM HEPES, 100 mM KCl, 2 mM EDTA, 1 mM DTT
and 10% glycerol. They were disrupted by sonication at 0oC in the presence of proteases inhibitors (leupeptin, aprotinin and
antipain, 2 [mu]g/ml each). Cell debris was removed by centrifugation (10 000
g
, 5 min, 4oC). Increasing amounts of cell extracts were incubated (final volume 100 [mu]l) for 30 min at 37oC with [
3
H]formamidopyrimidine (FaPy)-poly(dG[middot]dC) prepared as previously described (
6
). The radioactivity present in the ethanol-soluble fraction was quantitated by scintillation spectroscopy.
Characterization of the two rotameric forms of FaPy residues (
14
) in the ethanol fraction was by HPLC analysis after addition of authentic
markers, using a C18-[mu]Bondapack column eluted with 50 mM ammonium phosphate containing 5%
methanol.
Cells were lysed in 0.5 M NaCl, 0.05 M EDTA, 0.05 M Tris, pH 7.5, by addition of
SDS (2% final concentration). The lysate was treated for 2 h at 55oC with proteinase K (200 [mu]g/ml; Boehringer, Mannheim), then the DNA purified as described (
9
). Exon 9 (723 bp) or exons 2-3 (3 kb) of the
HPRT
gene were amplified using the primers described by Rossiter
et al
. (
15
) and a DNA Thermal Cycler (Perkin Elmer Cetus). Amplification products were
radioactively labeled by incorporation of [
32
P]dCTP (
16
) during the exponential increase in the PCR products (between 26 and 30 cycles)
and were analyzed by electrophoresis in 1.8% agarose gels. Quantitation of the
PCR products was performed by TCA precipitation and scintillation counting as
described (
17
). The number of lesions per strand was calculated as -ln(
A
d
/
A
o
), where
A
d
is c.p.m. incorporated in damaged DNA and
A
o
is c.p.m. incorporated in non-damaged DNA template (
10
).
Among the different G 418-resistant clones obtained after trans- fection of CHO cells with plasmid psV2-fpg some expressed a high level of FaPy-DNA glycosylase activity, which was ~40 times higher than the endogenous level for CHO
cells (Fig.
1
). Analysis of the reaction products by HPLC showed removal of FaPy residues
(Fig.
1
, insert). Inhibition of the activity in cell extracts incubated with specific
antibodies raised against the bacterial Fpg protein confirmed that the
increased activity was due to expression of the bacterial gene (
18
).
The survival of CHO cells was measured after exposure to increasing aziridine
concentrations. The survival curves showed that the transfected cells (Fig.
2
) were more resistant than the controls to the lethal effect of this compound.
It should be noted that the growth rate was identical in the two cell
populations and that no significant modification of cell survival was observed
in cells transfected with the control psV2-neo vector (Table
1
). To check that the enhanced resistance was specifically related to expression
of the Fpg protein CHO cells were transfected with psV2 vector carrying
APDG
cDNA. This cDNA was used because it expresses rat
N
3-methyladenine-DNA glycosylase, which specifically removes
N
3-alkyladenine and
N
7-alkylguanine residues from DNA (
12
): the transfected cells had an APDG activity ~10 times higher than the control value (Table
1
), but this enhanced activity did not modify cell sensitivity to the lethal
effect of aziridine (Fig.
2
and Table
1
). It should be recalled that expression of
APDG
cDNA could increase cell resistance to the toxic effect of MMS or EMS (
11
).
Results presented in Figure
3
show that aziridine is a potent mutagenic compound in mammalian cells. To check
whether the mutagenic lesion(s) produced by this compound were repaired by Fpg
protein the number of mutations in the
HPRT
locus was measured in control and transfected cells. The results showed that
the aziridine-induced mutation frequency was reduced in
fpg
gene-expressing cells (Fig.
3
). However, when control experiments were performed with cells transfected with
the psV2-APDG vector no decrease in aziridine-induced mutation frequency was observed (Fig.
3
).
Expression of the bacterial Fpg protein increases the resistance of CHO cells to
the toxic and mutagenic effects of aziridine, suggesting that this protein
removes aziridine-induced damage from cellular DNA. However, expression of this protein does
not change cell sensitivity to MMS (Laval, unpublished results). This suggests
that the biological effect of Fpg protein in the cells is due neither to its AP-nicking activity nor to its dRPase activity. Furthermore, the damage
produced by aziridine is not recognized by rat
N
3-methyladenine-DNA glycosylase, as expression of this protein does not modify cell
sensitivity to aziridine. As already described for HeLa cells (
20
), aziridine decreases DNA synthesis, suggesting formation of DNA lesions which
are a block to DNA replication. The aziridine concentration which blocks DNA
synthesis (1 mM) 2 h after treatment is not highly toxic to the cells. This
suggests that even control cells can sustain a level of aziridine-induced DNA damage and this result is expected, because CHO cells possess
a FaPy-DNA glycosylase activity. We have used the quantitative PCR technique to
measure the number of replication blocking lesions in DNA of aziridine-treated cells, as it has been shown that PCR could be used to quantify
damage induced in DNA by various agents, e.g.
cis
-platinum and UV light (
9
), even in a small gene segment (
17
). Using this analysis we could detect damage in aziridine-treated CHO cells and show that expression of the
E.coli
fpg
gene decreased the number of lesions present in cellular DNA. However, it is
possible that a low number of lesions are by-passed by
Taq
polymerase and therefore are not detected in our assay, lowering the number of
lesions measured in the DNA of the treated cells.
In vitro
ring-opened
N
7-guanine residues on the template strand of a DNA molecule prevent movement
of the replication fork and block DNA synthesis one base before the lesion (
21
). As Fapy residues are formed by aziridine (
4
) and repaired by the Fpg protein (
6
) they are good candidates to be one form of lethal damage produced by aziridine
in vivo.
The presence of Fapy residues in M13 phage DNA correlates with a decrease in
transfection efficiency and in mutagenicity when transfected into SOS-induced
E.coli
cells (
22
). In phage treated with dimethyl sulfate and alkali, treatment that produces
FaPy residues, most mutations were found not at G, suggesting that FaPy-guanine is not mutagenic, but at A, yielding A -> G transitions (
22
). These results suggest the formation of a mutagenic adenine derivative. As the
mutagenic effect of aziridine decreases in CHO cells expressing Fpg protein, a
modified adenine could be one form of damage responsible for the mutagenic
properties of aziridine.
In conclusion, the Fpg protein removes two types of lesions from the DNA of
aziridine-treated cells. (i) Lesions which are a block to DNA replication and are
probably responsible for the lethal effect of this compound. Due to the
specificity of the Fpg protein and the properties of FaPy-guanine residues, these residues are a good candidate for this lethal
lesion. (ii) Lesions which are mutagenic and are also recognized by the Fpg
protein. It has been suggested that FaPy-adenine residues could occur and be mutagenic in phage treated with
alkylating agents and alkali (
22
). This damage is repaired by the Fpg protein and could, therefore, be
responsible for the mutagenic effect of aziridine in mammalian cells.
This work was supported by grants from INSERM, ARC (Villejuif) and EEC (contract
EV5V-CT94-0396).
+
On leave from Unité 71 INSERM, Clermont-Ferrand, France
REFERENCES
Return