©
1996 Oxford University Press
1837-1841
Footnote
Mutagenicity of a unique thymine-thymine dimer or thymine-thymine pyrimidine pyrimidone (6-4) photoproduct in mammalian cells
Mutagenicity of a unique thymine-thymine dimer or thymine-thymine pyrimidine pyrimidone (6-4) photoproduct in mammalian cells
A.
Gentil*
,
F.
Le Page
,
A.
Margot
,
C. W.
Lawrence
1
,
A.
Borden
1
and
A.
Sarasin
Institut de Recherches sur le Cancer, CNRS, IFC 1, Laboratory of Molecular
Genetics, BP 8, 94801
Villejuif
Cedex,
France
,
1
University of Rochester School of Medicine and Dentistry, Department of
Biophysics,
Rochester
, NY 14642,
USA
Received February 16, 1996:
Revised and Accepted April 1, 1996
ABSTRACT
The mutagenic properties of UV-induced photoproducts, both the
cis-syn
thymine-thymine dimer (TT) and the thymine-thymine pyrimidine pyrimidone (6-4) photoproduct [T(6-4)T] were studied in mammalian cells using shuttle
vectors. A shuttle vector able to replicate in both mammalian cells and
bacteria was produced in its single-stranded DNA form. A unique photoproduct was inserted at a single
restriction site and after recircularization of the single-stranded DNA vector, this latter was transfected into simian COS7 cells.
After DNA replication the vector was extracted from cells and used to transform
bacteria. Amplified DNA was finally analyzed without any selective screening,
DNA from randomly picked bacterial colonies being directly sequenced. Our
results show clearly that both lesions are mutagenic, but at different levels.
Mutation frequencies of 2 and 60% respectively were observed with the TT dimer
and the T(6-4)T. With the TT dimer the mutations were targeted on the 3
'
-T. With the T(6-4)T a large variety of mutations were observed. A majority of G
->
T transversions were semi-targeted to the base before the 5
'
-T of the photoproduct. These kinds of mutations were not observed when the
same plasmid was transfected directly into SOS-induced JM105 bacteria or when the T(6-4)T oligonucleotide inserted in a different plasmid was replicated
in SOS-induced SMH10
Escherichia coli
bacteria. These semi-targeted mutations are therefore the specific result of bypass of the T(6-4)T lesion in COS7 cells by one of the eukaryotic DNA polymerases.
INTRODUCTION
The mutagenic potency of UV light has been clearly shown both
in vitro
and
in vivo
using a variety of experimental systems in mammalian cells as well as in
bacteria (
1
-
3
). Numerous lesions are induced by UV light, usually implicating a doublet of
adjacent pyrimidines, as in the pyrimidine dimers and the pyrimidine (6-4) pyrimidones (
4
). Other minor lesions are also produced that concern pyrimidines and purines (
5
-
7
). We designed experiments to characterize the mutagenicity induced by two of the main UV-induced lesions, the
cis-syn
thymine-thymine cyclobutane dimer (TT) and the thymine-thymine pyrimidine pyrimidone (6-4) photoproduct [(T(6-4)T)]. Studies have been carried out concerning the
mutagenic properties of these two adducts in prokaryotes and in yeast (
8
), but no data are available in mammalian cells. We therefore constructed a
single-stranded DNA shuttle vector carrying a unique photoproduct able to
replicate in both mammalian cells and bacteria. This vector was then replicated
in mammalian cells and mutations were analyzed by DNA sequencing of rescued
plasmids after individualization and amplification of DNA molecules in
bacteria, without any screening.
The main interest of using a single-stranded DNA plasmid is the fact that the lesion cannot be a substrate for
the excision/resynthesis DNA repair mechanism and that survival of the plasmid
is subordinated to replication of its genome across the photoproduct. Using a
single-stranded shuttle vector carrying a unique DNA lesion also removes any
ambiguity about the strand position of the lesion and on the targeting of the
mutation. Moreover, using the same DNA sequence for insertion of one of the two
lesions under analysis eliminates any potential effect of DNA structural
features that have been shown to play an important role in the production of
mutational hot spots (
9
). This modified DNA sequence is the same as the one used in
Escherichia coli
and yeast studies (
10
,
11
).
MATERIALS AND METHODS
Plasmid
The plasmid that has been used is pS189, a generous gift from Dr Seidman. As
described previously (
12
), it possesses the origins of SV40 and of [pi]AN7 which allow its replication in mammalian cells and in bacteria. Single-stranded DNA was produced using the M13 phage replication origin
present on the plasmid and the M13K07 helper phage according to the procedure
described previously in Analects (Pharmacia). Bacteria carrying the vector were
selected by their ampicillin resistance.
Cells and bacteria
Monkey COS7 cells which constitutively express the SV40 T antigen were used in
these experiments. They were grown in Dulbecco's modified Eagle's medium in 90
mm Petri dishes supplemented with 7% fetal calf serum in a water-saturated 5% CO
2
atmosphere.
DH
5
[alpha] were utilized to detect mutants and JM105 or XL1 Blue to produce the
single-stranded DNA plasmid.
SMH10 and JM105 bacteria were SOS induced as described previously (
13
), before being transformed with the damaged vector.
Oligonucleotides
A control oligonucleotide (GCAAGTTGGAG) or a modified oligonucleotide carrying
either the
cis-syn
TT dimer or the T(6-4)T photoproduct were synthesized and purified as described previously (
14
,
15
). Other oligonucleotides necessary for vector construction were prepared in our
laboratory by the solid phosphoramidite method using an Applied Biosystems 380B
DNA synthesizer.
Construction of the vector
An adaptation of the method described by Banerjee
et al
. (
14
) and Leclerc
et al
. (
15
) was used (
16
). The single-stranded DNA vector was linearized at the unique
Bgl
II site after hybridization of a 20mer oligonucleotide, creating a double-stranded structure at this restriction site. The oligonucleotide helper
was removed by heating at 80oC for 2 min then by cooling to 0oC and filtration through a Sephacryl S-400 column (Pharmacia). The linearized molecule was then re-circularized with a 50mer scaffold oligonucleotide, the
sequence of its central part corresponding to the single-stranded region of the gapped molecule thus formed being complementary to
the oligonucleotide carrying the unique photolesion. This latter was then
inserted in the gap and ligation was performed using 400 U T4 DNA ligase at 15oC for 30 min. The 50mer scaffold was finally removed as described above for
the 20mer. Finally, in order to remove any partially double-stranded intermediate molecules which may have been left, digestions with
Hin
cII and
Xba
I enzymes, whose restrictions sites are located at each site of the 50mer, were
performed (
16
).
Experimental protocol
COS7 cells were transfected with the plasmid DNA using the DEAE-Dextran method (
17
). After 3 days, replicated extrachromosomal DNA was extracted from cells using
an alkaline lysis method (
18
). This extracted DNA was used to transform DH5[alpha]recA
-
bacteria by Hanahan's procedure (
19
) in order to isolate and amplify DNA molecules. The ampicillin resistance
conferred by the plasmid allowed selection of transformants. Vector DNA was
then prepared by alkaline lysis and DNA sequencing of the plasmid DNA
containing the inserted oligonucleotide was performed by the chain elongation
termination method using Sequenase II kits (Amersham).
RESULTS AND DISCUSSION
The mutagenic properties of TT dimers and T(6-4)T in mammalian cells are shown in Table
1
. The mutagenicity of TT dimers is relatively low. A mutation frequency of only
2% has been observed. This mutagenicity is targeted opposite the photolesion,
leading to T -> C and T -> A substitutions. Two transitions T -> C and one transversion T -> A have been observed on the 3'-T implicated in the dimer.
As far as the T(6-4)T is concerned a very elevated mutation frequency was observed. This
frequency reaches 60% in the plasmid DNA progeny. Among these mutations a large
majority of semi-targeted transversions was obtained. Eighty percent of G -> T transversions occurred in the immediate vicinity 5' of the photolesion. Very few mutations are targeted directly
opposite the lesion (5%) and a few frameshifts are recorded near the original
lesion, particularly a -1 deletion of the G residue corresponding to the hot spot of G -> T substitutions of the 5' border of the lesion. No mutants were found when no
photolesion was present (Table
1
).
Results obtained with the
cis-syn
TT dimer are not surprising when compared with its mutagenic potency in bakers'
yeast,
Saccharomyces cerevisiae
. Indeed in this lower eukaryote, the mutation frequency of this lesion was <1% (
10
). In SOS-induced bacteria the error rate was higher and the mutation frequency
reached ~7%. Although we detected only a few mutants, the mutations we observed are
those expected from the observations in bacteria, that is to say, a base
substitution at the 3'-T of the photoproduct. As argued by Banerjee
et al
. (
14
) for bacteria, it seems that the low level of mutation induced by the TT dimer
in mammalian cells is most probably due to the fact that the dimer retains
enough correct base pairing information for polymerase bypass, rather than a
bias of adenine incorporation opposite a non-instructive lesion (
20
). In this latter case, the mutagenic potency of the photoadduct should be
greatly elevated, since the A insertion frequency is ~60-80% in the case of non-instructive apurinic sites (
21
). A mutation frequency of 20-40% should therefore be observed when TT dimers are replicated, if the `A
rule' should apply.
Mutations observed after replication of T(6-4)T in COS7 cells are more difficult to interpret in view of what has been
observed in bacteria and in yeast (
11
,
15
). Indeed a large variety of mutations was observed and, above all, a hot spot
located in the immediate vicinity of the 5'-T implicated in the photolesion. Moreover, it has to be noted that
in five cases (almost 8% of the mutants) this base 5' of the lesion has been deleted. It appears therefore that the nucleotide
adjacent to the 5'-T of the photolesion is a privileged target under our experimental
conditions. In order to test if some impurities in the modified
oligonucleotides could explain this result, we used the same batch of modified
oligonucleotides to transfect SOS-induced bacteria as described previously (
14
). The T(6-4)T and TT photoproduct-carrying oligonucleotides were inserted into the single-stranded DNA vector M13mpL1 (
13
). This DNA was then transfected directly into competent SOS-induced SMH10 bacteria, which is a uvr A6 derivative of the AB1157 strain.
As in experiments carried out in COS7 cells, no selection was applied, as
colonies were picked and sequenced at random. As shown in Table
2
, all the mutations were targeted. However we have no explanation for the higher
frequency of targeted deletions found in these experiments as compared with
that observed previously (
15
). Mutagenesis induced by the
cys-syn
dimer in SOS-induced bacteria was totally as expected, i.e. a low mutation frequency
(9.6%) and all mutants being targeted at the 3'-T of the dimer (Table
2
). It therefore demonstrates that the semi-targeted mutations observed using the COS7 cells are not artefacts due to
the T(6-4)T-carrying oligonucleotide. Another hypothesis to explain our results
in COS7 cells could be that our vector exhibits specific secondary structure at the site of the T(6-4)T lesions due to the DNA sequence present around it. This putative
structure could be a substrate for a specific type of error-prone replication at that particular site 5' of the T(6-4)T. Indeed, it has been shown in experiments in which UV DNA
lesions were randomly induced that hot spots of mutagenesis could be the
consequence of particular DNA structure features. The complete construction of
pS189 carrying the unique DNA lesion used to transfect COS7 cells was thus used
to transform SOS-induced JM105 bacteria. Table
3
shows that mutations are mainly those expected according to the results
observed previously (
15
). Only one mutation was identical to the semi-targeted mutational hot spot observed in mammalian cells, while the vast
majority (80%) of mutations are targeted at the lesion. In conclusion, the
specific mutational hot spot found in COS7 cells is really due to replication
of the lesion by the cells and not to a defect in the oligonucleotide or to a
conformational structure of the pS189 plasmid DNA.
Table 1
.
Mutagenesis of TT
cis-syn
cyclobutane dimer and T(6-4)T in mammalian cells
|
Sequences observed
|
|
TT
cis-syn
dimer
|
|
T(6-4)T
|
|
|
TT control
|
|
|
|
|
|
|
Number of
|
Mutation
|
Relative
|
Number of
|
Mutation
|
Relative
|
Number of
|
Mutation
|
|
|
sequences
|
frequency
|
mutation
|
sequences
|
frequency
|
mutation
|
sequences
|
frequency
|
|
|
analyzed
|
(%)
|
frequency
|
analyzed
|
(%)
|
frequency
|
analyzed
|
(%)
|
|
|
|
|
(%)
|
|
|
(%)
|
|
5'-GCAAGTTGGAG-3'
a
|
120
|
|
|
41
|
|
|
102
|
|
GCAA
T
TTGGAG
b
|
|
|
|
51
|
|
80.9
|
|
GCAAGT
C
GGAG
|
2
|
|
66.6
|
2
|
|
3.2
|
|
GCAAG
G
TGGAG
|
|
|
|
1
|
|
1.6
|
|
GCAAGTTGG
G
G
|
|
|
|
2
|
|
3.2
|
|
GCAAGTT
C
GAG
|
|
|
|
1
|
|
1.6
|
|
GCAA(
[Delta]
G
)TT(
+C
)GGAG
|
|
|
|
5
|
|
7.9
|
|
GCAAGTTGG([Delta]A)G
|
|
|
|
1
|
|
1.6
|
|
GCAAGT
A
GGAG
|
1
|
|
33.1
|
|
Total
|
123
|
2.4
|
|
104
|
60.5
|
|
102
|
<1%
|
The photoadduct was located at the T-T site in the sequence 5'-GCAAGTTGGAG-3'.
a
Non-mutated sequence.
b
Mutations are underlined.
Table 2
.
Mutagenesis of TT
cic-syn
cyclobutane dimer and T(6-4)T in SOS-induced SMH10
E.coli
bacteria
|
Sequences observed
|
Photolesion
|
|
|
TT
cis-syn
dimer
|
|
T(6-4)T
|
|
|
Number of
|
Mutation
|
Number of
|
Mutation
|
|
|
sequences
|
frequency (%)
|
sequences
|
frequency (%)
|
|
5'-GCAAGTTGGAG-3'
a
|
188
|
|
3
|
|
GCAAGT
A
GGAG
b
|
18
|
|
GCAAGT
C
GGAG
|
2
|
|
28
|
|
GCAAG
C
TGGAG
|
|
|
1
|
|
GCAAG(
[Delta]
T
)TGGAG
|
|
|
27
|
|
Total
|
208
|
9.6
|
59
|
95
|
The photolesion was located at the T-T site in the sequence 5'-GCAAGTTGGAG-3'.
a
Non-mutated sequence.
b
Mutations are underlined.
Table 3
.
Mutagenesis of the T(6-4)T photolesion inserted in pS189 replicated in SOS-induced JM105 bacteria
|
Sequences
|
Number of
|
Mutation
|
|
observed
|
sequences
|
frequency (%)
|
|
5'-GCAAGTTGGAG-3'
a
|
29
|
|
GCAAGT
C
GGAG
b
|
10
|
|
GCAAG
C
TGGAG
|
3
|
|
GCAA
T
TTGGAG
|
1
|
|
GCAAGT
GCAGCC
|
1
|
|
GCAAGTTGG
C
G
|
1
|
|
Total
|
45
|
35.5
|
The photolesion was located at the T-T site in the sequence 5'-GCAAGTTGGAG-3'.
a
Non-mutated sequence.
b
Mutations are underlined.
When present on a single-stranded DNA vector the T(6-4)T photoproduct led to a mutation frequency of 91% after
replication in uvr A6 bacteria (
15
) and all mutations were targeted opposite the lesion, with a great majority of
them being 3'-T -> C transitions. In mammalian cells a high mutagenic potency
was also obtained (60%). For mammalian cells, studies reported in the
literature concerning UV mutagenesis were mainly performed after random UV
irradiation of the target. Numerous studies were carried out with single- or double-stranded DNA shuttle vectors, lacZ' and supF tRNA for example being the mutagenesis targets. In
these cases, mutations may only be correlated with putative UV-induced lesions. The specificity of the mutations observed under these
experimental conditions is less striking, i.e. the number of mutations targeted
opposite pyrimidine-pyrimidine sequences varied between 60 and 80% in normal cells. Although
most of the mutations observed were targeted (
22
,
23
), frameshift and semi-targeted mutations were also reported (
24
). Indeed, using a single-stranded DNA shuttle vector a non-negligible number of semi-targeted mutations located at the 5'-side of the putative pyrimidine-pyrimidine lesion was found, these semi-targeted mutations being or not
being associated with a mutation opposite the 3' base of the pyrimidine doublet. The problem now is to understand how
such a mutation may have occurred. Several hypotheses may be put forward, but
up till now no mechanism that could explain this phenomenon has been
demonstrated. It is clear that the T(6-4)T photoproduct induces an important distortion of the DNA strand and
that this distortion may play a crucial role in the mutation spectrum observed,
where point mutations and also deletions or insertions occurred all around the
lesion. This suggests that some slippage phenomenon during replication may
intervene leading to this kind of error. Moreover, it has to be recalled that
the DNA replication mechanism of the single-stranded DNA vector transfected into mammalian cells is still largely
unknown. However, comparing replication in
S.cerevisiae
of a gapped duplex DNA shuttle vector with replication of a single-stranded DNA, both carrying a unique T(6-4)T photoproduct, Gibbs
et al
. (
11
) got similar results, indicating that replication of a single-stranded vector is very close if not identical to that of a double-stranded one. Mutation frequencies and the kinds of mutations were
different from those found in our work, as 60-70% of molecules were replicated accurately. Almost all the induced
mutations were targeted to the T-T site, 10-20% of them being 3'-T -> C substitutions. On the other hand, our results
concerning the high mutagenic potency of T(6-4)T are in agreement with those obtained in Chinese hamster cells using
the
hprt
gene as target (
25
,
26
). These studies showed that the T(6-4)T photoproduct was the main UV-induced mutagenic lesion and that no preferential repair of the
lesion in the transcribed gene was found. It seems, therefore, that the
mutagenic properties of the photoproduct is correlated with the organism and
with the replication conditions.
In bacteria, replication of both lesions requires a pretreatment in order to
induce SOS conditions in the cells. As in
S.cerevisiae
and unlike in
E.coli
, UV pretreatment is not necessary in mammalian cells to bypass the lesions. No
pretreatment of any kind in mammalian cells was necessary to obtain a mutated
plasmid progeny. However, we cannot eliminate the possibility of the induction
of SOS-like conditions in COS7 cells by the transfection protocol itself.
ACKNOWLEDGEMENTS
We thank L.Daya-Grosjean for critical reading of the manuscript. This work was supported by grants from the Association de la Recherche sur
le Cancer, ARC (Villejuif, France), the Commission of the European Community (Brussels, Belgium) and by US Public Service Grant
no. GM32885 from the National Institutes of Health. F.Le Page was supported by
a fellowship from the Ligue Française Nationale Contre le Cancer (Val de Marne).
REFERENCES
1 Witkin,E.M. (1976) Bacteriol. Rev., 40, 869-907. MEDLINE Abstract
2 Sarasin,A. and Benoit,A. (1980) Mutat. Res., 70, 71-81. MEDLINE Abstract
3 Hsia,C.H., Lebkowski ,J.J., Leong,P.M., Calos,M.P. and Miller,J.H. (1989) J. Mol. Biol., 205, 103-113. MEDLINE Abstract
4 Hélène,C. and Charlier,M. (1978) Biochimie, 60, 1110-1121.
5 Bose,S.N., Davies,R.J.H., Sethi,S.K. and McCloskey,J.A. (1983) Science, 220, 723-725. MEDLINE Abstract
6 Bourre,F., Renault,G. and Sarasin,A. (1987) Nucleic Acids Res., 15, 8861-8875. MEDLINE Abstract
7 Gallagher,P.E. and Duker,N.J. (1986) Mol. Cell. Biol., 6, 707-709. MEDLINE Abstract
8 Lawrence,C.W., Gibbs,P.E.M., Borden,A., Horsfall,M.J. and Kilbey,B.J. (1993) Mutat. Res., 299, 157-163. MEDLINE Abstract
9 Brash,D.E. and Haseltine,W.A. (1982) Nature, 298, 189-192. MEDLINE Abstract
10 Gibbs,P.E.M., Kilbey,B.J., Banerjee,S.K. and Lawrence,C.W. (1993) J. Bacteriol., 175, 2607-2612.
11 Gibbs,P.E.M., Borden,A. and Lawrence,C.W. (1995) Nucleic Acids Res., 23, 1919-1922. MEDLINE Abstract
12 Seidman,M. (1989) Mutat. Res., 220, 55-60. MEDLINE Abstract
13 Banerjee,S.K., Borden,A., Christensen,R.B., Leclerc,J.E. and Lawrence,C.W. (1990) J. Bacteriol., 172, 2105-2112. MEDLINE Abstract
14 Banerjee,S.K., Christensen,R.B., Lawrence,C.W. and Leclerc,J.E. (1988) Proc. Natl. Acad. Sci. USA, 85, 8141-8145. MEDLINE Abstract
15 Leclerc,J.E., Borden,A. and Lawrence,C.W. (1991) Proc. Natl. Acad. Sci. USA, 88, 9685-9689. MEDLINE Abstract
16 Cabral Neto,J.B., Caseira Cabral,R.E., Margot,A., Le Page,F., Sarasin,A. and Gentil,A. (1994) J. Mol. Biol., 240, 416-420.
17 MacCutcham,J.H. and Pagano,J.J. (1968) J. Natl. Cancer Inst., 42, 351-357.
18 Stary,A. and Sarasin,A. (1992) J. Gen. Virol., 73, 1679-1685. MEDLINE Abstract
19 Hanahan,D. (1983) J. Mol. Biol., 66, 557-580.
20 Strauss,B., Rabkin,S., Sagher,D. and Moore,P. (1982) Biochimie, 64, 829-838. MEDLINE Abstract
21 Lawrence,C.W., Borden,A., Banerjee,S.K. and Leclerc,J.E. (1990) Nucleic Acids Res., 18, 2153-2157. MEDLINE Abstract
22 Bourre,F. and Sarasin,A. (1983) Nature (Lond.), 305, 68-70.
23 Sarasin,A., Bourre,F., Daya-Grosjean,L., Gentil,A., Madzak,C. and Stary,A. (1990) Int. J. Radiat. Biol., 57, 665-676. MEDLINE Abstract
24 Madzak,C. and Sarasin,A. (1991) J. Mol. Biol., 218, 667-673. MEDLINE Abstract
25 Zdzienicka,M.Z., Venema,J., Mitchell,D.J., Van Hoffen,A., van Zeeland,A.A., Vrieling,H., Mullenders,L.H.F., Lohman,P.H.M. and Simons,J.W.I.M. (1992) Mutat. Res., 273, 73-83. MEDLINE Abstract
26 Mullenders,L.H.F., Hazekamp-van Dokkum,A.M., Kalle,W.H.J., Vrieling,H., Zdzienicka,M.Z. and Van Zeeland,A.A. (1993) Mutat. Res., 299, 271-276. MEDLINE Abstract
Return
*
To whom correspondence should be addressed
