Substitution and deletion mutations induced by 2-hydroxyadenine in
Escherichia coli
: effects of sequence contexts in leading and lagging strands
Substitution and deletion mutations induced by 2-hydroxyadenine in Escherichia coli : effects of sequence contexts in leading and lagging strands
Hiroyuki
Kamiya
and
Hiroshi
Kasai*
Department of Environmental Oncology, Institute of Industrial Ecological
Sciences, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku,
Kitakyushu
807,
Japan
Received October 10, 1996;
Revised and Accepted November 18, 1996
ABSTRACT
To evaluate the mutation frequency and the mutation spectrum of 2-hydroxyadenine (2-OH-Ade), an oxidative DNA lesion, the modified base was site-specifically incorporated into a unique restriction enzyme site (
Sal
I, GTCGA*C or
Afl
II, CTTA*AG where A* represents 2-OH-Ade) in single- and double-stranded vectors. The 2-OH-Ade residues were introduced into (+)- and (-)-strands of the double-stranded vectors
and into the (+)-strand of single-stranded vectors. When the vectors were transfected into
Escherichia coli
, the modified base showed little to no cytotoxicity. The mutation frequencies of 2-OH-Ade in the
Sal
I and
Afl
II sites were
~
0.8 and 0.07%, respectively, with double-stranded (+)-vectors. An increase in the mutation frequencies was not observed
with single-stranded vectors. When incorporated into the (-)-strand, the mutation frequencies of 2-OH-Ade in the
Sal
I and
Afl
II sites were
~
0.3 and 0.1%, respectively. The mutations observed most frequently were -1 deletions at both positions, in the case of the (+)-strand. On the other hand, we observed that 2-OH-Ade in the (-)-strand induced A
->
G and A
->
T substitutions. These results indicate that 2-OH-Ade residues in DNA induce substitution and deletion mutations
without blocking replication in
E.coli
.
INTRODUCTION
It is an accepted concept that reactive oxygen species (ROS) and the oxidative
DNA damage produced by ROS are involved in cellular processes such as
mutagenesis, carcinogenesis and ageing (
1
). ROS are produced by normal cellular respiration and oxygen metabolism, and
thus oxidative DNA damage contributes to spontaneous mutation. The damage is
also important because some mutagens/carcinogens produce these DNA lesions. Of
the many types of oxidative DNA damage known, only a few DNA lesions (e.g. 8-hydroxyguanine, abbreviated as 8-OH-Gua, hereafter) (
2
) have been studied in detail regarding their cellular lethality and mutation
inducing properties (
3
-
7
).
We recently found that hydroxylation of an adenine base at the C-2 position occurs when dA, dATP and DNA are treated with an ROS-forming system, Fe
2+
-EDTA-O
2
, to form 2-hydroxyadenine (2-OH-Ade, 1,2-dihydro-2-oxoadenine or isoguanine) (
8
). Rather than by direct oxidation of an adenine base in DNA, most of the 2-OH-Ade in DNA will be formed
in vivo
by the incorporation of 2-hydroxy-2'-deoxyadenosine 5'-triphosphate (2-OH-dATP) during DNA replication,
because DNA polymerases (mammalian DNA polymerase [alpha] and the Klenow fragment of
Escherichia coli
DNA polymerase I) incorporate the oxidized nucleotide in
in vitro
DNA synthesis (
8
). The formation of 2-OH- Ade in cellular DNA will cause mutations by misincorporation of
nucleotides, as observed in
in vitro
DNA synthesis (
9
,
10
), and may be a cause of mutations at various stages of carcinogenesis.
To study the frequency and the spectrum of mutations by 2-OH-Ade
in vivo
, we incorporated the oxidized base into unique, predetermined sites in single-stranded (ss) and double-stranded (ds) vectors and transfected these vectors into
E.coli
. We observed that (i) the 2-OH-Ade induced either a minimal or no DNA replication block and (ii)
the 2-OH-Ade was mutagenic in
E.coli
. Moreover, we revealed that (iii) the mutation spectra elicited by 2-OH-Ade were affected by the sequence contexts and the strands upon
which the base was located.
MATERIALS AND METHODS
Materials
T4 polynucleotide kinase,
Apa
I and
Kpn
I were purchased from Toyobo Co. T4 DNA ligase and
Eco
RI were obtained from Takara Shuzo Co.
Afl
II was from New England Biolabs. Snake venom phosphodiesterase was from
Boehringer Mannheim.
Sal
I was from either Nippon Gene or Takara Shuzo Co. pSVK3 was purchased from
Pharmacia Biotech Inc.
Escherichia coli
strain HB101 (
supE
44,
hsdS
20 (
r
B
-
m
B
-
),
recA
13,
ara
-14,
proA
2,
lacY
1,
galK
2,
rpsL
20,
xyl
-5,
mtl
-1,
leuB
6,
thi
-1) was used as the host bacteria. HB101 cells for electroporation were purchased from Nippon Gene or prepared according to the usual method
(
11
). Competent HB101 cells for the CaCl
2
method were obtained from Takara Shuzo Co. or prepared according to the method
described in the literature (
11
).
Oligonucleotide synthesis
A phosphoramidite derivative of 2-OH-Ade, 9-[2'-deoxy-5'-
O
- (4,4'-dimethoxytrityl)-[beta]-D-
erythro
-pentofuranosyl]-6-[(dimethylamino)methylidene]-9
H
-isoguanine 3'-[(2-cyanoethyl)
N,N
-diisopropylphosphoramidite], prepared as described previously (
12
), was a gift from Drs Tadaaki Ohgi and Toshihiro Ueda of Nippon Shinyaku Co.
Ltd. Oligonucleotides used for vector construction with or without 2-OH-Ade were synthesized by the phosphoramidite method with an Applied Biosystems model 394 DNA/RNA synthesizer (Perkin-Elmer) in the form including the dimethoxytrityl group.
The cleavage from the support and the deprotection of oligonucleotides with a
concentrated ammonia solution were carried out as described previously (
9
). The oligonucleotides with the dimethoxytrityl groups were purified by reverse-phase HPLC using an Inertsil ODS-2 column (10 * 250 mm, GL Sciences Inc.) with a linear gradient of
acetonitrile in 50 mM triethylammonium acetate. After removal of the 5'-dimethoxytrityl group with 80% aqueous acetic acid, the fully
deprotected oligonucleotides were purified by reverse-phase HPLC using the same kind of column, and by anion-exchange HPLC using a TSK-GEL DEAE-2SW column (4.6 * 250 mm, Tosoh Co.) with a linear gradient of ammonium formate in 20% aqueous acetonitrile.
Ammonium formate was removed by reverse-phase chromatography using Sep-Pak® plus C
18
cartridges (Millipore Co.). The oligonucleotides were further purified by reverse-phase HPLC using a YMC-Pack ODS-AM 303 column (4.6 * 250 mm, YMC Co.) and by anion- exchange HPLC.
The following oligonucleotides were synthesized:
A-17, (5'-dGGTCGACTTAAGGTACC-3');
2-OH-Sal, (5'-dGGTCGA*- CTTAAGGTACC-3');
2-OH-Afl, (5'-dGGTCGACTTA*AGGTACC-3');
splint-1, (5'-dTCGAGGTACCTTAAGTCGACCGTAC-3');
splint-2, (5'-dAATTGGTACCTTAAGTCGACCGG- CC-3');
where A* represents 2-OH-Ade.
The oligonucleotides with 2-OH-Ade (0.1 A
260
units) were digested with snake venom phosphodiesterase (2 [mu]g) and alkaline phosphatase (0.2 U) in 20 [mu]l of a buffer solution, containing 0.1 M Tris-HCl (pH 8.0) and 1 mM MgSO
4
, at 37oC for 24 h. The sample was passed through Ultra-free Probind (Millipore Co.) to remove the proteins, and was analyzed
by reverse-phase HPLC using an Ultrasphere ODS 5[mu] column (4.6 * 250 mm, Beckman) with an isocratic system consisting of 10 mM
NaH
2
PO
4
and 8.0% methanol.
Primer-1 (5'-dCCTCTGAGCTATTCCAGA-3') and primer-2 (5'-dAATTTCTGCCATTCATCCG-3') for the
sequencing reactions were purchased from Hokkaido System Science Co. (Sapporo, Japan) in purified
form.
Construction of vectors
The oligonucleotides, A-17, 2-OH-Sal and 2-OH-Afl (20 pmol), were each phosphorylated at their 5'-ends by incubation with T4
polynucleotide kinase and ATP in a buffer solution of 50 mM Tris-HCl (pH 9.5), 10 mM MgCl
2
, 10 mM 2-mercaptoethanol and 5% glycerol, in a total volume of 10 [mu]l, at 37oC for 30 min. Additional enzyme was added, and the mixture was
incubated at 37oC for a further 30 min. Splint-1 and -2 were also phosphorylated to construct the ds (+)- and (-)-vectors respectively.
Phosphorylated A-17, 2-OH-Sal and 2-OH-Afl (10 pmol) were each mixed with an equal molar
amount of phosphorylated or unphosphorylated splint oligonucleotide. The
mixture was heated at 80oC for 5 min and cooled slowly to room temperature. Before the ligation
reaction described below, this mixture (the annealed oligonucleotide) was kept
on ice for 10 min.
pSVK3 was digested with
Kpn
I and then with
Sal
I under the conditions recommended by the suppliers. After ethanol precipitation, the large fragment was purified by extraction from a low melting point agarose
gel and was recovered by ethanol precipitation. The gapped DNA, thus obtained,
was dissolved in 1 mM Tris-HCl (pH 8.0), 0.1 mM EDTA to a final concentration of 0.35 pmol/[mu]l, and was used for the construction of ss or ds (+)-vectors. Similar procedures were employed after digestion by
Eco
RI and
Apa
I, and this gapped DNA was used for the preparation of ds (-)-vectors.
Double-stranded vectors with 2-OH-Ade were constructed by ligation of the gapped vector and the
annealed oligonucleotide with phosphate groups at both 5'-ends. Specifically, the gapped vector (0.35 pmol) and the DNA
cassette (1.75 pmol) were mixed and joined by T4 DNA ligase in 66 mM Tris-HCl (pH 7.6), 6.6 mM MgCl
2
, 10 mM 2-mercaptoethanol and 200 [mu]M ATP, in a total volume of 20 [mu]l, at 16oC for 24 h. Single-stranded vectors were constructed by joining the gapped
vector and the annealed oligonucleotide obtained with unphosphorylated splint-1 under the same conditions. Proteins were removed by centrifugation with
Ultra-free Probind. The DNA was precipitated with ethanol and was dissolved in 1
mM Tris-HCl (pH 8.0), 0.1 mM EDTA. The vectors were quantitated with a DNA
Fluorometer (Hoechst). Heat denaturation (100oC, 3 min) was carried out for ss (+)-DNA just before transfection.
Mutagenesis experiments
The manipulated DNA was transfected into
E.coli
HB101 by electroporation using a Gene Pulser Transfection Apparatus with a
Pulse Controller (BioRad) or by the calcium chloride method (
11
). To recover plasmid DNA as described below, electroporation was used. The calcium chloride method or electroporation was employed to measure
the survival rate for double-stranded vectors. For single-stranded vectors, electroporation was used.
Recovery of plasmid DNA from transfected cells
Escherichia coli
cells were transfected with the vector by electroporation and were cultured in SOC medium at 37oC for 60 min. An aliquot of the culture was plated onto an agar plate for calculating the transformation efficiency. The remainder was transferred into L-broth medium containing ampicillin and was incubated at 37oC overnight. The plasmid DNA was recovered from the
E.coli
by the alkaline lysis method (
11
).
Mutant screening and sequencing
The recovered plasmid was digested with either
Sal
I or
Afl
II. The digested DNA was transfected again into
E.coli
strain HB101 by electroporation. Undigested DNA was also transfected, and the ratio of (colonies obtained with digested DNA) to (colonies obtained with undigested DNA) was calculated (f
1
).
The
E.coli
colonies obtained with the digested DNA were isolated and the plasmids in the
bacterial cells were recovered by the alkaline lysis method (
12
). The plasmids were heat-denatured and blotted onto a nitrocellulose membrane (Protoran, Schleicher
& Schuell Inc.). After ultraviolet-crosslinking, the DNAs were screened by hybridization using either A-17 or splint-1 and -2 as a probe. We labeled the probes with digoxigenin-2',3'-dideoxyuridine 5'-triphosphate at
their 3'-ends by the DIG Oligonucleotide 3' End Labeling Kit (Boehringer Mannheim) and detected the DNA
with the DIG Nucleic Acid Detection Kit (Boehringer Mannheim). Under our conditions (prehybridization at 60oC, hybridization at 50oC for 2 h, wash at 50oC), A-17 hybridizes with normal DNA but does not with any DNA
containing substitution or frameshift mutations, whereas splint-1 and -2 hybridize with mutated DNA with a substitution or a one-base deletion or addition, as well as with normal DNA. Thus,
the DNA from non-mutant colonies showed positive signals in both hybridization experiments
with A-17 and splint oligonucleotides as probes. On the other hand, the DNA from
mutant colonies showed positive signals only in the experiment with the splint
oligonucleotides. The existence of a mutation was also confirmed by the
digestion of the plasmid DNA with either
Sal
I or
Afl
II and the subsequent agarose gel electrophoresis analysis. The ratio of mutant
colonies to total colonies employed in the hybridization experiments (f
2
) was calculated and multiplied by f
1
to calculate the mutation frequency (MF). Specifically, the MFs were calculated as follows:
MF = f
1
* f
2
where f
1
= (colonies obtained with digested DNA)/ (colonies obtained with undigested
DNA); f
2
= (mutant colonies)/ (colonies with positive signals by labeled splint-1 or -2).
The nucleotide sequences of the mutants were analyzed by sequencing of the
plasmids with primer-1 or -2 and the
Taq
DyeDeoxy Terminator Cycle Sequencing Kit or the ABI PRISM Dye Terminator Cycle Sequencing Kit (Perkin-Elmer) in an Applied Biosystems model 373A DNA sequencer (Perkin-Elmer).
RESULTS
Construction of vectors
To evaluate the mutation inducibility and the mutation spectrum of an oxidative
DNA lesion, 2-OH-Ade, in
E.coli
, the modified base was incorporated into a predetermined position in a unique
restriction enzyme site. We have demonstrated that DNA polymerases incorporate
incorrect nucleotides, in addition to dTMP, opposite 2-OH-Ade
in vitro
(
9
,
10
). In these experiments it was found that DNA polymerases [alpha] and [beta] misinsert dCMP most frequently and that the Klenow fragment of DNA polymerase I misinserts dGMP (
10
). Moreover, 2-OH-Ade in a 5'-TA*A-3' context (A* represents 2-OH-Ade) induces misincorporation
of dAMP
in vitro
(
9
,
10
). Based on these findings, we incorporated the oxidized adenine into 5'-GTCGA*C-3' (a
Sal
I site) and 5'-CTTA*AG-3' (an
Afl
II site) sequences (Fig.
1
).
2-OH-Ade is not cytotoxic
The manipulated DNA was transfected into
E.coli
strain HB101 cells by the CaCl
2
method or by electroporation. When the ds (+)- or (-)-vectors with 2-OH-Ade were introduced into
E.coli
, colonies were obtained with transformation efficiencies similar to that of the
unmodified vector (Table
1
). On the other hand, the vectors with the modified base were transfected as ss
(+)-DNA, and fewer colonies were obtained than with the unmodified vector
(Table
1
). These results indicated that 2-OH-Ade in the ds DNA did not decrease the replication efficiency and
that the base in the ss DNA partially blocked DNA replication. Therefore, 2-OH-Ade is bypassed by DNA polymerases in
E.coli
. This is consistent with the
in vitro
results that the existence of 2-OH-Ade only partially retards DNA synthesis (
9
).
a
Colony-forming units/[mu]g DNA. Relative colony-forming units are represented in parentheses.
b
A* represents 2-OH-Ade.
c
20 ng DNA was transfected by electroporation.
d
5 ng DNA was transfected by electroporation.
e
2.5 ng DNA was transfected by by the CaCl
2
method.
f
0.4 ng DNA was transfected by electroporation.
g
5.3 ng DNA was transfected by the CaCl
2
method.
2-OH-Ade is mutagenic in
E.coli
We introduced 2-OH-Ade residues into unique restriction enzyme sites, and thus we could select mutants as bacterial cells with a
plasmid resistant to the restriction enzyme. The manipulated DNA was
transfected into HB101 cells by electroporation. The plasmid recovered from the
transfected bacteria was incubated with a restriction enzyme. The treated DNA
was then transfected into
E.coli
to obtain mutants. Undigested DNA in the same buffer solution was also
transfected, and the ratio of (colonies obtained with digested DNA) to
(colonies obtained with undigested DNA) was calculated. To correct for the
overestimation of mutation frequencies due to the presence of non-mutants, the ratio of (colonies with a mutation) to (colonies analyzed) was multiplied after a second
screening by dot blot hybridization (see Materials and Methods).
Table
2
shows the mutation frequencies (MFs) of ds and ss vectors with or without 2-OH-Ade. The 2-OH-Ade residues in the
Sal
I site in ds (+)-DNA induced mutations with an efficiency of 0.8%. The maximum MF was
expected to be 50% when the existence of the complementary strand was
considered. Thus, 1.6% of the incorporated 2-OH-Ade residues were estimated to induce misincorporation. This value
was comparable to the MF of 8-OH-Gua in a ds vector in
E.coli
(
4
) and with that of the modified guanine on a chromosome in NIH3T3 cells (
5
,
6
). On the other hand, the MF of the 2-OH-Ade residues in the
Afl
II site was 0.05-0.08%, a 10-fold lower value (Table
2
).
.
Mutation frequencies of 2-OH-Ade residues in
E.coli
a
Unmodified
GTCGA*C
b
Unmodified
CTTA*AG
(GTCGAC)
(CTTAAG)
Single strand (+)
Exp.1
<0.05%
0.59%
<0.02%
0.09%
Exp.2
0.01%
0.74%
0.0004
0.05%
Double strand (+)
Exp.1
0.04%
0.82%
<0.008%
0.05%
Exp.2
0.06%
0.78%
<0.003%
0.08%
Double strand (-)
Exp.1
0.02%
0.23%
0.02%
0.07%
Exp.2
0.02%
0.31%
0.03%
0.13%
a
Mutation frequencies were calculated as described in the Materials and Methods.
b
A* represents 2-OH-Ade.
We expected that the MFs of 2-OH-Ade residues in the ss (+)-vectors were higher than in the ds (+)-vectors because the complementary (-)-strand is not replicated in bacteria, and
because the DNA lesions in the ss DNA appeared to escape from the DNA repair
mechanism much more easily than those in the ds DNA. However, no increase in
the MF was observed with ss (+)-DNA (Table
2
). Considering that the maximum MF was 100% in this case, the actual MFs were
somewhat lower than those of the ds (+)-vectors. This finding suggests weak repair of 2-OH-Ade in bacteria and different `fidelities' (ratios of
misincorporation to proper incorporation of nucleotides) with ds and ss DNA
(see Discussion).
The oxidized adenine in the
Sal
I site of the ds (-)-vector induced mutations with three-fold lower frequency (0.2-0.3%), while the base in the
Afl
II site elicited mutagenic events at a similar level (0.1%, Table
2
). The fact that the MFs of 2-OH-Ade in the same sequence (the
Sal
I site) in different strands were not identical indicates that the strand
bearing the DNA lesion affects the MF. The difference in the MFs with ds (+)- and (-)-vectors reminds us of the different `fidelities' employed
during the replication of the plasmids.
Mutation spectra of 2-OH-Ade
We analyzed the sequences of a total of 28 mutants obtained with ss (+)-DNA containing 2-OH-Ade in the
Sal
I site. The mutation detected most frequently was a -1 deletion ([Delta]A, 23 colonies, Table
3
). The remaining five colonies contained A -> G transitions, which were consistent with the results (incorporation of dCMP opposite 2-OH-Ade) obtained in our previous
in vitro
experiments (
10
). The [Delta]A mutation also was detected most frequently, in a total of 35 analyzed colonies, which were derived from the ss (+)-vector with 2-OH-Ade in the
Afl
II site (Table
3
). As substitution mutations, A -> C transversions in addition to A -> G transitions were observed. We previously reported that DNA
polymerases incorporate dAMP opposite 2-OH-Ade in TA*A sequences (
9
,
10
). However, no colonies with an A -> T transversion were found. The mutation spectra of 2-OH-Ade residues in the ds (+)-vectors resembled those of the base in the ss (+)-DNA (Table
4
). A slightly lower ratio of substitutions was found in the case of the
Afl
II site.
a
A* represents 2-OH-Ade.
b
Four clones contain other mutations: GTCaACTTAG ([Delta]A + G -> A, two clones), GTCGaaCTTAG ([Delta]A + A addition, one clone) and GTCGAtTTAG ([Delta]A + C -> T, one clone).
c
CTTAAa (three clones) and TTAAG ([Delta]C, one clone).
a
A* represents 2-OH-Ade.
b
One clone has a double mutation: GTCggACTTAG ([Delta]A + G addition).
c
CTTAAa (two clones), TTAAG ([Delta]C, two clones) and CTTAAt (one clone).
Surprisingly, different spectra were found to be induced by 2-OH-Ade in the ds (-)-vectors (Table
5
). The oxidized adenine elicited mainly A -> G transitions when it was placed in the
Sal
I site (14 of 26 colonies analyzed). On the other hand, only five colonies
contained [Delta]A mutations. Thus, the ratio of substitutions to deletions was reversed.
This tendency was also observed with 2-OH-Ade in the
Afl
II site in the ds (-)-vector (Table
5
). In this case, A -> T transversions were detected most frequently (nine of 25 colonies
analyzed), while [Delta]A mutants were found in four colonies.
a
A* represents 2-OH-Ade.
b
GCGAC ([Delta]T, three clones), aTCGAC (one clone), GTtGAC (one clone) and GTccGAC (C
addition, one clone).
c
CTTAAa and TTAAG ([Delta]C).
In summary, the expected A -> G transitions in the
Sal
I site and the A -> T transversions in the
Afl
II site were the mutations most frequently observed with the ds (-)-vectors, whereas deletion mutations were predominant with the ds or
ss (+)-vectors.
It was found that other mutations in the same plasmid molecule were elicited in
four colonies of the 18 [Delta]A mutants, which were derived from the ss (+)-vector with 2-OH-Ade in the
Afl
II site (Table
3
). This probability is too high if the two mutations ([Delta]A and another mutation) occurred independently. The additional mutations
were located in the 5'-orientation near 2-OH-Ade. The DNA polymerase(s) may have adopted an
activated conformation(s), which was different from its usual state, to `read' the modified adenine, and
the putative conformer inserted an incorrect nucleotide opposite the normal
base(s). One clone with a double mutation was also detected with the ds (+)-vectors (Table
4
) while no clone had a double mutation in the case of the ds (-)-vectors.
DISCUSSION
The MFs of 2-OH-Ade residues were 0.8% and 0.07% in the
Sal
I and
Afl
II sequences respectively in ds (+)-vectors (Table
2
). These results show the importance of the sequence contexts used in mutation
studies. The MF of 2-OH-Ade in the
Sal
I sequence in the ds (-)-vector was lower than that in the ds (+)-vector (0.3%, Table
2
). The MF of a DNA lesion may be different when it is located in different
strands, even in the same sequence context. Therefore, the MF of 2-OH-Ade in the
Sal
I site was in the range of 0.8-0.3%. This value is similar to that reported for 8-OH-Gua (
4
). We think that 2-OH-Ade is as mutagenic as 8-OH-Gua in bacteria, because the
Afl
II (TA*A) sequence was a rather exceptional sequence, in which the incorporation
of dAMP and the lack of incorporation of dCMP opposite 2-OH-Ade were observed
in vitro
(
9
,
10
). Therefore, the formation of 2-OH-Ade opposite T in DNA will induce mutations at a frequency of ~0.5%.
The MF of thymine glycol or a
cis
-diamminedichloroplatinum (II)-GpG adduct in a ss vector is higher than in a ds vector (
13
,
14
). However, we did not observe an increase in the MF of 2-OH-Ade when ss (+)-vectors were used (Table
2
). The actual MFs were somewhat lower than those of the ds (+)-vectors, since there is no complementary strand. These unexpected results may be explained by the reannealing of the ss DNA. However, the transforming efficiency of
the `ss DNA' without heat denaturation (ds DNA nicked in the complementary
strand) was 100-fold higher than that of the ss DNA with heat treatment (data not shown). Moreover, the cytotoxicities of 2-OH-Ade residues were observed only with ss (+)-vectors (Table
1
). We assume that the fidelities of the DNA replication machinery differ with ds
and ss DNA in cells. Although DNA polymerase III holoenzyme is involved in
replication (
15
), different DNA polymerase(s) may participate in the replication of the ds and
ss plasmid DNAs used in this study. Alternatively, other proteins involved in replication may affect the MFs.
Furthermore, these results suggest weak repair of 2-OH-Ade in bacteria. The increases in the MFs of thymine glycol and the
cis
-diamminedichloroplatinum (II)-GpG adduct in a ss vector (
13
,
14
) are thought to be due to the minimal, if any, recognition of a DNA lesion in ss
DNA by DNA repair proteins. The lack of an increase in the MFs of 2-OH-Ade residues in ss vectors may indicate that the oxidized base is
weakly (or not) repaired, even in ds DNA. In fact, we found that a ds
oligonucleotide with 2-OH-Ade was nicked weakly by a crude extract of
E.coli
(Tsurudome
et al.
, unpublished results). Possibly, the different fidelities for ss and ds
plasmids with 2-OH-Ade and/or the resistance of the base to the repair mechanism in
cells explain(s) the similar MFs in ss (+)-and ds (+)-vectors.
The most frequently induced mutation by 2-OH-Ade in the ds (+)- and ss (+)-vectors was a -1 deletion ([Delta]A, Table
3
and
4
). On the other hand, the expected substitutions were found primarily in the
case of the ds (-)-vectors (Table
5
). The fact that the mutation spectra of the same modified base in the same
sequences were not identical between the (+)- and (-)-vectors, in addition to the lower MF of 2-OH-Ade in the
Sal
I sequence in the ds (-)-vector, suggests that different machinery may participate in the
replication of the ds (+)- and (-)-vectors. pSVK3, which was used for the insertion of the ds
oligonucleotides in this study, contains a unidirectional ColE1 ori (origin of
replication). The (-)-strand of the vectors around the multiple cloning site (Fig.
1
A) is expected to be synthesized as the lagging strand. Therefore, the 2-OH-Ade in the (+)-strand [ds (+)-vectors] was `read' during lagging strand synthesis. On
the other hand, the modified base in the (-)-strand [in the case of ds (-)-vectors] was `read' during leading strand synthesis.
The discrepancies between the ds (+)- and ds (-)-vectors may be due, at least in part, to whether a strand is
synthesized as a leading or lagging strand.
We assume that the observed -1 deletions were the results of the incorporation of dCMP opposite 2-OH-Ade in the
Sal
I site (5'-GA*C-3') and that of dAMP opposite the base in the
Afl
II site (5'-TA*A-3'). The 5'-flanking bases of 2-OH-Ade in the two sequences are
complementary to the putative incoming nucleotides. The slow extension of a 2-OH-Ade:C or 2-OH-Ade:A pair may induce a -1 deletion, which is thought to be mediated by a
loop-out mechanism (
16
-
18
). Figure
3
shows the putative pathways of an A -> G transition and a -1 deletion when 2-OH-Ade is located in the
Sal
I site. In this model, the product with 3'-C opposite 2-OH-Ade (product
2
, Fig.
3
) is the common intermediate of the A -> G transition and the deletion. When extension from the 2-OH-Ade:C pair occurs readily, the A -> G transition is formed (
2
->
3
, Fig.
3
). However, when the extension is slow and a DNA polymerase elongates the looped-out intermediate, the -1 deletion is generated (
4
->
5
, Fig.
3
). The observed -1 deletions at the
Afl
II site would be explained by a similar mechanism. Therefore, extensions from
the 2-OH-Ade:C pair in the
Sal
I site and the 2-OH-Ade:A pair in the
Afl
II site may be faster in leading strand synthesis than in lagging strand
synthesis, because A -> G transitions or A -> T transversions are `increased' in the case of ds (-)-vectors (Table
5
). This assumption may be supported by the fact that the ratios of ([Delta]A + A -> G) in the
Sal
I site and of ([Delta]A + A -> T) in the
Afl
II site to all of the observed targeted mutations were similar in the
experiments with the three different plasmids (~100% for the
Sal
I site and 60-80% for the
Afl
II site, Table
3
-
5
).
ACKNOWLEDGEMENTS
We thank Drs Tadaaki Ohgi and Toshihiro Ueda of Nippon Shinyaku Co. for
synthesis of the phosphoramidite derivative of 2-OH-Ade. This work was supported in part by Grants-in-Aid for Scientific Research on Priority Areas from the
Ministry of Education, Science, Sports and Culture of Japan (Nos. 05270102 and
08264237).
2 Kasai, H. and Nishimura, S. (1984) Nucleic Acids Res., 12, 2137-2145.MEDLINE Abstract
3 Wood, M. L., Dizdaroglu, M., Gajewski, E. and Essigmann, J. M. (1990) Biochemistry, 29, 7024-7032.
4 Cheng, K. C., Cahill, D. S., Kasai, H., Nishimura, S. and Loeb, L. A. (1992) J. Biol. Chem., 267, 166-172.
5 Kamiya, H., Miura, K., Ishikawa, H., Inoue, H., Nishimura, S. and Ohtsuka, E. (1992) Cancer Res., 52, 3483-3485.MEDLINE Abstract
6 Kamiya, H., Murata-Kamiya, N., Koizume, S., Inoue, H., Nishimura, S. and Ohtsuka, E. (1995) Carcinogenesis, 16, 883-889.MEDLINE Abstract
7 Moriya, M. (1993) Proc. Natl. Acad. USA, 90, 1122-1126.
8 Kamiya, H. and Kasai, H. (1995) J. Biol. Chem., 270, 19446-19450.MEDLINE Abstract
9 Kamiya, H., Ueda, T., Ohgi, T., Matsukage, A. and Kasai, H. (1995) Nucleic Acids Res., 23, 2893-2899.MEDLINE Abstract
10 Kamiya, H. and Kasai, H. (1996) FEBS Letts, 391, 113-116.
11 Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor.
12 Seela, F., Mertens, R. and Kazimierczuk, Z (1992) Helv. Chim. Acta, 75, 2298-2306.
13 Basu, A. K., Loechler, E. L., Leadon, S. A. and Essigmann, J. M. (1989) Proc. Natl. Acad. Sc. USA, 86, 7677-7681.
14 Bradley, L. J. N., Yarema, K. J., Lippard, S. J. and Essigmann, J. M. (1993) Biochemistry, 32, 982-988.
15 Maki, H., Maki, S. and Kornberg, A. (1988) J. Biol. Chem., 263, 6570-6578.MEDLINE Abstract
16 Kunkel, T. A. (1990) Biochemistry, 29, 8003-8011.
17 Lambert, I. B., Napolitano, R. L. and Fuchs, R. P. P. (1992) Proc. Natl. Acad. Sci. USA, 89, 1310-1314.
18 Shibutani, S. and Grollman, A. P. (1993) J. Biol. Chem., 268, 11703-11710.MEDLINE Abstract
19 Veaute, X. and Fuchs, R. P. P. (1993) Science, 261, 598-600.MEDLINE Abstract
20 Tajiri, T., Maki, H. and Sekiguchi, M. (1995) Mutat. Res., 336, 257-267.MEDLINE Abstract
21 Schaaper, R. M. and Dunn, R. L. (1991) Genomics, 129, 317-326.
22 Halliday, J. A. and Glickman, B. W. (1991) Mutat. Res., 250, 55-71.
23 Sargentini, N. J. and Smith, K. C. (1994) Mutat. Res., 309, 147-163.
24 Ono, T., Negishi, K. and Hayatsu, H. (1995) Mutat. Res., 326, 175-183.MEDLINE Abstract
25 Hauser, J., Seidman, M. M., Sidur, K. and Dixon, K. (1986) Mol. Cell. Biol., 6, 277-285.MEDLINE Abstract
26 Keyse, S. M., Amaudruz, F. and Tyrell, R. M. (1988) Mol. Cell. Biol., 8, 5425-5431.
27 Moraes, E. C., Keyse, S. M. and Tyrell, R. M. (1990) Carcinogenesis, 11, 283-293.
