Mutagenic effects of 2-hydroxy-dATP on replication in a HeLa extract: induction of substitution and deletion mutations

Abstract

The mutagenicity of an oxidized form of dATP, 2-hydroxydeoxyadenosine 5′-triphosphate (2-OH-dATP), was examined using an SV40 origin-dependent in vitro replication system with a HeLa extract. 2-OH-dATP induced mutations in a dose-dependent manner and elicited substitution and deletion mutations. Of the substitutions, a G·C→A·T transition including a tandem (CC→TT) mutation was mainly observed. This result agrees with our previous observation that mammalian DNA polymerase α misincorporates the oxidized nucleotide opposite C, but is in contrast to the finding that 2-OH-dATP elicits G·C→T·A transversions in Escherichia coli. This type of mutation was also elicited, but to a lesser extent. Interestingly, the mutagenicity of 2-OH-dATP was enhanced in the presence of 2-hydroxydeoxyadenosine 5′-diphosphate, an inhibitor of the MTH1 protein, suggesting that this protein functions in the hydrolysis of 2-OH-dATP in the replication reaction mixture, and probably in living cells. These results indicate that 2-OH-dATP is mutagenic and that its mutagenicity is suppressed by the MTH1 protein in mammalian cells.

INTRODUCTION

Reactive oxygen species (ROS) are generated endogenously by normal oxygen metabolism, and are also produced by many environmental mutagens and carcinogens. For these reasons, DNA oxidation is believed to be a very important source of mutations and to be one of the causative factors of carcinogenesis and aging (1–3). Several lines of evidence indicate that the oxidation of DNA precursors in the nucleotide pool is another important endogenous source in the mutation process. The mutT mutant of Escherichia coli, in which A·T→C·G transversions are induced with high frequency (4), lacks the ability to hydrolyze an oxidized form of dGTP, 8-hydroxy-2′-deoxyguanosine 5′-triphosphate (8-OH-dGTP) (5), indicating the incorporation of this oxidized nucleotide by DNA polymerases and its importance as a source of mutations. In terms of the accumulation of 8-hydroxyguanine in DNA, the contributions of the 8-OH-dGTP from the nucleotide pool and the direct oxidation of G bases in DNA were almost equal (4). Moreover, the presence of mammalian homologs of MutT (MTH1 proteins) supports this speculation (6). Indeed, a greater number of tumors were formed in the lungs, livers and stomachs of MTH1-deficient mice than wild-type mice (7). In addition, we found that the damaged nucleotides incorporated into bacterial cells elicited chromosomal gene mutations (8,9), providing direct evidence that damaged DNA precursors act as mutagens.

2-Hydroxy-2′-deoxyadenosine and 2-hydroxy-2′-deoxyadenosine 5′-triphosphate (2-OH-dATP) are produced by treating dA and dATP, respectively, with ROS-generating reagents (10–12). Nunoshiba et al. reported that one of the mutations found in an E.coli strain lacking superoxide dismutases and a repressor for iron-uptake systems might be induced by 2-OH-dATP (13). This oxidized form of dATP, 2-OH-dATP, is more mutagenic than 8-OH-dGTP when both are directly introduced into E.coli cells (8). Furthermore, we recently found that the human MTH1 protein hydrolyzes 2-OH-dATP more efficiently than 8-OH-dGTP (14). These findings suggest that 2-OH-dATP is also an important endogenous mutagen, and that the formation of 2-OH-dATP causes the tumor formation in the MTH1-deficient mouse (7). Thus, it is of great interest to examine the miscoding properties of 2-OH-dATP during DNA synthesis, particularly in mammalian cells.

Mammalian DNA polymerase (pol) α misincorporates 2-OH-dATP opposite C on a template DNA in vitro (10). This result suggests that the spectrum of mutations caused by 2-OH-dATP in mammalian cells only involves G·C→A·T transitions. However, E.coli DNA polymerases misincorporate 2-OH-dATP opposite G, in contrast to mammalian pol α (15,16). 2-OH-dATP induces G·C→T·A transversions in the chromosome when the oxidized nucleotide is introduced directly into E.coli cells (8). This raised the question of whether the actual mutation spectrum of 2-OH-dATP in mammalian cells is a G·C→A·T transition or a G·C→T·A transversion.

The SV40 origin-dependent in vitro replication system is a good model for replication in living cells. All of the factors required for bidirectional replication of double-stranded DNA, except for the SV40 large T antigen, are provided by the host cell extract (17,18). HeLa extracts are frequently used as a source of replication enzymes. The extract contains at least two DNA pols (α and δ). In addition, a specialized DNA pol, the XP-V protein (DNA pol η), is probably present, since it was purified from a HeLa extract (19).

We now report that 2-OH-dATP induces mutations during in vitro replication with a HeLa extract. This nucleotide induced substitution and deletion mutations. Of the substitutions, a G·C→A·T transition including a tandem (CC→TT) mutation was mainly observed. In addition, the mutagenicity of 2-OH-dATP was enhanced in the presence of 2-hydroxy-2′-deoxyadenosine 5′-diphosphate (2-OH-dADP), an inhibitor of the MTH1 protein, suggesting the role of this protein in the hydrolysis of 2-OH-dATP in the replication reaction mixture, and probably also in living cells.

Materials and methods

Materials

The FPLC-grade nucleoside triphosphates used in the replication reactions were from Amersham Biosciences (Piscataway, NJ). 2-OH-dATP was prepared by the treatment of dATP with Fe(II)–EDTA–O₂, and was purified by HPLC as described (10). The purified nucleotide was eluted as a single peak in both reverse-phase and anion-exchange HPLC (data not shown). The nucleotide purified by this procedure was eluted at the same time as the authentic 2-hydroxy-dA in reverse-phase HPLC after dephosphorylation (10). 2-OH-dADP was prepared from dADP by similar procedures. The SV40 large T antigen and the HeLa extract were purchased from Chimerx (Milwaukee, WI). Purified oligonucleotides were from Hokkaido System Science (Sapporo, Japan) and Sigma Genosys Japan (Ishikari, Japan). The E.coli strain KS40 [lacZ(am), CA7070, lacY1, hsdR, hsdM, Δ(araABC-leu)7679, galU, galK, rpsL, thi, gyrA]/pOF105 (20) was provided by Dr Tatsuo Nunoshiba, of Tohoku University, and was used as an indicator strain of the supF mutants.

In vitro DNA replication

The plasmid with the supF gene and the SV40 origin was replicated with the HeLa extract by an established method, as described (21,22). The double-stranded plasmid, pSVKAM189 (16) (40 ng), was replicated by the HeLa extract (287 µg) in a buffer (total 25 µl) containing 30 mM HEPES (pH 7.5), 7 mM MgCl₂, 0.5 mM dithiothreitol, 4 mM ATP, 100 µM each of dATP, dGTP, dCTP, dTTP, 50 µM each of GTP, CTP, UTP, 40 mM phosphocreatine, 0.625 U of creatine phosphokinase and 1 µg of large T antigen, with or without a damaged nucleotide (2-OH-dATP and/or 2-OH-dADP). The mixture was incubated at 37°C for 4 h, and the reaction was terminated by the addition of EDTA to a final concentration of 15 mM. The proteins were removed by phenol, phenol–chloroform and chloroform extractions, and the DNA was precipitated with ethanol in the presence of 10 µg of tRNA. The recovered DNA was treated with DpnI to digest the unreplicated plasmids. After the removal of proteins by passage through a Micropure EZ device (Millipore, Bedford, MA), the DNA was purified by ethanol precipitation.

Mutagenesis experiments

The DNAs obtained after the in vitro replication reactions were transfected into E.coli KS40/pOF105 cells by electroporation, using a Gene Pulser II Transfection Apparatus with a Pulse Controller II (Bio-Rad, Hercules, CA). The mutant frequency was calculated according to the number of colonies on a Luria–Bertani agar plate containing nalidixic acid (50 µg/ml), streptomycin (100 µg/ml), ampicillin (150 µg/ml) and chloramphenicol (30 µg/ml), and the number of colonies on an agar plate containing ampicillin and chloramphenicol, as described (20).

The nucleotide sequences of the supF gene were analyzed by sequencing as described previously (23), using an ABI PRISM Big Dye Terminator Cycle Sequencing Kit, and an ABI model 377 DNA Sequencer (Applera, Norwalk, CT).

RESULTS

Increased mutant frequency by the addition of 2-OH-dATP

The mutagenic potential of an oxidized form of dATP, 2-OH-dATP, was examined by in vitro replication reactions using a HeLa extract, as a good model for replication in living cells. The supF gene was chosen as the mutagenesis target, and the plasmid DNA containing the gene, pSVKAM189 (16), was added as the template to the replication reaction mixtures. The replicated DNA was purified and then transfected into the indicator strain, KS40/pOF105 (20).

When only the four unmodified dNTPs were present in the reaction mixture, the observed mutant frequency was about 3.5 × 10^–5 (Fig. 1). The mutant frequency increased when 100, 200 or 400 µM 2-OH-dATP was present during the DNA synthesis, in a dose-dependent manner. With the addition of 400 µM of 2-OH-dATP, the mutant frequency reached 6.4 × 10^–5 (Fig. 1). Thus, the mutations corresponding to the difference, 2.9 × 10^–5, were elicited by this oxidized dATP. These results suggest that 2-OH-dATP is incorporated by mammalian DNA pol(s) and is a potential mutagen in human cells. When the extract was omitted and this mocked ‘replicated’ DNA was transfected into the indicator bacteria, no supF mutant colony was detected (data not shown), excluding the possibility that the increased mutant frequency was derived from mutations elicited by 2-OH-dATP carried over to the bacterial cells.

Figure 1.

Mutant frequencies after replication in HeLa extracts. SV40 origin-dependent in vitro replication reactions of a plasmid containing the supF gene, with or without damaged nucleotides, and the work-up were conducted as described in the Materials and Methods. Mutant frequencies were calculated according to the numbers of supF mutant colonies on an agar plate containing nalidixic acid, streptomycin, ampicillin and chloramphenicol, and the numbers of colonies on an agar plate containing ampicillin and chloramphenicol. The in vitro replication experiments were carried out two to five times. Data are expressed as mean ± standard deviation. Asterisks indicate a significant difference, with P < 0.05 (*) or P < 0.01 (**).

Enhancement of mutant frequency by the addition of 2-OH-dADP

2-OH-dATP is a substrate of the human MTH1 protein (14), and this protein may hydrolyze the oxidized dATP during the replication reaction in vitro. The corresponding diphosphate derivative, 2-OH-dADP, is a potent inhibitor of the MTH1 protein, and its inhibition constant against the hydrolysis of 2-OH-dATP is ∼0.2 µM (14). We thus examined the effect of the addition of 2-OH-dADP into the replication reaction mixture on the mutations induced by 2-OH-dATP.

When 200 µM of 2-OH-dADP was added together with 200 µM of 2-OH-dATP into the replication reaction mixture, the mutant frequency was enhanced and reached 8.1 × 10^–5 (Fig. 1). This value was higher than the mutant frequencies elicited by 200 and 400 µM 2-OH-dATP. On the other hand, as expected, the addition of 200 µM 2-OH-dADP by itself increased the mutant frequency only slightly (Fig. 1). Thus, the human MTH1 protein appeared to degrade 2-OH-dATP during replication, and to contribute to the prevention of the mutagenesis by this nucleotide.

Mutation spectrum of 2-OH-dATP

We analyzed the sequences of the supF genes in 61 and 52 colonies, obtained with the control and 2-OH-dATP experiments, respectively. As shown in Figure 2 and Table 1, 2-OH-dATP elicited various types of mutations. Overall, the mutation spectrum of 2-OH-dATP resembled that of the control experiment without 2-OH-dATP, although the total mutant frequency was increased. We then multiplied the total mutant frequency and the ratio of each type of mutation, and subtracted the frequencies of each type of mutation for the control experiment from those induced by the 2-OH-dATP treatment. Several types of mutations were characteristically elicited (Table 1).

Figure 2.

The overall distribution of the point mutations detected in the supF gene. The sequence of the upper strand of the plasmid is shown. The mutations obtained with the control experiments and those induced by 400 µM 2-OH-dATP are shown below and above the sequence, respectively. The symbol Δ represents a deletion. Tandem two base substitutions or deletions are shown in parentheses. The symbol Δ below the midpoint of positions 65–66 represents the deletion of either of the G residues located at positions 64–67. The symbol Δ above and below position 136 represents the deletion of either of the C residues located at positions 134–138. The symbol (ΔΔ) below positions 97–99 represents the deletion of two of the three A residues located at positions 97–99. One supF mutant colony obtained with the 2-OH-dATP experiment contained two G→A mutations at positions 61 and 67. One supF mutant colony obtained with the 2-OH-dATP experiment contained one G→A mutation and one C→A mutation at positions 65 and 70, respectively. The anticodon corresponds to positions 95–97.

Table 1. Spectrum of mutations induced by 2-OH-dATP in a HeLa extract.

	Control Cases found (%)	MF (×10–6)a	400 µM 2-OH-dATP Cases found (%)	MF (×10–6)a	Subtracted MF (×10–6)b
Single base substitution
Transition<
G·C→A·T	9(15)	5.1	15(29)	18.3	13.2
A·T→G·C	3(5)	1.7	1(2)	1.2	–0.5
Transversion
G·C→T·A	5(8)	2.9	5(10)	6.1	3.2
A·T→C·G	4(7)	2.3	1(2)	1.2	–1.1
A·T→T·A	1(2)	0.6	1(2)	1.2	0.6
G·C→C·G	0(0)	0.0	2(4)	2.4	2.4
Tandem base substitution
CC→TT	5(8)	2.9	5(10)	6.1	3.2
GT→AG	0(0)	0.0	1(2)	1.2	1.2
–1 Deletion
ΔA·T	3(5)	1.7	6(12)	7.3	5.6
ΔG·C	21(34)	12.0	13(25)	15.9	3.9
Large deletion	6(10)	3.4	0(0)	0.0	–3.4
Others	4(7)c	2.3	4(8)d	4.9	2.6
Total	61(100)	34.9	52(100)e	63.5	28.6

^aMF means mutation frequency.

^bFrequencies of each type of control mutation are subtracted from those induced by 2-OH-dATP.

^cΔAA (two cases), ΔTC (one case) and TT→C (one case).

^dΔCC (one case), ΔTC (one case), GA→T (one case) and TT→C (one case).

^eA total of 52 supF mutant colonies were analyzed. Two colonies contained two base substitutions at non-tandem sites (see Fig. 2 legend).

With the 400 µM 2-OH-dATP treatment, transition and transversion mutations were induced at mutation frequencies (MFs) of 12.7 × 10^–6 and 5.1 × 10^–6, respectively (Table 1). Thus, 2-OH-dATP induced ∼2.5-fold more transitions than transversions. The expected substitutions, G·C→A·T transitions and G·C→T·A transversions, were induced by 13.2 × 10^–6 and 3.2 × 10^–6, respectively (Table 1). G·C→C·G transversions were detected with the extract treated with 2-OH-dATP, but not with the control. No increase in the MF was observed or evident for the other types of single base substitutions. To summarize, 2-OH-dATP primarily induced G·C→A·T transitions, and elicited G·C→T·A transversions to a lesser extent. These results suggest that 2-OH-dATP was misincorporated opposite C and G during the replication reaction conducted by the HeLa extract (see Discussion). Interestingly, tandem CC→TT (GG→AA) base substitutions seemed to be increased by 2-OH-dATP (Table 1). This result may be caused by the incorporation of 2-OH-dATP opposite tandem template sites. When we calculated this tandem CC→TT mutation as double G·C→A·T mutations, the predominance of the G·C→A·T transitions over the G·C→T·A transversions became more obvious.

The addition of 2-OH-dATP also elicited single base deletion mutations (Table 1). Their frequency was 9.5 × 10^–6 and was higher than the frequency of transversions (5.1 × 10^–6). The frequency of induced deletions was similar for A·T and G·C sites.

No major mutational hotspot was observed, but some minor hotspots existed (Fig. 2). A minor G·C→A·T transition hotspot was formed at position 130. This position was also a G·C→A·T transition hotspot in the control experiment. Positions 95–97 correspond to those of the anticodon in the encoded tRNA, and 9 out of the 54 mutations were found at these sites. Three out of the five tandem CC→TT substitutions elicited by 2-OH-dATP were observed at positions 84 and 85, and only one case out of the five same mutation types was found for the control. Interestingly, when a single-stranded moiety (gap) was filled with E.coli pol III in vitro in the presence of 2-OH-dATP, positions 95 and 130 were the major and minor hotspots of the induced G→T transversions, respectively (16).

DISCUSSION

An oxidized form of dATP, 2-OH-dATP, induces G·C→T·A transversions with high frequency in a chromosomal gene in E.coli (8). This in vivo result is supported by in vitro experiments, which revealed that the α subunit of E.coli DNA pol III incorporates 2-OH-dATP opposite G and T (15). On the other hand, mammalian DNA pol α inserts 2-OH-dATP opposite C and T (10). Recently, we observed that human DNA pol η, one of the specialized DNA polymerases (19), incorporates 2-OH-dATP opposite G and T (24). These results suggest that the misincorporation mode of 2-OH-dATP is DNA pol-specific. In addition, the incorporation of 2-OH-dATP by DNA pol(s) seems to be the major pathway for the accumulation of 2-hydroxyadenine (2-OH-Ade) bases in DNA and to be predominant over the direct oxidation of A bases, because the yields of 2-OH-Ade were much higher in the monomeric form (nucleoside and nucleotide) than in DNA in vitro (10). Thus, the mutagenic potentials of this damaged nucleotide in mammalian cells are of great interest.

In the present study, the well established, SV40 origin-dependent in vitro replication system, which has been used as a good model for replication in living cells, was employed to examine the mutagenicity of 2-OH-dATP. When 2-OH-dATP was present in the replication reaction mixture, we observed an increased mutant frequency (Fig. 1). In particular, G·C→A·T transition mutations were elicited by this nucleotide (Table 1). This finding implies that either 2-OH-dATP was incorporated opposite C, and then dTTP was inserted opposite the incorporated 2-OH-Ade residue during the next round of replication (Fig. 3, left pathway), or 2-OH-dATP was incorporated opposite G, and then dATP was inserted opposite the incorporated 2-OH-Ade residue during the next round of replication. Since 2-OH-Ade residues in plasmid vectors are ‘read’ as A with >99% probability in both mammalian and E.coli cells (25,26), the former explanation is most likely. Thus, it is probable that the DNA pol(s) present in the HeLa extract incorporated 2-OH-dATP opposite the C residues in the DNA. This conclusion agrees with our previous finding that the mammalian DNA pol α inserts 2-OH-dATP opposite the C residues in DNA (10). Since the eukaryotic replicating DNA pols (α, δ and ε) belong to the same B-family, the induction of the G·C→A·T transition mutations may be conducted by these pols. Likewise, the G·C→T·A transversion mutations elicited by 2-OH-dATP (Table 1) can be explained by a scenario in which 2-OH-dATP was incorporated opposite G, and then dTTP was inserted opposite the incorporated 2-OH-Ade residue during the next round of replication (Fig. 3, right pathway). This type of mutation may be related to the recent finding that human DNA pol η incorporated 2-OH-dATP opposite G (24). As discussed in the previous paper (27), the DNA pol-specific mispairing properties of 2-OH-dATP may be derived from the presence of the hydrophobicity-dependent enol-keto equilibrium of 2-OH-Ade and from the possibility of adopting the syn conformation.

Figure 3.

Proposed model for G·C→A·T and G·C→T·A mutations induced by 2-OH-dATP during in vitro replication reactions.

The addition of 2-OH-dATP also elicited single base deletion mutations (Table 1). The frequency of induced deletions was similar for A·T and G·C sites. 2-OH-Ade in DNA causes deletion mutations in both mammalian and E.coli cells (25,26). The deletions observed in this study may arise from incorporation of 2-OH-dATP, and subsequent induction of deletion by 2-OH-Ade in DNA.

In the absence of the large T antigen, the colonies formed on the titer plate were reduced to near zero (data not shown). Thus, most DNA molecules after the replication reaction and the work-up were ‘replicated’ during the incubation. However, ∼80% of DNA appeared to be the product of one round of replication because E.coli colonies after MboI treatment recognizing unmethylated GATC sequence decreased by ∼20%. This result suggests that the bacterial (mismatch) repair may contribute to removal of the incorporated 2-OH-Ade residues and may affect mutant frequency and mutation spectrum.

We added previously 2-OH-dATP in in vitro gap-filling reactions conducted by the E.coli pol III holoenzyme and the treated plasmid was transfected into repair-deficient E.coli strains (16). Slight increase in the mutant frequency was observed when alkA and mutY strains were used as hosts. Thus, the AlkA and MutY proteins in the indicator E.coli strain, KS40, might affect the mutation spectrum observed in this study, because the G·C→T·A transversion is the main mutation caused by 2-OH-dATP in E.coli. However, double-stranded oligonucleotides containing 2-OH-Ade were very poor substrates for the purified AlkA (H.Kamiya, H.Kasai, H.Terato and H.Ide, unpublished results) and MutY (28) proteins. Other DNA repair enzyme(s) might remove 2-OH-Ade opposite G selectively, and this putative removal might affect the spectrum observed in this study. Although we cannot exclude this possibility completely, it appears to be low since E.coli cell-free extract contains no nicking activity for double-stranded oligonucleotides containing 2-OH-Ade (29).

The human MTH1 protein hydrolyzes 2-OH-dATP more efficiently than 8-OH-dGTP (14). This protein might degrade 2-OH-dATP during the incubation, although the presence of this protein in the extract is unclear. As shown in Figure 1, 200 µM of 2-OH-dADP, a known potent inhibitor of the MTH1 protein (14), enhanced the mutagenicity of 2-OH-dATP (200 µM). This effect could be interpreted as (i) the inhibition of MTH1 or a similar protein specific for 2-OH-dATP, (ii) the inhibition of a non-specific nucleotidase or (iii) the formation of 2-OH-dATP from the added 2-OH-dADP by a phosphorylation reaction. The second possibility is unlikely, because a higher amount (4 mM) of ATP was present, and because the addition of CTP did not increase the damaged nucleotide-induced mutant frequency (data not shown). The third possibility may be supported by the finding that the eukaryotic nucleoside diphosphate kinase can convert 2-OH-dADP to 2-OH-dATP (30). However, this possibility is also unlikely, because the addition of 200 µM of 2-OH-dADP plus 200 µM of 2-OH-dATP enhanced the mutant frequency more strongly than 400 µM of 2-OH-dATP (Fig. 1). Thus, the first explanation is most probable, and this result suggests that a similar situation is present in living cells: the MTH1 protein (and/or a similar enzyme) prevents the mutagenesis induced by 2-OH-dATP. Interestingly, the E.coli Orf135 protein hydrolyzes 2-OH-dATP in vitro (31), and the spontaneous and H₂O₂-induced MFs in an orf135^– E.coli strain were higher than those in an orf135⁺ strain (H.Kamiya et al., unpublished results). Thus, bacterial and mammalian cells apparently share a similar mechanism to prevent mutagenesis by 2-OH-dATP.

In this study, we found that 2-OH-dATP was mutagenic during in vitro replication using a HeLa extract. This oxidized nucleotide caused G·C→A·T transitions, and G·C→T·A transversions to a lesser extent, suggesting misincorporations opposite C and G, respectively. These results agree with our previous finding, which showed that mammalian DNA pol α inserts 2-OH-dATP opposite C and T in DNA (10). The mutagenic potentials of 2-OH-dATP in living mammalian cells are of great interest. Experiments along this line are in progress.