Translesional synthesis on a DNA template containing N²-methyl-2′-deoxyguanosine catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I

Abstract

Formaldehyde is produced in most living systems and is present in the environment. Evidence that formaldehyde causes cancer in experimental animals infers that it may be a carcinogenic hazard to humans. Formaldehyde reacts with the exocyclic amino group of deoxyguanosine, resulting in the formation of N²-methyl-2′-deoxyguanosine (N²-Me-dG) via reduction of the Schiff base. The same reaction is likely to occur in living cells, because cells contain endogenous reductants such as ascorbic acid and gluthathione. To explore the miscoding properties of formaldehyde-derived DNA adducts a site-specifically modified oligodeoxynucleotide containing a N²-Me-dG was prepared and used as the template in primer extension reactions catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I. The primer extension reaction was slightly stalled one base before the N²-Me-dG lesion, but DNA synthesis past this lesion was readily completed. The fully extended products were analyzed to quantify the miscoding specificities of N²-Me-dG. Preferential incorporation of dCMP, the correct base, opposite the lesion was observed, along with small amounts of misincorporation of dTMP (9.4%). No deletions were detected. Steady-state kinetic studies indicated that the frequency of nucleotide insertion for dTMP was only 1.2 times lower than for dCMP and the frequency of chain extension from the 3′-terminus of a dT:N²-Me-dG pair was only 2.1 times lower than from a dC:N²-Me-dG pair. We conclude that N²-Me-dG is a miscoding lesion capable of generating G→A transition mutations.

INTRODUCTION

N²-methyl-2′-deoxyguanosine (N²-Me-dG) is one of the formaldehyde-derived DNA adducts. Formaldehyde is abundantly produced in industry and is used as a fixative and preservative for medical and biological specimens. The annual world production is apparently 12 000 000 tons. Therefore, humans are occupationally exposed to formaldehyde. Formaldehyde is also produced in most living systems, where it is derived from the metabolism of serine, choline and methionine (1). Common non-occupational sources of exposure include vehicle emissions, some building materials, food, tobacco smoke and its use as a disinfectant. Levels of formaldehyde in the indoor air of houses are in the range 0.02–0.06 mg/m³ (2,3).

Evidence that formaldehyde causes cancer in experimental animals infers that it may be a carcinogenic hazard to humans. Several studies in which formaldehyde was administered to rats by inhalation showed carcinogenicity, particularly induction of squamous cell carcinomas of the nasal cavities. In rats, administered formaldehyde in the drinking water increased incidences of forestomach papillomas, leukemias and gastrointestinal tract tumors were reported (4).

Aldehydes are known to produce DNA–protein and/or DNA–DNA crosslinks (5–8) and mainly react with dG in the exocyclic amino group producing N-alkyl-dG via an unstable Schiff base, which can be stabilized by reduction. For instance, acetaldehyde is known to form stable N²-ethyl-2′-deoxyguanosine (N²-Et-dG) (9,10). This adduct has been detected in granulocyte and lymphocyte DNA of alcoholic patients and in human buccal cells exposed to acetaldehyde (11,12). In addition, N²-Et-dG can be detected in human urine (9). Due to the structural similarities of acetaldehyde and formaldehyde, formaldehyde should produce the N²-Me-dG adduct in cells.

In this study to explore the miscoding properties of formaldehyde-derived DNA adducts site-specifically modified oligodeoxynucleotides containing a N²-Me-dG were prepared and used as DNA templates in primer extension reactions catalyzed by the Klenow fragment of Escherichia coli DNA polymerase I. Using these templates, miscoding specificities of N²-Me-dG were investigated with an in vitro experimental system that can quantify base substitutions and deletions (13) and determine steady-state kinetics. The results suggest that the N²-Me-dG adduct is a miscoding lesion capable of generating G→A transition mutations.

MATERIALS AND METHODS

Materials

[γ-³²P]ATP (sp. act. >6000 Ci/mmol) was obtained from Amersham Corp. The Klenow fragment of E.coli DNA polymerase I (exo⁺), T4 polynucleotide kinase, T4 DNA ligase and EcoRI restriction endonuclease were purchased from Takara Shuzo (Kyoto, Japan). Micrococcal nuclease and spleen phoshodiesterase were obtained from Worthington Biochemical Corp. Alkaline phosphatase was purchased from Sigma Chemical Corp. A Waters 996 HPLC instrument equipped with a photodiode array detector was used for separation and purification of modified and unmodified oligodeoxynucleotides. NMR spectroscopy was performed on a Bruker 600 MHz instrument. Mass spectroscopy was performed on a TSQ 7000 Triple-Stage Quadrupole (Finnigan MAT, San Jose).

Synthesis of N²-Me-dGTP and N²-Me-dG

N²-Me-dGTP was purchased from Trilink Biochemical Corp. (San Diego, CA) or synthesized by ourselves using dGTP and formalin (37% formaldehyde and 10–15% methanol) as the starting materials. dGTP (3 mg/reaction) was reacted overnight with 3.7% (v/v) formalin in 1 ml of 0.2 mM sodium acetate buffer (pH 4.5) and reduced by adding 1 mg NaBH₃CN (Fig. 1A). Five batches of the reactant were gathered and passed through a Sep-pak C₁₈ cartridge (Waters). The cartridge was washed with 5 ml of 50 mM triethylamine acetate (pH 7.0) and eluted with 5 ml of methanol. The extract was evaporated to dryness and redissolved in 100 µl of pure water. The sample containing N²-Me-dGTP was fractionated by HPLC on a reversed phase Wakosil 5C18-200T column (0.46 × 25 cm; Wako), eluted over 40 min at a flow rate of 1.0 ml/min with a linear gradient of 50 mM triethylamine acetate (pH 7.0) containing 0–40% acetonitrile. The yield of N²-Me-dGTP was ∼15% of the starting material dGTP. About 10 mg of the N²-Me-dGTP was prepared and identified using ³¹P-NMR, ¹H-NMR and ESI LC/MS. ³¹P-NMR and ¹H-NMR spectra were recorded in D₂O using orthophosphoric acid and 3-(trimethylsilyl)propionic acid-d₄ sodium salt as the internal standards, respectively. N²-Me-dGTP: ³¹P-NMR (D₂O) δ (p.p.m.), 4.50 (d, J = 18.3 Hz), 10.05 (d, J = 18.2 Hz), 19.66 (t); ¹H-NMR (D₂O) δ (p.p.m.), 2.51 (m, 1H, 2′-Ha), 2.86 (m, 1H, 2′-Hb), 2.97 (s, 3H, NHCH₃), 4.20 (m, 2H, 5′-H), 4.26 (m, 1H, 4′-H), 4.80 (overlapping with DOH, 3′-H), 6.41 (t, 1H, J = 6.8 Hz, 1′-H), 8.08 (s, 1H, 8-H). Using negative ion ESI LC/MS/MS the base peak was at m/z 520 (M–1).

Figure 1.

(A) Formation of N²-Me-dGTP via the Schiff base. dGTP was reacted with formaldehyde in sodium acetate buffer, pH 4.5, and reduced by NaBH₃CN as described in Materials and Methods. (B) HPLC separation of N²-Me-dGTP. N²-Me-dGTP was isolated from dGTP on a reversed phase Wakopak Wakosil 5C18-200T column (0.46 × 25 cm) using a linear gradient of 50 mM triethylamine acetate, pH 7.0, containing 0–40% acetonitrile over 40 min and a flow rate of 1.0 ml/min.

N²-Me-dG was prepared essentially by the same method as described above except that dG was the starting material instead of dGTP. About 1 mg of N²-Me-dG was prepared and identified using ¹H-NMR and ESI LC/MS/MS. N²-Me-dG: ¹H-NMR, δ 10.49 (brs, NH), 7.89 (s, H8), 6.30 (brs, 1H, J = 5.2 Hz, N₂H), 6.17 (t, J = 7.2 Hz, 1′-H), 5.26 (d, J = 3.6 Hz, 3′-OH), 4.85 (t, J = 6.0 Hz, 5′-OH), 4.36 (m, 3′-H), 3.81 (m, 4′-H), 3.55 and 3.50 (m, 5′-H), 3.30 (d, -CH₃), 2.20 and 2.60 (m, 2′-H). Me₂SO-d₆ solution was measured with tetramethylsilane as the internal standard. Using positive ion ESI LC/MS/MS the parent ion was at m/z 282 (M+1). This molecular ion gave a daughter ion at m/z 166 (base+1).

Synthesis and purification of oligodeoxynucleotides

DNA template, primers and standard markers, listed in Table 1, were purchased from Amersham Corp. These were purified on a reversed phase Wakopak WS-DNA column (0.46 × 25 cm; Wako) eluted over 60 min at a flow rate of 1.0 ml/min with a linear gradient of 50 mM triethylamine acetate (pH 7.0) containing 10–20% acetonitrile. DNA templates and primers were further purified by electrophoresis on a 20% polyacrylamide gel (20 × 40 × 0.035 cm) in the presence of 7 M urea. The oligomers recovered from PAGE were again subjected to HPLC to remove urea. Oligomers were labeled at the 5′-terminus by treatment with T4 polynucleotide kinase in the presence of [γ-³²P]ATP and subjected to electrophoresis to establish homogeneity.

Table 1. Sequences of oligodeoxynucletides^a.

Number	Sequence
1	5′-CATGCTGATGAATTCCTTC-3′
2	5′-CATGCTGATGAATTCCTTCX-3′
3	5′-CTTCTTTCCTCTCCCTTT-3′
4	5′-CATGCTGATGAATTCCTTCXCTTCTTTCCTCTCCCTTT-3′
5	5′-AAAGGGAGAGGAAAGAAGCGAAGGAATTCATCAGCATG-3′
6	5′-AGGAAAGAAGCGAAGGAATTCATC-3′
7	5′-AGAGGAAAGA-3′
8	5′-AGAGGAAAGAAG-3′
9	5′-AGAGGAAAGAAGN-3′
10	5′-AGAGGAAAGAAGNGAAGG-3′
11	5′-AGAGGAAAGAAGGAAGG-3′
12	5′-AGAGGAAAGAAGAAGG-3′
13	5′-AGAGGAAAGAAGNGAAGGAATTCATCAGCATG-3′

^aSequences of template, primers and standard markers. X = N²-Me-dG or dG; N = dC, dA, dG or dT.

The 38mer modified template (5′-CATGCTGATGAATTCCTTCG^MeCTTCTTTCCTCTCCCTTT, G^Me = N²-Me-dG) containing a single N²-Me-dG was prepared by enzymatic incorporation of N²-Me-dGTP using the 19mer oligodeoxynucleotide (5′-CATGCTGATGAATTCCTTC) and the Klenow fragment of E.coli DNA polymerase I following ligation to an 18mer (5′-CTTCTTTCCTCTCCCTTT) with T4 DNA ligase, as shown in Figure 2.

Figure 2.

Diagram of preparation of N²-Me-dG-modified 38mer template.

To synthesize a 20mer (sequence 2 in Table 1, X = N²-Me-dG) containing N²-Me-dG at the 3′-terminus, N²-Me-dGTP (2 µg) was mixed with a 19mer primer (1 µg, sequence 1 in Table 1) fully annealed to a 38mer template (10 µg, sequence 5 in Table 1). The mixture was incubated with Klenow fragment (0.1 U) at 25°C for 45 min in 50 µl of 10 mM Tris–HCl (pH 7.5) containing 7 mM MgCl₂ and 0.1 mM dithiothreitol (DTT). The reaction was stopped by heating the sample to 95°C for 5 min. Samples were subjected to electrophoresis on a 20% polyacrylamide gel containing 7 M urea (20 × 40 × 0.035 cm). The N²-Me-dG-modified 20mer recovered from PAGE was again subjected to HPLC to remove urea. Four micrograms of 18mer (sequence 3 in Table 1) was phosphorylated at the 5′-terminus for 40 min at 37°C using T4 kinase (35 U) and 2 mM ATP in 50 µl of 50 mM Tris–HCl (pH 8.0) containing 10 mM MgCl₂ and 5 mM DTT and subsequently ligated to N²-Me-dG-modified 20mer (2 µg, sequence 2 in Table 1) at 8°C overnight using T4 DNA ligase (1050 U), 0.3 mM ATP and a 24mer template (6 µg, sequence 6 in Table 1) in 50 µl of 66 mM Tris–HCl (pH 7.6) containing 6.6 mM MgCl₂ and 10 mM DTT. The resultant N²-Me-dG 38mer was also isolated by HPLC and its length was confirmed by PAGE.

Enzymatic digestion

A 20mer oligodeoxynucleotide containing N²-Me-dG at the 3′-terminus (sequence 2 in Table 1, X = N²-Me-dG) was digested to deoxynucleoside 3′-monophosphates at 37°C for 3 h in 20 µl of 17 mM sodium succinate (pH 6.0) containing 8 mM CaCl₂ using micrococcal nuclease (3 U) and spleen phoshodiesterase (0.1 U). The reaction mixture was further incubated at 37°C for 2 h with alkaline phosphatase (3 U) in 120 µl of 8 mM Tris–HCl (pH 8.5) containing 0.8 mM zinc sulfate. The mixture was evaporated to dryness and the resultant deoxynucleosides were extracted twice each with 300 µl of methanol. Methanol extracts were evaporated to dryness, dissolved in distilled water and analyzed by HPLC, using a Mightysil RP-18 GP Aqua column (0.46 × 25 cm; Kanto, Japan). Elution was carried out using a linear gradient of 50 mM sodium phosphate buffer (pH 6.0) containing 0–40% acetonitrile over 30 min at a flow rate of 1.0 ml/min.

Primer extension reactions

The primer (5′-AGAGGAAAGA) was labeled at the 5′-terminus by treatment with T4 polynucleotide kinase in the presence of [γ-³²P]ATP. Using a ³²P-labeled 10mer (0.5 pmol) primed with a 38mer template (0.75 pmol, 5′-CATGCTGATGAATTCCTTCXCTTCTTTCCTCTCCCTTT, X = dG or N²-Me-dG), primer extension reactions catalyzed by Klenow fragment (exo⁺) were carried out at 20°C for 1 h in 10 µl of buffer containing all four dNTPs (100 µM each). The reaction buffer consisted of 50 mM Tris–HCl (pH 8.0), 8 mM MgCl₂ and 5 mM 2-mercaptoethanol. Reactions were stopped by adding formamide dye and heating to 95°C for 3 min. Samples were subjected to electrophoresis on a 20% polyacrylamide gel containing 7 M urea (20 × 40 × 0.035 cm). Bands were identified by autoradiography and the radioactivities were measured with a FLA-2000 instrument (Fuji Photo Film Corp., Japan).

Quantitation of miscoding specificity

Reaction of a 38mer template (0.75 pmol, sequence 4 in Table 1) and a ³²P-labeled 10mer primer (0.5 pmol, sequence 7 in Table 1) was catalyzed by Klenow fragment (exo⁺) and carried out at 20°C for 1 h in the presence of all four dNTPs. Reactions were stopped by heating at 95°C for 3 min and then subjected to electrophoresis on a 20% polyacrylamide gel containing 7 M urea (20 × 40 × 0.035 cm). Fully extended products were recovered from the gel and then cleaved with EcoRI after annealing with an unmodified 38mer (sequence 4). Samples were incubated with EcoRI (100 U) for 1 h at 30°C and then for 1 h at 15°C to produce complete cleavage, as shown in Figure 3. Electrophoresis was carried out on two-phase 20% polyacrylamide gels (20 × 40 × 0.035 cm) containing 7 M urea in the upper phase and no urea in the lower phase (13). Standards representing products containing dC, dA, dG or dT opposite the lesion (sequence 10) or one and two base deletions (sequences 11 and 12) were completely resolved on the gel, based on their different migrations.

Figure 3.

Diagram of the primer extension methodology and analysis of the reaction products.

Kinetic studies of nucleotide insertion and chain extension

Kinetic parameters associated with nucleotide insertion opposite the lesion and chain extension from the primer 3′-terminus were determined in reactions containing a single dNTP (13–15). Reaction mixtures containing varying amounts (0.004–0.04 U) of Klenow fragment were incubated at 25°C for 1.0–10.0 min in 10 µl of Tris–HCl (pH 8.0) containing 1.0 pmol of 38mer template (sequence 4 in Table 1) primed with 0.5 pmol of ³²P-labeled 12mer (5′-AGAGGAAAGAAG) to measure nucleotide insertion kinetics or with a ³²P-labeled 13mer (5′-AGAGGAAAGAAGN, N = C, A, G or T) to measure chain extension kinetics. Base insertion was measured in reactions using 0.004 U Klenow fragment for 1 min (C:G, C:G^Me and T:G^Me), 0.01 U for 10 min (T:G) or 0.04 U for 10 min (other pairs).

Reaction mixtures were subjected to electrophoresis on 20% polyacrylamide gels (20 × 40 × 0.035 cm) in the presence of 7 M urea. The Michaelis constant (K_m) and maximum rate of reaction (V_max) were obtained from Hanes–Woolf plots of the kinetic data. Frequencies of insertion (F_ins) and extension (F_ext) were determined relative to the dC:dG base pair according to the equation developed by Mendelman et al. (14,15), where F = (V_max/K_m)_{[wrong pair]}/(V_max/K_m)_{[correct pair = dG:dC]} is defined for any base pair containing a single modified nucleoside. All reactions were linear over the course of the experiment. Data reported represent the averages of two to four separate experiments in which <20% of the primer was extended (14).

RESULTS

Preparation of site-specifically modified oligonucleotide containing a single N²-Me-dG

When dGTP was reacted with formalin and reduced by NaBH₃CN, N²-Me-dGTP (t_R = 14.8 min) could be completely isolated from dGTP (t_R = 12.5 min) by HPLC (Fig. 1B). The peak (t_R = 16.3 min) eluting after N²-Me-dGTP has not been identified.

N²-Me-dG-modified 20mer oligodeoxynucleotide (sequence 2 in Table 1, X = N²-Me-dG) was prepared by enzymatic incorporation of N²-Me-dGTP as described in Materials and Methods. The N²-Me-dG-modified 20mer was separable from unmodified 20mer by HPLC, and no contamination of the unmodified 20mer was observed. The N²-Me-dG-modified 20mer and unmodified 20mer (sequence 2 in Table 1, X = dG) were digested enzymatically and subjected to HPLC (Fig. 4A and B). The UV spectrum of the product detected at 19.2 min was identical to that of N²-Me-dG (Fig. 4C). The obtained molar ratio of dC:dG:dT:dA:dG^Me estimated from UV absorbance (5.0:3.5:7.4:4.4:1.0) was almost consistent with the theoretical value (5:3:7:4:1) of the nucleoside composition of the 20mer. Moreover, the N²-Me-dG-modified 20mer was characterized by the use of negative ion ESI LC/MS/MS. As shown in Figure 5, the spectrum of the N²-Me-dG-modified 20mer (mol. wt 6097) exhibited multivalent ion peaks from 6 to 9 electric charges at m/z 1015.17, 870.09, 765.89 and 676.19, respectively (Fig. 5A). The observed molecular weight of the N²-Me-dG-modified 20mer was determined to be 6097.0 Da using deconvolution of the mass spectrum, which is also consistent with the theoretical value (Fig. 5B).

Figure 4.

HPLC elution profiles of enzyme digests of unmodified and N²-Me-dG-modified 20mer. Three micrograms of unmodified (A) or N²-Me-dG-modified (B) 20mer (5′-CATGCTGATGAATTCCTTCX, X = dG or N²-Me-dG) was digested to nucleosides as described in Materials and Methods. Methanol extracts of the digested samples were evaporated to dryness and placed on a Mightysil RP-18 GP Aqua column (0.46 × 25 cm; Kanto). Elution was carried out using a linear gradient of 50 mM sodium phosphate buffer, pH 6.0, containing 0–40% acetonitrile over 40 min at a flow rate of 1.0 ml/min. (C) UV spectrum of N²-Me-dG and dG.

Figure 5.

(A) Mass spectrum of N²-Me-dG-modified 20mer. A gas sheath flow of 60 p.s.i. (483 kPa) nitrogen, an electrospray voltage of –4.5 kV, a tubelens voltage of –114 V and a capillary temperature of 175°C were used. (B) Deconvolution spectrum of the N²-Me-dG-modified 20mer. Using BioWorks software (Finnigan MAT) the molecular weight of N²-Me-dG-modified 20mer was calculated from the mass spectrum data.

The characterized N²-Me-dG-modified 20mer was ligated to an 18mer (sequence 3 in Table 1), resulting in formation of a 38mer containing a single N²-Me-dG (Fig. 6A). The ligation product was subjected to HPLC (Fig. 6B). Fractions 1–3 were collected and subjected to PAGE containing 7 M urea. This revealed that fr-1 (t_R = 28.1 min), fr-2 (t_R = 30.9 min) and fr-3 (t_R = 33.5 min) were 24mer, 18mer and N²-Me-dG-modified 38mer (ligation product), respectively. No peak of N²-Me-dG-modified 20mer was observed at its retention time (t_R = 29.5 min, indicated by an arrow in Fig. 6B), indicating that most of the modified 20mer was ligated to the 18mer, resulting in formation of the N²-Me-dG-modified 38mer oligonucleotide. Since we used very pure oligonucleotides, in terms of both HPLC and PAGE analysis, and we used a ligation method to make the N²-Me-dG-modified 38mer, contamination with undesired sequences was unlikely to occur. Moreover, there was no contamination with unmodified 38mer when checked by HPLC (modified and unmodified 38mer are separable by HPLC; an example of this separation is shown in Fig. 7).

Figure 6.

(A) Diagram of the method of ligation of N²-Me-dG-modified 20mer with 18mer. (B) HPLC elution profile after ligation. fr-1 and fr-2 were 24mer and 18mer, respectively. The N²-Me-dG-modified 20mer (t_R = 29.5 min) was completely consumed by the ligation reaction (no peak at 29.5 min). The desired N²-Me-dG-modified 38mer was fr-3.

Figure 7.

HPLC trace of N²-Me-dG-modified and unmodified oligomers. The two 38mers were separable on a reversed phase Wakopak WS-DNA column (0.46 × 25 cm) using a linear gradient of 50 mM triethylamine acetate, pH 7.0, containing 5–20% acetonitrile over 60 min and a flow rate of 1.0 ml/min.

Primer extension reactions

Primer extension reactions catalyzed by Klenow fragment (exo⁺) were carried out in the presence of four dNTPs using unmodified or N²-Me-dG-modified 38mer templates (sequence 4 in Table 1). As shown in Figure 8, primer extension on an unmodified template rapidly produced a single fully extended product (Fig. 8, lanes 1–4). Using N²-Me-dG-modified template the primer extension reaction was slightly stalled one and two bases before the lesion (Fig. 8, lanes 5–7). A small amount of 12mer (8.2% of the starting primer) and 11mer (9.3%) was observed after 60 min incubation (lane 7), but DNA synthesis past this lesion was readily completed.

Figure 8.

Primer extension reactions catalyzed by Klenow fragment (exo⁺). Using unmodified or N²-Me-dG-modified 38mer template (sequence 4 in Table 1) primed with a 5′-³²P-labeled 10mer (sequence 7 in Table 1) primer extension reactions were conducted at 20°C using 0.004 U Klenow fragment as described in Materials and Methods. One-third of the reaction mixture was subjected to denaturing 20% PAGE. 13× shows the location opposite the N²-Me-dG lesion.

Quantitation of base substitutions

Fully extended reaction products were analyzed by two-phase gel electrophoresis (13) to determine the miscoding specificities of N²-Me-dG. As shown in Figure 9, a standard mixture of ³²P-labeled oligodeoxynucleotides containing dC, dA, dG or dT at the position of the lesion and one or two base deletions served as standard markers for these experiments (Fig. 9, lane 2). DNA synthesis on an unmodified template led to the expected incorporation of dCMP opposite dG at position 13 (Fig. 9, lane 1). The full-length product represents 97.3% of the starting primer. When an N²-Me-dG-modified template was used (Fig. 9, lane 3) dCMP (84.0%) and dTMP (9.4%) were incorporated opposite the N²-Me-dG lesion. Incorporation of dAMP and dGMP was not observed. In addition, no deletions were detected.

Figure 9.

Quantitation of miscoding specificities induced by Klenow fragment (exo⁺). Using a 38mer template (sequence 4 in Table 1) primed with a 5′-³²P-labeled 10mer (sequence 7) primer extension reactions were conducted at 20°C for 1 h using 0.004 U Klenow fragment for unmodified and N²-Me-dG-modified templates. The fully extended products recovered by PAGE were cleaved by EcoRI restriction enzyme at 30°C for 1 h and subsequently at 15°C for 1 h as described in Materials and Methods. The reaction samples were subjected to two-phase 20% PAGE. Mobilities of reaction products were compared with those of 18mer standards (sequences 10–12) containing dC, dA, dG or dT opposite the lesion and one base (Δ¹) or two base (Δ²) deletions (lane 2).

Kinetic studies of nucleotide insertion and extension

Steady-state kinetic parameters were established for nucleotide insertion opposite N²-Me-dG and for chain extension from 3′-termini containing this lesion using Klenow fragment (exo⁺). The frequencies of nucleotide insertion (F_ins) and extension (F_ext) were defined as the ratio V_max/K_m relative to that of the correct base pair (dCTP opposite dG in the template). As shown in Table 2, F_ins for dCTP opposite N²-Me-dG was 0.704 and for dTTP opposite N²-Me-dG was 0.582. F_ins for dTTP opposite N²-Me-dG was 730 times higher than F_ins for dTTP opposite unmodified dG. The frequency of chain extension (F_ext) from 3′-termini containing a dC:N²-Me-dG pair was 0.795 and from those containing a dT:N²-Me-dG pair was 0.381. F_ext from 3′-termini containing a dT:N²-Me-dG pair was 9.1 times higher than that from a dT:dG pair. The relative frequency of translesional synthesis (F_ins × F_ext) past N²-Me-dG in reactions involving dT was 0.222.

Table 2. Kinetic parameters for the nucleotide insertion and chain extension reactions catalyzed by Klenow fragment^a.

N:X	Insertion			Extension			Fins×Fext
	dNTP			dGTP
	↓GAAGAAAGGAGA32P			↓NGAAGAAAGGAGA32P
	5′-CCTTCXCTT CTT TCC TCT CCCTTT			5′-CCTTCXCTT CTT TCC TCT CCCTTT
	Km (µM)	Vmax (% min–1)	Fins	Km (µM)	Vmax (% min–1)	Fext
C:dG	1.76 ± 0.7b	49.3 ± 10	1.0	5.31 ± 0.09	106 ± 5.4	1.0	1.0
C:N2-Me-dG	2.10 ± 0.3	41.3 ± 1.2	0.704	6.17 ± 0.7	98.1 ± 2.4	0.795	0.56
T:dG	51.3 ± 10.5	1.17 ± 0.3	7.96 × 10–4	18.5 ± 3.9	15.6 ± 2.1	4.20 × 10–2	3.34 × 10–5
T:N2-Me-dG	2.40 ± 0.4	39.1 ± 0.7	0.582	10.8 ± 2.7	82.6 ± 2.7	0.381	0.222
A:dGc	89.0 ± 37.3	0.82 ± 0.18	1.31 × 10–4	21.5 ± 9.6	0.10 ± 0.01	9.16 × 10–5	1.20 × 10–8
A:N2-Me-dG	71.9 ± 17.3	0.49 ± 0.20	2.44 × 10–4	96.7 ± 8.6	0.15 ± 0.08	7.85 × 10–5	1.92 × 10–8
G:dGb	25.1 ± 1.7	3.10 ± 0.42	1.53 × 10–3	8.64 ± 2.78	0.74 ± 0.02	1.42 × 10–3	2.17 × 10–6
G:N2-Me-dG	26.1 ± 1.4	2.68 ± 0.85	3.68 × 10–3	40.8±9.8	2.16 ± 0.17	2.65 × 10–3	9.75 × 10–6

^aKinetics of nucleotide insertion and chain extension reactions were determined as described in Materials and Methods. Frequencies of nucleotide insertion (F_ins) and chain extension (F_ext) were estimated by the equation: F = (V_max/K_m)_{[wrong pair]}/(V_max/K_m)_{[correct pair]}. X = dG or N²-Me-dG lesion.

^bData are expressed as means ± SD.

^cData are taken from Shibutani et al. (16).

DISCUSSION

The site-specifically modified 38mer template containing a single N²-Me-dG can be prepared by the enzymatic incorporation of N²-Me-dGTP into the 3′-terminus of a 19mer oligomer following enzymatic ligation to another 18mer oligomer. Our preliminary experiments indicate that N²-Me-dGTP was well utilized as a substrate during DNA synthesis catalyzed by Klenow fragment. The insertion of N²-Me-dGTP opposite template dC with this enzyme was almost equal to that of dGTP. In addition, most of the N²-Me-dG-modified 20mer was ligated to an 18mer, resulting in formation of a 38mer containing a single N²-Me-dG. This shows that T4 DNA ligase was able to efficiently execute its function even at the N²-Me-dG site. This is probably because the N²-Me-dG:dC pair does not significantly change the conformation of the DNA duplex, although the guanine N2 position is responsible for base pairing via a hydrogen bond.

The miscoding properties of N²-Me-dG observed during translesional synthesis in reactions catalyzed by Klenow fragment were analyzed by two-phase gel electrophoresis. Preferential targeted incorporation of dCMP, the correct base, was observed. In addition, 9.4% misincorporation of dTMP was detected opposite the N²-Me-dG lesion, generating G→A transitions.

The miscoding specificities of methylated DNA adducts such as 8-methyl-2′-deoxyguanosine (8-Me-dG) (17) and O⁶-methyl-2′-deoxuguanosine (O⁶-Me-dG) (13) have been determined using the same in vitro experimental system. 8-Me-dG is a weak miscoding lesion, generating G→C (1.1%) or G→T (0.41%) transitions or deletions (1.2%). In contrast, O⁶-Me-dG is a strong miscoding lesion, generating primarily G→A transitions (75.5%). The position of methylation on dG adducts significantly influences the miscoding specificity and frequency.

Analysis of full-length reaction products should reflect relative rates of nucleotide insertion opposite the lesion. This assumption was confirmed by steady-state kinetic analyses, in which the frequency of nucleotide insertion opposite N²-Me-dG and chain extension from the primer 3′-terminus were measured in the presence of a single dNTP. With Klenow fragment F_ins for dTTP was only 1.2 times lower than for dCTP and F_ext from a dT:N²-Me-dG terminus was only 2.1 times lower than from dC:N²-Me-dG. The relative translesional synthesis (F_ins × F_ext) for the dT:N²-Me-dG pair is only 2.5-fold lower than that for the dC:N²-Me-dG pair.

It is instructive to compare the miscoding properties of N²-Me-dG with those of another endogenously generated DNA lesion, 7,8-dihydro-8-oxo-2′-deoxyguanosine (8-oxo-dG) (18). As shown in Table 3, F_ins × F_ext for dCTP and dATP opposite 8-oxo-dG are 0.0116 and 0.025, respectively. F_ins × F_ext for dCTP and dTTP opposite N²-Me-dG are 0.56 and 0.222, respectively, which is much higher than that of 8-oxo-dG. This shows that Klenow fragment bypasses the N²-Me-dG lesion more easily than 8-oxo-dG. In addition, F_ins × F_ext for dTTP opposite N²-Me-dG is ∼9 times higher than that for dATP opposite 8-oxo-dG. Thus, N²-Me-dG is considered a more mutagenic lesion than 8-oxo-dG.

Table 3. Comparison of translesional synthesis of N²-Me-dG with 8-oxo-dG.

N:X	Fins	Fext	Fins × Fext
C:N2-Me-dG	0.704	0.795	0.56
T:N2-Me-dG	0.582	0.381	0.222
C:8-oxodGa	0.35	0.033	0.0116
A:8-oxodGa	0.05	0.5	0.025

^aData are taken from Shibutani et al. (18)

In a preliminary experiment utilization of N²-Me-dGTP opposite dN (dC, dA, dG or dT) embedded in templates during DNA synthesis catalyzed by Klenow fragment (exo⁺) was analyzed. Incorporation of N²-Me-dGTP opposite template dC and dT was observed (unpublished data), consistent with this study. Thus, N²-Me-dG tends to couple with dT. In chain extension from dT:dG, a mismatch base pair, most of the dT at the 3′-terminus was excised by Klenow fragment, but not from dT:N²-Me-dG. dT excision from dT:N²-Me-dG was markedly less than that from dT:dG. These results indicate that addition of a -CH₃ moiety to the N2 position of dG promotes misincorporation of dTMP, resulting in dT:N²-Me-dG base pairing, and evasion of the 3′→5′ exonuclease function of Klenow fragment in chain extension from dT:N²-Me-dG.

Rife et al. showed that N²-methylguanosine can form base pairs with C, U and A in RNA molecules in vitro (19). However N²-Me-dG paired with dC and dT but not with dA in this study. The incorporation rate (F_ins × F_ext) of dATP opposite N²-Me-dG was very low (Table 2). Thus, although thermodynamically N²-methylguanine can pair with adenine, the enzymatic reaction data show that N²-Me-dG does not pair with dA during DNA synthesis.

In this study, using a N²-Me-dG-modified template, the miscoding specificities of N²-Me-dG were investigated using an in vitro experimental system that can quantify base substitutions and steady-state kinetics. The results suggest that the formaldehyde-derived DNA adduct N²-Me-dG is a miscoding lesion capable of generating G→A transition mutations.