Genotoxic effects of the major alkylation damage N7-methylguanine and methyl formamidopyrimidine

Abstract

Various alkylating agents are known to preferentially modify guanine in DNA, resulting in the formation of N7-alkylguanine (N7-alkylG) and the imidazole-ring opened alkyl-formamidopyrimidine (alkyl-FapyG) lesions. Evaluating the mutagenic effects of N7-alkylG has been challenging due to the instability of the positively charged N7-alkylG. To address this issue, we developed a 2’-fluorine-mediated transition-state destabilization approach, which stabilizes N7-alkylG and prevents spontaneous depurination. We also developed a postsynthetic conversion of 2’-F-N7-alkylG DNA into 2’-F-alkyl-FapyG DNA. Using these methods, we incorporated site-specific N7-methylG and methyl-FapyG into pSP189 plasmid and determined their mutagenic properties in bacterial cells using the supF-based colony screening assay. The mutation frequency of N7-methylG was found to be less than 0.5%. Our crystal structure analysis revealed that N7-methylation did not significantly alter base pairing properties, as evidenced by a correct base pairing between 2’-F-N7-methylG and dCTP in Dpo4 polymerase catalytic site. In contrast, the mutation frequency of methyl-FapyG was 6.3%, highlighting the mutagenic nature of this secondary lesion. Interestingly, all mutations arising from methyl-FapyG in the 5’-GGT(methyl-FapyG)G-3’ context were single nucleotide deletions at the 5’-G of the lesion. Overall, our results demonstrate that 2’-fluorination technology is a useful tool for studying the chemically labile N7-alkylG and alkyl-FapyG lesions.

INTRODUCTION

Guanine N7 is the most nucleophilic atom within DNA, thereby reacting with a wide range of endogenous and exogenous alkylating agents (e.g., S-adenosylmethionine, temozolomide, nitrogen mustards) to generate N7-alkylguanine (N7-alkylG) as major adducts. N7-AlkylG lesions have a formal positive charge on guanine N7, which makes the lesions susceptible to further modifications such as the hydrolytic cleavage of glycosidic bond and the imidazole ring opening (Figure 1). The glycosidic bond scission gives rise to the highly genotoxic abasic sites, which promotes G-to-T transversions (Figure 1A) (1,2). The hydroxide-mediated opening of the imidazole ring in N7-alkylG lesions produces the conformationally flexible alkyl-formamidopyrimidine (alkyl-FapyG), which has been shown to be mutagenic in vitro and in vivo studies (3).

Figure 1. Processing of N7-alkylguanine and 2’-F-N7-alkylguanine.

(A) Further modification of N7-alkylguaine (N7-alkylG) adducts. The spontaneous depurination of N7-alkylG generates abasic sites (AP sites), while the imidazole ring opening of N7-alkylG produces alkyl-FapyG lesions. (B) Inhibition of the glycosidic bond cleavage by 2’-F-mediated destabilization of the transition state of spontaneous depurination. (C) Overall synthesis of 2’-F-N7-MeG-containing DNA from a commercially available ribose derivative. The 2’-F-N7-MeG phosphoramidite was incorporated into oligonucleotide via the solid phase DNA synthesis.

N7-AlkylG adducts in duplex DNA display half-lives of several hours to days, being capable of influencing replication and transcription (4). The mutagenic properties of the primary lesion N7-alkylG damage remain poorly understood except for few lesions such as aflatoxin B1-N7-G (5,6), where single G approaches have been used in the preparation of AFB1-N7-G and small N7-alkylG adducts (e.g., N7-MeG). This knowledge gap had been mainly due to the difficulty in preparing DNA containing a site-specific incorporated N7-alkylG via the solid-phase DNA synthesis: the positively charged N7-alkylG mononucleoside rapidly undergoes depurination. We have recently reported the synthesis, kinetic and structural studies of various N7-alkylG-containing DNA (7-12), where a 2’-fluorination approach is used to stabilize otherwise labile N7-alkylG lesions (Figure 1B and 1C). The introduction of fluorine at 2’ position of nucleoside enhances the chemical stability of the glycosidic bond by destabilizing the transition state of spontaneous depurination, thereby deterring the degradation of N7-alkylG adducts (11,13).

While alkyl-FapyG is less frequently produced than N7-alkylG, this secondary lesion is persistent in cells and is often more mutagenic than the primary N7-alkylG (3,14,15). Numerous alkyl-FapyG lesions are induced by alkylating agents including aflatoxin B1, N-methyl nitrosamines, cyclophosphamide, mitomycin C, among others (16-18). The imidazole ring-opened alkyl-FapyG lesions, also known as N5-substitued formamidopyrimidine, can take on cis/trans configuration, syn/anti base conformations, and α/β anomers (Figure 2A). Oligonucleotides containing alkyl-FapyG adducts (e.g., alkyl = aflatoxin B1) have been prepared via direct alkylation of a single dG-containing DNA followed by imidazole ring opening (19). Also, alkyl-FapyG phosphoramidites and carbocyclic nucleoside analogs of alkyl-FapyG have been synthesized and site-specifically incorporated into DNA via solid phase synthesis (20-23). Biochemical and/or structural studies of alkyl-FapyG lesions (alkyl = methyl, nitrogen half-mustard, aflatoxin B1) have provided insights into the mutagenic properties and mutagenesis mechanisms of these lesions (6,7,16,18,19,24).

Figure 2. Isomerization of alkyl-FapyG adducts.

(A) Equilibrium between β-alkyl-FapyG and α-alkyl-FapyG in DNA. (B) Production of α/β anomeric mixtures of methyl-FapyG via a postsynthetic modification of 2’-F-N7-MeG-containing DNA.

Herein, we report the application of the 2’-fluorination technology in studying the mutagenic properties of N7-MeG and Me-FapyG in bacterial cells. We have developed a postsynthetic DNA modification approach, where 2’-F-N7-alkylG-containing DNA is directly converted into 2’-F-alkyl-FapyG (Figure 2B). We have prepared pSP189 shuttle vector plasmids containing 2’-F-N7-MeG or 2’-F-Me-FapyG, replicated the plasmids in bacterial cells, and analyzed the mutation spectrum and frequency using the supF-based blue/white colony screening assay. Our studies revealed the mutagenic potential of N7-MeG and Me-FapyG in the absence of the lesion repair. Furthermore, we determined a crystal structure of the bacterial translesion synthesis (TLS) DNA polymerase Dpo4 incorporating dCTP opposite templating 2’-F-N7-MeG, which provided insights into the error-free bypass of the lesion by the enzyme. Overall, our developed chemical biology approaches provided important insights into the mutagenic properties of the major alkylative damage N7-alkylG and alkyl-FapyG.

Results and Discussion

N7-MeG is non-mutagenic in bacterial cells.

To systematically investigate the mutagenicity of N7-alkylG in bacteria, we began by site-specifically incorporating the smallest alkyl lesion N7-MeG, into pSP189 at the 6^th position of the supF gene. Successful ligation was confirmed using agarose gel electrophoresis and comparison to an insert-free negative control. Supercoiled plasmid was present only in the experimental reaction, confirming that the linearized vector did not self-ligate in the absence of an insert. Once the supercoiled plasmid was visually confirmed, it was gel-extracted and purified for transformation into the E. coli reporter strain MBM7070. The transformed bacteria were cultured on agar plates containing ampicillin, IPTG, and X-gal. Additionally, MBM7070 transformed with unmodified pSP189 was cultured, serving both as a positive ligation and negative mutagenesis control. Of more than >1,000 controls, only blue colonies were present. The use of psP189-N7MeG plasmid resulted in the exclusive formation of blue colonies (~200-300 colonies, three independent experiments) in MBM7070 cells, suggesting the miscoding potential of the smallest alkylG lesion in prokaryotes is low. To rule out any false negatives, we sequenced a random selection of colonies from each group. Sequencing indicated that there were no issues with ligation and the mutation frequency of N7-MeG is below 0.5%.

The observed low mutation frequency for 2’-F-N7-MeG is in consistent with the published reports (25,26) that describe non-mutagenic nature of N7-MeG in cells. While N7-MeG is the predominant methylated lesion, it does not significantly block replication (26). Also, N7-MeG is readily excised by the alkyladenine DNA glycosylase AlkA in E. coli and AAG in humans (27,28), which counteracts the promutagenicity of N7-MeG. As the fluorinated N7-MeG is not removed by E. coli AlkA (11), our studies indicate that the bypass of N7-MeG by bacterial replicative and/or translesion synthesis (TLS) DNA polymerases are error-free. This contrasts with recent reports showing that the bypass of N7-MeG by eucaryotic TLS pols is error-prone (8,29). The presence of templating N7-MeG dramatically slows replication by eukaryotic DNA pols such as polβ (10). It would be of interest to see whether 2’-F-N7-MeG is promutagenic in eukaryotic cells.

2’-F-N7-MeG forms Watson-Crick base pairing with dCTP in Dpo4 catalytic site.

An archaeal translesion synthesis (TLS) DNA polymerase Dpo4 belongs to the DinB subclass of Y family DNA polymerase that has been shown to bypass a wide variety of bulky DNA adducts, alongside with Dbh from Sulfolobus acidocaldarius, E. coli pol IV, and human polκ (hpolκ) (30,31). DinB polymerases possess a relatively spacious catalytic site, with tendency to make deletion mutations, and ability to bypass bulky DNA lesions. Dpo4 has been extensively used as a model enzyme for Y-family TLS polymerases. Dpo4 is known to efficiently replicate past N7-MeG (29). Dpo4 has been shown to bypass N7-MeG and other bulky dG lesions with relative efficiency similar to those for bypass of undamaged dG (6,32), suggesting the catalytic site of Dpo4 promotes an error-free bypass of damaged guanines.

To gain structural insights into the efficient and accurate replication of Dpo4 over N7-MeG, we solved a crystal structure of Dpo4 complexed with templating 2’-F-N7-MeG paired with incoming dCTP in the presence of Ca²⁺ ions (Figure 3). The Ca²⁺ has been widely used as a non-reactive substitute for Mg²⁺ in crystallographic studies of DNA polymerases, because Ca²⁺ allows the capture of a ternary complex for a catalytically active polymerase bound to replicating base pair with the same coordination mode as Mg²⁺ ions while it inhibits the nucleotidyl-transfer reactions (33,34). The structure of the Dpo4-N7MeG:dCTP ternary complexes was refined to 2.3 Å. This structure belonged to P2₁2₁2 space group with a single complex in the asymmetric unit and displayed the general structural features of ternary complexes of Y-family DNA polymerases, including the thumb, palm, finger, and little finger domains with a right-handed conformation (35).

Figure 3. Overall conformation of ternary Dpo4-N7MeG:dCTP complex structure.

(A) The ternary complex structure of Dpo4 captured with templating 2’-F-N7-MeG and incoming dCTP in the presence of Ca²⁺ ions. The protein is shown in cartoon with DNA duplex in stick mode. The Dpo4 domains were colored in orange (palm), green (thumb), cyan (finger), and yellow (little finger), respectively. Primer strand is colored in white and template strand in gray. The active-site Ca²⁺ ions are shown in green spheres. The incoming dGTP is shown in green. (B) Superposition of the Dpo4- N7MeG:dCTP and the Dpo4-dG:dCTP (colored in sand) structures.

The overall structure of the Dpo4-N7MeG:dCTP ternary complex was essentially identical to that of the Dpo4-dG:dCTP insertion complex, with RMSD of 0.243 Å when Cα atoms are aligned (Figure 3B). The N7MeG:dCTP base pair was positioned between the finger and little finger domains, coordinated by the palm and finger domains. The finger domain lies over the replicating N7MeG:dCTP base pairing in the active site of the ternary complex. The statistics for data collection and the refinement of the structural models were summarized in Table 1 (36).

Table 1.

Data collection and refinement statement statistics.

PDB CODE	Dpo4- N7MeG:dCTP (7KLE)
Data Collection
space group	P21212
Cell Constants
a (Å)	99.650
b	102.472
c	53.034
α (°)	90.00
β	90.00
γ	90.00
resolution (Å)a	20-3.00 (3.05-3.00)
Rmerge b (%)	0.113 (0.876)
<I/σ>	12.2 (1.42)
completeness (%)	100.0 (100.0)
redundancy	7.0 (7.2)
Refinement
Rwork c/Rfree d (%)	22.7/28.6
unique reflections	11345
Mean B Factor (Å2)
protein	49.94
ligand	44.57
solvent	44.60
Ramachandran Plot
most favored (%)	95.3
add. allowed (%)	3.8
RMSD
bond lengths (Å)	0.003
bond angles (degree)	0.44

^aValues in parentheses are for the highest resolution shell

^bR_merge = Σ∣I-<I>∣/ ΣI where I is the integrated intensity of a given reflection

^cR_work = Σ∣F(obs)-F(calc)∣/ΣF(obs)

^dR_free = Σ∣F(obs)-F(calc)∣/ΣF(obs), calculated using 5% of the data

Dpo4 accommodated N7MeG:dCTP base pair at the replicating base pair site without causing significant structural distortion on DNA and protein conformations (Figure 4). The crystal structure of the Dpo4-N7MeG:dCTP ternary complex disclosed the detailed base pairing mode of templating N7MeG to incoming dCTP at the nascent base pair site of Dpo4 (Figure 4). The N7 methyl moiety and 2′-β-fluorine atom of the templating N7MeG, and incoming dCTP were clearly observed in the 2F_o-F_c electron density map contoured at 1σ. Nearly all residues interacting with replicating N7MeG:dCTP from the palm and finger domain did not undergo significant conformational change relative to those observed in the corresponding undamaged structure. At the position of the insertion, the templating N7MeG formed a canonical Watson-Crick base pairing with incoming dCTP, as similarly observed in published Dpo4 structures with a correct base pair (2AGQ and 2ATL) (37,38). Furthermore, N7MeG base is also stabilized by the base stacking interactions with the adjacent 3′ base.

Figure 4. The catalytic site of Dpo4-N7MeG:dCTP ternary structures.

Close-up view for the replicating base pair of N7MeG:dCTP in the active site of Dpo4. The residues near the template base (S34, A42, R331 and R332) and those close to the incoming nucleotide (A44, T45, Y48, R51 and K159) are highlighted and labeled. A 2F_o-F_c electron density map contoured at 1σ around templating N7MeG and incoming dCTP in the Dpo4-N7MeG:dCTP ternary complex is shown. The active site Ca²⁺ ions are shown in green spheres.

The base-pairing properties of N7MeG:dCTP in Dpo4 catalytic site were very similar to those of correct base pairs (Figure 5). In the nascent base pair site of Dpo4, the incoming dCTP and template N7MeG formed three hydrogen bonds with a Watson-Crick geometry (Figure 5A). Specifically, the O6, N1, and N2 of N7MeG are hydrogen bonded to N4, N3, and O2 of dCTP, respectively, with distances of 2.9, 2.8, and 2.6 Å. The C1′-C1′ distance for N7MeG:dCTP was 10.6 Å and the λ angles for N7MeG and dCTP were 55.0° and 53.6°, respectively, which were essentially the same as observed in canonical Watson-Crick base pairing. While the N7MeG and dCTP base pairing was slightly sheared, buckled and propeller-twisted with respect to each other, the N7MeG did not significantly disturb base pair conformation (39). The formation of Watson-Crick N7MeG:dCTP base pair indicates that N7MeG preferentially uses a keto tautomer in pairing with dCTP (9). Taken together, N7MeG:dCTP base pair was virtually same with dG:dCTP base pair, indicating that the cationic N7-MeG did not significantly modify the base-paring characteristics for correct nucleotide incorporation.

Figure 5.

Base pairing properties of 2’-N7-MeG and dCTP at the TLS site of Dpo4. (A) Watson-Crick base pairing of N7MeG:dCTP. The three interbase hydrogen bonds are presented with dotted lines and the λ angles the C1′-C1′ distance for N7MeG:dCTP are shown. (B) Side view of N7MeG:dCTP with a coplanar base pair conformation.

2′-F-Me-FapyG-containing DNA is processed by Endo IV.

Endo IV is known to cleave the 5’-phosphodiester bond of the unnatural α anomeric nucleotides, but not that of the natural β anomers (21,40,41), and thus has been used for assessing the α/β anomeric ratio of various Fapy lesions (21,42). To evaluate whether the 2’-F postsynthetic approach prevents β-to-α anomerization, we treated 2’-F-Me-FapyG-containing duplex with Endo IV. Urea PAGE gel analysis of the Endo IV reaction revealed a fast-moving band (Figure 6), which suggested that α-Me-FapyG anomer was produced and processed by the enzyme. The incision product yield was about 24%, which was comparable to the incision yield for nonfluorinated Me-FapyG-containing duplex (42). This result indicates that the introduction of 2’-F in alkyl-FapyG does not preclude the interconversion between the α and β anomeric forms.

Figure 6. Urea PAGE analysis of the incision of Me-FapyG-containing duplex with Endo IV.

The formation of the fast-moving band on the gel suggests that Endo IV-mediated cleavage of the phosphodiester bond on the 5’-side of α-Me-FapyG has occurred.

2’-F-Me-FapyG is mutagenic in bacterial cells and induces deletion mutations in the 5’-GGTXGG-3’ context.

The transformation of 2’-F-Me-FapyG-containing pSP189 plasmid into MBM7070 indicator cells resulted in the formation of a mixture of blue, white, and light-blue colonies, which contrasted with the outcome of 2’-F-N7-MeG-containing pSP189 plasmid (Figure 6). Sequencing of all white and light-blue colonies and a random sample of blue colonies produced in three independent experiments showed that the mutation frequency for 2’-F-Me-FapyG was 6.3 ± 1.6%, indicating that 2’-F-Me-FapyG is significantly more mutagenic than N7-MeG in bacterial cells.

Interestingly, all the mutation observed in our 2’-F-Me-FapyG study was a single nucleotide deletion occurring at a 5’-G in the 5’-GGTXGG-3’ context. Published mutagenesis studies involving the use of mixture of α and β anomers of Me-FapyG in primate cells have shown that the anomeric mixtures induce G-to-T transversions (5.2%) and deletion mutations (6.1%) in the 5’-AGTXGG-3’ context (X = Me-FapyG) (43). The similar distribution of mutations is also observed in the 5’-AGGXGG-3’ context (43). The difference in mutational spectrum in bacterial and primate cells could be attributed to difference in arrays and fidelity of DNA polymerases involved in the bypass of Me-FapyG lesions.

Me-FapyG is known to be excised by various DNA glycosylases including bacterial Fpg (44), yeast 8-oxoguanine DNA glycosylase (45), and hNEIL1 (46) in vitro, which would counteract the mutagenic effect of the lesion. As the fluorinated Me-FapyG lesion would not be processed by E. coli Fpg, we expect the mutagenic property of the natural Me-FapyG lesion will be lesser than that of 2’-F-Me-FapyG.

In summary, we have successfully prepared 2’-F-N7MeG- and 2’-F-Me-FapyG-containing DNA, incorporated them into pSP189 plasmid, and performed the supF-based mutation assay. The 2’-F-mediated transition-state destabilization approach has been used to prevent the spontaneous depurination of the cationic N7-MeG. Our results reveal that 2’-F-N7MeG is nonmutagenic in bacterial cells, while 2’-F-Me-FapyG is significantly mutagenic. The transformation of 2’-F-Me-FapyG-containing pSP189 plasmid into MBM7070 cells results in single nucleotide deletion only. A crystal structure of Dpo4-DNA complex shows a canonical Watson-Crick base pairing between templating 2’-F-N7MeG and incoming dCTP in the polymerase active site, indicating that the cationic N7MeG uses its keto tautomer when paired with dCTP. Overall, these studies highlight that the 2’-F approach is a powerful means for studying chemically labile N7-alkylG lesions.

Materials and Methods

Synthesis of N7-MeG- and N7-Me-FapyG-containing DNA.

2’-F-N7-MeG phosphoramidite was synthesized using methods described previously (11). Briefly, a ribose derivative was converted into N2-Pac-protected 2’-F-dG nucleoside, which was then reacted with methyl iodide to produce the corresponding 2’-F-N7-Me-dG (Figure 1C). The sequential tritylation and phosphatidylation of 5’-OH and 3’-OH of 2’-F-N7-Me-dG produced 2’-F-N7-MeG phosphoramidite. The fluorinated N7-MeG phosphoramidite was then site-specifically incorporated into oligonucleotide (5’-pTCGAGCTGTGGTXGGGTTC-3’, where X denotes 2’-F-N7-MeG or 2’-F-N7-Me-FapyG) via solid-phase DNA synthesis. The ultramild deprotection conditions (50 mM K₂CO₃ in MeOH, 25 °C 16h) were used in the preparation of 2’-F-N7MeG DNA to avoid the imidazole ring opening. The resulting 2’-F-N7-MeG-containing DNA was treated with 0.1 N NaOH for 3h at an ambient temperature to produce 2’-F-N7-Me-FapyG DNA. Both 2’-F-N7-MeG- and 2’-F-Me-FapyG-containing DNA were PAGE-purified and their identity was confirmed by the MALDI-TOF mass spectrometry.

Oligonucleotide sequences used for pSP189.

A 19-mer template sequence, 5’-pTCGAGCTGTGGTXGGGTTC-3’, where X denotes N7-Me-Fapy-dG), and its complementary sequence, 5’-pTCGGGAACCCCACCACAGC-3’ were used. The 5’-phosphorylated two oligonucleotides were PAGE-purified and annealed for 5 minutes at 90 °C in 10 mM phosphate buffer (pH 7) and 100 mM NaCl, then slowly cooled to room temperature. Due to the sticky end overhangs on both the 5' and 3' ends of the annealed dsDNA, no digestion prior to ligation is required.

Construction of 2’-F-N7-MeG- and 2’-F-Me-FapyG-containing pSP189 plasmids.

The supF-based forward mutation assay has been used to determine the mutational specificity of various agents including aflatoxin B1 and 8-oxoguanine (47,48). The pSP189 plasmid containing supF gene, which encodes for a suppressor tRNA, was used as a vector for the modified dsDNA insert (49). Within supF there are two restriction sites for the enzyme, AvaI. pSP189 was digested with AvaI at 37°C overnight and the digested solution was visualized on a 1% agarose gel. Linearized plasmid was confirmed and purified with a Qiagen PCR Purification Kit. Mixtures containing 0.04 pmol of digested pSP189 and 0.12 pmol of modified DNA insert were incubated with 1 μL DNA ligase at 16°C overnight to achieve ligation. Negative controls that did not contain any insert in the reaction mixture were also included. Both ligation mixtures were run on a 0.8% low melt agarose gel, and the absence of self-ligation was verified via negative control. Supercoiled DNA bands were extracted and purified using the Qiagen Gel Extraction kit.

Transformation of the lesion-containing pSP189 plasmid into MBM7070.

Blue-white colony screening was performed using the E. coli reporter strain MBM7070 (49,50). The cells were electroporated with pSP189 containing the site-specific 2’-F-N7-MeG- or 2’-F-Me-Fapy-dG lesion and recovered at 37°C. After plating the cells on LB agar supplemented with 50 mg ampicillin, 100 mM IPTG, and 40 g/mL X-gal, they were incubated overnight at 37°C.

Assessing the mutation frequency and spectra of 2’-F-N7-MeG and 2’-F-Me-FapyG.

Colonies were counted automatically using a lab-developed algorithm and validated manually. The mutation frequency was calculated as (white colonies/total colonies) x100 based on three independent experiments and expressed as the mean ± standard error of the mean. Each white colony was subjected to sequencing to determine the mutational spectra. False positives were excluded from counting (UT DNA Sequencing Facility, Austin, TX). Colony sequencing results were aligned to the insert-containing plasmid using the Benchling alignment software (Benchling, San Francisco, CA). For each plate, a random sample of blue colonies was also sequenced to check for false negatives.

Glycosylase activity assay for 2’-F-Me-FapyG-containing DNA.

A 16-mer oligonucleotide (5’-CCGACXTCGCATCAGC-3’, X = N7-Me-FapyG or 7,8-dihydro-8-oxoguanine (oxoG)) was prepared by the solid phase DNA synthesis. The complementary oligonucleotide 5’-GCTGATGCGACGTCGG −3’ was obtained from Integrated DNA Technologies (Coralville, IA). The oligonucleotides were PAGE-purified and annealed for 5 minutes at 90 °C in 10 mM phosphate buffer (pH 7) and 100 mM NaCl, then slowly cooled to room temperature. N7-Me-FapyG-containing 16-mer duplexes and formamidopyrimidine DNA glycosylase (Fpg) were mixed in a reaction buffer (10 mM Tris-HCl, pH 8.0, 1mM EDTA, 0.1% BSA) and incubated for 60 minutes at 37° C. An oxoG-containing 16-mer was used as a positive control and was run under the equivalent conditions. Reactions were stopped at designated timepoints by adding 2 μL 0.5 N NaOH and boiled for 10 minutes. Also, 12 μL of loading buffer (98% formamide, 1mM EDTA, 1mg/mL Bromophenol Blue) were added and samples were boiled for an additional 10 minutes. Samples were immediately cooled and loaded onto a 10 cm x 10 cm 20% denaturing urea gel. Completed gels were stained with SYBR Gold for 45 minutes and imaged via Typhoon FLA9500 (GE Healthcare).

Cloning and Dop4 expression and purification.

Dpo4 (residues 1–341) was amplified from Sulfolobus solfataricus P2 genomic DNA (ATCC 35092) by using the forward and reverse primers in the PCR, then cloned into an expression vector of into pET28a plasmid with NdeI and BamHI restriction enzyme sites (30). The plasmid was transformed into E. coli BL21(DE3)pLysS cell, which were grown at 37 °C in LB medium supplemented with 10 μg ml⁻¹ Kanamycin until the OD600 of 0.8. Protein expression was induced for overnight at 20°C by addition of 0.1mM IPTG. The induced cells were collected by centrifugation at 8,000g for 20 min at 4 °C and stored at −80 °C until use. As described previously, the collected cell was suspended in buffer NA (20 mM sodium phosphate pH 7.5, 300 mM NaCl, 10% glycerol, 1mM PMSF and 50mM imidazole) and lysed by sonication for 2 min and then heated at 75°C for 10 min. The lysate was centrifuge at maximum speed (20,000xg) for 30 minutes at 4°C to collect the supernatant, which was cleared by 0.45um cellulose acetate filter. The filtrated supernatant was loaded Ni²⁺-NTA column and eluted from the column using a linear gradient of 1M imidazole. Fractions containing the Dpo4 were pooled and diluted with 3 volumes of buffer SA (50mM Tris, pH 7.5, 50 mM NaCl, 0.5 mM EDTA, 10% glycerol and 1 mM DTT), and then loaded onto SP column with a gradient of 1M NaCl. The Dpo4 fractions eluted the SP column were pooled and diluted with 3 volumes of buffer HA (20 mM HEPES, pH 7.5, 50 mM NaCl, 0.5 mM EDTA, 10% glycerol and 1 mM DTT), and then further purified from Heparin column with a gradient of 1M NaCl. Purified Dpo4 was concentrated to 20 mg/ml in storage buffer (20 mM HEPES at pH 7.5, 100 mM NaCl, 0.5 mM EDTA, 10% glycerol and 1 mM DTT) and aliquoted, and then flash frozen in liquid nitrogen to store at −80 °C. All these columns were purchased from GE Healthcare.

Crystallization, data collection, and refinement.

The template oligonucleotides for X-ray crystallographic studies, 5′-CTAAC(X)GAATCCTTCCCCC-3’, X = 2’-F-N7-MeG) were custom-synthesized by Midland Certified Reagent Co. (Midland, TX, USA). The primer, 5′-GGGGGAAGGATTC-3’ was synthesized by Integrated DNA Technologies (Coralville, IA, USA). The template and primer oligonucleotides mixed at a 1:1 molar ratio were annealed in a buffer containing 10 mM Tris (pH 8.0) and 50 mM NaCl by heating for 2 min at 95°C, incubating 10 min at 55°C and slowly cooling to room temperature. The crystal of Dpo4-N7MedG: dCTP ternary complex was grown at the similar conditions described previously (51). Briefly, the storage buffer for Dpo4 was exchanged into the crystallization buffer (20 mM HEPES (pH 7.0), 5 mM Ca(OAc)₂, 85 mM NaCl, 0.25 mM EDTA, and 1 mM DTT ) using an Amicon Ultra-15 centrifugal filter (10-kDa cutoff, Millipore). The ternary complex was prepared by combining 150 μM Dpo4 in the crystallization buffer, 180 μM DNA, and 1 mM dCTP (final concentrations). The ternary crystal was obtained at room temperature by using the hanging drop vapor-diffusion method after mixing 1.5 μL of the ternary complex and 1.5 μL of a reservoir buffer containing 12-18% polyethylene glycol 3350 (PEG 3350), 100 mM HEPES (pH 6.6),100 mM Ca(OAc)₂, and 2.5% glycerol.

Diffraction data were collected on the crystals cryo-protected with a reservation buffer with 20% glycerol and 1mM CTP at 100 K using the beamline beamline 5.0.3 at the Advanced Light Source, Lawrence Berkeley National Laboratory. All diffraction data were processed using HKL 2000 (HKL research, Charlottesville, VA, USA). Structures were solved by molecular replacement using the Dpo4-dT:dATP structure (PDB code 2AGQ) as the search model (37). The model was built using COOT (52) and refined using Phenix software (53). MolProbity was used to make Ramachandran plots (54).

Processing of N7-Me-FapyG-containing duplex with E. coli Endonuclease IV (Endo IV).

The oligonucleotide duplex (20 nM) containing 5’-pTCGAGCTGTGGT(N7-Me-FapyG)GGGTTC-3’ and 5’-pTCGGGAACCCCACCACAGC-3’ was incubated with 32 nM E. coli Endo IV (New England Biolabs, MA) in 100 mM NaCl, 50 mM Tris-HCl (pH 7.9), 10 mM MgCl₂, and 1 mM DTT at 37°C for one hour. The enzymatic reaction was quenched by adding loading buffer (95% formamide, 5 mM EDTA, xylene cyanol, and bromophenol blue dyes), and resolved on a 20% urea PAGE gel. The extent of the incision was determined by calculating the band intensities using ImageJ software.

Figure 7. Mutagenic properties of 2’-F-N7-MeG and 2’-F-Me-FapyG in the bacterial MBM7070 cells.

(A) Construction of pSP189 plasmid with a single N7-MeG or Me-FapyG lesion. The lesion-containing duplex insert was ligated into two AvaI restriction sites and the resulting plasmid was transformed into and replicated in MBM7070 cells. (B) Representative example for the supF-based blue/white colony screening assay of 2’-F-N7-MeG and 2’-F-Me-FapyG. (C) Deletion mutation caused by Me-FapyG.

Data Availability Statement:

The atomic coordinate of Dpo4-N7MeG:dCTP complex has been deposited in the Protein Data Bank with the accession codes of 7KLE (htpps://doi.org/10.2210/pdb7KLE/pdb)(36).