DEB-FAPy-dG Adducts of 1,3-Butadiene: Synthesis, Structural Characterization, and Formation in 1,2,3,4-Diepoxybutane Treated DNA

Abstract

Metabolic activation of the human carcinogen 1,3-butadiene (BD) by cytochrome 450 monooxygenases gives rise to a genotoxic diepoxide, 1,2,3,4-diepoxybutane (DEB). This reactive electrophile alkylates guanine bases in DNA to produce N7-(2-hydroxy-3,4-epoxy-1-yl)-dG (N7-HEB-dG) adducts. Because of the positive charge at the N7 position of the purine heterocycle, N7-HEB-dG adducts are inherently unstable and can undergo spontaneous depurination or base-catalyzed imidazole ring opening to give N⁶-[2-deoxy-D-erythro-pentofuranosyl]-2,6-diamino-3,4-dihydro-4-oxo-5-N-1-(oxiran-2-yl)propan-1-ol-formamidopyrimidine (DEB-FAPy-dG) adducts. Here we report the first synthesis and structural characterization of DEB-FAPy-dG adducts. Authentic standards of DEB-FAPy-dG and its ¹⁵N₃-labeled analogue were used for the development of a quantitative nanoLC-ESI⁺-HRMS/MS method, allowing for adduct detection in DEB-treated calf thymus DNA. DEB-FAPy-dG formation in DNA was dependent on DEB concentration and pH, with higher numbers observed under alkaline conditions.

Graphical Abstract

Formamidopyrimidnes (FAPy) are highly mutagenic DNA lesions. Here we describe the first synthesis and complete structural characterization of a diepoxybutane (DEB) derived formamidopyrimidine of dG (DEB-FAPy-dG). An isotope dilution nanoLC-ESI⁺-HRMS/MS method was developed and DEB-FAPy-dG adducts were detected in DEB treated DNA.

1,3-Butadiene (BD) is a known human carcinogen commonly found in automobile exhaust^[1] and wood burning smoke,^[2] and is widely used in the rubber and plastic industries.^[3] BD is one of the most abundant carcinogens present in cigarette smoke.^[4] It induces leukemia and lymphoma, as well as tumors of the lung, ovary, liver, and mammary gland, in laboratory animals.^[5] Furthermore, BD causes chromosomal abnormalities and increases the risk of lymphatic and hematopoietic cancer in occupationally exposed workers.^{[3, 6]}

BD is metabolically activated by cytochrome 450 monooxygenases CYP2E1 and 2A6 to reactive electrophile 3,4-epoxy-1-butene (EB)^[7]. EB can be further epoxidized to form 1,2,3,4-diepoxybutane (DEB) or hydrolyzed by epoxide hydrolase (EH) to form 1-butene-3,−4-diol (EB-diol)^[7]. 1,2-dihydroxy-3,4-epoxybutane (EBD) is formed both through the epoxidation of EB-diol and the hydrolysis of DEB (Scheme 1).^{[7b, 8]} EB, DEB, and EBD are genotoxic and can induce point mutations and deletions; among the three, DEB is by far the most mutagenic.^[9] BD-derived epoxides can enter cell nuclei and react with genomic DNA to form covalent DNA adducts at guanine and adenine bases.^[10] BD-DNA adducts can cause polymerase errors during DNA replication, potentially leading to mutations and cancer.^[11]

Scheme 1.

Metabolism of 1,3-butadiene to reactive species.

The N7-position of guanine in DNA is the favored reactive site for simple alkylating agents such as BD-derived epoxides.^[12] N7-guanine adducts retain the ability to pair with C,^{[12a, 12c, 12d]} but are hydrolytically labile due to the presence of a positive charge on the alkylated base.^{[12a, 12b, 13]} These adducts undergo two competing reactions: spontaneous depurination to release N7-alkylguanine bases (favored under acidic conditions) and imidazole ring opening to give the formamidopyrimidine (FAPy) lesions (preferred under basic conditions, Scheme 2).^[12d] Imidazole ring scission and alkyl-FAPy formation alters the molecular shape and base pairing preferences of the parent nucleobase, leading to mispairing during DNA replication.^[14] Alkyl-FAPy lesions are relatively stable and can accumulate in cells over time^{[12b, 12d, 15]} and have been shown to play a major role in genotoxicity of the liver carcinogen aflatoxin B1^[16]. Structurally related unsubstituted FAPy adducts can be formed by a free radical mechanism upon exposure to reactive oxygen species (ROS).^{[15, 17]}

Scheme 2.

Mechanism of apurinic site and FAPy lesion formation from dG

Previously, we reported the synthesis and structural characterization of N⁶-(2-deoxy-D-erythro-pentofuranosyl)-2,6-diamino-3,4-dihydro-4-oxo-5-N-(2-hydroxy-3-buten-1-yl)-formamidopyrimidine (EB-FAPy-dG) adducts in DNA exposed to butadiene monoepoxide (EB, Scheme 1).^[12c] EB-FAPy-dG lesions were detected in EB-treated calf thymus DNA and in murine cells in culture (MEF), with higher adduct numbers observed in cells deficient in NEIL1, a DNA repair gene involved in base excision repair.^[18] In the present work, we describe the synthesis and structural characterization of the corresponding adducts of butadiene derived diepoxide (DEB), N⁶-[2-deoxy-D-erythro-pentofuranosyl]-2,6-diamino-3,4-dihydro-4-oxo-5-N-1-(oxiran-2-yl)propan-1-ol-formamidopyrimidine (DEB-FAPy-dG). DEB-FAPy-dG adducts were detected in DEB-treated calf thymus DNA (CT DNA) using isotope dilution liquid chromatography-high resolution tandem mass spectrometry. To our knowledge, this is the first report of DEB-derived FAPy adducts, which are hypothesized to play a role in genotoxicity of DEB.

Chemical synthesis of DEB-FAPy-dG was achieved via direct alkylation of 2’-deoxyguanosine (dG, 1 in Scheme 3) with commercial 1,2,3,4-diepoxybutane (5 equiv.) in Tris-HCl buffer (10 mM, pH 7.2) overnight. The reaction mixture was separated by HPLC to obtain N7-(2-hydroxy-3,4-epoxy-1-yl)-dG (N7-HEB-dG) (compound 3, est. yield, 10%). HPLC separation was achieved via preparative HPLC using a Sunfire Prep C18 OBD column (Waters, 19 × 250 mm, 5 μm) and a gradient of acetonitrile/water (for details, see Supporting Information, Figure S1). Structural characterization data for compound 3 is given in Supporting Information (Figures S3-S8). Next, compound 3 was treated with 1 M NaOH at room temperature for 1 h and immediately neutralized with 20% AcOH, followed by HPLC purification to give compound 4 (est. yield, 70%) (Scheme 3). DEB-FAPy-dG was isolated by semipreparative HPLC on a Synergi Hydro RP column (Phenomenex, 4.6× 250 mm, 4 μm) using a gradient of acetonitrile in water (for details, see Supporting Information, Figure S2).

Scheme 3.

Synthesis of DEB-FAPy-dG nucleoside from 2’-deoxyguanosine.

It should be noted that under our reaction conditions, the 3,4- epoxy ring of DEB-FAPy-dG remained intact. The epoxide ring could be opened upon extended treatment of adduct 3 with 1 M NaOH to give trihydroxybutyl-FAPy-dG (compound 6) (Figure S3).

Complete structural characterization of DEB-FAPy-dG 4 was achieved via various spectroscopic techniques. DEB-FAPy-dG was initially characterized by UV spectrophotometry. Imidazole ring opening of N7-alkyl-dG to give alkyl-FAPy generates signatory UV-Vis spectra for these molecules.^{[12c, 12d, 19]} At neutral pH, the λ_max of DEB-FAPy-dG was determined to be at 272 nm (Figure 1). Under acidic conditions (pH 1.0), the λ_max of the nucleoside shifted to 271 nm. Conversely, when solution pH was adjusted to higher pH (12.0) using 1 M NaOH, λ_max moved to 265 nm (Figure 1). Overall, the UV absorption profile of this lesion matched the published spectra of closely related analogue EB-FAPy-dG and other alkyl-FAPy-dG adducts.^{[12c, 12d, 19]}

Figure 1.

UV absorption spectra of DEB-FAPy-dG in 0.1N HCl (pH: 1.0), λ_max 271 nm; water (pH: 7.0), λ_max 272 nm; and 0.1N NaOH (pH: 12.0). λ_max 265 nm.

Mass spectrometric characterization of synthetic DEB-FAPy-dG was performed on a Q Exactive Orbitrap mass spectrophotometer (Figure 2). The observed molecular ion signal of DEB-FAPy-dG at m/z 372.1509 ([M+H]⁺) is in agreement with the theoretical value for protonated DEB-FAPy-dG (372.1514). MS/MS fragmentation of DEB-FAPy-dG (m/z 372.1509) was characterized by the neutral loss of deoxyribose (m/z 372.1509 [M+H]⁺ → 256.1037 [M−dR+H]⁺) and a loss of 2’-deoxyribose and water (m/z 372.1509 [M+H]⁺ → 238.0932 [M+H−dR−H₂O]⁺) (Figure 2A).

Figure 2.

ESI⁺-MS/MS of DEB-FAPy-dG nucleoside (A) and its ¹⁵N₃-labeled analogue (B).

¹⁵N1, ¹⁵N², ¹⁵N3- N7-DEB-FAPy-dG was synthesized according to the above procedure starting with 25 mg of ¹⁵N1, ¹⁵N², ¹⁵N3-dG. ¹⁵N1, ¹⁵N², ¹⁵N3, N7-(2-hydroxy-3,4-epoxy-1-yl)-dG intermediate exhibited a molecular ion at m/z 375.2349, which is consistent with the theoretical value (375.1425, [M+H]⁺). ¹⁵N₃-N7-DEB-FAPy-dG was treated with base to obtain the corresponding DEB-FAPy adduct. MS/MS fragmentation of ¹⁵N₃-DEB-FAPy-dG (m/z 375.2349, [M+H]⁺) was characterized by two major fragments at 241.0845 [M+H−dR−H₂O]⁺ and 259.0952 [M+H − dR]⁺ (Figure 2B). This isotopically labelled analogue was subsequently used as an internal standard for quantitative mass spectrometry analyses (see below).

DEB-FAPy-dG was further characterized by ¹H-NMR,¹³C-NMR, 135-DEPT, and ¹H−¹³C HSQC (Figure 3 and S10-S12). ¹H−NMR signals at 8.20–8.10 and 7.95–7.85 ppm confirmed the presence of the formamido group (Figure 3). The complexity of the spectra is indicative of the presence of at least six different isomers of DEB-FAPy-dG. For the synthesis of DEB-FAPy-dG nucleoside, commercially available racemic DEB was used. As a result, the DEB-FAPy-dG adducts were formed as a mixture of stereoisomers. Additionally, the formyl N5-C7’ (N-CO bond) bond is characterized by restricted rotation, giving rise to stable rotamers.^{[12d, 20]} Finally, several E/Z formamide isomers of alkyl FAPy are possible as a result of hindered rotation around C5-N⁵ bond (Scheme 4).^[12d] Other N5-substituted-FAPy-adducts such as aflatoxin B1 FAPy adducts (AFB1-FAPy-dG) and Me-FAPy-dG adducts exist as a mixture of rotamers due to the hindered rotation around the C5-N⁵ as well as formamide (N5-C7’) bond. Similar isomers are possible for DEB-FAPy-dG (Scheme 4).

Figure 3.

¹H-NMR spectrum of DEB-FAPy-dG nucleoside (4) in D₂O.

Scheme 4.

Proposed structures of E/Z formamide isomers of DEB-FAPy-dG.

The multiple signals at 5.60–5.15 ppm (due to all the above discussed isomers) in the ¹H NMR spectrum of DEB-FAPy-dG (Figure 3) were assigned to C1’-H of the deoxyribose, while the signals in the range of 2.15–1.80 ppm correspond to the 2’ proton of the deoxyribose moiety. These signals were attributed to CH₂’-α and CH₂’-β protons. The ¹H-NMR signals in the range 4.30–3.20 ppm correspond to the sidechain hydrogens of C7 to C10-H (Figure 3). These ¹H-NMR assignments are consistent with the ¹³C-HSQC spectra (Figure S12). Signals at 168 ppm, 87 ppm, 84 ppm, and 39 ppm in ¹H-NMR-¹³C-HSQC correspond to formamide, C1’, side chain, and CH₂’-α and β carbons, respectively (Supporting Information, Figures S10-S12). The presence of multiple signals in ¹H-NMR and ¹³C-HSQC spectra reflects the presence of multiple stereoisomers and rotamers of DEB-FAPy-dG.

Collectively, the mass spectroscopy, NMR, and UV, and spectroscopy results for synthetic DEB-FAPy-dG are consistent with the structure of N⁶-[2-deoxy-D-erythro-pentofuranosyl]-2,6-diamino-3,4-dihydro-4-oxo-5-N-1-(oxiran-2-yl)propan-1-ol-formamidopyrimidine (DEB-FAPy-dG) shown in Scheme 2. Although the molecule was isolated as a mixture of isomers which were difficult or impossible to separate, this represents the structural complexity of DEB-FAPy-dG adducts formed in cells. All three stereoisomers of DEB (S,S, R,R and meso) are produced metabolically upon oxidation of 1,3-butadiene by liver enzymes such as P450 2E1.^{[8, 21]}

To allow for DEB-FAPy-dG adduct detection and quantification in DNA, a quantitative isotope dilution nanoLC-ESI⁺-MS/MS method was developed. ¹⁵N₃-DEB-FAPy-dG was used as an internal standard. We chose nanoLC for chromatographic separation since it allows for optimal sensitivity in mass spectrometry-based approaches for DNA adduct detection.^{[12c, 22]} nanoLC-ESI⁺-HR MS/MS analyses were conducted on an Orbitrap QExactive mass spectrometer interfaced with a Dionex Ultimate 3000 nano LC system (Thermo Fisher). Instrument resolution was set for 70,000 to allow for excellent selectivity. Quantitative analysis of DEB-FAPy-dG was conducted in the parallel reaction monitoring mode using mass transitions corresponding to the loss of deoxyribose sugar (m/z 372.1509 [M+H]⁺ → 256.1037 [M−dR+H]⁺) and a loss of 2’-deoxyribose and water (m/z 372.1509 [M+H]⁺ → 238.0923 [M+H−dR−H2O]⁺) (Figure 4). The ¹⁵N₃-labeled internal standard was monitored using the corresponding signals at m/z 375.2349 [M+H]⁺ → 241.0845 [M+H−dR−H₂O]⁺ and m/z 375.2349 [M+H]⁺ → 259.0952 [M+H− dR]⁺ (Figure 4). MS/MS transition corresponding to the neutral loss of both deoxyribose and water provided greater sensitivity and was thus chosen for quantitative analyses, while the second transition was used for confirmation purposes.

Figure 4.

Representative nanoLC-ESI⁺-HRMS/MS traces of DEB-FAPy-dG in CT DNA treated with 50 μM DEB.

To determine whether DEB-FAPy-dG adducts can form in DNA, calf thymus DNA was treated with increasing amounts of DEB (10 μM to 10 mM). High concentrations were chosen in order to facilitate adduct detection. DEB-treated DNA was subjected to enzymatic hydrolysis to 2’-nucleosides, followed by enrichment by solid phase extraction on Hypercarb Hypersep cartridges (Thermo Scientific). We investigated DEB-FAPy-dG retention on multiple SPE cartridges (Isolute ENV⁺ (40 mg/1 mL, Biotage, Charlotte, NC), Strata X polymeric C18 (30 mg/1 mL, Phenomenex, Torrance, CA) and Hypercarb Hypersep (100 mg/1 mL, Thermo Scientific, Waltham, MA) and found that Hypercarb Hypersep was the only stationary phase that retained DEB-FAPy-dG (SPE recovery 58%). SPE eluates containing DEB-FAPy-dG and ¹⁵N₃-DEB-FAPy-dG were concentrated under vacuum and reconstituted in water for nanoLC-ESI⁺-HRMS/MS analysis.

During nanoLC-ESI⁺-HRMS/MS method development for DEB-FAPy-dG, several HPLC stationary phases including Hypercarb (Thermo Scientific, Waltham, MA), Synergi Polar RP (Phenomenex, Torrance, CA) and Synergi Hydro RP (Phenomenex, Torrance, CA) were tested with a variety of solvent systems (5 mM ammonium formate, 0.1% formic acid, and 0.05% formic acid with acetonitrile) and various solvent gradients. The best HPLC peak shape and retention were achieved on a Synergi Hydro RP column with a gradient of 0.05% formic acid and acetonitrile. With this method, DEB-FAPy-dG eluted as a complex peak at 14.87 min (Figure 4). The quantitative nanoLC-ESI⁺-HRMS/MS method for DEB-FAPy-dG was validated by spiking known amounts of analyte (12.5–500 fmol) and internal standard (350 fmol) into calf thymus DNA. A linear correlation was observed between the spiked and calculated amounts of analyte, with an R² value of 0.999 (Supporting Information, Figure S13).

DEB-FAPy-dG formation was investigated at varying pH. Aliquots of treated CT DNA were taken and either analyzed directly (physiological pH) or treated with NaOH (pH = 12) to induce imidazole ring opening. Samples treated at high pH to induce ring opening exhibited a concentration dependent formation of DEB-FAPy-dG, with adduct numbers ranging from 0.7 to 428.0 DEB-FAPy-dG adducts per 10⁶ nucleotides (Figure 5A). However, DEB-FAPy-dG lesions were also observed at physiological pH, with amounts approximately 1300-lower than at pH 12 (Figure 5B).

Figure 5.

Dose dependent formation of DEB-FAPy-dG at pH 12. (A) and pH 7.4 (B).

In order to establish pH dependence for DEB-FAPy-dG adduct formation, CT DNA was treated with 5 mM DEB at 37 °C for 24 h and subsequently incubated at pH 7.5, 10, 11, and 12 at room temperature for 1 h. DEB-FAPy-dG adducts were observed in all samples, although the amounts were much greater in samples incubated at higher pH (pH 12) (Figure 6). This indicates that while DEB-FAPy-dG adducts can form at lower pH, the higher abundance only at pH 12 is consistent with their preferred formation under basic conditions.

Figure 6.

pH-dependent formation of DEB-FAPy-dG in CT DNA treated with 5 mM DEB.

In terms of their biological effects, ring open N⁵-alkyl-FAPy adducts are expected to hinder DNA synthesis and/or to mispair with purines during DNA replication.^{[14a, 23]} For instance, Me-FAPy-dG adducts block DNA polymerases α, β, and hPol δ/PCNA.^[14d] However, Me-FAPy-dG adducts were bypassed by TLS polymerases η, κ, hRev1/Pol ζ and polymerases η, κ, leading to misinsertion and −1 deletion products.^[14d] AFB1-FAPy adducts have been shown to be a potent block to DNA synthesis, even when DNA polymerase of lowered replication fidelity was used (MucAB).^[16] AFB1-FAPy-G adducts caused G → T transversions along with G → T mutations in the ras oncogene, making them the ultimate lesions responsible for mutagenesis and genotoxicity of aflatoxin.^[24] In light of these findings, we expect the newly discovered DEB-FAPy-dG lesions will exhibit mutagenetic and cytotoxic properties. Future research in our laboratory is directed towards investigating these lesions for their biological significance.

In summary, we have successfully synthesized and structurally characterized novel DEB-FAPy-dG and ¹⁵N₃-DEB-FAPy-dG adducts. A sensitive isotope dilution nanoLC-ESI⁺-MS/MS methodology was developed for the detection and quantitation of DEB-FAPy-dG lesions in DNA. We investigated the formation of DEB-FAPy-dG lesions in CT-DNA treated with DEB, the most genotoxic and carcinogenic metabolite of 1,3-butadiene, and observed concentration dependent adduct formation, which was enhanced at high pH conditions. We therefore conclude that DEB-FAPy-dG adducts can form under physiological conditions upon exposure to DEB or its metabolic precursor 1,3-butadiene and are likely to contribute to genotoxicity and mutagenicity of BD. Further studies are underway to detect these lesions in cells treated with DEB and to establish their effect on DNA replication, transcription, and repair.